From cdee45bba6ddc2f2cf2a4467ad34eac3b4deffa6 Mon Sep 17 00:00:00 2001
From: Miles Macklin <mmacklin@nvidia.com>
Date: Thu, 23 Nov 2023 09:04:34 +1300
Subject: [PATCH 01/50] Add content-type for PyPi setup

---
 pyproject.toml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pyproject.toml b/pyproject.toml
index 14ff5d984..6190136f0 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -37,7 +37,7 @@ find = {}
 
 [tool.setuptools.dynamic]
 version = {attr = "warp.config.version"}
-readme = {file = ["README.md"]}
+readme = {file = ["README.md"], content-type = "text/markdown"}
 
 [tool.black]
 line-length = 120

From cd77601f226537c14393decbf4fd48292bcf30e8 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Wed, 22 Nov 2023 15:19:59 +1300
Subject: [PATCH 02/50] Support passing the return value to adjoints

---
 warp/builtins.py            | 20 +++++++++++++++++++-
 warp/codegen.py             |  6 ++++--
 warp/context.py             |  5 +++++
 warp/native/builtin.h       | 26 ++++++++++++--------------
 warp/native/intersect.h     |  8 ++++----
 warp/native/intersect_adj.h | 16 ++++++++--------
 warp/native/mat.h           |  8 ++++----
 warp/native/quat.h          | 13 +++++++++----
 warp/native/vec.h           | 17 +++++++++--------
 9 files changed, 74 insertions(+), 45 deletions(-)

diff --git a/warp/builtins.py b/warp/builtins.py
index 5df6e0de2..5869800b1 100644
--- a/warp/builtins.py
+++ b/warp/builtins.py
@@ -116,6 +116,7 @@ def fn(arg_types, kwds, _):
     value_func=sametype_value_func(Float),
     doc="Return the square root of ``x``, where ``x`` is positive.",
     group="Scalar Math",
+    require_original_output_arg=True,
 )
 add_builtin(
     "cbrt",
@@ -123,6 +124,7 @@ def fn(arg_types, kwds, _):
     value_func=sametype_value_func(Float),
     doc="Return the cube root of ``x``.",
     group="Scalar Math",
+    require_original_output_arg=True,
 )
 add_builtin(
     "tan",
@@ -165,6 +167,7 @@ def fn(arg_types, kwds, _):
     value_func=sametype_value_func(Float),
     doc="Return the tanh of ``x``.",
     group="Scalar Math",
+    require_original_output_arg=True,
 )
 add_builtin(
     "degrees",
@@ -208,6 +211,7 @@ def fn(arg_types, kwds, _):
     value_func=sametype_value_func(Float),
     doc="Return the value of the exponential function :math:`e^x`.",
     group="Scalar Math",
+    require_original_output_arg=True,
 )
 add_builtin(
     "pow",
@@ -215,6 +219,7 @@ def fn(arg_types, kwds, _):
     value_func=sametype_value_func(Float),
     doc="Return the result of ``x`` raised to power of ``y``.",
     group="Scalar Math",
+    require_original_output_arg=True,
 )
 
 add_builtin(
@@ -405,6 +410,7 @@ def value_func_outer(arg_types, kwds, _):
     value_func=sametype_scalar_value_func,
     group="Vector Math",
     doc="Compute the length of a vector ``x``.",
+    require_original_output_arg=True,
 )
 add_builtin(
     "length",
@@ -412,6 +418,7 @@ def value_func_outer(arg_types, kwds, _):
     value_func=sametype_scalar_value_func,
     group="Vector Math",
     doc="Compute the length of a quaternion ``x``.",
+    require_original_output_arg=True,
 )
 add_builtin(
     "length_sq",
@@ -433,6 +440,7 @@ def value_func_outer(arg_types, kwds, _):
     value_func=sametype_value_func(vector(length=Any, dtype=Scalar)),
     group="Vector Math",
     doc="Compute the normalized value of ``x``. If ``length(x)`` is 0 then the zero vector is returned.",
+    require_original_output_arg=True,
 )
 add_builtin(
     "normalize",
@@ -465,6 +473,7 @@ def value_func_mat_inv(arg_types, kwds, _):
     value_func=value_func_mat_inv,
     group="Vector Math",
     doc="Return the inverse of a 2x2 matrix ``m``.",
+    require_original_output_arg=True,
 )
 
 add_builtin(
@@ -473,6 +482,7 @@ def value_func_mat_inv(arg_types, kwds, _):
     value_func=value_func_mat_inv,
     group="Vector Math",
     doc="Return the inverse of a 3x3 matrix ``m``.",
+    require_original_output_arg=True,
 )
 
 add_builtin(
@@ -481,6 +491,7 @@ def value_func_mat_inv(arg_types, kwds, _):
     value_func=value_func_mat_inv,
     group="Vector Math",
     doc="Return the inverse of a 4x4 matrix ``m``.",
+    require_original_output_arg=True,
 )
 
 
@@ -580,6 +591,7 @@ def value_func_get_diag(arg_types, kwds, _):
     value_func=sametype_value_func(vector(length=Any, dtype=Scalar)),
     group="Vector Math",
     doc="Component-wise division of two 2D vectors.",
+    require_original_output_arg=True,
 )
 
 add_builtin(
@@ -595,6 +607,7 @@ def value_func_get_diag(arg_types, kwds, _):
     value_func=sametype_value_func(matrix(shape=(Any, Any), dtype=Scalar)),
     group="Vector Math",
     doc="Component-wise division of two 2D vectors.",
+    require_original_output_arg=True,
 )
 
 
@@ -1082,6 +1095,7 @@ def quat_identity_value_func(arg_types, kwds, templates):
     value_func=lambda arg_types, kwds, _: quaternion(dtype=infer_scalar_type(arg_types)),
     group="Quaternion Math",
     doc="Linearly interpolate between two quaternions.",
+    require_original_output_arg=True,
 )
 add_builtin(
     "quat_to_matrix",
@@ -3065,7 +3079,11 @@ def mul_matmat_value_func(arg_types, kwds, _):
 )
 
 add_builtin(
-    "div", input_types={"x": Scalar, "y": Scalar}, value_func=sametype_value_func(Scalar), doc="", group="Operators"
+    "div",
+    input_types={"x": Scalar, "y": Scalar},
+    value_func=sametype_value_func(Scalar),
+    doc="", group="Operators",
+    require_original_output_arg=True,
 )
 add_builtin(
     "div",
diff --git a/warp/codegen.py b/warp/codegen.py
index 2e6a0eaf0..85184c671 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -683,10 +683,11 @@ def format_reverse_call_args(
         args_out,
         use_initializer_list,
         has_output_args=True,
+        require_original_output_arg=False,
     ):
         formatted_var = adj.format_args("var", args_var)
         formatted_out = []
-        if has_output_args and len(args_out) > 1:
+        if has_output_args and (require_original_output_arg or len(args_out) > 1):
             formatted_out = adj.format_args("var", args_out)
         formatted_var_adj = adj.format_args(
             "&adj" if use_initializer_list else "adj",
@@ -966,13 +967,14 @@ def add_call(adj, func, args, min_outputs=None, templates=[], kwds=None):
             adj.add_forward(forward_call, replay=replay_call)
 
         if not func.missing_grad and len(args):
-            reverse_has_output_args = len(output_list) > 1 and func.custom_grad_func is None
+            reverse_has_output_args = (func.require_original_output_arg or len(output_list) > 1) and func.custom_grad_func is None
             arg_str = adj.format_reverse_call_args(
                 args_var,
                 args,
                 output_list,
                 use_initializer_list,
                 has_output_args=reverse_has_output_args,
+                require_original_output_arg=func.require_original_output_arg,
             )
             if arg_str is not None:
                 reverse_call = f"{func.namespace}adj_{func.native_func}({arg_str});"
diff --git a/warp/context.py b/warp/context.py
index e0422b550..863067717 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -79,6 +79,7 @@ def __init__(
         overloaded_annotations=None,
         code_transformers=[],
         skip_adding_overload=False,
+        require_original_output_arg=False,
     ):
         self.func = func  # points to Python function decorated with @wp.func, may be None for builtins
         self.key = key
@@ -97,6 +98,7 @@ def __init__(
         self.native_snippet = native_snippet
         self.adj_native_snippet = adj_native_snippet
         self.custom_grad_func = None
+        self.require_original_output_arg = require_original_output_arg
 
         if initializer_list_func is None:
             self.initializer_list_func = lambda x, y: False
@@ -841,6 +843,7 @@ def add_builtin(
     missing_grad=False,
     native_func=None,
     defaults=None,
+    require_original_output_arg=False,
 ):
     # wrap simple single-type functions with a value_func()
     if value_func is None:
@@ -965,6 +968,7 @@ def typelist(param):
                     hidden=True,
                     skip_replay=skip_replay,
                     missing_grad=missing_grad,
+                    require_original_output_arg=require_original_output_arg,
                 )
 
     func = Function(
@@ -985,6 +989,7 @@ def typelist(param):
         generic=generic,
         native_func=native_func,
         defaults=defaults,
+        require_original_output_arg=require_original_output_arg,
     )
 
     if key in builtin_functions:
diff --git a/warp/native/builtin.h b/warp/native/builtin.h
index 4011312d9..ad33cf575 100644
--- a/warp/native/builtin.h
+++ b/warp/native/builtin.h
@@ -295,7 +295,7 @@ inline CUDA_CALLABLE T rshift(T a, T b) { return a>>b; } \
 inline CUDA_CALLABLE T invert(T x) { return ~x; } \
 inline CUDA_CALLABLE bool isfinite(T x) { return true; } \
 inline CUDA_CALLABLE void adj_mul(T a, T b, T& adj_a, T& adj_b, T adj_ret) { } \
-inline CUDA_CALLABLE void adj_div(T a, T b, T& adj_a, T& adj_b, T adj_ret) { } \
+inline CUDA_CALLABLE void adj_div(T a, T b, T ret, T& adj_a, T& adj_b, T adj_ret) { } \
 inline CUDA_CALLABLE void adj_add(T a, T b, T& adj_a, T& adj_b, T adj_ret) { } \
 inline CUDA_CALLABLE void adj_sub(T a, T b, T& adj_a, T& adj_b, T adj_ret) { } \
 inline CUDA_CALLABLE void adj_mod(T a, T b, T& adj_a, T& adj_b, T adj_ret) { } \
@@ -443,10 +443,10 @@ inline CUDA_CALLABLE T div(T a, T b)\
     })\
     return a/b;\
 }\
-inline CUDA_CALLABLE void adj_div(T a, T b, T& adj_a, T& adj_b, T adj_ret)\
+inline CUDA_CALLABLE void adj_div(T a, T b, T ret, T& adj_a, T& adj_b, T adj_ret)\
 {\
     adj_a += adj_ret/b;\
-    adj_b -= adj_ret*(a/b)/b;\
+    adj_b -= adj_ret*(ret)/b;\
     DO_IF_FPCHECK(\
     if (!isfinite(adj_a) || !isfinite(adj_b))\
     {\
@@ -859,11 +859,11 @@ inline CUDA_CALLABLE void adj_log10(T a, T& adj_a, T adj_ret)\
         assert(0);\
     })\
 }\
-inline CUDA_CALLABLE void adj_exp(T a, T& adj_a, T adj_ret) { adj_a += exp(a)*adj_ret; }\
-inline CUDA_CALLABLE void adj_pow(T a, T b, T& adj_a, T& adj_b, T adj_ret)\
+inline CUDA_CALLABLE void adj_exp(T a, T ret, T& adj_a, T adj_ret) { adj_a += ret*adj_ret; }\
+inline CUDA_CALLABLE void adj_pow(T a, T b, T ret, T& adj_a, T& adj_b, T adj_ret)\
 { \
     adj_a += b*pow(a, b-T(1))*adj_ret;\
-    adj_b += log(a)*pow(a, b)*adj_ret;\
+    adj_b += log(a)*ret*adj_ret;\
     DO_IF_FPCHECK(if (!isfinite(adj_a) || !isfinite(adj_b))\
     {\
         printf("%s:%d - adj_pow(%f, %f, %f, %f, %f)\n", __FILE__, __LINE__, float(a), float(b), float(adj_a), float(adj_b), float(adj_ret));\
@@ -962,24 +962,22 @@ inline CUDA_CALLABLE void adj_cosh(T x, T& adj_x, T adj_ret)\
 {\
     adj_x += sinh(x)*adj_ret;\
 }\
-inline CUDA_CALLABLE void adj_tanh(T x, T& adj_x, T adj_ret)\
+inline CUDA_CALLABLE void adj_tanh(T x, T ret, T& adj_x, T adj_ret)\
 {\
-    T tanh_x = tanh(x);\
-    adj_x += (T(1) - tanh_x*tanh_x)*adj_ret;\
+    adj_x += (T(1) - ret*ret)*adj_ret;\
 }\
-inline CUDA_CALLABLE void adj_sqrt(T x, T& adj_x, T adj_ret)\
+inline CUDA_CALLABLE void adj_sqrt(T x, T ret, T& adj_x, T adj_ret)\
 {\
-    adj_x += T(0.5)*(T(1)/sqrt(x))*adj_ret;\
+    adj_x += T(0.5)*(T(1)/ret)*adj_ret;\
     DO_IF_FPCHECK(if (!isfinite(adj_x))\
     {\
         printf("%s:%d - adj_sqrt(%f, %f, %f)\n", __FILE__, __LINE__, float(x), float(adj_x), float(adj_ret));\
         assert(0);\
     })\
 }\
-inline CUDA_CALLABLE void adj_cbrt(T x, T& adj_x, T adj_ret)\
+inline CUDA_CALLABLE void adj_cbrt(T x, T ret, T& adj_x, T adj_ret)\
 {\
-    T cbrt_x = cbrt(x);\
-    adj_x += (T(1)/T(3))*(T(1)/(cbrt_x*cbrt_x))*adj_ret;\
+    adj_x += (T(1)/T(3))*(T(1)/(ret*ret))*adj_ret;\
     DO_IF_FPCHECK(if (!isfinite(adj_x))\
     {\
         printf("%s:%d - adj_cbrt(%f, %f, %f)\n", __FILE__, __LINE__, float(x), float(adj_x), float(adj_ret));\
diff --git a/warp/native/intersect.h b/warp/native/intersect.h
index 7513772bd..e170dcd05 100644
--- a/warp/native/intersect.h
+++ b/warp/native/intersect.h
@@ -869,7 +869,7 @@ CUDA_CALLABLE inline void adj_closest_point_to_triangle(
     wp::adj_sub(var_9, var_71, adj_9, adj_71, adj_73);
     wp::adj_mul(var_21, var_70, adj_21, adj_70, adj_72);
     wp::adj_mul(var_41, var_70, adj_41, adj_70, adj_71);
-    wp::adj_div(var_9, var_69, adj_9, adj_69, adj_70);
+    wp::adj_div(var_9, var_69, var_70, adj_9, adj_69, adj_70);
     wp::adj_add(var_68, var_21, adj_68, adj_21, adj_69);
     wp::adj_add(var_53, var_41, adj_53, adj_41, adj_68);
     wp::adj_select(var_64, var_50, var_66, adj_64, adj_50, adj_66, adj_67);
@@ -881,7 +881,7 @@ CUDA_CALLABLE inline void adj_closest_point_to_triangle(
     }
     wp::adj_sub(var_32, var_33, adj_32, adj_33, adj_62);
     wp::adj_sub(var_13, var_12, adj_13, adj_12, adj_60);
-    wp::adj_div(var_54, var_57, adj_54, adj_57, adj_58);
+    wp::adj_div(var_54, var_57, var_58, adj_54, adj_57, adj_58);
     wp::adj_add(var_55, var_56, adj_55, adj_56, adj_57);
     wp::adj_sub(var_32, var_33, adj_32, adj_33, adj_56);
     wp::adj_sub(var_13, var_12, adj_13, adj_12, adj_55);
@@ -896,7 +896,7 @@ CUDA_CALLABLE inline void adj_closest_point_to_triangle(
     	wp::adj_vec2(var_48, var_5, adj_48, adj_5, adj_49);
     	wp::adj_sub(var_9, var_43, adj_9, adj_43, adj_48);
     }
-    wp::adj_div(var_4, var_42, adj_4, adj_42, adj_43);
+    wp::adj_div(var_4, var_42, var_43, adj_4, adj_42, adj_43);
     wp::adj_sub(var_4, var_33, adj_4, adj_33, adj_42);
     wp::adj_sub(var_39, var_40, adj_39, adj_40, adj_41);
     wp::adj_mul(var_3, var_33, adj_3, adj_33, adj_40);
@@ -917,7 +917,7 @@ CUDA_CALLABLE inline void adj_closest_point_to_triangle(
     	wp::adj_vec2(var_28, var_23, adj_28, adj_23, adj_29);
     	wp::adj_sub(var_9, var_23, adj_9, adj_23, adj_28);
     }
-    wp::adj_div(var_3, var_22, adj_3, adj_22, adj_23);
+    wp::adj_div(var_3, var_22, var_23, adj_3, adj_22, adj_23);
     wp::adj_sub(var_3, var_12, adj_3, adj_12, adj_22);
     wp::adj_sub(var_19, var_20, adj_19, adj_20, adj_21);
     wp::adj_mul(var_12, var_4, adj_12, adj_4, adj_20);
diff --git a/warp/native/intersect_adj.h b/warp/native/intersect_adj.h
index e64a9ad53..23c95af99 100644
--- a/warp/native/intersect_adj.h
+++ b/warp/native/intersect_adj.h
@@ -276,7 +276,7 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     label1:;
     adj_71 += adj_ret;
     wp::adj_vec3(var_61, var_62, var_70, adj_61, adj_62, adj_70, adj_71);
-    wp::adj_length(var_69, adj_69, adj_70);
+    wp::adj_length(var_69, var_70, adj_69, adj_70);
     wp::adj_sub(var_68, var_65, adj_68, adj_65, adj_69);
     wp::adj_add(var_p2, var_67, adj_p2, adj_67, adj_68);
     wp::adj_mul(var_66, var_62, adj_66, adj_62, adj_67);
@@ -297,7 +297,7 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     			wp::adj_select(var_51, var_49, var_54, adj_51, adj_49, adj_54, adj_55);
     			if (var_51) {
     				wp::adj_clamp(var_53, var_6, var_25, adj_53, adj_6, adj_25, adj_54);
-    				wp::adj_div(var_52, var_3, adj_52, adj_3, adj_53);
+    				wp::adj_div(var_52, var_3, var_53, adj_52, adj_3, adj_53);
     				wp::adj_sub(var_30, var_21, adj_30, adj_21, adj_52);
     			}
     		}
@@ -305,10 +305,10 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     		wp::adj_select(var_45, var_41, var_48, adj_45, adj_41, adj_48, adj_49);
     		if (var_45) {
     			wp::adj_clamp(var_47, var_6, var_25, adj_47, adj_6, adj_25, adj_48);
-    			wp::adj_div(var_46, var_3, adj_46, adj_3, adj_47);
+    			wp::adj_div(var_46, var_3, var_47, adj_46, adj_3, adj_47);
     			wp::adj_neg(var_21, adj_21, adj_46);
     		}
-    		wp::adj_div(var_43, var_4, adj_43, adj_4, adj_44);
+    		wp::adj_div(var_43, var_4, var_44, adj_43, adj_4, adj_44);
     		wp::adj_add(var_42, var_5, adj_42, adj_5, adj_43);
     		wp::adj_mul(var_30, var_41, adj_30, adj_41, adj_42);
     		wp::adj_select(var_34, var_6, var_40, adj_34, adj_6, adj_40, adj_41);
@@ -317,7 +317,7 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     		wp::adj_select(var_34, var_28, var_39, adj_34, adj_28, adj_39, adj_40);
     		if (var_34) {
     			wp::adj_clamp(var_38, var_6, var_25, adj_38, adj_6, adj_25, adj_39);
-    			wp::adj_div(var_37, var_33, adj_37, adj_33, adj_38);
+    			wp::adj_div(var_37, var_33, var_38, adj_37, adj_33, adj_38);
     			wp::adj_sub(var_35, var_36, adj_35, adj_36, adj_37);
     			wp::adj_mul(var_21, var_4, adj_21, adj_4, adj_36);
     			wp::adj_mul(var_30, var_5, adj_30, adj_5, adj_35);
@@ -332,7 +332,7 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     	if (var_22) {
     		wp::adj_cast_float(var_6, adj_6, adj_27);
     		wp::adj_clamp(var_24, var_6, var_25, adj_24, adj_6, adj_25, adj_26);
-    		wp::adj_div(var_23, var_3, adj_23, adj_3, adj_24);
+    		wp::adj_div(var_23, var_3, var_24, adj_23, adj_3, adj_24);
     		wp::adj_neg(var_21, adj_21, adj_23);
     	}
     	wp::adj_dot(var_0, var_2, adj_0, adj_2, adj_21);
@@ -341,7 +341,7 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     wp::adj_select(var_15, var_7, var_16, adj_15, adj_7, adj_16, adj_19);
     if (var_15) {
     	wp::adj_cast_float(var_17, adj_17, adj_18);
-    	wp::adj_div(var_5, var_4, adj_5, adj_4, adj_17);
+    	wp::adj_div(var_5, var_4, var_17, adj_5, adj_4, adj_17);
     	wp::adj_cast_float(var_6, adj_6, adj_16);
     }
     if (var_13) {
@@ -349,7 +349,7 @@ static CUDA_CALLABLE void adj_closest_point_edge_edge(vec3 var_p1,
     	adj_14 += adj_ret;
     	wp::adj_vec3(var_7, var_8, var_10, adj_7, adj_8, adj_10, adj_14);
     }
-    wp::adj_length(var_9, adj_9, adj_10);
+    wp::adj_length(var_9, var_10, adj_9, adj_10);
     wp::adj_sub(var_p2, var_p1, adj_p2, adj_p1, adj_9);
     wp::adj_cast_float(var_6, adj_6, adj_8);
     wp::adj_cast_float(var_6, adj_6, adj_7);
diff --git a/warp/native/mat.h b/warp/native/mat.h
index d5839f186..d3853f689 100644
--- a/warp/native/mat.h
+++ b/warp/native/mat.h
@@ -1105,10 +1105,10 @@ inline CUDA_CALLABLE void adj_determinant(const mat_t<4,4,Type>& m, mat_t<4,4,Ty
 }
 
 template<unsigned Rows, typename Type>
-inline CUDA_CALLABLE void adj_inverse(const mat_t<Rows,Rows,Type>& m, mat_t<Rows,Rows,Type>& adj_m, const mat_t<Rows,Rows,Type>& adj_ret)
+inline CUDA_CALLABLE void adj_inverse(const mat_t<Rows,Rows,Type>& m, mat_t<Rows,Rows,Type>& ret, mat_t<Rows,Rows,Type>& adj_m, const mat_t<Rows,Rows,Type>& adj_ret)
 {
     // todo: how to cache this from the forward pass?
-    mat_t<Rows,Rows,Type> invt = transpose(inverse(m));
+    mat_t<Rows,Rows,Type> invt = transpose(ret);
 
     // see https://people.maths.ox.ac.uk/gilesm/files/NA-08-01.pdf 2.2.3
     adj_m -= mul(mul(invt, adj_ret), invt);
@@ -1150,10 +1150,10 @@ inline CUDA_CALLABLE void adj_cw_mul(const mat_t<Rows,Cols,Type>& a, const mat_t
 }
 
 template<unsigned Rows, unsigned Cols, typename Type>
-inline CUDA_CALLABLE void adj_cw_div(const mat_t<Rows,Cols,Type>& a, const mat_t<Rows,Cols,Type>& b, mat_t<Rows,Cols,Type>& adj_a, mat_t<Rows,Cols,Type>& adj_b, const mat_t<Rows,Cols,Type>& adj_ret)
+inline CUDA_CALLABLE void adj_cw_div(const mat_t<Rows,Cols,Type>& a, const mat_t<Rows,Cols,Type>& b, mat_t<Rows,Cols,Type>& ret, mat_t<Rows,Cols,Type>& adj_a, mat_t<Rows,Cols,Type>& adj_b, const mat_t<Rows,Cols,Type>& adj_ret)
 {
   adj_a += cw_div(adj_ret, b);
-  adj_b -= cw_mul(adj_ret, cw_div(cw_div(a, b), b));
+  adj_b -= cw_mul(adj_ret, cw_div(ret, b));
 }
 
 // adjoint for the constant constructor:
diff --git a/warp/native/quat.h b/warp/native/quat.h
index f65ad321a..c554e243c 100644
--- a/warp/native/quat.h
+++ b/warp/native/quat.h
@@ -523,9 +523,14 @@ inline CUDA_CALLABLE void tensordot(const quat_t<Type>& a, const quat_t<Type>& b
 }
 
 template<typename Type>
-inline CUDA_CALLABLE void adj_length(const quat_t<Type>& a, quat_t<Type>& adj_a, const Type adj_ret)
+inline CUDA_CALLABLE void adj_length(const quat_t<Type>& a, Type ret, quat_t<Type>& adj_a, const Type adj_ret)
 {
-    adj_a += normalize(a)*adj_ret;
+    if (ret > Type(kEps))
+    {
+        Type inv_l = Type(1)/ret;
+
+        adj_a += quat_t<Type>(a.x*inv_l, a.y*inv_l, a.z*inv_l, a.w*inv_l) * adj_ret;
+    }
 }
 
 template<typename Type>
@@ -677,7 +682,7 @@ inline CUDA_CALLABLE void adj_quat_rotate_inv(const quat_t<Type>& q, const vec_t
 }
 
 template<typename Type>
-inline CUDA_CALLABLE void adj_quat_slerp(const quat_t<Type>& q0, const quat_t<Type>& q1, Type t, quat_t<Type>& adj_q0, quat_t<Type>& adj_q1, Type& adj_t, const quat_t<Type>& adj_ret)
+inline CUDA_CALLABLE void adj_quat_slerp(const quat_t<Type>& q0, const quat_t<Type>& q1, Type t, quat_t<Type>& ret, quat_t<Type>& adj_q0, quat_t<Type>& adj_q1, Type& adj_t, const quat_t<Type>& adj_ret)
 {
     vec_t<3,Type> axis;
     Type angle;
@@ -688,7 +693,7 @@ inline CUDA_CALLABLE void adj_quat_slerp(const quat_t<Type>& q0, const quat_t<Ty
     angle = angle * 0.5;
     
     // adj_t
-    adj_t += dot(mul(quat_slerp(q0, q1, t), quat_t<Type>(angle*axis[0], angle*axis[1], angle*axis[2], Type(0))), adj_ret);
+    adj_t += dot(mul(ret, quat_t<Type>(angle*axis[0], angle*axis[1], angle*axis[2], Type(0))), adj_ret);
 
     // adj_q0
     quat_t<Type> q_inc_x_q0;
diff --git a/warp/native/vec.h b/warp/native/vec.h
index 287b6366e..9c70864cb 100644
--- a/warp/native/vec.h
+++ b/warp/native/vec.h
@@ -735,9 +735,9 @@ inline CUDA_CALLABLE void adj_div(vec_t<Length, Type> a, Type s, vec_t<Length, T
 }
 
 template<unsigned Length, typename Type>
-inline CUDA_CALLABLE void adj_cw_div(vec_t<Length, Type> a, vec_t<Length, Type> b, vec_t<Length, Type>& adj_a, vec_t<Length, Type>& adj_b, const vec_t<Length, Type>& adj_ret) {
+inline CUDA_CALLABLE void adj_cw_div(vec_t<Length, Type> a, vec_t<Length, Type> b, vec_t<Length, Type>& ret, vec_t<Length, Type>& adj_a, vec_t<Length, Type>& adj_b, const vec_t<Length, Type>& adj_ret) {
   adj_a += cw_div(adj_ret, b);
-  adj_b -= cw_mul(adj_ret, cw_div(cw_div(a, b), b));
+  adj_b -= cw_mul(adj_ret, cw_div(ret, b));
 }
 
 template<unsigned Length, typename Type>
@@ -850,9 +850,12 @@ inline CUDA_CALLABLE void adj_extract(const vec_t<Length, Type> & a, int idx, ve
 }
 
 template<unsigned Length, typename Type>
-inline CUDA_CALLABLE void adj_length(vec_t<Length, Type> a, vec_t<Length, Type>& adj_a, const Type adj_ret)
+inline CUDA_CALLABLE void adj_length(vec_t<Length, Type> a, Type ret, vec_t<Length, Type>& adj_a, const Type adj_ret)
 {
-    adj_a += normalize(a)*adj_ret;
+    if (ret > Type(kEps))
+    {
+        adj_a += div(a, ret) * adj_ret;
+    }
 
 #if FP_CHECK
     if (!isfinite(adj_a))
@@ -880,7 +883,7 @@ inline CUDA_CALLABLE void adj_length_sq(vec_t<Length, Type> a, vec_t<Length, Typ
 }
 
 template<unsigned Length, typename Type>
-inline CUDA_CALLABLE void adj_normalize(vec_t<Length, Type> a, vec_t<Length, Type>& adj_a, const vec_t<Length, Type>& adj_ret)
+inline CUDA_CALLABLE void adj_normalize(vec_t<Length, Type> a, vec_t<Length, Type>& ret, vec_t<Length, Type>& adj_a, const vec_t<Length, Type>& adj_ret)
 {
     Type d = length(a);
     
@@ -888,9 +891,7 @@ inline CUDA_CALLABLE void adj_normalize(vec_t<Length, Type> a, vec_t<Length, Typ
     {
         Type invd = Type(1.0f)/d;
 
-        vec_t<Length, Type> ahat = normalize(a);
-
-        adj_a += (adj_ret*invd - ahat*(dot(ahat, adj_ret))*invd);
+        adj_a += (adj_ret*invd - ret*(dot(ret, adj_ret))*invd);
 
 #if FP_CHECK
         if (!isfinite(adj_a))

From db417403c0dd2635cfa261e5363d531499f6d0e5 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Tue, 28 Nov 2023 09:54:39 +1300
Subject: [PATCH 03/50] Edit the message for accessing non-existing attrs

---
 warp/codegen.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index 77cf87d80..726d8bf49 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -1344,7 +1344,7 @@ def emit_Attribute(adj, node):
 
             if isinstance(aggregate, Var):
                 raise WarpCodegenAttributeError(
-                    f"Error, `{node.attr}` is not an attribute of '{aggregate.label}' ({type_repr(aggregate.type)})"
+                    f"Error, `{node.attr}` is not an attribute of '{node.value.id}' ({type_repr(aggregate.type)})"
                 )
             raise WarpCodegenAttributeError(f"Error, `{node.attr}` is not an attribute of '{aggregate}'")
 

From 4885b43c9a802f7b89bec2b0f28755632436d974 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Thu, 30 Nov 2023 12:11:40 -0500
Subject: [PATCH 04/50] Only fetch prebuilt Clang/LLVM libraries when needed

---
 build_llvm.py | 50 ++++++++++++++++++++++++++------------------------
 1 file changed, 26 insertions(+), 24 deletions(-)

diff --git a/build_llvm.py b/build_llvm.py
index fc839bad0..24148bc2e 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -15,31 +15,33 @@
 llvm_build_path = os.path.join(llvm_project_path, "out/build/")
 llvm_install_path = os.path.join(llvm_project_path, "out/install/")
 
+
 # Fetch prebuilt Clang/LLVM libraries
-if os.name == "nt":
-    packman = "tools\\packman\\packman.cmd"
-    packages = {"x86_64": "15.0.7-windows-x86_64-ptx-vs142"}
-else:
-    packman = "./tools/packman/packman"
-    if sys.platform == "darwin":
-        packages = {
-            "arm64": "15.0.7-darwin-aarch64-macos11",
-            "x86_64": "15.0.7-darwin-x86_64-macos11",
-        }
+def fetch_prebuilt_libraries():
+    if os.name == "nt":
+        packman = "tools\\packman\\packman.cmd"
+        packages = {"x86_64": "15.0.7-windows-x86_64-ptx-vs142"}
     else:
-        packages = {"x86_64": "15.0.7-linux-x86_64-ptx-gcc7.5-cxx11abi0"}
-
-for arch in packages:
-    subprocess.check_call(
-        [
-            packman,
-            "install",
-            "-l",
-            f"./_build/host-deps/llvm-project/release-{arch}",
-            "clang+llvm-warp",
-            packages[arch],
-        ]
-    )
+        packman = "./tools/packman/packman"
+        if sys.platform == "darwin":
+            packages = {
+                "arm64": "15.0.7-darwin-aarch64-macos11",
+                "x86_64": "15.0.7-darwin-x86_64-macos11",
+            }
+        else:
+            packages = {"x86_64": "15.0.7-linux-x86_64-ptx-gcc7.5-cxx11abi0"}
+
+    for arch in packages:
+        subprocess.check_call(
+            [
+                packman,
+                "install",
+                "-l",
+                f"./_build/host-deps/llvm-project/release-{arch}",
+                "clang+llvm-warp",
+                packages[arch],
+            ]
+        )
 
 
 def build_from_source_for_arch(args, arch):
@@ -299,7 +301,7 @@ def build_warp_clang_for_arch(args, lib_name, arch):
             libpath = os.path.join(install_path, "lib")
         else:
             # obtain Clang and LLVM libraries from packman
-            assert os.path.exists("_build/host-deps/llvm-project"), "run build.bat / build.sh"
+            fetch_prebuilt_libraries()
             libpath = os.path.join(base_path, f"_build/host-deps/llvm-project/release-{arch}/lib")
 
         for _, _, libs in os.walk(libpath):

From 18aff4d89a94d48ae2ddc976bbd2341e28ca4bc7 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Thu, 30 Nov 2023 14:59:26 -0500
Subject: [PATCH 05/50] Support overriding the LLVM project source directory
 path

If `--build_llvm` is specified we previously cloned the LLVM project
Git repository from https://github.com/llvm/llvm-project.git into
`external/llvm-project`.

This change adds an additional `--llvm_source_path` build argument for
specifying a path to an existing LLVM project repo. Note that currently
it needs to be the `llvmorg-15.0.7` tag revision.
---
 build_lib.py  |  4 ++++
 build_llvm.py | 46 +++++++++++++++++++++++++---------------------
 2 files changed, 29 insertions(+), 21 deletions(-)

diff --git a/build_lib.py b/build_lib.py
index 1738b5f3e..c78cc6698 100644
--- a/build_lib.py
+++ b/build_lib.py
@@ -49,6 +49,10 @@
 parser.add_argument("--no_build_llvm", dest="build_llvm", action="store_false")
 parser.set_defaults(build_llvm=False)
 
+parser.add_argument(
+    "--llvm_source_path", type=str, help="Path to the LLVM project source code (optional, repo cloned if not set)"
+)
+
 parser.add_argument("--debug_llvm", action="store_true", help="Enable LLVM compiler code debugging, default disabled")
 parser.add_argument("--no_debug_llvm", dest="debug_llvm", action="store_false")
 parser.set_defaults(debug_llvm=False)
diff --git a/build_llvm.py b/build_llvm.py
index 24148bc2e..ed4f21b04 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -9,9 +9,7 @@
 base_path = os.path.dirname(os.path.realpath(__file__))
 build_path = os.path.join(base_path, "warp")
 
-llvm_project_dir = "external/llvm-project"
-llvm_project_path = os.path.join(base_path, llvm_project_dir)
-llvm_path = os.path.join(llvm_project_path, "llvm")
+llvm_project_path = os.path.join(base_path, "external/llvm-project")
 llvm_build_path = os.path.join(llvm_project_path, "out/build/")
 llvm_install_path = os.path.join(llvm_project_path, "out/install/")
 
@@ -44,29 +42,27 @@ def fetch_prebuilt_libraries():
         )
 
 
-def build_from_source_for_arch(args, arch):
-    # Install dependencies
-    subprocess.check_call([sys.executable, "-m", "pip", "install", "gitpython"])
-    subprocess.check_call([sys.executable, "-m", "pip", "install", "cmake"])
-    subprocess.check_call([sys.executable, "-m", "pip", "install", "ninja"])
+def build_from_source_for_arch(args, arch, llvm_source):
+    # Check out the LLVM project Git repository, unless it already exists
+    if not os.path.exists(llvm_source):
+        # Install dependencies
+        subprocess.check_call([sys.executable, "-m", "pip", "install", "gitpython"])
+        subprocess.check_call([sys.executable, "-m", "pip", "install", "cmake"])
+        subprocess.check_call([sys.executable, "-m", "pip", "install", "ninja"])
 
-    from git import Repo
+        from git import Repo
 
-    repo_url = "https://github.com/llvm/llvm-project.git"
+        repo_url = "https://github.com/llvm/llvm-project.git"
+        print(f"Cloning LLVM project from {repo_url}...")
 
-    if not os.path.exists(llvm_project_path):
-        print("Cloning LLVM project...")
         shallow_clone = True  # https://github.blog/2020-12-21-get-up-to-speed-with-partial-clone-and-shallow-clone/
         if shallow_clone:
-            repo = Repo.clone_from(
-                repo_url, to_path=llvm_project_path, single_branch=True, branch="llvmorg-15.0.7", depth=1
-            )
+            repo = Repo.clone_from(repo_url, to_path=llvm_source, single_branch=True, branch="llvmorg-15.0.7", depth=1)
         else:
-            repo = Repo.clone_from(repo_url, to_path=llvm_project_path)
+            repo = Repo.clone_from(repo_url, to_path=llvm_source)
             repo.git.checkout("tags/llvmorg-15.0.7", "-b", "llvm-15.0.7")
-    else:
-        print(f"Found existing {llvm_project_dir} directory")
-        repo = Repo(llvm_project_path)
+
+    print(f"Using LLVM project source from {llvm_source}")
 
     # CMake supports Debug, Release, RelWithDebInfo, and MinSizeRel builds
     if warp.config.mode == "release":
@@ -98,6 +94,7 @@ def build_from_source_for_arch(args, arch):
     else:
         host_triple = f"{arch}-pc-linux"
 
+    llvm_path = os.path.join(llvm_source, "llvm")
     build_path = os.path.join(llvm_build_path, f"{warp.config.mode}-{ arch}")
     install_path = os.path.join(llvm_install_path, f"{warp.config.mode}-{ arch}")
 
@@ -278,10 +275,17 @@ def build_from_source_for_arch(args, arch):
 
 
 def build_from_source(args):
-    build_from_source_for_arch(args, "x86_64")
+    print("Building Clang/LLVM from source...")
+
+    if args.llvm_source_path is not None:
+        llvm_source = args.llvm_source_path
+    else:
+        llvm_source = llvm_project_path
+
+    build_from_source_for_arch(args, "x86_64", llvm_source)
 
     if sys.platform == "darwin":
-        build_from_source_for_arch(args, "arm64")
+        build_from_source_for_arch(args, "arm64", llvm_source)
 
 
 # build warp-clang.dll

From 5f0633d91e43e4eae64e2198c6fc71adb18a4856 Mon Sep 17 00:00:00 2001
From: Eric Heiden <eheiden@nvidia.com>
Date: Sun, 3 Dec 2023 17:36:22 -0800
Subject: [PATCH 06/50] Flatten faces array for Mesh constructor in URDF parser

---
 warp/sim/import_urdf.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/warp/sim/import_urdf.py b/warp/sim/import_urdf.py
index 974f98bb5..b1458b6da 100644
--- a/warp/sim/import_urdf.py
+++ b/warp/sim/import_urdf.py
@@ -180,7 +180,7 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
                     # multiple meshes are contained in a scene
                     for geom in m.geometry.values():
                         vertices = np.array(geom.vertices, dtype=np.float32) * scaling
-                        faces = np.array(geom.faces, dtype=np.int32)
+                        faces = np.array(geom.faces.flatten(), dtype=np.int32)
                         mesh = Mesh(vertices, faces)
                         builder.add_shape_mesh(
                             body=link,
@@ -198,7 +198,7 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
                 else:
                     # a single mesh
                     vertices = np.array(m.vertices, dtype=np.float32) * scaling
-                    faces = np.array(m.faces, dtype=np.int32)
+                    faces = np.array(m.faces.flatten(), dtype=np.int32)
                     mesh = Mesh(vertices, faces)
                     builder.add_shape_mesh(
                         body=link,

From 77e99d013aed5e0e94772d55b56d1c0e30b54116 Mon Sep 17 00:00:00 2001
From: Gilles Daviet <gdaviet@nvidia.com>
Date: Sun, 3 Dec 2023 17:53:19 -0800
Subject: [PATCH 07/50] [warp.fem] Add point basis space and explicit
 point-based quadrature

---
 docs/modules/fem.rst                         |   8 +-
 examples/fem/example_apic_fluid.py           |   4 +-
 warp/fem/__init__.py                         |   4 +-
 warp/fem/integrate.py                        | 479 +++++++++++--------
 warp/fem/quadrature/__init__.py              |   2 +-
 warp/fem/quadrature/pic_quadrature.py        | 191 ++++++--
 warp/fem/quadrature/quadrature.py            |  90 +++-
 warp/fem/space/__init__.py                   |   2 +-
 warp/fem/space/basis_space.py                | 220 +++++++--
 warp/fem/space/grid_2d_function_space.py     |   4 +-
 warp/fem/space/grid_3d_function_space.py     |   4 +-
 warp/fem/space/hexmesh_function_space.py     |   4 +-
 warp/fem/space/quadmesh_2d_function_space.py |   4 +-
 warp/fem/space/tetmesh_function_space.py     |   4 +-
 warp/fem/space/topology.py                   |  10 +
 warp/fem/space/trimesh_2d_function_space.py  |   4 +-
 warp/tests/test_fem.py                       | 215 ++++++---
 17 files changed, 854 insertions(+), 395 deletions(-)

diff --git a/docs/modules/fem.rst b/docs/modules/fem.rst
index 359e42df9..75c4e6ed8 100644
--- a/docs/modules/fem.rst
+++ b/docs/modules/fem.rst
@@ -275,6 +275,9 @@ Geometry
 .. autoclass:: NodalQuadrature
    :show-inheritance:
 
+.. autoclass:: ExplicitQuadrature
+   :show-inheritance:
+
 .. autoclass:: PicQuadrature
    :show-inheritance:
 
@@ -300,6 +303,9 @@ Function Spaces
 .. autoclass:: SkewSymmetricTensorMapper
    :show-inheritance:
 
+.. autoclass:: PointBasisSpace
+   :show-inheritance:
+
 Fields
 ------
 
@@ -350,7 +356,7 @@ Interface classes are not meant to be constructed directly, but can be derived f
    :members: dimension, geometry, node_count, element_node_indices, trace
 
 .. autoclass:: BasisSpace
-   :members: topology, geometry, shape, node_positions
+   :members: topology, geometry, node_positions
 
 .. autoclass:: warp.fem.space.shape.ShapeFunction
 
diff --git a/examples/fem/example_apic_fluid.py b/examples/fem/example_apic_fluid.py
index 1470c638f..211602f3c 100644
--- a/examples/fem/example_apic_fluid.py
+++ b/examples/fem/example_apic_fluid.py
@@ -64,8 +64,6 @@ def update_particles(
     pos_proj = domain(lookup(domain, pos_adv))
     pos[s.qp_index] = pos_proj
 
-    return 0.0
-
 
 @integrand
 def velocity_boundary_projector_form(s: Sample, domain: Domain, u: Field, v: Field):
@@ -331,7 +329,7 @@ def update(self):
                 )
 
                 # (A)PIC advection
-                fem.integrate(
+                fem.interpolate(
                     update_particles,
                     quadrature=pic,
                     values={
diff --git a/warp/fem/__init__.py b/warp/fem/__init__.py
index 9a4091a1e..fe237a399 100644
--- a/warp/fem/__init__.py
+++ b/warp/fem/__init__.py
@@ -2,12 +2,12 @@
 from .geometry import GeometryPartition, LinearGeometryPartition, ExplicitGeometryPartition
 
 from .space import FunctionSpace, make_polynomial_space, ElementBasis
-from .space import BasisSpace, make_polynomial_basis_space, make_collocated_function_space
+from .space import BasisSpace, PointBasisSpace, make_polynomial_basis_space, make_collocated_function_space
 from .space import DofMapper, SkewSymmetricTensorMapper, SymmetricTensorMapper
 from .space import SpaceTopology, SpacePartition, SpaceRestriction, make_space_partition, make_space_restriction
 
 from .domain import GeometryDomain, Cells, Sides, BoundarySides, FrontierSides
-from .quadrature import Quadrature, RegularQuadrature, NodalQuadrature, PicQuadrature
+from .quadrature import Quadrature, RegularQuadrature, NodalQuadrature, ExplicitQuadrature, PicQuadrature
 from .polynomial import Polynomial
 
 from .field import FieldLike, DiscreteField, make_test, make_trial, make_restriction
diff --git a/warp/fem/integrate.py b/warp/fem/integrate.py
index 409591403..0230aa663 100644
--- a/warp/fem/integrate.py
+++ b/warp/fem/integrate.py
@@ -420,36 +420,6 @@ def visit_Call(self, call: ast.Call):
         return call
 
 
-def get_integrate_null_kernel(
-    integrand_func: wp.Function,
-    domain: GeometryDomain,
-    quadrature: Quadrature,
-    FieldStruct: wp.codegen.Struct,
-    ValueStruct: wp.codegen.Struct,
-):
-    def integrate_kernel_fn(
-        qp_arg: quadrature.Arg,
-        domain_arg: domain.ElementArg,
-        domain_index_arg: domain.ElementIndexArg,
-        fields: FieldStruct,
-        values: ValueStruct,
-    ):
-        element_index = domain.element_index(domain_index_arg, wp.tid())
-
-        test_dof_index = NULL_DOF_INDEX
-        trial_dof_index = NULL_DOF_INDEX
-
-        qp_point_count = quadrature.point_count(domain_arg, qp_arg, element_index)
-        for k in range(qp_point_count):
-            qp_index = quadrature.point_index(domain_arg, qp_arg, element_index, k)
-            qp_coords = quadrature.point_coords(domain_arg, qp_arg, element_index, k)
-            qp_weight = quadrature.point_weight(domain_arg, qp_arg, element_index, k)
-            sample = Sample(element_index, qp_coords, qp_index, qp_weight, test_dof_index, trial_dof_index)
-            integrand_func(sample, fields, values)
-
-    return integrate_kernel_fn
-
-
 def get_integrate_constant_kernel(
     integrand_func: wp.Function,
     domain: GeometryDomain,
@@ -477,7 +447,7 @@ def integrate_kernel_fn(
             qp_index = quadrature.point_index(domain_arg, qp_arg, element_index, k)
             coords = quadrature.point_coords(domain_arg, qp_arg, element_index, k)
             qp_weight = quadrature.point_weight(domain_arg, qp_arg, element_index, k)
-            
+
             sample = Sample(element_index, coords, qp_index, qp_weight, test_dof_index, trial_dof_index)
             vol = domain.element_measure(domain_arg, sample)
 
@@ -497,6 +467,7 @@ def get_integrate_linear_kernel(
     FieldStruct: wp.codegen.Struct,
     ValueStruct: wp.codegen.Struct,
     test: TestField,
+    output_dtype,
     accumulate_dtype,
 ):
     def integrate_kernel_fn(
@@ -506,32 +477,36 @@ def integrate_kernel_fn(
         test_arg: test.space_restriction.NodeArg,
         fields: FieldStruct,
         values: ValueStruct,
-        result: wp.array2d(dtype=accumulate_dtype),
+        result: wp.array2d(dtype=output_dtype),
     ):
-        local_node_index = wp.tid()
+        local_node_index, test_dof = wp.tid()
         node_index = test.space_restriction.node_partition_index(test_arg, local_node_index)
         element_count = test.space_restriction.node_element_count(test_arg, local_node_index)
 
         trial_dof_index = NULL_DOF_INDEX
 
+        val_sum = accumulate_dtype(0.0)
+
         for n in range(element_count):
             node_element_index = test.space_restriction.node_element_index(test_arg, local_node_index, n)
             element_index = domain.element_index(domain_index_arg, node_element_index.domain_element_index)
 
+            test_dof_index = DofIndex(node_element_index.node_index_in_element, test_dof)
+
             qp_point_count = quadrature.point_count(domain_arg, qp_arg, element_index)
             for k in range(qp_point_count):
                 qp_index = quadrature.point_index(domain_arg, qp_arg, element_index, k)
-                coords = quadrature.point_coords(domain_arg, qp_arg, element_index, k)
-
+                qp_coords = quadrature.point_coords(domain_arg, qp_arg, element_index, k)
                 qp_weight = quadrature.point_weight(domain_arg, qp_arg, element_index, k)
-                vol = domain.element_measure(domain_arg, make_free_sample(element_index, coords))
 
-                for i in range(test.space.VALUE_DOF_COUNT):
-                    test_dof_index = DofIndex(node_element_index.node_index_in_element, i)
-                    sample = Sample(element_index, coords, qp_index, qp_weight, test_dof_index, trial_dof_index)
-                    val = integrand_func(sample, fields, values)
+                vol = domain.element_measure(domain_arg, make_free_sample(element_index, qp_coords))
+
+                sample = Sample(element_index, qp_coords, qp_index, qp_weight, test_dof_index, trial_dof_index)
+                val = integrand_func(sample, fields, values)
+
+                val_sum += accumulate_dtype(qp_weight * vol * val)
 
-                    result[node_index, i] = result[node_index, i] + accumulate_dtype(qp_weight * vol * val)
+        result[node_index, test_dof] = output_dtype(val_sum)
 
     return integrate_kernel_fn
 
@@ -542,6 +517,7 @@ def get_integrate_linear_nodal_kernel(
     FieldStruct: wp.codegen.Struct,
     ValueStruct: wp.codegen.Struct,
     test: TestField,
+    output_dtype,
     accumulate_dtype,
 ):
     def integrate_kernel_fn(
@@ -550,7 +526,7 @@ def integrate_kernel_fn(
         test_restriction_arg: test.space_restriction.NodeArg,
         fields: FieldStruct,
         values: ValueStruct,
-        result: wp.array2d(dtype=accumulate_dtype),
+        result: wp.array2d(dtype=output_dtype),
     ):
         local_node_index, dof = wp.tid()
 
@@ -595,7 +571,7 @@ def integrate_kernel_fn(
 
                 val_sum += accumulate_dtype(node_weight * vol * val)
 
-        result[node_index, dof] = val_sum
+        result[node_index, dof] = output_dtype(val_sum)
 
     return integrate_kernel_fn
 
@@ -608,6 +584,7 @@ def get_integrate_bilinear_kernel(
     ValueStruct: wp.codegen.Struct,
     test: TestField,
     trial: TrialField,
+    output_dtype,
     accumulate_dtype,
 ):
     NODES_PER_ELEMENT = trial.space.topology.NODES_PER_ELEMENT
@@ -624,19 +601,26 @@ def integrate_kernel_fn(
         row_offsets: wp.array(dtype=int),
         triplet_rows: wp.array(dtype=int),
         triplet_cols: wp.array(dtype=int),
-        triplet_values: wp.array3d(dtype=accumulate_dtype),
+        triplet_values: wp.array3d(dtype=output_dtype),
     ):
-        test_local_node_index = wp.tid()
+        test_local_node_index, trial_node, test_dof, trial_dof = wp.tid()
 
         element_count = test.space_restriction.node_element_count(test_arg, test_local_node_index)
         test_node_index = test.space_restriction.node_partition_index(test_arg, test_local_node_index)
 
+        trial_dof_index = DofIndex(trial_node, trial_dof)
+
         for element in range(element_count):
             test_element_index = test.space_restriction.node_element_index(test_arg, test_local_node_index, element)
             element_index = domain.element_index(domain_index_arg, test_element_index.domain_element_index)
             qp_point_count = quadrature.point_count(domain_arg, qp_arg, element_index)
 
-            start_offset = (row_offsets[test_node_index] + element) * NODES_PER_ELEMENT
+            test_dof_index = DofIndex(
+                test_element_index.node_index_in_element,
+                test_dof,
+            )
+        
+            val_sum = accumulate_dtype(0.0)
 
             for k in range(qp_point_count):
                 qp_index = quadrature.point_index(domain_arg, qp_arg, element_index, k)
@@ -645,42 +629,28 @@ def integrate_kernel_fn(
                 qp_weight = quadrature.point_weight(domain_arg, qp_arg, element_index, k)
                 vol = domain.element_measure(domain_arg, make_free_sample(element_index, coords))
 
-                offset_cur = start_offset
-
-                for trial_n in range(NODES_PER_ELEMENT):
-                    for i in range(test.space.VALUE_DOF_COUNT):
-                        for j in range(trial.space.VALUE_DOF_COUNT):
-                            test_dof_index = DofIndex(
-                                test_element_index.node_index_in_element,
-                                i,
-                            )
-                            trial_dof_index = DofIndex(trial_n, j)
-                            sample = Sample(
-                                element_index,
-                                coords,
-                                qp_index,
-                                qp_weight,
-                                test_dof_index,
-                                trial_dof_index,
-                            )
-                            val = integrand_func(sample, fields, values)
-                            triplet_values[offset_cur, i, j] = triplet_values[offset_cur, i, j] + accumulate_dtype(
-                                qp_weight * vol * val
-                            )
-
-                    offset_cur += 1
-
-            # Set column indices
-            offset_cur = start_offset
-            for trial_n in range(NODES_PER_ELEMENT):
+                sample = Sample(
+                    element_index,
+                    coords,
+                    qp_index,
+                    qp_weight,
+                    test_dof_index,
+                    trial_dof_index,
+                )
+                val = integrand_func(sample, fields, values)
+                val_sum += accumulate_dtype(qp_weight * vol * val)
+
+            block_offset = (row_offsets[test_node_index] + element) * NODES_PER_ELEMENT + trial_node
+            triplet_values[block_offset, test_dof, trial_dof] = output_dtype(val_sum)
+
+            # Set row and column indices
+            if test_dof == 0 and trial_dof == 0:
                 trial_node_index = trial.space_partition.partition_node_index(
                     trial_partition_arg,
-                    trial.space.topology.element_node_index(domain_arg, trial_topology_arg, element_index, trial_n),
+                    trial.space.topology.element_node_index(domain_arg, trial_topology_arg, element_index, trial_node),
                 )
-
-                triplet_rows[offset_cur] = test_node_index
-                triplet_cols[offset_cur] = trial_node_index
-                offset_cur += 1
+                triplet_rows[block_offset] = test_node_index
+                triplet_cols[block_offset] = trial_node_index
 
     return integrate_kernel_fn
 
@@ -691,6 +661,7 @@ def get_integrate_bilinear_nodal_kernel(
     FieldStruct: wp.codegen.Struct,
     ValueStruct: wp.codegen.Struct,
     test: TestField,
+    output_dtype,
     accumulate_dtype,
 ):
     def integrate_kernel_fn(
@@ -701,7 +672,7 @@ def integrate_kernel_fn(
         values: ValueStruct,
         triplet_rows: wp.array(dtype=int),
         triplet_cols: wp.array(dtype=int),
-        triplet_values: wp.array3d(dtype=accumulate_dtype),
+        triplet_values: wp.array3d(dtype=output_dtype),
     ):
         local_node_index, test_dof, trial_dof = wp.tid()
 
@@ -729,7 +700,6 @@ def integrate_kernel_fn(
                     node_element_index.node_index_in_element,
                 )
 
-
                 test_dof_index = DofIndex(node_element_index.node_index_in_element, test_dof)
                 trial_dof_index = DofIndex(node_element_index.node_index_in_element, trial_dof)
 
@@ -746,7 +716,7 @@ def integrate_kernel_fn(
 
                 val_sum += accumulate_dtype(node_weight * vol * val)
 
-        triplet_values[local_node_index, test_dof, trial_dof] = val_sum
+        triplet_values[local_node_index, test_dof, trial_dof] = output_dtype(val_sum)
         triplet_rows[local_node_index] = node_index
         triplet_cols[local_node_index] = node_index
 
@@ -763,9 +733,12 @@ def _generate_integrate_kernel(
     trial: Optional[TrialField],
     trial_name: str,
     fields: Dict[str, FieldLike],
+    output_dtype: type,
     accumulate_dtype: type,
     kernel_options: Dict[str, Any] = {},
 ) -> wp.Kernel:
+    output_dtype = wp.types.type_scalar_type(output_dtype)
+
     # Extract field arguments from integrand
     field_args, value_args, domain_name, sample_name = _get_integrand_field_arguments(
         integrand, fields=fields, domain=domain
@@ -775,7 +748,7 @@ def _generate_integrate_kernel(
     ValueStruct = _gen_value_struct(value_args)
 
     # Check if kernel exist in cache
-    kernel_suffix = f"_itg_{wp.types.type_typestr(accumulate_dtype)}_{domain.name}_{FieldStruct.key}"
+    kernel_suffix = f"_itg_{wp.types.type_typestr(output_dtype)}{wp.types.type_typestr(accumulate_dtype)}_{domain.name}_{FieldStruct.key}"
     if nodal:
         kernel_suffix += "_nodal"
     else:
@@ -819,6 +792,7 @@ def _generate_integrate_kernel(
                 FieldStruct,
                 ValueStruct,
                 test=test,
+                output_dtype=output_dtype,
                 accumulate_dtype=accumulate_dtype,
             )
         else:
@@ -829,6 +803,7 @@ def _generate_integrate_kernel(
                 FieldStruct,
                 ValueStruct,
                 test=test,
+                output_dtype=output_dtype,
                 accumulate_dtype=accumulate_dtype,
             )
     else:
@@ -839,6 +814,7 @@ def _generate_integrate_kernel(
                 FieldStruct,
                 ValueStruct,
                 test=test,
+                output_dtype=output_dtype,
                 accumulate_dtype=accumulate_dtype,
             )
         else:
@@ -850,6 +826,7 @@ def _generate_integrate_kernel(
                 ValueStruct,
                 test=test,
                 trial=trial,
+                output_dtype=output_dtype,
                 accumulate_dtype=accumulate_dtype,
             )
 
@@ -949,32 +926,46 @@ def _launch_integrate_kernel(
 
     # Linear form
     if trial is None:
-        if test.space.VALUE_DOF_COUNT == 1:
-            accumulate_array_dtype = accumulate_dtype
-        else:
-            accumulate_array_dtype = cache.cached_vec_type(length=test.space.VALUE_DOF_COUNT, dtype=accumulate_dtype)
-
-        if output is not None and output.size < test.space_partition.node_count():
-            raise RuntimeError(f"Output array must be of size at least {test.space_partition.node_count()}")
-
-        accumulate_in_place = wp.types.types_equal(accumulate_array_dtype, output_dtype)
-
         # If an output array is provided with the correct type, accumulate directly into it
         # Otherwise, grab a temporary array
-        if output is not None and accumulate_in_place:
-            accumulate_array = output
-        else:
-            accumulate_temporary = cache.borrow_temporary(
+        if output is None:
+            if type_length(output_dtype) == test.space.VALUE_DOF_COUNT:
+                output_shape = (test.space_partition.node_count(),)
+            elif type_length(output_dtype) == 1:
+                output_shape = (test.space_partition.node_count(), test.space.VALUE_DOF_COUNT)
+            else:
+                raise RuntimeError(
+                    f"Incompatible output type {wp.types.type_repr(output_dtype)}, must be scalar or vector of length {test.space.VALUE_DOF_COUNT}"
+                )
+
+            output_temporary = cache.borrow_temporary(
                 temporary_store=temporary_store,
-                shape=test.space_partition.node_count(),
-                dtype=accumulate_array_dtype,
+                shape=output_shape,
+                dtype=output_dtype,
                 device=device,
-                requires_grad=output is not None and output.requires_grad,
             )
-            accumulate_array = accumulate_temporary.array
+
+            output = output_temporary.array
+
+        else:
+            output_temporary = None
+
+            if output.shape[0] < test.space_partition.node_count():
+                raise RuntimeError(f"Output array must have at least {test.space_partition.node_count()} rows")
+
+            output_dtype = output.dtype
+            if type_length(output_dtype) != test.space.VALUE_DOF_COUNT:
+                if type_length(output_dtype) != 1:
+                    raise RuntimeError(
+                        f"Incompatible output type {wp.types.type_repr(output_dtype)}, must be scalar or vector of length {test.space.VALUE_DOF_COUNT}"
+                    )
+                if output.ndim != 2 and output.shape[1] != test.space.VALUE_DOF_COUNT:
+                    raise RuntimeError(
+                        f"Incompatible output array shape, last dimension must be of size {test.space.VALUE_DOF_COUNT}"
+                    )
 
         # Launch the integration on the kernel on a 2d scalar view of the actual array
-        accumulate_array.zero_()
+        output.zero_()
 
         def as_2d_array(array):
             return wp.array(
@@ -984,11 +975,11 @@ def as_2d_array(array):
                 owner=False,
                 device=array.device,
                 shape=(test.space_partition.node_count(), test.space.VALUE_DOF_COUNT),
-                dtype=accumulate_dtype,
+                dtype=wp.types.type_scalar_type(output_dtype),
                 grad=None if array.grad is None else as_2d_array(array.grad),
             )
 
-        accumulate_2d_view = as_2d_array(accumulate_array)
+        output_view = output if output.ndim == 2 else as_2d_array(output)
 
         if nodal:
             wp.launch(
@@ -1000,14 +991,14 @@ def as_2d_array(array):
                     test_arg,
                     field_arg_values,
                     value_struct_values,
-                    accumulate_2d_view,
+                    output_view,
                 ],
                 device=device,
             )
         else:
             wp.launch(
                 kernel=kernel,
-                dim=test.space_restriction.node_count(),
+                dim=(test.space_restriction.node_count(), test.space.VALUE_DOF_COUNT),
                 inputs=[
                     qp_arg,
                     domain_elt_arg,
@@ -1015,47 +1006,23 @@ def as_2d_array(array):
                     test_arg,
                     field_arg_values,
                     value_struct_values,
-                    accumulate_2d_view,
+                    output_view,
                 ],
                 device=device,
             )
 
-        # Accumulate and output dtypes are simular -- return accumulation array directly
-        if output == accumulate_array:
-            return output
+        if output_temporary is not None:
+            return output_temporary.detach() 
 
-        if accumulate_in_place:
-            return accumulate_temporary.detach()
-
-        # Different types -- need to cast result
-        if output is not None:
-            cast_result = output
-        elif type_length(output_dtype) == test.space.VALUE_DOF_COUNT:
-            cast_result = wp.empty(
-                dtype=output_dtype,
-                shape=accumulate_array.shape,
-                device=device,
-                requires_grad=accumulate_array.requires_grad,
-            )
-        else:
-            cast_result = wp.empty(
-                dtype=output_dtype,
-                shape=accumulate_2d_view.shape,
-                device=device,
-                requires_grad=accumulate_array.requires_grad,
-            )
-
-        array_cast(in_array=accumulate_array, out_array=cast_result)
-        accumulate_temporary.release()  # Do not wait for garbage collection
-        return cast_result
+        return output
 
     # Bilinear form
 
     if test.space.VALUE_DOF_COUNT == 1 and trial.space.VALUE_DOF_COUNT == 1:
-        block_type = accumulate_dtype
+        block_type = output_dtype
     else:
         block_type = cache.cached_mat_type(
-            shape=(test.space.VALUE_DOF_COUNT, trial.space.VALUE_DOF_COUNT), dtype=accumulate_dtype
+            shape=(test.space.VALUE_DOF_COUNT, trial.space.VALUE_DOF_COUNT), dtype=output_dtype
         )
 
     if nodal:
@@ -1072,7 +1039,7 @@ def as_2d_array(array):
             test.space.VALUE_DOF_COUNT,
             trial.space.VALUE_DOF_COUNT,
         ),
-        dtype=accumulate_dtype,
+        dtype=output_dtype,
         device=device,
     )
     triplet_cols = triplet_cols_temp.array
@@ -1105,7 +1072,12 @@ def as_2d_array(array):
         trial_topology_arg = trial.space_partition.space_topology.topo_arg_value(device)
         wp.launch(
             kernel=kernel,
-            dim=test.space_restriction.node_count(),
+            dim=(
+                test.space_restriction.node_count(),
+                trial.space.topology.NODES_PER_ELEMENT,
+                test.space.VALUE_DOF_COUNT,
+                trial.space.VALUE_DOF_COUNT,
+            ),
             inputs=[
                 qp_arg,
                 domain_elt_arg,
@@ -1123,38 +1095,27 @@ def as_2d_array(array):
             device=device,
         )
 
-    compress_in_place = accumulate_dtype == output_dtype
-
     if output is not None:
         if output.nrow != test.space_partition.node_count() or output.ncol != trial.space_partition.node_count():
             raise RuntimeError(
                 f"Output matrix must have {test.space_partition.node_count()} rows and {trial.space_partition.node_count()} columns of blocks"
             )
 
-    if output is not None and compress_in_place:
-        bsr_matrix = output
     else:
-        bsr_matrix = bsr_zeros(
+        output = bsr_zeros(
             rows_of_blocks=test.space_partition.node_count(),
             cols_of_blocks=trial.space_partition.node_count(),
             block_type=block_type,
             device=device,
         )
-
-    bsr_set_from_triplets(bsr_matrix, triplet_rows, triplet_cols, triplet_values)
+    
+    bsr_set_from_triplets(output, triplet_rows, triplet_cols, triplet_values)
 
     # Do not wait for garbage collection
     triplet_values_temp.release()
     triplet_rows_temp.release()
     triplet_cols_temp.release()
 
-    if compress_in_place:
-        return bsr_matrix
-
-    if output is None:
-        return bsr_copy(bsr_matrix, scalar_type=output_dtype)
-
-    bsr_assign(src=bsr_matrix, dest=output)
     return output
 
 
@@ -1181,7 +1142,7 @@ def integrate(
         quadrature: Quadrature formula. If None, deduced from domain and fields degree.
         nodal: For linear or bilinear form only, use the test function nodes as the quadrature points. Assumes Lagrange interpolation functions are used, and no differential or DG operator is evaluated on the test or trial functions.
         fields: Discrete, test, and trial fields to be passed to the integrand. Keys in the dictionary must match integrand parameter names.
-        values: Additional variable values to be passed to the integrand, can by of any type accepted by warp kernel launchs. Keys in the dictionary must match integrand parameter names.
+        values: Additional variable values to be passed to the integrand, can be of any type accepted by warp kernel launchs. Keys in the dictionary must match integrand parameter names.
         temporary_store: shared pool from which to allocate temporary arrays
         accumulate_dtype: Scalar type to be used for accumulating integration samples
         output: Sparse matrix or warp array into which to store the result of the integration
@@ -1246,6 +1207,7 @@ def integrate(
         trial_name=trial_name,
         fields=fields,
         accumulate_dtype=accumulate_dtype,
+        output_dtype=output_dtype,
         kernel_options=kernel_options,
     )
 
@@ -1268,7 +1230,7 @@ def integrate(
     )
 
 
-def get_interpolate_kernel(
+def get_interpolate_to_field_kernel(
     integrand_func: wp.Function,
     domain: GeometryDomain,
     FieldStruct: wp.codegen.Struct,
@@ -1277,7 +1239,7 @@ def get_interpolate_kernel(
 ):
     value_type = dest.space.dtype
 
-    def interpolate_kernel_fn(
+    def interpolate_to_field_kernel_fn(
         domain_arg: domain.ElementArg,
         domain_index_arg: domain.ElementIndexArg,
         dest_node_arg: dest.space_restriction.NodeArg,
@@ -1331,12 +1293,82 @@ def interpolate_kernel_fn(
         if vol_sum > 0.0:
             dest.field.set_node_value(dest_eval_arg, node_index, val_sum / vol_sum)
 
-    return interpolate_kernel_fn
+    return interpolate_to_field_kernel_fn
+
+
+def get_interpolate_to_array_kernel(
+    integrand_func: wp.Function,
+    domain: GeometryDomain,
+    quadrature: Quadrature,
+    FieldStruct: wp.codegen.Struct,
+    ValueStruct: wp.codegen.Struct,
+    value_type: type,
+):
+    def interpolate_to_array_kernel_fn(
+        qp_arg: quadrature.Arg,
+        domain_arg: quadrature.domain.ElementArg,
+        domain_index_arg: quadrature.domain.ElementIndexArg,
+        fields: FieldStruct,
+        values: ValueStruct,
+        result: wp.array(dtype=value_type),
+    ):
+        element_index = domain.element_index(domain_index_arg, wp.tid())
+
+        test_dof_index = NULL_DOF_INDEX
+        trial_dof_index = NULL_DOF_INDEX
+
+        qp_point_count = quadrature.point_count(domain_arg, qp_arg, element_index)
+        for k in range(qp_point_count):
+            qp_index = quadrature.point_index(domain_arg, qp_arg, element_index, k)
+            coords = quadrature.point_coords(domain_arg, qp_arg, element_index, k)
+            qp_weight = quadrature.point_weight(domain_arg, qp_arg, element_index, k)
+
+            sample = Sample(element_index, coords, qp_index, qp_weight, test_dof_index, trial_dof_index)
+
+            result[qp_index] = integrand_func(sample, fields, values)
 
+    return interpolate_to_array_kernel_fn
 
-def _generate_interpolate_kernel(integrand: Integrand, dest: FieldLike, fields: Dict[str, FieldLike]) -> wp.Kernel:
-    domain = dest.domain
 
+def get_interpolate_nonvalued_kernel(
+    integrand_func: wp.Function,
+    domain: GeometryDomain,
+    quadrature: Quadrature,
+    FieldStruct: wp.codegen.Struct,
+    ValueStruct: wp.codegen.Struct,
+):
+    def interpolate_nonvalued_kernel_fn(
+        qp_arg: quadrature.Arg,
+        domain_arg: quadrature.domain.ElementArg,
+        domain_index_arg: quadrature.domain.ElementIndexArg,
+        fields: FieldStruct,
+        values: ValueStruct,
+    ):
+        element_index = domain.element_index(domain_index_arg, wp.tid())
+
+        test_dof_index = NULL_DOF_INDEX
+        trial_dof_index = NULL_DOF_INDEX
+
+        qp_point_count = quadrature.point_count(domain_arg, qp_arg, element_index)
+        for k in range(qp_point_count):
+            qp_index = quadrature.point_index(domain_arg, qp_arg, element_index, k)
+            coords = quadrature.point_coords(domain_arg, qp_arg, element_index, k)
+            qp_weight = quadrature.point_weight(domain_arg, qp_arg, element_index, k)
+
+            sample = Sample(element_index, coords, qp_index, qp_weight, test_dof_index, trial_dof_index)
+            integrand_func(sample, fields, values)
+
+    return interpolate_nonvalued_kernel_fn
+
+
+def _generate_interpolate_kernel(
+    integrand: Integrand,
+    domain: GeometryDomain,
+    dest: Optional[Union[FieldLike, wp.array]],
+    quadrature: Optional[Quadrature],
+    fields: Dict[str, FieldLike],
+    kernel_options: Dict[str, Any] = {},
+) -> wp.Kernel:
     # Extract field arguments from integrand
     field_args, value_args, domain_name, sample_name = _get_integrand_field_arguments(
         integrand, fields=fields, domain=domain
@@ -1354,9 +1386,14 @@ def _generate_interpolate_kernel(integrand: Integrand, dest: FieldLike, fields:
     ValueStruct = _gen_value_struct(value_args)
 
     # Check if kernel exist in cache
-    kernel_suffix = (
-        f"_itp_{FieldStruct.key}_{dest.domain.name}_{dest.space_restriction.space_partition.name}_{dest.space.name}"
-    )
+    if isinstance(dest, FieldRestriction):
+        kernel_suffix = (
+            f"_itp_{FieldStruct.key}_{dest.domain.name}_{dest.space_restriction.space_partition.name}_{dest.space.name}"
+        )
+    elif wp.types.is_array(dest):
+        kernel_suffix = f"_itp_{FieldStruct.key}_{quadrature.name}_{wp.types.type_repr(dest.dtype)}"
+    else:
+        kernel_suffix = f"_itp_{FieldStruct.key}_{quadrature.name}"
 
     kernel = cache.get_integrand_kernel(
         integrand=integrand,
@@ -1366,18 +1403,37 @@ def _generate_interpolate_kernel(integrand: Integrand, dest: FieldLike, fields:
         return kernel, FieldStruct, ValueStruct
 
     # Generate interpolation kernel
-    interpolate_kernel_fn = get_interpolate_kernel(
-        integrand_func,
-        domain,
-        dest=dest,
-        FieldStruct=FieldStruct,
-        ValueStruct=ValueStruct,
-    )
+    if isinstance(dest, FieldRestriction):
+        interpolate_kernel_fn = get_interpolate_to_field_kernel(
+            integrand_func,
+            domain,
+            dest=dest,
+            FieldStruct=FieldStruct,
+            ValueStruct=ValueStruct,
+        )
+    elif wp.types.is_array(dest):
+        interpolate_kernel_fn = get_interpolate_to_array_kernel(
+            integrand_func,
+            domain=domain,
+            quadrature=quadrature,
+            value_type=dest.dtype,
+            FieldStruct=FieldStruct,
+            ValueStruct=ValueStruct,
+        )
+    else:
+        interpolate_kernel_fn = get_interpolate_nonvalued_kernel(
+            integrand_func,
+            domain=domain,
+            quadrature=quadrature,
+            FieldStruct=FieldStruct,
+            ValueStruct=ValueStruct,
+        )
 
     kernel = cache.get_integrand_kernel(
         integrand=integrand,
         kernel_fn=interpolate_kernel_fn,
         suffix=kernel_suffix,
+        kernel_options=kernel_options,
         code_transformers=[
             PassFieldArgsToIntegrand(
                 arg_names=integrand.argspec.args,
@@ -1396,16 +1452,16 @@ def _launch_interpolate_kernel(
     kernel: wp.kernel,
     FieldStruct: wp.codegen.Struct,
     ValueStruct: wp.codegen.Struct,
-    dest: FieldLike,
+    domain: GeometryDomain,
+    dest: Optional[Union[FieldRestriction, wp.array]],
+    quadrature: Optional[Quadrature],
     fields: Dict[str, FieldLike],
     values: Dict[str, Any],
     device,
 ) -> wp.Kernel:
     # Set-up launch arguments
-    elt_arg = dest.domain.element_arg_value(device=device)
-    elt_index_arg = dest.domain.element_index_arg_value(device=device)
-    dest_node_arg = dest.space_restriction.node_arg(device=device)
-    dest_eval_arg = dest.field.eval_arg_value(device=device)
+    elt_arg = domain.element_arg_value(device=device)
+    elt_index_arg = domain.element_index_arg_value(device=device)
 
     field_arg_values = FieldStruct()
     for k, v in fields.items():
@@ -1415,37 +1471,65 @@ def _launch_interpolate_kernel(
     for k, v in values.items():
         setattr(value_struct_values, k, v)
 
-    wp.launch(
-        kernel=kernel,
-        dim=dest.space_restriction.node_count(),
-        inputs=[
-            elt_arg,
-            elt_index_arg,
-            dest_node_arg,
-            dest_eval_arg,
-            field_arg_values,
-            value_struct_values,
-        ],
-        device=device,
-    )
+    if isinstance(dest, FieldRestriction):
+        dest_node_arg = dest.space_restriction.node_arg(device=device)
+        dest_eval_arg = dest.field.eval_arg_value(device=device)
+
+        wp.launch(
+            kernel=kernel,
+            dim=dest.space_restriction.node_count(),
+            inputs=[
+                elt_arg,
+                elt_index_arg,
+                dest_node_arg,
+                dest_eval_arg,
+                field_arg_values,
+                value_struct_values,
+            ],
+            device=device,
+        )
+    elif wp.types.is_array(dest):
+        qp_arg = quadrature.arg_value(device)
+        wp.launch(
+            kernel=kernel,
+            dim=domain.element_count(),
+            inputs=[qp_arg, elt_arg, elt_index_arg, field_arg_values, value_struct_values, dest],
+            device=device,
+        )
+    else:
+        qp_arg = quadrature.arg_value(device)
+        wp.launch(
+            kernel=kernel,
+            dim=domain.element_count(),
+            inputs=[qp_arg, elt_arg, elt_index_arg, field_arg_values, value_struct_values],
+            device=device,
+        )
 
 
 def interpolate(
     integrand: Integrand,
-    dest: Union[DiscreteField, FieldRestriction],
+    dest: Optional[Union[DiscreteField, FieldRestriction, wp.array]] = None,
+    quadrature: Optional[Quadrature] = None,
     fields: Dict[str, FieldLike] = {},
     values: Dict[str, Any] = {},
     device=None,
+    kernel_options: Dict[str, Any] = {},
 ):
     """
-    Interpolates a function and assigns the result to a discrete field.
+    Interpolates a function at a finite set of sample points and optionally assigns the result to a discrete field or a raw warp array.
 
     Args:
         integrand: Function to be interpolated, must have :func:`integrand` decorator
-        dest: Discrete field, or restriction of a discrete field to a domain, to which the interpolation result will be assigned
+        dest: Where to store the interpolation result. Can be either
+
+         - a :class:`DiscreteField`, or restriction of a discrete field to a domain (from :func:`make_restriction`). In this case, interpolation will be performed at each node.
+         - a normal warp array. In this case, the `quadrature` argument defining the interpolation locations must be provided and the result of the `integrand` at each quadrature point will be assigned to the array.
+         - ``None``. In this case, the `quadrature` argument must also be provided and the `integrand` function is reponsible for dealing with the interpolation result.
+        quadrature: Quadrature formula defining the interpolation samples if `dest` is not a discrete field or field restriction.
         fields: Discrete fields to be passed to the integrand. Keys in the dictionary must match integrand parameters names.
-        values: Additional variable values to be passed to the integrand, can by of any type accepted by warp kernel launchs. Keys in the dictionary must match integrand parameter names.
+        values: Additional variable values to be passed to the integrand, can be of any type accepted by warp kernel launchs. Keys in the dictionary must match integrand parameter names.
         device: Device on which to perform the interpolation
+        kernel_options: Overloaded options to be passed to the kernel builder (e.g, ``{"enable_backward": True}``)
     """
     if not isinstance(integrand, Integrand):
         raise ValueError("integrand must be tagged with @integrand decorator")
@@ -1454,20 +1538,33 @@ def interpolate(
     if test is not None or trial is not None:
         raise ValueError("Test or Trial fields should not be used for interpolation")
 
-    if not isinstance(dest, FieldRestriction):
+    if isinstance(dest, DiscreteField):
         dest = make_restriction(dest)
 
+    if isinstance(dest, FieldRestriction):
+        domain = dest.domain
+    else:
+        if quadrature is None:
+            raise ValueError("When not interpolating to a field, a quadrature formula must be provided")
+
+        domain = quadrature.domain
+
     kernel, FieldStruct, ValueStruct = _generate_interpolate_kernel(
         integrand=integrand,
+        domain=domain,
         dest=dest,
+        quadrature=quadrature,
         fields=fields,
+        kernel_options=kernel_options,
     )
 
     return _launch_interpolate_kernel(
         kernel=kernel,
         FieldStruct=FieldStruct,
         ValueStruct=ValueStruct,
+        domain=domain,
         dest=dest,
+        quadrature=quadrature,
         fields=fields,
         values=values,
         device=device,
diff --git a/warp/fem/quadrature/__init__.py b/warp/fem/quadrature/__init__.py
index 6f7798827..5e5b91e5d 100644
--- a/warp/fem/quadrature/__init__.py
+++ b/warp/fem/quadrature/__init__.py
@@ -1,2 +1,2 @@
-from .quadrature import Quadrature, RegularQuadrature, NodalQuadrature
+from .quadrature import Quadrature, RegularQuadrature, NodalQuadrature, ExplicitQuadrature
 from .pic_quadrature import PicQuadrature
diff --git a/warp/fem/quadrature/pic_quadrature.py b/warp/fem/quadrature/pic_quadrature.py
index 746e48745..ebed52bfd 100644
--- a/warp/fem/quadrature/pic_quadrature.py
+++ b/warp/fem/quadrature/pic_quadrature.py
@@ -1,11 +1,11 @@
-from typing import Any
+from typing import Union, Tuple, Any, Optional
 
 import warp as wp
 
 from warp.fem.domain import GeometryDomain
-from warp.fem.types import ElementIndex, Coords
+from warp.fem.types import ElementIndex, Coords, make_free_sample
 from warp.fem.utils import compress_node_indices
-from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary
+from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary, dynamic_kernel
 
 from .quadrature import Quadrature
 
@@ -14,24 +14,36 @@
 
 
 class PicQuadrature(Quadrature):
-    """Particle-based quadrature formula, using a global set of points irregularely spread out over geometry elements.
+    """Particle-based quadrature formula, using a global set of points unevenly spread out over geometry elements.
 
     Useful for Particle-In-Cell and derived methods.
+
+    Args:
+        domain: Undelying domain for the qaudrature
+        positions: Either an array containing the world positions of all particles, or a tuple of arrays containing
+         the cell indices and coordinates for each particle. Note that the former requires the underlying geometry to
+         define a global :meth:`Geometry.cell_lookup` method; currently this is only available for :class:`Grid2D` and :class:`Grid3D`.
+        measures: Array containing the measure (area/volume) of each particle, used to defined the integration weights.
+         If ``None``, defaults to the cell measure divided by the number of particles in the cell.
+        temporary_store: shared pool from which to allocate temporary arrays
     """
 
     def __init__(
         self,
         domain: GeometryDomain,
-        positions: "wp.array()",
-        measures: "wp.array(dtype=float)",
+        positions: Union[
+            "wp.array(dtype=wp.vecXd)",
+            Tuple[
+                "wp.array(dtype=ElementIndex)",
+                "wp.array(dtype=Coords)",
+            ],
+        ],
+        measures: Optional["wp.array(dtype=float)"] = None,
         temporary_store: TemporaryStore = None,
     ):
         super().__init__(domain)
 
-        self.positions = positions
-        self.measures = measures
-
-        self._bin_particles(temporary_store)
+        self._bin_particles(positions, measures, temporary_store)
 
     @property
     def name(self):
@@ -61,12 +73,16 @@ def arg_value(self, device) -> Arg:
         arg = PicQuadrature.Arg()
         arg.cell_particle_offsets = self._cell_particle_offsets.array.to(device)
         arg.cell_particle_indices = self._cell_particle_indices.array.to(device)
-        arg.particle_fraction = self._particle_fraction.array.to(device)
-        arg.particle_coords = self._particle_coords.array.to(device)
+        arg.particle_fraction = self._particle_fraction.to(device)
+        arg.particle_coords = self._particle_coords.to(device)
         return arg
 
     def total_point_count(self):
-        return self.positions.shape[0]
+        return self._particle_coords.shape[0]
+
+    def active_cell_count(self):
+        """Number of cells containing at least one particle"""
+        return self._cell_count
 
     @wp.func
     def point_count(elt_arg: Any, qp_arg: Arg, element_index: ElementIndex):
@@ -109,52 +125,121 @@ def _fill_mask_kernel(
         i = wp.tid()
         element_mask[i] = wp.select(element_particle_offsets[i] == element_particle_offsets[i + 1], 1, 0)
 
-    def _bin_particles(self, temporary_store: TemporaryStore):
-        from warp.fem import cache
-
-        @cache.dynamic_kernel(suffix=f"{self.domain.name}")
-        def bin_particles(
-            cell_arg_value: self.domain.ElementArg,
-            positions: wp.array(dtype=self.positions.dtype),
-            measures: wp.array(dtype=float),
-            cell_index: wp.array(dtype=ElementIndex),
-            cell_coords: wp.array(dtype=Coords),
-            cell_fraction: wp.array(dtype=float),
-        ):
-            p = wp.tid()
-            sample = self.domain.element_lookup(cell_arg_value, positions[p])
+    @wp.kernel
+    def _compute_uniform_fraction(
+        cell_index: wp.array(dtype=ElementIndex),
+        cell_particle_offsets: wp.array(dtype=int),
+        cell_fraction: wp.array(dtype=float),
+    ):
+        p = wp.tid()
 
-            cell_index[p] = sample.element_index
+        cell = cell_index[p]
+        cell_particle_count = cell_particle_offsets[cell + 1] - cell_particle_offsets[cell]
 
-            cell_coords[p] = sample.element_coords
-            cell_fraction[p] = measures[p] / self.domain.element_measure(cell_arg_value, sample)
+        cell_fraction[p] = 1.0 / float(cell_particle_count)
 
-        device = self.positions.device
+    def _bin_particles(self, positions, measures, temporary_store: TemporaryStore):
+        if wp.types.is_array(positions):
+            # Initialize from positions
+            @dynamic_kernel(suffix=f"{self.domain.name}")
+            def bin_particles(
+                cell_arg_value: self.domain.ElementArg,
+                positions: wp.array(dtype=positions.dtype),
+                cell_index: wp.array(dtype=ElementIndex),
+                cell_coords: wp.array(dtype=Coords),
+            ):
+                p = wp.tid()
+                sample = self.domain.element_lookup(cell_arg_value, positions[p])
 
-        cell_index = borrow_temporary(temporary_store, shape=self.positions.shape, dtype=int, device=device)
-        self._particle_coords = borrow_temporary(
-            temporary_store, shape=self.positions.shape, dtype=Coords, device=device
-        )
-        self._particle_fraction = borrow_temporary(
-            temporary_store, shape=self.positions.shape, dtype=float, device=device
-        )
+                cell_index[p] = sample.element_index
+                cell_coords[p] = sample.element_coords
 
-        wp.launch(
-            dim=self.positions.shape[0],
-            kernel=bin_particles,
-            inputs=[
-                self.domain.element_arg_value(device),
-                self.positions,
-                self.measures,
-                cell_index.array,
-                self._particle_coords.array,
-                self._particle_fraction.array,
-            ],
-            device=device,
-        )
+            device = positions.device
+
+            cell_index_temp = borrow_temporary(temporary_store, shape=positions.shape, dtype=int, device=device)
+            cell_index = cell_index_temp.array
+
+            self._particle_coords_temp = borrow_temporary(
+                temporary_store, shape=positions.shape, dtype=Coords, device=device
+            )
+            self._particle_coords = self._particle_coords_temp.array
+
+            wp.launch(
+                dim=positions.shape[0],
+                kernel=bin_particles,
+                inputs=[
+                    self.domain.element_arg_value(device),
+                    positions,
+                    cell_index,
+                    self._particle_coords,
+                ],
+                device=device,
+            )
+
+        else:
+            cell_index, self._particle_coords = positions
+            if cell_index.shape != self._particle_coords.shape:
+                raise ValueError("Cell index and coordinates arrays must have the same shape")
+
+            cell_index_temp = None
+            self._particle_coords_temp = None
 
         self._cell_particle_offsets, self._cell_particle_indices, self._cell_count, _ = compress_node_indices(
-            self.domain.geometry_element_count(), cell_index.array
+            self.domain.geometry_element_count(), cell_index
         )
 
-        cell_index.release()
+        self._compute_fraction(cell_index, measures, temporary_store)
+
+    def _compute_fraction(self, cell_index, measures, temporary_store: TemporaryStore):
+        device = cell_index.device
+
+        self._particle_fraction_temp = borrow_temporary(
+            temporary_store, shape=cell_index.shape, dtype=float, device=device
+        )
+        self._particle_fraction = self._particle_fraction_temp.array
+
+        if measures is None:
+            # Split fraction uniformly over all particles in cell
+
+            wp.launch(
+                dim=cell_index.shape,
+                kernel=PicQuadrature._compute_uniform_fraction,
+                inputs=[
+                    cell_index,
+                    self._cell_particle_offsets.array,
+                    self._particle_fraction,
+                ],
+                device=device,
+            )
+
+        else:
+            # Fraction from particle measure
+
+            if measures.shape != cell_index.shape:
+                raise ValueError("Measures should be an 1d array or length equal to particle count")
+
+            @dynamic_kernel(suffix=f"{self.domain.name}")
+            def compute_fraction(
+                cell_arg_value: self.domain.ElementArg,
+                measures: wp.array(dtype=float),
+                cell_index: wp.array(dtype=ElementIndex),
+                cell_coords: wp.array(dtype=Coords),
+                cell_fraction: wp.array(dtype=float),
+            ):
+                p = wp.tid()
+                sample = make_free_sample(cell_index[p], cell_coords[p])
+
+                cell_fraction[p] = measures[p] / self.domain.element_measure(cell_arg_value, sample)
+
+            wp.launch(
+                dim=measures.shape[0],
+                kernel=compute_fraction,
+                inputs=[
+                    self.domain.element_arg_value(device),
+                    measures,
+                    cell_index,
+                    self._particle_coords,
+                    self._particle_fraction,
+                ],
+                device=device,
+            )
diff --git a/warp/fem/quadrature/quadrature.py b/warp/fem/quadrature/quadrature.py
index 3712080fc..e0208da3d 100644
--- a/warp/fem/quadrature/quadrature.py
+++ b/warp/fem/quadrature/quadrature.py
@@ -35,33 +35,37 @@ def arg_value(self, device) -> "Arg":
 
     def total_point_count(self):
         """Total number of quadrature points over the domain"""
-        pass
+        raise NotImplementedError()
+
+    def points_per_element(self):
+        """Number of points per element if constant, or ``None`` if varying"""
+        return None
 
     @staticmethod
     def point_count(elt_arg: "domain.GeometryDomain.ElementArg", qp_arg: Arg, element_index: ElementIndex):
         """Number of quadrature points for a given element"""
-        pass
+        raise NotImplementedError()
 
     @staticmethod
     def point_coords(
         elt_arg: "domain.GeometryDomain.ElementArg", qp_arg: Arg, element_index: ElementIndex, qp_index: int
     ):
         """Coordinates in element of the element's qp_index'th quadrature point"""
-        pass
+        raise NotImplementedError()
 
     @staticmethod
     def point_weight(
         elt_arg: "domain.GeometryDomain.ElementArg", qp_arg: Arg, element_index: ElementIndex, qp_index: int
     ):
         """Weight of the element's qp_index'th quadrature point"""
-        pass
+        raise NotImplementedError()
 
     @staticmethod
     def point_index(
         elt_arg: "domain.GeometryDomain.ElementArg", qp_arg: Arg, element_index: ElementIndex, qp_index: int
     ):
         """Global index of the element's qp_index'th  quadrature point"""
-        pass
+        raise NotImplementedError()
 
     def __str__(self) -> str:
         return self.name
@@ -98,13 +102,14 @@ def __init__(
 
     @property
     def name(self):
-        return (
-            f"{self.__class__.__name__}_{self.domain.name}_{self.family}_{self.order}"
-        )
+        return f"{self.__class__.__name__}_{self.domain.name}_{self.family}_{self.order}"
 
     def total_point_count(self):
         return len(self.points) * self.domain.geometry_element_count()
 
+    def points_per_element(self):
+        return self._N
+
     @property
     def points(self):
         return self._element_quadrature[0]
@@ -153,7 +158,7 @@ def point_index(elt_arg: self.domain.ElementArg, qp_arg: self.Arg, element_index
 class NodalQuadrature(Quadrature):
     """Quadrature using space node points as quadrature points
 
-    Note that in constant to the `nodal=True` flag for :func:`integrate`, this quadrature odes not make any assumption
+    Note that in contrast to the `nodal=True` flag for :func:`integrate`, this quadrature odes not make any assumption
     about orthogonality of shape functions, and is thus safe to use for arbitrary integrands.
     """
 
@@ -176,6 +181,9 @@ def name(self):
     def total_point_count(self):
         return self._space.node_count()
 
+    def points_per_element(self):
+        return self._space.topology.NODES_PER_ELEMENT
+
     def _make_arg(self):
         @cache.dynamic_struct(suffix=self.name)
         class Arg:
@@ -220,3 +228,67 @@ def point_index(elt_arg: self.domain.ElementArg, qp_arg: self.Arg, element_index
             return self._space.topology.element_node_index(elt_arg, qp_arg.topo_arg, element_index, qp_index)
 
         return point_index
+
+
+class ExplicitQuadrature(Quadrature):
+    """Quadrature using explicit per-cell points and weights. The number of quadrature points per cell is assumed
+    to be constant and deduced from the shape of the points and weights arrays.
+
+    Args:
+        domain: Domain of definition of the quadrature formula
+        points: 2d array of shape ``(domain.geometry_element-count(), points_per_cell)`` containing the coordinates of each quadrature point.
+        weights: 2d array of shape ``(domain.geometry_element-count(), points_per_cell)`` containing the weight for each quadrature point.
+
+    See also: :class:`PicQuadrature`
+    """
+
+    @wp.struct
+    class Arg:
+        points_per_cell: int
+        points: wp.array2d(dtype=Coords)
+        weights: wp.array2d(dtype=float)
+
+    def __init__(self, domain: domain.GeometryDomain, points: "wp.array2d(dtype=Coords)", weights: "wp.array2d(dtype=float)"):
+        super().__init__(domain)
+
+        if points.shape != weights.shape:
+            raise ValueError("Points and weights arrays must have the same shape")
+
+        self._points_per_cell = points.shape[1]
+        self._points = points
+        self._weights = weights
+
+    @property
+    def name(self):
+        return f"{self.__class__.__name__}"
+
+    def total_point_count(self):
+        return self._weights.size
+
+    def points_per_element(self):
+        return self._points_per_cell
+
+    @cache.cached_arg_value
+    def arg_value(self, device):
+        arg = self.Arg()
+        arg.points_per_cell = self._points_per_cell
+        arg.points = self._points.to(device)
+        arg.weights = self._weights.to(device)
+
+        return arg
+
+    @wp.func
+    def point_count(elt_arg: Any, qp_arg: Arg, element_index: ElementIndex):
+        return qp_arg.points_per_cell
+
+    @wp.func
+    def point_coords(elt_arg: Any, qp_arg: Arg, element_index: ElementIndex, qp_index: int):
+        return qp_arg.points[element_index, qp_index]
+
+    @wp.func
+    def point_weight(elt_arg: Any, qp_arg: Arg, element_index: ElementIndex, qp_index: int):
+        return qp_arg.weights[element_index, qp_index]
+
+    @wp.func
+    def point_index(elt_arg: Any, qp_arg: Arg, element_index: ElementIndex, qp_index: int):
+        return qp_arg.points_per_cell * element_index + qp_index
diff --git a/warp/fem/space/__init__.py b/warp/fem/space/__init__.py
index a984fd405..bfa455dc2 100644
--- a/warp/fem/space/__init__.py
+++ b/warp/fem/space/__init__.py
@@ -7,7 +7,7 @@
 
 from .function_space import FunctionSpace
 from .topology import SpaceTopology
-from .basis_space import BasisSpace
+from .basis_space import BasisSpace, PointBasisSpace
 from .collocated_function_space import CollocatedFunctionSpace
 
 from .grid_2d_function_space import (
diff --git a/warp/fem/space/basis_space.py b/warp/fem/space/basis_space.py
index 47e1a8161..ba8446788 100644
--- a/warp/fem/space/basis_space.py
+++ b/warp/fem/space/basis_space.py
@@ -4,9 +4,10 @@
 
 from warp.fem.types import ElementIndex, Coords, make_free_sample
 from warp.fem.geometry import Geometry
-from warp.fem import cache, utils
+from warp.fem.quadrature import Quadrature
+from warp.fem import cache
 
-from .topology import SpaceTopology
+from .topology import SpaceTopology, DiscontinuousSpaceTopology
 from .shape import ShapeFunction
 
 
@@ -25,19 +26,10 @@ class BasisArg:
 
         pass
 
-    def __init__(self, topology: SpaceTopology, shape: ShapeFunction):
+    def __init__(self, topology: SpaceTopology):
         self._topology = topology
-        self._shape = shape
 
         self.NODES_PER_ELEMENT = self._topology.NODES_PER_ELEMENT
-        self.ORDER = self._shape.ORDER
-
-        if hasattr(shape, "element_node_triangulation"):
-            self.node_triangulation = self._node_triangulation
-        if hasattr(shape, "element_node_tets"):
-            self.node_tets = self._node_tets
-        if hasattr(shape, "element_node_hexes"):
-            self.node_hexes = self._node_hexes
 
     @property
     def topology(self) -> SpaceTopology:
@@ -49,19 +41,10 @@ def geometry(self) -> Geometry:
         """Underlying geometry of the basis space"""
         return self._topology.geometry
 
-    @property
-    def shape(self) -> ShapeFunction:
-        """Shape functions used for defining individual element basis"""
-        return self._shape
-
     def basis_arg_value(self, device) -> "BasisArg":
         """Value for the argument structure to be passed to device functions"""
         return BasisSpace.BasisArg()
 
-    @property
-    def name(self):
-        return f"{self.topology.name}_{self._shape.name}"
-
     # Helpers for generating node positions
 
     def node_positions(self, out: Optional[wp.array] = None) -> wp.array:
@@ -117,6 +100,56 @@ def fill_node_positions(
 
         return node_positions
 
+    def make_node_coords_in_element(self):
+        raise NotImplementedError()
+
+    def make_node_quadrature_weight(self):
+        raise NotImplementedError()
+
+    def make_element_inner_weight(self):
+        raise NotImplementedError()
+
+    def make_element_outer_weight(self):
+        return self.make_element_inner_weight()
+
+    def make_element_inner_weight_gradient(self):
+        raise NotImplementedError()
+
+    def make_element_outer_weight_gradient(self):
+        return self.make_element_inner_weight_gradient()
+
+    def make_trace_node_quadrature_weight(self):
+        raise NotImplementedError()
+
+    def trace(self) -> "TraceBasisSpace":
+        return TraceBasisSpace(self)
+
+
+class ShapeBasisSpace(BasisSpace):
+    """Base class for defining shape-function-based basis spaces."""
+
+    def __init__(self, topology: SpaceTopology, shape: ShapeFunction):
+        super().__init__(topology)
+        self._shape = shape
+
+        self.ORDER = self._shape.ORDER
+
+        if hasattr(shape, "element_node_triangulation"):
+            self.node_triangulation = self._node_triangulation
+        if hasattr(shape, "element_node_tets"):
+            self.node_tets = self._node_tets
+        if hasattr(shape, "element_node_hexes"):
+            self.node_hexes = self._node_hexes
+
+    @property
+    def shape(self) -> ShapeFunction:
+        """Shape functions used for defining individual element basis"""
+        return self._shape
+
+    @property
+    def name(self):
+        return f"{self.topology.name}_{self._shape.name}"
+
     def make_node_coords_in_element(self):
         shape_node_coords_in_element = self._shape.make_node_coords_in_element()
 
@@ -156,14 +189,10 @@ def element_inner_weight(
             coords: Coords,
             node_index_in_elt: int,
         ):
-            a = self.BasisArg()
             return shape_element_inner_weight(coords, node_index_in_elt)
 
         return element_inner_weight
 
-    def make_element_outer_weight(self):
-        return self.make_element_inner_weight()
-
     def make_element_inner_weight_gradient(self):
         shape_element_inner_weight_gradient = self._shape.make_element_inner_weight_gradient()
 
@@ -179,11 +208,22 @@ def element_inner_weight_gradient(
 
         return element_inner_weight_gradient
 
-    def make_element_outer_weight_gradient(self):
-        return self.make_element_inner_weight_gradient()
+    def make_trace_node_quadrature_weight(self, trace_basis):
+        shape_trace_node_quadrature_weight = self._shape.make_trace_node_quadrature_weight()
 
-    def trace(self) -> "TraceBasisSpace":
-        return TraceBasisSpace(self)
+        @cache.dynamic_func(suffix=self.name)
+        def trace_node_quadrature_weight(
+            geo_side_arg: trace_basis.geometry.SideArg,
+            basis_arg: trace_basis.BasisArg,
+            element_index: ElementIndex,
+            node_index_in_elt: int,
+        ):
+            neighbour_elem, index_in_neighbour = trace_basis.topology.neighbor_cell_index(
+                geo_side_arg, element_index, node_index_in_elt
+            )
+            return shape_trace_node_quadrature_weight(index_in_neighbour)
+
+        return trace_node_quadrature_weight
 
     def _node_triangulation(self):
         element_node_indices = self._topology.element_node_indices().numpy()
@@ -211,7 +251,9 @@ class TraceBasisSpace(BasisSpace):
     """Auto-generated trace space evaluating the cell-defined basis on the geometry sides"""
 
     def __init__(self, basis: BasisSpace):
-        super().__init__(basis.topology.trace(), basis.shape)
+        super().__init__(basis.topology.trace())
+
+        self.ORDER = basis.ORDER
 
         self._basis = basis
         self.BasisArg = self._basis.BasisArg
@@ -247,21 +289,7 @@ def trace_node_coords_in_element(
         return trace_node_coords_in_element
 
     def make_node_quadrature_weight(self):
-        shape_trace_node_quadrature_weight = self._shape.make_trace_node_quadrature_weight()
-
-        @cache.dynamic_func(suffix=self._basis.name)
-        def trace_node_quadrature_weight(
-            geo_side_arg: self.geometry.SideArg,
-            basis_arg: self.BasisArg,
-            element_index: ElementIndex,
-            node_index_in_elt: int,
-        ):
-            neighbour_elem, index_in_neighbour = self.topology.neighbor_cell_index(
-                geo_side_arg, element_index, node_index_in_elt
-            )
-            return shape_trace_node_quadrature_weight(index_in_neighbour)
-
-        return trace_node_quadrature_weight
+        return self._basis.make_trace_node_quadrature_weight(self)
 
     def make_element_inner_weight(self):
         cell_inner_weight = self._basis.make_element_inner_weight()
@@ -337,9 +365,7 @@ def trace_element_inner_weight_gradient(
 
             cell_coords = self.geometry.side_inner_cell_coords(geo_side_arg, element_index, coords)
             geo_cell_arg = self.geometry.side_to_cell_arg(geo_side_arg)
-            return cell_inner_weight_gradient(
-                geo_cell_arg, basis_arg, cell_index, cell_coords, index_in_cell 
-            )
+            return cell_inner_weight_gradient(geo_cell_arg, basis_arg, cell_index, cell_coords, index_in_cell)
 
         return trace_element_inner_weight_gradient
 
@@ -361,11 +387,103 @@ def trace_element_outer_weight_gradient(
 
             cell_coords = self.geometry.side_outer_cell_coords(geo_side_arg, element_index, coords)
             geo_cell_arg = self.geometry.side_to_cell_arg(geo_side_arg)
-            return cell_outer_weight_gradient(
-                geo_cell_arg, basis_arg, cell_index, cell_coords, index_in_cell 
-            )
+            return cell_outer_weight_gradient(geo_cell_arg, basis_arg, cell_index, cell_coords, index_in_cell)
 
         return trace_element_outer_weight_gradient
 
     def __eq__(self, other: "TraceBasisSpace") -> bool:
         return self._topo == other._topo
+
+
+class PointBasisSpace(BasisSpace):
+    """An unstructured :class:`BasisSpace` that is non-zero at a finite set of points only.
+
+    The node locations and nodal quadrature weights are defined by a :class:`Quadrature` formula.
+    """
+
+    def __init__(self, quadrature: Quadrature):
+        self._quadrature = quadrature
+
+        if quadrature.points_per_element() is None:
+            raise NotImplementedError("Varying number of points per element is not supported yet")
+
+        topology = DiscontinuousSpaceTopology(
+            geometry=quadrature.domain.geometry, nodes_per_element=quadrature.points_per_element()
+        )
+        super().__init__(topology)
+
+        self.BasisArg = quadrature.Arg
+        self.basis_arg_value = quadrature.arg_value
+        self.ORDER = 0
+
+        self.make_element_outer_weight = self.make_element_inner_weight
+        self.make_element_outer_weight_gradient = self.make_element_outer_weight_gradient
+
+    @property
+    def name(self):
+        return f"{self._quadrature.name}_Point"
+
+    def make_node_coords_in_element(self):
+        @cache.dynamic_func(suffix=self.name)
+        def node_coords_in_element(
+            elt_arg: self._quadrature.domain.ElementArg,
+            basis_arg: self.BasisArg,
+            element_index: ElementIndex,
+            node_index_in_elt: int,
+        ):
+            return self._quadrature.point_coords(elt_arg, basis_arg, element_index, node_index_in_elt)
+
+        return node_coords_in_element
+
+    def make_node_quadrature_weight(self):
+        @cache.dynamic_func(suffix=self.name)
+        def node_quadrature_weight(
+            elt_arg: self._quadrature.domain.ElementArg,
+            basis_arg: self.BasisArg,
+            element_index: ElementIndex,
+            node_index_in_elt: int,
+        ):
+            return self._quadrature.point_weight(elt_arg, basis_arg, element_index, node_index_in_elt)
+
+        return node_quadrature_weight
+
+    def make_element_inner_weight(self):
+        @cache.dynamic_func(suffix=self.name)
+        def element_inner_weight(
+            elt_arg: self._quadrature.domain.ElementArg,
+            basis_arg: self.BasisArg,
+            element_index: ElementIndex,
+            coords: Coords,
+            node_index_in_elt: int,
+        ):
+            qp_coord = self._quadrature.point_coords(elt_arg, basis_arg, element_index, node_index_in_elt)
+            return wp.select(wp.length_sq(coords - qp_coord) < 0.001, 0.0, 1.0)
+
+        return element_inner_weight
+
+    def make_element_inner_weight_gradient(self):
+        gradient_vec = cache.cached_vec_type(length=self.geometry.dimension, dtype=float)
+
+        @cache.dynamic_func(suffix=self.name)
+        def element_inner_weight_gradient(
+            elt_arg: self._quadrature.domain.ElementArg,
+            basis_arg: self.BasisArg,
+            element_index: ElementIndex,
+            coords: Coords,
+            node_index_in_elt: int,
+        ):
+            return gradient_vec(0.0)
+
+        return element_inner_weight_gradient
+
+    def make_trace_node_quadrature_weight(self, trace_basis):
+        @cache.dynamic_func(suffix=self.name)
+        def trace_node_quadrature_weight(
+            elt_arg: trace_basis.geometry.SideArg,
+            basis_arg: trace_basis.BasisArg,
+            element_index: ElementIndex,
+            node_index_in_elt: int,
+        ):
+            return 0.0
+
+        return trace_node_quadrature_weight
diff --git a/warp/fem/space/grid_2d_function_space.py b/warp/fem/space/grid_2d_function_space.py
index 9b238202e..d808f9b75 100644
--- a/warp/fem/space/grid_2d_function_space.py
+++ b/warp/fem/space/grid_2d_function_space.py
@@ -7,7 +7,7 @@
 from warp.fem import cache
 
 from .topology import SpaceTopology, DiscontinuousSpaceTopologyMixin, forward_base_topology
-from .basis_space import BasisSpace, TraceBasisSpace
+from .basis_space import ShapeBasisSpace, TraceBasisSpace
 
 from .shape import ShapeFunction, ConstantShapeFunction
 from .shape import (
@@ -44,7 +44,7 @@ class Grid2DDiscontinuousSpaceTopology(
     pass
 
 
-class Grid2DBasisSpace(BasisSpace):
+class Grid2DBasisSpace(ShapeBasisSpace):
     def __init__(self, topology: Grid2DSpaceTopology, shape: ShapeFunction):
         super().__init__(topology, shape)
 
diff --git a/warp/fem/space/grid_3d_function_space.py b/warp/fem/space/grid_3d_function_space.py
index 877994dde..7d9dd1483 100644
--- a/warp/fem/space/grid_3d_function_space.py
+++ b/warp/fem/space/grid_3d_function_space.py
@@ -7,7 +7,7 @@
 from warp.fem import cache
 
 from .topology import SpaceTopology, DiscontinuousSpaceTopologyMixin, forward_base_topology
-from .basis_space import BasisSpace, TraceBasisSpace
+from .basis_space import ShapeBasisSpace, TraceBasisSpace
 
 from .shape import ShapeFunction, ConstantShapeFunction
 from .shape.cube_shape_function import (
@@ -45,7 +45,7 @@ class Grid3DDiscontinuousSpaceTopology(
     pass
 
 
-class Grid3DBasisSpace(BasisSpace):
+class Grid3DBasisSpace(ShapeBasisSpace):
     def __init__(self, topology: Grid3DSpaceTopology, shape: ShapeFunction):
         super().__init__(topology, shape)
 
diff --git a/warp/fem/space/hexmesh_function_space.py b/warp/fem/space/hexmesh_function_space.py
index d6b3c96f9..1afe472ca 100644
--- a/warp/fem/space/hexmesh_function_space.py
+++ b/warp/fem/space/hexmesh_function_space.py
@@ -6,7 +6,7 @@
 from warp.fem import cache
 
 from .topology import SpaceTopology, DiscontinuousSpaceTopologyMixin, forward_base_topology
-from .basis_space import BasisSpace, TraceBasisSpace
+from .basis_space import ShapeBasisSpace, TraceBasisSpace
 
 from .shape import ShapeFunction, ConstantShapeFunction
 from .shape import (
@@ -121,7 +121,7 @@ def __init__(self, mesh: Hexmesh, shape: ShapeFunction):
         super().__init__(mesh, shape.NODES_PER_ELEMENT)
 
 
-class HexmeshBasisSpace(BasisSpace):
+class HexmeshBasisSpace(ShapeBasisSpace):
     def __init__(self, topology: HexmeshSpaceTopology, shape: ShapeFunction):
         super().__init__(topology, shape)
 
diff --git a/warp/fem/space/quadmesh_2d_function_space.py b/warp/fem/space/quadmesh_2d_function_space.py
index 7c7881b8b..6258a1a12 100644
--- a/warp/fem/space/quadmesh_2d_function_space.py
+++ b/warp/fem/space/quadmesh_2d_function_space.py
@@ -6,7 +6,7 @@
 from warp.fem import cache
 
 from .topology import SpaceTopology, DiscontinuousSpaceTopologyMixin, forward_base_topology
-from .basis_space import BasisSpace, TraceBasisSpace
+from .basis_space import ShapeBasisSpace, TraceBasisSpace
 
 from .shape import ShapeFunction, ConstantShapeFunction
 from .shape import (
@@ -116,7 +116,7 @@ def __init__(self, mesh: Quadmesh2D, shape: ShapeFunction):
         super().__init__(mesh, shape.NODES_PER_ELEMENT)
 
 
-class Quadmesh2DBasisSpace(BasisSpace):
+class Quadmesh2DBasisSpace(ShapeBasisSpace):
     def __init__(self, topology: Quadmesh2DSpaceTopology, shape: ShapeFunction):
         super().__init__(topology, shape)
 
diff --git a/warp/fem/space/tetmesh_function_space.py b/warp/fem/space/tetmesh_function_space.py
index 0f8714ae0..4d6037ae2 100644
--- a/warp/fem/space/tetmesh_function_space.py
+++ b/warp/fem/space/tetmesh_function_space.py
@@ -5,7 +5,7 @@
 from warp.fem import cache
 
 from .topology import SpaceTopology, DiscontinuousSpaceTopologyMixin, forward_base_topology
-from .basis_space import BasisSpace, TraceBasisSpace
+from .basis_space import ShapeBasisSpace, TraceBasisSpace
 
 from .shape import ShapeFunction, ConstantShapeFunction
 from .shape import TetrahedronPolynomialShapeFunctions, TetrahedronNonConformingPolynomialShapeFunctions
@@ -136,7 +136,7 @@ def __init__(self, mesh: Tetmesh, shape: ShapeFunction):
         super().__init__(mesh, shape.NODES_PER_ELEMENT)
 
 
-class TetmeshBasisSpace(BasisSpace):
+class TetmeshBasisSpace(ShapeBasisSpace):
     def __init__(self, topology: TetmeshSpaceTopology, shape: ShapeFunction):
         super().__init__(topology, shape)
 
diff --git a/warp/fem/space/topology.py b/warp/fem/space/topology.py
index 23e866434..bc76a8102 100644
--- a/warp/fem/space/topology.py
+++ b/warp/fem/space/topology.py
@@ -227,6 +227,10 @@ def __init__(self, *args, **kwargs):
     def node_count(self):
         return self.geometry.cell_count() * self.NODES_PER_ELEMENT
 
+    @property
+    def name(self):
+        return f"{self.geometry.name}_D{self.NODES_PER_ELEMENT}"
+
     def _make_element_node_index(self):
         NODES_PER_ELEMENT = self.NODES_PER_ELEMENT
 
@@ -242,6 +246,12 @@ def element_node_index(
         return element_node_index
 
 
+class DiscontinuousSpaceTopology(DiscontinuousSpaceTopologyMixin, SpaceTopology):
+    """Topology for generic discontinuous spaces"""
+
+    pass
+
+
 class DeformedGeometrySpaceTopology(SpaceTopology):
     def __init__(self, geometry: DeformedGeometry, base_topology: SpaceTopology):
         super().__init__(geometry, base_topology.NODES_PER_ELEMENT)
diff --git a/warp/fem/space/trimesh_2d_function_space.py b/warp/fem/space/trimesh_2d_function_space.py
index 567a665d1..21f45ab6c 100644
--- a/warp/fem/space/trimesh_2d_function_space.py
+++ b/warp/fem/space/trimesh_2d_function_space.py
@@ -5,7 +5,7 @@
 from warp.fem import cache
 
 from .topology import SpaceTopology, DiscontinuousSpaceTopologyMixin, forward_base_topology
-from .basis_space import BasisSpace, TraceBasisSpace
+from .basis_space import ShapeBasisSpace, TraceBasisSpace
 
 from .shape import ShapeFunction, ConstantShapeFunction
 from .shape import Triangle2DPolynomialShapeFunctions, Triangle2DNonConformingPolynomialShapeFunctions
@@ -101,7 +101,7 @@ def __init__(self, mesh: Trimesh2D, shape: ShapeFunction):
         super().__init__(mesh, shape.NODES_PER_ELEMENT)
 
 
-class Trimesh2DBasisSpace(BasisSpace):
+class Trimesh2DBasisSpace(ShapeBasisSpace):
     def __init__(self, topology: Trimesh2DSpaceTopology, shape: ShapeFunction):
         super().__init__(topology, shape)
 
diff --git a/warp/tests/test_fem.py b/warp/tests/test_fem.py
index acf89fd34..70a6da628 100644
--- a/warp/tests/test_fem.py
+++ b/warp/tests/test_fem.py
@@ -14,13 +14,9 @@
 
 
 from warp.fem import Field, Sample, Domain, Coords
-from warp.fem import Grid2D, Grid3D, Trimesh2D, Tetmesh, Quadmesh2D, Hexmesh, Geometry
-from warp.fem import make_polynomial_space, SymmetricTensorMapper, SkewSymmetricTensorMapper
-from warp.fem import make_test
-from warp.fem import Cells, Sides, BoundarySides
-from warp.fem import integrate, interpolate
 from warp.fem import integrand, div, grad, curl, D, normal
-from warp.fem import RegularQuadrature, Polynomial
+import warp.fem as fem
+
 from warp.fem.types import make_free_sample
 from warp.fem.geometry.closest_point import project_on_tri_at_origin, project_on_tet_at_origin
 from warp.fem.geometry import DeformedGeometry
@@ -39,13 +35,13 @@ def linear_form(s: Sample, u: Field):
 def test_integrate_gradient(test_case, device):
     with wp.ScopedDevice(device):
         # Grid geometry
-        geo = Grid2D(res=wp.vec2i(5))
+        geo = fem.Grid2D(res=wp.vec2i(5))
 
         # Domain and function spaces
-        domain = Cells(geometry=geo)
-        quadrature = RegularQuadrature(domain=domain, order=3)
+        domain = fem.Cells(geometry=geo)
+        quadrature = fem.RegularQuadrature(domain=domain, order=3)
 
-        scalar_space = make_polynomial_space(geo, degree=3)
+        scalar_space = fem.make_polynomial_space(geo, degree=3)
 
         u = scalar_space.make_field()
         u.dof_values = wp.zeros_like(u.dof_values, requires_grad=True)
@@ -56,12 +52,12 @@ def test_integrate_gradient(test_case, device):
 
         # forward pass
         with tape:
-            integrate(linear_form, quadrature=quadrature, fields={"u": u}, output=result)
+            fem.integrate(linear_form, quadrature=quadrature, fields={"u": u}, output=result)
 
         tape.backward(result)
 
-        test = make_test(space=scalar_space, domain=domain)
-        rhs = integrate(linear_form, quadrature=quadrature, fields={"u": test})
+        test = fem.make_test(space=scalar_space, domain=domain)
+        rhs = fem.integrate(linear_form, quadrature=quadrature, fields={"u": test})
 
         err = np.linalg.norm(rhs.numpy() - u.dof_values.grad.numpy())
         test_case.assertLess(err, 1.0e-8)
@@ -87,14 +83,14 @@ def test_vector_divergence_theorem(test_case, device):
 
     with wp.ScopedDevice(device):
         # Grid geometry
-        geo = Grid2D(res=wp.vec2i(5))
+        geo = fem.Grid2D(res=wp.vec2i(5))
 
         # Domain and function spaces
-        interior = Cells(geometry=geo)
-        boundary = BoundarySides(geometry=geo)
+        interior = fem.Cells(geometry=geo)
+        boundary = fem.BoundarySides(geometry=geo)
 
-        vector_space = make_polynomial_space(geo, degree=2, dtype=wp.vec2)
-        scalar_space = make_polynomial_space(geo, degree=1, dtype=float)
+        vector_space = fem.make_polynomial_space(geo, degree=2, dtype=wp.vec2)
+        scalar_space = fem.make_polynomial_space(geo, degree=1, dtype=float)
 
         u = vector_space.make_field()
         u.dof_values = rng.random(size=(u.dof_values.shape[0], 2))
@@ -103,15 +99,15 @@ def test_vector_divergence_theorem(test_case, device):
         constant_one = scalar_space.make_field()
         constant_one.dof_values.fill_(1.0)
 
-        interior_quadrature = RegularQuadrature(domain=interior, order=vector_space.degree)
-        boundary_quadrature = RegularQuadrature(domain=boundary, order=vector_space.degree)
-        div_int = integrate(
+        interior_quadrature = fem.RegularQuadrature(domain=interior, order=vector_space.degree)
+        boundary_quadrature = fem.RegularQuadrature(domain=boundary, order=vector_space.degree)
+        div_int = fem.integrate(
             vector_divergence_form,
             quadrature=interior_quadrature,
             fields={"u": u, "q": constant_one},
             kernel_options={"enable_backward": False},
         )
-        boundary_int = integrate(
+        boundary_int = fem.integrate(
             vector_boundary_form,
             quadrature=boundary_quadrature,
             fields={"u": u.trace(), "q": constant_one.trace()},
@@ -124,21 +120,21 @@ def test_vector_divergence_theorem(test_case, device):
         q = scalar_space.make_field()
         q.dof_values = rng.random(size=q.dof_values.shape[0])
 
-        interior_quadrature = RegularQuadrature(domain=interior, order=vector_space.degree + scalar_space.degree)
-        boundary_quadrature = RegularQuadrature(domain=boundary, order=vector_space.degree + scalar_space.degree)
-        div_int = integrate(
+        interior_quadrature = fem.RegularQuadrature(domain=interior, order=vector_space.degree + scalar_space.degree)
+        boundary_quadrature = fem.RegularQuadrature(domain=boundary, order=vector_space.degree + scalar_space.degree)
+        div_int = fem.integrate(
             vector_divergence_form,
             quadrature=interior_quadrature,
             fields={"u": u, "q": q},
             kernel_options={"enable_backward": False},
         )
-        grad_int = integrate(
+        grad_int = fem.integrate(
             vector_grad_form,
             quadrature=interior_quadrature,
             fields={"u": u, "q": q},
             kernel_options={"enable_backward": False},
         )
-        boundary_int = integrate(
+        boundary_int = fem.integrate(
             vector_boundary_form,
             quadrature=boundary_quadrature,
             fields={"u": u.trace(), "q": q.trace()},
@@ -168,14 +164,14 @@ def test_tensor_divergence_theorem(test_case, device):
 
     with wp.ScopedDevice(device):
         # Grid geometry
-        geo = Grid2D(res=wp.vec2i(5))
+        geo = fem.Grid2D(res=wp.vec2i(5))
 
         # Domain and function spaces
-        interior = Cells(geometry=geo)
-        boundary = BoundarySides(geometry=geo)
+        interior = fem.Cells(geometry=geo)
+        boundary = fem.BoundarySides(geometry=geo)
 
-        tensor_space = make_polynomial_space(geo, degree=2, dtype=wp.mat22)
-        vector_space = make_polynomial_space(geo, degree=1, dtype=wp.vec2)
+        tensor_space = fem.make_polynomial_space(geo, degree=2, dtype=wp.mat22)
+        vector_space = fem.make_polynomial_space(geo, degree=1, dtype=wp.vec2)
 
         tau = tensor_space.make_field()
         tau.dof_values = rng.random(size=(tau.dof_values.shape[0], 2, 2))
@@ -184,15 +180,15 @@ def test_tensor_divergence_theorem(test_case, device):
         constant_vec = vector_space.make_field()
         constant_vec.dof_values.fill_(wp.vec2(0.5, 2.0))
 
-        interior_quadrature = RegularQuadrature(domain=interior, order=tensor_space.degree)
-        boundary_quadrature = RegularQuadrature(domain=boundary, order=tensor_space.degree)
-        div_int = integrate(
+        interior_quadrature = fem.RegularQuadrature(domain=interior, order=tensor_space.degree)
+        boundary_quadrature = fem.RegularQuadrature(domain=boundary, order=tensor_space.degree)
+        div_int = fem.integrate(
             tensor_divergence_form,
             quadrature=interior_quadrature,
             fields={"tau": tau, "v": constant_vec},
             kernel_options={"enable_backward": False},
         )
-        boundary_int = integrate(
+        boundary_int = fem.integrate(
             tensor_boundary_form,
             quadrature=boundary_quadrature,
             fields={"tau": tau.trace(), "v": constant_vec.trace()},
@@ -205,21 +201,21 @@ def test_tensor_divergence_theorem(test_case, device):
         v = vector_space.make_field()
         v.dof_values = rng.random(size=(v.dof_values.shape[0], 2))
 
-        interior_quadrature = RegularQuadrature(domain=interior, order=tensor_space.degree + vector_space.degree)
-        boundary_quadrature = RegularQuadrature(domain=boundary, order=tensor_space.degree + vector_space.degree)
-        div_int = integrate(
+        interior_quadrature = fem.RegularQuadrature(domain=interior, order=tensor_space.degree + vector_space.degree)
+        boundary_quadrature = fem.RegularQuadrature(domain=boundary, order=tensor_space.degree + vector_space.degree)
+        div_int = fem.integrate(
             tensor_divergence_form,
             quadrature=interior_quadrature,
             fields={"tau": tau, "v": v},
             kernel_options={"enable_backward": False},
         )
-        grad_int = integrate(
+        grad_int = fem.integrate(
             tensor_grad_form,
             quadrature=interior_quadrature,
             fields={"tau": tau, "v": v},
             kernel_options={"enable_backward": False},
         )
-        boundary_int = integrate(
+        boundary_int = fem.integrate(
             tensor_boundary_form,
             quadrature=boundary_quadrature,
             fields={"tau": tau.trace(), "v": v.trace()},
@@ -239,18 +235,18 @@ def test_grad_decomposition(test_case, device):
 
     with wp.ScopedDevice(device):
         # Grid geometry
-        geo = Grid3D(res=wp.vec3i(5))
+        geo = fem.Grid3D(res=wp.vec3i(5))
 
         # Domain and function spaces
-        domain = Cells(geometry=geo)
-        quadrature = RegularQuadrature(domain=domain, order=4)
+        domain = fem.Cells(geometry=geo)
+        quadrature = fem.RegularQuadrature(domain=domain, order=4)
 
-        vector_space = make_polynomial_space(geo, degree=2, dtype=wp.vec3)
+        vector_space = fem.make_polynomial_space(geo, degree=2, dtype=wp.vec3)
         u = vector_space.make_field()
 
         u.dof_values = rng.random(size=(u.dof_values.shape[0], 3))
 
-        err = integrate(grad_decomposition, quadrature=quadrature, fields={"u": u, "v": u})
+        err = fem.integrate(grad_decomposition, quadrature=quadrature, fields={"u": u, "v": u})
         test_case.assertLess(err, 1.0e-8)
 
 
@@ -300,7 +296,7 @@ def _gen_hexmesh(N):
     return wp.array(positions, dtype=wp.vec3), wp.array(vidx, dtype=int)
 
 
-def _launch_test_geometry_kernel(geo: Geometry, device):
+def _launch_test_geometry_kernel(geo: fem.Geometry, device):
     @dynamic_kernel(suffix=geo.name, kernel_options={"enable_backward": False})
     def test_geo_cells_kernel(
         cell_arg: geo.CellArg,
@@ -368,7 +364,7 @@ def test_geo_sides_kernel(
 
     cell_measures = wp.zeros(dtype=float, device=device, shape=geo.cell_count())
 
-    cell_quadrature = RegularQuadrature(Cells(geo), order=2)
+    cell_quadrature = fem.RegularQuadrature(fem.Cells(geo), order=2)
     cell_qps = wp.array(cell_quadrature.points, dtype=Coords, device=device)
     cell_qp_weights = wp.array(cell_quadrature.weights, dtype=float, device=device)
 
@@ -381,7 +377,7 @@ def test_geo_sides_kernel(
 
     side_measures = wp.zeros(dtype=float, device=device, shape=geo.side_count())
 
-    side_quadrature = RegularQuadrature(Sides(geo), order=2)
+    side_quadrature = fem.RegularQuadrature(fem.Sides(geo), order=2)
     side_qps = wp.array(side_quadrature.points, dtype=Coords, device=device)
     side_qp_weights = wp.array(side_quadrature.weights, dtype=float, device=device)
 
@@ -398,7 +394,7 @@ def test_geo_sides_kernel(
 def test_grid_2d(test_case, device):
     N = 3
 
-    geo = Grid2D(res=wp.vec2i(N))
+    geo = fem.Grid2D(res=wp.vec2i(N))
 
     test_case.assertEqual(geo.cell_count(), N**2)
     test_case.assertEqual(geo.vertex_count(), (N + 1) ** 2)
@@ -417,7 +413,7 @@ def test_triangle_mesh(test_case, device):
     with wp.ScopedDevice(device):
         positions, tri_vidx = _gen_trimesh(N)
 
-    geo = Trimesh2D(tri_vertex_indices=tri_vidx, positions=positions)
+    geo = fem.Trimesh2D(tri_vertex_indices=tri_vidx, positions=positions)
 
     test_case.assertEqual(geo.cell_count(), 2 * (N) ** 2)
     test_case.assertEqual(geo.vertex_count(), (N + 1) ** 2)
@@ -436,7 +432,7 @@ def test_quad_mesh(test_case, device):
     with wp.ScopedDevice(device):
         positions, quad_vidx = _gen_quadmesh(N)
 
-    geo = Quadmesh2D(quad_vertex_indices=quad_vidx, positions=positions)
+    geo = fem.Quadmesh2D(quad_vertex_indices=quad_vidx, positions=positions)
 
     test_case.assertEqual(geo.cell_count(), N**2)
     test_case.assertEqual(geo.vertex_count(), (N + 1) ** 2)
@@ -452,7 +448,7 @@ def test_quad_mesh(test_case, device):
 def test_grid_3d(test_case, device):
     N = 3
 
-    geo = Grid3D(res=wp.vec3i(N))
+    geo = fem.Grid3D(res=wp.vec3i(N))
 
     test_case.assertEqual(geo.cell_count(), (N) ** 3)
     test_case.assertEqual(geo.vertex_count(), (N + 1) ** 3)
@@ -472,7 +468,7 @@ def test_tet_mesh(test_case, device):
     with wp.ScopedDevice(device):
         positions, tet_vidx = _gen_tetmesh(N)
 
-    geo = Tetmesh(tet_vertex_indices=tet_vidx, positions=positions)
+    geo = fem.Tetmesh(tet_vertex_indices=tet_vidx, positions=positions)
 
     test_case.assertEqual(geo.cell_count(), 5 * (N) ** 3)
     test_case.assertEqual(geo.vertex_count(), (N + 1) ** 3)
@@ -492,7 +488,7 @@ def test_hex_mesh(test_case, device):
     with wp.ScopedDevice(device):
         positions, tet_vidx = _gen_hexmesh(N)
 
-    geo = Hexmesh(hex_vertex_indices=tet_vidx, positions=positions)
+    geo = fem.Hexmesh(hex_vertex_indices=tet_vidx, positions=positions)
 
     test_case.assertEqual(geo.cell_count(), (N) ** 3)
     test_case.assertEqual(geo.vertex_count(), (N + 1) ** 3)
@@ -518,15 +514,15 @@ def test_deformed_geometry(test_case, device):
     with wp.ScopedDevice(device):
         positions, tet_vidx = _gen_tetmesh(N)
 
-    geo = Tetmesh(tet_vertex_indices=tet_vidx, positions=positions)
+    geo = fem.Tetmesh(tet_vertex_indices=tet_vidx, positions=positions)
 
     translation = [1.0, 2.0, 3.0]
     rotation = [0.0, math.pi / 4.0, 0.0]
     scale = 2.0
 
-    vector_space = make_polynomial_space(geo, dtype=wp.vec3, degree=2)
+    vector_space = fem.make_polynomial_space(geo, dtype=wp.vec3, degree=2)
     pos_field = vector_space.make_field()
-    interpolate(
+    fem.interpolate(
         _rigid_deformation_field,
         dest=pos_field,
         values={"translation": translation, "rotation": rotation, "scale": scale},
@@ -705,9 +701,9 @@ def test_dof_mapper(test_case, device):
     matrix_types = [wp.mat22, wp.mat33]
 
     # Symmetric mapper
-    for mapping in SymmetricTensorMapper.Mapping:
+    for mapping in fem.SymmetricTensorMapper.Mapping:
         for dtype in matrix_types:
-            mapper = SymmetricTensorMapper(dtype, mapping=mapping)
+            mapper = fem.SymmetricTensorMapper(dtype, mapping=mapping)
             dof_dtype = mapper.dof_dtype
 
             for k in range(dof_dtype._length_):
@@ -727,7 +723,7 @@ def test_dof_mapper(test_case, device):
 
     # Skew-symmetric mapper
     for dtype in matrix_types:
-        mapper = SkewSymmetricTensorMapper(dtype)
+        mapper = fem.SkewSymmetricTensorMapper(dtype)
         dof_dtype = mapper.dof_dtype
 
         if hasattr(dof_dtype, "_length_"):
@@ -879,9 +875,9 @@ def square_coord_delta_sampler(epsilon: float, state: wp.uint32):
         param_delta = wp.normalize(wp.vec2(wp.randf(state), wp.randf(state))) * epsilon
         return param_delta, Coords(param_delta[0], param_delta[1], 0.0)
 
-    Q_1 = shape.SquareBipolynomialShapeFunctions(degree=1, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
-    Q_2 = shape.SquareBipolynomialShapeFunctions(degree=2, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
-    Q_3 = shape.SquareBipolynomialShapeFunctions(degree=3, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    Q_1 = shape.SquareBipolynomialShapeFunctions(degree=1, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    Q_2 = shape.SquareBipolynomialShapeFunctions(degree=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    Q_3 = shape.SquareBipolynomialShapeFunctions(degree=3, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
 
     test_shape_function_weight(test_case, Q_1, square_coord_sampler, SQUARE_CENTER_COORDS)
     test_shape_function_weight(test_case, Q_2, square_coord_sampler, SQUARE_CENTER_COORDS)
@@ -893,8 +889,8 @@ def square_coord_delta_sampler(epsilon: float, state: wp.uint32):
     test_shape_function_gradient(test_case, Q_2, square_coord_sampler, square_coord_delta_sampler)
     test_shape_function_gradient(test_case, Q_3, square_coord_sampler, square_coord_delta_sampler)
 
-    S_2 = shape.SquareSerendipityShapeFunctions(degree=2, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
-    S_3 = shape.SquareSerendipityShapeFunctions(degree=3, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    S_2 = shape.SquareSerendipityShapeFunctions(degree=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    S_3 = shape.SquareSerendipityShapeFunctions(degree=3, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
 
     test_shape_function_weight(test_case, S_2, square_coord_sampler, SQUARE_CENTER_COORDS)
     test_shape_function_weight(test_case, S_3, square_coord_sampler, SQUARE_CENTER_COORDS)
@@ -930,9 +926,9 @@ def cube_coord_delta_sampler(epsilon: float, state: wp.uint32):
         param_delta = wp.normalize(wp.vec3(wp.randf(state), wp.randf(state), wp.randf(state))) * epsilon
         return param_delta, param_delta
 
-    Q_1 = shape.CubeTripolynomialShapeFunctions(degree=1, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
-    Q_2 = shape.CubeTripolynomialShapeFunctions(degree=2, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
-    Q_3 = shape.CubeTripolynomialShapeFunctions(degree=3, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    Q_1 = shape.CubeTripolynomialShapeFunctions(degree=1, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    Q_2 = shape.CubeTripolynomialShapeFunctions(degree=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    Q_3 = shape.CubeTripolynomialShapeFunctions(degree=3, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
 
     test_shape_function_weight(test_case, Q_1, cube_coord_sampler, CUBE_CENTER_COORDS)
     test_shape_function_weight(test_case, Q_2, cube_coord_sampler, CUBE_CENTER_COORDS)
@@ -944,8 +940,8 @@ def cube_coord_delta_sampler(epsilon: float, state: wp.uint32):
     test_shape_function_gradient(test_case, Q_2, cube_coord_sampler, cube_coord_delta_sampler)
     test_shape_function_gradient(test_case, Q_3, cube_coord_sampler, cube_coord_delta_sampler)
 
-    S_2 = shape.CubeSerendipityShapeFunctions(degree=2, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
-    S_3 = shape.CubeSerendipityShapeFunctions(degree=3, family=Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    S_2 = shape.CubeSerendipityShapeFunctions(degree=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    S_3 = shape.CubeSerendipityShapeFunctions(degree=3, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
 
     test_shape_function_weight(test_case, S_2, cube_coord_sampler, CUBE_CENTER_COORDS)
     test_shape_function_weight(test_case, S_3, cube_coord_sampler, CUBE_CENTER_COORDS)
@@ -1054,6 +1050,81 @@ def tet_coord_delta_sampler(epsilon: float, state: wp.uint32):
     wp.synchronize()
 
 
+def test_point_basis(test_case, device):
+    geo = fem.Grid2D(res=wp.vec2i(2))
+
+    domain = fem.Cells(geo)
+
+    quadrature = fem.RegularQuadrature(domain, order=2, family=fem.Polynomial.GAUSS_LEGENDRE)
+    point_basis = fem.PointBasisSpace(quadrature)
+
+    point_space = fem.make_collocated_function_space(point_basis)
+    point_test = fem.make_test(point_space, domain=domain)
+
+    # Sample at particle positions
+    ones = fem.integrate(linear_form, fields={"u": point_test}, nodal=True)
+    test_case.assertAlmostEqual(np.sum(ones.numpy()), 1.0, places=5)
+
+    # Sampling outside of particle positions
+    other_quadrature = fem.RegularQuadrature(domain, order=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
+    zeros = fem.integrate(linear_form, quadrature=other_quadrature, fields={"u": point_test})
+
+    test_case.assertAlmostEqual(np.sum(zeros.numpy()), 0.0, places=5)
+
+
+@fem.integrand
+def _bicubic(s: Sample, domain: Domain):
+    x = domain(s)
+    return wp.pow(x[0], 3.0) * wp.pow(x[1], 3.0)
+
+
+@fem.integrand
+def _piecewise_constant(s: Sample):
+    return float(s.element_index)
+
+
+def test_particle_quadratures(test_case, device):
+    geo = fem.Grid2D(res=wp.vec2i(2))
+
+    domain = fem.Cells(geo)
+    points, weights = domain.reference_element().instantiate_quadrature(order=4, family=fem.Polynomial.GAUSS_LEGENDRE)
+    points_per_cell = len(points)
+
+    points = points * domain.element_count()
+    weights = weights * domain.element_count()
+
+    points = wp.array(points, shape=(domain.element_count(), points_per_cell), dtype=Coords, device=device)
+    weights = wp.array(weights, shape=(domain.element_count(), points_per_cell), dtype=float, device=device)
+
+    explicit_quadrature = fem.ExplicitQuadrature(domain, points, weights)
+
+    test_case.assertEqual(explicit_quadrature.points_per_element(), points_per_cell)
+    test_case.assertEqual(explicit_quadrature.total_point_count(), points_per_cell * geo.cell_count())
+
+    val = fem.integrate(_bicubic, quadrature=explicit_quadrature)
+    test_case.assertAlmostEqual(val, 1.0 / 16, places=5)
+
+    element_indices = wp.array([3, 3, 2], dtype=int, device=device)
+    element_coords = wp.array(
+        [
+            [0.25, 0.5, 0.0],
+            [0.5, 0.25, 0.0],
+            [0.5, 0.5, 0.0],
+        ],
+        dtype=Coords,
+        device=device,
+    )
+
+    pic_quadrature = fem.PicQuadrature(domain, positions=(element_indices, element_coords))
+
+    test_case.assertIsNone(pic_quadrature.points_per_element())
+    test_case.assertEqual(pic_quadrature.total_point_count(), 3)
+    test_case.assertEqual(pic_quadrature.active_cell_count(), 2)
+
+    val = fem.integrate(_piecewise_constant, quadrature=pic_quadrature)
+    test_case.assertAlmostEqual(val, 1.25, places=5)
+
+
 def register(parent):
     devices = get_test_devices()
 
@@ -1078,6 +1149,8 @@ class TestFem(parent):
     add_function_test(TestFem, "test_cube_shape_functions", test_cube_shape_functions)
     add_function_test(TestFem, "test_tri_shape_functions", test_tri_shape_functions)
     add_function_test(TestFem, "test_tet_shape_functions", test_tet_shape_functions)
+    add_function_test(TestFem, "test_point_basis", test_point_basis)
+    add_function_test(TestFem, "test_particle_quadratures", test_particle_quadratures)
 
     return TestFem
 

From cb179c28ec9255c628f706496349a314587cf162 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Tue, 21 Nov 2023 11:43:06 +1300
Subject: [PATCH 08/50] Expose user-friendly mesh query overloads

These overloads return structs instead of writing the results
to arguments passed by reference.
---
 warp/__init__.py                    |   4 +-
 warp/builtins.py                    | 131 ++++++++++++++++++++++++++++
 warp/native/exports.h               |   6 ++
 warp/native/mesh.h                  | 131 ++++++++++++++++++++++++++--
 warp/tests/test_mesh_query_point.py |  28 ++++++
 warp/types.py                       |  35 +++++++-
 6 files changed, 328 insertions(+), 7 deletions(-)

diff --git a/warp/__init__.py b/warp/__init__.py
index 685bdea87..3ef997ad6 100644
--- a/warp/__init__.py
+++ b/warp/__init__.py
@@ -26,7 +26,9 @@
 
 # geometry types
 from warp.types import Bvh, Mesh, HashGrid, Volume, MarchingCubes
-from warp.types import bvh_query_t, mesh_query_aabb_t, hash_grid_query_t
+from warp.types import bvh_query_t, hash_grid_query_t, mesh_query_aabb_t, mesh_query_point_t, mesh_query_ray_t
+
+
 
 # device-wide gemms
 from warp.types import matmul, adj_matmul, batched_matmul, adj_batched_matmul, from_ptr
diff --git a/warp/builtins.py b/warp/builtins.py
index 5869800b1..d655dc819 100644
--- a/warp/builtins.py
+++ b/warp/builtins.py
@@ -1541,6 +1541,27 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param bary_v: Returns the barycentric v coordinate of the closest point""",
 )
 
+add_builtin(
+    "mesh_query_point",
+    input_types={
+        "id": uint64,
+        "point": vec3,
+        "max_dist": float,
+    },
+    value_type=mesh_query_point_t,
+    group="Geometry",
+    doc="""Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
+
+   Identifies the sign of the distance using additional ray-casts to determine if the point is inside or outside.
+   This method is relatively robust, but does increase computational cost.
+   See below for additional sign determination methods.
+
+   :param id: The mesh identifier
+   :param point: The point in space to query
+   :param max_dist: Mesh faces above this distance will not be considered by the query""",
+    require_original_output_arg=True,
+)
+
 add_builtin(
     "mesh_query_point_no_sign",
     input_types={
@@ -1565,6 +1586,25 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param bary_v: Returns the barycentric v coordinate of the closest point""",
 )
 
+add_builtin(
+    "mesh_query_point_no_sign",
+    input_types={
+        "id": uint64,
+        "point": vec3,
+        "max_dist": float,
+    },
+    value_type=mesh_query_point_t,
+    group="Geometry",
+    doc="""Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
+
+   This method does not compute the sign of the point (inside/outside) which makes it faster than other point query methods.
+
+   :param id: The mesh identifier
+   :param point: The point in space to query
+   :param max_dist: Mesh faces above this distance will not be considered by the query""",
+    require_original_output_arg=True,
+)
+
 add_builtin(
     "mesh_query_furthest_point_no_sign",
     input_types={
@@ -1589,6 +1629,25 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param bary_v: Returns the barycentric v coordinate of the furthest point""",
 )
 
+add_builtin(
+    "mesh_query_furthest_point_no_sign",
+    input_types={
+        "id": uint64,
+        "point": vec3,
+        "min_dist": float,
+    },
+    value_type=mesh_query_point_t,
+    group="Geometry",
+    doc="""Computes the furthest point on the mesh with identifier `id` to the given point in space.
+
+   This method does not compute the sign of the point (inside/outside).
+
+   :param id: The mesh identifier
+   :param point: The point in space to query
+   :param min_dist: Mesh faces below this distance will not be considered by the query""",
+    require_original_output_arg=True,
+)
+
 add_builtin(
     "mesh_query_point_sign_normal",
     input_types={
@@ -1622,6 +1681,31 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
                    fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3""",
 )
 
+add_builtin(
+    "mesh_query_point_sign_normal",
+    input_types={
+        "id": uint64,
+        "point": vec3,
+        "max_dist": float,
+        "epsilon": float,
+    },
+    defaults={"epsilon": 1.0e-3},
+    value_type=mesh_query_point_t,
+    group="Geometry",
+    doc="""Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
+
+   Identifies the sign of the distance (inside/outside) using the angle-weighted pseudo normal.
+   This approach to sign determination is robust for well conditioned meshes that are watertight and non-self intersecting.
+   It is also comparatively fast to compute.
+
+   :param id: The mesh identifier
+   :param point: The point in space to query
+   :param max_dist: Mesh faces above this distance will not be considered by the query
+   :param epsilon: Epsilon treating distance values as equal, when locating the minimum distance vertex/face/edge, as a
+                   fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3""",
+    require_original_output_arg=True,
+)
+
 add_builtin(
     "mesh_query_point_sign_winding_number",
     input_types={
@@ -1658,6 +1742,34 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param threshold: The threshold of the winding number to be considered inside, default 0.5""",
 )
 
+add_builtin(
+    "mesh_query_point_sign_winding_number",
+    input_types={
+        "id": uint64,
+        "point": vec3,
+        "max_dist": float,
+        "accuracy": float,
+        "threshold": float,
+    },
+    defaults={"accuracy": 2.0, "threshold": 0.5},
+    value_type=mesh_query_point_t,
+    group="Geometry",
+    doc="""Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given point in space.
+
+   Identifies the sign using the winding number of the mesh relative to the query point. This method of sign determination is robust for poorly conditioned meshes
+   and provides a smooth approximation to sign even when the mesh is not watertight. This method is the most robust and accurate of the sign determination meshes
+   but also the most expensive.
+
+   .. note:: The :class:`Mesh` object must be constructed with ``support_winding_number=True`` for this method to return correct results.
+
+   :param id: The mesh identifier
+   :param point: The point in space to query
+   :param max_dist: Mesh faces above this distance will not be considered by the query
+   :param accuracy: Accuracy for computing the winding number with fast winding number method utilizing second-order dipole approximation, default 2.0
+   :param threshold: The threshold of the winding number to be considered inside, default 0.5""",
+    require_original_output_arg=True,
+)
+
 add_builtin(
     "mesh_query_ray",
     input_types={
@@ -1688,6 +1800,25 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param face: Returns the index of the hit face""",
 )
 
+add_builtin(
+    "mesh_query_ray",
+    input_types={
+        "id": uint64,
+        "start": vec3,
+        "dir": vec3,
+        "max_t": float,
+    },
+    value_type=mesh_query_ray_t,
+    group="Geometry",
+    doc="""Computes the closest ray hit on the :class:`Mesh` with identifier ``id``.
+
+   :param id: The mesh identifier
+   :param start: The start point of the ray
+   :param dir: The ray direction (should be normalized)
+   :param max_t: The maximum distance along the ray to check for intersections""",
+    require_original_output_arg=True,
+)
+
 add_builtin(
     "mesh_query_aabb",
     input_types={"id": uint64, "lower": vec3, "upper": vec3},
diff --git a/warp/native/exports.h b/warp/native/exports.h
index d737c1fa9..78b2f14d2 100644
--- a/warp/native/exports.h
+++ b/warp/native/exports.h
@@ -747,11 +747,17 @@ WP_API void builtin_bvh_query_aabb_uint64_vec3f_vec3f(uint64 id, vec3f lower, ve
 WP_API void builtin_bvh_query_ray_uint64_vec3f_vec3f(uint64 id, vec3f start, vec3f dir, bvh_query_t* ret) { *ret = wp::bvh_query_ray(id, start, dir); }
 WP_API void builtin_bvh_query_next_bvh_query_t_int32(bvh_query_t query, int32 index, bool* ret) { *ret = wp::bvh_query_next(query, index); }
 WP_API void builtin_mesh_query_point_uint64_vec3f_float32_float32_int32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_point(id, point, max_dist, inside, face, bary_u, bary_v); }
+WP_API void builtin_mesh_query_point_uint64_vec3f_float32(uint64 id, vec3f point, float32 max_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_point(id, point, max_dist); }
 WP_API void builtin_mesh_query_point_no_sign_uint64_vec3f_float32_int32_float32_float32(uint64 id, vec3f point, float32 max_dist, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_point_no_sign(id, point, max_dist, face, bary_u, bary_v); }
+WP_API void builtin_mesh_query_point_no_sign_uint64_vec3f_float32(uint64 id, vec3f point, float32 max_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_no_sign(id, point, max_dist); }
 WP_API void builtin_mesh_query_furthest_point_no_sign_uint64_vec3f_float32_int32_float32_float32(uint64 id, vec3f point, float32 min_dist, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_furthest_point_no_sign(id, point, min_dist, face, bary_u, bary_v); }
+WP_API void builtin_mesh_query_furthest_point_no_sign_uint64_vec3f_float32(uint64 id, vec3f point, float32 min_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_furthest_point_no_sign(id, point, min_dist); }
 WP_API void builtin_mesh_query_point_sign_normal_uint64_vec3f_float32_float32_int32_float32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, float32 epsilon, bool* ret) { *ret = wp::mesh_query_point_sign_normal(id, point, max_dist, inside, face, bary_u, bary_v, epsilon); }
+WP_API void builtin_mesh_query_point_sign_normal_uint64_vec3f_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 epsilon, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_sign_normal(id, point, max_dist, epsilon); }
 WP_API void builtin_mesh_query_point_sign_winding_number_uint64_vec3f_float32_float32_int32_float32_float32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, float32 accuracy, float32 threshold, bool* ret) { *ret = wp::mesh_query_point_sign_winding_number(id, point, max_dist, inside, face, bary_u, bary_v, accuracy, threshold); }
+WP_API void builtin_mesh_query_point_sign_winding_number_uint64_vec3f_float32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 accuracy, float32 threshold, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_sign_winding_number(id, point, max_dist, accuracy, threshold); }
 WP_API void builtin_mesh_query_ray_uint64_vec3f_vec3f_float32_float32_float32_float32_float32_vec3f_int32(uint64 id, vec3f start, vec3f dir, float32 max_t, float32 t, float32 bary_u, float32 bary_v, float32 sign, vec3f normal, int32 face, bool* ret) { *ret = wp::mesh_query_ray(id, start, dir, max_t, t, bary_u, bary_v, sign, normal, face); }
+WP_API void builtin_mesh_query_ray_uint64_vec3f_vec3f_float32(uint64 id, vec3f start, vec3f dir, float32 max_t, mesh_query_ray_t* ret) { *ret = wp::mesh_query_ray(id, start, dir, max_t); }
 WP_API void builtin_mesh_query_aabb_uint64_vec3f_vec3f(uint64 id, vec3f lower, vec3f upper, mesh_query_aabb_t* ret) { *ret = wp::mesh_query_aabb(id, lower, upper); }
 WP_API void builtin_mesh_query_aabb_next_mesh_query_aabb_t_int32(mesh_query_aabb_t query, int32 index, bool* ret) { *ret = wp::mesh_query_aabb_next(query, index); }
 WP_API void builtin_mesh_eval_position_uint64_int32_float32_float32(uint64 id, int32 face, float32 bary_u, float32 bary_v, vec3f* ret) { *ret = wp::mesh_eval_position(id, face, bary_u, bary_v); }
diff --git a/warp/native/mesh.h b/warp/native/mesh.h
index b3febe53c..c536ad132 100644
--- a/warp/native/mesh.h
+++ b/warp/native/mesh.h
@@ -1181,7 +1181,7 @@ CUDA_CALLABLE inline bool mesh_query_point_sign_winding_number(uint64_t id, cons
     }
 }
 
-CUDA_CALLABLE inline void adj_mesh_query_point_no_sign(uint64_t id, const vec3& point, float max_dist, int& face, float& u, float& v, 
+CUDA_CALLABLE inline void adj_mesh_query_point_no_sign(uint64_t id, const vec3& point, float max_dist, const int& face, const float& u, const float& v,
                                                uint64_t adj_id, vec3& adj_point, float& adj_max_dist, int& adj_face, float& adj_u, float& adj_v, bool& adj_ret)
 {
     Mesh mesh = mesh_get(id);
@@ -1202,7 +1202,7 @@ CUDA_CALLABLE inline void adj_mesh_query_point_no_sign(uint64_t id, const vec3&
     adj_closest_point_to_triangle(p, q, r, point, adj_p, adj_q, adj_r, adj_point, adj_uv);	
 }
 
-CUDA_CALLABLE inline void adj_mesh_query_furthest_point_no_sign(uint64_t id, const vec3& point, float min_dist, int& face, float& u, float& v, 
+CUDA_CALLABLE inline void adj_mesh_query_furthest_point_no_sign(uint64_t id, const vec3& point, float min_dist, const int& face, const float& u, const float& v,
                                                uint64_t adj_id, vec3& adj_point, float& adj_min_dist, int& adj_face, float& adj_u, float& adj_v, bool& adj_ret)
 {
     Mesh mesh = mesh_get(id);
@@ -1223,24 +1223,116 @@ CUDA_CALLABLE inline void adj_mesh_query_furthest_point_no_sign(uint64_t id, con
     adj_closest_point_to_triangle(p, q, r, point, adj_p, adj_q, adj_r, adj_point, adj_uv);	 // Todo for Miles :>
 }
 
-CUDA_CALLABLE inline void adj_mesh_query_point(uint64_t id, const vec3& point, float max_dist, float& inside, int& face, float& u, float& v,
+CUDA_CALLABLE inline void adj_mesh_query_point(uint64_t id, const vec3& point, float max_dist, const float& inside, const int& face, const float& u, const float& v,
                                                uint64_t adj_id, vec3& adj_point, float& adj_max_dist, float& adj_inside, int& adj_face, float& adj_u, float& adj_v, bool& adj_ret)
 {
     adj_mesh_query_point_no_sign(id, point, max_dist, face, u, v, adj_id, adj_point, adj_max_dist, adj_face, adj_u, adj_v, adj_ret);
 }
 
-CUDA_CALLABLE inline void adj_mesh_query_point_sign_normal(uint64_t id, const vec3& point, float max_dist, float& inside, int& face, float& u, float& v, const float epsilon,
+CUDA_CALLABLE inline void adj_mesh_query_point_sign_normal(uint64_t id, const vec3& point, float max_dist, const float& inside, const int& face, const float& u, const float& v, const float epsilon,
                                                uint64_t adj_id, vec3& adj_point, float& adj_max_dist, float& adj_inside, int& adj_face, float& adj_u, float& adj_v, float& adj_epsilon, bool& adj_ret)
 {
     adj_mesh_query_point_no_sign(id, point, max_dist, face, u, v, adj_id, adj_point, adj_max_dist, adj_face, adj_u, adj_v, adj_ret);
 }
 
-CUDA_CALLABLE inline void adj_mesh_query_point_sign_winding_number(uint64_t id, const vec3& point, float max_dist, float& inside, int& face, float& u, float& v, const float accuracy, const float winding_number_threshold,
+CUDA_CALLABLE inline void adj_mesh_query_point_sign_winding_number(uint64_t id, const vec3& point, float max_dist, const float& inside, const int& face, const float& u, const float& v, const float accuracy, const float winding_number_threshold,
                                                uint64_t adj_id, vec3& adj_point, float& adj_max_dist, float& adj_inside, int& adj_face, float& adj_u, float& adj_v, float& adj_accuracy, float& adj_winding_number_threshold, bool& adj_ret)
 {
     adj_mesh_query_point_no_sign(id, point, max_dist, face, u, v, adj_id, adj_point, adj_max_dist, adj_face, adj_u, adj_v, adj_ret);
 }
 
+
+// Stores the result of querying the closest point on a mesh.
+struct mesh_query_point_t
+{
+    CUDA_CALLABLE mesh_query_point_t()
+    {
+    }
+
+    CUDA_CALLABLE mesh_query_point_t(int)
+    {
+        // For backward pass.
+    }
+
+    bool result;
+    float sign;
+    int face;
+    float u;
+    float v;
+};
+
+CUDA_CALLABLE inline mesh_query_point_t mesh_query_point(uint64_t id, const vec3& point, float max_dist)
+{
+    mesh_query_point_t query;
+    query.result = mesh_query_point(id, point, max_dist, query.sign, query.face, query.u, query.v);
+    return query;
+}
+
+CUDA_CALLABLE inline mesh_query_point_t mesh_query_point_no_sign(uint64_t id, const vec3& point, float max_dist)
+{
+    mesh_query_point_t query;
+    query.sign = 0.0;
+    query.result = mesh_query_point_no_sign(id, point, max_dist, query.face, query.u, query.v);
+    return query;
+}
+
+CUDA_CALLABLE inline mesh_query_point_t mesh_query_furthest_point_no_sign(uint64_t id, const vec3& point, float min_dist)
+{
+    mesh_query_point_t query;
+    query.sign = 0.0;
+    query.result = mesh_query_furthest_point_no_sign(id, point, min_dist, query.face, query.u, query.v);
+    return query;
+}
+
+CUDA_CALLABLE inline mesh_query_point_t mesh_query_point_sign_normal(uint64_t id, const vec3& point, float max_dist, const float epsilon = 1e-3f)
+{
+    mesh_query_point_t query;
+    query.result = mesh_query_point_sign_normal(id, point, max_dist, query.sign, query.face, query.u, query.v, epsilon);
+    return query;
+}
+
+CUDA_CALLABLE inline mesh_query_point_t mesh_query_point_sign_winding_number(uint64_t id, const vec3& point, float max_dist, float accuracy, float winding_number_threshold)
+{
+    mesh_query_point_t query;
+    query.result = mesh_query_point_sign_winding_number(id, point, max_dist, query.sign, query.face, query.u, query.v, accuracy, winding_number_threshold);
+    return query;
+}
+
+CUDA_CALLABLE inline void adj_mesh_query_point(uint64_t id, const vec3& point, float max_dist, const mesh_query_point_t& ret,
+                                               uint64_t adj_id, vec3& adj_point, float& adj_max_dist, mesh_query_point_t& adj_ret)
+{
+    adj_mesh_query_point(id, point, max_dist, ret.sign, ret.face, ret.u, ret.v,
+                         adj_id, adj_point, adj_max_dist, adj_ret.sign, adj_ret.face, adj_ret.u, adj_ret.v, adj_ret.result);
+}
+
+CUDA_CALLABLE inline void adj_mesh_query_point_no_sign(uint64_t id, const vec3& point, float max_dist, const mesh_query_point_t& ret,
+                                                       uint64_t adj_id, vec3& adj_point, float& adj_max_dist, mesh_query_point_t& adj_ret)
+{
+    adj_mesh_query_point_no_sign(id, point, max_dist, ret.face, ret.u, ret.v,
+                                adj_id, adj_point, adj_max_dist, adj_ret.face, adj_ret.u, adj_ret.v, adj_ret.result);
+}
+
+CUDA_CALLABLE inline void adj_mesh_query_furthest_point_no_sign(uint64_t id, const vec3& point, float min_dist, const mesh_query_point_t& ret,
+                                                                uint64_t adj_id, vec3& adj_point, float& adj_min_dist, mesh_query_point_t& adj_ret)
+{
+    adj_mesh_query_furthest_point_no_sign(id, point, min_dist, ret.face, ret.u, ret.v,
+                                          adj_id, adj_point, adj_min_dist, adj_ret.face, adj_ret.u, adj_ret.v, adj_ret.result);
+}
+
+CUDA_CALLABLE inline void adj_mesh_query_point_sign_normal(uint64_t id, const vec3& point, float max_dist, float epsilon, const mesh_query_point_t& ret,
+                                                           uint64_t adj_id, vec3& adj_point, float& adj_max_dist, float& adj_epsilon, mesh_query_point_t& adj_ret)
+{
+    adj_mesh_query_point_sign_normal(id, point, max_dist, ret.sign, ret.face, ret.u, ret.v, epsilon,
+                                     adj_id, adj_point, adj_max_dist, adj_ret.sign, adj_ret.face, adj_ret.u, adj_ret.v, epsilon, adj_ret.result);
+}
+
+CUDA_CALLABLE inline void adj_mesh_query_point_sign_winding_number(uint64_t id, const vec3& point, float max_dist, float accuracy, float winding_number_threshold, const mesh_query_point_t& ret,
+                                                                   uint64_t adj_id, vec3& adj_point, float& adj_max_dist, float& adj_accuracy, float& adj_winding_number_threshold, mesh_query_point_t& adj_ret)
+{
+    adj_mesh_query_point_sign_winding_number(id, point, max_dist, ret.sign, ret.face, ret.u, ret.v, accuracy, winding_number_threshold,
+                                             adj_id, adj_point, adj_max_dist, adj_ret.sign, adj_ret.face, adj_ret.u, adj_ret.v, adj_accuracy, adj_winding_number_threshold, adj_ret.result);
+}
+
 CUDA_CALLABLE inline bool mesh_query_ray(uint64_t id, const vec3& start, const vec3& dir, float max_t, float& t, float& u, float& v, float& sign, vec3& normal, int& face)
 {
     Mesh mesh = mesh_get(id);
@@ -1353,6 +1445,35 @@ CUDA_CALLABLE inline void adj_mesh_query_ray(
 }
 
 
+// Stores the result of querying the closest point on a mesh.
+struct mesh_query_ray_t
+{
+    CUDA_CALLABLE mesh_query_ray_t()
+    {
+    }
+
+    CUDA_CALLABLE mesh_query_ray_t(int)
+    {
+        // For backward pass.
+    }
+
+    bool result;
+    float sign;
+    int face;
+    float t;
+    float u;
+    float v;
+    vec3 normal;
+};
+
+CUDA_CALLABLE inline mesh_query_ray_t mesh_query_ray(uint64_t id, const vec3& start, const vec3& dir, float max_t)
+{
+    mesh_query_ray_t query;
+    query.result = mesh_query_ray(id, start, dir, max_t, query.t, query.u, query.v, query.sign, query.normal, query.face);
+    return query;
+}
+
+
 // determine if a point is inside (ret < 0 ) or outside the mesh (ret > 0)
 CUDA_CALLABLE inline float mesh_query_inside(uint64_t id, const vec3& p)
 {
diff --git a/warp/tests/test_mesh_query_point.py b/warp/tests/test_mesh_query_point.py
index b04f81b33..9650287ed 100644
--- a/warp/tests/test_mesh_query_point.py
+++ b/warp/tests/test_mesh_query_point.py
@@ -44,6 +44,12 @@ def sample_mesh_query(
     query_faces[tid] = face_index
     query_dist[tid] = wp.length(cp - p)
 
+    query = wp.mesh_query_point(mesh, p, max_dist)
+    wp.expect_eq(query.sign, sign)
+    wp.expect_eq(query.face, face_index)
+    wp.expect_eq(query.u, face_u)
+    wp.expect_eq(query.v, face_v)
+
 
 @wp.kernel
 def sample_mesh_query_no_sign(
@@ -69,6 +75,11 @@ def sample_mesh_query_no_sign(
     query_faces[tid] = face_index
     query_dist[tid] = wp.length(cp - p)
 
+    query = wp.mesh_query_point_no_sign(mesh, p, max_dist)
+    wp.expect_eq(query.face, face_index)
+    wp.expect_eq(query.u, face_u)
+    wp.expect_eq(query.v, face_v)
+
 
 @wp.kernel
 def sample_mesh_query_sign_normal(
@@ -97,6 +108,12 @@ def sample_mesh_query_sign_normal(
     query_faces[tid] = face_index
     query_dist[tid] = wp.length(cp - p)
 
+    query = wp.mesh_query_point_sign_normal(mesh, p, max_dist)
+    wp.expect_eq(query.sign, sign)
+    wp.expect_eq(query.face, face_index)
+    wp.expect_eq(query.u, face_u)
+    wp.expect_eq(query.v, face_v)
+
 
 @wp.kernel
 def sample_mesh_query_sign_winding_number(
@@ -125,6 +142,12 @@ def sample_mesh_query_sign_winding_number(
     query_faces[tid] = face_index
     query_dist[tid] = wp.length(cp - p)
 
+    query = wp.mesh_query_point_sign_winding_number(mesh, p, max_dist)
+    wp.expect_eq(query.sign, sign)
+    wp.expect_eq(query.face, face_index)
+    wp.expect_eq(query.u, face_u)
+    wp.expect_eq(query.v, face_v)
+
 
 @wp.func
 def triangle_closest_point(a: wp.vec3, b: wp.vec3, c: wp.vec3, p: wp.vec3):
@@ -574,6 +597,11 @@ def sample_furthest_points(mesh: wp.uint64, query_points: wp.array(dtype=wp.vec3
 
         query_result[tid] = wp.length_sq(p - closest)
 
+    query = wp.mesh_query_furthest_point_no_sign(mesh, p, 0.0)
+    wp.expect_eq(query.face, face)
+    wp.expect_eq(query.u, bary_u)
+    wp.expect_eq(query.v, bary_v)
+
 
 @wp.kernel
 def sample_furthest_points_brute(
diff --git a/warp/types.py b/warp/types.py
index 61e8ba27b..c3a419290 100644
--- a/warp/types.py
+++ b/warp/types.py
@@ -2979,6 +2979,37 @@ def allocate_by_tiles(
         return volume
 
 
+# definition just for kernel type (cannot be a parameter), see mesh.h
+# NOTE: its layout must match the corresponding struct defined in C.
+# NOTE: it needs to be defined after `indexedarray` to workaround a circular import issue.
+class mesh_query_point_t:
+    from warp.codegen import Var
+
+    vars = {
+        "result": Var("result", bool),
+        "sign": Var("sign", float32),
+        "face": Var("face", int32),
+        "u": Var("u", float32),
+        "v": Var("v", float32),
+    }
+
+
+# definition just for kernel type (cannot be a parameter), see mesh.h
+# NOTE: its layout must match the corresponding struct defined in C.
+class mesh_query_ray_t:
+    from warp.codegen import Var
+
+    vars = {
+        "result": Var("result", bool),
+        "sign": Var("sign", float32),
+        "face": Var("face", int32),
+        "t": Var("t", float32),
+        "u": Var("u", float32),
+        "v": Var("v", float32),
+        "normal": Var("normal", vec3),
+    }
+
+
 def matmul(
     a: array2d,
     b: array2d,
@@ -3957,7 +3988,7 @@ def infer_argument_types(args, template_types, arg_names=None):
             arg_types.append(arg._cls)
         # elif arg_type in [warp.types.launch_bounds_t, warp.types.shape_t, warp.types.range_t]:
         #     arg_types.append(arg_type)
-        # elif arg_type in [warp.hash_grid_query_t, warp.mesh_query_aabb_t, warp.bvh_query_t]:
+        # elif arg_type in [warp.hash_grid_query_t, warp.mesh_query_aabb_t, warp.mesh_query_point_t, warp.mesh_query_ray_t, warp.bvh_query_t]:
         #     arg_types.append(arg_type)
         elif arg is None:
             # allow passing None for arrays
@@ -3995,6 +4026,8 @@ def infer_argument_types(args, template_types, arg_names=None):
     launch_bounds_t: "lb",
     hash_grid_query_t: "hgq",
     mesh_query_aabb_t: "mqa",
+    mesh_query_point_t: "mqp",
+    mesh_query_ray_t: "mqr",
     bvh_query_t: "bvhq",
 }
 

From 172591ea003abc6a02b36ebe902a00738c585dae Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Tue, 5 Dec 2023 08:29:54 +1300
Subject: [PATCH 09/50] Add documentation for query types

---
 warp/builtins.py | 11 +++++------
 warp/context.py  |  8 ++++++++
 warp/types.py    | 33 +++++++++++++++++++++++++++++++++
 3 files changed, 46 insertions(+), 6 deletions(-)

diff --git a/warp/builtins.py b/warp/builtins.py
index d655dc819..a7ac4e132 100644
--- a/warp/builtins.py
+++ b/warp/builtins.py
@@ -1483,7 +1483,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
     value_type=bvh_query_t,
     group="Geometry",
     doc="""Construct an axis-aligned bounding box query against a BVH object. This query can be used to iterate over all bounds
-   inside a BVH. Returns an object that is used to track state during BVH traversal.
+   inside a BVH.
 
    :param id: The BVH identifier
    :param lower: The lower bound of the bounding box in BVH space
@@ -1496,7 +1496,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
     value_type=bvh_query_t,
     group="Geometry",
     doc="""Construct a ray query against a BVH object. This query can be used to iterate over all bounds
-   that intersect the ray. Returns an object that is used to track state during BVH traversal.
+   that intersect the ray.
 
    :param id: The BVH identifier
    :param start: The start of the ray in BVH space
@@ -1786,7 +1786,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
     },
     value_type=builtins.bool,
     group="Geometry",
-    doc="""Computes the closest ray hit on the :class:`Mesh` with identifier ``id``, returns ``True`` if a point < ``max_t`` is found.
+    doc="""Computes the closest ray hit on the :class:`Mesh` with identifier ``id``, returns ``True`` if a hit < ``max_t`` is found.
 
    :param id: The mesh identifier
    :param start: The start point of the ray
@@ -1795,7 +1795,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param t: Returns the distance of the closest hit along the ray
    :param bary_u: Returns the barycentric u coordinate of the closest hit
    :param bary_v: Returns the barycentric v coordinate of the closest hit
-   :param sign: Returns a value > 0 if the hit ray hit front of the face, returns < 0 otherwise
+   :param sign: Returns a value > 0 if the ray hit in front of the face, returns < 0 otherwise
    :param normal: Returns the face normal
    :param face: Returns the index of the hit face""",
 )
@@ -1826,7 +1826,6 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
     group="Geometry",
     doc="""Construct an axis-aligned bounding box query against a :class:`Mesh`.
    This query can be used to iterate over all triangles inside a volume.
-   Returns an object that is used to track state during mesh traversal.
 
    :param id: The mesh identifier
    :param lower: The lower bound of the bounding box in mesh space
@@ -1864,7 +1863,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
     value_type=hash_grid_query_t,
     group="Geometry",
     doc="Construct a point query against a :class:`HashGrid`. This query can be used to iterate over all neighboring points "
-    "within a fixed radius from the query point. Returns an object that is used to track state during neighbor traversal.",
+    "within a fixed radius from the query point.",
 )
 
 add_builtin(
diff --git a/warp/context.py b/warp/context.py
index a9688a56a..c1ccc20ee 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -4036,6 +4036,14 @@ def print_builtins(file):  # pragma: no cover
     print(".. class:: Transformation", file=file)
     print(".. class:: Array", file=file)
 
+    print("\nQuery Types", file=file)
+    print("-------------", file=file)
+    print(".. autoclass:: bvh_query_t", file=file)
+    print(".. autoclass:: hash_grid_query_t", file=file)
+    print(".. autoclass:: mesh_query_aabb_t", file=file)
+    print(".. autoclass:: mesh_query_point_t", file=file)
+    print(".. autoclass:: mesh_query_ray_t", file=file)
+
     # build dictionary of all functions by group
     groups = {}
 
diff --git a/warp/types.py b/warp/types.py
index c3a419290..3e056174b 100644
--- a/warp/types.py
+++ b/warp/types.py
@@ -910,18 +910,21 @@ def __init__(self):
 
 # definition just for kernel type (cannot be a parameter), see bvh.h
 class bvh_query_t:
+    """Object used to track state during BVH traversal."""
     def __init__(self):
         pass
 
 
 # definition just for kernel type (cannot be a parameter), see mesh.h
 class mesh_query_aabb_t:
+    """Object used to track state during mesh traversal."""
     def __init__(self):
         pass
 
 
 # definition just for kernel type (cannot be a parameter), see hash_grid.h
 class hash_grid_query_t:
+    """Object used to track state during neighbor traversal."""
     def __init__(self):
         pass
 
@@ -2983,6 +2986,22 @@ def allocate_by_tiles(
 # NOTE: its layout must match the corresponding struct defined in C.
 # NOTE: it needs to be defined after `indexedarray` to workaround a circular import issue.
 class mesh_query_point_t:
+    """Output for the mesh query point functions.
+
+    Attributes:
+        result (bool): Whether a point is found within the given constraints.
+        sign (float32): A value < 0 if query point is inside the mesh, >=0 otherwise.
+                        Note that mesh must be watertight for this to be robust
+        face (int32): Index of the closest face.
+        u (float32): Barycentric u coordinate of the closest point.
+        v (float32): Barycentric v coordinate of the closest point.
+
+    See Also:
+        :func:`mesh_query_point`, :func:`mesh_query_point_no_sign`,
+        :func:`mesh_query_furthest_point_no_sign`,
+        :func:`mesh_query_point_sign_normal`,
+        and :func:`mesh_query_point_sign_winding_number`.
+    """
     from warp.codegen import Var
 
     vars = {
@@ -2997,6 +3016,20 @@ class mesh_query_point_t:
 # definition just for kernel type (cannot be a parameter), see mesh.h
 # NOTE: its layout must match the corresponding struct defined in C.
 class mesh_query_ray_t:
+    """Output for the mesh query ray functions.
+
+    Attributes:
+        result (bool): Whether a hit is found within the given constraints.
+        sign (float32): A value > 0 if the ray hit in front of the face, returns < 0 otherwise.
+        face (int32): Index of the closest face.
+        t (float32): Distance of the closest hit along the ray.
+        u (float32): Barycentric u coordinate of the closest hit.
+        v (float32): Barycentric v coordinate of the closest hit.
+        normal (vec3f): Face normal.
+
+    See Also:
+        :func:`mesh_query_ray`.
+    """
     from warp.codegen import Var
 
     vars = {

From bde240e0ade3f258b1b0c84e1022ba1d1b45f8a9 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Tue, 5 Dec 2023 08:36:21 +1300
Subject: [PATCH 10/50] Hide the previous mesh query builtins

---
 warp/builtins.py | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/warp/builtins.py b/warp/builtins.py
index a7ac4e132..c7eb52014 100644
--- a/warp/builtins.py
+++ b/warp/builtins.py
@@ -1539,6 +1539,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param face: Returns the index of the closest face
    :param bary_u: Returns the barycentric u coordinate of the closest point
    :param bary_v: Returns the barycentric v coordinate of the closest point""",
+    hidden=True,
 )
 
 add_builtin(
@@ -1584,6 +1585,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param face: Returns the index of the closest face
    :param bary_u: Returns the barycentric u coordinate of the closest point
    :param bary_v: Returns the barycentric v coordinate of the closest point""",
+    hidden=True,
 )
 
 add_builtin(
@@ -1627,6 +1629,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param face: Returns the index of the furthest face
    :param bary_u: Returns the barycentric u coordinate of the furthest point
    :param bary_v: Returns the barycentric v coordinate of the furthest point""",
+    hidden=True,
 )
 
 add_builtin(
@@ -1679,6 +1682,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param bary_v: Returns the barycentric v coordinate of the closest point
    :param epsilon: Epsilon treating distance values as equal, when locating the minimum distance vertex/face/edge, as a
                    fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3""",
+    hidden=True,
 )
 
 add_builtin(
@@ -1740,6 +1744,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param bary_v: Returns the barycentric v coordinate of the closest point
    :param accuracy: Accuracy for computing the winding number with fast winding number method utilizing second-order dipole approximation, default 2.0
    :param threshold: The threshold of the winding number to be considered inside, default 0.5""",
+    hidden=True,
 )
 
 add_builtin(
@@ -1798,6 +1803,7 @@ def spatial_vector_constructor_value_func(arg_types, kwds, templates):
    :param sign: Returns a value > 0 if the ray hit in front of the face, returns < 0 otherwise
    :param normal: Returns the face normal
    :param face: Returns the index of the hit face""",
+    hidden=True,
 )
 
 add_builtin(

From 14d1f09fd06b09ab349d36e3856bba8393899242 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Mon, 4 Dec 2023 16:15:20 -0500
Subject: [PATCH 11/50] Add a backward test for returning a reference

Currently `test_grad()` fails when `lookup3()` directly returns a
reference to an array element.
---
 warp/tests/test_lvalue.py | 54 ++++++++++++++++++++++++++++++++++++++-
 1 file changed, 53 insertions(+), 1 deletion(-)

diff --git a/warp/tests/test_lvalue.py b/warp/tests/test_lvalue.py
index 789bf44dc..53fec71bd 100644
--- a/warp/tests/test_lvalue.py
+++ b/warp/tests/test_lvalue.py
@@ -1,4 +1,5 @@
 import warp as wp
+import numpy as np
 from warp.tests.test_base import *
 
 import unittest
@@ -88,6 +89,56 @@ def test_lookup(test, device):
             raise AssertionError(f"Unexpected result, got: {f} expected: {1}")
 
 
+@wp.func
+def lookup3(foos: wp.array(dtype=wp.float32), index: int):
+    # FIXME: return foos[index]
+    x = foos[index]
+    return x
+
+
+@wp.kernel
+def grad_kernel(foos: wp.array(dtype=wp.float32), bars: wp.array(dtype=wp.float32)):
+    i = wp.tid()
+
+    x = lookup3(foos, i)
+    bars[i] = x * wp.float32(i) + 1.0
+
+
+def test_grad(test, device):
+    num = 10
+    data = np.linspace(20, 20 + num, num, endpoint=False, dtype=np.float32)
+    input = wp.array(data, device=device, requires_grad=True)
+    output = wp.zeros(num, dtype=wp.float32, device=device)
+
+    ones = wp.array(np.ones(len(output)), dtype=wp.float32, device=device)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel=grad_kernel,
+            dim=(num,),
+            inputs=[input],
+            outputs=[output],
+            device=device,
+        )
+
+    tape.backward(grads={output: ones})
+
+    wp.synchronize()
+
+    # test forward results
+    for i, f in enumerate(output.list()):
+        expected = data[i] * i + 1
+        if f != expected:
+            raise AssertionError(f"Unexpected result, got: {f} expected: {expected}")
+
+    # test backward results
+    for i, f in enumerate(tape.gradients[input].list()):
+        expected = i
+        if f != expected:
+            raise AssertionError(f"Unexpected result, got: {f} expected: {expected}")
+
+
 @wp.func
 def lookup2(foos: wp.array(dtype=wp.uint32), index: int):
     if index % 2 == 0:
@@ -102,7 +153,7 @@ def lookup2(foos: wp.array(dtype=wp.uint32), index: int):
 def lookup2_kernel(foos: wp.array(dtype=wp.uint32)):
     i = wp.tid()
 
-    x = lookup(foos, i)
+    x = lookup2(foos, i)
     foos[i] = x + wp.uint32(1)
 
 
@@ -425,6 +476,7 @@ class TestLValue(parent):
     add_function_test(TestLValue, "test_rmw_array_struct", test_rmw_array_struct, devices=devices)
     add_function_test(TestLValue, "test_lookup", test_lookup, devices=devices)
     add_function_test(TestLValue, "test_lookup2", test_lookup2, devices=devices)
+    add_function_test(TestLValue, "test_grad", test_grad, devices=devices)
     add_function_test(TestLValue, "test_unary", test_unary, devices=devices)
     add_function_test(TestLValue, "test_rvalue", test_rvalue, devices=devices)
     add_function_test(TestLValue, "test_intermediate", test_intermediate, devices=devices)

From ceb4dbe9b0d72e542ad03e0420ab92b65f903fdf Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Mon, 4 Dec 2023 16:50:54 -0500
Subject: [PATCH 12/50] Fix backpropagation for returning array element
 references

In the forward pass returning a reference would result in a `load()`
operation, but this isn't differentiable. By instead emitting a `copy()`
into a temporary variable, we can reuse the value from the forward pass
in the backward pass.
---
 warp/codegen.py           | 9 ++++++---
 warp/tests/test_lvalue.py | 4 +---
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index f1db21796..30ab0fdb4 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -967,7 +967,9 @@ def add_call(adj, func, args, min_outputs=None, templates=[], kwds=None):
             adj.add_forward(forward_call, replay=replay_call)
 
         if not func.missing_grad and len(args):
-            reverse_has_output_args = (func.require_original_output_arg or len(output_list) > 1) and func.custom_grad_func is None
+            reverse_has_output_args = (
+                func.require_original_output_arg or len(output_list) > 1
+            ) and func.custom_grad_func is None
             arg_str = adj.format_reverse_call_args(
                 args_var,
                 args,
@@ -1903,7 +1905,8 @@ def emit_Return(adj, node):
         if var is not None:
             adj.return_var = tuple()
             for ret in var:
-                ret = adj.load(ret)
+                if is_reference(ret.type):
+                    ret = adj.add_builtin_call("copy", [ret])
                 adj.return_var += (ret,)
 
         adj.add_return(adj.return_var)
@@ -1947,7 +1950,7 @@ def emit_Pass(adj, node):
         ast.AugAssign: emit_AugAssign,
         ast.Tuple: emit_Tuple,
         ast.Pass: emit_Pass,
-        ast.Ellipsis: emit_Ellipsis
+        ast.Ellipsis: emit_Ellipsis,
     }
 
     def eval(adj, node):
diff --git a/warp/tests/test_lvalue.py b/warp/tests/test_lvalue.py
index 53fec71bd..0eb7471c4 100644
--- a/warp/tests/test_lvalue.py
+++ b/warp/tests/test_lvalue.py
@@ -91,9 +91,7 @@ def test_lookup(test, device):
 
 @wp.func
 def lookup3(foos: wp.array(dtype=wp.float32), index: int):
-    # FIXME: return foos[index]
-    x = foos[index]
-    return x
+    return foos[index]
 
 
 @wp.kernel

From cb6e90c75ea612f6b13b0bff77a07103d5198059 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 09:19:25 -0500
Subject: [PATCH 13/50] Copy linux-x86_64 CI build script to linux-aarch64

Just a verbatim copy as a starting point. `step.sh` is based on a very
outdated script. It's contents are added to `build.sh` as comment just
for reference.
---
 .../ci/building/build-linux-aarch64/build.sh  | 73 +++++++++++++++++++
 tools/ci/building/build-linux-aarch64/step.sh | 22 ------
 2 files changed, 73 insertions(+), 22 deletions(-)
 create mode 100755 tools/ci/building/build-linux-aarch64/build.sh
 delete mode 100755 tools/ci/building/build-linux-aarch64/step.sh

diff --git a/tools/ci/building/build-linux-aarch64/build.sh b/tools/ci/building/build-linux-aarch64/build.sh
new file mode 100755
index 000000000..013e990a5
--- /dev/null
+++ b/tools/ci/building/build-linux-aarch64/build.sh
@@ -0,0 +1,73 @@
+##!/usr/bin/env bash
+#
+#set -e
+#
+#SCRIPT_DIR="$(dirname "${BASH_SOURCE}")"
+#
+## Verify formatting
+## "$SCRIPT_DIR/../../../../format_code.sh" --verify
+#
+## Full rebuild
+#"$SCRIPT_DIR/../../../../build.sh" -x --platform-target linux-aarch64
+#
+## Package all
+#"$SCRIPT_DIR/../../../../tools/package.sh" -a -c release --platform-target linux-aarch64
+#
+## Package all
+#"$SCRIPT_DIR/../../../../tools/package.sh" -a -c debug --platform-target linux-aarch64
+#
+## publish artifacts to teamcity
+#echo "##teamcity[publishArtifacts '_build/packages']"
+
+#!/bin/bash
+set -e
+
+USE_LINBUILD=1
+BUILD_MODE="release"
+
+# scan command line for options
+for arg; do
+    shift
+    case $arg in
+    --no-docker)
+        USE_LINBUILD=0
+        ;;
+    --debug)
+        BUILD_MODE="debug"
+        ;;
+    esac
+
+    # keep all options (including --no-docker) to pass to repo.sh
+    if [[ $arg != "--debug" ]]; then
+        set -- "$@" "$arg"
+    fi 
+done
+
+SCRIPT_DIR=$(dirname ${BASH_SOURCE})
+#source "$SCRIPT_DIR/repo.sh" build --fetch-only $@ || exit $?
+
+"$SCRIPT_DIR/../../../../repo.sh" build --fetch-only $@ 
+
+PYTHON="$SCRIPT_DIR/../../../../_build/target-deps/python/python"
+LINBUILD="$SCRIPT_DIR/../../../../_build/host-deps/linbuild/linbuild.sh"
+CUDA="$SCRIPT_DIR/../../../../_build/target-deps/cuda"
+
+# pip deps
+$PYTHON -m pip install --upgrade pip
+$PYTHON -m pip install numpy
+$PYTHON -m pip install gitpython
+$PYTHON -m pip install cmake
+$PYTHON -m pip install ninja
+
+if [[ "$OSTYPE" == "darwin"* ]]; then
+    $PYTHON "$SCRIPT_DIR/../../../../build_lib.py"
+else
+    if [ ${USE_LINBUILD} -ne 0 ]; then
+        # build with docker for increased compatibility
+        $LINBUILD -- $PYTHON "$SCRIPT_DIR/../../../../build_lib.py" --cuda_path=$CUDA --mode=$BUILD_MODE
+    else
+        # build without docker
+        $PYTHON "$SCRIPT_DIR/../../../../build_lib.py" --cuda_path=$CUDA --mode=$BUILD_MODE
+    fi
+fi
+
diff --git a/tools/ci/building/build-linux-aarch64/step.sh b/tools/ci/building/build-linux-aarch64/step.sh
deleted file mode 100755
index 3ea1d4b82..000000000
--- a/tools/ci/building/build-linux-aarch64/step.sh
+++ /dev/null
@@ -1,22 +0,0 @@
-#!/usr/bin/env bash
-
-set -e
-
-SCRIPT_DIR="$(dirname "${BASH_SOURCE}")"
-
-# Verify formatting
-# "$SCRIPT_DIR/../../../../format_code.sh" --verify
-
-# Full rebuild
-"$SCRIPT_DIR/../../../../build.sh" -x --platform-target linux-aarch64
-
-# Package all
-"$SCRIPT_DIR/../../../../tools/package.sh" -a -c release --platform-target linux-aarch64
-
-# Package all
-"$SCRIPT_DIR/../../../../tools/package.sh" -a -c debug --platform-target linux-aarch64
-
-# publish artifacts to teamcity
-echo "##teamcity[publishArtifacts '_build/packages']"
-
-

From bb57da3d785d2ec7483a8d00353f2eceeb8a952a Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 09:39:34 -0500
Subject: [PATCH 14/50] Support building LLVM for AArch64 on non-Apple systems

---
 build_llvm.py | 25 +++++++++++++++++++++++--
 1 file changed, 23 insertions(+), 2 deletions(-)

diff --git a/build_llvm.py b/build_llvm.py
index ed4f21b04..f916eae76 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -1,6 +1,7 @@
 import os
 import subprocess
 import sys
+import platform
 
 import warp
 from warp.build_dll import build_dll_for_arch, run_cmd
@@ -14,6 +15,18 @@
 llvm_install_path = os.path.join(llvm_project_path, "out/install/")
 
 
+# return a canonical machine architecture string
+# - "x86_64" for x86-64, aka. AMD64, aka. x64
+# - "aarch64" for AArch64, aka. ARM64
+def machine_architecture() -> str:
+    machine = platform.machine()
+    if machine == "x86_64" or machine == "AMD64":
+        return "x86_64"
+    if machine == "aarch64" or machine == "arm64":
+        return "aarch64"
+    raise RuntimeError(f"Unrecognized machine architecture {machine}")
+
+
 # Fetch prebuilt Clang/LLVM libraries
 def fetch_prebuilt_libraries():
     if os.name == "nt":
@@ -282,10 +295,18 @@ def build_from_source(args):
     else:
         llvm_source = llvm_project_path
 
-    build_from_source_for_arch(args, "x86_64", llvm_source)
+    # build for the machine's architecture
+    if machine_architecture() == "x86_64":
+        build_from_source_for_arch(args, "x86_64", llvm_source)
+    else:
+        build_from_source_for_arch(args, "arm64", llvm_source)
 
+    # for Apple systems also cross-compile for building a universal binary
     if sys.platform == "darwin":
-        build_from_source_for_arch(args, "arm64", llvm_source)
+        if machine_architecture() == "x86_64":
+            build_from_source_for_arch(args, "arm64", llvm_source)
+        else:
+            build_from_source_for_arch(args, "x86_64", llvm_source)
 
 
 # build warp-clang.dll

From 4a28d93ffdd190f237c64ac6b9dbc197fb8b2d28 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 11:04:13 -0500
Subject: [PATCH 15/50] Fix LLVM build of BLAKE3 for non-OSX AArch64

LLVM 15's build files for the BLAKE3 hash functions assume that if
`CMAKE_OSX_ARCHITECTURES` is defined we want to compile the x86 SSE/AVX
optimized version. This depends on having a suitable cross-architecture
assembler, which typically isn't available on Linux AArch64.
---
 build_llvm.py | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/build_llvm.py b/build_llvm.py
index f916eae76..7c453cc07 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -102,14 +102,17 @@ def build_from_source_for_arch(args, arch, llvm_source):
 
     if sys.platform == "darwin":
         host_triple = f"{arch}-apple-macos11"
+        osx_architectures = arch  # build one architecture only
     elif os.name == "nt":
         host_triple = f"{arch}-pc-windows"
+        osx_architectures = ""
     else:
         host_triple = f"{arch}-pc-linux"
+        osx_architectures = ""
 
     llvm_path = os.path.join(llvm_source, "llvm")
-    build_path = os.path.join(llvm_build_path, f"{warp.config.mode}-{ arch}")
-    install_path = os.path.join(llvm_install_path, f"{warp.config.mode}-{ arch}")
+    build_path = os.path.join(llvm_build_path, f"{warp.config.mode}-{arch}")
+    install_path = os.path.join(llvm_install_path, f"{warp.config.mode}-{arch}")
 
     # Build LLVM and Clang
     cmake_gen = [
@@ -143,7 +146,7 @@ def build_from_source_for_arch(args, arch, llvm_source):
         "-D", "CMAKE_CXX_FLAGS=-D_GLIBCXX_USE_CXX11_ABI=0",  # The pre-C++11 ABI is still the default on the CentOS 7 toolchain
         "-D", f"CMAKE_INSTALL_PREFIX={install_path}",
         "-D", f"LLVM_HOST_TRIPLE={host_triple}",
-        "-D", f"CMAKE_OSX_ARCHITECTURES={ arch}",
+        "-D", f"CMAKE_OSX_ARCHITECTURES={osx_architectures}",
 
         # Disable unused tools and features
         "-D", "CLANG_BUILD_TOOLS=FALSE",

From 2156cc2b4551e2eb9ba4dcfdeed2a6be410b1d72 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 11:39:03 -0500
Subject: [PATCH 16/50] Support building warp-clang.so for AArch64 on
 non-Darwin platforms

---
 build_llvm.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/build_llvm.py b/build_llvm.py
index 7c453cc07..8ca188b2e 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -380,4 +380,7 @@ def build_warp_clang(args, lib_name):
         os.remove(f"{dylib_path}-arm64")
 
     else:
-        build_warp_clang_for_arch(args, lib_name, "x86_64")
+        if machine_architecture() == "x86_64":
+            build_warp_clang_for_arch(args, lib_name, "x86_64")
+        else:
+            build_warp_clang_for_arch(args, lib_name, "arm64")

From 04a9332ee14c3155b00df47bb8eb9f7d8be4e6f9 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 13:42:26 -0500
Subject: [PATCH 17/50] Remove deprecated Packman credentials configuration

The referenced repository has been removed. See
https://confluence.nvidia.com/pages/viewpage.action?pageId=1822616759
for details.

Instead we can now use https://omnipackages.nvidia.com/packages/publish
to manually upload packages created by
`tools/packman/packman pack <path>`.
---
 tools/packman/config.packman.xml | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tools/packman/config.packman.xml b/tools/packman/config.packman.xml
index 7402744d3..e72195910 100644
--- a/tools/packman/config.packman.xml
+++ b/tools/packman/config.packman.xml
@@ -1,7 +1,6 @@
 <config remotes="cloudfront urm">
     <remote2 name="cloudfront">
         <transport actions="download" protocol="https" packageLocation="d4i3qtqj3r0z5.cloudfront.net/${name}@${version}" />
-        <transport actions="list upload" configUrl="https://gitlab-master.nvidia.com/hfannar/packman-transports/raw/main/d4i3qtqj3r0z5.cloudfront.toml" />
     </remote2>
     <remote2 name="urm">
         <transport actions="download" protocol="https" packageLocation="urm.nvidia.com/artifactory/ct-omniverse-generic/pkgs/${name}/${name}@${version}" />

From df4c9ba7e1b1438919bde20c5d6dce90c31ef23a Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 13:48:47 -0500
Subject: [PATCH 18/50] Build warp-clang.so for Linux AArch64 from prebuilt
 LLVM libraries

---
 build_llvm.py                                 |  5 ++++-
 .../ci/building/build-linux-aarch64/build.sh  | 21 -------------------
 2 files changed, 4 insertions(+), 22 deletions(-)

diff --git a/build_llvm.py b/build_llvm.py
index 8ca188b2e..975f96e7b 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -40,7 +40,10 @@ def fetch_prebuilt_libraries():
                 "x86_64": "15.0.7-darwin-x86_64-macos11",
             }
         else:
-            packages = {"x86_64": "15.0.7-linux-x86_64-ptx-gcc7.5-cxx11abi0"}
+            packages = {
+                "arm64": "15.0.7-linux-aarch64-gcc7.5",
+                "x86_64": "15.0.7-linux-x86_64-ptx-gcc7.5-cxx11abi0",
+            }
 
     for arch in packages:
         subprocess.check_call(
diff --git a/tools/ci/building/build-linux-aarch64/build.sh b/tools/ci/building/build-linux-aarch64/build.sh
index 013e990a5..1087cf5dd 100755
--- a/tools/ci/building/build-linux-aarch64/build.sh
+++ b/tools/ci/building/build-linux-aarch64/build.sh
@@ -1,24 +1,3 @@
-##!/usr/bin/env bash
-#
-#set -e
-#
-#SCRIPT_DIR="$(dirname "${BASH_SOURCE}")"
-#
-## Verify formatting
-## "$SCRIPT_DIR/../../../../format_code.sh" --verify
-#
-## Full rebuild
-#"$SCRIPT_DIR/../../../../build.sh" -x --platform-target linux-aarch64
-#
-## Package all
-#"$SCRIPT_DIR/../../../../tools/package.sh" -a -c release --platform-target linux-aarch64
-#
-## Package all
-#"$SCRIPT_DIR/../../../../tools/package.sh" -a -c debug --platform-target linux-aarch64
-#
-## publish artifacts to teamcity
-#echo "##teamcity[publishArtifacts '_build/packages']"
-
 #!/bin/bash
 set -e
 

From 08f4d6b7fd551abe444c45a4018d35b7043d1372 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Tue, 5 Dec 2023 15:21:58 -0500
Subject: [PATCH 19/50] Copy the Linux x86-64 CI test script for AArch64

---
 tools/ci/testing/test-linux-aarch64/step.sh |  8 --------
 tools/ci/testing/test-linux-aarch64/test.sh | 19 +++++++++++++++++++
 2 files changed, 19 insertions(+), 8 deletions(-)
 delete mode 100755 tools/ci/testing/test-linux-aarch64/step.sh
 create mode 100755 tools/ci/testing/test-linux-aarch64/test.sh

diff --git a/tools/ci/testing/test-linux-aarch64/step.sh b/tools/ci/testing/test-linux-aarch64/step.sh
deleted file mode 100755
index 8268c323c..000000000
--- a/tools/ci/testing/test-linux-aarch64/step.sh
+++ /dev/null
@@ -1,8 +0,0 @@
-#!/usr/bin/env bash
-
-set -e
-
-SCRIPT_DIR="$(dirname "${BASH_SOURCE}")"
-
-# tests
-"$SCRIPT_DIR/../../../../repo.sh" test --config release --from-package
diff --git a/tools/ci/testing/test-linux-aarch64/test.sh b/tools/ci/testing/test-linux-aarch64/test.sh
new file mode 100755
index 000000000..44da31bd3
--- /dev/null
+++ b/tools/ci/testing/test-linux-aarch64/test.sh
@@ -0,0 +1,19 @@
+#!/bin/bash
+set -e
+
+SCRIPT_DIR=$(dirname ${BASH_SOURCE})
+"$SCRIPT_DIR/../../../../repo.sh" build --fetch-only $@
+
+PYTHON="$SCRIPT_DIR/../../../../_build/target-deps/python/python"
+
+echo "Installing test dependencies"
+#$PYTHON -m pip install matplotlib
+#$PYTHON -m pip install usd-core
+#$PYTHON -m pip install torch --extra-index-url https://download.pytorch.org/whl/cu113
+#$PYTHON -m pip install "jax[cuda]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+
+echo "Installing Warp to Python"
+$PYTHON -m pip install -e "$SCRIPT_DIR/../../../../."
+
+echo "Running tests"
+$PYTHON "$SCRIPT_DIR/../../../../warp/tests/test_all.py"

From 4a5697aaffe66b67b88bd54b37fa412fe8f67239 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Wed, 6 Dec 2023 16:10:36 -0500
Subject: [PATCH 20/50] Fix wp.synchronize() errors if CUDA is unavailable

If the Warp DLLs were built with CUDA enabled (i.e. the SDK is present),
but there is no runtime CUDA support, the `wp.synchronize()` call
previously produced an error:

`Warp CUDA error: Function cuCtxGetCurrent_f: a suitable driver entry point was not found`

This change properly checks whether the CUDA driver has been
successfully initialized.
---
 warp/native/cuda_util.cpp | 11 ++++++++---
 warp/native/cuda_util.h   |  1 +
 warp/native/warp.cu       |  7 +------
 3 files changed, 10 insertions(+), 9 deletions(-)

diff --git a/warp/native/cuda_util.cpp b/warp/native/cuda_util.cpp
index 3b07043dd..1448357af 100644
--- a/warp/native/cuda_util.cpp
+++ b/warp/native/cuda_util.cpp
@@ -89,6 +89,7 @@ static PFN_cuGraphicsResourceGetMappedPointer_v3020 pfn_cuGraphicsResourceGetMap
 static PFN_cuGraphicsGLRegisterBuffer_v3000 pfn_cuGraphicsGLRegisterBuffer;
 static PFN_cuGraphicsUnregisterResource_v3000 pfn_cuGraphicsUnregisterResource;
 
+static bool cuda_driver_initialized = false;
 
 bool ContextGuard::always_restore = false;
 
@@ -196,11 +197,15 @@ bool init_cuda_driver()
     get_driver_entry_point("cuGraphicsUnregisterResource", &(void*&)pfn_cuGraphicsUnregisterResource);
 
     if (pfn_cuInit)
-        return check_cu(pfn_cuInit(0));
-    else
-        return false;
+        cuda_driver_initialized = check_cu(pfn_cuInit(0));
+    
+    return cuda_driver_initialized;
 }
 
+bool is_cuda_driver_initialized()
+{
+    return cuda_driver_initialized;
+}
 
 bool check_cuda_result(cudaError_t code, const char* file, int line)
 {
diff --git a/warp/native/cuda_util.h b/warp/native/cuda_util.h
index 446109554..27acaff5d 100644
--- a/warp/native/cuda_util.h
+++ b/warp/native/cuda_util.h
@@ -83,6 +83,7 @@ CUresult cuGraphicsUnregisterResource_f(CUgraphicsResource resource);
 
 
 bool init_cuda_driver();
+bool is_cuda_driver_initialized();
 
 bool check_cuda_result(cudaError_t code, const char* file, int line);
 inline bool check_cuda_result(uint64_t code, const char* file, int line)
diff --git a/warp/native/warp.cu b/warp/native/warp.cu
index c48ba430b..9ae10de1d 100644
--- a/warp/native/warp.cu
+++ b/warp/native/warp.cu
@@ -1143,12 +1143,7 @@ int cuda_toolkit_version()
 
 bool cuda_driver_is_initialized()
 {
-    CUcontext ctx;
-
-    // result can be: CUDA_SUCCESS, CUDA_ERROR_DEINITIALIZED, CUDA_ERROR_NOT_INITIALIZED
-    CUresult result = cuCtxGetCurrent_f(&ctx);
-
-    return result == CUDA_SUCCESS;
+    return is_cuda_driver_initialized();
 }
 
 int nvrtc_supported_arch_count()

From 48ad930a8fcbfeb6755378b2618db04d504c75a0 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Wed, 6 Dec 2023 23:12:36 -0500
Subject: [PATCH 21/50] Upgrade to Python 3.9 for testing Linux AArch64 builds

PyPI currently only has binary wheels of numpy for Python 3.9+. While
there's also a source distribution, it fails to build properly on Linux
AArch64 due to dependencies that can't be met.
---
 deps/target-deps.packman.xml | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/deps/target-deps.packman.xml b/deps/target-deps.packman.xml
index 1c2f42321..2620a1ae4 100644
--- a/deps/target-deps.packman.xml
+++ b/deps/target-deps.packman.xml
@@ -23,10 +23,11 @@
    -->
 
 <dependency name="python" linkPath="../_build/target-deps/python" tags="slim-package">
-    <package name="python" version="3.7.9-windows-x86_64" platforms="windows-x86_64" />
+    <package name="python" version="3.7.9-windows-x86_64" platforms="windows-x86_64"/>
     <!-- https://teamcity.nvidia.com/project.html?projectId=Omniverse_Externals_Python -->
-    <package name="python" version="3.7.10-187.9b4bf28b-${platform}" platforms="linux-x86_64 linux-aarch64"/>
-    <package name="python" version="3.7.12+nv1-${platform}" platforms="macos-x86_64" />
+    <package name="python" version="3.7.10-187.9b4bf28b-${platform}" platforms="linux-x86_64"/>
+    <package name="python" version="3.9.16+nv1-${platform}" platforms="linux-aarch64"/>
+    <package name="python" version="3.7.12+nv1-${platform}" platforms="macos-x86_64"/>
 </dependency>   
 
   <dependency name="cuda" linkPath="../_build/target-deps/cuda">

From 79cd8c118786230e8c0cda573ae3f9d1ab0b65f8 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Thu, 7 Dec 2023 00:27:45 -0500
Subject: [PATCH 22/50] Skip testing FEM examples if USD is unavailable

The FEM examples depend on the `usd-core` package, which isn't available
for AArch64 at this time.
---
 warp/tests/test_examples.py | 55 ++++++++++++++++++++++---------------
 1 file changed, 33 insertions(+), 22 deletions(-)

diff --git a/warp/tests/test_examples.py b/warp/tests/test_examples.py
index dd0040edd..a87497841 100644
--- a/warp/tests/test_examples.py
+++ b/warp/tests/test_examples.py
@@ -82,28 +82,39 @@ class TestExamples(parent):
         add_example_test(TestExamples, name="example_wave", options={"resx": 256, "resy": 256})
         add_example_test(TestExamples, name="fem.example_diffusion_mgpu", options={"quiet": True, "num_frames": 1})
 
-    # The following examples do not need cuda
-    add_example_test(TestExamples, name="fem.example_apic_fluid", options={"quiet": True, "res": [16, 16, 16]})
-    add_example_test(
-        TestExamples,
-        name="fem.example_diffusion",
-        options={"quiet": True, "resolution": 10, "mesh": "tri", "num_frames": 1},
-    )
-    add_example_test(
-        TestExamples, name="fem.example_diffusion_3d", options={"quiet": True, "resolution": 10, "num_frames": 1}
-    )
-    add_example_test(
-        TestExamples,
-        name="fem.example_deformed_geometry",
-        options={"quiet": True, "resolution": 10, "num_frames": 1, "mesh": "tri"},
-    )
-    add_example_test(TestExamples, name="fem.example_convection_diffusion", options={"quiet": True, "resolution": 20})
-    add_example_test(
-        TestExamples,
-        name="fem.example_mixed_elasticity",
-        options={"quiet": True, "nonconforming_stresses": True, "mesh": "quad", "num_frames": 1},
-    )
-    add_example_test(TestExamples, name="fem.example_stokes_transfer", options={"quiet": True, "num_frames": 1})
+    # USD import failures should not count as a test failure
+    try:
+        from pxr import Usd, UsdGeom
+
+        have_usd = True
+    except:
+        have_usd = False
+
+    # The following examples do not need CUDA, but they need USD
+    if have_usd:
+        add_example_test(TestExamples, name="fem.example_apic_fluid", options={"quiet": True, "res": [16, 16, 16]})
+        add_example_test(
+            TestExamples,
+            name="fem.example_diffusion",
+            options={"quiet": True, "resolution": 10, "mesh": "tri", "num_frames": 1},
+        )
+        add_example_test(
+            TestExamples, name="fem.example_diffusion_3d", options={"quiet": True, "resolution": 10, "num_frames": 1}
+        )
+        add_example_test(
+            TestExamples,
+            name="fem.example_deformed_geometry",
+            options={"quiet": True, "resolution": 10, "num_frames": 1, "mesh": "tri"},
+        )
+        add_example_test(
+            TestExamples, name="fem.example_convection_diffusion", options={"quiet": True, "resolution": 20}
+        )
+        add_example_test(
+            TestExamples,
+            name="fem.example_mixed_elasticity",
+            options={"quiet": True, "nonconforming_stresses": True, "mesh": "quad", "num_frames": 1},
+        )
+        add_example_test(TestExamples, name="fem.example_stokes_transfer", options={"quiet": True, "num_frames": 1})
 
     return TestExamples
 

From c46c4c5712416480ddcbd1493ce75abb32b3b671 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Fri, 8 Dec 2023 15:18:16 -0500
Subject: [PATCH 23/50] Fix getting the source segment the AST was constructed
 from

`ast.get_source_segment(source, node)` expects that the `source` is the
same as the one the `node` was parsed from. `adj.raw_source` does not
include the `dedent()` operation that happens in `Adjoint`'s constructor
prior to AST construction.
---
 warp/codegen.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index 30ab0fdb4..7cb7ab885 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -2083,7 +2083,7 @@ def set_lineno(adj, lineno):
 
     def get_node_source(adj, node):
         # return the Python code corresponding to the given AST node
-        return ast.get_source_segment("".join(adj.raw_source), node)
+        return ast.get_source_segment(adj.source, node)
 
 
 # ----------------

From dd000d7f7e17af453802edd44adf4eee48d53e90 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Fri, 8 Dec 2023 15:56:13 -0500
Subject: [PATCH 24/50] Use only the source code used for AST parsing

`adj.raw_source` was only used for annotating the generated code with
the original Python code. It is identical to `adj.source` except the
latter is dedented and not split into lines yet. The original
indentation doesn't matter because it is stripped.

Other than eliminating `adj.raw_source` the code was restructured for
readability.
---
 warp/codegen.py | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index 7cb7ab885..d70553458 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -518,21 +518,16 @@ def __init__(
         # whether the generation of the adjoint code is skipped for this function
         adj.skip_reverse_codegen = skip_reverse_codegen
 
-        # build AST from function object
-        adj.source = inspect.getsource(func)
-
-        # get source code lines and line number where function starts
-        adj.raw_source, adj.fun_lineno = inspect.getsourcelines(func)
-
-        # keep track of line number in function code
-        adj.lineno = None
+        # extract name of source file
+        adj.filename = inspect.getsourcefile(func) or "unknown source file"
+        # get source file line number where function starts
+        _, adj.fun_lineno = inspect.getsourcelines(func)
 
+        # get function source code
+        adj.source = inspect.getsource(func)
         # ensures that indented class methods can be parsed as kernels
         adj.source = textwrap.dedent(adj.source)
 
-        # extract name of source file
-        adj.filename = inspect.getsourcefile(func) or "unknown source file"
-
         # build AST and apply node transformers
         adj.tree = ast.parse(adj.source)
         adj.transformers = transformers
@@ -541,6 +536,9 @@ def __init__(
 
         adj.fun_name = adj.tree.body[0].name
 
+        # for keeping track of line number in function code
+        adj.lineno = None
+
         # whether the forward code shall be used for the reverse pass and a custom
         # function signature is applied to the reverse version of the function
         adj.custom_reverse_mode = custom_reverse_mode
@@ -2076,7 +2074,7 @@ def resolve_static_expression(adj, root_node, eval_types=True):
     def set_lineno(adj, lineno):
         if adj.lineno is None or adj.lineno != lineno:
             line = lineno + adj.fun_lineno
-            source = adj.raw_source[lineno].strip().ljust(80 - len(adj.indentation), " ")
+            source = adj.source.splitlines()[lineno].strip().ljust(80 - len(adj.indentation), " ")
             adj.add_forward(f"// {source}       <L {line}>")
             adj.add_reverse(f"// adj: {source}  <L {line}>")
         adj.lineno = lineno

From af651ae72d2a22e1acbf06fc36adae083170e33d Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Fri, 8 Dec 2023 16:01:08 -0500
Subject: [PATCH 25/50] Cache line-splitted source code

This is an optimization which eliminates multiple occurrences where the
source code was being split into an array of lines.

Note this doesn't add new storage overhead because the previous commit
removed `adj.raw_source` which was the line-splitted undedented source.
---
 warp/codegen.py | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index d70553458..87b9e3316 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -528,6 +528,8 @@ def __init__(
         # ensures that indented class methods can be parsed as kernels
         adj.source = textwrap.dedent(adj.source)
 
+        adj.source_lines = adj.source.splitlines()
+
         # build AST and apply node transformers
         adj.tree = ast.parse(adj.source)
         adj.transformers = transformers
@@ -623,7 +625,7 @@ def build(adj, builder):
                 else:
                     msg = "Error"
                 lineno = adj.lineno + adj.fun_lineno
-                line = adj.source.splitlines()[adj.lineno]
+                line = adj.source_lines[adj.lineno]
                 msg += f' while parsing function "{adj.fun_name}" at {adj.filename}:{lineno}:\n{line}\n'
                 ex, data, traceback = sys.exc_info()
                 e = ex(";".join([msg] + [str(a) for a in data.args])).with_traceback(traceback)
@@ -1415,7 +1417,7 @@ def materialize_redefinitions(adj, symbols):
             if var1 != var2:
                 if warp.config.verbose and not adj.custom_reverse_mode:
                     lineno = adj.lineno + adj.fun_lineno
-                    line = adj.source.splitlines()[adj.lineno]
+                    line = adj.source_lines[adj.lineno]
                     msg = f'Warning: detected mutated variable {sym} during a dynamic for-loop in function "{adj.fun_name}" at {adj.filename}:{lineno}: this may not be a differentiable operation.\n{line}\n'
                     print(msg)
 
@@ -1598,7 +1600,7 @@ def check_tid_in_func_error(adj, node):
         if adj.is_user_function:
             if hasattr(node.func, "attr") and node.func.attr == "tid":
                 lineno = adj.lineno + adj.fun_lineno
-                line = adj.source.splitlines()[adj.lineno]
+                line = adj.source_lines[adj.lineno]
                 raise WarpCodegenError(
                     "tid() may only be called from a Warp kernel, not a Warp function. "
                     "Instead, obtain the indices from a @wp.kernel and pass them as "
@@ -1820,7 +1822,7 @@ def emit_Assign(adj, node):
 
                 if warp.config.verbose and not adj.custom_reverse_mode:
                     lineno = adj.lineno + adj.fun_lineno
-                    line = adj.source.splitlines()[adj.lineno]
+                    line = adj.source_lines[adj.lineno]
                     node_source = adj.get_node_source(lhs.value)
                     print(
                         f"Warning: mutating {node_source} in function {adj.fun_name} at {adj.filename}:{lineno}: this is a non-differentiable operation.\n{line}\n"
@@ -1877,7 +1879,7 @@ def emit_Assign(adj, node):
 
                 if warp.config.verbose and not adj.custom_reverse_mode:
                     lineno = adj.lineno + adj.fun_lineno
-                    line = adj.source.splitlines()[adj.lineno]
+                    line = adj.source_lines[adj.lineno]
                     msg = f'Warning: detected mutated struct {attr.label} during function "{adj.fun_name}" at {adj.filename}:{lineno}: this is a non-differentiable operation.\n{line}\n'
                     print(msg)
 
@@ -2074,7 +2076,7 @@ def resolve_static_expression(adj, root_node, eval_types=True):
     def set_lineno(adj, lineno):
         if adj.lineno is None or adj.lineno != lineno:
             line = lineno + adj.fun_lineno
-            source = adj.source.splitlines()[lineno].strip().ljust(80 - len(adj.indentation), " ")
+            source = adj.source_lines[lineno].strip().ljust(80 - len(adj.indentation), " ")
             adj.add_forward(f"// {source}       <L {line}>")
             adj.add_reverse(f"// adj: {source}  <L {line}>")
         adj.lineno = lineno

From f1128b89a0c015899fca0ecac92b76822449ab4a Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Wed, 13 Dec 2023 07:17:27 +1300
Subject: [PATCH 26/50] Bump omni.warp.core to version 1.0.0-beta.5

---
 exts/omni.warp.core/config/extension.toml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/exts/omni.warp.core/config/extension.toml b/exts/omni.warp.core/config/extension.toml
index 0a287a390..ec8d094d4 100644
--- a/exts/omni.warp.core/config/extension.toml
+++ b/exts/omni.warp.core/config/extension.toml
@@ -1,6 +1,6 @@
 [package]
 # Semantic Versioning is used: https://semver.org/
-version = "1.0.0-beta.4"
+version = "1.0.0-beta.5"
 authors = ["NVIDIA"]
 title = "Warp Core"
 description="The core Warp Python module"

From ebee05d0d7b3a4e865ddf0ff2685e3f11784c1c2 Mon Sep 17 00:00:00 2001
From: Gilles Daviet <gdaviet@nvidia.com>
Date: Tue, 12 Dec 2023 16:33:14 -0800
Subject: [PATCH 27/50] Address multiple differentiability issues

---
 warp/codegen.py                               |  26 ++++-
 warp/fem/cache.py                             |   3 +-
 warp/fem/field/nodal_field.py                 |  35 +++---
 warp/fem/integrate.py                         |  72 ++++++++++--
 warp/fem/space/shape/cube_shape_function.py   |  42 ++++---
 warp/fem/space/shape/square_shape_function.py |  47 +++++---
 warp/tests/test_fem.py                        | 106 ++++++++++++++++++
 warp/tests/test_grad.py                       |  51 +++++++++
 8 files changed, 317 insertions(+), 65 deletions(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index 30ab0fdb4..f852f0b00 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -1295,6 +1295,12 @@ def vector_component_index(adj, component, vector_type):
         index = adj.add_constant(index)
         return index
 
+    @staticmethod
+    def is_differentiable_value_type(var_type):
+        # checks that the argument type is a value type (i.e, not an array)
+        # possibly holding differentiable values (for which gradients must be accumulated)
+        return type_scalar_type(var_type) in float_types or isinstance(var_type, Struct)
+
     def emit_Attribute(adj, node):
         if hasattr(node, "is_adjoint"):
             node.value.is_adjoint = True
@@ -1331,9 +1337,12 @@ def emit_Attribute(adj, node):
 
                 if is_reference(aggregate.type):
                     adj.add_forward(f"{attr.emit()} = &({aggregate.emit()}->{node.attr});")
-                    adj.add_reverse(f"{aggregate.emit_adj()}.{node.attr} = {attr.emit_adj()};")
                 else:
                     adj.add_forward(f"{attr.emit()} = &({aggregate.emit()}.{node.attr});")
+
+                if adj.is_differentiable_value_type(strip_reference(attr_type)):
+                    adj.add_reverse(f"{aggregate.emit_adj()}.{node.attr} += {attr.emit_adj()};")
+                else:
                     adj.add_reverse(f"{aggregate.emit_adj()}.{node.attr} = {attr.emit_adj()};")
 
                 return attr
@@ -1455,6 +1464,10 @@ def eval_num(adj, a):
         # try and resolve the expression to an object
         # e.g.: wp.constant in the globals scope
         obj, path = adj.resolve_static_expression(a)
+
+        if isinstance(obj, Var) and obj.constant is not None:
+            obj = obj.constant
+
         return warp.types.is_int(obj), obj
 
     # detects whether a loop contains a break (or continue) statement
@@ -2135,7 +2148,9 @@ def get_node_source(adj, node):
     {{
     }}
 
-    CUDA_CALLABLE {name}& operator += (const {name}&) {{ return *this; }}
+    CUDA_CALLABLE {name}& operator += (const {name}& rhs)
+    {{{prefix_add_body}
+        return *this;}}
 
 }};
 
@@ -2362,6 +2377,7 @@ def codegen_struct(struct, device="cpu", indent_size=4):
     forward_initializers = []
     reverse_body = []
     atomic_add_body = []
+    prefix_add_body = []
 
     # forward args
     for label, var in struct.vars.items():
@@ -2375,6 +2391,11 @@ def codegen_struct(struct, device="cpu", indent_size=4):
         prefix = f"{indent_block}," if forward_initializers else ":"
         forward_initializers.append(f"{indent_block}{prefix} {label}{{{label}}}\n")
 
+    # prefix-add operator
+    for label, var in struct.vars.items():
+        if not is_array(var.type):
+            prefix_add_body.append(f"{indent_block}{label} += rhs.{label};\n")
+
     # reverse args
     for label, var in struct.vars.items():
         reverse_args.append(var.ctype() + " & adj_" + label)
@@ -2392,6 +2413,7 @@ def codegen_struct(struct, device="cpu", indent_size=4):
         forward_initializers="".join(forward_initializers),
         reverse_args=indent(reverse_args),
         reverse_body="".join(reverse_body),
+        prefix_add_body="".join(prefix_add_body),
         atomic_add_body="".join(atomic_add_body),
     )
 
diff --git a/warp/fem/cache.py b/warp/fem/cache.py
index 93faab52e..5345fc4dd 100644
--- a/warp/fem/cache.py
+++ b/warp/fem/cache.py
@@ -95,6 +95,7 @@ def wrap_struct(struct: type):
 def get_integrand_function(
     integrand: "warp.fem.operator.Integrand",
     suffix: str,
+    func=None,
     annotations=None,
     code_transformers=[],
 ):
@@ -102,7 +103,7 @@ def get_integrand_function(
 
     if key not in _func_cache:
         _func_cache[key] = wp.Function(
-            func=integrand.func,
+            func=integrand.func if func is None else func,
             key=key,
             namespace="",
             module=integrand.module,
diff --git a/warp/fem/field/nodal_field.py b/warp/fem/field/nodal_field.py
index 2f657a2c1..ccec0da63 100644
--- a/warp/fem/field/nodal_field.py
+++ b/warp/fem/field/nodal_field.py
@@ -84,15 +84,14 @@ def _make_eval_grad_inner(self, world_space: bool):
         if not self.gradient_valid():
             return None
 
-        @cache.dynamic_func(suffix=self.name + ("W" if world_space else "R"))
-        def eval_grad_inner(args: self.ElementEvalArg, s: Sample):
+        @cache.dynamic_func(suffix=self.name)
+        def eval_grad_inner_ref_space(args: self.ElementEvalArg, s: Sample):
             res = utils.generalized_outer(
                 self._read_node_value(args, s.element_index, 0),
                 self.space.element_inner_weight_gradient(
                     args.elt_arg, args.eval_arg.space_arg, s.element_index, s.element_coords, 0
                 ),
             )
-
             for k in range(1, NODES_PER_ELEMENT):
                 res += utils.generalized_outer(
                     self._read_node_value(args, s.element_index, k),
@@ -100,14 +99,15 @@ def eval_grad_inner(args: self.ElementEvalArg, s: Sample):
                         args.elt_arg, args.eval_arg.space_arg, s.element_index, s.element_coords, k
                     ),
                 )
-
-            if world_space:
-                grad_transform = self.space.element_inner_reference_gradient_transform(args.elt_arg, s)
-                return utils.apply_right(res, grad_transform)
-
             return res
 
-        return eval_grad_inner
+        @cache.dynamic_func(suffix=self.name)
+        def eval_grad_inner_world_space(args: self.ElementEvalArg, s: Sample):
+            grad_transform = self.space.element_inner_reference_gradient_transform(args.elt_arg, s)
+            res = eval_grad_inner_ref_space(args, s)
+            return utils.apply_right(res, grad_transform)
+
+        return eval_grad_inner_world_space if world_space else eval_grad_inner_ref_space
 
     def _make_eval_div_inner(self):
         NODES_PER_ELEMENT = self.space.topology.NODES_PER_ELEMENT
@@ -173,8 +173,8 @@ def _make_eval_grad_outer(self, world_space: bool):
         if not self.gradient_valid():
             return None
 
-        @cache.dynamic_func(suffix=self.name + ("W" if world_space else "R"))
-        def eval_grad_outer(args: self.ElementEvalArg, s: Sample):
+        @cache.dynamic_func(suffix=self.name)
+        def eval_grad_outer_ref_space(args: self.ElementEvalArg, s: Sample):
             res = utils.generalized_outer(
                 self._read_node_value(args, s.element_index, 0),
                 self.space.element_outer_weight_gradient(
@@ -188,14 +188,15 @@ def eval_grad_outer(args: self.ElementEvalArg, s: Sample):
                         args.elt_arg, args.eval_arg.space_arg, s.element_index, s.element_coords, k
                     ),
                 )
-
-            if world_space:
-                grad_transform = self.space.element_outer_reference_gradient_transform(args.elt_arg, s)
-                return utils.apply_right(res, grad_transform)
-
             return res
 
-        return eval_grad_outer
+        @cache.dynamic_func(suffix=self.name)
+        def eval_grad_outer_world_space(args: self.ElementEvalArg, s: Sample):
+            grad_transform = self.space.element_outer_reference_gradient_transform(args.elt_arg, s)
+            res = eval_grad_outer_ref_space(args, s)
+            return utils.apply_right(res, grad_transform)
+
+        return eval_grad_outer_world_space if world_space else eval_grad_outer_ref_space
 
     def _make_eval_div_outer(self):
         NODES_PER_ELEMENT = self.space.topology.NODES_PER_ELEMENT
diff --git a/warp/fem/integrate.py b/warp/fem/integrate.py
index 0230aa663..ebee7db93 100644
--- a/warp/fem/integrate.py
+++ b/warp/fem/integrate.py
@@ -619,7 +619,7 @@ def integrate_kernel_fn(
                 test_element_index.node_index_in_element,
                 test_dof,
             )
-        
+
             val_sum = accumulate_dtype(0.0)
 
             for k in range(qp_point_count):
@@ -1012,7 +1012,7 @@ def as_2d_array(array):
             )
 
         if output_temporary is not None:
-            return output_temporary.detach() 
+            return output_temporary.detach()
 
         return output
 
@@ -1108,7 +1108,7 @@ def as_2d_array(array):
             block_type=block_type,
             device=device,
         )
-    
+
     bsr_set_from_triplets(output, triplet_rows, triplet_cols, triplet_values)
 
     # Do not wait for garbage collection
@@ -1230,7 +1230,7 @@ def integrate(
     )
 
 
-def get_interpolate_to_field_kernel(
+def get_interpolate_to_field_function(
     integrand_func: wp.Function,
     domain: GeometryDomain,
     FieldStruct: wp.codegen.Struct,
@@ -1239,7 +1239,8 @@ def get_interpolate_to_field_kernel(
 ):
     value_type = dest.space.dtype
 
-    def interpolate_to_field_kernel_fn(
+    def interpolate_to_field_fn(
+        local_node_index: int,
         domain_arg: domain.ElementArg,
         domain_index_arg: domain.ElementIndexArg,
         dest_node_arg: dest.space_restriction.NodeArg,
@@ -1247,19 +1248,15 @@ def interpolate_to_field_kernel_fn(
         fields: FieldStruct,
         values: ValueStruct,
     ):
-        local_node_index = wp.tid()
         node_index = dest.space_restriction.node_partition_index(dest_node_arg, local_node_index)
-
         element_count = dest.space_restriction.node_element_count(dest_node_arg, local_node_index)
-        if element_count == 0:
-            return
 
         test_dof_index = NULL_DOF_INDEX
         trial_dof_index = NULL_DOF_INDEX
         node_weight = 1.0
 
-        # Volume-weighted average accross elements
-        # Superfluous if the function is continuous, but we might as well
+        # Volume-weighted average across elements
+        # Superfluous if the interpolated function is continuous, but helpful for visualizing discontinuous spaces
 
         val_sum = value_type(0.0)
         vol_sum = float(0.0)
@@ -1290,7 +1287,34 @@ def interpolate_to_field_kernel_fn(
                 vol_sum += vol
                 val_sum += vol * val
 
+        return val_sum, vol_sum
+
+    return interpolate_to_field_fn
+
+
+def get_interpolate_to_field_kernel(
+    interpolate_to_field_fn: wp.Function,
+    domain: GeometryDomain,
+    FieldStruct: wp.codegen.Struct,
+    ValueStruct: wp.codegen.Struct,
+    dest: FieldRestriction,
+):
+    def interpolate_to_field_kernel_fn(
+        domain_arg: domain.ElementArg,
+        domain_index_arg: domain.ElementIndexArg,
+        dest_node_arg: dest.space_restriction.NodeArg,
+        dest_eval_arg: dest.field.EvalArg,
+        fields: FieldStruct,
+        values: ValueStruct,
+    ):
+        local_node_index = wp.tid()
+
+        val_sum, vol_sum = interpolate_to_field_fn(
+            local_node_index, domain_arg, domain_index_arg, dest_node_arg, dest_eval_arg, fields, values
+        )
+
         if vol_sum > 0.0:
+            node_index = dest.space_restriction.node_partition_index(dest_node_arg, local_node_index)
             dest.field.set_node_value(dest_eval_arg, node_index, val_sum / vol_sum)
 
     return interpolate_to_field_kernel_fn
@@ -1404,13 +1428,37 @@ def _generate_interpolate_kernel(
 
     # Generate interpolation kernel
     if isinstance(dest, FieldRestriction):
-        interpolate_kernel_fn = get_interpolate_to_field_kernel(
+        # need to split into kernel + function for diffferentiability
+        interpolate_fn = get_interpolate_to_field_function(
             integrand_func,
             domain,
             dest=dest,
             FieldStruct=FieldStruct,
             ValueStruct=ValueStruct,
         )
+
+        interpolate_fn = cache.get_integrand_function(
+            integrand=integrand,
+            func=interpolate_fn,
+            suffix=kernel_suffix,
+            code_transformers=[
+                PassFieldArgsToIntegrand(
+                    arg_names=integrand.argspec.args,
+                    field_args=field_args.keys(),
+                    value_args=value_args.keys(),
+                    sample_name=sample_name,
+                    domain_name=domain_name,
+                )
+            ],
+        )
+
+        interpolate_kernel_fn = get_interpolate_to_field_kernel(
+            interpolate_fn,
+            domain,
+            dest=dest,
+            FieldStruct=FieldStruct,
+            ValueStruct=ValueStruct,
+        )
     elif wp.types.is_array(dest):
         interpolate_kernel_fn = get_interpolate_to_array_kernel(
             integrand_func,
diff --git a/warp/fem/space/shape/cube_shape_function.py b/warp/fem/space/shape/cube_shape_function.py
index bc16edcbe..91e7b6f62 100644
--- a/warp/fem/space/shape/cube_shape_function.py
+++ b/warp/fem/space/shape/cube_shape_function.py
@@ -41,6 +41,7 @@ def __init__(self, degree: int, family: Polynomial):
         self.LOBATTO_COORDS = wp.constant(NodeVec(lobatto_coords))
         self.LOBATTO_WEIGHT = wp.constant(NodeVec(lobatto_weight))
         self.LAGRANGE_SCALE = wp.constant(NodeVec(lagrange_scale))
+        self.ORDER_PLUS_ONE = wp.constant(self.ORDER + 1)
 
         self._node_ijk = self._make_node_ijk()
         self.node_type_and_type_index = self._make_node_type_and_type_index()
@@ -57,21 +58,21 @@ def _vertex_coords_f(vidx_in_cell: int):
         return wp.vec3(float(x), float(y), float(z))
 
     def _make_node_ijk(self):
-        ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
 
         def node_ijk(
             node_index_in_elt: int,
         ):
-            node_i = node_index_in_elt // ((ORDER + 1) * (ORDER + 1))
-            node_jk = node_index_in_elt - (ORDER + 1) * (ORDER + 1) * node_i
-            node_j = node_jk // (ORDER + 1)
-            node_k = node_jk - (ORDER + 1) * node_j
+            node_i = node_index_in_elt // (ORDER_PLUS_ONE * ORDER_PLUS_ONE)
+            node_jk = node_index_in_elt - ORDER_PLUS_ONE * ORDER_PLUS_ONE * node_i
+            node_j = node_jk // ORDER_PLUS_ONE
+            node_k = node_jk - ORDER_PLUS_ONE * node_j
             return node_i, node_j, node_k
 
         return cache.get_func(node_ijk, self.name)
 
     def _make_node_type_and_type_index(self):
-        ORDER = wp.constant(self.ORDER)
+        ORDER = self.ORDER
 
         @cache.dynamic_func(suffix=self.name)
         def node_type_and_type_index(
@@ -190,13 +191,21 @@ def trace_node_quadrature_weight_linear(
         ):
             return 0.25
 
+        def trace_node_quadrature_weight_open(
+            node_index_in_elt: int,
+        ):
+            return 0.0
+
+        if not is_closed(self.family):
+            return cache.get_func(trace_node_quadrature_weight_open, self.name)
+
         if ORDER == 1:
             return cache.get_func(trace_node_quadrature_weight_linear, self.name)
 
         return cache.get_func(trace_node_quadrature_weight, self.name)
 
     def make_element_inner_weight(self):
-        ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
 
@@ -207,7 +216,7 @@ def element_inner_weight(
             node_i, node_j, node_k = self._node_ijk(node_index_in_elt)
 
             w = float(1.0)
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_i:
                     w *= coords[0] - LOBATTO_COORDS[k]
                 if k != node_j:
@@ -230,13 +239,13 @@ def element_inner_weight_linear(
             wz = (1.0 - coords[2]) * (1.0 - v[2]) + v[2] * coords[2]
             return wx * wy * wz
 
-        if ORDER == 1:
+        if self.ORDER == 1 and is_closed(self.family):
             return cache.get_func(element_inner_weight_linear, self.name)
 
         return cache.get_func(element_inner_weight, self.name)
 
     def make_element_inner_weight_gradient(self):
-        ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
 
@@ -249,7 +258,7 @@ def element_inner_weight_gradient(
             prefix_xy = float(1.0)
             prefix_yz = float(1.0)
             prefix_zx = float(1.0)
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_i:
                     prefix_yz *= coords[0] - LOBATTO_COORDS[k]
                 if k != node_j:
@@ -265,7 +274,7 @@ def element_inner_weight_gradient(
             grad_y = float(0.0)
             grad_z = float(0.0)
 
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_i:
                     delta_x = coords[0] - LOBATTO_COORDS[k]
                     grad_x = grad_x * delta_x + prefix_x
@@ -308,7 +317,7 @@ def element_inner_weight_gradient_linear(
 
             return wp.vec3(dx * wy * wz, dy * wz * wx, dz * wx * wy)
 
-        if ORDER == 1:
+        if self.ORDER == 1 and is_closed(self.family):
             return cache.get_func(element_inner_weight_gradient_linear, self.name)
 
         return cache.get_func(element_inner_weight_gradient, self.name)
@@ -356,6 +365,7 @@ def __init__(self, degree: int, family: Polynomial):
         self.LOBATTO_COORDS = wp.constant(NodeVec(lobatto_coords))
         self.LOBATTO_WEIGHT = wp.constant(NodeVec(lobatto_weight))
         self.LAGRANGE_SCALE = wp.constant(NodeVec(lagrange_scale))
+        self.ORDER_PLUS_ONE = wp.constant(self.ORDER + 1)
 
         self.node_type_and_type_index = self._get_node_type_and_type_index()
         self._node_lobatto_indices = self._get_node_lobatto_indices()
@@ -466,6 +476,7 @@ def trace_node_quadrature_weight(
 
     def make_element_inner_weight(self):
         ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
 
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
@@ -511,7 +522,7 @@ def element_inner_weight(
             w *= wp.select(node_all[1] == 0, local_coords[1], 1.0 - local_coords[1])
             w *= wp.select(node_all[2] == 0, local_coords[2], 1.0 - local_coords[2])
 
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_all[0]:
                     w *= local_coords[0] - LOBATTO_COORDS[k]
             w *= LAGRANGE_SCALE[node_all[0]]
@@ -522,6 +533,7 @@ def element_inner_weight(
 
     def make_element_inner_weight_gradient(self):
         ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
 
@@ -585,7 +597,7 @@ def element_inner_weight_gradient(
             w_alt = LAGRANGE_SCALE[node_all[0]]
             g_alt = float(0.0)
             prefix_alt = LAGRANGE_SCALE[node_all[0]]
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_all[0]:
                     delta_alt = local_coords[0] - LOBATTO_COORDS[k]
                     w_alt *= delta_alt
diff --git a/warp/fem/space/shape/square_shape_function.py b/warp/fem/space/shape/square_shape_function.py
index ba205ceb7..0cf680f0c 100644
--- a/warp/fem/space/shape/square_shape_function.py
+++ b/warp/fem/space/shape/square_shape_function.py
@@ -25,6 +25,7 @@ def __init__(self, degree: int, family: Polynomial):
         self.LOBATTO_COORDS = wp.constant(NodeVec(lobatto_coords))
         self.LOBATTO_WEIGHT = wp.constant(NodeVec(lobatto_weight))
         self.LAGRANGE_SCALE = wp.constant(NodeVec(lagrange_scale))
+        self.ORDER_PLUS_ONE = wp.constant(self.ORDER + 1)
 
     @property
     def name(self) -> str:
@@ -93,13 +94,21 @@ def trace_node_quadrature_weight_linear(
         ):
             return 0.5
 
+        def trace_node_quadrature_weight_open(
+            node_index_in_elt: int,
+        ):
+            return 0.0
+
+        if not is_closed(self.family):
+            return cache.get_func(trace_node_quadrature_weight_open, self.name)
+
         if ORDER == 1:
             return cache.get_func(trace_node_quadrature_weight_linear, self.name)
 
         return cache.get_func(trace_node_quadrature_weight, self.name)
 
     def make_element_inner_weight(self):
-        ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
 
@@ -107,11 +116,11 @@ def element_inner_weight(
             coords: Coords,
             node_index_in_elt: int,
         ):
-            node_i = node_index_in_elt // (ORDER + 1)
-            node_j = node_index_in_elt - (ORDER + 1) * node_i
+            node_i = node_index_in_elt // ORDER_PLUS_ONE
+            node_j = node_index_in_elt - ORDER_PLUS_ONE * node_i
 
             w = float(1.0)
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_i:
                     w *= coords[0] - LOBATTO_COORDS[k]
                 if k != node_j:
@@ -131,13 +140,13 @@ def element_inner_weight_linear(
             wy = (1.0 - coords[1]) * (1.0 - v[1]) + v[1] * coords[1]
             return wx * wy
 
-        if ORDER == 1:
+        if self.ORDER == 1 and is_closed(self.family):
             return cache.get_func(element_inner_weight_linear, self.name)
 
         return cache.get_func(element_inner_weight, self.name)
 
     def make_element_inner_weight_gradient(self):
-        ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
 
@@ -145,12 +154,12 @@ def element_inner_weight_gradient(
             coords: Coords,
             node_index_in_elt: int,
         ):
-            node_i = node_index_in_elt // (ORDER + 1)
-            node_j = node_index_in_elt - (ORDER + 1) * node_i
+            node_i = node_index_in_elt // ORDER_PLUS_ONE
+            node_j = node_index_in_elt - ORDER_PLUS_ONE * node_i
 
             prefix_x = float(1.0)
             prefix_y = float(1.0)
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_i:
                     prefix_y *= coords[0] - LOBATTO_COORDS[k]
                 if k != node_j:
@@ -159,7 +168,7 @@ def element_inner_weight_gradient(
             grad_x = float(0.0)
             grad_y = float(0.0)
 
-            for k in range(ORDER + 1):
+            for k in range(ORDER_PLUS_ONE):
                 if k != node_i:
                     delta_x = coords[0] - LOBATTO_COORDS[k]
                     grad_x = grad_x * delta_x + prefix_x
@@ -187,7 +196,7 @@ def element_inner_weight_gradient_linear(
 
             return wp.vec2(dx * wy, dy * wx)
 
-        if ORDER == 1:
+        if self.ORDER == 1 and is_closed(self.family):
             return cache.get_func(element_inner_weight_gradient_linear, self.name)
 
         return cache.get_func(element_inner_weight_gradient, self.name)
@@ -230,6 +239,7 @@ def __init__(self, degree: int, family: Polynomial):
         self.LOBATTO_COORDS = wp.constant(NodeVec(lobatto_coords))
         self.LOBATTO_WEIGHT = wp.constant(NodeVec(lobatto_weight))
         self.LAGRANGE_SCALE = wp.constant(NodeVec(lagrange_scale))
+        self.ORDER_PLUS_ONE = wp.constant(self.ORDER + 1)
 
         self.node_type_and_type_index = self._get_node_type_and_type_index()
         self._node_lobatto_indices = self._get_node_lobatto_indices()
@@ -328,6 +338,7 @@ def trace_node_quadrature_weight(
 
     def make_element_inner_weight(self):
         ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
 
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
@@ -361,7 +372,7 @@ def element_inner_weight(
             if node_type == SquareSerendipityShapeFunctions.EDGE_Y:
                 w *= wp.select(node_i == 0, coords[0], 1.0 - coords[0])
             else:
-                for k in range(ORDER + 1):
+                for k in range(ORDER_PLUS_ONE):
                     if k != node_i:
                         w *= coords[0] - LOBATTO_COORDS[k]
 
@@ -370,7 +381,7 @@ def element_inner_weight(
             if node_type == SquareSerendipityShapeFunctions.EDGE_X:
                 w *= wp.select(node_j == 0, coords[1], 1.0 - coords[1])
             else:
-                for k in range(ORDER + 1):
+                for k in range(ORDER_PLUS_ONE):
                     if k != node_j:
                         w *= coords[1] - LOBATTO_COORDS[k]
                 w *= LAGRANGE_SCALE[node_j]
@@ -381,6 +392,7 @@ def element_inner_weight(
 
     def make_element_inner_weight_gradient(self):
         ORDER = self.ORDER
+        ORDER_PLUS_ONE = self.ORDER_PLUS_ONE
         LOBATTO_COORDS = self.LOBATTO_COORDS
         LAGRANGE_SCALE = self.LAGRANGE_SCALE
 
@@ -424,7 +436,7 @@ def element_inner_weight_gradient(
                 prefix_x = wp.select(node_j == 0, coords[1], 1.0 - coords[1])
             else:
                 prefix_x = LAGRANGE_SCALE[node_j]
-                for k in range(ORDER + 1):
+                for k in range(ORDER_PLUS_ONE):
                     if k != node_j:
                         prefix_x *= coords[1] - LOBATTO_COORDS[k]
 
@@ -432,7 +444,7 @@ def element_inner_weight_gradient(
                 prefix_y = wp.select(node_i == 0, coords[0], 1.0 - coords[0])
             else:
                 prefix_y = LAGRANGE_SCALE[node_i]
-                for k in range(ORDER + 1):
+                for k in range(ORDER_PLUS_ONE):
                     if k != node_i:
                         prefix_y *= coords[0] - LOBATTO_COORDS[k]
 
@@ -441,7 +453,7 @@ def element_inner_weight_gradient(
             else:
                 prefix_y *= LAGRANGE_SCALE[node_j]
                 grad_y = float(0.0)
-                for k in range(ORDER + 1):
+                for k in range(ORDER_PLUS_ONE):
                     if k != node_j:
                         delta_y = coords[1] - LOBATTO_COORDS[k]
                         grad_y = grad_y * delta_y + prefix_y
@@ -452,7 +464,7 @@ def element_inner_weight_gradient(
             else:
                 prefix_x *= LAGRANGE_SCALE[node_i]
                 grad_x = float(0.0)
-                for k in range(ORDER + 1):
+                for k in range(ORDER_PLUS_ONE):
                     if k != node_i:
                         delta_x = coords[0] - LOBATTO_COORDS[k]
                         grad_x = grad_x * delta_x + prefix_x
@@ -530,7 +542,6 @@ def make_node_quadrature_weight(self):
         NODES_PER_ELEMENT = self.NODES_PER_ELEMENT
 
         if self.ORDER == 2:
-
             # Intrinsic quadrature (order 2)
             @cache.dynamic_func(suffix=self.name)
             def node_quadrature_weight_quadratic(
diff --git a/warp/tests/test_fem.py b/warp/tests/test_fem.py
index 70a6da628..36f5fb110 100644
--- a/warp/tests/test_fem.py
+++ b/warp/tests/test_fem.py
@@ -63,6 +63,89 @@ def test_integrate_gradient(test_case, device):
         test_case.assertLess(err, 1.0e-8)
 
 
+@fem.integrand
+def bilinear_field(s: fem.Sample, domain: fem.Domain):
+    x = domain(s)
+    return x[0] * x[1]
+
+
+@fem.integrand
+def grad_field(s: fem.Sample, p: fem.Field):
+    return fem.grad(p, s)
+
+
+def test_interpolate_gradient(test_case, device):
+    with wp.ScopedDevice(device):
+        # Quad mesh with single element
+        # so we can test gradient with respect to vertex positions
+        positions = wp.array([[0.0, 0.0], [0.0, 2.0], [2.0, 0.0], [2.0, 2.0]], dtype=wp.vec2, requires_grad=True)
+        quads = wp.array([[0, 2, 3, 1]], dtype=int)
+        geo = fem.Quadmesh2D(quads, positions)
+
+        # Quadratic scalar space
+        scalar_space = fem.make_polynomial_space(geo, degree=2)
+
+        # Point-based vector space
+        # So we can test gradient with respect to inteprolation point position
+        point_coords = wp.array([[[0.5, 0.5, 0.0]]], dtype=fem.Coords, requires_grad=True)
+        interpolation_nodes = fem.PointBasisSpace(
+            fem.ExplicitQuadrature(domain=fem.Cells(geo), points=point_coords, weights=wp.array([[1.0]], dtype=float))
+        )
+        vector_space = fem.make_collocated_function_space(interpolation_nodes, dtype=wp.vec2)
+
+        # Initialize scalar field with known function
+        scalar_field = scalar_space.make_field()
+        scalar_field.dof_values.requires_grad = True
+        fem.interpolate(bilinear_field, dest=scalar_field)
+
+        # Interpolate gradient at center point
+        vector_field = vector_space.make_field()
+        vector_field.dof_values.requires_grad = True
+        tape = wp.Tape()
+        with tape:
+            fem.interpolate(grad_field, dest=vector_field, fields={"p": scalar_field})
+
+        assert_np_equal(vector_field.dof_values.numpy(), np.array([[1.0, 1.0]]))
+
+        vector_field.dof_values.grad.assign([1.0, 0.0])
+        tape.backward()
+
+        assert_np_equal(scalar_field.dof_values.grad.numpy(), np.array([0.0, 0.0, 0.0, 0.0, 0.0, -0.5, 0.0, 0.5, 0.0]))
+        assert_np_equal(
+            geo.positions.grad.numpy(),
+            np.array(
+                [
+                    [0.25, 0.25],
+                    [0.25, 0.25],
+                    [-0.25, -0.25],
+                    [-0.25, -0.25],
+                ]
+            ),
+        )
+        assert_np_equal(point_coords.grad.numpy(), np.array([[[0.0, 2.0, 0.0]]]))
+
+        tape.zero()
+        scalar_field.dof_values.grad.zero_()
+        geo.positions.grad.zero_()
+        point_coords.grad.zero_()
+
+        vector_field.dof_values.grad.assign([0.0, 1.0])
+        tape.backward()
+
+        assert_np_equal(scalar_field.dof_values.grad.numpy(), np.array([0.0, 0.0, 0.0, 0.0, -0.5, 0.0, 0.5, 0.0, 0.0]))
+        assert_np_equal(
+            geo.positions.grad.numpy(),
+            np.array(
+                [
+                    [0.25, 0.25],
+                    [-0.25, -0.25],
+                    [0.25, 0.25],
+                    [-0.25, -0.25],
+                ]
+            ),
+        )
+        assert_np_equal(point_coords.grad.numpy(), np.array([[[2.0, 0.0, 0.0]]]))
+
 @integrand
 def vector_divergence_form(s: Sample, u: Field, q: Field):
     return div(u, s) * q(s)
@@ -889,6 +972,17 @@ def square_coord_delta_sampler(epsilon: float, state: wp.uint32):
     test_shape_function_gradient(test_case, Q_2, square_coord_sampler, square_coord_delta_sampler)
     test_shape_function_gradient(test_case, Q_3, square_coord_sampler, square_coord_delta_sampler)
 
+    Q_1 = shape.SquareBipolynomialShapeFunctions(degree=1, family=fem.Polynomial.GAUSS_LEGENDRE)
+    Q_2 = shape.SquareBipolynomialShapeFunctions(degree=2, family=fem.Polynomial.GAUSS_LEGENDRE)
+    Q_3 = shape.SquareBipolynomialShapeFunctions(degree=3, family=fem.Polynomial.GAUSS_LEGENDRE)
+
+    test_shape_function_weight(test_case, Q_1, square_coord_sampler, SQUARE_CENTER_COORDS)
+    test_shape_function_weight(test_case, Q_2, square_coord_sampler, SQUARE_CENTER_COORDS)
+    test_shape_function_weight(test_case, Q_3, square_coord_sampler, SQUARE_CENTER_COORDS)
+    test_shape_function_gradient(test_case, Q_1, square_coord_sampler, square_coord_delta_sampler)
+    test_shape_function_gradient(test_case, Q_2, square_coord_sampler, square_coord_delta_sampler)
+    test_shape_function_gradient(test_case, Q_3, square_coord_sampler, square_coord_delta_sampler)
+
     S_2 = shape.SquareSerendipityShapeFunctions(degree=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
     S_3 = shape.SquareSerendipityShapeFunctions(degree=3, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
 
@@ -940,6 +1034,17 @@ def cube_coord_delta_sampler(epsilon: float, state: wp.uint32):
     test_shape_function_gradient(test_case, Q_2, cube_coord_sampler, cube_coord_delta_sampler)
     test_shape_function_gradient(test_case, Q_3, cube_coord_sampler, cube_coord_delta_sampler)
 
+    Q_1 = shape.CubeTripolynomialShapeFunctions(degree=1, family=fem.Polynomial.GAUSS_LEGENDRE)
+    Q_2 = shape.CubeTripolynomialShapeFunctions(degree=2, family=fem.Polynomial.GAUSS_LEGENDRE)
+    Q_3 = shape.CubeTripolynomialShapeFunctions(degree=3, family=fem.Polynomial.GAUSS_LEGENDRE)
+
+    test_shape_function_weight(test_case, Q_1, cube_coord_sampler, CUBE_CENTER_COORDS)
+    test_shape_function_weight(test_case, Q_2, cube_coord_sampler, CUBE_CENTER_COORDS)
+    test_shape_function_weight(test_case, Q_3, cube_coord_sampler, CUBE_CENTER_COORDS)
+    test_shape_function_gradient(test_case, Q_1, cube_coord_sampler, cube_coord_delta_sampler)
+    test_shape_function_gradient(test_case, Q_2, cube_coord_sampler, cube_coord_delta_sampler)
+    test_shape_function_gradient(test_case, Q_3, cube_coord_sampler, cube_coord_delta_sampler)
+
     S_2 = shape.CubeSerendipityShapeFunctions(degree=2, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
     S_3 = shape.CubeSerendipityShapeFunctions(degree=3, family=fem.Polynomial.LOBATTO_GAUSS_LEGENDRE)
 
@@ -1135,6 +1240,7 @@ class TestFem(parent):
     add_function_test(TestFem, "test_closest_point_queries", test_closest_point_queries)
     add_function_test(TestFem, "test_grad_decomposition", test_grad_decomposition, devices=devices)
     add_function_test(TestFem, "test_integrate_gradient", test_integrate_gradient, devices=devices)
+    add_function_test(TestFem, "test_interpolate_gradient", test_interpolate_gradient, devices=devices)
     add_function_test(TestFem, "test_vector_divergence_theorem", test_vector_divergence_theorem, devices=devices)
     add_function_test(TestFem, "test_tensor_divergence_theorem", test_tensor_divergence_theorem, devices=devices)
     add_function_test(TestFem, "test_grid_2d", test_grid_2d, devices=devices)
diff --git a/warp/tests/test_grad.py b/warp/tests/test_grad.py
index 6e15c748f..d8ed5d062 100644
--- a/warp/tests/test_grad.py
+++ b/warp/tests/test_grad.py
@@ -6,6 +6,7 @@
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
 import unittest
+from typing import Any
 
 import numpy as np
 
@@ -560,6 +561,55 @@ def test_name_clash(test, device):
         assert_np_equal(input_b.grad.numpy(), np.array([1.0, 1.0, 0.0]))
 
 
+@wp.struct
+class NestedStruct:
+    v: wp.vec2
+
+
+@wp.struct
+class ParentStruct:
+    a: float
+    n: NestedStruct
+
+
+@wp.func
+def noop(a: Any):
+    pass
+
+
+@wp.func
+def sum2(v: wp.vec2):
+    return v[0] + v[1]
+
+
+@wp.kernel
+def test_struct_attribute_gradient_kernel(src: wp.array(dtype=float), res: wp.array(dtype=float)):
+    tid = wp.tid()
+
+    p = ParentStruct(src[tid], NestedStruct(wp.vec2(2.0 * src[tid])))
+
+    # test that we are not losing gradients when accessing attributes
+    noop(p.a)
+    noop(p.n)
+    noop(p.n.v)
+
+    res[tid] = p.a + sum2(p.n.v)
+
+
+def test_struct_attribute_gradient(test_case, device):
+    src = wp.array([1], dtype=float, requires_grad=True)
+    res = wp.empty_like(src)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(test_struct_attribute_gradient_kernel, dim=1, inputs=[src, res])
+
+    res.grad.fill_(1.0)
+    tape.backward()
+
+    test_case.assertEqual(src.grad.numpy()[0], 5.0)
+
+
 def register(parent):
     devices = get_test_devices()
 
@@ -579,6 +629,7 @@ class TestGrad(parent):
     add_function_test(TestGrad, "test_multi_valued_function_grad", test_multi_valued_function_grad, devices=devices)
     add_function_test(TestGrad, "test_mesh_grad", test_mesh_grad, devices=devices)
     add_function_test(TestGrad, "test_name_clash", test_name_clash, devices=devices)
+    add_function_test(TestGrad, "test_struct_attribute_gradient", test_struct_attribute_gradient, devices=devices)
 
     return TestGrad
 

From f6fb39684d0176b910315e27044ffde49fc6a2c8 Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Tue, 12 Dec 2023 16:42:38 -0800
Subject: [PATCH 28/50] Add github-linguist exclusions for accurate language
 breakdown

---
 .gitattributes | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/.gitattributes b/.gitattributes
index b495b445b..c3c477560 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -13,3 +13,7 @@
 *.grid filter=lfs diff=lfs merge=lfs -text
 *.nvdb filter=lfs diff=lfs merge=lfs -text
 *.gif filter=lfs diff=lfs merge=lfs -text
+
+# Exclude vendored code from project language stats
+warp/native/cutlass/** linguist-vendored
+tools/** linguist-vendored

From 0a5c46a48b65bf71d3c53ce2409d0a5d3fd88c0a Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Tue, 12 Dec 2023 16:43:02 -0800
Subject: [PATCH 29/50] Make clear when tid() breaking change happened in
 CHANGELOG

---
 CHANGELOG.md | 42 +++++++++++++++++-------------------------
 1 file changed, 17 insertions(+), 25 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index c452e8f7f..d98de5de3 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,15 +5,15 @@
 - Fix for kernel caching when function argument types change
 - Fix code-gen ordering of dependent structs
 - Fix for `wp.Mesh` build on MGPU systems
-- Fix for name clash bug with adjoint code: https://github.com/NVIDIA/warp/issues/154 
+- Fix for name clash bug with adjoint code: https://github.com/NVIDIA/warp/issues/154
 - Add `wp.frac()` for returning the fractional part of a floating point value
 - Add support for custom native CUDA snippets using `@wp.func_native` decorator
 - Add support for batched matmul with batch size > 2^16-1
-- Add support for tranposed CUTLASS `wp.matmul()` and additional error checking
+- Add support for transposed CUTLASS `wp.matmul()` and additional error checking
 - Add support for quad and hex meshes in `wp.fem`
-- Detect and warn when C++ runtime doesn't match compiler during build, e.g.: libstdc++.so.6: version `GLIBCXX_3.4.30' not found
+- Detect and warn when C++ runtime doesn't match compiler during build, e.g.: ``libstdc++.so.6: version `GLIBCXX_3.4.30' not found``
 - Documentation update for `wp.BVH`
-- Documentaiton and simplified API for runtime kernel specialization `wp.Kernel`
+- Documentation and simplified API for runtime kernel specialization `wp.Kernel`
 
 ## [1.0.0-beta.4] - 2023-11-01
 
@@ -43,7 +43,11 @@
 - Fix for `wp.func` to return a default value if function does not return on all control paths
 - Refactor `wp.fem` support for new basis functions, decoupled function spaces
 - Optimizations for `wp.noise` functions, up to 10x faster in most cases
-- Optimizations for `type_size_in_bytes()` used in array construction
+- Optimizations for `type_size_in_bytes()` used in array construction'
+
+### Breaking Changes
+
+- To support grid-stride kernels, `wp.tid()` can no longer be called inside `wp.func` functions.
 
 ## [1.0.0-beta.2] - 2023-09-01
 
@@ -181,7 +185,7 @@
 ## [0.8.1] - 2023-04-13
 
 - Fix for regression when passing flattened numeric lists as matrix arguments to kernels
-- Fix for regressions when passing wp.struct types with uninitialized (None) member attributes
+- Fix for regressions when passing `wp.struct` types with uninitialized (`None`) member attributes
 
 ## [0.8.0] - 2023-04-05
 
@@ -215,20 +219,18 @@
 - Optimizations for `wp.launch()`, up to 3x faster launches in common cases
 - Fix `wp.randf()` conversion to float to reduce bias for uniform sampling
 - Fix capture of `wp.func` and `wp.constant` types from inside Python closures
-- Fix for CUDA on WSL 
+- Fix for CUDA on WSL
 - Fix for matrices in structs
 - Fix for transpose indexing for some non-square matrices
 - Enable Python faulthandler by default
 - Update to VS2019
 
-Breaking Changes
-----------------
+### Breaking Changes
 
 - `wp.constant` variables can now be treated as their true type, accessing the underlying value through `constant.val` is no longer supported
 - `wp.sim.model.ground_plane` is now a `wp.array` to support gradient, users should call `builder.set_ground_plane()` to create the ground 
 - `wp.sim` capsule, cones, and cylinders are now aligned with the default USD up-axis
 
-  
 ## [0.7.2] - 2023-02-15
 
 - Reduce test time for vec/math types
@@ -248,9 +250,9 @@ Breaking Changes
 - Add support for slicing `wp.array` types in Python
 - Add `wp.from_ptr()` helper to construct arrays from an existing allocation
 - Add support for `break` statements in ranged-for and while loops (backward pass support currently not implemented)
-- Add built-in mathematic constants, see `wp.pi`, `wp.e`, `wp.log2e`, etc
+- Add built-in mathematic constants, see `wp.pi`, `wp.e`, `wp.log2e`, etc.
 - Add built-in conversion between degrees and radians, see `wp.degrees()`, `wp.radians()`
-- Add security pop-up for Kernel Node 
+- Add security pop-up for Kernel Node
 - Improve error handling for kernel return values
 
 ## [0.6.3] - 2023-01-31
@@ -336,7 +338,7 @@ Breaking Changes
 - Fix for hashing of `wp.constants()` not invalidating kernels
 - Fix for reload when multiple `.ptx` versions are present
 - Improved error reporting during code-gen
- 
+
 ## [0.4.3] - 2022-09-20
 
 - Update all samples to use GPU interop path by default
@@ -360,7 +362,6 @@ Breaking Changes
 - Fix for debug flags not being set correctly on CUDA when `wp.config.mode == "debug"`, this enables bounds checking on CUDA kernels in debug mode
 - Fix for code gen of functions that do not return a value
 
-
 ## [0.4.0] - 2022-08-09
 
 - Fix for FP16 conversions on GPUs without hardware support
@@ -373,18 +374,15 @@ Breaking Changes
 - Add support for cross-module `@wp.struct` references
 - Support running even if CUDA initialization failed, use `wp.is_cuda_available()` to check availability
 - Statically linking with the CUDA runtime library to avoid deployment issues
-  
 
 ### Breaking Changes
 
 - Removed `wp.runtime` reference from the top-level module, as it should be considered private
 
-
 ## [0.3.2] - 2022-07-19
 
 - Remove Torch import from `__init__.py`, defer import to `wp.from_torch()`, `wp.to_torch()`
 
-
 ## [0.3.1] - 2022-07-12
 
 - Fix for marching cubes reallocation after initialization
@@ -396,7 +394,6 @@ Breaking Changes
 - Add support for using arbitrary external CUDA contexts, see `wp.map_cuda_device()`, `wp.unmap_cuda_device()`
 - Add PyTorch device aliasing functions, see `wp.device_from_torch()`, `wp.device_to_torch()`
 
-
 ### Breaking Changes
 
 - A CUDA device is used by default, if available (aligned with `wp.get_preferred_device()`)
@@ -404,7 +401,6 @@ Breaking Changes
 - `wp.synchronize()` now synchronizes all devices; for finer-grained control, use `wp.synchronize_device()`
 - Device alias `"cuda"` now refers to the current CUDA context, rather than a specific device like `"cuda:0"` or `"cuda:1"`
 
-
 ## [0.3.0] - 2022-07-08
 
 - Add support for FP16 storage type, see `wp.float16`
@@ -429,7 +425,6 @@ Breaking Changes
 - Fix for reload of generated CPU kernel code on Linux
 - Fix for example scripts to output USD at 60 timecodes per-second (better Kit compatibility)
 
-
 ## [0.2.3] - 2022-06-13
 
 - Fix for incorrect 4d array bounds checking
@@ -438,7 +433,6 @@ Breaking Changes
 - Array gradients are now allocated along with the arrays and accessible as `wp.array.grad`, users should take care to always call `wp.Tape.zero()` to clear gradients between different invocations of `wp.Tape.backward()`
 - Added `wp.array.fill_()` to set all entries to a scalar value (4-byte values only currently)
 
-
 ### Breaking Changes
 
 - Tape `capture` option has been removed, users can now capture tapes inside existing CUDA graphs (e.g.: inside Torch)
@@ -471,7 +465,7 @@ Breaking Changes
 ## [0.2.1] - 2022-05-11
 
 - Fix for unit tests in Kit
-- 
+
 ## [0.2.0] - 2022-05-02
 
 ### Warp Core
@@ -509,7 +503,7 @@ Breaking Changes
 - Fix for URDF importer and floating base support
 - Add examples showing how to use differentiable forward kinematics to solve inverse kinematics
 - Add examples for URDF cartpole and quadruped simulation
- 
+
 ### Breaking Changes
 
 - `wp.volume_sample_world()` is now replaced by `wp.volume_sample_f/i/vec()` which operate in index (local) space. Users should use `wp.volume_world_to_index()` to transform points from world space to index space before sampling.
@@ -517,7 +511,6 @@ Breaking Changes
 - `wp.array.length` member has been removed, please use `wp.array.shape` to access array dimensions, or use `wp.array.size` to get total element count
 - Marking `dense_gemm()`, `dense_chol()`, etc methods as experimental until we revisit them
 
-
 ## [0.1.25] - 2022-03-20
 
 - Add support for class methods to be Warp kernels
@@ -528,7 +521,6 @@ Breaking Changes
 - Add support for floored division on integer types
 - Move tests into core library so they can be run in Kit environment
 
-
 ## [0.1.24] - 2022-03-03
 
 ### Warp Core

From 5e305f837348034df2c1a7fb1959644da3d26d86 Mon Sep 17 00:00:00 2001
From: Eric Heiden <eheiden@nvidia.com>
Date: Tue, 12 Dec 2023 16:47:42 -0800
Subject: [PATCH 30/50] Make atomic_min, atomic_max differentiable

---
 warp/native/array.h       | 57 +++++++++++++++++++++++++++++++++------
 warp/native/builtin.h     | 19 +++++++++++++
 warp/native/mat.h         | 12 +++++++++
 warp/native/vec.h         | 23 +++++++++++-----
 warp/stubs.py             | 16 +++++++++++
 warp/tests/test_atomic.py | 54 ++++++++++++++++++++++++++++++++-----
 6 files changed, 160 insertions(+), 21 deletions(-)

diff --git a/warp/native/array.h b/warp/native/array.h
index 86ac4d98d..8c6ea2b98 100644
--- a/warp/native/array.h
+++ b/warp/native/array.h
@@ -951,23 +951,64 @@ inline CUDA_CALLABLE void adj_atomic_sub(const A1<T>& buf, int i, int j, int k,
 template<template<typename> class A1, template<typename> class A2, typename T>
 inline CUDA_CALLABLE void adj_atomic_sub(const A1<T>& buf, int i, int j, int k, int l, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, int& adj_l, T& adj_value, const T& adj_ret) {}
 
+// generic handler for scalar values
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, T value, const A2<T>& adj_buf, int& adj_i, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, T value, const A2<T>& adj_buf, int& adj_i, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i), &index_grad(buf, i), value, adj_value);
+
+    FP_VERIFY_ADJ_1(value, adj_value)
+}
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, int j, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, int j, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i, j), &index_grad(buf, i, j), value, adj_value);
+
+    FP_VERIFY_ADJ_2(value, adj_value)
+}
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, int j, int k, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, int j, int k, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i, j, k), &index_grad(buf, i, j, k), value, adj_value);
+
+    FP_VERIFY_ADJ_3(value, adj_value)
+}
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, int j, int k, int l, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, int& adj_l, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_min(const A1<T>& buf, int i, int j, int k, int l, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, int& adj_l, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i, j, k, l), &index_grad(buf, i, j, k, l), value, adj_value);
+
+    FP_VERIFY_ADJ_4(value, adj_value)
+}
 
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, T value, const A2<T>& adj_buf, int& adj_i, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, T value, const A2<T>& adj_buf, int& adj_i, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i), &index_grad(buf, i), value, adj_value);
+
+    FP_VERIFY_ADJ_1(value, adj_value)
+}
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, int j, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, int j, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i, j), &index_grad(buf, i, j), value, adj_value);
+
+    FP_VERIFY_ADJ_2(value, adj_value)
+}
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, int j, int k, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, int j, int k, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i, j, k), &index_grad(buf, i, j, k), value, adj_value);
+
+    FP_VERIFY_ADJ_3(value, adj_value)
+}
 template<template<typename> class A1, template<typename> class A2, typename T>
-inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, int j, int k, int l, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, int& adj_l, T& adj_value, const T& adj_ret) {}
+inline CUDA_CALLABLE void adj_atomic_max(const A1<T>& buf, int i, int j, int k, int l, T value, const A2<T>& adj_buf, int& adj_i, int& adj_j, int& adj_k, int& adj_l, T& adj_value, const T& adj_ret) {
+    if (buf.grad)
+        adj_atomic_minmax(&index(buf, i, j, k, l), &index_grad(buf, i, j, k, l), value, adj_value);
+
+    FP_VERIFY_ADJ_4(value, adj_value)
+}
 
 } // namespace wp
 
diff --git a/warp/native/builtin.h b/warp/native/builtin.h
index ad33cf575..d5b7c8b2d 100644
--- a/warp/native/builtin.h
+++ b/warp/native/builtin.h
@@ -1271,6 +1271,25 @@ inline CUDA_CALLABLE int atomic_min(int* address, int val)
 #endif
 }
 
+// default behavior for adjoint of atomic min/max operation that accumulates gradients for all elements matching the min/max value
+template <typename T>
+CUDA_CALLABLE inline void adj_atomic_minmax(T *addr, T *adj_addr, const T &value, T &adj_value)
+{
+    if (value == *addr)
+        adj_value += *adj_addr;
+}
+
+// for integral types we do not accumulate gradients
+CUDA_CALLABLE inline void adj_atomic_minmax(int8* buf, int8* adj_buf, const int8 &value, int8 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(uint8* buf, uint8* adj_buf, const uint8 &value, uint8 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(int16* buf, int16* adj_buf, const int16 &value, int16 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(uint16* buf, uint16* adj_buf, const uint16 &value, uint16 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(int32* buf, int32* adj_buf, const int32 &value, int32 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(uint32* buf, uint32* adj_buf, const uint32 &value, uint32 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(int64* buf, int64* adj_buf, const int64 &value, int64 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(uint64* buf, uint64* adj_buf, const uint64 &value, uint64 &adj_value) { }
+CUDA_CALLABLE inline void adj_atomic_minmax(bool* buf, bool* adj_buf, const bool &value, bool &adj_value) { }
+
 
 } // namespace wp
 
diff --git a/warp/native/mat.h b/warp/native/mat.h
index d3853f689..40dfbdcc7 100644
--- a/warp/native/mat.h
+++ b/warp/native/mat.h
@@ -297,6 +297,18 @@ inline CUDA_CALLABLE mat_t<Rows,Cols,Type> atomic_max(mat_t<Rows,Cols,Type> * ad
     return m;
 }
 
+template<unsigned Rows, unsigned Cols, typename Type>
+inline CUDA_CALLABLE void adj_atomic_minmax(
+    mat_t<Rows,Cols,Type> *addr,
+    mat_t<Rows,Cols,Type> *adj_addr,
+    const mat_t<Rows,Cols,Type> &value,
+    mat_t<Rows,Cols,Type> &adj_value)
+{
+    for (unsigned i=0; i < Rows; ++i)
+        for (unsigned j=0; j < Cols; ++j)
+            adj_atomic_minmax(&addr->data[i][j], &adj_addr->data[i][j], value.data[i][j], adj_value.data[i][j]);
+}
+
 template<unsigned Rows, unsigned Cols, typename Type>
 inline CUDA_CALLABLE vec_t<Cols,Type> extract(const mat_t<Rows,Cols,Type>& m, int row)
 {
diff --git a/warp/native/vec.h b/warp/native/vec.h
index 9c70864cb..e5abeca9b 100644
--- a/warp/native/vec.h
+++ b/warp/native/vec.h
@@ -952,8 +952,8 @@ inline CUDA_CALLABLE void adj_max(const vec_t<Length,Type> &v, vec_t<Length,Type
 
 // Do I need to specialize these for different lengths?
 template<unsigned Length, typename Type>
-inline CUDA_CALLABLE vec_t<Length, Type> atomic_add(vec_t<Length, Type> * addr, vec_t<Length, Type> value) {
-
+inline CUDA_CALLABLE vec_t<Length, Type> atomic_add(vec_t<Length, Type> * addr, vec_t<Length, Type> value)
+{
     vec_t<Length, Type> ret;
     for( unsigned i=0; i < Length; ++i )
     {
@@ -964,8 +964,8 @@ inline CUDA_CALLABLE vec_t<Length, Type> atomic_add(vec_t<Length, Type> * addr,
 }
 
 template<unsigned Length, typename Type>
-inline CUDA_CALLABLE vec_t<Length, Type> atomic_min(vec_t<Length, Type> * addr, vec_t<Length, Type> value) {
-
+inline CUDA_CALLABLE vec_t<Length, Type> atomic_min(vec_t<Length, Type> * addr, vec_t<Length, Type> value)
+{
     vec_t<Length, Type> ret;
     for( unsigned i=0; i < Length; ++i )
     {
@@ -976,8 +976,8 @@ inline CUDA_CALLABLE vec_t<Length, Type> atomic_min(vec_t<Length, Type> * addr,
 }
 
 template<unsigned Length, typename Type>
-inline CUDA_CALLABLE vec_t<Length, Type> atomic_max(vec_t<Length, Type> * addr, vec_t<Length, Type> value) {
-
+inline CUDA_CALLABLE vec_t<Length, Type> atomic_max(vec_t<Length, Type> * addr, vec_t<Length, Type> value)
+{
     vec_t<Length, Type> ret;
     for( unsigned i=0; i < Length; ++i )
     {
@@ -987,6 +987,17 @@ inline CUDA_CALLABLE vec_t<Length, Type> atomic_max(vec_t<Length, Type> * addr,
     return ret;
 }
 
+template<unsigned Length, typename Type>
+inline CUDA_CALLABLE void adj_atomic_minmax(
+    vec_t<Length,Type> *addr,
+    vec_t<Length,Type> *adj_addr,
+    const vec_t<Length,Type> &value,
+    vec_t<Length,Type> &adj_value)
+{
+    for (unsigned i=0; i < Length; ++i)
+        adj_atomic_minmax(&(addr->c[i]), &(adj_addr->c[i]), value[i], adj_value[i]);
+}
+
 // ok, the original implementation of this didn't take the absolute values.
 // I wouldn't consider this expected behavior. It looks like it's only
 // being used for bounding boxes at the moment, where this doesn't matter,
diff --git a/warp/stubs.py b/warp/stubs.py
index 9d5afa21f..a3856efdd 100644
--- a/warp/stubs.py
+++ b/warp/stubs.py
@@ -1863,6 +1863,8 @@ def atomic_min(a: Array[Any], i: int32, value: Any):
     Compute the minimum of ``value`` and ``a[i]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1873,6 +1875,8 @@ def atomic_min(a: Array[Any], i: int32, j: int32, value: Any):
     Compute the minimum of ``value`` and ``a[i,j]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1883,6 +1887,8 @@ def atomic_min(a: Array[Any], i: int32, j: int32, k: int32, value: Any):
     Compute the minimum of ``value`` and ``a[i,j,k]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1893,6 +1899,8 @@ def atomic_min(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any
     Compute the minimum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1983,6 +1991,8 @@ def atomic_max(a: Array[Any], i: int32, value: Any):
     Compute the maximum of ``value`` and ``a[i]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
@@ -1993,6 +2003,8 @@ def atomic_max(a: Array[Any], i: int32, j: int32, value: Any):
     Compute the maximum of ``value`` and ``a[i,j]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
@@ -2003,6 +2015,8 @@ def atomic_max(a: Array[Any], i: int32, j: int32, k: int32, value: Any):
     Compute the maximum of ``value`` and ``a[i,j,k]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
@@ -2013,6 +2027,8 @@ def atomic_max(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any
     Compute the maximum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
+
+    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
diff --git a/warp/tests/test_atomic.py b/warp/tests/test_atomic.py
index 6dc0d4ea2..5cba5f746 100644
--- a/warp/tests/test_atomic.py
+++ b/warp/tests/test_atomic.py
@@ -53,20 +53,60 @@ def test_atomic(test, device):
             base = rng.random(size=(1, *type._shape_), dtype=float)
             val = rng.random(size=(n, *type._shape_), dtype=float)
 
-        add_array = wp.array(base, dtype=type, device=device)
-        min_array = wp.array(base, dtype=type, device=device)
-        max_array = wp.array(base, dtype=type, device=device)
+        add_array = wp.array(base, dtype=type, device=device, requires_grad=True)
+        min_array = wp.array(base, dtype=type, device=device, requires_grad=True)
+        max_array = wp.array(base, dtype=type, device=device, requires_grad=True)
+        add_array.zero_()
+        min_array.fill_(10000)
+        max_array.fill_(-10000)
 
-        val_array = wp.array(val, dtype=type, device=device)
+        val_array = wp.array(val, dtype=type, device=device, requires_grad=True)
 
-        wp.launch(kernel, n, inputs=[add_array, min_array, max_array, val_array], device=device)
-
-        val = np.append(val, [base[0]], axis=0)
+        tape = wp.Tape()
+        with tape:
+            wp.launch(kernel, n, inputs=[add_array, min_array, max_array, val_array], device=device)
 
         assert_np_equal(add_array.numpy(), np.sum(val, axis=0), tol=1.0e-2)
         assert_np_equal(min_array.numpy(), np.min(val, axis=0), tol=1.0e-2)
         assert_np_equal(max_array.numpy(), np.max(val, axis=0), tol=1.0e-2)
 
+        if type != wp.int32:
+            add_array.grad.fill_(1)
+            tape.backward()
+            assert_np_equal(val_array.grad.numpy(), np.ones_like(val))
+            tape.zero()
+
+            min_array.grad.fill_(1)
+            tape.backward()
+            min_grad_array = np.zeros_like(val)
+            argmin = val.argmin(axis=0)
+            if val.ndim == 1:
+                min_grad_array[argmin] = 1
+            elif val.ndim == 2:
+                for i in range(val.shape[1]):
+                    min_grad_array[argmin[i], i] = 1
+            elif val.ndim == 3:
+                for i in range(val.shape[1]):
+                    for j in range(val.shape[2]):
+                        min_grad_array[argmin[i, j], i, j] = 1
+            assert_np_equal(val_array.grad.numpy(), min_grad_array)
+            tape.zero()
+
+            max_array.grad.fill_(1)
+            tape.backward()
+            max_grad_array = np.zeros_like(val)
+            argmax = val.argmax(axis=0)
+            if val.ndim == 1:
+                max_grad_array[argmax] = 1
+            elif val.ndim == 2:
+                for i in range(val.shape[1]):
+                    max_grad_array[argmax[i], i] = 1
+            elif val.ndim == 3:
+                for i in range(val.shape[1]):
+                    for j in range(val.shape[2]):
+                        max_grad_array[argmax[i, j], i, j] = 1
+            assert_np_equal(val_array.grad.numpy(), max_grad_array)
+
     return test_atomic
 
 

From ebaf4adb18bc89c47bfcec85e7a17ba9716501f1 Mon Sep 17 00:00:00 2001
From: Eric Heiden <eheiden@nvidia.com>
Date: Tue, 12 Dec 2023 16:48:34 -0800
Subject: [PATCH 31/50] Add depth rendering to OpenGLRenderer, document
 warp.render

---
 docs/index.rst                    |   1 +
 docs/modules/render.rst           |  38 +++
 examples/example_render_opengl.py |  25 +-
 warp/render/render_opengl.py      | 370 +++++++++++++++++++++---------
 4 files changed, 317 insertions(+), 117 deletions(-)
 create mode 100644 docs/modules/render.rst

diff --git a/docs/index.rst b/docs/index.rst
index 1c7c8f690..383df0f66 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -180,6 +180,7 @@ Full Table of Contents
     modules/sim
     modules/sparse
     modules/fem
+    modules/render
 
 .. toctree::
     :hidden:
diff --git a/docs/modules/render.rst b/docs/modules/render.rst
new file mode 100644
index 000000000..4988c8757
--- /dev/null
+++ b/docs/modules/render.rst
@@ -0,0 +1,38 @@
+warp.render
+===============================
+
+.. currentmodule:: warp.render
+
+The ``warp.render`` module provides a set of renderers that can be used for visualizing scenes involving shapes of various types.
+
+Built on top of these stand-alone renderers, the ``warp.sim.render`` module provides renderers that can be used to visualize scenes directly from ``warp.sim.ModelBuilder`` objects and update them from ``warp.sim.State`` objects.
+
+..
+    .. toctree::
+    :maxdepth: 2
+
+Stand-alone renderers
+---------------------
+
+The ``OpenGLRenderer`` provides an interactive renderer to play back animations in real time, the ``UsdRenderer`` provides a renderer that exports the scene to a USD file that can be rendered in a renderer of your choice.
+
+.. autoclass:: UsdRenderer
+    :members:
+
+
+.. autoclass:: OpenGLRenderer
+    :members:
+
+Simulation renderers
+--------------------
+
+Based on these renderers from ``warp.render``, the ``SimRendererUsd`` (which equals ``SimRenderer``) and ``SimRendererOpenGL`` classes from ``warp.sim.render`` are derived to populate the renderers directly from ``warp.sim.ModelBuilder`` scenes and update them from ``warp.sim.State`` objects.
+
+.. currentmodule:: warp.sim.render
+
+.. autoclass:: SimRendererUsd
+    :members:
+
+.. autoclass:: SimRendererOpenGL
+    :members:
+
diff --git a/examples/example_render_opengl.py b/examples/example_render_opengl.py
index 759db719c..8bf2ffa38 100644
--- a/examples/example_render_opengl.py
+++ b/examples/example_render_opengl.py
@@ -27,6 +27,8 @@
 custom_tile_arrangement = False
 # whether to display the pixels in a matplotlib figure
 show_plot = True
+# whether to render depth image to a Warp array
+render_mode = "depth"
 
 renderer = wp.render.OpenGLRenderer(vsync=False)
 instance_ids = []
@@ -51,11 +53,12 @@
 
 renderer.render_ground()
 
+channels = 1 if render_mode == "depth" else 3
 if show_plot:
     import matplotlib.pyplot as plt
 
     if split_up_tiles:
-        pixels = wp.zeros((num_tiles, renderer.tile_height, renderer.tile_width, 3), dtype=wp.float32)
+        pixels = wp.zeros((num_tiles, renderer.tile_height, renderer.tile_width, channels), dtype=wp.float32)
         ncols = int(np.ceil(np.sqrt(num_tiles)))
         nrows = int(np.ceil(num_tiles / float(ncols)))
         img_plots = []
@@ -70,17 +73,23 @@
             sharey=True,
             num=1,
         )
-        tile_temp = np.zeros((renderer.tile_height, renderer.tile_width, 3), dtype=np.float32)
+        tile_temp = np.zeros((renderer.tile_height, renderer.tile_width, channels), dtype=np.float32)
         for dim in range(ncols * nrows):
             ax = axes[dim // ncols, dim % ncols]
             if dim >= num_tiles:
                 ax.axis("off")
                 continue
-            img_plots.append(ax.imshow(tile_temp))
+            if render_mode == "depth":
+                img_plots.append(ax.imshow(tile_temp, vmin=renderer.camera_near_plane, vmax=renderer.camera_far_plane))
+            else:
+                img_plots.append(ax.imshow(tile_temp))
     else:
         fig = plt.figure(1)
-        pixels = wp.zeros((renderer.screen_height, renderer.screen_width, 3), dtype=wp.float32)
-        img_plot = plt.imshow(pixels.numpy())
+        pixels = wp.zeros((renderer.screen_height, renderer.screen_width, channels), dtype=wp.float32)
+        if render_mode == "depth":
+            img_plot = plt.imshow(pixels.numpy(), vmin=renderer.camera_near_plane, vmax=renderer.camera_far_plane)
+        else:
+            img_plot = plt.imshow(pixels.numpy())
 
     plt.ion()
     plt.show()
@@ -110,15 +119,15 @@
 
     if show_plot and plt.fignum_exists(1):
         if split_up_tiles:
-            pixel_shape = (num_tiles, renderer.tile_height, renderer.tile_width, 3)
+            pixel_shape = (num_tiles, renderer.tile_height, renderer.tile_width, channels)
         else:
-            pixel_shape = (renderer.screen_height, renderer.screen_width, 3)
+            pixel_shape = (renderer.screen_height, renderer.screen_width, channels)
 
         if pixel_shape != pixels.shape:
             # make sure we resize the pixels array to the right dimensions if the user resizes the window
             pixels = wp.zeros(pixel_shape, dtype=wp.float32)
 
-        renderer.get_pixels(pixels, split_up_tiles=split_up_tiles)
+        renderer.get_pixels(pixels, split_up_tiles=split_up_tiles, mode=render_mode)
 
         if split_up_tiles:
             pixels_np = pixels.numpy()
diff --git a/warp/render/render_opengl.py b/warp/render/render_opengl.py
index 0efc77805..6cff3a98d 100644
--- a/warp/render/render_opengl.py
+++ b/warp/render/render_opengl.py
@@ -13,12 +13,14 @@
 
 from collections import defaultdict
 from typing import List, Tuple, Union, Optional
+from enum import Enum
 
 import numpy as np
 import ctypes
 
 Mat44 = Union[List[float], List[List[float]], np.ndarray]
 
+
 wp.set_module_options({"enable_backward": False})
 
 shape_vertex_shader = """
@@ -175,12 +177,13 @@
 
 uniform vec3 color1;
 uniform vec3 color2;
+uniform float farPlane;
 
 uniform vec3 sunDirection;
 
 void main()
 {
-    float y = tanh(FragPos.y*0.01)*0.5+0.5;
+    float y = tanh(FragPos.y/farPlane*10.0)*0.5+0.5;
     float height = sqrt(1.0-y);
 
     float s = pow(0.5, 1.0 / 10.0);
@@ -222,6 +225,42 @@
 }
 """
 
+frame_depth_fragment_shader = """
+#version 330 core
+in vec2 TexCoord;
+
+out vec4 FragColor;
+
+uniform sampler2D textureSampler;
+
+vec3 bourkeColorMap(float v) {
+    vec3 c = vec3(1.0, 1.0, 1.0);
+
+    v = clamp(v, 0.0, 1.0); // Ensures v is between 0 and 1
+
+    if (v < 0.25) {
+        c.r = 0.0;
+        c.g = 4.0 * v;
+    } else if (v < 0.5) {
+        c.r = 0.0;
+        c.b = 1.0 + 4.0 * (0.25 - v);
+    } else if (v < 0.75) {
+        c.r = 4.0 * (v - 0.5);
+        c.b = 0.0;
+    } else {
+        c.g = 1.0 + 4.0 * (0.75 - v);
+        c.b = 0.0;
+    }
+
+    return c;
+}
+
+void main() {
+    float depth = texture(textureSampler, TexCoord).r;
+    FragColor = vec4(bourkeColorMap(sqrt(1.0 - depth)), 1.0);
+}
+"""
+
 
 @wp.kernel
 def update_vbo_transforms(
@@ -411,7 +450,7 @@ def assemble_gfx_vertices(
 
 
 @wp.kernel
-def copy_frame(
+def copy_rgb_frame(
     input_img: wp.array(dtype=wp.uint8),
     width: int,
     height: int,
@@ -432,7 +471,26 @@ def copy_frame(
 
 
 @wp.kernel
-def copy_frame_tiles(
+def copy_depth_frame(
+    input_img: wp.array(dtype=wp.float32),
+    width: int,
+    height: int,
+    near: float,
+    far: float,
+    # outputs
+    output_img: wp.array(dtype=wp.float32, ndim=3),
+):
+    w, v = wp.tid()
+    pixel = v * width + w
+    # flip vertically (OpenGL coordinates start at bottom)
+    v = height - v - 1
+    d = 2.0 * input_img[pixel] - 1.0
+    d = 2.0 * near * far / ((far - near) * d - near - far)
+    output_img[v, w, 0] = -d
+
+
+@wp.kernel
+def copy_rgb_frame_tiles(
     input_img: wp.array(dtype=wp.uint8),
     positions: wp.array(dtype=int, ndim=2),
     screen_width: int,
@@ -463,7 +521,34 @@ def copy_frame_tiles(
 
 
 @wp.kernel
-def copy_frame_tile(
+def copy_depth_frame_tiles(
+    input_img: wp.array(dtype=wp.float32),
+    positions: wp.array(dtype=int, ndim=2),
+    screen_width: int,
+    screen_height: int,
+    tile_height: int,
+    near: float,
+    far: float,
+    # outputs
+    output_img: wp.array(dtype=wp.float32, ndim=4),
+):
+    tile, x, y = wp.tid()
+    p = positions[tile]
+    qx = x + p[0]
+    qy = y + p[1]
+    pixel = qy * screen_width + qx
+    # flip vertically (OpenGL coordinates start at bottom)
+    y = tile_height - y - 1
+    if qx >= screen_width or qy >= screen_height:
+        output_img[tile, y, x, 0] = far
+        return  # prevent out-of-bounds access
+    d = 2.0 * input_img[pixel] - 1.0
+    d = 2.0 * near * far / ((far - near) * d - near - far)
+    output_img[tile, y, x, 0] = -d
+
+
+@wp.kernel
+def copy_rgb_frame_tile(
     input_img: wp.array(dtype=wp.uint8),
     offset_x: int,
     offset_y: int,
@@ -768,15 +853,19 @@ def __init__(
         up_axis="Y",
         screen_width=1024,
         screen_height=768,
-        near_plane=0.01,
-        far_plane=1000.0,
+        near_plane=1.0,
+        far_plane=100.0,
         camera_fov=45.0,
+        camera_pos=(0.0, 2.0, 10.0),
+        camera_front=(0.0, 0.0, -1.0),
+        camera_up=(0.0, 1.0, 0.0),
         background_color=(0.53, 0.8, 0.92),
         draw_grid=True,
         draw_sky=True,
         draw_axis=True,
         show_info=True,
         render_wireframe=False,
+        render_depth=False,
         axis_scale=1.0,
         vsync=False,
         headless=False,
@@ -804,6 +893,7 @@ def __init__(
         self.draw_axis = draw_axis
         self.show_info = show_info
         self.render_wireframe = render_wireframe
+        self.render_depth = render_depth
         self.enable_backface_culling = enable_backface_culling
 
         self._device = wp.get_cuda_device()
@@ -819,9 +909,9 @@ def __init__(
 
         self.screen_width, self.screen_height = self.window.get_framebuffer_size()
 
-        self._camera_pos = PyVec3(0.0, 2.0, 10.0)
-        self._camera_front = PyVec3(0.0, 0.0, -1.0)
-        self._camera_up = PyVec3(0.0, 1.0, 0.0)
+        self._camera_pos = PyVec3(*camera_pos)
+        self._camera_front = PyVec3(*camera_front)
+        self._camera_up = PyVec3(*camera_up)
         self._camera_speed = 0.04
         if isinstance(up_axis, int):
             self._camera_axis = up_axis
@@ -832,6 +922,7 @@ def __init__(
         self._first_mouse = True
         self._left_mouse_pressed = False
         self._keys_pressed = defaultdict(bool)
+        self._key_callbacks = []
 
         self.update_view_matrix()
         self.update_projection_matrix()
@@ -888,6 +979,7 @@ def __init__(
         self._tile_projection_matrices = None
 
         self._frame_texture = None
+        self._frame_depth_texture = None
         self._frame_fbo = None
         self._frame_pbo = None
 
@@ -903,6 +995,8 @@ def __init__(
 
         gl.glClearColor(*self.background_color, 1)
         gl.glEnable(gl.GL_DEPTH_TEST)
+        gl.glDepthMask(True)
+        gl.glDepthRange(0.0, 1.0)
 
         self._shape_shader = ShaderProgram(
             Shader(shape_vertex_shader, "vertex"), Shader(shape_fragment_shader, "fragment")
@@ -969,15 +1063,17 @@ def __init__(
 
             self._loc_sky_color1 = gl.glGetUniformLocation(self._sky_shader.id, str_buffer("color1"))
             self._loc_sky_color2 = gl.glGetUniformLocation(self._sky_shader.id, str_buffer("color2"))
+            self._loc_sky_far_plane = gl.glGetUniformLocation(self._sky_shader.id, str_buffer("farPlane"))
             gl.glUniform3f(self._loc_sky_color1, *background_color)
             # glUniform3f(self._loc_sky_color2, *np.clip(np.array(background_color)+0.5, 0.0, 1.0))
             gl.glUniform3f(self._loc_sky_color2, 0.8, 0.4, 0.05)
+            gl.glUniform1f(self._loc_sky_far_plane, self.camera_far_plane)
             self._loc_sky_view_pos = gl.glGetUniformLocation(self._sky_shader.id, str_buffer("viewPos"))
             gl.glUniform3f(
                 gl.glGetUniformLocation(self._sky_shader.id, str_buffer("sunDirection")), *self._sun_direction
             )
 
-        # Create VAO, VBO, and EBO
+        # create VAO, VBO, and EBO
         self._sky_vao = gl.GLuint()
         gl.glGenVertexArrays(1, self._sky_vao)
         gl.glBindVertexArray(self._sky_vao)
@@ -995,7 +1091,7 @@ def __init__(
         gl.glBindBuffer(gl.GL_ELEMENT_ARRAY_BUFFER, self._sky_ebo)
         gl.glBufferData(gl.GL_ELEMENT_ARRAY_BUFFER, indices.nbytes, indices.ctypes.data, gl.GL_STATIC_DRAW)
 
-        # Set up vertex attributes
+        # set up vertex attributes
         vertex_stride = vertices.shape[1] * vertices.itemsize
         # positions
         gl.glVertexAttribPointer(0, 3, gl.GL_FLOAT, gl.GL_FALSE, vertex_stride, ctypes.c_void_p(0))
@@ -1029,6 +1125,7 @@ def __init__(
 
         # create frame buffer for rendering to a texture
         self._frame_texture = None
+        self._frame_depth_texture = None
         self._frame_fbo = None
         self._setup_framebuffer()
 
@@ -1076,7 +1173,15 @@ def __init__(
         gl.glUseProgram(self._frame_shader.id)
         self._frame_loc_texture = gl.glGetUniformLocation(self._frame_shader.id, str_buffer("textureSampler"))
 
-        # Unbind the VBO and VAO
+        self._frame_depth_shader = ShaderProgram(
+            Shader(frame_vertex_shader, "vertex"), Shader(frame_depth_fragment_shader, "fragment")
+        )
+        gl.glUseProgram(self._frame_depth_shader.id)
+        self._frame_loc_depth_texture = gl.glGetUniformLocation(
+            self._frame_depth_shader.id, str_buffer("textureSampler")
+        )
+
+        # unbind the VBO and VAO
         gl.glBindBuffer(gl.GL_ARRAY_BUFFER, 0)
         gl.glBindVertexArray(0)
 
@@ -1322,7 +1427,11 @@ def _setup_framebuffer(self):
         if self._frame_texture is None:
             self._frame_texture = gl.GLuint()
             gl.glGenTextures(1, self._frame_texture)
+        if self._frame_depth_texture is None:
+            self._frame_depth_texture = gl.GLuint()
+            gl.glGenTextures(1, self._frame_depth_texture)
 
+        # set up RGB texture
         gl.glBindFramebuffer(gl.GL_FRAMEBUFFER, 0)
         gl.glBindBuffer(gl.GL_PIXEL_UNPACK_BUFFER, 0)
         gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_texture)
@@ -1339,6 +1448,22 @@ def _setup_framebuffer(self):
         )
         gl.glTexParameteri(gl.GL_TEXTURE_2D, gl.GL_TEXTURE_MIN_FILTER, gl.GL_LINEAR)
         gl.glTexParameteri(gl.GL_TEXTURE_2D, gl.GL_TEXTURE_MAG_FILTER, gl.GL_LINEAR)
+
+        # set up depth texture
+        gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_depth_texture)
+        gl.glTexImage2D(
+            gl.GL_TEXTURE_2D,
+            0,
+            gl.GL_DEPTH_COMPONENT32,
+            self.screen_width,
+            self.screen_height,
+            0,
+            gl.GL_DEPTH_COMPONENT,
+            gl.GL_FLOAT,
+            None,
+        )
+        gl.glTexParameteri(gl.GL_TEXTURE_2D, gl.GL_TEXTURE_MIN_FILTER, gl.GL_LINEAR)
+        gl.glTexParameteri(gl.GL_TEXTURE_2D, gl.GL_TEXTURE_MAG_FILTER, gl.GL_LINEAR)
         gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
 
         # create a framebuffer object (FBO)
@@ -1351,15 +1476,9 @@ def _setup_framebuffer(self):
             gl.glFramebufferTexture2D(
                 gl.GL_FRAMEBUFFER, gl.GL_COLOR_ATTACHMENT0, gl.GL_TEXTURE_2D, self._frame_texture, 0
             )
-
-            self._frame_depth_renderbuffer = gl.GLuint()
-            gl.glGenRenderbuffers(1, self._frame_depth_renderbuffer)
-            gl.glBindRenderbuffer(gl.GL_RENDERBUFFER, self._frame_depth_renderbuffer)
-            gl.glRenderbufferStorage(gl.GL_RENDERBUFFER, gl.GL_DEPTH_COMPONENT, self.screen_width, self.screen_height)
-
-            # attach the depth renderbuffer to the FBO
-            gl.glFramebufferRenderbuffer(
-                gl.GL_FRAMEBUFFER, gl.GL_DEPTH_ATTACHMENT, gl.GL_RENDERBUFFER, self._frame_depth_renderbuffer
+            # attach the depth texture to the FBO as its depth attachment
+            gl.glFramebufferTexture2D(
+                gl.GL_FRAMEBUFFER, gl.GL_DEPTH_ATTACHMENT, gl.GL_TEXTURE_2D, self._frame_depth_texture, 0
             )
 
             if gl.glCheckFramebufferStatus(gl.GL_FRAMEBUFFER) != gl.GL_FRAMEBUFFER_COMPLETE:
@@ -1367,13 +1486,6 @@ def _setup_framebuffer(self):
                 gl.glBindFramebuffer(gl.GL_FRAMEBUFFER, 0)
                 sys.exit(1)
 
-            gl.glBindRenderbuffer(gl.GL_RENDERBUFFER, 0)
-        else:
-            # rescale framebuffer
-            gl.glBindRenderbuffer(gl.GL_RENDERBUFFER, self._frame_depth_renderbuffer)
-            gl.glRenderbufferStorage(gl.GL_RENDERBUFFER, gl.GL_DEPTH_COMPONENT, self.screen_width, self.screen_height)
-            gl.glBindRenderbuffer(gl.GL_RENDERBUFFER, 0)
-
         # unbind the FBO (switch back to the default framebuffer)
         gl.glBindFramebuffer(gl.GL_FRAMEBUFFER, 0)
 
@@ -1383,7 +1495,11 @@ def _setup_framebuffer(self):
         gl.glBindBuffer(gl.GL_PIXEL_PACK_BUFFER, self._frame_pbo)  # binding to this buffer
 
         # allocate memory for PBO
-        pixels = np.zeros((self.screen_height, self.screen_width, 3), dtype=np.uint8)
+        rgb_bytes_per_pixel = 3
+        depth_bytes_per_pixel = 4
+        pixels = np.zeros(
+            (self.screen_height, self.screen_width, rgb_bytes_per_pixel + depth_bytes_per_pixel), dtype=np.uint8
+        )
         gl.glBufferData(gl.GL_PIXEL_PACK_BUFFER, pixels.nbytes, pixels.ctypes.data, gl.GL_DYNAMIC_DRAW)
         gl.glBindBuffer(gl.GL_PIXEL_PACK_BUFFER, 0)
 
@@ -1542,7 +1658,6 @@ def _draw(self):
 
         if self._frame_fbo is not None:
             gl.glBindFramebuffer(gl.GL_FRAMEBUFFER, self._frame_fbo)
-            gl.glBindBuffer(gl.GL_PIXEL_UNPACK_BUFFER, self._frame_fbo)
 
         gl.glClearColor(*self.background_color, 1)
         gl.glClear(gl.GL_COLOR_BUFFER_BIT | gl.GL_DEPTH_BUFFER_BIT)
@@ -1580,15 +1695,26 @@ def _draw(self):
 
         # render frame buffer texture to screen
         if self._frame_fbo is not None:
-            with self._frame_shader:
-                gl.glActiveTexture(gl.GL_TEXTURE0)
-                gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_texture)
-                gl.glUniform1i(self._frame_loc_texture, 0)
+            if self.render_depth:
+                with self._frame_depth_shader:
+                    gl.glActiveTexture(gl.GL_TEXTURE0)
+                    gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_depth_texture)
+                    gl.glUniform1i(self._frame_loc_depth_texture, 0)
+
+                    gl.glBindVertexArray(self._frame_vao)
+                    gl.glDrawElements(gl.GL_TRIANGLES, len(self._frame_indices), gl.GL_UNSIGNED_INT, None)
+                    gl.glBindVertexArray(0)
+                    gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
+            else:
+                with self._frame_shader:
+                    gl.glActiveTexture(gl.GL_TEXTURE0)
+                    gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_texture)
+                    gl.glUniform1i(self._frame_loc_texture, 0)
 
-                gl.glBindVertexArray(self._frame_vao)
-                gl.glDrawElements(gl.GL_TRIANGLES, len(self._frame_indices), gl.GL_UNSIGNED_INT, None)
-                gl.glBindVertexArray(0)
-                gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
+                    gl.glBindVertexArray(self._frame_vao)
+                    gl.glDrawElements(gl.GL_TRIANGLES, len(self._frame_indices), gl.GL_UNSIGNED_INT, None)
+                    gl.glBindVertexArray(0)
+                    gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
 
         # check for OpenGL errors
         # check_gl_error()
@@ -1767,9 +1893,17 @@ def _key_press_callback(self, symbol, modifiers):
             self.show_info = not self.show_info
         if symbol == pyglet.window.key.X:
             self.render_wireframe = not self.render_wireframe
+        if symbol == pyglet.window.key.T:
+            self.render_depth = not self.render_depth
         if symbol == pyglet.window.key.B:
             self.enable_backface_culling = not self.enable_backface_culling
 
+        for cb in self._key_callbacks:
+            cb(symbol, modifiers)
+
+    def register_key_press_callback(self, callback):
+        self._key_callbacks.append(callback)
+
     def _window_resize_callback(self, width, height):
         self._first_mouse = True
         self.screen_width, self.screen_height = self.window.get_framebuffer_size()
@@ -1861,7 +1995,7 @@ def allocate_shape_instances(self):
             gl.glDeleteBuffers(1, self._instance_color1_buffer)
             gl.glDeleteBuffers(1, self._instance_color2_buffer)
 
-        # Create instance buffer and bind it as an instanced array
+        # create instance buffer and bind it as an instanced array
         self._instance_transform_gl_buffer = gl.GLuint()
         gl.glGenBuffers(1, self._instance_transform_gl_buffer)
         gl.glBindBuffer(gl.GL_ARRAY_BUFFER, self._instance_transform_gl_buffer)
@@ -1869,7 +2003,7 @@ def allocate_shape_instances(self):
         transforms = np.tile(np.diag(np.ones(4, dtype=np.float32)), (len(self._instances), 1, 1))
         gl.glBufferData(gl.GL_ARRAY_BUFFER, transforms.nbytes, transforms.ctypes.data, gl.GL_DYNAMIC_DRAW)
 
-        # Create CUDA buffer for instance transforms
+        # create CUDA buffer for instance transforms
         self._instance_transform_cuda_buffer = wp.RegisteredGLBuffer(
             int(self._instance_transform_gl_buffer.value), self._device
         )
@@ -1897,7 +2031,7 @@ def allocate_shape_instances(self):
         gl.glBindBuffer(gl.GL_ARRAY_BUFFER, self._instance_color2_buffer)
         gl.glBufferData(gl.GL_ARRAY_BUFFER, colors2.nbytes, colors2.ctypes.data, gl.GL_STATIC_DRAW)
 
-        # Set up instance attribute pointers
+        # set up instance attribute pointers
         matrix_size = transforms[0].nbytes
 
         instance_ids = []
@@ -2018,9 +2152,27 @@ def save(self):
             self.clear()
             self.app.event_loop.exit()
 
-    def get_pixels(self, target_image: wp.array, split_up_tiles=True):
+    def get_pixels(self, target_image: wp.array, split_up_tiles=True, mode="rgb"):
+        """
+        Read the pixels from the frame buffer (RGB or depth are supported) into the given array.
+
+        If `split_up_tiles` is False, array must be of shape (screen_height, screen_width, 3) for RGB mode or
+        (screen_height, screen_width, 1) for depth mode.
+        If `split_up_tiles` is True, the pixels will be split up into tiles (see :attr:`tile_width` and :attr:`tile_height` for dimensions):
+        array must be of shape (num_tiles, tile_height, tile_width, 3) for RGB mode or (num_tiles, tile_height, tile_width, 1) for depth mode.
+
+        Args:
+            target_image (array): The array to read the pixels into. Must have float32 as dtype and be on a CUDA device.
+            split_up_tiles (bool): Whether to split up the viewport into tiles, see :meth:`setup_tiled_rendering`.
+            mode (str): can be either "rgb" or "depth"
+
+        Returns:
+            bool: Whether the pixels were successfully read.
+        """
         from pyglet import gl
 
+        channels = 3 if mode == "rgb" else 1
+
         if split_up_tiles:
             assert (
                 self._tile_width is not None and self._tile_height is not None
@@ -2035,20 +2187,26 @@ def get_pixels(self, target_image: wp.array, split_up_tiles=True):
                 self.num_tiles,
                 self._tile_height,
                 self._tile_width,
-                3,
-            ), f"Shape of `target_image` array does not match {self.num_tiles} x {self.screen_height} x {self.screen_width} x 3"
+                channels,
+            ), f"Shape of `target_image` array does not match {self.num_tiles} x {self.screen_height} x {self.screen_width} x {channels}"
         else:
             assert target_image.shape == (
                 self.screen_height,
                 self.screen_width,
-                3,
-            ), f"Shape of `target_image` array does not match {self.screen_height} x {self.screen_width} x 3"
+                channels,
+            ), f"Shape of `target_image` array does not match {self.screen_height} x {self.screen_width} x {channels}"
 
         gl.glBindBuffer(gl.GL_PIXEL_PACK_BUFFER, self._frame_pbo)
-        gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_texture)
+        if mode == "rgb":
+            gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_texture)
+        if mode == "depth":
+            gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_depth_texture)
         try:
             # read screen texture into PBO
-            gl.glGetTexImage(gl.GL_TEXTURE_2D, 0, gl.GL_RGB, gl.GL_UNSIGNED_BYTE, ctypes.c_void_p(0))
+            if mode == "rgb":
+                gl.glGetTexImage(gl.GL_TEXTURE_2D, 0, gl.GL_RGB, gl.GL_UNSIGNED_BYTE, ctypes.c_void_p(0))
+            elif mode == "depth":
+                gl.glGetTexImage(gl.GL_TEXTURE_2D, 0, gl.GL_DEPTH_COMPONENT, gl.GL_FLOAT, ctypes.c_void_p(0))
         except gl.GLException:
             # this can happen if the window is closed/being moved to a different display
             gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
@@ -2061,63 +2219,54 @@ def get_pixels(self, target_image: wp.array, split_up_tiles=True):
             int(self._frame_pbo.value), self._device, wp.RegisteredGLBuffer.WRITE_DISCARD
         )
         screen_size = self.screen_height * self.screen_width
-        img = pbo_buffer.map(dtype=wp.uint8, shape=(screen_size * 3))
+        if mode == "rgb":
+            img = pbo_buffer.map(dtype=wp.uint8, shape=(screen_size * channels))
+        elif mode == "depth":
+            img = pbo_buffer.map(dtype=wp.float32, shape=(screen_size * channels))
         img = img.to(target_image.device)
         if split_up_tiles:
             positions = wp.array(self._tile_viewports, ndim=2, dtype=wp.int32, device=target_image.device)
-            wp.launch(
-                copy_frame_tiles,
-                dim=(self.num_tiles, self._tile_width, self._tile_height),
-                inputs=[img, positions, self.screen_width, self.screen_height, self._tile_height],
-                outputs=[target_image],
-                device=target_image.device,
-            )
+            if mode == "rgb":
+                wp.launch(
+                    copy_rgb_frame_tiles,
+                    dim=(self.num_tiles, self._tile_width, self._tile_height),
+                    inputs=[img, positions, self.screen_width, self.screen_height, self._tile_height],
+                    outputs=[target_image],
+                    device=target_image.device,
+                )
+            elif mode == "depth":
+                wp.launch(
+                    copy_depth_frame_tiles,
+                    dim=(self.num_tiles, self._tile_width, self._tile_height),
+                    inputs=[
+                        img,
+                        positions,
+                        self.screen_width,
+                        self.screen_height,
+                        self._tile_height,
+                        self.camera_near_plane,
+                        self.camera_far_plane,
+                    ],
+                    outputs=[target_image],
+                    device=target_image.device,
+                )
         else:
-            wp.launch(
-                copy_frame,
-                dim=(self.screen_width, self.screen_height),
-                inputs=[img, self.screen_width, self.screen_height],
-                outputs=[target_image],
-                device=target_image.device,
-            )
-        pbo_buffer.unmap()
-        return True
-
-    def get_tile_pixels(self, tile_id: int, target_image: wp.array):
-        from pyglet import gl
-
-        viewport = self._tile_viewports[tile_id]
-        assert target_image.shape == (
-            viewport[3],
-            viewport[2],
-            3,
-        ), f"Shape of `target_image` array does not match {viewport[3]} x {viewport[2]} x 3"
-        gl.glBindBuffer(gl.GL_PIXEL_PACK_BUFFER, self._frame_pbo)
-        gl.glBindTexture(gl.GL_TEXTURE_2D, self._frame_texture)
-        try:
-            # read screen texture into PBO
-            gl.glGetTexImage(gl.GL_TEXTURE_2D, 0, gl.GL_RGB, gl.GL_UNSIGNED_BYTE, ctypes.c_void_p(0))
-        except gl.GLException:
-            # this can happen if the window is closed/being moved to a different display
-            gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
-            gl.glBindBuffer(gl.GL_PIXEL_PACK_BUFFER, 0)
-            return False
-        gl.glBindTexture(gl.GL_TEXTURE_2D, 0)
-        gl.glBindBuffer(gl.GL_PIXEL_PACK_BUFFER, 0)
-
-        pbo_buffer = wp.RegisteredGLBuffer(
-            int(self._frame_pbo.value), self._device, wp.RegisteredGLBuffer.WRITE_DISCARD
-        )
-        screen_size = self.screen_height * self.screen_width
-        img = pbo_buffer.map(dtype=wp.uint8, shape=(screen_size * 3))
-        img = img.to(target_image.device)
-        wp.launch(
-            copy_frame_tiles,
-            dim=(self.num_tiles, self._tile_width, self._tile_height),
-            inputs=[img, viewport[0], viewport[1], self.screen_width, self.screen_height, self._tile_height],
-            outputs=[target_image],
-            device=target_image.device,
-        )
+            if mode == "rgb":
+                wp.launch(
+                    copy_rgb_frame,
+                    dim=(self.screen_width, self.screen_height),
+                    inputs=[img, self.screen_width, self.screen_height],
+                    outputs=[target_image],
+                    device=target_image.device,
+                )
+            elif mode == "depth":
+                wp.launch(
+                    copy_depth_frame,
+                    dim=(self.screen_width, self.screen_height),
+                    inputs=[img, self.screen_width, self.screen_height, self.camera_near_plane, self.camera_far_plane],
+                    outputs=[target_image],
+                    device=target_image.device,
+                )
         pbo_buffer.unmap()
         return True
 
@@ -2570,7 +2719,7 @@ def _render_lines(self, name: str, lines, color: tuple, radius: float = 0.01):
         if name not in self._shape_instancers:
             instancer = ShapeInstancer(self._shape_shader, self._device)
             vertices, indices = self._create_capsule_mesh(radius, 0.5)
-            if color is None or isinstance(color, list):
+            if color is None or isinstance(color, list) and len(color) > 0 and isinstance(color[0], list):
                 color = tab10_color_map(len(self._shape_geo_hash))
             instancer.register_shape(vertices, indices, color, color)
             instancer.allocate_instances(np.zeros((len(lines), 3)))
@@ -2640,7 +2789,12 @@ def update_shape_vertices(self, shape, points):
         cuda_buffer.unmap()
 
     @staticmethod
-    def _create_sphere_mesh(radius=1.0, num_latitudes=default_num_segments, num_longitudes=default_num_segments, reverse_winding=False):
+    def _create_sphere_mesh(
+        radius=1.0,
+        num_latitudes=default_num_segments,
+        num_longitudes=default_num_segments,
+        reverse_winding=False,
+    ):
         vertices = []
         indices = []
 
@@ -2754,7 +2908,7 @@ def _create_cylinder_mesh(radius, half_height, up_axis=1, segments=default_num_s
             top_radius = radius
         side_slope = -np.arctan2(top_radius - radius, 2 * half_height)
 
-        # Create the cylinder base and top vertices
+        # create the cylinder base and top vertices
         for j in (-1, 1):
             center_index = max(j, 0)
             if j == 1:
@@ -2782,12 +2936,10 @@ def _create_cylinder_mesh(radius, half_height, up_axis=1, segments=default_num_s
                 vertex = np.hstack([position[[x_dir, y_dir, z_dir]], normal[[x_dir, y_dir, z_dir]], uv])
                 cap_vertices.append(vertex)
 
-                indices.extend(
-                    [center_index, i + center_index * segments + 2,
-                        (i + 1) % segments + center_index * segments + 2][::-j]
-                )
+                cs = center_index * segments
+                indices.extend([center_index, i + cs + 2, (i + 1) % segments + cs + 2][::-j])
 
-        # Create the cylinder side indices
+        # create the cylinder side indices
         for i in range(segments):
             index1 = len(cap_vertices) + i + segments
             index2 = len(cap_vertices) + ((i + 1) % segments) + segments

From 276ce36a61715d0a99ce8b2d7e8c77c76b05fb98 Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Tue, 12 Dec 2023 16:49:24 -0800
Subject: [PATCH 32/50] Fix multi-gpu synchronization issue in sparse.py

---
 warp/sparse.py | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/warp/sparse.py b/warp/sparse.py
index 70d831596..148064f28 100644
--- a/warp/sparse.py
+++ b/warp/sparse.py
@@ -1,9 +1,9 @@
-from typing import Tuple, Any, Optional, Union, TypeVar, Generic
+from typing import Any, Generic, Optional, Tuple, TypeVar, Union
 
 import warp as wp
 import warp.types
-from warp.types import Matrix, Vector, Rows, Cols, Scalar, Array
 import warp.utils
+from warp.types import Array, Cols, Matrix, Rows, Scalar, Vector
 
 # typing hints
 
@@ -54,7 +54,7 @@ def block_size(self) -> int:
 
     @property
     def shape(self) -> Tuple[int, int]:
-        """Shape of the matrix, i.e. number of rows/columns of blocks times number of rows/columsn per block"""
+        """Shape of the matrix, i.e. number of rows/columns of blocks times number of rows/columns per block"""
         block_shape = self.block_shape
         return (self.nrow * block_shape[0], self.ncol * block_shape[1])
 
@@ -193,7 +193,7 @@ def bsr_set_from_triplets(
     elif values.ndim == 3:
         if values.shape[1:] != dest.block_shape:
             raise ValueError(
-                f"Last two dimensions in values array ({values.shape[1:]}) shoudl correspond to matrix block shape {(dest.block_shape)})"
+                f"Last two dimensions in values array ({values.shape[1:]}) should correspond to matrix block shape {(dest.block_shape)})"
             )
 
         if warp.types.type_scalar_type(values.dtype) != dest.scalar_type:
@@ -266,7 +266,7 @@ def bsr_assign(dest: BsrMatrix[BlockType[Rows, Cols, Scalar]], src: BsrMatrix[Bl
 
 
 def bsr_copy(A: BsrMatrix, scalar_type: Optional[Scalar] = None):
-    """Returns a copy of matrix ``A``, possibly chaning its scalar type.
+    """Returns a copy of matrix ``A``, possibly changing its scalar type.
 
     Args:
        scalar_type: If provided, the returned matrix will use this scalar type instead of the one from `A`.
@@ -992,7 +992,7 @@ def bsr_mm(
     # Get back total counts on host
     if device.is_cuda:
         wp.copy(dest=work_arrays._pinned_count_buffer, src=work_arrays._mm_row_counts, src_offset=z.nrow, count=1)
-        wp.synchronize_stream(wp.get_stream())
+        wp.synchronize_stream(wp.get_stream(device))
         mm_nnz = int(work_arrays._pinned_count_buffer.numpy()[0])
     else:
         mm_nnz = int(work_arrays._mm_row_counts.numpy()[z.nrow])

From c19a704864e269f4d1244032e5e9e4d89e311844 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Thu, 26 Oct 2023 15:23:00 +1300
Subject: [PATCH 33/50] Add operators with scalar as the dividend

---
 warp/builtins.py        | 21 +++++++++++++++++
 warp/native/exports.h   | 51 +++++++++++++++++++++++++++++++++++++++++
 warp/native/mat.h       | 33 ++++++++++++++++++++++++--
 warp/native/quat.h      | 19 +++++++++++++++
 warp/native/vec.h       | 50 ++++++++++++++++++++++++++++++++++++++++
 warp/tests/test_mat.py  | 28 ++++++++++++++++++++++
 warp/tests/test_quat.py | 28 ++++++++++++++++++++++
 warp/tests/test_vec.py  | 28 ++++++++++++++++++++++
 warp/types.py           | 16 ++-----------
 9 files changed, 258 insertions(+), 16 deletions(-)

diff --git a/warp/builtins.py b/warp/builtins.py
index c7eb52014..0e72173a7 100644
--- a/warp/builtins.py
+++ b/warp/builtins.py
@@ -3228,6 +3228,13 @@ def mul_matmat_value_func(arg_types, kwds, _):
     doc="",
     group="Operators",
 )
+add_builtin(
+    "div",
+    input_types={"x": Scalar, "y": vector(length=Any, dtype=Scalar)},
+    value_func=scalar_mul_value_func(vector(length=Any, dtype=Scalar)),
+    doc="",
+    group="Operators",
+)
 add_builtin(
     "div",
     input_types={"x": matrix(shape=(Any, Any), dtype=Scalar), "y": Scalar},
@@ -3235,6 +3242,13 @@ def mul_matmat_value_func(arg_types, kwds, _):
     doc="",
     group="Operators",
 )
+add_builtin(
+    "div",
+    input_types={"x": Scalar, "y": matrix(shape=(Any, Any), dtype=Scalar)},
+    value_func=scalar_mul_value_func(matrix(shape=(Any, Any), dtype=Scalar)),
+    doc="",
+    group="Operators",
+)
 add_builtin(
     "div",
     input_types={"x": quaternion(dtype=Scalar), "y": Scalar},
@@ -3242,6 +3256,13 @@ def mul_matmat_value_func(arg_types, kwds, _):
     doc="",
     group="Operators",
 )
+add_builtin(
+    "div",
+    input_types={"x": Scalar, "y": quaternion(dtype=Scalar)},
+    value_func=scalar_mul_value_func(quaternion(dtype=Scalar)),
+    doc="",
+    group="Operators",
+)
 
 add_builtin(
     "floordiv",
diff --git a/warp/native/exports.h b/warp/native/exports.h
index 78b2f14d2..931680e4c 100644
--- a/warp/native/exports.h
+++ b/warp/native/exports.h
@@ -1302,6 +1302,42 @@ WP_API void builtin_div_vec4ul_uint64(vec4ul x, uint64 y, vec4ul* ret) { *ret =
 WP_API void builtin_div_vec2ub_uint8(vec2ub x, uint8 y, vec2ub* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_vec3ub_uint8(vec3ub x, uint8 y, vec3ub* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_vec4ub_uint8(vec4ub x, uint8 y, vec4ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_vec2h(float16 x, vec2h y, vec2h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_vec3h(float16 x, vec3h y, vec3h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_vec4h(float16 x, vec4h y, vec4h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_spatial_vectorh(float16 x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_vec2f(float32 x, vec2f y, vec2f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_vec3f(float32 x, vec3f y, vec3f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_vec4f(float32 x, vec4f y, vec4f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_spatial_vectorf(float32 x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_vec2d(float64 x, vec2d y, vec2d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_vec3d(float64 x, vec3d y, vec3d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_vec4d(float64 x, vec4d y, vec4d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_spatial_vectord(float64 x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int16_vec2s(int16 x, vec2s y, vec2s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int16_vec3s(int16 x, vec3s y, vec3s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int16_vec4s(int16 x, vec4s y, vec4s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int32_vec2i(int32 x, vec2i y, vec2i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int32_vec3i(int32 x, vec3i y, vec3i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int32_vec4i(int32 x, vec4i y, vec4i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int64_vec2l(int64 x, vec2l y, vec2l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int64_vec3l(int64 x, vec3l y, vec3l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int64_vec4l(int64 x, vec4l y, vec4l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int8_vec2b(int8 x, vec2b y, vec2b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int8_vec3b(int8 x, vec3b y, vec3b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int8_vec4b(int8 x, vec4b y, vec4b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint16_vec2us(uint16 x, vec2us y, vec2us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint16_vec3us(uint16 x, vec3us y, vec3us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint16_vec4us(uint16 x, vec4us y, vec4us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint32_vec2ui(uint32 x, vec2ui y, vec2ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint32_vec3ui(uint32 x, vec3ui y, vec3ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint32_vec4ui(uint32 x, vec4ui y, vec4ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint64_vec2ul(uint64 x, vec2ul y, vec2ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint64_vec3ul(uint64 x, vec3ul y, vec3ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint64_vec4ul(uint64 x, vec4ul y, vec4ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint8_vec2ub(uint8 x, vec2ub y, vec2ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint8_vec3ub(uint8 x, vec3ub y, vec3ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint8_vec4ub(uint8 x, vec4ub y, vec4ub* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_mat22h_float16(mat22h x, float16 y, mat22h* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_mat33h_float16(mat33h x, float16 y, mat33h* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_mat44h_float16(mat44h x, float16 y, mat44h* ret) { *ret = wp::div(x, y); }
@@ -1314,9 +1350,24 @@ WP_API void builtin_div_mat22d_float64(mat22d x, float64 y, mat22d* ret) { *ret
 WP_API void builtin_div_mat33d_float64(mat33d x, float64 y, mat33d* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_mat44d_float64(mat44d x, float64 y, mat44d* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_spatial_matrixd_float64(spatial_matrixd x, float64 y, spatial_matrixd* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_mat22h(float16 x, mat22h y, mat22h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_mat33h(float16 x, mat33h y, mat33h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_mat44h(float16 x, mat44h y, mat44h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_spatial_matrixh(float16 x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_mat22f(float32 x, mat22f y, mat22f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_mat33f(float32 x, mat33f y, mat33f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_mat44f(float32 x, mat44f y, mat44f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_spatial_matrixf(float32 x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_mat22d(float64 x, mat22d y, mat22d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_mat33d(float64 x, mat33d y, mat33d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_mat44d(float64 x, mat44d y, mat44d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_spatial_matrixd(float64 x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_quath_float16(quath x, float16 y, quath* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_quatf_float32(quatf x, float32 y, quatf* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_quatd_float64(quatd x, float64 y, quatd* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_quath(float16 x, quath y, quath* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_quatf(float32 x, quatf y, quatf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_quatd(float64 x, quatd y, quatd* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_floordiv_float16_float16(float16 x, float16 y, float16* ret) { *ret = wp::floordiv(x, y); }
 WP_API void builtin_floordiv_float32_float32(float32 x, float32 y, float32* ret) { *ret = wp::floordiv(x, y); }
 WP_API void builtin_floordiv_float64_float64(float64 x, float64 y, float64* ret) { *ret = wp::floordiv(x, y); }
diff --git a/warp/native/mat.h b/warp/native/mat.h
index 40dfbdcc7..da3b9c9df 100644
--- a/warp/native/mat.h
+++ b/warp/native/mat.h
@@ -437,7 +437,22 @@ inline CUDA_CALLABLE mat_t<Rows,Cols,Type> div(const mat_t<Rows,Cols,Type>& a, T
         }
     }
 
-    return t;   
+    return t;
+}
+
+template<unsigned Rows, unsigned Cols, typename Type>
+inline CUDA_CALLABLE mat_t<Rows,Cols,Type> div(Type b, const mat_t<Rows,Cols,Type>& a)
+{
+    mat_t<Rows,Cols,Type> t;
+    for (unsigned i=0; i < Rows; ++i)
+    {
+        for (unsigned j=0; j < Cols; ++j)
+        {
+            t.data[i][j] = b / a.data[i][j];
+        }
+    }
+
+    return t;
 }
 
 template<unsigned Rows, unsigned Cols, typename Type>
@@ -452,7 +467,7 @@ inline CUDA_CALLABLE mat_t<Rows,Cols,Type> mul(const mat_t<Rows,Cols,Type>& a, T
         }
     }
 
-    return t;   
+    return t;
 }
 
 template<unsigned Rows, unsigned Cols, typename Type>
@@ -944,6 +959,20 @@ inline CUDA_CALLABLE void adj_div(const mat_t<Rows,Cols,Type>& a, Type s, mat_t<
     }
 }
 
+template<unsigned Rows, unsigned Cols, typename Type>
+inline CUDA_CALLABLE void adj_div(Type s, const mat_t<Rows,Cols,Type>& a, Type& adj_s, mat_t<Rows,Cols,Type>& adj_a, const mat_t<Rows,Cols,Type>& adj_ret)
+{
+    adj_s -= tensordot(a , adj_ret)/ (s * s); // - a / s^2
+
+    for (unsigned i=0; i < Rows; ++i)
+    {
+        for (unsigned j=0; j < Cols; ++j)
+        {
+            adj_a.data[i][j] += s / adj_ret.data[i][j];
+        }
+    }
+}
+
 template<unsigned Rows, unsigned Cols, typename Type>
 inline CUDA_CALLABLE void adj_mul(const mat_t<Rows,Cols,Type>& a, Type b, mat_t<Rows,Cols,Type>& adj_a, Type& adj_b, const mat_t<Rows,Cols,Type>& adj_ret)
 {
diff --git a/warp/native/quat.h b/warp/native/quat.h
index c554e243c..dc55b0f1b 100644
--- a/warp/native/quat.h
+++ b/warp/native/quat.h
@@ -225,12 +225,24 @@ inline CUDA_CALLABLE quat_t<Type> div(quat_t<Type> q, Type s)
     return quat_t<Type>(q.x/s, q.y/s, q.z/s, q.w/s);
 }
 
+template<typename Type>
+inline CUDA_CALLABLE quat_t<Type> div(Type s, quat_t<Type> q)
+{
+    return quat_t<Type>(s/q.x, s/q.y, s/q.z, s/q.w);
+}
+
 template<typename Type>
 inline CUDA_CALLABLE quat_t<Type> operator / (quat_t<Type> a, Type s)
 {
     return div(a,s);
 }
 
+template<typename Type>
+inline CUDA_CALLABLE quat_t<Type> operator / (Type s, quat_t<Type> a)
+{
+    return div(s,a);
+}
+
 template<typename Type>
 inline CUDA_CALLABLE quat_t<Type> operator*(Type s, const quat_t<Type>& a)
 {
@@ -613,6 +625,13 @@ inline CUDA_CALLABLE void adj_div(quat_t<Type> a, Type s, quat_t<Type>& adj_a, T
     adj_a += adj_ret / s;
 }
 
+template<typename Type>
+inline CUDA_CALLABLE void adj_div(Type s, quat_t<Type> a, Type& adj_s, quat_t<Type>& adj_a, const quat_t<Type>& adj_ret)
+{
+    adj_s -= dot(a, adj_ret)/ (s * s); // - a / s^2
+    adj_a += s / adj_ret;
+}
+
 template<typename Type>
 inline CUDA_CALLABLE void adj_quat_rotate(const quat_t<Type>& q, const vec_t<3,Type>& p, quat_t<Type>& adj_q, vec_t<3,Type>& adj_p, const vec_t<3,Type>& adj_ret)
 {
diff --git a/warp/native/vec.h b/warp/native/vec.h
index e5abeca9b..52840f71c 100644
--- a/warp/native/vec.h
+++ b/warp/native/vec.h
@@ -284,12 +284,41 @@ inline CUDA_CALLABLE vec_t<2, Type> div(vec_t<2, Type> a, Type s)
     return vec_t<2, Type>(a.c[0]/s,a.c[1]/s);
 }
 
+template<unsigned Length, typename Type>
+inline CUDA_CALLABLE vec_t<Length, Type> div(Type s, vec_t<Length, Type> a)
+{
+    vec_t<Length, Type> ret;
+    for (unsigned i=0; i < Length; ++i)
+    {
+        ret[i] = s / a[i];
+    }
+    return ret;
+}
+
+template<typename Type>
+inline CUDA_CALLABLE vec_t<3, Type> div(Type s, vec_t<3, Type> a)
+{
+    return vec_t<3, Type>(s/a.c[0],s/a.c[1],s/a.c[2]);
+}
+
+template<typename Type>
+inline CUDA_CALLABLE vec_t<2, Type> div(Type s, vec_t<2, Type> a)
+{
+    return vec_t<2, Type>(s/a.c[0],s/a.c[1]);
+}
+
 template<unsigned Length, typename Type>
 inline CUDA_CALLABLE vec_t<Length, Type> operator / (vec_t<Length, Type> a, Type s)
 {
     return div(a,s);
 }
 
+template<unsigned Length, typename Type>
+inline CUDA_CALLABLE vec_t<Length, Type> operator / (Type s, vec_t<Length, Type> a)
+{
+    return div(s, a);
+}
+
 // component wise division
 template<unsigned Length, typename Type>
 inline CUDA_CALLABLE vec_t<Length, Type> cw_div(vec_t<Length, Type> a, vec_t<Length, Type> b)
@@ -734,6 +763,27 @@ inline CUDA_CALLABLE void adj_div(vec_t<Length, Type> a, Type s, vec_t<Length, T
 #endif
 }
 
+template<unsigned Length, typename Type>
+inline CUDA_CALLABLE void adj_div(Type s, vec_t<Length, Type> a, Type& adj_s, vec_t<Length, Type>& adj_a, const vec_t<Length, Type>& adj_ret)
+{
+
+    adj_s -= dot(a , adj_ret)/ (s * s); // - a / s^2
+
+    for( unsigned i=0; i < Length; ++i )
+    {
+        adj_a[i] += s / adj_ret[i];
+    }
+
+#if FP_CHECK
+    if (!isfinite(a) || !isfinite(s) || !isfinite(adj_a) || !isfinite(adj_s) || !isfinite(adj_ret))
+    {
+        // \TODO: How shall we implement this error message?
+        // printf("adj_div((%f %f %f %f), %f, (%f %f %f %f), %f, (%f %f %f %f)\n", a.x, a.y, a.z, a.w, s, adj_a.x, adj_a.y, adj_a.z, adj_a.w, adj_s, adj_ret.x, adj_ret.y, adj_ret.z, adj_ret.w);
+        assert(0);
+    }
+#endif
+}
+
 template<unsigned Length, typename Type>
 inline CUDA_CALLABLE void adj_cw_div(vec_t<Length, Type> a, vec_t<Length, Type> b, vec_t<Length, Type>& ret, vec_t<Length, Type>& adj_a, vec_t<Length, Type>& adj_b, const vec_t<Length, Type>& adj_ret) {
   adj_a += cw_div(adj_ret, b);
diff --git a/warp/tests/test_mat.py b/warp/tests/test_mat.py
index 10d3b0dea..e74748fbc 100644
--- a/warp/tests/test_mat.py
+++ b/warp/tests/test_mat.py
@@ -159,6 +159,31 @@ def test_components(test, device, dtype):
     test.assertEqual(m[1, 2], 18)
 
 
+def test_py_arithmetic_ops(test, device, dtype):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def make_mat(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(tuple(wptype._type_(x).value for x in row) for row in args)
+
+        return args
+
+    mat_cls = wp.mat((3, 3), wptype)
+
+    m = mat_cls(((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
+    test.assertSequenceEqual(+m, make_mat((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
+    test.assertSequenceEqual(-m, make_mat((1, -2, -3), (-4, 5, -6), (-7, -8, 9)))
+    test.assertSequenceEqual(m + mat_cls((5, 5, 5) * 3), make_mat((4, 7, 8), (9, 0, 11), (12, 13, -4)))
+    test.assertSequenceEqual(m - mat_cls((5, 5, 5) * 3), make_mat((-6, -3, -2), (-1, -10, 1), (2, 3, -14)))
+
+    m = mat_cls(((2, 4, 6), (8, 10, 12), (14, 16, 18)))
+    test.assertSequenceEqual(m * wptype(2), make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
+    test.assertSequenceEqual(wptype(2) * m, make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
+    test.assertSequenceEqual(m / wptype(2), make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
+    test.assertSequenceEqual(wptype(5040) / m, make_mat((2520, 1260, 840), (630, 504, 420), (360, 315, 280)))
+
+
 def test_constants(test, device, dtype, register_kernels=False):
     wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
     mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
@@ -4412,6 +4437,9 @@ class TestMat(parent):
         )
 
     for dtype in np_float_types:
+        add_function_test(
+            TestMat, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
+        )
         add_function_test_register_kernel(
             TestMat, f"test_quat_constructor_{dtype.__name__}", test_quat_constructor, devices=devices, dtype=dtype
         )
diff --git a/warp/tests/test_quat.py b/warp/tests/test_quat.py
index ef415ebe6..b7828ece7 100644
--- a/warp/tests/test_quat.py
+++ b/warp/tests/test_quat.py
@@ -1932,6 +1932,31 @@ def test_constructor_default():
     wp.expect_eq(qeye[3], 1.0)
 
 
+def test_py_arithmetic_ops(test, device, dtype):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def make_quat(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(wptype._type_(x).value for x in args)
+
+        return args
+
+    quat_cls = wp.types.quaternion(wptype)
+
+    v = quat_cls(1, -2, 3, -4)
+    test.assertSequenceEqual(+v, make_quat(1, -2, 3, -4))
+    test.assertSequenceEqual(-v, make_quat(-1, 2, -3, 4))
+    test.assertSequenceEqual(v + quat_cls(5, 5, 5, 5), make_quat(6, 3, 8, 1))
+    test.assertSequenceEqual(v - quat_cls(5, 5, 5, 5), make_quat(-4, -7, -2, -9))
+
+    v = quat_cls(2, 4, 6, 8)
+    test.assertSequenceEqual(v * wptype(2), make_quat(4, 8, 12, 16))
+    test.assertSequenceEqual(wptype(2) * v, make_quat(4, 8, 12, 16))
+    test.assertSequenceEqual(v / wptype(2), make_quat(1, 2, 3, 4))
+    test.assertSequenceEqual(wptype(24) / v, make_quat(12, 6, 4, 3))
+
+
 def register(parent):
     devices = get_test_devices()
 
@@ -2020,6 +2045,9 @@ class TestQuat(parent):
         add_function_test_register_kernel(
             TestQuat, f"test_quat_to_matrix_{dtype.__name__}", test_quat_to_matrix, devices=devices, dtype=dtype
         )
+        add_function_test(
+            TestQuat, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
+        )
 
     return TestQuat
 
diff --git a/warp/tests/test_vec.py b/warp/tests/test_vec.py
index 362e0ead7..88f5d9d2d 100644
--- a/warp/tests/test_vec.py
+++ b/warp/tests/test_vec.py
@@ -176,6 +176,31 @@ def test_components(test, device, dtype):
     test.assertEqual(v[2], 15)
 
 
+def test_py_arithmetic_ops(test, device, dtype):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def make_vec(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(wptype._type_(x).value for x in args)
+
+        return args
+
+    vec_cls = wp.vec(3, wptype)
+
+    v = vec_cls(1, -2, 3)
+    test.assertSequenceEqual(+v, make_vec(1, -2, 3))
+    test.assertSequenceEqual(-v, make_vec(-1, 2, -3))
+    test.assertSequenceEqual(v + vec_cls(5, 5, 5), make_vec(6, 3, 8))
+    test.assertSequenceEqual(v - vec_cls(5, 5, 5), make_vec(-4, -7, -2))
+
+    v = vec_cls(2, 4, 6)
+    test.assertSequenceEqual(v * wptype(2), make_vec(4, 8, 12))
+    test.assertSequenceEqual(wptype(2) * v, make_vec(4, 8, 12))
+    test.assertSequenceEqual(v / wptype(2), make_vec(1, 2, 3))
+    test.assertSequenceEqual(wptype(24) / v, make_vec(12, 6, 4))
+
+
 def test_anon_type_instance(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
@@ -3205,6 +3230,9 @@ class TestVec(parent):
     for dtype in np_scalar_types:
         add_function_test(TestVec, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
         add_function_test(TestVec, f"test_components_{dtype.__name__}", test_components, devices=None, dtype=dtype)
+        add_function_test(
+            TestVec, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
+        )
         add_function_test_register_kernel(
             TestVec, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
         )
diff --git a/warp/types.py b/warp/types.py
index 3e056174b..148109422 100644
--- a/warp/types.py
+++ b/warp/types.py
@@ -177,15 +177,9 @@ def __setattr__(self, name, value):
         def __add__(self, y):
             return warp.add(self, y)
 
-        def __radd__(self, y):
-            return warp.add(self, y)
-
         def __sub__(self, y):
             return warp.sub(self, y)
 
-        def __rsub__(self, x):
-            return warp.sub(x, self)
-
         def __mul__(self, y):
             return warp.mul(self, y)
 
@@ -195,7 +189,7 @@ def __rmul__(self, x):
         def __truediv__(self, y):
             return warp.div(self, y)
 
-        def __rdiv__(self, x):
+        def __rtruediv__(self, x):
             return warp.div(x, self)
 
         def __pos__(self):
@@ -293,15 +287,9 @@ def __init__(self, *args):
         def __add__(self, y):
             return warp.add(self, y)
 
-        def __radd__(self, y):
-            return warp.add(self, y)
-
         def __sub__(self, y):
             return warp.sub(self, y)
 
-        def __rsub__(self, x):
-            return warp.sub(x, self)
-
         def __mul__(self, y):
             return warp.mul(self, y)
 
@@ -317,7 +305,7 @@ def __rmatmul__(self, x):
         def __truediv__(self, y):
             return warp.div(self, y)
 
-        def __rdiv__(self, x):
+        def __rtruediv__(self, x):
             return warp.div(x, self)
 
         def __pos__(self):

From 579683b236da2f29b7c88371f2c0f5514ffbed41 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Tue, 31 Oct 2023 14:31:19 +1300
Subject: [PATCH 34/50] Implement the vector @ matrix operator

---
 warp/builtins.py          |  24 ++++++
 warp/native/exports.h     |  12 +++
 warp/native/mat.h         |  18 ++++
 warp/tests/test_ctypes.py |  28 ++++---
 warp/tests/test_mat.py    | 171 ++++++++++++++++++++++++++++++++++++++
 5 files changed, 241 insertions(+), 12 deletions(-)

diff --git a/warp/builtins.py b/warp/builtins.py
index 0e72173a7..b1072779e 100644
--- a/warp/builtins.py
+++ b/warp/builtins.py
@@ -3106,6 +3106,23 @@ def mul_matvec_value_func(arg_types, kwds, _):
     return vector(length=arg_types[0]._shape_[0], dtype=arg_types[0]._wp_scalar_type_)
 
 
+def mul_vecmat_value_func(arg_types, kwds, _):
+    if arg_types is None:
+        return vector(length=Any, dtype=Scalar)
+
+    if arg_types[1]._wp_scalar_type_ != arg_types[0]._wp_scalar_type_:
+        raise RuntimeError(
+            f"Can't multiply vector and matrix with different types {arg_types[1]._wp_scalar_type_}, {arg_types[0]._wp_scalar_type_}"
+        )
+
+    if arg_types[1]._shape_[0] != arg_types[0]._length_:
+        raise RuntimeError(
+            f"Can't multiply vector with length {arg_types[0]._length_} and matrix of shape {arg_types[1]._shape_}"
+        )
+
+    return vector(length=arg_types[1]._shape_[1], dtype=arg_types[1]._wp_scalar_type_)
+
+
 def mul_matmat_value_func(arg_types, kwds, _):
     if arg_types is None:
         return matrix(length=Any, dtype=Scalar)
@@ -3180,6 +3197,13 @@ def mul_matmat_value_func(arg_types, kwds, _):
     doc="",
     group="Operators",
 )
+add_builtin(
+    "mul",
+    input_types={"x": vector(length=Any, dtype=Scalar), "y": matrix(shape=(Any, Any), dtype=Scalar)},
+    value_func=mul_vecmat_value_func,
+    doc="",
+    group="Operators",
+)
 add_builtin(
     "mul",
     input_types={"x": matrix(shape=(Any, Any), dtype=Scalar), "y": matrix(shape=(Any, Any), dtype=Scalar)},
diff --git a/warp/native/exports.h b/warp/native/exports.h
index 931680e4c..fa35feff6 100644
--- a/warp/native/exports.h
+++ b/warp/native/exports.h
@@ -1223,6 +1223,18 @@ WP_API void builtin_mul_mat22d_vec2d(mat22d x, vec2d y, vec2d* ret) { *ret = wp:
 WP_API void builtin_mul_mat33d_vec3d(mat33d x, vec3d y, vec3d* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_mat44d_vec4d(mat44d x, vec4d y, vec4d* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_spatial_matrixd_spatial_vectord(spatial_matrixd x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2h_mat22h(vec2h x, mat22h y, vec2h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3h_mat33h(vec3h x, mat33h y, vec3h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4h_mat44h(vec4h x, mat44h y, vec4h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectorh_spatial_matrixh(spatial_vectorh x, spatial_matrixh y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2f_mat22f(vec2f x, mat22f y, vec2f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3f_mat33f(vec3f x, mat33f y, vec3f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4f_mat44f(vec4f x, mat44f y, vec4f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectorf_spatial_matrixf(spatial_vectorf x, spatial_matrixf y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2d_mat22d(vec2d x, mat22d y, vec2d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3d_mat33d(vec3d x, mat33d y, vec3d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4d_mat44d(vec4d x, mat44d y, vec4d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectord_spatial_matrixd(spatial_vectord x, spatial_matrixd y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_mat22h_mat22h(mat22h x, mat22h y, mat22h* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_mat33h_mat33h(mat33h x, mat33h y, mat33h* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_mat44h_mat44h(mat44h x, mat44h y, mat44h* ret) { *ret = wp::mul(x, y); }
diff --git a/warp/native/mat.h b/warp/native/mat.h
index da3b9c9df..42cf586f6 100644
--- a/warp/native/mat.h
+++ b/warp/native/mat.h
@@ -500,6 +500,17 @@ inline CUDA_CALLABLE vec_t<Rows,Type> mul(const mat_t<Rows,Cols,Type>& a, const
     return r;
 }
 
+template<unsigned Rows, unsigned Cols, typename Type>
+inline CUDA_CALLABLE vec_t<Cols,Type> mul(const vec_t<Rows,Type>& b, const mat_t<Rows,Cols,Type>& a)
+{
+    vec_t<Cols,Type> r = a.get_row(0)*b[0];
+    for( unsigned i=1; i < Rows; ++i )
+    {
+        r += a.get_row(i)*b[i];
+    }
+    return r;
+}
+
 template<unsigned Rows, unsigned Cols, unsigned ColsOut, typename Type>
 inline CUDA_CALLABLE mat_t<Rows,ColsOut,Type> mul(const mat_t<Rows,Cols,Type>& a, const mat_t<Cols,ColsOut,Type>& b)
 {
@@ -1006,6 +1017,13 @@ inline CUDA_CALLABLE void adj_mul(const mat_t<Rows,Cols,Type>& a, const vec_t<Co
     adj_b += mul(transpose(a), adj_ret);
 }
 
+template<unsigned Rows, unsigned Cols, typename Type>
+inline CUDA_CALLABLE void adj_mul(const vec_t<Rows,Type>& b, const mat_t<Rows,Cols,Type>& a, vec_t<Rows,Type>& adj_b, mat_t<Rows,Cols,Type>& adj_a, const vec_t<Cols,Type>& adj_ret)
+{
+    adj_a += outer(b, adj_ret);
+    adj_b += mul(adj_ret, transpose(a));
+}
+
 template<unsigned Rows, unsigned Cols, unsigned ColsOut, typename Type>
 inline CUDA_CALLABLE void adj_mul(const mat_t<Rows,Cols,Type>& a, const mat_t<Cols,ColsOut,Type>& b, mat_t<Rows,Cols,Type>& adj_a, mat_t<Cols,ColsOut,Type>& adj_b, const mat_t<Rows,ColsOut,Type>& adj_ret)
 {
diff --git a/warp/tests/test_ctypes.py b/warp/tests/test_ctypes.py
index 63e6e92ca..9717f3496 100644
--- a/warp/tests/test_ctypes.py
+++ b/warp/tests/test_ctypes.py
@@ -24,11 +24,11 @@ def add_vec2(dest: wp.array(dtype=wp.vec2), c: wp.vec2):
 
 
 @wp.kernel
-def transform_vec2(dest: wp.array(dtype=wp.vec2), m: wp.mat22, v: wp.vec2):
+def transform_vec2(dest_right: wp.array(dtype=wp.vec2), dest_left: wp.array(dtype=wp.vec2), m: wp.mat22, v: wp.vec2):
     tid = wp.tid()
 
-    p = wp.mul(m, v)
-    dest[tid] = p
+    dest_right[tid] = wp.mul(m, v)
+    dest_left[tid] = wp.mul(v, m)
 
 
 @wp.kernel
@@ -38,11 +38,11 @@ def add_vec3(dest: wp.array(dtype=wp.vec3), c: wp.vec3):
 
 
 @wp.kernel
-def transform_vec3(dest: wp.array(dtype=wp.vec3), m: wp.mat33, v: wp.vec3):
+def transform_vec3(dest_right: wp.array(dtype=wp.vec3), dest_left: wp.array(dtype=wp.vec3), m: wp.mat33, v: wp.vec3):
     tid = wp.tid()
 
-    p = wp.mul(m, v)
-    dest[tid] = p
+    dest_right[tid] = wp.mul(m, v)
+    dest_left[tid] = wp.mul(v, m)
 
 
 @wp.kernel
@@ -63,12 +63,14 @@ def test_vec2_arg(test, device, n):
 
 
 def test_vec2_transform(test, device, n):
-    dest = wp.zeros(n=n, dtype=wp.vec2, device=device)
+    dest_right = wp.zeros(n=n, dtype=wp.vec2, device=device)
+    dest_left = wp.zeros(n=n, dtype=wp.vec2, device=device)
     c = np.array((1.0, 2.0))
     m = np.array(((3.0, -1.0), (2.5, 4.0)))
 
-    wp.launch(transform_vec2, dim=n, inputs=[dest, m, c], device=device)
-    test.assertTrue(np.array_equal(dest.numpy(), np.tile(m @ c, (n, 1))))
+    wp.launch(transform_vec2, dim=n, inputs=[dest_right, dest_left, m, c], device=device)
+    test.assertTrue(np.array_equal(dest_right.numpy(), np.tile(m @ c, (n, 1))))
+    test.assertTrue(np.array_equal(dest_left.numpy(), np.tile(c @ m, (n, 1))))
 
 
 def test_vec3_arg(test, device, n):
@@ -80,12 +82,14 @@ def test_vec3_arg(test, device, n):
 
 
 def test_vec3_transform(test, device, n):
-    dest = wp.zeros(n=n, dtype=wp.vec3, device=device)
+    dest_right = wp.zeros(n=n, dtype=wp.vec3, device=device)
+    dest_left = wp.zeros(n=n, dtype=wp.vec3, device=device)
     c = np.array((1.0, 2.0, 3.0))
     m = np.array(((1.0, 2.0, 3.0), (4.0, 5.0, 6.0), (7.0, 8.0, 9.0)))
 
-    wp.launch(transform_vec3, dim=n, inputs=[dest, m, c], device=device)
-    test.assertTrue(np.array_equal(dest.numpy(), np.tile(m @ c, (n, 1))))
+    wp.launch(transform_vec3, dim=n, inputs=[dest_right, dest_left, m, c], device=device)
+    test.assertTrue(np.array_equal(dest_right.numpy(), np.tile(m @ c, (n, 1))))
+    test.assertTrue(np.array_equal(dest_left.numpy(), np.tile(c @ m, (n, 1))))
 
 
 def test_transform_multiply(test, device, n):
diff --git a/warp/tests/test_mat.py b/warp/tests/test_mat.py
index e74748fbc..321bc1cf9 100644
--- a/warp/tests/test_mat.py
+++ b/warp/tests/test_mat.py
@@ -169,19 +169,35 @@ def make_mat(*args):
 
         return args
 
+    def make_vec(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(wptype._type_(x).value for x in args)
+
+        return args
+
     mat_cls = wp.mat((3, 3), wptype)
+    vec_cls = wp.vec(3, wptype)
 
     m = mat_cls(((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
     test.assertSequenceEqual(+m, make_mat((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
     test.assertSequenceEqual(-m, make_mat((1, -2, -3), (-4, 5, -6), (-7, -8, 9)))
     test.assertSequenceEqual(m + mat_cls((5, 5, 5) * 3), make_mat((4, 7, 8), (9, 0, 11), (12, 13, -4)))
     test.assertSequenceEqual(m - mat_cls((5, 5, 5) * 3), make_mat((-6, -3, -2), (-1, -10, 1), (2, 3, -14)))
+    test.assertSequenceEqual(m * vec_cls(5, 5, 5), make_vec(20, 25, 30))
+    test.assertSequenceEqual(m @ vec_cls(5, 5, 5), make_vec(20, 25, 30))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) * m, make_vec(50, 25, 0))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) @ m, make_vec(50, 25, 0))
 
     m = mat_cls(((2, 4, 6), (8, 10, 12), (14, 16, 18)))
     test.assertSequenceEqual(m * wptype(2), make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
     test.assertSequenceEqual(wptype(2) * m, make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
     test.assertSequenceEqual(m / wptype(2), make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
     test.assertSequenceEqual(wptype(5040) / m, make_mat((2520, 1260, 840), (630, 504, 420), (360, 315, 280)))
+    test.assertSequenceEqual(m * vec_cls(5, 5, 5), make_vec(60, 150, 240))
+    test.assertSequenceEqual(m @ vec_cls(5, 5, 5), make_vec(60, 150, 240))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) * m, make_vec(120, 150, 180))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) @ m, make_vec(120, 150, 180))
 
 
 def test_constants(test, device, dtype, register_kernels=False):
@@ -1600,6 +1616,153 @@ def check_mat_vec_mul(
                 idx = idx + 1
 
 
+def test_vecmat_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 2.0e-2,
+        np.float32: 5.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat23 = wp.types.matrix(shape=(2, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_vec_mat_mul(
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v32: wp.array(dtype=vec2),
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        m23: wp.array(dtype=mat23),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        v2result = v2[0] * m2[0]
+        v3result = v3[0] * m3[0]
+        v4result = v4[0] * m4[0]
+        v5result = v5[0] * m5[0]
+        v32result = v32[0] * m23[0]
+        v2result_2 = v2[0] @ m2[0]
+        v3result_2 = v3[0] @ m3[0]
+        v4result_2 = v4[0] @ m4[0]
+        v5result_2 = v5[0] @ m5[0]
+        v32result_2 = v32[0] @ m23[0]
+
+        idx = 0
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        for i in range(2):
+            outcomponents[idx] = wptype(2) * v2result[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v3result[i]
+            idx = idx + 1
+
+        for i in range(4):
+            outcomponents[idx] = wptype(2) * v4result[i]
+            idx = idx + 1
+
+        for i in range(5):
+            outcomponents[idx] = wptype(2) * v5result[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v32result[i]
+            idx = idx + 1
+
+        for i in range(2):
+            outcomponents[idx] = wptype(2) * v2result_2[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v3result_2[i]
+            idx = idx + 1
+
+        for i in range(4):
+            outcomponents[idx] = wptype(2) * v4result_2[i]
+            idx = idx + 1
+
+        for i in range(5):
+            outcomponents[idx] = wptype(2) * v5result_2[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v32result_2[i]
+            idx = idx + 1
+
+    kernel = getkernel(check_vec_mat_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
+    v32 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    m23 = wp.array(randvals(rng, [1, 2, 3], dtype), dtype=mat23, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * (2 + 3 + 4 + 5 + 3), dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m23], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:2], 2 * np.matmul(v2.numpy()[0], m2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[2:5], 2 * np.matmul(v3.numpy()[0], m3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[5:9], 2 * np.matmul(v4.numpy()[0], m4.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[9:14], 2 * np.matmul(v5.numpy()[0], m5.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[14:17], 2 * np.matmul(v32.numpy()[0], m23.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[17:19], 2 * np.matmul(v2.numpy()[0], m2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[19:22], 2 * np.matmul(v3.numpy()[0], m3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[22:26], 2 * np.matmul(v4.numpy()[0], m4.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[26:31], 2 * np.matmul(v5.numpy()[0], m5.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[31:34], 2 * np.matmul(v32.numpy()[0], m23.numpy()[0]), tol=5 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, inmat, invec in [(2, m2, v2), (3, m3, v3), (4, m4, v4), (5, m5, v5), (3, m23, v32)]:
+            for i in range(dim):
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(
+                        kernel,
+                        dim=1,
+                        inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m23],
+                        outputs=[outcomponents],
+                        device=device,
+                    )
+                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+                tape.backward(loss=out)
+
+                assert_np_equal(tape.gradients[invec].numpy()[0], 2 * inmat.numpy()[0, :, i], tol=2 * tol)
+                expectedresult = np.zeros(inmat.dtype._shape_, dtype=dtype)
+                expectedresult[:, i] = 2 * invec.numpy()[0]
+                assert_np_equal(tape.gradients[inmat].numpy()[0], expectedresult, tol=2 * tol)
+
+                tape.zero()
+
+                idx = idx + 1
+
+
 def test_matmat_multiplication(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
@@ -4165,6 +4328,7 @@ def test_constructors_explicit_precision():
 mat32d = wp.mat(shape=(3, 2), dtype=wp.float64)
 
 
+
 @wp.kernel
 def test_matrix_constructor_value_func():
     a = wp.mat22()
@@ -4389,6 +4553,13 @@ class TestMat(parent):
             devices=devices,
             dtype=dtype,
         )
+        add_function_test_register_kernel(
+            TestMat,
+            f"test_vecmat_multiplication_{dtype.__name__}",
+            test_vecmat_multiplication,
+            devices=devices,
+            dtype=dtype,
+        )
         add_function_test_register_kernel(
             TestMat,
             f"test_matmat_multiplication_{dtype.__name__}",

From d7ec437bb5833b15a0b1f55fab79f49c36d6de60 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Wed, 1 Nov 2023 15:31:14 +1300
Subject: [PATCH 35/50] Fix `float16` types being interpreted as `ushort`

---
 warp/context.py | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/warp/context.py b/warp/context.py
index c1ccc20ee..514302485 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -244,7 +244,10 @@ class ValueArg(ctypes.Structure):
                         else:
                             try:
                                 # try to pack as a scalar type
-                                params.append(arg_type._type_(a))
+                                if arg_type == warp.types.float16:
+                                    params.append(arg_type._type_(warp.types.float_to_half_bits(a)))
+                                else:
+                                    params.append(arg_type._type_(a))
                             except Exception:
                                 raise RuntimeError(
                                     f"Error calling function {f.key}, unable to pack function parameter type {type(a)} for param {arg_name}, expected {arg_type}"
@@ -264,10 +267,11 @@ def type_ctype(dtype):
                             # scalar type
                             return dtype._type_
 
-                    value_type = type_ctype(f.value_func(None, None, None))
+                    value_type = f.value_func(None, None, None)
+                    value_ctype = type_ctype(value_type)
 
                     # construct return value (passed by address)
-                    ret = value_type()
+                    ret = value_ctype()
                     ret_addr = ctypes.c_void_p(ctypes.addressof(ret))
 
                     params.append(ret_addr)
@@ -275,9 +279,11 @@ def type_ctype(dtype):
                     c_func = getattr(warp.context.runtime.core, f.mangled_name)
                     c_func(*params)
 
-                    if issubclass(value_type, ctypes.Array) or issubclass(value_type, ctypes.Structure):
+                    if issubclass(value_ctype, ctypes.Array) or issubclass(value_ctype, ctypes.Structure):
                         # return vector types as ctypes
                         return ret
+                    elif value_type == warp.types.float16:
+                        return warp.types.half_bits_to_float(ret.value)
 
                     # return scalar types as int/float
                     return ret.value

From 9c8bc86fac9a53bc2c99e6a720d1213528b3a687 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Wed, 6 Dec 2023 16:34:55 -0500
Subject: [PATCH 36/50] Fix support for function calls in range()

Support for the `type()` operator in kernels introduced a regression for
`range()` expressions that contain function calls. In the case of e.g.
`wp.min()` the full function name is not an `ast.Name` but rather an
`ast.Attribute`, and in the case of just `min()` we should not
immediately treat it as an invalid operator but continue parsing it as a
potential builtin function call.
---
 warp/codegen.py             | 15 ++++-----------
 warp/tests/test_codegen.py  | 34 ++++++++++++++++++++++++++++++++++
 warp/tests/test_generics.py |  2 +-
 3 files changed, 39 insertions(+), 12 deletions(-)

diff --git a/warp/codegen.py b/warp/codegen.py
index b04f74e46..4dd04fec7 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -1463,7 +1463,7 @@ def eval_num(adj, a):
 
         # try and resolve the expression to an object
         # e.g.: wp.constant in the globals scope
-        obj, path = adj.resolve_static_expression(a)
+        obj, _ = adj.resolve_static_expression(a)
 
         if isinstance(obj, Var) and obj.constant is not None:
             obj = obj.constant
@@ -1630,7 +1630,7 @@ def emit_Call(adj, node):
 
         if not isinstance(func, warp.context.Function):
             if len(path) == 0:
-                raise WarpCodegenError(f"Unrecognized syntax for function call, path not valid: '{node.func}'")
+                raise WarpCodegenError(f"Unknown function or operator: '{node.func.func.id}'")
 
             attr = path[-1]
             caller = func
@@ -2027,16 +2027,11 @@ def resolve_static_expression(adj, root_node, eval_types=True):
             attributes.append(node.attr)
             node = node.value
 
-        if eval_types and isinstance(node, ast.Call):
+        if eval_types and isinstance(node, ast.Call) and isinstance(node.func, ast.Name):
             # support for operators returning modules
             # i.e. operator_name(*operator_args).x.y.z
             operator_args = node.args
-            operator_name = getattr(node.func, "id", None)
-
-            if operator_name is None:
-                raise WarpCodegenError(
-                    f"Invalid operator call syntax, expected a plain name, got {ast.dump(node.func, annotate_fields=False)}"
-                )
+            operator_name = node.func.id
 
             if operator_name == "type":
                 if len(operator_args) != 1:
@@ -2061,8 +2056,6 @@ def resolve_static_expression(adj, root_node, eval_types=True):
                 else:
                     raise WarpCodegenError(f"Cannot deduce the type of {var}")
 
-            raise WarpCodegenError(f"Unknown operator '{operator_name}'")
-
         # reverse list since ast presents it backward order
         path = [*reversed(attributes)]
         if isinstance(node, ast.Name):
diff --git a/warp/tests/test_codegen.py b/warp/tests/test_codegen.py
index 5c0d0deef..a078e5e76 100644
--- a/warp/tests/test_codegen.py
+++ b/warp/tests/test_codegen.py
@@ -317,6 +317,33 @@ def test_range_constant_dynamic_nested(m: int):
     wp.expect_eq(s, N * m * N)
 
 
+@wp.kernel
+def test_range_expression():
+    idx = 1
+    batch_size = 100
+
+    a = wp.float(0.0)
+    c = wp.float(1.0)
+
+    # constant expression with a function
+    for i in range(4 * idx, wp.min(4 * idx + 4, batch_size)):
+        a += c
+
+    for i in range(4 * idx, min(4 * idx + 4, batch_size)):
+        a += c
+
+    tid = wp.tid()
+
+    # dynamic expression with a function
+    for i in range(4 * idx, wp.min(4 * idx, tid + 1000)):
+        a += c
+
+    for i in range(4 * idx, min(4 * idx, tid + 1000)):
+        a += c
+
+    wp.expect_eq(a, 8.0)
+
+
 def test_unresolved_func(test, device):
     # kernel with unresolved function must be in a separate module, otherwise the current module would fail to load
     from warp.tests.test_unresolved_func import unresolved_func_kernel
@@ -450,6 +477,13 @@ class TestCodeGen(parent):
         inputs=[4],
         devices=devices,
     )
+    add_kernel_test(
+        TestCodeGen,
+        name="test_range_expression",
+        kernel=test_range_expression,
+        dim=1,
+        devices=devices,
+    )
 
     add_kernel_test(TestCodeGen, name="test_while_zero", kernel=test_while, dim=1, inputs=[0], devices=devices)
     add_kernel_test(TestCodeGen, name="test_while_positive", kernel=test_while, dim=1, inputs=[16], devices=devices)
diff --git a/warp/tests/test_generics.py b/warp/tests/test_generics.py
index 33d58029e..dcddea20c 100644
--- a/warp/tests/test_generics.py
+++ b/warp/tests/test_generics.py
@@ -463,7 +463,7 @@ def kernel():
         i = wp.tid()
         _ = typez(i)(0)
 
-    with test.assertRaisesRegex(RuntimeError, r"Unknown operator 'typez'$"):
+    with test.assertRaisesRegex(RuntimeError, r"Unknown function or operator: 'typez'$"):
         wp.launch(
             kernel,
             dim=1,

From d9d9670e6692b94c0790492178cc58378deac969 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Wed, 8 Nov 2023 11:25:31 +1300
Subject: [PATCH 37/50] Improve native built-ins overload resolution

Python types such as `int` and `float` are strictly resolving
to `wp.int32` and `wp.float32`, which might require users to be
more explicit in some code paths that were working until now.

For example, calls like `wp.mul(np.array((1.0, 2.0, 3.0)), 1.23)`
won't work anymore since the default float type of NumPy is
`np.float64`, which isn't compatibles with `1.23`'s `wp.float32`.

As a result, users are now required to either call something like
`wp.mul(np.array((1.0, 2.0, 3.0), dtype=np.float32), 1.23)` or
`wp.mul(np.array((1.0, 2.0, 3.0)), wp.float64(1.23))`.

Additionally, array types such as lists, tuples, NumPy arrays,
and others, are now matched by correctly preserving the precision of
their element type. However, using Warp types such as `wp.vec`,
`wp.mat`, `wp.quat`, or `wp.transform`, is preferred, so support
for non-Warp array types is now marked as deprecated and
will be dropped in the future.
---
 warp/context.py                        |  323 +++---
 warp/tests/test_all.py                 |    2 +
 warp/tests/test_builtins_resolution.py | 1374 ++++++++++++++++++++++++
 warp/tests/test_types.py               |  221 +++-
 warp/types.py                          |   12 +
 5 files changed, 1784 insertions(+), 148 deletions(-)
 create mode 100644 warp/tests/test_builtins_resolution.py

diff --git a/warp/context.py b/warp/context.py
index 514302485..1b788d79a 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -178,118 +178,16 @@ def __call__(self, *args, **kwargs):
         # from within a kernel (experimental).
 
         if self.is_builtin() and self.mangled_name:
-            for f in self.overloads:
-                if f.generic:
+            # For each of this function's existing overloads, we attempt to pack
+            # the given arguments into the C types expected by the corresponding
+            # parameters, and we rinse and repeat until we get a match.
+            for overload in self.overloads:
+                if overload.generic:
                     continue
 
-                # try and find builtin in the warp.dll
-                if not hasattr(warp.context.runtime.core, f.mangled_name):
-                    raise RuntimeError(
-                        f"Couldn't find function {self.key} with mangled name {f.mangled_name} in the Warp native library"
-                    )
-
-                try:
-                    # try and pack args into what the function expects
-                    params = []
-                    for i, (arg_name, arg_type) in enumerate(f.input_types.items()):
-                        a = args[i]
-
-                        # try to convert to a value type (vec3, mat33, etc)
-                        if issubclass(arg_type, ctypes.Array):
-                            # wrap the arg_type (which is an ctypes.Array) in a structure
-                            # to ensure parameter is passed to the .dll by value rather than reference
-                            class ValueArg(ctypes.Structure):
-                                _fields_ = [("value", arg_type)]
-
-                            x = ValueArg()
-
-                            # force conversion to ndarray first (handles tuple / list, Gf.Vec3 case)
-                            if isinstance(a, ctypes.Array) is False:
-                                # assume you want the float32 version of the function so it doesn't just
-                                # grab an override for a random data type:
-                                if arg_type._type_ != ctypes.c_float:
-                                    raise RuntimeError(
-                                        f"Error calling function '{f.key}', parameter for argument '{arg_name}' does not have c_float type."
-                                    )
-
-                                a = np.array(a)
-
-                                # flatten to 1D array
-                                v = a.flatten()
-                                if len(v) != arg_type._length_:
-                                    raise RuntimeError(
-                                        f"Error calling function '{f.key}', parameter for argument '{arg_name}' has length {len(v)}, but expected {arg_type._length_}. Could not convert parameter to {arg_type}."
-                                    )
-
-                                for i in range(arg_type._length_):
-                                    x.value[i] = v[i]
-
-                            else:
-                                # already a built-in type, check it matches
-                                if not warp.types.types_equal(type(a), arg_type):
-                                    raise RuntimeError(
-                                        f"Error calling function '{f.key}', parameter for argument '{arg_name}' has type '{type(a)}' but expected '{arg_type}'"
-                                    )
-
-                                if isinstance(a, arg_type):
-                                    x.value = a
-                                else:
-                                    # Cast the value to its argument type to make sure that it can be assigned to the field of the `ValueArg` struct.
-                                    # This could error otherwise when, for example, the field type is set to `vec3i` while the value is of type
-                                    # `vector(length=3, dtype=int)`, even though both types are semantically identical.
-                                    x.value = arg_type(a)
-
-                            params.append(x)
-
-                        else:
-                            try:
-                                # try to pack as a scalar type
-                                if arg_type == warp.types.float16:
-                                    params.append(arg_type._type_(warp.types.float_to_half_bits(a)))
-                                else:
-                                    params.append(arg_type._type_(a))
-                            except Exception:
-                                raise RuntimeError(
-                                    f"Error calling function {f.key}, unable to pack function parameter type {type(a)} for param {arg_name}, expected {arg_type}"
-                                )
-
-                    # returns the corresponding ctype for a scalar or vector warp type
-                    def type_ctype(dtype):
-                        if dtype == float:
-                            return ctypes.c_float
-                        elif dtype == int:
-                            return ctypes.c_int32
-                        elif issubclass(dtype, ctypes.Array):
-                            return dtype
-                        elif issubclass(dtype, ctypes.Structure):
-                            return dtype
-                        else:
-                            # scalar type
-                            return dtype._type_
-
-                    value_type = f.value_func(None, None, None)
-                    value_ctype = type_ctype(value_type)
-
-                    # construct return value (passed by address)
-                    ret = value_ctype()
-                    ret_addr = ctypes.c_void_p(ctypes.addressof(ret))
-
-                    params.append(ret_addr)
-
-                    c_func = getattr(warp.context.runtime.core, f.mangled_name)
-                    c_func(*params)
-
-                    if issubclass(value_ctype, ctypes.Array) or issubclass(value_ctype, ctypes.Structure):
-                        # return vector types as ctypes
-                        return ret
-                    elif value_type == warp.types.float16:
-                        return warp.types.half_bits_to_float(ret.value)
-
-                    # return scalar types as int/float
-                    return ret.value
-                except Exception:
-                    # couldn't pack values to match this overload
-                    continue
+                success, return_value = call_builtin(overload, *args)
+                if success:
+                    return return_value
 
             # overload resolution or call failed
             raise RuntimeError(
@@ -297,7 +195,7 @@ def type_ctype(dtype):
                 f"the arguments '{', '.join(type(x).__name__ for x in args)}'"
             )
 
-        elif hasattr(self, "user_overloads") and len(self.user_overloads):
+        if hasattr(self, "user_overloads") and len(self.user_overloads):
             # user-defined function with overloads
 
             if len(kwargs):
@@ -306,28 +204,26 @@ def type_ctype(dtype):
                 )
 
             # try and find a matching overload
-            for f in self.user_overloads.values():
-                if len(f.input_types) != len(args):
+            for overload in self.user_overloads.values():
+                if len(overload.input_types) != len(args):
                     continue
-                template_types = list(f.input_types.values())
-                arg_names = list(f.input_types.keys())
+                template_types = list(overload.input_types.values())
+                arg_names = list(overload.input_types.keys())
                 try:
                     # attempt to unify argument types with function template types
                     warp.types.infer_argument_types(args, template_types, arg_names)
-                    return f.func(*args)
+                    return overload.func(*args)
                 except Exception:
                     continue
 
             raise RuntimeError(f"Error calling function '{self.key}', no overload found for arguments {args}")
 
-        else:
-            # user-defined function with no overloads
+        # user-defined function with no overloads
+        if self.func is None:
+            raise RuntimeError(f"Error calling function '{self.key}', function is undefined")
 
-            if self.func is None:
-                raise RuntimeError(f"Error calling function '{self.key}', function is undefined")
-
-            # this function has no overloads, call it like a plain Python function
-            return self.func(*args, **kwargs)
+        # this function has no overloads, call it like a plain Python function
+        return self.func(*args, **kwargs)
 
     def is_builtin(self):
         return self.func is None
@@ -444,6 +340,187 @@ def __repr__(self):
         return f"<Function {self.key}({inputs_str})>"
 
 
+def call_builtin(func: Function, *params) -> Tuple[bool, Any]:
+    uses_non_warp_array_type = False
+
+    # Retrieve the built-in function from Warp's dll.
+    c_func = getattr(warp.context.runtime.core, func.mangled_name)
+
+    # Try gathering the parameters that the function expects and pack them
+    # into their corresponding C types.
+    c_params = []
+    for i, (_, arg_type) in enumerate(func.input_types.items()):
+        param = params[i]
+
+        try:
+            iter(param)
+        except TypeError:
+            is_array = False
+        else:
+            is_array = True
+
+        if is_array:
+            if not issubclass(arg_type, ctypes.Array):
+                return (False, None)
+
+            # The argument expects a built-in Warp type like a vector or a matrix.
+            # We need to wrap the given argument into a ctypes' structure
+            # to ensure that it's passed by value to the dll rather than by reference.
+            class Param(ctypes.Structure):
+                _fields_ = [("value", arg_type)]
+
+            c_param = Param()
+
+            if isinstance(param, ctypes.Array):
+                # The given parameter is also a built-in Warp type, so we only need
+                # to make sure that it matches with the argument.
+                if not warp.types.types_equal(type(param), arg_type):
+                    return (False, None)
+
+                if isinstance(param, arg_type):
+                    c_param.value = param
+                else:
+                    # Cast the value to its argument type to make sure that it
+                    # can be assigned to the field of the `Param` struct.
+                    # This could error otherwise when, for example, the field type
+                    # is set to `vec3i` while the value is of type `vector(length=3, dtype=int)`,
+                    # even though both types are semantically identical.
+                    c_param.value = arg_type(param)
+            else:
+                # Flatten the parameter values into a flat 1-D array.
+                arr = []
+                ndim = 1
+                stack = [(0, param)]
+                while stack:
+                    depth, elem = stack.pop(0)
+                    try:
+                        # If `elem` is a sequence, then it should be possible
+                        # to add its elements to the stack for later processing.
+                        stack.extend((depth + 1, x) for x in elem)
+                    except TypeError:
+                        # Since `elem` doesn't seem to be a sequence,
+                        # we must have a leaf value that we need to add to our
+                        # resulting array.
+                        arr.append(elem)
+                        ndim = max(depth, ndim)
+
+                assert ndim > 0
+
+                # Ensure that if the given parameter value is, say, a 2-D array,
+                # then we try to resolve it against a matrix argument rather than
+                # a vector.
+                if ndim > len(arg_type._shape_):
+                    return (False, None)
+
+                elem_count = len(arr)
+                if elem_count != arg_type._length_:
+                    return (False, None)
+
+                # Retrieve the element type of the sequence while ensuring
+                # that it's homogeneous.
+                elem_type = type(arr[0])
+                for i in range(1, elem_count):
+                    if type(arr[i]) is not elem_type:
+                        raise ValueError("All array elements must share the same type.")
+
+                expected_elem_type = arg_type._wp_scalar_type_
+                if not (
+                    elem_type is expected_elem_type
+                    or (elem_type is float and expected_elem_type is warp.types.float32)
+                    or (elem_type is int and expected_elem_type is warp.types.int32)
+                    or (
+                        issubclass(elem_type, np.number)
+                        and warp.types.np_dtype_to_warp_type[np.dtype(elem_type)] is expected_elem_type
+                    )
+                ):
+                    # The parameter value has a type not matching the type defined
+                    # for the corresponding argument.
+                    return (False, None)
+
+                if elem_type in warp.types.int_types:
+                    # Pass the value through the expected integer type
+                    # in order to evaluate any integer wrapping.
+                    # For example `uint8(-1)` should result in the value `-255`.
+                    arr = tuple(elem_type._type_(x.value).value for x in arr)
+                elif elem_type in warp.types.float_types:
+                    # Extract the floating-point values.
+                    arr = tuple(x.value for x in arr)
+
+                if warp.types.type_is_matrix(arg_type):
+                    rows, cols = arg_type._shape_
+                    for i in range(rows):
+                        idx_start = i * cols
+                        idx_end = idx_start + cols
+                        c_param.value[i] = arr[idx_start:idx_end]
+                else:
+                    c_param.value[:] = arr
+
+                uses_non_warp_array_type = True
+
+            c_params.append(c_param)
+        else:
+            if issubclass(arg_type, ctypes.Array):
+                return (False, None)
+
+            if not (
+                isinstance(param, arg_type)
+                or (type(param) is float and arg_type is warp.types.float32)
+                or (type(param) is int and arg_type is warp.types.int32)
+                or warp.types.np_dtype_to_warp_type.get(getattr(param, "dtype", None)) is arg_type
+            ):
+                return (False, None)
+
+            if type(param) in warp.types.scalar_types:
+                param = param.value
+
+            # try to pack as a scalar type
+            if arg_type == warp.types.float16:
+                c_params.append(arg_type._type_(warp.types.float_to_half_bits(param)))
+            else:
+                c_params.append(arg_type._type_(param))
+
+    # returns the corresponding ctype for a scalar or vector warp type
+    value_type = func.value_func(None, None, None)
+    if value_type == float:
+        value_ctype = ctypes.c_float
+    elif value_type == int:
+        value_ctype = ctypes.c_int32
+    elif issubclass(value_type, (ctypes.Array, ctypes.Structure)):
+        value_ctype = value_type
+    else:
+        # scalar type
+        value_ctype = value_type._type_
+
+    # construct return value (passed by address)
+    ret = value_ctype()
+    ret_addr = ctypes.c_void_p(ctypes.addressof(ret))
+    c_params.append(ret_addr)
+
+    # Call the built-in function from Warp's dll.
+    c_func(*c_params)
+
+    # TODO: uncomment when we have a way to print warning messages only once.
+    # if uses_non_warp_array_type:
+    #     warp.utils.warn(
+    #         "Support for built-in functions called with non-Warp array types, "
+    #         "such as lists, tuples, NumPy arrays, and others, will be dropped "
+    #         "in the future. Use a Warp type such as `wp.vec`, `wp.mat`, "
+    #         "`wp.quat`, or `wp.transform`.",
+    #         DeprecationWarning,
+    #         stacklevel=3
+    #     )
+
+    if issubclass(value_ctype, ctypes.Array) or issubclass(value_ctype, ctypes.Structure):
+        # return vector types as ctypes
+        return (True, ret)
+
+    if value_type == warp.types.float16:
+        return (True, warp.types.half_bits_to_float(ret.value))
+
+    # return scalar types as int/float
+    return (True, ret.value)
+
+
 class KernelHooks:
     def __init__(self, forward, backward):
         self.forward = forward
diff --git a/warp/tests/test_all.py b/warp/tests/test_all.py
index bbd014dff..812aef902 100644
--- a/warp/tests/test_all.py
+++ b/warp/tests/test_all.py
@@ -14,6 +14,7 @@
 import warp.tests.test_array
 import warp.tests.test_atomic
 import warp.tests.test_bool
+import warp.tests.test_builtins_resolution
 import warp.tests.test_bvh
 import warp.tests.test_closest_point_edge_edge
 import warp.tests.test_codegen
@@ -85,6 +86,7 @@ def register_tests(parent):
     tests.append(warp.tests.test_mesh_query_aabb.register(parent))
     tests.append(warp.tests.test_mesh_query_point.register(parent))
     tests.append(warp.tests.test_mesh_query_ray.register(parent))
+    tests.append(warp.tests.test_builtins_resolution.register(parent))
     tests.append(warp.tests.test_bvh.register(parent))
     tests.append(warp.tests.test_conditional.register(parent))
     tests.append(warp.tests.test_operators.register(parent))
diff --git a/warp/tests/test_builtins_resolution.py b/warp/tests/test_builtins_resolution.py
new file mode 100644
index 000000000..bfeb6fa2c
--- /dev/null
+++ b/warp/tests/test_builtins_resolution.py
@@ -0,0 +1,1374 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import contextlib
+import unittest
+
+import numpy as np
+
+from warp.tests.test_base import *
+
+
+wp.init()
+
+
+def nps(dtype, value):
+    """Creates a NumPy scalar value based on the given data type."""
+    # Workaround to avoid deprecation warning messages for integer overflows.
+    return np.array((value,)).astype(dtype)[0]
+
+
+def npv(dtype, values):
+    """Creates a vector of NumPy scalar values based on the given data type."""
+    return tuple(nps(dtype, x) for x in values)
+
+
+def npm(dtype, dim, values):
+    """Creates a matrix of NumPy scalar values based on the given data type."""
+    return tuple(npv(dtype, values[i * dim : (i + 1) * dim]) for i in range(dim))
+
+
+def wpv(dtype, values):
+    """Creates a vector of Warp scalar values based on the given data type."""
+    return tuple(dtype(x) for x in values)
+
+
+def wpm(dtype, dim, values):
+    """Creates a matrix of Warp scalar values based on the given data type."""
+    return tuple(wpv(dtype, values[i * dim : (i + 1) * dim]) for i in range(dim))
+
+
+def test_int_arg_overflow(test, device):
+    value = -1234567890123456789
+
+    test.assertEqual(wp.invert(wp.int8(value)), 20)
+    test.assertEqual(wp.invert(wp.int16(value)), -32492)
+    test.assertEqual(wp.invert(wp.int32(value)), 2112454932)
+    test.assertEqual(wp.invert(wp.int64(value)), 1234567890123456788)
+
+    test.assertEqual(wp.invert(wp.uint8(value)), 20)
+    test.assertEqual(wp.invert(wp.uint16(value)), 33044)
+    test.assertEqual(wp.invert(wp.uint32(value)), 2112454932)
+    test.assertEqual(wp.invert(wp.uint64(value)), 1234567890123456788)
+
+    test.assertEqual(wp.invert(value), wp.invert(wp.int32(value)))
+
+
+def test_float_arg_precision(test, device):
+    value = 1.23
+    expected = 0.94248880193169748409
+
+    result = wp.sin(wp.float64(value))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.sin(wp.float32(value))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.sin(wp.float16(value))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    test.assertEqual(wp.sin(value), wp.sin(wp.float32(value)))
+
+
+def test_int_int_args_overflow(test, device):
+    value = -1234567890
+
+    test.assertEqual(wp.mul(wp.int8(value), wp.int8(value)), 68)
+    test.assertEqual(wp.mul(wp.int16(value), wp.int16(value)), -3004)
+    test.assertEqual(wp.mul(wp.int32(value), wp.int32(value)), 304084036)
+    test.assertEqual(wp.mul(wp.int64(value), wp.int64(value)), 1524157875019052100)
+
+    test.assertEqual(wp.mul(wp.uint8(value), wp.uint8(value)), 68)
+    test.assertEqual(wp.mul(wp.uint16(value), wp.uint16(value)), 62532)
+    test.assertEqual(wp.mul(wp.uint32(value), wp.uint32(value)), 304084036)
+    test.assertEqual(wp.mul(wp.uint64(value), wp.uint64(value)), 1524157875019052100)
+
+    test.assertEqual(wp.mul(value, value), wp.mul(wp.int32(value), wp.int32(value)))
+
+
+def test_mat22_arg_precision(test, device):
+    values = (1.23, 2.34, 3.45, 4.56)
+    values_2d = (values[0:2], values[2:4])
+    expected = 5.78999999999999914735
+
+    result = wp.trace(wp.mat22d(*values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.trace(wp.mat22f(*values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.trace(wp.mat22h(*values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.trace(values), wp.trace(wp.mat22f(*values)))
+        test.assertEqual(wp.trace(values_2d), wp.trace(wp.mat22f(*values)))
+
+
+def test_mat33_arg_precision(test, device):
+    values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
+    values_2d = (values[0:3], values[3:6], values[6:9])
+    expected = 15.91000000000000014211
+
+    result = wp.trace(wp.mat33d(*values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.trace(wp.mat33f(*values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.trace(wp.mat33h(*values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.trace(values), wp.trace(wp.mat33f(*values)))
+        test.assertEqual(wp.trace(values_2d), wp.trace(wp.mat33f(*values)))
+
+
+def test_mat44_arg_precision(test, device):
+    values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01, 10.12, 11.23, 12.34, 13.45, 14.56, 15.67, 16.78)
+    values_2d = (values[0:4], values[4:8], values[8:12], values[12:16])
+    expected = 36.02000000000000312639
+
+    result = wp.trace(wp.mat44d(*values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.trace(wp.mat44f(*values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.trace(wp.mat44h(*values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.trace(values), wp.trace(wp.mat44f(*values)))
+        test.assertEqual(wp.trace(values_2d), wp.trace(wp.mat44f(*values)))
+
+
+def test_mat22_mat22_args_precision(test, device):
+    a_values = (0.12, 1.23, 0.12, 1.23)
+    a_values_2d = (a_values[0:2], a_values[2:4])
+    b_values = (1.23, 0.12, 1.23, 0.12)
+    b_values_2d = (b_values[0:2], b_values[2:4])
+    expected = 0.59039999999999992486
+
+    result = wp.ddot(wp.mat22d(*a_values), wp.mat22d(*b_values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.ddot(wp.mat22h(*a_values), wp.mat22h(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+        test.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+        test.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+        test.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+
+
+def test_mat33_mat33_args_precision(test, device):
+    a_values = (0.12, 1.23, 2.34, 0.12, 1.23, 2.34, 0.12, 1.23, 2.34)
+    a_values_2d = (a_values[0:3], a_values[3:6], a_values[6:9])
+    b_values = (2.34, 1.23, 0.12, 2.34, 1.23, 0.12, 2.34, 1.23, 0.12)
+    b_values_2d = (b_values[0:3], b_values[3:6], b_values[6:9])
+    expected = 6.22350000000000047606
+
+    result = wp.ddot(wp.mat33d(*a_values), wp.mat33d(*b_values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.ddot(wp.mat33h(*a_values), wp.mat33h(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+        test.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+        test.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+        test.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+
+
+def test_mat44_mat44_args(test, device):
+    a_values = (0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45)
+    a_values_2d = (a_values[0:4], a_values[4:8], a_values[8:12], a_values[12:16])
+    b_values = (3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12)
+    b_values_2d = (b_values[0:4], b_values[4:8], b_values[8:12], b_values[12:16])
+    expected = 26.33760000000000189857
+
+    result = wp.ddot(wp.mat44d(*a_values), wp.mat44d(*b_values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.ddot(wp.mat44h(*a_values), wp.mat44h(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+        test.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+        test.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+        test.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+
+
+def test_mat22_float_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45, 4.56)
+    a_values_2d = (a_values[0:2], a_values[2:4])
+    b_value = 0.12
+    expected_00 = 0.14759999999999998122
+    expected_01 = 0.28079999999999999405
+    expected_10 = 0.41399999999999997913
+    expected_11 = 0.54719999999999990870
+
+    result = wp.mul(wp.mat22d(*a_values), wp.float64(b_value))
+    test.assertAlmostEqual(result[0][0], expected_00, places=12)
+    test.assertAlmostEqual(result[0][1], expected_01, places=12)
+    test.assertAlmostEqual(result[1][0], expected_10, places=12)
+    test.assertAlmostEqual(result[1][1], expected_11, places=12)
+
+    result = wp.mul(wp.mat22f(*a_values), wp.float32(b_value))
+    test.assertNotAlmostEqual(result[0][0], expected_00, places=12)
+    test.assertNotAlmostEqual(result[0][1], expected_01, places=12)
+    test.assertNotAlmostEqual(result[1][0], expected_10, places=12)
+    test.assertNotAlmostEqual(result[1][1], expected_11, places=12)
+    test.assertAlmostEqual(result[0][0], expected_00, places=5)
+    test.assertAlmostEqual(result[0][1], expected_01, places=5)
+    test.assertAlmostEqual(result[1][0], expected_10, places=5)
+    test.assertAlmostEqual(result[1][1], expected_11, places=5)
+
+    result = wp.mul(wp.mat22h(*a_values), wp.float16(b_value))
+    test.assertNotAlmostEqual(result[0][0], expected_00, places=5)
+    test.assertNotAlmostEqual(result[0][1], expected_01, places=5)
+    test.assertNotAlmostEqual(result[1][0], expected_10, places=5)
+    test.assertNotAlmostEqual(result[1][1], expected_11, places=5)
+    test.assertAlmostEqual(result[0][0], expected_00, places=1)
+    test.assertAlmostEqual(result[0][1], expected_01, places=1)
+    test.assertAlmostEqual(result[1][0], expected_10, places=1)
+    test.assertAlmostEqual(result[1][1], expected_11, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        # Multiplying a 1-D tuple of length 4 is ambiguous because it could match
+        # either the `vec4f` or `mat22f` overload. As a result, only the 2-D variant
+        # of the tuple is expected to resolve correctly.
+        test.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat22f(*a_values), wp.float32(b_value)))
+
+
+def test_mat33_float_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
+    a_values_2d = (a_values[0:3], a_values[3:6], a_values[6:9])
+    b_value = 0.12
+    expected_00 = 0.14759999999999998122
+    expected_01 = 0.28079999999999999405
+    expected_02 = 0.41399999999999997913
+    expected_10 = 0.54719999999999990870
+    expected_11 = 0.68040000000000000480
+    expected_12 = 0.81359999999999998987
+    expected_20 = 0.94679999999999997495
+    expected_21 = 1.06800000000000006040
+    expected_22 = 1.08119999999999993889
+
+    result = wp.mul(wp.mat33d(*a_values), wp.float64(b_value))
+    test.assertAlmostEqual(result[0][0], expected_00, places=12)
+    test.assertAlmostEqual(result[0][1], expected_01, places=12)
+    test.assertAlmostEqual(result[0][2], expected_02, places=12)
+    test.assertAlmostEqual(result[1][0], expected_10, places=12)
+    test.assertAlmostEqual(result[1][1], expected_11, places=12)
+    test.assertAlmostEqual(result[1][2], expected_12, places=12)
+    test.assertAlmostEqual(result[2][0], expected_20, places=12)
+    test.assertAlmostEqual(result[2][1], expected_21, places=12)
+    test.assertAlmostEqual(result[2][2], expected_22, places=12)
+
+    result = wp.mul(wp.mat33f(*a_values), wp.float32(b_value))
+    test.assertNotAlmostEqual(result[0][0], expected_00, places=12)
+    test.assertNotAlmostEqual(result[0][1], expected_01, places=12)
+    test.assertNotAlmostEqual(result[0][2], expected_02, places=12)
+    test.assertNotAlmostEqual(result[1][0], expected_10, places=12)
+    test.assertNotAlmostEqual(result[1][1], expected_11, places=12)
+    test.assertNotAlmostEqual(result[1][2], expected_12, places=12)
+    test.assertNotAlmostEqual(result[2][0], expected_20, places=12)
+    test.assertNotAlmostEqual(result[2][1], expected_21, places=12)
+    test.assertNotAlmostEqual(result[2][2], expected_22, places=12)
+    test.assertAlmostEqual(result[0][0], expected_00, places=5)
+    test.assertAlmostEqual(result[0][1], expected_01, places=5)
+    test.assertAlmostEqual(result[0][2], expected_02, places=5)
+    test.assertAlmostEqual(result[1][0], expected_10, places=5)
+    test.assertAlmostEqual(result[1][1], expected_11, places=5)
+    test.assertAlmostEqual(result[1][2], expected_12, places=5)
+    test.assertAlmostEqual(result[2][0], expected_20, places=5)
+    test.assertAlmostEqual(result[2][1], expected_21, places=5)
+    test.assertAlmostEqual(result[2][2], expected_22, places=5)
+
+    result = wp.mul(wp.mat33h(*a_values), wp.float16(b_value))
+    test.assertNotAlmostEqual(result[0][0], expected_00, places=5)
+    test.assertNotAlmostEqual(result[0][1], expected_01, places=5)
+    test.assertNotAlmostEqual(result[0][2], expected_02, places=5)
+    test.assertNotAlmostEqual(result[1][0], expected_10, places=5)
+    test.assertNotAlmostEqual(result[1][1], expected_11, places=5)
+    test.assertNotAlmostEqual(result[1][2], expected_12, places=5)
+    test.assertNotAlmostEqual(result[2][0], expected_20, places=5)
+    test.assertNotAlmostEqual(result[2][1], expected_21, places=5)
+    test.assertNotAlmostEqual(result[2][2], expected_22, places=5)
+    test.assertAlmostEqual(result[0][0], expected_00, places=1)
+    test.assertAlmostEqual(result[0][1], expected_01, places=1)
+    test.assertAlmostEqual(result[0][2], expected_02, places=1)
+    test.assertAlmostEqual(result[1][0], expected_10, places=1)
+    test.assertAlmostEqual(result[1][1], expected_11, places=1)
+    test.assertAlmostEqual(result[1][2], expected_12, places=1)
+    test.assertAlmostEqual(result[2][0], expected_20, places=1)
+    test.assertAlmostEqual(result[2][1], expected_21, places=1)
+    test.assertAlmostEqual(result[2][2], expected_22, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.mat33f(*a_values), wp.float32(b_value)))
+        test.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat33f(*a_values), wp.float32(b_value)))
+
+
+def test_mat44_float_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01, 10.12, 11.23, 12.34, 13.45, 14.56, 15.67, 16.78)
+    a_values_2d = (a_values[0:4], a_values[4:8], a_values[8:12], a_values[12:16])
+    b_value = 0.12
+    expected_00 = 0.14759999999999998122
+    expected_01 = 0.28079999999999999405
+    expected_02 = 0.41399999999999997913
+    expected_03 = 0.54719999999999990870
+    expected_10 = 0.68040000000000000480
+    expected_11 = 0.81359999999999998987
+    expected_12 = 0.94679999999999997495
+    expected_13 = 1.06800000000000006040
+    expected_20 = 1.08119999999999993889
+    expected_21 = 1.21439999999999992397
+    expected_22 = 1.34759999999999990905
+    expected_23 = 1.48079999999999989413
+    expected_30 = 1.61399999999999987921
+    expected_31 = 1.74720000000000008633
+    expected_32 = 1.88039999999999984936
+    expected_33 = 2.01360000000000027853
+
+    result = wp.mul(wp.mat44d(*a_values), wp.float64(b_value))
+    test.assertAlmostEqual(result[0][0], expected_00, places=12)
+    test.assertAlmostEqual(result[0][1], expected_01, places=12)
+    test.assertAlmostEqual(result[0][2], expected_02, places=12)
+    test.assertAlmostEqual(result[0][3], expected_03, places=12)
+    test.assertAlmostEqual(result[1][0], expected_10, places=12)
+    test.assertAlmostEqual(result[1][1], expected_11, places=12)
+    test.assertAlmostEqual(result[1][2], expected_12, places=12)
+    test.assertAlmostEqual(result[1][3], expected_13, places=12)
+    test.assertAlmostEqual(result[2][0], expected_20, places=12)
+    test.assertAlmostEqual(result[2][1], expected_21, places=12)
+    test.assertAlmostEqual(result[2][2], expected_22, places=12)
+    test.assertAlmostEqual(result[2][3], expected_23, places=12)
+    test.assertAlmostEqual(result[3][0], expected_30, places=12)
+    test.assertAlmostEqual(result[3][1], expected_31, places=12)
+    test.assertAlmostEqual(result[3][2], expected_32, places=12)
+    test.assertAlmostEqual(result[3][3], expected_33, places=12)
+
+    result = wp.mul(wp.mat44f(*a_values), wp.float32(b_value))
+    test.assertNotAlmostEqual(result[0][0], expected_00, places=12)
+    test.assertNotAlmostEqual(result[0][1], expected_01, places=12)
+    test.assertNotAlmostEqual(result[0][2], expected_02, places=12)
+    test.assertNotAlmostEqual(result[0][3], expected_03, places=12)
+    test.assertNotAlmostEqual(result[1][0], expected_10, places=12)
+    test.assertNotAlmostEqual(result[1][1], expected_11, places=12)
+    test.assertNotAlmostEqual(result[1][2], expected_12, places=12)
+    test.assertNotAlmostEqual(result[1][3], expected_13, places=12)
+    test.assertNotAlmostEqual(result[2][0], expected_20, places=12)
+    test.assertNotAlmostEqual(result[2][1], expected_21, places=12)
+    test.assertNotAlmostEqual(result[2][2], expected_22, places=12)
+    test.assertNotAlmostEqual(result[2][3], expected_23, places=12)
+    test.assertNotAlmostEqual(result[3][0], expected_30, places=12)
+    test.assertNotAlmostEqual(result[3][1], expected_31, places=12)
+    test.assertNotAlmostEqual(result[3][2], expected_32, places=12)
+    test.assertNotAlmostEqual(result[3][3], expected_33, places=12)
+    test.assertAlmostEqual(result[0][0], expected_00, places=5)
+    test.assertAlmostEqual(result[0][1], expected_01, places=5)
+    test.assertAlmostEqual(result[0][2], expected_02, places=5)
+    test.assertAlmostEqual(result[0][3], expected_03, places=5)
+    test.assertAlmostEqual(result[1][0], expected_10, places=5)
+    test.assertAlmostEqual(result[1][1], expected_11, places=5)
+    test.assertAlmostEqual(result[1][2], expected_12, places=5)
+    test.assertAlmostEqual(result[1][3], expected_13, places=5)
+    test.assertAlmostEqual(result[2][0], expected_20, places=5)
+    test.assertAlmostEqual(result[2][1], expected_21, places=5)
+    test.assertAlmostEqual(result[2][2], expected_22, places=5)
+    test.assertAlmostEqual(result[2][3], expected_23, places=5)
+    test.assertAlmostEqual(result[3][0], expected_30, places=5)
+    test.assertAlmostEqual(result[3][1], expected_31, places=5)
+    test.assertAlmostEqual(result[3][2], expected_32, places=5)
+    test.assertAlmostEqual(result[3][3], expected_33, places=5)
+
+    result = wp.mul(wp.mat44h(*a_values), wp.float16(b_value))
+    test.assertNotAlmostEqual(result[0][0], expected_00, places=5)
+    test.assertNotAlmostEqual(result[0][1], expected_01, places=5)
+    test.assertNotAlmostEqual(result[0][2], expected_02, places=5)
+    test.assertNotAlmostEqual(result[0][3], expected_03, places=5)
+    test.assertNotAlmostEqual(result[1][0], expected_10, places=5)
+    test.assertNotAlmostEqual(result[1][1], expected_11, places=5)
+    test.assertNotAlmostEqual(result[1][2], expected_12, places=5)
+    test.assertNotAlmostEqual(result[1][3], expected_13, places=5)
+    test.assertNotAlmostEqual(result[2][0], expected_20, places=5)
+    test.assertNotAlmostEqual(result[2][1], expected_21, places=5)
+    test.assertNotAlmostEqual(result[2][2], expected_22, places=5)
+    test.assertNotAlmostEqual(result[2][3], expected_23, places=5)
+    test.assertNotAlmostEqual(result[3][0], expected_30, places=5)
+    test.assertNotAlmostEqual(result[3][1], expected_31, places=5)
+    test.assertNotAlmostEqual(result[3][2], expected_32, places=5)
+    test.assertNotAlmostEqual(result[3][3], expected_33, places=5)
+    test.assertAlmostEqual(result[0][0], expected_00, places=1)
+    test.assertAlmostEqual(result[0][1], expected_01, places=1)
+    test.assertAlmostEqual(result[0][2], expected_02, places=1)
+    test.assertAlmostEqual(result[0][3], expected_03, places=1)
+    test.assertAlmostEqual(result[1][0], expected_10, places=1)
+    test.assertAlmostEqual(result[1][1], expected_11, places=1)
+    test.assertAlmostEqual(result[1][2], expected_12, places=1)
+    test.assertAlmostEqual(result[1][3], expected_13, places=1)
+    test.assertAlmostEqual(result[2][0], expected_20, places=1)
+    test.assertAlmostEqual(result[2][1], expected_21, places=1)
+    test.assertAlmostEqual(result[2][2], expected_22, places=1)
+    test.assertAlmostEqual(result[2][3], expected_23, places=1)
+    test.assertAlmostEqual(result[3][0], expected_30, places=1)
+    test.assertAlmostEqual(result[3][1], expected_31, places=1)
+    test.assertAlmostEqual(result[3][2], expected_32, places=1)
+    test.assertAlmostEqual(result[3][3], expected_33, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.mat44f(*a_values), wp.float32(b_value)))
+        test.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat44f(*a_values), wp.float32(b_value)))
+
+
+def test_vec2_arg_precision(test, device):
+    values = (1.23, 2.34)
+    expected = 2.64357712200722438922
+
+    result = wp.length(wp.vec2d(*values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.length(wp.vec2f(*values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.length(wp.vec2h(*values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.length(values), wp.length(wp.vec2f(*values)))
+
+
+def test_vec2_arg_overflow(test, device):
+    values = (-1234567890, -1234567890)
+
+    test.assertEqual(wp.length_sq(wp.vec2b(*values)), -120)
+    test.assertEqual(wp.length_sq(wp.vec2s(*values)), -6008)
+    test.assertEqual(wp.length_sq(wp.vec2i(*values)), 608168072)
+    test.assertEqual(wp.length_sq(wp.vec2l(*values)), 3048315750038104200)
+
+    test.assertEqual(wp.length_sq(wp.vec2ub(*values)), 136)
+    test.assertEqual(wp.length_sq(wp.vec2us(*values)), 59528)
+    test.assertEqual(wp.length_sq(wp.vec2ui(*values)), 608168072)
+    test.assertEqual(wp.length_sq(wp.vec2ul(*values)), 3048315750038104200)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec2i(*values)))
+
+
+def test_vec3_arg_precision(test, device):
+    values = (1.23, 2.34, 3.45)
+    expected = 4.34637780226247727455
+
+    result = wp.length(wp.vec3d(*values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.length(wp.vec3f(*values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.length(wp.vec3h(*values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.length(values), wp.length(wp.vec3f(*values)))
+
+
+def test_vec3_arg_overflow(test, device):
+    values = (-1234567890, -1234567890, -1234567890)
+
+    test.assertEqual(wp.length_sq(wp.vec3b(*values)), -52)
+    test.assertEqual(wp.length_sq(wp.vec3s(*values)), -9012)
+    test.assertEqual(wp.length_sq(wp.vec3i(*values)), 912252108)
+    test.assertEqual(wp.length_sq(wp.vec3l(*values)), 4572473625057156300)
+
+    test.assertEqual(wp.length_sq(wp.vec3ub(*values)), 204)
+    test.assertEqual(wp.length_sq(wp.vec3us(*values)), 56524)
+    test.assertEqual(wp.length_sq(wp.vec3ui(*values)), 912252108)
+    test.assertEqual(wp.length_sq(wp.vec3ul(*values)), 4572473625057156300)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec3i(*values)))
+
+
+def test_vec4_arg_precision(test, device):
+    values = (1.23, 2.34, 3.45, 4.56)
+    expected = 6.29957141399317777086
+
+    result = wp.length(wp.vec4d(*values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.length(wp.vec4f(*values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.length(wp.vec4h(*values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.length(values), wp.length(wp.vec4f(*values)))
+
+
+def test_vec4_arg_overflow(test, device):
+    values = (-1234567890, -1234567890, -1234567890, -1234567890)
+
+    test.assertEqual(wp.length_sq(wp.vec4b(*values)), 16)
+    test.assertEqual(wp.length_sq(wp.vec4s(*values)), -12016)
+    test.assertEqual(wp.length_sq(wp.vec4i(*values)), 1216336144)
+    test.assertEqual(wp.length_sq(wp.vec4l(*values)), 6096631500076208400)
+
+    test.assertEqual(wp.length_sq(wp.vec4ub(*values)), 16)
+    test.assertEqual(wp.length_sq(wp.vec4us(*values)), 53520)
+    test.assertEqual(wp.length_sq(wp.vec4ui(*values)), 1216336144)
+    test.assertEqual(wp.length_sq(wp.vec4ul(*values)), 6096631500076208400)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec4i(*values)))
+
+
+def test_vec2_vec2_args_precision(test, device):
+    a_values = (1.23, 2.34)
+    b_values = (3.45, 4.56)
+    expected = 14.91389999999999815827
+
+    result = wp.dot(wp.vec2d(*a_values), wp.vec2d(*b_values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.dot(wp.vec2f(*a_values), wp.vec2f(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.dot(wp.vec2h(*a_values), wp.vec2h(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec2f(*a_values), wp.vec2f(*b_values)))
+
+
+def test_vec2_vec2_args_overflow(test, device):
+    values = (-1234567890, -1234567890)
+
+    test.assertEqual(wp.dot(wp.vec2b(*values), wp.vec2b(*values)), -120)
+    test.assertEqual(wp.dot(wp.vec2s(*values), wp.vec2s(*values)), -6008)
+    test.assertEqual(wp.dot(wp.vec2i(*values), wp.vec2i(*values)), 608168072)
+    test.assertEqual(wp.dot(wp.vec2l(*values), wp.vec2l(*values)), 3048315750038104200)
+
+    test.assertEqual(wp.dot(wp.vec2ub(*values), wp.vec2ub(*values)), 136)
+    test.assertEqual(wp.dot(wp.vec2us(*values), wp.vec2us(*values)), 59528)
+    test.assertEqual(wp.dot(wp.vec2ui(*values), wp.vec2ui(*values)), 608168072)
+    test.assertEqual(wp.dot(wp.vec2ul(*values), wp.vec2ul(*values)), 3048315750038104200)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(values, values), wp.dot(wp.vec2i(*values), wp.vec2i(*values)))
+
+
+def test_vec3_vec3_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45)
+    b_values = (4.56, 5.67, 6.78)
+    expected = 42.26760000000000161435
+
+    result = wp.dot(wp.vec3d(*a_values), wp.vec3d(*b_values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.dot(wp.vec3f(*a_values), wp.vec3f(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.dot(wp.vec3h(*a_values), wp.vec3h(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec3f(*a_values), wp.vec3f(*b_values)))
+
+
+def test_vec3_vec3_args_overflow(test, device):
+    values = (-1234567890, -1234567890, -1234567890)
+
+    test.assertEqual(wp.dot(wp.vec3b(*values), wp.vec3b(*values)), -52)
+    test.assertEqual(wp.dot(wp.vec3s(*values), wp.vec3s(*values)), -9012)
+    test.assertEqual(wp.dot(wp.vec3i(*values), wp.vec3i(*values)), 912252108)
+    test.assertEqual(wp.dot(wp.vec3l(*values), wp.vec3l(*values)), 4572473625057156300)
+
+    test.assertEqual(wp.dot(wp.vec3ub(*values), wp.vec3ub(*values)), 204)
+    test.assertEqual(wp.dot(wp.vec3us(*values), wp.vec3us(*values)), 56524)
+    test.assertEqual(wp.dot(wp.vec3ui(*values), wp.vec3ui(*values)), 912252108)
+    test.assertEqual(wp.dot(wp.vec3ul(*values), wp.vec3ul(*values)), 4572473625057156300)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(values, values), wp.dot(wp.vec3i(*values), wp.vec3i(*values)))
+
+
+def test_vec4_vec4_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45, 4.56)
+    b_values = (5.67, 6.78, 7.89, 8.90)
+    expected = 90.64379999999999881766
+
+    result = wp.dot(wp.vec4d(*a_values), wp.vec4d(*b_values))
+    test.assertAlmostEqual(result, expected, places=12)
+
+    result = wp.dot(wp.vec4f(*a_values), wp.vec4f(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=12)
+    test.assertAlmostEqual(result, expected, places=5)
+
+    result = wp.dot(wp.vec4h(*a_values), wp.vec4h(*b_values))
+    test.assertNotAlmostEqual(result, expected, places=5)
+    test.assertAlmostEqual(result, expected, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec4f(*a_values), wp.vec4f(*b_values)))
+
+
+def test_vec4_vec4_args_overflow(test, device):
+    values = (-1234567890, -1234567890, -1234567890, -1234567890)
+
+    test.assertEqual(wp.dot(wp.vec4b(*values), wp.vec4b(*values)), 16)
+    test.assertEqual(wp.dot(wp.vec4s(*values), wp.vec4s(*values)), -12016)
+    test.assertEqual(wp.dot(wp.vec4i(*values), wp.vec4i(*values)), 1216336144)
+    test.assertEqual(wp.dot(wp.vec4l(*values), wp.vec4l(*values)), 6096631500076208400)
+
+    test.assertEqual(wp.dot(wp.vec4ub(*values), wp.vec4ub(*values)), 16)
+    test.assertEqual(wp.dot(wp.vec4us(*values), wp.vec4us(*values)), 53520)
+    test.assertEqual(wp.dot(wp.vec4ui(*values), wp.vec4ui(*values)), 1216336144)
+    test.assertEqual(wp.dot(wp.vec4ul(*values), wp.vec4ul(*values)), 6096631500076208400)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(values, values), wp.dot(wp.vec4i(*values), wp.vec4i(*values)))
+
+
+def test_vec2_float_args_precision(test, device):
+    a_values = (1.23, 2.34)
+    b_value = 3.45
+    expected_x = 4.24350000000000004974
+    expected_y = 8.07300000000000039790
+
+    result = wp.mul(wp.vec2d(*a_values), wp.float64(b_value))
+    test.assertAlmostEqual(result[0], expected_x, places=12)
+    test.assertAlmostEqual(result[1], expected_y, places=12)
+
+    result = wp.mul(wp.vec2f(*a_values), wp.float32(b_value))
+    test.assertNotAlmostEqual(result[0], expected_x, places=12)
+    test.assertNotAlmostEqual(result[1], expected_y, places=12)
+    test.assertAlmostEqual(result[0], expected_x, places=5)
+    test.assertAlmostEqual(result[1], expected_y, places=5)
+
+    result = wp.mul(wp.vec2h(*a_values), wp.float16(b_value))
+    test.assertNotAlmostEqual(result[0], expected_x, places=5)
+    test.assertNotAlmostEqual(result[1], expected_y, places=5)
+    test.assertAlmostEqual(result[0], expected_x, places=1)
+    test.assertAlmostEqual(result[1], expected_y, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec2f(*a_values), wp.float32(b_value)))
+
+
+def test_vec3_float_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45)
+    b_value = 4.56
+    expected_x = 5.60879999999999956373
+    expected_y = 10.67039999999999899671
+    expected_z = 15.73199999999999931788
+
+    result = wp.mul(wp.vec3d(*a_values), wp.float64(b_value))
+    test.assertAlmostEqual(result[0], expected_x, places=12)
+    test.assertAlmostEqual(result[1], expected_y, places=12)
+    test.assertAlmostEqual(result[2], expected_z, places=12)
+
+    result = wp.mul(wp.vec3f(*a_values), wp.float32(b_value))
+    test.assertNotAlmostEqual(result[0], expected_x, places=12)
+    test.assertNotAlmostEqual(result[1], expected_y, places=12)
+    test.assertNotAlmostEqual(result[2], expected_z, places=12)
+    test.assertAlmostEqual(result[0], expected_x, places=5)
+    test.assertAlmostEqual(result[1], expected_y, places=5)
+    test.assertAlmostEqual(result[2], expected_z, places=5)
+
+    result = wp.mul(wp.vec3h(*a_values), wp.float16(b_value))
+    test.assertNotAlmostEqual(result[0], expected_x, places=5)
+    test.assertNotAlmostEqual(result[1], expected_y, places=5)
+    test.assertNotAlmostEqual(result[2], expected_z, places=5)
+    test.assertAlmostEqual(result[0], expected_x, places=1)
+    test.assertAlmostEqual(result[1], expected_y, places=1)
+    test.assertAlmostEqual(result[2], expected_z, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec3f(*a_values), wp.float32(b_value)))
+
+
+def test_vec4_float_args_precision(test, device):
+    a_values = (1.23, 2.34, 3.45, 4.56)
+    b_value = 5.67
+    expected_x = 6.97409999999999996589
+    expected_y = 13.26779999999999937188
+    expected_z = 19.56150000000000233058
+    expected_w = 25.85519999999999640750
+
+    result = wp.mul(wp.vec4d(*a_values), wp.float64(b_value))
+    test.assertAlmostEqual(result[0], expected_x, places=12)
+    test.assertAlmostEqual(result[1], expected_y, places=12)
+    test.assertAlmostEqual(result[2], expected_z, places=12)
+    test.assertAlmostEqual(result[3], expected_w, places=12)
+
+    result = wp.mul(wp.vec4f(*a_values), wp.float32(b_value))
+    test.assertNotAlmostEqual(result[0], expected_x, places=12)
+    test.assertNotAlmostEqual(result[1], expected_y, places=12)
+    test.assertNotAlmostEqual(result[2], expected_z, places=12)
+    test.assertNotAlmostEqual(result[3], expected_w, places=12)
+    test.assertAlmostEqual(result[0], expected_x, places=5)
+    test.assertAlmostEqual(result[1], expected_y, places=5)
+    test.assertAlmostEqual(result[2], expected_z, places=5)
+    test.assertAlmostEqual(result[3], expected_w, places=5)
+
+    result = wp.mul(wp.vec4h(*a_values), wp.float16(b_value))
+    test.assertNotAlmostEqual(result[0], expected_x, places=5)
+    test.assertNotAlmostEqual(result[1], expected_y, places=5)
+    test.assertNotAlmostEqual(result[2], expected_z, places=5)
+    test.assertNotAlmostEqual(result[3], expected_w, places=5)
+    test.assertAlmostEqual(result[0], expected_x, places=1)
+    test.assertAlmostEqual(result[1], expected_y, places=1)
+    test.assertAlmostEqual(result[2], expected_z, places=1)
+    test.assertAlmostEqual(result[3], expected_w, places=1)
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec4f(*a_values), wp.float32(b_value)))
+
+
+def test_int_arg_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    value = -1234567890123456789
+    expected = wp.invert(dtype(value))
+
+    test.assertEqual(wp.invert(nps(np_type, value)), expected)
+
+
+def test_float_arg_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    value = 1.23
+    expected = wp.sin(dtype(value))
+
+    test.assertEqual(wp.sin(nps(np_type, value)), expected)
+
+
+def test_int_int_args_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    value = -1234567890
+    expected = wp.mul(dtype(value), dtype(value))
+
+    test.assertEqual(wp.mul(dtype(value), dtype(value)), expected)
+    test.assertEqual(wp.mul(dtype(value), nps(np_type, value)), expected)
+
+    test.assertEqual(wp.mul(nps(np_type, value), dtype(value)), expected)
+    test.assertEqual(wp.mul(nps(np_type, value), nps(np_type, value)), expected)
+
+    if dtype is wp.int32:
+        test.assertEqual(wp.mul(dtype(value), value), expected)
+        test.assertEqual(wp.mul(nps(np_type, value), value), expected)
+        test.assertEqual(wp.mul(value, value), expected)
+
+        test.assertEqual(wp.mul(value, dtype(value)), expected)
+        test.assertEqual(wp.mul(value, nps(np_type, value)), expected)
+    else:
+        with test.assertRaisesRegex(
+            RuntimeError,
+            rf"Couldn't find a function 'mul' compatible with "
+            rf"the arguments '{dtype.__name__}, int'$",
+        ):
+            wp.mul(dtype(value), value)
+
+        with test.assertRaisesRegex(
+            RuntimeError,
+            rf"Couldn't find a function 'mul' compatible with "
+            rf"the arguments '{np_type.__name__}, int'$",
+        ):
+            wp.mul(nps(np_type, value), value)
+
+        with test.assertRaisesRegex(
+            RuntimeError,
+            rf"Couldn't find a function 'mul' compatible with "
+            rf"the arguments 'int, {dtype.__name__}'$",
+        ):
+            wp.mul(value, dtype(value))
+
+        with test.assertRaisesRegex(
+            RuntimeError,
+            rf"Couldn't find a function 'mul' compatible with "
+            rf"the arguments 'int, {np_type.__name__}'$",
+        ):
+            wp.mul(value, nps(np_type, value))
+
+
+def test_mat_arg_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    mat_cls = wp.types.matrix((3, 3), dtype)
+    values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
+    expected = wp.trace(mat_cls(*values))
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.trace(wpv(dtype, values)), expected)
+        test.assertEqual(wp.trace(wpm(dtype, 3, values)), expected)
+        test.assertEqual(wp.trace(npv(np_type, values)), expected)
+        test.assertEqual(wp.trace(npm(np_type, 3, values)), expected)
+        test.assertEqual(wp.trace(np.array(npv(np_type, values))), expected)
+
+
+def test_mat_mat_args_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    mat_cls = wp.types.matrix((3, 3), dtype)
+    a_values = (0.12, 1.23, 2.34, 0.12, 1.23, 2.34, 0.12, 1.23, 2.34)
+    b_values = (2.34, 1.23, 0.12, 2.34, 1.23, 0.12, 2.34, 1.23, 0.12)
+    expected = wp.ddot(mat_cls(*a_values), mat_cls(*b_values))
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.ddot(mat_cls(*a_values), mat_cls(*b_values)), expected)
+        test.assertEqual(wp.ddot(mat_cls(*a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.ddot(mat_cls(*a_values), wpm(dtype, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(mat_cls(*a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.ddot(mat_cls(*a_values), npm(np_type, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(mat_cls(*a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.ddot(wpv(dtype, a_values), mat_cls(*b_values)), expected)
+        test.assertEqual(wp.ddot(wpv(dtype, a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.ddot(wpv(dtype, a_values), wpm(dtype, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(wpv(dtype, a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.ddot(wpv(dtype, a_values), npm(np_type, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(wpv(dtype, a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), mat_cls(*b_values)), expected)
+        test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), wpm(dtype, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), npm(np_type, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.ddot(npv(np_type, a_values), mat_cls(*b_values)), expected)
+        test.assertEqual(wp.ddot(npv(np_type, a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.ddot(npv(np_type, a_values), wpm(dtype, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(npv(np_type, a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.ddot(npv(np_type, a_values), npm(np_type, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(npv(np_type, a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.ddot(npm(np_type, 3, a_values), mat_cls(*b_values)), expected)
+        test.assertEqual(wp.ddot(npm(np_type, 3, a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.ddot(npm(np_type, 3, a_values), wpm(dtype, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(npm(np_type, 3, a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.ddot(npm(np_type, 3, a_values), npm(np_type, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(npm(np_type, 3, a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), mat_cls(*b_values)), expected)
+        test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), wpm(dtype, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), npm(np_type, 3, b_values)), expected)
+        test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), np.array(npv(np_type, b_values))), expected)
+
+        if dtype is wp.float32:
+            test.assertEqual(wp.ddot(mat_cls(*a_values), b_values), expected)
+            test.assertEqual(wp.ddot(wpv(dtype, a_values), b_values), expected)
+            test.assertEqual(wp.ddot(wpm(dtype, 3, a_values), b_values), expected)
+            test.assertEqual(wp.ddot(npv(np_type, a_values), b_values), expected)
+            test.assertEqual(wp.ddot(npm(np_type, 3, a_values), b_values), expected)
+            test.assertEqual(wp.ddot(a_values, b_values), expected)
+            test.assertEqual(wp.ddot(np.array(npv(np_type, a_values)), b_values), expected)
+
+            test.assertEqual(wp.ddot(a_values, mat_cls(*b_values)), expected)
+            test.assertEqual(wp.ddot(a_values, wpv(dtype, b_values)), expected)
+            test.assertEqual(wp.ddot(a_values, wpm(dtype, 3, b_values)), expected)
+            test.assertEqual(wp.ddot(a_values, npv(np_type, b_values)), expected)
+            test.assertEqual(wp.ddot(a_values, npm(np_type, 3, b_values)), expected)
+            test.assertEqual(wp.ddot(a_values, np.array(npv(np_type, b_values))), expected)
+        else:
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'mat_t, tuple'$",
+            ):
+                wp.ddot(mat_cls(*a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(wpv(dtype, a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(wpm(dtype, 3, a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(npv(np_type, a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(npm(np_type, 3, a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'ndarray, tuple'$",
+            ):
+                wp.ddot(np.array(npv(np_type, a_values)), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, mat_t'$",
+            ):
+                wp.ddot(a_values, mat_cls(*b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(a_values, wpv(dtype, b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(a_values, wpm(dtype, 3, b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(a_values, npv(np_type, b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.ddot(a_values, npm(np_type, 3, b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'ddot' compatible with "
+                r"the arguments 'tuple, ndarray'$",
+            ):
+                wp.ddot(a_values, np.array(npv(np_type, b_values)))
+
+
+def test_mat_float_args_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    mat_cls = wp.types.matrix((3, 3), dtype)
+    a_values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
+    b_value = 0.12
+    expected = wp.mul(mat_cls(*a_values), dtype(b_value))
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(mat_cls(*a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(mat_cls(*a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(wpv(dtype, a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(wpv(dtype, a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(wpm(dtype, 3, a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(wpm(dtype, 3, a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(npv(np_type, a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(npv(np_type, a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(npm(np_type, 3, a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(npm(np_type, 3, a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(np.array(npv(np_type, a_values)), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(np.array(npv(np_type, a_values)), nps(np_type, b_value)), expected)
+
+        if dtype is wp.float32:
+            test.assertEqual(wp.mul(mat_cls(*a_values), b_value), expected)
+            test.assertEqual(wp.mul(wpv(dtype, a_values), b_value), expected)
+            test.assertEqual(wp.mul(wpm(dtype, 3, a_values), b_value), expected)
+            test.assertEqual(wp.mul(npv(np_type, a_values), b_value), expected)
+            test.assertEqual(wp.mul(npm(np_type, 3, a_values), b_value), expected)
+            test.assertEqual(wp.mul(a_values, b_value), expected)
+            test.assertEqual(wp.mul(np.array(npv(np_type, a_values)), b_value), expected)
+
+            test.assertEqual(wp.mul(a_values, dtype(b_value)), expected)
+            test.assertEqual(wp.mul(a_values, nps(np_type, b_value)), expected)
+        else:
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'mat_t, float'$",
+            ):
+                wp.mul(mat_cls(*a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'tuple, float'$",
+            ):
+                wp.mul(wpv(dtype, a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'tuple, float'$",
+            ):
+                wp.mul(wpm(dtype, 3, a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'tuple, float'$",
+            ):
+                wp.mul(npv(np_type, a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'tuple, float'$",
+            ):
+                wp.mul(npm(np_type, 3, a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'ndarray, float'$",
+            ):
+                wp.mul(np.array(npv(np_type, a_values)), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                fr"Couldn't find a function 'mul' compatible with "
+                fr"the arguments 'tuple, {dtype.__name__}'$",
+            ):
+                wp.mul(a_values, dtype(b_value))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                fr"Couldn't find a function 'mul' compatible with "
+                fr"the arguments 'tuple, {np_type.__name__}'$",
+            ):
+                wp.mul(a_values, nps(np_type, b_value))
+
+
+def test_vec_arg_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    vec_cls = wp.types.vector(3, dtype)
+    values = (1.23, 2.34, 3.45)
+    expected = wp.length(vec_cls(*values))
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertAlmostEqual(wp.length(wpv(dtype, values)), expected)
+        test.assertAlmostEqual(wp.length(npv(np_type, values)), expected)
+        test.assertAlmostEqual(wp.length(np.array(npv(np_type, values))), expected)
+
+
+def test_vec_vec_args_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    vec_cls = wp.types.vector(3, dtype)
+    a_values = (1.23, 2.34, 3.45)
+    b_values = (4.56, 5.67, 6.78)
+    expected = wp.dot(vec_cls(*a_values), vec_cls(*b_values))
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.dot(vec_cls(*a_values), vec_cls(*b_values)), expected)
+        test.assertEqual(wp.dot(vec_cls(*a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.dot(vec_cls(*a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.dot(vec_cls(*a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.dot(wpv(dtype, a_values), vec_cls(*b_values)), expected)
+        test.assertEqual(wp.dot(wpv(dtype, a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.dot(wpv(dtype, a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.dot(wpv(dtype, a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.dot(npv(np_type, a_values), vec_cls(*b_values)), expected)
+        test.assertEqual(wp.dot(npv(np_type, a_values), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.dot(npv(np_type, a_values), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.dot(npv(np_type, a_values), np.array(npv(np_type, b_values))), expected)
+
+        test.assertEqual(wp.dot(np.array(npv(np_type, a_values)), vec_cls(*b_values)), expected)
+        test.assertEqual(wp.dot(np.array(npv(np_type, a_values)), wpv(dtype, b_values)), expected)
+        test.assertEqual(wp.dot(np.array(npv(np_type, a_values)), npv(np_type, b_values)), expected)
+        test.assertEqual(wp.dot(np.array(npv(np_type, a_values)), np.array(npv(np_type, b_values))), expected)
+
+        if dtype is wp.float32:
+            test.assertEqual(wp.dot(vec_cls(*a_values), b_values), expected)
+            test.assertEqual(wp.dot(wpv(dtype, a_values), b_values), expected)
+            test.assertEqual(wp.dot(npv(np_type, a_values), b_values), expected)
+            test.assertEqual(wp.dot(a_values, b_values), expected)
+            test.assertEqual(wp.dot(np.array(npv(np_type, a_values)), b_values), expected)
+
+            test.assertEqual(wp.dot(a_values, vec_cls(*b_values)), expected)
+            test.assertEqual(wp.dot(a_values, wpv(dtype, b_values)), expected)
+            test.assertEqual(wp.dot(a_values, npv(np_type, b_values)), expected)
+            test.assertEqual(wp.dot(a_values, np.array(npv(np_type, b_values))), expected)
+        else:
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'vec_t, tuple'$",
+            ):
+                wp.dot(vec_cls(*a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.dot(wpv(dtype, a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.dot(npv(np_type, a_values), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'ndarray, tuple'$",
+            ):
+                wp.dot(np.array(npv(np_type, a_values)), b_values)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'tuple, vec_t'$",
+            ):
+                wp.dot(a_values, vec_cls(*b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.dot(a_values, wpv(dtype, b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'tuple, tuple'$",
+            ):
+                wp.dot(a_values, npv(np_type, b_values))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'dot' compatible with "
+                r"the arguments 'tuple, ndarray'$",
+            ):
+                wp.dot(a_values, np.array(npv(np_type, b_values)))
+
+
+def test_vec_float_args_support(test, device, dtype):
+    np_type = wp.types.warp_type_to_np_dtype[dtype]
+    vec_cls = wp.types.vector(3, dtype)
+    a_values = (1.23, 2.34, 3.45)
+    b_value = 4.56
+    expected = wp.mul(vec_cls(*a_values), dtype(b_value))
+
+    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+        test.assertEqual(wp.mul(vec_cls(*a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(vec_cls(*a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(wpv(dtype, a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(wpv(dtype, a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(npv(np_type, a_values), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(npv(np_type, a_values), nps(np_type, b_value)), expected)
+
+        test.assertEqual(wp.mul(np.array(npv(np_type, a_values)), dtype(b_value)), expected)
+        test.assertEqual(wp.mul(np.array(npv(np_type, a_values)), nps(np_type, b_value)), expected)
+
+        if dtype is wp.float32:
+            test.assertEqual(wp.mul(vec_cls(*a_values), b_value), expected)
+            test.assertEqual(wp.mul(wpv(dtype, a_values), b_value), expected)
+            test.assertEqual(wp.mul(npv(np_type, a_values), b_value), expected)
+            test.assertEqual(wp.mul(a_values, b_value), expected)
+            test.assertEqual(wp.mul(np.array(npv(np_type, a_values)), b_value), expected)
+
+            test.assertEqual(wp.mul(a_values, dtype(b_value)), expected)
+            test.assertEqual(wp.mul(a_values, nps(np_type, b_value)), expected)
+        else:
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'vec_t, float'$",
+            ):
+                wp.mul(vec_cls(*a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'tuple, float'$",
+            ):
+                wp.mul(wpv(dtype, a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'tuple, float'$",
+            ):
+                wp.mul(npv(np_type, a_values), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                r"Couldn't find a function 'mul' compatible with "
+                r"the arguments 'ndarray, float'$",
+            ):
+                wp.mul(np.array(npv(np_type, a_values)), b_value)
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                fr"Couldn't find a function 'mul' compatible with "
+                fr"the arguments 'tuple, {dtype.__name__}'$",
+            ):
+                wp.mul(a_values, dtype(b_value))
+
+            with test.assertRaisesRegex(
+                RuntimeError,
+                fr"Couldn't find a function 'mul' compatible with "
+                fr"the arguments 'tuple, {np_type.__name__}'$",
+            ):
+                wp.mul(a_values, nps(np_type, b_value))
+
+
+def register(parent):
+    class TestBuiltinsResolution(parent):
+        pass
+
+    add_function_test(TestBuiltinsResolution, "test_int_arg_overflow", test_int_arg_overflow)
+    add_function_test(TestBuiltinsResolution, "test_float_arg_precision", test_float_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_int_int_args_overflow", test_int_int_args_overflow)
+    add_function_test(TestBuiltinsResolution, "test_mat22_arg_precision", test_mat22_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat33_arg_precision", test_mat33_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat44_arg_precision", test_mat44_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat22_mat22_args_precision", test_mat22_mat22_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat33_mat33_args_precision", test_mat33_mat33_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat44_mat44_args", test_mat44_mat44_args)
+    add_function_test(TestBuiltinsResolution, "test_mat22_float_args_precision", test_mat22_float_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat33_float_args_precision", test_mat33_float_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_mat44_float_args_precision", test_mat44_float_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec2_arg_precision", test_vec2_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec2_arg_overflow", test_vec2_arg_overflow)
+    add_function_test(TestBuiltinsResolution, "test_vec3_arg_precision", test_vec3_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec3_arg_overflow", test_vec3_arg_overflow)
+    add_function_test(TestBuiltinsResolution, "test_vec4_arg_precision", test_vec4_arg_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec4_arg_overflow", test_vec4_arg_overflow)
+    add_function_test(TestBuiltinsResolution, "test_vec2_vec2_args_precision", test_vec2_vec2_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec2_vec2_args_overflow", test_vec2_vec2_args_overflow)
+    add_function_test(TestBuiltinsResolution, "test_vec3_vec3_args_precision", test_vec3_vec3_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec3_vec3_args_overflow", test_vec3_vec3_args_overflow)
+    add_function_test(TestBuiltinsResolution, "test_vec4_vec4_args_precision", test_vec4_vec4_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec4_vec4_args_overflow", test_vec4_vec4_args_overflow)
+    add_function_test(TestBuiltinsResolution, "test_vec2_float_args_precision", test_vec2_float_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec3_float_args_precision", test_vec3_float_args_precision)
+    add_function_test(TestBuiltinsResolution, "test_vec4_float_args_precision", test_vec4_float_args_precision)
+
+    for dtype in wp.types.int_types:
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_int_arg_support_{dtype.__name__}",
+            test_int_arg_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_int_int_args_support_{dtype.__name__}",
+            test_int_int_args_support,
+            dtype=dtype,
+        )
+
+    for dtype in wp.types.float_types:
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_float_arg_support_{dtype.__name__}",
+            test_float_arg_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_mat_arg_support_{dtype.__name__}",
+            test_mat_arg_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_mat_mat_args_support_{dtype.__name__}",
+            test_mat_mat_args_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_mat_float_args_support_{dtype.__name__}",
+            test_mat_float_args_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_vec_arg_support_{dtype.__name__}",
+            test_vec_arg_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_vec_vec_args_support_{dtype.__name__}",
+            test_vec_vec_args_support,
+            dtype=dtype,
+        )
+        add_function_test(
+            TestBuiltinsResolution,
+            f"test_vec_float_args_support_{dtype.__name__}",
+            test_vec_float_args_support,
+            dtype=dtype,
+        )
+
+    return TestBuiltinsResolution
+
+
+if __name__ == "__main__":
+    wp.build.clear_kernel_cache()
+    _ = register(unittest.TestCase)
+    unittest.main(verbosity=2)
diff --git a/warp/tests/test_types.py b/warp/tests/test_types.py
index 424a68c4d..9d140367e 100644
--- a/warp/tests/test_types.py
+++ b/warp/tests/test_types.py
@@ -5,9 +5,6 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import contextlib
-import io
-import inspect
 import unittest
 
 from warp.tests.test_base import *
@@ -44,39 +41,213 @@ def test_constant_error_invalid_type(test, device):
         wp.constant((1, 2, 3))
 
 
-def test_vector(test, device):
-    for dtype in (int, float, wp.float16):
-        vec_cls = wp.vec(3, dtype)
+def test_vector(test, device, dtype):
+    def make_scalar(x):
+        # Cast to the correct integer type to simulate wrapping.
+        if dtype in wp.types.int_types:
+            return dtype._type_(x).value
 
-        v = vec_cls(1, 2, 3)
-        test.assertEqual(v[0], 1)
-        test.assertSequenceEqual(v[0:2], (1, 2))
-        test.assertSequenceEqual(v, (1, 2, 3))
+        return x
 
-        v[0] = -1
-        test.assertEqual(v[0], -1)
-        test.assertSequenceEqual(v[0:2], (-1, 2))
-        test.assertSequenceEqual(v, (-1, 2, 3))
+    def make_vec(*args):
+        if dtype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(dtype._type_(x).value for x in args)
 
-        v[1:3] = (-2, -3)
-        test.assertEqual(v[0], -1)
-        test.assertSequenceEqual(v[0:2], (-1, -2))
-        test.assertSequenceEqual(v, (-1, -2, -3))
+        return args
 
-        v += vec_cls(1, 1, 1)
-        test.assertSequenceEqual(v, (0, -1, -2))
+    vec3_cls = wp.vec(3, dtype)
+    vec4_cls = wp.vec(4, dtype)
+
+    v = vec4_cls(1, 2, 3, 4)
+    test.assertEqual(v[0], make_scalar(1))
+    test.assertEqual(v.x, make_scalar(1))
+    test.assertEqual(v.y, make_scalar(2))
+    test.assertEqual(v.z, make_scalar(3))
+    test.assertEqual(v.w, make_scalar(4))
+    test.assertSequenceEqual(v[0:2], make_vec(1, 2))
+    test.assertSequenceEqual(v, make_vec(1, 2, 3, 4))
+
+    v[0] = -1
+    test.assertEqual(v[0], make_scalar(-1))
+    test.assertEqual(v.x, make_scalar(-1))
+    test.assertEqual(v.y, make_scalar(2))
+    test.assertEqual(v.z, make_scalar(3))
+    test.assertEqual(v.w, make_scalar(4))
+    test.assertSequenceEqual(v[0:2], make_vec(-1, 2))
+    test.assertSequenceEqual(v, make_vec(-1, 2, 3, 4))
+
+    v[1:3] = (-2, -3)
+    test.assertEqual(v[0], make_scalar(-1))
+    test.assertEqual(v.x, make_scalar(-1))
+    test.assertEqual(v.y, make_scalar(-2))
+    test.assertEqual(v.z, make_scalar(-3))
+    test.assertEqual(v.w, make_scalar(4))
+    test.assertSequenceEqual(v[0:2], make_vec(-1, -2))
+    test.assertSequenceEqual(v, make_vec(-1, -2, -3, 4))
+
+    v.x = 1
+    test.assertEqual(v[0], make_scalar(1))
+    test.assertEqual(v.x, make_scalar(1))
+    test.assertEqual(v.y, make_scalar(-2))
+    test.assertEqual(v.z, make_scalar(-3))
+    test.assertEqual(v.w, make_scalar(4))
+    test.assertSequenceEqual(v[0:2], make_vec(1, -2))
+    test.assertSequenceEqual(v, make_vec(1, -2, -3, 4))
+
+    v = vec3_cls(2, 4, 6)
+    test.assertSequenceEqual(+v, make_vec(2, 4, 6))
+    test.assertSequenceEqual(-v, make_vec(-2, -4, -6))
+    test.assertSequenceEqual(v + vec3_cls(1, 1, 1), make_vec(3, 5, 7))
+    test.assertSequenceEqual(v - vec3_cls(1, 1, 1), make_vec(1, 3, 5))
+    test.assertSequenceEqual(v * dtype(2), make_vec(4, 8, 12))
+    test.assertSequenceEqual(dtype(2) * v, make_vec(4, 8, 12))
+    test.assertSequenceEqual(v / dtype(2), make_vec(1, 2, 3))
+    test.assertSequenceEqual(dtype(12) / v, make_vec(6, 3, 2))
+
+
+def test_matrix(test, device, dtype):
+    def make_scalar(x):
+        # Cast to the correct integer type to simulate wrapping.
+        if dtype in wp.types.int_types:
+            return dtype._type_(x).value
+
+        return x
+
+    def make_vec(*args):
+        if dtype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(dtype._type_(x).value for x in args)
+
+        return args
+
+    def make_mat(*args):
+        if dtype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(tuple(dtype._type_(x).value for x in row) for row in args)
+
+        return args
+
+    mat22_cls = wp.mat((2, 2), dtype)
+    mat33_cls = wp.mat((3, 3), dtype)
+
+    m = mat33_cls(((1, 2, 3), (4, 5, 6), (7, 8, 9)))
+    test.assertEqual(m[0][0], make_scalar(1))
+    test.assertEqual(m[0][1], make_scalar(2))
+    test.assertEqual(m[0][2], make_scalar(3))
+    test.assertEqual(m[1][0], make_scalar(4))
+    test.assertEqual(m[1][1], make_scalar(5))
+    test.assertEqual(m[1][2], make_scalar(6))
+    test.assertEqual(m[2][0], make_scalar(7))
+    test.assertEqual(m[2][1], make_scalar(8))
+    test.assertEqual(m[2][2], make_scalar(9))
+    test.assertEqual(m[0, 0], make_scalar(1))
+    test.assertEqual(m[0, 1], make_scalar(2))
+    test.assertEqual(m[0, 2], make_scalar(3))
+    test.assertEqual(m[1, 0], make_scalar(4))
+    test.assertEqual(m[1, 1], make_scalar(5))
+    test.assertEqual(m[1, 2], make_scalar(6))
+    test.assertEqual(m[2, 0], make_scalar(7))
+    test.assertEqual(m[2, 1], make_scalar(8))
+    test.assertEqual(m[2, 2], make_scalar(9))
+    test.assertSequenceEqual(m[0], make_vec(1, 2, 3))
+    test.assertSequenceEqual(m[1], make_vec(4, 5, 6))
+    test.assertSequenceEqual(m[2], make_vec(7, 8, 9))
+    test.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
+    test.assertSequenceEqual(m[1][0:2], make_vec(4, 5))
+    test.assertSequenceEqual(m[2][0:3], make_vec(7, 8, 9))
+    # test.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
+    # test.assertSequenceEqual(m[1, 0:2], make_vec(4, 5))
+    # test.assertSequenceEqual(m[2, 0:3], make_vec(7, 8, 9))
+    test.assertSequenceEqual(m, make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
+
+    m[1, 0] = -4
+    test.assertEqual(m[0][0], make_scalar(1))
+    test.assertEqual(m[0][1], make_scalar(2))
+    test.assertEqual(m[0][2], make_scalar(3))
+    test.assertEqual(m[1][0], make_scalar(-4))
+    test.assertEqual(m[1][1], make_scalar(5))
+    test.assertEqual(m[1][2], make_scalar(6))
+    test.assertEqual(m[2][0], make_scalar(7))
+    test.assertEqual(m[2][1], make_scalar(8))
+    test.assertEqual(m[2][2], make_scalar(9))
+    test.assertEqual(m[0, 0], make_scalar(1))
+    test.assertEqual(m[0, 1], make_scalar(2))
+    test.assertEqual(m[0, 2], make_scalar(3))
+    test.assertEqual(m[1, 0], make_scalar(-4))
+    test.assertEqual(m[1, 1], make_scalar(5))
+    test.assertEqual(m[1, 2], make_scalar(6))
+    test.assertEqual(m[2, 0], make_scalar(7))
+    test.assertEqual(m[2, 1], make_scalar(8))
+    test.assertEqual(m[2, 2], make_scalar(9))
+    test.assertSequenceEqual(m[0], make_vec(1, 2, 3))
+    test.assertSequenceEqual(m[1], make_vec(-4, 5, 6))
+    test.assertSequenceEqual(m[2], make_vec(7, 8, 9))
+    test.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
+    test.assertSequenceEqual(m[1][0:2], make_vec(-4, 5))
+    test.assertSequenceEqual(m[2][0:3], make_vec(7, 8, 9))
+    # test.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
+    # test.assertSequenceEqual(m[1, 0:2], make_vec(-4, 5))
+    # test.assertSequenceEqual(m[2, 0:3], make_vec(7, 8, 9))
+    test.assertSequenceEqual(m, make_mat((1, 2, 3), (-4, 5, 6), (7, 8, 9)))
+
+    m[2] = (-7, 8, -9)
+    test.assertEqual(m[0][0], make_scalar(1))
+    test.assertEqual(m[0][1], make_scalar(2))
+    test.assertEqual(m[0][2], make_scalar(3))
+    test.assertEqual(m[1][0], make_scalar(-4))
+    test.assertEqual(m[1][1], make_scalar(5))
+    test.assertEqual(m[1][2], make_scalar(6))
+    test.assertEqual(m[2][0], make_scalar(-7))
+    test.assertEqual(m[2][1], make_scalar(8))
+    test.assertEqual(m[2][2], make_scalar(-9))
+    test.assertEqual(m[0, 0], make_scalar(1))
+    test.assertEqual(m[0, 1], make_scalar(2))
+    test.assertEqual(m[0, 2], make_scalar(3))
+    test.assertEqual(m[1, 0], make_scalar(-4))
+    test.assertEqual(m[1, 1], make_scalar(5))
+    test.assertEqual(m[1, 2], make_scalar(6))
+    test.assertEqual(m[2, 0], make_scalar(-7))
+    test.assertEqual(m[2, 1], make_scalar(8))
+    test.assertEqual(m[2, 2], make_scalar(-9))
+    test.assertSequenceEqual(m[0], make_vec(1, 2, 3))
+    test.assertSequenceEqual(m[1], make_vec(-4, 5, 6))
+    test.assertSequenceEqual(m[2], make_vec(-7, 8, -9))
+    test.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
+    test.assertSequenceEqual(m[1][0:2], make_vec(-4, 5))
+    test.assertSequenceEqual(m[2][0:3], make_vec(-7, 8, -9))
+    # test.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
+    # test.assertSequenceEqual(m[1, 0:2], make_vec(-4, 5))
+    # test.assertSequenceEqual(m[2, 0:3], make_vec(-7, 8, -9))
+    test.assertSequenceEqual(m, make_mat((1, 2, 3), (-4, 5, 6), (-7, 8, -9)))
+
+    m = mat22_cls(2, 4, 6, 8)
+    test.assertSequenceEqual(+m, make_mat((2, 4), (6, 8)))
+    test.assertSequenceEqual(-m, make_mat((-2, -4), (-6, -8)))
+    test.assertSequenceEqual(m + mat22_cls(1, 1, 1, 1), make_mat((3, 5), (7, 9)))
+    test.assertSequenceEqual(m - mat22_cls(1, 1, 1, 1), make_mat((1, 3), (5, 7)))
+    test.assertSequenceEqual(m * dtype(2), make_mat((4, 8), (12, 16)))
+    test.assertSequenceEqual(dtype(2) * m, make_mat((4, 8), (12, 16)))
+    test.assertSequenceEqual(m / dtype(2), make_mat((1, 2), (3, 4)))
+    test.assertSequenceEqual(dtype(24) / m, make_mat((12, 6), (4, 3)))
 
 
 def register(parent):
     devices = get_test_devices()
 
-    class TestUtils(parent):
+    class TestTypes(parent):
         pass
 
-    add_function_test(TestUtils, "test_constant", test_constant)
-    add_function_test(TestUtils, "test_constant_error_invalid_type", test_constant_error_invalid_type)
-    add_function_test(TestUtils, "test_vector", test_vector)
-    return TestUtils
+    add_function_test(TestTypes, "test_constant", test_constant)
+    add_function_test(TestTypes, "test_constant_error_invalid_type", test_constant_error_invalid_type)
+
+    for dtype in tuple(wp.types.scalar_types) + (int, float):
+        add_function_test(TestTypes, f"test_vector_{dtype.__name__}", test_vector, devices=None, dtype=dtype)
+
+    for dtype in tuple(wp.types.float_types) + (float,):
+        add_function_test(TestTypes, f"test_matrix_{dtype.__name__}", test_matrix, devices=None, dtype=dtype)
+
+    return TestTypes
 
 
 if __name__ == "__main__":
diff --git a/warp/types.py b/warp/types.py
index 148109422..05f931548 100644
--- a/warp/types.py
+++ b/warp/types.py
@@ -177,9 +177,15 @@ def __setattr__(self, name, value):
         def __add__(self, y):
             return warp.add(self, y)
 
+        def __radd__(self, y):
+            return warp.add(y, self)
+
         def __sub__(self, y):
             return warp.sub(self, y)
 
+        def __rsub__(self, y):
+            return warp.sub(y, self)
+
         def __mul__(self, y):
             return warp.mul(self, y)
 
@@ -287,9 +293,15 @@ def __init__(self, *args):
         def __add__(self, y):
             return warp.add(self, y)
 
+        def __radd__(self, y):
+            return warp.add(y, self)
+
         def __sub__(self, y):
             return warp.sub(self, y)
 
+        def __rsub__(self, y):
+            return warp.sub(y, self)
+
         def __mul__(self, y):
             return warp.mul(self, y)
 

From a69872f79cd269eaac08eab358d16b3f7c565f5c Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Wed, 15 Nov 2023 15:46:01 +1300
Subject: [PATCH 38/50] Increase coverage of vector and matrix types

---
 warp/tests/test_types.py | 147 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 147 insertions(+)

diff --git a/warp/tests/test_types.py b/warp/tests/test_types.py
index 9d140367e..450584aa1 100644
--- a/warp/tests/test_types.py
+++ b/warp/tests/test_types.py
@@ -105,6 +105,50 @@ def make_vec(*args):
     test.assertSequenceEqual(v / dtype(2), make_vec(1, 2, 3))
     test.assertSequenceEqual(dtype(12) / v, make_vec(6, 3, 2))
 
+    test.assertTrue(v != vec3_cls(1, 2, 3))
+    test.assertEqual(str(v), "[{}]".format(", ".join(str(x) for x in v)))
+
+    # Check added purely for coverage reasons but is this really a desired
+    # behaviour? Not allowing to define new attributes using systems like
+    # `__slots__` could help improving memory usage.
+    v.foo = 123
+    test.assertEqual(v.foo, 123)
+
+
+def test_vector_error_invalid_arg_count(test, device):
+    with test.assertRaisesRegex(
+        ValueError,
+        r"Invalid number of arguments in vector constructor, expected 3 elements, got 2$",
+    ):
+        wp.vec3(1, 2)
+
+
+def test_vector_error_invalid_ptr(test, device):
+    with test.assertRaisesRegex(
+        RuntimeError,
+        r"NULL pointer exception",
+    ):
+        wp.vec3.from_ptr(0)
+
+
+def test_vector_error_invalid_get_item_key(test, device):
+    v = wp.vec3(1, 2, 3)
+
+    with test.assertRaisesRegex(
+        KeyError,
+        r"Invalid key None, expected int or slice",
+    ):
+        v[None]
+
+
+def test_vector_error_invalid_set_item_key(test, device):
+    v = wp.vec3(1, 2, 3)
+    with test.assertRaisesRegex(
+        KeyError,
+        r"Invalid key None, expected int or slice",
+    ):
+        v[None] = 0
+
 
 def test_matrix(test, device, dtype):
     def make_scalar(x):
@@ -130,6 +174,7 @@ def make_mat(*args):
 
     mat22_cls = wp.mat((2, 2), dtype)
     mat33_cls = wp.mat((3, 3), dtype)
+    vec2_cls = wp.vec(2, dtype)
 
     m = mat33_cls(((1, 2, 3), (4, 5, 6), (7, 8, 9)))
     test.assertEqual(m[0][0], make_scalar(1))
@@ -231,6 +276,94 @@ def make_mat(*args):
     test.assertSequenceEqual(m / dtype(2), make_mat((1, 2), (3, 4)))
     test.assertSequenceEqual(dtype(24) / m, make_mat((12, 6), (4, 3)))
 
+    test.assertSequenceEqual(m * vec2_cls(1, 2), make_vec(10, 22))
+    test.assertSequenceEqual(m @ vec2_cls(1, 2), make_vec(10, 22))
+    test.assertSequenceEqual(vec2_cls(1, 2) * m, make_vec(14, 20))
+    test.assertSequenceEqual(vec2_cls(1, 2) @ m, make_vec(14, 20))
+
+    test.assertTrue(m != mat22_cls(1, 2, 3, 4))
+    test.assertEqual(
+        str(m), "[{}]".format(",\n ".join("[{}]".format(", ".join(str(y) for y in m[x])) for x in range(m._shape_[0])))
+    )
+
+    # Check added purely for coverage reasons but is this really a desired
+    # behaviour? Not allowing to define new attributes using systems like
+    # `__slots__` could help improving memory usage.
+    m.foo = 123
+    test.assertEqual(m.foo, 123)
+
+
+def test_matrix_error_invalid_arg_count(test, device):
+    with test.assertRaisesRegex(
+        ValueError,
+        r"Invalid number of arguments in matrix constructor, expected 4 elements, got 3$",
+    ):
+        wp.mat22(1, 2, 3)
+
+
+def test_matrix_error_invalid_row_count(test, device):
+    with test.assertRaisesRegex(
+        TypeError,
+        r"Invalid argument in matrix constructor, expected row of length 2, got \(1, 2, 3\)$",
+    ):
+        wp.mat22((1, 2, 3), (3, 4, 5))
+
+
+def test_matrix_error_invalid_ptr(test, device):
+    with test.assertRaisesRegex(
+        RuntimeError,
+        r"NULL pointer exception",
+    ):
+        wp.mat22.from_ptr(0)
+
+
+def test_matrix_error_invalid_set_row_index(test, device):
+    m = wp.mat22(1, 2, 3, 4)
+    with test.assertRaisesRegex(
+        IndexError,
+        r"Invalid row index$",
+    ):
+        m.set_row(2, (0, 0))
+
+
+def test_matrix_error_invalid_get_item_key(test, device):
+    m = wp.mat22(1, 2, 3, 4)
+
+    with test.assertRaisesRegex(
+        KeyError,
+        r"Invalid key None, expected int or pair of ints",
+    ):
+        m[None]
+
+
+def test_matrix_error_invalid_get_item_key_length(test, device):
+    m = wp.mat22(1, 2, 3, 4)
+
+    with test.assertRaisesRegex(
+        KeyError,
+        r"Invalid key, expected one or two indices, got 3",
+    ):
+        m[0, 1, 2]
+
+
+def test_matrix_error_invalid_set_item_key(test, device):
+    m = wp.mat22(1, 2, 3, 4)
+    with test.assertRaisesRegex(
+        KeyError,
+        r"Invalid key None, expected int or pair of ints",
+    ):
+        m[None] = 0
+
+
+def test_matrix_error_invalid_set_item_key_length(test, device):
+    m = wp.mat22(1, 2, 3, 4)
+
+    with test.assertRaisesRegex(
+        KeyError,
+        r"Invalid key, expected one or two indices, got 3",
+    ):
+        m[0, 1, 2] = (0, 0)
+
 
 def register(parent):
     devices = get_test_devices()
@@ -244,9 +377,23 @@ class TestTypes(parent):
     for dtype in tuple(wp.types.scalar_types) + (int, float):
         add_function_test(TestTypes, f"test_vector_{dtype.__name__}", test_vector, devices=None, dtype=dtype)
 
+    add_function_test(TestTypes, "test_vector_error_invalid_arg_count", test_vector_error_invalid_arg_count)
+    add_function_test(TestTypes, "test_vector_error_invalid_ptr", test_vector_error_invalid_ptr)
+    add_function_test(TestTypes, "test_vector_error_invalid_get_item_key", test_vector_error_invalid_get_item_key)
+    add_function_test(TestTypes, "test_vector_error_invalid_set_item_key", test_vector_error_invalid_set_item_key)
+
     for dtype in tuple(wp.types.float_types) + (float,):
         add_function_test(TestTypes, f"test_matrix_{dtype.__name__}", test_matrix, devices=None, dtype=dtype)
 
+    add_function_test(TestTypes, "test_matrix_error_invalid_arg_count", test_matrix_error_invalid_arg_count)
+    add_function_test(TestTypes, "test_matrix_error_invalid_row_count", test_matrix_error_invalid_row_count)
+    add_function_test(TestTypes, "test_matrix_error_invalid_ptr", test_matrix_error_invalid_ptr)
+    add_function_test(TestTypes, "test_matrix_error_invalid_set_row_index", test_matrix_error_invalid_set_row_index)
+    add_function_test(TestTypes, "test_matrix_error_invalid_get_item_key", test_matrix_error_invalid_get_item_key)
+    add_function_test(TestTypes, "test_matrix_error_invalid_get_item_key_length", test_matrix_error_invalid_get_item_key_length)
+    add_function_test(TestTypes, "test_matrix_error_invalid_set_item_key", test_matrix_error_invalid_set_item_key)
+    add_function_test(TestTypes, "test_matrix_error_invalid_set_item_key_length", test_matrix_error_invalid_set_item_key_length)
+
     return TestTypes
 
 

From 5c332973231fe9d243b92807c37b5bdc2e590f4c Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Mon, 20 Nov 2023 15:11:06 +1300
Subject: [PATCH 39/50] Address code using incorrect types

---
 examples/env/env_ant.py                       |   2 +-
 examples/env/env_cartpole.py                  |   2 +-
 examples/env/env_humanoid.py                  |   2 +-
 examples/example_jacobian_ik.py               |   2 +-
 examples/example_mesh_intersect.py            |  12 +-
 examples/example_nvdb.py                      |   6 +-
 examples/example_render_opengl.py             |   4 +-
 examples/example_sim_cartpole.py              |   2 +-
 examples/example_sim_cloth.py                 |  18 +--
 examples/example_sim_fk_grad.py               |   4 +-
 examples/example_sim_fk_grad_torch.py         |   6 +-
 examples/example_sim_grad_bounce.py           |   4 +-
 examples/example_sim_grad_cloth.py            |   6 +-
 examples/example_sim_granular.py              |   4 +-
 .../example_sim_granular_collision_sdf.py     |  26 ++--
 examples/example_sim_neo_hookean.py           |   6 +-
 examples/example_sim_particle_chain.py        |   4 +-
 examples/example_sim_quadruped.py             |   2 +-
 examples/example_sim_rigid_chain.py           |   2 +-
 examples/example_sim_rigid_contact.py         |  12 +-
 examples/example_sim_rigid_fem.py             |   4 +-
 examples/example_sim_rigid_gyroscopic.py      |   4 +-
 examples/example_sim_rigid_kinematics.py      |   2 +-
 examples/example_sim_trajopt.py               |   2 +-
 examples/fem/example_apic_fluid.py            |   8 +-
 warp/sim/import_mjcf.py                       |  20 ++-
 warp/sim/import_urdf.py                       |  24 ++--
 warp/sim/inertia.py                           |  35 +++--
 warp/sim/model.py                             | 134 +++++++++---------
 warp/sim/render.py                            |   2 +-
 warp/tests/test_func.py                       |   4 +-
 warp/tests/test_model.py                      |   8 +-
 warp/types.py                                 |   4 +-
 33 files changed, 192 insertions(+), 185 deletions(-)

diff --git a/examples/env/env_ant.py b/examples/env/env_ant.py
index 1937d79ea..2491364c0 100644
--- a/examples/env/env_ant.py
+++ b/examples/env/env_ant.py
@@ -54,7 +54,7 @@ def create_articulation(self, builder):
             up_axis="y",
         )
         builder.joint_q[7:] = [0.0, 1.0, 0.0, -1.0, 0.0, -1.0, 0.0, 1.0]
-        builder.joint_q[:7] = [0.0, 0.7, 0.0, *wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.5)]
+        builder.joint_q[:7] = [0.0, 0.7, 0.0, *wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5)]
 
 
 if __name__ == "__main__":
diff --git a/examples/env/env_cartpole.py b/examples/env/env_cartpole.py
index ba51f1cf3..e96c67f43 100644
--- a/examples/env/env_cartpole.py
+++ b/examples/env/env_cartpole.py
@@ -39,7 +39,7 @@ def create_articulation(self, builder):
         wp.sim.parse_urdf(
             os.path.join(os.path.dirname(__file__), "../assets/cartpole.urdf"),
             builder,
-            xform=wp.transform((0.0, 0.0, 0.0), wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.5)),
+            xform=wp.transform((0.0, 0.0, 0.0), wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5)),
             floating=False,
             armature=0.1,
             stiffness=0.0,
diff --git a/examples/env/env_humanoid.py b/examples/env/env_humanoid.py
index 7e052881b..59ac91e51 100644
--- a/examples/env/env_humanoid.py
+++ b/examples/env/env_humanoid.py
@@ -58,7 +58,7 @@ def create_articulation(self, builder):
             up_axis="y",
         )
 
-        builder.joint_q[:7] = [0.0, 1.7, 0.0, *wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.5)]
+        builder.joint_q[:7] = [0.0, 1.7, 0.0, *wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5)]
 
 
 if __name__ == "__main__":
diff --git a/examples/example_jacobian_ik.py b/examples/example_jacobian_ik.py
index 81959f66c..48cb4a8ba 100644
--- a/examples/example_jacobian_ik.py
+++ b/examples/example_jacobian_ik.py
@@ -87,7 +87,7 @@ def __init__(self, stage=None, enable_rendering=True, num_envs=1, device=None):
             builder.add_builder(
                 articulation_builder,
                 xform=wp.transform(
-                    np.array((i * 2.0, 4.0, 0.0)), wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.5)
+                    wp.vec3(i * 2.0, 4.0, 0.0), wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5)
                 ),
             )
             self.target_origin.append((i * 2.0, 4.0, 0.0))
diff --git a/examples/example_mesh_intersect.py b/examples/example_mesh_intersect.py
index fce3e83a2..e5e1b6de9 100644
--- a/examples/example_mesh_intersect.py
+++ b/examples/example_mesh_intersect.py
@@ -83,6 +83,8 @@ def intersect(
 
 class Example:
     def __init__(self, stage):
+        rng = np.random.default_rng()
+
         self.query_count = 1024
         self.has_queried = False
 
@@ -102,10 +104,10 @@ def __init__(self, stage):
 
         for i in range(self.query_count):
             # random offset
-            p = (np.random.rand(3) * 0.5 - 0.5) * 5.0
+            p = wp.vec3(rng.random(3) * 0.5 - 0.5) * 5.0
 
             # random orientation
-            axis = wp.normalize(np.random.rand(3) * 0.5 - 0.5)
+            axis = wp.normalize(wp.vec3(rng.random(3) * 0.5 - 0.5))
             angle = float(np.random.rand(1)[0])
 
             q = wp.quat_from_axis_angle(wp.normalize(axis), angle)
@@ -143,14 +145,14 @@ def render(self):
                     os.path.join(os.path.dirname(__file__), self.path_0),
                     pos=wp.vec3(xform.p[0] + offset, xform.p[1], xform.p[2]),
                     rot=xform.q,
-                    scale=(1.0, 1.0, 1.0),
+                    scale=wp.vec3(1.0, 1.0, 1.0),
                 )
                 self.renderer.render_ref(
                     f"mesh_{i}_1",
                     os.path.join(os.path.dirname(__file__), self.path_1),
-                    pos=(offset, 0.0, 0.0),
+                    pos=wp.vec3(offset, 0.0, 0.0),
                     rot=wp.quat_identity(),
-                    scale=(1.0, 1.0, 1.0),
+                    scale=wp.vec3(1.0, 1.0, 1.0),
                 )
 
                 # if pair intersects then draw a small box above the pair
diff --git a/examples/example_nvdb.py b/examples/example_nvdb.py
index 53448eaa4..5e3a32b9a 100644
--- a/examples/example_nvdb.py
+++ b/examples/example_nvdb.py
@@ -143,9 +143,9 @@ def render(self, is_live=False):
             self.renderer.render_ref(
                 name="collision",
                 path=os.path.join(os.path.dirname(__file__), "assets/rocks.usd"),
-                pos=(0.0, 0.0, 0.0),
-                rot=wp.quat_from_axis_angle((1.0, 0.0, 0.0), math.pi),
-                scale=(1.0, 1.0, 1.0),
+                pos=wp.vec3(0.0, 0.0, 0.0),
+                rot=wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), math.pi),
+                scale=wp.vec3(1.0, 1.0, 1.0),
             )
             self.renderer.render_points(name="points", points=self.positions.numpy(), radius=self.sim_margin)
 
diff --git a/examples/example_render_opengl.py b/examples/example_render_opengl.py
index 8bf2ffa38..3c5d2e23a 100644
--- a/examples/example_render_opengl.py
+++ b/examples/example_render_opengl.py
@@ -104,14 +104,14 @@
     renderer.render_cylinder(
         "cylinder",
         [3.2, 1.0, np.sin(time + 0.5)],
-        np.array(wp.quat_from_axis_angle((1.0, 0.0, 0.0), np.sin(time + 0.5))),
+        np.array(wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), wp.sin(time + 0.5))),
         radius=0.5,
         half_height=0.8,
     )
     renderer.render_cone(
         "cone",
         [-1.2, 1.0, 0.0],
-        np.array(wp.quat_from_axis_angle((0.707, 0.707, 0.0), time)),
+        np.array(wp.quat_from_axis_angle(wp.vec3(0.707, 0.707, 0.0), time)),
         radius=0.5,
         half_height=0.8,
     )
diff --git a/examples/example_sim_cartpole.py b/examples/example_sim_cartpole.py
index 9c80337b6..5c2724d8f 100644
--- a/examples/example_sim_cartpole.py
+++ b/examples/example_sim_cartpole.py
@@ -37,7 +37,7 @@ def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=T
         wp.sim.parse_urdf(
             os.path.join(os.path.dirname(__file__), "assets/cartpole.urdf"),
             articulation_builder,
-            xform=wp.transform(np.zeros(3), wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.5)),
+            xform=wp.transform(wp.vec3(), wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5)),
             floating=False,
             density=100,
             armature=0.1,
diff --git a/examples/example_sim_cloth.py b/examples/example_sim_cloth.py
index 83caef096..5ec7dfcf7 100644
--- a/examples/example_sim_cloth.py
+++ b/examples/example_sim_cloth.py
@@ -68,9 +68,9 @@ def __init__(self, stage):
 
         if self.integrator_type == IntegratorType.EULER:
             builder.add_cloth_grid(
-                pos=(0.0, 4.0, 0.0),
-                rot=wp.quat_from_axis_angle((1.0, 0.0, 0.0), math.pi * 0.5),
-                vel=(0.0, 0.0, 0.0),
+                pos=wp.vec3(0.0, 4.0, 0.0),
+                rot=wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), math.pi * 0.5),
+                vel=wp.vec3(0.0, 0.0, 0.0),
                 dim_x=self.sim_width,
                 dim_y=self.sim_height,
                 cell_x=0.1,
@@ -83,9 +83,9 @@ def __init__(self, stage):
             )
         else:
             builder.add_cloth_grid(
-                pos=(0.0, 4.0, 0.0),
-                rot=wp.quat_from_axis_angle((1.0, 0.0, 0.0), math.pi * 0.5),
-                vel=(0.0, 0.0, 0.0),
+                pos=wp.vec3(0.0, 4.0, 0.0),
+                rot=wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), math.pi * 0.5),
+                vel=wp.vec3(0.0, 0.0, 0.0),
                 dim_x=self.sim_width,
                 dim_y=self.sim_height,
                 cell_x=0.1,
@@ -109,9 +109,9 @@ def __init__(self, stage):
         builder.add_shape_mesh(
             body=-1,
             mesh=mesh,
-            pos=(1.0, 0.0, 1.0),
-            rot=wp.quat_from_axis_angle((0.0, 1.0, 0.0), math.pi * 0.5),
-            scale=(2.0, 2.0, 2.0),
+            pos=wp.vec3(1.0, 0.0, 1.0),
+            rot=wp.quat_from_axis_angle(wp.vec3(0.0, 1.0, 0.0), math.pi * 0.5),
+            scale=wp.vec3(2.0, 2.0, 2.0),
             ke=1.0e2,
             kd=1.0e2,
             kf=1.0e1,
diff --git a/examples/example_sim_fk_grad.py b/examples/example_sim_fk_grad.py
index ba4c5edaa..92d711147 100644
--- a/examples/example_sim_fk_grad.py
+++ b/examples/example_sim_fk_grad.py
@@ -84,12 +84,12 @@ def __init__(self, stage, device=None, verbose=False):
 
             if i == chain_length - 1:
                 # create end effector
-                builder.add_shape_sphere(pos=(0.0, 0.0, 0.0), radius=0.1, density=10.0, body=b)
+                builder.add_shape_sphere(pos=wp.vec3(0.0, 0.0, 0.0), radius=0.1, density=10.0, body=b)
 
             else:
                 # create shape
                 builder.add_shape_box(
-                    pos=(chain_width * 0.5, 0.0, 0.0), hx=chain_width * 0.5, hy=0.1, hz=0.1, density=10.0, body=b
+                    pos=wp.vec3(chain_width * 0.5, 0.0, 0.0), hx=chain_width * 0.5, hy=0.1, hz=0.1, density=10.0, body=b
                 )
 
         self.device = wp.get_device(device)
diff --git a/examples/example_sim_fk_grad_torch.py b/examples/example_sim_fk_grad_torch.py
index 6fb5b453d..f047c5b7f 100644
--- a/examples/example_sim_fk_grad_torch.py
+++ b/examples/example_sim_fk_grad_torch.py
@@ -95,7 +95,7 @@ def __init__(self, stage, device=None, verbose=False):
             builder.add_joint_revolute(
                 parent=parent,
                 child=b,
-                axis=(0.0, 0.0, 1.0),
+                axis=wp.vec3(0.0, 0.0, 1.0),
                 parent_xform=parent_joint_xform,
                 child_xform=wp.transform_identity(),
                 limit_lower=-np.deg2rad(60.0),
@@ -108,12 +108,12 @@ def __init__(self, stage, device=None, verbose=False):
 
             if i == chain_length - 1:
                 # create end effector
-                builder.add_shape_sphere(pos=(0.0, 0.0, 0.0), radius=0.1, density=10.0, body=b)
+                builder.add_shape_sphere(pos=wp.vec3(0.0, 0.0, 0.0), radius=0.1, density=10.0, body=b)
 
             else:
                 # create shape
                 builder.add_shape_box(
-                    pos=(chain_width * 0.5, 0.0, 0.0), hx=chain_width * 0.5, hy=0.1, hz=0.1, density=10.0, body=b
+                    pos=wp.vec3(chain_width * 0.5, 0.0, 0.0), hx=chain_width * 0.5, hy=0.1, hz=0.1, density=10.0, body=b
                 )
 
         # finalize model
diff --git a/examples/example_sim_grad_bounce.py b/examples/example_sim_grad_bounce.py
index de524a7b0..526f86742 100644
--- a/examples/example_sim_grad_bounce.py
+++ b/examples/example_sim_grad_bounce.py
@@ -71,8 +71,8 @@ def __init__(self, stage=None, enable_rendering=True, profile=False, adapter=Non
         mu = 0.2
 
         builder = wp.sim.ModelBuilder()
-        builder.add_particle(pos=(-0.5, 1.0, 0.0), vel=(5.0, -5.0, 0.0), mass=1.0)
-        builder.add_shape_box(body=-1, pos=(2.0, 1.0, 0.0), hx=0.25, hy=1.0, hz=1.0, ke=ke, kf=kf, kd=kd, mu=mu)
+        builder.add_particle(pos=wp.vec3(-0.5, 1.0, 0.0), vel=wp.vec3(5.0, -5.0, 0.0), mass=1.0)
+        builder.add_shape_box(body=-1, pos=wp.vec3(2.0, 1.0, 0.0), hx=0.25, hy=1.0, hz=1.0, ke=ke, kf=kf, kd=kd, mu=mu)
 
         self.device = wp.get_device(adapter)
         self.profile = profile
diff --git a/examples/example_sim_grad_cloth.py b/examples/example_sim_grad_cloth.py
index b1d1944cd..d0afb6905 100644
--- a/examples/example_sim_grad_cloth.py
+++ b/examples/example_sim_grad_cloth.py
@@ -80,9 +80,9 @@ def __init__(self, stage, profile=False, adapter=None, verbose=False):
         dim_y = 16
 
         builder.add_cloth_grid(
-            pos=(0.0, 0.0, 0.0),
-            vel=(0.1, 0.1, 0.0),
-            rot=wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.25),
+            pos=wp.vec3(0.0, 0.0, 0.0),
+            vel=wp.vec3(0.1, 0.1, 0.0),
+            rot=wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.25),
             dim_x=dim_x,
             dim_y=dim_y,
             cell_x=1.0 / dim_x,
diff --git a/examples/example_sim_granular.py b/examples/example_sim_granular.py
index 3eb29181c..b868e850b 100644
--- a/examples/example_sim_granular.py
+++ b/examples/example_sim_granular.py
@@ -44,9 +44,9 @@ def __init__(self, stage):
             cell_x=self.radius * 2.0,
             cell_y=self.radius * 2.0,
             cell_z=self.radius * 2.0,
-            pos=(0.0, 1.0, 0.0),
+            pos=wp.vec3(0.0, 1.0, 0.0),
             rot=wp.quat_identity(),
-            vel=(5.0, 0.0, 0.0),
+            vel=wp.vec3(5.0, 0.0, 0.0),
             mass=0.1,
             jitter=self.radius * 0.1,
         )
diff --git a/examples/example_sim_granular_collision_sdf.py b/examples/example_sim_granular_collision_sdf.py
index 60fcd2e92..175b1067d 100644
--- a/examples/example_sim_granular_collision_sdf.py
+++ b/examples/example_sim_granular_collision_sdf.py
@@ -48,9 +48,9 @@ def __init__(self, stage):
             cell_x=self.radius * 2.0,
             cell_y=self.radius * 2.0,
             cell_z=self.radius * 2.0,
-            pos=(0.0, 20.0, 0.0),
+            pos=wp.vec3(0.0, 20.0, 0.0),
             rot=wp.quat_identity(),
-            vel=(2.0, 0.0, 0.0),
+            vel=wp.vec3(2.0, 0.0, 0.0),
             mass=0.1,
             jitter=self.radius * 0.1,
         )
@@ -67,9 +67,9 @@ def __init__(self, stage):
             mu=0.5,
             sdf=rock_sdf,
             body=-1,
-            pos=(0.0, 0.0, 0.0),
-            rot=wp.quat_from_axis_angle((1.0, 0.0, 0.0), -0.5 * math.pi),
-            scale=(0.01, 0.01, 0.01),
+            pos=wp.vec3(0.0, 0.0, 0.0),
+            rot=wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -0.5 * math.pi),
+            scale=wp.vec3(0.01, 0.01, 0.01),
         )
 
         mins = np.array([-3.0, -3.0, -3.0])
@@ -89,7 +89,7 @@ def __init__(self, stage):
         sphere_vdb = wp.Volume.load_from_numpy(sphere_sdf_np, mins, voxel_size, rad + 3.0 * voxel_size)
         sphere_sdf = wp.sim.SDF(sphere_vdb)
 
-        self.sphere_pos = (3.0, 15.0, 0.0)
+        self.sphere_pos = wp.vec3(3.0, 15.0, 0.0)
         self.sphere_scale = 1.0
         self.sphere_radius = rad
         builder.add_shape_sdf(
@@ -100,7 +100,7 @@ def __init__(self, stage):
             sdf=sphere_sdf,
             body=-1,
             pos=self.sphere_pos,
-            scale=(self.sphere_scale, self.sphere_scale, self.sphere_scale),
+            scale=wp.vec3(self.sphere_scale, self.sphere_scale, self.sphere_scale),
         )
 
         self.model = builder.finalize()
@@ -136,21 +136,21 @@ def render(self, is_live=False):
 
             self.renderer.begin_frame(time)
 
-            # Note the extra wp.quat_from_axis_angle((1.0, 0.0, 0.0), math.pi) is because .usd is oriented differently from .nvdb
+            # Note the extra wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), math.pi) is because .usd is oriented differently from .nvdb
             self.renderer.render_ref(
                 name="collision",
                 path=os.path.join(os.path.dirname(__file__), "assets/rocks.usd"),
-                pos=(0.0, 0.0, 0.0),
-                rot=wp.quat_from_axis_angle((1.0, 0.0, 0.0), -0.5 * math.pi)
-                * wp.quat_from_axis_angle((1.0, 0.0, 0.0), math.pi),
-                scale=(0.01, 0.01, 0.01),
+                pos=wp.vec3(0.0, 0.0, 0.0),
+                rot=wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -0.5 * math.pi)
+                * wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), math.pi),
+                scale=wp.vec3(0.01, 0.01, 0.01),
             )
 
             self.renderer.render_sphere(
                 name="sphere",
                 pos=self.sphere_pos,
                 radius=self.sphere_scale * self.sphere_radius,
-                rot=(0.0, 0.0, 0.0, 1.0),
+                rot=wp.quat(0.0, 0.0, 0.0, 1.0),
             )
 
             self.renderer.render(self.state_0)
diff --git a/examples/example_sim_neo_hookean.py b/examples/example_sim_neo_hookean.py
index 19cb463cf..99e9506d4 100644
--- a/examples/example_sim_neo_hookean.py
+++ b/examples/example_sim_neo_hookean.py
@@ -80,9 +80,9 @@ def __init__(self, stage):
         center = cell_size * cell_dim * 0.5
 
         builder.add_soft_grid(
-            pos=(-center, 0.0, -center),
+            pos=wp.vec3(-center, 0.0, -center),
             rot=wp.quat_identity(),
-            vel=(0.0, 0.0, 0.0),
+            vel=wp.vec3(0.0, 0.0, 0.0),
             dim_x=cell_dim,
             dim_y=cell_dim,
             dim_z=cell_dim,
@@ -117,7 +117,7 @@ def update(self):
         with wp.ScopedTimer("simulate"):
             xform = wp.transform(
                 (0.0, self.lift_speed * self.sim_time, 0.0),
-                wp.quat_from_axis_angle((0.0, 1.0, 0.0), self.rot_speed * self.sim_time),
+                wp.quat_from_axis_angle(wp.vec3(0.0, 1.0, 0.0), self.rot_speed * self.sim_time),
             )
             wp.launch(
                 kernel=twist_points,
diff --git a/examples/example_sim_particle_chain.py b/examples/example_sim_particle_chain.py
index d93a82c19..712f40d82 100644
--- a/examples/example_sim_particle_chain.py
+++ b/examples/example_sim_particle_chain.py
@@ -38,13 +38,13 @@ def __init__(self, stage):
         builder = wp.sim.ModelBuilder()
 
         # anchor
-        builder.add_particle((0.0, 1.0, 0.0), (0.0, 0.0, 0.0), 0.0)
+        builder.add_particle(wp.vec3(0.0, 1.0, 0.0), wp.vec3(0.0, 0.0, 0.0), 0.0)
 
         # chain
         for i in range(1, 10):
             radius = math.sqrt(i) * 0.2
             mass = math.pi * radius * radius * radius
-            builder.add_particle((i, 1.0, 0.0), (0.0, 0.0, 0.0), mass, radius=radius)
+            builder.add_particle(wp.vec3(i, 1.0, 0.0), wp.vec3(0.0, 0.0, 0.0), mass, radius=radius)
             builder.add_spring(i - 1, i, 1.0e6, 0.0, 0)
 
         self.model = builder.finalize()
diff --git a/examples/example_sim_quadruped.py b/examples/example_sim_quadruped.py
index c982d07cc..f0413f40b 100644
--- a/examples/example_sim_quadruped.py
+++ b/examples/example_sim_quadruped.py
@@ -75,7 +75,7 @@ def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=T
         wp.sim.parse_urdf(
             os.path.join(os.path.dirname(__file__), "assets/quadruped.urdf"),
             articulation_builder,
-            xform=wp.transform([0.0, 0.7, 0.0], wp.quat_from_axis_angle((1.0, 0.0, 0.0), -math.pi * 0.5)),
+            xform=wp.transform([0.0, 0.7, 0.0], wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5)),
             floating=True,
             density=1000,
             armature=0.01,
diff --git a/examples/example_sim_rigid_chain.py b/examples/example_sim_rigid_chain.py
index a33f34c32..4309536ca 100644
--- a/examples/example_sim_rigid_chain.py
+++ b/examples/example_sim_rigid_chain.py
@@ -65,7 +65,7 @@ def __init__(self, stage):
 
                 # create shape
                 builder.add_shape_box(
-                    pos=(self.chain_width * 0.5, 0.0, 0.0),
+                    pos=wp.vec3(self.chain_width * 0.5, 0.0, 0.0),
                     hx=self.chain_width * 0.5,
                     hy=0.1,
                     hz=0.1,
diff --git a/examples/example_sim_rigid_contact.py b/examples/example_sim_rigid_contact.py
index 051fe0346..53450ce12 100644
--- a/examples/example_sim_rigid_contact.py
+++ b/examples/example_sim_rigid_contact.py
@@ -50,7 +50,7 @@ def __init__(self, stage):
             b = builder.add_body(origin=wp.transform((i, 1.0, 0.0), wp.quat_identity()))
 
             builder.add_shape_box(
-                pos=(0.0, 0.0, 0.0),
+                pos=wp.vec3(0.0, 0.0, 0.0),
                 hx=0.5 * self.scale,
                 hy=0.2 * self.scale,
                 hz=0.2 * self.scale,
@@ -65,7 +65,7 @@ def __init__(self, stage):
             b = builder.add_body(origin=wp.transform((i, 1.0, 2.0), wp.quat_identity()))
 
             builder.add_shape_sphere(
-                pos=(0.0, 0.0, 0.0), radius=0.25 * self.scale, body=b, ke=self.ke, kd=self.kd, kf=self.kf
+                pos=wp.vec3(0.0, 0.0, 0.0), radius=0.25 * self.scale, body=b, ke=self.ke, kd=self.kd, kf=self.kf
             )
 
         # capsules
@@ -73,7 +73,7 @@ def __init__(self, stage):
             b = builder.add_body(origin=wp.transform((i, 1.0, 6.0), wp.quat_identity()))
 
             builder.add_shape_capsule(
-                pos=(0.0, 0.0, 0.0),
+                pos=wp.vec3(0.0, 0.0, 0.0),
                 radius=0.25 * self.scale,
                 half_height=self.scale * 0.5,
                 up_axis=0,
@@ -93,15 +93,15 @@ def __init__(self, stage):
             b = builder.add_body(
                 origin=wp.transform(
                     (i * 0.5 * self.scale, 1.0 + i * 1.7 * self.scale, 4.0 + i * 0.5 * self.scale),
-                    wp.quat_from_axis_angle((0.0, 1.0, 0.0), math.pi * 0.1 * i),
+                    wp.quat_from_axis_angle(wp.vec3(0.0, 1.0, 0.0), math.pi * 0.1 * i),
                 )
             )
 
             builder.add_shape_mesh(
                 body=b,
                 mesh=bunny,
-                pos=(0.0, 0.0, 0.0),
-                scale=(self.scale, self.scale, self.scale),
+                pos=wp.vec3(0.0, 0.0, 0.0),
+                scale=wp.vec3(self.scale, self.scale, self.scale),
                 ke=self.ke,
                 kd=self.kd,
                 kf=self.kf,
diff --git a/examples/example_sim_rigid_fem.py b/examples/example_sim_rigid_fem.py
index a62a72e7a..80e34f6c3 100644
--- a/examples/example_sim_rigid_fem.py
+++ b/examples/example_sim_rigid_fem.py
@@ -40,9 +40,9 @@ def __init__(self, stage):
         builder.default_particle_radius = 0.01
 
         builder.add_soft_grid(
-            pos=(0.0, 0.0, 0.0),
+            pos=wp.vec3(0.0, 0.0, 0.0),
             rot=wp.quat_identity(),
-            vel=(0.0, 0.0, 0.0),
+            vel=wp.vec3(0.0, 0.0, 0.0),
             dim_x=20,
             dim_y=10,
             dim_z=10,
diff --git a/examples/example_sim_rigid_gyroscopic.py b/examples/example_sim_rigid_gyroscopic.py
index 89cf18a81..de1c1b61d 100644
--- a/examples/example_sim_rigid_gyroscopic.py
+++ b/examples/example_sim_rigid_gyroscopic.py
@@ -36,7 +36,7 @@ def __init__(self, stage):
 
         # axis shape
         builder.add_shape_box(
-            pos=(0.3 * self.scale, 0.0, 0.0),
+            pos=wp.vec3(0.3 * self.scale, 0.0, 0.0),
             hx=0.25 * self.scale,
             hy=0.1 * self.scale,
             hz=0.1 * self.scale,
@@ -46,7 +46,7 @@ def __init__(self, stage):
 
         # tip shape
         builder.add_shape_box(
-            pos=(0.0, 0.0, 0.0), hx=0.05 * self.scale, hy=0.2 * self.scale, hz=1.0 * self.scale, density=100.0, body=b
+            pos=wp.vec3(0.0, 0.0, 0.0), hx=0.05 * self.scale, hy=0.2 * self.scale, hz=1.0 * self.scale, density=100.0, body=b
         )
 
         # initial spin
diff --git a/examples/example_sim_rigid_kinematics.py b/examples/example_sim_rigid_kinematics.py
index be794e9ac..cdb6c8ca1 100644
--- a/examples/example_sim_rigid_kinematics.py
+++ b/examples/example_sim_rigid_kinematics.py
@@ -59,7 +59,7 @@ def __init__(self, stage, num_envs=1, device=None, verbose=False):
             # base
             builder.joint_q[coord_start : coord_start + 3] = [i * 2.0, 0.70, 0.0]
             builder.joint_q[coord_start + 3 : coord_start + 7] = wp.quat_from_axis_angle(
-                (1.0, 0.0, 0.0), -math.pi * 0.5
+                wp.vec3(1.0, 0.0, 0.0), -math.pi * 0.5
             )
 
             # joints
diff --git a/examples/example_sim_trajopt.py b/examples/example_sim_trajopt.py
index fc01472fd..279fb5f32 100644
--- a/examples/example_sim_trajopt.py
+++ b/examples/example_sim_trajopt.py
@@ -72,7 +72,7 @@ def __init__(self, stage, device=None, verbose=False):
         builder = wp.sim.ModelBuilder(gravity=0.0)
         builder.add_articulation()
         b = builder.add_body(origin=wp.transform())
-        builder.add_shape_box(pos=(0.0, 0.0, 0.0), hx=0.5, hy=0.5, hz=0.5, density=100.0, body=b)
+        builder.add_shape_box(pos=wp.vec3(0.0, 0.0, 0.0), hx=0.5, hy=0.5, hz=0.5, density=100.0, body=b)
 
         # compute reference trajectory
         rad = np.linspace(0.0, np.pi * 2, self.episode_frames)
diff --git a/examples/fem/example_apic_fluid.py b/examples/fem/example_apic_fluid.py
index 211602f3c..84c2535ec 100644
--- a/examples/fem/example_apic_fluid.py
+++ b/examples/fem/example_apic_fluid.py
@@ -169,10 +169,10 @@ def __init__(self, stage, num_frames=1000, res=[32, 64, 16], quiet=False):
         grid_cell_volume = np.prod(grid_cell_size)
 
         PARTICLES_PER_CELL_DIM = 3
-        self.radius = np.max(grid_cell_size) / (2 * PARTICLES_PER_CELL_DIM)
+        self.radius = float(np.max(grid_cell_size) / (2 * PARTICLES_PER_CELL_DIM))
 
         particle_grid_res = np.array(particle_fill_frac * grid_res * PARTICLES_PER_CELL_DIM, dtype=int)
-        particle_grid_offset = self.radius * np.ones(3)
+        particle_grid_offset = wp.vec3(self.radius, self.radius, self.radius)
 
         np.random.seed(0)
         builder = wp.sim.ModelBuilder()
@@ -183,9 +183,9 @@ def __init__(self, stage, num_frames=1000, res=[32, 64, 16], quiet=False):
             cell_x=self.radius * 2.0,
             cell_y=self.radius * 2.0,
             cell_z=self.radius * 2.0,
-            pos=(0.0, 0.0, 0.0) + particle_grid_offset,
+            pos=wp.vec3(0.0, 0.0, 0.0) + particle_grid_offset,
             rot=wp.quat_identity(),
-            vel=(0.0, 0.0, 0.0),
+            vel=wp.vec3(0.0, 0.0, 0.0),
             mass=grid_cell_volume / PARTICLES_PER_CELL_DIM**3,
             jitter=self.radius * 1.0,
             radius_mean=self.radius,
diff --git a/warp/sim/import_mjcf.py b/warp/sim/import_mjcf.py
index 355fa669f..f84017dfe 100644
--- a/warp/sim/import_mjcf.py
+++ b/warp/sim/import_mjcf.py
@@ -150,9 +150,15 @@ def parse_float(attrib, key, default):
 
     def parse_vec(attrib, key, default):
         if key in attrib:
-            return np.fromstring(attrib[key], sep=" ")
+            out = np.fromstring(attrib[key], sep=" ", dtype=np.float32)
         else:
-            return np.array(default)
+            out = np.array(default, dtype=np.float32)
+
+        length = len(out)
+        if length == 1:
+            return wp.vec(len(default), wp.float32)(out[0], out[0], out[0])
+
+        return wp.vec(length, wp.float32)(out)
 
     def parse_orientation(attrib):
         if "quat" in attrib:
@@ -397,19 +403,19 @@ def parse_body(body, parent, incoming_defaults: dict):
                 if "fromto" in geom_attrib:
                     geom_fromto = parse_vec(geom_attrib, "fromto", (0.0, 0.0, 0.0, 1.0, 0.0, 0.0))
 
-                    start = geom_fromto[0:3] * scale
-                    end = geom_fromto[3:6] * scale
+                    start = wp.vec3(geom_fromto[0:3]) * scale
+                    end = wp.vec3(geom_fromto[3:6]) * scale
 
                     # compute rotation to align the Warp capsule (along x-axis), with mjcf fromto direction
                     axis = wp.normalize(end - start)
-                    angle = math.acos(np.dot(axis, (0.0, 1.0, 0.0)))
-                    axis = wp.normalize(np.cross(axis, (0.0, 1.0, 0.0)))
+                    angle = math.acos(wp.dot(axis, wp.vec3(0.0, 1.0, 0.0)))
+                    axis = wp.normalize(wp.cross(axis, wp.vec3(0.0, 1.0, 0.0)))
 
                     geom_pos = (start + end) * 0.5
                     geom_rot = wp.quat_from_axis_angle(axis, -angle)
 
                     geom_radius = geom_size[0]
-                    geom_height = np.linalg.norm(end - start) * 0.5
+                    geom_height = wp.length(end - start) * 0.5
                     geom_up_axis = 1
 
                 else:
diff --git a/warp/sim/import_urdf.py b/warp/sim/import_urdf.py
index b1458b6da..a80786413 100644
--- a/warp/sim/import_urdf.py
+++ b/warp/sim/import_urdf.py
@@ -100,8 +100,8 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
                 size = [float(x) for x in size.split()]
                 builder.add_shape_box(
                     body=link,
-                    pos=tf.p,
-                    rot=tf.q,
+                    pos=wp.vec3(tf.p),
+                    rot=wp.quat(tf.q),
                     hx=size[0] * 0.5 * scale,
                     hy=size[1] * 0.5 * scale,
                     hz=size[2] * 0.5 * scale,
@@ -117,8 +117,8 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
             for sphere in geo.findall("sphere"):
                 builder.add_shape_sphere(
                     body=link,
-                    pos=tf.p,
-                    rot=tf.q,
+                    pos=wp.vec3(tf.p),
+                    rot=wp.quat(tf.q),
                     radius=float(sphere.get("radius") or "1") * scale,
                     density=density,
                     ke=shape_ke,
@@ -132,8 +132,8 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
             for cylinder in geo.findall("cylinder"):
                 builder.add_shape_capsule(
                     body=link,
-                    pos=tf.p,
-                    rot=tf.q,
+                    pos=wp.vec3(tf.p),
+                    rot=wp.quat(tf.q),
                     radius=float(cylinder.get("radius") or "1") * scale,
                     half_height=float(cylinder.get("length") or "1") * 0.5 * scale,
                     density=density,
@@ -184,8 +184,8 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
                         mesh = Mesh(vertices, faces)
                         builder.add_shape_mesh(
                             body=link,
-                            pos=tf.p,
-                            rot=tf.q,
+                            pos=wp.vec3(tf.p),
+                            rot=wp.quat(tf.q),
                             mesh=mesh,
                             density=density,
                             ke=shape_ke,
@@ -252,22 +252,22 @@ def parse_shapes(link, collisions, density, incoming_xform=None):
             I_m[2, 0] = I_m[0, 2]
             I_m[2, 1] = I_m[1, 2]
             rot = wp.quat_to_matrix(inertial_frame.q)
-            I_m = rot @ I_m
+            I_m = rot @ wp.mat33(I_m)
             m = float(inertial.find("mass").get("value") or "0")
             builder.body_mass[link] = m
             builder.body_inv_mass[link] = 1.0 / m
             builder.body_com[link] = com
             builder.body_inertia[link] = I_m
-            builder.body_inv_inertia[link] = np.linalg.inv(I_m)
+            builder.body_inv_inertia[link] = wp.inverse(I_m)
         if m == 0.0 and ensure_nonstatic_links:
             # set the mass to something nonzero to ensure the body is dynamic
             m = static_link_mass
             # cube with side length 0.5
-            I_m = np.eye(3) * m / 12.0 * (0.5 * scale) ** 2 * 2.0
+            I_m = wp.mat33(np.eye(3)) * m / 12.0 * (0.5 * scale) ** 2 * 2.0
             builder.body_mass[link] = m
             builder.body_inv_mass[link] = 1.0 / m
             builder.body_inertia[link] = I_m
-            builder.body_inv_inertia[link] = np.linalg.inv(I_m)
+            builder.body_inv_inertia[link] = wp.inverse(I_m)
 
     end_shape_count = len(builder.shape_geo_type)
 
diff --git a/warp/sim/inertia.py b/warp/sim/inertia.py
index 5be776104..9569c8166 100644
--- a/warp/sim/inertia.py
+++ b/warp/sim/inertia.py
@@ -149,9 +149,9 @@ def compute_sphere_inertia(density: float, r: float) -> tuple:
     m = density * v
     Ia = 2.0 / 5.0 * m * r * r
 
-    I = np.array([[Ia, 0.0, 0.0], [0.0, Ia, 0.0], [0.0, 0.0, Ia]])
+    I = wp.mat33([[Ia, 0.0, 0.0], [0.0, Ia, 0.0], [0.0, 0.0, Ia]])
 
-    return (m, np.zeros(3), I)
+    return (m, wp.vec3(), I)
 
 
 def compute_capsule_inertia(density: float, r: float, h: float) -> tuple:
@@ -177,9 +177,9 @@ def compute_capsule_inertia(density: float, r: float, h: float) -> tuple:
     Ia = mc * (0.25 * r * r + (1.0 / 12.0) * h * h) + ms * (0.4 * r * r + 0.375 * r * h + 0.25 * h * h)
     Ib = (mc * 0.5 + ms * 0.4) * r * r
 
-    I = np.array([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ia]])
+    I = wp.mat33([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ia]])
 
-    return (m, np.zeros(3), I)
+    return (m, wp.vec3(), I)
 
 
 def compute_cylinder_inertia(density: float, r: float, h: float) -> tuple:
@@ -200,9 +200,9 @@ def compute_cylinder_inertia(density: float, r: float, h: float) -> tuple:
     Ia = 1 / 12 * m * (3 * r * r + h * h)
     Ib = 1 / 2 * m * r * r
 
-    I = np.array([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ia]])
+    I = wp.mat33([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ia]])
 
-    return (m, np.zeros(3), I)
+    return (m, wp.vec3(), I)
 
 
 def compute_cone_inertia(density: float, r: float, h: float) -> tuple:
@@ -223,9 +223,9 @@ def compute_cone_inertia(density: float, r: float, h: float) -> tuple:
     Ia = 1 / 20 * m * (3 * r * r + 2 * h * h)
     Ib = 3 / 10 * m * r * r
 
-    I = np.array([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ia]])
+    I = wp.mat33([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ia]])
 
-    return (m, np.zeros(3), I)
+    return (m, wp.vec3(), I)
 
 
 def compute_box_inertia(density: float, w: float, h: float, d: float) -> tuple:
@@ -249,17 +249,16 @@ def compute_box_inertia(density: float, w: float, h: float, d: float) -> tuple:
     Ib = 1.0 / 12.0 * m * (w * w + d * d)
     Ic = 1.0 / 12.0 * m * (w * w + h * h)
 
-    I = np.array([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ic]])
+    I = wp.mat33([[Ia, 0.0, 0.0], [0.0, Ib, 0.0], [0.0, 0.0, Ic]])
 
-    return (m, np.zeros(3), I)
+    return (m, wp.vec3(), I)
 
 
 def compute_mesh_inertia(
     density: float, vertices: list, indices: list, is_solid: bool = True, thickness: Union[List[float], float] = 0.001
 ) -> tuple:
     """Computes mass, center of mass, 3x3 inertia matrix, and volume for a mesh."""
-    com = np.mean(vertices, 0)
-    com_warp = wp.vec3(com[0], com[1], com[2])
+    com = wp.vec3(np.mean(vertices, 0))
 
     num_tris = len(indices) // 3
 
@@ -279,7 +278,7 @@ def compute_mesh_inertia(
             kernel=compute_solid_mesh_inertia,
             dim=num_tris,
             inputs=[
-                com_warp,
+                com,
                 weight,
                 wp.array(indices, dtype=int),
                 wp.array(vertices, dtype=wp.vec3),
@@ -294,7 +293,7 @@ def compute_mesh_inertia(
             kernel=compute_hollow_mesh_inertia,
             dim=num_tris,
             inputs=[
-                com_warp,
+                com,
                 weight,
                 wp.array(indices, dtype=int),
                 wp.array(vertices, dtype=wp.vec3),
@@ -304,14 +303,14 @@ def compute_mesh_inertia(
         )
 
     # Extract mass and inertia and save to class attributes.
-    mass = mass_warp.numpy()[0] * density
-    I = I_warp.numpy()[0] * density
+    mass = float(mass_warp.numpy()[0] * density)
+    I = wp.mat33(*(I_warp.numpy()[0] * density))
     volume = vol_warp.numpy()[0]
     return mass, com, I, volume
 
 
 def transform_inertia(m, I, p, q):
-    R = np.array(wp.quat_to_matrix(q)).reshape(3, 3)
+    R = wp.quat_to_matrix(q)
 
     # Steiner's theorem
-    return R @ I @ R.T + m * (np.dot(p, p) * np.eye(3) - np.outer(p, p))
+    return R @ I @ wp.transpose(R) + m * (wp.dot(p, p) * wp.mat33(np.eye(3)) - wp.outer(p, p))
diff --git a/warp/sim/model.py b/warp/sim/model.py
index 6cd1e2ac8..8c2ee1de4 100644
--- a/warp/sim/model.py
+++ b/warp/sim/model.py
@@ -113,8 +113,8 @@ class JointAxis:
     def __init__(
         self,
         axis,
-        limit_lower=-np.inf,
-        limit_upper=np.inf,
+        limit_lower=-math.inf,
+        limit_upper=math.inf,
         limit_ke=100.0,
         limit_kd=10.0,
         target=None,
@@ -133,7 +133,7 @@ def __init__(
             self.target_kd = axis.target_kd
             self.mode = axis.mode
         else:
-            self.axis = np.array(wp.normalize(np.array(axis, dtype=np.float32)))
+            self.axis = wp.normalize(wp.vec3(axis))
             self.limit_lower = limit_lower
             self.limit_upper = limit_upper
             self.limit_ke = limit_ke
@@ -159,7 +159,7 @@ class SDF:
         mass (float): The total mass of the SDF
         com (Vec3): The center of mass of the SDF
     """
-    def __init__(self, volume=None, I=np.eye(3, dtype=np.float32), mass=1.0, com=np.array((0.0, 0.0, 0.0))):
+    def __init__(self, volume=None, I=wp.mat33(np.eye(3)), mass=1.0, com=wp.vec3()):
         self.volume = volume
         self.I = I
         self.mass = mass
@@ -234,9 +234,9 @@ def __init__(self, vertices: List[Vec3], indices: List[int], compute_inertia=Tru
         if compute_inertia:
             self.mass, self.com, self.I, _ = compute_mesh_inertia(1.0, vertices, indices, is_solid=is_solid)
         else:
-            self.I = np.eye(3, dtype=np.float32)
+            self.I = wp.mat33(np.eye(3))
             self.mass = 1.0
-            self.com = np.array((0.0, 0.0, 0.0))
+            self.com = wp.vec3()
 
     def finalize(self, device=None):
         """
@@ -325,7 +325,7 @@ def flatten(self):
 
 def compute_shape_mass(type, scale, src, density, is_solid, thickness):
     if density == 0.0 or type == GEO_PLANE:  # zero density means fixed
-        return 0.0, np.zeros(3), np.zeros((3, 3))
+        return 0.0, wp.vec3(), wp.mat33()
 
     if type == GEO_SPHERE:
         solid = compute_sphere_inertia(density, scale[0])
@@ -335,7 +335,7 @@ def compute_shape_mass(type, scale, src, density, is_solid, thickness):
             hollow = compute_sphere_inertia(density, scale[0] - thickness)
             return solid[0] - hollow[0], solid[1], solid[2] - hollow[2]
     elif type == GEO_BOX:
-        w, h, d = np.array(scale[:3]) * 2.0
+        w, h, d = scale * 2.0
         solid = compute_box_inertia(density, w, h, d)
         if is_solid:
             return solid
@@ -370,13 +370,12 @@ def compute_shape_mass(type, scale, src, density, is_solid, thickness):
         if src.has_inertia and src.mass > 0.0 and src.is_solid == is_solid:
             m, c, I = src.mass, src.com, src.I
 
-            s = np.array(scale[:3])
-            sx, sy, sz = s
+            sx, sy, sz = scale
 
             mass_ratio = sx * sy * sz * density
             m_new = m * mass_ratio
 
-            c_new = c * s
+            c_new = wp.cw_mul(c, scale)
 
             Ixx = I[0, 0] * (sy**2 + sz**2) / 2 * mass_ratio
             Iyy = I[1, 1] * (sx**2 + sz**2) / 2 * mass_ratio
@@ -385,12 +384,12 @@ def compute_shape_mass(type, scale, src, density, is_solid, thickness):
             Ixz = I[0, 2] * sx * sz * mass_ratio
             Iyz = I[1, 2] * sy * sz * mass_ratio
 
-            I_new = np.array([[Ixx, Ixy, Ixz], [Ixy, Iyy, Iyz], [Ixz, Iyz, Izz]])
+            I_new = wp.mat33([[Ixx, Ixy, Ixz], [Ixy, Iyy, Iyz], [Ixz, Iyz, Izz]])
 
             return m_new, c_new, I_new
         elif type == GEO_MESH:
             # fall back to computing inertia from mesh geometry
-            vertices = np.array(src.vertices) * np.array(scale[:3])
+            vertices = np.array(src.vertices) * np.array(scale)
             m, c, I, vol = compute_mesh_inertia(density, vertices, src.indices, is_solid, thickness)
             return m, c, I
     raise ValueError("Unsupported shape type: {}".format(type))
@@ -1131,8 +1130,8 @@ def __init__(self, up_vector=(0.0, 1.0, 0.0), gravity=-9.80665):
         self.joint_coord_count = 0
         self.joint_axis_total_count = 0
 
-        self.up_vector = np.array(up_vector)
-        self.up_axis = np.argmax(np.abs(up_vector))
+        self.up_vector = wp.vec3(up_vector)
+        self.up_axis = wp.vec3(np.argmax(np.abs(up_vector)))
         self.gravity = gravity
         # indicates whether a ground plane has been created
         self._ground_created = False
@@ -1366,8 +1365,8 @@ def add_body(
         self,
         origin: Transform = wp.transform(),
         armature: float = 0.0,
-        com: Vec3 = np.zeros(3),
-        I_m: Mat33 = np.zeros((3, 3)),
+        com: Vec3 = wp.vec3(),
+        I_m: Mat33 = wp.mat33(),
         m: float = 0.0,
         name: str = None,
     ) -> int:
@@ -1392,7 +1391,7 @@ def add_body(
         body_id = len(self.body_mass)
 
         # body data
-        inertia = I_m + np.eye(3) * armature
+        inertia = I_m + wp.mat33(np.eye(3)) * armature
         self.body_inertia.append(inertia)
         self.body_mass.append(m)
         self.body_com.append(com)
@@ -1402,8 +1401,8 @@ def add_body(
         else:
             self.body_inv_mass.append(0.0)
 
-        if inertia.any():
-            self.body_inv_inertia.append(np.linalg.inv(inertia))
+        if any(x for x in inertia):
+            self.body_inv_inertia.append(wp.inverse(inertia))
         else:
             self.body_inv_inertia.append(inertia)
 
@@ -1459,8 +1458,8 @@ def add_joint(
         else:
             self.joint_parents[child].append(parent)
         self.joint_child.append(child)
-        self.joint_X_p.append([*parent_xform])
-        self.joint_X_c.append([*child_xform])
+        self.joint_X_p.append(wp.transform(parent_xform))
+        self.joint_X_c.append(wp.transform(child_xform))
         self.joint_name.append(name or f"joint {self.joint_count}")
         self.joint_axis_start.append(len(self.joint_axis))
         self.joint_axis_dim.append((len(linear_axes), len(angular_axes)))
@@ -2331,7 +2330,7 @@ def add_shape_plane(
                 angle = np.arcsin(np.linalg.norm(c))
                 axis = c / np.linalg.norm(c)
                 rot = wp.quat_from_axis_angle(axis, angle)
-        scale = (width, length, 0.0)
+        scale = wp.vec3(width, length, 0.0)
 
         return self._add_shape(
             body,
@@ -2390,10 +2389,10 @@ def add_shape_sphere(
 
         return self._add_shape(
             body,
-            pos,
-            rot,
+            wp.vec3(pos),
+            wp.quat(rot),
             GEO_SPHERE,
-            (radius, 0.0, 0.0, 0.0),
+            wp.vec3(radius, 0.0, 0.0),
             None,
             density,
             ke,
@@ -2450,10 +2449,10 @@ def add_shape_box(
 
         return self._add_shape(
             body,
-            pos,
-            rot,
+            wp.vec3(pos),
+            wp.quat(rot),
             GEO_BOX,
-            (hx, hy, hz, 0.0),
+            wp.vec3(hx, hy, hz),
             None,
             density,
             ke,
@@ -2508,19 +2507,19 @@ def add_shape_capsule(
 
         """
 
-        q = rot
+        q = wp.quat(rot)
         sqh = math.sqrt(0.5)
         if up_axis == 0:
-            q = wp.mul(rot, wp.quat(0.0, 0.0, -sqh, sqh))
+            q = wp.mul(q, wp.quat(0.0, 0.0, -sqh, sqh))
         elif up_axis == 2:
-            q = wp.mul(rot, wp.quat(sqh, 0.0, 0.0, sqh))
+            q = wp.mul(q, wp.quat(sqh, 0.0, 0.0, sqh))
 
         return self._add_shape(
             body,
-            pos,
-            q,
+            wp.vec3(pos),
+            wp.quat(q),
             GEO_CAPSULE,
-            (radius, half_height, 0.0, 0.0),
+            wp.vec3(radius, half_height, 0.0),
             None,
             density,
             ke,
@@ -2584,10 +2583,10 @@ def add_shape_cylinder(
 
         return self._add_shape(
             body,
-            pos,
-            q,
+            wp.vec3(pos),
+            wp.quat(q),
             GEO_CYLINDER,
-            (radius, half_height, 0.0, 0.0),
+            wp.vec3(radius, half_height, 0.0),
             None,
             density,
             ke,
@@ -2651,10 +2650,10 @@ def add_shape_cone(
 
         return self._add_shape(
             body,
-            pos,
-            q,
+            wp.vec3(pos),
+            wp.quat(q),
             GEO_CONE,
-            (radius, half_height, 0.0, 0.0),
+            wp.vec3(radius, half_height, 0.0),
             None,
             density,
             ke,
@@ -2670,10 +2669,10 @@ def add_shape_cone(
     def add_shape_mesh(
         self,
         body: int,
-        pos: Vec3 = (0.0, 0.0, 0.0),
-        rot: Quat = (0.0, 0.0, 0.0, 1.0),
+        pos: Vec3 = wp.vec3(0.0, 0.0, 0.0),
+        rot: Quat = wp.quat(0.0, 0.0, 0.0, 1.0),
         mesh: Mesh = None,
-        scale: Vec3 = (1.0, 1.0, 1.0),
+        scale: Vec3 = wp.vec3(1.0, 1.0, 1.0),
         density: float = default_shape_density,
         ke: float = default_shape_ke,
         kd: float = default_shape_kd,
@@ -2712,7 +2711,7 @@ def add_shape_mesh(
             pos,
             rot,
             GEO_MESH,
-            (scale[0], scale[1], scale[2], 0.0),
+            wp.vec3(scale[0], scale[1], scale[2]),
             mesh,
             density,
             ke,
@@ -2765,10 +2764,10 @@ def add_shape_sdf(
         """
         return self._add_shape(
             body,
-            pos,
-            rot,
+            wp.vec3(pos),
+            wp.quat(rot),
             GEO_SDF,
-            (scale[0], scale[1], scale[2], 0.0),
+            wp.vec3(scale[0], scale[1], scale[2]),
             sdf,
             density,
             ke,
@@ -2859,7 +2858,7 @@ def _add_shape(
 
         (m, c, I) = compute_shape_mass(type, scale, src, density, is_solid, thickness)
 
-        self._update_body_mass(body, m, I, np.array(pos) + c, np.array(rot))
+        self._update_body_mass(body, m, I, pos + c, rot)
         return shape
 
     # particles
@@ -2954,9 +2953,9 @@ def add_triangle(
 
         """
         # compute basis for 2D rest pose
-        p = np.array(self.particle_q[i])
-        q = np.array(self.particle_q[j])
-        r = np.array(self.particle_q[k])
+        p = self.particle_q[i]
+        q = self.particle_q[j]
+        r = self.particle_q[k]
 
         qp = q - p
         rp = r - p
@@ -3153,13 +3152,13 @@ def add_edge(
         """
         # compute rest angle
         if rest is None:
-            x1 = np.array(self.particle_q[i])
-            x2 = np.array(self.particle_q[j])
-            x3 = np.array(self.particle_q[k])
-            x4 = np.array(self.particle_q[l])
+            x1 = self.particle_q[i]
+            x2 = self.particle_q[j]
+            x3 = self.particle_q[k]
+            x4 = self.particle_q[l]
 
-            n1 = wp.normalize(np.cross(x3 - x1, x4 - x1))
-            n2 = wp.normalize(np.cross(x4 - x2, x3 - x2))
+            n1 = wp.normalize(wp.cross(x3 - x1, x4 - x1))
+            n2 = wp.normalize(wp.cross(x4 - x2, x3 - x2))
             e = wp.normalize(x4 - x3)
 
             d = np.clip(np.dot(n2, n1), -1.0, 1.0)
@@ -3299,8 +3298,8 @@ def grid_index(x, y, dim_x):
 
         for y in range(0, dim_y + 1):
             for x in range(0, dim_x + 1):
-                g = np.array((x * cell_x, y * cell_y, 0.0))
-                p = np.array(wp.quat_rotate(rot, g)) + pos
+                g = wp.vec3(x * cell_x, y * cell_y, 0.0)
+                p = wp.quat_rotate(rot, g) + pos
                 m = mass
 
                 if x == 0 and fix_left:
@@ -3425,7 +3424,7 @@ def add_cloth_mesh(
 
         # particles
         for v in vertices:
-            p = np.array(wp.quat_rotate(rot, v * scale)) + pos
+            p = wp.quat_rotate(rot, v * scale) + pos
 
             self.add_particle(p, vel, 0.0)
 
@@ -3498,13 +3497,14 @@ def add_particle_grid(
         radius_mean: float = default_particle_radius,
         radius_std: float = 0.0,
     ):
+        rng = np.random.default_rng()
         for z in range(dim_z):
             for y in range(dim_y):
                 for x in range(dim_x):
-                    v = np.array((x * cell_x, y * cell_y, z * cell_z))
+                    v = wp.vec3(x * cell_x, y * cell_y, z * cell_z)
                     m = mass
 
-                    p = np.array(wp.quat_rotate(rot, v)) + pos + np.random.rand(3) * jitter
+                    p = wp.quat_rotate(rot, v) + pos + wp.vec3(rng.random(3) * jitter)
 
                     if radius_std > 0.0:
                         r = radius_mean + np.random.randn() * radius_std
@@ -3570,7 +3570,7 @@ def add_soft_grid(
         for z in range(dim_z + 1):
             for y in range(dim_y + 1):
                 for x in range(dim_x + 1):
-                    v = np.array((x * cell_x, y * cell_y, z * cell_z))
+                    v = wp.vec3(x * cell_x, y * cell_y, z * cell_z)
                     m = mass
 
                     if fix_left and x == 0:
@@ -3585,7 +3585,7 @@ def add_soft_grid(
                     if fix_bottom and y == 0:
                         m = 0.0
 
-                    p = np.array(wp.quat_rotate(rot, v)) + pos
+                    p = wp.quat_rotate(rot, v) + pos
 
                     self.add_particle(p, vel, m)
 
@@ -3752,8 +3752,8 @@ def _update_body_mass(self, i, m, I, p, q):
         else:
             self.body_inv_mass[i] = 0.0
 
-        if new_inertia.any():
-            self.body_inv_inertia[i] = np.linalg.inv(new_inertia)
+        if any(x for x in new_inertia):
+            self.body_inv_inertia[i] = wp.inverse(new_inertia)
         else:
             self.body_inv_inertia[i] = new_inertia
 
diff --git a/warp/sim/render.py b/warp/sim/render.py
index 726060377..020269897 100644
--- a/warp/sim/render.py
+++ b/warp/sim/render.py
@@ -268,7 +268,7 @@ def populate(self, model: warp.sim.Model):
                             q = wp.quat(tf[3:])
                             # compute rotation between axis and y
                             axis = axis / np.linalg.norm(axis)
-                            q = q * wp.quat_between_vectors(wp.vec3(*axis), y_axis)
+                            q = q * wp.quat_between_vectors(wp.vec3(axis), y_axis)
                             name = f"joint_{i}_{a}"
                             self.add_shape_instance(name, shape, body, p, q, scale, color1=color, color2=color)
                             self.instance_count += 1
diff --git a/warp/tests/test_func.py b/warp/tests/test_func.py
index 9acfd4a8a..109d70a15 100644
--- a/warp/tests/test_func.py
+++ b/warp/tests/test_func.py
@@ -89,7 +89,7 @@ def test_native_func_export(test, device):
     q = wp.quat(0.0, 0.0, 0.0, 1.0)
     assert_np_equal(np.array([*q]), np.array([0.0, 0.0, 0.0, 1.0]))
 
-    r = wp.quat_from_axis_angle((1.0, 0.0, 0.0), 2.0)
+    r = wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), 2.0)
     assert_np_equal(np.array([*r]), np.array([0.8414709568023682, 0.0, 0.0, 0.5403022170066833]), tol=1.0e-3)
 
     q = wp.quat(1.0, 2.0, 3.0, 4.0)
@@ -149,7 +149,7 @@ def test_native_func_export(test, device):
         wp.quat(4.0, 5.0, 6.0, 7.0),
     )
     test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t * (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0))
+    test.assertSequenceEqual(t * wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0))
     test.assertSequenceEqual(
         t * wp.transform((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0)), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0)
     )
diff --git a/warp/tests/test_model.py b/warp/tests/test_model.py
index 750efb5e8..f88e96657 100644
--- a/warp/tests/test_model.py
+++ b/warp/tests/test_model.py
@@ -35,8 +35,8 @@ def test_add_triangles(self):
             builder1 = ModelBuilder()
             builder2 = ModelBuilder()
             for pt in pts:
-                builder1.add_particle(pt, [0.0, 0.0, 0.0], 1.0)
-                builder2.add_particle(pt, [0.0, 0.0, 0.0], 1.0)
+                builder1.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
+                builder2.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
 
             # test add_triangle(s) with default arguments:
             areas = builder2.add_triangles(tris[:, 0], tris[:, 1], tris[:, 2])
@@ -85,8 +85,8 @@ def test_add_edges(self):
             builder1 = ModelBuilder()
             builder2 = ModelBuilder()
             for pt in pts:
-                builder1.add_particle(pt, [0.0, 0.0, 0.0], 1.0)
-                builder2.add_particle(pt, [0.0, 0.0, 0.0], 1.0)
+                builder1.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
+                builder2.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
 
             # test defaults:
             for i in range(2):
diff --git a/warp/types.py b/warp/types.py
index 05f931548..e7a50e1c1 100644
--- a/warp/types.py
+++ b/warp/types.py
@@ -582,11 +582,11 @@ def __init__(self, *args, **kwargs):
 
         @property
         def p(self):
-            return self[0:3]
+            return vec3(self[0:3])
 
         @property
         def q(self):
-            return self[3:7]
+            return quat(self[3:7])
 
     return transform_t
 

From 676e69d9570e27f3c19788fce78d1073dfb046ab Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Fri, 15 Dec 2023 10:35:16 +1300
Subject: [PATCH 40/50] Pass compound types to built-ins by reference

---
 warp/context.py       |   33 +-
 warp/native/exports.h | 2354 ++++++++++++++++++++---------------------
 2 files changed, 1197 insertions(+), 1190 deletions(-)

diff --git a/warp/context.py b/warp/context.py
index 1b788d79a..d196b20f3 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -364,12 +364,8 @@ def call_builtin(func: Function, *params) -> Tuple[bool, Any]:
                 return (False, None)
 
             # The argument expects a built-in Warp type like a vector or a matrix.
-            # We need to wrap the given argument into a ctypes' structure
-            # to ensure that it's passed by value to the dll rather than by reference.
-            class Param(ctypes.Structure):
-                _fields_ = [("value", arg_type)]
 
-            c_param = Param()
+            c_param = None
 
             if isinstance(param, ctypes.Array):
                 # The given parameter is also a built-in Warp type, so we only need
@@ -378,14 +374,14 @@ class Param(ctypes.Structure):
                     return (False, None)
 
                 if isinstance(param, arg_type):
-                    c_param.value = param
+                    c_param = param
                 else:
                     # Cast the value to its argument type to make sure that it
                     # can be assigned to the field of the `Param` struct.
                     # This could error otherwise when, for example, the field type
                     # is set to `vec3i` while the value is of type `vector(length=3, dtype=int)`,
                     # even though both types are semantically identical.
-                    c_param.value = arg_type(param)
+                    c_param = arg_type(param)
             else:
                 # Flatten the parameter values into a flat 1-D array.
                 arr = []
@@ -446,18 +442,19 @@ class Param(ctypes.Structure):
                     # Extract the floating-point values.
                     arr = tuple(x.value for x in arr)
 
+                c_param = arg_type()
                 if warp.types.type_is_matrix(arg_type):
                     rows, cols = arg_type._shape_
                     for i in range(rows):
                         idx_start = i * cols
                         idx_end = idx_start + cols
-                        c_param.value[i] = arr[idx_start:idx_end]
+                        c_param[i] = arr[idx_start:idx_end]
                 else:
-                    c_param.value[:] = arr
+                    c_param[:] = arr
 
                 uses_non_warp_array_type = True
 
-            c_params.append(c_param)
+            c_params.append(ctypes.byref(c_param))
         else:
             if issubclass(arg_type, ctypes.Array):
                 return (False, None)
@@ -4232,7 +4229,17 @@ def export_stubs(file):  # pragma: no cover
 
 
 def export_builtins(file: io.TextIOBase):  # pragma: no cover
-    def ctype_str(t):
+    def ctype_arg_str(t):
+        if isinstance(t, int):
+            return "int"
+        elif isinstance(t, float):
+            return "float"
+        elif t in warp.types.vector_types:
+            return f"{t.__name__}&"
+        else:
+            return t.__name__
+
+    def ctype_ret_str(t):
         if isinstance(t, int):
             return "int"
         elif isinstance(t, float):
@@ -4259,7 +4266,7 @@ def ctype_str(t):
             if not simple or f.variadic:
                 continue
 
-            args = ", ".join(f"{ctype_str(v)} {k}" for k, v in f.input_types.items())
+            args = ", ".join(f"{ctype_arg_str(v)} {k}" for k, v in f.input_types.items())
             params = ", ".join(f.input_types.keys())
 
             return_type = ""
@@ -4267,7 +4274,7 @@ def ctype_str(t):
             try:
                 # todo: construct a default value for each of the functions args
                 # so we can generate the return type for overloaded functions
-                return_type = ctype_str(f.value_func(None, None, None))
+                return_type = ctype_ret_str(f.value_func(None, None, None))
             except Exception:
                 continue
 
diff --git a/warp/native/exports.h b/warp/native/exports.h
index fa35feff6..8f7079d5e 100644
--- a/warp/native/exports.h
+++ b/warp/native/exports.h
@@ -13,78 +13,78 @@ WP_API void builtin_min_uint16_uint16(uint16 x, uint16 y, uint16* ret) { *ret =
 WP_API void builtin_min_uint32_uint32(uint32 x, uint32 y, uint32* ret) { *ret = wp::min(x, y); }
 WP_API void builtin_min_uint64_uint64(uint64 x, uint64 y, uint64* ret) { *ret = wp::min(x, y); }
 WP_API void builtin_min_uint8_uint8(uint8 x, uint8 y, uint8* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2h_vec2h(vec2h x, vec2h y, vec2h* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4h_vec4h(vec4h x, vec4h y, vec4h* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2f_vec2f(vec2f x, vec2f y, vec2f* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4f_vec4f(vec4f x, vec4f y, vec4f* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2d_vec2d(vec2d x, vec2d y, vec2d* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4d_vec4d(vec4d x, vec4d y, vec4d* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2s_vec2s(vec2s x, vec2s y, vec2s* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4s_vec4s(vec4s x, vec4s y, vec4s* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2i_vec2i(vec2i x, vec2i y, vec2i* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4i_vec4i(vec4i x, vec4i y, vec4i* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2l_vec2l(vec2l x, vec2l y, vec2l* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4l_vec4l(vec4l x, vec4l y, vec4l* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2b_vec2b(vec2b x, vec2b y, vec2b* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4b_vec4b(vec4b x, vec4b y, vec4b* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2us_vec2us(vec2us x, vec2us y, vec2us* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4us_vec4us(vec4us x, vec4us y, vec4us* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2ui_vec2ui(vec2ui x, vec2ui y, vec2ui* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4ui_vec4ui(vec4ui x, vec4ui y, vec4ui* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2ul_vec2ul(vec2ul x, vec2ul y, vec2ul* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4ul_vec4ul(vec4ul x, vec4ul y, vec4ul* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2ub_vec2ub(vec2ub x, vec2ub y, vec2ub* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec4ub_vec4ub(vec4ub x, vec4ub y, vec4ub* ret) { *ret = wp::min(x, y); }
-WP_API void builtin_min_vec2h(vec2h v, float16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3h(vec3h v, float16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4h(vec4h v, float16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_spatial_vectorh(spatial_vectorh v, float16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2f(vec2f v, float32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3f(vec3f v, float32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4f(vec4f v, float32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_spatial_vectorf(spatial_vectorf v, float32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2d(vec2d v, float64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3d(vec3d v, float64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4d(vec4d v, float64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_spatial_vectord(spatial_vectord v, float64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2s(vec2s v, int16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3s(vec3s v, int16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4s(vec4s v, int16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2i(vec2i v, int32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3i(vec3i v, int32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4i(vec4i v, int32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2l(vec2l v, int64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3l(vec3l v, int64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4l(vec4l v, int64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2b(vec2b v, int8* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3b(vec3b v, int8* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4b(vec4b v, int8* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2us(vec2us v, uint16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3us(vec3us v, uint16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4us(vec4us v, uint16* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2ui(vec2ui v, uint32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3ui(vec3ui v, uint32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4ui(vec4ui v, uint32* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2ul(vec2ul v, uint64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3ul(vec3ul v, uint64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4ul(vec4ul v, uint64* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec2ub(vec2ub v, uint8* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec3ub(vec3ub v, uint8* ret) { *ret = wp::min(v); }
-WP_API void builtin_min_vec4ub(vec4ub v, uint8* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2h_vec2h(vec2h& x, vec2h& y, vec2h* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4h_vec4h(vec4h& x, vec4h& y, vec4h* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2f_vec2f(vec2f& x, vec2f& y, vec2f* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4f_vec4f(vec4f& x, vec4f& y, vec4f* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2d_vec2d(vec2d& x, vec2d& y, vec2d* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4d_vec4d(vec4d& x, vec4d& y, vec4d* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2s_vec2s(vec2s& x, vec2s& y, vec2s* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4s_vec4s(vec4s& x, vec4s& y, vec4s* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2i_vec2i(vec2i& x, vec2i& y, vec2i* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4i_vec4i(vec4i& x, vec4i& y, vec4i* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2l_vec2l(vec2l& x, vec2l& y, vec2l* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4l_vec4l(vec4l& x, vec4l& y, vec4l* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2b_vec2b(vec2b& x, vec2b& y, vec2b* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4b_vec4b(vec4b& x, vec4b& y, vec4b* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2us_vec2us(vec2us& x, vec2us& y, vec2us* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4us_vec4us(vec4us& x, vec4us& y, vec4us* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2ui_vec2ui(vec2ui& x, vec2ui& y, vec2ui* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4ui_vec4ui(vec4ui& x, vec4ui& y, vec4ui* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2ul_vec2ul(vec2ul& x, vec2ul& y, vec2ul* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4ul_vec4ul(vec4ul& x, vec4ul& y, vec4ul* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2ub_vec2ub(vec2ub& x, vec2ub& y, vec2ub* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec4ub_vec4ub(vec4ub& x, vec4ub& y, vec4ub* ret) { *ret = wp::min(x, y); }
+WP_API void builtin_min_vec2h(vec2h& v, float16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3h(vec3h& v, float16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4h(vec4h& v, float16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_spatial_vectorh(spatial_vectorh& v, float16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2f(vec2f& v, float32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3f(vec3f& v, float32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4f(vec4f& v, float32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_spatial_vectorf(spatial_vectorf& v, float32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2d(vec2d& v, float64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3d(vec3d& v, float64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4d(vec4d& v, float64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_spatial_vectord(spatial_vectord& v, float64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2s(vec2s& v, int16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3s(vec3s& v, int16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4s(vec4s& v, int16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2i(vec2i& v, int32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3i(vec3i& v, int32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4i(vec4i& v, int32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2l(vec2l& v, int64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3l(vec3l& v, int64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4l(vec4l& v, int64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2b(vec2b& v, int8* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3b(vec3b& v, int8* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4b(vec4b& v, int8* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2us(vec2us& v, uint16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3us(vec3us& v, uint16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4us(vec4us& v, uint16* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2ui(vec2ui& v, uint32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3ui(vec3ui& v, uint32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4ui(vec4ui& v, uint32* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2ul(vec2ul& v, uint64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3ul(vec3ul& v, uint64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4ul(vec4ul& v, uint64* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec2ub(vec2ub& v, uint8* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec3ub(vec3ub& v, uint8* ret) { *ret = wp::min(v); }
+WP_API void builtin_min_vec4ub(vec4ub& v, uint8* ret) { *ret = wp::min(v); }
 WP_API void builtin_max_float16_float16(float16 x, float16 y, float16* ret) { *ret = wp::max(x, y); }
 WP_API void builtin_max_float32_float32(float32 x, float32 y, float32* ret) { *ret = wp::max(x, y); }
 WP_API void builtin_max_float64_float64(float64 x, float64 y, float64* ret) { *ret = wp::max(x, y); }
@@ -96,78 +96,78 @@ WP_API void builtin_max_uint16_uint16(uint16 x, uint16 y, uint16* ret) { *ret =
 WP_API void builtin_max_uint32_uint32(uint32 x, uint32 y, uint32* ret) { *ret = wp::max(x, y); }
 WP_API void builtin_max_uint64_uint64(uint64 x, uint64 y, uint64* ret) { *ret = wp::max(x, y); }
 WP_API void builtin_max_uint8_uint8(uint8 x, uint8 y, uint8* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2h_vec2h(vec2h x, vec2h y, vec2h* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4h_vec4h(vec4h x, vec4h y, vec4h* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2f_vec2f(vec2f x, vec2f y, vec2f* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4f_vec4f(vec4f x, vec4f y, vec4f* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2d_vec2d(vec2d x, vec2d y, vec2d* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4d_vec4d(vec4d x, vec4d y, vec4d* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2s_vec2s(vec2s x, vec2s y, vec2s* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4s_vec4s(vec4s x, vec4s y, vec4s* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2i_vec2i(vec2i x, vec2i y, vec2i* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4i_vec4i(vec4i x, vec4i y, vec4i* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2l_vec2l(vec2l x, vec2l y, vec2l* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4l_vec4l(vec4l x, vec4l y, vec4l* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2b_vec2b(vec2b x, vec2b y, vec2b* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4b_vec4b(vec4b x, vec4b y, vec4b* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2us_vec2us(vec2us x, vec2us y, vec2us* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4us_vec4us(vec4us x, vec4us y, vec4us* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2ui_vec2ui(vec2ui x, vec2ui y, vec2ui* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4ui_vec4ui(vec4ui x, vec4ui y, vec4ui* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2ul_vec2ul(vec2ul x, vec2ul y, vec2ul* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4ul_vec4ul(vec4ul x, vec4ul y, vec4ul* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2ub_vec2ub(vec2ub x, vec2ub y, vec2ub* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec4ub_vec4ub(vec4ub x, vec4ub y, vec4ub* ret) { *ret = wp::max(x, y); }
-WP_API void builtin_max_vec2h(vec2h v, float16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3h(vec3h v, float16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4h(vec4h v, float16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_spatial_vectorh(spatial_vectorh v, float16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2f(vec2f v, float32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3f(vec3f v, float32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4f(vec4f v, float32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_spatial_vectorf(spatial_vectorf v, float32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2d(vec2d v, float64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3d(vec3d v, float64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4d(vec4d v, float64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_spatial_vectord(spatial_vectord v, float64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2s(vec2s v, int16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3s(vec3s v, int16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4s(vec4s v, int16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2i(vec2i v, int32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3i(vec3i v, int32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4i(vec4i v, int32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2l(vec2l v, int64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3l(vec3l v, int64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4l(vec4l v, int64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2b(vec2b v, int8* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3b(vec3b v, int8* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4b(vec4b v, int8* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2us(vec2us v, uint16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3us(vec3us v, uint16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4us(vec4us v, uint16* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2ui(vec2ui v, uint32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3ui(vec3ui v, uint32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4ui(vec4ui v, uint32* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2ul(vec2ul v, uint64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3ul(vec3ul v, uint64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4ul(vec4ul v, uint64* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec2ub(vec2ub v, uint8* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec3ub(vec3ub v, uint8* ret) { *ret = wp::max(v); }
-WP_API void builtin_max_vec4ub(vec4ub v, uint8* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2h_vec2h(vec2h& x, vec2h& y, vec2h* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4h_vec4h(vec4h& x, vec4h& y, vec4h* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2f_vec2f(vec2f& x, vec2f& y, vec2f* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4f_vec4f(vec4f& x, vec4f& y, vec4f* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2d_vec2d(vec2d& x, vec2d& y, vec2d* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4d_vec4d(vec4d& x, vec4d& y, vec4d* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2s_vec2s(vec2s& x, vec2s& y, vec2s* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4s_vec4s(vec4s& x, vec4s& y, vec4s* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2i_vec2i(vec2i& x, vec2i& y, vec2i* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4i_vec4i(vec4i& x, vec4i& y, vec4i* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2l_vec2l(vec2l& x, vec2l& y, vec2l* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4l_vec4l(vec4l& x, vec4l& y, vec4l* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2b_vec2b(vec2b& x, vec2b& y, vec2b* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4b_vec4b(vec4b& x, vec4b& y, vec4b* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2us_vec2us(vec2us& x, vec2us& y, vec2us* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4us_vec4us(vec4us& x, vec4us& y, vec4us* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2ui_vec2ui(vec2ui& x, vec2ui& y, vec2ui* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4ui_vec4ui(vec4ui& x, vec4ui& y, vec4ui* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2ul_vec2ul(vec2ul& x, vec2ul& y, vec2ul* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4ul_vec4ul(vec4ul& x, vec4ul& y, vec4ul* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2ub_vec2ub(vec2ub& x, vec2ub& y, vec2ub* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec4ub_vec4ub(vec4ub& x, vec4ub& y, vec4ub* ret) { *ret = wp::max(x, y); }
+WP_API void builtin_max_vec2h(vec2h& v, float16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3h(vec3h& v, float16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4h(vec4h& v, float16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_spatial_vectorh(spatial_vectorh& v, float16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2f(vec2f& v, float32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3f(vec3f& v, float32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4f(vec4f& v, float32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_spatial_vectorf(spatial_vectorf& v, float32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2d(vec2d& v, float64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3d(vec3d& v, float64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4d(vec4d& v, float64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_spatial_vectord(spatial_vectord& v, float64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2s(vec2s& v, int16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3s(vec3s& v, int16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4s(vec4s& v, int16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2i(vec2i& v, int32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3i(vec3i& v, int32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4i(vec4i& v, int32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2l(vec2l& v, int64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3l(vec3l& v, int64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4l(vec4l& v, int64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2b(vec2b& v, int8* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3b(vec3b& v, int8* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4b(vec4b& v, int8* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2us(vec2us& v, uint16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3us(vec3us& v, uint16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4us(vec4us& v, uint16* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2ui(vec2ui& v, uint32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3ui(vec3ui& v, uint32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4ui(vec4ui& v, uint32* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2ul(vec2ul& v, uint64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3ul(vec3ul& v, uint64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4ul(vec4ul& v, uint64* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec2ub(vec2ub& v, uint8* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec3ub(vec3ub& v, uint8* ret) { *ret = wp::max(v); }
+WP_API void builtin_max_vec4ub(vec4ub& v, uint8* ret) { *ret = wp::max(v); }
 WP_API void builtin_clamp_float16_float16_float16(float16 x, float16 a, float16 b, float16* ret) { *ret = wp::clamp(x, a, b); }
 WP_API void builtin_clamp_float32_float32_float32(float32 x, float32 a, float32 b, float32* ret) { *ret = wp::clamp(x, a, b); }
 WP_API void builtin_clamp_float64_float64_float64(float64 x, float64 a, float64 b, float64* ret) { *ret = wp::clamp(x, a, b); }
@@ -298,494 +298,494 @@ WP_API void builtin_ceil_float64(float64 x, float64* ret) { *ret = wp::ceil(x);
 WP_API void builtin_frac_float16(float16 x, float16* ret) { *ret = wp::frac(x); }
 WP_API void builtin_frac_float32(float32 x, float32* ret) { *ret = wp::frac(x); }
 WP_API void builtin_frac_float64(float64 x, float64* ret) { *ret = wp::frac(x); }
-WP_API void builtin_dot_vec2h_vec2h(vec2h x, vec2h y, float16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3h_vec3h(vec3h x, vec3h y, float16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4h_vec4h(vec4h x, vec4h y, float16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, float16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2f_vec2f(vec2f x, vec2f y, float32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3f_vec3f(vec3f x, vec3f y, float32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4f_vec4f(vec4f x, vec4f y, float32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, float32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2d_vec2d(vec2d x, vec2d y, float64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3d_vec3d(vec3d x, vec3d y, float64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4d_vec4d(vec4d x, vec4d y, float64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, float64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2s_vec2s(vec2s x, vec2s y, int16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3s_vec3s(vec3s x, vec3s y, int16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4s_vec4s(vec4s x, vec4s y, int16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2i_vec2i(vec2i x, vec2i y, int32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3i_vec3i(vec3i x, vec3i y, int32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4i_vec4i(vec4i x, vec4i y, int32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2l_vec2l(vec2l x, vec2l y, int64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3l_vec3l(vec3l x, vec3l y, int64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4l_vec4l(vec4l x, vec4l y, int64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2b_vec2b(vec2b x, vec2b y, int8* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3b_vec3b(vec3b x, vec3b y, int8* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4b_vec4b(vec4b x, vec4b y, int8* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2us_vec2us(vec2us x, vec2us y, uint16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3us_vec3us(vec3us x, vec3us y, uint16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4us_vec4us(vec4us x, vec4us y, uint16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2ui_vec2ui(vec2ui x, vec2ui y, uint32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3ui_vec3ui(vec3ui x, vec3ui y, uint32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4ui_vec4ui(vec4ui x, vec4ui y, uint32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2ul_vec2ul(vec2ul x, vec2ul y, uint64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3ul_vec3ul(vec3ul x, vec3ul y, uint64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4ul_vec4ul(vec4ul x, vec4ul y, uint64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec2ub_vec2ub(vec2ub x, vec2ub y, uint8* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec3ub_vec3ub(vec3ub x, vec3ub y, uint8* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_vec4ub_vec4ub(vec4ub x, vec4ub y, uint8* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_quath_quath(quath x, quath y, float16* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_quatf_quatf(quatf x, quatf y, float32* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_dot_quatd_quatd(quatd x, quatd y, float64* ret) { *ret = wp::dot(x, y); }
-WP_API void builtin_ddot_mat22h_mat22h(mat22h x, mat22h y, float16* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat33h_mat33h(mat33h x, mat33h y, float16* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat44h_mat44h(mat44h x, mat44h y, float16* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_spatial_matrixh_spatial_matrixh(spatial_matrixh x, spatial_matrixh y, float16* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat22f_mat22f(mat22f x, mat22f y, float32* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat33f_mat33f(mat33f x, mat33f y, float32* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat44f_mat44f(mat44f x, mat44f y, float32* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_spatial_matrixf_spatial_matrixf(spatial_matrixf x, spatial_matrixf y, float32* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat22d_mat22d(mat22d x, mat22d y, float64* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat33d_mat33d(mat33d x, mat33d y, float64* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_mat44d_mat44d(mat44d x, mat44d y, float64* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_ddot_spatial_matrixd_spatial_matrixd(spatial_matrixd x, spatial_matrixd y, float64* ret) { *ret = wp::ddot(x, y); }
-WP_API void builtin_argmin_vec2h(vec2h v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3h(vec3h v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4h(vec4h v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_spatial_vectorh(spatial_vectorh v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2f(vec2f v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3f(vec3f v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4f(vec4f v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_spatial_vectorf(spatial_vectorf v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2d(vec2d v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3d(vec3d v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4d(vec4d v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_spatial_vectord(spatial_vectord v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2s(vec2s v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3s(vec3s v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4s(vec4s v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2i(vec2i v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3i(vec3i v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4i(vec4i v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2l(vec2l v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3l(vec3l v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4l(vec4l v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2b(vec2b v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3b(vec3b v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4b(vec4b v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2us(vec2us v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3us(vec3us v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4us(vec4us v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2ui(vec2ui v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3ui(vec3ui v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4ui(vec4ui v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2ul(vec2ul v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3ul(vec3ul v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4ul(vec4ul v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec2ub(vec2ub v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec3ub(vec3ub v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmin_vec4ub(vec4ub v, uint32* ret) { *ret = wp::argmin(v); }
-WP_API void builtin_argmax_vec2h(vec2h v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3h(vec3h v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4h(vec4h v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_spatial_vectorh(spatial_vectorh v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2f(vec2f v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3f(vec3f v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4f(vec4f v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_spatial_vectorf(spatial_vectorf v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2d(vec2d v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3d(vec3d v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4d(vec4d v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_spatial_vectord(spatial_vectord v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2s(vec2s v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3s(vec3s v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4s(vec4s v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2i(vec2i v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3i(vec3i v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4i(vec4i v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2l(vec2l v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3l(vec3l v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4l(vec4l v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2b(vec2b v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3b(vec3b v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4b(vec4b v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2us(vec2us v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3us(vec3us v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4us(vec4us v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2ui(vec2ui v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3ui(vec3ui v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4ui(vec4ui v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2ul(vec2ul v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3ul(vec3ul v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4ul(vec4ul v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec2ub(vec2ub v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec3ub(vec3ub v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_argmax_vec4ub(vec4ub v, uint32* ret) { *ret = wp::argmax(v); }
-WP_API void builtin_outer_vec2h_vec2h(vec2h x, vec2h y, mat22h* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec3h_vec3h(vec3h x, vec3h y, mat33h* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec4h_vec4h(vec4h x, vec4h y, mat44h* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_matrixh* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec2f_vec2f(vec2f x, vec2f y, mat22f* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec3f_vec3f(vec3f x, vec3f y, mat33f* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec4f_vec4f(vec4f x, vec4f y, mat44f* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_matrixf* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec2d_vec2d(vec2d x, vec2d y, mat22d* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec3d_vec3d(vec3d x, vec3d y, mat33d* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_vec4d_vec4d(vec4d x, vec4d y, mat44d* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_outer_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_matrixd* ret) { *ret = wp::outer(x, y); }
-WP_API void builtin_cross_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_cross_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::cross(x, y); }
-WP_API void builtin_skew_vec3h(vec3h x, mat33h* ret) { *ret = wp::skew(x); }
-WP_API void builtin_skew_vec3f(vec3f x, mat33f* ret) { *ret = wp::skew(x); }
-WP_API void builtin_skew_vec3d(vec3d x, mat33d* ret) { *ret = wp::skew(x); }
-WP_API void builtin_length_vec2h(vec2h x, float16* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec3h(vec3h x, float16* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec4h(vec4h x, float16* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_spatial_vectorh(spatial_vectorh x, float16* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec2f(vec2f x, float32* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec3f(vec3f x, float32* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec4f(vec4f x, float32* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_spatial_vectorf(spatial_vectorf x, float32* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec2d(vec2d x, float64* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec3d(vec3d x, float64* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_vec4d(vec4d x, float64* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_spatial_vectord(spatial_vectord x, float64* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_quath(quath x, float16* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_quatf(quatf x, float32* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_quatd(quatd x, float64* ret) { *ret = wp::length(x); }
-WP_API void builtin_length_sq_vec2h(vec2h x, float16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3h(vec3h x, float16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4h(vec4h x, float16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_spatial_vectorh(spatial_vectorh x, float16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2f(vec2f x, float32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3f(vec3f x, float32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4f(vec4f x, float32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_spatial_vectorf(spatial_vectorf x, float32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2d(vec2d x, float64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3d(vec3d x, float64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4d(vec4d x, float64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_spatial_vectord(spatial_vectord x, float64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2s(vec2s x, int16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3s(vec3s x, int16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4s(vec4s x, int16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2i(vec2i x, int32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3i(vec3i x, int32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4i(vec4i x, int32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2l(vec2l x, int64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3l(vec3l x, int64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4l(vec4l x, int64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2b(vec2b x, int8* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3b(vec3b x, int8* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4b(vec4b x, int8* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2us(vec2us x, uint16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3us(vec3us x, uint16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4us(vec4us x, uint16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2ui(vec2ui x, uint32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3ui(vec3ui x, uint32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4ui(vec4ui x, uint32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2ul(vec2ul x, uint64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3ul(vec3ul x, uint64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4ul(vec4ul x, uint64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec2ub(vec2ub x, uint8* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec3ub(vec3ub x, uint8* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_vec4ub(vec4ub x, uint8* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_quath(quath x, float16* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_quatf(quatf x, float32* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_length_sq_quatd(quatd x, float64* ret) { *ret = wp::length_sq(x); }
-WP_API void builtin_normalize_vec2h(vec2h x, vec2h* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec3h(vec3h x, vec3h* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec4h(vec4h x, vec4h* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_spatial_vectorh(spatial_vectorh x, spatial_vectorh* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec2f(vec2f x, vec2f* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec3f(vec3f x, vec3f* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec4f(vec4f x, vec4f* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_spatial_vectorf(spatial_vectorf x, spatial_vectorf* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec2d(vec2d x, vec2d* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec3d(vec3d x, vec3d* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_vec4d(vec4d x, vec4d* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_spatial_vectord(spatial_vectord x, spatial_vectord* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_quath(quath x, quath* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_quatf(quatf x, quatf* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_normalize_quatd(quatd x, quatd* ret) { *ret = wp::normalize(x); }
-WP_API void builtin_transpose_mat22h(mat22h m, mat22h* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat33h(mat33h m, mat33h* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat44h(mat44h m, mat44h* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_spatial_matrixh(spatial_matrixh m, spatial_matrixh* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat22f(mat22f m, mat22f* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat33f(mat33f m, mat33f* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat44f(mat44f m, mat44f* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_spatial_matrixf(spatial_matrixf m, spatial_matrixf* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat22d(mat22d m, mat22d* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat33d(mat33d m, mat33d* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_mat44d(mat44d m, mat44d* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_transpose_spatial_matrixd(spatial_matrixd m, spatial_matrixd* ret) { *ret = wp::transpose(m); }
-WP_API void builtin_inverse_mat22h(mat22h m, mat22h* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat22f(mat22f m, mat22f* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat22d(mat22d m, mat22d* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat33h(mat33h m, mat33h* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat33f(mat33f m, mat33f* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat33d(mat33d m, mat33d* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat44h(mat44h m, mat44h* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat44f(mat44f m, mat44f* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_inverse_mat44d(mat44d m, mat44d* ret) { *ret = wp::inverse(m); }
-WP_API void builtin_determinant_mat22h(mat22h m, float16* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat22f(mat22f m, float32* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat22d(mat22d m, float64* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat33h(mat33h m, float16* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat33f(mat33f m, float32* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat33d(mat33d m, float64* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat44h(mat44h m, float16* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat44f(mat44f m, float32* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_determinant_mat44d(mat44d m, float64* ret) { *ret = wp::determinant(m); }
-WP_API void builtin_trace_mat22h(mat22h m, float16* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat33h(mat33h m, float16* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat44h(mat44h m, float16* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_spatial_matrixh(spatial_matrixh m, float16* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat22f(mat22f m, float32* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat33f(mat33f m, float32* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat44f(mat44f m, float32* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_spatial_matrixf(spatial_matrixf m, float32* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat22d(mat22d m, float64* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat33d(mat33d m, float64* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_mat44d(mat44d m, float64* ret) { *ret = wp::trace(m); }
-WP_API void builtin_trace_spatial_matrixd(spatial_matrixd m, float64* ret) { *ret = wp::trace(m); }
-WP_API void builtin_diag_vec2h(vec2h d, mat22h* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec3h(vec3h d, mat33h* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec4h(vec4h d, mat44h* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_spatial_vectorh(spatial_vectorh d, spatial_matrixh* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec2f(vec2f d, mat22f* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec3f(vec3f d, mat33f* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec4f(vec4f d, mat44f* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_spatial_vectorf(spatial_vectorf d, spatial_matrixf* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec2d(vec2d d, mat22d* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec3d(vec3d d, mat33d* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_vec4d(vec4d d, mat44d* ret) { *ret = wp::diag(d); }
-WP_API void builtin_diag_spatial_vectord(spatial_vectord d, spatial_matrixd* ret) { *ret = wp::diag(d); }
-WP_API void builtin_get_diag_mat22h(mat22h m, vec2h* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat33h(mat33h m, vec3h* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat44h(mat44h m, vec4h* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_spatial_matrixh(spatial_matrixh m, spatial_vectorh* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat22f(mat22f m, vec2f* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat33f(mat33f m, vec3f* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat44f(mat44f m, vec4f* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_spatial_matrixf(spatial_matrixf m, spatial_vectorf* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat22d(mat22d m, vec2d* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat33d(mat33d m, vec3d* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_mat44d(mat44d m, vec4d* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_get_diag_spatial_matrixd(spatial_matrixd m, spatial_vectord* ret) { *ret = wp::get_diag(m); }
-WP_API void builtin_cw_mul_vec2h_vec2h(vec2h x, vec2h y, vec2h* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4h_vec4h(vec4h x, vec4h y, vec4h* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2f_vec2f(vec2f x, vec2f y, vec2f* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4f_vec4f(vec4f x, vec4f y, vec4f* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2d_vec2d(vec2d x, vec2d y, vec2d* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4d_vec4d(vec4d x, vec4d y, vec4d* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2s_vec2s(vec2s x, vec2s y, vec2s* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4s_vec4s(vec4s x, vec4s y, vec4s* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2i_vec2i(vec2i x, vec2i y, vec2i* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4i_vec4i(vec4i x, vec4i y, vec4i* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2l_vec2l(vec2l x, vec2l y, vec2l* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4l_vec4l(vec4l x, vec4l y, vec4l* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2b_vec2b(vec2b x, vec2b y, vec2b* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4b_vec4b(vec4b x, vec4b y, vec4b* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2us_vec2us(vec2us x, vec2us y, vec2us* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4us_vec4us(vec4us x, vec4us y, vec4us* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2ui_vec2ui(vec2ui x, vec2ui y, vec2ui* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4ui_vec4ui(vec4ui x, vec4ui y, vec4ui* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2ul_vec2ul(vec2ul x, vec2ul y, vec2ul* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4ul_vec4ul(vec4ul x, vec4ul y, vec4ul* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec2ub_vec2ub(vec2ub x, vec2ub y, vec2ub* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_vec4ub_vec4ub(vec4ub x, vec4ub y, vec4ub* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat22h_mat22h(mat22h x, mat22h y, mat22h* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat33h_mat33h(mat33h x, mat33h y, mat33h* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat44h_mat44h(mat44h x, mat44h y, mat44h* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_spatial_matrixh_spatial_matrixh(spatial_matrixh x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat22f_mat22f(mat22f x, mat22f y, mat22f* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat33f_mat33f(mat33f x, mat33f y, mat33f* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat44f_mat44f(mat44f x, mat44f y, mat44f* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_spatial_matrixf_spatial_matrixf(spatial_matrixf x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat22d_mat22d(mat22d x, mat22d y, mat22d* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat33d_mat33d(mat33d x, mat33d y, mat33d* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_mat44d_mat44d(mat44d x, mat44d y, mat44d* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_mul_spatial_matrixd_spatial_matrixd(spatial_matrixd x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::cw_mul(x, y); }
-WP_API void builtin_cw_div_vec2h_vec2h(vec2h x, vec2h y, vec2h* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4h_vec4h(vec4h x, vec4h y, vec4h* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2f_vec2f(vec2f x, vec2f y, vec2f* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4f_vec4f(vec4f x, vec4f y, vec4f* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2d_vec2d(vec2d x, vec2d y, vec2d* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4d_vec4d(vec4d x, vec4d y, vec4d* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2s_vec2s(vec2s x, vec2s y, vec2s* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4s_vec4s(vec4s x, vec4s y, vec4s* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2i_vec2i(vec2i x, vec2i y, vec2i* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4i_vec4i(vec4i x, vec4i y, vec4i* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2l_vec2l(vec2l x, vec2l y, vec2l* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4l_vec4l(vec4l x, vec4l y, vec4l* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2b_vec2b(vec2b x, vec2b y, vec2b* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4b_vec4b(vec4b x, vec4b y, vec4b* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2us_vec2us(vec2us x, vec2us y, vec2us* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4us_vec4us(vec4us x, vec4us y, vec4us* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2ui_vec2ui(vec2ui x, vec2ui y, vec2ui* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4ui_vec4ui(vec4ui x, vec4ui y, vec4ui* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2ul_vec2ul(vec2ul x, vec2ul y, vec2ul* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4ul_vec4ul(vec4ul x, vec4ul y, vec4ul* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec2ub_vec2ub(vec2ub x, vec2ub y, vec2ub* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_vec4ub_vec4ub(vec4ub x, vec4ub y, vec4ub* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat22h_mat22h(mat22h x, mat22h y, mat22h* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat33h_mat33h(mat33h x, mat33h y, mat33h* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat44h_mat44h(mat44h x, mat44h y, mat44h* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_spatial_matrixh_spatial_matrixh(spatial_matrixh x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat22f_mat22f(mat22f x, mat22f y, mat22f* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat33f_mat33f(mat33f x, mat33f y, mat33f* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat44f_mat44f(mat44f x, mat44f y, mat44f* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_spatial_matrixf_spatial_matrixf(spatial_matrixf x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat22d_mat22d(mat22d x, mat22d y, mat22d* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat33d_mat33d(mat33d x, mat33d y, mat33d* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_mat44d_mat44d(mat44d x, mat44d y, mat44d* ret) { *ret = wp::cw_div(x, y); }
-WP_API void builtin_cw_div_spatial_matrixd_spatial_matrixd(spatial_matrixd x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_dot_vec2h_vec2h(vec2h& x, vec2h& y, float16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3h_vec3h(vec3h& x, vec3h& y, float16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4h_vec4h(vec4h& x, vec4h& y, float16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, float16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2f_vec2f(vec2f& x, vec2f& y, float32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3f_vec3f(vec3f& x, vec3f& y, float32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4f_vec4f(vec4f& x, vec4f& y, float32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, float32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2d_vec2d(vec2d& x, vec2d& y, float64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3d_vec3d(vec3d& x, vec3d& y, float64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4d_vec4d(vec4d& x, vec4d& y, float64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, float64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2s_vec2s(vec2s& x, vec2s& y, int16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3s_vec3s(vec3s& x, vec3s& y, int16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4s_vec4s(vec4s& x, vec4s& y, int16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2i_vec2i(vec2i& x, vec2i& y, int32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3i_vec3i(vec3i& x, vec3i& y, int32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4i_vec4i(vec4i& x, vec4i& y, int32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2l_vec2l(vec2l& x, vec2l& y, int64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3l_vec3l(vec3l& x, vec3l& y, int64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4l_vec4l(vec4l& x, vec4l& y, int64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2b_vec2b(vec2b& x, vec2b& y, int8* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3b_vec3b(vec3b& x, vec3b& y, int8* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4b_vec4b(vec4b& x, vec4b& y, int8* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2us_vec2us(vec2us& x, vec2us& y, uint16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3us_vec3us(vec3us& x, vec3us& y, uint16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4us_vec4us(vec4us& x, vec4us& y, uint16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2ui_vec2ui(vec2ui& x, vec2ui& y, uint32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3ui_vec3ui(vec3ui& x, vec3ui& y, uint32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4ui_vec4ui(vec4ui& x, vec4ui& y, uint32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2ul_vec2ul(vec2ul& x, vec2ul& y, uint64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3ul_vec3ul(vec3ul& x, vec3ul& y, uint64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4ul_vec4ul(vec4ul& x, vec4ul& y, uint64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec2ub_vec2ub(vec2ub& x, vec2ub& y, uint8* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec3ub_vec3ub(vec3ub& x, vec3ub& y, uint8* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_vec4ub_vec4ub(vec4ub& x, vec4ub& y, uint8* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_quath_quath(quath& x, quath& y, float16* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_quatf_quatf(quatf& x, quatf& y, float32* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_dot_quatd_quatd(quatd& x, quatd& y, float64* ret) { *ret = wp::dot(x, y); }
+WP_API void builtin_ddot_mat22h_mat22h(mat22h& x, mat22h& y, float16* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat33h_mat33h(mat33h& x, mat33h& y, float16* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat44h_mat44h(mat44h& x, mat44h& y, float16* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_spatial_matrixh_spatial_matrixh(spatial_matrixh& x, spatial_matrixh& y, float16* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat22f_mat22f(mat22f& x, mat22f& y, float32* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat33f_mat33f(mat33f& x, mat33f& y, float32* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat44f_mat44f(mat44f& x, mat44f& y, float32* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_spatial_matrixf_spatial_matrixf(spatial_matrixf& x, spatial_matrixf& y, float32* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat22d_mat22d(mat22d& x, mat22d& y, float64* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat33d_mat33d(mat33d& x, mat33d& y, float64* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_mat44d_mat44d(mat44d& x, mat44d& y, float64* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_ddot_spatial_matrixd_spatial_matrixd(spatial_matrixd& x, spatial_matrixd& y, float64* ret) { *ret = wp::ddot(x, y); }
+WP_API void builtin_argmin_vec2h(vec2h& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3h(vec3h& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4h(vec4h& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_spatial_vectorh(spatial_vectorh& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2f(vec2f& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3f(vec3f& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4f(vec4f& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_spatial_vectorf(spatial_vectorf& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2d(vec2d& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3d(vec3d& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4d(vec4d& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_spatial_vectord(spatial_vectord& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2s(vec2s& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3s(vec3s& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4s(vec4s& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2i(vec2i& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3i(vec3i& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4i(vec4i& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2l(vec2l& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3l(vec3l& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4l(vec4l& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2b(vec2b& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3b(vec3b& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4b(vec4b& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2us(vec2us& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3us(vec3us& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4us(vec4us& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2ui(vec2ui& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3ui(vec3ui& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4ui(vec4ui& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2ul(vec2ul& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3ul(vec3ul& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4ul(vec4ul& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec2ub(vec2ub& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec3ub(vec3ub& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmin_vec4ub(vec4ub& v, uint32* ret) { *ret = wp::argmin(v); }
+WP_API void builtin_argmax_vec2h(vec2h& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3h(vec3h& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4h(vec4h& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_spatial_vectorh(spatial_vectorh& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2f(vec2f& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3f(vec3f& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4f(vec4f& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_spatial_vectorf(spatial_vectorf& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2d(vec2d& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3d(vec3d& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4d(vec4d& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_spatial_vectord(spatial_vectord& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2s(vec2s& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3s(vec3s& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4s(vec4s& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2i(vec2i& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3i(vec3i& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4i(vec4i& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2l(vec2l& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3l(vec3l& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4l(vec4l& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2b(vec2b& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3b(vec3b& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4b(vec4b& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2us(vec2us& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3us(vec3us& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4us(vec4us& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2ui(vec2ui& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3ui(vec3ui& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4ui(vec4ui& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2ul(vec2ul& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3ul(vec3ul& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4ul(vec4ul& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec2ub(vec2ub& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec3ub(vec3ub& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_argmax_vec4ub(vec4ub& v, uint32* ret) { *ret = wp::argmax(v); }
+WP_API void builtin_outer_vec2h_vec2h(vec2h& x, vec2h& y, mat22h* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec3h_vec3h(vec3h& x, vec3h& y, mat33h* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec4h_vec4h(vec4h& x, vec4h& y, mat44h* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_matrixh* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec2f_vec2f(vec2f& x, vec2f& y, mat22f* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec3f_vec3f(vec3f& x, vec3f& y, mat33f* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec4f_vec4f(vec4f& x, vec4f& y, mat44f* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_matrixf* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec2d_vec2d(vec2d& x, vec2d& y, mat22d* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec3d_vec3d(vec3d& x, vec3d& y, mat33d* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_vec4d_vec4d(vec4d& x, vec4d& y, mat44d* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_outer_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_matrixd* ret) { *ret = wp::outer(x, y); }
+WP_API void builtin_cross_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_cross_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::cross(x, y); }
+WP_API void builtin_skew_vec3h(vec3h& x, mat33h* ret) { *ret = wp::skew(x); }
+WP_API void builtin_skew_vec3f(vec3f& x, mat33f* ret) { *ret = wp::skew(x); }
+WP_API void builtin_skew_vec3d(vec3d& x, mat33d* ret) { *ret = wp::skew(x); }
+WP_API void builtin_length_vec2h(vec2h& x, float16* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec3h(vec3h& x, float16* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec4h(vec4h& x, float16* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_spatial_vectorh(spatial_vectorh& x, float16* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec2f(vec2f& x, float32* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec3f(vec3f& x, float32* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec4f(vec4f& x, float32* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_spatial_vectorf(spatial_vectorf& x, float32* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec2d(vec2d& x, float64* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec3d(vec3d& x, float64* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_vec4d(vec4d& x, float64* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_spatial_vectord(spatial_vectord& x, float64* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_quath(quath& x, float16* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_quatf(quatf& x, float32* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_quatd(quatd& x, float64* ret) { *ret = wp::length(x); }
+WP_API void builtin_length_sq_vec2h(vec2h& x, float16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3h(vec3h& x, float16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4h(vec4h& x, float16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_spatial_vectorh(spatial_vectorh& x, float16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2f(vec2f& x, float32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3f(vec3f& x, float32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4f(vec4f& x, float32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_spatial_vectorf(spatial_vectorf& x, float32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2d(vec2d& x, float64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3d(vec3d& x, float64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4d(vec4d& x, float64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_spatial_vectord(spatial_vectord& x, float64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2s(vec2s& x, int16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3s(vec3s& x, int16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4s(vec4s& x, int16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2i(vec2i& x, int32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3i(vec3i& x, int32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4i(vec4i& x, int32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2l(vec2l& x, int64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3l(vec3l& x, int64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4l(vec4l& x, int64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2b(vec2b& x, int8* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3b(vec3b& x, int8* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4b(vec4b& x, int8* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2us(vec2us& x, uint16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3us(vec3us& x, uint16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4us(vec4us& x, uint16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2ui(vec2ui& x, uint32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3ui(vec3ui& x, uint32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4ui(vec4ui& x, uint32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2ul(vec2ul& x, uint64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3ul(vec3ul& x, uint64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4ul(vec4ul& x, uint64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec2ub(vec2ub& x, uint8* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec3ub(vec3ub& x, uint8* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_vec4ub(vec4ub& x, uint8* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_quath(quath& x, float16* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_quatf(quatf& x, float32* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_length_sq_quatd(quatd& x, float64* ret) { *ret = wp::length_sq(x); }
+WP_API void builtin_normalize_vec2h(vec2h& x, vec2h* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec3h(vec3h& x, vec3h* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec4h(vec4h& x, vec4h* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_spatial_vectorh(spatial_vectorh& x, spatial_vectorh* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec2f(vec2f& x, vec2f* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec3f(vec3f& x, vec3f* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec4f(vec4f& x, vec4f* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_spatial_vectorf(spatial_vectorf& x, spatial_vectorf* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec2d(vec2d& x, vec2d* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec3d(vec3d& x, vec3d* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_vec4d(vec4d& x, vec4d* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_spatial_vectord(spatial_vectord& x, spatial_vectord* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_quath(quath& x, quath* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_quatf(quatf& x, quatf* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_normalize_quatd(quatd& x, quatd* ret) { *ret = wp::normalize(x); }
+WP_API void builtin_transpose_mat22h(mat22h& m, mat22h* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat33h(mat33h& m, mat33h* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat44h(mat44h& m, mat44h* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_spatial_matrixh(spatial_matrixh& m, spatial_matrixh* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat22f(mat22f& m, mat22f* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat33f(mat33f& m, mat33f* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat44f(mat44f& m, mat44f* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_spatial_matrixf(spatial_matrixf& m, spatial_matrixf* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat22d(mat22d& m, mat22d* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat33d(mat33d& m, mat33d* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_mat44d(mat44d& m, mat44d* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_transpose_spatial_matrixd(spatial_matrixd& m, spatial_matrixd* ret) { *ret = wp::transpose(m); }
+WP_API void builtin_inverse_mat22h(mat22h& m, mat22h* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat22f(mat22f& m, mat22f* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat22d(mat22d& m, mat22d* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat33h(mat33h& m, mat33h* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat33f(mat33f& m, mat33f* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat33d(mat33d& m, mat33d* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat44h(mat44h& m, mat44h* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat44f(mat44f& m, mat44f* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_inverse_mat44d(mat44d& m, mat44d* ret) { *ret = wp::inverse(m); }
+WP_API void builtin_determinant_mat22h(mat22h& m, float16* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat22f(mat22f& m, float32* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat22d(mat22d& m, float64* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat33h(mat33h& m, float16* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat33f(mat33f& m, float32* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat33d(mat33d& m, float64* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat44h(mat44h& m, float16* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat44f(mat44f& m, float32* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_determinant_mat44d(mat44d& m, float64* ret) { *ret = wp::determinant(m); }
+WP_API void builtin_trace_mat22h(mat22h& m, float16* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat33h(mat33h& m, float16* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat44h(mat44h& m, float16* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_spatial_matrixh(spatial_matrixh& m, float16* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat22f(mat22f& m, float32* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat33f(mat33f& m, float32* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat44f(mat44f& m, float32* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_spatial_matrixf(spatial_matrixf& m, float32* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat22d(mat22d& m, float64* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat33d(mat33d& m, float64* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_mat44d(mat44d& m, float64* ret) { *ret = wp::trace(m); }
+WP_API void builtin_trace_spatial_matrixd(spatial_matrixd& m, float64* ret) { *ret = wp::trace(m); }
+WP_API void builtin_diag_vec2h(vec2h& d, mat22h* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec3h(vec3h& d, mat33h* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec4h(vec4h& d, mat44h* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_spatial_vectorh(spatial_vectorh& d, spatial_matrixh* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec2f(vec2f& d, mat22f* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec3f(vec3f& d, mat33f* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec4f(vec4f& d, mat44f* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_spatial_vectorf(spatial_vectorf& d, spatial_matrixf* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec2d(vec2d& d, mat22d* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec3d(vec3d& d, mat33d* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_vec4d(vec4d& d, mat44d* ret) { *ret = wp::diag(d); }
+WP_API void builtin_diag_spatial_vectord(spatial_vectord& d, spatial_matrixd* ret) { *ret = wp::diag(d); }
+WP_API void builtin_get_diag_mat22h(mat22h& m, vec2h* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat33h(mat33h& m, vec3h* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat44h(mat44h& m, vec4h* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_spatial_matrixh(spatial_matrixh& m, spatial_vectorh* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat22f(mat22f& m, vec2f* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat33f(mat33f& m, vec3f* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat44f(mat44f& m, vec4f* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_spatial_matrixf(spatial_matrixf& m, spatial_vectorf* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat22d(mat22d& m, vec2d* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat33d(mat33d& m, vec3d* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_mat44d(mat44d& m, vec4d* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_get_diag_spatial_matrixd(spatial_matrixd& m, spatial_vectord* ret) { *ret = wp::get_diag(m); }
+WP_API void builtin_cw_mul_vec2h_vec2h(vec2h& x, vec2h& y, vec2h* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4h_vec4h(vec4h& x, vec4h& y, vec4h* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2f_vec2f(vec2f& x, vec2f& y, vec2f* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4f_vec4f(vec4f& x, vec4f& y, vec4f* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2d_vec2d(vec2d& x, vec2d& y, vec2d* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4d_vec4d(vec4d& x, vec4d& y, vec4d* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2s_vec2s(vec2s& x, vec2s& y, vec2s* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4s_vec4s(vec4s& x, vec4s& y, vec4s* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2i_vec2i(vec2i& x, vec2i& y, vec2i* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4i_vec4i(vec4i& x, vec4i& y, vec4i* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2l_vec2l(vec2l& x, vec2l& y, vec2l* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4l_vec4l(vec4l& x, vec4l& y, vec4l* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2b_vec2b(vec2b& x, vec2b& y, vec2b* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4b_vec4b(vec4b& x, vec4b& y, vec4b* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2us_vec2us(vec2us& x, vec2us& y, vec2us* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4us_vec4us(vec4us& x, vec4us& y, vec4us* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2ui_vec2ui(vec2ui& x, vec2ui& y, vec2ui* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4ui_vec4ui(vec4ui& x, vec4ui& y, vec4ui* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2ul_vec2ul(vec2ul& x, vec2ul& y, vec2ul* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4ul_vec4ul(vec4ul& x, vec4ul& y, vec4ul* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec2ub_vec2ub(vec2ub& x, vec2ub& y, vec2ub* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_vec4ub_vec4ub(vec4ub& x, vec4ub& y, vec4ub* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat22h_mat22h(mat22h& x, mat22h& y, mat22h* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat33h_mat33h(mat33h& x, mat33h& y, mat33h* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat44h_mat44h(mat44h& x, mat44h& y, mat44h* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_spatial_matrixh_spatial_matrixh(spatial_matrixh& x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat22f_mat22f(mat22f& x, mat22f& y, mat22f* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat33f_mat33f(mat33f& x, mat33f& y, mat33f* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat44f_mat44f(mat44f& x, mat44f& y, mat44f* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_spatial_matrixf_spatial_matrixf(spatial_matrixf& x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat22d_mat22d(mat22d& x, mat22d& y, mat22d* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat33d_mat33d(mat33d& x, mat33d& y, mat33d* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_mat44d_mat44d(mat44d& x, mat44d& y, mat44d* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_mul_spatial_matrixd_spatial_matrixd(spatial_matrixd& x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::cw_mul(x, y); }
+WP_API void builtin_cw_div_vec2h_vec2h(vec2h& x, vec2h& y, vec2h* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4h_vec4h(vec4h& x, vec4h& y, vec4h* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2f_vec2f(vec2f& x, vec2f& y, vec2f* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4f_vec4f(vec4f& x, vec4f& y, vec4f* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2d_vec2d(vec2d& x, vec2d& y, vec2d* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4d_vec4d(vec4d& x, vec4d& y, vec4d* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2s_vec2s(vec2s& x, vec2s& y, vec2s* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4s_vec4s(vec4s& x, vec4s& y, vec4s* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2i_vec2i(vec2i& x, vec2i& y, vec2i* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4i_vec4i(vec4i& x, vec4i& y, vec4i* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2l_vec2l(vec2l& x, vec2l& y, vec2l* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4l_vec4l(vec4l& x, vec4l& y, vec4l* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2b_vec2b(vec2b& x, vec2b& y, vec2b* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4b_vec4b(vec4b& x, vec4b& y, vec4b* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2us_vec2us(vec2us& x, vec2us& y, vec2us* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4us_vec4us(vec4us& x, vec4us& y, vec4us* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2ui_vec2ui(vec2ui& x, vec2ui& y, vec2ui* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4ui_vec4ui(vec4ui& x, vec4ui& y, vec4ui* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2ul_vec2ul(vec2ul& x, vec2ul& y, vec2ul* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4ul_vec4ul(vec4ul& x, vec4ul& y, vec4ul* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec2ub_vec2ub(vec2ub& x, vec2ub& y, vec2ub* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_vec4ub_vec4ub(vec4ub& x, vec4ub& y, vec4ub* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat22h_mat22h(mat22h& x, mat22h& y, mat22h* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat33h_mat33h(mat33h& x, mat33h& y, mat33h* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat44h_mat44h(mat44h& x, mat44h& y, mat44h* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_spatial_matrixh_spatial_matrixh(spatial_matrixh& x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat22f_mat22f(mat22f& x, mat22f& y, mat22f* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat33f_mat33f(mat33f& x, mat33f& y, mat33f* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat44f_mat44f(mat44f& x, mat44f& y, mat44f* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_spatial_matrixf_spatial_matrixf(spatial_matrixf& x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat22d_mat22d(mat22d& x, mat22d& y, mat22d* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat33d_mat33d(mat33d& x, mat33d& y, mat33d* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_mat44d_mat44d(mat44d& x, mat44d& y, mat44d* ret) { *ret = wp::cw_div(x, y); }
+WP_API void builtin_cw_div_spatial_matrixd_spatial_matrixd(spatial_matrixd& x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::cw_div(x, y); }
 WP_API void builtin_quat_identity(quatf* ret) { *ret = wp::quat_identity(); }
-WP_API void builtin_quat_from_axis_angle_vec3h_float16(vec3h axis, float16 angle, quath* ret) { *ret = wp::quat_from_axis_angle(axis, angle); }
-WP_API void builtin_quat_from_axis_angle_vec3f_float32(vec3f axis, float32 angle, quatf* ret) { *ret = wp::quat_from_axis_angle(axis, angle); }
-WP_API void builtin_quat_from_axis_angle_vec3d_float64(vec3d axis, float64 angle, quatd* ret) { *ret = wp::quat_from_axis_angle(axis, angle); }
-WP_API void builtin_quat_from_matrix_mat33h(mat33h m, quath* ret) { *ret = wp::quat_from_matrix(m); }
-WP_API void builtin_quat_from_matrix_mat33f(mat33f m, quatf* ret) { *ret = wp::quat_from_matrix(m); }
-WP_API void builtin_quat_from_matrix_mat33d(mat33d m, quatd* ret) { *ret = wp::quat_from_matrix(m); }
+WP_API void builtin_quat_from_axis_angle_vec3h_float16(vec3h& axis, float16 angle, quath* ret) { *ret = wp::quat_from_axis_angle(axis, angle); }
+WP_API void builtin_quat_from_axis_angle_vec3f_float32(vec3f& axis, float32 angle, quatf* ret) { *ret = wp::quat_from_axis_angle(axis, angle); }
+WP_API void builtin_quat_from_axis_angle_vec3d_float64(vec3d& axis, float64 angle, quatd* ret) { *ret = wp::quat_from_axis_angle(axis, angle); }
+WP_API void builtin_quat_from_matrix_mat33h(mat33h& m, quath* ret) { *ret = wp::quat_from_matrix(m); }
+WP_API void builtin_quat_from_matrix_mat33f(mat33f& m, quatf* ret) { *ret = wp::quat_from_matrix(m); }
+WP_API void builtin_quat_from_matrix_mat33d(mat33d& m, quatd* ret) { *ret = wp::quat_from_matrix(m); }
 WP_API void builtin_quat_rpy_float16_float16_float16(float16 roll, float16 pitch, float16 yaw, quath* ret) { *ret = wp::quat_rpy(roll, pitch, yaw); }
 WP_API void builtin_quat_rpy_float32_float32_float32(float32 roll, float32 pitch, float32 yaw, quatf* ret) { *ret = wp::quat_rpy(roll, pitch, yaw); }
 WP_API void builtin_quat_rpy_float64_float64_float64(float64 roll, float64 pitch, float64 yaw, quatd* ret) { *ret = wp::quat_rpy(roll, pitch, yaw); }
-WP_API void builtin_quat_inverse_quath(quath q, quath* ret) { *ret = wp::quat_inverse(q); }
-WP_API void builtin_quat_inverse_quatf(quatf q, quatf* ret) { *ret = wp::quat_inverse(q); }
-WP_API void builtin_quat_inverse_quatd(quatd q, quatd* ret) { *ret = wp::quat_inverse(q); }
-WP_API void builtin_quat_rotate_quath_vec3h(quath q, vec3h p, vec3h* ret) { *ret = wp::quat_rotate(q, p); }
-WP_API void builtin_quat_rotate_quatf_vec3f(quatf q, vec3f p, vec3f* ret) { *ret = wp::quat_rotate(q, p); }
-WP_API void builtin_quat_rotate_quatd_vec3d(quatd q, vec3d p, vec3d* ret) { *ret = wp::quat_rotate(q, p); }
-WP_API void builtin_quat_rotate_inv_quath_vec3h(quath q, vec3h p, vec3h* ret) { *ret = wp::quat_rotate_inv(q, p); }
-WP_API void builtin_quat_rotate_inv_quatf_vec3f(quatf q, vec3f p, vec3f* ret) { *ret = wp::quat_rotate_inv(q, p); }
-WP_API void builtin_quat_rotate_inv_quatd_vec3d(quatd q, vec3d p, vec3d* ret) { *ret = wp::quat_rotate_inv(q, p); }
-WP_API void builtin_quat_slerp_quath_quath_float16(quath q0, quath q1, float16 t, quath* ret) { *ret = wp::quat_slerp(q0, q1, t); }
-WP_API void builtin_quat_slerp_quatf_quatf_float32(quatf q0, quatf q1, float32 t, quatf* ret) { *ret = wp::quat_slerp(q0, q1, t); }
-WP_API void builtin_quat_slerp_quatd_quatd_float64(quatd q0, quatd q1, float64 t, quatd* ret) { *ret = wp::quat_slerp(q0, q1, t); }
-WP_API void builtin_quat_to_matrix_quath(quath q, mat33h* ret) { *ret = wp::quat_to_matrix(q); }
-WP_API void builtin_quat_to_matrix_quatf(quatf q, mat33f* ret) { *ret = wp::quat_to_matrix(q); }
-WP_API void builtin_quat_to_matrix_quatd(quatd q, mat33d* ret) { *ret = wp::quat_to_matrix(q); }
+WP_API void builtin_quat_inverse_quath(quath& q, quath* ret) { *ret = wp::quat_inverse(q); }
+WP_API void builtin_quat_inverse_quatf(quatf& q, quatf* ret) { *ret = wp::quat_inverse(q); }
+WP_API void builtin_quat_inverse_quatd(quatd& q, quatd* ret) { *ret = wp::quat_inverse(q); }
+WP_API void builtin_quat_rotate_quath_vec3h(quath& q, vec3h& p, vec3h* ret) { *ret = wp::quat_rotate(q, p); }
+WP_API void builtin_quat_rotate_quatf_vec3f(quatf& q, vec3f& p, vec3f* ret) { *ret = wp::quat_rotate(q, p); }
+WP_API void builtin_quat_rotate_quatd_vec3d(quatd& q, vec3d& p, vec3d* ret) { *ret = wp::quat_rotate(q, p); }
+WP_API void builtin_quat_rotate_inv_quath_vec3h(quath& q, vec3h& p, vec3h* ret) { *ret = wp::quat_rotate_inv(q, p); }
+WP_API void builtin_quat_rotate_inv_quatf_vec3f(quatf& q, vec3f& p, vec3f* ret) { *ret = wp::quat_rotate_inv(q, p); }
+WP_API void builtin_quat_rotate_inv_quatd_vec3d(quatd& q, vec3d& p, vec3d* ret) { *ret = wp::quat_rotate_inv(q, p); }
+WP_API void builtin_quat_slerp_quath_quath_float16(quath& q0, quath& q1, float16 t, quath* ret) { *ret = wp::quat_slerp(q0, q1, t); }
+WP_API void builtin_quat_slerp_quatf_quatf_float32(quatf& q0, quatf& q1, float32 t, quatf* ret) { *ret = wp::quat_slerp(q0, q1, t); }
+WP_API void builtin_quat_slerp_quatd_quatd_float64(quatd& q0, quatd& q1, float64 t, quatd* ret) { *ret = wp::quat_slerp(q0, q1, t); }
+WP_API void builtin_quat_to_matrix_quath(quath& q, mat33h* ret) { *ret = wp::quat_to_matrix(q); }
+WP_API void builtin_quat_to_matrix_quatf(quatf& q, mat33f* ret) { *ret = wp::quat_to_matrix(q); }
+WP_API void builtin_quat_to_matrix_quatd(quatd& q, mat33d* ret) { *ret = wp::quat_to_matrix(q); }
 WP_API void builtin_transform_identity(transformf* ret) { *ret = wp::transform_identity(); }
-WP_API void builtin_transform_get_translation_transformh(transformh t, vec3h* ret) { *ret = wp::transform_get_translation(t); }
-WP_API void builtin_transform_get_translation_transformf(transformf t, vec3f* ret) { *ret = wp::transform_get_translation(t); }
-WP_API void builtin_transform_get_translation_transformd(transformd t, vec3d* ret) { *ret = wp::transform_get_translation(t); }
-WP_API void builtin_transform_get_rotation_transformh(transformh t, quath* ret) { *ret = wp::transform_get_rotation(t); }
-WP_API void builtin_transform_get_rotation_transformf(transformf t, quatf* ret) { *ret = wp::transform_get_rotation(t); }
-WP_API void builtin_transform_get_rotation_transformd(transformd t, quatd* ret) { *ret = wp::transform_get_rotation(t); }
-WP_API void builtin_transform_multiply_transformh_transformh(transformh a, transformh b, transformh* ret) { *ret = wp::transform_multiply(a, b); }
-WP_API void builtin_transform_multiply_transformf_transformf(transformf a, transformf b, transformf* ret) { *ret = wp::transform_multiply(a, b); }
-WP_API void builtin_transform_multiply_transformd_transformd(transformd a, transformd b, transformd* ret) { *ret = wp::transform_multiply(a, b); }
-WP_API void builtin_transform_point_transformh_vec3h(transformh t, vec3h p, vec3h* ret) { *ret = wp::transform_point(t, p); }
-WP_API void builtin_transform_point_transformf_vec3f(transformf t, vec3f p, vec3f* ret) { *ret = wp::transform_point(t, p); }
-WP_API void builtin_transform_point_transformd_vec3d(transformd t, vec3d p, vec3d* ret) { *ret = wp::transform_point(t, p); }
-WP_API void builtin_transform_point_mat44h_vec3h(mat44h m, vec3h p, vec3h* ret) { *ret = wp::transform_point(m, p); }
-WP_API void builtin_transform_point_mat44f_vec3f(mat44f m, vec3f p, vec3f* ret) { *ret = wp::transform_point(m, p); }
-WP_API void builtin_transform_point_mat44d_vec3d(mat44d m, vec3d p, vec3d* ret) { *ret = wp::transform_point(m, p); }
-WP_API void builtin_transform_vector_transformh_vec3h(transformh t, vec3h v, vec3h* ret) { *ret = wp::transform_vector(t, v); }
-WP_API void builtin_transform_vector_transformf_vec3f(transformf t, vec3f v, vec3f* ret) { *ret = wp::transform_vector(t, v); }
-WP_API void builtin_transform_vector_transformd_vec3d(transformd t, vec3d v, vec3d* ret) { *ret = wp::transform_vector(t, v); }
-WP_API void builtin_transform_vector_mat44h_vec3h(mat44h m, vec3h v, vec3h* ret) { *ret = wp::transform_vector(m, v); }
-WP_API void builtin_transform_vector_mat44f_vec3f(mat44f m, vec3f v, vec3f* ret) { *ret = wp::transform_vector(m, v); }
-WP_API void builtin_transform_vector_mat44d_vec3d(mat44d m, vec3d v, vec3d* ret) { *ret = wp::transform_vector(m, v); }
-WP_API void builtin_transform_inverse_transformh(transformh t, transformh* ret) { *ret = wp::transform_inverse(t); }
-WP_API void builtin_transform_inverse_transformf(transformf t, transformf* ret) { *ret = wp::transform_inverse(t); }
-WP_API void builtin_transform_inverse_transformd(transformd t, transformd* ret) { *ret = wp::transform_inverse(t); }
-WP_API void builtin_spatial_dot_spatial_vectorh_spatial_vectorh(spatial_vectorh a, spatial_vectorh b, float16* ret) { *ret = wp::spatial_dot(a, b); }
-WP_API void builtin_spatial_dot_spatial_vectorf_spatial_vectorf(spatial_vectorf a, spatial_vectorf b, float32* ret) { *ret = wp::spatial_dot(a, b); }
-WP_API void builtin_spatial_dot_spatial_vectord_spatial_vectord(spatial_vectord a, spatial_vectord b, float64* ret) { *ret = wp::spatial_dot(a, b); }
-WP_API void builtin_spatial_cross_spatial_vectorh_spatial_vectorh(spatial_vectorh a, spatial_vectorh b, spatial_vectorh* ret) { *ret = wp::spatial_cross(a, b); }
-WP_API void builtin_spatial_cross_spatial_vectorf_spatial_vectorf(spatial_vectorf a, spatial_vectorf b, spatial_vectorf* ret) { *ret = wp::spatial_cross(a, b); }
-WP_API void builtin_spatial_cross_spatial_vectord_spatial_vectord(spatial_vectord a, spatial_vectord b, spatial_vectord* ret) { *ret = wp::spatial_cross(a, b); }
-WP_API void builtin_spatial_cross_dual_spatial_vectorh_spatial_vectorh(spatial_vectorh a, spatial_vectorh b, spatial_vectorh* ret) { *ret = wp::spatial_cross_dual(a, b); }
-WP_API void builtin_spatial_cross_dual_spatial_vectorf_spatial_vectorf(spatial_vectorf a, spatial_vectorf b, spatial_vectorf* ret) { *ret = wp::spatial_cross_dual(a, b); }
-WP_API void builtin_spatial_cross_dual_spatial_vectord_spatial_vectord(spatial_vectord a, spatial_vectord b, spatial_vectord* ret) { *ret = wp::spatial_cross_dual(a, b); }
-WP_API void builtin_spatial_top_spatial_vectorh(spatial_vectorh a, vec3h* ret) { *ret = wp::spatial_top(a); }
-WP_API void builtin_spatial_top_spatial_vectorf(spatial_vectorf a, vec3f* ret) { *ret = wp::spatial_top(a); }
-WP_API void builtin_spatial_top_spatial_vectord(spatial_vectord a, vec3d* ret) { *ret = wp::spatial_top(a); }
-WP_API void builtin_spatial_bottom_spatial_vectorh(spatial_vectorh a, vec3h* ret) { *ret = wp::spatial_bottom(a); }
-WP_API void builtin_spatial_bottom_spatial_vectorf(spatial_vectorf a, vec3f* ret) { *ret = wp::spatial_bottom(a); }
-WP_API void builtin_spatial_bottom_spatial_vectord(spatial_vectord a, vec3d* ret) { *ret = wp::spatial_bottom(a); }
-WP_API void builtin_bvh_query_aabb_uint64_vec3f_vec3f(uint64 id, vec3f lower, vec3f upper, bvh_query_t* ret) { *ret = wp::bvh_query_aabb(id, lower, upper); }
-WP_API void builtin_bvh_query_ray_uint64_vec3f_vec3f(uint64 id, vec3f start, vec3f dir, bvh_query_t* ret) { *ret = wp::bvh_query_ray(id, start, dir); }
+WP_API void builtin_transform_get_translation_transformh(transformh& t, vec3h* ret) { *ret = wp::transform_get_translation(t); }
+WP_API void builtin_transform_get_translation_transformf(transformf& t, vec3f* ret) { *ret = wp::transform_get_translation(t); }
+WP_API void builtin_transform_get_translation_transformd(transformd& t, vec3d* ret) { *ret = wp::transform_get_translation(t); }
+WP_API void builtin_transform_get_rotation_transformh(transformh& t, quath* ret) { *ret = wp::transform_get_rotation(t); }
+WP_API void builtin_transform_get_rotation_transformf(transformf& t, quatf* ret) { *ret = wp::transform_get_rotation(t); }
+WP_API void builtin_transform_get_rotation_transformd(transformd& t, quatd* ret) { *ret = wp::transform_get_rotation(t); }
+WP_API void builtin_transform_multiply_transformh_transformh(transformh& a, transformh& b, transformh* ret) { *ret = wp::transform_multiply(a, b); }
+WP_API void builtin_transform_multiply_transformf_transformf(transformf& a, transformf& b, transformf* ret) { *ret = wp::transform_multiply(a, b); }
+WP_API void builtin_transform_multiply_transformd_transformd(transformd& a, transformd& b, transformd* ret) { *ret = wp::transform_multiply(a, b); }
+WP_API void builtin_transform_point_transformh_vec3h(transformh& t, vec3h& p, vec3h* ret) { *ret = wp::transform_point(t, p); }
+WP_API void builtin_transform_point_transformf_vec3f(transformf& t, vec3f& p, vec3f* ret) { *ret = wp::transform_point(t, p); }
+WP_API void builtin_transform_point_transformd_vec3d(transformd& t, vec3d& p, vec3d* ret) { *ret = wp::transform_point(t, p); }
+WP_API void builtin_transform_point_mat44h_vec3h(mat44h& m, vec3h& p, vec3h* ret) { *ret = wp::transform_point(m, p); }
+WP_API void builtin_transform_point_mat44f_vec3f(mat44f& m, vec3f& p, vec3f* ret) { *ret = wp::transform_point(m, p); }
+WP_API void builtin_transform_point_mat44d_vec3d(mat44d& m, vec3d& p, vec3d* ret) { *ret = wp::transform_point(m, p); }
+WP_API void builtin_transform_vector_transformh_vec3h(transformh& t, vec3h& v, vec3h* ret) { *ret = wp::transform_vector(t, v); }
+WP_API void builtin_transform_vector_transformf_vec3f(transformf& t, vec3f& v, vec3f* ret) { *ret = wp::transform_vector(t, v); }
+WP_API void builtin_transform_vector_transformd_vec3d(transformd& t, vec3d& v, vec3d* ret) { *ret = wp::transform_vector(t, v); }
+WP_API void builtin_transform_vector_mat44h_vec3h(mat44h& m, vec3h& v, vec3h* ret) { *ret = wp::transform_vector(m, v); }
+WP_API void builtin_transform_vector_mat44f_vec3f(mat44f& m, vec3f& v, vec3f* ret) { *ret = wp::transform_vector(m, v); }
+WP_API void builtin_transform_vector_mat44d_vec3d(mat44d& m, vec3d& v, vec3d* ret) { *ret = wp::transform_vector(m, v); }
+WP_API void builtin_transform_inverse_transformh(transformh& t, transformh* ret) { *ret = wp::transform_inverse(t); }
+WP_API void builtin_transform_inverse_transformf(transformf& t, transformf* ret) { *ret = wp::transform_inverse(t); }
+WP_API void builtin_transform_inverse_transformd(transformd& t, transformd* ret) { *ret = wp::transform_inverse(t); }
+WP_API void builtin_spatial_dot_spatial_vectorh_spatial_vectorh(spatial_vectorh& a, spatial_vectorh& b, float16* ret) { *ret = wp::spatial_dot(a, b); }
+WP_API void builtin_spatial_dot_spatial_vectorf_spatial_vectorf(spatial_vectorf& a, spatial_vectorf& b, float32* ret) { *ret = wp::spatial_dot(a, b); }
+WP_API void builtin_spatial_dot_spatial_vectord_spatial_vectord(spatial_vectord& a, spatial_vectord& b, float64* ret) { *ret = wp::spatial_dot(a, b); }
+WP_API void builtin_spatial_cross_spatial_vectorh_spatial_vectorh(spatial_vectorh& a, spatial_vectorh& b, spatial_vectorh* ret) { *ret = wp::spatial_cross(a, b); }
+WP_API void builtin_spatial_cross_spatial_vectorf_spatial_vectorf(spatial_vectorf& a, spatial_vectorf& b, spatial_vectorf* ret) { *ret = wp::spatial_cross(a, b); }
+WP_API void builtin_spatial_cross_spatial_vectord_spatial_vectord(spatial_vectord& a, spatial_vectord& b, spatial_vectord* ret) { *ret = wp::spatial_cross(a, b); }
+WP_API void builtin_spatial_cross_dual_spatial_vectorh_spatial_vectorh(spatial_vectorh& a, spatial_vectorh& b, spatial_vectorh* ret) { *ret = wp::spatial_cross_dual(a, b); }
+WP_API void builtin_spatial_cross_dual_spatial_vectorf_spatial_vectorf(spatial_vectorf& a, spatial_vectorf& b, spatial_vectorf* ret) { *ret = wp::spatial_cross_dual(a, b); }
+WP_API void builtin_spatial_cross_dual_spatial_vectord_spatial_vectord(spatial_vectord& a, spatial_vectord& b, spatial_vectord* ret) { *ret = wp::spatial_cross_dual(a, b); }
+WP_API void builtin_spatial_top_spatial_vectorh(spatial_vectorh& a, vec3h* ret) { *ret = wp::spatial_top(a); }
+WP_API void builtin_spatial_top_spatial_vectorf(spatial_vectorf& a, vec3f* ret) { *ret = wp::spatial_top(a); }
+WP_API void builtin_spatial_top_spatial_vectord(spatial_vectord& a, vec3d* ret) { *ret = wp::spatial_top(a); }
+WP_API void builtin_spatial_bottom_spatial_vectorh(spatial_vectorh& a, vec3h* ret) { *ret = wp::spatial_bottom(a); }
+WP_API void builtin_spatial_bottom_spatial_vectorf(spatial_vectorf& a, vec3f* ret) { *ret = wp::spatial_bottom(a); }
+WP_API void builtin_spatial_bottom_spatial_vectord(spatial_vectord& a, vec3d* ret) { *ret = wp::spatial_bottom(a); }
+WP_API void builtin_bvh_query_aabb_uint64_vec3f_vec3f(uint64 id, vec3f& lower, vec3f& upper, bvh_query_t* ret) { *ret = wp::bvh_query_aabb(id, lower, upper); }
+WP_API void builtin_bvh_query_ray_uint64_vec3f_vec3f(uint64 id, vec3f& start, vec3f& dir, bvh_query_t* ret) { *ret = wp::bvh_query_ray(id, start, dir); }
 WP_API void builtin_bvh_query_next_bvh_query_t_int32(bvh_query_t query, int32 index, bool* ret) { *ret = wp::bvh_query_next(query, index); }
-WP_API void builtin_mesh_query_point_uint64_vec3f_float32_float32_int32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_point(id, point, max_dist, inside, face, bary_u, bary_v); }
-WP_API void builtin_mesh_query_point_uint64_vec3f_float32(uint64 id, vec3f point, float32 max_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_point(id, point, max_dist); }
-WP_API void builtin_mesh_query_point_no_sign_uint64_vec3f_float32_int32_float32_float32(uint64 id, vec3f point, float32 max_dist, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_point_no_sign(id, point, max_dist, face, bary_u, bary_v); }
-WP_API void builtin_mesh_query_point_no_sign_uint64_vec3f_float32(uint64 id, vec3f point, float32 max_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_no_sign(id, point, max_dist); }
-WP_API void builtin_mesh_query_furthest_point_no_sign_uint64_vec3f_float32_int32_float32_float32(uint64 id, vec3f point, float32 min_dist, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_furthest_point_no_sign(id, point, min_dist, face, bary_u, bary_v); }
-WP_API void builtin_mesh_query_furthest_point_no_sign_uint64_vec3f_float32(uint64 id, vec3f point, float32 min_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_furthest_point_no_sign(id, point, min_dist); }
-WP_API void builtin_mesh_query_point_sign_normal_uint64_vec3f_float32_float32_int32_float32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, float32 epsilon, bool* ret) { *ret = wp::mesh_query_point_sign_normal(id, point, max_dist, inside, face, bary_u, bary_v, epsilon); }
-WP_API void builtin_mesh_query_point_sign_normal_uint64_vec3f_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 epsilon, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_sign_normal(id, point, max_dist, epsilon); }
-WP_API void builtin_mesh_query_point_sign_winding_number_uint64_vec3f_float32_float32_int32_float32_float32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, float32 accuracy, float32 threshold, bool* ret) { *ret = wp::mesh_query_point_sign_winding_number(id, point, max_dist, inside, face, bary_u, bary_v, accuracy, threshold); }
-WP_API void builtin_mesh_query_point_sign_winding_number_uint64_vec3f_float32_float32_float32(uint64 id, vec3f point, float32 max_dist, float32 accuracy, float32 threshold, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_sign_winding_number(id, point, max_dist, accuracy, threshold); }
-WP_API void builtin_mesh_query_ray_uint64_vec3f_vec3f_float32_float32_float32_float32_float32_vec3f_int32(uint64 id, vec3f start, vec3f dir, float32 max_t, float32 t, float32 bary_u, float32 bary_v, float32 sign, vec3f normal, int32 face, bool* ret) { *ret = wp::mesh_query_ray(id, start, dir, max_t, t, bary_u, bary_v, sign, normal, face); }
-WP_API void builtin_mesh_query_ray_uint64_vec3f_vec3f_float32(uint64 id, vec3f start, vec3f dir, float32 max_t, mesh_query_ray_t* ret) { *ret = wp::mesh_query_ray(id, start, dir, max_t); }
-WP_API void builtin_mesh_query_aabb_uint64_vec3f_vec3f(uint64 id, vec3f lower, vec3f upper, mesh_query_aabb_t* ret) { *ret = wp::mesh_query_aabb(id, lower, upper); }
+WP_API void builtin_mesh_query_point_uint64_vec3f_float32_float32_int32_float32_float32(uint64 id, vec3f& point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_point(id, point, max_dist, inside, face, bary_u, bary_v); }
+WP_API void builtin_mesh_query_point_uint64_vec3f_float32(uint64 id, vec3f& point, float32 max_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_point(id, point, max_dist); }
+WP_API void builtin_mesh_query_point_no_sign_uint64_vec3f_float32_int32_float32_float32(uint64 id, vec3f& point, float32 max_dist, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_point_no_sign(id, point, max_dist, face, bary_u, bary_v); }
+WP_API void builtin_mesh_query_point_no_sign_uint64_vec3f_float32(uint64 id, vec3f& point, float32 max_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_no_sign(id, point, max_dist); }
+WP_API void builtin_mesh_query_furthest_point_no_sign_uint64_vec3f_float32_int32_float32_float32(uint64 id, vec3f& point, float32 min_dist, int32 face, float32 bary_u, float32 bary_v, bool* ret) { *ret = wp::mesh_query_furthest_point_no_sign(id, point, min_dist, face, bary_u, bary_v); }
+WP_API void builtin_mesh_query_furthest_point_no_sign_uint64_vec3f_float32(uint64 id, vec3f& point, float32 min_dist, mesh_query_point_t* ret) { *ret = wp::mesh_query_furthest_point_no_sign(id, point, min_dist); }
+WP_API void builtin_mesh_query_point_sign_normal_uint64_vec3f_float32_float32_int32_float32_float32_float32(uint64 id, vec3f& point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, float32 epsilon, bool* ret) { *ret = wp::mesh_query_point_sign_normal(id, point, max_dist, inside, face, bary_u, bary_v, epsilon); }
+WP_API void builtin_mesh_query_point_sign_normal_uint64_vec3f_float32_float32(uint64 id, vec3f& point, float32 max_dist, float32 epsilon, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_sign_normal(id, point, max_dist, epsilon); }
+WP_API void builtin_mesh_query_point_sign_winding_number_uint64_vec3f_float32_float32_int32_float32_float32_float32_float32(uint64 id, vec3f& point, float32 max_dist, float32 inside, int32 face, float32 bary_u, float32 bary_v, float32 accuracy, float32 threshold, bool* ret) { *ret = wp::mesh_query_point_sign_winding_number(id, point, max_dist, inside, face, bary_u, bary_v, accuracy, threshold); }
+WP_API void builtin_mesh_query_point_sign_winding_number_uint64_vec3f_float32_float32_float32(uint64 id, vec3f& point, float32 max_dist, float32 accuracy, float32 threshold, mesh_query_point_t* ret) { *ret = wp::mesh_query_point_sign_winding_number(id, point, max_dist, accuracy, threshold); }
+WP_API void builtin_mesh_query_ray_uint64_vec3f_vec3f_float32_float32_float32_float32_float32_vec3f_int32(uint64 id, vec3f& start, vec3f& dir, float32 max_t, float32 t, float32 bary_u, float32 bary_v, float32 sign, vec3f& normal, int32 face, bool* ret) { *ret = wp::mesh_query_ray(id, start, dir, max_t, t, bary_u, bary_v, sign, normal, face); }
+WP_API void builtin_mesh_query_ray_uint64_vec3f_vec3f_float32(uint64 id, vec3f& start, vec3f& dir, float32 max_t, mesh_query_ray_t* ret) { *ret = wp::mesh_query_ray(id, start, dir, max_t); }
+WP_API void builtin_mesh_query_aabb_uint64_vec3f_vec3f(uint64 id, vec3f& lower, vec3f& upper, mesh_query_aabb_t* ret) { *ret = wp::mesh_query_aabb(id, lower, upper); }
 WP_API void builtin_mesh_query_aabb_next_mesh_query_aabb_t_int32(mesh_query_aabb_t query, int32 index, bool* ret) { *ret = wp::mesh_query_aabb_next(query, index); }
 WP_API void builtin_mesh_eval_position_uint64_int32_float32_float32(uint64 id, int32 face, float32 bary_u, float32 bary_v, vec3f* ret) { *ret = wp::mesh_eval_position(id, face, bary_u, bary_v); }
 WP_API void builtin_mesh_eval_velocity_uint64_int32_float32_float32(uint64 id, int32 face, float32 bary_u, float32 bary_v, vec3f* ret) { *ret = wp::mesh_eval_velocity(id, face, bary_u, bary_v); }
-WP_API void builtin_hash_grid_query_uint64_vec3f_float32(uint64 id, vec3f point, float32 max_dist, hash_grid_query_t* ret) { *ret = wp::hash_grid_query(id, point, max_dist); }
+WP_API void builtin_hash_grid_query_uint64_vec3f_float32(uint64 id, vec3f& point, float32 max_dist, hash_grid_query_t* ret) { *ret = wp::hash_grid_query(id, point, max_dist); }
 WP_API void builtin_hash_grid_query_next_hash_grid_query_t_int32(hash_grid_query_t query, int32 index, bool* ret) { *ret = wp::hash_grid_query_next(query, index); }
 WP_API void builtin_hash_grid_point_id_uint64_int32(uint64 id, int32 index, int* ret) { *ret = wp::hash_grid_point_id(id, index); }
-WP_API void builtin_intersect_tri_tri_vec3f_vec3f_vec3f_vec3f_vec3f_vec3f(vec3f v0, vec3f v1, vec3f v2, vec3f u0, vec3f u1, vec3f u2, int* ret) { *ret = wp::intersect_tri_tri(v0, v1, v2, u0, u1, u2); }
+WP_API void builtin_intersect_tri_tri_vec3f_vec3f_vec3f_vec3f_vec3f_vec3f(vec3f& v0, vec3f& v1, vec3f& v2, vec3f& u0, vec3f& u1, vec3f& u2, int* ret) { *ret = wp::intersect_tri_tri(v0, v1, v2, u0, u1, u2); }
 WP_API void builtin_mesh_get_uint64(uint64 id, Mesh* ret) { *ret = wp::mesh_get(id); }
 WP_API void builtin_mesh_eval_face_normal_uint64_int32(uint64 id, int32 face, vec3f* ret) { *ret = wp::mesh_eval_face_normal(id, face); }
 WP_API void builtin_mesh_get_point_uint64_int32(uint64 id, int32 index, vec3f* ret) { *ret = wp::mesh_get_point(id, index); }
 WP_API void builtin_mesh_get_velocity_uint64_int32(uint64 id, int32 index, vec3f* ret) { *ret = wp::mesh_get_velocity(id, index); }
 WP_API void builtin_mesh_get_index_uint64_int32(uint64 id, int32 index, int* ret) { *ret = wp::mesh_get_index(id, index); }
-WP_API void builtin_closest_point_edge_edge_vec3f_vec3f_vec3f_vec3f_float32(vec3f p1, vec3f q1, vec3f p2, vec3f q2, float32 epsilon, vec3f* ret) { *ret = wp::closest_point_edge_edge(p1, q1, p2, q2, epsilon); }
+WP_API void builtin_closest_point_edge_edge_vec3f_vec3f_vec3f_vec3f_float32(vec3f& p1, vec3f& q1, vec3f& p2, vec3f& q2, float32 epsilon, vec3f* ret) { *ret = wp::closest_point_edge_edge(p1, q1, p2, q2, epsilon); }
 WP_API void builtin_iter_next_range_t(range_t range, int* ret) { *ret = wp::iter_next(range); }
 WP_API void builtin_iter_next_hash_grid_query_t(hash_grid_query_t query, int* ret) { *ret = wp::iter_next(query); }
 WP_API void builtin_iter_next_mesh_query_aabb_t(mesh_query_aabb_t query, int* ret) { *ret = wp::iter_next(query); }
-WP_API void builtin_volume_sample_f_uint64_vec3f_int32(uint64 id, vec3f uvw, int32 sampling_mode, float* ret) { *ret = wp::volume_sample_f(id, uvw, sampling_mode); }
-WP_API void builtin_volume_sample_grad_f_uint64_vec3f_int32_vec3f(uint64 id, vec3f uvw, int32 sampling_mode, vec3f grad, float* ret) { *ret = wp::volume_sample_grad_f(id, uvw, sampling_mode, grad); }
+WP_API void builtin_volume_sample_f_uint64_vec3f_int32(uint64 id, vec3f& uvw, int32 sampling_mode, float* ret) { *ret = wp::volume_sample_f(id, uvw, sampling_mode); }
+WP_API void builtin_volume_sample_grad_f_uint64_vec3f_int32_vec3f(uint64 id, vec3f& uvw, int32 sampling_mode, vec3f& grad, float* ret) { *ret = wp::volume_sample_grad_f(id, uvw, sampling_mode, grad); }
 WP_API void builtin_volume_lookup_f_uint64_int32_int32_int32(uint64 id, int32 i, int32 j, int32 k, float* ret) { *ret = wp::volume_lookup_f(id, i, j, k); }
-WP_API void builtin_volume_sample_v_uint64_vec3f_int32(uint64 id, vec3f uvw, int32 sampling_mode, vec3f* ret) { *ret = wp::volume_sample_v(id, uvw, sampling_mode); }
+WP_API void builtin_volume_sample_v_uint64_vec3f_int32(uint64 id, vec3f& uvw, int32 sampling_mode, vec3f* ret) { *ret = wp::volume_sample_v(id, uvw, sampling_mode); }
 WP_API void builtin_volume_lookup_v_uint64_int32_int32_int32(uint64 id, int32 i, int32 j, int32 k, vec3f* ret) { *ret = wp::volume_lookup_v(id, i, j, k); }
-WP_API void builtin_volume_sample_i_uint64_vec3f(uint64 id, vec3f uvw, int* ret) { *ret = wp::volume_sample_i(id, uvw); }
+WP_API void builtin_volume_sample_i_uint64_vec3f(uint64 id, vec3f& uvw, int* ret) { *ret = wp::volume_sample_i(id, uvw); }
 WP_API void builtin_volume_lookup_i_uint64_int32_int32_int32(uint64 id, int32 i, int32 j, int32 k, int* ret) { *ret = wp::volume_lookup_i(id, i, j, k); }
-WP_API void builtin_volume_index_to_world_uint64_vec3f(uint64 id, vec3f uvw, vec3f* ret) { *ret = wp::volume_index_to_world(id, uvw); }
-WP_API void builtin_volume_world_to_index_uint64_vec3f(uint64 id, vec3f xyz, vec3f* ret) { *ret = wp::volume_world_to_index(id, xyz); }
-WP_API void builtin_volume_index_to_world_dir_uint64_vec3f(uint64 id, vec3f uvw, vec3f* ret) { *ret = wp::volume_index_to_world_dir(id, uvw); }
-WP_API void builtin_volume_world_to_index_dir_uint64_vec3f(uint64 id, vec3f xyz, vec3f* ret) { *ret = wp::volume_world_to_index_dir(id, xyz); }
+WP_API void builtin_volume_index_to_world_uint64_vec3f(uint64 id, vec3f& uvw, vec3f* ret) { *ret = wp::volume_index_to_world(id, uvw); }
+WP_API void builtin_volume_world_to_index_uint64_vec3f(uint64 id, vec3f& xyz, vec3f* ret) { *ret = wp::volume_world_to_index(id, xyz); }
+WP_API void builtin_volume_index_to_world_dir_uint64_vec3f(uint64 id, vec3f& uvw, vec3f* ret) { *ret = wp::volume_index_to_world_dir(id, uvw); }
+WP_API void builtin_volume_world_to_index_dir_uint64_vec3f(uint64 id, vec3f& xyz, vec3f* ret) { *ret = wp::volume_world_to_index_dir(id, xyz); }
 WP_API void builtin_rand_init_int32(int32 seed, uint32* ret) { *ret = wp::rand_init(seed); }
 WP_API void builtin_rand_init_int32_int32(int32 seed, int32 offset, uint32* ret) { *ret = wp::rand_init(seed, offset); }
 WP_API void builtin_randi_uint32(uint32 state, int* ret) { *ret = wp::randi(state); }
@@ -804,116 +804,116 @@ WP_API void builtin_sample_unit_square_uint32(uint32 state, vec2f* ret) { *ret =
 WP_API void builtin_sample_unit_cube_uint32(uint32 state, vec3f* ret) { *ret = wp::sample_unit_cube(state); }
 WP_API void builtin_poisson_uint32_float32(uint32 state, float32 lam, uint32* ret) { *ret = wp::poisson(state, lam); }
 WP_API void builtin_noise_uint32_float32(uint32 state, float32 x, float* ret) { *ret = wp::noise(state, x); }
-WP_API void builtin_noise_uint32_vec2f(uint32 state, vec2f xy, float* ret) { *ret = wp::noise(state, xy); }
-WP_API void builtin_noise_uint32_vec3f(uint32 state, vec3f xyz, float* ret) { *ret = wp::noise(state, xyz); }
-WP_API void builtin_noise_uint32_vec4f(uint32 state, vec4f xyzt, float* ret) { *ret = wp::noise(state, xyzt); }
+WP_API void builtin_noise_uint32_vec2f(uint32 state, vec2f& xy, float* ret) { *ret = wp::noise(state, xy); }
+WP_API void builtin_noise_uint32_vec3f(uint32 state, vec3f& xyz, float* ret) { *ret = wp::noise(state, xyz); }
+WP_API void builtin_noise_uint32_vec4f(uint32 state, vec4f& xyzt, float* ret) { *ret = wp::noise(state, xyzt); }
 WP_API void builtin_pnoise_uint32_float32_int32(uint32 state, float32 x, int32 px, float* ret) { *ret = wp::pnoise(state, x, px); }
-WP_API void builtin_pnoise_uint32_vec2f_int32_int32(uint32 state, vec2f xy, int32 px, int32 py, float* ret) { *ret = wp::pnoise(state, xy, px, py); }
-WP_API void builtin_pnoise_uint32_vec3f_int32_int32_int32(uint32 state, vec3f xyz, int32 px, int32 py, int32 pz, float* ret) { *ret = wp::pnoise(state, xyz, px, py, pz); }
-WP_API void builtin_pnoise_uint32_vec4f_int32_int32_int32_int32(uint32 state, vec4f xyzt, int32 px, int32 py, int32 pz, int32 pt, float* ret) { *ret = wp::pnoise(state, xyzt, px, py, pz, pt); }
-WP_API void builtin_curlnoise_uint32_vec2f_uint32_float32_float32(uint32 state, vec2f xy, uint32 octaves, float32 lacunarity, float32 gain, vec2f* ret) { *ret = wp::curlnoise(state, xy, octaves, lacunarity, gain); }
-WP_API void builtin_curlnoise_uint32_vec3f_uint32_float32_float32(uint32 state, vec3f xyz, uint32 octaves, float32 lacunarity, float32 gain, vec3f* ret) { *ret = wp::curlnoise(state, xyz, octaves, lacunarity, gain); }
-WP_API void builtin_curlnoise_uint32_vec4f_uint32_float32_float32(uint32 state, vec4f xyzt, uint32 octaves, float32 lacunarity, float32 gain, vec3f* ret) { *ret = wp::curlnoise(state, xyzt, octaves, lacunarity, gain); }
-WP_API void builtin_extract_vec2h_int32(vec2h a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3h_int32(vec3h a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4h_int32(vec4h a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_spatial_vectorh_int32(spatial_vectorh a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2f_int32(vec2f a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3f_int32(vec3f a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4f_int32(vec4f a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_spatial_vectorf_int32(spatial_vectorf a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2d_int32(vec2d a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3d_int32(vec3d a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4d_int32(vec4d a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_spatial_vectord_int32(spatial_vectord a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2s_int32(vec2s a, int32 i, int16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3s_int32(vec3s a, int32 i, int16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4s_int32(vec4s a, int32 i, int16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2i_int32(vec2i a, int32 i, int32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3i_int32(vec3i a, int32 i, int32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4i_int32(vec4i a, int32 i, int32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2l_int32(vec2l a, int32 i, int64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3l_int32(vec3l a, int32 i, int64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4l_int32(vec4l a, int32 i, int64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2b_int32(vec2b a, int32 i, int8* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3b_int32(vec3b a, int32 i, int8* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4b_int32(vec4b a, int32 i, int8* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2us_int32(vec2us a, int32 i, uint16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3us_int32(vec3us a, int32 i, uint16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4us_int32(vec4us a, int32 i, uint16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2ui_int32(vec2ui a, int32 i, uint32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3ui_int32(vec3ui a, int32 i, uint32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4ui_int32(vec4ui a, int32 i, uint32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2ul_int32(vec2ul a, int32 i, uint64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3ul_int32(vec3ul a, int32 i, uint64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4ul_int32(vec4ul a, int32 i, uint64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec2ub_int32(vec2ub a, int32 i, uint8* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec3ub_int32(vec3ub a, int32 i, uint8* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_vec4ub_int32(vec4ub a, int32 i, uint8* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_quath_int32(quath a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_quatf_int32(quatf a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_quatd_int32(quatd a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat22h_int32(mat22h a, int32 i, vec2h* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat33h_int32(mat33h a, int32 i, vec3h* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat44h_int32(mat44h a, int32 i, vec4h* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_spatial_matrixh_int32(spatial_matrixh a, int32 i, spatial_vectorh* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat22f_int32(mat22f a, int32 i, vec2f* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat33f_int32(mat33f a, int32 i, vec3f* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat44f_int32(mat44f a, int32 i, vec4f* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_spatial_matrixf_int32(spatial_matrixf a, int32 i, spatial_vectorf* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat22d_int32(mat22d a, int32 i, vec2d* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat33d_int32(mat33d a, int32 i, vec3d* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat44d_int32(mat44d a, int32 i, vec4d* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_spatial_matrixd_int32(spatial_matrixd a, int32 i, spatial_vectord* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_mat22h_int32_int32(mat22h a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat33h_int32_int32(mat33h a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat44h_int32_int32(mat44h a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_spatial_matrixh_int32_int32(spatial_matrixh a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat22f_int32_int32(mat22f a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat33f_int32_int32(mat33f a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat44f_int32_int32(mat44f a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_spatial_matrixf_int32_int32(spatial_matrixf a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat22d_int32_int32(mat22d a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat33d_int32_int32(mat33d a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_mat44d_int32_int32(mat44d a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_spatial_matrixd_int32_int32(spatial_matrixd a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
-WP_API void builtin_extract_transformh_int32(transformh a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_transformf_int32(transformf a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
-WP_API void builtin_extract_transformd_int32(transformd a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_pnoise_uint32_vec2f_int32_int32(uint32 state, vec2f& xy, int32 px, int32 py, float* ret) { *ret = wp::pnoise(state, xy, px, py); }
+WP_API void builtin_pnoise_uint32_vec3f_int32_int32_int32(uint32 state, vec3f& xyz, int32 px, int32 py, int32 pz, float* ret) { *ret = wp::pnoise(state, xyz, px, py, pz); }
+WP_API void builtin_pnoise_uint32_vec4f_int32_int32_int32_int32(uint32 state, vec4f& xyzt, int32 px, int32 py, int32 pz, int32 pt, float* ret) { *ret = wp::pnoise(state, xyzt, px, py, pz, pt); }
+WP_API void builtin_curlnoise_uint32_vec2f_uint32_float32_float32(uint32 state, vec2f& xy, uint32 octaves, float32 lacunarity, float32 gain, vec2f* ret) { *ret = wp::curlnoise(state, xy, octaves, lacunarity, gain); }
+WP_API void builtin_curlnoise_uint32_vec3f_uint32_float32_float32(uint32 state, vec3f& xyz, uint32 octaves, float32 lacunarity, float32 gain, vec3f* ret) { *ret = wp::curlnoise(state, xyz, octaves, lacunarity, gain); }
+WP_API void builtin_curlnoise_uint32_vec4f_uint32_float32_float32(uint32 state, vec4f& xyzt, uint32 octaves, float32 lacunarity, float32 gain, vec3f* ret) { *ret = wp::curlnoise(state, xyzt, octaves, lacunarity, gain); }
+WP_API void builtin_extract_vec2h_int32(vec2h& a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3h_int32(vec3h& a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4h_int32(vec4h& a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_spatial_vectorh_int32(spatial_vectorh& a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2f_int32(vec2f& a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3f_int32(vec3f& a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4f_int32(vec4f& a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_spatial_vectorf_int32(spatial_vectorf& a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2d_int32(vec2d& a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3d_int32(vec3d& a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4d_int32(vec4d& a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_spatial_vectord_int32(spatial_vectord& a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2s_int32(vec2s& a, int32 i, int16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3s_int32(vec3s& a, int32 i, int16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4s_int32(vec4s& a, int32 i, int16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2i_int32(vec2i& a, int32 i, int32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3i_int32(vec3i& a, int32 i, int32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4i_int32(vec4i& a, int32 i, int32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2l_int32(vec2l& a, int32 i, int64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3l_int32(vec3l& a, int32 i, int64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4l_int32(vec4l& a, int32 i, int64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2b_int32(vec2b& a, int32 i, int8* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3b_int32(vec3b& a, int32 i, int8* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4b_int32(vec4b& a, int32 i, int8* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2us_int32(vec2us& a, int32 i, uint16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3us_int32(vec3us& a, int32 i, uint16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4us_int32(vec4us& a, int32 i, uint16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2ui_int32(vec2ui& a, int32 i, uint32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3ui_int32(vec3ui& a, int32 i, uint32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4ui_int32(vec4ui& a, int32 i, uint32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2ul_int32(vec2ul& a, int32 i, uint64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3ul_int32(vec3ul& a, int32 i, uint64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4ul_int32(vec4ul& a, int32 i, uint64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec2ub_int32(vec2ub& a, int32 i, uint8* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec3ub_int32(vec3ub& a, int32 i, uint8* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_vec4ub_int32(vec4ub& a, int32 i, uint8* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_quath_int32(quath& a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_quatf_int32(quatf& a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_quatd_int32(quatd& a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat22h_int32(mat22h& a, int32 i, vec2h* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat33h_int32(mat33h& a, int32 i, vec3h* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat44h_int32(mat44h& a, int32 i, vec4h* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_spatial_matrixh_int32(spatial_matrixh& a, int32 i, spatial_vectorh* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat22f_int32(mat22f& a, int32 i, vec2f* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat33f_int32(mat33f& a, int32 i, vec3f* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat44f_int32(mat44f& a, int32 i, vec4f* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_spatial_matrixf_int32(spatial_matrixf& a, int32 i, spatial_vectorf* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat22d_int32(mat22d& a, int32 i, vec2d* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat33d_int32(mat33d& a, int32 i, vec3d* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat44d_int32(mat44d& a, int32 i, vec4d* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_spatial_matrixd_int32(spatial_matrixd& a, int32 i, spatial_vectord* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_mat22h_int32_int32(mat22h& a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat33h_int32_int32(mat33h& a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat44h_int32_int32(mat44h& a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_spatial_matrixh_int32_int32(spatial_matrixh& a, int32 i, int32 j, float16* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat22f_int32_int32(mat22f& a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat33f_int32_int32(mat33f& a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat44f_int32_int32(mat44f& a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_spatial_matrixf_int32_int32(spatial_matrixf& a, int32 i, int32 j, float32* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat22d_int32_int32(mat22d& a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat33d_int32_int32(mat33d& a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_mat44d_int32_int32(mat44d& a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_spatial_matrixd_int32_int32(spatial_matrixd& a, int32 i, int32 j, float64* ret) { *ret = wp::extract(a, i, j); }
+WP_API void builtin_extract_transformh_int32(transformh& a, int32 i, float16* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_transformf_int32(transformf& a, int32 i, float32* ret) { *ret = wp::extract(a, i); }
+WP_API void builtin_extract_transformd_int32(transformd& a, int32 i, float64* ret) { *ret = wp::extract(a, i); }
 WP_API void builtin_extract_shape_t_int32(shape_t s, int32 i, int* ret) { *ret = wp::extract(s, i); }
 WP_API void builtin_lerp_float16_float16_float16(float16 a, float16 b, float16 t, float16* ret) { *ret = wp::lerp(a, b, t); }
 WP_API void builtin_lerp_float32_float32_float32(float32 a, float32 b, float32 t, float32* ret) { *ret = wp::lerp(a, b, t); }
 WP_API void builtin_lerp_float64_float64_float64(float64 a, float64 b, float64 t, float64* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec2h_vec2h_float16(vec2h a, vec2h b, float16 t, vec2h* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec3h_vec3h_float16(vec3h a, vec3h b, float16 t, vec3h* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec4h_vec4h_float16(vec4h a, vec4h b, float16 t, vec4h* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_spatial_vectorh_spatial_vectorh_float16(spatial_vectorh a, spatial_vectorh b, float16 t, spatial_vectorh* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec2f_vec2f_float32(vec2f a, vec2f b, float32 t, vec2f* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec3f_vec3f_float32(vec3f a, vec3f b, float32 t, vec3f* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec4f_vec4f_float32(vec4f a, vec4f b, float32 t, vec4f* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_spatial_vectorf_spatial_vectorf_float32(spatial_vectorf a, spatial_vectorf b, float32 t, spatial_vectorf* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec2d_vec2d_float64(vec2d a, vec2d b, float64 t, vec2d* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec3d_vec3d_float64(vec3d a, vec3d b, float64 t, vec3d* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_vec4d_vec4d_float64(vec4d a, vec4d b, float64 t, vec4d* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_spatial_vectord_spatial_vectord_float64(spatial_vectord a, spatial_vectord b, float64 t, spatial_vectord* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat22h_mat22h_float16(mat22h a, mat22h b, float16 t, mat22h* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat33h_mat33h_float16(mat33h a, mat33h b, float16 t, mat33h* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat44h_mat44h_float16(mat44h a, mat44h b, float16 t, mat44h* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_spatial_matrixh_spatial_matrixh_float16(spatial_matrixh a, spatial_matrixh b, float16 t, spatial_matrixh* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat22f_mat22f_float32(mat22f a, mat22f b, float32 t, mat22f* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat33f_mat33f_float32(mat33f a, mat33f b, float32 t, mat33f* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat44f_mat44f_float32(mat44f a, mat44f b, float32 t, mat44f* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_spatial_matrixf_spatial_matrixf_float32(spatial_matrixf a, spatial_matrixf b, float32 t, spatial_matrixf* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat22d_mat22d_float64(mat22d a, mat22d b, float64 t, mat22d* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat33d_mat33d_float64(mat33d a, mat33d b, float64 t, mat33d* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_mat44d_mat44d_float64(mat44d a, mat44d b, float64 t, mat44d* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_spatial_matrixd_spatial_matrixd_float64(spatial_matrixd a, spatial_matrixd b, float64 t, spatial_matrixd* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_quath_quath_float16(quath a, quath b, float16 t, quath* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_quatf_quatf_float32(quatf a, quatf b, float32 t, quatf* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_quatd_quatd_float64(quatd a, quatd b, float64 t, quatd* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_transformh_transformh_float16(transformh a, transformh b, float16 t, transformh* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_transformf_transformf_float32(transformf a, transformf b, float32 t, transformf* ret) { *ret = wp::lerp(a, b, t); }
-WP_API void builtin_lerp_transformd_transformd_float64(transformd a, transformd b, float64 t, transformd* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec2h_vec2h_float16(vec2h& a, vec2h& b, float16 t, vec2h* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec3h_vec3h_float16(vec3h& a, vec3h& b, float16 t, vec3h* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec4h_vec4h_float16(vec4h& a, vec4h& b, float16 t, vec4h* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_spatial_vectorh_spatial_vectorh_float16(spatial_vectorh& a, spatial_vectorh& b, float16 t, spatial_vectorh* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec2f_vec2f_float32(vec2f& a, vec2f& b, float32 t, vec2f* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec3f_vec3f_float32(vec3f& a, vec3f& b, float32 t, vec3f* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec4f_vec4f_float32(vec4f& a, vec4f& b, float32 t, vec4f* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_spatial_vectorf_spatial_vectorf_float32(spatial_vectorf& a, spatial_vectorf& b, float32 t, spatial_vectorf* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec2d_vec2d_float64(vec2d& a, vec2d& b, float64 t, vec2d* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec3d_vec3d_float64(vec3d& a, vec3d& b, float64 t, vec3d* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_vec4d_vec4d_float64(vec4d& a, vec4d& b, float64 t, vec4d* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_spatial_vectord_spatial_vectord_float64(spatial_vectord& a, spatial_vectord& b, float64 t, spatial_vectord* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat22h_mat22h_float16(mat22h& a, mat22h& b, float16 t, mat22h* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat33h_mat33h_float16(mat33h& a, mat33h& b, float16 t, mat33h* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat44h_mat44h_float16(mat44h& a, mat44h& b, float16 t, mat44h* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_spatial_matrixh_spatial_matrixh_float16(spatial_matrixh& a, spatial_matrixh& b, float16 t, spatial_matrixh* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat22f_mat22f_float32(mat22f& a, mat22f& b, float32 t, mat22f* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat33f_mat33f_float32(mat33f& a, mat33f& b, float32 t, mat33f* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat44f_mat44f_float32(mat44f& a, mat44f& b, float32 t, mat44f* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_spatial_matrixf_spatial_matrixf_float32(spatial_matrixf& a, spatial_matrixf& b, float32 t, spatial_matrixf* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat22d_mat22d_float64(mat22d& a, mat22d& b, float64 t, mat22d* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat33d_mat33d_float64(mat33d& a, mat33d& b, float64 t, mat33d* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_mat44d_mat44d_float64(mat44d& a, mat44d& b, float64 t, mat44d* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_spatial_matrixd_spatial_matrixd_float64(spatial_matrixd& a, spatial_matrixd& b, float64 t, spatial_matrixd* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_quath_quath_float16(quath& a, quath& b, float16 t, quath* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_quatf_quatf_float32(quatf& a, quatf& b, float32 t, quatf* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_quatd_quatd_float64(quatd& a, quatd& b, float64 t, quatd* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_transformh_transformh_float16(transformh& a, transformh& b, float16 t, transformh* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_transformf_transformf_float32(transformf& a, transformf& b, float32 t, transformf* ret) { *ret = wp::lerp(a, b, t); }
+WP_API void builtin_lerp_transformd_transformd_float64(transformd& a, transformd& b, float64 t, transformd* ret) { *ret = wp::lerp(a, b, t); }
 WP_API void builtin_smoothstep_float16_float16_float16(float16 edge0, float16 edge1, float16 x, float16* ret) { *ret = wp::smoothstep(edge0, edge1, x); }
 WP_API void builtin_smoothstep_float32_float32_float32(float32 edge0, float32 edge1, float32 x, float32* ret) { *ret = wp::smoothstep(edge0, edge1, x); }
 WP_API void builtin_smoothstep_float64_float64_float64(float64 edge0, float64 edge1, float64 x, float64* ret) { *ret = wp::smoothstep(edge0, edge1, x); }
@@ -928,60 +928,60 @@ WP_API void builtin_add_uint16_uint16(uint16 x, uint16 y, uint16* ret) { *ret =
 WP_API void builtin_add_uint32_uint32(uint32 x, uint32 y, uint32* ret) { *ret = wp::add(x, y); }
 WP_API void builtin_add_uint64_uint64(uint64 x, uint64 y, uint64* ret) { *ret = wp::add(x, y); }
 WP_API void builtin_add_uint8_uint8(uint8 x, uint8 y, uint8* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2h_vec2h(vec2h x, vec2h y, vec2h* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4h_vec4h(vec4h x, vec4h y, vec4h* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2f_vec2f(vec2f x, vec2f y, vec2f* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4f_vec4f(vec4f x, vec4f y, vec4f* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2d_vec2d(vec2d x, vec2d y, vec2d* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4d_vec4d(vec4d x, vec4d y, vec4d* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2s_vec2s(vec2s x, vec2s y, vec2s* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4s_vec4s(vec4s x, vec4s y, vec4s* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2i_vec2i(vec2i x, vec2i y, vec2i* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4i_vec4i(vec4i x, vec4i y, vec4i* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2l_vec2l(vec2l x, vec2l y, vec2l* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4l_vec4l(vec4l x, vec4l y, vec4l* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2b_vec2b(vec2b x, vec2b y, vec2b* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4b_vec4b(vec4b x, vec4b y, vec4b* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2us_vec2us(vec2us x, vec2us y, vec2us* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4us_vec4us(vec4us x, vec4us y, vec4us* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2ui_vec2ui(vec2ui x, vec2ui y, vec2ui* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4ui_vec4ui(vec4ui x, vec4ui y, vec4ui* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2ul_vec2ul(vec2ul x, vec2ul y, vec2ul* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4ul_vec4ul(vec4ul x, vec4ul y, vec4ul* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec2ub_vec2ub(vec2ub x, vec2ub y, vec2ub* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_vec4ub_vec4ub(vec4ub x, vec4ub y, vec4ub* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_quath_quath(quath x, quath y, quath* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_quatf_quatf(quatf x, quatf y, quatf* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_quatd_quatd(quatd x, quatd y, quatd* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat22h_mat22h(mat22h x, mat22h y, mat22h* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat33h_mat33h(mat33h x, mat33h y, mat33h* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat44h_mat44h(mat44h x, mat44h y, mat44h* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_spatial_matrixh_spatial_matrixh(spatial_matrixh x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat22f_mat22f(mat22f x, mat22f y, mat22f* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat33f_mat33f(mat33f x, mat33f y, mat33f* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat44f_mat44f(mat44f x, mat44f y, mat44f* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_spatial_matrixf_spatial_matrixf(spatial_matrixf x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat22d_mat22d(mat22d x, mat22d y, mat22d* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat33d_mat33d(mat33d x, mat33d y, mat33d* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_mat44d_mat44d(mat44d x, mat44d y, mat44d* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_spatial_matrixd_spatial_matrixd(spatial_matrixd x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_transformh_transformh(transformh x, transformh y, transformh* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_transformf_transformf(transformf x, transformf y, transformf* ret) { *ret = wp::add(x, y); }
-WP_API void builtin_add_transformd_transformd(transformd x, transformd y, transformd* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2h_vec2h(vec2h& x, vec2h& y, vec2h* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4h_vec4h(vec4h& x, vec4h& y, vec4h* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2f_vec2f(vec2f& x, vec2f& y, vec2f* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4f_vec4f(vec4f& x, vec4f& y, vec4f* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2d_vec2d(vec2d& x, vec2d& y, vec2d* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4d_vec4d(vec4d& x, vec4d& y, vec4d* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2s_vec2s(vec2s& x, vec2s& y, vec2s* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4s_vec4s(vec4s& x, vec4s& y, vec4s* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2i_vec2i(vec2i& x, vec2i& y, vec2i* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4i_vec4i(vec4i& x, vec4i& y, vec4i* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2l_vec2l(vec2l& x, vec2l& y, vec2l* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4l_vec4l(vec4l& x, vec4l& y, vec4l* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2b_vec2b(vec2b& x, vec2b& y, vec2b* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4b_vec4b(vec4b& x, vec4b& y, vec4b* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2us_vec2us(vec2us& x, vec2us& y, vec2us* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4us_vec4us(vec4us& x, vec4us& y, vec4us* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2ui_vec2ui(vec2ui& x, vec2ui& y, vec2ui* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4ui_vec4ui(vec4ui& x, vec4ui& y, vec4ui* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2ul_vec2ul(vec2ul& x, vec2ul& y, vec2ul* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4ul_vec4ul(vec4ul& x, vec4ul& y, vec4ul* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec2ub_vec2ub(vec2ub& x, vec2ub& y, vec2ub* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_vec4ub_vec4ub(vec4ub& x, vec4ub& y, vec4ub* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_quath_quath(quath& x, quath& y, quath* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_quatf_quatf(quatf& x, quatf& y, quatf* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_quatd_quatd(quatd& x, quatd& y, quatd* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat22h_mat22h(mat22h& x, mat22h& y, mat22h* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat33h_mat33h(mat33h& x, mat33h& y, mat33h* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat44h_mat44h(mat44h& x, mat44h& y, mat44h* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_spatial_matrixh_spatial_matrixh(spatial_matrixh& x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat22f_mat22f(mat22f& x, mat22f& y, mat22f* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat33f_mat33f(mat33f& x, mat33f& y, mat33f* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat44f_mat44f(mat44f& x, mat44f& y, mat44f* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_spatial_matrixf_spatial_matrixf(spatial_matrixf& x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat22d_mat22d(mat22d& x, mat22d& y, mat22d* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat33d_mat33d(mat33d& x, mat33d& y, mat33d* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_mat44d_mat44d(mat44d& x, mat44d& y, mat44d* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_spatial_matrixd_spatial_matrixd(spatial_matrixd& x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_transformh_transformh(transformh& x, transformh& y, transformh* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_transformf_transformf(transformf& x, transformf& y, transformf* ret) { *ret = wp::add(x, y); }
+WP_API void builtin_add_transformd_transformd(transformd& x, transformd& y, transformd* ret) { *ret = wp::add(x, y); }
 WP_API void builtin_sub_float16_float16(float16 x, float16 y, float16* ret) { *ret = wp::sub(x, y); }
 WP_API void builtin_sub_float32_float32(float32 x, float32 y, float32* ret) { *ret = wp::sub(x, y); }
 WP_API void builtin_sub_float64_float64(float64 x, float64 y, float64* ret) { *ret = wp::sub(x, y); }
@@ -993,60 +993,60 @@ WP_API void builtin_sub_uint16_uint16(uint16 x, uint16 y, uint16* ret) { *ret =
 WP_API void builtin_sub_uint32_uint32(uint32 x, uint32 y, uint32* ret) { *ret = wp::sub(x, y); }
 WP_API void builtin_sub_uint64_uint64(uint64 x, uint64 y, uint64* ret) { *ret = wp::sub(x, y); }
 WP_API void builtin_sub_uint8_uint8(uint8 x, uint8 y, uint8* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2h_vec2h(vec2h x, vec2h y, vec2h* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3h_vec3h(vec3h x, vec3h y, vec3h* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4h_vec4h(vec4h x, vec4h y, vec4h* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_spatial_vectorh_spatial_vectorh(spatial_vectorh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2f_vec2f(vec2f x, vec2f y, vec2f* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3f_vec3f(vec3f x, vec3f y, vec3f* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4f_vec4f(vec4f x, vec4f y, vec4f* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_spatial_vectorf_spatial_vectorf(spatial_vectorf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2d_vec2d(vec2d x, vec2d y, vec2d* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3d_vec3d(vec3d x, vec3d y, vec3d* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4d_vec4d(vec4d x, vec4d y, vec4d* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_spatial_vectord_spatial_vectord(spatial_vectord x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2s_vec2s(vec2s x, vec2s y, vec2s* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3s_vec3s(vec3s x, vec3s y, vec3s* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4s_vec4s(vec4s x, vec4s y, vec4s* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2i_vec2i(vec2i x, vec2i y, vec2i* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3i_vec3i(vec3i x, vec3i y, vec3i* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4i_vec4i(vec4i x, vec4i y, vec4i* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2l_vec2l(vec2l x, vec2l y, vec2l* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3l_vec3l(vec3l x, vec3l y, vec3l* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4l_vec4l(vec4l x, vec4l y, vec4l* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2b_vec2b(vec2b x, vec2b y, vec2b* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3b_vec3b(vec3b x, vec3b y, vec3b* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4b_vec4b(vec4b x, vec4b y, vec4b* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2us_vec2us(vec2us x, vec2us y, vec2us* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3us_vec3us(vec3us x, vec3us y, vec3us* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4us_vec4us(vec4us x, vec4us y, vec4us* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2ui_vec2ui(vec2ui x, vec2ui y, vec2ui* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3ui_vec3ui(vec3ui x, vec3ui y, vec3ui* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4ui_vec4ui(vec4ui x, vec4ui y, vec4ui* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2ul_vec2ul(vec2ul x, vec2ul y, vec2ul* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3ul_vec3ul(vec3ul x, vec3ul y, vec3ul* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4ul_vec4ul(vec4ul x, vec4ul y, vec4ul* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec2ub_vec2ub(vec2ub x, vec2ub y, vec2ub* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec3ub_vec3ub(vec3ub x, vec3ub y, vec3ub* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_vec4ub_vec4ub(vec4ub x, vec4ub y, vec4ub* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat22h_mat22h(mat22h x, mat22h y, mat22h* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat33h_mat33h(mat33h x, mat33h y, mat33h* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat44h_mat44h(mat44h x, mat44h y, mat44h* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_spatial_matrixh_spatial_matrixh(spatial_matrixh x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat22f_mat22f(mat22f x, mat22f y, mat22f* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat33f_mat33f(mat33f x, mat33f y, mat33f* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat44f_mat44f(mat44f x, mat44f y, mat44f* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_spatial_matrixf_spatial_matrixf(spatial_matrixf x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat22d_mat22d(mat22d x, mat22d y, mat22d* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat33d_mat33d(mat33d x, mat33d y, mat33d* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_mat44d_mat44d(mat44d x, mat44d y, mat44d* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_spatial_matrixd_spatial_matrixd(spatial_matrixd x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_quath_quath(quath x, quath y, quath* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_quatf_quatf(quatf x, quatf y, quatf* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_quatd_quatd(quatd x, quatd y, quatd* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_transformh_transformh(transformh x, transformh y, transformh* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_transformf_transformf(transformf x, transformf y, transformf* ret) { *ret = wp::sub(x, y); }
-WP_API void builtin_sub_transformd_transformd(transformd x, transformd y, transformd* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2h_vec2h(vec2h& x, vec2h& y, vec2h* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3h_vec3h(vec3h& x, vec3h& y, vec3h* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4h_vec4h(vec4h& x, vec4h& y, vec4h* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_spatial_vectorh_spatial_vectorh(spatial_vectorh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2f_vec2f(vec2f& x, vec2f& y, vec2f* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3f_vec3f(vec3f& x, vec3f& y, vec3f* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4f_vec4f(vec4f& x, vec4f& y, vec4f* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_spatial_vectorf_spatial_vectorf(spatial_vectorf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2d_vec2d(vec2d& x, vec2d& y, vec2d* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3d_vec3d(vec3d& x, vec3d& y, vec3d* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4d_vec4d(vec4d& x, vec4d& y, vec4d* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_spatial_vectord_spatial_vectord(spatial_vectord& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2s_vec2s(vec2s& x, vec2s& y, vec2s* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3s_vec3s(vec3s& x, vec3s& y, vec3s* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4s_vec4s(vec4s& x, vec4s& y, vec4s* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2i_vec2i(vec2i& x, vec2i& y, vec2i* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3i_vec3i(vec3i& x, vec3i& y, vec3i* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4i_vec4i(vec4i& x, vec4i& y, vec4i* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2l_vec2l(vec2l& x, vec2l& y, vec2l* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3l_vec3l(vec3l& x, vec3l& y, vec3l* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4l_vec4l(vec4l& x, vec4l& y, vec4l* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2b_vec2b(vec2b& x, vec2b& y, vec2b* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3b_vec3b(vec3b& x, vec3b& y, vec3b* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4b_vec4b(vec4b& x, vec4b& y, vec4b* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2us_vec2us(vec2us& x, vec2us& y, vec2us* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3us_vec3us(vec3us& x, vec3us& y, vec3us* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4us_vec4us(vec4us& x, vec4us& y, vec4us* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2ui_vec2ui(vec2ui& x, vec2ui& y, vec2ui* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3ui_vec3ui(vec3ui& x, vec3ui& y, vec3ui* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4ui_vec4ui(vec4ui& x, vec4ui& y, vec4ui* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2ul_vec2ul(vec2ul& x, vec2ul& y, vec2ul* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3ul_vec3ul(vec3ul& x, vec3ul& y, vec3ul* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4ul_vec4ul(vec4ul& x, vec4ul& y, vec4ul* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec2ub_vec2ub(vec2ub& x, vec2ub& y, vec2ub* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec3ub_vec3ub(vec3ub& x, vec3ub& y, vec3ub* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_vec4ub_vec4ub(vec4ub& x, vec4ub& y, vec4ub* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat22h_mat22h(mat22h& x, mat22h& y, mat22h* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat33h_mat33h(mat33h& x, mat33h& y, mat33h* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat44h_mat44h(mat44h& x, mat44h& y, mat44h* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_spatial_matrixh_spatial_matrixh(spatial_matrixh& x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat22f_mat22f(mat22f& x, mat22f& y, mat22f* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat33f_mat33f(mat33f& x, mat33f& y, mat33f* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat44f_mat44f(mat44f& x, mat44f& y, mat44f* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_spatial_matrixf_spatial_matrixf(spatial_matrixf& x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat22d_mat22d(mat22d& x, mat22d& y, mat22d* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat33d_mat33d(mat33d& x, mat33d& y, mat33d* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_mat44d_mat44d(mat44d& x, mat44d& y, mat44d* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_spatial_matrixd_spatial_matrixd(spatial_matrixd& x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_quath_quath(quath& x, quath& y, quath* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_quatf_quatf(quatf& x, quatf& y, quatf* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_quatd_quatd(quatd& x, quatd& y, quatd* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_transformh_transformh(transformh& x, transformh& y, transformh* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_transformf_transformf(transformf& x, transformf& y, transformf* ret) { *ret = wp::sub(x, y); }
+WP_API void builtin_sub_transformd_transformd(transformd& x, transformd& y, transformd* ret) { *ret = wp::sub(x, y); }
 WP_API void builtin_bit_and_int16_int16(int16 x, int16 y, int16* ret) { *ret = wp::bit_and(x, y); }
 WP_API void builtin_bit_and_int32_int32(int32 x, int32 y, int32* ret) { *ret = wp::bit_and(x, y); }
 WP_API void builtin_bit_and_int64_int64(int64 x, int64 y, int64* ret) { *ret = wp::bit_and(x, y); }
@@ -1106,156 +1106,156 @@ WP_API void builtin_mul_uint16_uint16(uint16 x, uint16 y, uint16* ret) { *ret =
 WP_API void builtin_mul_uint32_uint32(uint32 x, uint32 y, uint32* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_uint64_uint64(uint64 x, uint64 y, uint64* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mul_uint8_uint8(uint8 x, uint8 y, uint8* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2h_float16(vec2h x, float16 y, vec2h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3h_float16(vec3h x, float16 y, vec3h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4h_float16(vec4h x, float16 y, vec4h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_vectorh_float16(spatial_vectorh x, float16 y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2f_float32(vec2f x, float32 y, vec2f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3f_float32(vec3f x, float32 y, vec3f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4f_float32(vec4f x, float32 y, vec4f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_vectorf_float32(spatial_vectorf x, float32 y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2d_float64(vec2d x, float64 y, vec2d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3d_float64(vec3d x, float64 y, vec3d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4d_float64(vec4d x, float64 y, vec4d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_vectord_float64(spatial_vectord x, float64 y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2s_int16(vec2s x, int16 y, vec2s* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3s_int16(vec3s x, int16 y, vec3s* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4s_int16(vec4s x, int16 y, vec4s* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2i_int32(vec2i x, int32 y, vec2i* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3i_int32(vec3i x, int32 y, vec3i* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4i_int32(vec4i x, int32 y, vec4i* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2l_int64(vec2l x, int64 y, vec2l* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3l_int64(vec3l x, int64 y, vec3l* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4l_int64(vec4l x, int64 y, vec4l* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2b_int8(vec2b x, int8 y, vec2b* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3b_int8(vec3b x, int8 y, vec3b* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4b_int8(vec4b x, int8 y, vec4b* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2us_uint16(vec2us x, uint16 y, vec2us* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3us_uint16(vec3us x, uint16 y, vec3us* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4us_uint16(vec4us x, uint16 y, vec4us* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2ui_uint32(vec2ui x, uint32 y, vec2ui* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3ui_uint32(vec3ui x, uint32 y, vec3ui* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4ui_uint32(vec4ui x, uint32 y, vec4ui* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2ul_uint64(vec2ul x, uint64 y, vec2ul* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3ul_uint64(vec3ul x, uint64 y, vec3ul* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4ul_uint64(vec4ul x, uint64 y, vec4ul* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2ub_uint8(vec2ub x, uint8 y, vec2ub* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3ub_uint8(vec3ub x, uint8 y, vec3ub* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4ub_uint8(vec4ub x, uint8 y, vec4ub* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_vec2h(float16 x, vec2h y, vec2h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_vec3h(float16 x, vec3h y, vec3h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_vec4h(float16 x, vec4h y, vec4h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_spatial_vectorh(float16 x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_vec2f(float32 x, vec2f y, vec2f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_vec3f(float32 x, vec3f y, vec3f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_vec4f(float32 x, vec4f y, vec4f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_spatial_vectorf(float32 x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_vec2d(float64 x, vec2d y, vec2d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_vec3d(float64 x, vec3d y, vec3d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_vec4d(float64 x, vec4d y, vec4d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_spatial_vectord(float64 x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int16_vec2s(int16 x, vec2s y, vec2s* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int16_vec3s(int16 x, vec3s y, vec3s* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int16_vec4s(int16 x, vec4s y, vec4s* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int32_vec2i(int32 x, vec2i y, vec2i* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int32_vec3i(int32 x, vec3i y, vec3i* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int32_vec4i(int32 x, vec4i y, vec4i* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int64_vec2l(int64 x, vec2l y, vec2l* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int64_vec3l(int64 x, vec3l y, vec3l* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int64_vec4l(int64 x, vec4l y, vec4l* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int8_vec2b(int8 x, vec2b y, vec2b* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int8_vec3b(int8 x, vec3b y, vec3b* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_int8_vec4b(int8 x, vec4b y, vec4b* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint16_vec2us(uint16 x, vec2us y, vec2us* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint16_vec3us(uint16 x, vec3us y, vec3us* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint16_vec4us(uint16 x, vec4us y, vec4us* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint32_vec2ui(uint32 x, vec2ui y, vec2ui* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint32_vec3ui(uint32 x, vec3ui y, vec3ui* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint32_vec4ui(uint32 x, vec4ui y, vec4ui* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint64_vec2ul(uint64 x, vec2ul y, vec2ul* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint64_vec3ul(uint64 x, vec3ul y, vec3ul* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint64_vec4ul(uint64 x, vec4ul y, vec4ul* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint8_vec2ub(uint8 x, vec2ub y, vec2ub* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint8_vec3ub(uint8 x, vec3ub y, vec3ub* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_uint8_vec4ub(uint8 x, vec4ub y, vec4ub* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_quath_float16(quath x, float16 y, quath* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_quatf_float32(quatf x, float32 y, quatf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_quatd_float64(quatd x, float64 y, quatd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_quath(float16 x, quath y, quath* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_quatf(float32 x, quatf y, quatf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_quatd(float64 x, quatd y, quatd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_quath_quath(quath x, quath y, quath* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_quatf_quatf(quatf x, quatf y, quatf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_quatd_quatd(quatd x, quatd y, quatd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_mat22h(float16 x, mat22h y, mat22h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_mat33h(float16 x, mat33h y, mat33h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_mat44h(float16 x, mat44h y, mat44h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_spatial_matrixh(float16 x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_mat22f(float32 x, mat22f y, mat22f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_mat33f(float32 x, mat33f y, mat33f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_mat44f(float32 x, mat44f y, mat44f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_spatial_matrixf(float32 x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_mat22d(float64 x, mat22d y, mat22d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_mat33d(float64 x, mat33d y, mat33d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_mat44d(float64 x, mat44d y, mat44d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_spatial_matrixd(float64 x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22h_float16(mat22h x, float16 y, mat22h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33h_float16(mat33h x, float16 y, mat33h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44h_float16(mat44h x, float16 y, mat44h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixh_float16(spatial_matrixh x, float16 y, spatial_matrixh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22f_float32(mat22f x, float32 y, mat22f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33f_float32(mat33f x, float32 y, mat33f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44f_float32(mat44f x, float32 y, mat44f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixf_float32(spatial_matrixf x, float32 y, spatial_matrixf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22d_float64(mat22d x, float64 y, mat22d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33d_float64(mat33d x, float64 y, mat33d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44d_float64(mat44d x, float64 y, mat44d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixd_float64(spatial_matrixd x, float64 y, spatial_matrixd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22h_vec2h(mat22h x, vec2h y, vec2h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33h_vec3h(mat33h x, vec3h y, vec3h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44h_vec4h(mat44h x, vec4h y, vec4h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixh_spatial_vectorh(spatial_matrixh x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22f_vec2f(mat22f x, vec2f y, vec2f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33f_vec3f(mat33f x, vec3f y, vec3f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44f_vec4f(mat44f x, vec4f y, vec4f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixf_spatial_vectorf(spatial_matrixf x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22d_vec2d(mat22d x, vec2d y, vec2d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33d_vec3d(mat33d x, vec3d y, vec3d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44d_vec4d(mat44d x, vec4d y, vec4d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixd_spatial_vectord(spatial_matrixd x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2h_mat22h(vec2h x, mat22h y, vec2h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3h_mat33h(vec3h x, mat33h y, vec3h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4h_mat44h(vec4h x, mat44h y, vec4h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_vectorh_spatial_matrixh(spatial_vectorh x, spatial_matrixh y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2f_mat22f(vec2f x, mat22f y, vec2f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3f_mat33f(vec3f x, mat33f y, vec3f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4f_mat44f(vec4f x, mat44f y, vec4f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_vectorf_spatial_matrixf(spatial_vectorf x, spatial_matrixf y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec2d_mat22d(vec2d x, mat22d y, vec2d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec3d_mat33d(vec3d x, mat33d y, vec3d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_vec4d_mat44d(vec4d x, mat44d y, vec4d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_vectord_spatial_matrixd(spatial_vectord x, spatial_matrixd y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22h_mat22h(mat22h x, mat22h y, mat22h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33h_mat33h(mat33h x, mat33h y, mat33h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44h_mat44h(mat44h x, mat44h y, mat44h* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixh_spatial_matrixh(spatial_matrixh x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22f_mat22f(mat22f x, mat22f y, mat22f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33f_mat33f(mat33f x, mat33f y, mat33f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44f_mat44f(mat44f x, mat44f y, mat44f* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixf_spatial_matrixf(spatial_matrixf x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat22d_mat22d(mat22d x, mat22d y, mat22d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat33d_mat33d(mat33d x, mat33d y, mat33d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_mat44d_mat44d(mat44d x, mat44d y, mat44d* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_spatial_matrixd_spatial_matrixd(spatial_matrixd x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_transformh_transformh(transformh x, transformh y, transformh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_transformf_transformf(transformf x, transformf y, transformf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_transformd_transformd(transformd x, transformd y, transformd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float16_transformh(float16 x, transformh y, transformh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float32_transformf(float32 x, transformf y, transformf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_float64_transformd(float64 x, transformd y, transformd* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_transformh_float16(transformh x, float16 y, transformh* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_transformf_float32(transformf x, float32 y, transformf* ret) { *ret = wp::mul(x, y); }
-WP_API void builtin_mul_transformd_float64(transformd x, float64 y, transformd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2h_float16(vec2h& x, float16 y, vec2h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3h_float16(vec3h& x, float16 y, vec3h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4h_float16(vec4h& x, float16 y, vec4h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectorh_float16(spatial_vectorh& x, float16 y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2f_float32(vec2f& x, float32 y, vec2f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3f_float32(vec3f& x, float32 y, vec3f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4f_float32(vec4f& x, float32 y, vec4f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectorf_float32(spatial_vectorf& x, float32 y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2d_float64(vec2d& x, float64 y, vec2d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3d_float64(vec3d& x, float64 y, vec3d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4d_float64(vec4d& x, float64 y, vec4d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectord_float64(spatial_vectord& x, float64 y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2s_int16(vec2s& x, int16 y, vec2s* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3s_int16(vec3s& x, int16 y, vec3s* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4s_int16(vec4s& x, int16 y, vec4s* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2i_int32(vec2i& x, int32 y, vec2i* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3i_int32(vec3i& x, int32 y, vec3i* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4i_int32(vec4i& x, int32 y, vec4i* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2l_int64(vec2l& x, int64 y, vec2l* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3l_int64(vec3l& x, int64 y, vec3l* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4l_int64(vec4l& x, int64 y, vec4l* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2b_int8(vec2b& x, int8 y, vec2b* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3b_int8(vec3b& x, int8 y, vec3b* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4b_int8(vec4b& x, int8 y, vec4b* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2us_uint16(vec2us& x, uint16 y, vec2us* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3us_uint16(vec3us& x, uint16 y, vec3us* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4us_uint16(vec4us& x, uint16 y, vec4us* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2ui_uint32(vec2ui& x, uint32 y, vec2ui* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3ui_uint32(vec3ui& x, uint32 y, vec3ui* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4ui_uint32(vec4ui& x, uint32 y, vec4ui* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2ul_uint64(vec2ul& x, uint64 y, vec2ul* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3ul_uint64(vec3ul& x, uint64 y, vec3ul* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4ul_uint64(vec4ul& x, uint64 y, vec4ul* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2ub_uint8(vec2ub& x, uint8 y, vec2ub* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3ub_uint8(vec3ub& x, uint8 y, vec3ub* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4ub_uint8(vec4ub& x, uint8 y, vec4ub* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_vec2h(float16 x, vec2h& y, vec2h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_vec3h(float16 x, vec3h& y, vec3h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_vec4h(float16 x, vec4h& y, vec4h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_spatial_vectorh(float16 x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_vec2f(float32 x, vec2f& y, vec2f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_vec3f(float32 x, vec3f& y, vec3f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_vec4f(float32 x, vec4f& y, vec4f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_spatial_vectorf(float32 x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_vec2d(float64 x, vec2d& y, vec2d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_vec3d(float64 x, vec3d& y, vec3d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_vec4d(float64 x, vec4d& y, vec4d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_spatial_vectord(float64 x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int16_vec2s(int16 x, vec2s& y, vec2s* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int16_vec3s(int16 x, vec3s& y, vec3s* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int16_vec4s(int16 x, vec4s& y, vec4s* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int32_vec2i(int32 x, vec2i& y, vec2i* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int32_vec3i(int32 x, vec3i& y, vec3i* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int32_vec4i(int32 x, vec4i& y, vec4i* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int64_vec2l(int64 x, vec2l& y, vec2l* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int64_vec3l(int64 x, vec3l& y, vec3l* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int64_vec4l(int64 x, vec4l& y, vec4l* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int8_vec2b(int8 x, vec2b& y, vec2b* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int8_vec3b(int8 x, vec3b& y, vec3b* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_int8_vec4b(int8 x, vec4b& y, vec4b* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint16_vec2us(uint16 x, vec2us& y, vec2us* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint16_vec3us(uint16 x, vec3us& y, vec3us* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint16_vec4us(uint16 x, vec4us& y, vec4us* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint32_vec2ui(uint32 x, vec2ui& y, vec2ui* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint32_vec3ui(uint32 x, vec3ui& y, vec3ui* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint32_vec4ui(uint32 x, vec4ui& y, vec4ui* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint64_vec2ul(uint64 x, vec2ul& y, vec2ul* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint64_vec3ul(uint64 x, vec3ul& y, vec3ul* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint64_vec4ul(uint64 x, vec4ul& y, vec4ul* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint8_vec2ub(uint8 x, vec2ub& y, vec2ub* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint8_vec3ub(uint8 x, vec3ub& y, vec3ub* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_uint8_vec4ub(uint8 x, vec4ub& y, vec4ub* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_quath_float16(quath& x, float16 y, quath* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_quatf_float32(quatf& x, float32 y, quatf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_quatd_float64(quatd& x, float64 y, quatd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_quath(float16 x, quath& y, quath* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_quatf(float32 x, quatf& y, quatf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_quatd(float64 x, quatd& y, quatd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_quath_quath(quath& x, quath& y, quath* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_quatf_quatf(quatf& x, quatf& y, quatf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_quatd_quatd(quatd& x, quatd& y, quatd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_mat22h(float16 x, mat22h& y, mat22h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_mat33h(float16 x, mat33h& y, mat33h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_mat44h(float16 x, mat44h& y, mat44h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_spatial_matrixh(float16 x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_mat22f(float32 x, mat22f& y, mat22f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_mat33f(float32 x, mat33f& y, mat33f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_mat44f(float32 x, mat44f& y, mat44f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_spatial_matrixf(float32 x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_mat22d(float64 x, mat22d& y, mat22d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_mat33d(float64 x, mat33d& y, mat33d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_mat44d(float64 x, mat44d& y, mat44d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_spatial_matrixd(float64 x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22h_float16(mat22h& x, float16 y, mat22h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33h_float16(mat33h& x, float16 y, mat33h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44h_float16(mat44h& x, float16 y, mat44h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixh_float16(spatial_matrixh& x, float16 y, spatial_matrixh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22f_float32(mat22f& x, float32 y, mat22f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33f_float32(mat33f& x, float32 y, mat33f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44f_float32(mat44f& x, float32 y, mat44f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixf_float32(spatial_matrixf& x, float32 y, spatial_matrixf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22d_float64(mat22d& x, float64 y, mat22d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33d_float64(mat33d& x, float64 y, mat33d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44d_float64(mat44d& x, float64 y, mat44d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixd_float64(spatial_matrixd& x, float64 y, spatial_matrixd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22h_vec2h(mat22h& x, vec2h& y, vec2h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33h_vec3h(mat33h& x, vec3h& y, vec3h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44h_vec4h(mat44h& x, vec4h& y, vec4h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixh_spatial_vectorh(spatial_matrixh& x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22f_vec2f(mat22f& x, vec2f& y, vec2f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33f_vec3f(mat33f& x, vec3f& y, vec3f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44f_vec4f(mat44f& x, vec4f& y, vec4f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixf_spatial_vectorf(spatial_matrixf& x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22d_vec2d(mat22d& x, vec2d& y, vec2d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33d_vec3d(mat33d& x, vec3d& y, vec3d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44d_vec4d(mat44d& x, vec4d& y, vec4d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixd_spatial_vectord(spatial_matrixd& x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2h_mat22h(vec2h& x, mat22h& y, vec2h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3h_mat33h(vec3h& x, mat33h& y, vec3h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4h_mat44h(vec4h& x, mat44h& y, vec4h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectorh_spatial_matrixh(spatial_vectorh& x, spatial_matrixh& y, spatial_vectorh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2f_mat22f(vec2f& x, mat22f& y, vec2f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3f_mat33f(vec3f& x, mat33f& y, vec3f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4f_mat44f(vec4f& x, mat44f& y, vec4f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectorf_spatial_matrixf(spatial_vectorf& x, spatial_matrixf& y, spatial_vectorf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec2d_mat22d(vec2d& x, mat22d& y, vec2d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec3d_mat33d(vec3d& x, mat33d& y, vec3d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_vec4d_mat44d(vec4d& x, mat44d& y, vec4d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_vectord_spatial_matrixd(spatial_vectord& x, spatial_matrixd& y, spatial_vectord* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22h_mat22h(mat22h& x, mat22h& y, mat22h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33h_mat33h(mat33h& x, mat33h& y, mat33h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44h_mat44h(mat44h& x, mat44h& y, mat44h* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixh_spatial_matrixh(spatial_matrixh& x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22f_mat22f(mat22f& x, mat22f& y, mat22f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33f_mat33f(mat33f& x, mat33f& y, mat33f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44f_mat44f(mat44f& x, mat44f& y, mat44f* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixf_spatial_matrixf(spatial_matrixf& x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat22d_mat22d(mat22d& x, mat22d& y, mat22d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat33d_mat33d(mat33d& x, mat33d& y, mat33d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_mat44d_mat44d(mat44d& x, mat44d& y, mat44d* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_spatial_matrixd_spatial_matrixd(spatial_matrixd& x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_transformh_transformh(transformh& x, transformh& y, transformh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_transformf_transformf(transformf& x, transformf& y, transformf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_transformd_transformd(transformd& x, transformd& y, transformd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float16_transformh(float16 x, transformh& y, transformh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float32_transformf(float32 x, transformf& y, transformf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_float64_transformd(float64 x, transformd& y, transformd* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_transformh_float16(transformh& x, float16 y, transformh* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_transformf_float32(transformf& x, float32 y, transformf* ret) { *ret = wp::mul(x, y); }
+WP_API void builtin_mul_transformd_float64(transformd& x, float64 y, transformd* ret) { *ret = wp::mul(x, y); }
 WP_API void builtin_mod_float16_float16(float16 x, float16 y, float16* ret) { *ret = wp::mod(x, y); }
 WP_API void builtin_mod_float32_float32(float32 x, float32 y, float32* ret) { *ret = wp::mod(x, y); }
 WP_API void builtin_mod_float64_float64(float64 x, float64 y, float64* ret) { *ret = wp::mod(x, y); }
@@ -1278,108 +1278,108 @@ WP_API void builtin_div_uint16_uint16(uint16 x, uint16 y, uint16* ret) { *ret =
 WP_API void builtin_div_uint32_uint32(uint32 x, uint32 y, uint32* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_uint64_uint64(uint64 x, uint64 y, uint64* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_div_uint8_uint8(uint8 x, uint8 y, uint8* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2h_float16(vec2h x, float16 y, vec2h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3h_float16(vec3h x, float16 y, vec3h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4h_float16(vec4h x, float16 y, vec4h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_spatial_vectorh_float16(spatial_vectorh x, float16 y, spatial_vectorh* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2f_float32(vec2f x, float32 y, vec2f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3f_float32(vec3f x, float32 y, vec3f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4f_float32(vec4f x, float32 y, vec4f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_spatial_vectorf_float32(spatial_vectorf x, float32 y, spatial_vectorf* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2d_float64(vec2d x, float64 y, vec2d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3d_float64(vec3d x, float64 y, vec3d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4d_float64(vec4d x, float64 y, vec4d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_spatial_vectord_float64(spatial_vectord x, float64 y, spatial_vectord* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2s_int16(vec2s x, int16 y, vec2s* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3s_int16(vec3s x, int16 y, vec3s* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4s_int16(vec4s x, int16 y, vec4s* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2i_int32(vec2i x, int32 y, vec2i* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3i_int32(vec3i x, int32 y, vec3i* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4i_int32(vec4i x, int32 y, vec4i* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2l_int64(vec2l x, int64 y, vec2l* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3l_int64(vec3l x, int64 y, vec3l* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4l_int64(vec4l x, int64 y, vec4l* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2b_int8(vec2b x, int8 y, vec2b* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3b_int8(vec3b x, int8 y, vec3b* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4b_int8(vec4b x, int8 y, vec4b* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2us_uint16(vec2us x, uint16 y, vec2us* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3us_uint16(vec3us x, uint16 y, vec3us* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4us_uint16(vec4us x, uint16 y, vec4us* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2ui_uint32(vec2ui x, uint32 y, vec2ui* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3ui_uint32(vec3ui x, uint32 y, vec3ui* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4ui_uint32(vec4ui x, uint32 y, vec4ui* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2ul_uint64(vec2ul x, uint64 y, vec2ul* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3ul_uint64(vec3ul x, uint64 y, vec3ul* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4ul_uint64(vec4ul x, uint64 y, vec4ul* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec2ub_uint8(vec2ub x, uint8 y, vec2ub* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec3ub_uint8(vec3ub x, uint8 y, vec3ub* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_vec4ub_uint8(vec4ub x, uint8 y, vec4ub* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_vec2h(float16 x, vec2h y, vec2h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_vec3h(float16 x, vec3h y, vec3h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_vec4h(float16 x, vec4h y, vec4h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_spatial_vectorh(float16 x, spatial_vectorh y, spatial_vectorh* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_vec2f(float32 x, vec2f y, vec2f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_vec3f(float32 x, vec3f y, vec3f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_vec4f(float32 x, vec4f y, vec4f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_spatial_vectorf(float32 x, spatial_vectorf y, spatial_vectorf* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_vec2d(float64 x, vec2d y, vec2d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_vec3d(float64 x, vec3d y, vec3d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_vec4d(float64 x, vec4d y, vec4d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_spatial_vectord(float64 x, spatial_vectord y, spatial_vectord* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int16_vec2s(int16 x, vec2s y, vec2s* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int16_vec3s(int16 x, vec3s y, vec3s* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int16_vec4s(int16 x, vec4s y, vec4s* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int32_vec2i(int32 x, vec2i y, vec2i* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int32_vec3i(int32 x, vec3i y, vec3i* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int32_vec4i(int32 x, vec4i y, vec4i* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int64_vec2l(int64 x, vec2l y, vec2l* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int64_vec3l(int64 x, vec3l y, vec3l* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int64_vec4l(int64 x, vec4l y, vec4l* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int8_vec2b(int8 x, vec2b y, vec2b* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int8_vec3b(int8 x, vec3b y, vec3b* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_int8_vec4b(int8 x, vec4b y, vec4b* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint16_vec2us(uint16 x, vec2us y, vec2us* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint16_vec3us(uint16 x, vec3us y, vec3us* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint16_vec4us(uint16 x, vec4us y, vec4us* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint32_vec2ui(uint32 x, vec2ui y, vec2ui* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint32_vec3ui(uint32 x, vec3ui y, vec3ui* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint32_vec4ui(uint32 x, vec4ui y, vec4ui* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint64_vec2ul(uint64 x, vec2ul y, vec2ul* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint64_vec3ul(uint64 x, vec3ul y, vec3ul* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint64_vec4ul(uint64 x, vec4ul y, vec4ul* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint8_vec2ub(uint8 x, vec2ub y, vec2ub* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint8_vec3ub(uint8 x, vec3ub y, vec3ub* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_uint8_vec4ub(uint8 x, vec4ub y, vec4ub* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat22h_float16(mat22h x, float16 y, mat22h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat33h_float16(mat33h x, float16 y, mat33h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat44h_float16(mat44h x, float16 y, mat44h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_spatial_matrixh_float16(spatial_matrixh x, float16 y, spatial_matrixh* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat22f_float32(mat22f x, float32 y, mat22f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat33f_float32(mat33f x, float32 y, mat33f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat44f_float32(mat44f x, float32 y, mat44f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_spatial_matrixf_float32(spatial_matrixf x, float32 y, spatial_matrixf* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat22d_float64(mat22d x, float64 y, mat22d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat33d_float64(mat33d x, float64 y, mat33d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_mat44d_float64(mat44d x, float64 y, mat44d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_spatial_matrixd_float64(spatial_matrixd x, float64 y, spatial_matrixd* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_mat22h(float16 x, mat22h y, mat22h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_mat33h(float16 x, mat33h y, mat33h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_mat44h(float16 x, mat44h y, mat44h* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_spatial_matrixh(float16 x, spatial_matrixh y, spatial_matrixh* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_mat22f(float32 x, mat22f y, mat22f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_mat33f(float32 x, mat33f y, mat33f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_mat44f(float32 x, mat44f y, mat44f* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_spatial_matrixf(float32 x, spatial_matrixf y, spatial_matrixf* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_mat22d(float64 x, mat22d y, mat22d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_mat33d(float64 x, mat33d y, mat33d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_mat44d(float64 x, mat44d y, mat44d* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_spatial_matrixd(float64 x, spatial_matrixd y, spatial_matrixd* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_quath_float16(quath x, float16 y, quath* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_quatf_float32(quatf x, float32 y, quatf* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_quatd_float64(quatd x, float64 y, quatd* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float16_quath(float16 x, quath y, quath* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float32_quatf(float32 x, quatf y, quatf* ret) { *ret = wp::div(x, y); }
-WP_API void builtin_div_float64_quatd(float64 x, quatd y, quatd* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2h_float16(vec2h& x, float16 y, vec2h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3h_float16(vec3h& x, float16 y, vec3h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4h_float16(vec4h& x, float16 y, vec4h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_spatial_vectorh_float16(spatial_vectorh& x, float16 y, spatial_vectorh* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2f_float32(vec2f& x, float32 y, vec2f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3f_float32(vec3f& x, float32 y, vec3f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4f_float32(vec4f& x, float32 y, vec4f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_spatial_vectorf_float32(spatial_vectorf& x, float32 y, spatial_vectorf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2d_float64(vec2d& x, float64 y, vec2d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3d_float64(vec3d& x, float64 y, vec3d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4d_float64(vec4d& x, float64 y, vec4d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_spatial_vectord_float64(spatial_vectord& x, float64 y, spatial_vectord* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2s_int16(vec2s& x, int16 y, vec2s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3s_int16(vec3s& x, int16 y, vec3s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4s_int16(vec4s& x, int16 y, vec4s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2i_int32(vec2i& x, int32 y, vec2i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3i_int32(vec3i& x, int32 y, vec3i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4i_int32(vec4i& x, int32 y, vec4i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2l_int64(vec2l& x, int64 y, vec2l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3l_int64(vec3l& x, int64 y, vec3l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4l_int64(vec4l& x, int64 y, vec4l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2b_int8(vec2b& x, int8 y, vec2b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3b_int8(vec3b& x, int8 y, vec3b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4b_int8(vec4b& x, int8 y, vec4b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2us_uint16(vec2us& x, uint16 y, vec2us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3us_uint16(vec3us& x, uint16 y, vec3us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4us_uint16(vec4us& x, uint16 y, vec4us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2ui_uint32(vec2ui& x, uint32 y, vec2ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3ui_uint32(vec3ui& x, uint32 y, vec3ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4ui_uint32(vec4ui& x, uint32 y, vec4ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2ul_uint64(vec2ul& x, uint64 y, vec2ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3ul_uint64(vec3ul& x, uint64 y, vec3ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4ul_uint64(vec4ul& x, uint64 y, vec4ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec2ub_uint8(vec2ub& x, uint8 y, vec2ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec3ub_uint8(vec3ub& x, uint8 y, vec3ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_vec4ub_uint8(vec4ub& x, uint8 y, vec4ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_vec2h(float16 x, vec2h& y, vec2h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_vec3h(float16 x, vec3h& y, vec3h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_vec4h(float16 x, vec4h& y, vec4h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_spatial_vectorh(float16 x, spatial_vectorh& y, spatial_vectorh* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_vec2f(float32 x, vec2f& y, vec2f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_vec3f(float32 x, vec3f& y, vec3f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_vec4f(float32 x, vec4f& y, vec4f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_spatial_vectorf(float32 x, spatial_vectorf& y, spatial_vectorf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_vec2d(float64 x, vec2d& y, vec2d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_vec3d(float64 x, vec3d& y, vec3d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_vec4d(float64 x, vec4d& y, vec4d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_spatial_vectord(float64 x, spatial_vectord& y, spatial_vectord* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int16_vec2s(int16 x, vec2s& y, vec2s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int16_vec3s(int16 x, vec3s& y, vec3s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int16_vec4s(int16 x, vec4s& y, vec4s* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int32_vec2i(int32 x, vec2i& y, vec2i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int32_vec3i(int32 x, vec3i& y, vec3i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int32_vec4i(int32 x, vec4i& y, vec4i* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int64_vec2l(int64 x, vec2l& y, vec2l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int64_vec3l(int64 x, vec3l& y, vec3l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int64_vec4l(int64 x, vec4l& y, vec4l* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int8_vec2b(int8 x, vec2b& y, vec2b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int8_vec3b(int8 x, vec3b& y, vec3b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_int8_vec4b(int8 x, vec4b& y, vec4b* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint16_vec2us(uint16 x, vec2us& y, vec2us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint16_vec3us(uint16 x, vec3us& y, vec3us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint16_vec4us(uint16 x, vec4us& y, vec4us* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint32_vec2ui(uint32 x, vec2ui& y, vec2ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint32_vec3ui(uint32 x, vec3ui& y, vec3ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint32_vec4ui(uint32 x, vec4ui& y, vec4ui* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint64_vec2ul(uint64 x, vec2ul& y, vec2ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint64_vec3ul(uint64 x, vec3ul& y, vec3ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint64_vec4ul(uint64 x, vec4ul& y, vec4ul* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint8_vec2ub(uint8 x, vec2ub& y, vec2ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint8_vec3ub(uint8 x, vec3ub& y, vec3ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_uint8_vec4ub(uint8 x, vec4ub& y, vec4ub* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat22h_float16(mat22h& x, float16 y, mat22h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat33h_float16(mat33h& x, float16 y, mat33h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat44h_float16(mat44h& x, float16 y, mat44h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_spatial_matrixh_float16(spatial_matrixh& x, float16 y, spatial_matrixh* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat22f_float32(mat22f& x, float32 y, mat22f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat33f_float32(mat33f& x, float32 y, mat33f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat44f_float32(mat44f& x, float32 y, mat44f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_spatial_matrixf_float32(spatial_matrixf& x, float32 y, spatial_matrixf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat22d_float64(mat22d& x, float64 y, mat22d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat33d_float64(mat33d& x, float64 y, mat33d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_mat44d_float64(mat44d& x, float64 y, mat44d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_spatial_matrixd_float64(spatial_matrixd& x, float64 y, spatial_matrixd* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_mat22h(float16 x, mat22h& y, mat22h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_mat33h(float16 x, mat33h& y, mat33h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_mat44h(float16 x, mat44h& y, mat44h* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_spatial_matrixh(float16 x, spatial_matrixh& y, spatial_matrixh* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_mat22f(float32 x, mat22f& y, mat22f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_mat33f(float32 x, mat33f& y, mat33f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_mat44f(float32 x, mat44f& y, mat44f* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_spatial_matrixf(float32 x, spatial_matrixf& y, spatial_matrixf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_mat22d(float64 x, mat22d& y, mat22d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_mat33d(float64 x, mat33d& y, mat33d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_mat44d(float64 x, mat44d& y, mat44d* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_spatial_matrixd(float64 x, spatial_matrixd& y, spatial_matrixd* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_quath_float16(quath& x, float16 y, quath* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_quatf_float32(quatf& x, float32 y, quatf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_quatd_float64(quatd& x, float64 y, quatd* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float16_quath(float16 x, quath& y, quath* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float32_quatf(float32 x, quatf& y, quatf* ret) { *ret = wp::div(x, y); }
+WP_API void builtin_div_float64_quatd(float64 x, quatd& y, quatd* ret) { *ret = wp::div(x, y); }
 WP_API void builtin_floordiv_float16_float16(float16 x, float16 y, float16* ret) { *ret = wp::floordiv(x, y); }
 WP_API void builtin_floordiv_float32_float32(float32 x, float32 y, float32* ret) { *ret = wp::floordiv(x, y); }
 WP_API void builtin_floordiv_float64_float64(float64 x, float64 y, float64* ret) { *ret = wp::floordiv(x, y); }
@@ -1402,57 +1402,57 @@ WP_API void builtin_pos_uint16(uint16 x, uint16* ret) { *ret = wp::pos(x); }
 WP_API void builtin_pos_uint32(uint32 x, uint32* ret) { *ret = wp::pos(x); }
 WP_API void builtin_pos_uint64(uint64 x, uint64* ret) { *ret = wp::pos(x); }
 WP_API void builtin_pos_uint8(uint8 x, uint8* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2h(vec2h x, vec2h* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3h(vec3h x, vec3h* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4h(vec4h x, vec4h* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_spatial_vectorh(spatial_vectorh x, spatial_vectorh* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2f(vec2f x, vec2f* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3f(vec3f x, vec3f* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4f(vec4f x, vec4f* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_spatial_vectorf(spatial_vectorf x, spatial_vectorf* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2d(vec2d x, vec2d* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3d(vec3d x, vec3d* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4d(vec4d x, vec4d* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_spatial_vectord(spatial_vectord x, spatial_vectord* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2s(vec2s x, vec2s* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3s(vec3s x, vec3s* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4s(vec4s x, vec4s* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2i(vec2i x, vec2i* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3i(vec3i x, vec3i* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4i(vec4i x, vec4i* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2l(vec2l x, vec2l* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3l(vec3l x, vec3l* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4l(vec4l x, vec4l* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2b(vec2b x, vec2b* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3b(vec3b x, vec3b* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4b(vec4b x, vec4b* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2us(vec2us x, vec2us* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3us(vec3us x, vec3us* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4us(vec4us x, vec4us* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2ui(vec2ui x, vec2ui* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3ui(vec3ui x, vec3ui* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4ui(vec4ui x, vec4ui* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2ul(vec2ul x, vec2ul* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3ul(vec3ul x, vec3ul* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4ul(vec4ul x, vec4ul* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec2ub(vec2ub x, vec2ub* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec3ub(vec3ub x, vec3ub* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_vec4ub(vec4ub x, vec4ub* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_quath(quath x, quath* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_quatf(quatf x, quatf* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_quatd(quatd x, quatd* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat22h(mat22h x, mat22h* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat33h(mat33h x, mat33h* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat44h(mat44h x, mat44h* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_spatial_matrixh(spatial_matrixh x, spatial_matrixh* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat22f(mat22f x, mat22f* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat33f(mat33f x, mat33f* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat44f(mat44f x, mat44f* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_spatial_matrixf(spatial_matrixf x, spatial_matrixf* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat22d(mat22d x, mat22d* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat33d(mat33d x, mat33d* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_mat44d(mat44d x, mat44d* ret) { *ret = wp::pos(x); }
-WP_API void builtin_pos_spatial_matrixd(spatial_matrixd x, spatial_matrixd* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2h(vec2h& x, vec2h* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3h(vec3h& x, vec3h* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4h(vec4h& x, vec4h* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_spatial_vectorh(spatial_vectorh& x, spatial_vectorh* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2f(vec2f& x, vec2f* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3f(vec3f& x, vec3f* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4f(vec4f& x, vec4f* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_spatial_vectorf(spatial_vectorf& x, spatial_vectorf* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2d(vec2d& x, vec2d* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3d(vec3d& x, vec3d* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4d(vec4d& x, vec4d* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_spatial_vectord(spatial_vectord& x, spatial_vectord* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2s(vec2s& x, vec2s* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3s(vec3s& x, vec3s* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4s(vec4s& x, vec4s* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2i(vec2i& x, vec2i* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3i(vec3i& x, vec3i* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4i(vec4i& x, vec4i* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2l(vec2l& x, vec2l* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3l(vec3l& x, vec3l* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4l(vec4l& x, vec4l* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2b(vec2b& x, vec2b* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3b(vec3b& x, vec3b* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4b(vec4b& x, vec4b* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2us(vec2us& x, vec2us* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3us(vec3us& x, vec3us* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4us(vec4us& x, vec4us* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2ui(vec2ui& x, vec2ui* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3ui(vec3ui& x, vec3ui* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4ui(vec4ui& x, vec4ui* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2ul(vec2ul& x, vec2ul* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3ul(vec3ul& x, vec3ul* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4ul(vec4ul& x, vec4ul* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec2ub(vec2ub& x, vec2ub* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec3ub(vec3ub& x, vec3ub* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_vec4ub(vec4ub& x, vec4ub* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_quath(quath& x, quath* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_quatf(quatf& x, quatf* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_quatd(quatd& x, quatd* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat22h(mat22h& x, mat22h* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat33h(mat33h& x, mat33h* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat44h(mat44h& x, mat44h* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_spatial_matrixh(spatial_matrixh& x, spatial_matrixh* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat22f(mat22f& x, mat22f* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat33f(mat33f& x, mat33f* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat44f(mat44f& x, mat44f* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_spatial_matrixf(spatial_matrixf& x, spatial_matrixf* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat22d(mat22d& x, mat22d* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat33d(mat33d& x, mat33d* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_mat44d(mat44d& x, mat44d* ret) { *ret = wp::pos(x); }
+WP_API void builtin_pos_spatial_matrixd(spatial_matrixd& x, spatial_matrixd* ret) { *ret = wp::pos(x); }
 WP_API void builtin_neg_float16(float16 x, float16* ret) { *ret = wp::neg(x); }
 WP_API void builtin_neg_float32(float32 x, float32* ret) { *ret = wp::neg(x); }
 WP_API void builtin_neg_float64(float64 x, float64* ret) { *ret = wp::neg(x); }
@@ -1464,57 +1464,57 @@ WP_API void builtin_neg_uint16(uint16 x, uint16* ret) { *ret = wp::neg(x); }
 WP_API void builtin_neg_uint32(uint32 x, uint32* ret) { *ret = wp::neg(x); }
 WP_API void builtin_neg_uint64(uint64 x, uint64* ret) { *ret = wp::neg(x); }
 WP_API void builtin_neg_uint8(uint8 x, uint8* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2h(vec2h x, vec2h* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3h(vec3h x, vec3h* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4h(vec4h x, vec4h* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_spatial_vectorh(spatial_vectorh x, spatial_vectorh* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2f(vec2f x, vec2f* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3f(vec3f x, vec3f* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4f(vec4f x, vec4f* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_spatial_vectorf(spatial_vectorf x, spatial_vectorf* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2d(vec2d x, vec2d* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3d(vec3d x, vec3d* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4d(vec4d x, vec4d* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_spatial_vectord(spatial_vectord x, spatial_vectord* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2s(vec2s x, vec2s* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3s(vec3s x, vec3s* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4s(vec4s x, vec4s* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2i(vec2i x, vec2i* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3i(vec3i x, vec3i* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4i(vec4i x, vec4i* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2l(vec2l x, vec2l* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3l(vec3l x, vec3l* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4l(vec4l x, vec4l* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2b(vec2b x, vec2b* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3b(vec3b x, vec3b* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4b(vec4b x, vec4b* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2us(vec2us x, vec2us* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3us(vec3us x, vec3us* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4us(vec4us x, vec4us* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2ui(vec2ui x, vec2ui* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3ui(vec3ui x, vec3ui* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4ui(vec4ui x, vec4ui* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2ul(vec2ul x, vec2ul* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3ul(vec3ul x, vec3ul* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4ul(vec4ul x, vec4ul* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec2ub(vec2ub x, vec2ub* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec3ub(vec3ub x, vec3ub* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_vec4ub(vec4ub x, vec4ub* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_quath(quath x, quath* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_quatf(quatf x, quatf* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_quatd(quatd x, quatd* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat22h(mat22h x, mat22h* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat33h(mat33h x, mat33h* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat44h(mat44h x, mat44h* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_spatial_matrixh(spatial_matrixh x, spatial_matrixh* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat22f(mat22f x, mat22f* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat33f(mat33f x, mat33f* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat44f(mat44f x, mat44f* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_spatial_matrixf(spatial_matrixf x, spatial_matrixf* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat22d(mat22d x, mat22d* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat33d(mat33d x, mat33d* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_mat44d(mat44d x, mat44d* ret) { *ret = wp::neg(x); }
-WP_API void builtin_neg_spatial_matrixd(spatial_matrixd x, spatial_matrixd* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2h(vec2h& x, vec2h* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3h(vec3h& x, vec3h* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4h(vec4h& x, vec4h* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_spatial_vectorh(spatial_vectorh& x, spatial_vectorh* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2f(vec2f& x, vec2f* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3f(vec3f& x, vec3f* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4f(vec4f& x, vec4f* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_spatial_vectorf(spatial_vectorf& x, spatial_vectorf* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2d(vec2d& x, vec2d* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3d(vec3d& x, vec3d* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4d(vec4d& x, vec4d* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_spatial_vectord(spatial_vectord& x, spatial_vectord* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2s(vec2s& x, vec2s* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3s(vec3s& x, vec3s* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4s(vec4s& x, vec4s* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2i(vec2i& x, vec2i* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3i(vec3i& x, vec3i* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4i(vec4i& x, vec4i* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2l(vec2l& x, vec2l* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3l(vec3l& x, vec3l* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4l(vec4l& x, vec4l* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2b(vec2b& x, vec2b* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3b(vec3b& x, vec3b* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4b(vec4b& x, vec4b* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2us(vec2us& x, vec2us* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3us(vec3us& x, vec3us* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4us(vec4us& x, vec4us* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2ui(vec2ui& x, vec2ui* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3ui(vec3ui& x, vec3ui* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4ui(vec4ui& x, vec4ui* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2ul(vec2ul& x, vec2ul* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3ul(vec3ul& x, vec3ul* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4ul(vec4ul& x, vec4ul* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec2ub(vec2ub& x, vec2ub* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec3ub(vec3ub& x, vec3ub* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_vec4ub(vec4ub& x, vec4ub* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_quath(quath& x, quath* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_quatf(quatf& x, quatf* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_quatd(quatd& x, quatd* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat22h(mat22h& x, mat22h* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat33h(mat33h& x, mat33h* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat44h(mat44h& x, mat44h* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_spatial_matrixh(spatial_matrixh& x, spatial_matrixh* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat22f(mat22f& x, mat22f* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat33f(mat33f& x, mat33f* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat44f(mat44f& x, mat44f* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_spatial_matrixf(spatial_matrixf& x, spatial_matrixf* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat22d(mat22d& x, mat22d* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat33d(mat33d& x, mat33d* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_mat44d(mat44d& x, mat44d* ret) { *ret = wp::neg(x); }
+WP_API void builtin_neg_spatial_matrixd(spatial_matrixd& x, spatial_matrixd* ret) { *ret = wp::neg(x); }
 WP_API void builtin_unot_bool(bool b, bool* ret) { *ret = wp::unot(b); }
 WP_API void builtin_unot_int8(int8 b, bool* ret) { *ret = wp::unot(b); }
 WP_API void builtin_unot_uint8(uint8 b, bool* ret) { *ret = wp::unot(b); }

From 060e5f860ac23cedf656228b3d2f5ca1f8c5d725 Mon Sep 17 00:00:00 2001
From: Christopher Crouzet <christopher@crouzet.pm>
Date: Mon, 18 Dec 2023 11:46:02 +1300
Subject: [PATCH 41/50] Update the repository URL

---
 .teamcity/settings.kts | 2 +-
 PACKAGE-INFO.yaml      | 6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/.teamcity/settings.kts b/.teamcity/settings.kts
index 61211edf1..94d1d4c66 100644
--- a/.teamcity/settings.kts
+++ b/.teamcity/settings.kts
@@ -42,7 +42,7 @@ project {
 
 object GitlabMasterWarp : GitVcsRoot({
     name = "gitlab-master-warp"
-    url = "ssh://git@gitlab-master.nvidia.com:12051/mmacklin/warp.git"
+    url = "ssh://git@gitlab-master.nvidia.com:12051/omniverse/warp.git"
     branch = "refs/heads/tc"
     branchSpec = """
         +:refs/heads/WARForTeamCityNotFillingParamsUnlessThereIsABranchSpecDefined
diff --git a/PACKAGE-INFO.yaml b/PACKAGE-INFO.yaml
index 8df6f24db..fa9860079 100644
--- a/PACKAGE-INFO.yaml
+++ b/PACKAGE-INFO.yaml
@@ -1,6 +1,6 @@
 Package : warp
-Description : 
-Maintainers : 
+Description :
+Maintainers :
 SWIPAT NvBug :
-Repository : https://gitlab-master.nvidia.com/mmacklin/warp
+Repository : https://gitlab-master.nvidia.com/omniverse/warp
 License Type : NVIDIA

From 9ae387756329f4cc20b528bd65bc6d7d5478176c Mon Sep 17 00:00:00 2001
From: Zach Corse <zcorse@nvidia.com>
Date: Sun, 17 Dec 2023 18:54:17 -0800
Subject: [PATCH 42/50] matmul adjoints now accumulate

---
 warp/tests/test_matmul.py |  30 +++++++++
 warp/types.py             | 129 +++++++++++++-------------------------
 warp/utils.py             |  14 +++++
 3 files changed, 86 insertions(+), 87 deletions(-)

diff --git a/warp/tests/test_matmul.py b/warp/tests/test_matmul.py
index 27b404485..225770b12 100644
--- a/warp/tests/test_matmul.py
+++ b/warp/tests/test_matmul.py
@@ -331,6 +331,7 @@ def test_tape(test, device):
 
     tape.backward(loss=loss)
     A_grad = A.grad.numpy()
+    tape.reset()
 
     # test adjoint
     D.grad = wp.array2d(np.ones((m, n)), dtype=float, device=device)
@@ -420,6 +421,34 @@ def test_large_batch_count(test, device):
     assert np.array_equal(adj_C_np, C.grad.numpy())
 
 
+def test_adjoint_accumulation(test, device):
+    a_np = np.ones(shape=(2,3))
+    b_np = np.ones(shape=(3,2))
+    c_np = np.zeros(shape=(2,2))
+    d_np = np.zeros(shape=(2,2))
+
+    a_wp = wp.from_numpy(a_np, dtype=float, requires_grad=True)
+    b_wp = wp.from_numpy(b_np, dtype=float, requires_grad=True)
+    c_wp = wp.from_numpy(c_np, dtype=float, requires_grad=True)
+    d1_wp = wp.from_numpy(d_np, dtype=float, requires_grad=True)
+    d2_wp = wp.from_numpy(d_np, dtype=float, requires_grad=True)
+
+    tape = wp.Tape()
+
+    with tape:
+        wp.matmul(a_wp, b_wp, c_wp, d1_wp, alpha=1.0, beta=1.0)
+        wp.matmul(a_wp, b_wp, d1_wp, d2_wp, alpha=1.0, beta=1.0)
+
+    d_grad = wp.zeros_like(d2_wp)
+    d_grad.fill_(1.)
+    grads = {d2_wp : d_grad}
+    tape.backward(grads=grads) 
+
+    assert np.array_equal(a_wp.grad.numpy(), 4.0 * np.ones(shape=(2,3)))
+    assert np.array_equal(b_wp.grad.numpy(), 4.0 * np.ones(shape=(3,2)))
+    assert np.array_equal(c_wp.grad.numpy(), np.ones(shape=(2,2)))
+
+
 def register(parent):
     devices = [d for d in get_test_devices()]
 
@@ -437,6 +466,7 @@ class TestMatmul(parent):
             add_function_test(TestMatmul, "test_tape", test_tape, devices=devices)
             add_function_test(TestMatmul, "test_operator", test_operator, devices=devices)
             add_function_test(TestMatmul, "test_large_batch_count", test_large_batch_count, devices=devices)
+            add_function_test(TestMatmul, "test_adjoint_accumulation", test_adjoint_accumulation, devices=devices)
         else:
             print("Skipping matmul tests because CUTLASS is not supported in this build")
 
diff --git a/warp/types.py b/warp/types.py
index 3e056174b..d396780b6 100644
--- a/warp/types.py
+++ b/warp/types.py
@@ -3221,9 +3221,9 @@ def adj_matmul(
 
     # cpu fallback if no cuda devices found
     if device == "cpu":
-        adj_a.assign(alpha * np.matmul(adj_d.numpy(), b.numpy().transpose()))
-        adj_b.assign(alpha * (a.numpy().transpose() @ adj_d.numpy()))
-        adj_c.assign(beta * adj_d.numpy())
+        adj_a.assign(alpha * np.matmul(adj_d.numpy(), b.numpy().transpose()) + adj_a.numpy())
+        adj_b.assign(alpha * (a.numpy().transpose() @ adj_d.numpy()) + adj_b.numpy())
+        adj_c.assign(beta * adj_d.numpy() + adj_c.numpy())
         return
 
     cc = device.arch
@@ -3238,10 +3238,10 @@ def adj_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(adj_d.ptr),
             ctypes.c_void_p(b.ptr),
-            ctypes.c_void_p(a.ptr),
+            ctypes.c_void_p(adj_a.ptr),
             ctypes.c_void_p(adj_a.ptr),
             alpha,
-            0.0,
+            1.0,
             True,
             b.is_transposed,
             allow_tf32x3_arith,
@@ -3258,10 +3258,10 @@ def adj_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(b.ptr),
             ctypes.c_void_p(adj_d.ptr),
-            ctypes.c_void_p(a.ptr),
+            ctypes.c_void_p(adj_a.ptr),
             ctypes.c_void_p(adj_a.ptr),
             alpha,
-            0.0,
+            1.0,
             not b.is_transposed,
             False,
             allow_tf32x3_arith,
@@ -3280,10 +3280,10 @@ def adj_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(a.ptr),
             ctypes.c_void_p(adj_d.ptr),
-            ctypes.c_void_p(b.ptr),
+            ctypes.c_void_p(adj_b.ptr),
             ctypes.c_void_p(adj_b.ptr),
             alpha,
-            0.0,
+            1.0,
             a.is_transposed,
             True,
             allow_tf32x3_arith,
@@ -3300,10 +3300,10 @@ def adj_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(adj_d.ptr),
             ctypes.c_void_p(a.ptr),
-            ctypes.c_void_p(b.ptr),
+            ctypes.c_void_p(adj_b.ptr),
             ctypes.c_void_p(adj_b.ptr),
             alpha,
-            0.0,
+            1.0,
             False,
             not a.is_transposed,
             allow_tf32x3_arith,
@@ -3313,25 +3313,13 @@ def adj_matmul(
             raise RuntimeError("adj_matmul failed.")        
 
     # adj_c
-    ret = runtime.core.cutlass_gemm(
-        cc,
-        m,
-        n,
-        k,
-        type_typestr(a.dtype).encode(),
-        ctypes.c_void_p(a.ptr),
-        ctypes.c_void_p(b.ptr),
-        ctypes.c_void_p(adj_d.ptr),
-        ctypes.c_void_p(adj_c.ptr),
-        0.0,
-        beta,
-        not a.is_transposed,
-        not b.is_transposed,
-        allow_tf32x3_arith,
-        1,
+    warp.launch(
+        kernel=warp.utils.add_kernel_2d,
+        dim=adj_c.shape,
+        inputs=[adj_c, adj_d, adj_d.dtype(beta)],
+        device=device,
+        record_tape=False
     )
-    if not ret:
-        raise RuntimeError("adj_matmul failed.")
 
 
 def batched_matmul(
@@ -3540,9 +3528,9 @@ def adj_batched_matmul(
 
     # cpu fallback if no cuda devices found
     if device == "cpu":
-        adj_a.assign(alpha * np.matmul(adj_d.numpy(), b.numpy().transpose((0, 2, 1))))
-        adj_b.assign(alpha * np.matmul(a.numpy().transpose((0, 2, 1)), adj_d.numpy()))
-        adj_c.assign(beta * adj_d.numpy())
+        adj_a.assign(alpha * np.matmul(adj_d.numpy(), b.numpy().transpose((0, 2, 1))) + adj_a.numpy())
+        adj_b.assign(alpha * np.matmul(a.numpy().transpose((0, 2, 1)), adj_d.numpy()) + adj_b.numpy())
+        adj_c.assign(beta * adj_d.numpy() + adj_c.numpy())
         return
 
     # handle case in which batch_count exceeds max_batch_count, which is a CUDA array size maximum
@@ -3566,10 +3554,10 @@ def adj_batched_matmul(
                 type_typestr(a.dtype).encode(),
                 ctypes.c_void_p(adj_d[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(b[idx_start:idx_end,:,:].ptr),
-                ctypes.c_void_p(a[idx_start:idx_end,:,:].ptr),
+                ctypes.c_void_p(adj_a[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(adj_a[idx_start:idx_end,:,:].ptr),
                 alpha,
-                0.0,
+                1.0,
                 True,
                 b.is_transposed,
                 allow_tf32x3_arith,
@@ -3586,10 +3574,10 @@ def adj_batched_matmul(
                 type_typestr(a.dtype).encode(),
                 ctypes.c_void_p(b[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(adj_d[idx_start:idx_end,:,:].ptr),
-                ctypes.c_void_p(a[idx_start:idx_end,:,:].ptr),
+                ctypes.c_void_p(adj_a[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(adj_a[idx_start:idx_end,:,:].ptr),
                 alpha,
-                0.0,
+                1.0,
                 not b.is_transposed,
                 False,
                 allow_tf32x3_arith,
@@ -3608,10 +3596,10 @@ def adj_batched_matmul(
                 type_typestr(a.dtype).encode(),
                 ctypes.c_void_p(a[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(adj_d[idx_start:idx_end,:,:].ptr),
-                ctypes.c_void_p(b[idx_start:idx_end,:,:].ptr),
+                ctypes.c_void_p(adj_b[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(adj_b[idx_start:idx_end,:,:].ptr),
                 alpha,
-                0.0,
+                1.0,
                 a.is_transposed,
                 True,
                 allow_tf32x3_arith,
@@ -3628,10 +3616,10 @@ def adj_batched_matmul(
                 type_typestr(a.dtype).encode(),
                 ctypes.c_void_p(adj_d[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(a[idx_start:idx_end,:,:].ptr),
-                ctypes.c_void_p(b[idx_start:idx_end,:,:].ptr),
+                ctypes.c_void_p(adj_b[idx_start:idx_end,:,:].ptr),
                 ctypes.c_void_p(adj_b[idx_start:idx_end,:,:].ptr),
                 alpha,
-                0.0,
+                1.0,
                 False,
                 not a.is_transposed,
                 allow_tf32x3_arith,
@@ -3639,27 +3627,6 @@ def adj_batched_matmul(
             )
             if not ret:
                 raise RuntimeError("adj_matmul failed.")   
-
-        # adj_c
-        ret = runtime.core.cutlass_gemm(
-            cc,
-            m,
-            n,
-            k,
-            type_typestr(a.dtype).encode(),
-            ctypes.c_void_p(a[idx_start:idx_end,:,:].ptr),
-            ctypes.c_void_p(b[idx_start:idx_end,:,:].ptr),
-            ctypes.c_void_p(adj_d[idx_start:idx_end,:,:].ptr),
-            ctypes.c_void_p(adj_c[idx_start:idx_end,:,:].ptr),
-            0.0,
-            beta,
-            not a.is_transposed,
-            not b.is_transposed,
-            allow_tf32x3_arith,
-            max_batch_count,
-        )
-        if not ret:
-            raise RuntimeError("adj_batched_matmul failed.")
     
     idx_start = iters * max_batch_count
     
@@ -3673,10 +3640,10 @@ def adj_batched_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(adj_d[idx_start:,:,:].ptr),
             ctypes.c_void_p(b[idx_start:,:,:].ptr),
-            ctypes.c_void_p(a[idx_start:,:,:].ptr),
+            ctypes.c_void_p(adj_a[idx_start:,:,:].ptr),
             ctypes.c_void_p(adj_a[idx_start:,:,:].ptr),
             alpha,
-            0.0,
+            1.0,
             True,
             b.is_transposed,
             allow_tf32x3_arith,
@@ -3693,10 +3660,10 @@ def adj_batched_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(b[idx_start:,:,:].ptr),
             ctypes.c_void_p(adj_d[idx_start:,:,:].ptr),
-            ctypes.c_void_p(a[idx_start:,:,:].ptr),
+            ctypes.c_void_p(adj_a[idx_start:,:,:].ptr),
             ctypes.c_void_p(adj_a[idx_start:,:,:].ptr),
             alpha,
-            0.0,
+            1.0,
             not b.is_transposed,
             False,
             allow_tf32x3_arith,
@@ -3715,10 +3682,10 @@ def adj_batched_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(a[idx_start:,:,:].ptr),
             ctypes.c_void_p(adj_d[idx_start:,:,:].ptr),
-            ctypes.c_void_p(b[idx_start:,:,:].ptr),
+            ctypes.c_void_p(adj_b[idx_start:,:,:].ptr),
             ctypes.c_void_p(adj_b[idx_start:,:,:].ptr),
             alpha,
-            0.0,
+            1.0,
             a.is_transposed,
             True,
             allow_tf32x3_arith,
@@ -3735,10 +3702,10 @@ def adj_batched_matmul(
             type_typestr(a.dtype).encode(),
             ctypes.c_void_p(adj_d[idx_start:,:,:].ptr),
             ctypes.c_void_p(a[idx_start:,:,:].ptr),
-            ctypes.c_void_p(b[idx_start:,:,:].ptr),
+            ctypes.c_void_p(adj_b[idx_start:,:,:].ptr),
             ctypes.c_void_p(adj_b[idx_start:,:,:].ptr),
             alpha,
-            0.0,
+            1.0,
             False,
             not a.is_transposed,
             allow_tf32x3_arith,
@@ -3748,25 +3715,13 @@ def adj_batched_matmul(
             raise RuntimeError("adj_matmul failed.")   
 
     # adj_c
-    ret = runtime.core.cutlass_gemm(
-        cc,
-        m,
-        n,
-        k,
-        type_typestr(a.dtype).encode(),
-        ctypes.c_void_p(a[idx_start:,:,:].ptr),
-        ctypes.c_void_p(b[idx_start:,:,:].ptr),
-        ctypes.c_void_p(adj_d[idx_start:,:,:].ptr),
-        ctypes.c_void_p(adj_c[idx_start:,:,:].ptr),
-        0.0,
-        beta,
-        not a.is_transposed,
-        not b.is_transposed,
-        allow_tf32x3_arith,
-        remainder,
+    warp.launch(
+        kernel=warp.utils.add_kernel_3d,
+        dim=adj_c.shape,
+        inputs=[adj_c, adj_d, adj_d.dtype(beta)],
+        device=device,
+        record_tape=False
     )
-    if not ret:
-        raise RuntimeError("adj_batched_matmul failed.")
 
 class HashGrid:
     def __init__(self, dim_x, dim_y, dim_z, device=None):
diff --git a/warp/utils.py b/warp/utils.py
index e38c14326..1ce30d52c 100644
--- a/warp/utils.py
+++ b/warp/utils.py
@@ -666,3 +666,17 @@ def __exit__(self, exc_type, exc_value, traceback):
                 print("{}{} took {:.2f} ms".format(indent, self.name, self.elapsed))
 
             ScopedTimer.indent -= 1
+
+
+# helper kernels for adj_matmul
+@wp.kernel
+def add_kernel_2d(x: wp.array2d(dtype=Any), acc: wp.array2d(dtype=Any), beta: Any):
+    i, j = wp.tid()
+
+    x[i,j] = x[i,j] + beta * acc[i,j]
+
+@wp.kernel
+def add_kernel_3d(x: wp.array3d(dtype=Any), acc: wp.array3d(dtype=Any), beta: Any):
+    i, j, k = wp.tid()
+
+    x[i,j,k] = x[i,j,k] + beta * acc[i,j,k]

From 647fd35c5303f038558cb6b4b8543deaff8ceac5 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Fri, 15 Dec 2023 21:30:52 -0500
Subject: [PATCH 43/50] Support building pip wheels for Linux AArch64

Previously we relied solely on the extension of the `warp/bin/`
libraries to determine the wheel's binary distribution platform. This no
longer works since Linux builds for x86-64 and AArch64 both use the
`.so` extension.

Instead this change relies on the libraries to be placed in a
subdirectory of `warp/bin/` which indicates one of (currently) four
platforms:
- windows-x86_64
- linux-x86_64
- linux-aarch64
- macos-universal

CD systems are responsible for putting the libraries in these
subdirectories.

Creating a wheel from a local build is still supported, by assuming the
wheel's architecture should match the system's architecture.
---
 PACKAGING.md |  7 +++--
 setup.py     | 89 ++++++++++++++++++++++++++++++++++++++--------------
 2 files changed, 70 insertions(+), 26 deletions(-)

diff --git a/PACKAGING.md b/PACKAGING.md
index 10ef78bd2..db64ff7e6 100644
--- a/PACKAGING.md
+++ b/PACKAGING.md
@@ -73,9 +73,10 @@ Release Steps:
 
 3) Run
     ```bash
-    python -m build --wheel -C--build-option=-Pwindows &&
-    python -m build --wheel -C--build-option=-Plinux &&
-    python -m build --wheel -C--build-option=-Pmacos
+    python -m build --wheel -C--build-option=-Pwindows-x86_64 &&
+    python -m build --wheel -C--build-option=-Plinux-x86_64 &&
+    python -m build --wheel -C--build-option=-Plinux-aarch64 &&
+    python -m build --wheel -C--build-option=-Pmacos-universal
     ```
 
 4) Run `python -m twine upload dist/*`
diff --git a/setup.py b/setup.py
index b3897edf0..9a80c2dc1 100644
--- a/setup.py
+++ b/setup.py
@@ -1,6 +1,8 @@
 import argparse
 import os
 import shutil
+import pathlib
+import platform
 from typing import NamedTuple
 
 import setuptools
@@ -10,46 +12,80 @@
 # ourselves because when bdist_wheel runs it's too late to select a subset of libraries for package_data.
 parser = argparse.ArgumentParser()
 parser.add_argument("command")
-parser.add_argument("--platform", "-P", type=str, default="", help="Wheel platform: windows|linux|macos")
+parser.add_argument(
+    "--platform", "-P", type=str, default="", help="Wheel platform: windows|linux|macos-x86_64|aarch64|universal"
+)
 args = parser.parse_known_args()[0]
 
 
+# return a canonical machine architecture string
+# - "x86_64" for x86-64, aka. AMD64, aka. x64
+# - "aarch64" for AArch64, aka. ARM64
+def machine_architecture() -> str:
+    machine = platform.machine()
+    if machine == "x86_64" or machine == "AMD64":
+        return "x86_64"
+    if machine == "aarch64" or machine == "arm64":
+        return "aarch64"
+    raise RuntimeError(f"Unrecognized machine architecture {machine}")
+
+
 class Platform(NamedTuple):
-    name: str
+    os: str
+    arch: str
     fancy_name: str
     extension: str
     tag: str
 
+    def name(self) -> str:
+        return self.os + "-" + self.arch
+
 
 platforms = [
-    Platform("windows", "Windows", ".dll", "win_amd64"),
-    Platform("linux", "Linux", ".so", "manylinux2014_x86_64"),
-    Platform("macos", "macOS", ".dylib", "macosx_10_13_universal2"),
+    Platform("windows", "x86_64", "Windows x86-64", ".dll", "win_amd64"),
+    Platform("linux", "x86_64", "Linux x86-64", ".so", "manylinux2014_x86_64"),
+    Platform("linux", "aarch64", "Linux AArch64", ".so", "manylinux2014_aarch64"),
+    Platform("macos", "universal", "macOS universal", ".dylib", "macosx_10_13_universal2"),
 ]
 
 
-# Determine supported platforms of warp/bin libraries based on their extension
-def detect_warp_platforms():
-    detected_platforms = set()
-    for filename in os.listdir("warp/bin"):
-        for p in platforms:
-            if os.path.splitext(filename)[1] == p.extension:
-                detected_platforms.add(p)
+class Library(NamedTuple):
+    file: str
+    directory: str
+    platform: Platform
+
 
-    if len(detected_platforms) == 0:
+# Enumerate warp/bin libraries
+def detect_warp_libraries():
+    detected_libraries = set()
+    warp_bin = pathlib.Path("warp/bin")
+    for file in warp_bin.rglob("*.*"):
+        for p in platforms:
+            if os.path.splitext(file.name)[1] == p.extension:
+                # If this is a local build, assume we want a wheel for this machine's architecture
+                if file.parent.name == "bin" and (p.arch == machine_architecture() or p.arch == "universal"):
+                    detected_libraries.add(Library(file.name, "bin/", p))
+                else:
+                    # Excpect libraries to be in a subdirectory named after the wheel platform
+                    platform_name = p.name()
+                    if file.parent.name == platform_name:
+                        detected_libraries.add(Library(file.name, "bin/" + platform_name + "/", p))
+
+    if len(detected_libraries) == 0:
         raise Exception("No libraries found in warp/bin. Please run build_lib.py first.")
 
-    return detected_platforms
+    return detected_libraries
 
 
-detected_platforms = detect_warp_platforms()
+detected_libraries = detect_warp_libraries()
+detected_platforms = set([lib.platform for lib in detected_libraries])
 
 wheel_platform = None  # The one platform for which we're building a wheel
 
 if args.command == "bdist_wheel":
     if args.platform != "":
         for p in platforms:
-            if args.platform == p.name or args.platform == p.fancy_name:
+            if args.platform == p.name():
                 wheel_platform = p
                 print(f"Platform argument specified for building {p.fancy_name} wheel")
                 break
@@ -64,7 +100,9 @@ def detect_warp_platforms():
     if wheel_platform is None:
         if len(detected_platforms) > 1:
             print("Libraries for multiple platforms were detected. Picking the first one.")
-            print("Run `python -m build --wheel -C--build-option=-P[windows|linux|macos]` to select a specific one.")
+            print(
+                "Run `python -m build --wheel -C--build-option=-P[windows|linux|macos]-[x86_64|aarch64|universal]` to select a specific one."
+            )
         wheel_platform = next(iter(detected_platforms))
 
     print("Creating Warp wheel for " + wheel_platform.fancy_name)
@@ -77,7 +115,7 @@ class WarpBDistWheel(bdist_wheel):
     # Even though we parse the platform argument ourselves, we need to declare it here as well so
     # setuptools.Command can validate the command line options.
     user_options = bdist_wheel.user_options + [
-        ("platform=", "P", "Wheel platform: windows|linux|macos"),
+        ("platform=", "P", "Wheel platform: windows|linux|macos-x86_64|aarch64|universal"),
     ]
 
     def initialize_options(self):
@@ -109,17 +147,22 @@ def has_ext_modules(self):
         return True
 
 
-def get_warp_libraries(extension):
+def get_warp_libraries(platform):
     libraries = []
-    for filename in os.listdir("warp/bin"):
-        if os.path.splitext(filename)[1] == extension:
-            libraries.append("bin/" + filename)
+    for library in detected_libraries:
+        if library.platform == platform:
+            src = "warp/" + library.directory + library.file
+            dst = "warp/bin/" + library.file
+            if src != dst:
+                shutil.copyfile(src, dst)
+
+            libraries.append("bin/" + library.file)
 
     return libraries
 
 
 if wheel_platform is not None:
-    warp_binary_libraries = get_warp_libraries(wheel_platform.extension)
+    warp_binary_libraries = get_warp_libraries(wheel_platform)
 else:
     warp_binary_libraries = []  # Not needed during egg_info command
 

From 72072ef2368887e1e2ccd563cb2b31802236f656 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Thu, 14 Dec 2023 16:22:27 -0500
Subject: [PATCH 44/50] Build CUDA static code for Jetson into warp.so

`warp.so` contains compiled versions of `warp/native/warp.cu` for each
supported GPU architecture. Add the ones for Tegra/Jetson when we're on
an AArch64 Linux system.
---
 warp/build_dll.py | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/warp/build_dll.py b/warp/build_dll.py
index 60e54325e..65d8c9df5 100644
--- a/warp/build_dll.py
+++ b/warp/build_dll.py
@@ -182,12 +182,15 @@ def build_dll_for_arch(dll_path, cpp_paths, cu_path, libs, mode, arch, verify_fp
                 "-gencode=arch=compute_75,code=sm_75",  # Turing
                 "-gencode=arch=compute_80,code=sm_80",  # Ampere
                 "-gencode=arch=compute_86,code=sm_86",
-                # SASS for supported mobile architectures (e.g. Tegra/Jetson)
-                # "-gencode=arch=compute_53,code=sm_53",
-                # "-gencode=arch=compute_62,code=sm_62",
-                # "-gencode=arch=compute_72,code=sm_72",
-                # "-gencode=arch=compute_87,code=sm_87",
             ]
+            if arch == "arm64" and sys.platform() == "linux":
+                gencode_opts += [
+                    # SASS for supported mobile architectures (e.g. Tegra/Jetson)
+                    "-gencode=arch=compute_53,code=sm_53",  # X1
+                    "-gencode=arch=compute_62,code=sm_62",  # X2
+                    "-gencode=arch=compute_72,code=sm_72",  # Xavier
+                    "-gencode=arch=compute_87,code=sm_87",  # Orin
+                ]
 
             # support for Ada and Hopper is available with CUDA Toolkit 11.8+
             if ctk_version >= (11, 8):

From 089bb3d4e9ca160e290465656ecaa8ce9c6ee8e5 Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Tue, 19 Dec 2023 13:42:21 -0800
Subject: [PATCH 45/50] Add parallel testing runner

---
 CHANGELOG.md                                  |    2 +-
 docs/debugging.rst                            |    2 +-
 examples/example_dem.py                       |   54 +-
 examples/example_diffray.py                   |   16 +-
 examples/example_fluid.py                     |   10 +-
 examples/example_mesh_intersect.py            |    5 +-
 examples/example_sim_cartpole.py              |   10 +-
 examples/example_sim_cloth.py                 |   36 +-
 examples/example_sim_grad_bounce.py           |   15 +-
 examples/example_sim_grad_cloth.py            |   15 +-
 examples/example_sim_quadruped.py             |    9 +-
 examples/example_sim_rigid_chain.py           |    9 +-
 examples/example_sim_rigid_contact.py         |    9 +-
 examples/example_sim_trajopt.py               |    3 +-
 examples/fem/example_diffusion_mgpu.py        |   34 +-
 .../omni/warp/core/tests/test_ext.py          |  156 +-
 pyproject.toml                                |   10 +-
 tools/ci/testing/test-linux-aarch64/test.sh   |    2 +-
 tools/ci/testing/test-linux-x86_64/test.sh    |   23 +-
 tools/ci/testing/test-windows-x86_64/test.bat |    5 +-
 warp/config.py                                |    2 +
 warp/context.py                               |    5 +-
 warp/tests/__main__.py                        |    9 +-
 ...ass_kernel.py => aux_test_class_kernel.py} |   10 +-
 ...my.py => aux_test_compile_consts_dummy.py} |    0
 ..._test_conditional_unequal_types_kernels.py |   21 +
 ...est_dependent.py => aux_test_dependent.py} |    4 +-
 ...est_reference.py => aux_test_reference.py} |    2 +-
 ...nce.py => aux_test_reference_reference.py} |    0
 .../{test_square.py => aux_test_square.py}    |    0
 warp/tests/aux_test_unresolved_func.py        |   14 +
 warp/tests/aux_test_unresolved_symbol.py      |   14 +
 ...t_kinematics.py => disabled_kinematics.py} |   22 +-
 warp/tests/run_coverage_serial.py             |   31 +
 warp/tests/test_adam.py                       |   24 +-
 warp/tests/test_all.py                        |  241 --
 warp/tests/test_arithmetic.py                 |   79 +-
 warp/tests/test_array.py                      |   94 +-
 warp/tests/test_array_reduce.py               |   44 +-
 warp/tests/test_atomic.py                     |   34 +-
 warp/tests/test_bool.py                       |   27 +-
 warp/tests/test_builtins_resolution.py        | 1706 ++++-----
 warp/tests/test_bvh.py                        |   21 +-
 warp/tests/test_closest_point_edge_edge.py    |  110 +-
 warp/tests/test_codegen.py                    |  278 +-
 warp/tests/test_compile_consts.py             |   32 +-
 warp/tests/test_conditional.py                |   62 +-
 .../test_conditional_unequal_types_kernels.py |   14 -
 warp/tests/test_copy.py                       |   21 +-
 warp/tests/test_coverage.py                   |   38 -
 warp/tests/test_ctypes.py                     |  160 +-
 warp/tests/test_dense.py                      |   34 +-
 warp/tests/test_devices.py                    |   69 +-
 warp/tests/test_dlpack.py                     |  149 +-
 warp/tests/test_examples.py                   |  313 +-
 warp/tests/test_fabricarray.py                |   36 +-
 warp/tests/test_fast_math.py                  |   25 +-
 warp/tests/test_fem.py                        |   68 +-
 warp/tests/test_fp16.py                       |   34 +-
 warp/tests/test_func.py                       |  371 +-
 warp/tests/test_generics.py                   |  146 +-
 warp/tests/test_grad.py                       |   66 +-
 warp/tests/test_grad_customs.py               |   16 +-
 warp/tests/test_hash_grid.py                  |   16 +-
 warp/tests/test_import.py                     |   32 +-
 warp/tests/test_indexedarray.py               |   40 +-
 warp/tests/test_intersect.py                  |   24 +-
 warp/tests/test_large.py                      |   36 +-
 warp/tests/test_launch.py                     |   31 +-
 warp/tests/test_lerp.py                       |  126 +-
 warp/tests/test_lvalue.py                     |   71 +-
 warp/tests/test_marching_cubes.py             |   22 +-
 warp/tests/test_mat.py                        | 3409 ++---------------
 warp/tests/test_mat_lite.py                   |   21 +-
 warp/tests/test_mat_scalar_ops.py             | 2889 ++++++++++++++
 warp/tests/test_math.py                       |   21 +-
 warp/tests/test_matmul.py                     |  176 +-
 warp/tests/test_matmul_lite.py                |  169 +-
 warp/tests/test_mesh.py                       |   67 +-
 warp/tests/test_mesh_query_aabb.py            |   43 +-
 warp/tests/test_mesh_query_point.py           |   34 +-
 warp/tests/test_mesh_query_ray.py             |   30 +-
 warp/tests/test_mlp.py                        |   17 +-
 warp/tests/test_model.py                      |  182 +-
 warp/tests/test_modules_lite.py               |   40 +-
 warp/tests/test_multigpu.py                   |  201 +-
 warp/tests/test_noise.py                      |   22 +-
 warp/tests/test_operators.py                  |   35 +-
 warp/tests/test_options.py                    |   21 +-
 warp/tests/test_pinned.py                     |   34 +-
 warp/tests/test_print.py                      |   36 +-
 warp/tests/test_quat.py                       |  187 +-
 warp/tests/test_rand.py                       |   25 +-
 warp/tests/test_reload.py                     |   59 +-
 warp/tests/test_rounding.py                   |   17 +-
 warp/tests/test_runlength_encode.py           |  211 +-
 warp/tests/test_smoothstep.py                 |   17 +-
 warp/tests/test_snippet.py                    |   35 +-
 warp/tests/test_sparse.py                     |   59 +-
 warp/tests/test_spatial.py                    |  353 +-
 warp/tests/test_streams.py                    |  207 +-
 warp/tests/test_struct.py                     |  165 +-
 warp/tests/test_tape.py                       |   28 +-
 warp/tests/test_torch.py                      |  175 +-
 warp/tests/test_transient_module.py           |   21 +-
 warp/tests/test_types.py                      |  730 ++--
 warp/tests/test_unresolved_func.py            |    7 -
 warp/tests/test_unresolved_symbol.py          |    7 -
 warp/tests/test_utils.py                      |  435 ++-
 warp/tests/test_vec.py                        | 2308 +----------
 warp/tests/test_vec_lite.py                   |   19 +-
 warp/tests/test_vec_scalar_ops.py             | 2099 ++++++++++
 warp/tests/test_volume.py                     |  773 ++--
 warp/tests/test_volume_write.py               |  257 +-
 warp/tests/unittest_serial.py                 |   35 +
 warp/tests/unittest_suites.py                 |  291 ++
 .../tests/{test_base.py => unittest_utils.py} |  163 +-
 .../{test_misc.py => unused_test_misc.py}     |   18 +-
 .../{test_debug.py => walkthough_debug.py}    |   17 +-
 warp/thirdparty/unittest_parallel.py          |  313 +-
 120 files changed, 11044 insertions(+), 10703 deletions(-)
 rename warp/tests/{test_class_kernel.py => aux_test_class_kernel.py} (54%)
 rename warp/tests/{test_compile_consts_dummy.py => aux_test_compile_consts_dummy.py} (100%)
 create mode 100644 warp/tests/aux_test_conditional_unequal_types_kernels.py
 rename warp/tests/{test_dependent.py => aux_test_dependent.py} (85%)
 rename warp/tests/{test_reference.py => aux_test_reference.py} (72%)
 rename warp/tests/{test_reference_reference.py => aux_test_reference_reference.py} (100%)
 rename warp/tests/{test_square.py => aux_test_square.py} (100%)
 create mode 100644 warp/tests/aux_test_unresolved_func.py
 create mode 100644 warp/tests/aux_test_unresolved_symbol.py
 rename warp/tests/{test_kinematics.py => disabled_kinematics.py} (94%)
 create mode 100644 warp/tests/run_coverage_serial.py
 delete mode 100644 warp/tests/test_all.py
 delete mode 100644 warp/tests/test_conditional_unequal_types_kernels.py
 delete mode 100644 warp/tests/test_coverage.py
 create mode 100644 warp/tests/test_mat_scalar_ops.py
 delete mode 100644 warp/tests/test_unresolved_func.py
 delete mode 100644 warp/tests/test_unresolved_symbol.py
 create mode 100644 warp/tests/test_vec_scalar_ops.py
 create mode 100644 warp/tests/unittest_serial.py
 create mode 100644 warp/tests/unittest_suites.py
 rename warp/tests/{test_base.py => unittest_utils.py} (54%)
 rename warp/tests/{test_misc.py => unused_test_misc.py} (65%)
 rename warp/tests/{test_debug.py => walkthough_debug.py} (88%)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index d98de5de3..53b680d60 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -148,7 +148,7 @@
 ## [0.9.0] - 2023-06-01
 
 - Add support for in-place modifications to vector, matrix, and struct types inside kernels (will warn during backward pass with `wp.verbose` if using gradients)
-- Add support for step-through VSCode debugging of kernel code with standalone LLVM compiler, see `wp.breakpoint()`, and `test_debug.py`
+- Add support for step-through VSCode debugging of kernel code with standalone LLVM compiler, see `wp.breakpoint()`, and `walkthrough_debug.py`
 - Add support for default values on built-in functions
 - Add support for multi-valued `@wp.func` functions
 - Add support for `pass`, `continue`, and `break` statements
diff --git a/docs/debugging.rst b/docs/debugging.rst
index c006bd226..4ce66e0d1 100644
--- a/docs/debugging.rst
+++ b/docs/debugging.rst
@@ -42,7 +42,7 @@ the debugger should be attached, and a breakpoint inserted into the generated co
 
 .. note:: Generated kernel code is not a 1:1 correspondence with the original Python code, but individual operations can still be replayed and variables inspected.
 
-Also see :github:`warp/tests/test_debug.py` for an example of how to debug Warp kernel code running on the CPU.
+Also see :github:`warp/tests/walkthrough_debug.py` for an example of how to debug Warp kernel code running on the CPU.
 
 Generated Code
 --------------
diff --git a/examples/example_dem.py b/examples/example_dem.py
index 44485c882..250f691e7 100644
--- a/examples/example_dem.py
+++ b/examples/example_dem.py
@@ -118,6 +118,8 @@ def integrate(
 
 class Example:
     def __init__(self, stage):
+        self.device = wp.get_device()
+
         self.frame_dt = 1.0 / 60
         self.frame_count = 400
 
@@ -150,32 +152,32 @@ def __init__(self, stage):
         self.use_graph = wp.get_device().is_cuda
 
         if self.use_graph:
-            wp.capture_begin()
-
-            for _ in range(self.sim_substeps):
-                with wp.ScopedTimer("forces", active=False):
-                    wp.launch(
-                        kernel=apply_forces,
-                        dim=len(self.x),
-                        inputs=[
-                            self.grid.id,
-                            self.x,
-                            self.v,
-                            self.f,
-                            self.point_radius,
-                            self.k_contact,
-                            self.k_damp,
-                            self.k_friction,
-                            self.k_mu,
-                        ],
-                    )
-                    wp.launch(
-                        kernel=integrate,
-                        dim=len(self.x),
-                        inputs=[self.x, self.v, self.f, (0.0, -9.8, 0.0), self.sim_dt, self.inv_mass],
-                    )
-
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                for _ in range(self.sim_substeps):
+                    with wp.ScopedTimer("forces", active=False):
+                        wp.launch(
+                            kernel=apply_forces,
+                            dim=len(self.x),
+                            inputs=[
+                                self.grid.id,
+                                self.x,
+                                self.v,
+                                self.f,
+                                self.point_radius,
+                                self.k_contact,
+                                self.k_damp,
+                                self.k_friction,
+                                self.k_mu,
+                            ],
+                        )
+                        wp.launch(
+                            kernel=integrate,
+                            dim=len(self.x),
+                            inputs=[self.x, self.v, self.f, (0.0, -9.8, 0.0), self.sim_dt, self.inv_mass],
+                        )
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True):
diff --git a/examples/example_diffray.py b/examples/example_diffray.py
index c02af5739..bca2c01f7 100644
--- a/examples/example_diffray.py
+++ b/examples/example_diffray.py
@@ -304,6 +304,8 @@ class Example:
     """
 
     def __init__(self, stage=None, rot_array=[0.0, 0.0, 0.0, 1.0], verbose=False):
+        self.device = wp.get_device()
+
         self.verbose = verbose
         cam_pos = wp.vec3(0.0, 0.75, 7.0)
         cam_rot = wp.quat(0.0, 0.0, 0.0, 1.0)
@@ -416,12 +418,14 @@ def __init__(self, stage=None, rot_array=[0.0, 0.0, 0.0, 1.0], verbose=False):
         self.loss = wp.zeros(1, dtype=float, requires_grad=True)
 
         # capture graph
-        wp.capture_begin()
-        self.tape = wp.Tape()
-        with self.tape:
-            self.compute_loss()
-        self.tape.backward(self.loss)
-        self.graph = wp.capture_end()
+        wp.capture_begin(self.device)
+        try:
+            self.tape = wp.Tape()
+            with self.tape:
+                self.compute_loss()
+            self.tape.backward(self.loss)
+        finally:
+            self.graph = wp.capture_end(self.device)
 
         self.optimizer = SGD(
             [self.render_mesh.rot],
diff --git a/examples/example_fluid.py b/examples/example_fluid.py
index 50a046ba0..ddd77edc9 100644
--- a/examples/example_fluid.py
+++ b/examples/example_fluid.py
@@ -165,6 +165,8 @@ def init(rho: wp.array2d(dtype=float), u: wp.array2d(dtype=wp.vec2), radius: int
 
 class Example:
     def __init__(self, **kwargs):
+        self.device = wp.get_device()
+
         self.sim_fps = 60.0
         self.sim_substeps = 2
         self.iterations = 100
@@ -187,9 +189,11 @@ def __init__(self, **kwargs):
 
         # capture pressure solve as a CUDA graph
         if self.device.is_cuda:
-            wp.capture_begin()
-            self.pressure_iterations()
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                self.pressure_iterations()
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def update(self):
         with wp.ScopedTimer("update"):
diff --git a/examples/example_mesh_intersect.py b/examples/example_mesh_intersect.py
index e5e1b6de9..c399ad9ce 100644
--- a/examples/example_mesh_intersect.py
+++ b/examples/example_mesh_intersect.py
@@ -85,6 +85,7 @@ class Example:
     def __init__(self, stage):
         rng = np.random.default_rng()
 
+        self.device = wp.get_device()
         self.query_count = 1024
         self.has_queried = False
 
@@ -117,8 +118,8 @@ def __init__(self, stage):
         self.array_result = wp.zeros(self.query_count, dtype=int)
         self.array_xforms = wp.array(self.xforms, dtype=wp.transform)
 
-        # force module load (for accurate profiling)
-        wp.force_load()
+        # compile and load the module up front (for accurate profiling)
+        wp.load_module(device=self.device)
 
     def update(self):
         with wp.ScopedTimer("intersect", active=True):
diff --git a/examples/example_sim_cartpole.py b/examples/example_sim_cartpole.py
index 5c2724d8f..7f1d2779a 100644
--- a/examples/example_sim_cartpole.py
+++ b/examples/example_sim_cartpole.py
@@ -28,6 +28,8 @@
 
 class Example:
     def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=True):
+        self.device = wp.get_device()
+
         builder = wp.sim.ModelBuilder()
 
         self.num_envs = num_envs
@@ -98,9 +100,11 @@ def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=T
 
         if self.use_graph:
             # create update graph
-            wp.capture_begin()
-            self.update()
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                self.update()
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True, print=self.print_timers):
diff --git a/examples/example_sim_cloth.py b/examples/example_sim_cloth.py
index 5ec7dfcf7..40b5982e3 100644
--- a/examples/example_sim_cloth.py
+++ b/examples/example_sim_cloth.py
@@ -37,19 +37,10 @@ def __str__(self):
 
 
 class Example:
-    def __init__(self, stage):
-        self.parser = argparse.ArgumentParser()
-        self.parser.add_argument(
-            "--integrator",
-            help="Type of integrator",
-            type=IntegratorType,
-            choices=list(IntegratorType),
-            default=IntegratorType.EULER,
-        )
-
-        args = self.parser.parse_args()
+    def __init__(self, stage, integrator=IntegratorType.EULER):
+        self.device = wp.get_device()
 
-        self.integrator_type = args.integrator
+        self.integrator_type = integrator
 
         self.sim_width = 64
         self.sim_height = 32
@@ -135,9 +126,11 @@ def __init__(self, stage):
 
         if self.sim_use_graph:
             # create update graph
-            wp.capture_begin()
-            self.update()
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                self.update()
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def update(self):
         with wp.ScopedTimer("simulate", dict=self.profiler):
@@ -168,9 +161,20 @@ def render(self, is_live=False):
 
 
 if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--integrator",
+        help="Type of integrator",
+        type=IntegratorType,
+        choices=list(IntegratorType),
+        default=IntegratorType.EULER,
+    )
+
+    args = parser.parse_args()
+
     stage_path = os.path.join(os.path.dirname(__file__), "outputs/example_sim_cloth.usd")
 
-    example = Example(stage_path)
+    example = Example(stage_path, integrator=args.integrator)
 
     for i in range(example.sim_frames):
         example.update()
diff --git a/examples/example_sim_grad_bounce.py b/examples/example_sim_grad_bounce.py
index 526f86742..833f7b6da 100644
--- a/examples/example_sim_grad_bounce.py
+++ b/examples/example_sim_grad_bounce.py
@@ -45,6 +45,7 @@ def step_kernel(x: wp.array(dtype=wp.vec3), grad: wp.array(dtype=wp.vec3), alpha
 
 class Example:
     def __init__(self, stage=None, enable_rendering=True, profile=False, adapter=None, verbose=False):
+        self.device = wp.get_device()
         self.verbose = verbose
 
         # seconds
@@ -106,12 +107,14 @@ def __init__(self, stage=None, enable_rendering=True, profile=False, adapter=Non
             self.renderer = wp.sim.render.SimRenderer(self.model, stage, scaling=1.0)
 
         # capture forward/backward passes
-        wp.capture_begin()
-        self.tape = wp.Tape()
-        with self.tape:
-            self.compute_loss()
-        self.tape.backward(self.loss)
-        self.graph = wp.capture_end()
+        wp.capture_begin(self.device)
+        try:
+            self.tape = wp.Tape()
+            with self.tape:
+                self.compute_loss()
+            self.tape.backward(self.loss)
+        finally:
+            self.graph = wp.capture_end(self.device)
 
     def compute_loss(self):
         # run control loop
diff --git a/examples/example_sim_grad_cloth.py b/examples/example_sim_grad_cloth.py
index d0afb6905..694461c0a 100644
--- a/examples/example_sim_grad_cloth.py
+++ b/examples/example_sim_grad_cloth.py
@@ -53,6 +53,7 @@ def step_kernel(x: wp.array(dtype=wp.vec3), grad: wp.array(dtype=wp.vec3), alpha
 
 class Example:
     def __init__(self, stage, profile=False, adapter=None, verbose=False):
+        self.device = wp.get_device()
         self.verbose = verbose
 
         # seconds
@@ -115,12 +116,14 @@ def __init__(self, stage, profile=False, adapter=None, verbose=False):
         self.renderer = wp.sim.render.SimRenderer(self.model, stage, scaling=4.0)
 
         # capture forward/backward passes
-        wp.capture_begin()
-        self.tape = wp.Tape()
-        with self.tape:
-            self.compute_loss()
-        self.tape.backward(self.loss)
-        self.graph = wp.capture_end()
+        wp.capture_begin(self.device)
+        try:
+            self.tape = wp.Tape()
+            with self.tape:
+                self.compute_loss()
+            self.tape.backward(self.loss)
+        finally:
+            self.graph = wp.capture_end(self.device)
 
     def compute_loss(self):
         # run control loop
diff --git a/examples/example_sim_quadruped.py b/examples/example_sim_quadruped.py
index f0413f40b..9fea24476 100644
--- a/examples/example_sim_quadruped.py
+++ b/examples/example_sim_quadruped.py
@@ -70,6 +70,7 @@ def compute_env_offsets(num_envs, env_offset=(5.0, 0.0, 5.0), up_axis="Y"):
 
 class Example:
     def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=True):
+        self.device = wp.get_device()
         self.num_envs = num_envs
         articulation_builder = wp.sim.ModelBuilder()
         wp.sim.parse_urdf(
@@ -135,9 +136,11 @@ def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=T
 
         if self.use_graph:
             # create update graph
-            wp.capture_begin()
-            self.update()
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                self.update()
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True, print=self.print_timers):
diff --git a/examples/example_sim_rigid_chain.py b/examples/example_sim_rigid_chain.py
index 4309536ca..fd33ec03e 100644
--- a/examples/example_sim_rigid_chain.py
+++ b/examples/example_sim_rigid_chain.py
@@ -27,6 +27,7 @@
 
 class Example:
     def __init__(self, stage):
+        self.device = wp.get_device()
         self.chain_length = 8
         self.chain_width = 1.0
         self.chain_types = [
@@ -146,9 +147,11 @@ def __init__(self, stage):
 
         if self.use_graph:
             # create update graph
-            wp.capture_begin()
-            self.update()
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                self.update()
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True):
diff --git a/examples/example_sim_rigid_contact.py b/examples/example_sim_rigid_contact.py
index 53450ce12..d3c84917e 100644
--- a/examples/example_sim_rigid_contact.py
+++ b/examples/example_sim_rigid_contact.py
@@ -28,6 +28,7 @@
 
 class Example:
     def __init__(self, stage):
+        self.device = wp.get_device()
         builder = wp.sim.ModelBuilder()
 
         self.sim_time = 0.0
@@ -126,9 +127,11 @@ def __init__(self, stage):
 
         if self.use_graph:
             # create update graph
-            wp.capture_begin()
-            self.update()
-            self.graph = wp.capture_end()
+            wp.capture_begin(self.device)
+            try:
+                self.update()
+            finally:
+                self.graph = wp.capture_end(self.device)
 
     def load_mesh(self, filename, path):
         asset_stage = Usd.Stage.Open(filename)
diff --git a/examples/example_sim_trajopt.py b/examples/example_sim_trajopt.py
index 279fb5f32..d5f832a37 100644
--- a/examples/example_sim_trajopt.py
+++ b/examples/example_sim_trajopt.py
@@ -16,7 +16,6 @@
 
 import os
 
-import matplotlib.pyplot as plt
 import numpy as np
 
 import warp as wp
@@ -194,6 +193,8 @@ def render(self):
 
 
 if __name__ == "__main__":
+    import matplotlib.pyplot as plt
+
     stage_path = os.path.join(os.path.dirname(__file__), "outputs/example_sim_trajopt.usd")
     example = Example(stage_path, device=wp.get_preferred_device(), verbose=True)
 
diff --git a/examples/fem/example_diffusion_mgpu.py b/examples/fem/example_diffusion_mgpu.py
index 32b6b9fad..5b672d433 100644
--- a/examples/fem/example_diffusion_mgpu.py
+++ b/examples/fem/example_diffusion_mgpu.py
@@ -6,19 +6,18 @@
 
 import warp as wp
 import warp.fem as fem
-
 from warp.sparse import bsr_axpy, bsr_mv
 from warp.utils import array_cast
 
 # Import example utilities
 # Make sure that works both when imported as module and run as standalone file
 try:
-    from .example_diffusion import linear_form, diffusion_form
     from .bsr_utils import bsr_cg
+    from .example_diffusion import diffusion_form, linear_form
     from .plot_utils import Plot
 except ImportError:
-    from example_diffusion import linear_form, diffusion_form
     from bsr_utils import bsr_cg
+    from example_diffusion import diffusion_form, linear_form
     from plot_utils import Plot
 
 
@@ -100,14 +99,18 @@ def __init__(self, stage=None, quiet=False):
         self._quiet = quiet
 
         self._geo = fem.Grid2D(res=wp.vec2i(25))
-        self._scalar_space = fem.make_polynomial_space(self._geo, degree=3)
-        self._scalar_field = self._scalar_space.make_field()
+
+        self._main_device = wp.get_device("cuda")
+
+        with wp.ScopedDevice(self._main_device):
+            self._scalar_space = fem.make_polynomial_space(self._geo, degree=3)
+            self._scalar_field = self._scalar_space.make_field()
 
         self.renderer = Plot(stage)
 
     def update(self):
         devices = wp.get_cuda_devices()
-        main_device = devices[0]
+        main_device = self._main_device
 
         rhs_vecs = []
         res_vecs = []
@@ -141,16 +144,15 @@ def update(self):
         A.tmp_buf = tmp
         A.rank_data = (matrices, rhs_vecs, res_vecs, indices)
 
-        with wp.ScopedDevice(main_device):
-            bsr_cg(
-                A,
-                x=global_res,
-                b=glob_rhs,
-                use_diag_precond=False,
-                quiet=self._quiet,
-                mv_routine=mv_routine,
-                device=main_device,
-            )
+        bsr_cg(
+            A,
+            x=global_res,
+            b=glob_rhs,
+            use_diag_precond=False,
+            quiet=self._quiet,
+            mv_routine=mv_routine,
+            device=main_device,
+        )
 
         array_cast(in_array=global_res, out_array=self._scalar_field.dof_values)
 
diff --git a/exts/omni.warp.core/omni/warp/core/tests/test_ext.py b/exts/omni.warp.core/omni/warp/core/tests/test_ext.py
index 189d62c4f..46fb22efb 100644
--- a/exts/omni.warp.core/omni/warp/core/tests/test_ext.py
+++ b/exts/omni.warp.core/omni/warp/core/tests/test_ext.py
@@ -5,117 +5,73 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-"""Tests for the Warp library itself.
+"""Tests for the Warp core library in Kit.
 
 Only a trimmed down list of tests is run since the full suite is too slow.
+
+More information about testing in Kit:
+    https://docs.omniverse.nvidia.com/kit/docs/kit-manual/latest/guide/testing_exts_python.html
 """
 
-import omni.kit
+import importlib
 
-import warp as wp
-import warp.tests.test_array
-import warp.tests.test_array_reduce
-import warp.tests.test_bvh
-import warp.tests.test_codegen
-import warp.tests.test_compile_consts
-import warp.tests.test_conditional
-import warp.tests.test_ctypes
-import warp.tests.test_devices
-import warp.tests.test_dlpack
-import warp.tests.test_fabricarray
-import warp.tests.test_func
-import warp.tests.test_generics
-import warp.tests.test_grad_customs
-import warp.tests.test_hash_grid
-import warp.tests.test_indexedarray
-import warp.tests.test_launch
-import warp.tests.test_marching_cubes
-import warp.tests.test_mat_lite
-import warp.tests.test_math
-import warp.tests.test_matmul_lite
-import warp.tests.test_mesh
-import warp.tests.test_mesh_query_aabb
-import warp.tests.test_mesh_query_point
-import warp.tests.test_mesh_query_ray
-import warp.tests.test_modules_lite
-import warp.tests.test_noise
-import warp.tests.test_operators
-import warp.tests.test_quat
-import warp.tests.test_rand
-import warp.tests.test_reload
-import warp.tests.test_rounding
-import warp.tests.test_runlength_encode
-import warp.tests.test_sparse
-import warp.tests.test_streams
-import warp.tests.test_tape
-import warp.tests.test_transient_module
-import warp.tests.test_types
-import warp.tests.test_utils
-import warp.tests.test_vec_lite
-import warp.tests.test_volume
-import warp.tests.test_volume_write
+import omni.kit.test
 
-initialized = False
+TEST_DESCS = (
+    ("test_array", "TestArray"),
+    ("test_array_reduce", "TestArrayReduce"),
+    ("test_bvh", "TestBvh"),
+    ("test_codegen", "TestCodeGen"),
+    ("test_compile_consts", "TestConstants"),
+    ("test_conditional", "TestConditional"),
+    ("test_ctypes", "TestCTypes"),
+    ("test_devices", "TestDevices"),
+    ("test_dlpack", "TestDLPack"),
+    ("test_fabricarray", "TestFabricArray"),
+    ("test_func", "TestFunc"),
+    ("test_generics", "TestGenerics"),
+    ("test_grad_customs", "TestGradCustoms"),
+    ("test_hash_grid", "TestHashGrid"),
+    ("test_indexedarray", "TestIndexedArray"),
+    ("test_launch", "TestLaunch"),
+    ("test_marching_cubes", "TestMarchingCubes"),
+    ("test_mat_lite", "TestMatLite"),
+    ("test_math", "TestMath"),
+    ("test_matmul_lite", "TestMatmulLite"),
+    ("test_mesh", "TestMesh"),
+    ("test_mesh_query_aabb", "TestMeshQueryAABBMethods"),
+    ("test_mesh_query_point", "TestMeshQueryPoint"),
+    ("test_mesh_query_ray", "TestMeshQueryRay"),
+    ("test_modules_lite", "TestModuleLite"),
+    ("test_noise", "TestNoise"),
+    ("test_operators", "TestOperators"),
+    ("test_quat", "TestQuat"),
+    ("test_rand", "TestRand"),
+    ("test_reload", "TestReload"),
+    ("test_rounding", "TestRounding"),
+    ("test_runlength_encode", "TestRunlengthEncode"),
+    ("test_sparse", "TestSparse"),
+    ("test_streams", "TestStreams"),
+    ("test_tape", "TestTape"),
+    ("test_transient_module", "TestTransientModule"),
+    ("test_types", "TestTypes"),
+    ("test_utils", "TestUtils"),
+    ("test_vec_lite", "TestVecLite"),
+    ("test_volume", "TestVolume"),
+    ("test_volume_write", "TestVolumeWrite"),
+)
 
 
-class BaseTestCase(omni.kit.test.AsyncTestCase):
-    @classmethod
-    def setUpClass(cls):
-        global initialized
+test_clss = []
+for module_name, cls_name in TEST_DESCS:
+    module = importlib.import_module(f"warp.tests.{module_name}")
+    cls = getattr(module, cls_name)
 
-        if not initialized:
-            # Load all the Warp modules just once. This needs to be done
-            # within the `setUpClass` method instead of at the module level
-            # to avoid the reload to be done as soon as this module is imported,
-            # which might happen when scanning for existing tests without
-            # actually needing to run these ones.
-            wp.force_load()
-            initialized = True
+    # Change the base class from unittest.TestCase
+    cls.__bases__ = (omni.kit.test.AsyncTestCase,)
 
+    test_clss.append(cls)
 
-test_clss = (
-    warp.tests.test_array.register(BaseTestCase),
-    warp.tests.test_array_reduce.register(BaseTestCase),
-    warp.tests.test_bvh.register(BaseTestCase),
-    warp.tests.test_codegen.register(BaseTestCase),
-    warp.tests.test_compile_consts.register(BaseTestCase),
-    warp.tests.test_conditional.register(BaseTestCase),
-    warp.tests.test_ctypes.register(BaseTestCase),
-    warp.tests.test_devices.register(BaseTestCase),
-    warp.tests.test_dlpack.register(BaseTestCase),
-    warp.tests.test_fabricarray.register(BaseTestCase),
-    warp.tests.test_func.register(BaseTestCase),
-    warp.tests.test_generics.register(BaseTestCase),
-    warp.tests.test_grad_customs.register(BaseTestCase),
-    warp.tests.test_hash_grid.register(BaseTestCase),
-    warp.tests.test_indexedarray.register(BaseTestCase),
-    warp.tests.test_launch.register(BaseTestCase),
-    warp.tests.test_marching_cubes.register(BaseTestCase),
-    warp.tests.test_mat_lite.register(BaseTestCase),
-    warp.tests.test_math.register(BaseTestCase),
-    warp.tests.test_matmul_lite.register(BaseTestCase),
-    warp.tests.test_mesh.register(BaseTestCase),
-    warp.tests.test_mesh_query_aabb.register(BaseTestCase),
-    warp.tests.test_mesh_query_point.register(BaseTestCase),
-    warp.tests.test_mesh_query_ray.register(BaseTestCase),
-    warp.tests.test_modules_lite.register(BaseTestCase),
-    warp.tests.test_noise.register(BaseTestCase),
-    warp.tests.test_operators.register(BaseTestCase),
-    warp.tests.test_quat.register(BaseTestCase),
-    warp.tests.test_rand.register(BaseTestCase),
-    warp.tests.test_reload.register(BaseTestCase),
-    warp.tests.test_rounding.register(BaseTestCase),
-    warp.tests.test_runlength_encode.register(BaseTestCase),
-    warp.tests.test_sparse.register(BaseTestCase),
-    warp.tests.test_streams.register(BaseTestCase),
-    warp.tests.test_tape.register(BaseTestCase),
-    warp.tests.test_transient_module.register(BaseTestCase),
-    warp.tests.test_types.register(BaseTestCase),
-    warp.tests.test_utils.register(BaseTestCase),
-    warp.tests.test_vec_lite.register(BaseTestCase),
-    warp.tests.test_volume_write.register(BaseTestCase),
-    warp.tests.test_volume.register(BaseTestCase)
-)
 
 # Each test class needs to be defined at the module level to be found by
 # the test runners.
diff --git a/pyproject.toml b/pyproject.toml
index 6190136f0..ae3994153 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -30,7 +30,7 @@ Documentation = "https://nvidia.github.io/warp"
 Changelog = "https://github.com/NVIDIA/warp/blob/main/CHANGELOG.md"
 
 [project.optional-dependencies]
-dev = ["flake8", "black", "isort", "nvtx", "furo", "sphinx-copybutton"]
+dev = ["flake8", "black", "isort", "nvtx", "furo", "sphinx-copybutton", "coverage[toml]"]
 
 [tool.setuptools.packages]
 find = {}
@@ -45,3 +45,11 @@ line-length = 120
 [tool.isort]
 profile = "black"
 skip_gitignore = true
+
+[tool.coverage.run]
+source = ["warp", "warp.sim", "warp.render"]
+disable_warnings = ["module-not-measured", "module-not-imported", "no-data-collected", "couldnt-parse"]
+
+[tool.coverage.report]
+exclude_lines = ["pragma: no cover", "@wp", "@warp", "if 0:", "if __name__ == .__main__.:"]
+omit =["*/warp/tests/*", "*/warp/fem/*", "appdirs.py", "render_opengl.py", "build_dll.py", "config.py", "stubs.py"]
diff --git a/tools/ci/testing/test-linux-aarch64/test.sh b/tools/ci/testing/test-linux-aarch64/test.sh
index 44da31bd3..c87575574 100755
--- a/tools/ci/testing/test-linux-aarch64/test.sh
+++ b/tools/ci/testing/test-linux-aarch64/test.sh
@@ -16,4 +16,4 @@ echo "Installing Warp to Python"
 $PYTHON -m pip install -e "$SCRIPT_DIR/../../../../."
 
 echo "Running tests"
-$PYTHON "$SCRIPT_DIR/../../../../warp/tests/test_all.py"
+$PYTHON -m warp.tests
diff --git a/tools/ci/testing/test-linux-x86_64/test.sh b/tools/ci/testing/test-linux-x86_64/test.sh
index c6c9755f8..4ac6c9314 100755
--- a/tools/ci/testing/test-linux-x86_64/test.sh
+++ b/tools/ci/testing/test-linux-x86_64/test.sh
@@ -5,15 +5,32 @@ SCRIPT_DIR=$(dirname ${BASH_SOURCE})
 "$SCRIPT_DIR/../../../../repo.sh" build --fetch-only $@
 
 PYTHON="$SCRIPT_DIR/../../../../_build/target-deps/python/python"
+CUDA_BIN="$SCRIPT_DIR/../../../../_build/target-deps/cuda/bin"
+
+# Make sure ptxas can be run by JAX
+export PATH="$CUDA_BIN:$PATH"
 
 echo "Installing test dependencies"
+
+if [ -n "$TEAMCITY_VERSION" ]; then
+    echo "##teamcity[blockOpened name='Installing test dependencies']"
+fi
+
+$PYTHON -m pip install --upgrade pip
 $PYTHON -m pip install matplotlib
 $PYTHON -m pip install usd-core
-$PYTHON -m pip install torch --extra-index-url https://download.pytorch.org/whl/cu113
-$PYTHON -m pip install "jax[cuda]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+$PYTHON -m pip install coverage
+if [[ "$OSTYPE" == "linux-gnu" ]]; then
+    $PYTHON -m pip install torch --extra-index-url https://download.pytorch.org/whl/cu115
+    $PYTHON -m pip install --upgrade "jax[cuda]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
+fi
+
+if [ -n "$TEAMCITY_VERSION" ]; then
+    echo "##teamcity[blockClosed name='Installing test dependencies']"
+fi
 
 echo "Installing Warp to Python"
 $PYTHON -m pip install -e "$SCRIPT_DIR/../../../../."
 
 echo "Running tests"
-$PYTHON "$SCRIPT_DIR/../../../../warp/tests/test_all.py"
+$PYTHON -m warp.tests
diff --git a/tools/ci/testing/test-windows-x86_64/test.bat b/tools/ci/testing/test-windows-x86_64/test.bat
index ba575a82f..6470f1d2a 100644
--- a/tools/ci/testing/test-windows-x86_64/test.bat
+++ b/tools/ci/testing/test-windows-x86_64/test.bat
@@ -6,7 +6,8 @@ SET PYTHON="%~dp0..\..\..\..\_build\target-deps\python\python.exe"
 echo "Installing test dependencies"
 call %PYTHON% -m pip install matplotlib
 call %PYTHON% -m pip install usd-core
-call %PYTHON% -m pip install torch --extra-index-url https://download.pytorch.org/whl/cu113
+call %PYTHON% -m pip install coverage
+call %PYTHON% -m pip install torch --extra-index-url https://download.pytorch.org/whl/cu115
 :: According to Jax docs, the pip packages don't work on Windows and may fail silently
 ::call %PYTHON% -m pip install "jax[cuda]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
 
@@ -14,4 +15,4 @@ echo "Installing Warp to Python"
 call %PYTHON% -m pip install -e "%~dp0..\..\..\..\."
 
 echo "Running tests"
-call %PYTHON% "%~dp0..\..\..\..\warp\tests\test_all.py"
+call %PYTHON% -m warp.tests
diff --git a/warp/config.py b/warp/config.py
index 73bad3dd2..d231b0825 100644
--- a/warp/config.py
+++ b/warp/config.py
@@ -33,3 +33,5 @@
 enable_backward = True  # whether to compiler the backward passes of the kernels
 
 llvm_cuda = False  # use Clang/LLVM instead of NVRTC to compile CUDA
+
+graph_capture_module_load_default = True  # Default value of force_module_load for capture_begin()
diff --git a/warp/context.py b/warp/context.py
index d196b20f3..6c90780b1 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -3778,7 +3778,7 @@ def get_module_options(module: Optional[Any] = None) -> Dict[str, Any]:
     return get_module(m.__name__).options
 
 
-def capture_begin(device: Devicelike = None, stream=None, force_module_load=True):
+def capture_begin(device: Devicelike = None, stream=None, force_module_load=None):
     """Begin capture of a CUDA graph
 
     Captures all subsequent kernel launches and memory operations on CUDA devices.
@@ -3792,6 +3792,9 @@ def capture_begin(device: Devicelike = None, stream=None, force_module_load=True
 
     """
 
+    if force_module_load is None:
+        force_module_load = warp.config.graph_capture_module_load_default
+
     if warp.config.verify_cuda is True:
         raise RuntimeError("Cannot use CUDA error verification during graph capture")
 
diff --git a/warp/tests/__main__.py b/warp/tests/__main__.py
index e331aff81..98f9af5ca 100644
--- a/warp/tests/__main__.py
+++ b/warp/tests/__main__.py
@@ -1,7 +1,4 @@
-from .test_all import run
+from warp.thirdparty.unittest_parallel import main
 
-ret = run()
-
-import sys
-
-sys.exit(ret)
+if __name__ == "__main__":
+    main()
diff --git a/warp/tests/test_class_kernel.py b/warp/tests/aux_test_class_kernel.py
similarity index 54%
rename from warp/tests/test_class_kernel.py
rename to warp/tests/aux_test_class_kernel.py
index 7b251a0e5..0b258fef3 100644
--- a/warp/tests/test_class_kernel.py
+++ b/warp/tests/aux_test_class_kernel.py
@@ -1,7 +1,15 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+"""Dummy class used in test_reload.py"""
+
 import warp as wp
 
 
-# dummy class used in test_reload.py
 class ClassKernelTest:
     def __init__(self, device):
         # 3x3 frames in the rest pose:
diff --git a/warp/tests/test_compile_consts_dummy.py b/warp/tests/aux_test_compile_consts_dummy.py
similarity index 100%
rename from warp/tests/test_compile_consts_dummy.py
rename to warp/tests/aux_test_compile_consts_dummy.py
diff --git a/warp/tests/aux_test_conditional_unequal_types_kernels.py b/warp/tests/aux_test_conditional_unequal_types_kernels.py
new file mode 100644
index 000000000..2a843e389
--- /dev/null
+++ b/warp/tests/aux_test_conditional_unequal_types_kernels.py
@@ -0,0 +1,21 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+"""This file defines a kernel that fails on codegen.py"""
+
+import warp as wp
+
+
+@wp.kernel
+def unequal_types_kernel():
+    x = wp.int32(10)
+    y = 10
+    z = True
+
+    # Throws a TypeError
+    if x == y == z:
+        pass
diff --git a/warp/tests/test_dependent.py b/warp/tests/aux_test_dependent.py
similarity index 85%
rename from warp/tests/test_dependent.py
rename to warp/tests/aux_test_dependent.py
index 2df0f05b4..ece37e95e 100644
--- a/warp/tests/test_dependent.py
+++ b/warp/tests/aux_test_dependent.py
@@ -5,10 +5,10 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# This file is used to test reloading module references.
+"""This file is used to test reloading module references."""
 
 import warp as wp
-import warp.tests.test_reference as ref
+import warp.tests.aux_test_reference as ref
 
 wp.init()
 
diff --git a/warp/tests/test_reference.py b/warp/tests/aux_test_reference.py
similarity index 72%
rename from warp/tests/test_reference.py
rename to warp/tests/aux_test_reference.py
index 375183272..dbf0f358d 100644
--- a/warp/tests/test_reference.py
+++ b/warp/tests/aux_test_reference.py
@@ -1,7 +1,7 @@
 # This file is used to test reloading module references.
 
 import warp as wp
-import warp.tests.test_reference_reference as refref
+import warp.tests.aux_test_reference_reference as refref
 
 wp.init()
 
diff --git a/warp/tests/test_reference_reference.py b/warp/tests/aux_test_reference_reference.py
similarity index 100%
rename from warp/tests/test_reference_reference.py
rename to warp/tests/aux_test_reference_reference.py
diff --git a/warp/tests/test_square.py b/warp/tests/aux_test_square.py
similarity index 100%
rename from warp/tests/test_square.py
rename to warp/tests/aux_test_square.py
diff --git a/warp/tests/aux_test_unresolved_func.py b/warp/tests/aux_test_unresolved_func.py
new file mode 100644
index 000000000..1b9e433f8
--- /dev/null
+++ b/warp/tests/aux_test_unresolved_func.py
@@ -0,0 +1,14 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import warp as wp
+
+
+@wp.kernel
+def unresolved_func_kernel():
+    # this should trigger an exception due to unresolved function
+    x = wp.missing_func(42)
diff --git a/warp/tests/aux_test_unresolved_symbol.py b/warp/tests/aux_test_unresolved_symbol.py
new file mode 100644
index 000000000..a81ce4bf5
--- /dev/null
+++ b/warp/tests/aux_test_unresolved_symbol.py
@@ -0,0 +1,14 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import warp as wp
+
+
+@wp.kernel
+def unresolved_symbol_kernel():
+    # this should trigger an exception due to unresolved symbol
+    x = missing_symbol
diff --git a/warp/tests/test_kinematics.py b/warp/tests/disabled_kinematics.py
similarity index 94%
rename from warp/tests/test_kinematics.py
rename to warp/tests/disabled_kinematics.py
index 25894cf51..524395398 100644
--- a/warp/tests/test_kinematics.py
+++ b/warp/tests/disabled_kinematics.py
@@ -5,12 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
+import math
 import unittest
 
 import warp as wp
 import warp.sim
-import math
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -221,21 +221,19 @@ def test_fk_ik_complex_joint_mechanism(test, device):
     check_fk_ik(builder, device)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestGrad(parent):
-        pass
+class TestKinematics(unittest.TestCase):
+    pass
 
-    add_function_test(TestGrad, "test_fk_ik_ant", test_fk_ik_ant, devices=devices)
-    add_function_test(
-        TestGrad, "test_fk_ik_complex_joint_mechanism", test_fk_ik_complex_joint_mechanism, devices=devices
-    )
 
-    return TestGrad
+add_function_test(TestKinematics, "test_fk_ik_ant", test_fk_ik_ant, devices=devices)
+add_function_test(
+    TestKinematics, "test_fk_ik_complex_joint_mechanism", test_fk_ik_complex_joint_mechanism, devices=devices
+)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/run_coverage_serial.py b/warp/tests/run_coverage_serial.py
new file mode 100644
index 000000000..aa5f600ac
--- /dev/null
+++ b/warp/tests/run_coverage_serial.py
@@ -0,0 +1,31 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+"""Serial code-coverage runner
+
+This script is used to generate code-coverage reports by running Warp tests.
+It runs in serial so can take over an hour to finish. To generate a coverage
+report in parallel, use the warp/thirdparty./unittest_parallel.py script
+instead with the --coverage option, e.g. python -m warp.tests --coverage
+"""
+
+import coverage
+
+cover = coverage.Coverage(config_file=True, messages=True)
+
+cover.start()
+
+with cover.collect():
+    import unittest_serial  # noqa: E402
+
+    unittest_serial.run_specified()
+
+cover.save()
+
+cover.report()
+
+cover.html_report(title="Warp Testing Code Coverage Report")
diff --git a/warp/tests/test_adam.py b/warp/tests/test_adam.py
index 0eb099a47..c104df306 100644
--- a/warp/tests/test_adam.py
+++ b/warp/tests/test_adam.py
@@ -5,16 +5,14 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
+import unittest
+
 import numpy as np
-import math
 
 import warp as wp
-from warp.tests.test_base import *
-import unittest
-
 import warp.optim
 import warp.sim
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -142,20 +140,18 @@ def gradient_func():
             test.assertLessEqual(v, tol)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestArray(parent):
-        pass
+class TestAdam(unittest.TestCase):
+    pass
 
-    add_function_test(TestArray, "test_adam_solve_float", test_adam_solve_float, devices=devices)
-    add_function_test(TestArray, "test_adam_solve_vec3", test_adam_solve_vec3, devices=devices)
-    add_function_test(TestArray, "test_adam_solve_two_inputs", test_adam_solve_two_inputs, devices=devices)
 
-    return TestArray
+add_function_test(TestAdam, "test_adam_solve_float", test_adam_solve_float, devices=devices)
+add_function_test(TestAdam, "test_adam_solve_vec3", test_adam_solve_vec3, devices=devices)
+add_function_test(TestAdam, "test_adam_solve_two_inputs", test_adam_solve_two_inputs, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_all.py b/warp/tests/test_all.py
deleted file mode 100644
index 812aef902..000000000
--- a/warp/tests/test_all.py
+++ /dev/null
@@ -1,241 +0,0 @@
-# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
-# NVIDIA CORPORATION and its licensors retain all intellectual property
-# and proprietary rights in and to this software, related documentation
-# and any modifications thereto.  Any use, reproduction, disclosure or
-# distribution of this software and related documentation without an express
-# license agreement from NVIDIA CORPORATION is strictly prohibited.
-
-import os
-import unittest
-
-import warp as wp
-import warp.tests.test_adam
-import warp.tests.test_arithmetic
-import warp.tests.test_array
-import warp.tests.test_atomic
-import warp.tests.test_bool
-import warp.tests.test_builtins_resolution
-import warp.tests.test_bvh
-import warp.tests.test_closest_point_edge_edge
-import warp.tests.test_codegen
-import warp.tests.test_compile_consts
-import warp.tests.test_conditional
-import warp.tests.test_copy
-import warp.tests.test_ctypes
-import warp.tests.test_devices
-import warp.tests.test_dlpack
-import warp.tests.test_examples
-import warp.tests.test_fabricarray
-import warp.tests.test_fast_math
-import warp.tests.test_fem
-import warp.tests.test_fp16
-import warp.tests.test_func
-import warp.tests.test_generics
-import warp.tests.test_grad
-import warp.tests.test_grad_customs
-import warp.tests.test_hash_grid
-import warp.tests.test_import
-import warp.tests.test_indexedarray
-import warp.tests.test_intersect
-import warp.tests.test_large
-import warp.tests.test_launch
-import warp.tests.test_lerp
-import warp.tests.test_lvalue
-import warp.tests.test_mat
-import warp.tests.test_math
-import warp.tests.test_matmul
-import warp.tests.test_mesh
-import warp.tests.test_mesh_query_aabb
-import warp.tests.test_mesh_query_point
-import warp.tests.test_mesh_query_ray
-import warp.tests.test_mlp
-import warp.tests.test_model
-import warp.tests.test_modules_lite
-import warp.tests.test_multigpu
-import warp.tests.test_noise
-import warp.tests.test_operators
-import warp.tests.test_options
-import warp.tests.test_pinned
-import warp.tests.test_print
-import warp.tests.test_quat
-import warp.tests.test_rand
-import warp.tests.test_reload
-import warp.tests.test_rounding
-import warp.tests.test_smoothstep
-import warp.tests.test_snippet
-import warp.tests.test_sparse
-import warp.tests.test_spatial
-import warp.tests.test_streams
-import warp.tests.test_struct
-import warp.tests.test_tape
-import warp.tests.test_torch
-import warp.tests.test_transient_module
-import warp.tests.test_vec
-import warp.tests.test_volume
-from warp.tests.test_base import get_test_devices
-
-# Uncomment to run the tests on all devices
-# import warp.tests.test_base
-# warp.tests.test_base.test_mode = "all"
-
-
-def register_tests(parent):
-    tests = []
-
-    tests.append(warp.tests.test_codegen.register(parent))
-    tests.append(warp.tests.test_mesh_query_aabb.register(parent))
-    tests.append(warp.tests.test_mesh_query_point.register(parent))
-    tests.append(warp.tests.test_mesh_query_ray.register(parent))
-    tests.append(warp.tests.test_builtins_resolution.register(parent))
-    tests.append(warp.tests.test_bvh.register(parent))
-    tests.append(warp.tests.test_conditional.register(parent))
-    tests.append(warp.tests.test_operators.register(parent))
-    tests.append(warp.tests.test_rounding.register(parent))
-    tests.append(warp.tests.test_hash_grid.register(parent))
-    tests.append(warp.tests.test_ctypes.register(parent))
-    tests.append(warp.tests.test_rand.register(parent))
-    tests.append(warp.tests.test_noise.register(parent))
-    tests.append(warp.tests.test_tape.register(parent))
-    tests.append(warp.tests.test_compile_consts.register(parent))
-    tests.append(warp.tests.test_volume.register(parent))
-    tests.append(warp.tests.test_mlp.register(parent))
-    tests.append(warp.tests.test_grad.register(parent))
-    tests.append(warp.tests.test_grad_customs.register(parent))
-    tests.append(warp.tests.test_intersect.register(parent))
-    tests.append(warp.tests.test_array.register(parent))
-    tests.append(warp.tests.test_launch.register(parent))
-    tests.append(warp.tests.test_large.register(parent))
-    tests.append(warp.tests.test_import.register(parent))
-    tests.append(warp.tests.test_func.register(parent))
-    tests.append(warp.tests.test_fp16.register(parent))
-    tests.append(warp.tests.test_reload.register(parent))
-    tests.append(warp.tests.test_struct.register(parent))
-    tests.append(warp.tests.test_closest_point_edge_edge.register(parent))
-    tests.append(warp.tests.test_multigpu.register(parent))
-    tests.append(warp.tests.test_quat.register(parent))
-    tests.append(warp.tests.test_atomic.register(parent))
-    tests.append(warp.tests.test_adam.register(parent))
-    tests.append(warp.tests.test_transient_module.register(parent))
-    tests.append(warp.tests.test_lerp.register(parent))
-    tests.append(warp.tests.test_smoothstep.register(parent))
-    tests.append(warp.tests.test_model.register(parent))
-    tests.append(warp.tests.test_fast_math.register(parent))
-    tests.append(warp.tests.test_streams.register(parent))
-    tests.append(warp.tests.test_torch.register(parent))
-    tests.append(warp.tests.test_pinned.register(parent))
-    tests.append(warp.tests.test_print.register(parent))
-    tests.append(warp.tests.test_matmul.register(parent))
-    tests.append(warp.tests.test_options.register(parent))
-    tests.append(warp.tests.test_dlpack.register(parent))
-    tests.append(warp.tests.test_vec.register(parent))
-    tests.append(warp.tests.test_mat.register(parent))
-    tests.append(warp.tests.test_arithmetic.register(parent))
-    tests.append(warp.tests.test_spatial.register(parent))
-    tests.append(warp.tests.test_sparse.register(parent))
-    tests.append(warp.tests.test_math.register(parent))
-    tests.append(warp.tests.test_generics.register(parent))
-    tests.append(warp.tests.test_indexedarray.register(parent))
-    tests.append(warp.tests.test_copy.register(parent))
-    tests.append(warp.tests.test_mesh.register(parent))
-    tests.append(warp.tests.test_fabricarray.register(parent))
-    tests.append(warp.tests.test_bool.register(parent))
-    tests.append(warp.tests.test_examples.register(parent))  # Needs to be before test_fem for now
-    tests.append(warp.tests.test_fem.register(parent))
-    tests.append(warp.tests.test_lvalue.register(parent))
-    tests.append(warp.tests.test_devices.register(parent))
-    tests.append(warp.tests.test_modules_lite.register(parent))
-    tests.append(warp.tests.test_snippet.register(parent))
-
-    return tests
-
-
-class TeamCityTestResult(unittest.TextTestResult):
-    """This class will report each test result to TeamCity"""
-
-    def __init__(self, stream, descriptions, verbosity):
-        super(TeamCityTestResult, self).__init__(stream, descriptions, verbosity)
-
-    def addSuccess(self, test):
-        super(TeamCityTestResult, self).addSuccess(test)
-        self.reportSuccess(test)
-
-    def addError(self, test, err):
-        super(TeamCityTestResult, self).addError(test, err)
-        self.reportFailure(test)
-
-    def addFailure(self, test, err):
-        super(TeamCityTestResult, self).addFailure(test, err)
-        self.reportFailure(test)
-
-    def addSkip(self, test, reason):
-        super(TeamCityTestResult, self).addSkip(test, reason)
-
-    def addExpectedFailure(self, test, err):
-        super(TeamCityTestResult, self).addExpectedFailure(test, err)
-        self.reportSuccess(test)
-
-    def addUnexpectedSuccess(self, test):
-        super(TeamCityTestResult, self).addUnexpectedSuccess(test)
-        self.reportFailure(test)
-
-    def reportSuccess(self, test):
-        test_id = test.id()
-        print(f"##teamcity[testStarted name='{test_id}']")
-        print(f"##teamcity[testFinished name='{test_id}']")
-
-    def reportFailure(self, test):
-        test_id = test.id()
-        print(f"##teamcity[testStarted name='{test_id}']")
-        print(f"##teamcity[testFailed name='{test_id}']")
-        print(f"##teamcity[testFinished name='{test_id}']")
-
-
-class TeamCityTestRunner(unittest.TextTestRunner):
-    """Test runner that will report test results to TeamCity if running in TeamCity"""
-
-    def __init__(self, **kwargs):
-        self.running_in_teamcity = os.environ.get("TEAMCITY_VERSION") is not None
-        if self.running_in_teamcity:
-            kwargs["resultclass"] = TeamCityTestResult
-        super(TeamCityTestRunner, self).__init__(**kwargs)
-
-    def run(self, test, name):
-        if self.running_in_teamcity:
-            print(f"##teamcity[testSuiteStarted name='{name}']")
-
-        result = super(TeamCityTestRunner, self).run(test)
-
-        if self.running_in_teamcity:
-            print(f"##teamcity[testSuiteFinished name='{name}']")
-            if not result.wasSuccessful():
-                print("##teamcity[buildStatus status='FAILURE']")
-
-        return result
-
-
-def run():
-    test_suite = unittest.TestSuite()
-
-    tests = register_tests(unittest.TestCase)
-
-    for test in tests:
-        test_suite.addTest(unittest.defaultTestLoader.loadTestsFromTestCase(test))
-
-    # force rebuild of all kernels
-    wp.build.clear_kernel_cache()
-
-    # load all modules
-    for device in get_test_devices():
-        wp.force_load(device)
-
-    runner = TeamCityTestRunner(verbosity=2, failfast=False)
-    ret = not runner.run(test_suite, "WarpTests").wasSuccessful()
-    return ret
-
-
-if __name__ == "__main__":
-    ret = run()
-
-    import sys
-
-    sys.exit(ret)
diff --git a/warp/tests/test_arithmetic.py b/warp/tests/test_arithmetic.py
index e9b24a326..790219853 100644
--- a/warp/tests/test_arithmetic.py
+++ b/warp/tests/test_arithmetic.py
@@ -8,8 +8,9 @@
 import unittest
 
 import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -1040,52 +1041,50 @@ def check_clamp(
             tape.zero()
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestArithmetic(parent):
-        pass
 
-    # these unary ops only make sense for signed values:
-    for dtype in np_signed_int_types + np_float_types:
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_unary_ops_{dtype.__name__}", test_unary_ops, devices=devices, dtype=dtype
-        )
+class TestArithmetic(unittest.TestCase):
+    pass
 
-    for dtype in np_float_types:
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_special_funcs_{dtype.__name__}", test_special_funcs, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestArithmetic,
-            f"test_special_funcs_2arg_{dtype.__name__}",
-            test_special_funcs_2arg,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_interp_{dtype.__name__}", test_interp, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_float_to_int_{dtype.__name__}", test_float_to_int, devices=devices, dtype=dtype
-        )
 
-    for dtype in np_scalar_types:
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_clamp_{dtype.__name__}", test_clamp, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_nonzero_{dtype.__name__}", test_nonzero, devices=devices, dtype=dtype
-        )
-        add_function_test(TestArithmetic, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
-        add_function_test_register_kernel(
-            TestArithmetic, f"test_binary_ops_{dtype.__name__}", test_binary_ops, devices=devices, dtype=dtype
-        )
+# these unary ops only make sense for signed values:
+for dtype in np_signed_int_types + np_float_types:
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_unary_ops_{dtype.__name__}", test_unary_ops, devices=devices, dtype=dtype
+    )
 
-    return TestArithmetic
+for dtype in np_float_types:
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_special_funcs_{dtype.__name__}", test_special_funcs, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestArithmetic,
+        f"test_special_funcs_2arg_{dtype.__name__}",
+        test_special_funcs_2arg,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_interp_{dtype.__name__}", test_interp, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_float_to_int_{dtype.__name__}", test_float_to_int, devices=devices, dtype=dtype
+    )
+
+for dtype in np_scalar_types:
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_clamp_{dtype.__name__}", test_clamp, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_nonzero_{dtype.__name__}", test_nonzero, devices=devices, dtype=dtype
+    )
+    add_function_test(TestArithmetic, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
+    add_function_test_register_kernel(
+        TestArithmetic, f"test_binary_ops_{dtype.__name__}", test_binary_ops, devices=devices, dtype=dtype
+    )
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_array.py b/warp/tests/test_array.py
index dd923e324..9085ae282 100644
--- a/warp/tests/test_array.py
+++ b/warp/tests/test_array.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -2085,55 +2085,53 @@ def test_array_from_numpy(test, device):
     assert_np_equal(result.numpy(), expected.numpy())
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestArray(parent):
-        pass
-
-    add_function_test(TestArray, "test_shape", test_shape, devices=devices)
-    add_function_test(TestArray, "test_flatten", test_flatten, devices=devices)
-    add_function_test(TestArray, "test_reshape", test_reshape, devices=devices)
-    add_function_test(TestArray, "test_slicing", test_slicing, devices=devices)
-    add_function_test(TestArray, "test_transpose", test_transpose, devices=devices)
-    add_function_test(TestArray, "test_view", test_view, devices=devices)
-
-    add_function_test(TestArray, "test_1d_array", test_1d, devices=devices)
-    add_function_test(TestArray, "test_2d_array", test_2d, devices=devices)
-    add_function_test(TestArray, "test_3d_array", test_3d, devices=devices)
-    add_function_test(TestArray, "test_4d_array", test_4d, devices=devices)
-    add_function_test(TestArray, "test_4d_array_transposed", test_4d_transposed, devices=devices)
-
-    add_function_test(TestArray, "test_fill_scalar", test_fill_scalar, devices=devices)
-    add_function_test(TestArray, "test_fill_vector", test_fill_vector, devices=devices)
-    add_function_test(TestArray, "test_fill_matrix", test_fill_matrix, devices=devices)
-    add_function_test(TestArray, "test_fill_struct", test_fill_struct, devices=devices)
-    add_function_test(TestArray, "test_fill_slices", test_fill_slices, devices=devices)
-    add_function_test(TestArray, "test_full_scalar", test_full_scalar, devices=devices)
-    add_function_test(TestArray, "test_full_vector", test_full_vector, devices=devices)
-    add_function_test(TestArray, "test_full_matrix", test_full_matrix, devices=devices)
-    add_function_test(TestArray, "test_full_struct", test_full_struct, devices=devices)
-    add_function_test(TestArray, "test_empty_array", test_empty_array, devices=devices)
-    add_function_test(TestArray, "test_empty_from_numpy", test_empty_from_numpy, devices=devices)
-    add_function_test(TestArray, "test_empty_from_list", test_empty_from_list, devices=devices)
-    add_function_test(TestArray, "test_to_list_scalar", test_to_list_scalar, devices=devices)
-    add_function_test(TestArray, "test_to_list_vector", test_to_list_vector, devices=devices)
-    add_function_test(TestArray, "test_to_list_matrix", test_to_list_matrix, devices=devices)
-    add_function_test(TestArray, "test_to_list_struct", test_to_list_struct, devices=devices)
-
-    add_function_test(TestArray, "test_lower_bound", test_lower_bound, devices=devices)
-    add_function_test(TestArray, "test_round_trip", test_round_trip, devices=devices)
-    add_function_test(TestArray, "test_array_to_bool", test_array_to_bool, devices=devices)
-    add_function_test(TestArray, "test_array_of_structs", test_array_of_structs, devices=devices)
-    add_function_test(TestArray, "test_array_of_structs_grad", test_array_of_structs_grad, devices=devices)
-    add_function_test(TestArray, "test_array_of_structs_from_numpy", test_array_of_structs_from_numpy, devices=devices)
-    add_function_test(TestArray, "test_array_of_structs_roundtrip", test_array_of_structs_roundtrip, devices=devices)
-    add_function_test(TestArray, "test_array_from_numpy", test_array_from_numpy, devices=devices)
-
-    return TestArray
+devices = get_test_devices()
+
+
+class TestArray(unittest.TestCase):
+    pass
+
+
+add_function_test(TestArray, "test_shape", test_shape, devices=devices)
+add_function_test(TestArray, "test_flatten", test_flatten, devices=devices)
+add_function_test(TestArray, "test_reshape", test_reshape, devices=devices)
+add_function_test(TestArray, "test_slicing", test_slicing, devices=devices)
+add_function_test(TestArray, "test_transpose", test_transpose, devices=devices)
+add_function_test(TestArray, "test_view", test_view, devices=devices)
+
+add_function_test(TestArray, "test_1d_array", test_1d, devices=devices)
+add_function_test(TestArray, "test_2d_array", test_2d, devices=devices)
+add_function_test(TestArray, "test_3d_array", test_3d, devices=devices)
+add_function_test(TestArray, "test_4d_array", test_4d, devices=devices)
+add_function_test(TestArray, "test_4d_array_transposed", test_4d_transposed, devices=devices)
+
+add_function_test(TestArray, "test_fill_scalar", test_fill_scalar, devices=devices)
+add_function_test(TestArray, "test_fill_vector", test_fill_vector, devices=devices)
+add_function_test(TestArray, "test_fill_matrix", test_fill_matrix, devices=devices)
+add_function_test(TestArray, "test_fill_struct", test_fill_struct, devices=devices)
+add_function_test(TestArray, "test_fill_slices", test_fill_slices, devices=devices)
+add_function_test(TestArray, "test_full_scalar", test_full_scalar, devices=devices)
+add_function_test(TestArray, "test_full_vector", test_full_vector, devices=devices)
+add_function_test(TestArray, "test_full_matrix", test_full_matrix, devices=devices)
+add_function_test(TestArray, "test_full_struct", test_full_struct, devices=devices)
+add_function_test(TestArray, "test_empty_array", test_empty_array, devices=devices)
+add_function_test(TestArray, "test_empty_from_numpy", test_empty_from_numpy, devices=devices)
+add_function_test(TestArray, "test_empty_from_list", test_empty_from_list, devices=devices)
+add_function_test(TestArray, "test_to_list_scalar", test_to_list_scalar, devices=devices)
+add_function_test(TestArray, "test_to_list_vector", test_to_list_vector, devices=devices)
+add_function_test(TestArray, "test_to_list_matrix", test_to_list_matrix, devices=devices)
+add_function_test(TestArray, "test_to_list_struct", test_to_list_struct, devices=devices)
+
+add_function_test(TestArray, "test_lower_bound", test_lower_bound, devices=devices)
+add_function_test(TestArray, "test_round_trip", test_round_trip, devices=devices)
+add_function_test(TestArray, "test_array_to_bool", test_array_to_bool, devices=devices)
+add_function_test(TestArray, "test_array_of_structs", test_array_of_structs, devices=devices)
+add_function_test(TestArray, "test_array_of_structs_grad", test_array_of_structs_grad, devices=devices)
+add_function_test(TestArray, "test_array_of_structs_from_numpy", test_array_of_structs_from_numpy, devices=devices)
+add_function_test(TestArray, "test_array_of_structs_roundtrip", test_array_of_structs_roundtrip, devices=devices)
+add_function_test(TestArray, "test_array_from_numpy", test_array_from_numpy, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_array_reduce.py b/warp/tests/test_array_reduce.py
index 73f70dcc3..a299ddf80 100644
--- a/warp/tests/test_array_reduce.py
+++ b/warp/tests/test_array_reduce.py
@@ -1,9 +1,17 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
+
 import numpy as np
-import warp as wp
 
-from warp.utils import array_sum, array_inner
-from warp.tests.test_base import *
+import warp as wp
+from warp.tests.unittest_utils import *
+from warp.utils import array_inner, array_sum
 
 wp.init()
 
@@ -118,27 +126,25 @@ def test_array_inner_empty(test, device):
     assert_np_equal(array_inner(values, values, axis=0).numpy(), np.zeros(3))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestArraySym(parent):
-        pass
+class TestArrayReduce(unittest.TestCase):
+    pass
 
-    add_function_test(TestArraySym, "test_array_sum_double", make_test_array_sum(wp.float64), devices=devices)
-    add_function_test(TestArraySym, "test_array_sum_vec3", make_test_array_sum(wp.vec3), devices=devices)
-    add_function_test(TestArraySym, "test_array_sum_axis_float", make_test_array_sum_axis(wp.float32), devices=devices)
-    add_function_test(TestArraySym, "test_array_sum_empty", test_array_sum_empty, devices=devices)
-    add_function_test(TestArraySym, "test_array_inner_double", make_test_array_inner(wp.float64), devices=devices)
-    add_function_test(TestArraySym, "test_array_inner_vec3", make_test_array_inner(wp.vec3), devices=devices)
-    add_function_test(
-        TestArraySym, "test_array_inner_axis_float", make_test_array_inner_axis(wp.float32), devices=devices
-    )
-    add_function_test(TestArraySym, "test_array_inner_empty", test_array_inner_empty, devices=devices)
 
-    return TestArraySym
+add_function_test(TestArrayReduce, "test_array_sum_double", make_test_array_sum(wp.float64), devices=devices)
+add_function_test(TestArrayReduce, "test_array_sum_vec3", make_test_array_sum(wp.vec3), devices=devices)
+add_function_test(TestArrayReduce, "test_array_sum_axis_float", make_test_array_sum_axis(wp.float32), devices=devices)
+add_function_test(TestArrayReduce, "test_array_sum_empty", test_array_sum_empty, devices=devices)
+add_function_test(TestArrayReduce, "test_array_inner_double", make_test_array_inner(wp.float64), devices=devices)
+add_function_test(TestArrayReduce, "test_array_inner_vec3", make_test_array_inner(wp.vec3), devices=devices)
+add_function_test(
+    TestArrayReduce, "test_array_inner_axis_float", make_test_array_inner_axis(wp.float32), devices=devices
+)
+add_function_test(TestArrayReduce, "test_array_inner_empty", test_array_inner_empty, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_atomic.py b/warp/tests/test_atomic.py
index 5cba5f746..6a5faccdf 100644
--- a/warp/tests/test_atomic.py
+++ b/warp/tests/test_atomic.py
@@ -5,14 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
+import unittest
+
 import numpy as np
-import math
 
 import warp as wp
-from warp.tests.test_base import *
-
-import unittest
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -121,25 +119,23 @@ def test_atomic(test, device):
 test_atomic_mat44 = make_atomic_test(wp.mat44)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestAtomic(parent):
-        pass
+class TestAtomic(unittest.TestCase):
+    pass
 
-    add_function_test(TestAtomic, "test_atomic_int", test_atomic_int, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_float", test_atomic_float, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_vec2", test_atomic_vec2, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_vec3", test_atomic_vec3, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_vec4", test_atomic_vec4, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_mat22", test_atomic_mat22, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_mat33", test_atomic_mat33, devices=devices)
-    add_function_test(TestAtomic, "test_atomic_mat44", test_atomic_mat44, devices=devices)
 
-    return TestAtomic
+add_function_test(TestAtomic, "test_atomic_int", test_atomic_int, devices=devices)
+add_function_test(TestAtomic, "test_atomic_float", test_atomic_float, devices=devices)
+add_function_test(TestAtomic, "test_atomic_vec2", test_atomic_vec2, devices=devices)
+add_function_test(TestAtomic, "test_atomic_vec3", test_atomic_vec3, devices=devices)
+add_function_test(TestAtomic, "test_atomic_vec4", test_atomic_vec4, devices=devices)
+add_function_test(TestAtomic, "test_atomic_mat22", test_atomic_mat22, devices=devices)
+add_function_test(TestAtomic, "test_atomic_mat33", test_atomic_mat33, devices=devices)
+add_function_test(TestAtomic, "test_atomic_mat44", test_atomic_mat44, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_bool.py b/warp/tests/test_bool.py
index 94c4d1283..848bed5ad 100644
--- a/warp/tests/test_bool.py
+++ b/warp/tests/test_bool.py
@@ -1,9 +1,16 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -71,24 +78,22 @@ def check_compile_constant(result: wp.array(dtype=wp.bool)):
 
 
 def test_bool_constant(test, device):
-    compile_constant_value = wp.zeros(1, dtype=wp.bool)
-    wp.launch(check_compile_constant, 1, inputs=[compile_constant_value])
+    compile_constant_value = wp.zeros(1, dtype=wp.bool, device=device)
+    wp.launch(check_compile_constant, 1, inputs=[compile_constant_value], device=device)
     test.assertTrue(compile_constant_value.numpy()[0])
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestBool(parent):
-        pass
+class TestBool(unittest.TestCase):
+    pass
 
-    add_function_test(TestBool, "test_bool_identity_ops", test_bool_identity_ops, devices=devices)
-    add_function_test(TestBool, "test_bool_constant", test_bool_constant, devices=devices)
 
-    return TestBool
+add_function_test(TestBool, "test_bool_identity_ops", test_bool_identity_ops, devices=devices)
+add_function_test(TestBool, "test_bool_constant", test_bool_constant, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_builtins_resolution.py b/warp/tests/test_builtins_resolution.py
index bfeb6fa2c..c40902521 100644
--- a/warp/tests/test_builtins_resolution.py
+++ b/warp/tests/test_builtins_resolution.py
@@ -10,8 +10,7 @@
 
 import numpy as np
 
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -42,733 +41,6 @@ def wpm(dtype, dim, values):
     return tuple(wpv(dtype, values[i * dim : (i + 1) * dim]) for i in range(dim))
 
 
-def test_int_arg_overflow(test, device):
-    value = -1234567890123456789
-
-    test.assertEqual(wp.invert(wp.int8(value)), 20)
-    test.assertEqual(wp.invert(wp.int16(value)), -32492)
-    test.assertEqual(wp.invert(wp.int32(value)), 2112454932)
-    test.assertEqual(wp.invert(wp.int64(value)), 1234567890123456788)
-
-    test.assertEqual(wp.invert(wp.uint8(value)), 20)
-    test.assertEqual(wp.invert(wp.uint16(value)), 33044)
-    test.assertEqual(wp.invert(wp.uint32(value)), 2112454932)
-    test.assertEqual(wp.invert(wp.uint64(value)), 1234567890123456788)
-
-    test.assertEqual(wp.invert(value), wp.invert(wp.int32(value)))
-
-
-def test_float_arg_precision(test, device):
-    value = 1.23
-    expected = 0.94248880193169748409
-
-    result = wp.sin(wp.float64(value))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.sin(wp.float32(value))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.sin(wp.float16(value))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    test.assertEqual(wp.sin(value), wp.sin(wp.float32(value)))
-
-
-def test_int_int_args_overflow(test, device):
-    value = -1234567890
-
-    test.assertEqual(wp.mul(wp.int8(value), wp.int8(value)), 68)
-    test.assertEqual(wp.mul(wp.int16(value), wp.int16(value)), -3004)
-    test.assertEqual(wp.mul(wp.int32(value), wp.int32(value)), 304084036)
-    test.assertEqual(wp.mul(wp.int64(value), wp.int64(value)), 1524157875019052100)
-
-    test.assertEqual(wp.mul(wp.uint8(value), wp.uint8(value)), 68)
-    test.assertEqual(wp.mul(wp.uint16(value), wp.uint16(value)), 62532)
-    test.assertEqual(wp.mul(wp.uint32(value), wp.uint32(value)), 304084036)
-    test.assertEqual(wp.mul(wp.uint64(value), wp.uint64(value)), 1524157875019052100)
-
-    test.assertEqual(wp.mul(value, value), wp.mul(wp.int32(value), wp.int32(value)))
-
-
-def test_mat22_arg_precision(test, device):
-    values = (1.23, 2.34, 3.45, 4.56)
-    values_2d = (values[0:2], values[2:4])
-    expected = 5.78999999999999914735
-
-    result = wp.trace(wp.mat22d(*values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.trace(wp.mat22f(*values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.trace(wp.mat22h(*values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.trace(values), wp.trace(wp.mat22f(*values)))
-        test.assertEqual(wp.trace(values_2d), wp.trace(wp.mat22f(*values)))
-
-
-def test_mat33_arg_precision(test, device):
-    values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
-    values_2d = (values[0:3], values[3:6], values[6:9])
-    expected = 15.91000000000000014211
-
-    result = wp.trace(wp.mat33d(*values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.trace(wp.mat33f(*values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.trace(wp.mat33h(*values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.trace(values), wp.trace(wp.mat33f(*values)))
-        test.assertEqual(wp.trace(values_2d), wp.trace(wp.mat33f(*values)))
-
-
-def test_mat44_arg_precision(test, device):
-    values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01, 10.12, 11.23, 12.34, 13.45, 14.56, 15.67, 16.78)
-    values_2d = (values[0:4], values[4:8], values[8:12], values[12:16])
-    expected = 36.02000000000000312639
-
-    result = wp.trace(wp.mat44d(*values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.trace(wp.mat44f(*values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.trace(wp.mat44h(*values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.trace(values), wp.trace(wp.mat44f(*values)))
-        test.assertEqual(wp.trace(values_2d), wp.trace(wp.mat44f(*values)))
-
-
-def test_mat22_mat22_args_precision(test, device):
-    a_values = (0.12, 1.23, 0.12, 1.23)
-    a_values_2d = (a_values[0:2], a_values[2:4])
-    b_values = (1.23, 0.12, 1.23, 0.12)
-    b_values_2d = (b_values[0:2], b_values[2:4])
-    expected = 0.59039999999999992486
-
-    result = wp.ddot(wp.mat22d(*a_values), wp.mat22d(*b_values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.ddot(wp.mat22h(*a_values), wp.mat22h(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
-        test.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
-        test.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
-        test.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
-
-
-def test_mat33_mat33_args_precision(test, device):
-    a_values = (0.12, 1.23, 2.34, 0.12, 1.23, 2.34, 0.12, 1.23, 2.34)
-    a_values_2d = (a_values[0:3], a_values[3:6], a_values[6:9])
-    b_values = (2.34, 1.23, 0.12, 2.34, 1.23, 0.12, 2.34, 1.23, 0.12)
-    b_values_2d = (b_values[0:3], b_values[3:6], b_values[6:9])
-    expected = 6.22350000000000047606
-
-    result = wp.ddot(wp.mat33d(*a_values), wp.mat33d(*b_values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.ddot(wp.mat33h(*a_values), wp.mat33h(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
-        test.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
-        test.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
-        test.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
-
-
-def test_mat44_mat44_args(test, device):
-    a_values = (0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45)
-    a_values_2d = (a_values[0:4], a_values[4:8], a_values[8:12], a_values[12:16])
-    b_values = (3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12)
-    b_values_2d = (b_values[0:4], b_values[4:8], b_values[8:12], b_values[12:16])
-    expected = 26.33760000000000189857
-
-    result = wp.ddot(wp.mat44d(*a_values), wp.mat44d(*b_values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.ddot(wp.mat44h(*a_values), wp.mat44h(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
-        test.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
-        test.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
-        test.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
-
-
-def test_mat22_float_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45, 4.56)
-    a_values_2d = (a_values[0:2], a_values[2:4])
-    b_value = 0.12
-    expected_00 = 0.14759999999999998122
-    expected_01 = 0.28079999999999999405
-    expected_10 = 0.41399999999999997913
-    expected_11 = 0.54719999999999990870
-
-    result = wp.mul(wp.mat22d(*a_values), wp.float64(b_value))
-    test.assertAlmostEqual(result[0][0], expected_00, places=12)
-    test.assertAlmostEqual(result[0][1], expected_01, places=12)
-    test.assertAlmostEqual(result[1][0], expected_10, places=12)
-    test.assertAlmostEqual(result[1][1], expected_11, places=12)
-
-    result = wp.mul(wp.mat22f(*a_values), wp.float32(b_value))
-    test.assertNotAlmostEqual(result[0][0], expected_00, places=12)
-    test.assertNotAlmostEqual(result[0][1], expected_01, places=12)
-    test.assertNotAlmostEqual(result[1][0], expected_10, places=12)
-    test.assertNotAlmostEqual(result[1][1], expected_11, places=12)
-    test.assertAlmostEqual(result[0][0], expected_00, places=5)
-    test.assertAlmostEqual(result[0][1], expected_01, places=5)
-    test.assertAlmostEqual(result[1][0], expected_10, places=5)
-    test.assertAlmostEqual(result[1][1], expected_11, places=5)
-
-    result = wp.mul(wp.mat22h(*a_values), wp.float16(b_value))
-    test.assertNotAlmostEqual(result[0][0], expected_00, places=5)
-    test.assertNotAlmostEqual(result[0][1], expected_01, places=5)
-    test.assertNotAlmostEqual(result[1][0], expected_10, places=5)
-    test.assertNotAlmostEqual(result[1][1], expected_11, places=5)
-    test.assertAlmostEqual(result[0][0], expected_00, places=1)
-    test.assertAlmostEqual(result[0][1], expected_01, places=1)
-    test.assertAlmostEqual(result[1][0], expected_10, places=1)
-    test.assertAlmostEqual(result[1][1], expected_11, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        # Multiplying a 1-D tuple of length 4 is ambiguous because it could match
-        # either the `vec4f` or `mat22f` overload. As a result, only the 2-D variant
-        # of the tuple is expected to resolve correctly.
-        test.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat22f(*a_values), wp.float32(b_value)))
-
-
-def test_mat33_float_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
-    a_values_2d = (a_values[0:3], a_values[3:6], a_values[6:9])
-    b_value = 0.12
-    expected_00 = 0.14759999999999998122
-    expected_01 = 0.28079999999999999405
-    expected_02 = 0.41399999999999997913
-    expected_10 = 0.54719999999999990870
-    expected_11 = 0.68040000000000000480
-    expected_12 = 0.81359999999999998987
-    expected_20 = 0.94679999999999997495
-    expected_21 = 1.06800000000000006040
-    expected_22 = 1.08119999999999993889
-
-    result = wp.mul(wp.mat33d(*a_values), wp.float64(b_value))
-    test.assertAlmostEqual(result[0][0], expected_00, places=12)
-    test.assertAlmostEqual(result[0][1], expected_01, places=12)
-    test.assertAlmostEqual(result[0][2], expected_02, places=12)
-    test.assertAlmostEqual(result[1][0], expected_10, places=12)
-    test.assertAlmostEqual(result[1][1], expected_11, places=12)
-    test.assertAlmostEqual(result[1][2], expected_12, places=12)
-    test.assertAlmostEqual(result[2][0], expected_20, places=12)
-    test.assertAlmostEqual(result[2][1], expected_21, places=12)
-    test.assertAlmostEqual(result[2][2], expected_22, places=12)
-
-    result = wp.mul(wp.mat33f(*a_values), wp.float32(b_value))
-    test.assertNotAlmostEqual(result[0][0], expected_00, places=12)
-    test.assertNotAlmostEqual(result[0][1], expected_01, places=12)
-    test.assertNotAlmostEqual(result[0][2], expected_02, places=12)
-    test.assertNotAlmostEqual(result[1][0], expected_10, places=12)
-    test.assertNotAlmostEqual(result[1][1], expected_11, places=12)
-    test.assertNotAlmostEqual(result[1][2], expected_12, places=12)
-    test.assertNotAlmostEqual(result[2][0], expected_20, places=12)
-    test.assertNotAlmostEqual(result[2][1], expected_21, places=12)
-    test.assertNotAlmostEqual(result[2][2], expected_22, places=12)
-    test.assertAlmostEqual(result[0][0], expected_00, places=5)
-    test.assertAlmostEqual(result[0][1], expected_01, places=5)
-    test.assertAlmostEqual(result[0][2], expected_02, places=5)
-    test.assertAlmostEqual(result[1][0], expected_10, places=5)
-    test.assertAlmostEqual(result[1][1], expected_11, places=5)
-    test.assertAlmostEqual(result[1][2], expected_12, places=5)
-    test.assertAlmostEqual(result[2][0], expected_20, places=5)
-    test.assertAlmostEqual(result[2][1], expected_21, places=5)
-    test.assertAlmostEqual(result[2][2], expected_22, places=5)
-
-    result = wp.mul(wp.mat33h(*a_values), wp.float16(b_value))
-    test.assertNotAlmostEqual(result[0][0], expected_00, places=5)
-    test.assertNotAlmostEqual(result[0][1], expected_01, places=5)
-    test.assertNotAlmostEqual(result[0][2], expected_02, places=5)
-    test.assertNotAlmostEqual(result[1][0], expected_10, places=5)
-    test.assertNotAlmostEqual(result[1][1], expected_11, places=5)
-    test.assertNotAlmostEqual(result[1][2], expected_12, places=5)
-    test.assertNotAlmostEqual(result[2][0], expected_20, places=5)
-    test.assertNotAlmostEqual(result[2][1], expected_21, places=5)
-    test.assertNotAlmostEqual(result[2][2], expected_22, places=5)
-    test.assertAlmostEqual(result[0][0], expected_00, places=1)
-    test.assertAlmostEqual(result[0][1], expected_01, places=1)
-    test.assertAlmostEqual(result[0][2], expected_02, places=1)
-    test.assertAlmostEqual(result[1][0], expected_10, places=1)
-    test.assertAlmostEqual(result[1][1], expected_11, places=1)
-    test.assertAlmostEqual(result[1][2], expected_12, places=1)
-    test.assertAlmostEqual(result[2][0], expected_20, places=1)
-    test.assertAlmostEqual(result[2][1], expected_21, places=1)
-    test.assertAlmostEqual(result[2][2], expected_22, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.mat33f(*a_values), wp.float32(b_value)))
-        test.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat33f(*a_values), wp.float32(b_value)))
-
-
-def test_mat44_float_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01, 10.12, 11.23, 12.34, 13.45, 14.56, 15.67, 16.78)
-    a_values_2d = (a_values[0:4], a_values[4:8], a_values[8:12], a_values[12:16])
-    b_value = 0.12
-    expected_00 = 0.14759999999999998122
-    expected_01 = 0.28079999999999999405
-    expected_02 = 0.41399999999999997913
-    expected_03 = 0.54719999999999990870
-    expected_10 = 0.68040000000000000480
-    expected_11 = 0.81359999999999998987
-    expected_12 = 0.94679999999999997495
-    expected_13 = 1.06800000000000006040
-    expected_20 = 1.08119999999999993889
-    expected_21 = 1.21439999999999992397
-    expected_22 = 1.34759999999999990905
-    expected_23 = 1.48079999999999989413
-    expected_30 = 1.61399999999999987921
-    expected_31 = 1.74720000000000008633
-    expected_32 = 1.88039999999999984936
-    expected_33 = 2.01360000000000027853
-
-    result = wp.mul(wp.mat44d(*a_values), wp.float64(b_value))
-    test.assertAlmostEqual(result[0][0], expected_00, places=12)
-    test.assertAlmostEqual(result[0][1], expected_01, places=12)
-    test.assertAlmostEqual(result[0][2], expected_02, places=12)
-    test.assertAlmostEqual(result[0][3], expected_03, places=12)
-    test.assertAlmostEqual(result[1][0], expected_10, places=12)
-    test.assertAlmostEqual(result[1][1], expected_11, places=12)
-    test.assertAlmostEqual(result[1][2], expected_12, places=12)
-    test.assertAlmostEqual(result[1][3], expected_13, places=12)
-    test.assertAlmostEqual(result[2][0], expected_20, places=12)
-    test.assertAlmostEqual(result[2][1], expected_21, places=12)
-    test.assertAlmostEqual(result[2][2], expected_22, places=12)
-    test.assertAlmostEqual(result[2][3], expected_23, places=12)
-    test.assertAlmostEqual(result[3][0], expected_30, places=12)
-    test.assertAlmostEqual(result[3][1], expected_31, places=12)
-    test.assertAlmostEqual(result[3][2], expected_32, places=12)
-    test.assertAlmostEqual(result[3][3], expected_33, places=12)
-
-    result = wp.mul(wp.mat44f(*a_values), wp.float32(b_value))
-    test.assertNotAlmostEqual(result[0][0], expected_00, places=12)
-    test.assertNotAlmostEqual(result[0][1], expected_01, places=12)
-    test.assertNotAlmostEqual(result[0][2], expected_02, places=12)
-    test.assertNotAlmostEqual(result[0][3], expected_03, places=12)
-    test.assertNotAlmostEqual(result[1][0], expected_10, places=12)
-    test.assertNotAlmostEqual(result[1][1], expected_11, places=12)
-    test.assertNotAlmostEqual(result[1][2], expected_12, places=12)
-    test.assertNotAlmostEqual(result[1][3], expected_13, places=12)
-    test.assertNotAlmostEqual(result[2][0], expected_20, places=12)
-    test.assertNotAlmostEqual(result[2][1], expected_21, places=12)
-    test.assertNotAlmostEqual(result[2][2], expected_22, places=12)
-    test.assertNotAlmostEqual(result[2][3], expected_23, places=12)
-    test.assertNotAlmostEqual(result[3][0], expected_30, places=12)
-    test.assertNotAlmostEqual(result[3][1], expected_31, places=12)
-    test.assertNotAlmostEqual(result[3][2], expected_32, places=12)
-    test.assertNotAlmostEqual(result[3][3], expected_33, places=12)
-    test.assertAlmostEqual(result[0][0], expected_00, places=5)
-    test.assertAlmostEqual(result[0][1], expected_01, places=5)
-    test.assertAlmostEqual(result[0][2], expected_02, places=5)
-    test.assertAlmostEqual(result[0][3], expected_03, places=5)
-    test.assertAlmostEqual(result[1][0], expected_10, places=5)
-    test.assertAlmostEqual(result[1][1], expected_11, places=5)
-    test.assertAlmostEqual(result[1][2], expected_12, places=5)
-    test.assertAlmostEqual(result[1][3], expected_13, places=5)
-    test.assertAlmostEqual(result[2][0], expected_20, places=5)
-    test.assertAlmostEqual(result[2][1], expected_21, places=5)
-    test.assertAlmostEqual(result[2][2], expected_22, places=5)
-    test.assertAlmostEqual(result[2][3], expected_23, places=5)
-    test.assertAlmostEqual(result[3][0], expected_30, places=5)
-    test.assertAlmostEqual(result[3][1], expected_31, places=5)
-    test.assertAlmostEqual(result[3][2], expected_32, places=5)
-    test.assertAlmostEqual(result[3][3], expected_33, places=5)
-
-    result = wp.mul(wp.mat44h(*a_values), wp.float16(b_value))
-    test.assertNotAlmostEqual(result[0][0], expected_00, places=5)
-    test.assertNotAlmostEqual(result[0][1], expected_01, places=5)
-    test.assertNotAlmostEqual(result[0][2], expected_02, places=5)
-    test.assertNotAlmostEqual(result[0][3], expected_03, places=5)
-    test.assertNotAlmostEqual(result[1][0], expected_10, places=5)
-    test.assertNotAlmostEqual(result[1][1], expected_11, places=5)
-    test.assertNotAlmostEqual(result[1][2], expected_12, places=5)
-    test.assertNotAlmostEqual(result[1][3], expected_13, places=5)
-    test.assertNotAlmostEqual(result[2][0], expected_20, places=5)
-    test.assertNotAlmostEqual(result[2][1], expected_21, places=5)
-    test.assertNotAlmostEqual(result[2][2], expected_22, places=5)
-    test.assertNotAlmostEqual(result[2][3], expected_23, places=5)
-    test.assertNotAlmostEqual(result[3][0], expected_30, places=5)
-    test.assertNotAlmostEqual(result[3][1], expected_31, places=5)
-    test.assertNotAlmostEqual(result[3][2], expected_32, places=5)
-    test.assertNotAlmostEqual(result[3][3], expected_33, places=5)
-    test.assertAlmostEqual(result[0][0], expected_00, places=1)
-    test.assertAlmostEqual(result[0][1], expected_01, places=1)
-    test.assertAlmostEqual(result[0][2], expected_02, places=1)
-    test.assertAlmostEqual(result[0][3], expected_03, places=1)
-    test.assertAlmostEqual(result[1][0], expected_10, places=1)
-    test.assertAlmostEqual(result[1][1], expected_11, places=1)
-    test.assertAlmostEqual(result[1][2], expected_12, places=1)
-    test.assertAlmostEqual(result[1][3], expected_13, places=1)
-    test.assertAlmostEqual(result[2][0], expected_20, places=1)
-    test.assertAlmostEqual(result[2][1], expected_21, places=1)
-    test.assertAlmostEqual(result[2][2], expected_22, places=1)
-    test.assertAlmostEqual(result[2][3], expected_23, places=1)
-    test.assertAlmostEqual(result[3][0], expected_30, places=1)
-    test.assertAlmostEqual(result[3][1], expected_31, places=1)
-    test.assertAlmostEqual(result[3][2], expected_32, places=1)
-    test.assertAlmostEqual(result[3][3], expected_33, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.mat44f(*a_values), wp.float32(b_value)))
-        test.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat44f(*a_values), wp.float32(b_value)))
-
-
-def test_vec2_arg_precision(test, device):
-    values = (1.23, 2.34)
-    expected = 2.64357712200722438922
-
-    result = wp.length(wp.vec2d(*values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.length(wp.vec2f(*values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.length(wp.vec2h(*values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.length(values), wp.length(wp.vec2f(*values)))
-
-
-def test_vec2_arg_overflow(test, device):
-    values = (-1234567890, -1234567890)
-
-    test.assertEqual(wp.length_sq(wp.vec2b(*values)), -120)
-    test.assertEqual(wp.length_sq(wp.vec2s(*values)), -6008)
-    test.assertEqual(wp.length_sq(wp.vec2i(*values)), 608168072)
-    test.assertEqual(wp.length_sq(wp.vec2l(*values)), 3048315750038104200)
-
-    test.assertEqual(wp.length_sq(wp.vec2ub(*values)), 136)
-    test.assertEqual(wp.length_sq(wp.vec2us(*values)), 59528)
-    test.assertEqual(wp.length_sq(wp.vec2ui(*values)), 608168072)
-    test.assertEqual(wp.length_sq(wp.vec2ul(*values)), 3048315750038104200)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec2i(*values)))
-
-
-def test_vec3_arg_precision(test, device):
-    values = (1.23, 2.34, 3.45)
-    expected = 4.34637780226247727455
-
-    result = wp.length(wp.vec3d(*values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.length(wp.vec3f(*values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.length(wp.vec3h(*values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.length(values), wp.length(wp.vec3f(*values)))
-
-
-def test_vec3_arg_overflow(test, device):
-    values = (-1234567890, -1234567890, -1234567890)
-
-    test.assertEqual(wp.length_sq(wp.vec3b(*values)), -52)
-    test.assertEqual(wp.length_sq(wp.vec3s(*values)), -9012)
-    test.assertEqual(wp.length_sq(wp.vec3i(*values)), 912252108)
-    test.assertEqual(wp.length_sq(wp.vec3l(*values)), 4572473625057156300)
-
-    test.assertEqual(wp.length_sq(wp.vec3ub(*values)), 204)
-    test.assertEqual(wp.length_sq(wp.vec3us(*values)), 56524)
-    test.assertEqual(wp.length_sq(wp.vec3ui(*values)), 912252108)
-    test.assertEqual(wp.length_sq(wp.vec3ul(*values)), 4572473625057156300)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec3i(*values)))
-
-
-def test_vec4_arg_precision(test, device):
-    values = (1.23, 2.34, 3.45, 4.56)
-    expected = 6.29957141399317777086
-
-    result = wp.length(wp.vec4d(*values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.length(wp.vec4f(*values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.length(wp.vec4h(*values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.length(values), wp.length(wp.vec4f(*values)))
-
-
-def test_vec4_arg_overflow(test, device):
-    values = (-1234567890, -1234567890, -1234567890, -1234567890)
-
-    test.assertEqual(wp.length_sq(wp.vec4b(*values)), 16)
-    test.assertEqual(wp.length_sq(wp.vec4s(*values)), -12016)
-    test.assertEqual(wp.length_sq(wp.vec4i(*values)), 1216336144)
-    test.assertEqual(wp.length_sq(wp.vec4l(*values)), 6096631500076208400)
-
-    test.assertEqual(wp.length_sq(wp.vec4ub(*values)), 16)
-    test.assertEqual(wp.length_sq(wp.vec4us(*values)), 53520)
-    test.assertEqual(wp.length_sq(wp.vec4ui(*values)), 1216336144)
-    test.assertEqual(wp.length_sq(wp.vec4ul(*values)), 6096631500076208400)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec4i(*values)))
-
-
-def test_vec2_vec2_args_precision(test, device):
-    a_values = (1.23, 2.34)
-    b_values = (3.45, 4.56)
-    expected = 14.91389999999999815827
-
-    result = wp.dot(wp.vec2d(*a_values), wp.vec2d(*b_values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.dot(wp.vec2f(*a_values), wp.vec2f(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.dot(wp.vec2h(*a_values), wp.vec2h(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec2f(*a_values), wp.vec2f(*b_values)))
-
-
-def test_vec2_vec2_args_overflow(test, device):
-    values = (-1234567890, -1234567890)
-
-    test.assertEqual(wp.dot(wp.vec2b(*values), wp.vec2b(*values)), -120)
-    test.assertEqual(wp.dot(wp.vec2s(*values), wp.vec2s(*values)), -6008)
-    test.assertEqual(wp.dot(wp.vec2i(*values), wp.vec2i(*values)), 608168072)
-    test.assertEqual(wp.dot(wp.vec2l(*values), wp.vec2l(*values)), 3048315750038104200)
-
-    test.assertEqual(wp.dot(wp.vec2ub(*values), wp.vec2ub(*values)), 136)
-    test.assertEqual(wp.dot(wp.vec2us(*values), wp.vec2us(*values)), 59528)
-    test.assertEqual(wp.dot(wp.vec2ui(*values), wp.vec2ui(*values)), 608168072)
-    test.assertEqual(wp.dot(wp.vec2ul(*values), wp.vec2ul(*values)), 3048315750038104200)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.dot(values, values), wp.dot(wp.vec2i(*values), wp.vec2i(*values)))
-
-
-def test_vec3_vec3_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45)
-    b_values = (4.56, 5.67, 6.78)
-    expected = 42.26760000000000161435
-
-    result = wp.dot(wp.vec3d(*a_values), wp.vec3d(*b_values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.dot(wp.vec3f(*a_values), wp.vec3f(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.dot(wp.vec3h(*a_values), wp.vec3h(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec3f(*a_values), wp.vec3f(*b_values)))
-
-
-def test_vec3_vec3_args_overflow(test, device):
-    values = (-1234567890, -1234567890, -1234567890)
-
-    test.assertEqual(wp.dot(wp.vec3b(*values), wp.vec3b(*values)), -52)
-    test.assertEqual(wp.dot(wp.vec3s(*values), wp.vec3s(*values)), -9012)
-    test.assertEqual(wp.dot(wp.vec3i(*values), wp.vec3i(*values)), 912252108)
-    test.assertEqual(wp.dot(wp.vec3l(*values), wp.vec3l(*values)), 4572473625057156300)
-
-    test.assertEqual(wp.dot(wp.vec3ub(*values), wp.vec3ub(*values)), 204)
-    test.assertEqual(wp.dot(wp.vec3us(*values), wp.vec3us(*values)), 56524)
-    test.assertEqual(wp.dot(wp.vec3ui(*values), wp.vec3ui(*values)), 912252108)
-    test.assertEqual(wp.dot(wp.vec3ul(*values), wp.vec3ul(*values)), 4572473625057156300)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.dot(values, values), wp.dot(wp.vec3i(*values), wp.vec3i(*values)))
-
-
-def test_vec4_vec4_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45, 4.56)
-    b_values = (5.67, 6.78, 7.89, 8.90)
-    expected = 90.64379999999999881766
-
-    result = wp.dot(wp.vec4d(*a_values), wp.vec4d(*b_values))
-    test.assertAlmostEqual(result, expected, places=12)
-
-    result = wp.dot(wp.vec4f(*a_values), wp.vec4f(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=12)
-    test.assertAlmostEqual(result, expected, places=5)
-
-    result = wp.dot(wp.vec4h(*a_values), wp.vec4h(*b_values))
-    test.assertNotAlmostEqual(result, expected, places=5)
-    test.assertAlmostEqual(result, expected, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec4f(*a_values), wp.vec4f(*b_values)))
-
-
-def test_vec4_vec4_args_overflow(test, device):
-    values = (-1234567890, -1234567890, -1234567890, -1234567890)
-
-    test.assertEqual(wp.dot(wp.vec4b(*values), wp.vec4b(*values)), 16)
-    test.assertEqual(wp.dot(wp.vec4s(*values), wp.vec4s(*values)), -12016)
-    test.assertEqual(wp.dot(wp.vec4i(*values), wp.vec4i(*values)), 1216336144)
-    test.assertEqual(wp.dot(wp.vec4l(*values), wp.vec4l(*values)), 6096631500076208400)
-
-    test.assertEqual(wp.dot(wp.vec4ub(*values), wp.vec4ub(*values)), 16)
-    test.assertEqual(wp.dot(wp.vec4us(*values), wp.vec4us(*values)), 53520)
-    test.assertEqual(wp.dot(wp.vec4ui(*values), wp.vec4ui(*values)), 1216336144)
-    test.assertEqual(wp.dot(wp.vec4ul(*values), wp.vec4ul(*values)), 6096631500076208400)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.dot(values, values), wp.dot(wp.vec4i(*values), wp.vec4i(*values)))
-
-
-def test_vec2_float_args_precision(test, device):
-    a_values = (1.23, 2.34)
-    b_value = 3.45
-    expected_x = 4.24350000000000004974
-    expected_y = 8.07300000000000039790
-
-    result = wp.mul(wp.vec2d(*a_values), wp.float64(b_value))
-    test.assertAlmostEqual(result[0], expected_x, places=12)
-    test.assertAlmostEqual(result[1], expected_y, places=12)
-
-    result = wp.mul(wp.vec2f(*a_values), wp.float32(b_value))
-    test.assertNotAlmostEqual(result[0], expected_x, places=12)
-    test.assertNotAlmostEqual(result[1], expected_y, places=12)
-    test.assertAlmostEqual(result[0], expected_x, places=5)
-    test.assertAlmostEqual(result[1], expected_y, places=5)
-
-    result = wp.mul(wp.vec2h(*a_values), wp.float16(b_value))
-    test.assertNotAlmostEqual(result[0], expected_x, places=5)
-    test.assertNotAlmostEqual(result[1], expected_y, places=5)
-    test.assertAlmostEqual(result[0], expected_x, places=1)
-    test.assertAlmostEqual(result[1], expected_y, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec2f(*a_values), wp.float32(b_value)))
-
-
-def test_vec3_float_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45)
-    b_value = 4.56
-    expected_x = 5.60879999999999956373
-    expected_y = 10.67039999999999899671
-    expected_z = 15.73199999999999931788
-
-    result = wp.mul(wp.vec3d(*a_values), wp.float64(b_value))
-    test.assertAlmostEqual(result[0], expected_x, places=12)
-    test.assertAlmostEqual(result[1], expected_y, places=12)
-    test.assertAlmostEqual(result[2], expected_z, places=12)
-
-    result = wp.mul(wp.vec3f(*a_values), wp.float32(b_value))
-    test.assertNotAlmostEqual(result[0], expected_x, places=12)
-    test.assertNotAlmostEqual(result[1], expected_y, places=12)
-    test.assertNotAlmostEqual(result[2], expected_z, places=12)
-    test.assertAlmostEqual(result[0], expected_x, places=5)
-    test.assertAlmostEqual(result[1], expected_y, places=5)
-    test.assertAlmostEqual(result[2], expected_z, places=5)
-
-    result = wp.mul(wp.vec3h(*a_values), wp.float16(b_value))
-    test.assertNotAlmostEqual(result[0], expected_x, places=5)
-    test.assertNotAlmostEqual(result[1], expected_y, places=5)
-    test.assertNotAlmostEqual(result[2], expected_z, places=5)
-    test.assertAlmostEqual(result[0], expected_x, places=1)
-    test.assertAlmostEqual(result[1], expected_y, places=1)
-    test.assertAlmostEqual(result[2], expected_z, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec3f(*a_values), wp.float32(b_value)))
-
-
-def test_vec4_float_args_precision(test, device):
-    a_values = (1.23, 2.34, 3.45, 4.56)
-    b_value = 5.67
-    expected_x = 6.97409999999999996589
-    expected_y = 13.26779999999999937188
-    expected_z = 19.56150000000000233058
-    expected_w = 25.85519999999999640750
-
-    result = wp.mul(wp.vec4d(*a_values), wp.float64(b_value))
-    test.assertAlmostEqual(result[0], expected_x, places=12)
-    test.assertAlmostEqual(result[1], expected_y, places=12)
-    test.assertAlmostEqual(result[2], expected_z, places=12)
-    test.assertAlmostEqual(result[3], expected_w, places=12)
-
-    result = wp.mul(wp.vec4f(*a_values), wp.float32(b_value))
-    test.assertNotAlmostEqual(result[0], expected_x, places=12)
-    test.assertNotAlmostEqual(result[1], expected_y, places=12)
-    test.assertNotAlmostEqual(result[2], expected_z, places=12)
-    test.assertNotAlmostEqual(result[3], expected_w, places=12)
-    test.assertAlmostEqual(result[0], expected_x, places=5)
-    test.assertAlmostEqual(result[1], expected_y, places=5)
-    test.assertAlmostEqual(result[2], expected_z, places=5)
-    test.assertAlmostEqual(result[3], expected_w, places=5)
-
-    result = wp.mul(wp.vec4h(*a_values), wp.float16(b_value))
-    test.assertNotAlmostEqual(result[0], expected_x, places=5)
-    test.assertNotAlmostEqual(result[1], expected_y, places=5)
-    test.assertNotAlmostEqual(result[2], expected_z, places=5)
-    test.assertNotAlmostEqual(result[3], expected_w, places=5)
-    test.assertAlmostEqual(result[0], expected_x, places=1)
-    test.assertAlmostEqual(result[1], expected_y, places=1)
-    test.assertAlmostEqual(result[2], expected_z, places=1)
-    test.assertAlmostEqual(result[3], expected_w, places=1)
-
-    with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
-        test.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec4f(*a_values), wp.float32(b_value)))
-
-
 def test_int_arg_support(test, device, dtype):
     np_type = wp.types.warp_type_to_np_dtype[dtype]
     value = -1234567890123456789
@@ -806,29 +78,25 @@ def test_int_int_args_support(test, device, dtype):
     else:
         with test.assertRaisesRegex(
             RuntimeError,
-            rf"Couldn't find a function 'mul' compatible with "
-            rf"the arguments '{dtype.__name__}, int'$",
+            rf"Couldn't find a function 'mul' compatible with " rf"the arguments '{dtype.__name__}, int'$",
         ):
             wp.mul(dtype(value), value)
 
         with test.assertRaisesRegex(
             RuntimeError,
-            rf"Couldn't find a function 'mul' compatible with "
-            rf"the arguments '{np_type.__name__}, int'$",
+            rf"Couldn't find a function 'mul' compatible with " rf"the arguments '{np_type.__name__}, int'$",
         ):
             wp.mul(nps(np_type, value), value)
 
         with test.assertRaisesRegex(
             RuntimeError,
-            rf"Couldn't find a function 'mul' compatible with "
-            rf"the arguments 'int, {dtype.__name__}'$",
+            rf"Couldn't find a function 'mul' compatible with " rf"the arguments 'int, {dtype.__name__}'$",
         ):
             wp.mul(value, dtype(value))
 
         with test.assertRaisesRegex(
             RuntimeError,
-            rf"Couldn't find a function 'mul' compatible with "
-            rf"the arguments 'int, {np_type.__name__}'$",
+            rf"Couldn't find a function 'mul' compatible with " rf"the arguments 'int, {np_type.__name__}'$",
         ):
             wp.mul(value, nps(np_type, value))
 
@@ -915,85 +183,73 @@ def test_mat_mat_args_support(test, device, dtype):
         else:
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'mat_t, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'mat_t, tuple'$",
             ):
                 wp.ddot(mat_cls(*a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(wpv(dtype, a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(wpm(dtype, 3, a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(npv(np_type, a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(npm(np_type, 3, a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'ndarray, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'ndarray, tuple'$",
             ):
                 wp.ddot(np.array(npv(np_type, a_values)), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, mat_t'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, mat_t'$",
             ):
                 wp.ddot(a_values, mat_cls(*b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(a_values, wpv(dtype, b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(a_values, wpm(dtype, 3, b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(a_values, npv(np_type, b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.ddot(a_values, npm(np_type, 3, b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'ddot' compatible with "
-                r"the arguments 'tuple, ndarray'$",
+                r"Couldn't find a function 'ddot' compatible with " r"the arguments 'tuple, ndarray'$",
             ):
                 wp.ddot(a_values, np.array(npv(np_type, b_values)))
 
@@ -1038,57 +294,49 @@ def test_mat_float_args_support(test, device, dtype):
         else:
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'mat_t, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'mat_t, float'$",
             ):
                 wp.mul(mat_cls(*a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'tuple, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'tuple, float'$",
             ):
                 wp.mul(wpv(dtype, a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'tuple, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'tuple, float'$",
             ):
                 wp.mul(wpm(dtype, 3, a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'tuple, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'tuple, float'$",
             ):
                 wp.mul(npv(np_type, a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'tuple, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'tuple, float'$",
             ):
                 wp.mul(npm(np_type, 3, a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'ndarray, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'ndarray, float'$",
             ):
                 wp.mul(np.array(npv(np_type, a_values)), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                fr"Couldn't find a function 'mul' compatible with "
-                fr"the arguments 'tuple, {dtype.__name__}'$",
+                rf"Couldn't find a function 'mul' compatible with " rf"the arguments 'tuple, {dtype.__name__}'$",
             ):
                 wp.mul(a_values, dtype(b_value))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                fr"Couldn't find a function 'mul' compatible with "
-                fr"the arguments 'tuple, {np_type.__name__}'$",
+                rf"Couldn't find a function 'mul' compatible with " rf"the arguments 'tuple, {np_type.__name__}'$",
             ):
                 wp.mul(a_values, nps(np_type, b_value))
 
@@ -1147,57 +395,49 @@ def test_vec_vec_args_support(test, device, dtype):
         else:
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'vec_t, tuple'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'vec_t, tuple'$",
             ):
                 wp.dot(vec_cls(*a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.dot(wpv(dtype, a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.dot(npv(np_type, a_values), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'ndarray, tuple'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'ndarray, tuple'$",
             ):
                 wp.dot(np.array(npv(np_type, a_values)), b_values)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'tuple, vec_t'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'tuple, vec_t'$",
             ):
                 wp.dot(a_values, vec_cls(*b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.dot(a_values, wpv(dtype, b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'tuple, tuple'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'tuple, tuple'$",
             ):
                 wp.dot(a_values, npv(np_type, b_values))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'dot' compatible with "
-                r"the arguments 'tuple, ndarray'$",
+                r"Couldn't find a function 'dot' compatible with " r"the arguments 'tuple, ndarray'$",
             ):
                 wp.dot(a_values, np.array(npv(np_type, b_values)))
 
@@ -1234,141 +474,819 @@ def test_vec_float_args_support(test, device, dtype):
         else:
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'vec_t, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'vec_t, float'$",
             ):
                 wp.mul(vec_cls(*a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'tuple, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'tuple, float'$",
             ):
                 wp.mul(wpv(dtype, a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'tuple, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'tuple, float'$",
             ):
                 wp.mul(npv(np_type, a_values), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                r"Couldn't find a function 'mul' compatible with "
-                r"the arguments 'ndarray, float'$",
+                r"Couldn't find a function 'mul' compatible with " r"the arguments 'ndarray, float'$",
             ):
                 wp.mul(np.array(npv(np_type, a_values)), b_value)
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                fr"Couldn't find a function 'mul' compatible with "
-                fr"the arguments 'tuple, {dtype.__name__}'$",
+                rf"Couldn't find a function 'mul' compatible with " rf"the arguments 'tuple, {dtype.__name__}'$",
             ):
                 wp.mul(a_values, dtype(b_value))
 
             with test.assertRaisesRegex(
                 RuntimeError,
-                fr"Couldn't find a function 'mul' compatible with "
-                fr"the arguments 'tuple, {np_type.__name__}'$",
+                rf"Couldn't find a function 'mul' compatible with " rf"the arguments 'tuple, {np_type.__name__}'$",
             ):
                 wp.mul(a_values, nps(np_type, b_value))
 
 
-def register(parent):
-    class TestBuiltinsResolution(parent):
-        pass
-
-    add_function_test(TestBuiltinsResolution, "test_int_arg_overflow", test_int_arg_overflow)
-    add_function_test(TestBuiltinsResolution, "test_float_arg_precision", test_float_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_int_int_args_overflow", test_int_int_args_overflow)
-    add_function_test(TestBuiltinsResolution, "test_mat22_arg_precision", test_mat22_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat33_arg_precision", test_mat33_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat44_arg_precision", test_mat44_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat22_mat22_args_precision", test_mat22_mat22_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat33_mat33_args_precision", test_mat33_mat33_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat44_mat44_args", test_mat44_mat44_args)
-    add_function_test(TestBuiltinsResolution, "test_mat22_float_args_precision", test_mat22_float_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat33_float_args_precision", test_mat33_float_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_mat44_float_args_precision", test_mat44_float_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec2_arg_precision", test_vec2_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec2_arg_overflow", test_vec2_arg_overflow)
-    add_function_test(TestBuiltinsResolution, "test_vec3_arg_precision", test_vec3_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec3_arg_overflow", test_vec3_arg_overflow)
-    add_function_test(TestBuiltinsResolution, "test_vec4_arg_precision", test_vec4_arg_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec4_arg_overflow", test_vec4_arg_overflow)
-    add_function_test(TestBuiltinsResolution, "test_vec2_vec2_args_precision", test_vec2_vec2_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec2_vec2_args_overflow", test_vec2_vec2_args_overflow)
-    add_function_test(TestBuiltinsResolution, "test_vec3_vec3_args_precision", test_vec3_vec3_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec3_vec3_args_overflow", test_vec3_vec3_args_overflow)
-    add_function_test(TestBuiltinsResolution, "test_vec4_vec4_args_precision", test_vec4_vec4_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec4_vec4_args_overflow", test_vec4_vec4_args_overflow)
-    add_function_test(TestBuiltinsResolution, "test_vec2_float_args_precision", test_vec2_float_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec3_float_args_precision", test_vec3_float_args_precision)
-    add_function_test(TestBuiltinsResolution, "test_vec4_float_args_precision", test_vec4_float_args_precision)
-
-    for dtype in wp.types.int_types:
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_int_arg_support_{dtype.__name__}",
-            test_int_arg_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_int_int_args_support_{dtype.__name__}",
-            test_int_int_args_support,
-            dtype=dtype,
-        )
+class TestBuiltinsResolution(unittest.TestCase):
+    def test_int_arg_overflow(self):
+        value = -1234567890123456789
 
-    for dtype in wp.types.float_types:
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_float_arg_support_{dtype.__name__}",
-            test_float_arg_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_mat_arg_support_{dtype.__name__}",
-            test_mat_arg_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_mat_mat_args_support_{dtype.__name__}",
-            test_mat_mat_args_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_mat_float_args_support_{dtype.__name__}",
-            test_mat_float_args_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_vec_arg_support_{dtype.__name__}",
-            test_vec_arg_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_vec_vec_args_support_{dtype.__name__}",
-            test_vec_vec_args_support,
-            dtype=dtype,
-        )
-        add_function_test(
-            TestBuiltinsResolution,
-            f"test_vec_float_args_support_{dtype.__name__}",
-            test_vec_float_args_support,
-            dtype=dtype,
+        self.assertEqual(wp.invert(wp.int8(value)), 20)
+        self.assertEqual(wp.invert(wp.int16(value)), -32492)
+        self.assertEqual(wp.invert(wp.int32(value)), 2112454932)
+        self.assertEqual(wp.invert(wp.int64(value)), 1234567890123456788)
+
+        self.assertEqual(wp.invert(wp.uint8(value)), 20)
+        self.assertEqual(wp.invert(wp.uint16(value)), 33044)
+        self.assertEqual(wp.invert(wp.uint32(value)), 2112454932)
+        self.assertEqual(wp.invert(wp.uint64(value)), 1234567890123456788)
+
+        self.assertEqual(wp.invert(value), wp.invert(wp.int32(value)))
+
+    def test_float_arg_precision(self):
+        value = 1.23
+        expected = 0.94248880193169748409
+
+        result = wp.sin(wp.float64(value))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.sin(wp.float32(value))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.sin(wp.float16(value))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        self.assertEqual(wp.sin(value), wp.sin(wp.float32(value)))
+
+    def test_int_int_args_overflow(self):
+        value = -1234567890
+
+        self.assertEqual(wp.mul(wp.int8(value), wp.int8(value)), 68)
+        self.assertEqual(wp.mul(wp.int16(value), wp.int16(value)), -3004)
+        self.assertEqual(wp.mul(wp.int32(value), wp.int32(value)), 304084036)
+        self.assertEqual(wp.mul(wp.int64(value), wp.int64(value)), 1524157875019052100)
+
+        self.assertEqual(wp.mul(wp.uint8(value), wp.uint8(value)), 68)
+        self.assertEqual(wp.mul(wp.uint16(value), wp.uint16(value)), 62532)
+        self.assertEqual(wp.mul(wp.uint32(value), wp.uint32(value)), 304084036)
+        self.assertEqual(wp.mul(wp.uint64(value), wp.uint64(value)), 1524157875019052100)
+
+        self.assertEqual(wp.mul(value, value), wp.mul(wp.int32(value), wp.int32(value)))
+
+    def test_mat22_arg_precision(self):
+        values = (1.23, 2.34, 3.45, 4.56)
+        values_2d = (values[0:2], values[2:4])
+        expected = 5.78999999999999914735
+
+        result = wp.trace(wp.mat22d(*values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.trace(wp.mat22f(*values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.trace(wp.mat22h(*values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.trace(values), wp.trace(wp.mat22f(*values)))
+            self.assertEqual(wp.trace(values_2d), wp.trace(wp.mat22f(*values)))
+
+    def test_mat33_arg_precision(self):
+        values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
+        values_2d = (values[0:3], values[3:6], values[6:9])
+        expected = 15.91000000000000014211
+
+        result = wp.trace(wp.mat33d(*values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.trace(wp.mat33f(*values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.trace(wp.mat33h(*values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.trace(values), wp.trace(wp.mat33f(*values)))
+            self.assertEqual(wp.trace(values_2d), wp.trace(wp.mat33f(*values)))
+
+    def test_mat44_arg_precision(self):
+        values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01, 10.12, 11.23, 12.34, 13.45, 14.56, 15.67, 16.78)
+        values_2d = (values[0:4], values[4:8], values[8:12], values[12:16])
+        expected = 36.02000000000000312639
+
+        result = wp.trace(wp.mat44d(*values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.trace(wp.mat44f(*values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.trace(wp.mat44h(*values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.trace(values), wp.trace(wp.mat44f(*values)))
+            self.assertEqual(wp.trace(values_2d), wp.trace(wp.mat44f(*values)))
+
+    def test_mat22_mat22_args_precision(self):
+        a_values = (0.12, 1.23, 0.12, 1.23)
+        a_values_2d = (a_values[0:2], a_values[2:4])
+        b_values = (1.23, 0.12, 1.23, 0.12)
+        b_values_2d = (b_values[0:2], b_values[2:4])
+        expected = 0.59039999999999992486
+
+        result = wp.ddot(wp.mat22d(*a_values), wp.mat22d(*b_values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.ddot(wp.mat22h(*a_values), wp.mat22h(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+            self.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+            self.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+            self.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat22f(*a_values), wp.mat22f(*b_values)))
+
+    def test_mat33_mat33_args_precision(self):
+        a_values = (0.12, 1.23, 2.34, 0.12, 1.23, 2.34, 0.12, 1.23, 2.34)
+        a_values_2d = (a_values[0:3], a_values[3:6], a_values[6:9])
+        b_values = (2.34, 1.23, 0.12, 2.34, 1.23, 0.12, 2.34, 1.23, 0.12)
+        b_values_2d = (b_values[0:3], b_values[3:6], b_values[6:9])
+        expected = 6.22350000000000047606
+
+        result = wp.ddot(wp.mat33d(*a_values), wp.mat33d(*b_values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.ddot(wp.mat33h(*a_values), wp.mat33h(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+            self.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+            self.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+            self.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat33f(*a_values), wp.mat33f(*b_values)))
+
+    def test_mat44_mat44_args(self):
+        a_values = (0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45, 0.12, 1.23, 2.34, 3.45)
+        a_values_2d = (a_values[0:4], a_values[4:8], a_values[8:12], a_values[12:16])
+        b_values = (3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12, 3.45, 2.34, 1.23, 0.12)
+        b_values_2d = (b_values[0:4], b_values[4:8], b_values[8:12], b_values[12:16])
+        expected = 26.33760000000000189857
+
+        result = wp.ddot(wp.mat44d(*a_values), wp.mat44d(*b_values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.ddot(wp.mat44h(*a_values), wp.mat44h(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.ddot(a_values, b_values), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+            self.assertEqual(wp.ddot(a_values_2d, b_values_2d), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+            self.assertEqual(wp.ddot(a_values, b_values_2d), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+            self.assertEqual(wp.ddot(a_values_2d, b_values), wp.ddot(wp.mat44f(*a_values), wp.mat44f(*b_values)))
+
+    def test_mat22_float_args_precision(self):
+        a_values = (1.23, 2.34, 3.45, 4.56)
+        a_values_2d = (a_values[0:2], a_values[2:4])
+        b_value = 0.12
+        expected_00 = 0.14759999999999998122
+        expected_01 = 0.28079999999999999405
+        expected_10 = 0.41399999999999997913
+        expected_11 = 0.54719999999999990870
+
+        result = wp.mul(wp.mat22d(*a_values), wp.float64(b_value))
+        self.assertAlmostEqual(result[0][0], expected_00, places=12)
+        self.assertAlmostEqual(result[0][1], expected_01, places=12)
+        self.assertAlmostEqual(result[1][0], expected_10, places=12)
+        self.assertAlmostEqual(result[1][1], expected_11, places=12)
+
+        result = wp.mul(wp.mat22f(*a_values), wp.float32(b_value))
+        self.assertNotAlmostEqual(result[0][0], expected_00, places=12)
+        self.assertNotAlmostEqual(result[0][1], expected_01, places=12)
+        self.assertNotAlmostEqual(result[1][0], expected_10, places=12)
+        self.assertNotAlmostEqual(result[1][1], expected_11, places=12)
+        self.assertAlmostEqual(result[0][0], expected_00, places=5)
+        self.assertAlmostEqual(result[0][1], expected_01, places=5)
+        self.assertAlmostEqual(result[1][0], expected_10, places=5)
+        self.assertAlmostEqual(result[1][1], expected_11, places=5)
+
+        result = wp.mul(wp.mat22h(*a_values), wp.float16(b_value))
+        self.assertNotAlmostEqual(result[0][0], expected_00, places=5)
+        self.assertNotAlmostEqual(result[0][1], expected_01, places=5)
+        self.assertNotAlmostEqual(result[1][0], expected_10, places=5)
+        self.assertNotAlmostEqual(result[1][1], expected_11, places=5)
+        self.assertAlmostEqual(result[0][0], expected_00, places=1)
+        self.assertAlmostEqual(result[0][1], expected_01, places=1)
+        self.assertAlmostEqual(result[1][0], expected_10, places=1)
+        self.assertAlmostEqual(result[1][1], expected_11, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            # Multiplying a 1-D tuple of length 4 is ambiguous because it could match
+            # either the `vec4f` or `mat22f` overload. As a result, only the 2-D variant
+            # of the tuple is expected to resolve correctly.
+            self.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat22f(*a_values), wp.float32(b_value)))
+
+    def test_mat33_float_args_precision(self):
+        a_values = (1.23, 2.34, 3.45, 4.56, 5.67, 6.78, 7.89, 8.90, 9.01)
+        a_values_2d = (a_values[0:3], a_values[3:6], a_values[6:9])
+        b_value = 0.12
+        expected_00 = 0.14759999999999998122
+        expected_01 = 0.28079999999999999405
+        expected_02 = 0.41399999999999997913
+        expected_10 = 0.54719999999999990870
+        expected_11 = 0.68040000000000000480
+        expected_12 = 0.81359999999999998987
+        expected_20 = 0.94679999999999997495
+        expected_21 = 1.06800000000000006040
+        expected_22 = 1.08119999999999993889
+
+        result = wp.mul(wp.mat33d(*a_values), wp.float64(b_value))
+        self.assertAlmostEqual(result[0][0], expected_00, places=12)
+        self.assertAlmostEqual(result[0][1], expected_01, places=12)
+        self.assertAlmostEqual(result[0][2], expected_02, places=12)
+        self.assertAlmostEqual(result[1][0], expected_10, places=12)
+        self.assertAlmostEqual(result[1][1], expected_11, places=12)
+        self.assertAlmostEqual(result[1][2], expected_12, places=12)
+        self.assertAlmostEqual(result[2][0], expected_20, places=12)
+        self.assertAlmostEqual(result[2][1], expected_21, places=12)
+        self.assertAlmostEqual(result[2][2], expected_22, places=12)
+
+        result = wp.mul(wp.mat33f(*a_values), wp.float32(b_value))
+        self.assertNotAlmostEqual(result[0][0], expected_00, places=12)
+        self.assertNotAlmostEqual(result[0][1], expected_01, places=12)
+        self.assertNotAlmostEqual(result[0][2], expected_02, places=12)
+        self.assertNotAlmostEqual(result[1][0], expected_10, places=12)
+        self.assertNotAlmostEqual(result[1][1], expected_11, places=12)
+        self.assertNotAlmostEqual(result[1][2], expected_12, places=12)
+        self.assertNotAlmostEqual(result[2][0], expected_20, places=12)
+        self.assertNotAlmostEqual(result[2][1], expected_21, places=12)
+        self.assertNotAlmostEqual(result[2][2], expected_22, places=12)
+        self.assertAlmostEqual(result[0][0], expected_00, places=5)
+        self.assertAlmostEqual(result[0][1], expected_01, places=5)
+        self.assertAlmostEqual(result[0][2], expected_02, places=5)
+        self.assertAlmostEqual(result[1][0], expected_10, places=5)
+        self.assertAlmostEqual(result[1][1], expected_11, places=5)
+        self.assertAlmostEqual(result[1][2], expected_12, places=5)
+        self.assertAlmostEqual(result[2][0], expected_20, places=5)
+        self.assertAlmostEqual(result[2][1], expected_21, places=5)
+        self.assertAlmostEqual(result[2][2], expected_22, places=5)
+
+        result = wp.mul(wp.mat33h(*a_values), wp.float16(b_value))
+        self.assertNotAlmostEqual(result[0][0], expected_00, places=5)
+        self.assertNotAlmostEqual(result[0][1], expected_01, places=5)
+        self.assertNotAlmostEqual(result[0][2], expected_02, places=5)
+        self.assertNotAlmostEqual(result[1][0], expected_10, places=5)
+        self.assertNotAlmostEqual(result[1][1], expected_11, places=5)
+        self.assertNotAlmostEqual(result[1][2], expected_12, places=5)
+        self.assertNotAlmostEqual(result[2][0], expected_20, places=5)
+        self.assertNotAlmostEqual(result[2][1], expected_21, places=5)
+        self.assertNotAlmostEqual(result[2][2], expected_22, places=5)
+        self.assertAlmostEqual(result[0][0], expected_00, places=1)
+        self.assertAlmostEqual(result[0][1], expected_01, places=1)
+        self.assertAlmostEqual(result[0][2], expected_02, places=1)
+        self.assertAlmostEqual(result[1][0], expected_10, places=1)
+        self.assertAlmostEqual(result[1][1], expected_11, places=1)
+        self.assertAlmostEqual(result[1][2], expected_12, places=1)
+        self.assertAlmostEqual(result[2][0], expected_20, places=1)
+        self.assertAlmostEqual(result[2][1], expected_21, places=1)
+        self.assertAlmostEqual(result[2][2], expected_22, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.mat33f(*a_values), wp.float32(b_value)))
+            self.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat33f(*a_values), wp.float32(b_value)))
+
+    def test_mat44_float_args_precision(self):
+        a_values = (
+            1.23,
+            2.34,
+            3.45,
+            4.56,
+            5.67,
+            6.78,
+            7.89,
+            8.90,
+            9.01,
+            10.12,
+            11.23,
+            12.34,
+            13.45,
+            14.56,
+            15.67,
+            16.78,
         )
+        a_values_2d = (a_values[0:4], a_values[4:8], a_values[8:12], a_values[12:16])
+        b_value = 0.12
+        expected_00 = 0.14759999999999998122
+        expected_01 = 0.28079999999999999405
+        expected_02 = 0.41399999999999997913
+        expected_03 = 0.54719999999999990870
+        expected_10 = 0.68040000000000000480
+        expected_11 = 0.81359999999999998987
+        expected_12 = 0.94679999999999997495
+        expected_13 = 1.06800000000000006040
+        expected_20 = 1.08119999999999993889
+        expected_21 = 1.21439999999999992397
+        expected_22 = 1.34759999999999990905
+        expected_23 = 1.48079999999999989413
+        expected_30 = 1.61399999999999987921
+        expected_31 = 1.74720000000000008633
+        expected_32 = 1.88039999999999984936
+        expected_33 = 2.01360000000000027853
+
+        result = wp.mul(wp.mat44d(*a_values), wp.float64(b_value))
+        self.assertAlmostEqual(result[0][0], expected_00, places=12)
+        self.assertAlmostEqual(result[0][1], expected_01, places=12)
+        self.assertAlmostEqual(result[0][2], expected_02, places=12)
+        self.assertAlmostEqual(result[0][3], expected_03, places=12)
+        self.assertAlmostEqual(result[1][0], expected_10, places=12)
+        self.assertAlmostEqual(result[1][1], expected_11, places=12)
+        self.assertAlmostEqual(result[1][2], expected_12, places=12)
+        self.assertAlmostEqual(result[1][3], expected_13, places=12)
+        self.assertAlmostEqual(result[2][0], expected_20, places=12)
+        self.assertAlmostEqual(result[2][1], expected_21, places=12)
+        self.assertAlmostEqual(result[2][2], expected_22, places=12)
+        self.assertAlmostEqual(result[2][3], expected_23, places=12)
+        self.assertAlmostEqual(result[3][0], expected_30, places=12)
+        self.assertAlmostEqual(result[3][1], expected_31, places=12)
+        self.assertAlmostEqual(result[3][2], expected_32, places=12)
+        self.assertAlmostEqual(result[3][3], expected_33, places=12)
+
+        result = wp.mul(wp.mat44f(*a_values), wp.float32(b_value))
+        self.assertNotAlmostEqual(result[0][0], expected_00, places=12)
+        self.assertNotAlmostEqual(result[0][1], expected_01, places=12)
+        self.assertNotAlmostEqual(result[0][2], expected_02, places=12)
+        self.assertNotAlmostEqual(result[0][3], expected_03, places=12)
+        self.assertNotAlmostEqual(result[1][0], expected_10, places=12)
+        self.assertNotAlmostEqual(result[1][1], expected_11, places=12)
+        self.assertNotAlmostEqual(result[1][2], expected_12, places=12)
+        self.assertNotAlmostEqual(result[1][3], expected_13, places=12)
+        self.assertNotAlmostEqual(result[2][0], expected_20, places=12)
+        self.assertNotAlmostEqual(result[2][1], expected_21, places=12)
+        self.assertNotAlmostEqual(result[2][2], expected_22, places=12)
+        self.assertNotAlmostEqual(result[2][3], expected_23, places=12)
+        self.assertNotAlmostEqual(result[3][0], expected_30, places=12)
+        self.assertNotAlmostEqual(result[3][1], expected_31, places=12)
+        self.assertNotAlmostEqual(result[3][2], expected_32, places=12)
+        self.assertNotAlmostEqual(result[3][3], expected_33, places=12)
+        self.assertAlmostEqual(result[0][0], expected_00, places=5)
+        self.assertAlmostEqual(result[0][1], expected_01, places=5)
+        self.assertAlmostEqual(result[0][2], expected_02, places=5)
+        self.assertAlmostEqual(result[0][3], expected_03, places=5)
+        self.assertAlmostEqual(result[1][0], expected_10, places=5)
+        self.assertAlmostEqual(result[1][1], expected_11, places=5)
+        self.assertAlmostEqual(result[1][2], expected_12, places=5)
+        self.assertAlmostEqual(result[1][3], expected_13, places=5)
+        self.assertAlmostEqual(result[2][0], expected_20, places=5)
+        self.assertAlmostEqual(result[2][1], expected_21, places=5)
+        self.assertAlmostEqual(result[2][2], expected_22, places=5)
+        self.assertAlmostEqual(result[2][3], expected_23, places=5)
+        self.assertAlmostEqual(result[3][0], expected_30, places=5)
+        self.assertAlmostEqual(result[3][1], expected_31, places=5)
+        self.assertAlmostEqual(result[3][2], expected_32, places=5)
+        self.assertAlmostEqual(result[3][3], expected_33, places=5)
+
+        result = wp.mul(wp.mat44h(*a_values), wp.float16(b_value))
+        self.assertNotAlmostEqual(result[0][0], expected_00, places=5)
+        self.assertNotAlmostEqual(result[0][1], expected_01, places=5)
+        self.assertNotAlmostEqual(result[0][2], expected_02, places=5)
+        self.assertNotAlmostEqual(result[0][3], expected_03, places=5)
+        self.assertNotAlmostEqual(result[1][0], expected_10, places=5)
+        self.assertNotAlmostEqual(result[1][1], expected_11, places=5)
+        self.assertNotAlmostEqual(result[1][2], expected_12, places=5)
+        self.assertNotAlmostEqual(result[1][3], expected_13, places=5)
+        self.assertNotAlmostEqual(result[2][0], expected_20, places=5)
+        self.assertNotAlmostEqual(result[2][1], expected_21, places=5)
+        self.assertNotAlmostEqual(result[2][2], expected_22, places=5)
+        self.assertNotAlmostEqual(result[2][3], expected_23, places=5)
+        self.assertNotAlmostEqual(result[3][0], expected_30, places=5)
+        self.assertNotAlmostEqual(result[3][1], expected_31, places=5)
+        self.assertNotAlmostEqual(result[3][2], expected_32, places=5)
+        self.assertNotAlmostEqual(result[3][3], expected_33, places=5)
+        self.assertAlmostEqual(result[0][0], expected_00, places=1)
+        self.assertAlmostEqual(result[0][1], expected_01, places=1)
+        self.assertAlmostEqual(result[0][2], expected_02, places=1)
+        self.assertAlmostEqual(result[0][3], expected_03, places=1)
+        self.assertAlmostEqual(result[1][0], expected_10, places=1)
+        self.assertAlmostEqual(result[1][1], expected_11, places=1)
+        self.assertAlmostEqual(result[1][2], expected_12, places=1)
+        self.assertAlmostEqual(result[1][3], expected_13, places=1)
+        self.assertAlmostEqual(result[2][0], expected_20, places=1)
+        self.assertAlmostEqual(result[2][1], expected_21, places=1)
+        self.assertAlmostEqual(result[2][2], expected_22, places=1)
+        self.assertAlmostEqual(result[2][3], expected_23, places=1)
+        self.assertAlmostEqual(result[3][0], expected_30, places=1)
+        self.assertAlmostEqual(result[3][1], expected_31, places=1)
+        self.assertAlmostEqual(result[3][2], expected_32, places=1)
+        self.assertAlmostEqual(result[3][3], expected_33, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.mat44f(*a_values), wp.float32(b_value)))
+            self.assertEqual(wp.mul(a_values_2d, b_value), wp.mul(wp.mat44f(*a_values), wp.float32(b_value)))
+
+    def test_vec2_arg_precision(self):
+        values = (1.23, 2.34)
+        expected = 2.64357712200722438922
+
+        result = wp.length(wp.vec2d(*values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.length(wp.vec2f(*values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.length(wp.vec2h(*values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.length(values), wp.length(wp.vec2f(*values)))
+
+    def test_vec2_arg_overflow(self):
+        values = (-1234567890, -1234567890)
+
+        self.assertEqual(wp.length_sq(wp.vec2b(*values)), -120)
+        self.assertEqual(wp.length_sq(wp.vec2s(*values)), -6008)
+        self.assertEqual(wp.length_sq(wp.vec2i(*values)), 608168072)
+        self.assertEqual(wp.length_sq(wp.vec2l(*values)), 3048315750038104200)
+
+        self.assertEqual(wp.length_sq(wp.vec2ub(*values)), 136)
+        self.assertEqual(wp.length_sq(wp.vec2us(*values)), 59528)
+        self.assertEqual(wp.length_sq(wp.vec2ui(*values)), 608168072)
+        self.assertEqual(wp.length_sq(wp.vec2ul(*values)), 3048315750038104200)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec2i(*values)))
+
+    def test_vec3_arg_precision(self):
+        values = (1.23, 2.34, 3.45)
+        expected = 4.34637780226247727455
+
+        result = wp.length(wp.vec3d(*values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.length(wp.vec3f(*values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.length(wp.vec3h(*values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.length(values), wp.length(wp.vec3f(*values)))
+
+    def test_vec3_arg_overflow(self):
+        values = (-1234567890, -1234567890, -1234567890)
+
+        self.assertEqual(wp.length_sq(wp.vec3b(*values)), -52)
+        self.assertEqual(wp.length_sq(wp.vec3s(*values)), -9012)
+        self.assertEqual(wp.length_sq(wp.vec3i(*values)), 912252108)
+        self.assertEqual(wp.length_sq(wp.vec3l(*values)), 4572473625057156300)
+
+        self.assertEqual(wp.length_sq(wp.vec3ub(*values)), 204)
+        self.assertEqual(wp.length_sq(wp.vec3us(*values)), 56524)
+        self.assertEqual(wp.length_sq(wp.vec3ui(*values)), 912252108)
+        self.assertEqual(wp.length_sq(wp.vec3ul(*values)), 4572473625057156300)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec3i(*values)))
+
+    def test_vec4_arg_precision(self):
+        values = (1.23, 2.34, 3.45, 4.56)
+        expected = 6.29957141399317777086
+
+        result = wp.length(wp.vec4d(*values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.length(wp.vec4f(*values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.length(wp.vec4h(*values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.length(values), wp.length(wp.vec4f(*values)))
+
+    def test_vec4_arg_overflow(self):
+        values = (-1234567890, -1234567890, -1234567890, -1234567890)
+
+        self.assertEqual(wp.length_sq(wp.vec4b(*values)), 16)
+        self.assertEqual(wp.length_sq(wp.vec4s(*values)), -12016)
+        self.assertEqual(wp.length_sq(wp.vec4i(*values)), 1216336144)
+        self.assertEqual(wp.length_sq(wp.vec4l(*values)), 6096631500076208400)
+
+        self.assertEqual(wp.length_sq(wp.vec4ub(*values)), 16)
+        self.assertEqual(wp.length_sq(wp.vec4us(*values)), 53520)
+        self.assertEqual(wp.length_sq(wp.vec4ui(*values)), 1216336144)
+        self.assertEqual(wp.length_sq(wp.vec4ul(*values)), 6096631500076208400)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.length_sq(values), wp.length_sq(wp.vec4i(*values)))
+
+    def test_vec2_vec2_args_precision(self):
+        a_values = (1.23, 2.34)
+        b_values = (3.45, 4.56)
+        expected = 14.91389999999999815827
+
+        result = wp.dot(wp.vec2d(*a_values), wp.vec2d(*b_values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.dot(wp.vec2f(*a_values), wp.vec2f(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.dot(wp.vec2h(*a_values), wp.vec2h(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec2f(*a_values), wp.vec2f(*b_values)))
+
+    def test_vec2_vec2_args_overflow(self):
+        values = (-1234567890, -1234567890)
+
+        self.assertEqual(wp.dot(wp.vec2b(*values), wp.vec2b(*values)), -120)
+        self.assertEqual(wp.dot(wp.vec2s(*values), wp.vec2s(*values)), -6008)
+        self.assertEqual(wp.dot(wp.vec2i(*values), wp.vec2i(*values)), 608168072)
+        self.assertEqual(wp.dot(wp.vec2l(*values), wp.vec2l(*values)), 3048315750038104200)
+
+        self.assertEqual(wp.dot(wp.vec2ub(*values), wp.vec2ub(*values)), 136)
+        self.assertEqual(wp.dot(wp.vec2us(*values), wp.vec2us(*values)), 59528)
+        self.assertEqual(wp.dot(wp.vec2ui(*values), wp.vec2ui(*values)), 608168072)
+        self.assertEqual(wp.dot(wp.vec2ul(*values), wp.vec2ul(*values)), 3048315750038104200)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.dot(values, values), wp.dot(wp.vec2i(*values), wp.vec2i(*values)))
+
+    def test_vec3_vec3_args_precision(self):
+        a_values = (1.23, 2.34, 3.45)
+        b_values = (4.56, 5.67, 6.78)
+        expected = 42.26760000000000161435
+
+        result = wp.dot(wp.vec3d(*a_values), wp.vec3d(*b_values))
+        self.assertAlmostEqual(result, expected, places=12)
 
-    return TestBuiltinsResolution
+        result = wp.dot(wp.vec3f(*a_values), wp.vec3f(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.dot(wp.vec3h(*a_values), wp.vec3h(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec3f(*a_values), wp.vec3f(*b_values)))
+
+    def test_vec3_vec3_args_overflow(self):
+        values = (-1234567890, -1234567890, -1234567890)
+
+        self.assertEqual(wp.dot(wp.vec3b(*values), wp.vec3b(*values)), -52)
+        self.assertEqual(wp.dot(wp.vec3s(*values), wp.vec3s(*values)), -9012)
+        self.assertEqual(wp.dot(wp.vec3i(*values), wp.vec3i(*values)), 912252108)
+        self.assertEqual(wp.dot(wp.vec3l(*values), wp.vec3l(*values)), 4572473625057156300)
+
+        self.assertEqual(wp.dot(wp.vec3ub(*values), wp.vec3ub(*values)), 204)
+        self.assertEqual(wp.dot(wp.vec3us(*values), wp.vec3us(*values)), 56524)
+        self.assertEqual(wp.dot(wp.vec3ui(*values), wp.vec3ui(*values)), 912252108)
+        self.assertEqual(wp.dot(wp.vec3ul(*values), wp.vec3ul(*values)), 4572473625057156300)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.dot(values, values), wp.dot(wp.vec3i(*values), wp.vec3i(*values)))
+
+    def test_vec4_vec4_args_precision(self):
+        a_values = (1.23, 2.34, 3.45, 4.56)
+        b_values = (5.67, 6.78, 7.89, 8.90)
+        expected = 90.64379999999999881766
+
+        result = wp.dot(wp.vec4d(*a_values), wp.vec4d(*b_values))
+        self.assertAlmostEqual(result, expected, places=12)
+
+        result = wp.dot(wp.vec4f(*a_values), wp.vec4f(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=12)
+        self.assertAlmostEqual(result, expected, places=5)
+
+        result = wp.dot(wp.vec4h(*a_values), wp.vec4h(*b_values))
+        self.assertNotAlmostEqual(result, expected, places=5)
+        self.assertAlmostEqual(result, expected, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.dot(a_values, b_values), wp.dot(wp.vec4f(*a_values), wp.vec4f(*b_values)))
+
+    def test_vec4_vec4_args_overflow(self):
+        values = (-1234567890, -1234567890, -1234567890, -1234567890)
+
+        self.assertEqual(wp.dot(wp.vec4b(*values), wp.vec4b(*values)), 16)
+        self.assertEqual(wp.dot(wp.vec4s(*values), wp.vec4s(*values)), -12016)
+        self.assertEqual(wp.dot(wp.vec4i(*values), wp.vec4i(*values)), 1216336144)
+        self.assertEqual(wp.dot(wp.vec4l(*values), wp.vec4l(*values)), 6096631500076208400)
+
+        self.assertEqual(wp.dot(wp.vec4ub(*values), wp.vec4ub(*values)), 16)
+        self.assertEqual(wp.dot(wp.vec4us(*values), wp.vec4us(*values)), 53520)
+        self.assertEqual(wp.dot(wp.vec4ui(*values), wp.vec4ui(*values)), 1216336144)
+        self.assertEqual(wp.dot(wp.vec4ul(*values), wp.vec4ul(*values)), 6096631500076208400)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.dot(values, values), wp.dot(wp.vec4i(*values), wp.vec4i(*values)))
+
+    def test_vec2_float_args_precision(self):
+        a_values = (1.23, 2.34)
+        b_value = 3.45
+        expected_x = 4.24350000000000004974
+        expected_y = 8.07300000000000039790
+
+        result = wp.mul(wp.vec2d(*a_values), wp.float64(b_value))
+        self.assertAlmostEqual(result[0], expected_x, places=12)
+        self.assertAlmostEqual(result[1], expected_y, places=12)
+
+        result = wp.mul(wp.vec2f(*a_values), wp.float32(b_value))
+        self.assertNotAlmostEqual(result[0], expected_x, places=12)
+        self.assertNotAlmostEqual(result[1], expected_y, places=12)
+        self.assertAlmostEqual(result[0], expected_x, places=5)
+        self.assertAlmostEqual(result[1], expected_y, places=5)
+
+        result = wp.mul(wp.vec2h(*a_values), wp.float16(b_value))
+        self.assertNotAlmostEqual(result[0], expected_x, places=5)
+        self.assertNotAlmostEqual(result[1], expected_y, places=5)
+        self.assertAlmostEqual(result[0], expected_x, places=1)
+        self.assertAlmostEqual(result[1], expected_y, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec2f(*a_values), wp.float32(b_value)))
+
+    def test_vec3_float_args_precision(self):
+        a_values = (1.23, 2.34, 3.45)
+        b_value = 4.56
+        expected_x = 5.60879999999999956373
+        expected_y = 10.67039999999999899671
+        expected_z = 15.73199999999999931788
+
+        result = wp.mul(wp.vec3d(*a_values), wp.float64(b_value))
+        self.assertAlmostEqual(result[0], expected_x, places=12)
+        self.assertAlmostEqual(result[1], expected_y, places=12)
+        self.assertAlmostEqual(result[2], expected_z, places=12)
+
+        result = wp.mul(wp.vec3f(*a_values), wp.float32(b_value))
+        self.assertNotAlmostEqual(result[0], expected_x, places=12)
+        self.assertNotAlmostEqual(result[1], expected_y, places=12)
+        self.assertNotAlmostEqual(result[2], expected_z, places=12)
+        self.assertAlmostEqual(result[0], expected_x, places=5)
+        self.assertAlmostEqual(result[1], expected_y, places=5)
+        self.assertAlmostEqual(result[2], expected_z, places=5)
+
+        result = wp.mul(wp.vec3h(*a_values), wp.float16(b_value))
+        self.assertNotAlmostEqual(result[0], expected_x, places=5)
+        self.assertNotAlmostEqual(result[1], expected_y, places=5)
+        self.assertNotAlmostEqual(result[2], expected_z, places=5)
+        self.assertAlmostEqual(result[0], expected_x, places=1)
+        self.assertAlmostEqual(result[1], expected_y, places=1)
+        self.assertAlmostEqual(result[2], expected_z, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec3f(*a_values), wp.float32(b_value)))
+
+    def test_vec4_float_args_precision(self):
+        a_values = (1.23, 2.34, 3.45, 4.56)
+        b_value = 5.67
+        expected_x = 6.97409999999999996589
+        expected_y = 13.26779999999999937188
+        expected_z = 19.56150000000000233058
+        expected_w = 25.85519999999999640750
+
+        result = wp.mul(wp.vec4d(*a_values), wp.float64(b_value))
+        self.assertAlmostEqual(result[0], expected_x, places=12)
+        self.assertAlmostEqual(result[1], expected_y, places=12)
+        self.assertAlmostEqual(result[2], expected_z, places=12)
+        self.assertAlmostEqual(result[3], expected_w, places=12)
+
+        result = wp.mul(wp.vec4f(*a_values), wp.float32(b_value))
+        self.assertNotAlmostEqual(result[0], expected_x, places=12)
+        self.assertNotAlmostEqual(result[1], expected_y, places=12)
+        self.assertNotAlmostEqual(result[2], expected_z, places=12)
+        self.assertNotAlmostEqual(result[3], expected_w, places=12)
+        self.assertAlmostEqual(result[0], expected_x, places=5)
+        self.assertAlmostEqual(result[1], expected_y, places=5)
+        self.assertAlmostEqual(result[2], expected_z, places=5)
+        self.assertAlmostEqual(result[3], expected_w, places=5)
+
+        result = wp.mul(wp.vec4h(*a_values), wp.float16(b_value))
+        self.assertNotAlmostEqual(result[0], expected_x, places=5)
+        self.assertNotAlmostEqual(result[1], expected_y, places=5)
+        self.assertNotAlmostEqual(result[2], expected_z, places=5)
+        self.assertNotAlmostEqual(result[3], expected_w, places=5)
+        self.assertAlmostEqual(result[0], expected_x, places=1)
+        self.assertAlmostEqual(result[1], expected_y, places=1)
+        self.assertAlmostEqual(result[2], expected_z, places=1)
+        self.assertAlmostEqual(result[3], expected_w, places=1)
+
+        with open(os.devnull, "w") as f, contextlib.redirect_stdout(f):
+            self.assertEqual(wp.mul(a_values, b_value), wp.mul(wp.vec4f(*a_values), wp.float32(b_value)))
+
+
+for dtype in wp.types.int_types:
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_int_arg_support_{dtype.__name__}",
+        test_int_arg_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_int_int_args_support_{dtype.__name__}",
+        test_int_int_args_support,
+        dtype=dtype,
+    )
+
+for dtype in wp.types.float_types:
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_float_arg_support_{dtype.__name__}",
+        test_float_arg_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_mat_arg_support_{dtype.__name__}",
+        test_mat_arg_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_mat_mat_args_support_{dtype.__name__}",
+        test_mat_mat_args_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_mat_float_args_support_{dtype.__name__}",
+        test_mat_float_args_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_vec_arg_support_{dtype.__name__}",
+        test_vec_arg_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_vec_vec_args_support_{dtype.__name__}",
+        test_vec_vec_args_support,
+        dtype=dtype,
+    )
+    add_function_test(
+        TestBuiltinsResolution,
+        f"test_vec_float_args_support_{dtype.__name__}",
+        test_vec_float_args_support,
+        dtype=dtype,
+    )
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_bvh.py b/warp/tests/test_bvh.py
index a67e56dfb..d749b9a37 100644
--- a/warp/tests/test_bvh.py
+++ b/warp/tests/test_bvh.py
@@ -10,8 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -118,8 +117,8 @@ def test_bvh(test, type, device):
         if test_case == 0:
             lowers = rng.random(size=(num_bounds, 3)) * 5.0
             uppers = lowers + rng.random(size=(num_bounds, 3)) * 5.0
-            wp.copy(device_lowers, wp.array(lowers, dtype=wp.vec3))
-            wp.copy(device_uppers, wp.array(uppers, dtype=wp.vec3))
+            wp.copy(device_lowers, wp.array(lowers, dtype=wp.vec3, device=device))
+            wp.copy(device_uppers, wp.array(uppers, dtype=wp.vec3, device=device))
             bvh.refit()
             bounds_intersected.zero_()
 
@@ -132,19 +131,17 @@ def test_bvh_query_ray(test, device):
     test_bvh(test, "ray", device)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices(mode="basic")  # TODO: Fix the illegal address error when two CUDA devices are used
+
 
-    class TestBvh(parent):
-        pass
+class TestBvh(unittest.TestCase):
+    pass
 
-    add_function_test(TestBvh, "test_bvh_aabb", test_bvh_query_aabb, devices=devices)
-    add_function_test(TestBvh, "test_bvh_ray", test_bvh_query_ray, devices=devices)
 
-    return TestBvh
+add_function_test(TestBvh, "test_bvh_aabb", test_bvh_query_aabb, devices=devices)
+add_function_test(TestBvh, "test_bvh_ray", test_bvh_query_ray, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_closest_point_edge_edge.py b/warp/tests/test_closest_point_edge_edge.py
index 8762421b3..0c0f4a415 100644
--- a/warp/tests/test_closest_point_edge_edge.py
+++ b/warp/tests/test_closest_point_edge_edge.py
@@ -5,12 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import numpy as np
-
 import unittest
 
+import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 epsilon = 0.00001
@@ -171,62 +171,58 @@ def test_edge_edge_perpendicular_s0_t1(test, device):
     test.assertAlmostEqual(st0[1], 1.0)  # t value
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestClosestPointEdgeEdgeMethods(parent):
-        pass
-
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods,
-        "test_edge_edge_middle_crossing",
-        test_edge_edge_middle_crossing,
-        devices=devices,
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods, "test_edge_edge_parallel_s1_t0", test_edge_edge_parallel_s1_t0, devices=devices
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods, "test_edge_edge_parallel_s0_t1", test_edge_edge_parallel_s0_t1, devices=devices
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods,
-        "test_edge_edge_both_degenerate_case",
-        test_edge_edge_both_degenerate_case,
-        devices=devices,
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods,
-        "test_edge_edge_degenerate_first_edge",
-        test_edge_edge_degenerate_first_edge,
-        devices=devices,
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods,
-        "test_edge_edge_degenerate_second_edge",
-        test_edge_edge_degenerate_second_edge,
-        devices=devices,
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods, "test_edge_edge_parallel", test_edge_edge_parallel, devices=devices
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods,
-        "test_edge_edge_perpendicular_s1_t0",
-        test_edge_edge_perpendicular_s1_t0,
-        devices=devices,
-    )
-    add_function_test(
-        TestClosestPointEdgeEdgeMethods,
-        "test_edge_edge_perpendicular_s0_t1",
-        test_edge_edge_perpendicular_s0_t1,
-        devices=devices,
-    )
-
-    return TestClosestPointEdgeEdgeMethods
+devices = get_test_devices()
+
+
+class TestClosestPointEdgeEdgeMethods(unittest.TestCase):
+    pass
+
+
+add_function_test(
+    TestClosestPointEdgeEdgeMethods,
+    "test_edge_edge_middle_crossing",
+    test_edge_edge_middle_crossing,
+    devices=devices,
+)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods, "test_edge_edge_parallel_s1_t0", test_edge_edge_parallel_s1_t0, devices=devices
+)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods, "test_edge_edge_parallel_s0_t1", test_edge_edge_parallel_s0_t1, devices=devices
+)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods,
+    "test_edge_edge_both_degenerate_case",
+    test_edge_edge_both_degenerate_case,
+    devices=devices,
+)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods,
+    "test_edge_edge_degenerate_first_edge",
+    test_edge_edge_degenerate_first_edge,
+    devices=devices,
+)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods,
+    "test_edge_edge_degenerate_second_edge",
+    test_edge_edge_degenerate_second_edge,
+    devices=devices,
+)
+add_function_test(TestClosestPointEdgeEdgeMethods, "test_edge_edge_parallel", test_edge_edge_parallel, devices=devices)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods,
+    "test_edge_edge_perpendicular_s1_t0",
+    test_edge_edge_perpendicular_s1_t0,
+    devices=devices,
+)
+add_function_test(
+    TestClosestPointEdgeEdgeMethods,
+    "test_edge_edge_perpendicular_s0_t1",
+    test_edge_edge_perpendicular_s0_t1,
+    devices=devices,
+)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_codegen.py b/warp/tests/test_codegen.py
index a078e5e76..958e83c22 100644
--- a/warp/tests/test_codegen.py
+++ b/warp/tests/test_codegen.py
@@ -5,12 +5,11 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
 import sys
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 # wp.config.mode = "debug"
 
@@ -346,7 +345,7 @@ def test_range_expression():
 
 def test_unresolved_func(test, device):
     # kernel with unresolved function must be in a separate module, otherwise the current module would fail to load
-    from warp.tests.test_unresolved_func import unresolved_func_kernel
+    from warp.tests.aux_test_unresolved_func import unresolved_func_kernel
 
     # ensure that an appropriate exception is raised when the bad module gets loaded
     with test.assertRaisesRegex(RuntimeError, "Could not find function wp.missing_func"):
@@ -354,13 +353,13 @@ def test_unresolved_func(test, device):
 
     # remove all references to the bad module so that subsequent calls to wp.force_load()
     # won't try to load it unless we explicitly re-import it again
-    del wp.context.user_modules["warp.tests.test_unresolved_func"]
-    del sys.modules["warp.tests.test_unresolved_func"]
+    del wp.context.user_modules["warp.tests.aux_test_unresolved_func"]
+    del sys.modules["warp.tests.aux_test_unresolved_func"]
 
 
 def test_unresolved_symbol(test, device):
     # kernel with unresolved symbol must be in a separate module, otherwise the current module would fail to load
-    from warp.tests.test_unresolved_symbol import unresolved_symbol_kernel
+    from warp.tests.aux_test_unresolved_symbol import unresolved_symbol_kernel
 
     # ensure that an appropriate exception is raised when the bad module gets loaded
     with test.assertRaisesRegex(KeyError, "Referencing undefined symbol: missing_symbol"):
@@ -368,143 +367,140 @@ def test_unresolved_symbol(test, device):
 
     # remove all references to the bad module so that subsequent calls to wp.force_load()
     # won't try to load it unless we explicitly re-import it again
-    del wp.context.user_modules["warp.tests.test_unresolved_symbol"]
-    del sys.modules["warp.tests.test_unresolved_symbol"]
-
-
-def register(parent):
-    class TestCodeGen(parent):
-        pass
-
-    devices = get_test_devices()
-
-    add_kernel_test(TestCodeGen, name="test_inplace", kernel=test_inplace, dim=1, devices=devices)
-    add_kernel_test(TestCodeGen, name="test_rename", kernel=test_rename, dim=1, devices=devices)
-    add_kernel_test(TestCodeGen, name="test_constant", kernel=test_constant, inputs=[1.0], dim=1, devices=devices)
-    add_kernel_test(
-        TestCodeGen, name="test_dynamic_for_rename", kernel=test_dynamic_for_rename, inputs=[10], dim=1, devices=devices
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_dynamic_for_inplace",
-        kernel=test_dynamic_for_inplace,
-        inputs=[10],
-        dim=1,
-        devices=devices,
-    )
-    add_kernel_test(TestCodeGen, name="test_reassign", kernel=test_reassign, dim=1, devices=devices)
-    add_kernel_test(
-        TestCodeGen, name="test_dynamic_reassign", kernel=test_dynamic_reassign, inputs=[2], dim=1, devices=devices
-    )
-
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_forward",
-        kernel=test_range_dynamic,
-        dim=1,
-        inputs=[0, 4, 1],
-        expect=[0, 1, 2, 3],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_reverse",
-        kernel=test_range_dynamic,
-        dim=1,
-        inputs=[4, 0, -1],
-        expect=[4, 3, 2, 1],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_foward_step",
-        kernel=test_range_dynamic,
-        dim=1,
-        inputs=[0, 8, 2],
-        expect=[0, 2, 4, 6],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_reverse_step",
-        kernel=test_range_dynamic,
-        dim=1,
-        inputs=[8, 0, -2],
-        expect=[8, 6, 4, 2],
-        devices=devices,
-    )
-
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_static_sum",
-        kernel=test_range_static_sum,
-        dim=1,
-        expect=[10, 10, 10],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_sum",
-        kernel=test_range_dynamic_sum,
-        dim=1,
-        inputs=[0, 10, 2],
-        expect=[10, 10, 10, 10],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_sum_zero",
-        kernel=test_range_dynamic_sum,
-        dim=1,
-        inputs=[0, 0, 1],
-        expect=[0, 0, 0, 0],
-        devices=devices,
-    )
-    add_kernel_test(TestCodeGen, name="test_range_constant", kernel=test_range_constant, dim=1, devices=devices)
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_constant_dynamic_nested",
-        kernel=test_range_constant_dynamic_nested,
-        dim=1,
-        inputs=[10],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_dynamic_nested",
-        kernel=test_range_dynamic_nested,
-        dim=1,
-        inputs=[4],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCodeGen,
-        name="test_range_expression",
-        kernel=test_range_expression,
-        dim=1,
-        devices=devices,
-    )
-
-    add_kernel_test(TestCodeGen, name="test_while_zero", kernel=test_while, dim=1, inputs=[0], devices=devices)
-    add_kernel_test(TestCodeGen, name="test_while_positive", kernel=test_while, dim=1, inputs=[16], devices=devices)
-    add_kernel_test(TestCodeGen, name="test_pass", kernel=test_pass, dim=1, inputs=[16], devices=devices)
-
-    add_kernel_test(TestCodeGen, name="test_break", kernel=test_break, dim=1, inputs=[10], devices=devices)
-    add_kernel_test(TestCodeGen, name="test_break_early", kernel=test_break_early, dim=1, inputs=[10], devices=devices)
-    add_kernel_test(TestCodeGen, name="test_break_unroll", kernel=test_break_unroll, dim=1, devices=devices)
-    add_kernel_test(
-        TestCodeGen, name="test_break_multiple", kernel=test_break_multiple, dim=1, inputs=[10], devices=devices
-    )
-    add_kernel_test(TestCodeGen, name="test_continue", kernel=test_continue, dim=1, inputs=[10], devices=devices)
-    add_kernel_test(TestCodeGen, name="test_continue_unroll", kernel=test_continue_unroll, dim=1, devices=devices)
-
-    add_function_test(TestCodeGen, func=test_unresolved_func, name="test_unresolved_func", devices=devices)
-    add_function_test(TestCodeGen, func=test_unresolved_symbol, name="test_unresolved_symbol", devices=devices)
-
-    return TestCodeGen
+    del wp.context.user_modules["warp.tests.aux_test_unresolved_symbol"]
+    del sys.modules["warp.tests.aux_test_unresolved_symbol"]
+
+
+class TestCodeGen(unittest.TestCase):
+    pass
+
+
+devices = get_test_devices()
+
+add_kernel_test(TestCodeGen, name="test_inplace", kernel=test_inplace, dim=1, devices=devices)
+add_kernel_test(TestCodeGen, name="test_rename", kernel=test_rename, dim=1, devices=devices)
+add_kernel_test(TestCodeGen, name="test_constant", kernel=test_constant, inputs=[1.0], dim=1, devices=devices)
+add_kernel_test(
+    TestCodeGen, name="test_dynamic_for_rename", kernel=test_dynamic_for_rename, inputs=[10], dim=1, devices=devices
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_dynamic_for_inplace",
+    kernel=test_dynamic_for_inplace,
+    inputs=[10],
+    dim=1,
+    devices=devices,
+)
+add_kernel_test(TestCodeGen, name="test_reassign", kernel=test_reassign, dim=1, devices=devices)
+add_kernel_test(
+    TestCodeGen, name="test_dynamic_reassign", kernel=test_dynamic_reassign, inputs=[2], dim=1, devices=devices
+)
+
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_forward",
+    kernel=test_range_dynamic,
+    dim=1,
+    inputs=[0, 4, 1],
+    expect=[0, 1, 2, 3],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_reverse",
+    kernel=test_range_dynamic,
+    dim=1,
+    inputs=[4, 0, -1],
+    expect=[4, 3, 2, 1],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_forward_step",
+    kernel=test_range_dynamic,
+    dim=1,
+    inputs=[0, 8, 2],
+    expect=[0, 2, 4, 6],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_reverse_step",
+    kernel=test_range_dynamic,
+    dim=1,
+    inputs=[8, 0, -2],
+    expect=[8, 6, 4, 2],
+    devices=devices,
+)
+
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_static_sum",
+    kernel=test_range_static_sum,
+    dim=1,
+    expect=[10, 10, 10],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_sum",
+    kernel=test_range_dynamic_sum,
+    dim=1,
+    inputs=[0, 10, 2],
+    expect=[10, 10, 10, 10],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_sum_zero",
+    kernel=test_range_dynamic_sum,
+    dim=1,
+    inputs=[0, 0, 1],
+    expect=[0, 0, 0, 0],
+    devices=devices,
+)
+add_kernel_test(TestCodeGen, name="test_range_constant", kernel=test_range_constant, dim=1, devices=devices)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_constant_dynamic_nested",
+    kernel=test_range_constant_dynamic_nested,
+    dim=1,
+    inputs=[10],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_dynamic_nested",
+    kernel=test_range_dynamic_nested,
+    dim=1,
+    inputs=[4],
+    devices=devices,
+)
+add_kernel_test(
+    TestCodeGen,
+    name="test_range_expression",
+    kernel=test_range_expression,
+    dim=1,
+    devices=devices,
+)
+
+add_kernel_test(TestCodeGen, name="test_while_zero", kernel=test_while, dim=1, inputs=[0], devices=devices)
+add_kernel_test(TestCodeGen, name="test_while_positive", kernel=test_while, dim=1, inputs=[16], devices=devices)
+add_kernel_test(TestCodeGen, name="test_pass", kernel=test_pass, dim=1, inputs=[16], devices=devices)
+
+add_kernel_test(TestCodeGen, name="test_break", kernel=test_break, dim=1, inputs=[10], devices=devices)
+add_kernel_test(TestCodeGen, name="test_break_early", kernel=test_break_early, dim=1, inputs=[10], devices=devices)
+add_kernel_test(TestCodeGen, name="test_break_unroll", kernel=test_break_unroll, dim=1, devices=devices)
+add_kernel_test(
+    TestCodeGen, name="test_break_multiple", kernel=test_break_multiple, dim=1, inputs=[10], devices=devices
+)
+add_kernel_test(TestCodeGen, name="test_continue", kernel=test_continue, dim=1, inputs=[10], devices=devices)
+add_kernel_test(TestCodeGen, name="test_continue_unroll", kernel=test_continue_unroll, dim=1, devices=devices)
+
+add_function_test(TestCodeGen, func=test_unresolved_func, name="test_unresolved_func", devices=devices)
+add_function_test(TestCodeGen, func=test_unresolved_symbol, name="test_unresolved_symbol", devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_compile_consts.py b/warp/tests/test_compile_consts.py
index a98c0dfe6..d7e2ec400 100644
--- a/warp/tests/test_compile_consts.py
+++ b/warp/tests/test_compile_consts.py
@@ -5,14 +5,11 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import os
-import sys
-
 import unittest
-from warp.tests.test_base import *
-import warp.tests.test_compile_consts_dummy
 
 import warp as wp
+import warp.tests.aux_test_compile_consts_dummy
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -40,7 +37,7 @@ def test_constants_bool():
 @wp.kernel
 def test_constants_int(a: int):
     if Foobar.ONE > 0:
-        a = 123 + Foobar.TWO + warp.tests.test_compile_consts_dummy.MINUS_ONE
+        a = 123 + Foobar.TWO + warp.tests.aux_test_compile_consts_dummy.MINUS_ONE
     else:
         a = 456 + LOCAL_ONE
     expect_eq(a, 124)
@@ -82,26 +79,23 @@ def closure_kernel_fn(expected: int):
     wp.launch(two_closure, dim=(1), inputs=[2], device=device)
 
 
-def register(parent):
-    class TestConstants(parent):
-        pass
+class TestConstants(unittest.TestCase):
+    pass
 
-    a = 0
-    x = 0.0
 
-    devices = get_test_devices()
+a = 0
+x = 0.0
 
-    add_kernel_test(TestConstants, test_constants_bool, dim=1, inputs=[], devices=devices)
-    add_kernel_test(TestConstants, test_constants_int, dim=1, inputs=[a], devices=devices)
-    add_kernel_test(TestConstants, test_constants_float, dim=1, inputs=[x], devices=devices)
+devices = get_test_devices()
 
-    add_function_test(TestConstants, "test_constant_math", test_constant_math, devices=devices)
-    add_function_test(TestConstants, "test_constant_closure_capture", test_constant_closure_capture, devices=devices)
+add_kernel_test(TestConstants, test_constants_bool, dim=1, inputs=[], devices=devices)
+add_kernel_test(TestConstants, test_constants_int, dim=1, inputs=[a], devices=devices)
+add_kernel_test(TestConstants, test_constants_float, dim=1, inputs=[x], devices=devices)
 
-    return TestConstants
+add_function_test(TestConstants, "test_constant_math", test_constant_math, devices=devices)
+add_function_test(TestConstants, "test_constant_closure_capture", test_constant_closure_capture, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_conditional.py b/warp/tests/test_conditional.py
index b0a1aaa9c..b9c7f081b 100644
--- a/warp/tests/test_conditional.py
+++ b/warp/tests/test_conditional.py
@@ -8,7 +8,7 @@
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -203,46 +203,44 @@ def test_conditional_chain_mixed():
 
 def test_conditional_unequal_types(test: unittest.TestCase, device):
     # The bad kernel must be in a separate module, otherwise the current module would fail to load
-    from warp.tests.test_conditional_unequal_types_kernels import unequal_types_kernel
+    from warp.tests.aux_test_conditional_unequal_types_kernels import (
+        unequal_types_kernel,
+    )
 
     with test.assertRaises(TypeError):
         wp.launch(unequal_types_kernel, dim=(1,), inputs=[], device=device)
 
     # remove all references to the bad module so that subsequent calls to wp.force_load()
     # won't try to load it unless we explicitly re-import it again
-    del wp.context.user_modules["warp.tests.test_conditional_unequal_types_kernels"]
-    del sys.modules["warp.tests.test_conditional_unequal_types_kernels"]
-
-
-def register(parent):
-    devices = get_test_devices()
-
-    class TestConditional(parent):
-        pass
-
-    add_kernel_test(TestConditional, kernel=test_conditional_if_else, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_if_else_nested, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_boolean_and, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_boolean_or, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_boolean_compound, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_boolean_literal, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_int_logical_not, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_int_conditional_assign_overload, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_bool_param_conditional, dim=1, inputs=[True], devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_chain_basic, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_chain_empty_range, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_chain_faker, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_chain_and, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_chain_eqs, dim=1, devices=devices)
-    add_kernel_test(TestConditional, kernel=test_conditional_chain_mixed, dim=1, devices=devices)
-    add_function_test(
-        TestConditional, "test_conditional_unequal_types", test_conditional_unequal_types, devices=devices
-    )
+    del wp.context.user_modules["warp.tests.aux_test_conditional_unequal_types_kernels"]
+    del sys.modules["warp.tests.aux_test_conditional_unequal_types_kernels"]
+
+
+devices = get_test_devices()
+
+
+class TestConditional(unittest.TestCase):
+    pass
+
 
-    return TestConditional
+add_kernel_test(TestConditional, kernel=test_conditional_if_else, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_if_else_nested, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_boolean_and, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_boolean_or, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_boolean_compound, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_boolean_literal, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_int_logical_not, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_int_conditional_assign_overload, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_bool_param_conditional, dim=1, inputs=[True], devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_chain_basic, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_chain_empty_range, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_chain_faker, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_chain_and, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_chain_eqs, dim=1, devices=devices)
+add_kernel_test(TestConditional, kernel=test_conditional_chain_mixed, dim=1, devices=devices)
+add_function_test(TestConditional, "test_conditional_unequal_types", test_conditional_unequal_types, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_conditional_unequal_types_kernels.py b/warp/tests/test_conditional_unequal_types_kernels.py
deleted file mode 100644
index c2d16f6d5..000000000
--- a/warp/tests/test_conditional_unequal_types_kernels.py
+++ /dev/null
@@ -1,14 +0,0 @@
-# This file defines a kernel that fails on codegen.py
-
-import warp as wp
-
-
-@wp.kernel
-def unequal_types_kernel():
-    x = wp.int32(10)
-    y = 10
-    z = True
-
-    # Throws a TypeError
-    if x == y == z:
-        pass
diff --git a/warp/tests/test_copy.py b/warp/tests/test_copy.py
index 1b956a25b..63942ef46 100644
--- a/warp/tests/test_copy.py
+++ b/warp/tests/test_copy.py
@@ -5,13 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
+import unittest
+
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
-
-import unittest
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -200,19 +199,17 @@ def test_copy_indexed(test, device):
         assert_np_equal(a4.numpy(), expected4 * s)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestCopy(parent):
-        pass
+class TestCopy(unittest.TestCase):
+    pass
 
-    add_function_test(TestCopy, "test_copy_strided", test_copy_strided, devices=devices)
-    add_function_test(TestCopy, "test_copy_indexed", test_copy_indexed, devices=devices)
 
-    return TestCopy
+add_function_test(TestCopy, "test_copy_strided", test_copy_strided, devices=devices)
+add_function_test(TestCopy, "test_copy_indexed", test_copy_indexed, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_coverage.py b/warp/tests/test_coverage.py
deleted file mode 100644
index 8ffb6d410..000000000
--- a/warp/tests/test_coverage.py
+++ /dev/null
@@ -1,38 +0,0 @@
-import os
-
-import coverage
-
-cover = coverage.Coverage(
-    source=["warp", "warp.sim", "warp.render"],
-    omit=[
-        "build*.py",
-        "copyright.py",
-        "setup.py",
-        os.path.join("warp", "stubs.py"),
-        os.path.join("warp", "tests", "*.py"),
-        os.path.join("warp", "thirdparty", "appdirs.py"),
-        os.path.join("warp", "fem", "**", "*.py"),
-        os.path.join("warp", "render", "render_opengl.py"),
-    ],
-)
-
-cover.config.disable_warnings = [
-    "module-not-measured",
-    "module-not-imported",
-    "no-data-collected",
-    "couldnt-parse",
-]
-
-cover.exclude("@wp")
-cover.exclude("@warp")
-cover.start()
-
-import test_all  # noqa: E402
-
-test_all.run()
-
-
-cover.stop()
-cover.save()
-
-cover.html_report()
diff --git a/warp/tests/test_ctypes.py b/warp/tests/test_ctypes.py
index 9717f3496..660ef8f01 100644
--- a/warp/tests/test_ctypes.py
+++ b/warp/tests/test_ctypes.py
@@ -5,14 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
+import unittest
+
 import numpy as np
-import math
 
 import warp as wp
-from warp.tests.test_base import *
-
-import unittest
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -556,89 +554,79 @@ def test_transform_matrix():
     wp.expect_near(r_2, r_0 - t, 1.0e-4)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestCTypes(parent):
-        pass
-
-    inputs = [
-        wp.vec2(1.0, 2.0),
-        wp.vec3(1.0, 2.0, 3.0),
-        wp.vec4(1.0, 2.0, 3.0, 4.0),
-        wp.mat22(1.0, 2.0, 3.0, 4.0),
-        wp.mat33(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0),
-        wp.mat44(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0),
-    ]
-
-    add_function_test(TestCTypes, "test_mat22", test_mat22, devices=devices)
-    add_function_test(TestCTypes, "test_mat33", test_mat33, devices=devices)
-    add_function_test(TestCTypes, "test_mat44", test_mat44, devices=devices)
-    add_kernel_test(
-        TestCTypes,
-        name="test_transformation_constructor",
-        kernel=test_transformation_constructor,
-        dim=1,
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCTypes,
-        name="test_spatial_vector_constructor",
-        kernel=test_spatial_vector_constructor,
-        dim=1,
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCTypes,
-        name="test_scalar_arg_types",
-        kernel=test_scalar_arg_types,
-        dim=1,
-        inputs=[-64, 255, -64, 255, -64, 255, -64, 255, 3.14159, 3.14159],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCTypes,
-        name="test_scalar_arg_types_explicit",
-        kernel=test_scalar_arg_types,
-        dim=1,
-        inputs=[
-            wp.int8(-64),
-            wp.uint8(255),
-            wp.int16(-64),
-            wp.uint16(255),
-            wp.int32(-64),
-            wp.uint32(255),
-            wp.int64(-64),
-            wp.uint64(255),
-            wp.float32(3.14159),
-            wp.float64(3.14159),
-        ],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestCTypes, name="test_vector_arg_types", kernel=test_vector_arg_types, dim=1, inputs=inputs, devices=devices
-    )
-    add_kernel_test(TestCTypes, name="test_type_convesrions", kernel=test_type_conversions, dim=1, devices=devices)
-
-    add_function_test(
-        TestCTypes, "test_scalar_array_load", test_scalar_array_types, devices=devices, load=True, store=False
-    )
-    add_function_test(
-        TestCTypes, "test_scalar_array_store", test_scalar_array_types, devices=devices, load=False, store=True
-    )
-    add_function_test(TestCTypes, "test_vec2_arg", test_vec2_arg, devices=devices, n=8)
-    add_function_test(TestCTypes, "test_vec2_transform", test_vec2_transform, devices=devices, n=8)
-    add_function_test(TestCTypes, "test_vec3_arg", test_vec3_arg, devices=devices, n=8)
-    add_function_test(TestCTypes, "test_vec3_transform", test_vec3_transform, devices=devices, n=8)
-    add_function_test(TestCTypes, "test_transform_multiply", test_transform_multiply, devices=devices, n=8)
-    add_kernel_test(TestCTypes, name="test_transform_matrix", kernel=test_transform_matrix, dim=1, devices=devices)
-    add_function_test(TestCTypes, "test_scalar_array", test_scalar_array, devices=devices)
-    add_function_test(TestCTypes, "test_vector_array", test_vector_array, devices=devices)
-
-    return TestCTypes
+devices = get_test_devices()
+
+
+class TestCTypes(unittest.TestCase):
+    pass
+
+
+inputs = [
+    wp.vec2(1.0, 2.0),
+    wp.vec3(1.0, 2.0, 3.0),
+    wp.vec4(1.0, 2.0, 3.0, 4.0),
+    wp.mat22(1.0, 2.0, 3.0, 4.0),
+    wp.mat33(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0),
+    wp.mat44(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0),
+]
+
+add_function_test(TestCTypes, "test_mat22", test_mat22, devices=devices)
+add_function_test(TestCTypes, "test_mat33", test_mat33, devices=devices)
+add_function_test(TestCTypes, "test_mat44", test_mat44, devices=devices)
+add_kernel_test(
+    TestCTypes, name="test_transformation_constructor", kernel=test_transformation_constructor, dim=1, devices=devices
+)
+add_kernel_test(
+    TestCTypes, name="test_spatial_vector_constructor", kernel=test_spatial_vector_constructor, dim=1, devices=devices
+)
+add_kernel_test(
+    TestCTypes,
+    name="test_scalar_arg_types",
+    kernel=test_scalar_arg_types,
+    dim=1,
+    inputs=[-64, 255, -64, 255, -64, 255, -64, 255, 3.14159, 3.14159],
+    devices=devices,
+)
+add_kernel_test(
+    TestCTypes,
+    name="test_scalar_arg_types_explicit",
+    kernel=test_scalar_arg_types,
+    dim=1,
+    inputs=[
+        wp.int8(-64),
+        wp.uint8(255),
+        wp.int16(-64),
+        wp.uint16(255),
+        wp.int32(-64),
+        wp.uint32(255),
+        wp.int64(-64),
+        wp.uint64(255),
+        wp.float32(3.14159),
+        wp.float64(3.14159),
+    ],
+    devices=devices,
+)
+add_kernel_test(
+    TestCTypes, name="test_vector_arg_types", kernel=test_vector_arg_types, dim=1, inputs=inputs, devices=devices
+)
+add_kernel_test(TestCTypes, name="test_type_convesrions", kernel=test_type_conversions, dim=1, devices=devices)
+
+add_function_test(
+    TestCTypes, "test_scalar_array_load", test_scalar_array_types, devices=devices, load=True, store=False
+)
+add_function_test(
+    TestCTypes, "test_scalar_array_store", test_scalar_array_types, devices=devices, load=False, store=True
+)
+add_function_test(TestCTypes, "test_vec2_arg", test_vec2_arg, devices=devices, n=8)
+add_function_test(TestCTypes, "test_vec2_transform", test_vec2_transform, devices=devices, n=8)
+add_function_test(TestCTypes, "test_vec3_arg", test_vec3_arg, devices=devices, n=8)
+add_function_test(TestCTypes, "test_vec3_transform", test_vec3_transform, devices=devices, n=8)
+add_function_test(TestCTypes, "test_transform_multiply", test_transform_multiply, devices=devices, n=8)
+add_kernel_test(TestCTypes, name="test_transform_matrix", kernel=test_transform_matrix, dim=1, devices=devices)
+add_function_test(TestCTypes, "test_scalar_array", test_scalar_array, devices=devices)
+add_function_test(TestCTypes, "test_vector_array", test_vector_array, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_dense.py b/warp/tests/test_dense.py
index 0841bb1a3..a489d52e5 100644
--- a/warp/tests/test_dense.py
+++ b/warp/tests/test_dense.py
@@ -1,11 +1,15 @@
-import numpy as np
-import math
-
-import warp as wp
-from warp.tests.test_base import *
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
 
 import unittest
 
+import warp as wp
+from warp.tests.unittest_utils import *
+
 wp.init()
 
 
@@ -42,20 +46,22 @@ def eval_dense_solve(
     wp.dense_solve(n, A, L, b, x)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestDense(parent):
-        pass
-
+def test_dense_compilation(test, device):
     # just testing compilation of the dense matrix routines
     # most are deprecated / WIP
-    wp.force_load()
+    wp.load_module(device=device)
+
+
+devices = get_test_devices()
+
+
+class TestDense(unittest.TestCase):
+    pass
+
 
-    return TestDense
+add_function_test(TestDense, "test_dense_compilation", test_dense_compilation, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_devices.py b/warp/tests/test_devices.py
index 2125c43bc..c597eb472 100644
--- a/warp/tests/test_devices.py
+++ b/warp/tests/test_devices.py
@@ -8,33 +8,31 @@
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
 
-def test_devices_get_device_functions(test, device):
-    # save default device
-    saved_device = wp.get_device()
-
-    test.assertTrue(saved_device.is_cuda)
+def test_devices_get_cuda_device_functions(test, device):
+    test.assertTrue(device.is_cuda)
     test.assertTrue(wp.is_device_available(device))
 
     device_ordinal = device.ordinal
     current_device = wp.get_cuda_device(device_ordinal)
     test.assertEqual(current_device, device)
-
     current_device = wp.get_cuda_device()  # No-ordinal version
     test.assertTrue(wp.is_device_available(current_device))
 
+    if device == current_device:
+        test.assertEqual(device, "cuda")
+    else:
+        test.assertNotEqual(device, "cuda")
+
     preferred_device = wp.get_preferred_device()
     test.assertTrue(wp.is_device_available(preferred_device))
 
-    # restore default device
-    wp.set_device(saved_device)
 
-
-def test_devices_map_device(test, device):
+def test_devices_map_cuda_device(test, device):
     with wp.ScopedDevice(device):
         saved_alias = device.alias
         # Map alias twice to check code path
@@ -58,42 +56,43 @@ def test_devices_verify_cuda_device(test, device):
     wp.config.verify_cuda = verify_cuda_saved
 
 
+@unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
 def test_devices_can_access_self(test, device):
     test.assertTrue(device.can_access(device))
 
-    # Also test CPU access
-    cpu_device = wp.get_device("cpu")
-    if device != cpu_device:
-        test.assertFalse(device.can_access(cpu_device))
+    for warp_device in wp.get_devices():
+        device_str = str(warp_device)
+
+        if (device.is_cpu and warp_device.is_cuda) or (device.is_cuda and warp_device.is_cpu):
+            test.assertFalse(device.can_access(warp_device))
+            test.assertNotEqual(device, warp_device)
+            test.assertNotEqual(device, device_str)
 
-    test.assertNotEqual(cpu_device, "cuda")
 
+devices = get_test_devices()
 
-def register(parent):
-    devices = get_test_devices()
 
-    class TestDevices(parent):
-        pass
+class TestDevices(unittest.TestCase):
+    pass
 
-    add_function_test(
-        TestDevices,
-        "test_devices_get_device_functions",
-        test_devices_get_device_functions,
-        devices=wp.get_cuda_devices(),
-    )
-    add_function_test(TestDevices, "test_devices_map_device", test_devices_map_device, devices=wp.get_cuda_devices())
-    add_function_test(
-        TestDevices, "test_devices_unmap_imaginary_device", test_devices_unmap_imaginary_device, devices=devices
-    )
-    add_function_test(TestDevices, "test_devices_verify_cuda_device", test_devices_verify_cuda_device, devices=devices)
 
-    if wp.is_cuda_available():
-        add_function_test(TestDevices, "test_devices_can_access_self", test_devices_can_access_self, devices=devices)
+add_function_test(
+    TestDevices,
+    "test_devices_get_cuda_device_functions",
+    test_devices_get_cuda_device_functions,
+    devices=get_unique_cuda_test_devices(),
+)
+add_function_test(
+    TestDevices, "test_devices_map_cuda_device", test_devices_map_cuda_device, devices=get_unique_cuda_test_devices()
+)
+add_function_test(
+    TestDevices, "test_devices_unmap_imaginary_device", test_devices_unmap_imaginary_device, devices=devices
+)
+add_function_test(TestDevices, "test_devices_verify_cuda_device", test_devices_verify_cuda_device, devices=devices)
 
-    return TestDevices
+add_function_test(TestDevices, "test_devices_can_access_self", test_devices_can_access_self, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_dlpack.py b/warp/tests/test_dlpack.py
index 4c14c4bc4..2d514485c 100644
--- a/warp/tests/test_dlpack.py
+++ b/warp/tests/test_dlpack.py
@@ -5,13 +5,14 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import numpy as np
-import unittest
-import os
 import ctypes
+import os
+import unittest
+
+import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -299,79 +300,77 @@ def test_dlpack_jax_to_warp(test, device):
         assert_np_equal(a2.numpy(), np.asarray(j))
 
 
-def register(parent):
-    class TestDLPack(parent):
-        pass
-
-    devices = get_test_devices()
-
-    add_function_test(TestDLPack, "test_dlpack_warp_to_warp", test_dlpack_warp_to_warp, devices=devices)
-    add_function_test(TestDLPack, "test_dlpack_dtypes_and_shapes", test_dlpack_dtypes_and_shapes, devices=devices)
-
-    # torch interop via dlpack
-    try:
-        import torch
-        import torch.utils.dlpack
-
-        # check which Warp devices work with Torch
-        # CUDA devices may fail if Torch was not compiled with CUDA support
-        test_devices = get_test_devices()
-        torch_compatible_devices = []
-        for d in test_devices:
-            try:
-                t = torch.arange(10, device=wp.device_to_torch(d))
-                t += 1
-                torch_compatible_devices.append(d)
-            except Exception as e:
-                print(f"Skipping Torch DLPack tests on device '{d}' due to exception: {e}")
-
-        if torch_compatible_devices:
-            add_function_test(
-                TestDLPack, "test_dlpack_warp_to_torch", test_dlpack_warp_to_torch, devices=torch_compatible_devices
-            )
-            add_function_test(
-                TestDLPack, "test_dlpack_torch_to_warp", test_dlpack_torch_to_warp, devices=torch_compatible_devices
-            )
-
-    except Exception as e:
-        print(f"Skipping Torch DLPack tests due to exception: {e}")
-
-    # jax interop via dlpack
-    try:
-        # prevent Jax from gobbling up GPU memory
-        os.environ["XLA_PYTHON_CLIENT_PREALLOCATE"] = "false"
-
-        import jax
-        import jax.dlpack
-
-        # check which Warp devices work with Jax
-        # CUDA devices may fail if Jax cannot find a CUDA Toolkit
-        test_devices = get_test_devices()
-        jax_compatible_devices = []
-        for d in test_devices:
-            try:
-                with jax.default_device(wp.device_to_jax(d)):
-                    j = jax.numpy.arange(10, dtype=jax.numpy.float32)
-                    j += 1
-                jax_compatible_devices.append(d)
-            except Exception as e:
-                print(f"Skipping Jax DLPack tests on device '{d}' due to exception: {e}")
-
-        if jax_compatible_devices:
-            add_function_test(
-                TestDLPack, "test_dlpack_warp_to_jax", test_dlpack_warp_to_jax, devices=jax_compatible_devices
-            )
-            add_function_test(
-                TestDLPack, "test_dlpack_jax_to_warp", test_dlpack_jax_to_warp, devices=jax_compatible_devices
-            )
-
-    except Exception as e:
-        print(f"Skipping Jax DLPack tests due to exception: {e}")
-
-    return TestDLPack
+class TestDLPack(unittest.TestCase):
+    pass
+
+
+devices = get_test_devices()
+
+add_function_test(TestDLPack, "test_dlpack_warp_to_warp", test_dlpack_warp_to_warp, devices=devices)
+add_function_test(TestDLPack, "test_dlpack_dtypes_and_shapes", test_dlpack_dtypes_and_shapes, devices=devices)
+
+# torch interop via dlpack
+try:
+    import torch
+    import torch.utils.dlpack
+
+    # check which Warp devices work with Torch
+    # CUDA devices may fail if Torch was not compiled with CUDA support
+    test_devices = get_test_devices()
+    torch_compatible_devices = []
+    for d in test_devices:
+        try:
+            t = torch.arange(10, device=wp.device_to_torch(d))
+            t += 1
+            torch_compatible_devices.append(d)
+        except Exception as e:
+            print(f"Skipping Torch DLPack tests on device '{d}' due to exception: {e}")
+
+    if torch_compatible_devices:
+        add_function_test(
+            TestDLPack, "test_dlpack_warp_to_torch", test_dlpack_warp_to_torch, devices=torch_compatible_devices
+        )
+        add_function_test(
+            TestDLPack, "test_dlpack_torch_to_warp", test_dlpack_torch_to_warp, devices=torch_compatible_devices
+        )
+
+except Exception as e:
+    print(f"Skipping Torch DLPack tests due to exception: {e}")
+
+# jax interop via dlpack
+try:
+    # prevent Jax from gobbling up GPU memory
+    os.environ["XLA_PYTHON_CLIENT_PREALLOCATE"] = "false"
+    os.environ["XLA_PYTHON_CLIENT_ALLOCATOR"] = "platform"
+
+    import jax
+    import jax.dlpack
+
+    # check which Warp devices work with Jax
+    # CUDA devices may fail if Jax cannot find a CUDA Toolkit
+    test_devices = get_test_devices()
+    jax_compatible_devices = []
+    for d in test_devices:
+        try:
+            with jax.default_device(wp.device_to_jax(d)):
+                j = jax.numpy.arange(10, dtype=jax.numpy.float32)
+                j += 1
+            jax_compatible_devices.append(d)
+        except Exception as e:
+            print(f"Skipping Jax DLPack tests on device '{d}' due to exception: {e}")
+
+    if jax_compatible_devices:
+        add_function_test(
+            TestDLPack, "test_dlpack_warp_to_jax", test_dlpack_warp_to_jax, devices=jax_compatible_devices
+        )
+        add_function_test(
+            TestDLPack, "test_dlpack_jax_to_warp", test_dlpack_jax_to_warp, devices=jax_compatible_devices
+        )
+
+except Exception as e:
+    print(f"Skipping Jax DLPack tests due to exception: {e}")
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_examples.py b/warp/tests/test_examples.py
index a87497841..b0999d2c6 100644
--- a/warp/tests/test_examples.py
+++ b/warp/tests/test_examples.py
@@ -1,130 +1,225 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import importlib
 import os
 import unittest
 
 import warp as wp
+from warp.tests.unittest_utils import get_unique_cuda_test_devices
+
+wp.init()
 
 
 # registers an example to run as a TestCase
-def add_example_test(cls, name, options):
+def add_example_test(cls, name, devices=None, options={}):
     def run(test, device):
-        # disable scoped timer to avoid log spam
-        wp.ScopedTimer.enabled = False
-
         try:
             module = importlib.import_module(f"examples.{name}")
+
+            torch_cuda_required = options.setdefault("torch_cuda_required", False)
+            options.pop("torch_cuda_required", None)
+            if torch_cuda_required and wp.get_device(device).is_cuda:
+                # Ensure torch has CUDA support
+                import torch
+
+                if not torch.cuda.is_available():
+                    test.skipTest("Torch not compiled with CUDA support")
+
         except Exception as e:
-            print(f"Skipping example: {name}\n    Reason: {e}")
-            return
+            test.skipTest(f"{e}")
 
         # create default USD stage output path which many examples expect
-        if "stage" not in options:
-            stage_path = os.path.join(os.path.dirname(__file__), f"outputs/{name}.usd")
-            options["stage"] = stage_path
-
-        if "num_frames" in options:
-            num_frames = options["num_frames"]
-            del options["num_frames"]
-        else:
-            num_frames = 10
-
-        e = module.Example(**options)
-
-        for _ in range(num_frames):
-            e.update()
-            e.render()
-
-        wp.ScopedTimer.enabled = True
-
-    from warp.tests.test_base import add_function_test
-
-    add_function_test(cls, f"test_{name}", run, check_output=False)
-
-
-def register(parent):
-    class TestExamples(parent):
-        pass
-
-    # Exclude unless we can run headless somehow
-    # add_example_test(TestExamples, name="example_render_opengl", options={})
-
-    # TODO: Test CPU and GPU versions
-    if wp.is_cuda_available():
-        add_example_test(TestExamples, name="example_dem", options={})
-        add_example_test(TestExamples, name="example_diffray", options={})
-        add_example_test(TestExamples, name="example_fluid", options={})
-        add_example_test(TestExamples, name="example_jacobian_ik", options={})
-        add_example_test(TestExamples, name="example_marching_cubes", options={})
-        add_example_test(TestExamples, name="example_mesh", options={})
-        add_example_test(TestExamples, name="example_mesh_intersect", options={"num_frames": 1})
-        add_example_test(TestExamples, name="example_nvdb", options={})
-        add_example_test(TestExamples, name="example_raycast", options={})
-        add_example_test(TestExamples, name="example_raymarch", options={})
-        add_example_test(TestExamples, name="example_sim_cartpole", options={})
-        add_example_test(TestExamples, name="example_sim_cloth", options={})
-        add_example_test(TestExamples, name="example_sim_fk_grad", options={})
-        # add_example_test(TestExamples, name="example_sim_fk_grad_torch", options={})   # disabling due to failure on TC machines that have torch but not CUDA torch
-        add_example_test(TestExamples, name="example_sim_grad_bounce", options={})
-        add_example_test(TestExamples, name="example_sim_grad_cloth", options={})
-        add_example_test(TestExamples, name="example_sim_granular", options={})
-        add_example_test(TestExamples, name="example_sim_granular_collision_sdf", options={})
-        add_example_test(TestExamples, name="example_sim_neo_hookean", options={})
-        add_example_test(TestExamples, name="example_sim_particle_chain", options={})
-        add_example_test(TestExamples, name="example_sim_quadruped", options={})
-        add_example_test(TestExamples, name="example_sim_rigid_chain", options={})
-        add_example_test(TestExamples, name="example_sim_rigid_contact", options={})
-        add_example_test(TestExamples, name="example_sim_rigid_fem", options={})
-        add_example_test(TestExamples, name="example_sim_rigid_force", options={})
-        add_example_test(TestExamples, name="example_sim_rigid_gyroscopic", options={})
-        add_example_test(TestExamples, name="example_sim_rigid_kinematics", options={})
-        add_example_test(TestExamples, name="example_sim_trajopt", options={})
-        add_example_test(TestExamples, name="example_sph", options={})
-        add_example_test(TestExamples, name="example_wave", options={"resx": 256, "resy": 256})
-        add_example_test(TestExamples, name="fem.example_diffusion_mgpu", options={"quiet": True, "num_frames": 1})
-
-    # USD import failures should not count as a test failure
-    try:
-        from pxr import Usd, UsdGeom
-
-        have_usd = True
-    except:
-        have_usd = False
-
-    # The following examples do not need CUDA, but they need USD
-    if have_usd:
-        add_example_test(TestExamples, name="fem.example_apic_fluid", options={"quiet": True, "res": [16, 16, 16]})
-        add_example_test(
-            TestExamples,
-            name="fem.example_diffusion",
-            options={"quiet": True, "resolution": 10, "mesh": "tri", "num_frames": 1},
-        )
-        add_example_test(
-            TestExamples, name="fem.example_diffusion_3d", options={"quiet": True, "resolution": 10, "num_frames": 1}
-        )
-        add_example_test(
-            TestExamples,
-            name="fem.example_deformed_geometry",
-            options={"quiet": True, "resolution": 10, "num_frames": 1, "mesh": "tri"},
-        )
-        add_example_test(
-            TestExamples, name="fem.example_convection_diffusion", options={"quiet": True, "resolution": 20}
-        )
-        add_example_test(
-            TestExamples,
-            name="fem.example_mixed_elasticity",
-            options={"quiet": True, "nonconforming_stresses": True, "mesh": "quad", "num_frames": 1},
-        )
-        add_example_test(TestExamples, name="fem.example_stokes_transfer", options={"quiet": True, "num_frames": 1})
-
-    return TestExamples
+        options.setdefault("stage", os.path.join(os.path.dirname(__file__), f"outputs/{name}.usd"))
 
+        num_frames = options.get("num_frames", 10)
+        options.pop("num_frames", None)
 
-if __name__ == "__main__":
-    wp.init()
+        # Don't want to force load all modules by default for serial test runner
+        wp.config.graph_capture_module_load_default = False
+
+        try:
+            enable_backward = options.get("enable_backward", True)
+            wp.set_module_options({"enable_backward": enable_backward}, module)
+            options.pop("enable_backward", None)
 
+            wp.load_module(module, device=wp.get_device())
+            extra_load_modules = options.get("load_modules", [])
+            for module_name in extra_load_modules:
+                wp.load_module(module_name, device=wp.get_device())
+            options.pop("load_modules", None)
+
+            e = module.Example(**options)
+
+            # disable scoped timer to avoid log spam from time steps
+            wp.ScopedTimer.enabled = False
+
+            for _ in range(num_frames):
+                e.update()
+                e.render()
+        except Exception as e:
+            test.fail(f"{e}")
+        finally:
+            wp.ScopedTimer.enabled = True
+            wp.config.graph_capture_module_load_default = True
+
+    from warp.tests.unittest_utils import add_function_test
+
+    add_function_test(cls, f"test_{name}", run, devices=devices, check_output=False)
+
+
+# TODO: Make the example classes use the passed in device
+cuda_test_devices = get_unique_cuda_test_devices()
+
+# NOTE: To give the parallel test runner more opportunities to parallelize test cases,
+# we break up the tests into multiple TestCase classes that should be non-conflicting
+# w.r.t. kernel compilation
+
+
+class TestExamples(unittest.TestCase):
+    pass
+
+
+# Exclude unless we can run headless somehow
+# add_example_test(TestExamples, name="example_render_opengl", options={})
+
+add_example_test(TestExamples, name="example_dem", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_diffray", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_fluid", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_jacobian_ik", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_marching_cubes", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_mesh", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_mesh_intersect", devices=cuda_test_devices, options={"num_frames": 1})
+add_example_test(TestExamples, name="example_nvdb", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_raycast", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_raymarch", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_sph", devices=cuda_test_devices)
+add_example_test(TestExamples, name="example_wave", devices=cuda_test_devices, options={"resx": 256, "resy": 256})
+
+
+class TestSimExamples(unittest.TestCase):
+    pass
+
+
+add_example_test(
+    TestSimExamples,
+    name="example_sim_cartpole",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.integrator_euler"]},
+)
+add_example_test(
+    TestSimExamples,
+    name="example_sim_cloth",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.collide", "warp.sim.integrator_euler", "warp.sim.particles"]},
+)
+add_example_test(TestSimExamples, name="example_sim_fk_grad", devices=cuda_test_devices)
+add_example_test(
+    TestSimExamples, name="example_sim_fk_grad_torch", devices=cuda_test_devices, options={"torch_cuda_required": True}
+)
+add_example_test(
+    TestSimExamples,
+    name="example_sim_grad_bounce",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.integrator_euler", "warp.sim.particles"]},
+)
+add_example_test(
+    TestSimExamples,
+    name="example_sim_grad_cloth",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.integrator_euler", "warp.sim.particles"]},
+)
+add_example_test(TestSimExamples, name="example_sim_granular", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_granular_collision_sdf", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_neo_hookean", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_particle_chain", devices=cuda_test_devices)
+add_example_test(
+    TestSimExamples,
+    name="example_sim_quadruped",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.integrator_xpbd", "warp.sim.integrator_euler"]},
+)
+add_example_test(
+    TestSimExamples,
+    name="example_sim_rigid_chain",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.integrator_xpbd", "warp.sim.integrator_euler"]},
+)
+add_example_test(
+    TestSimExamples,
+    name="example_sim_rigid_contact",
+    devices=cuda_test_devices,
+    options={"load_modules": ["warp.sim.integrator_euler"]},
+)
+add_example_test(TestSimExamples, name="example_sim_rigid_fem", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_rigid_force", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_rigid_gyroscopic", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_rigid_kinematics", devices=cuda_test_devices)
+add_example_test(TestSimExamples, name="example_sim_trajopt", devices=cuda_test_devices)
+
+
+class TestFemExamples(unittest.TestCase):
+    pass
+
+
+add_example_test(
+    TestFemExamples,
+    name="fem.example_diffusion_mgpu",
+    devices=cuda_test_devices,
+    options={"quiet": True, "num_frames": 1, "enable_backward": False},
+)
+
+# The following examples do not need CUDA, but they need USD
+add_example_test(
+    TestFemExamples,
+    name="fem.example_apic_fluid",
+    devices=cuda_test_devices,
+    options={"quiet": True, "res": [16, 16, 16], "enable_backward": False},
+)
+add_example_test(
+    TestFemExamples,
+    name="fem.example_diffusion",
+    devices=cuda_test_devices,
+    options={"quiet": True, "resolution": 10, "mesh": "tri", "num_frames": 1, "enable_backward": False},
+)
+add_example_test(
+    TestFemExamples,
+    name="fem.example_diffusion_3d",
+    devices=cuda_test_devices,
+    options={"quiet": True, "resolution": 10, "num_frames": 1, "enable_backward": False},
+)
+add_example_test(
+    TestFemExamples,
+    name="fem.example_deformed_geometry",
+    devices=cuda_test_devices,
+    options={"quiet": True, "resolution": 10, "num_frames": 1, "mesh": "tri", "enable_backward": False},
+)
+add_example_test(
+    TestFemExamples,
+    name="fem.example_convection_diffusion",
+    devices=cuda_test_devices,
+    options={"quiet": True, "resolution": 20, "enable_backward": False},
+)
+add_example_test(
+    TestFemExamples,
+    name="fem.example_mixed_elasticity",
+    devices=cuda_test_devices,
+    options={"quiet": True, "nonconforming_stresses": True, "mesh": "quad", "num_frames": 1, "enable_backward": False},
+)
+add_example_test(
+    TestFemExamples,
+    name="fem.example_stokes_transfer",
+    devices=cuda_test_devices,
+    options={"quiet": True, "num_frames": 1, "enable_backward": False},
+)
+
+
+if __name__ == "__main__":
     # force rebuild of all kernels
     wp.build.clear_kernel_cache()
 
-    _ = register(unittest.TestCase)
-
     unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_fabricarray.py b/warp/tests/test_fabricarray.py
index 09642c1ea..7a23c204f 100644
--- a/warp/tests/test_fabricarray.py
+++ b/warp/tests/test_fabricarray.py
@@ -12,7 +12,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -930,32 +930,26 @@ def test_fabricarrayarray(test, device):
     wp.overload(fa_generic_sums_kernel_indexed, [wp.indexedfabricarrayarray(dtype=T), wp.array(dtype=T)])
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestFabricArray(parent):
-        pass
 
-    # fabric arrays
-    add_function_test(TestFabricArray, "test_fabricarray_kernel", test_fabricarray_kernel, devices=devices)
-    add_function_test(TestFabricArray, "test_fabricarray_empty", test_fabricarray_empty, devices=devices)
-    add_function_test(
-        TestFabricArray, "test_fabricarray_generic_dtype", test_fabricarray_generic_dtype, devices=devices
-    )
-    add_function_test(
-        TestFabricArray, "test_fabricarray_generic_array", test_fabricarray_generic_array, devices=devices
-    )
-    add_function_test(TestFabricArray, "test_fabricarray_fill_scalar", test_fabricarray_fill_scalar, devices=devices)
-    add_function_test(TestFabricArray, "test_fabricarray_fill_vector", test_fabricarray_fill_vector, devices=devices)
-    add_function_test(TestFabricArray, "test_fabricarray_fill_matrix", test_fabricarray_fill_matrix, devices=devices)
+class TestFabricArray(unittest.TestCase):
+    pass
 
-    # fabric arrays of arrays
-    add_function_test(TestFabricArray, "test_fabricarrayarray", test_fabricarrayarray, devices=devices)
 
-    return TestFabricArray
+# fabric arrays
+add_function_test(TestFabricArray, "test_fabricarray_kernel", test_fabricarray_kernel, devices=devices)
+add_function_test(TestFabricArray, "test_fabricarray_empty", test_fabricarray_empty, devices=devices)
+add_function_test(TestFabricArray, "test_fabricarray_generic_dtype", test_fabricarray_generic_dtype, devices=devices)
+add_function_test(TestFabricArray, "test_fabricarray_generic_array", test_fabricarray_generic_array, devices=devices)
+add_function_test(TestFabricArray, "test_fabricarray_fill_scalar", test_fabricarray_fill_scalar, devices=devices)
+add_function_test(TestFabricArray, "test_fabricarray_fill_vector", test_fabricarray_fill_vector, devices=devices)
+add_function_test(TestFabricArray, "test_fabricarray_fill_matrix", test_fabricarray_fill_matrix, devices=devices)
+
+# fabric arrays of arrays
+add_function_test(TestFabricArray, "test_fabricarrayarray", test_fabricarrayarray, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_fast_math.py b/warp/tests/test_fast_math.py
index 1fe4b578e..0347c1086 100644
--- a/warp/tests/test_fast_math.py
+++ b/warp/tests/test_fast_math.py
@@ -1,11 +1,14 @@
-import warp as wp
-import numpy as np
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
 
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -33,19 +36,19 @@ def test_fast_math(test, device):
             with CheckOutput():
                 wp.launch(test_pow, dim=1, inputs=[-2.0, 2.0, 2.0], device=device)
 
+        # Turn fast math back off
+        wp.set_module_options({"fast_math": False})
+
 
-def register(parent):
-    class TestFastMath(parent):
-        pass
+class TestFastMath(unittest.TestCase):
+    pass
 
-    devices = get_test_devices()
 
-    add_function_test(TestFastMath, "test_fast_math", test_fast_math, devices=devices)
+devices = get_test_devices()
 
-    return TestFastMath
+add_function_test(TestFastMath, "test_fast_math", test_fast_math, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_fem.py b/warp/tests/test_fem.py
index 36f5fb110..c62c49006 100644
--- a/warp/tests/test_fem.py
+++ b/warp/tests/test_fem.py
@@ -1,4 +1,4 @@
-# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
 # NVIDIA CORPORATION and its licensors retain all intellectual property
 # and proprietary rights in and to this software, related documentation
 # and any modifications thereto.  Any use, reproduction, disclosure or
@@ -10,7 +10,7 @@
 
 import numpy as np
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 
 from warp.fem import Field, Sample, Domain, Coords
@@ -1230,38 +1230,42 @@ def test_particle_quadratures(test_case, device):
     test_case.assertAlmostEqual(val, 1.25, places=5)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestFem(parent):
-        pass
-
-    add_function_test(TestFem, "test_regular_quadrature", test_regular_quadrature)
-    add_function_test(TestFem, "test_closest_point_queries", test_closest_point_queries)
-    add_function_test(TestFem, "test_grad_decomposition", test_grad_decomposition, devices=devices)
-    add_function_test(TestFem, "test_integrate_gradient", test_integrate_gradient, devices=devices)
-    add_function_test(TestFem, "test_interpolate_gradient", test_interpolate_gradient, devices=devices)
-    add_function_test(TestFem, "test_vector_divergence_theorem", test_vector_divergence_theorem, devices=devices)
-    add_function_test(TestFem, "test_tensor_divergence_theorem", test_tensor_divergence_theorem, devices=devices)
-    add_function_test(TestFem, "test_grid_2d", test_grid_2d, devices=devices)
-    add_function_test(TestFem, "test_triangle_mesh", test_triangle_mesh, devices=devices)
-    add_function_test(TestFem, "test_quad_mesh", test_quad_mesh, devices=devices)
-    add_function_test(TestFem, "test_grid_3d", test_grid_3d, devices=devices)
-    add_function_test(TestFem, "test_tet_mesh", test_tet_mesh, devices=devices)
-    add_function_test(TestFem, "test_hex_mesh", test_hex_mesh, devices=devices)
-    add_function_test(TestFem, "test_deformed_geometry", test_deformed_geometry, devices=devices)
-    add_function_test(TestFem, "test_dof_mapper", test_dof_mapper)
-    add_function_test(TestFem, "test_square_shape_functions", test_square_shape_functions)
-    add_function_test(TestFem, "test_cube_shape_functions", test_cube_shape_functions)
-    add_function_test(TestFem, "test_tri_shape_functions", test_tri_shape_functions)
-    add_function_test(TestFem, "test_tet_shape_functions", test_tet_shape_functions)
-    add_function_test(TestFem, "test_point_basis", test_point_basis)
-    add_function_test(TestFem, "test_particle_quadratures", test_particle_quadratures)
-
-    return TestFem
+devices = get_test_devices()
+
+
+class TestFem(unittest.TestCase):
+    pass
+
+
+add_function_test(TestFem, "test_regular_quadrature", test_regular_quadrature)
+add_function_test(TestFem, "test_closest_point_queries", test_closest_point_queries)
+add_function_test(TestFem, "test_grad_decomposition", test_grad_decomposition, devices=devices)
+add_function_test(TestFem, "test_integrate_gradient", test_integrate_gradient, devices=devices)
+add_function_test(TestFem, "test_interpolate_gradient", test_interpolate_gradient, devices=devices)
+add_function_test(TestFem, "test_vector_divergence_theorem", test_vector_divergence_theorem, devices=devices)
+add_function_test(TestFem, "test_tensor_divergence_theorem", test_tensor_divergence_theorem, devices=devices)
+add_function_test(TestFem, "test_grid_2d", test_grid_2d, devices=devices)
+add_function_test(TestFem, "test_triangle_mesh", test_triangle_mesh, devices=devices)
+add_function_test(TestFem, "test_quad_mesh", test_quad_mesh, devices=devices)
+add_function_test(TestFem, "test_grid_3d", test_grid_3d, devices=devices)
+add_function_test(TestFem, "test_tet_mesh", test_tet_mesh, devices=devices)
+add_function_test(TestFem, "test_hex_mesh", test_hex_mesh, devices=devices)
+add_function_test(TestFem, "test_deformed_geometry", test_deformed_geometry, devices=devices)
+add_function_test(TestFem, "test_dof_mapper", test_dof_mapper)
+add_function_test(TestFem, "test_point_basis", test_point_basis)
+add_function_test(TestFem, "test_particle_quadratures", test_particle_quadratures)
+
+
+class TestFemShapeFunctions(unittest.TestCase):
+    pass
+
+
+add_function_test(TestFemShapeFunctions, "test_square_shape_functions", test_square_shape_functions)
+add_function_test(TestFemShapeFunctions, "test_cube_shape_functions", test_cube_shape_functions)
+add_function_test(TestFemShapeFunctions, "test_tri_shape_functions", test_tri_shape_functions)
+add_function_test(TestFemShapeFunctions, "test_tet_shape_functions", test_tet_shape_functions)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_fp16.py b/warp/tests/test_fp16.py
index 09b08870e..4ccdf5114 100644
--- a/warp/tests/test_fp16.py
+++ b/warp/tests/test_fp16.py
@@ -1,9 +1,16 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -102,25 +109,22 @@ def test_fp16_grad(test, device):
     assert_np_equal(input.grad.numpy(), np.ones(len(s)) * 2.0)
 
 
-def register(parent):
-    class TestFp16(parent):
-        pass
+class TestFp16(unittest.TestCase):
+    pass
 
-    devices = []
-    if wp.is_cpu_available():
-        devices.append("cpu")
-    for cuda_device in wp.get_cuda_devices():
-        if cuda_device.arch >= 70:
-            devices.append(cuda_device)
 
-    add_function_test(TestFp16, "test_fp16_conversion", test_fp16_conversion, devices=devices)
-    add_function_test(TestFp16, "test_fp16_grad", test_fp16_grad, devices=devices)
-    add_function_test(TestFp16, "test_fp16_kernel_parameter", test_fp16_kernel_parameter, devices=devices)
+devices = []
+if wp.is_cpu_available():
+    devices.append("cpu")
+for cuda_device in get_unique_cuda_test_devices():
+    if cuda_device.arch >= 70:
+        devices.append(cuda_device)
 
-    return TestFp16
+add_function_test(TestFp16, "test_fp16_conversion", test_fp16_conversion, devices=devices)
+add_function_test(TestFp16, "test_fp16_grad", test_fp16_grad, devices=devices)
+add_function_test(TestFp16, "test_fp16_kernel_parameter", test_fp16_kernel_parameter, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_func.py b/warp/tests/test_func.py
index 109d70a15..a42d1e039 100644
--- a/warp/tests/test_func.py
+++ b/warp/tests/test_func.py
@@ -5,14 +5,13 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
-import numpy as np
 import math
+import unittest
 
-import warp as wp
-from warp.tests.test_base import *
+import numpy as np
 
-import unittest
+import warp as wp
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -83,169 +82,6 @@ def test_override_func():
     wp.expect_eq(i, 3)
 
 
-def test_native_func_export(test, device):
-    # tests calling native functions from Python
-
-    q = wp.quat(0.0, 0.0, 0.0, 1.0)
-    assert_np_equal(np.array([*q]), np.array([0.0, 0.0, 0.0, 1.0]))
-
-    r = wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), 2.0)
-    assert_np_equal(np.array([*r]), np.array([0.8414709568023682, 0.0, 0.0, 0.5403022170066833]), tol=1.0e-3)
-
-    q = wp.quat(1.0, 2.0, 3.0, 4.0)
-    q = wp.normalize(q) * 2.0
-    assert_np_equal(
-        np.array([*q]),
-        np.array([0.18257418274879456, 0.3651483654975891, 0.547722578048706, 0.7302967309951782]) * 2.0,
-        tol=1.0e-3,
-    )
-
-    v2 = wp.vec2(1.0, 2.0)
-    v2 = wp.normalize(v2) * 2.0
-    assert_np_equal(np.array([*v2]), np.array([0.4472135901451111, 0.8944271802902222]) * 2.0, tol=1.0e-3)
-
-    v3 = wp.vec3(1.0, 2.0, 3.0)
-    v3 = wp.normalize(v3) * 2.0
-    assert_np_equal(
-        np.array([*v3]), np.array([0.26726123690605164, 0.5345224738121033, 0.8017836809158325]) * 2.0, tol=1.0e-3
-    )
-
-    v4 = wp.vec4(1.0, 2.0, 3.0, 4.0)
-    v4 = wp.normalize(v4) * 2.0
-    assert_np_equal(
-        np.array([*v4]),
-        np.array([0.18257418274879456, 0.3651483654975891, 0.547722578048706, 0.7302967309951782]) * 2.0,
-        tol=1.0e-3,
-    )
-
-    v = wp.vec2(0.0)
-    v += wp.vec2(1.0, 1.0)
-    assert v == wp.vec2(1.0, 1.0)
-    v -= wp.vec2(1.0, 1.0)
-    assert v == wp.vec2(0.0, 0.0)
-    v = wp.vec2(2.0, 2.0) - wp.vec2(1.0, 1.0)
-    assert v == wp.vec2(1.0, 1.0)
-    v *= 2.0
-    assert v == wp.vec2(2.0, 2.0)
-    v = v * 2.0
-    assert v == wp.vec2(4.0, 4.0)
-    v = v / 2.0
-    assert v == wp.vec2(2.0, 2.0)
-    v /= 2.0
-    assert v == wp.vec2(1.0, 1.0)
-    v = -v
-    assert v == wp.vec2(-1.0, -1.0)
-    v = +v
-    assert v == wp.vec2(-1.0, -1.0)
-
-    m22 = wp.mat22(1.0, 2.0, 3.0, 4.0)
-    m22 = m22 + m22
-
-    test.assertEqual(m22[1, 1], 8.0)
-    test.assertEqual(str(m22), "[[2.0, 4.0],\n [6.0, 8.0]]")
-
-    t = wp.transform(
-        wp.vec3(1.0, 2.0, 3.0),
-        wp.quat(4.0, 5.0, 6.0, 7.0),
-    )
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t * wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0))
-    test.assertSequenceEqual(
-        t * wp.transform((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0)), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0)
-    )
-
-    t = wp.transform()
-    test.assertSequenceEqual(t, (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0))
-
-    t = wp.transform(p=(1.0, 2.0, 3.0), q=(4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform(q=(4.0, 5.0, 6.0, 7.0), p=(1.0, 2.0, 3.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform((1.0, 2.0, 3.0), q=(4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform(p=(1.0, 2.0, 3.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 0.0, 0.0, 0.0, 1.0))
-
-    t = wp.transform(q=(4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (0.0, 0.0, 0.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform(p=wp.vec3(1.0, 2.0, 3.0), q=wp.quat(4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0)
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform(wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    t = wp.transform(*wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
-
-    transformf = wp.types.transformation(dtype=float)
-
-    t = wp.transformf((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0))
-    test.assertSequenceEqual(
-        t + transformf((2.0, 3.0, 4.0), (5.0, 6.0, 7.0, 8.0)),
-        (3.0, 5.0, 7.0, 9.0, 11.0, 13.0, 15.0),
-    )
-    test.assertSequenceEqual(
-        t - transformf((2.0, 3.0, 4.0), (5.0, 6.0, 7.0, 8.0)),
-        (-1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0),
-    )
-
-    f = wp.sin(math.pi * 0.5)
-    test.assertAlmostEqual(f, 1.0, places=3)
-
-    m = wp.mat22(0.0, 0.0, 0.0, 0.0)
-    m += wp.mat22(1.0, 1.0, 1.0, 1.0)
-    assert m == wp.mat22(1.0, 1.0, 1.0, 1.0)
-    m -= wp.mat22(1.0, 1.0, 1.0, 1.0)
-    assert m == wp.mat22(0.0, 0.0, 0.0, 0.0)
-    m = wp.mat22(2.0, 2.0, 2.0, 2.0) - wp.mat22(1.0, 1.0, 1.0, 1.0)
-    assert m == wp.mat22(1.0, 1.0, 1.0, 1.0)
-    m *= 2.0
-    assert m == wp.mat22(2.0, 2.0, 2.0, 2.0)
-    m = m * 2.0
-    assert m == wp.mat22(4.0, 4.0, 4.0, 4.0)
-    m = m / 2.0
-    assert m == wp.mat22(2.0, 2.0, 2.0, 2.0)
-    m /= 2.0
-    assert m == wp.mat22(1.0, 1.0, 1.0, 1.0)
-    m = -m
-    assert m == wp.mat22(-1.0, -1.0, -1.0, -1.0)
-    m = +m
-    assert m == wp.mat22(-1.0, -1.0, -1.0, -1.0)
-    m = m * m
-    assert m == wp.mat22(2.0, 2.0, 2.0, 2.0)
-
-
-def test_native_function_error_resolution(test, device):
-    a = wp.mat22f(1.0, 2.0, 3.0, 4.0)
-    b = wp.mat22d(1.0, 2.0, 3.0, 4.0)
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"^Couldn't find a function 'mul' compatible with " r"the arguments 'mat22f, mat22d'$",
-    ):
-        a * b
-
-
-def test_user_func_export(test, device):
-    # tests calling overloaded user-defined functions from Python
-    i = custom(1)
-    f = custom(1.0)
-    v = custom(wp.vec3(1.0, 0.0, 0.0))
-
-    test.assertEqual(i, 2)
-    test.assertEqual(f, 2.0)
-    assert_np_equal(np.array([*v]), np.array([2.0, 0.0, 0.0]))
-
-
 def test_func_closure_capture(test, device):
     def make_closure_kernel(func):
         def closure_kernel_fn(data: wp.array(dtype=float), expected: float):
@@ -321,34 +157,181 @@ def test_builtin_shadowing():
     wp.expect_eq(sign(1.23), 123.0)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestFunc(parent):
-        pass
-
-    add_kernel_test(TestFunc, kernel=test_overload_func, name="test_overload_func", dim=1, devices=devices)
-    add_function_test(TestFunc, func=test_return_func, name="test_return_func", devices=devices)
-    add_kernel_test(TestFunc, kernel=test_override_func, name="test_override_func", dim=1, devices=devices)
-    add_function_test(TestFunc, func=test_native_func_export, name="test_native_func_export", devices=["cpu"])
-    add_function_test(
-        TestFunc,
-        func=test_native_function_error_resolution,
-        name="test_native_function_error_resolution",
-        devices=["cpu"],
-    )
-    add_function_test(TestFunc, func=test_user_func_export, name="test_user_func_export", devices=["cpu"])
-    add_function_test(TestFunc, func=test_func_closure_capture, name="test_func_closure_capture", devices=devices)
-    add_function_test(TestFunc, func=test_multi_valued_func, name="test_multi_valued_func", devices=devices)
-    add_kernel_test(TestFunc, kernel=test_func_defaults, name="test_func_defaults", dim=1, devices=devices)
-    add_kernel_test(TestFunc, kernel=test_builtin_shadowing, name="test_builtin_shadowing", dim=1, devices=devices)
-
-    return TestFunc
+devices = get_test_devices()
+
+
+class TestFunc(unittest.TestCase):
+    def test_user_func_export(self):
+        # tests calling overloaded user-defined functions from Python
+        i = custom(1)
+        f = custom(1.0)
+        v = custom(wp.vec3(1.0, 0.0, 0.0))
+
+        self.assertEqual(i, 2)
+        self.assertEqual(f, 2.0)
+        assert_np_equal(np.array([*v]), np.array([2.0, 0.0, 0.0]))
+
+    def test_native_func_export(self):
+        # tests calling native functions from Python
+
+        q = wp.quat(0.0, 0.0, 0.0, 1.0)
+        assert_np_equal(np.array([*q]), np.array([0.0, 0.0, 0.0, 1.0]))
+
+        r = wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), 2.0)
+        assert_np_equal(np.array([*r]), np.array([0.8414709568023682, 0.0, 0.0, 0.5403022170066833]), tol=1.0e-3)
+
+        q = wp.quat(1.0, 2.0, 3.0, 4.0)
+        q = wp.normalize(q) * 2.0
+        assert_np_equal(
+            np.array([*q]),
+            np.array([0.18257418274879456, 0.3651483654975891, 0.547722578048706, 0.7302967309951782]) * 2.0,
+            tol=1.0e-3,
+        )
+
+        v2 = wp.vec2(1.0, 2.0)
+        v2 = wp.normalize(v2) * 2.0
+        assert_np_equal(np.array([*v2]), np.array([0.4472135901451111, 0.8944271802902222]) * 2.0, tol=1.0e-3)
+
+        v3 = wp.vec3(1.0, 2.0, 3.0)
+        v3 = wp.normalize(v3) * 2.0
+        assert_np_equal(
+            np.array([*v3]), np.array([0.26726123690605164, 0.5345224738121033, 0.8017836809158325]) * 2.0, tol=1.0e-3
+        )
+
+        v4 = wp.vec4(1.0, 2.0, 3.0, 4.0)
+        v4 = wp.normalize(v4) * 2.0
+        assert_np_equal(
+            np.array([*v4]),
+            np.array([0.18257418274879456, 0.3651483654975891, 0.547722578048706, 0.7302967309951782]) * 2.0,
+            tol=1.0e-3,
+        )
+
+        v = wp.vec2(0.0)
+        v += wp.vec2(1.0, 1.0)
+        assert v == wp.vec2(1.0, 1.0)
+        v -= wp.vec2(1.0, 1.0)
+        assert v == wp.vec2(0.0, 0.0)
+        v = wp.vec2(2.0, 2.0) - wp.vec2(1.0, 1.0)
+        assert v == wp.vec2(1.0, 1.0)
+        v *= 2.0
+        assert v == wp.vec2(2.0, 2.0)
+        v = v * 2.0
+        assert v == wp.vec2(4.0, 4.0)
+        v = v / 2.0
+        assert v == wp.vec2(2.0, 2.0)
+        v /= 2.0
+        assert v == wp.vec2(1.0, 1.0)
+        v = -v
+        assert v == wp.vec2(-1.0, -1.0)
+        v = +v
+        assert v == wp.vec2(-1.0, -1.0)
+
+        m22 = wp.mat22(1.0, 2.0, 3.0, 4.0)
+        m22 = m22 + m22
+
+        self.assertEqual(m22[1, 1], 8.0)
+        self.assertEqual(str(m22), "[[2.0, 4.0],\n [6.0, 8.0]]")
+
+        t = wp.transform(
+            wp.vec3(1.0, 2.0, 3.0),
+            wp.quat(4.0, 5.0, 6.0, 7.0),
+        )
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t * wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0))
+        self.assertSequenceEqual(
+            t * wp.transform((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0)), (396.0, 432.0, 720.0, 56.0, 70.0, 84.0, -28.0)
+        )
+
+        t = wp.transform()
+        self.assertSequenceEqual(t, (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0))
+
+        t = wp.transform(p=(1.0, 2.0, 3.0), q=(4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform(q=(4.0, 5.0, 6.0, 7.0), p=(1.0, 2.0, 3.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform((1.0, 2.0, 3.0), q=(4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform(p=(1.0, 2.0, 3.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 0.0, 0.0, 0.0, 1.0))
+
+        t = wp.transform(q=(4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (0.0, 0.0, 0.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform(p=wp.vec3(1.0, 2.0, 3.0), q=wp.quat(4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0)
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform(wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        t = wp.transform(*wp.transform(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(t, (1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0))
+
+        transformf = wp.types.transformation(dtype=float)
+
+        t = wp.transformf((1.0, 2.0, 3.0), (4.0, 5.0, 6.0, 7.0))
+        self.assertSequenceEqual(
+            t + transformf((2.0, 3.0, 4.0), (5.0, 6.0, 7.0, 8.0)),
+            (3.0, 5.0, 7.0, 9.0, 11.0, 13.0, 15.0),
+        )
+        self.assertSequenceEqual(
+            t - transformf((2.0, 3.0, 4.0), (5.0, 6.0, 7.0, 8.0)),
+            (-1.0, -1.0, -1.0, -1.0, -1.0, -1.0, -1.0),
+        )
+
+        f = wp.sin(math.pi * 0.5)
+        self.assertAlmostEqual(f, 1.0, places=3)
+
+        m = wp.mat22(0.0, 0.0, 0.0, 0.0)
+        m += wp.mat22(1.0, 1.0, 1.0, 1.0)
+        assert m == wp.mat22(1.0, 1.0, 1.0, 1.0)
+        m -= wp.mat22(1.0, 1.0, 1.0, 1.0)
+        assert m == wp.mat22(0.0, 0.0, 0.0, 0.0)
+        m = wp.mat22(2.0, 2.0, 2.0, 2.0) - wp.mat22(1.0, 1.0, 1.0, 1.0)
+        assert m == wp.mat22(1.0, 1.0, 1.0, 1.0)
+        m *= 2.0
+        assert m == wp.mat22(2.0, 2.0, 2.0, 2.0)
+        m = m * 2.0
+        assert m == wp.mat22(4.0, 4.0, 4.0, 4.0)
+        m = m / 2.0
+        assert m == wp.mat22(2.0, 2.0, 2.0, 2.0)
+        m /= 2.0
+        assert m == wp.mat22(1.0, 1.0, 1.0, 1.0)
+        m = -m
+        assert m == wp.mat22(-1.0, -1.0, -1.0, -1.0)
+        m = +m
+        assert m == wp.mat22(-1.0, -1.0, -1.0, -1.0)
+        m = m * m
+        assert m == wp.mat22(2.0, 2.0, 2.0, 2.0)
+
+
+    def test_native_function_error_resolution(self):
+        a = wp.mat22f(1.0, 2.0, 3.0, 4.0)
+        b = wp.mat22d(1.0, 2.0, 3.0, 4.0)
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"^Couldn't find a function 'mul' compatible with " r"the arguments 'mat22f, mat22d'$",
+        ):
+            a * b
+
+
+add_kernel_test(TestFunc, kernel=test_overload_func, name="test_overload_func", dim=1, devices=devices)
+add_function_test(TestFunc, func=test_return_func, name="test_return_func", devices=devices)
+add_kernel_test(TestFunc, kernel=test_override_func, name="test_override_func", dim=1, devices=devices)
+add_function_test(TestFunc, func=test_func_closure_capture, name="test_func_closure_capture", devices=devices)
+add_function_test(TestFunc, func=test_multi_valued_func, name="test_multi_valued_func", devices=devices)
+add_kernel_test(TestFunc, kernel=test_func_defaults, name="test_func_defaults", dim=1, devices=devices)
+add_kernel_test(TestFunc, kernel=test_builtin_shadowing, name="test_builtin_shadowing", dim=1, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
-    wp.force_load()
-
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_generics.py b/warp/tests/test_generics.py
index dcddea20c..e1e158519 100644
--- a/warp/tests/test_generics.py
+++ b/warp/tests/test_generics.py
@@ -5,12 +5,13 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import numpy as np
 import unittest
 from typing import Any
 
+import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -487,81 +488,74 @@ def kernel():
         )
 
 
-def register(parent):
-    class TestGenerics(parent):
-        pass
-
-    devices = get_test_devices()
-
-    add_kernel_test(TestGenerics, name="test_generic_adder", kernel=test_generic_adder, dim=1, devices=devices)
-    add_kernel_test(TestGenerics, name="test_specialized_func", kernel=test_specialized_func, dim=1, devices=devices)
-
-    add_function_test(TestGenerics, "test_generic_array_kernel", test_generic_array_kernel, devices=devices)
-    add_function_test(TestGenerics, "test_generic_accumulator_kernel", test_generic_accumulator_kernel, devices=devices)
-    add_function_test(TestGenerics, "test_generic_fill", test_generic_fill, devices=devices)
-    add_function_test(TestGenerics, "test_generic_fill_overloads", test_generic_fill_overloads, devices=devices)
-    add_function_test(TestGenerics, "test_generic_transform_kernel", test_generic_transform_kernel, devices=devices)
-    add_function_test(
-        TestGenerics, "test_generic_transform_array_kernel", test_generic_transform_array_kernel, devices=devices
-    )
-    add_function_test(TestGenerics, "test_generic_type_cast", test_generic_type_cast, devices=devices)
-    add_function_test(TestGenerics, "test_generic_type_construction", test_generic_type_construction, devices=devices)
-    add_function_test(
-        TestGenerics, "test_generic_scalar_construction", test_generic_scalar_construction, devices=devices
-    )
-    add_function_test(TestGenerics, "test_generic_type_as_argument", test_generic_type_as_argument, devices=devices)
-
-    foo = Foo()
-    foo.x = 17.0
-    foo.y = 25.0
-    foo.z = 42.0
-
-    bar = Bar()
-    bar.x = wp.vec3(1, 2, 3)
-    bar.y = wp.vec3(10, 20, 30)
-    bar.z = wp.vec3(11, 22, 33)
-
-    add_kernel_test(
-        TestGenerics,
-        name="test_generic_struct_kernel",
-        kernel=test_generic_struct_kernel,
-        dim=1,
-        inputs=[foo],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestGenerics,
-        name="test_generic_struct_kernel",
-        kernel=test_generic_struct_kernel,
-        dim=1,
-        inputs=[bar],
-        devices=devices,
-    )
-
-    add_kernel_test(
-        TestGenerics,
-        name="test_generic_struct_construction_kernel",
-        kernel=test_generic_struct_construction_kernel,
-        dim=1,
-        inputs=[foo],
-        devices=devices,
-    )
-    add_kernel_test(
-        TestGenerics,
-        name="test_generic_struct_construction_kernel",
-        kernel=test_generic_struct_construction_kernel,
-        dim=1,
-        inputs=[bar],
-        devices=devices,
-    )
-
-    add_function_test(TestGenerics, "test_type_operator_mispell", test_type_operator_mispell, devices=devices)
-    add_function_test(TestGenerics, "test_type_attribute_error", test_type_attribute_error, devices=devices)
-
-    return TestGenerics
-
+class TestGenerics(unittest.TestCase):
+    pass
+
+
+devices = get_test_devices()
+
+add_kernel_test(TestGenerics, name="test_generic_adder", kernel=test_generic_adder, dim=1, devices=devices)
+add_kernel_test(TestGenerics, name="test_specialized_func", kernel=test_specialized_func, dim=1, devices=devices)
+
+add_function_test(TestGenerics, "test_generic_array_kernel", test_generic_array_kernel, devices=devices)
+add_function_test(TestGenerics, "test_generic_accumulator_kernel", test_generic_accumulator_kernel, devices=devices)
+add_function_test(TestGenerics, "test_generic_fill", test_generic_fill, devices=devices)
+add_function_test(TestGenerics, "test_generic_fill_overloads", test_generic_fill_overloads, devices=devices)
+add_function_test(TestGenerics, "test_generic_transform_kernel", test_generic_transform_kernel, devices=devices)
+add_function_test(
+    TestGenerics, "test_generic_transform_array_kernel", test_generic_transform_array_kernel, devices=devices
+)
+add_function_test(TestGenerics, "test_generic_type_cast", test_generic_type_cast, devices=devices)
+add_function_test(TestGenerics, "test_generic_type_construction", test_generic_type_construction, devices=devices)
+add_function_test(TestGenerics, "test_generic_scalar_construction", test_generic_scalar_construction, devices=devices)
+add_function_test(TestGenerics, "test_generic_type_as_argument", test_generic_type_as_argument, devices=devices)
+
+foo = Foo()
+foo.x = 17.0
+foo.y = 25.0
+foo.z = 42.0
+
+bar = Bar()
+bar.x = wp.vec3(1, 2, 3)
+bar.y = wp.vec3(10, 20, 30)
+bar.z = wp.vec3(11, 22, 33)
+
+add_kernel_test(
+    TestGenerics,
+    name="test_generic_struct_kernel",
+    kernel=test_generic_struct_kernel,
+    dim=1,
+    inputs=[foo],
+    devices=devices,
+)
+add_kernel_test(
+    TestGenerics,
+    name="test_generic_struct_kernel",
+    kernel=test_generic_struct_kernel,
+    dim=1,
+    inputs=[bar],
+    devices=devices,
+)
+
+add_kernel_test(
+    TestGenerics,
+    name="test_generic_struct_construction_kernel",
+    kernel=test_generic_struct_construction_kernel,
+    dim=1,
+    inputs=[foo],
+    devices=devices,
+)
+add_kernel_test(
+    TestGenerics,
+    name="test_generic_struct_construction_kernel",
+    kernel=test_generic_struct_construction_kernel,
+    dim=1,
+    inputs=[bar],
+    devices=devices,
+)
+add_function_test(TestGenerics, "test_type_operator_mispell", test_type_operator_mispell, devices=devices)
+add_function_test(TestGenerics, "test_type_attribute_error", test_type_attribute_error, devices=devices)
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_grad.py b/warp/tests/test_grad.py
index d8ed5d062..6b7d74cae 100644
--- a/warp/tests/test_grad.py
+++ b/warp/tests/test_grad.py
@@ -11,7 +11,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -67,26 +67,26 @@ def test_for_loop_grad(test, device):
 
 
 def test_for_loop_graph_grad(test, device):
+    wp.load_module(device=device)
+
     n = 32
     val = np.ones(n, dtype=np.float32)
 
     x = wp.array(val, device=device, requires_grad=True)
     sum = wp.zeros(1, dtype=wp.float32, device=device, requires_grad=True)
 
-    wp.force_load()
-
-    wp.capture_begin()
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(for_loop_grad, dim=1, inputs=[n, x, sum], device=device)
-
-    tape.backward(loss=sum)
+    wp.capture_begin(device, force_module_load=False)
+    try:
+        tape = wp.Tape()
+        with tape:
+            wp.launch(for_loop_grad, dim=1, inputs=[n, x, sum], device=device)
 
-    graph = wp.capture_end()
+        tape.backward(loss=sum)
+    finally:
+        graph = wp.capture_end(device)
 
     wp.capture_launch(graph)
-    wp.synchronize()
+    wp.synchronize_device(device)
 
     # ensure forward pass outputs persist
     assert_np_equal(sum.numpy(), 2.0 * np.sum(x.numpy()))
@@ -94,7 +94,7 @@ def test_for_loop_graph_grad(test, device):
     assert_np_equal(x.grad.numpy(), 2.0 * val)
 
     wp.capture_launch(graph)
-    wp.synchronize()
+    wp.synchronize_device(device)
 
 
 @wp.kernel
@@ -610,31 +610,31 @@ def test_struct_attribute_gradient(test_case, device):
     test_case.assertEqual(src.grad.numpy()[0], 5.0)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestGrad(parent):
-        pass
+class TestGrad(unittest.TestCase):
+    pass
 
-    # add_function_test(TestGrad, "test_while_loop_grad", test_while_loop_grad, devices=devices)
-    add_function_test(TestGrad, "test_for_loop_nested_for_grad", test_for_loop_nested_for_grad, devices=devices)
-    add_function_test(TestGrad, "test_scalar_grad", test_scalar_grad, devices=devices)
-    add_function_test(TestGrad, "test_for_loop_grad", test_for_loop_grad, devices=devices)
-    add_function_test(TestGrad, "test_for_loop_graph_grad", test_for_loop_graph_grad, devices=wp.get_cuda_devices())
-    add_function_test(TestGrad, "test_for_loop_nested_if_grad", test_for_loop_nested_if_grad, devices=devices)
-    add_function_test(TestGrad, "test_preserve_outputs_grad", test_preserve_outputs_grad, devices=devices)
-    add_function_test(TestGrad, "test_vector_math_grad", test_vector_math_grad, devices=devices)
-    add_function_test(TestGrad, "test_matrix_math_grad", test_matrix_math_grad, devices=devices)
-    add_function_test(TestGrad, "test_3d_math_grad", test_3d_math_grad, devices=devices)
-    add_function_test(TestGrad, "test_multi_valued_function_grad", test_multi_valued_function_grad, devices=devices)
-    add_function_test(TestGrad, "test_mesh_grad", test_mesh_grad, devices=devices)
-    add_function_test(TestGrad, "test_name_clash", test_name_clash, devices=devices)
-    add_function_test(TestGrad, "test_struct_attribute_gradient", test_struct_attribute_gradient, devices=devices)
 
-    return TestGrad
+# add_function_test(TestGrad, "test_while_loop_grad", test_while_loop_grad, devices=devices)
+add_function_test(TestGrad, "test_for_loop_nested_for_grad", test_for_loop_nested_for_grad, devices=devices)
+add_function_test(TestGrad, "test_scalar_grad", test_scalar_grad, devices=devices)
+add_function_test(TestGrad, "test_for_loop_grad", test_for_loop_grad, devices=devices)
+add_function_test(
+    TestGrad, "test_for_loop_graph_grad", test_for_loop_graph_grad, devices=get_unique_cuda_test_devices()
+)
+add_function_test(TestGrad, "test_for_loop_nested_if_grad", test_for_loop_nested_if_grad, devices=devices)
+add_function_test(TestGrad, "test_preserve_outputs_grad", test_preserve_outputs_grad, devices=devices)
+add_function_test(TestGrad, "test_vector_math_grad", test_vector_math_grad, devices=devices)
+add_function_test(TestGrad, "test_matrix_math_grad", test_matrix_math_grad, devices=devices)
+add_function_test(TestGrad, "test_3d_math_grad", test_3d_math_grad, devices=devices)
+add_function_test(TestGrad, "test_multi_valued_function_grad", test_multi_valued_function_grad, devices=devices)
+add_function_test(TestGrad, "test_mesh_grad", test_mesh_grad, devices=devices)
+add_function_test(TestGrad, "test_name_clash", test_name_clash, devices=devices)
+add_function_test(TestGrad, "test_struct_attribute_gradient", test_struct_attribute_gradient, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_grad_customs.py b/warp/tests/test_grad_customs.py
index a4d7bbf6e..eee0353c8 100644
--- a/warp/tests/test_grad_customs.py
+++ b/warp/tests/test_grad_customs.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -160,19 +160,17 @@ def test_custom_overload_grad(test, device):
     # fmt: on
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestGradCustoms(parent):
-        pass
 
-    add_function_test(TestGradCustoms, "test_custom_replay_grad", test_custom_replay_grad, devices=devices)
-    add_function_test(TestGradCustoms, "test_custom_overload_grad", test_custom_overload_grad, devices=devices)
+class TestGradCustoms(unittest.TestCase):
+    pass
 
-    return TestGradCustoms
+
+add_function_test(TestGradCustoms, "test_custom_replay_grad", test_custom_replay_grad, devices=devices)
+add_function_test(TestGradCustoms, "test_custom_overload_grad", test_custom_overload_grad, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_hash_grid.py b/warp/tests/test_hash_grid.py
index 177d1b81a..ff9e45f2c 100644
--- a/warp/tests/test_hash_grid.py
+++ b/warp/tests/test_hash_grid.py
@@ -10,8 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -73,6 +72,7 @@ def count_neighbors_reference(
 
 
 def test_hashgrid_query(test, device):
+    wp.load_module(device=device)
     rng = np.random.default_rng(123)
 
     grid = wp.HashGrid(dim_x, dim_y, dim_z, device)
@@ -134,19 +134,15 @@ def particle_grid(dim_x, dim_y, dim_z, lower, radius, jitter):
         test.assertTrue(np.array_equal(counts, counts_ref))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestHashGrid(parent):
-        pass
 
-    add_function_test(TestHashGrid, "test_hashgrid_query", test_hashgrid_query, devices=devices)
+class TestHashGrid(unittest.TestCase):
+    pass
 
-    return TestHashGrid
 
+add_function_test(TestHashGrid, "test_hashgrid_query", test_hashgrid_query, devices=devices)
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    wp.force_load()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_import.py b/warp/tests/test_import.py
index d0d60298a..11025cc36 100644
--- a/warp/tests/test_import.py
+++ b/warp/tests/test_import.py
@@ -5,20 +5,14 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
-import numpy as np
-import math
+import unittest
 
 import warp as wp
-from warp.tests.test_base import *
-
-import unittest
+import warp.tests.test_func as test_func
+from warp.tests.unittest_utils import *
 
 wp.init()
 
-# from test_func import sqr
-import warp.tests.test_func as test_func
-
 
 @wp.kernel
 def test_import_func():
@@ -30,24 +24,16 @@ def test_import_func():
     wp.expect_eq(y, 8.0)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestImport(parent):
-        pass
+class TestImport(unittest.TestCase):
+    pass
 
-    add_kernel_test(TestImport, kernel=test_import_func, name="test_import_func", dim=1, devices=devices)
 
-    return TestImport
+add_kernel_test(TestImport, kernel=test_import_func, name="test_import_func", dim=1, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    c = register(unittest.TestCase)
-    # unittest.main(verbosity=2)
-
-    wp.force_load()
-
-    loader = unittest.defaultTestLoader
-    testSuite = loader.loadTestsFromTestCase(c)
-    testSuite.debug()
+    unittest.main(verbosity=2)
diff --git a/warp/tests/test_indexedarray.py b/warp/tests/test_indexedarray.py
index 4e1c9af6f..d33be60eb 100644
--- a/warp/tests/test_indexedarray.py
+++ b/warp/tests/test_indexedarray.py
@@ -12,7 +12,7 @@
 
 import warp as wp
 from warp.tests.test_array import FillStruct
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -1106,31 +1106,29 @@ def test_indexedarray_fill_struct(test, device):
     assert_np_equal(a4.numpy(), np.zeros(a4.shape, dtype=nptype))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestIndexedArray(parent):
-        pass
 
-    add_function_test(TestIndexedArray, "test_indexedarray_1d", test_indexedarray_1d, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_2d", test_indexedarray_2d, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_3d", test_indexedarray_3d, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_4d", test_indexedarray_4d, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_mixed", test_indexedarray_mixed, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_shape", test_indexedarray_shape, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_getitem", test_indexedarray_getitem, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_slicing", test_indexedarray_slicing, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_generics", test_indexedarray_generics, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_empty", test_indexedarray_empty, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_fill_scalar", test_indexedarray_fill_scalar, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_fill_vector", test_indexedarray_fill_vector, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_fill_matrix", test_indexedarray_fill_matrix, devices=devices)
-    add_function_test(TestIndexedArray, "test_indexedarray_fill_struct", test_indexedarray_fill_struct, devices=devices)
+class TestIndexedArray(unittest.TestCase):
+    pass
 
-    return TestIndexedArray
+
+add_function_test(TestIndexedArray, "test_indexedarray_1d", test_indexedarray_1d, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_2d", test_indexedarray_2d, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_3d", test_indexedarray_3d, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_4d", test_indexedarray_4d, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_mixed", test_indexedarray_mixed, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_shape", test_indexedarray_shape, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_getitem", test_indexedarray_getitem, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_slicing", test_indexedarray_slicing, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_generics", test_indexedarray_generics, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_empty", test_indexedarray_empty, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_fill_scalar", test_indexedarray_fill_scalar, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_fill_vector", test_indexedarray_fill_vector, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_fill_matrix", test_indexedarray_fill_matrix, devices=devices)
+add_function_test(TestIndexedArray, "test_indexedarray_fill_struct", test_indexedarray_fill_struct, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_intersect.py b/warp/tests/test_intersect.py
index c19214350..1f0a64be0 100644
--- a/warp/tests/test_intersect.py
+++ b/warp/tests/test_intersect.py
@@ -1,8 +1,16 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
-import warp as wp
 import numpy as np
-from warp.tests.test_base import *
+
+import warp as wp
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -44,18 +52,16 @@ def test_intersect_tri(test, device):
     assert_np_equal(result.numpy(), np.array([0]))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestIntersect(parent):
-        pass
+class TestIntersect(unittest.TestCase):
+    pass
 
-    add_function_test(TestIntersect, "test_intersect_tri", test_intersect_tri, devices=devices)
 
-    return TestIntersect
+add_function_test(TestIntersect, "test_intersect_tri", test_intersect_tri, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_large.py b/warp/tests/test_large.py
index f35f0fde1..e32e6bdd3 100644
--- a/warp/tests/test_large.py
+++ b/warp/tests/test_large.py
@@ -1,4 +1,4 @@
-# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
 # NVIDIA CORPORATION and its licensors retain all intellectual property
 # and proprietary rights in and to this software, related documentation
 # and any modifications thereto.  Any use, reproduction, disclosure or
@@ -9,7 +9,7 @@
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -114,30 +114,28 @@ def test_large_arrays_fast(test, device):
     assert_np_equal(a1.numpy(), np.zeros_like(a1.numpy()))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestLarge(parent):
-        pass
 
-    add_function_test(
-        TestLarge, "test_large_launch_large_kernel", test_large_launch_large_kernel, devices=wp.get_cuda_devices()
-    )
+class TestLarge(unittest.TestCase):
+    pass
 
-    add_function_test(TestLarge, "test_large_launch_max_blocks", test_large_launch_max_blocks, devices=devices)
-    add_function_test(
-        TestLarge,
-        "test_large_launch_very_large_kernel",
-        test_large_launch_very_large_kernel,
-        devices=wp.get_cuda_devices(),
-    )
 
-    add_function_test(TestLarge, "test_large_arrays_fast", test_large_arrays_fast, devices=devices)
+add_function_test(
+    TestLarge, "test_large_launch_large_kernel", test_large_launch_large_kernel, devices=get_unique_cuda_test_devices()
+)
+
+add_function_test(TestLarge, "test_large_launch_max_blocks", test_large_launch_max_blocks, devices=devices)
+add_function_test(
+    TestLarge,
+    "test_large_launch_very_large_kernel",
+    test_large_launch_very_large_kernel,
+    devices=get_unique_cuda_test_devices(),
+)
 
-    return TestLarge
+add_function_test(TestLarge, "test_large_arrays_fast", test_large_arrays_fast, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_launch.py b/warp/tests/test_launch.py
index ea578adec..6a13d33b0 100644
--- a/warp/tests/test_launch.py
+++ b/warp/tests/test_launch.py
@@ -7,11 +7,10 @@
 
 import unittest
 
-# include parent path
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -306,27 +305,25 @@ def test_launch_tuple_args(test, device):
     assert_np_equal(out.numpy(), np.array((0, 3, 6, 9)))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestLaunch(parent):
-        pass
 
-    add_function_test(TestLaunch, "test_launch_1d", test1d, devices=devices)
-    add_function_test(TestLaunch, "test_launch_2d", test2d, devices=devices)
-    add_function_test(TestLaunch, "test_launch_3d", test3d, devices=devices)
-    add_function_test(TestLaunch, "test_launch_4d", test4d, devices=devices)
+class TestLaunch(unittest.TestCase):
+    pass
 
-    add_function_test(TestLaunch, "test_launch_cmd", test_launch_cmd, devices=devices)
-    add_function_test(TestLaunch, "test_launch_cmd_set_param", test_launch_cmd_set_param, devices=devices)
-    add_function_test(TestLaunch, "test_launch_cmd_set_ctype", test_launch_cmd_set_ctype, devices=devices)
-    add_function_test(TestLaunch, "test_launch_cmd_set_dim", test_launch_cmd_set_dim, devices=devices)
-    add_function_test(TestLaunch, "test_launch_cmd_empty", test_launch_cmd_empty, devices=devices)
 
-    return TestLaunch
+add_function_test(TestLaunch, "test_launch_1d", test1d, devices=devices)
+add_function_test(TestLaunch, "test_launch_2d", test2d, devices=devices)
+add_function_test(TestLaunch, "test_launch_3d", test3d, devices=devices)
+add_function_test(TestLaunch, "test_launch_4d", test4d, devices=devices)
+
+add_function_test(TestLaunch, "test_launch_cmd", test_launch_cmd, devices=devices)
+add_function_test(TestLaunch, "test_launch_cmd_set_param", test_launch_cmd_set_param, devices=devices)
+add_function_test(TestLaunch, "test_launch_cmd_set_ctype", test_launch_cmd_set_ctype, devices=devices)
+add_function_test(TestLaunch, "test_launch_cmd_set_dim", test_launch_cmd_set_dim, devices=devices)
+add_function_test(TestLaunch, "test_launch_cmd_empty", test_launch_cmd_empty, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_lerp.py b/warp/tests/test_lerp.py
index 52f0f8db8..187b2d663 100644
--- a/warp/tests/test_lerp.py
+++ b/warp/tests/test_lerp.py
@@ -5,14 +5,16 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
+import unittest
 from dataclasses import dataclass
 from typing import Any
-import unittest
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
+
+wp.init()
 
 
 @dataclass
@@ -162,8 +164,6 @@ def check_backwards(self):
     ),
 }
 
-wp.init()
-
 
 def test_lerp(test, device):
     def make_kernel_fn(data_type):
@@ -184,78 +184,78 @@ def fn(
             key=f"test_lerp_{data_type.__name__}_kernel",
         )
 
-        for test_data in TEST_DATA[data_type]:
-            a = wp.array(
-                [test_data.a],
-                dtype=data_type,
-                device=device,
-                requires_grad=True,
-            )
-            b = wp.array(
-                [test_data.b],
-                dtype=data_type,
-                device=device,
-                requires_grad=True,
-            )
-            t = wp.array(
-                [test_data.t],
-                dtype=float,
-                device=device,
-                requires_grad=True,
-            )
-            out = wp.array(
-                [0] * wp.types.type_length(data_type),
-                dtype=data_type,
-                device=device,
-                requires_grad=True,
-            )
-
-            tape = wp.Tape()
-            with tape:
-                wp.launch(
-                    kernel,
-                    dim=1,
-                    inputs=[a, b, t, out],
+        with test.subTest(data_type=data_type):
+            for test_data in TEST_DATA[data_type]:
+                a = wp.array(
+                    [test_data.a],
+                    dtype=data_type,
                     device=device,
+                    requires_grad=True,
                 )
-
-            assert_np_equal(
-                out.numpy(),
-                np.array([test_data.expected]),
-                tol=1e-6,
-            )
-
-            if test_data.check_backwards():
-                tape.backward(out)
-
-                assert_np_equal(
-                    tape.gradients[a].numpy(),
-                    np.array([test_data.expected_adj_a]),
-                    tol=1e-6,
+                b = wp.array(
+                    [test_data.b],
+                    dtype=data_type,
+                    device=device,
+                    requires_grad=True,
                 )
-                assert_np_equal(
-                    tape.gradients[b].numpy(),
-                    np.array([test_data.expected_adj_b]),
-                    tol=1e-6,
+                t = wp.array(
+                    [test_data.t],
+                    dtype=float,
+                    device=device,
+                    requires_grad=True,
+                )
+                out = wp.array(
+                    [0] * wp.types.type_length(data_type),
+                    dtype=data_type,
+                    device=device,
+                    requires_grad=True,
                 )
+
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(
+                        kernel,
+                        dim=1,
+                        inputs=[a, b, t, out],
+                        device=device,
+                    )
+
                 assert_np_equal(
-                    tape.gradients[t].numpy(),
-                    np.array([test_data.expected_adj_t]),
+                    out.numpy(),
+                    np.array([test_data.expected]),
                     tol=1e-6,
                 )
 
+                if test_data.check_backwards():
+                    tape.backward(out)
+
+                    assert_np_equal(
+                        tape.gradients[a].numpy(),
+                        np.array([test_data.expected_adj_a]),
+                        tol=1e-6,
+                    )
+                    assert_np_equal(
+                        tape.gradients[b].numpy(),
+                        np.array([test_data.expected_adj_b]),
+                        tol=1e-6,
+                    )
+                    assert_np_equal(
+                        tape.gradients[t].numpy(),
+                        np.array([test_data.expected_adj_t]),
+                        tol=1e-6,
+                    )
+
+
+devices = get_test_devices()
+
 
-def register(parent):
-    devices = get_test_devices()
+class TestLerp(unittest.TestCase):
+    pass
 
-    class TestLerp(parent):
-        pass
 
-    add_function_test(TestLerp, "test_lerp", test_lerp, devices=devices)
-    return TestLerp
+add_function_test(TestLerp, "test_lerp", test_lerp, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_lvalue.py b/warp/tests/test_lvalue.py
index 0eb7471c4..5cea866f1 100644
--- a/warp/tests/test_lvalue.py
+++ b/warp/tests/test_lvalue.py
@@ -1,11 +1,16 @@
-import warp as wp
-import numpy as np
-from warp.tests.test_base import *
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
 
 import unittest
 
+import warp as wp
+from warp.tests.unittest_utils import *
+
 wp.init()
-wp.config.mode = "debug"
 
 
 @wp.kernel
@@ -450,45 +455,39 @@ def test_swizzle(test, device):
             raise AssertionError(f"Unexpected result, got: {f} expected: {expected}")
 
 
-def test_swizzle_error_invalid_attribute(test, device):
-    v = wp.vec3(1, 2, 3)
-    with test.assertRaisesRegex(
-        AttributeError,
-        r"'vec3f' object has no attribute 'foo'$",
-    ):
-        v.foo
-
-    try:
-        v.bar = 123
-    except AttributeError:
-        test.fail()
+devices = get_test_devices()
 
 
-def register(parent):
-    devices = get_test_devices()
+class TestLValue(unittest.TestCase):
+    def test_swizzle_error_invalid_attribute(self):
+        v = wp.vec3(1, 2, 3)
+        with self.assertRaisesRegex(
+            AttributeError,
+            r"'vec3f' object has no attribute 'foo'$",
+        ):
+            v.foo
 
-    class TestLValue(parent):
-        pass
+        try:
+            v.bar = 123
+        except AttributeError:
+            self.fail()
 
-    add_function_test(TestLValue, "test_rmw_array", test_rmw_array, devices=devices)
-    add_function_test(TestLValue, "test_rmw_array_struct", test_rmw_array_struct, devices=devices)
-    add_function_test(TestLValue, "test_lookup", test_lookup, devices=devices)
-    add_function_test(TestLValue, "test_lookup2", test_lookup2, devices=devices)
-    add_function_test(TestLValue, "test_grad", test_grad, devices=devices)
-    add_function_test(TestLValue, "test_unary", test_unary, devices=devices)
-    add_function_test(TestLValue, "test_rvalue", test_rvalue, devices=devices)
-    add_function_test(TestLValue, "test_intermediate", test_intermediate, devices=devices)
-    add_function_test(TestLValue, "test_array_assign", test_array_assign, devices=devices)
-    add_function_test(TestLValue, "test_array_struct_assign", test_array_struct_assign, devices=devices)
-    add_function_test(TestLValue, "test_array_struct_struct_assign", test_array_struct_struct_assign, devices=devices)
-    add_function_test(TestLValue, "test_complex", test_complex, devices=devices)
-    add_function_test(TestLValue, "test_swizzle", test_swizzle, devices=devices)
-    add_function_test(TestLValue, "test_swizzle_error_invalid_attribute", test_swizzle_error_invalid_attribute)
 
-    return TestLValue
+add_function_test(TestLValue, "test_rmw_array", test_rmw_array, devices=devices)
+add_function_test(TestLValue, "test_rmw_array_struct", test_rmw_array_struct, devices=devices)
+add_function_test(TestLValue, "test_lookup", test_lookup, devices=devices)
+add_function_test(TestLValue, "test_lookup2", test_lookup2, devices=devices)
+add_function_test(TestLValue, "test_grad", test_grad, devices=devices)
+add_function_test(TestLValue, "test_unary", test_unary, devices=devices)
+add_function_test(TestLValue, "test_rvalue", test_rvalue, devices=devices)
+add_function_test(TestLValue, "test_intermediate", test_intermediate, devices=devices)
+add_function_test(TestLValue, "test_array_assign", test_array_assign, devices=devices)
+add_function_test(TestLValue, "test_array_struct_assign", test_array_struct_assign, devices=devices)
+add_function_test(TestLValue, "test_array_struct_struct_assign", test_array_struct_struct_assign, devices=devices)
+add_function_test(TestLValue, "test_complex", test_complex, devices=devices)
+add_function_test(TestLValue, "test_swizzle", test_swizzle, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_marching_cubes.py b/warp/tests/test_marching_cubes.py
index 2d63e8a91..5b112e2cb 100644
--- a/warp/tests/test_marching_cubes.py
+++ b/warp/tests/test_marching_cubes.py
@@ -5,14 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-# include parent path
+import unittest
+
 import numpy as np
-import math
 
 import warp as wp
-from warp.tests.test_base import *
-
-import unittest
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -39,7 +37,7 @@ def test_marching_cubes(test, device):
 
     radius = dim / 4.0
 
-    wp.launch(make_field, dim=field.shape, inputs=[field, wp.vec3(dim / 2, dim / 2, dim / 2), radius], device="cuda")
+    wp.launch(make_field, dim=field.shape, inputs=[field, wp.vec3(dim / 2, dim / 2, dim / 2), radius], device=device)
 
     iso.surface(field=field, threshold=0.0)
 
@@ -52,18 +50,16 @@ def test_marching_cubes(test, device):
     iso.resize(nx=dim * 2, ny=dim * 2, nz=dim * 2, max_verts=max_verts, max_tris=max_tris)
 
 
-def register(parent):
-    devices = ["cuda"]
+devices = get_unique_cuda_test_devices()
+
 
-    class TestMarchingCubes(parent):
-        pass
+class TestMarchingCubes(unittest.TestCase):
+    pass
 
-    add_function_test(TestMarchingCubes, "test_marching_cubes", test_marching_cubes, devices=devices)
 
-    return TestMarchingCubes
+add_function_test(TestMarchingCubes, "test_marching_cubes", test_marching_cubes, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_mat.py b/warp/tests/test_mat.py
index 321bc1cf9..ee36736c5 100644
--- a/warp/tests/test_mat.py
+++ b/warp/tests/test_mat.py
@@ -8,8 +8,9 @@
 import unittest
 
 import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -21,20 +22,8 @@
     np.byte,
 ]
 
-np_unsigned_int_types = [
-    np.uint8,
-    np.uint16,
-    np.uint32,
-    np.uint64,
-    np.ubyte,
-]
-
-np_int_types = np_signed_int_types + np_unsigned_int_types
-
 np_float_types = [np.float16, np.float32, np.float64]
 
-np_scalar_types = np_int_types + np_float_types
-
 
 def randvals(rng, shape, dtype):
     if dtype in np_float_types:
@@ -64,415 +53,9 @@ def output_select_kernel_fn(
 
     return getkernel(output_select_kernel_fn, suffix=dtype.__name__)
 
-
-def test_arrays(test, device, dtype):
-    rng = np.random.default_rng(123)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
-
-    v2_np = randvals(rng, [10, 2, 2], dtype)
-    v3_np = randvals(rng, [10, 3, 3], dtype)
-    v4_np = randvals(rng, [10, 4, 4], dtype)
-    v5_np = randvals(rng, [10, 5, 5], dtype)
-    v32_np = randvals(rng, [10, 3, 2], dtype)
-
-    v2 = wp.array(v2_np, dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(v3_np, dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(v4_np, dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(v5_np, dtype=mat55, requires_grad=True, device=device)
-    v32 = wp.array(v32_np, dtype=mat32, requires_grad=True, device=device)
-
-    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
-    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
-    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
-    assert_np_equal(v5.numpy(), v5_np, tol=1.0e-6)
-    assert_np_equal(v32.numpy(), v32_np, tol=1.0e-6)
-
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-
-    v2 = wp.array(v2_np, dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(v3_np, dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(v4_np, dtype=mat44, requires_grad=True, device=device)
-
-    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
-    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
-    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
-
-
-def test_components(test, device, dtype):
-    # test accessing matrix components from Python - this is especially important
-    # for float16, which requires special handling internally
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat23 = wp.types.matrix(shape=(2, 3), dtype=wptype)
-
-    m = mat23(1, 2, 3, 4, 5, 6)
-
-    # test __getitem__ for row vectors
-    r0 = m[0]
-    r1 = m[1]
-    test.assertEqual(r0[0], 1)
-    test.assertEqual(r0[1], 2)
-    test.assertEqual(r0[2], 3)
-    test.assertEqual(r1[0], 4)
-    test.assertEqual(r1[1], 5)
-    test.assertEqual(r1[2], 6)
-
-    # test __getitem__ for individual components
-    test.assertEqual(m[0, 0], 1)
-    test.assertEqual(m[0, 1], 2)
-    test.assertEqual(m[0, 2], 3)
-    test.assertEqual(m[1, 0], 4)
-    test.assertEqual(m[1, 1], 5)
-    test.assertEqual(m[1, 2], 6)
-
-    # test __setitem__ for row vectors
-    m[0] = [7, 8, 9]
-    m[1] = [10, 11, 12]
-    test.assertEqual(m[0, 0], 7)
-    test.assertEqual(m[0, 1], 8)
-    test.assertEqual(m[0, 2], 9)
-    test.assertEqual(m[1, 0], 10)
-    test.assertEqual(m[1, 1], 11)
-    test.assertEqual(m[1, 2], 12)
-
-    # test __setitem__ for individual components
-    m[0, 0] = 13
-    m[0, 1] = 14
-    m[0, 2] = 15
-    m[1, 0] = 16
-    m[1, 1] = 17
-    m[1, 2] = 18
-    test.assertEqual(m[0, 0], 13)
-    test.assertEqual(m[0, 1], 14)
-    test.assertEqual(m[0, 2], 15)
-    test.assertEqual(m[1, 0], 16)
-    test.assertEqual(m[1, 1], 17)
-    test.assertEqual(m[1, 2], 18)
-
-
-def test_py_arithmetic_ops(test, device, dtype):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    def make_mat(*args):
-        if wptype in wp.types.int_types:
-            # Cast to the correct integer type to simulate wrapping.
-            return tuple(tuple(wptype._type_(x).value for x in row) for row in args)
-
-        return args
-
-    def make_vec(*args):
-        if wptype in wp.types.int_types:
-            # Cast to the correct integer type to simulate wrapping.
-            return tuple(wptype._type_(x).value for x in args)
-
-        return args
-
-    mat_cls = wp.mat((3, 3), wptype)
-    vec_cls = wp.vec(3, wptype)
-
-    m = mat_cls(((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
-    test.assertSequenceEqual(+m, make_mat((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
-    test.assertSequenceEqual(-m, make_mat((1, -2, -3), (-4, 5, -6), (-7, -8, 9)))
-    test.assertSequenceEqual(m + mat_cls((5, 5, 5) * 3), make_mat((4, 7, 8), (9, 0, 11), (12, 13, -4)))
-    test.assertSequenceEqual(m - mat_cls((5, 5, 5) * 3), make_mat((-6, -3, -2), (-1, -10, 1), (2, 3, -14)))
-    test.assertSequenceEqual(m * vec_cls(5, 5, 5), make_vec(20, 25, 30))
-    test.assertSequenceEqual(m @ vec_cls(5, 5, 5), make_vec(20, 25, 30))
-    test.assertSequenceEqual(vec_cls(5, 5, 5) * m, make_vec(50, 25, 0))
-    test.assertSequenceEqual(vec_cls(5, 5, 5) @ m, make_vec(50, 25, 0))
-
-    m = mat_cls(((2, 4, 6), (8, 10, 12), (14, 16, 18)))
-    test.assertSequenceEqual(m * wptype(2), make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
-    test.assertSequenceEqual(wptype(2) * m, make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
-    test.assertSequenceEqual(m / wptype(2), make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
-    test.assertSequenceEqual(wptype(5040) / m, make_mat((2520, 1260, 840), (630, 504, 420), (360, 315, 280)))
-    test.assertSequenceEqual(m * vec_cls(5, 5, 5), make_vec(60, 150, 240))
-    test.assertSequenceEqual(m @ vec_cls(5, 5, 5), make_vec(60, 150, 240))
-    test.assertSequenceEqual(vec_cls(5, 5, 5) * m, make_vec(120, 150, 180))
-    test.assertSequenceEqual(vec_cls(5, 5, 5) @ m, make_vec(120, 150, 180))
-
-
-def test_constants(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
-
-    cm22 = wp.constant(mat22(22))
-    cm33 = wp.constant(mat33(33))
-    cm44 = wp.constant(mat44(44))
-    cm55 = wp.constant(mat55(55))
-    cm32 = wp.constant(mat32(32))
-
-    def check_matrix_constants():
-        wp.expect_eq(cm22, mat22(wptype(22)))
-        wp.expect_eq(cm33, mat33(wptype(33)))
-        wp.expect_eq(cm44, mat44(wptype(44)))
-        wp.expect_eq(cm55, mat55(wptype(55)))
-        wp.expect_eq(cm32, mat32(wptype(32)))
-
-    kernel = getkernel(check_matrix_constants, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
     wp.launch(kernel, dim=1, inputs=[])
 
 
-def test_constructors(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_scalar_mat_constructor(
-        input: wp.array(dtype=wptype),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        m2result = wptype(2) * mat22(input[0])
-        m3result = wptype(2) * mat33(input[0])
-        m4result = wptype(2) * mat44(input[0])
-        m5result = wptype(2) * mat55(input[0])
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = m2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = m3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = m4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = m5result[i, j]
-                idx = idx + 1
-
-    def check_component_mat_constructor(
-        input: wp.array(dtype=wptype),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        m2result = wptype(2) * mat22(input[0], input[1], input[2], input[3])
-        m3result = wptype(2) * mat33(
-            input[4],
-            input[5],
-            input[6],
-            input[7],
-            input[8],
-            input[9],
-            input[10],
-            input[11],
-            input[12],
-        )
-        m4result = wptype(2) * mat44(
-            input[13],
-            input[14],
-            input[15],
-            input[16],
-            input[17],
-            input[18],
-            input[19],
-            input[20],
-            input[21],
-            input[22],
-            input[23],
-            input[24],
-            input[25],
-            input[26],
-            input[27],
-            input[28],
-        )
-        m5result = wptype(2) * mat55(
-            input[29],
-            input[30],
-            input[31],
-            input[32],
-            input[33],
-            input[34],
-            input[35],
-            input[36],
-            input[37],
-            input[38],
-            input[39],
-            input[40],
-            input[41],
-            input[42],
-            input[43],
-            input[44],
-            input[45],
-            input[46],
-            input[47],
-            input[48],
-            input[49],
-            input[50],
-            input[51],
-            input[52],
-            input[53],
-        )
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = m2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = m3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = m4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = m5result[i, j]
-                idx = idx + 1
-
-    def check_vector_mat_constructor(
-        input: wp.array(dtype=wptype),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        m2result = wptype(2) * mat22(vec2(input[0], input[2]), vec2(input[1], input[3]))
-        m3result = wptype(2) * mat33(
-            vec3(input[4], input[7], input[10]),
-            vec3(input[5], input[8], input[11]),
-            vec3(input[6], input[9], input[12]),
-        )
-        m4result = wptype(2) * mat44(
-            vec4(input[13], input[17], input[21], input[25]),
-            vec4(input[14], input[18], input[22], input[26]),
-            vec4(input[15], input[19], input[23], input[27]),
-            vec4(input[16], input[20], input[24], input[28]),
-        )
-        m5result = wptype(2) * mat55(
-            vec5(input[29], input[34], input[39], input[44], input[49]),
-            vec5(input[30], input[35], input[40], input[45], input[50]),
-            vec5(input[31], input[36], input[41], input[46], input[51]),
-            vec5(input[32], input[37], input[42], input[47], input[52]),
-            vec5(input[33], input[38], input[43], input[48], input[53]),
-        )
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = m2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = m3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = m4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = m5result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_scalar_mat_constructor, suffix=dtype.__name__)
-    compkernel = getkernel(check_component_mat_constructor, suffix=dtype.__name__)
-    veckernel = getkernel(check_vector_mat_constructor, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    input = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    val = input.numpy()[0]
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * val * np.ones(2 * 2), tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * val * np.ones(3 * 3), tol=tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * val * np.ones(4 * 4), tol=tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * val * np.ones(5 * 5), tol=tol)
-
-    if dtype in np_float_types:
-        for idx in range(len(outcomponents)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(kernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-            tape.backward(loss=out)
-            test.assertEqual(tape.gradients[input].numpy()[0], 2)
-            tape.zero()
-
-    input = wp.array(randvals(rng, [2 * 2 + 3 * 3 + 4 * 4 + 5 * 5], dtype), requires_grad=True, device=device)
-
-    wp.launch(compkernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
-    assert_np_equal(2 * input.numpy(), outcomponents.numpy(), tol=10 * tol)
-
-    if dtype in np_float_types:
-        for idx in range(len(outcomponents)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(compkernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-            tape.backward(loss=out)
-            expectedgrads = np.zeros(len(input))
-            expectedgrads[idx] = 2
-            assert_np_equal(tape.gradients[input].numpy(), expectedgrads)
-            tape.zero()
-
-    wp.launch(veckernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
-    assert_np_equal(2 * input.numpy(), outcomponents.numpy(), tol=10 * tol)
-
-    if dtype in np_float_types:
-        for idx in range(len(outcomponents)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(veckernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-            tape.backward(loss=out)
-            expectedgrads = np.zeros(len(input))
-            expectedgrads[idx] = 2
-            assert_np_equal(tape.gradients[input].numpy(), expectedgrads)
-            tape.zero()
-
-
 def test_anon_constructor_error_shape_keyword_missing(test, device):
     @wp.kernel
     def kernel():
@@ -645,6 +228,47 @@ def test_tpl_ops_with_anon(test, device):
     test.assertSequenceEqual(m, ((0.0, 1.0), (2.0, 3.0)))
 
 
+def test_py_arithmetic_ops(test, device, dtype):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def make_mat(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(tuple(wptype._type_(x).value for x in row) for row in args)
+
+        return args
+
+    def make_vec(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(wptype._type_(x).value for x in args)
+
+        return args
+
+    mat_cls = wp.mat((3, 3), wptype)
+    vec_cls = wp.vec(3, wptype)
+
+    m = mat_cls(((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
+    test.assertSequenceEqual(+m, make_mat((-1, 2, 3), (4, -5, 6), (7, 8, -9)))
+    test.assertSequenceEqual(-m, make_mat((1, -2, -3), (-4, 5, -6), (-7, -8, 9)))
+    test.assertSequenceEqual(m + mat_cls((5, 5, 5) * 3), make_mat((4, 7, 8), (9, 0, 11), (12, 13, -4)))
+    test.assertSequenceEqual(m - mat_cls((5, 5, 5) * 3), make_mat((-6, -3, -2), (-1, -10, 1), (2, 3, -14)))
+    test.assertSequenceEqual(m * vec_cls(5, 5, 5), make_vec(20, 25, 30))
+    test.assertSequenceEqual(m @ vec_cls(5, 5, 5), make_vec(20, 25, 30))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) * m, make_vec(50, 25, 0))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) @ m, make_vec(50, 25, 0))
+
+    m = mat_cls(((2, 4, 6), (8, 10, 12), (14, 16, 18)))
+    test.assertSequenceEqual(m * wptype(2), make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
+    test.assertSequenceEqual(wptype(2) * m, make_mat((4, 8, 12), (16, 20, 24), (28, 32, 36)))
+    test.assertSequenceEqual(m / wptype(2), make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
+    test.assertSequenceEqual(wptype(5040) / m, make_mat((2520, 1260, 840), (630, 504, 420), (360, 315, 280)))
+    test.assertSequenceEqual(m * vec_cls(5, 5, 5), make_vec(60, 150, 240))
+    test.assertSequenceEqual(m @ vec_cls(5, 5, 5), make_vec(60, 150, 240))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) * m, make_vec(120, 150, 180))
+    test.assertSequenceEqual(vec_cls(5, 5, 5) @ m, make_vec(120, 150, 180))
+
+
 def test_quat_constructor(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
@@ -744,11 +368,11 @@ def check_mat_quat_constructor(
             idx = idx + 1
 
 
-def test_indexing(test, device, dtype, register_kernels=False):
+def test_negation(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
     tol = {
-        np.float16: 1.0e-3,
+        np.float16: 1.0e-2,
         np.float32: 1.0e-6,
         np.float64: 1.0e-8,
     }.get(dtype, 0)
@@ -761,391 +385,41 @@ def test_indexing(test, device, dtype, register_kernels=False):
 
     output_select_kernel = get_select_kernel(wptype)
 
-    def check_mat_indexing(
+    def check_mat_negation(
         m2: wp.array(dtype=mat22),
         m3: wp.array(dtype=mat33),
         m4: wp.array(dtype=mat44),
         m5: wp.array(dtype=mat55),
         outcomponents: wp.array(dtype=wptype),
     ):
+        mat2 = -m2[0]
+        mat3 = -m3[0]
+        mat4 = -m4[0]
+        mat5 = -m5[0]
+
         # multiply outputs by 2 so we've got something to backpropagate:
         idx = 0
         for i in range(2):
             for j in range(2):
-                outcomponents[idx] = wptype(2) * m2[0][i, j]
+                outcomponents[idx] = wptype(2) * mat2[i, j]
                 idx = idx + 1
 
         for i in range(3):
             for j in range(3):
-                outcomponents[idx] = wptype(2) * m3[0][i, j]
+                outcomponents[idx] = wptype(2) * mat3[i, j]
                 idx = idx + 1
 
         for i in range(4):
             for j in range(4):
-                outcomponents[idx] = wptype(2) * m4[0][i, j]
+                outcomponents[idx] = wptype(2) * mat4[i, j]
                 idx = idx + 1
 
         for i in range(5):
             for j in range(5):
-                outcomponents[idx] = wptype(2) * m5[0][i, j]
+                outcomponents[idx] = wptype(2) * mat5[i, j]
                 idx = idx + 1
 
-    kernel = getkernel(check_mat_indexing, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * m2.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * m3.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * m4.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * m5.numpy().reshape(-1), tol=tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, input in [(2, m2), (3, m3), (4, m4), (5, m5)]:
-            for i in range(dim):
-                for j in range(dim):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5], outputs=[outcomponents], device=device)
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2
-                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult)
-                    tape.zero()
-                    idx = idx + 1
-
-
-def test_equality(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    def check_mat_equality():
-        wp.expect_eq(
-            mat22(wptype(1.0), wptype(2.0), wptype(3.0), wptype(4.0)),
-            mat22(wptype(1.0), wptype(2.0), wptype(3.0), wptype(4.0)),
-        )
-        wp.expect_neq(
-            mat22(wptype(1.0), wptype(2.0), wptype(3.0), -wptype(4.0)),
-            mat22(wptype(1.0), wptype(2.0), wptype(3.0), wptype(4.0)),
-        )
-
-        wp.expect_eq(
-            mat33(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-            ),
-            mat33(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-            ),
-        )
-        wp.expect_neq(
-            mat33(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-            ),
-            mat33(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                -wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-            ),
-        )
-
-        wp.expect_eq(
-            mat44(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-            ),
-            mat44(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-            ),
-        )
-
-        wp.expect_neq(
-            mat44(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-            ),
-            mat44(
-                -wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-            ),
-        )
-
-        wp.expect_eq(
-            mat55(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-                wptype(17.0),
-                wptype(18.0),
-                wptype(19.0),
-                wptype(20.0),
-                wptype(21.0),
-                wptype(22.0),
-                wptype(23.0),
-                wptype(24.0),
-                wptype(25.0),
-            ),
-            mat55(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-                wptype(17.0),
-                wptype(18.0),
-                wptype(19.0),
-                wptype(20.0),
-                wptype(21.0),
-                wptype(22.0),
-                wptype(23.0),
-                wptype(24.0),
-                wptype(25.0),
-            ),
-        )
-
-        wp.expect_neq(
-            mat55(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-                wptype(17.0),
-                wptype(18.0),
-                wptype(19.0),
-                wptype(20.0),
-                wptype(21.0),
-                wptype(22.0),
-                wptype(23.0),
-                wptype(24.0),
-                wptype(25.0),
-            ),
-            mat55(
-                wptype(1.0),
-                wptype(2.0),
-                wptype(3.0),
-                wptype(4.0),
-                wptype(5.0),
-                wptype(6.0),
-                wptype(7.0),
-                wptype(8.0),
-                wptype(9.0),
-                wptype(10.0),
-                wptype(11.0),
-                wptype(12.0),
-                wptype(13.0),
-                wptype(14.0),
-                wptype(15.0),
-                wptype(16.0),
-                -wptype(17.0),
-                wptype(18.0),
-                wptype(19.0),
-                wptype(20.0),
-                wptype(21.0),
-                wptype(22.0),
-                wptype(23.0),
-                wptype(24.0),
-                wptype(25.0),
-            ),
-        )
-
-    kernel = getkernel(check_mat_equality, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    wp.launch(kernel, dim=1, inputs=[], outputs=[], device=device)
-
-
-def test_negation(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_negation(
-        m2: wp.array(dtype=mat22),
-        m3: wp.array(dtype=mat33),
-        m4: wp.array(dtype=mat44),
-        m5: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        mat2 = -m2[0]
-        mat3 = -m3[0]
-        mat4 = -m4[0]
-        mat5 = -m5[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * mat2[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * mat3[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * mat4[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * mat5[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_negation, suffix=dtype.__name__)
+    kernel = getkernel(check_mat_negation, suffix=dtype.__name__)
 
     if register_kernels:
         return
@@ -1178,1410 +452,8 @@ def check_mat_negation(
                     tape.backward(loss=out)
                     expectedresult = np.zeros((dim, dim), dtype=dtype)
                     expectedresult[i, j] = -2
-                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult)
-                    tape.zero()
-                    idx = idx + 1
-
-
-def test_transpose(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_transpose(
-        m2: wp.array(dtype=mat22),
-        m3: wp.array(dtype=mat33),
-        m4: wp.array(dtype=mat44),
-        m5: wp.array(dtype=mat55),
-        m32: wp.array(dtype=mat32),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        mat2 = wptype(2) * wp.transpose(m2[0])
-        mat3 = wptype(2) * wp.transpose(m3[0])
-        mat4 = wptype(2) * wp.transpose(m4[0])
-        mat5 = wptype(2) * wp.transpose(m5[0])
-        mat32 = wptype(2) * wp.transpose(m32[0])
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = mat2[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = mat3[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = mat4[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = mat5[i, j]
-                idx = idx + 1
-
-        for i in range(2):
-            for j in range(3):
-                outcomponents[idx] = mat32[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_transpose, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    m32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 2 * 3, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * m2.numpy()[0].T.reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * m3.numpy()[0].T.reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * m4.numpy()[0].T.reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * m5.numpy()[0].T.reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[54:], 2 * m32.numpy()[0].T.reshape(-1), tol=tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for input in [m2, m3, m4, m5]:
-            for i in range(input.dtype._shape_[0]):
-                for j in range(input.dtype._shape_[1]):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((input.dtype._shape_[1], input.dtype._shape_[0]), dtype=dtype)
-                    expectedresult[j, i] = 2
-                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult)
-                    tape.zero()
-                    idx = idx + 1
-
-
-def test_scalar_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_scalar_mul(
-        s: wp.array(dtype=wptype),
-        m2: wp.array(dtype=mat22),
-        m3: wp.array(dtype=mat33),
-        m4: wp.array(dtype=mat44),
-        m5: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-        outcomponents_rightmul: wp.array(dtype=wptype),
-    ):
-        m2result = s[0] * m2[0]
-        m3result = s[0] * m3[0]
-        m4result = s[0] * m4[0]
-        m5result = s[0] * m5[0]
-
-        m2resultright = m2[0] * s[0]
-        m3resultright = m3[0] * s[0]
-        m4resultright = m4[0] * s[0]
-        m5resultright = m5[0] * s[0]
-
-        m2result_2 = s[0] * m2[0]
-        m3result_2 = s[0] * m3[0]
-        m4result_2 = s[0] * m4[0]
-        m5result_2 = s[0] * m5[0]
-
-        m2resultright_2 = m2[0] * s[0]
-        m3resultright_2 = m3[0] * s[0]
-        m4resultright_2 = m4[0] * s[0]
-        m5resultright_2 = m5[0] * s[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * m2result[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m2resultright[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * m3result[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m3resultright[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * m4result[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m4resultright[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * m5result[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m5resultright[i, j]
-                idx = idx + 1
-
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * m2result_2[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m2resultright_2[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * m3result_2[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m3resultright_2[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * m4result_2[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m4resultright_2[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * m5result_2[i, j]
-                outcomponents_rightmul[idx] = wptype(2) * m5resultright_2[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_scalar_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * (2 * 2 + 3 * 3 + 4 * 4 + 5 * 5), dtype=wptype, requires_grad=True, device=device)
-    outcomponents_rightmul = wp.zeros(
-        2 * (2 * 2 + 3 * 3 + 4 * 4 + 5 * 5), dtype=wptype, requires_grad=True, device=device
-    )
-
-    wp.launch(kernel, dim=1, inputs=[s, m2, m3, m4, m5], outputs=[outcomponents, outcomponents_rightmul], device=device)
-
-    sval = s.numpy()[0]
-    assert_np_equal(outcomponents.numpy()[:4], 2 * sval * m2.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
-
-    assert_np_equal(outcomponents_rightmul.numpy()[:4], 2 * sval * m2.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents_rightmul.numpy()[4:13], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents_rightmul.numpy()[13:29], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents_rightmul.numpy()[29:54], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
-
-    assert_np_equal(outcomponents.numpy()[54:58], 2 * sval * m2.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[58:67], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[67:83], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[83:108], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
-
-    assert_np_equal(outcomponents_rightmul.numpy()[54:58], 2 * sval * m2.numpy().reshape(-1), tol=tol)
-    assert_np_equal(outcomponents_rightmul.numpy()[58:67], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents_rightmul.numpy()[67:83], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
-    assert_np_equal(outcomponents_rightmul.numpy()[83:108], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, input in [(2, m2), (3, m3), (4, m4), (5, m5)]:
-            for i in range(dim):
-                for j in range(dim):
-                    # test left mul gradient:
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[s, m2, m3, m4, m5],
-                            outputs=[outcomponents, outcomponents_rightmul],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2 * sval
-                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult, tol=10 * tol)
-                    assert_np_equal(tape.gradients[s].numpy()[0], 2 * input.numpy()[0, i, j], tol=10 * tol)
-                    tape.zero()
-
-                    # test right mul gradient:
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[s, m2, m3, m4, m5],
-                            outputs=[outcomponents, outcomponents_rightmul],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel,
-                            dim=1,
-                            inputs=[outcomponents_rightmul, idx],
-                            outputs=[out],
-                            device=device,
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2 * sval
-                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult, tol=10 * tol)
-                    assert_np_equal(tape.gradients[s].numpy()[0], 2 * input.numpy()[0, i, j], tol=10 * tol)
-                    tape.zero()
-
-                    idx = idx + 1
-
-
-def test_matvec_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 2.0e-2,
-        np.float32: 5.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_vec_mul(
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v32: wp.array(dtype=vec2),
-        m2: wp.array(dtype=mat22),
-        m3: wp.array(dtype=mat33),
-        m4: wp.array(dtype=mat44),
-        m5: wp.array(dtype=mat55),
-        m32: wp.array(dtype=mat32),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        v2result = m2[0] * v2[0]
-        v3result = m3[0] * v3[0]
-        v4result = m4[0] * v4[0]
-        v5result = m5[0] * v5[0]
-        v32result = m32[0] * v32[0]
-        v2result_2 = m2[0] @ v2[0]
-        v3result_2 = m3[0] @ v3[0]
-        v4result_2 = m4[0] @ v4[0]
-        v5result_2 = m5[0] @ v5[0]
-        v32result_2 = m32[0] @ v32[0]
-
-        idx = 0
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        for i in range(2):
-            outcomponents[idx] = wptype(2) * v2result[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v3result[i]
-            idx = idx + 1
-
-        for i in range(4):
-            outcomponents[idx] = wptype(2) * v4result[i]
-            idx = idx + 1
-
-        for i in range(5):
-            outcomponents[idx] = wptype(2) * v5result[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v32result[i]
-            idx = idx + 1
-
-        for i in range(2):
-            outcomponents[idx] = wptype(2) * v2result_2[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v3result_2[i]
-            idx = idx + 1
-
-        for i in range(4):
-            outcomponents[idx] = wptype(2) * v4result_2[i]
-            idx = idx + 1
-
-        for i in range(5):
-            outcomponents[idx] = wptype(2) * v5result_2[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v32result_2[i]
-            idx = idx + 1
-
-    kernel = getkernel(check_mat_vec_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
-    v32 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    m32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * (2 + 3 + 4 + 5 + 3), dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:2], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[2:5], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[5:9], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[9:14], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[14:17], 2 * np.matmul(m32.numpy()[0], v32.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[17:19], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[19:22], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[22:26], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[26:31], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[31:34], 2 * np.matmul(m32.numpy()[0], v32.numpy()[0]), tol=5 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, invec, inmat in [(2, v2, m2), (3, v3, m3), (4, v4, m4), (5, v5, m5), (3, v32, m32)]:
-            for i in range(dim):
-                tape = wp.Tape()
-                with tape:
-                    wp.launch(
-                        kernel,
-                        dim=1,
-                        inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32],
-                        outputs=[outcomponents],
-                        device=device,
-                    )
-                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-                tape.backward(loss=out)
-
-                assert_np_equal(tape.gradients[invec].numpy()[0], 2 * inmat.numpy()[0, i, :], tol=2 * tol)
-                expectedresult = np.zeros(inmat.dtype._shape_, dtype=dtype)
-                expectedresult[i, :] = 2 * invec.numpy()[0]
-                assert_np_equal(tape.gradients[inmat].numpy()[0], expectedresult, tol=2 * tol)
-
-                tape.zero()
-
-                idx = idx + 1
-
-
-def test_vecmat_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 2.0e-2,
-        np.float32: 5.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat23 = wp.types.matrix(shape=(2, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_vec_mat_mul(
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v32: wp.array(dtype=vec2),
-        m2: wp.array(dtype=mat22),
-        m3: wp.array(dtype=mat33),
-        m4: wp.array(dtype=mat44),
-        m5: wp.array(dtype=mat55),
-        m23: wp.array(dtype=mat23),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        v2result = v2[0] * m2[0]
-        v3result = v3[0] * m3[0]
-        v4result = v4[0] * m4[0]
-        v5result = v5[0] * m5[0]
-        v32result = v32[0] * m23[0]
-        v2result_2 = v2[0] @ m2[0]
-        v3result_2 = v3[0] @ m3[0]
-        v4result_2 = v4[0] @ m4[0]
-        v5result_2 = v5[0] @ m5[0]
-        v32result_2 = v32[0] @ m23[0]
-
-        idx = 0
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        for i in range(2):
-            outcomponents[idx] = wptype(2) * v2result[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v3result[i]
-            idx = idx + 1
-
-        for i in range(4):
-            outcomponents[idx] = wptype(2) * v4result[i]
-            idx = idx + 1
-
-        for i in range(5):
-            outcomponents[idx] = wptype(2) * v5result[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v32result[i]
-            idx = idx + 1
-
-        for i in range(2):
-            outcomponents[idx] = wptype(2) * v2result_2[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v3result_2[i]
-            idx = idx + 1
-
-        for i in range(4):
-            outcomponents[idx] = wptype(2) * v4result_2[i]
-            idx = idx + 1
-
-        for i in range(5):
-            outcomponents[idx] = wptype(2) * v5result_2[i]
-            idx = idx + 1
-
-        for i in range(3):
-            outcomponents[idx] = wptype(2) * v32result_2[i]
-            idx = idx + 1
-
-    kernel = getkernel(check_vec_mat_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
-    v32 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    m23 = wp.array(randvals(rng, [1, 2, 3], dtype), dtype=mat23, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * (2 + 3 + 4 + 5 + 3), dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m23], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:2], 2 * np.matmul(v2.numpy()[0], m2.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[2:5], 2 * np.matmul(v3.numpy()[0], m3.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[5:9], 2 * np.matmul(v4.numpy()[0], m4.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[9:14], 2 * np.matmul(v5.numpy()[0], m5.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[14:17], 2 * np.matmul(v32.numpy()[0], m23.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[17:19], 2 * np.matmul(v2.numpy()[0], m2.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[19:22], 2 * np.matmul(v3.numpy()[0], m3.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[22:26], 2 * np.matmul(v4.numpy()[0], m4.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[26:31], 2 * np.matmul(v5.numpy()[0], m5.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[31:34], 2 * np.matmul(v32.numpy()[0], m23.numpy()[0]), tol=5 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, inmat, invec in [(2, m2, v2), (3, m3, v3), (4, m4, v4), (5, m5, v5), (3, m23, v32)]:
-            for i in range(dim):
-                tape = wp.Tape()
-                with tape:
-                    wp.launch(
-                        kernel,
-                        dim=1,
-                        inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m23],
-                        outputs=[outcomponents],
-                        device=device,
-                    )
-                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-                tape.backward(loss=out)
-
-                assert_np_equal(tape.gradients[invec].numpy()[0], 2 * inmat.numpy()[0, :, i], tol=2 * tol)
-                expectedresult = np.zeros(inmat.dtype._shape_, dtype=dtype)
-                expectedresult[:, i] = 2 * invec.numpy()[0]
-                assert_np_equal(tape.gradients[inmat].numpy()[0], expectedresult, tol=2 * tol)
-
-                tape.zero()
-
-                idx = idx + 1
-
-
-def test_matmat_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 2.0e-2,
-        np.float32: 5.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_mat_mul(
-        a2: wp.array(dtype=mat22),
-        a3: wp.array(dtype=mat33),
-        a4: wp.array(dtype=mat44),
-        a5: wp.array(dtype=mat55),
-        a32: wp.array(dtype=mat32),
-        b2: wp.array(dtype=mat22),
-        b3: wp.array(dtype=mat33),
-        b4: wp.array(dtype=mat44),
-        b5: wp.array(dtype=mat55),
-        b32: wp.array(dtype=mat32),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        c2result = b2[0] * a2[0]
-        c3result = b3[0] * a3[0]
-        c4result = b4[0] * a4[0]
-        c5result = b5[0] * a5[0]
-        c32result = b32[0] * a2[0]
-        c32result2 = b3[0] * a32[0]
-        c2result_2 = b2[0] @ a2[0]
-        c3result_2 = b3[0] @ a3[0]
-        c4result_2 = b4[0] @ a4[0]
-        c5result_2 = b5[0] @ a5[0]
-        c32result_2 = b32[0] @ a2[0]
-        c32result2_2 = b3[0] @ a32[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * c2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * c3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * c4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * c5result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * c32result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * c32result2[i, j]
-                idx = idx + 1
-
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * c2result_2[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * c3result_2[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * c4result_2[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * c5result_2[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * c32result_2[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * c32result2_2[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_mat_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    v32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    m32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
-    outcomponents = wp.zeros(
-        2 * (2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2 + 3 * 2), dtype=wptype, requires_grad=True, device=device
-    )
-
-    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=2 * tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[54:60], 2 * np.matmul(m32.numpy()[0], v2.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[60:66], 2 * np.matmul(m3.numpy()[0], v32.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[66:70], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[70:79], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
-    assert_np_equal(outcomponents.numpy()[79:95], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=2 * tol)
-    assert_np_equal(outcomponents.numpy()[95:120], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[120:126], 2 * np.matmul(m32.numpy()[0], v2.numpy()[0]), tol=5 * tol)
-    assert_np_equal(outcomponents.numpy()[126:132], 2 * np.matmul(m3.numpy()[0], v32.numpy()[0]), tol=5 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for v, m in [(v2, m2), (v3, m3), (v4, m4), (v5, m5), (v2, m32), (v32, m3)]:
-            rows, cols = m.dtype._shape_[0], v.dtype._shape_[1]
-            for i in range(rows):
-                for j in range(cols):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32],
-                            outputs=[outcomponents],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-
-                    expected = np.zeros(v.dtype._shape_, dtype=dtype)
-                    expected[:, j] = 2 * m.numpy()[0, i, :]
-                    assert_np_equal(tape.gradients[v].numpy()[0], expected, tol=10 * tol)
-
-                    expected = np.zeros(m.dtype._shape_, dtype=dtype)
-                    expected[i, :] = 2 * v.numpy()[0, :, j]
-                    assert_np_equal(tape.gradients[m].numpy()[0], expected, tol=10 * tol)
-
-                    tape.zero()
-                    idx = idx + 1
-
-
-def test_cw_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_cw_mul(
-        s2: wp.array(dtype=mat22),
-        s3: wp.array(dtype=mat33),
-        s4: wp.array(dtype=mat44),
-        s5: wp.array(dtype=mat55),
-        v2: wp.array(dtype=mat22),
-        v3: wp.array(dtype=mat33),
-        v4: wp.array(dtype=mat44),
-        v5: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        v2result = wptype(2) * wp.cw_mul(v2[0], s2[0])
-        v3result = wptype(2) * wp.cw_mul(v3[0], s3[0])
-        v4result = wptype(2) * wp.cw_mul(v4[0], s4[0])
-        v5result = wptype(2) * wp.cw_mul(v5[0], s5[0])
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = v2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = v3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = v4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = v5result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_cw_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(
-        kernel,
-        dim=1,
-        inputs=[
-            s2,
-            s3,
-            s4,
-            s5,
-            v2,
-            v3,
-            v4,
-            v5,
-        ],
-        outputs=[outcomponents],
-        device=device,
-    )
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() * s2.numpy()).reshape(-1), tol=50 * tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() * s3.numpy()).reshape(-1), tol=50 * tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() * s4.numpy()).reshape(-1), tol=50 * tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() * s5.numpy()).reshape(-1), tol=50 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, in1, in2 in [(2, s2, v2), (3, s3, v3), (4, s4, v4), (5, s5, v5)]:
-            for i in range(dim):
-                for j in range(dim):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[
-                                s2,
-                                s3,
-                                s4,
-                                s5,
-                                v2,
-                                v3,
-                                v4,
-                                v5,
-                            ],
-                            outputs=[outcomponents],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2 * in1.numpy()[0][i, j]
-                    assert_np_equal(tape.gradients[in2].numpy()[0], expectedresult, tol=5 * tol)
-                    expectedresult[i, j] = 2 * in2.numpy()[0][i, j]
-                    assert_np_equal(tape.gradients[in1].numpy()[0], expectedresult, tol=5 * tol)
-                    tape.zero()
-
-                    idx = idx + 1
-
-
-def test_cw_division(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_cw_div(
-        s2: wp.array(dtype=mat22),
-        s3: wp.array(dtype=mat33),
-        s4: wp.array(dtype=mat44),
-        s5: wp.array(dtype=mat55),
-        v2: wp.array(dtype=mat22),
-        v3: wp.array(dtype=mat33),
-        v4: wp.array(dtype=mat44),
-        v5: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        v2result = wptype(2) * wp.cw_div(v2[0], s2[0])
-        v3result = wptype(2) * wp.cw_div(v3[0], s3[0])
-        v4result = wptype(2) * wp.cw_div(v4[0], s4[0])
-        v5result = wptype(2) * wp.cw_div(v5[0], s5[0])
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = v2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = v3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = v4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = v5result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_cw_div, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = randvals(rng, [1, 2, 2], dtype)
-    s3 = randvals(rng, [1, 3, 3], dtype)
-    s4 = randvals(rng, [1, 4, 4], dtype)
-    s5 = randvals(rng, [1, 5, 5], dtype)
-
-    # set denominators to 1 if their magnitudes are small
-    # to prevent divide by zero, or overflows if we're testing
-    # float16:
-    s2[np.abs(s2) < 1.0e-2] = 1
-    s3[np.abs(s3) < 1.0e-2] = 1
-    s4[np.abs(s4) < 1.0e-2] = 1
-    s5[np.abs(s5) < 1.0e-2] = 1
-
-    s2 = wp.array(s2, dtype=mat22, requires_grad=True, device=device)
-    s3 = wp.array(s3, dtype=mat33, requires_grad=True, device=device)
-    s4 = wp.array(s4, dtype=mat44, requires_grad=True, device=device)
-    s5 = wp.array(s5, dtype=mat55, requires_grad=True, device=device)
-
-    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(
-        kernel,
-        dim=1,
-        inputs=[
-            s2,
-            s3,
-            s4,
-            s5,
-            v2,
-            v3,
-            v4,
-            v5,
-        ],
-        outputs=[outcomponents],
-        device=device,
-    )
-
-    if dtype in np_float_types:
-        assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() / s2.numpy()).reshape(-1), tol=50 * tol)
-        assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() / s3.numpy()).reshape(-1), tol=50 * tol)
-        assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() / s4.numpy()).reshape(-1), tol=50 * tol)
-        assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() / s5.numpy()).reshape(-1), tol=50 * tol)
-    else:
-        assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() // s2.numpy()).reshape(-1), tol=50 * tol)
-        assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() // s3.numpy()).reshape(-1), tol=50 * tol)
-        assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() // s4.numpy()).reshape(-1), tol=50 * tol)
-        assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() // s5.numpy()).reshape(-1), tol=50 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, s, v in [(2, s2, v2), (3, s3, v3), (4, s4, v4), (5, s5, v5)]:
-            for i in range(dim):
-                for j in range(dim):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[
-                                s2,
-                                s3,
-                                s4,
-                                s5,
-                                v2,
-                                v3,
-                                v4,
-                                v5,
-                            ],
-                            outputs=[outcomponents],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-
-                    # y = v/s
-                    # dy/dv = 1.0/s
-                    # dy/ds = -v/s^2
-
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2.0 / (s.numpy()[0, i, j])
-                    assert_np_equal(tape.gradients[v].numpy()[0], expectedresult, tol=50 * tol)
-                    expectedresult[i, j] = -2.0 * v.numpy()[0, i, j] / (s.numpy()[0, i, j] ** 2)
-                    assert_np_equal(
-                        tape.gradients[s].numpy()[0], expectedresult, tol=abs(outcomponents.numpy()[idx]) * 50 * tol
-                    )
-                    tape.zero()
-
-                    idx = idx + 1
-
-
-def test_outer_product(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_outer_product(
-        s2: wp.array(dtype=vec2),
-        s3: wp.array(dtype=vec3),
-        s4: wp.array(dtype=vec4),
-        s5: wp.array(dtype=vec5),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        m22result = wptype(2) * wp.outer(s2[0], v2[0])
-        m33result = wptype(2) * wp.outer(s3[0], v3[0])
-        m44result = wptype(2) * wp.outer(s4[0], v4[0])
-        m55result = wptype(2) * wp.outer(s5[0], v5[0])
-        m25result = wptype(2) * wp.outer(s2[0], v5[0])
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = m22result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = m33result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = m44result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = m55result[i, j]
-                idx = idx + 1
-
-        for i in range(2):
-            for j in range(5):
-                outcomponents[idx] = m25result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_outer_product, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 2 * 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[s2, s3, s4, s5, v2, v3, v4, v5], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * s2.numpy()[0, :, None] * v2.numpy()[0, None, :], tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * s3.numpy()[0, :, None] * v3.numpy()[0, None, :], tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * s4.numpy()[0, :, None] * v4.numpy()[0, None, :], tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * s5.numpy()[0, :, None] * v5.numpy()[0, None, :], tol=10 * tol)
-    assert_np_equal(outcomponents.numpy()[54:], 2 * s2.numpy()[0, :, None] * v5.numpy()[0, None, :], tol=10 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for s, v in [(s2, v2), (s3, v3), (s4, v4), (s5, v5), (s2, v5)]:
-            rows = s.dtype._length_
-            cols = v.dtype._length_
-            for i in range(rows):
-                for j in range(cols):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[
-                                s2,
-                                s3,
-                                s4,
-                                s5,
-                                v2,
-                                v3,
-                                v4,
-                                v5,
-                            ],
-                            outputs=[outcomponents],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-
-                    # this component's gonna be s_i * v_j, so its s gradient is gonna be nozero
-                    # at the ith component and its v gradient will be nonzero at the jth component:
-
-                    expectedresult = np.zeros((rows), dtype=dtype)
-                    expectedresult[i] = 2 * v.numpy()[0, j]
-                    assert_np_equal(tape.gradients[s].numpy()[0], expectedresult, tol=10 * tol)
-
-                    expectedresult = np.zeros((cols), dtype=dtype)
-                    expectedresult[j] = 2 * s.numpy()[0, i]
-                    assert_np_equal(tape.gradients[v].numpy()[0], expectedresult, tol=10 * tol)
-                    tape.zero()
-
-                    idx = idx + 1
-
-
-def test_scalar_division(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_scalar_div(
-        s: wp.array(dtype=wptype),
-        m2: wp.array(dtype=mat22),
-        m3: wp.array(dtype=mat33),
-        m4: wp.array(dtype=mat44),
-        m5: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        m2result = m2[0] / s[0]
-        m3result = m3[0] / s[0]
-        m4result = m4[0] / s[0]
-        m5result = m5[0] / s[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * m2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * m3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * m4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * m5result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_scalar_div, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[s, m2, m3, m4, m5], outputs=[outcomponents], device=device)
-
-    sval = s.numpy()[0]
-    if dtype in np_float_types:
-        assert_np_equal(outcomponents.numpy()[:4], 2 * m2.numpy().reshape(-1) / sval, tol=tol)
-        assert_np_equal(outcomponents.numpy()[4:13], 2 * m3.numpy().reshape(-1) / sval, tol=10 * tol)
-        assert_np_equal(outcomponents.numpy()[13:29], 2 * m4.numpy().reshape(-1) / sval, tol=10 * tol)
-        assert_np_equal(outcomponents.numpy()[29:54], 2 * m5.numpy().reshape(-1) / sval, tol=10 * tol)
-    else:
-        assert_np_equal(outcomponents.numpy()[:4], 2 * (m2.numpy().reshape(-1) // sval), tol=tol)
-        assert_np_equal(outcomponents.numpy()[4:13], 2 * (m3.numpy().reshape(-1) // sval), tol=10 * tol)
-        assert_np_equal(outcomponents.numpy()[13:29], 2 * (m4.numpy().reshape(-1) // sval), tol=10 * tol)
-        assert_np_equal(outcomponents.numpy()[29:54], 2 * (m5.numpy().reshape(-1) // sval), tol=10 * tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, input in [(2, m2), (3, m3), (4, m4), (5, m5)]:
-            for i in range(dim):
-                for j in range(dim):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(kernel, dim=1, inputs=[s, m2, m3, m4, m5], outputs=[outcomponents], device=device)
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2.0 / sval
-                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult, tol=10 * tol)
-                    assert_np_equal(
-                        tape.gradients[s].numpy()[0], -2 * input.numpy()[0, i, j] / (sval * sval), tol=10 * tol
-                    )
-                    tape.zero()
-
-                    idx = idx + 1
-
-
-def test_addition(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 2.0e-2,
-        np.float32: 5.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_add(
-        s2: wp.array(dtype=mat22),
-        s3: wp.array(dtype=mat33),
-        s4: wp.array(dtype=mat44),
-        s5: wp.array(dtype=mat55),
-        v2: wp.array(dtype=mat22),
-        v3: wp.array(dtype=mat33),
-        v4: wp.array(dtype=mat44),
-        v5: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        v2result = v2[0] + s2[0]
-        v3result = v3[0] + s3[0]
-        v4result = v4[0] + s4[0]
-        v5result = v5[0] + s5[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                outcomponents[idx] = wptype(2) * v2result[i, j]
-                idx = idx + 1
-
-        for i in range(3):
-            for j in range(3):
-                outcomponents[idx] = wptype(2) * v3result[i, j]
-                idx = idx + 1
-
-        for i in range(4):
-            for j in range(4):
-                outcomponents[idx] = wptype(2) * v4result[i, j]
-                idx = idx + 1
-
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = wptype(2) * v5result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_add, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(
-        kernel,
-        dim=1,
-        inputs=[
-            s2,
-            s3,
-            s4,
-            s5,
-            v2,
-            v3,
-            v4,
-            v5,
-        ],
-        outputs=[outcomponents],
-        device=device,
-    )
-
-    assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() + s2.numpy()).reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() + s3.numpy()).reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() + s4.numpy()).reshape(-1), tol=tol)
-    assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() + s5.numpy()).reshape(-1), tol=tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for dim, in1, in2 in [(2, s2, v2), (3, s3, v3), (4, s4, v4), (5, s5, v5)]:
-            for i in range(dim):
-                for j in range(dim):
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            kernel,
-                            dim=1,
-                            inputs=[
-                                s2,
-                                s3,
-                                s4,
-                                s5,
-                                v2,
-                                v3,
-                                v4,
-                                v5,
-                            ],
-                            outputs=[outcomponents],
-                            device=device,
-                        )
-                        wp.launch(
-                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
-                        )
-                    tape.backward(loss=out)
-                    expectedresult = np.zeros((dim, dim), dtype=dtype)
-                    expectedresult[i, j] = 2
-                    assert_np_equal(tape.gradients[in2].numpy()[0], expectedresult, tol=10 * tol)
-                    expectedresult[i, j] = 2
-                    assert_np_equal(tape.gradients[in1].numpy()[0], expectedresult, tol=10 * tol)
+                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult)
                     tape.zero()
-
                     idx = idx + 1
 
 
@@ -2715,129 +587,6 @@ def check_mat_sub(
                     idx = idx + 1
 
 
-def test_ddot(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    def check_mat_dot(
-        s2: wp.array(dtype=mat22),
-        s3: wp.array(dtype=mat33),
-        s4: wp.array(dtype=mat44),
-        s5: wp.array(dtype=mat55),
-        v2: wp.array(dtype=mat22),
-        v3: wp.array(dtype=mat33),
-        v4: wp.array(dtype=mat44),
-        v5: wp.array(dtype=mat55),
-        dot2: wp.array(dtype=wptype),
-        dot3: wp.array(dtype=wptype),
-        dot4: wp.array(dtype=wptype),
-        dot5: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        dot2[0] = wptype(2) * wp.ddot(v2[0], s2[0])
-        dot3[0] = wptype(2) * wp.ddot(v3[0], s3[0])
-        dot4[0] = wptype(2) * wp.ddot(v4[0], s4[0])
-        dot5[0] = wptype(2) * wp.ddot(v5[0], s5[0])
-
-    kernel = getkernel(check_mat_dot, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    dot2 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    dot3 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    dot4 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    dot5 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s2,
-                s3,
-                s4,
-                s5,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[dot2, dot3, dot4, dot5],
-            device=device,
-        )
-
-    assert_np_equal(dot2.numpy()[0], 2 * (v2.numpy() * s2.numpy()).sum(), tol=10 * tol)
-    assert_np_equal(dot3.numpy()[0], 2 * (v3.numpy() * s3.numpy()).sum(), tol=10 * tol)
-    assert_np_equal(dot4.numpy()[0], 2 * (v4.numpy() * s4.numpy()).sum(), tol=50 * tol)
-    assert_np_equal(dot5.numpy()[0], 2 * (v5.numpy() * s5.numpy()).sum(), tol=200 * tol)
-
-    if dtype in np_float_types:
-        tape.backward(loss=dot2)
-        sgrads = tape.gradients[s2].numpy()[0]
-        expected_grads = 2.0 * v2.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v2].numpy()[0]
-        expected_grads = 2.0 * s2.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
-
-        tape.zero()
-
-        tape.backward(loss=dot3)
-        sgrads = tape.gradients[s3].numpy()[0]
-        expected_grads = 2.0 * v3.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v3].numpy()[0]
-        expected_grads = 2.0 * s3.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
-
-        tape.zero()
-
-        tape.backward(loss=dot4)
-        sgrads = tape.gradients[s4].numpy()[0]
-        expected_grads = 2.0 * v4.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v4].numpy()[0]
-        expected_grads = 2.0 * s4.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
-
-        tape.zero()
-
-        tape.backward(loss=dot5)
-        sgrads = tape.gradients[s5].numpy()[0]
-        expected_grads = 2.0 * v5.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v5].numpy()[0]
-        expected_grads = 2.0 * s5.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
-
-        tape.zero()
-
-
 def test_determinant(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
@@ -2978,268 +727,119 @@ def check_mat_det(
                         det4,
                     ],
                     device=device,
-                )
-                dplus = det3.numpy()[0]
-                v3test[0, i, j] -= 2.0 * dx
-                wp.launch(
-                    kernel,
-                    dim=1,
-                    inputs=[
-                        v2,
-                        wp.array(v3test, dtype=v3.dtype, requires_grad=True, device=device),
-                        v4,
-                    ],
-                    outputs=[
-                        det2,
-                        det3,
-                        det4,
-                    ],
-                    device=device,
-                )
-                dminus = det3.numpy()[0]
-                assert_np_equal((dplus - dminus) / (2.0 * dx * dplus), v3grads[i, j] / dplus, tol=fdtol)
-
-        for i in range(4):
-            for j in range(4):
-                v4test = v4.numpy()
-                v4test[0, i, j] += dx
-                wp.launch(
-                    kernel,
-                    dim=1,
-                    inputs=[
-                        v2,
-                        v3,
-                        wp.array(v4test, dtype=v4.dtype, requires_grad=True, device=device),
-                    ],
-                    outputs=[
-                        det2,
-                        det3,
-                        det4,
-                    ],
-                    device=device,
-                )
-                dplus = det4.numpy()[0]
-                v4test[0, i, j] -= 2.0 * dx
-                wp.launch(
-                    kernel,
-                    dim=1,
-                    inputs=[
-                        v2,
-                        v3,
-                        wp.array(v4test, dtype=v4.dtype, requires_grad=True, device=device),
-                    ],
-                    outputs=[
-                        det2,
-                        det3,
-                        det4,
-                    ],
-                    device=device,
-                )
-                dminus = det4.numpy()[0]
-                assert_np_equal((dplus - dminus) / (2.0 * dx * dplus), v4grads[i, j] / dplus, tol=fdtol)
-
-
-def test_trace(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    def check_mat_trace(
-        v2: wp.array(dtype=mat22),
-        v3: wp.array(dtype=mat33),
-        v4: wp.array(dtype=mat44),
-        v5: wp.array(dtype=mat55),
-        tr2: wp.array(dtype=wptype),
-        tr3: wp.array(dtype=wptype),
-        tr4: wp.array(dtype=wptype),
-        tr5: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        tr2[0] = wptype(2) * wp.trace(v2[0])
-        tr3[0] = wptype(2) * wp.trace(v3[0])
-        tr4[0] = wptype(2) * wp.trace(v4[0])
-        tr5[0] = wptype(2) * wp.trace(v5[0])
-
-    kernel = getkernel(check_mat_trace, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
-    tr2 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tr3 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tr4 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tr5 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[
-                tr2,
-                tr3,
-                tr4,
-                tr5,
-            ],
-            device=device,
-        )
-
-    assert_np_equal(tr2.numpy()[0], 2 * np.trace(v2.numpy()[0]), tol=10 * tol)
-    assert_np_equal(tr3.numpy()[0], 2 * np.trace(v3.numpy()[0]), tol=10 * tol)
-    assert_np_equal(tr4.numpy()[0], 2 * np.trace(v4.numpy()[0]), tol=200 * tol)
-    assert_np_equal(tr4.numpy()[0], 2 * np.trace(v4.numpy()[0]), tol=200 * tol)
-
-    if dtype in np_float_types:
-        tape.backward(loss=tr2)
-        vgrads = tape.gradients[v2].numpy()[0]
-        assert_np_equal(vgrads, 2.0 * np.eye(2), tol=10 * tol)
-        tape.zero()
-
-        tape.backward(loss=tr3)
-        vgrads = tape.gradients[v3].numpy()[0]
-        assert_np_equal(vgrads, 2.0 * np.eye(3), tol=10 * tol)
-        tape.zero()
-
-        tape.backward(loss=tr4)
-        vgrads = tape.gradients[v4].numpy()[0]
-        assert_np_equal(vgrads, 2.0 * np.eye(4), tol=10 * tol)
-        tape.zero()
-
-        tape.backward(loss=tr5)
-        vgrads = tape.gradients[v5].numpy()[0]
-        assert_np_equal(vgrads, 2.0 * np.eye(5), tol=10 * tol)
-        tape.zero()
-
-
-def test_diag(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_diag(
-        s5: wp.array(dtype=vec5),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        m55result = wptype(2) * wp.diag(s5[0])
-
-        idx = 0
-        for i in range(5):
-            for j in range(5):
-                outcomponents[idx] = m55result[i, j]
-                idx = idx + 1
-
-    kernel = getkernel(check_mat_diag, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
-    outcomponents = wp.zeros(5 * 5, dtype=wptype, requires_grad=True, device=device)
-    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[s5], outputs=[outcomponents], device=device)
-
-    assert_np_equal(outcomponents.numpy(), 2 * np.diag(s5.numpy()[0]), tol=tol)
-
-    if dtype in np_float_types:
-        idx = 0
-        for i in range(5):
-            for j in range(5):
-                tape = wp.Tape()
-                with tape:
-                    wp.launch(kernel, dim=1, inputs=[s5], outputs=[outcomponents], device=device)
-                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-                tape.backward(loss=out)
-                expectedresult = np.zeros(5, dtype=dtype)
-                if i == j:
-                    expectedresult[i] = 2
-                assert_np_equal(tape.gradients[s5].numpy()[0], expectedresult, tol=10 * tol)
-                tape.zero()
-
-                idx = idx + 1
-
-
-def test_get_diag(test, device, dtype, register_kernels=False):
-    tol = {
-        np.float16: 1.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    mat55 = wp.types.vector(shape=(5, 5), dtype=wptype)
-
-    output_select_kernel = get_select_kernel(wptype)
-
-    def check_mat_diag(
-        m55: wp.array(dtype=mat55),
-        outcomponents: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        vec5result = wptype(2) * wp.get_diag(m55[0])
-
-        idx = 0
-        for i in range(5):
-            outcomponents[idx] = vec5result[i]
-            idx = idx + 1
-
-    kernel = getkernel(check_mat_diag, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    m55 = wp.array(randvals((1, 5, 5), dtype), dtype=mat55, requires_grad=True, device=device)
-    outcomponents = wp.zeros(5, dtype=wptype, requires_grad=True, device=device)
-    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(kernel, dim=1, inputs=[m55], outputs=[outcomponents], device=device)
+                )
+                dplus = det3.numpy()[0]
+                v3test[0, i, j] -= 2.0 * dx
+                wp.launch(
+                    kernel,
+                    dim=1,
+                    inputs=[
+                        v2,
+                        wp.array(v3test, dtype=v3.dtype, requires_grad=True, device=device),
+                        v4,
+                    ],
+                    outputs=[
+                        det2,
+                        det3,
+                        det4,
+                    ],
+                    device=device,
+                )
+                dminus = det3.numpy()[0]
+                assert_np_equal((dplus - dminus) / (2.0 * dx * dplus), v3grads[i, j] / dplus, tol=fdtol)
 
-    assert_np_equal(outcomponents.numpy(), 2 * np.diag(m55.numpy()[0]), tol=tol)
+        for i in range(4):
+            for j in range(4):
+                v4test = v4.numpy()
+                v4test[0, i, j] += dx
+                wp.launch(
+                    kernel,
+                    dim=1,
+                    inputs=[
+                        v2,
+                        v3,
+                        wp.array(v4test, dtype=v4.dtype, requires_grad=True, device=device),
+                    ],
+                    outputs=[
+                        det2,
+                        det3,
+                        det4,
+                    ],
+                    device=device,
+                )
+                dplus = det4.numpy()[0]
+                v4test[0, i, j] -= 2.0 * dx
+                wp.launch(
+                    kernel,
+                    dim=1,
+                    inputs=[
+                        v2,
+                        v3,
+                        wp.array(v4test, dtype=v4.dtype, requires_grad=True, device=device),
+                    ],
+                    outputs=[
+                        det2,
+                        det3,
+                        det4,
+                    ],
+                    device=device,
+                )
+                dminus = det4.numpy()[0]
+                assert_np_equal((dplus - dminus) / (2.0 * dx * dplus), v4grads[i, j] / dplus, tol=fdtol)
 
-    if dtype in np_float_types:
-        idx = 0
-        for i in range(5):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(kernel, dim=1, inputs=[m55], outputs=[outcomponents], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
-            tape.backward(loss=out)
-            expectedresult = np.zeros((5, 5), dtype=dtype)
-            expectedresult[i, i] = 2
-            assert_np_equal(tape.gradients[m55].numpy()[0], expectedresult, tol=10 * tol)
-            tape.zero()
 
-            idx = idx + 1
+# Unused. Why?
+# def test_get_diag(test, device, dtype, register_kernels=False):
+#     tol = {
+#         np.float16: 1.0e-3,
+#         np.float32: 1.0e-6,
+#         np.float64: 1.0e-8,
+#     }.get(dtype, 0)
+#
+#     wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+#     mat55 = wp.types.vector(shape=(5, 5), dtype=wptype)
+#
+#     output_select_kernel = get_select_kernel(wptype)
+#
+#     def check_mat_diag(
+#         m55: wp.array(dtype=mat55),
+#         outcomponents: wp.array(dtype=wptype),
+#     ):
+#         # multiply outputs by 2 so we've got something to backpropagate:
+#         vec5result = wptype(2) * wp.get_diag(m55[0])
+#
+#         idx = 0
+#         for i in range(5):
+#             outcomponents[idx] = vec5result[i]
+#             idx = idx + 1
+#
+#     kernel = getkernel(check_mat_diag, suffix=dtype.__name__)
+#
+#     if register_kernels:
+#         return
+#
+#     m55 = wp.array(randvals((1, 5, 5), dtype), dtype=mat55, requires_grad=True, device=device)
+#     outcomponents = wp.zeros(5, dtype=wptype, requires_grad=True, device=device)
+#     out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+#
+#     wp.launch(kernel, dim=1, inputs=[m55], outputs=[outcomponents], device=device)
+#
+#     assert_np_equal(outcomponents.numpy(), 2 * np.diag(m55.numpy()[0]), tol=tol)
+#
+#     if dtype in np_float_types:
+#         idx = 0
+#         for i in range(5):
+#             tape = wp.Tape()
+#             with tape:
+#                 wp.launch(kernel, dim=1, inputs=[m55], outputs=[outcomponents], device=device)
+#                 wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+#             tape.backward(loss=out)
+#             expectedresult = np.zeros((5, 5), dtype=dtype)
+#             expectedresult[i, i] = 2
+#             assert_np_equal(tape.gradients[m55].numpy()[0], expectedresult, tol=10 * tol)
+#             tape.zero()
+#
+#             idx = idx + 1
 
 
 def test_inverse(test, device, dtype, register_kernels=False):
@@ -3978,330 +1578,6 @@ def check_mat_transform_vector(
             tape.zero()
 
 
-def test_anon_type_instance(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    def check_scalar_init(
-        input: wp.array(dtype=wptype),
-        output: wp.array(dtype=wptype),
-    ):
-        m2result = wp.matrix(input[0], shape=(2, 2))
-        m3result = wp.matrix(input[1], shape=(3, 3))
-        m4result = wp.matrix(input[2], shape=(4, 4))
-        m5result = wp.matrix(input[3], shape=(5, 5))
-        m32result = wp.matrix(input[4], shape=(3, 2))
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                output[idx] = wptype(2) * m2result[i, j]
-                idx = idx + 1
-        for i in range(3):
-            for j in range(3):
-                output[idx] = wptype(2) * m3result[i, j]
-                idx = idx + 1
-        for i in range(4):
-            for j in range(4):
-                output[idx] = wptype(2) * m4result[i, j]
-                idx = idx + 1
-        for i in range(5):
-            for j in range(5):
-                output[idx] = wptype(2) * m5result[i, j]
-                idx = idx + 1
-        for i in range(3):
-            for j in range(2):
-                output[idx] = wptype(2) * m32result[i, j]
-                idx = idx + 1
-
-    def check_component_init(
-        input: wp.array(dtype=wptype),
-        output: wp.array(dtype=wptype),
-    ):
-        m2result = wp.matrix(input[0], input[1], input[2], input[3], shape=(2, 2))
-        m3result = wp.matrix(
-            input[4], input[5], input[6], input[7], input[8], input[9], input[10], input[11], input[12], shape=(3, 3)
-        )
-        m4result = wp.matrix(
-            input[13],
-            input[14],
-            input[15],
-            input[16],
-            input[17],
-            input[18],
-            input[19],
-            input[20],
-            input[21],
-            input[22],
-            input[23],
-            input[24],
-            input[25],
-            input[26],
-            input[27],
-            input[28],
-            shape=(4, 4),
-        )
-        m5result = wp.matrix(
-            input[29],
-            input[30],
-            input[31],
-            input[32],
-            input[33],
-            input[34],
-            input[35],
-            input[36],
-            input[37],
-            input[38],
-            input[39],
-            input[40],
-            input[41],
-            input[42],
-            input[43],
-            input[44],
-            input[45],
-            input[46],
-            input[47],
-            input[48],
-            input[49],
-            input[50],
-            input[51],
-            input[52],
-            input[53],
-            shape=(5, 5),
-        )
-        m32result = wp.matrix(input[54], input[55], input[56], input[57], input[58], input[59], shape=(3, 2))
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                output[idx] = wptype(2) * m2result[i, j]
-                idx = idx + 1
-        for i in range(3):
-            for j in range(3):
-                output[idx] = wptype(2) * m3result[i, j]
-                idx = idx + 1
-        for i in range(4):
-            for j in range(4):
-                output[idx] = wptype(2) * m4result[i, j]
-                idx = idx + 1
-        for i in range(5):
-            for j in range(5):
-                output[idx] = wptype(2) * m5result[i, j]
-                idx = idx + 1
-        for i in range(3):
-            for j in range(2):
-                output[idx] = wptype(2) * m32result[i, j]
-                idx = idx + 1
-
-    scalar_kernel = getkernel(check_scalar_init, suffix=dtype.__name__)
-    component_kernel = getkernel(check_component_init, suffix=dtype.__name__)
-    output_select_kernel = get_select_kernel(wptype)
-
-    if register_kernels:
-        return
-
-    input = wp.array(randvals(rng, [5], dtype), requires_grad=True, device=device)
-    output = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-
-    assert_np_equal(output.numpy()[:4], 2 * np.array([input.numpy()[0]] * 2 * 2), tol=1.0e-6)
-    assert_np_equal(output.numpy()[4:13], 2 * np.array([input.numpy()[1]] * 3 * 3), tol=1.0e-6)
-    assert_np_equal(output.numpy()[13:29], 2 * np.array([input.numpy()[2]] * 4 * 4), tol=1.0e-6)
-    assert_np_equal(output.numpy()[29:54], 2 * np.array([input.numpy()[3]] * 5 * 5), tol=1.0e-6)
-    assert_np_equal(output.numpy()[54:], 2 * np.array([input.numpy()[4]] * 3 * 2), tol=1.0e-6)
-
-    if dtype in np_float_types:
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for i in range(len(output)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
-
-            tape.backward(loss=out)
-            expected = np.zeros_like(input.numpy())
-            if i < 4:
-                expected[0] = 2
-            elif i < 13:
-                expected[1] = 2
-            elif i < 29:
-                expected[2] = 2
-            elif i < 54:
-                expected[3] = 2
-            else:
-                expected[4] = 2
-
-            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
-
-            tape.reset()
-            tape.zero()
-
-    input = wp.array(randvals(rng, [2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2], dtype), requires_grad=True, device=device)
-    output = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-
-    assert_np_equal(output.numpy(), 2 * input.numpy(), tol=1.0e-6)
-
-    if dtype in np_float_types:
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for i in range(len(output)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
-
-            tape.backward(loss=out)
-            expected = np.zeros_like(input.numpy())
-            expected[i] = 2
-
-            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
-
-            tape.reset()
-            tape.zero()
-
-
-def test_identity(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    def check_identity_mat(
-        output: wp.array(dtype=wptype),
-    ):
-        m2result = wp.identity(dtype=wptype, n=2)
-        m3result = wp.identity(dtype=wptype, n=3)
-        m4result = wp.identity(dtype=wptype, n=4)
-        m5result = wp.identity(dtype=wptype, n=5)
-
-        idx = 0
-        for i in range(2):
-            for j in range(2):
-                output[idx] = wptype(2) * m2result[i, j]
-                idx = idx + 1
-        for i in range(3):
-            for j in range(3):
-                output[idx] = wptype(2) * m3result[i, j]
-                idx = idx + 1
-        for i in range(4):
-            for j in range(4):
-                output[idx] = wptype(2) * m4result[i, j]
-                idx = idx + 1
-        for i in range(5):
-            for j in range(5):
-                output[idx] = wptype(2) * m5result[i, j]
-                idx = idx + 1
-
-    id_kernel = getkernel(check_identity_mat, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    output = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
-    wp.launch(id_kernel, dim=1, inputs=[], outputs=[output], device=device)
-    assert_np_equal(output.numpy()[:4], 2 * np.eye(2), tol=1.0e-6)
-    assert_np_equal(output.numpy()[4:13], 2 * np.eye(3), tol=1.0e-6)
-    assert_np_equal(output.numpy()[13:29], 2 * np.eye(4), tol=1.0e-6)
-    assert_np_equal(output.numpy()[29:], 2 * np.eye(5), tol=1.0e-6)
-
-
-def test_equivalent_types(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    # matrix types
-    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    # matrix types equivalent to the above
-    mat22_equiv = wp.types.matrix(shape=(2, 2), dtype=wptype)
-    mat33_equiv = wp.types.matrix(shape=(3, 3), dtype=wptype)
-    mat44_equiv = wp.types.matrix(shape=(4, 4), dtype=wptype)
-    mat55_equiv = wp.types.matrix(shape=(5, 5), dtype=wptype)
-
-    # declare kernel with original types
-    def check_equivalence(
-        m2: mat22,
-        m3: mat33,
-        m4: mat44,
-        m5: mat55,
-    ):
-        wp.expect_eq(m2, mat22(wptype(42)))
-        wp.expect_eq(m3, mat33(wptype(43)))
-        wp.expect_eq(m4, mat44(wptype(44)))
-        wp.expect_eq(m5, mat55(wptype(45)))
-
-        wp.expect_eq(m2, mat22_equiv(wptype(42)))
-        wp.expect_eq(m3, mat33_equiv(wptype(43)))
-        wp.expect_eq(m4, mat44_equiv(wptype(44)))
-        wp.expect_eq(m5, mat55_equiv(wptype(45)))
-
-    kernel = getkernel(check_equivalence, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    # call kernel with equivalent types
-    m2 = mat22_equiv(42)
-    m3 = mat33_equiv(43)
-    m4 = mat44_equiv(44)
-    m5 = mat55_equiv(45)
-
-    wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5], device=device)
-
-
-def test_conversions(test, device, dtype, register_kernels=False):
-    def check_matrices_equal(
-        m0: wp.mat22,
-        m1: wp.mat22,
-        m2: wp.mat22,
-        m3: wp.mat22,
-        m4: wp.mat22,
-        m5: wp.mat22,
-        m6: wp.mat22,
-    ):
-        wp.expect_eq(m1, m0)
-        wp.expect_eq(m2, m0)
-        wp.expect_eq(m3, m0)
-        wp.expect_eq(m4, m0)
-        wp.expect_eq(m5, m0)
-        wp.expect_eq(m6, m0)
-
-    kernel = getkernel(check_matrices_equal, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    m0 = wp.mat22(1, 2, 3, 4)
-
-    # test explicit conversions - constructing matrices from different containers
-    m1 = wp.mat22(((1, 2), (3, 4)))  # nested tuples
-    m2 = wp.mat22([[1, 2], [3, 4]])  # nested lists
-    m3 = wp.mat22(np.array([[1, 2], [3, 4]], dtype=dtype))  # 2d array
-    m4 = wp.mat22((1, 2, 3, 4))  # flat tuple
-    m5 = wp.mat22([1, 2, 3, 4])  # flat list
-    m6 = wp.mat22(np.array([1, 2, 3, 4], dtype=dtype))  # 1d array
-
-    wp.launch(kernel, dim=1, inputs=[m0, m1, m2, m3, m4, m5, m6], device=device)
-
-    # test implicit conversions - passing different containers as matrices to wp.launch()
-    m1 = ((1, 2), (3, 4))  # nested tuples
-    m2 = [[1, 2], [3, 4]]  # nested lists
-    m3 = np.array([[1, 2], [3, 4]], dtype=dtype)  # 2d array
-    m4 = (1, 2, 3, 4)  # flat tuple
-    m5 = [1, 2, 3, 4]  # flat list
-    m6 = np.array([1, 2, 3, 4], dtype=dtype)  # 1d array
-
-    wp.launch(kernel, dim=1, inputs=[m0, m1, m2, m3, m4, m5, m6], device=device)
-
-
 # Test matrix constructors using explicit type (float16)
 # note that these tests are specifically not using generics / closure
 # args to create kernels dynamically (like the rest of this file)
@@ -4328,7 +1604,6 @@ def test_constructors_explicit_precision():
 mat32d = wp.mat(shape=(3, 2), dtype=wp.float64)
 
 
-
 @wp.kernel
 def test_matrix_constructor_value_func():
     a = wp.mat22()
@@ -4404,239 +1679,149 @@ def test_constructors_constant_shape():
             m[i, j] = float(i * j)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestMat(parent):
-        pass
-
-    add_kernel_test(TestMat, test_constructors_explicit_precision, dim=1, devices=devices)
-    add_kernel_test(TestMat, test_constructors_default_precision, dim=1, devices=devices)
-    add_kernel_test(TestMat, test_constructors_constant_shape, dim=1, devices=devices)
-    add_kernel_test(TestMat, test_matrix_constructor_value_func, dim=1, devices=devices)
-
-    mat103 = wp.types.matrix(shape=(10, 3), dtype=float)
-    add_kernel_test(
-        TestMat,
-        test_matrix_mutation,
-        dim=1,
-        inputs=[
-            mat103(
-                1.0,
-                2.0,
-                3.0,
-                2.0,
-                4.0,
-                6.0,
-                3.0,
-                6.0,
-                9.0,
-                4.0,
-                8.0,
-                12.0,
-                5.0,
-                10.0,
-                15.0,
-                6.0,
-                12.0,
-                18.0,
-                7.0,
-                14.0,
-                21.0,
-                8.0,
-                16.0,
-                24.0,
-                9.0,
-                18.0,
-                27.0,
-                10.0,
-                20.0,
-                30.0,
-            )
-        ],
-        devices=devices,
-    )
-
-    for dtype in np_signed_int_types + np_float_types:
-        add_function_test_register_kernel(
-            TestMat, f"test_negation_{dtype.__name__}", test_negation, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_subtraction_{dtype.__name__}", test_subtraction, devices=devices, dtype=dtype
+devices = get_test_devices()
+
+
+class TestMat(unittest.TestCase):
+    pass
+
+
+add_kernel_test(TestMat, test_constructors_explicit_precision, dim=1, devices=devices)
+add_kernel_test(TestMat, test_constructors_default_precision, dim=1, devices=devices)
+add_kernel_test(TestMat, test_constructors_constant_shape, dim=1, devices=devices)
+add_kernel_test(TestMat, test_matrix_constructor_value_func, dim=1, devices=devices)
+
+mat103 = wp.types.matrix(shape=(10, 3), dtype=float)
+add_kernel_test(
+    TestMat,
+    test_matrix_mutation,
+    dim=1,
+    inputs=[
+        mat103(
+            1.0,
+            2.0,
+            3.0,
+            2.0,
+            4.0,
+            6.0,
+            3.0,
+            6.0,
+            9.0,
+            4.0,
+            8.0,
+            12.0,
+            5.0,
+            10.0,
+            15.0,
+            6.0,
+            12.0,
+            18.0,
+            7.0,
+            14.0,
+            21.0,
+            8.0,
+            16.0,
+            24.0,
+            9.0,
+            18.0,
+            27.0,
+            10.0,
+            20.0,
+            30.0,
         )
+    ],
+    devices=devices,
+)
 
-    add_function_test(
-        TestMat,
-        "test_anon_constructor_error_shape_keyword_missing",
-        test_anon_constructor_error_shape_keyword_missing,
-        devices=devices,
-    )
-    add_function_test(
-        TestMat,
-        "test_anon_constructor_error_dtype_keyword_missing",
-        test_anon_constructor_error_dtype_keyword_missing,
-        devices=devices,
+for dtype in np_signed_int_types + np_float_types:
+    add_function_test_register_kernel(
+        TestMat, f"test_negation_{dtype.__name__}", test_negation, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestMat,
-        "test_anon_constructor_error_shape_mismatch",
-        test_anon_constructor_error_shape_mismatch,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestMat, f"test_subtraction_{dtype.__name__}", test_subtraction, devices=devices, dtype=dtype
     )
+
+add_function_test(
+    TestMat,
+    "test_anon_constructor_error_shape_keyword_missing",
+    test_anon_constructor_error_shape_keyword_missing,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_anon_constructor_error_dtype_keyword_missing",
+    test_anon_constructor_error_dtype_keyword_missing,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_anon_constructor_error_shape_mismatch",
+    test_anon_constructor_error_shape_mismatch,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_anon_constructor_error_invalid_arg_count",
+    test_anon_constructor_error_invalid_arg_count,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_tpl_constructor_error_incompatible_sizes",
+    test_tpl_constructor_error_incompatible_sizes,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_tpl_constructor_error_invalid_scalar_type",
+    test_tpl_constructor_error_invalid_scalar_type,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_tpl_constructor_error_invalid_vector_count",
+    test_tpl_constructor_error_invalid_vector_count,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_tpl_constructor_error_invalid_vector_shape",
+    test_tpl_constructor_error_invalid_vector_shape,
+    devices=devices,
+)
+add_function_test(
+    TestMat,
+    "test_tpl_constructor_error_invalid_arg_count",
+    test_tpl_constructor_error_invalid_arg_count,
+    devices=devices,
+)
+add_function_test(TestMat, "test_tpl_ops_with_anon", test_tpl_ops_with_anon)
+
+for dtype in np_float_types:
     add_function_test(
-        TestMat,
-        "test_anon_constructor_error_invalid_arg_count",
-        test_anon_constructor_error_invalid_arg_count,
-        devices=devices,
+        TestMat, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
     )
-    add_function_test(
-        TestMat,
-        "test_tpl_constructor_error_incompatible_sizes",
-        test_tpl_constructor_error_incompatible_sizes,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestMat, f"test_quat_constructor_{dtype.__name__}", test_quat_constructor, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestMat,
-        "test_tpl_constructor_error_invalid_scalar_type",
-        test_tpl_constructor_error_invalid_scalar_type,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestMat, f"test_inverse_{dtype.__name__}", test_inverse, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestMat,
-        "test_tpl_constructor_error_invalid_vector_count",
-        test_tpl_constructor_error_invalid_vector_count,
-        devices=devices,
+    add_function_test_register_kernel(TestMat, f"test_svd_{dtype.__name__}", test_svd, devices=devices, dtype=dtype)
+    add_function_test_register_kernel(TestMat, f"test_qr_{dtype.__name__}", test_qr, devices=devices, dtype=dtype)
+    add_function_test_register_kernel(TestMat, f"test_eig_{dtype.__name__}", test_eig, devices=devices, dtype=dtype)
+    add_function_test_register_kernel(
+        TestMat, f"test_transform_point_{dtype.__name__}", test_transform_point, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestMat,
-        "test_tpl_constructor_error_invalid_vector_shape",
-        test_tpl_constructor_error_invalid_vector_shape,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestMat, f"test_transform_vector_{dtype.__name__}", test_transform_vector, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestMat,
-        "test_tpl_constructor_error_invalid_arg_count",
-        test_tpl_constructor_error_invalid_arg_count,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestMat, f"test_determinant_{dtype.__name__}", test_determinant, devices=devices, dtype=dtype
     )
-    add_function_test(TestMat, "test_tpl_ops_with_anon", test_tpl_ops_with_anon)
-
-    for dtype in np_scalar_types:
-        add_function_test(TestMat, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
-        add_function_test(TestMat, f"test_components_{dtype.__name__}", test_components, devices=None, dtype=dtype)
-        add_function_test_register_kernel(
-            TestMat, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_anon_type_instance_{dtype.__name__}", test_anon_type_instance, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_identity_{dtype.__name__}", test_identity, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_indexing_{dtype.__name__}", test_indexing, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_equality_{dtype.__name__}", test_equality, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat,
-            f"test_scalar_multiplication_{dtype.__name__}",
-            test_scalar_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestMat,
-            f"test_matvec_multiplication_{dtype.__name__}",
-            test_matvec_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestMat,
-            f"test_vecmat_multiplication_{dtype.__name__}",
-            test_vecmat_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestMat,
-            f"test_matmat_multiplication_{dtype.__name__}",
-            test_matmat_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_cw_multiplication_{dtype.__name__}", test_cw_multiplication, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_cw_division_{dtype.__name__}", test_cw_division, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_outer_product_{dtype.__name__}", test_outer_product, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_transpose_{dtype.__name__}", test_transpose, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_scalar_division_{dtype.__name__}", test_scalar_division, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_addition_{dtype.__name__}", test_addition, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_ddot_{dtype.__name__}", test_ddot, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_trace_{dtype.__name__}", test_trace, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_diag_{dtype.__name__}", test_diag, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_get_diag_{dtype.__name__}", test_diag, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_equivalent_types_{dtype.__name__}", test_equivalent_types, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_conversions_{dtype.__name__}", test_conversions, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_constants_{dtype.__name__}", test_constants, devices=devices, dtype=dtype
-        )
-
-    for dtype in np_float_types:
-        add_function_test(
-            TestMat, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_quat_constructor_{dtype.__name__}", test_quat_constructor, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_inverse_{dtype.__name__}", test_inverse, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(TestMat, f"test_svd_{dtype.__name__}", test_svd, devices=devices, dtype=dtype)
-        add_function_test_register_kernel(TestMat, f"test_qr_{dtype.__name__}", test_qr, devices=devices, dtype=dtype)
-        add_function_test_register_kernel(TestMat, f"test_eig_{dtype.__name__}", test_eig, devices=devices, dtype=dtype)
-        add_function_test_register_kernel(
-            TestMat, f"test_transform_point_{dtype.__name__}", test_transform_point, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_transform_vector_{dtype.__name__}", test_transform_vector, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_determinant_{dtype.__name__}", test_determinant, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestMat, f"test_skew_{dtype.__name__}", test_skew, devices=devices, dtype=dtype
-        )
-
-    return TestMat
+    add_function_test_register_kernel(TestMat, f"test_skew_{dtype.__name__}", test_skew, devices=devices, dtype=dtype)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_mat_lite.py b/warp/tests/test_mat_lite.py
index f894dde09..9e2d436cf 100644
--- a/warp/tests/test_mat_lite.py
+++ b/warp/tests/test_mat_lite.py
@@ -1,4 +1,4 @@
-# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
 # NVIDIA CORPORATION and its licensors retain all intellectual property
 # and proprietary rights in and to this software, related documentation
 # and any modifications thereto.  Any use, reproduction, disclosure or
@@ -7,9 +7,8 @@
 
 import unittest
 
-import numpy as np
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -99,20 +98,18 @@ def test_matrix_mutation(expected: wp.types.matrix(shape=(10, 3), dtype=float)):
     wp.expect_eq(m, expected)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestMat(parent):
-        pass
 
-    add_kernel_test(TestMat, test_matrix_constructor_value_func, dim=1, devices=devices)
-    add_kernel_test(TestMat, test_constructors_explicit_precision, dim=1, devices=devices)
-    add_kernel_test(TestMat, test_constructors_default_precision, dim=1, devices=devices)
+class TestMatLite(unittest.TestCase):
+    pass
 
-    return TestMat
+
+add_kernel_test(TestMatLite, test_matrix_constructor_value_func, dim=1, devices=devices)
+add_kernel_test(TestMatLite, test_constructors_explicit_precision, dim=1, devices=devices)
+add_kernel_test(TestMatLite, test_constructors_default_precision, dim=1, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_mat_scalar_ops.py b/warp/tests/test_mat_scalar_ops.py
new file mode 100644
index 000000000..2d5930c6c
--- /dev/null
+++ b/warp/tests/test_mat_scalar_ops.py
@@ -0,0 +1,2889 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import unittest
+
+import numpy as np
+
+import warp as wp
+from warp.tests.unittest_utils import *
+
+wp.init()
+
+np_signed_int_types = [
+    np.int8,
+    np.int16,
+    np.int32,
+    np.int64,
+    np.byte,
+]
+
+np_unsigned_int_types = [
+    np.uint8,
+    np.uint16,
+    np.uint32,
+    np.uint64,
+    np.ubyte,
+]
+
+np_int_types = np_signed_int_types + np_unsigned_int_types
+
+np_float_types = [np.float16, np.float32, np.float64]
+
+np_scalar_types = np_int_types + np_float_types
+
+
+def randvals(rng, shape, dtype):
+    if dtype in np_float_types:
+        return rng.standard_normal(size=shape).astype(dtype)
+    elif dtype in [np.int8, np.uint8, np.byte, np.ubyte]:
+        return rng.integers(1, high=3, size=shape, dtype=dtype)
+    return rng.integers(1, high=5, size=shape, dtype=dtype)
+
+
+kernel_cache = dict()
+
+
+def getkernel(func, suffix=""):
+    key = func.__name__ + "_" + suffix
+    if key not in kernel_cache:
+        kernel_cache[key] = wp.Kernel(func=func, key=key)
+    return kernel_cache[key]
+
+
+def get_select_kernel(dtype):
+    def output_select_kernel_fn(
+        input: wp.array(dtype=dtype),
+        index: int,
+        out: wp.array(dtype=dtype),
+    ):
+        out[0] = input[index]
+
+    return getkernel(output_select_kernel_fn, suffix=dtype.__name__)
+
+    wp.launch(kernel, dim=1, inputs=[])
+
+
+def test_arrays(test, device, dtype):
+    rng = np.random.default_rng(123)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
+
+    v2_np = randvals(rng, [10, 2, 2], dtype)
+    v3_np = randvals(rng, [10, 3, 3], dtype)
+    v4_np = randvals(rng, [10, 4, 4], dtype)
+    v5_np = randvals(rng, [10, 5, 5], dtype)
+    v32_np = randvals(rng, [10, 3, 2], dtype)
+
+    v2 = wp.array(v2_np, dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(v3_np, dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(v4_np, dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(v5_np, dtype=mat55, requires_grad=True, device=device)
+    v32 = wp.array(v32_np, dtype=mat32, requires_grad=True, device=device)
+
+    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
+    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
+    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
+    assert_np_equal(v5.numpy(), v5_np, tol=1.0e-6)
+    assert_np_equal(v32.numpy(), v32_np, tol=1.0e-6)
+
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+
+    v2 = wp.array(v2_np, dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(v3_np, dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(v4_np, dtype=mat44, requires_grad=True, device=device)
+
+    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
+    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
+    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
+
+
+def test_components(test, device, dtype):
+    # test accessing matrix components from Python - this is especially important
+    # for float16, which requires special handling internally
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat23 = wp.types.matrix(shape=(2, 3), dtype=wptype)
+
+    m = mat23(1, 2, 3, 4, 5, 6)
+
+    # test __getitem__ for row vectors
+    r0 = m[0]
+    r1 = m[1]
+    test.assertEqual(r0[0], 1)
+    test.assertEqual(r0[1], 2)
+    test.assertEqual(r0[2], 3)
+    test.assertEqual(r1[0], 4)
+    test.assertEqual(r1[1], 5)
+    test.assertEqual(r1[2], 6)
+
+    # test __getitem__ for individual components
+    test.assertEqual(m[0, 0], 1)
+    test.assertEqual(m[0, 1], 2)
+    test.assertEqual(m[0, 2], 3)
+    test.assertEqual(m[1, 0], 4)
+    test.assertEqual(m[1, 1], 5)
+    test.assertEqual(m[1, 2], 6)
+
+    # test __setitem__ for row vectors
+    m[0] = [7, 8, 9]
+    m[1] = [10, 11, 12]
+    test.assertEqual(m[0, 0], 7)
+    test.assertEqual(m[0, 1], 8)
+    test.assertEqual(m[0, 2], 9)
+    test.assertEqual(m[1, 0], 10)
+    test.assertEqual(m[1, 1], 11)
+    test.assertEqual(m[1, 2], 12)
+
+    # test __setitem__ for individual components
+    m[0, 0] = 13
+    m[0, 1] = 14
+    m[0, 2] = 15
+    m[1, 0] = 16
+    m[1, 1] = 17
+    m[1, 2] = 18
+    test.assertEqual(m[0, 0], 13)
+    test.assertEqual(m[0, 1], 14)
+    test.assertEqual(m[0, 2], 15)
+    test.assertEqual(m[1, 0], 16)
+    test.assertEqual(m[1, 1], 17)
+    test.assertEqual(m[1, 2], 18)
+
+
+def test_constants(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
+
+    cm22 = wp.constant(mat22(22))
+    cm33 = wp.constant(mat33(33))
+    cm44 = wp.constant(mat44(44))
+    cm55 = wp.constant(mat55(55))
+    cm32 = wp.constant(mat32(32))
+
+    def check_matrix_constants():
+        wp.expect_eq(cm22, mat22(wptype(22)))
+        wp.expect_eq(cm33, mat33(wptype(33)))
+        wp.expect_eq(cm44, mat44(wptype(44)))
+        wp.expect_eq(cm55, mat55(wptype(55)))
+        wp.expect_eq(cm32, mat32(wptype(32)))
+
+    kernel = getkernel(check_matrix_constants, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+
+def test_constructors(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_scalar_mat_constructor(
+        input: wp.array(dtype=wptype),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        m2result = wptype(2) * mat22(input[0])
+        m3result = wptype(2) * mat33(input[0])
+        m4result = wptype(2) * mat44(input[0])
+        m5result = wptype(2) * mat55(input[0])
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = m2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = m3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = m4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = m5result[i, j]
+                idx = idx + 1
+
+    def check_component_mat_constructor(
+        input: wp.array(dtype=wptype),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        m2result = wptype(2) * mat22(input[0], input[1], input[2], input[3])
+        m3result = wptype(2) * mat33(
+            input[4],
+            input[5],
+            input[6],
+            input[7],
+            input[8],
+            input[9],
+            input[10],
+            input[11],
+            input[12],
+        )
+        m4result = wptype(2) * mat44(
+            input[13],
+            input[14],
+            input[15],
+            input[16],
+            input[17],
+            input[18],
+            input[19],
+            input[20],
+            input[21],
+            input[22],
+            input[23],
+            input[24],
+            input[25],
+            input[26],
+            input[27],
+            input[28],
+        )
+        m5result = wptype(2) * mat55(
+            input[29],
+            input[30],
+            input[31],
+            input[32],
+            input[33],
+            input[34],
+            input[35],
+            input[36],
+            input[37],
+            input[38],
+            input[39],
+            input[40],
+            input[41],
+            input[42],
+            input[43],
+            input[44],
+            input[45],
+            input[46],
+            input[47],
+            input[48],
+            input[49],
+            input[50],
+            input[51],
+            input[52],
+            input[53],
+        )
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = m2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = m3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = m4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = m5result[i, j]
+                idx = idx + 1
+
+    def check_vector_mat_constructor(
+        input: wp.array(dtype=wptype),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        m2result = wptype(2) * mat22(vec2(input[0], input[2]), vec2(input[1], input[3]))
+        m3result = wptype(2) * mat33(
+            vec3(input[4], input[7], input[10]),
+            vec3(input[5], input[8], input[11]),
+            vec3(input[6], input[9], input[12]),
+        )
+        m4result = wptype(2) * mat44(
+            vec4(input[13], input[17], input[21], input[25]),
+            vec4(input[14], input[18], input[22], input[26]),
+            vec4(input[15], input[19], input[23], input[27]),
+            vec4(input[16], input[20], input[24], input[28]),
+        )
+        m5result = wptype(2) * mat55(
+            vec5(input[29], input[34], input[39], input[44], input[49]),
+            vec5(input[30], input[35], input[40], input[45], input[50]),
+            vec5(input[31], input[36], input[41], input[46], input[51]),
+            vec5(input[32], input[37], input[42], input[47], input[52]),
+            vec5(input[33], input[38], input[43], input[48], input[53]),
+        )
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = m2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = m3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = m4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = m5result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_scalar_mat_constructor, suffix=dtype.__name__)
+    compkernel = getkernel(check_component_mat_constructor, suffix=dtype.__name__)
+    veckernel = getkernel(check_vector_mat_constructor, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    input = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    val = input.numpy()[0]
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * val * np.ones(2 * 2), tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * val * np.ones(3 * 3), tol=tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * val * np.ones(4 * 4), tol=tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * val * np.ones(5 * 5), tol=tol)
+
+    if dtype in np_float_types:
+        for idx in range(len(outcomponents)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(kernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+            tape.backward(loss=out)
+            test.assertEqual(tape.gradients[input].numpy()[0], 2)
+            tape.zero()
+
+    input = wp.array(randvals(rng, [2 * 2 + 3 * 3 + 4 * 4 + 5 * 5], dtype), requires_grad=True, device=device)
+
+    wp.launch(compkernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
+    assert_np_equal(2 * input.numpy(), outcomponents.numpy(), tol=10 * tol)
+
+    if dtype in np_float_types:
+        for idx in range(len(outcomponents)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(compkernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+            tape.backward(loss=out)
+            expectedgrads = np.zeros(len(input))
+            expectedgrads[idx] = 2
+            assert_np_equal(tape.gradients[input].numpy(), expectedgrads)
+            tape.zero()
+
+    wp.launch(veckernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
+    assert_np_equal(2 * input.numpy(), outcomponents.numpy(), tol=10 * tol)
+
+    if dtype in np_float_types:
+        for idx in range(len(outcomponents)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(veckernel, dim=1, inputs=[input], outputs=[outcomponents], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+            tape.backward(loss=out)
+            expectedgrads = np.zeros(len(input))
+            expectedgrads[idx] = 2
+            assert_np_equal(tape.gradients[input].numpy(), expectedgrads)
+            tape.zero()
+
+
+def test_anon_type_instance(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def check_scalar_init(
+        input: wp.array(dtype=wptype),
+        output: wp.array(dtype=wptype),
+    ):
+        m2result = wp.matrix(input[0], shape=(2, 2))
+        m3result = wp.matrix(input[1], shape=(3, 3))
+        m4result = wp.matrix(input[2], shape=(4, 4))
+        m5result = wp.matrix(input[3], shape=(5, 5))
+        m32result = wp.matrix(input[4], shape=(3, 2))
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                output[idx] = wptype(2) * m2result[i, j]
+                idx = idx + 1
+        for i in range(3):
+            for j in range(3):
+                output[idx] = wptype(2) * m3result[i, j]
+                idx = idx + 1
+        for i in range(4):
+            for j in range(4):
+                output[idx] = wptype(2) * m4result[i, j]
+                idx = idx + 1
+        for i in range(5):
+            for j in range(5):
+                output[idx] = wptype(2) * m5result[i, j]
+                idx = idx + 1
+        for i in range(3):
+            for j in range(2):
+                output[idx] = wptype(2) * m32result[i, j]
+                idx = idx + 1
+
+    def check_component_init(
+        input: wp.array(dtype=wptype),
+        output: wp.array(dtype=wptype),
+    ):
+        m2result = wp.matrix(input[0], input[1], input[2], input[3], shape=(2, 2))
+        m3result = wp.matrix(
+            input[4], input[5], input[6], input[7], input[8], input[9], input[10], input[11], input[12], shape=(3, 3)
+        )
+        m4result = wp.matrix(
+            input[13],
+            input[14],
+            input[15],
+            input[16],
+            input[17],
+            input[18],
+            input[19],
+            input[20],
+            input[21],
+            input[22],
+            input[23],
+            input[24],
+            input[25],
+            input[26],
+            input[27],
+            input[28],
+            shape=(4, 4),
+        )
+        m5result = wp.matrix(
+            input[29],
+            input[30],
+            input[31],
+            input[32],
+            input[33],
+            input[34],
+            input[35],
+            input[36],
+            input[37],
+            input[38],
+            input[39],
+            input[40],
+            input[41],
+            input[42],
+            input[43],
+            input[44],
+            input[45],
+            input[46],
+            input[47],
+            input[48],
+            input[49],
+            input[50],
+            input[51],
+            input[52],
+            input[53],
+            shape=(5, 5),
+        )
+        m32result = wp.matrix(input[54], input[55], input[56], input[57], input[58], input[59], shape=(3, 2))
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                output[idx] = wptype(2) * m2result[i, j]
+                idx = idx + 1
+        for i in range(3):
+            for j in range(3):
+                output[idx] = wptype(2) * m3result[i, j]
+                idx = idx + 1
+        for i in range(4):
+            for j in range(4):
+                output[idx] = wptype(2) * m4result[i, j]
+                idx = idx + 1
+        for i in range(5):
+            for j in range(5):
+                output[idx] = wptype(2) * m5result[i, j]
+                idx = idx + 1
+        for i in range(3):
+            for j in range(2):
+                output[idx] = wptype(2) * m32result[i, j]
+                idx = idx + 1
+
+    scalar_kernel = getkernel(check_scalar_init, suffix=dtype.__name__)
+    component_kernel = getkernel(check_component_init, suffix=dtype.__name__)
+    output_select_kernel = get_select_kernel(wptype)
+
+    if register_kernels:
+        return
+
+    input = wp.array(randvals(rng, [5], dtype), requires_grad=True, device=device)
+    output = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+
+    assert_np_equal(output.numpy()[:4], 2 * np.array([input.numpy()[0]] * 2 * 2), tol=1.0e-6)
+    assert_np_equal(output.numpy()[4:13], 2 * np.array([input.numpy()[1]] * 3 * 3), tol=1.0e-6)
+    assert_np_equal(output.numpy()[13:29], 2 * np.array([input.numpy()[2]] * 4 * 4), tol=1.0e-6)
+    assert_np_equal(output.numpy()[29:54], 2 * np.array([input.numpy()[3]] * 5 * 5), tol=1.0e-6)
+    assert_np_equal(output.numpy()[54:], 2 * np.array([input.numpy()[4]] * 3 * 2), tol=1.0e-6)
+
+    if dtype in np_float_types:
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for i in range(len(output)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
+
+            tape.backward(loss=out)
+            expected = np.zeros_like(input.numpy())
+            if i < 4:
+                expected[0] = 2
+            elif i < 13:
+                expected[1] = 2
+            elif i < 29:
+                expected[2] = 2
+            elif i < 54:
+                expected[3] = 2
+            else:
+                expected[4] = 2
+
+            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
+
+            tape.reset()
+            tape.zero()
+
+    input = wp.array(randvals(rng, [2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2], dtype), requires_grad=True, device=device)
+    output = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+
+    assert_np_equal(output.numpy(), 2 * input.numpy(), tol=1.0e-6)
+
+    if dtype in np_float_types:
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for i in range(len(output)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
+
+            tape.backward(loss=out)
+            expected = np.zeros_like(input.numpy())
+            expected[i] = 2
+
+            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
+
+            tape.reset()
+            tape.zero()
+
+
+def test_identity(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def check_identity_mat(
+        output: wp.array(dtype=wptype),
+    ):
+        m2result = wp.identity(dtype=wptype, n=2)
+        m3result = wp.identity(dtype=wptype, n=3)
+        m4result = wp.identity(dtype=wptype, n=4)
+        m5result = wp.identity(dtype=wptype, n=5)
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                output[idx] = wptype(2) * m2result[i, j]
+                idx = idx + 1
+        for i in range(3):
+            for j in range(3):
+                output[idx] = wptype(2) * m3result[i, j]
+                idx = idx + 1
+        for i in range(4):
+            for j in range(4):
+                output[idx] = wptype(2) * m4result[i, j]
+                idx = idx + 1
+        for i in range(5):
+            for j in range(5):
+                output[idx] = wptype(2) * m5result[i, j]
+                idx = idx + 1
+
+    id_kernel = getkernel(check_identity_mat, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    output = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+    wp.launch(id_kernel, dim=1, inputs=[], outputs=[output], device=device)
+    assert_np_equal(output.numpy()[:4], 2 * np.eye(2), tol=1.0e-6)
+    assert_np_equal(output.numpy()[4:13], 2 * np.eye(3), tol=1.0e-6)
+    assert_np_equal(output.numpy()[13:29], 2 * np.eye(4), tol=1.0e-6)
+    assert_np_equal(output.numpy()[29:], 2 * np.eye(5), tol=1.0e-6)
+
+
+def test_indexing(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_indexing(
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * m2[0][i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * m3[0][i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * m4[0][i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * m5[0][i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_indexing, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * m2.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * m3.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * m4.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * m5.numpy().reshape(-1), tol=tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, input in [(2, m2), (3, m3), (4, m4), (5, m5)]:
+            for i in range(dim):
+                for j in range(dim):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5], outputs=[outcomponents], device=device)
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2
+                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult)
+                    tape.zero()
+                    idx = idx + 1
+
+
+def test_equality(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    def check_mat_equality():
+        wp.expect_eq(
+            mat22(wptype(1.0), wptype(2.0), wptype(3.0), wptype(4.0)),
+            mat22(wptype(1.0), wptype(2.0), wptype(3.0), wptype(4.0)),
+        )
+        wp.expect_neq(
+            mat22(wptype(1.0), wptype(2.0), wptype(3.0), -wptype(4.0)),
+            mat22(wptype(1.0), wptype(2.0), wptype(3.0), wptype(4.0)),
+        )
+
+        wp.expect_eq(
+            mat33(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+            ),
+            mat33(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+            ),
+        )
+        wp.expect_neq(
+            mat33(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+            ),
+            mat33(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                -wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+            ),
+        )
+
+        wp.expect_eq(
+            mat44(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+            ),
+            mat44(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+            ),
+        )
+
+        wp.expect_neq(
+            mat44(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+            ),
+            mat44(
+                -wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+            ),
+        )
+
+        wp.expect_eq(
+            mat55(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+                wptype(17.0),
+                wptype(18.0),
+                wptype(19.0),
+                wptype(20.0),
+                wptype(21.0),
+                wptype(22.0),
+                wptype(23.0),
+                wptype(24.0),
+                wptype(25.0),
+            ),
+            mat55(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+                wptype(17.0),
+                wptype(18.0),
+                wptype(19.0),
+                wptype(20.0),
+                wptype(21.0),
+                wptype(22.0),
+                wptype(23.0),
+                wptype(24.0),
+                wptype(25.0),
+            ),
+        )
+
+        wp.expect_neq(
+            mat55(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+                wptype(17.0),
+                wptype(18.0),
+                wptype(19.0),
+                wptype(20.0),
+                wptype(21.0),
+                wptype(22.0),
+                wptype(23.0),
+                wptype(24.0),
+                wptype(25.0),
+            ),
+            mat55(
+                wptype(1.0),
+                wptype(2.0),
+                wptype(3.0),
+                wptype(4.0),
+                wptype(5.0),
+                wptype(6.0),
+                wptype(7.0),
+                wptype(8.0),
+                wptype(9.0),
+                wptype(10.0),
+                wptype(11.0),
+                wptype(12.0),
+                wptype(13.0),
+                wptype(14.0),
+                wptype(15.0),
+                wptype(16.0),
+                -wptype(17.0),
+                wptype(18.0),
+                wptype(19.0),
+                wptype(20.0),
+                wptype(21.0),
+                wptype(22.0),
+                wptype(23.0),
+                wptype(24.0),
+                wptype(25.0),
+            ),
+        )
+
+    kernel = getkernel(check_mat_equality, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    wp.launch(kernel, dim=1, inputs=[], outputs=[], device=device)
+
+
+def test_scalar_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_scalar_mul(
+        s: wp.array(dtype=wptype),
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        outcomponents: wp.array(dtype=wptype),
+        outcomponents_rightmul: wp.array(dtype=wptype),
+    ):
+        m2result = s[0] * m2[0]
+        m3result = s[0] * m3[0]
+        m4result = s[0] * m4[0]
+        m5result = s[0] * m5[0]
+
+        m2resultright = m2[0] * s[0]
+        m3resultright = m3[0] * s[0]
+        m4resultright = m4[0] * s[0]
+        m5resultright = m5[0] * s[0]
+
+        m2result_2 = s[0] * m2[0]
+        m3result_2 = s[0] * m3[0]
+        m4result_2 = s[0] * m4[0]
+        m5result_2 = s[0] * m5[0]
+
+        m2resultright_2 = m2[0] * s[0]
+        m3resultright_2 = m3[0] * s[0]
+        m4resultright_2 = m4[0] * s[0]
+        m5resultright_2 = m5[0] * s[0]
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * m2result[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m2resultright[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * m3result[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m3resultright[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * m4result[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m4resultright[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * m5result[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m5resultright[i, j]
+                idx = idx + 1
+
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * m2result_2[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m2resultright_2[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * m3result_2[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m3resultright_2[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * m4result_2[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m4resultright_2[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * m5result_2[i, j]
+                outcomponents_rightmul[idx] = wptype(2) * m5resultright_2[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_scalar_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * (2 * 2 + 3 * 3 + 4 * 4 + 5 * 5), dtype=wptype, requires_grad=True, device=device)
+    outcomponents_rightmul = wp.zeros(
+        2 * (2 * 2 + 3 * 3 + 4 * 4 + 5 * 5), dtype=wptype, requires_grad=True, device=device
+    )
+
+    wp.launch(kernel, dim=1, inputs=[s, m2, m3, m4, m5], outputs=[outcomponents, outcomponents_rightmul], device=device)
+
+    sval = s.numpy()[0]
+    assert_np_equal(outcomponents.numpy()[:4], 2 * sval * m2.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
+
+    assert_np_equal(outcomponents_rightmul.numpy()[:4], 2 * sval * m2.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents_rightmul.numpy()[4:13], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents_rightmul.numpy()[13:29], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents_rightmul.numpy()[29:54], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
+
+    assert_np_equal(outcomponents.numpy()[54:58], 2 * sval * m2.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[58:67], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[67:83], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[83:108], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
+
+    assert_np_equal(outcomponents_rightmul.numpy()[54:58], 2 * sval * m2.numpy().reshape(-1), tol=tol)
+    assert_np_equal(outcomponents_rightmul.numpy()[58:67], 2 * sval * m3.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents_rightmul.numpy()[67:83], 2 * sval * m4.numpy().reshape(-1), tol=10 * tol)
+    assert_np_equal(outcomponents_rightmul.numpy()[83:108], 2 * sval * m5.numpy().reshape(-1), tol=10 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, input in [(2, m2), (3, m3), (4, m4), (5, m5)]:
+            for i in range(dim):
+                for j in range(dim):
+                    # test left mul gradient:
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[s, m2, m3, m4, m5],
+                            outputs=[outcomponents, outcomponents_rightmul],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2 * sval
+                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult, tol=10 * tol)
+                    assert_np_equal(tape.gradients[s].numpy()[0], 2 * input.numpy()[0, i, j], tol=10 * tol)
+                    tape.zero()
+
+                    # test right mul gradient:
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[s, m2, m3, m4, m5],
+                            outputs=[outcomponents, outcomponents_rightmul],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel,
+                            dim=1,
+                            inputs=[outcomponents_rightmul, idx],
+                            outputs=[out],
+                            device=device,
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2 * sval
+                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult, tol=10 * tol)
+                    assert_np_equal(tape.gradients[s].numpy()[0], 2 * input.numpy()[0, i, j], tol=10 * tol)
+                    tape.zero()
+
+                    idx = idx + 1
+
+
+def test_matvec_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 2.0e-2,
+        np.float32: 5.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_vec_mul(
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v32: wp.array(dtype=vec2),
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        m32: wp.array(dtype=mat32),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        v2result = m2[0] * v2[0]
+        v3result = m3[0] * v3[0]
+        v4result = m4[0] * v4[0]
+        v5result = m5[0] * v5[0]
+        v32result = m32[0] * v32[0]
+        v2result_2 = m2[0] @ v2[0]
+        v3result_2 = m3[0] @ v3[0]
+        v4result_2 = m4[0] @ v4[0]
+        v5result_2 = m5[0] @ v5[0]
+        v32result_2 = m32[0] @ v32[0]
+
+        idx = 0
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        for i in range(2):
+            outcomponents[idx] = wptype(2) * v2result[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v3result[i]
+            idx = idx + 1
+
+        for i in range(4):
+            outcomponents[idx] = wptype(2) * v4result[i]
+            idx = idx + 1
+
+        for i in range(5):
+            outcomponents[idx] = wptype(2) * v5result[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v32result[i]
+            idx = idx + 1
+
+        for i in range(2):
+            outcomponents[idx] = wptype(2) * v2result_2[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v3result_2[i]
+            idx = idx + 1
+
+        for i in range(4):
+            outcomponents[idx] = wptype(2) * v4result_2[i]
+            idx = idx + 1
+
+        for i in range(5):
+            outcomponents[idx] = wptype(2) * v5result_2[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v32result_2[i]
+            idx = idx + 1
+
+    kernel = getkernel(check_mat_vec_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
+    v32 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    m32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * (2 + 3 + 4 + 5 + 3), dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:2], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[2:5], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[5:9], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[9:14], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[14:17], 2 * np.matmul(m32.numpy()[0], v32.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[17:19], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[19:22], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[22:26], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[26:31], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[31:34], 2 * np.matmul(m32.numpy()[0], v32.numpy()[0]), tol=5 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, invec, inmat in [(2, v2, m2), (3, v3, m3), (4, v4, m4), (5, v5, m5), (3, v32, m32)]:
+            for i in range(dim):
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(
+                        kernel,
+                        dim=1,
+                        inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32],
+                        outputs=[outcomponents],
+                        device=device,
+                    )
+                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+                tape.backward(loss=out)
+
+                assert_np_equal(tape.gradients[invec].numpy()[0], 2 * inmat.numpy()[0, i, :], tol=2 * tol)
+                expectedresult = np.zeros(inmat.dtype._shape_, dtype=dtype)
+                expectedresult[i, :] = 2 * invec.numpy()[0]
+                assert_np_equal(tape.gradients[inmat].numpy()[0], expectedresult, tol=2 * tol)
+
+                tape.zero()
+
+                idx = idx + 1
+
+
+def test_vecmat_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 2.0e-2,
+        np.float32: 5.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat23 = wp.types.matrix(shape=(2, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_vec_mat_mul(
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v32: wp.array(dtype=vec2),
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        m23: wp.array(dtype=mat23),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        v2result = v2[0] * m2[0]
+        v3result = v3[0] * m3[0]
+        v4result = v4[0] * m4[0]
+        v5result = v5[0] * m5[0]
+        v32result = v32[0] * m23[0]
+        v2result_2 = v2[0] @ m2[0]
+        v3result_2 = v3[0] @ m3[0]
+        v4result_2 = v4[0] @ m4[0]
+        v5result_2 = v5[0] @ m5[0]
+        v32result_2 = v32[0] @ m23[0]
+
+        idx = 0
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        for i in range(2):
+            outcomponents[idx] = wptype(2) * v2result[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v3result[i]
+            idx = idx + 1
+
+        for i in range(4):
+            outcomponents[idx] = wptype(2) * v4result[i]
+            idx = idx + 1
+
+        for i in range(5):
+            outcomponents[idx] = wptype(2) * v5result[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v32result[i]
+            idx = idx + 1
+
+        for i in range(2):
+            outcomponents[idx] = wptype(2) * v2result_2[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v3result_2[i]
+            idx = idx + 1
+
+        for i in range(4):
+            outcomponents[idx] = wptype(2) * v4result_2[i]
+            idx = idx + 1
+
+        for i in range(5):
+            outcomponents[idx] = wptype(2) * v5result_2[i]
+            idx = idx + 1
+
+        for i in range(3):
+            outcomponents[idx] = wptype(2) * v32result_2[i]
+            idx = idx + 1
+
+    kernel = getkernel(check_vec_mat_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
+    v32 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    m23 = wp.array(randvals(rng, [1, 2, 3], dtype), dtype=mat23, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * (2 + 3 + 4 + 5 + 3), dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m23], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:2], 2 * np.matmul(v2.numpy()[0], m2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[2:5], 2 * np.matmul(v3.numpy()[0], m3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[5:9], 2 * np.matmul(v4.numpy()[0], m4.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[9:14], 2 * np.matmul(v5.numpy()[0], m5.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[14:17], 2 * np.matmul(v32.numpy()[0], m23.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[17:19], 2 * np.matmul(v2.numpy()[0], m2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[19:22], 2 * np.matmul(v3.numpy()[0], m3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[22:26], 2 * np.matmul(v4.numpy()[0], m4.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[26:31], 2 * np.matmul(v5.numpy()[0], m5.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[31:34], 2 * np.matmul(v32.numpy()[0], m23.numpy()[0]), tol=5 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, inmat, invec in [(2, m2, v2), (3, m3, v3), (4, m4, v4), (5, m5, v5), (3, m23, v32)]:
+            for i in range(dim):
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(
+                        kernel,
+                        dim=1,
+                        inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m23],
+                        outputs=[outcomponents],
+                        device=device,
+                    )
+                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+                tape.backward(loss=out)
+
+                assert_np_equal(tape.gradients[invec].numpy()[0], 2 * inmat.numpy()[0, :, i], tol=2 * tol)
+                expectedresult = np.zeros(inmat.dtype._shape_, dtype=dtype)
+                expectedresult[:, i] = 2 * invec.numpy()[0]
+                assert_np_equal(tape.gradients[inmat].numpy()[0], expectedresult, tol=2 * tol)
+
+                tape.zero()
+
+                idx = idx + 1
+
+
+def test_matmat_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 2.0e-2,
+        np.float32: 5.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_mat_mul(
+        a2: wp.array(dtype=mat22),
+        a3: wp.array(dtype=mat33),
+        a4: wp.array(dtype=mat44),
+        a5: wp.array(dtype=mat55),
+        a32: wp.array(dtype=mat32),
+        b2: wp.array(dtype=mat22),
+        b3: wp.array(dtype=mat33),
+        b4: wp.array(dtype=mat44),
+        b5: wp.array(dtype=mat55),
+        b32: wp.array(dtype=mat32),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        c2result = b2[0] * a2[0]
+        c3result = b3[0] * a3[0]
+        c4result = b4[0] * a4[0]
+        c5result = b5[0] * a5[0]
+        c32result = b32[0] * a2[0]
+        c32result2 = b3[0] * a32[0]
+        c2result_2 = b2[0] @ a2[0]
+        c3result_2 = b3[0] @ a3[0]
+        c4result_2 = b4[0] @ a4[0]
+        c5result_2 = b5[0] @ a5[0]
+        c32result_2 = b32[0] @ a2[0]
+        c32result2_2 = b3[0] @ a32[0]
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * c2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * c3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * c4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * c5result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * c32result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * c32result2[i, j]
+                idx = idx + 1
+
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * c2result_2[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * c3result_2[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * c4result_2[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * c5result_2[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * c32result_2[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * c32result2_2[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_mat_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    v32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    m32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
+    outcomponents = wp.zeros(
+        2 * (2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 3 * 2 + 3 * 2), dtype=wptype, requires_grad=True, device=device
+    )
+
+    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=2 * tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[54:60], 2 * np.matmul(m32.numpy()[0], v2.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[60:66], 2 * np.matmul(m3.numpy()[0], v32.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[66:70], 2 * np.matmul(m2.numpy()[0], v2.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[70:79], 2 * np.matmul(m3.numpy()[0], v3.numpy()[0]), tol=tol)
+    assert_np_equal(outcomponents.numpy()[79:95], 2 * np.matmul(m4.numpy()[0], v4.numpy()[0]), tol=2 * tol)
+    assert_np_equal(outcomponents.numpy()[95:120], 2 * np.matmul(m5.numpy()[0], v5.numpy()[0]), tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[120:126], 2 * np.matmul(m32.numpy()[0], v2.numpy()[0]), tol=5 * tol)
+    assert_np_equal(outcomponents.numpy()[126:132], 2 * np.matmul(m3.numpy()[0], v32.numpy()[0]), tol=5 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for v, m in [(v2, m2), (v3, m3), (v4, m4), (v5, m5), (v2, m32), (v32, m3)]:
+            rows, cols = m.dtype._shape_[0], v.dtype._shape_[1]
+            for i in range(rows):
+                for j in range(cols):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[v2, v3, v4, v5, v32, m2, m3, m4, m5, m32],
+                            outputs=[outcomponents],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+
+                    expected = np.zeros(v.dtype._shape_, dtype=dtype)
+                    expected[:, j] = 2 * m.numpy()[0, i, :]
+                    assert_np_equal(tape.gradients[v].numpy()[0], expected, tol=10 * tol)
+
+                    expected = np.zeros(m.dtype._shape_, dtype=dtype)
+                    expected[i, :] = 2 * v.numpy()[0, :, j]
+                    assert_np_equal(tape.gradients[m].numpy()[0], expected, tol=10 * tol)
+
+                    tape.zero()
+                    idx = idx + 1
+
+
+def test_cw_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_cw_mul(
+        s2: wp.array(dtype=mat22),
+        s3: wp.array(dtype=mat33),
+        s4: wp.array(dtype=mat44),
+        s5: wp.array(dtype=mat55),
+        v2: wp.array(dtype=mat22),
+        v3: wp.array(dtype=mat33),
+        v4: wp.array(dtype=mat44),
+        v5: wp.array(dtype=mat55),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        v2result = wptype(2) * wp.cw_mul(v2[0], s2[0])
+        v3result = wptype(2) * wp.cw_mul(v3[0], s3[0])
+        v4result = wptype(2) * wp.cw_mul(v4[0], s4[0])
+        v5result = wptype(2) * wp.cw_mul(v5[0], s5[0])
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = v2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = v3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = v4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = v5result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_cw_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(
+        kernel,
+        dim=1,
+        inputs=[
+            s2,
+            s3,
+            s4,
+            s5,
+            v2,
+            v3,
+            v4,
+            v5,
+        ],
+        outputs=[outcomponents],
+        device=device,
+    )
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() * s2.numpy()).reshape(-1), tol=50 * tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() * s3.numpy()).reshape(-1), tol=50 * tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() * s4.numpy()).reshape(-1), tol=50 * tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() * s5.numpy()).reshape(-1), tol=50 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, in1, in2 in [(2, s2, v2), (3, s3, v3), (4, s4, v4), (5, s5, v5)]:
+            for i in range(dim):
+                for j in range(dim):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[
+                                s2,
+                                s3,
+                                s4,
+                                s5,
+                                v2,
+                                v3,
+                                v4,
+                                v5,
+                            ],
+                            outputs=[outcomponents],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2 * in1.numpy()[0][i, j]
+                    assert_np_equal(tape.gradients[in2].numpy()[0], expectedresult, tol=5 * tol)
+                    expectedresult[i, j] = 2 * in2.numpy()[0][i, j]
+                    assert_np_equal(tape.gradients[in1].numpy()[0], expectedresult, tol=5 * tol)
+                    tape.zero()
+
+                    idx = idx + 1
+
+
+def test_cw_division(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_cw_div(
+        s2: wp.array(dtype=mat22),
+        s3: wp.array(dtype=mat33),
+        s4: wp.array(dtype=mat44),
+        s5: wp.array(dtype=mat55),
+        v2: wp.array(dtype=mat22),
+        v3: wp.array(dtype=mat33),
+        v4: wp.array(dtype=mat44),
+        v5: wp.array(dtype=mat55),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        v2result = wptype(2) * wp.cw_div(v2[0], s2[0])
+        v3result = wptype(2) * wp.cw_div(v3[0], s3[0])
+        v4result = wptype(2) * wp.cw_div(v4[0], s4[0])
+        v5result = wptype(2) * wp.cw_div(v5[0], s5[0])
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = v2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = v3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = v4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = v5result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_cw_div, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = randvals(rng, [1, 2, 2], dtype)
+    s3 = randvals(rng, [1, 3, 3], dtype)
+    s4 = randvals(rng, [1, 4, 4], dtype)
+    s5 = randvals(rng, [1, 5, 5], dtype)
+
+    # set denominators to 1 if their magnitudes are small
+    # to prevent divide by zero, or overflows if we're testing
+    # float16:
+    s2[np.abs(s2) < 1.0e-2] = 1
+    s3[np.abs(s3) < 1.0e-2] = 1
+    s4[np.abs(s4) < 1.0e-2] = 1
+    s5[np.abs(s5) < 1.0e-2] = 1
+
+    s2 = wp.array(s2, dtype=mat22, requires_grad=True, device=device)
+    s3 = wp.array(s3, dtype=mat33, requires_grad=True, device=device)
+    s4 = wp.array(s4, dtype=mat44, requires_grad=True, device=device)
+    s5 = wp.array(s5, dtype=mat55, requires_grad=True, device=device)
+
+    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(
+        kernel,
+        dim=1,
+        inputs=[
+            s2,
+            s3,
+            s4,
+            s5,
+            v2,
+            v3,
+            v4,
+            v5,
+        ],
+        outputs=[outcomponents],
+        device=device,
+    )
+
+    if dtype in np_float_types:
+        assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() / s2.numpy()).reshape(-1), tol=50 * tol)
+        assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() / s3.numpy()).reshape(-1), tol=50 * tol)
+        assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() / s4.numpy()).reshape(-1), tol=50 * tol)
+        assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() / s5.numpy()).reshape(-1), tol=50 * tol)
+    else:
+        assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() // s2.numpy()).reshape(-1), tol=50 * tol)
+        assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() // s3.numpy()).reshape(-1), tol=50 * tol)
+        assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() // s4.numpy()).reshape(-1), tol=50 * tol)
+        assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() // s5.numpy()).reshape(-1), tol=50 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, s, v in [(2, s2, v2), (3, s3, v3), (4, s4, v4), (5, s5, v5)]:
+            for i in range(dim):
+                for j in range(dim):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[
+                                s2,
+                                s3,
+                                s4,
+                                s5,
+                                v2,
+                                v3,
+                                v4,
+                                v5,
+                            ],
+                            outputs=[outcomponents],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+
+                    # y = v/s
+                    # dy/dv = 1.0/s
+                    # dy/ds = -v/s^2
+
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2.0 / (s.numpy()[0, i, j])
+                    assert_np_equal(tape.gradients[v].numpy()[0], expectedresult, tol=50 * tol)
+                    expectedresult[i, j] = -2.0 * v.numpy()[0, i, j] / (s.numpy()[0, i, j] ** 2)
+                    assert_np_equal(
+                        tape.gradients[s].numpy()[0], expectedresult, tol=abs(outcomponents.numpy()[idx]) * 50 * tol
+                    )
+                    tape.zero()
+
+                    idx = idx + 1
+
+
+def test_outer_product(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_outer_product(
+        s2: wp.array(dtype=vec2),
+        s3: wp.array(dtype=vec3),
+        s4: wp.array(dtype=vec4),
+        s5: wp.array(dtype=vec5),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        m22result = wptype(2) * wp.outer(s2[0], v2[0])
+        m33result = wptype(2) * wp.outer(s3[0], v3[0])
+        m44result = wptype(2) * wp.outer(s4[0], v4[0])
+        m55result = wptype(2) * wp.outer(s5[0], v5[0])
+        m25result = wptype(2) * wp.outer(s2[0], v5[0])
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = m22result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = m33result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = m44result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = m55result[i, j]
+                idx = idx + 1
+
+        for i in range(2):
+            for j in range(5):
+                outcomponents[idx] = m25result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_outer_product, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, [1, 2], dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3], dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4], dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 2 * 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[s2, s3, s4, s5, v2, v3, v4, v5], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * s2.numpy()[0, :, None] * v2.numpy()[0, None, :], tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * s3.numpy()[0, :, None] * v3.numpy()[0, None, :], tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * s4.numpy()[0, :, None] * v4.numpy()[0, None, :], tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * s5.numpy()[0, :, None] * v5.numpy()[0, None, :], tol=10 * tol)
+    assert_np_equal(outcomponents.numpy()[54:], 2 * s2.numpy()[0, :, None] * v5.numpy()[0, None, :], tol=10 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for s, v in [(s2, v2), (s3, v3), (s4, v4), (s5, v5), (s2, v5)]:
+            rows = s.dtype._length_
+            cols = v.dtype._length_
+            for i in range(rows):
+                for j in range(cols):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[
+                                s2,
+                                s3,
+                                s4,
+                                s5,
+                                v2,
+                                v3,
+                                v4,
+                                v5,
+                            ],
+                            outputs=[outcomponents],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+
+                    # this component's gonna be s_i * v_j, so its s gradient is gonna be nozero
+                    # at the ith component and its v gradient will be nonzero at the jth component:
+
+                    expectedresult = np.zeros((rows), dtype=dtype)
+                    expectedresult[i] = 2 * v.numpy()[0, j]
+                    assert_np_equal(tape.gradients[s].numpy()[0], expectedresult, tol=10 * tol)
+
+                    expectedresult = np.zeros((cols), dtype=dtype)
+                    expectedresult[j] = 2 * s.numpy()[0, i]
+                    assert_np_equal(tape.gradients[v].numpy()[0], expectedresult, tol=10 * tol)
+                    tape.zero()
+
+                    idx = idx + 1
+
+
+def test_transpose(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat32 = wp.types.matrix(shape=(3, 2), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_transpose(
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        m32: wp.array(dtype=mat32),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        mat2 = wptype(2) * wp.transpose(m2[0])
+        mat3 = wptype(2) * wp.transpose(m3[0])
+        mat4 = wptype(2) * wp.transpose(m4[0])
+        mat5 = wptype(2) * wp.transpose(m5[0])
+        mat32 = wptype(2) * wp.transpose(m32[0])
+
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = mat2[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = mat3[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = mat4[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = mat5[i, j]
+                idx = idx + 1
+
+        for i in range(2):
+            for j in range(3):
+                outcomponents[idx] = mat32[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_transpose, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    m32 = wp.array(randvals(rng, [1, 3, 2], dtype), dtype=mat32, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5 + 2 * 3, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * m2.numpy()[0].T.reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * m3.numpy()[0].T.reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * m4.numpy()[0].T.reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * m5.numpy()[0].T.reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[54:], 2 * m32.numpy()[0].T.reshape(-1), tol=tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for input in [m2, m3, m4, m5]:
+            for i in range(input.dtype._shape_[0]):
+                for j in range(input.dtype._shape_[1]):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5, m32], outputs=[outcomponents], device=device)
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((input.dtype._shape_[1], input.dtype._shape_[0]), dtype=dtype)
+                    expectedresult[j, i] = 2
+                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult)
+                    tape.zero()
+                    idx = idx + 1
+
+
+def test_scalar_division(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_scalar_div(
+        s: wp.array(dtype=wptype),
+        m2: wp.array(dtype=mat22),
+        m3: wp.array(dtype=mat33),
+        m4: wp.array(dtype=mat44),
+        m5: wp.array(dtype=mat55),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        m2result = m2[0] / s[0]
+        m3result = m3[0] / s[0]
+        m4result = m4[0] / s[0]
+        m5result = m5[0] / s[0]
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * m2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * m3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * m4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * m5result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_scalar_div, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    m2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    m3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    m4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    m5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[s, m2, m3, m4, m5], outputs=[outcomponents], device=device)
+
+    sval = s.numpy()[0]
+    if dtype in np_float_types:
+        assert_np_equal(outcomponents.numpy()[:4], 2 * m2.numpy().reshape(-1) / sval, tol=tol)
+        assert_np_equal(outcomponents.numpy()[4:13], 2 * m3.numpy().reshape(-1) / sval, tol=10 * tol)
+        assert_np_equal(outcomponents.numpy()[13:29], 2 * m4.numpy().reshape(-1) / sval, tol=10 * tol)
+        assert_np_equal(outcomponents.numpy()[29:54], 2 * m5.numpy().reshape(-1) / sval, tol=10 * tol)
+    else:
+        assert_np_equal(outcomponents.numpy()[:4], 2 * (m2.numpy().reshape(-1) // sval), tol=tol)
+        assert_np_equal(outcomponents.numpy()[4:13], 2 * (m3.numpy().reshape(-1) // sval), tol=10 * tol)
+        assert_np_equal(outcomponents.numpy()[13:29], 2 * (m4.numpy().reshape(-1) // sval), tol=10 * tol)
+        assert_np_equal(outcomponents.numpy()[29:54], 2 * (m5.numpy().reshape(-1) // sval), tol=10 * tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, input in [(2, m2), (3, m3), (4, m4), (5, m5)]:
+            for i in range(dim):
+                for j in range(dim):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(kernel, dim=1, inputs=[s, m2, m3, m4, m5], outputs=[outcomponents], device=device)
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2.0 / sval
+                    assert_np_equal(tape.gradients[input].numpy()[0], expectedresult, tol=10 * tol)
+                    assert_np_equal(
+                        tape.gradients[s].numpy()[0], -2 * input.numpy()[0, i, j] / (sval * sval), tol=10 * tol
+                    )
+                    tape.zero()
+
+                    idx = idx + 1
+
+
+def test_addition(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 2.0e-2,
+        np.float32: 5.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_add(
+        s2: wp.array(dtype=mat22),
+        s3: wp.array(dtype=mat33),
+        s4: wp.array(dtype=mat44),
+        s5: wp.array(dtype=mat55),
+        v2: wp.array(dtype=mat22),
+        v3: wp.array(dtype=mat33),
+        v4: wp.array(dtype=mat44),
+        v5: wp.array(dtype=mat55),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        v2result = v2[0] + s2[0]
+        v3result = v3[0] + s3[0]
+        v4result = v4[0] + s4[0]
+        v5result = v5[0] + s5[0]
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        idx = 0
+        for i in range(2):
+            for j in range(2):
+                outcomponents[idx] = wptype(2) * v2result[i, j]
+                idx = idx + 1
+
+        for i in range(3):
+            for j in range(3):
+                outcomponents[idx] = wptype(2) * v3result[i, j]
+                idx = idx + 1
+
+        for i in range(4):
+            for j in range(4):
+                outcomponents[idx] = wptype(2) * v4result[i, j]
+                idx = idx + 1
+
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = wptype(2) * v5result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_add, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    outcomponents = wp.zeros(2 * 2 + 3 * 3 + 4 * 4 + 5 * 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(
+        kernel,
+        dim=1,
+        inputs=[
+            s2,
+            s3,
+            s4,
+            s5,
+            v2,
+            v3,
+            v4,
+            v5,
+        ],
+        outputs=[outcomponents],
+        device=device,
+    )
+
+    assert_np_equal(outcomponents.numpy()[:4], 2 * (v2.numpy() + s2.numpy()).reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[4:13], 2 * (v3.numpy() + s3.numpy()).reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[13:29], 2 * (v4.numpy() + s4.numpy()).reshape(-1), tol=tol)
+    assert_np_equal(outcomponents.numpy()[29:54], 2 * (v5.numpy() + s5.numpy()).reshape(-1), tol=tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for dim, in1, in2 in [(2, s2, v2), (3, s3, v3), (4, s4, v4), (5, s5, v5)]:
+            for i in range(dim):
+                for j in range(dim):
+                    tape = wp.Tape()
+                    with tape:
+                        wp.launch(
+                            kernel,
+                            dim=1,
+                            inputs=[
+                                s2,
+                                s3,
+                                s4,
+                                s5,
+                                v2,
+                                v3,
+                                v4,
+                                v5,
+                            ],
+                            outputs=[outcomponents],
+                            device=device,
+                        )
+                        wp.launch(
+                            output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device
+                        )
+                    tape.backward(loss=out)
+                    expectedresult = np.zeros((dim, dim), dtype=dtype)
+                    expectedresult[i, j] = 2
+                    assert_np_equal(tape.gradients[in2].numpy()[0], expectedresult, tol=10 * tol)
+                    expectedresult[i, j] = 2
+                    assert_np_equal(tape.gradients[in1].numpy()[0], expectedresult, tol=10 * tol)
+                    tape.zero()
+
+                    idx = idx + 1
+
+
+def test_ddot(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    def check_mat_dot(
+        s2: wp.array(dtype=mat22),
+        s3: wp.array(dtype=mat33),
+        s4: wp.array(dtype=mat44),
+        s5: wp.array(dtype=mat55),
+        v2: wp.array(dtype=mat22),
+        v3: wp.array(dtype=mat33),
+        v4: wp.array(dtype=mat44),
+        v5: wp.array(dtype=mat55),
+        dot2: wp.array(dtype=wptype),
+        dot3: wp.array(dtype=wptype),
+        dot4: wp.array(dtype=wptype),
+        dot5: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        dot2[0] = wptype(2) * wp.ddot(v2[0], s2[0])
+        dot3[0] = wptype(2) * wp.ddot(v3[0], s3[0])
+        dot4[0] = wptype(2) * wp.ddot(v4[0], s4[0])
+        dot5[0] = wptype(2) * wp.ddot(v5[0], s5[0])
+
+    kernel = getkernel(check_mat_dot, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    dot2 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    dot3 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    dot4 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    dot5 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s2,
+                s3,
+                s4,
+                s5,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[dot2, dot3, dot4, dot5],
+            device=device,
+        )
+
+    assert_np_equal(dot2.numpy()[0], 2 * (v2.numpy() * s2.numpy()).sum(), tol=10 * tol)
+    assert_np_equal(dot3.numpy()[0], 2 * (v3.numpy() * s3.numpy()).sum(), tol=10 * tol)
+    assert_np_equal(dot4.numpy()[0], 2 * (v4.numpy() * s4.numpy()).sum(), tol=50 * tol)
+    assert_np_equal(dot5.numpy()[0], 2 * (v5.numpy() * s5.numpy()).sum(), tol=200 * tol)
+
+    if dtype in np_float_types:
+        tape.backward(loss=dot2)
+        sgrads = tape.gradients[s2].numpy()[0]
+        expected_grads = 2.0 * v2.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v2].numpy()[0]
+        expected_grads = 2.0 * s2.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
+
+        tape.zero()
+
+        tape.backward(loss=dot3)
+        sgrads = tape.gradients[s3].numpy()[0]
+        expected_grads = 2.0 * v3.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v3].numpy()[0]
+        expected_grads = 2.0 * s3.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
+
+        tape.zero()
+
+        tape.backward(loss=dot4)
+        sgrads = tape.gradients[s4].numpy()[0]
+        expected_grads = 2.0 * v4.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v4].numpy()[0]
+        expected_grads = 2.0 * s4.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
+
+        tape.zero()
+
+        tape.backward(loss=dot5)
+        sgrads = tape.gradients[s5].numpy()[0]
+        expected_grads = 2.0 * v5.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v5].numpy()[0]
+        expected_grads = 2.0 * s5.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
+
+        tape.zero()
+
+
+def test_trace(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    def check_mat_trace(
+        v2: wp.array(dtype=mat22),
+        v3: wp.array(dtype=mat33),
+        v4: wp.array(dtype=mat44),
+        v5: wp.array(dtype=mat55),
+        tr2: wp.array(dtype=wptype),
+        tr3: wp.array(dtype=wptype),
+        tr4: wp.array(dtype=wptype),
+        tr5: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        tr2[0] = wptype(2) * wp.trace(v2[0])
+        tr3[0] = wptype(2) * wp.trace(v3[0])
+        tr4[0] = wptype(2) * wp.trace(v4[0])
+        tr5[0] = wptype(2) * wp.trace(v5[0])
+
+    kernel = getkernel(check_mat_trace, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v2 = wp.array(randvals(rng, [1, 2, 2], dtype), dtype=mat22, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, [1, 3, 3], dtype), dtype=mat33, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, [1, 4, 4], dtype), dtype=mat44, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, [1, 5, 5], dtype), dtype=mat55, requires_grad=True, device=device)
+    tr2 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tr3 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tr4 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tr5 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[
+                tr2,
+                tr3,
+                tr4,
+                tr5,
+            ],
+            device=device,
+        )
+
+    assert_np_equal(tr2.numpy()[0], 2 * np.trace(v2.numpy()[0]), tol=10 * tol)
+    assert_np_equal(tr3.numpy()[0], 2 * np.trace(v3.numpy()[0]), tol=10 * tol)
+    assert_np_equal(tr4.numpy()[0], 2 * np.trace(v4.numpy()[0]), tol=200 * tol)
+    assert_np_equal(tr4.numpy()[0], 2 * np.trace(v4.numpy()[0]), tol=200 * tol)
+
+    if dtype in np_float_types:
+        tape.backward(loss=tr2)
+        vgrads = tape.gradients[v2].numpy()[0]
+        assert_np_equal(vgrads, 2.0 * np.eye(2), tol=10 * tol)
+        tape.zero()
+
+        tape.backward(loss=tr3)
+        vgrads = tape.gradients[v3].numpy()[0]
+        assert_np_equal(vgrads, 2.0 * np.eye(3), tol=10 * tol)
+        tape.zero()
+
+        tape.backward(loss=tr4)
+        vgrads = tape.gradients[v4].numpy()[0]
+        assert_np_equal(vgrads, 2.0 * np.eye(4), tol=10 * tol)
+        tape.zero()
+
+        tape.backward(loss=tr5)
+        vgrads = tape.gradients[v5].numpy()[0]
+        assert_np_equal(vgrads, 2.0 * np.eye(5), tol=10 * tol)
+        tape.zero()
+
+
+def test_diag(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    output_select_kernel = get_select_kernel(wptype)
+
+    def check_mat_diag(
+        s5: wp.array(dtype=vec5),
+        outcomponents: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        m55result = wptype(2) * wp.diag(s5[0])
+
+        idx = 0
+        for i in range(5):
+            for j in range(5):
+                outcomponents[idx] = m55result[i, j]
+                idx = idx + 1
+
+    kernel = getkernel(check_mat_diag, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s5 = wp.array(randvals(rng, [1, 5], dtype), dtype=vec5, requires_grad=True, device=device)
+    outcomponents = wp.zeros(5 * 5, dtype=wptype, requires_grad=True, device=device)
+    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(kernel, dim=1, inputs=[s5], outputs=[outcomponents], device=device)
+
+    assert_np_equal(outcomponents.numpy(), 2 * np.diag(s5.numpy()[0]), tol=tol)
+
+    if dtype in np_float_types:
+        idx = 0
+        for i in range(5):
+            for j in range(5):
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(kernel, dim=1, inputs=[s5], outputs=[outcomponents], device=device)
+                    wp.launch(output_select_kernel, dim=1, inputs=[outcomponents, idx], outputs=[out], device=device)
+                tape.backward(loss=out)
+                expectedresult = np.zeros(5, dtype=dtype)
+                if i == j:
+                    expectedresult[i] = 2
+                assert_np_equal(tape.gradients[s5].numpy()[0], expectedresult, tol=10 * tol)
+                tape.zero()
+
+                idx = idx + 1
+
+
+def test_equivalent_types(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    # matrix types
+    mat22 = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33 = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44 = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55 = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    # matrix types equivalent to the above
+    mat22_equiv = wp.types.matrix(shape=(2, 2), dtype=wptype)
+    mat33_equiv = wp.types.matrix(shape=(3, 3), dtype=wptype)
+    mat44_equiv = wp.types.matrix(shape=(4, 4), dtype=wptype)
+    mat55_equiv = wp.types.matrix(shape=(5, 5), dtype=wptype)
+
+    # declare kernel with original types
+    def check_equivalence(
+        m2: mat22,
+        m3: mat33,
+        m4: mat44,
+        m5: mat55,
+    ):
+        wp.expect_eq(m2, mat22(wptype(42)))
+        wp.expect_eq(m3, mat33(wptype(43)))
+        wp.expect_eq(m4, mat44(wptype(44)))
+        wp.expect_eq(m5, mat55(wptype(45)))
+
+        wp.expect_eq(m2, mat22_equiv(wptype(42)))
+        wp.expect_eq(m3, mat33_equiv(wptype(43)))
+        wp.expect_eq(m4, mat44_equiv(wptype(44)))
+        wp.expect_eq(m5, mat55_equiv(wptype(45)))
+
+    kernel = getkernel(check_equivalence, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    # call kernel with equivalent types
+    m2 = mat22_equiv(42)
+    m3 = mat33_equiv(43)
+    m4 = mat44_equiv(44)
+    m5 = mat55_equiv(45)
+
+    wp.launch(kernel, dim=1, inputs=[m2, m3, m4, m5], device=device)
+
+
+def test_conversions(test, device, dtype, register_kernels=False):
+    def check_matrices_equal(
+        m0: wp.mat22,
+        m1: wp.mat22,
+        m2: wp.mat22,
+        m3: wp.mat22,
+        m4: wp.mat22,
+        m5: wp.mat22,
+        m6: wp.mat22,
+    ):
+        wp.expect_eq(m1, m0)
+        wp.expect_eq(m2, m0)
+        wp.expect_eq(m3, m0)
+        wp.expect_eq(m4, m0)
+        wp.expect_eq(m5, m0)
+        wp.expect_eq(m6, m0)
+
+    kernel = getkernel(check_matrices_equal, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    m0 = wp.mat22(1, 2, 3, 4)
+
+    # test explicit conversions - constructing matrices from different containers
+    m1 = wp.mat22(((1, 2), (3, 4)))  # nested tuples
+    m2 = wp.mat22([[1, 2], [3, 4]])  # nested lists
+    m3 = wp.mat22(np.array([[1, 2], [3, 4]], dtype=dtype))  # 2d array
+    m4 = wp.mat22((1, 2, 3, 4))  # flat tuple
+    m5 = wp.mat22([1, 2, 3, 4])  # flat list
+    m6 = wp.mat22(np.array([1, 2, 3, 4], dtype=dtype))  # 1d array
+
+    wp.launch(kernel, dim=1, inputs=[m0, m1, m2, m3, m4, m5, m6], device=device)
+
+    # test implicit conversions - passing different containers as matrices to wp.launch()
+    m1 = ((1, 2), (3, 4))  # nested tuples
+    m2 = [[1, 2], [3, 4]]  # nested lists
+    m3 = np.array([[1, 2], [3, 4]], dtype=dtype)  # 2d array
+    m4 = (1, 2, 3, 4)  # flat tuple
+    m5 = [1, 2, 3, 4]  # flat list
+    m6 = np.array([1, 2, 3, 4], dtype=dtype)  # 1d array
+
+    wp.launch(kernel, dim=1, inputs=[m0, m1, m2, m3, m4, m5, m6], device=device)
+
+
+devices = get_test_devices()
+
+
+class TestMatScalarOps(unittest.TestCase):
+    pass
+
+
+for dtype in np_scalar_types:
+    add_function_test(TestMatScalarOps, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
+    add_function_test(TestMatScalarOps, f"test_components_{dtype.__name__}", test_components, devices=None, dtype=dtype)
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps,
+        f"test_anon_type_instance_{dtype.__name__}",
+        test_anon_type_instance,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_identity_{dtype.__name__}", test_identity, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_indexing_{dtype.__name__}", test_indexing, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_equality_{dtype.__name__}", test_equality, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps,
+        f"test_scalar_multiplication_{dtype.__name__}",
+        test_scalar_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps,
+        f"test_matvec_multiplication_{dtype.__name__}",
+        test_matvec_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps,
+        f"test_vecmat_multiplication_{dtype.__name__}",
+        test_vecmat_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps,
+        f"test_matmat_multiplication_{dtype.__name__}",
+        test_matmat_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps,
+        f"test_cw_multiplication_{dtype.__name__}",
+        test_cw_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_cw_division_{dtype.__name__}", test_cw_division, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_outer_product_{dtype.__name__}", test_outer_product, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_transpose_{dtype.__name__}", test_transpose, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_scalar_division_{dtype.__name__}", test_scalar_division, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_addition_{dtype.__name__}", test_addition, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_ddot_{dtype.__name__}", test_ddot, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_trace_{dtype.__name__}", test_trace, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_diag_{dtype.__name__}", test_diag, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_get_diag_{dtype.__name__}", test_diag, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_equivalent_types_{dtype.__name__}", test_equivalent_types, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_conversions_{dtype.__name__}", test_conversions, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestMatScalarOps, f"test_constants_{dtype.__name__}", test_constants, devices=devices, dtype=dtype
+    )
+
+
+if __name__ == "__main__":
+    wp.build.clear_kernel_cache()
+    unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_math.py b/warp/tests/test_math.py
index 018dde5f5..84f896c32 100644
--- a/warp/tests/test_math.py
+++ b/warp/tests/test_math.py
@@ -5,13 +5,13 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-from typing import NamedTuple
 import unittest
+from typing import NamedTuple
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -176,19 +176,18 @@ def test_mat_type(test, device):
         raise ValueError("mat to string error")
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
+
+class TestMath(unittest.TestCase):
+    pass
 
-    class TestMath(parent):
-        pass
 
-    add_function_test(TestMath, "test_scalar_math", test_scalar_math, devices=devices)
-    add_function_test(TestMath, "test_vec_type", test_vec_type, devices=devices)
-    add_function_test(TestMath, "test_mat_type", test_mat_type, devices=devices)
-    return TestMath
+add_function_test(TestMath, "test_scalar_math", test_scalar_math, devices=devices)
+add_function_test(TestMath, "test_vec_type", test_vec_type, devices=devices)
+add_function_test(TestMath, "test_mat_type", test_mat_type, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_matmul.py b/warp/tests/test_matmul.py
index 225770b12..1a4bb4cdc 100644
--- a/warp/tests/test_matmul.py
+++ b/warp/tests/test_matmul.py
@@ -1,11 +1,21 @@
-import numpy as np
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
+import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
+from warp.context import runtime  # noqa: E402
+
 
 class gemm_test_bed_runner:
     def __init__(self, dtype, device):
@@ -21,63 +31,54 @@ def alloc(self, m, n, k, batch_count):
                 np.ceil(rng.uniform(low=low, high=high, size=(m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
-            D = wp.array2d(
-                np.zeros((m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True)
+            D = wp.array2d(np.zeros((m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
         else:
             A = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
-            )
-            D = wp.array3d(
-                np.zeros((batch_count, m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
+            D = wp.array3d(np.zeros((batch_count, m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
         return A, B, C, D
 
     def run_and_verify(self, m, n, k, batch_count, alpha, beta):
         A, B, C, D = self.alloc(m, n, k, batch_count)
         ones = wp.zeros_like(D)
         ones.fill_(1.0)
-        
+
         if batch_count == 1:
             tape = wp.Tape()
             with tape:
                 wp.matmul(A, B, C, D, alpha, beta, False, self.device)
-            tape.backward(grads={D : ones})
-            
+            tape.backward(grads={D: ones})
+
             D_np = alpha * (A.numpy() @ B.numpy()) + beta * C.numpy()
             assert np.array_equal(D_np, D.numpy())
 
@@ -89,8 +90,8 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
             tape = wp.Tape()
             with tape:
                 wp.batched_matmul(A, B, C, D, alpha, beta, False, self.device)
-            tape.backward(grads={D : ones})
-            
+            tape.backward(grads={D: ones})
+
             D_np = alpha * np.matmul(A.numpy(), B.numpy()) + beta * C.numpy()
             assert np.array_equal(D_np, D.numpy())
 
@@ -132,75 +133,45 @@ def alloc(self, m, n, k, batch_count):
                 np.ceil(rng.uniform(low=low, high=high, size=(m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
-            )
-            D = wp.array2d(
-                np.zeros((m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            AT = wp.array2d(
-                A.numpy().transpose([1, 0]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            BT = wp.array2d(
-                B.numpy().transpose([1, 0]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
+            D = wp.array2d(np.zeros((m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
+            AT = wp.array2d(A.numpy().transpose([1, 0]), dtype=self.dtype, device=self.device, requires_grad=True)
+            BT = wp.array2d(B.numpy().transpose([1, 0]), dtype=self.dtype, device=self.device, requires_grad=True)
         else:
             A = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
-            )
-            D = wp.array3d(
-                np.zeros((batch_count, m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            AT = wp.array3d(
-                A.numpy().transpose([0, 2, 1]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            BT = wp.array3d(
-                B.numpy().transpose([0, 2, 1]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
+            D = wp.array3d(np.zeros((batch_count, m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
+            AT = wp.array3d(A.numpy().transpose([0, 2, 1]), dtype=self.dtype, device=self.device, requires_grad=True)
+            BT = wp.array3d(B.numpy().transpose([0, 2, 1]), dtype=self.dtype, device=self.device, requires_grad=True)
         return A, B, C, D, AT, BT
 
     def run_and_verify(self, m, n, k, batch_count, alpha, beta):
@@ -219,17 +190,17 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
         ones3.fill_(1.0)
 
         if batch_count == 1:
-            ATT1 = AT1.transpose([1, 0])        
+            ATT1 = AT1.transpose([1, 0])
             BTT1 = BT1.transpose([1, 0])
-            ATT2 = AT2.transpose([1, 0])        
+            ATT2 = AT2.transpose([1, 0])
             BTT2 = BT2.transpose([1, 0])
             tape = wp.Tape()
             with tape:
                 wp.matmul(A, BTT1, C1, D1, alpha, beta, False, self.device)
                 wp.matmul(ATT1, B, C2, D2, alpha, beta, False, self.device)
                 wp.matmul(ATT2, BTT2, C3, D3, alpha, beta, False, self.device)
-            tape.backward(grads={D1 : ones1, D2 : ones2, D3 : ones3})
-            
+            tape.backward(grads={D1: ones1, D2: ones2, D3: ones3})
+
             D_np = alpha * (A.numpy() @ B.numpy()) + beta * C1.numpy()
             assert np.array_equal(D_np, D1.numpy())
             assert np.array_equal(D_np, D2.numpy())
@@ -240,7 +211,7 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
             adj_C_np = beta * ones1.numpy()
 
         else:
-            ATT1 = AT1.transpose([0, 2, 1])        
+            ATT1 = AT1.transpose([0, 2, 1])
             BTT1 = BT1.transpose([0, 2, 1])
             ATT2 = AT2.transpose([0, 2, 1])
             BTT2 = BT2.transpose([0, 2, 1])
@@ -249,8 +220,8 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
                 wp.batched_matmul(A, BTT1, C1, D1, alpha, beta, False, self.device)
                 wp.batched_matmul(ATT1, B, C2, D2, alpha, beta, False, self.device)
                 wp.batched_matmul(ATT2, BTT2, C3, D3, alpha, beta, False, self.device)
-            tape.backward(grads={D1 : ones1, D2 : ones2, D3 : ones3})
-            
+            tape.backward(grads={D1: ones1, D2: ones2, D3: ones3})
+
             D_np = alpha * np.matmul(A.numpy(), B.numpy()) + beta * C1.numpy()
             assert np.array_equal(D_np, D1.numpy())
             assert np.array_equal(D_np, D2.numpy())
@@ -288,11 +259,13 @@ def test_f16(test, device):
     gemm_test_bed_runner_transpose(wp.float16, device).run()
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_f32(test, device):
     gemm_test_bed_runner(wp.float32, device).run()
     gemm_test_bed_runner_transpose(wp.float32, device).run()
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_f64(test, device):
     gemm_test_bed_runner(wp.float64, device).run()
     gemm_test_bed_runner_transpose(wp.float64, device).run()
@@ -304,6 +277,7 @@ def matrix_sum_kernel(arr: wp.array2d(dtype=float), loss: wp.array(dtype=float))
     wp.atomic_add(loss, 0, arr[i, j])
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_tape(test, device):
     rng = np.random.default_rng(42)
     low = -4.5
@@ -343,6 +317,7 @@ def test_tape(test, device):
     assert_array_equal(A.grad, wp.zeros_like(A))
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_operator(test, device):
     rng = np.random.default_rng(42)
     low = -4.5
@@ -378,6 +353,7 @@ def test_operator(test, device):
     assert_array_equal(A.grad, wp.zeros_like(A))
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_large_batch_count(test, device):
     rng = np.random.default_rng(42)
     low = -4.5
@@ -387,31 +363,38 @@ def test_large_batch_count(test, device):
     k = 4
     batch_count = 65535 * 2 + int(65535 / 2)
     A = wp.array3d(
-        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))), dtype=float, device=device, requires_grad=True
+        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))),
+        dtype=float,
+        device=device,
+        requires_grad=True,
     )
     B = wp.array3d(
-        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))), dtype=float, device=device, requires_grad=True
+        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))),
+        dtype=float,
+        device=device,
+        requires_grad=True,
     )
     C = wp.array3d(
-        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))), dtype=float, device=device, requires_grad=True
-    )
-    D = wp.array3d(
-        np.zeros((batch_count, m, n)), dtype=float, device=device, requires_grad=True
+        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))),
+        dtype=float,
+        device=device,
+        requires_grad=True,
     )
+    D = wp.array3d(np.zeros((batch_count, m, n)), dtype=float, device=device, requires_grad=True)
     ones = wp.zeros_like(D)
     ones.fill_(1.0)
 
     alpha = 1.0
     beta = 1.0
-    
+
     tape = wp.Tape()
     with tape:
         wp.batched_matmul(A, B, C, D, alpha=alpha, beta=beta, allow_tf32x3_arith=False, device=device)
-    tape.backward(grads={D : ones})
+    tape.backward(grads={D: ones})
 
     D_np = alpha * np.matmul(A.numpy(), B.numpy()) + beta * C.numpy()
     assert np.array_equal(D_np, D.numpy())
-    
+
     adj_A_np = alpha * np.matmul(ones.numpy(), B.numpy().transpose((0, 2, 1)))
     adj_B_np = alpha * np.matmul(A.numpy().transpose((0, 2, 1)), ones.numpy())
     adj_C_np = beta * ones.numpy()
@@ -449,31 +432,22 @@ def test_adjoint_accumulation(test, device):
     assert np.array_equal(c_wp.grad.numpy(), np.ones(shape=(2,2)))
 
 
-def register(parent):
-    devices = [d for d in get_test_devices()]
+devices = get_test_devices()
 
-    class TestMatmul(parent):
-        pass
 
-    if devices:
-        # check if CUTLASS is available
-        from warp.context import runtime
+class TestMatmul(unittest.TestCase):
+    pass
 
-        if runtime.core.is_cutlass_enabled():
-            # add_function_test(TestMatmul, "test_f16", test_f16, devices=devices)
-            add_function_test(TestMatmul, "test_f32", test_f32, devices=devices)
-            add_function_test(TestMatmul, "test_f64", test_f64, devices=devices)
-            add_function_test(TestMatmul, "test_tape", test_tape, devices=devices)
-            add_function_test(TestMatmul, "test_operator", test_operator, devices=devices)
-            add_function_test(TestMatmul, "test_large_batch_count", test_large_batch_count, devices=devices)
-            add_function_test(TestMatmul, "test_adjoint_accumulation", test_adjoint_accumulation, devices=devices)
-        else:
-            print("Skipping matmul tests because CUTLASS is not supported in this build")
 
-    return TestMatmul
+# add_function_test(TestMatmul, "test_f16", test_f16, devices=devices)
+add_function_test(TestMatmul, "test_f32", test_f32, devices=devices)
+add_function_test(TestMatmul, "test_f64", test_f64, devices=devices)
+add_function_test(TestMatmul, "test_tape", test_tape, devices=devices)
+add_function_test(TestMatmul, "test_operator", test_operator, devices=devices)
+add_function_test(TestMatmul, "test_large_batch_count", test_large_batch_count, devices=devices)
+add_function_test(TestMatmul, "test_adjoint_accumulation", test_adjoint_accumulation, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_matmul_lite.py b/warp/tests/test_matmul_lite.py
index 55e995af5..d3e00d5a8 100644
--- a/warp/tests/test_matmul_lite.py
+++ b/warp/tests/test_matmul_lite.py
@@ -1,11 +1,21 @@
-import numpy as np
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
+import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
+from warp.context import runtime  # noqa: E402
+
 
 class gemm_test_bed_runner:
     def __init__(self, dtype, device):
@@ -21,63 +31,54 @@ def alloc(self, m, n, k, batch_count):
                 np.ceil(rng.uniform(low=low, high=high, size=(m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
-            D = wp.array2d(
-                np.zeros((m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True)
+            D = wp.array2d(np.zeros((m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
         else:
             A = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
-            )
-            D = wp.array3d(
-                np.zeros((batch_count, m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
+            D = wp.array3d(np.zeros((batch_count, m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
         return A, B, C, D
 
     def run_and_verify(self, m, n, k, batch_count, alpha, beta):
         A, B, C, D = self.alloc(m, n, k, batch_count)
         ones = wp.zeros_like(D)
         ones.fill_(1.0)
-        
+
         if batch_count == 1:
             tape = wp.Tape()
             with tape:
                 wp.matmul(A, B, C, D, alpha, beta, False, self.device)
-            tape.backward(grads={D : ones})
-            
+            tape.backward(grads={D: ones})
+
             D_np = alpha * (A.numpy() @ B.numpy()) + beta * C.numpy()
             assert np.array_equal(D_np, D.numpy())
 
@@ -89,8 +90,8 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
             tape = wp.Tape()
             with tape:
                 wp.batched_matmul(A, B, C, D, alpha, beta, False, self.device)
-            tape.backward(grads={D : ones})
-            
+            tape.backward(grads={D: ones})
+
             D_np = alpha * np.matmul(A.numpy(), B.numpy()) + beta * C.numpy()
             assert np.array_equal(D_np, D.numpy())
 
@@ -132,75 +133,45 @@ def alloc(self, m, n, k, batch_count):
                 np.ceil(rng.uniform(low=low, high=high, size=(m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array2d(
                 np.ceil(rng.uniform(low=low, high=high, size=(m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
-            )
-            D = wp.array2d(
-                np.zeros((m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            AT = wp.array2d(
-                A.numpy().transpose([1, 0]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            BT = wp.array2d(
-                B.numpy().transpose([1, 0]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
+            D = wp.array2d(np.zeros((m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
+            AT = wp.array2d(A.numpy().transpose([1, 0]), dtype=self.dtype, device=self.device, requires_grad=True)
+            BT = wp.array2d(B.numpy().transpose([1, 0]), dtype=self.dtype, device=self.device, requires_grad=True)
         else:
             A = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             B = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
             C = wp.array3d(
                 np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))),
                 dtype=self.dtype,
                 device=self.device,
-                requires_grad=True
-            )
-            D = wp.array3d(
-                np.zeros((batch_count, m, n)),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            AT = wp.array3d(
-                A.numpy().transpose([0, 2, 1]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
-            )
-            BT = wp.array3d(
-                B.numpy().transpose([0, 2, 1]),
-                dtype=self.dtype,
-                device=self.device,
-                requires_grad=True
+                requires_grad=True,
             )
+            D = wp.array3d(np.zeros((batch_count, m, n)), dtype=self.dtype, device=self.device, requires_grad=True)
+            AT = wp.array3d(A.numpy().transpose([0, 2, 1]), dtype=self.dtype, device=self.device, requires_grad=True)
+            BT = wp.array3d(B.numpy().transpose([0, 2, 1]), dtype=self.dtype, device=self.device, requires_grad=True)
         return A, B, C, D, AT, BT
 
     def run_and_verify(self, m, n, k, batch_count, alpha, beta):
@@ -219,17 +190,17 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
         ones3.fill_(1.0)
 
         if batch_count == 1:
-            ATT1 = AT1.transpose([1, 0])        
+            ATT1 = AT1.transpose([1, 0])
             BTT1 = BT1.transpose([1, 0])
-            ATT2 = AT2.transpose([1, 0])        
+            ATT2 = AT2.transpose([1, 0])
             BTT2 = BT2.transpose([1, 0])
             tape = wp.Tape()
             with tape:
                 wp.matmul(A, BTT1, C1, D1, alpha, beta, False, self.device)
                 wp.matmul(ATT1, B, C2, D2, alpha, beta, False, self.device)
                 wp.matmul(ATT2, BTT2, C3, D3, alpha, beta, False, self.device)
-            tape.backward(grads={D1 : ones1, D2 : ones2, D3 : ones3})
-            
+            tape.backward(grads={D1: ones1, D2: ones2, D3: ones3})
+
             D_np = alpha * (A.numpy() @ B.numpy()) + beta * C1.numpy()
             assert np.array_equal(D_np, D1.numpy())
             assert np.array_equal(D_np, D2.numpy())
@@ -240,7 +211,7 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
             adj_C_np = beta * ones1.numpy()
 
         else:
-            ATT1 = AT1.transpose([0, 2, 1])        
+            ATT1 = AT1.transpose([0, 2, 1])
             BTT1 = BT1.transpose([0, 2, 1])
             ATT2 = AT2.transpose([0, 2, 1])
             BTT2 = BT2.transpose([0, 2, 1])
@@ -249,8 +220,8 @@ def run_and_verify(self, m, n, k, batch_count, alpha, beta):
                 wp.batched_matmul(A, BTT1, C1, D1, alpha, beta, False, self.device)
                 wp.batched_matmul(ATT1, B, C2, D2, alpha, beta, False, self.device)
                 wp.batched_matmul(ATT2, BTT2, C3, D3, alpha, beta, False, self.device)
-            tape.backward(grads={D1 : ones1, D2 : ones2, D3 : ones3})
-            
+            tape.backward(grads={D1: ones1, D2: ones2, D3: ones3})
+
             D_np = alpha * np.matmul(A.numpy(), B.numpy()) + beta * C1.numpy()
             assert np.array_equal(D_np, D1.numpy())
             assert np.array_equal(D_np, D2.numpy())
@@ -282,6 +253,7 @@ def run(self):
             self.run_and_verify(m, n, k, batch_count, alpha, beta)
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_f32(test, device):
     gemm_test_bed_runner(wp.float32, device).run()
     gemm_test_bed_runner_transpose(wp.float32, device).run()
@@ -293,6 +265,7 @@ def matrix_sum_kernel(arr: wp.array2d(dtype=float), loss: wp.array(dtype=float))
     wp.atomic_add(loss, 0, arr[i, j])
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_tape(test, device):
     rng = np.random.default_rng(42)
     low = -4.5
@@ -331,6 +304,7 @@ def test_tape(test, device):
     assert_array_equal(A.grad, wp.zeros_like(A))
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_operator(test, device):
     rng = np.random.default_rng(42)
     low = -4.5
@@ -366,6 +340,7 @@ def test_operator(test, device):
     assert_array_equal(A.grad, wp.zeros_like(A))
 
 
+@unittest.skipUnless(runtime.core.is_cutlass_enabled(), "Warp was not built with CUTLASS support")
 def test_large_batch_count(test, device):
     rng = np.random.default_rng(42)
     low = -4.5
@@ -375,31 +350,38 @@ def test_large_batch_count(test, device):
     k = 4
     batch_count = 65535 * 2 + int(65535 / 2)
     A = wp.array3d(
-        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))), dtype=float, device=device, requires_grad=True
+        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, k))),
+        dtype=float,
+        device=device,
+        requires_grad=True,
     )
     B = wp.array3d(
-        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))), dtype=float, device=device, requires_grad=True
+        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, k, n))),
+        dtype=float,
+        device=device,
+        requires_grad=True,
     )
     C = wp.array3d(
-        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))), dtype=float, device=device, requires_grad=True
-    )
-    D = wp.array3d(
-        np.zeros((batch_count, m, n)), dtype=float, device=device, requires_grad=True
+        np.ceil(rng.uniform(low=low, high=high, size=(batch_count, m, n))),
+        dtype=float,
+        device=device,
+        requires_grad=True,
     )
+    D = wp.array3d(np.zeros((batch_count, m, n)), dtype=float, device=device, requires_grad=True)
     ones = wp.zeros_like(D)
     ones.fill_(1.0)
 
     alpha = 1.0
     beta = 1.0
-    
+
     tape = wp.Tape()
     with tape:
         wp.batched_matmul(A, B, C, D, alpha=alpha, beta=beta, allow_tf32x3_arith=False, device=device)
-    tape.backward(grads={D : ones})
+    tape.backward(grads={D: ones})
 
     D_np = alpha * np.matmul(A.numpy(), B.numpy()) + beta * C.numpy()
     assert np.array_equal(D_np, D.numpy())
-    
+
     adj_A_np = alpha * np.matmul(ones.numpy(), B.numpy().transpose((0, 2, 1)))
     adj_B_np = alpha * np.matmul(A.numpy().transpose((0, 2, 1)), ones.numpy())
     adj_C_np = beta * ones.numpy()
@@ -409,28 +391,19 @@ def test_large_batch_count(test, device):
     assert np.array_equal(adj_C_np, C.grad.numpy())
 
 
-def register(parent):
-    devices = [d for d in get_test_devices()]
+devices = get_test_devices()
 
-    class TestMatmul(parent):
-        pass
 
-    if devices:
-        # check if CUTLASS is available
-        from warp.context import runtime
+class TestMatmulLite(unittest.TestCase):
+    pass
 
-        if runtime.core.is_cutlass_enabled():
-            add_function_test(TestMatmul, "test_f32", test_f32, devices=devices)
-            add_function_test(TestMatmul, "test_tape", test_tape, devices=devices)
-            add_function_test(TestMatmul, "test_operator", test_operator, devices=devices)
-            add_function_test(TestMatmul, "test_large_batch_count", test_large_batch_count, devices=devices)
-        else:
-            print("Skipping matmul tests because CUTLASS is not supported in this build")
 
-    return TestMatmul
+add_function_test(TestMatmulLite, "test_f32", test_f32, devices=devices)
+add_function_test(TestMatmulLite, "test_tape", test_tape, devices=devices)
+add_function_test(TestMatmulLite, "test_operator", test_operator, devices=devices)
+add_function_test(TestMatmulLite, "test_large_batch_count", test_large_batch_count, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_mesh.py b/warp/tests/test_mesh.py
index fc08f7eec..c586072ee 100644
--- a/warp/tests/test_mesh.py
+++ b/warp/tests/test_mesh.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 # fmt: off
 
@@ -222,9 +222,9 @@ def query_ray_kernel(
 
 
 def test_mesh_query_ray(test, device):
-    points = wp.array(POINT_POSITIONS, dtype=wp.vec3)
+    points = wp.array(POINT_POSITIONS, dtype=wp.vec3, device=device)
 
-    indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+    indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
     mesh = wp.Mesh(points=points, indices=indices)
     expected_sign = -1.0
     wp.launch(
@@ -234,9 +234,10 @@ def test_mesh_query_ray(test, device):
             mesh.id,
             expected_sign,
         ],
+        device=device,
     )
 
-    indices = wp.array(LEFT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+    indices = wp.array(LEFT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
     mesh = wp.Mesh(points=points, indices=indices)
     expected_sign = 1.0
     wp.launch(
@@ -246,76 +247,78 @@ def test_mesh_query_ray(test, device):
             mesh.id,
             expected_sign,
         ],
+        device=device,
     )
 
 
 def test_mesh_refit_graph(test, device):
-    points = wp.array(POINT_POSITIONS, dtype=wp.vec3)
+    points = wp.array(POINT_POSITIONS, dtype=wp.vec3, device=device)
 
-    indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+    indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
     mesh = wp.Mesh(points=points, indices=indices)
 
-    wp.capture_begin()
-
-    mesh.refit()
-
-    graph = wp.capture_end()
+    wp.capture_begin(device, force_module_load=False)
+    try:
+        mesh.refit()
+    finally:
+        graph = wp.capture_end(device)
 
     # replay
     num_iters = 10
     for _ in range(num_iters):
         wp.capture_launch(graph)
 
+    wp.synchronize_device(device)
+
 
 def test_mesh_exceptions(test, device):
     # points and indices must be on same device
     with test.assertRaises(RuntimeError):
         points = wp.array(POINT_POSITIONS, dtype=wp.vec3, device="cpu")
-        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
         wp.Mesh(points=points, indices=indices)
 
     # points must be vec3
     with test.assertRaises(RuntimeError):
-        points = wp.array(POINT_POSITIONS, dtype=wp.vec3d)
-        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+        points = wp.array(POINT_POSITIONS, dtype=wp.vec3d, device=device)
+        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
         wp.Mesh(points=points, indices=indices)
 
     # velocities must be vec3
     with test.assertRaises(RuntimeError):
-        points = wp.array(POINT_POSITIONS, dtype=wp.vec3)
-        velocities = wp.zeros(points.shape, dtype=wp.vec3d)
-        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+        points = wp.array(POINT_POSITIONS, dtype=wp.vec3, device=device)
+        velocities = wp.zeros(points.shape, dtype=wp.vec3d, device=device)
+        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
         wp.Mesh(points=points, indices=indices, velocities=velocities)
 
     # indices must be int32
     with test.assertRaises(RuntimeError):
-        points = wp.array(POINT_POSITIONS, dtype=wp.vec3)
-        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=wp.int64)
+        points = wp.array(POINT_POSITIONS, dtype=wp.vec3, device=device)
+        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=wp.int64, device=device)
         wp.Mesh(points=points, indices=indices)
 
     # indices must be 1d
     with test.assertRaises(RuntimeError):
-        points = wp.array(POINT_POSITIONS, dtype=wp.vec3)
-        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int)
+        points = wp.array(POINT_POSITIONS, dtype=wp.vec3, device=device)
+        indices = wp.array(RIGHT_HANDED_FACE_VERTEX_INDICES, dtype=int, device=device)
         indices = indices.reshape((3, -1))
         wp.Mesh(points=points, indices=indices)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
+
+class TestMesh(unittest.TestCase):
+    pass
 
-    class TestMesh(parent):
-        pass
 
-    add_function_test(TestMesh, "test_mesh_read_properties", test_mesh_read_properties, devices=devices)
-    add_function_test(TestMesh, "test_mesh_query_point", test_mesh_query_point, devices=devices)
-    add_function_test(TestMesh, "test_mesh_query_ray", test_mesh_query_ray, devices=devices)
-    add_function_test(TestMesh, "test_mesh_refit_graph", test_mesh_refit_graph, devices=wp.get_cuda_devices())
-    add_function_test(TestMesh, "test_mesh_exceptions", test_mesh_exceptions, devices=wp.get_cuda_devices())
-    return TestMesh
+add_function_test(TestMesh, "test_mesh_read_properties", test_mesh_read_properties, devices=devices)
+add_function_test(TestMesh, "test_mesh_query_point", test_mesh_query_point, devices=devices)
+add_function_test(TestMesh, "test_mesh_query_ray", test_mesh_query_ray, devices=devices)
+add_function_test(TestMesh, "test_mesh_refit_graph", test_mesh_refit_graph, devices=get_unique_cuda_test_devices())
+add_function_test(TestMesh, "test_mesh_exceptions", test_mesh_exceptions, devices=get_unique_cuda_test_devices())
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_mesh_query_aabb.py b/warp/tests/test_mesh_query_aabb.py
index 3536e9c6b..9feeb21d2 100644
--- a/warp/tests/test_mesh_query_aabb.py
+++ b/warp/tests/test_mesh_query_aabb.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -98,7 +98,6 @@ def test_compute_bounds(test, device):
 
     lower_view = lowers.numpy()
     upper_view = uppers.numpy()
-    wp.synchronize()
 
     # Confirm the bounds of each triangle are correct.
     test.assertTrue(lower_view[0][0] == 0)
@@ -150,8 +149,6 @@ def test_mesh_query_aabb_count_overlap(test, device):
         device=device,
     )
 
-    wp.synchronize()
-
     view = counts.numpy()
 
     # 2 triangles that share a vertex having overlapping AABBs.
@@ -190,8 +187,6 @@ def test_mesh_query_aabb_count_nonoverlap(test, device):
         device=device,
     )
 
-    wp.synchronize()
-
     view = counts.numpy()
 
     # AABB query only returns one triangle at a time, the triangles are not close enough to overlap.
@@ -199,30 +194,28 @@ def test_mesh_query_aabb_count_nonoverlap(test, device):
         test.assertTrue(c == 1)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestMeshQueryAABBMethods(parent):
-        pass
 
-    add_function_test(TestMeshQueryAABBMethods, "test_compute_bounds", test_compute_bounds, devices=devices)
-    add_function_test(
-        TestMeshQueryAABBMethods,
-        "test_mesh_query_aabb_count_overlap",
-        test_mesh_query_aabb_count_overlap,
-        devices=devices,
-    )
-    add_function_test(
-        TestMeshQueryAABBMethods,
-        "test_mesh_query_aabb_count_nonoverlap",
-        test_mesh_query_aabb_count_nonoverlap,
-        devices=devices,
-    )
+class TestMeshQueryAABBMethods(unittest.TestCase):
+    pass
+
 
-    return TestMeshQueryAABBMethods
+add_function_test(TestMeshQueryAABBMethods, "test_compute_bounds", test_compute_bounds, devices=devices)
+add_function_test(
+    TestMeshQueryAABBMethods,
+    "test_mesh_query_aabb_count_overlap",
+    test_mesh_query_aabb_count_overlap,
+    devices=devices,
+)
+add_function_test(
+    TestMeshQueryAABBMethods,
+    "test_mesh_query_aabb_count_nonoverlap",
+    test_mesh_query_aabb_count_nonoverlap,
+    devices=devices,
+)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_mesh_query_point.py b/warp/tests/test_mesh_query_point.py
index 9650287ed..21bd85fb3 100644
--- a/warp/tests/test_mesh_query_point.py
+++ b/warp/tests/test_mesh_query_point.py
@@ -5,14 +5,13 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
+import math
 import unittest
 
 import numpy as np
-import math
 
 import warp as wp
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -288,6 +287,7 @@ def triangulate(face_counts, face_indices):
     return tri_indices
 
 
+@unittest.skipUnless(USD_AVAILABLE, "Requires usd-core")
 def test_mesh_query_point(test, device):
     from pxr import Usd, UsdGeom
 
@@ -503,6 +503,7 @@ def mesh_query_point_loss(
     loss[tid] = dist
 
 
+@unittest.skipUnless(USD_AVAILABLE, "Requires usd-core")
 def test_adj_mesh_query_point(test, device):
     from pxr import Usd, UsdGeom
 
@@ -621,6 +622,7 @@ def sample_furthest_points_brute(
     query_result[tid] = max_dist_sq
 
 
+@unittest.skipUnless(USD_AVAILABLE, "Requires usd-core")
 def test_mesh_query_furthest_point(test, device):
     from pxr import Usd, UsdGeom
 
@@ -654,32 +656,18 @@ def test_mesh_query_furthest_point(test, device):
     assert_np_equal(dist_query.numpy(), dist_brute.numpy(), tol=1.0e-3)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestMeshQuery(parent):
-        pass
 
-    # USD import failures should not count as a test failure
-    try:
-        from pxr import Usd, UsdGeom
+class TestMeshQueryPoint(unittest.TestCase):
+    pass
 
-        have_usd = True
-    except:
-        have_usd = False
 
-    if have_usd:
-        add_function_test(TestMeshQuery, "test_mesh_query_point", test_mesh_query_point, devices=devices)
-        add_function_test(
-            TestMeshQuery, "test_mesh_query_furthest_point", test_mesh_query_furthest_point, devices=devices
-        )
-        add_function_test(TestMeshQuery, "test_adj_mesh_query_point", test_adj_mesh_query_point, devices=devices)
-
-    return TestMeshQuery
+add_function_test(TestMeshQueryPoint, "test_mesh_query_point", test_mesh_query_point, devices=devices)
+add_function_test(TestMeshQueryPoint, "test_mesh_query_furthest_point", test_mesh_query_furthest_point, devices=devices)
+add_function_test(TestMeshQueryPoint, "test_adj_mesh_query_point", test_adj_mesh_query_point, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
-
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_mesh_query_ray.py b/warp/tests/test_mesh_query_ray.py
index 029204678..12d547d61 100644
--- a/warp/tests/test_mesh_query_ray.py
+++ b/warp/tests/test_mesh_query_ray.py
@@ -10,8 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -66,8 +65,9 @@ def mesh_query_ray_loss(
     loss[tid] = l
 
 
+@unittest.skipUnless(USD_AVAILABLE, "Requires usd-core")
 def test_mesh_query_ray_grad(test, device):
-    from pxr import Usd, UsdGeom, Gf, Sdf
+    from pxr import Usd, UsdGeom
 
     # test tri
     # print("Testing Single Triangle")
@@ -209,7 +209,7 @@ def raycast_kernel(
     sign = float(0.0)  # hit face sign
     n = wp.vec3()  # hit face normal
     f = int(0)  # hit face index
-    max_dist = 1e6  # max raycast disance
+    max_dist = 1e6  # max raycast distance
 
     # ray cast against the mesh
     tid = wp.tid()
@@ -258,29 +258,17 @@ def test_mesh_query_ray_edge(test, device):
     test.assertEqual(counts.numpy()[0], n)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestMeshQueryRay(parent):
-        pass
-
-    add_function_test(TestMeshQueryRay, "test_mesh_query_ray_edge", test_mesh_query_ray_edge, devices=devices)
+devices = get_test_devices()
 
-    # USD import failures should not count as a test failure
-    try:
-        from pxr import Usd, UsdGeom
 
-        have_usd = True
-    except:
-        have_usd = False
+class TestMeshQueryRay(unittest.TestCase):
+    pass
 
-    if have_usd:
-        add_function_test(TestMeshQueryRay, "test_mesh_query_ray_grad", test_mesh_query_ray_grad, devices=devices)
 
-    return TestMeshQueryRay
+add_function_test(TestMeshQueryRay, "test_mesh_query_ray_edge", test_mesh_query_ray_edge, devices=devices)
+add_function_test(TestMeshQueryRay, "test_mesh_query_ray_grad", test_mesh_query_ray_grad, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_mlp.py b/warp/tests/test_mlp.py
index ef241cc3e..2af60f470 100644
--- a/warp/tests/test_mlp.py
+++ b/warp/tests/test_mlp.py
@@ -8,8 +8,9 @@
 import unittest
 
 import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -259,19 +260,17 @@ def profile_mlp_warp(device):
 # profile_mlp_torch("cuda")
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestMLP(parent):
-        pass
+class TestMLP(unittest.TestCase):
+    pass
 
-    add_function_test(TestMLP, "test_mlp", test_mlp, devices=devices)
-    add_function_test(TestMLP, "test_mlp_grad", test_mlp_grad, devices=devices)
 
-    return TestMLP
+add_function_test(TestMLP, "test_mlp", test_mlp, devices=devices)
+add_function_test(TestMLP, "test_mlp_grad", test_mlp_grad, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_model.py b/warp/tests/test_model.py
index f88e96657..b5caeb547 100644
--- a/warp/tests/test_model.py
+++ b/warp/tests/test_model.py
@@ -7,108 +7,104 @@
 
 import unittest
 
+import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
 from warp.sim import ModelBuilder
-
-import numpy as np
+from warp.tests.unittest_utils import *
 
 wp.init()
 
 
-def register(parent):
-    class TestModel(parent):
-        def test_add_triangles(self):
-            rng = np.random.default_rng(123)
-
-            pts = np.array(
-                [
-                    [-0.00585869, 0.34189449, -1.17415233],
-                    [-1.894547, 0.1788074, 0.9251329],
-                    [-1.26141048, 0.16140787, 0.08823282],
-                    [-0.08609255, -0.82722546, 0.65995427],
-                    [0.78827592, -1.77375711, -0.55582718],
-                ]
-            )
-            tris = np.array([[0, 3, 4], [0, 2, 3], [2, 1, 3], [1, 4, 3]])
-
-            builder1 = ModelBuilder()
-            builder2 = ModelBuilder()
-            for pt in pts:
-                builder1.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
-                builder2.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
-
-            # test add_triangle(s) with default arguments:
-            areas = builder2.add_triangles(tris[:, 0], tris[:, 1], tris[:, 2])
-            for i, t in enumerate(tris):
-                area = builder1.add_triangle(t[0], t[1], t[2])
-                self.assertAlmostEqual(area, areas[i], places=6)
-
-            # test add_triangle(s) with non default arguments:
-            tri_ke = rng.standard_normal(size=pts.shape[0])
-            tri_ka = rng.standard_normal(size=pts.shape[0])
-            tri_kd = rng.standard_normal(size=pts.shape[0])
-            tri_drag = rng.standard_normal(size=pts.shape[0])
-            tri_lift = rng.standard_normal(size=pts.shape[0])
-            for i, t in enumerate(tris):
-                builder1.add_triangle(
-                    t[0],
-                    t[1],
-                    t[2],
-                    tri_ke[i],
-                    tri_ka[i],
-                    tri_kd[i],
-                    tri_drag[i],
-                    tri_lift[i],
-                )
-            builder2.add_triangles(tris[:, 0], tris[:, 1], tris[:, 2], tri_ke, tri_ka, tri_kd, tri_drag, tri_lift)
-
-            assert_np_equal(np.array(builder1.tri_indices), np.array(builder2.tri_indices))
-            assert_np_equal(np.array(builder1.tri_poses), np.array(builder2.tri_poses), tol=1.0e-6)
-            assert_np_equal(np.array(builder1.tri_activations), np.array(builder2.tri_activations))
-            assert_np_equal(np.array(builder1.tri_materials), np.array(builder2.tri_materials))
-
-        def test_add_edges(self):
-            rng = np.random.default_rng(123)
-
-            pts = np.array(
-                [
-                    [-0.00585869, 0.34189449, -1.17415233],
-                    [-1.894547, 0.1788074, 0.9251329],
-                    [-1.26141048, 0.16140787, 0.08823282],
-                    [-0.08609255, -0.82722546, 0.65995427],
-                    [0.78827592, -1.77375711, -0.55582718],
-                ]
+class TestModel(unittest.TestCase):
+    def test_add_triangles(self):
+        rng = np.random.default_rng(123)
+
+        pts = np.array(
+            [
+                [-0.00585869, 0.34189449, -1.17415233],
+                [-1.894547, 0.1788074, 0.9251329],
+                [-1.26141048, 0.16140787, 0.08823282],
+                [-0.08609255, -0.82722546, 0.65995427],
+                [0.78827592, -1.77375711, -0.55582718],
+            ]
+        )
+        tris = np.array([[0, 3, 4], [0, 2, 3], [2, 1, 3], [1, 4, 3]])
+
+        builder1 = ModelBuilder()
+        builder2 = ModelBuilder()
+        for pt in pts:
+            builder1.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
+            builder2.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
+
+        # test add_triangle(s) with default arguments:
+        areas = builder2.add_triangles(tris[:, 0], tris[:, 1], tris[:, 2])
+        for i, t in enumerate(tris):
+            area = builder1.add_triangle(t[0], t[1], t[2])
+            self.assertAlmostEqual(area, areas[i], places=6)
+
+        # test add_triangle(s) with non default arguments:
+        tri_ke = rng.standard_normal(size=pts.shape[0])
+        tri_ka = rng.standard_normal(size=pts.shape[0])
+        tri_kd = rng.standard_normal(size=pts.shape[0])
+        tri_drag = rng.standard_normal(size=pts.shape[0])
+        tri_lift = rng.standard_normal(size=pts.shape[0])
+        for i, t in enumerate(tris):
+            builder1.add_triangle(
+                t[0],
+                t[1],
+                t[2],
+                tri_ke[i],
+                tri_ka[i],
+                tri_kd[i],
+                tri_drag[i],
+                tri_lift[i],
             )
-            edges = np.array([[0, 4, 3, 1], [3, 2, 4, 1]])
-
-            builder1 = ModelBuilder()
-            builder2 = ModelBuilder()
-            for pt in pts:
-                builder1.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
-                builder2.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
-
-            # test defaults:
-            for i in range(2):
-                builder1.add_edge(edges[i, 0], edges[i, 1], edges[i, 2], edges[i, 3])
-            builder2.add_edges(edges[:, 0], edges[:, 1], edges[:, 2], edges[:, 3])
-
-            # test non defaults:
-            rest = rng.standard_normal(size=2)
-            edge_ke = rng.standard_normal(size=2)
-            edge_kd = rng.standard_normal(size=2)
-            for i in range(2):
-                builder1.add_edge(edges[i, 0], edges[i, 1], edges[i, 2], edges[i, 3], rest[i], edge_ke[i], edge_kd[i])
-            builder2.add_edges(edges[:, 0], edges[:, 1], edges[:, 2], edges[:, 3], rest, edge_ke, edge_kd)
-
-            assert_np_equal(np.array(builder1.edge_indices), np.array(builder2.edge_indices))
-            assert_np_equal(np.array(builder1.edge_rest_angle), np.array(builder2.edge_rest_angle), tol=1.0e-4)
-            assert_np_equal(np.array(builder1.edge_bending_properties), np.array(builder2.edge_bending_properties))
-
-    return TestModel
+        builder2.add_triangles(tris[:, 0], tris[:, 1], tris[:, 2], tri_ke, tri_ka, tri_kd, tri_drag, tri_lift)
+
+        assert_np_equal(np.array(builder1.tri_indices), np.array(builder2.tri_indices))
+        assert_np_equal(np.array(builder1.tri_poses), np.array(builder2.tri_poses), tol=1.0e-6)
+        assert_np_equal(np.array(builder1.tri_activations), np.array(builder2.tri_activations))
+        assert_np_equal(np.array(builder1.tri_materials), np.array(builder2.tri_materials))
+
+    def test_add_edges(self):
+        rng = np.random.default_rng(123)
+
+        pts = np.array(
+            [
+                [-0.00585869, 0.34189449, -1.17415233],
+                [-1.894547, 0.1788074, 0.9251329],
+                [-1.26141048, 0.16140787, 0.08823282],
+                [-0.08609255, -0.82722546, 0.65995427],
+                [0.78827592, -1.77375711, -0.55582718],
+            ]
+        )
+        edges = np.array([[0, 4, 3, 1], [3, 2, 4, 1]])
+
+        builder1 = ModelBuilder()
+        builder2 = ModelBuilder()
+        for pt in pts:
+            builder1.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
+            builder2.add_particle(wp.vec3(pt), wp.vec3(), 1.0)
+
+        # test defaults:
+        for i in range(2):
+            builder1.add_edge(edges[i, 0], edges[i, 1], edges[i, 2], edges[i, 3])
+        builder2.add_edges(edges[:, 0], edges[:, 1], edges[:, 2], edges[:, 3])
+
+        # test non defaults:
+        rest = rng.standard_normal(size=2)
+        edge_ke = rng.standard_normal(size=2)
+        edge_kd = rng.standard_normal(size=2)
+        for i in range(2):
+            builder1.add_edge(edges[i, 0], edges[i, 1], edges[i, 2], edges[i, 3], rest[i], edge_ke[i], edge_kd[i])
+        builder2.add_edges(edges[:, 0], edges[:, 1], edges[:, 2], edges[:, 3], rest, edge_ke, edge_kd)
+
+        assert_np_equal(np.array(builder1.edge_indices), np.array(builder2.edge_indices))
+        assert_np_equal(np.array(builder1.edge_rest_angle), np.array(builder2.edge_rest_angle), tol=1.0e-4)
+        assert_np_equal(np.array(builder1.edge_bending_properties), np.array(builder2.edge_bending_properties))
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_modules_lite.py b/warp/tests/test_modules_lite.py
index b35a2ae1f..b12983f11 100644
--- a/warp/tests/test_modules_lite.py
+++ b/warp/tests/test_modules_lite.py
@@ -8,42 +8,32 @@
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
 
-def test_module_lite_load(test, device):
-    # Load current module
-    wp.load_module()
+devices = get_test_devices()
 
-    # Load named module
-    wp.load_module(wp.config)
 
-    # Load named module (string)
-    wp.load_module(wp.config, recursive=True)
+class TestModuleLite(unittest.TestCase):
+    def test_module_lite_load(self):
+        # Load current module
+        wp.load_module()
 
+        # Load named module
+        wp.load_module(wp.config)
 
-def test_module_lite_options(test, device):
-    wp.set_module_options({"max_unroll": 8})
-    module_options = wp.get_module_options()
-    test.assertIsInstance(module_options, dict)
-    test.assertEqual(module_options["max_unroll"], 8)
+        # Load named module (string)
+        wp.load_module(wp.config, recursive=True)
 
-
-def register(parent):
-    devices = get_test_devices()
-
-    class TestModuleLite(parent):
-        pass
-
-    add_function_test(TestModuleLite, "test_module_lite_load", test_module_lite_load, devices=devices)
-    add_function_test(TestModuleLite, "test_module_lite_get_options", test_module_lite_options, devices=devices)
-
-    return TestModuleLite
+    def test_module_lite_options(self):
+        wp.set_module_options({"max_unroll": 8})
+        module_options = wp.get_module_options()
+        self.assertIsInstance(module_options, dict)
+        self.assertEqual(module_options["max_unroll"], 8)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_multigpu.py b/warp/tests/test_multigpu.py
index 42216f2bb..533cf35c1 100644
--- a/warp/tests/test_multigpu.py
+++ b/warp/tests/test_multigpu.py
@@ -5,16 +5,12 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import numpy as np
-import warp as wp
+import unittest
 
-import math
+import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
-
-import unittest
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -31,157 +27,134 @@ def arange(start: int, step: int, a: wp.array(dtype=int)):
     a[tid] = start + step * tid
 
 
-def test_multigpu_set_device(test, device):
-    assert len(wp.get_cuda_devices()) > 1, "At least two CUDA devices are required"
-
-    # save default device
-    saved_device = wp.get_device()
-
-    n = 32
-
-    wp.set_device("cuda:0")
-    a0 = wp.empty(n, dtype=int)
-    wp.launch(arange, dim=a0.size, inputs=[0, 1, a0])
-
-    wp.set_device("cuda:1")
-    a1 = wp.empty(n, dtype=int)
-    wp.launch(arange, dim=a1.size, inputs=[0, 1, a1])
-
-    # restore default device
-    wp.set_device(saved_device)
+class TestMultiGPU(unittest.TestCase):
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_multigpu_set_device(self):
+        # save default device
+        saved_device = wp.get_device()
 
-    assert a0.device == "cuda:0"
-    assert a1.device == "cuda:1"
+        n = 32
 
-    expected = np.arange(n, dtype=int)
-
-    assert_np_equal(a0.numpy(), expected)
-    assert_np_equal(a1.numpy(), expected)
-
-
-def test_multigpu_scoped_device(test, device):
-    assert len(wp.get_cuda_devices()) > 1, "At least two CUDA devices are required"
-
-    n = 32
-
-    with wp.ScopedDevice("cuda:0"):
+        wp.set_device("cuda:0")
         a0 = wp.empty(n, dtype=int)
         wp.launch(arange, dim=a0.size, inputs=[0, 1, a0])
 
-    with wp.ScopedDevice("cuda:1"):
+        wp.set_device("cuda:1")
         a1 = wp.empty(n, dtype=int)
         wp.launch(arange, dim=a1.size, inputs=[0, 1, a1])
 
-    assert a0.device == "cuda:0"
-    assert a1.device == "cuda:1"
-
-    expected = np.arange(n, dtype=int)
+        # restore default device
+        wp.set_device(saved_device)
 
-    assert_np_equal(a0.numpy(), expected)
-    assert_np_equal(a1.numpy(), expected)
+        assert a0.device == "cuda:0"
+        assert a1.device == "cuda:1"
 
+        expected = np.arange(n, dtype=int)
 
-def test_multigpu_nesting(test, device):
-    assert len(wp.get_cuda_devices()) > 1, "At least two CUDA devices are required"
+        assert_np_equal(a0.numpy(), expected)
+        assert_np_equal(a1.numpy(), expected)
 
-    initial_device = wp.get_device()
-    initial_cuda_device = wp.get_cuda_device()
-
-    with wp.ScopedDevice("cuda:1"):
-        assert wp.get_device() == "cuda:1"
-        assert wp.get_cuda_device() == "cuda:1"
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_multigpu_scoped_device(self):
+        n = 32
 
         with wp.ScopedDevice("cuda:0"):
-            assert wp.get_device() == "cuda:0"
-            assert wp.get_cuda_device() == "cuda:0"
-
-            with wp.ScopedDevice("cpu"):
-                assert wp.get_device() == "cpu"
-                assert wp.get_cuda_device() == "cuda:0"
+            a0 = wp.empty(n, dtype=int)
+            wp.launch(arange, dim=a0.size, inputs=[0, 1, a0])
 
-                wp.set_device("cuda:1")
+        with wp.ScopedDevice("cuda:1"):
+            a1 = wp.empty(n, dtype=int)
+            wp.launch(arange, dim=a1.size, inputs=[0, 1, a1])
 
-                assert wp.get_device() == "cuda:1"
-                assert wp.get_cuda_device() == "cuda:1"
+        assert a0.device == "cuda:0"
+        assert a1.device == "cuda:1"
 
-            assert wp.get_device() == "cuda:0"
-            assert wp.get_cuda_device() == "cuda:0"
+        expected = np.arange(n, dtype=int)
 
-        assert wp.get_device() == "cuda:1"
-        assert wp.get_cuda_device() == "cuda:1"
+        assert_np_equal(a0.numpy(), expected)
+        assert_np_equal(a1.numpy(), expected)
 
-    assert wp.get_device() == initial_device
-    assert wp.get_cuda_device() == initial_cuda_device
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_multigpu_nesting(self):
+        initial_device = wp.get_device()
+        initial_cuda_device = wp.get_cuda_device()
 
+        with wp.ScopedDevice("cuda:1"):
+            assert wp.get_device() == "cuda:1"
+            assert wp.get_cuda_device() == "cuda:1"
 
-def test_multigpu_pingpong(test, device):
-    assert len(wp.get_cuda_devices()) > 1, "At least two CUDA devices are required"
+            with wp.ScopedDevice("cuda:0"):
+                assert wp.get_device() == "cuda:0"
+                assert wp.get_cuda_device() == "cuda:0"
 
-    n = 1024 * 1024
+                with wp.ScopedDevice("cpu"):
+                    assert wp.get_device() == "cpu"
+                    assert wp.get_cuda_device() == "cuda:0"
 
-    a0 = wp.zeros(n, dtype=float, device="cuda:0")
-    a1 = wp.zeros(n, dtype=float, device="cuda:1")
+                    wp.set_device("cuda:1")
 
-    iters = 10
+                    assert wp.get_device() == "cuda:1"
+                    assert wp.get_cuda_device() == "cuda:1"
 
-    for _ in range(iters):
-        wp.launch(inc, dim=a0.size, inputs=[a0], device=a0.device)
-        wp.synchronize_device(a0.device)
-        wp.copy(a1, a0)
+                assert wp.get_device() == "cuda:0"
+                assert wp.get_cuda_device() == "cuda:0"
 
-        wp.launch(inc, dim=a1.size, inputs=[a1], device=a1.device)
-        wp.synchronize_device(a1.device)
-        wp.copy(a0, a1)
+            assert wp.get_device() == "cuda:1"
+            assert wp.get_cuda_device() == "cuda:1"
 
-    expected = np.full(n, iters * 2, dtype=np.float32)
+        assert wp.get_device() == initial_device
+        assert wp.get_cuda_device() == initial_cuda_device
 
-    assert_np_equal(a0.numpy(), expected)
-    assert_np_equal(a1.numpy(), expected)
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_multigpu_pingpong(self):
+        n = 1024 * 1024
 
+        a0 = wp.zeros(n, dtype=float, device="cuda:0")
+        a1 = wp.zeros(n, dtype=float, device="cuda:1")
 
-def test_multigpu_pingpong_streams(test, device):
-    assert len(wp.get_cuda_devices()) > 1, "At least two CUDA devices are required"
+        iters = 10
 
-    n = 1024 * 1024
+        for _ in range(iters):
+            wp.launch(inc, dim=a0.size, inputs=[a0], device=a0.device)
+            wp.synchronize_device(a0.device)
+            wp.copy(a1, a0)
 
-    a0 = wp.zeros(n, dtype=float, device="cuda:0")
-    a1 = wp.zeros(n, dtype=float, device="cuda:1")
+            wp.launch(inc, dim=a1.size, inputs=[a1], device=a1.device)
+            wp.synchronize_device(a1.device)
+            wp.copy(a0, a1)
 
-    stream0 = wp.get_stream("cuda:0")
-    stream1 = wp.get_stream("cuda:1")
+        expected = np.full(n, iters * 2, dtype=np.float32)
 
-    iters = 10
+        assert_np_equal(a0.numpy(), expected)
+        assert_np_equal(a1.numpy(), expected)
 
-    for _ in range(iters):
-        wp.launch(inc, dim=a0.size, inputs=[a0], stream=stream0)
-        stream1.wait_stream(stream0)
-        wp.copy(a1, a0, stream=stream1)
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_multigpu_pingpong_streams(self):
+        n = 1024 * 1024
 
-        wp.launch(inc, dim=a1.size, inputs=[a1], stream=stream1)
-        stream0.wait_stream(stream1)
-        wp.copy(a0, a1, stream=stream0)
+        a0 = wp.zeros(n, dtype=float, device="cuda:0")
+        a1 = wp.zeros(n, dtype=float, device="cuda:1")
 
-    expected = np.full(n, iters * 2, dtype=np.float32)
+        stream0 = wp.get_stream("cuda:0")
+        stream1 = wp.get_stream("cuda:1")
 
-    assert_np_equal(a0.numpy(), expected)
-    assert_np_equal(a1.numpy(), expected)
+        iters = 10
 
+        for _ in range(iters):
+            wp.launch(inc, dim=a0.size, inputs=[a0], stream=stream0)
+            stream1.wait_stream(stream0)
+            wp.copy(a1, a0, stream=stream1)
 
-def register(parent):
-    class TestMultigpu(parent):
-        pass
+            wp.launch(inc, dim=a1.size, inputs=[a1], stream=stream1)
+            stream0.wait_stream(stream1)
+            wp.copy(a0, a1, stream=stream0)
 
-    if wp.get_cuda_device_count() > 1:
-        add_function_test(TestMultigpu, "test_multigpu_set_device", test_multigpu_set_device)
-        add_function_test(TestMultigpu, "test_multigpu_scoped_device", test_multigpu_scoped_device)
-        add_function_test(TestMultigpu, "test_multigpu_nesting", test_multigpu_nesting)
-        add_function_test(TestMultigpu, "test_multigpu_pingpong", test_multigpu_pingpong)
-        add_function_test(TestMultigpu, "test_multigpu_pingpong_streams", test_multigpu_pingpong_streams)
+        expected = np.full(n, iters * 2, dtype=np.float32)
 
-    return TestMultigpu
+        assert_np_equal(a0.numpy(), expected)
+        assert_np_equal(a1.numpy(), expected)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_noise.py b/warp/tests/test_noise.py
index 99e682f8a..4f0d4cdcd 100644
--- a/warp/tests/test_noise.py
+++ b/warp/tests/test_noise.py
@@ -7,11 +7,11 @@
 
 import unittest
 
-import warp as wp
-from warp.tests.test_base import *
-
 import numpy as np
 
+import warp as wp
+from warp.tests.unittest_utils import *
+
 # import matplotlib.pyplot as plt
 
 wp.init()
@@ -231,20 +231,18 @@ def test_adj_noise(test, device):
     test.assertTrue(err < 1.0e-8)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestNoise(parent):
-        pass
+class TestNoise(unittest.TestCase):
+    pass
 
-    add_function_test(TestNoise, "test_pnoise", test_pnoise, devices=devices)
-    add_function_test(TestNoise, "test_curlnoise", test_curlnoise, devices=devices)
-    add_function_test(TestNoise, "test_adj_noise", test_adj_noise, devices=devices)
 
-    return TestNoise
+add_function_test(TestNoise, "test_pnoise", test_pnoise, devices=devices)
+add_function_test(TestNoise, "test_curlnoise", test_curlnoise, devices=devices)
+add_function_test(TestNoise, "test_adj_noise", test_adj_noise, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_operators.py b/warp/tests/test_operators.py
index c4c729bda..7f8ab7017 100644
--- a/warp/tests/test_operators.py
+++ b/warp/tests/test_operators.py
@@ -7,11 +7,8 @@
 
 import unittest
 
-import numpy as np
-import math
-
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -229,29 +226,25 @@ def test_operators_mat44():
     expect_eq(r0[3], wp.vec4(39.0, 42.0, 45.0, 48.0))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestOperators(parent):
-        pass
+class TestOperators(unittest.TestCase):
+    pass
 
-    add_kernel_test(TestOperators, test_operators_scalar_float, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_scalar_int, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_matrix_index, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_vector_index, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_vec3, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_vec4, dim=1, devices=devices)
 
-    add_kernel_test(TestOperators, test_operators_mat22, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_mat33, dim=1, devices=devices)
-    add_kernel_test(TestOperators, test_operators_mat44, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_scalar_float, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_scalar_int, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_matrix_index, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_vector_index, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_vec3, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_vec4, dim=1, devices=devices)
 
-    return TestOperators
+add_kernel_test(TestOperators, test_operators_mat22, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_mat33, dim=1, devices=devices)
+add_kernel_test(TestOperators, test_operators_mat44, dim=1, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    wp.force_load()
-
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_options.py b/warp/tests/test_options.py
index 1d1d32bed..ae92d66df 100644
--- a/warp/tests/test_options.py
+++ b/warp/tests/test_options.py
@@ -8,7 +8,7 @@
 import unittest
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -93,20 +93,19 @@ def test_options_4(test, device):
     assert_np_equal(tape.gradients[x].numpy(), np.array(0.0))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestOptions(parent):
-        pass
 
-    add_function_test(TestOptions, "test_options_1", test_options_1, devices=devices)
-    add_function_test(TestOptions, "test_options_2", test_options_2, devices=devices)
-    add_function_test(TestOptions, "test_options_3", test_options_3, devices=devices)
-    add_function_test(TestOptions, "test_options_4", test_options_4, devices=devices)
-    return TestOptions
+class TestOptions(unittest.TestCase):
+    pass
+
+
+add_function_test(TestOptions, "test_options_1", test_options_1, devices=devices)
+add_function_test(TestOptions, "test_options_2", test_options_2, devices=devices)
+add_function_test(TestOptions, "test_options_3", test_options_3, devices=devices)
+add_function_test(TestOptions, "test_options_4", test_options_4, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_pinned.py b/warp/tests/test_pinned.py
index c3908101a..81f781194 100644
--- a/warp/tests/test_pinned.py
+++ b/warp/tests/test_pinned.py
@@ -5,16 +5,17 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
+import unittest
+
 import numpy as np
-import warp as wp
-from warp.tests.test_base import *
 
-import unittest
+import warp as wp
+from warp.tests.unittest_utils import *
 
 wp.init()
 
 
-def test_pinned(test, device):
+def test_pinned(test: unittest.TestCase, device):
     assert wp.get_device(device).is_cuda, "Test device must be a CUDA device"
 
     n = 1024 * 1024
@@ -25,20 +26,20 @@ def test_pinned(test, device):
     a_pageable1 = wp.array(ones, dtype=float, device="cpu")
     a_pageable2 = wp.zeros_like(a_pageable1)
 
-    assert a_pageable1.pinned == False
-    assert a_pageable2.pinned == False
+    test.assertFalse(a_pageable1.pinned)
+    test.assertFalse(a_pageable2.pinned)
 
     # pinned host arrays for asynchronous transfers
     a_pinned1 = wp.array(ones, dtype=float, device="cpu", pinned=True)
     a_pinned2 = wp.zeros_like(a_pinned1)
 
-    assert a_pinned1.pinned == True
-    assert a_pinned2.pinned == True
+    test.assertTrue(a_pinned1.pinned)
+    test.assertTrue(a_pinned2.pinned)
 
     # device array
     a_device = wp.zeros(n, dtype=float, device=device)
 
-    assert a_device.pinned == False
+    test.assertFalse(a_device.pinned)
 
     wp.synchronize_device(device)
 
@@ -59,22 +60,19 @@ def test_pinned(test, device):
     assert_np_equal(a_pinned2.numpy(), ones)
 
     # ensure that launching asynchronous transfers took less CPU time
-    assert pinned_timer.elapsed < pageable_timer.elapsed, "Pinned transfers did not take less CPU time"
+    test.assertTrue(pinned_timer.elapsed < pageable_timer.elapsed, "Pinned transfers did not take less CPU time")
+
 
+devices = get_unique_cuda_test_devices()
 
-def register(parent):
-    cuda_devices = wp.get_cuda_devices()
 
-    class TestPinned(parent):
-        pass
+class TestPinned(unittest.TestCase):
+    pass
 
-    if cuda_devices:
-        add_function_test(TestPinned, "test_pinned", test_pinned, devices=cuda_devices)
 
-    return TestPinned
+add_function_test(TestPinned, "test_pinned", test_pinned, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_print.py b/warp/tests/test_print.py
index d5c33a445..c90ad4722 100644
--- a/warp/tests/test_print.py
+++ b/warp/tests/test_print.py
@@ -5,13 +5,10 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import os
 import unittest
 
 import warp as wp
-
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -25,33 +22,32 @@ def test_print_kernel():
 
 
 def test_print(test, device):
+    wp.load_module(device=device)
     capture = StdOutCapture()
     capture.begin()
     wp.launch(kernel=test_print_kernel, dim=1, inputs=[], device=device)
-    wp.synchronize()
+    wp.synchronize_device(device)
     s = capture.end()
 
-    test.assertRegex(
-        s,
-        rf"1{os.linesep}"
-        rf"this is a string{os.linesep}"
-        rf"this is a float 457\.500000{os.linesep}"
-        rf"this is an int 123",
-    )
+    # The CPU kernel printouts don't get captured by StdOutCapture()
+    if device.is_cuda:
+        test.assertRegex(
+            s,
+            rf"1{os.linesep}"
+            rf"this is a string{os.linesep}"
+            rf"this is a float 457\.500000{os.linesep}"
+            rf"this is an int 123",
+        )
 
 
-def register(parent):
-    devices = get_test_devices()
-    devices = tuple(x for x in devices if not x.is_cpu)
+class TestPrint(unittest.TestCase):
+    pass
 
-    class TestPrint(parent):
-        pass
 
-    add_function_test(TestPrint, "test_print", test_print, devices=devices, check_output=False)
-    return TestPrint
+devices = get_test_devices()
+add_function_test(TestPrint, "test_print", test_print, devices=devices, check_output=False)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_quat.py b/warp/tests/test_quat.py
index b7828ece7..fea6955ca 100644
--- a/warp/tests/test_quat.py
+++ b/warp/tests/test_quat.py
@@ -1,9 +1,16 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -1957,102 +1964,100 @@ def make_quat(*args):
     test.assertSequenceEqual(wptype(24) / v, make_quat(12, 6, 4, 3))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestQuat(parent):
-        pass
 
-    add_kernel_test(TestQuat, test_constructor_default, dim=1, devices=devices)
+class TestQuat(unittest.TestCase):
+    pass
 
-    for dtype in np_float_types:
-        add_function_test_register_kernel(
-            TestQuat, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat,
-            f"test_casting_constructors_{dtype.__name__}",
-            test_casting_constructors,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_anon_type_instance_{dtype.__name__}", test_anon_type_instance, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_inverse_{dtype.__name__}", test_inverse, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_quat_identity_{dtype.__name__}", test_quat_identity, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_dotproduct_{dtype.__name__}", test_dotproduct, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_length_{dtype.__name__}", test_length, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_normalize_{dtype.__name__}", test_normalize, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_addition_{dtype.__name__}", test_addition, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_subtraction_{dtype.__name__}", test_subtraction, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat,
-            f"test_scalar_multiplication_{dtype.__name__}",
-            test_scalar_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_scalar_division_{dtype.__name__}", test_scalar_division, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat,
-            f"test_quat_multiplication_{dtype.__name__}",
-            test_quat_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_indexing_{dtype.__name__}", test_indexing, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_quat_lerp_{dtype.__name__}", test_quat_lerp, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat,
-            f"test_quat_to_axis_angle_grad_{dtype.__name__}",
-            test_quat_to_axis_angle_grad,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_slerp_grad_{dtype.__name__}", test_slerp_grad, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_quat_rpy_grad_{dtype.__name__}", test_quat_rpy_grad, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_quat_from_matrix_{dtype.__name__}", test_quat_from_matrix, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_quat_rotate_{dtype.__name__}", test_quat_rotate, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestQuat, f"test_quat_to_matrix_{dtype.__name__}", test_quat_to_matrix, devices=devices, dtype=dtype
-        )
-        add_function_test(
-            TestQuat, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
-        )
 
-    return TestQuat
+add_kernel_test(TestQuat, test_constructor_default, dim=1, devices=devices)
+
+for dtype in np_float_types:
+    add_function_test_register_kernel(
+        TestQuat, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat,
+        f"test_casting_constructors_{dtype.__name__}",
+        test_casting_constructors,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_anon_type_instance_{dtype.__name__}", test_anon_type_instance, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_inverse_{dtype.__name__}", test_inverse, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_quat_identity_{dtype.__name__}", test_quat_identity, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_dotproduct_{dtype.__name__}", test_dotproduct, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_length_{dtype.__name__}", test_length, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_normalize_{dtype.__name__}", test_normalize, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_addition_{dtype.__name__}", test_addition, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_subtraction_{dtype.__name__}", test_subtraction, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat,
+        f"test_scalar_multiplication_{dtype.__name__}",
+        test_scalar_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_scalar_division_{dtype.__name__}", test_scalar_division, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat,
+        f"test_quat_multiplication_{dtype.__name__}",
+        test_quat_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_indexing_{dtype.__name__}", test_indexing, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_quat_lerp_{dtype.__name__}", test_quat_lerp, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat,
+        f"test_quat_to_axis_angle_grad_{dtype.__name__}",
+        test_quat_to_axis_angle_grad,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_slerp_grad_{dtype.__name__}", test_slerp_grad, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_quat_rpy_grad_{dtype.__name__}", test_quat_rpy_grad, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_quat_from_matrix_{dtype.__name__}", test_quat_from_matrix, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_quat_rotate_{dtype.__name__}", test_quat_rotate, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestQuat, f"test_quat_to_matrix_{dtype.__name__}", test_quat_to_matrix, devices=devices, dtype=dtype
+    )
+    add_function_test(
+        TestQuat, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
+    )
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_rand.py b/warp/tests/test_rand.py
index 4fe21f58c..2f169329e 100644
--- a/warp/tests/test_rand.py
+++ b/warp/tests/test_rand.py
@@ -9,10 +9,11 @@
 
 import numpy as np
 
+import warp as wp
+from warp.tests.unittest_utils import *
+
 # import matplotlib.pyplot as plt
 
-import warp as wp
-from warp.tests.test_base import *
 
 wp.init()
 
@@ -62,7 +63,7 @@ def test_rand(test, device):
     wp.copy(int_ab_host, int_ab_device)
     wp.copy(float_01_host, float_01_device)
     wp.copy(float_ab_host, float_ab_device)
-    wp.synchronize()
+    wp.synchronize_device(device)
 
     int_a = int_a_host.numpy()
     int_ab = int_ab_host.numpy()
@@ -269,21 +270,19 @@ def test_poisson(test, device):
     test.assertTrue(np.abs(poisson_high_std - np_poisson_high_std) <= 2e-1)
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestNoise(parent):
-        pass
+class TestRand(unittest.TestCase):
+    pass
 
-    add_function_test(TestNoise, "test_rand", test_rand, devices=devices)
-    add_function_test(TestNoise, "test_sample_cdf", test_sample_cdf, devices=devices)
-    add_function_test(TestNoise, "test_sampling_methods", test_sampling_methods, devices=devices)
-    add_function_test(TestNoise, "test_poisson", test_poisson, devices=devices)
 
-    return TestNoise
+add_function_test(TestRand, "test_rand", test_rand, devices=devices)
+add_function_test(TestRand, "test_sample_cdf", test_sample_cdf, devices=devices)
+add_function_test(TestRand, "test_sampling_methods", test_sampling_methods, devices=devices)
+add_function_test(TestRand, "test_poisson", test_poisson, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_reload.py b/warp/tests/test_reload.py
index bb301e9d5..4694c1587 100644
--- a/warp/tests/test_reload.py
+++ b/warp/tests/test_reload.py
@@ -5,25 +5,22 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import numpy as np
-import warp as wp
+import importlib
+import os
+import unittest
 
-import math
+import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
 
-import unittest
-import importlib
-import os
+# dummy modules used for testing reload with dependencies
+import warp.tests.aux_test_dependent as test_dependent
+import warp.tests.aux_test_reference as test_reference
+import warp.tests.aux_test_reference_reference as test_reference_reference
 
 # dummy module used for testing reload
-import warp.tests.test_square as test_square
-
-# dummy modules used for testing reload with dependencies
-import warp.tests.test_dependent as test_dependent
-import warp.tests.test_reference as test_reference
-import warp.tests.test_reference_reference as test_reference_reference
+import warp.tests.aux_test_square as test_square
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -110,7 +107,7 @@ def run(expect, device):
 
 def test_reload(test, device):
     # write out the module python and import it
-    f = open(os.path.abspath(os.path.join(os.path.dirname(__file__), "test_square.py")), "w")
+    f = open(os.path.abspath(os.path.join(os.path.dirname(__file__), "aux_test_square.py")), "w")
     f.writelines(square_two)
     f.flush()
     f.close()
@@ -118,7 +115,7 @@ def test_reload(test, device):
     reload_module(test_square)
     test_square.run(expect=4.0, device=device)  # 2*2=4
 
-    f = open(os.path.abspath(os.path.join(os.path.dirname(__file__), "test_square.py")), "w")
+    f = open(os.path.abspath(os.path.join(os.path.dirname(__file__), "aux_test_square.py")), "w")
     f.writelines(square_four)
     f.flush()
     f.close()
@@ -130,12 +127,12 @@ def test_reload(test, device):
 
 def test_reload_class(test, device):
     def test_func():
-        import warp.tests.test_class_kernel
-        from warp.tests.test_class_kernel import ClassKernelTest
-
         import importlib as imp
 
-        imp.reload(warp.tests.test_class_kernel)
+        import warp.tests.aux_test_class_kernel
+        from warp.tests.aux_test_class_kernel import ClassKernelTest
+
+        imp.reload(warp.tests.aux_test_class_kernel)
 
         ctest = ClassKernelTest(device)
         expected = np.zeros((10, 3, 3), dtype=np.float32)
@@ -149,7 +146,7 @@ def test_func():
 template_ref = """# This file is used to test reloading module references.
 
 import warp as wp
-import warp.tests.test_reference_reference as refref
+import warp.tests.aux_test_reference_reference as refref
 
 wp.init()
 
@@ -173,8 +170,8 @@ def more_magic():
 
 
 def test_reload_references(test, device):
-    path_ref = os.path.abspath(os.path.join(os.path.dirname(__file__), "test_reference.py"))
-    path_refref = os.path.abspath(os.path.join(os.path.dirname(__file__), "test_reference_reference.py"))
+    path_ref = os.path.abspath(os.path.join(os.path.dirname(__file__), "aux_test_reference.py"))
+    path_refref = os.path.abspath(os.path.join(os.path.dirname(__file__), "aux_test_reference_reference.py"))
 
     # rewrite both dependency modules and reload them
     with open(path_ref, "w") as f:
@@ -202,21 +199,19 @@ def test_reload_references(test, device):
     test_dependent.run(expect=4.0, device=device)  # 2 * 2 = 4
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestReload(parent):
-        pass
+class TestReload(unittest.TestCase):
+    pass
 
-    add_function_test(TestReload, "test_redefine", test_redefine, devices=devices)
-    add_function_test(TestReload, "test_reload", test_reload, devices=devices)
-    add_function_test(TestReload, "test_reload_class", test_reload_class, devices=devices)
-    add_function_test(TestReload, "test_reload_references", test_reload_references, devices=devices)
 
-    return TestReload
+add_function_test(TestReload, "test_redefine", test_redefine, devices=devices)
+add_function_test(TestReload, "test_reload", test_reload, devices=devices)
+add_function_test(TestReload, "test_reload_class", test_reload_class, devices=devices)
+add_function_test(TestReload, "test_reload_references", test_reload_references, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=False)
diff --git a/warp/tests/test_rounding.py b/warp/tests/test_rounding.py
index 31c9b4671..999a59056 100644
--- a/warp/tests/test_rounding.py
+++ b/warp/tests/test_rounding.py
@@ -7,11 +7,11 @@
 
 import unittest
 
-import warp as wp
-from warp.tests.test_base import *
-
 import numpy as np
 
+import warp as wp
+from warp.tests.unittest_utils import *
+
 compare_to_numpy = False
 print_results = False
 
@@ -165,18 +165,15 @@ def test_rounding(test, device):
         print("----------------------------------------------")
 
 
-def register(parent):
-    class TestRounding(parent):
-        pass
+class TestRounding(unittest.TestCase):
+    pass
 
-    devices = get_test_devices()
 
-    add_function_test(TestRounding, "test_rounding", test_rounding, devices=devices)
+devices = get_test_devices()
 
-    return TestRounding
+add_function_test(TestRounding, "test_rounding", test_rounding, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_runlength_encode.py b/warp/tests/test_runlength_encode.py
index a7f29fe09..49b7e3946 100644
--- a/warp/tests/test_runlength_encode.py
+++ b/warp/tests/test_runlength_encode.py
@@ -1,11 +1,18 @@
-from functools import partial
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
+from functools import partial
 
 import numpy as np
-import warp as wp
 
+import warp as wp
+from warp.tests.unittest_utils import *
 from warp.utils import runlength_encode
-from warp.tests.test_base import *
 
 wp.init()
 
@@ -29,39 +36,10 @@ def test_runlength_encode_int(test, device, n):
     assert_np_equal(unique_counts.numpy()[:run_count], unique_counts_np[:run_count])
 
 
-def test_runlength_encode_error_devices_mismatch(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty_like(values, device="cuda:0")
-    run_lengths = wp.empty_like(values, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        runlength_encode(values, run_values, run_lengths)
-
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty_like(values, device="cpu")
-    run_lengths = wp.empty_like(values, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        runlength_encode(values, run_values, run_lengths)
-
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty_like(values, device="cuda:0")
-    run_lengths = wp.empty_like(values, device="cpu")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        runlength_encode(values, run_values, run_lengths)
-
-
 def test_runlength_encode_error_insufficient_storage(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty(1, dtype=int, device="cpu")
-    run_lengths = wp.empty(123, dtype=int, device="cpu")
+    values = wp.zeros(123, dtype=int, device=device)
+    run_values = wp.empty(1, dtype=int, device=device)
+    run_lengths = wp.empty(123, dtype=int, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Output array storage sizes must be at least equal to value_count$",
@@ -79,9 +57,9 @@ def test_runlength_encode_error_insufficient_storage(test, device):
 
 
 def test_runlength_encode_error_dtypes_mismatch(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty(123, dtype=float, device="cpu")
-    run_lengths = wp.empty_like(values)
+    values = wp.zeros(123, dtype=int, device=device)
+    run_values = wp.empty(123, dtype=float, device=device)
+    run_lengths = wp.empty_like(values, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"values and run_values data types do not match$",
@@ -90,9 +68,9 @@ def test_runlength_encode_error_dtypes_mismatch(test, device):
 
 
 def test_runlength_encode_error_run_length_unsupported_dtype(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty(123, dtype=int, device="cpu")
-    run_lengths = wp.empty(123, dtype=float, device="cpu")
+    values = wp.zeros(123, dtype=int, device=device)
+    run_values = wp.empty(123, dtype=int, device=device)
+    run_lengths = wp.empty(123, dtype=float, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"run_lengths array must be of type int32$",
@@ -100,23 +78,11 @@ def test_runlength_encode_error_run_length_unsupported_dtype(test, device):
         runlength_encode(values, run_values, run_lengths)
 
 
-def test_runlength_encode_error_run_count_device_mismatch(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty_like(values, device="cpu")
-    run_lengths = wp.empty_like(values, device="cpu")
-    run_count = wp.empty(shape=(1,), dtype=int, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"run_count storage device does not match other arrays$",
-    ):
-        runlength_encode(values, run_values, run_lengths, run_count=run_count)
-
-
 def test_runlength_encode_error_run_count_unsupported_dtype(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    run_values = wp.empty_like(values, device="cpu")
-    run_lengths = wp.empty_like(values, device="cpu")
-    run_count = wp.empty(shape=(1,), dtype=float, device="cpu")
+    values = wp.zeros(123, dtype=int, device=device)
+    run_values = wp.empty_like(values, device=device)
+    run_lengths = wp.empty_like(values, device=device)
+    run_count = wp.empty(shape=(1,), dtype=float, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"run_count array must be of type int32$",
@@ -135,57 +101,90 @@ def test_runlength_encode_error_unsupported_dtype(test, device):
         runlength_encode(values, run_values, run_lengths)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestRunlengthEncode(parent):
-        pass
-
-    add_function_test(
-        TestRunlengthEncode, "test_runlength_encode_int", partial(test_runlength_encode_int, n=100), devices=devices
-    )
-    add_function_test(
-        TestRunlengthEncode, "test_runlength_encode_empty", partial(test_runlength_encode_int, n=0), devices=devices
-    )
-    add_function_test(
-        TestRunlengthEncode,
-        "test_runlength_encode_error_devices_mismatch",
-        test_runlength_encode_error_devices_mismatch,
-    )
-    add_function_test(
-        TestRunlengthEncode,
-        "test_runlength_encode_error_insufficient_storage",
-        test_runlength_encode_error_insufficient_storage,
-    )
-    add_function_test(
-        TestRunlengthEncode, "test_runlength_encode_error_dtypes_mismatch", test_runlength_encode_error_dtypes_mismatch
-    )
-    add_function_test(
-        TestRunlengthEncode,
-        "test_runlength_encode_error_run_length_unsupported_dtype",
-        test_runlength_encode_error_run_length_unsupported_dtype,
-    )
-    add_function_test(
-        TestRunlengthEncode,
-        "test_runlength_encode_error_run_count_device_mismatch",
-        test_runlength_encode_error_run_count_device_mismatch,
-    )
-    add_function_test(
-        TestRunlengthEncode,
-        "test_runlength_encode_error_run_count_unsupported_dtype",
-        test_runlength_encode_error_run_count_unsupported_dtype,
-    )
-    add_function_test(
-        TestRunlengthEncode,
-        "test_runlength_encode_error_unsupported_dtype",
-        test_runlength_encode_error_unsupported_dtype,
-        devices=devices,
-    )
-
-    return TestRunlengthEncode
+devices = get_test_devices()
+
+
+class TestRunlengthEncode(unittest.TestCase):
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_runlength_encode_error_devices_mismatch(self):
+        values = wp.zeros(123, dtype=int, device="cpu")
+        run_values = wp.empty_like(values, device="cuda:0")
+        run_lengths = wp.empty_like(values, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            runlength_encode(values, run_values, run_lengths)
+
+        values = wp.zeros(123, dtype=int, device="cpu")
+        run_values = wp.empty_like(values, device="cpu")
+        run_lengths = wp.empty_like(values, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            runlength_encode(values, run_values, run_lengths)
+
+        values = wp.zeros(123, dtype=int, device="cpu")
+        run_values = wp.empty_like(values, device="cuda:0")
+        run_lengths = wp.empty_like(values, device="cpu")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            runlength_encode(values, run_values, run_lengths)
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_runlength_encode_error_run_count_device_mismatch(self):
+        values = wp.zeros(123, dtype=int, device="cpu")
+        run_values = wp.empty_like(values, device="cpu")
+        run_lengths = wp.empty_like(values, device="cpu")
+        run_count = wp.empty(shape=(1,), dtype=int, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"run_count storage device does not match other arrays$",
+        ):
+            runlength_encode(values, run_values, run_lengths, run_count=run_count)
+
+
+add_function_test(
+    TestRunlengthEncode, "test_runlength_encode_int", partial(test_runlength_encode_int, n=100), devices=devices
+)
+add_function_test(
+    TestRunlengthEncode, "test_runlength_encode_empty", partial(test_runlength_encode_int, n=0), devices=devices
+)
+add_function_test(
+    TestRunlengthEncode,
+    "test_runlength_encode_error_insufficient_storage",
+    test_runlength_encode_error_insufficient_storage,
+    devices=devices,
+)
+add_function_test(
+    TestRunlengthEncode,
+    "test_runlength_encode_error_dtypes_mismatch",
+    test_runlength_encode_error_dtypes_mismatch,
+    devices=devices,
+)
+add_function_test(
+    TestRunlengthEncode,
+    "test_runlength_encode_error_run_length_unsupported_dtype",
+    test_runlength_encode_error_run_length_unsupported_dtype,
+    devices=devices,
+)
+add_function_test(
+    TestRunlengthEncode,
+    "test_runlength_encode_error_run_count_unsupported_dtype",
+    test_runlength_encode_error_run_count_unsupported_dtype,
+    devices=devices,
+)
+add_function_test(
+    TestRunlengthEncode,
+    "test_runlength_encode_error_unsupported_dtype",
+    test_runlength_encode_error_unsupported_dtype,
+    devices=devices,
+)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_smoothstep.py b/warp/tests/test_smoothstep.py
index 2f1070716..d12b9718f 100644
--- a/warp/tests/test_smoothstep.py
+++ b/warp/tests/test_smoothstep.py
@@ -5,14 +5,14 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
+import unittest
 from dataclasses import dataclass
 from typing import Any
-import unittest
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 
 @dataclass
@@ -153,17 +153,16 @@ def fn(
                 )
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
+
+class TestSmoothstep(unittest.TestCase):
+    pass
 
-    class TestSmoothstep(parent):
-        pass
 
-    add_function_test(TestSmoothstep, "test_smoothstep", test_smoothstep, devices=devices)
-    return TestSmoothstep
+add_function_test(TestSmoothstep, "test_smoothstep", test_smoothstep, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_snippet.py b/warp/tests/test_snippet.py
index d1205a6ef..299150a9f 100644
--- a/warp/tests/test_snippet.py
+++ b/warp/tests/test_snippet.py
@@ -1,8 +1,10 @@
-import warp as wp
-from warp.tests.test_base import *
-import numpy as np
 import unittest
 
+import numpy as np
+
+import warp as wp
+from warp.tests.unittest_utils import *
+
 wp.init()
 
 
@@ -101,11 +103,11 @@ def reverse_kernel(d: wp.array(dtype=int), N: int):
 
 
 def test_cpu_snippet(test, device):
-    
     snippet = """
     int inc = 1;
     out[tid] = x[tid] + inc;
     """
+
     @wp.func_native(snippet)
     def increment_snippet(
         x: wp.array(dtype=wp.int32),
@@ -115,10 +117,7 @@ def increment_snippet(
         ...
 
     @wp.kernel
-    def increment(
-        x: wp.array(dtype=wp.int32),
-        out: wp.array(dtype=wp.int32)
-    ):
+    def increment(x: wp.array(dtype=wp.int32), out: wp.array(dtype=wp.int32)):
         tid = wp.tid()
         increment_snippet(x, out, tid)
 
@@ -128,25 +127,17 @@ def increment(
 
     wp.launch(kernel=increment, dim=N, inputs=[x], outputs=[out], device=device)
 
-    assert_np_equal(out.numpy(), np.arange(1, N+1, 1, dtype=np.int32))
+    assert_np_equal(out.numpy(), np.arange(1, N + 1, 1, dtype=np.int32))
 
 
-def register(parent):
-    
-    class TestSnippets(parent):
-        pass
+class TestSnippets(unittest.TestCase):
+    pass
 
-    if wp.is_cuda_available():
-        cuda_device = [wp.get_cuda_device()]
-        add_function_test(TestSnippets, "test_basic", test_basic, devices=cuda_device)
-        add_function_test(TestSnippets, "test_shared_memory", test_shared_memory, devices=cuda_device)
-    
-    if wp.is_cpu_available():
-        add_function_test(TestSnippets, "test_cpu_snippet", test_cpu_snippet, devices=["cpu"])
 
-    return TestSnippets
+add_function_test(TestSnippets, "test_basic", test_basic, devices=get_unique_cuda_test_devices())
+add_function_test(TestSnippets, "test_shared_memory", test_shared_memory, devices=get_unique_cuda_test_devices())
+add_function_test(TestSnippets, "test_cpu_snippet", test_cpu_snippet, devices=["cpu"])
 
 
 if __name__ == "__main__":
-    c = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_sparse.py b/warp/tests/test_sparse.py
index 2b6a325ec..6cbf8f21b 100644
--- a/warp/tests/test_sparse.py
+++ b/warp/tests/test_sparse.py
@@ -1,13 +1,20 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
 import numpy as np
-import warp as wp
-import unittest
 
+import warp as wp
 from warp.sparse import bsr_zeros, bsr_set_from_triplets, bsr_get_diag, bsr_diag, bsr_identity, bsr_copy, bsr_scale
 from warp.sparse import bsr_set_transpose, bsr_transposed
 from warp.sparse import bsr_axpy, bsr_mm, bsr_axpy_work_arrays, bsr_mm_work_arrays, bsr_mv
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
+
 
 wp.init()
 
@@ -419,41 +426,35 @@ def test_bsr_mv(test, device):
     return test_bsr_mv
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestSparse(parent):
-        pass
 
-    add_function_test(TestSparse, "test_csr_from_triplets", test_csr_from_triplets, devices=devices)
-    add_function_test(TestSparse, "test_bsr_from_triplets", test_bsr_from_triplets, devices=devices)
-    add_function_test(TestSparse, "test_bsr_get_diag", test_bsr_get_set_diag, devices=devices)
-    add_function_test(TestSparse, "test_bsr_copy_scale", test_bsr_copy_scale, devices=devices)
+class TestSparse(unittest.TestCase):
+    pass
 
-    add_function_test(TestSparse, "test_csr_transpose", make_test_bsr_transpose((1, 1), wp.float32), devices=devices)
-    add_function_test(
-        TestSparse, "test_bsr_transpose_1_3", make_test_bsr_transpose((1, 3), wp.float32), devices=devices
-    )
-    add_function_test(
-        TestSparse, "test_bsr_transpose_3_3", make_test_bsr_transpose((3, 3), wp.float64), devices=devices
-    )
 
-    add_function_test(TestSparse, "test_csr_axpy", make_test_bsr_axpy((1, 1), wp.float32), devices=devices)
-    add_function_test(TestSparse, "test_bsr_axpy_1_3", make_test_bsr_axpy((1, 3), wp.float32), devices=devices)
-    add_function_test(TestSparse, "test_bsr_axpy_3_3", make_test_bsr_axpy((3, 3), wp.float64), devices=devices)
+add_function_test(TestSparse, "test_csr_from_triplets", test_csr_from_triplets, devices=devices)
+add_function_test(TestSparse, "test_bsr_from_triplets", test_bsr_from_triplets, devices=devices)
+add_function_test(TestSparse, "test_bsr_get_diag", test_bsr_get_set_diag, devices=devices)
+add_function_test(TestSparse, "test_bsr_copy_scale", test_bsr_copy_scale, devices=devices)
+
+add_function_test(TestSparse, "test_csr_transpose", make_test_bsr_transpose((1, 1), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_transpose_1_3", make_test_bsr_transpose((1, 3), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_transpose_3_3", make_test_bsr_transpose((3, 3), wp.float64), devices=devices)
 
-    add_function_test(TestSparse, "test_csr_mm", make_test_bsr_mm((1, 1), wp.float32), devices=devices)
-    add_function_test(TestSparse, "test_bsr_mm_1_3", make_test_bsr_mm((1, 3), wp.float32), devices=devices)
-    add_function_test(TestSparse, "test_bsr_mm_3_3", make_test_bsr_mm((3, 3), wp.float64), devices=devices)
+add_function_test(TestSparse, "test_csr_axpy", make_test_bsr_axpy((1, 1), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_axpy_1_3", make_test_bsr_axpy((1, 3), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_axpy_3_3", make_test_bsr_axpy((3, 3), wp.float64), devices=devices)
 
-    add_function_test(TestSparse, "test_csr_mv", make_test_bsr_mv((1, 1), wp.float32), devices=devices)
-    add_function_test(TestSparse, "test_bsr_mv_1_3", make_test_bsr_mv((1, 3), wp.float32), devices=devices)
-    add_function_test(TestSparse, "test_bsr_mv_3_3", make_test_bsr_mv((3, 3), wp.float64), devices=devices)
+add_function_test(TestSparse, "test_csr_mm", make_test_bsr_mm((1, 1), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_mm_1_3", make_test_bsr_mm((1, 3), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_mm_3_3", make_test_bsr_mm((3, 3), wp.float64), devices=devices)
 
-    return TestSparse
+add_function_test(TestSparse, "test_csr_mv", make_test_bsr_mv((1, 1), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_mv_1_3", make_test_bsr_mv((1, 3), wp.float32), devices=devices)
+add_function_test(TestSparse, "test_bsr_mv_3_3", make_test_bsr_mv((3, 3), wp.float64), devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_spatial.py b/warp/tests/test_spatial.py
index 9be8a03c1..a557a8d58 100644
--- a/warp/tests/test_spatial.py
+++ b/warp/tests/test_spatial.py
@@ -1,9 +1,16 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import unittest
 
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -1956,188 +1963,186 @@ def transform_create_test(input: wp.array(dtype=wptype), output: wp.array(dtype=
         tape.zero()
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestSpatial(parent):
-        pass
 
-    for dtype in np_float_types:
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_vector_constructors_{dtype.__name__}",
-            test_spatial_vector_constructors,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_vector_indexing_{dtype.__name__}",
-            test_spatial_vector_indexing,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_vector_scalar_multiplication_{dtype.__name__}",
-            test_spatial_vector_scalar_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_vector_add_sub_{dtype.__name__}",
-            test_spatial_vector_add_sub,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial, f"test_spatial_dot_{dtype.__name__}", test_spatial_dot, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestSpatial, f"test_spatial_cross_{dtype.__name__}", test_spatial_cross, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_top_bottom_{dtype.__name__}",
-            test_spatial_top_bottom,
-            devices=devices,
-            dtype=dtype,
-        )
+class TestSpatial(unittest.TestCase):
+    pass
 
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_constructors_{dtype.__name__}",
-            test_transform_constructors,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_anon_type_instance_{dtype.__name__}",
-            test_transform_anon_type_instance,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_identity_{dtype.__name__}",
-            test_transform_identity,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_indexing_{dtype.__name__}",
-            test_transform_indexing,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_get_trans_rot_{dtype.__name__}",
-            test_transform_get_trans_rot,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_multiply_{dtype.__name__}",
-            test_transform_multiply,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_inverse_{dtype.__name__}",
-            test_transform_inverse,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_point_vector_{dtype.__name__}",
-            test_transform_point_vector,
-            devices=devices,
-            dtype=dtype,
-        )
 
-        # are these two valid? They don't seem to be doing things you'd want to do,
-        # maybe they should be removed
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_scalar_multiplication_{dtype.__name__}",
-            test_transform_scalar_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_transform_add_sub_{dtype.__name__}",
-            test_transform_add_sub,
-            devices=devices,
-            dtype=dtype,
-        )
+for dtype in np_float_types:
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_vector_constructors_{dtype.__name__}",
+        test_spatial_vector_constructors,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_vector_indexing_{dtype.__name__}",
+        test_spatial_vector_indexing,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_vector_scalar_multiplication_{dtype.__name__}",
+        test_spatial_vector_scalar_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_vector_add_sub_{dtype.__name__}",
+        test_spatial_vector_add_sub,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial, f"test_spatial_dot_{dtype.__name__}", test_spatial_dot, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestSpatial, f"test_spatial_cross_{dtype.__name__}", test_spatial_cross, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_top_bottom_{dtype.__name__}",
+        test_spatial_top_bottom,
+        devices=devices,
+        dtype=dtype,
+    )
 
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_matrix_constructors_{dtype.__name__}",
-            test_spatial_matrix_constructors,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_matrix_indexing_{dtype.__name__}",
-            test_spatial_matrix_indexing,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_matrix_scalar_multiplication_{dtype.__name__}",
-            test_spatial_matrix_scalar_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_matrix_add_sub_{dtype.__name__}",
-            test_spatial_matrix_add_sub,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_matvec_multiplication_{dtype.__name__}",
-            test_spatial_matvec_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_matmat_multiplication_{dtype.__name__}",
-            test_spatial_matmat_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial,
-            f"test_spatial_outer_product_{dtype.__name__}",
-            test_spatial_outer_product,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestSpatial, f"test_spatial_adjoint_{dtype.__name__}", test_spatial_adjoint, devices=devices, dtype=dtype
-        )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_constructors_{dtype.__name__}",
+        test_transform_constructors,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_anon_type_instance_{dtype.__name__}",
+        test_transform_anon_type_instance,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_identity_{dtype.__name__}",
+        test_transform_identity,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_indexing_{dtype.__name__}",
+        test_transform_indexing,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_get_trans_rot_{dtype.__name__}",
+        test_transform_get_trans_rot,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_multiply_{dtype.__name__}",
+        test_transform_multiply,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_inverse_{dtype.__name__}",
+        test_transform_inverse,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_point_vector_{dtype.__name__}",
+        test_transform_point_vector,
+        devices=devices,
+        dtype=dtype,
+    )
 
-        # \TODO: test spatial_mass and spatial_jacobian
+    # are these two valid? They don't seem to be doing things you'd want to do,
+    # maybe they should be removed
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_scalar_multiplication_{dtype.__name__}",
+        test_transform_scalar_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_transform_add_sub_{dtype.__name__}",
+        test_transform_add_sub,
+        devices=devices,
+        dtype=dtype,
+    )
+
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_matrix_constructors_{dtype.__name__}",
+        test_spatial_matrix_constructors,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_matrix_indexing_{dtype.__name__}",
+        test_spatial_matrix_indexing,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_matrix_scalar_multiplication_{dtype.__name__}",
+        test_spatial_matrix_scalar_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_matrix_add_sub_{dtype.__name__}",
+        test_spatial_matrix_add_sub,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_matvec_multiplication_{dtype.__name__}",
+        test_spatial_matvec_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_matmat_multiplication_{dtype.__name__}",
+        test_spatial_matmat_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial,
+        f"test_spatial_outer_product_{dtype.__name__}",
+        test_spatial_outer_product,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestSpatial, f"test_spatial_adjoint_{dtype.__name__}", test_spatial_adjoint, devices=devices, dtype=dtype
+    )
 
-    return TestSpatial
+    # \TODO: test spatial_mass and spatial_jacobian
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_streams.py b/warp/tests/test_streams.py
index 2f76861f2..5cb0f5082 100644
--- a/warp/tests/test_streams.py
+++ b/warp/tests/test_streams.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -284,145 +284,138 @@ def test_stream_scope_wait_stream(test, device):
         assert_np_equal(d.numpy(), np.full(N, fill_value=4.0))
 
 
-def test_stream_exceptions(test, device):
-    cpu_device = wp.get_device("cpu")
+devices = get_unique_cuda_test_devices()
 
-    # Can't set the stream on a CPU device
-    with test.assertRaises(RuntimeError):
-        stream0 = wp.Stream()
-        cpu_device.stream = stream0
 
-    # Can't create a stream on the CPU
-    with test.assertRaises(RuntimeError):
-        wp.Stream(device="cpu")
+class TestStreams(unittest.TestCase):
+    def test_stream_exceptions(self):
+        cpu_device = wp.get_device("cpu")
 
-    # Can't create an event with CPU device
-    with test.assertRaises(RuntimeError):
-        wp.Event(device=cpu_device)
+        # Can't set the stream on a CPU device
+        with self.assertRaises(RuntimeError):
+            stream0 = wp.Stream()
+            cpu_device.stream = stream0
 
-    # Can't get the stream on a CPU device
-    with test.assertRaises(RuntimeError):
-        cpu_stream = cpu_device.stream  # noqa: F841
+        # Can't create a stream on the CPU
+        with self.assertRaises(RuntimeError):
+            wp.Stream(device="cpu")
 
+        # Can't create an event with CPU device
+        with self.assertRaises(RuntimeError):
+            wp.Event(device=cpu_device)
 
-def test_stream_arg_graph_mgpu(test, device):
-    # resources on GPU 0
-    stream0 = wp.get_stream("cuda:0")
-    a0 = wp.zeros(N, dtype=float, device="cuda:0")
-    b0 = wp.empty(N, dtype=float, device="cuda:0")
-    c0 = wp.empty(N, dtype=float, device="cuda:0")
+        # Can't get the stream on a CPU device
+        with self.assertRaises(RuntimeError):
+            cpu_stream = cpu_device.stream  # noqa: F841
 
-    # resources on GPU 1
-    stream1 = wp.get_stream("cuda:1")
-    a1 = wp.zeros(N, dtype=float, device="cuda:1")
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_stream_arg_graph_mgpu(self):
+        wp.load_module(device="cuda:0")
+        wp.load_module(device="cuda:1")
 
-    # start recording on stream0
-    wp.capture_begin(stream=stream0)
+        # resources on GPU 0
+        stream0 = wp.get_stream("cuda:0")
+        a0 = wp.zeros(N, dtype=float, device="cuda:0")
+        b0 = wp.empty(N, dtype=float, device="cuda:0")
+        c0 = wp.empty(N, dtype=float, device="cuda:0")
 
-    # branch into stream1
-    stream1.wait_stream(stream0)
+        # resources on GPU 1
+        stream1 = wp.get_stream("cuda:1")
+        a1 = wp.zeros(N, dtype=float, device="cuda:1")
 
-    # launch concurrent kernels on each stream
-    wp.launch(inc, dim=N, inputs=[a0], stream=stream0)
-    wp.launch(inc, dim=N, inputs=[a1], stream=stream1)
-
-    # wait for stream1 to finish
-    stream0.wait_stream(stream1)
-
-    # copy values from stream1
-    wp.copy(b0, a1, stream=stream0)
+        # start recording on stream0
+        wp.capture_begin(stream=stream0, force_module_load=False)
+        try:
+            # branch into stream1
+            stream1.wait_stream(stream0)
 
-    # compute sum
-    wp.launch(sum, dim=N, inputs=[a0, b0, c0], stream=stream0)
+            # launch concurrent kernels on each stream
+            wp.launch(inc, dim=N, inputs=[a0], stream=stream0)
+            wp.launch(inc, dim=N, inputs=[a1], stream=stream1)
 
-    # finish recording on stream0
-    g = wp.capture_end(stream=stream0)
+            # wait for stream1 to finish
+            stream0.wait_stream(stream1)
 
-    # replay
-    num_iters = 10
-    for _ in range(num_iters):
-        wp.capture_launch(g, stream=stream0)
+            # copy values from stream1
+            wp.copy(b0, a1, stream=stream0)
 
-    # check results
-    assert_np_equal(c0.numpy(), np.full(N, fill_value=2 * num_iters))
+            # compute sum
+            wp.launch(sum, dim=N, inputs=[a0, b0, c0], stream=stream0)
+        finally:
+            # finish recording on stream0
+            g = wp.capture_end(stream=stream0)
 
+        # replay
+        num_iters = 10
+        for _ in range(num_iters):
+            wp.capture_launch(g, stream=stream0)
 
-def test_stream_scope_graph_mgpu(test, device):
-    # resources on GPU 0
-    with wp.ScopedDevice("cuda:0"):
-        stream0 = wp.get_stream()
-        a0 = wp.zeros(N, dtype=float)
-        b0 = wp.empty(N, dtype=float)
-        c0 = wp.empty(N, dtype=float)
+        # check results
+        assert_np_equal(c0.numpy(), np.full(N, fill_value=2 * num_iters))
 
-    # resources on GPU 1
-    with wp.ScopedDevice("cuda:1"):
-        stream1 = wp.get_stream()
-        a1 = wp.zeros(N, dtype=float)
+    @unittest.skipUnless(len(wp.get_cuda_devices()) > 1, "Requires at least two CUDA devices")
+    def test_stream_scope_graph_mgpu(self):
+        wp.load_module(device="cuda:0")
+        wp.load_module(device="cuda:1")
 
-    # capture graph
-    with wp.ScopedDevice("cuda:0"):
-        # start recording
-        wp.capture_begin()
+        # resources on GPU 0
+        with wp.ScopedDevice("cuda:0"):
+            stream0 = wp.get_stream()
+            a0 = wp.zeros(N, dtype=float)
+            b0 = wp.empty(N, dtype=float)
+            c0 = wp.empty(N, dtype=float)
 
+        # resources on GPU 1
         with wp.ScopedDevice("cuda:1"):
-            # branch into stream1
-            wp.wait_stream(stream0)
+            stream1 = wp.get_stream()
+            a1 = wp.zeros(N, dtype=float)
 
-            wp.launch(inc, dim=N, inputs=[a1])
+        # capture graph
+        with wp.ScopedDevice("cuda:0"):
+            # start recording
+            wp.capture_begin(force_module_load=False)
+            try:
+                with wp.ScopedDevice("cuda:1"):
+                    # branch into stream1
+                    wp.wait_stream(stream0)
 
-        wp.launch(inc, dim=N, inputs=[a0])
+                    wp.launch(inc, dim=N, inputs=[a1])
 
-        # wait for stream1 to finish
-        wp.wait_stream(stream1)
+                wp.launch(inc, dim=N, inputs=[a0])
 
-        # copy values from stream1
-        wp.copy(b0, a1)
+                # wait for stream1 to finish
+                wp.wait_stream(stream1)
 
-        # compute sum
-        wp.launch(sum, dim=N, inputs=[a0, b0, c0])
+                # copy values from stream1
+                wp.copy(b0, a1)
 
-        # finish recording
-        g = wp.capture_end()
+                # compute sum
+                wp.launch(sum, dim=N, inputs=[a0, b0, c0])
+            finally:
+                # finish recording
+                g = wp.capture_end()
 
-    # replay
-    with wp.ScopedDevice("cuda:0"):
-        num_iters = 10
-        for _ in range(num_iters):
-            wp.capture_launch(g)
-
-    # check results
-    assert_np_equal(c0.numpy(), np.full(N, fill_value=2 * num_iters))
-
-
-def register(parent):
-    devices = wp.get_cuda_devices()
+        # replay
+        with wp.ScopedDevice("cuda:0"):
+            num_iters = 10
+            for _ in range(num_iters):
+                wp.capture_launch(g)
 
-    class TestStreams(parent):
-        pass
+        # check results
+        assert_np_equal(c0.numpy(), np.full(N, fill_value=2 * num_iters))
 
-    add_function_test(TestStreams, "test_stream_arg_implicit_sync", test_stream_arg_implicit_sync, devices=devices)
-    add_function_test(TestStreams, "test_stream_scope_implicit_sync", test_stream_scope_implicit_sync, devices=devices)
 
-    add_function_test(TestStreams, "test_stream_arg_synchronize", test_stream_arg_synchronize, devices=devices)
-    add_function_test(TestStreams, "test_stream_arg_wait_event", test_stream_arg_wait_event, devices=devices)
-    add_function_test(TestStreams, "test_stream_arg_wait_stream", test_stream_arg_wait_stream, devices=devices)
-    add_function_test(TestStreams, "test_stream_scope_synchronize", test_stream_scope_synchronize, devices=devices)
-    add_function_test(TestStreams, "test_stream_scope_wait_event", test_stream_scope_wait_event, devices=devices)
-    add_function_test(TestStreams, "test_stream_scope_wait_stream", test_stream_scope_wait_stream, devices=devices)
-    add_function_test(TestStreams, "test_stream_exceptions", test_stream_exceptions, devices=devices)
+add_function_test(TestStreams, "test_stream_arg_implicit_sync", test_stream_arg_implicit_sync, devices=devices)
+add_function_test(TestStreams, "test_stream_scope_implicit_sync", test_stream_scope_implicit_sync, devices=devices)
 
-    if len(devices) > 1:
-        add_function_test(TestStreams, "test_stream_arg_graph_mgpu", test_stream_arg_graph_mgpu)
-        add_function_test(TestStreams, "test_stream_scope_graph_mgpu", test_stream_scope_graph_mgpu)
-
-    return TestStreams
+add_function_test(TestStreams, "test_stream_arg_synchronize", test_stream_arg_synchronize, devices=devices)
+add_function_test(TestStreams, "test_stream_arg_wait_event", test_stream_arg_wait_event, devices=devices)
+add_function_test(TestStreams, "test_stream_arg_wait_stream", test_stream_arg_wait_stream, devices=devices)
+add_function_test(TestStreams, "test_stream_scope_synchronize", test_stream_scope_synchronize, devices=devices)
+add_function_test(TestStreams, "test_stream_scope_wait_event", test_stream_scope_wait_event, devices=devices)
+add_function_test(TestStreams, "test_stream_scope_wait_stream", test_stream_scope_wait_stream, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
-
-    wp.force_load(wp.get_cuda_devices())
-
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_struct.py b/warp/tests/test_struct.py
index e41c7a677..9b26b628f 100644
--- a/warp/tests/test_struct.py
+++ b/warp/tests/test_struct.py
@@ -5,15 +5,16 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
+import unittest
 from typing import Any
+
 import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 from warp.fem import Sample as StructFromAnotherModule
 
-import unittest
-
 wp.init()
 
 
@@ -291,7 +292,7 @@ def test_struct_math_conversions(test, device):
     s.m5 = [10, 20, 30, 40]
     s.m6 = np.array([100, 200, 300, 400])
 
-    wp.launch(check_math_conversions, dim=1, inputs=[s])
+    wp.launch(check_math_conversions, dim=1, inputs=[s], device=device)
 
 
 @wp.struct
@@ -416,9 +417,9 @@ def test_nested_array_struct(test, device):
     var2.i = 2
 
     struct = ArrayStruct()
-    struct.array = wp.array([var1, var2], dtype=InnerStruct)
+    struct.array = wp.array([var1, var2], dtype=InnerStruct, device=device)
 
-    wp.launch(struct2_reader, dim=2, inputs=[struct])
+    wp.launch(struct2_reader, dim=2, inputs=[struct], device=device)
 
 
 @wp.struct
@@ -564,81 +565,111 @@ def test_dependent_module_import(c: DependentModuleImport_C):
     wp.tid()  # nop, we're just testing codegen
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestStruct(parent):
-        pass
 
-    add_function_test(TestStruct, "test_step", test_step, devices=devices)
-    add_function_test(TestStruct, "test_step_grad", test_step_grad, devices=devices)
-    add_kernel_test(TestStruct, kernel=test_empty, name="test_empty", dim=1, inputs=[Empty()], devices=devices)
-    add_kernel_test(
-        TestStruct,
-        kernel=test_uninitialized,
-        name="test_uninitialized",
-        dim=1,
-        inputs=[Uninitialized()],
-        devices=devices,
-    )
-    add_kernel_test(TestStruct, kernel=test_return, name="test_return", dim=1, inputs=[], devices=devices)
-    add_function_test(TestStruct, "test_struct_attribute_error", test_struct_attribute_error, devices=devices)
-    add_function_test(TestStruct, "test_nested_struct", test_nested_struct, devices=devices)
-    add_function_test(TestStruct, "test_nested_array_struct", test_nested_array_struct, devices=devices)
-    add_function_test(TestStruct, "test_nested_empty_struct", test_nested_empty_struct, devices=devices)
-    add_function_test(TestStruct, "test_struct_math_conversions", test_struct_math_conversions, devices=devices)
-    add_function_test(
-        TestStruct, "test_struct_default_attributes_python", test_struct_default_attributes_python, devices=devices
-    )
-    add_kernel_test(
-        TestStruct,
-        name="test_struct_default_attributes",
-        kernel=test_struct_default_attributes_kernel,
-        dim=1,
-        inputs=[],
-        devices=devices,
-    )
+class TestStruct(unittest.TestCase):
+    pass
+
 
+add_function_test(TestStruct, "test_step", test_step, devices=devices)
+add_function_test(TestStruct, "test_step_grad", test_step_grad, devices=devices)
+add_kernel_test(TestStruct, kernel=test_empty, name="test_empty", dim=1, inputs=[Empty()], devices=devices)
+add_kernel_test(
+    TestStruct,
+    kernel=test_uninitialized,
+    name="test_uninitialized",
+    dim=1,
+    inputs=[Uninitialized()],
+    devices=devices,
+)
+add_kernel_test(TestStruct, kernel=test_return, name="test_return", dim=1, inputs=[], devices=devices)
+add_function_test(TestStruct, "test_nested_struct", test_nested_struct, devices=devices)
+add_function_test(TestStruct, "test_nested_array_struct", test_nested_array_struct, devices=devices)
+add_function_test(TestStruct, "test_nested_empty_struct", test_nested_empty_struct, devices=devices)
+add_function_test(TestStruct, "test_struct_math_conversions", test_struct_math_conversions, devices=devices)
+add_function_test(
+    TestStruct, "test_struct_default_attributes_python", test_struct_default_attributes_python, devices=devices
+)
+add_kernel_test(
+    TestStruct,
+    name="test_struct_default_attributes",
+    kernel=test_struct_default_attributes_kernel,
+    dim=1,
+    inputs=[],
+    devices=devices,
+)
+
+add_kernel_test(
+    TestStruct,
+    name="test_struct_mutate_attributes",
+    kernel=test_struct_mutate_attributes_kernel,
+    dim=1,
+    inputs=[],
+    devices=devices,
+)
+add_kernel_test(
+    TestStruct,
+    kernel=test_uninitialized,
+    name="test_uninitialized",
+    dim=1,
+    inputs=[Uninitialized()],
+    devices=devices,
+)
+add_kernel_test(TestStruct, kernel=test_return, name="test_return", dim=1, inputs=[], devices=devices)
+add_function_test(TestStruct, "test_nested_struct", test_nested_struct, devices=devices)
+add_function_test(TestStruct, "test_nested_array_struct", test_nested_array_struct, devices=devices)
+add_function_test(TestStruct, "test_nested_empty_struct", test_nested_empty_struct, devices=devices)
+add_function_test(TestStruct, "test_struct_math_conversions", test_struct_math_conversions, devices=devices)
+add_function_test(
+    TestStruct, "test_struct_default_attributes_python", test_struct_default_attributes_python, devices=devices
+)
+add_kernel_test(
+    TestStruct,
+    name="test_struct_default_attributes",
+    kernel=test_struct_default_attributes_kernel,
+    dim=1,
+    inputs=[],
+    devices=devices,
+)
+
+add_kernel_test(
+    TestStruct,
+    name="test_struct_mutate_attributes",
+    kernel=test_struct_mutate_attributes_kernel,
+    dim=1,
+    inputs=[],
+    devices=devices,
+)
+
+for device in devices:
     add_kernel_test(
         TestStruct,
-        name="test_struct_mutate_attributes",
-        kernel=test_struct_mutate_attributes_kernel,
+        kernel=test_struct_instantiate,
+        name="test_struct_instantiate",
         dim=1,
-        inputs=[],
-        devices=devices,
+        inputs=[wp.array([1], dtype=int, device=device)],
+        devices=[device],
     )
-
-    for device in devices:
-        add_kernel_test(
-            TestStruct,
-            kernel=test_struct_instantiate,
-            name="test_struct_instantiate",
-            dim=1,
-            inputs=[wp.array([1], dtype=int, device=device)],
-            devices=[device],
-        )
-        add_kernel_test(
-            TestStruct,
-            kernel=test_return_struct,
-            name="test_return_struct",
-            dim=1,
-            inputs=[wp.zeros(10, dtype=int, device=device)],
-            devices=[device],
-        )
-
     add_kernel_test(
         TestStruct,
-        kernel=test_dependent_module_import,
-        name="test_dependent_module_import",
+        kernel=test_return_struct,
+        name="test_return_struct",
         dim=1,
-        inputs=[DependentModuleImport_C()],
-        devices=devices,
+        inputs=[wp.zeros(10, dtype=int, device=device)],
+        devices=[device],
     )
 
-    return TestStruct
+add_kernel_test(
+    TestStruct,
+    kernel=test_dependent_module_import,
+    name="test_dependent_module_import",
+    dim=1,
+    inputs=[DependentModuleImport_C()],
+    devices=devices,
+)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_tape.py b/warp/tests/test_tape.py
index 1f28de38c..598ef2446 100644
--- a/warp/tests/test_tape.py
+++ b/warp/tests/test_tape.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -127,28 +127,22 @@ def test_tape_dot_product(test, device):
     assert_np_equal(tape.gradients[y].numpy(), x.numpy())
 
 
-def test_tape_no_nested_tapes(test, device):
-    with test.assertRaises(RuntimeError):
-        with wp.Tape():
-            with wp.Tape():
-                pass
-
+devices = get_test_devices()
 
-def register(parent):
-    devices = get_test_devices()
 
-    class TestTape(parent):
-        pass
+class TestTape(unittest.TestCase):
+    def test_tape_no_nested_tapes(self):
+        with self.assertRaises(RuntimeError):
+            with wp.Tape():
+                with wp.Tape():
+                    pass
 
-    add_function_test(TestTape, "test_tape_mul_constant", test_tape_mul_constant, devices=devices)
-    add_function_test(TestTape, "test_tape_mul_variable", test_tape_mul_variable, devices=devices)
-    add_function_test(TestTape, "test_tape_dot_product", test_tape_dot_product, devices=devices)
-    add_function_test(TestTape, "test_tape_no_nested_tapes", test_tape_no_nested_tapes, devices=devices)
 
-    return TestTape
+add_function_test(TestTape, "test_tape_mul_constant", test_tape_mul_constant, devices=devices)
+add_function_test(TestTape, "test_tape_mul_variable", test_tape_mul_variable, devices=devices)
+add_function_test(TestTape, "test_tape_dot_product", test_tape_dot_product, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_torch.py b/warp/tests/test_torch.py
index 107dc90e3..bc69a3301 100644
--- a/warp/tests/test_torch.py
+++ b/warp/tests/test_torch.py
@@ -7,11 +7,10 @@
 
 import unittest
 
-# include parent path
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -470,6 +469,8 @@ def backward(ctx, adj_y):
 def test_torch_graph_torch_stream(test, device):
     """Capture Torch graph on Torch stream"""
 
+    wp.load_module(device=device)
+
     import torch
 
     torch_device = wp.device_to_torch(device)
@@ -551,12 +552,14 @@ def test_warp_graph_warp_stream(test, device):
 
     # capture graph
     with wp.ScopedDevice(device), torch.cuda.stream(torch_stream):
-        wp.capture_begin()
-        t += 1.0
-        wp.launch(inc, dim=n, inputs=[a])
-        t += 1.0
-        wp.launch(inc, dim=n, inputs=[a])
-        g = wp.capture_end()
+        wp.capture_begin(force_module_load=False)
+        try:
+            t += 1.0
+            wp.launch(inc, dim=n, inputs=[a])
+            t += 1.0
+            wp.launch(inc, dim=n, inputs=[a])
+        finally:
+            g = wp.capture_end()
 
     # replay graph
     num_iters = 10
@@ -570,6 +573,8 @@ def test_warp_graph_warp_stream(test, device):
 def test_warp_graph_torch_stream(test, device):
     """Capture Warp graph on Torch stream"""
 
+    wp.load_module(device=device)
+
     import torch
 
     torch_device = wp.device_to_torch(device)
@@ -587,12 +592,14 @@ def test_warp_graph_torch_stream(test, device):
 
     # capture graph
     with wp.ScopedStream(warp_stream), torch.cuda.stream(torch_stream):
-        wp.capture_begin()
-        t += 1.0
-        wp.launch(inc, dim=n, inputs=[a])
-        t += 1.0
-        wp.launch(inc, dim=n, inputs=[a])
-        g = wp.capture_end()
+        wp.capture_begin(force_module_load=False)
+        try:
+            t += 1.0
+            wp.launch(inc, dim=n, inputs=[a])
+            t += 1.0
+            wp.launch(inc, dim=n, inputs=[a])
+        finally:
+            g = wp.capture_end()
 
     # replay graph
     num_iters = 10
@@ -603,83 +610,79 @@ def test_warp_graph_torch_stream(test, device):
     assert passed.item()
 
 
-def register(parent):
-    class TestTorch(parent):
-        pass
-
-    try:
-        import torch
-
-        # check which Warp devices work with Torch
-        # CUDA devices may fail if Torch was not compiled with CUDA support
-        test_devices = get_test_devices()
-        torch_compatible_devices = []
-        torch_compatible_cuda_devices = []
-
-        for d in test_devices:
-            try:
-                t = torch.arange(10, device=wp.device_to_torch(d))
-                t += 1
-                torch_compatible_devices.append(d)
-                if d.is_cuda:
-                    torch_compatible_cuda_devices.append(d)
-            except Exception as e:
-                print(f"Skipping Torch tests on device '{d}' due to exception: {e}")
-
-        if torch_compatible_devices:
-            add_function_test(TestTorch, "test_from_torch", test_from_torch, devices=torch_compatible_devices)
-            add_function_test(
-                TestTorch, "test_from_torch_slices", test_from_torch_slices, devices=torch_compatible_devices
-            )
-            add_function_test(
-                TestTorch,
-                "test_from_torch_zero_strides",
-                test_from_torch_zero_strides,
-                devices=torch_compatible_devices,
-            )
-            add_function_test(TestTorch, "test_to_torch", test_to_torch, devices=torch_compatible_devices)
-            add_function_test(TestTorch, "test_torch_zerocopy", test_torch_zerocopy, devices=torch_compatible_devices)
-            add_function_test(TestTorch, "test_torch_autograd", test_torch_autograd, devices=torch_compatible_devices)
-
-        if torch_compatible_cuda_devices:
-            add_function_test(
-                TestTorch,
-                "test_torch_graph_torch_stream",
-                test_torch_graph_torch_stream,
-                devices=torch_compatible_cuda_devices,
-            )
-            add_function_test(
-                TestTorch,
-                "test_torch_graph_warp_stream",
-                test_torch_graph_warp_stream,
-                devices=torch_compatible_cuda_devices,
-            )
-            add_function_test(
-                TestTorch,
-                "test_warp_graph_warp_stream",
-                test_warp_graph_warp_stream,
-                devices=torch_compatible_cuda_devices,
-            )
-            add_function_test(
-                TestTorch,
-                "test_warp_graph_torch_stream",
-                test_warp_graph_torch_stream,
-                devices=torch_compatible_cuda_devices,
-            )
+class TestTorch(unittest.TestCase):
+    pass
 
-        # multi-GPU tests
-        if len(torch_compatible_cuda_devices) > 1:
-            add_function_test(TestTorch, "test_torch_mgpu_from_torch", test_torch_mgpu_from_torch)
-            add_function_test(TestTorch, "test_torch_mgpu_to_torch", test_torch_mgpu_to_torch)
-            add_function_test(TestTorch, "test_torch_mgpu_interop", test_torch_mgpu_interop)
 
-    except Exception as e:
-        print(f"Skipping Torch tests due to exception: {e}")
+test_devices = get_test_devices()
+
+try:
+    import torch
 
-    return TestTorch
+    # check which Warp devices work with Torch
+    # CUDA devices may fail if Torch was not compiled with CUDA support
+    torch_compatible_devices = []
+    torch_compatible_cuda_devices = []
+
+    for d in test_devices:
+        try:
+            t = torch.arange(10, device=wp.device_to_torch(d))
+            t += 1
+            torch_compatible_devices.append(d)
+            if d.is_cuda:
+                torch_compatible_cuda_devices.append(d)
+        except Exception as e:
+            print(f"Skipping Torch tests on device '{d}' due to exception: {e}")
+
+    if torch_compatible_devices:
+        add_function_test(TestTorch, "test_from_torch", test_from_torch, devices=torch_compatible_devices)
+        add_function_test(TestTorch, "test_from_torch_slices", test_from_torch_slices, devices=torch_compatible_devices)
+        add_function_test(
+            TestTorch,
+            "test_from_torch_zero_strides",
+            test_from_torch_zero_strides,
+            devices=torch_compatible_devices,
+        )
+        add_function_test(TestTorch, "test_to_torch", test_to_torch, devices=torch_compatible_devices)
+        add_function_test(TestTorch, "test_torch_zerocopy", test_torch_zerocopy, devices=torch_compatible_devices)
+        add_function_test(TestTorch, "test_torch_autograd", test_torch_autograd, devices=torch_compatible_devices)
+
+    if torch_compatible_cuda_devices:
+        add_function_test(
+            TestTorch,
+            "test_torch_graph_torch_stream",
+            test_torch_graph_torch_stream,
+            devices=torch_compatible_cuda_devices,
+        )
+        add_function_test(
+            TestTorch,
+            "test_torch_graph_warp_stream",
+            test_torch_graph_warp_stream,
+            devices=torch_compatible_cuda_devices,
+        )
+        add_function_test(
+            TestTorch,
+            "test_warp_graph_warp_stream",
+            test_warp_graph_warp_stream,
+            devices=torch_compatible_cuda_devices,
+        )
+        add_function_test(
+            TestTorch,
+            "test_warp_graph_torch_stream",
+            test_warp_graph_torch_stream,
+            devices=torch_compatible_cuda_devices,
+        )
+
+    # multi-GPU tests
+    if len(torch_compatible_cuda_devices) > 1:
+        add_function_test(TestTorch, "test_torch_mgpu_from_torch", test_torch_mgpu_from_torch)
+        add_function_test(TestTorch, "test_torch_mgpu_to_torch", test_torch_mgpu_to_torch)
+        add_function_test(TestTorch, "test_torch_mgpu_interop", test_torch_mgpu_interop)
+
+except Exception as e:
+    print(f"Skipping Torch tests due to exception: {e}")
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_transient_module.py b/warp/tests/test_transient_module.py
index e35b44ba6..326622c85 100644
--- a/warp/tests/test_transient_module.py
+++ b/warp/tests/test_transient_module.py
@@ -5,14 +5,13 @@
 # distribution of this software and related documentation without an express
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
-import importlib
 import os
 import tempfile
 import unittest
+from importlib import util
 
 import warp as wp
-from warp.tests.test_base import *
-from importlib import util
+from warp.tests.unittest_utils import *
 
 CODE = """# -*- coding: utf-8 -*-
 
@@ -64,27 +63,25 @@ def test_transient_module(test, device):
     assert len(module.compute.module.functions) == 1
 
     data = module.Data()
-    data.x = wp.array([123], dtype=int)
+    data.x = wp.array([123], dtype=int, device=device)
 
     wp.set_module_options({"foo": "bar"}, module=module)
     assert wp.get_module_options(module=module).get("foo") == "bar"
     assert module.compute.module.options.get("foo") == "bar"
 
-    wp.launch(module.compute, dim=1, inputs=[data])
+    wp.launch(module.compute, dim=1, inputs=[data], device=device)
     assert_np_equal(data.x.numpy(), np.array([124]))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
+
 
-    class TestTransientModule(parent):
-        pass
+class TestTransientModule(unittest.TestCase):
+    pass
 
-    add_function_test(TestTransientModule, "test_transient_module", test_transient_module, devices=devices)
-    return TestTransientModule
 
+add_function_test(TestTransientModule, "test_transient_module", test_transient_module, devices=devices)
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_types.py b/warp/tests/test_types.py
index 450584aa1..bc91c13c2 100644
--- a/warp/tests/test_types.py
+++ b/warp/tests/test_types.py
@@ -7,397 +7,357 @@
 
 import unittest
 
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
 
-def test_constant(test, device):
-    const = wp.constant(123)
-    test.assertEqual(const, 123)
-
-    const = wp.constant(1.25)
-    test.assertEqual(const, 1.25)
-
-    const = wp.constant(True)
-    test.assertEqual(const, True)
-
-    const = wp.constant(wp.float16(1.25))
-    test.assertEqual(const.value, 1.25)
-
-    const = wp.constant(wp.int16(123))
-    test.assertEqual(const.value, 123)
-
-    const = wp.constant(wp.vec3i(1, 2, 3))
-    test.assertEqual(const, wp.vec3i(1, 2, 3))
-
-
-def test_constant_error_invalid_type(test, device):
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Invalid constant type: <class 'tuple'>$",
-    ):
-        wp.constant((1, 2, 3))
-
-
-def test_vector(test, device, dtype):
-    def make_scalar(x):
-        # Cast to the correct integer type to simulate wrapping.
-        if dtype in wp.types.int_types:
-            return dtype._type_(x).value
-
-        return x
-
-    def make_vec(*args):
-        if dtype in wp.types.int_types:
-            # Cast to the correct integer type to simulate wrapping.
-            return tuple(dtype._type_(x).value for x in args)
-
-        return args
-
-    vec3_cls = wp.vec(3, dtype)
-    vec4_cls = wp.vec(4, dtype)
-
-    v = vec4_cls(1, 2, 3, 4)
-    test.assertEqual(v[0], make_scalar(1))
-    test.assertEqual(v.x, make_scalar(1))
-    test.assertEqual(v.y, make_scalar(2))
-    test.assertEqual(v.z, make_scalar(3))
-    test.assertEqual(v.w, make_scalar(4))
-    test.assertSequenceEqual(v[0:2], make_vec(1, 2))
-    test.assertSequenceEqual(v, make_vec(1, 2, 3, 4))
-
-    v[0] = -1
-    test.assertEqual(v[0], make_scalar(-1))
-    test.assertEqual(v.x, make_scalar(-1))
-    test.assertEqual(v.y, make_scalar(2))
-    test.assertEqual(v.z, make_scalar(3))
-    test.assertEqual(v.w, make_scalar(4))
-    test.assertSequenceEqual(v[0:2], make_vec(-1, 2))
-    test.assertSequenceEqual(v, make_vec(-1, 2, 3, 4))
-
-    v[1:3] = (-2, -3)
-    test.assertEqual(v[0], make_scalar(-1))
-    test.assertEqual(v.x, make_scalar(-1))
-    test.assertEqual(v.y, make_scalar(-2))
-    test.assertEqual(v.z, make_scalar(-3))
-    test.assertEqual(v.w, make_scalar(4))
-    test.assertSequenceEqual(v[0:2], make_vec(-1, -2))
-    test.assertSequenceEqual(v, make_vec(-1, -2, -3, 4))
-
-    v.x = 1
-    test.assertEqual(v[0], make_scalar(1))
-    test.assertEqual(v.x, make_scalar(1))
-    test.assertEqual(v.y, make_scalar(-2))
-    test.assertEqual(v.z, make_scalar(-3))
-    test.assertEqual(v.w, make_scalar(4))
-    test.assertSequenceEqual(v[0:2], make_vec(1, -2))
-    test.assertSequenceEqual(v, make_vec(1, -2, -3, 4))
-
-    v = vec3_cls(2, 4, 6)
-    test.assertSequenceEqual(+v, make_vec(2, 4, 6))
-    test.assertSequenceEqual(-v, make_vec(-2, -4, -6))
-    test.assertSequenceEqual(v + vec3_cls(1, 1, 1), make_vec(3, 5, 7))
-    test.assertSequenceEqual(v - vec3_cls(1, 1, 1), make_vec(1, 3, 5))
-    test.assertSequenceEqual(v * dtype(2), make_vec(4, 8, 12))
-    test.assertSequenceEqual(dtype(2) * v, make_vec(4, 8, 12))
-    test.assertSequenceEqual(v / dtype(2), make_vec(1, 2, 3))
-    test.assertSequenceEqual(dtype(12) / v, make_vec(6, 3, 2))
-
-    test.assertTrue(v != vec3_cls(1, 2, 3))
-    test.assertEqual(str(v), "[{}]".format(", ".join(str(x) for x in v)))
-
-    # Check added purely for coverage reasons but is this really a desired
-    # behaviour? Not allowing to define new attributes using systems like
-    # `__slots__` could help improving memory usage.
-    v.foo = 123
-    test.assertEqual(v.foo, 123)
-
-
-def test_vector_error_invalid_arg_count(test, device):
-    with test.assertRaisesRegex(
-        ValueError,
-        r"Invalid number of arguments in vector constructor, expected 3 elements, got 2$",
-    ):
-        wp.vec3(1, 2)
-
-
-def test_vector_error_invalid_ptr(test, device):
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"NULL pointer exception",
-    ):
-        wp.vec3.from_ptr(0)
-
-
-def test_vector_error_invalid_get_item_key(test, device):
-    v = wp.vec3(1, 2, 3)
-
-    with test.assertRaisesRegex(
-        KeyError,
-        r"Invalid key None, expected int or slice",
-    ):
-        v[None]
-
-
-def test_vector_error_invalid_set_item_key(test, device):
-    v = wp.vec3(1, 2, 3)
-    with test.assertRaisesRegex(
-        KeyError,
-        r"Invalid key None, expected int or slice",
-    ):
-        v[None] = 0
-
-
-def test_matrix(test, device, dtype):
-    def make_scalar(x):
-        # Cast to the correct integer type to simulate wrapping.
-        if dtype in wp.types.int_types:
-            return dtype._type_(x).value
-
-        return x
-
-    def make_vec(*args):
-        if dtype in wp.types.int_types:
-            # Cast to the correct integer type to simulate wrapping.
-            return tuple(dtype._type_(x).value for x in args)
-
-        return args
-
-    def make_mat(*args):
-        if dtype in wp.types.int_types:
-            # Cast to the correct integer type to simulate wrapping.
-            return tuple(tuple(dtype._type_(x).value for x in row) for row in args)
-
-        return args
-
-    mat22_cls = wp.mat((2, 2), dtype)
-    mat33_cls = wp.mat((3, 3), dtype)
-    vec2_cls = wp.vec(2, dtype)
-
-    m = mat33_cls(((1, 2, 3), (4, 5, 6), (7, 8, 9)))
-    test.assertEqual(m[0][0], make_scalar(1))
-    test.assertEqual(m[0][1], make_scalar(2))
-    test.assertEqual(m[0][2], make_scalar(3))
-    test.assertEqual(m[1][0], make_scalar(4))
-    test.assertEqual(m[1][1], make_scalar(5))
-    test.assertEqual(m[1][2], make_scalar(6))
-    test.assertEqual(m[2][0], make_scalar(7))
-    test.assertEqual(m[2][1], make_scalar(8))
-    test.assertEqual(m[2][2], make_scalar(9))
-    test.assertEqual(m[0, 0], make_scalar(1))
-    test.assertEqual(m[0, 1], make_scalar(2))
-    test.assertEqual(m[0, 2], make_scalar(3))
-    test.assertEqual(m[1, 0], make_scalar(4))
-    test.assertEqual(m[1, 1], make_scalar(5))
-    test.assertEqual(m[1, 2], make_scalar(6))
-    test.assertEqual(m[2, 0], make_scalar(7))
-    test.assertEqual(m[2, 1], make_scalar(8))
-    test.assertEqual(m[2, 2], make_scalar(9))
-    test.assertSequenceEqual(m[0], make_vec(1, 2, 3))
-    test.assertSequenceEqual(m[1], make_vec(4, 5, 6))
-    test.assertSequenceEqual(m[2], make_vec(7, 8, 9))
-    test.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
-    test.assertSequenceEqual(m[1][0:2], make_vec(4, 5))
-    test.assertSequenceEqual(m[2][0:3], make_vec(7, 8, 9))
-    # test.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
-    # test.assertSequenceEqual(m[1, 0:2], make_vec(4, 5))
-    # test.assertSequenceEqual(m[2, 0:3], make_vec(7, 8, 9))
-    test.assertSequenceEqual(m, make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
-
-    m[1, 0] = -4
-    test.assertEqual(m[0][0], make_scalar(1))
-    test.assertEqual(m[0][1], make_scalar(2))
-    test.assertEqual(m[0][2], make_scalar(3))
-    test.assertEqual(m[1][0], make_scalar(-4))
-    test.assertEqual(m[1][1], make_scalar(5))
-    test.assertEqual(m[1][2], make_scalar(6))
-    test.assertEqual(m[2][0], make_scalar(7))
-    test.assertEqual(m[2][1], make_scalar(8))
-    test.assertEqual(m[2][2], make_scalar(9))
-    test.assertEqual(m[0, 0], make_scalar(1))
-    test.assertEqual(m[0, 1], make_scalar(2))
-    test.assertEqual(m[0, 2], make_scalar(3))
-    test.assertEqual(m[1, 0], make_scalar(-4))
-    test.assertEqual(m[1, 1], make_scalar(5))
-    test.assertEqual(m[1, 2], make_scalar(6))
-    test.assertEqual(m[2, 0], make_scalar(7))
-    test.assertEqual(m[2, 1], make_scalar(8))
-    test.assertEqual(m[2, 2], make_scalar(9))
-    test.assertSequenceEqual(m[0], make_vec(1, 2, 3))
-    test.assertSequenceEqual(m[1], make_vec(-4, 5, 6))
-    test.assertSequenceEqual(m[2], make_vec(7, 8, 9))
-    test.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
-    test.assertSequenceEqual(m[1][0:2], make_vec(-4, 5))
-    test.assertSequenceEqual(m[2][0:3], make_vec(7, 8, 9))
-    # test.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
-    # test.assertSequenceEqual(m[1, 0:2], make_vec(-4, 5))
-    # test.assertSequenceEqual(m[2, 0:3], make_vec(7, 8, 9))
-    test.assertSequenceEqual(m, make_mat((1, 2, 3), (-4, 5, 6), (7, 8, 9)))
-
-    m[2] = (-7, 8, -9)
-    test.assertEqual(m[0][0], make_scalar(1))
-    test.assertEqual(m[0][1], make_scalar(2))
-    test.assertEqual(m[0][2], make_scalar(3))
-    test.assertEqual(m[1][0], make_scalar(-4))
-    test.assertEqual(m[1][1], make_scalar(5))
-    test.assertEqual(m[1][2], make_scalar(6))
-    test.assertEqual(m[2][0], make_scalar(-7))
-    test.assertEqual(m[2][1], make_scalar(8))
-    test.assertEqual(m[2][2], make_scalar(-9))
-    test.assertEqual(m[0, 0], make_scalar(1))
-    test.assertEqual(m[0, 1], make_scalar(2))
-    test.assertEqual(m[0, 2], make_scalar(3))
-    test.assertEqual(m[1, 0], make_scalar(-4))
-    test.assertEqual(m[1, 1], make_scalar(5))
-    test.assertEqual(m[1, 2], make_scalar(6))
-    test.assertEqual(m[2, 0], make_scalar(-7))
-    test.assertEqual(m[2, 1], make_scalar(8))
-    test.assertEqual(m[2, 2], make_scalar(-9))
-    test.assertSequenceEqual(m[0], make_vec(1, 2, 3))
-    test.assertSequenceEqual(m[1], make_vec(-4, 5, 6))
-    test.assertSequenceEqual(m[2], make_vec(-7, 8, -9))
-    test.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
-    test.assertSequenceEqual(m[1][0:2], make_vec(-4, 5))
-    test.assertSequenceEqual(m[2][0:3], make_vec(-7, 8, -9))
-    # test.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
-    # test.assertSequenceEqual(m[1, 0:2], make_vec(-4, 5))
-    # test.assertSequenceEqual(m[2, 0:3], make_vec(-7, 8, -9))
-    test.assertSequenceEqual(m, make_mat((1, 2, 3), (-4, 5, 6), (-7, 8, -9)))
-
-    m = mat22_cls(2, 4, 6, 8)
-    test.assertSequenceEqual(+m, make_mat((2, 4), (6, 8)))
-    test.assertSequenceEqual(-m, make_mat((-2, -4), (-6, -8)))
-    test.assertSequenceEqual(m + mat22_cls(1, 1, 1, 1), make_mat((3, 5), (7, 9)))
-    test.assertSequenceEqual(m - mat22_cls(1, 1, 1, 1), make_mat((1, 3), (5, 7)))
-    test.assertSequenceEqual(m * dtype(2), make_mat((4, 8), (12, 16)))
-    test.assertSequenceEqual(dtype(2) * m, make_mat((4, 8), (12, 16)))
-    test.assertSequenceEqual(m / dtype(2), make_mat((1, 2), (3, 4)))
-    test.assertSequenceEqual(dtype(24) / m, make_mat((12, 6), (4, 3)))
-
-    test.assertSequenceEqual(m * vec2_cls(1, 2), make_vec(10, 22))
-    test.assertSequenceEqual(m @ vec2_cls(1, 2), make_vec(10, 22))
-    test.assertSequenceEqual(vec2_cls(1, 2) * m, make_vec(14, 20))
-    test.assertSequenceEqual(vec2_cls(1, 2) @ m, make_vec(14, 20))
-
-    test.assertTrue(m != mat22_cls(1, 2, 3, 4))
-    test.assertEqual(
-        str(m), "[{}]".format(",\n ".join("[{}]".format(", ".join(str(y) for y in m[x])) for x in range(m._shape_[0])))
-    )
-
-    # Check added purely for coverage reasons but is this really a desired
-    # behaviour? Not allowing to define new attributes using systems like
-    # `__slots__` could help improving memory usage.
-    m.foo = 123
-    test.assertEqual(m.foo, 123)
-
-
-def test_matrix_error_invalid_arg_count(test, device):
-    with test.assertRaisesRegex(
-        ValueError,
-        r"Invalid number of arguments in matrix constructor, expected 4 elements, got 3$",
-    ):
-        wp.mat22(1, 2, 3)
-
-
-def test_matrix_error_invalid_row_count(test, device):
-    with test.assertRaisesRegex(
-        TypeError,
-        r"Invalid argument in matrix constructor, expected row of length 2, got \(1, 2, 3\)$",
-    ):
-        wp.mat22((1, 2, 3), (3, 4, 5))
-
-
-def test_matrix_error_invalid_ptr(test, device):
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"NULL pointer exception",
-    ):
-        wp.mat22.from_ptr(0)
-
-
-def test_matrix_error_invalid_set_row_index(test, device):
-    m = wp.mat22(1, 2, 3, 4)
-    with test.assertRaisesRegex(
-        IndexError,
-        r"Invalid row index$",
-    ):
-        m.set_row(2, (0, 0))
-
-
-def test_matrix_error_invalid_get_item_key(test, device):
-    m = wp.mat22(1, 2, 3, 4)
-
-    with test.assertRaisesRegex(
-        KeyError,
-        r"Invalid key None, expected int or pair of ints",
-    ):
-        m[None]
-
-
-def test_matrix_error_invalid_get_item_key_length(test, device):
-    m = wp.mat22(1, 2, 3, 4)
-
-    with test.assertRaisesRegex(
-        KeyError,
-        r"Invalid key, expected one or two indices, got 3",
-    ):
-        m[0, 1, 2]
-
-
-def test_matrix_error_invalid_set_item_key(test, device):
-    m = wp.mat22(1, 2, 3, 4)
-    with test.assertRaisesRegex(
-        KeyError,
-        r"Invalid key None, expected int or pair of ints",
-    ):
-        m[None] = 0
-
-
-def test_matrix_error_invalid_set_item_key_length(test, device):
-    m = wp.mat22(1, 2, 3, 4)
-
-    with test.assertRaisesRegex(
-        KeyError,
-        r"Invalid key, expected one or two indices, got 3",
-    ):
-        m[0, 1, 2] = (0, 0)
-
-
-def register(parent):
-    devices = get_test_devices()
-
-    class TestTypes(parent):
-        pass
-
-    add_function_test(TestTypes, "test_constant", test_constant)
-    add_function_test(TestTypes, "test_constant_error_invalid_type", test_constant_error_invalid_type)
-
-    for dtype in tuple(wp.types.scalar_types) + (int, float):
-        add_function_test(TestTypes, f"test_vector_{dtype.__name__}", test_vector, devices=None, dtype=dtype)
-
-    add_function_test(TestTypes, "test_vector_error_invalid_arg_count", test_vector_error_invalid_arg_count)
-    add_function_test(TestTypes, "test_vector_error_invalid_ptr", test_vector_error_invalid_ptr)
-    add_function_test(TestTypes, "test_vector_error_invalid_get_item_key", test_vector_error_invalid_get_item_key)
-    add_function_test(TestTypes, "test_vector_error_invalid_set_item_key", test_vector_error_invalid_set_item_key)
-
-    for dtype in tuple(wp.types.float_types) + (float,):
-        add_function_test(TestTypes, f"test_matrix_{dtype.__name__}", test_matrix, devices=None, dtype=dtype)
-
-    add_function_test(TestTypes, "test_matrix_error_invalid_arg_count", test_matrix_error_invalid_arg_count)
-    add_function_test(TestTypes, "test_matrix_error_invalid_row_count", test_matrix_error_invalid_row_count)
-    add_function_test(TestTypes, "test_matrix_error_invalid_ptr", test_matrix_error_invalid_ptr)
-    add_function_test(TestTypes, "test_matrix_error_invalid_set_row_index", test_matrix_error_invalid_set_row_index)
-    add_function_test(TestTypes, "test_matrix_error_invalid_get_item_key", test_matrix_error_invalid_get_item_key)
-    add_function_test(TestTypes, "test_matrix_error_invalid_get_item_key_length", test_matrix_error_invalid_get_item_key_length)
-    add_function_test(TestTypes, "test_matrix_error_invalid_set_item_key", test_matrix_error_invalid_set_item_key)
-    add_function_test(TestTypes, "test_matrix_error_invalid_set_item_key_length", test_matrix_error_invalid_set_item_key_length)
-
-    return TestTypes
+devices = get_test_devices()
+
+
+class TestTypes(unittest.TestCase):
+    def test_constant(self):
+        const = wp.constant(123)
+        self.assertEqual(const, 123)
+
+        const = wp.constant(1.25)
+        self.assertEqual(const, 1.25)
+
+        const = wp.constant(True)
+        self.assertEqual(const, True)
+
+        const = wp.constant(wp.float16(1.25))
+        self.assertEqual(const.value, 1.25)
+
+        const = wp.constant(wp.int16(123))
+        self.assertEqual(const.value, 123)
+
+        const = wp.constant(wp.vec3i(1, 2, 3))
+        self.assertEqual(const, wp.vec3i(1, 2, 3))
+
+    def test_constant_error_invalid_type(self):
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Invalid constant type: <class 'tuple'>$",
+        ):
+            wp.constant((1, 2, 3))
+
+    def test_vector(self):
+        for dtype in tuple(wp.types.scalar_types) + (int, float):
+
+            def make_scalar(x):
+                # Cast to the correct integer type to simulate wrapping.
+                if dtype in wp.types.int_types:
+                    return dtype._type_(x).value
+
+                return x
+
+            def make_vec(*args):
+                if dtype in wp.types.int_types:
+                    # Cast to the correct integer type to simulate wrapping.
+                    return tuple(dtype._type_(x).value for x in args)
+
+                return args
+
+            vec3_cls = wp.vec(3, dtype)
+            vec4_cls = wp.vec(4, dtype)
+
+            v = vec4_cls(1, 2, 3, 4)
+            self.assertEqual(v[0], make_scalar(1))
+            self.assertEqual(v.x, make_scalar(1))
+            self.assertEqual(v.y, make_scalar(2))
+            self.assertEqual(v.z, make_scalar(3))
+            self.assertEqual(v.w, make_scalar(4))
+            self.assertSequenceEqual(v[0:2], make_vec(1, 2))
+            self.assertSequenceEqual(v, make_vec(1, 2, 3, 4))
+
+            v[0] = -1
+            self.assertEqual(v[0], make_scalar(-1))
+            self.assertEqual(v.x, make_scalar(-1))
+            self.assertEqual(v.y, make_scalar(2))
+            self.assertEqual(v.z, make_scalar(3))
+            self.assertEqual(v.w, make_scalar(4))
+            self.assertSequenceEqual(v[0:2], make_vec(-1, 2))
+            self.assertSequenceEqual(v, make_vec(-1, 2, 3, 4))
+
+            v[1:3] = (-2, -3)
+            self.assertEqual(v[0], make_scalar(-1))
+            self.assertEqual(v.x, make_scalar(-1))
+            self.assertEqual(v.y, make_scalar(-2))
+            self.assertEqual(v.z, make_scalar(-3))
+            self.assertEqual(v.w, make_scalar(4))
+            self.assertSequenceEqual(v[0:2], make_vec(-1, -2))
+            self.assertSequenceEqual(v, make_vec(-1, -2, -3, 4))
+
+            v.x = 1
+            self.assertEqual(v[0], make_scalar(1))
+            self.assertEqual(v.x, make_scalar(1))
+            self.assertEqual(v.y, make_scalar(-2))
+            self.assertEqual(v.z, make_scalar(-3))
+            self.assertEqual(v.w, make_scalar(4))
+            self.assertSequenceEqual(v[0:2], make_vec(1, -2))
+            self.assertSequenceEqual(v, make_vec(1, -2, -3, 4))
+
+            v = vec3_cls(2, 4, 6)
+            self.assertSequenceEqual(+v, make_vec(2, 4, 6))
+            self.assertSequenceEqual(-v, make_vec(-2, -4, -6))
+            self.assertSequenceEqual(v + vec3_cls(1, 1, 1), make_vec(3, 5, 7))
+            self.assertSequenceEqual(v - vec3_cls(1, 1, 1), make_vec(1, 3, 5))
+            self.assertSequenceEqual(v * dtype(2), make_vec(4, 8, 12))
+            self.assertSequenceEqual(dtype(2) * v, make_vec(4, 8, 12))
+            self.assertSequenceEqual(v / dtype(2), make_vec(1, 2, 3))
+            self.assertSequenceEqual(dtype(12) / v, make_vec(6, 3, 2))
+
+            self.assertTrue(v != vec3_cls(1, 2, 3))
+            self.assertEqual(str(v), "[{}]".format(", ".join(str(x) for x in v)))
+
+            # Check added purely for coverage reasons but is this really a desired
+            # behaviour? Not allowing to define new attributes using systems like
+            # `__slots__` could help improving memory usage.
+            v.foo = 123
+            self.assertEqual(v.foo, 123)
+
+    def test_vector_error_invalid_arg_count(self):
+        with self.assertRaisesRegex(
+            ValueError,
+            r"Invalid number of arguments in vector constructor, expected 3 elements, got 2$",
+        ):
+            wp.vec3(1, 2)
+
+    def test_vector_error_invalid_ptr(self):
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"NULL pointer exception",
+        ):
+            wp.vec3.from_ptr(0)
+
+    def test_vector_error_invalid_get_item_key(self):
+        v = wp.vec3(1, 2, 3)
+
+        with self.assertRaisesRegex(
+            KeyError,
+            r"Invalid key None, expected int or slice",
+        ):
+            v[None]
+
+    def test_vector_error_invalid_set_item_key(self):
+        v = wp.vec3(1, 2, 3)
+        with self.assertRaisesRegex(
+            KeyError,
+            r"Invalid key None, expected int or slice",
+        ):
+            v[None] = 0
+
+    def test_matrix(self):
+        for dtype in tuple(wp.types.float_types) + (float,):
+
+            def make_scalar(x):
+                # Cast to the correct integer type to simulate wrapping.
+                if dtype in wp.types.int_types:
+                    return dtype._type_(x).value
+
+                return x
+
+            def make_vec(*args):
+                if dtype in wp.types.int_types:
+                    # Cast to the correct integer type to simulate wrapping.
+                    return tuple(dtype._type_(x).value for x in args)
+
+                return args
+
+            def make_mat(*args):
+                if dtype in wp.types.int_types:
+                    # Cast to the correct integer type to simulate wrapping.
+                    return tuple(tuple(dtype._type_(x).value for x in row) for row in args)
+
+                return args
+
+            mat22_cls = wp.mat((2, 2), dtype)
+            mat33_cls = wp.mat((3, 3), dtype)
+            vec2_cls = wp.vec(2, dtype)
+
+            m = mat33_cls(((1, 2, 3), (4, 5, 6), (7, 8, 9)))
+            self.assertEqual(m[0][0], make_scalar(1))
+            self.assertEqual(m[0][1], make_scalar(2))
+            self.assertEqual(m[0][2], make_scalar(3))
+            self.assertEqual(m[1][0], make_scalar(4))
+            self.assertEqual(m[1][1], make_scalar(5))
+            self.assertEqual(m[1][2], make_scalar(6))
+            self.assertEqual(m[2][0], make_scalar(7))
+            self.assertEqual(m[2][1], make_scalar(8))
+            self.assertEqual(m[2][2], make_scalar(9))
+            self.assertEqual(m[0, 0], make_scalar(1))
+            self.assertEqual(m[0, 1], make_scalar(2))
+            self.assertEqual(m[0, 2], make_scalar(3))
+            self.assertEqual(m[1, 0], make_scalar(4))
+            self.assertEqual(m[1, 1], make_scalar(5))
+            self.assertEqual(m[1, 2], make_scalar(6))
+            self.assertEqual(m[2, 0], make_scalar(7))
+            self.assertEqual(m[2, 1], make_scalar(8))
+            self.assertEqual(m[2, 2], make_scalar(9))
+            self.assertSequenceEqual(m[0], make_vec(1, 2, 3))
+            self.assertSequenceEqual(m[1], make_vec(4, 5, 6))
+            self.assertSequenceEqual(m[2], make_vec(7, 8, 9))
+            self.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
+            self.assertSequenceEqual(m[1][0:2], make_vec(4, 5))
+            self.assertSequenceEqual(m[2][0:3], make_vec(7, 8, 9))
+            # self.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
+            # self.assertSequenceEqual(m[1, 0:2], make_vec(4, 5))
+            # self.assertSequenceEqual(m[2, 0:3], make_vec(7, 8, 9))
+            self.assertSequenceEqual(m, make_mat((1, 2, 3), (4, 5, 6), (7, 8, 9)))
+
+            m[1, 0] = -4
+            self.assertEqual(m[0][0], make_scalar(1))
+            self.assertEqual(m[0][1], make_scalar(2))
+            self.assertEqual(m[0][2], make_scalar(3))
+            self.assertEqual(m[1][0], make_scalar(-4))
+            self.assertEqual(m[1][1], make_scalar(5))
+            self.assertEqual(m[1][2], make_scalar(6))
+            self.assertEqual(m[2][0], make_scalar(7))
+            self.assertEqual(m[2][1], make_scalar(8))
+            self.assertEqual(m[2][2], make_scalar(9))
+            self.assertEqual(m[0, 0], make_scalar(1))
+            self.assertEqual(m[0, 1], make_scalar(2))
+            self.assertEqual(m[0, 2], make_scalar(3))
+            self.assertEqual(m[1, 0], make_scalar(-4))
+            self.assertEqual(m[1, 1], make_scalar(5))
+            self.assertEqual(m[1, 2], make_scalar(6))
+            self.assertEqual(m[2, 0], make_scalar(7))
+            self.assertEqual(m[2, 1], make_scalar(8))
+            self.assertEqual(m[2, 2], make_scalar(9))
+            self.assertSequenceEqual(m[0], make_vec(1, 2, 3))
+            self.assertSequenceEqual(m[1], make_vec(-4, 5, 6))
+            self.assertSequenceEqual(m[2], make_vec(7, 8, 9))
+            self.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
+            self.assertSequenceEqual(m[1][0:2], make_vec(-4, 5))
+            self.assertSequenceEqual(m[2][0:3], make_vec(7, 8, 9))
+            # self.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
+            # self.assertSequenceEqual(m[1, 0:2], make_vec(-4, 5))
+            # self.assertSequenceEqual(m[2, 0:3], make_vec(7, 8, 9))
+            self.assertSequenceEqual(m, make_mat((1, 2, 3), (-4, 5, 6), (7, 8, 9)))
+
+            m[2] = (-7, 8, -9)
+            self.assertEqual(m[0][0], make_scalar(1))
+            self.assertEqual(m[0][1], make_scalar(2))
+            self.assertEqual(m[0][2], make_scalar(3))
+            self.assertEqual(m[1][0], make_scalar(-4))
+            self.assertEqual(m[1][1], make_scalar(5))
+            self.assertEqual(m[1][2], make_scalar(6))
+            self.assertEqual(m[2][0], make_scalar(-7))
+            self.assertEqual(m[2][1], make_scalar(8))
+            self.assertEqual(m[2][2], make_scalar(-9))
+            self.assertEqual(m[0, 0], make_scalar(1))
+            self.assertEqual(m[0, 1], make_scalar(2))
+            self.assertEqual(m[0, 2], make_scalar(3))
+            self.assertEqual(m[1, 0], make_scalar(-4))
+            self.assertEqual(m[1, 1], make_scalar(5))
+            self.assertEqual(m[1, 2], make_scalar(6))
+            self.assertEqual(m[2, 0], make_scalar(-7))
+            self.assertEqual(m[2, 1], make_scalar(8))
+            self.assertEqual(m[2, 2], make_scalar(-9))
+            self.assertSequenceEqual(m[0], make_vec(1, 2, 3))
+            self.assertSequenceEqual(m[1], make_vec(-4, 5, 6))
+            self.assertSequenceEqual(m[2], make_vec(-7, 8, -9))
+            self.assertSequenceEqual(m[0][1:3], make_vec(2, 3))
+            self.assertSequenceEqual(m[1][0:2], make_vec(-4, 5))
+            self.assertSequenceEqual(m[2][0:3], make_vec(-7, 8, -9))
+            # self.assertSequenceEqual(m[0, 1:3], make_vec(2, 3))
+            # self.assertSequenceEqual(m[1, 0:2], make_vec(-4, 5))
+            # self.assertSequenceEqual(m[2, 0:3], make_vec(-7, 8, -9))
+            self.assertSequenceEqual(m, make_mat((1, 2, 3), (-4, 5, 6), (-7, 8, -9)))
+
+            m = mat22_cls(2, 4, 6, 8)
+            self.assertSequenceEqual(+m, make_mat((2, 4), (6, 8)))
+            self.assertSequenceEqual(-m, make_mat((-2, -4), (-6, -8)))
+            self.assertSequenceEqual(m + mat22_cls(1, 1, 1, 1), make_mat((3, 5), (7, 9)))
+            self.assertSequenceEqual(m - mat22_cls(1, 1, 1, 1), make_mat((1, 3), (5, 7)))
+            self.assertSequenceEqual(m * dtype(2), make_mat((4, 8), (12, 16)))
+            self.assertSequenceEqual(dtype(2) * m, make_mat((4, 8), (12, 16)))
+            self.assertSequenceEqual(m / dtype(2), make_mat((1, 2), (3, 4)))
+            self.assertSequenceEqual(dtype(24) / m, make_mat((12, 6), (4, 3)))
+
+            self.assertSequenceEqual(m * vec2_cls(1, 2), make_vec(10, 22))
+            self.assertSequenceEqual(m @ vec2_cls(1, 2), make_vec(10, 22))
+            self.assertSequenceEqual(vec2_cls(1, 2) * m, make_vec(14, 20))
+            self.assertSequenceEqual(vec2_cls(1, 2) @ m, make_vec(14, 20))
+
+            self.assertTrue(m != mat22_cls(1, 2, 3, 4))
+            self.assertEqual(
+                str(m),
+                "[{}]".format(",\n ".join("[{}]".format(", ".join(str(y) for y in m[x])) for x in range(m._shape_[0]))),
+            )
+
+            # Check added purely for coverage reasons but is this really a desired
+            # behaviour? Not allowing to define new attributes using systems like
+            # `__slots__` could help improving memory usage.
+            m.foo = 123
+            self.assertEqual(m.foo, 123)
+
+    def test_matrix_error_invalid_arg_count(self):
+        with self.assertRaisesRegex(
+            ValueError,
+            r"Invalid number of arguments in matrix constructor, expected 4 elements, got 3$",
+        ):
+            wp.mat22(1, 2, 3)
+
+    def test_matrix_error_invalid_row_count(self):
+        with self.assertRaisesRegex(
+            TypeError,
+            r"Invalid argument in matrix constructor, expected row of length 2, got \(1, 2, 3\)$",
+        ):
+            wp.mat22((1, 2, 3), (3, 4, 5))
+
+    def test_matrix_error_invalid_ptr(self):
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"NULL pointer exception",
+        ):
+            wp.mat22.from_ptr(0)
+
+    def test_matrix_error_invalid_set_row_index(self):
+        m = wp.mat22(1, 2, 3, 4)
+        with self.assertRaisesRegex(
+            IndexError,
+            r"Invalid row index$",
+        ):
+            m.set_row(2, (0, 0))
+
+    def test_matrix_error_invalid_get_item_key(self):
+        m = wp.mat22(1, 2, 3, 4)
+
+        with self.assertRaisesRegex(
+            KeyError,
+            r"Invalid key None, expected int or pair of ints",
+        ):
+            m[None]
+
+    def test_matrix_error_invalid_get_item_key_length(self):
+        m = wp.mat22(1, 2, 3, 4)
+
+        with self.assertRaisesRegex(
+            KeyError,
+            r"Invalid key, expected one or two indices, got 3",
+        ):
+            m[0, 1, 2]
+
+    def test_matrix_error_invalid_set_item_key(self):
+        m = wp.mat22(1, 2, 3, 4)
+        with self.assertRaisesRegex(
+            KeyError,
+            r"Invalid key None, expected int or pair of ints",
+        ):
+            m[None] = 0
+
+    def test_matrix_error_invalid_set_item_key_length(self):
+        m = wp.mat22(1, 2, 3, 4)
+
+        with self.assertRaisesRegex(
+            KeyError,
+            r"Invalid key, expected one or two indices, got 3",
+        ):
+            m[0, 1, 2] = (0, 0)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_unresolved_func.py b/warp/tests/test_unresolved_func.py
deleted file mode 100644
index 8122be812..000000000
--- a/warp/tests/test_unresolved_func.py
+++ /dev/null
@@ -1,7 +0,0 @@
-import warp as wp
-
-
-@wp.kernel
-def unresolved_func_kernel():
-    # this should trigger an exception due to unresolved function
-    x = wp.missing_func(42)
diff --git a/warp/tests/test_unresolved_symbol.py b/warp/tests/test_unresolved_symbol.py
deleted file mode 100644
index ce2aa241e..000000000
--- a/warp/tests/test_unresolved_symbol.py
+++ /dev/null
@@ -1,7 +0,0 @@
-import warp as wp
-
-
-@wp.kernel
-def unresolved_symbol_kernel():
-    # this should trigger an exception due to unresolved symbol
-    x = missing_symbol
diff --git a/warp/tests/test_utils.py b/warp/tests/test_utils.py
index 7bc5ca869..19ea4469e 100644
--- a/warp/tests/test_utils.py
+++ b/warp/tests/test_utils.py
@@ -6,36 +6,15 @@
 # license agreement from NVIDIA CORPORATION is strictly prohibited.
 
 import contextlib
-import io
 import inspect
+import io
 import unittest
 
-from warp.tests.test_base import *
-
+from warp.tests.unittest_utils import *
 
 wp.init()
 
 
-def test_warn(test, device):
-    with contextlib.redirect_stdout(io.StringIO()) as f:
-        frame_info = inspect.getframeinfo(inspect.currentframe())
-        wp.utils.warn("hello, world!")
-
-    expected = '{}:{}: UserWarning: hello, world!\n  wp.utils.warn("hello, world!")\n'.format(
-        frame_info.filename,
-        frame_info.lineno + 1,
-    )
-    test.assertEqual(f.getvalue(), expected)
-
-
-def test_transform_expand(test, device):
-    t = (1.0, 2.0, 3.0, 4.0, 3.0, 2.0, 1.0)
-    test.assertEqual(
-        wp.utils.transform_expand(t),
-        wp.transformf(p=(1.0, 2.0, 3.0), q=(4.0, 3.0, 2.0, 1.0)),
-    )
-
-
 def test_array_scan(test, device):
     rng = np.random.default_rng(123)
 
@@ -71,19 +50,9 @@ def test_array_scan_empty(test, device):
     wp.utils.array_scan(values, result)
 
 
-def test_array_scan_error_devices_mismatch(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    result = wp.zeros_like(values, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        wp.utils.array_scan(values, result, True)
-
-
 def test_array_scan_error_sizes_mismatch(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    result = wp.zeros(234, dtype=int, device="cpu")
+    values = wp.zeros(123, dtype=int, device=device)
+    result = wp.zeros(234, dtype=int, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Array storage sizes do not match$",
@@ -92,8 +61,8 @@ def test_array_scan_error_sizes_mismatch(test, device):
 
 
 def test_array_scan_error_dtypes_mismatch(test, device):
-    values = wp.zeros(123, dtype=int, device="cpu")
-    result = wp.zeros(123, dtype=float, device="cpu")
+    values = wp.zeros(123, dtype=int, device=device)
+    result = wp.zeros(123, dtype=float, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Array data types do not match$",
@@ -125,19 +94,9 @@ def test_radix_sort_pairs_empty(test, device):
     wp.utils.radix_sort_pairs(keys, values, 0)
 
 
-def test_radix_sort_pairs_error_devices_mismatch(test, device):
-    keys = wp.array((1, 2, 3), dtype=int, device="cpu")
-    values = wp.array((1, 2, 3), dtype=int, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        wp.utils.radix_sort_pairs(keys, values, 1)
-
-
 def test_radix_sort_pairs_error_insufficient_storage(test, device):
-    keys = wp.array((1, 2, 3), dtype=int, device="cpu")
-    values = wp.array((1, 2, 3), dtype=int, device="cpu")
+    keys = wp.array((1, 2, 3), dtype=int, device=device)
+    values = wp.array((1, 2, 3), dtype=int, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Array storage must be large enough to contain 2\*count elements$",
@@ -157,28 +116,19 @@ def test_radix_sort_pairs_error_unsupported_dtype(test, device):
 
 def test_array_sum(test, device):
     for dtype in (wp.float32, wp.float64):
-        values = wp.array((1.0, 2.0, 3.0), dtype=dtype, device=device)
-        test.assertEqual(wp.utils.array_sum(values), 6.0)
-
-        values = wp.array((1.0, 2.0, 3.0), dtype=dtype, device=device)
-        result = wp.empty(shape=(1,), dtype=dtype, device=device)
-        wp.utils.array_sum(values, out=result)
-        test.assertEqual(result.numpy()[0], 6.0)
-
+        with test.subTest(dtype=dtype):
+            values = wp.array((1.0, 2.0, 3.0), dtype=dtype, device=device)
+            test.assertEqual(wp.utils.array_sum(values), 6.0)
 
-def test_array_sum_error_out_device_mismatch(test, device):
-    values = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    result = wp.empty(shape=(1,), dtype=wp.float32, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"out storage device should match values array$",
-    ):
-        wp.utils.array_sum(values, out=result)
+            values = wp.array((1.0, 2.0, 3.0), dtype=dtype, device=device)
+            result = wp.empty(shape=(1,), dtype=dtype, device=device)
+            wp.utils.array_sum(values, out=result)
+            test.assertEqual(result.numpy()[0], 6.0)
 
 
 def test_array_sum_error_out_dtype_mismatch(test, device):
-    values = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    result = wp.empty(shape=(1,), dtype=wp.float64, device="cpu")
+    values = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    result = wp.empty(shape=(1,), dtype=wp.float64, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"out array should have type float32$",
@@ -187,8 +137,8 @@ def test_array_sum_error_out_dtype_mismatch(test, device):
 
 
 def test_array_sum_error_out_shape_mismatch(test, device):
-    values = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    result = wp.empty(shape=(2,), dtype=wp.float32, device="cpu")
+    values = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    result = wp.empty(shape=(2,), dtype=wp.float32, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"out array should have shape \(1,\)$",
@@ -219,8 +169,8 @@ def test_array_inner(test, device):
 
 
 def test_array_inner_error_sizes_mismatch(test, device):
-    a = wp.array((1.0, 2.0), dtype=wp.float32, device="cpu")
-    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
+    a = wp.array((1.0, 2.0), dtype=wp.float32, device=device)
+    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Array storage sizes do not match$",
@@ -228,19 +178,9 @@ def test_array_inner_error_sizes_mismatch(test, device):
         wp.utils.array_inner(a, b)
 
 
-def test_array_inner_error_devices_mismatch(test, device):
-    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        wp.utils.array_inner(a, b)
-
-
 def test_array_inner_error_dtypes_mismatch(test, device):
-    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float64, device="cpu")
+    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float64, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Array data types do not match$",
@@ -248,21 +188,10 @@ def test_array_inner_error_dtypes_mismatch(test, device):
         wp.utils.array_inner(a, b)
 
 
-def test_array_inner_error_out_device_mismatch(test, device):
-    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    result = wp.empty(shape=(1,), dtype=wp.float32, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"out storage device should match values array$",
-    ):
-        wp.utils.array_inner(a, b, result)
-
-
 def test_array_inner_error_out_dtype_mismatch(test, device):
-    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    result = wp.empty(shape=(1,), dtype=wp.float64, device="cpu")
+    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    result = wp.empty(shape=(1,), dtype=wp.float64, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"out array should have type float32$",
@@ -271,9 +200,9 @@ def test_array_inner_error_out_dtype_mismatch(test, device):
 
 
 def test_array_inner_error_out_shape_mismatch(test, device):
-    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
-    result = wp.empty(shape=(2,), dtype=wp.float32, device="cpu")
+    a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device=device)
+    result = wp.empty(shape=(2,), dtype=wp.float32, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"out array should have shape \(1,\)$",
@@ -313,27 +242,17 @@ def test_array_cast(test, device):
     test.assertEqual(result.shape, (2,))
     assert_np_equal(result.numpy(), np.array((1.0, 2.0), dtype=float))
 
-    values = wp.array(((1, 2), (3, 4)), dtype=int, device="cpu")
-    result = wp.zeros((2, 2), dtype=int, device="cpu")
+    values = wp.array(((1, 2), (3, 4)), dtype=int, device=device)
+    result = wp.zeros((2, 2), dtype=int, device=device)
     wp.utils.array_cast(values, result)
     test.assertEqual(result.dtype, wp.int32)
     test.assertEqual(result.shape, (2, 2))
     assert_np_equal(result.numpy(), np.array(((1, 2), (3, 4)), dtype=int))
 
 
-def test_array_cast_error_devices_mismatch(test, device):
-    values = wp.array((1, 2, 3), dtype=int, device="cpu")
-    result = wp.empty(3, dtype=float, device="cuda:0")
-    with test.assertRaisesRegex(
-        RuntimeError,
-        r"Array storage devices do not match$",
-    ):
-        wp.utils.array_cast(values, result)
-
-
 def test_array_cast_error_unsupported_partial_cast(test, device):
-    values = wp.array(((1, 2), (3, 4)), dtype=int, device="cpu")
-    result = wp.zeros((2, 2), dtype=float, device="cpu")
+    values = wp.array(((1, 2), (3, 4)), dtype=int, device=device)
+    result = wp.zeros((2, 2), dtype=float, device=device)
     with test.assertRaisesRegex(
         RuntimeError,
         r"Partial cast is not supported for arrays with more than one dimension$",
@@ -341,112 +260,192 @@ def test_array_cast_error_unsupported_partial_cast(test, device):
         wp.utils.array_cast(values, result, count=1)
 
 
-def test_mesh_adjacency(test, device):
-    triangles = (
-        (0, 3, 1),
-        (0, 2, 3),
-    )
-    adj = wp.utils.MeshAdjacency(triangles, len(triangles))
-    expected_edges = {
-        (0, 3): (0, 3, 1, 2, 0, 1),
-        (1, 3): (3, 1, 0, -1, 0, -1),
-        (0, 1): (1, 0, 3, -1, 0, -1),
-        (0, 2): (0, 2, 3, -1, 1, -1),
-        (2, 3): (2, 3, 0, -1, 1, -1),
-    }
-    edges = {k: (e.v0, e.v1, e.o0, e.o1, e.f0, e.f1) for k, e in adj.edges.items()}
-    test.assertDictEqual(edges, expected_edges)
-
-
-def test_mesh_adjacency_error_manifold(test, device):
-    triangles = (
-        (0, 3, 1),
-        (0, 2, 3),
-        (3, 0, 1),
-    )
-
-    with contextlib.redirect_stdout(io.StringIO()) as f:
-        wp.utils.MeshAdjacency(triangles, len(triangles))
-
-    test.assertEqual(f.getvalue(), "Detected non-manifold edge\n")
-
-
-def test_scoped_timer(test, device):
-    with contextlib.redirect_stdout(io.StringIO()) as f:
-        with wp.ScopedTimer("hello"):
-            pass
-
-    test.assertRegex(f.getvalue(), r"^hello took \d+\.\d+ ms$")
-
-    with contextlib.redirect_stdout(io.StringIO()) as f:
-        with wp.ScopedTimer("hello", detailed=True):
-            pass
-
-    test.assertRegex(f.getvalue(), r"^         2 function calls in \d+\.\d+ seconds")
-    test.assertRegex(f.getvalue(), r"hello took \d+\.\d+ ms$")
-
-
-def register(parent):
-    devices = get_test_devices()
-
-    class TestUtils(parent):
-        pass
-
-    add_function_test(TestUtils, "test_warn", test_warn)
-    add_function_test(TestUtils, "test_transform_expand", test_transform_expand)
-    add_function_test(TestUtils, "test_array_scan", test_array_scan, devices=devices)
-    add_function_test(TestUtils, "test_array_scan_empty", test_array_scan_empty, devices=devices)
-    add_function_test(TestUtils, "test_array_scan_error_devices_mismatch", test_array_scan_error_devices_mismatch)
-    add_function_test(TestUtils, "test_array_scan_error_sizes_mismatch", test_array_scan_error_sizes_mismatch)
-    add_function_test(TestUtils, "test_array_scan_error_dtypes_mismatch", test_array_scan_error_dtypes_mismatch)
-    add_function_test(
-        TestUtils, "test_array_scan_error_unsupported_dtype", test_array_scan_error_unsupported_dtype, devices=devices
-    )
-    add_function_test(TestUtils, "test_radix_sort_pairs", test_radix_sort_pairs, devices=devices)
-    add_function_test(TestUtils, "test_radix_sort_pairs_empty", test_radix_sort_pairs_empty, devices=devices)
-    add_function_test(
-        TestUtils, "test_radix_sort_pairs_error_devices_mismatch", test_radix_sort_pairs_error_devices_mismatch
-    )
-    add_function_test(
-        TestUtils, "test_radix_sort_pairs_error_insufficient_storage", test_radix_sort_pairs_error_insufficient_storage
-    )
-    add_function_test(
-        TestUtils,
-        "test_radix_sort_pairs_error_unsupported_dtype",
-        test_radix_sort_pairs_error_unsupported_dtype,
-        devices=devices,
-    )
-    add_function_test(TestUtils, "test_array_sum", test_array_sum, devices=devices)
-    add_function_test(TestUtils, "test_array_sum_error_out_device_mismatch", test_array_sum_error_out_device_mismatch)
-    add_function_test(TestUtils, "test_array_sum_error_out_dtype_mismatch", test_array_sum_error_out_dtype_mismatch)
-    add_function_test(TestUtils, "test_array_sum_error_out_shape_mismatch", test_array_sum_error_out_shape_mismatch)
-    add_function_test(
-        TestUtils, "test_array_sum_error_unsupported_dtype", test_array_sum_error_unsupported_dtype, devices=devices
-    )
-    add_function_test(TestUtils, "test_array_inner", test_array_inner, devices=devices)
-    add_function_test(TestUtils, "test_array_inner_error_sizes_mismatch", test_array_inner_error_sizes_mismatch)
-    add_function_test(TestUtils, "test_array_inner_error_devices_mismatch", test_array_inner_error_devices_mismatch)
-    add_function_test(TestUtils, "test_array_inner_error_dtypes_mismatch", test_array_inner_error_dtypes_mismatch)
-    add_function_test(
-        TestUtils, "test_array_inner_error_out_device_mismatch", test_array_inner_error_out_device_mismatch
-    )
-    add_function_test(TestUtils, "test_array_inner_error_out_dtype_mismatch", test_array_inner_error_out_dtype_mismatch)
-    add_function_test(TestUtils, "test_array_inner_error_out_shape_mismatch", test_array_inner_error_out_shape_mismatch)
-    add_function_test(
-        TestUtils, "test_array_inner_error_unsupported_dtype", test_array_inner_error_unsupported_dtype, devices=devices
-    )
-    add_function_test(TestUtils, "test_array_cast", test_array_cast, devices=devices)
-    add_function_test(TestUtils, "test_array_cast_error_devices_mismatch", test_array_cast_error_devices_mismatch)
-    add_function_test(
-        TestUtils, "test_array_cast_error_unsupported_partial_cast", test_array_cast_error_unsupported_partial_cast
-    )
-    add_function_test(TestUtils, "test_mesh_adjacency", test_mesh_adjacency)
-    add_function_test(TestUtils, "test_mesh_adjacency_error_manifold", test_mesh_adjacency_error_manifold)
-    add_function_test(TestUtils, "test_scoped_timer", test_scoped_timer)
-    return TestUtils
+devices = get_test_devices()
+
+
+class TestUtils(unittest.TestCase):
+    def test_warn(self):
+        with contextlib.redirect_stdout(io.StringIO()) as f:
+            frame_info = inspect.getframeinfo(inspect.currentframe())
+            wp.utils.warn("hello, world!")
+
+        expected = '{}:{}: UserWarning: hello, world!\n  wp.utils.warn("hello, world!")\n'.format(
+            frame_info.filename,
+            frame_info.lineno + 1,
+        )
+        self.assertEqual(f.getvalue(), expected)
+
+    def test_transform_expand(self):
+        t = (1.0, 2.0, 3.0, 4.0, 3.0, 2.0, 1.0)
+        self.assertEqual(
+            wp.utils.transform_expand(t),
+            wp.transformf(p=(1.0, 2.0, 3.0), q=(4.0, 3.0, 2.0, 1.0)),
+        )
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_array_scan_error_devices_mismatch(self):
+        values = wp.zeros(123, dtype=int, device="cpu")
+        result = wp.zeros_like(values, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            wp.utils.array_scan(values, result, True)
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_radix_sort_pairs_error_devices_mismatch(self):
+        keys = wp.array((1, 2, 3), dtype=int, device="cpu")
+        values = wp.array((1, 2, 3), dtype=int, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            wp.utils.radix_sort_pairs(keys, values, 1)
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_array_inner_error_out_device_mismatch(self):
+        a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
+        b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
+        result = wp.empty(shape=(1,), dtype=wp.float32, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"out storage device should match values array$",
+        ):
+            wp.utils.array_inner(a, b, result)
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_array_sum_error_out_device_mismatch(self):
+        values = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
+        result = wp.empty(shape=(1,), dtype=wp.float32, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"out storage device should match values array$",
+        ):
+            wp.utils.array_sum(values, out=result)
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_array_inner_error_devices_mismatch(self):
+        a = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cpu")
+        b = wp.array((1.0, 2.0, 3.0), dtype=wp.float32, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            wp.utils.array_inner(a, b)
+
+    @unittest.skipUnless(wp.is_cuda_available(), "Requires CUDA")
+    def test_array_cast_error_devices_mismatch(self):
+        values = wp.array((1, 2, 3), dtype=int, device="cpu")
+        result = wp.empty(3, dtype=float, device="cuda:0")
+        with self.assertRaisesRegex(
+            RuntimeError,
+            r"Array storage devices do not match$",
+        ):
+            wp.utils.array_cast(values, result)
+
+    def test_mesh_adjacency(self):
+        triangles = (
+            (0, 3, 1),
+            (0, 2, 3),
+        )
+        adj = wp.utils.MeshAdjacency(triangles, len(triangles))
+        expected_edges = {
+            (0, 3): (0, 3, 1, 2, 0, 1),
+            (1, 3): (3, 1, 0, -1, 0, -1),
+            (0, 1): (1, 0, 3, -1, 0, -1),
+            (0, 2): (0, 2, 3, -1, 1, -1),
+            (2, 3): (2, 3, 0, -1, 1, -1),
+        }
+        edges = {k: (e.v0, e.v1, e.o0, e.o1, e.f0, e.f1) for k, e in adj.edges.items()}
+        self.assertDictEqual(edges, expected_edges)
+
+    def test_mesh_adjacency_error_manifold(self):
+        triangles = (
+            (0, 3, 1),
+            (0, 2, 3),
+            (3, 0, 1),
+        )
+
+        with contextlib.redirect_stdout(io.StringIO()) as f:
+            wp.utils.MeshAdjacency(triangles, len(triangles))
+
+        self.assertEqual(f.getvalue(), "Detected non-manifold edge\n")
+
+    def test_scoped_timer(self):
+        with contextlib.redirect_stdout(io.StringIO()) as f:
+            with wp.ScopedTimer("hello"):
+                pass
+
+        self.assertRegex(f.getvalue(), r"^hello took \d+\.\d+ ms$")
+
+        with contextlib.redirect_stdout(io.StringIO()) as f:
+            with wp.ScopedTimer("hello", detailed=True):
+                pass
+
+        self.assertRegex(f.getvalue(), r"^         2 function calls in \d+\.\d+ seconds")
+        self.assertRegex(f.getvalue(), r"hello took \d+\.\d+ ms$")
+
+
+add_function_test(TestUtils, "test_array_scan", test_array_scan, devices=devices)
+add_function_test(TestUtils, "test_array_scan_empty", test_array_scan_empty, devices=devices)
+add_function_test(
+    TestUtils, "test_array_scan_error_sizes_mismatch", test_array_scan_error_sizes_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_scan_error_dtypes_mismatch", test_array_scan_error_dtypes_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_scan_error_unsupported_dtype", test_array_scan_error_unsupported_dtype, devices=devices
+)
+add_function_test(TestUtils, "test_radix_sort_pairs", test_radix_sort_pairs, devices=devices)
+add_function_test(TestUtils, "test_radix_sort_pairs_empty", test_radix_sort_pairs, devices=devices)
+add_function_test(
+    TestUtils,
+    "test_radix_sort_pairs_error_insufficient_storage",
+    test_radix_sort_pairs_error_insufficient_storage,
+    devices=devices,
+)
+add_function_test(
+    TestUtils,
+    "test_radix_sort_pairs_error_unsupported_dtype",
+    test_radix_sort_pairs_error_unsupported_dtype,
+    devices=devices,
+)
+add_function_test(TestUtils, "test_array_sum", test_array_sum, devices=devices)
+add_function_test(
+    TestUtils, "test_array_sum_error_out_dtype_mismatch", test_array_sum_error_out_dtype_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_sum_error_out_shape_mismatch", test_array_sum_error_out_shape_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_sum_error_unsupported_dtype", test_array_sum_error_unsupported_dtype, devices=devices
+)
+add_function_test(TestUtils, "test_array_inner", test_array_inner, devices=devices)
+add_function_test(
+    TestUtils, "test_array_inner_error_sizes_mismatch", test_array_inner_error_sizes_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_inner_error_dtypes_mismatch", test_array_inner_error_dtypes_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_inner_error_out_dtype_mismatch", test_array_inner_error_out_dtype_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_inner_error_out_shape_mismatch", test_array_inner_error_out_shape_mismatch, devices=devices
+)
+add_function_test(
+    TestUtils, "test_array_inner_error_unsupported_dtype", test_array_inner_error_unsupported_dtype, devices=devices
+)
+add_function_test(TestUtils, "test_array_cast", test_array_cast, devices=devices)
+add_function_test(
+    TestUtils,
+    "test_array_cast_error_unsupported_partial_cast",
+    test_array_cast_error_unsupported_partial_cast,
+    devices=devices,
+)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_vec.py b/warp/tests/test_vec.py
index 88f5d9d2d..dd6ba4e2b 100644
--- a/warp/tests/test_vec.py
+++ b/warp/tests/test_vec.py
@@ -8,8 +8,9 @@
 import unittest
 
 import numpy as np
+
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -29,12 +30,8 @@
     np.ubyte,
 ]
 
-np_int_types = np_signed_int_types + np_unsigned_int_types
-
 np_float_types = [np.float16, np.float32, np.float64]
 
-np_scalar_types = np_int_types + np_float_types
-
 
 def randvals(rng, shape, dtype):
     if dtype in np_float_types:
@@ -54,546 +51,6 @@ def getkernel(func, suffix=""):
     return kernel_cache[key]
 
 
-def get_select_kernel(dtype):
-    def output_select_kernel_fn(
-        input: wp.array(dtype=dtype),
-        index: int,
-        out: wp.array(dtype=dtype),
-    ):
-        out[0] = input[index]
-
-    return getkernel(output_select_kernel_fn, suffix=dtype.__name__)
-
-
-def get_select_kernel2(dtype):
-    def output_select_kernel2_fn(
-        input: wp.array(dtype=dtype, ndim=2),
-        index0: int,
-        index1: int,
-        out: wp.array(dtype=dtype),
-    ):
-        out[0] = input[index0, index1]
-
-    return getkernel(output_select_kernel2_fn, suffix=dtype.__name__)
-
-
-def test_arrays(test, device, dtype):
-    rng = np.random.default_rng(123)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    v2_np = randvals(rng, (10, 2), dtype)
-    v3_np = randvals(rng, (10, 3), dtype)
-    v4_np = randvals(rng, (10, 4), dtype)
-    v5_np = randvals(rng, (10, 5), dtype)
-
-    v2 = wp.array(v2_np, dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(v3_np, dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(v4_np, dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(v5_np, dtype=vec5, requires_grad=True, device=device)
-
-    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
-    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
-    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
-    assert_np_equal(v5.numpy(), v5_np, tol=1.0e-6)
-
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-
-    v2 = wp.array(v2_np, dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(v3_np, dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(v4_np, dtype=vec4, requires_grad=True, device=device)
-
-    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
-    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
-    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
-
-
-def test_components(test, device, dtype):
-    # test accessing vector components from Python - this is especially important
-    # for float16, which requires special handling internally
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-
-    v = vec3(1, 2, 3)
-
-    # test __getitem__ for individual components
-    test.assertEqual(v[0], 1)
-    test.assertEqual(v[1], 2)
-    test.assertEqual(v[2], 3)
-
-    # test __getitem__ for slices
-    s = v[:]
-    test.assertEqual(s[0], 1)
-    test.assertEqual(s[1], 2)
-    test.assertEqual(s[2], 3)
-
-    s = v[1:]
-    test.assertEqual(s[0], 2)
-    test.assertEqual(s[1], 3)
-
-    s = v[:2]
-    test.assertEqual(s[0], 1)
-    test.assertEqual(s[1], 2)
-
-    s = v[::2]
-    test.assertEqual(s[0], 1)
-    test.assertEqual(s[1], 3)
-
-    # test __setitem__ for individual components
-    v[0] = 4
-    v[1] = 5
-    v[2] = 6
-    test.assertEqual(v[0], 4)
-    test.assertEqual(v[1], 5)
-    test.assertEqual(v[2], 6)
-
-    # test __setitem__ for slices
-    v[:] = [7, 8, 9]
-    test.assertEqual(v[0], 7)
-    test.assertEqual(v[1], 8)
-    test.assertEqual(v[2], 9)
-
-    v[1:] = [10, 11]
-    test.assertEqual(v[0], 7)
-    test.assertEqual(v[1], 10)
-    test.assertEqual(v[2], 11)
-
-    v[:2] = [12, 13]
-    test.assertEqual(v[0], 12)
-    test.assertEqual(v[1], 13)
-    test.assertEqual(v[2], 11)
-
-    v[::2] = [14, 15]
-    test.assertEqual(v[0], 14)
-    test.assertEqual(v[1], 13)
-    test.assertEqual(v[2], 15)
-
-
-def test_py_arithmetic_ops(test, device, dtype):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    def make_vec(*args):
-        if wptype in wp.types.int_types:
-            # Cast to the correct integer type to simulate wrapping.
-            return tuple(wptype._type_(x).value for x in args)
-
-        return args
-
-    vec_cls = wp.vec(3, wptype)
-
-    v = vec_cls(1, -2, 3)
-    test.assertSequenceEqual(+v, make_vec(1, -2, 3))
-    test.assertSequenceEqual(-v, make_vec(-1, 2, -3))
-    test.assertSequenceEqual(v + vec_cls(5, 5, 5), make_vec(6, 3, 8))
-    test.assertSequenceEqual(v - vec_cls(5, 5, 5), make_vec(-4, -7, -2))
-
-    v = vec_cls(2, 4, 6)
-    test.assertSequenceEqual(v * wptype(2), make_vec(4, 8, 12))
-    test.assertSequenceEqual(wptype(2) * v, make_vec(4, 8, 12))
-    test.assertSequenceEqual(v / wptype(2), make_vec(1, 2, 3))
-    test.assertSequenceEqual(wptype(24) / v, make_vec(12, 6, 4))
-
-
-def test_anon_type_instance(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    def check_scalar_init(
-        input: wp.array(dtype=wptype),
-        output: wp.array(dtype=wptype),
-    ):
-        v2result = wp.vector(input[0], length=2)
-        v3result = wp.vector(input[1], length=3)
-        v4result = wp.vector(input[2], length=4)
-        v5result = wp.vector(input[3], length=5)
-
-        idx = 0
-        for i in range(2):
-            output[idx] = wptype(2) * v2result[i]
-            idx = idx + 1
-        for i in range(3):
-            output[idx] = wptype(2) * v3result[i]
-            idx = idx + 1
-        for i in range(4):
-            output[idx] = wptype(2) * v4result[i]
-            idx = idx + 1
-        for i in range(5):
-            output[idx] = wptype(2) * v5result[i]
-            idx = idx + 1
-
-    def check_component_init(
-        input: wp.array(dtype=wptype),
-        output: wp.array(dtype=wptype),
-    ):
-        v2result = wp.vector(input[0], input[1])
-        v3result = wp.vector(input[2], input[3], input[4])
-        v4result = wp.vector(input[5], input[6], input[7], input[8])
-        v5result = wp.vector(input[9], input[10], input[11], input[12], input[13])
-
-        idx = 0
-        for i in range(2):
-            output[idx] = wptype(2) * v2result[i]
-            idx = idx + 1
-        for i in range(3):
-            output[idx] = wptype(2) * v3result[i]
-            idx = idx + 1
-        for i in range(4):
-            output[idx] = wptype(2) * v4result[i]
-            idx = idx + 1
-        for i in range(5):
-            output[idx] = wptype(2) * v5result[i]
-            idx = idx + 1
-
-    scalar_kernel = getkernel(check_scalar_init, suffix=dtype.__name__)
-    component_kernel = getkernel(check_component_init, suffix=dtype.__name__)
-    output_select_kernel = get_select_kernel(wptype)
-
-    if register_kernels:
-        return
-
-    input = wp.array(randvals(rng, [4], dtype), requires_grad=True, device=device)
-    output = wp.zeros(2 + 3 + 4 + 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-
-    assert_np_equal(output.numpy()[:2], 2 * np.array([input.numpy()[0]] * 2), tol=1.0e-6)
-    assert_np_equal(output.numpy()[2:5], 2 * np.array([input.numpy()[1]] * 3), tol=1.0e-6)
-    assert_np_equal(output.numpy()[5:9], 2 * np.array([input.numpy()[2]] * 4), tol=1.0e-6)
-    assert_np_equal(output.numpy()[9:], 2 * np.array([input.numpy()[3]] * 5), tol=1.0e-6)
-
-    if dtype in np_float_types:
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for i in range(len(output)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
-
-            tape.backward(loss=out)
-            expected = np.zeros_like(input.numpy())
-            if i < 2:
-                expected[0] = 2
-            elif i < 5:
-                expected[1] = 2
-            elif i < 9:
-                expected[2] = 2
-            else:
-                expected[3] = 2
-
-            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
-
-            tape.reset()
-            tape.zero()
-
-    input = wp.array(randvals(rng, [2 + 3 + 4 + 5], dtype), requires_grad=True, device=device)
-    output = wp.zeros(2 + 3 + 4 + 5, dtype=wptype, requires_grad=True, device=device)
-
-    wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-
-    assert_np_equal(output.numpy(), 2 * input.numpy(), tol=1.0e-6)
-
-    if dtype in np_float_types:
-        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-        for i in range(len(output)):
-            tape = wp.Tape()
-            with tape:
-                wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
-                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
-
-            tape.backward(loss=out)
-            expected = np.zeros_like(input.numpy())
-            expected[i] = 2
-
-            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
-
-            tape.reset()
-            tape.zero()
-
-
-def test_constants(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    cv2 = wp.constant(vec2(1, 2))
-    cv3 = wp.constant(vec3(1, 2, 3))
-    cv4 = wp.constant(vec4(1, 2, 3, 4))
-    cv5 = wp.constant(vec5(1, 2, 3, 4, 5))
-
-    def check_vector_constants():
-        wp.expect_eq(cv2, vec2(wptype(1), wptype(2)))
-        wp.expect_eq(cv3, vec3(wptype(1), wptype(2), wptype(3)))
-        wp.expect_eq(cv4, vec4(wptype(1), wptype(2), wptype(3), wptype(4)))
-        wp.expect_eq(cv5, vec5(wptype(1), wptype(2), wptype(3), wptype(4), wptype(5)))
-
-    kernel = getkernel(check_vector_constants, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    wp.launch(kernel, dim=1, inputs=[])
-
-
-def test_constructors(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_scalar_constructor(
-        input: wp.array(dtype=wptype),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = vec2(input[0])
-        v3result = vec3(input[0])
-        v4result = vec4(input[0])
-        v5result = vec5(input[0])
-
-        v2[0] = v2result
-        v3[0] = v3result
-        v4[0] = v4result
-        v5[0] = v5result
-
-        # multiply outputs by 2 so we've got something to backpropagate
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    def check_vector_constructors(
-        input: wp.array(dtype=wptype),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = vec2(input[0], input[1])
-        v3result = vec3(input[2], input[3], input[4])
-        v4result = vec4(input[5], input[6], input[7], input[8])
-        v5result = vec5(input[9], input[10], input[11], input[12], input[13])
-
-        v2[0] = v2result
-        v3[0] = v3result
-        v4[0] = v4result
-        v5[0] = v5result
-
-        # multiply the output by 2 so we've got something to backpropagate:
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    vec_kernel = getkernel(check_vector_constructors, suffix=dtype.__name__)
-    kernel = getkernel(check_scalar_constructor, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    input = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    v2 = wp.zeros(1, dtype=vec2, device=device)
-    v3 = wp.zeros(1, dtype=vec3, device=device)
-    v4 = wp.zeros(1, dtype=vec4, device=device)
-    v5 = wp.zeros(1, dtype=vec5, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[input],
-            outputs=[v2, v3, v4, v5, v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    if dtype in np_float_types:
-        for l in [v20, v21]:
-            tape.backward(loss=l)
-            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
-            tape.zero()
-
-        for l in [v30, v31, v32]:
-            tape.backward(loss=l)
-            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
-            tape.zero()
-
-        for l in [v40, v41, v42, v43]:
-            tape.backward(loss=l)
-            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
-            tape.zero()
-
-        for l in [v50, v51, v52, v53, v54]:
-            tape.backward(loss=l)
-            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
-            tape.zero()
-
-    val = input.numpy()[0]
-    assert_np_equal(v2.numpy()[0], np.array([val, val]), tol=1.0e-6)
-    assert_np_equal(v3.numpy()[0], np.array([val, val, val]), tol=1.0e-6)
-    assert_np_equal(v4.numpy()[0], np.array([val, val, val, val]), tol=1.0e-6)
-    assert_np_equal(v5.numpy()[0], np.array([val, val, val, val, val]), tol=1.0e-6)
-
-    assert_np_equal(v20.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v21.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v30.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v31.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v32.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v40.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v41.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v42.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v43.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v50.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v51.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v52.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v53.numpy()[0], 2 * val, tol=1.0e-6)
-    assert_np_equal(v54.numpy()[0], 2 * val, tol=1.0e-6)
-
-    input = wp.array(randvals(rng, [14], dtype), requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            vec_kernel,
-            dim=1,
-            inputs=[input],
-            outputs=[v2, v3, v4, v5, v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
-            tape.backward(loss=l)
-            grad = tape.gradients[input].numpy()
-            expected_grad = np.zeros_like(grad)
-            expected_grad[i] = 2
-            assert_np_equal(grad, expected_grad, tol=tol)
-            tape.zero()
-
-    assert_np_equal(v2.numpy()[0, 0], input.numpy()[0], tol=tol)
-    assert_np_equal(v2.numpy()[0, 1], input.numpy()[1], tol=tol)
-    assert_np_equal(v3.numpy()[0, 0], input.numpy()[2], tol=tol)
-    assert_np_equal(v3.numpy()[0, 1], input.numpy()[3], tol=tol)
-    assert_np_equal(v3.numpy()[0, 2], input.numpy()[4], tol=tol)
-    assert_np_equal(v4.numpy()[0, 0], input.numpy()[5], tol=tol)
-    assert_np_equal(v4.numpy()[0, 1], input.numpy()[6], tol=tol)
-    assert_np_equal(v4.numpy()[0, 2], input.numpy()[7], tol=tol)
-    assert_np_equal(v4.numpy()[0, 3], input.numpy()[8], tol=tol)
-    assert_np_equal(v5.numpy()[0, 0], input.numpy()[9], tol=tol)
-    assert_np_equal(v5.numpy()[0, 1], input.numpy()[10], tol=tol)
-    assert_np_equal(v5.numpy()[0, 2], input.numpy()[11], tol=tol)
-    assert_np_equal(v5.numpy()[0, 3], input.numpy()[12], tol=tol)
-    assert_np_equal(v5.numpy()[0, 4], input.numpy()[13], tol=tol)
-
-    assert_np_equal(v20.numpy()[0], 2 * input.numpy()[0], tol=tol)
-    assert_np_equal(v21.numpy()[0], 2 * input.numpy()[1], tol=tol)
-    assert_np_equal(v30.numpy()[0], 2 * input.numpy()[2], tol=tol)
-    assert_np_equal(v31.numpy()[0], 2 * input.numpy()[3], tol=tol)
-    assert_np_equal(v32.numpy()[0], 2 * input.numpy()[4], tol=tol)
-    assert_np_equal(v40.numpy()[0], 2 * input.numpy()[5], tol=tol)
-    assert_np_equal(v41.numpy()[0], 2 * input.numpy()[6], tol=tol)
-    assert_np_equal(v42.numpy()[0], 2 * input.numpy()[7], tol=tol)
-    assert_np_equal(v43.numpy()[0], 2 * input.numpy()[8], tol=tol)
-    assert_np_equal(v50.numpy()[0], 2 * input.numpy()[9], tol=tol)
-    assert_np_equal(v51.numpy()[0], 2 * input.numpy()[10], tol=tol)
-    assert_np_equal(v52.numpy()[0], 2 * input.numpy()[11], tol=tol)
-    assert_np_equal(v53.numpy()[0], 2 * input.numpy()[12], tol=tol)
-    assert_np_equal(v54.numpy()[0], 2 * input.numpy()[13], tol=tol)
-
-
 def test_anon_constructor_error_dtype_keyword_missing(test, device):
     @wp.kernel
     def kernel():
@@ -726,1102 +183,30 @@ def kernel():
         r"All numeric arguments to vec\(\) constructor should have the same "
         r"type, expected 2 arg_types of type <class 'warp.types.float32'>, "
         r"received <class 'warp.types.float32'>,<class 'warp.types.int32'>$",
-    ):
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[],
-            device=device,
-        )
-
-
-def test_tpl_ops_with_anon(test, device):
-    vec3i = wp.vec(3, dtype=int)
-
-    v = wp.vec3i(1, 2, 3)
-    v += vec3i(2, 3, 4)
-    v -= vec3i(3, 4, 5)
-    test.assertSequenceEqual(v, (0, 1, 2))
-
-    v = vec3i(1, 2, 3)
-    v += wp.vec3i(2, 3, 4)
-    v -= wp.vec3i(3, 4, 5)
-    test.assertSequenceEqual(v, (0, 1, 2))
-
-
-def test_indexing(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_indexing(
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        # multiply outputs by 2 so we've got something to backpropagate:
-        v20[0] = wptype(2) * v2[0][0]
-        v21[0] = wptype(2) * v2[0][1]
-
-        v30[0] = wptype(2) * v3[0][0]
-        v31[0] = wptype(2) * v3[0][1]
-        v32[0] = wptype(2) * v3[0][2]
-
-        v40[0] = wptype(2) * v4[0][0]
-        v41[0] = wptype(2) * v4[0][1]
-        v42[0] = wptype(2) * v4[0][2]
-        v43[0] = wptype(2) * v4[0][3]
-
-        v50[0] = wptype(2) * v5[0][0]
-        v51[0] = wptype(2) * v5[0][1]
-        v52[0] = wptype(2) * v5[0][2]
-        v53[0] = wptype(2) * v5[0][3]
-        v54[0] = wptype(2) * v5[0][4]
-
-    kernel = getkernel(check_indexing, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[v2, v3, v4, v5],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
-            tape.backward(loss=l)
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
-            expected_grads = np.zeros_like(allgrads)
-            expected_grads[i] = 2
-            assert_np_equal(allgrads, expected_grads, tol=tol)
-            tape.zero()
-
-    assert_np_equal(v20.numpy()[0], 2.0 * v2.numpy()[0, 0], tol=tol)
-    assert_np_equal(v21.numpy()[0], 2.0 * v2.numpy()[0, 1], tol=tol)
-    assert_np_equal(v30.numpy()[0], 2.0 * v3.numpy()[0, 0], tol=tol)
-    assert_np_equal(v31.numpy()[0], 2.0 * v3.numpy()[0, 1], tol=tol)
-    assert_np_equal(v32.numpy()[0], 2.0 * v3.numpy()[0, 2], tol=tol)
-    assert_np_equal(v40.numpy()[0], 2.0 * v4.numpy()[0, 0], tol=tol)
-    assert_np_equal(v41.numpy()[0], 2.0 * v4.numpy()[0, 1], tol=tol)
-    assert_np_equal(v42.numpy()[0], 2.0 * v4.numpy()[0, 2], tol=tol)
-    assert_np_equal(v43.numpy()[0], 2.0 * v4.numpy()[0, 3], tol=tol)
-    assert_np_equal(v50.numpy()[0], 2.0 * v5.numpy()[0, 0], tol=tol)
-    assert_np_equal(v51.numpy()[0], 2.0 * v5.numpy()[0, 1], tol=tol)
-    assert_np_equal(v52.numpy()[0], 2.0 * v5.numpy()[0, 2], tol=tol)
-    assert_np_equal(v53.numpy()[0], 2.0 * v5.numpy()[0, 3], tol=tol)
-    assert_np_equal(v54.numpy()[0], 2.0 * v5.numpy()[0, 4], tol=tol)
-
-
-def test_equality(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_equality(
-        v20: wp.array(dtype=vec2),
-        v21: wp.array(dtype=vec2),
-        v22: wp.array(dtype=vec2),
-        v30: wp.array(dtype=vec3),
-        v31: wp.array(dtype=vec3),
-        v32: wp.array(dtype=vec3),
-        v33: wp.array(dtype=vec3),
-        v40: wp.array(dtype=vec4),
-        v41: wp.array(dtype=vec4),
-        v42: wp.array(dtype=vec4),
-        v43: wp.array(dtype=vec4),
-        v44: wp.array(dtype=vec4),
-        v50: wp.array(dtype=vec5),
-        v51: wp.array(dtype=vec5),
-        v52: wp.array(dtype=vec5),
-        v53: wp.array(dtype=vec5),
-        v54: wp.array(dtype=vec5),
-        v55: wp.array(dtype=vec5),
-    ):
-        wp.expect_eq(v20[0], v20[0])
-        wp.expect_neq(v21[0], v20[0])
-        wp.expect_neq(v22[0], v20[0])
-
-        wp.expect_eq(v30[0], v30[0])
-        wp.expect_neq(v31[0], v30[0])
-        wp.expect_neq(v32[0], v30[0])
-        wp.expect_neq(v33[0], v30[0])
-
-        wp.expect_eq(v40[0], v40[0])
-        wp.expect_neq(v41[0], v40[0])
-        wp.expect_neq(v42[0], v40[0])
-        wp.expect_neq(v43[0], v40[0])
-        wp.expect_neq(v44[0], v40[0])
-
-        wp.expect_eq(v50[0], v50[0])
-        wp.expect_neq(v51[0], v50[0])
-        wp.expect_neq(v52[0], v50[0])
-        wp.expect_neq(v53[0], v50[0])
-        wp.expect_neq(v54[0], v50[0])
-        wp.expect_neq(v55[0], v50[0])
-
-    kernel = getkernel(check_equality, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v20 = wp.array([1.0, 2.0], dtype=vec2, requires_grad=True, device=device)
-    v21 = wp.array([1.0, 3.0], dtype=vec2, requires_grad=True, device=device)
-    v22 = wp.array([3.0, 2.0], dtype=vec2, requires_grad=True, device=device)
-
-    v30 = wp.array([1.0, 2.0, 3.0], dtype=vec3, requires_grad=True, device=device)
-    v31 = wp.array([-1.0, 2.0, 3.0], dtype=vec3, requires_grad=True, device=device)
-    v32 = wp.array([1.0, -2.0, 3.0], dtype=vec3, requires_grad=True, device=device)
-    v33 = wp.array([1.0, 2.0, -3.0], dtype=vec3, requires_grad=True, device=device)
-
-    v40 = wp.array([1.0, 2.0, 3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
-    v41 = wp.array([-1.0, 2.0, 3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
-    v42 = wp.array([1.0, -2.0, 3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
-    v43 = wp.array([1.0, 2.0, -3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
-    v44 = wp.array([1.0, 2.0, 3.0, -4.0], dtype=vec4, requires_grad=True, device=device)
-
-    v50 = wp.array([1.0, 2.0, 3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
-    v51 = wp.array([-1.0, 2.0, 3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
-    v52 = wp.array([1.0, -2.0, 3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
-    v53 = wp.array([1.0, 2.0, -3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
-    v54 = wp.array([1.0, 2.0, 3.0, -4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
-    v55 = wp.array([1.0, 2.0, 3.0, 4.0, -5.0], dtype=vec5, requires_grad=True, device=device)
-    wp.launch(
-        kernel,
-        dim=1,
-        inputs=[
-            v20,
-            v21,
-            v22,
-            v30,
-            v31,
-            v32,
-            v33,
-            v40,
-            v41,
-            v42,
-            v43,
-            v44,
-            v50,
-            v51,
-            v52,
-            v53,
-            v54,
-            v55,
-        ],
-        outputs=[],
-        device=device,
-    )
-
-
-def test_negation(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_negation(
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v2out: wp.array(dtype=vec2),
-        v3out: wp.array(dtype=vec3),
-        v4out: wp.array(dtype=vec4),
-        v5out: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = -v2[0]
-        v3result = -v3[0]
-        v4result = -v4[0]
-        v5result = -v5[0]
-
-        v2out[0] = v2result
-        v3out[0] = v3result
-        v4out[0] = v4result
-        v5out[0] = v5result
-
-        # multiply these outputs by 2 so we've got something to backpropagate:
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    kernel = getkernel(check_negation, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5_np = randvals(rng, (1, 5), dtype)
-    v5 = wp.array(v5_np, dtype=vec5, requires_grad=True, device=device)
-
-    v2out = wp.zeros(1, dtype=vec2, device=device)
-    v3out = wp.zeros(1, dtype=vec3, device=device)
-    v4out = wp.zeros(1, dtype=vec4, device=device)
-    v5out = wp.zeros(1, dtype=vec5, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[v2, v3, v4, v5],
-            outputs=[v2out, v3out, v4out, v5out, v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
-            tape.backward(loss=l)
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
-            expected_grads = np.zeros_like(allgrads)
-            expected_grads[i] = -2
-            assert_np_equal(allgrads, expected_grads, tol=tol)
-            tape.zero()
-
-    assert_np_equal(v2out.numpy()[0], -v2.numpy()[0], tol=tol)
-    assert_np_equal(v3out.numpy()[0], -v3.numpy()[0], tol=tol)
-    assert_np_equal(v4out.numpy()[0], -v4.numpy()[0], tol=tol)
-    assert_np_equal(v5out.numpy()[0], -v5.numpy()[0], tol=tol)
-
-
-def test_scalar_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_mul(
-        s: wp.array(dtype=wptype),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = s[0] * v2[0]
-        v3result = s[0] * v3[0]
-        v4result = s[0] * v4[0]
-        v5result = s[0] * v5[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    kernel = getkernel(check_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    assert_np_equal(v20.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 0], tol=tol)
-    assert_np_equal(v21.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 1], tol=tol)
-
-    assert_np_equal(v30.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v31.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v32.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 2], tol=10 * tol)
-
-    assert_np_equal(v40.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v41.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v42.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 2], tol=10 * tol)
-    assert_np_equal(v43.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 3], tol=10 * tol)
-
-    assert_np_equal(v50.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v51.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v52.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 2], tol=10 * tol)
-    assert_np_equal(v53.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 3], tol=10 * tol)
-    assert_np_equal(v54.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 4], tol=10 * tol)
-
-    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43]):
-            tape.backward(loss=l)
-            sgrad = tape.gradients[s].numpy()[0]
-            assert_np_equal(sgrad, 2 * incmps[i], tol=10 * tol)
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4]])
-            expected_grads = np.zeros_like(allgrads)
-            expected_grads[i] = s.numpy()[0] * 2
-            assert_np_equal(allgrads, expected_grads, tol=10 * tol)
-            tape.zero()
-
-
-def test_scalar_multiplication_rightmul(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_rightmul(
-        s: wp.array(dtype=wptype),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = v2[0] * s[0]
-        v3result = v3[0] * s[0]
-        v4result = v4[0] * s[0]
-        v5result = v5[0] * s[0]
-
-        # multiply outputs by 2 so we've got something to backpropagate:
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    kernel = getkernel(check_rightmul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    assert_np_equal(v20.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 0], tol=tol)
-    assert_np_equal(v21.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 1], tol=tol)
-
-    assert_np_equal(v30.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v31.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v32.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 2], tol=10 * tol)
-
-    assert_np_equal(v40.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v41.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v42.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 2], tol=10 * tol)
-    assert_np_equal(v43.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 3], tol=10 * tol)
-
-    assert_np_equal(v50.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v51.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v52.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 2], tol=10 * tol)
-    assert_np_equal(v53.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 3], tol=10 * tol)
-    assert_np_equal(v54.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 4], tol=10 * tol)
-
-    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43]):
-            tape.backward(loss=l)
-            sgrad = tape.gradients[s].numpy()[0]
-            assert_np_equal(sgrad, 2 * incmps[i], tol=10 * tol)
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4]])
-            expected_grads = np.zeros_like(allgrads)
-            expected_grads[i] = s.numpy()[0] * 2
-            assert_np_equal(allgrads, expected_grads, tol=10 * tol)
-            tape.zero()
-
-
-def test_cw_multiplication(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_cw_mul(
-        s2: wp.array(dtype=vec2),
-        s3: wp.array(dtype=vec3),
-        s4: wp.array(dtype=vec4),
-        s5: wp.array(dtype=vec5),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = wp.cw_mul(s2[0], v2[0])
-        v3result = wp.cw_mul(s3[0], v3[0])
-        v4result = wp.cw_mul(s4[0], v4[0])
-        v5result = wp.cw_mul(s5[0], v5[0])
-
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    kernel = getkernel(check_cw_mul, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s2,
-                s3,
-                s4,
-                s5,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    assert_np_equal(v20.numpy()[0], 2 * s2.numpy()[0, 0] * v2.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v21.numpy()[0], 2 * s2.numpy()[0, 1] * v2.numpy()[0, 1], tol=10 * tol)
-
-    assert_np_equal(v30.numpy()[0], 2 * s3.numpy()[0, 0] * v3.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v31.numpy()[0], 2 * s3.numpy()[0, 1] * v3.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v32.numpy()[0], 2 * s3.numpy()[0, 2] * v3.numpy()[0, 2], tol=10 * tol)
-
-    assert_np_equal(v40.numpy()[0], 2 * s4.numpy()[0, 0] * v4.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v41.numpy()[0], 2 * s4.numpy()[0, 1] * v4.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v42.numpy()[0], 2 * s4.numpy()[0, 2] * v4.numpy()[0, 2], tol=10 * tol)
-    assert_np_equal(v43.numpy()[0], 2 * s4.numpy()[0, 3] * v4.numpy()[0, 3], tol=10 * tol)
-
-    assert_np_equal(v50.numpy()[0], 2 * s5.numpy()[0, 0] * v5.numpy()[0, 0], tol=10 * tol)
-    assert_np_equal(v51.numpy()[0], 2 * s5.numpy()[0, 1] * v5.numpy()[0, 1], tol=10 * tol)
-    assert_np_equal(v52.numpy()[0], 2 * s5.numpy()[0, 2] * v5.numpy()[0, 2], tol=10 * tol)
-    assert_np_equal(v53.numpy()[0], 2 * s5.numpy()[0, 3] * v5.numpy()[0, 3], tol=10 * tol)
-    assert_np_equal(v54.numpy()[0], 2 * s5.numpy()[0, 4] * v5.numpy()[0, 4], tol=10 * tol)
-
-    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
-    scmps = np.concatenate([v.numpy()[0] for v in [s2, s3, s4, s5]])
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
-            tape.backward(loss=l)
-            sgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [s2, s3, s4, s5]])
-            expected_grads = np.zeros_like(sgrads)
-            expected_grads[i] = incmps[i] * 2
-            assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
-            expected_grads = np.zeros_like(allgrads)
-            expected_grads[i] = scmps[i] * 2
-            assert_np_equal(allgrads, expected_grads, tol=10 * tol)
-
-            tape.zero()
-
-
-def test_scalar_division(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 5.0e-3,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_div(
-        s: wp.array(dtype=wptype),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = v2[0] / s[0]
-        v3result = v3[0] / s[0]
-        v4result = v4[0] / s[0]
-        v5result = v5[0] / s[0]
-
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    kernel = getkernel(check_div, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    if dtype in np_int_types:
-        assert_np_equal(v20.numpy()[0], 2 * (v2.numpy()[0, 0] // (s.numpy()[0])), tol=tol)
-        assert_np_equal(v21.numpy()[0], 2 * (v2.numpy()[0, 1] // (s.numpy()[0])), tol=tol)
-
-        assert_np_equal(v30.numpy()[0], 2 * (v3.numpy()[0, 0] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v31.numpy()[0], 2 * (v3.numpy()[0, 1] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v32.numpy()[0], 2 * (v3.numpy()[0, 2] // (s.numpy()[0])), tol=10 * tol)
-
-        assert_np_equal(v40.numpy()[0], 2 * (v4.numpy()[0, 0] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v41.numpy()[0], 2 * (v4.numpy()[0, 1] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v42.numpy()[0], 2 * (v4.numpy()[0, 2] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v43.numpy()[0], 2 * (v4.numpy()[0, 3] // (s.numpy()[0])), tol=10 * tol)
-
-        assert_np_equal(v50.numpy()[0], 2 * (v5.numpy()[0, 0] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v51.numpy()[0], 2 * (v5.numpy()[0, 1] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v52.numpy()[0], 2 * (v5.numpy()[0, 2] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v53.numpy()[0], 2 * (v5.numpy()[0, 3] // (s.numpy()[0])), tol=10 * tol)
-        assert_np_equal(v54.numpy()[0], 2 * (v5.numpy()[0, 4] // (s.numpy()[0])), tol=10 * tol)
-
-    else:
-        assert_np_equal(v20.numpy()[0], 2 * v2.numpy()[0, 0] / (s.numpy()[0]), tol=tol)
-        assert_np_equal(v21.numpy()[0], 2 * v2.numpy()[0, 1] / (s.numpy()[0]), tol=tol)
-
-        assert_np_equal(v30.numpy()[0], 2 * v3.numpy()[0, 0] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v31.numpy()[0], 2 * v3.numpy()[0, 1] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v32.numpy()[0], 2 * v3.numpy()[0, 2] / (s.numpy()[0]), tol=10 * tol)
-
-        assert_np_equal(v40.numpy()[0], 2 * v4.numpy()[0, 0] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v41.numpy()[0], 2 * v4.numpy()[0, 1] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v42.numpy()[0], 2 * v4.numpy()[0, 2] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v43.numpy()[0], 2 * v4.numpy()[0, 3] / (s.numpy()[0]), tol=10 * tol)
-
-        assert_np_equal(v50.numpy()[0], 2 * v5.numpy()[0, 0] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v51.numpy()[0], 2 * v5.numpy()[0, 1] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v52.numpy()[0], 2 * v5.numpy()[0, 2] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v53.numpy()[0], 2 * v5.numpy()[0, 3] / (s.numpy()[0]), tol=10 * tol)
-        assert_np_equal(v54.numpy()[0], 2 * v5.numpy()[0, 4] / (s.numpy()[0]), tol=10 * tol)
-
-    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
-
-    if dtype in np_float_types:
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
-            tape.backward(loss=l)
-            sgrad = tape.gradients[s].numpy()[0]
-
-            # d/ds v/s = -v/s^2
-            assert_np_equal(sgrad, -2 * incmps[i] / (s.numpy()[0] * s.numpy()[0]), tol=10 * tol)
-
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
-            expected_grads = np.zeros_like(allgrads)
-            expected_grads[i] = 2 / s.numpy()[0]
-
-            # d/dv v/s = 1/s
-            assert_np_equal(allgrads, expected_grads, tol=tol)
-            tape.zero()
-
-
-def test_cw_division(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_cw_div(
-        s2: wp.array(dtype=vec2),
-        s3: wp.array(dtype=vec3),
-        s4: wp.array(dtype=vec4),
-        s5: wp.array(dtype=vec5),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        v20: wp.array(dtype=wptype),
-        v21: wp.array(dtype=wptype),
-        v30: wp.array(dtype=wptype),
-        v31: wp.array(dtype=wptype),
-        v32: wp.array(dtype=wptype),
-        v40: wp.array(dtype=wptype),
-        v41: wp.array(dtype=wptype),
-        v42: wp.array(dtype=wptype),
-        v43: wp.array(dtype=wptype),
-        v50: wp.array(dtype=wptype),
-        v51: wp.array(dtype=wptype),
-        v52: wp.array(dtype=wptype),
-        v53: wp.array(dtype=wptype),
-        v54: wp.array(dtype=wptype),
-    ):
-        v2result = wp.cw_div(v2[0], s2[0])
-        v3result = wp.cw_div(v3[0], s3[0])
-        v4result = wp.cw_div(v4[0], s4[0])
-        v5result = wp.cw_div(v5[0], s5[0])
-
-        v20[0] = wptype(2) * v2result[0]
-        v21[0] = wptype(2) * v2result[1]
-
-        v30[0] = wptype(2) * v3result[0]
-        v31[0] = wptype(2) * v3result[1]
-        v32[0] = wptype(2) * v3result[2]
-
-        v40[0] = wptype(2) * v4result[0]
-        v41[0] = wptype(2) * v4result[1]
-        v42[0] = wptype(2) * v4result[2]
-        v43[0] = wptype(2) * v4result[3]
-
-        v50[0] = wptype(2) * v5result[0]
-        v51[0] = wptype(2) * v5result[1]
-        v52[0] = wptype(2) * v5result[2]
-        v53[0] = wptype(2) * v5result[3]
-        v54[0] = wptype(2) * v5result[4]
-
-    kernel = getkernel(check_cw_div, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s2,
-                s3,
-                s4,
-                s5,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
-            device=device,
-        )
-
-    if dtype in np_int_types:
-        assert_np_equal(v20.numpy()[0], 2 * (v2.numpy()[0, 0] // s2.numpy()[0, 0]), tol=tol)
-        assert_np_equal(v21.numpy()[0], 2 * (v2.numpy()[0, 1] // s2.numpy()[0, 1]), tol=tol)
-
-        assert_np_equal(v30.numpy()[0], 2 * (v3.numpy()[0, 0] // s3.numpy()[0, 0]), tol=tol)
-        assert_np_equal(v31.numpy()[0], 2 * (v3.numpy()[0, 1] // s3.numpy()[0, 1]), tol=tol)
-        assert_np_equal(v32.numpy()[0], 2 * (v3.numpy()[0, 2] // s3.numpy()[0, 2]), tol=tol)
-
-        assert_np_equal(v40.numpy()[0], 2 * (v4.numpy()[0, 0] // s4.numpy()[0, 0]), tol=tol)
-        assert_np_equal(v41.numpy()[0], 2 * (v4.numpy()[0, 1] // s4.numpy()[0, 1]), tol=tol)
-        assert_np_equal(v42.numpy()[0], 2 * (v4.numpy()[0, 2] // s4.numpy()[0, 2]), tol=tol)
-        assert_np_equal(v43.numpy()[0], 2 * (v4.numpy()[0, 3] // s4.numpy()[0, 3]), tol=tol)
-
-        assert_np_equal(v50.numpy()[0], 2 * (v5.numpy()[0, 0] // s5.numpy()[0, 0]), tol=tol)
-        assert_np_equal(v51.numpy()[0], 2 * (v5.numpy()[0, 1] // s5.numpy()[0, 1]), tol=tol)
-        assert_np_equal(v52.numpy()[0], 2 * (v5.numpy()[0, 2] // s5.numpy()[0, 2]), tol=tol)
-        assert_np_equal(v53.numpy()[0], 2 * (v5.numpy()[0, 3] // s5.numpy()[0, 3]), tol=tol)
-        assert_np_equal(v54.numpy()[0], 2 * (v5.numpy()[0, 4] // s5.numpy()[0, 4]), tol=tol)
-    else:
-        assert_np_equal(v20.numpy()[0], 2 * v2.numpy()[0, 0] / s2.numpy()[0, 0], tol=tol)
-        assert_np_equal(v21.numpy()[0], 2 * v2.numpy()[0, 1] / s2.numpy()[0, 1], tol=tol)
-
-        assert_np_equal(v30.numpy()[0], 2 * v3.numpy()[0, 0] / s3.numpy()[0, 0], tol=tol)
-        assert_np_equal(v31.numpy()[0], 2 * v3.numpy()[0, 1] / s3.numpy()[0, 1], tol=tol)
-        assert_np_equal(v32.numpy()[0], 2 * v3.numpy()[0, 2] / s3.numpy()[0, 2], tol=tol)
-
-        assert_np_equal(v40.numpy()[0], 2 * v4.numpy()[0, 0] / s4.numpy()[0, 0], tol=tol)
-        assert_np_equal(v41.numpy()[0], 2 * v4.numpy()[0, 1] / s4.numpy()[0, 1], tol=tol)
-        assert_np_equal(v42.numpy()[0], 2 * v4.numpy()[0, 2] / s4.numpy()[0, 2], tol=tol)
-        assert_np_equal(v43.numpy()[0], 2 * v4.numpy()[0, 3] / s4.numpy()[0, 3], tol=tol)
-
-        assert_np_equal(v50.numpy()[0], 2 * v5.numpy()[0, 0] / s5.numpy()[0, 0], tol=tol)
-        assert_np_equal(v51.numpy()[0], 2 * v5.numpy()[0, 1] / s5.numpy()[0, 1], tol=tol)
-        assert_np_equal(v52.numpy()[0], 2 * v5.numpy()[0, 2] / s5.numpy()[0, 2], tol=tol)
-        assert_np_equal(v53.numpy()[0], 2 * v5.numpy()[0, 3] / s5.numpy()[0, 3], tol=tol)
-        assert_np_equal(v54.numpy()[0], 2 * v5.numpy()[0, 4] / s5.numpy()[0, 4], tol=tol)
-
-    if dtype in np_float_types:
-        incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
-        scmps = np.concatenate([v.numpy()[0] for v in [s2, s3, s4, s5]])
-
-        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
-            tape.backward(loss=l)
-            sgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [s2, s3, s4, s5]])
-            expected_grads = np.zeros_like(sgrads)
+    ):
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[],
+            device=device,
+        )
 
-            # d/ds v/s = -v/s^2
-            expected_grads[i] = -incmps[i] * 2 / (scmps[i] * scmps[i])
-            assert_np_equal(sgrads, expected_grads, tol=20 * tol)
 
-            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
-            expected_grads = np.zeros_like(allgrads)
+def test_tpl_ops_with_anon(test, device):
+    vec3i = wp.vec(3, dtype=int)
 
-            # d/dv v/s = 1/s
-            expected_grads[i] = 2 / scmps[i]
-            assert_np_equal(allgrads, expected_grads, tol=tol)
+    v = wp.vec3i(1, 2, 3)
+    v += vec3i(2, 3, 4)
+    v -= vec3i(3, 4, 5)
+    test.assertSequenceEqual(v, (0, 1, 2))
 
-            tape.zero()
+    v = vec3i(1, 2, 3)
+    v += wp.vec3i(2, 3, 4)
+    v -= wp.vec3i(3, 4, 5)
+    test.assertSequenceEqual(v, (0, 1, 2))
 
 
-def test_addition(test, device, dtype, register_kernels=False):
+def test_negation(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
     tol = {
@@ -1836,15 +221,15 @@ def test_addition(test, device, dtype, register_kernels=False):
     vec4 = wp.types.vector(length=4, dtype=wptype)
     vec5 = wp.types.vector(length=5, dtype=wptype)
 
-    def check_add(
-        s2: wp.array(dtype=vec2),
-        s3: wp.array(dtype=vec3),
-        s4: wp.array(dtype=vec4),
-        s5: wp.array(dtype=vec5),
+    def check_negation(
         v2: wp.array(dtype=vec2),
         v3: wp.array(dtype=vec3),
         v4: wp.array(dtype=vec4),
         v5: wp.array(dtype=vec5),
+        v2out: wp.array(dtype=vec2),
+        v3out: wp.array(dtype=vec3),
+        v4out: wp.array(dtype=vec4),
+        v5out: wp.array(dtype=vec5),
         v20: wp.array(dtype=wptype),
         v21: wp.array(dtype=wptype),
         v30: wp.array(dtype=wptype),
@@ -1860,11 +245,17 @@ def check_add(
         v53: wp.array(dtype=wptype),
         v54: wp.array(dtype=wptype),
     ):
-        v2result = v2[0] + s2[0]
-        v3result = v3[0] + s3[0]
-        v4result = v4[0] + s4[0]
-        v5result = v5[0] + s5[0]
+        v2result = -v2[0]
+        v3result = -v3[0]
+        v4result = -v4[0]
+        v5result = -v5[0]
+
+        v2out[0] = v2result
+        v3out[0] = v3result
+        v4out[0] = v4result
+        v5out[0] = v5result
 
+        # multiply these outputs by 2 so we've got something to backpropagate:
         v20[0] = wptype(2) * v2result[0]
         v21[0] = wptype(2) * v2result[1]
 
@@ -1883,19 +274,21 @@ def check_add(
         v53[0] = wptype(2) * v5result[3]
         v54[0] = wptype(2) * v5result[4]
 
-    kernel = getkernel(check_add, suffix=dtype.__name__)
+    kernel = getkernel(check_negation, suffix=dtype.__name__)
 
     if register_kernels:
         return
 
-    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
     v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
     v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
     v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v5_np = randvals(rng, (1, 5), dtype)
+    v5 = wp.array(v5_np, dtype=vec5, requires_grad=True, device=device)
+
+    v2out = wp.zeros(1, dtype=vec2, device=device)
+    v3out = wp.zeros(1, dtype=vec3, device=device)
+    v4out = wp.zeros(1, dtype=vec4, device=device)
+    v5out = wp.zeros(1, dtype=vec5, device=device)
     v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
     v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
     v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
@@ -1910,57 +303,31 @@ def check_add(
     v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
     v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
     v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
     tape = wp.Tape()
     with tape:
         wp.launch(
             kernel,
             dim=1,
-            inputs=[
-                s2,
-                s3,
-                s4,
-                s5,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            inputs=[v2, v3, v4, v5],
+            outputs=[v2out, v3out, v4out, v5out, v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
             device=device,
         )
 
-    assert_np_equal(v20.numpy()[0], 2 * (v2.numpy()[0, 0] + s2.numpy()[0, 0]), tol=tol)
-    assert_np_equal(v21.numpy()[0], 2 * (v2.numpy()[0, 1] + s2.numpy()[0, 1]), tol=tol)
-
-    assert_np_equal(v30.numpy()[0], 2 * (v3.numpy()[0, 0] + s3.numpy()[0, 0]), tol=tol)
-    assert_np_equal(v31.numpy()[0], 2 * (v3.numpy()[0, 1] + s3.numpy()[0, 1]), tol=tol)
-    assert_np_equal(v32.numpy()[0], 2 * (v3.numpy()[0, 2] + s3.numpy()[0, 2]), tol=tol)
-
-    assert_np_equal(v40.numpy()[0], 2 * (v4.numpy()[0, 0] + s4.numpy()[0, 0]), tol=tol)
-    assert_np_equal(v41.numpy()[0], 2 * (v4.numpy()[0, 1] + s4.numpy()[0, 1]), tol=tol)
-    assert_np_equal(v42.numpy()[0], 2 * (v4.numpy()[0, 2] + s4.numpy()[0, 2]), tol=tol)
-    assert_np_equal(v43.numpy()[0], 2 * (v4.numpy()[0, 3] + s4.numpy()[0, 3]), tol=tol)
-
-    assert_np_equal(v50.numpy()[0], 2 * (v5.numpy()[0, 0] + s5.numpy()[0, 0]), tol=tol)
-    assert_np_equal(v51.numpy()[0], 2 * (v5.numpy()[0, 1] + s5.numpy()[0, 1]), tol=tol)
-    assert_np_equal(v52.numpy()[0], 2 * (v5.numpy()[0, 2] + s5.numpy()[0, 2]), tol=tol)
-    assert_np_equal(v53.numpy()[0], 2 * (v5.numpy()[0, 3] + s5.numpy()[0, 3]), tol=tol)
-    assert_np_equal(v54.numpy()[0], 2 * (v5.numpy()[0, 4] + s5.numpy()[0, 4]), tol=2 * tol)
-
     if dtype in np_float_types:
         for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
             tape.backward(loss=l)
-            sgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [s2, s3, s4, s5]])
-            expected_grads = np.zeros_like(sgrads)
-
-            expected_grads[i] = 2
-            assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
             allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
+            expected_grads = np.zeros_like(allgrads)
+            expected_grads[i] = -2
             assert_np_equal(allgrads, expected_grads, tol=tol)
-
             tape.zero()
 
+    assert_np_equal(v2out.numpy()[0], -v2.numpy()[0], tol=tol)
+    assert_np_equal(v3out.numpy()[0], -v3.numpy()[0], tol=tol)
+    assert_np_equal(v4out.numpy()[0], -v4.numpy()[0], tol=tol)
+    assert_np_equal(v5out.numpy()[0], -v5.numpy()[0], tol=tol)
+
 
 def test_subtraction_unsigned(test, device, dtype, register_kernels=False):
     wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
@@ -2156,127 +523,6 @@ def check_subtraction(
             tape.zero()
 
 
-def test_dotproduct(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    tol = {
-        np.float16: 1.0e-2,
-        np.float32: 1.0e-6,
-        np.float64: 1.0e-8,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    def check_dot(
-        s2: wp.array(dtype=vec2),
-        s3: wp.array(dtype=vec3),
-        s4: wp.array(dtype=vec4),
-        s5: wp.array(dtype=vec5),
-        v2: wp.array(dtype=vec2),
-        v3: wp.array(dtype=vec3),
-        v4: wp.array(dtype=vec4),
-        v5: wp.array(dtype=vec5),
-        dot2: wp.array(dtype=wptype),
-        dot3: wp.array(dtype=wptype),
-        dot4: wp.array(dtype=wptype),
-        dot5: wp.array(dtype=wptype),
-    ):
-        dot2[0] = wptype(2) * wp.dot(v2[0], s2[0])
-        dot3[0] = wptype(2) * wp.dot(v3[0], s3[0])
-        dot4[0] = wptype(2) * wp.dot(v4[0], s4[0])
-        dot5[0] = wptype(2) * wp.dot(v5[0], s5[0])
-
-    kernel = getkernel(check_dot, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
-    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
-    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
-    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
-    dot2 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    dot3 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    dot4 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    dot5 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(
-            kernel,
-            dim=1,
-            inputs=[
-                s2,
-                s3,
-                s4,
-                s5,
-                v2,
-                v3,
-                v4,
-                v5,
-            ],
-            outputs=[dot2, dot3, dot4, dot5],
-            device=device,
-        )
-
-    assert_np_equal(dot2.numpy()[0], 2.0 * (v2.numpy() * s2.numpy()).sum(), tol=10 * tol)
-    assert_np_equal(dot3.numpy()[0], 2.0 * (v3.numpy() * s3.numpy()).sum(), tol=10 * tol)
-    assert_np_equal(dot4.numpy()[0], 2.0 * (v4.numpy() * s4.numpy()).sum(), tol=10 * tol)
-    assert_np_equal(dot5.numpy()[0], 2.0 * (v5.numpy() * s5.numpy()).sum(), tol=10 * tol)
-
-    if dtype in np_float_types:
-        tape.backward(loss=dot2)
-        sgrads = tape.gradients[s2].numpy()[0]
-        expected_grads = 2.0 * v2.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v2].numpy()[0]
-        expected_grads = 2.0 * s2.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=tol)
-
-        tape.zero()
-
-        tape.backward(loss=dot3)
-        sgrads = tape.gradients[s3].numpy()[0]
-        expected_grads = 2.0 * v3.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v3].numpy()[0]
-        expected_grads = 2.0 * s3.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=tol)
-
-        tape.zero()
-
-        tape.backward(loss=dot4)
-        sgrads = tape.gradients[s4].numpy()[0]
-        expected_grads = 2.0 * v4.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v4].numpy()[0]
-        expected_grads = 2.0 * s4.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=tol)
-
-        tape.zero()
-
-        tape.backward(loss=dot5)
-        sgrads = tape.gradients[s5].numpy()[0]
-        expected_grads = 2.0 * v5.numpy()[0]
-        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
-
-        vgrads = tape.gradients[v5].numpy()[0]
-        expected_grads = 2.0 * s5.numpy()[0]
-        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
-
-        tape.zero()
-
-
 def test_length(test, device, dtype, register_kernels=False):
     rng = np.random.default_rng(123)
 
@@ -2738,139 +984,6 @@ def check_cross(
         tape.zero()
 
 
-def test_minmax(test, device, dtype, register_kernels=False):
-    rng = np.random.default_rng(123)
-
-    # \TODO: not quite sure why, but the numbers are off for 16 bit float
-    # on the cpu (but not cuda). This is probably just the sketchy float16
-    # arithmetic I implemented to get all this stuff working, so
-    # hopefully that can be fixed when we do that correctly.
-    tol = {
-        np.float16: 1.0e-2,
-    }.get(dtype, 0)
-
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    # \TODO: Also not quite sure why: this kernel compiles incredibly
-    # slowly though...
-    def check_vec_min_max(
-        a: wp.array(dtype=wptype, ndim=2),
-        b: wp.array(dtype=wptype, ndim=2),
-        mins: wp.array(dtype=wptype, ndim=2),
-        maxs: wp.array(dtype=wptype, ndim=2),
-    ):
-        for i in range(10):
-            # multiplying by 2 so we've got something to backpropagate:
-            a2read = vec2(a[i, 0], a[i, 1])
-            b2read = vec2(b[i, 0], b[i, 1])
-            c2 = wptype(2) * wp.min(a2read, b2read)
-            d2 = wptype(2) * wp.max(a2read, b2read)
-
-            a3read = vec3(a[i, 2], a[i, 3], a[i, 4])
-            b3read = vec3(b[i, 2], b[i, 3], b[i, 4])
-            c3 = wptype(2) * wp.min(a3read, b3read)
-            d3 = wptype(2) * wp.max(a3read, b3read)
-
-            a4read = vec4(a[i, 5], a[i, 6], a[i, 7], a[i, 8])
-            b4read = vec4(b[i, 5], b[i, 6], b[i, 7], b[i, 8])
-            c4 = wptype(2) * wp.min(a4read, b4read)
-            d4 = wptype(2) * wp.max(a4read, b4read)
-
-            a5read = vec5(a[i, 9], a[i, 10], a[i, 11], a[i, 12], a[i, 13])
-            b5read = vec5(b[i, 9], b[i, 10], b[i, 11], b[i, 12], b[i, 13])
-            c5 = wptype(2) * wp.min(a5read, b5read)
-            d5 = wptype(2) * wp.max(a5read, b5read)
-
-            mins[i, 0] = c2[0]
-            mins[i, 1] = c2[1]
-
-            mins[i, 2] = c3[0]
-            mins[i, 3] = c3[1]
-            mins[i, 4] = c3[2]
-
-            mins[i, 5] = c4[0]
-            mins[i, 6] = c4[1]
-            mins[i, 7] = c4[2]
-            mins[i, 8] = c4[3]
-
-            mins[i, 9] = c5[0]
-            mins[i, 10] = c5[1]
-            mins[i, 11] = c5[2]
-            mins[i, 12] = c5[3]
-            mins[i, 13] = c5[4]
-
-            maxs[i, 0] = d2[0]
-            maxs[i, 1] = d2[1]
-
-            maxs[i, 2] = d3[0]
-            maxs[i, 3] = d3[1]
-            maxs[i, 4] = d3[2]
-
-            maxs[i, 5] = d4[0]
-            maxs[i, 6] = d4[1]
-            maxs[i, 7] = d4[2]
-            maxs[i, 8] = d4[3]
-
-            maxs[i, 9] = d5[0]
-            maxs[i, 10] = d5[1]
-            maxs[i, 11] = d5[2]
-            maxs[i, 12] = d5[3]
-            maxs[i, 13] = d5[4]
-
-    kernel = getkernel(check_vec_min_max, suffix=dtype.__name__)
-    output_select_kernel = get_select_kernel2(wptype)
-
-    if register_kernels:
-        return
-
-    a = wp.array(randvals(rng, (10, 14), dtype), dtype=wptype, requires_grad=True, device=device)
-    b = wp.array(randvals(rng, (10, 14), dtype), dtype=wptype, requires_grad=True, device=device)
-
-    mins = wp.zeros((10, 14), dtype=wptype, requires_grad=True, device=device)
-    maxs = wp.zeros((10, 14), dtype=wptype, requires_grad=True, device=device)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(kernel, dim=1, inputs=[a, b], outputs=[mins, maxs], device=device)
-
-    assert_np_equal(mins.numpy(), 2 * np.minimum(a.numpy(), b.numpy()), tol=tol)
-    assert_np_equal(maxs.numpy(), 2 * np.maximum(a.numpy(), b.numpy()), tol=tol)
-
-    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
-    if dtype in np_float_types:
-        for i in range(10):
-            for j in range(14):
-                tape = wp.Tape()
-                with tape:
-                    wp.launch(kernel, dim=1, inputs=[a, b], outputs=[mins, maxs], device=device)
-                    wp.launch(output_select_kernel, dim=1, inputs=[mins, i, j], outputs=[out], device=device)
-
-                tape.backward(loss=out)
-                expected = np.zeros_like(a.numpy())
-                expected[i, j] = 2 if (a.numpy()[i, j] < b.numpy()[i, j]) else 0
-                assert_np_equal(tape.gradients[a].numpy(), expected, tol=tol)
-                expected[i, j] = 2 if (b.numpy()[i, j] < a.numpy()[i, j]) else 0
-                assert_np_equal(tape.gradients[b].numpy(), expected, tol=tol)
-                tape.zero()
-
-                tape = wp.Tape()
-                with tape:
-                    wp.launch(kernel, dim=1, inputs=[a, b], outputs=[mins, maxs], device=device)
-                    wp.launch(output_select_kernel, dim=1, inputs=[maxs, i, j], outputs=[out], device=device)
-
-                tape.backward(loss=out)
-                expected = np.zeros_like(a.numpy())
-                expected[i, j] = 2 if (a.numpy()[i, j] > b.numpy()[i, j]) else 0
-                assert_np_equal(tape.gradients[a].numpy(), expected, tol=tol)
-                expected[i, j] = 2 if (b.numpy()[i, j] > a.numpy()[i, j]) else 0
-                assert_np_equal(tape.gradients[b].numpy(), expected, tol=tol)
-                tape.zero()
-
-
 def test_casting_constructors(test, device, dtype, register_kernels=False):
     np_type = np.dtype(dtype)
     wp_type = wp.types.np_dtype_to_warp_type[np_type]
@@ -2974,85 +1087,6 @@ def cast_float64(a: wp.array(dtype=wp_type, ndim=2), b: wp.array(dtype=wp64, ndi
     assert_np_equal(out, a_grad.numpy())
 
 
-def test_equivalent_types(test, device, dtype, register_kernels=False):
-    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
-
-    # vector types
-    vec2 = wp.types.vector(length=2, dtype=wptype)
-    vec3 = wp.types.vector(length=3, dtype=wptype)
-    vec4 = wp.types.vector(length=4, dtype=wptype)
-    vec5 = wp.types.vector(length=5, dtype=wptype)
-
-    # vector types equivalent to the above
-    vec2_equiv = wp.types.vector(length=2, dtype=wptype)
-    vec3_equiv = wp.types.vector(length=3, dtype=wptype)
-    vec4_equiv = wp.types.vector(length=4, dtype=wptype)
-    vec5_equiv = wp.types.vector(length=5, dtype=wptype)
-
-    # declare kernel with original types
-    def check_equivalence(
-        v2: vec2,
-        v3: vec3,
-        v4: vec4,
-        v5: vec5,
-    ):
-        wp.expect_eq(v2, vec2(wptype(1), wptype(2)))
-        wp.expect_eq(v3, vec3(wptype(1), wptype(2), wptype(3)))
-        wp.expect_eq(v4, vec4(wptype(1), wptype(2), wptype(3), wptype(4)))
-        wp.expect_eq(v5, vec5(wptype(1), wptype(2), wptype(3), wptype(4), wptype(5)))
-
-        wp.expect_eq(v2, vec2_equiv(wptype(1), wptype(2)))
-        wp.expect_eq(v3, vec3_equiv(wptype(1), wptype(2), wptype(3)))
-        wp.expect_eq(v4, vec4_equiv(wptype(1), wptype(2), wptype(3), wptype(4)))
-        wp.expect_eq(v5, vec5_equiv(wptype(1), wptype(2), wptype(3), wptype(4), wptype(5)))
-
-    kernel = getkernel(check_equivalence, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    # call kernel with equivalent types
-    v2 = vec2_equiv(1, 2)
-    v3 = vec3_equiv(1, 2, 3)
-    v4 = vec4_equiv(1, 2, 3, 4)
-    v5 = vec5_equiv(1, 2, 3, 4, 5)
-
-    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5], device=device)
-
-
-def test_conversions(test, device, dtype, register_kernels=False):
-    def check_vectors_equal(
-        v0: wp.vec3,
-        v1: wp.vec3,
-        v2: wp.vec3,
-        v3: wp.vec3,
-    ):
-        wp.expect_eq(v1, v0)
-        wp.expect_eq(v2, v0)
-        wp.expect_eq(v3, v0)
-
-    kernel = getkernel(check_vectors_equal, suffix=dtype.__name__)
-
-    if register_kernels:
-        return
-
-    v0 = wp.vec3(1, 2, 3)
-
-    # test explicit conversions - constructing vectors from different containers
-    v1 = wp.vec3((1, 2, 3))
-    v2 = wp.vec3([1, 2, 3])
-    v3 = wp.vec3(np.array([1, 2, 3], dtype=dtype))
-
-    wp.launch(kernel, dim=1, inputs=[v0, v1, v2, v3], device=device)
-
-    # test implicit conversions - passing different containers as vectors to wp.launch()
-    v1 = (1, 2, 3)
-    v2 = [1, 2, 3]
-    v3 = np.array([1, 2, 3], dtype=dtype)
-
-    wp.launch(kernel, dim=1, inputs=[v0, v1, v2, v3], device=device)
-
-
 @wp.kernel
 def test_vector_constructor_value_func():
     a = wp.vec2()
@@ -3122,175 +1156,113 @@ def test_constructors_constant_length():
         v[i] = float(i)
 
 
-def register(parent):
-    devices = get_test_devices()
-
-    class TestVec(parent):
-        pass
+devices = get_test_devices()
 
-    add_kernel_test(TestVec, test_vector_constructor_value_func, dim=1, devices=devices)
-    add_kernel_test(TestVec, test_constructors_explicit_precision, dim=1, devices=devices)
-    add_kernel_test(TestVec, test_constructors_default_precision, dim=1, devices=devices)
-    add_kernel_test(TestVec, test_constructors_constant_length, dim=1, devices=devices)
 
-    vec10 = wp.types.vector(length=10, dtype=float)
-    add_kernel_test(
-        TestVec,
-        test_vector_mutation,
-        dim=1,
-        inputs=[vec10(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0)],
-        devices=devices,
-    )
+class TestVec(unittest.TestCase):
+    pass
 
-    for dtype in np_unsigned_int_types:
-        add_function_test_register_kernel(
-            TestVec,
-            f"test_subtraction_unsigned_{dtype.__name__}",
-            test_subtraction_unsigned,
-            devices=devices,
-            dtype=dtype,
-        )
 
-    for dtype in np_signed_int_types + np_float_types:
-        add_function_test_register_kernel(
-            TestVec, f"test_negation_{dtype.__name__}", test_negation, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_subtraction_{dtype.__name__}", test_subtraction, devices=devices, dtype=dtype
-        )
+add_kernel_test(TestVec, test_vector_constructor_value_func, dim=1, devices=devices)
+add_kernel_test(TestVec, test_constructors_explicit_precision, dim=1, devices=devices)
+add_kernel_test(TestVec, test_constructors_default_precision, dim=1, devices=devices)
+add_kernel_test(TestVec, test_constructors_constant_length, dim=1, devices=devices)
 
-    for dtype in np_float_types:
-        add_function_test_register_kernel(
-            TestVec, f"test_crossproduct_{dtype.__name__}", test_crossproduct, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_length_{dtype.__name__}", test_length, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_normalize_{dtype.__name__}", test_normalize, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec,
-            f"test_casting_constructors_{dtype.__name__}",
-            test_casting_constructors,
-            devices=devices,
-            dtype=dtype,
-        )
+vec10 = wp.types.vector(length=10, dtype=float)
+add_kernel_test(
+    TestVec,
+    test_vector_mutation,
+    dim=1,
+    inputs=[vec10(1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0)],
+    devices=devices,
+)
 
-    add_function_test(
+for dtype in np_unsigned_int_types:
+    add_function_test_register_kernel(
         TestVec,
-        "test_anon_constructor_error_dtype_keyword_missing",
-        test_anon_constructor_error_dtype_keyword_missing,
+        f"test_subtraction_unsigned_{dtype.__name__}",
+        test_subtraction_unsigned,
         devices=devices,
+        dtype=dtype,
     )
-    add_function_test(
-        TestVec,
-        "test_anon_constructor_error_length_mismatch",
-        test_anon_constructor_error_length_mismatch,
-        devices=devices,
-    )
-    add_function_test(
-        TestVec,
-        "test_anon_constructor_error_numeric_arg_missing_1",
-        test_anon_constructor_error_numeric_arg_missing_1,
-        devices=devices,
+
+for dtype in np_signed_int_types + np_float_types:
+    add_function_test_register_kernel(
+        TestVec, f"test_negation_{dtype.__name__}", test_negation, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestVec,
-        "test_anon_constructor_error_numeric_arg_missing_2",
-        test_anon_constructor_error_numeric_arg_missing_2,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestVec, f"test_subtraction_{dtype.__name__}", test_subtraction, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestVec,
-        "test_anon_constructor_error_dtype_keyword_extraneous",
-        test_anon_constructor_error_dtype_keyword_extraneous,
-        devices=devices,
+
+for dtype in np_float_types:
+    add_function_test_register_kernel(
+        TestVec, f"test_crossproduct_{dtype.__name__}", test_crossproduct, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestVec,
-        "test_anon_constructor_error_numeric_args_mismatch",
-        test_anon_constructor_error_numeric_args_mismatch,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestVec, f"test_length_{dtype.__name__}", test_length, devices=devices, dtype=dtype
     )
-    add_function_test(
-        TestVec,
-        "test_tpl_constructor_error_incompatible_sizes",
-        test_tpl_constructor_error_incompatible_sizes,
-        devices=devices,
+    add_function_test_register_kernel(
+        TestVec, f"test_normalize_{dtype.__name__}", test_normalize, devices=devices, dtype=dtype
     )
-    add_function_test(
+    add_function_test_register_kernel(
         TestVec,
-        "test_tpl_constructor_error_numeric_args_mismatch",
-        test_tpl_constructor_error_numeric_args_mismatch,
+        f"test_casting_constructors_{dtype.__name__}",
+        test_casting_constructors,
         devices=devices,
+        dtype=dtype,
     )
-    add_function_test(TestVec, "test_tpl_ops_with_anon", test_tpl_ops_with_anon)
-
-    for dtype in np_scalar_types:
-        add_function_test(TestVec, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
-        add_function_test(TestVec, f"test_components_{dtype.__name__}", test_components, devices=None, dtype=dtype)
-        add_function_test(
-            TestVec, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_anon_type_instance_{dtype.__name__}", test_anon_type_instance, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_indexing_{dtype.__name__}", test_indexing, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_equality_{dtype.__name__}", test_equality, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec,
-            f"test_scalar_multiplication_{dtype.__name__}",
-            test_scalar_multiplication,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestVec,
-            f"test_scalar_multiplication_rightmul_{dtype.__name__}",
-            test_scalar_multiplication_rightmul,
-            devices=devices,
-            dtype=dtype,
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_cw_multiplication_{dtype.__name__}", test_cw_multiplication, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_scalar_division_{dtype.__name__}", test_scalar_division, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_cw_division_{dtype.__name__}", test_cw_division, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_addition_{dtype.__name__}", test_addition, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_dotproduct_{dtype.__name__}", test_dotproduct, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_equivalent_types_{dtype.__name__}", test_equivalent_types, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_conversions_{dtype.__name__}", test_conversions, devices=devices, dtype=dtype
-        )
-        add_function_test_register_kernel(
-            TestVec, f"test_constants_{dtype.__name__}", test_constants, devices=devices, dtype=dtype
-        )
-
-        # the kernels in this test compile incredibly slowly...
-        # add_function_test_register_kernel(TestVec, f"test_minmax_{dtype.__name__}", test_minmax, devices=devices, dtype=dtype)
 
-    return TestVec
+add_function_test(
+    TestVec,
+    "test_anon_constructor_error_dtype_keyword_missing",
+    test_anon_constructor_error_dtype_keyword_missing,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_anon_constructor_error_length_mismatch",
+    test_anon_constructor_error_length_mismatch,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_anon_constructor_error_numeric_arg_missing_1",
+    test_anon_constructor_error_numeric_arg_missing_1,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_anon_constructor_error_numeric_arg_missing_2",
+    test_anon_constructor_error_numeric_arg_missing_2,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_anon_constructor_error_dtype_keyword_extraneous",
+    test_anon_constructor_error_dtype_keyword_extraneous,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_anon_constructor_error_numeric_args_mismatch",
+    test_anon_constructor_error_numeric_args_mismatch,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_tpl_constructor_error_incompatible_sizes",
+    test_tpl_constructor_error_incompatible_sizes,
+    devices=devices,
+)
+add_function_test(
+    TestVec,
+    "test_tpl_constructor_error_numeric_args_mismatch",
+    test_tpl_constructor_error_numeric_args_mismatch,
+    devices=devices,
+)
+add_function_test(TestVec, "test_tpl_ops_with_anon", test_tpl_ops_with_anon)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_vec_lite.py b/warp/tests/test_vec_lite.py
index 71e706eda..4eda8a5af 100644
--- a/warp/tests/test_vec_lite.py
+++ b/warp/tests/test_vec_lite.py
@@ -7,9 +7,8 @@
 
 import unittest
 
-import numpy as np
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -57,20 +56,18 @@ def test_constructors_default_precision():
         wp.expect_eq(custom[i], float(i))
 
 
-def register(parent):
-    devices = get_test_devices()
+devices = get_test_devices()
 
-    class TestVec(parent):
-        pass
 
-    add_kernel_test(TestVec, test_vector_constructor_value_func, dim=1, devices=devices)
-    add_kernel_test(TestVec, test_constructors_explicit_precision, dim=1, devices=devices)
-    add_kernel_test(TestVec, test_constructors_default_precision, dim=1, devices=devices)
+class TestVecLite(unittest.TestCase):
+    pass
 
-    return TestVec
+
+add_kernel_test(TestVecLite, test_vector_constructor_value_func, dim=1, devices=devices)
+add_kernel_test(TestVecLite, test_constructors_explicit_precision, dim=1, devices=devices)
+add_kernel_test(TestVecLite, test_constructors_default_precision, dim=1, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_vec_scalar_ops.py b/warp/tests/test_vec_scalar_ops.py
new file mode 100644
index 000000000..c3b9fdfc9
--- /dev/null
+++ b/warp/tests/test_vec_scalar_ops.py
@@ -0,0 +1,2099 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import unittest
+
+import numpy as np
+
+import warp as wp
+from warp.tests.unittest_utils import *
+
+wp.init()
+
+np_signed_int_types = [
+    np.int8,
+    np.int16,
+    np.int32,
+    np.int64,
+    np.byte,
+]
+
+np_unsigned_int_types = [
+    np.uint8,
+    np.uint16,
+    np.uint32,
+    np.uint64,
+    np.ubyte,
+]
+
+np_int_types = np_signed_int_types + np_unsigned_int_types
+
+np_float_types = [np.float16, np.float32, np.float64]
+
+np_scalar_types = np_int_types + np_float_types
+
+
+def randvals(rng, shape, dtype):
+    if dtype in np_float_types:
+        return rng.standard_normal(size=shape).astype(dtype)
+    elif dtype in [np.int8, np.uint8, np.byte, np.ubyte]:
+        return rng.integers(1, high=3, size=shape, dtype=dtype)
+    return rng.integers(1, high=5, size=shape, dtype=dtype)
+
+
+kernel_cache = dict()
+
+
+def getkernel(func, suffix=""):
+    key = func.__name__ + "_" + suffix
+    if key not in kernel_cache:
+        kernel_cache[key] = wp.Kernel(func=func, key=key)
+    return kernel_cache[key]
+
+
+def get_select_kernel(dtype):
+    def output_select_kernel_fn(
+        input: wp.array(dtype=dtype),
+        index: int,
+        out: wp.array(dtype=dtype),
+    ):
+        out[0] = input[index]
+
+    return getkernel(output_select_kernel_fn, suffix=dtype.__name__)
+
+
+def get_select_kernel2(dtype):
+    def output_select_kernel2_fn(
+        input: wp.array(dtype=dtype, ndim=2),
+        index0: int,
+        index1: int,
+        out: wp.array(dtype=dtype),
+    ):
+        out[0] = input[index0, index1]
+
+    return getkernel(output_select_kernel2_fn, suffix=dtype.__name__)
+
+
+def test_arrays(test, device, dtype):
+    rng = np.random.default_rng(123)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    v2_np = randvals(rng, (10, 2), dtype)
+    v3_np = randvals(rng, (10, 3), dtype)
+    v4_np = randvals(rng, (10, 4), dtype)
+    v5_np = randvals(rng, (10, 5), dtype)
+
+    v2 = wp.array(v2_np, dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(v3_np, dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(v4_np, dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(v5_np, dtype=vec5, requires_grad=True, device=device)
+
+    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
+    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
+    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
+    assert_np_equal(v5.numpy(), v5_np, tol=1.0e-6)
+
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+
+    v2 = wp.array(v2_np, dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(v3_np, dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(v4_np, dtype=vec4, requires_grad=True, device=device)
+
+    assert_np_equal(v2.numpy(), v2_np, tol=1.0e-6)
+    assert_np_equal(v3.numpy(), v3_np, tol=1.0e-6)
+    assert_np_equal(v4.numpy(), v4_np, tol=1.0e-6)
+
+
+def test_components(test, device, dtype):
+    # test accessing vector components from Python - this is especially important
+    # for float16, which requires special handling internally
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+
+    v = vec3(1, 2, 3)
+
+    # test __getitem__ for individual components
+    test.assertEqual(v[0], 1)
+    test.assertEqual(v[1], 2)
+    test.assertEqual(v[2], 3)
+
+    # test __getitem__ for slices
+    s = v[:]
+    test.assertEqual(s[0], 1)
+    test.assertEqual(s[1], 2)
+    test.assertEqual(s[2], 3)
+
+    s = v[1:]
+    test.assertEqual(s[0], 2)
+    test.assertEqual(s[1], 3)
+
+    s = v[:2]
+    test.assertEqual(s[0], 1)
+    test.assertEqual(s[1], 2)
+
+    s = v[::2]
+    test.assertEqual(s[0], 1)
+    test.assertEqual(s[1], 3)
+
+    # test __setitem__ for individual components
+    v[0] = 4
+    v[1] = 5
+    v[2] = 6
+    test.assertEqual(v[0], 4)
+    test.assertEqual(v[1], 5)
+    test.assertEqual(v[2], 6)
+
+    # test __setitem__ for slices
+    v[:] = [7, 8, 9]
+    test.assertEqual(v[0], 7)
+    test.assertEqual(v[1], 8)
+    test.assertEqual(v[2], 9)
+
+    v[1:] = [10, 11]
+    test.assertEqual(v[0], 7)
+    test.assertEqual(v[1], 10)
+    test.assertEqual(v[2], 11)
+
+    v[:2] = [12, 13]
+    test.assertEqual(v[0], 12)
+    test.assertEqual(v[1], 13)
+    test.assertEqual(v[2], 11)
+
+    v[::2] = [14, 15]
+    test.assertEqual(v[0], 14)
+    test.assertEqual(v[1], 13)
+    test.assertEqual(v[2], 15)
+
+
+def test_py_arithmetic_ops(test, device, dtype):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def make_vec(*args):
+        if wptype in wp.types.int_types:
+            # Cast to the correct integer type to simulate wrapping.
+            return tuple(wptype._type_(x).value for x in args)
+
+        return args
+
+    vec_cls = wp.vec(3, wptype)
+
+    v = vec_cls(1, -2, 3)
+    test.assertSequenceEqual(+v, make_vec(1, -2, 3))
+    test.assertSequenceEqual(-v, make_vec(-1, 2, -3))
+    test.assertSequenceEqual(v + vec_cls(5, 5, 5), make_vec(6, 3, 8))
+    test.assertSequenceEqual(v - vec_cls(5, 5, 5), make_vec(-4, -7, -2))
+
+    v = vec_cls(2, 4, 6)
+    test.assertSequenceEqual(v * wptype(2), make_vec(4, 8, 12))
+    test.assertSequenceEqual(wptype(2) * v, make_vec(4, 8, 12))
+    test.assertSequenceEqual(v / wptype(2), make_vec(1, 2, 3))
+    test.assertSequenceEqual(wptype(24) / v, make_vec(12, 6, 4))
+
+
+def test_constructors(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_scalar_constructor(
+        input: wp.array(dtype=wptype),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = vec2(input[0])
+        v3result = vec3(input[0])
+        v4result = vec4(input[0])
+        v5result = vec5(input[0])
+
+        v2[0] = v2result
+        v3[0] = v3result
+        v4[0] = v4result
+        v5[0] = v5result
+
+        # multiply outputs by 2 so we've got something to backpropagate
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    def check_vector_constructors(
+        input: wp.array(dtype=wptype),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = vec2(input[0], input[1])
+        v3result = vec3(input[2], input[3], input[4])
+        v4result = vec4(input[5], input[6], input[7], input[8])
+        v5result = vec5(input[9], input[10], input[11], input[12], input[13])
+
+        v2[0] = v2result
+        v3[0] = v3result
+        v4[0] = v4result
+        v5[0] = v5result
+
+        # multiply the output by 2 so we've got something to backpropagate:
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    vec_kernel = getkernel(check_vector_constructors, suffix=dtype.__name__)
+    kernel = getkernel(check_scalar_constructor, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    input = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    v2 = wp.zeros(1, dtype=vec2, device=device)
+    v3 = wp.zeros(1, dtype=vec3, device=device)
+    v4 = wp.zeros(1, dtype=vec4, device=device)
+    v5 = wp.zeros(1, dtype=vec5, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[input],
+            outputs=[v2, v3, v4, v5, v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    if dtype in np_float_types:
+        for l in [v20, v21]:
+            tape.backward(loss=l)
+            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
+            tape.zero()
+
+        for l in [v30, v31, v32]:
+            tape.backward(loss=l)
+            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
+            tape.zero()
+
+        for l in [v40, v41, v42, v43]:
+            tape.backward(loss=l)
+            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
+            tape.zero()
+
+        for l in [v50, v51, v52, v53, v54]:
+            tape.backward(loss=l)
+            test.assertEqual(tape.gradients[input].numpy()[0], 2.0)
+            tape.zero()
+
+    val = input.numpy()[0]
+    assert_np_equal(v2.numpy()[0], np.array([val, val]), tol=1.0e-6)
+    assert_np_equal(v3.numpy()[0], np.array([val, val, val]), tol=1.0e-6)
+    assert_np_equal(v4.numpy()[0], np.array([val, val, val, val]), tol=1.0e-6)
+    assert_np_equal(v5.numpy()[0], np.array([val, val, val, val, val]), tol=1.0e-6)
+
+    assert_np_equal(v20.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v21.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v30.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v31.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v32.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v40.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v41.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v42.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v43.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v50.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v51.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v52.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v53.numpy()[0], 2 * val, tol=1.0e-6)
+    assert_np_equal(v54.numpy()[0], 2 * val, tol=1.0e-6)
+
+    input = wp.array(randvals(rng, [14], dtype), requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            vec_kernel,
+            dim=1,
+            inputs=[input],
+            outputs=[v2, v3, v4, v5, v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
+            tape.backward(loss=l)
+            grad = tape.gradients[input].numpy()
+            expected_grad = np.zeros_like(grad)
+            expected_grad[i] = 2
+            assert_np_equal(grad, expected_grad, tol=tol)
+            tape.zero()
+
+    assert_np_equal(v2.numpy()[0, 0], input.numpy()[0], tol=tol)
+    assert_np_equal(v2.numpy()[0, 1], input.numpy()[1], tol=tol)
+    assert_np_equal(v3.numpy()[0, 0], input.numpy()[2], tol=tol)
+    assert_np_equal(v3.numpy()[0, 1], input.numpy()[3], tol=tol)
+    assert_np_equal(v3.numpy()[0, 2], input.numpy()[4], tol=tol)
+    assert_np_equal(v4.numpy()[0, 0], input.numpy()[5], tol=tol)
+    assert_np_equal(v4.numpy()[0, 1], input.numpy()[6], tol=tol)
+    assert_np_equal(v4.numpy()[0, 2], input.numpy()[7], tol=tol)
+    assert_np_equal(v4.numpy()[0, 3], input.numpy()[8], tol=tol)
+    assert_np_equal(v5.numpy()[0, 0], input.numpy()[9], tol=tol)
+    assert_np_equal(v5.numpy()[0, 1], input.numpy()[10], tol=tol)
+    assert_np_equal(v5.numpy()[0, 2], input.numpy()[11], tol=tol)
+    assert_np_equal(v5.numpy()[0, 3], input.numpy()[12], tol=tol)
+    assert_np_equal(v5.numpy()[0, 4], input.numpy()[13], tol=tol)
+
+    assert_np_equal(v20.numpy()[0], 2 * input.numpy()[0], tol=tol)
+    assert_np_equal(v21.numpy()[0], 2 * input.numpy()[1], tol=tol)
+    assert_np_equal(v30.numpy()[0], 2 * input.numpy()[2], tol=tol)
+    assert_np_equal(v31.numpy()[0], 2 * input.numpy()[3], tol=tol)
+    assert_np_equal(v32.numpy()[0], 2 * input.numpy()[4], tol=tol)
+    assert_np_equal(v40.numpy()[0], 2 * input.numpy()[5], tol=tol)
+    assert_np_equal(v41.numpy()[0], 2 * input.numpy()[6], tol=tol)
+    assert_np_equal(v42.numpy()[0], 2 * input.numpy()[7], tol=tol)
+    assert_np_equal(v43.numpy()[0], 2 * input.numpy()[8], tol=tol)
+    assert_np_equal(v50.numpy()[0], 2 * input.numpy()[9], tol=tol)
+    assert_np_equal(v51.numpy()[0], 2 * input.numpy()[10], tol=tol)
+    assert_np_equal(v52.numpy()[0], 2 * input.numpy()[11], tol=tol)
+    assert_np_equal(v53.numpy()[0], 2 * input.numpy()[12], tol=tol)
+    assert_np_equal(v54.numpy()[0], 2 * input.numpy()[13], tol=tol)
+
+
+def test_anon_type_instance(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    def check_scalar_init(
+        input: wp.array(dtype=wptype),
+        output: wp.array(dtype=wptype),
+    ):
+        v2result = wp.vector(input[0], length=2)
+        v3result = wp.vector(input[1], length=3)
+        v4result = wp.vector(input[2], length=4)
+        v5result = wp.vector(input[3], length=5)
+
+        idx = 0
+        for i in range(2):
+            output[idx] = wptype(2) * v2result[i]
+            idx = idx + 1
+        for i in range(3):
+            output[idx] = wptype(2) * v3result[i]
+            idx = idx + 1
+        for i in range(4):
+            output[idx] = wptype(2) * v4result[i]
+            idx = idx + 1
+        for i in range(5):
+            output[idx] = wptype(2) * v5result[i]
+            idx = idx + 1
+
+    def check_component_init(
+        input: wp.array(dtype=wptype),
+        output: wp.array(dtype=wptype),
+    ):
+        v2result = wp.vector(input[0], input[1])
+        v3result = wp.vector(input[2], input[3], input[4])
+        v4result = wp.vector(input[5], input[6], input[7], input[8])
+        v5result = wp.vector(input[9], input[10], input[11], input[12], input[13])
+
+        idx = 0
+        for i in range(2):
+            output[idx] = wptype(2) * v2result[i]
+            idx = idx + 1
+        for i in range(3):
+            output[idx] = wptype(2) * v3result[i]
+            idx = idx + 1
+        for i in range(4):
+            output[idx] = wptype(2) * v4result[i]
+            idx = idx + 1
+        for i in range(5):
+            output[idx] = wptype(2) * v5result[i]
+            idx = idx + 1
+
+    scalar_kernel = getkernel(check_scalar_init, suffix=dtype.__name__)
+    component_kernel = getkernel(check_component_init, suffix=dtype.__name__)
+    output_select_kernel = get_select_kernel(wptype)
+
+    if register_kernels:
+        return
+
+    input = wp.array(randvals(rng, [4], dtype), requires_grad=True, device=device)
+    output = wp.zeros(2 + 3 + 4 + 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+
+    assert_np_equal(output.numpy()[:2], 2 * np.array([input.numpy()[0]] * 2), tol=1.0e-6)
+    assert_np_equal(output.numpy()[2:5], 2 * np.array([input.numpy()[1]] * 3), tol=1.0e-6)
+    assert_np_equal(output.numpy()[5:9], 2 * np.array([input.numpy()[2]] * 4), tol=1.0e-6)
+    assert_np_equal(output.numpy()[9:], 2 * np.array([input.numpy()[3]] * 5), tol=1.0e-6)
+
+    if dtype in np_float_types:
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for i in range(len(output)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(scalar_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
+
+            tape.backward(loss=out)
+            expected = np.zeros_like(input.numpy())
+            if i < 2:
+                expected[0] = 2
+            elif i < 5:
+                expected[1] = 2
+            elif i < 9:
+                expected[2] = 2
+            else:
+                expected[3] = 2
+
+            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
+
+            tape.reset()
+            tape.zero()
+
+    input = wp.array(randvals(rng, [2 + 3 + 4 + 5], dtype), requires_grad=True, device=device)
+    output = wp.zeros(2 + 3 + 4 + 5, dtype=wptype, requires_grad=True, device=device)
+
+    wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+
+    assert_np_equal(output.numpy(), 2 * input.numpy(), tol=1.0e-6)
+
+    if dtype in np_float_types:
+        out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+        for i in range(len(output)):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(component_kernel, dim=1, inputs=[input], outputs=[output], device=device)
+                wp.launch(output_select_kernel, dim=1, inputs=[output, i], outputs=[out], device=device)
+
+            tape.backward(loss=out)
+            expected = np.zeros_like(input.numpy())
+            expected[i] = 2
+
+            assert_np_equal(tape.gradients[input].numpy(), expected, tol=tol)
+
+            tape.reset()
+            tape.zero()
+
+
+def test_indexing(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_indexing(
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        # multiply outputs by 2 so we've got something to backpropagate:
+        v20[0] = wptype(2) * v2[0][0]
+        v21[0] = wptype(2) * v2[0][1]
+
+        v30[0] = wptype(2) * v3[0][0]
+        v31[0] = wptype(2) * v3[0][1]
+        v32[0] = wptype(2) * v3[0][2]
+
+        v40[0] = wptype(2) * v4[0][0]
+        v41[0] = wptype(2) * v4[0][1]
+        v42[0] = wptype(2) * v4[0][2]
+        v43[0] = wptype(2) * v4[0][3]
+
+        v50[0] = wptype(2) * v5[0][0]
+        v51[0] = wptype(2) * v5[0][1]
+        v52[0] = wptype(2) * v5[0][2]
+        v53[0] = wptype(2) * v5[0][3]
+        v54[0] = wptype(2) * v5[0][4]
+
+    kernel = getkernel(check_indexing, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[v2, v3, v4, v5],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
+            tape.backward(loss=l)
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
+            expected_grads = np.zeros_like(allgrads)
+            expected_grads[i] = 2
+            assert_np_equal(allgrads, expected_grads, tol=tol)
+            tape.zero()
+
+    assert_np_equal(v20.numpy()[0], 2.0 * v2.numpy()[0, 0], tol=tol)
+    assert_np_equal(v21.numpy()[0], 2.0 * v2.numpy()[0, 1], tol=tol)
+    assert_np_equal(v30.numpy()[0], 2.0 * v3.numpy()[0, 0], tol=tol)
+    assert_np_equal(v31.numpy()[0], 2.0 * v3.numpy()[0, 1], tol=tol)
+    assert_np_equal(v32.numpy()[0], 2.0 * v3.numpy()[0, 2], tol=tol)
+    assert_np_equal(v40.numpy()[0], 2.0 * v4.numpy()[0, 0], tol=tol)
+    assert_np_equal(v41.numpy()[0], 2.0 * v4.numpy()[0, 1], tol=tol)
+    assert_np_equal(v42.numpy()[0], 2.0 * v4.numpy()[0, 2], tol=tol)
+    assert_np_equal(v43.numpy()[0], 2.0 * v4.numpy()[0, 3], tol=tol)
+    assert_np_equal(v50.numpy()[0], 2.0 * v5.numpy()[0, 0], tol=tol)
+    assert_np_equal(v51.numpy()[0], 2.0 * v5.numpy()[0, 1], tol=tol)
+    assert_np_equal(v52.numpy()[0], 2.0 * v5.numpy()[0, 2], tol=tol)
+    assert_np_equal(v53.numpy()[0], 2.0 * v5.numpy()[0, 3], tol=tol)
+    assert_np_equal(v54.numpy()[0], 2.0 * v5.numpy()[0, 4], tol=tol)
+
+
+def test_equality(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_equality(
+        v20: wp.array(dtype=vec2),
+        v21: wp.array(dtype=vec2),
+        v22: wp.array(dtype=vec2),
+        v30: wp.array(dtype=vec3),
+        v31: wp.array(dtype=vec3),
+        v32: wp.array(dtype=vec3),
+        v33: wp.array(dtype=vec3),
+        v40: wp.array(dtype=vec4),
+        v41: wp.array(dtype=vec4),
+        v42: wp.array(dtype=vec4),
+        v43: wp.array(dtype=vec4),
+        v44: wp.array(dtype=vec4),
+        v50: wp.array(dtype=vec5),
+        v51: wp.array(dtype=vec5),
+        v52: wp.array(dtype=vec5),
+        v53: wp.array(dtype=vec5),
+        v54: wp.array(dtype=vec5),
+        v55: wp.array(dtype=vec5),
+    ):
+        wp.expect_eq(v20[0], v20[0])
+        wp.expect_neq(v21[0], v20[0])
+        wp.expect_neq(v22[0], v20[0])
+
+        wp.expect_eq(v30[0], v30[0])
+        wp.expect_neq(v31[0], v30[0])
+        wp.expect_neq(v32[0], v30[0])
+        wp.expect_neq(v33[0], v30[0])
+
+        wp.expect_eq(v40[0], v40[0])
+        wp.expect_neq(v41[0], v40[0])
+        wp.expect_neq(v42[0], v40[0])
+        wp.expect_neq(v43[0], v40[0])
+        wp.expect_neq(v44[0], v40[0])
+
+        wp.expect_eq(v50[0], v50[0])
+        wp.expect_neq(v51[0], v50[0])
+        wp.expect_neq(v52[0], v50[0])
+        wp.expect_neq(v53[0], v50[0])
+        wp.expect_neq(v54[0], v50[0])
+        wp.expect_neq(v55[0], v50[0])
+
+    kernel = getkernel(check_equality, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v20 = wp.array([1.0, 2.0], dtype=vec2, requires_grad=True, device=device)
+    v21 = wp.array([1.0, 3.0], dtype=vec2, requires_grad=True, device=device)
+    v22 = wp.array([3.0, 2.0], dtype=vec2, requires_grad=True, device=device)
+
+    v30 = wp.array([1.0, 2.0, 3.0], dtype=vec3, requires_grad=True, device=device)
+    v31 = wp.array([-1.0, 2.0, 3.0], dtype=vec3, requires_grad=True, device=device)
+    v32 = wp.array([1.0, -2.0, 3.0], dtype=vec3, requires_grad=True, device=device)
+    v33 = wp.array([1.0, 2.0, -3.0], dtype=vec3, requires_grad=True, device=device)
+
+    v40 = wp.array([1.0, 2.0, 3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
+    v41 = wp.array([-1.0, 2.0, 3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
+    v42 = wp.array([1.0, -2.0, 3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
+    v43 = wp.array([1.0, 2.0, -3.0, 4.0], dtype=vec4, requires_grad=True, device=device)
+    v44 = wp.array([1.0, 2.0, 3.0, -4.0], dtype=vec4, requires_grad=True, device=device)
+
+    v50 = wp.array([1.0, 2.0, 3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
+    v51 = wp.array([-1.0, 2.0, 3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
+    v52 = wp.array([1.0, -2.0, 3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
+    v53 = wp.array([1.0, 2.0, -3.0, 4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
+    v54 = wp.array([1.0, 2.0, 3.0, -4.0, 5.0], dtype=vec5, requires_grad=True, device=device)
+    v55 = wp.array([1.0, 2.0, 3.0, 4.0, -5.0], dtype=vec5, requires_grad=True, device=device)
+    wp.launch(
+        kernel,
+        dim=1,
+        inputs=[
+            v20,
+            v21,
+            v22,
+            v30,
+            v31,
+            v32,
+            v33,
+            v40,
+            v41,
+            v42,
+            v43,
+            v44,
+            v50,
+            v51,
+            v52,
+            v53,
+            v54,
+            v55,
+        ],
+        outputs=[],
+        device=device,
+    )
+
+
+def test_scalar_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_mul(
+        s: wp.array(dtype=wptype),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = s[0] * v2[0]
+        v3result = s[0] * v3[0]
+        v4result = s[0] * v4[0]
+        v5result = s[0] * v5[0]
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    kernel = getkernel(check_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    assert_np_equal(v20.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 0], tol=tol)
+    assert_np_equal(v21.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 1], tol=tol)
+
+    assert_np_equal(v30.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v31.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v32.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 2], tol=10 * tol)
+
+    assert_np_equal(v40.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v41.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v42.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 2], tol=10 * tol)
+    assert_np_equal(v43.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 3], tol=10 * tol)
+
+    assert_np_equal(v50.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v51.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v52.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 2], tol=10 * tol)
+    assert_np_equal(v53.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 3], tol=10 * tol)
+    assert_np_equal(v54.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 4], tol=10 * tol)
+
+    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43]):
+            tape.backward(loss=l)
+            sgrad = tape.gradients[s].numpy()[0]
+            assert_np_equal(sgrad, 2 * incmps[i], tol=10 * tol)
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4]])
+            expected_grads = np.zeros_like(allgrads)
+            expected_grads[i] = s.numpy()[0] * 2
+            assert_np_equal(allgrads, expected_grads, tol=10 * tol)
+            tape.zero()
+
+
+def test_scalar_multiplication_rightmul(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_rightmul(
+        s: wp.array(dtype=wptype),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = v2[0] * s[0]
+        v3result = v3[0] * s[0]
+        v4result = v4[0] * s[0]
+        v5result = v5[0] * s[0]
+
+        # multiply outputs by 2 so we've got something to backpropagate:
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    kernel = getkernel(check_rightmul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    assert_np_equal(v20.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 0], tol=tol)
+    assert_np_equal(v21.numpy()[0], 2 * s.numpy()[0] * v2.numpy()[0, 1], tol=tol)
+
+    assert_np_equal(v30.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v31.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v32.numpy()[0], 2 * s.numpy()[0] * v3.numpy()[0, 2], tol=10 * tol)
+
+    assert_np_equal(v40.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v41.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v42.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 2], tol=10 * tol)
+    assert_np_equal(v43.numpy()[0], 2 * s.numpy()[0] * v4.numpy()[0, 3], tol=10 * tol)
+
+    assert_np_equal(v50.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v51.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v52.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 2], tol=10 * tol)
+    assert_np_equal(v53.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 3], tol=10 * tol)
+    assert_np_equal(v54.numpy()[0], 2 * s.numpy()[0] * v5.numpy()[0, 4], tol=10 * tol)
+
+    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43]):
+            tape.backward(loss=l)
+            sgrad = tape.gradients[s].numpy()[0]
+            assert_np_equal(sgrad, 2 * incmps[i], tol=10 * tol)
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4]])
+            expected_grads = np.zeros_like(allgrads)
+            expected_grads[i] = s.numpy()[0] * 2
+            assert_np_equal(allgrads, expected_grads, tol=10 * tol)
+            tape.zero()
+
+
+def test_cw_multiplication(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_cw_mul(
+        s2: wp.array(dtype=vec2),
+        s3: wp.array(dtype=vec3),
+        s4: wp.array(dtype=vec4),
+        s5: wp.array(dtype=vec5),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = wp.cw_mul(s2[0], v2[0])
+        v3result = wp.cw_mul(s3[0], v3[0])
+        v4result = wp.cw_mul(s4[0], v4[0])
+        v5result = wp.cw_mul(s5[0], v5[0])
+
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    kernel = getkernel(check_cw_mul, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s2,
+                s3,
+                s4,
+                s5,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    assert_np_equal(v20.numpy()[0], 2 * s2.numpy()[0, 0] * v2.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v21.numpy()[0], 2 * s2.numpy()[0, 1] * v2.numpy()[0, 1], tol=10 * tol)
+
+    assert_np_equal(v30.numpy()[0], 2 * s3.numpy()[0, 0] * v3.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v31.numpy()[0], 2 * s3.numpy()[0, 1] * v3.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v32.numpy()[0], 2 * s3.numpy()[0, 2] * v3.numpy()[0, 2], tol=10 * tol)
+
+    assert_np_equal(v40.numpy()[0], 2 * s4.numpy()[0, 0] * v4.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v41.numpy()[0], 2 * s4.numpy()[0, 1] * v4.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v42.numpy()[0], 2 * s4.numpy()[0, 2] * v4.numpy()[0, 2], tol=10 * tol)
+    assert_np_equal(v43.numpy()[0], 2 * s4.numpy()[0, 3] * v4.numpy()[0, 3], tol=10 * tol)
+
+    assert_np_equal(v50.numpy()[0], 2 * s5.numpy()[0, 0] * v5.numpy()[0, 0], tol=10 * tol)
+    assert_np_equal(v51.numpy()[0], 2 * s5.numpy()[0, 1] * v5.numpy()[0, 1], tol=10 * tol)
+    assert_np_equal(v52.numpy()[0], 2 * s5.numpy()[0, 2] * v5.numpy()[0, 2], tol=10 * tol)
+    assert_np_equal(v53.numpy()[0], 2 * s5.numpy()[0, 3] * v5.numpy()[0, 3], tol=10 * tol)
+    assert_np_equal(v54.numpy()[0], 2 * s5.numpy()[0, 4] * v5.numpy()[0, 4], tol=10 * tol)
+
+    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
+    scmps = np.concatenate([v.numpy()[0] for v in [s2, s3, s4, s5]])
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
+            tape.backward(loss=l)
+            sgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [s2, s3, s4, s5]])
+            expected_grads = np.zeros_like(sgrads)
+            expected_grads[i] = incmps[i] * 2
+            assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
+            expected_grads = np.zeros_like(allgrads)
+            expected_grads[i] = scmps[i] * 2
+            assert_np_equal(allgrads, expected_grads, tol=10 * tol)
+
+            tape.zero()
+
+
+def test_scalar_division(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_div(
+        s: wp.array(dtype=wptype),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = v2[0] / s[0]
+        v3result = v3[0] / s[0]
+        v4result = v4[0] / s[0]
+        v5result = v5[0] / s[0]
+
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    kernel = getkernel(check_div, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s = wp.array(randvals(rng, [1], dtype), requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    if dtype in np_int_types:
+        assert_np_equal(v20.numpy()[0], 2 * (v2.numpy()[0, 0] // (s.numpy()[0])), tol=tol)
+        assert_np_equal(v21.numpy()[0], 2 * (v2.numpy()[0, 1] // (s.numpy()[0])), tol=tol)
+
+        assert_np_equal(v30.numpy()[0], 2 * (v3.numpy()[0, 0] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v31.numpy()[0], 2 * (v3.numpy()[0, 1] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v32.numpy()[0], 2 * (v3.numpy()[0, 2] // (s.numpy()[0])), tol=10 * tol)
+
+        assert_np_equal(v40.numpy()[0], 2 * (v4.numpy()[0, 0] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v41.numpy()[0], 2 * (v4.numpy()[0, 1] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v42.numpy()[0], 2 * (v4.numpy()[0, 2] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v43.numpy()[0], 2 * (v4.numpy()[0, 3] // (s.numpy()[0])), tol=10 * tol)
+
+        assert_np_equal(v50.numpy()[0], 2 * (v5.numpy()[0, 0] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v51.numpy()[0], 2 * (v5.numpy()[0, 1] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v52.numpy()[0], 2 * (v5.numpy()[0, 2] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v53.numpy()[0], 2 * (v5.numpy()[0, 3] // (s.numpy()[0])), tol=10 * tol)
+        assert_np_equal(v54.numpy()[0], 2 * (v5.numpy()[0, 4] // (s.numpy()[0])), tol=10 * tol)
+
+    else:
+        assert_np_equal(v20.numpy()[0], 2 * v2.numpy()[0, 0] / (s.numpy()[0]), tol=tol)
+        assert_np_equal(v21.numpy()[0], 2 * v2.numpy()[0, 1] / (s.numpy()[0]), tol=tol)
+
+        assert_np_equal(v30.numpy()[0], 2 * v3.numpy()[0, 0] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v31.numpy()[0], 2 * v3.numpy()[0, 1] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v32.numpy()[0], 2 * v3.numpy()[0, 2] / (s.numpy()[0]), tol=10 * tol)
+
+        assert_np_equal(v40.numpy()[0], 2 * v4.numpy()[0, 0] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v41.numpy()[0], 2 * v4.numpy()[0, 1] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v42.numpy()[0], 2 * v4.numpy()[0, 2] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v43.numpy()[0], 2 * v4.numpy()[0, 3] / (s.numpy()[0]), tol=10 * tol)
+
+        assert_np_equal(v50.numpy()[0], 2 * v5.numpy()[0, 0] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v51.numpy()[0], 2 * v5.numpy()[0, 1] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v52.numpy()[0], 2 * v5.numpy()[0, 2] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v53.numpy()[0], 2 * v5.numpy()[0, 3] / (s.numpy()[0]), tol=10 * tol)
+        assert_np_equal(v54.numpy()[0], 2 * v5.numpy()[0, 4] / (s.numpy()[0]), tol=10 * tol)
+
+    incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
+            tape.backward(loss=l)
+            sgrad = tape.gradients[s].numpy()[0]
+
+            # d/ds v/s = -v/s^2
+            assert_np_equal(sgrad, -2 * incmps[i] / (s.numpy()[0] * s.numpy()[0]), tol=10 * tol)
+
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
+            expected_grads = np.zeros_like(allgrads)
+            expected_grads[i] = 2 / s.numpy()[0]
+
+            # d/dv v/s = 1/s
+            assert_np_equal(allgrads, expected_grads, tol=tol)
+            tape.zero()
+
+
+def test_cw_division(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_cw_div(
+        s2: wp.array(dtype=vec2),
+        s3: wp.array(dtype=vec3),
+        s4: wp.array(dtype=vec4),
+        s5: wp.array(dtype=vec5),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = wp.cw_div(v2[0], s2[0])
+        v3result = wp.cw_div(v3[0], s3[0])
+        v4result = wp.cw_div(v4[0], s4[0])
+        v5result = wp.cw_div(v5[0], s5[0])
+
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    kernel = getkernel(check_cw_div, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s2,
+                s3,
+                s4,
+                s5,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    if dtype in np_int_types:
+        assert_np_equal(v20.numpy()[0], 2 * (v2.numpy()[0, 0] // s2.numpy()[0, 0]), tol=tol)
+        assert_np_equal(v21.numpy()[0], 2 * (v2.numpy()[0, 1] // s2.numpy()[0, 1]), tol=tol)
+
+        assert_np_equal(v30.numpy()[0], 2 * (v3.numpy()[0, 0] // s3.numpy()[0, 0]), tol=tol)
+        assert_np_equal(v31.numpy()[0], 2 * (v3.numpy()[0, 1] // s3.numpy()[0, 1]), tol=tol)
+        assert_np_equal(v32.numpy()[0], 2 * (v3.numpy()[0, 2] // s3.numpy()[0, 2]), tol=tol)
+
+        assert_np_equal(v40.numpy()[0], 2 * (v4.numpy()[0, 0] // s4.numpy()[0, 0]), tol=tol)
+        assert_np_equal(v41.numpy()[0], 2 * (v4.numpy()[0, 1] // s4.numpy()[0, 1]), tol=tol)
+        assert_np_equal(v42.numpy()[0], 2 * (v4.numpy()[0, 2] // s4.numpy()[0, 2]), tol=tol)
+        assert_np_equal(v43.numpy()[0], 2 * (v4.numpy()[0, 3] // s4.numpy()[0, 3]), tol=tol)
+
+        assert_np_equal(v50.numpy()[0], 2 * (v5.numpy()[0, 0] // s5.numpy()[0, 0]), tol=tol)
+        assert_np_equal(v51.numpy()[0], 2 * (v5.numpy()[0, 1] // s5.numpy()[0, 1]), tol=tol)
+        assert_np_equal(v52.numpy()[0], 2 * (v5.numpy()[0, 2] // s5.numpy()[0, 2]), tol=tol)
+        assert_np_equal(v53.numpy()[0], 2 * (v5.numpy()[0, 3] // s5.numpy()[0, 3]), tol=tol)
+        assert_np_equal(v54.numpy()[0], 2 * (v5.numpy()[0, 4] // s5.numpy()[0, 4]), tol=tol)
+    else:
+        assert_np_equal(v20.numpy()[0], 2 * v2.numpy()[0, 0] / s2.numpy()[0, 0], tol=tol)
+        assert_np_equal(v21.numpy()[0], 2 * v2.numpy()[0, 1] / s2.numpy()[0, 1], tol=tol)
+
+        assert_np_equal(v30.numpy()[0], 2 * v3.numpy()[0, 0] / s3.numpy()[0, 0], tol=tol)
+        assert_np_equal(v31.numpy()[0], 2 * v3.numpy()[0, 1] / s3.numpy()[0, 1], tol=tol)
+        assert_np_equal(v32.numpy()[0], 2 * v3.numpy()[0, 2] / s3.numpy()[0, 2], tol=tol)
+
+        assert_np_equal(v40.numpy()[0], 2 * v4.numpy()[0, 0] / s4.numpy()[0, 0], tol=tol)
+        assert_np_equal(v41.numpy()[0], 2 * v4.numpy()[0, 1] / s4.numpy()[0, 1], tol=tol)
+        assert_np_equal(v42.numpy()[0], 2 * v4.numpy()[0, 2] / s4.numpy()[0, 2], tol=tol)
+        assert_np_equal(v43.numpy()[0], 2 * v4.numpy()[0, 3] / s4.numpy()[0, 3], tol=tol)
+
+        assert_np_equal(v50.numpy()[0], 2 * v5.numpy()[0, 0] / s5.numpy()[0, 0], tol=tol)
+        assert_np_equal(v51.numpy()[0], 2 * v5.numpy()[0, 1] / s5.numpy()[0, 1], tol=tol)
+        assert_np_equal(v52.numpy()[0], 2 * v5.numpy()[0, 2] / s5.numpy()[0, 2], tol=tol)
+        assert_np_equal(v53.numpy()[0], 2 * v5.numpy()[0, 3] / s5.numpy()[0, 3], tol=tol)
+        assert_np_equal(v54.numpy()[0], 2 * v5.numpy()[0, 4] / s5.numpy()[0, 4], tol=tol)
+
+    if dtype in np_float_types:
+        incmps = np.concatenate([v.numpy()[0] for v in [v2, v3, v4, v5]])
+        scmps = np.concatenate([v.numpy()[0] for v in [s2, s3, s4, s5]])
+
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
+            tape.backward(loss=l)
+            sgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [s2, s3, s4, s5]])
+            expected_grads = np.zeros_like(sgrads)
+
+            # d/ds v/s = -v/s^2
+            expected_grads[i] = -incmps[i] * 2 / (scmps[i] * scmps[i])
+            assert_np_equal(sgrads, expected_grads, tol=20 * tol)
+
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
+            expected_grads = np.zeros_like(allgrads)
+
+            # d/dv v/s = 1/s
+            expected_grads[i] = 2 / scmps[i]
+            assert_np_equal(allgrads, expected_grads, tol=tol)
+
+            tape.zero()
+
+
+def test_addition(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 5.0e-3,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_add(
+        s2: wp.array(dtype=vec2),
+        s3: wp.array(dtype=vec3),
+        s4: wp.array(dtype=vec4),
+        s5: wp.array(dtype=vec5),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        v20: wp.array(dtype=wptype),
+        v21: wp.array(dtype=wptype),
+        v30: wp.array(dtype=wptype),
+        v31: wp.array(dtype=wptype),
+        v32: wp.array(dtype=wptype),
+        v40: wp.array(dtype=wptype),
+        v41: wp.array(dtype=wptype),
+        v42: wp.array(dtype=wptype),
+        v43: wp.array(dtype=wptype),
+        v50: wp.array(dtype=wptype),
+        v51: wp.array(dtype=wptype),
+        v52: wp.array(dtype=wptype),
+        v53: wp.array(dtype=wptype),
+        v54: wp.array(dtype=wptype),
+    ):
+        v2result = v2[0] + s2[0]
+        v3result = v3[0] + s3[0]
+        v4result = v4[0] + s4[0]
+        v5result = v5[0] + s5[0]
+
+        v20[0] = wptype(2) * v2result[0]
+        v21[0] = wptype(2) * v2result[1]
+
+        v30[0] = wptype(2) * v3result[0]
+        v31[0] = wptype(2) * v3result[1]
+        v32[0] = wptype(2) * v3result[2]
+
+        v40[0] = wptype(2) * v4result[0]
+        v41[0] = wptype(2) * v4result[1]
+        v42[0] = wptype(2) * v4result[2]
+        v43[0] = wptype(2) * v4result[3]
+
+        v50[0] = wptype(2) * v5result[0]
+        v51[0] = wptype(2) * v5result[1]
+        v52[0] = wptype(2) * v5result[2]
+        v53[0] = wptype(2) * v5result[3]
+        v54[0] = wptype(2) * v5result[4]
+
+    kernel = getkernel(check_add, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v20 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v21 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v30 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v31 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v32 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v40 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v41 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v42 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v43 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v50 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v51 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v52 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v53 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    v54 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s2,
+                s3,
+                s4,
+                s5,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54],
+            device=device,
+        )
+
+    assert_np_equal(v20.numpy()[0], 2 * (v2.numpy()[0, 0] + s2.numpy()[0, 0]), tol=tol)
+    assert_np_equal(v21.numpy()[0], 2 * (v2.numpy()[0, 1] + s2.numpy()[0, 1]), tol=tol)
+
+    assert_np_equal(v30.numpy()[0], 2 * (v3.numpy()[0, 0] + s3.numpy()[0, 0]), tol=tol)
+    assert_np_equal(v31.numpy()[0], 2 * (v3.numpy()[0, 1] + s3.numpy()[0, 1]), tol=tol)
+    assert_np_equal(v32.numpy()[0], 2 * (v3.numpy()[0, 2] + s3.numpy()[0, 2]), tol=tol)
+
+    assert_np_equal(v40.numpy()[0], 2 * (v4.numpy()[0, 0] + s4.numpy()[0, 0]), tol=tol)
+    assert_np_equal(v41.numpy()[0], 2 * (v4.numpy()[0, 1] + s4.numpy()[0, 1]), tol=tol)
+    assert_np_equal(v42.numpy()[0], 2 * (v4.numpy()[0, 2] + s4.numpy()[0, 2]), tol=tol)
+    assert_np_equal(v43.numpy()[0], 2 * (v4.numpy()[0, 3] + s4.numpy()[0, 3]), tol=tol)
+
+    assert_np_equal(v50.numpy()[0], 2 * (v5.numpy()[0, 0] + s5.numpy()[0, 0]), tol=tol)
+    assert_np_equal(v51.numpy()[0], 2 * (v5.numpy()[0, 1] + s5.numpy()[0, 1]), tol=tol)
+    assert_np_equal(v52.numpy()[0], 2 * (v5.numpy()[0, 2] + s5.numpy()[0, 2]), tol=tol)
+    assert_np_equal(v53.numpy()[0], 2 * (v5.numpy()[0, 3] + s5.numpy()[0, 3]), tol=tol)
+    assert_np_equal(v54.numpy()[0], 2 * (v5.numpy()[0, 4] + s5.numpy()[0, 4]), tol=2 * tol)
+
+    if dtype in np_float_types:
+        for i, l in enumerate([v20, v21, v30, v31, v32, v40, v41, v42, v43, v50, v51, v52, v53, v54]):
+            tape.backward(loss=l)
+            sgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [s2, s3, s4, s5]])
+            expected_grads = np.zeros_like(sgrads)
+
+            expected_grads[i] = 2
+            assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+            allgrads = np.concatenate([tape.gradients[v].numpy()[0] for v in [v2, v3, v4, v5]])
+            assert_np_equal(allgrads, expected_grads, tol=tol)
+
+            tape.zero()
+
+
+def test_dotproduct(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    tol = {
+        np.float16: 1.0e-2,
+        np.float32: 1.0e-6,
+        np.float64: 1.0e-8,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    def check_dot(
+        s2: wp.array(dtype=vec2),
+        s3: wp.array(dtype=vec3),
+        s4: wp.array(dtype=vec4),
+        s5: wp.array(dtype=vec5),
+        v2: wp.array(dtype=vec2),
+        v3: wp.array(dtype=vec3),
+        v4: wp.array(dtype=vec4),
+        v5: wp.array(dtype=vec5),
+        dot2: wp.array(dtype=wptype),
+        dot3: wp.array(dtype=wptype),
+        dot4: wp.array(dtype=wptype),
+        dot5: wp.array(dtype=wptype),
+    ):
+        dot2[0] = wptype(2) * wp.dot(v2[0], s2[0])
+        dot3[0] = wptype(2) * wp.dot(v3[0], s3[0])
+        dot4[0] = wptype(2) * wp.dot(v4[0], s4[0])
+        dot5[0] = wptype(2) * wp.dot(v5[0], s5[0])
+
+    kernel = getkernel(check_dot, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    s2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    s3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    s4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    s5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    v2 = wp.array(randvals(rng, (1, 2), dtype), dtype=vec2, requires_grad=True, device=device)
+    v3 = wp.array(randvals(rng, (1, 3), dtype), dtype=vec3, requires_grad=True, device=device)
+    v4 = wp.array(randvals(rng, (1, 4), dtype), dtype=vec4, requires_grad=True, device=device)
+    v5 = wp.array(randvals(rng, (1, 5), dtype), dtype=vec5, requires_grad=True, device=device)
+    dot2 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    dot3 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    dot4 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    dot5 = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    tape = wp.Tape()
+    with tape:
+        wp.launch(
+            kernel,
+            dim=1,
+            inputs=[
+                s2,
+                s3,
+                s4,
+                s5,
+                v2,
+                v3,
+                v4,
+                v5,
+            ],
+            outputs=[dot2, dot3, dot4, dot5],
+            device=device,
+        )
+
+    assert_np_equal(dot2.numpy()[0], 2.0 * (v2.numpy() * s2.numpy()).sum(), tol=10 * tol)
+    assert_np_equal(dot3.numpy()[0], 2.0 * (v3.numpy() * s3.numpy()).sum(), tol=10 * tol)
+    assert_np_equal(dot4.numpy()[0], 2.0 * (v4.numpy() * s4.numpy()).sum(), tol=10 * tol)
+    assert_np_equal(dot5.numpy()[0], 2.0 * (v5.numpy() * s5.numpy()).sum(), tol=10 * tol)
+
+    if dtype in np_float_types:
+        tape.backward(loss=dot2)
+        sgrads = tape.gradients[s2].numpy()[0]
+        expected_grads = 2.0 * v2.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v2].numpy()[0]
+        expected_grads = 2.0 * s2.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=tol)
+
+        tape.zero()
+
+        tape.backward(loss=dot3)
+        sgrads = tape.gradients[s3].numpy()[0]
+        expected_grads = 2.0 * v3.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v3].numpy()[0]
+        expected_grads = 2.0 * s3.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=tol)
+
+        tape.zero()
+
+        tape.backward(loss=dot4)
+        sgrads = tape.gradients[s4].numpy()[0]
+        expected_grads = 2.0 * v4.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v4].numpy()[0]
+        expected_grads = 2.0 * s4.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=tol)
+
+        tape.zero()
+
+        tape.backward(loss=dot5)
+        sgrads = tape.gradients[s5].numpy()[0]
+        expected_grads = 2.0 * v5.numpy()[0]
+        assert_np_equal(sgrads, expected_grads, tol=10 * tol)
+
+        vgrads = tape.gradients[v5].numpy()[0]
+        expected_grads = 2.0 * s5.numpy()[0]
+        assert_np_equal(vgrads, expected_grads, tol=10 * tol)
+
+        tape.zero()
+
+
+def test_equivalent_types(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+
+    # vector types
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    # vector types equivalent to the above
+    vec2_equiv = wp.types.vector(length=2, dtype=wptype)
+    vec3_equiv = wp.types.vector(length=3, dtype=wptype)
+    vec4_equiv = wp.types.vector(length=4, dtype=wptype)
+    vec5_equiv = wp.types.vector(length=5, dtype=wptype)
+
+    # declare kernel with original types
+    def check_equivalence(
+        v2: vec2,
+        v3: vec3,
+        v4: vec4,
+        v5: vec5,
+    ):
+        wp.expect_eq(v2, vec2(wptype(1), wptype(2)))
+        wp.expect_eq(v3, vec3(wptype(1), wptype(2), wptype(3)))
+        wp.expect_eq(v4, vec4(wptype(1), wptype(2), wptype(3), wptype(4)))
+        wp.expect_eq(v5, vec5(wptype(1), wptype(2), wptype(3), wptype(4), wptype(5)))
+
+        wp.expect_eq(v2, vec2_equiv(wptype(1), wptype(2)))
+        wp.expect_eq(v3, vec3_equiv(wptype(1), wptype(2), wptype(3)))
+        wp.expect_eq(v4, vec4_equiv(wptype(1), wptype(2), wptype(3), wptype(4)))
+        wp.expect_eq(v5, vec5_equiv(wptype(1), wptype(2), wptype(3), wptype(4), wptype(5)))
+
+    kernel = getkernel(check_equivalence, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    # call kernel with equivalent types
+    v2 = vec2_equiv(1, 2)
+    v3 = vec3_equiv(1, 2, 3)
+    v4 = vec4_equiv(1, 2, 3, 4)
+    v5 = vec5_equiv(1, 2, 3, 4, 5)
+
+    wp.launch(kernel, dim=1, inputs=[v2, v3, v4, v5], device=device)
+
+
+def test_conversions(test, device, dtype, register_kernels=False):
+    def check_vectors_equal(
+        v0: wp.vec3,
+        v1: wp.vec3,
+        v2: wp.vec3,
+        v3: wp.vec3,
+    ):
+        wp.expect_eq(v1, v0)
+        wp.expect_eq(v2, v0)
+        wp.expect_eq(v3, v0)
+
+    kernel = getkernel(check_vectors_equal, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    v0 = wp.vec3(1, 2, 3)
+
+    # test explicit conversions - constructing vectors from different containers
+    v1 = wp.vec3((1, 2, 3))
+    v2 = wp.vec3([1, 2, 3])
+    v3 = wp.vec3(np.array([1, 2, 3], dtype=dtype))
+
+    wp.launch(kernel, dim=1, inputs=[v0, v1, v2, v3], device=device)
+
+    # test implicit conversions - passing different containers as vectors to wp.launch()
+    v1 = (1, 2, 3)
+    v2 = [1, 2, 3]
+    v3 = np.array([1, 2, 3], dtype=dtype)
+
+    wp.launch(kernel, dim=1, inputs=[v0, v1, v2, v3], device=device)
+
+
+def test_constants(test, device, dtype, register_kernels=False):
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    cv2 = wp.constant(vec2(1, 2))
+    cv3 = wp.constant(vec3(1, 2, 3))
+    cv4 = wp.constant(vec4(1, 2, 3, 4))
+    cv5 = wp.constant(vec5(1, 2, 3, 4, 5))
+
+    def check_vector_constants():
+        wp.expect_eq(cv2, vec2(wptype(1), wptype(2)))
+        wp.expect_eq(cv3, vec3(wptype(1), wptype(2), wptype(3)))
+        wp.expect_eq(cv4, vec4(wptype(1), wptype(2), wptype(3), wptype(4)))
+        wp.expect_eq(cv5, vec5(wptype(1), wptype(2), wptype(3), wptype(4), wptype(5)))
+
+    kernel = getkernel(check_vector_constants, suffix=dtype.__name__)
+
+    if register_kernels:
+        return
+
+    wp.launch(kernel, dim=1, inputs=[])
+
+
+def test_minmax(test, device, dtype, register_kernels=False):
+    rng = np.random.default_rng(123)
+
+    # \TODO: not quite sure why, but the numbers are off for 16 bit float
+    # on the cpu (but not cuda). This is probably just the sketchy float16
+    # arithmetic I implemented to get all this stuff working, so
+    # hopefully that can be fixed when we do that correctly.
+    tol = {
+        np.float16: 1.0e-2,
+    }.get(dtype, 0)
+
+    wptype = wp.types.np_dtype_to_warp_type[np.dtype(dtype)]
+    vec2 = wp.types.vector(length=2, dtype=wptype)
+    vec3 = wp.types.vector(length=3, dtype=wptype)
+    vec4 = wp.types.vector(length=4, dtype=wptype)
+    vec5 = wp.types.vector(length=5, dtype=wptype)
+
+    # \TODO: Also not quite sure why: this kernel compiles incredibly
+    # slowly though...
+    def check_vec_min_max(
+        a: wp.array(dtype=wptype, ndim=2),
+        b: wp.array(dtype=wptype, ndim=2),
+        mins: wp.array(dtype=wptype, ndim=2),
+        maxs: wp.array(dtype=wptype, ndim=2),
+    ):
+        for i in range(10):
+            # multiplying by 2 so we've got something to backpropagate:
+            a2read = vec2(a[i, 0], a[i, 1])
+            b2read = vec2(b[i, 0], b[i, 1])
+            c2 = wptype(2) * wp.min(a2read, b2read)
+            d2 = wptype(2) * wp.max(a2read, b2read)
+
+            a3read = vec3(a[i, 2], a[i, 3], a[i, 4])
+            b3read = vec3(b[i, 2], b[i, 3], b[i, 4])
+            c3 = wptype(2) * wp.min(a3read, b3read)
+            d3 = wptype(2) * wp.max(a3read, b3read)
+
+            a4read = vec4(a[i, 5], a[i, 6], a[i, 7], a[i, 8])
+            b4read = vec4(b[i, 5], b[i, 6], b[i, 7], b[i, 8])
+            c4 = wptype(2) * wp.min(a4read, b4read)
+            d4 = wptype(2) * wp.max(a4read, b4read)
+
+            a5read = vec5(a[i, 9], a[i, 10], a[i, 11], a[i, 12], a[i, 13])
+            b5read = vec5(b[i, 9], b[i, 10], b[i, 11], b[i, 12], b[i, 13])
+            c5 = wptype(2) * wp.min(a5read, b5read)
+            d5 = wptype(2) * wp.max(a5read, b5read)
+
+            mins[i, 0] = c2[0]
+            mins[i, 1] = c2[1]
+
+            mins[i, 2] = c3[0]
+            mins[i, 3] = c3[1]
+            mins[i, 4] = c3[2]
+
+            mins[i, 5] = c4[0]
+            mins[i, 6] = c4[1]
+            mins[i, 7] = c4[2]
+            mins[i, 8] = c4[3]
+
+            mins[i, 9] = c5[0]
+            mins[i, 10] = c5[1]
+            mins[i, 11] = c5[2]
+            mins[i, 12] = c5[3]
+            mins[i, 13] = c5[4]
+
+            maxs[i, 0] = d2[0]
+            maxs[i, 1] = d2[1]
+
+            maxs[i, 2] = d3[0]
+            maxs[i, 3] = d3[1]
+            maxs[i, 4] = d3[2]
+
+            maxs[i, 5] = d4[0]
+            maxs[i, 6] = d4[1]
+            maxs[i, 7] = d4[2]
+            maxs[i, 8] = d4[3]
+
+            maxs[i, 9] = d5[0]
+            maxs[i, 10] = d5[1]
+            maxs[i, 11] = d5[2]
+            maxs[i, 12] = d5[3]
+            maxs[i, 13] = d5[4]
+
+    kernel = getkernel(check_vec_min_max, suffix=dtype.__name__)
+    output_select_kernel = get_select_kernel2(wptype)
+
+    if register_kernels:
+        return
+
+    a = wp.array(randvals(rng, (10, 14), dtype), dtype=wptype, requires_grad=True, device=device)
+    b = wp.array(randvals(rng, (10, 14), dtype), dtype=wptype, requires_grad=True, device=device)
+
+    mins = wp.zeros((10, 14), dtype=wptype, requires_grad=True, device=device)
+    maxs = wp.zeros((10, 14), dtype=wptype, requires_grad=True, device=device)
+
+    tape = wp.Tape()
+    with tape:
+        wp.launch(kernel, dim=1, inputs=[a, b], outputs=[mins, maxs], device=device)
+
+    assert_np_equal(mins.numpy(), 2 * np.minimum(a.numpy(), b.numpy()), tol=tol)
+    assert_np_equal(maxs.numpy(), 2 * np.maximum(a.numpy(), b.numpy()), tol=tol)
+
+    out = wp.zeros(1, dtype=wptype, requires_grad=True, device=device)
+    if dtype in np_float_types:
+        for i in range(10):
+            for j in range(14):
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(kernel, dim=1, inputs=[a, b], outputs=[mins, maxs], device=device)
+                    wp.launch(output_select_kernel, dim=1, inputs=[mins, i, j], outputs=[out], device=device)
+
+                tape.backward(loss=out)
+                expected = np.zeros_like(a.numpy())
+                expected[i, j] = 2 if (a.numpy()[i, j] < b.numpy()[i, j]) else 0
+                assert_np_equal(tape.gradients[a].numpy(), expected, tol=tol)
+                expected[i, j] = 2 if (b.numpy()[i, j] < a.numpy()[i, j]) else 0
+                assert_np_equal(tape.gradients[b].numpy(), expected, tol=tol)
+                tape.zero()
+
+                tape = wp.Tape()
+                with tape:
+                    wp.launch(kernel, dim=1, inputs=[a, b], outputs=[mins, maxs], device=device)
+                    wp.launch(output_select_kernel, dim=1, inputs=[maxs, i, j], outputs=[out], device=device)
+
+                tape.backward(loss=out)
+                expected = np.zeros_like(a.numpy())
+                expected[i, j] = 2 if (a.numpy()[i, j] > b.numpy()[i, j]) else 0
+                assert_np_equal(tape.gradients[a].numpy(), expected, tol=tol)
+                expected[i, j] = 2 if (b.numpy()[i, j] > a.numpy()[i, j]) else 0
+                assert_np_equal(tape.gradients[b].numpy(), expected, tol=tol)
+                tape.zero()
+
+
+devices = get_test_devices()
+
+
+class TestVecScalarOps(unittest.TestCase):
+    pass
+
+
+for dtype in np_scalar_types:
+    add_function_test(TestVecScalarOps, f"test_arrays_{dtype.__name__}", test_arrays, devices=devices, dtype=dtype)
+    add_function_test(TestVecScalarOps, f"test_components_{dtype.__name__}", test_components, devices=None, dtype=dtype)
+    add_function_test(
+        TestVecScalarOps, f"test_py_arithmetic_ops_{dtype.__name__}", test_py_arithmetic_ops, devices=None, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_constructors_{dtype.__name__}", test_constructors, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps,
+        f"test_anon_type_instance_{dtype.__name__}",
+        test_anon_type_instance,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_indexing_{dtype.__name__}", test_indexing, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_equality_{dtype.__name__}", test_equality, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps,
+        f"test_scalar_multiplication_{dtype.__name__}",
+        test_scalar_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps,
+        f"test_scalar_multiplication_rightmul_{dtype.__name__}",
+        test_scalar_multiplication_rightmul,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps,
+        f"test_cw_multiplication_{dtype.__name__}",
+        test_cw_multiplication,
+        devices=devices,
+        dtype=dtype,
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_scalar_division_{dtype.__name__}", test_scalar_division, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_cw_division_{dtype.__name__}", test_cw_division, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_addition_{dtype.__name__}", test_addition, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_dotproduct_{dtype.__name__}", test_dotproduct, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_equivalent_types_{dtype.__name__}", test_equivalent_types, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_conversions_{dtype.__name__}", test_conversions, devices=devices, dtype=dtype
+    )
+    add_function_test_register_kernel(
+        TestVecScalarOps, f"test_constants_{dtype.__name__}", test_constants, devices=devices, dtype=dtype
+    )
+
+    # the kernels in this test compile incredibly slowly...
+    # add_function_test_register_kernel(TestVecScalarOps, f"test_minmax_{dtype.__name__}", test_minmax, devices=devices, dtype=dtype)
+
+
+if __name__ == "__main__":
+    wp.build.clear_kernel_cache()
+    unittest.main(verbosity=2, failfast=True)
diff --git a/warp/tests/test_volume.py b/warp/tests/test_volume.py
index 6c26e7946..40bbb3808 100644
--- a/warp/tests/test_volume.py
+++ b/warp/tests/test_volume.py
@@ -10,7 +10,7 @@
 import numpy as np
 
 import warp as wp
-from warp.tests.test_base import *
+from warp.tests.unittest_utils import *
 
 wp.init()
 
@@ -330,398 +330,407 @@ def test_volume_store_i(volume: wp.uint64, points: wp.array(dtype=wp.vec3), valu
     values[tid] = wp.volume_lookup_i(volume, i, j, k)
 
 
-def register(parent):
-    devices = get_test_devices()
-    rng = np.random.default_rng(101215)
-
-    # Note about the test grids:
-    # test_grid and test_int32_grid
-    #   active region: [-10,10]^3
-    #   values: v[i,j,k] = i * j * k
-    #   voxel size: 0.25
-    #
-    # test_vec_grid
-    #   active region: [-10,10]^3
-    #   values: v[i,j,k] = (i + 2*j + 3*k, 4*i + 5*j + 6*k, 7*i + 8*j + 9*k)
-    #   voxel size: 0.25
-    #
-    # torus
-    #   index to world transformation:
-    #      [0.1, 0, 0, 0]
-    #      [0, 0, 0.1, 0]
-    #      [0, 0.1, 0, 0]
-    #      [1, 2, 3, 1]
-    #   (-90 degrees rotation along X)
-    #   voxel size: 0.1
-    volume_paths = {
-        "float": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_grid.nvdb")),
-        "int32": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_int32_grid.nvdb")),
-        "vec3f": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_vec_grid.nvdb")),
-        "torus": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/torus.nvdb")),
-        "float_write": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_grid.nvdb")),
-    }
-
-    test_volume_tiles = (
-        np.array([[i, j, k] for i in range(-2, 2) for j in range(-2, 2) for k in range(-2, 2)], dtype=np.int32) * 8
-    )
+devices = get_test_devices()
+rng = np.random.default_rng(101215)
+
+# Note about the test grids:
+# test_grid and test_int32_grid
+#   active region: [-10,10]^3
+#   values: v[i,j,k] = i * j * k
+#   voxel size: 0.25
+#
+# test_vec_grid
+#   active region: [-10,10]^3
+#   values: v[i,j,k] = (i + 2*j + 3*k, 4*i + 5*j + 6*k, 7*i + 8*j + 9*k)
+#   voxel size: 0.25
+#
+# torus
+#   index to world transformation:
+#      [0.1, 0, 0, 0]
+#      [0, 0, 0.1, 0]
+#      [0, 0.1, 0, 0]
+#      [1, 2, 3, 1]
+#   (-90 degrees rotation along X)
+#   voxel size: 0.1
+volume_paths = {
+    "float": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_grid.nvdb")),
+    "int32": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_int32_grid.nvdb")),
+    "vec3f": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_vec_grid.nvdb")),
+    "torus": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/torus.nvdb")),
+    "float_write": os.path.abspath(os.path.join(os.path.dirname(__file__), "assets/test_grid.nvdb")),
+}
+
+test_volume_tiles = (
+    np.array([[i, j, k] for i in range(-2, 2) for j in range(-2, 2) for k in range(-2, 2)], dtype=np.int32) * 8
+)
+
+volumes = {}
+for value_type, path in volume_paths.items():
+    volumes[value_type] = {}
+    volume_data = open(path, "rb").read()
+    for device in devices:
+        try:
+            volume = wp.Volume.load_from_nvdb(volume_data, device)
+        except RuntimeError as e:
+            raise RuntimeError(f'Failed to load volume from "{path}" to {device} memory:\n{e}')
+
+        volumes[value_type][device.alias] = volume
+
+axis = np.linspace(-1, 1, 3)
+point_grid = np.array([[x, y, z] for x in axis for y in axis for z in axis], dtype=np.float32)
+
+
+def test_volume_sample_linear_f_gradient(test, device):
+    points = rng.uniform(-10.0, 10.0, size=(100, 3))
+    values = wp.array(np.zeros(1), dtype=wp.float32, device=device, requires_grad=True)
+    for test_case in points:
+        uvws = wp.array(test_case, dtype=wp.vec3, device=device, requires_grad=True)
+        xyzs = wp.array(test_case * 0.25, dtype=wp.vec3, device=device, requires_grad=True)
+
+        tape = wp.Tape()
+        with tape:
+            wp.launch(
+                test_volume_sample_local_f_linear_values,
+                dim=1,
+                inputs=[volumes["float"][device.alias].id, uvws, values],
+                device=device,
+            )
+        tape.backward(values)
+
+        x, y, z = test_case
+        grad_expected = np.array([y * z, x * z, x * y])
+        grad_computed = tape.gradients[uvws].numpy()[0]
+        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+
+        tape = wp.Tape()
+        with tape:
+            wp.launch(
+                test_volume_sample_world_f_linear_values,
+                dim=1,
+                inputs=[volumes["float"][device.alias].id, xyzs, values],
+                device=device,
+            )
+        tape.backward(values)
 
-    volumes = {}
-    points = {}
-    points_jittered = {}
-    for value_type, path in volume_paths.items():
-        volumes[value_type] = {}
-        volume_data = open(path, "rb").read()
-        for device in devices:
-            try:
-                volume = wp.Volume.load_from_nvdb(volume_data, device)
-            except RuntimeError as e:
-                raise RuntimeError(f'Failed to load volume from "{path}" to {device} memory:\n{e}')
-
-            volumes[value_type][device.alias] = volume
-
-    axis = np.linspace(-1, 1, 3)
-    point_grid = np.array([[x, y, z] for x in axis for y in axis for z in axis], dtype=np.float32)
-
-    class TestVolumes(parent):
-        def test_volume_sample_linear_f_gradient(self):
-            for device in devices:
-                points = rng.uniform(-10.0, 10.0, size=(100, 3))
-                values = wp.array(np.zeros(1), dtype=wp.float32, device=device, requires_grad=True)
-                for case in points:
-                    uvws = wp.array(case, dtype=wp.vec3, device=device, requires_grad=True)
-                    xyzs = wp.array(case * 0.25, dtype=wp.vec3, device=device, requires_grad=True)
-
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            test_volume_sample_local_f_linear_values,
-                            dim=1,
-                            inputs=[volumes["float"][device.alias].id, uvws, values],
-                            device=device,
-                        )
-                    tape.backward(values)
-
-                    x, y, z = case
-                    grad_expected = np.array([y * z, x * z, x * y])
-                    grad_computed = tape.gradients[uvws].numpy()[0]
-                    np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            test_volume_sample_world_f_linear_values,
-                            dim=1,
-                            inputs=[volumes["float"][device.alias].id, xyzs, values],
-                            device=device,
-                        )
-                    tape.backward(values)
-
-                    x, y, z = case
-                    grad_expected = np.array([y * z, x * z, x * y]) / 0.25
-                    grad_computed = tape.gradients[xyzs].numpy()[0]
-                    np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-
-        def test_volume_sample_grad_linear_f_gradient(self):
-            for device in devices:
-                points = rng.uniform(-10.0, 10.0, size=(100, 3))
-                values = wp.array(np.zeros(1), dtype=wp.float32, device=device, requires_grad=True)
-                for case in points:
-                    uvws = wp.array(case, dtype=wp.vec3, device=device, requires_grad=True)
-                    xyzs = wp.array(case * 0.25, dtype=wp.vec3, device=device, requires_grad=True)
-
-                    for case_num in range(4):
-                        tape = wp.Tape()
-                        with tape:
-                            wp.launch(
-                                test_volume_sample_grad_local_f_linear_values,
-                                dim=1,
-                                inputs=[volumes["float"][device.alias].id, uvws, values, case_num],
-                                device=device,
-                            )
-                        tape.backward(values)
-
-                        x, y, z = case
-                        grad_computed = tape.gradients[uvws].numpy()[0]
-                        if case_num == 0:
-                            grad_expected = np.array([y * z, x * z, x * y])
-                        elif case_num == 1:
-                            grad_expected = np.array([0.0, z, y])
-                        elif case_num == 2:
-                            grad_expected = np.array([z, 0.0, x])
-                        elif case_num == 3:
-                            grad_expected = np.array([y, x, 0.0])
-
-                        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-                        tape.zero()
-
-                    for case_num in range(4):
-                        tape = wp.Tape()
-                        with tape:
-                            wp.launch(
-                                test_volume_sample_grad_world_f_linear_values,
-                                dim=1,
-                                inputs=[volumes["float"][device.alias].id, xyzs, values, case_num],
-                                device=device,
-                            )
-                        tape.backward(values)
-
-                        x, y, z = case
-                        grad_computed = tape.gradients[xyzs].numpy()[0]
-                        if case_num == 0:
-                            grad_expected = np.array([y * z, x * z, x * y]) / 0.25
-                        elif case_num == 1:
-                            grad_expected = np.array([0.0, z, y]) / 0.25
-                        elif case_num == 2:
-                            grad_expected = np.array([z, 0.0, x]) / 0.25
-                        elif case_num == 3:
-                            grad_expected = np.array([y, x, 0.0]) / 0.25
-
-                        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-                        tape.zero()
-
-        def test_volume_sample_linear_v_gradient(self):
-            for device in devices:
-                points = rng.uniform(-10.0, 10.0, size=(100, 3))
-                values = wp.zeros(1, dtype=wp.float32, device=device, requires_grad=True)
-                for case in points:
-                    uvws = wp.array(case, dtype=wp.vec3, device=device, requires_grad=True)
-                    xyzs = wp.array(case * 0.25, dtype=wp.vec3, device=device, requires_grad=True)
-
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            test_volume_sample_local_v_linear_values,
-                            dim=1,
-                            inputs=[volumes["vec3f"][device.alias].id, uvws, values],
-                            device=device,
-                        )
-                    tape.backward(values)
-
-                    grad_expected = np.array([6.0, 15.0, 24.0])
-                    grad_computed = tape.gradients[uvws].numpy()[0]
-                    np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            test_volume_sample_world_v_linear_values,
-                            dim=1,
-                            inputs=[volumes["vec3f"][device.alias].id, xyzs, values],
-                            device=device,
-                        )
-                    tape.backward(values)
-
-                    grad_expected = np.array([6.0, 15.0, 24.0]) / 0.25
-                    grad_computed = tape.gradients[xyzs].numpy()[0]
-                    np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-
-        def test_volume_transform_gradient(self):
-            for device in devices:
-                values = wp.zeros(1, dtype=wp.float32, device=device, requires_grad=True)
-                grad_values = wp.zeros(1, dtype=wp.vec3, device=device)
-                points = rng.uniform(-10.0, 10.0, size=(10, 3))
-                for case in points:
-                    points = wp.array(case, dtype=wp.vec3, device=device, requires_grad=True)
-                    tape = wp.Tape()
-                    with tape:
-                        wp.launch(
-                            test_volume_index_to_world,
-                            dim=1,
-                            inputs=[volumes["torus"][device.alias].id, points, values, grad_values],
-                            device=device,
-                        )
-                    tape.backward(values)
-
-                    grad_computed = tape.gradients[points].numpy()
-                    grad_expected = grad_values.numpy()
-                    np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-
-                    grad_computed = tape.gradients[points].numpy()
-                    grad_expected = grad_values.numpy()
-                    np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
-
-        def test_volume_store(self):
-            values_ref = np.array([x + 100 * y + 10000 * z for x, y, z in point_grid])
-            for device in devices:
-                points = wp.array(point_grid, dtype=wp.vec3, device=device)
-                values = wp.empty(len(point_grid), dtype=wp.float32, device=device)
-                wp.launch(
-                    test_volume_store_f,
-                    dim=len(point_grid),
-                    inputs=[volumes["float_write"][device.alias].id, points, values],
-                    device=device,
-                )
+        x, y, z = test_case
+        grad_expected = np.array([y * z, x * z, x * y]) / 0.25
+        grad_computed = tape.gradients[xyzs].numpy()[0]
+        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
 
-                values_res = values.numpy()
-                np.testing.assert_equal(values_res, values_ref)
 
-        def test_volume_allocation_f(self):
-            bg_value = -123.0
-            points_np = np.append(point_grid, [[8096, 8096, 8096]], axis=0)
-            values_ref = np.append(np.array([x + 100 * y + 10000 * z for x, y, z in point_grid]), bg_value)
-            for device in devices:
-                if device.is_cpu:
-                    continue
+def test_volume_sample_grad_linear_f_gradient(test, device):
+    points = rng.uniform(-10.0, 10.0, size=(100, 3))
+    values = wp.array(np.zeros(1), dtype=wp.float32, device=device, requires_grad=True)
+    for test_case in points:
+        uvws = wp.array(test_case, dtype=wp.vec3, device=device, requires_grad=True)
+        xyzs = wp.array(test_case * 0.25, dtype=wp.vec3, device=device, requires_grad=True)
 
-                volume = wp.Volume.allocate(
-                    min=[-11, -11, -11], max=[11, 11, 11], voxel_size=0.1, bg_value=bg_value, device=device
+        for case_num in range(4):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(
+                    test_volume_sample_grad_local_f_linear_values,
+                    dim=1,
+                    inputs=[volumes["float"][device.alias].id, uvws, values, case_num],
+                    device=device,
                 )
-                points = wp.array(points_np, dtype=wp.vec3, device=device)
-                values = wp.empty(len(points_np), dtype=wp.float32, device=device)
-                wp.launch(test_volume_store_f, dim=len(points_np), inputs=[volume.id, points, values], device=device)
-
-                values_res = values.numpy()
-                np.testing.assert_equal(values_res, values_ref)
-
-        def test_volume_allocation_v(self):
-            bg_value = (-1, 2.0, -3)
-            points_np = np.append(point_grid, [[8096, 8096, 8096]], axis=0)
-            values_ref = np.append(point_grid, [bg_value], axis=0)
-            for device in devices:
-                if device.is_cpu:
-                    continue
-
-                volume = wp.Volume.allocate(
-                    min=[-11, -11, -11], max=[11, 11, 11], voxel_size=0.1, bg_value=bg_value, device=device
+            tape.backward(values)
+
+            x, y, z = test_case
+            grad_computed = tape.gradients[uvws].numpy()[0]
+            if case_num == 0:
+                grad_expected = np.array([y * z, x * z, x * y])
+            elif case_num == 1:
+                grad_expected = np.array([0.0, z, y])
+            elif case_num == 2:
+                grad_expected = np.array([z, 0.0, x])
+            elif case_num == 3:
+                grad_expected = np.array([y, x, 0.0])
+
+            np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+            tape.zero()
+
+        for case_num in range(4):
+            tape = wp.Tape()
+            with tape:
+                wp.launch(
+                    test_volume_sample_grad_world_f_linear_values,
+                    dim=1,
+                    inputs=[volumes["float"][device.alias].id, xyzs, values, case_num],
+                    device=device,
                 )
-                points = wp.array(points_np, dtype=wp.vec3, device=device)
-                values = wp.empty(len(points_np), dtype=wp.vec3, device=device)
-                wp.launch(test_volume_store_v, dim=len(points_np), inputs=[volume.id, points, values], device=device)
-
-                values_res = values.numpy()
-                np.testing.assert_equal(values_res, values_ref)
-
-        def test_volume_allocation_i(self):
-            bg_value = -123
-            points_np = np.append(point_grid, [[8096, 8096, 8096]], axis=0)
-            values_ref = np.append(
-                np.array([x + 100 * y + 10000 * z for x, y, z in point_grid], dtype=np.int32), bg_value
+            tape.backward(values)
+
+            x, y, z = test_case
+            grad_computed = tape.gradients[xyzs].numpy()[0]
+            if case_num == 0:
+                grad_expected = np.array([y * z, x * z, x * y]) / 0.25
+            elif case_num == 1:
+                grad_expected = np.array([0.0, z, y]) / 0.25
+            elif case_num == 2:
+                grad_expected = np.array([z, 0.0, x]) / 0.25
+            elif case_num == 3:
+                grad_expected = np.array([y, x, 0.0]) / 0.25
+
+            np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+            tape.zero()
+
+
+def test_volume_sample_linear_v_gradient(test, device):
+    points = rng.uniform(-10.0, 10.0, size=(100, 3))
+    values = wp.zeros(1, dtype=wp.float32, device=device, requires_grad=True)
+    for test_case in points:
+        uvws = wp.array(test_case, dtype=wp.vec3, device=device, requires_grad=True)
+        xyzs = wp.array(test_case * 0.25, dtype=wp.vec3, device=device, requires_grad=True)
+
+        tape = wp.Tape()
+        with tape:
+            wp.launch(
+                test_volume_sample_local_v_linear_values,
+                dim=1,
+                inputs=[volumes["vec3f"][device.alias].id, uvws, values],
+                device=device,
+            )
+        tape.backward(values)
+
+        grad_expected = np.array([6.0, 15.0, 24.0])
+        grad_computed = tape.gradients[uvws].numpy()[0]
+        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+
+        tape = wp.Tape()
+        with tape:
+            wp.launch(
+                test_volume_sample_world_v_linear_values,
+                dim=1,
+                inputs=[volumes["vec3f"][device.alias].id, xyzs, values],
+                device=device,
+            )
+        tape.backward(values)
+
+        grad_expected = np.array([6.0, 15.0, 24.0]) / 0.25
+        grad_computed = tape.gradients[xyzs].numpy()[0]
+        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+
+
+def test_volume_transform_gradient(test, device):
+    values = wp.zeros(1, dtype=wp.float32, device=device, requires_grad=True)
+    grad_values = wp.zeros(1, dtype=wp.vec3, device=device)
+    test_points = rng.uniform(-10.0, 10.0, size=(10, 3))
+    for test_case in test_points:
+        points = wp.array(test_case, dtype=wp.vec3, device=device, requires_grad=True)
+        tape = wp.Tape()
+        with tape:
+            wp.launch(
+                test_volume_index_to_world,
+                dim=1,
+                inputs=[volumes["torus"][device.alias].id, points, values, grad_values],
+                device=device,
             )
-            for device in devices:
-                if device.is_cpu:
-                    continue
+        tape.backward(values)
+
+        grad_computed = tape.gradients[points].numpy()
+        grad_expected = grad_values.numpy()
+        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+
+        grad_computed = tape.gradients[points].numpy()
+        grad_expected = grad_values.numpy()
+        np.testing.assert_allclose(grad_computed, grad_expected, rtol=1e-4)
+
+
+def test_volume_store(test, device):
+    values_ref = np.array([x + 100 * y + 10000 * z for x, y, z in point_grid])
+    points = wp.array(point_grid, dtype=wp.vec3, device=device)
+    values = wp.empty(len(point_grid), dtype=wp.float32, device=device)
+    wp.launch(
+        test_volume_store_f,
+        dim=len(point_grid),
+        inputs=[volumes["float_write"][device.alias].id, points, values],
+        device=device,
+    )
 
-                volume = wp.Volume.allocate(
-                    min=[-11, -11, -11], max=[11, 11, 11], voxel_size=0.1, bg_value=bg_value, device=device
-                )
-                points = wp.array(points_np, dtype=wp.vec3, device=device)
-                values = wp.empty(len(points_np), dtype=wp.int32, device=device)
-                wp.launch(test_volume_store_i, dim=len(points_np), inputs=[volume.id, points, values], device=device)
-
-                values_res = values.numpy()
-                np.testing.assert_equal(values_res, values_ref)
-
-        def test_volume_introspection(self):
-            for volume_names in ("float", "vec3f"):
-                for device in devices:
-                    volume = volumes[volume_names][device.alias]
-                    tiles_actual = volume.get_tiles().numpy()
-                    tiles_sorted = tiles_actual[np.lexsort(tiles_actual.T[::-1])]
-                    voxel_size = np.array(volume.get_voxel_size())
-
-                    np.testing.assert_equal(test_volume_tiles, tiles_sorted)
-                    np.testing.assert_equal([0.25] * 3, voxel_size)
-
-        def test_volume_from_numpy(self):
-            # Volume.allocate_from_tiles() is only available with CUDA
-            if wp.is_cuda_available():
-                mins = np.array([-3.0, -3.0, -3.0])
-                voxel_size = 0.2
-                maxs = np.array([3.0, 3.0, 3.0])
-                nums = np.ceil((maxs - mins) / (voxel_size)).astype(dtype=int)
-                center = np.array([0.0, 0.0, 0.0])
-                rad = 2.5
-                sphere_sdf_np = np.zeros(tuple(nums))
-                for x in range(nums[0]):
-                    for y in range(nums[1]):
-                        for z in range(nums[2]):
-                            pos = mins + voxel_size * np.array([x, y, z])
-                            dis = np.linalg.norm(pos - center)
-                            sphere_sdf_np[x, y, z] = dis - rad
-                sphere_vdb = wp.Volume.load_from_numpy(sphere_sdf_np, mins, voxel_size, rad + 3.0 * voxel_size)
-
-                self.assertNotEqual(sphere_vdb.id, 0)
-
-                sphere_vdb_array = sphere_vdb.array()
-                self.assertEqual(sphere_vdb_array.dtype, wp.uint8)
-                self.assertFalse(sphere_vdb_array.owner)
+    values_res = values.numpy()
+    np.testing.assert_equal(values_res, values_ref)
+
+
+def test_volume_allocation_f(test, device):
+    bg_value = -123.0
+    points_np = np.append(point_grid, [[8096, 8096, 8096]], axis=0)
+    values_ref = np.append(np.array([x + 100 * y + 10000 * z for x, y, z in point_grid]), bg_value)
+
+    volume = wp.Volume.allocate(min=[-11, -11, -11], max=[11, 11, 11], voxel_size=0.1, bg_value=bg_value, device=device)
+    points = wp.array(points_np, dtype=wp.vec3, device=device)
+    values = wp.empty(len(points_np), dtype=wp.float32, device=device)
+    wp.launch(test_volume_store_f, dim=len(points_np), inputs=[volume.id, points, values], device=device)
+
+    values_res = values.numpy()
+    np.testing.assert_equal(values_res, values_ref)
+
+
+def test_volume_allocation_v(test, device):
+    bg_value = (-1, 2.0, -3)
+    points_np = np.append(point_grid, [[8096, 8096, 8096]], axis=0)
+    values_ref = np.append(point_grid, [bg_value], axis=0)
+
+    volume = wp.Volume.allocate(min=[-11, -11, -11], max=[11, 11, 11], voxel_size=0.1, bg_value=bg_value, device=device)
+    points = wp.array(points_np, dtype=wp.vec3, device=device)
+    values = wp.empty(len(points_np), dtype=wp.vec3, device=device)
+    wp.launch(test_volume_store_v, dim=len(points_np), inputs=[volume.id, points, values], device=device)
+
+    values_res = values.numpy()
+    np.testing.assert_equal(values_res, values_ref)
+
+
+def test_volume_allocation_i(test, device):
+    bg_value = -123
+    points_np = np.append(point_grid, [[8096, 8096, 8096]], axis=0)
+    values_ref = np.append(np.array([x + 100 * y + 10000 * z for x, y, z in point_grid], dtype=np.int32), bg_value)
+
+    volume = wp.Volume.allocate(min=[-11, -11, -11], max=[11, 11, 11], voxel_size=0.1, bg_value=bg_value, device=device)
+    points = wp.array(points_np, dtype=wp.vec3, device=device)
+    values = wp.empty(len(points_np), dtype=wp.int32, device=device)
+    wp.launch(test_volume_store_i, dim=len(points_np), inputs=[volume.id, points, values], device=device)
+
+    values_res = values.numpy()
+    np.testing.assert_equal(values_res, values_ref)
+
+
+def test_volume_introspection(test, device):
+    for volume_names in ("float", "vec3f"):
+        with test.subTest(volume_names=volume_names):
+            volume = volumes[volume_names][device.alias]
+            tiles_actual = volume.get_tiles().numpy()
+            tiles_sorted = tiles_actual[np.lexsort(tiles_actual.T[::-1])]
+            voxel_size = np.array(volume.get_voxel_size())
+
+            np.testing.assert_equal(test_volume_tiles, tiles_sorted)
+            np.testing.assert_equal([0.25] * 3, voxel_size)
+
+
+def test_volume_from_numpy(test, device):
+    # Volume.allocate_from_tiles() is only available with CUDA
+    mins = np.array([-3.0, -3.0, -3.0])
+    voxel_size = 0.2
+    maxs = np.array([3.0, 3.0, 3.0])
+    nums = np.ceil((maxs - mins) / (voxel_size)).astype(dtype=int)
+    center = np.array([0.0, 0.0, 0.0])
+    rad = 2.5
+    sphere_sdf_np = np.zeros(tuple(nums))
+    for x in range(nums[0]):
+        for y in range(nums[1]):
+            for z in range(nums[2]):
+                pos = mins + voxel_size * np.array([x, y, z])
+                dis = np.linalg.norm(pos - center)
+                sphere_sdf_np[x, y, z] = dis - rad
+    sphere_vdb = wp.Volume.load_from_numpy(sphere_sdf_np, mins, voxel_size, rad + 3.0 * voxel_size, device=device)
+
+    test.assertNotEqual(sphere_vdb.id, 0)
+
+    sphere_vdb_array = sphere_vdb.array()
+    test.assertEqual(sphere_vdb_array.dtype, wp.uint8)
+    test.assertFalse(sphere_vdb_array.owner)
+
+
+class TestVolume(unittest.TestCase):
+    pass
+
+
+add_function_test(
+    TestVolume, "test_volume_sample_linear_f_gradient", test_volume_sample_linear_f_gradient, devices=devices
+)
+add_function_test(
+    TestVolume, "test_volume_sample_grad_linear_f_gradient", test_volume_sample_grad_linear_f_gradient, devices=devices
+)
+add_function_test(
+    TestVolume, "test_volume_sample_linear_v_gradient", test_volume_sample_linear_v_gradient, devices=devices
+)
+add_function_test(TestVolume, "test_volume_transform_gradient", test_volume_transform_gradient, devices=devices)
+add_function_test(TestVolume, "test_volume_store", test_volume_store, devices=devices)
+add_function_test(
+    TestVolume, "test_volume_allocation_f", test_volume_allocation_f, devices=get_unique_cuda_test_devices()
+)
+add_function_test(
+    TestVolume, "test_volume_allocation_v", test_volume_allocation_v, devices=get_unique_cuda_test_devices()
+)
+add_function_test(
+    TestVolume, "test_volume_allocation_i", test_volume_allocation_i, devices=get_unique_cuda_test_devices()
+)
+add_function_test(TestVolume, "test_volume_introspection", test_volume_introspection, devices=devices)
+add_function_test(TestVolume, "test_volume_from_numpy", test_volume_from_numpy, devices=get_unique_cuda_test_devices())
+
+points = {}
+points_jittered = {}
+for device in devices:
+    points_jittered_np = point_grid + rng.uniform(-0.5, 0.5, size=point_grid.shape)
+    points[device.alias] = wp.array(point_grid, dtype=wp.vec3, device=device)
+    points_jittered[device.alias] = wp.array(points_jittered_np, dtype=wp.vec3, device=device)
+
+    add_kernel_test(
+        TestVolume,
+        test_volume_lookup_f,
+        dim=len(point_grid),
+        inputs=[volumes["float"][device.alias].id, points[device.alias]],
+        devices=[device],
+    )
+    add_kernel_test(
+        TestVolume,
+        test_volume_sample_closest_f,
+        dim=len(point_grid),
+        inputs=[volumes["float"][device.alias].id, points_jittered[device.alias]],
+        devices=[device.alias],
+    )
+    add_kernel_test(
+        TestVolume,
+        test_volume_sample_linear_f,
+        dim=len(point_grid),
+        inputs=[volumes["float"][device.alias].id, points_jittered[device.alias]],
+        devices=[device.alias],
+    )
+    add_kernel_test(
+        TestVolume,
+        test_volume_sample_grad_linear_f,
+        dim=len(point_grid),
+        inputs=[volumes["float"][device.alias].id, points_jittered[device.alias]],
+        devices=[device.alias],
+    )
 
-    for device in devices:
-        points_jittered_np = point_grid + rng.uniform(-0.5, 0.5, size=point_grid.shape)
-        points[device.alias] = wp.array(point_grid, dtype=wp.vec3, device=device)
-        points_jittered[device.alias] = wp.array(points_jittered_np, dtype=wp.vec3, device=device)
-
-        add_kernel_test(
-            TestVolumes,
-            test_volume_lookup_f,
-            dim=len(point_grid),
-            inputs=[volumes["float"][device.alias].id, points[device.alias]],
-            devices=[device],
-        )
-        add_kernel_test(
-            TestVolumes,
-            test_volume_sample_closest_f,
-            dim=len(point_grid),
-            inputs=[volumes["float"][device.alias].id, points_jittered[device.alias]],
-            devices=[device.alias],
-        )
-        add_kernel_test(
-            TestVolumes,
-            test_volume_sample_linear_f,
-            dim=len(point_grid),
-            inputs=[volumes["float"][device.alias].id, points_jittered[device.alias]],
-            devices=[device.alias],
-        )
-        add_kernel_test(
-            TestVolumes,
-            test_volume_sample_grad_linear_f,
-            dim=len(point_grid),
-            inputs=[volumes["float"][device.alias].id, points_jittered[device.alias]],
-            devices=[device.alias],
-        )
-
-        add_kernel_test(
-            TestVolumes,
-            test_volume_lookup_v,
-            dim=len(point_grid),
-            inputs=[volumes["vec3f"][device.alias].id, points[device.alias]],
-            devices=[device.alias],
-        )
-        add_kernel_test(
-            TestVolumes,
-            test_volume_sample_closest_v,
-            dim=len(point_grid),
-            inputs=[volumes["vec3f"][device.alias].id, points_jittered[device.alias]],
-            devices=[device.alias],
-        )
-        add_kernel_test(
-            TestVolumes,
-            test_volume_sample_linear_v,
-            dim=len(point_grid),
-            inputs=[volumes["vec3f"][device.alias].id, points_jittered[device.alias]],
-            devices=[device.alias],
-        )
-
-        add_kernel_test(
-            TestVolumes,
-            test_volume_lookup_i,
-            dim=len(point_grid),
-            inputs=[volumes["int32"][device.alias].id, points[device.alias]],
-            devices=[device.alias],
-        )
-        add_kernel_test(
-            TestVolumes,
-            test_volume_sample_i,
-            dim=len(point_grid),
-            inputs=[volumes["int32"][device.alias].id, points_jittered[device.alias]],
-            devices=[device.alias],
-        )
-
-    return TestVolumes
+    add_kernel_test(
+        TestVolume,
+        test_volume_lookup_v,
+        dim=len(point_grid),
+        inputs=[volumes["vec3f"][device.alias].id, points[device.alias]],
+        devices=[device.alias],
+    )
+    add_kernel_test(
+        TestVolume,
+        test_volume_sample_closest_v,
+        dim=len(point_grid),
+        inputs=[volumes["vec3f"][device.alias].id, points_jittered[device.alias]],
+        devices=[device.alias],
+    )
+    add_kernel_test(
+        TestVolume,
+        test_volume_sample_linear_v,
+        dim=len(point_grid),
+        inputs=[volumes["vec3f"][device.alias].id, points_jittered[device.alias]],
+        devices=[device.alias],
+    )
+
+    add_kernel_test(
+        TestVolume,
+        test_volume_lookup_i,
+        dim=len(point_grid),
+        inputs=[volumes["int32"][device.alias].id, points[device.alias]],
+        devices=[device.alias],
+    )
+    add_kernel_test(
+        TestVolume,
+        test_volume_sample_i,
+        dim=len(point_grid),
+        inputs=[volumes["int32"][device.alias].id, points_jittered[device.alias]],
+        devices=[device.alias],
+    )
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    wp.force_load()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/test_volume_write.py b/warp/tests/test_volume_write.py
index 48532f22e..3c460e214 100644
--- a/warp/tests/test_volume_write.py
+++ b/warp/tests/test_volume_write.py
@@ -7,11 +7,11 @@
 
 import unittest
 
-import warp as wp
-from warp.tests.test_base import *
-
 import numpy as np
 
+import warp as wp
+from warp.tests.unittest_utils import *
+
 wp.init()
 # wp.config.cache_kernels = False
 
@@ -139,140 +139,127 @@ def test_volume_tile_readback_v(volume: wp.uint64, tiles: wp.array2d(dtype=wp.in
         values[tid * 512 + r] = wp.volume_lookup_v(volume, ii, jj, kk)
 
 
-rng = np.random.default_rng(101215)
-
-
-def register(parent):
-    devices = get_test_devices()
-
-    class TestVolumes(parent):
-        def test_volume_allocation(self):
-            voxel_size = 0.125
-            background_value = 123.456
-            translation = wp.vec3(-12.3, 4.56, -789)
-
-            axis = np.linspace(-11, 11, 23)
-            points_ref = np.array([[x, y, z] for x in axis for y in axis for z in axis])
-            values_ref = np.array([x + 100 * y + 10000 * z for x in axis for y in axis for z in axis])
-            num_points = len(points_ref)
-            bb_max = np.array([11, 11, 11])
-            for device in devices:
-                if device.is_cpu:
-                    continue
-
-                volume_a = wp.Volume.allocate(
-                    -bb_max,
-                    bb_max,
-                    voxel_size=voxel_size,
-                    bg_value=background_value,
-                    translation=translation,
-                    device=device,
-                )
-                volume_b = wp.Volume.allocate(
-                    -bb_max * voxel_size + translation,
-                    bb_max * voxel_size + translation,
-                    voxel_size=voxel_size,
-                    bg_value=background_value,
-                    translation=translation,
-                    points_in_world_space=True,
-                    device=device,
-                )
-                points = wp.array(points_ref, dtype=wp.vec3, device=device)
-                values_a = wp.empty(num_points, dtype=wp.float32, device=device)
-                values_b = wp.empty(num_points, dtype=wp.float32, device=device)
-                wp.launch(test_volume_store_f, dim=num_points, inputs=[volume_a.id, points], device=device)
-                wp.launch(test_volume_store_f, dim=num_points, inputs=[volume_b.id, points], device=device)
-                wp.launch(test_volume_readback_f, dim=num_points, inputs=[volume_a.id, points, values_a], device=device)
-                wp.launch(test_volume_readback_f, dim=num_points, inputs=[volume_b.id, points, values_b], device=device)
-
-                np.testing.assert_equal(values_a.numpy(), values_ref)
-                np.testing.assert_equal(values_b.numpy(), values_ref)
-
-        def test_volume_allocate_by_tiles_f(self):
-            voxel_size = 0.125
-            background_value = 123.456
-            translation = wp.vec3(-12.3, 4.56, -789)
-
-            num_tiles = 1000
-            tiles = rng.integers(-512, 512, size=(num_tiles, 3), dtype=np.int32)
-            points_is = tiles * 8  # points in index space
-            points_ws = points_is * voxel_size + translation  # points in world space
-
-            values_ref = np.empty(num_tiles * 512)
-            for t in range(num_tiles):
-                ti, tj, tk = points_is[t]
-                for i in range(8):
-                    for j in range(8):
-                        for k in range(8):
-                            values_ref[t * 512 + i * 64 + j * 8 + k] = float(100 * (ti + i) + 10 * (tj + j) + (tk + k))
-
-            for device in devices:
-                if device.is_cpu:
-                    continue
-
-                points_is_d = wp.array(points_is, dtype=wp.int32, device=device)
-                points_ws_d = wp.array(points_ws, dtype=wp.vec3, device=device)
-                volume_a = wp.Volume.allocate_by_tiles(
-                    points_is_d, voxel_size, background_value, translation, device=device
-                )
-                volume_b = wp.Volume.allocate_by_tiles(
-                    points_ws_d, voxel_size, background_value, translation, device=device
-                )
-                values_a = wp.empty(num_tiles * 512, dtype=wp.float32, device=device)
-                values_b = wp.empty(num_tiles * 512, dtype=wp.float32, device=device)
-
-                wp.launch(test_volume_tile_store_f, dim=num_tiles, inputs=[volume_a.id, points_is_d], device=device)
-                wp.launch(test_volume_tile_store_ws_f, dim=num_tiles, inputs=[volume_b.id, points_ws_d], device=device)
-                wp.launch(
-                    test_volume_tile_readback_f,
-                    dim=num_tiles,
-                    inputs=[volume_a.id, points_is_d, values_a],
-                    device=device,
-                )
-                wp.launch(
-                    test_volume_tile_readback_f,
-                    dim=num_tiles,
-                    inputs=[volume_b.id, points_is_d, values_b],
-                    device=device,
-                )
-
-                np.testing.assert_equal(values_a.numpy(), values_ref)
-                np.testing.assert_equal(values_b.numpy(), values_ref)
-
-        def test_volume_allocate_by_tiles_v(self):
-            num_tiles = 1000
-            tiles = rng.integers(-512, 512, size=(num_tiles, 3), dtype=np.int32)
-            points_is = tiles * 8
-
-            values_ref = np.empty((len(tiles) * 512, 3))
-            for t in range(len(tiles)):
-                ti, tj, tk = points_is[t]
-                for i in range(8):
-                    for j in range(8):
-                        for k in range(8):
-                            values_ref[t * 512 + i * 64 + j * 8 + k] = [ti + i, tj + j, tk + k]
-
-            for device in devices:
-                if device.is_cpu:
-                    continue
-
-                points_d = wp.array(points_is, dtype=wp.int32, device=device)
-                volume = wp.Volume.allocate_by_tiles(points_d, 0.1, wp.vec3(1, 2, 3), device=device)
-                values = wp.empty(len(points_d) * 512, dtype=wp.vec3, device=device)
-
-                wp.launch(test_volume_tile_store_v, dim=len(points_d), inputs=[volume.id, points_d], device=device)
-                wp.launch(
-                    test_volume_tile_readback_v, dim=len(points_d), inputs=[volume.id, points_d, values], device=device
-                )
-
-                values_res = values.numpy()
-                np.testing.assert_equal(values_res, values_ref)
-
-    return TestVolumes
+def test_volume_allocation(test, device):
+    voxel_size = 0.125
+    background_value = 123.456
+    translation = wp.vec3(-12.3, 4.56, -789)
+
+    axis = np.linspace(-11, 11, 23)
+    points_ref = np.array([[x, y, z] for x in axis for y in axis for z in axis])
+    values_ref = np.array([x + 100 * y + 10000 * z for x in axis for y in axis for z in axis])
+    num_points = len(points_ref)
+    bb_max = np.array([11, 11, 11])
+    volume_a = wp.Volume.allocate(
+        -bb_max,
+        bb_max,
+        voxel_size=voxel_size,
+        bg_value=background_value,
+        translation=translation,
+        device=device,
+    )
+    volume_b = wp.Volume.allocate(
+        -bb_max * voxel_size + translation,
+        bb_max * voxel_size + translation,
+        voxel_size=voxel_size,
+        bg_value=background_value,
+        translation=translation,
+        points_in_world_space=True,
+        device=device,
+    )
+    points = wp.array(points_ref, dtype=wp.vec3, device=device)
+    values_a = wp.empty(num_points, dtype=wp.float32, device=device)
+    values_b = wp.empty(num_points, dtype=wp.float32, device=device)
+    wp.launch(test_volume_store_f, dim=num_points, inputs=[volume_a.id, points], device=device)
+    wp.launch(test_volume_store_f, dim=num_points, inputs=[volume_b.id, points], device=device)
+    wp.launch(test_volume_readback_f, dim=num_points, inputs=[volume_a.id, points, values_a], device=device)
+    wp.launch(test_volume_readback_f, dim=num_points, inputs=[volume_b.id, points, values_b], device=device)
+
+    np.testing.assert_equal(values_a.numpy(), values_ref)
+    np.testing.assert_equal(values_b.numpy(), values_ref)
+
+
+def test_volume_allocate_by_tiles_f(test, device):
+    voxel_size = 0.125
+    background_value = 123.456
+    translation = wp.vec3(-12.3, 4.56, -789)
+
+    num_tiles = 1000
+    rng = np.random.default_rng(101215)
+    tiles = rng.integers(-512, 512, size=(num_tiles, 3), dtype=np.int32)
+    points_is = tiles * 8  # points in index space
+    points_ws = points_is * voxel_size + translation  # points in world space
+
+    values_ref = np.empty(num_tiles * 512)
+    for t in range(num_tiles):
+        ti, tj, tk = points_is[t]
+        for i in range(8):
+            for j in range(8):
+                for k in range(8):
+                    values_ref[t * 512 + i * 64 + j * 8 + k] = float(100 * (ti + i) + 10 * (tj + j) + (tk + k))
+
+    points_is_d = wp.array(points_is, dtype=wp.int32, device=device)
+    points_ws_d = wp.array(points_ws, dtype=wp.vec3, device=device)
+    volume_a = wp.Volume.allocate_by_tiles(points_is_d, voxel_size, background_value, translation, device=device)
+    volume_b = wp.Volume.allocate_by_tiles(points_ws_d, voxel_size, background_value, translation, device=device)
+    values_a = wp.empty(num_tiles * 512, dtype=wp.float32, device=device)
+    values_b = wp.empty(num_tiles * 512, dtype=wp.float32, device=device)
+
+    wp.launch(test_volume_tile_store_f, dim=num_tiles, inputs=[volume_a.id, points_is_d], device=device)
+    wp.launch(test_volume_tile_store_ws_f, dim=num_tiles, inputs=[volume_b.id, points_ws_d], device=device)
+    wp.launch(
+        test_volume_tile_readback_f,
+        dim=num_tiles,
+        inputs=[volume_a.id, points_is_d, values_a],
+        device=device,
+    )
+    wp.launch(
+        test_volume_tile_readback_f,
+        dim=num_tiles,
+        inputs=[volume_b.id, points_is_d, values_b],
+        device=device,
+    )
+
+    np.testing.assert_equal(values_a.numpy(), values_ref)
+    np.testing.assert_equal(values_b.numpy(), values_ref)
+
+
+def test_volume_allocate_by_tiles_v(test, device):
+    num_tiles = 1000
+    rng = np.random.default_rng(101215)
+    tiles = rng.integers(-512, 512, size=(num_tiles, 3), dtype=np.int32)
+    points_is = tiles * 8
+
+    values_ref = np.empty((len(tiles) * 512, 3))
+    for t in range(len(tiles)):
+        ti, tj, tk = points_is[t]
+        for i in range(8):
+            for j in range(8):
+                for k in range(8):
+                    values_ref[t * 512 + i * 64 + j * 8 + k] = [ti + i, tj + j, tk + k]
+
+    points_d = wp.array(points_is, dtype=wp.int32, device=device)
+    volume = wp.Volume.allocate_by_tiles(points_d, 0.1, wp.vec3(1, 2, 3), device=device)
+    values = wp.empty(len(points_d) * 512, dtype=wp.vec3, device=device)
+
+    wp.launch(test_volume_tile_store_v, dim=len(points_d), inputs=[volume.id, points_d], device=device)
+    wp.launch(test_volume_tile_readback_v, dim=len(points_d), inputs=[volume.id, points_d, values], device=device)
+
+    values_res = values.numpy()
+    np.testing.assert_equal(values_res, values_ref)
+
+
+devices = get_unique_cuda_test_devices()
+
+
+class TestVolumeWrite(unittest.TestCase):
+    pass
+
+
+add_function_test(TestVolumeWrite, "test_volume_allocation", test_volume_allocation, devices=devices)
+add_function_test(TestVolumeWrite, "test_volume_allocate_by_tiles_f", test_volume_allocate_by_tiles_f, devices=devices)
+add_function_test(TestVolumeWrite, "test_volume_allocate_by_tiles_v", test_volume_allocate_by_tiles_v, devices=devices)
 
 
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    wp.force_load()
-    _ = register(unittest.TestCase)
     unittest.main(verbosity=2)
diff --git a/warp/tests/unittest_serial.py b/warp/tests/unittest_serial.py
new file mode 100644
index 000000000..80146b059
--- /dev/null
+++ b/warp/tests/unittest_serial.py
@@ -0,0 +1,35 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import warp as wp
+from warp.tests.unittest_utils import TeamCityTestRunner
+
+
+def run_suite() -> bool:
+    """Run a test suite"""
+
+    import warp.tests.unittest_suites
+
+    # force rebuild of all kernels
+    wp.build.clear_kernel_cache()
+    print("Cleared Warp kernel cache")
+
+    runner = TeamCityTestRunner(verbosity=2, failfast=False)
+
+    # Can swap out warp.tests.unittest_suites.explicit_suite()
+    suite = warp.tests.unittest_suites.auto_discover_suite()
+    print(f"Test suite has {suite.countTestCases()} tests")
+
+    ret = not runner.run(suite, "WarpTests").wasSuccessful()
+    return ret
+
+
+if __name__ == "__main__":
+    ret = run_suite()
+    import sys
+
+    sys.exit(ret)
diff --git a/warp/tests/unittest_suites.py b/warp/tests/unittest_suites.py
new file mode 100644
index 000000000..e64e107c0
--- /dev/null
+++ b/warp/tests/unittest_suites.py
@@ -0,0 +1,291 @@
+# Copyright (c) 2023 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+"""Warp Test Suites
+
+This file is intended to define functions that return TestSuite objects, which
+can be used in parallel or serial unit tests (with optional code coverage)
+"""
+
+import os
+import unittest
+
+START_DIRECTORY = os.path.realpath(os.path.dirname(__file__))
+TOP_LEVEL_DIRECTORY = os.path.realpath(os.path.join(START_DIRECTORY, "..", ".."))
+
+
+def _create_suite_from_test_classes(test_classes):
+    suite = unittest.TestSuite()
+
+    for test in test_classes:
+        sub_suite = unittest.TestSuite()
+        sub_suite.addTest(unittest.defaultTestLoader.loadTestsFromTestCase(test))
+        suite.addTest(sub_suite)
+
+    return suite
+
+
+def auto_discover_suite(loader=unittest.defaultTestLoader, pattern="test*.py"):
+    """Uses unittest auto-discovery to build a test suite (test_*.py pattern)"""
+
+    return loader.discover(start_dir=START_DIRECTORY, pattern=pattern, top_level_dir=TOP_LEVEL_DIRECTORY)
+
+
+def explicit_suite():
+    """Example of a manually constructed test suite.
+
+    Intended to be modified to create additional test suites
+    """
+    from warp.tests.test_adam import TestAdam
+    from warp.tests.test_arithmetic import TestArithmetic
+    from warp.tests.test_array import TestArray
+    from warp.tests.test_array_reduce import TestArrayReduce
+    from warp.tests.test_atomic import TestAtomic
+    from warp.tests.test_bool import TestBool
+    from warp.tests.test_builtins_resolution import TestBuiltinsResolution
+    from warp.tests.test_bvh import TestBvh
+    from warp.tests.test_closest_point_edge_edge import TestClosestPointEdgeEdgeMethods
+    from warp.tests.test_codegen import TestCodeGen
+    from warp.tests.test_compile_consts import TestConstants
+    from warp.tests.test_conditional import TestConditional
+    from warp.tests.test_copy import TestCopy
+    from warp.tests.test_ctypes import TestCTypes
+    from warp.tests.test_dense import TestDense
+    from warp.tests.test_devices import TestDevices
+    from warp.tests.test_dlpack import TestDLPack
+    from warp.tests.test_examples import TestExamples, TestFemExamples, TestSimExamples
+    from warp.tests.test_fabricarray import TestFabricArray
+    from warp.tests.test_fast_math import TestFastMath
+    from warp.tests.test_fem import TestFem
+    from warp.tests.test_fp16 import TestFp16
+    from warp.tests.test_func import TestFunc
+    from warp.tests.test_generics import TestGenerics
+    from warp.tests.test_grad import TestGrad
+    from warp.tests.test_grad_customs import TestGradCustoms
+    from warp.tests.test_hash_grid import TestHashGrid
+    from warp.tests.test_import import TestImport
+    from warp.tests.test_indexedarray import TestIndexedArray
+    from warp.tests.test_intersect import TestIntersect
+    from warp.tests.test_large import TestLarge
+    from warp.tests.test_launch import TestLaunch
+    from warp.tests.test_lerp import TestLerp
+    from warp.tests.test_lvalue import TestLValue
+    from warp.tests.test_marching_cubes import TestMarchingCubes
+    from warp.tests.test_mat import TestMat
+    from warp.tests.test_mat_lite import TestMatLite
+    from warp.tests.test_mat_scalar_ops import TestMatScalarOps
+    from warp.tests.test_math import TestMath
+    from warp.tests.test_matmul import TestMatmul
+    from warp.tests.test_matmul_lite import TestMatmulLite
+    from warp.tests.test_mesh import TestMesh
+    from warp.tests.test_mesh_query_aabb import TestMeshQueryAABBMethods
+    from warp.tests.test_mesh_query_point import TestMeshQueryPoint
+    from warp.tests.test_mesh_query_ray import TestMeshQueryRay
+    from warp.tests.test_mlp import TestMLP
+    from warp.tests.test_model import TestModel
+    from warp.tests.test_modules_lite import TestModuleLite
+    from warp.tests.test_multigpu import TestMultiGPU
+    from warp.tests.test_noise import TestNoise
+    from warp.tests.test_operators import TestOperators
+    from warp.tests.test_options import TestOptions
+    from warp.tests.test_pinned import TestPinned
+    from warp.tests.test_print import TestPrint
+    from warp.tests.test_quat import TestQuat
+    from warp.tests.test_rand import TestRand
+    from warp.tests.test_reload import TestReload
+    from warp.tests.test_rounding import TestRounding
+    from warp.tests.test_runlength_encode import TestRunlengthEncode
+    from warp.tests.test_smoothstep import TestSmoothstep
+    from warp.tests.test_snippet import TestSnippets
+    from warp.tests.test_sparse import TestSparse
+    from warp.tests.test_spatial import TestSpatial
+    from warp.tests.test_streams import TestStreams
+    from warp.tests.test_struct import TestStruct
+    from warp.tests.test_tape import TestTape
+    from warp.tests.test_torch import TestTorch
+    from warp.tests.test_transient_module import TestTransientModule
+    from warp.tests.test_types import TestTypes
+    from warp.tests.test_utils import TestUtils
+    from warp.tests.test_vec import TestVec
+    from warp.tests.test_vec_lite import TestVecLite
+    from warp.tests.test_vec_scalar_ops import TestVecScalarOps
+    from warp.tests.test_volume import TestVolume
+    from warp.tests.test_volume_write import TestVolumeWrite
+
+    test_classes = [
+        TestAdam,
+        TestArithmetic,
+        TestArray,
+        TestArrayReduce,
+        TestAtomic,
+        TestBool,
+        TestBuiltinsResolution,
+        TestBvh,
+        TestClosestPointEdgeEdgeMethods,
+        TestCodeGen,
+        TestConstants,
+        TestConditional,
+        TestCopy,
+        TestCTypes,
+        TestDense,
+        TestDevices,
+        TestDLPack,
+        TestExamples,
+        TestFemExamples,
+        TestSimExamples,
+        TestFabricArray,
+        TestFastMath,
+        TestFem,
+        TestFp16,
+        TestFunc,
+        TestGenerics,
+        TestGrad,
+        TestGradCustoms,
+        TestHashGrid,
+        TestImport,
+        TestIndexedArray,
+        TestIntersect,
+        TestLarge,
+        TestLaunch,
+        TestLerp,
+        TestLValue,
+        TestMarchingCubes,
+        TestMat,
+        TestMatLite,
+        TestMatScalarOps,
+        TestMath,
+        TestMatmul,
+        TestMatmulLite,
+        TestMesh,
+        TestMeshQueryAABBMethods,
+        TestMeshQueryPoint,
+        TestMeshQueryRay,
+        TestMLP,
+        TestModel,
+        TestModuleLite,
+        TestMultiGPU,
+        TestNoise,
+        TestOperators,
+        TestOptions,
+        TestPinned,
+        TestPrint,
+        TestQuat,
+        TestRand,
+        TestReload,
+        TestRounding,
+        TestRunlengthEncode,
+        TestSmoothstep,
+        TestSparse,
+        TestSnippets,
+        TestSpatial,
+        TestStreams,
+        TestStruct,
+        TestTape,
+        TestTorch,
+        TestTransientModule,
+        TestTypes,
+        TestUtils,
+        TestVec,
+        TestVecLite,
+        TestVecScalarOps,
+        TestVolume,
+        TestVolumeWrite,
+    ]
+
+    return _create_suite_from_test_classes(test_classes)
+
+
+def kit_suite():
+    """Tries to mimic the test suite used for testing omni.warp.core in Kit
+
+    Requires manual updates with test_ext.py for now.
+    """
+    from warp.tests.test_array import TestArray
+    from warp.tests.test_array_reduce import TestArrayReduce
+    from warp.tests.test_bvh import TestBvh
+    from warp.tests.test_codegen import TestCodeGen
+    from warp.tests.test_compile_consts import TestConstants
+    from warp.tests.test_conditional import TestConditional
+    from warp.tests.test_ctypes import TestCTypes
+    from warp.tests.test_devices import TestDevices
+    from warp.tests.test_dlpack import TestDLPack
+    from warp.tests.test_fabricarray import TestFabricArray
+    from warp.tests.test_func import TestFunc
+    from warp.tests.test_generics import TestGenerics
+    from warp.tests.test_grad_customs import TestGradCustoms
+    from warp.tests.test_hash_grid import TestHashGrid
+    from warp.tests.test_indexedarray import TestIndexedArray
+    from warp.tests.test_launch import TestLaunch
+    from warp.tests.test_marching_cubes import TestMarchingCubes
+    from warp.tests.test_mat_lite import TestMatLite
+    from warp.tests.test_math import TestMath
+    from warp.tests.test_matmul_lite import TestMatmulLite
+    from warp.tests.test_mesh import TestMesh
+    from warp.tests.test_mesh_query_aabb import TestMeshQueryAABBMethods
+    from warp.tests.test_mesh_query_point import TestMeshQuery
+    from warp.tests.test_mesh_query_ray import TestMeshQueryRay
+    from warp.tests.test_modules_lite import TestModuleLite
+    from warp.tests.test_noise import TestNoise
+    from warp.tests.test_operators import TestOperators
+    from warp.tests.test_quat import TestQuat
+    from warp.tests.test_rand import TestRand
+    from warp.tests.test_rounding import TestRounding
+    from warp.tests.test_runlength_encode import TestRunlengthEncode
+    from warp.tests.test_sparse import TestSparse
+    from warp.tests.test_streams import TestStreams
+    from warp.tests.test_tape import TestTape
+    from warp.tests.test_transient_module import TestTransientModule
+    from warp.tests.test_types import TestTypes
+    from warp.tests.test_utils import TestUtils
+    from warp.tests.test_vec_lite import TestVecLite
+    from warp.tests.test_volume import TestVolume
+    from warp.tests.test_volume_write import TestVolumeWrite
+
+    test_classes = [
+        TestArray,
+        TestArrayReduce,
+        TestBvh,
+        TestCodeGen,
+        TestConstants,
+        TestConditional,
+        TestCTypes,
+        TestDevices,
+        TestDLPack,
+        TestFabricArray,
+        TestFunc,
+        TestGenerics,
+        TestGradCustoms,
+        TestHashGrid,
+        TestIndexedArray,
+        TestLaunch,
+        TestMarchingCubes,
+        TestMatLite,
+        TestMath,
+        TestMatmulLite,
+        TestMesh,
+        TestMeshQueryAABBMethods,
+        TestMeshQuery,
+        TestMeshQueryRay,
+        TestModuleLite,
+        TestNoise,
+        TestOperators,
+        TestQuat,
+        TestRand,
+        TestRounding,
+        TestRunlengthEncode,
+        TestSparse,
+        TestStreams,
+        TestTape,
+        TestTransientModule,
+        TestTypes,
+        TestUtils,
+        TestVecLite,
+        TestVolume,
+        TestVolumeWrite,
+    ]
+
+    return _create_suite_from_test_classes(test_classes)
diff --git a/warp/tests/test_base.py b/warp/tests/unittest_utils.py
similarity index 54%
rename from warp/tests/test_base.py
rename to warp/tests/unittest_utils.py
index e5eb830f1..39f50fcaa 100644
--- a/warp/tests/test_base.py
+++ b/warp/tests/unittest_utils.py
@@ -9,15 +9,25 @@
 import ctypes.util
 import os
 import sys
+import unittest
 
 import numpy as np
+
 import warp as wp
 
+try:
+    import pxr  # noqa: F401
+
+    USD_AVAILABLE = True
+except ImportError as e:
+    USD_AVAILABLE = False
+    print(f"Skipping USD tests for reason: {e}")
+
 # default test mode (see get_test_devices())
 #   "basic" - only run on CPU and first GPU device
 #   "unique" - run on CPU and all unique GPU arches
 #   "all" - run on all devices
-test_mode = "basic"
+test_mode = "unique"
 
 try:
     if sys.platform == "win32":
@@ -29,7 +39,40 @@
     LIBC = None
 
 
+def get_unique_cuda_test_devices(mode=None):
+    """Returns a list of unique CUDA devices according to the CUDA arch.
+
+    If ``mode`` is ``None``, the ``global test_mode`` value will be used and
+    this list will be a subset of the devices returned from ``get_test_devices()``.
+    """
+
+    if mode is None:
+        global test_mode
+        mode = test_mode
+
+    if mode == "basic":
+        cuda_devices = [wp.get_device("cuda:0")]
+    else:
+        cuda_devices = wp.get_cuda_devices()
+
+    unique_cuda_devices = {}
+    for d in cuda_devices:
+        if d.arch not in unique_cuda_devices:
+            unique_cuda_devices[d.arch] = d
+
+    return list(unique_cuda_devices.values())
+
+
 def get_test_devices(mode=None):
+    """Returns a list of devices based on the mode selected.
+
+    Args:
+        mode (str, optional): The testing mode to specify which devices to include. If not provided or ``None``, the
+          ``global test_mode`` value will be used.
+          "basic" (default): Returns the CPU and the first GPU device when available.
+          "unique": Returns the CPU and all unique GPU architectures.
+          "all": Returns all available devices.
+    """
     if mode is None:
         global test_mode
         mode = test_mode
@@ -48,14 +91,7 @@ def get_test_devices(mode=None):
         if wp.is_cpu_available():
             devices.append(wp.get_device("cpu"))
 
-        cuda_devices = wp.get_cuda_devices()
-
-        unique_cuda_devices = {}
-        for d in cuda_devices:
-            if d.arch not in unique_cuda_devices:
-                unique_cuda_devices[d.arch] = d
-
-        devices.extend(list(unique_cuda_devices.values()))
+        devices.extend(get_unique_cuda_test_devices())
 
     # run on all devices
     elif mode == "all":
@@ -78,7 +114,8 @@ def begin(self):
         self.target = os.dup(self.saved.fileno())
 
         # create temporary capture stream
-        import io, tempfile
+        import io
+        import tempfile
 
         self.tempfile = io.TextIOWrapper(
             tempfile.TemporaryFile(buffering=0), encoding="utf-8", errors="replace", newline="", write_through=True
@@ -126,12 +163,8 @@ def __exit__(self, exc_type, exc_value, traceback):
             self.test.fail(f"Unexpected output:\n'{s.rstrip()}'")
 
 
-def assert_array_equal(result, expect):
-    a = result.numpy()
-    b = expect.numpy()
-
-    if (a == b).all() == False:
-        raise AssertionError(f"Unexpected result, got: {a} expected: {b}")
+def assert_array_equal(result: wp.array, expect: wp.array):
+    np.testing.assert_equal(result.numpy(), expect.numpy())
 
 
 def assert_np_equal(result, expect, tol=0.0):
@@ -139,7 +172,7 @@ def assert_np_equal(result, expect, tol=0.0):
     b = expect.flatten()
 
     if tol == 0.0:
-        if (a == b).all() == False:
+        if (a == b).all() is False:
             raise AssertionError(f"Unexpected result, got: {a} expected: {b}")
 
     else:
@@ -153,7 +186,6 @@ def assert_np_equal(result, expect, tol=0.0):
 
 # if check_output is True any output to stdout will be treated as an error
 def create_test_func(func, device, check_output, **kwargs):
-    
     # pass args to func
     def test_func(self):
         if check_output:
@@ -165,6 +197,11 @@ def test_func(self):
     return test_func
 
 
+def skip_test_func(self):
+    # A function to use so we can tell unittest that the test was skipped.
+    self.skipTest("No suitable devices to run the test.")
+
+
 def sanitize_identifier(s):
     """replace all non-identifier characters with '_'"""
 
@@ -174,15 +211,25 @@ def sanitize_identifier(s):
     else:
         import re
 
-        return re.sub("\W|^(?=\d)", "_", s)
+        return re.sub(r"\W|^(?=\d)", "_", s)
 
 
 def add_function_test(cls, name, func, devices=None, check_output=True, **kwargs):
     if devices is None:
         setattr(cls, name, create_test_func(func, None, check_output, **kwargs))
+    elif isinstance(devices, list):
+        if not devices:
+            # No devices to run this test
+            setattr(cls, name, skip_test_func)
+        else:
+            for device in devices:
+                setattr(
+                    cls,
+                    name + "_" + sanitize_identifier(device),
+                    create_test_func(func, device, check_output, **kwargs),
+                )
     else:
-        for device in devices:
-            setattr(cls, name + "_" + sanitize_identifier(device), create_test_func(func, device, check_output, **kwargs))
+        setattr(cls, name + "_" + sanitize_identifier(devices), create_test_func(func, devices, check_output, **kwargs))
 
 
 def add_kernel_test(cls, kernel, dim, name=None, expect=None, inputs=None, devices=None):
@@ -217,13 +264,79 @@ def test_func(self, device):
 
     # register test func with class for the given devices
     for d in devices:
-        # use a lambda to forward the device to the inner test function
-        test_lambda = lambda test, device=d: test_func(test, device)
-        setattr(cls, name + "_" + sanitize_identifier(d), test_lambda)
+        # use a function to forward the device to the inner test function
+        def test_func_wrapper(test, device=d):
+            test_func(test, device)
+
+        setattr(cls, name + "_" + sanitize_identifier(d), test_func_wrapper)
 
 
-# helper that first calls the test function to generate all kernel permuations
+# helper that first calls the test function to generate all kernel permutations
 # so that compilation is done in one-shot instead of per-test
 def add_function_test_register_kernel(cls, name, func, devices=None, **kwargs):
     func(None, None, **kwargs, register_kernels=True)
     add_function_test(cls, name, func, devices=devices, **kwargs)
+
+
+class TeamCityTestResult(unittest.TextTestResult):
+    """This class will report each test result to TeamCity"""
+
+    def __init__(self, stream, descriptions, verbosity):
+        super(TeamCityTestResult, self).__init__(stream, descriptions, verbosity)
+
+    def addSuccess(self, test):
+        super(TeamCityTestResult, self).addSuccess(test)
+        self.reportSuccess(test)
+
+    def addError(self, test, err):
+        super(TeamCityTestResult, self).addError(test, err)
+        self.reportFailure(test)
+
+    def addFailure(self, test, err):
+        super(TeamCityTestResult, self).addFailure(test, err)
+        self.reportFailure(test)
+
+    def addSkip(self, test, reason):
+        super(TeamCityTestResult, self).addSkip(test, reason)
+
+    def addExpectedFailure(self, test, err):
+        super(TeamCityTestResult, self).addExpectedFailure(test, err)
+        self.reportSuccess(test)
+
+    def addUnexpectedSuccess(self, test):
+        super(TeamCityTestResult, self).addUnexpectedSuccess(test)
+        self.reportFailure(test)
+
+    def reportSuccess(self, test):
+        test_id = test.id()
+        print(f"##teamcity[testStarted name='{test_id}']")
+        print(f"##teamcity[testFinished name='{test_id}']")
+
+    def reportFailure(self, test):
+        test_id = test.id()
+        print(f"##teamcity[testStarted name='{test_id}']")
+        print(f"##teamcity[testFailed name='{test_id}']")
+        print(f"##teamcity[testFinished name='{test_id}']")
+
+
+class TeamCityTestRunner(unittest.TextTestRunner):
+    """Test runner that will report test results to TeamCity if running in TeamCity"""
+
+    def __init__(self, **kwargs):
+        self.running_in_teamcity = os.environ.get("TEAMCITY_VERSION") is not None
+        if self.running_in_teamcity:
+            kwargs["resultclass"] = TeamCityTestResult
+        super(TeamCityTestRunner, self).__init__(**kwargs)
+
+    def run(self, test, name):
+        if self.running_in_teamcity:
+            print(f"##teamcity[testSuiteStarted name='{name}']")
+
+        result = super(TeamCityTestRunner, self).run(test)
+
+        if self.running_in_teamcity:
+            print(f"##teamcity[testSuiteFinished name='{name}']")
+            if not result.wasSuccessful():
+                print("##teamcity[buildStatus status='FAILURE']")
+
+        return result
diff --git a/warp/tests/test_misc.py b/warp/tests/unused_test_misc.py
similarity index 65%
rename from warp/tests/test_misc.py
rename to warp/tests/unused_test_misc.py
index af84ba7e1..7bcded557 100644
--- a/warp/tests/test_misc.py
+++ b/warp/tests/unused_test_misc.py
@@ -1,8 +1,17 @@
-import warp as wp
+# Copyright (c) 2021 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import numpy as np
 
+import warp as wp
+
 wp.init()
 
+
 @wp.kernel
 def arange(out: wp.array(dtype=int)):
     tid = wp.tid()
@@ -28,7 +37,7 @@ def arange(out: wp.array(dtype=int)):
 
 graph = wp.capture_end()
 
-#---------------------------------------
+# ---------------------------------------
 
 ref = np.arange(0, n, dtype=int)
 wp.capture_launch(graph)
@@ -37,7 +46,7 @@ def arange(out: wp.array(dtype=int)):
     print(arrays[i].numpy())
 
 
-#---------------------------------------
+# ---------------------------------------
 
 n = 16
 arrays = []
@@ -49,7 +58,7 @@ def arange(out: wp.array(dtype=int)):
 for i in range(5):
     cmd.set_dim(n)
     cmd.set_param(arrays[i])
-    
+
     cmd.update_graph()
 
 
@@ -60,4 +69,3 @@ def arange(out: wp.array(dtype=int)):
 
 for i in range(5):
     print(arrays[i].numpy())
-
diff --git a/warp/tests/test_debug.py b/warp/tests/walkthough_debug.py
similarity index 88%
rename from warp/tests/test_debug.py
rename to warp/tests/walkthough_debug.py
index af2570016..2d501956c 100644
--- a/warp/tests/test_debug.py
+++ b/warp/tests/walkthough_debug.py
@@ -49,10 +49,7 @@
 #
 ####################################################################################################
 
-import unittest
-
 import warp as wp
-from warp.tests.test_base import *
 
 # The init() function prints the directory of the kernel cache which contains the .cpp files
 # generated from Warp kernels. You can put breakpoints in these C++ files through Visual Studio Code,
@@ -67,7 +64,7 @@
 
 
 @wp.kernel
-def test_breakpoint(n: int):
+def example_breakpoint(n: int):
     a = int(0)
 
     for i in range(0, n):
@@ -83,16 +80,6 @@ def test_breakpoint(n: int):
     wp.expect_eq(a, 5)
 
 
-def register(parent):
-    class TestDebug(parent):
-        pass
-
-    add_kernel_test(TestDebug, name="test_breakpoint", kernel=test_breakpoint, dim=1, inputs=[10], devices=["cpu"])
-
-    return TestDebug
-
-
 if __name__ == "__main__":
     wp.build.clear_kernel_cache()
-    _ = register(unittest.TestCase)
-    unittest.main(verbosity=2, failfast=True)
+    wp.launch(example_breakpoint, dim=1, inputs=[10], device="cpu")
diff --git a/warp/thirdparty/unittest_parallel.py b/warp/thirdparty/unittest_parallel.py
index 4face9684..a5eefca0d 100644
--- a/warp/thirdparty/unittest_parallel.py
+++ b/warp/thirdparty/unittest_parallel.py
@@ -1,11 +1,22 @@
 # Licensed under the MIT License
 # https://github.com/craigahobbs/unittest-parallel/blob/main/LICENSE
 
+# SPDX-FileCopyrightText: Copyright (c) 2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: LicenseRef-NvidiaProprietary
+#
+# NVIDIA CORPORATION, its affiliates and licensors retain all intellectual
+# property and proprietary rights in and to this material, related
+# documentation and any modifications thereto. Any use, reproduction,
+# disclosure or distribution of this material and related documentation
+# without an express license agreement from NVIDIA CORPORATION or
+# its affiliates is strictly prohibited.
+
 """
 unittest-parallel command-line script main module
 """
 
 import argparse
+import concurrent.futures  # NVIDIA Modification
 import multiprocessing
 import os
 import sys
@@ -15,7 +26,20 @@
 from contextlib import contextmanager
 from io import StringIO
 
-import coverage
+import warp.tests.unittest_suites  # NVIDIA Modification
+
+try:
+    import coverage
+
+    COVERAGE_AVAILABLE = True  # NVIDIA Modification
+except ImportError:
+    COVERAGE_AVAILABLE = False  # NVIDIA Modification
+
+
+# The following variables are NVIDIA Modifications
+RUNNING_IN_TEAMCITY = os.environ.get("TEAMCITY_VERSION") is not None
+TEST_SUITE_NAME = "WarpTests"
+START_DIRECTORY = os.path.dirname(__file__)  # The directory to start test discovery
 
 
 def main(argv=None):
@@ -25,8 +49,8 @@ def main(argv=None):
 
     # Command line arguments
     parser = argparse.ArgumentParser(prog="unittest-parallel")
-    parser.add_argument("-v", "--verbose", action="store_const", const=2, default=1, help="Verbose output")
-    parser.add_argument("-q", "--quiet", dest="verbose", action="store_const", const=0, default=1, help="Quiet output")
+    # parser.add_argument("-v", "--verbose", action="store_const", const=2, default=1, help="Verbose output")
+    parser.add_argument("-q", "--quiet", dest="verbose", action="store_const", const=0, default=2, help="Quiet output")
     parser.add_argument("-f", "--failfast", action="store_true", default=False, help="Stop on first fail or error")
     parser.add_argument(
         "-b", "--buffer", action="store_true", default=False, help="Buffer stdout and stderr during tests"
@@ -38,9 +62,6 @@ def main(argv=None):
         type=_convert_select_pattern,
         help="Only run tests which match the given substring",
     )
-    parser.add_argument(
-        "-s", "--start-directory", metavar="START", default=".", help="Directory to start discovery ('.' default)"
-    )
     parser.add_argument(
         "-p", "--pattern", metavar="PATTERN", default="test*.py", help="Pattern to match tests ('test*.py' default)"
     )
@@ -59,11 +80,19 @@ def main(argv=None):
         default=0,
         help="The number of test processes (default is 0, all cores)",
     )
+    group_parallel.add_argument(
+        "-m",
+        "--maxjobs",
+        metavar="MAXCOUNT",
+        type=int,
+        default=8,
+        help="The maximum number of test processes (default is 8)",
+    )  # NVIDIA Modification
     group_parallel.add_argument(
         "--level",
         choices=["module", "class", "test"],
-        default="module",
-        help="Set the test parallelism level (default is 'module')",
+        default="class",
+        help="Set the test parallelism level (default is 'class')",
     )
     group_parallel.add_argument(
         "--disable-process-pooling",
@@ -71,40 +100,45 @@ def main(argv=None):
         default=False,
         help="Do not reuse processes used to run test suites",
     )
+    group_parallel.add_argument(
+        "--disable-concurrent-futures",
+        action="store_true",
+        default=False,
+        help="Use multiprocessing instead of concurrent.futures.",
+    )  # NVIDIA Modification
+    group_parallel.add_argument(
+        "--serial-fallback",
+        action="store_true",
+        default=False,
+        help="Run in a single-process (no spawning) mode without multiprocessing or concurrent.futures.",
+    )  # NVIDIA Modification
     group_coverage = parser.add_argument_group("coverage options")
     group_coverage.add_argument("--coverage", action="store_true", help="Run tests with coverage")
     group_coverage.add_argument("--coverage-branch", action="store_true", help="Run tests with branch coverage")
-    group_coverage.add_argument("--coverage-rcfile", metavar="RCFILE", help="Specify coverage configuration file")
-    group_coverage.add_argument(
-        "--coverage-include",
-        metavar="PAT",
-        action="append",
-        help="Include only files matching one of these patterns. Accepts shell-style (quoted) wildcards.",
-    )
-    group_coverage.add_argument(
-        "--coverage-omit",
-        metavar="PAT",
-        action="append",
-        help="Omit files matching one of these patterns. Accepts shell-style (quoted) wildcards.",
-    )
     group_coverage.add_argument(
-        "--coverage-source",
-        metavar="SRC",
-        action="append",
-        help="A list of packages or directories of code to be measured",
+        "--coverage-html",
+        metavar="DIR",
+        help="Generate coverage HTML report",
+        default=os.path.join(START_DIRECTORY, "..", "..", "htmlcov"),
     )
-    group_coverage.add_argument("--coverage-html", metavar="DIR", help="Generate coverage HTML report")
     group_coverage.add_argument("--coverage-xml", metavar="FILE", help="Generate coverage XML report")
     group_coverage.add_argument(
         "--coverage-fail-under", metavar="MIN", type=float, help="Fail if coverage percentage under min"
     )
     args = parser.parse_args(args=argv)
+
     if args.coverage_branch:
         args.coverage = args.coverage_branch
 
+    if args.coverage and not COVERAGE_AVAILABLE:
+        parser.exit(
+            status=2, message="--coverage was used, but coverage was not found. Is it installed?\n"
+        )  # NVIDIA Modification
+
     process_count = max(0, args.jobs)
     if process_count == 0:
         process_count = multiprocessing.cpu_count()
+    process_count = min(process_count, args.maxjobs)  # NVIDIA Modification
 
     # Create the temporary directory (for coverage files)
     with tempfile.TemporaryDirectory() as temp_dir:
@@ -113,9 +147,10 @@ def main(argv=None):
             test_loader = unittest.TestLoader()
             if args.testNamePatterns:
                 test_loader.testNamePatterns = args.testNamePatterns
-            discover_suite = test_loader.discover(
-                args.start_directory, pattern=args.pattern, top_level_dir=args.top_level_directory
-            )
+            discover_suite = warp.tests.unittest_suites.auto_discover_suite(
+                test_loader, args.pattern
+            )  # NVIDIA Modification
+            # discover_suite = warp.tests.unittest_suites.explicit_suite()
 
         # Get the parallelizable test suites
         if args.level == "test":
@@ -128,26 +163,69 @@ def main(argv=None):
         # Don't use more processes than test suites
         process_count = max(1, min(len(test_suites), process_count))
 
-        # Report test suites and processes
-        print(
-            f"Running {len(test_suites)} test suites ({discover_suite.countTestCases()} total tests) across {process_count} processes",
-            file=sys.stderr,
-        )
-        if args.verbose > 1:
-            print(file=sys.stderr)
+        if RUNNING_IN_TEAMCITY:
+            print(f"##teamcity[testSuiteStarted name='{TEST_SUITE_NAME}']")  # NVIDIA Modification for TC
+
+        if not args.serial_fallback:
+            # Report test suites and processes
+            print(
+                f"Running {len(test_suites)} test suites ({discover_suite.countTestCases()} total tests) across {process_count} processes",
+                file=sys.stderr,
+            )
+            if args.verbose > 1:
+                print(file=sys.stderr)
+
+            # Run the tests in parallel
+            start_time = time.perf_counter()
+
+            if args.disable_concurrent_futures:
+                multiprocessing_context = multiprocessing.get_context(method="spawn")
+                maxtasksperchild = 1 if args.disable_process_pooling else None
+                with multiprocessing_context.Pool(
+                    process_count,
+                    maxtasksperchild=maxtasksperchild,
+                    initializer=set_worker_cache,
+                    initargs=(args, temp_dir),
+                ) as pool, multiprocessing.Manager() as manager:
+                    test_manager = ParallelTestManager(manager, args, temp_dir)
+                    results = pool.map(test_manager.run_tests, test_suites)
+            else:
+                # NVIDIA Modification added concurrent.futures
+                with concurrent.futures.ProcessPoolExecutor(
+                    max_workers=process_count,
+                    mp_context=multiprocessing.get_context(method="spawn"),
+                    initializer=set_worker_cache,
+                    initargs=(args, temp_dir),
+                ) as executor, multiprocessing.Manager() as manager:
+                    test_manager = ParallelTestManager(manager, args, temp_dir)
+                    results = list(executor.map(test_manager.run_tests, test_suites, timeout=3600))
+        else:
+            # This entire path is an NVIDIA Modification
+
+            # Report test suites and processes
+            print(f"Running {discover_suite.countTestCases()} total tests (serial fallback)", file=sys.stderr)
+            if args.verbose > 1:
+                print(file=sys.stderr)
+
+            import warp as wp
+
+            # force rebuild of all kernels
+            wp.build.clear_kernel_cache()
+            print("Cleared Warp kernel cache")
+
+            # Run the tests in serial
+            start_time = time.perf_counter()
+
+            with multiprocessing.Manager() as manager:
+                test_manager = ParallelTestManager(manager, args, temp_dir)
+                results = [test_manager.run_tests(discover_suite)]
 
-        # Run the tests in parallel
-        start_time = time.perf_counter()
-        multiprocessing_context = multiprocessing.get_context(method="spawn")
-        maxtasksperchild = 1 if args.disable_process_pooling else None
-        with multiprocessing_context.Pool(
-            process_count, maxtasksperchild=maxtasksperchild
-        ) as pool, multiprocessing.Manager() as manager:
-            test_manager = ParallelTestManager(manager, args, temp_dir)
-            results = pool.map(test_manager.run_tests, test_suites)
         stop_time = time.perf_counter()
         test_duration = stop_time - start_time
 
+        if RUNNING_IN_TEAMCITY:
+            print(f"##teamcity[testSuiteFinished name='{TEST_SUITE_NAME}']")  # NVIDIA Modification for TC
+
         # Aggregate parallel test run results
         tests_run = 0
         errors = []
@@ -195,14 +273,15 @@ def main(argv=None):
 
         # Return an error status on failure
         if not is_success:
+            if RUNNING_IN_TEAMCITY:
+                print("##teamcity[buildStatus status='FAILURE']")  # NVIDIA Modification for TC
             parser.exit(status=len(errors) + len(failures) + unexpected_successes)
 
         # Coverage?
         if args.coverage:
             # Combine the coverage files
             cov_options = {}
-            if args.coverage_rcfile is not None:
-                cov_options["config_file"] = args.coverage_rcfile
+            cov_options["config_file"] = True  # Grab configuration from pyproject.toml (must install coverage[toml])
             cov = coverage.Coverage(**cov_options)
             cov.combine(data_paths=[os.path.join(temp_dir, x) for x in os.listdir(temp_dir)])
 
@@ -225,7 +304,7 @@ def main(argv=None):
 
 
 def _convert_select_pattern(pattern):
-    if not "*" in pattern:
+    if "*" not in pattern:
         return f"*{pattern}*"
     return pattern
 
@@ -242,12 +321,9 @@ def _coverage(args, temp_dir):
         cov_options = {
             "branch": args.coverage_branch,
             "data_file": coverage_file.name,
-            "include": args.coverage_include,
-            "omit": (args.coverage_omit if args.coverage_omit else []) + [__file__],
-            "source": args.coverage_source,
+            # NVIDIA Modification removed unneeded options
         }
-        if args.coverage_rcfile is not None:
-            cov_options["config_file"] = args.coverage_rcfile
+        cov_options["config_file"] = True  # Grab configuration from pyproject.toml (must install coverage[toml])
         cov = coverage.Coverage(**cov_options)
         try:
             # Start measuring code coverage
@@ -307,7 +383,9 @@ def run_tests(self, test_suite):
         with _coverage(self.args, self.temp_dir):
             runner = unittest.TextTestRunner(
                 stream=StringIO(),
-                resultclass=ParallelTextTestResult,
+                resultclass=ParallelTeamCityTestResult
+                if RUNNING_IN_TEAMCITY
+                else ParallelTextTestResult,  # NVIDIA Modification for TC
                 verbosity=self.args.verbose,
                 failfast=self.args.failfast,
                 buffer=self.args.buffer,
@@ -384,3 +462,130 @@ def addUnexpectedSuccess(self, test):
 
     def printErrors(self):
         pass
+
+
+# NVIDIA Modifications from here until end of file
+
+
+def set_worker_cache(args, temp_dir):
+    """This function is run at the start of ever new process. It changes the Warp cache to avoid conflicts."""
+
+    with _coverage(args, temp_dir):
+        import warp as wp
+        from warp.thirdparty import appdirs
+
+        pid = os.getpid()
+        cache_root_dir = appdirs.user_cache_dir(
+            appname="warp", appauthor="NVIDIA Corporation", version=f"{wp.config.version}-{pid}"
+        )
+
+        wp.config.kernel_cache_dir = cache_root_dir
+
+        # wp.config.verify_cuda = True  # TESTING
+
+        wp.build.clear_kernel_cache()
+
+
+def _tc_escape(s):
+    s = s.replace("|", "||")
+    s = s.replace("\n", "|n")
+    s = s.replace("\r", "|r")
+    s = s.replace("'", "|'")
+    s = s.replace("[", "|[")
+    s = s.replace("]", "|]")
+    return s
+
+
+class ParallelTeamCityTestResult(unittest.TextTestResult):
+    def __init__(self, stream, descriptions, verbosity):
+        stream = type(stream)(sys.stderr)
+        super().__init__(stream, descriptions, verbosity)
+
+    def startTest(self, test):
+        if self.showAll:
+            self.stream.writeln(f"{self.getDescription(test)} ...")
+            self.stream.flush()
+        self.start_time = time.perf_counter_ns()
+        super(unittest.TextTestResult, self).startTest(test)
+
+    def _add_helper(self, test, dots_message, show_all_message):
+        if self.showAll:
+            self.stream.writeln(f"{self.getDescription(test)} ... {show_all_message}")
+        elif self.dots:
+            self.stream.write(dots_message)
+        self.stream.flush()
+
+    def addSuccess(self, test):
+        super(unittest.TextTestResult, self).addSuccess(test)
+        self._add_helper(test, ".", "ok")
+        self.reportSuccess(test)
+
+    def addError(self, test, err):
+        super(unittest.TextTestResult, self).addError(test, err)
+        self._add_helper(test, "E", "ERROR")
+        self.reportFailure(test, err)
+
+    def addFailure(self, test, err):
+        super(unittest.TextTestResult, self).addFailure(test, err)
+        self._add_helper(test, "F", "FAIL")
+        self.reportFailure(test, err)
+
+    def addSkip(self, test, reason):
+        super(unittest.TextTestResult, self).addSkip(test, reason)
+        self._add_helper(test, "s", f"skipped {reason!r}")
+        self.reportIgnored(test, reason)
+
+    def addExpectedFailure(self, test, err):
+        super(unittest.TextTestResult, self).addExpectedFailure(test, err)
+        self._add_helper(test, "x", "expected failure")
+        self.reportSuccess(test)
+
+    def addUnexpectedSuccess(self, test):
+        super(unittest.TextTestResult, self).addUnexpectedSuccess(test)
+        self._add_helper(test, "u", "unexpected success")
+        self.reportFailure(test, "unexpected success")
+
+    def addSubTest(self, test, subtest, err):
+        super(unittest.TextTestResult, self).addSubTest(test, subtest, err)
+        if err is not None:
+            self._add_helper(test, "E", "ERROR")
+            self.reportSubTestFailure(test, err)
+
+    def printErrors(self):
+        pass
+
+    def reportIgnored(self, test, reason):
+        test_id = test.id()
+        reason_str = str(reason)
+        print(reason_str)
+        self.stream.writeln(f"##teamcity[testIgnored name='{test_id}' message='{_tc_escape(str(reason))}']")
+        self.stream.flush()
+
+    def reportSuccess(self, test):
+        duration = round((time.perf_counter_ns() - self.start_time) / 1e6)  # [ms]
+        test_id = test.id()
+        self.stream.writeln(f"##teamcity[testStarted name='{test_id}']")
+        self.stream.writeln(f"##teamcity[testFinished name='{test_id}' duration='{duration}']")
+        self.stream.flush()
+
+    def reportFailure(self, test, err):
+        test_id = test.id()
+        self.stream.writeln(f"##teamcity[testStarted name='{test_id}']")
+        self.stream.writeln(
+            f"##teamcity[testFailed name='{test_id}' message='{_tc_escape(str(err[1]))}' details='{_tc_escape(str(err[2]))}']"
+        )
+        self.stream.writeln(f"##teamcity[testFinished name='{test_id}']")
+        self.stream.flush()
+
+    def reportSubTestFailure(self, test, err):
+        test_id = test.id()
+        self.stream.writeln(f"##teamcity[testStarted name='{test_id}']")
+        self.stream.writeln(
+            f"##teamcity[testFailed name='{test_id}' message='{_tc_escape(str(err[1]))}' details='{_tc_escape(self._exc_info_to_string(err, test))}']"
+        )
+        self.stream.writeln(f"##teamcity[testFinished name='{test_id}']")
+        self.stream.flush()
+
+
+if __name__ == "__main__":  # pragma: no cover
+    main()

From e41700fe55699fbd65677a19561350e252bec422 Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Tue, 19 Dec 2023 13:55:03 -0800
Subject: [PATCH 46/50] CI/CD Pipelines for Documentation

---
 .github/workflows/sphinx.yml                  |   30 +
 .gitignore                                    |    4 +-
 .gitlab-ci.yml                                |   24 +
 build_docs.py                                 |   41 +-
 build_lib.py                                  |    7 +
 docs/_build/html/.buildinfo                   |    4 -
 .../_build/html/_images/compiler_pipeline.png |  Bin 36891 -> 0 bytes
 docs/_build/html/_images/header.png           |  Bin 1646470 -> 0 bytes
 docs/_build/html/_images/joint_transforms.png |  Bin 479613 -> 0 bytes
 docs/_build/html/_sources/basics.rst.txt      |  214 -
 .../html/_sources/configuration.rst.txt       |  138 -
 docs/_build/html/_sources/debugging.rst.txt   |   86 -
 docs/_build/html/_sources/faq.rst.txt         |  112 -
 docs/_build/html/_sources/index.rst.txt       |  192 -
 .../_build/html/_sources/installation.rst.txt |   93 -
 docs/_build/html/_sources/limitations.rst.txt |   56 -
 .../html/_sources/modules/devices.rst.txt     |  200 -
 docs/_build/html/_sources/modules/fem.rst.txt |  386 --
 .../html/_sources/modules/functions.rst.txt   | 1992 ---------
 .../_sources/modules/interoperability.rst.txt |  146 -
 .../html/_sources/modules/profiling.rst.txt   |   44 -
 .../html/_sources/modules/runtime.rst.txt     | 1491 -------
 docs/_build/html/_sources/modules/sim.rst.txt |  157 -
 .../html/_sources/modules/sparse.rst.txt      |   19 -
 docs/_build/html/_static/basic.css            |  925 -----
 docs/_build/html/_static/check-solid.svg      |    4 -
 docs/_build/html/_static/clipboard.min.js     |    7 -
 docs/_build/html/_static/copy-button.svg      |    5 -
 docs/_build/html/_static/copybutton.css       |   94 -
 docs/_build/html/_static/copybutton.js        |  248 --
 docs/_build/html/_static/copybutton_funcs.js  |   73 -
 docs/_build/html/_static/custom.css           |   37 -
 docs/_build/html/_static/debug.css            |   69 -
 docs/_build/html/_static/doctools.js          |  156 -
 .../html/_static/documentation_options.js     |   13 -
 docs/_build/html/_static/file.png             |  Bin 286 -> 0 bytes
 docs/_build/html/_static/language_data.js     |  199 -
 docs/_build/html/_static/logo-dark-mode.png   |  Bin 8504 -> 0 bytes
 docs/_build/html/_static/logo-light-mode.png  |  Bin 10324 -> 0 bytes
 docs/_build/html/_static/minus.png            |  Bin 90 -> 0 bytes
 docs/_build/html/_static/plus.png             |  Bin 90 -> 0 bytes
 docs/_build/html/_static/pygments.css         |  258 --
 .../html/_static/scripts/furo-extensions.js   |    0
 docs/_build/html/_static/scripts/furo.js      |    3 -
 .../html/_static/scripts/furo.js.LICENSE.txt  |    7 -
 docs/_build/html/_static/scripts/furo.js.map  |    1 -
 docs/_build/html/_static/searchtools.js       |  574 ---
 docs/_build/html/_static/skeleton.css         |  296 --
 docs/_build/html/_static/sphinx_highlight.js  |  154 -
 .../html/_static/styles/furo-extensions.css   |    2 -
 .../_static/styles/furo-extensions.css.map    |    1 -
 docs/_build/html/_static/styles/furo.css      |    2 -
 docs/_build/html/_static/styles/furo.css.map  |    1 -
 docs/_build/html/basics.html                  |  525 ---
 docs/_build/html/configuration.html           |  530 ---
 docs/_build/html/debugging.html               |  406 --
 docs/_build/html/faq.html                     |  422 --
 docs/_build/html/genindex.html                | 1980 ---------
 docs/_build/html/index.html                   |  566 ---
 docs/_build/html/installation.html            |  396 --
 docs/_build/html/limitations.html             |  382 --
 docs/_build/html/modules/devices.html         |  511 ---
 docs/_build/html/modules/fem.html             | 2063 ----------
 docs/_build/html/modules/functions.html       | 3001 --------------
 .../_build/html/modules/interoperability.html |  659 ---
 docs/_build/html/modules/profiling.html       |  388 --
 docs/_build/html/modules/runtime.html         | 2743 ------------
 docs/_build/html/modules/sim.html             | 3661 -----------------
 docs/_build/html/modules/sparse.html          |  758 ----
 docs/_build/html/objects.inv                  |  Bin 5391 -> 0 bytes
 docs/_build/html/py-modindex.html             |  322 --
 docs/_build/html/search.html                  |  291 --
 docs/_build/html/searchindex.js               |    1 -
 docs/conf.py                                  |    7 +-
 docs/modules/functions.rst                    |   66 +-
 docs/requirements.txt                         |    2 +
 .../ci/building/build-linux-aarch64/build.sh  |    1 -
 .../building/build-windows-x86_64/build.bat   |    1 -
 warp/__init__.py                              |    1 -
 warp/context.py                               |    2 +-
 warp/stubs.py                                 |  116 +-
 81 files changed, 136 insertions(+), 28230 deletions(-)
 create mode 100644 .github/workflows/sphinx.yml
 create mode 100644 .gitlab-ci.yml
 delete mode 100644 docs/_build/html/.buildinfo
 delete mode 100644 docs/_build/html/_images/compiler_pipeline.png
 delete mode 100644 docs/_build/html/_images/header.png
 delete mode 100644 docs/_build/html/_images/joint_transforms.png
 delete mode 100644 docs/_build/html/_sources/basics.rst.txt
 delete mode 100644 docs/_build/html/_sources/configuration.rst.txt
 delete mode 100644 docs/_build/html/_sources/debugging.rst.txt
 delete mode 100644 docs/_build/html/_sources/faq.rst.txt
 delete mode 100644 docs/_build/html/_sources/index.rst.txt
 delete mode 100644 docs/_build/html/_sources/installation.rst.txt
 delete mode 100644 docs/_build/html/_sources/limitations.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/devices.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/fem.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/functions.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/interoperability.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/profiling.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/runtime.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/sim.rst.txt
 delete mode 100644 docs/_build/html/_sources/modules/sparse.rst.txt
 delete mode 100644 docs/_build/html/_static/basic.css
 delete mode 100644 docs/_build/html/_static/check-solid.svg
 delete mode 100644 docs/_build/html/_static/clipboard.min.js
 delete mode 100644 docs/_build/html/_static/copy-button.svg
 delete mode 100644 docs/_build/html/_static/copybutton.css
 delete mode 100644 docs/_build/html/_static/copybutton.js
 delete mode 100644 docs/_build/html/_static/copybutton_funcs.js
 delete mode 100644 docs/_build/html/_static/custom.css
 delete mode 100644 docs/_build/html/_static/debug.css
 delete mode 100644 docs/_build/html/_static/doctools.js
 delete mode 100644 docs/_build/html/_static/documentation_options.js
 delete mode 100644 docs/_build/html/_static/file.png
 delete mode 100644 docs/_build/html/_static/language_data.js
 delete mode 100644 docs/_build/html/_static/logo-dark-mode.png
 delete mode 100644 docs/_build/html/_static/logo-light-mode.png
 delete mode 100644 docs/_build/html/_static/minus.png
 delete mode 100644 docs/_build/html/_static/plus.png
 delete mode 100644 docs/_build/html/_static/pygments.css
 delete mode 100644 docs/_build/html/_static/scripts/furo-extensions.js
 delete mode 100644 docs/_build/html/_static/scripts/furo.js
 delete mode 100644 docs/_build/html/_static/scripts/furo.js.LICENSE.txt
 delete mode 100644 docs/_build/html/_static/scripts/furo.js.map
 delete mode 100644 docs/_build/html/_static/searchtools.js
 delete mode 100644 docs/_build/html/_static/skeleton.css
 delete mode 100644 docs/_build/html/_static/sphinx_highlight.js
 delete mode 100644 docs/_build/html/_static/styles/furo-extensions.css
 delete mode 100644 docs/_build/html/_static/styles/furo-extensions.css.map
 delete mode 100644 docs/_build/html/_static/styles/furo.css
 delete mode 100644 docs/_build/html/_static/styles/furo.css.map
 delete mode 100644 docs/_build/html/basics.html
 delete mode 100644 docs/_build/html/configuration.html
 delete mode 100644 docs/_build/html/debugging.html
 delete mode 100644 docs/_build/html/faq.html
 delete mode 100644 docs/_build/html/genindex.html
 delete mode 100644 docs/_build/html/index.html
 delete mode 100644 docs/_build/html/installation.html
 delete mode 100644 docs/_build/html/limitations.html
 delete mode 100644 docs/_build/html/modules/devices.html
 delete mode 100644 docs/_build/html/modules/fem.html
 delete mode 100644 docs/_build/html/modules/functions.html
 delete mode 100644 docs/_build/html/modules/interoperability.html
 delete mode 100644 docs/_build/html/modules/profiling.html
 delete mode 100644 docs/_build/html/modules/runtime.html
 delete mode 100644 docs/_build/html/modules/sim.html
 delete mode 100644 docs/_build/html/modules/sparse.html
 delete mode 100644 docs/_build/html/objects.inv
 delete mode 100644 docs/_build/html/py-modindex.html
 delete mode 100644 docs/_build/html/search.html
 delete mode 100644 docs/_build/html/searchindex.js

diff --git a/.github/workflows/sphinx.yml b/.github/workflows/sphinx.yml
new file mode 100644
index 000000000..7f80e9324
--- /dev/null
+++ b/.github/workflows/sphinx.yml
@@ -0,0 +1,30 @@
+name: Deploy Sphinx documentation to Pages
+
+# Runs on pushes targeting the default branch
+on:
+  push:
+    branches: [github-pages]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write
+    steps:
+      - uses: actions/checkout@v4
+      - name: Build HTML
+        uses: ammaraskar/sphinx-action@master
+        with:
+          docs-folder: "docs/"
+          build-command: "python ../build_docs.py"
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: html-docs
+          path: docs/_build/html/
+      - name: Deploy
+        uses: peaceiris/actions-gh-pages@v3
+        if: github.ref == 'refs/heads/github-pages'
+        with:
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          publish_dir: docs/_build/html
diff --git a/.gitignore b/.gitignore
index 32462e43d..4e9d2c85e 100644
--- a/.gitignore
+++ b/.gitignore
@@ -18,8 +18,7 @@ archive
 /_repo
 examples/assets/.thumbs
 /examples/tmp
-/docs/_build/doctrees
-/docs/_build/html/_static/fonts
+/docs/_build
 /warp_lang.egg-info
 exts/omni.warp/omni/warp/ogn/tests/usd
 build/lib/
@@ -29,4 +28,5 @@ exts/omni.warp/config/extension.gen.toml
 /build
 /dist
 .coverage
+.cache
 warp/tests/outputs
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
new file mode 100644
index 000000000..9fe2f2c2c
--- /dev/null
+++ b/.gitlab-ci.yml
@@ -0,0 +1,24 @@
+stages:
+  - deploy
+
+pages:
+  stage: deploy
+  image: python:3.11-slim
+  before_script:
+    - echo -e "\\e[0Ksection_start:`date +%s`:my_first_section[collapsed=true]\\r\\e[0KSet up docs environment"
+    - apt-get update && apt-get install make --no-install-recommends -y
+    - python -m pip install --upgrade pip
+    - python -m pip install -r docs/requirements.txt
+    - echo -e "\\e[0Ksection_end:`date +%s`:my_first_section\\r\\e[0K"
+  script:
+    - cd docs && make clean
+    - python ../build_docs.py --no-color
+  after_script:
+    - mv docs/_build/html/ ./public/
+  artifacts:
+    paths:
+    - public
+  rules:
+    - if: $CI_COMMIT_BRANCH == "master"
+  tags:
+    - pages
diff --git a/build_docs.py b/build_docs.py
index 4bcdd57de..2f492dd39 100644
--- a/build_docs.py
+++ b/build_docs.py
@@ -1,38 +1,47 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 import os
-import sys
 import subprocess
+import sys
 
-import warp as wp
-
-
-wp.init()
+from warp.context import export_functions_rst, export_stubs
 
 # docs
 
 # disable sphinx color output
 os.environ["NO_COLOR"] = "1"
 
-with open("docs/modules/functions.rst", "w") as function_ref:
-    wp.print_builtins(function_ref)
+base_path = os.path.dirname(os.path.realpath(__file__))
+
+with open(os.path.join(base_path, "docs", "modules", "functions.rst"), "w") as function_ref:
+    export_functions_rst(function_ref)
 
 # run Sphinx build
 try:
-    if os.name == 'nt':
-        subprocess.check_output("make.bat html", cwd="docs", shell=True)
+    docs_folder = os.path.join(base_path, "docs")
+    make_html_cmd = ["make.bat" if os.name == "nt" else "make", "html"]
+
+    if os.name == "nt" or len(sys.argv) == 1:
+        subprocess.check_output(make_html_cmd, cwd=docs_folder)
     else:
-        subprocess.run("make clean", cwd="docs", shell=True)
-        subprocess.check_output("make html", cwd="docs", shell=True)
+        # Sphinx options were passed via the argument list
+        make_html_cmd.append("-e")
+        sphinx_options = " ".join(sys.argv[1:])
+        subprocess.check_output(make_html_cmd, cwd=docs_folder, env=dict(os.environ, SPHINXOPTS=sphinx_options))
 except subprocess.CalledProcessError as e:
     print(e.output.decode())
     raise e
 
-
 # generate stubs for autocomplete
-stub_file = open("warp/stubs.py", "w")
-wp.export_stubs(stub_file)
-stub_file.close()
+with open(os.path.join(base_path, "warp", "stubs.py"), "w") as stub_file:
+    export_stubs(stub_file)
 
 # code formatting
-subprocess.run([sys.executable, "-m", "black", "warp/stubs.py"])
+subprocess.run([sys.executable, "-m", "black", os.path.join(base_path, "warp", "stubs.py")])
 
 print("Finished")
diff --git a/build_lib.py b/build_lib.py
index c78cc6698..ce8f7c1cb 100644
--- a/build_lib.py
+++ b/build_lib.py
@@ -1,3 +1,10 @@
+# Copyright (c) 2022 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
 # This script is an 'offline' build of the core warp runtime libraries
 # designed to be executed as part of CI / developer workflows, not
 # as part of the user runtime (since it requires CUDA toolkit, etc)
diff --git a/docs/_build/html/.buildinfo b/docs/_build/html/.buildinfo
deleted file mode 100644
index 7e3e1db1b..000000000
--- a/docs/_build/html/.buildinfo
+++ /dev/null
@@ -1,4 +0,0 @@
-# Sphinx build info version 1
-# This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
-config: 3801b32913435da995d554ffb86411be
-tags: 645f666f9bcd5a90fca523b33c5a78b7
diff --git a/docs/_build/html/_images/compiler_pipeline.png b/docs/_build/html/_images/compiler_pipeline.png
deleted file mode 100644
index f3475577b971e46a42ee1bf622def4b3a05048d5..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 36891
zcmeFZbx>Sg)-Mbp!QCxDaDsK?4#5c$0)*hCfrb!h+%*aAAq2NT@ZjzQcL*+xySwW>
z<e8awroOknnyI??zd0#5^y$6V*-Q6Yzr89!sw#3==;Y`K2nbjT^3rMu2#*F45D*<v
zk%1>9y(oeR2spppG_)MlAkH*4cGjlmmQWf8R~sk|)WzHs0l{TDGv3^Zz6Q(p{+Lkg
zkpiw>c<ZU?{IAY)6jKHE+%fJLxh>6@JgypqPk{)lXWn<7GffvS!S1o;qce7A9+@#Z
zJM}qtms{s%2b*{GhQ}8dZs#M{g&W(@mFIe^m*3ac$>-+0MGu^?lHXoIz6%*rhYnJg
zpXeFia{Vr&;;f`~Z!kcrn2^9v#c8qlcxTzWuG!nk9(t>1I4BNTXey)5^)fiKZX!vq
z`sPaC!Tv%-T*d$c_nK2hh!bDjr<mDm%k$|)d`2TO9nF0E=IzgtbYn&3`MDDo*8QSJ
zlb?-}i;bqN?Gx-r3|*JESG(678?iVmo_80vYZ~*AqfO8LI}z3^e9vp|fxPv-c>R-a
zXpUa6Em+7o(K%LDaqdrJ^C^-$R;h2-vloTm-5&X^W>b3%{+6WIXUZs+#9`WTPllAX
z5mtrxk$?x7&-<S<R0^~o#qJ0vX=QYOeff1=+5PK9V~kBU%&S&d7rI7N&5MkTA#${A
zeLjLRP@;y~Y^1Q&8Z}SKKFQX~WhPz|7V3M9o;Qck<mr7|T8~bq5mFj0PglH5X7v7|
z{jnVZ`%I1=^@3li0TM24mL{>((+r96oZv4$oZuK-R1=05$(eErwqHJdCnpIBRiKYb
zB!_*VVp*Ic=RK-i_UsB($}wSC$x>j_W?{Kt$^D#Gn6B`qjJROB+^n0$VxY!^uH7Jm
zKCoa#?InA`$#dQ}1vN1`7C#%>8NXyz&b<6tP%#%@#%;I0Ie^s=Az+e`X0`HK_9^TH
z&BHW7=$Nf+fbTk3k@%+M==smW$}4ON=G}6~b$zE<*g4W@cy&~GibqdwcJf_H#akWk
z>$?}@Evy!Yrae|{``XL1YIFXmvP>TY=^#Uy5*;C#2oiF=@kP%v-^Z6FT_WEG$r9(r
zlo7ERn}Bw4PkpOqYA#Ghxy~bw)?QwHtGc*19njwLimRDFKGPqM>Hb|$=AQTa^OoP%
zatame{qMZ_qP<ttlW&X|xo~moPNA{&6Jq8CLIdgV$zMRV1w)6w3d7n{dM1lENgWm0
zlm~(fp6FZjiSh<ib{omeZlCQI3&o%IiMo++3FB;XnM8ep)bqD&xddzRwm%A4v*XFo
z53qZsE^b5(b@Juk;P@&1`K=|{<KpUaMOs_qQV%=oa2t#jN4~_DKX6SrGepdk%St==
z@>HC(g@!45;bQtD@?sV~lXW}G?3j6woUiPYZaa}xCvLvCsjW(7=kcGF_=Y@}Y&CU%
zSC9wr6XMHi^WxUG`|~g-ejMIm7(1+bdtSx=0s2c|I$^~WF6JiW-u>R}2gvN{OCzhd
z+uap~Ll7xJx?TN&xQr|%r$#rH6`O4P4rKW7Ewxp^jVmpdjXU}q0kguWgc3J~^#G=B
z<mf4kkI`Rr^?G3VMsHs_n3ld-pBk`<KH7N94mQ{Owm&JYP4arC^vixJ#D~Z^b_wNR
zhsu9MCv3=V;}<=DRTg5OMD?p}pAXOGiK+wVu{wsvwd%iP_r1*_i`;a_=uBeu-^MhY
zXdX*%@sD#0eU6=jgm$2~`x}QW+VVSnoBsS~=ou%0_WKDN6^6t_8w0n=Hpsq<?eGA-
zDaO7$UIpR~a!OEf^gH4&Unz*+ec=_$?C56@E^NC<Mh}xv%+=%**F;Ey<LnqWyqU3S
zyho5rE&Sx#J8f6a?^Gz5$@?}jaVuao2Z<+>6Go?tsWE9ZLHG`0Gy1M03JoJNeT}u`
z)e|L2w$1Wy&n=%)28a55zSEr>gTLpJjxC}0<D=0Nu;&za*^}5>u`EL5Ok(l;j`{QJ
z6!%O~^TZd`&cZSTny650H20J569^Pl6&M#j*s>d}zbisX83o@nB#F*6CbC#ix->96
zl@DMV#-++i>&PiU_J!s*BcJ&n!X?F!KXbht@|B?6m{s0>nuFcNBs;9^pKbZGjQiVA
zOTflS7Sa()xKgJU1yoJT<z+``i<&11_ZLE}j|^8fhnszR1V|2$oARk2oLm?hu{~KX
zzaIB`8Tq~!oW!bV9Sz^G-wb*3s8>DgLY|@byT{6R-(zOVM*Miez&C2ac`5qeaGV^J
znV%4HiuHUtnYE}w-YzOCGxk_?o)iq-(~=emlSpw@t0YpjLmG!(M{Wk5FG<PM@owd;
zRU9G(OX~0kWe+`Ji?r0&;9tnNQmAYG?m_9az$QOT>spQYlAHqEmqJPCT<iw=g{!Z(
zr+)o;T2g7$iH$;e&^Mt^6<I0dg?9ZbZr)zBCp|b4_NBuGk8u&{Sw_jzQoH=}yRDE{
z4BZEryAB2YLql@yF^UPgpb-uq#vxwmua55on4J_W1mzEk-=KG7oac7pkBFqHlP1Sx
zoiLh1Hy)wCw%~O8fUJD*=#+;NEjgm~%?{tXQgN?GH~6NZU)5zWil%fHweor(y=R~Y
z(<-N;_)*kjnu2eRtua-qTC%gm+>c0<q`AJKw!d&?{MCV(jE<-S{Z*SqPMG#-zS*#K
zEw)G|gi)e#0~%B7oG!GUU~bD<!MS!5tW}qiJ^dJ4fl>C$_V;wcRk>f|nQ*e9$Uq`b
z!vdW$m{?0~)r6hGntt(SLV@<{b0+yFN$HQn&~n3{91@J=ErpLYC_8*DY}~sc%t`|C
zxGG&Mt;}to6Rs!nESQ#bzW;cZ5~^F*RJQP1-mm)UkM6R~me8jb@p7%fpBY~UGp(hm
z$`dRa#S9(wq2MS@269DO)+sqLh8GxXSBljYC_26xZg?w^N#Gx+YnzolIMazxIr$+;
zH`BDZ#f>%tEnenri94K<_mQj^*sPjah7$DRoeKipXY|?6bRx}L7N!ve{esHYJ_mf$
zV9Ld+U(XIp(no`GkqK|tx%fPNLTz;7py-kK(PPiN9TLckMzB>G2NJ>swWc?3<9emV
zzBlHZm#DXxRCrD9sy~T&1=bTNrdiZ6#I1;R3?aLu;vVuUf4_t*f+nT|jyIw?Uy$sx
zbM3R%klJA`$8DpN9iQG@P}Ojn3r|>;UkkUfaMDm6dRe*K&mG@{>2&u6oiEGS?~1%|
zxyAf8lCz6{JgN(~#`t|iG;@yGb5tZw)Tx{fOjS`449~};jEi5<3~e-UpcVlf3XdQJ
z1k_;rOkDldyH*vtq?ZwqlYW9z+$kSk#L-@TRUs{8vj^+0yM2BeM*R6r@vT8gPc}ic
zWx;RFjn3y{OJhpf@Hx}paU8;1d#wmn0`Iu!Ser#rwI9RD6eaqd6{%I-BPo@e2QoR_
z_}<8QqYkB)NvP3I`uQH04v>95@oBgyr)zkQE_#3yr?d0HOv(XKM8TvUG9I5>r+49p
z$ny1llea07N8fj*3~L$cK4a<pXNU$HI_4AIWMeu3p|^amL@jaOqvsFs5%;ekrx+PU
zFr<B9Ciop(yO<$Anw8tENz0wi5Qkv&lJ~W@YiESm6T_MOrb?={dV8%wMlkbW$C&0)
z(?|V~nm{JG-@~ZF^uc8;ag2peVUVKe4VmpND^N_b!a%_Vst}Ax0ypXmCaomVW4#g~
zTAqx^5q!yp(M2=qNpDH$<#(pj_594Xy(9}WiEsP~d7GGMeF}%jXWnby*EX*Xo|a<L
z=k@Z~`*SQ8)@22iw(@f4x#)>{VRDl<1jg@tJnV(<Sn!mo>&Q;#k_X2;`<dxy${6q@
zMwSDidOR6aw-Q}}oD*oL(}4G+{j)lW`2~U|(!!%Jb=4}P&PcjHxGAtcIU27q{3g_R
z>X^mnoX}7R3!sG8u6Z@+N@3wsCR}|W>mC3H;D5UQ0C(m%l{G39+00wit3-G^>_1ee
z{*-k3gQpKMDat(WV<CxaG*URGE8z+K9_tXCtFtdS67@?m^?BS61kpTc#r<zct6td|
zXdRsbht)Ul45(R&R9idtuG%Wzb+O(NIPj1*rBf>=1SpsO-jovk_NXD9bRH7%grC~F
z+9eR({ArFBUi^x54r{N<CsZ}F>KOR~U##Wjn}#XwLs<mgkCx!vHVsB*0+T3Y^0zlx
z+WShVxIDrZbI4(necsG5pNZ)Q8umBEyM)@hR6Al&lAj~?A8iPQG4#KZQL-Q-{rUO}
zd94C5%8N|O->ZGT%OWn$IWM#17e};H?0Awle`skx)|kj4$5!S`YWfksvxr8(l7-3m
zSU?-kA?~$n-1coKvVqH`2+ZQOzy^AgiRX65Nv_aLyi@%XhSEj4bQ!rmcKwa9{gCM~
zYoE6?iacyooh_vr0czk$D>jnvDgACrm^T?E_F9L$2qcgvFO)u0XG(*5PW(rVJxH(?
z5rs!|1C%j*g`Os_;iQB0^ql&z#<=bhQ%8IgmTWCS8`$R|B{i@W=Bl64_&dHd{9vvh
zy6(!!!S1?`Ki9SvON#eun5KXJRG2&ULW&Qg2K`9B%C`t<Zb8$xCwJ}ZAxgNZLP~P~
zQDfOng_NZA0I%sp>=d}m(X^p+v>54it$QskMq||wMD|%#M5mVn!!#Pig+K}6NjI69
zBx+z6yd-J+w%9U$?c`ml>Fm?;ALHxJ!rytcwY>Hs0vlYQT6D`KsQO^OMgx>nwEQqr
zZf*AgRMpG!apzg_s0BRi1cbeWG<?%mE<)kzNZv{dt((V|?}|U1ZBr301lxV(oYZXE
z(@f*J(S@UiOs5X*cwI*gZa=9nR>@^7rQ*mP9}al#kfV8ug=@7!kRSkmpCfMTi@f;^
zU%9$$`KgX+hMz;RBir)iTcg64-(RfUEjTq&BQM4a)M}cS4UY|6Q_dDm2<)ujkFjd4
z&QrBDe%haV_x1|wn7lktX(2f6m1o*1(nc5uf)Do*Ob)?q6sh7fS$-pZpEV)9hw2Fd
zzv5eKBC8Om!8~?fGUFirGZ(>~O`fp+h=#xjY5r~78rOGcPDrmV0#7O(&vzQw^KLt~
zaDRpwcVDyC7I`$H?8Li0dqodsVXdW*DI#f;F_b@LOX5xb0wowra==99@wE%<99s0)
z5m0{71!b&flI>l4U51T?`^y)Hyk%a`TPj|C-hWE5_{%HMwHL1Uch12y4J5ts{5Qgb
z_)ymzaCpM{L^w_p{B|!J=VsFjsII&k(=$|iQKN{{m}vvlWXq79t~lE7xb*G1)}-pV
zxLBWOy--CT?OM)dcR~6Q|JqDAB}=&3{ApgJc9rZ1`j06=MOkwEf{b>9-3S!X1<=k7
zj8^jMAob~2pQvmK39Y_<IQ4W`S7X!s?a3$c*?lT6KP_s1&kNqh`TK1v%Q82CxBK(B
zU|Ft7b-f~jSJ~pi<hRRX@y>1I7ifdQ-zUUQ$+<g7is-4nc8s;GmFov`fSWpvK+Iv+
zA7{*Y3la|tOHI|$AFuDeT5=bS3cWAU+Z*#dM)a#r0E%DFxu}VhOfjI)_cfQ2QdN+W
z`g?5;RP;%nk)rag5|r(RBg8T~lx*osU($rMafxh4pA)7@<`U^Rf@)4M5_srYIKn?$
zG&e`I=lO(`6=D~Drnx{|+FHukpk=t@WGN3kT&Skr-xJ-NLzMi{qQMQq&Pjk~5ZHnf
zb&2W>@FYlYQH@lMvczIvGAs%<?kz5F>xY`Tre53oM%Q;Eq{@r*I<(Tx(WK0kI}BX6
z;c0#)T(VyR`*6tLbU99~@(;)aJV_w~b8WgVp_XXvysPoi?p6)T4*U_E_qy?mqxNEb
zyLQ2-)^^H`Us`=-X`f?4tNc4WMfn%Lq2xU|za&F7!w3n?rz6}i0w}$M^|)$!-mPd!
z5Iv{w#JNLhs^i6oBCt|p#(r#Qd2Rf7WBN&gBm7RRa(4}F!5g`-ys~K;!BkA4sH$QE
z?{^)Ro?aEND^B$cB65mD_ZQ0USL)1!AJ4h3vM7H$W1QsIG*q0~ULly|{9M7Zu|W+b
zW^tcgo}Eo@xOLAOo5!KzwezD;O^*aH7;<v}Ytd3x5;C@iaX?J0ji4MZFdG1iK|l}@
zcd>yOTR<IXjG$)bR-&}~bxpK1<|d-FZ+VqL$~IEa_vZ3$c2IRU6%AuI3u8eOT5&OS
z5f>qV01WB?p>cs(TG<P^h|>O{D+K)g@R*a9=FboZ3sG7vWmOs}Yda_n4+jqih+W3T
z+=+`;44p>A&cswmO<MMEB!Ew%wC^1pY=k&Dot>RIoVhuy?aVmAf`WpaATCZWE_Pr9
zyS=NG1H^^h%AW3l#9uU|q4vgh<~9!I)>bqRG$BUTjt-)<w7~Z?e_tQWMp^kE<gM)g
zCIvtqoGuU>PA~_E69(h_=Na}6GEM-Izb)v0onfy5KvtY;P<v}fJ7cJf6V%Fq?w?7R
z82@9wjia6AA9hTPIiZ$N7%<cxcq{l{UMZ)bton}`4<ax#huQp@1ql0JMmm_A{*SQ!
zWo-}N{9)&x7XnQG2i<=e{deB~3<gFiD+@_m8#_KMPeEFg_Tl|PCf3I0CPIJyGzEjX
zxdn^`*#!i7P1$*jO~C8|AT9xRL0&Mgu^=z6DU_f4pGYZK**ie2jG+&t0OB0x03AU=
z9ure8K5ljp#F(F*habetF2KhPVdpUdL-~!21wde9qkkgt%FZ0HN{Hn@uj+x62|$XQ
zn}>^sM*sr69%RDKV<N!CZe(O?!VWbO1RH_4c_Dlt=pRxh#zL~zb}$H_PIDN<49aO^
zW%lQb2Z0N{P*o76<>CPS&l6Qkh=VCGL6la>+{)4Af4<N#he6dHAP>?6^9g`?xVQyD
zg4`fMF3?|W|DmS|wX+8-@qs27#KHX+-3PS@0n7k`g*@0PK;REqfEOVtJ1E4#+D^mT
z+ESGEfdtLN%s=K$Bl4G0<jw7Y5v~sw|2LaghuZ%2=`R<sH2?Dz4b3036@nQ5^%8rC
z6V&9-i-385eP#R}Vr2#e{QGYP_4nuI|BJGWjZL`tcp-qQK}LM+JP<BG^H6R;!u$dz
z`~dG@9`3)C^$+Uy)}{{55Ig7#Gk{Zo8^D16a6`lVN0co8x)$g6&<8nzxBzqDVFz<-
zfO&<$U?Cnsb`ZZ12t>>IH-<SMy!!7Pi*WvbWFqosz(0inVBTNPfbarDE6#s}tG~(Y
zL5%+oufOTy|KS?|(*J$Qe<a`kitE4P`i~^=9}WJmb^TXd|B(d#qrv~RuK#P|LjRwF
z2WkamLC!!qv!?t;5h#U_jTGgi5$+%UeQ(H)0-m7R$ZOjpAYc$Y{2(GEB#{CSQ5+PM
zWl+{1V-aJ3KiplGBOuTqC`iB1aGBnjb#|Z}YwlQ_@8dHcOB{nL#z1S9gTcs1{s>u8
z_%!ssxGk}`lK7~fB;J3L_^7JJ{0!HvC<~4qw`4Wr9u>l-t+@|3YZTXvVJCT6T;eyk
zPWe=!j=QYPd2gW6etqC+YbA^VFOn)JE$ve<3=YPW@*6^$s#Ml_$u%{l{TvhTayaSv
zBfQPr=RYgeDT3(4z<9iLCnAwHaUJZ5gefx8i`vpFy~VJ=S4sfrp|Dtkf*#PyGrF1Y
z6UZzQ66leM(6B7$Hdjf4cbTU|^eOP+YAO47Zx1AZ>kp+wGUPZz+ag$k@H$v|!QLvz
zOenqRFms~wW^0>{HvNE&mUdp=X0EZ|Yd5E;;;;Sd;p2SWhTwiE#d~Xu)i9Ejth7Sn
z*6+t=0rK>HUoVf}=lr}@L!9P;EGB99CEd*z5mw7cKXA12Yd>E(Cy#l3ESs8+^kC7!
z39G5D5`jXt{LLh2A<EpE`+6DYy#;JBGy6{26NZJ&l&@bC>TKwZ*IEo#LPfr*Kbu*$
zoGz{`dos<_ToN(<e#F<6TWo3Be1FMi&s_*F(wZVWSVB6Z4iu<`H(sdo393za%E~~;
zO<-6T0V7&nAtGF|tIH{vQ(-1-Sv~|^6oXeth#_!3<RJr!7v8?-b1w1Gio{>vs>(x}
z)X29!#5;_;62@YurxiwiY)Xz<GyPEY_O-}6T=Pw`$up{-QUZiVm+1@_({SdeB81j@
zHz^}Qe`MOlPnE3flNVf2e>3`{yn=6Je7dK?yOE>{-kQv>*PgoP?z>ZOIhT5a{w8{r
z(s_7u#<lfL()~zpoJ|YMQw;EXb91BFyj9Cty^e4EhKuebccBlAPRUU(?0JhyU9NGZ
zcwJZ%o^21F!HBA7sjM2$4e7Hp9MC$l%(>P_aL}IkXneEIa5#Y6tX+vOJJghyNU!KB
z&hkl1*P_8oH>0Ebqxt8WXYbv@=VaPNj*FTIFfYt4%$rN|-}l=6l5@Afpu?u_{Fv5y
zSW)HZlv@l-3^m9N72lKg4;M*dwZ6gdygD;p%HURbsq%~v9Ut>C2-C;V(3^sP{ung&
z2CmiQc1cBoq7x&txd73-k8L`>7N2cxso|c|ERi9>3;tuNl=z2avG6b^Mt|Qnt#F|`
z`bw**z@mnyq53Gpfp<htv5cwfF2ltrW##gm1Spfe;g!mEAIQ;y6zCbVzkG?4*Hu(p
z7pcp~CX3g0Sr*BFnk+a7nOj?2&)|tl|7C9!t2WXB(H7RQc8b`KOpM6z(&AlIY!vC1
zw1yL5zCba;P1I>*P5RKyVzBb0ryacVM{n__9mDtebvk_PeXiL7YCJqYGdw&&6H<SL
z-LvOnF6Xd0uT@?H?>n?`(L)Wpqw58S8kghDbr<U+T@Rkq<Yd>0aOB16i^nroIO>c|
z{zdFcnpR@a?Qzg<L3h4k12R(al>4JfnDF%5+T7@E&U-ZL`is$lLbAFHQ(ZJ2X*Wrv
zj(+}mkFwww(r)Qn2|6u>?|XOCsFR*8oEs-#=ly)HU15pWnc_g6onkL$dw94)Oj!DE
zAe)DukIQbyC0G$<sLF&!EH&bDi&5|uQB-_#XNRK&Dab^?ubtsB5~5h*EelKTCgB+?
z10gENW>VhB$S6=z{=rg)CQdIm#9c@gx>*4CPD)==L%)t*IKS-f>PfR!V2Obep!<<}
zOfoMm5!AMyRk2}sFFT<hm7YRyBRQMIE5|fwXhlblqJJK`mMJK`+~S+XPG|4oi+O^b
z{Rt*R;jK)-*1K1I88kE%ZM5$9SvVV!?y4G5vu$r@+mck6;_LY-|73FSyyKL+it`2r
z5bCV-dM;Y!Hd<+D4HNwH2=F##(-(Qjy$ZSpI@wvTF7|FCoNwJbd6FaG!knZtp3|PU
z4a>Y|9(uPaG8ty7sesym>p^?}-tDyu@x#b}|MoyTvZx<~T`+NPINr(AG<OZ^FKWX4
z^U_GF@0m4Zy`|Va7J%96TCi3&9G}8<ben8Eov><Sh#;p&P##`Buhj#K1YGPVQ>C_a
z+nNO#xoL|FloAfU&VLb+`Ua73*TuN6Js-3zV(SE*dw5NNrdPd<_XqdI?n%%tJa4Y8
zz0>lHShLW)422_ZmpK3n5n7+Wb+m+=u6}JFwVd}im;6P&U|n?|Io3<OY~yh`Zam^!
z>S$4O*pq61OGIMdyU(#)+v_bUbD72Q!s&cse!^IPuIlh|WcR8XBGDex4+LS4)S@zR
zo%O%Mc;kS|u<x5*;^V91SAgw>4*ViLxJC>3T+YKkH2GTsgiAwsGBP**(i6rW16@47
z%qb|$YG}Y}{b<ANTh!2CKNUli)=+>2uRMP`r~KaP!$Rx%oXJ%Br9e_KY@0Ys$;+>Y
zJrUf$VMo5bZJzV9wUUlh)*<EE`kqcvEHPw$9mVqVCxk_tFaEW%)9~p=kRCZQ&*+$K
z=V?>>uATk%?fu%GT*_)d>|JnMn~9~cj#%@@)P&tP<y94qF%jmnx}6q3s|e7a2#t@>
z5@G_v&n?JhTsVio@wT=<m|Ixv>ZLZYkCa*(cx}A~9KCx$^IR1q58B9XC<vQQ_@JYk
zP**;1Jq<?-kjiVLEj1K;^O%<QnmFiJr}})F$pWsa=ZT?;(6W2=j8%%~Nvg)%H^Y;v
za@`-Bfe2ONa%X7?r{xffR38wGV4|vz?YI&$GHSTF>DfMP>o}S*#yQv)zX=*W(OB~k
zii*gp?<exECaUTzJ3p}cb=HGk`Q{t<dK8YpWuLuPtqqkfzqQGTe?=%Ie~KB^B?`BA
zf-F6yY2Sr$^~KRV*t$b5U#B*3{r;FHdNon-_$DM!nzsI`&)em4v$L2zq_?o~({QuY
z)Z2ZW5yWFt4{2l1r`~thKTJ%`kiN__#mbXfQj8bsH811LNJ)9JQ@z5Rzdy{Rblo}Z
zyV^i9aJhldMuNI?H}$R(ErLQ`Eu^{MDJdzjZEJEeafp3tS-Jz_2qi22V(I*B*8*Ov
zcAXtV#QiZ9AZ8nPBZ2+&dpf2j{xY!y|M>|Odi+JRrbhouZgJCrY-*zf#z^_cAIk>r
z!9T)D<zS9a?=O18ZVrZ7H*|+neMXmMr5nS-o&)()q~{Keat_;k+t#XL^-A#1Db3d^
zQ?+ApjFxl7;cpa44o7;6*4#-9Uhba3Jov5&1_aI}Ick<Zl-eC#>_u=a;#AGQQt`4*
zao<7Y;wsCTDGic2HuFIonfHzYqPUTf=(Qe7>qljZ@!=u|T=?^z+)1^8;U~uemVJ1Z
zT2<um&#fJy%T?a>$*Ff+AMUOPwgHL9y<QH_Dy|k^zd6WGI`v5f_N5@-R>Ek^g=zR6
z12<DdY^vMIu~a73wr0_H=hxuPhSFlT$#Mu%s?zmGEQ^WZ>ERwE=lh_>TNJwyL+Yh2
z<EB=+^ZDjrh7i|-Jzc>MOA0eC=RUx`e|H?wv~*m-?Y-7{Le1kVesk%S_q@m4+-jv(
z{64HaRLQ`5QDItV+V!AN{I%9hYT}+(;18yhHmUO&e<sG5NaId!)0>i;`RIwrn8<D*
zm*`79Ma%X&M|nsaMp#(`Sgae-f0fD{=gidl8N^R{sI)fzNi)0dj7UMtt+^TJJtly;
zb@g*DjNr`eyWn>`br(Sjz-DA%N9(N=usqR{4Sy=9Ba2*OYpY>7*JU3e_G=AmJ11`@
zn4cqsj*)VZX=D!SzUnkyV9-6(nL0Cd?rhCQ!I!V!KKUc!!9MOi&mDnGOyGD1BX_bs
z^5c9!=knt0gu%OMc+t@4qy)^7$o3Ge+E|1>J1a6N0rew%88TXZc*4;GZHuGJDa^ZR
zvBHW0K--gS5T;gx2by=oDtU^>hF1@NbctQ}qlZ@!;LP;`OqyH3po>Xlla=j<*-0;~
zy!bj(?U1)8^1o;12Klzp`-f91Lu|n{Pa4QTE1)aFbuODp8z4qamf9lq#<G34oc)mE
zvSARz7N=-o&%FyPa9Lgs3W;E9cq^o0<}m9YM%mg`BKI{ie!!L~bs&LAF^9S6-Cp(x
znf=OR{3rsV6d98ywJI#Yq#v(WjynO#29{(R4bHs1aI|hO%Y1$0vG)-MX6vu*u+B<|
z9AC(3hvUHgG<O8|_VT^mB`<%;hbf~u?_~x{_*za7(_uHt{_);6=htuFByBfo*euqW
z@$0G{2RwoRj2|YC3^XI$*{=6ff+4hY9aQbYMGe1tGBI2i3Uvh*%k4=;tsW)zjvH%7
zZoHiO2y^aUi+NJ{gip4-sN(7FTNvrnh<$P%EavAdt0VV~a{9_d4W(s3vOM*Wv0m}A
z0_NV<<|cUejk~*V;0cA%^3r5IMBl+_q;1X-lDz%y-PF)T>2CPBrCcgb05G(T%XT>4
zWOeNmmU!(6`q4EZM25{0zIEhf)L={M|3Uz>J=3SFt8Xp&ZECjVEV;_{+!OQx+S2*l
zRc&b#YhgKp7>*=h6};;PC-#8B4Z4b73KsPDv*`=q*q?((oP5XL8$j)+gs!kirBfRj
zcz8#bEv)HCQGaJ!qg^&&xGpuj3*O`yA9l1q0`2bdyKG0V*XbP)Hyjj2UL|}u0-81f
zOGfaIAd@py;jpR-fMTo@99)wDLh^_Fo*8|}TC;nDdRI+nx1vWai|7c^_w;Hw_D4*1
zgI7iF2Cso#Zp#iYjp$Qrgi!cowaZ@JF0v-OuIyAT77!yJ!7IDB_2p$v`JfhWBcmn~
z<0iH0ZEZ10?MP#w+}A&FCJnS(%d6(UoBYPb$Bz=4OPyty*=Lr3mv*slNEPMl0XD4Q
zWPk`Z;Iu_UYwM^Dd?jz9IEDPYN^^L8Yn$!tsgcnvrTtQ_$oZrXJw4;=QR|OaLLpsV
z>(!TQiPoCnK>1f1`s#|4(Lj=H8nKRvBo=>0MW#KaXVcBOF;c}T!;q=*Mgte!@1U*i
zYkZVk_pfFI%PB4Y-d)p+I64rfy!^{WA7>`+?H~6`HbTHA^9`F^*2#;pfW4qE@etf}
zF?$MBClJzup>NJ{XJ-~kE?3UBLi)*wjg_Xt0ME8JKqaBLF9VcQGhn7K2{8j}rKc+^
z=g)v`q{`4Wywve$_eOLWOokM26+J5)>xLV)UgCit_unBJR<kV&i$A)1`ZL{hiTahp
zLWS2JS6a<=`FJ-(+jb^NxD^@bA31t>P$C6{KX{ufpxX<VEn%%iRI?Osyi6(o-*wh#
z`%Q;j>ho^VgQnh+DpuCWIMgb3AB2jU4hVsr03&;(JF}^V*jkU$;w`(`&QNFMm$<fs
zRmBZ`J;`}52Fuyf=y$&c!SnIb5L@rL)R~e#?bI*HZzWP2+vypyx$Uns#1f8(%{Lot
zYV}C{y8~MzLan%I#GXYyFvs0^FXbOj?sqG=I91Aa69#AL>KzYwA>GK+TIr~!DD9TD
z%KCQs*fc8cTlvJub>kTyB{quH;DF6j)Qb{YJj~WMhq&t5p^#f3E2@Mo1_kY6&MSvu
zJ-+c$8yPvQxVEX8eb=h^O04CIx3}hS55-m>DUR{xJei+Gw<=h%usW+(NSYS%$jjU8
z-1ptDh~C`}kkTVBYJYR11mx3-injK40s;ytU0vZIh3=0sQu<G0ppoCczGNnC+dW&b
z5kd=)^p$*8Z);F#IrD4e{z|y%=2jVq#e1)bg*$=@f97@|tEr6~bhrxFC6-rI4v$U-
zPEg26s&V;49P007J$7Lk8Ci=#VbAa@53Z+!f=+?GAIN2~&C2~m2r==)eS^YIo!H94
zvO4@lW2UBum4IYAsGB7&D+sFH<^Z(x<C{*kT#DFCu+KOQN)ev0+|0zR+XyHvcVvc&
zd-#fB{y{K~99zpE#p}x*19Nrm?p70CkA-U`CY?_9<po;k&hP!A6-jKS#`w8xHXvHW
z7oK7d`ZDgImeP?e5{;>wIGx-*b@FDv#-Q5Q;{aJ)@u-b(lY&l)2VuFp^)Esj*6p0~
zLmDE^o2t%BCr(S6mbiqoz)ugl%UMKD4{np}1}>9q_xud2Z-F5H0PL8E`>)^jSHhlG
zSv`C9>{G_KWcK=@*<wF)A?%(fRxw1)i#871A={WpTNV~TmJdY6BPhovP@6>F`^my_
z@h&?arM&RkpG*Fsv~wjuS%2VGEO~t=3i>p9;;=!=D=qy8J<^r<Ax=4o5H^q*+=ZBx
zwgo)?t1$jIZf?kM*{a~J5(+~Zpa6^Sa3zkUV&&`T&x6|A)aZ$Bc~+HI;s;5~tGwJA
zf{Rm@F!w1;pa4dTKbxW}?PleI9)R}vk*GzsHL2xnEm70_D_dKz=q>B7Rl7x}x<56e
zGW#Iyu)Xs~6($w4K{%FMgMlTyWBW7@KQPckX6u-ZcDR+)|C5SP@E=6w4`T3uZ#_U(
zf6%Lk4-Y?-rNUk#ZKZ5PzJGo8h<`w=&<AO0kDh#PreW30f7kr^iH4?Tp2T0=n-~Jc
z(dE(V&ARw~^ZIcv@VJq7mR?3W1Le>~gwlD&mY#_zDlzfrBRsszU<Y?A0L5d(MO`?)
zlvh#Fg#5Jx06e24eOM)M9Wmq%mU*zkW*VdO){rq<RUs!5m9VJzFOsh{=ftwn{G{?H
zb^OIC)jplhBuL)(JyOH}szLZ*m!k0c0t>Kli%SwVkt_}&YQU7*>*JcmDUyTZ7wvXe
zW*P5t&vhb<$&nKi$b;<LANU4XZe64ieioyZo#*mu1tVaJ7}IOf!Pbi36)gyTP?SgH
zTJqWNDVB#y7I9}I*WE9js`j#J$Em7WO?0y)KU|20_9z;=@u6w~t_L~&d$&J8Anjjx
z;J<bIJMC*9bs3tYQ_`$7Nr5kx#W3AdiPP`8qeu@cj2y(nd~E*V{p%N$IrSz!OdT?6
zFVnYPyedMWB2mZ<0$@^L&}Jd#$UFEr`Hm`aw>jAdB4OOL?$W7xFBT;K_655-oGCZe
zH?I}xCqqWd4Ycx5P-sNpK=%I)@i4$j6!)P_zihE~2=rwywLb#(iPlOD`fG8X7s{t@
zY}>y%X5<AVO`XK1TrYu{IaH&X)-?DfHn{>1E|ihzU84I=o0;e7_ur%X<aRK3GyN|}
zZ*WHQIB8K*9~|*NMhqTi{9gbi0N?@P$1;l3-5ys^$@+cWz_HB8zR+M})_uL}8E<N!
zyz{yS9$=z6K!t{koYji>kqFHg2|IwG^|80RaKCbYTU!S%Nviy>{nk~dWrtd!*hLHL
z&PAK0v+9x(bIbSiKxIslDv{4$*t_l5`t0*3nz`zYWdLPJO<l`kq*OrxP@*_f2YpO}
zmNZ;tEGz^dtys3J6JED~nP(PM_yHNr(lP;9sbc7lQ##Q3BUTpJZE#RxUmV;z+X_$u
zpeGB0hsdcXNQqG;p)I&w@$!8BPEuV;v}8`e%e{41$?MP@{fh8#mRdg)CZ%pn=iRXR
zVW4rr2pOAVfIH$Qij2AuhxDYUXR(K3pN5+W1s>cu_l=)vWM;IhM!fNHZ)v>XEs}x9
zk|BVz;9~m;?5&Jk9&&nxjF^`?9-H3XRtZm)dtv#TgWP5wbKLe*+Xy^W^4!KUaJ_mB
zgbk~R07dU(6xlFgeAe{c=MT1Q7s#!Q8qj*-i6{c5N>X8pw0Vq)8qo4lS;-PJ=1B|`
zkq@X@S@EAS@*#r=YjGaIAFA1`lE&~K5#nT^xU+$6APE95`YMQA9^9IknA)#mIjidG
z%V?@lZ8mVZdLQ*F30;dU;n(Num42-F855cO_SbQQXq)%EYcgwLU37kuM*{hlyu6Zs
zIBfeIC0@|;dDMUyP5^<}wF9H`tOKuEiu*IUR8Uau)^2%K6(j-3zu0FyN5%kM9|geo
zJ3*K&iy)QQD8W>ZF7f?(KkEA>{!mKc6#pwByJL4_-RZ19-W?#&`f)Xc<Y+kRJ}WmZ
zyxbuEW@lo8%X>}EgU=lq0B=gt!U<kob7Xs*dzh-E_Ka2;wC_5p>sN#U`PMH%@p{rK
z_>Yzw|0j=L@WHft0sJv1ciMks)u{VMaOISxW#1yZATz%Hv)Ib{rt*IJoECAY_?dA1
z;Z4*7?zs4)69gno{^0_Y0j<sOBC)~truwfx7^&pwG5v=k<mt(Wl?nI?41MO;jSnDn
zXvl1U#=sqvf}cV-u6g~<)wsTUo8=gS%or&qidIu!HN8x2^ekaZM=@?jEfPH!2#65(
zx97M3zf;xH3x#gx2R!U1^8Ba+@Txa<6Ie~gKn7~l%+QokS&Lr8A0D<nuKKntxwoh$
z2vEuKr1)OC^Sqa3(M-=;siU4}@9gLs>-ALUIi;c<%0~MwGU?hy7UvNj?P?c1S#9<X
zu|hqI^9uDus*bKfF%$e0h=jc{(k={^k!4nFIb`OtycDPycxGV*GhgdF_UaSAv!e?;
zbxd*}x__-F^yLA>RYc*kzv8(%+3f@>)S451@9%E942Vq;F9O17@fpdToZTC0xW&>P
z4phIYeSM!}@qh>E9c|Aq+r&M-fs7X#N`Xi-05D=(lbwmF>0|!Pqn5?w-uEyV#D2;x
zyl%%qX=KAgMS2A;|JE+h11>g-tyTMugO5+V{)Fo#3XbWGZD&+cOeho%oP(Gg96*xJ
zW{;WK_6C<O%Mh8FcY2E>e+b-tQzzrydoPxFs!#>vE&p^;Ex^UaCI9Nx{4x_Aolp76
z2?xpKQ~bp#-`!{;7ZEDPWKWJ`Qx&tl&`qTz2V}IMZp+#F$e`?`j@^5Yq<5fS6;1Vl
ze)eXIL5?<kLD{7?drGF>H;7rA@45nqd?jBfjkImr?egYSRhE}}44F}SjoKn;MWO=k
z7u$=43}n73_k&V5YfJkjD`iuYm0`WB9W%SMK)qDVj!`9N_WVzJRNEUc6>4w;MG!L-
z2B0mDDzicL^~x~lCy2Dv8v(*Cm&rZLcn3$3Iaq5Mbf~C0?r^zt|9Cx+E}Ta28QwLn
zTa-VdE?*%Nn8go`DXHaZjP=6uEL-VffF&dAJRJZ7D6vUJZa`L7+enyVmV`icLB#s#
zM+cSP{KNnNICKJ406hb5JOI)l@L6k76i^^1Hz^a0=W7LzY-Uv7OXQ;h_<=(2Mte_Y
z$S?nDUl)$1HAXk6DFNz5_rALH8yP`mtctks_mlkn;g@dEQq%U~fpdtH2-9CqOxfy7
zq)2?&!~N>_9MyM|U+)0eE$+Bm4#>ev&0F1c5~hJ88{kAsM|WDgc7gi=1_6$H5Z_zF
zGBa!99e+*~16U<8HD8`?0Cg(}+~05yWMi8GkLQ!HD8>sIg_}NpZfP`GCk9Spq;0zq
zfEu~f>lD3b-0IzOE!WnE5AR`CEd&HUFezD0J}LhQ)7ymT&{P-NN-ROwCl5iVle#*B
zm-X$4>f6qf^Z87msz(5H{7g#jX*fRzG2A~+3x_H7(sRyrRWPgJmN*RS6ez6I3o4$#
zOUWGDh6IvS0I<bc!x}zNd1MEiq>-n+`|=Q#_8SiIxy5ciW_8FC4<z^3lcm0|2Qq;1
z*QsWa)~&CL5kuAo7Yp8|tO6LxkoRxUK;6c=$W@iF>*}T)7}E!EbTT0B%#e-L98QS;
z_yE@o7v6M`rMBe(bZxdeEOWED)eFS4j*n?{aV|Y%?kADaNUW5V=GHLCLNQ+TjLl4U
z_6UGWDdk7q`S>3!x>LBGpORd<^a#5c8Otm0w@-;z3+SLI#T_K8Pc{~O{)*^R!e&!V
zsdohzx4DOTuUz$df94mXitX)RotROwP^7?e<W&;dzD_nHwZwP;mPf}ljt?5D7Q+;2
zD6wBp7JW<qifCd2wOI{wK3LHAUg}9r3E2d(FteGSRHG$%jGyPu)W*A8H-rNi7VRR)
z%<zB!rU6I<Y^ggT*;_Tp8Y0HXIwli({Oc(-xqq|&m~K~(NK)TSy^j$6+w%Jl80^Yn
zCTlB>dh}+m-)Jk?8;j;{HUMbMlC^i^r?WG@f$Q!;pr~9~SzivaJ=|%cUMrj8zE~aO
z|BMJ=Wwv#qC=GCbSnKDSeYHt(s#i*Vu_=Sp*y?Ynm+D%XcI*?@Dv48c5AlqTN{bsi
zD0r28O2>3J#w;+WYAy}8KLW?-0Q?6aGkDvlOXt_cw8V4O0*}1J+4~=I7_~RA!(mcZ
zFA(Q0E(U5zFp=L|S<*DT9p~6aK}?M(NaV-EMD8BwWfEkFwqaIMS4Zk*G1tEW$-N-#
z^Zt7;K<ASx0ANwjhw%bHqRIN2bEQopBhUDRdQAmoEL_>(1+|fpF(QC!09b=05WdaH
z;(@fI1Jrk`hC@8y9x-=YpImf$iX(tE&O6EtYUGr1BeBS40gkP7Qe+yRSW+wz-~+q%
zBtw0?$@*$!RO|$qn^J(!h415FMa5!ThqC}aWfXuZ^y|}B5GEGb={NRg%waArp{iUa
zM&cjK)>TaYCk8`EfnG?+PAX;FFB0Lq&*->+_Ewg$+jO?6=?Ni(ZroR5VP{8`pM;85
z+Mjqw<=2ap9v}b8QuHillYgneAjr;pmDF%7IO3rCI~zB&MRZ*GPY}eeL1>V}Efw1L
z`vh$^pRYrJR|feyqe<bBiXYPZhf#&VwmX;uTAXGs2M2bmw(V1^!p^Sr<hpLvM)orI
zb}9;YjZvi>Pygh!m*c*rElXbRkCQh1pqHewjh`ccDha4pc8ZH^>Z>#@fYf&?>ftC4
zxPY|&-@N@TU5>mD<0eEeaq=Wf3$$WCQ??^jRnznBTz(A{WPx3=#Hf$8>Ka1>t*lPN
z>;#ow28Y;O-Fo$G!@GBdQF}E%-Z92yDL|*lk37Zxq~Dh>J;bP^`yqD0SH{{sR`c>s
z+)6$C`QLx-N6e+bfkZgS_Q%nERtvX)KF2J!Hc&A)a>2vbP$T4~b`$%oumiZSs)Cnk
z-gx|Jy<nDeXrd{g{a@+<K~kX0BeJLy^n(OH=YHzW4E~PPA*6*jEWGx6zmvokqQ4mA
zAz%;pt8;&BKWew1EPsZkd+2%}z=)xvlgI4vjdIJP=ZL|6QlnDmbU|XwXYDz<;!6$E
zKa)J0e#1(S+#qppq-lADhs$@?RTCmdFk*Wh11EeiK^i#NCZh6#n|{5!YOfzuA@95v
z7Lu8hA^*gd##o)76jyZor`vZonuqi@ozw9(P%cGsIKck&hWsh6|570E`~UaTpL+Zs
zb@9J`a-dTo`qXtl0K)vG`GRs%t)mh)hBW3Q)kyV+rpJqSAM&l!O?A1kvoTKE8|SC^
z1oVk7Wu~4Yp8n29QO4@mc*~bdMfuBZ!A6^q&=e9kd+uxQ=*U}vrA#`VfwJTih<GwE
z$iT;^x%v#?B4BNRe9Iz1UV%&!07lsq)64-$CK4%T6JR0()qQ}aXwFY_T0vLb_$gPQ
z;H#x-Yjc2~zpku#q=sQrGc%oY9dDkX@z&}YOWNy2ZJ=ufd)E5oeTY{6grc5D5;-M~
z{QE}$h`3R;HeI>Z%&&@5>sf_`1e}T6)DfkeyAP4$E{Amvqo@+5K&{z&TV{4gq0EG|
z_(2PHH6Z0EdHT|GtUU)O?Qp2J`xF|P0A;w;+O5Cjvqzs04iABPHBi=U8%N{)mep#W
zo6Yy06Zr3`<_(dJm|N-;$Ei=uBE1Hew_d~?dR-esKWuCC5c2iwVj{%v;wmfokg+Md
ziXp9WTu^kNUrB=Iqj^HL9*3H`dOVN0+B!=Cv^NH9Hb4m+AnczLJnYQyP*U*cp&xS8
zT>GxQG3<Sb{1N-!_tya$IV6B<`k4u~z0uG(03Ic_v`Chgm6^j}tC_I|OD7|*n1qEB
zC1@nn)bI(IfTO#hgBA=<(EHys*upy?N^H5<5kJqXlV|MgVYRhFDIP}(95pNW_yhz&
zAtBpIfB~@eZebQg%4aBktbo<^`svl(HT`(wey(|>A8QWhtAshno1R}v0)?fr^7I9O
zP5}O4m;!^d=~jmB>{x0R=*T!Z@ujH-j*n{&kBoG8b!{Z^@$-}PmdYzC(s6S~&CJY*
z#>9U6CJihQ3<m!N=O_Xn5CAmjdtqTXATFTsfmFl=#BDZ^(b$+08j2Y%dZq=$gb)%=
z(-vQ}>RBs3tC_B|T{s!<2SPwR#_hpGaarAgcy!F{?9kPo=+V*9-*a=}oy%z%8NS+O
zroc`^ChEr1-roKoNnVTb1sefUek)4P>$5R`61+$xT2f4-1k+fKFj_-teuYiht?ljH
zoSX#|@e4$Ne(U;i<hO5a)watnJ0+c3<?r!6NxX~rbqn>mEPT$&8ZY9?B_kt)Nx_Gi
zJ)!`CJY!0-dBn@h+n2x_Qe7=T%xTb934<fVahqXNMFbGBYJ!~i^ca=megO-1IUReO
zo1dSNmlxa}#iXvK)%80~DqhS}NF_T8jeybN%q%=2Vm;0{JgvGqF4oYCEOpSi3;Q)A
zE$u8kl#E-e(sGLb%nfKN$;*2lrd@b&)5MKwBk*f6^p3BuNUUkfXX$=x>}Pj>f2LuR
z7jRtT>MEFJZg<qpWWN?~#s%#mG;rI@6n%7Z-%<$+7g$2QI^Tcw;luNwpdfmB`eLtJ
zH=y3ut+Z_3KIL*dd9RtT^#TTC$HKxQBqDP3-vC~kkdW}O{l0mFXw!x#r>dj#(#=f>
zkjA&@=!gBir>6yI2Z)W8gF+c9Dk?1Dm0^H_AHw|9K5Zl_5(;)kadGkSy^TQVAkf(p
z^iG=?V8%vZM(~vYB_n>&&f!J)`ZOGV9ybR7@LuA8WmQ;A9L>4!<Dou2b8Gf;en<MW
z1}5jk-FS5~D_SRrW4c%cq|rcvtdp21ucQJ*RPp#J!O6;!LD%$V4)Su=Doe=W%@n0~
zk9KwAs?`1CM~J3<aW6DAi4|ixGHYt$WkQIDiVVf|I7p=YG1I<(|9HJu7ZDtc9vmFp
zcxJ&!Pw#No;BshW+Mm#_z?AF@gXim3VZ$rQ0E^{pyf9i<Vdmfn2h9JaqN0L=LUSqG
zfcU*fX=!Pku(MSx9yhl5jm_7nDA(mMeo{V5($)Z+_1x@~rGr+S?DFz&m&faXaS(o&
zkv8!LOw0Iar8~)WEgrCL6l#x;y{kqeS;`rinUN_e16j(60Nk5o+d<aW*?D}mU1Y3H
z+~41ipp?K9R9`QW_w#2w5N~#OciTET?3RL<hJU{0%u<SfBmroUfq~({V*%G#bPy8G
z-n442Da=a9#8|^afAVnJGe7(Z07Ule3CUVFNKxCv{nr2#zRXjXfj}j06t*4i3gZ7&
z8Ej{Dd%IeHxTG#}w;+0RmA(&N_Q@4U($`H>=5}*X*MHO9`%3&J^G56D$D@8|{ZXH>
zuOwp9K%ty?a=~;RB@vmdLyd6EMDmdqV3t5YAtWIIc=#a~VY)e-$-~Dt{QI{A4ekbI
z%AUmI{x?50!A5PUSEt(n6WhBY33Ef)Zd;?dm{cO9fY;48dYW5XXTM4n11vK#Dr$rw
zaDU<anuUdBaa8>k^zP=;a;~1N6+{Yzfz8d$WD??9Jq~gTio@&MX?RFrmsezJ>MI}$
z0juzBg_sT`0jfh=PjTk}jG~=Kc|i1B4?w27zcOgWv;#}noNp3ePxTh_^Yg1e>R~M}
zFF)A&S@bnFw(|*zo`jVZhjrs6WAnQ&SQLD!hK7Ti*{K77vp?wDxFdA&AxiV|R$TkZ
zqOj?MBH)F$MEmt;kG$`D!}L5ip8&mRb`g}O2}9$XKr7GT9R(*Ofe-s(XzV#tlkd>s
zePgNay7Af^_iVF+s(Y7ij0|vQ<9%`L3FN-3`_406#}VSvPF~hPatoaCB7QHePXgLM
zZ{kxATD+-fxmjZYfTCqXeiArE9vPmIe%*8jKwe0{hiBRnh=#nBZc(tFfmwp4w>|)8
z-0aHAKnKsT^{;no!^5@hafue=`HMh2TI`~Wc)YPAR06~Z%_0LpZpvmk1HHZ5Nt|Ht
zAqL6{zj72H7LGLc&SFR5hwUs^JzMR~8yR5BB}5+@9+tGT<D!!drD0^uvX=uvS{@sC
zp1z@SUBMHf!U4oe>Aq9qB9g*E0%g-F4+J(yyddcZ<pH6EO{@50tbtq5<fInhNV$1=
z=|GgfxNtEYOzCe6A{awwr=>O2RZ_wPw(aD3M=~<9m9;fsi<8=)Yv}Inr2~T_)6>%j
z@KS*{#mC13fevK1BLARszzpG)u;u#W0Xi|Ulr*V;BS3S*<KzB-<Fy6iWn^R^;8IY8
zOl&iN!6k67SO9)?BRO3j^!5$~4jCPrYUB6~T!PEIZSXeZwdoldjCJ{d{z5#8?bYPB
zhVO2-hyVpa7k)1SEsOM&kg?%RpyL1S{cv*p!-oi{p^aJ-qT3kg)wpax0PF$p>9wuZ
zv;4`-gpWPqeV|%D=i@Gw=i}v?3;Ec)OT)t8pLFx6XS`quXzwuF&f+SpTlTW-T}yMf
zGX7m0X{pl`1GH8txgHZO5k%PtP=2$1z8n;=N4Iefz!Wu?DFL=Z$q&7)zMo%M{@OjW
znSFZBef-0`{tWof26iS=VY<CRuas7@n#80fjh3EKe~%E_%Heo6x$9(qce!qyz?hkt
z$-~dTeCQ&tps-2#a)3vfQ8{5$km{t7o`J!3(T~JwQ?2A$U=RrzlZKTwM3q_k<;#a~
z<kDb(fR2vd)!n^Gsl&m*;75JCAvZcP0gwdTmDImE=k<?THv7If@X7WB(1P#+m<H^I
z7#!6fJ33^)e*Fq0YpTq?6WgmxA)LVB53>N^zldd37?+r>7c&?)PMAZr*Tn_GjRDTo
z-9PHi9Ot6V$*X6v@xOA(4|dTLq8}|qMf|SIp$7UHNJhsejwcem2L(xPz5$(g5B(zQ
z^B#WrNg}{Wt|Cd_YE38AJ9p?4;0!{~`TEvUugU+~sv+fDxdZvO#U{{_*;KnmY~XeE
zT9ZRt9%w+Im?(g(Odjfa?XtUL`oN7_$`An^AB-z<zJv1t_y;Y}2H{O9cBf@kbH(+%
zm){s^m4g4E%+OG3{tijI!3b?)&WTeYSE286#Tan7%Q4=nF9|8lZ^VUD8B+VxD&2$&
z*!yQr54eFi^d>fGFkrudOmsLjY&9=&4`@+L`RU~(mdKY6o$PwOv+VB?5DGMKh*H6M
zclHW&+xA9wLRuFmjW30*GD^1xj3?BU>X2QtM`}6;nwJm1Bq!h2^q#ngm_VV9u6^65
zPX0&w>p@{*v&LR{K#qi$bWRlGJmX9wC>ZaDLGm>vMdif{M00cVeC;xSz*kBwr?~9w
z>;gkV;KqVLYNzq`?Z_l>qP*R{eCSYT$A#&$1%ta>=5{_{9t;?3cCE?x84V55Z&SR!
z0jMt;Yb;wRwg?D98?DP91@O#UH5<|k1v;5PGPf^l&MMt=f99?w80vNTSNwE24Dx=h
zq1H22HPfGkWl?ihO3i(^!VR2KcW<kAIab|=U6zTh0ijSH=$kVqBSR@z9C7_pg^>0x
zaEvwPOG;$-V9^y_XNt>nmDrG*Ihnq@6)7MZ)HGoH!SC9~-#0;9e{JN<04>cKS=q~B
zA#@D9$yxqx<3$F|ADdg{U#k2nO#@O3-JbENiR{3YQHLdbe1t@xFKt>^x1tlMHa9;3
z`7e-Fw=UZJu5-2q%*h`Dh*%kjOSrJ4>7hJI?0|S7#}8&m!`f!aGLn*y%#t06)w7f!
zhfD1Ol-L9Rr?Iz=iu(KB#Z?eg8UzO^2}PwtIz$jeN(H4s7`nSdrCUV03_@xEk%ple
zkQllK7`g@+x`y~2-k<mVeb-%g-E}|n&uhV&IInZ|*=O%(Kl^!!%o9)A32SR>r+0m{
zkj6;)o@}j`A^F|b$0U0@+!D7YM@E{%?-RXHM9#$c9~1s;XHWt74^r~rUe}$TOZ@f{
z%4;>TUzQNpU9$Z$I1m~UgQ(y#t~apCuXqj5i`?a|xt)l*SNx5v^HJVc`{_`lB)}PK
z-Qeq3p|*7(BEajX8Owlx;Qj%CXgKWa0zNfMy$>*bE}qN;2rG5xr*T!1jBw|P@v6!G
z65HpT#DncF>w5Bg9Y25GCJ6vjp`)wYctzZ}e{$@peDQjEu7R7f{fDS_o!ek`!lFf&
zARytHb#OQsH=lQ_1puF7+yhs)kM#T|FU%dHqoNis&q-ERycopbI2XUPbS4o`9XTjh
z+t}DF&)0Gj{xE-y6kVo2_4BW^aSgVpjFa+Js+xuQQ{G`BLB!t3%F5>3*<$rv9KYy$
zrP)<nTwBhXKr7eGtUTB*=uVtU!ysy_fbT4l-s)BVr=nuU;VJJ-3wD>{4wLP=#8Ax$
z<3j9#*pC8RsBV9ArTDeIj?($O;vzGpaWDC3$?zBWfM)=D0p_Mg4b4_R-PfY9Mw(t#
zVDje&USqx}(Qt0qJTnVGr_mCdt&(aj4~0SIzNf{V2pB72Mi6kc^^Ux*L`~6pxo@OE
z=;%<M-)A>Rsj8>!z)4jr0mnGG`St_`2J+M;UA&^U4frE2-)d}LsfiaYWkl&Lb^duw
zi^o@q;Jx<eblt|2Ck06-BPJuo<5$+#W$mbJ@uR}_orYTDJeP;>9BtL-fwGpggGf3t
zt(4e=!sIkpEyIG~hD1_OX4;qbPII4_szBJC_3@`V#V#lDq^(r**_!W!VtMygQ85vI
zj`sy})uN*N99sK9ML=HR`4fontOo#<`Fy%5PWQsmyPbdB`BVrs^$ytyNEw2Q>x;_X
z)ar3ZiyssJ{=M55<p&U&g<X)ffhcOS-XLH|eT?SzZ5lqXVI9Qu^>^b)K*z%UN_VDq
z2Ip&h24t}}A|+2mgkXj(Ay}36MYnzsd>D(vY@Mp7zyKc9^Ow;|_^<_AZcPIU-RG~}
zPb#Nn<;@6M5cG<LqY=~}TNiF$TBXN46{!{8iVO*oHhEm{v;h4)#&|tlf%|Gh=JQ>%
z(t0bpJBH`t$^u>{fW5DJaZV<(n6@h(Ly9|K>J!DTqA!e=(d=nn1c-DOFGRWxF((lq
znSs&(!U#2U&<g5MiLMncmvGE_SG1A7H-FbB04l#0^UzA1v4Qjr^>rWyr`s@5^Px8p
zz=Uj;qax;W8%_X4$CZ??jy1wlg70+8ydiS>wD$wpiC?j2Ta?NQe_~wNuCb@N(Tw%e
zCjs7PjQ@5jth$svXjxy@CGoH2Q(u)QSop!U+%#WCRzx(GpEB6A6y?C{ZJ0nkE;<F9
z^)MW~98-rT(k}QWAV`{wC=4ETE;g_hyJz;UmqT@+iM8kru855rt-+RVB#49M2Dcbu
zq_=o}XLsj(u=vGC_V%Cei~!0IE1N|9k(srf*W}1zl5w<B3M5x;@%$=OgGDo(EWdfg
zRU0=ZZb4IAGR`N@{w%?e8)Ei*HU@tdEy!wJtRvh$Vvc_Naar1Yl}{*ME3`Q&C0>z7
zJ&tl;KgYnTeoC$4blmhb#;Mjmsi*47W#O*pRE)gi<aqYX_>CaG0?SwX3FVhMVy?p|
z#h#AP=?9|(UEw9wHmtx^2|gK7*Q{q~5d<S<6&H>fNq7ecP9JP_lXAu#sp}c)RUFZj
zlXmQ=N?Yjs$rBNJRj)fVaT&?kB#t+R0%}sMK>JbeuTq6lw?%#cA3tK`{Se-xJNB+L
zXcij1J}}HjzEuX;?mM`>Dd_Fci^Y8uY66f|v%aYrM)G(fDv?A7|J0mc`1Ws3uvB|@
zWgE#CXC=sHS~iEb5gpCccb*=^R8q|E4~$Iy)fYaG&OF_DY~Xt!efyB9QLXrhA(&(_
zRmgELKGCU-$!>9*aMK7i!qv+DU|?jR87w$m_)CJVLzMkJ@*A&8O<8X$b^@l_=QwnT
zwcQzqvvfLw^>9iCSh_8Wm`Na7$=`cQrd|>zg`9C|Zg(v>9<BE`>x(PiZjVrljgRa4
z@EuR<_j_)@qN8k)k8bPlXIXJ*obxO@@|dXAa~({_6$0JKeFp^8TRfEy^V70WuO`$`
z-@D6T)NvFouG5M_=_K#2x?EOq16y0UeGS~x4a|iB1EXeBLj!+@#J^u?X!wn8eR^P)
z37}~Mg7Y<SU_6O`(fXqzKYZb<;N+Cc%z4HG%A33GL}~+o)hT_nifpQ|w6$zrb>W`_
z9{R9+S@iHnW?a&3q<NWi+eX-Tz`O=zI=`K+0+EBPK@c?f?{+s_bmAfFn)Jz}WqHYG
zPB^&Dn%z}r|6tF^TX-!BivbpoZd^szesK2lxS2rBOn>W<h`%BLtSpbG3<=d#NTP;@
z->DmFhS-j$7Y~d2ZF2i#-)r-Njx1V18+^OUpIvQ0T8yoGglkc**@(YvH*~)pu&92G
zy-~e4!_Ll=sb#Nc{);4t1<)ob6c}*~^rQz`2C}}Ape^vAf~N5vyEIc`SKHtF5gseW
zm9d{c%WZ#<Ny-U$->BwQ%-NVzlOGOT_O_y;p`*y}GsBS(7ij30z-H$d+*}zaNVLAc
zvOuO*uI~pJue*77oo&yG3nseWOBfvt*d@fOE@iVbHEnFqxYQgiV(pFFL)32&;GLf4
zTU*-MEb;i)eaX~!Bnb#6!ye?SleB!Rs0fat6W%Q6rVW;HwkA6N(13y~bX9Z$o-QDR
zs*8C%JR=}BL-Y`~t0Z+q#7K7o$}>dDiykok+pBIOs4|lRqf~74IP+GwLDdBcApl|0
z!q9C1260!?+3QI>vQ)fMchRooA;FjCc^=f-=6z^!e|?h^wYHLDyz%5n2e;CA{2aM)
zqX#om)Pg+}@YtH+DKEa@Z}M#W^89p{M;Mw!&=VDfAMlYU@^$dlVTQ`5=jSmzZ4+Jj
zb*Fn!khPcs;8t{Ys7w@iWslR3-pSG!DCpeUz8Q9Ua-u7~Mq)Ew6?-yqH#m~MOhH<6
z@G6hgHZM3BKNyUQf}F{Guj*&@Lb!F3I8;m*u&M#C-f_a+|EWPH3u85w6jr@&H=(0@
z$=R<yO-WT$R-M3jZQ?%gKQlE|*=x5(|2D%gRFa@eFRW~A*vN@jh*W2*&4ZFpYP>Q|
z7WC^Fa%CEnAxAZ5gKii1HdiJaq^tB`XElBonDj)q$xpsU&Z4=h5_O2tCxd}psX{v)
zXa2Jai3Ux8UaqPR#mlp#6%>9#dFwh~=(JrqydwBL?@^IO7q4<ac7Y~^-rH&<9-m>O
z8s@{-mEEbWh>;3TCn7KBlSBX3II<}u_2^k1R2jE2e?~|KUD~tEGB;B%HKmfq^_2G$
zU1^~JtPAPx)2)U=19+eS633_5ndfbQZxLYHvJ{2s?BoIT)?;0F=Fcs&@B8bJNf&4}
z0O5I)G<kcUMl_e1YBNA5?p~tTWYVXf>Fn!b|8(n7-a0>2)B4jD5aU1PEuJk)3eK{c
zl$acN4`>_M3L^QFgGYnRR55eo9v40KcIm$rBHMlBaElee*ha=vf2~BD*PLK*+Am6X
z^X!i3@%71_+AfccYWAwrL*e&HTjT3$fUka^OVT%{#^_293CZ4u^oIhwnmE8Nu0H;o
zDX<oLGtFUbygN-Qs=&TFqVCf(;=!ycmwRN>zI5;dw{}Kf`nGJm1}9%4{D3;~Mg1>v
zU(R#4d>#^R>W}vD5(5D75xw7SEAuxVn|wC+xTP|nF_H;c(PY(U*A@O|R{lJ%ONc>w
z`xzSEy6D<riS5wTX+WZ0Z)jB(*RZ1Y_vfEI$+wWh@!4zDQ&ysSO4&0&8JN`GM-S0K
z<oyV3-x)pjo3lOYZcDOk^yIS4J7*=LV!Mo~BHdqwp_lJ&4t`Q_X~Q=OKJ_!y5?%Hu
zo;TFvqQ7A3kP;WZ*do1Q+VtAku(TJU$@~FHiX6$m2b|0c_hS(p>80)x73L$1QSFT0
zOJnmM?Z0ib5eb&D)?r_hp7NRVeZYTljrd+{f7GV@vfga4`^7JLswg>@+pkC@GMyc+
zxV)|R#nd<Cdu{4seX&#M3({3($3V~7xonS&7)Rowja^#WyKL&(ojiWCUG$ivCS?W)
z4jxlJ8|kWp3=_L~O$UXtD1fow_{0jI5@=UeQxi=_kw#0P{cszFh7B1T-eRhG(~%!D
zV5Bu-hud@whlCoVCKeV}%#>k%3()#!jn+Jph_|EDfpmNd%B7+5KYdOTFUZc1YyE?;
z-ah-94V^gqhUd7`AWr@26s>NfGgHl-2_>Ex95Wody=2X^JI-^K3MCZ5b0vw=(8$rQ
zK6tjoW6&IJYd<(S5{(gK`|{{V=?$8rwPL@3;)@;IZvCnxr2is1L*s&N-VuJ(#sgY*
zoSM&Ux_the^lgZX?9>%`>+LpEXCRCB(q2B$swxRfNm&~&8#g_FEa`$Ht<$DIo~=ty
z)Rq_vpP8%TmLVq(^eYv&;kzvw9Y7*WPZ$YB7L{G+VxzptpWIzac61{wsiV;UM>O4H
zZ?S*e#c5P=>HL|ygI*%<Lxb1*gy=J+&0_?12Xird_<{}Uq3#@CT({}<&d`JJ0yzrr
z9*mLdHupQ`px@V;X#7HL?XkJ7u(Gk4<vOirbK@fuBh<EbT=;YU*g*G==bf+X?t+4X
zGa{t1F9IdgC378~LT(@0LMV?IjmU2s)-J6u<7NZ|dxwNy5$0;?2|9-dk;q0uOGX?9
z&83h7)CI?znHPNi2k3LEu6ITEiLZV0nEBdemJ2s^IhvU|>{#Ay!v%)e=FTfKh2240
zhdb{%k9V#*KMoIIxY};CmNhaJC}iSMl8HBkORjz2f0IylUPQy>aWt+*I#CYi+%f)L
zK&p2x^}xn?r4HkCIxo^2;@^F&I}we?BYAk+e+z~FhJ7mCrCH2M8RayJKC`c|^-axX
z^i1llXZTZrG)JNAw?}e>J)Bws_FP=&R>a@O;+Os}E<oe>p#f#^(D3N4^~LOksj4rl
z5OFEZ+8U;M>kG$?fDaMHbzWYZzshI4ulCsX_-t$XMb%!++K<&*AMSmOFK2?w1YOlT
zAR2wGi*CF)vA;;x%=G`5%QV%9J|Ab|6O(XJ^)c03i-Q@Ljy+O|b*A^Q$I$s;gd2LN
z(`zj(Q<b1fbQe4I4Mtsbx}M8L!r_C(k>7JFqJeiG7<}?n#g;)o9S1(63n4V{-PWct
zD2c2&5gjbph>W2|x`JCegWyT~pC|?g6KIznu$+*(l&v+Z_0=~7m^1nYjv9`4IIBrC
zXJ^($zrZ)}I#7HEs}BwT8@_r#KzMT0f<Ar#rf_Jlsx2d3+`}DHd%6%J{jH#^%EjP;
zdKlKz#%J5}V14Yl=Ca@UCeL|N@%fr?^<LcQwx<87AuZ!wsjw4%t=^dLTGlcFz2-6U
z8P9v45ZpSJfc|7Xf~%=3Q8k=xNY92`Yzd$4UmVdqadoyZED6Qf9Sfo=>wDZU)?>74
zoJcg$9^%yMm}^SM{thF(dFC7J+^HvKS4Z@HXfHM^{W6u?h6W4?V6}^bU|Xi3!dcs4
z7^m5`M0NC-@=VKqrC)G5wgFL5i8kK+bX_F6C1pRcZ)ikSggVO_vvyQ}v8{PJBktei
ze!fQ*${wfZx}nlab#o(BV9)D7BBno<*$2t0A3G=Pk8RkE@!wrO-MBcx4hai6_GfDX
zc(`gK5#D|7>pk{+^({Ai?k=qEK;FP_Q!$}PM*)Jb=R9w#K<^P175y1l{HIMfzw>Tr
z>-^2l({g|0Nf&I+8hsjLbJQt$l!166Jl$)XdBIX$s3v%K{x8NZZQGB6;ObA-mfKFE
zylqz#r^Kh1vFE2T4{zV{nXhbU`$An%LIXeIxQNa?+f!Jd-oC1&2{T%p*gu_PFwFUk
z+&%O^miDyoqSN!+FZYuEvg4y29UVQPIX&J|XIkJI^i81E@jW`)w$M<pHpjO6!6AH`
zUZwmsW$Xap{Xoa6ZJAV}5(ma6-n}bhEF{^M%v>1?ON_=dXkyJoQpF&`<RYBe0`A?*
z?23Ky;D&6(e3fbWKv;Rn>ac=ViJ<c{Bd%rK9<kS?<Loe^4!P0_ac|h=O#JhNBIB=w
z2BO3oC3W{M_xI}Zv&i}w9qXF`@4ZOSbI5CjXAa!3a%LIw?j$O@BR@we*!+h_1R1<H
z$ntJ7LDOiJ`bA0c{Tf?a@qs(i5OZx!X~>2J^w9aY*Acy8`_ehMD`i8B3L(&r)oxw$
zj6inmdMxe>D^?aI`RQkLzz^!1hvvKYFx9WNLfPlEtR1Irw12>QY7zBoedQ(ImWa_3
zlqbipqDX8}v)QV<2E)vWaF6IXjdQ{<<Z2H})CJJL$LWp8KflbqA`4;E_8#1zz9sjA
zf`uaF;uzPU@;LobM&T_s$`Aw5s-s$ZM4JdH3^P;39ZZJ&ek|f#O;*&u8BIzvTK%Wu
zPugy#qpWO;&BS=iWLwwQsowW)s9IZ&#G<s~So2utonK=j0#e^v$R}TXbrBR}-4lsc
z;!h{gt^4q)JIB{1`UX#Xl>mHAsGS+;tOr@PLv2b4=$Bj4u{8>QoPo>`1eZbIHU02o
zn{|u@r(Y{Zls2mOXmfgE4oP;7f`p6>%}+v!+2nqf(l)Rst5WA;1DVcY9uilV`XZ8&
zpq0Y=a446?=uqrmlNS%>@tbB{p)a!D2rM%yt2`H}d+_{kRC8H}=fW@6pc5Sp)oi~7
z&dtvPV@o&BVZ4Dnz8j-_?GD{R=^AQIQX$I$qj{A^tJ8X2i%#h^HopK;pioelKd`(g
zx^A=ly{uZU!C6emucRU7u)@wLE706fEmDnFOv2|~qSEfI>nGmcv~;xn1@ZvV5$Rp1
z^AIibL7RUoj*0#Lz3fZZtz-I@fB+`U38ewEbbXW&`3x?oj4mSz4<~-GQ-<^JhU<%a
z?42_S@M{qF5nlLsER$XMub`ROZSL^k)dVo1_a&XG*IzTaJM7`K6yy;nm7=|3$suTe
z6OYiu4Bs@|MFZ2T4r|}%0G@c{-_AR)*H25k%*H>ya0H?AKPQ$xX+jGKi8<{ln?<K1
zL~Oi8J}SF8Xt(10j(iDDNM~vge%(*eR~h|}asy>})8`qzXQX_vL)Z55q9k<3_I4ul
zeA?KnHxi}PHqyKffHW#WJEA=a!jx$s=;Gv)wq44%Q_UGiMde|Tk#UwzytjXCR)MkJ
zpv2}rb^I`XhJqV^*GfmnkoN2q=CgXjm5IujXHvpbVQ{Q~82>ZK$?aRB1ScoqcOq_F
zd%*snWBIp>t4sITc3W9Ik&C-qA8Ig)Bw)3eU`bHWY1iB@S^4?!P5pGWOZ3d<M$2dL
z>%Xa})RX`6O9}P19*I=zB}^{h2R)141VcGE+<qO`UljCr?XKO)9>dUx@Qn9#B&z<|
zM|cJ!smliYR=sg>$^-$LXWHG#$H&Y28k${PREAqzaTM+w857mpVPdp0h??s8xXE4j
zEQ6QW($YHDcKX?ac7@fol{CgGNeKyipKWb^YHGRp*g3_*a#vz9lC36h`^<qSL#md&
z&fi~sd(|P3`#nqn>;Kt(dQwhbmKA<yYtZ{VOq4`r@+cyckcdZO|Dms>zMjwG6N$4j
z!cGm1H?QA5mrBeo&5kmxIv=4`Vr96W3yo|#BsyKmY*652Ts|;7xfMS=EP*>BAn+n1
zk6;JWrsv+3YiVI;oMrz3!B~};pQ5vKOrN3ozUO&jpO8KCq0fbQ>zw1-8Xbg*ZgGvy
z?x*d!MDPir=UObGgWer9$LG9X3+k1AX6?znD`J#@#PVjTd#L6@`g|G7=iX!tt3P&Y
z=3Q9LoI)<Nidc^r{vDL|{+z$2!yg~N@POSC9E=>W>RHxAuJVHp5tAa*!2+Iv(Ls`@
zV7TCGgL!Xwt~YrnS32-zR-MMxm2Tg6!+K=$oTnN>eQdt^pXFk|z<y=r=Ol!c^1&K5
z$^?ak<I=l#x_^=ZjaC^>8mK?Q00y6Z_fEemdVK*uMtYAbMHfZOK)V!|Src>|L8DT<
z#xELcimRwR-Wm1t_1aH*(OTlNcvq<5#Q9xO?aF$fIbE>aPrieVX+`oQJ;xrBs%Rya
z)!!bgx8AH0ry6+1^4_?NXr7x=gIH{K9KmSZKd`9hIz@7O=Q`tQKiPP#wRl7Vr{ogl
zi}t77iHqAI{(mQrBdT`Pg=1B7$7+sU#SYi!+SSl9f;98#!ou6l>M=Xj#izP!badL)
zc9JpDr&RD0f8T)igm>V~j#UOX{QfxKRqv#;V64uOm8q^azM5zWv=zj>di}QTd7^<L
zIljd1ifrB4(t{C0zw6%#QlHDonIx3Zyjf)fBguv5C1*<SiX?}>)R|Z^Dl$D@Bl0}`
zj?;pBG2AgcR&*MtCp%v;Tv?k`!qpwK*KuagL*Bvkk*{>6WoCB4==jIb3QD>}AnG+O
zX&^%c?I-HDfly&fB4$F=`Guy&@AH8IEnHZH$70Z(vF-EPKYxVFOn2qy!tjOt4~QU$
ztu-Q~Di72`wks%6uW3=!u&37EcJ|DD+W7p(^TA|C6AyCqZeBhtJ;U3cOR)UU?kQEw
zs}Ao$3@WpKE0Rx(?#)~MmV?9T9A~WEJD4sRO^Oa`{NNyUk7<SZhAdd#s`ukoSMsf_
zw5v>Rwm;>k{`tK+Y_{p+Dzm$}Rd3gK0oF)9VzPrc<f1m*iXmC~7V#eAN=|WQ$mOie
z4t<TB5Zw=y38ixlmHVmZ*|Xmb+N;qVFe582jdyPc#oQT3o{6b9JvHPpx-pf`Aot_7
zd@w=j2dtXwW<R`**CLPM(xY0d6m6`JA75qgc8^ug`2{P2?g!n07G=As;(_Yxl@A&Z
zWb}O{f40c-`k(U&jPtQD%i`flocGmd4N$XvQ6#{5m#W{KbWB@K)2w?KEx=$=U~Oe-
znW<#K9N+sHxgQX8K8?p|Sbi7rB=D-9*YJ}ofmgHmlUL~zUz$gc(V%y4w2(V@+;pUI
zm}KRlaglFv7QEY{qg2Dvl52M+D2Y299GeJcYD^1$P|<KGM8vHXjH$8l_9jX+^u-%#
z-s-!Dl-`C2Po8_WBjm?woe72gPNL_Yof3c;kk+e4tZLucFvY-O>d<$MPPb|)2X%Bj
zsP_KjlRNs&bwPLf{J(|hO5wWL)GdF<Q7MC<eKdi%>CorXy{yk<;dk&IA~av`_%fl%
zN^TdYy{%iL;nXg@i*T*pnV{osH!`tsSmSFeVDvP&iay+~yT&K&H$-#`;P9be@vOfC
z<0UzqYWeGWCpJCJ<nZ9IdZ`_Cbp^XId(6%>pOR{4?(r=X)=SD=w~)V@M~al(e)yAB
zcJg$UN7;~`{l+yPbZW0GoQXWDdfnqf|85EjxJnd+k=_V_!_MKPXhP+FEI5l62H{Oj
z(XUL9jnO6(i^q57d^t%A2Cd#(QoA%<Z4g;a1wt9}J}LjJ@a0+W5Q+A~n`E>`utM0Q
zj7bqbdDl8p>n1*4!9FEdTZVqEy)4ZD(d>ogn&9V3np*>9YNhAD3vRT0^SJ*buYWQp
z!7YK)@HyLEQyY^&<(yBve}Kt@IP`K$I?#u76~i-vbC&WrSekT;8=7l%g`CfdGVjyy
z1TTa5KLqPN)NA>Eu_nUfXu9^MPp0b~Lsfr!GO|;QKAZX@5Z{5=`TO_0WCqmqV$Y~S
z^R5KJiI?_o<0l0p`Y+Tpg{RVSWBX0sfBqnUg56Xjb^s2<PSXYl1~uuEd(y|V(sS}G
z_<oLJFI`ng>}`C)>T4?o%Nha@Om#6M)w|UT`PPoEmO_^XKe(m$T3bckn(i`MH};`N
zA1dYy;LAKk)dhhBhOzzvAqVA?I?hYY%PUCLvFiK1x_D=0b#389S6Ycp=k(xwiYs+|
zds-YMt2^sZvcaK&_W04Mj=G=-&x37})4dFT=ENT3me02K4T{bvy%f~a&IqR10j+-_
zCzXjuEUBNK^~Mt1!0+8N_q+o8_x=dRV1kq6q>eJTvuug|4!lVm51O4ui5y=l@uE2w
zvVy(5W6HeSMb67*y8(w|mFDbt_4VDB7D*{aub!h$gVs-HBth=;_|vzsJJ^P@cxr2&
zW=AP6!85Jym6cz<i|LZxx)mo`+|_r3uVU-Ulh1Gmuf%gRTr;cEv?QclP3F|n@hxt&
zTp5ib4`!y@tW7O|)yC$hh(grq;nI8V%g)*oNLKaNqGtldukQU}-uC!(@5aLg=PH;I
zXGS9gL_ZJXsOC-hsk1mZ8t+OJf6#oWOdWHdCW0!eEr$tP#SK32OiBR~4Dvd91T!&S
zFag2MYnl0LcY&uJc_x+Ea#){-3#^L05hk}2i0l40DH@VfQ7JbVzwdyhL-uRX-KTgo
zC5+b&akE1ysLy8$R#d1kWp}Om4x;zgSUI&)m@{7ZtRohEGBadi#JTHEk`8IpA@iH=
z8H-;k9yF_K<qnAi>!S|(`y6ie6m8VFxQ4C0-)+z2!+yFQX0r3=&nl$l;JeH_6Pi-5
z={bQN2?NhFl>u9hoQBu&g9yiOwLb&eT)9C5-}31$!`-`PgT0JP+y)1wXy@?c?&~&u
zEmagg=;c|ai_dg^+huI(wDgGg^ycl$+928@$IlX=9+c)-*!3{Dif}CX-sZ?K7seJ1
zO-f68@8j}S+4O#?7BTWrk?jivZ+x8%%F3$ocJw_xA(WUv_WNgqD=Ua>K(qs>rYH`x
z-&YHR@>A{_mp;r8j-RJb)RItT&SWRXGv039`;ar=n30YeKbmv)IWK7VgsCAWA>G=l
zci%lOKE1W7yc8&(!9aLmhx%*Ev7C8AO~8OK-8hNRu05a2<2{?T4N<jp00~*+v_5M5
zxJd|L1((_XepHmg`jH2}B1>}Gb{45`nWKK9$z$IRx`OY&^VdQPXrKS&`_a#e&LZ{q
z;mk<deqbA^r0;vScR^NhM~y8swRLEK$pU_<h1h+I$&5b$L$l0;oj7N#b+ei8Bq7@(
z5}tf%JC(*t@aaq5P^8V<ZWTx4up<i34xJY{gINQstGL?Ol|9=qxo{JHS2+t6M*co(
zX{!g;N$36$#^m{Ep(@H?;NvKg!yh-k|7gPBv*GTu7#OfStJ+mEiI7|?rO`nx@0Q!2
zE0iP*l?qK~cyLQj-Ki2a^FAnP03nB*8LhR^bTiS>u>VEvhx_SyB31{Yvl4q0w3{Y7
z6P;!fP<jwok{UP#da%f_B#SI|d$c3nBHHF#skXo18vv(5aqKa-jUSfx8s+>Xj^Vsc
z6|o!nwdJ423JTNT<BhNHTpkPm>68EG@|D^-fAldWD7cP4OGHqK2K5nxvv9la{Rw53
z)5-0``?t)S-113Z>JzqcL?>&XOK@grHbT@7h=gb#!Vaw?G(T{-N6*jTuHVbM6A1|Z
zU$N}Wt#jO|hB)o*H~VX`gXp&J_cKbW>})+pWmuZ&iltxfaShAlJu~NIqzoJCUR!>x
z9&3yc=b2C1xfmFw+*xZM1xk<E_4N#N6Z~ic2Zk3t4KH{gpe^LlGTGizh5Tdd$AS}D
z&q_VgUI@OKcI}&Z@z&MsUE~+!rdVk2+Ok8_^$6`Mho-`@k$X2*u&c&BDNhP@%B|i-
zO3^_><!)PFwr>|!$KH559B(n=Y}sOYJo-^7du^t!I3T8wYBU}0;!sW0^4UW0B_|E5
zSQnYjrDWjWO<4w)PQ}*@p`==Jy<|GJvS79TX_WadFPAcx;PrpI^a>w*ZXmm;l<eI|
zTiD39h~u|acjd($j2{~i{?oRnKuwUaO_7<7s_8IiAYEt5c1B(?kZOh<*3#1MKo}0j
z_x*&y-{<uJyNv#f6kDGhX=~N$cyE{>i6Wb!{;@vQuXOQI@d1KMd>W<3jdA#99T`1^
z=Mj`AL>OVa>C+I=tdTmE&br{d$!xnMW#epK+Ov+|Cd5usuMc%u-$Zc8jql5lwFF#m
z4~{_o99trvVraRYGeOTp0$D?rdgu4}isM!=$w}j8hawEM;kKLS-$G_TUW2kSRg9No
zEo_&@&t%j}d-MfR-Kyo)_L_G_CU%MgUT3D8+hn>XZb)G;jOJ)&TT+q*TBe)fwo7_7
z2PA{#*v~==*MbR1t;@XA*0UZ}h+g%bqFSG-Zdk`W1?ynu;xYu^0UnS`R|S$U+uK1A
zfroRmfu8+%U}cp*%=|w-JM8Ob8<d!#w$`DWo}c<Hh(WKR5jskOG})hd$dgab;8?1N
zwk-)VC^0~jo_&8Rc=tZB?5BJ8XkuomCFl`Nj`Ve(6a*Q?y;2E$up+h-pBtrBQ$NO#
z#7yO2aaTghEpPhxo29}65_u|U39cGJ7)NrN4(p&5mPK3#J+D!blmyHf(#mgb9dXjI
zH0+?jiX$ROlWAA#Clz$^thY>WU_{R2*7t6REFy!OUOYs3(Q$1?V#1t#z>Mu~P|RCG
zR*m5ls{DX?I2qml91q6zmW+N+A;FW*NJ+p0kB;Zi%$NP!-PgyP>UO~EzRho1DUdYv
zua~u6&|X;3{*rgQ)PsK1^=7~oXgJg6N2Zw?0BaSNgEN#IUnDw!FZkZWX-3t0Of^ht
zhH$Xy9<l6XhvNjX^bZvrg{{-6P!ykn<sjp@*_Kgin{Z+aiZ}+9BWah)JktCT#&}IS
z>CDw3W4(;ZcF=EKY~?}Ht!Wk(^CvF~h~!ia*zSU9+_QCB%^G+V`DhpD0XgUm%UELW
zsf^$Gbm@~iwjsfUAxZR^edb<yUSm%m$F+;7YTGY#^pR7|&*ukx8If80qcTo@*(4@P
zR9HUrWJvw>+dQh|mtzRYbbWl8V^vhi_iPwE<8S_B;AAl03UjglbQ94h@&#}Azy2Mk
z9T`loG=k{aD5`?#j-W(@#N^$a-nhOR5VPRGh>^$cD&RMYv`Lq7DKed=rPu~YM&q4?
zN(ztl8UG=$T|F6X9_xU-Eq`kYPpA3G|1hsK_I1A1<Ez&#C>}1|=rIusIFpQzzseFl
z&)bh~K@~@L*SM~b?Hy1?wf!o}Fzu?Rl}(~U=nwp`N*iwKeWR@(WcXqgb26~cf9%7k
z+iuNAjrm3?T`cx+QAUF~Q#$+p_2<3!VHn=A`o#Y4S6y;;lErl#PdTH@KEL2kUP$5K
zNj7kI5%N6UCe7#gqIVKTUA88Rnfa}}>Y~#oQV{cu<gDJeioj}Z>7hvcg2-Us*kmAP
z;x{bd<Fp^~s{*ly!yb4froh9PzY$AcIw{~zHaLIXtLzpb|C24QbFRbr?c3o(AC~y|
zctEFA*VQeke{8-Or14f?1(M}DW7sUUxwgWl-yf(B+tts7f4t#4EOWCq2^OtNqbX=P
zU3R1G$*i@N)ewbwXJIDS&)1Biy2RuY<aEJ0KMGBJZGOS>xUt!5SLkSp^K-=>zGi@c
z|1|NX%z)=hAX&=9CZb}ZW4o4ysP9WWb7YO>zt9=rl@e0rP8~jf%q%>&hOxY~gXEq+
zy}QY)^KiWS#roc|@}Xg5zZi1ali@#IWcbi<O$a>`-Ds%RO<a|$o%eY0u3ljP-=%fj
z`*GZR>0J2>p0iT0oIK?Lr*BDdLt$b*UZ@;vRcxr6Rv6|U=kBFh1Wa!{S3SX-`qG_s
z<CdeZzTwN9x<tP6kuNWEx8D-phv5al6xdi6pRyuZ*m2sM5CR+lglTwmKf7D0+<DDG
zKerI`{#$V-u$N&OMa(xE#IQobrO9{e2>lrv8vc^^t1lFqcQU4PfW@@(EK?DXjmTuL
zeRFZWf;)7jJQ6c{b$ND}aY2cJr64&hT1x|c_0VG<)Y~x<bkK(jyJb6%i^_6Izrjmi
zgE3s*6>70eR0654@bxnm_tl%NXK}~l`sH?eddBh31gDfIOo+Ogircn8oQGojpl7Fj
zsII&ih2Y?ue?8)_TXU7WmTP?N|3d;LwqxU;kD0Hj+UzEZ5`l^P?#n}oH-~&DW&I2o
zeTi9RTLM46Ly^apsTJ6UragN~yMjURz0z0Le0OyM(z2^2rN~J$|GS_#!(gok(jpKe
z^^!k6WqUWE67W$=W~aT67jmVZJA31_4Ox02vP#}`Fzc=wQUV+yFHOER(vJZG1?lSI
zexgA$$L;W0DkYFJeZ-d9?BBJrvMwFjX)h>#qy4t+N48ROTAImV7f+e{GU)(yd||^n
z|HLwqC!$6u*qA|1Fjr5mS1L*(B25tB=PZmdiHjq)Q-8t$*33L7pQAv}wY6kjygi-3
z1?i6Lm!B6|_^qAced-8M)4#^}6FMS&!}?1YoX(d#wo*CIp^x(->(iwH*o%F`vekIa
zvEQBo_Vw_vjNY@lok=2zy`N08BBZqy_RBD=U*f0b*dDXvX~R=dfDhvK4i9$0&#a%!
z1Q=U!g)mBK5#9wlO8nmd+^<d^C4W;TOJbB+C^)Rt5ftNSPaDm9o?|*S-l<N#JB>?E
zGQrT<^YTfx4^VT2vB#MTH;BY4fAwc*{egeI(q`=*@wF5k-J>Y<1&_u%O$q|$XiSMS
zN(jE2>)+g_Y0wGyTL^Yo-51k>cYv2dwmpJLJyP@ELvZTU+(sWpd@qkA)uHTUTk%o;
zXM45faCfbz+<4*|_M51$9b({rYzjbkEG)h5cugLDuG_Fby8-d@LP~<WcK(AjY<y}j
zJv!;5UhT@493D<(2&_w9L1A@cLvpDO+h59e%V748%|c%zd9-6*ky3&J1Egi46(8u8
zvZSjD<mSaz6Vd$k80QCQ1poOQ*~N#7>fNeE=S)`OX9m8ro}k#r7m>(?$JuX@MU1(p
zK2iYtcc0VgtgS7Rd&HZ1Hh&|L9np7sgg%-=FK!4=&v^SEFB6jxD~Bjr%K|hSz<FVY
z$4kOsLizcG-Wed9bZW#8)n5;WWBg9LzKKV-6xiDszn`ywGHRLZiq@5@Bn9oNu2Ker
zo%*Dk`JqPq@XdgqEiDq|MtCsHP8bL5{pRU)kTqg>Y9n!*Cs!+d&q=K?OP>MUVqy`R
z&7nD4Ln=cH%ahb69A<w<_>Z=dX{)kIL}Po0M(f{&*2e-+@y7*^!`SCEMn~Su0yW$8
zTSx*Zz0_G7kMpQ|Ls#Z$B6>br{6dX}OZT0d@8qQ0+{Z`6crY!0zJ{uu9cV~bDIeqH
z#mOUVWq|AZ00?#b;%h=7VoMvl*^BdZboP&YtD1{z=;gND%pcW7=~2@D9}DbjVwIoY
z4>S2v+TTyB*&tPyU<L(F^Xn%#9+l{I83<+YA_Z?@IRM&Ki%NK()accHlvR)A0|0Yl
zV>1!WuDJhEh|gSj!BG?1FJ!h!Ch3++4rc}xfb%mwLd!n)n*1s%l3dG7Nx5saW))h~
zKlnS{h}_uLGril4K17aR=5ah5O$4AgVC#+_U~R{n?aveKV_%vBScZuzN?sf8#g%I4
znf!6ufTjQ8h|Ys{MAzDJng$r*WtKr^&Gq)l+bsUovIdFd`9AlVE9jFb^a22rTO1kI
zV?8OM^g0tL&$H~rqnUJfQ2!=igdZwI5|dFzv?*o_fQ_*`hDXLL#{Y(Z!%oX-+(tXS
z%I%NXufjZ*c-2Jhsxkr6Bv2)(y2$XZI!&?Gf(JDj7hJhQzH|AE0JA5V7i*fHi`C1X
z?p?Y*YZa)__PNtl|1+AyY1cdG<2X>uM&od+RD@l#_Z}zqQ0f#_FteCv;x_F4ai}ua
z`AU1VDt|I)KhXDnW9Ct5&Xb&_loiMhp+)pf^#>%n102X)SOQOm9MrzLzr=G`T39s)
zCnOxWrJ!_5M<~_Us^7OcE%CrVJY!bN`SD3Ztfitzx0FU|H=C!-q4WBm-gLx!e`z8b
z!>3|#qP55_$N)7TrK0n*I~!Y|l4H{mzI2SJQ=8@}DYqvpil<!{eM6|89!$bT{dSns
z03P{*?M+$^KmRjk!<mUxd*<YqZ^X?>0r*XiSp2lurCsgDz}mvPJ`E2D|NBzPvn2`?
z!4ulP1hqdEqXNft?nagjlsKT8bH7yQZh`fFj_22a=1R0*YEI6ByLaynF|&lg{l~OV
zCC7qE=zitp&4zQp`ZB)KZu-uPrda)CW64e(u_q=zh%q{|GTy%ZiojXF=sJ`)#HOcd
z1)6(dTcTB1F47Lro5q<cRmV_d678M0q9|l86zi>5@*~WG^PrR|ao5+2KytHcYYjSW
z?qVESi}p-Vnn<~!Hfy`}o*~fh$+lBoK+Az|DNlqnufm?=wkQP16+~chRTOtd9sioK
zP3f$R;kY9-1u4Woiit2HAQfaJzgfychj(c(D!aT!5GSPMVTs+8*gKeB9w~)rpep3H
zD1%Jf*XL#_6mDc>ic6olga!}GnW7gMh+mE*^xv-`c{@A|5g;;4(vIN!+V;kE3BsH?
ze>$mZd(mXN_7@YL{OYZK6jX5V6*#2oslw8p$$?_$eevYsa=Mq0lpm!J+xg0VkY>63
zZ6N(8@r>@8m^*>Tzm3<q{w>gV66O4jbCgl)2atl-Ak2p0Nxw3`^o$)grjbS_4oP9#
z+x#g@4x9Uqn8}LY2Ix^pOGh8?JQM`gqcU=sfAr%z3YX;d=oOe&vW>Zj56BYs-?(0Z
z(KWg5yXmZn-H_#a%MOhA@<vz5T$IG(m>;t3x%w*NM6=RrE42tMeTHQ^akTrzp<l0(
z#D}J_%|pP+NlV9Y83nCD@toH%U0xdQHKg)je1!pko+45sQTa~{WbmID2-X3@8Vz3?
zu5_H(R@ZT0Q=V9d#=BpnFW(WC_#1|RUk1(RvEv{~Vm53%^obHlD2c()(W#=|E`qlN
zU4t*@k-i*3+{QfJ{d=jVXyVpoR^}1(N$f}9r+W`^u78~4v~OMBFnA~im>Ev9qHH_W
zbU2L1k%^3qnDXoPW1<lqtL35Buk-fdN}~u~30^sfMLEeo=U_P~fgF(W3W?<uW0b}a
zT+XO2zw~8ehLymkfqlL|7U|~yylsOs%;b}nL=NoXTQXh1q<XZw@5=?8tQIGX;oX{8
zUq>03o}yh&r-9g;#@iwv@AU7K-4~=IRW;sDTKl%(TNpq^@rVpW1EPxU5P(l6L+BLH
zBZ>Jj0O86n)I`0{Cv8_iXb|2n3<%GEdFfYffmk`2gv)i``PZ+3Fys3es5K~>a$@3X
zNJnH$%h$erb?uufHT-vpZOkG8*8E9ND3_`YWKAkXTve)`nY<#R-^CvKG6*K_$3OT-
zh*B!J@v!UEq+;CO$(T~fGiK!kUK}^x`vUmiG+z*nF0g4^=j9hYc$hUz3w{y+h>gLk
zAhTFYxqJFdUc4~Nl~)n6R{r3h;Uzc{<JF&$6PP^C2Z1%d--q&q7lPrM4`<|AAmOZW
z{aQ9Bl)ZLrpEMtmhsl>3y&w;igT9tPwm8y*%==K{N9hW}Vh{m|9x!_<NR!&1v+UXW
z<MpQ*)``8dIl9*G|7PPxp(HI-HiMM(6T;B}7?Ingmd>hBU(e#*b;t?sqy(<^d=XI;
zFt9Gra<fAFhp(I5Wh$>_OQZSdcttEpwslzSKXJ3Sc85mJigcfX+9+_ix)27&rd<h8
zTLiP3l>@8rpooI|=`2<B`+p-7o_nLOD?18Ty_=E>_-m0d(fyMn&yeSKZ&q)2x`bz(
zP5t>($uA}W?)4RV^2I3JI5qk*md$mwqq|JaqLo~yLTqVvJmnSSEiLUOfA>_pYnNB(
zT;2s`*;0Q)_9x{}L|{t(_Dn_2D9FkNLL@k)MyKY$^$mJ5RC5aqE!j>W`DJWx*6K9+
zxw9_E{Lh~Y&YpL5Q&uKM72n#D*tKR~G-{;!o<RhL%@BUzF1G5(P!fH^DlExt`mx_@
z%OZfv4zQ@EqoENVZmig^ha*iN56SWZK52gPefO=jy0i+pX62AZHEF(jdlmHLm`mzU
zv5}$8q4Nu6)yAm|xV-#4&??{O{w*=#KnTR_SAM0dD`+aP??;gw!I2$(TN(^|fHBiS
zKd{s~CX&Ekj7K5rPk>zKl#=vm2_^w*2@Y-4nigt*iT{Up#K@#$4R}Vp#18a?))8gS
z4D+BM(EcdnsTrUbi3AF@zLL25n3yXXT2e6<J+Ca;4vC!ojFPrXZ9^&b=QtAkqzQ2`
zP5Wu3j?gUQXIiATI|A+}`yWFmxEk^sg5i#VZ#OPdHPZ1=9O<N}OEvolPLkon@YOEU
zcU?%S1gTopBb%uJG475nrkGEnf(6VwMjGMz?s4!(Xi2JlKXv#&{XgKesV=SzLV%wp
ze{24WR0D*0%U3V=CwPHu?7zHRZkNl~e}DNuzS}7U;`;HRx`S^`uAGI1o%ILzXXn^*
zW^FVgpQBev6pfhJV#O|xe<;LK>Ad}W+|C!AhAm*dJW}xA`<dd^DtI|E1pU~3O&-ko
z%SL3=_ecLjA}SIWvL(K|0TNiqJ&LSfikiO!E<dm^*Tzwu6u)*uG{=Oq$=15~l>(sj
z-0V;LuOF`Qui_=h<1fMM|Nio?_RoKF`R~8<-*^5uU;4k4I+wN2@#z=+!PtG_;c2w7
z<LCGX>*FnHqRu5FR`%h*SrUM<=hPGcp{<8#S(e{;WJvo<A*FDfYEs8p2U%dcERBn@
zfB*h6F`?rc++h@%-X>cC$Ig3Z$fpCuSk?V0Z?K01Y?bcvv;RzzXq=14qyN&J{xy#O
zmpM5}5q<j5*$H?xSHPA9O}l#+O4~RuZ|`jhY|<g%!ZkI?1P2F4dpAHqKf95MiJZ9R
zKQYL{-rB~-q9I6uAKkunr4FdQ)E&)5+$Ta->qXaM;i>)o{ZKn3nco)P`7G1<$4eC%
z_HwS``h&Zv!uBJ*)D)t~=R?<7oN7EX6x`i~uhhC@moAlLFoqXm(drTghrXFUQfd;6
z?37{37u}BZJI4}0%Manm+pc^eeK)GtqT$s;W$9@hCGJFVt9!j>Wv5#q(kA1-O?_6C
z44ZY7Ai)sm72ucyjb2DFW=_LpEBR<5D5B)IB|rj^u}pz~HB8R<1a1WR(zWQIp9m)U
z+CShTg(zP;q`ke}<7hUd`s@IeS6oc^t)L)+J;UP@n5QqDY{FA6R6tTh;%Fv7Lqnr)
zc(`pb(Xcl|(i>EuqoYYb-J2=PJmZXtigIyt+gXTH`_f~knEhtbXSc_MKN)I=*qj2u
zJ?$*eZFms>vZu-xbpA{152<@(ASSDws}~b=Up$CgpQ=qX&+rTeznGt&Uj@byCdUGl
zB$DB9sS_No`e4FZyWZ;w(Cm>o+ZzNwbnViUmH=&#RR?b1erQwxCi4ZbOb|@|X<!Kx
zIJK`eE-bJ~A9t#0m6?&aPkXWhRhy-cgpBu6dv+$Qi;apFi42dffj;><AON5B<`H}^
zo2S38udzLfE|<w4wADK;r@Qm}pZhTSY{}dTYUXRP0_(QBjATz%W^~l`^_9%TO)ekp
zBsO$|xfp}{ibJ8$N!L+*P=W~;$JWJRiW@RC9zTZ!oAA@xRWH)OYB$A%Y{o8!g_ORK
z$XMQvFZKyawTCAI$&t^~(+1QZLOVoX0!Yhgf)+Yf<;dD;@`V>jz&W!Qyes-qS(ykP
ziWHZSKy!daBge)dva+(bZ(Rq|rB&<3MNLgTs4hmssrB?hoY8D^5Kx?s-`unVvNg|i
zb?F!w7$T#i`9(xLCbxVf`YpazPJ2bf#!`T(1xxcCEVmp>6WF(*ZN^}Uue9aSDkX4!
z&&e_KdfB<P#Y-n_msMUKf5*@#_TG&Sv$DvjsQ$6BU(fzyK*t`S|M+-Kv*B-J;}tT=
zJ#r!%&PE+2FhO_pUHbRZW&@nVfA+8Joi6_ja<t|GP6Es_(Duk(ra>6<0dX2oLf9__
zx)zrkYH(13bjLRnhi%9N1a4UUfsPLvFf%|N6fFQf-z}w2gVwuk=q&Z7N0cXYao0uy
zX`KJe+n|u#!k@Acejw{&Wo@ku_YnmmEvG=>Iv0o?DL-hLnYs7&?OOo;v&1&KPQ~b*
zu@K7w`5|Dc#AIY@(o)IFAYm>UahhrT$jT@T)*YM%Y7*P{V4C+14t`q>g1j67Bu$Kf
zI7#cgBXAO6CV+SfznIux-~hS+J6BRta>1#ZRlcIA9CR%J5uYYt8831N!5=KKp{jRp
zQ;3FCqt6d28;b16V+S84cJT=bweKu;g~|UEeDb8!6$756r`tZ)76GJa%*TG45&<uf
zQ&be?BBo+!cqz;DAXyps(Bk5l>qIo4RdZ>Xm=L=XATd(}i>xMnwzoHrLZR|tum>~|
zrC`Jf-1?xjmpYxG#d-Oy<8___Jat>wbSv%d$90yNbW#qi3WDs89(Rb%&B+P25$XaW
zM!Eh7lzl_bbiKEB$@{CvJ{czbib_gEAVufpZ5+r}(gwOrOg>vzc^Xdn!Cq!fR~?+3
zECj-Dq6)OS`9Ye`7wV{kFUibW1~A~Wcw{sm-p)z<C@%66<H;EN!o(|ZbXP)ew9S#e
znU{^YWz~}5*vHQ+uO>j|c*e;!gh$jJpt#Das&h5mc^Xz;k2}DIiJt%bf-PrRhe?kG
z+lQXTNj2^bc)+o0_QhD7Ntre<qA<B<9Jc>|xi@qPq<Q~a_4ohvSBKHLzlem7P{sAS
TI`GiMD~ium<zUa=fBOFb%Ll@B

diff --git a/docs/_build/html/_images/header.png b/docs/_build/html/_images/header.png
deleted file mode 100644
index 89cec360a26b08b1ae58ab459b7752d92129fd92..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 1646470
zcmeFYWmuG58#Rp5ji7X=bT>E>(k;?8DBU0pgOr4HHz*~N(mgZ?(%ndR!w?hS+}_Xg
zzQ=bQ-@otg^8=X81=n1&_uA*V&ULOGrKO>Wk3)%rf`WpttR%04f`ZYBf`VFug@OEy
zEfbm`^2ZYo9YtA`sxj&V<N(b^MqLI4<!3xD6o`%-W4kIDdZ3^X^#1epWYFdN2NV=U
zv9i33u8-L<>^`04m*+b6TGN^FaZ7coVWy$sV~Fh6A6hS;<Iw+%98v1cOoPXr-bcX-
zcdz3;jHa^+3lEqY<W(xQn_fnJlMi{`kIsx0^7`59cRQVfogIPsj~#HXsjJy9kz|?A
z5}zfmYlcaCH?962T^&7n5XOyE(#K~ABa+2pczkkg?u~n!jQ#I0EFn;#ritb6rT_cw
zgSlKc*?<1vpMU!2ia3$h`I~=V{+}xZXwL(l{pb3B?}z#X-TeRF>F=xU*RTI~yZ_DI
z|A6xUnt}iAivLBF{~zHKZRTT#>A!%wmD&fNk{+{>{2kV6Vx!S^N_=cPULv{VlKD^M
zUDW(<_#P6BG*3WxJyv`9Tz~`*gN2$PJ6JkSGB>W?h+A;fdK&K-1>G`?TJ7Z><KO80
zh>g?vpA=vr_JkDeorcfO?JB<-*gPl$E|z1C)kMGV{<Oxc&B@q%#a8Hd^LyDLcti;A
zzYmjNu<_+@3Q3U03sWaxt8rg4a6W=;#pTYC;>7`TjDs>jIlF1Nd;9bIJ^U-Ve0Z;k
z=6K0N^$a@9(v2#uZ($civbQ45xaU9q%~FryH|t(abpOd<LxS-nQupC{V`JD2P_HpD
zNwP7QGt5B^dURBP#><^!6zCngmh-wxd<~3VYSY8*e#_PEvA28GE=2r<&t~N(NnA?+
zvFt7Be-<JEhW6j=CB;f~E%qeE++wE*(gs|2V%T;cggF!JVW2{vR@|K9e*3iU64VR=
zi><ws>T(0q@eSn{0@lS8s5K<<jBUfj|B!_N7Y0e;CO)!5(@*qy%=iDM+#%4P$)@St
zaM*``=cn|5OzR&1IHzN}^-@zB3l$0sUWP2oWNh8iXk=`Yaa8UNYJ%w$!Rt5B8^M>c
zvE;04H=7@UKRtW+J_WMf^^h)EF7(jI3fR0PV*23SmD<`MiTZCU){OtN25Q3m??hE?
zpm7k%KB1NjPC=0OaDX>(FmA_2>BgLL-7cEB!vr*lw&G|)(EvP-02l}cZ{*AyB~Kai
z!y`FUxNu4dY6jwN9g;rf1@qw<i<&Y*EmW}<&0JFng7oSC&G82U(B-q_UXZ}wclYJW
zaqR<pKIH&g3}?#7a@m+rqdbn$-7<)E&z%ka2@?%n>9Sr@1pOKJbHfV-(4I59IVl7g
zg)Wp`W6qK_j;(>;tV%LuWD87EuwhY^aD^pwZ{TddkpBzBTU<#<8uZPAChOm~c6c56
zGtW(yLo}5h?XB}sm@ePK`b!>^;0Oe{@q)O_5Ixv=9elGzxU~-veYpbaHK)_WWs5<B
zw_oR@>aU;22igXUZ_N{z&)R$80WkkiKdYtMP}!RO&)&jqyC)s__fri$8RRW+efr~{
z&x5%IwKyW&6NTZ`dUTCqJJRYm{~_GEHO(P7mh3hTqC8#L0apLv9aK2(Z%6g!5#fv&
z0W|VCOUSd~^^*4Ri8?XH<(kV5$+;F^{P=eUAH)Bw`TU>tx@90b6+@Ll4bum8zlb40
z>0Z6}7@+H1#~&a>BN*c$dYwbqmXPo!;|tvJ4c`T`Y)ouVE3`Kru-+B!aT<F5XdWLs
z<t?Ql)f@Jtw+Kw*sH;iE@3@!?K)h<yko?U{6ndeVx*ulQ_Kqy<_>JrAInMvrL*y~S
zp9Ham9g>*=k05*AS5(d7-g|U+=(sSk?k7DEL8b2-*qx5mP2stFE}XKFl{j1vBU^(l
z_+w{ey-wRf<a4SBxAVap=8NO!@oyE4*)caEk7tBOaSSLPA$(0|VHO&1e^)Z?qS=H|
ze{}V%DgGP45@0>&F439<kH7w|?y?#ZLilTKfsVOt>(y6Vghw%SeUrTOC`b6Wl8@a$
zi?dtsel=jM0K0R8U<>LGY+W5=%2f?B<voG>i1y;1Z5jY7$WG)SylLp35lm3}papY4
z=jiJknAI}J(d;DXE5gKCQ5@>V*a?4cLb4#cfsX3UhUGEPB15!$j3ut*<IVpcC}6hT
z5BwJj63tJ%M284ge%AHQVgGPFMl%Hcd2v3t6BmrT=lTdkg@GQ>@d*|nkLZLa(xT(I
zn?AM?@3FAD{M=AHj{HvZ01fj93%-pCe~|DiGV9k6FD(j_)}yxjt&7L*P7la!hrrP(
z<!RS_zcX)~BJ@en%Em<rl!}xLQsBo%crjL{uei|DEOLbDH8zS3wtzZ(7T;R&x`Yzf
zzTL3582N9gTn*T=0%ZOZf5-7a-N__k_f2CP<RP3*&ZLb;_k*2+u;%XTSd3ZB@MULO
z9KD=f@;G2Zqdvwn7aZ}8@nkSw&F2CArt%%nZei;7Fxfm-@(N*A*vV2p+qKTt8^Pyh
zMub-zlV1j3e+Xr~hWY>>Va?|te=<6seX@H{-?&_9!^1+gV?)Qu&gD0*-d#eF>0)jT
zZV0bhJXSDyj5!eo5<&_?W<9Hu|7Ms)LdO%Q>t7cCKJz1iXyge-?UP$r$jgl|y+oxC
z>XAEWdN2%Nj^z2^7REe3(HdZa0C0rg35md5@V=EKNSZ9_xQ5wnF4SFx0rM2q-eI*v
zeRqsT+FHNbhaAqFSmN_nkF_7*x8=S5eYHpu5Ny^u22_H2hQ*?v`3=%`V9nw*67@;T
zFF$#j`C*J@#X~=cFx;r)s=RG%tk~{p{93D(S!>7n`N3hY>ur|{Sk!H287Dr?d+#Fu
z@dV>y^Fe`f45bMfoE^VX><abOc!uyf<_5{;i8SjhQ2r+Y0BzQ;G5!-M^CX-LmH`_l
z7-dfyK@4fjkciM+K{2mCun;ahgU4PJ1WI!qDo;-Dy$5K%mkzV*;_W2$Psi)ahsbk`
zE?R}7G8>njLtnQnU~VJZz8pnJa9mN+Ght;zBX-rwxg;ea4w=d`4NTeK#&CW=&#-+H
z72|I6h=o9dvzoz{a8+JCRkXdi@Posaqmxzv`%h+Re#U2A#ak=w@6+|hb%X>~-z~5G
zT<$sR6{1_ERcrTKG&J#FjvQT~fkXYAHYT?~8*AnYCT$uUJ9F4q4w2$A6#Im#LmHua
zzx5RV6RLYVemN2T%S}EEx`vJrF#^}Y;h%7pZG~DJ&q|(ughK~h2v<lgx)#rx=xju7
zEtbO3HC_w9q$TsJ>vjq<AcI(k@ZK$Qt#Q9M-0yE;eg41*Z<lC?7QAYGDqW?;M;{(w
zWEODIo+Ih40ebjlbJ+`i_^Dei95rA`iIT(`HtGHl{Z*uK^~VD5r>hEM+|}uqN9=_+
zOP`v~gl5|7849L|4ol1&9SoRHyRN3{u2WY<aK>`I&ud;uu^W*UpNHRU29Q_VoUIVB
zF0UzIZd7I7EDTlBrT;^J1n{B>K^7Bp4_mGwd|<by{N!r6k=$sOu1p`;WXh!$ylMW1
z4)+9dxYOw`;J>goeu~<~r$Mw+Uy(C0+#)XNeN4C&2Xh|7vy}G#)V(8vx)K-L^EAv8
z1Vx>NyE?BI*F6iJT6RQf2mIzcaQxhV@a-gi>=;<)8D|Rfw6i=)iZIqPN*XWTt`s#-
zbP3uszMLAV9-4C3hF<{SY?t{U2&DarB2KxDCdIF?rSUeM<?IbluII4e_bR*w#aCJ`
z!3_bKp<zsB*&!1nl90neG?%~-Ui1*jzPdCGvTtJ4{t-(LT-?<>)PX%JSok*!xK|VF
zz3hxnrhFF&2swJwTvo3n(BgBd*|yy-m+49XfY3_}fSK`i;{;0R2JQfMw?&SY<3hl*
zS_<1(NA<r^{R6m0+QqN8`)^{sj0p`k0lyxm0P{IWi-vh(+``{)98-s41mu3|rgP@3
zz<&AQvvP-Wc`(j7e8F6it_aXfS_fWidhf{Q6-;i@JkhtFzQ$n$92Od<3zo(WufxCr
zny>CQ0yD+@bOM=oZb!$rmuAb|A*a&!TL5@4^jr~;YF)?l%!x|RjM-|oM3S-+J11b(
zgc`MC5FMXaZaO02_ndYQ|I|D_UhQXc*{?MNgx<VTmoe^ts2SE#d@+YGg_;w1Hcz}Z
zG*|r=KW&tged&MVK=Qr28heJ`B2>`{LW(&je)ZJf>FKGMf~g;WrD9O_QhD=3fRUrk
zgK?!xP%BSG-4TE`Y@PS^=Nqc9?#^tUN#5rkN0Hj}3bdH@VvUr4p;K#afN-*Z{1-an
z78b8*F?!gW)?N5s;-UgzPkH{yFg!6-GvQPVO5eA79xez^E!X=oG49u6VllqZvAv1_
ztg%I5`aI2u-SU|d66x(n@V!VOq|~B4H4#HaoKvN=nuvL;!MoQH;9>tWFw;SR4EOXQ
z@s&xr)jb3OU;hPw2EjFTc=TC_qwU3b%}6i}pVmw+aB1+qQSWa|i(8Z%!3n93t2cJS
z{@eijTp#OcG<71xLLIFnl$PLE$Z)c@^ga^(C;?rLane@lq>YmFr~UG*S1e=Gk|+Lp
z5|I?TtcVpOnz4*D0eY{4EDzR3!W)z)Z(eR=*BNndKc6Q->&TPzF(4)cFVmZJJWzZb
zgMlzw&X{5M{*?5DZSPJ`emIbM6+VR0fdMuXb2;ZDl~kVt!3mCqKELYljh-80{w1V~
zyvJkNd6<7m#SrZPZ@W8c=@`MI#@#klj3JH0RvZB5K9A_{T&EqtAC~a<mWsSNhPmCl
zxrM-m@-W~~F_GabdXbSfk9DrR6KgHMF%0AF;hmM*?*JmHCk}Le)VNLCn;wcgZR?L|
z?XaDBmh2H_0YRQ_C5^y0Gx9s#u=2}l#Er{HsK225P8|}8W71`!##6J`U!S3q^v!(A
z`mRsb@$gB{kXNyUV)}MUq+3^M*+Yz(>ApTkJkTVL4!{trpbFpCM3|70rE{5-uO>N)
zF!7pz%rwgLjYUs~%%MA)#X4mwT$FVJeVa0wSsY4V`B=;QYggJqsvNVeyMxmP9~W9J
z(*TLC%r7Gd<A=Amn^!s{nVOu}n~vCtqpEh+MWie6B8QJna{SKu{BY;jH4@;?O;I<l
zzDMI*&Y{=&^%^F3z>!VOGq>S0O_z>#$v;hPXCWXkkykfeBOj3;&`+L?y!U6_`hF<t
z<?$!d>yOn2|8l3#Hs{MzV_ko@do8ZTr)dhdPwt_4V`CI-^&u$b<|HF6_b1e1UR|J$
zCnlm6I0zVsXe`{+g=l`UrN(*Lh0w$V^ssddnLa}(?`t|3h>UmumJ@B#2b@#sc$9Eu
zS;zAm*;g$oyEWRl6Kj;XC{qGhd3ZZ~%+>-S+~QuFHAP8w;sM+B%vTXYcxdJ{2le(M
zufetL96?ZXggpEn@VLIcTiHe&-6!;_Xln-@!_LpxbD~n+^Lvof%>ut?Wn=x4dXxf_
zfP@62OXo#2^pZD_i`S%+ixas`D%T(MFD1cOXI9Q$8pxGmK5M^Xxt|jSv!wHYUjh*K
zyxaNk=^&n<TN^-vbs^6$c@s0=o1}odKmP95`)-D2Jio?4x45GYBVo24f`U3c=^sFL
zhO}zq_NMhLUk_-dy$m`pCL=C~IeCAbNJ4gk5Tb}u=&0ta$)jgILRe}1m6q$CGm=hb
zi#4+jLv^({T_8k@r{`uV!bi*heCqBSEV-7%&ACsbc<<14;SS~r7xP^@rhXda?}nF5
zx+!ab)t2M71O$!p-MNJbA{Q>Lu}z=*-zqYGoNPVCi{Nh+nJ|bFns`Kk=P0p7W+@(=
zD*4&`)B7NUCc-T`^8O2JF+}g(Dh~W^r4!&1FK7^uU{DPWJ+zk=k+6w322=H8li?+0
zd%b(TY)$|2L{E`&^z*6r_U~VE?(k!N@sB<xV>#07He)#nK2Jlc3YKz{eh&dx3CqjA
z)El*CPr;ghYnrDc;6ZmmLv)CXd{sW`c`2WhwsN}`?p@YC-7sLiH>@rg=7sB@G|sYf
zchT#<*~c&LdD(2z8KBK$D#6>rkF})ZzV$G^jkt<Oklk;W0ur0n{oZw6owyA5tMaP8
ze8qlZZ3sGpzzd<{h<!Tcax-K$(fl>vdHcfm@noJQpLgGrEZwMs`;V>3;cV7Qi>I8P
zUfT9R<>Ms`0Ds#Ogs94qdd1?>4$lg>o)q2Q&fIIbl)E%WZ07$$90kG6KLCGz_nkO;
zeiYLnoafqCWzIRhBz{?CtlN>Q@AUh61z9M>x2=g5BA{*Oj}+TM83K530r4YVvkYU%
z{nx&H^Tk228h|hY63I$d^0VQEGGG##&^$hBy#7nJ@>sl1Pfowo<f~yZY-bMQS6Cj2
zcJW5`bWZm~%QHGW)|}CGdCBtF?1dSFR*>YigAUPktN{1Q8l76Jda?~0_S^Bkxh>ke
za69tm4t%^pRp)Ft!;i!)2rOo;I1x%79h&)yeSg?yp5(m-ddJGD!1DaF-djb>sEkuZ
z7Rp#v=Cql+pH+I(J)gf}$BHrvLY$GwBLa~Oz0XHzZ4HksEQB>6REjN)i<+A2KTJOp
zp;RxMr}U#i|L}pCH#|TuPtsqf^S;t$=u0}UD(Hj!^uAK1kuHEgidC^l+c-5J0kv^I
z+VvEd5@#jj%}dt>s!VIB)S0D$;gk6zXRx)i;mKabott^HR3=4MVXpdL^2EO?`liJ@
zYJR@ToKrF84t-`PoPD_mxC@5vD+aNX@#duMjpwFcM9C#7#s2`;Lr*jjy6_V)46r?x
zi&<OK%=HBqcgnH8C`LfYnD~oNh+_r7I+dnNy1NZ?;bXkgXs+DqM#XDPJRohU^mQFw
zHDKE}BcPDLI|MjQ_c?cXOQB%(dT*UZ5=hnv0%n-2{>JWYEGHP6?rux}x35wQ@VRVS
z|MNG!4+Zz*zkF=cu-ysMS%_CC^g0fEH-<leU19NQeH5SkA)Y^BzX@4~??cuBcl(ed
z@()eMc-dY-;(0ITI3dotbutG`PU)Q<;Q;al2UUG4tT3a)i4;c;M&3>z<u$~izNf#T
zg5qV8S|2kdwo0LPdYZ`W@qQ7S9Q#3|zWPvD^O6(RSSxYd=OUxZWlD1Z>|lNP&|X_3
zSAvO*x1(TQTRvW&g*IN5nKF?MwqHF1ZCL3H)U@6s{SZ(=PX~epwwy!7#f8Kfm<pMa
zH1?Qtx+E-V>}K_gRzp6&GnoFaQa774yICtQ3#6c;Nwu|?hrQW>=Z^101YHd3yxiWd
zFF_8t1M1#|I%@PJHZ#99<PI{ZmhOO)9MXO&F;e`IYL}Va!It^Fenx&Xs-o{LR(?wA
zo6)`Fdgd+m%*@P0I>g;zZGAmgKJ<Do3-N|MW?Lw>M9Sy<HRVl7z3^#V*i2RBBaxe3
zu7jFL<@?uX!rcia<)j0*=7L8JinXGKXRrA8r6eS!zSNw>!RFiitHd@iofjm4NEzzc
zQ+~u89oypLM$DDt-<B6rrq~e8&kg=7<UJBl)t_Pi;3JBw6htv*>pGkt3JGpo_I7)+
zN7x)JOEwlradr&zqIwd<lOy^02@lWSZd~v>h<D?%fx@4Cylu3gmeANmbJ)+W?YhCL
zOdD`W_zLw-n<O^!ISz$5=%Wfx9)SC-^RY2U64Ab`s8Brn!?-=sz+`DOOP51<`V)>~
zt)CQc>Qsn&Vuo7r%_jD*;i-z1R?p=yxx)7(#CU_uCD_bK>>e(94y?9wbcUuG={haU
zH-6J$=IXJCaMisJj*jv@7a8^CMD^%GPWV*&O67(PwP?cQ!nW_ewpAb)I$ph>{Eqf5
z3o~luaMP*9cZ<yZZX#vd9sQtQGVAPHT5*Y0tX*QLPPqs(rRsNO6#>cO`!LtasuJc^
z>o?Osl*@GI{1r{6TH4WM?GC)M?2QgZ*^PV*P4m+A3fVn8RCOxtKp{_j`L149fk1C|
z_sDkEU=%HWCndjskEuvyrhJref-r{3te9S$zME+7rMS>Y&f{GG>dW}jLHzvEs(WEW
zevl}cy=PuW_uhBC;$S%67|l2zR!>f_0(m1*H1KGin8}|6Asza!8r?BA;3BY$`<D;q
zagfJ&U_52Vi+R7yTjMUvP5pk|1?pz;dOX|>Lrio7g8&v9O0c)ya^H)(K!_t62?~NY
zTZnwIS@<Z@Q{3{bw@6S;%0lw8J^vy-XhFh53=wY}0sBL=Zj{sUXVMuHRMYQgZ0dV(
zDSkgUcs3G_yZ6D~UPnP;fSFWtKOBQW-L#4EkFBgkWps+39!nyJ5_4W!YWT7a2agh0
zT8bk`?ORmE49AH3<A5joa~uKC-S?7Y?P<Nx1BB41wei!$@3cq2@~X!SXFisbxoBM#
zPU858;$>%LrD?~8C1Sic!tp@%AI62+sd!<EUn6PLSyFj?mJEVRK2p6nu(a=1>e;M(
zH$tpg!dw9P>|Dn>f;-Kcg6q&XY4E##s&+AlIf){wc#zp5gm<(s?WZ=cs+;HJX4ivh
z3pBD%m>b(~T1>WFjpW_iX9oo)S%#J6&TZT~DvMzjqY8?rYb`9`D@laLE7&#J<>d(P
zuOpXs)<reFD$@=I#OXSM?H;=1G%a@IwPbyq_u^PR;&8068ips)1ZsnOL7JT+rD*1=
zrfI}nc$^3Q?8N*XR~2c9Kr;L2s6Yd7ngQ6FVlmLw%eNufCdrYNsdPLiQSNO-kx}(<
z$0KBC>_qYbrTRADpB4W=c)7p)YpmxV2Abg1k6lu8vOb7H-AH5)TtC}I_8o9B2AwCy
zoPw_;BjGaAh@ZGuVxgY!>xb#{Ey&)~#d2l(S!j=fXPy4yn8zja&U#rR@FxZzch;_^
z1Z&a3cV(9Mm#b$1FA<RSySSJ*(iu&q!<#hbBs*nh`UHE1jIhFIWvQyQtwu@Ep)h$8
z@v>Cfp`MjMmxtct$ZeeW$W<z8uYFmmCs(~*)Av>!|45xFvh2AugSJk5+m!QjPcY|K
zKA{Go_b1V>kIJ>H?OTanM)hs7M-6PUUfcf&6rxSnV&-X3QQ!O(aw=!>F4@{iW@$97
z<i)Im#uZhZ3SY>_{t_jVoDR;}sc&;~%N$IUUsU9;D6tfnrQMabM(whM?SwgWDl2}l
zt1!}Px0Upf7?G7}V=D+|nfGD;A*(t{Vr$1$sC3rOmr*)?N_6|CXxT8eJRf0Fa<-l<
z7s0x7o%xnhvHS9(8a@k74Z74sgmZC|-I$fUadHSSg2=2~2?@oax=w@wji*8Fr+Fg&
zWn(K1fR~yu2>Kr8Eewu~j;op?oNZqQ=`izb_%u+E@)=Z?pY3Jut3?{u&mym~!>C^~
zCVp?fT`Ir45Pg;QC7nmPOebv|@Tdm5{Ix!O<1qih+D}hUFa2gLyxQd-MpPSNfZ(?J
zw|E|Ey~ZuJSeXgYU@Bjyi$PX=Ptxizd1#SB13Xy75f<vP;62C_2M=xm0bqAqAD4j#
znl?l*JAzBmm*$vJm2*dC7Lwz?l2Z^DSHZ4FMO2@v$W}UTKH=e6G#Mu;vu&q?Z}Omr
zkXI=m1Q>6u-)>%n73mmaD}I}rdXM$q6@T7dIF^$+d{T}$&PpgQ>9^*3;;iLl3Epq3
ztE91bmDgK5FVc&eTR9ewzZc<bWzmPd$>6}>(wxY_Q7UZxzFZ|wf=zOrfnoI0(O6CR
zd)i@DOQVFHLmI<^iV|PaY`rj#(zXzQ=#`8G3xjR+z^M7l?^Xo<S%$HPKeD33u0h3l
z#U+C`e(wtB?-==i#nYrJadGd=inEjL6`64wQXlB5=%uAIXHz9Y_xZU!%%(pjXn0H;
zXa{kY<Wl0<OE+j&n;N(ME|E=X&RXjVYGSW6(muJNjcj+wb2fwN!XZIdLBm0h({w|}
z$KULt7A5Sos;50Z*r*<@c%=Z1XF%<@d940roqw17NLt_G46;r)XbnjoWDg3;T8N-Q
zvjm;+u@QyRgjp<j%hnAlr&kb;2{@!kd3^8?zi%*q^t?ay2D9yr=O@Rl8R+The{gn|
zsakJl_*+!b!TSlfPfv{fm^Mu+kpj~Yt>soz%E{j@o3BFGhAx<B_@uXM)rIf-QJ4u&
z&hlCOV7MEqku%@#MSRMABbW1$aG4btXyu~dni>S4;;#V5zNYRCTaT+Bdbk@80%*34
zJzh1!?qvdaG>UdK2~S9tBS24A*Mt|Hm6^FNH*wi^%FjBvA_!wugh<mSC$*J}r@Uao
z#!h_}PZ@2si>JhutrVPo^g5UFcuYe+fIu3=7a1{e>M9a%Yl59^)n&@L5{H$A@?K!e
zV4?rfdsi3cudYRTlZ9O<OX*!YtA#MdmMO4)i*4Ca`S~N|J1Q<nkI1ZxPDshltPbbR
zbT_5Eurz5Di*POq+FNzI*>vXT^1l4<j1{l)C#2|48wXnGJ?bDcW!#x+8cB=T9nV;c
zYyA|hPeB~TQ?=MotvJ69h6qBs<cl)Fq#-qX{!2<l)!{GRAKdKpj@Yf<@>q88s8LZw
z?FaHxPv7x{7wT1ELxpH7!bx%LJ6MVgtBs)}sw0=BLAR*)>!3#jTRsRRCmUR2JGyXy
zYr9sN+f@Lw4eWL%`c5}rBGO7QM7B8guvB&Ar<M1x<pn~50X!?^s0WAN(ZH{R5r(mq
zxDky9cxYyPL<PMY5AGPUKV=RQ{&X=<fDcZHFOL(HSa*u`%@Whi0+b2p#95W4HLN`5
zqB1iwQZ9b-|Ju2nI`yhA6ma2K`A2MAB!R|voH+j#AC3k2zHf(VXkei~d4U)9j_;>|
z+U7-b2WC&vwB0Z3x}@>F(QWq@Pj%L0elQ>Tf;uxXslE`U@YdYMC1LG71cqv`a3<Ot
zBK8=&4w4q}@WwlGVQM6XD^D}V4^CHk&yVFSf=qHKgIpeoiWZlbIz5h5Bq&lRWmVi4
zGtd~g>}OesKhslXQHePZA0GbT;^j47NLiDk4-M%~2@QU&logwnqRP_7789qrBor3w
z*IVCM0DBQv>a^73Rmai&$?VaK@KgH_rz5A#1f?x(GzsE;AxB(%VOwiYN$!rpa29g)
zY|?jF?#vBu<ELh!o<cz)G!{fltXPF@4682w(=1CC?YmTz?+uP{X(ditcZ;ha)ivsS
z<gskRr&vD}jc&f#{rGxZY^i*@+Exl_%h<jKQADvK8J+FD5Xnh&iauG93D44N5M$!c
zh8|g9C5f8<!if*1tr5mZXqtnSulCz{0S7OKhUkIfi#Xcky*x=KTwD)IMKvC<tBiso
z&Y_ub@O$Ht?aIO3)UV!*{yw)0*SE)mx#mIs%LIaWp*`J>0|k6A>*~5`-{bgob6#ig
z`ndCkW)tE*u8~mGaZ*%f{Rxctnr(tk27qdO%|D^&7Kl4veKDvif+5^*#5hjP-;vpV
zl?vHL+)@Q;u00-NCPlKT9pRFpN}#8@8=kJTa08$tbi)wD24p803@hW};aboa{|6_&
zKaO{^|1!eMlnX-ICrxMdKUU`O>BYG#K%$~UXp%PT@c^0)j8u#GK*Zb$VJN^X%+0O&
zeZ=~@!X4~pu!GEimkzc=we7MX;X`S)`qIU}_%#2Gh=br-Cbkk5{2GM`u(luXy+%aq
z%x!D%v6UF_<97d-(w$0CvxL2TNvk8nRFyBpIE^gX>Z($Sd|#(7-6zou-)|)gNW5Yh
zwN^~ZY^SA5(&A#9vK!i5i7A;(O1!|v4%3-(PEoHxKVil0Lzc;jehpl|bX8=v)jY_%
zO@b*){@C%Ohxi$!y1OY0E~hm=;G!{3;fJwAAE5gindIKq)*rqgdF9F1P>{w%{ly_m
z*=~L$OT*C8t>OY5|78c^K>KxZtQ~$w8JbhHoH|qLFBSTn7k0BnX2z)sg^a0}(e+uL
zQMR|c)~6#j`EXeUMTMQdvgxEzP5TuJstgxfT0I85nRfkheD)aH#GX6tMEXmLsk<*a
zyssuIm81kqtZ;^>Hc_!ZW+$nX&_A1SFEh^;;ire3naw&WDk!jSFQpz`@*;M?iR(!5
zCwW%qO^?i(SzP^bK$U5z*y;YywNn3Td{ZpE`R;(OMY~uaoty8d)!V%7SGPio<B*Ej
z%g(o+V6&{=B1{1MoaeW~&Zle-@5c$0Ug6^smA<)?wgY0=A1vX{Awe8chTOcoxrp;f
zDaeZ9NwY;9PznjB@K?4>+DASyu`;NRvUHnnWiy95a}k^kU?jsxD8&&*nqxlXhsw~A
zv!z@*yuSwr4Cs$f)P{JaE=MTL4Gl$R%6_#Nu(Ru?>r7HGD2)nlwpbyXRvN+Xd58tL
zWM=|l5<wv7APf{?RSx$i3Ce_)<O89{EF#0}UfZkYgv(B!fm+QS7SIZt`S>f$X<=`y
zdOzR|T?lau0xgutFG~KJEjLm{;8~iy+P11oCo^5pR-o@qgFaGz5+t5~91K0cq-5M!
z&EHg3W|>YJA!ANdQ*JP-d$O;_%FXb${!pcDdder%k4`14k?gaDact}EljK{wLm4$U
zy{1QPp(iqgCGe~IdU~NFSqD`K9iGZz=y4!>6_@Q@mO3t0Wrdbli5u{+xACKq71dt!
zfN-F#oo9z<_lACJRvK+Y55x2Zj)IUH3Ww!)EZYd+utH_NmJ*|OA1*hqtxZFbu1Q+z
zj@6?UwUGspP?^5!diEwiYp%xYtkK3|^K!l7s+$D0V&TPvzV}4Wt?Sh?)`%dhRiC@B
zqk$yD=popZ;{~(eH!(T3Mh1#%<!>w}3+C$9dXo(^GbFz5{XlN@a<G89WLvz0sSK9g
zo?-7->l5l~CvFAKL%#~~8z?~P)JgQo!F4vQt03TU)XZ87l)K1|QH);q3<L=6xnBN?
zPw4I)$TA<{7RU@9ka$o6aepMif%E`ILiliyn8Dq}*I*?<s=_6r(J>uZ=%Gu`CN!`l
zD4%s?GB=<KBGuvu0`&uRW=>U3&f4aBW`q5df2t8p^nW~69|)%aueUJl_+D?(<+#1Y
zkTnQ@k5x5C@_`{HJ8l9X3<T4U;Gf@JJwbg0HNl_I?_oe6m)`{Ch^_q=dx)nAJ<bz0
zSv@XUjFAC0T6!hP9Nc7g43k~$k9&^~56T4*2FIG-chLO4Cc*LR$#<<`QY0kXuAE2K
zV7!3pN2bfoFMB)C`?{n_|CBH$5m0??`-dKG3rK;138Q^a;`c!2j`sSOiLQ#Uv>PA8
zG}2u)`jp?>jxcW9x5D->&X9a^wkylbDT+lWE4FBWQul4TGcKJ;lbX^CEJdf-Y8AI%
zc=dIbKO^y-BWgtSQF?0Q(!L$6diO?hGUR%_F{;cOARJ|D2r*_ZSr#J@Qc~Ueq5687
z&{jI^+k}{E<~go!$VTO~o$+FA%d=mk3Js3@o>yqTnK5hSilx_P(j{dPYFM8e#<i+5
zc7JMRq{wrWTg@F#*JAV%LB9srKb!Hdb>=rM#I7m`E`hjXt!S~S8pT)oy&a)Qbrs?o
z{X7@7i#^Tq&AIe!wT_*TKh;uBeF4q$iJi9uYr$2c5_XT2^dEbh9?g2P(O|*&0@A#R
zil+6e>Pe|Oyw*6{h}cbE9~zg@W2-#o{OoYI(}hhxMFhq?*@GM(Qs~R(d4}<n|N85M
zjjVy~XK7BP^)8PBWA)WxLu&=8{s&L1H}SH!{RGY!x4+^)ci}UaiN^IrLY|UFwGxa+
z3mqD0{RK%I-alTn+AkIfXrV0Ry4@DMYw0Tm(SjNUJ81d_l8e}NIiE%XV`F3}n)>PS
zV@NFT(GB@P3|W%5TXq;eZy)bgR|*nPPsf^rnG_Kh&cPq&Rom7_X&}p?H^<oG8O>gu
z@CH;-lcUg2l4y(>_XK7?t!E7wSm@dxTLbP@c}EuCMp!ZDPx+J0ovW%8P9_zHzbVt@
zApRonWG@$~OCNx9*nuNk<io}A{ymAGOU&@~H4TZMj27x2Y&6-64+>2g--sYAwq`QR
z?5roGB@~W_9r&S>k(~4=3fg+<WHXvGb0{QHsvWJ2`qfZb=hgOakhZi+M5@wZpkUH-
zY?&?G#V%<gH=nF`Q<Xn*k*c^dwmp=H_8q5{=Yi3~l$}HSiyKt#nmmjQ?8&9ZG5yln
zB!Vz&%BN4+u}bLCf6P1RT|KMBwe?3A4qORk2neC)E4pAwHjFCv*j@XCD43Ibe*e@d
z7BdU@!qBkqI?>cK6lI)bD7**Bjiza2zkk2&&Y08e3JY3wUd<y*H}{cQ^WCHQ_2N=M
zP?D*<e1v$WeSqX{h_UL1WU85d#nI5fB}=X%z67GYYAk4rV!vU}^AKN(u|gecJXuJe
zH~Jf7rDfP?FA}}yw)oO9i1H}&#F}mMW^ca#<bj_L1`!8SJ*ZWk&G8a>1PaU(%7I7v
zo7Or!z_Q(BzSiU)AKBAK+>hp*?)!OPrK&WP@H!Z>?~fQ#doUdJ>bZ~|mOCt&PvB(d
z<kS!-wCJR^WMbZi|Mi=Az4#~8Z1Z;&d?sNOXGoqN2%*Or!2fJ<6lO?dBF@caLm2wx
zd60qVr;ti>CS2<TWXu`6!!#73{OINk666}=ojS5fSk-ggS@_}RxGjQbd;_N2ua1|o
zZlg#U(EIHr@Dk1@+{^cp`uDUoObUq5_Fl;R!n^Es192v+A;2g=UrOZ*<LUma7F&$e
z=6Qy<4KxmKqU{(PmbvYOY6jlD8{SFMRe574bmT|(TbY?UDZKy9&3sz@8ZS^_v3>1|
zkrV^cYs<{@Q|vUtFlvCfL_C~fp#U2sRGslqjSHMfLbjfxma@$K%xA<H)S~yLpTF11
zBP!=6JY7v(ZmLq1!Wbn`<NprPTa}gLoN!vy{j6szyecB}tF_VEpbJeU@|&HJ&ctT)
z=U&!|5ChH%h|OrMz93Nm10m7pvhJD4m);$EQ#nmVWmd+fhFW#DzV`Nk1q3xFSZKhB
zg3vCWyfJ$z2Bo~BSx1A$9wX7TnB5n?au4L1y(v+NwmPfClmMZfw00YkG$w7)8H)Oq
zA6BD^4VL3+iBVeuo_F;*uRL_(sPhtv@9i?l$A@_lQ;J9-37Qme-reTDxy|3=4@uJ%
zG9jOqwKrv~)<Qq3msp~1&3Y4SEQrp{kG1E>@w=gQpMts@g!&#nS}|VW3bSS-Pa*g%
zW6|M`XWV&G6|ovQ?h+AbBhqs(Lx=a2XCC5VbPaz+-B_b&PJjUu3tP6&ci=@Dg?)Fs
zQJZwwM>_l=AB~fvekrr>tD5S%pm_n0q|pdpOC`8@4DI96?kbxZWHj*x)il%p(<Ai4
z{|8|5!RJHxAG$V7`|+)tpWuroO!5x(kPa-iaEF;W8&NlTqB2BMp|bKwgU^TY+nXz?
zm>0*olYpCU2ZTNO`e0v^Io$R9@m!_e@#H?RzAvSf-;_`j(|F|j(&5w1wV{eZdT^T^
zbm76k8dG%qR9i}Pc{=t!yn7;!lFJD=y-?O)wv~K&iOgs{S8O{oQF4AKB39N9!Cc(j
zN<z{89BI8oQ3^eWZSSY48DX<iKTO)(`-Q&M2w{Z1GT6hx4wI!(E+I}sB{`J0EI~)-
z=2qj}R|k^61wT7zRMp|;Whx~eq%;ryZWj{HWH*#wwA@W;PFXsTKLKU71X9q^rK1T=
zW^qULzZFscrcM_5)M4esdKOJ;ViQM6%V;!<_PvS>^K^s1*rtQMp=){1>W^hTL~}kS
zL(a5oa7};NGnIEEOLp_KLPg&m4P5X-cpXQIzoqfp%1RfZkJj>)sKr>czDcz1;JWdf
z^>=>luev3IDb*$vWS~k%=_UgE7FO@fG4&}3udzH|gKB5+!Af2Cb4l@zu0L-zFi<3R
z*ke35q3l<VNhdU?9Rb(h-qBxAs2ev;V~SAR*#kVv>*?$7Z2#JF-yQ;%li40KIHUa0
zC+pWG+5gsa(>yK1^ixV!3S)0{61t(eoa*2)Cgn6Y5(<c?%Ip|R-u2bG8x+<YF8HS=
zUZ;P6Jw;p;z@xytQ6Sof*nM1y(9VL#3G5I)h&D+NN=(_AmDZ)As``+iS`{BFUEtRH
z`x|qFaG-0aEe%b;jwNAWhm24IlXki8_CtfirDD)^;5{$m9xzgx53-)7BmPIq)hAz>
zUZ?oS{(Fq$eBMnVs`aNSwEJ~vP{Gpy?82N5K*h3m^Onq0_B+V(eXqnLoDVh#uW!sA
zox(UO29w+y@s3q)J4l;Q!(Br+(R=R}HSvt~mrds?^9p68?D86!2r-`pb?lp#>5td+
z&ds)fPo+emKsrjKJ=++ogU`u~ej=^915pxiH_Q90PhJM+K4!+=Dzh_4f7iOy>hPQa
ziHJIXPTj(+lScJ<QBg_WtE($y@K&GhY<2z9{yG1-8fJ!4A5wU;Uoo{4`ilJ`RY|DW
zjhUHPRYmN1x|*}RWhVBhvO!jZU0Dh#%MTE7rrRjT+<izIiTlz7qpnc(&00!yR_sHx
z(2?-tU8_l%vQ%Q?2Z?gK^0uW}4i=KgVgX<5{;4UahGHGFR9Cd|F9q0Ni7&#i*C=PR
z>;ix|a%v8dv#)xHI<ZTb9qlDJb*j^f^tiLBN}Rd$qo<sOg>!2wD!3M>7)X*Mn3K7c
zbf-sl2bi;l{X;68-wkutA$y|CxuZ&AuYO#73%GNoo_=GAD?>iKTz4cSoEujfHQS>b
zd1d6ypj@Iu+*cph9l0c0T+B+{&!ujNAJGl5Doahfp&(7Qtt}YDF4IffAp{`+!<E@h
z+kF>XN#kVOx4CTsl7hUe8v5gZLROl6<%V~ng_cZ_P6>O%Bcsu!S3(%J?VhNuqJ1ot
zn#R(5YkBj#v=~dn(!jPeyA9<tcF9DxTFC|J>NBW|*DF)<`Jii5QCmCNiEw|S+w9Sj
zHT}cJe*R;iYrB4mycSNQrEI`+WUUE)#6-OAyoin8UikSw#5Y$?TS|?eN?_*OO)$8X
zalXnt{VWJU_&Wf8uy?yd@HhD7u=$5BQ~u*`q{S6=Md^Ot7<%lAw}ob)kv!ZY3FbTB
zqpRuScLkxW%u~&{I`i?E+5oWj2G3li!SSJAAj<pW$pCQN$!`z9Q7TX?Ygv13byNA!
z5s;dEOb0y~zO9e)^Xl-Ybi4siUo4$wQQh8D7lZiB-pjsd=&$fE46hHk{<D0!rf(+t
zWPWHU!w!v`du)ge0HabaPEHytDK;xpX)qzSb*7PY(;+1#wE}^(vqnE-hhN^u#5}*$
z;i$B?cT#5kmZVWyR2k*n<`*=O5;0ld_tsarSb^D`JYC689;Q&dY?5SO{4&wNi1C8I
zp|^v&W#AM4ZT8d;ZoAGb;yC{iH>I{hjl&WuHKs^`aC+PFcS3|#*q@n_=rR*5m`ndi
zQg+%3k~Rp(skj*uM-{!Yv7E9MQg`4;n`#$1nsS~|n^xf#?ntYs;xSsIp9vnInv>;B
z)mMMSDKX(tZqY$Hol0+X`M%B%-SK0~UuIR6tQe)t78_zmFcQbJAvG_%HGy5$5aB-h
z3DtvFYU-%Z8}^N=h~g3>UKnPylzEz@YD)v6ey1$S@s;{k@@D31;aSDl!hd1}z3F*|
zMGnOLV`|J%Y$)wro}|(5*!s(_k^w)UehnL`Wmh2Ax>ToHw0P}s_zZd)l$a0S-Xm_h
z8F+fw@1T0LN(j4s5P~e_z#)&Pga~HHd}+UeFYRp};AI@Jw`venF|vwwGi^RtHHI<U
zQMatp!H5TyBmc~eC-lDDZ&j8a14xc{Opg@`UY#S7D_hXPx(KdzH}l<6zkiqqr0{m!
zq)`KWP3U$hvLzSWjOwS0$Q<n&Ug}qyB?6$KXGqgS>3V9msLLrul*l58ulPx4Qmx&y
z8J`~rLh~xrTh~v<F4JvwmZ9uKtxH$^*H})gn_mP28Lv%{)X?$S*+84*xq>>@Pk5f7
z#t>R?@^UMj{+6NHozftKU;w+8j~pck*29H)`-n+pLTMo_Dmt;T>Fw1;2kyleyW8@*
z+oIU+G!1<bdU+$(?EY{N@1Kx=Vu@rO$<-ZWTQkcbPNqeGj}tB<CzPXE^B{O$ZMB2&
zY}vHU`(V{*y+v9k-Oi^>dTwVp{yB4Q<gY#-GmShkF5b@boE$L`_6vEPq_X;cQzO%)
zb=WEb<v#xT3<7#wQGX~b?)iY^O@VqUsnOV~1A_xX@J3xuD?5=x0ZCD(p6bcT_89tH
zHI1{YXR&M?W%Wz3G)cN3mAa?{%p;x3cdl<|ZRCgQQ)Gd*@_%S?)vHTLpJ!<?jxI&l
z8xyGj?fEN`i+2tzP!_XAo>`qWQr6bi7r5y;Eyf)WH7k)neYrAtq^CV<v^dwM!upmt
zj<wjV(&(;PePkh3uaMPgln|p#n{ekcTHPjc&p?Vm#?W*`zrGABG3d8nm=tp|wledT
z4tZ_@QVJUt3G>SuoC`sAY#*dzj#{S3`SYHh2uRy>Oj~8ybZhBAj;8EVDyp77!x}u#
zsZ{L{;8ex+C^1YQi?_Q^&a#%vN_{bEqqzAi2AlNANJK4@JWd8x?PzEi|B1djNu7e_
zQJCj4jk?@&M3#t<4w9c(YiVlVBjCwFj0k8xOx4iP>_&hbPYHL+Kg)aOr?JcE7yNfi
zoMDv5-3DUNt(<GA+jUj9NH2T%FOr{Y9>Z;Pj$<0Pgtts{W0fa<fimR!>2L{8&`Kv_
z;h)n^In<CBw!tze7^q*#8^`iVqEFC<#K(mg0-2}9eg9m3xty&Iq(1{gkR~Y)Sz<S-
zgp01Xg=DM}qUD^N?mNSKmux|of9=ZNJ4+1?B_iqs1jx3O`;u>%z0{nXlHDL5@oh<0
zADK4!!s}9Uqx0rszo-0XyEqfR6|*(APH%>1qrF5%bSmd6wj%FK$kIt0rj?vzkox~q
zLjN^MLb^j_199a`YY-hq&r7^4iSCs3t~KliM~m6L_3*(?^BOeNIsZ&W@(F;&2Kjbe
zV0xP5tqbxz!_l7}$K80T;euAL8XgGmvfHVMJ(c9BMq|e+*Sz1(ClNAtphJKisD_jT
zL-nUM*(<GZA_bZoTAJ|RvS*1z!xqw%wo@28GiAw_yV=zv+kF`y-tsG#8S|m2hqHzl
z8mN}8)?ehNroLm3i5WT^u9X*tX1jDUw3wKYlBOda0z(T68DT<S-OISj``?uvJxWj6
z&2diIH#y>T1pswVHbZVIrWW0lcjK7Q*?HCyvgZL!!mY@z2+skOL7gs3g{VHo-*?hK
zfUGAoi}q03JYSk<?Rom1`llCBVccdF)!}7S>7)g#x=AT%+?hivUdI`BNep9pg_c{-
zA_Y@_Df4BeI+t@5r=oKPO$_iAyqFXDN-H1UV6U7;M&|f1petcD0<s9KXe@bGo~}-U
zdlSK*WJ~>)hxEgTk19Rbn%SfE&1xIPk0Rl~-Z|(G-5<n*OCOXqC$O{>Y!k!3VRM&e
zJ3kvAEy;`jmgXxn?l~P6c2Y9%wwx)WBzwQ|8EK+CM>UY)?kSO<uuw%~Jk}iHwA4oj
zqun1LTh~zh^$Q6jZhUFOKN}kDmQ;_pAm&_t(Lv#mn`X`n`JR&GF#3wn_>6|!St=vU
z(X>X<YPwH9Y45_t@TQIZQ04540jEjbgX3PXi~_Z~;u8aZybg&o->EBHK(9%!aR;er
z^p)LO^(`Z2<tyl3C)jb!i3uP^I|IsDEqE~EjX{Mt-nsSg9dG%CX;tWzSK37{3jc5(
zlfoVv?Utwub0o>&y+@iJhDAFs@O{rZAluiO(+||s>BfN>e}wBQb1A+zrANOZ@d1uj
z1Ma60topCozU86>4kZU-4iE}tRy?l{jXoeiO6m<)dNocUX5JRZwcT_A2*&FP?Pz<<
z=iWLW{aoe3MaUkc=D6pnUUdqTJ77LD4*|k$KZ;$j=ygo;bss=ia(?sj46htl{2m=e
zdN)!<t&n>n<Hmn_nb74|*s<VwPwW$R$6SkH>}rb~ku*uKQ=STdR3jh3-l<vN>d)RY
z3?kCUP&9M6>k)rUK3n&WN9pt;L8D}R)iGw`$@lV7L%`!#Y3IJ=nOx3Cz0HNQn;p#l
z#?7w%P6TZ@VT^j*Q{0BL#;+&LB9#_A`E?ED!#lN|=DbSf@6wTlNz(kfk233k;!P)A
z5(!(}$i<c-OKQw=y-)FGG?EZ3FW`p}0dGj+`&mfimoa}3U0PT)nw2YYktRk(xlN~q
z7%NXll>^kSFoXJu(KOg8Xt<T?v{Q?V(Mg7wlkKA9NsLmE<kykob>Lb`n)3+sw{+Bi
z&{?>fa?v1*ZI%vEz`Oc{38O=$(su$f$-k7Y$z_|V=BrgI^^$rc2pt}L1M}WAXh@Xg
zqMv+UJn*kI;u^u_4Kj?Q0dN8X+bb)$GKknx$5aFolavJ5O>J}EMeCTD@&oY(#9vp&
zc~(h?-26ElPO=;On)ekpV{iGHN$uTSl1j1T9#&8H2aZxJWfcy=ZzUE-S;|CKt{CH8
zSs#};DP7+7{ER^hA#!b~9}<pj0*(W7T(W6(j;t!5ea~01TPtxt{&=Xvso!8`MBH;l
z5sr)I6kD%rbd@k{u<E&-q4F%GK7+Q+vrmX{wY8B+^EJ-K;J{rEkp#w09T=7vaC=b7
zI-_RamIdNt0>F^giAxpjao;zf+pF>I?(2#);VLKA<5-w2ZSPmJ_1EuZDB6X0gcMXU
zt>%#(>Donyv!{SVG<=~CrSs$GwN`G912vFHwhjPZ_LxioxS}o4Wwol6Hsx%PovP<p
zo`&Ac=94OFpcOe*e*ymP9DXSMZ-CQg!K)hu`dgdQ7bF7D33+jUqXxh?(eGo|@9^*U
zspzt``ut_)Fthy*Y;ABk)`&o>ogT*U>v6bzCn91!Js)v0?)X)+7SL57&y$|aq!>SZ
zFkQNX_dB!B-PT@3U@^w)<NJ8LBp-+kw9DtDczQ0oUSfljJs{AU#zW2RyQtg8;HApO
zKo<1ivbL}9L3UU@^yjU<Tde`0ooMcw*7ol-=SHIvxqKbFN>~hi?;rg)j(R2e7&M=;
zeF{NpD8|==&;jIm%%gGuN3*+|1s$*keQfQISr3f&5hRJ2>+oswp&45pr<V+B{b$eE
zgr8J>k=grD{i)VD1zbAG^SO+-QNPyob(fViU%e!3aBy(#H;2!-^HHD~b5YZ301pe&
zo0>Rp9we71329#edz-^|z<1of=NrM$!)lOm=gq}DORn;Pr049Cs}uo#&G*0|lk)i@
zT@K~))r6Y0-k)ibLJi*<G4vJ&{JE%11I|tAK1%=5(m!48C_ISxa5(Wy8YhDu?`iX-
ztj6JNtx1R9Y#&WwVIi?*5!&=+c%fT10nrvt>Qqu?qSezBEnUl~_M$6AA+#`gbyd06
zckJBbSS8|pOx9Lb%A|XHo^6)z(cf=v8PU3kWFgy1{M_@y!{4NugN$q1dnQT9KDZxd
zw7!=vw=2`55_t8h*a}yvp_T(Us>qcm#6W?O%5TMRNKrS5rpAdQIKwq^d6|?FJt9I*
zLlDkFjK4e+ard3fmyRksrDy2-Og4?N9A`G?(Ec3*wgP`Recoios=~-j|MNKV<q4zy
z>(5WsxlsJEYUB@h&?~`7b{TftB2^)i&+)vDzV4Q>t6q83whOUM?8?G?-}1AvB&sdm
z(UiP(bO>x(qN#kWANZ<`*=rx)(3U_a#8v;PafWk(*+G?2_$x?Y>Ibjg!w9Lq)>fsy
zeJvl@*6oG<m&<DS4LA|;V2;qu@-yK%@|}F>Y9?Iz*5=}`P8zhIZTnRS)fIUPY89jU
zr~rl`R%ZM%EKqYlLgEv}E=V$l?z{-H1Dv8@jg`4L)?ND?>%AbcqAKMqeH%d~@;f=N
z@aFXjk0uX3bsMOS<Wu*UT|`SINu)S8zo^xwAN?)Ny99gDjn<5EA1`3<3`%o9i!{oJ
zKdyKrjefWZt|ohbwi%|<q3&;BlISpBHB1PHj_(j6&Ixy}GyzQnXp6o3?0o@iHsg7z
zqU!;wJm#HTit)6GvY<YRs$XcjdrWIB3O36cEseMxWtOW|X3yum<st@2Z0*>7Hsz5u
zuB~h~m-&bAlSFpElYD<$n#)?BZ29x$moHyxYd*!TBzbA<ZO@GwUmon`XItRqBGV4T
z=hawmY~*2rMeFkr+&?MV0yXAZJ1aHeZntCc>vWAg8XJ%;v+tOHppdPeyCy!KS7c+0
zsOf78Vqmz~XJmtvkYm}D@HFT6jnCMxI=r>Y6UqtI=OKh02I_5?M*0UIk-V~EQkWxx
zP+jR&enRqK>Gn>#d0OG1=J0i>;r+=v{Qu$TEZCy#wlJ*pNJ{sBAm~Ro(gRA)kV*{Q
z4Bbdeqm&@sA<fX;-Q6J|(%l`N*YgA5x`;h%@3rpdUN*p`HzrE_P6CJw#<a=^n7W%T
zhza#ny^fP#5*EaO(nGs&*x8_~8ilg2K*-$Y>dUm2k2-OmKW3=WM|)l$^>;p$9EGqx
zCShK`{-HS3u+1UQ7(ae}{eh@Ekgru56{Y6Wy#=K|Hx{g%mN;*-i}>ped2QkGU-Ie^
zk{9sOpJ<X?DKEGP1GZG&59ve?4vIFM-7@Dn#E+Sq0PV_Tg!^f30xwQ=|NH-1cCjna
zP+7m*I+&_3G3EW_bk#gho)|okR~bS<Ruv8~a0`e71I{q<cx*C`{t09ApR*R8E&<!f
zS-#oFPG^4ttXZm0EK0>$w=*IPqTvXWO07c6N0Jvr@L&~WSP7^oIU-Lir7Q1}<Xe>1
ziqN5Hgi_b8`}nYho^f(QIVeKe_U07LhmGyL`Xh@1;hYmqCvl_)EnTQpEuC|KeTlJE
zIV1qrZNR$Lpi|sSC;UA8hRY|Q5Ua`WJE&U{;&nh6ZNXi<3mL+ppD4AbjlU#tr<`C5
zOa>n0Mba!pHu%h|vsCA3X4kgM@%sXH1Nsj=`U7qEX@BYBEY<w|$~}VYl`@X6G<PGe
zEt(TiULK;1EKKc5>Z6D`gI$*WsP1^-*Y+Q(Lj3Yw=J%??#(rEb3_X51%xf!As?sgd
zI4+lIO+CGLvTX-7*y9HQ@fJl+i~w)*DCM!yBqkG!U$FKjkx>yhz1GX2)##|UTi84I
zz3QKY4#ypVm|A2pJMk`9boKSFm){;X^`C;g{^CC=mwH_#xg__my!uHqpCc35Ut!$0
zIh-_g_Ha(N@zAghTgRr<+bLPVG<_y8C!<K;$~{HJ7A}+|@OgrY3%Vp!i}f4fo<ui1
z!mwP`2X2)-3pCebBnXA{w5b#?65}ci4J|rk#HR?@_TwJvEH}FV^lX#ee!xq~exAr>
z4>yolH%JS!`O6(3wAkk@`W3cMw-$GP9KPDNx?27jNFyl$czM^IgF_>-D=4Jpyh$q+
zk8G~|sYCt2R@ztVibz-szpGo+dfCQYEF+M4q|e8K<;}W}?bOo0x6EPD_Q;vVZX%hI
z#1}nsvnqxBKaK}C^zNHI3XZ#kyh(Tp{0Tqo%h@1^WDHe7!g%=xqjfHV=?zE3+l?wh
zrS8KP;mk3TZi*LZrmn8)TiaVRO0z3}i)Sjo>gVM}%KA8~4#pMzYH*rDqtsNgGINzC
zB<Ky~<?}W#`}ERRPg=U~XI9_!(XZ&9pbsu{URH@;zXIW%z3T34u3a?An9)et|I)bV
z=U)WcC%0J$W?v<eau)Vnph*rm>xHG0>=JKIWjsoh>fXLTpNR(eWAVq0{L040SH6G2
z58g?ssnu&(!y@y$Xf_YeZvBHN2G6o_AB_HMt<D2Mz`Qvzg(Vy3#*W7;Hn&QDPge(X
z$=LhXr21-XQTSUJ9t`jLEptu$JJkC`SKsel7WFo<#;MA!Pk7K!C72Z(FX|DnCn<A0
zr`MJcV)J@HjyY)CvCttKw`7L)&3ZvuQ3T&K$xB9ZDaOyUgC?A&tYd$wu>3hl0c#o!
zgvx(7<ONU2_M<t)Sq^kYMu8_T)b?77ww-48@)2rM7*H#mPMK+`S4q*<KC0)j73Wj7
zmOhJ#dMZmAtc44K<JfAiB=1Z9+k>3z@X^1{E5WgPgfKuEv!pk!Mly!}-nBhB-Hx?g
z%kT%8Fp}1A8+gb*PVk4^$y$h245QBq?GH|O7o>`M!p}wM!@`Ens`PcDZ+kx3tgeL2
zu28yA-Y&kbtgXT*jY8^$t-O0M()n^dCiAZUazej0%}Wm_ppC7m*#NsK;w<<YH6#qR
zH=?dnVrP@gAL9I&q6bizQnW4OC*S*w;>U540X7c}P5;QhtEQ;w3Z=~FPer2l+rW2G
z+uPf@nyl_m{poZGV?zoYEpLw7{(idM!vy!=AGdInH9r)&rZdBahDZfx@EArppt5$2
zwwCBsbKY1z<=@#3tOwCAjl&KW*{=*&u~82tY`}V-m-~K4_T<}OIABVVCe(qaM|CuX
z(^W<k;L<X^8^A}U;|>97?}(hL@XthXN_|{DDqJg%K!P)zGx4k1rswwq8%ZdVZ?H-^
zM|kq^J4yEw83{phmVZefiOegtO2XjfL5EzXyy1VzkCLEfBXl^k5)hG4{L?Nv(L^75
zay`9A>x);vqQslUe{nio+-n_^FI^6MSZ}S2sXx#<r~hQXTk;piAH|XKO}reo<$SYz
z$h)6gK@rf2!=1wQW|*#uq%F!g9V05G7JI$Zu5O{UULdS*d}t=}Q0rM(yqI>(*_C!#
zzl^@G{)*2%Zjy+j;%XFjjNG4hg#1~+yM@mXBhX~PWEhRpPeS9**F;3v8cnla*#dC@
zfe0iK4VkA%m5jo6xvC17cQ_a$Jk%&*C2yx<Eo&^{hBA@VG1eIS-XXc1LhJx|c7+pA
zZi+w*1oElC%n=R~y-})H!wUON-f+iTkes<A3LCQp7vxL)z}8Hu0SW_k9n4@lTA8tg
zWITFilLEBwYVF+J`-1M5Kb^UyqvZKMLcl=x&f-<Xi>c=KQ1XzPQaaiFYD>$Ax1<br
z)Jhf$G4VbvDpTBX&=e#bsd@WS{OHGg2o;)GA}we1pu8aAOk)s+X5Xx2hKTdW$psa;
zybG-iaGZtCK4bDpbs-8l^Ar!p8d0wVTGbjaY@TG|DKMhOf|kIXU{NpGtXQx16gND@
zA}@F8l(!BAsAX@g_=_hkCHG6Rx6!APXAV%;HCb~XBi?>x<k_OCyR|1bW@{`v?KVy}
zdqbD2I<Y4b4Yd;EgJC5$%A$JHrsBh>jD;qr?HA2{m#$T)Yd3_K6ZjyS>8o+E`Z)7q
zW2BL+;ty+4hXJNuqd0y#^a(L@dJT;Fk9!)`X{~Sb6*H#2EU6;U{_%e?cE0;k?#GXA
zn`hVwC5pDc>T@=C5kD=STZdzjnDgs*>oOd9P|W@tgFbF)nJcG>L;?}cnui;QN1t-7
z75g;7^}b&C%x_VY<QC4O<sd-6G>eW;jj!{u;=L*mtYi+!!;@o5n^44Jmu;E1??8kM
zvMV3#^#mull~*?YcYU-nqX?o6+CcmmA=|Lh2MCa3&($B-E1vqJY2J!VOiYZ0X{uhi
z0gB{Xd$lP{82V6&Tzm{3g3r4VAL|5X1XpLpRm*p4Fi`L9&xQ8Wm9gVjTFuQdB)LST
zna6%jw<(C@_>tSrF3R;}HvmKwqHi}G){`jDR5BH^ivlekb{msf3?y)D&-xwq9cItr
zyYu7E<`^i6uHUhf6)P*10Nej9R*?Z+y(0>1n;#7td_F{+x!ut!tCLRjCm=Yq@=g~+
zz0-#abOkNpw6!4vDNLvV>(jQm=smM()B19u*Aa$N^7W&#>&65{$m-}ZfB%YIdf??B
zDwAV@rd{*=k*UD|<L^t-y+I3a@1{Hnc?i5|6R93{vaWHud}IhKzNCt}$m;oCqBl67
zy36I2tE(x<4^36_{*_n_p~vc=<gmHVBwn+u`6f*g8z{C-XWe0s)B4bN@Yd=4{0a}R
zBewPXlFU$0XDFMv{q?ABgU@3lZ;A^A4F#7WNqRInPrpPLG{Y&&{v%KcO-+EqtGjHg
z{VKT^==~vgXO5~b9-E8%PI!zISl%0}z>1CcHR4!4P_eXo7=Mn$$N_=mbq9Og-zLPv
z3}z5>u<;*q`5WVF43u+$P#6t2tPH@rA#BEigma&3yfmBS_d)ym5POX`;o9z#Z$)y}
z@2qEU0#x|u3qKWCa*F4hjyD3R0_q*Aa(yLV1t(u-JiZ#L5e^#4yJ2WRA#q=|wI(W!
zG0}Lo{Oc(~IkVT~k%js=Y+VM>pY%wNnTLXa2rUb~B}(lMPN*hQ)P^04nJdUhFz;#)
zk6ulx%P2f27Xw9(ygMKlEWcUw(FzxrDpr;iF%(mfGuu#0gj6VD<L4GKMd^J1%2Z9E
zwh8EB?td}mg)UwNz@;}Z*ZD{+zVgxO6fb^^N;G=8O$;~FT*n#VPKP2|J|iFmRw#qd
zFjj@(+b^|rriEtC@8jI`658(fSL4cjFZkTok#p58dgD&~>@rc%0jshWMh?AJMm6At
zu#vTW=@mJ++*D^HDul)5H)nU&OW!iI7B&a_dunFg!{Hu#SR$Mr!trPEOmQY6ECD!V
zt<TsaLnJw5RKcnFqp5N0eM&pnzKY&!O5nMb_pNu|yET~7zhI}a$wHuPJ|1vz;s=+d
z#UarJG5^JvqCS6Oy-sTIH>kkgQhM&UrteB<WSm`>flK84T3~jSVQ0Y#7iUMp^7o0C
zyISkbU5bAXhOem0D-=`VaAM15o~Ywe`|#>=R5uIH-BMR#UCQL<I{Ym1Kz&0yS_h~a
zCwGhn&ws}2i-ok{nNi(eAOD7Z7(Rska{&>1MVr59qZ^>YnSO#O6`&iE1}cq8B+hVj
zJ7S>Pj{-&n{qI2?qHX-7Xj<$v3AJ<J%F#%hiMGq7+xw<R*HJLeLGUhBli(NMHn1<C
zD|rydD}88>^CVyg14Jv}*@9970;ES=M_)>_;){?4dFqr~7x%?soUTlJZN9GVswJGk
zF1o7!*F2>Zi2LHcnSRbfpEe&Fe(g$hJSJQ?+FpWPUx*FtdlO0hPJ}6OlHhm@$g=1(
zona^W*e_K(FCz0m<_G&v%I&;ObT(h#bB!2kP~u`K3t5uHn5kV~n$oE4=5pVqvW6#{
zk@EJ6-9PX&xF05OoKXNCfC4f6sruU`o|h=q9#u9>v|&~&oj~j!%1|<6E4OTB;`f`8
zwJ$PbeA{v*J%hj=-Cy@@T)e3}LVd5MvO9a+F_=_CMSUyZ6tUJCtFdkBVhUM3*5GX;
zDU2QB|3DuooH$G1UpTo}1dB@?f7m)>NIq^BD_&PEjvtm7M}3aJ@n(ocfk_z@LjNGY
zc;Pwj)9Ae&?>&rth_n`CzH=5Zgp0)U`sq<SST;t$Km6;nv+-WAjJJLHUH-9D;>Z|`
zD>bEna<j8^f6}zXu8^-%>ky^PnzN}Wo39%u-IBd6x4$~dgCS+`KKZg};X^ioWFCPZ
z2(#tVgZHH~(S?X~4%gRYWxUJi$>xn}9WJ75LmiB%R7egh@rae{8QhL5f{gO>o<Q21
zkY8;AnH~y3BfYl?Q|7J#yeoOD41rRw{QewiXtnuCdmGm~t9sUrz>v>(60$@%5EVWU
z#-SHG#T+HaOZYu(du&iZ{HaK%*<O08gKwecN>~W~y~-RoYu%xX>&@!{`+<Q{x>`Dk
z(MsZvLA;#iuYU$mI{bAx#IavOo(LK~+?;dv;|yh%3)|EV{s>+Tf4aFJ;lv@b{hr}-
zUhV#?`KIf(>l%FjuWb1J2Ho!redZT<wVlyM$E=iE4MFjGy$<5PculS(=IlgwIPQC_
z8US~d`qq;TRT|SY_ZOjwq!p$Id!6vz-mFJ$29fD~?+kSn^4EltY%7mt*wJBb6QjIE
z-6q>6veivBuLd1>O0qgzpW6KytvN4E<a^kg<zHz9ReQ8rkJ%1}k{s>qcsjJ>`U~(R
zi&HsI;yE)?R~Axo>aNev7k~^VqI;{c;<&dmuUd&;1%5}xL)SN1UB9CNyiZff{um?}
z$8@3o5`!F>owEqzOYF+jXG1WSP{CAw(m!s9cgByjho~Jl=hOHHza%Gp#OE8!+Z*1k
zlf~XXK5uI?9C77@iabgpA-i_OW1NU*WXIBZ89q%T5A~0brz@&AQoaiq3%F8k;=Vfj
zNrv`McU^c^)B%9stlz@w_%L%h%wSLw>L4JN9Dm(<7Vq_FWw-CdMvr}dp7#<hqBLpu
zUX`_U=N#R?G#Hsr+`;R2=ZZ%kk7Oq4c6MJ8`X!5wQcRsk4mD>?l278qnxm-h;G{{v
z{VKBov}9`|$<btb{#5o(kV0Q$Wsy~fo!muUHqc}F=V>-Y>|&pe(ngz5PCbj7_b~36
zAxVLkuZ)Xm45^mQ9KMU-tD*bV;P6<a({fgrSMhW0av>*5UL=5SXJAvBF(+?fr+?6L
zV$kRBlPFAqF?w{#!$7m?KjKJWkQM_|!SYj5IIqv%U1@SsuuQJAR=)@3sKxY?FcMp6
zr(-h!?|>5(7(<y)_0iyKZ3Wz0e}3Y;RjDUselN!GLR$KbCg0=8l0*q)sc9kn(PPn?
ztGLNkZ7B3S^Y(i7=37xl@U;98e~*dG7)vAScx{ZtN0jCGG&_q3$d_45$ZrkRrc4m`
zp|KGzK}7FmGb3ZJOUl4TH^RxVrGh+<b6QA3PXxw&Smm6OhX^jj?K(}Z1$cViN?Yyq
z-ed%ub{Mw*xGb>)LLap5X$YMOeTyqh7#n<<MOV;ypww?nh4N1Hn4mR#K)M<Rg~0|l
z6#%-MqlG<R5v;bKhkQ5NPHCv=#zs(ino^(6p^XV$dU1vhS<)4i5x*T^Rz(`__0UNN
zI`6OP5|Pt+SYZCDGR(oZe-IgPGMT>#;Zh2K+<sXo#u>C~qv+=Jx!c9(v95>Ud>gWJ
zjR8*KR4{Pn2*9-`xrNz%jHECN8LWMRpN!(F->bIA2CzAF;-y0HmJB4J;BG5_oR-9K
zSZ5a|4Qqk)ddIggoG`1=k<odh{F}wl1NNY3ko$EpxL*eBy*!!$YB@0_7)^66@L%i-
zc+rONz8uX>)37*chzwp<I7Ouro)Rk_l_E>^C9jgubSeis8?{i`jTP4QNH)z5909T}
zH5uG5zi(8r+=B?|S^Yi2U0O2_`LYik9la#af|^}3;j>942=inDNjyQI42lMMRyg|8
zTH8D;VIJ;m{|cCBK;^)FQ;nhFZ46yEa1>46(UW#I)zlJX5LtB4vHjc?coMmd^1Br9
zSc3FMVFL=ti;(9552CpIL-%z2=X2nEo>7~!@r=5bU~%6rN?*Z{)s?noEV7VQ>)`V#
z@Tgs?@Pg0Xe)6T)_(?j0Ao4SI-~Qc>th`?RXLTSZr7n$4;fYxqs{z{y+BRg@FE=;i
zPKu9<BWZBju4T;B%{`u)!ao`q{L|cOU5B;^Ppl+RyXeVruo51>v8+vIYTk3zzGIEp
zyLW+ej(}(HI^imbouxY6IAU%;;kt2}#Nb>L#%!N*e$8C<S^M%XA<dT{=}xB50kVrn
z5%E_eaM*dx8z-l!Nz0joclzWMb0T>#&tiCC{|FUO;c(x%9)vP?``I(+jy>W8Ix!GN
z<nm?E5Y>*1j11Hi&zV?1lmo<y-g@$+SFiQq@xlToFb|s*k%6_>)^p}RnL{;s#lG@2
z^rBh08jc_fw%+L{##SdQe_SuqFm-WP9NaV=F`ue#v42IUGq^2X6Hbp1u;f3~FRU&y
zFU&PmM?z&+0~IKl(X=}3B_2ayUl^Z5JFxatr|MNgG_@6)7}f3k=)%+OFX$!7?=9qX
z&m-||R2Q!SY<MN4Wefe0D0~hgxIrWJo9t^YnpK!#OS!Ggm}yhtlduOK6)Ah5Gs^g4
zSP|7MPq8*lJ548)ipW2cg&_J1f8(?2x$1*Z7DgJjg!KK$ErH<6i5R9E{}a3^O4npf
z=F(i#Lu6QTV0#H->eKB0_&2pS^8JZK!hz<zACp(=b>}1z!MuuI-PG~=4+mr(Po&ji
zND~x8mcef?nag}H%kB7gTDRMnh1UCMuJ0*&sXyX5E~E$c6($&V9(?q^?rYK$D!(kH
zlT3FwF%b6#@E2jXl~!dx`)&-%NjZuM#li0GSb;ZO+)NGD(!#HmRq*zt#;Bm}sEb>4
z(wOq#pEoENmOFKJx(VWc&%yoRrwQ;-#D1MeVf3<riM6WO(`?<wx}BF4yKW;R`XO~l
z&dm$xgs{GZ|DuXYj+^9M;mL_@#7;pXF?J{Oe7lm8tf~Hsr0+!jEKctee%AG9NZU_|
zQd@n#`6N?y-kJ=wS@yd<PU+Bpn(I=Bq3Scz(>D2D=S>E#*5)pdLZ_4qXH5`>|Bvqi
z4`lrT{l}T8&5nR~`@=+%{m;?edY1hKrp&1?PC3W_1S-r6`TZbzV~9;pcC~gC?u2<n
zE+{PUeeHVt!2aNh=CKct%(ZDpsV}1rf<SJs;vYO*VO8n5(_(7WrJOSE;;CJ=$oejW
zO=gALFl+E!tDs6v$YW`(>#ob@m!1s~Wp27co^?_9p;8^CI<3x%Vox`s@7%5mc=t(C
z))<pQ<#JWIZOlr_(AQpy(ZQIAn4kr-MTAiwF1;JrGM@gB@xyc}!BHkMCe9B9L4M*`
zZsqa(yx-_Dzi^^hX@AH80o;CNGbamhW)-r&AkY}@T6HfbmCQCT7Q*~%+Re>qclhbn
zjYH7Xj`>nZ-hNVtSwxS7%ZgC`nAvpK(o6mUN3z~vkuJ)QxB8D`s&R>Zw%2&NV)lLA
zprbP??yCYec}b;#6_UerA<u~I?~;3R1df#EJ?7I&&zSh^7(*pmWGef$dPtR>UB00@
z+ofQ0va!pNiXeuKm383j5A%(Ug(1SP+dtLela;U~K@I8imtz~qBeMMC{=G@mw{fID
z)$Imvrd94g3&#C7@2-{H#u4Up{WHrY<j$hjQni>hg;d?=A~cYJVql}MCcO`d6TG#6
z@cU{(n{YG$_zjR5!L!DSl|gb}Usw7s*zmq2^5U~1K_&!yd(rkzhfKzw4f+QolE;ZI
zZL2N|e6>?VOAtYH)}1$s+He<(Z-t+T)tHQ$(Iv|ZY-g0}-fqrl9frBQF1=)4{_bp}
zTVFa)X-Zj_V7@_SY9U`-`STE&?Jk!XOoOID5QEXEsaE<5V7<365pFi{0TV*F@v9F=
za|<LoDvhWHI?PT-HOc=<EgL?Do%aDV%Z8VWU14$d>GbbCc)<52s;z9zQeC>*>R!SV
zIc7nHY8vwV$toZG_maQgTeI29(RY>1E#S$iLue_r51G@47@NL%H71fMc{IGhbGCAS
zJTxbM2Lx#j>o)%YJ@mCs^oEPh*YAAS08yo{UL;tw6yFJ}`41@urR3et45q=j2?LE+
zQkEzQ7Go;GFmKyM7Re+gPo9Vu=VSZF-ht5{V2T8{RZk^AS9RPU#?fk_A7I#O^Xtvr
z?4BbktbTj(QO&J!$6yDQR$bIDA<dvlS)c{7{tH#GYw>7-l!Lzo@5R&b3eOFHA!<%K
z8~)<kL!)9JRx#{jijlaUwGYqSx)+8L+=DR4|D0f~#(WL*hTzfLa8uOb**Un(q_F&}
zm$bl77!@QcE|so4sO{D@mDw4hs95zvTL4K{U25>42I?GviH9PmdOIbkasC9J&Z3^?
zkzFY9g!<vsxr{xI%pYvpCkod3i|aqt&nfiXy7fk_maf_e+n?IKw)IU!Cgyi!dRfoZ
ze@-X48#*35s>S>opqY31Iezf2{-91s(nmLanH22NYxMkfa;a=U2JG`ekv64z;WOb|
z5hs~T8sF(z?asu(hhHV2Qlrz*ZsV8U?B*8>R4PqHpbqY&YOy-(80c=5rwMe~NVQmf
z8fljNiG1m6AK{(IRUq#8A4zsPV(bsuF+)*S9z>K%qh_D9^tAZc%7cUTb#*JAq?l0s
zWasm0iis1fbxQ)bFuAd@<_QbKp)_J|bBOw06T)PhW@_xoL*B&6pBdbx112I_RFn(}
zJ0vTTg2;c(ShGY8BO*)F=;IKV<ywHDh1v(3P&#Tn3yK=??AwSS`r%?N3Z5ht;C&j%
z=tQ(9Kwr#K<@omqH-ftJ19WpPwh^?S95eV6iiL(-96g~j@Uucpqr122+t9h=dG9tq
zY`uFUmaR>h9YSO)J+uzH4~cSDi7GHA`s|WL5?s@|=V7}vSrW&FBSs+e7Zat^tT>X$
z&O2K@hqY$#Tbw|!2UfOMqEJOq+YYLGvT}KVFDf%OMrVPdHVXX80Lv=Kkn#i6L2Dmj
z3Q^YJg+cD^-^O!^_mfrE2_chtdCSJ1y6SPN`$~+O2*O>U=Qzp6<=6(^oDogxdNECP
zwu`aBnqAgrck#SbAF!t;Yo9@A?)>_Nl3nw)csg{7wI=RfBr{R;64Q$Wavq(tWLVw9
zYh&Pxw$xaZ2ui=zi>mZxk0&=OUp<VZ0100$Jk<r)7aB6!d{L<A`Ws2+)?U!o`{m<-
z)m!1VSNeC8OO6jWAkWMwrGF^Vf{r?2MSWqZ%;j^E9?73;<zmB<cjGS>=I9|J2^X&d
z>$|58{C*cB)0i#3Kt44Xr<LUPyw>{I#qrFQv|!2iHqC8gY4tI(`Ql6=hK0BpkI*A?
zQMi(4rV04uQpFw3N+VegX#$o1{^DKr(O4hPy@`~%SH}hOkx2L~P|YLeP-LAW6uD5?
zRExhP^Tw_Z#Y4dH5Sry=?7A`=Xk>|rtWN(0gOtD(y>@?YcEU|}zWfQRJw0`x3hk;6
zxfvWv8A}CDA2HylM-5<p+C?)=<;3oMBhZqmpSVe8ahp%=3<${ydO_6$;6ERP$ot!8
zL1n3@0o!f)%ZG=kf3Z`Rgv3+UI`7t>J@ze)xOTVvxOKXi`zu{9<VnpF3VEuXveuv+
zdl4{rG^}T;tQW%9G?00bF>Q_B_7@pUL+ZP7m9sEp9jjkwcXbz^BHy?1CukcG_vHZn
z3_uk&G=CGzs7B#Ub3BM#s@sqef5`gP<ffg`N1yUFf1w~Hn5;HJT{i>U2!%S)nTe&A
zG`>}M-Fqnz_l~-+QTg*!AM1Y+9=hwch{Z*_eoJ3N6X-oU`XRmq%Ueb_U?6xujxw~(
zOUr@ti<8Yz8+;BBu#Wq8+m72;xc?FtK&L!!?C+NiY=#)<I_j^V)x6Mr<XW}e^A3$%
z)wn-{G2Y`ZY|1rp;O1N@Dp-r3p2Ozie}_t)pcvxuMmMUKCFH{2dUUeHg4D>1R$Z%%
zr)b!R$L0l}@8K*O8p5bL3k)*9PRc6-|EDTNn=vJpX;~R4u$<gY5~AL$)c|7){ZGZ8
z3S$#CJG`OSnpIDPW`fBX83|SzW%Cyy1VX(;a-g6Z{9p$jSktF6xg4!%_Q0?3GC)MI
zPp-yLv?`b<NF0H=?-|6aL<1E2nVYb^_(>tzBH`eo8d2vZS9bbmk_n(Ia<Jc)UR}i4
zw;E~Al`L0?h3~?}w;iJD+@+!-<36ggsaB(DG}uegVA8}|UaXaAaC|G4W$XAA!EAib
z4lE>;l;W$DGR6qE0xaUL)5F@YHjuASlegfoy1=F4O%)G*32vmtU&6zZg3GTC)K~&!
zTV|$;%TigPCN;Dwc;Sx**z}s3OJ{@f6J{3`t)}f%=M|NRpm#O`GJjpFX}j*UqYOKR
z#RG<=kEj}MNB-p;`MFKZfyECZ=hVUuI!`OB_qgttGjr8Ha%iZ$$FmsxIn+<uJ5b*g
zFqPmJ6o^g3<}+>Sy=9ZcMeq)#gG9ZF|GdM7goq_ArPYFHGi{?fa9OcDj*n3H`@g3>
z|L=HvO#ge7*pp`K*{Ftzm$i<t=86VHP-f?IX23qpl62I#Lj_G4$2zsnYVWV<C{U+I
zNNw>p#WrP=>om_Iy?`f=$I<1Oa$gNS`*K|eX9VUUCL1XFl^#~NwH3ypvfKE|1wIl7
z@jHx_Xp}y@F$=b6)VQcJViiK|6qVD$Hq4Qz53ECyP(E>a2FbwgWMuC5|1tooG&u2c
zl6;F*2()e5T>M$<?<26c(X>LxW{r1<j26!8>s&Z8@iVs59d54Yx_Zzb9E5|v8WTJ{
z;D%xIgfFmR<UZD$36RGg*(M$uH2rzke=&D&n`E8@xqQxp)%Ga;)c=H$Phw8$?>}id
zF?-Z0%r>juxd`of6(3ViUuU~AeZ3r-D(KqD+djCN4TweNd3_0>E{inUV?(^(FSNq8
zIqt|N60cr!yC0%<=T2B8pA@Q>EzIMJIYPj%O0~VZo12^80L<m^7;Gr&@0|>{urM3Y
z883PKmf90H>6}A}f2PA;)*s+mXNquuY;BuwRw?<G9v|oqSgfKRpN%v5t83=#PNBF$
zcLH*`%5BNxr@s6inoZbqoo>zE6@c9YURElDVYo|{tO_-iZ;DyM8ds;w4M8$L_9$Z%
zi;a@C07DJXV$~a>jkXptcE@riATvYm%In;Vk#iKP%kzt&+c06xTGk}T7MmV|Kg{S$
z4B`42!!pwXTL}Y$3(ArjFqtkoiGA}%Xz}FuGvYe?6G3Dz^*K?Vf1KmZ;1o5^pW?P*
zNnL<4S#`37>t{x;1@Ge1ms|1g^9TEt1*(`~O_o<<%|1%D(1{W~6;SzP@ded%LuK=}
z$^GA0RLSKVCVzTnm5466a;K9OdZ$(wIcuXmCX=gL>qG`gm(pB?Ek&d!n6<D+z!%O?
zA!*FcfxTp^wg@=zXRwCxLwzZfNEqY%B$PKg0uVN+VetILD_-BtZ?lGIk0U>;;WnbB
zaBrhLTJ5XsASa^uDf7|A7#7(-9Rc@?qg)ZM%wDu(Uca;>01b3q*k5;y46ENIk8*T2
zAbB49w7K@V`t8~%W>@K|R>?mGFZ2EzBy*Y(0Bwm#3XFfE0^HWjhJv&NeDQ}yYEh42
zq4$TjOCM{NJa8d5K^$Ty?I1(8H!i*w9%LzxyE$Zi1Rl>tI5X2rv{Rmx!InF`)j~Fh
zoFkUzCP|&4H1u@9oH$PZ?y7FH<$5KKp)Ydb*K$2gcI}+)cR>xa;@*YbXqOz1d)Fi9
zzRrcRy8O{}swWQ=_PV2ltnI~L4nVV|kFu@@U7Rw+rrG%<OuPgYqXeXdA;L<_mtK`X
zgzXnjhgWm50N)n;`n+}JH?1BaDH&-?kRgbE!$6{AjuvyQpa#O^jdvfUNv;XF79HUl
zoIW-sKI?&Sto1-G#qXPenzBUNUqnho?@xdILizmE`1}jRns`;(_DN9Sn{U(UT$i*N
zi_i5p`V^jm6$@Iv`y|Nbf^Ag_j{K?}b-*v$x$i9~f%xOk;6Q}fZs;O_;w`dNR&YGF
z@bqOzwFd#zz${%YUva<b?@Q0wG9W4E>`LI%QzQP{`^&Ew%GoYGJ?;*o+1YP>RfGyA
zj3tj4$Lkxl?`y^g!ajH!KVPY$STrW6gkXdGf2IhOC1vi^%<0XEK?nczR@9Z&*)AnA
zm9aFi-n}MRajtHQ;R<k|<E;Q>E9d9<!?<O~ah1$yz~pe|;s?Lj$LY$IRu|Y5o9APJ
zvTQQ&g}8=LmAvESnB1Osa<A3$bc1TKA&M>P{3EOuv_ZU%Q-NW@u%Q?GN`=fs_HxSW
zR=hB}peMjR(f(O=-A<eQ{tG|hEN{zO&|wz?8%OnFQO!4<Cf5(=>TyK*qx}O!%&;#3
zcU`=Z(Xzj4HHn7MwBdSKHRWDt@sD2)n0~NnisQOq(r<QlhR<-6N^CGk%UX#LR70?_
zlP&i&hVTaP_g=dD8;pv*m${V^JuJuLZMR%e-^WgN>39=p^cFWvj9yx68u_W_JS-}V
zD?Z?ya8eo_`7V4Z3-<XlLgEejI%3IvidkJDthSc|P0i3o$xm55`&31A&iyO1QQlh&
z`m9Zb^$2Fi7>;lSCXMmC4oU=UL~3X-v}~)QRYypGiRcIO#?2|<agLvQs)%nvg;geR
zaov=+g?`&xJJP0FZVNs%Z@kac5C}Mqc)aqd`hg>FM}E#|SZ7K7*msZXn}m_nXDYnD
zXP#IOEJzyTHS{J=AIF;-;%IwiY`pIAVF$I5dD#BHz7>u3tQu@fSdS*I^+6r8?5Xwd
zpZMhG+5DC#^&pF1p>-1lO9=|a#jnp5YNM=eKdyM(51lErU9h=7-EO}o43Teh!?R9f
zUCtgu4f06YQy!ygln}Ja6E&3(*>oYWr?lah?0!!6SRpE?%}i#4E+)!y&leS%mD41u
zy__gs4`8Eu&E@Y?7YE8k1qD&9majUsL>hImXRu!=IR7GeyZE2b54l5}?b9AOcPY&v
zdXk2*J6+)SM|^bsCam<V9qf+Zw|~gY-i@&0gbPCArz(LRFt<?^W~70d^KPB2`O)aW
zoY5q72~^qKnkevl?N0si^VmiRzEIuq1Kx|OA2;W_N=tUfW5<f?N^T6PZOfdVNs!y+
zZwrbTJ#C+)uJ{8+7dU&@R#$;(yN}L!y@=er;m+9hdUF0i=EDIAQ1O-)YwM!i&W$P<
zua8#~q;mknw>TSzj>pQ~3j$t4kH689dfsliaJGTT$$x;f1^~-3M=L`olTr;VAzshm
z|HUD}e!Dkx)t#6lX?U|fXR{gdOf?)tpcDzrsrxks3#e&<F5cG<44)_Xv+!qU#N6*n
zdY;qcJxB&%?gwbCzS)9>>6FD;Q}_=ING-Nea6t&rBy%0*%<{pB!Iagb@WWD29-W1M
zu(DcILPEmk-Y00%XPLodyI2yJ=mDe0%EaMeXtB2`><11CtCCbn)@|U`N40Yl!xwid
zTqAC}-!?1XSL?^kh-u_geQ#L61gK<lNg-hHaQgP6h}eyE8e|F2C2m4Qavju#GoD#I
zv(J=DMU@dt#5}4=W4v4@-;izkGJ>V_k|eHn3$&kBv&4t|5xUi8ObI0~leKtv`BAp%
z(~J`j3;snava}@e1oM=%H=mK3W1bbNXEolMA1_>TiV?5~4)319RU@&?tHuwaNTeeU
zC?7!6BH$1D&XgzJ&HI&LLxk819R-F9{$apq>_T|J&qs|GZNnmPXFBm&5{N0}QA|`?
zcv{r)0TMELEi4L^Rb7@cM1a<>MH$?DMscD47Gi6kLwgEt``kyjdl}LF^hL^Scvx%X
zNB)HOOu@<eS?=oKi*&PvDX~Kd`msYCI@N#ln_OW_b|01luCl7<XlYw&$MFn~*K}p1
z0&EK~AR-dR7e3w-_0A81O%PkppeK=|*wZOr=oQAlu{Na$c@(wi9;lV67b#_O?bBko
zHiR#tLocX;eN6x5KJjRM&FHE<wZ&M|{pmLbeW0F+XqJ|YC+bZly4q_aJcP4lCS0jD
zYMbL9lXJgxHmSfNrzq@tGue+8=kZE}jl)@dy!d1^DvjkjOffn_d^dUyt2qKx-fS$}
zI*#PRx37-44*6S?IjXp8u`SnRv}JJmCOz!h%0_A_$jTgxY@Cyy6obV?V`^qLkcK0w
zrCHOxL}&5kq<B@`$Hv)PMW)kF-m1dgE?Sy8)BJTswtxZVIP3R5SH;seF`~y_j%j~_
zA%uy$K8pBsuVb<%Q1di%?y`>~*X9r=^t8-$Qj~gs_tvWYIob6e1%;W%DNXo08|-3x
z75y(#PZP+1mdleI)R!_~xT-f(h<sM$ZNMKdovH-#*tQmv^gzhzVZ6I)Ekk6>sL$J+
zzs*8b+6QKE?c6rS2X~kI`9||P(Wz-(Vch{TZosxcwz&Zs9!C|w^C?}DuVG~w!v3d+
z`ZJmB^>j;;l%~ar%+8TfN~Fty|7VrK$iiT}&IAno)T{KO9UUNMc|ebGubD>bCP2%F
z(}lkB*>n5@5u5VRU@KYfxJj|G`YU%}OyJo8M$;#loQE4^C4wBp{vOzwdffR1D-?R`
zao;+~?bX5yWf7(WD`QTUn7D<TaUa5j0ZJ<=^KIVlERYAqh839d5PYLa7P3y(BIsXe
z)VDmTs8NrWn1U8}^`^*gW^WZ*cne~gHIBcTtXErP?wbBiPz_UBO}K5K_-X|2Qf}9#
zf^`|e{<8>(lmAf6Vv&U|1;0{PfLj8FPii6O_joj7W}9|=OL@`L2sao5j%&YKA+TrQ
zoJXC#B+72ljX-ku7B|G75YIJNT{|u*={~OoGon)S71Cj&CjA8*-vNnfp$}Oq%)v!n
zr&C!r7=$E-CP$`&k)9D%Q<llW$(QEFiHn?B?!^l3E|9Le*VHhpFn>DSp$J~jwF<L@
zTB&(Nnz0Gi-|?HuXfmvo#^)*D2&{_uX4+ra5K<xp_32a02)A*jz_0YGbq}=}#u`7Q
z0b%!QFD+1@aXgb)b5u5%TGCg0y_|oc;vrdCaV$hua_OL8!wCei>7N>$u3NWIC7VO$
zK!K(!q^1&+vP(^_7X@{fJIDBChO?v|BPHIft)z1XtpplTH6pm2m~jx(poM-LQ7PB|
zUp+^L%39c0@8k5N$@D&VkS11lkdy!VJiWZ^brPoC2BI{Ws}f+NdK6eK=Vrd??gH42
zH8c{7FZRlxLOgu(m*TTs$IZ;mkZK=lpCecQ<rI8&6MzN~_+2(lQ@I=IG_CI<-%c7)
zRpt3}Bu6TLeUiQ0SlX?Mf-p-I^+l&4CJop5{=4>3D;5vyzvaj1do?f3t+X}Q+Fv<E
z`=^g%H&M3r6yKf^N{5i@-+fLWfn6IOpg-%guv?@!uO_VV41quoiH*l!>NhSo`DzG<
zE^;vdg%THCAGQ6w9aR1=GB?%l&p8OHOx=njvVhnJ@iCPL(G4{SKaUr948MNO^T!Es
z6G3GCE`ZF&n#M<f`6uuMyzjboSWBZY*%s)+3b+H9h8J^ocUoh%O~ZP|%Cz?*id9sI
z1B}PZE8Y=3bLI39zWeCM#B0Oz{lTjsZlgN4o1Xd$@?)))bUy(WJAL2#=JoP1Oa1en
zyLtV2;qpI-_MP)3sBHGy_l-)5{SW|m80`Gx-x|4p|GDXPF$)16n|w>KOF}3A@~?9;
zj-j3T`sx2Pw*fyWs~JQt$xOXahl?Im-Ce5JV6^WxLp6S3X*ur?3N5$&gjA1zm4|gu
zi#|0K)TMHWj}RZcdk;cPYpWK{PP65C7nYa+y_iuFI`5c2dokV*$pE{}FTm8$S<k4d
zMMM{;@Oo-~&l#n%vhh`&tq$LFmFGS&S_7LhCXSA(_a;Q#?mMA~pdjAE)ZVC8(!s;t
zSWc$rxExqjgSpv2957|PXvR|07uQN2@;&J6bFzjJ*r1D}7rz|OL!qk<4x6eUv{;4+
z2K$hU4ePRz=4QKu;y0)*;s;jn_3^$64=}`KiKPKkFY$i!;9K*ka;cDfj-30piPW1D
z>Z3G;`)o8!UG8Bie0k5{aOf6^(A|8EhWnprNWVtD*3^03t7UbV(sMLP`Uu*6Ysi<G
zc*Umwa7DLIumz^ERYN2|Qw**ZqZ-AwE<rWz2a?CLHP(s#1eC@EUTf<nqUv=9jD?r6
z(>HXoUDaRNBCv<j46Ms-HX-+xHFQ5MwP-ZqD1Py^zwl;pi!Hd5HIUj<i9-NJlRn#M
zy_9SOaq0RPMvm1<>7;^mDha`O&NhmjxBlJ}J1BDoJ<CKU5B%##@m#3_33!+xgFi8M
zAUhAN$7yLAFDh&YHW;JHN_CA+`Y=WsKN@gUUy)bSKL>32id}L)r4}`wNDSc7@5UCM
zKcxFJC|u&6$|-F4faB<^Gc`sh-jxWk-oO+7_lN|wZoPZ_<#d*I?4u-hw?oCv$7hQA
z-lK3wbY-9Vx*M;>9F-OY>ucBm^`T+rGMi}d=|QS~S1V~ph3mX?7$!Gu^GVnTp@|NO
zd}RF#64zshBrOqg`UkVp5T^cpa`R<w9a8GW$?q%P*6K-@RombliG6Sp^>&XZg@2Q6
z%Ma^X!L^lz<J<PHRmRJ&SY1bO`LW9WT7Fe5O;g0`vtB(9A181$vT{gbKG0S^mcezd
z`*J(28-{viwcC~AUVTvlwJ?TDr(9Fo-h{X<LsF#r$XtO1Im5%l2+X8?@0XHSe3*JC
zL7q+C{h7U>Dmr`rXM@PF%{<~j2Y^B?GoWjB%+n{rUVDk3jyo!781K!t#KZ1w{xDya
zZ%%DB)=c}@*M55)a$c1#dIGErrEqK37Ey#ZCABe|3ZFHS4L0ezONc}n(<U5-hkXZT
z9sCCZ&&G)BR*#RM;y#`tS_kvy$2hdYS34cEViJM9q3+#Lx6ij$X^eeShUqYFmq1Q*
z<&Hb<`*9$f+=W0=RXNgxZ$NO~rNRBYwuMI~haSMsn3M;%u>p|}*E^rHYU+2cs`u{_
zr4=t)>TJ*Jesa4{ZTYu@{3)98*Ox`o6ahwy`?<E5xY53mAq@FvFMC*|o*_<tf;VY0
ziEE#VFXc62)=Cx0$5p`a^>b%+Cua@)$}QDpQH!rJEdOg^PRu}rJY({7HBrn6iHV7|
z{JO<upo?kB5&l~KZvmcHr_OpY3;tVH@0svvelYZn$r_pr7h$-3ajBrAM$`e*!%rp%
zgdFJ>nTesH;(s$S25F_AMNFq|Aob4MqiirM$SZghSITIG^%Rk0JBA-oDK7LpQGOeW
z^{H6h6!8*npU#Q9S$7%;1ie4$BxY0~u<5(~nc&>$hAQs?6wv1VK7&oNts$G*r>@b_
zUq3wE?BrW2EjI(7z1&|hzU#59$=c0UWO|MC)*`Y+FlI@Nk7^fxij7Cw&Cb?7drpTT
zA?e{g$LsrA6ffoXOG-Tg<r>QL6D-N1?|JBkser4RuJA433;R@N^<uG<H+-UlH>w^}
zB8|n;_;Zc?z#!SoO$D*M%=&xnz+jCx2R*yF0j}$yb)mx;VlwlRNIH{hmr+SX1P{~5
z-nU`sYW!x9@MHlSg2qC8DbH%!E08O~SoJ))7C}JeQf)38H4({c^@+$sx52T8%6L*+
ztEmuCM}N|-Nt8@?x9x#+b~siG@T84YVO$PiL36k=6lYxn_UqAz`@<Vbn<9b=#f7^|
z7xvWGrR>{gle!jgF|PJ}Z8C1O+E&tOg_H8rzTHn;?HsBBrb0RHJA=m??bR6kwx_IA
z5%<BZSHT3>G;{?owZRl~p<>h4zopqu@)IJ1^o6;V($fxiQ5TlGvizfZ6lVx!uF5My
zB9jYT@lKT6f3+pLC9BbjQG7L#7kfC7x?P-WW$KHh;h|?z49s`f1_d7E%{84itgd7U
zn+hU4R%RrGX(V4CzBTZ8R_|%9gY>iwUZTUnyI2)^J?{rTFHU<se|bRvS^Hj9=8(17
z)*&<!WNpB^ztZaN*933>5hwMx=5XypdvYdIpd+01X6`Xv;Bz%cK4&h9(1qKPWqz0%
zEwJrZ;9mVC=$=O}mfoV1m(Nxnf|Zl0wTT8N`1md>ZmM&S%!odmN09*=;}mal@C>_3
zNADLlr0q))*i-TH0cCsDu8!m&Nsw;j@@yg_l8V>GjKJp_nK;sEyCv3t04Sy!R!%+4
z$i)E?mV(%|X;^(1|GfTG1hka4@<T5_!2Yt5YD>#V`8I^hGs5s}qs<q!t}VLbzG`WX
zV7AsXYi86HCk25{x_H5bWevvLYx<tMiyfj-@l?JPB)*D&-YkqG?^%|Z<?yupI5es)
z^Iu%!@aV`;n(*~T$K!oWsEjCExgc+_G<3h9n3;4@_nifP5U`hn*aj|+rRCW0QZ8@O
z|Lp?-Paq$#8YE^;kThY-#K%=}>1+Nwv^4_WflSdLfrHNedMJ%ho}oXtxRVTEhPSpk
zJ(cQvo#kDwSXFq-GBnxQJUlkCfcfEvS_r`WeD5Ygutm*Xf0Lvm0mngw-d0%1_V!Bu
zTXWIdm(SW-4*JCFWLe(PXpor4MHysJXltba!NOXDH|X%PWoZtd{}{>R(>?r)0CSez
z2GN+j8X8J>)lt7s+`+D0Wb?}-4YWTI78hr^6ynW)H97m9>j-^KUvZx+_qIfx1ptir
z*2vIIB7|n^{DDXm-nZGQnAxdsptV=EwVIMcVNYffq@;P9c29R9X9?mrYKXc$$DMIQ
zr7DGe6u5KFIuSRw(CgdIg7lQwndpclL@)i%qF4>yL(`Ptm;SWjHIQ<1@LwYjNK-Jq
z*IAwzFa-K&#TN!1il4_EtKX`?r;pnry9}kUw`UI+f0*=gU9urcHcW<uC!qP;6!RRv
zFQCfnv#xzj-7y_S;bH|}LLoTGYH>~eprtq7X#FR(TZD~gLt#lNr;$~_&WlN-^Z_cA
z7Hm?@CT*-)jCcKJg9z?SS(lEANZ{j0ROiY*d~JSN0SuR!6@gt@ytex*hC<RC;5qcX
z$nF}HwKBJw&vTT6dt%S<3o9yK+1$Wi?`^eUwin;3xgFe;eX<_W6E$QVs`9iDv3Iec
zTq#`(3OVf*L31cx)idBo4?*A=KLDCX)<*{%ZeTBs$U*Hwp_jErbHJb7{MB@Sjx(r#
z4`kK*+S`3MHv#zr?U9VgV_}Y{u}6dY?u(5@U-q)9uy?`l*@#kZA=DcotoxN3Ao0Y>
zS!1w#xi$Cc?^QXPsHaO`R~PNhXRLiNDM**iND(k9TERC?6O)BlSD!}RHbw+iVq;w`
zmZesN*jrL)N56%*g{y^BS#0;=c?8(IJ{e@S%?HJaUR0E6JJrUIpNH97%?ff05XHke
zy-cUW@G%e!c4mGz%_nBYA%fD@>1o?#a}FJm@3dZ+9sgR}3Q}#eRCW#Yqd7Ce?zHq6
zDrVu%r*Rza1K9Zc*nIYzv-DjE)%!2sK>^f?H)zjE166!rtmIh8J9JIOQ1mve?Ld93
zwsbn|-771T26LJdaTx$=i1c|t2#P<gj~oEE_p_g;*V``N7p!mwe^i|}9riRIliclZ
zA*N>(e>S8!R6ZWyVCL$Du!XFGuLOiJK9^N!aCujBX56U?A1D3f5HVZ#uK7lD?2azZ
zKNT!wM6itg6<=j!fT#BK!It^)7Bk?$)qh#uoLSQ|YKA+L{Kz_m?)|EjVZht9lQny!
z`j5>x)1>tD=#N4S+Zo~<3tppvN-UZ@#h>P`&*G8BE&wiTMcd#7k-WLl{7YBTe)nk?
zs-B`OW4qj#x0MHaZCO)vRbf=SMVxo4z@%2IPJ{-Dkp5t1ew4i2y}|X>oT=Ko2P#~o
z`Cm%Ba#c_vYD-NAon|MwQgIN1T+~pTE_V9b5&D59+S&G_V~*P5_7FUoT-0Bp_@mDE
zyu6@?kMhN;g_6(aRK@kgIVmHwXgT@l^JH0xdl_@RWx@?1pFfeOaKd;?%+U{7ElMI_
zP`GI9H<B_XW*Y#*hXD_)Va^DK-aLYkI4#^+;{EC+rW;9$i=s+ni_sv-mr3sSN@DjH
z%$0(^`iEAvu4JV=7L0~P{lywS?Bcg6ZSngWWr$WOcn12&kF6pt0s4i0*^WWHH3Q{!
z985{3k6pI`v${xlH1JDsaGSAsMBHYX2J^KmAN!SVY3x&tprUYEs6Zr`;TnmwFZazB
z8Xxl_Lf(QHCe?+jiNa&v^TMJ-SHzR^<jAHUq|~5xEbwca<zO3Jc8mBdy%R33S&=Sp
zGyrR&l<5vA1Zl%p?(vV3$9ZXGgw>jY8{McV?fMqc7bS(#rq20GZ8%M@d1I@JH*tf%
z)8^Ti6s84(n5)x%ry1%fX1rr4Yq`p6zS>i8yt<Hcf9iCy%$T`*tMK2;smEqe7bU!w
z@1XJY$h)7jl*6tBdoF==ys7`>`0$#af9vNH=gezj6lL4)fL(TaiF136d|{t{vSY|7
zc+$o8$a0mE{lRs48`uLMfU(z;bl43({?Yv?c>2#}EXV1_lh-Zmc9hDfDyNpw`_LC*
zHq=U}m8u`x(}SK;M}hCwl6?W8_gD{+Z@yx(mKbO<Tz`%-!jARopwekBebI2PiL`}-
z?V$5udE17K%RBTpy(TM(se`*5cs%Nw;B&m+)yk5-k4)>`6VNURiHu+-KI3h{QF;|b
zBMeTLkRN+KN)Wl$;-+YW>Xz!Xq$yFw?MCE5U*g%f2Wn)kpW{C$tLyo@7_o4UaGgC>
z7T$P%dwRgfUQH$1OD0cW?>=j<i|XLsDe4TL&fdnXSSQfuTk9W65O{orA@>|jk7sI%
zTR_;Z59qz&FYt9}@TWYEKq$55B@H&wv^WWjZkO+2ZNSDL%E3oT)<g)6yLrgHM_ZF{
zZ|y%$1^id+r0JS!`>6KuK83;)<LDXtlaG!^z%N(Sgt9)qNQf=x{+ioT|ElZy*pePs
z%G$L+_}`I}KJ<Y7Mc;$)u0woBRG<5N(<|<PvnKuFADnMm^G$)F)LlduP182KjjC5`
zWQR+y>L5+}8b&^O9=6of5?1U-gVA_80bW_Ez-J>V2|$awf`@={iwfO$-6yQS4Xv7)
zwQMiZwJq8F`?kXUg5}3|@1Jdd^nINa4j658#(m}rYA{PeTrPHAf`U-<Myfc)Sy-^0
zY88p@G!nD%Q4*}#?T$ysS2{e@D?SgR0ZoK3V431}WxXJCl67aGVUBSK?-NK{EDM+_
zJAc!}#-4g$E25yFnetf<-C`@TAp*;J!h;&T)Qu)WW5ZZHe#L1rGQlj!XNr?Glo&Ag
zg|xzE|47u+L$DmJKT6g;FN{?;aovinWx#@>s7}|H(^8;hdeDgke1D<#&PcTCJvW0P
zEMKo|Xm}X)Ekf`7e<YoSTa)kGhef(U8l<E|y1PpnMY_9Na)dO}A>AO|!T^EM-HnXy
z7^6qWyYKHkj{OBYw)?rB`}%y&Q>Z3eZ757XMxO-ln^fNzQj}5CO!I12kaFO~m3ou$
z3OARya|rjiG*ThvVp#|ynK(w>QYEA5Q8NYOVemOul_#Nce@V43Tiv8$aa+-iR$NQP
z-n{$2RvDc%VjC;PIhQQ!tg^U>E3j>Wnlib>lRQjR4p~|N?Qell1n``Ti2j&8Wl|{+
z-rUQ<yDP*h_f}Dk65Rsh*cz?;MPul<G^9j}zt)9}F236Pj*0q!4fTKOTVu1DiD4t4
zO<(6}ED7Tv3739}kly5uDBvO;yzluQ;g<(I!81pu`EjBe90j4AGnD?US4#*DOjQ|F
z`O#9(S#@w3j4o7YMG`j%jEhwg7O~c(hyaQw;-WbyS1HeOo7v1MQ)tr2Ps?jA64S2c
z<Os7KOi`vBaOC_4UP?Nc%u8D`0m!ZU?|2G*zpF1b;bN1^_Ea(;+&;4J#IEsv<TO}8
z2IF7Gw*8=PB61PcJZfsfs-Lgl*Aauz^5>zDo)-VByvXTj#za&XtB<jm5`8c(t^l9y
zY$gm%^)*EZz^TxNU9{@Bua9f;2T0!Z|J@;=qU5goU$F^!+*AjY7`gCTDN@(o+1oz!
zG#Pfe8PyuMGs06pa2lF=E-5d?Vf`VZOdW$L&B<0FNw$$e8!({N!DbMNzJ4hlt2411
zo7&4Oe<`xM=01Et3MdoL5mUcTLM+VZ&CP3*4^RudPY=}lk7}X^brfE3g3PxD@Hr&t
zi24A6UMg{H%lg!+Z#Mv3y~IqI;;zPGfEb0AJnX<Vcn`NQ$V|YD6j{I&Gdw=z<(TU!
z(}{yS>;jxn<U5p)-tPw;yF?877fIrzv&f8K8wnN(=2W4`wq9b%_A+bA)pxiaEui>r
zO)ctUYSLxqrVR&fMV@4BLdG+Bx{j7cFCDD^GNa5S{CI<DvSpq)K=vvraJu(AI1>cd
z4hvsDd%alL9GtJf9^DM4UCz8c1OZ(uet}CCEP;^&`KUT&FW;M4RMRKx_Z63GeIk1D
z&xZH_NmBRM6p=x%if(RBgk=)XfkAy?OjOmypdRZ3gwuq(2BYVhM<Dh~RrXrj!f1{#
z0-&ykV||Mg4^v78E|0MMRC&gYZO6+M6JAz`hvUhR9Y2y(OaOPzrthwKuO|}>;f$S(
zH>u)B1iOxff-Gqf0$z^Wla;6U&XPoLNG%nbsdHW<4|*U9o?teDG+8>hb@DvNAF_k{
zzqHt0106S?_@_eGB_v+m6b!^yU6)V92_tYLsH6pB=A;rmIq=rJLxm}Y>?hM4KNXL}
ztt32RQq2*`SI3>%0Xy6Ng2q998DsJU*8tDu4?_(E_SQiw!*QjW&7PH<dqypD(>{f9
zM1-ttKfEtzkE1F3Br#F%d%k^NbN)ORk*2r0J|)6dtp)sKv}PVxDf!Vf>uXh&GMrnX
zIt=`yKkwfJ%qLs8y-Nf7{zmP=FBR|q9+{Db4@qKU9naWAh8UYg_x^bZ=0wl>$~>-w
zr~a8W_=qsu9X48aDG+o;r#QIXye@W3No}zs{h2|Nu{yXLCnWo;+X8RXo2G^jGl^+@
zMF%Cd_Z-S4MN8Tm(64qiOR>(4@Dw>4W$S{yr=7!EaemI|rVO`wAGX)@$K&Js8rxjy
z6x;)(w=G(10;8f_Zl7oAA{of37kE27>uAbDOvl*uFG?A%2O*p{GrSND))F_Xrdq_b
zA$ef+mTd2rY6pr2_kz<Ru-bg_J_6Vvqm8rDtN=LzATu_LfmD1W#TRCQ*L-Bxa7`l{
zcPNEM?%LX7OZmaMgTj=p@xQi-caZ*AHjk78Z`Sl9@i7Y6FiY*D-5I9mgh4yDQ)785
zS+M{fQbec|8&kvy!XRyx7jDb|CbzKHx5+$!TAoOd+N9`hPG?=tp9-EU-+t<5&r#QP
z70z;4h#OrFu5F84F9Oe<?i(w~ATmg$lqOgD&#@{FF%B2dePxbBkoJn1%F03q&}?^s
z;@fO=f64D4K}4?_KY;y%o0quv+NPms$7F7PXlR$CdY>o@)+Z5_E(&-KbXkVt<*xsC
z$Ngfq`{g_{;4@K-+~0?LqWk5kAX@LU^`y+dzHq5z#M5<(`tA)>r1Wb-ypD_rYpd(I
zCq8dawFL@oICv{~E5IoNfpx7dTS)lgSOEm?>v_3&1APvM^~~Gz0W}*db8C_9GF{`a
zWEfLY1dM}Z#GejjgaWt20^~xkAe}gUI?N6SQd9X$-h<*;UT)g;J7P_4VU#C;_Vr-O
zs^VME7%#c+)Y1%?Hjiit&dwY->X%l~VjBymw`AwDe3*?xwIG)x+#4uF?x}AQntgE@
zH_xhcVm<`UcAlW@t&hBK;59VR^mZr)a~bOD{z9GO0W03@mokPW1`O7scg9nS++?do
zg+G?BJjCS(J_dS6Zh$r%E#ctd<~)<G=lavh>LKgX>@D20o0h)TyQk}Rmxzs5zkNHd
zj~pvbUMW&XXTO%Qy7FJ@6CB>d6-;Z(h_9Xbu#=V$2@uGV-$DZpu1MkjIB7QS#rbjF
zY{J+a^yHYKej8tX_VEU19^l<4g0-l{FQQux0xeZM&%n=-$#jzxfejxywS4i~vruLO
zeo3*>fdlms5f^w|_vI3D33KV-*mzo0fG6dJDzq!zZ+l|r8`N-lr8-*Zp68w_q@GNR
zJ8%p(&K0iY8i`ak!*6x#-2Uk*iV`PdPUWZFc3oD!7yFj@U6?kqY>**c9rDP=8>ByL
z%YPEOV~HV6XZuEuFC@|QSj;85zs-)vs!=Gr7iVV4o&nn=eJ1G(8Aa}VGk%(14I^w_
zUmG4d0AA3lghwW%F)<BVk@3JjE1N*M_$bo`f4o`D9<4PqtIEP<YavDqL2WX)W@xb1
z`H;jHjEX{NKOm>4KkcR0_ptJyw&AP{H2mW(c^BqJzl%(6pgV~`U0++vIt&NU*tu`*
z*r8c-mY7tuI)<32l*9oG;1m8xL^m!F7d!v6%2;){{EK$leu*J#*3IH)1h4}F!K|Km
zpK*q}ublIG2ct)0qhDCUQ1Nk{#T=(OSi!c-sLM|aex0hq;&3@_He#tTZ!*fuo}@Ex
z1h=pH&@D!)(Gzg<dG;?#T$TFFKiQH|=tut4?Ol8f*&OFx`KmbKA^l9>#7=Wqt6cSE
z!Gn;{7LwVlji*UxPw5aJicp;zp0gc@P{V0!UlE>XV!LIAPB72VG|uTs&)fQeQPLqZ
z5p#|UrP$B;Z3g2XHkkdFzHTbG2iXZ)<JDL>Bt>!Cr_f){5zAyRS;f=jiB?DBe~o6l
zrt+CzqSh+%&EnkKe%`1pv{TtTXEKMmJ%&uOxl2Z<VFBegCFOUEm$cYN_JGjEf;b!6
zsegJD0kf@hj8ARlx9yrt32-qT+wylMxO=_GNu<uHjQ!qFI#0q|-Z<d?&l(9+6B`|3
ziTMbaJ<RN^>#^+#)b(g|zc0{5>(X&i+FX_sz>7^w#k<K_LO;msv^!e+1?27Uf}a%E
zvW%{=wC<T8Gu3J_1)`n`4BF<%=J~I%PCkfbp4e5BC53f)D`<`P9Qa=t=siJ7%H<2{
ze?N48O?g<4KD*9|b>Q>9r@_Cg-JdD&Tlp`Lr%mP5Xs?f-G}>#;^R_Aj$hK7}Nz(Ax
z>A(N-bdW|4c+(7(+N{6>WQjd!d~0Fs#P~8dGl?$z)-yn|&$MgC`E!MSQn-~$o$lg+
zb@zVYU|-+2)UG>68hwFuW1lbLcbka5_a!<VLd8%zLiJ6n10-D2D$%ouD34v|ct*TA
z>i2&9c_!JH)53&)yF#Q6W!*)g5+^cGB_e2ZS3jD~Lev{&3|udErbYN!sf|xNJ|-G8
z`nAfrkeN`x-e!t%m%aYG?cHwCQDTYJ8O#F$rC2h3aWXcmj9oI$AurFUUU2pJW2_N@
zKc%?;oq$mp*&ur7kCg|vm#0UHOsQ%(kmEI8>NIqlj5z3D;1207f2<pp`wBb!e)J3}
zbapu!z8Ssce!kzBLkW6uatVT0L*UZAL6@_Q+idvpW1feMlDd*9Ma_DhJ!GX}Xuzl5
zTX2^*U#(-pF6NSW6=Qe1BJ=>wAy{8>Y{m#s2qsRqqvoaHkJ@Oc80omRvpwo&BRo-}
z%}=cJ-&2K)2OgXKVWGEsaRM8&-nTn%*)l7B;@=<kdGA)4JSW4izGdqJHAJ-%zxk;!
z4k_%-sn-oZqc*-9{AEkt07`{j0&i*;#&SSIu&veMoh@*;_23h@?$$W|=-};aB3!Pp
zOW_wdCFVUIrll)W#mW7l!qPCn{&A)`?C8n;Z%5wFkMXIgv6GWn^a*fo+W6EY_}bSE
zEooM9&a`5I1xyTnhy5RYf4INjj38F(AA%G@Q>{7UccpEakG`iTdSql2EzFNJrqUw4
zS!&ZA(klz&mAM%OABYEjEsB|`Hda7MZp?a`!V2+57dd+{N;UdtN$_5L^!4Skzv4lu
zm&ltU&l@;Osbm`+?bbQhrycnmA3836x{*;*Ci%RdLvB71qb4^Wv6TL^9816L7K01X
zm)Z=mBd8*(>9QTJO)}(n82eF>V++gJLbLm}@Y=f?yV?#|;T<~hTi~V%n!^iYWvkeT
zAdH#T4so@CC$l;pg1ea=tuiw^ncy>i+IL|Ng@n9B%TH#W0>b>vbj6zdS{9<jTnJQ1
z8sXCGO@pqHQmdp)RSRE&C<*wbSkOIBRrShmnPE}PY<!6CehxB_s-c(CI$8v9)~J6}
zUT!<kzX$%wV~WGnBm;vFvhtW4%(_vjX|g5yW1cMD*H%H{tqpJ2IZ-?DnMD;+5><QS
zkmIsK=-iwnrmca!t$`$<djuo02plonEN{q~sU<<q_`U&Ph^+a}zJ30X*MB;!^0=HV
zI*X=erNUv2y=o&ZPv#e<Z%nxjmJfnIYl7Q*!nZnDwW(3`$wETap3hHz=iHBGg?xA7
zmiF#351O!<b_Yn<M+Ca=J?}RisE7Z<Jd>~+Cti%8V^`ige}hBhcrG2fE}eW~sU!@U
z)Tuu1hJzS-bEX{uD-NiX<H_^N_}KDO5&Ozlrb~6+j<x$+Uvg6K`_!qQZ$qT4>4FT*
zH11&Ijs0q=L1o*N#Q(i+4JE-Lb!&AUEK?Fe>r@?wglnAv%5W`n74B+Oj<~0Mat1u$
z8#^{O#vSn5DUNphV2Y*K@GbdorXKSVaiG(r>2qjwa~b36Or(^;F<7tM7PYuMH`>b9
zosDC7b-SZx2dHnH`nSZ_ms5g`$D@(E>PES0?CT<`W#dZnH-3g4O#*H=pRdV(W5vJ0
zXe+0>!@T^8(ES(QTJ5bbQJS2jNbtbp0+*@Y8!7*l+eE$&oy0JLO2E{eZ~qunytID$
zv$Gx?AXbjL4kgahNOzqN@R&P|Ik3GUR+*}T8XO=Q@a^wE_e_MD3UYTf$I!W;j%0*0
z5z2@IsQgarkM>4NYA6Hx?T*TG53CW!PicK<f1vh<&a$`;6+#W_Z~}xma`f3*S%(ku
zO&aR}|C#(6975`f6%ycMRCkKkya42+?!b8}&PfTtvvYnX;q-9TnsfyBA*Iab?_2LH
zk`dZ;EwiHxfV6m64KCU(XfTIS<)<AXWKT{^p`&1z7UO`EpMKyS$8QR_YxO|mKbPe}
ziw#58FhN5xz}@rhnT+73ssJe`t&FU?GHrfRlC+xpa;)VaJBotAP0t!8EM<ni(izxI
zx^e2kWbP%(Fr6Y9^~Biyf!aSppP~48I}B`!&7mx~Z%c_FdRg*1=+PyJ(c5{WQ3|YG
z<<yIEM%n2!bUcGH3@u(%=)M@VVp@q@;6K)X6vOk{zho8?%}gOsklziDDYJ6t=2&K3
zR8d7=Jc}MI3h@NHIVHrYFFzRFW!#_U8o0xbZOHWu{$&30aWI@ZY<E%J!Ooy=O7R1J
zn8q0+8HC4@?ArT;zapTAhrXNRJ@<<Dwyj>!Cne((QBf2~rk1O!7d)^&bqEw;v<*o_
zn)@iN5pE7AQfIy?*Jjo?!O@qS&ahupW=vEDKh!Z3)JyhN{)BbwPpehg-qrr*wwv1z
z#zU5FwEV<WT)dO4La#c;3SX#r=bgLw?FT>I)xGKmKJ8E|tB$3PSt2_W!O6^HaTEeT
zHD@I^PpP@O%u<ptZ$*%NL3miEhp?ZRxhjuJ0vh%oc=D3*jj{t3=bk`q=9EO&CyDzD
zb;nP=noLu}6+8!4j@$Kn4J+98nb#^|y^QmEdYLV-nfShqiUTBv6uE1Evx&-)iXC#s
zl08;9Te>ONBS%MNq@{zG!{ds_E5I-0tXa~6t<(x<Ao*CJ+$+QSj~zTYf``^_`mqQ~
zW2tYJv5yOjtU#q54&kQ^`-=B{u!vU7#C#4N#jTSr%UKtqF=xkD-<&mAc7U&`C7nTv
zW#(T|d?fXTWYOzw9O!o{fi!<Ck+cB@)DBLk8S5x|p}8I{r_K>@`Sg7k5v*TGvwd8%
zhb&dzmc6Lo?(7?(t);sYDv1%&YN@}X|FLFkhny<QQlAMLy<cPB`<Vjq*e^r}Z8A}w
zwHB5cKYZ6;{$42X8@ovxJqPGA`MrqOYUw3!&&6SRPM7N!b|b;E+w~@YSd~dy`T}lk
zF#ECm)epXRFciBb+%St)bp;FjxgsZCeLrL$YNfEI_znp1eZvg*+7JRx635!jc*^EB
zInrtbAN<_%f!`CQn@ytsSLQ>!B3D<XOqPi(J;97NXmK7n&=F5J5O7)cpYSQzXbqHk
z(HFs%c&O@1gZHfLhd$Ki9)AhoT!NXs{uOO~?e6^4V>qo_jCp+M2~BK)1^=5qXBi;M
z%*;;(c+9y6-E^*Y0N|UBF1Fk6L0)Fs&Z{7{C3rD@r44koS&QCVppUa~HuICqAjkc9
z*i13OlA(mMd{tM9F<wp4cx`+6x9mGE^r7N+QO9y|zPC*t5o7JQjzXISbmZxCHdA9d
z8$l+<zI$3D`vr^%b*4?kuEjC|1q7zKG-8W>4UhRef`fw|m))vg*+j9K^L3tlwwd1R
zMGg}f`}Qih82hX=P9`4IK7tXX6v#{}X`94e%<51mg=JS}sQ=MH*CV<23A%`jogE)k
zL{RXI#RwyMx?mpT=tEg~$Y-AX<LgzJELD4CN488xB3t=S*Df`x@JhY7g$`e*U%S_a
z=O9<P2M9m=-ehX6y~~BICNKh9LPcLaE59ZpD5r@^Ac1{V-nt4addF>C9_#W|&{|-5
z@_KQ!AVBi*!z$U+k#>;iTwhDo<|aPo)f)-pr9UL;$d}_;x!%8DI(n#eeU1P_xX)Wt
z;Wt?4k7WJDH!stH#a*2~i#*R+qU<@1-b3wHof&y!`EUu&Kfggt3$KJbBaeXAy6i~)
z)2YAWyyu4{5*<egR`vODWQ5i}SM`&NeP>-ZSJT~-nZ_>@FPJEI3Dn}K)FB-{fEV5+
zaQZVO0TB5ms5GwltUJZR_kP{Gi}`hJ;}!dU5kBA7bzM|;)~T15)zYVMDVChG=cOr=
z<fo9ZoA9a<bu?Ola({FN4_)U95@Ul+xk}R<fb9=?dC9WEbw{4NlhBn*8@SRsGcPj<
zzS5=(ukQ37QblCQR?MbslHdeVq7)R`ZI4C)W10v-PXH9FH?^_NE`zVP(_I^hac*pw
zXEnN2hYLFFqe~J#3XFq|vvx}=w!t1YPRc7mg2X&*OlS`93@LwRJS+S7=mccR((j=&
zHX&{}{D5Ql?(TQZBXTfWQ<};ZOXjcTWj#H8eYjHo9d{2dQptk8)^l;du1a>LKKn{!
zah&!_m>S-DLcUr!p%@Ifg|i^`nLiG2izmkGzmUgOuC@orz*l?Da6+<*_T@4@&;_2n
z{bT1O6@$(JvGpd^lx`oq{@Kb;a|hfnImJ6&fE<_BUFlX~YF@4V^%%2u4K~Fj*Y;?f
z<Bx!)2VJD=*9sy1`ug`nzW=4ie|E0hJ&wS0L_K3;UhL+i88-sgO^Y+_Mz)N%oOW3z
zTm-`2*W}&b0lF&={^rtygM)11Dh{npP4qLavTC{Fbdm9SyiOdpD^eb;`0C1{!LaHg
z{UI(;AWs%??Oq~!6wP?A&f0Xyag%UqxHlB+X<&l2L!&B|KB28Vi+669Ywa#HrynX$
zI{4!x#x)Lc0Y!=gNrBYmVsSgN9_1FHdW3<$6D63aRIdM50<A6Xu^94>Jl>^Lxi#b^
zvs7#Rm=NM5m5tG>`cDS!CoQ4>IgTppDy{~x`lHgBlM(h@dFJLjhaxjVxRutfQ-MBf
zJX>Stg9eRCW42^j#l@kR5`_8v)KTESMcPo#MFY&qY(IQyziJAS=z>+eEW1K1Lk4Vu
zQhtZ3tjJOz!CZ&?91Tk1S=)S*DA4}Pl*8w%lzGNlXFPJRm3Q-3@6iU}^5@Z?7MkuS
zyKrU#{98BpPT@(!9lRNq_o_@XH3F%F-<Jd)IDygzugO0dd~pqm95fa!JdnO^z@bEn
z?jHPmy>yyVe`vJou=5QSi6wt3u}uB-{=YJ&2QT^h)0hvbNf-PCJ?wf(dV(bc(l6B-
zb3to;q4eV%<12jxxjFUXyXRJ@^9!_xZN89ruPhselqov}dleBybh~B5I%nLY)XU0M
zc=eXbb_h|Er1qEg@87sN$>M0V@2eaXtV>#1F~Q5s_GMM2TPmDIxOVElXOf!5J7_o2
z>>@uI43F&2|B5|EWGs86YZLo^5*4ea_Su@A<H%5dcDWh}*tjt}DItu9dZKrU%qega
z!_md+5^&@sR7_c0o<{<zIzlAM#XXacj3=O~^y-`AT(?-_Id{<dows#NrQQq;AHrX^
zn--F08tryIlN$<F`zwHlA@O<Uufu!SIf<A)dUf2L2(C4;MwwfrBVTE5)G(U_*~mC&
zwyA&8<qOJvNQU*i>cqW`sA;8xH9wb0zC1+ADM>t^XhK`f)!g9M_&TCQJxeqlfQK8d
zkK#7BzwNLos9yMa;7h&G|2B30K~F1oROwP~LNZh(UB*wIe8YG5MItBTJu3=K!#4nw
zAlG^X+NnB%SYKP0kp~UeP|63z^_MIX<TVLfgo1Wd2bY^SxB?$>cEbC9RJXv82ol<a
z5F>#BJ{Od@{Lg`+!66&#7t>Rw?eGm@f&}h*U{D}6*WIvFB8Z&H){UCk=b(4sfDhR*
z7!xrjZ-MdpVNr}KFVt9+c&wP~ehoEqXeb=k9A)2lM-6VT*n&~_nm26$J09aANd|9D
zNDXdZc2u)*u(9s0V($;3cUPuoa{p$oowDkj;o#<t$908ESi*X0N?H6!e}B1(YW&zp
z%Ftd3iAiEh*Y$s}3xVB<7Wk-|AEYKPuYjyMObfOq&>Z)Lyf0MA$YReNRbowQu*2y~
zs%BhFqgbjPHi6T%kK6r>)`QN*E?G97(*|QadyRoR0>%Mf-1P#=biC1M&XYbYa{qBh
z1(6wdkk86qu!sLC!mb7Kx!lAltc2>^1pX(EcS=7(J#b5Wrg?gKaucsOLiO6xkYr?X
zzZ_&a*t;z8`1;XO#bb;Yc<f2xICLF?XDiwI1L5WGQ9B&YLU#+70DD&S7Q)d*UWRa7
zHO8V&2p+;MBAWPzq*lv6LO^9@(Nk5JPEG3dnvXH$H^>Y3b}<YxJRS<TJSxl2dgDl+
zuOq=mpfX_2EcAW%mtFR9d$Tca6sA)}D>nyRM4G5n^5MQW>Rzz~bH8q(7SilAHzZ{n
zZii=a@q3@i3`#+6kYV+?j<qJK7Cp|<|MOM6!|nR>%4OXvtSkBT3DCtNP-i0GR9f5Y
z>Hp7&c>$jE();WAm}dy`(iDFl(7GQeTiTs8LhBQDKb|KO^15zCe?2C$#YU#jlogZQ
zVKaB?sD&BSnUVLU+0~cu2<4E0#K9r3LHGNdU9JVoT2lx2k^-Z&(SJkm1Xy@q#s}`Y
z7tRR2$|d{0(KlQ4VY+zDw9aVgs-JVOc{gS=Lo6HV+9+qfd%TOch6&>=`x&+_gz;7{
zSzbROaiI{-d6eh~AuZ~c;YFbI>5{O1gxQZEW#X?zqEMH1H$B8i_pw-nP@|E#QHRiA
zhUBS{gh&dgeyFUfI%0xTKEq^fDRdu@JW+Z`n{|1xA(mU>czr_``f26Om<G3)x3~o6
z6|Z<<rbSyoV_<-!!PE)wng;1DzF3gc+51SKYH{2L^=T|DBvYEi+a^uK;CeW4LNB9>
z1)Gft$;0bZC2ENvuwtWRfJ`r=y=u=+@iVz!XL?PIDaMz13i0=nl>O%T6V9`}p5SS!
zIpLU1-LZ1dbc&H9kr^%__v<iYO@-ZXIB6|z$vDlQ--_HX5?4l0xClz7DaOwmC~Cor
z+p2R+9Vm8%_bO)EJ~3yh5qMMNoA$r+catSLb)krroqY`c2Vs<W+$G4&&*kDfHFuMv
zri%Ic?e>>fZ2Xbx^I=1A^28p@_jh`Xr~iPkVr2Y#?vxSh8UnFKL6#U6<v)gosU>?>
zdh8?sebe;^<HK@EI;>tU|9}wR7>GvhBC0eNCI9y~*THdW2>X7`b!OLkNuL?m;-bW3
z%|Xv$UJo>3%B4KX)Raj{Kiw2b{(yUQ+Et|cd}4rLUax;boSO*isYwm!_xmlK40WKF
zw$3M0F71GkAKBud8V_3m4-TX-1qT7If5qZ{L)I>^o#k6EP{38O$Q?JDGz;b-LU8Iq
zrI(mla-qSAFGq+NVm7cKuN@QeyoIdTa;#JQr*#4`(65)31h*O#O#F*Pd>-2v8&;S$
zImtV-*<y151?0NLsjVC)uYv$NCK&s`wL(NW__bn0*@ruK4vfI7-`7Ktj{#k|R|2fh
z@D4><%^hVLZ2S6PicG*3l=Pnt-3xr;iL&QgmoY&~xL<hm7eKWj_`1l}B%TOM_4C>%
zAa1+(0rH0W3M#A}cacQeB-CX(O$i$kr#MYqZ};S5cfi0GSs*eW*bM-6y;E+-Nnr5)
zV(6X|XF3RL9LE5{0T{yZcJy-aJm;;?xgIyAYNP4W9|E}-e8-gmI9nGF^em!$6wA(&
z$_KMlHMRuwFIC0#4Emms^-<ojc~?<A+r51izbk5gE@Z7gOGO`YZPe!xu@M!psWxhF
z;|pcn$<w2q!uE`F-3Mbjc;$($&vke(y^BM#bfl*%Q>SY7{2TwCa>-k#x(vJKXZ-JE
zbTSH-;obSx_d?ti)=KgU3Qml~SXjj%bNyE0CTq?;UF8ioUTQp9+4>hX`q~5W!sV!=
zM}5`6%apDpIF56yF7PaUFK)>sxzR>#v;uylUmp=<!J$&N?N7a!m%V#q?x(AInZlxG
zXFKr9Dxj`cQg~K&+4O|9*zn;1UD0nT{n<AVg{%tgs;hI^iC?J^j`(?o7D-^QM!UA>
zl245)r*o>Vb1C*)@I%R)fSjLkX)y*>IM--P-w@7e=%>MT)r-V%(YuiQU6h@<oP{C`
z5*H^~g^>kUdT4VW<;2Fvj=}AgvLB=>7Nf~sw9L(1?1j;6xWq%552o3<plhECX)>{z
z`h@udHS5nK9Ula2qkGhaQ)WDBoNRoxiTBGE*k~2va})go4Odh@e@d;a;Xr}P+8Rnl
z!NU{uCet^$VXn1RXU8$%PKRVAK<~#`RmlRBEj%Q6k*l$JkvPmA&fhZn4+YjFqdBBv
zJL~Alms?7xOr%=I{L}IqeDT1}(HV6Uv^ZO78z5~PcC%s+RN2%53W5^4Kna6|e8>ni
zc}}zh{tSF3<x4*AdTYyhc^!#(M!tt8Ym1;Y_GY2ls+p}~91uv?!flr*4uAn(3B?JH
zQhv%A;rFmb^sxjbRxQr0%oal#1>r1-U2KSLc(0-pr(cbfYpnqfxlrjq|F3eqw#`*q
zBM{qXgDC-UO5BDmB{<*{;Sy)wBF*AO?;XV`Ihy=JT{DXlVtGPh+SsWR?S%3hO|&9a
zUdJb<-v@BGo8dg3qZ!87ckrB-OdcncZnoF=C?kU8G$C)F`1tJ{r8yhF!O&zWx@=G%
zfxKMNvBDS`+6U}d6MkOxasyQZRmnbc!?G_NihrAIJwpQ(+iUJ;{acx-+mBoHzuclM
z_7gbSSixUq&UxpXa?L44nsm@QHBb@QD4ITH1G6F)NI$erHN&wcc2r+bjn|lVXfs7c
znVHn`(=>-lPdXnKZZ}eOi~|+Pfu*Y;Z*}{H%I2Q(1%zWXPp2u&`Owp1!_f0pD!$z#
z?!z3%0Y|3LZEDM>&5Zx0R_j9d51Q!ya2|MC`xhXF3`ndzNGjX#+~7r&5($5MU`-m4
zfluB2ZqVIbE2Ue4qYOLGwKjl(_5S|v*=f{d@eX}FucH)}YIQPO*CS1t7N<w<;eJy4
zNDlX+hs5y8j=v7ZrUg{7-u9yRoZg&prO>eqIwelO&Ju6d!;iRu;;a_@0Q6yH#`GB+
zV>OH&WJO3IJ^rGN`xuqDk>yi2UZSQ=UW&<cs!8e<N!}8wy{^M#bxm<Wi?G;vn51^3
z;3*OHt)zNEIO4Mo;4V0hOf=SMsIk8`R~Q=h??==Nz{?8-%7FG()k8xBt=l!08Mh&!
zV4uVbyc|OP925rR#k%_<qFWgg06HAlrtE7Gx_PTB(JRuN&gK~^<^OrDXhx|=GHxMK
z-X(;C&vxg^;|eas2nYx+?`5J5J0ZD=9bB?*eZ2Iu#)B-o^jo_*heM-w6#v;jJnsd9
z?>AyjN`&<|IGzvO#5y-ohbFJP^C)lHl-GwOo67>kTz+g^{fH;@(+9oe9~2)Db=`gH
zDmg<Tb?n{LP!2UMzzo$po<sT55lC`Lbb|eYICo>x*Y!$&&HYmpT$?^LaUQPEQicri
zl&n!(@cGg^m6B;z9|*kstozu=iUkh_`UQD85{pr|;QC0Ty*G2M_jU}TdI2#ZZsqz8
z(^-+vV)c>q35hu7y1DrV)&uF*Zn5+wM6-ZS4GzuX4t2rT>uy+>C7b^MA{WzMbxR10
zh_WfSDBnQ{CLC&>1}nF<K(8vsa7V-S`&pLDE!4AN<@Ee?^njr7G^rB(G;2hM_eYPM
zC;+%BXl~xy##YoeyXTi_20nvj7V(3z<6I>XJ^7HnN|yFNJ4M?p+KE)-6w){G1G2qY
zJuxJcV<Q9#s!`#d0L~KHbr4r~_5`H;xnakBf1X(gaGGP-;iGqXU+sT6r*3IwC0F0s
zBH&+&3{<)rJt{#?%8OiReEzEs(5(U_MLJp3FQfCT%hVZ*9AFE<l`DCG)26(#HaJC?
z`E3Kv`@uII71y(V38(*j`-Lj`R+2*`9sH$BaVWWZegV-{c*;<+Ov016vL{f%l6b;f
zMwD;A*&2t><iR>|5H#QfPA>*#`>f9-4s?{_e&05(R{k_`+CaIpC%`Jwd)x52W>tU|
zHLb2Y05zmuAH;qX+g#H(M4wK*&|LYlgU)7QFvcDN`@4OSv5yK#WvB*9-Zj??a%5m|
z)!0eLd@4G%V<QPe$J><B;1cKmG?YM@N~eIM%$>@+HF5Xv*LyLquW&ZLr8#^hGN`~8
z|9hYL16(99I-oo?26w+%IhwByW>jV=_>;zS)!SQXMGsin@!2<1>3qnX(Zo;IKI_w)
zQ<@W%ZVlxwnbm85$&M@irA(BK((t8983mF&Kr0*KT4wi_+5bAu-G=txtRTF07*`?4
zd+}9hE*YZ=K!db}5ytA*QtyUo8v%s<!Td}*hbwIxUR7h%2tW#xAEujhk#y06`N_E$
zVgKw0RovoH?fB@XW8Px9ILha5;90Uv@5=;i`y<O7m%2$4aJ=gyHzpgiXxGy!zQ91z
z^&>%Z0AOsr|A6CEmPM_CPZn>}6tk%+qqpRD*wpUD!PAqPh5+fx<YXZgpfzs)vLwu*
zR29#R8S|By%|TptB{+>*li3`^zG}jY+m_UpM~&9DiEFk=5~9%Ly^H)->)M+p3>#@h
z|A0Y^+RDiQk2JzbNxkMPT%p|DKeoDB4TMB2>e+eDl**nQd!jfknHH2l-#E-@UF$yb
zo4I__IMNpDp}8cCauE4E+R&Hce%MbY<bLu?MKyMGp-N}>G<RivM><HuE7qkALU}JA
zfcW$fXGA!dH9^;I=}qPz6eujJX#C%joU;}n990`|#4RAe!NbEdI~@tvQV)cbz&C!*
z9a^BRgu;|s^?#2%xR!x-QlodH3Vr4XW9-BO^QfP`HC~v<#)~5vP2WkQO64nod+b+2
zvLDc<Q3>WNL24>S*hucBZ3!Iyyb4aRgcwSF3qEM9WcRzddfi5>TyL}~PLK>|EF3VP
znf@94p_&cquT}YgCrZ;>=N~A&?*FKKx#zv!P{9bQcsws(((l)tU&95qwhC-dw=4Le
zuCv=4PDMG1opbN`UK{V%A2?Chw7g0HUVA9LlXLES5el$I3o+K;(Ht&EK-<eEu6^uy
zd1jT`0UJ}rQ`dP{d24XXMxzZ-Uy%i>L>G5tr>_3vT(S^@>h@kt4>`m!QMEaHUxgtN
z=6%G&&CUIRZg!kto!S4LJHs|QE&?j;i8qE^CY}#)!sK6)%~;vICOh-U;#<uGe`{LL
zDc}xXV$5-k!oiGrEotrU$C#h!_r?B%NJU)e$MqFm9KBw2;`h6iKx&cN9pT!YybrhE
z&^q6}1aNzQeZL-%7c`wqxbOD7w5K)-p??)T{e7<f?)A&54_JL0lzkg^OBvJ^^o>H|
zWX0_^^Xz5r(IQAx>Q<6dh@UrweOg7V#%G<9Ejh|{KQcPvdrqhqhF@8}evpZ%$Gx;+
zn=2zD(@^pHzi42b9|nzZ+^-aQW%KhsTg}T6At`@9sm|4GXBCb+r@vnRnnhlqdCD;7
z4eHwpV}HNt9q<M?zO2}%#@i%S<kJ{e&?+!Lp{wlq>%W}~f*Qcom0E5IOY9M$$n*C6
zpZQU<nmsxHhwh874>->G4;8sOe!g7}LG2`^P%5mAQ<2ff<?mEOVi`+{FK%m~^R`kQ
zzx|*1iR-R>#X7eNE(Yfqt8~}Wk2GQ}OC&F<KSxSD@y?K^9Ko_0g#~0a%0?(>ZJL5S
zb-}*R%&b|cQ^q+z^PiV-*{;c{D(%Fu45DA?Bx<sMod++D+Y0G6q9SBm3Qp@sz{c5+
zMT|0sNl<~Ol=Gyguq+By4$^~)&alF1eE{|8z~P}Q7S6n(L|??<r_x1T4*+%vH+e)D
zo;!%Bra02#v(=7&JzVikAK^wU-&Dm=Cj50v8{qS;1w8F?N6T^A^>*oWi36gp1aVuv
zs2|!1#DBF-@1=>AYJ>+bGc&Tt6H-&dbxll~)3j!3l(tMq5p`@C3omqPux0;#%8(xZ
zdem|+nGghXV@zq$#y?Q|5(@*S{?j!fFoX^)dhuwdX;)YLY5wyr2J`Bnv$APxZ5n82
z!J@C4L2GDa6uVa<myRGI5yX>5?97}H1zi(!z$^*~Wz=~`sD;QAnQ_K)qgAYMj_yxa
zF!X?X@RPP1p|b%e3=aqQjg@=&$F#~|AxR%a$KUHZ0b2xS<PBY28{kBs{)DtxSOD5?
zeXM3x5{Go-x*(qz5qwh~t3*Ul5V=pVB?u06`_rhEu(nn+To|YIu?l;v7bw)%&d1Wu
z^95Qpa;u=%se9tF+A$hjfTha0FUcFt`8xE4hS;!T94l^Jnjw0x@)RHAIQw`EpO_3Z
z#7YD64xdo(A*?)?QfF^la1l39p8C0g$S%d%S>qO?rz8iN6lU#ZZc5{}m17pQX7Ek;
z+j}?((X$B5Zu>}<Y@YJ>Ax->r`^0|@Ag+_QI9*ic#s7=ssVSfj{~+YPn#)4toXAJf
zN)v?FzZ{Zs94kqv$~lWXMWewJCuxLlPw6}a2CwP10#@F>I>7|A5b|)sc7A1zrt;2|
zzwCphS=rZP??nUxI(6jEjVj_ESh;+nf1$h|@OAnm=V;%-OAd)3_~bzA_?Kg}B>I8A
zM^8TK&oVpgIf8F^x~*nO$HKy5AcTW|>j3|4QE;SYrV*|{fqmV1#_KoQ@E8v+-hT%#
z1){b*wJwfm^2yfkp#y^Er-kuv(jI?E{D=h@|8pB?Ip+?7`Xp>o39M2#I<y=VKk4kL
zDLpDar@NV^#cx*a2%jRMew8l?@U+sBoocIPX{hHUX%Xyl*iv(ZI>|K)5J0lN?-L&s
z1wV+88d$u8vWnttll8`CM(rW?ci|2-*`O+XaK(!@!@^Gu@Vc$%2~iBr<x|O6w<_gD
z!@RV4D}|jk5dQN`ICNjqBnngYa2MUI95*jVP!?vI{!Vo5bbHTcXGr43&9LoS<(KS}
zUirLqL>_0R)B?TJ5O1b_=N^s!S_p<hyAaM9k=hi_r1V7kb&}}a?+BF`(fe!ZbZz0E
zKa<us2+B*`@wpkYM-RO}^t$$ca3D4f+!0+NqA@C)Wl!E|)?^~w7+FCQ+cHE!!w47<
zK{(iU6=h`V1^uO5a#@Z3D{>Vv77uBKf6Drw*O5&Wt|sZ81h&*hNnXocHoRU1-~O(f
z3rM=|I^}+*{hI<l_UURqO8^|EH}BjF#%9o8dVkDa6v;U8L)EfG_}RylU>(snY_`rc
z7eGp|Bg4h30Yhnm=0*&E%rJlC;x&x(UiTGy*`9i3?R*4z?~JA%a2?}iXT!ZwU@RgP
zL0ivDXCCv3Z{9eILfm!==pYC1H5r2@n(=X-RY-2tQ93iFM5!N8K)_AWx17pvkQ$n!
zJ%1w-R3_S~esnan&RmLH@jibCd~tyrKZV*Q5wC2fDeQ#u57b&y#olo#(82;M!3Ot^
z5n^GapO!(*xU(<HT`O`bCTzd*s1lW7D>H9s8{J9z-E>GSLk=3xY<sRp@?@#7uH>H5
zy=!<|eVmeV=9c(?eGQ($>n&-F5np<R#=ii1cNyM5VnH8Mx!q33zl**;@m#Dnl2|;<
zhw+hUYo`+?@Ug-p)4%yD_^5anO}$d2jTMbLI4L)LcTx|b6UW(u6kw5zKRm^xnx@wz
zDqj~n0mdYKD(he{zf7(L3h9y%k`X%vWXEVgO8G59$oh$+EJ0kZ{F<k{L1bK!?c{E>
zRb9y(n%KMu>GpNW@oC3}UijHhb6N`D)Vf8J#(`LC0tg%0L~@1CqOBqiL&=`^rL)~L
zb!NXyKyeNA{@2$s8Yf-b$GTB5IBlqO4d=FXE5DzT1NAP^O*MaVGIn5$!zjMzr4m^a
zO$P-v5{bojrGN}roao3qIS_v?5D>*I{Xq7VgOt5uz3*YXFnOAGfpoxJm>F)@9pGu&
zuxjH2=MihY^1}khdbM1AE53cpLX4Q$+uO^?Znr2E&(eiGShKRRJ?7NL0qHHfuRmme
z!s9npN_@gIEtEp}_=a)5p~BbF`P9p?5-WpK7Jy-f34g6f4-*GeQfF%cuJ+0oIEOMc
zMVow@S*smB8X<YcQdSefi198<8uKvh>5#sFKyoGdNR~SU1YCSVN3D>mv4~}EYs?si
ztE_nf@|c_y9S=RcbIvZVgm-D2X!D#znOd1b#nEx#1^b}+2&N*J^Kakg6~rT*aP!g!
z)$&xq=OFGyTN#m*+asgJfThq1B_t{F>}dI~CY2I#%GZ<=AExhVePk^b!A2qJ&(Mg$
z+bp2;Rtz&Oux7C2cqwl1oBF$q9?Tv8>fZ>J*D#kj4J5}%v`&L3`Y;tG#Lu!aaXfQ`
zDpRP<ZDHrA{CLvLJMX8)loYzKVN9H9ljy6O)Yo?z$?QN217%Nd8+Fe167ExGBdL31
zp2ekfB#G?B+CEYG^H>d67mh847S*NTbFzad+Kjbd8^kZ8%Q;N@jLaV1>OIoNlf*I_
z&0TE&3_HjuAS_8^XeLnTzY~iu^~vJ{=E4M0)(hl*-Puz{zYO&Be%{{;^?V?YBUk_X
zM*tpEtZL9Rxe|kknmXDUg8LbyWoFbi2SzF`TE8C^jp{Tko_J(2u}gT2Sa<{ktnJyb
z9-25wJ(e`6EARLwIP4(xzROZLUAfz)4_eK=e#RY<U^~72vWU78owR}P?`b$BHtXrD
zFfiL|xcb*u%$Z{~Hb6B#HMn#5`li2>F}00R=wpT>1WxTx{dF?PRaj*+i5mLGB1GXM
zav}B?7v3|i%2L4&JYfQll^0%?M5=HLl0+!Xe(<(8e#DrdhNLaaUy6aTjgNqU#QZ0^
zw+<%udZz9S_1QAp?R8si>t(O<MIZI6Abb{F<+poh89e9>-1`F1u)*g#Vk^9k_fm5*
z+e&@CPu_n2_mRrAVgf8e?6!7nP*z2aW`&uv`&mp5&c{ioqxXg%JWdGu(u1EQj_{g@
zHW<l#6$Ce2>!?B6Sdrfa#a(fcq4G4Q5aZV)2v?cQuT(!*PDqV`_clE3|3R`@9|1o7
z!atp`QwjoqRW-F2SZF`b%UD~=?9+!Av(qmdozPRa(ZCMuPWxB6Rg$<{xL6qS@OX~P
z0-s@b;w3h2*T3WPlJc7Pxsa>#5={Lht8p#LRy7#8z!_AHVw4c$!IyY!q#lCQfe2+n
zpRMtVg54jkWjh{5JuZr=I|!yseA(rP)Q5%-eR3ou2rT*#ts`g!`#Kh>_G8C+BRfNr
zP~ytG303o44B29PvO|YN-fE!wSE-4%9A~W`rnd$vtO9zAu7Lq{l0!M+t}lZG63kCC
z19keiOs&P-O}+1f#80W8ZZ@wJ&ysh<7nnc}K}^;8r;nxNM|oa5B2%37fTlskgW2<7
zn~a5~FsWX(rXcl}&YpsK65wA=N2^_l5-F<0DW`)DgIlMic<<G#n;vhIZ~;fr2mmKl
ze5)dw@*8_qi#PclnntimX$rn~4it5idOII~O}u-j7GyU;=L4DaIpK8X&OAM=+TC7O
z=D9d~Jtnz0d)Yp?*vWtTXdLux1?1Yo-JYM9%@Gr!D3R{5ZX(sp`6(;RDT;ZY=W}zJ
z>2s5bIaB4rD`8b)E4HS$R|snY3Yf+^Jg>;uedJmEI{kZDyA^SO4T*8H10~ui`}#wn
zvxBDT*9Izws;WhIQ=X5EC%VeF*V;||XN`^!(-yN}MfNd0;ZbFXXU6agX@(}&R;YW&
z`S#VtcnPN4)!;-%c2`xy8e9IcvN6?(bI-gNJ0}S^ZmIP;Gd=K3rzQax;}`X{e~@9u
z-G|fPqSV(LzdVd99kh*HPseo)MFs5keSr+FQuSfPL>+{iWGn)<%6t^6m~!~BE{g)@
zW`UC`e!LlYN~WW6suR~_JBW-D^+{)J%THAG!QCi@i`|=17Xc->Y^6oW;&Fa+=VAMk
z)0?K(w;q!|>a43E2hAv3SUp0fN3~L_5}_YpqH9@0Z~t^P_93D-vE3B8uur@A6nnu|
z7CQk;0T`rVGMA~}?Mob(6u=){BA!}ea)BGbNJ6d@i_Pn_xU@KL8~Npv4r}!X_H4l)
z>C!GUZ~2g12R@Ly1`z)!Jo0*6^z3<6`&NY*j4pTgDMvYc^S5qo1*;f}!zk9vHEA-_
zTGI^qKO18PH7@SUnFUBAOlhEF)G%Jjlk;*QxfD!^>G)Y<!$6x9STX4QPN}Er>$mKi
zJ_y27$zQ@POA_(|`F9UPm<gQX_-(9{YtesKmgXZ?_yKmo*2*-hzHQtY=Hr9wlcnL}
z=l#>B(yfsrQscGG&aCBHBZ!85WQHTMJ6?@;v_#SzQ?!{Ov!DKAiJz%@j#@G&p2vZy
zHPdrm<&0uwm>XB3xDRrh+f>e5DccP^vp>n+{an;)FpTd!*0_dYx0yIKYeF?ytEOgz
zMOI__o_}wpjWr|1*&9p(l&^};+ov{__D~12VFSAQtrhOhWK-NIi`waHIF5gYLI|(k
zTv4?KtmwOnMG(bn^>E{bOavER43j*tV#(`9;|bB-SyfNVhaD}D_K4DXBFGD+JDd9O
zPUdIFr5pQ+uiu{<TyDf=9##09fo@}EjsU=9X{M*yLFY}a!C|qE6xWZgOM@~{hau0^
z9bxP1{~xcDPN|ZXb0P6b6%B#`TPiB5>aq=Ar~g1`Zun>$H6z#fR9te-!rORy>tx@+
zqCP{28m5cUmBUon@pn~Kv$U5Yhf#u3QWJ%xIlKaCkV4DH8f0W+92n`oJ)#fG4!rFS
z!BuTmd4tDv65((byRr_R?e8eQRSeJT@M@M*aVlGTubvNyx2ZGy+0uEH#<)eRnqi0E
zU6igkgM*$7rsuPH)FCK!xnIl^y_ar3O5JD<M4xpJJV^D3-fUjHnz4j<9|z%M>LR*P
z_1g!iZK2I5&65Ov9XYwqAYQFO4VnR_^6xo{JuKhiEL5g+q`BE+t~HJBxkPUo)zxq}
zsm9(Jd;Talz3_!Tz*+-g^79bSA?SZ&c02%MRp8T|{o49Lca6zl(AsJC@D9Wkw^?)O
zHOc4UWtgcAdib|WM2tFoNxA(-b@teNiPI9tlE%!G(VV*mtP^eCp!5Br7T6RGS|+7B
zB{5ocLRiXw8o}MEds*Yy9*yA`Og+ep^RQ7Hf>U1cTtF~+(XxYv2Clg~aa##N_0o5Z
zjXdiUqO~s5MHvc_nFH~GO$J|c-|>td#8|c!e`P+-Yz(9bKH^t$8C0*5p7uS`!)5%0
zzgf_OJD1G>s3hH}Ns_@>WeJ)U{6qidSwS$!+dNR+|8jH0Zq``z{1R!&8Pe?1{ypeq
z0Fiy5eWPWgWegm6BbHzo?=M1CH<w>w4YD0NP_5CAy|Ph%)Og<7qSXrZoU+Ekh7IlZ
z`}`V11}TH7#3jhEL-+~4j!5x8=1Gu7ingruLS+M)@w%K^RpS0dvwGKc-1<Yt&^w{=
zE%z6<PU)Q~^EPt>z&T|rl?~0iX61nD+qTG;viJy!RFknDQmXR7lOs(o4?lS*Dq_B;
z)LN|8tke%>#}?n)V7~l8`Ny5hy{ijw6~{pPSqE8N*Ngsj0+xg0x28!sT~T6=P9KKA
z^U$*ac+Me@PnmO1of>pGs_b?=|BGzoU@CXE_0c#s7UvW6o7d=YTUzDZY`5DxnwCLH
z+;LdPQV(asvS74Oonz<k(O%Y#7^gUwY;loQ_)|>?&fODo02S!<k~X-q%$eg*pAJqg
zwh6q|Pc)x+NB(tWIhG8|Ym#VGw(uWyXkQ<{g4grIgs!pvqBpO{v>{%to5<e}UZU+F
zq76lNEy~y_eY3filz>nDV(II^S-Z7YS`jcu%29NulDK`U@hx5N2Y%KcHra8i3ErNS
zuH1s1hp@Ywn><qnh(|j=|M95~m|_cWA01zownM54(-7N`VO=R*mFg5tf&>G^#mQN~
zi`mfJw$IZbgi9f=Bz0e(DU=S;>hd<2&g1n#a_BB~1x8z%5go84F&%{Q@e8nJRQ!6(
z2{O)WkP=hMqNoH$Z(FmvN=k;wYqO_+X;b_h*^K`j-;IpmqQ*HwX0zMS`d#N}X+hjp
z)jeKoEUXNS!N$isH@cVHx%Z8|r>j#l4<4}XU(H>~YA3Z(%8pxkd|V{2{AM2v$6A08
z9mJEUdZR!fv8!@zo%B5&(Xv0w<cXGu0z0ljWm|5CJEGItv+rBmc*|~#wAaaQ(R*_b
z*=<=XgK#)=(;wLE9Jx0dMhtviT>6{$bud%DIrFv_VM6QIY^&LZ;wW@10#73jOdjwI
z&v1SSlu9Ic8gXq6iP}S%iRLUmw{D(Gy8h|>4apmlFk?HMwYgI3Rv469ohGq+a;Kv8
z<?iBn@nys|E!C*~#1>``h~Sx&2-OL{dEfg~y-DsQ+l0M^PwOh{8fE>SJfQjdcc&HN
zn*h_sD0>p{r*ve2kc2C73pN8*2tfS=TCgR*Y%4;fbyxSdp~!2Ecs!LgCuAv<G4n0i
zWV3flrt)Y@N7MET5{KFB<>{JGz)}22#StlaUaX;=<lZDk#I!#&;Vbo21}U;}Sb7bu
zE?m{2oHv2-30HLLYsv2igj=E+lgzua*XIN`=&@F`S*sZ$zNdNoLW1u%!6k*MzzUaB
z#0Q0hAS}W2m2I%cq+8I_^V~Km8D)<S{26c{M7!@_BVf%wIXJ+QZn1yj`0cT(s8Oy7
zZ7a0;i=6zX6(4|hzYPsxehL&LH%$Ej)g0|@%$i7s^y<v-p;~%08%727KOY{x=+(vi
zFhk9IHz$D_`p1<Rfqsf7MK0UtqE<|u_{5oDtE*;2*G?v<ekW3$>~axl!_o2Q4<BJ5
zpnmN62~(%Z$Bz#M6$OM>=!}Q05)M_=w9Qzay~VRT+6jIPX)abJ(&d;d($EzVYo5E{
zF!3bnQS`fwq6*Lu-jxIY9{?CZ=f3QaN#PvDamk0UymDH6P7f&!Jsh7rAP|n81q7|2
z_tKU3#S11~{iA>9kwQ8y@kF>{a+CMK6Q3#1K4s;P;CTHsv?%!a<4@z0Pd@doo{sCd
zj_bIm#C>Ceuc@i2Z?(qP!~Ah~()VwxZ<qIJb5BE=kGJ#h+dt@_*wUZMX@CD8?z7)o
z&!T%^FlR*L!uT|kyc>FNFUH0O<DIwPi07ZZ6oo=Ac6OREdwL>%{KI$R_19mDmtMZ2
z>nk?%)062obG242d-r{<f50Z7Fh9hh_-Pk4%y5`rn0p$Ev**rwM{|FzY(d&qV^ehe
z9wevAhL#aba|Gz4@#?EDM9s`iQT^<=K2SI_!J*~T88xO|^}X6(RQ@G12lw_wQ>tFj
zZN93$M;xP8OPI7%{3NL1i7@wn!|}u|QPWqadK3lM?w{0G+@-k5@s5|Rz;f+AJ?<ao
z+TR>5-SGbqPZ-_UwO-1>jD-33&YLgAfB5%5i=Y4er|RRk)K~P+PW<Vg{2+e%r$3Ho
zE}zr8yu!nj-4D!8A%OH+zyg+3FSm#9sA;0zaYy;Pd7$-=il8f$dkS;1d)WKqF6Hlh
z?(^+*`TnSMyn1l>drsx&nRN7*#ubljkP|%W3Ilt1pb|g+!5i@({==Wg&wugrc=!9?
zjj@q_Yb<PQjG#4<1T$<f+}HTX#l?#=@l&<!U;VqE$6MchC(fL`WK9kYIu*saL@w7A
zh2C6LYkg6u4BOn*GndZCYZ_y(f9H*O^|hB`a(uwrlov0a6@0@$2DX)#s(T<<=iC|0
zh;jDNzVfELNwAazP}GBSgAf|P#Li_4ULPSS#{AoMb3Haz=3N)yJ%iO0w8!LOMt6HG
zGmdv-Z)aWM)^vYphBFpg>}bE#kn*%MeZe)ScB;kF;$kc=+%*#yZ!=+Ps~xSoflmv(
zJeE@*wfoY{Z_+(VKRNCbleeAcy>G`I-;c-pLv_>;eu)}WeOfn6^$+w}AjG*^oAs4w
zY_8d~AZ9F+>B3|(ui8H>_#28ssVHBG!J(0;l=?Ius1MXHm`DV#E-lPkQw!lxUhyzt
zxxBa_*xQMm(nF|0KmYf?`5*CL{_B5Jo|`dq`g9Bpk2>80;b99ba=pUm!Z9i{M=<na
z{;vnE#Z0WNt(w=9(|x7PTv8VP*^fqj!x~~AefUA#x_Kk2nsah_l}Y(;tYONou5kdb
z20`AY$E5i?x>9f21nn~56>T!kds<q&E7<4}ELH@|f_K3Q!ZSQVZ3M+zDht|VS;mdV
z#irK&zkTm_@#!a@#`5x#VQXM;Fa`&QEr=mKrW^Pjwp1hBiRnAq22HcIwZ&LlSyUYn
zOabFNvAwDBr?JOei}^)_VU7Bl>iTPiowHyFL1U&nqhY9VqdE2Ppuyl*pCi<2X?&?g
z4`W+#5jw6%R`7tvX41w)c(t!MyVhjFR4Qk!(TAjC6M_~>?<V1!lA|sN$-8=58zNff
z2B8%~@f(c|k6YS{`I(`cRU4$Uv*_bwDihud+F}-xDSz4@T!gTUGX{B&W*1sojA?b8
z1*70FghG^sIf;2Fn=eGA$~kHM${X#veN`(Rd9o`w;e3D26zdWEbai3c6X6$Hu+Six
z$o2~V_F&eKHlvO$9HPEP@P%na&Q3&FE36c|!nH-gZnc&)#eiRgt>D;=wKWSgft8Jo
z)wp&0y5V(t`kb}Sx-)3(vN4D9D8Fvz)_n_h5tK2mD90|F?r9H|9qUM3eRf-@tn}&O
zPTW!2%YvJt#&x-zH|$VOHtDm0-}9~K6kw0}0S!rTb8CK+`(wa^kK@{N!q@3fa>w64
z82;PpJA$u{>$uNznxLv<{c-l03vuT1g*Y`k6Scv9O?sMWh0OPv0GWt25uswg*+AHJ
z=zAs-gj`TV5Wlu7t_d2#0Ro1rZJ+B3!H+P7$pa#(xuy8K>Ah_bQxxUFBns{3$w%)X
z<h*FrfJVEFc%C#=27OHVu!zw8-mboHU<<j)-0{>oY)L==-1G6`%P&WNf6bGi`@r>M
zXCjLY77uC}^&2kj8cCI<4Pc5`Xw`bBo>SW~G1HoIil3S@AxSd}^evtDC9gDypLPLh
zGHyllSNg*5lTR*9`t9WPL76J;Llv-<7vB^^9|=VHc$wZDMRa-6t&`Th5A61h>MZZN
z60iMUDe#VZyUcv{+w`tcZ^VsjH{$BmYq7GjsxKYaaUIujPlfAfe6`$3X+P+{33pPy
z%u6S|$Lr@a@7n%J9HHC(t8nDsW<WmF^Y_3z*Y)$_eGB}z#eJr1qD$`eC4Uapdb$@L
z3Wi^3Hy+bjT8VDrzFaQGz(6(76qufwil6`NPvg&j{?mBvt+(U#H(rY?&s~UTM9;tU
z;`8ysOIKoQYD}~?hs&?TuIOiW(<df7G$XsUIcH-(a)q+eFSg3W&T=d#(BFlq)`sKU
z+4J$r%P++2?6i0Nudl9&u3A=ccVqVSc>Kvvei;AZKm0{}?}tB(x8HulnlGGl!Fe(;
zk1)*yvnP`|jPdco_{A^&G=BbzpT^Aew3$(u*GgsLjFQwAiLOh89rtJ)umhW&)=HF@
zDkHwthVnep)+xMd0IlpaT4^cWHvIcaJjd}TUCEp8t><Gt<#o^TL|(^T3fFS2i0xPH
zdg0s7w?|2^CZb+@F+5m_fB%<%7O%bfVpImlqtrJZr>95anaii+%JY|F=FDmJ|Eaiq
z>3qC!<x0Hz+AHz;>#w^#Y3sICjl^4nCRQJ)-~HGsQ!=wh8>9tAT3@h<Fg+hH?|q)w
zC%r%Nsb0yaw18sryys|i@;(;N@A;+g`Q<%n_=DWNwoajd(6-}=?|ac#>Q$Ruh#&pn
zjd)&dRT-FyN@YZ_J034Qe<_}Q_HvAjjz+%F>)rYYbl-jF)%e*@e-N*|@<NQC8joCW
zzuKpwWP8kD)e2Pi2%FIkY1oWcCL<U_E1_DAVnt&zSJFDJ5>t~SF*z}u%)lu=S}F*g
zV8FVbYGa=%<h3tmEcbncz#dvDW?gk@yspLS%95&t7Fa=JcsDi>GAQpZ)p2j9X^kJ?
zn(;uNAxP<BO^(yJ#QfL(4rcAhKdZI#o-MAhZt_}NFaZ-i3B%froX*5KE!cYf^;hF{
z!POgYyb>?J{E`{oXc)m%2d;dSf!y&i>ovQ)`;2$G(3jls{m6OkM_65sy>5p&y7H&{
zG-lKLHuL+QZ@+)EcWSzR-u(Avol;Rc2x9L~_jvkLn|9({xJo%Eczq^beBrt1ua$h>
z8tqG;lMdQm@MM;jmxNC@VqkF80w{Pe;2*U85VnAqYqkCu9vafKaP<0G+`e_&S}q6y
zY-+Ht;yB>$8`p2d&6_vl-~RR2@#|mzR`W`~@Xad<H=;BVHuMM<^2)1dlZ0IRYXgG&
zoccMUK)waEe3Q^TuX{`&az5N<z3#KZ{`dd>Kk?81{4eHdu%Wsx_}SgYd~RzKv*xB1
z;r?Qw5F^T$4b%u~mzEY(-^FNd?Rf5*yK^VDHXAmVSFBVlbhGeP86(I+D{Ecx2)m7T
zqhPgA$SX{j+u_EI>v8w)yoH6QXJ>8ZF>RRTy!fm&dXi~L#$~t1OIcwDD0i)@cIdTm
zrD$Oan>fP48qOPo-+;N(-~8%d)t)z$W_R@FF>{C|BF5QPY&F&l4?TGVVF*bQ9HcY3
zG$!|xAZm~Cs6RBH5MCqrE~qcd^e=s(ez9ht(q$~yh36@=j~h|guGZ!i#^}~o>x@p}
z3(h8-pPN&>yy7G9Yw_BIpDbuosX6w<8r|woC9g0CM|{3w1;J$jiohCG2=0AENssEN
z@-QzWOe6jtT7GDO?d}U_q%+Sj`-tg4gswI{sJwv_@Cl|nfg#KY8mAmKY_b!f9{rFu
zC}2jBdw7m%b50yI)NaJZ`jl{@#u8yM3ySuaQfk#PdHQ2ua44$6*PJ=GCb;My7>kL?
zS@k(ubI9;^T|eLxGm~p8f~T!5mjk!}p5TLV)|!Q53<TzFYpl`s<h2vyV<R3*Km5@T
zG-qFm;o-qJd-hC>j*i6Y`ij!(ilO15=u^8kZM|yC?d*9HJV8P&T!G{E_<AF~yZ8J>
z{gb+r()cpI<Gxzg5qx!A$34sm-(>PVQ5Y=6g)5ii*%zLV!O>w86-?#`ys(FkV}fM{
z*=%eYK@0tAGD*7=AQ+e^Y+6DS0c!FVc<rY}>VYN?Pdb{>5m;dgamNQ=ve;&qVwaFT
z#J?tbO^_@GC>!q+QJ4-Ig&1OSfVmY>8EIFH%7n(3)7svT-KL0ubxZ@OjLbz(o704T
z>6y#%g8HIT>C;4-?5nf;C=s}5g=xXmBKmQ(vlkk)xh7zQgL@EjA|fE^QZwjEa316x
zmuDx@ddcK%J_6}zF{G<MDU5|nq}TGazivp#am`H7)eF#8*_;LCavYN5Dq_c12S19J
ze7dCYO#QlVg-um8>xTgLgh}-#0QVGM@m#dD$mf9QWf3hmZr+Nyxw|CNaUIuj9rrZ2
zj>cEZKaM*o-dF8Ua*rzABfp)H{^xo9eCFL3@$LK{=~8z0`ShAUt}S<T^UXMDzIOg6
zdi{Ln-3k7ya3`gE&+C_)_Or_E`&MHn3I`^!shU=~^(UGSVT69o9D)wZ_P|s`XtW;#
zL)G~CFa9+C<A40i_~DO!81KCEcD!)q*%+G~@nQO<GN#2N`Um=>$nIoOMLxrKN%RkC
z<<jBrFvNDYIXoUl*Ov2Smo$g#+wE_Nc5j-Al$cDM{gIDCsTPxy<8ki%`IwnG70qT{
z^yqfH{_1n_=9{m?xr-OHhUkld!2t_kp1piA&YhWwVm=%74R*Be$FtAO#9#dRPvZOE
ze>W~(ycDyuXX4VO3vpR~c&J}xU9`5o3T(QL#_17p*-iPi7QtlQ*yxbzc4(%lX<u4V
z|1cI|hGIe;im+6|d>E{NB=jLHyW(N`A&8(}rXk6vorb=7bD5lb37WIzO#doSd#85W
zIb7>i8e2yX_N;%}Oe>y`#-Ebr-p_}-*62EaZYrMRY^cj;<L!6e^cf=1)$~J7-BYW=
zq|K^aXb)%&SLu(zk#ROK#Hmv=F?;rGJpcUj@#-rtsgEYr7BDIs(cEU&deXo{7z2|r
z(c1iB-iwWiY_3C#4+b3NOKsvZ>XEiMD(^#GYA?c&7B0a%@Azrmw9+{|NZ)857>i^H
zZGSIc-#CgLQPU-TS6?_s<>hmU`pfa=YtO}Z-gr@MGHf&axn4CKOk}}Dbzs0k+0!%A
zap}VOc=gqn;^mj0i!0BakLl@A3k>L9+SZyGFjcoVkblXe5zy<gv9Ix?6`IyyNw}Ca
z5`eXg;ugJ;5$!vKMF<MeK!I_)qhZZ40b0FStD;coi*ny!lq!SPIw%xagL8%v=HN62
zRWBH}tbOP2&YQuH2G8D3I?E@Y%cw0_w<YE|Cbd*R-+O#`+{3g@X2pEAm`cgoIWecx
zc~4#4Xj`;|X0faCRAXdp%Hw@%`gEK=eKyWrIP3O$<&`TcXO_{Udi7`xz7soI+apLo
zh{k%8JYbTiYq~ca`yFE<g~6pztvK%UbZ<Cw-}B^NZn>0B^0@Z(^Ib~o(e81=Jj~rI
zv!DD9m)5)YTssf?p}$&)7oWc@e0av1o<KC3fe4=Z`$wWWIIQs97FIO`2mQ5v!(}>C
zYt2Fmw1Y?yrhis2(Zt-W#sen(`Z(XJ(&uxqN~MD4mfcuhT#ogPhWZb)nu%{SJv|xE
zJ@-s-RvqhgOx7YC#RMpV0E7wsDhqQ3+ERU0!L05XS7;hmDm9ItKBq~Z0}5LzmW^|O
zp;OamRG(36689D{O^5c&zJ(u@vs9|8Jq}`fhcl_xwQj$o`ZfhCEdfiG^X`&aN8lu{
zdxSw)7NV?}TWkwa(C9>y2n|V0BrYs2#HPZGjE=?3=~-*<Y;CiN`oQVi{GL2oM!+fe
z%w^OO4ZD)?EN6lh%Y#vn&lf7npSDhB7vWKHju>M`NyfU$!o0J&u@oyy^B#u?*0MQ`
zGtaYa7-qgX^x0?#&wPHH`h)Y_to0=v1a4tt3g@4dN_~;hM39$<pSH5HqH(w5aY*{$
z6;ice>tI)5wuB4zVpny|Db1?jsaPJ0EN76Z+|6b~-x}^?Ob#M^+1uOFoUvhA0i4V^
zgFQLS4Jv)XRI^^U^(T0INpm(?YeKNq7Od002os@Wthw65e9q>9F10~UWl{U}ayFUp
zw&Fk|<pjuVGH7fp#I>s*#^&a#=6vb{f00c&Nw9?0-NEjf!l4<t6l<$@)ed*n?wo<f
z#*Kq$Zlc}R73Ff-_>K0bpZjV<>PyacQ)=pq;o(!FmnQVR+nQWB1cTsUc$Un0mA<Ol
ztg7^Sj9+P!)E))3F?`J}!x5V<tfkm}5S408urMQdn~AAO!Q81+@r=gSl^3qWbI(63
z9QtDX_{ZOici(wSb>54sSFhSEBRtg9e=QuMhT&>l%DWaGG*1uzC0)wr@%>lrPRip!
z-*J!cI)bl`>$nFUI7IVpp;nBEGm|lO=5)+z5j!Hp%x<wgO%8iP^q^LRP6%AmA)1=d
zG~qExvBSs7pw%OS8<PwOT$p6i!c(YO=`i8h{2>by($U0#=UlEW5i=~f_jJ#$CU!rR
zIm}gyG4e-<#ts`V8oU5uv6j+5fWY2C*rvR<Fb9Fwm!`t4ov5#E#7<*J<$C%YMB2>E
z>A3vtvswt1E%0NKV`1(-P@^OG@~%E619g&?^Qu3@%`PF$D1e#59PGs->B{rYltpU}
zNt21*>6Ps_D`Fzd)772{S68&ok}!<;gmI}H)^MO-CnpafWJUqu_$S}ib1O(Itm3rp
zt&OHIzV{#RR2M2iHjco%RHya=qFsiR7VoJeuk?<8UD^3{`{peXQP+>!qvJZR<2vqJ
z?!HmMSL^nP-SNfAgZ`UzUnai?%l)L@jJgNBe(WE7-}c8HPoMFvaP73)SN$A)`&#`8
zf2j;E!jZ>u`8@w+9ch?Nq??EPRyqCq@V|D1&(!$|y?##2OCR`j<oqTDPx1W5EOS#Z
zCY8)Ifr*4rM&CutK<f?-^u^gTr($TR7Nt@?UVr12_|rf6Q9SqJOEEriAx1_|MX`*D
zqRrUeT#5DNh1jgG2PU3i4#NCdyn83^+`bhY`VJ!;Q$~&EhMqTKZEZ>P<4#oiIG3U~
zx<rQ`?l)p%<!<c25El*ZL+9Q4IMMVyy$U}Vr?q~VJv|-IJ##kB%}z#NwQMFQw7iH^
z)AB>?az1@3&YqpsbtcZAn~9fSx~zJ<93$i73YN|@8XX;t3+HF!%<O5OnUw3zdp9=p
z8_Y7!sljB>bI)9kKmUuL$M?VYZd|^6F{Y;`qgX1Mfo~y!N(hrL3CITH-FM%NpZw%|
zaplVMF)}=?{Cc!rfl1GrL+c7?LX(Z6A7R?NLnwZtv6D|d)^_baymxKUu7ve7_j<=#
zNPW;|(mK4o(&xSZcqLd6ecg4q8_!-k8-M=upTu`we<7ZK?oy1OIvqVoyP(kzx1+hS
z7>kRy)UWDG)q{1xQ5ky>xTrk~<ro+ok14Gs&Z&>ioSE_YQzK&|QP3I-Aq8tT)@ELl
z?Q<>|^nKD`AU$Yl=xfS{^Od%dOJ&w0c^>me0a}har+O#YIpkhf*BhoR@AN~P^PaE>
za&RW)6^4Gn(I0Lzr`JkX?{%Ru6)OEz%Efs7m1p8d-+v<}Cx;@PO?9X?$*XQEM^`%k
z88|=yFf=l#aWt-RR@NG_5&L^hZ=_62UCa(5%tE`Wt`VU2=wVks4AM2NW7k|>*1oL4
zC?CxE#Cl3KB`%IK)3#`2aE>15c=_zETp{x6uY9RLa)n{ReJwCm2BVhzSWhJ_6toL6
zQL2=qR;xuX<6Ft2(bB7NhUVthHfKWBtvNtnXa$wvO5x~dEQesdg?^{eX==I#JVW|P
z@I_zI_^kQ42Y%JwsuyL)eAvXqOq@M89n+`A<I<(`@|S#8)V9{&t7~gM5`jGNq~Jc*
zl^0-4f1oXShocaLS3Yhh`n?r~f8QQC;30)Gd=T#md~!!OUyDm+BM<qu1hD1WYge4Z
z$g}<iGzSJM*2Y3}1)&#XKmZs0gCkMy8&EQb9`n3M;|Jl@`r5L!Y%*#)%=shS0cX<B
z+k0Eq`WzY>jbV9&Q}74S4r0DR+oq=e93P*sW)t%U0*YcWnb_OWT+wK5MqPNAv$EEe
z{;tY}_6{Zxx3+gJfU%(0V^HPNGXgNiK)J6LgDU&z=y;5bj79&zpiSOYDg#af=rBJ+
z@K`LBqfg-x4mTQGHZ8~4LPHTNHhq;^jEzp%q+_{^wiofRGK8j)a8|e0`e+(yp2)Ot
z1><;i1x>NJn46o6#igadauGtWQ`4uF7n`<eWcmfjN-#x-c{=8<JEPHnb{F+Spxa|X
zSH9S<x?>(#^P|GC0U1GeZFnMv$4*5-<D2sqRTR~sp?+8w4l0<R)uXXwZ7#PM_(5~N
z!tH2IuGd#R=F8=((o3dAlhYNAr%XxttG<d#y8#nyj7H;=cHkY!GUuRyME-H0`ECc(
zd^_%gLa|Tnz(#;nM$A!Sb`oKZHM3ObUC|R;&2`6N%p&m7Z%S7Hh;~=d=G&VV4l!l{
zObfQ;Xp3HGfvjkg{T;#9MxdR_9F0(qe(uQ@1e0ZNOvq<B*KXUwON99CbKejsVNR0G
zD*4`$>d8igooLjTV`Fte?T(OhSM9qJ8)`euFmlG51)cNrN~=qF9u2z#!EQSD3)7DD
zE5ol`#7yW!6sn`f^_iXn^{M(teTyl_ZqYl?F}Y$@&*Xoo@s<R!)~xHws*FulU1Lsd
z0uL{pgO=ulqydINu}3shN$@`u)&5ay>YtgNi0SF6n4Fr8{=Q<&&EJhXb9WV2kfylI
zKXw96a7BC0a|iRwgI~15bzH~EL49-tUme$R$6QygEBc1|V*1?am^pVgY9k=xJ|XsA
zAyT1v2xKANBw*0AETRlz0z%0MSrZQnWOjwH+s5V-G&%0Iw6`Dtn3P#Ku^R=4@NCaQ
zFebM>6DUlA98^iVAbAJ{4sJ!2oE=NOg^~#cgkKzR$=P7rBIx%-)FCw6WkD=Le!HoK
zJZkgXJ5gU1;jqfV#cA?=`rKiRO-{u{EhL7uNT>dZi6EkjMwRn!Oii%K7O=jlPdHBv
zAtQZ;hftv}5+TPmErRGz^5EiyE$bVlYX*!p$LM=@7t?|INS~9h@*ungQOeROD_y5W
zq9RZ{^>XX9lXp7p7jkLZPCH8Z#B1m>O5mVg_@r>;MJ)1Z698?Ymp$QH=@KawhZ=K9
z<>d@H4m!Pd?V9a-@2QR**Kr-!aZiE!)(gJU9PsUMl;w;1FY1W-u-JZoLS83%{e1SF
zqaS<bs`-(WkH7z-{_`9`Qf|jQagY7+)@eP$x4*aZ{i0q!NBKVdPr>AmZXV{pBIo4v
z^Pqn)k3=tVlRlbxBpL|vg+7H*$9Wi-CW862zrQ7VxI3P`d@jEC?wj%QE0`-BjnR=>
zy!_JhF)=wI+7Q7OO!<tNKb-ZW8zsnjL72y)Jr2F=e+!K+<-^W*&PZ95U&f5o(voQ;
zXfRA6qCvE|u@WD=|C?CT72z4QcM`@yZ$i(biN#^@O1}%i!BkAoo{fQi7)6*mFGnWd
zr|(6t7n+;PqTlvoXnZQpUN|2YE}V1uG0DZ5F=(dj@3P~(?sh2GhT@#U&7Pi#=~JiT
z!o~A}iKdP9<#_Xrm*dZW@zZ$cyYIxz`3rIG+-#gbf7a%M5bQkn>@zVmG!P$s^nuFV
z6+ivykKz}<_>=g~Td&14`gZ2bnK+|3XV0FE%a<?2aDO#cmltgsOO&tciCHFZka8bU
zM^(iRCKeC$mEMa0H4~pVt+rN-6i0tJ)-|~FlxS0ER)>IeRr`3;m*T)YtQ52Ho!76#
zyYIXilWL>kv5Dx(vooE3h8{-%v#mbDs>P1ln?4csto2!!z9CAbj}V|mH+|LR^HGLI
zPpR!M#N~^Z;-cDfbZppbsHNo<Gej}hH#<9Np(?x3i=}+b-?`&5VMT&JI~Dq9Jhq(c
z)T-N&1Fl`Ib{i^$>TNTN)So(&q^_Ve7`=x71Fho(y9j7_Z<Cg~A@0#;KN?QgHWQFM
z+|Hu+fwPO}X5xF_eLXImpVoUfu>z^j_VituTK3gi+t9~)(wed!jm_1luP^Df)|K>&
z;GHvXY>h<ixw~8UnM0gObfEiY1I>(0#aEs5i<SdxFv^+MW_abHxYX;AH4??wcbMO~
zqSo96&eg$$T|s3*fY8<$lUJEDf??LouA>EYmFOFsjNam?*2y`QW!ueIC>EkSlhyi$
zHbxM4)MKg(j99g;JOWMx8(TYCd&8_&nrM~edW%V*=e(&6>u1I|4mgqrrg+b1hs#B+
ze<&tTO?w=k(-@qcIUVIn$=anb`hi!1CkAlF<w~$X8m$u2DC!yar|Jp=J+1eED*d%9
z;(T%G<{k{T13Zpv$2m&up$?eCwc+<kd^=vsNB)8A+F|_3aS)E|3J)Q;T8@o6$4Q_$
zv!{1!X2|DrTJvW(E2wHlJVL27XQ+)7pSD7y7IT2>!e<DAIR6p;LABbiKFC;z!B{~P
z2iOM2%9XN(#Rz8>7w*LB>auW4BN|Q2;f93<S!=%R$F_w~hvvPkEHBxNZL!oB#iGLG
z5F$|?_%sO2nihPNDt(4Ad~*vS$yVeuf(79$G@IZJ;f97sqT1hYm_%5|krrjaJ7xq|
zR@PLv4L$d1z8Q(Ws^%Ku0SmeW$J@<DV7?c8g?T|V(OB!V!5YCC0zu5}A|M**AJ80h
zDW<2-sN9?r#yr}i)@$)M<UkLJhrks+0)nG$<4V#%aFypAEt|Fjhp?&pFpXoI3hq_f
zS+!knYHNgCgi}YGN29)Fu8TwAn*B}T4B!E)M#@;_#^htUZ&+ipqWpzB_BRA7YA3@@
zD;<QAnimmV6-#J(SM?f!mBy)XjW>xKXDS=vCEAIw%<0sDbGy6|Kw}v}6T+KjgR`&T
zEip#X_G`K?NrSoWfVB0Vd9qvccel#4yfhzcf*CZ!$j4gh3Jshgs7G)M{)E?uKopHV
z&S~ciHs-CK9^p@_6g%FS1B{l+rO1^Bqg)+Pe{#ka0@bo$s87iuFx-p!+Oo=5S6c~}
z3ie13{Feo;mHwJwrcvKidmmVHaDINyrZ?d+GH;?a3vR?*D4MO^xuhMB_LlLL9nZA8
z<{yn81yKF&Y2g}YnfFEC_sW;lp9gA>6~lLraX#(0rTJ$IWZWvhZq@I5Xbr0$5k&Xb
z24h=wxOVl1XdU&pXgg+A+MbLllW9ghG1l$;4R>_YaUJK`xg+@MxQ=VNOd%V?lS45v
zb1J50r?ns&iHr!UJt62rAyx=-#0^R4Af(F#sQY`n3e&C<cG)lyVpEujl-+|4(?ZD*
z3`W*M)Y#nJ6G7B=Rz25b$~jznLiC_)glFtvg1}=J9A^I55eLDQ!z_aWL3ptH2tn6Y
zqtOzci5SrWkaV^-o3XOE5={|}DgUR_^;K&zHGL{(re~5Z?$*L5O@MOrfe$}s7ZMXQ
z1TH=at-NcMIF5I0LR6cyAGxqg(eD&55o4*JE#P!|>Pw~NJZMnAgQ!V;jKB;S;hBDS
zV&v&ND94na-|*RvH1v*LNjxcShzfsKU<e6z6_K);4>-N^AIbO~f-`dAuKkU2rm|9J
z?x`O{+uG_{eER99@$ttWd$(H0bzH}F+*9DbF~QfVQ>RWeb@+K6chdWB%BRKPz2`6L
zj;HY?UO(yngg?Hu-GlK?^e6pkKGny2o_Fo=U$fUwD#Ly6T!bTk!mAVgSLK}FgNZ-j
zkC*phUO)CHeS2d6U;*&0Z|T~4ZI>RPVxYH7gXsw#Sh?667tWoDS6_cArl-fFTF%Bx
zS1!f7-+MDo%}n~-m%eg7_IJ0uAqveMXm=RY+fB4jnr0HgoL*a7jyi|Y?{b(u%&hJh
z85#Dso7b<#-0fT52~QbnwE+uLu6^=Bz|_y>vp#g5UDGgF*`1BX3CyX3z4h3tuPNQM
zyO}dLQk^o9$!k53Q<-`>CyWE$`401DEB1DidGc<hmC4s^;s{2PwNVZ*w<fxNyAJuN
z{FI{T!&+@L&R=*gUVQPn80atCWZh4G_-?%L@{7@1EGm!!t4w*d!|*6!FU9jOyb|Zn
zo{nN66X(yJj-UL=4`S-nsDkbX=i^{{=+qP&%g)EkFTND#&dkO@wG?dbTHRQWjXJ_N
zcABR?O8urg{X@6nZ3V-cGd*U~cj+bNPAn-%;MDgBZ1{toCN<VVZFZUyGQFB8X!-$W
z^FXZ-FTHRvPM@Au8=(oLcJzVy3WX3sd1f&ORw#PYA?q&~jw?%xYQs7y$&)r~F<=X(
zA~_&X$=Nj5@W`Z1DFPS2{?%`_Hrt8UUVScp{G)f=UT?kqw(_5`*3hk+x6Fu4f}m8l
zHUkbv%G51+S4mgs_;mFcXwmwqTzJ}wTh(91wWfrMH+%Y2yz=UcK6|UyUsZc-MO|ey
zlU8A2h$AqzMi_BhX*nnQL3J1C#&eg?#&geIh-#(R49;>{Ysp-%8IiP$`&Myb-~*4u
zYYL{fG=_jvghTog1+_D9k2R4U!6xRk78Y*B^5UEYvA`4zW*Ev|Bdao`$66M%WtbpI
zVf61lNM{fsfC3gUnHXJVHAd3BU<{@Hqu&UxFsd(&L|XuFg<?6nvlV?qP^L9+W6k|&
zxPbABCYgmAdBZWxXP=`a*rHxePI=I#Zdc{Q`0nn`s;r5zn|epkv!^)12WjIYFw9zw
zPm~jQN-*HF02QWG9*XJH7u1gvYOA&AS0A35nly8JV?#IylacNxmBH&i9)TQPb?<YA
zxX}kzfwU=MT3y_7>AS-?>~TlGrt8D}XXAXnBhDA`t=J^cy1zdyKkKi(p%r1s#>Q66
z%`L|K+)~WnU67wMEcFi{T*?dQSA5nYn+EpOKj2|BYv_B%DMC#^tw(V%nRnyHHHFPv
zSjO7dXF;jI!GmZTGOjm;H=0{@t&`E7*;3eEw>^R<&IFs9Lfd%6xZOfh#)M#gYjfRc
zjgL)hybcB?J=s_RPTv>oqcI7du2ie)+@xiD`_{EsTfVD0B^&`ifjJP|TqvgV#NZX-
z*<g(@XM1Nq{`Ft}CH~<b{vqzn&)fP8f*J4@g5R9xA<m3Ks78DCX+F#4ODf~O$6|l~
zKorX*3x|e=&{mv`$*CEYd01_bG$|Qp!YGDI(l8F!AO9XP8Y{rDg>>pKHiLkB5k`SC
zc;BtDhj~-YNb+nJZz$-Y+boZnK(!U-8;OrFk2KI`+S*!=ET#f`^2TkHq@XxGy?qL+
zc~|rMPIFa$6J+VS+w`H@uGw5y8}3G5wO{QwXe~0%9z@HP@#?vhx*USD86W%ljtNQT
z@Ga$=)qKpEgc_Us%~)SswXn+PhtX%-+kzYQqiBIbL2z-X((I@m1Rn^15d33CnKQ&N
zr;Fw_LPvO0@c8H#=J#YOGatoDzviHt;fA)}*=|I0b5&_74dEy@wzKgBxI>5po}<2b
z#VrUH67B;}iLO~k(=nZ02p&P+<_!h8qSEdbKHsuNFEkNC)8(ZlwfC;-hDKo{78V7E
zYRZD(g1Umo!RrYBb_Dxq!+T?a$_`#Z`^-mm00Room;;VI*8)R5XqD0TY~)cra+nd`
zi}~B1Mtv2{GooQ?7p<_Z*jQh*U=f_3>B-sbWKHb@U-W|yKGIvXaN)D6x_FSc@TccJ
zzeq63-xqbqanVVx{r&6qkK;~?_xteQl>0JyLUDBjUme$Ra@kTg#!ipMnTzLRXmmJA
zTJYt1waC?iD(w<c7!d*x7W=}j7Ov=JTbR%*0?Ug%A#Vr@4kfif%Zmt2^a#F$oS9g_
z&Fm&AX#!`*4GTNcfH-3(9Rd|JaS&jkwdDy!1OXE&vr$HAkXBnWXj79ulLT5xOhgb^
zTbm--HX;5r5sNr#Y&5heS@-?Z>qJ<n4cPgVcCQYD0sX}jp1!j?#Kao!*zx6wnRa1O
zp0vLB^FR0UgoQZpLPPI;E|>x!^x`}H!(!aw`GbnkI}@g)*}7M364i4Ghr`4I<${1u
zC2c!Bxsd5L{iGLE$?^GD*7Us>)lBRP!{RK}SKm`G$2pn|AOcdNj_Z5M<=Aq%he5$E
zp!M~Q_~fIH<AV=B5xkOF$8}uCb=<e!ed`2YZIp{UdiMx-Jnq-wkGqq;Kd%3x@zb1j
z^0)gu>3@zt>E+|b{W`p7*au#w`XnaG@54V<TSMo3C1q5d?^XRX{)hs3@Pj;;doHEt
zJC}>Q{V0g`o|qVk-rk<rTwji<iQ)LkkG~f``^8V=#TTCQSy+`~PmGNXdzW%=Z#s0p
z+1wU=IOLu9^|iaPIQL03>Z_5@Wkr`I(^o4iOA3eCHkeaQ(e0r$dc9M5Zf@RXVRIRW
zyhCF;UD4N^VKFm%Hcp?rY=+^176LFJdy6pPGNRQnLAB-eflaX;nLxw>!<y;DOqO?T
zfmoN;0i1IL!wj>0iD?OwowGtR*+L+kU_H>&ExI<FGmX!>1G9bXoQKIiJu@MC_@WPm
z&m#0f)6E)a(0(}bf*GH+NwAk+dpVwa=AyN~AWvb!ac&I#0L@J3o@_PxY9sN?^Uqkj
ztW+wRA-1@<==8j?Oyx>Dx(_0!I!ufW#ksSmea09j<}eS1768m?!odJ-Zycj&P^7Gv
zEGW`*8-CMw#gP}Ks<!No$%)~3;mW1Bta1+w^(!1{LenES+>E{LM#_iX_F8}C3W*6#
z`(qjm?GWg4)?o<dpy8=IyZ+VK)Y$dH!F*d?z8in{KmX3mz5na~`ipqy?Kfh4VlsM5
zV=+25tu`4~di&NqLlCvOxnbtDg*N(Y)93By<Wl|AdabU}8qV#QOjaH4qlw5TwT#QD
zb?3oOOrIKwpZ)Yl@$dirzl(R@eLK#cITO`tHA=;TwNYSh!W4zk?^H$qlOz4(cC^rh
zwa9)<O^ztNF~K|M&7die)qBFB@szY4&?>Vw8_aCMG=dtN0rdvXJ-q{tb7lft+e2-?
zzP1>v9#`!4?+PrFAW%!^q5)IgdRLK0+fn_tv#Bf*M5rxPUQ8&?-?^!FXhxp1bgXgH
z9w#&wSA<cR>E{{G`b%SY3$1{h`i6~=?J);~HCa2cnS1pWxG!xY6}~{5C68$?G{Y$4
zuGaUtq!FRKG|aq?#B3Yd1xI58&4$PXrXA5J0X{J+XH5kQy>RU{HtWP){ZnR*lWsJU
zioz>>@xqHQ#Vaqr6a!kzuB@*4tYsL-F!_Ol_9jX?F0JR11_iHxE%O3sRlf1w?ZIt=
zv6I~Aee>1x!{>PS#l1LR)ZLqhxW7NWc3yU>8hC>KKp?xSHS3+ZIcs}?n<gekqpwo3
z5I4h^hY7EC2R{4y27T@mW0iyk1_mvNT3uU>4}SNa+VUW(`i`~@JQT7-Ly$IJUtfv2
z`P;E@cisYx;gQjpnwn-vX+GGpSv)kcm;(rpFc_^vO3fHzUP7A;_~48v&Wi-+fpZdH
zV$a_g>j-EWKRbfgY`&oROyj4w*FwBrl?M}dY}jB7B9K84hUFi$wSM`_U&jCZ+rNw7
zy!U?GxsBi%As5;@h3KzURK6bLJ%miP8g<PI-Z8(|r7{-e%bL#yV`OYR28Tv1P(uj8
zd}Lvkafs$1^%wYxHcB`;eYfCC{lXbm;2O{lxC^s{>uU=R<8#JrzEHEpep7W-Ik(oL
zsCnMzQk53MX6oBl9n$zODW1l!=4ph!8R0mD)ZiNMcG8w|a*Dbg8*3}ZMF{%3*?6Am
z7Di{&K+a}Tqo=Ak!ZQj(It-tL9}#Y&Ig8Nbz~)fV(Av|ymCYcO<Jgh|3!X4X2uvc>
zqYUH+UaJW15Ili{(y;9ET!|(X^+HgF=2V~B7!76myuQ8|%S(53)wmZdAkbnwBmp7k
zw4qIVm;_ClS2$aYP46B9!dLrx$EF_6F{BaDC^UZDudA%9${%_|utZtaXTY2{%CiAL
zeGA`;v%NUe4m^$dU$hg!Z)<`n&W&R{fya|T53N2-Mv4Yeozaq~`4HrBb}DUmfIv(2
z$mY>J1Ak+23e7RWj{BI+4m;}OB*+4W?%cX&4K84=SgN_bwiqj#&oJ?9As;k?aB!tM
zrkuCq@BaSp3>OHK&1s_`YI9g!fF<xroA?IrcL|?7iEq1;zTv+qcT(OD`frOnscc_|
zKkiQY{wUuOe05yMiEb<pl;hO7nK*T3CaQx~5gNUkfVyK(v)HaCe+Emp5V~zY>yMpc
zXmx>QA#fBl$|4>xduSWaBEBGIIG`5-mECSk3`}fHM7nWWOk_+znBqIwWf#nzCT@0A
z3cYJ$vAt@2S5y-&TbgK5QAhiX1u+u}3vw1c+gfmMYtg~Nn;jVt`w%X>TRYL**wkdc
z9$On6FsWJKDR!u9v%~6*H(rYo5uHqG7G}vqw6j}={-V8c)R#p>`&p<@FVn)5KZJ^y
z&%tCzhDE5pTd?IY2+Fu8{j^}GzkJ{7S3GDqE!xvWOdeLF7jWsdoZflz<vA^4c%dBp
zoglXF&ka{D{OUiYmYzu)V4w?Ket8^8rhJGXFF3SDn9HdYT11`OKmmzIy;!KPuC2$1
zAAA@eeDFzYG1YM$*Kr;9t#{w3;H!1}NC%E?6W5R99w%Hoz2jkzzk6K&MdP*oN%7n6
z!S`*ipVTG~`BQ+$|7)e!_7CRut?(}AeZ~Ct`LC#q`fc5FeR8iK`$Q9wo*qS~uwgrw
zJ&b3cJs-dL`Jcq(#89j)&&8?9p}6w=C7*RwtMx^;8wUKQXqkQQ9)_U<18-`2Hu{Gq
z&4Bscum36j?O*;TcAE8IXFp8p`sTVdOE{<=M)vMb(+nnPrDg-cL3d0|p7L3;<qC}Y
zVhj!q$JqF(7#x|1N_8-@y*1Hym={NQ0;31*6f|f!mna>eZ}W1l*lsj8=VN1S*|a8S
zcO>R~tI}wGb?5a>ML_VOGNK7%I`NQi1*KmxgNgP*fRvcsc$V-8M?}}F{Ah_mbGC?0
z(|MwZb>9{S;K*0y$Y;IAF_RFQQ4VIJ=~m~%JJQd^;P6<y{L0HQIW`iXeEMlDEG`CT
z?7-ziNKq-}W9HOY{PZX9#$WyA&*G(*uBgpw!Pz6VS}CSZjmN-XzZq?87=#f@@(2Ps
zSI3*Mws%xt=;%bx!#IU8jNtS6=P$-z{ngLoPyg&EacX+nRYyp}j(V8W%?3OB)prL7
zYm|=YU$i4IPnYejL~URs`g{gZm$mn-g=D5EJIvKSWNHWf3xj+w?%ci}zx?IDM5Wvt
zKluKearWG7^r#QeCZJuawOYLT>Xmr=?YDeB+nqb}v9Pe<^%?mjr#rQ0+rhx)9i%xX
z_e`SG5bC>HUrB8^Zuq7TU~nU#E0?nI?z^v9<LuP*RAdXiF+4I7S6+A_zVq5Earxpo
zt%G1@vxZ4y3z~ml>q!`-<UxOO#?{`=rUl5SPmf#hh1n_uhG-SghTB_px9RRK>&GNW
z<Gi~~v=(5nDlwR-tUED3md?q80la1SM$59+k9I{#ZCi|NhVy+0Pd}q=RgPStqICA6
zSzqzGHLJRIX9e|2WA641m19kPj#dTd>9qw(s)XBAu+Er2psx`WZdqW3Fa)MPeQj+o
zuanS<>9OXY&7%Srt#x3xa7DH-pgPnv4ls?U@uxVf$9uYSf}umL<quU4&K48gbTht5
z7Pz86fdQ@CchNH0@p$j*Mx%lE+seA`vLS#04#5*>eW*C9Q{KY#;gaCnPXGY_^hrcP
zRN;vj8lJKSA7_xg|NaNDy0RFxYDutKw15nJ!Fg6}_GIY>1KN<KXT}Im790XXyq8Pk
z_GEX2dt0vkd|w>je~u>%fW^~;@#P+TKFOspZGX?T(>w0zKX4v6wM&f%T+kK>HO##4
z%~~TE;RrmEjg2*LY~ZY`Qh6XsX6P#|rBmsvMX6FVucfiEYT+*DcyWFf=O59~Sm9y(
zZRq<NCJR?rqfdB$_RK{KzgRbKHFsiga8PNNEbJvcgd~{wgD0>p+?O^iM{I6lif~zB
zw*8IG$c(eb#!Aa19i|)iqgbj$|KP9%5ZT^djh7@mMliCzx)|Hd4YvV2q^!mh0s;6$
zXuo~%;fHbi)}r!ZPTUjT$g52>Ce&UClsJ2<+*eik^2)mz^}5Ep=0(>0j7_xV3d*-o
zC~N$2wiFu2d&1%A3_vy^&<Ea#!SI5|pzq$ypg%P0(40eiE43fmS{v57#KbJ~Gv`VI
zu547eaq~mL`z_^{jos!}EUBGw2oAv$2s#j`slk-C#;58^JJp3Zz*|7PH#$%rg<D=g
zIN9_D4A2G|oQ!|c3baQxkI}bif+1-0h6B~Zo0?U=1ZyOF80%|GN^{%&fEF<CXe-(p
z{6RSC4KE07q~2UnX(oJS4MWPW_Gz07Oq;P0s%uPV5`J1)yb*V9eH<Hg&Oy}Nptc2u
z5HJ#+c?R>W`JDP&FaXT<ma36axb^il!#Fq#SfCBqIIvZ}t8y;tyXI$=AB}Xk9plEs
zmG%cmVJ1{_AY&_&_a=a9|A=t+h~|Zo#|dy!8yGU(k?58leXH2KVxcnNd0(js@0Oz{
z*QfT{5uLSWc=!2W$`heK+H=4U8-Wh?(;1O$OyO*=#CJ<)jP)t+Oo0t82u63Df3eiB
zHbB#nE=U^52vhIgoi}`=X~$T^#Au>n<W*1Z?EFZ@kKS}#$8`i>9rwrWdU8E6FgBQk
zUZ<yp=xYhdYGgoIGHEe+_K2XeU_^-8+C7?#nGDg!V)BOoDGPZ+&}?n(80j;K;h7xq
zCjOZiTRT%Npi!E<XMsZ)6BhCoXz+eV1d<33c6=bnLJ-HK#7-ax5OyJ<!hS$LN`r}c
zbA3H_Af%KYyF+)I?1ov3^%ZtSvZ#ByUA2EOo`2y=y!y&3F)*Y>2~-Hpr8a{oWM_#;
z7=>&Jh9J-o7K*-U6F0OAT6jFu-VhUM(GN<b-xU}RjN-fR^k)H@-utXCu9}2N)WRfk
zBuhteMt-m`On0Q&{@!}w3&rz9f9xGU0>PGpv{0ayBc~+%O6j!)YbiQqP-sy&Nw}Hb
zro2f-<yX9vBK5)s__kHt-~Zi*@zEz&Rj+$_bzH}FT*rNj-8Uxqnm#rC*I@F`{S%&L
zk(ikI=^2kZ{`@4q9k+db@cv2s7me5U54wlFKkoSKASd|rt=(pi_WJqEyJJax#lAIn
z{{ebe@GEdm>VdSM#OvptZ_z<)CW1Pdm>7(o{n-!WM?d^t3=P(z(AyKkgJla)5FGie
zt0Fs}w_<I5!^}Iz9OkBA5OtZkv=^U#^sD&ry<d7)JUf*UcA@R{!TY~6Jx$hIP0o^l
z&W6c%VCG=0raX&f3%;=Ou-Vv(dcEPY`s^1`Vq5i%Xl$l)Ud+6QbnHS<rLo^S+vH$x
zQ}p=VShzc98Xhe&Xwk&jf~l_p*bxW8mFTk!!UT54s|K8rq&&O25xBry&nh1lD53;S
z;)xdB$0|fqX)13H<cG0E#T3ArCY0w$@Wp$yX0{?=%u{%cJs50gJa7hC(rzI<8Nl@1
za&jD0k9OTae>Ij@mYveR)+XIudvRXtryu^{o%rr|--uJ^FUH8&Ow7(s$A$B!<NM$L
zZhZIMcjM)kUy2uAyb{yX)6v&g_Bm#iN;x=hhs}e-L;W!%Un;^tEm~`Ac%<sI|Kcxx
z8gISzY7CB!MX_9oVgVs%L1pUpc{i-n;4#z2Xs@vjJJgy8#vt#t@KPN!Hlf9tItVja
zf3Y4z@OQ5-s52f$^NQQ&op;`fXP&vF{4mo+|F(jXwqPog?Nhl2MO*L2wd>bKPcLg7
z!sbQr8f{P8396MWp5#3S^yjms4t8T)?fRF0`SbYy{KvnHH{W{0LPCTgw8zZpsrb$t
zFT|-S!9MH7UG<4zs3)6=oa!??IvnTDUx?RUdo`|Hxe}uzXrN$uq`SrR-mw+~f~+Bp
zq1Rr!66eoO3yxJ66$hFhZMdD?eR-@vpe2<U`IHy!MHrdA`GT44Fraw{tdLK$r6IAl
zvg$r5Ah2@%Sj(y%G%zR&VY$zvoz`6Uc}ZwkGL~U}(jPFBmzNh5|4?aFBa;K}>6><=
zYYogfVmn)@ueJVVydkiHLCu+ZW^m)W1QW^=>nI4l(QLzl2PWn)F=|1w&l^f(inDZl
zej3bnRUDxTX1;84kxo-PBE)K&UBvuhF0XipNqYoXAYIDHS!X*<&WqA}#<SXEf2S#Y
zwPWE-DkBJkai5J^e^F`dSr{=hGZ}Bb`Fgyr{u~}2RG75ZqmA}ec3_+G>Mf<xH`)fr
z>X2sxXL1B+IbPrU#D~_J)a}}G4g<`lckV~uKj>R`_Yyn`+kSWa`6Qp5Q%SdvaE_yE
zdkiJv<DuZE5zTr%CdP+jY;?eV%ej0$qj7OTV|Y!loX#AYo4XnH^;M01&Xmf=))pq)
z5E30k|G<#O+_2B{Voa{Dt$WjQp|>}x!c{&~3cOS5Gb~#>Fq@4^U!Mi5+#~Qo@P!Zw
zfi8mkQbq8iIbnHmN%PLK;R~Jwf-+3)t}bibC@)}VU|`rnFV1c!Zws7+*Ep99K|S~Y
zEzhFABUC{!z2~{F*=WS=+qV=q7tcQXOgwY>Vhjp*u=xW)KVt=<4EPVB$sL77!wk(a
zG|SMc@mcY%3+HobJn0_%g@dm`8kqK@{kxfK+y(opug0{(7cIOPj#3}m)|i^@(R{0U
z1!0%kAfq|i_(S0lP~Eupk@{*r^2%>zc{Og{xaF}@Vg3des0|pe@VYX2&gVi)4Z#z6
zs$P2s=6@lpY-rqk@WK1h&^*sLXl&LE8^ChX76RrNyQ+&KfLGFHl7j074y3AS1ICux
zb3^zi9|J?<ktqx&c+=RWJ$g0wq~lANYj)gTy=;6@;*3uLki)}1!i-==^aQj7V;$b0
zDyILY>O8MLTvr=mYO^mm@2y*O17)D!nWs5Z6cf$hV%n}tI2VEbnqUc<8Ouj#vvGdb
z{_dvWeJ(cYm~mdWkdVobaf9V3G|;FkW+H)C=1evWa2^{p4fwEEIJDHKIk;T*oPZ`H
zbtSJ}HWw+bGIKj381y-Mhe<HKN4quai8KKpB56||hN~ii=d9kt$3#nxjTsH$@}}ya
z@wm&VKcWlGv9v+MCMy-PeRf#Vre9atd*aIT&&C_y5v?(OTCmK9lSTJ6@C;tCOVL}8
zx_JQqJc<8(xhKj0czK^B-XEi5=Isc+I_{6svD<NYax`WypNr9{NiV_}5eP&eCNOCz
z^dUl?g^}<lldvWXCLsg{OuEJ`@>z%qO$vKH;1t`viq8ZlDgi<PffZMF7_xw1hZzeO
zCOGnQc}#q#-C?XOm}@avh^jM$&7LMAh~2!1mug>4VKw>d9V+Vxxzc0;@x0-9>}1*#
zfwZ^1D}rDvHrFtP*qRicVt3)v<@mu5eh`-~UNmC@1mCVzGBv3M7KmRBPP<r$C1MD5
z_4Z^y|MEVgXNZBWOj>vmmp&(yiN0fLSi?(%g8P(vuH;O|I$7fRjo+nI<S2`7EresC
zlLDkPT0!{MO2Xd{y;|2*`L&9CFB>}+K9rHPm1_%DsW%#5&O*0oHz41YpYzbQ{Ve!I
zU43C`$u^Tey?V{{>$r~VxQ=@Y+&3rqn#R<jQRySSux-M@a_RTc?r|0;?Jy69eVn*B
z!X5Wt)IFHaV|%y6$-ezn0gs*@_UG^?r2Dn}De3$2dQ7jV-+HH#o>=y8+UqCzcBym^
zO$5X2?YCZu*I#=f%B5a2qF{QIikTP~s719lV8#v1N6xfeSzEO*3mUCfOJ)h#QNFpc
z8dpF0C~n`p8JnA1-ay6<;bx;Azy8&~nQ2(5_K6}&<tUeHKL4bIIkrMcv{g^q84iQL
zn>2T0b#+DIlUXV=?e)E&G-2eyEW>0WJCu{wNq6itmSf@W?Z6TPjMIW>|J?vWLA}21
zrkO&}|1i+0yC_n5c6(D6;^JZ0LF2)&gQ0_C$?lhqhE~aTb5-k+b>kkJ6LOpK1k*_L
zEP?}=es<f6B=3B9hTh|*b+wQIp%=`3cEs<P3B`Aqa-7j}YHBntUpgPpT)v<+&UVx{
z7US&MQ}M<dug2u537?4)T?h<IF*w{8wShsO%R4wQ5M$$GF*9>2E?>SJZ@l?dyz<Hm
z);5|sH5xzt>5t;=w_lHEpS={{`OYix^Pl}V-hKDYICpkBPEU^q=aE3uyLRf2o?KCD
zAxt&(N3qx^`W>bro7}V(GQGQ}exprcAi|K;6?(ZRuRbcMoSX-xc6Qt1NttWv6NGQG
zv$JvOnP;PaU{rBo*rH{$t#yLh%bZ_4{e(8cTKr0Ib?45UncuiXey8?L8|QS@6XZQz
zh`_Se*BgKG<9FkyfBJ(sbMACtesWrE^zutD#+z@ejb6MGH5llE5d@<w5O!4t7~094
zo7ON{O!rlNmfP%^Gw}i&17GGWya|n;q`@^lIuJkj{&zI4UW{6`Y{oLCG}*yl8=Qzr
zZ8Uldm;=ixuO!sV_Lc(Egw@)JJ*L$x%*tSbs?Xzsv4EBd>s@x~v!>Y6yVOsNX@ru%
zqUwvL1~8@aU?z~WwKy;DPz#~nV$G&RF=a+%cKf4QnZ^R}%Q{YLZq{j8&ePGTQJ$Dw
z%&Xj8oCoH<qlL9r-mlwJ4urv!o3e1;H9}(j3*1SUKHpVes2=O9!WHWZ!&yvhRP7GV
z2|PU5GGiAeadcB2*G_qA9d-aOU`K6&V2}LiBG#H(i)PTqNG4F5N*BQg0yqn&4Lg*T
zvwtcvJ2M#<E(qQ<u3mrRwRrx@6>EcCJbzB#reb6e)2chdbt{3@lYL+um{J;E<LcS#
zUjL;Mw$tb0clzO)ppa*tC`Zfbu}!}aD&6Tf1y7N2H2x!8O80Rb;qDJh+z0(V$GFyo
zfkmAK|D(eL@$w7L#njZeQ$kB|MQhsYSFgsx{G9Obt+;yi!}#$1U&qqoTpVcJH0m2R
zF}JV2qmOOo4edVl4ZNM^_Kx6N&w1fOjW>i);CsxrApk<4iqH!o2xA(pnS5_iu#!yO
zp^d}bRIBz|u(Y*>K=mLAepGgNN6Bnl)|zC4gX6-Jqmj)Q)t@~ogT`!gL;cN0>MhqC
z;cl<SBf<y-P@9dUC1!1zs_?DG!R(otc<t4f<AU&WUsYpS<rI*(Tm``xX7X;|ydHPw
zZdu4xtqpoi!`H&>q0PstJ<`09;2XgV?Evls*4i4ZwDZm`V<Ky9%ZzXWXNPrVl4(Hj
za0b(Qn)57Z5{}~hMdMM8d&YiWxi9*w0~FCPUFa<fANIS<v^i~+$&q(!J`@h*ND(yC
zz&-H04)^wBLo~pR8`mvtEUG>T`q==1(1s>p+;w+p4nUKd!gqn6;p?fMJqT`u(^L;k
z7dAFGm3EhCiT)_`4QfCj9A}In>{L5y9;YA4i+=EC8?__+q8;_wW?kc4?E>7`T%q8o
zC#Shac~H*P)w{w42xKwYm=VsaS<|Z8KdQ6`ErdezwxM|ubB5Fdff{hyt@0zRZK~}M
z2oks3f_XH|1=pJk)_{gCF^>^px$vts3e_K&#ATxa?F+{U&D^rvG8+enJ=RF;$+O|0
z<gq~CX5=xaNZZgZ2cmU$MQ_y6mNagMXGWU}ld9}!kAaiP&vS+5*>t8U8$B?UtGqQ2
zEG^uLyYn{{CT%974@jf2u@tLIbFFfcFnEahNcj*}^WDIp#>~i+#?4Ut<~Q%f`@ehN
zT8-2j{6IKmn&NWX`un2pejN18xi6dVaUA+}`Ums+Cft|Fk6FJX`0BVnI>+QaJUtd?
zo;e?5lUl3^$?s~RgvmjK0{4hN&3l^A8qJ0#k2H~j&=Gzy;IKh!qJc?C6Z)1W2!veh
zmg$v8m;w&Q{;o}GK(Gk`XGJ8NDA9d72%24Q887l7W)W*ZOf%td;3q^EyP22=IIOzA
zzu&tg*uCN%Y9a_A0Cq)SVfqlEm$iQOl#YnF`g%RK>S?#m)9t!M)Ln|Vv>=?Gp7BmT
z_m@9)dtz=2tdfaH76f>GmrwH{yK^BTM1Vq^>7LzMEQTyJBMp5MN>x-$W8l*w0|ArU
zmF!56yjyvacKX6(3(@EYz2~MKKNSQ5GKEgxd~Llt)NMN3fEO(Hc*z||UcdS|Ixozh
z^rLs40Q`-Rd?$qJ-gZjU3nWL<y*xXXIJ4oS4?fc3;D+CIT*q}>$2|q^8xwrZaK@Le
zpXd2CG0sy9Cm;Ftw`1P(j3$l;-+z_9?OetazJD3-bWeES_GZML^zJil(JtV_Jz?&9
zrJwf3kFUdj5vQ0ZruB#Fzl>|gCLQP}o}rJ*#e97Co!4Swe9-hShsGmN+TO+t+qM~!
z2(9kU&zqjXG!Jx6n=uHDzO&bi#k+H6_zd(9D3k73ZJh@*K0am!{r1kT8FsZQT1Mrl
zRH|lDqD2CY#HKQs)-Yw*x!q>|<?<Po`Ow;11A_z6tMcw{ZCa}gp&QJ<#8`uVtVd2{
z{=o29RBGd9Y+{m*UGmA?VO!vW7u{+BRf@}V5?Z-sTAs4u^$ra%E{B*y-!U$Mfu&6K
zkF|$=CC7O;iBW&BzY&`oOImMm9un=X^kCGJCt2YALz&Xr1fdbkGqhV$*>vBNje((R
z%&0A4%+JhBM6Fh_z@}6wnijVJj&km^*)C~6!0<$1lFMh+cBLql$1G5qp6rijE}xI*
zpMO5ioIM+-Pfy3`+36S>n^ry}QR{1&L^6G@@*Z}JPIX~b8oh&|s=k2{2osmSf);08
zWT7eTVY*f{v94_RG7YbP)lgS%_>NQh)Ie*&jQhB(wCPV6wlH4VWVjvOn5(q55VeJ_
zZp3@<{U(0>t6#;+%BuU#8hGjxzI*+UilBFChrHWgZFG8iG=A~(pQz5K%n-y(7`x#6
zYMik)9Myq==qs~6(RxMe&Ru<{<g8)n|Fr(Y9~eHfxq{Y){V_B$60=J4^5qL^+h^m-
zbC)fw931FV9M-V=W>O9gPey5AJPPF@Yr=IQAY?pgEacI8EZ3scH>h?l3wBx0VrngE
z39#vMZ@U?Nm5P~>oN?4>)V+3PjKLIz=}f=Tb!b8%2u7=>QYuGp>l`+i$GJjH?OM^=
zS$Qj;TrUh%G`@g4)^Z545IUd{nM{jivR%O$%1J{KPk*Y<_cvp!u@cSAy1Ett7Ft(o
z?VV<9HrAC^I@2o&py?ylMf+Mex5g4C)b!0lJHY}17Bh00uYFDtWlC{;cAfgg%ym6m
zYg1#SCtFa_*~r?oNjsSF`Ci5aFi1S*0iZgqZuJFPkmJI8!WGq8WYi{{i8eYu8fVX*
ziOZKR#@OgkY}7a6_S{@FnoYrb-kK=_HP?ytspmwcXz&8)4F4*t^20o(^Y=$!5zqK$
ztOHMiO2ep1)WXTMso|G9l4}8S8~<~C(6`?|+O@;C)A;;2E;V5(lVGUWn~B$7dnsOe
z>Dd?==~sJz>keb?&h5B<?V8{e6L1Kv*4<V^!q>Z+Lw@t?Uu(U4HOhU$;lm>yGYMAr
z&1<n>OMSo?K_eA@4QDjb#=E;3H^MWRSj55eL7S8@fj|q*uDhChID2k%Y%F?}_V(_s
z@lv(gXPy!K9?}~c8I8fAG2z0R=N{mOO}RcVu({!5FW?)w&E?@S!J}Z_V<QLe`}!)z
zQ)pWb_Scl2@C90GrJ`_c&T|qQIXKIS_C*^AnBe?MHj&Vt)+E$;z)UO8@B!d=vDt&o
z+E_aRb}TSPm@8OeZbe88WFRovKVXi*>W;>l+8B_t&`Y>7i<v{s+eu?-H+Gtvo=-<d
zC!%j)#QjiG8LQR7DA$G}t9iLgW6s)*8c*O3Hf*5%1S~T)!24(^WiaO_`k>UCj{)Jz
zkr6ca(z#}gSKyd>)HMd_OE#rrb`NvH2WUO3&Mp@lI~dFKr^W+)0j^=LHs42Wf?y|Y
zmY_Tb=06oH32%8S*h7nYU*$^L({!5!Lz+9l*J!%V-M((vU?WUFbywf#yosZta;V<m
zgaggdYH#p!zStLe^#K~Y^^LWt3#ZWsd9@dUsUGl1*N)1KrKeTpxo2~;wD0y-)413h
zbUZe~8|gPT)=fisV~=3UmVlI2PH84wn)(|JD%ub{zq89ZrEKt7kJZ(s*jUHB@PXk9
zItrNfnTePt1PIiB2>+p>QiFjT5oD^|8Y|nI%NnDf8a6nSwMTRt!c@w~d8-A{S7`Sm
zFeJJKvVzTS<!S9bwH43TuV1z4L&mP>HQ@!%uZq*cTKdQ9_8V{smwh$<X>m_l)(7kP
zMf{g>%(xxFSI7NPxo)jQ$Fx{G|J*Y%c52GR_W_d@xKRTNWSSOxY0_lEYqA4ONR`Qi
z-KOklK*KAkgA2iej5YINUm8IcJ3_p>r$ta&U07Ep3a*}%6@v+f2?GL$^f!g{ZI(|S
zRc&@uAe`9T;E-q*6TKo7`ZZyS$SY~G5rX%mEW(?GI$=2!a(64)K0g$ZhA^zL+0a5j
z3no*JPq#aN;ZnTuo$n+DhT4@Xwc8Y;0o7~T5&{6?gasx92>qJ~IEYkK@YUw}L*NK;
zr$hh%=iy&Pg205>Rce~Z<(-hy(>pI*S}>CumYjfs<Os@hS|I7p&nXQ>_xBW+GCPQF
zQw-nAlL0V;G|LMSJ@btSyiWP6?()QC=OlTM7qDePmYD>+<$EIFhzOCeDY&?C<7T|~
zyZ7Vf&AIfg<2tV6I_}%=zCn$z>A!9v+t2Y_()B_2nV@+7+Se!fQzS)uLQlM7_b2_4
z?nC@ZPG9t`o#to#NiM~?-+wmd<9SED&*|Ma`KNjJ!Mc3g|M)-NhF{$C`^Ep%Sf;hP
zO^HFP9Gc-XF)|S40%t#Q?nxm=h6l~4`5*t|Z{u(O_V45G|Nb9hX=zC`;ABiroiame
zv%VQOu3Z(ayB@c0-iYPJyD=y_d3<6r5Q?$8d1Q3lrVRW0hhlYg&AZ1rrv-CTZQ2_%
zRWS8nEW!}2_76uz;Y*5#iM@<SE$D89ESL>Kc(K)34NS@*q|4eu1XN(sID+=i^!AH>
z<NONfPtlGn8qBDN5ssq~c*hkQ41e$WV1GS!cUJX;xi!7AfCg;@tt&Dx7AZ$}G&eVG
z(ulKUpm8}Q|46uIqHxc&v6|Qn;qIP9Z>FEN@BjzxoP<_v{7NR?oUvSsp6x0{sWM~@
zE|{)noO=BO6K*q>7jD>8BRnw12|_Q-B__wqRIX}248Z=#7BF#DQMh7@S%y==RAvsp
zCyjiuqG;@%S6ceIzq_q!QPw1+L?eye_1RpXnQ+A6kbblyXd46}Xj4%4p6&y+gD895
zqs_y4Z!oZFTh0@K@w~UcZQ)9)yV~JU`RTD6flMH9l{-WuPH?KSojZHp!l`T5+SU`$
z2x5a78Xv4l6kI<r8`GNQFs3Gk;`P^GjDdkZ#bLuD)<X8(mT0PVAxIGnVP*;@GHe@?
zwpUpC6I1t`A+?LP)Pdr4!F0qNVsBIjMq+AaIwmG2qFO1d=eh$ES1^tHYJ*X%RFwyC
zLc8W-yRj7WcWy>=dpmkd6*FjQYnYxeQmKR2B7%dB^%bpCH)42XNMj0<U_GiUnsu2d
zDISb>)?@{n_g6b?ZN&9YKURAln(>ZltwO05h4O&<kM``=dbO@L2n)kl(?u8O`E2Sv
z414vFwcZje3gY8nyRNzdqb>O4TrGW13|-Ff+gA6Wl~vbzXEoM0R-G=)W6r~-ufY$P
z;>zcQnSdwNZQsJK9jC!sk2+H@D{C>9>{xTc90zXH7h3z%F%~fCAM=Kr>r1h|wyeHN
z+A4eME6(HYp<lZZ=%LX=e~|`^@Aep>)v1&4q~N-@I3Qd%71J{_f(_2{UXK^9T#g@n
z@9nsJ>8#qTCsvl1yg2~$(F3kg+cO@~++_0rd6FuyY0EMi+h`S`jl?>U)oT*QB_=cN
zNQcwsyd~kb)}=n`n%1L6;|`zJsSc-i#QQu)oJTpnwtY%Y`KbJijs99G-hS)Vm_3a~
zXx=;#G<e9oRvU<^Q>P3UoW0Ib6rAJ5n7MlGdaQ2L<Jsq5j&tYH=p0d%d*aURJJvLs
zoSaf$6=HpT-NF(~-wq9rMOP+g?3egJ8Fe+<V1g0)TzLXBgkM=%j2sq)hKJl9@Lw!E
zE0xp_Xb_?u$=ouaXH3{7c%<BDP+}Sv&Ez8CvpMxQ8e&@szXF58bu=62Gy$VqK7VS{
z-#0ff!K<+%d<Kt$GqX5*i~6t$+UHvepA4#Oj8SVjDK9ombFLji8{ibpw|ad|$sl0E
ztZ3qaS%|IijX(txdWzyXh;~?AS+alxflP02C3<{Tm+-QD!ZE6&#uLIC#(lOhr1xy#
zW^+kaI2-)USnP>Cjn|E}1<$R``jW=cyztGk%Cc{M8Ctstz=2or0OyAl%Vps+G`7pe
zZ<Gt1>5c6cdaECl7ee{<)g_;)M>{h9m5|2^LL96rCEgoCdyO0QO;5MV<@q%6nRu5g
zRFww;P>o~N8-YG&vt^5zA4E%+4Kz*7Q=73Z+_SVeA8YHYK4)%Z^i))9V-^}BFlB@L
z0cjvSGwmRlP@i@8pqY%|OXb%bM)}ba>@D_rfHmriu{8g2tgp<e%n0*^$<+pIOkk`4
zKWKF!gv31L-mdVJI<3)y88)SW*Q_Nboa^yJeGnkgcB<!|@b=pByv76P(XD%oVD{4*
z@(BI}6NCp{fijr?(L7_sBB(mp)w~4lVtlSVL<i*aN#m@4_>}NJ<4<@`W17wD)?`C~
z=EM??LJQaDfGQINXJ|WbYpnG5SL3zUUW!*<dfqhH^&8g|1^knarAhNG^Hsu6ZQphe
zK0nMIkNYV9b+~UTpRdZDRNmwM_w71@ua5hpa+tImo*IkuFFYT!=g&q3lMh195W@Lh
zcCn?MVGwK#Pdqya_Jn-d1!P;{nk2FLjKB<`6@&we9S9m0f9&qD$vr)LpzF`Ob|Bl7
zFy9~^AtIonObF>2^Mp)*Odt?6oY%#~fgo$MzM(r2a{~k3MO5z70#SKG1h80VH`$gJ
zco4mN?9@>@crS=dpkNY(1tP8R^g2}MUwrwc_|BVeiVz)6i+kF*wHt)?wIEP!n*=-%
z$SJNDn(7C_df_8a|M8A*?s$5MZW3XT1WqY31Ppz~owLz|O4vknafhdebmvJ{Pwk9#
zam=H|QVQ#L@-DZYxFnwhqI^v$wF1h!Yyz5AvHXs*vvUyw)amGVkFuv1a@0v)&yMR?
zhlbhJ)zx_a{rBTHzx`b-EG#>8$8}uCb=*_nzB$3y%*<m6zD(4&d}18kd&l+oyX5sr
z>8E8`x_>OM9~O4^r}Y?Kf8<3u+Mf5i`E1z7@J`FAp8V(1D%^43zHNWTf3{s7(?9sf
zyDwOmFXGe2rl--TKKOF}v+3&XgMl9E`O}Sm6Ojf@#ik`pbFbK>Pl27kqRTVc9?|Cc
z_~@fg;vfFue?_HSjtdtr#_7{%VrU5ShS|7v>t_6?|KI;NK6vjp!Fec?6BDMDVTQ4j
zo!!fvqk@nO+MM0|wQ5aa^Rc|V94kx9W;*s{(q^)z=x+<YYC};dSJl@&v9Y=w^Ygbw
zJHjN(SV(s3<|nbZcw6fN%9*7G8WY>1x7Spq{a_=UH!j%(qG+(a4b$a`=7ft51=)5m
zb`ss6gj=)@-x0cOMK?ky-Z?G|a+dJ&-7rsKF6tIWU6w=k(Mn2!9MXaQR1>JJ2+KNW
zFSZ+))8epygmp0WT1@xD&Dd$ID2&S8lXm#)hY#ty%hgJ^yVn|AXfmX;l(1Y;kF~WW
zp9{hUK{Re`29ACv>QTK(k9XW7)M?e(T5MZ-4Q&gZjs}p@QGcW|k{0VL#TC53q(s{Q
zhAHQhb#<52k7xw2qrE#atjBZ(#&&i%@1(Bq=?o^yM0{vuuMcQ5wKFE{hyvsDaKGV$
z`(Z$*_G=3*DSuOOx}sbekBQ0Cg43?Jdi7H?&28?!MR#Kv1g0|mcSOhUZwm%<ab^}%
z`!EC%+PDQ$+R*U|M;{%kz33;Heyl|anCSY}cgoxfpO}SMFWDtGw+l>6&X>+9K4#f`
z))`tD-RfUUbEs3d)`~EjH>}A}=^M~Hiov-wSR4^-3C?zQn$f7Q#oFqUJZEyT#?5IR
ziYY@HRbxn%U=2mv(uSPRv$eSzH?Du8_pD`yRG*UCL~Ge>zv3u=m3OCE*Sbz?@gf4b
zOzds1M}2+HFh#qdEdf(}Uw(UYMfu;guz_*M{zcY3X0~TAhqx0f%lfV`4Tas_PG>r$
zvy75poi@y8HDZNt*4CE24mX@^W96l;awGhrUA3-aZBPHGiWUYNUTHrxbJC_t&p(7$
ze;}B}%p~<<osU)`0uRO~<we5`bD^F`G|rO4>?v(X+6K^;*I3Xv8Xg&s@zF~3Rr2x1
z>o3PU-+eQldFFD=o}P}uLG_owW_^7_Wkp~YF*!LLZ@%?Py#3Z2arW%2H9z)u_oKg7
zi|JDn)*^!^(QKr%@$3L=d0;hNlW#khmp=pR{dCAZml8SY+r4+K7Y}hqpMmq%Z5xH4
z9hvPph|@FUF*YU~UC6|)U;_aTxO#YaIQj<$1!JjCn$2x*^yGSHeoo=`;>DNVh|{y@
zt+ledv*)wH5NM5$O?ZA-US1L$G&M%izC=5UvysxKSKu1QS{)c7k9=NzruPV;F{@ka
z!`vWdP|<o#M`&!;>w?!+*9Tz)V}`XjFab{jEme3v;4U;UNgEy$<9T1>kmEE`o5OqB
zwt3FEJ2zu>dC`0+G`%SE;=-KpTqA0lzxt~E?qjr)>zkX#iFv^yn|>J&;K8JcoQ*z}
zxvF`QzS&UP>+6dx!5ZfBG{&-=FP~HUATUAWOyO&T=8L8AwV%udmo=uDQ>?X2O99J*
zakU-CjCA#|c>u3u4%lp>He&1&R=EGh^^fD`jZf5;8>+`j+_`-<KKbNBm!0#(7{lN#
z1Y@^ueyqM)WazsNOu3v12A)H6m9fdVVxx4qG7y=3QSI0?3^5+StI0}I5-K8yGfxHK
zq51`F*RUy7)j>GL+DvML&|E|1l8|nhZV}Buxu`eec}H!y7|V<EqAwN%e;dZT!^7i(
zr%?q#z@t7;IoNyy&S9eia}_kp{;uY1mEHJTunx@ttmaChM|zT0+t%it_~`xrqyD}f
zm10T#BD|t@*i|ENP8V8J;2b5Vd=T0e)IVwy)zAH|^nojNrRE0J+hbe(z&K-`>&^=&
z3vXY0-y0$@Z>g`UN7b7>7;E4fXqvRyT=0w;N5(^ejVy{UxbQio!nvt~6;|^$0>E66
zaRnSG4Yea11L!-`3lvIiVS$<2oQh%MxLfmvzv<n;U^Pyio{e)C&PBDasCHQLIe+Qw
zG3Wu{?j8<md)$NfPm^nx{dgUo#62qB@pK>b-$r**IUnW01|7jy$NkZ`>hNI9K65F~
zfFuWNR+IMPN;3c^OIQTkM203o6B;6nw|2Ho_;MgBo=KENG@kRhu(u1rSWsMcQh>N2
zcF+d1z^jETOdM&@F>!mMP})52ZnL|vX+f3d3V%b}i`_d+eh9M=sP$^HN~-aLCeB<=
zm2FFe9*26OdfeO;(MMkVu~^w*;iWVh<neTCd@+HYo;?@e|G^LA^*3I(ia%|MgZaix
zZK0Ry%bW;t%H&P^!wg7^MgAcqK!EK<H4`iMq@q6v1RBa|$P<>10ckon$O3-Tw^Rnl
z)HA`jrnFoDzP8dLoZ&&wsh`>@%DYTGdPhBY(TbaDCnrM83##<ii+=7En0f<ytxVho
z{-#VEgf{V=zAJp%&9og$OUv<_U;ifFfB(a%*R??FxQ^?%j(a-XHzxR+ojLv2$Av`?
z^Tz!OkN13{XWjVfFvi;_`GmS3+w1QrzbEyc%UXpzmfneBKgUy_FH(+Hj`#j}-)g5X
z=asb5`=b8jG#}x=$v@|rVCviIp*FZYnJ2Dj6a6!0pi?*}gY&csg<O=%Inl_?5Gk#2
z(9999z4l5Bpe2-r304TsUgB(yrR60Hww`<LN}N4+A$ki*GiY6D&Ckz8t=4A-((PL}
zV{K*C^g9eUn1Nikw_$`8gTv@y<YQ`#UCEp)fgtYM)sJmjr&i@msqR=?U5<}F{J`g?
zkPnQV-QBIYedE*k*I)i`GeKc2?TAM2&1-G2y%u-xToo-|*Eg*bGKXG!sAlqLn$+WP
z2Mrb)rp;KV-UvHjsPRrt=_aC5be`kUhOAAB-a(FuxFqz#!KgzMD4m_;UGnU@R=z%q
zL@zNLhv~#5NTV;b9?|*&2B9@^%*1P#6FL$Hjj1&SFxky6c@FWfdbfN=^e<_aD*a~e
z5(NQ5G9#y`B%{7eH{^-G*A9n=bqnQK!(cWs+sh7mm}8=IVZ0rp)kM}w;Ac~ejxYEF
zzd0Q)TC3$tkuPF4u4L0uY@R#7GJ;JeDu(D`Xn2^;%8=?fUGMTIPu1O<5x3B!!Q7Yf
z)VGwM+Q9w5eX0kh5V4|gQ!qjMoSGbo*Is=oUVrW7xbnhtF*H07oZ$k)wo=T*>1j+7
zPO5yIQKmJY){?9}>0?2R?^$2&Ydy$@KHATwA5~t?UqL7cUa(*-hgryWTT(aG(`y0R
zn(LvzBNWZ5U##J#FiFdZHp4{OPGCNCXk@}?>0ky8Eiwd}&QE!xEpbaQe)IY#8cWw>
zO>0jWx;_I3rmDiFx@hgJK4gx-6yT25%6IPE<_K$>S>tRa&VobBFTuLt@nAP{1>m!y
zB56IoyAewZH)C~WA?h2;ap%^xxc2D>dPbO~@ic!Y=1Fr8^I%Fs?Pbj-#@%*3ZeIIX
zVQ#2)2%*ueqWuu&cDcN)(_!q--@Y0vON+)I2q;<CgNOF^7^7)ypmmxw%d}o34>vRI
z#tj0n))+N#3X`cUz&%~UM%lbzF{5#Uxv!+P!~991m4eS8+HJvh3w}M;NSiS2cYD!a
z%f-lWEvBX>qc%98e(I00iRrjxbBJS6t(IeGuo{DdmH5tgUX8!_i=V~2-+MQnxqLaM
zCMM;_;)NHUkJn#+C0>5{#W*uNYZGyt{YL*cHaEORB!B({A<jO{(Ym_ayh`&w-|UaM
zB!tC%_T#<!J)WZ#aKy9MJ%VHS9|)dG#k@Bm%+1~M<^;4sC?7CE`=i-6KR+M0Z_mZ8
zJ9F{LC)d=!JDMk+w<aW7CTuQ7)0a61SYKIQ@L5=UJHp+Vv@3B|5yB=kPZ~DKhQPp^
zXoVAiB{pPGS9mrE94w^B_bLu1*!I0f-)z*q(Sm2Ri?(+-$BDDDk_l0SVBjM();JRq
zEi!mLz$C&R`UTBP@Px~w@IHS`_yKdHBO_?fMBJI1)3~e)Rxo{7QMm0`U01$}$C>ot
zJ@RFYViFSG7V}rq5G$M4u(q-k%S)IP1_va;3&JEegy1MsZD1t&yg8&#4ZiES8leSd
zhm+<mcn5)qyt`d^ZBIBH{+EGhN4S33LT438b!1!|2)|s7YuB!t2Zk^Sd0%~V%k5eg
z9><~*rbw3;=QIcZ)<Q~zEeQ5$69y;vr7Kf$J259A=<mtXkMQ#Joi^<io)P}D=Hp&N
z<x2Al?E(zbCkW9Q--m~y5jGbDQ`8Yy1pmWl<2aH{3*LbTYe{STGZ&y$R<C2)u^4^T
z0c#G@RtStd3>2=LGz6!#8=6#{tq5NB8Hk**sWIvCr1oKx54eN5jX89yeoJljJH?F{
z8%M)2BUsoAG@YB){#Ku``F*qQIV72`RFws92`3dajxcq3X!D5(Ha)MBzUhsO%GR`C
z5wo62Ll{9^GP&8Md6aY5Y%UbIqF)uOw~Ww>uxw~RaJ%b1z%2Cq+;xvH+R*DFwE^d{
zqN&IjvEl&jYfo8JXZ086e!JBlY?4TGH1n<I47DE{aLUErSXdNI_VIPqiFp9NUcwiT
z<BsslBmB43omAFG`L=u1x0BpS;oJUcaX@uP@YQjD98TkMbb2agFPw|v(c$RsWzoP+
zI4zu*1Vo^KSkbnENN2}SvfT{=&3IwK#P5kj!~>Hq6DrRv64~(xF+)5E6z~27aq}mn
zz#<zJebmp1yQj3<0x;B^*<s4!du=`0@n!P1TDvB?Vjl;XR&BetpnNEo7wj9G-j%Z>
zA`Z1|COC)(7AyNg{NxQ`&n}_O&8BIPr`17_efN9ckMGNKR)Y^1Rog)j81Yg$1UxMd
zk)ZBzA3;#0!!fl;M+l~Vx4`PCe?;E9Z6NbWM^_q>G(_Cc?}P_Vn5<!XFzGV+)5o5?
zl`h1qCu|mzM3!Sxr+hq-y2BF2@5rP5tN^UUxjKH!>4q}7I4Mbo^@2kmwT5?D+iKms
zO-xt5@|3=D<<DkR1m6(nOCo08|K0E6(@$@BCtSyMT*r0XQ{uik!Pn`TzcxO7ET3?D
z@+p4nlAcfW`guan<>a*5UOy-K69W6w6aGE6&`W=xN&SH^`Z@mfOa8TM_2k~=cmjW7
zB43kNj>jKQ_WJpKdfZ9v^eEq^W7_x7LTx4>=eA5vveD}>7Uu87=EhnK4iCoY*r@20
zhDe&UTOFno=X*SV<%PKP%yTh3JZ1(ROg^+z*lCR!Cd?tB#c}Q0C)UV;hNR&zg~hoo
z>~u#!!cJr~Qu;V+M|3>Q`({IFn5nvN4JWk27Vpl-ifAZI@W626Yzj<6&ClJ4n>Vi7
z<PchGJ_}{15!=lbpK$~e9<w}@yDQKJp(^#*Sh{USS*|y4flnez5pvP8lobzAt4A1#
z{-$SZA8{o-A7Ku)-W_5N4wF^vCPzrcH<tdG`KG-|4>}emWs6Z6oC~5brXva4PLH%<
zh~pAPYX+S2<{R;~uFB=2P%8T@7R-Y|gIhpnh9G^dN5WfctzDj0`41ZwcI+TD(K-jM
zgsvVKj|!vsFtyXBL4;H=0QJpgs1kjj+N{lFC0&)9{LySuIxrPsNOPtVj9F-FX!fMp
zrML=f6OrD9j=&=c*AOb1DTvU^LPgSU*BPJM<LbvB#^3+LKgONgH)5cgkDvVbT?-z*
z_x*R{mDgU0vu9_br)y8O-HGR)ITtU!c**Ca0Xvi(VIPcU81^vk5uVUqtaDhGBD?_}
zVGyb%Rc@_q)CXRhsO&tm#zb&yZ9UpVb@ZCZ>mt%&C7`~y07&(41_)mERmXhTOjr(~
zaVB<}oDH`kI7tf-`ebE!E<XI=w{ds=w#Oe@ePa`oHr>WLjkdzH7y@3#GR#%_eS5oM
zQ(rmdQLAw#Q=h_?Y|6{-Q0rHO7V3+1)(o10TH7|zJXtc6dtv^D#@%h3f+O!9;7PU7
z(003dldAI1<a-ovGnVeoX}#W184yat_{F3lS}STN!2}yvDJN#t&<GhGozVCzdAuXs
zpwElNn(CqTrpCV4VQPEQ>JmJ{8&2>IJa7djc}ZJs8UN6{2tY8qW=#>nB-$#NYeT5E
zznui3oDr6o=D;>^+)fvd+f&~4*xg=NebDO68pokM!J3w{?j|QE;@M{|$e)e#=h8WR
zr%p{pM))wDBQ_S7FI|WW=g-9H*_oI+eL5~)JQr7<yA&5L%*H@}UsTFPpBuHlwr&PG
zu$ACQ712e*#kX9NQxAORUE8%j9CN_J<2pu_8&E#Y6$V4~y$Ygs#(W(bobz|@#L~jN
z;6eQ__^EHK#tq@%>sPPGZ-4XKxOwwd)a&)QyRa0C%WD=AUA}zDXA=TH%qPIqf!Y|e
zYIo;v30L2-RwHI`nHMmThu~|o4m_@!*W<J7EO<|XRK^wBUBI}7z=AK%XQw6zia76G
z5e%ax^#}0Bc&1Glw`heSu;#ox+8yBs{2_!tj3ef<wdEzV3XI3-*8<v&%30|eKKXHa
zNf?f>tgbR|vw=io9<ztQ*@nh-eH~3!1QUWebuPRs`gdb(E#~HKs;(=>P2TjaxWyv4
z6Kx{|Fg>bYUS-i$W1yit=5F6qe{5O%ESKx{7}!-EY?c=I-D|qP&`3%=vO{Z3@h+>f
zqOk@(@OgWxWVMQwqsbT^n~JgVDUVS!y4g^{_ol`!0uxM0R{J#H_M)vNT&U+fLarXP
zwOZq@wLyXn7zk)Q$J%_M;s_r%8%t`7hJ~G2T!Nnmc1c=S!2fn^H&%o{(cr`UBO6N6
z8U1KG@qEA>Wjx&Daf$Y9Mq_+*{8aQ0O^7Z@f;i^>W@9Zd|GKxct?&p7%aK$2BJ{Tw
z_Q6&()&7{7q)zmyM*#SLUuE1}h#Oab8|xc$F*-UH{r#iLA0CzPKVw#Z&?^YMI4>1p
z9%hi?HSX?V-Y{tZb8na{pn+Rb6QP~1xW+~BtTYpfCK(zT7k(U9TQC+0aS-`JIm&%Q
zYFbPcvq?mAl*(t`n6>><e9u#Y?=9gvG!!|rtXxul!o>hH!b2)oXqop^=JpmYDj&~z
z>JMe0kwm!^M(wd9*ilx><{&=!;G_75fBfgTe)EpzAHpJ-Q#vH9|HpBOJ~`QU+@sw|
zZT5BhFXEU@JA$u{`=fA$zG9p{e>TQB*Gq&QlkSl)L`0)dza{`JI3Q>>pu8Z|J0^6d
z5bjxQFd<?N0d$I46o?9l?X+75b@oF|xR_1kV8+&hMCoJp0>T+W-6`vy#RzGTABSyo
za4owu*u}&|Q!1A=nTrtR@J}tGO&B4#(BcC^t*-D55p(PcJlNA>MtL(qvRLv?Jr$C2
zZEx*FeT|(X%J`{ssUOOfYW(C+e->}N_PR~f&?VHD$<h0Sb!E}$1)h6{AZlCsfIb$H
z44&j_!iaaoVPa%GNEDt)SNHTdo}aYyq!mZrJ4x^+E|{Ek@3a+;2qYlidXnQgU3q4=
ztrJ!l;we#oR_^+xI4M3sjNExtT~qlTp>-t~a7E0N0%b`=T6*WAw4NYtA<UAnbvqUp
zm*V~3y|2Z=O+CX%>9~&TxQ=@&+&3opI*qBr`&`28#`mA&^<#h1w|k!X>7@4$_mtsb
zp1(&5rI^o-Ju@F3fBq`Hem?W=LGQ9YP^RC9e`K5ep8c`zeW+K+6XSkU9(U6FM|u5-
zmTJRbI+uY>MN?A~@!|{5ni-qZ^X*$VVs&-hJEEtjXXDiLv;|Wzr=i)01_v!@ViOvh
z!O$FGr!ow4Hj0^k7A*`Ngk~3vqS@J5$3Xx#G&B?w6O*Qw5t>yhHRY8UKrpEgbQO!`
z7@wH3woJ3hX0>h6`bE8)j)|#LW*k9p_4W0~*|X<kY+^=fady)|aMlp?dr@I1H=0Cj
zR<m}`;cjei)?;;L)jQjXPitZd&x{|_bBVcnkFieM@xILzZ_~YPPk6!+hP4LFB7{f?
z6mVO?8APl#y!+gFAb`*{b)}jx2_jNo`HkNn348pKym0Mrt#>qpA7MCh-bz7zEc%n3
z?a;p#v?WGla>V0434GGVvcw36#!lxZQGNslDlg%kf4kn?^R9J!^p5sX8C+(C75z(_
zu%Qq^0nBK$(+q>q^cIE^mbjRkT2nSHA)5tldhH*|jwdX6;P>PI{GWe}|L4E{*Qjr-
z#QF0x@n?Vb(>QzXT;v3|oM(gS%>GIy`ulq0`R6Xg)a0mOvt))Y>qZ1mFw1FkuW97n
z{;CfQKQjPf{)vWXQzn8B3&i%-ZtAP!{gCjgiro@$TF_;|3w_L*MQJKt7XqULU6JBg
zKuA3WCrS!aYAdTSgfW-DDY$2ix)Ga#*J`yd&R@6~Q>UhNUsb>D3Z}7ElFakh>+2qS
zSTI>$!X)96>epkFqX@`)fJMQw&!IZN>Pge}NCG9bahKL}yLF!xN1YMa6%}u2Xw+BE
zE*hVl(Kwqj19f9VVFfRhT1|bcvACZ$DV9nFGyD-=4GxVcF6Wu0vzBN>7`(JU-%9<X
z4!5?vV4(v-aPp!3%y`vWi*-B}GZ4=1b0!a_3<2FXC>!?FZ!NQB2p}-;n2s`_ZmiLO
zA?lx3x|qLPc0W=U)^rx=k)HY}rKh~<|J|kqA(V$9miQ6ddXJD)ZKqDl3vc!HkHpyc
zcvSm2Ur=>b-w2ouM|7;B3uJ)<%GnjULT?NX55$?7>3Ht>=i}VD^A<YysgG;@Rd1FA
z2PT25N<=ccOSFq?>-YHdql>Obc;3(VUmwR2(&^}Gc-CKgOdz~O_=Iq7VS%&MFo&6r
zF@X8LxVR8=^K<d(r=Mz^=VE5&bPNm*c&>Q%nP)UU&kN6EIR>F@)8+yjf^A>|J_~u>
zRT{T$-!ebPW+K&&SlNMJlM$T2mjJI~vXFDG5IE#=2z1oSDuV?L8aJ5I1NPA#10FFq
zm&;~^d(ap|P@Occs4N?pfj!Eyy1J-(t#}T>gyHq8pIZB-zrPm4f>+`IFKNRzf*|2_
zm7jCT(wIb0y%cLo1K|+DGw>=fKpvDC;hWDb!>ddPeAZT1oDYIq%r=_Gp>{<$MOo4r
zQ7WUxHfNXte``x~-td8z7TQgD&ep<oBN}SVa~d}YUl=>UWS%p?%#TTYH^vF&pl|bP
zlaY~$m^^(x`Ub`=<ST2;V?Gy+R0NR-SJ<#nDVHorM@!BEF~JAl%!31;s~?l6-tfoL
z#&znGKJBZ`H)2EhiMB-8jV3Ca2R!E}j)#izCBh>%zcWtBU-OX$6I$osL{-KDeY83;
zHB0+Z7Hg5KJdAhpMHpY-K+BM`(=-&p%?29gI`9Kpy$3EUa}{T4<wQ$hW)eKq5Pk(-
z=kMMW-rTZzNd#KU%PY<cGlU4WQkxc{m(2{i2gfsXsS{%uT;_8D^&TOr1yss^Pw)&r
z64vwhfc7bKCZy<>G>@a5i`F6nVa{4pT6zbpLSF$>^bNv#zIA1&I|5W;y<E4?S*=uu
zq9>2pMl|vc)nB=^;X-o|!a6otU^R;bd~}QEBX87D!I{gz-|K5=)UCzGA6<?A@?ZZZ
z{?|YKv-+WBy$HS-l?B&=M2<ON>?Hqr?or_$H;v=29p}sVKW6tu+5t+fBlzmLKX}JZ
zgQ1Dhm=ICdH`wn*RfY*%=-)eCg!IwkLe(8DD<({K=0K3xq@gAn2pc9L2n=Qp7Ad%0
zCUiO5>P=Wy6DPZKlGYasdh#O@=`dL^Mj<Lm8$tr2n_Y(x48*B$0QW$@2%8}hRvd_$
zu^<dButpP`;Uc_6;OsO-?CRa#HmcI7$cxYy;Q|UsLLsgZn;Xs8T+>AL)H%(Vu1)v=
z+8_PsC-Lla&saq}?eyc15D+Bnf>hg5VZUd#6#-!)B85b#r6;wMlT1#1=5EtB6L;Ja
zj<~vE&h)&=53cpj3y9V$-f`On1;cwe7z!>I4=NLRq(y>Yjt6-Nc@jhH#lncNKL3q6
z5S~Bc2vW(wbs;_C$pLRZM}V|UD2b@iH=ZF1NTyu|(uAnviEquumKO!T{?)JJ`psKP
z$4Pcv$8}uCJq_*~)c896*T)HLqu+#rAL<Bq+&wBv!t=!WYCIWv5q-kPRQ@y%wY`3x
z*gtN(#|h_hd<*_umZ$iS|KF!OUY5uHGYquh5$fwuba7Yo_xJhynQ{rCM2m**%EaLC
zc#Kb;j&fg3(K0rP$nJialF&9Vnr_|rC{~tl$98kwrqPyHS7Uf=G6sf5MBDc&%)aRE
zd=!gl^}zpyIgXG^wDKkz8ks2cmPL2xL^E~i-DY4)Yj9{Z2FA#n-PHpYV2vnUnD4dP
zU|`;9|DZc+{Ub3vIvs`Tc=VRWEL?+W29uw2lZurAg{_!r+M~K|H8=ciXly$Ahfn()
zi)MX|1`W=1N`e^H5D#`RbR_E$d4EC>gX5tcz9lTudLx}nLRz9RcQ!@yVP1({`!HM+
zQ!sg$O$ft8U+%)#XKj%f?<t?-<zfAUN9fWD!_I!w!s=IRm?cwDxP!e-rKi3|i_2-L
z?2e10kJ&A+I~Z^(7mV3-9+%c>h86Nn<uifXuD{cP;H7MMITOIC4N|>C|7s0~R^H(r
z8iQIdxE$<qPnI}v2Ezl+Twad#)%nP(K;FTBpga!H&QJiGe{8ow+xNuW?Q8LG|N1Y2
zv+j8FJFmrSue}~s!2@Tdq%vThY9S7GH)D9XKYG#hav2kRAau!Q3x+)kvd?-z<wB^%
z)nyhec}c0X+CndDh-nR!cJ?Qxr3EQ)3l)dt1x!j3CS^(esdepM-CBs<>Mz!&Y2zUE
z6RdDPkIJ0QvYyzpKx^Ud4UM%WkDKv{35~^(D3z<;m<fY)@$OvQzIjz^i<P4_HfI{4
z1^4j>zlrANy2kXd1x#jcYVES8bt???-5rh$XsVuCXVXtR_1M{5jJh>$8fJ7Biv7xS
zQt`(%2KysltobaVN_9YW%f)&ft;VjX)G#sEul)Noj?rM@D1%D$4~}{KvEtfNnp=%D
z2Fzp@Ug#}pJZ3e9Rj$RmcPzjg7#OkuJZa(pCTRmDeL<PGwl;L#cKtE2XcL}j&k6Pr
zAab@E40O^&0I<KiX9h20GQk7uzyy<+?!){XLMr+g^I!)D+gcMMuu8Cub|1`a6<XzE
z-H(8OTj_EJQ|fQpK54q(1z!j=S+uH52v0T}8x~|8?2s3NMudM!D91TuFu2DjPsPhG
zJ|Azs^=izXnO6DtVrgkf_;yYCA%InS9tWgLHInGdYg#*A;&^W1dG3zBpX@&KUC;jW
zm1n1>A8<>T&1UZGcEwtKH|FP8wWeJ)Zoho_e2k3`n;%jt7vsYD3pVvQJv9~MBSSGc
zIToj8PHSvx9#L5r7v?orT(?lSR4T{F=(zBEf2^*r#a(>^mJyy&EjCba9unscqLEZW
zsI9(4cxLm8f~m%4-5W6K>&wRFXgeaH*w|Q$)fK%H?9mtISqKNi4?_3>3?Mw%++6j%
zfM(h1@{-Mq-nn%n-h1!YN@Fz!2Zk-&Y6~)wX~kq}k+IyXHoyuI&53D5giX1OaB6EV
z#e$IqtZtu*&%(s?B5jw1Y&+I&^Z1~JXn(~;(-HHB%ggf?Y9Y`=V~H~nSC(!EXa7|+
zuMLfiS-6L=o(<*9u?W@RQ32bOmklJ87hw{@aSigwYaTa#Jydwjr(Kzv!XzBYeB|=0
zttby?oY8)4I6(k|#uRaxs}Ue0P)VAq+ZwQJA^~oa_E)#&6XpU;2{s#Q>#e$RUs354
zE5ng3h*s#Ws?X65HL}%ML5RsYRJ+1O72y)WSH9vp(#E8ZP`y`p3XNfe`J0<7s=wM0
z;8j^YHhEB9%qb<|oN`6DN_|f|feR5JVR8|iKv*=hvN`o_kLU>DZ+MMt4nU}1E)@f-
zP6$gIjWw08DY#JC^J+V_A@mDo2r)&>*=4{sLPPkTm~#c*z+a4BpZlghVsi@Rz#<lT
zGcPeewrK=RRc@><c&uQ)6+R|r9XThFjr-s>Z!S<bqV0DeJeBA-wW06~{RjQi<#SI(
zV+qE{6W(2#TQ%>gJ~Ww48)_cv$@ChRqK%zQ&&msL-MMo;{`sH(IsWs1`S0-$|Mai1
zvVxgb<~&7H0gg7Ali24t<CtUqLHDrtZTF~e4|DA_zo_rHCwGkIj^L}~{@@)Z0cs+4
zvPCV9!4*v4BIG=|i--dev(uAF`G7^9o>_Exvd(2S(H&}XU^kYQ*76Xpn0FBQXCe^7
z8iIkz7?%^FYa9CPI3Rz7Xqq?>K0uU^pM_maf-a7pHNloNiD9yZ!&DKzk&h<~l>^}y
zJ3Sy^wwo-<AXHHS#_U4U@Pcqf5az?PAtF>pc0eKcYHWzGc<NkgPY%g^=J_k}{qMaS
zr!jMAJJ6yX)SfI>t?}l?oy9tunl&k?&6xPqIZlJ5AWrEkNA@cw52k#_(0wAN+Y^e3
zH1ArGb9|EHHxc>CL0D<R_U@9l*K0nv(jkmv=#lv3=6H@PLIS1;fXP*kv|Vh4wctwg
zxZmoTdlokqtWiHX2t0@%!qXn)PdPmQD*`am&XR=^yF^#l*5cETKaF4h@;5QRu;z3+
zuH!nc<DL@tjS0Ta%+5X__<DS=A78y&`C;C3j%J0EzI{Re$oy2N`@G&gGW^%<op)>A
z{eAjxOF91de4@#q*XRRi7X%*ch+aVpO7sf5*ZV4cqLC|6(G^DF#>TqOA%TfxGvzRj
zx0-Qx{$^bN<ioge?Ngh%n!mdkGqY!-e_+rwJ(?&s@wBrQ9P$p?g?aH$uU@rDEa-Ea
zOw#Cpfd;b+b3?WMfynj3)Z$zsGyvGR#yKJ7$YPpG^?<?MTPQ`wOg-ua(>W{JYd_ZO
z8=~L0BF|227;0J3u_`+z+AtZED~zZa5<_QebJesQw7i)@-6|(cb-ZauXg`f!g{A(q
zp9izwCm~bog&a&PnB_dPQ7$oM*>p#_F%_8N@Es3pfPOjD`U8`BiMF)qw(U)|PfG~p
zykLfL4<m6$>yX3%?CL?lf?$!fC^N)pVnB9hVa8YVj&^BU2*-I=PSQd%-78OoFKh%v
zIA~#xyy^8s9P2;94e1|x{j<}0-71$&DN;Th;nd#xhIvV!sRGWm;EQlH<Bn-%@5omh
zt*tItpxRsNQ+a!=A%_3}^IjGxI88kL!BsU<cDZbKJahS?)+^tQvB^o*SMdXQ5L`DF
z;@Z{sW3#>-!^48dEaytY97c$hw3=X6vVMaJnbwG`9uBph)cQ~L^w~Trs{#t1l`H@(
z_(DJhBbfCNV*y4yT1hZPVRXXCY@b(yu(46+oGigX7Oks2uPGb#O^?A$j=n>?uGdUV
z&eL04Ui4T**oEe$&#+d>h)bCeK(hlMn5^{=Mql5c8RZBSR+g7-{%mw~JbH_0JaSG8
zeUvt9qWuB`mHwd}Fd=&P_Kj$0t&3o5czimlL(@^}8?a!U_F$Y9RhEo})b#TK-K6oQ
z^zy(@kD4W&GsebIo6wR&2(h8@kG7OoW(XY3RLyxkk7g6F1Y<eVT89JE;3sS62tLxf
zk-kDJFl{{L86kkr2io6Nal8OiostlcGBEbxF8~YVvA45fSj^^B28E;Vl8M%BGm&XO
zaF8`a)wf>P0vA2PJ!lK5jA)x6q(h^}8()EI<}qNyV^oczI<l5!EWwXp2%+hLkN`$L
zf`=^x1}?MmQoXWyOqLGB`1q-~c=4?A%f$ShIdAG^4Gt`Tr@_|=AK-vR)y<FmwBOU!
z_wJ87<BOJWKew-~JO6R@J?~WkWkVnfzgYhBu2kraQ&S^x>EcY(swH)SU_$MZSO0p0
zv&xIL9I^m+<2YaH+NU4K($Zb6qq9-%t9p(B=iI)1%hplG$0n>%#D?1}aG^9Aql827
zMLQ!LX5$6vudS`d+WLz6khw#1na0)n`bK>6Dd#`_TU@(#&Dv6E(xCaYe}EuoFB)h^
zAz0B^L351r=N2@6Zr!>SS3kX~^4wP1>%teD`_&tm(#3oyLM}FM0GDVeGj|kv(IQP7
zuqiL|8t}wDVJZKf((yU*f_b#qEG!V<fJdk|=e;5HMXLv`tBl$hfiZ14G&B<ZgM(^=
zn#;<$O<9ePiOJJ3I6PsP2fj#~u7gLBgh&XlaxM?>0^X%xa+-^5Y7apRbJ~H*!nhVZ
zf{zha)92KEZEeXW^AI2+xMM>L<@Oo*2+~zgHj-06a1%m0gvX5YG&g7tI~2a#RKL{M
z)Lw^bucF!pjU#x0grTgAb><Dho91b7DnkS<Vb0S;7?d=fR&~FpzCg=L_(I`#w>PZm
zYK=HW0xl4iZ#OrrJqs?O|Jmq5xfs9TFxq#WBTq#CI4{jN=1GF=8PvvC8S}+TeS$p&
zVv_@B^bLr{85kIiA~Xr<soitgLgb5RHnKUNBoIPkwi2_>RpC+KPcyGz5UnrLhtK45
zAhQV1Q+pxM-w{0Axp_TS)&Ag`T5Z_yYk|AUSE<&bSQ+vhiJ9DnXfzyi%2rc!SAA8*
zMSzJAS!oJCZ))BE-+`ZjNeXF`zWQ#nw=KbAa^N#OI7hfz^Vfg+&;KR<$KU>=(njM_
zbBOX{Zf^@2!SRH@-+@c+cWu1WcE&TueA~6}kH5nmfBz_-#@M~*FXN7<^L6{j&F}Z+
zPAbbu9%$+azB=v?-mycXP_0;-tEmZ)14Xwq$ulW35r7y$vP|}%M|M>44zmf^YSxm@
z#0%=$5PMEj$QUA-GynrZ{J!|^!;ytdd5?-Q6Ce{iXdnq82$c|UTFf!&i)ceAA>_`x
zLb2e<h%_KN$d7x3WC%J)gZ$WmdZ@fv?|9dj!eN#KbzqyoXf58bWzJb&>#Ox>Za%FV
zUx(g3^yb@d$7`>?qJ`6l-a_=MHR&_z4w0m1h&~7dT3HV!K1?XICOLoCJCGmg^6Ptr
zF;a9O2n(+M=rkah+DMal&YN76UMdH96B)<z(XEqlSkln1)1X|fa7WW}E0XimXO~Xz
zl5j75BP93K8<!Sh5cLpSz*=gTRHs%9MI{ZVsq2miBm};5b93>V-~1-td+(#zY&?yy
z(QzHuaeu__n-hGUp8acZ>l6L8=Nb9--0@J)UkdsadVW;+$M7la@6R8#!SDIcW$FBV
zvwv#y_GToLMS2I_hbf`8)s47$V=h)#S7T&&AcltqP3xhdz!@ypKKU@3&2`b+qL-Un
z3VS1d^{@XLAN=;War651*sO1wZik6DF*P9q=pYv7@5D#%zZY}2K8?+_Wiw`APQuWF
zF^Rb%n8|E_+uz@|ZzzvOV_o!CR}}K1-E)*tG*}lqxY<FTm~}AOz3YD;noTsRXmNI&
zBlyDf5sbF=m0OYN#)J^&iq@SUjHN=cPwNz=bEy2*=gc^#ZIWrNt=Lk!FaR?>5RGX4
zu--rbgFs84V6-15?SbT=87Ze5RgP!Mgn72a;54%^gMgzYq&sqn?oJM2LGt^YD}pwU
zci!uj%4a5`J^3ns#Y=)R1S|wnTkJ2!Zu7R*C+lX|b!Ewu@5Iv@3g&2iLu;9a>Ibt6
zMq8>!VmjuvmQp(_ZEx`NPI$qA3!$=g!Dv*UAv{vMr==Kax68nqO?7t}@dzrmlW?aS
zEeXO3{$McD2NudeJMX5o8VL$^lJFs|t-Q9dK#%nC$rVezF+DvVXJ${w@bFO5YOr<~
z;aR64{M?M*e8x;in0kjQdpehedSm$`3BjP(iG{EhhF~%ahGv@Rc&!f+^0G63Yim>2
zX6y<^*u9T=HEV&%A6ddsLNGgfG*)0@?(M8uXbiKR_rQ)#<jbSkf(BH(U%S+ox!y`t
zYcRUURmZB=#;k?u0*3EaGcnIOt7m;<&Eo>0)Zrma?yOB)O-`MT+Q7Kh;aSzYAedL%
ztKMyMn=tZwF%y<4m{E;p41GU7c_wltw8Sc|$J+9Bt!Wo5kSegbQ|mCp2Fy~`IhSGF
zY-_CDjm5>=QLl3@kJ@5SZL*t$L@<#N{3Zss@R0hGaqqrW9x!#$fItY@*xK<Jq}^fQ
z(wD67&E!_y(89tYprY+D8;N#cyB`=zX2QahMU$@^=5jW+Hq{5K%fcC(D!1BFX~Q%(
zlUw)n8w_oqT_(8py<w2ewn{>4K41o|8yNl60a!)IiNF(lVRK^^>;j`nxQu|WM`cE+
zjpmZaP77!4?Whko*IjpOmN2%||HV>&%+5^2OD|oCx4!eb@bRF9I~&4b;1csQls`C2
zP1bTpH%HF2#v@$&yRSxJJRkQsR>Elw%sqYHo9~K?=O*KgS1-l6GZWD(9F4h61gDtD
ztJVesZ5Ln>&6>uh@IB?1zj@<YEH5v)|NAPHDCYCRa~tu|habeh{rcaORw2$`xTJ4Y
z<Lr&qMV}doun-u;yxz9P5keT=AuQU|IKY)FwLnW56Q^h$-nnx({`Hst8uND-qEhXP
zSxn-NpN>))?X7N&*<Hb9AqED<RGuD{aasL+5aSc4tzC>(F#?{Xk+rQlX>8_%9~p>(
zzxw)$*7mn7fTooYCRm`Ldh5r+UIcfwYQN$Yg-5nxWo022?#}xg0x>ik5iTW7=WW3`
z0yecvp{Oz?VRvt#5(6WX8qcFq?i=)&Wv(r%jYd?TLf@$S4)cNYb90)%IA^aQUozgs
z`y4-YDW+C|+pKVoasR<~Y-zqlK$6K91>Y*;;+*QN`l9LCTeT+cwNHNQHZNBY!Xp4i
z=%{vA{qia+;}hJnx5L=cSknB!ri7y@Qa6lgcq*7?ER=^-kGyb?aH!Ak+7?J7Q3ckR
z$5loIRBYVfmCD}DQ6<V_7aZK&@E9_$PyM*9vTY!|QWDIy;D@EfJFzf#!@@}f$q2*S
z=KR>~K^t^gI}Nk0eM!(S9ET8=vk1Ea;f@9dW4cSYPxuSE!di6}oM|kAkd-M~YzX<;
z#83XvFDjto!Ea1wIreBSM(~!+S41cEMQ>?9->NDH=3qhY%8RoXX%FGEjkT4ydFzJL
z937i-|7XxH6Z{zuaPRXm*$^Pu>CG9ZbLQ64(n2)1HY}v=6}_`B8jthxm`8z|OfMTE
zx>ZE*r{-HB5)1K_|Nb85#L*{cphF8aL{rq`{SQBl|Nh_qKCWM%Q@1DFX-q*p+QF_(
z%ixSuTK=B<JiLTEDa`SA$8p!A{L}1Cs?Tx%G&n{{NAT5gfAFqHi`8;}RS2GgMiUXs
z#O?`7lirRd-li4~5Q{9r*?EQl0S&98$rGZQD~kigMTmwiTZn(=5{-Hj?>1qG5KG7%
zqKHKdws&_#bTfyrBL)OP-pbnE_yVF)lNR*LE^)MwU=f!HE6h@eD3!PU=SI_IYioc(
zAY+r4-DZ@L^b_&i6V1(r769v7Bqf2})8`JgkUSGV_|Xq+mz&*KXh>0KFLYVhsU27(
zLU1S4n;O*<pZkT$j){;qPL94vZO)RZ6_3f1xzc7F_-}n}N760Cr1Mh8SpemC3ZiyT
zju5RY@03y7e4)#yg&^-7T;<`CqWXo3q<6Y?9=cjEs(;Fr7F{W%VOL=e_76o=YLU0T
zW+ISXPOgW;-dltio<(>;oHiPbxOVME{L4T8GHzVIqp`pOzT-Ns<2vrCao?EW>+G2`
zUs&VovAh{Zt@-II^44ZJ;r+>8bjOMBzZuVh{K?9Dveys)_q}@(?>2B>bo@8!`GZa|
z-NT5}gXpB)-Givt8<7!FjK;y(*oc{mXh?kc{s-~?@7_xcf6={n?%a;+pI(hSw{OL+
zXy<aNY*XR2{#pzS4*6^sn38wr=VE<zB{E$H(VK&jg|&rT6pEbt3%!o^7|icO(e0bD
zFn`BPPRtoW=tK9xIA`}a^nP~_Odd@3=-d95>q9#2O>dl|l+M*EM^EM;8XI$>$&-eC
zT0bNU6YPLz95i99uBY=@U?#Dli38tla!F~KxhFagflRkGwGd=z>_kgbN@?)y#<VlF
z*z&9a4g&K*X0X8wbyytXxNpa8$EOys3b-OTt9zWl<nk+(DLKZkyy~K#);k3_Y{uf9
zkK*>tkGzKAED|$eEtn!Ko3dET!2Hx3t(&X~1Z_>5lLtRaleEbeZ2)L~n?`bdQ+fz3
zlHekRM`OuxZ3dBY@=ghe4n9yHdiTGXbR<eRzEdvZvndjDO9)hw)<6;<DILYP5P|Z!
zK;(l1CJqymxLh<d*6n+U&<%4=8=}AKs@e{F9aUDPPWc7vwl1Ntu7mCqZQoj>x;)*{
z(~S@=BbeZx_Es7y4@o2(N;qW&t*^4OHlpmbm9-yGjnVs^?e*XcGQzMq((o?0KwE|7
z26@1Yb=BOzsjn~{m@f=QzE|t$Ezb7=7BEek81ZPZAfQ3xB9|*ffB$e43Pp`Y&b8||
zb05ZczPA#kO25y}LTgL4Q9LZE<Re!c&^o=UcUY#_QoO<FADLA-5KyWP+bePV`bV+2
zcuQ@NtePOOLz@D#aA;DX2}Kze=5O0990DWGS*mYnZ0VUYGTuF=U>v&-_~&dO*3HVU
zN2Ss|>nF7E9Dp^u+6>U7ya`t^_V}FES7_o{;I+4_7^(jdDshGRez1pNH}xHSWwfto
zd&-8!mCq$2Z<y@Bk5Zxhd%(QmF5w_+SWr%V$5W`b=}2%&!W}IWr&LJs;t&&R8Vg$Q
zkw#Cxpti24477#9s=UAjJd!@Nn=%N*8Hcz7wRI*&Mn_`u)Y&+9b~Yx)M^(<QSYBPT
zIZ8A>tPN+tQQ1_<2YvhfqukMVU(&eHue*RB#-P{Ua#U?(urFS^axvb1^ToJ)aVGj}
zWrg0i;0DvR<ud1_R|R{?l$#Q8`yos~+lfsDz4^Q~63Lr+VovkOoNzyyPUkLMjMJyj
zD7UVdpSvA5t_q*szG)uNpyrmo+Msb0ns=Botk>&?onByHFpf~A($}xCP}2N!H$M63
zBlYKg%$%NyGv_YE(C~ycemw?N7H~7|56)Q<4EJh&nwdQxL!;y7O;OIiN=f0m&1b??
zEwGDu#ksjVaqHG~j|JKd;WGKQg-WCe+?4blAuEEdW@FvsgL!Lcc+_nLEYVID8urzU
zw-FR0=w<GqylBD#JMfv{r(x0$;bL`QB)TzC2tE_u1Fw)?U#%9s!U5fw?DNKm9`z%p
z5}Rqx1m2YA{!Y@I^E=@j>e17kv^F_2llgRiXD!eIL+AyLV7{!Y?e8w!bvc>y(cpqV
z0yve+mFVSMFtrteQcU<Ve*g#cIYRobZqm-FAJ|-hkUR;>!7qR`^Cj~==7*X8(LiPl
zL%V_x5jLXjX#tSN0UO|x&~?BjF_BBHt=tuS)y>xfKHx7h7L&QugvT+zh*`>#aAr@g
zUwNXvy6?>Z-92gJJVLKD@33J3-sFJ=rN9B3BH#zz73?quIKvcy9r#K`O*l#S;JyrI
zb@LU$x%yZ5pv#*y)HmwW1I#VTAMT(ryeE92{t#|Oum(&I43AqdxwC^t9>OZ&9hHN!
zq8*J;mGe^38e_vtkKSjpT~SfJ5Uv4_2m4Lsg^5+ce|K4>XH$me5fv8r>`^;`|7bJP
zW1|rC*pmABH@|)_{_p?tH}SjQeQH{f^l<bi?U?R;Id%n~a0Lh7!%OXr=f@rIA9SXB
z<c`N{-?!aK@%|9p<JPO)9^ZujA`U9ABlzmLKWwMLQ5x)zv6)j*)#5iJ<PE_B0%Mm1
zc5WdS+k&n%5r)kbVlKf8U4?@fg@{kYa<>;0w%M&XOmg1&rS~k_Sgf*`z$^ue0kpCb
z!Il<fHW8qAkZ#7$`XPVPB6IR*0>eC_7eOMDh_|x^;@{MSyX##k9BOMqR}&qCK8M~i
zp=C7z<UGMcP>bMS<KSS3=BLTEky=+=e&+f3+0XwhW@e@}VWvef1Pu#2b~So&r+!Ps
za3XX}e0y@EPawc}nc9|jrdS{ZQYJ}2B;TI?9KO3v^y@AE>LY(wK3<ptGr)_V6Z{-G
zJ*NxdQabI)mas04{#q%Lp`3_57UiVLpTZkH<XPzYECxBgLuhWR|Je<F<K~U{@WYQ{
z;qHQ9i$%VYqt2=$(J4Hz4xC`lWOZ#lKK|&F_~k$UG8UGTDtgCtT*r0XQ{=um!Po4Q
z3BF7RwC|6-*U!<t=cLwsTjS;E+mrfZX??zTR`+?&6|K4HiM-n8zUO20elnlx|0v%M
z`{;VQKo7ioGXMDPn7n>I^X{AU50+Oy_dP$3SEMKM561J~vGgA4TXF99Cpdvb(!h}K
zK8(IfIVL71&D^PPHsW9Z?Y;QNfBLsrSXhc&PV}xkG!mwmF#l9h_`#vU7#p9A(ebh1
z>=5X2HgBT=hFNhosIj}gRDn6kPIAs9!CX&&)CLA!j1PYIUi|jgzlwd)PLPayd;6x}
zyF@UuV>-jx8`g+YU#Ofg0yFtq^kgfh*X{|TGKwR*zgbs&&KPT*)uGSHN`ngKG1#qg
z6kHaV>o!MN<Sa0DneT}HU-t$#n>j;pBX@YPsr3M6pjx!Ae_C{<9c`@i5x}?~l9m+0
zB5S4{s6Hk%VF}<w$6}VSwbMPtSJ}GJ5@AO+ZJ_p09$gMYx^3E*G+OB#?x<YJtRxI8
zw6_p+KtEHK?r!3V)}^miA54!WAt4MbrJ=IIs6W_k3eJ_j@<)5FTWd4cSqK-Pvn}JI
zoGuTHZG>68Y-@UPr4G=~gyH-bwS&?~+F0}tVX%BaczNa9RGHUSR$^mqS!*rKhap5l
zu+TDPr#8j(nJS`t(r@oZr(7zp`#sd>>(&OqEE7!T-CY>F>ND1r#8;U}mpe@U**qDv
zzN*C95kZ74dk99-xhskJOc*p+Ea2cin>0JDfy9DF@zGdfxd9*BFsM4iTt_nzII~$!
z!65>j6|E)II!BCho9M*3Zz)~NtIR!JXl0;LfFL@pzYyABYHMg{)Rs#SD!0!S;y#n3
z+?+*JwLqY!P|-Wih)KdG`roFtLS<6cy}53mhnLAS7AZ?J=5Ku**RQ^>wa>cBiO?=%
z0o?lPqQ?g1E)>dYW5x;$=Dg3+V;zrn(#q1Jir*ELa@ovh>YH4;vR-4o?~RNKM_XDT
zN7>Y$)~4DKY_!Y}rga$tGAxa#zq$IJ&8=xoh@c*=EyfLuaZKXDSkGGfWxxWNt(|79
ztSqN-ol(9*Hl%9-q{={Xz&$=+%j@2jaE<Y+kH7(fc={OuCt7(nDXDJ=61JNN1vgYL
z)~Z;$L8BD>3X_|@<_sPB2kp~J<ct0Ko{wF%^X}HFYQ~vEEv>hnQsjF3VtRHq&Yqo(
zTeok=M<0JgbGf|asp3A;wdaHO{fSq5{iesGYnxU<b7^F_Kfe3UEAh_TFUQQOQQ>6F
z-3fQ81TfnXE?^=TLDbGRnsI3ZGg^(9$U^`Md|<Y(+*cuU&l3oW5r&*Se<{wLyAajB
ze)apV1xjc%?x~$Q^Ks(TtYCgb_0CzS)M%)URleH5kovWtrcgPQrujko{>djF#i#Ol
zwe^{^XJT@4%I4*O5wyFO7VkPvS@YNC`bx~tU2|OqhlV}Q5$pm>@IHV)nEf8B!h3A8
zM#BvuG`uEEln%7cqJ(cn-!LrzKWvz8C@uIW^gDQ<uQsf97;|}mUtk{dhVZ=*3|bJO
z{Llu>ws>Ql>$JAEsPQKpDHu-Q!5`pY#=ORl`iZ&2CW<xQ(^yt#3J!ryaH&lLcPozC
zYI}1nb_L%SnxK&^I7F)`lNW9Pw;eRq$4$@U{Vk2NtiB--Ma!Epf<O@SmW)@%PeFa%
zokh5Xz->R47H<3eHR5QrrN1<Ml_xkhZA77+s26hvn$Gw5N@=42xFF4kXq&OI0gWu-
zVzqUz+Pt@-w&#2;(Fxg%+Dl=;?=`_pr8<<%=&A&0fN>rc^DaC+ng<w98#$Dp<|m(Z
z2~P05kKl_@ruV`Dj7x8bQ2RqlfP2^^zqZD?UkwY7@Cc{DanK8jq<$)>zO*aj7QsB{
zeF--pP(zT#pJ@Ztow+855V_BF%k~lvA*5ibzNYqEh~=d_ad+-&EZn{24c1tz$`=sa
z2{)+C*kpvT4B?vVr+Eayr{^|s9eJresaabCyS?EhBN$=~{pPnH#DD((|M&QBfAhcM
zz4tyA9IIZM*NEdZ6^?r1C_C?x=Uuvg(DCZHrx*#(BwXV-`cGkxr+L5sqVA-$zAAr`
zJ1N}5{FAto;{QJVNe;@eBlzmLKWNvb1@XYxNKBqS9Rou{BG}SmbfBL@GrN3PD-#40
z9JaMJT#P^=R9IvqXksT81a?&uHv~KgkU@pOt5-yacM|Cy<jthZgu<??&4v~jniLR<
zAxI;h7r6S)#Dqo{CL^qdEf2ZK!iHUdOxWzEvV(XMB1cmVO)#iigdUie@UE0Cm0#bm
zr|ij$MUW<Y2(^vXb)WI2*H5AAiowyT_}&k`AK!cT-RQ5?OcWl9u-EcLZKO8SWKExe
z<5Fuu$Pl%K_~gwTqAj_17#zLKZ@Oxl(*r3bM-HS2amT#{Sp4ak{@@v6!&j2v$6oQ2
z59Nfw=1QXX(rvl+4Pk(Z^x*J7U?Q+u`ICkmX%H4${}eCn`eYHRiCzmP4xI*0StQ%8
zxrh$F16vTuETA#-3Cw@=;m7fBzj;salmulR*Kr-!aZizZDm1?IbL5Y|PXddCzx8wU
z{)_muNIUu4XS@QR^j<$t<Q-3KbALK|en|R{_V*=yvR7Vd+4^wrUj;rnkH_=+@fB9X
zqyD~%yuJ#rAD4&na_{FS^lrI((I3Mrp7Tr3>GocF5A`Xa?jQ8W!=-m{=wO0C$3Ww*
ziH_a!MxW*Q_>)iL)2p{*VJR_@E}T0PXGPx+4Gl!CucGUqg<j=8G*l2G6(S3bF1i@g
zXq*>@Nv9C4y}bhsPNpz~ME7TNF*-gaKM_^MUtYLtMkQKCrDDkpV;CZtT)}3i@_J{J
zgE+W~FpwfK{gNgUbRrCDc7n66$XVM3EiRaD2(dUM9;Pfi)XRPSku40Exn<#s!lVsP
zT2HL6+SE|eDuLm>t72`&c4JL+>4A6Q!!TEaDLiGNUM%xlBj>1bDI-^YMQaj<_Xum0
zCK^cy!yq!P;pDKSgWv>)73UyvE=N<<W6h#Z2+I(VI9b2Y{ekWd-Tp<D&Dsa5H#97!
zue!ThJLw%*^s?S3+Lv}im<ZEPgzh0UDr+<9MA`_x@PZFljW*s1OdwL<t}gn7^1`Uc
zJQ7TP>Y?%>-~#W$0Cie8n0%@W`OpT`yIXbDA6gs;-cY_l5Q|k1Haw!$LpT*$?-i$8
zY3yOLF$uR|PW!X^ssSTc_1!@Lv5i^6y5J55HuV>Mt~DcC1>^?|rj3%}o%)Q4w80O7
zN9xlyL(a@i#YG^p$2vlBv>H&F8Lv%Y1_C>16`<LLIn5G`*xsDdV7<v2HlsZ0kF-`r
z=+!3}L=(sYM7I&`PwF_8alamGOLIQIX4@v+U}*M6v0T=*<THPmk1)fCKn)Z8Xnf=t
zhdIo76;*Dmr=YEtG-aqb<%Zcwo@g4NZ2>GKCb}T$pdL4_z8|YAa~kKxsP$L%{xFsn
z=VE2~uFba%4vq!dW@tTN0umEnXj|m6S*5`Gd?ztug*OtT)d1x5aWJv*Ft9QG#(R!C
zpq=!={p0nz$CyTbp0tuS4fq87L7o<RGo}QaXfwie2F_Ts((eeEZ2lBjg|RP?2d2Ok
z2v*(CN|Qe1YW%Ley{=Rk@X)T>#{w+(Bg}S8bRwM2Yn=6~En)NngP7o}yY859g?UW-
zkQM?|3xFBhDn~Y7R6kKJ#-9qczZpyO*NvmlPQ#`3H1`>6+iY*w<I_(+iH&tku_?nA
zEK-Lwwv%hGV~;zxQS18I`xO5u4Wi3?W07uP%nuHf<IOj(#KrTIk;`k$tB))wRy!kj
z;#@AYkXYN3KS{7IR|Xc~0Po<XQbl7!@Qa`fff(mG4h#-QpKvwsMBHX0X|W)rM<CQc
zG#bMr6H23^HrTbc7aF;I8yX(-xTCE##NGdx6MXNtzm2)ux6R9#J$uIIxl<`Nr6SY<
zW~n=PAqm3n#M07|;fZn1cWX8&EzIa4u&VVBS@Q|u7MeY1H!;T{7^2=-5ZYJy`Ccj(
z-JZ+K3$e1YsQA5(k07R29X5_&9GXw1kXSHELRBRN{HtsTufZ>rWp#BaR+r~3`~s$S
z_XO|C)53efICB{MHO3kO)g*XgtkDxmm;<~Zd}AE&ZC~(e4PdpO=RR*v;2SV`Ae^&g
zejRlO-(Z%nD}yjra4eYM+%EIx5IhPsqp=CjM0*IB#*A&ViN@=e1@xGz>{gkH%Z3ck
zhdp`Md3$TaXSE^ZSD_T2v4+5ya_(v##GLY$`<F8o;muH2p*!`LFzlglf^p1&#}Foc
z$*)uyuy7h-9^)0WfZN;aN?SD*46=d2^MdE3q<Q8IHy#_Hnlu)1-E0O?7SKuHCh!lM
zW^C?Bf<28?(3ftA56*y&pgssGp=C%k&HK!K%0uaaOMKR5kJ?Rrzb`y)yvEpfIB*7d
zRq(sLAvm}#Jn*}?bNf@9ltjzX8ytiKR~P1Eb#2vS5<EiN?Fer)HD7uzRpG!zNxSXP
zb!Rh2cQ?<#U=r#<Fa7G*AH;w9&wmqFug$9%yw9e(@uzwaw%tzn<anoSe4oPM+wOSy
z<at&<KlsjlneI<=4}1S8_eJ7-S@-+M^Y`J}_2k<2P=+1BSI7N9yNZb8@zXOgGBFm#
zK24HBo*ZDxE+B*o_4W1GX*OfG$;6_?6O+5%qou=y%>mjFQS5kuaIhdkh?yM^We6q<
z8_=#w#4@{Wm{=N0v#AM@Nt6XMf;8SE=z^GF0Qy-ASiBc~LfA}_OrEF?gQgQBN$8r1
zl-+;qKvCKd<&=-zcBowQ3}Hk*O!_z$(`zer)h#Wgo-W5E(-UXUUl3vYW<2-YbDE?%
z2wL@pkW*dRxdhR~<ivagLGD>z&Bb)@iQDbRJ(F0Pv>@Irv`V4$r>UE`5XACKj`*}Y
z4zG83^#RXNHR+ibBJ_MsVVu6NEKm~B!4J9Vk=HJdb7{v({|-xO+RmZ$o!y6+Jfz+U
z&LA4LMc}Y&tsK=_UpmyZzu(067R0$=2~~UQy(6Lt;tZizeWUJ!a^HLJck%JZ*96-t
zd&hNL$93G(<-Re&*SWLj{`&aNzejuUJ9EuR-`f8EFueF|7#N%n|JL>v4p6rCJoiYi
zJWmSOnsttQX37WOJ?M{Xw&C@2bpP<WeO}v~?7v8yC-tdqj+KwUW8aT_s|{&{&wTrg
ze_XivVxF>o*|x!*^!}l~)h?f@|L6TV4%hqR9WOtC=WFXZ!FN|QMW5fkb6d2_?Wk{V
zM5R)R*_l&u@%-sHJu@zPzG%&na$h+{#wTKIVk*kj{@Az4#68jcdDHALS75%_%upt0
zdYqlf_04)<LXeGZ!=n>ElL!-2oRP7zyewL7-{um{BrUPKobpAW^@JIwgT1=e5_7S)
zy{<eMkI>*SyG7%R_QLbt=bPk<j<dP3tT?MS-v~`d`HHF+ZLqzWw2Zu)dw(Yu@7{@z
zKl;F0A*>tNA-;;)q{fEZ2nL>aQi~o%zyfXT@#%8lVQ#}L<Enz&>{(Oab9!tGilAdf
zQ$x?nL2G+snY#lD6|4YFjftYIbcUMN9^?rv>~o5gMwcxm?595De6Fn8yHxV-_MR-}
zrjoFQv4NRC)>>>jL=%R!4CkL=(#b-*uB<M~Q{_m)A()SN@=<-vfaeuuWeo+R4CX53
zMHnX97t0(7QAn41uvl<i2!n$=s<8Xe+scymu?9>bN`<`Iw&L?vIR9mDFP&`#ZBFs&
z3rtY$E1#|AhP4O~d{GZHt2i6(o|XpvCRkfv)3+@%snOP;xmhQ|z+)}Aj@dh{=YU0@
z%cb^$=?Bx%Ydf_o>j%~bXb)N_<~6IT!VhZ(g`xk6hp;tY;Ow4C<V#hhi8fYGYzyw2
zn+W0>D$l0IMb_GnXfpbo7BixO3)%{}A?@v0U%nl81&;_|F)`*%qVlY(X&=t@N({mz
zSThW&euW}hNHBX#feAX8#)@V~AH%)0mO&S5TJ@XS`S3t#fCG92mwOGvJ52TdLCl-w
zJZ3pN2B!Pu)TyXe2i>cU=9(GJtcNhMh~}A=L^Pf9>@v)GZPl8?m_v1^sv<b;aUZi@
zA}yYQZG@QgdD0eC+v=XQk`|=Otnbt(pG%u&*~Drlxgr5qcoINxfUu;8vw~XfYk_#F
z*RY3G4%T~=@kkh~F!VjVgCr2+d?Mg)&wK$G$>1oOJqh|@W+y=cW>853#y{{{-&pnf
z5qLwW*PT(l8T+nn8e8ceLF8d<tSnkvE~k36>yb>j(#M&&apTkY&2N7ftE<9AH1HR2
z#`(vtIN!D&xw!cdu3JFQ(Gfe*SIWeh*|8X?72F2K3;gaWj2Z*&C0prfjUV74jfs2S
zCTj&_^cgrG{9Y_G<}trUn}JieV{L6&<Fa9a62dKneuIO<PL~bRn9&3Wq1l8M8Rdsh
zfp7^v&En#s+GthZDskrg`Iw$QZQ%_Ym6?}V&s!ji&_nBT`kwUG)|ejz|4I-Z2<^X7
z*E_*zS?mA4zI6T;@evj=7g)PW>A;6tU0JaiL`)nqUct4@X&k{ZK0X;kBjXA`pteZ@
zE$}jCHOYte5+ug@+LAoNqC>YO^9vedTTL{C)PBO#2+JHta~xWoz{{?1B1bV`W-yya
z$Op`WOMtO$rGv)|tnj}20NM(nJY#8l3qcNN5T-IY&S4{N-}pGzR#(&(2(s8Dp5S3;
zYeS%ca3-gA(U=#m>!$zU7b%_YF6JC~D3}#IP(N~Z9okq5mtjMP#*@m529h<ec7>xf
zH`fLG;A1q8z-897MM$H+H18$840APt{ZdJJs}0aJgD1DSxfUx+chr8X8gIQ(siwB3
zjp5VSgraaBX9EJOxdIyFq7l>=7G|lx*dV~z7tnk2J7a;dp70lfb~HOTQ+pwJSKNKg
z<psfWZD7C~YA|a^Ux72&G*3U~^JSGqbyrz9j}ScEZIhGWY;XY%UZwg2Eitv-?iO0>
zm;pwcPHn1A@6nu(ARY@-2sqE4y%gh<XEompT3EHbyr_OSOxk&wqS8ZSxZ%EGQwHq?
z4y292rHPjb|Ee`X$YIv{>eWx<pa1z^<HL`yt8LIeW!^~hHf@vOgjasy_sOSwcd~q<
z-@p_+f^`$7B#y)An&z;3j&N;S6_@h(tZS$Jxc=*QUpCJ#66f3Gj@KIxM0EsT9rp+A
zh9|~i`s|q~i8$c!$DWp5e+>sAWDbr7k!#_i`vVb>yaVYXNLtkbq_NpBal~X|fr|(c
zc6Aj+u=I)mVlrZ4!p3!@!FdBaCR`gr(kxEOg|cm&6PI(u7<dpUT=$h9$kPP4JVXhI
z6++gBS?eA<--Km>$)vWojmFN37fxXE0}H;`{grkMGKnY^y`v60(b!PkG(kL#j`|D^
zjl}u$7h+~sM4~2f2rf*7cz2QNjb<F}oS1M#N)w)DEw>+wmh?@Yd5kb#A*%U!6rtsC
zJg9Gdp-+ykEI=Rvxymap#3+jl{VEFYbf<Up2k9L3snSJA+K%fyOf0ydPKWq<aWtV)
z9urdv@5PwYP@2G=;SY8GjaXh;Hep*VVy7L#i3680djeD8UjDqNt+3_ILi6t3h4|I4
zeiOg?x8KCl%9_GarjF~lj_bIm%6)@^ud^+UFHaT6d<$oP#w*ezyyqg{`{5q$f57YK
zq2E5^`SblT!Ocnjq(2xA4}?<*KlAQEub=z;(YIrN`f%^Pz22OV-{*S$J$74sQSU!W
z-13z5B(HBcFS7OG#I%n5v3OsDr!cL{y>dOF$DLgElYKklN$)<-JLs`^j(g-pj6ftE
zeM?MZYdmyk%;+l>d*g*G&&5kGJs%g&&BWw50xGN<G<}XPjK7>{ZqAJ8Emo{$LON)1
zBtcNO=rv{04YLu3Q4+$SWr2VKI=a$_mO#afJ$C7{yB<c-(BM!EkBnJsj5ARXKCu&g
zcPs67{`iAm*(@H1&!Zdx125G<`GshGl^+vl2fLd-od3?PYd+5lMh|otOrE@!CcBNL
z_~5s{inX;R#X~66jJdg6!TB;UNz0`o>?_5E2FyfXAEpq~Zs_LbhMBXRw*n1l!J!Ml
zav$Gf=3vSYO{i`(Y+xV)%PzE;U#VNreu1vtV@08FZNV7~V;JwEM@5U<?3(HZ;|%(|
zi}2m4AKmY!rilw99tLABGC8$76g2cRbgkAmW`N4??`iF%^22xXZgcW9JWw9iGYIW=
z4HrJ^1mT$SV;6i^x1NtNR$-EFsZDH}k@a7eHc(x<*fa+-nf3r5LysqzgZY-)i%;r5
z7^At8;3bp5Q}6P+r(JjUlEz+RbKSew(dgp*FqnYAB-%_#Fo9+k!Y~A6n4U@+DQFFH
zz8hKq2rS%>tZ~$y2q|bg)|;#~F*(L})*dj%Z3a|*1(P$G1f?$R^$AR9+Ff<h*dgs4
zro(JXlAZNRN9AcCWaEsVEzAODEF7cF`JS|X=)gAIuwL2HI7ylbwOTbW700H!dVS4n
zRkUMZq{0BEPQB!-d&XL+G8mcSpvqHHF*#-+3BHPbHNiv8{mmEzezL^N(>8(=m9y($
zTjkx3Qn?tMFGHKHudT)6!kpT?>2**AZMJ;T?Z?>I-`~<Yo%3r-u7x%3%rpk353J#4
z;TCHpno6M%Zo+V-Jpo??S44(UyDvBwfV%%_XPCeADMGQ$`nrV!`x>jX4dc|XBsglH
z%|<-;wQy6e7ws3u1aOMWyKm4!LRbSlGKTjMxCvI$x>JxX7)PUnb_OnZ3BSQ`#@JIH
znDI-RGAbupp(;19O`m}LIE#ofm(^Z*YvE8IjRkA=Q63tjeP#U|hH*@EQog*(&lk>N
z+Z8VRPygxvj=%f+e^N7n>qtvkGoF0+c<%mrAUVIGZ~W&6{qIM)oQvtnkvKIm91|n`
zF*00HcW!$V0r(Oj5Mu(=1>Yu_^Q5g+LiHzgNTxW2hXjD+&3HjD?Ty5&$<=SbQm^I`
z!XY$YU0GB+G#nRmntf`&ieMM`VS_aAfCeDEA%tmI;2~W$KDWnWfB!(to;e${vuC4P
z8}!&<OrwR^*1l^C!IJ<>@T>9~=P2!JwdOGgY}vG($`7t!jZYl-KfoSWG!fwmP%kzW
zfCItP@VCGpd><W~(wG`h%cl92QbK2ev*=Io23l-$w>4L7thg<E)h-rt$s^<#92}3q
zp)tWZnqIw8?H{sm2W{O(W6dzpD|qkjDJymG5539p8-gi~UFHqutV5M4%cc<xM)1-`
zy{@`&jvtzgz!mf7dfd8!cHx@R%tcoHOug1u@5b`N9o1n=cxgDwwNZ7U^30&=4gBmY
z92*ok?xCrEVQyACBk)qVWbRXRf^Y`s+%@XU@@#tF^>}i<H7=PS!LwQ636EC<Rcdc=
z8F&cIyJ~GjIBifcQZkNRShyYY^Vi)M<CABi+&`wcXb`4(iShy0I5fR`GH8LSj|D@R
zIHdHsLSOXc1lOA9*suUx0hbA%!!u%Y4d#RszTev3iM18sBaH_H|CoYA8y%q_@J8Qe
z1be+2%hsr63<y^+Xwtl@XVoW}8HOK)wk&NY+C?-DFl7_9N~tYyX9EEEc53Qe3=W@)
z?);F_0rt^$-}hN(g>rgFSo+##RZGJEsy7?Wm7MECJ<=S++#|2Kh<^E>|MPF-zy3e}
zO|%i4Nzz=`=Et@9X($e)`OoOXu8m(z3Z+c$bJqBCzN!d;Kmm?(43F{LN(Yz5^w;Oy
zmyhYc33pPy5BjIU0Z|>nSI7N9yJAJe!(hKAE^N}K-7-w@2)}IhZDmE1uMo2k9uu00
zOd(Kq!tsp*s7(hkQ3%mvjs)#5FQWB+yQxWcZ6lVKmo*u$D@|0M*>To10Wvr+q=_?W
zDuJj?Y-vIRx00#_D2f1~mgk$UM&2q1Q$5Ip?`anpW+QfDb3==51auHFdKSu2*o1^r
z+Fc?fRJSGU2D8BPr>D|ks%GlcDG`dZ)?o5kRq7K=Jt7QbvZF8DAF4NkA=+6qlpH$_
zAtE8jL3eVr#SxN*V9?~rOHF)^qG$TwJ6x3(J6`#~BnP2~0E(SGz)6~PlYoiT(le<#
z3*H>1PYA!lS(3$D922YlJn`bvdx!Qa8yGoIy;(T3vv^OF{#q<7E}B?|IL<+YDSTQ4
z!{nj<DoZQ9z3sGnaeaLwZrr#Tzy8&)<KvG%Rr|Gf^mJUubzH|iRqh)Td|fzq{;#1A
zjOUMgP2!(<_dvMM2y0k49}n~3^L?H**u%b^;M1(}P{+3ug<Rxu>^)_EA|aFF>gMSA
z(Qy-V?Bs6`_eW{+ckKN)?vJ-0bce2w^0fMW@lWuCJL&!7ddKJQfquBpQ<CHN$NdQ~
z@u1hwXWo5P-mX=c)>Dgi5&WC7Npvzi=(A_0<NNQv9p}%_M!8ZjGZ51`Fc8g%=|-Re
zd;!xiHAVL+F=(j7C{d#NW-XXrn8!)$0meC*vk{JXFi4w?daR1BM^M3;A*D*iOhf1>
zcKE|+wyE*`Lo@EsJb~GX<p+U=%L6kKbB*1x7tQ*bnNhcIU5oj-8!<nBOW`@oBr(ts
zjw~<E$F-{;#Rng}7YO0fIW%bX55&~uRE&>}M5R(x9O|k?yy^=hiw$=^k7|8Qv|dwb
zVIGOid|JOC;1V61gi)+2U?iz+R5q@zxE;)IXhJ$d?I(IN2>|fvT@ql?Z&}giFqL80
zrk&Z4%2*Jft%<{$2jLn5CPziE=J-6od}ZyT*ZK(4ls_DOq^^>Kp4D3_iV4YX=<jp}
zinTP<he@l2di=la{Reks$C2-gCV(nb0p;LLI~tJ`jU*ar+?l!WK6BUoT;~hDx7IuB
z-u335b7n?!MssE~qmd&enkL&Dc$D`5-v9THs%|tI-E4|vOC&3StlG7A?lh5+k&zJ@
zk*Dj&w}O>>th=->;taSMEEQsX$4%B)xRxWYuq;FHOr9vtsEbzwx&%hbQ|YYy(yu{(
zWMQVdu`Z+k8}+L3flZ986;L!FB&Kex;kd_oWLEtR4P<S}I*#U{Ai{eH_-w(X(&JJQ
zO9NH}tXHw1Ab%_#5p-j9!@DTpeD;^}_PKS6jDzU2jD?%?#f8K<Gtm~+UG2sN73c3@
z$@0n7kKAuqYM`VFnnJ$;gHikS4Sa09m!p4CMs-uI3QphBbFJ63)}^iFML@vLS{N5-
zSXs`@hLzLNk=mv)#5xOwUYahr`*7O{p5g)w1sUhyA%{KcGp^c#0qx6qoxQz%Uu*OO
zkK;M5gDA40vB?<h;Mz~H0`@2nkB&KSuH!YM^HAS(G}f`^z@kS9J+^Tl&$vZ7M!&MA
zLTM6$^PZJaY_{Z>0xS@?2Z5oV4)*s|pIu#tt2LA=Ugv>t>I%jZ79Ci7O=URSM-4;R
z%$Re(C<+g;1Otu;+@W8}!fg|!7i_>c=#KhNV}nh!xZXpM46UK>a>}2xaTs^~PF1m?
zyaV?*|EkeA5Kf?=SQVU?^o%}H9jpju{GjlpT+u?|2xACWiR6Lnz}b1MoQle;lYag$
zKTrSlzx_kHzoqr4tyq-aW7kc8WNv#}c=B_tkI7GunFHA5@T0e0P5=A<_P?Ybyzxp}
zT`3zE2MFcW&y1l;8A~4&#4!e~IH5<~pD2(Rf9g>bh59y)p^l@nSSy3G^z$s^b(S^t
zz*a_h(ZLlR$^~0C3TeS9ta%vUWzIj!7fc7i30sFMZ|E78WGl<dR)pEIXLio{kOp^z
zNC{AE2EM5-u3N_MvDU|%5N!k}kcE&BxDVbU?+JINn$wtXtc1>{yxNca30I%c6_hb3
z>q@1H=Dgy_GB6`ppva947_<Y6K<b0S9hx~ln~HPIIQtA-L$STMw3?PyH|<^#7=bU;
ziSyNZYz9zteMC1**G!|4TQ$;9<8j(FeW>}gDHsW#RZsegxfd8zmX?)QN$)@eIA@MC
zw5HR%Xi%|;d&;u&JUlo|9gTbXq<{-SHcW`NSdl0g&ScdGrUNIg2XHn$Q+m&FAT485
z_38JUrVDM+bkgDSa$#lD8)+;SqGqbA=MCWp?XC+q+Nv{3$vO4El}8#=!(KBK;;Q@l
z#)e?OqGwoxshkeVGL+K6n`kysTQsk$-*KCWn?qdWg169gXawN1<sPy^^=)ig5shu^
zr>4rm8WmX=cmf^djL@9sp#Jd0@*a-|;R*ReORU%xPC(Zg2f&C(n_7*UV1cDN8(p%x
zCwju!l%gS6t}$24sIOH$;b33%l;c(g#(&@+io2uKYT}-`XC;@FdH`B;3(7x_d!q6_
zfgezvu-f(}7J%(}hc;$~ll64%vyXhH;DyZ#>C&YOX?1let*tDj<t3ER!ij?Bf_!WQ
zw^b_rD#pZeJZO_(6>>U$GqXl*POnJ-GWX95mIIm~QElFg@6+*V;cVKc#V-=RX!@@!
zoNdF`9siKPOgmA0O~SW2WVJ9WEmXwhRnlBx-po!@3t5<R7#7a&GLd9b7IF7ZL=iri
z<$fnz3BqvTip0thF$K1;5|KZw*3$moeki?a?D*59Ai@vBgn2hRb#Oli^Mc~SjI;<l
z%rvx<g)>YRRMwNR7!^n=lOS#cB3&pkI3U#wiPE!ckV%alp7?#AVK&s8R<4PO@2TF+
zYCClrEc^_HZ*1tRO}MRj>&>^Uykv51plB4MvA2H^?lLqniwdg!g0Uv}n4BPr>Q}*#
zXhJD4nQ8HaN?=+HMxEqw`{Bd};}Vl3_r#n4UoV_tQuSxpMp`d+N#hAM?!)YYXQPA>
zTrZyLz6m$)saEP+*TK_pq~WoNyPngwP{?7yg$vGx7(FqX!l$u&1lk6^vXhY=Td~`B
zCUun_|KvNnyXpPkevm%={m1Fxut5zbVG<@`626Jy`6<3$ef2+|ryLm5>4$;)W4b+x
z=e>72@{T{xi!9zAzVq0-r{Xy8ar&pmGk>~y>X&}=dG?)iVk|Br{pok~5z)CR&FI^I
zy3KDwq<FSS@U=evxbak<jO+BI$j9jWsgX2aHu5pPJ}1UUgmFKO`|4T#$l&Yxa~;Hd
z8J9F(#(W9S5J*^IBi;uA=Swecq}N`1DHV(CBo>D@5*VU5z@4FaHQj@+;(hah%2P~W
zT>Bn;nQn=9vfCndL5oL6P=iN@p4JEKYGzXz0w<JWmBj_+pG~bsJ=IWfh@VH<hA^zW
zuw+ZH<)t;fpHj0`PjmAKHW4T*414Kte>>f~b0h6;-%3|MdOzK~$;P9B%FkGU$oZ_e
zRl0ih!*uW7O+b<^UVcgO<+QSPAr%%bsQt6)aBnN^DY9<pi`T2y>#3#pSP!t4iStJK
z0+-eyM69v2mSN|*ek7D5R_@p;3gI8TDV82x0kf$j9sjnni8NCPut*2&_~JgI1Ae0@
zLU~wJ8%k~m?&pB#06w(a5}b#+N0|bQ-$hZSYm_P~+loQ%tHD^qWXMBs<$lzGeo|Rh
z46ycM&2yyZ2vc<*;iKC^+A*KaJ<3C=p!E^fF|$SU!U()rCn1OfW*`IWt`Nu)XcyN+
z%GH;7dPNH!^SY(AV^?b{)o*5o9r`nZZ(nPXhVc<R#9C>tP|=#Dp!Okz=S(ZrTWb|?
z6~Q=4oBH9d+SK*>j=qWY=E$1`NUCxXIuC~k;<4=Lh?mDbVM{zd77RqzrleIB)ee-p
z;GV5ra`UNBUQRRlWj#db$mY&=>UL^+PyIY~taZUo`sFWwWhH8<sC6<{2kI{rP2f@p
zD911%3V%*|sk(nx_;|+_V{8&dz>N|b0WjwfF=lxW1rxLer5;w0;VyCC8%|5*Qp#%m
z0-TQqZMO?r5Y{g$ilBV3uB814XK5(zNo%#L>%6hCsqs@v)mklmcI~R*Ka=L>a$W<`
zC!Etmxvc5IJfFcOTm!~zphRhiLI7m~ZN?&GsD7Y)=mK@~x=)ccIMJ>w1%K*_0s&<f
z`S0%Bvtk<@Lt(-uNt8`|gU{p_Rwh^`BFrA+7B!=7#29y^qfRJD^1>%*6$+kq^H6Kx
zJ<B0@PDQ~LlwO?emB}wEk9pOzle&$))U5BQO>GOzaleU>8jIF^zMQ7B-~{6v<k6SM
z5@Qsln{KGDx)xjy{SNLbr!ys|p-umv|M&mb)?UZQc&9-)PCFEMl`ryezE2NMcTO1P
zrf>Zz@73jEdhear(yOmrN|lOe{+#9q;T^8$*gQalE})2}t>#hm8;d7gwV~9+a)-We
ziB8iu%o(HsFK`J8S>fzDls4dYuhX<e7Z!#nQLs2f@e-~_)u)0#9)c)~O$u1ApbYTZ
z|3RM$%Jh{UFu$PSC@-!WZ!*GH6k%EV*|;>X25TM~ZMw!XlpJt#+pIT)uc7d?6{ynV
zTCZpqrWy5<`U~r>k%B^~YI?wU!WtT7MWIl(!VA2DMxY#mHqr-HpbJKZhtgvKO8e}B
zR_XGp?@VUS{XMNQWS548qF{Kex$SN`&^!kn2HpqEQ-U29QF>WuK2+ZHBk~C34@32Z
z=4@56*+N+yN<S2;;1=!172nhh^Qh_|T89h<h2H#p+4MC-eN+#WJ<BVbX?6Wdnp2%j
zLsf?iihIVE`hwoT8g&+!XXpHmABo06dwlMp#v<bbjHA8a1m~V*a%I7(;JFDrA8&fW
z)iHA|ZZ13OSKKc$pA3Y5&H7%dRd+P6*HqrL`zV72n&};N7Kb9ZxSc4Aa78Ev0;|_{
zOFRaO`aW|3N=W8!%{eNUKBcep@Apv}wd(uA`9_*k`QY<dUP=F;a0=ds<1es)15f(q
z8<>-31^2Lo)~qbr$0i!^2oF`r$Dq%=-)iEbw<XvYLb*7{MhZ43s4c=5^0eh4^Sx32
zNHA$w@dwY(CLZd|W(iyn+dZskXtTNR`eV^+UIBcezeMjXtMlAXGaIV=FunQu_tXFL
z|MkD7zy9mLN<aDW+lud`_ul<sdh5;C()V9`IlXZCLRwv^q?MHg@nq#xF3)?jP_dxC
zWRuY>7M7u49K%tCsI350|H5|^n`2K;r^8t~74JV9pLC!2b%w99jMH`gD)arS!+G^Q
z8_x!?X`=X=gl~0V=L>$2D@(X9VP}IFaV-vEcr_|wm))EgVNLQGO&U4<W&(!TLcpPW
zCtBo&`$0^q*|k~Kg0q>nx3<#3?yfDcICqO3dYHm@JM26;5uwKq^nxZ!&JxhnrAZ4i
z$__1N+>RI<7TRWrHOazA*&nf(qw&r?%7@_&7j|MO*hQrJcUv$wECe9ZN~8%B*NN?>
z76gY)F}zHC0fTRHK=`n+c`^O@U;c#_No6sk9a}ry-`-AJ_qS5L!5IhaKGb4F%sKF6
z0<*#kW=<0_3=V1N?|?}F#zGQZ0zZ*RZp1xJ_Mym96qjMEJb$y(k9|4u4h$Y#2ns1r
zkV-dJe(6*Bk<aMaD8DG1YvrM<xK`AU_zkh<iw}f8@!P*NA;VOG7wl5p74EU9#OHWc
zi%J6ipIH<V!bQT2EfEh6s_EwS+v#8a_2=o<t$W`2GzpV136t;*4bM&S_5D|0d%_uC
zkm<Acf@WLbGWz{^nn&Zq1=yK;nzK$n)5nAB(~-F?=DE+0q}PYjKY0I~-^%g9?<Yl<
zt8ab8y(iS+^gBeqC%B%y|0f^a9|7xgpPv&6=iNUY4Wnm(_vb}r6yK-fgJ)5ZKfD~}
zt&7L+e+6+=wGmgwHTU9K@J>hK-DlH_FI-69|Ne_<d8w4Jyujij&J_v%4Zg_dVCWUt
z8+bw#Fz{{k9a$kDMu6(~`zVEQZ3TbNZgBBzdKn@6%^TOe>mSPqgw<>=n%6o(#i*Q2
z;ykW$sbV4M!qSS}Ai{^i`(yEeKo126n7+TamHz$T{yF{pUw@Xi?%hs%yF1>5L^I8!
z38xVB4OAa?V&lRImsaZ+u4oOhsC5B~pp5vv`{~!e`nPoB#%HOg`m#fx`eG@DAQyos
zyeDTiA^1HW(7te&rdB%-tlOqI!$UQTz&FD8emYk9v>oM{y(B2Q!K6X>G!j5XIQ49#
zMDzI28iup65dOy5EJ~rZi_dU!I=+Fk-qxDO^bxu-jb)tRHN$y4w3l?j8+&KH%N9Ne
zPOOK(jp$2uj-x0bZ|VczjgSzf)<Cb}CJTWZyl}J$<;iG&z27y?D3BNL$HqA7IBvVX
zxwL|SJmRb!TQI4fMA`~|VL8EWca$Ns;0}sFm1`xI0_|aAr1KQs;Z7?oNf0z51k_qm
z{fdwrcWS_<)vVeA0>ujL9t#$YvlfI3fBUgU1t%VQdmt^=P4lBZ5BOung+df18oT-#
z2QR<;vf`^gv#3~JvO7M?V3Qrz19le%4H)%bw|kI2{rGoj``!(`6N=wb8Nu#M+S=ao
zS$xpSwe`)kw6d;ovZ#87vJWKzmK1YZkAwfX<wsDD+fdp<DU45qtejiZYscAXxZ1+v
zqc^~!t7YN;%F3$O*a)Z*x(<h3mAR5umzUkI?M^e{eiBQPpcQ~Vl-%GR_p#Ezs>YUJ
zs=vqLAY34ZH4u2Hcj-3;+^Yc`#unCBxIQgcIMeA^WBs1S>6R7qczXc`C|DVvlm|W_
z{GQgBHf|wY#nRzeV_z{8#Pl9epq@FUolnymPsj~?^nu`yBBs^EVn)A}FE|x$6cI$v
zkKPqr+xzME&8ul|Z&&Yf-Vt;|w1RPJy2AKF@h03?UG$-A`WRe64swzjwO!GLwyl{^
zZfgMO85T(G^!NYpkLlWHH<S;`V*sc?SHJ4a*wUx!q>s_P^8?SuaU90ik(2&tpYRTa
z1d55Jg?zevVKJ38ALQmThI3`1l1iM(g$0!8GWZEC6+n%1D8R7B1V5nLl)=W)hH#&8
z<;};cHF6r19lS>uX9XMDeRz1NSQRdHRhO{90-wtZi!O6sw1B`y2&|n{fPQi81T;eZ
zCwQUgMj5d$x_|ffEz5&0Uc6!@7c}2vSah8CIQNOMjx`W@QD@x3p*Z49?<LJO%u80p
zo=n-rCzg2NX`^;%3(l6>g&Y99z(Nu_hXNkO4cDIPRIk_=ESfgLBVdE|Ac`{H!<v`6
zpnSoFWO+gJTmg$hlvKI6&gKQjj@or2?e6SYhE-KPu$E@N;k>h)Xc>9grJZ0H8#CzB
zY10@i)==U>5179w&k8%{2v8{v-B<JzbeFR-=_|%1&<5@~pCKt)KaFw+dOXlO4fj8^
zkzAMq!%|p%jlv9CDtuw{fX@|U<A`_y>RM2JuwEkF7#D3}JPn@2aU`7G8J1+6874ZQ
zF@o2nz5Ts(@9v$nuWvS_P%et;?5uDU3o9&)!D;1*1u%;6>1ocn!dh4Tsyvxv=5R-v
z%NizFnjZ~?o1|g9=`ZF+@iQ8WSZpHWWIlp+L7PxILf60zZ?XaY0vJkQ>elXHL8mm(
z6{YuP5YZsUTv&^RH9BJ&<sLK|-Ufv!Q>4omJec2SRZr@QYgb!XE03n&PaPIH3onON
z;8E&pOt~Kf<9@H1Ze0IVJkhT6Wjrw-H0zwJ#u=m<4<n@<7O3-cv+3f6&9t_*ZdbdD
zi<Pvxx|}xF*V5%n7t%{FUP-UL_PzAhTW_TI-bE?)Ui!(Oy_^2*C-0^A-g_tg;H@{s
zL%y0WUEDN32fWyj0o?`9LY{hToELbbT;>gZGd7q4jSKuf75^B)e6MNEqk((p#ixa5
zNqaWUdGSv%oUJ3+G*Ns_!nZo09AbeyB?h)EBFU~L7J;0-G$+DWsg$hvV)smbo}E=J
zkQl%)09x?s9t#*IQd^0kSkMH-0mWE&p_pNUVd27V9Q;hPlPzDEPs<|CtE;P?%)R4=
z2}uMNmjJfrQYo0Cv&)8ckxjxO9t=)Rh*&Z)*Nzo0TIHA_Z;NTOpLk8s@X2H05)*f$
z-b~#l3qLpiTM=G-@#XY4|I7a^EiNr-V(zEIS}ipju}ESe(?JO+xWE9wym_%K_#&Xl
zXcDGB&9ex!z}$fs!2=f0T<Trk8{Jk6fl9z$<uLh;qRBf6b$!IbMQOc&aeA)Vy-YS@
z*qxZ`hrxMpUuEiJ6#3LQ|MVw+sxwR!xIl437%~_o;V4WDR&N~o4ufj<n`$#mCktYj
zFcxIMWqWHo-MV=zee}^sw)Sdw#86MdBuv62d~?I|v-o=L)&Cq6c|@&8^X>VO=OL`&
zcMx>H^hmTo$Cvun<Fg8)P=@e4e&hB=7x8<P_c%Hsd2pY<Q_s%Ev9gN%9Y^m_i*e*n
zi5_|HNl_nE{z3Zj`~EwxKBI}4DnD}nPd%#Mp|BhMe#ZJU@$2fTe@=1EyI(kt^E-9#
zL4QUmkKx&~M)2>v`(H(bZ;IMQEiHJ3`O!0Y+}Z3&di}L4>2Ln}&(bTezARoHw>NAs
zYoyqT9`4By>fx^L=vaA)Z-W0q$ctqayR*gJxjYnP2>VzUAV5Y@g|f$jIPp-tv%imf
zq9d&Vu)N5mg@whmy2cLuywCkbScjk!VL#`3Aml+9kBcM@azEVPOZ&S!w#=y4cGK4V
zo9XH&pQNAv{8#Db&AV=1rCbtzaCXdY+QUs^w`;)<XLfB~ysWj!irR)NtiIdI85iv4
zr;Z4t5Pl*!rTqxrP>4`xTwxy|)M^8to}P&})q;S{bqK7odN-6fg8K=)AVMDT$umW7
z02H@5dWP#W+<_sGB_J{&2F^x28@Pv{4F%S`>V%+EeXM>YORghO2D~z5>z-XUQK!-O
z6l)*&ZQc{VY=r<qT7<Ejg%tukuYpib*eZ=Ez$)L9VaN5iJ4aiva37q~zw<@G!e0nN
zv4RU)%)5Q>N;h0=nTBG3c8~H=!Fc50EG+N?B~>mzFW6%7hLCVNHCwpJ+%t~Z(nC0o
zbqVV>T%aL{L~(-~GH6RreT)K%3bI~g(<D|62o~u_lw$DtD8q1Fg%Se+GQwy2lrhA)
zD+v6tC<*0dST}+1tbYT>z#%Ib>V36qfA^kn^KN?i<?p2p;atABl%}%_N{i);M$8ds
zS_5-jHtJRPTTg4?y;QC4S$Md4;YwP;-C23T!g1U{a<243c`<G1S+2a{Jh5!Cpm%mY
z&6k!^uCT1UX6!1l+iqE?kA)n4YKslE1*OaVJ6F@q>mS<^^5%`t6tAXMqnb(@cZ*AF
zUJs!F3?)~`?yeBnvKC@21B0^aM*o0Uwr~+}EC^OQ1iZ{)kfE^r%GnyoZVQzsIAH~t
z{^(o!JjUHf7_4!R#UAJTQAgH;7u4?6l?}t0O@_E}LlI{jL1;fxPGS9WG{n-ap4!ct
zXprdCFvcXp?QlZ~?rE$DXHX93=F}F|Q3*oXGplr>BaAI9pJq5S2m-@-MYV0gq$xVJ
zn3h&9B)g<jzt`$DD_R&|Co2DFAQ~vxA;iZGTPUi*6~@HSRw$gkHa8cz+-X(QuYUFG
z^bbG%r?j`TA2?<hf*+b^R5z;dqyQa0H!!wFm!qsqmoqa*>D8Anrayb{&9uH!68^_{
zf;OP6M1hHA3Um>n^}+s*=q}1G&e#j985EPqSWx~ke<bDWIY6+WIt8tn`9?uTJM4uA
z3mMUyDdo>bilN@YI(KDtEtM<F##e;pcn<=Pm=jP=FiiN3g=Z+eInxM1e<$6%b={V_
zo12>|XK1A^)@69`p-xyfcDjt=p74fqvKoRh)>OOR{JXk#LGYMS-Ld+Nv)ZtnL^+5D
z9Ap~cC$cPX+%RR_2|ls8`q;{K<_d!6VSPhCcppU|b<KMdc*Y6{Hd2shp~QKTvmOVy
z+C!m*a+bOhP(l}r6_2lCaUor}_(ED-z971?mS$$lDol8*F$l~o_tLzDVtu|)6#ZC9
znS4cX(;VE#+6ns0hHfm0pqp$&5$(@KK3Kf&?`~UppU<H%60RasfsTWJ8bhQ(_QrW_
zDE^==C(2Kcl?K>|9-xE~t;4bwMH%w|^mnefC^#_wz(LImf{%Gx%?0F-(xxf?0sPF1
z#*{e2s<33739DVsqpPIFg%#1N#e`)ZE@v~dSXrtMG^S!xf%5Zsqf(3=6yYP@150J*
z6s+XBec@h3{Zckg;yzLj-3HEkW4HxfpB6qf(~WDNq@AssY0wi+XikHUf!|nWF&;Xd
zp7UE;UbpfMUI8A4w(&d8w+nfw=S8*u=y;&HDrhs~3z~r2W%CcyVYLWuvata=1n-j5
zy`c5s+S7_V)t^43ABLLSuzXxuyJU;+R;yzkhB_j{W4<FT<I&@sxle!Kd2B#8*o;&3
zb#5*f3UQ5<r6uugD=TSzV=Z0S+(=h0U({UpVtVzJm(%Otf5pnNzxeYX+iL8cx8JnV
z3`^bZoh|nbxZurh{HYAwCq`hzJRXAnJ-_foI(klcTG}s?&(p$JmCx517-<v5*Cc$q
zgXj=$DOhaa>M)z1717oLK@1-Z%=lnv`>_?#Wb))4TkVL5agRj<N-LXXD@*S44sIP-
zq@WmssW{dj%*3o1sP#*i)8!Xl5VN^rYn<3sBcebC6rns&fN&<284x|fv^)q$EYSU&
zXDDZM4O7H~iq#prCXO|+a^@2O;>_-zW>XAoNB5Nc+Y(-S`IYq7fAcrq1<OPPlU)|0
zx3I9F#Z^U<T<q4PFYs|qAF^`>6LR|1yEyco9z2+?3}5<!esG7MBG}y1E&2$?6Q*Yz
zeNS+P$@Fi<*h$Dcen}Zb9}+Q{Cb<7}`Jnt0)kk&UkMBq?q743VpD+G+rpWHEsu&HJ
zQkZLYBf(Ik_=1sP$05ougdE=GC8nFj2Rle_-@2VXc>j0lH^2F9+R-A#^_+xBn1o6A
z=7#5I@%0)OU(n>B$DgY_;HA;WgS1|hKTN}Y(u}W9fB!*|zsT~DaBkX1V$6x>KYO;Q
z4@cME^%%v|*Jtm2$uggJe5`K_-Omqytaj?-(f)Xbdg$ZHF>?4CBA=u3`K|A$3jzmy
zi^nOK=hBbgc_Y2^_8Vz+ZACoYjD<P45JCV$I`i$8W+7b9C=V20@MQ39xVIaamqQ@u
zGf&tBtr9s4s8&5lamEN%G{aQ&c^z$^gTmQCY&asoOQN{Ab^BJrV+8y>$~D#l?N%$*
z>s7IFC{*IiuY0#|q`P;nEB`$U(D(KZ($%ZCQnxouo15#lOhI6Uu#SyNVReG5q+F_$
zmsIa5)f-_If=XNv9;r@63qKJ6=EdJ42x0`s8YR{w^W^1Ce5_{#LalY|W?j#%IN^|f
zpXCDtQ1HN;at~!nF3n_ty<m6DdBp<@;ShXTvSvP&$PeE*6Gj3&=S_diD$=JwO7C&o
zN~DaU%D@#PLTS@h?v39^n5M)kA4?ytZuA=EAut;$Fa#G1JJqhpbENnh2_~K1f+`{f
zvnUcCC_e<C;2G9gVSO_yk2vjK^$vIhd$urR?Pi!!PUIs1>7L2}-=^7FkIPB~rX{^I
zOpW@X;m-y%aB|!aC?Vm^eclRvsBdGE(z2FBnSnqZ0V>LyrtloaDne8&5Nut808w@4
zH~F}Y#ywgZI1j#CxigF9ncAm!cpCkN^##JrmdD0C3LgaH2vezu;Eu4=7I8UTT7{w}
zU@7?Za3eNt<*i*6DsS9>vCN5c%cfJwXYf>n&wAI2MU)h5>J-coDASJ`J1BjqbD@Ax
zJFhxv-K{>z%$%gdy?g1#XTMLi!#%+Y3j_q8Y6ES{Vj+N#9jmHND4q9r6Ur_uH*iVH
zMo^9e0G=q^up;r9d!hrZ=}k8X`a;-Go<psx4-a-cZUqEGj5U?#aMKu+U+T}Je!73}
zR@&aWX90Vw9cQm`CeYH-nv#W*0hlq*qIjl!l&lDN!B-TzCr9nnY1dMxi_!vGbK<lp
zcOQ;Z@<CCd-&6Ft*0|6zly-tE791$}p?SQY(fvWErMlfxJCD-F=4H{1WwmcMwK|P-
z@7`_G3tU2S{#m<;0<R++5U!}5z|`ew{Gc;MF|)tFWiKP2T>UKl^FRMQ{oUXHWBTl~
zo2Gv#^}NRRSXW#9euNlSv?0F7*G~%;Xe+Yeg0tlE*^~6%k6uZC^Pk>J>#LPi;A}zl
z0dkL_V1|&IR~Qql%{v-1D3y@k0235MwszC{nX$n}3FZKZK1$N1B`j<5i6bixIQL!Y
zjr+n=1n&sb=>cc~LUmhMseiDzLYY;!0t6fhxdOO^(4MkdnoG!w@qjh2-G>T~3d*OU
zHU0g2chvt);lYvTvu3lQ@`tHVD)<~PtQfbpZl~+lK2zR$PxBI9e`*K2datc<Q+D04
z)?q^iWfuzbwq&G@BN2q!fI=8W4s(`JUid>DnLjKy0B%;A4}@QAte8`IC#i)Ms`6zn
zoYDB`_Hc#Ac~P9N8H#&0mXJ3JX5dL5;})^3c`R3`sIZdC1(#7)LhspdfFeXcHP2X)
zY=scYkYk@$1>Eq+ge8%W{?Om7=B{bG37t`2#E~D+1eA236?9hwUYP>Rj{3YG3Y>f>
z?52g&M}jdj6I?BNUI9L$C%95%PPe6-X=czh?qRj;xdtp?E>QW*4e%FiKrx-msf~<R
zHELGri|QAwT^ANtgs-ZD<_9b$nPVY19xloct90rFJ;4Paiafkasb=PFJa6ikE$oCt
zKEH3IK#ah6Ra&P(sg(8R{Gb2Ne@y@Q5C2u;c3X6kjSQmA!fTY9Y!pDrzr4Dc=1Z%p
zOE&epRd3pXp2J&UxefAAH{90_MB{vxpz>jKvsnS$pA|iY?+98IG>x-G;aR{dEL^QX
zq`?Z_#Nxm_dz5m*!Ln%6Y*Awh3s=z`l!3Feq3EI>Q?ys}kF#Lj9n%L8nQvcnADJ8q
zebFu+>5uN>JvJ}EZ_Vbj9v@hXt*x)7%?lgp%B79;;)|EkE3bYpUC?jf!x1(pFL7-N
zbVI31oyUJ}!YEME^9|2JN6&^a{XH9>7M>-|>2NyVKZSTUz`;!vUz6~i2t5=8m{;o>
zMvX}yR}1)j^^P135D26f$GTuLXVPI8WqfC>D1+JZ;<6aNnDa#ztc6spV9w3X1~Em;
zE9uHhFQv<uE~lavsb<a~$SNDg2%>5iEMkgbD8`H<%r3)}-_!$U4Mxl5t3E7(J6d3~
zNbj~gwk$gm6B7A~vg=}~MRl*sd(WE;e@#JcfA#xsq(A@5zf8pv=LpQE%ECfgSy@ht
z`d(UCOzUf_sjO%81A+ni2L_%UPd2F*lWC?!dFh`_W0K-Su#S6N$Mj`|7C-sc1d3v7
ztoVYFjfpiTTitnZ?ZY%(Mr5d4&WHZgM|>;9b0$!IIw=Vib>EDR@Fh}GQzg(p4EsT~
zY5@yvfoUG^Zo}>2($bP}8K#ze#Fz-rc(%W{mwx;H`{@_I_(i&Q?bc`^HVKn336t=R
z4$n>T_2%oZ+tneoIq35f;)6NnVSJe0<o%Hdq2}wU`JR*Klj6Dg=>13UJt05-PXE^Y
z@x(ff<EgrRZrX21d}2G!=?C);?!~9XFVKHajmk?G&lFGpJ#vrlr$&-KeqSGt-E%zM
z&ObQn4}SYpo)JcUp{u!kCT*-Qr%Hu$z1aNKwnY=?ZGq2N$~2ly3t7zj&E~{cVI_|s
zGRBvAHSu8Z{Rlr0tRT3ri>JDN{hF=f;kDt-;jdxL78aLMsT`ZFaAAkCq<XlY?%d`K
zt|_}bYqlEc5DPC0A)6K?a~2o8Am@bPp3p9~P*mj$X?^ojdj0j+)AwF^*>2jFm)Fvz
zi!a#i6&t%)iXgnR!UJox4(BZOVt2ptfuCiA*Nl}r;1&XWtx2+256z}xp_mqw4?FMK
zMUSAJ-TCYeXV*KzOOyau^jLf~?g)P%5*Xc>26oC{X<$u(A3`b)+_y`%De;klf%EV?
zgweZ?gwZ{OMhFVULtA-Z%P#fH5!NQ5d?Qa_9Rf50c@MN!!gc(S)<AmCmR+NIf*%MG
zEfnJ(f$~OqaQ{U7alav$w)JG7^7@7&xZ$;d)^oHolnO-fQ0-P(t|#SD4vIC`YgD!a
z!Zprl#xg<hVEs0O8?HjaB@?^e5%{sgU75KY<6RczA}*maN`u7_ir87z2Und~W8obH
zfj!nIU0W2P<gmLi+~kew=JRTVQzs|Ww$!0bC_E6ZBK!oGw6fH1+%BTnR6ECYbrU!@
zFXMFsJN>Z~!*$mYXM6dao!CH^rV%vLU%&}rD~b(a24GIY)qcw^{u=dt-4LE=Jo9{Z
zR<HtuQ}kUgb?Q4lJBfT5AGq|w+NskxRQ)-RNo!rz*seLXR%e{!78C^*o>~|qn-?yp
zwe?FWD_q91jIo6hAS^aHyDfI<V_`KMXuYg9;pUFEDO1&#br}6Nx@Ju$+-960WX0-|
zF<z@3TA>0hv64%`BwvKS2tZL5wVH?N*3IkbQ0WmYGqzS%Hq(U*FIW)XSNXP<>2y+G
z-_RlYmNHRrWo8%)T00&^{Z%F9pG`U8k6ld)2k1ff7vs)KIQs5b{nL&yp$gIOC>{o)
z8Kz6Zi)^l}_mZtmaU*H>aH0{w2xUpLiR;1zuc=W29PVu=+~(U#EsvrC`arp1b%QM8
z!;e2nfBRqmA^pq0{yP1~uRl!JZ`|?PGmAS>aEoA?1ah$^*872Ix~3gx!{hU3SXbzp
z4_)z%f?>Hlo4)_*#q`n(xF}<jD@sqT^-&~f9FvefVxurPY}b7A^X`W@KPd#~C~P<z
zjXw8zR>E)G(?NT2Y1i#w^#U$Ii<Boy38lyCsM)A0);zX(XQ*c=sjwI`j-hxG&LR&$
z7Qsf;F%F|-4aFDk5c57ei?NQ92`@I_H)lnnY@r>vNW{YpN^t5@-QN|i-LM-`&bFhx
zCL3XUJ+}c^b0!mDgU0BFa8&iyTw>ZSxLTpEvJG{$4f=<o1lNk5Yc*~u8~VaIXjsxQ
z9+-!qM}<;YpkghEG7Or8>rLFIwwiVSrfl$u`O3;j_hp=)MzOulzUWKCay7gAgNAB8
zF-=q1c44IcZZ$a9kWDX(R-AE0BYhnkGO(D%O<>YknN4$ra+=EE&UV_g9eOz!ofD=E
z-NvB7Do^D?^TBP-)MRtQk?@5vjj}H{r+oAd3L=!RR*X|O-u0#c!Bx10Wu`a7=QuNL
zUN~5ESw6o?xB?z~o)-Pk>uM98p%|ZJ2N6Lo1#J<p0NmA2DDUAh$cIf7&>irrUOg~g
zTb{)nqx8@KtdW5obF!6v;H>Z*SWx7Lzx!?a&98ppKG?i`(aJ&U4~@hPCl<P#Uxs^O
zLZ`Wx?%%sBI*426g3Agw%3b#_XM*AK7OUP_)f+{7vn9GJ7;z8xm+%IIUfoJZ;E2)~
zg`FB;#o};)`&!%(BmWf5Qk!WHN<rKbGhxhVUWl{PXrt~hPYc%2i{y<m>MxK9VCp}-
zVJJO`)n}lQuldOzzqzmc*yO>OfY$@p$$;{a>2WMhe!h~{*Vod<>Z&a=Uw!3!>B7c_
z+m4kceM6*t=5v?D_;Q~_sitA7(}Oxt0R`?oDV`V3OGhNcm_Gl#Nq97zg&&wSQG89p
zcO!`Qbj0kzRO3<s7bQ$2n65MFSlOjNTWpCr^}(T9tj0o?g^DIW<pF^+!=TB_yDKy~
zg;H!T70NI#3-)KbvbJj9({{Naf;%`unIdLO@7ZKt6C8Al3E#WB^v&}J7Dgk*75BuJ
zc!KYD#e9k37>Gf|l)0yCTv6CcNXgAOEB_(PRF@rz&f*&#GU|gje(;0z<M)1SmoiMa
z`8jrwup1<V04Oing~cuo7XK)OfV+2u2{ue<F<BAx$t3B?2%yD1elpQ|r=%iyz`}wF
z(bj)V+={vgK<h>F2oOhJloEwdhN5o!rxQirdWQ<-8JT=#e1IU21vGJAZ|K9j?>xbt
zm>I)gG?k@|xKTi1%A%$uoc1DK?~L1xAM=kt`6T`Hr~jCK|G~#%RA8zmVG<@`629@_
zIV!&12#c@Ji54C_;T{zI>>tn0d*^I?mNZ1C6Wu?1UmvHUtMVW|xa9hD)IBcF=JiM%
zzx#E@$I2T0{xb1&{%51pJPyOh{~|n=?rbE@R}cy3-GA12?#;)dKF*2fwo8$4-u*8e
z$A7*yv!sm0`_K=%<Xj&FDMD@WK7+8*Iy^9s!`UwEl4lpQQd+PFe{L>|JTmatEOmpo
zg2%JlLzGL(_wJoL>En++v~YTPY1w=v)>rWVoEuUov1>hCxS`NO@H`L?Ca%rO3T1Bv
z*TKQA1$X6A(Q6FOZNd#2Zrs>e{r&I1=`)nxc=HG8l~-R;o(N>~UMt|bDqOK4SWAiX
zM*1oPrI#%^da-jIz7L^FW-2pUL!i7F3HH=xb|xcOClj8-Gor9T=!*afAwB~B+$?7d
z#jbhc$#F-ZV;8=DB20T=&4iE*fgj%?c(USXfLkjz2$Cj*Qy~xvoH#$gH<LHQe$s~U
z@fbI2eY_T6U8VM@exZ~jfD`cEJ;Ri78A40SArIP$LWX+;g!JQQ1h7dwGu9tNTwZAn
zr$5$ve2-y8xoVi2toOy=M=O*EJL(MsuE~=K9Qfv$zE4zl^$|Fwg@f*ad1{BAq3i_S
ze#dkfUOp6+tQoNa!?mQ{5h|T674@8RZ4i#qe^FQZoqj~2f%~#Zo56)to-=K5+u2S{
z@%t=KGFtnPXSnf;G+D)fSM)*jTdX?}zHe{cPOVne<)NfuOo3ZiKZIgo40j^+%LxiS
z!L_aN&`8&=eV{RL-y1>;8k3;Q@QAaxnzj}~xwWG*4)=G2&pj)hvA)1+6PJRx1Vz9a
zRx2oa+a5!V0W2X`S2yf33`HMGAtK`ttE^Tttb14!q3l2*Qk<`-9-Ox`W$Pb=_pI>|
z%p>?jpl?@d>Mq9JSSd7iU5EnR7Bj5<)qcjEm3;<*)^Wl|*SE`AW0|zDv}h&E%Id{b
zDz60G=x^4VN=uuKBNV7S5O5;|Mo^BD#)5X?$(-7SE3>It`U|TTjSrN9pdiXu_1V!d
zEG?)Cw2gV+n=>`$=WXdDoS4yikUl!Wo$H~M`D3dd6fTwWVp<UG0pD7FeyH_yJ}s&K
zc6EtcPb+(MDR^N8^^5=bP5S$P{Fij|_Fn3A!6DHGwUMzjrsW8sSrv!S8v3J8;qmDZ
zX-C(<g8*23Ur|6dZ#1oS?#xWOa%nUD@XZ&}#SPr~g;Ir$0mmnd7se3diMo&x@Cnzp
z2)0p>777(x@1k5{ejuGKz)&){Ova|h*9@`)Xg^0^C<-r7h(L>5^=jI>cgGtXP;%l%
zjSUwZc|lp=7JZE%A8Rd?+1%@N8p7>b&|sBKi@~Qj&k9SjIA0FeiAyW1BL%g{#i(9@
zk+})wN2}TJ_&`3vnM{l!Z_HL?ZIAUCa~Ja>3X_9_eY>NxJd3$(KF+uYc65nqOy`JX
zg@T6&?V+w#_zLzY?!cYA>I?j^Aj5kS`I|PW&FywQ?QU<mUA%+R30I-Gf24kFLeAxJ
zPdJ_Woqdgmj^G15;k>7Yl~q`@f-~Oot9-JVoa<~Cg`TUh=t2?G5*@kYIXA0xC~0DI
z0*W<_ZH-Bk8`GI7!IJm$K}R*VE#n)_Y2Zw^T@~#)5DW$?OJiL$V`^G35{*PDgNxGW
z2W1UiXX697BVGh~klhNR)KR-)P8j1o9mr3{<WTbuR)nH)&AP`zM(Lw2j4x;@avZ%Q
z8ipmJT`bOVUgA{hiGD&GQ1sFk>VlFB7~-0dxfmYA${Uv~xK3kLHl1!?yP9rX`!wZ*
zubY=dmqwey$pgh&&X%}YmCqcnb`6`}OvE_gTsHcXK-;mBhAtAYC<iyO{1bpImz&cV
zLoqn$)~#%3?9OxMT26BT3Pa&0ePgS!EO?6x=QwAQxsJ^uY+}IzoVfyvdD;Um!r$0c
zulngk?;j6C!4D2w>7_8H8@wZorU&Q^w-u=HobZqqkw*2<*ueUjwD4N=J2JP$#l`f-
z>)%g5di#gz<rl9A=eETwayC9RS2*PSP!uEArF>`Va+xa#<C4a~bsQxeU!T5rHazJ*
z;c3q%;pu@{bfWm0gzrqig#yg(v=%@xAxsWTitLtzielD^uWru#GO-nHLE!{L0Fwkm
zj)fL!tz3YK)+E%|L@R3OeC)DB6AZg8ne<^uQG8L3Nuj<WStg3QW^!i{!!8WcvcrP3
zcEg~An7>m7XFb8Nqxh<cp}`!Q2>>F^8YU)ORYOFV2_SaAp{VigZ)|{>yYk|T>794q
zOG}!JnEb;%j3yWu2EmyA0bWGr9CjWVex9IWGGe0eBuN*Lub7&#lF5oN`p63dlv9ci
zzIBg&;h7a*qi1@TNME@x9UfjgZ=PGxrTZ$Iq9~A5@KIE5Jc03)Kc4xKAJa~*SOCYu
zhTkal5W2t+2@0{;6VAd+5zH_v?Ps5Un*Qs5|8MCxzxr)zw%AoW36n4hlklwwe-y=6
zEQHUG=RVWi#QpQ)=`@eLb0+HJv3uvor|0v;_=NJmNFBdSe59PS@w{@sOkL>E^X@+q
zo#b<3Ts;2n7muFsW5$1CJny|P5-H^IT)$*Ir<}7<ACKL8ES|~>wh8`~>#+a<LDtO7
zNm^T5wtE^jI>8U)1_^}}LKK8FD1r#^@d!4(i&vS$tCOFDzOih8KSTJ98)y!BUtL{G
zue|bFoKaFLTgXSgyvw;>oc|l{2tydonM=4{TU=O7t#&KjzrSTZbZvdjeBajg{j|Tg
z<MqPE=Ed~-8$V3T2%1$s!u&?FW_}G}B)gf9Plg@`-snU%^@kD&B?khwxqO`2a;$c-
z+a9-y<YWF{FK|{(Q+d{Fhv~3-V1XW1B(&NJI<-TrtGP=yjb(EvGX(5mH68BUOnW=`
zQm<1L+?wKXfe*Yc0w(fD88xSffJ}v{eMcyZw7B3rAZ6lfU<*G=?H>&wxbjolIID{^
zME&MH9&=l-0SlLB@q})kaK2eB;o1^D+?GPZg~)?6L;{%*#&sZo^SCzb3QmIOYzQgG
z?%}bN2vM1W3O^S9gq6xjxp{;e%~AV&U%?8XasKexEem0RKW8itdTlFd$JRK&J(OIr
zMzLiB5x64wJkfoG_g>3n6>+o0*<q|d)fRAKCWPhSoj1H$P>Dc!-qsLULa6d~^H<lp
zP|I5jCvj_ul_3^ic8~QynQ<CM$5wP9BuD7Vxjf-wk1@!mJS@62{?u;>Gaq)X)+B?&
z)T-XI0Qulx+mEo2^qFJFgT#3`je6Z{TGG2W1mk118Sf~-1qBcSQ1EtlcPrhval=X~
zd)iRhz#YM^71*rx1c$N9FqF@n>&3d9wKwY^EE%wPvLIV&hQnTJs()B3EiJC3%EFRi
zi~y6h8-i67U##_Pp`vohr(9XG;1&Tqf?=;O1@pkqP!iz2tT<9&GnQFz;_6b6ve2LL
z#aM%Gph#dn-suUS?d9cIeT8CNb<=MwO6imNVku?jO1hp`WX+Geuz~7{V4pFFqOt`z
zQ*IswV}{|#Pr*kpp#KE_=wHG82!-gO;3Qfgc+>WR>R4G^P1(66;rEQ_NmVdDNh>QG
z3Ab=zX$GD%w#a{&e*4?^(?9+5zp9T=G6DnPtYF}EE-4|jV1RK}m!tdVg`mZwbiN9D
zKf3Za@TCQ%VlMsFU%r?A@-N;_OAA;(4c%7oh&~0kco&%fu!bh1=wYp1Z#GiDH&7cx
zCV+wx>qh$5R>3NP^SI(HxH$8Q(rr1*1_)@q@Pl!T<=$cSF!eOHaU)9}p=?T?FHo%6
zt)OloCLn?X;3f+8d8{=JSJg%J#N{D%vSp+CbINjtL8?~w)ArV#w7q>#^T%##;o(J<
zqK(Lk*4H=E#Y>mfS1YPR!L$N+pp43k9#j^WEFb8qzxOn+VbO$j8k;YeM}Qe_^dM)p
zqS4EEfr&91<v=5h<Bb17*E9u%U~DE1+C`v`P`=>?@8->$dPY0v-AB0o1UHMgI~9%c
z`Cj1DvF0qna|Xq(@Cdh}DopLbY6?Z8%Adwks;qtxyo7sJI(dE(yp?zLu$p#um|Ht4
zGv;=hLg%3ff?jgu#Y~p*ftA>lH@G89LxF<@;nXzlFsrtlQx>XcQDs6SQH&vfLm`F2
zg*pHq&Cd`b0Swwsxtw9?%>?w_k;b&>6|MlMGI_nH`BwCV^FvWww%Z-m?N~4j*R&^s
z_lch4uCd>(TbWDKJxAzWI-U84`4Ov3&P1HaidWFMa}b`<m-Ge7-87A&{Y137ulLnI
z$0x?6J<)agvQ$RVI;Sy`)1%lJ17BdvGL=W0r%teP3+pQsUsIkRsS1kZw(tf8JjaYI
z=w2pqMi`3Zg@ww<O(m|NIVY{IJa9>i@>eiJ4u=w6G-gKgX;!#sOKp`0jpyhR&N5?u
zIuieZ<uIF6iYNe4M%uL{?FMF=lj$3^16sj&13LiT_>UWmOYqY`HNNs4>Ad+wv=Y3r
zbr$dzPO#YnxmnPA3=~9ng@YU;0zJ96wUu^u_mnqsxS)MxPl427JWgG>3VTquQSh_T
z_xSpB7~g-=^Km$v?rCYis&HPJ&l>;K0;71M_?m?8RA6z#fr~Tje1ITpVB)d_!p216
z#g!*_4R#Sum>f?eB6e5~k%q|?rU6rMA6zT)1|tO{68_n7cY$wZc47rX;RUUJ!(cPH
zp$}m)g<<FZKofwKBqElig~^9xFv(IE-m$s8782}s?KH*w!t^PPEs<0U$U9k>v2fkk
z-OA!Y55A#+9iA_L@0IlSJMW~`)m2Zlv1nw{3c^aenPOr>qOm|l&-f+{8Tk(BSls$T
zlf2&1O&B5pU$BGujC{P4gT<E^ExR{V8Y{d;e`N9ytgz7WO?iC7Krr!Hk;ZNAjq3Zb
zPSFPE!|nKbEDa?YH*^E$dQJ?eX77ZH1aOpvl^H;#W+B6Z?(pz1{rkWFBK=?g*MAcO
zU2{&8FbR_|3Ez_N+!SALq4*MY^`GeNn2w)~`UtxHMWf32yhu3r-A5wy|GfK8i6g$`
z+-IjF<$qn#Wjq0{<G=In{ULu8I4<{`_(j5b@Bbl@K)t@s7-i7@=NI)61%FN;?HC8)
zh5W7{<KOV(Q^#q2Z6&?)_7Btg`l<yj2%&4$s;$})cB2qLn2qua&j|4ID8F!nMs8!}
zm<2J)H_lEOq{D*)@!*-Xxp^_ItZi7(WjERRdGmYhJm%~o&gn(4hX6*m&5yAQo89^d
zLq7iKBj>kyVKbGA1=kIs<?`}Mdhw-K66fo5I-RtA|8BZ=?NhtIDi-HcWg*Uk^4Tg~
zT>7y-$XYmtYdhQ#aWMY;T-p3;wYr<`+__=*j3_#AHHB~t0Vn4Y9qjL<L*2)61x1X7
zzG@%vM6i$Lly}M<Ypv6-ru%n3PTTiBOVvZpG}^Z?5k(I|pJ5Nd8}B1@N63c)3~Moj
z;mRYSSQKxna-uBt9`Lo0>U3az5{du;PVZ{nf)YgmB^Lrm&NXwl>p6-~eehIids^!p
z3N~%^??Ar?2{%ZjbvoCOv_#70_c)B|h@uSsn`cx3hTOu<Y4i9N<U$AaoV6ePc3bf4
z_gX144gTph;gj(a>jIZUKsnR48z`)oP_p%L6L{1$&MHgAP+Enh7p}kp53xM(J^i5o
zvqDF3_QpPh?eO!4!HL&#uD>$0i%Z={fSI4Gs5%w>#_joum6yAUQ^&o;87=vIQE){m
zqcvs*W#YK*e2=cxB$UbsusM%sacLu!Dx3v{K)b4S#DUhG0}Dyn9UuMH6Fll^fBUm^
z@79NE|DozToi1N~!2(k@74|x<RM)#ht#MKKa_&`SX+_T$Rd1AWYOB`12q0PSA|SqT
z^G5pY>c`4^kjlyjk1Hr|@7=wrG4ok!D^Je1E31B3$^cu|*L0^XS`h9dtVXa)j|fM1
zceedLeYd)P!EFnLprVxl2!gftxAJM|{z2)DfEr<~$D-2Mx<h$Hq%X5+zQB1#CDmOx
z+Qtpotd(Wp7w0k^Glsl2=UkaNjp?9K9pN&HZoIY_ca&#X@y&#S0y;Bt-#Imf#h-A=
z?#PJhFx|67(Tr%qG;qLbEtB%aMc<=54Q>tb(b-|`rtl<drQbZR|J1f=6c1A<1`6sg
z)ko=ZWyiR)#f-8D{1YDP>)(I=Z|T4P{XePhoP8x2>kk^B+88FL-`wK{gN#mjLhv{u
zI7z(6-#A2`8WS2rm2xiq`A>eBzW@D~Q=vE~T3EKS5qykt7<XEK>m}qgR!r)eH9PJt
z5oY&=TUc#j;cItuf;03T7l^nwq-@edf4ds<C}P8+3`j<v%mFRI5qzBCoI2qNH3i=}
z!wKt4^0r(^;}`{nK_K9jDixpi1a9Fz6N@xk0x3Vvh`avT@6*=38);|zZn}B%({$$+
zw0%c39BZ<nk1=5U6iW-XDnvO74uexDF;L3kevrDXt!@Yh3kD6!FcfP`izuUGgEjIi
za2*<iCnMxfDBfuc^+oPM|AhSM#MYW@;6UjGT(NSq3sAj>)f(~$lw3R8JE~JDRn%62
z!1IjpS>p`LHWU>ohd4J8#q88HxLXq33Rc`q&0wX70!LL*$(#*|+q`b;&{lKIODI=K
zZ#Sa)jWsFrAIdP?4`ydj=!9Yv+_k0A6xL5+5ltVX1n&1x#&=S+wx9O)w%p&G8IL8F
zH+(2e<5HG7O=ArNeNg$LlgO8d-jw6HfT%f8?ZuTJb6hA~kQH`pMTEkMGdnrQs!&==
zS-}KCz`V((j(Tm!XA&~*&2xw@Fn0T@7qlPRFr)g!xnAlA<d!NTo68#Sf>waTed9Sa
z4S27wt)>eX!qSWMC@)K;71KTN506P4*MgNF3PG$Dy(xjYV_x(@@a{Hi;tRmVJ;7jT
zC1!DP&3)KyYJ92h@qPu3gkFUkVK(v5k?I2>jpe6?(dZd-k&umzJ8TTXa*wm)*i>?S
z)KMFnK87Z1E5T5R1y6x2Q+TDcQ=(h&YJi&D>2x;9t2AF*aW4EDJ#_%T;R3L-#>xwJ
z1ipoQj}2?A__3-k_!<Qs=UiqqSFNqCdUMFlTQ`)i@Ka?mmQe6{U?_ld9`k%~YjmHw
zJe)hmY2*8=3a9fQ$1fVr%VQkBOgJyUFB`v104pYnuSxih1^iWunbagX1(PH~Z)Q!j
zib=mW=#QAy6HTh@%+Um>fyN|J67$01j)SfHVqRbxVi#V{E?4YsPm>IjiA{4AAcibh
zY=xvrdPEsDL2EMAqJTw$YQhxZ0S&QcBH<3>o9kIkh`4NMHCy&Q-fL^ZQ62DQjq(h}
zJ%|BJs~9%@u>in5<Ddsq5R>^gG0cm}eeI1m(%V0LJFRH)LD)clpj4vIm|xX{zV?Ei
z8OJbIfBI8Dqboeo#d|O>!K5QtL5N}HmcC&;Z0V)Q<Yx;!r{Vd7OcZs=`-Xe`Hq)bL
zsKKnP@~vQ$n3SDlG$Gs7;V5|L2ze?i_oJDHhk1|t)(TN}$G}-CF=gQ1hKPdA3uECd
z`1!jJf0ustvwuz>e|Xi*(<Ds7Buv7$C_FdC*AL!&^O-EZ#_^Gb4)p%iv#&LtF83Lt
z>-h!W`C8)htOFf--u*vfj4GY~?5m2;ql|OoGnRWgguvC@l6fP9n=2~|>D_mKm@Z!2
zw1A5<iLQR~ak_Koj`(d9RdZ>+5IgbN2!t{VUJe&y2q?TUNaY|HV+TBL3$b*e9IUoD
zH+F12!46~&i^qxw_e7le5(3p}3u`!w15Xfaio@DtYildL|J(OdwR)K5a#^nxEZjai
zPHSr$X>H@8nlfWW*v`&<rTul{;PllMlz5Bc-EgzhO>Ob@d}~06f6s`Q3xORLDF_-7
z=i?gg{@vSdJLhq5W(-znC<5V)u{^*{50-1V&+~5nBluh32_J|%uC5hhoIiH^=4Z-#
zI$gSWNibM6Z;3^~coQHn3wVTJi#|cvXE$nEo8WpZ&X+>KBN&-C4MFO|fQ6OTAn@6O
zKZ+FUfxAib#|WsV)INmJdLLyJb@ke5iZvJt2;I@zhE^CQ+?-H~>O()k8DrZhI;AdE
zCus0^y$4LKWDst_<FoUgd=;V7NzZMObv&^3maPqW-~2jtQh8njpuj^2p0ToHu2{4H
z*j5Ox59<e%7ATew;)i#UF`N+GqpU@M4G!U63WWj+7S@#frrxASnTxg1khPv*h|&e2
z{jA;{yM-JMYU#%HPtxH&JJMNOvTi|8jF34j7ei@9q*lO-h~+}gW5;LxP!IKYzuORg
z-1C{j2=*72Q3{W&Hckeq->#*7;SX-~P+B1r<Xoe2X~9MCdw*}A#-_q(V;(|ttY82S
zLU62M7~5F7Ah2ajFI39ut+(DvmoL947|nYeUAy*4`ps{DsqxYBxhtHF2Tmf?x3F3L
zn41lEZn*11iOP6oyxqM1nU&4+f-wS51g)&au%66X@UQWzzCh@VVi;u<IEvy0eDQib
zqqRFYp|)@YL0Fcu*2<=#`U49!lnGe=%;R1W95SsD{TXncH*FN8g?DU>UC{WPFLDM{
z-in+`1%;LJQvafi&FI}Jl(kqP^y+D71uXYa_AqaUGAsn%;0buCF@q~ME64aoDT{Sz
z!vb`lM=sikTmafayXYI*aXf6KdTlQ?1ut+0y5-HYN~g?I?O;3o?4SQN{r3IeYwWTf
z)wopkj78LyKYr+AyuLmg{CFIlm}~yxo?p>L+#xVNkJ4OzD!uyZ<+QxW27#{L)tXoJ
z!m<mOgW2qi@T%`K9<i$F^#hl{t$E=?sZ>%{oOeA;J<%P`#Oks!LLCAepyPP>VLm{K
z!~K@>snzQvD;?bI_7dm5ZEfGNB`vVwJU6Vt7{kESYj{<iqcNCYN*V{Y08>BYP)g=!
zRsSQszo&Tu3sH@kz3ueD`@c>f{PtJr`n6Bf&i1wySkRl*)pgP2D`{zY*-GGWxjEyy
zV5!Vp<#U*PukQ^UD7zOb71OGY#tQB*=S4eL*Va?9h_Vpe7+E1|Jb?FX-ewFSz&D+P
zHmOaxHiRBvUA(t{Ao@^L`-BUE3l`4E6u=WUfur=Ht&D%h3Y$%^KJK)!zQSc_Npvln
zI#_a2zVZ)TWv-y(L*53>#5$Mx1YI2G{?!h5(zfPIHlVn_bK%0XFfTZlI7<;F#j)lr
z+ySC22@9`5Uo>5PVws!DVIIK}tlz1+Zom-5V69d+zA$fMeT#C3Il(j={HKnLEzNVr
zRpkvm4Y?g-5Tqq7^NPlmegb!&XNA&4_fVu~a-s`Zwb`0kaK_zVRyk<SKw;%GIm@C0
zg7N;&wrL&mA7~kKK@=qzsV=x1ZZ!^6hoRsame|l_6lw8}a9(s^MznBwbuks;Q<Trz
z`h|4i(u-+%WnFv<iapH>dLNpP#Uo<`WhzSHY2n6kubDpk^mpl#Pu>?EaLyu@u*(Un
zP&(*H?X1@i1?OYWV>4=>Wuj^ua!BSZ>W>n2sQzm=)h0CoT<dm(SG#Pq>08l;;Q#<0
zD$l`D&$JTjvfw8I{wIPH01k^(0%I=59p{T-sqA@P_rL+>c=s1kcy0cK`y-wJ+(qdb
zbC3Y6x&ae#7WwV5#u_-zrYDqPC<lw>LVE4{uNt15?`m1G%NVU;_|rG{2<AN$AJk)X
zKj_87>#rl6F7sL9Hz}M~|0l(>0gRd`z9!+j7c^=0+g&pZUbw)Ri@@RnW58lXbES6x
ziSWTd9EmVu6$X<4)5FeAm~;p$XM4q=ks{9`U@*N*5KN1ZGw(n%fsR24k;ClTJBo;U
z#$*TKW>yFxSr9|&MC5&#Xbw?5*gMb!(y*B{?NCMiUJP({D+U2)!pG#F18)an266NF
zO$<zcFMaQO>Fu}Q_Rc355@5q_J{HZeHMHFA7CauDqeXLpH(&nnPe7$_CL(T;o+xHQ
z?=XQoGX<`JIm~w`!pMYqR|!UGxfAcuXZoeQe2?^8@q;pXhfAmPrEWjO2j1hhKm87q
z^OP7-{Z@=goJA=MFx7oDkp^?dWKRF<o?eBKKoIla|L*V7&;IG3wIFCy+$2oGBuv7$
zDg03sU(n_!#hAO#zxQ0C>U!?|Cq?EF-uYuk>T=%wuOpsnKjE?Ge}qUlU9YbrevvZt
z;q+fcI4}S6qCOnCHV*|)#BW_&`JmP6=EiDTsLY9P#%=Ra%FWGMu!(D`W~=$&93M8K
z!Pg^nLMVkG+d?Ao?g&iz-rwIz)!KoT4k&yOf?J3IKc|Q=45gQMP>VN5NXqW*z1@3>
zGl}>OZ;ycI&fR<I`i<+ZCu;-FH^G7ntBl2^l~kB7IUNeKt*v|M(@(ED&&!vu==pL&
zIl<;7_;f7T5SWR{^cuwLg+bpg-F%kV-cC9^*iVgW&3P`0Cth7&OO?gtR46anqNY&5
zU8An`+XAWS@MyvMh8@?qf9rOeiF1??{2~<1<gmJ$Pqq@Wk_)Aa>Hs_xIBz6&Zqp+~
z#0LmD>vn;J!Un+?@Qy$ldJ?4<?;+SiNUuw`i}eq#mDnAwbfQHT*xQ|=(t!_tU-xw{
z&a;|RTTpoLh`tKf^j1bu4zD~|5sV}KiE^Z8zTNe6J=is#arr)fiMqgRvkptM7IJZ}
zkI#BhpQ11Y7AUp`J(MFY!I9ndD3%Zk&)MpLGg;#Nr;Zgu)G^j^D3~7B%L)#`Y+5)n
zt+x4^Kc%%5hSojog2$a9?x=84=S^;UhJcq{`_l-rmH%*Xn3{FmqSdVUU07JQE6!@|
z&_YJ=3pa0reoe`n9!t>?)P;ajOfVTKD6x>kN(MJ)3kxg4vCTA-Thz0W@)IGl>cd#V
ziV30N-rjxV7I=-|5UYc^LfLE3p5VnU{x;)6b;Lr9^S-JFJA!|>Q3NLuMsq&c=7r0`
z-zD|F@JnL>SW;c`pPj`jN;sf?K?%&}&snX@$I2Ye0LvFbXkS0vPuutISip(Z1Ngqb
zw=Fny?F!AFGE`6CJTt=?Uq|Zm16xlbOh-XrxBB3e*IywNWywZ>ULYf2jAGSNDq)F+
z8!x33&LX%)dBobxn-7Jn&?NdJlgT?RiZT=u2%R}g30Glpu3bMhYkL-OqYMDYr+xNb
z*VZ8P4dVuvZ3%aXef5{{6ucRaqiOXK?}bvLWNSgw0pTY&isEo;3W0t<^-yH)tE{7`
zw6eOP_AD7ka6R@f|MO?*zx~}$^}MRSLbxC6Pta$~-|)?i<M`s=_fJQHKlvSB58mTz
z{LJx$G)JjYnoXPQi-OltYBvvj1`$GYZ}#O4)rB^Lhhf!&n^MjrYP+BCzJn6Siel0T
zKj|-IKhRn7QDG`F6q!~)3y-{JS6)ZZ%67xnv?#3xy}n>sOe?1A1=9tNoj^&M6K=9(
zpAznYD{b}V&h~BL;hy@CbJttW18Ydy))cPYzI9z=qUCXmn^0VAE-i1Q_01Rby<!^Q
zP@CyD=AYf2`{_`505{I5FRCt(Nt9Dq#WI%a)xC7%+ErTw;|g(kWmRJp3+CiGh`EsF
zAuHfF^A@;IMA5)GV~is<PJsLTUR+!gEt=C<40^~c4g64yW0A!e<9Dltd(%VZ+w=TK
zo3L1(;+!s30Q|?AD(Eyy=(g~=Ay{*6mS{v%V_NxSHQz7}I?bxuyyZ;@D84v*)8k8h
zMIAFa`WDMGjpbP^o3MZsjS{T}$59A@-ySFGBV6Vpp8+0NVO3PWIn&cpnlCJhc9aDJ
zQ9`wo@j+IMp&9kpl;{j@7@0o+`hziKQp%VU-sqms<rG~2#wZo3Yb1hjs2+L;MGoKe
z1@p6AjApPf)0h)|Lata|Sk-;O4!jngQUAVhpYss+cDLNua~juCAAN})Fi$)<Bl84{
z?XbSW<?Z6~x_E*W&C_Mk9??jp&14Ir?Pc9V`J^Z~Olyug8Ftd{{cGui-~2q?x_MQ!
z<2d2s5Gy~s4OZFB23GBLy_c~i+jOSlx-fr4eK|G;%dD6qu*AiZ3<dr`v>tpc6&FMU
z$M~VKE4X2q$Xw4DGB=3pWq3XTG?f2J%5VA>ygh=pqRc{3$$W@KEjGV;7m({2R%W<*
zwam}&gmoOO2Z~;-4zYZLHt=V5L;<oLG%krxET>mr{a*Uf+wY{;^vzhfas8HP4RzKF
zB%(Z(;^6l*f|xrLAJD+j{c$|+-si>B?@Yqe!)dq?VSgryuSxh$28d<9!(oe>82H0v
zS~P^6jxYp1b4$d#FGlAC1<28f7BVmm5OvIvwE$8Yh&Kunc59+!!c_*79kdT-3Ze-i
z)sQnew4z0Mz|gQ0jCC35*+F2XhSFPYqI|J9;{w0EOr-26WhY%v3^D}YIg?O<36o$&
zi~>xcnRH#-e|4XU$dA5Jft^*azwu^z_uY5J4B(zd%!~SkZ`$gKk#-3Vf~?{Q9)hDj
zM)1(JIu}s+=?G@5gAG5JB+~ghV9G=dgJh)`&y_b>I*?{m4#B-k5u{STJm{0rvj^9%
z(}>ykf4uk59U=*15#e(#{3pBw=U~uB<vOhrqxg!Qd9j-XJmj$UpZ@e8(m(wDKc>5P
zx5;r5CSejL;oBCTo8s$7Z~fqb;_GQq9}nsGR~=bAop=As#1X&oM~{D`dg$Y^dw+~b
zIPd=Hs1H*8!7;KpljjqH-#C-@Y4NKFQOi+?3JBj37^n>KUI^9@gw5w?Qlohw-sZM`
zpQPpG<-`tccD%Do8orSC*kMj&S3ZJY@`0zrJv8pyu)IRxhsA}@;}x$A4+YPS&<;Ku
zejBzLt1Nbwvo>IdGvy%kM4;7ZG?mwG;vA2qrG<p`83GY_e_V@|D~l|h1zQwcw1uU+
zVdxeZy1S&6Ziengx>LGCrMr=C7`j8cyOBm3$$>lHz4t$y=bW?lUh7?ZE9!Pkn-^O(
zrvW3@uLf=YbX|}A3T%EG+MkI14{QQvD0O9{#n128TVy3Z$DqkQN^F!D3F!YZAY8=S
z6_unt4WGdY#sr{%VrSWUV?wMuTO^#PS4}}ach4i*npD3gCZ{q+26<_P|JD;p1@}!^
zpW=Pr07)(K25O8=vj2JmOIeWj>%sNmXVokqCM&f_N(dAS*~S0KS)DJ65A@v=iC5^H
z=D&;^-y=DQTPuAV^w_^^@^)Ly^C0eL?hw@s#Q99wwwHS8jV&LHjNIm9n#^!kjBcBf
zcQw#4QRWhy`lXE}1~r*ot|bJ{5%{!1hbiI@AY?mc@h(;d?j_u_@UhG}wPc$^1ok$|
zQ<qs{>W62Xo{bI)+#4HUgY&Slx+!FE<x|?8e9!_`$@3PSJoC-3-Q0^kcHa_iD2rX&
zY|g6Mo*D>P9Rdvuw?7Zv{2+jvJT$fA71Vds>nV3@@_RL`ywD>OsJ7?R4GB8G8YyJe
z-H9^!-n3J3#hj<`M7EF6YR=o5<z|STKSd~CVZL?n_j2RjN{Iz{cHhp*`VFa0LL$x;
zO@V_29(%@>4ZCLbn}XA)wIa>(WF<IERkB~~**bq`VD0u>oR6eW`mmfS&yetkkWE%O
zbXf=x>}OHpA$HBDH&Q!-{R~~P*d%>iF$8>Nhyg{_$Xez_)TkcGi92#T6W?8e{}pz+
zPjh+0(;lOswqXZ`QjQYAZ_cX!JOad^mByV|6`4N6PMPk8lX50|01*VU$Vkkc*xRu%
z&DvJ&`KuJqADnffc(LRSPXYTj^!GW0s3OO>F(^|=@Zxpq1(@LDs^xz&%kA1iset8s
zCbr?yxoT3K`S+*ym-V~1mmZxQT_@N|qDG0AXahfsHi!#*4f1!IkXi&dW5H;#M+!Pv
z(yoZ5bX@}=lv-5zFqikfiZ#FHDq{}1`}_#2REpROUn|&4BoygIG1WdSsY&iclHjrB
zU!N}D&p3>oMdN%u(oXywY*50!!tI3ZO2Ux+0N+h;cuO+<#ku??oB28FfsiomYxQ^p
z+OdmVz>j%>e3W~=$t7!V__g(Q<-CCB7k9|oqUuUH`^%{?pXxm29jiII{XU_Qzo4Pd
zuHQ++UU)2faY+hBVt~P(QICsKlkYDv(lyCwUyfGOHrN`s_;X}?^TRPeFUPX1)|$4&
z3`%ESR17;?Ota7^OW;_MUbTKX;#U1GtRPMd<Bf$F{1`}%CiI%vD#S(hL!U3kfiVbV
zMHVc08H@*vi$#bx?}GdL&8k3kU@yi;KnX>>pr=q5lmKfy)o8~3v;^^MK`#+?xJYoD
zlZIPhv$4y?AOU)XweAZJ)yS!-W|cBC1r0NBiEf<jp5m<q|3LK(_P{yXm?R)nm+)W#
z$lexTC|L_|<ByBugE9x;y=+h*e+1XulO~5}5TWW%l=xJvt5&Z+=vO=lXKkoxEjW_#
zYIV@42|czh;n>XhuVrrJ@CU46Gvtzl0U>TmQl@?_TDQ^)TCs2SF(te;KQbN$5NK`z
zu5!bG2Aauhd4(LSY7?DzLDr^VchhSb+I&JstjpP7ru^g%+h6hw1H_?}zPa>FrSX1R
z{5()(!LOnOva;Yp)BzZLFW_{@g~|7rX2UBQ&$F{OMQ6N#LQLYx#YW!?3^HBlGl!y}
zamYVa+~h&##ychxpfY>8<zJ9-2YStqf?~9i!=3^JunlCeRzLI~A|7%LgEvbnswc|t
z(g<4fP2DqG;yq#foD*FTYhc_CKicqq$=E&&X9ixT>ZHPN2OsYt6_zO_YCYquUHS)?
z-C&TEn~5-ppzptg^_Dz}c8ft?JDs1OPP~-=cVxT|m94DE=KuSsULfM5cBKN6%}RMf
zM89zM`M_-AxrAvmV*J!rQ8c|z`Yo@>fV9veY<9X{u1ip4Zq;4;#m6YjWom*{@pl{p
zwWXvOjX$<HzA-jhH+4l42EEy(dU(_sB039REJF|{yCKe1U~Af`*^2$ki!GIG?8}sn
z?FcL<kxEbHgFp6F6gKmB`ajNR*Gg!XoZ;e+yNPptmz`(ffLID5@8q8x7%dBP{BhK1
zUkub{V64<l8f@N>|Cv{^Ix<<_Rd1P^TGFLuq=SmrCJS!phM9er^FOltt{9dirXQP4
zP9k}IQ3$%qGMY7{mw>n%+qxcQK8Uygold7ZI8?2|UNN02KA-MBn}*Ah%l)jJ^W!XW
z@~ksGxvv#}U%oSb8xQ)jT)X|R<NtTJ{a2G}{t_?KTfzjC3PDD%*&=UAkBSFkWsh>@
zvCscDUZw^Avq@POJw-(V{~-->KblV*zLon0nfR-VJ$s*TJ#pe2gze-HKTsZ?8McpP
z6NZh`ce35Fn(Px_bmENCG)aF6R!!Wfd+yEoxk))i$bv5mhQC~K(C(7jBFhxRF4Y2P
ztCuu5h`sJOZ>MS@-#ZgDL-DayEZ>~wh1<@vy`Gm&G0-NR*+>17`SGz(Rrtgy9jDb2
z0`9IyPRwPJzpxwmaCSY;J)G~M)AbgN*|m~7s`wN*P~(Phc#A#?!G~1K4!}05PDLHz
z#g5h-E2P!@3M7r*d;&JsarMhfphG)=otc(kkzo<A!iuwL2cM+HgdMt**YxJyn@c5j
zRS#du;8U22DqV4Y0IIO{TU1rasvf`g{Y4(fPoWhfsr)!4g;J2r<T<Aj$_Beu#}?(G
zgv?kJiBm@%m8r$Onj@AH0hCIp7egr^?^u_@mV3tLdlpS045|9gJBi+}9p+PPPD^$v
z-3~7MKodQJB!(<Lq#C<*+%1kxWcN4aw|&=8^C^!^u3L2DNCPm_%GTKG;}x1ts9T6J
zUl|`W`8}4Y`LgYblDIJL3?|}8h|>sUT#yq=Rn-Y2KcY_7;O$gaGcHHZE3uL&HSyTs
z-tsFGl1KGCng-0rZR)EzW_!iyvsR_agU9za%0;aCEyLw*Z7nTn4-V6k4b=|9w2&^U
z8^##mN8opGty6qhg^XP~h@oVY`~cGwBoSNYZRvRZs_;5>-0tMD*#2Fgo=_PANpBW|
zUMGX^<mKCk$zGN&w5+U#n&hg(&Falt3+;|O_5JDj#)3U%Y6Elza$hB9z-B~Hs!?&m
z(m8Qo1}|9_BmtvaZBtVPe(|ri*4a(KL59K^X%^?3R_&I_@ilV%4jH#9-dVjwPjy4)
z7}kSDfq-m*WFWHqZ#=SwqqOgHN&l2iDBKYbw5t?wT0(lN{w|iC>hC9vwWz0D7uDN}
z2%OSQiJ*4os0YBeil8EdbpuDS;jJC|h6Q9sWA`$f%fm<Q95n*82e{Op9yclkt7M6=
z3_a`(_p$9sSUx8%cvk0B?_|uT8!3k0#Eh|^<Z9_pv$B|cZ94RgARWrMV}E~EIk|kh
zC_9y^vMui!&Rv7w^tjULrFei`Ain#m?-(9`2pGz`4w=8ra>-BE+U|5dkbDk$-F_G8
zx`q%gCW1r$5XG!GbhftU{gcUa$%}w81M@lg?JX5OJ!J!CmHayVw42NbVm^sZomd$~
z#SdpISXhQi&aA2aGLwly(_%rvS)tV^*jC*!xerOfc7%NM>DVluqX?Qw+uY1?cdoN9
zv1knqH>ZUf6LfzM5&!MQL?C-C!So2Kt}&L+A~c?R7;4J*DQ)W2RV6^%?(n4?;>rR7
zQd2eAtYgCOwHg`yWuj_Z%-iDn9f8?>u2``u&bj;i&cU&k+FjI{RYI;oP}@&+>RIN<
z%7?yEBxm9SlG~3NS(w0}<k2hJr#63Jx8j4E&{So%Ku}y;A&t<t9~5BDDF>8Id!MRr
ze52yS+VTPMg2NOSRDz^m1(0>oIl?0(rkP@ypL2DLz+W`|;JNYnwKZES%PWx$TeSdL
zY@g^4#7hK^q<hqqDKJ#jjX10?CYX$wnWS~pT5OWQm*=99?a^$KBdZ2AvNl`uk1{y5
zEfluO3Sbh;Dbrk&>{G+3mbvijjQVIT(%UH3szv%?PM&XVXmn)YVT0^}lH#>^qd32-
zkE38Bz)tn-{9hO=l7Z?__lK$KLE>OgN9a!`#c8738Mj3kG_}Q{5#wXSWdsFY%a{Xd
z{<3d9YmR3=hZz#{bt1$(>K*lYo-IyK6Q7S))`k)G>G!1gFkP(3FPIaRN9DkYa~k)^
zLq)@7&!nI|yK4|TXYRMjbP9F0Bh^e<SZkAnLd-ykhklx*f5^xqduYjdiLfw4Y@vZL
zmmaKdFcc;$PBF$EPWEzBho@0&Exew+IM1PzI1?Sf_ODviZAg{+I|DA9^1tl=G>one
zHhyqQ=u@kY3t;xJE?o$m^*WBXk;{Fm-29@yzV3-<&jizHTUZ#{&Fc4&HvIY_y;hWb
z>g}z;)L^)ASez$2ZTZBTFFWnf{k0vA=>PIQVkT-sAv{$*cceq9>1Tm*cs&3%IS8%3
zH*ZC}=6Y?(a{<v%1fXcqS7>E*5mJ#cwTMCVY>A`kDMM{zB2i0b`axQ;UNa=U5v|{<
znL|^4pWG;uh&me^N@@y6pmoPR3>I3%#L|#b#1QUD|4h`2t)0#|mto9tO#oA1fIyFX
zWA)?UoPjx_TLNbOba~sCx$9*saR}ss?F1|l=p~~M{esvt5%Im}#WNHd)S!(|G4Iu*
z3a;_J!Kr(sFG(>9X-iRPAL;*yZ(K}0M~Q7wa%3puim3beC3i|cs5du%4<D%bF0vf(
zY#T?|e&dJP9@Q{p;bX=?j)rb1ESjNu@Oc1>-CK)2UBlmTyC;Ypa!2{j3HiJI_O;yy
z(sFu%`oBBckN%_sw;KMComG5emAJ9}^6O5oso~o#VbYX~SYXoKH3QC*@a5alFgxkg
z_-$pvG|}g$rq9TqA*Mk`;kV9vw%6ZDL|xxsJMVU_^9Zy5Ewpamp6$P?;A-m=vQ?E)
zzFoz;KiybWXvzbg|2ljv)5f!}q|3_*{g6a{U)KYOiM5s-9dh(|bgat5_HHa9twezr
zyaJsovI7WNUndV5Tj+NU2ygWC^y&{{Zvm9?DNA#E4G8yFrpOXF$8~TdFrpoGtkM>Q
zZDUd&Uiy%F8ZB4;<C0gm2JYPQ)vcU*b=W#t3;#X80I&Xur$Hff0Jm_9Gm2p$5a9$a
zz1FTJ&^X=pBQRHCZ<tDi&ydRyv#kvn;=Qu?v(;tK3Jao%Du8|TK=pMCSyy-6tOE<(
z=A2bmwV)IiEdCgv%3%DwpzTcMDh6o`elD#y>Vte=$IWG9YiR^vt-pK90>>gcQOxcH
zQtG46Qt3~RBT_qk)mwoTNM>p}9byI_;mJ-=NsbVvNxOTD-|-ylOWv1zqw$7ZvT;Eh
z2Dy4g{v+@b>XEzT@VEWOQm%E$%I*P0QUHxpFetpqoQ~9c1R|KChTN+69NmvPq!UQd
zstin2#fHwD=`|oHjp)`Qq1IsF$DTSCiSquJt$49d5+NSnPN~TXjiq)!#OZc0Qna%2
zpi9(IOkSa<$DfU^eXyt4vaDt#$|oC7t@<{!#%cTsr(YO&0a~WoDRJB=y0}BDj&BTd
zenL=Gi?X;SIQU)KUZWm2bt~~J6Y+uzKLBzVj1c!E3AQW3(5g<Xo@l1St&w!OOlzca
zznzOcnLE(o)n1N|-C=#4AeLpWJfjGDl{Jmn`@l}w0)plk{UyJ(RdBYnv{&3P*(#e)
zr|OU8xq?Pa7u~@<DFQ(bfbdT~Y8)S0xI;RYqDbW0+Nuy*?vQFuN$zbTWur%+TS8Sq
zo9OkLw*fPx@NfBF{qYD~-c3>@Kh=@MXyHdo`1PXWs{ti&&h^1;WOBWRmE_<Y@S4R2
zOwUjW?<~P?%GSX-6F%8Ncs}nAa+do^r2&PEUISNlV8-MLWXx>r4?wf<*&Rr9zSSJt
zT~t()VuaC*mE9LWw;v=_*+kMhBKNhq%dRtb7G!nn1kMyze%}vzWr5u2^ppsZa3fz3
z6I#Q@lOfb5669v1KRZ8DN0ICJXB?s(IhVD7L5kx3`*gDlKMX6h?6sObfFk{dUx)9$
zd#Uo7{SK@AoIG(SZk;7@R4|!sjX*&=W7X)=Lpc_e)DvZ`o)<O|QFGRjrv__Yq7Vp*
z4<?*_v)xzKcv{w?#JAi*iJ|%<(7CQHs><gYQSi-v$e;~h+hW%jC#?bG3{VUuOkJAy
zkM@xZh;vo*M%^?IETRZABaY?eQ?EmjunoVQFck|lCkw~+rZD4{4I=s~6K!^-*Yd%^
zTBNkMnTC`^o<z#cCoeK0R+G<rVshV?yQ<p%EyFhFkSX3WnL8P7vgusPeN_9?gbr{L
zUoH2R4YC{GUKkX|I+X$++2^6Za{Sk*_Pa01(Tw>O#d4m5ri5-JE}PCT8+7(#Bk=gZ
zB0U@P0B_~9(WN#oq3JcU;m9V7bL18+9Ct<e-x<P<#{vV{Xvr=Xa*}`W`*m}n7f>~~
z!zql1s*!9o&)rL9J8|)AT3^f>^T9{{7nhkBYV2kebXCMfi$OA;v*OG}d7FOoX;;Y@
zRFFO}#!@B}xh^#hTVPqk02lzV)pO>5j{wL0-Q#PqF988JN7X$5Hxb!3#e*(}A)W<S
zGOHFtwI||DsvbJzf^rx#x<xHn7hBIfP(ce2smLr6EaEvuff(W^=;<<N!Ub5s!QDD#
zUjoKdlJi>}Gz5WYKMysM2FG+4)fTmUx)uNwpR#Nt`Ge;ZLm0=*ym_}Q@g{H_lX)*}
zu>?l=-F}^wigXnqw%2g2O$Nj-2WVXaZC5&2syId()NBw$keAPsohk7U;D%Wmboc2X
zv>P6xVe|F7GH#W*5^2CRk>hDe7YY9jyLOK=*|oKhDH`UmQe+-B{=j*uvXfHVjpGNO
zhiS2()W<TiZX3htQgPBA2NO+2D9?ferBJEh{1mi1bRl0;|11=&Hn-QLWxv)?=>kGS
zSTGfBwvS)1-5Y^?OaTDqZrX_+=rDPQ2tt6hTvp3y@H64Zc?hVAXmQ%;7PF=$^$oH=
z89K1eBR*kOJ<jl=lBb(*$nXp;c()>>@t2swMuMf>O-HLV`|Li`ajo{UBw@wjptFvN
zo?a`VoC|chKZW1zQ@(?KUj87GRUFmroQ_@%3m$x>0R13oiTSODZcHe?&BC|(6!J;c
znOG@=gRVe&Pk*8p)K_Bv+&@NF61z$1l1Ms6H#Qq!f)>`ME#Fn!0Let#p_w}kWhWxo
zqT*;1LvEioi_h`C6IW#KsOzP?g7bF~^IebO)(`6&Y~&y*)@x36+-|QP`dtf6A<R$D
zMfSXZtl#(kzqzqK*?)86zm%_61$*On-tGbPdUei13G-qHIY`&hw&^re9q-;2My3gV
zcW>OukmliZq&$igWoOTY$n|UV?^i2b(12_%Hf8M@VcZdc7d_MNk!--}nc|jpp%_KS
zgA9#G^f>W|+gx_Sv$sgX3hKMRbHQD=zX+DuN8$HX0X^>i*7=ZH12tndlZu47JxRf?
zhb^r6ET`^FZEZqz*>hg`l$@L#cOXU7=5%(f8WX-o0nyQ6Z^Z5%ic!pMRMoc`VGtKJ
z&-E(1n~M=HtQEY}{5->0Ys$rDm!*zw97EUz4|y&cjC@Ih$v84*aak^Mt<wfL74zS}
zxfEs{S;vCIxrTX`k{a*XE{N!Dg;VE~cKd5TnrBx;2lEmPaH6HPb>fWL*(E21JA^11
zM;p9bqt>~*A|D*@3o3F6L>nX67+S562m<Q1|E=4T{IIGK*2BwQezaAR$F4gQC0W1d
z_zNRa&o4|^e8U(6p5ZhS8&^PS+T`z0g6rdA%XC*C6C8{o)wjLR(IGZ4%oVl}iw9oP
z`vfHNzV1>FEv@?jJuDc)7{WnpAl%9#IU5<ypMeJrYxrL6s`xn>3#El<Gl=5s7<04n
zt1oE&7rSkxS0e54KC$qr+8&VbdbVEqn8<Nno9R1$lAO(#L^q}Uj~m9#8nqg{31(jj
zwK<z~&%$<tuOhiUli4M$`^W^WlO2_3LlmpFluN&+M6LUWoqA&iKa@E+^!;jAeN%hq
zXeRXwfu<Z);W{2&GvN8u3h~dt%GHOtCWSZQ%YY?DO8l;fdc^)ZN?5B)tGwa8;7-ZM
zb5Ex2jEU@C{J$t5Rm$OCJ4JU*DBj;Gtvm15k^ZY&upd)qvSW=HXNE7(e{d|b&kOj0
ziRBr7bkGF#7B7W^yUomPN({99`e^ZUq?q$n%~*FEH1-!Hy{fz1#uB-2>XL1i>?y3l
zZk^sIG1Z?0wygpwbV`uTToQ?nMCzwDx7T?oiWGtPgemMbU)(}?&!np8Vk+FeJf=29
zR4`oz48ZEC0{0s<d8jL!ONL3$IFk(>>fqT|4T$9}ccV>UDAcTL?qMp3{W?<zpMTbu
zi9Zt#k>Bc<jr*AUl`rD5s=w4>4>o-D45VXuao(iQ{&bZ$9y|4Gx>Hc&Y1b~+Elst*
z{zUCllf7Uv;ykzI1F8IPqmB%_d$?Uyoxg|Y67jZuNR`SRtXglylNO7JB?TxTW_84d
zgv7x2jyo0c7oT>wwA-*Sg;Gi$?zP{E;F}7a;4;;(0a-QT-~;S(S;m4Vm)&D8`1D8O
zGySl+6x0jE`niGdjN^o1c3S{bN>kmpMW&j&8@*eK65pcowjB=gn5Df#+-Kw(i2;_M
zYNXP7^PebzVa~x`Jn9s?Q}Iu5<E&TsA}6ScHq8EpqgDcF_Qy|fl{M{$<ikX%2|g@j
zT0xx%-}z$z$?l5-aEt6&Y)0J<;|7NxhJPabDcql+yLCHNIG7~%YC%+QiWP^xvHVow
z>#93u{E?@8r?Eez3AnrDv|-ZgVeNVA#b(oWC`MycUw2(wJF&HYkA6<)Rwddm#aqR8
zgyPfrGy)}Vqn%@w;DJ+CUnL?#^0=>j`^!_igB5^_zgjIn$_Hl|cC2Y4AXV_!0BuP*
zN-*vy30<_Q$C#&@9q6VUX%jn#o$052euBRY9ptZ-y#S%#NvlAv>3qew*!j`bJ2X8f
zVnWoS4XD57dH?%8U9-;nx>CWRDxb^_B&u5jQ9jXfZH6O~%~VjCQQHCylTE{eK)W=s
zte~y_$Yr7}4VY38G{L~F%h&%a4sWuD6<-sWP{(BH2wO{yUih`2GT99n=Wq}!sqOR0
ztH5Wq^O<oxA|bu?Sk5`vbfh$8Y%jC=EMnm2!apm|W3-#f?(075FBOI#NW)DMp9nLs
z1^Q_hMGvLB2A;l!jP$*EwtGVGcJFhIEXCr4r1ty|Q=j1|!X6638^xx%i`q>CG8@-N
zZ4a>?5-FY_N#g~dq@xE#tLy~JrpAO*%n74Uk*+cW!)xfSF=;-xq>}x5ud~b2&YMX+
z?Q1c-;d<v*p=jJS^l=IN=ng>Osqx&+IOJg<N3kJk*Pd_%kR7E<wyFWS9zMZl)T%|y
zD{Dzysczf6N=BB8-_6%pgrAkBJ1T&z*jawPF$^?yw=iwp-8X^3RL@lDu_HKo?bG=X
zP*S)ep}`(0)iqX+!a@M5Fu_#EL-AGBNjHWmdYU*77uSE2>GUwQUY6Ec1oK*2$z@Ax
z7&JQ0&q|)xJE$;?`brg<Ghet>N$BxCwI`xs^z{r|?7N9M@vu#iffJ!+s2BqD_Da*w
zOEaO$1xd4_`kc|f8W#BRU^M<5nf{d_r@HtzYhZpP$9_<4Fh59JIAJ2Y3WK-zMdc|I
zJ*M{u5_;37`>|K{gG^30aj@a{%CoE!2Jt5sy;&McdQ=TRf1l|L7xOfx4mZg*BoXtQ
z5SQ4^Fh!QF68I`n9J~JIVp}jXc@OM{2&t~`wSK_oroC~L<dBRwBE~)Pn$e%cOE9qG
z?qzR_ev0lFs?S;~n@63sL6<dJ8OU5FA9XE;F1W2^8)6=k1&L48V*__J70?b(`%K$?
zUbLU@odUDo4ual93?Y9q8Ls}LK&FQ|`FID$)KZFR`iqus_@@36$tB5SoIH}lmCY!-
zgx(>XZxrPtKg|u|PN$2#Pao#@;2sl-97iDY-(QSJwi^-V0*w4qowxjqKMIMUrZbF`
zA^dpU*KBXROf(nqm-l}>*zp?X-wD?GflxdWO}?GR{<YLquvQLeZ3p;#g>MQAD;mb2
zjd^fb<-Ny3=q+{AFQe4#oAl4C)qyv-7eN8lfD|B4mlm4U>)Ck9)9SR{ltgd<Ag$Wz
z*?~BJNl3!(!`h(ef2`CI|4bPG<@KG#rm^`j@TW4Tv2~{}EwMf3I9RP6egA4Ob19Na
zNv$Yjf|7`ph~zY`?0Q@L*#-(>i=<kT9MgT=!v5H=P7?)dbYAt~^RXRi0^5;R@yIev
z=-euzh*!>b%Y%OMFcwfs^HKWPSagLj{bldxszZb-rA}zbUm!btrn$^-cS+7ZL=k*;
zxo3Jv{gOno3bp?SFH6DRP3pgA`k=+5R=ZH%xd|3qo6u7WYC702Hy2^eX<X7a$4X(q
zmJQe%92yLQ-I3kP=G<$yu&j#hiP6pEt<WPZI?TVz!q26(I8hDrq$OnX;0cx(_S631
z-qf4Lc6obv{-Mjtm=UGwKF`I7;P4_)ZrJ^zs{J5qHDRTZb?@{%WfeR2sgh+_q7Oc!
zDNKax&*%IPVR@0;NY)@ibcAJ3NH>A}{b@IXd((FwV)zDR1VhHtRf24it!iW84Zyh&
zlQ9I#KRJzcx3rFahpevd+t}pZzvymw<z5z(+Yc?>n1rb=7?@NGBYPO$bES&?GRF5i
zxl2Xs(T6iC$j>g55e<0Rk`i^`&P+BV|5ZmKTO>&Ay<ByB*)_PfU5B-*%wZS-oY<Sq
z%Y(`a|Hr{>V>TUX-mJ>*#n`xuTRV7es1(Qd2mbAscU?0lTS*;<Qlp*qc$Dh)q`YJM
z3DXSY7!z=N)!}7OWBA~U3qDxnu!*P#f{<5yRh)$-S1Lq^^nr$|OFk170Tf6;%H}Ok
z%3qaWIX&u#i=lHEObkRhCVRHi(iUszNpz*FVrXb-`E8}ZgxBg$Z6n8-mS+r@@KABD
zo45THs8NUlM6IhJWppk>ETofn<<eX`2}ApImCy}$**m-Pv52$S-UjjL!z4kf!2{{C
zxYGN``*l$1cISJ+Fv`6Z@3?m;eK28e`F<u}1!fbk-chVv=!jjXM3{1pGQRzvS*(z`
zkeBM=;M3*CLN0a?Kk0o_7Xll9*m5Clqu{O-Nd0@sW`?an(cQpl3a2u;3&V&v!jTUH
z?M_`^@a>ed1c~EgsX2tzck-3P7`S4_!e1Fk$ToZF5NC33sunR5<l`aw)xk)!XOcAW
zAC_lV!iYRk$^74S{^aS-5tw});LclN0PqZ68)4R&x$3k6sZKU9SO)FLHXM7~`mo%Q
zHy--apw|mQ8#eD?AAF7r97HIy)G5pXzNk)dv@`d+jTsvlU5nVt5Rc*R3V45$Os-G9
z&+bKxbj}-Vi<n{GDZ%7&NL*kzF3QIO`#3bO=MEqP9~wR9Gqgl`7@}KT0rcPb(kW0|
zT}k}Fi9?WF*^o7j<{{2h%a%AUrJT8qV6p^%yYkfv;_exb8DxY05PYHD40AW8v`I2u
z7VR>rYs-Uj)9P~~zc{2x8SSmJ$TJ}g=MFW=0_^dE>?J2Ypb18}dhGRCJ#0XRw*GyV
zU}#K}-r|B9`FC;C@&hX0SS{T9iHi4psTs2UgttVVx6!nx1i|IoLX)r0E+Gi)Sxl-l
zrY>OwrNgnqoPeOl+y0{L<8XpoH6(8`f=l20bW0)^F_M6&i67r+BD>4Be}%D{q`1o8
zx*G5ATx@39+P}2enCJ>%iQ}_T7bF5(It2&Kwe#`FoIOUN-yecb_%Zn^Idix?65D4+
zY3S-MaCoE7szi2eeG@P_5tGQ2F)Y=CVcuoxm9!H+);|{?n4v|o<Kn+C>tA0T#p^u(
zRd&hNU%P`7*!$x~nuqj2bpL}caaOph_7l41Wc278kreQ?Q}^?|#<d+oYU~6hCqFFQ
zb4BHwsfZ-pcg_Yu(J9~1Qm^?_b-}LNsl>UU=d2!O_O2_*3R~q4pIcAKWt(ZG3fh5<
zJC%4IqRlS`a5|#*N}D?xCm~|W{mVF?vZD9wp_U>CY)CM+`SsAB$vtGN?_VNvXt!XP
zB+<C-?Wxl-Im<!*TB**IN#`+PhT8NM&*uHsuJ~VWrF#?bh!8J7%oqC({=rUGWmFx_
zcNV-U{IQ7>FolOLF;q?CNVtL!oDe&Z%Bb3;<#YnUF+Ygb3)=1eHpGlpRz-yF-@v`w
zUa}ok=W}i5`;y}|-$2A~AjF{yE?2~EH;FNkym6F+%x<mvb2um~e4r@$Z8^~i=knZ(
zhoLg60hywS7{$m%f*bPlWka0AOaE{o^2+9uM227uB+mGlQ$D6S+x<~BkCS^CW^9M}
z@2^EWF{0%Sn2FFirsy%e+;2!R91B1b)8|zQGWptv`JVUnpyEoMT3ejK5Rd<t%#{C0
z<{Zi{NS<g}pwV67<?dyAWWfQ}r6^~v&x5JO6x3n0fBoCTUwvzuR?Hq!Ui3}Lx0}N6
z4gLF+(B<3Dy~ew7JZOCA9{p><(0}Ddp`?$l)bh`-zhI}k-s{AUM1&cH3VQPJpQy6i
ze-nzN<mwjC8O2zDbo0~u7fFuZH+uZmUvV!m($cgi{HQ;ixVP_>m|Js8V!*dw?LsSv
zPA)cuwF6Dbhn4-pSgwZ7`yZ4ae-;fwW0t;?Sre*!E)h#9%<TH#!()MHor9OV%GiPa
zUPq4YUPr0n4<KWxkwoTFNY-Y|M+TuoQHqM&8U`priz@V}kW72!>dmpTvZDDWbC8Ec
zt%aBtT7?%8fp^e9UFa;aqsc^zu5vQ<SJ;tojC+ez+&bW!4*Z%4GuJ^l_9y^o2Xjd?
z`p2JOgT?28SMpJP=pK<Cw6KVP7FG=Swkrj60U#aG@Edx0UKd@VRi<Ez=~ETEpZiZw
zWrKk62V-P0=!*!kWbj(aemsM8<{A(ZN6d005znHU&UNzL51l;Ve5Ny@K|mn^;1i9B
zG$&6}KGPJ=03A+0_g>T3v0ay@ipI#8{b1RE3avop5_Sb89ZL#`R1ioWQwY)l)+21<
zlxlGQy+qt1z?MXn)drX6gc=!g*jDC(2x-80_zRf5rX_t%OD1=w-Bbejtb@gQ0K&BJ
zfpJD-16)Y%7yXBYyaz|0NgQutz&GU5YGh3c&q<qu2$TGV$kRm6vrlyvgaO)h=@@jY
z=&2;gswxXx?aw{TBDlGfBZ!|(gyNq*FQ}`5T?X4EKp*cgA_Wv=rghV1#E}uXk*{9L
zb=N(%4_vv$_AP>fbnb14x9hniOcyE1Y|noMw2yeVF&K2Yy_Am7UP!mLUX+k7I~_yn
z5MkuL7tuPDU?9ialQe=*{@f5V7X~1vt}#||H~@dT)zlS#TK(3Hpv{Pu;;N@-q_t}5
zBKPL$v{sT=swuZy@{Afwc^b>R26dXc2oB3vL3)V?Y<LoH*kHY%L;Z>QtHgEt8(XX=
zofdGuRdYy1RfzMMLg%W51WNsI__{zAqHRx^3V+pX&cI_heCl>iCdxAc8`3e>;Uewa
zQ;p>4IMSWqD?lfudxh=G+t)>F8x_WE9E1SVS4uJpWMxn`{@s>hXDwF)Nttc?FV{$D
zdFUI{e*`$Gvx)YxA(@r>Cp3Z@d$>QQ9|AV|k;Fg$v*Gyc?Ua4to~hw<dia<ak5N4w
z*bc7$VH5=6y`PpmJ=tnLTb>`G4Ej;0UUKR3H2pX^jdB@g9d}e{qVVPbKg2DMyzOE;
zDj1i2U2<LXwle|F$@!W$>EZ2c63c25@q=HOlvI^FgrR}-A(pFz^124@Tw7ZQior~q
z?>zRG67YV1Gwiz*`0D78KDFKEp}^B{;30g}TSTRAU~r#~)N@u?fDsS3!Huhmh#k6>
zxL;-H)bQIY$Z36l_-LZsQS_OU*&s8I5OaEUh`IrV@AIdgBm#-|RrC-)Dg5J9>t8+A
z@L0)kaXHxMu{FOg)lk3D_OkJlqVm`v?-M^`P)P7+O=b?G00e;m$pT92G|R6z6IQ4u
zGNj*V4L@`jou`i#eiXuaGyDE=jS(!hEFi_s(iKri*m1f~zRwUa0+Oun<JWH7X{&DG
zq5sB9Q(d1WlH3n+n?4lua$fSOvxR^lo;;fJ_eK14AefJ0pKkH@<CS3ML13k6T`#;E
zcE7P(fh!4%Bx+l_im{kUJd8w-HGyLkN+eqQQZwBXRhnr2K3A2x4`MAXnvitkRq3FQ
zz7v!@6okJge1OLStrgo&pzq@=^&+u8%FCy~_|@`cLX{9<eLe$@GF|Gbw=&~8^+Jl=
z?S*%+RWVyKF`G_-;MkXY=*gEATG_^Ba{^O-7bXbPIFOeKK&+1LmLX2OCtTVSd*`Zc
zsCLkiok56~k(*~h{~&$IBj*;locxSXLfeTeKrW)A-Zi9>AE#;J&(&IlevTYT<W+wn
z`#v&ppv3IEN(O_a0|b1%U(@itn#qb;a2Zu1*4Tj;A@C^Mmdst^|H?_GVg=94H@4I5
zy_IU)o1#Rfaq9NS#79=QOH;{LR2+M|u_I{1<c}$DJd3o5cfeJ{BM!N%c&4eEkt{EJ
zBRCfopE;)T^i8W4fWLUz!@|URx^?n;ZGf4`hNg;4adQFE@m1N_QoVl!2Lq5h>{beb
z5AYW>&0V?;<n7bgipTaskjZ`$&8O3Ia;k1&6h^~^qqCtBcbcz_IMM(BTb!lQIFivm
zH3U}e^cZnK;gHO8*Ts*!In`r%cL8`?RHeXas&hBR&*zdFUEF3dSHe|-x(_RK{B*KK
z%!r>n-Xq{yx}oU7$)<*eJ(QncLQp2_U>%K$vDIj<!!{Moh!EgtgmtQ-aI-TWgT({(
zeIl^BB#k)2_QMr=^Nc=Xps7&F<qh>mQU+=r++fTPxCI@{7B5qteL?6pIGUvlC}KC(
zaI7|(?|0dbyr%h6-~`7|ceO`W+hVwve+9KZ;)Qmlh`k=;F?U0QV(mnz0Ga>yWJUkS
z)3|&){%b9?f5tiV^N=jyRBygtqevs?X&y2^dy-4><b6lcGvWimfvVu~pEM_*Ay>IU
zgCm4!MGut2E?Y;CfK_ZCf2ab!vdNiF{H|>I7VMJT$1M|(>-X;No|}i4^OKzx(l#p-
zU;9+*?9FpM4WJL_CG01oyfGdt@ux5Gy;q&CI|kZkTTUumbA^bbjSYScPZE4I_qMy>
zrco%m^<!DkmiK(pz?l*!L6gRYFFM!NUHyD^MmTR^+V?7a1b_x4SU07vmjv)@BP><r
z!FC<~2M8UeQnON9$+4f{P0_VKBA?gcOMoXcKNO$>(79JnnrxQ8rdT}00xqa~2LgOl
z?yZ(x`I70tf*NmE_ueR<Y+{ofTuf&}RBAc7fOmCf{j`!GVw4zKG5(zyyO^|%t~AVt
zRt%jfs|<G5Qc09$eao(*mKJ;H?>51EyxXw+;C6K|16UlH6R|vOM_u-;3D9ny>eWLi
zFl4^Zda+#@!j%X}pNxeW=2aD(oJc*ds9djE&)nbH={^fV$R-d4OqFO>FR@0+eu~HC
zg!edQrA5Y&2%bd_G;aMh4aZC_S5u}@)|Pd#{l=4Xmj%dtHUQi^wT_rR`t=w5)MP67
z5XJ4W)3Hh*mcw{<pZ(*9c*jy!c?+~`rYm5Q6o=UopJR6`PN=?Hn1be;C80p7($y6h
z9uM<a@C=2In+BYnIoze8>pW0G$VQC=NOKp@tB72OijH+9dba915ts;ME@DbgpYms!
zmlf&NFFD%0i<m8%R&w{fcGwCqp0z*OU1U8bi&xQ)Ji=pT;OCIj>)jjCh6k-iTXt|S
zuyKqEKTFl|i-tDP;imwtt4s#JT14LJ+!+bUcXT{dKT4X0lS(>c@)x%{ZYi=>+hT{j
zUqG<toB|=^?vTbxfZY`Gzh2xvuKn-L<mNbMbjihS{t{OMG%j*yg;E{gC3U0n7yA`U
zEFm7=;4DnfOTD(pA>Hcsyt*AGd^Lw~UI2g)_{D`T5*DkyUkKA-FN?bo=+r$rtTk0*
z8d|huT!p3^8H{u~{s>p>YkW}j?b9*kWAqGQsK05v30ASZ4+sC8ijgN;scA=ujljj_
zGkM$+D)6$AIAzx9kxX<cB>#S)=|&?6WQ)|@c%}tI7@n;bwul&`aZ_ZBl%ogrW>if3
zNiv!-W5%N;Pc2{!dAD``IM-J4;AUZ|%59)}PkX<n{EwAh;#?2yHUb}d)OJs6_!Ix=
zCcmedDfO13{<Q0k#TO)o>_YA{=z$y+X<MJI;?(^ns{b;D5~+qtrcL@^&~8<R4xr*d
z2GTQf7uFqt%g1HIM9a4P`~r(~EdESDlDN)=Pjx5f@aY%nNSxGWsg1EA$D+Wje@MtE
z8@u&tQvP)%CwkXan^c9H@`3vO*|NBTm*WPu<6cNAH_p!+>dVqA%Ava%Yhg#DXCo$7
z7PLg`0LRrOh;TIb;b>U{*aATgJSwe>O<LtfD}%#D)y`ud3;yuo{qBpjgk~Qpr*k~x
zbDyAB9OZDfRZa5p%LKv>tE6u!jm-x~xeNut#);I<5IQ#3dNE0AM>OHt+<^hO-HhSH
zVJcE;XX?`_&G%sA`PGKwK$W|W!}V3Nr6eFqxM>C14ZYLYMGLxIl{88m#&l5|YR{Z%
zQ|YG2=j_Xc<^!S|c6`gsm%EWy>1?g{)bYxS=|UTNxB04J7h^#({oUUpt6%vS*`{ns
z7~>b%rYf^!pQgz^J|k>}3l9n=D(%31ZTvy}0EVl2GeeK86uTkO`xi17(-!p-=h4BN
zy*u&LeVAVHjC>}#fjhif02Iux1|uhUd~h4Dcfl#?YN%5uUG|%J0j)`>IM^1G6f~Iy
zp4*Xys({SR`r^#0m#bp|^J0s<a2NHzCEVo>Xv8Je;K`gZejg$av)^!3C68wL01tK7
zkn@rVSP7^2{~_o@S0<*6k4&8A6e7}qVT|rNi+<mjn`&=-gZgt=S^srdQVWTCL`puJ
zEV12?^y~3MvTn8Zkf1BPJA|d$(KCNQnK$#r>RJ*f^qozz%+CLuJDI{WytK(6yuNPd
zrR81Gc`xRUoo!;?CMwB@WkX7uzw+9btpe$#O@40~E9&kSjo>-pI^-I)ihrPvEOAOY
zHO+uR5Ee-;uy+~Frlg6752<a008k66wXeqepPhI(k>KGB#Np?pHR+jvcW-ZNTnenj
zDRN^Gj`9C06*y3)i8{S!;gQtts1s;mhl+c^^{Gqj#!WBq_{NGvEw8f{2Lm|w>#@B!
z*P%20*o|5A0B}8Xw2kEW3$|=PBbEJ<^yAP%V-d!!KAvZ~8;(~p8(%q{@qtq-5zXV)
zY`^rBlji|Y(q<I~!F-?ioD@4XReQd&wkMa5kAGsOex;NWgrEN1x=+66SPTu{WPH%B
zl`u0K!)@re@tat{kqkxv69{&?zAj`Zha^Ux=lQP+o}s6^0MeE@j^nSJ-=Y_TqUs26
z(^`K9vd`EKAEE+7RXBPdE;m)YE;l&$qg`WBa=uWMtn|J0*+hP2{`>c`Y<ln!s+6}&
z?DIBp-HN;azn$nJ?5OO~_XTR&G7p{-;JWmleG$Ai73v;Q>+#4F+;h)WiJdx9B&W&s
zx>R8mdlwQTZrOnD^q)MmE8H&g5}e`k2&eomV+YxPM2byhW<NNP-S3L@xI2p#*#yJ4
z${rkN3IjT$TA5RBP8cpbQ;huXhng6TlH3|rs$)G6%@JWTyz&NaZC?eB4R6fk*s~_E
zHSLR!TOPS{Pe_6hVO2emi@=~?ETzr&82a}0>#JJ?R@kHN^Evv<b8?h{Ndjrmq9fkw
znwnB5e^gT!)O3%z4V<X^^@)?9pZB`3Q80emhV*DEpC`CyV|`r-JSlk`YQ`LEqI|NN
zGbJeAM46k{zDJRyxZ~%S<GJ@rv_Y223<Vv8G*CYVgt0UwdIYlys%B17Co$?jqGUVs
zd)*FOLoYpX9An*kwzu#h0)4<0W=yJM>-V}xNye6TPtF6h*fB~3HDm={>m&yGp1AM~
zphYR(1vDa9Cpn-6fP9E6CzUUY3#>=$WHh0ok}41BC|x0n>YWVX=(F;%6=bb7{=E7c
z^Gvu~IoFsa?xHMj_<9|1y$m#mr|Uxs?g{r{qVE54+JZF)>EL(YY|STvS-JkL3YpV*
zapqiXChhij(Y=i95HNs`=wL1WX;QGI?XxTn0ct(8YLtT%!9B%Z$0}&dWoRuN>FVpM
zoeOr6+GcgA`R}@fbw(!qzF=nQ)vy^=djvw9150b?B7v_5SW7q{x`L0<9@3z<-M#V4
z+&lHZJ4S1Os(QY@iWs<uMvz7D_Vc&}JZVbsUd4PxNlp%HbBDd9S$Su1J1Y}=^B*wX
zt7AKsm75IOvrF@#%f}TO(C|JyN2zEw!xH}1&J;HD@9Mv-`uoX*sXT$HVN$<ihNf!*
zGsRvZktb<A&z8?ec@Twm-~0A0VICG6J^^UWY3i2u{v@<mg}v0(hv#s?YRL&mnRzee
zHBfj3Geq)8WR4qdVhlDQR4&Yyz#&2{JE>}x@tq?6EY+HE{n>eoB(aJV=U*s8)4aA1
zzzp^*pK+o*P84<_<e1r}$NBdXfu5XHHBk_<295}=gB2#jj_3Hk+DV&SS3PR{?bkXs
zQa!a1p}OTWed?(3MQGhkAo_E}dM(8fn(g0d#A)JZcfF?=N6u8?5E%2SQC;Ew#5iXg
z*?oj?OpahS62zDiv%XZaRS~v9)v5>+0c;$F?g41g5wuciYq7oQ<XP%xXM=V?2v;7)
zzTc53?Wg&BMsWPRw@v&HVU@noEF7MH1u5eSC3i>SPo}<qn+ZLf(lhNC*a^m6*Z=@y
z2gEM-6KM6@qK=ICQKxcuP>_2>(pnRnP<KQ7A+)8v<(LUTn+>B=O;J(Hb<#hiG!BRf
zL0jw_qD(agHJX(pOe8hWIJE#Y-fzg|i;ETr0(moY++UWIpb%wNBsFQZJhUuA0(KiR
zZU*!@|MDM8d~DA?Dt9o<qk(zx@SyZ^$_26Kl?(Im&_5&X+m8BLg2}B8*}a=J1W$H9
zao5ZcOVD_}#ghX^ZG&{1Dkd^K3Im4J9kxErj))LiD=L?C&ZR2Q%}WhmbzZr8c!;vr
zIu`%O?nDGsjj7mDW|!OWJ7rt)*>}o*$0RpB2WE_5Xj=rtt1w|2TAl9!oP103n?|&K
zd1YEvjXqFD=q4&{&QNViec3+AHyJIHwEeQ58jn^HI}At#{R!s6%{F_bvc8ptEm*;R
zjLzkv?<0?wWCoTxuQNT1X<s$1V+75oHD<z@yP=W-RLwrEe;<ezLQ`p|puGrWwZ=7^
z)vSE8-uNf#G!>gG+1Q7d7sMwZ-^di{M%sp?9BK?av>~<JiEyHq6E{ndUvL^zZ#82T
z<rfzg60baeyNt%<INCo<YFw#1#&~8GD)nCl8RKG<h=Rr%`aUfv1#yQYjzzvji4Bs$
z8=41xLWy=@<jbe340ALq7+qlqn&;w!&yIrjwa}+aojFXNT1#Ll$=%%&4SkGX<h_zm
z`;zI?SEj0~eA_>sz4k2xbG+flVf@Bx;fq)~Yg{M|o5|6buQ(raA5x!6{j9Bz&wv}o
z9WojhAF>4huS@4|4p_^xw-<zbJl<2d0SK3UO3X+@WtQlE!}N}<1;i7?I}OKzB3Vak
zD>Ja-GCt6rR$H$JBc^Di^c;QF0BIe>ueXtL42duIi+4BU?8TJb5I6{1v7i3>mNV{{
zEo|>>S@2-aBl$G)b(RT*)cxMem+KC|RA}GNy;Y9?%7j1qBT)aGE{$hSn>!6x;^fO*
zFioe{e5Ux>T(55FKP?QryEG&_x0#WigvV7`RV#>&`7J$|Z77WW#@GEi6h<B&#qZNk
zRqtK*L&2x;QD{y$o8XgS!l?3712b|O76&5>lbS!dZHTyx?Yjk>FtV{9<lg;Yj)3)i
z#S%|moW96h$*R|*QvhdF79I;ETi351i*O0N!N6B95Rr<uLP*W+YeLwse!asWlY%Mz
zox1F;lee|H_j-rt81b5%0AHMk3#UT9F=B-b>4khCr;lgg+uY-~bWOjiNx%L}Za}=?
zUnexi-ocNLx8>*|+mV{sw6m>5`TFK{ayqhnj%oFOvuxr2jSY9TZu`&=rP&JQO<ceJ
zn+88L?ePCL_YXLwAl!fL$@wnm;;*FNlld<KH_2NlNJ@*d|Kufc*hMrKaDR=yd)-Ny
zh&3;&w>&>P(VqtrsLf53roB8W7G9*>?}fOyK{Q|0`Giy09BWEjw|wcoI#hkc+!W6+
zL#n|$CdZc|MpJVj2q_Gg1Ii5I6>>mXRFcS^IQy_kJVm%HKq(k$s4FPT*2PwfWs8cO
z;sVsSaM}77^?0p}vP8D=J0Bl=$)o~%v8&}5p9oSWZ+UN*Fk@g3w#7a!jtr&<t>nnY
zd3Jab_)dbu`+X|sE@N|#XE2X8NeV>x7@LokOKz$n*C`Ggn9Kwe!o+$QeBn)Tl=T38
zzP+`qrjw0fFQL4X^tMY2xhmch6`Z&42DSW%mAVVQ?b}ke<G0i1nm`dl*23X%35*5n
zwK{!;s8_IjEiTVM=yRRw@9yBQ)jLsR(@|Hp9C<R9_Dxo@$&Q_kJh&GPvz`Aeox&hb
zq=^ht>7Cv?MV29jt?~U!rDgweS{wI5CEQx&g?&X`^Qo7R?Xk{ZP+(=2B)^gkvw&Gr
zoHeU(I-a=7N%dn?t4_{01nm&yHs?vX+@U=)J<yisrY0uapAAeV@^pe_>ZY7=r4m=_
zG$u3<`tXHQSnxS|V)zIZM!QQ5Hu#%IbP7pK?^=xD&bN<?e@!8hqIVUJIb_WmJsMi%
z1O18jdDldy7r&ca#6rW9Qv5XRgpX!7n?rEMQ*ZVvoUy!qEFdlCeccUy#_{@L#|<D_
zKnnJe@b1kJvPGrlF~@mO0#8UKMb{-d4H2jCi&l0zKa_CNlkM??WBUP>i4cgCWx?BC
zr${xotxkvTJa|=Q(T;V|F1{&9m0Q}_`&!=j<RA@C-rm54jkR{==SV-UfQWT=9}UtF
z!WqR<mxJP;bBAH#RE#?WuvAb;=nfP15CLO9B>X_DmiQ<vO|WmghL4v5XDmWBU`pO1
z*$VVf7ty3H#bKQhK>!<oq!diwXoWc>1%!ase_{oM(jkCD?&$#t7A6M;&ajqzlF+hq
z4IeR$lg~w{{xkZ+)>~SBe%UC+!ZS!y6@7&f_dsiE2?b|!ZSjqJKcID&cJoemi3pPR
zVBxQk805aTBu0+~as5I9aIsYH20ljLBC|FQ2vw<ax|HYQ<m~#<%wKYx%_xu#a{RZo
z8Bz3et4z^?_dtleUaDz!w6-)66k^Yhx!;6EYaZDmEVaw0NxL*uRTSA@(soBE?jsB-
z>+xkb1lLvQNY)$Tcj+2{wAb+Fm6AoFfOI?G-EczDbEwrZH}ECPxWrzkR!>_UXDc5W
z)~Xk`gvnZFv&#<}IM^V_ezEniWMySV^-7tF(LXS*Fp<#bAeZF~hYh|hRi#17iFOe$
zbKw!zY3R?uKt3Y{uJw1(k16Rp+fLj`$ExMe`WsAJHqO(geHHBxwHSb>f+bY-EMWPB
z7|blE5B|lHECJ-Icc;yNl_@~MufHp|&D~@9g6}=bYr~L5rb;xDp}6Zwh~RAKM$UFe
zWKBX;!<-@&J>EGW=nE=VcInOzwW&I>+xar{OCSwPm*57=namP)pB5kAH=$6<9T81=
zD%1?!nO=;ika=Us(p2r1I>!5hiO!I3nuu?q4(NM>njc%Opn`d5!!LGJV-_!mFVP0l
z=2QlK4T92c*Pq2KygX*`WD=F;WxEDIOClBm_Eq`rgQ$chMAj~*PmyS{Q)zH=UD0|1
zVFY;l7~#a$->BV+CP6WUGg1W>{81s6DoxQCu;_BbtKNabN*t+6GuG2+{dCO!==veY
z$Tcv$fR$dclwb;tmMcg1;q*<pxqq|5aEr~0B6Y5@f<^>{dCvOR-3OX8G`BFxWg=2O
zzr8P*dvcwxRh=E9DfV&yho!UdYx@7cwvwYkS{furH_|N)A}~Ty7|rPJ?vQ4*NW)0!
z7%AOd1Er;r#+%Rm{rwSp@9{d%b6uy29TY|Svq<>OXHH%7$9+#rvvplW%iC{`Q$oAQ
zru{9bHDS_1yMez_KtqT(&bnTx4i_Et*)fd#Fa6*4WyW`xVvRc7;S)t}+fcrfia%2h
zYhTh;;wIDhbNjAMuOwR{I&kil1IAFBLq1VyF=a2X;@z{=)&&i4$9UW)I!DH()g{Pa
zm(X0k>pou^cyw6oQ5NDb;-SJ{gR$tAS8zR!M-SSPF@4b?6^u6xzu#8bRN?A=?0msZ
zX<qR1Q@Q-)+9I2xXL8Z2c*`hEWAw8<{w`vhZqulnateIEs~4TLcYmGz-$dex1$6A(
z{~vUBec4LdO&h;%Z0!23B-6jD%Bx{MB`Zf~f81k6Z@T?sMmesY*#*#SReFxAc3Pp0
zW`#mb*O7T4*ci2M;lW|`T7`#1?^wOw#AuqN-Fx~QHlqKZpL&1(IuqsJOf4$2SDh*!
zuFVlkM^UXL$$>7n+x>r!0F-0*_c>1AGmQOy#<728gqZV;tMePqor|wj!|54iRJr*>
zm3ROUmYSekj-e}uiR6;HfLX5NXR6IoJTI=O@AVhrAdi!QZwDral8V2HEETBys#SnI
zt-3*l@joX}A0&QHPktiVg?eztVn|16PlY6{1FwKJRE`@R$r6ue3>=RGnIe(@D|4NG
z`mfC8nWuJM5;$}llyW(Av)6wfn>cX(`+&D={Jhe(Ht3#K?4nO{GH}n$FTzJ`?4<l$
zKV+6|vb1w+sC4vS2+xJr%gwWV%Vg(4BWmyWCFmV9v!v@wm~FFNOilE&=~*Gxm5h^!
zHT=%r=RAD8(c4VqWigO1mQ*Gx_zjjP@0mS`1km8jJ~oLQOTmA*HLq`#j#Bm@l!Dz0
zc=|}p6`z6*WV3|AI^dtRs{@i8*IKeq*4uYnbxt<ASo7?JI6jfg><$3SK?ZImh#eXI
zIc~f}UPh0dSp`9$^M(STp5Q$lfcU|-aJAZS@nyMvF?X_J0q-o}?D6=$0xY^`_M{ck
z&bYBr<N8U@wEx76$XkN>!y;gN8Ih2Li+I7F(rsEE>on!vWedWma$L5GMPeCqvM^E8
z59JLt;X0Kd`YS@t8Ea;lopKwGjF*XPD$N<%ZaMPMi*bMJuAHtu^I|76ab?LvTJ<#~
z_OJ+01D&7rvQ)R3jg&8JST?H{*KcA>nujn%Iqb57Duy9U9dv8@Cp;zT8c)V{O-bEb
zyo7h@b*)=V#Wwr930p^g+vtWb_>bF<bosLSJ>*O*pR6fD<hE89+YwA`0ijOt9B2Su
zwm~cvrqEvweubpydWf>+Jl;)p8G@aR{aAA&*hDuSumGT0p2ssN&Xl?x6>3r6+!{PQ
zGgb8d6t7V=iN6stdVG@qv*@ZQZ_ciu%rns!9xPIw93~224@Z_2NKi|eKdd1x0sN{`
z{g{VRh0~iASIvwc^f5j?&-E`cl0om<_3R<iQYBm)#5`uzv{2$_<YiXElY?23f_gI=
z(^$J|SBcYp&^$QOG%2BCXefHP3ES>@vl}n8@u9Z+`RRp^`|WnVUDDW9=Lj;h36Y`!
zN=#6~L&61N9twc&mN0n+8QE#nUM#VJ&s56VsggY?h4fxAAMxlUA+k38ty3*UZoWkD
zYtowLHa29tx?nSN&!9uh+J9lisnR9`24AlkH{lim4QM)F)(XGH-GH=Zz%bg`Yzj|P
zpNULt#`b%V1!g#1jK@stt7$+4vy6e0I<b0P%9mq{k|_ERS1*Wf`ed=!XQhTiyO;7_
zkwdP>><-mH{O4>s!?OsbiOr+$cXvhxAbhnqP-*e4S7^D=xRo#2a56#jwtttd27xv9
zF(-eb(KZ^l6^OAi&lE!$8nnp%dHvdDfqEvUC>)vONrHxDRug7q%9CEU88!)PI=)GW
z*F9Qzf(={dTn#lpx;VbEv#M%a&|Umg%8HmO+&>Ti_7pJpBso;Fs1QV4QZN<^iKybp
z-^4^X{}3YTXiJ}j1Q4XNUdbB^Q=eQWR7JgBY|E_?0?LkRC8dqp2S@>Gt)_^sfmhVA
zG=gh_*;Cln^YfD|X1zOSAcl_F)3h}AL`7@j8t-Hg&B>6%BOHqCof!ElWWMI2Ur@1x
zF4JNZOoo+sc^p@wkv9|X?HYKbV5YLKb-O(L*(l#hLXuiz&J$c9!hjND3cK~m@9vi&
zb&%v%t;Pi~+l<B_;b#ki0g$OGCLF}JtEK)LJuRF!S^|~RJJq-W^hU#QSBA=&c%-Gk
zq516$%?N7MF&T;1Uo+R&7nFzfK_53`oYHE=GDQoubp33q&Qmz(3I|wpdn^zv7X7?J
zI_tWkasK-hO~Ny4bEmiN_wS5`2SL<sM-?jTJ5d4VRTU4Z87r?edhhhvvc0LX^bO${
z0pD{{fh&nAF?_dVd{fA~n5i!IUHdldp6@|9g~Dce1xD6k^<<5+Cr`fR8&#L1H=0$q
zM%?4&8`!iRiVhRE@qaOyLKF7vQ^m$;K(9mF?d)<eezZ<WSrc91Z1iGERh5QgjS&xJ
ztE6WPm=qQe)oksi@V)74)4Zx{Nm9nL+t`=!A5^Z~!b;NS-K%h#pQHlXAAMJOxx5TT
zu_d+|Hf~Vd<LDWXUJ*k@$X!mdOi$GAr0srk6?<y=S!x-0U!q?rYE7Q~m3s?8%S$*P
zYYEC5Jm>1cjRU;z@dkfGzgY$)5Hnf~r?-tQ#>C09GdR;=Je!sSy5WQAR)iZY6PIBb
zrg^2^S&73L-sobc$fzB<7+6B*hV2i&5)<#lf(EQ*P9jq>P2ZRKpqLy};DvZ2CNTm|
zKT&RqKB)=&KXt%c)J=3Ad0%4+{_hF=^VPKBnvqKEQ8pAy6$Rfem3wTcumzLJ07b)D
z400Ee$=fsl|FlUT_AOj^uxVovMN4$f{0NpEK|hzLnmP%ZhWn{x2D9vod+5;OZ$Y_e
z6siEPIk|13qWHA9>)ckdZ++Pd@`&;zgT=VmZM1Gee*~r)d;j(|4<hNnKQBVwl_BgN
zA4YW|l$U<j!^4zRb{CY)Bt@hi<4mPo6WOir*Z=O<L))I@y(N^vh-2{zYxCD+#(h?h
zOzaG{fVitkrKAzp%o&?LYj5bqKvgh}co^RySo~bZL+Q^K(zoqDKrk3tg<Y@P_1DS+
zPnt*0_4XeO-xrMf{%51~<oM70oVe+K$}^y6@95>_|F?A~^RJ|=pE-2`k^9Ee(9Z1T
zopvJHL6)yZ0*VApqS)e1*424J7mdn4bh(T71EUAqYhG=~%me2S&%R=@T7H?&c+pz@
zS-v|DdtwhAWr>k6uaCbh{GP(dX#89W#4eVAAwpCKPL{3E4gZJht3$gC?UEFY<97;P
z781)Lt8q25XZF-8ercDUpIax%Y1ikf7CBgMa8Hjv*rehisP6{!36U#7OKF|m-MP4b
zJ#*1(G9fyl_`l!-qc&zX`W76xsfvhTcPu39Tl*??^JC`o)niAiuLku<0JG7bN-cZE
z)Bsn46yp{*5eK#+zqi=F<;In5;yVJb9Hf&%Enh8&hr0}LqSbB4;2y013CXEmR)9;G
zhQiRP8)E58BIyC6d^)O@7S*9Hg+}^~XY=4#OBihr$CUGheuJ`5vzv4wq2n~HW;*h4
z?xeB5o1Q(Kl=%3&h_C!9v|s;y%jJlbCO>B!itdQ4H}$fu%4yLUAx+{G(V~+xHIFh+
zm^YiOLw=q@W`rhwGU}O_`jaC|B<#zUYQ3MC(y7ZJ*}{k1xr#;h7V{t7;hNjkKN&I<
z0XZ72><jL5s74|L5-fqIMx-<w>neYC3m5c5^2}33RTVq`H2UUWuGf@y=88sN+6_}+
zwItMddj43mYVUmGI3U7u$DgCPHK9+xC9nVVnwr1UDvQM4-X<L;C{q|&Vhr+^+id6>
zJN(Q}59=ZWLV7Uf4GCWJk`S|m^bJtQn9i#>3|63=5_}0ONO4D@KqgpWm7V7rq)5jB
zyw+MUT@Iq)Oyyuj7*|v;;RPe+J_w5qth~|Fh_&hJ;DmB{*%xX`DaFRNboy88+X0+P
zXP{?3g&zs>I7*aljOz;A*F#Y6w}ehAG;s@{UmQkSz&J)TAOrnnRn#Vp+!7BQ0SSos
zqHE5YdNsJ$8(3&hTqICy_KJO#WZ0bb<^7>;jZ9<gi4lnplfkOGedF@r$GBi+e2=kE
zL34&?ZhS+~%Jm2~G8K=YeKD;GT0~sGd%cu0(KrlWCj>6lb=e*m{-`bkrf2+7PL=rb
z*=e9C0PLd0ysz55sVYzOx5N^`*z#^gppP6Z8e%doW0Klpa0cYirBnkNLIz^CO#`~N
z+swmPNI72V0#0Pj)ckGsJ|JwM`RenbomW5SHD;d?g#pOggDQ>aWD;+w+j%h8@Vw$e
zwG$lUg_2wd4fu@nGs2i%T|wKq-xU7P`rke~y$HL6=5XU#f$iXG{yo)>;Cm8OrW%X$
z)R4%W3={8x+t8HqQn#&zjg6;IqHDwf(52t(j<2*{Ssch2zu0Zu)@qhNNkcHD-^0QM
zS%JMCv~7`79mNPqqNT9VrvxE+ZoF5%M7I`<*8J>M6$=Yo-x;l|_o!D&%R~SmZlR?W
z6-yWovb;0KkAxy;mb9!zVW2(UZ`dyS+BL_gks8w11El=KdNuQh0V|uUf>37*vIQ=d
zAby&l)w$J1JfZL{3dj(LwZ;zCK#uj_4v6dx*<w<xCYikcAR!TRkx<Kt?CT(Sm}_;*
zq<82P&#r8K24kB&RN71-(2gDQOiWR)we$*T+TRg3@R8t0a5f*=%|OU*>rBc8#u4Az
zoQp2aWk_{K=GRou?#&t5eX-xx{1r39T`<#dCT{epXPDQUKQQ2C-<;lE$D5q*iXT{C
z)^Yof%bt3IJToV?IHe6;vVYWPH_Y(s)-ZdJ0#3@L)7z(y8LAF7qvEy_wy}Xl-)`!3
z<(Vx;H3id`P~SmjRM}QWjb^6BdY4Xs94Fxi7GB6(mKiy474{O%j>2DX^BP<n9YbgI
zw#LR3l8l8}RpZB)-M3Y1J$ytF&APd8-YeP{23a?-DGYcCw>fKd8epqu=~IQtiHv6H
zjv@GFr?dtgda}In+whDLF6pkYw;>#E8GNg2R_EyBhsN)hR)46*su4A&d1^`-kS{p`
z1jp_e7|XXTmh;ZEhl}pnLcUgKZx%aFtu0gP#=yx_2O8A~+M;=hpGQxE)9~}KT!z}O
z(u8WPPoge6Lh_k4Qk|Ms@>aYkElmEPc}ChYxaD1)<!;Fg)7j7{bCu+B@In6=kQodr
zl*W<@&Zh_O>nhvXjRs6oK#~xwb)yTSxcB2?JDc|f>iML(R~R=#hz{RL258Ns9znNt
z9=K2LTma+i88Hf3%x^3T8yg25><lK{55<G4pfmZi40@DgJSK3erq$IJOe?6zD~cRk
zc)ul1pw3j+_aqMNP8e`Q4C2!#U;#cB@-uMZdWTo|W)vQ}71`mca#%a}PSBIF&+w6B
zw<B}lo>NS~r{tf0Mx@POU-kXiRIly4w>2;-v!?GSeU8g!+f{;G1XrkvwIzgOXyaqG
zEs=lNdid~_8Y}RWeY&%7YpB&k_x+)dJ+VTd*Y>#+39@;AtE{p4o1q)PH`gH~(C7P?
z^KH#YX@F|9`uIc4FU#-@($c=U`FF#d>>gmApQMV8K>U#{<)FXxZmoegj`!EKfnQ&S
zgZMi8kSzZ%e<sh%f9X<t{^KKBkK^B56Z7>qrWF6TR@S(;ChqQie{aJx;>XYZ;FnxL
zX_W6I<u&EFUG^dd)y8<bbVWXmXS!sT_wP3eYe*bA*N2BTc1;Z}5>VX+kPUhM%S6;=
z$8&BC;QK*>ZtZHjlAdQxo>(}zh~MZc0MhGFq(%Ccq-oLd5{iaKu-9dmFalL009*ub
zdSmG{;r>LTtB6WXCNsKFMtqqPVbWq&V_z<TLnv3urHJvsn-h$N4Mbv&ERp&#rsF=9
z(Mh4HRFZlCpcb)HhaI%|i+aHj5)8U5;1-!_2YBYhD{LDTXLQ}fTo?mTu|r1tp5|n@
zzw?7QdyeP}8XsiSa7?U8B>=Ty6=x~BXlSM7v|XU<0cG@FY}4>=xjVG(xO%i7;KyD2
z`rnodobpYvu014#U<e>axmaN6XKHve;C`t*F$7JTG{iitb3NcC0XLv~rNAAY^7;sX
z0{uOroDHO};wrEG1im(_{jBnG%x{8W%WK(u(TV86FZ@)z+}K0#yU5!g1#B_Om9}2t
z;Jqh$N=#JB3&7DD8EdJ2g^EhE=Lox`EMyR^B|tK#LaU)=Q4CU+JbSop$-mF-o6}im
zUy{ozY=%OwK6ZEJzqkd2*V|onepBy%CelW{!F$cM0Nx1a%G-2I*EEH01z=eI>hK)Y
z0OPOGcM$#yoFPYWx+nN(Ol~RsSz6%mZEsf}ns8r(kOy`&Q2VH3={H=x*^{85)a}%j
zYQgHzdP`4KnlECPprfU9Ysjf1gJ11Z{fST_nOY2ff%Q9na*7-yrSv<9#(loly;gep
zr}{k$jhUc3g5vdI9mHcO<Oc3Hff*ih#?Q`65QevI0`(tUf@8+>M7i*;<xu&4rKve~
z){oIkts3mA#-8hr(;*UUxvg?l$h=3jBLLoi3hmg6uMcvLrX@5Jv#|EO=^ClfPY%|6
z(J*)Cj2gmX&t^ij;)y2^Z?Qo%KB$tmG<Jf66@%#zK=!YOl6mSu=-#XOc(MfyhPIKb
zUesEk$=1)|UsQDQ7x)_(iFF-vt;snix}O`H=WnH=Y=1z=Q~g=wmu}mNJw9p6Ux7Vz
zy}osMNbc*mY{1veM1Ere2l>v(*A(i|Plzyy=jl+MC4fdx`U_se#E|o!{tJ!v9d3UJ
z9oJw|k9U)SdCyOz-HnXu!)!&L9WCF96AlUux$V0N`*WFw%#3z>jnfRP%9wWk_`z{s
zUvdo*bkA;byDr`|tZVBSBRAQyQI0G!6LK<cSe1xzjJ4Y0_#H2_sl4g?JME<$YBkC_
z@*1xqvVBmzUAXvEhk7I+73C^CqShvBaVJE}<TmIHZ{#BiV>^WXMTm(b%#0uIFrRMq
z;42#QkcZ_*op*0qSN}yXbPDyXpaW(;K$>3PGqp!pM<;!a;oYNUA2RO+^Hq&kB65&m
zR)-BW0e6A~!(z3K0K9mrrER8KgcHRqt)Bmyt%ucLd_48Uc&_*t0K-h~lX=p{;8rb3
zi-ajBoxU64_PT<O13p0^O<@~NUVS%Jb42J+_aJ)BpJ4&4=Se7l+P8U2uC63?&_+iJ
z_AN)Dap-i1AR32}FQjgeOBkC&R*NTX(SQ0K1gRwFp!*&?MWPDdj_GpEo$GFKmoF!|
z!hT!OT)`#Ir062Jg`$c)DHEMA6TT}|oay=<_D)y+z_tC8uEbb|eSp~p6&qD!Sck9%
z4|^U=%+?q88&{zj3nJ6^9&WjhZE-GlZT5^bq$w2)2^X2}fI{5f>IL=)SBg;E3EyVc
z%+#H~00u!B)RpWLx_v8xImL;Yl5!8>PE+;}|1NH3qrPj0pLaap_qp<RwguSNke=6f
zrB^oEO4{_&gmS`GDm=BlkW>@mHXkF@SF$y1wvxAmmRP83p0B<Kg!Zxg+={>YYn|8o
z9)~gg4AkA*2gx6X3R6)qZbkPX|Ngr8__9UPQRn;N-YyI%Cum6q4%0@Y5QHBA0&GRa
zq7JIQ7Jmy#ugFh%k%;vwr#pEe?FK_36FK?)c-%DQ44)NBSEIa%<nv=BwP=!9>p$gB
zmPzPn^C!|eSvC-KsGYHFQIfc^5Fie|hksQ`YrQ%g&6vU4@ZsaAt%~uunneWfGM(3%
zp`yv-vy(IP=h2~=C%_MmPF<qBcb0dSJ$t_Wk7ZJcP?fL{io~L~YDMbL7$_{O>2e;3
z1mcp^pzu*)OJsiW2PV@iS7}0ey}6anyQ1^++*fnmT7K9KkhaV6;eW_<BQW=Hjv2G9
zHyuK+ky%15?kJp7)~ottUojSzyBPB7IsLOdOJ|2{Szm(C$lWZC-3u|i?zyQ}8<X5r
zDa(YR@hc>BWf*1ur;H7b*7^<C)7gahlYUmXsT4Mp*eE+GlQOGcf31ipR>|ibckn2`
zX{wC>&hszv&&36e8pz>Zk(-Y(UB#wh?9<&oL*517{90~<e})9Zn_i1zAJ}CQ+$W>5
zDJ#bWWqql8K@~D3?rv@GZ~HW6%+#=Q|1VZ{^m2@;&5vX6Dma+(sPP{VvwzQ3apEMY
z`%7r^o@^k}W-_+;OWut7Qvk&O{h!%liHUw^@j-WAA4+C%-+y=V{$dt>jeh{VZVz?c
zL&ldweV&7hr8_KtW}nqdJV&&<X%xgSW}`H)i&01v#2@cI#1)7s>Y4Bh)A&OZ{LZeI
zLTAr0dNn3g0+|lAO)R{iG}qFO)C^X|S0rMDe)_GT549aH;^&LbSd0w6D}CtEtHkPB
zg8aB6;H}B&mOMr}1|9hwgysg6HnI=-_dFG%gW|<PjK5ENB~=jRuKFX9mh&s@mA<jD
z)<#2ggH9O*iC)@({AB@)$C5W5O=LatZ+uK)pJ|yViF79BiP(637F38O`som-xsg(D
ztcya%WWz`SzqY?y?Yu&ZkmkB9%P1P_L~x0GtA3N3%`u(SW(WEyAQIygvo_isXM@C>
zVJ~<rHeU`Dde#h*1C2OZR(VHkSnEu`?<HZl^R6b2Az`;bZT!(*^7zqUd+eLJ5+h*g
ztZE3kJ28k%gRl_;;nbk7m)G+Kp}$Uo!y=6@rY~(?$jXVkYA{I&@PKZH91XMv&0f{5
z{}#yMz-k=%ZAk+;b(JfqY&<Mu1Rb}<DVHBV3(}klkXXDOkfvL{r@hv`hD^c*j1C7B
z>U+36#u-cpsf;;IKIY1ZMAch$HCmIZ0sbyM#TR5z9Yb#x>}jw1#z`E3vif8v6Uv8#
zDA^$5BKF=tyN%&?B)luOIF$44<zVRL?gztcHO7gD!$#bhi6_y6Er`kK_}f~)yx1AK
z4}k@becRjLN}CEH9)Wt^*YkUTVI33su#~y*CMC;Y0aN{JU<8u7`^Ac)1wbW~d?5Ag
z6CBs?5Q$Q&C_b!sN@+c7#(<gXdJe~|t_*LzgtJXz0aSJZFn;P6@wHsYeTyeE8T6T9
zf<s{{aeG+2`~p2ha)gN;OHfEid;_*1{Zo~Gq5t#s=V}jAc|c7Oh>LhHEl_L6!D__L
zJQX$00T`wOu420R+ngFOxfA@{9OB^ht;v~5XC&<tzA9`)mc>CuabB@c2c{4lP9KZ0
zOjb%}F)CafAnSE3P-0DFr{K>L{_!01NV`*mBAC%@Sl{Bypto8aaXp&T-tHx}XK-ba
zU5xhZufEkQ<dC!CImd#eT3<ptXd2N|$P6)^qq79NnnPBSoUHXM{|k<TS~}40t#(B!
z;}CYc^`)-9s;ibo+~RB3mvM@x#H5S{?0<hY|MfXu7s^jB-c!!n8+*TlcAYQDcJh9|
zOPp-q`86dQ@CLuRbjQ&LF@ieFa>7;1UK$^dWh1`{_$Pwr8yGm`6$C}KOT;1-njKte
zvjK6^aPNvy8=}f+$PdBr&B!GwXBxEcb5$6p<#(eY-Wa&iBnVtJe%!lPI}vJt7O7ui
z?*|z%4KJj&)PCiLhK9QR+YsMj+zW8dqpHC;YHG7Vi}@gEa%{i;cj-3>(=b{lP@b!h
z97;%b0uUG!Wl0O;@;W1M3^j?1^)Te;FoA8S#!rG!i^v?K&}b+r?^Ww!NjY0_?LTf=
z<5(kE5S5TOil)*OW%$rMqU~i3i8Z5W0CNN)diYY6W%ScD(9L)v($Z^CDXKCGd|!Y(
z0%`>3&6eq;Yu@KoHY(0%^(&}1BRe{=CKASC)vr;Ry*&#@EkjLXS8D1U(hrgD3^f-9
z@f8;i6#@1a6Xo|?K;=x$g7Gm03Hzy@G%kM1Z2T4l3LVb$AwGK+fZn&QzU0Aw&nJ0j
zHHCf%oF&yzQSErM&cu)Y3YRFl4Iq3Zm6{=cUU81>3ZGl5aaNRxogo=aOG;gdEkf8U
zjpRFm?Wlx;)7OS{-w(5TE0q5*Iqh1qH;|9Apm(;)@Z1@(??(*oId3*D=b%Fr_xVPx
zbeMa!UD+0ln@`+Y&(zoyEelxSaDQ%V!#uJbF`}lp1eiMfK4Tova)>Rm6TJ^<G0|Ng
z_wD`Kd&<o<C}KGLdoTy3L&W4dAIM&=wnY`(7Kt4{P616zpKkZSh#qLVFrTnMCU8XS
zHQeXmonp_ChNWCWKjgrb$yHRX<2VX>fiVOp1_TxN_2j78lVVP9@y4|RwlE-Y1FE&q
zAF<va=e2%(fq1hOnQds@UB<<0ZAh67<m}?d*D=h$01{xdvPHXKyhB5xFQcQeHrH_C
z{3rLS0Iz>fcyf!Pg8~rjjW71eIG;btZm+aFa7)+;KTI5Kn&>@hV}Xy*cmLOxHjy5V
z`dJZ04&;o1*lA&4VPKrmtH`bMtD;~>r2R}seIG^=GMXSvGEAwU{Pr}3FtQhAc~Q=r
z*2>BW%q-9ZJMKe?#^+l$6$le)jA+2%ZD&<|&LWdb8dX8~X7iGrsm?q(atTyif2!5h
zuH5+0$Jril-J9>&;&rEdfAsHa@t}m%m?(7egoCy++K`YeqcK0i5Cq5Jdh6yn#|>=H
zEF(YzSlypFkvtVP%J49LkFnb;<hqVbGw>Ot$79O!dt6&S^nd!U`dMUhX@CgUTofm@
zJduB4mZG0j?=|yIEeo^vs)v2x%h(rX#t-PT@vBLzKb%Sf|NSueX87Yo+%)sQ1qZV?
zUBw8Iw)cLEelAmgwKvyXe{sS4Hg|mQfBmcTq#i_<mZVPSNdN~1PTvk}(C8g@fK37)
zWmiT-WAJEy-W=R|dEN4FyyprC@XEZHzJKv;on4j{*FJe0PeCX8Tp4JL4s$vlkFd}#
zgb8WpM&<KwZmxPRCu8$TjbaW>H`D^n5BxHEW&$NdbJ%{182S>c0J!Sz&4<KAZI`9`
zv$Q}>v{IPT#35Bt?9OQoyaNWPDy;dnC1AjqpQC)?H_*a$mA61x#MuT^va!xv`uDhg
zR-dB}yApnOUYY#_Y-x414k9d>s8R?~3D>mt68q3s9MGfpD4|@Y(2(BTd}27IuhkYm
zK5F)}=>8+v%N6Cz%P|-`(nL<GO;i8Mr*PfgfL76v03LScZ<?`}8eQ_Dq9n+Bg?-wk
z&{7_iMBoL{eA=EdVAD&V+vxBfUC(VVFKN@)M;XH>8V_@SCAz=*=errKu|C;lz|vhA
zWMJyh@91MxN|!3jGO{dkn3Ouc^<7*rtMGPP(-%niNr9z%PLI6f(?0bYA0`D>-1c$S
zcY1eylkU{rQHlzzbS}UztvQ=<Ua9l`goiJTv@*+GD@vYNaWA9|4uTws&&a4EzN!g@
z`PKvEPcIyLG(5JZJTjK9jms#Pz~@YoN{)%bz5l@>Q!RE>rG}J57j!tjtJ3x(OIK&|
z0$zXlUTjT3m7oPT(!$IcDuA?vZ(F!!tLw=Thugxil&#(upzYPnAT0Q?oCRv|O$_B*
z=$zSNHygC62?M6-s4aUfsn-%<Ur_`{pT`SN3vvq2<sd}K)I15JO=@BoAx)<NocDbu
zJFAa^5E&Wz*M04Lx0bKe`B(ZaCNoH}jw$BSKMF8ci9zaeCAxzG{cb-v{C7fpk*4?L
z4VbjHx+JYBzOl`jj)SuXq>q}wKqmh>=8T3B^@!ETfBp!)`3j*bKqMH8RcRxg?Ae3N
z4y>yP@(B6?P^zPq!gv)gHGL$I_GT(@>Qz{gc5@h@kpJ#N&4fH+P$xZj3a^Zi_II{`
zS=9|0`gbF^ch$e3@>dRIzQvXFCn$UPySt<u$JEwwK}SIuYKd`(xU&cs4yd>^-E<?2
zrkDRu;Tc3o$)tj2o*Ber`g)k0AH-{>6`V$Qt<a*8Pl@&1T)<idqzRi+jCKv*mlTUJ
zjgAV}zj_jeRR2AafCplBbG0=!b#k29<W}5E6{{jsW9(4)N&Y?iL?$Uo#ef2}P^(r!
zjJmCWOq3<>dnpd)7md|OxuU6wFyWzH-2`W@3Al}%rA>Sb^iu|=D@Tw)d@wfVh7T^s
z;CsOgu1vITq=_1xYw%r=^}9)D(P($*tLp57tb3Gu6e>m&qEZRI;dv5EGHR@RO!!ON
z+udRIB3+?6ug3pgX=QmCw<HVo@N=i@_TJRBk2}};?UgLn0hdu*3LDg{^cm}Io~oqq
zq>2C_yfe+bNX-X#Tqn|iJlMrd&6430_$55Xk3oY_uC&*;*&4;+V1UFu<U9@11;V$e
zQ8%EMb^?aVuD{MJ%>{-8X6a4>@NT4~dNE=fYfPtkJ4Lj-`S=dGO5_Rp@3e<}a9wC;
zgVOczWLTA`gjEjCAgkV>@-KS6bX$wVL=%ON^N*~LPL)(1j0fIwmX%dzc%yu&+|Z|?
zwNQCJ7+`=j2Zt=1?+iV=v>-AR@d1NAQ~IAb*Dmg$@PfgJEcX_6y*<;ONctUjKIv!t
z=R!&p&md9E+;Z}&R1~{-jWg-+%ETTj3yu$k=Y`F`A(AZbWdm6Qf%`<7M1{X55gw^H
z=(aSvq4sD!BiO%{0h#@)aYP?ed-YRW;WF7HoZqsRi>$E$0BYWRUUh*<ytn1RE|g_+
zmeU>b^iSgB!DpDDnQ2lGu&CwRNi&g<xafs1@i}@6UPOd!tbZ5O`eyn()mmTQ0&moe
zsF=JEwT4GDFA~0@WC_MRQK8Q<C(BRbu|~Kt5xHN1Tl~Iv!wq#5yd&mG>eZ?^4@9S>
z0<aczed6~WK_8-Zy=y|BJ#v|C3hrKh|C)FPWfF>aU}G(S`MqTQlFanLU|do!2B*Z0
zy6{v8ct1-~E-R*qVIlU4@G|=x!>i&%k!7-wG-k+WqC>J+0!4DE=c{(R5?-g`U0gkR
zwzNt2zur^+F5I-l6$>;pa(Z!G+^L}+;td-)*E{O<y$#|1%b}eE$lsgAm--c=GQmWU
z+u(oREiUpgn9X^82VF!Aa)mr-dKW#EeNZX+{7}|io;?dTUSLPp_v<wuZlMJjRh0L_
zd<HCkT;R{mGHxKCbMw+_f!B(oyq^rKMPW}dlQWZ8s_U!yY4J4Z@UxACM3rSDtq3Dt
zE>~A{9xrvXSNf2@pQzq6XfO{WDPvETNmA0N%}wO#x5;C8cUBbTCy^*xd%}O_pJ~?6
zj~!DfBH&8?v=I|Aoa?b<G4LfitiEr1+@NQwSBMq=;kQxG?<!se0;@TToD|)oO5XSr
z_@6kXQM4<lk0lkI(VGBSb+8FJzp{O$_LuwidCqRC56^jsa0i@ktkV5_{ZL-M`F~jP
z`Pt(O?~D~&>rBSC1t=I=)QC%tIzXI*!RCNert>wx<1OJox4+~XZoU5)ZfKtW?(k{d
ztY^I>)2A1En|M2Y{6=E35aZ&<!Q|Dw^}vsm;O)l=G2g-4xt^V&6zF)U;U3fD8?#g=
zqiik`3(Jl8)icW{V--JbML!o7@RAJ;=c9QByV%<k^N2HVhAulJJ(RRuTqFW^8LKNt
zXfP^XLVc{ODn2{h@rOZ^+unqlZOX!;(D6zXRr?>0RnC%h-Tu`{;*TnRe2(9Q^Rb@&
za7U5OkL&7eOr_sgF1tg|npTfWD!jZV_DgDxj&f|w%=G8`I`gu0F2h-7aR)MXaB+nk
zfxt=|8=H9~APR&94gacF2p7NSq9%`yOd5i8nm9EBHY!jhL_t`NKU5euqr4Gg#e4lY
z@wfA8I<otRI}(z?>6z~Cw<AwO4=jb1rfV$Q2UMO)P~?|SOd-x`)y+Lv7Rx<t-#}26
zmmy<aD_W3Z)Zm>8NN1o`2|(b?Bcgg-p}t)h8LGbhm})FE@$P*O=5bj*NrvvnWY^Dr
z0V5!8JSc=FI{fnn2^BPO&6?jG&+N^@8#=L&Q)X2)oS7@7$q>k4uj_*9xPsa;;!Q>7
z#U*2tgAp!|uy<$m4M7J2&dYq8uC3%iY1mFT1<k2_*B&?P-oD=8tC~Mq!ZszYR{cX*
z(t7L$+%QY>c83H9h@V91nb+c9^iLg)qx5w`%e>r&NXyq?ODY<T)wi}*R*j=3?y_}y
z!{gsVstxASi#DWx7M%z8)UG>F{L{Q0Fb0qFRf(fw9sX*<g9{6gBOH3v{6polh8SU*
zpY%Hyi~-fCu~mn}HKZ0)?>}|c?IQ(3{880S9l}LN6{tECm9iS<88OAzx?k)P2E9fL
zOYr&qr)po4sbVRa4?jT)-s8gkY%UJh<|}I*UrI<rti%4W;MFbkR&b+Ba3}B^<5fOz
zW}J7`*QY2mP7v-?F|n@-pY91?hl|D~LU@Sn#0}%Dp~1{o=x&K^$mB?uiSpM|Y}Bu~
zI>Dp-3#SG<=rv)mITEoJCOHI<+e)}r-TTZ)vq1(wR%jOKnzY54wk6o4k!lnThq~9?
zx<y7ChC7<HNA$Z)QlqWhV4YZPMVtC?ws6ow@3u(d8;3JmD*bQUV}t|_kT5l1dJ<6J
zd;fhC<4R;?#%N^;q%P!Wq(KvK6&+yVP397I85-?qB*6n!yX5BTo7-f8w!BvtL2wAm
zc3(dz9t@5f9z5Lyf*1d7iFw&!^|{Ey^0XbbF@!=v3uYx)YQ4NfM=aLZt@S+Kx{!X%
z<`?M{CxaL!X$zEVuc6C8Em6f<l)#eky57(zO|1(gQ0Rx008%nTRj`9$b_SlRzn7D)
zIoH^T+evHEnGIlX7dCVTA9YRo00ExMw84z5x}6|J>V%F5(`u!MEYAy99#2avEvqt`
zM2Mcg`lyP^AWl`KW-2$eFxL7BrcP~{=P<QLq3BzhH88RqsKT5stq%CHW7+hz%D>eT
z^LB*yn{e6-plQR)q{U$qvOATd>j9rMvB8F-$Rrs)TJR4BM%0Bg(04O|m)gMMwN=Aw
zT_xS&3NG`Vx{H}xsjS#4v~1M1%+tc$I8UN~Ef4c${L}`)xza!BJBT<ey1>-45SKg9
z#@74f%LceD=rL%YzxJ!7jb^<jcwZ)f<R-jG1kReOJ5E47FlgDN@~2KGVVCSC@-Xt^
ztMgt%o!dBNEU}<6WvYqu29bLxIzmi(eTd42#52SH`mW0vrG#Z++yYCLNCfc_->$4<
zP-2Jp*6sMdp%grILAIH$T29@v*Z+pn&Pj_+zWq{%0vQ*ogjS{5^QzSTFjlf+UO1<Q
zaqSCV(&2ahIB6rnT>;!APqW-nrVKH6X!h*G%RVD%?TIr%OEhn?_v^RxP{RD+CV}1&
zVgKSVIO8{O-dSTuctb0p-*}eQ7%EcsqI}OrjBqb!DBSpz=cbRwAgkO1*>I3X>?gA4
zx1!=P=Tw>Qsl*7ZYYX;D>Mh}S@fYQG+OEO}f?U=$1e55JmM-r)>J4Rj*-fh6<$JTX
zKm-|g2z<=2)&hne3+l<P4ST}h3-zP(?(Sh-hP(iX0&#*;qFaRp91+MkS_K=gGtRj_
zy9#_v5>F-yjO{D7OPNdrBY!h1y3Gz#x~-S)M>lD*%TR2W8(`6Ir9*Q19r<QSqq7LT
z7^60%X592h=qp8}Y5njuT<0e9F{=@TFN9lB&Y^%6G8PsG>^LCwu#9MZ{}q#mK_Q>Y
zFWluP78uvcQjuyeq^=S}Ub)ynMTcjxlSI_k^maW)av6evTiu^)1XP{IS~;RhZ{lSO
zet*seTv?)_;$pVAnq=rg%DH4-54egK2ll_C`vx;4TUV5N6-qXy{jG0(qK|j`jqpWI
z*Dt{X5bpU@yV5l@;0NuOROW-*zwK7gtB5WRnAC^D=kL-Tist`~EG>CDZZ@??=lpQA
z+m?s$VEp&@dX|ejsdj!1*7eI08-ngQW!g@kp)ZzNOt}E-hfS|Gb`GfH|65DSck!zU
z-ZuKR^SeT@_&s?TfdI3lxST{{hwFP=3Fb|JpJW`K<aKqwM(i#fy-ml26X!KSvRhpy
z6(gB{e!Pi<80%uBJNdeo@|mwXWw#_3G-BY+IaQNxFr}z1VU<L4UqJz=<6+uY>WnWg
zV|r_4gum-@X`@bn2=QhP4{Gz`P3*HdR}G!BZK%x7-C)zU(fm#aLr|kOuQs_I?Sc>b
zHUmuuhPUa#2VbhfWR~ihJ(QgJi1$=<?dK{U%GWDmN%NJ$;T?Crx9%tl0Au_&^pDwD
zA$yUfic#+~gn`)7Ov;3}5-YDyBV8PPk6!d!r_Gb|GMVii9cDm|utUHX=cX{7<u$QA
ziE)Dhg|DHJ1MaB-W&SMatQ2MkV!Akdl!7K|=uoF(4|A;$_lZtXA?@q4DH{@!9ZIsb
z0yB<2a*h@~aCKdOD=57FVsHjf_d2)#z3iX7+-7{DLRO?sb7$7^U&;XrQo{B(wxKm6
zYfAH-?dAod7<$P)dF$7J0Q?a%;_%<Nf7oB%c(AaflVDy<uW(w7A8H*#_vpXn3x~bs
zffb9;h|H54yMwaCoEJ2Jx*QkKV=+rB(_`SS2b0};2dao~iJ&knUUrzf?O>R=Idq@T
zKUBPC4f%X3rU7+c5K=7zVzREunGwS}I`T%+bPyWSVq(W)%2+6#hjx4A_c(pCXfl@U
zufA80%{08M8ML+x_I+j$o&7_iV)r_g#oij~A*pF{?W+N+rD>VuyEDJg2Wzu<W(I=V
zVQ<YjVlHOWrXPBEin~ti{IziG*bz1!XumUOC<ih%WjgUF+u9bDo;Y>w#hX-i)GWh^
zjki#0koH_y6vD#eW{p=ZP)Ng#BUyPb%9#H&#;U$936uZG38ve7dhiT?@IBNfGWdL&
zp!hSQee@%_{C@r5TWTE_I3c^F@M1Jya?>Dj4Ym(5u%d3I=lU(C?1h5$q-)qOz<8Y)
z0F}4q7=~R@*#)%;EugH#GTIKe#DN>Nt5ID`%4kgZoogY(k=YwSDo6(Y<brY6&NQ-t
z*o25_XBs*Pi-~3j7=Ir~Wuw_2#IQxok5WRhgQ22M{5z;9|Gs#MKhIElS-YERdE{)i
zlBP0XYV_-zy^b1JZ)lcHYO*pLDEh6Bktv3}uU5L->S3{jjMe?yN<3og@z;$9!cX^$
z-Rd-tC$iTJS7g|_fq4`P#QwHkE^nv8s*)3XL<z{WBO|ZfPEVRw(EYozPc}CHUY3N*
zRHVk})-&8l)nSbT(}2-@e+rUWdMgAb{=SFJb7DH3p*@-!wU;9$2z`iB;FmwWxoNKE
zAGfjRRo6sG<`PESA-@)FRKAVSBiV|2LnLHNsaV(ESzR|nsP_3{5H%I`vf^b$-p4?H
zo%{KvC`#P8Qp7Rr%fEr&vh$p#29P&IHM#N&O-XDRD>GpNE|vLB(3`W9U`@%gx7pex
zULLP-%!{t=gba-(vILU*vV=LUcUZO<w?4WU48{nVG}oh+U|j;P%09_VM}(kQz8(=+
zOV_Q!e~qS#Em&b&3&~<lM{pQb>i*EKPC#N5WPSnT9Z`SlvFI3S2;v02e8f`cDGAc`
zwf}zU7mF_<%>p>GyW0A)h<q2=)r4CcO>p$DnOiDuHk!T{{m33$BfAIJp;X4K%GHW8
zjVEkmTYHBb%rXQfYvv!avo|M*#s-iRUaq^5WY09wzg_$qr<2FmoaL+Psw2@zSOVh!
zrwMV=kh}7QJY=fhhzd-9ub^h=#(FQaQv7;kB;qjZ&B9BL6>~$@&_gyVwBp12>2pqG
zc};7(F=_4`G$swK>&k3FJV3=g_z^Q)_=$_D*I(MK$7RGVbO8&Tb@^T3rVOucOmuDY
z&7UY|bpKs!-ut(){1D{HPx`$mE947ZZ;^c_G{YT3xOAC{G)dzU?r{*5ds#(6U`{hf
zVZj>g%#7;^cGvV>2|UkrPfM8j?86ux%b>jt`nV2#;A~_F>t)zyt3wHY9t#61zYI)_
zOjeCwMOUaXY-gUd*;9l#!(Yx(l6d^GoVFMW$i2{DZ$6#|iCS2)IZ}haD@LG9d4;lB
zv$W(8O9V6j5~@8H{TfV~wcxc+u@QIvG9|r%|B(#=2X)NLh&=5pN`80=YVi0?sK2lN
zI3*(;qrfETVf?PKi%&e%1c_>mt>mG3ii|jac<Dtmp;$Hl?_3N7Fw=iL(&KzVOG)L&
z20k4bvlulVb0ffeC@09~H&QLDv}qq&0KixWk9UQR!CSq?^+{IA>-ammLts~AtAGM9
zLMSgBnNS={fsB>!Z8nNox>?dF7M8@<R|BM?yguf0rf~+>KDP7y$DXFtMYw?@2>8p(
zJO7&sX|30gqN;Gk4MP}-JmEHOZ<<mthVCYh5)2p?MgPWY;H30SJPw0D*4mWJ-`ujK
zJFR-;Oh|f><h!QF;}yC;HZry=zI#DnnzB30MeRu=!0&Pj%k-M(YYwyh(=Nx1uV?Ht
z$P}jH;S?z8f7XkBpx6D_l$py^s{#M76z#Y8|96&dwFdOY-kV=320HJV`F)}5>8`!r
zA8WYW&cM*Rzl^Oov$+1}!w{ZF__#GD`}E7BpZWLox8>AGb;yKov1&eR*vx2ae3#;K
z9!BX!z(+FiKyyk)Ut7(#hGb7gJgO(5wtrjGq4+k`M&ciCe%;E4d}WYFcJzw@hj<$6
z>wl2+I-38)TKqwZClm`f@gXE%hC8K66FZNsg)8g<Zxm0O!;dzdRrb_Na%olN=44&X
z-NQ9DQi0oN=N^2pP@Pl0-b;^k1dF?Ks*)vP?}74l#|I%5zO7k+@k>Xc<3**tb}T3O
zM$i94-T)1i2(z%B!$o*<L>X}IL-LCihXPs=`CwyrRO9`L9UFfxV@XG>@6H^QE0DX8
zj$T25!~`UmQ@GW=)u*SZgjeTz>7X#%ikzMtuc?Tx-Z<E7eI76NsC}^tkQ)1lI2gBB
zmry#GX9%?)mv2&X2G5~+nHovfiAu_$|BA!M{E-;R8R~<KWhA~Fg4*T##U|2_X3c87
z<J`s^%;zAz92tbuQZUu09r4+z+$dP_R8ewJgdfn(LfA{j8GU$Q8aQ-nvCOS!JYAfU
z*QbK_v9j<w<=**T&ad?zGn}~na>()Yj3xZ~6?Z>9)`qWDqoC|^IB@9eoLZ=3ON8uQ
z%s@6?H!0`udEM0T*w{+lIXX?`ZkTFswo?E#h$K}KzR{IeJT_7S9k~>dRt6vS5uB*o
z_($2bsaM(tGd)VewLC5xCbW|!W-X7+F~|y1EB<V%&;;HV-aiVDkK3w6<%~?fMFlbq
z;@MFTA}#QNK6_W~3n3jo7gS5^bd(aed#U=ENa=mcC&2b(Z~W7IWb7M)-l8BIMb<;F
znQuQ*-|ml@i7}+$*e3IHO6|cH9p1g}2GML1P`i4P5{?SjRoQ`A$;;u&yVRLlqCw&t
z&QJ`S&=54&3S@~#nql=@?_#8h%s^o<v>l_;qKlr7adi=lID1sV5-cAyzl2@zhs`}a
zv`a{0V~GvojU&+nQ9&r3)Ef&7Hv#PBiiYlGdN%?m@Fo(ljq3!(lX8274_dXE(Tre6
zB`n|_*Gh;v_KjYvy|H_)p3;;k`+7p0A^=|Z!js@(awk{`_t{~Jkt$=qjS$fNI;7ar
znWI5L0Wxy=zF4)`ByL)zjvT3RMRCVxQ#p@p=*EppPBX5c!#z;_wo%0^5=-v0P!9S5
z>A&aGgPp`$%#$8@i<AxST<V*DdW5RAJ>oY_hgI|i$bUS9UdA9x{#5^I41{cT`CrB_
zhS@O>Qq4@0ER1eRfU`(3Q4M9yFm$I{3h{c||NO~qbf?K^co@VY+Do4W>;k^mz58rr
z(0G9j?Zd>jt0PM+lJ>)jvNzvSg{9n68obKeO87gGYv_R!&HMMime*;=!_L*=&E^M+
zoSZ?tUxITDPNP3nx9ta-n%seh_;U<NV8Wnk`;7FTg6o~G%rBpqhlBfTy9n)oJns)~
z<Bi0<J{qn=FWqlugfYR&WM*){)(U}VR}Tu*(I%ki1eB-cxm&TXCCExyfr|`;WNMRx
zLHwboQJ*y5!w!&c%BE5sxh5(ta21m4xb7zu2nNOcqp{44+N3MVM0sjp{2xuQx~;+%
zr;LE9a)$NryMwqhT%e&XN{5zpIb=f~Q9vq{$WPfS`h`;VvOwRyhIXlmQ`fX8&@7m*
zJ}4wAR=p5Egfr9*iX7pBCWe%pG&{98JQIY2DP%e>hVrEFz5g*hjVaAM!m{R)T#t>g
zWHqU-#yS2?V)M33GS4obh)vek009$-u6V5uuJK1f=|gwc^UN^o^of~}0E+e&W~W8y
zD_1s5HX^qv0Q=C@8`ScNx3dj!t$4GwG$ctQG9(Klm4wN!a%L9K)-m2k!<V?16#QZ1
z%)^+~*t+7C^A%}zTOSu(v$sk;Qm~$~EuQ%jXpsq_sFJtBf~ZsIc}_YsUL?Ce$uCKq
z<K>|bpo5^~2Lq|TBwEqXdR)&sh7Nq?rO$;BoLIMeMlQdxTR)C{y2*Xn`?A^&1Wvf@
z6vw<>)fN9*IM5ytlCYD2pP8u}g(t0lJdPa_8pUjl@Ds)&v}Ccst>m1@3biVLBA2qy
zYKH}yAJ<`^iox4zYP}=WuFmw%lCKC|<+R*v-TlQKj7+t4q%EvXEXY(IG{<~q0Z%C3
z4^+B|=4ivWoT8nnx@qusz<UF=@1c5wi{$x?Q2$6ek^gga1oO1GAY=@d*YN2gl44)Q
zPI)I+Mx*Y&239Wxqc+gNnInL>%e#8YRBD+_&G)F%OjnK8WmF3vD~(mCcyyc3c%A3B
zHT%~@13#gm@C?Zf2X|&$Yj!Q0xJR#ZE2)adUZgYKdK^aI*(@|pX>L@k<ogBL_}z5g
z%nXcX2=K%vf3yhYVX_TIiX%uV<57+13HMdq|AsnI7)FVy3!1jsri78<T9SQ*ioT-d
zoj$Ae#S1uG;Uq{qy%i<0+>yX7?CbOOQT%P;GU1<598gljjjdY5UeHk~8%rL=<6gkd
z_wTB+rhVPG&HpIfbM5ofRoi7rb};$>h0x3Yxh(^qzveH^BZ{XR8oNS`^Ap6#H^Jg>
zZ|I?25)@wW4#|nV<bAPxx<d`dz|O#m{Z`AsaV?q<{EnHd2i%=g=dGLBodYrIXdfS@
zgT696lavYnN*dL*Niwg8=?m)Z#C*w%W3#J<(gD{Ex<grwd}?Z?Q;Hq0Ly3+^uSIkl
zYL2I9=f~$zKXz_^F^G^Ln{!FyqacHO3foH=??Q!iW^hnBI|ociO40IGIa-BIj2wG>
zlWsUD|D*H-{xg1F5)FU>pT#^noOYJP)PF6SE-F(<W03PCG_<d=p-)8Rsj;IYn+?mk
zhA;-}Rv6{?XOAlqOpk*fjU9O;B#Zr<lUZDHJg+w$NM)51Aq_#Jtfxgo0w!GRZQcxK
z^Qf0<>muFO|4b)C&Psl#LH3$kep&J7;FPSZ<vRAs@S6f8M{r$H6hRq1Dn+!?roV7>
zw?#Orv`3vyh7vA1pK?D{7JWh-W2Z_hmV;PUep5RBVs{&L#x+pt*8AAn_@x9{#y>Um
zI7I;@3>03fPIYB|W#XaBFefYZbA{2%z(a;1NGAC8aUEy7?@a;%y*oA*zDBuMT&-U)
zvg~lyJ-=dne>FS#v2y_?^NLgkKuu7f(Aqa=Z~vl6^RKE;F&!XRW?Toa4|vH;4^EuM
ztp+?Sl-fEmbFkNz+-<vTzf7diP2cPu4CL(8jEF9afUPj)%^3tLZo;sB%vr&ltyPkI
z{!GEE9{u!Ts~-ktinbEu7kd7ndV(9gOQfB%8(>Yd)mRd;wac15%fQGBDHMyFxP}(7
z_e%bVFP&XE2Gz(2B+)5-?X3%yrz*GQH5mY=G<IL`bzkZrvaVBIdE&)7-yp{VRg0yZ
z_v-AF>~t^^_<e3!{2$zr(male-58Y-7Zh2KW5a2(m>$Aj++g~a<Q$;<x_)hoYL!G8
z-a%yf>YCQ(af!my6qqPHoUkP*jcRWVLGLzXr{&&+q&!k_IMY~U>+a6s0821lKNE(h
zxUl&3^5Rzx=pJcWxG_^5|9M$B|9=3eKv%zVjve=%W?k)U$BuoigJ~y<y}S4BrkmG3
z^DcYf0}iA7qfS^-cG}(4ZMVG6X16})y%0NX)v2X5vhv0?7fobd1wZLSXv(4LvA4IC
zPBiWoL@RI|M}N}ySXp3&gAxEbmYL340Wqg>g_}a~pUti)dMgVn>f^l1!n&g^Jed|O
z5GI3*SOSjD?lQa(W>Y?bdzGU;*xR{p3y_7y#kBnYXYW6o8_lvjPw<c+6rl)e(naWu
zjL6CoU0I^5yQg<&Y|Ped_sx8x{wnio->t1}@3i%_*y`>QsYPVy2oIOkBq*Rk66F4W
z{|mU&J={G!GDJpJ;krKr@QC|daqc<i9=UWaW@c;lK0uLms`YD6Hyg@=E7fO>HF&b3
z_7*jt6bt32vQeE|j%<N2ti0!3^L0Od{No?T?YsB<KKK)_MVImC*HKbkrV-2zpZd>u
zL-8T8;8b#1y?kRWUVHUr3yC*2Htki4K%71o5iIM#ixPzw2O%!!gsW>hh=2orNFFUc
zjdxFYT5Bwq)wNmS2bA#jsLe0w-7^k7NPcw;Etpr}25%P>*W`OdxnpIg#uvo|dF4cP
zV(t3%c<qhv$NG(zh0~X;Y+@~RgwF|gcGTEc8@dj|xwy2Z>*cWWR%;a)uGDH)Rsq91
z+dHwfwPk!+5dEvorJR%~hmb3fZ?JK25PLg2mh(Wb2;pi)h2RQh2J{bkIr5KWWUamn
z-=OfpBL=vj{ahpT?C=7mQWD_LOy{&dG)}@NCE>eLZAE<+-V#89H=z;Wkz^FWT%q(L
z;?SDz7i<f!(^ha91|#4#E1+hQw;JmbItY$K5i>m3nqkeM&;kzeEFHl*;XpA;;C;+F
zN;eF_!P(F}3f`x9k^w__KaqzM&m_ErHHC&_UDTaJzXv=CJt&ms+y)eow8uH)EWaTn
zF3NbkZ8M%nXi=U}h2m$<)F;-s->bgDD^LfXr_dd6*i2sZRWu5v3Lc~2v%c0LG#roN
zQhCns23(@-icp>8I_nUwlN2YXrmKYXwNVTT8qF_CFME%vFBcj&axkJW0mhR@E`}2N
zE_{UCD|Maq#e!lC(lzJ@Fp9wj=b&+fzyt^5{l`(RF6q7Eb@YC`?g6N2!PnFj%6Sx+
z1LH81c#H$N)(D9U{UYyS;!&U}YM<N4yn)ox{MoQXI2Z+H3Y*EA)|%6EjJM%ka}JKg
zV-uvn_)QCrP;B9y2VS*2Pj?#^&lEk!c!%+-)ZU@TA6PGFUW(<qmAM!?O{>1Nu1<B|
zOyo2sD1nm_O#KC31;6bk^g;ZK@CG<)TKCTAeKVR9Mzt;8g@R(tUgE%r;2gylur#B2
zga5D$Q}tq8M2^54H?E04>c;(z`;ZvAsKcwr|EMp`9|vQ|r(fmr=-Q*>pK^Ixf1h=X
ziy(fY_?nbS`8o>AYJQy@Un|~pB_Tn&2KI7j10iv|j6k3uH5dsLV-PAfuMqBJtqG&d
zIT2YQXEs|stkzl`|7NqxLdN0+B1|11<eZqXf`~JpFh01b6YCvUTbL%|$DU~ea@OzL
zI*BB)PSmRZ-{k`1Mt-lGFTHF_EQB1yR346q2K^r^wo)FgwCNKGGy~)NM`KO+448a>
z>M&z7QfNVB`0j9r`jNsHq~9p8%;1RtT3M#L^vCyco_0`R1XFwT+$qW~2w!5#AdnDh
z;KN>h5P!i5_fQgo)^zZ!>Lf43M<09`@4ov%G>%TvJvzSQ(={oRGAWbtO_y&^@%86_
z{_nuMk00P_HXi!^ltX{{^YDX{NvZYdo+l6PrQW9&UmxB0j05_i7WwlVJyMH)FGa8a
z-1}(}=bk)%=c0!n)$s(i^~q7~{(xW8SN(a~F}}#SJ-v)?;<jhqGb*FHzHAwf&sQma
ze)R5#&G;QEP0yzBaDPX2dYt{0^XVRc*ZIu!Fe>G&6<_5_@}hyCLWq`lr(`%pv1RV`
zjCm~-Y4FGhEf82?X^S!f{uu#8rCN^p`NgQuFGXFvGD<b_dBD5fxN$RHfBkzgUte@x
zC~KVfzJsOtxrOTpg(}r~D?d<b=(Z>p^R_~!JA>1{6%Gha5GcX(9Ubn+&en$8hF2BJ
zH1p+Wrxv^}kdLTZcUw{LA*3c8n=Quavv^BJW$b<8PxYC8A>4i#2}1m%g`54Z1@L%e
zImc3UF32&1Fgz*zayF>K>j|k?p@5eSUMmP8FXQCo#GVnQ*?i2Q1XQ~aLZx+qq6MK}
z3Mqv^Gx5S|pZ6|(jlrqz!MZu|-ea#6!adSf3)K)7S=oYxGks1W-REixu|Q}rweumE
zs9X$NR<1C&vqc2E2$FJfX`vA9V*c|MwoR#hTMs_-7*^*HeAf*NtXg>aez&i^7CBwY
zdJjTz1pd5?d^_9G*x&NJGWYNP_8oist*l;CT}x^|?IyfY-(!qdkCl)*Qckmr*x%cU
zfBeV)5gQNgsNp*2y@B?32nDeq&(Epbj5YK@aC(VQx2^XzQwTz}4G)Hyd`0&V1`A;(
z1$f8PU?#Gf4!nN0w>LcRwYj;-sBH-Sw8gN6JIWoxrjuU;Wekx&Xg6V@%4O{n3CpB5
zC9x2NabzK4sE2?ttIK#7)V$Z6x;WDwvKjqeQ?M_H%xI;acjLIRCHTG<ckkY|7m7Uq
zv>q=mr!+o<B5cM<6L`c+DTRO<<52cPUMppkYuV^_`kI52=(J9P5Quww+xA+->w&y8
z#R6g1y3y2}5SorPgy$84QVbjMl(Yv;ZbooIh_2*4hv3_kQe%!1k9^sLD<#hg3I)8*
z2op&@xt_)guTL#l{`;V@tnHoMO|@yu>joTRg}30?`$w%;6bT5#F?c1MXkG{NLhc_t
zGP1c*E}=qQzssKgUYGnbM2HbSVSe@L&|jgw<Pd>4FrE~=C@4|FoE&b(gZrPwx!|on
zzoP3ID=KI=`^`$VqIITu(>qwFjs2aNLAg?0Q2lws0m6OY!1<oWIpp*Xl%(y>iC}9t
zT5XgEjE#q2G;3d18P_%bq@U0N^;73q|0wI~wMxAD^0j#B#x*OA%H^`~MAdkgwMAYk
za285GJU#(qJp90SC=yYGRVy_;Q`3B-c+nWLhBWU?Ew~@$2g)ofIRszeYLsVqNZFfD
zHx9v<8XuGuc;y09cn%RtmIyn9HJqDY(42#7)i&Wql&^+I&3#Ar?(c8wT^LHtsU5lK
z2(FtgLR4PZgN;ygR)h%XP_&WntD^p91<#C<<rZ4+9F~LV6Z$}#Xe;;{g9G}Ak_cgb
zp@71ld|oJ_H|<G;CmJ4dC=0>KcvWLC;T&OBgd|xLh6JVBl3;Z~@T2u&58rg)xr)ck
z!s4=(UeFP8G!pKW{y@j57ll5Pg3^XuQ8qNtZz~#vXu$#Gs3W)HG+|q<=uPl{Iv}T7
zSNA~w@cJcvh3Kf(_B8lOa{!=RoPl!$H~NiXfNIoOof$u)I72CYen##!j9>b!B((<|
zOM42MqZ5x8G~G{ww{?$k#|RH#6jD?_F>Ii2WJ;`8D^7${$j?Q%;GsUbm$f&Y0Vj(-
z&xrQu-FTZ>aavsr@&Ez@z&Uua-)+TFV^1)J@khbRQ<S&RFxD+zf98D%Z<xgx0*^QH
zayCpF(oR5!ak@U8+Yj%?k@C-oOhwNz9$d3bYBYx6dn<QE7fZqm@Feg$dFZEb#!xiM
zIl2t{cx~~{Gmla7N~Ip*5XRa{N@yGRq<Klk73!?lclv%Q+K)H7y<XK$cqaxv$<Xjn
z)14Y4yvH@m8Yh%*8g-2~d?Jcz43ALE!=srGL@B6xq5tp{UJGh>tDW+w&Wd-N$!Yyz
zK%(;~RQaS%jB@Nb2tC(%@FZ$@lutG`xx}yu9*#ER^;OWl1uNh%`k1=JNJ4O_`DTo9
zI!Jj&QI-k+3+8}F^7#>l7`nH-a$RdKYlHU`;VIE6#(@5Fc-(x9%hm6b@<m0^H&J{|
z%A|Z91){xlZ8es(shbf~4nYLTvJq#a3I+%H4@4l@48oXu^Q#RwNET+wp$vrCVD_*!
zp@4|n%~tfa`6R~|;SfRWJPWeK<0uh8F{xr^v9Kr#fi8&YsxQpD9WYRAu9H>7X$(3k
zWa8MO76u9e^Ia{3i|A{U`uZDhnxTZbV<T<FAxxr(Am6Rf5)%SJM%iVrD~+FXB<O$`
zLinwS(kBW*7^hR6hp16Ua`P6U*4W^^<#UX)7+t<Ichtpu?cL>1y=z2_lJ%%IDf-Ub
zz=a?<0S;iK2<PH(m)x(pPN5SuePgPoI?vp0HYf-2%U}H}KKbO+=xftBDU&iOlQJnT
z-<sm<udaA}J@UxH$}fG6uYLX)RiIDbD?0G>`yL(o^VGxFQ&rDD^yiyDo_g||Js$Qk
z{eE;QjZ6A5zR4HHA5TAc=9z^XM3;5r=l@-q!!grDWzRVL%wyjl9jS~ij>^^RR}bE#
zZ|P3`{@n45!j*JOy6>@i(=Cq`-~Ytn!t>+8cRx=(4}U)O_<Z5yqvPr=57*YEMtLNJ
zK>#=$l4FUmeDg6|o((LM;iC{p_4^2%lE4g}7)BOF2G1s=BC}K~V{fo&3l7IZJ3^`<
zBusgh2={anco@C<%FD67zHUqYluwAzItWhkBS9wum#H9<1HumkX~#!<wgjdR@OlVd
z2sPI4A;7t`66#=Ym&U}>@@kZ;bqi`z=q2)=6fFp6-@w2yHVC6EaGOGzLW2K(;!njR
zCgn#G5Yu0TLU>5wErKAn)9nOd+;(@j^;|FN_4!y`y&m&Rgx0FtgNKk)2v?|yJ|R$K
z>!GiD-<cJC@TiLmdS+o=7v^;5SP-8=tRW~uuuCW(ggS+MIdZdA^$o>Z)<P#j!T}H~
zYu93J{btO~Em}dyn2=M9T*e5q_(Z5g7{Qg*n~^Ue$VUK*xB>yMt`A7+KNZzVMRh0P
zCA>JRWCrOPhuery5Cs?SNEX=nym)W5ll=jq_7GSh<Stf5-c*#!UXb^w*FN+ffxISh
znTe+MVdT#NF35L-kd8S=NNKMg&1Fva4g{0?`#Z7m;C6iQ!8`GjAOAz_@9jsWG#lh@
zLz&?`Q~BALLg8fLF7T?oVHZ!S-5?Ytw1xS^Iy{?0n3#l^Y2F1>nuEB&cu{-8DWZS$
z+!s7srn4yV5VQ&|rqVtM<=yeofyOv_CejYzm%Ryf&de}&X`EC2>^b_n)YH55PKU`F
zD?ZNz>*TpRimcWFbB6~OIcT=GHe!2c$LkY?0g4#%6rqs7FcIN8>l=%E)+TF={*!;2
zy)J@7Ji?9+@El{_1eY2Egt(Ww508VT<&~(_>gspa@XPGmJ4Jgd1dMoN;kiXPy7l!N
zF;|<{wXEPo)m>6w)&gmVsm$S-koWpSDU63EifVGH4NkRBWdG<rVmEq%WjtF@z8xMO
zXkRu6^6c2K6+AFZVGU7~(+>8B4sofvS)<<HWALc;j6l@G?fFNckB3Uj%1Lky;&$eB
zDzr~MBiak$vM_Q5KL7_~9EzvWo|Qc)b?{O{LCE0Zg@)kT%2mxB;||=g_Ok_@S3AxH
z<KP**KEbyrJ(6G=uf<X<EL3B2YbS2qx@WxOy(;hK<13d@T~8F?!p6g%S@zjFp_7$z
zKCZ7V#L{9dmX_)<KToK@rKry*MLMtod=o9D+v!EK)sCa%lwXOka#m;vClOw7?fMO^
z6_isb5kj!nw4xKEV{joMM^R>BfPiA0vBcnkrr5aF$C=(k&h%X?vQShY(?P+OyekpL
zPkU|PF$%7N)*x@#+uzgtlQ#)v6S9S#jRWkVC;ZUX8o)~u`clz4f?naJd7$SG8^TS4
zQ!6Jb!d>NRO?73gn74OWliZ0H#^K;tyd*JxsL#&_-a_C#WE>a=I42>YW`%3W1rK~P
z=HORo5nf@)QSgdFQ8Ny$M|jq(=qRD?z*D8#Ld@2d>^-B$J!kZZJb&O{6v-$v=ucT~
zPVk3u!iC4$axd0~@UwYk;~NfSIT(PneclRWjTy=iby4pw#S}`1X{{qoB+4<u>^aYt
z-bK4n*b-iqb=x6N)(N=lSnG=S3expn=qC%Zr};#tgTi`R_;IN9fl|`S7j+C+0oG>J
zXYy*PrJA>)<wDb!@FU1?gv$g=6mYF3Lv5fRDNJ9lccOKM0!{Ql_{TU^M+$8l<~MeB
zH{$fX=Un~Og*A=!yvD2=GlC6r?@g&aC^@I8A9^Pm4F7>_5xjzzsYC5)y-$fwK*zy5
zC|!nXD<1ABaqv9stIvcxY_$$;bitUz2PB$~cN>NpDD^a3qWSG{SWMOL{9WKFaK{Wv
zIN$@=YewsMSGZ|E2EDe%R6K+3gD0568_jft16r+@Y3vwW3}Nh5Hbe%eD{3>|(cn^t
z78Jjf6|PM3JLV34PoLlm(FpnsPl90-#snx)QS7m1%})Wh;^V;mi6<1_q4@&F)dO8G
zm>w921qaS4rmqfHt9lq)!{JDwj`0A^$1|C614fXuB9~<SJ5QP02%netT8F2E&(wf;
z9ori@qrQnBsL#zCM{RFo<dH%;+7O>nGTy&h_^p03SH{bqFXpNL8600!#`ghn6UEo0
zOv=|$wC?IFczi88%pfsp*?gu=xKN;w9n5o5x(tnkVN4*PtXYUAOBZDcjHS;Db+XCr
z3Hi0LYCAbHlEexN#@Nc7)3m9<$_)k%kFWXy?(<7#mc|yPgmPfhjHg#$3?YQrgaYaY
zidG#E^zUM^2i4hmy#2#Jj_WsWh-is$tIZG)!pT5jP+nmjbt+<vg)d%Nj1^2AN+dS9
z5VpP;A{1cE$M7sE&$u6jmNzXLQ#Lcq9d-2d-J5C+6YXHLCfcjvOYdjUAW6&>1!hk}
z>B4woaaSm6;{&l46Nb<w5y7$GG(E{N{!LriDDLg=$1i^I%eZ$R3t?9|DU&iOlQJpa
zs^aUB#}_3r(0sCyOgQ`VN4oFIL;m`OhyFb6c>2wcALFmzX1VhEZ}doiQc35=|HkF>
z<9qe4@xO;xHK<p9J#p}NRGyrOhX>an%Kq}~gO72!dWqui`g8Sof#S!WIL3dUS3dWQ
z-uJnuo@~$Po~y^h;uHGwtmDfR-H!k%{nJnWiAPF&(M8Nvv7%oT50$KY5%?uRXS$za
zrHObu&LNON5QsvFyjc83NI*D2=XJG#)R}6+I|;)xynotO6cLUP%Xow@2pq`Wg@*^i
z^Towg3rvQ(pO8B?ViS+Jzqb`1fACIhZ9K5j%lRU(&eyxm_w!Ep$zIp*tj{k7;SBMv
zLeN9|P~3qt;6WX}OS~=pC3N0wft*k|ojVn8sP^<zxIhd45R#ieL~yA7=2ECVjJpsx
z60Qp&Ey7!bo_H9{mT6zj!haM{`6BtDQofjO8w0ye5FQF6F0B3u6F5_t*E<jhA#BkL
zG5FGf(31PvO4w_pvY`D)HKwyQN1wPn!`tZ4Laf<xRr@0ZPUJx%wB$}49&C%BK6hUk
z=iX`0iU&dvO=k$zJ{FMa1$rOCFFf`Texb;rkIW^4InKLnW5s}Ar<k7>TnsGG>-Py8
zg;FQuca+rDd}&TUfe*@A95puL{r7$o2m9N$3a``_wWnIr+%H7AG^>46M$I^P&a&OD
z`;pOW@C?Mm0U<EL(Bs3s`0UoZu7i-0Yinz<y0Q`r3k$ZK&uPvPKqduk>RYe1uV>zm
zyLWC0KF<a7bAmy_l+9Yemb}^2FRv|>Q|vzwY9cU9;q(#CBN)~ORM!~`XcVxA@b*~q
zb!ONjtP)Bo>OxQo>}!-F%lM)6xYWLBaFmpPy=F`cz?}C<ZKlqPVK+L>L+wBHoSJXr
zO4TtuBYzxTOBXS(zAk7l%DxA0vz^^t@0;-W1NITT4+ZnUMYlVMlarRCo8aL^I5>p&
zj15-sH*dZc%d0n{tg)lbz*_PunbUYCSjN)^uNQ>W90*@)mAd*l?|Q}_W$ejNc45(v
z@{f^Zk2gdaptXdcHJ{U((AaZtF3&TpPa=4Xgx{*0e$CCzIiC{?FzxZvKGL+GdRd!n
z2nCPKo)t&a{MO#I*VaJ@u7t0kCn&iv-~@b8?t-V>pXRZ~Gzpq9aAgb-Xvalg@WDP)
z{nI&=EX+ew<Hx<^0J0~7>KgMU;I`!5C|GsKxr8$i2;<p<fOm>Es>O^OiiwXt`ZPZH
z@DsH`aNxS=9o^PF;~&3M8CS%A>f%Q_2WID+5C55tT;@W!ka@fm%;1gJ@kP8o@o*aB
zRbU$$LHk+ngq1@<MM%JYw<ow<j+JXSlJZ~uv9YYiJ!e5RN|KJo9S`58y?{?oHE$@a
zHRr$)3L-q9@E+OTx)+~({9bH4xMw91`+Ef5Hi#8oW8Mh4hXDWzFY+d#zye1T+7?el
zyaSm(6d^?FA@nOo&YZ^+7$ppHf!*!x*xB0gHSVpcujCZ0*6M~$a(r!%6koI(MHog0
z`;9%*6AUgK+EmXH<`!=+lr7FXHNe0F<q%4$Gt)(6DtLtfhba_n&h5nO%IiS=(-PKq
z;Te?pDBJa1GT_G8qpjx~T2CmfvxZ5BBh72vtw7Vni>8h1$Ge75x`qH(%s1h$Gton>
zD=UhG4}k@g8_0d|T4i118R?a&bI`7y4%*232&+h&&P0n)5_iaHH5_PtYJN0sG{fQQ
zG#5jA9pM>8J2WnWAJ%m88Wq0K8fGnn8yvb%2QY*I0%3~@xym!NpS4Bh@CtYjqWHvM
zqNj1;JVpWF4gK-`$c56lbv2ipaVgx5r&l3gb^jO-lxHo`z2tF~_#Dge)J8med4B;#
zzUTmBIaHrfn%fXRhxas!FZD@n#N&vuIy)6?_Kux56GIe%ua#cF0|+|tNXWn1R-yvu
z(3^2?D=UYf8=koo_V7e>^!Rwkv==(e{4qZ$hEW1RM^fQA3bb9g&|X$Wy%U2b6vu>1
z6#T2rcn*T&z_;*vC@TqF%$PyrJf_Tv+D;BS6p&W%sg2A93aw18Y6T)jAy!shrqGtj
z5kYedZd4uTJTF{`0Tag%FE7D`m021~hf<tUUvep&;^E=G=1uEWFp4)nFyQqs8ZZN2
zCjJ0p7rZ{fM=4JlIW6IFk_R<oP%JHoMlRb><len|(b4_D;aEQCGWtkg9{zZA{Yv?2
z`1zk+f!mrWz9wZ-zNRu;)y8tZ9y#2QN1JwVPI5yhOi8kyfQh#914a&4STR~C!eE-Q
z9N{+`I2ak2wC3?~H1_tKmuq+iLn`7YBn^{(agJq}h;T6%C2aua*la8=iSZLM)yvuN
z_O)>*zWxb`g|!Jj6VpVSK&JLg6EkP!5(JpazJmn<zPh$<#?}^sB6Rc(s}B@=Fp^yp
za4}JgwHP-x!&qJr(gaJYfe0rGK)W-mugnKZFBDxQaYcd1m=ZFA;#k<aX74V~A#<cn
zjjq|_j`3nc!<gwe<H?vK5Vdum+RP>c3uPE)1Pivh5p3AgOUw$4fIaq9OrZ3Yc4Do&
zv$GrTyz@@%?H~L4q)f`BOv<FVd~1rYfB6?r<oMFLN575opLy`d-}(AQhi+sm^gI6>
zz4TGrzPUqx)Z!~`|6LFMyg>1LdE)BBSBigf`C<Fgx$*CG&NVzd`2F;Qk0<VZtUNk%
zwf?6b-1C$gIQ8hB@qPU9g-6F@1wP^Gxkt=ex_wmNmE(DZtK9RbmS+^|xpMfqbma>Q
z_ot6%AEUPFtLx_*|KXGL8@_9<T8uyX!E3R$x@5u8*4Cy4h=eACx8eaCEs0;VkV-rn
zJT$@yETZAP9HQzX<=r7H8Xf?6cfkvC9)Z!(QNw~lcszte2-M)mo8r6SmlvJ$Yr(=E
z1S;euvQd@pX&mgwCm(+pjl;trHyFZrge(Ycct1xG0cNqJXRrpwTa0$4=n#&+Abt>D
z9B-F9cW%YW$+3kagrO>HZ$V!XxWg|Zut(^HFq4pY4z;A`;W-f~!T%x{C!89dEah_5
zUMH}&j45M{;0_CZgr9^Uq#X!*cXu{oe}BtX@d%Xg2C#MX3^_WIvJvqC7WN1^QFa_3
zH`Isoz=JN6Lja1v4}sQnWTsN+D)Y=2B=Sc{IJCmb-UVWXQ)2&u4cF*X8G*gSLLpG2
zE{fhcm8{NVNT2aUy7b)U@*1B~U1MJH_#kk_`(k+7)iNP;Qpyd2u+*M$rCOA0%Q2l@
zP#t)~&?oWp8ehDO;JwezhE}LpsZ}a#d_(X<2tuN6<SY~<9}eSNpI?rg#xav657t8D
zbC+>^xML5mKl<UH#`nMfN3o2TABsVZC3(KiwRf|J2cerZcPF}kd+UMr8D|#a18cw!
zZL?BEunJrz&$Kkp<Z)ry;N8?z+mfP2^Fn{rM)KmcPVk&MjDv$M&E>2;pqO*Ujd@}J
z#Qq26M`0E(3_R>O-?GOEdrar1(fuB#i*B@zcC;UD=zYoSk$n=d!!wLMp)OOXG(!7L
zX)gNOgHl(5Cp>3}Y{7G_7HhAm=lbgBsn^GuLvLoIR$q!6FTJL{-*v-3-bujc-p;n+
zsa!54#Xzy(HQMcTwT5P*ib6~W9&i8{%3?gM2>S=DQ`jq_@E-5iJkEM;+SZNsa9R_r
zTSDWp4%vIMf2S>Tb9F0OQm9Sv3(7M*|5#TQ!!To<3?>n-Y7c2a=$Xd#Fpe4rh93~H
zr3E86MNo^77Nsi+Ir99Vh{50!MOG$9pK_XqQwwMDZpmk-qoy|EDFn_dXujCXc2TnE
zS-cWY`^VZ3w_{4}1V1JvreJ`4NP;PpM66+yD$SEap`&4BGnaAuvs>}^|L~L8+}LuP
zQm8RCmu>QB@0tMP(Z`o7dPwWo=wJ0c+z<P&cgSjscq73P!gE46Hk%#oiw|Ra_rRZr
zf-RIp8P$Vens!#J)tH~3w-Inn_?SKa=H_<n?d=OM;$fQ?tO2*81O1N1EEn@ycPO2Z
z4FDJRzEw*I)#>7PPCeY$Bcs{h+qID@b8qA00%M60!Kua`FSB#E9~yu_oiK@nU|YoN
zY+Z1z_Zc?MM0d1qSWn0;P+aANzvt_7`pg<$fp5GESJ&6<^~j*%kp^ABqm%WqFuxef
zD{ES_!owO$V2Hl6hU@hO!!br91+5uir7s!=4Z+J4!98#fKBl0AMp1`1*sQ}~f_pSB
z&=~;$hSV*=3PvaNk5GZog$(ONcoBsO9^8aqwfCX$F5a{K!MXa<*7L_&aOAo}fq+p3
z3K6_}ga=S+YJE;?-O>?zWr2jj<)XdEM~9m8fzMk33p%6*)kD8f;`YfECHl;~WDBJz
z7P2}&G|ncsR=N*`lW1|FY{f)o`a<u<OPKJrC{{0x$BPb&SJY2rPsol=)vTdCkWh-w
z=sEB(bOVDIXe-JmXgF*BOt6b$1v+d`?-_8$lyKmg4Fm8fE#$?ss3YLy{rz3l55AmN
zn=~%;5jrY3A!MxKNbRT3c!Hj3?(ootW;4fG<MO<{x!S`2&;%4P&e5s+Ek7Id)c@oi
zH(p=BH~5rKdk*SJj08?Ko^9a~JQ0h9N>oLk2w91t$HVZenlErT^K;nP*ZVKL_VCaL
zAL+&*>??{b?!V9&(q5hL`z#R?{?k{VxAGB>D!s$XW7UfhIXk6u(Gu2LCP&EOORYDI
z+(jeR51zM@QR@Lbh2qLyx9}#0F}+9aBqS}$)(lEKjg5vlji>1W-n&VGX=SI{1@6bd
zg88<QMXROv=^>3LV^40_<UKeFLk(;&N19XkHP)k*#iKeoL8wk>K0Ga9qH}@)@Kfug
z8SlOSp5~yfXOoeP$5;iLY8j$ho5sqgT|ACH5)b{)r9A5Q^AARGqWGGWN%^`843{<*
zWig<Rc@D-ElgZ`=BnPsxC58|%j2Vn92o7rmJiUlX&88H_9byR*W)pm<P4(g4Ui7rF
zAY9?4h%PwKUPF2hR(iyTt<EpR!t$~iU^dneD3~Q;pzmyKMzhg~LAPfH3uXz#4P$Rh
zQV_9s#bD-9f4)P7aZ$G|WI#{>^TK8V;l=)bO5EMo_|Y$zNCW|Fq;M%F*VaID07IWP
zc>@tNl#B>12oZucs2C&|3m7XEd($y*Vp_4PMR{hzt+8=_F}>deQizm|x`h!UY_=NG
zn8Mh=95c@*T3ASF{wXjRdQ5c*M$|8!WiBc7Gcn^GTW-Dg-bZnCj3vmVOv<E8%A}0S
zx2X8~i^mjSAn!-piAUeQsEjT4o;>vD(f6-dzIqKWR{S2d_gQ`7(4XgC=b)w+VK*M#
z&lBn2xNv%WJXyxiQ26$+o~PaXWbp%|!yMK4tuOHGxP0Nexc2D!_}AlSDOX47@hi`}
zc6HpIRlYR4H|o>#YgO@W>HRNqxFLLs6>^RaPKY;JUaG~P{PFjqRwazwAhtF);^^o=
z{M3bcXoN?U<b9L`=gI2=&wwO=fWIOP(B9U4=eWVc1;Geb!dO4!v4lVsVG#GCz(GK-
zmiau98&LjCiI+PSFOM}b?;x}|yf+rccoV>1BG4>Xs^S+B#-PZ-;EnK#Weeo-@Je|`
zP||Vz-2KLan0y{6Xzty+9Y==^3!unF#{K|H{C2mabET-)7cE%B3cK5D*~^1EZ6&St
zBS^W7WSxwcRi#RPE4^3$u$*RJz<sl&nuS;>+z|HQm9V$}Aolk*<M?P_4C$G@2`u2!
zv()F!hXpW<JJ!&bF;E*2pe2DG!XB>SeW!K|$>lVq_8-N`$#!I>hMJ2L3&x0mut4a=
z+Cq>!1UPV%47~$I0t$iY3<5uZBMG5VPz^e}(QGziCRd31g2sDpU3D$#a#Cs#wh4uS
z!#MStwz6hFh5#2ORc&r5re?6T9xL)j@D@4E{X$&3_L3DugwsPwfL9yJz2oMQ=V@VK
zRbxjkv?TCjj1h_^B^?T<zTVl?9F^6-wOClh%V}BT$Xq7xM3ha0R?p^8c;-BQB-p>U
zem#~~*Hx#+UGRmH6@fNm<9TonCd4<42lJNIe0QyQLP(pu3zPLaW7X?4oYUm^Xjd?p
zkGa~S<|u^<OKV{0_v1Z9o*R$J6kZY<7r`zbXGQgIDueL#Jo?SO*tq+tLs%7xIq!2c
zv=-bpQL?p;Ja>4-AaKWH(tBHhL%XGMJZVNt=YU1_N5k{W=yVB7g+hVcQ>D26(yQ^>
z8*l0JwOCxarm@XMw~fcnPHf!29h)2Xv`%`ypD{yVi<cxqWO6Ib3ns|tM(!Rw+8T#@
z>T55`<yqqlLMjrDlJIvZ(1BZ&5?PH4Yk?>&j5~7#yrb~Kw2^%y$|~N+eR$$AmNWKJ
zK#7eKwxBkl04b?Hl;k|q*FD-75*AG3esY*R6$|-{4m=hFH<~-bxnaafI}n-^mE}~h
z2h4zfW;I7s8RiJ@7WSZ8|LtSVE%<OIYV+zhaHIFLuf>Bd3Gmy|Y2m%sHQu0IX$g;o
zsT^9%2y3-hZJq23R<w=<0{8Dfh>t(~Bn}Q5x(CHV&hv}1o;d?YCuLRoOy8ehjPD;S
zkLW9KK`{wX@s-i-Y{5I-v)FI6Vtel>4jRqa*=xkc#zyRHZ^TJsSL^gpuz4hW)3I_D
z#U!Bs8Bg~92M2_~#5+-I5SZHEj;5Z$U;-RS_&?S*d5tg_AYXb?7J=*FX#@!<-&LIY
zK<fuZ@{N~Xj_cQ7vNAN!7^$%i2d8I%3Dz!(*s|zKRrn6YHhJN|m&kQ!7ckA3qtt1T
z<EhscZR%={fw>FQgQeA#xOU^Z--)sdJlO9xtvtcg_?1^)7w)<lOH0c!kJ1lio;`+9
zicXtuIhQ=L3axu+tzlj7B5WP`>J}E(+`r?KJ<+aAlq!VGO!+LyTZW<t*nw7^2v&v{
z<g(QI$)f}Troe-FduPpxrje&`DhvIr>HKsMGPKz|jKjlS<L83#sB<z|(Fne6t3M|i
z3*iyg$Mme|fbfZ)>j~al!owIL02e6wfjPz+nF`)P<U}maGOj3o3D+y=#;Z~NEee0n
z)mNiBFC2@<pJ?WI&dKS7M<yOs`v+TUCm{)^y_Qi_SxJ70LSb6-LO9comhG8-qpV3V
zjE5!kEMHnoiZs!|q+Eg);{`pZF+?e*?`jV+AY?u%9~f<%Rez=&b`ymWd<eJ)86s_=
zA4RQUXg7R?y-?LZXcY1=JQ1lE90JbnbwqcCt9cJ}8F?Y&%-Ev%IXfprAD&n#M`%*|
zqb9?Eq-Pvg)EemY;E_6BOY`%qk)K`Ey(pbD->iAnM}9W&YzkcoK1bQ8xzj(^3AGl4
z$5L)Ly!OcXrgjN;CYnYBCZ0vnguIQOt8?&ns+V&pWyz&R#3n+sqG+Q3&`5jGsy?1W
zo`)d_MgU2vHBvsAe;H`N`W|QU)uD)EP6@NiToF}^s9mH>f%ihmIab7}vHG8L%9xg5
zl(wRPLovm@D9DicF-JMU7kJ<DNY#zft1zpz2%o36LPzi(&uBeav8-`-K3DOj_wU_{
zcYgg_;X90j#=H-Qnm~s~-<dlK<NXM)jz1^m$pVy36kn4vDPL2;qGPsJiIwZugq$HD
zNvK0OJ`fTi5FJn4h%;HF!gPSxA>1fEVEQs5Bn2@}pe|cCiHU0+9ouv3T*#gR;+q!3
zPl%Fots3>^Wo-~v#h3`u3c15*A2)DwKhkD@JN7m=<Ft?Ug5D$KoV=j)TW@h#6o`)y
zG_Kyd!$f0*Prnm|u|c`@Kuq}U+jqq9Hnf>(#E$y5xwRQhZF+E5X3VhIZE9oPI5-q@
z)Ut&p8yJ|!W=jmB#&IxKWI<^3$vt!#W))!tgg?!X<_YEkMqk%q9M~`-uy}O9FeBK2
zIY!Bq=GCs?Sgb%WnP&)-KbdnKqjQf++BCofoiv+b=C|U*4?cFB@=2MLNtu*Md04(R
z#n)f|#g882lcBxiO$O-uiw^$yAK#x+sK<4VzW<>dFTx%?qstG=xX=3YSlv$^Pb;6h
z&tv@g;|m+{0_9;n59{^Ubp8d(xGj$zqid<+rx)(yGI#jv!{^7p(|xJ<uJrXwlt*=b
z{?MP%T~8@b-uo5H)w(P|6?=w-?R>ow%Zt@ONrmSN0%xpyv97hy#r%{$EbxGjz7#))
zrTv9@Ff7qgh#em7$HCsVg$lK~x(^GM`yIc}7V;|ad4!}}T3itih_F8b3tohU<m?)#
zE(DZ31J6fZ9fbUZXG8G?502ssJ`X`_w~Js&eZd24hGWKjBzz+VSKOnU&1)jWLeOdN
z9@R-090cc|eEMm;|H1olcz9%i=HlXV%KO4TL;ncLO?Ye+4cN3I+hI(S5mw5nl$38t
zz=QCU@kJ4Y1v3I2+JsUJfn)L#>8hUtD{II{hTxPv4Z?Phwdb7qLZCIUzzT!3#4n4F
z9%v3UPbs_;Ic>HqphxI_eB6xvgNEjEM)NRdp(7qg%ol<ncyxqP7DTH)HGlLzglr`K
zzZd&^JLa`50Gz4lTJlCfiGW}U<z=y$)4r)~!4v#8sx*`w4(&#L7APSKp{P&(jyGLV
z^Hoch?)`QgYi<c;gg|U{?S}STt9nPp!f@yP(70$+)pmrSf{(6Xp?#>jlHeA>p%tDQ
zuVh5Wo)Cd+ve-vSg(poegSCAU!sd${vq>mCqdsv@o?Ks8tn(BleP(BXeT}(blRb~B
zcZf=IuH>=^(@+TaHJ3%tBQTE8mQZdC1KutkXKF;b$>;Q0<Y|-IJ(b0K0L9*Aw43{}
z`QWoSKyjoxfi;w82<ge^gZCrKE|f&Hg;1L)fq~1-&7Jt<)~A~P6X&jZ<Bd0BesR$X
zXVycxRE+O^|9kPT|IPmrfAuf_r?~#|YaTZ|XUHA4t2yV{uJ*UUCb@7>_+cD+ba<$_
z&7>SD?jx4>cm|;O+IJXJ)*gnV+uAqYyZf0HO?YzwlPFu+^P#jrflS?p+GDmmDa=}z
zlmj~0AL1p#8gjT)jU@_Dgw$y*WsNg{U-pWO0pUwgDBuN0E*F$mtli`_wjBA{CBOS@
zK&ZTIROgm7?lq0s(8}nfv`y=j95jR#1I`g_pADLE@6Kn@I6Blj$Q!8hdIPvJSIF5b
zZmx#+-K@?E{sl|44<YrXh2HF6K^V+?l1-FLz?=56+1XfHzGh?Ey<N@w)|T$mvuclx
za8=LM0(ST^v>@G^_Pygm9exR8$hT2-HY9fa*|=0&>C51Zd^tp`0pAQfUt3!{u_ZWr
z@L<#4x>}k+7(?(ra8AGcTjOK6+TMu0y}f8@pA9}{u6QrTs*E|gt}2zX;Q~Agz#AR}
z<J8X_a~+z2cMIdqSmTKZzJ*>Pc!!37-)#6inoA68X*ZrDd%HVY!bJR;ahNmUhPCwI
z{(Zfxt9r}Am4v5F2DQ|MqKaIT%r!W-R+)=>y>0^vXvdu30Xm1L7vUf2b4&H2hz1W?
zai{)5AHhF(wmLkfU<}1<^LWqiDHib<7A{i1rhzZKe5MJ(dMVu6GK?m#7-TSK8gTL@
zC3y_^+(y4f;<UM7FSn}tf>LOE^MP>@v=%QqU=c$K>MTINL^l)8yR=vAiQpXLXB&Kv
z6p7$QU;txg-jzZh;yD9uB8A0Ja4*`RH3_KUv4gUMwj`wmJOngYFeV(nuesP0uYl6o
zxHm^wJFO3ty}EZs&#>l^sU4Goq)*5^J+FCYZfEisC#c<*DJ1Tm@D`DTP;x@oQMLit
zz*xSJjoE_mP9B+2-Ub=jd|7i=iG`)bkw+lL>guyPM1C_XoXBS-n1Y7T9}0@l#5<8I
zo=~X5QQ}n`3CS`n=nIM-lw!gqQx}2(6jf(_H)8=MV=kc?tQqDVx`9ls**X!d<H3c(
z5GAMP|Fom_K?BJb*3&r@w4!0cb>Ku`8lwnkForJV#zJwVPN;n-^EC3J^ZKE&$BXV<
zeFC>>{E&&NEzCC)m!f3}_sz`ebK1NGVI0vz!^^?@af}TUP+DWmfnt+9UIXFunGDJ_
z6u&6=cz0UUtRKb-C9w0%0Vg_7J$cOm22>c+6LJ)<OY=hD7PXJHjM5NqVt6X%0GfyV
zkMF=Yp*MLS>&Z$o!6msMp(!4Ft=%-ws$Vd_ySp2|`t`3w-vonWn&MezC=$+0_><$w
zGCu#bYu`*6^X}io@pTp;ZKC*^lu7xT3JRlAeJ<8-T(^Z#O8Uq~hwvUKvyw6lL~8`D
zPn0wq6p$hcCxl>Jhgdo0vNjIK`v+EhL3lv~AV3b{STC<%i&Y(LxT%Yzt}QW6BnNKn
z>_lTv8<8eqKk!}>Apwt=0TA*OVuw+Ni6Hz0h#N)&Vr3=U7}I_S3Pc#I-R;dyZF&!4
zYhyF++`cUa_r4W`r-L)QW}lo8;~ge8C5J_kL^m{aW@;hqgc`BSx6eZ?Q+#DC#Hh!L
zNr<t=8bZoI;AU?^>0e0!2_c6_rO*c?CP#oE<f`Y{WZ5!|-+B&a2ZE>9nV`BZeNvc<
z^vUM+T>WTileifle{xHVs^)!CCS_74Wm3lF+w=JP%O^a(zGQ)7BwYWDV|3piTv6LC
zRpRq0hkG{u{T(Z`=&5s|Lx0juPc42jeH|4)H~Rji%J_YW^gpw>c6Q;TU+H*M9<_aB
zNW<|&d34X`uRX8$ZC9SX`uvkMJ$c{NdtbD8Gykw=zN7rY;u_DcVH`9_zMk^BVEKyx
z5J7~!FvR;Z1_+7>FXQ|jIuM4Po~LL42zFMl!7pF8mkx$Y?e?)g&n-l=5tMkzY7GG*
zVes@`wF$3&6iWySv*PjGMoo#6`l}sSMsM7|8$05+35!*&Rm~4lELarpnuKe3Tp?)0
z>K?(d_)DF|;0h~g!cCz-s8y?R{l?2NKfi3j$Y-B^8u#wqkF&ub<`Ha)SEn5*vVg_`
z-jn@>m`(bE^}EMoR5wZj=EL3|dIsSt??%{!0ETgb_hipO=qs*~yNK{qXXglj%dxb2
zBgl{8X}vtv*qrGY>RFV4gyCw6_dfPJ5%{G3U5cO8ex%h<J;PXBSdHb?YcX4_c&^}|
z@wj1cl*wT1hjOtU4zHzqs&mWgS6%O6PtsAJ$UC(g2m71R>9(RYTZ^&|Ecp>$nc3Fi
zkcd+V*boq>aFOIU!$U7ws1G&X82+8>9?bxKA-TTS8T*37m*WL9L%2vihd>k$o%#Aw
zOk2@b)Hq(m;lWPq?cCFTYeW04Ms#%SY~FUhuD$)8$ZJnk6FfNtp$3ofN6BjMnru?+
zIV2KeHR!0Hr$N}q6ygaXDc)E5PQMWPvB3xu;v;0Fy(rPxH?p79UQ=UQ)4b(0eh7@G
zGlVHC>mVdwPOv%%^37aa3@nUeUxk(VaF{I8@yI(H9L0cKFYICPa3Ft=J<KAy?R_1(
z-fu<YaL0-#!pspqlw)^y+qsAi8-!SFYMdEgLaLp{XLs+!#^#oVzzd7>v9z-2z0vVe
zL+v?_H^2YA_|cF4D!%vT52LDaW-c&VMCgjZ9<LqXXJz%ez8AC~Jc^y|&Dh=Ah@(UD
zH?=k22xQscB9zqnZjtYZec-8c<spE*d*`#b^Vz3@*_Ln!Ld6S@gR!jsupa2k*46{f
z@2<y-XesR9D$b);^7?Wvyg^sb60L=KAitC5%=<iq%Kac8AkQo=U-xsY2RvSiv(+e<
z>snJ~%_-{&rHo)rZOr86G<UPANBBZ8f#Mhs4I0P#a|98Msj#D-nGtBB^x4-~H8o$%
zCwZc~hB=gND6fm<Ie}Bw-c78><0EqQoT#0I@WR7^>U&<#(}GK%m_m+Gi<Pz2z;o=K
zcYYH+wdG>yk<$Gj1Kq}W>)?~)MN7JOREIxNPP>=-m)79b+Bl10@K^!Gf%j3NP_e?s
zIopQ5%Zj<9MkDriw_<;P+g?#9LZ`G3XYH*Bj!2!r8ra+4i^hT0Zi`TL7{L;r5X{Wj
zpbShNsAEMuQDoXE!|}=z%rQSGKO8zvutQ<35n>PT8`dWRe1zoGk5UI&1^6D?1<pn}
zHdm_~S0R77I3pA&3hOf~9_#g*+kbR)7~9)h!qo$xuaMtP_=ItSK9Ca*r4_h0DMgS)
z5N;P7nZ^~`iOi=*C`Q34Fk+mMpH)9minldx8}|IA9=w=XPvD1X#zy0T^6OGKq_OuP
z?%%zoe&E?jKDtKqI)W?V6OEesOTI9{lllbS!c%)n_<VZ0;t-vu7zt?nD%AyhL?I))
zU_EI)qSV6!(c>q)*wL8bIRhRqmP@8fglC<eAy202Vl9e}&SZvaBg)@1%dzl)bY4Y`
zG3%DK4(?70JRRy6%1}ZE!mFd;VNRh<(3rg7i~7N}&`gJ~n-R^@z4q!;z2xiUllK!Q
zlr>5i;MU2J6&%J3Iv7JdsR=te=pG4<4xAGb*a56D${<WFYZcs2+xnudDGa6H3J++)
z>xp(je_1W24O$B*Z@@e7HQ+DmL~&J{TU7h-a8~{LO-M=<bF>k~5i|ifPX<1kCE-e*
zMJZ;lD!mx~1Rg5!SKv(XMQ14_q+YDLrVRt+A<Yu9(H?T5X@Yg~S0%+A-n7u)BqQ@U
zPNUeIR$Um|wABWk*L@-SopQcW7qn9IMBg+|v<G-{KDG0nerF=5doVypNr}-2(X42r
zm1@Enc(FD_qcM0X2=@kgMxlG^7oD~joNjfUTBGoljwT=+hr*J2Fq+`~tWR>4!CR?G
zo&oxUVG70qz!9OA@ifO17%yDB`G$gP42tZzCcXfLTCda9*q+<j0SH6s2Yesy@5ej8
zeJ_qr7!&9?YmT)rKDe7ze-%UO*Td`MgL6+i|Mec<WO>^Ao^^at(G*V<Uz0K^Ut7_d
zE6vqnd41K^HW@M0Y?}LGP*O<26knLd>5ZYTkw*&TX`zLFW4Yt&LbA?VrA;qNtH$1L
z^xJJSEv)TJHacQzDs%OC>GjuRacNnEWhe%3KaRz~wGNN;DMY`E!mMjT0h*+ZSV^Wh
z$jO?fq)mEdj+p3GA!C?{Vhn}!yIov?h1I`v<<gs(1C7b9Ha&a0dm<);(7+=Mt3bk_
zv>Z<3==j*e3Yb+}S(yRM`+#{)RxnucKpe%WXf7O@Mg%5V;3B9X&UA{qD?UN@LgWwz
z<TPd_GpYzwl0YPFK$&k2n5ePP)6Dv){V*zg@(tqbN%i*>`?oCyd}CuXKK=BwXt%rR
z+N4a%q)f`BP`=erhoi4AR)ovPOn5p*=f6t%T%BM4p+BRSeQ`0}qCDr3admt=?R^~j
zlWuy!^0{ZPK0ZG8MA6-ku08J<6@Sz3^q$f0hlO8{pZUt=;b1<mjBk40`_p;fF>1d%
z@I>+XuhNHc-QPe--jt&sblLnF`B=q|c6zbBBc5+}KhAZYJV?da+2{|t7Wh#R+#r~-
z6}RfanjV1*(H3lJE4~}cTogqJGVrc)ZVmCg3kxgyM9>SbPA(e+Lo==i#TQ0p2x{iA
zRM&eD79rF_z`D1)9l!m}uj9_0yB;6-al+iecOnp>JPgBVK^gB5&#3n}G+LHWJm*PB
ze{vEfwJ+tfomC%Bt&l<>grYB5s3(IuJka1J=?`JE*mqR)UU*XW8|){J#RC%-3c-={
zOX(Bu1sm6$5nbRg+RhC!ZT_E-cZ65MlMRIp!pMcil_=NO)i*qjX7oygZb*DEjzjo^
z(D2;tV;@me9fWgY%n`1CcKcI}b3bn0d{yJOV2?NygCvZ%74><tkY+CN)<bAaIIvuC
zLH$EWr~OcCD|R+jkLsJrPFYYpzp$qF%&9#{?b)jUG>o@_N5??zB99gdthDE0{828V
zK()2DSaGbu5hSCAM8M=ex*#|cENHH#qZiYcr<xO#RtS+Vqtj}{KmPrHi=Y1Zzs2T*
z+tEC3ICs<i`?ungPu{nOQmI<8R~u`FoWj5zb!0V<)SJ!WNi`Lxy<Rj=_SLqQz@9PH
zTDa7a$;MQ+s4ikSIc}HsiE1mt;BhEP{?TU?IViah7+Q+W-ja8s+|V8b;UWqfgpj~J
zaAMD?PE+uL2gg}74)B0!Mp4%(Y+Wp%WE6Tw78U~5C_53}Be+MH%w7gB8axIOI-j26
z6@@Y1Fb)on9X1c+Lvk<8&y%|dMI0Uny?E=bH{-AW=3m8YuYJ!7t^pzHQ9`Is?3XK*
zio=L5F0WaUgd&J}?An8+>$S!Btgm19T1HsPdHMlfnV(Pjjj-z9cyK>9Hy>CCnFOXO
zQV7b8v4R>QF8QB;Z+rXbCU8ql>(ZVt8gAwQ?*i^cF@82cX^f>m2Bq3xYi!fHtHx|~
zP48hmd(jY#%w)9pMgS{V1J6+GEd$H}e?A1CLH<JW_cZm%b(CY8Uxd$kKf>j_)<dJQ
z6TkiKFFltQz-um9kKh>c;S2>&78q-tl1prJ^S;LrMPgQK2+z%q;BRkdQ*g(aBc!Gc
zY2Dntb1Q!Gv!7`lqeMpzL0z<3w~TQ%F!=EM7njHB2;X`B@d{W2oEx9@d-a>uO}ji+
zml`F_ojt%b2gp)N#e(Lg9fyZ|f<^L3F&{IdInKn|`n6bIS+x;?jS;}j`h*^!R0nQ=
zO^j+=ZNZu75BLJlAC$BRwaFJrXvKs5v|jKsPH7m(H;I8m3Uvz%u{KbS0%uOipzroU
zRa^4;yzwm_#B|@@n1W5h(=M;z;kzg}LP?Nl+VRn`;f_3JDZiW`L2YxI1mQK-AD+oQ
z(GEP0p;cDYs*lK<inHXn!ef<ORD{l**Sl(JOV;%t9PGrd;DC^<&{#q=YS;{0$-@-|
zV8%*~?TtHeNBDiVR8(C>)z3Ry28{}IfDpmLmzR1!q2w-we@*9A2VSyMmp#9uzO-&7
zFbW+E2*4|NX))dj-n1$iKaDF2J(L5ih3P3gVzoX6t0*h+Iy%>U1KT$A*4TjOz@I2U
zpj9ZO$?pfMr=XPNJ2Ev44wx76j@9uD7p{Q@+Vf0!7sHrywE>)zlsBP!i;*v(Sk9OR
z5Q=ep9$bO99&m=Y6M&3DzSC_AZnoot_kSB7eej;%k70`Fv-$#TFpeBdomE^^VcYKM
zPN|{0q@}w-LPC-5k?tP4hL9GKl1}N6cBr8nq`O0E7`oZK-}~+T+Xri%t;6*^cU;&1
zIzt(V!rc}I5xH`?t`bST&$s6@?w&*~JpB?EZ;1-y!KsG}G!lZV|FYbN4ggd=Y;SjK
zOQi_!4W8ZvTIOii$}A5V97>o`aK!nrF^|~{o3K1@Qi+hYlS-rDZ+l-YchI8Js+thn
z>ZV&?28|VVX>POl#ofFUmH*(-Xss}>{ia|Bn4_IpZ++S0QG<1P0Sy4l<GWMuGCN>c
zwfJ&DbhOiDIZ9D(OeRQxY74+P4BoQ|F-ZPg-2p&}<ji`K;XZEbO%Bd3z!mRBQJgoR
z`TT>EEImfHOhWo}CTfkP@y=O|J>e0JRJhhmSFwCxdHbZvUjJipzoX%V8J5|ZM2y`b
zWG2fTB-?3&DjeS3dK>1AW0F=~!G(j}cf=4FSm1j7Pw%;mY;WIp;57&H)wS$JN?&MP
zw$U<d#Yf`Uq!H0&5vElj!(@>~XV0X0bweoEmx+RgSrDm6tAL^q^T+>_y2iu7`?$1p
zycAwTh~v&tteV?N|5!`VQ=9M>x@Af8>3ZVc=C@O_x)m3FIvMeWFZGKkZqs-{7kB^;
zxer~pwHAG1c5DnX(Y?;!39t)00$J7UsT63p8atcG{wy;^uO;~3oc8VYBfkoJ*OF?#
zfrsQ55Vjrm6fZ)oI7s8!-Z+&N<0V#o%*y(>jaA><##nPfKpjntgSv@B)$wurNTjFD
zy4OB3dLInVFrv<iVzZwU6=hH>dm`Y+0d2t8!MLJ!jWf?<Em!|eY<>QNTmPRCImfQi
zr+}w1v*%qPjhPuJF2`Rd$9*DAQt2w-d(?~6$?Akio@s8)?vyNfuvCR>p!!XRqYA9`
zZwXv9Ft}a<a-GHaq_R8Z`$Ro2?Kc1vP8gZ>r=clyFw-#&8E`jRd7X?W21~!y>nw$4
z48T;Fv+p18$k?l=&8n|IqwnXinRz}9K_%A~j&j!v_isP2b!sR-x}xKANl-xPv-b7|
zeZEU%o#CiNevyYev5~pU9ynJn4}FdBA%%_+!T*A#d$fj-DB@qTs(EVG8uSR{oH<`^
zFc%|>>q&HxFZx%a-XM3o{q`*?Ec0Kk``t(a>wKstF=Nv=4caA%;tLP%4NJSG!;l3C
z>IUp?QfvEtW&N-@`+g(gCiC`m>hR;u6ZLLy)*3c`V&0n^qD^w$f1tZO40(DbkT^I6
z!+FCFbJo+{otSkXiiYym@X+#ggeZQbR1s(!N*2T4j~PMamF`1AYxH)$z|bt(G2xLD
zqb!N^{Hs;~3kEFeX7J8e(<v)au9aqoJ~I>g9hH7sG2n9Z0Z;p8$)Uywqb!nlYHhnK
zmeCk`nEQmc{fV9EnLsUw7!~7|G2z-}puas)Kx3LuxRdJQ;^-R>R|WUdj}(MNJd{dS
zK3ka4S^rUp49>FFWZsV4d);bIr4Id@n-uZ%r+J+XRByM0e0mEu!XI?JPWGCfbO0ve
zqU$fOv+3LL_1+iw%Tw^c1FXHUgWwdarYc8rog!-IkNvVk$hI+zX&vXWGr~$^IB_dn
z%AV72%r-~B(#2)uE@*+)veG7hQOC;DOk)6a3}>=%$%jpg@uV}o;I65F=x_;T3W$l~
z#HoJl4T$OsjK5|A1jq5yAlmpfP0<PKor}omC!A%qV;?12CPK>^*4+fK>76-J8_)=L
zh&I8KALGa#Os<V=N_%W=dHh^m0r)hI`gSJVWBqEvE`Wp#7w9xeqswz_=dJeR26`~n
z4pbhrbV_B%30Lp&(B_wrU!x%Y3n1TW{WixeYI}IaypQzf?(fakkNIGNi;;Oxvp3E;
zN4_ae>}2Fq0vc3)ajx!{IQil<Zfr)F*ux!(a0ffiMd9S%b~KxHdtt3TY7?U75alm{
zg(y~m__JnmE}Ksg^@2dfhKOEKxwYaIWAbj&VP$g~NQ4ZzENoYC{tlEiviHV<4QZX<
z2?K5@`!x)))I9Z378OEgBamAqazbT)YOq{xirm=JP!88uaYc02`Kb>at-7gFY(!o0
z3AH&4nAxu>yR&mbW<{r<B>ow3?D-`gEEpwrgD-vO3%Oawm$2S=yNzhp8Mb{Tp1n0P
zd!Z?vV~FAcM8?>GRVViBFNnvooSBdfbW$H`0cZN&%WUhE5pVxUt5Iweql|d;tgPg;
z8?8@3@bCmNWvxzKsEi|ZAS8T?sI0~)wavkFS(LCdY>__5RRX@y37txx<xr(7B20jY
z42fZE5;S}OSLap(lufpQ<SEYiJbGR0N#Vw2*Fr2#?2msU%G*uT$?ijynd~^v?B!I9
zt1KL42PL~}4RM@Bi4%BNesPtg8zzFrkYeq#*A&+z$0SvGH<@s7lEWB%@%I`M^2IlR
zkkERz`3PlRB^ftl8WyU5hZIxbJ)CuFX#e4(7R)`!bLuUWKJ1g3-(9@EQRlz21h$D8
zFMkTF*b952VGJBifi;S7kH&q=XbFxg7x%JV>*+u?*~|fS`v*FEFrNY<AEWVCEd6a;
z#THk@yUP%44uTVlK2a#$wrsd17NgA^|L{0*Bw-I3?WZs@IMVi&edcjvT$<CcSA@my
ziBJ!LQxM32u%Ho)L<s0(<nInUou)v4Ei8%RiF_Bo{r+4z%IImKv8_JNlfry0F(a}O
zsKCu5_b|tYrms($!xrQ0slY=C4KIi)-H<-7lzpg3C1Su<Aa^cKL!)5kRAq(Yu=D_k
zBd4_)*NyFDpYS<`xw+9J>vrI%CX8gvudL9es>nO)jqih+T}#(<&c}{KG4<>KIXo|m
zjtb+j=AjC%p^a{uh%Ht}1iE`$!=jsSJ-EK47uazYGUJBK#7<b~L-#~^7^J)m^gbLd
zM8#h^m))ts)3dcki$fxLqhT6bl2$IC+r?|z`1v>dU2_o3_U{xHv7VqJ2#mC6RY3>A
znc}yU{!ZlkzsRADU1Vr-pyqIs>zn4@eQLiIcf;k=@I&Z(6lR7Pp*SnvUkJjJX0})X
zr9F&W+N?`>ILR(JsALX=2Hb0LgQ%ZqZG&%vmJG^XbHB~%jIZZxnr9Ovn@xg0(ms{-
zJgKD7Y|czIy41zF-3z+~WLy33U>$oj*s7yCz5qbNM*<#4F3cpTTeKpooPb+eGmZ-b
zm<RCY6LS1p(ZTp0K??+kj)H~jgz|wbks{JWp9zO{XI}+0CVcLFpM$qgS4HO9sX2x<
z)Oq(?T&((ZDI5=L9|;o@B}m^f#5KL2WgVpWf{iG-<kCB|Sw|qcDX=mMvn6L#5V8Ri
z63DZZ#K$9}030|aI`R6LpgY`L-+s%+Ro+JwbB^+p%JN3@^6?VK1rmm$#Qo^t{{W7V
zB^S~)uY*$>jfycb#G#}tqUd_rsb@%MMlq|pe<MR8qvD+lLwo-&1$r?aI0XH)@>PQ$
zzu~7=RI~FcDPQ*g$oF4o>eVSOf-{>34sYr1c-+qNdCYp?AJR(OFGSH_Ry%EI@A=za
z7W>{Nz}qJzl!QBNeeJ2K{+)kX%(OWZlXxff3#ija6cLpRIj{83Wt#xOfL1FAaQw0Z
z^uO=MB?F!#)~<gA3@zt!I*<R@#vb7xI8`OY|Dp8e7S6a8&)uhdB%-|V>g5jl+__ny
z`a6Q{NAdt#>M2GeVSpHV+Mz5#;yW}%>40~4O0`Nb&*iO~@ubX*o8kQt?XpoL!U#H?
z!m_9S9FPQr<DB#q7y{lW3j4t6<BR?~l#U9q6u0_RI@&Oh`}uO>nKLylZOo@0FPWY=
zwQHz?4xXt%I2arJMUze{D-iLiQ^Kc{a1GvvlxEx=wkRj%lCM*yj==?v*3T<lk;gD3
zT-(M>2IsKJ)c#dk<mX)<(a*&H{0G!DD;C$knFHGMxT3|ci5Mgqw(m)K734L&2WI_L
z@L~;J1d$RNS|A1zW;b24Npg<D1vp~i)a*8%OuDxMwV(lcL<{Q8z%$zr7)vk)RImRV
z->}t%F63KQE?=F%fjfcGjZ{n(db6=+U=UZ#F6tGk6;kRgMD4~`iS0?`M1T`S1gyII
z(C;`(^gw~qiP5Y|0wo<UGlS$lCEfq(*kbpV647<a#J+&pOxZj3MZum}f*!wj#;!kK
zlWk4FdY{B|MQ?~nmq7FSpsGN`7{zn==-7nfEA56sbpgjU1nM#O6=qYfzgjq2unLa1
zuLuO&NY`%R;A4UFF^ax7lWdPIcucbNEA0;k%XhFsO|yNfjwvpe!#v)=9KN8K0nizE
zSJoHLDG(z^GF_s0KYaj;%ik)eOM)x(J4-tV#?}2bPJW)Frg(NCnBweca@?DMG`s|E
z`sBDVXk%oEyZP2EpPua;;~@gg%|2YCre<dG!luDoq@;Vw4_W9-KgR}^r9npfsi+lT
z5~u#Mexy(V{{~_?0!3OZ)RavMz9$POx@Z%VPQv->&NmhFrnC1(aL^`p7Y*(%id@g$
z(6_mvuab0pDPN@3#kI92#L@p6>o|CENa+iMS&^^bk`(06K+=>qLy}SOk@{IqUI^|>
zjKjJ}EEu&#!u7O<7yrnd4XwS~MBe0%zi+*M#*l$ynSa>2>R&1Rl^5Kab*N-s@xW<J
zibIs()g(}Jvqj$1CKCT?jNdh3^2bWB!_j<|D}s_n!c91p*NzWVt(q=f8L{?vW8vM;
z#)4(ha7cAT?7JS3>8Bliq-yxu^7`V}&L@{(TIpvv54HD*5@fBp#Fw{femaHEGpM)#
zE95lF>tp8xyo{%CouXk1R3fhJvppL*Flb!G3Gin(ZEGcUaEPsIYhuWJ{ou^s2WB+%
z=T&b*NqNI%CJWc9(g9l7^|gdu;XHN(T##&5X7<<8{zp$F?sIyu)-*3-`UyhWifl1D
zhZ3&Fadf@n{;b$ddemQ!NV0LFbMZb%1V76#D`og!h|n3OAtH{Y#26hx450z4NOAx$
z<!RM~gN<-6e15>8nMDN6v{hI!zne(efi((B()tF=sQm-5`yuHe340=jvdU;~?s?67
zrQ@oO<S*#C;8rewn$Y;KHL+ZTgz4;SqG-Xo=(z|dR~5F9LqFN1J_Kikq^gv`6bdzC
zJ8a~p$F(Uf=q`|CbI`)jZ!JQG>5f?vod39z3oglL{q(h@C`&(ru1`5;g)zW214kxH
zTx~+8Szq%v+i3kxft+?IQtgpj?v&vNdMwRV=edK0@2tKH<)PDdr+L&0QVgsVuA8nt
z@>LG5kbXeT$GyQE2tym&%y>k&K@Gb?)miac-(!Nu)*YG1huq&`KUIC_cq0R1uA)l?
z=-m9!a#|bsG_0H#TSw~)#`w!!D)n<`MRpX^v}r8kLjQt;5bzs*xP^`970KScgLGAh
zb^|;X#g<4-J6QFXVot{V>)c*)aQnKSMqTu7DY9nlg>-_ACe+KSyry^{UfL<;q+hZ;
zm9EZ%<aEmIa*WR3BB6t-mwQy7X!hx?^h-(wS)geEtGq!*6gCJ|0ClZu11FC6FKN~V
zbMA~I;HIG5TYJB_0rzW~Y2#n(AUn%P3#zKtJr-9{r8k)-JwvE5AJ=#u6Vg~>-b105
zMGB%PxprCK<2-rklz}|#r3b&4`V1L(n74;PAN7(i>QeJt4|6xri`i(tunm0WRy@_X
zWs5@Bm5wo_9Q|QcXIO~zW_`j{?h{r)AGRTKr|*8V>@SBiYF??JM8BjBU3^z9nx{QO
zvE1>%-vh!1IoBPf)I1-%`v2Dpu3uZ9MaWlS*YmCr_cw_FFQuFuLD!dAe#@|kKQhOh
zdPX=0ru&&&U{XFxgsse6Dw_Tb3?3W~OGJX5$G-(VWwmZBv@_x8H<tU|$VP=(-tWgn
zNU%rRY`i{?IuFfRDaZF5G)9NUg&nN)`c0~R{j3655|gSEJl;gyBz!o;`RVF&ByO6s
z`ab<<O3}ki7_v|)L7))lhRUDUX2M*<!q}506bYP!1P>VsP3l&6qfoO5G6Lm^rR(jb
z>GIZq%Zj8<=j|q#DFd*DK=?i}baK6C*W03ie5P~y9qiQ4s|&M{@M%q##e19o-GI}i
z{)eBY7kmd(=k!ePeolT-XUft0INN!%9h-UgrFCgw|7rV$ELEE_$LpZR#eWLAswsZ^
zvnAs4rtbVUUX$x|f`8;wbs4<uX@XtTZg_vVN32>04Ozwa_K~dp01jQ{X)}0iQnR16
zt-eHG!=OhR&c0L`<f+`^*RCRVssx)8PN1Ox&KA$~^{-gy={76R2>xZ@GYY(+FSx~%
zjx#6fe&4xHO1t8E@HWL-7-LMYM)}WbLi<x7JP@O8&`9=IbH=ZkLsAqA{m4m_zQ1E>
z&b`Ynog`puH93Fya|W?pcK}0R3kEy`WU|Xir`+YDZ<Cm$M!MgCjY#o;3a-RxJ3v+?
zLD^u7BWM%<CVAuc!b^2U6l#k#5MEpSF?&Arhxa%ZRQ`k=kMf}|xRz)@8GTEXr=ap#
zcNha8?;p0ku<lEs9Q8n>eDP<gkS<Pz&Z9b<hmi&pqZJB9Vq`$IA@@hI)HlVq)V&Cg
zT9mnv&m}0>lnmKO2(V|fIA?S?Xi%0*7bd~TD=i%_EE76CaXeTI{Xu~7g}jh`ug6rY
zAp-eX48E{K&A5dM>`M2%k~*i%f0UscZOKci)tY-spSoonRl;;7eY%$Qgx`xE_4U~e
zY&H|%1f35lkMHjJ4v_)fET*8&Q1DdKZt<&`$gJVrUjLY$XLfBG+xZ)uNzMHQKF%rK
zB}w9A>g8eT9^u3Ot)E&3qqy$`&wmtm81zct@0kCX4C0&&<$(Bl>iWx^7m--+H`u8A
zh9@MzcSM@IezdEYJ;`l`)kofCbJ-?>Ej4m4QJI-y$T_@G(kxK3mc)xCKPJ=~lZ`ty
zcQ#kpSU2K*%wdM0rjImT7l@1dyZvfsUJu<CUVk)woWc(#sDa#<OBxV}b$V_W>%WGZ
z{}enQZCwgu{W1R!M{dEL7$tWnUUnuLA;l*}YXh@76t?&-CqZ#LraKRaD3T8i5jCe4
z89-(^xzoxVRZs~uHl7RM;x-yI{Va)5Iiy^=EK$ma9%?7IC)wh|+SDZ6Zps&gQgYJ0
zeU0#^a)=z;NBbN;bWBIZ(!1oDbvO_*1aGs#kA@;4a(Qh%g<9-5+G)Hdkd4DtWwnMo
zrgQ8lz7wCn0nM5Uc#s+yu~|t?XVELv{Cc*lsR@6g4y5h>dv5s36q(77F%2y9TGbN&
zN)<kU%sA}FEJ1E+m2q%2vE}k|4ciii1yI2ruM2Nqql!q3G6(9A);+heP5`-uCy#<_
z&&d%>G}K7raGuUs)I7fd7&Z0tndE&<-K7V@H}h2sq_jN&>#)Mnl=dV%{KT-VMbNRM
zlhZ&rI=+CWWIVkpnE3iT)yIurZD-G8X=X94cWujc9GPABZkJ@#$NzW)Y#<dn+y<do
zr~L$}M|QYOB7Ls(<ji@$$k`=lY0d=Ck><pYl7on2{)z5=Jp!3-VX9+%a__R3jOF9G
z0RK&E4AaVPrN_dZuqWU=lIb=nn>*&(kq9m>kH@mb7eSIoaF5v*9QAKPUG99I<=L{4
zezH40J;mVB16o-bIFvcGg%V<HN|hl$VoFI7wllH53GBfLKg(l(q+a{-+Yg@}^_!0j
z{+kYH9zrp-sUCnsetyU;YMy|md<O`miKW0c=O6l@Ae~<bGQFEznr^9k=3POEf#($j
z**fULG#Wx1bKK+e9RX33zyNCN7KCv*(JDdgz`uU-o8W>-9|Y{c_;^J*KeQ(7rC2#<
zdkP7^e_qRuoB65$1IX6+EZHW~rsl{pZ$r~)6)PxTdN42f-J1YZE}1(kW_nZm!)h7b
zZM|rNsW$c06`KMwvcSeVtCkRrRNe#jf>D;vz#d<`{0I4X4mGqEv=VZXGx`%?000RG
zG3eoEZXZuDJP{6$f-M4{=2&wVc#)W;R5R+f@NStjKjWer`^s+XstTyPo)Y{lsJSFx
zi>nLOl?ev;OTf}?&QGqys+fd7Sx6?E5d&!tl`&DNCT;o)v~u*8i8Mn3aEbvQJt%n~
z^d>FvU){O(T~Ik0Bt8_y;@#)EaqJtcfqb?Im$%Vqe1{;v>|8`LwqP%l?UYbjtv}0S
zJB#l>iQ1AmxW4ZYxe_gGRMZf0<y_m|cp)ct727aKNaz)|O;Cr;YDoE#dEiQX=ZW`V
zV;NeEDy6LL7U`mu!3XH2G@_#qz^8Mc8N9vudc)SDJNzA$MfA+3yIb8<=BMEkr4mK*
z6QW&A|M&it@!pqvo<*&!Xs|<^^6mbbgc)8BFV;6GSa0gu5nK4YZsN%k9sh{{{an%w
zw?A`CFix;;{3|+6SY#d*7JE}dLc<hMJqeOohc}mL(34l^zPTpl%1QP?<6_s>TfBxT
zm>qD1i15U-aM@nDhCQ-)@zTfgi#%Hp+1}M3!&+WRN2-2^P?>|nQBw$iVP%~SjbnES
zUa7B2FH|9REWS#yH3}C`UFW0aD+v<9)Q#rFH=rVeATMMM75EdxGFzT@3y}ihd(;Hy
z_!*hnsd0qid<t-#d^L@9kg&35zWlE+oIK5iZ)Vf%=?}QLdl}OwIGg!fxZ7%HXLx64
z^Em?#t?(5`s+&vh73_{jWxn4>vhCt-LIPHfEpyfYG^Pn^_d8X=I~i`2o+*%;j*XJs
zj@5{3of2SrFVqXvQ>$x(5MX=BHPP2j80m*>pyV?{ZsD69s2OK|JFQ|8OU+4P5|WLr
zs24tb5?=2fdiry2*NHgZ6C_tvnkDYeQf{3LUQ9L?OUF>BH)0E-=O>}Eqq^s%0A(S&
zhf9W)W=8nuA>I6cDWkOHG0Gn{@d+pBy1G0nv<8<bGMco7k;GAAFeV8V{<3i95H`I*
zEGT{$ER?%#D6>mJGFUZoUH7GhQY^*n3PyUcJtnh`Bwx^M_Q?RmQNCap)hZ~_hy5)u
zEK|}5aQM9#J>eV($PK*Dm&c-iZ>3JeQ+Z=6@wYcGAV15_%!6WQh=CR}h8Z#k2u8`<
z`AR>52K_`#=sUa58-GaVWb;UnVq$9?{n715sdY_{&(c@nv4Frei1@J>lA@8SgE5gg
zQ~q115MGWVeGJ3D9k7;$vMwk6jv^qQAl;_@hN4@~PxrUCXf(RdT4L_g4om?3&(^=j
z7deVDJ>C+IiVC-jezm&#`aPgq(9oT^^9qRBwC&Dt6z(4+nyM8xO9pKE^%S`jeK0c2
z4MQf}n9}SglHIA^;MtumPqAB$YnRk`DCH}B;MxF=ndKb%5_bZ3$#NT{*}1jO>v9M8
zhop@Y%c%URp-y88y2xoGd?jRVo)FprRyLubzOVv}8Z-!&7Yyf2Oqb<5|67B8|KK8D
zS&T7#dWBP=#N&M3oY{v4nh-a0!kYe`p6{UA)+!&tS!k@jzHzkl^R#rGGE+>=tMi(w
z26C>t0`cxq)q+ki>Srnb(_=F_%jmgCy!xISuhxicr6YTUoZe_o%n`7PjAJek&6^{*
zVQiP_5Q%}vaJNl>k~Y#vhe95Ma2>vbQ-MX6cZ<ZL)$&*PP@^_OcxC0!gMp)yBtej^
zFU^|;gbD}+NS;r7hVSTkn_g+5jlWwg#t#ads{S;_z8&*sz%*($a}%*x^2Fx!2-|_e
zMy_-BDkgNT#;e>HNxHZAI+YVaSf*AikV0JCXZ!br6k(RJlI&5+jgThB(BfZE0?Lj7
z)broq2A}G}z$XrN&W%}*G_pe8x<$xF0p-hG=cCEvF&V5s+kbiL(404<02k3COo;4?
zen=Sbd#V%<8$|o(x{rBw!|+ml>E}`6#92q(-&p>CN~hZwP@h}!lYBy3AAjU+_^5!k
zGs<#i@EPIuXaIw237#uc3?i}E@iJjG@DbbjZJZo#J)WQp4XfPG&AzYuRqL2slhRG+
z$IC+_2?akMPcWv*dHQhys5SCD`UBT~PR(4)7V(9~)}u)_)68=Bh&F5~`ef`_V+|80
zr_Ki7La2aFfhH%M_a-w#OVYXza9f4wgn7KRj4oyi9C$%rAyFX?B_1I1QmXqP)OM^i
z_m!3WxF-z%H;|i&wgzz<qseSQ*Hq5QfO}97ZSti=I>{Z_RTXw24HUfh$&rN^2newm
z@Uac4<3<Q3BZ)+yf9cQ?%L7b9epSUy@A|Ii<7vm`*DCxISFkABm)v74c7X;$U!B39
zRfQ$=E`M||_H5DgMC8bNhYRCQVy`k0vnA%gp4PQHLH2Q&UYR>0pfNkVS?8r_IaJg!
zfDpO#9(G}(F1FuW309FF^AI<sZ&u6^&&&MA4Kx#58FHVGk|YAzXdiTqw|*!`tGSY{
zEwTCA+ORO{|L*R#x-|Mz_vd%97(Axjp1b@o<m}RpBLT1JO=qF6JB`p%DR4BZ!DwL5
zB_>mvIB2KnVIcT+`8KNa<)yD(+{e<zhKF7FVo~47<hPIbEhg$YbgSO=>Sk-HalTKD
z)9+OK?5`BgL8!Q`7<J!I@)gWra9lDn4;KjUG^65k{GV<wV)M6!t-01x4K$(Ikf4D$
zTXWr7js<D5$0TJEk+(mroeo|DRqEvVGIolG@-P|k<v!s>DGX9zpeV&mE@mF}aKPkh
zaq;?ph1T|fbsOgCyv)rDY>wMmZi})jEzDPx=)$=AzpI>yQg_LrwM<^D<fi<a)+V4z
z8o|jF9H)=zMjYm&T!^J1oQ>6r0pud^MEr#u4jFvAtDD<&3A!+aJe)=uqMK~u-9n0P
zEuF7kJo@%)9+90{cC|x2GX9HL8E2sKK)f(u7c1j)8(n16=VM^HnQW8o#fYu>@+6av
z49%4>o-RRT!Bx{DrUQ$G7=-H+f2JjU&vY2k&G?FfUvn0~Kf{`ieBXO;Kf=@J<-{iQ
zs%kZ=>ue+n*I+23yL1!(zRQlQp-!P5a@H19dL*@R5GV06iJhZER_FcX2f06VsC!%D
zMr!dDu#gp9%*CFbJGJU~ZV_qWTGyb4!^zkEJ%SJ=wlzFdKC~~(ZCo8Kmqj(x=Vh++
ztwyKUhfWivPn$joW|f-J6hG-IRiz&{7@g$)|E01{z|$@;z<tRTa?t%W5xhru9;`vW
zevQ=pd>V0m+k?J7Fffa>=(jaPmdeW2NGq;$D=w9z+iG}ywthQ0ow!f74o7X4hU}Lq
za!oc0`26F~ehPy=71n5`DtnnoiU|8zKB~w@zvF&ZZQu6Ls}B5nzVN5RGC#G(nCJxF
zsSPJX`KgkGQA0ttubuldcHXp|e10x1j7pV#XTxq|7LkA1P7vs0WAr%v71e)?#Z=vh
zbc6m@!TdcNAGE6*h>vDNdT)8hVXc{~&t;~c5gzqstZRvSQzT$JJxZBwIFz?cmywf#
z^lFH|KXu?<Bbo}#qwl*crvF&-eYJuHB^AtGXoA<u{~bCah}Q%JZkuLNF{(4cRHp;z
zvBbegI|tpaczf^IyEZ=e4IE}?De*>p_O=%DjW+)p{|-@;25Jlc3G+H(9{do#zf^WF
zdx_z$o1G;f*x%OF;uz}GDgiqSBP?`{VlxkOl`q9(R%~HLISquO*d5o74rEf~^S8=p
zHOomfdu-TJ5$H2?^*w~TDabC)Wm(jX1Ae^q;QRF6Vl}2Xh9q0DUkyeNrbJpRV@Ih3
zYloH6xR{D0YmIIG(D+VAuuHdwk8weu#T;66>lbf!uJnF_jzaW^y1ypq*#{1Lk$$PM
zZA<`KdGV?G;Sc9qmQnF^o=3|2e|w43(os`$fS-@l45;yR9tPMR&k{^d3CYRDw?i-z
z`-I{kvH_N(@1I#e6dX!TiDV5>(YWsJEP8l*=Qo+@C&i>ie#TlVGBB_5)v%;eU234G
zW2rRzGduT;ifCgB>}_v*WfD&;YMp5Sp0-7(@wn2Dy&l}ETix-vHYR}b9r+$tFTK#6
z!91$>rdUgZB|fH}z)2_j7XhMO)kWT4u}coju#u{hlNOJCQ!4Dp)?lYV=I6(WmxvyU
zxqUKJY{7lF%27!<ywWycI$iX*Xm8#$x$E~FexOCJAU_sk>X7AC;?}1|hv2Av#F!5-
zuSRb|{d3-!IAh%-4`<VXWD>JRF;idRX}`Y<=ckuzl@Y@p_Ov<Z2JMLf;;;2o-GHT?
z+th>09x6WJ^$(mV$<i4q>_x(^3Jq5bi=D35W)Mz6H!AhcS580m?N$1a-Yb76hsvG{
zqr+OeNY4ZxwfN|=I7d{OFXE6G(PgPd3BxDrxOeR|TtrI3^^%nMIF6$w?kw}z;CH6<
zk|l7A9lErMx=6Ct%`fX^|3qk2v}(29O8ZLlcMCY1vYB%9!bMj>n8>tPs|tEW(LVq5
zK7dGz>OeMkS#wuLhAyAr;TQjoZZF<)g<+*j)hS&Kb!7H@{+GL4h%9l!azJhxnEy-{
za9XaZp#G1ZQoA#hWDcdZVsHPb+(XPSkT{tQ$i4kuWOu`S-8F82IA2T~t#&@QLicK=
zz`U^p{=tG_=v#lw<yFA%)uuF^LbnZ92evp&p0^0_Tfc6~fxKfx-J}QGXECR3LpjWM
zox~fB)BThe0((c`t;*OPT%f;y&)Sd46y~_x2v!NxL-hUhwD70E5j;S65%H&pHguRg
z@lX5c2pC&>_g669&UeHw5Z;!JyhN9$4K7Oe4Y=ARy#!bHIs_Y_6HySzMmBzRCC~=F
zn52WOgM=Nv@hK)me0;08VW&7O|2{JjdtEbnhPcgJ_r`WjC((NL$sV)BXsk)n(N_8#
z;O>_l!%}#gS6j_yusP4ZFE=wxU?NK%H7|VoZpi6lhj$9&q)G~5oR`uqA<&BWX8~Yp
zc4hNQ6q##QI>n6t3|pe~daF#~{TFuwZ5|rSN#-?qDUNfiEhq7COVg9t?dw}rG^!yr
z75wsG!?FALD0~PKeQZ7-+xIScycdj>cXMg`1-{GI5{hNf&7v`mkfm|74^=PJi7<T?
z`OZPwbWlkx5}_4?D0XZk^Jb~>D^-pN|8(`b%^*ED(P6%N;HajPQ2OzEFyXwj+(@cA
z?JzU+x?0K2S<?NthduV$;=A_q%$b&Vxf=v?DQ*$>YVDI~GQ_Q97QsVl`sLg$!IR0H
zkpQ+!)EdA<o~QbeA$~;=p({rnA0E^k8K%0nDfm}ZJ9pwkt|Lwn{N0CFX!@0SNB;;{
z6c8s=WOkiQ;vF}hs5)K~d3VQQsRksun{d(q>z%m7cX(GMX;ztrIkuFW?j_=LeqrN8
zC;|z&Q^o~UrBW0v_l^n=y<ljoCkC|#zszaFqr9<Po+8{OuRD(Va%0s;Iy(4vAh&OT
zc72ZGqjUs?y4uW5DVb}`pN9omKD>5!34<KS^MgdkKcv~L|K8AhVNiR+!45HL8*u6^
zUXshl=&qI{uq#tsYX^2>#wKByKw-6m#umFA0unfMV0e<x><aTuB%gx=Bg6|9L<JMv
z(ut!67mJz29`STi{b&_6dOCJKQNZuR6v>J-=F;2(@X70oeBN&u(vc(y+!}TicIQ_Q
zUO=@z`Jrc)D7@~+@v7ogp%moIZ^)lPL!_701txAFS~@zgn|Q?e^U0mDO&d>>orZq1
zK!~qYyrrugrJi5>760`+*T>Jf6Y;0WXi5S1wphQS{Be<_G`X8rv)gYs!V?|2{x5SB
zf26te*8P`!AUW+bAn*?AzMEh3QKv6GK%PT{CKV2=*bgeT)DACKdHI_G&($XHR33<h
z;NAVZNHN`x#|s{a8tLIFqnlLOJ<Q$?fA{>j_3o7R<@V*vopEkwUJE|Yx!;ZO>iy#~
z?W)By6j+32^4YI74mQ2%pJUR0Wh|M)i9aE|yA+$Ee09j5eV>h;WV=Zx{W_nHrhhNU
zdgM=RvFlDxdbrC*>OtB-OG~JMZo#NB|LGJT<zt#<2w{uhwReG+`<LRbg|M}lRAWB-
z3l42F{ifsyK9Q-wfhhmB{oY|pl)xQM`YGu<BU|h+Ch+F?14GV~>T5_(0R2KfJ_ZLt
z6a6n%M|@R2N`r3f*BsPlmXyD6c)p~7#%zwp5!<!5aG_f*CCHC{FsyXv>EepHQF36A
zX#xos)7O@82abnwdaXlfHp0*L1Q0}V>pQn3R>j_ezhV{qL9LaQ#)*B!4u?J;U}Ev@
z?YJD6F2=#4Ykm)!+xPLWaxjxx^W*@170<3vEmlmaAvKf{qbe?2s<g|m_3A)TqxroO
zS|6#01&`lF@V;r5-Kt^z1x;>(%4h#W|KGk}ESLl7R+e88X^;gXtE&BO4>Zl3?f!&_
z?<#PC8V7c&H%L5Spqt7}jB{45_li*edaKIW`@iiNf`nq?6PYB*m&%1-BF23i5`DJ5
z;|CKh8{s?VceK@4`j-~?EXCH}w|GB!jYag=>+7z@3DQur>yQZ2xk^a)T{!27s>~-h
zD}QuHG-KEMFok4sA1#CDYQ8QzPMN~H$f??I7nMCcHYb2Xazz4T*^`z`+e6By!5YUl
zit!OFAr&bX#Jiby>JO1{%m#M%lx?CN-L{H?Z?K>7Bp5vvWn2)ASa{OAj>-v&-{#+t
zdeJ&C7S^4XwX>GP=<x8o1%laOxV%+Zj=dSgW=3xNG`>KL@2Ow)%v#A^I-k5N%{r^$
zGNKpn`5ZQ@$=VKE6o!(oRf$uSNKMP7ud`<Os@$@N=Pt|Omg84z9^H)oY<mrPzVxG{
z9^Kb|K|zA4sWk%<4#vFENRGZ&C3BVLhA@klNGLuvDo_)oh<()9kIB{i`?p<$jP6;u
z9;v_Eb(BGS&x7RI_iI0ajHYeTkOAD*NCNx%NOH0$O`V^(JikF+gr-NFZ6`Zi{Dpxn
zM;Uvj_r+e`d>?u7d9KEZb8u@AP$qq({H5`{Vf}}1V%q#hF{L@HR$M!zAe36~cv(bU
zbM`Sngr=EtgTnwDu*3Cs`;F9V{QM(Z&~vQWz(nxf^1_lR0<pyDO5zl^8cO0U9}a8Q
z#JuNNH{{W1-ZvqZIQSL{L<a>u9^bmRecBJe{iSYhfhgWNaC1xgwE5j%^p~oze_~Z+
zUXqx-_jcYCo+l%?rt0&Eb``?B0Y@W^<FV5AuVO*nNl?)Z7dPkP128yveIVm{8^8=C
zi0z^@#(LzP1IXp+_hMM+U0V68x&C!lx@g&)EML`_ZYX75@dh>6v>|a4Sbrw~4~8UY
zCT#EAGGT=gozq#$gq%}Y;nM@)G4z#7*Vo=z0zV@^N3|1YU6{skzb(MUIM;4FfniBg
zz(Q%cxp?IM$)8q6y6%S_V0k8vBW;t;RY(HRbl!q~ZD^Z(eLXb!Q&Up|@LO{M_HtVV
z#K;L4)asoYjUAd19%t`*jeM&k=T+a~C^;({k;5)zg~xdn09$4JuCEfwHwl89erLBb
z9&!7L8(Ci5`t@;2M>rTg7pYeXbr?$AlTr{hc0d|IH7QkJrWh^WWp$l0BYi}uZ`HgV
zq)I@jfOW@3OqwKT&VLa0zL&|C_io}?RH^SU){Vgd&kEB2rjN`5eL?}4S2q)zMs{?+
zh17??;A=SV5B&}fe$BD#PuIl6S1$_$M&Mut*a}~Uk(Jb@KodoY9}b(>h74VVZz>Vw
zR1(jzt*Eg~Y|(x1@;oLL%&`)S)fz`)m_MtgE6)+B_6N|<W={~D*6>gXQP0omZk^CR
z@&?_qO2?#`1*_GH(%yYE5qazD`k^A8FQNcfQ}morXr?rD50gTSam9;e?T{%_OjySZ
z{#Y^rklkoZAjCT1-YC;04*0cN8VDu3wp6{LJ7Q92@>MP)Yl(~s9|)5wtZbiw$giV^
z(Sy))ET}eO-?e7;V*rb)`)`b&h&mWqGu?PZbZ%vN)x?u;1X?Zm7zo>m0=k0D^TOW{
zAtT~2=gkj397mfMdnRv}VM&jCN!CRZ`69WAYrUqm9Y;pqFEL4XMkcz<D^x3|bvXvf
zxR1hlrAdT-^l@=nw`dX;@_Aij1hWR+9o<;^l+xd#VezpciZidJPxQJ-8{pIxP7wa*
zdasi3oq!#IzMU7`NLI)^{kKuA`Gdip&zVuBQ(#CYN2<-cNnD45DR<jHh?!4t4AK3a
z#4$>lD^EE4DSKWL+DO-0mj=k?<<n0i+;VmC;PCsv$NRSUPPK9!Lil4T3!^f667J@y
z5Ut@4_SZ6jSdS7P`q$}x{a8iXS^&W36ZdjD)vf+dHL%)XY3pWyleo4hW>+r7iTYBD
zJtyH30DFMX;XmvDyrXP7HV;pUS|nurx;5#x<P~B}Z>fJ|K4i}!Z$z;v#3d)RBw;R!
zt{o`~bf&X*#1@zg3Ef}0;hhjuZqK><wikQ1M}dz>+SH(9@u}lAoOf(+bE<CQH6G<Z
zOw*W==91fcz2Gahna$hvX4A`I_+D3>bQGq_{cvv0S~*AHh4TGV>rKKib~`M>*>BiM
zGYinzTRJcXf`}G&-;Srb^(af{8g!gRx>TO+Rd_r%v<92{o;va9qce{QncY%N@F0Dt
z`@Ln89lKjew!3Jo(>G%18umsFLf)D-_#-SVc2W|9hd_)zy$}y2??i*u+q-G|@zHo<
z!oW5w<>Pb{@!HZ-^md^<wij=v9AeUMIm~B*Zgf%cT7ga+=rmy%J}RmzD~dPqq0%5G
zY#<b2N^X$QY|)uU95G-U0r99BH5Efj3-m25npgiU4fM_A&vSBn490Ba2Ab?~#C@_t
zP~xo`9_QCh?V*Q=TB>|1EDDq+u+rZu3Lca&ii*YhaaNwnZaZnQ>=M-Y3JzgQ3%6OQ
zKUjIECv!Nj#X`!HF77v3pH32XY~2KzO*pS2_>=s5rL|P!ptnu50i?=NQFTR<sK7Xc
zn0C|^-W+X=-rAymweUbZXyvi`6l|B8Ti*SmLEvSl3o9No^>4*D4z9ShA5cAJ<|Olc
z|1G#9m5{^>+gs1VCOI&VoR%n1)~e}!%dcc^J{wn9o7926d3<duL`A=emWH}3Z{TnV
z_axrgmd0ikeQ(`^v&CXR3L%NW3WV&{h8%CC@>C(dkE1G#VgdqhdPR4w^pr_8Y9>^E
zk)@BF<nf)jS`@QSum&o@u>JXKZ#LSgBt}^yT3=JK85)q+G^<=0jGtdou`!?heY2D0
zu{UXv{v}(axt>;%7}aAp?H$H05seRcCOlbuxJl&v_f#D#LMwp!ht2+H*i=Hqux?Uq
z%@iO{OBU4H+E^^uRNUqP?|0xCpnaO6^@T;~DM*=3`AtGFPmd)T@=(wgkPYChfCf6F
zm%*W4Kg1>jvuQC)!-2sE*SCJAFQ9ZSt%P`w+kE}}>btxDq<%1NI*1M*n&sCQiN4SA
zjJsQ8A2x;%VBji0S6t9h(QWPmq_Mw5_{U_5Ls&xD=kiIo?J0p>-OpKd6j3Em4xHk5
z%dRQr4HNW>5~V6hIrIU=8wgh2BCt1`^}q1(B_VG^l~X85<-T{?9!dmVR0PLqO8E&_
zSJm`2I?NZCnnkNRb+7SY>}fY-<=|d*gl$9<IQZ!X_I(PS!HgTL(sUUJeoTB}54!s4
zvwdxYzjXkBF^)Cg_qJ$NdRiu2iE{rNZYgSittRk=xX-1>55KSL<$OTRP`;-kb5_}J
zkq*3nIoLpd_f7k^(oK2J)X3e5PiqyZlR)WQz%aW*%FFh{W=2>P?BAkQRs;H!vx8$J
zn*EQDGiS$!`n;!WWTvl9N<2BMYXzROVaq_|G?15wGBGe+K9XSS@4XlS&A@Ujl1BgU
zaIaBA$D~v_P|~%<tM{eo?zBAU`Vj=!(w=i{+<p#!n9?%o`Y^oO_sS%_IJ{0piG;sF
z+CK}LBNm9^3!9@)XH{Qy#GF`*aWgNSj(Ih(^qwr|VsYn1NT?lX&0(n$f5DF%_-c`A
z9J6!xdVSqwc%U2-J<2}!dS5UlD3oe62=S&KG%Z8fNvi5Zr)a8J<MT=6#G(RvC+rIx
zb>bij4x?CrO>>S)Ft9RdRuUBE7Lqn>k52QKjo@$SIsFtjhqX>F%hK>$*F=(k<eLT9
zRianO#mOY>nc{Oimt+fF8yQ|*pimxdF{7Nq(As%3m6N%9^Nzvq!THyPRI^A8?h#Tz
zb3BNiLfQrN0?gYWXK;0>5hX;Obv=W>`NXiM2>9248s*f8pM^GRI9Uq(RX0T5sV?tQ
z%I6+ZQs%-I0rucy<Heh4k@K+)*1dgpPd|5LHdO?DGU=_D0M>Yn3;#mnBV}Xl{x|DB
zoETGZ6D9q)efwSdj?C?8R=7~#h*-beghCmIs|1I#Aw2vQW&)o9S{Emor#EeoW5N1B
zc`a7$lWT4*|6c8K;18npnFId11luRhZR`@t>B_|IjrYkGSyzahg~ljo-BAF66Gs@}
zW*mEc7k>R0gfPJlv3N)PcT8+xAy1#k@oYJgbn*u_$M^1vQSsZRVxb=XVj=rS1I==x
zME;bDu2(NYy4bv<kRqj6Dw*yPD47+46)$s#6O3WvssA~9eQ}jobDqG3af$IB>#@+b
za15-nva}NK6ZGKxc%0T5m;3sLuRT;wCG1_wCH4RhblWqooEElXb3UL)tf#<Tu=e|v
z-vX%|seE^Dxl?Q!PY&Uqv5iOnYstbT79t?XVc5iLWRL15ZW7c6!l9Ki5DWJx0$h44
z04*n!ZnTMo7dpR35mpSN3otIApL4WMz)#!)z0hg=S*j%VsB`y_e6{v8rW1xTl+N$1
zK1?~t853>I#4-vS#X*GJ7cHf#+Y90^f4baW>^Z#t%W0-D5hSm;{oa=7qyt(~swl9w
zPVn<2_GlMJO((ev1a4hfS)GpYv-R%^-AYLG(nT!HGAOx=k-f7LWxH8QJ1jIvG+B<5
zPir)Bs{VQ5PvX6KL4S+kc5X*05gYxd(9Nk}YR9<ris{hw)AsP&Dhzr|U)KKtpRnv&
z+>uIlJRkPme#!T6y$8nKAN;+DaejF8(4=~mXfM~O<Z7ByK3$7Wee;Z2|M!5M2H%#r
z*^wm=ObC{0hD?6YY&{tw@(g>mqxWYA-rk*%#h)K6&wuVXz3=e5Gc-i`evxGf`e3F6
zn>GvHY@H%E32&DBC1aawX~FXLGeo>Lg}ZlU<xw8n<2Cdw&8oDo@GJcQ7Rc92axX14
zj8_-;Tq7%2Q7X2S7$Pdy(-C3NzD7qn<m7fif{n0AW``8k&7`mJE3GaMp99p9NMZuY
zAn?YO%C+py%qe^rG{x8l>&=hPg)gkHH<?G@zgno1Bt%4_Y3869$J|=K)GN%e$Nz**
z4u|Q&tEm9f^gDID>3rmG-o&)`Ml6kZsMWR9)3EK9aC@FWh@P(Vf5-`7OaO@8kNFAi
zwWMKX<6MR<-hOj>?KZN{HF38VC@N)VUo8<QzURCydEbT!!Xe~mc=H+3DOy}th!NYB
zsQRnpb*RRU$-}dEZ1aa(jo52M-K0IA_;tstlBd#hauu4}T}FJNiWaSAobv@bnR-9p
z`r*f9%+S!<6PaRzM3Ttj3BifVXJV70{TSUWF<tCYu@oFB2ehhC)lr_&WpR{<k>}@x
zo^*P!NoUX}|03V3J7}w|5Y0+Fy4cy<FOQlJfr)}N4!?Kre$9{;{(^ntXA?a``P+fe
z&D>dvO*))yAF#v)Sx-inrJaZ?d;-<_nt7;1zjx0@2rlFkG^WSg7LQ0WdsO;Y)LH)B
z4a#B3`$|K%NyJMj>EerFa<8TbDNZSy>7#>*<qx8}t%rX`Sh8b^PiSItea)NeoZ*hQ
z%BA`eIF@0phwh3uimZw84wEo?zWe8r&GS+4Q_c2?^gZLgWKYFAvph`?-RciEYJ7!H
zB<)Bg+HVbq;zD#h#xm*1X!>K=^u8x~NRT!#?>5SZ#1ME#u1J8C2$wVi6+~k7uv<df
zTRg57j%rI_AO*Ujtok&}wM@}QC)Kl!q@=MPTY>9_5_C+d_8NJ%L|1AUdX-N+pTGMa
zU7<C7`mO{JwK8n%xE+b+4HZjqoL`bf<-x$D2%<=FCq}0B@H?f-ZjA?#OjiF>d}t5~
ztcxAGpC|=-t^P-HjYR&`Lw@-xVd4<bfluHR{=s9jX^cMydT^{NUK@FQRG<>?)ke3{
z<GCKJ{|Q~3!ls`mBH&WwU)zsEPIWQxWdKT4YF}I2nP9B06(k|hS?5;j`RPDh-@=o}
zE3x{lT4%aYd;s;9f@9kl#pf9y%~y`T$?LKE<g4piG<9+{E|9xw+l2B`O1QdKjOR{h
znMKE$G(pi1!lu90k_r9!wF~Ok?SU#Gnu?xx_OC$$#S)o%r^UYi+j#y;@DaLnobcMC
zEGK6TtKzw%3P`y&EKKo2Ek`WH?gqBjC=>9sqz1pIT=!WMmH3Ehpqg%!@|lgEs_8>K
z6X(Gr(q>@A>GGs$0*DkC@bhc(S~#KdIaZi%QGgK$SvGv|BpPX435`Phocany{4(*(
zi<RxZp`@&{gU+we3snBR-tNyD1i#4bTNQ7f*)KVBn_in{yMt7t5CK%qMj!uXlK2nS
zQ^W%svs+bsU%I4+j?$vVP~)+WI!lfKG$DU7<e@GFT7(@I?Iak$G=A>xUm^ZlRGLQ!
zsZN)9<fj<>gr><W&Y*Z%^!)q`V}qmj=nc^;AIhWZGT#LK9REclDSpr#$DzdpM>-==
zwVW32?vZ&?n3Cq_$}AF8%UyNZ4l5cR`XznYJ<LfTrWPgB>(6C_YElwnlYABwA|CPc
z!O4Hv3;dv&aHEimrlNE}s)fFe0x#%B?#2QkuVBRY^CF>gdvfPa=yyjZr#607Hr4i!
zeXp021QxO828hprC373E;ebN?4GJ8Gw<8YFfgxnl8exgbv+55X@<wcN4j+&6beQ7I
zG-JPd;bf&@Fjt78rsez-8>T1zGbr#asZIh;)2DCkv8{UyljL>u&jHiI(+hI4Z1;(M
zCoiQAuwaPF9<=DLn@LHB$br_7u9It&xh&(YlhRge95_}#FKVJ|KM`mOIR{ye16sjb
z0>DY(z}_S2{?|VX*syJ{*X8Evn=YdT5N#SGXm;Pr*{LsUyZ@ZCUG4sv$x0;2fW!38
zlsFaaYK^O2sMRj<>E$MS94uk_>f(HqhrCR2!6^Vg@^<`tT|(!Nzm<DSe+(#w-gp{6
z1zqJR-+ason(!%<SZrcVtTEg;|8H~bQV=~(R1Ylq^#LrT)$1h)EHJ?wBnU&iD0yLt
zBS1}OkV9Q~Gx8^`2%*OwEfSlQ@by7b!>Ux!4ze(*uWfoGY#JmwH$T_c>j}JgeiB)6
zh%Q5%Dt*VyB>8Dnxr1}>i!b|}O9yeq{9HlFHHl}nmgOhy+H+Jnui9oZqZ?X<BN1kv
zCwU%FK>OVsegmfe2wz9t8_O(<ditLPN*t}8ct<=@g1q6eA&82pdwp>UIh(mbzYW6A
z-0`f}2MunD#=f1-`T#uka8kwsj`*m9w&fL^k%pvA_Cw0-sHpqVVePd354h|6|1Ua|
zypN1+1sx!jf5Cy8>AIU!N_Btq!eeN<qkHPM`fuWgPrD5HD*Bk8sc-(F=~;H}pA*I9
z!*fNhCwWdYdxU@+;Ljeioa6Oiw`JvlPs>kdGd+LaycGGX{l3#TE>~_;JuUi_{fD?S
za#o!Fnc&$cYz4dL)42?>rLzTJs;(Z_iV|<4d7ZC2)v5+-?XdDfIPdo?h2lkb<4$<}
zp;_Zuq7TEUQZ(XdH<@K!ViyYX+sbcsNGwGJuoAj-unSYl8(-#H3T1k?+Gdhw=?5m~
z)%%4oF9{dYLu}p3Gfw&{g6hg@lh`yfU&Y+{!pM5m&{1-FZ-Suc+=)0a_m&`+`b`eK
zK@g)rHx_lVUPlAVvrwt`!Rj*D@u<HN^n_*VY%d*F>@5tx1L2)<gRL)P?+NmH3g6&S
zg%(6?kkS{4X5{>R-eJHbb{ZYnzoRRO48+a@l-@(b4zp(eh}YFT7Z%mpXd=S6i37I$
zHR~8yC77Xt;MQ`p3%{Fd*19@3Jy;h9S|}aR3{S$tS_?^EM!~;d@kmD6V{v3~OdVB~
z8s$%WkT6=NP30eL5M6D<RcHUXeQ<+8Oy%B|`@*ybMkgV!hJmk6*=Xx(xGB&7;pr{k
zqWr%9ZRwQmZU$*-7`haYl*XZ?yN9l!O9iD{K%}K%Xc)SryF+s5hI>B0_xHGu<N6D(
z2W#)O_G_Idf@??mveoh8^hp1*v1ZI)3$6R<y(5(S9iT6i+h;^F?4_I94Q3Sl$UZF$
z@xq)(EBk%3Sx&sYcW|YS8ppnhW40aL9jj^HnL$-*6Gi2aARV@mhkrC2f+2PGgNX*D
zeM!)bEz01T(H62x#^ou$?p%~Gi`{K9D?^nmSk>7=%AP@>%vrOpzcnZMaSCJc;<J}2
z(XWz6PVSYerD@)jwlUu#;%UX_45K2NHUF(E@Rq*|VvG03*|l2w>@%nStPOT<5$~#o
zuX=qX;SiDYS6AnjpMFtx^cBrrI$yu&mS$m9`iEtQ-V6$0YM{A5!_SzMm)>6}JcxDE
z*~bee>K!mJWXoxOSi-^(FS>i9T&$_HVwJ$k_-;^a&P~#ik&Vt6=gN1_pj#e8mRsfH
zjD;*$EYEt{gfGV>m2dEy!3(*I#4ud6_r`}IzA$FeWjA+E7K7d;u82tH%l%7F#j9`5
zd6X5qq3aAB5UlUJ&04Jk50_=8Pe6o|wkNG>E-M`ORGqDeGp&SR6}PB2McxUdOSC5m
z+g{but!Zv6*@H8wfE>!v?~BMKA43-3>IJ-IgR*lpPx<zQJu{s0ac@5EC&MnB5R%{9
z^V`q&71H;cHuNBk4$8*<H}JO0H`!`8z<_t3W&#cgFzuXqA}mX`j+ltEbJqQ2H29oJ
zrTRRSq$RRJbzaqMX*#>fUWzZ8e(N>8;xPY%_QZ-&41@(W=;-u>5^ryheCxsGiJ@`@
z%(^>Z_Yqdsu)jx;)@V$DY(j;<)^cS>BaFATc$L}mVpSo%!n}x@ot{}H&dVf9cRxw~
zB(rdk75Y0)NbWG56siwzlKZDODc-Y$p@@AdclY6eUsy&ARq8W@;Dnx*MwFK#mJK?m
z`M81PxwWHjX}8mo)9Ohd{9H@lk8#>T?e8x!7+&XQCN?{S4Pz9^iYkxnUtBJ@*cD^p
zqbmyYcRv(6{k}Nj8J23m4%Q#TWTW@v^BaRmqs9X&j&@?W6mQ8`Ud5oie2$lXTX0C>
zn`K9%sV$2Vsn;N1i%=RQRTt`re><1gbW7h&-b+)36U4f>bB%4+J!5mw@C;GFMs93B
z<asi9o-Fv9j&Di|5{u!|K(+=~_NQ4g1YEHgeb3aKiP;QAg{y=MJJrqY^OeDJLJ5Uh
z8AN{P&reH{wCAD7gCnmm6!$yrx-u@p^mY}42XZ5CT#X!Jl(y@Gct@z$u|5DSZ;9fq
z^Q-_>n&QYhfIjQ|_Xi?#Aw>(?iXP$Y#5ZwtSW?<Sd?X*%-|Z*a#y=<-bX3}5dc^6I
zxqra5RGINj5HROFo3;8tNY_Giu?^E{MX=c7@s;SDHU!<PWpnK}6+5r+@HywPX838&
zK2O*}FS`!m=7WbpDxzl!9myEd0l7QH2gHFPdJ7lYIH*Y?uG8s@32ubr^TUT~hq4~M
zO@@@*kJ8CyF+PkUpL+fnK>z$3Z$4oq5y&Z+!1tFUyU<fJ&0=oMKD!O;$=3nb)z{PI
zGf(8$SMZ4Q?OkD4p*-6jo%U%2-(TFG2H$MmURAA;Kd&S`b1*vRKMln93yyu#CVadX
z7?r#npuPNc->oxLJNB|cw)^ZONii9->7dQFL;Yrtz=`oCmI68Y^$>9d(6sii>O4Lw
zZWoZ4Q3#)d1!hedN!jSgC0NZAB8%MDHIk>sC@i#UK(HR0R@gT~>at(9m`3=<>Y^)L
z^w>_|jm193)uLl5LG6|O2scmP`IQ4dZ)A}PG5=rmQ#DMIKKx89$g9`<u1hzk{U@`<
z5$G=@*t2B8=Qb})?Jw6$HF+OIsVQ&-7-%xW?Fi)rI>gP`95>!{+pzwKrX>5ImdhQ;
zWh(>klw{g%Q4%s~bXVJJN=l{}(1Lg?ua$oG{Hu5>Z1U4%I>HMtQz&iItzhy0qS$VC
zZJg~g3w(cJOV3}&r1b)P1D*_(7h9mVlgFkeFt3M)t&R`jIX=Ij@BEb-r8@rWe9>|r
z)JYRg5pEn_(<w_2bb&j6p3C)%i~GAEN#9-b;r=aOG&2N145yu&Gp9nzH16)so`v`K
z*1NEqc(eh`ID$X<@p2UW9Zf{FJF$l+(IpLM^OLd!KM)C2NEBv1{FZ5L5E)$?ami1&
zpRvST8$eS9myuLrSrO(aY^%%Nbv+O=<!`^wFM+S0&bQDj(&SJ=ni%&;bGR_G4m~bv
zpit{Byk?NpBb*pVB^Vw~|3591cX_2rt*iaesz`MiX3dz8U~1@J93)~D6!=%)p!#Ge
zIoSW}9B1*cANKf<{QfYw^HtH+z!(p@n37+%lB3gV&6k)7S@@7JTszP}f%yL5@$s!S
z&i6D~8R1c`gc~#i|Ktmf6`ueob;TCSI9ixwlr(}oqM3sqCHrwnv(BKp6vcco1@?tc
zJCQ4Ykafh|uE84V6I<@hGlne3`|ArT=6R=UWk9W_Uv)hbM)cBWD*`78;_;yTmSUZO
z4jrx<2Ru^clf*cjgviy?0EYPNJFk>?Br~Dp#-|FVQDi2Yc91al24G2Zrw(9i+2!VB
zH6i!;l09eb`%IAmf{Cem%YFV#P=u0itR273ACa0EqU&E#!z$??)c|Wp=tmyxr!>9F
zl>$Sny;R2-BNPK0m-Ojp@HmUXV`f=7HWL8^k}SUU^+L~(#O{GA76Vv#+@p49Tu<#k
znKMN&I|(bm#(G9MB@A#yFjCV#!HGAn1EM?61)ugj%#_5%foSNCeCe<-25T{ALqDMI
zLu@%yRce~5_XFl2g!0kZ<EfAT+H{nvJ@fC+z<(DrGL-BDsK{vn3z&WBBpBQETGE2v
z4T~k?MlxpTyAyOeVPZKS1!i6<Jz#mC6-(m_BKfNu-;qU7``(UiI!<WoF8Uun&JQPC
zK(vcSi`)kpKVp%_6jU1wz|My<Tzj+NBaR>Ig9e47Zd7<QPKv^2gN7b{qGO+g7`jGW
z-~X;f@Q~Fub#>gwMDO4BIt*05<Kc*ZTR%&vUYM0ZMKQ)Z#tV^dfr(wOkK|#}h+|N$
ziu0~P?fQheP34pdWZcHwx`h8k;D!Zv5y$3y450+d*t<D7am>!mZSJaywBaQsCMI`P
ztnNs@ZKNVEq_8|*|6bS<$Vhq`ew7Rr(O36h?ziF!QG+H2C3Wcs-%@0&-@r9QOxkkK
z;f7LAUY_*Qo{5cjt%+xuci-c6(g_!nmUoyU0T_afRJ2z&Z;i4qMD6Te+omJiU^ai9
z%(LyJCgObqb=Q1z?CjEiz5By;PY<ztWS=<pdhhVCe8snm;WncYs{vIR(8r(H<n<u1
zF-SbBs^k#m8m{Q?{`JDfB~|<sCsAXoD6^{f+kUqr8dUe}Azfp4CMT5YX*)gTwR`*N
z@rh{Igh^Ya2ng)NgfA{GF6dNCj?Y)zz(+QkTr4?^poVbhmW^Y*j8ofi@m8SSy=WHX
zW*uR*C(giPzngG&wl6Bs&&|%!QW1Schleoo#rnyiT1T+Q?T~Jawb&-k?)5R?QMcCA
zwvryJ>o`qgPg4{syuT%ABEeH?*1-Qr&zDMKilr=Znd6H3ls68DB6`+{giToNQ%6!$
zR(Qd%En^+^F7oaMKA%Bxr!1g7TtB8HBE-~;--2_oFsjwDJp4FXtSvoP2AV-7B|dps
z3~Ggo7Wm8?%o{k}*T<(W_K7!zzIF7OH_oLA2>?<dW$2bkKo}_iLvJY03?&y|{cQEX
zmiEUnQbMYpo+xFLe0n$4IF@(upMovQRdxhtV@QP4`dbv+*4X`c!RNW#o~b-8MXi+X
z;&1ZbCcb%H5iS>qO@cF&uo}0XN@Q<<$n|hi6OCA2Yn+UVyUDuGio*hitMX0yuJQI$
z&D)X*MAduC=If2xUZY~i>!~4b_iN6yib|)SKy8F&4k9y=1$KhKiFmW@I~XyQ#*&4d
zo>!TvJ^o~!x)9qsTS~I-q!AR?=A85^az&gDkN{0O%&qc5%8a<B56ZZPT3ovT!Qp$W
zOc+IoCED{=8m58~$O>UmVk*~l#RZc>!!cNMAtnm{^vbbMW0kJ&*~1!h%&)Z_GP6R~
z-BW%7z`R#dv(MRxMTOIs=aMu0m)kf*Eg0W~H&uv*NZF6@AHiEGJj_i4;XGUV^=Dem
zu`881!5{y6<Vpvb?v%B^*wY84qTmW>1t4Be19TKc|7pLgix&AVj0C7fjH3Ns)r5VH
zAfSiQk@X-`ZHX&j%VD}jy86^>LA1;*k%>PLshON`es@o?*Cj+J3a}JQCuJgI_70YZ
zM15ctI*{y32XIesEpqbqkLsHy^nEAgKUmBnW%L^?v{OO^ig5(o6u$LkXYXg%(QnL3
z!?mZ`wv>a<^v`COqzQT=y<#XhHr60vee(xVZRZp?gbbL?ZH+QN?mt(SUlN_k0z&$-
zZM60n-i=sl$9AdlRb^<<#?xb_Vz)KVx>T|zb$s>uS#_vmOSuSBDx5lit@!(VFnIr0
z;{Wf5L5Woz&t<_FenIE&Om8opIy~<FY+rG<pEz*xPA8H(9_=|lRC=DPrc^Z!npO(;
zeDQ4y{*AT@NQhBP56}%x)^%4XO_}59PS{8fkGO~@)qw}VwJ|elXy54mCL<u!qv<0b
zD!D*Hwgon76L~`0)<X$9Q~H=Kxf764Rw)#z7y~qaGkPL;#aCd{Q^f(udd35ow>f!z
zhcJvGz@_vxxR>EEov^Cku}gWA)^!UuTmx-7XrQO1e`vb<%XzkYgBcic{+gG<-N%+b
zw5a1A2xahn(L7ZHvOJbbc8nxj7n)?~&<}l`6i^=;zZ>~T;+_sjndXg3Rgrt%Sa?q(
zBFxRw{gxhy<*EQ8dlmtIZDn8;#9DmZhMKHm+ZYO4|A>2R_vK>eR}Ha`b8y}kv#~(3
zG=$H=VA43BugI-IixPUB5>Yz3B~ShRt3&jef?+hg@+U%}l{tI6cd=dh_dyF__c*{F
zxwkC(Sr@5CE`fE3hpWww8iWF&mNSS7i>r~5(Sv1vSgWaATUnWwDziFmr%d1+FC5sq
zyZHKtDAZN)SoUlH%7hCGN4f=IIq%vF;RIDxRbe#i!G$|}dUB#N##s~!okp4#!@}fU
zJ59d2pmJ;)=9;)5Dp<W!nxS#9;i-}48j)H7K;esWWoEg&mYsTyV<Ux4`GAZ8j4pyN
zFTB$OH%(1!XZ{d#w+<6slIvRM*O&khPjDQEik$0Yp<|sfjH%2(W@DHR-6c9Pr=MBj
zbehKSRf$jb)zakr(v>I*vwT;u{zaHiY+ED2+e6T`B-*i%VJg6vg4$opz9W{2peWb7
zi;5i*^Pa)s-|x8ll4on{0f+E8^qW_W%ZMSw`_n1SpLSzd^eU^2>tuc8PTg^ARe=Id
z8@?pDQ+RIas~(g9(sa%PtDMMgZVnd;Y&qB$H^rFhdf&CXn<Ea!rpe1+RcQXUgtLEQ
z*Nq<BNa_J5$v1;3C)8N0g1Xn0TG?I|#{q7#Ua7cSHYM-HWGQCdD0mko5;XM!<r!Wl
z)--bXetk<?fwj>zEHFcOc5S@YNKW;{{E-FZi|LIu7bTZ=H?pJ25F!&HG1eOSz4RZ>
z2E-kDzId}aDzZirpubs|B!w_}M(z>b)_euXqtH!;ZZdKJqwJ~4JL~SJ>$DyyOdpoN
zcT>3jJpCv2qEb>xq7D=n3)=lnG}<@RG#Oav4`?@yTH9NTKA)n?eW<~oc=cN%URgLY
zHJu_$1IwGaPC@X+zWGGf5X;$0eeeYh@6V<xc&)j^Pm*(2;{kk!m;?`?`>nKlsE7wo
zzOwSfwpp?lvh=r6cJL$M93FN<5S=--IPir}U^b0Xu|46RIHjv9I#@&rvwn}3KP}QQ
zZvQzUzQr?%`AcP_93<l|e_eMy)cqd3a&mgC5Vve>Ur37Hf?@UQ58$(69nvy#fJ89i
zp|K@1pUzVYyD14x7-!4DL;3#n>#qKrt%SKX&&;z`1*F)$(d%85y^Ee<0)s*8!5agu
zB}1D6_ORog@a*~dCUWH{4Dnp|olqP&&9L>?xP+92ZnRGF2}@sW2U4r1@luK#{!EH^
zd?9@*W7jXUxLZm6PHdcuHdfL~Oxq|kRC8&D=~$~>nu>wSVwOc&u#YZNl7J4%t?lm$
zzf2CsV!hq*8L7_Tf3$VBK7-{r3WU5uwI5KIrzUTL2*#S&@5A~9l`!K2?Ro2TDAcY(
zlsio+cDcB?9iDg;YP9tqB2G~?!ljg`X<uob?&&fR%rF}z^wWq=i<M0~uZIY?AoQnV
z@x+lggT_CAvE4}B_}5Yw2DJpQJS%Jt6t0V|a#9K}KRnfa4Q;;&3hb2M;J|i<t)m&m
zBO&wcG;2MKLHtWa1PafxU}=nHQu*-*ai@z{Ss!<QAch9k+rt+|N!w1Oxs0MC+jdKJ
zw(~V7Q}eaOUrMn`HoNd?n3H?TW}DAdyjDQdRGky~8wly#?BO>X9Z`e8@Q#qwSEi+L
z2T{AKFJtQY;V5m#-JV~g4JnlKf7!=MJI1;6a8@v+4UsA`d(jIfmdH8*H4L=O!oGaR
zKUs69lO~yEo&4xe#k{7dRUx(deB}L@`f_i&#V&oP__&*mZ|Ts=xSX8w3#R;=82Alt
zB7W-(6wWcurB6L;KHE!<xXA4gt~_0mg6;UuwGS2$=KL499+U$bkf<8G`Y~)^o3Htv
znvsBaVCj8!ABh++7EYOoCQJcnGt0Q|O(m0y6f5<U|DI+@!fsHraQ+d3gpMTFSX)mR
z)r!&m#5c6KsAp}tQ*>ApaBNimX>RWz_9D@$#=z`?^`~gXC^IC~m6Y4*9>2k7)76G|
z%x@7@iq_@1i!`mA{%NNpzH!x0HB%pzfsfN-b4406E;=99GeRf`Vo#&r9rS=PcK3SW
z38Sp0Me~Qqx+d~T5BnTgIhL}|^fCBRC@MU0Kiw4PW;W&ejfBL=G$^c}t}+IP2=6hi
zBjjBF`!>(H-?Ts9uO)oB-ReIr4WM=j`sxG5Y~P3dzZ2EJYp>ly@%Y&<<2g$FRnpK*
z#a++l=pzo<ig>$30(wAlVy|B57Xb+0Iwq+Y`ew%;gbvm=6jP%sZX-f&Aie8fvzc(j
znOyV-u;lOVMVjHOyE`<TQO$sjg$;Q4wl+u6<EK|ZV1wk~B^?;J+9jPrynQ7`&Z|N4
z3J)@kNnKOSK1ak@>9}B8u>i*^GevdT&J2927onlUJ_SJBA)=nPqV=3u-V5ZPuTG%8
z>Z{q;cs6A17NW|nM^<c7bZD6d?@kG6p7}A7#6EsnVW3Gy({DNlqY@zhqtP3+7YS1X
zu74w^o;^poXn|A@jq@Bhqk`;!`@uw>8&|>;O|<{3KZ*Ej{-$1+mQGK8TwFZHc6+`a
z)Y{I1P_QkATo5p1z&!FTYQE%_%nJDQi*nfv+t0~MfLoAlZn)1xo&4KxsM5wPxSz%9
zBc%R;P<6Ho#Kcz{|AtH$-z~&uR5tvQwy)kTDs{_>vH#@cS;=<~VIasCnMgH5Qw@L?
z6Z{gV$uYxVW%Q7vC_(6$D!^N#r~b>rbij%B*Aah<%%i<?3XiC~h4e@Fe}5qF3Eg9j
zAj+DRI^!CM2y-Vva;?zQu;j~-7XJ_=$hB5N!}K+Xt?8Y|fRB_`5aGRw|5v*d)`Q(I
z1X=7YmW$@E1_!-7!Jo%rK^<G-l>YahyQ-vASw|CA-{**(&$^D8KV0I=(pU8i_nM+U
z0P-}{6S-titqt%M*g5Q$wRz}|zrJ0)n)gqrsnKo-@%dX9MC=_e$;CJx!V8Jr=I(h2
zZ+B<-qvBy<HGi9BhP{*C|30wI<1o4DY9UJRu_np-WMHy?v4$Y~iOOHp|0^aLv}E`I
zD2jrZW#|X?Nb>d5Olh=F@y^mVPjpHge0zwS!vblq<C7G{OES(O@^Lweq46PCBSU&A
z4V|c~7633IZ1b+<z{^2;k=C8NUiVbf-!o-pod`&-IgoQmV=jfmUnqnKoYW@rMrQBh
zjB{~KdvPsJ3G8?f^JuohJzd|;X3gri7uZbS42RAtGe?H&eHNhF!*%6d{$2q)*Om-=
zuFYLcwksFyn?ge3y0v8Upe*C`a67)YFzINZnVYQ-s`HT&7uPn%q>lWfGvnJ107~2!
z^9vk+!Q6pt`v`~!+&=&a`lcxwnAm>zPN>@Y<I=67Esm0I{OR>WV4}l68Mm+7Z^MxG
zd(m!5@fR*cY3&pI5+M&xkPz+Lw~>2;`yp=V>U<<-Yl-^gf*AN_Cn%`}TzwN^9*@N%
zTeRNC8MFSnhXusOOPOu~;9tT&;ZBe`<~f71sn~+ufJ$MTaO_>_q@$8AK$QGAz^~R7
zy{uUD{4Y#g+}I*5xJUBznh^T9hS%d`6}THxDR+SLGglZjwzyw_r(*>EVe^b5rnkRe
zl8k~Y4qaH}q;R!k)m(!zrAmW*pqvN%q$?z>-bKTtVlu~-dUXbp8KKV2Q$eAw^(fPE
zrpa6)9_!XkcB577s{WL~0S}v9`!W#{*h8j-2EgGHmZ|9!N7pwVtZEIk<>}H9Xj$8M
z<3%ruR)~VjMj=5GwQa2M?Ec%$m*sSF0xMT$YB5`WgXrA%2UAqPtJIFZe0rpI{8gff
zjS_yoV`cR0^g*2P{j0T0beqUbZ8R4ilBxA3zV@}IVD^T_0gss^do=tv@C2jMntlnp
z5dGR!a}?^p@iN+}40bqEUfj9Znc}*@mF!{q#1NM>=d@;GB2*rYWm4Sk{@El{{PHov
z?BgXzLh6C_qY1xCEu(JBD#WL27zGL0Cp=01*n#>)6VHuohYp<s(2meQb0%~gi7Ar0
z@ieidepbKe;4&bP(te^C3sF8R7vwF-+{j%ojF)(D8ub{micyxt<~1WvdQTEUnc)9+
z`W;NBQxrVsG{?P8?k{oJR!;58k8Rv7e2z$!jufw;o5dD0Y1Ho(z^HTjK%ebVZmRLV
z$tC)EjY0u;rlI+c(|diX(d~?ul4PLXxR<f)hqIFhdRXvX)eHB@%95u+&#<Z)8E@70
z#&_3DrEgt7rALmVF+w6sx+SwzZroTmfNR3{Ie2S)9QZhz694s~$q}PfX(fv$y8R|5
znD0T@`3MjgD!LL*suuxU#A^XmMqX=*UbSzNt9GVy<3|?w<`4r9=>b}$RXUZHT$=?<
zdLgQYg5J>|@W2a)VYjXNISX^&L32WzVOAHutLk);`4#?i2)`{Ieg6r}n&$W<hCy!P
z%JgbpK#dwpdL!z&dT9C4;xRM$!Ua(qgVz+|YNGPeG~r1aqaXe}PWt4QbYij=*zL^P
zt}=L!M5J3)YSF7D)2YSlX#bWS_$4yao9me%eka1@5rhZp!4}Bwbw7B9DJ^&0FsNs6
z=VbjqZ)19j%fgn!<&}gM(TDpLoN?(>@T7%bAMv>VJ{sravXf~s<68cC#MW8&4!-BB
z_E{#jcQ4%2XGl{qd6U6j=(3?;y|d<Qb?HQn7P5DsjvBgpD#okzl0Isi-o^u&Nvons
z66EsV5l^h;aeZ-yW}g6yosKpbym<|z!gG}?)fi}uXycI_E@T1Tn=`${iKtEi1tW43
zVT@E>Ov?%kr`58XNHvb^AL&6%5c<XC#n_d$Ho)i)3#+S)98Q?#Q`cttjfEDzskA6!
zEk6kl;sy5l_vA<4L3qgk>2jxUuL)t$Jnc^#^ZmmUrHA)hh^i*Y(^z>@Z|z$Y`o_=5
z`(D-90QEr$3>=#%FjYM@vjijxT=)4*s`@X|t%6<nW{A9L$*<FH^K<pY#PErb!?{Xh
z->4G#Di0z^z%~dHval(a;xNpN^y8X!=5FA}0U36I=l*rZB8nGE5*n%uWYcmjth;o{
z4l!gGZ*<(yR-5F3#26e*(Uq6XY@KKwojTd$s@Dg<F&v=G2tJV4jM_BMas5$?<3|=j
z6gb8Iy`FD+BL@1Cjs>*+)0~2UKz;`MG92-2B6SC*XI1oRHRQiSLqG6>WEtXc)hsKg
zN$d+egD)_>1@R!qL!aZcrrVP=J-0EDytP0~xV^@m$U?`&h6G0f14so*2kSJ4VW%m|
zxe)Mn=bvYThpumnNpn<u3ei<Pg(zM9={H3RxHanrfec;2pbkQH@0P%XnjnXOVjBWY
zds<tKq1(yp5jE~Nx^il-QHPiGG?}tKYl!RTddxR!m)a|OVt38&X;no8gnkLZ%v#XF
zmJ$MSl#mv1ta!#??4-}7i{#@+1Z}Mqw%c=|R25BTEU&C0;4yy6hgGCy4iILOIyh0p
z+uP9kt+v{eeq*MYaH53MP9^NXzMMopi*-E6a6C2R`>Kj$7nd<hFdLYyT(4t^NUZ@V
zf+_c~l03iMWI6uoT(K~%AvN19H2_3HqE&8)n#u7xto@%PfD?Wo7jlv%K*j05g5|7P
zP#a`R9Ojba(V<ND*8UG^%C$X8!zKO#5*SsHn(=js&F+<VN*00|N?YngE2-S1ecg@j
z-R6R@LU)Q{n~@-|8$##a9-LCUm?U-FWlG0%++H*yW=PLHw^P!-suw2lN`>a;ZG8EA
z$qo697)yxV7k%(9*T{}eg(rOz?@D?sZ(?FU&P^*t66^hXqb7St^B8wa&NG(6Z6yjX
zK1teu)zK9=@h#D*Oa=c{@`X{mTTp(DL~FEqFa>+Xl&I(t41}WSC&AGW1!>LvBNq$w
z)=kl7h08475dZmm{MX>E2-vSaVmQ)>lWgQLk3tHvrd6GF?GN*qpo1p=!yh+C{luMG
z?%K^6Ym=|s$*S^Ifcn$ASNbj$R1ej31C4_M10!FvD>Iqq6JR$lYWow*Dg9wCJX){U
zK2WgP#Yg9&nQ}B4WfHhey-w8K{vjQ-n^zL%Gm8z6SsqfUS%kU{WWmgpEPs2y_A}D$
zrY-@LoQ?7HJwCc~P3vd-eEAYIp#rB>dE3*vi)lq|u4x1Dm|C0$|NFE@C7tp%$KrY4
z+16ZEkR`MKo#I8Y8>){sJvJjyapvO`@h)|W1-|px)>jBdMNjO(Xhs|dFj8w96O_RY
zVOgCqQZvUI<lpz#5~cQr1818kD!&YFJaLE%x3Vl&r^T-d!7kzswMwhXq$flx{IASy
z9)bek&s7%wTa_EA22{2qdBGug*n}i&>vU#Um*wnIXspF>Yjh3%tM{ii&jP=R{a&5O
zPQJU9qt1F2K_$LNF=cV&BP6`ZF;Nw&G>|#?w%hWDxS_~iSflzww4qK1?C?j;zvBBD
zy?v%FeDznRub9Cnlj=~_d}BB_bQv^uuP)s=$9-OH_f_m?p)Gx^c(=5Vztda<GG+Vg
zB0#g@nW&WQ(ZBRcxWW5N1u$U`Ru{!n7iK7b(Re4@s_wcZRB+HI%7(?O8^6mmO%F%J
zXNFwPL{CUssEZkt1VHBhn9(wFyRdNZ=iZO9tzoga9XaZfx~iI^OaE~Do#!`TYAUP-
zhqEgBgq!+0e-CgRbX{u-d%4SfR)xN-M%Ao|5=u?ZV|IX-(c*v5{<emb%*Vg4{xiG=
zvq{IO)FL_ofBY+;!C7}@*zZ#TtVp2IuL_(&_Tm5UK<(KviFX94sVY*-m;~l`>R1?<
zYc41xiHtak)9Q8Up;9s*cu)wI_w`N?Bu0Q699Ly89p5dOXkn;*V0x2^H)Vlt_^w${
zKc?@h-oVH>#*p6z?EArH>M2iq10?X)x>uokUWT5YuJ0r~!Iqw;7;irZN|AvOXZ{)`
zX5upx4oyw=>h`G*VVMf;+}s_m>U+5ge$L&xnlrsHV5P}eOcG*Z4-g^0EpUsn>;D~d
z7gF70X8yhp#=|kF8Tk*7kQXxz=W4(8(CsRQP-BJ-&4|dvRr>P>ky5+^w-NuG8CmDX
zwv{$;-s=H`^0yWE_rFGT&Ek%|@Rmv|ft<(m`ni7yhk$TDMLO3FSZ=8>m7^PwS!o)0
zZP%|eEWrEd0nk_4rDnQcL+3r4z6+%51l9+%7g=__!U!M@AaS%E*V~%X;$ZI0#ryW<
zLey7ls3^AWaU-7)xnR3{lNJdX*$Q{0WiwEtSHiMB5$W&MA+H!N>0Ka!!bY~@1tF^C
zPQnCXzWU}df$Nv8-yF|iqf?eBZ63^1*u%yGVS64!D2B3(FL=H>j{<Rb_k#yvf}##^
zRhS`s*Y}eWyqD@J945+Ol(}~ugadH=xI6WdvbOFbv=~TV_W@m50N4`;jDf-_0e=o{
zRA2~yrAo{HqG3ufwB;L;%}@4#nYL_$y1=;lQJPEb^Q{-z$-LrLf_wDO%fmSxJPNM!
zvYD2ORSC&%Y-i=FCHJOlWyWej`nx~!l_jq6n2aR_b5gN-7Xxw#So_x`E)G-caMvFg
z{>G~)8!Jk^sHkQf^xvp|43UU=)s_5ZCCS^}p7{(!Z>d1c-txaN_PY)Fgt>m-`$v4^
zVRlq<H*5hYX-mGheg0&LjjsV_-aD4b`*j}@b8jMv*ak=F33dFXpz|s@(x_1pay-^o
zQYp!Lu)Hjp_vF9zt4b<Hu)a!ZieKKhiQ3YLU$P==rKDpzYQ?@P)C>4zPgg<!N%(Fy
z!Ww3Nok9WfW=>VMt4)(rwvHpK)aO*Zq3@A29u;WJ2v=+PSv7)|&lF|PxdFc$rD4od
zm9shV_@$QMFQxIW(pfMYMb)45Dqox1xHPVjJdo+-HA3ijzEDwaPg6T4pAeFQmPOJs
zJ$PHlwO=}xC`n&D_`yn+ArMG=mD*fz3J$U5f<ttS#~v!0g7=7D_ZULYhp4r(jb}0}
zl77_urOvPFNh}>#5H)|qj4G!7NUwr3TJf8Cn{w6baz`z27}I4Nb?xq4g{$;wDizWq
zan}At66i27Bwt@PBbE4Yb7DV+&Tm|hQ$9xo;=$u6ZD&gBL&3-w%9kT-Al@YyfA2qx
zl|*X8iM$oNC}Nq<>ENy8(d#}#w#=vSi38UPwiWmWs{D)FeTD#*D9I8_dM-fNlK91}
zGm~`2k9{>65Lv2YW*^Wi<zr{}{t&znTmBS|G85Mi^M#(>*Wa!^H@+A@UYWkc8GJH~
zcR1v%DeynM@T$1kZdSO}bj{*Z%wQM(`!91fb1dPdnDdQzu4`*B@bPofsK7}3`0d^0
zpqOiLi+Abup(cB$&AKY|v!-aGG^NWdDaa7)PY???aKOPYA(k6CnCd@!|0VeAX+OeP
zGSVQmRMX)@fG^x&=|Q~n^&V5yeyocSv%4!DWIPi)>nD5h9m;Ko6YIB)S(nVOG$t>B
z)z_d*SJ}fXeKF)v)o(0)YYsg>FVm|0SSK9SqSP+cz5M|T6haMEN0k5^dWNh~y1<aL
zk~(wGWSc|ePUQ~K3sL1<EaufPRc%4_dQ*$)`SWlPf7@SnY#iv6cZ0}rVDuclC-I`q
z&ISSgnuouC=v0?7PF!4A`c7I|vJO$zJJ{*OQ_o@aM;burHw>@tLdrG=dnK@LVRwsK
zzUW!ILS;lvy=!6um7w#4Q8sgGDybEC$}T}?U&NtTP~oD>C{)TdyFNQUIo}k_AKKb=
z)vY>`3q$jb+xuD%R0F}*1&SRI=G&9{8)vrDGi5y#ZiX8&Z1I#LK!!kzbGU~YS;o&C
zM&6!7!Ay2umS*8D#dWt0eaX0PM>}8E*{YT4`HzxS@%VJ|MF)StNZM(?0^<{AMLm$-
zsEN8zlo+{8;Na^1it`=AN_;D}ftP>*vKt|6O7vulX22+EF%|5#<OV8tXp73XDc`<Y
z^#2RIezPrE&IxCMaf1J3;Mx8C#*YL{gnw~XN{n%QZD4Lf=^^KRzC_n3X1l|OcimWz
z=dV@D^?MxAC5ftY-gn;$PU=sxud849dN!5Rh<Ss}{VmI0WAgEOnJrtom`tKEePZNQ
zti&?Q!(=;RU9AOA%Y_dt^xY4{trz@)5uhmYMXDUnhHj9xC-|avY<VfOaJjw##cbR3
zs8TnvW-$@A+VVzXPYHPJ79uh$;q!}&F-y6U(JyfNXAgujz8}VzP-vZ%zADG@m*Owk
zh1r`y4?-xW@aJVoPLqkfTd0PK7(ib-7nlPApVBX$=`OuS|G&ZRKc-%^(A*hVW`GHS
znk%8g59DqSB1Oi3<rGZDNwJ8@f58WoXv+4l6c4qKyorlI3Oau5)k|6U6;U+WsoXy=
z9=4>W8{KX?u5TuPEYLCHFa@X$;z~iNs4>5=7%@4r>@GCYdY&He3xfYfD3#?;@i*2H
z%$c_QTQ=hp)#aj8d^hdRAMB`2^!K@v4D&A{g`@X4+Hs%xKXVhI*&DuDD#<67XPyq(
zcue9ZG~~|_DX@^c>H^iQaBjg%W9cd1_L74De6?2|T`sFuO7Gb3wtrU1ucx6ryCogk
zSPlw}Mx^PbohKxW&Mz;LWNB{vXFrsFIM*238s1(P>z_H_KIPQ~eSye4RJOkihRf+{
z21@1UYS2x0DN5z|#ODSdZHs(d4!q5-GN7A9?0t4r<vgr7bS(M?{$@gzEzHO@e*B5P
z9FTzV{BK7Q_xoS&4u4*7zfNut>@MoK-Kgni<7rC+T(Z4`wq~F6KB=pNHR=8zL)ONm
zmOw+pwWUickZ6!7`+yifD&pMWHln)4hWC4yA+{n!lZOLs*TT+sW>uwb#JqoSCUg(4
zyB#z<PIS%5Q=@s(zcCrK+!4HlEY_Wo%85y8GrA^_geh7LQtFJp66hvI1a7w8d{-GM
zP>_`=5oQWd_#!8rJmaH%zyLYie<CAdI+|R-;T49?o@;TLf0ClFZ>}F8C1=+E1;!8M
z!`>nNF;z!>=*M+x(?`B#rv&(QJ!notcaquh$p92%?_>yUU#x2QmY4g*_=rE&7x*F4
zk6l|HCh0a($Gc<kTf$RMFYpt{C}FWX7QF9B<Y-q8R~Xn+Ap;xT?(fz7WQ!;4UJS~z
z<<eUI64j)cp>s398CmeQO#k(CQgK^lb=?geATH|10e|p^*1t3tV$7m)ua1D!<bX$U
zyx{p0!NOlP%D&dG2v0{JrrhdXrLg!v@gl=U)IP^84qG883P~kB;gW4<Z}nzpg@Wb^
z42-8hs_<kEY&;CIS+Y6|`O=s@n(p&bTuT`VH0Eu#!oV;NlH~1R4lJJAUY(fs-^|zI
z05XRIdm^5e56*9cVS#gdJI%W-WkGbl&xF4Vpi`Y?oW3}h0k$vm_8YoITmNcay?VKP
zuORy|^*n84d*F%9;ImszRi+Y0-!m&#+(K=2)Q2Jdv43Ph0YTt$l$SKf)S$JnskfLw
zgU~&VmX^+l!6dWHp-O^ybl2Qd!l^4*$od?bDz3|Omgvy?Jf<gzmWtNJq5_&-L~FK;
zlx(A~iq=ZwS+qtgn7vn&b`E?DEzDVl-YJ!X{B<4h(sXhK`?Hi6P-&RBY}(WdADNcv
zv4s*c<IuW}#?@aZ3;)R4ptZU(7ycjEVzzDRUW!z419z7d>+;M~EJ1#?FlY}4tg2Vo
z4wh8e@x3Ch3gd`Es3n2>jRh<A2%Ac!hXhh8<|nKeCD9(Y4g$I@ZX+&kTgg!R;u+qQ
zuK<h?dML|zqyUUmLz?7lDKj(2G+*|nD<}Iv#5IgyQBY}IU?zaIgk-2;aanu0$+suJ
z&3Uce$!hWbz~oW-`SeBSabH<Fx>?LDSmh9-@BR6%?_H5YO?$f|+G#22hmo(d><&7v
zIgnOOHM8$C-BOueeO=88IMCc#-_y*lQ~gBtxW1lZ(tpH^;#<;su7p8f8$6CqPy3!%
zK{Q7qkD0EufNG=TL>O&Ewp7I7UGHI|W!0C;!m2OqWfG!O{PA0}hDS#tTV2-74L*2V
z`#-5AC+QF)sZb~^txMr7Lc4d~ZJGQ@RjIP`hm3Eu4>lcVpMjxa(p5e6*b{n?QnM!*
zOhX_1XczD}Sb-=!Wbrif;lx$?t^!5`u=k(l02ZF1Zpu_IY}mM6l%B90Iq@X7i?K0v
zK|E@`faGw@9SX}tCRCFJpt|L(o!2sGQBKO2v*ZdtEfjshRza}u2GfV20dI$_4zns9
zAwzkHe~bGl!0Z?PhdxbOY=QVmZ6jC!p!kv9GOC}mE*4L0FQ<owNr1X5=0JiR6rPj!
z5xF7n0x)QEv=LE*Ot%V{QmFEB(QA@lnM&EG4SO;{LOG8GxjEyr$&D9m|54lfLr+cZ
zW8v@+dM0eCol$Y{t2D9PjESSH7BCzUtktxxn39P$hq%?V>HxOaUw3}!1lILo+XRBJ
ze=K4Q7OITTI(Q&l>JHsHRT9b!uHln7=7eW|iaG!TaKKLBz>9%0V9elVqHCKm!tJ%v
z#k%#j&bk7v-h_ydlw{O<h;wps%Gm(6DN7MKSpRS@Ya;f;8HSxB|6-n0ypt{GQzUu8
zc)33`uS8H9)Dxy?hH=L<F30?uchmB3zR$kQb=vp`tegX_bmPg=JHb=9B(OvgNvi7>
zriAI|bGR+qAjp249C-F#?1oa&TOHh{$;Rc%c3*Hv?$<$!JuMbb&BZ)(L!wP_XFcVa
zamLKTjs2b_s9XSyvc7Hr?Fv1SJ&PqGj1U?Udg7}>5A=txH>|gllWttN2VUhDP-;x5
zPUb%NoWNe5J6@!>Y@{DPJRazu_OCoF6`)aBtXdiJt_jf$5bE-RQ@yfWv)Vo6Fx*!>
zc{Z?Lu5Io!+XLy8rH*gqd3pc%_<luZ&G_#%Hi7A<9>ImCaRxbmDAFVh>kx-gHyTp~
z&sB$s#Bst)23Ip6#<qH&v2zgH6E;Bn(Yh~<uJ?*I^c5MFV}{xDyWO}FGHgDsHu$B*
zR6=*`O}u<VwM(Si(7{4b33cWTeecHZY|kU8&rj?}RRy9?Q<*;SVA2!f{EJ=Nn+BuO
zq?R@(22Qn?VORf2t882il6c-!mX4jN*Uc?lQ!1h|rVt}h)kpgf&#LlHc-kO-aY)(2
zqBtu&^OBvA_03ju@5n=LP{uwS-~Qym{NOGdl_sWZ&V{!0?b<dQUk=}j4<ed6F4cN`
z{eM`DA(rV49P+QP_b2C!ex2Y@b?=bd9g;UV2K1z!+&1daIlbvP03<0N*(-OLB(y(J
zQukIlY~4c7p+3OFp(&9axB`C1ohemnFK<HsfXySZIgOHu`FKfoTN?t%7cJ*3{)QCj
zOmJ@Tw3_tvw89WT*G&(}Y?B?2jZjg9Iu#vEL~cJa?O{P_!+cQ(@bu>n#b`+Hz=W{|
zufS5S;LhXYg!vW1wE*v30_~YMSW6{G_2Hx<gj88u`p~ym*)&2UmSb~=W4ubYf~flM
z?}RpSMmb|$#`7MBj0SSLiy5c8dXBZ{2n&iaiNGm@!hFUu1j!_dEdP}A<X1_KNgnaK
zp9wdUvOeiRsx#Z;pRyC0d|oPVSE~P3MO(1HCCi{)TJ$sHeS6NzT13i(iVYO_aWlf#
z{r;G9dQ}?%P@YLlx7c3KC!GG^U>auTNL3c}A?rJ3$)w%!!mtCIN2TODW6S2E!$ko*
zd`k^Zoak>%OadPHdxXvcss2H{$l~$$-;h(h8AK%$bY{FNAK(Q@j^~PUos3+HO+$~R
zS9{?0Um!t2bhg+~6^z)QlQ+Y%COJOg(9z{6IZk93RQI;J-ooJWnuC6(eBui#e^U+Q
zzT~YoQTJbL5U<4^JI9ivm7i_I=J%I$=jrRRlXi```d}ozy76wE?v05{F1Zc&qRKY?
z4>JV>vUw%p8uWIL-0*1%;sz`YPGBRpeR9OCA2nZ5oSzT&oYJk(^*6NLfJpBREg0R`
zXl9N%GGzTEZaG~XmGt1%lDoO(f+Yq|K>ja8-C%8e-?&oajwLI37lD}@t&c6|bDmIB
z#YnT^*-+4uW;r_I%c<GTl-19Wd}Zz;)I}<{j(6AlksXA)8ut%~tec~l_c+*3k8b{8
z8n{(qU}Si6&z4Gt9tUApm^ZD|viSh#EA)%ZFiyd|$Ir7}nc5qUbI!t7=Mvn7ekb)K
zKV5)6Xe~Dv-Pq#p9T<@tbpXoYY<<Q!Mch#y{B}pGJ1tUY&lZ7}z(G2Jk3Po!C#`4y
z*y;Zl1gvnBWiTdt9<;|>A!&^9MS<Cb2MBMwslqGoroiF3#e$Mxi97~{yE?v(ZG>qQ
zOWRiU6H$2~>W0?uHUdzu0DAhXup>3RDiT)=2X6YiBVR|y)y8ip33D#JrY|$Ouygkn
zY~~d4Iu{?GIqV-rMa>9naD1jnB?1-N_Vf8`55D4lx)n*?%*%e<gU3$m^VwPsRls|j
zW%l4^Z91u~u8s7`_T>&<Np`(RTy#APF^wO1jfzgi+#HvTJ}=G@Tm%r=W$A;y>l1U}
z{CrIIisb2n4kpMI!<{-IA1|J601YO(RTJh`OLc^@{6VUhTEvupOg>$q^-Q7lX(9V|
zFe#X-<8i0;Y@uSQrn#B&!N1kbwffW_aQ&NR#%^4p4F?SXa`Z3onV<&s&2QS(*49!d
z!19AV=pxV}XP}uT-vPg3Xgp(R-aD~jP+rDdg&)PphFgb}94(aivUqZUF&>#l3cTjs
z_Qin#hSZ&sD(qlL_|xe8LVFqiUw<cXSH{Evc@DArbemYuclH5O1CLioizvWK!ODY+
z_<J#{0I&Qzr<UWfafqYpZ5H2806iu7XB~j(m!yq!D-Cv%cKJ!`DV+`d4SqO3NcMFC
zFC$}D2dTTq_uodT<v6=~4oWYY^9Fh)L4=h*!umO4m6T7lB*xGH+chJ*T^crS?oG^6
z=9WLBhf38`b~%nXuktzzSC7=Xy!-`)q}OX}Hq&l;roe`_0U<SVrA#y`n7x!Yl=v>d
z%z;$2AXnR24-zFfGcEoh9Jh}W^s}6{{?mmbp7^RM$DD~x7`|;vv6A4cEcutbEwL*x
zUwelf&Fg<W{@qyUSY-o#mnOC+p?7C3td~V?Iaro4B@G9VM>*rxFkS!g-*m<F8dFM^
zrcKvdwsgm(pA;{o21K=rt>Z`&B!@Ut+`Ff`jlnPvd~A--DNF!yUa8}4hHHFpMVw?t
zB-p!kJQADoQ2DcXnu#+KtFmnqr_1oWFzfCPS%GW~-D95t#RL?=6GT`q7GLU?LF#$C
z_V(X+FtguW%rl3IN_<l14xIA+!cKt(oWOFEzR)9{U8VA4rO8EeS%)DtuO8jT2)V>_
zDOuBx{}y<R0R+Dt`RU_^Ic49P6LF2j-a!b%(z+<HmoWL&DP`r%sw=mte&Ov<WJ;4*
z0dZtQu%hxeF)}eYt2BT9`^4wX*AHmLWJc~PR&S+=o0kJHdIy!b(m-s%;w0u`MDoER
ziQauIxv>V(F=?j8F+~oCTAzNEG{k)oEiGJ3!Z2#qBTNR(_%oO>O)-IJ))KxEfZN&$
z5c@pAWWm=Z!Blaoss|P5d`}|DKVru-QbKOcP3CcX0t}Ff&urp_EDBry@yrp&Y}&m{
znWULc3AxKppmS3ACSEuKE^)`TbXqJwD*a%6{>=|0O>S>)w(27v>HM$v{i#M`$_r93
zXLxI>`o3;-=l_n=i;Qczp^T~u7t>0UKyFXz?>3hPPq&-aQck@fs7L#SBJ7a*&B)B$
z^>tJaA6+tgztu<2`wg}gWUD65`khVSP*FfR*CtX#J|%Q<!rV{ldH<=&ha5jL#ik%F
zHvu_^)IQ)3)0h}4iFrx&xipZ;<--#8H@DB|&PvtqJ<%Xv5a~w4{76k=a{;}F8I={e
zD)mOnidik1To!x+C|6?AwWW<e=Kx3K+;@*Z(aOqva+tz$DcRXKzBt+W9-C`gizJtL
zUe+wGy~)<tMUWg-pu^i=%pJJM>`hB?waFBoSl?%6ZbGjyn*rbLwYYeH839Y*-e=7=
zO^Temnkd!e`B-5ldpe&2x+h78uMiS|1UcXL<19OvtNf|k^7+|RrYUGNsXgAf<BRt}
zm2S#SV8pL&Xruut)qMx*legFeKR5?mcDZc+swwmN`=gG1ka?cwNhw4mTUK+#zpimo
zY;o2;`&X6W-t++3K})eD=8l}Dz&<)-PWAmM{xsRu*JbAjrZCvI>vyRQ&q{V;h3&gS
z`0q=*z)8^JqnJpUQ$cnA*9pwNS+@A2oA{?X)VzL^`fHAqRW`ItuIQn*_5!{Z85th;
zUGuFKdOQrMQ@{`Tm?ORXAgwSF5@O6fpt`N{N6zZV>P92uiPc13`Sk%I1u+`^4i7W8
zB$+cH=GUmsT*)0cuA=D>XWyxJ3F)?HPs=LV^CErm-ak{3v4?)Q#MJLFPa`k8^n{_?
zK)WYG<SAHsgkk6&-~ZH4cV^7q^-`EOKe|P{E%E}{K#Vg8jHb~<lm535((!PuT!P|F
z-(E?NdSA|axLkK3K^Jp#wdU>4k!;uv^@ACSl-(%9uiQDQB!YzPpKo?&^eeoO3Mws0
zem}Xose9Ru)mvjqD)kpz5Bas38Ly3F^t%OAQ&uEPcB#BYR{R%FK#^2Cb?tWKEFBH&
zs%&FWJ*pA%ruZ@T!+~Gl^v|BU;UT)>JN;_;)>0IjuygLJHZccsG<lbaq_(H0c^8^r
z354Kz+hLw2hrWM#JN1!aV|+y#_#I|x7gMRv<`GinEd;2AA=$<Tap%k5;D}^_JkC50
zIKdZqMxxUS$<!SAW5r&AH^d_cZ0R{l<dwL7XKY;2|8n2)%zM^$`ZD1fw%N|P`~z5U
z-;!NzVMLulpTc@LHyeSoS70!L0~_FoVu~<Yc(8YR1(NzJWk7qz-m%{?PBrIFY<ib7
zblZB?mLrm-!``8s3(DpBjV~5^@8Uc3-LJi}fR(gjmG2t=zNS98n>DN?m5jVFa_-Ha
zoUeJb!0AK$f+<JiUm6dXc1;aIJ^6fCg{_Sb40zW*g1pL$FD||Qr-2kVC$tW)^E=5S
zC!xg1UO4#G=yab3B_IS5m*1bawBF~i>_$OBOU9WsZOycq)Jrw3ttCQhq#z7lA0V4*
z_c4SKO&<a2&ZsDU5H`;7qU&^Tc7)E(39PHbQmdghe(Wk0?!M}W6%?rlYehi<Bk^_L
zq2f`SXjK``dkU8~$@6OqbwF&xl<%TbVf3!!b#NxQkMA2l*|h;vp>D6#-&-_0=yWC-
zrv5@je06|Tt=lgi_SWcVXs*mT3WB1Ahhkd|lrtPO`NDhdelpd_qikw!C0<7@{9nwE
z;3`TRLKy8cP}xZOp@~joJ@7hOlvddOy`~2<FqA0aL7*?=-DksNc9ze|6&U!9qv5WJ
zS@s?{W!*@+Tc6|U&{4X{8idGXr2o_jOtr+jnyX(C8i}w2W6oJ3Zu<Z%zFO(4{=i}<
zl1~4znOvwVu^>^z;`6E=Vi|C5`H<HBYa3O>ON?NI4Zlj+XIy@#h3%9aiDP;CM4{*I
zQ+te@T8PI8((rUQ-(isdq8*ok7a$f|DV9V|GsNG;`iNnUL+L{DWxj2lvOM!j@rhu<
zY+By<O??sM#w+2`^mPTsKCip#t>CO$*9N-~o_)td)o+-JE7Z}$3JO=CMm5d9vuni9
z=EMaN%py~;X0b)eJ$d&t%3L%Xa^@diDMk&zl5b9qm)uUrxFtOpL6%)0P;F7m9$^l8
z(La}<2hd~5%VXo&uxUGSZa^eb_&U;JLA_ZsBTCOcI7BIMw!aA|*gN<7kvdovUo+?N
zr^(av71{Wji032Ej@{KGPgHT>f6iaMSPC?`*PB9$;MdTfG#7>+dwu{c0Q2t(pET4=
zU}Od10_D19^D;3;bxg}>ge#sWa=moDzg1ft71JnC$F{gp&jt&!h}=4LX&rQc2_ugC
zQY2!_hy;!91UX7wM;o#6jRFPYQfli_N`Q0dP*dGNHFKXO46u01kO=)4j~2+7`zoIk
zj?Sz$Qb!T{S+}oPl)tv)RPJ#*_>T8Q96|l%?q5?{_-@b#(IoE4LS?tk6B3px@mGWP
zvtVtPuJMeJ&e0JGolJH~;^JpfC{la1<{SpW!U=1NNIg!Qw<y~@1hxOA7#CZLw^(RL
z*Q^p~sq1<C_mKNiKCEzdm5`#am!nJ9dXM&c!@XrEj)Y00SzPMs5u@Y!8|a4B)IOuK
zy`@%zK83vOY%*H%jcEb*E@cj={c}?LN6NXUWglRP6XNMKujKy&)<7x0kKZ%P==Nk(
zl=v(7w)$Kpe(=_7@!BhEQ7pnUqfk1Hjg9;9;fL?X-X2!JdFSz3U%PI;6+RAO3@t_Z
z1%HNM5Q}I8VR*BULk9sj_md+GB?v+-JdYX(to8o++G4D(T@#;KN7iFYWRk>FAKoAc
zp~(w!bleb6cNCpgPuDV0ER}5q&0c`IP&l2Di$irJ!?uhCgASJn|E~HFbYj_$F#Q4-
z1Qm?k@Xo<1zSB91!-H*I#~WlAc%ES?O+FHY+!jb#6cCZmP3v6<J(KqaUQj6Nj*boj
zi|B<VggrI2p`yDHdWjE)uhn4>qabHVC(imUd*!s7NA~hz{O~www-DU6-EKlHqMY2_
z*^K}AAO3y(<i~%f`N+n#>(`@dA#2Iwgko}dPT!Iui{$nw%n-a`k&W;Q1p&f+a%-UM
zaeo93@ac>JN&`YuPHVq4J6nm2#*emR!Hu^~L352&H=(MWvx<H_gk}^)2<%b(;Ps?E
zmG)1(0PA#eHFH18h}k0fnX?WjNoX|o282s;7)Ff^3Yba-&mTPIP<o|$()cru><5RD
z(>usLf<kU@Z_i4FW~&w3TN~c5Az;oqOd{>%dFFk*FW?hJ9ff@f0`Q^7ns8>!Bc5NG
z9Q(Mk_c?@(MDfLb4dp6=Pv-|yvrtS91Ve-iV-H05Gz9ZjL~0&Wh%DN#qGwox+SyoG
z%Vpk&M-_s|vi6|lfSQ|Mh~<?veG-Zjp*5E9%p31yO!QqamBCXmPe^@~QS>#<E$b+o
zn-Uz7$A)WoC~*ysh(YAEP8XK1$Kvw3V0Sm}eD=Qjl-3Av=zX)+56>sK1@9nu2cDV#
zG%x3;&1fF)33iTkM-m285IQ3iE>}x}%YoMLj>e^{b$hP1ke3F}ndEuNTo(!@!`klN
zmc6p@_`|Cdfj?pAdY!HnzKl(obO`qJt6355d^n9=d;YNQ5xff^)jB-g@_Jtye?n%0
zJEjQLbwWPqDJ$=1G6l_7Qj$*#u1ckf+E%u(-OB5!8Qs~76O9ug+;}(M30BZBub0gG
zsoGYES6^C>*WY*}^5u1nK{f`x6e148T0GtGMns4`HV&w2U6SAIq<IpzKD`}#`zN}N
zyaEVO@gE9b(w>;FSGk4Joi2~gJ}Ry&J;pD8_uiT>KC6-NcfKYAVb9YyRA^)W>@pfh
zeZksktgbAG7A$*i36qAx)sc9}@29aqF_k<(G<M``6O7rzR&$Qm5jYM7)y_7Gs+MS3
zE|!*7Vt!#kZ<-21HEwNf#@6PB!?S_6@x)`TgQr-}D7LtE_wMcZ<l|4{-rWarqUTUz
zIqz9d^Deq}?fT7_tIg^BxoF0IY;SK1@0^&<E$V%Q<K>;Ewfc`pDurp(n!214UIJz-
z?jzx2p`jRY;~kCC9ECAB2E~4_-PAja_97!9ks}s?yEG28nfi|!+YUL1QlX%}WKdKw
zm0JINkd}jdTPV2fol2gjY({W=B0AUhd|N+^%qvd_Ug0~fm*H72di|uxM-GQ#1~}f^
zPRd4rhzLRT`f}74*46*A4bw43!DA4f0OfG5fLAs75TV0;^$+}$qAgkRg*Ov<5w#{V
zq7Nw4@PvgAfrqHo341!{Hk}DSU^s!1h>g<0gI3g@d#zCy;Z>nwR`hbfqhx2bgoU&8
zPUdl_xq$9~!#xI5Lo1=-9iU6C=Aq}HUPGZbr*>lU4nL67JctJpU1A;67vKX0VYhv3
zg&ZD%#aWb_v&L8OAoeWdoyfq1gRFQ1LLeR^e5o&L2RNR!sIPjKJ|XXdPf4j?m|OKn
z{mW(v^;{4=*4WGts#%vbcF@wYU}je9hdIB5k5C&?>fyzWw;Ud|7zYu0lJx|B1E%nT
zo0^%@Jaz>84Z#3;BY`74#RMb*W?<>^GGlM|g_Uc1kKjTuP8&%b1FryDgonr#MZ?vH
zL-BnZT6@3x^>5>E|NGy^`yYIwHQ9|m`S`$}yqiwDsLjA63>}KEbZT^r?->_lkLmOA
z^60l$fu2h^@WqbrO8I@ZMYA$dd`-%vd}~FRtXwTeL7N{GPeV~<E1rZ5A^dhBH(MdL
zR==IhSOnLuxY`6JBLOxtDGoF4ybw4xVAKKP$I??@^;^hQ?}LD{5g@#xLmQ5mETdg9
zD`qnFA3{3lcB8R(5Dnb*#XM1F=PI#$eLd!v7ag`_wj^d;IP08EAOuWpVZ(qTjNd56
z90rjMo|yQGp3lz~;u1ooHu-(vJg~aX@zcM*;d1@vOYv|1^}mW&UwO?|B99bH<B}L;
z4lBNNMqhPDnh!R4T;nrsfD(HSd150r{!K~sB!cHYi+XYg_wlUjL13VBZU%1{6G--s
zyjQ>V$M-YOB-h3A3k&2)nUqPHlz--BqWHR6Oh@zw8uArOqDe1MzTj0{=15KZf=ZvL
zE8YE~4S7_)@Cmi!5v6`g`LeGZ^~=rw%40tsv8(0t4f2^6o}@2-|CcTv)8~{gx;`r0
z^RV|GOZcMn_3H8Hrg5#DNZ$$Dg|cIPeJS34>*ZKmsfk|}KYe)`&6ATj5g&yx1mVW3
zue}+syz*u&F0DCo0fuSf#se!}F>U6pvE-(DghvQhQm7yCWi#-7;;+TW^93U{Jhc!8
zAt=KTjQmVIOK2gi^lKGzUoAzk0RMSmf#lKgu`Tj>24z{LQZpajYM#W-_PuDgj?CL@
zJ77T)LKuXJwmP@)ln4(Igi=BPi2%uA)d+hee!t)AJDePmAqsh{=%=*TxU>=g1qLiQ
zhIR=4@f<-(*K$6urt5}x#VWk0{X?z3pm*z?D0>i8B`#S9N(T;oR~-}-D@lnmrMk$I
zGYTiAHZrd8v<R$D&oARxbHdo8^jcZH9+fINKW01+4lmj7YOGqZyZazE@87ZljR+dl
zN3wq`zb(k0MrlF2Qkzf=s0%S1kk_m~+9T<<DRP~)1P@0RlwsLDTUpeFk@uUAM}0aU
z5PI<37WbLloG$0phjZtAyEyM?U!}1;8w9zs$R~$@w9~=TKIIU}WGfcNO_4{6+65!*
z!z`3@TgQ;Ey<L7vu#ok<;3aq@m@T5H(44YIM7cqjz-cxLflw47rS^%0BfL0O{{#%o
zrM=bIO!VXYv}0i_<4*gE<(l4)vIvhK?Fn^1W6eI3@nDhBPWF;E*wgo34^PDm@>Q(~
z6k|#8<*krmfx9s9)0#z~FBkz%XUmnSRZ|E+a#oQuuRgyRggz@fFUpMPhVeW<?diR0
zhlhpsQ>a%k+&k1*BN*<g0SL7<mYz)A>~Ulad+G;*aFjt6y}uSGN1O5MU;cgU?`}t4
zYlm?Dj14<rycDN}7cPf-?yTp1wCf!82jw0iHBlI+97w_gS_cUCGx@58(K8zFRyQeg
znd{kdU3=iF_v_@V!SGZgVudg3ll8r@xURkNs-7<d`EnWugb{5UU#zX)jO#aEiM4C5
zXslK|-=J>15$g+UdVj(l3IE`gh4PO6A|2LVKd(7LsciVTu%Z%NLf#sL@`Q3N+Oq@r
zqo3fEUi3SMZXff6@}D*nI+OyAtIaPsrx(Ir3$-;iDBsxsS5+tYh&AbGHTn)5pfJIc
za(Ay0@4ov{Z0``-RqImM*ndA1gab1io*RzQISsh4bIzC2eGjkeu(FK8AdP6iqeA^@
zUu|Pzt+`~}ilW-Jrn~&?*IO9<R!cZXI2Lc-3h>U@7x9>xwIKy400nBHnDg3Z{4q=b
z*Wnq742bywSEBqv2?kuDK;v-eTfrpo&=!t9Iy{QKy#tN&uD!sJ5fMJM-9{1AjiW{*
zwzhWkyQ6j#9b%Uds*E9uxuwMg;nS>P^`Nou@Rh(0awF&&3OneAJ&grp;9noYcND_s
zGXE%Upo2^eW6T&pSD?+{7?e&ZahI0Yqg<_P-BgX6p_6zPVJy+>;7z%qHt*~G8Q}*)
z<f0@5uM6gZLE$r$P=p$W=A$%+=Aod_kXTUzy@KX3mX^KPD+!z?x~4HEY%xX#7_jsQ
z!k>B``h&-2PW_pl85tM|KY<I$%a{>eLq>#>*W-^;U1I{>HBJ!BpPfSEXjfXR7)+oX
z!XvI&Ag?BCHjNKnpoHaPEzJr)fJ4aD*A$(jFVG5m8AB8pRCNwR5Rf$M(Mlz)M-)^j
zy&YC_GzTakQDkEjgO?!hz!P<4^@d<_LHI=dp%bbM*n+=ceG@{~`9S*}k1w!Bh+{%2
z;^pj6hI;oo^Dn&3+QpMwjq~?}Tl7oMUI>19Me?l0ph!msqXojx3WuS<_c~6V)q;3+
z5gMa4sy^dM2P~ruTUc0%`6b|HK2EeI_Vy3dKj9(X34MgN&Lq$If%vOa!7?F=8GptJ
zP86jcxC{OZo{i;CuCoFje0ra{Cb-GydRzAr!Z(+U{)JmH41%xGXnDQ;$N%#`$4`Iu
z>-hCM@5Svq_qA@t-{`8n&qf7IL+>798KzMGeAWZQ)z24XvSU2=tizRkiSo2Mo_Bnc
z<!SGG-tl!6AakPlnv_ZTmI{b3BSr&DtYWbsq^r%f*0UG2{=<O5gm4ZoEFU635E86Y
zaq0DM-D8EB4lKfkA}my3q9sH+rRT83K{*A{v_*tB%-ZCLxWNDrdVv-RrI?_C4B98H
z*xTHU!|fdrjjoyFDlxm)*P<lEPl!X}HGAVMVx2-|Tt?q+%cmkz88JuN;D~{*FV<sm
zd0EF&%q`A~$Px~*V#QgekPY&K+4D`~L7S5$#MjS&j>}6gzZ(Dc-~L~;xxS&z1fe3+
zyHMDTJ+V9u8f%RpSM-}hW25$^O<p?O-t@G`&ky<x2WR-ruXK)A`Ii4bd;irX*^%UV
zg5vJRIdvW$Jiro2PAW62x~glsrh4}5JnX#Ahndf_5A(EpPR-fvKE=C|A$cMJ1a;1g
zQ#UhrpZ`zPEEobL03;GwWg2na%r4QZDk>@>DokDAmH>ocp?X|$(KQdvX59QHy+fam
z|8Y(S3z*ZBV^7aRG~c|;%e>6Xf8a7Ve7&%I^wrO`eC(^>f1j{CU6-@Xmw0ws{E3fw
zF5Kf^KZi1{Xv(9XF8umf@O+@W=&8@v*wEk4t;5fhS*;&c`zMy^o~q12=`fvrK6iLQ
zy3)<F`~3_5No!=byezr3nzLjaCF|Hd*YS?eWD9ov`YpSrqg+|>w09`SP;QOSN1i&b
z+r}Gyq$<h6Ky*=XN@L^lW~*VxlEa7?YwBKPK^)FdiWTybd8<_;!-9KIOI}7kW`7mZ
zN3mxHN<NhR>^p_GJz_X1q)2nMzP{yN^C-noHq~nfdakc}$Phs^fI8-#MGlR4DncTE
zFi_!<?h!q@FJCBIuCVL|7YrsTWNp0gJt`yS%ADh4Qe?5`6#c^pHqd$k0|-i|vG2)+
z@eIW*N=FnyF^o4Zzp_rznnUt&FtE@jlwOioxz64sOBfP^H$O@`?x0Q-v1d9^l%q^7
zlveE4o!9MEGBFhLdKRT!*eriMW^b;rKNk^OCr7(>dc5mOxujr=0uUTP0f(ZH$SVDq
zwueIdv|iG<(l->57#N78Al(*9=7h#0nbvz3!2y(2ZgqL;KrTlqN#C;ijTb*K?AhM-
zJv%sfXsHBI3VQctSZmQPqFzucrrZFT@piRN5A5XRz;TJS*iwY`1%@ixIfJF}PVnb&
zW1GgaT3xqFWz|Y$a4X}HA4F)e7K_AkyiRxqjL-ENj8!wG8<97H_pswB)4*wr9?@io
z^a=grFW&T8yXjph#EF^!@6`#xly@Zc9YaXU66~3xarEWO6uzLPH^S#9@FKbbgAA9$
z+7-M5H%S`>9!-RIIRD`QigA;al)YZtjbCA(9pHc=Z(^f<-Hs2~W9rn#!<LOShNR9^
zD||ggzO2hj&8L3b@i#;t(7z>X)OYQdKmS+XAEqq)%4LO*f-PwW9S^!q>$U5ei=<&=
zZ4Dl4T+m0V4z?#C+JF>-R7vKPSd{3CU=$(^YQ0voQYmj&u3QVoDa|=>c68Qr<t1Q3
z$<A7Qb#+7QV2rH7B@9-YM_sjHs4*hFS4rcXE3WvFOf3s`L@B1MR9zEIsi<8Tt<VX0
zGxa?Eoc++4X<$NN413@WT%k@>D27kglHe0CN8!(YdyM^{TeG%civb1NbD{M;QE5>k
zR=kZU=}}O}yxz;nP<0W(7|~nVGwn>{7NQIpV+`^bOrZUIN8w$owQYCzNc6SmxF52h
zV2lXxnXw7l7)TteUw!!5>3PO_THJ`ijec_|JTd#nfBs<bLia)XS7;g0BH2vVjdd7a
z6KXH*1RtkO_}t35Dm<-q96xYyXlvdQF@zC<NE!B0nrLk37j1C-Q!8BWpy$CC=sd*@
zsTv#f21MbaZ=u0N)RoF5@r|s;gDBQfNSAq1)4PV652Th026EEscDr47fVg(`s@=G8
z%?ia3^#eU-o??LRP4`11H7Gm>J_av<S+VaGdI)%$?=z`(j6Gc=`^F_5ju;bYJAJ~)
zfB^*@AB_jj8-uYgZK*8Aw2a0lKIN?^c6_++$53+<r-r<EBfN|L`abKT85nOcVh|~W
z!Nd84VB#tBpm~BL?};Z1YDY@#z7TEbIsFSdu%zp<h(II#YII5SLTz`WbalhhxvJg~
z54>Nr0;d2BF~BX6De2z4;G*|v9*~ZasL%7Ua6JL7XOBeoS346NH3xJL=|34)TDKJC
zI{bjNhm4QtCShs9V`w9FEG%aHe)eV@i!L$HZg|%xG+32*yKuH3jUy}%dcfX=tya^G
z*VGjeY}rE#Tp~>~#u%cya)pXTQ#f!$eT*GXp!I?!^BbIF{BWoMADE)R&7FLlA3_|V
z3wp8e3&#N8n@w=UAgan7hjC7T#!^3c<PliTU+FXa0y?1QICn9$e)G^8;_qF_Sis)x
zXj7R4If^k!W(t-rRy0NxuVOH0t3{!W5D{?%^<l~3%6Ow6B^p-{bl!Y%At;Kzoe3xJ
z-o5K{s$7!1R9v>0WG{?to%V_F=}h<(BFN}7#!`6q&h{hCqp{VS9s9-4e`A05cmLhq
zdH25L%f99^x?AS5ps4QCXD35yD72LlXp=XVA1ZXxzdTdIS)Es(X2t2<GlkzDe0E+w
zx*)F14PWyzFJDkW5FizIu~ZP#XG2CU*k)cVBCNqvA`mj}B~LfMAo?syFnn>|uj?S(
z7(oD`6X7ZB_Z5s@c%YI$8is}7Co1g16W@bCH6^%u9}5sD_37b}J$m<jtM4C)XPt`?
z#I3Zt;(L4%$#sbksrDk=u+T=(VyQOPZwNhJp}d2GFzSqi4}^AvOoUA~{7w48A?gbV
zy#CUew6_HT(9`%mq4ST+=Rw-<y!M*?)nERVRaaI59Q~iO78Siza1G!`slXXs({C5B
zVPM?IZ<w_&EW)#l8J|>vV92wa34$#3a*mV1<A?8V<bpv@VUHl;#xR|8k2IL)v`W8I
z?{FANcp>*To+ezImwB0&dHD}o{)7x)Q|9rZ$IBIe;sXo-<=^z{CoUg+;H9eg$fm0%
zDpR9AQ4hIk`pBo~$RDq`U=Rx5W(R-Z?=RPr$4@_1K0Pe{kg%N9{i$mY75G=mtiD<Q
z0}S=|quM{cG=o@B3?O~6A%Dz#508t94TeL>O{8+W^b~9CeS=Krsl?AnT_jnVNT*KA
z>a|l>76;FA)pT+0ibj<4r2O02d1MbBJg`=~W27|00~t?u<Tea9#bU|W<HzN6$#J9=
zLJmj1V{Z(yJdhd@4}6SCcu^C5#@;R-ArTSIpnUAN+}k>*bwZ`OD!Cbj7zPQV&?c&%
zs1=E#$nvgmQa^NH7%3L3lGoY4E3IenP)FXEq^`9A1{tEc_V#vM35t^y?P<5`*6lTo
zXfO7uVo%&caoMAEkUfdiB02?SIz|dMnHL20JjNEm++mOMJ1D?IK~RpUEy-yLx3rJP
zXf)CHoZwX!4AT}(vgZ)jE`9GC6l&~c<$Kk*EFACx#TDrzS5`KyfbuSuu#v`wF`P`a
zUJ`twuJxxas^=2rFdpYb^9W`O7z4sySEPU=l8QPVjTkQV8-;#Qum@aHqjYYlZy5Se
z+M%d3jcL>Fz4wmo@9(NDMN13r;0neEA{$VYvTq9eF{@c>L*Jv-4i2^*H;DW|q3i}4
z_0jZ<rwY6n*_h}I;Tn+-q<u`Mi?)a&Gn(=>UZ*A8IO_{9*43nv>Sv6#MgV6Qh<f9o
zoj4V84n<#IxHGnB6r5y@2u{-Tb64zobb#=dd(Sl&P&^NPKNZ1M?<ATbmMp3bq`K0Y
z6%eUo3%o<0Zm1Icz%imKNExUhw8<I!pA6O4m`AF`*&{}EU7V91OlvO9G1iidJJEsB
zNK$pjgv)}VaFw)IgI?Da?#xl}9h$)0<vF68Q0}ALY`2?1@rsgBZ6Ph@-p>2>n_vCd
zcDEl|MtI1+X`~c8IX<*zz2^Jdc*L1{;`;<eG1_X4tf6<D7A~Hv{`$$0=1b2FLQ9%c
ztRWXJhIaqn?`-$ceOsxP?aGy_9>D_bAZmv6r}Qx%Puf5@N_#Opp_~<Ju$a_*XV^7o
zQ1q&A>`8`0ERK#Cy)o1Vg*|u_l&}~_*>ep8wi`Q9_Mi8(X731xHLk+p3ylW`K8%c{
zO@z*1+$ZX_-6XmT<5h?xfo_mSpHzpWo?|YO(rm$_0)3z#AMCf(@1mH-FbjRW^k^ft
zWQYoO>>ozkfA`y8ISh$%N~g8{&*rVJwd?-=wymvJbS@soP_#@C(4ip<GtuGB;aO|+
z4(7{jP0g+0?x%{*qgdu?Q8=H69_^$yI*#c3EOi_A5}D$93Jn6rCg?qe74$5hcft5|
z(j3^``$u-~-UHj&-nZT~;wqC4PQ<NNPjunL>h-o@n6{$kJPLVNylY-^FXQ5=D}^u^
z>^tqY!=qFxiB=cfp%~*ADGU39k)76>8Y}o<TyskERR?>7!;6Z=qFuYTW$Wu}nwJ^}
z&XZk8Fv5@(A}}wVcBRu9y_fy?0v~9!n$9=C9qOhHsBPG53cPW8L-b54ul^BP8gYI@
zlrA)dw6vtKJ=3`GY>ZTc`eg26&?mYLJqWl|5)WEkS#dZvnhn2(zQi-~K;|*f2p&RV
z;ZjIRJP}<Bj>JK);`eZ~umfBGPZkz1=!t(~d=-5IJ{b2&CFs#)>@%oq7_@+)vlB3o
z_c~$F|4ZtE*FgiB#DOl&qOwTKhv5gK-@ti!TVp-edGv3>ZPgtdN;Edm9JQ6mI5#w5
z*c0AhtaKRbEzDP54ecbQAbiTB?lfLFf<XT<))1YI>_c=M1~}+2O@n{Y52B_Pab&Qh
zzl)TUAuT5DTC|bVVIns*KZJvEk9u1YPLLLOtam}9)n89(2u09*Vhlv%NY#r$ov~zJ
z!h!e~2GYP^hyX^W(i3VY2E7Z>a~wQKkKC>w+VSC@)#|4%uhD3XcOG3w%G+d4auv~O
zz*S=rO*(F7Mbk0Zbvs>`i!i8Oo}(+faHpGiI_<Oy`2`wEv>A>*>={P-&yDqU(eA26
zC5I4s##lA#$9i|d^5s?aE8{%oVE@2=@{^z0Pk#1Gd+_MMe*fM>``O#Swr0Dhw!&l4
z+X-{QJ&o~9*B9iXY0-DbNlpn@plRWsUkaZ((E0g4-Sc#LrY<kxi5b28w8ux47v29w
z9nY5+-ScS<sKMOuH81n>MHSI8LZXvm{-h1buz|LSGXg0BB&6_MOzhkjP<Ytk$p}#m
zQAZx(rSmXqzb=Lj6Y5LQ)RcCQ9qxr*A|wzxkASxTW76{;Z6)CvhN~N0fDm<+bf$NT
zfu8Q~+4lSQY|!fXnl)1?+WNICwj$=h;&mn%4#eEph!-YA2sA>Pd+~@_z?{h`j_}^?
zkW)B#4;>vG_~Hd`qDH-G9WiJa6^vr6{<0wTuwo*>R0*jX7Q`&$ELOB&j(C`{_@p->
zwByC)g9I~`E!eld^=<q94}K8RH2@>vtxqrhF*ST4-0;tt7Vv=)@yWprfp~^Cgimh`
zm-x_qeqEIzz|D+SZqRyMC+ARDgmbg&EZq6@D75K2#N+)F!IcO((x}vGTBJyL{wOHS
z%e>6Xy!<;Ze<Fr25pXBsA93*U5(Sp+yx~(bfbig_e|}m%VY@$R_d>TmYoosSG95+z
z34Gg+X|G?dH;-$2s(i49*}Fbb;lA1Nc@{P6ll1Mw-}*fM@>|mFw4Jl&ei%ADeLBns
zzDTxGL79kw2Zyk;!N6LQXZyXrr<|%(Dn{xdJiReW@jh3SUXYFngV&;!iljQika6Mn
zx$<g2Dyb9O+u3&KqssD{z5d!;cJ=yoSLoAGlt(DGsT27ZBNZ3elQ)&jxLn(A*Cf9k
zyP^)IaG_9^T%PmiF?8XH-0L;1etK*~iE+J9sv1#g$gH##BNx$2VZRWR$7deV)9tk-
zBW8>Z`SGL=`xODmv0aSXHauhRsjhXq>;Xdr8j4c9?K{?N9@*jHBOUujdbR4xhT2!s
z8icfZ3zFO0YWI<KTBmk$v}2F%zw2q>h!l|cD!8FsYP;9DEB$bZ^Fh?O3bqSsH$PDm
zqKqWcE*ekkI&v_I!?=X|i{Kru=keIb2tk@5<bD){#o~%pm$!^)wpb#ocSiIG%0D6}
zNaL4Kzmuv<^<54<#o=fWl&*dC5j0_}NM)5TtO{0H!7>=~MrQ-PABCUVgJMtXkMr}M
zvA<Whd1x29j>0?`#YQ&lHytJ@^)Ys0h{SL>)EM@OB<gi^zgaRKQym(g2C2igCAZ^|
zpR{siS?j>6+L~2;c+#sly!75NDV&H7J9K5XM+stJT4HUawU+P~C&M^<vMjP!7;WO9
zhrugfl}$fUXr@(9(#NaUZ@Yn)Xa|&$tSc55QIe8=DkePXTfbe`SgP+AL-p%I_^WYp
zqYDRXPZXFeaWtCjRkncAlQos_OnAMBL2dC|b@hbDM|OI8;9bMWnaLJ~cV$awFxV}5
z<Q416Otx(0<qgY~);;QhwNM13p~e(kBB}^jptO&v?~fy91Y0*Wst%$Ah!A8vYkKeY
z!}sm(`@gom?FZIvHyu~=Inpbd^}B7&pOKG`r=Uv<&cexbuBi9t9Y+}pQZ8QT9o-g6
z+=1{XV=>J!*3qLu!?y3gW1Z%yt*@<EsZ`N<wN2|-Qo!{!m!t83Z{RqYS%Q&esJY(J
zZ*Y!vr{JaefdQ9i*xL=bu=korT?n^>kpRW5Xu!DZ#yiFs<tU2d(XeH0_U6+u8e>@R
zI{hKSgMIm8!h7b?$;pu&>_61JC*w+1>v{}bMZNw~bYQ6O?Bl~;gHPiYxJ0?oebB?M
z@Mh8V53FaGV(hV#P+yZq1Zun4w0n1dZ)>Y7f_KSbSt=F1?L_c>^P69>E1MfY*BXsk
zWSDT9SLhGxVFrow`tNkl|LK!ddl+g+;mV#ygMQx)F&rrUyPb|`EC&(3O+PTWH5v_T
z)*GVXL|^H1dJaP%+XqZ|(I2o-dw+Ooho^nbo{7Eo+8w)j>!$i0xBdN7`^k@gWpDrd
z_i7UnNKuQ4*0^p(_>XSIBZ!k2liB+ZdfjxR*v7_&70X4xM&vL&uGbsd(Q#elkI^RL
zkzRv7@WQFUjJ@_US@w$xnPe~;VC)JJPQif4+;roIXavSFj9>6X@Se{V_UIFDVr-$!
zq%=*F%C}e*zgo3YW!-Y+Rn66m4Kz0F`v`x7hT=eRbbRPI9vnT!j`Ps_)zwYgx_Zl_
za6OZekQPH<quFx2z&YeX^Tp{55rfcc_=jUth)l-70e=HOFp3a;>XEDBD;Sd-^+WOS
zL)F{XGYf9iCo(F~H`Q~|7b_s`Dmu7?r}raT34;kyoX}@@FS?pYJfk+`ylozdsyPk%
zV!VnaqLM*G18WD?X&-5xVr<4J1{cJsgZk0`X|63@+F7rr`a>GXg-BY@=LP$$+A`MI
zVDt}qBj>NvDTxe&K@A#${DMI=8Y2QZ7zl$ieK3T(yfEd%j4xP;(ah7gif=;edz#zu
zLf#Q!53?oGwu-0dxzTXsG{}uaE?cM{^hIs*$SpT24K<c^8=Q63mb%8Er#8d$(Oimq
z@Lu5<{0Dr1UXymX)7AJ27s};|(_`u*;xviVM>MbB;xY8+QgV-I0MTFUZ5Gor=)#D+
z<sC#tV{~2893qvXI_iCaieW(WD0sn#=s@=RF9@$`pSmX)VqmRXr`HmHt_m)p9VbWo
z_7DH`&-Rm_{mSm|oY?(`NA_sv$m-3mU=}2ObfMrHc;IIMGVs-3|D)gjOaGZ);DMj^
ze|~xT`^h=~#(f`g=#Ss|S;~v*efs!gl|NKF!2G%4YhLE%3oI^Jhzh6kc?tRwl&2nt
zzM#_NX0eT@Qy`>M4+-7iLZ28QWR=Am42%T>i;tcb=OPUrwoiefQU~53?qR0i7}N+E
z$E$-15f*p`!5hZP{$2+UcWnQ`witBq+RT&+cID<ZD^=MzI$|RU(ymazGid5PL-ZB`
zm`9(9$&5!6Eh>5vep?c<PweFQ*iMg+Y=39hP7aU#d`rUaK=-h48LK{55;*4}wMOs|
z%~wjcyt!$m<*KC$VPjYUMGIRNDtKyYA))sJuV)J^?nyx~zjgbL{on^bus7d)(-+h_
z8!oE2a4jK(Gr$P~8$5=%${!v4*LlV#{L~-DME^Xbo8p7W7q0ppp)QSI5J1AgZ-fs2
z%|Cs*2a^u{(KEDP=e;iK2yMh`HznZ~!LzGH_1^BT7W>cleO~5eUgqULMETP(d<AAO
z{d~dklErz$V+Qe(2V85)vOY@*4jCU^9<#)cE01sb92)jaXJ2$sUUYtTe3CNb@Y8P^
z3qSnK2g~eySo((ho<5$G51#ojm^^>ii^{CF7uEXtmJc5{WTa=yN4Lb~tWQ#q+nz(@
z$G`PS`X%<xVsF%TyC+#`!SeZ{EA1wd8BvU(972&s+C2BXmVAkPNOS~pGkZX^@LnFB
z`E)~0Meb$qkehd2wL7o8=AQ5<<x#q#jHSoOa<m0UE)=@y6!L1^<z~F!>-D;>pSz-x
z^iU{IP!2X~l8;+;SJL<N8)vvoj<o1$J&TNtcQyNZMf7{(%1T#i>b)qmPft&by~@(r
zys>u+G9!^k{m!xKKCqVR?&u(bq>#&5xm2`xBI<^-Uaw_4+YjvU@W2X%l1El8VBnzR
zYAZ%Rys<I70oT*xQ@tyzK9HU;ruGcogSyjfSaebFANF(}T-W{VJ)-A34Lz^>l1UV~
zQHSYl%uqhjUJN=!4n*R4tvL$DK1i2DCrR$j<f|4-<AqJOjF^@K$n8XoT)GmU6iy3~
zynaU=&G0k_<|tKLt(J}SofPvK!3X0LirzuLE&OO|TzanX&lHyREc+;-L_Aiz4{d+<
zzAGeAQiCUW>$9&7ir*{OZdtapqHBx3X+9|ki3A{}Bd|k>$^L67*fE|GO-H)TN=45k
zFiuX53mEG>av&l&KqoG=PGBy8OJQB8FNP5fqXug)l;Oc>$R%)(ck08}K>p>aGI9e6
z9{98?oylq)nH7vN@?gBeU<SSne7uNwK*<iy`5F?1rB_bfL)E1=Xg(~G@(-maMzDpD
zqVRk)a6?R|)$sYl+#sTh=%ZYrD4fqYT;eet{WM2Ze^xk`Ai7XvPP(wA7zR77l`nuL
z(TpR@q;sA|pE*VM9KOOM<|Cf|o+>P|Xom*}_TwM_$o}nr{EJ6=5lxrZJiW4YRde=L
zTU*=Iye&DNv*(yQp=wN0Y0dj&K^3z`1g`3<`g1uRczc&u*R8aAMeiy)Znx`uc6{(a
zYsCxCaKbzopI-`>LX;cPEv3@38zaDXjMNw{Fg#+Ifi7X(W-L)QlO7NQ4s(}?OQL+(
zlZz-nvVAbeCK$E)>?yV=d=;GZPVj9km}5}i*?s87wTaqsaiJLuy%5eiHt0P>I`ldX
zTV7fA{RU%kqC$oD*^(6t7?Y_Nydc^OBj*f{#(MtD$AoAIqFtevD3yVGB85ScJ)lH)
zwLVR!6ILo$^jvUQjVEG)b<S?zx@KQ_;}zT7T-R8pt*LndtqXJj4oSQ9S9Hd89_Sm!
z6Q8>n#erMH(@UP(@$reLoIO4g?bGjY)I_K1eqbzVi{5s+!h6xIo}M3wFK`SsKZ!aT
z3x~hRRdgo4U~j#C-M;y)w`^slESQ=-e7I--_M>0h?;iE-;r6ke9_~6GSF6ib6pbcI
zxYcUA@d{XgLxaJ<jlf%5TaH_dc~rCjqYsf|2M5QRb0f_?qREbIZ|_jB?)X?hOQ7N4
z8fj<w&6vXX3VI)rG{M-$er*w-pJeD^8w<t<8VKLZ7vN3epSnh*F^=4M&8MvDVITEi
zEIPOP$-Z!|<`F>fPmD8+Bb`ToL^KZVNvHBY$1=q==X(>;TyPq?11#1yHZ4;s3-_1Q
z|DZoYPs0`;=x#i!Ncg-M3F-YXm=V1MZN@l*eP(>t6_4BVbd9tH9UJ`#M&%_t7j11f
zkBxmL(aXSR+$lqJ58g%e>6z*!S`mCDGR>pJL<2B<^)xrpr@@<C&jZ9TPVua0U|h5>
zL^6>|FgUW~bP%G$n!<OYG;u^=?m=5k{JW_!Z@PXBn#UYNk4IWjj4c?KcrON_KDt2B
zEga^NKit@bbB7y+g+scR$T*J(BRy~|>9Ps1Veaz}9UtmAPi^ceUG+nJfqtNeTTq{!
zKDr4(4HVti4LXF(dXPwajHOIzbdWgdKo5}yQsOxa{!aBp_YxiN&M29*Z;6r-4^K-b
zn~aItybw)E2o~)3c_w^lNv<HxE$w04^?o5VG8OyeRjU=3N!Wi8_>J7KcIG-(H!#vk
zFBtKHL(~4=w%_CXvuPYq!!at*PsV)E>)X$M_H+Ah|LyPXo%ilrr$=9KilOJ|@Kn@M
z#>KM)=o7sppy0cgDQ)9lSOK874cEEFKOf)m<Qo@fzy0w4PkvKt_&zT$Q^10`;cH&z
z<qIw>SR)96<+6x0gdML076B0QK+a|%)fhMM+;Yz@2`CUr(nhclU{MFz>ucc)lTNBG
zh%SN(%<oJDolt4fXHO|W2y`IM7+R+9(*5NB?MSfNd-%vsclHAhQ%UDjws!5B79}~m
z(0c>34=F_uyjeK9k;uPaIOFX~m?x(-(OtEZ6RYd9aeC^E7X<-cBNt+7mwG22J*2eI
zlg^ytX$etUBvjYe?CQ;%wt4NU7$G6faf^u=5_+A<WwdxAgr0oXVnP_{U=Y6WVo^42
z{ItM0zw@2%+V{WzeP8%sq;sVm@Kr(J2DUS!m;U<^zW9GX&ohDPOs_d(^FN&MPk&zZ
z`d$8;55Y$VOcFQ&_wXB|823yKU*7gmmp6@jLbZMxdnM^el4-mv&1#Lh)f;T`Eu7He
z19%_+q9CB|$4a1inU{H)mw$idPr~p89sXqhoTYr|L4m(KU1sNBra14TO!>;EK2*Z{
zpKFDFhTcrei=KV%a9;IXH6K;{))(@(4_5xnZJa_FdhQ;0aF89II;ZF0qt1h$y5azQ
ze1v;G&=!79ug?y5J`%5Xd4qT9912f9XYclExULU>u@_FhxIB*!A6J5W^U=?IoZJxb
z`RU?~4qsnT!t0*Du9MTYIQ*U}Pn!4P?F4Sr__#kh!tZcxc8(~9(d5#GXBRq<rFEZV
zEu6xThsm7a`&gkslH7z63=ex!;2~=fMRj_5r27|qBmT9uE!)_<YAb7-wz{%rd9nxS
z&3Kl#JM6pEHH>M=c+$ntGmLAJYeNdCpd_K*kRc%`+gzEYe;A4ywIerrq4YuFOPVy4
zPZ(VA>?ZoDP^?&%bZTllGB1iw+D%pk_903nla@#kd9mb*@>cW6-hcPkcK_~sc6_vJ
zM+e((>|rf{ViV;Y{l^f8(){$~)XI7Xay5}*i@p!gxy~P2v)Q&>uI#A>t1DNmvb<s0
zY{89p^bY5_vq4vDiDP@X{Xl&WQ3?zP%JALoyIL3Q2p%XQ86OnSDE?7y4-MrlGBTe3
z>}3>E_R+tv|Hx7-t8qmkk5Pj4hThd_A}UKRE?7K?hxSY<7t^{b?3t2?p*S16a`0?4
z&^kge*L#R`!w`V{T`VnIv9h6OO8z`Jw8SR+=L5BQUoh)<{X`~3Vi~O~G9E31H~PW?
z(G!b?r+Y*A!d_D#WE^EMMo7VzgS8mSw<M|G@<sI(<4*9L4+RX-cXf#IOK9v_|8-R#
zkrAYSB;^!FIP6`l8%e`-c`?+ws^c&u_2i=Ob;E%|QU}p*<DNc41O*0@4Exwb1y7H1
zz~}`H(zry}BabyCdwnH+OeW*L)^0t493$h%Bc1r|ic$=Af*)<ayc~EW9O*LIFArSA
zz=n|tLn>oFQ5_S{^Z}h??Wu{WZ|=dbdFBdVkJwW`dnn60!bOzx7*B$75->Rx(ikx}
zFr-D|IUSnI!o~faNA`Dr_kY=se*9za@8-suUAunGZrr$Sn_D-mq;aPIDUD53_mgT2
zV^18VYy_A9kJ(sUsCyv|A<<rb9b={7Tsyoc`tV3?Af@GO&+bL1Edz}QY21iXi>l4+
zG1n6gF{jAJLaI(un3Db!1wD9JZ`R!KTrRKJ`ubH%<yL|LSh#=YQJKdcb%{Ze^o2{Y
zf?%H(4q}|3t@<yR0P8}Y{opX-ExPdoV;7MwwUZ+^j+0t75!YN27=!T?I0#2w=?+dZ
zrr-~sivnX;=+n_!6JCi@Ugct>=QJ=x18QC_U_2GBkv?*feTeiN(S$gvMg@zO=7*>6
z)EX9J1Wv3Qn`@%YL%V<ffz|5`%}<Qc1J`Y!<G@(Z6J0^M&6*eEJNsO>J6(^|g2uXW
zOt1(39_@?a4D|n!m0Wz$zWw#<_Vusdu}mr=+H&S?_q1t0e`wv2`k{6o*GWBjB7BS6
zl`HES)=N7&Ja*a)oySN7J}xgW3s=g*i-v7)?--G8eZlVV@Wc)dPX$LJ3@3v7*cy$I
zb$Wg0Ehz4B`~a`AxvcLIR;ez#eq`&~4e`1a$1k?1z&XRWs7R%R+v0V?Rh%C%SYeQ5
zUpn+Nq-=%PVYGp^V*FVY4d`|2_Wpanwf(&Zj^h})(0_q<7|qbffTI{_*-wecXg9EF
z!gt#ZJ=1c1Rw|S8Il36l2v3uO2`Pg~_j%!whm1dbhITA0VC=;h#$IYewFQTOkop#b
zDs;G5D0#FlQEl)a=qR)k=Y(eSM5CoX>OR+(F$D~oCCVV4&AyNr+I@xyzNCyL@(F`H
zQDN`_BHUt{PtjCS^G)|B0-ev~N|qGQT+p2H2+%Xqn1=d^Cc`lTy%Ka$-O{`DezgT!
z%bs)yrp=ako*IR5k$DD;aC!kxh~h?mLFPba!0BKtc$4Zjq?2s>JJ2_Vu~NGvM`-fs
zdG#TcSKo@FF>$BGq;wn)YoZ&w)@s*%ykc?o_9N{fduR<s7f3Iw0o6fTQI~NN?BQ1M
zafM%v#{-8uI!gGK2GV61qUjbRk{{HbAh(IE2zUCTlh91~gea=D8z;86D;m*29wW_j
zNF%(Ucllm)dJe~M-Ut7t?}vxGPD3!ZEiUC<1|X6d=X$XsJ3QF4pS}HWmM`XPd1YPa
zRi9wd?e=Z|;K<&6@2<V`&hPEVKl!;me5krao6yq%JCE8Eo}-sTwx<E`q>vR08J{r?
zL<pINzH!XH&#pZwoO|{NvhnobdB?b?&gpo*mOn(9wf&EE%*u<deVRkqH#dCE%e;Ky
zMGIQIEX7?Ubd<GtM3@%AmyjMJm;wPs$-p9EEJi?qnaALgqITrLg~>pC`9$cz0}$gF
z-WDw85Df9;fvJR4Amn?7X$7+B3l|Y=3|6g1(@qecj_Up{Zpc&%wz_r2^YgPzp&bxz
z>W?F^uuy@6sE}5sC86x}^wb(BC+^|MCgE&uiRUFL9f{;}Zx$^S^eB`}j7tnei$?Yk
ztE{iuwVOBWiWVvC#g)xvTv5TIlEu7xiR$-QeLT}5ukS+Gg~d{&IDt`Xfvc7Zj4JeU
z(JEqSKm3ay+SkAF4bO*-cMtEKDckf_2XNrifAFxw2PQ{!o)b(z6D~Z-z59fis$qVQ
zE3y23F2kxg$7L3Yd<h4JD$a4gtGYPHZvlo|f*~%n!QUY%$kQE4&_v;a@vqj9AUteY
zu9UMvF>fpDq=KnP5M6QuqvtfAmwB0&dHMHO{$%#}nsJNIQp5^AP+qRy!0)EtKELv$
z+ApNI!t+B5P<j!ho-VU{o*eMm@cjvZ{1jiOyngn3D!5FKY0={Rsb4Bj?s=we)ul5*
z4xmssyxo6K&#44?V)lF5CY2}OrsV@CzKrGhMn2ou4;0`LzWCo4m*?TlW3w}}Ysg{9
znK~{>Gsa#tLu<C%R;xFyc1qM0kuMpSWnJEK8Ewg3r%=wJOhVy`B6(};hOJ1JEEX#+
zL!toWAR5e5Y;`+^gBr>y3|mBKxJS8U{F!nOS<ek@I<Slp0fLg&_uXpM?O^Yb*F&lw
zJePq1_2Ah~iffcQ7!mLg=Q-MeVI`L>cq9|b7-V$z+(B{2K2R8(P|lQ#1xv&h?f7`t
zcD5gS)C8&3Fw~)xV;`elujP>|ufO&ct5%3qA)<?vk7FZ!ABw_jH}2Tl=51SEz3wTk
zFl>?P3<V+?C)8c5)eda`VBc2PuUMhHEV(`HY1rG%Bir4%=Y3(VgA(2M_z~<GHB?+7
zLIuaY;AQXQxiHc<$+Et86h?_~e$Z>!!w2^?4nsF41&?jujA9-v;1Imk@8}XraUvjy
zyy|N>7J}kjV@O0DDUzZ|U_Pr~?>q0iGIBEPs$Yv%EU#HAzp83UTZzIDjpYpQw01iu
zYNzIrUyOsN|H|a`Y|NE`-~h@q(q}QgEcMh{tp#T7rG)u{@rrBILt3aLsp*nQJx|n2
z*rSQ{9uWtmnPY8;!jQ-z6s8`D<q<m=t0q2%;6t-s7mi1?rehB{u7islyqEO?%1lx|
z#)MCCU%L(2_oP|(_Cm8l8oe`rCo3FJ)wghA{cg+l_eqzz<Mm~<uJtIotk$?Wt-Z9?
zBQ+gS3E;d(dUT0;0tYbEUI<qs!f&<sjFk8_U;knxB#jzg`9!=heyop)+KVMK>TAqx
zdIGOtHMD>Gmw&eZ_TT=y@bJ{u*Vpap)vF%a1uj;q8<sDwTOw0(r9F7Zd%-~r%Zq{u
zDJ6BnVKMF#fms*socsC`MYzN5d}Q^LZELkLP_PG6UUMbyk!vZ@0F>1z)JZjWsYN7h
zZnYcUpK_^eE3504Ev*Z0OPU)|t?M!D#a$Oby028GD7+=|OmjhXjRcQwPvb6p%?cK&
zR9@|Tf~ynZ14c@WZAHN=TUho!dPM2PM0B9(d*~1$MKoDDO{6B{t9}ZW!NAQp3KtwN
zRPQBDrTQxT5R?Spk-qC5(XGj#>-)E{{-3RRxp^rPQm5q#RUh|c3Zt*;6D%-Nw?uc?
zLk>e&u8_0MtqogUUbXG*U3>fIztDAz(L^S8-I;*B6kDRfL~OQ2yL)}%%Q$4a`BN*`
z*9-RNKm5Agy0tD`!>NIJw`e;%NA}LUqWx#*78Mw^d>6f)XpZ%4xgz?ZdDHEY0+aM{
zUFSJ&fDzp$nsocow%xsZUp%eh(QUO_)0z#@ED>XL1voh{PQl3`=21KOT-H`stG0Dz
z)7IA4tg>7U(T~C%r|IGg+=~H)Ei%x-g#EwR_lkYEE}f5g1|f~V%VjA4H7=wV#(@K8
z0h|knWUki^J>m%3556VESKu#h@DN>RZ@9%p^jFMb@f-0lQuC5Bba*z_`)Lyqw>S#~
zgF{sB^C(ZEy!722q2amujxpEM9xh^xA-WGE8BP{v>6G{yMsL+eyKxv`UM8oJSj;!@
zmK32xk1>CS-KJ0&=M&Mq7={|M?;uy;%uPx~bw5h`%=yx2)K}Zm)3m8kjk4yjXje!H
z3V+kwbw$;M=px3osUwKm?{vC(s&*_0E?M!5v>OT8`eZ>oEYSXh>d){}=S|2QX^j((
zJQ%)W;$_g=sKy;UNs1=1=OlYOGF<@#kzc@S0e#v)JXxPafGsRxOh?zI{)z6K;4pKv
zBisrGZH!mYG7Lu;wH(gk{TTBYhYPjOw=EIR!)T4slkq1)7-t^(g^|<uw}j1tSCTJ=
zy|#^X|FAFo(fkL$;-Zb<F}#A5#zcz~X~-U-i}2J%45o37eMh{cDY`I~BlIkWvq0C0
zTok{?KzSLW<rpg>$;aXs)JMPi{jnV!9@|fT^0xi`KmN1*>Q}$DM~{dOYzha1u9bR`
zF&;axL-*8Y%8VkpzV7ky+;9I!`EdF*yZ7n(^f&yNm(N_Fx^u(Vyv)lVp-jZ2G19DU
zZnzNxBIld1O28(RdB}A!M$c~zw-)h(!5}E$4TG_Y#X5oji$^DpBET$^SP)ZRUyC3Y
zVqnm!7Y4|8m_7SQ`6jPi8;$Lhv;mKHt>5f;`y$zdRj+K=^2VA+V!4q<?~kc&7&V`s
z;!vM6;~7z7cw-GD;6S!9y0Q4lO86q}LAhGBG6Gh$Y`L-)BFkl~Zm!$P#=5Otxndhz
zTefv=%SshCy%u9qmpp73hNFHWh++J~fR{!&kkNuVo%ATMOb#zbG5s_l_EEtWV_Mp7
z+`8ovU$<|+;;9v=#~F*m^a+!BqPz-!otJUV7t*tPo?OE#Cq#FJYkrw$!+-Ab@!<m3
zXWxO9|Klcp`}c78@29@c&O4*V@TE`24`GzPnillweAcR~%T`@m)?#YK%7PscdI&c|
z2@1ZboR@i-mwEa3RlYF8*Z05solhU}r2|s`M=9cQy#JFG<cw)s;XfZ*!hKK5Cx64^
zqDNk)Wt4EYKll-ac89jl3|22LT;p&}E+?j+PoBX6WmeaddY?Y0A{ih&p})?n7hR^~
z9{6^+iU&FlO^7jJ(d>2j|L`wJ9#86>@qeD5waD-I$UePHd2u~Y!|cOe_0%2Hn?B^m
zY5CyQX<r||fZyupGdI%z*|+Hz?wNk0_;(%x?*5xIPv0@EgX^xGm5h3RZoU4%nys##
zoHVW78M+Lc%aN)nIhCbHR;!=r`mwR!4vM7Z)io=ZszD~bn0Puu(j{TwAleHDFFfdT
z`GU*4v<1g3KB=F2X_<Qos~yPbq#Po8hm?32(CW2gJJ`oVdF;_;Maj#^+3d-6cu1P6
zmfD!HLa8EI``nEO?u3?1OCHQxM(rf}rP)B~d7|G*fkU#N`iwDaWf_CbrS0rIvb*oU
zZ~Hqt*6(yI8ONE5bZMmfS+dQIE!$YXE*X4H>ynBq-iCS?9@IgGU$w<pNidxne{g&Q
zr%?pP3zE4r-cJ;ARlzMu)QL$Zj|j%exS@;cbcxg$S-DJ93hA=g>sT^>9ODS}qIe&x
z?QOxkX@l;uEB+a8HyZkN6sCCe2j@3;=u_o*s7K<=F^o9OZ`FgcOCx}1J}H+*f?qry
zB5}YOJoRxLMB%C7nZAEwX9M;Cs;P?$f?-kPTUFn4y2f*qfvT>1J3}-9(LtgC`eYwu
z6mO)0!!87>MxseKNM!RRm*Yn$(^QM^p~8zb1{hmr`$`Q4ZC8S_XB6u!@Q8h>NOevM
zsZ<gJX~Aac3W-=ix$6c%jZbK&;I)83kZ1%p>EBZwXI2!>vnPu0zqZKUK7Abn-7lE2
zcvxVqiP1LV=b&HFXv&RE5w(@~gy^%LuHg`gfqB6riW;?=J1XXjDDBy6C8aS-Y3+vM
z^<4Nqb_F*^CL+sX!ePj5w+Bu~T$#;&F*y36Jod=CrGlD;!dY}tYr=4i$ceCa?)Mtj
z?>6lCaM%9N|MUNCzyJMj?Dn0T_QN0iz^-4v<w|1kd3kNq(gopgl!z%(ZfVZrM5%cO
zjzhzVFk#RpXAL{tduWaNsm>8q6>!Yga2RfhJWzX~-w}b#_v|aNcO3i12!mYF4t-j(
z3zWWErvux34#lnToOFT<S+!%yjU5qHlFkUmqBGp@%5<%{!#+4&e>a|(*(~#tJ+n~Z
z36KEaKp?+@Kam_5vlka8!c)CJ&OSWs+ZNGyX`X1T*n^I;y1J$J6g9T&+bw!D9tp=u
zr;kyOu?PQkU<7kDp$fP)>R7*dYNLVZ-C4_ef;rKKDC{wkVLYRN=aC3|17%d_g{Q3J
z9YmgB(6|&_ip7H3S@Z~>Qn{#l&h7no-?#VPyDPffvq6u&@`91HCmKsSS9A=|@`*o<
z!W_>DoNv6kVPAdgj#bNyf67Y5veR#LJ*Tw;PqVkNlC$euWy@!lES-vqZj0_{t?zt9
zV~GLPqXxxqNaJ{TbmXZFYqgf>s^Ha!K0sg3MH}K)Mc1Oap3C7doAIMiV6V%ft*@`x
zm8}iE11ATJtc>5e8-XyeV&wAq0BsfE!Dq%0L%eUbfv!sD7=NS{^dh9SBk~xBXN)X3
zcYvQ5%7;WA_C+srFTAo`UA9yv?*=A}#zJ=2WkFY<dv0uAB2{qK3W60x0z(-4Nbx@O
zK}(6GP&Pz9LEkWbK_iKf<s4*{{-8U<*b>En=JcI;8FVY~E9hDxjj<U+@tN}qqJ_gA
zjG|fnhQ_I<`k0sKepClEO7jL9h%w6bGn&I~%@L0fR@=zVbTR1*pP}V~=T!g0bIy}Q
z$Jncixr);V@&WsNGMCw>inNgn3*zO1jj3Iuv$kM4v3L@LACadxiik!EMmPb$SAY}7
zduR~GFZLmeh;N0oh&V=|buuGTkTk#;?&ur)fJK2D{79_}m@D~2W672zj8mboV(DDT
zjb@~jg{C2Ecv?is14KyIkJ-zz;dByX8Ij(n$NRz!BD~dS$tmnZit&)W)EcLUwzGZT
zj*s@#28}CYk<41QShYB`U*oKby*bcL-!qYDSC0(VSZMqvf=i=*;B*GZ1Bbol)`I9I
zymCQwF~Yu+Q=G(z8dp1r676))?04@xv|s-EJ^StN?%LbG_>H~y?nA5ByBbF#o-x=7
z6kec@=mhyxgE(lIpA+2uc>4SEzo(Dz)Z=r{9#5C&f4@w5zTS^J{-YE)+T8FpFZ1%p
zD83lFu%d(x1l1)mD5ACSWP+KTjb=gzA;A$bEiqLWM8(Jv0AUbQL(o_P0`<WJ&dwyn
ziuq!&fdR6A0R^TL7I%8y-va~3Yaw{b9EpjY>>k=#r|<V?Bp6jT*Tj4ZVz6S$de#>)
z)5SY`df~;%LXY;6bGq4RSWk;-3Ze0&ax7J<S^%+V4Ef3lq0j0&xuh#AE4H?=p@sgc
z@9C8l96UEPb*hac)$0*lE_4x!&Ehqp_r`poPe^=R3oN`@(%Fy#ksR^}6D%sz66Rie
z?RES5*S_Ytu!G?bCIq+yuo<m~$5Wg^>8Jh##>RiX&p!D}VW~4c?<c0;k^I69(`)|6
z>!aWyqBrox?_1RPa6SCv6u-IS$!`~wxMy0|L?0swb^|Sjj0KbWpO6q;&^WR1N|Opw
zaKy;x5q5Dcd=g>52NupOvJi&Iq3^$iD>TF5F)#BnFZ1%JQNB3C*LOZc#21A83oi&4
z{{JOQc*2`J<ypuSA1p6f^K-ZR&zG#iANbg3Xfsm6lk&kwID>*Gn43c3T)6Ye{m4)H
z2*B_o@L_#JldU5>g?xg85)maq@yFf4h=eD-Y0-=`6G~OT?AJmo!`~0^-lt(dg+K3p
z?ygT+R158YVVRu|@BjGseG;ibcRap{dZ&GRzP#u(@ScA4@<Lox0bX44n}HM0=~KVC
z1nf}4;4~(w*EAH8he^1QknCBqwY9SEF~NNv!PO_ypl<AS!oDJfe94vJfDMHp`)%Ny
zmQE!kQ^t*mBOpP#DCA_6U?^>QCyr+rSX`zQGSG$58T&N#BsYf$3h;w{e~55dBAsEW
zs(NDj-LUuXzGwH|zvuN{zj4corHW^J;Qe?y$K&j4l+bs*XDMpij~?2;{_Edcy?$){
zZrhd2)P>?^Z+FLj^Xp&Q;laLjTP<rfL+UXMc!$UP-tJtMJzk1hH;_6m=>`wlfD_f~
z>W2EnUMQ?DP##hQOfM+QP~b@(m+X(h#5efY8pZdJvbe>PIV(v{MLCC(zgR3uMlTqf
z_|FU*vuKlxN!!}3BWtxt&$eg7L0jhqLk;SZ@7cmyBq>=}>jjJ<tby#@MQc@U7w1I4
zsQ>3ZkA@+F4+SlT6FlN`IlUWidcl?X$aT^&A){kJM5%Uh(YHqJ#J%CO>Aa;<Inx3p
z$nP4ri&4vt_wQM&aimkMhe+**5_Z9(jcO-{t{hIMNSzti+9v0FltrQ;?H`6Vl-Y|>
z(haKL3n=%fo79X$PnpHqifA#8(U`q(P*755*o%&6x)242A&_+-Mj^h3luu5_I0RQ$
zg6bXg)6-AsAd+u5AVpfsyB3KNg@NHF7#7)Mi-<-HI2pA+sdX0yU{ZsDW1VhYZ6Gp)
zJ)SUPCM-#Ni2~F(p}Yk@XNCP_h_C~eD817}C25{PQ!)CY&?Q=deW#)^_Qv8_U_w8L
zLSSq)c?JJSl)k4NWEYc;k;3!j^uXSE=hv3cXYFtQ_HXPj|MEX;-rTWrb<LJnw9YIp
zJFbC`9MFWt1+Axr$3!N@i6DzG9uw;c|CpC)t)mKBQ^7abn}~HYQF{sDL^fX)?)2=z
z-FH0oe6g}-Zn)Ff!WU5DdKyKoJrkOvjE&PG>cH@dLR;^;#4x0}IvV$FsBvjDTE0HR
zkrN}wndk-44j2eh*{arW^dXG914Zu}^jd05*ZZ+_sWr6lg*0UF9*ld7OXtFwqE)Nw
zs)y8|8P$X0JY*B`etV=PdIn(5XD||wda|i`_-OmC@UkYlqpP}%aTnbJ1`ue{`8XJ9
z)j@xb^p_9szwdi;xuRM341~j;CQoAxJ!B^2wQfB*I<cSr<mX!3YBbPMh-g6@QBM1f
ze}W~eA-Zv;Xm7l-Zm++3-PTttmdz4zmDhTmG?q)E=ZtyJvZ;uzuT*TcTC#E}W0i7F
z;}s%^Lv+ca8xBdSI~<&OG+evY6@IWE8U}X^(&$PwZv`*rVxf@HXH4@<-!&&V@_Eu1
zV)PAdM0R0+N}@VR2g@E@J3AWdM~~d7MzkFU45FXevVw7@I8TFriAMR@pA)5q5tO~(
z!Ea1E&^Xat5QbT%dzi2|l;p+BNZni1oMex^Fi*HoWA6s$B_mB>b(u)pGF(CL3-pb7
zYaT5v8H?}2TO7XO1#WN?-Qv4msn;!trjW{$xgGZZx>TFNDQGFY8~W@9dR=8tEyiOB
zV-yik>O-${YR97I7*>c}LKc9QEii9Hci|rxSV*sUa<p%?(^H?n^rc=u5sz&;-$!5O
z^h09_UZdk;{w^)yT%J}tqJn4G$1kRLv*)O%t`%HIqHi6|z0qW-YvTK&vqWVDI?us;
z6^uP4AG}g{p!(hT&igb+(JK-`1<07Q=-QA`;Ma^D=R!(e=C5Q{y_<-^jCf5kU(q;;
zcg2L0@g<insE4rz_n>#(=Be%NysKu6BtIksmoqm6PKef$T)^IKL_y;m;89qD?ReN0
z|LOSH;y6T9yGP_DG44Z);XybT1(^qD33s}ICO83!sILRz>OD^bS*@;Wyn(0g5snZo
zh9PuG{EYXIN;O70W$~ql_aE9n{nL-^M?d+c-MzbQ2ZuGygTCOP*G~B-IHtF#edx-V
zhr%Fza%Oh;Isbh2{Onpd!kJH6LY*&PX78JoC-490<CEO`d6XB``~2~AdC~RHaxhEg
zhOc><mp@WL;E3XJnZ<C!hO#U+{E8;%Q&CQ0h6s`vop^?YFQfrO4gw@X4*8d1JQGoD
zPhWyyJ~dd8=XrpqjGjd~gMf{o0t3c?M&9TX1RM#Tqux-I#w=5kP+!$zMvHs)$%5f{
znn1k^rpmM22V?6<7-}`z){uZiTkyCbA`2$Qrl^Ezqae_OHc=a4sxdL!l=_|0qQX<2
z=sZGzhkEPNH}lpx&)Y3w$%S_@KJp^-+ZX@(By8L_0M?=auM<4*h7zpnwT3SazxUnm
z*;`-zs;w?B`-0$cVe#Q#^yvgq-*k>o{y-qTgN%irhv9KapHnvI3IJ!koPYWmeRr)(
zIM0Lrsl^Ca`TaOd+UxRLc`AGd3H&;OBX=;q5<@h`<U>DV6XT;jYe+!D049h?$RPK!
zgaUnH3?u5RQmuM4SP8GOvILh((aLL8{VrKSV_VeuOetpx3B{3=7ODvhVQK}&L0(7+
z&6}5bnU{I_GL<hd;_FLl_=5MmpnPJI{UOg&X3Y*uzLzL3xZO{E#JgTp{Jo#>ktes&
zY8`qPatP(=Dng$^kNo@ekF(*&C0uo0%z4#=hb-SiYPyM7GX^fA?$2}%14z`Cm&IQT
z8Ox;EpJF6A<kD{5ykd85U$bhZAUO!{<8#SEq#p|M5UH@-gPHcuT8gX?zCBYOw^`-s
z@6Q)xK)*vby!akq6Fy&5o?Mzf?QNSWxn^f)&%U^hXNu&AXYYFI*G$Gxt&iL4gYo#+
z3(IV5{l4koJ(j<8?dhi<zfE6wj3^1t#UhKgxw&k&Z*5t*j7L400En!oPWC_<dfF$H
zD0qV-Cli^~?KEs8*|Sut*yh%CE0veM-pR$e?=gV^hP`Z&xsly5NFnQy;>S~Tp{x@Q
zosCeg1O*}uViWbjU}VLT<g?PUM`G;l?b_XYcO~nN?e?A5?Tt6SYU~k0`|&*Yh#ieL
zsi#n=VMHN1>}Nm!nHyeCj}NWg!Wpfny7uk8-~ZmWAKiCZ8d-RBHnMuNp=*2g@FDwq
z)m0aj2ZuP|j68mNQqwvkrFNm@#~WU2kthmE%FOTu-T`@y0|t}xfey)~u2kpVprps6
zKAU21n21N<p|~Rzk*7U!Mbm8bB-KvX>*mz%-@9kW$EO+-BC3{rtVsXm3NKO}MMHK4
zyu9PdyiTWmu=^UTljA+DP1qL*1#V4qc|r3}a&M+6Ss6z-ly>L7mm6sUQMAV5>=l;M
zHTC5(tVvjlIlRwNUJA$5X67$Se4<fUYbBGcwR*mn+QSET-H^#%GG*aNHbZJTlx@HP
zTCn7b=RjM44{H<pA3k}P>D{FFBhoCPeo=TA5Xx%pgo2W?;D(!_4SOxE?Qob>y~6L1
z@{v`JV~qb#X5zLWxD#E45eGvNxXykwq+g6gF{mVTui!Fi+0otut7)wfk1bh2a9hNY
zQg9g!ySBTx?P(X6t3&~;=oyS+I3Egjp%J2UIl(dPXG1DRl*Q3FhUB7fNbq2-fx+)$
z?Cab_A|;pz4%4Wk%S)~KMlJQ?%)bwNL@6|c8zZ#|ya5IywWp~*5s9QR)SC76*S})l
z{`Pn5>b2XJDpV|2EL$p5R^3FxQ8!QoE@Mr1y<2!dYRX05_YBpT*2u%Yw}*83?0o}$
zfnKxsSp?WGs-MxKP0pJ3tDpayZEtUTI#lj!G)}Cq=ZR=a3#Y&z)~!VAG&LuQ_+&lH
z-fZ1w-P$^j0i@A9(dR&|6Q9VJwU#Ri=Ps?+Yw8@)Tm_G;412TD&TtU97Fl9{thTZ3
zfRDSTB+-XM)mvA4h{R#u=W;~i#cUDy=^yriu(ukx$-!J^PdL@%&bAz9M!Li<#&oSW
zbEMnPk(nj0cV~1iV*S48-b6f3up!EpXBIApc6_kyktS?45MY53m&m^}H{xL6%4n_b
z`%yJWDL3*x4@p)+OTNfLtJ4}<j8KuKVBkm!B$<?GxzO@*a_R6QpqA9JY+EtAC!dK~
zA|goWL@^(?S8lA?x4&`6zWKFR?dsO5<#9$8-D3>E2k0x2yr(B8HqyK<<O(`6PTz>c
zP3S%`6v1UAxNE+4+QNO|%^4As=tVH}2Sa@@c55z8brm8~ns>}CVa+5&DPr)zh=F5l
zR&Bw^(P^<gMcwxhYc?@1kQNgo!jbK4@7m$Pp?J!%?d~4f_QO5f+1=B*W8sr%jm8Ul
zO#51`rW?%QZAA8xnw9r?M3wrAyo2)y#!gaog2&(?29A(o5Xb6Z;AhWyj6~ojG>)h#
z9Ce9Ua!2q4x-!y~F8X{QYA&h!dKZRejD+XjCUh84(I9xE03p~g?}^$RInTt&gs3oh
z1#?qG#UpK@i$o+7#f_nsd543DM}cde5*dmkHvMDYc~YQaU}0`EKlb)^Y-eX%ymFxV
zKzc|bsWe7<7im#3W|5wec|}w?MmFbRv_-f>IzC7fvI9=;_2YdzIea7-AE{kE^-oU<
zuJEmk2@W01GY-<?65S+Nz@OFcGw~<d1!ITCAvXZq3-K15G)Q}z6`jHPfU&x`SoHUL
z8dcpxlqqyNDw>1i#QCTznDkT~X&KLizwEasdVGGNd9f(^)3QeGK;tYN7aeD8XfM&?
zIJ#iO1@AE?VaR73*{drd{<5^B^TKy`%+^CVU3iMu_KD<x1Jz9mXAG?unu{1eftUJ8
z)SGBu?fAetoxa{xb$I|ejJXNk#S&@NtudPJgXd{-)t3ABAKKsl!@t=52fG?4!3G)a
zlCcoJxjd*j1YL0anc|Vo`|qa?%G00EACHT^&VE0C&-2HpDYNH3O+8OP^I^v)z3-!r
zkE`PkUEpSO!`Hmb%O9!WIWiD)Ou~Sw6-#0S5^+TzfS}?;fI~tK%-O$6*uet~K__@q
z;^7jm`)y)yp3@f(7ZwmYk4Ne(FS;vP#H5|EO32^~`*9ub?_2Bm#Ll(Iro^)nP^u*>
zRVreT62|l{m@R^>dtvEiETCHwcIx%I79DIFDWO_|3@I+~XiBEkA73y^j3>fK3=U5~
zn14iq5Da7_VFP0r3;`hu??L*BP)0v7Zk?;mVmVW0$A-=mJ7eLag)YJq{qS@Fr#0K(
z-Bw=?B~bTm>*{s;+rRlQcJt;f_cn3wO1Q9IpuYjO2pW9n=k#At^>~x~9+;@#>sP~h
zzox>yj28?xoDAdi_$ENjb@jl{hi}ts(>LlnsV7}<<p<yW9*rqasPp2*HXNPV;mM&r
z)PnP<riGOjtP*u}kAwk@2Z9Um0iGC9*+4xm0hDOC0vp(_t#~@bWsbE~TVB_1eXeb+
z+v?_q|E_MVT4kAZiD?N7!VkSyqwR~Yd6}1anU^m~`HyM%5-*&Um!il|JvS>eKKGf5
zs(7}R51e@E;_dnbbv$nW(;w4*zUXo1zy2p&oed4jBYwC+MW2Bui%%nT=rec`leTQp
zs>?;&++4M4rKrX8k|h&McH{cGz469veQw&u28-^ArPDF{vp@T$ed8NnvFkd&vQm^x
zl(hA=s$IRhZdJ)vw1RXq$Wwvi^Il#W%B&C1^yR5>q>WF0@ojdfxIprP!t*b{g;{YJ
z_+<ygv`GG%$tF*gC-DEU`e@(u85e?{eV*%_38(2hfBkDX`PBU{e1oq0t&hKBtl<ZG
zs9EcfwsePnbIz3)l4lsBwe_-H-CD5}N*b4^Nu{=Ey#XG`eaWe5OQ&+K1o8BBgP!#|
zEy-Hvmd{tMTv^vRXN|o(Q2L=9BNY_*gu<Ntu@K(QP@qXh2NozPkwN=|j<s=I>Jxo4
zw&7qTJd3GKS6p#}!y1M*p4-^GYF~Nt8~VH^xjv-5B08wkX}B`zlJtIy5y^X!(<S>C
z3K(JY9sx2Mk~XPn@BaQ>PjiT&3<F&%ot8`<b$H-Mk5~AmM-1RRhhmI0Zo=<m$|_~a
zgw+j!Sbb4_z7~kh3<$teQYxPQ7%+Uy^!>c=kS7XlX6y;~BEw6j9}`(e^b!)TAfvGX
z_85fx_d-xsv9C&}gA#th;xUw|XgHa(muf$XJpI<Q7+PZSw5LEKS_q}s#i(gJ58oGz
zn(CKec^Mq#670PaVXeix)MpG|CkOgAv0}06N<KF}=^Yq+Jawb`ct)zXe#7x9o+{~i
z(oC{>KWPm~Llx3Gz5m{C?C9`7;}qB0<BsqEWk0FgsZ)K#gWqN9iC!=0`<w?pkFnYN
zh5R1F5ELP`B+62v`?x0DR+}-xU_2d<`Znw}?L=z>=#(2Z1P|b8!gJDClIqDWhN}5e
zYb@4>DYbcIt>&TEv%axqsVw_`VMx+TG=|;gw*Bhuf3t(V`<70|Ek$}h;H7sG9ddZM
zD?Ax_ddzsTq-Pmtjfr68xRc0e%_mrrs!Qik)H)q;FMi%3I9zmX&~3RVe%R}VXg7@~
zFc-YWXLV~e4qU-IQ9nEFhUQ8~_Yhsdj2dZ9p4jox{#1dyB-*-c8$OR#94Fj}=7u!h
zOB=yC!9!ypKngyjxD2=lT-8Ir&U%rkt_3%S0aL+U@b$gE)V~l_7qhc&%^ux<M{`4P
zT;FhhL8>z5FEk~t1G<XSVoUVC-aK`MF!PKQgq{LTeZ`>EYMr>DV7ankh2n~jJ9}as
zAMSXx2Zl7FJ=6%j8@!=RDc!=z&IB7ATfuQsqxyPUFemjIhA(&x`+%{s#4wHF2?Ium
z9-%5CcQh|2q`nl4Q7(g<n%hy~Ey`;26h!PLl0{oeh*rgmYI9Wmp&rqp6w$L<r)xg6
z8%Lh%kbQKB?#ULa9?{AWK)ZkgGzS$xO0?^mc;k)hSHvsRqF){BiVmaR_(K<FT~@q&
zb3JFT+`eLWZi$z#6*UeiyCxoSTfa9~%eGujif5GVl{;H@Lv39yr)_OT&)nLw*IvD5
zTjD9}>jiuB^&9r*-~YOO{cEpSwUiZ1&qO~+C919q7jQ%-y6WiQ&`u7I1*5V0Ov+kl
z6H!4L3gM;eLB^Vk-M+EM*O~5fb%_RxJ)0Oecj$Gx8~8Ny$g}h4I&>ZSMS4Rb7tyDn
zAIjy#*VJna>9v!F9jG0oN@V^{CKy`#Zm<C|qtUtV8Fp}ZqWg3l9ox|nhE(>o#4vZ}
zI-q*}RQQL{_QK}?DH6jxR@<0g7{bNBa)pv-{&B;-P#64TONNjRRD4!EvDX$}iq4Sc
z^W^x*TJ;k*nk+41Xu>d0L=TZkI3Hl}?t7#!IxlxjP{kNn-9RRIpAErh@g?E+WF-D3
z7_xWPf=7XoUXXdCdC0s_uZal4Fx~5uZWlviFvbz}2ED{M4XiWhxG<ope?*yi1lg1}
zqxZ^YQl8ScT3Hh<N_d|!x-7;8fAxboHW*+;9H@WLY~)V%-6X9nyrpmZyZ7zhy<b`F
zbYJaSbi+QhVlZHDQKIl#f<fD56QY8dixKe(=<#IS*R{Um3R{+_dflf!=su4a7map0
zrumBQ2qzTZW0P}|D~MzZJUb<NF|hj41C7V7(~-0B$WCkEO~Z|}X|<W?*M@isJc=~V
zybC#s6v@a~L~dau#i+=B^$E2TgCsnRbf1SuyMl2`?-ak4tihg{q_I64O3s1ys{PRA
zcr<BCPTPeGm@*c1O>+hLiu8~;81+Tl;=nwEOrZW~(8p&J*A-Gf5oEu8=RNz~JNK;J
z5`L&2_G5&u23Q2c7w;5es83To_PO?KnVsYF<Bm^q--q4vapjZV_i4-1ZTT$6$Ca14
z_hk#x+}!XrFZ1$8EE5SA2?=?nWxV?Eb_?lHVAgm<VqAjhVyj{?!#|evFh{>f*C~8w
zLrDlUkxrx`ggId-!+Qk}7&k28StADI%rzbuHp1`G-oCX?YT|WBRc492nAO^fRn;!i
zEs!pST)G&|5ZqX>lA3~zv|%6!S}-#5QRnh`7mi7Z!MhN=5r7cT+$)p!ig{sBA*VBP
zE$s=)hq!xt(D#wrOuz6jWwGyq5&|`nu;2lMM|dYC%+daV9Xxtu$B*`GBmt4|_#gi8
zFFeQl^769Yk1&bw6<`<&Mkp59s*XQm*4}vuivo)uGldWAFo;tOj;{Ir+{=gm3C~Us
zU(C9I%Fi%r{>cfh>G$Jn(@*{JfBdHq#TP>H&CbNsk84NvVCRwT9Uo{x(6ciHEiIgB
zJ8)u=!=efAV=t0&7&8o#q=h8<4lf@Jt)xOk_{`{&{n3aJD;5ftFBP>gDOi!uV%~D4
zyroGe69eBQIIvh}z~*IM=4D>~c;!#V@CE(;yvobf_3Yg<Q^02~PgUbz#d|)lGUKnG
zs62n4KRRWeFDOs0eDDQa<K0BXd(@N;&cXkYp->Xvx^>n5?EBxcZ+!htyL0D;8?lJy
zx~=n9HrH%zwW@O~;;~6jF?99nrsc&u(`nLsC2VD-VjCN49(`V}l=O<R9UT#AGL}fd
zatuDsTXlGYrtQ?{3m1T#`^2yh&m$*%d=VUU8^#FaC1BuZr_Z`9^SB!S<)5>6ys-GC
z#}9e&KDkzbACLcd;oFm6DlcsJ>{)&4I(PV8eE0i<y!A|7I_;ScChU=vvSN|Fv4T8?
zEKLMYvsrhgUldPglxK@k$)$WJQYU0pz!AuCL$X${r)L^<S9YVsOQlj)DwkZICxujh
zFmTxnr?6SH#c(uqqbDBxIKc6)tgd79!YkgDmwI0&leM*V460jxoyaKCqT?Nn(wKts
zh&?{y8N9dG?fT7I!m+a1`A~J38}aTxcxdfT%gSZRc#`+PM_|KVvFq#WwzYLb_`4ZW
z3`ssmDTV@x_T-8cizbPp!eBL28y0mBhA+wEu0{|{M!^dl#U0AL(EvpwY0o78qiu7U
zmvI`{neUf^(iF!wtpT)-z=NLkjl@;m%GjaU$Jmm#%`02#(;Zt`TXz{ALm2y<p=?7j
z*l9Ha41KMk=iGS1I;w5;(?jp;#>N#}Tia59D>}dA3SZ_p5ekeU#uQR>c?u>Bbqk@d
zo@GL9WZVaXrW?K(zm(wQYmsqZ?>N>PXsAB*HJ=ZBTv7I|tZdk-#*S#e!7xO_P}eM)
z?MyHNc0@e+dcoHptmhE0QJ~`a&)f&b0Y@263^OP_`@%as<bfY+CGhZE>obgl^nrtE
zKvXJqg7*t3j-$f)cu}~L7Yxl>t&T0PUbScfgWr@6U9{|Q@4o%!x4*KzgB>?g_6BFV
z#GJyI%UW|mYlfngN-MS$2?|W!bH;vI>SxrQ8#yOj^VF2kRSZjNl%Bmf?^v_8ZwCk4
zYQw4C86rr6(<gA5xY5Lw&7(6tAEGyW9isk@Fod;E?C@w`upj#RGnGvGcoQi{goUaR
zSn0Rm=k0?MO!o;=8!*Z(F6j`gh%#co8H{W`?_(JaF_9U%@4RPY;Rw+JtX(k(1;e8G
zo>$O~Y`*9QMCMIA!Mn7E9R+1Lid4pe{smgTpn01R{VCe=%BGc8u4?Y(v}D8BnovKo
zuHX%deWG%Xyw9xRh-ASCM^s-Z;FRXj8PSJbz2{PMyx=rzK`_CPTPWr%lZp#JQ23JK
zEGW2%ZtzvG+Bi`ETFoQTB=!Q+IB6UlhTtE2c<Ep&b1w!8loHMw7$@OB#`m<|%l?1N
zL5$6N+YdbA#1+tL3+sE*@}W;)J&%Jf26r;JK=ZC#xngg9^-X)@E3aElJk%p$gj0;~
z2;GEAz|c?vB%eKARGvOhlcPE@gI+eVXj^MV`})`J*!RDOvFbH@<@PnZeRIoh-MlLP
ze$B3nmtVWKVXwV<)4uZhZSnu>cJunC%LH${cFSJXd46BrTKDH}>$$a+vgX&K+OE0U
zB}E|Vjza_%hLL_x^y2W))8K*A=`@kxNjF%b1A!j6EaLM;|Me`WgDxemiAz)!#&PBg
zQAi%Or0-6jQlwqQ`64O)v1(uW%4?!Wt4^EH_xKzZkL$F1c6`zh-8!{;vn{?vgzLm<
z2{f7LB#%&vYTVQa@R)J*2u9&SaHbCFd%NA9wc0Ifw>!r685pwA4MA&z6GzD0qYHXJ
zdL*L$hQe=M6JAmRFVLJ(rHmN{rhE7A+M|c}#XmEGPh5DY`7HhitzTTi@J9sdQ12WF
zKQ9D3bVh2c_@^xry(C`Q>*!tmu)T!ZOQS(+#4e(EF~XuxN|H)6L)yiha4+hzA1QzX
z518%|DOxlcUplWO+mPdt`huf(CSMgT-Lkdis;yKd4~n-ID$7n^*tZTofpHCZ5;YuE
z+s_8#Q?*l{(@Eb?pMB~%BzHuRO9sA2*5Tp4;|B5t*vt4}Fr2_&giAOh0B;|om|z2a
zkc`RulIeuVA5IjaDS{s~34{AUG@Y><2{t(6Ag_2tmGE{fo{n=3DH7wFFoSN*9^U_*
z?d@(mt-`T~s9WZ7A_l!(5*@GE?(Rc9!&vhk;iG6V<3yT8#t7pp^cZ-<&vlc}2MobP
z)ivrTuEWIm$krz4?w}K-_ATTqmM>NX8ypx^W^#q-75dCvzAWBMbeiC+ej<bUUW&R0
zxkbdrPL5CP!NYBP=biWMAOGop*v{Ue&>=+3!(Zq}ka^*~)U8AB3H*ZJA1utLa4+AV
zz31b8zo>A}^T&ssfBKl+^QT-stWTfU0sojAzUF0K{>UYs(V|z2kdy?KMG1fK<q6(C
z5)xn_;vK#ygrPZ6)}a|}Loux38D3?<Sj6H70~rf|sE9ZQFN6~}fQh*T<5a?hPS%!q
z&=CHNm(JndzO_%-sF{_5S-P0D^2WMV9vi+`afvDDz*t5bT_Ggqiy%}KQ({3*E_1w=
z@FvAG6eE|XYNY*QCZX-f9SAxLE<nKqAxIGx1v*RpupM6vt6yFR;mR(2Nyri-rk%lH
zcVLh1-nZkOJ?qt5dPoc~nXw=K`48=X{-6KGZrr%#5o-~>Bj}XW4juj?2^xBppI;P^
zoq6efEX+cibkH7{pw3hLG5z5=esKr?ILU&<@8UUb;uiwg^z-pK?)7v2%MDjEqI@}h
z4}M21V9tdH^tZ0X{ezu{wtJ*;k`O1+OJgk<QomptkrYQH9k?K|Ftpd(t-1s>(ng{h
zqF>;Xw+AJKzG5hoh#~=j1v8=UQBQ@CE#%!0Rw`or(qdVQIfTdyArJU8j{=*Qd6}0#
zQu&iGd_i-+cq*wkhX3LFXDKgu{F50^KD_wTFImTvyY=nE9(Dd23g4bT9>3S`6Yuk{
z;R<}86Ua)ekCO=$$4gp}k6l?&spjmv-+9Y!-n?o=2qhB{EymezG@{=z{YHt0H?Mf4
z7U5h+=|q$e`|{v9OOybM^N<3_tlMqa-oE6PF6j<~Ocz=jinr{e;Pg~~L!nR8`*q2m
z4_|fYpMG#1hnm@Ms_=Bp+^Y|dz~PGyzPsFsEEMjT-W#sFaN#u}x9K~l9=9pHHJp6D
z_>KOL|6bMeg_GVm>fj@MQl8vDle4IC_Bp$#+I7v}!2>#wSCCn`1%(kCw3G6ilte_5
zpl~5-1SJiTTqsTOfJPBbiY`3Z@!GG~PINEc?Oj*sp+G`rMbRAgpukuboVQSvqwvE^
zJDEyLKF+9~xVH-<Cpdv8K6}=(=M=IuavN>OQy=3O=TLal4mW%WK{1*l|0UumCDVrY
zc+B^EB9aLMKvc5vJv%<ywt8KCYqs3U>E_KFRxTISMj|xEo@#t$b<Ly7);6wsYR6Kk
zZ23ajJ@eU*0~n0ePxq`Jqoia5diF;mf=2MY5DZ-5KG8Y>fpfuS^IHAT#^YcBL%BE^
z<CT6gO;tzAM8x>u*hLhQrygX?7*jX)aD}<a8UdwU&XsIFhAK(jkMXbFCJh`4&oJI$
zbN-R_JI8ugOMNVObY3F2CK+~F?^x2fHC4|@FbE1`qPA%N;v)Uh`}AJk!zZKxLuu_~
z*Dl6A$;KJ=5&Y@9fe+_BoXAKOh_aMuishAcD_4m)N@$*UJ*x&0KTJCII$}+PvYPaN
z7kb{8MWj$AT@^-_vyL0y^dg-{={i>X`nops$OKm)>N!%~liuCa$xHtCH1=vY`{iKV
zT5!Woi2Ct#k{*2pe0X8ZNWDfKY0>~jf<jfjwaKV!t&U(M97BQH>a?xas98a4g5|YM
zE0k8PT&fBd>{pf)l!TWN>LtZwOMh7x5}Bb>UMI9g{Xyx8HEJ5Eb-4e~BN@)m`VLb8
z-Pe1-iF90^W?%D#C=_s(v{-!N{5TkPtfz5=c2}!wf<fBHl6OLrRF&X?(<f>E*#dyZ
z2;atMj*}r()-+;3$7I~Glha*&pI9=Zenm?9juBO>qcMhyGk>@5jiq_X9yfNuo;guh
z#+K9uK0R%i#`5TJ&yJ4|yuTRZFy>-J2OcPep%;Z>$r4$j<=E#=^IZ3#fEBPbr`U&W
zAUF`ouqb|#();7Su1A3zqPb{0@C0|X)*q36mz27q19}hi9(>jxRXK9SHb%X?=60sI
zp|$<0-e=a;9BQ{}*69(AL1dEXrJjulcTl#Qo?TG=j0=%%M3;10r`Bm6S+8C5He!HZ
zie+^Vc!a?WS}T64maqpn#(wCTU19_!qLZvN30H*2w6-KG3u|urSS%H^cE4_4|N2+#
z=JhR4dAhrEtX2t%Q<_g<Zt{x*ESr6&0GU!cVOLg5wz*!ia$a;?JiA&<*_Dl|t?B&6
zYR<m-^;hk$|MQ>Opa0;S_S!4gZMjNxcEYlmxacw|OvUG8i=xjl;d@-Pm537zOgI@w
z{d^+6#FpEZRzHaHiKE96{(>`ZC=tJ*?eJXqYD~NYeFoPszLVaKNGy!itjjT~GVjoj
zxZXy10j<U<fjyas4oj<VQSvXU!_0f;gu_qwQ&&uLoB6%GQnuT-Zo5<R;n9)R8+FHn
z+G#`hS+iQLZOwL1M_=Qwal*(b97F#?okU=Ih9A8HM_STcmX|B;9GlOR#+kIM%=HD~
zCedl61nyf)Jf~5wTf5nG<8{B+v4QZ0=ql(bN+{=L{L_3ye+4|?Eg>B(d;cxEQK#E&
zX{>`02q%S+=pQ;AQjS93u<^jpxd%Rgv4+Sl49W0@3-L=}nM@ar{rS!(?CIFE1<@pr
z9#y}HM)Ops&;;>_z<&i_WQQs4V_2ho%he5AUAt;A(OII>dcC&eB6Pu1@2OAVA^JL;
zG(4qcjmW-Us7p6{M3Hc2D3}}_?D-ta`IaZp=!D3jXuRek#$V{y>FJ5)isl(|NsOpV
z9G6KSns+*YVF^)ZrVm^a&wxL&pD6U0XjmK(pbPYkXe=U7;eGBz(AHQ<4niNesQde!
zy3=OT+Aa#0+pS~Ue)z6v<guqMEaZ!tPm+_ODaUPek<F%PNVD#=I9Cu&EU!8*3NmES
z_qicepTVgr;y4ezN<&v&?lD_gUeP;Ngja$udO+a;{bf(c1)|MGUnlC9b9_&CtA42-
zQYo9{9~=yjk&we?`clg2(V_k8zy6#3^S}PYe*Dv4xX}xhX9x%gvejef3*i)>e0i)J
z<fk7&MtgER|NEm4%8SlFeLP=YbkB>9r^}13f70X2QeM=y&vL*Q=7z6%nU_CunTV-G
zlX2f1JCzPh#217TZXu`;V^6msT7+<iAPOTO|1XQKw5Rcd@saj|4JUP%COF=Tc+Rj8
z!i$6FT=3L2JU%dvVQgYSc6@kb9W9VA@Z=LCSW3k`;_J%QEiqR%8zwih1PHZ`Fk=6%
zx@liZ0xZTZH*~qNE9Ccu>9g2ku`ufsi#5Xh5vX8-2uzSz`h&3y;feyYLLi1=BH+7M
z$#l`%@AM>ew>_od!J{4P9G^<QralDvf_?29-?YE}>;K8V{jG0%UUA;<iY@vDY+zt|
zhrr;B(;ZBwL=pD<*bt@D2oiqYfBREYrl{{+b7mMAqyG5ya6iBGQ#bL|&vRmWdU~h+
zaghdu=lRV$fi$oNW_2w}whwk~|9IcJ5=w*NN_|nkFk(8Kfu96kgkEZ6(Snc)95J5x
z0#<Fs<A>0G+USvcI`2kS3059`C*h7~Jnf7GaBsW5v*|3+VufPCN|mBVfU%$)4}~xD
zGB5KoFJEl=;tXHk`|kfSCD<>0K#RZV0{;l*v%fvm@Y(A5Kpihxo!|bXQy)<RU*mw6
zJ~^hBAJ@SN9r}sd1#TOZu@{nGBDS`+Y&UOQu}Zn%$^?vCZ@h8Su3ldke<e*FDQ4KL
zoPAmbswWuB;FWmw?(gjydn*yqvbVcqEqzD1jTbfs5D8GO_`!(KXt2)?`$oWPQMl7y
zwdrwF`Tf-KxHBItkBwf`GP`$r%H<OkWQtkgU7=d;4c_}G@$jt1h=;-m<^0U+-0RnO
z6jh;`@Qx=nA`i}fJvsZqLc7AVv!9QT=gTwasFM@3&(Qu*)GEE(WoG?DJPVg-`?<y$
zBc8r%tbjer6jI)>89Mu65tV?kvQn++Sh0MrXfer?zy$dUIhlwj3}cO2&5fDJvdF^7
zrYM1k&Z1?cAF9=BE{mg(!YkTkV8IWg7asX|I=dn+!Ct&^S5`%IV4y(eLoRd08%}EK
zq@SaYssn}ikz~Q+qeFi`ayK62DAGf^IgQErnf>lpKek5?-gnP_Way1eA{1`<{U|aS
zcNA18F4?>uWg|zqT=u9HJj*fC0$U7Oy>82mF?i}DXGYQV3Z`Sy5Mk&@kv2(d1>e(Y
zVB6dGh2v-5H`;KZd2w*CqxUjKv-hWh5<a)A`igoM!+_xGhO&-bs2&u<q|VYs)vx~Q
zwM6ShF%%{<T7M8-f}<b`GQmufanP^Z&i1>WPO@02TCTKg3z4ul5;AZ+ilIzntMA|l
zQBhzrky`Ad7KyU?eu!)UwmAMB?Qi>@VT}Fp$*GY(9^7Dk1iVO3g))<SF`Q;|6+efP
z8RHj*&*kM6%M>cM6eS`@V?{KU#xxQa+}tR2Zlgg@w1s_PLR1lJBGPv;u0D@ZLXX*h
zg|!9IAv&izkmvc1;RxqH_Fh6Es9x&edFF!d_1_fM2!e>(MT9{x0&&0Q4=@(3%4RYF
z2UTw&pSSDRuUWaeE?mXAQF!O$#hPV7umu0u({47tdZ*7vs)RP^yXKSzK&NRx{X$Wz
zb!e3RdWah7YHUR(Ny{ZTl48r#MC$&K6@)#Kv}RPL?1{A)xwPf#s&ID24S2kh`52F<
z++fBkiYO?L@Dq)N4ztz)PHqqa01gv=Gv_A89!qid^jaj^i)$EgM!r6cq8wjJ>zdZ8
zqWwfUMuhjt%(Cy56JxU{^$Wgmu=mJ~>eXu5ip7fHOe6{E2k8&Ew;=pTI)2S~7VBjU
zLhM0^(bkPcT2EsHz~GAGDVbJK($bcNh1nQ`E3E%O4|E5Yj-NPga4&1Rrs&*(*7+FG
zJq==FYA7ciCQ)wccUHKJVqGYvK9jN)+=%Nt%6k1Lbz=%AfF<kr9&yLk>kpic<b~G@
zk*ppJ5tlv|iC|>Q7t3y_7>_SJIyI^~7KNXT1CeYYk|m=xKjYQc_qbq|^OVB${o(yb
zcK7|e!ux<vk2!!N*ON~Q*BI5S>qYy^|MX|}U;p;6?FT>jw!QKCZM%JQ(<8LL_06~J
zTi^Psed`-<+IRK&%AM=Np`0bf>xp*3U_{D69~RXJ{Q)mr7o&0GH)HPx9N`#-8|W1G
zazG<-TE+lF<P){{e0F`of{($ncuh7Z8m9S*k=XY}*BTqcitl}=d*K0OYhheCGP+)=
zlx=xc_?F8#{em~3>){<CvW~sqF({xr5{=auux{twj2VVYBC_hWruBNlVT~1$e}myj
zbo0y|Ie;Rx20j7pCL+%>@aPJWO|*sdl|(GgwwvI*8^J^uFqZKS>SBERnt#1s*VuQ7
z{j~=AWa|d_HHILM1``dZzeIJR%fKl_v(gP_?3sA)?t9j1)*SAGK9Nc1ZfL|vqmw>f
zQPGXXSjr~Df%y8d?mrEN8^IIUkfJxFC{;U1pQ<rpUo#?v78jU9&;-?}=b<@K99A?R
zf+6`redwr<DNkLPO6P^<5zVo#ot_+P{58&M2fS}@XUB~|@G|z$KR&6sVF=ntdRpJV
zUT|3ehcutzDc}lDCParp=f;}*jFqPk6?{pl9CT))4{9%Sj{VU%=MG9jdT3xu=ZU7m
zDV_cPzzxO$Sz&PjL!O!~yvB)VjB#)&;5SA!V9iz{%y&`*x0)wT!+-^;oR5z%<{t@O
z%<H1&spugoY1N0GWMcY^@e;jTQ8YQ9FR5;#U4!ESIKnv49;Pf#J#GkddXy5lMv}|=
zJ|`Ft#)fEP@XqNZ7_p#!4LjCpwbairrcRs$>&prFr9Nx?z&8xIp5aPQwi|W(=}+Ib
zzx%)b|MuSd_k`2pKY}$8t+dgNUix&Sv*?n~eJ3nDP5mnHG;cY_lM?O=b2~iu^zp%S
z;h7H}q418+UP7Bbefg}<e_0E>VQ%=EmwEXk7Z@K4-g1>RCrO)#-mvhc;JM+#z61dT
z5ofLtN&SPFaE_e)Fg^rgJb$2|kWLI9oM%CVaD-4X(RU1HESeBH5juM9o}Iekt7+$F
z!DB9#lYq6cZtFT&7;%n;r3(tYM@$>fG7k^e`$=Osi;{wQurQ__w3n{1=M)Plh^c$5
zsX`Xx2<5bsx?q?rd|;v;CXZmCIuY(rJjC^RHaN4>qa!=sJFuheJ?kEwh(}*Yuq)cN
z>o@F;ue@o0`B#4>q4rxIW{!uTFLLzWV5AC45ehI23Yb~pI)ZDs=2!H`f9pB_>5MO&
z<FEhY9uDf#0@bhSHtvL7dYxS7y#Ec~7(e~-`xr4k8JDqO#sUCga5z4*!`g}M97>Rq
z;DON#SWcIdQwFWK=mn$#;Vcnc^x2JkD%3;m+&k;M3(LY41P~02Tn`ak;01N@EMp+y
zRpSs0kU?<CWl5JPfh&q}5}^vuN3V8X=4D>y<%=j^nBnVt-}%n}5h&xQJN{5*W{~*c
z@p1(=^r6qaM0L*3KBw|&>SA#kj@iG!SHm&ox>W0jemVce8$T);jHs)Tr#AZbH{Y@w
z*Ea3S)|%b8dBv_?TeIsouG+@Ns*$$v(%5XiWe0nEwzK`vNJWKmwO&Kn{>bj$eczRq
z$V8+gL%HCd`{I?pw-Fnpi|1lkxLk<Y&cUhmhUdB-JgaFJif$I*G#Xw_d%Y(<=ISLf
zjq`rpp#Lss>A3(motM0X?1WtK_&m4zGrFd`Bx(dW=pGdR<#Nit_r0&#U;M@Q?6uc!
z8&ThbVTf2GyMRMc@24gH-+8{yA}TEC2a0W~(>dP9JDwg7|E8N~zqrc3r$sR25HI0j
z)#K(Ked1sf>I_9sKJGuy=-ouja1XHKMzsM)so2tmEmzXET+X^;)A!anCo<;3`op1}
z)S52SZf$PZ^=sFReV$N03i6U|`o>-}7&+O$0&jEPnaiQ@E9t(35ov=mj5e_`KKsnp
z>or%{vUxn4^8;T}TcNyngAxV+N*Dw^8pf0kB6DFhMdoKuD*6^u>48f_8%;2zH0<Q`
z$Z?720}MgI0gC>E&t3b)+y7?w?!K$~Caz$rR;sqLvZlTRs|71pmL)fr-2m9`V7v-Z
zB&1ZNjkJL@U1Pxj1z5XLQ&WQx#-kyJ9m(TOZxgaT#v3!>fU{ddb;R_JRm%}slFC_A
zb&%E$FMJfmq<YI0Di+UF)$XF6116eX{hA%_J#aYUiB9UOMI#z0q?<!wh|y>viqbx3
zkyuXmI0!ASb~A5?L^(Nrr0OPCDlhB(MAWGM=^j>#*t-_a)s72qKT1H4q)>aJ@to?)
z=z22Tj|X`qZTVcvJ>-A&+uvEOQMcvlvL&=O!m#J5@g#ra{m)n@lNsSeFm|Ab1W)os
z(lk{pq|akbMKlevw#H5MW0VW$QWi_lFFfVdP{C@EC_#*hiyCL)kf()LojBNmFYRvA
zdZI%p`5YHe<}p7|<hpPU{8)Es9)$_311CdpL4EX;gu#e~YG^_2z;H%XNm4kN%jB(6
zsaQ$)Ld4pF<`oJ?o_Fthp)gU03kwNfuVMTlnlh}Ln7^dlqI!&LL?x1WByFh-DHjVq
zex&3i+Qieb3$7uZAn7DW?ocqxXt6{JQXQ<DF}P)kc(~MBWWqWU$4{a)bf5ap8fP@_
z`#Lq9E`{eVCYsAc%#pfJuyt@iFJl_djGm2XEu`~HtcAVRT=wh1sDptLIyP|w1;)8>
z4!E7Et(p^O>_awm174v}vZUG=Ph#xB@Qu@|_<-mN#*&!YM)Vjwh$tUmL-vgx(N{Xp
znhPT&Ygjj?3)n<Qd32QlPddN>UPslZ3#U`y!;)Z_%oNndjPszAqXUf(TL-YFW{s$^
z5j{b%?MCDrI9nF}idJGgjT421(OzRM8ca0l!h+~mBB61@336S>nq>=RizM>X_JBu%
zn}7yv79yITG4)?G$oC@Dr+Jiy>T>)<j}UR#qo^j;Y%I<ghPlC(3CG7Lw!O10zS`Ef
zfl<#Mf%1n>%4#KP|NH;`zuEu#|NL8f_0?;(yqvSu)uLV7TD7ZJNH<)xVj*Ld68mbQ
zP!EyaIJ&t00lLiC!>@FJKj2<ss<?+%<Gc%g#uFh@jcFAVe+dR#%|G=iI0(cYpD~bQ
zl=5siaf}=6pQZJ+`pi55zq!Zn6D^uK{PJFyEpr22SY2MWE8<gyGPERO!+wbVgl3aY
z7o!?FnZ!)LqH8#IV4NYXB9qnW#)zyj!W@l>AgyyCx%iL{_ETrQ&v*e>{vew`FI~5w
zdoc17=^l4o55^OI&x~|A;eI;ILCH0vF~(TU7(EX^8hF2hQH;4j|EZ7uo$FpAVYx;`
zHF#XF!#j4xhX+oRc(y<23nh`K8sa4(!_b1}7c_f8^oPBu-hKBMc65B;(QXOxB<4E&
zi~5PqiitkQ5;!`9y&5kk=sM1=-Ky)|9XG;Z%w3FRG{Y{e)j08XW<?KyNig!84Fq3|
zMi}=SwNv}?kN?&7_a9g~1ziztTt-AYi*Ar~99G{(l0$HCAet3<Nt8s-gFEnj@I#-G
zq#LNCqUp>}j7{lG)(!Y*fn2{Q{t%o)g7Fys&Ax(5qU$ly9M|6oS4kBLt^%jA=6t6^
zbSHa)1$`a5ry(%aTqjLwSM=xT=+IMF23garDY@p|cYp7D`xEs=`a#B+RHy7Ah;s~c
z$LW)B59ew2SA(_(Bj<&#2V*Kc1Kf584*kP%eKZ)jGZJ$PL)=8;ivu^~13vWBW(;t&
z9lUgBlQ^_ja2&L(ulEl<`W)DZ&I<RKi#SF7<flKk|Ih#T|7Q;#?KtgcUSdpxZ-dY1
zz2IezB+O?zOv@kx(}VK#?_kL0yWTVX9w^aAl@GuEq;Be>vhW$cf3otj_kUEG)%C|d
zKF=~Ye9g<ed>IOKP=q=|{(l$-f|Dk@@Ashp7=1j-OSfSlf+@f?AP6qxBY5bydjTT)
z@tB_FS`g6mj2I=Jjw}{1cEA+4jxnp%YFq7ylm&QL;@Nd!#g&S!Y;8%XExP~)v%(PN
z!U|p_=Xkqbc-weCUXJ{B1&rEB{V;O`G9rl(_V82*UQfJRg0}<+?#Ba$#T#BcM4x#i
z8MW}KHj<7J_z;cOJgV8=!)-e`II=;zYbi0lYuB&aTVMT#ee;{&vTuItTlTfDf5SG{
zH{44GJ_g-HsBp%lwk;s-srX-i7lmuQ)4w|W_;iCAr?|uyeY%&pp7RTo>AkL8(tnSn
z(gVIn#;i_0Bou{vf}j(=&VK7dL&wm1S^+IG5Qyty@cYMmb|isogn>%1xb!g%DOzCA
z?$M%8;EF<FtT9BmL2(kipo36~aPFSN`VBm2Gr~T~Av`znEW%67y;g-QoX0B*;fN&#
z$`Wvsx)BJY2`$pKh(pl?j&VPIn3s8(mwEXEmoLun_1*7$dE{gm#iuDB`54S+`t?(k
z@aSi#U|K$HJ>qcs3!jD|&zSP8%&v+Ti+}lb@oVz`vpCLWQg-{!bt{#!Mhc@FH#hCh
zt!tJ}#VwJHS}M&(;%wT#<dIM)%lrM1w&&>Z$PNz<4DWPO3po!I??#CLzeT<x!ehBo
zb;S&udbb+v6NEQv-&4@td$6O0_<@bY%8^feA0BuuvVuNwfP&AvOo5_JJf42S$LWnb
zNP!d6Zo9l9K2O1rMjsspx&=6(WSISiZ^P>W3(B5bw>IoM-+9aK+_`D&1(itRG(<}6
zp|KYPvOzAFwqhagiaX>$6qU$VywaQCe}Fl%&(zrE|3dMFEdQoAPg}zeFaGWE&rlDI
z<xXUr*>fEJSASfV(m$_Fa0@WdFZG@516vzad+p8@yRx}#)oRJAm6AJ35wSw{0rp;6
zUar`iZ@gj~8|$vvL~g~{)vO=;o)wQC-n9VxKm@;?oktq?uB}P7+*rTrMiLa0GdY%Y
zcKiE#c6xFmTuR!;=9cewOF!^pCfyeXNuoSZ<RD2B88GWGV@W?;k;Pd1*yvcIFWhIw
zzAoS)hBo9r_P8Sb6AEsWX0Axr*f1XFqrP?9Ewy>z#*p>3b*rwdxUwgi&Zun#;d$1Q
ziL~$>g>HyO!)u)WG$eDfe-)nS$g-sFV(rlGwAG)U9USgktKBf7I`H%lX`WDIqwI_7
z9Z8ENFgWGZKD{HkU>VJ^d_E{4eIE`ywn;A-0UkViX2ag8WY~da)X%)eNno&(lS4Z>
z+BZ@+F;6@ik9C22?=#<YPBJO`m@r4jXDziCV^-2K8MQ&}0q0RF0w-Wj>NgaS{eIVu
zHD(|asr!iZ3XyZDNSI4tn$|h-oZxw4@4ox4-gl|_a>GcY?h0eoP5Qxu!(Hoi+g2=*
z@~-U0If=+FLjxaI2C6#n!0Q&isv$bZLWOpL<J_a~evUQVSf{y$fh8DdSaZd+p29f5
zo=i=R11(Lbh@!#Z&RS|9IB8xiNQUpXTsi5+uM6tqydLmA&S38+rk|nDj<2L)6K-Ir
z^!23X3wv6zFAYY2Xdlr*zP|Eyg0sS3(TgRFlS>%HXdUyD>r>d0b_+w92AGtXq$Ol+
zK?V~VHyn<H!x-0y7KmFsUDV@gRWJP3IEKhJ6uc;2i5Ou`ih^BF=)kZIo=z@D9>JN)
zEek*Lg58oUPr2VC3uu=c1Zf*-vTz2i+i1|%`W6EiZ3RXcb$}Q5vmQd(sb?HlI7fQ5
zrk!fNc6QdbxCY)MF9b^;vq;kC6_LI~0TanUBnfFUQ5F*+1MgtXOMA}Lc5sI4D24mo
zo~Qi7NQH5ZG=#x$#QhX#J+N0>&WD04P8}J|Pw}lqa78`UdF~^^4qh-3{l@SM?^^P-
zh8QJL24g@552+iR)cilkAT*;_3vQT=CCaXl2d3eCxS#9cb9P>5bP5{n&ZGVVn1ah8
zs&c`j`*2Jch*owqXGjwpcpm9jzw*W#qNy96lD6Gy`&x7MCf*R<VwOuU*?<14Z`=R&
z|N4K}=4Mqivg?Zc$kMrJ+oF$4G{RHx&WZOQ{R4YjVU)pG$~bv6B}#F%iE$*#h4G<1
z@MHK6w7gWRi2flzU<^#T0gC7?cW_X@c!qK2H~5NCfD9_oY8*;{F+7KB>}iPMD<>XX
z5PimZ0zC$XgGRIA?cyBt`$EqViPdd&yx-Xj5n&mhryLld3I)-4)j5+bBASQZC-6q!
z!Mn*M=xm4{BGq0t2Y*R=-)I+egudfA&0LMeFiNu5I{U(rhOjF-K+4~dAbO#D)L%FB
z<@D~H)2U0n3%^=$yw&MS{^_~lYM_3(enEqYo@k`&7$wjN^+oGXYvMuTL(q3%yScGp
zw{PF_z31V7`}@0g_q})3ZyX80ZydbQg`IhH?#}Lg@zRNp5i|sPPikD~m#5cF<OCa{
zz|gs2oDY$}@PtmM<@mu^`5uKj9~C_fPAJS9(Hpfn$QwZLLSqbX7k$z5OQId5&0Mao
zS+TtCP9{;!$$I@%caqXI7&VDb#gJAIj%N!ccczF$MKc#k@0-%6Xq@;Yb%PritWS>*
ztRdKVB%NT-KBCZA_&oJ7m*IItXfEm<3BihY#Zo!vx$prDf<yo!7sL0NO*jt3;}PL@
zpr_y+j?D>;A*qrZ^_u<Sm%nnuT)9#aepMWAm_rzx&=nTT6(4&Xsxguh0T?1x;kUuy
z=rmh=1bkpTkaIGk(WE;@wgJze-N<j5ERnKV&EqrS!l~pE9nG5XQ}a>%bDbiN9g>lF
zPSaImme3r*AtvH-_x|Cb{q$$QvfupnJtq{dAJYMCU+{enMVB<b&ZnF)&JN0qvZ%~X
zP0Q0pnWw&fPUYDOpMKXzl}~!#f7Aj;nj606WnR7{C6<v;B*x~RGB6^iUl3ghSyN*T
zGz|d;rh*Uz)5Z97E+GtM2n+^+jKy+7j1Fp!R|!TGlo&jN*A2oOsSa7h^xAE!YvISj
z&6O^q{mWY$wt4Nk7;#SgkcA0~0K9O<dJiG}7}oFxLQr5~l<+ok9WON&?g&P-l}5t=
zz5V)3!<@x*5FTNM!Rv?n#IVFrlVY;I2p6-)=tcdPlS``|o!a)j`*!eX&y_AK)fM~Z
zx4v!v%YXf^_UAwRbNlMozG^pb-?4J3B&LRd8s6(Udj$_*rB|xO0ghU?NZ1zBLnzcg
z2e&TpPrT6YnHj*i&V%%VQ?6_w{EG8jP~Tlu#S1v7!@u&Kr@5b7`8~abKP)JCqu<N;
z@P%d1rS*n=+dJB~11<Wy5(3Y~+==!=fJ8ut!36`C7&z@#U%>-}KlbV38jRm@L}SJM
z2p+-f7>}@!B8FyQBxBD6o(ZAz0iQh85e*WzaIZH6f5Flh2I{^GCK8krN!29?Ah=F-
zH1jeq^D-}gsPcswzW(g97`~v#)32XTd5PL|>yx{m`SF>`2YU1w>wR3`OV$&3Yj_Nv
z?F@-eFSE)*?eIQV%<7YWL@G<fuvcEaW?%c-YqqjlvRo!^*=$fi5lw-k4#o$J7AQ|h
z%|#?vM+@l?DShq>{k8RVTV7f9jkQsDSM?q5cJ@MI4;RueqkJcwVI-Q;eI=WS*T3`b
zU3>fIzqaw@!j<T6y#9*ays>4=)sib2`~8lsu9WP}H(%5FtGe&Z*d#xQRLB=%v@TDC
zKG9K^btI&q93|3>s0^eV6ap0FJ`@^@ixb=2sMwodxotPEZP>MIYxdTguiKTaO(V4r
zm3j2i-i{j`&Luw<i&=a9_1l8kO<P;V=(DIc_B>?_ilLcd5qV|?eQ%flVjT48@=h3&
zS@Ch44x8(b_|zTY)Xd<8azbaP!kfRrKY3nY04RP3%UU8(xN>1-!SdBtuBm^^Zrsf0
ziNeZSL7%yN(n|T1tqI0AuC3ektDA0A!a->;XsiFHc6hLD@4fqb`^|5DV|zQh>T|+w
zOU5L18DnrZB1)y}aA-@8M%G4Ay|s1KmY3IE0fpB(3PO5~0tp2a${G}_L28`1oJB`a
znqd@jSq?`wbpu(J{ga4RVb85}I%jc|V@px>8;|j!+JX{@{l1763Xwr&^%r>!*x+rx
zer3Z7rJ^gPO2X4rDkoe|yVrI>{lSwQ?|s$)rzc0ASpoUCSXj2o@)awxPgW{#z24C7
z-MeRZ@BUu>>RARQspbsuM`s`KK>az#VG4r+V?q?zlI3(S>E2MJ&axVK>O}Ogx@|nH
zTd!O9wiK&Z)t9QVpUI%t^=K;;k<^2t(Y@RSFE=8AI00dV^FaysSfQ?eDD*HaqCj;+
zhM~-i=pK}h7zF}+BTETi7h%+}1r)LaeX1{B`N;0R`z!nPuYPWc#FAaTdPU=@evE_i
zuGy+vt;Rkq6X97=<A?!L?K~5XxEDHOHSm3y7EmmXx;E@#kYc|xwJX9L1rNGUy;GkH
zQ)3kiGmoYL*JnaD(NLrZ%vz<oZn?sW+8@#@4up3oYKMcS?d?Bs1AwQs0}g^$h$y8#
z3=pAzMD#tLgno`FAHf`jqTt4?L_v)bes}kQ`n#_)!BC5#hW*|aFNUIDqA@5HJzXG%
zEUh6&ZQX~Fmciz|^i}W}_pD1c3$-hg$+;nxwG}iL!#e8@(r7L&%_53`1BNxVM|Aw;
z_{f!;-~@HMLKS1i*}$VJpnaKaUU;OntlHucOu|i~!V;;XsskU0hNJHnb}{K$zt>bf
zS~CkNF<b-#D~Dinsebys7<vQVn{{n6tXZqRueJ1^#<k(fV4{Fej!73rB%1Jn+0Nc@
z!U>`kh=`1(OI|I}4D8{?-aHtnpx;T&<wUBYR}g`zwR*3k@k1$1v>T~YMHd~PDHsVw
zPXyz$LCgE%sR}VFJAMh@)z5YZ!!iz!to79*!5!WLKf`%(aSGdD45cu?=)16CG*$G2
z3C>c$m66a>{XgqzO${vqXL+6*`Z@x<F7yuiN2)S66lnhOjwi}y4A?1MAA4jd<L3r_
z)q~;dwbx#^Yg=31=WIG=x2|u9XT4&tzk0(~s|APKayesv^%vi?zxnGwx0}~DZP0C5
zQ*)9%1!*gNhN!yoU3HU2lRckMpyLSP4#tkRi~0`zfG(hm3Fgtji+~xpwj|mIKVhFG
z(M~<jy%>TyAcLbb(&;hpg<tA3`~#yHMkONO;_h^iaYGR@15O7xm7*ibPK{^Gi*Bdu
zw6apEiVjy@27u>~3XeWv2#0S|KYf6Y-~cf*mf);XDVN<4hRg!YkrmL_c=V0>hQT3|
z)sewS7Ytssg*JGKIz2~PQg|e3H>m@sidMU8oh}A2_V()eK4C<YiNdHY?0<@(PtS|i
z(mzjQ2#wV^aG$4w)pKpt1D~9TzCc6W_=SAY@2VY9t5(Y5<9F=pwcCQjs;&*~7jOUE
z-ud0{ESpI?EHH{<<R=QO)u<ckPS@71cp6d-fVDaetJja6KF~kcXFyMdf{+2^Ch)`~
zMMneYP0(uS7RD9!?rXAlS)(I-3z?YULoR;^FQ8q}Fn20Rr#<?rSkxTPR)p@N6{3wN
zM~7|<tSqmIPlykTpQAr>*ocpMx<y9B=LCh(0h*eZY7fRc(kV|Q`(-jEy+d^`qU&MI
z#sA}?m4Y>L!I@-8-!D>h(v9kl%jnaFb}G8r5nhpMkd%b*zi3o4QMsz$71dktc2XzC
zEN~BLWhlAi&Yf3n^U5`sdE;@xUU-C#5JP8Gd?=$nk!lhB8?tX)at5*yGArqG!Bs@p
zq39rvH8|kl`0uGy<5AZ|5|vBam>=+Qrv;*+o?S{bg!xAVoA1pgJX^p3CXnGk0qnr7
z-~aCS_TT^EpX}|oe`UwVr-D<!b99TYlk|BV<WZMf9Og5)ho6s6`FGL27ae+5AD`0^
z>eRQ#-^1@G*FNfaQTeEAe+uREY&)E3Zupv)dHFIF&FdvG;!LsN-aelD8$!+kSb`4`
zTinBMgszc@JD(USoTyO+*YzhDUr_V}B?lh69(5&#iZR1Ie=u%{Sd(+P+iFQjsM(;y
zrlF*?F<ZW}Zp)i2bn(8x14P0Ai&Q<!+gvDg!7%EA5sVvQ9B*b8EKa6qj}{#W7%-`@
zV8ehCa%H<0lIry~=^E!}UNE%X--UptiiQ%5b{{;l{YN`mAavcc@q6F<p8d^#`5XJ{
zSHC7Ed_@eqEGDZ(HTCOx_hz0Nyc|OMkC0D&kCiFN=HkG3mhnE?;{f1~fAS$ApPxL>
zKi%gZUx)$7(SED`@k532;J1JGL;w5>@71R>$)KPL0;oRq$J?VeU+Ox>!w~ItXsv!*
zI3R&XFJh6;IC|c8Z+v(!?FOE}je1<j(Pt1M$Dx1IdmLx<&pp(qiyy{^F+tG7i;3o=
zBw(C_@sKvF@d5-<W!^jL(lZFxY0@92wK&o(<kP>vkU1~&GB5M;hbey&hOfW{!rw2f
zOkd|b=yQIxSNC~$=Oyc#-Tw4r@G~(eorA%0Wp*{l81PumaUY7Li;FX>RMYm>TX$@2
z9dBUqyG8gpau#WaSa4%VAo5u<RKMrZFYIrCL5Qeq<es8r4x&Erz$R4_&*7bl0)PcO
z%6E8s)D;r!k;xuXYnDi7jSb}wkB%*`1^=J_@cZ`N?|j{E-@ay7Hr7-(N{B^!{k7Zn
zpZ?SL?JHk-U9wHePETsCOm#jk{_MR~SH=11slLP0k<-}QWl`@c+1lEwD+}1PoRogZ
z1gonB``&lnwC{fVt9D24ymo!lQ#n#E*YWh-+1<9iy=}4iOON!bmJ9aU8+Y6|x4vGr
zN;zk2W`BH2>K?=aWRC!Upik>u7D9MJZs07JDKuJbo{bZyT~6ch-?M-EP93@pry!4z
z@+0&$yl5%NV-5(lL4q2x4aN?PKUc4=SRs#M2*)tw%mpL8CGFqXSQfnL3n}R^IGHPj
z+nt8hPWEkY_n|#_@X&U*cfGF|Ij>#4=8^B+9?H798#RLG_o6GhR#(@oSgPn9X|+9K
zc!;A|MsWyE`M#`+QExK_FOLQrhDfHgWGS*3;IM?ujy#8w2qkGSE(RkQpU8Bym1qVG
zU+Hwgvbmy1wYU;Fin38~5k3^OrXZ><mcSc7C7dg(zd79_d0VnO`}*KOg@J?AhWGEi
zZ_P%-s+AR6-MDVW%BCf9YnINJB-@Sbz4v}=zxnkqJq1-H%APe@SKj)5x*C^}`a_B%
z0Y>dbz|~L*1}Iv=)u6;@ue`P^{LhB^-Z`<8lRZ~TW>Lh&E9!^(y?|1Xam8s1gU^NE
zL$n0zBzF?jxyiWa#w7MvLJ5jd3zQ&AudDGyzDF4x6xkR+fFmCIxv;MneIK)bnPh8Y
zzmvLe_>UZnqOIRKwSWAFf3TnZ^hcVDG0lgf%fdvV`Ti{$KacDXJ#b?Q3Q?UuIXTjt
zJr=BvG=KNi&ao@ZJ*}M(ILOywtwO}x<+!V7HK9m3r9Npwdt?rVchY%YUi5X~q$t_`
ztmpkIl|@rB%c_(8cQp<Xt+SR$UAZ9KYKm^{x(7XUfqCMd@eoC=Cx8R+!{CEa1$q^X
z9-2dAtqBJhTbMT(27uX;=m@F$TAu1sjq(0ss3FDP`Iz?Ueyvj$qZq1U`aM+tcSTp)
zt^jvqjNXYMQ?tPPj?*1VRMJLzbdu`O%N##3PAqzwMAnhu019fxsO6DotYO$^8iNB)
zfkbo-*n3WWCDH?>DbRMEGGNx&L}Nr$Wc6G0V%!#uZ|VF{upVe$j@4hrg|#ou5DvJZ
zMr(1w1lW_3Fc`xapN7?sckFQQo;5VaL}PTdo+b+m1_s~rEn9G<?)ijhEYS+#Tq;x6
z7^E~uN7gt!u-fUqosWdaf-iIoLo@rqF%N(-5hEDw*h`0$a2QCI#9N*iFBuNfEjHXJ
zO@t$kvCK)LoEfw2?YnNQN2v`S63vW}0mF*p1#3Qy5#ttUG%yb|2j%*L#*P$bM3vbk
zhL?dGn7Zwz@Ug3VE`);`bK&pBq^(wA4AcB1x=i$nxdfdJ`}t8nL>!|X-(#Ye?2~z4
zV;S>Eomq-RQnF>m12@DU;fvq>&e!Y*fA%f=v+sRd{P7jLac$M!eEo*=w42wj=$Q*^
zHyffaE!8=4!yXxKi0~axcqh?IqLHGvM4A9!*1>N0QUgOK8*rXTTHOyG=)Ch8{XjoM
zWC@W&JjYmbz^^bM5&=wf595FihG!BQt64g_T!D1C1$RW?oa<S5j|zSDXkxX2s4ip<
z_{Y|@YrY31Lyhqe8N?kSFr26@;UFDoSkI$-!RaBNlT4y}kx4Ro7g1nD>Ow<_8p6O;
zDC9)Hh!mo7K}Y=nCd?_|0CpjBu*VqtK+;c)X2Su|Ptd_LYqh#|dfKqly7;G#R;wdg
z4PJ@w2+rtdcs~be8TsVeNW8W0d@vm7l7z4>%bZ`3eh{N!T=RwfAc>rc+0OQk{r1<t
z5uZMGy;ZAOx1*yyz4JtP-ViL!R+rc9%GPx^sv!3rA0OyiTm6byTyqB!L?lrZ8n1bb
zV-tH~VfYGq6ZQe6AEz2`=7A{oIeYUV`(d<7JKZCC&Gjy#UEr0=DmuU`_D>Xb6Ri=B
zXnx>Ggp62SUiBzeQp8TwhO>UxB7$j{7eE_?(}({NPHFCR+6`|HJOdeHbzS(da>Wft
zPPK)TiA=%gD0_6Bik}`I?TeR-UPzvBM{dzJG8PbRmDJp-F0UBroiV~<xXbH%656Cb
zLO+Ng%NI1}m8#&F_0MZJZb_bCpR<C~a!9o(Q<zW5toI%KN9ObhX3bw5F`#RVBWZ_m
zc!xJ47vd-aenX?kN;T*OXER0xgBtxurp_0ZEtk(ZZ-aLxCBq<dFOWJ_&k}9_=+Q$@
zjmo}&|L~9hY(M(Z+ro>Q+L}_Eg*z}uJ<Hre4~HxQu7S_`qifF<KTS)f|M>pY!41!!
z`z*(k@`AR!WO=^Mr{D8sKb|iyy64jzbHmrX%*&Uhh{s`0$cU&HOJeR>38Wr!j5ijc
z_YxMJz=j0@2Lg2*A}&D%Atgj~xj9DlVa#EHfYJcgBBms!?2B#H8HhfM{E>~O#wr$w
zgN_)ugt*e$ij`N$jUN`_h+}RT(ladZ5vn{)TlKKm!}Fyp=I^2Vy4TwyVS?-22UBuG
zi`qxqQ5+$N&~e&C-w^y*EKxTC34(Y`jOnaDw1b^JYo6Au-|UKU#qIUiU$_5>y}Z8t
z9jlZpS|CU`mVn9v(nIoDNOGvHz>M}dko6fz>P(qJR1@tX@`+w@CH$FvhG)Vem499j
zA8IBaUZeiI(TOk9bN<Td*zg^q8}6V@`m28WcirQLFBTPQD^Tfbk$R-XGDjb;7%d#e
zf*%WyxP(vyW3}1^Y{n%-00ubHD(Vl45iMv~IO{u?i0a~5?#CMqL7R5Fz^_k~UY_Gz
zV}k(9Ki~qcIW8g;N{|gb2h;GPWpPUGcD&~>stK<o$cFpoWnSiGUOxBo#TmZ7^X)IE
zoD2zmL7DNf&!fC-U7zQ>DS_WURrK>kZ*Z<Bp7;DU&hrGH&g(BHu0+r0lWzRk*j&-?
zkRpy}Nm~ZLhSy>6K}m<Vc#qUK?Bk$?JiHO32I=zf>Lr?<)JyPAlxBETqtIgG_lS5X
zat~2mv?ipK39<)#7<i#@dh4yP*t%pclq)Dx*iUC|b=9tHZQ1Jjnx(Q?TdtICsQ2&g
zZ2R6$$T-MONRI3uMig7vt1D;MuW#C|n_KqQS6`Q`^op^MNvB<NkLY|hWncT+>-Nek
zH?1IflRZ1xHv^^j_M`h&Z?M0`z&er#@y@;T$}6_Dbwzz5Vk%{+M8w!IpL9{RM#r{y
zj@-KzIRLrC3-XM<hhNjb{6-$rcOR$c2J7*&(-Ou$-0jXi>L+Q8a@mAsGNexfu8f=R
z@n>~EGOETBMQ<XpXr&^l855G3P<jT%3+*JeADi;$vN<Cq6nmBMJO+$byJ@x46FWTE
zwS$9wYc<<8KAWh%q-|_$Sf$FIREzGNj=~KYGo3`47^1N}V}mP2lP))5*kZ3R<SWMC
z9gp-3_(GJ-Y3<OCj}G+ui7N)(U?!QHG-mNQM%Aoj=&+AWkm-Z0iG0U6u}2q%3KZ)^
zwmeZNpWCosv%7cSG4?4dmh!f!F+(|r!U!c5kq#)lFldnK4R3RlnT-Y!BTKfjx+Z*D
zb&u@@(t~L(U{t{1_3nGWx4ZZ58EHF9<rUl9yyl8YqC?Jw?--!Qf;TW9^hvo$O0E!f
zMWkF!bpT`fj}Z)yfBHkZrd+<Lwh=899Ojmm#u`)7YED$Q`lWRSQCit7>FcuI-$B1^
zwUZN{tIR9bI2799DV@|G(t&x3MK@vqBNXlw$=-PM3&t3z1qfG|w>uq;XHH{As=1E6
z{q|4oy?1}Bdf4kWqrRb_4NAE}q3jAmqB=X>o;4fIsnS~gQhR&dwkt?8nSw2^U9o7Y
ztTkTAB6^?CjYYNBBcar17FwV)j%oVR6@2Qo6Km8CJOYAD7|;q$3$;yhx7x=#LVxtU
z)=;rai!W(SFpYEzQE}??C5n6Iitf`pg+_ucYXhwVg`><^6tThRA?TcGTv{ki#}*R~
zEUK-*k4QI+h^+6xxlFd~1}gTBfeyLRKy)yyiPVpAPxYMmSR^oR39tPf>MsT@#u@_z
zMro~Lofff=BJlDQU8;{R*uq#S*wnnkNF|y#ozGRR0hz}{yD;C_7YMwejUlpyxu*50
z@bKc2C>+gO)`@O_)0*e3W1UXh=Ls-j4=fA|z8BBvOk{<KErIU>d>-l^wUZOTQFwXQ
zhh`kw&ck~?H<;VWSi++^u3o)q>l@cBE1U-AjGwRLH1EJS`b1>L(ZP-z+kp{yMZ__*
zAVlWTchZ#spOHrwGVb8pVw6ZW#*6l#WY&6akw~7Q8x&Y0dfI*A6Z7o!<iOX3(5ga#
zc^0BpcnQWF&mbZk4#qUS8$=W=y*M6$9~z}e&kfvKV5+|t9-TwI828{!qoMGA(AM`5
z9R&{2_aTvCju%?Px^i1{NO<aL6&GT9n`pYmmb81&XpDRCZq}aELxdH3kFP9O#KV_u
zO*Hwc_~i9#S8Q#4S$KfMN>q4=^Mjt#J9_=THATmJ;-9qD_p4Gbhz?=D$FTm_f&N6-
zD8K@GO-lp64iP@|oiSk_NazgfTyO;1MLm2!Il)QhSU3e924~Zhcm%Qm#)gvkJBBZJ
zI!L8`oWp!U2Y^$Hc&z#jj?ms#OZcuaKsSUT%J(QzTYUdO*RO<8!gxS5mPcp__xXg!
zGf$Wo&=U9)ju1HI5T%SkqEIYYsZ`W_!1ymYLikHNfHwt0Fb*de##UFWE(^f_kb6Q3
z-$k{Hw0(hY_#Rw6(OuEOc1v@sOH?%cfN>Tdl*|PUA^HZ#p)^ht>JRk*e`ppFL(G3<
z9a0mL77<=U+i=dnpo8NC(O%#O&oTd@gN!?RBJi7WVxMvhS+p6L<nv|8L^y~nIlsf8
z$Gl@7F$|Ub#z2Bj3TFsaFBp|IR#~+<Z^eq_#>}$lEJhz9Q*ma<JI}`vW^aGTBb0rg
zG|>U(CEwZ0uOoPP^rdi%x#5`vG@pqwCv(msdu1}m1phFq;2b>b(rGl&e;8ZYJCMj|
zqPQ@IGZ&$|(G<~mQ5$q?p6;>JCgN+FvQ-6RaGO+tf<KuG3hK*J0=@_xPKr(F-Fhv2
z7I^?U0pn7F2yq=4Y1soXksxxBG@4U~dhr$J6$Y<%Q@l^G=Di%yQO1pXX0i|2p1=`s
z9@uz>0=+Y$_w!C1A;1@K0M?dLUl(Hq!R1`_AGr>TC^cvlxUsXnXTSKxFYQ;q{<Yn|
z|46dQBiq?MR2>tiX~`rGG&o@(PcXmIKChX-$d|yIe-uFPG>R|&pLS1=r~b|E4WE4Z
z91j1OzG3#x=}Vse{Yj2L<?@H>3$tKu_?nk_`I3~0m}gv!FHcB1N|UGrb}@)EF-td&
zOea6W7T*y{5RkZrpy{FGVkS^dlm-Yev($CENsJCJQJ6R$dJ?j9&J|TUSVW&mFhRhG
zB%)SWUbeE1jF^q8b%y4`v)%&(MQKB-6+RJIAl5Kg>SBXa76vT-5fo_)ObRCEjL?O{
zkuL`M?scjy^vRV@=Mp}}0CmwKTBy}ePVL_B-?OC)yk0IOxTQSO>pS21uC1)BNU#X)
zAk8Dp$~_H%huT1Wd{^aOKVk%hej_ZA76Y$M@4#b5!&jaU$INJkf<<S%Z&c~$bsrmJ
z&X}a1n)Zdext|pWzx?G?#WmIwv!DFsL49g;Y$8FZE<tGLaMw;7r`9`@urByff^aA}
z2sUmo439Z;XS_*cNlOvnXRxL{yc`(wJk4-}8Fe58sO@evlrWF65!hqULO=jUyaG=m
z>U7UJ;UxD3VTbogXwavc<@K|-Lp09*6YK}4cF%x5FY_`l^YZzWFU;`uy)T2|>jgB_
zdG||{kG=aP>wCU@o@5>#=ASQ7X?W(@2LjL6IaLJDVjr)oS6A)!oh_?WSj3Zd4y6Qp
z8?i9PdmF{c(4$b;!zL(lgFz~j%emr%^g%>AyOWMMBvH=rVj`=kgZ*51Em21(0N4YA
z`%tDJYY{QQ{YlAtt7|K^y1L?uBUd=-UlQdMp4x@{l;4y32c?nvF&ViciIj8bGB8Fh
zMI@gsoLerRwr_pwP5a)TeM5b`W}8>mtX#^ey(q_tF1fJVx31Y+Z@%VU;4{Sp%H~e{
z)E?Y_*ZPC5CB(zopq}(LYpd(FELji_@yRr`A8o)u_TKvsZD;#P?~b_A%!LtkEXW*y
zeJY~_$;r?Am_9p9{*0IDfb*Q+HoIBX#U)damhctz>&Ep>wJ+g&koE?^3<bV!x|rzJ
zbE|4>Z{NJ4wihi1Y*1KFaP$dIJ1AE0x<>hdB9UhqD-1Olzc6yLpGB+La@h(=4%r69
zBOb@d)h^>MMl6-cxG^J<%(+4eTv^h+u_)<=h@K-&712sTK1FVK?|x*xF$%n0-$#sS
z2v_t9zu40wrg}+7$R0hel)}kP?RSrO;SdHal;bgs1}MqZ50~#QQCQK<uALk{vfuso
z*KR}rUx_?pj567x##nWuz)Iy^;T2UK9$le#=W->hRF{Q=C086GD^7+zJ2~06gZ*8-
zBkcK$!4sp;_3N+L)oXVwmLvrv#*vidaLJOF7!A~B;Z&#F5Z+*@xD-q}c6_{Vr>FZq
zE_i`^iYU?|#mM*&l;;<hTF+ecjp!@-KwZEPCq4{KC?AR5Lb0e7kFhyFYZVl*jzf}B
zJz`6Du;w9hjz}kD)sRk5@X)-W___uq@*;bD30L)QJl<33vc?0?@|g8H4Lds6^YJW|
z1w*wP1vROnk{T=0Of5xHl6|$t5qzKpg;K@6>OJjPCavdI1^<c=mNa=N?vUv<U><oz
z)FJy~#rz;zAQ&Ti!pB29(s&|km&;YH({i>LAuYbzjNA@Q(fme1dN~oz>DH{<>)3)|
zmCdma5qqVe@J3P2oDvWgNKYR2K++(q?HHIac%W<rUs)sd9fpZyL3M|;h2TEMeAWep
zl4zH%frFEC_F)NzA2$rBT_PI732021dvPMt64}S^2&yQ)FA2RdK%MF!%A)3xd%&3J
zs|(GQvtGmMwLNPJ78pk~lxnx2hVfQ#@@UO=Lo|&2vl5;eM9UTx_Sx(DJnIq-;?aW(
zYI86Y;pD))8EI}2UE-0CBO*VWZlq;x<T%gV!)T`WvL1rFEC@!>Ln8l9jt|_Skd$%-
z(TbGdoy#v<wpg`A;G_A1B6mguQ1*^AH;8&{HU&5J5gaFK4WkB;CBV=lwxF+<7=HWC
zKitr+!=oJ>ep4DMII3!`OW8N9=W~t}t5LxyWRWpF9}H_4{(%iV3TH+5r7NPr5e%_j
zC&ppu2A>#(I8Relx8^}CmKE&?5hp};p<Gs#uA~q1Js5uAGthyN;K`WUC3~%@zrq#9
z7d*vi4*kIp!#M^I-h_dJh(+jbOmm8MVQc_j-N3D8!r#Gv@Q!&)+J84rs>9>)#Z=!A
z*1R|qqcgzRIUEgC5Bd}g!}_lST>{E-bP4b#;DW)5l&&~;Vr*hxL++huyl~3ky+l+&
zpLez&x={&(G|_%U<iJnqH!=cyc@YhyRkPX$E@0G*={l(jal}D6PpUy^f7E$$r`vRz
zNFy--F~@Lr4f`Z@+^|HSpnbF#n84S-(Ua37^*=;XL6_jJM0FpZp7=T)V-VXTFuoz>
zC6k6Z!NELVv_d{>o13e)wpR6bVCW+y9rrQUD&>+p>r|I3o_&Xuzjm?UM!R~W;l@Dr
zllO6QIR>7Lz6yg~++mAz3DIcmk?H%0O@uS*Yp>IB$AOM`0Q<i&muVL%i`O?+ZGC-R
z?_HC;vn5$&Mg0NSBbwKe%|<<I={fp>lXgjSF(-PMN*C3Sl;FeolHRcAQMqH$LgYet
zF?-~tQw4YaNP4<L;kj_);*w{;JB^(ib~UE#56OJJU@yLXhYt~XS;+=5$p+x$qMm~W
zpgTDe-a$7mL^C^`nj7WmH~Wr-)S`m@G;ORqYUq9O6mlw2SY+9O{x7+V+Q31&*VGsy
zmoWaCGi}K)qM_hwxhxta*%cbMySwfCcVZwdR+iOf41n-FoF`_qMB@*Aga)}>6ZZaF
zSP-6Qeun)3;ZK1Ua69@y=-KfhdmQ2nQ<3~y5{9Eo8u=C`7|ehfMp(vTd3jav%~_*)
zYE8A5IhD<pEt)Cm-IDLRL`!z{9?=lJi}w;8TRS<B>`?RL;NZ~SeebUC;q~iZ|JL^Q
z4z1Q`TdRYUn)=}SLCJT*CEtfma|Spuzk|`5c@*?!p#&Njbb7NhoYPG5-+><Z+Cn(R
z`EcLkn?LrLJ^xWhxcRe{P|uV1%?sRdZupv)dHM1b(OpaqaaVfeC3lL!PRTLeeL-oU
z<+2z)#f?QQ_H~0V`bF3fV))H{q-G(_0X1OgaK<EI3C|azhFplEay^HK8J;ymEz(Kb
zv%Ilk%WG@CP(ub9V~7&nj7gY4ctcR(c?2MQB|Y5T6*IwG&b=%24&IHR<z7T;1BF+j
z$iWkf?+AVL1?I!P2JRW6&x^CMr(ZldII?3gH}?SQ4kSqA?7#e%zp*=a?nqFQkgIpX
zq}aGMC>O+}AoR|-rUYOA1rlGZATas6+;9UNVQM^ww-Ae<X+3&~x@Qank8;i-tIq0_
zyySiG@BHzveDZ$2aaBjS{mE78(tG&rMkf7Z>@i#owBXu3+PC)~-3w_B5s$^VQEJde
zlo0O8r}hW0blQyIuRb9psaAw|-T|%zoKf9^pQpXlv%Hu35H9II`;lPOVw`9X(St6G
z>mHBj^KnKIA=s%d_lyG%^?p|Z2`}hx$StmCBuJAw=t2uMa(GMLnwNQ*mwEZT%Abhg
z%c=b*I6l5W$3JcHCqKjUUu1EP`6-(E1d2~!3a_7ivH;)OTD5O{<2Ac|bIS^aY~W#f
z7rc^xD7sKS;I)n?^vUtD%S;(kRr$URSy!N=7{aR=nFhrMMglz1F-9ONjfP==3gj*#
z!iW?@NlaQJ6hfprLJ`5jf8)xQ*U6qHD5g-Bp>T-!-UUhLp|134H=Mtcx{%H1G18Pu
zIlFWFx*NEbm#cR3<~5J-T3IQp9`%TRiASd-C$V?fl`ETe?b?>=Sa!pgF1vz`y<-~n
zV}Dmd@?XA)vVPSSugGJr=pGHM**tYmYoZBIzU}QES*_mnF>`4KuYHGi_=BPaSj{5+
zpWra<dP=e2@PD@On{h)<BW>fx#)`f9#%=qG4pKAb^J%rcZ^y?CS8`&w;C;DF%)a*K
zEB3XwUa?|9^6O;aO4yJkLA4@xNxpMK2ufNNV2S*S?2_^bD~++*jAD-TSSX^AyYR|K
z0fXF$Oym&~ank#w^bB6w$h%}vz(9DW&ojMe$vxzG-((UZ?A^$!-)gdvY2mb1a~Z5!
zS+z{AAUs6r86rMVLWfjVD5Zo8y`C#;@c1q*vv1cGj~)PCXJ;tYv|fmjDWPfazxR8`
zL2!l}y4|kJf#4uxyR<}<embPbi>cr7tfjMA_h_$F8UKXvc<9fu+5h(TeS5U?$c?hs
zu3on{-grav>D%_otFJrGJW-rUzKv)tpy%TWjKx_uB$L{R=)}JIz`jh3SGiOcu5DPc
zwCeM(*{HjM)U`|+dxXcP?may{GQ9CwgJ4)-oGb#&8AFuo?BzsSCX|{PJ&GX{Z+=p;
z<CTsB9)>W+FnEdsGnBoK6a2a$Lgq}ka^~~`V|R!sL+KZ@d}hg`CXWvHHTQ@ZM1dW1
z0|e<yF@B|URlx#9?1tspQw><BgomTP8&MqQn&Yn6cjE%@MgfdrNI>)76lfR9RpCOC
zeScV|of#hYrOKMtei>a0hN#K7tFcFcFIimcyt978jt{prKU#Ws#^IlcXEYXRSNMU0
z9vz`M(P}l^00Az8V_^#OZj3ra$S&#qq7_TC$Rgo`M~|KgPQoS8vc(9-$|W1C4fpQd
z)p)V@8!4$mR4eNgPqU~oU>%Bq_=#a97{oBHbejim2<~@#YI9C>Y(ryOa(JQO?@X27
zZQ-KRb{#ksg!Nm9RzcBDWD|3kJz{WbU@Zd8Iy^kk7|<t-j(WGAht|`66r`j=Wj><N
zN8t`_<UJT=!5fTy)WdrD+~;yImV!^NT-ABrOC(EZ+cfG*-`&Yl|J+HFi7!}SBtr?z
zdYII<DA6&zqp*cnpy0;zvxvbes_(+-Zl~!=b@+*|bJY`%=9uBI<MY_@7egS1=9KVr
zF|Izu*>?zK|Dqc~!KI1ld#zq`Bao+a^Tj0XRNYKs6{1p@x4bXB2c|hSSj95J4QQs;
zjC${^f2^@tqjE2_bha-wQI|=K1hKrP8}y2O-1J)B?RBH;AR`Sa8^xP6-bCS27sg|H
z7}5ulMvZmvSl@N8;0Ufc%@n=Fm<;TACkBW@UOYFG@pW!X<BAT$+o*A2Ee!8w)$b98
zanfU|9iknq)4^Hh21YC2_et+1y+8Pk@t9`ksb91c!)Bw=^s!@ZP#*>YXbF+v;KxXK
zG!~(Doe7{6?$S=r0;9S;LQ%K`EubG5iJ-s88$_jth&A1>@u5FG%@^mBY9IL4AZlAM
zaf74AmwoUUGI%jYGI%$UL*O?6rB6IZp}+oq_GHp;aJgD3iRWIGtU-#qW#?rWs@Qvt
zJxqZwYko2n;T*DDt?Idq*8T%GPC#EUwql$kMKEn<Jcr^}g8>m!=z{3~lIt6QBe0@w
zBEparM(PK6<W9Tle^J-g*WrH^y)!r$<i(fZO&DF^0myFfLmW8@;tPpnMld6KDj1L8
zd*CqXO$BC$JvaotNfL=GdMB)Q9#8r}=rPVSZSkXiSN!V&!yns-2uCm$Yiy!XvMUL0
zf=Nj-VWv<OZs&DPJ;YcqSR=>rGH4d_1&0lcyXbA=q9vY6^b-7U={+4GIZ=rNy>IBU
z5RrmQM22bnnIq77p2bi~e=j7ncyy)aL@JflI8_~YdC!jcce8OUc;p;*qzZN;9=zvL
zIN%7WwhFJo4WiFNI$WI>9Ro(tJ;x`FA$S`XTydaT)H6iQ(kIekf;;eGaFaPj+F7FS
zXs<{5i7zqk&}Gs+udZ&op$+5WMDhZuDZv+>Wz6(?r%|L7eemFs{p2UVu(yBy8~e>~
z-j!^!t2&1MEOMw&#pf{5{K)j+7CJKMu+t@7Cz>zJCH^raK`y1=PU}Li_3m&!Tz~4|
z-}F4seBk<r9NhbnNBHyXGtd40xZ~;jXXWYZU-skq@}he_%^}R48@}dcUcO8P4-kwT
zaWT!nY$b|`39|6@$SpBt7H|6S14GA*P$Ce7h%c4teHg?rR3IoIRPhNRXEDI@2s{ve
zlsFv;K?wEfe8DQKD^{*bcu+eKyxf?iPmEnW8}j4^0}ILl1St>ymw<;c3`T=;00Rob
z9R@kxLmx=f0Ymc5r}dk2&OFqhh!|Fi`eEpuj-Bow*vbB(_1f&CpaqD888*57iy!`l
zt#4@Yr}hwWMd}FZM38`Q>8}$x3Lo^9jtdg%b5Ou=l{V8aUpNWQ2)^zm#Y4{QbWjP`
z^~I~CF648r@SQ)?uN>UM`~8cb<T{6+^)e&EUY`oLk%AK8t}TIO@A%Mm4tA_N=psu!
zHa7BFS03rPz^YkPhW3WEM;IRQW?|tF3|}bINIfG2@kI^?(RX~hAVc4Pn?Ftcf)T={
zaKjgN5;PE)d9NG9G&a5;fnJ9)$qjOjTeJuRD~g?vW`PB}e!38i0%>06WnSjxb0~ij
zhA&Sm70C2Q9QyOj=L?QcQJg<~%4a{Ucy&`g^GV9b-3iZme2?n=;Qie3$&c`F_AS(g
z$2fJ9-tN||4ZC@LO>zX$F*wbTE(`_XG!={PM=_No8V1D(%0=-zWFTaW=#u!VxUkDp
zI`5vv(S(&ME0!;>TQZr4kxLFca~_Jb%tXIY{NUwXt*%%sTku6Sd^ARMPpsr}K%;T2
zMSjbbGx1DSEg{t#?;{eSr9O<Eujg_ZTV1Kxl`E@u_3FB9ZLLWrdc~Deq)1Fp_Xk5R
zz!Mo>jO-F3#snAr;CDp#a1FT$Pu*6#ZZY;%DX&>BUsXHO-e>m`ANF0Lv9tHk_x(!5
z5|&A2)lZanm)7mFFAXpWif#bOXu4ui$E9myP!y}*YFBVT;$UckB9U?P`@+M14+@G4
z$(Q*;#@=}Swq3urZpEC&Wf28+TkVmYw2MI`M7yJ8S*c{~tv7Dj))i7mUKpu|P=q7n
zkWNj)is1l+p;3e7GBolp%041TGC8s>kj4tUp*<)9&uuU~^L`TD=LS8sjW(hj#Te#$
zc$|;yw0>aQ+YjvY^hEE}DB|5OJR)V)*$|lv<qc`Qk=rybu8g_VeK}iQyCNJ)xq^l<
zVr<w?EF$?IM>v#j7~ohNkc}Y{N1>M0m=Y<V2Sjg3$K(dnN#E|h|9gl3wQJYi`x-eL
zxsP`<7VJ5*gb~0MTT`QzE1gjs>$m#J{m7ueinL&?309Wz>=({Y6_YWHXg)^}zWrtu
z6scM#u%FgqS|aes6?tgpm_sO<i-oeStZr(pvEfD|j2pF+BUg4~tZ)NYuBvxbgf|nP
zBPdguqe1a681g!hM}S4)MZ_8Kh{YE)UqX5^lwCwWA^)?tQY^LXZF1v_>hOLW25k%;
zr=Cj7WnIbAC>K%2^ODHo$o6+1*~5GH!dh*Sy^qJ9j<Q%LH9Qdkq@~2Dto7SMMB^|B
zsku(}JPjTzFqCj87D?e~mwn-4Lv0j}O0MP^l%AKK+L4ZdgY4G@ZNjNAMC#chV@A6L
z(lJW$k;Z`t(Y$rrC-&YuZ(FT?D%hfgR=afnqQ)7xIZpAZK90_415Rst?}g^kh_tO;
zo$CqT)JM?}16CfLrn!6Gvh7E|vz<qGZPDqfu3cgri1_&9MUQbW&pt7tkI)vB_88ch
zH$G+;MC<6&VT<8wta=C5Zy&3!u@y_pR;g}UGF!E!L`lOW9MifIV=?ao*O-4e7qE}g
zg`Oe(Dh7&RY#O;?4?JOQ;!c->B~A*Z!kSe?tAQzzhoqzBK7OBzZWEORj(Mst;S&n<
zLSfkw!ogS+M}XiUh%%OOCqmDJbFoB7&F3k!&Pi?Au%n{`d-uI}?CqbwZSTJOo~PU^
zY940wTzKC|_z}{?dPFTL9@QrdB^i(A0#8Vb$ed<u8N+kmJ7*|-#u))R!PvR+%kdO=
zLYqRQC7&3%F`RK7I+)P+bhfB@Q1-e>E$J!VBEi@{6iz_1qH)Gu&pEvdktab3&e#D`
zz1(qBMfiGo8H`b!hqgm|Z~%aAVubKLt};dS3!`mD@4#r=5e_mp1@9>RWj<=Yc2BKd
zJ2dv^gO}q>i{62hl|<2A2sgnkjDRzK9M$;5)Hjsh>~}@<6!<n<2lE{#glehcMjQ5>
zs-4ztXM0z8)Y9|fe<h+fLz-HCtD%lVtdoHak*7F{09PC)h;TeUI`npUYBu2?^m%1<
z!)fC{d<G*FG%AS^TH}gb0S>tVRk(??0uKS^%uCnz=-p^3fG51j`Im4GT=jl2-hwf7
z9GXnOfH{1J6s_n=Nc9JwZ8g{<jP$<ZZJHyx&7C+f*5wPq(StQQ`^KqaNVK<!D7DSa
zO}la9nq-YFE0-&3S1@1@<q8a$2k3tqt&rZ1_R>~pIYwxlKQLzH)fQ+(x6`v`vne?C
zt<%Msrep0^*ZTU6<1=j|!wqd9;;K}sx?_hMXjHGKqt;l%FQF4evO0|x9D;5~^I7m?
z9%7J&UY3gMmM^Ue{w2*d3_u#I(NOIHN5OsYR#<icF0)@CQSxQ+*NoFNa0jOdeRUaw
zeZme8ckJYp=q7N1Z7B4-=KY{2I53AbJ_#bJ-N-L`HR@Tfb)t8Q*JraD%ar4n%S`Gl
zJfHEyA)K^@qz7hi*$cfJhxC)99Z&JcutWbBJz^Mmh&PDd7V;SOa?V$scL`^SxC4fy
z31t2e@lEtFIGapG9Y4WC=8}5p`#GLZIM`cqB>ZB|>xG*0(1^?Pp^t&nAqNK0@WRpv
z4j4p+3AY3T=r2ZnoCq)k<81Hn)jwn{iA|~_Vm;BEnPWV7RV<cWz8Uu0dRFsZuTDu`
zPG^{tIJ+P(r|m*Gb@$%3{pvS&#aG+H70n;vFz;7`RJrD?+Kq0JcJgle5oik~$fn?$
z8?|&i(Q$^~{)`&{c|PdKpdZ26V0N8X@!a&h-#feZ>^-xC69LO!e1!7k{%4-$$_viF
z==f7Ff2jTlrRIjOd6}0lQz6YC3~B-6Q7$b8Cjm^v8*em(BJp%LAYe>FI0%XYgiHiG
z{TAnh;fhi?6Lk+62)uqf1JFN&54=CP?%obMd|@x4ocanC358X1WlJDR@P082f0p}U
z1fec(uLNQ|n%%nyM(+k13@>=WAtW!j&_G*x2Emn2+Q5AnO}waHHn5GWj|m;;gOQyc
z9NPZ&p4E<P;;-kbJK_6H{O|`qbWhWQ78ya2B4(v`V<bdKaT2LN^pyg$okb|oXSiKp
zcHA!^N^K-^&KG1bKc3~Bu2Y-Vbk7i8zRrFlm{LD!COs9TpXHqX`nz-$=1FdSeLB>*
ziGifh4!=L-M<#X8aieAjr$^S%!if!Efl=7-T!J0&r`;~GoeQoQsDc-cf2u#8@<y=b
zJ&r39u4iKgey~@VD|#fDaFZ)x^bU-ez>{{n!bxKb+%U*d2RO$X9DE3A1n5mz#Pd)v
zmg!;rN4bU9R4UEl5RYPVX_JQ<WyrkD%e>6XXDVNy;p;n+lcCoyc{u&&`xjf_N7FB#
zxp-Zlr5<OvA5majA9)Wv=4qI|96aIW$VMoIQ^}ZZte31>Nr!Z2OUP8nC}yP8V)2bE
zfKn&yHxnXy;EQhH&@&-YL%dZL0~?emMCZ6NE1nT=B$6jYN3q~1DggdY`eGDuq?{T`
z&T7;fuINA!Mrt)xEneMK8%X<>a^uBJsl9M{X6L#Vi?FG>>KAW!Z`{kVCB#?Rn<Jmk
zc~k-lI~MDUizv!zf7siAhzsO9ylq{sSyCIvXGXdsSG)ix{bNtpT)rgv5KmorI`ZMf
z+ReHfBTi3`y&qS%uDQ_yWg5y^_A)`y1+JjTb7kJbx!^2WK(ONd9-WRn<njtB&)7r5
zTmgmB6FC*R14CsnR^s7}+(F8qsJ-^;P1{;umTbv8P*C^m@bJVBlq|sR(nhwiTC`g?
zHzdEZ7s|lcupSS1`khe!z#}5)$Z#-IHsS4#0R!hEG76v^<hfp_?d?KtqK^j$N49@>
z;0kxnM>XE0??J{%Bx0&>?CH*kY$9R{MIl)eE+r44R6~|S<|Lw|*=)GH2OhHj&&uki
z6^j)gGjI{57Y<{{?%<>wVUeqJzo+5yb2Fn=kX`X4Ux);Q>EhDZ+L9&N3us++vquSv
zOcbgpn~=xiaUy_7O^ytQ5rX-}nnLArWGy{Ulnsh#WJC6H&8ZHQ_i5oSawsw>GQRHv
z!}tj=?Q*EE1HnylKiVs`k>A5!!{LU~vQl2LbgrOw(tgr=X&}Jq_>yG%tYBC2Xeo>6
zp2duCK<kPnB5qK2qex!#eqfYc(EIFybZp4L^h1}`9u$(uq~HpvrI-ug_hN+TCGJz(
z{F6g4To4|dwcWS@?QpqXU%*#VBKF!1+u45LW6XY5q!nbIWpXHiN%_aVO{|wfeReT4
z_O|Lumao<HDSRS@A&PP4NUM2b_1b|)Qi2muJ9k>s7>DRd#)5vaw^U#E*PW(dK!9$A
zc4=NM0B7|=V~OE%Z|7b6(ZBwK-rKe6>YC<tR<IR}&ovh{>4@@!RuLhF@}E?Q@pMLX
zrf-KwdqI&u7^rVrOZdB{{ZZ-n4y}H&<LkR}sp7_B@Q5fIl+)R4!Ky1)gl{Xty@=L^
zea)|-uQ5>QlNt^OOV(^r(V2k1qNfwl#nWw#Rm&BeL?i*LkoF4#srjw(B!ZIl4C@`7
z6+Fv{;DizvxahRQ#trN`Pb9_h(XlJ{iNGlqmMx<rCOSPjgXWS>^U~=uw21nN)*vdA
zd57^Xl_?0U;0=AnVIVR!Ff-R^f54*%(PL_3zt{3TyYAh;XFvbNFYKTH<zMY*Z~ww}
zclK<V6r%aOaJc6*27?LFFw7wg^8G<b{fU7O_%lYd17kMtVvZAqcR{;t;xHxR!VS63
zGeTQ1%A!2y9b{~A+AF$wF(UFxYjHhii{J|pPAGgYgkvFnov$N_Vij(9`c2^&^#!9B
z<H0-hE{_x;dP4JX(6(-`>FY#{DDYJ174<+P)543S8@*Jg&{mDHp=fZsCEOqNeNMQu
zuiiBtb?m|Y_dF6OmlZ!G+EwjGkHDPsbb*@7&}CmA3szz8q2Q=H(HLOtEr`BAH-~CV
zPjkWXPHkdODflIOBsCjtd;4d<u)Y0#E0qhP!>bN|j1ttzJcNeoi^d7V!O-amQ7h0!
z_$upT-UH4+_lZDa?{Rovqh5EX3?jd1pQix6KtaE#WNcZ(LvDeEf8m35I{FNZc#=;v
zpW*rFO^`)!wBaBEFhsDzOWCtB<&F^0L*55`Hm__5&xjt@v%-Chvg|FE*E#A85nBbz
z6)?iUyDlHkQQE{DxPJYbz541MTW6n9!Mi7(L^K;TnEqgJV^2np6cvAEE};*?IfV3v
zqzc9uL0Z^+E~mDV7JA@@GIT~{3hH+H)@*e{<Y`~HCoo`W1cy4hCeUfF;pn0HE%?$;
zQZPe%JStK+3Jic5V^2gU&PF^>JCN5fB4fl|Si*r^yh!5+KWEONGs2L_UXI9P{Pyur
zAe-@w_>5>$y<US#`M4lQ;9w-lPWW5*%mB@L%T7c$8{(PJDR($hUG7*j=-TnYp4Cqe
zbxyNKxOFbx0?fyw5N!uN#MufQAbJ;^O*t-=C10d8myn%sM34+1ynr*{Ou{@ujwMCn
zg41<yDCl*-C6^g>xG_rf&5gIBPfmZMacg&aZgh+Zeq3v}>+V$Kb~E+E$3ytW*a^&f
zDZD^u_36A3WT7#77uVx@KTZu8|KR^bm;2tM>Ns;7!!COhGIki}ptsQF<z*bMir&sf
z<5c&Vu@@bY<QVCh3yd#1$_IBJ*pGhnGk1DY<D3@(Yc)~u2n-+?vS=`;;FG$Q267MQ
zbj%D@+{d`WuV|C&0_kG-<8b8bZ2iIC@Xastclz(iwb?QIJ1bAlz4!>VJoz>|LW^In
z%=++o9seE*oNI3QnwNR`k`)%?3$aDZYT=j8lYRt6hX}TWJVJ)qkQIgwQwpAH7`M({
zphL(&KyrdFCc}Lg(qMuxV(#Udr#aE5lTr~hqNYfbqKme|hL>!jDuKp5IK(vYe#0Y{
zeaKwl!?SwE4J49GW<)VCGL%91DdX+MUKs-yzI@|Ok#O4YiX`33Qz#rfwF1mVeeBem
zw*PS3b{{;lmV`Jq5U2lZ>l^m1Z++Xo^X+eI@v!0{%%nYG!AGAsydC`a?uYL%Wdvl7
z(AIE2Oq&vjdSG-YzIaBr`fsmm`uy}}{nz*DZ)f`a4nL-UroUM<`}Oei1B_!@cTlWo
z(IEk*D_A|+-?4*}LpzgjhOrE<RTwg!lnBtX0_+hq-J3@JK}kiszzGC5nn*ebTH{7W
z;RU#G=}M^u)rkQM1xfIFpGAyu9?w$7j5dOElrsrUEMk`~gOG{R1J7aiEONL_4TaOX
z9IpDuxaSKJ%5)!!w$Wg2{F;|}nU~L0zA(ercfXC{>r)h8Bu#n2$CeM?7tVar=btJs
z@d&*9<zGR+I9^=5x+yRDxT*t$X7xRPh_5{UuA(|$RNsvKKE4L?pTZutgc6&`g+e}V
zn;RG+Fo;YYagZM*cc92&(T}1OZ{m=;Di|qH8u@qeJg%YOg6GYY>Wk6ffsMoy3@uF^
z4OfQE(()jLpu9tAiJ}&TG)gK`HxYS7Dk+qFlM(5ONOOlGNza~lt<yfXLBH*?5xkjx
zk+P=OZ|R*2miJU(+%Ne_a@S}q+3<4WQ5|@`d&D+EkbvR8)m~R5>RA-7q}2ij81V4A
zB@zuDjv^Q(1ct9xyCHdG=!%+hY1x*S*Uk4j%aGm;<(!@qOxW0btY?rX6Y)i>Rx*Nl
z)k=kwWRgYIhj;!+@>|}nZY^7>lu?~1dpnXFfi+5piQp2kN+oNBA{*BaJyK}7QqX&g
z>OH9+@gyeQ)ljkx`;R0owsdaAT*8fB8yoEXbZ$L8hj%$fF5o~vF<SN79fw81_xmX8
z{Gwnw-Oo+(5C*K{<2`%y;I8fO?_0aww0fNg4h$4$R<4jX@`|l(T(kA{Esx|wsYyXD
zLup#AuBv}{$Rl@~%f%kGFPv<5>K+XO{-EqB6ql`BUb8HQAl<<D9v$ubK3GJvVGP9>
zu&~5_Tv?6LrC^DnQL;EVp}wL74^t&5rr7T$l}HHI8&;@Zv3R=Rj%*i~!iRIoX~=QH
zH8+F`ciD>vft1J_j8x-M+YK_{-lgCoM9{c~C>os5=nu*>j}~mze2=)%S<BAHO|`#n
z{cgjJQZ6G37par9lcXrj=Bu6pD#89)8i!WvRL`FX?jhxrd&DP6OC|Xf;{s7=`W}mF
zEh0JhLT$%TR;wL*R0v+|z7G=cKpYK!Tn0ty$i7d^M^aJ}Q5j@g=nF@n75=IIla7jr
z8um`XI2_fzj2VX9dZT9hd%Nn(nU#upE0)W)b?utQdfoe_o(isHJz(Fbjx|qr?et_<
zIMWl%P-3pw^2$}=&!#Jm`@&;SHzu4x&W~yive(*TB%}5)N7XXbb1@zmdtd#_zx=)M
zanDv(mz{RGjLn`di%G#4<ACP#xUG43tUmN5cbBYGUI_}~SkmXUM@)GA7@b5nA{ehj
z<kFI8kpL^1pShS02K^=0fUKLKA!9o|ek9t{Rv(r%juoxLv_?(kwfIvVOIkC=+4HFo
z>Wna-gRvdMM@)Fk8U%nkO?8DO1{+ec_3dEyJ*%DW8L1&rn1YuW{eVGFbB5Gij0r|V
zPbHTt`dXHC8;+AimQc<$1|0fk_2VOZaQ~j@R!E(Oa(h9rVhq`r3u6XG=JqtQh)5Zt
zyub)>3|#S4j-K)oJR}m3^<zT6c^~tQ!um^X)mRONUEAK?w)^*YY;W(#I$aEXQKvod
z8s7Eb;d>r6f<e}!ZG;2RC+PHSt>|e^8CT&d^pR`eY$OIP!w7^iov0GV5BdXt;2!RU
zj*B1pUO#p5q?*3h?Of})nCgNq`8?76;5T!eC|C?u7>J=&ydS>e#ThO4W9;G`)AtC*
zmlLfGiImekq%NE|;c=|1J>8$?#!?K!2QVZGvTtq8+uG@&M^TgJG?9T0$J_u~KiRWi
z{QM`P>+p{a=aak*e#%-lgW*W+CxUKCW8>*Wwa#R14(&!i!QMgfOi}Plsn29RAq{0z
zu!<50;|8Rh@PV{=>}S5Vwqk3VR~T66Ge$T!Qe|>(EGKG)_mg%I<AWRB)D}0U7AuzZ
z)SeiqNS7@7=*s`D^B#|QB2}{9NnbD&xg!JjGsar4!+()40v~7nO@w2J@WR;v++YvJ
zq}qcK0A4&AoohZ!ydCL`=9XlNuHFfZFs^wfAH5%gA;u%p4`Kv)lCqby&I#c<(Xnxl
z5Ci{#6)8ZkZf%Nx5b20PX<*Gp!}r^AXIh?7KXB|IGMCiI710FJLNo5cK_Tn|Cy+RQ
zK^8*xA_5mW1noyw6W0-+lU$_nL>H1#U(h8Sv(3ffk?Sllu%WR+cZ4yAXf=%X-i~Qx
z713T8y?`TQ)or(|-=qI$qU%I4iXN$-zMrM=8$JrnbVrgzQR9)e#S6`0bSlV2;&0$L
z5w!K%iN-GNPA<^TctYcujM@IquKnXb{G<K&fBe{f`s1J2_WcLqffz#3>)^03aHHPQ
z;jYgS#wv!fP;>|$2Tv6$3BNL;r=qvaH}N-&_|Uf~ssGf!WIW~$RNbCnq*ftkv9H%~
zz`kij55uP<8|oZU^0Xb944E+K41=>ykO@gunH66ss_#jE7ZH5SAKHoGo_S14M;sv1
z%q=&(!;?e<Ft{-xHO}ar;BTa1WH^_g8|nix4*Zlo)e7SE%op@WMB}+3Qg8q_7$>6A
za7b8U??)W1T;3cAM=?5!PU)Q^wWBLJ?BwX!e*Dw7?RW3I7vw*W^u=%wCg5Nt`mFH>
zjy(5F1Q-9DzN+ovZXL+KIJGh7L{~ZJ{PcX#4RSn@7D64m&Vl?i{mns{efJ}r(X*k>
zaNfT@bx$}qJMZuDYy1o=B+Bfb*};Y9&&`fMM0wGBK6p&~^b*G>DKC1C<KxQn_cNpC
zhOc><moHlpbJEhoQhB_j*bo(u%phoa7_<Zm7sy4Z5sF}L{tv-OzaJB8{X@_Wiv)z&
zs04su#9-rG-85a$pa3HDek{g%X?Y2e$utU?u(0Ji!U=*o-g*=SONc#8FBp?pM8jn8
zA_?t+q4l)*Vu6kqBZ3ve^GHG!eR0o7Jx9J?+NjRiutPX_$J+Ipo$en>01JvE1cNuf
z@|Jz~yWg`rw_kA~k2X7_a>l0q&X^m-SO1({sNE>D5SH9~S^ubvd+_QAu$n&4Nfm?&
zp646y<ySZ;I^mb}#}@+n<hwroR=#nH^9WV?<hIbB@LtOFNq;Avi;T-qaH)6OwzI!$
zhqWUcX{=%Vj6pCag@u?$dFdL@%@%xkr6Dj<=nDcM3rv)2cs21p1S16hkmFzDi4kmy
zM?oPI+UUZ)4uojN4i7#QG^8PL!AWCGI!<5@j_^EXqWZ>Kz<HV{y)fWT=p$_cFK7!3
zqKh#tnwNQ*mwEYA<%^5>`m~0xXUg;R4ob_pkD`OR_N60}&b6k;hd1u!n*DU0Pal-0
zzrVPIJDj@)eysz!!etwj^@|s_zFM?bZeMkSNIXI6vq@mzdd3xddJtt6MhjBXxzb5I
z5#=C?B8(a+h~R^+xYbKs9@6(kEzGIMBdEG9->4gf7yQv%re}~r@Tx}X=6u%qdEnP9
z^xb=3zu}R{Q1IW$sAuiQk?rsASf|}`UXBt11qSK8@Z?A7mcff&{2#?sqtyuS*FE%U
zCSM>MxSWQJsD98lHEjwjj2Wa>LIyndv`9pj;7v~S5j>m}IT%<-m6Xqy)kc)<>?x7Z
zvk@cB7;@;*(V?y_2;Nm&->BHl8yoi8tJm$;jVr#dSVnSXxtO!tx31_KN;SOC**7-k
z=^Ag|x^CC6uG^hkTXyG_>$bVEV!3S6QmTJ#wPKYD9>GKPv2N_a#(US-s<yFK@f3*V
zV%mzuj4do-9PHVUeU%0w^%)9D3XW93f@rLb%`Gc<e->O>&8GS8;3YYQs67pik3GuJ
zbSh(+Y}T${zhPhd+Sly%9rjY&uySQZ@@PhN63K$mh?L&QMJOvnR1ylvsNfU>S2Pwx
zIItI3I-T=KtW*kb^Smp>H27|4tJRNPVdsj~QZNz)V+Hk&ZO|j#*VGB^VqoWIU0Xn5
zw}1gG;3(ek@oYu#ASEKv9l>zq3RqII0S~neMW@S}KHidrFG>H{*Q(;eBMmX?f-cCA
zj<$Nx<!{y$S~rl^4FlQPur1u&vy<a}$8lhXfrlxEF&M>Pw!p^qM9UCOk@5FFdiXm#
z+P`l^p0EbNYahih%6p8lVec?frsK#(YOiU8-G$^+U`dpU`i9~e2fC2H5d}7KE&B_H
zRO`@(5a}ry*bRNE7K5Q5Ktr-Gh6};g>pdf4Kr(Y8>54y6sbSn0pY?=i1-p9vrrp$f
zq*z?m`<HwTMQW<ac<8hM&-sh<U}zvMSGm0I(MC(rsw-*}@dZmK6PC>|9-5bN6u2pk
z7fMFx2kST@*M#$pV>>+DbGTQ^!Uwf!Np+BxGNd7l*sy<Uy>`vwq9Yh%SJyYKQeCr5
zPVbB7e0=GDh)4_cDSU=OWv(Zaardy#<(AcN;f=1HyD{7BOk;fa{olFa3ne%EhlTyW
zFr?AuS;P_bqP`Pto1my&5H7OjTyo<TX*p4RvJUa6MLh%Fwrbn<%U}M)qn?;kI1#k!
zN7gtc{bh*G@^Mss7_ivChIL#>dB-|4?lig85iW`z3;3d;4OiGRcQKaeadnk6b?li#
zBo2nXvC~$X4NW7<3Ql@Nrx3A$a{zRcbqsij@dJ=yjGgt<G)ANcQ*(U1o7DJV=p*IY
zl`C8J+Uu{W%_%!NKD2{F_O59N#wpRobvKYR&r#%4ACW}+2m8XKb5EBB-G+Y9F0PXz
zv9B=*nN&z^CYqoze`18Au*M#$t$p>aqrR|T8POqO-OHqeb_C-Wcy{5+WegS=W^l;R
z9l9d^rTN3ecU^$SF!Y(iIAaVFzV|!&d7*v~c^Bs?y+<@pIK32w_5iCXeq3TGYiX`7
zTA{om8dX&v!2@W`$a>8ad-vVnd6ZC22S9~yx}&G&7i&vsUq?7WG>RJ|)TC~=>8S-V
z-Y~yN!G<FKQqR#p3^aMoTj%+zk1;J2OSZPL?$IU8Q|Ogvy3ssAZ-U{A`k3F%X3c3c
zdKQd6=w{Nw!Av^ub-`!Y8;-R#1}`Fw(S>YoTycKG{30D7xQZr*d5D}4)4PZQn(+-!
zk*far_5i{?WE0<FK=ape3ZscWp})RAx9AS=A<BkGVt6CCL3CP6vIB4n4iajB4s-y)
z5hq0QKsT8mxhzIC+O(kW;-wghMGr8_^C)d%y-zAmeSIuffKM0_^ZBB1x#SU9)oR7z
zjWGmR5Fy82Ui6QxFv8y9A)<}wGjI#J2YI59&p93MwA$YOwmxg8r*^7?dFjSapU+|b
z;dsQn9*=P_A_`68&o$vYV|H*L8A@}CzS4K19V0}WY99AQW6*=dW1>&$N2hV(W8CkF
z9v~-bP7gGH=&Qp=?e{sLV^KVRZ~Kw`(?9-`{q#pavY)^G3w!_lyDkScRga`a^;7Ti
zxrP&P!g*JUxsEY7LR2U7MYLaI8i{9pPGImP-8MMvR$3j<J&y#_;W`|hZW1wEKXDxK
z)ZF5M$c3{EOS5e$FvOvk2~mt$H`HRNER<F)C;2BU{Eq1P1@-6nXwT}kV~>=@=?4Q_
zF|WBHe1RW_=uY+}#K0^#=p7!JFL~7?`!(k-_5PjRdrs@)q>B_kW3RWpy&WHO_OM0H
z?g}q3UbF97JSzAL#yh*)_Rs(PPlDsS#(t(a{;(hD-tMl>y>EB#-4`DepX(EGjU2M*
z4)Dl#%q7MhC#P8^DA!jBSd2a7%4wL(<g6O}gkPH-&=GWr{^{rZi&H%P3g`Il_XfJ8
zio!LWr}*{R!F}Ar;lKT3dj09ALh)bv6Rzu+UY`AC+#gr>%y(}@IQW>Ir*R)sLO-88
z^GS~n*3C)3@KS~M|DnqBbqizXhOc><moH<PNI-znCbb|UJqZdE=m|nS!hstSV5XYX
z?1u$mg;*n~L5ydH9FO?XbM8Sb8ile31qI5S89{I_CF+L_=oI1Wz7Zt?5JH&ha+A?i
z$l*B`n00Ut#SU%rbGqh?ThAjTg@p=CcxFJ;t!fa00A5V)g{$8fOxpE^v7cABDM5`5
zwcCWEyLc!DU9|6i|A+RCuYbd;E6ZNL8~nu7y{|Af{R>Qpeucg{gL~@mv)rQtRu;bM
z`|Nfu`=99#H(pJ@>zsF$KYotqR4)&6-s|H0lgB8$DSS?^PrvKQsqh$#SslE58>hB+
zv~P6@V|Z;NsHm0t&I@>+?--^8Q^6B&Iahq>I)WcJA)wO+_Vt<wUpRPgLc$_KCgaNW
zuyDg5r8*=ucsq2jNAdYs2Ll+29LA6~yQdrN=DhA_Oc@980>O$!2X%mhw4HGQ=eX`h
zT8%qiK6p+0zGL$;FY_`lFJJy-3}4X8XA4As`t=Jc;dL)~>2s%M_sw|L2MgzC*JeF?
zVI?n7p04xx@)6hkJyV{p?=H`T|I>3Q<VK!)<&`^E?Uh%qSuTh272X(lQt<MGhoKB-
zNrdtwI0v~?hj=AQ1{C<nNXRp8+z>zH;5X@cFl@n4iMB$4isB3T2IVv|6xZQ}7}Q)2
znzq{wEGWw*2ykyIlk@X?yN_)95r%-Um=9jBX5C(Qsw@az*YH%L`cG=dZt&{jDLWjf
z4Ww9WS+h~Mj?PWQn|T-M3xmUvTA+6#m*H_vJwYja9_rS;YzSYfta|hX2Ea&!2#bOd
z<%Nua5rouEDAX|Oc|`x>xvi{}?8c1^TU#w!u@pAOUtL+TOgg1DgtUbl8*5gnvd4;^
zS6fij-@0?tu3p)&_0@`%%cMFZvMA~bxYgC*k)KW{-H3!@2haIRxoo9k(G?#kcTk+5
z?BEkQggr};Eje6XqOX!2ISNJg`zd?BiA3x5nqGJ48?tG{@PNN^^_pF~cH1^K|8MsG
z^SjX`N%sVe0*U}g5Y%MoRHduByZV;yd;9L3**P<3_OILzvmfTn*PT5(dqs=tE}1ei
zBSSE#H31NU1W1DW{2m7)qB0{@b#``EHJ$MWfLHkIZf<U7Zpv@R(#ndy>-XJzF+E*X
ze<j~Yo<WJ;Zj-i*)J$p5xTa*dqr<(}M(L(8utG-f2Onso4IlJhbs*=W*a9y|eRf>O
zv;M@_XEX*S!5xKBjG!E*uny>XOxTBn#X_&svXOLs{fVPvfJs)*CxtT#abukNm-a)l
zlblY`eq=wQ3C1$2%X@MWfk0}m6d{024921i#$%j~{)w~zms1*btrH>>ozt!rh8VW0
zGdE+hGOvD(SV6tFvlR{Xi%10U8G{#yc-8O3qYvM;VQgjPrtqt(cO@J>?cy<y(M`0b
zPnHX^SEO`Y$dce4h6@C9l(pa^it91;Wf+Y^>Zd$L1J!{-ll_LALQHcyrSJn?&h#I~
z4EBCwuNm+l$-fwE*tkDNr_|=M+Lwv`d0XFw$ICZ`--{Ycy{j)dSTqM?rK5K=*P$C2
z=os&mE)4~;)-CEkDj4WXUO0`RfDSwQQsV}kQwlu+0Yj%?&RTC~W<HjeZbi8?ZG#i(
z3e(u8@$H`09XX0Jp1sUU<!a>1MADGj5d#YAIpG<7C4!4`C3<yU*Bm(#$s{sKEz1=3
zPK;hjF-}S^qPIwQy1M$nMh+r965UK|5E~P$R-#_^d+G-Jn?Va%4^WsNM1eqS=`(9V
z940Z!dMz^6jSn9@u<{r>hhd7n14TInN7B5N^_i6Z;Oj7&68ede1%n-U$$r(WAFr<1
zqif20=RvzjH|nS>wSfphXg>$E*=rxs4c3&<aZ;F_GM|O}j?B?~A^Ox&jTnM7#zg&M
z#JSReVMU0i;VL>xgMA-oMVlte<(QwJkGYwu>t9`c;&f<Bi;MBYzxrXk@s01r+}sU~
z*_81yt+}E5WbP=KHk_WNeWUUdX&i~@!1)1Mk=9?D!<xtA!gD=rm=Gy<-W5#~P5>WZ
zH(i<4T$m6KBf1Nm$Js)>Wo$xl)11=aL;uD#r!e{vjaWM(nv>ML&``mdY3o<;9|t-B
zqGU+B2p^!XK|fR%Y?{5`t}qq~FPz4c=uPzpCA_14R4-$o2B|s>)hS{Hhk%P75hd!I
zo`F9d9E#V~YU(@29O0kl5Ph+8h<Gz?B)SM38}S%{?|tJ7Y12^j(>8P?&|ajNbI}<h
z-KdXw10M!wG5+_}7K{L-UpqWZdwJzBW-y1<KO5488yGKu2iQLjZmVBhBU%jShT&K;
z2Ts*qjAR&W+2f5i(h}y8qsnyv4>enLjEWy8JqqIj?IR124MyryM)-~s18Z{j_X77v
zMVpB3B5f}`%Ek|k4}BaO4lS23E&@A-C>)-7A^qbw2FIbCgprDOF}4^2ozhi^Kp$Fq
zAC5NE%Lu|(h+JY1$c4ov8{mg~-!V5CZ+LfkvZQyG#bYOh`vtetX+}$W_Y_8<Wc))0
zq7b<z@x}W~-ne<k#=oX`BxAy!YnG+d9^mK5LG=wp#qp%Bxy->F2VaXA{V?=s{E&%|
z;n+KlXlIOt@SdU0i|-hNN#Tgo2z04?hV$;3=2lzquA8qx$8O%b5wE}gjkt5~l~}%c
zJ8s>1RkTO+T6BqY#=x%+JvF^le-VBdYr!NrBWNxPZ|$Ivnb5Vw7ksEa-tSBE65Jfu
z`=C!at}rH3Q?oWO!>b+DtNAi6zBHB*ofJ=W8cy+ZA{U8h&1vq9>m9WFqDLxF_SI^|
z!NHEBijg(Z@i;31_Xq~eEu6^&n1tW(GvPP}H=I38Kh@5?-JSURfB5h5bZtYhX^B?V
z<NXgF#IJwzW*i>Xqkhtg)u-$6%YXWn==~ELMByJ8ir;+mZ9V(T*xA|FxFAQh<4Dgm
z8Yh;kz+H@+zz_#;=1zi#Xo}`X;z#r!xn48?{1+a;#em}Q06nZX1-$xgz^UyV{N|eH
zBmCMoq+#$-`1|15XXo^szjHp_^z2&t`Rx3&Mht1SPd_`E{s#5DcnofQUi6dGzBGJ}
zKc5_2AN>Bb!)0*)=RRIqKK0_KmY22bqsPn2%kF#e_-H{wxi)-Vm+SI%ETX?>?M`GQ
zK(RS7LfaLD-h>%LhrWpkV}xOD!{jy1Mf_3JB!d!)4ZM?CXklnFQ_?3w6T%0K2Ihc(
z$)cQtg&IN+f*e8?US{@emQckzFoHnX5XK-XW+xEuA)(=`>V+u{83^ri#E|Ns;BARV
z8Oj!-m0&QZTBK|}-H6?d&8Qz7#wnZOiW#zaLBRRpU;TCb^k+YdS6+G5o`Cj@hN;nc
z-sKnZru5zH%K|T7l5htD9QN5QA3no1zH=|%_2D0Mfs#(7GO9kg+5~Uv#`xtre4}dC
zqr*q+X~KBvyRUN1Uo@8bM5x4Lt1F?l)~d(a&PMFl_8k?6Q10;oo_Yw_Coffm*h?F|
z?Cq%U2-FCPz#W(-Z}Ujwraxevm+*s;lg3aVcz{xmHjr=IxT1FgTLg%t)S!z=(4ox~
z-pyViDNPb7Hq|!OfilYYq5BY8sfQuZKjRBHhq3_A98wCMbrAZ2%yqdg*X6o=q4Kpd
ze7&5C=`3{f6Rv(%Nl*Q8&wi|Y?rK=<559h`0(beB7MRT7>&J?jW%~G}0uR;)d=+ks
z0?CX=d<zfOa=8$9?kvW_0{aArFY12t-Anc*N-1~Xbyi4=e{zlmb~0wEK6tA5DSLLc
zJ1srV`M#AbL<f=1S<;Q)k0O+PT7e--wB(sB9xFbH(hfy21)~*;LT<zhAH^zrOpN8q
zG1@<m-JMmZ)>&Cuwt=OPFS>oyBXPhy6nRZ7-)%Z+pa^V>pN~W#M@pa+SwVyiN<?G{
zJGkim?yLUn;G(YBkfC!ZfUhLn^e%dm4RSHRupE;!E0N7ttf1r`WH&p{pj=R0HXQ0Z
zyuMJ#MrFDbb8}NMStgB;<h_v#%N{(3r~A?2w$IParw9l7q1Wo3@tCbvBA=6NaDHk<
zYqQx<`_8R+sm`LrFWR|==na&Bjs}>Tj9#}FyW88bv%RVBUCXcKQaOlf!?^|z@$>Gv
z<tX;db-E(SpUDX*MJ%D@MzM)(gMyfGOR{oa@SlrnbuqjrOtGMLq*Q3gl5BEcJKk4Y
zNyFAQUL*q*Uha+9+IXUNZ5Xx~+DNg4(Zx|rBbgv_!|AYephQGK>`9hG7E3UlP@9SJ
z5|D&<R_1Cffh$U=@hs78!i~|q#_>x1IEbU;UE$-o+mTcMl5$6gqOnCmkJ1jj^E??r
zt|uxI!-DVzuj@fn2F5SarxBeaAgN8HaLlOxQ;{!E$2fR4MigT<*aN9pOer^k-OZb?
z#q9hY;m)+%j$yGQTttE_sz0jH2AclWsoH!f7_B?<>CT;3RX>X9tl-cUeDw_DhN4pM
z25#UJ_)9_QkL-(K7twgQ7n;`w$hzF8{sFQv^}o|;SjkDG-Gn10lW|OfsNf*@J9<U3
zsm9U<FyTYFtZ~EOFp?D8WCO(bff9LYa$2}l(mPOqj|oO+-amy&S}s>T2Bc6$84bP=
zDZzSx_7Qaieq$Iy*?f3(80}Ni!m}P@eU`?-2m3H-4OpHeJ=IKPC)g{Au}Vr|_fz9J
z#(dK|NIf{3t)%fB?4dQh=jx&@IUVIGr+`l{I<5WK+<2_>q}3!!j`t_y90emFxJ{}&
zq6~<lEf$mV(PN=`hVpYn^BEWcM~sP1`<R)q0`}_aT=RsKlcF!0?>3f+fEhnEW+;7g
z6Jy2;_RLXZVsT;7h6u)BdS*`F1#iI>!>VB^m_mb?>rSOOA=nf|SKQSrwXqV5E3d@L
zo!2y06`hP|X|7^qC3PcfL<~HMUZZdpJ_!ayYmn*^BSYGwW->Cu!A|SMbDeGiXOzyQ
zoW1M|ZW<q=43oi9$7oUXjM2u?@63u@#<||Wsj(lgzy3!2^$&j>H*dVE_e|;>h8AFx
z_Z+*ia?6o@tf^28<A4Cp67@;H`<fpNAA4`Hx0X|qi7t(fB|J4Q_iC~F<OAW>aV#$0
zjQP1GE5IGqdd_}VnlqwDqoN}*Qq-Kra7-i~d?lr2B*GS>3h=^k;PhiQI1cd>y#i_f
zP(0@ZQygAJZ#6e<91?J;1IGZIGKhK^8HH|58ed4`2|Y{mRd^=&!&4TQSHwpu!V`=b
zBjz{E_ZE?qqUr2EHa!ik7}Izi*~tU`frAD6R}+zfA<X<mG$&sYeG)IDuZRJpbMyOX
zuXqTK2NR_!rw4^jqiZ=n-t%|^3nIEPsL_6mT+nYd+HD19rq!A|q#Vt#_m%hsku@dJ
z8mHca+-ks5N?A=A(JdlvF({DQ7H808n1PP5_a69@qW$dDa@zEJfN_d?gdd0=i#E;8
z%-bLV4&W?3zqDw%fv70xiFpQ3Xg*?~WqqE*NTPwZkw8F#AHZ{oMypn59Jxh%;7iaW
z45UO5(=YZ^qixJx@Q8k3#DmtkpXx8|glA%`$Dv_q>4x`)gm+Gq3gQ#&5yqaHL{kGh
z^i7$_7xTK#{&M1z!U-Gs3OUXBoMgDA_~y628+Y%$s%J|Ym!?w^Ggon<L6$?Ty1XRi
zp~hByE0l!O8gPu=9r4+&=sNTQ2cFR^dxVvo7S^fxgb$1_*U*7L4{-dz5XGEgzOl6g
z{H)g{y&rqYm12H=E?$4*Ui|I<?LWo;@PGfG;-^3VPw|5v{D-*v$~T2i^X6L^pQxbT
zXxLDP(T3F4>`lq=B;%)V(W{sXyi=e2R!Mu<hf`2v@M2F>XeoTBEdE;(ts?>$9*9%T
zct+!+_p*;4@&QIM(@fEJ&r!`qXzi(Gm?Il6*^^5b{XTGKX?am_&fBm@rXI$e@f#Oz
z64{lQNI$yj&;G$)?C<Tx#`;>k^UhoG)|>B$<~5_)I*q;k!&qC}jMdeRI5@0{hcse)
zXE)w||6y!zNq%aPh7o!EDt34Gqt<A7zGJ{7iWg^mcpv+IvM(QwUEWs~+yrMV%cZPV
zIFlr~6dfT(EA(j0JHDqegl8GdTr<#td(z`EPoP=8#xuGhomb(W!B;=Shc0s6_YBSr
zpGgO2hMzj`Yv~xCPhVd+&!=np!z11ILV4l$;IVWMpMyXCeb(Xke&*jt3eWM!v;V^5
z^Ok?fv%><%x;A`Wm+SI1Ef}^mvtvRF?UEMx87okRN*ghCF{lxLPXauI2%*CpJ?c4(
z8Qg1vFXjjlN1(zZ2XACZ2KDI?6MqCa1br5na9Eh5l_S*2^I~#Nokz+$FDAtp#9TO^
zyfDOg^!rlR@Svh*3~uznBvK#A69W%SJW%u=Ro-c`@#kI~iV;TlB3gCA+K=>ZLibPO
z_U*gzKmE`DbA0<--;pp^dB)IC=1}R7$qZQ%ix~c6t<MEMT=XRwzjzKNr|<q#5&yOk
z(O>lYS$%`2c_f{DR?=0@4eA*F?^8PGpXnS0??(O889sFYU%iukyH2$**gDva4_6;X
zTS84%LLdS?<L7n?j5cmrf!6mzgqjRbT^t%F={p7}|JHNBpZB2LpguOgXD>JeXEszt
z8P$;h#NrC!Vp0n{1W*<{ClZ25nLt0hfKdS-I6=RW-x+rnu#AO;Xcmte8`=!05TSV9
zhEKKMURb&U{<cm?QIiDy>vCPL%XRqz<!fX3GD|RB`?8Db{MV@Vh4QiLKl6R?GmbB_
zr21a^YMq3Cy!4emFTUdkhc_gd1%6a2<zi)dE^gjhh}oHP6bh0X2J#9r1UwI~R(K^m
z8QVfyND+ZenqVS%kfAQT-@wt~ffW`QvXD98oya&ScqkZ9Y#h?!8y?Fd8!u_zm!i5-
z3Os}j3^XV`;D<Pb*rU2ov;ugfkN0uI?;v%MPKtlxJc9>((FzGGFp#<Q3<@B!0-&Hs
zN=qUICW1}Tt1~lEl>CST6MJzLazt;TP$I=NDSS>*=a^^1tx;T|%mq$($7iy0I*;U=
z*m3{>|MW>jK~w>xpl6VokOfgHWOWXuOp+~;CrE>h?8!|;C?G@72j&LG1}kX<i`1?;
z8~X5o#z<8mYAcHue#go;_U{^wBXUcM@W;^9^4?f@XOlt*`El*(YCIw>PU|ElQ8ub>
zBDL@^=lv+1ogz=N6JtwH8AIexaDdcQ^+wIl)3;=-B%+A1Oi^u&D@rrQpZC%p`iz0@
z=x9&p+k<grd@$e@+{T@q?dVF*x_SF{EG@05kEBsc3VS@wQG}y(;~r9uF_x~QNIIfv
z{VrI^i((gr7yA;SC?hIDkXK!j%LaxGj8K)zgbj$K?_#Xjm-T|a>p7x!Fc6_&WzOIs
zPoHRm<#)FYc|GB%mBOTOBXXqIjK1VIA`5aTqsOayUsAHCC@|qzt{Rh57zqkmvs6^q
zwEBrxIflmZC}F@<pHTePYsuIJJn`BG=h!!meWSpsi;JXiM<%!Ocxq11p+H92hEYbc
zCWbflpZ=V7FtQ}M61+oJM!}0ho=6y?9oRpNJ;pHb;T_LjR7tt08KC-7G}MG)aV9(i
zA5i?3we~_DSA{ypB)9IxgZsa8M5a?j3de{X;J{=4`0!A0PW!xJ)ZE|QjIGUeE9Ow-
zV;n@Ohe4Wji0r?#x3d$!`lnxdUpG>JA}gn;6^Ih!z`nRW8wCP`)CiG3Y22^+ZH+PM
zy4o?G6<kQ+CAd2}kg>e##bx(I*O}`?Vrfl5Uu_^k8LPHjwA7zf<1~sz_LyPn0vGV$
zve%AQ^U!-zVHljBn-dJH>XYgaE-|;EC(M@=g$6944bTaU_F8wSKN!|9s?mR*1EkCw
zXowd%#<?rgy*qbai5oZWSlN!D20DvUca)T5?wf7}Kh=KHW_sT;8z(ili}{$Ylx)Pv
zk|LBSMvPi#tyo)oY-2R3w&*Xk3Azjo1vfa~6g^;Xw}r(UQJvGgotRY{N#S*FoM4`2
z^CdseJO@lE;s>BHe`p-5q)yb1cVc~gEgB~$&-R47apP9J`sy3;$}6wOM1>Tl&;$6*
znP8ZVDWpQn>ODlmpnT7Xh7Y4YF}f5ch@9b_iM}#sq}&@NYSTt3(dn^Er>1=T=uylP
z;X1ABjEiwL8LPZMq{j(myNYNF5v-XE#`>z$tg@dH259<Dw3Ab?4fp$Dl?NXNFNIg$
z`%(vr{;@2nOmpIOXIf9z?aYdj8Y4D*jpv~`PGzcj4{bBu*Eo!hS0aPc1<|e;-}N3F
zB{WV%QB|w+>OT=<L}G)}7@MaZ{em-x9WzB^iE<guRaE~FE~@EL?WrHySmsDU;XLW~
z>}0Cn7&@RCMDy^8lLiJH_zFCoxo-SblNo!Ag6KZbv9M=35!2{L^1xT)IaIiV=hU4X
zLz4pux*KRayp8oXGz>if1)2dJTwh;}jm=ff!4uO==;x_EG3bycwIKRLTi|PR3yT^v
z(kc>>McP#&yf6p~mx`pGB~q+pgB<X~`4uM~QV~Ln;di9poD(gE&m{pXMKMET1tRgr
z#FNn{z@sqw0z38{!x+&KPlfNQb2>k&I@y=4><I0JdC6n5M0;yas*Th&bmRdiffp%3
znNRRw;DO$Qc46q8SN|rbD#k--ik%(=)A{)uF<F^aTX1?|jF1IMag1I9{)q95$n2)B
zjXLVfGK6p+^zwTJOF+-~p}{~Gk}WU~XT*cVhbBn}NxKCR^^HDM)Rt-S+*u;q1-GgB
z71cAV=P(X2zBo2?Ja-t2y}cwa;KXv?V;|CF^d!;@I3t@;cn*4rQwHY~uIa_XN77m&
zD-pSgqX+YzeY(Ii0Y&qfISr0tT$K!P-VskbiL+MIve2pUjmT%R7YvQQHnwUk(XkPs
zI=?U{IZg7XhX3gJ*bdf~CrioMhPitoIdx}aEml__#fCrE;{Ep@#M7tig587-p6qQW
zBr@%Wu9qdhP7>*-_sq>MSoQ-p(7hSek30_l2EQgX?-rNlV|jVWxQjfqx4*A<Z3Zbl
zkq?3WbhRpZzi3(sJ`yolVXGKDL&P<74qO`IvUOG(ba*`lx!*=E<|%)wmurbua_BqP
zsRImw-VL7fn!%^^A2fl_!8PCX!uP??PbjqQc}e#^e}Zq%>gJcfzw8)%(|?28KX%89
z2Zefv_kYr%|9$Tt6i&T(Jpb~wEsXfJ;p@6wm#=ZLXRzo<X=+k}Fu8KcBQF62Yn1*W
zl)*HhH86Gz3=nS^8R7K$X`*dIg|0j5MnVP)eD1Mfg`4yY0*Hyd4v7qDfdnHJ)kBEm
zCA<rdXfp<whnWI(>-&IVqclp9W=Yt9oMXhpc!u`|ynCE_X``6${`PjPKYASdo7?IT
z%7CtfQu3GgV`=$D{Osqyh`;%p|ETXvNwGses01dZcRV}X!UUIR=?GsezK?s-pa1!}
z^gUhV3nhK}^x*GGOw*t;G$|;82G6*jRQ&1R!!tZI{H6ajpsq{5E>z)}7CQ&^qj<Wt
z9$Q)@pI=>gj2KG{jf@enVt+S`oD>_H^k+dwW5A*b0TFy4EdtLnb{76Lj+BhX2jvRF
z7YZ*c2=tvMA#^6#l2!qQ2!a&?r500Ke0e_z)k7*E1a%9*;EUTyIudfPlL`PKHNBgZ
zm(DA$7YvPcHqc_c(>)k9J8j`;YTk9ZF4yI{eBSbv8NPn>!@vI0<z#aw(ZH7$_zC3`
zF8?v*<yGly`u7D2Z5y=j+4s+JeCm6+=aWi$-h5qOKX#{ijXwsT@aPLGCtiR3ZhZd-
z--sJGX0`auJgWn~l~VU5MIZbUae{-2(gJr}cOxGW)tsVoh7mnHZ==C!N7ZRTpG#?V
za0o(qh6gjhQSxCxp>1r!jUvwJoh5sr+=E}jPgzW}r>dj+`zXU}vA?%&uUOJDSwWU7
zL^jWc{wRFe&kLDvB-mG@Se$ZdA@-rdfYR=CBbzHa)ew7fAu68rQ0|ULAy<klig$Gu
zB`XRy??0hFrxe5;eb2;1k*Ft(7G?E0E%wz+J(s-2ld&Tk8I+hPILFyPHW~Ju_AO8z
zrg~e=<{-u0M9g|$iG^TO_Z0o4L-3<cHZ(sx+Oy*n8}ZXd<O=Uas`n#nlHv<P5fLLq
z2;nKey|WdqMk8jb)wpx#uJ_z3l`2-A@qYID!nlA(GyTI@fx;0*8ip?vbjZ}C^I-#h
z&L{N;=zwf<Za6g?bsLJ<y9cEb9`2;TBBdcI)bM7vk)}k7xRE$MI*uL5wtQ!gACz^;
zIVeTi`vv@`aYWHWK}kdu5z1BMG4SYce>dt!`+`4;JQSJ28Qp6mgW8euYI8`Ca^wXD
z(7e65_x3j9_;_D$Nzq~$EK#VVghB~O>OXHHuX>V^@<Mg0Nu<LI3~Zg~ogS+%J66mU
z3hb+r*LxUaWII}dQHIW-%%(pYt<kK$q8uOIBbinoC~}E}d;0i(?Ch-RcZ!r?WB)0Q
z--PA?ie4M~G*&1}9rcl!P#weiFs@?s!61u4K(Z_ISx5Iw^QUv-c}$d{4JE=Y=5n)n
z6zz6hb0cR3c}h2kBAoVdKo3wLPpCf0!R&E1LIfX*YpsP;2XNV1e;B{}^*_YM`eQ54
z$H&=otfVm$zCkwton&3oteu`V;`neUjt=%j3yMx3%Dz}**(t$!Tzze6el(&{tHp;8
z-dFocA()H@7-kDu&CBzXXgBvIHe)EK9Vj-3u;@jrwi9bl-wUG9pf}?fo@`V|O8)2z
z|7&}KOT+!h<WXE!b%FlVZ|Gr)Ai^vRO<_+T48@N2(;F`?I<dL&FrGgB(3vJk@yK3(
z8Sl@9aXA@KQD8%-Qc6zVKZGy!kWMp2vM}#OQ9Mq3;9TH+>Gb<b<4_Wu=I2*pvN9vQ
z$yhn;wW@F{MLp^{M<?iW1Op6;Sl-1v(3-iAp;rST8kKMbSe-{rWAXUW13NK*dqc+r
zaBvD^`NX8=LNa<31WRbucz&?Ap6YjWBll!6q=-%o+D<f~jg%<Nd6kWoTHiDdVq@b;
z9BBMW*Z<9LeLKGS?H|M|uYFT6yrHq4)R-jYKGA~b=c30Ni>&4tbQlMTjw90OtLoHY
zzBQa@Ue9C1&kL@}$n-q@rS2hWvws#1(U1OlGjbWxeesU|!09p>;lWwK7RLsC@3UW5
z3xg2RO_Mh8LF3Uw@C<Z^uBM1YeKE`_{KkL=uMqyL{T>_OB=}(b0)MSMzQP#8`?v>u
zWeiCjOq4Iqy5JBn=e#(#=nzrW1@_V^VekTO0yOJi;d(wVIx#gH#mU)VpCmHnFb}ih
zMVZ`$+5$|di?#~S>0`So8ni3;<zjA*I!OVUQV>Ev;Ss>PT%3$*Wlnq@=M3s$U7g14
zQuLTeM)rVWPe|rW(#fQ#Ab8I?{nj6K0Y~14^9kdd99QWdX_8ZvInE>A%U!%2gAsDV
z*4EQ_@cz5;<ng08Jlqq_I9A;!f(-KpS)%Lrq8>0`RiB9xM1N!%1-#W7p7xIkzrimI
z6a0o}19RGpEI_oA@lx%zK}4Tpnr|Jw4_d|khwva_gRY6jQjcgK4kPde`U0=Tallct
zYCn9Iy3zlThLt_-W@Z+mDt<X7`i}kqnTQB#^hd~1L~v15o#D;AggK!D!EkiaJc<4N
zJsXmsCriusqB^^5x&+@Ka=b3Sg)_nC*0$uLjw)g#gm;2D*>hldWXJ(e^0Y(p5qlC6
z#f=^W`Y?tONqmMmh2zPE@B^nOQneD9d7*JSt{n>+X@9XX4D-y3%QHta!jFhj1jdl7
zP6q?mmHJKe?v(15Jl93flTw9}0+i`JjG#k?VNZ38?%h_)GKi#2&Bc_hfE2FPY0+Zg
z*16h}QfCSmB;R#)FMDscTcr3s(R(n!!LNboi24mZo3L{S4)4y|qjuVnLGK1u?5_s=
zm=}|J7x;?U(Gd<lU40mDzw@T>bJKMu2WiQfq8Z3B;1SV*Q&U7gA|DaGTh{L>8^72O
zkSKcg*TbNR^E~^wkzy5HB!)?xPB1Lv97DA1@o`P^bWe4&pKC?Ol%KCw*$cHIyqb=M
zg*nN;80<??(RJPh*+JHKh6N2g4pn@nH3O~JOwmpB&vF_vk?<}Vz0f&|PwI216ZPky
z3@IRmvC}{QXDQM@h9589NqO=2A6~N&j>iXws~UVsAL)X1n1ds|B;7MCFJAxXq5r7O
zw?BX4<7L>gbocYWj~C8;T1_u6FKgS&j+d2}-S=q+?D^X8bzQE@*SK71LCgYmTnwQs
z;!YYDm=Xo14WmK;gQ_D)Aqc{_P_kT}ODGbtMp(sL1A=6xrE>@~5P?MS5kMibiLh7*
zV&SIW+=uYPZ*Jxe7>ETn9gd*VeVpe$QuJL7o_)q3s8sK<!U@Kq@0Z>4XxC5TaC<Ly
zH@9v4Y8=)44%R@IlCmq6YW(aMzlgv6+y5M|y!xsc-7xYDC5YSX_fx<B)5qZVu;^Vd
zZ~oWa`W#%;P5d5_@PiR5u`mN86&4lxbA4(PpS**GSGp_JMPE3lJJs|A!}MgjXL$GU
ze*O*5km?X&_pnim_3h2r+TV&!_tc(=UONkx-b;%`iVhewgKh<e7Hz<R@uWtKk%O1&
zUf`eJ&Ub`P8_F~u2rol#Y|;gRFDMOAJh4Xx9%qD+6V8tUkjOkd^pe4~>u13aiUk6-
z0YQp%PfnL8!HX3KLIrK)Im2MkKH7`0LG&AVOP==L8PB-N>vCPL%XRrY<*PG%{q+|$
zd_jl@#MtG<?}PG{(>uNJCGY&=&G=~FzDOB9`_bdm%HSUJ*kKv8df0g65?_SJ!F%yc
zEywr1|BYCfXU`Wrp-~oIiB~0u97Gd4d)V`QBrW(+D)TD+hjJX{i{&KM46lNhp|t8r
zt~x$C4E7X(msz=vY@@pHsz#Au<rj)8JNk@RDU2c?K1<!4Ls3G#$X4)LHm0u~A2=F;
zeN2$&P#&N#LqUPU5m^Jn0E#8vjj;=*r~9FPvwscICMdvBFj0q<bK>>LOI9AKN2JO*
zJlwGo0i_r`oejj<NS?GplT#(h3@E&iu?9-i5tNzw(0hsg39e@+b*>m}%#SgOR8j0J
zM0zk3J9zElO-^*%{K7&^&(2HkA=R<QgeZm_SqzB0I1kc~5#5ybpeXwt7yy75{b9^d
z03+7`6Q_u_axfQjv$HW-F6lr4T8)z4kF1Q6gMP8!ndK)v3%tG8m!4z1k&n2~5%GEk
zZ*lLPitM6CPV_ucX2?J6Rl>W8NW*iSaYg<Cm+)LhMg=}72-zoty}D5TROc21WA->Y
zOM7qW-55dH_k~UD1xzb;h<c+xyc>`s$K%;g3a`1jIV*gD4SN}(;6;YSV2RQgWk1nf
zLxoAFU5oqo-}b&4^Rw)QHXGvwqOhi{2xgvm?^fMQM;Rj`jjRyLNKV%u4ppSw+$K7!
z)ponV-SKRa{RKVAjctw98R(#TQdIxwD9S%@je^rt+v&$h&)BFR#`^jr<L`}?J2AJg
z5=51hr)F#vJZ-hSkv-7|DZ*kR*az4i>^A7I7SS3Agb0*UX?{;=jijc#4Hz)X>hD+{
zV<Tg7<$cfC3#ce~7m7NML1a(|@Jk23C8Lf@X1+q1d@dZrBR(n7F^Hd@wuB33j<mq&
zRjE{BhJ9?NH9pe|QK>FDijg@De&cZt+}JN~c6P!0KUt}-$&|_TJ+Fy2dR1$jS6{sw
zQ&a54r16?v);JX6^rRL~o;<dZ9Nchvc+E?9{OUX!$2;-(@q21_Pq<OiIeirjlGB;)
z8-W&K^lj-Ud+lPlO$z)}l0k`-uw;AyA5r#du=F18BXaU8_IKA}W8<mlN-<VeRvcNA
z(b$k;?ez3S_<b&T5N(=O{fxZ_LkIZDp)rHTy8Rf|a4cXx>#Fb^I)Tx)FhM%2nPjk_
z4H!){$Cw}BDeFtpb7AyGc@54yOXqmi(JRzfj79p5K^_B{_n<oOING74_LItuh)NW%
zDe4L17sloZ(@1Dv7Nfilz9+Pxr6^%Bpd|QV^yr<DI`YC}<CUOr28H)Uw;ji|y*N17
z()oTYuiS_mH*O0*<{e>_)%Xy3(b2qO-qA;t+KESrM(W)0@sW8!N)yN!YW=1A*v~FS
zXJTl=(4j+gJUK$rPUZ;(W4`!?XhWlMr1!ChTZ%ly@J9+k_O#@k7<xv>HP*sg44pVf
zY9*I&K|R0$07rzr`iwII;7AAa2I~YsN;-&ij*IXm#!z%+?8^L)C?6t&tV0k#9p2+T
z;VyB|AX;T;=vK|=ns4em(LJQh^xVO)4qbsj>!15?+D23tbo10}QKI%ziaDM1_z4d$
zyLvn;Iv~EKc|@u@B3<Cw;24G=qRJeFE#5OWiX+HmWW`SnUlk&*?JOex#MnZoF<#KG
zq~{pdSalsm1z!ix$1|*_Q%W}O0roTr8hfh#p_E5KU!5fydP+Q0_$hwMxPqS;!l(}#
zigOLaOvDa*U)5@L*N4%0T=z2u7}iVdEj0<PgJ06}Bm)pd$w51qV}Nl&{Fn%A%OV<k
zaF5Tj+C;QCK=j^l1A`j87anY*lwd}`pu6;eG{oQ+ZB9pWV6h%ag*iubFwtK`mn|+X
ziqDTbU0xj-L~WVCh$H}|h;Qa5(QiFx1KQr+6fIgeUEEmTFb$YnSQ39M4PeTCW=+Aq
z6+62-@w?x@84o^qV1pS28P7DOD?An4kn>32m*#g*yh7uq@grhTWJ<IE`9pGCyA@}e
zLpU-Kah5Mt#MjPaZ+F*o2%{lJ%`2iw1tXkzNVhyHT8KkIS@RCwKwU{@n2D(w(z%v}
z8{_Jy@Xd4&gCrT6n3ICN=FFMofW4hv$%w}`5OO%mQ8Ed0m%bsB03)YE#OMmG5d3gD
zazvCqkx#*W<Sz!^*$DKBfqiRp+s5FH_08Da+H%^#_uhLye*Np;#YwXhQ{r_*SJs7Z
zPaZ#sz5QL&V8$OK+|10JQ?vpHQo9}<;c!GWs>Vd`0N%_o%f}po8L}q^FL)1ZjXl(G
zoCn{mi`2Y8KD2z!pKfAaV(eO6Oi^hi`Xd<;hmGW1L3A**CZZxdo1dGp0S%)ie24qs
zyKa}>@l3y_f5?|S$Ge^t&H%tbKfi0ttaBS2`ei)g@GU&gQ~cyS&kyS19_ztAv!r{{
z(;xpE{7c{cgU{#1cYXZyCl=o}xcAc@d3N7NkB^m4yzsRtNF~>Xuj_JMzUBo3)|r?^
zd1}%Ndl&`=pp;fd%Qb{hF>n7?;WzpC5nxzAb3Z~@Ujh>g4g_X|PLw5SaSwAxDuS>h
z1SLZkLY<9XVxErH(S0x`dq1e|^c$g;G(2j^@HxGY#R5hpGYtJ@#TX%nPVb{S&m8Hs
z6YEb_<D^!Lf`ld(=6D1lh*z{Q`N>ay8bAN}&*F`5eM<~=){KgF6NN<N5<<}*5L$oe
zKfhCl+V=;HNq?{_e$!$bwbDa6M-TMfKiwPkkN*ef2InoTa&V8HL6GFO!7I{veR9Q*
z4eEVX_uyuP0N~q{V74Y9ZhdDX8m$wL4LPlol8d@AR3^caF$azoR#g=SFyKLd7)MeA
zAS6>50}JJ92*M9uKw|M09{e!(XdE0-HV{UA4cO^A-b5I`Q0(BbO^$Om(dP#56(*rJ
ztpU_O;Q-Ho3s%tSV2o0#OawnInDh=5CKfOykR!a<qfG5=Hd|Vh0<r6IU9QV@`CR2|
zX!wG#LKOK-2lV%o%k%p_e*DV{{DoH!kI!8O&kcFz$BWMEhL_$mc=pRJ19+I%BBMNi
zgWsli4!Enn3f{Vf2(8JfeB8Us9w7zED=02fN}&t!K^$FNrT#Dsp{zxrgEHIiP`}t1
z99aY<9nm~^rW477Vh7{J&hD0@h)|djEr7BPMFPq)Qq|!7jWRQb;#B+&qgUI;D$?2M
z9K2inkJN?05hL2e58sRZy)E-)JliW%Rmpv%17<HIl()386K9>1IB6UO`*#&`>}!B0
z@_De&1>WFz=prwmL~&}nE>UHYz0OIcbz#NAj^quz#}5t;+^?im!owIN4xYQ>+o}_J
zZG5yZ7y@?;4w5O5FMC8bpp-$;2+k9IgiKJ<f%2^wx$IbE)Hh^|R{O+;A);*d4~_(j
zj^IVx#|BDD;MKD%QK_*1*K|@k=|EwD$8+-5?MD{HMxhv_dBiZdxVT~^Nb=V2>N<NN
zO&Z21s7N7&9703~3Yos}#0pFw$SBCmDCQ(5={};_P~xGu<3NFQbi4<K1*w*hyBhUl
zr<<~(RQI4z;(d5jH&2dZe}6k_!Vi>t_g;M?=H`|K+rFdPkUhbzG)6_?&{T|R42k$a
zNeaA>r7&t(+1)#{@sB8&e4!G>@?4DNX4Q6ZXG|EYF+o<BEX8;+Mx(+PA_AVSK32aj
z1}TT;G)yRzQ?wTC1^#&b(+?sbP<$54GaB0|)sM0-6PJ2FbA_nMi%Sf+!ehoEhcvCR
zRi7G-L$%>lxB{Ml3n&qpl@#G`uaU;MFPTy>0;Yw+WGpV+is|ZdWQ((!C&CZm4r>c<
zl&^lF48<^s(JZdm$BUGDL{~KRj!U&qlUs1=Uw}g$8xD|ry+4lN8e=KF8L~8nK8)0p
zq)p5avBP=^*_Y!?U~6jn^B$Drqy@axd_l<{8sEt%2<{_ez<=6%pP*377qY5vI;N%<
zVnVoRxQ%K&bBPWR2?0#ONuuQnnrowul0yBWCI|(A|3tA8lhgBhPu_d)lqS_i(u}h2
zQg%E#nzQf{6py(H3}?(aXqDbE&fZ=)L!ywzX)Ud>^m>gkN^q?eHHS%gMd~bIhC-F|
z>a6~$ZuQ;k#SBVW4wT}w6Qc)f4B&U_lxwVM=4PvEH)|16zP94{_#nMg!~^-Bs3j8}
z8xt{lK|5%#&il~4qJJ4u;b{(dy@S!f>k8hBqMiG2jH_!6QWuV%Yg|#xM;zA<gh%XY
zsCTIij(pU+dS?xR<=h4mq6CSEIcx4mtD*L34T1uFO7w+DSo#a?q5Z8EdykzN_8zm+
zOyr9ThF@RjZCn;^L!S&Y7gB3>Yz)GPfFT-Mj$;DOsa-wSIYohfqC5LhoG6J-%?NHN
zw$)$F(R%$zILqE#W$$aGRgv10Xg2FF(~~B9;&5Nv^cmXA97#&?=VJs+83WpD5Su>>
zH#8Q-37i`)MTfh>+u5kjEeelBt3?apXABh%2Puta+DFfOBZ_7@LRoELf3p+OwjM@Q
z&1Jnx)hC?*?`DiKu;_=sCxez6(id$ddUAXua>8@=jLhalpT!$fq!;r~u+{xz7`;e)
z==niiSE5S?f(0-pt*TyT+~*^$dv%NYiveedHyRhw1M4!>=TRmOb6WHxGzT!$5(x<G
zh{7#R6%A+79zr|W3l77i(+R6C)~E2c9K3qC{?<WTkh6$zcG_MWomhh>=LvKmtf?_r
zlQOXl4x5e<^@ia?aMoHMyv4{5R$&OK%n+$Pr}3z$U!u7h1EQ}N7nfssdD(j;67@to
zh}fg_MB}IDmZB`&20xM!trCk%ccW5WjJ#ky0$sNLM{^Y2kLDlibm~FJfP)6|3;Kes
z?X6f_TeB=ebPmy6<#NU8k7*bEgf60I!T<@Xkx>SmB;_8d6v1hpgFjNJ7dZ&r!C)-*
z<2GQJWdAyb97hqNeV{f(P*P_y)UX9Z-ZBivzvH7LAMBlo{DHBJw1k(UeJT31?J*((
z8#q6H{5byM@BcC0d*{CTwj1ph&I^)@)Na&N%ni~;X0pt{BvS%oqWVD>BFIJrL-x(e
ziEoi|6PXem8qZ2Dlq`cBfx``{W=C*>7VY4BHdivu&gqjhr0m(reu(HlFmRF*7+kvS
zve%u?5A-8y!T{HklfX%>9)|~qmIsih%<s=H)ket;k`a+}KqBZexWpVr4kF?TnT4Jq
zf1<ZT-o;P`40m?-9I^HGJMYGS{V)G5{^^&$inrc+N59{XcXhmf|8Z<=?CF{0Fw|;w
zVsG!zj?Oq7;RJy}eRFG9O-^zJdLyTp7L1eJiP0Q63m%2DK1Q!(Jf#mdLW)N49CHc>
zGR76T1l+=qg_9Ta6?vNsQh;W1Qg|tO8r>MOJ+f`>_|VQy&}EF}9#`mua2cIcL33_O
z_=54z0OOzD$pIcfGk`sFpNqq5zR?Bj(D<rhCb8V>@2nfrIq=FCUMRkoX9xGDbA#V6
zet+T6e?0R1pbWlq$v=np^4-VaZ+Kp9Pv<{=@Snf?-eI3!ehhB^!{deP=`EDu=a*kz
zeBbc+q6N-!ZTPw_*X8S8L~qZ}(!LAhELb4I!}$(T6VtTdXWA@=EdmccV9^c}!DvS4
zJl;>7gE+I8gb89eK{z8)3t<{%3Dw|jW`;FTa=;Ai@t}&i4`FHWdcT90LxfO9p@70$
zMP!MKdY7KnJ21ZBad9q2eYAfN`#U?)uGgc}Y)cS8z$u!6A~1dLd*6@0`J2CqZ+-jQ
z&L@vY2!xsU4-W%@PyM6Kyq|Tq6<hpPJM>S5wsCzZOh7PG`_ki_OwYjJc#gA<1fy>h
zTx<F2pNYil>htdsbo_<?bPx4X4eg}9=LRR1-~G6dfZdT0ymxdMPdC=$pm7vEFRlQY
z7OB8{@MmL@4l1>G3dTj4v+4ofbc-|sR8RY1x<dhnG31`1*Z7zgCM?Rp3A|holAPWL
zk80J=BB(8)5#a<c=+uW~cq4yuS!0933b2rd1H}>{`MjHVAV8z&<HUu=+u-Az$1cIk
z-Zi?8m(q9+L0msb*B~5qx30@|xh~h`)63V!@ReAC|NoQ%h5RB1RGW{_s`p=~(1K4c
z>FTH7^99Ri^>e_}Use*A)Tj81`S;j`WQZ{>ZVOr%<Jp>=M$A9Whw(f|5g;C~YT<40
zEIWHh4&pwPHKg%D34yYR>!i3tK7!|wDhtCCX<Jb4ka8-?6q41n*hfi7dM<H3%K+?u
z_2PIYMIHO1raDpD5VcY+7lZvdW@nbHd^^M8<SgwWgL3ZZXy5zKpj;&~8-9&pgVZ~u
zRl)<DR66Fpk`=h_lsKxuQQkPYAgif0ymNMD&S}eLX6MY;ktt9>pgcnHW0;CxJ593s
zp&xN2nF2K@p5;kN316q@$b>I^7f<hYj-ysPRC{_cRh_m{5XBqH3sMoHY-V%%6ZMId
zHyEc#Rm9#SNrvG~%moyPC|gMboeB23A#w!+nA0%pk5oQq>c0(JdX{&#+i5=w`hb_b
z_i+*I@zBSa=j`;<3TEoHceVwP5y3*Tj>Z<F#Q6ozKQ-TjQqRt0gVa)_jY3|<LmhAH
zZKCxKc8vqonc28|@3mN7xvhIDeh$L}N_4)1U!<f2Ce&T8A39151y&w~sKCgWk+R$h
zR+Mh}$;cHj+Mwi3aAwb>QH?{|Po!%lKPml;4;p_w&GE!0wUwQ=bf3qGIdzp%CXv31
zv{c|+u22?!Vjx3~)7%*CM@B+rE<Y|Pr+#B7%@>LGL!o*Z&DOCEPdHy;NPZsiL+RR(
zL?2P`k-}<HV}Y``T$+h2h8!Epcm`$cSQK(&Y6tiQ&PmqPx`D{MgM%F#9Z7wJ(l{A*
zG{5xj?X8X2-Q5+=TzcFborVDhyp3~pH$_5@v)_+X<Iz5@4RH3VYr`RM!{eXyR75}$
z*)&<16DZ2M4`X01GWH(F5J6f|;lQMDTkyqTo|O5}q_k#Y?c|6M^#M7UC^3uyh6{6%
zaq!%_G(Lj!Iie3n(->fUBx(oaDu%OUlvLZPLvw8OO1P}?-WT3=7;oV&^on{C9tmH&
zZNcx*#xIok%pot~^!&N5bw$fC)S+lk<G=x2Fj`UQPd9$`&%ca~^)<nkXczXb(b`h$
z32=h+a?nbYnb000T+%vmh;P&fZ1}ChZ%5rR&M4mn3&FsPINz(4(nU)!!SFC*BRBiY
z06*v)<ILK~k*R{;6-Iv1N(_aLMD-l$X??kG8pOVPM1(MB87Jo9ndlrww1b0f{qBoj
zu&*9UZjJQ_2U7$iEQ;!knnk-X<_*?g{gfFamoMpF#w-^^ztVS%yS2u?aOl`H2%`{A
z2DHOUUzEIp8~9b8o{zEol;${*za8T#>A^}<WwnRYXQFV%Um_XO-bAD|B$~sxz+9!V
zjEUZ>CaGOR*o&9tF&?0-XCJ`Xq&(J|mFN}PiE#|0hliZ87W_vw@1d`S0`1b+s?W^v
zZm*?wvlm>&(Y-{Z;9x;Iwb7BT;0s-048bAVrExrmE)(@77z+N3cQQN=+HH<Pd%+LR
z@jDfi*F+J51FUC>c*5A?$MhNvf@xYZx>1L?uHXWEn1dcN!KKq~7&a)kN7WYZEe0-(
zT}E9vP_Nb0KK5xOQur!*noB6&q1EvCv7E*jgRq{%FpmKOKEwS)-4Iy?oN$yG)0n{D
z*c;3#chy$#_+Vc&ersFr!q9~tBmC~wr%%L(uEhU{M4XRug~-q;jbqxUD@C~w#Z9_G
z(F1)hEG~)1R8>C)ylSl6ycd;eB8vqZ^^eqyg~>U!SD)fhq)lf3%V|ABv?_a%H5v`$
zJ@f*4={BjqI9eo~gJ#e<`zH1bUyL@)!HEKUQBK(S1Dyd*y{`DE>bg|_#coYw;63OU
za1cO;1D@F##gUDo6T0tGcnIGma&32S*W<t*@eD7<j+yG59agxP_n>b;jzRy>YPZcB
zFsuxtphz(ZE%Tmcn#)8SJzaeiAAInEV817M<1lu2wlx;JHbgQ8DI#Aq#X5)!_ZOoV
zx)#Q;I9U}en1|3~^+T{>exS%8t#3Aq1Iehy496eQZD^?~#+XZFaf+H(5v|6N1mg@w
zAtK2z`m^`evCe^;IKP~V_arA4Qbg)G##G0oXhL;XZ4o|a^iKBRBYKs{XbhB76(a7D
zThNy!eG&a?o;2g|=s4ClHsaC4$MNLx<9PV!VXUq`jV<Bh{{DgQK_|mLTtp7zcr`b>
z7z+z4v9x^4_%`WCzKrEoXwWblkUD6(Nt8W%86F)S8F!r0Szl-e<HvNRk2Wy*1n`5m
zr1Xl6D^d0A^@T$K&ep((C}<41j496@A0J1v*)(lne)Al36FHdo4RI2B0Z(nz8<roy
zISgL(gSMx8X@1g$+OSS5^*r>-g&Vj}ZA1>EZa<^`0pm2zN#-=$(!DN&5BJk$a6*M|
zmZt~b`HxR<aqx5S|KP+6B~^#q0NnoYE`JVRHz?2A`I3YA&Qtp8&wpuo**iYZ@h>Sa
ztMA3*qXiCeZTPw_*X2*4NRmL<hnQQr<5O300b@k?6$61`GtpUWSkWUY%A%fV8}29V
z155-(3=CbOt!Z5{Zs7^YCZq_97^@_}=$@gG1;v6F7zmUSk|6RF&a-Yq*o0|#zX%wB
z&S2=`dHtvF2nniF0!8ESD0akv>jy`Y10=vmz$E2e+K9H`2(O?1^r!Ld?|vud=NGi7
zNS<C0Ea=!XVwPS!AoPh5@ZzE2>e)VDI?vyW#cQr~b2>5j@C#sMz*y%}RSDk1TPVPY
z|M_rpI0hg5<_QyC{Q)DCG5*wH1&fZl1cc4Kt=N(P<`jow;<{pG2V;T&%KilyD;XOU
z2N<L<5GA33113x(#Qeq75oTe`2o~u+HegNZ3K0eoK&YE%STvGDoUn0wje}zdmAa15
z%efpvjs$L$6{H@>>wZUy319HuLO4NSVc}^H!~|Hs2mAp~z(wG4uJ+^g!y?&`kwiyW
zdKu;91WF7&$Df|<UjX0ha$T;=b$My|8W_HYH0sl6)yJ>-3^e<5l*D=d6_1ZJ<FnsK
z8Ge4jqIbaqKVF9C_<Zpge2D+zfHLIC@D5+X*+eDDJbWij+ej=eRO8O=c`cH2;#KF8
zOi<=Xb{Rv7kI@1Hm-s8ZNiq?NdK9)+Qj1^mPDd?B*5O(B?r={FI5-M%QW9SF*=K-M
zgDC&?taz@Yn=eqvoJ6f&(>Wp=(tA)8!5>lB@O~5qcwX~L>L41RIy3JyR74)2JVVJz
zG(n-D1;5?_|F;rMzfm-JUkCMv{VS0B(q3O$)azXs<LE06FIG^vuaez#9vKCt6?-ns
zO2#UXwoLtnH}{BC5O3#Ml(Z;$^bi0K(s&JxLn*2ssmzj@j-f(0jmu?kKQ^Ae8@oH}
zfl&)(1JM*Vc$C#f)m=N<k9Xhst<&kTp?Za=|6Eb;WK2+WqI|??!agdkfkI5<IfB9)
z@B2}=8RH=G2ud}KeJD6k=-Ini{jhQZrG`G~ybVenA~TXw5x9(~|H*hcYS{=S5C%q+
zWb13IQLoo6J0ag%@iUQ>f8YQSF17k$>`RtoOqQ2##f=+xV~&mPHCBDqN~9B!8x8LX
zM(V3ED+R|SH+er2T{|MhrC>KZH>diEj7X`QFiHVeJjAL0xsn&Sr!g3btnNFB#~=PC
zPU`!9FGfTZ>lkNPFA!CyQPyw@$*+2GNd#e+eM3;hUI^Dn?IjsrZF4F>y_5ZVhyY9x
zB|Mw4Aup510HgNnULq`rG;$hlHHq|ehP4X{=3I=EMlMs-bCZrP$!a{@dPGM52i`E@
znX{LX8_z{{90h;DxQ)WIRy%SA0`M7S^H_#7k!gKkBc}S7&#4{^df=(rG!`B8v*qTD
zN!Cq<sicsje+hmVHIh<(d^|-JvCqrS-gZn^XY9!j4yB?`D==F@L>_y*T}XEJ+=3P{
zQw7_;U`wP6(L~1wyEfQe2u{t$VQg;-pK5!8TSxuZ`bso|y;v9r=nXP64u)>@mB#+E
zC-`B=!xJ8Z4sDhkujh$|u@Svcj_H{twRh2YjPZI@_Yhq`q#?$!Q{fSgi7A><cr;4%
z5s?ebXN=seAtAEBg}IF3i+MeWAQMeuP8?~z9vvQ9_D3OEERmkDs`hEl*l7-@I?Ykh
zu@r?fFm}06_ppB9i3^HQ&0Ed&3maC5*cr^tY&i<WX{`|$7xi0pvv!64<H$#J>}ZxK
zIL&Jecxpp>uik-SPjIl36x<XohF%a2fN=o&i~)mwm)T#hGOPFE2qD}QF6ckg1tL2}
zZ43r)XfucI(eT)~f@Q--8ypBRj<Vle%c<K~gG1LQ%T>)oB4voKZ0LO7w6HieC%9pZ
zWu2$`)GiF1dEK8;zevT0qML|G=nV!ZXg@|&4Cw-q-<NQL2)U-(fZ<WUO(%do#@bvm
z{+Fh8DknJi96f_m2J@cCx&EcbfYhb^j^?uHV4mndj=a7TIYF=c`pjtl&FI{;-X;F0
zHgwwiadfbz{<W;|hR^u@!lx_M%N}#5t)`wo7cM1!?uj;-^Qq8YU<rIV4{rya`X+oZ
zpC`>BfY&{JjU&!1Hn3{Whj7UIzj-V)Us!uW*GKzte7GB1TWjJWEj@o8S&YUbU4gbO
z97%gyVid+<1tmGg3DW&dl!`WLrSzcS#7Oi`bzSwbCnxJ<BCp}I@L}i^@N@)~#+Ou)
z%xC6Xzds^4oC)qtM|t7!@%ZtF@%Z5*y=OA++<i4>XP07Jv>)0E{$KUa^ql(1+!S7t
zVh`g+GIEHQPF7=Pc3Ja&$>RZz19Ob0qz-lDooG2}EQ!R<<jRpN7Cq0lw>Aw+96f9d
z7f<JtwLV5-90+h4Av&npY>4+?SZ>Ix4d?{W7uhJPvAz`8&xO<Iq0+uyq}junTJuS?
z5#0`vU8Dkp<{<ZQKk^0!C>*T4J;$YJzi3BwW-b<&Zp7T&g8Ag3XbuLm>FQL>EzEnK
zqhnw$)4p1rNU)k@ievBp1wVj3v&SWo@bC|45_NmeJmD7+&&X^?8fOe{PUR_BLyLJH
zXK`dH-~-=Fc-m9jyK1XG)yKYQJO*?OVTL`pC^%j9+WI8IP&DnVVTY6x(fqBAC-L^%
zztddV)jYsRnv7FK_F>R$wOVm_u&;RxuTX!Ecg<^rW>KzGHGd~f(}<uW0-b$Roeo%I
zhYk&Agqma(X!N#Zm&cDD#|IxijJ37(*xA`rUj-k%1A{6?EaVqbz7F@^A|)bvAsj6*
z5*3NKbDG40kKx{2Y<sc2vlDy!`wUqS-Hxt~eYg&f4mA(j(Up9K9CnH9D9Bm2!yz0&
zmVgAHcS?#j^pAx>cBgC{9AF6SOO7uZF39BQoeuYrvko-2XQH2Ko7w`Nf(!5-=3y3k
zh74!j8L7Ry(a@YbmE4Ei4Zq~I=;5p<(;w-ZF|c#ME1nzZ)JEM0PfLfIs{1V~YVJUj
zIG4`T55Dt5-}P>;L#}+8V%n8TdWKV}u9RiUXFl=Z@kIYPo|klv&liq#<`WP8Tn5j6
zYVjlb?-LIH@%cv_uKpvxzpV0Ny&Q0yYs1%dxh{Vq<pM7h362N~_9_!2Kovt|5(Xv+
zIRX|QK?q1J`e7;%dKNX@%fifFa=MPN2w}&&#l|r`ugfa>PlVpYo?jCFZPe0%k;F=Z
zbffBJftD6&yn{AT35K+u>LtBIMnpZUd*ZwwO)=!1&CS@`+Ky&T3&t)&F#Ai1bL#o;
z|KKm<7eD`beCIpglK_YkL<?~}i^Av`@zQ%y0#VX0e^LASFMXHz@r=><4m}4`VIApb
zxX4ZEm%qEO>6c6DV)~8QYIuozQr$4W)D|xN!L#}s-oyCt9cZF_uyIT9*_U9lA;F~H
zl0d1mEL0HG&$XCgL1y7Y!Y6_?9CFAcsU27&w2=yty)bN~)OgVX#vSh|lshO05MD4y
zGR`Q57%S2Tu>Z)=6BXfvdSUwBCqu%7amI|Flpd;I2Z9Ge#E`M0D6|)=+DJP|Jz{}E
zLK&S-W2dE|>H!~~8E5qjb?^=^5-h;r8HNz8g@x|D6uw-S>vCPL%S*~vX88IKKm6gR
z7`~nn<(HO3CqG_3{V7V~DE}JAA2#Xb_k8;CDMgikYTsT44+M|TfC$|5`P=osZoh(u
z=oc_fUYO@mt(M~3-+Copxwj&|H%#9L&p`%4HgN^%c|C^=gts#XvQu&jQhnmT6rut!
zs*q9#<pdE0q}{>@18=i(31vBQk9ZvW&S0n^O%&-?9Qj~{5)oY}in8L3L};D55qRbz
zLlM1%Qk0a5xcHp5+BV=!R?>c1r1it#k@h81Js86#h&E$Smb4cYi+}b-#Zwng{iNL1
zfnfz0@E+Q01)SR2N&DciXB35V$PLJsz>@_#(FQ0?fi+H6m&hiPRe+x)vLFRv8b86m
z-%%Z;RYaINkImJ0;?3XwT^t|n*;AMOPxOG=Fk#pMhbK=yj9>ryAFV(ql@rQBl#j@2
z7#lks_N*X1RDv0DB~DanKUL~rpHNc$bW}%IZO8k&XBihIu@!UbKkfD2rg|RZm1R=_
z4JC=kMF%MqE!*l|`cFM<TEF__VXUt|welRVV#dK~6N?xNF(wj0b{0E3n^w3jE-c5b
z+xIk16M9#Y+mHcC#dsoF4`UmOexlvjBSixvxLw$2gM5SnioKvPBoKWCIL|sQwf$W0
z#t4L=F7@9XV?2!5BhhIe#)J3%QSZKrrR6)`SBf-yC@dWnC^!)@0si2*u2&1r9l<80
zQL_?R@)xotGQXq#HKym?Q%BTM2MXeRz8q;U8QRd-d+WlTn&5gWxsCmldcqGK7}(e+
z%HHzHC}2grntP#lVl1>#5BXE?(#pX~y_8m)Ihz!#z**?4TY9>O=)H23%A|=(kz#n(
z=L?xA<TV#APaPRa+QT+cTMfL~j|4aM%gRfdjbdD#*Epn4;K#ZK`4b}@*O|*b8@-rE
zD049ajgFk_I|hPN;TKMTNx8-RAv%J+PLPFBz+Z|M9I<ar?Z{({0Scw(FaPns#ryBS
z9qo1_3OS+z7~{5R3wt?rEN^ERXVs5^63=yv%(N{T2WT@2=C*M)lbwj^+2trt-PHV=
z@%iJ!ZQ)Z-YcC@81mDpM;WNgniy%EC#`BaelRmRv>}VWYZZpPC&ZWX9%EDYMF5eV>
z&p0BEz2nNog7>W|m1mt=5(PJg9{PjR9_ivpnVF0-D4RKk1=#DrNO&RKY@awa;EBe{
zk#v5(q_)hcKE_$kGEY2zv<4D9aLUUP1!v=o)(vM3-7Bi5CJXu+3(*P`uX+aDZ?&7k
zQzAo}o^MPa<{Z%zL>gdNCSnC;zq;anfaA#s%#q;E`xD-2ZKJlc$DJJnY;fTke1-K*
z#B{Y1w{G6mTwD>I$wlkrP;;rS@1>ZkE)I-dJjZ&Hbev>%8AOn&p1#`ItRGp4@2Cf0
z4P8(_(Ffq9N<^29OjkrKYCWlEtb^CQN$sXh829y^lz_}d46f>X0mX8;qI*ThL~l{Z
zce|t_6|G@^A>*5>7ENbAvkTE<=3~AfTCKU6=FPZh#9{33J`wEB9f9O%O!$KEp~sx=
z)Q$h7?*%vDAB;cgQk{m7!LXq&(Uf7EfqDP3CEDNgcbpy=7Yv9PSKxn+-ePPt-p~gl
z0QGRA=oiL`w1?!0#)Nw={kwj+8%G#LM4RD*&|8na9wnLtet}}#DKND*_VBBJ%n?$_
z){l?OC%^;J%{9qhqxO(`7aok!soi2e_k$=YqG=|H0uyYZ2h4Zq6g2<dE3e1xJFkk~
zOfztLR<ulTI1`O<l<+uoIi=wpI>Hi3goDIHaawC|A`XR<>KibJ-y1fjdB750!2x`M
zX_^m2aIrnXg?JVB+Mz|)kx57~NVFBkE8u{>0R9CZhDKrV!a#;0c3ik^L!Ib&SA2rB
zbI45aU8nCw#u6PRY7BV*V+}Hwbqtb|NdH^W93;{a1I*&mO5D)AnVVm*-s$0o4`Oxg
zNz5(Gissz3OoEQX>FsKg$t2I<h>L>?21{UjcB(drt^&WQ$&#L1j@!3y#*LMgSYBL;
znVG6-HF9F7Eg3?e@MYu<WG3JSy;PU1pMbVN`!RYUGXgL8KtuGM)T`jpmF7oRbo4~$
z8nrzeNJ(Lh98|9p-S;ql`|E!aeco5?j0eV@lIX*N<{|SYe2s|E$B!Pxn{WME^r|+{
z<q*jl@!*5|@xg-+gd*(YR<S|r-S_UtqsObUiGFQw*HKYNYQu58A%049FtT3JvJ3kg
zl6o4NhnxgnY3S6q<P=6>T)=G(@FqnX7TgYSq23S=P~8}~yuZ7iB~9h--kv`7PK|pK
zZ<3sD9g@ZwewFZ0z1Iiw9Osc|(038lh+&whF^sFz)oIITIE)<~l1BQ#d>Or#_l=#d
zidJVlPL3i}%@pPdxJ&dJ_zi!sLxbi8ZAHc~?m=tR5Oh()KBA8@|KWbY4VVkzTrYAX
z{Yu}{v(Jie(SL(a3|b;s!wb*amMWwS>wkVv`i*|}`L%R?a8333ZvDr%^edI&Z{Mds
zZhEnfj~}VdbnDBm505{n{iT2WOE>!FFH~N9*GGT9czhWJ>UVAUx-Qq{PpI?>I~Q{%
z4M<+ZIU`}r3L6ngf9ep0Fp*{f%Xt=JUI>Uli+q^CBb>3wB?TgaGt7y5@j!vF!T=F+
zU{)}DhXLy#S1!bxx~K!s2!HwzgA+%vvPp|{>K|8IvItXpFGAvzhmT`>eJvVd%G`qo
zVRkGlMpKs1cQ1bO<DbO0zx&;oouAi&5Jiyc<=v*I>d)Xa5yHe6ptJOASW>FPbjFN_
zd->COGamk<pWHBDX2bifUW1Y8BVF;C^o$>Pv2gF89^c@b{6C*GgQsnd;-7DqAS*#)
zR}1s?-OboNJczay{`7-|reUT9RARaqlkh}E87Bs-dsuLw%t_v*jBR>9${Z9)9^>@7
z^g_4fdCRB=1ro>5(*^H5yb2L$fbXTGJPcLjlxHsvr<V|JpiJ_3sGSpe3B?G!1wEr{
zR*>i&7}$7Q5^$2C70)abd^xq9{#e*mKQXp)4}Br>i^V73N!>7(VV{)qxahJtzAo40
zx?Gn(C|_N~*AKtQ9$$&(zE~1J_%|x>vM*AG_YZi-=PrJ1$X~zUJ3sMNy5MVr8>m-b
zS#<ghYNHKR{p1ZFBNwr}G85nU#w#&1Q?YUlUUMN{W<D)G#gp(alvpTRY=lsqX|wCZ
zbCDTRv=H9E7%nhcbgkIJ%aqN^hdhncO+-?mv_XNwLfQ%g@nVc$@Mx5?7z@lp^&E<F
z98fUmP&dXX_%E^>&*F7_${rM2#N*A1LN&=b@OqR-1@U6i5+Um$w_)TdYJrcEYC5u{
z-BPEJ`|wm}A1QTJawAb@dEdjHLMUESdLtAMC^*`&v$bYV^kFI+?^7iCfHu%>ls>&K
z-oiv0jO+Yp^w?kNqT^Io>VcJ8lq1PE4?n!0B10q-;vJ0A3&j>5*X>TzkzT;HI$aI#
zcc#`LF=1$7Ula^fz>mFo)FZ<XEVE)@F!mQHqEOc8H{-#*q#2y9%t)3%aXuPnf^`dH
zqS}%6mB4d7$vyPb5fA!pCnFF-;{wbv^sui+yGay9Y8{bl)0L`au|}g72YcJGxw#(e
zYfts;Xsj&Xa{3{z?b;Yt(|HtoS;@2$HcXNp5u+E%8q2^r_Ir^$pCx^e<l%CP=1bWN
zl7e(=YbTCsb;;nmS9mniKNp-D!sCv{LAU80#|NwN=)<>TcBUGO%XhU#AjM)z`-ox?
z;~)y5VtH0g#bAJA84)yTdZGLy1*9F41jnJ$6onosm~dzk7Ft%-i}ek|BSw}zM@X^H
z+*uo=ZAX_$MJIXDQ3oi<DL6!Rbeu)Au^;c<e>-;fc5E<0;fjLG%3GAL!Z+ZJfsL`D
z4=B!&M=R49eMvh;`)ECiYEm+u+sL)Izvq6U5Tic_`}<zcAXh^`ShrArN;im%ncxfz
zlWIh5AH#cG^Ht-+I;bUl+}qvKI%dy`Z(zjys?02DysD}eIhfRE>|0fikumlw%jo=7
zOq8Z#cHxHbzZ8v=R@56Ox}I@_5vk|$SyIo1jf*J4ncpRi9ra+Sz|j$-GjL1MMd~w3
ze&*m9;{?orGlq(cjVs%mYl7LO=M@fR%-y~H9kpjf_^Eelj(dG%_^~cHNy=izPV>pM
zN_D4*Ef5T4d?qGDlRC|#c)Gf3CE<;mHw7mQDbO{P|0h~!@Vv%OFiO#Uv>DiP4BGEc
z9aqdp(N?X4I!;%n{#$7|sqy9gL)_`c)g>wQ*u$se{f;r(j0jfLb9LE{{rz=~Ge+D?
zE9Wr)IL*3VcO?9L`0xYG)iu#C(xI`Qfo2n>o1z~`kLTz@%@YhGr|o0aajNkUCJNT6
zfG^+(@5LwvkHB#hu7s0?K7=#S+A`4*q`}M*iNj~shFxe$X>ulV6QoeXzy&;#L7Zrq
z7=ea^+nRT+BjHTV(5hDFJa<QPI0%ec8BA(Dr&`op=KWnO?bE)Fq*G1jsecG_6xsTw
z@q+GU3+yq({&d;s2?wBKq$`8Y;T%cd&d)S1&ARZjVWTJf0KUi2TIV3#Ai7WkFZ^f!
zoN2+o?+857kYd=vsLlR|M|OmOCz1}b=?FCpSfI7h(d%~#Kd4)etIgi;5W~WN?y!9V
zeMCQ?38Lq~XAFSg6nzxUuW!cs<9B0gbJch@U8z{v4{v_^t+&jRh-5-ZPk%5hqDaRu
zOjI)2Qcg7=iDn%Wofz{t>vifsM)advEgJO`>tw(=<|%L^N(LIuq{g@e4<d4i$R%)q
z$n?=s(SMB-k<Q=z_IKju?YnLpe2x8{yoaH>0^K7jJu7+)fRGtjpYy()=tfR70>iiG
zfv(#qFZ^bWjZ>_nZ^6+%xPn6g`zEEdv%n6T%O}xjHcm0F`h@;NJ23o{0vFztcqIF<
zGnN>(Z~!3ftayrb2or`4Ml%eFeT^RxMnum6M@EwQm>g4*4uSo^*dL8lyqS#npU!9V
zY2U+{*~O^LE{Km!*}!%G{kLrh2DZQ*zKkq_E(6$M%z@8hJhl!+bRLvPPlilz<HoJH
zB{&mV_sXkp_;YD##q)~iCq;8`Hei-GHKWFjzT!MHHAU1PG78!(;hvovlH<|2XcxwG
zaEU$P(w^u<Qa7VkKaK<O$0n&Zg~Q+j`>=vvL|ZX7z;}9PL3NU*a3prOx8v~eIR5dU
zeii@M|M$P>drjjxCOm`Rp2pL)wRrk;O?(7{a!+`?7pqU6#_=)ECq3an-|_|WYfj@<
zou08w0!?MC&$^_BJhg$TUdK6v=tCS2Fa{;r2cE})42@ir93ifvqxX|y650t&aF#ZH
z3a^PGLxxS!V;E$M9&2PrU`ajP$Ft+`ctcts5KbV60S|OX$c&_d#OQ@_5;+y4CdNE;
zSBwqkXVfQjQt%AsG|wUncF!=(lGOoYm)b9!#_0gOflQyBCV3u%sC8Ou4|9#{IDS{F
zRkdGqO!!agV9RH^##qB2EDLIWv4&ut@hs;(r`%uhJ^fifNB?Mt`p9#szZ}p2&ZY0P
zYjD|xCrt?l=lJfg+{>rG`r*MZ6}}FC4laG_F}#@{pIUtLrxxn+pX&Ow<Fm@ko*5o5
zyZ%KB9Oc^ZbzQE@pIGUO`H`MLR8@i&Mgu%svu0=}1UewV_B0+aE0`cm3_{DMnlK#-
z#5xPnS6w!c=~)PR3Q<m;A4DRtNI;0-c?2&C@52zr!UkcGHTs2^4&HUtlhgH_>XuXz
zhkJYR<l&=Of3m7YStBk^b&upR7X1kB%QtSukAM8r`2P3*A{LjH)F(pk)lSt7^FY~Q
zCIX#P8@%@82cLst9f2C+j~6eP81o@m=+jr!U$@!qR^5JBg?GSK{D6wi8zy=*^;(ZU
ztB+sYuj~5Grwetc;`9vPeVd;l6=6*a==H6Q*wiBQghh(R(wofc2C-2sf+cVX)-Z93
znYqS}+7TRJ;#|igl(yP~S;8On!{ieKW&G2^4&@sssRv;hB_3nTy^JeLKbUp$MDO^z
zQ(WpdB_mw%J|rj-bU$_5P>R5yF-Cac85WNS8g!P>b`*yw8(8=wTnvQ)ok;i4KiW(i
zRSyL`L@AlmLLE<_ljF3h_jS20*X6prP`);XuSA~o|Cd3hzGPze$JE9%f3TqV9BuL)
z|B|*3U-!}~(n(|hzNP<F^o(VLs|$PcuB^<)&07mGKUYo0CM$*EQQ|w|S(aPGjQQ;d
zEj^0?1jQ0{CPxztSQu{b^d-d`DGG7GNegE9F3O-nk%;KTTk$}5<WQbS7}5)+NC3Ko
z;*scTl*~vD=8tMaGLZG{DU0VahBSKS{lSP-X{Lw-<Tv;#Jf7#X*<wtT)W3X6fd>mk
zxt`CZ&F9B+)8e5h$oiI3az&J_7?sMt4-bA)gJlbIF;SjZo3pY0^uu`k=)Kt8erlru
zN`}f53Waeiqlw_ap$Wx%@9e}1NDOd9M|hL_J{yyhq6~!uDHT!X<YJ<biH>Bwt<6pM
z3q`~1+(MKli$SDIvsE`-?%aJ%GT*XVdu2~?QiOFn7`=#CF9sXbm(^$51<X)X8$Yj5
z%%HqL$%{8U-pEBCPBUer7tvTfM|q%}94hrJpGu}d`9@ltVP*mFgdsv|<`o$1S|?2_
z(Rn{9pr$Gnd#0lhd%F5K9zA?N_9frqEVjI~61Q&MwRivF!M3h{;QfVg5W-kiE>>f5
zl8C$sy+ksP`h#M~>C5cxjiKaRZQ;_D#~!6H2A_k&BjMm^%+D`cnNA-tR-E@vtn@$Y
z9LCP}str)HC=6$oB%7yA_vr_9F{UUN!2#n53Txp-PBP&5Sk`hJ`=p>)N*nJJNkfW4
z$%?>Oa|b*He^KIf+G!s##()S3l+o>G&F$3$F<#@SHEQF;r8<RCMeyu4ykFPDCl74w
zA$rSuxm1LA!q-f;>~~r=B$8n)AHCCN>`P|Gv!1y||B*jw5-A(If(0J_DAyUA`Gus=
zB>mEJLo@59QOVe(@T4zl(+G15V+P7{K1D0~r>d{*IRs81vl1Ed!Tq;me|IyAg{;$2
zW;C8LHm&{<{f9C<X9agM^yLNr361Mylq-u-DBXxkbtRUUm#lpK?)UyGUVrU7>i0qv
zgtN%Uz?AWtuFS^N^n&UxyB<=N^}5HxvklKPA~2kqaU|_2lZ;+MSR0lu1>ctE9BoO8
z!Dj3lb{)+PBAbfpqaM+{M0fEhk$Obm5!pjqZ1DEJW{DoAx;Y~1N4s(Vy>~Spd!8>C
z&oLmf7I0<<^?~$(j#Lw_u;*2R33QISIQR}7)||d18k0R*1an;?vH_TQ{gsvmDjFL#
zIvL4ubWjb#U2qPgef@YR4iDFbhbd(khGWy`z#M-W&E|1zY_54vFH)YNWQ9-AR^FY~
z&%!z3@wwh{c(fBo>SyP)uJ&pE3tuqSxamVfAQ*!2STF`Q;2{Q!=lfd?KL_8%L;4q|
zNufG{A{|=IJ-}6c5^d9<>l|}a6tU0#YYn|W8R|)EH(n6VMT+nmS6f-Pk7b}I@Ex_s
zsqvVP@VT7cLH)p*qBi#G`o#F5zBol9hPTwl5z@SA{$tFWEER=2S<?;-l6}z+r>$e%
zC|oCsl=ZnIKh-zVOQP^D6xFZs3F`qcdX0^qtL-DvX{EG&M+aNpf0I3QaC|sEJ`mn_
zJdR^xvfjfBI6uP?eN!LD$KlcmS1}qfm%){k>1Bvtz)j^!Fc7cWSbHztdFR)0w6~*?
z!cauy7JG^Jyq_Y5TcWd|A&q+7FeCME2yXC83{>qFhKfCnMMrc|vOumZu#w(2C%UV)
zs@?F>$;!0xR@B3N$6yFwB!ZL3?4sxZhDMa-PM2KUSO1Tks&4tlEzMONR3rlkw`~lB
z|BD8K$65CK8ZU^)2&S52PHmc-@ZN&-CmFrye^zynHn0s~1S`0z*58s<;3H?<rsind
z@(+9(!!)`D<P;2DHf)Ip4UNF)VW1(%MNCBY9we%qNF4M77@TM`&%jGD{J~!^;><4J
z628s|MuLOpHF5)yQkUYTPVJ}fL<J(#jH_>a#{~m<$nO#lot&B#>6h%1RbK_;!^17r
zb0Qo)jgxxQhQwR9ZpF&-O5D5mO5C}9CvMz?2j7mx#YK-VbCZ5d5e+}hepoXxGq)sJ
z@fFeGdoic)oLgiZR#r4lMZ*NW2u=p*N~WhpGc`V>n#H(E+D#0p%t6wOjs=cR;%n*?
z4!wQ#1UZHM3gKgP34ABzCbGbSaO~dQS8QNGKI!!|za?9Y>dC#Gop}6YRs69Q4<D|^
zhmRzewHdoh8>Y@gUz=^Bw0m}<=$>^Ymz0cS7}Ls<eViSDG1XYuV5#}aIGsqY+}+)i
zOtvlg<Tx5N91Sqibwy`-qN}8oPWz(5GcEHnIn*EcC3F?z5B1G2%nNq&`kstl&{NV9
zVtkyPnbG}{Wd;UXWN)5FZ^X0KBPBwrnn(N`u)tVJ8`*mfLlco#7}gjMvMvyH58f3D
zIU6duztwEn;74joj8Mp`ycaztGFi%o#hM{G<s-9Tu%&Z2$Js~?4Wo@BD9D8q6M2o(
z!0@ZKdA~TPdnD6B$`D|K-vCG8GmJ4jglK#|fgAcw8^iRSx#)vVL$Zdy6Bzu-Z!Ypl
ziz%t)sl50-71ite)AL`r_;sJ}!StR_K8AOt<CDuLKl|C`WzT)|+-Dtd>TAQ-b-6Bo
zf~D6H0dKX$Cj0h!V^NL}Xhjc)4GjZBmHuGH7h)JN9`<cAG1h$r6gDFCFboJkqy#Ti
z7BC3LEaninVL~aQEO{v;LZ1i;lsGAJ^d`MWC!?TeifWUj<2c;kkNa=G6AwRl5bZ`Y
z?On{FcOks7q3d_Q_r3V}&wegJ<hB_R?MDbT)8$6qqc2x_7*bD5-C8p=eIVrd>!3|s
z)hCzvIe46Zx>xsbEBEVwt&^7IivH*iIhs3@K+Rk9M+H}dH*;El+-SkuFHY_8HD2mY
zruzF@*l~xJg;Ccc|0zbVz3r&A>fZMX0}^9JJ-i3_;?ZSqbk$|pGB$&Dr|@dMLgUeu
z(1hTE0E)s6xb;xv={nwOB{6P$z3Vy=b_j8dB}|?Dx@-*7c;iWDPf8Y)syrFkQ0OH^
zpuKa{UvP+pEn(sq+ZgX8{7TT!_|w1ycP&OGq@&nFP@z5cQo|cb*J!6<rMB4%mj#&@
zHR_{<kG!9$9oOUH3}wW1xh~h`x;!gipNKCJDv_!e3RL=!C?Bim->#V7y-<eEE?;(`
z$9~h`b$-i7-t}Tzo?p>9$skq;W=G<C-}^@V)epZL^K;X>=|acA8(a%)7O3!43`yQ|
zRE6J8eW-fOr|?|XA{qGzB~IeUY2$hnLMVbzn%OH@2g-aLcAP$$y7V3t0GvlzkD`dE
z9*kgILkZw+=s#eIv4JQpJgiZ$!7oubqC9cRp7yB~4;&aHXpi@lvN1x3o*5mLJg4{L
zS&fovY&@sxaePt#`{%k3IgiKzE$T;dR;=XnoJ)#I`i?T{pa1^<jQ`_*|9{1Y@4pud
zb8~U;-s>@4o%PssPTM+2Thxw5y=DajUft{?lrJv|UX$u~A;`#<k!*;e!%BG7KQ6cx
za-;Ud%?XY-QARDVxQ(PEBBH!pnu?W`+rj4Sj2T7{3`HomNx?*FD7?ZM2b4-Etx+Z-
zpCG%i9|rdV!{P*OK}mulTd*5Mw$ld`Z9Id~iajiF=1B@Z<V@O)0!w3~I+1Cp6NP8Z
z-s(rj55^hAASrf`pJ@Nnr;ojl4~oE>H*d!6J9lGlcEMx3v%MZ0>#M@A%UE8%;r&W7
zD5$|(l<gW@BB_vbF?=z$2$bxF!^ZoaHaOWAJdT`l5hKpGzWv=;Sh^LN+<c6&Usk4P
zMe^#C4`XL{Q|-Tu+1dG+U%aV0XEY8O!GRQdbwt^i6ikP>PgFgMJ7(1gN<7J5={UCn
z8xQ<sup)&Oa-d|hKF&;%IZscLF|m6dINr_7EyV2HN|3pMJ$pN+UE@62Es>M)z*jeo
z=bY*azwwGk0X8{R5f0o^zgG0VvaTlF94Zx4WX-tx))ii#X^h7_9_(d8>OU0!op#fT
z>qg_)hA-x&jjW}zmDPMwz=LddA}ZA-RnNXqLxTX56P1WQHGg`ic+$6YZ`uPFCHB7e
zOZm`V{j;;vv9P%8Xg0>0eVH)wn4W2lrzi{RV|9U1gF??SK3>I?;5$9H6vfG<$QOmT
zW5Nf$6TAZ_i9E>*@BO?DBO{v9bsML_%L&b6<ny%e(4gH!++3dNbv?_^DAcEC<}7n>
zY-|XJY96PB1;JTxWu1p1jP#W3GZVr|3_hgWV?JOcPDTx2ME#_r!#H>m6QZfizqR$X
zc=F_lWqam8S$I4(Rn^*p{qfkhtP|&&Pc|$+Gok?#jE&R@{z02tHUfY<C>wDW0Jp$f
zi~yMoDL9F06RpwQ%LtiRi*R5JVB8N5c1`OS7nh=3UC}$|)c%6TM!2G}#5m60G#H&Q
zxI;5YYs)+&#Utw&3iN>WFKGn90pNpj86|3#z0iP_+OKJ>lG<;hk?JQ6CWgE;uQ<3@
zV{$3jjp{qN#TaYOjAaE6aO8@*)z8z0uiH67b2Qa^u6d*NkoXE|aJ}~-D5_`Ji;HMi
zoIg@Z$@D(@iy_+k4GlatNr`4Wx$3E}hnh<mjZnz*++e=x#xc!tB4g&}mjvetk1=aF
z_EID5ZeempI9t^kGB4QmH7;tCV4TKPbzywd7_sjc=|VB|c%LcNzq6(J^Z2o60sN)k
z9}^8wdpcdg7h@}qBU-PjZoaFNdW^e)p}y)I^pkhe+P3IV4?1Hdv-$^MfQ@Lx#o0-m
z)DPo?sA1tDk(oqflA4W368P@zJFms-Z+urg>4y4Kh$pL0G`H(9Cmxuj5)|h>M=P;E
zAI7apWj@9W;#H&w6+TWBaxpVI7qg3^ec~-y&E2ke4pBYup);+MMZ%*zHRGtzarF;6
zh=HDcLWy9hh_0gpAxejc8E6fLGxjU8@l9<@22$v7qS44Ag$c<4>~RRLXguA2a7ObV
zLo!k5@D%p+8j&1;!!yxftkp42rD$Jt7AYDnrBl}YQr)D-#farSzmRE!ufzSnFd`8J
z!yKI=;&r+zd}h2Sz##K?N0I3_jyTgZ(@`lGH6OE@V?@B6I--zNdFU?SCm2yNb`n_y
zpT?mZXPvz0iSbj{;cXM5BN(gEJ)k$Z_v$P0-S7QH+<oP>xN$S_klT0fIht!pvPNEY
z&?jUq@5QRVvbP@kt0Y&DewR6NB_1&;x=S>Ah*pXg32&ULQSuL|0Wl=$PGJUVqU(<Q
z?Y1>%g(E~k;urzlz^MTG>Aq<^;XOnUB0GTmz_2<!A2XuOIKW`cIy!C&E*N8X<K1`P
zkG=f^!2yF3hT4L7Ic=MAloQ4vj6Lw!toWa!+cgHFKIV%=M)8TPi!6%Vi42xfK-T;$
zw2t$@8p9mXv&gv6ahwl$9zKSQ1W#ih&js<5Tet404Kt#n7^<cXKj<e>W2E`C(N8o3
zm>sdt7c^A!!DB31?QK<r*TB&;8(dTyXXh|TPFW^kuQiN>INwkgv=?I)QJFTNjS+#y
z-eed|GkULZNwO(pg|5r{erXKQYg+c!ISLtskYkB*!wCqPa;931X~~{WW2y$R2EfV5
zdMot>81OuDfKwo1{1V=${m(eiqu~sYbaS|_BSR)#9#fL_fi8ell6`my=c$+bQyF~Y
zJb&EabI<Siq%y4K#s4X;e{g)B^7-!iJcXLRcx_*_y~Fb2bD#GB8?Fss*X6qWiI&SA
zf>kFOtu&!wlx%#;!q9Y91PG%Q%A91JK;fc^j{qZKk;9Qjx(>1SMw@yT?_vl9OwZmK
zI_LZRi3bu3A{L@ziFy%6854W8Nu(t*sy8pjG%kUnws#Qk{r;_3d-OQkwWd=wUUU&+
zuDqG>y;ok1U;N@1@spqa)Jg+V`#8tBm<<a6+U7_vedmwViS)-r%us+WKPR=EefAq5
za9;w3PbOm)7rCTs{4{6cN#8L1=KK78@Wb~ah`TQ8e&O)Vx`(QC02fYiGw{Jc+17$*
zV`nQqc=R9+*xLZ(qTmXQQ?!?!5gZvOdk$$Fl1FQT0SuTou>mZCA0E`CpCJ#qnY-S@
zd4z_Nm@}R#JjZ<q8!UwIVq(8FnucMG&EgTXhlWgGZf`KXmzNQ-g@+OM;W14JIRXem
z4EHcr2reu*@xrBjv>zeN#xvocH?@_p%1aFwT{n#M8ysU%?8PEatG(1geF!b~K$bw&
zzu@e3xh~h`y0`$#Ks3L6g?oH`lHu!RGzrRT!vDpJ=;BvK;h@2T;&1#O9MAdcS64pk
zeY6Q)VctN42kjTD6u*kNu{;ys`OYiedq!u)*Sfk#i|a8w-cjx)S%PaUz~N)Om4z`2
zT6iCea}*Wu!~}uVD(*osf=4cV*a{T!GL$=<hevTQpD0sN*g4(VL`e(zs?KN4pW(qg
zk0A_&D26M%_env6mo9R}?)JL)YbT~AryS8kdLBH(={KIr7}C<7Dv}j0Pt_*Ut%-jY
zrUN4g{p9@lSv$#j>=R?7!4Sg8iI;+n4ht${(K|hkH~;y+#IOGG@4VmYYp=c@3k%EQ
z$$7&sdG}_lR3IG=Y1w9~%aNa`N_Lr29rPciuRbo$B-bSe7-s<($y!M|KacnCn2lcS
zU4`KczF*KM3V!xrz$jU-9a~X}p@v8T?>!>fA~|LC9T~8-wI-ZJzDW@)?G_4=u9c;1
z^v@WeL<4RpUf5_I1r`w=<;g*MJi!Th32$E{Gd#}$k?~BGlA;JD{J}x;$Y<XQqMAsP
zh$0Jl5p*IlYFF|cDW5QeVZ6fwo$=?~_U4AhbjJ$(Y4v+ykv)FW`%yf1CA(p`1HLGg
ziI`wtv!lbEc=Y&TTxyswy6*1n+5>v!#;sUdx~=;d*Q89ox@g4xcYhnd`pqw`P=Do>
z*J9=7t1&&Z64`uJaG8+o)^-#YFr}Rs7jTY3FOEWszJoVJ+MP6xqftMy@gb$Q()ijq
zgusTf{WMygW*pb|1UqE-s=fOsN=q8!s${@qSg<lw_-h56CNJLTnVj%^6ooMnc6i{=
z3fIeqdoD|q9uY9$EAInmJ=xQ{RmL+TT37H*a_;$=m8ivuisl8L-+AK$hFr#;_kw3h
zw!}a#xmpKCuxfQ7a-})d0}K=Xrgt(oc^Agd3*(%P2O4blTWQoAP6vey%O1F;vT#r`
zJZTS6DC1DaSVBM8=V){kywY*WK3qie5S4c9adIXLE9+74pGWNg!<1mExd~13{8K%B
z(W5bFhQKj4($(}mrFK25gZE+ZuyJqzFASI%Sb!tx@<v3f;#}*5vzXJozIp3atqY(_
zC<RlhIa2xwIHsGV<-{0<(jGX_PLz7C{+VG^(9$@s+`Jw0vsKNJW5Wt#>CE(k<}`ay
zi5BW!@1>`2DC&tqNKq+Nh5-#@D{u$5E^KH57pMn^M`*#;)@mFcY(`FFKT(2CPUwE&
zrRo~d93t9uJWFcAw5K#tl~dD;!Xx(0Vto6eCkLuy*UEjOEz9L8<A!=>oMF90sx0=U
zVSgX`%f4W&nP@36K#9yXjDlBv@et97t4rpP*1f>T>NKLMnsKRiq^KoTe#t&_2Zr&L
zpTBb3GkFyHL|R2O>ua&Kem~B89gjWi4a3DKbqha6$0zhmNw_s`L*hx}AkI3+QLF9R
zPy;Sta2eJ6uAD;jD2N(ne!?R$<P*7s5gCKo<%MXD)@9xk%Hsjf(+7<?YuT|0wS7v5
zJ~f8WBpfTUc?|R5R7q%?x3V05l#J8#|HU$xQ$(&Y=f{aA!XQ$Jg{9@F&Ma!~&O~)~
zIp$`UG%hnfij^6mb2&Z7;9TKc*->36lNrxsWCM;Dnny$h4^!~D3&Mwf??iKT-{~Rq
zqANFV+>OPRTWUi#NUz)LwLQO<mtKkT?A^!|M8AX^+ncL?&XFet!9~F9U34uQz*j2M
zb9#>GFd{jDoA|i+8tLg`R0pu?V!T#+)xH$zEuN#JI5j6SN?OZ|a0esFuJG#0Q8Tl1
z;-``;61^86({u14)6HT<xI*1@j&=2=@e_XM6qYtxGJmv=*L*U5lIBtC>1-aEfM<al
z^9+Ypt=XYb&>h`Ftr_v8v1~=WfPEW<XM#UG2BUt$V`xlH=P`N}?a+e55%UAOj)4jT
zTVZlea)D^LDn$+;8r-Q@HJ52O#x(X7go2<i>UGb=qYv!p0q)T+<TPXnI~w5_<UJhe
zr$$}0d}d}o3fY2sav5{;({bnCD={;FBZ}o|8@aN2hG=;Dj?o+b2Y;9(8ea64EHgOr
zz|V0`c0`wrjF;{kMpbB^=epWLq%5)<Mkn+g=X$Q&If?z<_4wfZw_<(miROG)vdcld
zfB!8TV{!6nwL6ZiYIoRDB1Bu-vAwk$YisMVvALypUU@EHfV{jM(Y?p<;KQ|ey0&Qt
z1>}^;%9Q$PG}QYi9ohEED|Z|*j6gghIfJpph}&qM#Ky+jV1KVYrwSwu=L!38Nv=vM
z0?(2RfU!o5l*qH{OeN;#NP#F=Yn(6$aUdThr>~saQjpxSB>Cf}cs%Ky*_W=Vwh5M!
zVYVgXtgf!vXorkgbNbJ^^;e_}M~6dXq;*t!iua@pKpJn2>*OT+n^lbyI9D;(aiVbg
zX`ZpXCHk*+V5B1Q3R*~;nPZF>#xES$kilkWk(EoS?V8sd@D+?%IF+?EHprXQS0*i~
z>OoIMdywOidoj3K$A;sSp24_>4wZR12H({^$dnkqi9X|#C^4tk)HUk~)9C4$<QUFe
zvn~%D*LMnQ2xM%3e|Gls6&q}sY=h@`o}WJTLZR=Ub_{O%*!j<MyjWg#{bk2zm6ts;
zJid$qw66_c*X6qW375-m+VmUc00K1&TN6?-M)7GeDg;mrDG0UtC&HWd@gfq;m-J76
zlD2L95OLwVjT#a@hz{Z~F^5sXwDg@%7?q2jfidxXM)zP)yXsxW{^m}+|IWMd;d}Sv
zq{f0#_oxxvM>zC1zVWU2=}&(iKl;&+;_khBTG)sIiLpA;OMOU$UCaW42csXf`#JqX
z5ain6n=TBmSvQvc<Fc>&i?8#aeyDo=OFwm_XZgZ+J;D11=U8m=zYS+Zhxx2NxmW-3
z`Rq;!l)gjn=RTm+6{FqS-;IY)AH~z{^*Gb-OfGrZn9;&i`zg{2#>gVwX>MSc>T^a6
z8D5IlA>%{aP~e?N&|>@<FJP9&2ZNKwT@R4s8!sLL(DcQ>Nv#O1c(=z`%$hMl0n{eM
z9ehFxS0Pl}QFlB~m-L4x4DVx+&)BeVL{WgZ4-sf=nor~yMl<@rGrZ3cXTlkr91VL3
zKL~hu=(u`7tA0^8LT(OWM9(?ZhU!FLcU`W_b-6AoU!CFWuYUL`ax#=_SYQGJM)1*M
zUi3961D^DWB{94&t$b#$x#8I&@#Vz}9`l*OZym@IoL~X}op0TXJ9ic%FMbODKrSK8
zjCm(Ajs7q;z#HK|=RFh~cvgp=Py0u~d*NS155V8J#zGffh}Y_g7S^;8ewiFoFodOu
zXcp0q43X?Ll$$0>Wi5zN{D@C$!Ht5zGMeNur_3u(#w3vuj@HVW_p*Nw3L+v^3KM0$
z1A|%1ZJe83jEwH3Vck<w!5n!pjzVrcJE^vz*h*0c-n@T|2s4z#X;F<*u77c&XGx`o
z@(lT4G-?M=;<vy02lLqf@S`8cx4->8D~;HjgNQK{o+yKoL8~A0^DC0w=92*q<(Wt0
zJkHw3!5%P<<kq`^4TdiiUxgxjm`w!f>}bygDJ1oN6j(ftcRWTflu;-%yf;LN^fNh+
z7ltd8Y$?@EN=HSc1`b+`1qb&!N=7hee;bt8z#ZiWQ69snKgI)jrl9jUg^^;bagvlY
zPK~B(Esgd5{(kZzJ|%iWcp=#d#X9?g3B;DyntBKOf3=z&%U-+-Z}Ov~T5NA`dtWI$
zp;3Ziys%fc@B^=Kl-%@(P55~hL#MhVT+=y}qG|{Gku1zFSy2UC%9B%8kRsnFW%E_^
zI)|~f@z@H)nb~T*{`xnp_(yq}3?F##k48?zjZDZ?YN*|(!W|TuC<)m^3MD8;BNSyg
zIpNhzd#va|_SbWlg6o;&!xWX+(3lpZSei1fjyqB-;nesT>GM!3>O0DA;V%Z`@n;b~
z!b#x{(f*pU8ebyY(tcp1Q*^{45jBA!mA!#fR4;oiowkV-WUrV@E4Rw!ic?~ZW+x;g
zvt~ePrn#8USrLXji<1`nt}vdA@6_as4L2!5h84i@pr~y+j}elH7v?5WHYi1_GxLt>
zyLI!Hm1|@!pneP=DBts>Bg?S&(zy3=+TU4=wZ~ffG-_%)Q2{m0+nQwR@E*C4JyOBd
zkN4y7Xy3{+D|Q7>(x>7bPjr`~jK)!%o$HyDs+?4k7)_Dei9{iK!|)&Acv^Mn1<d4;
zi@=bHks+I#Rr`p7zz`{Xh@RI3tS^xHSxYeIylxb&0q>F1Qz}-*j!$TcU^HG5&P_+D
zkQI)fID-UxhZQGCzu0$_6(05&`-$YS!B}I2@x6c1(m^U$>S7H<)Pdfh_t+o}Y(*E2
z1se<rL|hQjnf7#J+<3oQg5#i`B|RuK3<E0kJgpPy|7p|@w>+Ooort2Ic}g@4Y5Twx
z<^^jh6t=t*xOvSvDDWQ}%Y~cp7wA1ap{4oT6aCQ2No~-L!p~;?SiGVhP0b@BjyBdF
zd7s1Goo&&uGt(g}^c|^3U;44L`9VB>__me7q<zyHJ;u<Ow3PsSME7d!QLyjtZF_z(
zCeS@-Ic>Vqff64j@Q~lI;%47NjL=@sYTjT38ylxi<|Az&lCtUfrO9I0C1baa5%7aK
zsCs}sK=(7k9Y^2|`yqaMN%}<a4x<f5m>~`V%Ms%-%4VW&q1DAwHFCwd$W6?O-m~8%
zG(q+I-S7_D!9V(;diZVJqrQ~3wLn@o<}$_^7M9{W;43sQuQ`d4b8;}JFm__BE{eBI
zRpw%HdN~=b0s~J!wl<##?(JAwSvC$cw=kB#t8CDQ25DYGd%Z^%>F(S|+9cVcr#ac!
z7u`A2`*8}@*boh!!&pRgDX_YVPV3P7LJE__PZuNy%y@oa^g?C;FO7raxhRxp^gj5w
z-i1+{wL83(b2fVE{Z~ZvihpX&o%ZS+ZaXll_fr>eW6oiC1x&!0SLht$4IOl9P3}#7
zNe8$$9FG)n<uP|ZE=9XA&O#efboYqU;*w$-{gECcS{RtAJs7$2@CzHAH4oKKaDirc
zZzj<i<{7#Q3>ir;!93>uv_<{YC#eRD;wLy@Fy!z}Xf@-`_6iaOjBywZag0z)j0ZS-
z6M0<{-XUKQ{YD?56-2@_zK()Z8;~DpN7AS0wZbb>7vo&v{^&iYqDkz-g@X%58s^;7
zr>lZRPBPm?{N}fB##?W_E56)_o$dX2@4XM={{0W)!v~LIb7NcMG-87sQTnH6=drhc
z6ziKiHk3gNNvG%c2sStygL}Y?{Xd!4>uVdTcUN?ms87v3bsV_{ye;9B!F&=7p=}s`
zFq}<J5_!h{dlP!^SX3PSo}xODrI109gW&hjNO&kbo;~O=wjl=|9;%(|>z134Klk?b
zV|RB~v=f@cI1B$ohj3WpS?CxJWz6mB3<fETR`6*k3Mp+TTtDx^Sca34bwBz|9keCM
zAgR3=rIAf=zQ@qSbF`Thh~PH&FprYshNy^m4`YjrNaR?`R0N$8whKN(S)Wg25a>R<
z5849Eq4!{vuS*8?zVp<rHqj5VD$yPsc!*e|eH?tppvRbHGpU}$N0RQ;I!^H}#*E+Q
zRWyXRr4JXL!J)<z<;CJtoJ`laFa3Vu;m=$A?3YyrZTghrwtnRJDhm|u+VFK<uFIcz
zxfC&m!NBGbgkjWBPZJdVV=>O+5Au?{po!?wFWuv4Ehz1PURvNcF%p3oC~;t*j=(a3
z;w=(F#7L;$f{z#kUJwv*zF&#?9qjJMyKlZ3AH4l;G{qFTpXYfEA>)Lm|M(|Ai68y=
z$8r1i9WCyY$B8|t_~c#shwzyimk4BPLptRq`)9hBOLT>|__o0f{_`z^oBRRe9WWK{
z;@q?Q^{CHt5B=456m6+O|Ia&oUI%rhJE>3oL>5-zgLCZbb*zQZqqQfoy7AQedXbuf
zec4zvGd9EVfl0&Usgs2@UMHzj2yzlmB#^;GF+MWp2#uV#m#77uB<yfMLM`<&)GTTp
zW-cL#4p?xIK;-=wR96ZqKZ~Xo(=2{bav^vSsl}dS^jXbOQ-n*Z3!#y*;2meZQ+q`j
z2Zd`W5ZNCE!xt%s7<)=Gz77m-v{&__fbzmk&)C?ewjyv~00kD{DuNKoI*eb}kzd#4
zx?C5PugviE*FXF;!&l+~>7Wch2L=8*`1)^N)DX7?e)6wSEJyq+6nNnfh;}^D0k4K9
z;BkLrr7Auq-pgWCyu}{-Fe?>!3hf%RF-F%$$Mrioji4w+F;Chao<rek!;PPe=$y7<
zYjZ7jb~dfp&1eD3f|yOgG1kDhNRfhqorO9I1{wxGg$J|vMoAAJBnru$&_9%)>^G4k
z72Rk_GJyEo1zy9)RyYs^27e{RUc1%MB6>R5>jDK8o2WNV4z<v4SPmey+}QYw1}-(h
zi)@r)9I3EZOtWz)(4i<V>sb_Cm$9??I6nB`-MIJ4-T2YZejW?<jAx%BHd|ke($t*3
zPe_hPp6@7*O69U%LR3skRh&4YVQuY^BeGEFXCyNu;|$6rwM%Wu6)Jk~q-BQQ8OphH
z8zP=Oeh?2Hyk|q+j8kka>mBS%R?s+w-q{XPdOJE@axSt1irj;P!)U8L$ULNjM7D5*
zkKj(49SmPQ&wdxQ7g-jMUledCj&SHf<~eE9)echQk?Ls5V~BBwXfb3fa0Q$}rsg|J
zVLYL!+^H1R7nE#6<tQ>15n5Z@TUPq_E|7zI=*x@~-aA8mB2vf-U1VmBIWiAXN?aon
zjHnlkjHKv7P9#bT&+YEH<S7;K1jTZ*z9(3qRPQ<k(cIi(<T9kw$_RxphNN-GWF_sY
z-x!%_mypg%!`v`!m&OR0ET?h9lNmXYbWh9^_WlCbi5y80Cv2eKvEqr;lB7&So^%ue
zkv!}-cG?syI_|gAH4O|%Nm)xFDpO-|ekxe+3-=FoFNSL73~AqLan?T4Sk!}kVQhSh
z3*qo_931YtA5Lc^Ji+LMf);u}3P}{lcyN=Za%*!#ZEo2}iNXsHaZ*~@kqil4@b{!t
z-RYCUIdGTuCRrLq^H^p&3Z>bYMEO>nP<`h%ULr#m3R5wj7c9N^(uj>M|Mma(|B3(Y
zfBS!WOlN1N2PwKxQliLHZzMMp<+ZoFX&B9uDpK>D@f$`SV9+EDU`DWEEu<$MaksX<
zD!R1obwEk@Nb0-<Tj&53tQ-9va6I<hB|`^Mq4CTZ&Kl!F^MEM0i71w5)VFETlS!=$
zNOOr2_{v5w+Cpl(6rscz1E=JL=5%Qz-MIhWTc($lsVO}>YWyHAX|0Yi96E_}f#%7$
z#&JyRf;0A}LqYnSPUtoj)&WGQwSqK<%wbX|5~*b+p{_$;!C#EX;3aFZbHS3RE!Hq7
z=y71^U$Sov2B>2zHE9F110zvWbd_i~B4nURmng|KkD;m5IaE+mm`lK@P$(I%Jg3!m
zr$bdciP9y~h&|Feos&2g?O0!b8t=dJcD(uf-^AL}$C|&VcFZ6PN8&foMaJPuxW5s#
z!%e}xsD92z8641evpo!t8l#ahoGnEAMKg&CoSm7A>DiSiPR-k)kThl*DvuQkS@vUO
zokp9%DGUax%*HWC@EBJxU|_I0h{o|#!Mtfh8-`f{!PkIK|6F}kKd&xGiK#itJa!q*
z6WvSQHg=JfM0g`u5Umw`Qs=WLq}HyQf7F9v#73?R26x&(d-T%2?ia46Z31{dxR&N0
z5rL*Hw2lAxXTuu?Z0H<q8MH@zBT|S6BVdQV0s~04FzJ+o7+>H=?3<O|bEdYnRM)7D
z!$hlMV0J3Uajk!I7)CIhz_-+YAtiWqCcc*)SKS(WXj!|y>&Q(a%EmO`a74~#g)`!N
zA^KrB>0PI-S{xnj8&@XT_bZDrK=UZ0PK~LbEZEf-jH8(Xk*R{WqeUB@kKCKaPw=_w
zMEhhrHa8v#AI}UP_F_}xEh7;%I4ayB!uVWx${3C5H}sBu;Ls01kCH)oIIhD&ZM2z%
zf7dV>whda0(H@%1943ODeHeKLnyyNPbNtrx!bRjT(malgl_RV9%Up&I&&)2wbag?$
zv*z2R&qbS(6^>b+RQ=$K24^r<=!g2Tcd#q?VNl6OM|0~`ePd3d3$f8d&(W5Yo?2rq
zdPZa{(RJ0?CCL~wj#kF#SiqnzT%vE_E~zk)>1^!QJZQH`<)<+f+#S&`p1@q&-8+o^
zeNrM`#{S`<X!3^m!C9<r?8jSgzaLMZZtA(F=-`Ru5%EGjM|;oDuOu^}*C7HeL?ch)
z;OID->JQz_7lb#G5$#}wZcoRVWhTvCjRnT-T21t;!`#AoD;LwEyVKQa$vY*};JMj3
zAG5Qwk|X8@QBtJPEn6N!w*!1g`-f9VGJ3Js7}*f6Y|KlBoMfy*zS`N@i`6HqYH!VT
zIX$!*4BhPa)$pq%AKMV5HsF+n(F^AxqW6i$<s8r8xJ2=OU~_Y(VLL>~>3(D|`ULDT
z;v$0}m*W5iZ3efXfyji&G917L-j0I=vJri_5N>mTS44RsubrKe!VrC#<bB};bDX~Q
zx=Dx1cRRr8ymf>F9%G#;v{CJF`_%^Op$+IzDew#W&pN_J!x9F^a#U<^gs+mKm-=*{
zV23PgxrQkP=BXZ6rQcrj=!<oWgTv>C=lJpL7<@}#RnUJ9?t6an6U%4c@yX@m?{Q1g
z*Dp{$UeC+Ri}(Gp2UO(R@O53T%b!6Z?IC&d5sV<rD2I|~LNd>YzxqH(F~Jo<7a8~i
z5(SZmFrW~i9tfz3y%zy#QD6csrT~$)G0b%!$ml+OUWw7v4vymfJMYA!_dkf{Q6s(B
zB^ig<lkEFH_>1__kA56??%uUmmQyK8u=xW5rT0EJOzAB!S$(%a<a;2ogB~E|sn5gy
z@Qt?7X4l6b_xn3H@c-~x>h|M<--AyWk$<GR`9<NID)*IC8;Ul4a#?*Hwi$67Lsh-i
zil>`v@pNM?>W!L&OcY}X`=d#)(E`fCr{HOYo$A3L1|vlPLkYoR(B79ihwy{PE(#jP
ztXwR`gcz+`sONzj$`ZzoXGv2;N(hgc-iuJf`|ZUnfdS_2E{O@#28@`)e$f?h0ioK8
znX^H<A6*3=2)-<^z!?^h2sS929BPi>F9Fty4=ZYt(F+e?jA<+Y0U2Jic;j4HPyx4e
z()~<$MarmrF|R&IaOsg|F~R7%T$k(e@3wq(5nrEZ_)0YL509@=F(G}S_~)M>;S}E(
zc4f$yP_#*uiOF)#M=q<g;u-Km3=CGN>dzj;cyj8zz5K;{_|8Jvn~$p=l$jVUQ24|D
zs2+tfir>>SEv(y3@!ym>k!G-v#|T7Z2j?(4I6ZTQ)L|W`*JHm-QtF_L!6_xlA$r(6
zcsxqQsmP#sv+^-Ln;pA~(gf*%CdBJ{QLi6~A9iD=I-@ogRd?6vn=rz`YvKC^9Vyy_
z<s7Y3p9~`uSrSrQ`f<^%t37Qk_Sr0an4*yDDB{oJl~-Sjm0Nd%=o!-0O;zWkFu7=D
zKPijGM_EjxWGPwhYt;5*ck4;8pHe0}=I9|(=b*FzmdG|%hCNHQ2wWwTsPEZa#*uEM
zgG7;f+3yA($;&IZyyq82N|Zt<EaCYmcu@B6{&J~oc>-k@y=4y~WD*>9P;M|T7_u-(
zpe!QA4&pF4fLAaPR$NQbG;JFePHaR>`zocqGOoO*L`rXoLOyGG1x35}mC`2)Y@*9B
z3W1v#H<8tlzmSEHyHJp${9L}V60g7UTHLsCBeh>L&`{3H7fER}9ZoGMJVa@dC5mMn
zV--pw6!mA`55@{747G(~OcYaEz^jXv4JR1IP|Bd#N5O`H80D6g+$uyIolACdG=Olq
zsCHmfOKB5PcnaS~nFGQc4b^bp3~+a9YDPH0y%-j$UvP<(I-9=ZSclBWymPdOaEkrv
zh+@JhY2$?OKRL&#@+--*XU*ug7>84j$#Lz_GOnhR+u3Z^93h2*ixhX+g8DlSy!Ad^
zM@i@P$DnRri(wO4m$61E)$Mh3&!`oF<w-o<+2@A6r;@@7B^QwyPE{xPkRptz7t+Ns
z*2Y_u_Xzcldf*)W1HF-Vbw%_g?HQ$cNEFRLd5n>OXXocg!}@JI_f1U;jv6!4CLSE_
z#o^(uzO(lfp6atUI7#C5xQ<9R2A(M7Ynp49f*JFYXp@efvGGs!RI3$x`LoYjioO}Z
zzjY7~?!OuL-+SA5N;*;IRWeQr!~3VA32l!}CWk_reFMvm+{if+?8<WwI3d%g^l0=A
z7+k7v7&f%N1O~!SqL7Bsn2v(cc@&GR?J$;`hG-sTybd5uKJeFE(tE&LPj124-u9XY
zz~!{v(wLQFc6QNo1A2`?i>2O}Xt|9h8W+Z_d2*yVdt%&X-9x$H9l{lGl(8pjgZgoJ
zsEJ0buRV^f?RDcpL9k_SyqVcq&27>OjtPIX*63f^c!c50ifFHu`kogkr<nuHRV#o6
zmz>57#VdyEg@yUJsku*zvK#uWinchCL4EOhR<PnzZ!+!2xHny0ifnOC*HBaomVK>>
zx*hlNOuUFxePz)e;F_737VZ%3d!f0L48eusq{f>FSCs1PyQc5OWRRtOSG`k>HDd(4
zlk&TFx)*ypj{+wR`ii07Ykg2jxYQOOB2^@YLVB5>z{m<739!CN_p)}!fH_#hGX^#s
z3QtnnMxDPJjjoLw4c$YkS=D2sr`E~p+o|X<jtS?&>k-YJv<D&lLvLk&Keaza8}V+Q
z<I{~)5zg6o%_rlJQ4OJhz4O|Fg`N<tV^2yVsfb8U&L6tB-*I$eR<N3xTT#36!oRk=
zhfxZ{r^$iFojx({tyVo&A3x9->2cu&k%H?_-`DXV@n`kldvdZ5*EsKC?g&oM7U-Yo
z4f_(J^T_8X%{PheUf+DGx;x1+#`BW?lERp%=zZZ|U*{)OKm8X(&YR-Jj7{6~6x?I{
zoB~g;gAO`UQFD*<roae8ID3~tLsOg5c-hF$Z^Mwn`0~ALha324U+o@d!vT$K05NUQ
z^RAz{KvbG)GQHrMF7b@oz?fXZha?x})aSyK=38aS^98<uLjZ<u@Eg9*-j+B@XG!<y
zu_rnyMXq5~zzKy6JIpB?r$uuxe4|T9d|LCNOxoz8=3d?g&yn$x#!zyQ=pxZ#@C2e>
z&jl;8;o!vEkt_jSIP3J{cW=HE|L_0v|BCgEEgP~hR1u-|{s)iZ@sp>bd+dpY(Woyv
ztvRhRBW)*6D5R(^mn$AmWDC<$(Jp8`dwybsIzcX})toH{T7n!!+u+}9YtiZuWsPo+
zJqCfdD2C(^XwKBsWXy?X5$S|;N!rq4ET$w+O^Y{7s(uV<IPn}E9a%ow+ue0UI+0qW
zKt!KIggKF*L?}&9S40cvER(?>xL(INw|}TQxF5U$w<POGzPh*|B1*WTzJoL9vVcE^
zB+`7NcLKk_CTCC(yu6?Mndk!NRnP9;zF@lN_ERroU}L6W!kokC1x=z};J|n4HNU@1
zGCYy}zySjo_`&!=uW2uBK|aJ;3S%$zk;2hI0>b%Z#7c%ToFp*RS`Vr@55KYAP4Kon
zs5V=U(M!0GcO=~y`|%=U4YEzBX2t@87ki?SF4IwVgZ=JUqgc1f6jWb$pHng;L;9VG
zmJaofA1}Ot<M}rq>C{(Wp1t_rssW#`FyhySuj_JM{!9u?k~DLU*wJr<0HS&zo?R9U
z`otiD_W}$Fh5<oGz|OEJk|e_-7Qzi7Hqm9(FeJUh|9p<G5I8*xGZ0gNp)`(a@$}J?
zc<|P{QQJLwE<<svp2O(%t?zs{e)hAU#dp5_U3-+k1P}micu_sPH!&K0>flq~k}#!u
z_{C>>zY0QC`l7FDo9};qZ~DQ%Q~~v+yZMi2(>*Y_^o+`&uYCCSVV9r%&$|R4-%q`q
zA26u_?C5&Ec@nEzYw_XZ4`cU8j9m<o#g`dALYx>o3s@8(L$4BG{><Pd4hU1nqc+`>
zm@iC{F{I5XBLuEAmU^Cd!@$jq#fbSX=xLm_P|!6Ccsd7@rTsjQk&43D6An%q6+Eg?
zCU6d6gg&ycwa}mg1)lq_0|uXrTx|Z1at%*-2`Oqb_oJlqc<Mgd2$(D|(;kE!37ls7
zFn?VeDp7=@5@aOM^S&(mvScKLOOWZH;JYr@<+}VkEnfq}*H=`A{PqhLKlMjE%M)g2
zA1NQL>BW2gxE2n}uqHm!F|5y5#ANsr=kx#Y$BcNvEBBV-`#*Ri=H@Emsp%dRGDQ94
zb0`>-lJ}W&hYAWI?>U(lzq`8h;u?=lekVSPk_|qTA_4rTPoiPGCyz^V;9!4%JW8-;
zLksznXnIeYjaXk_8yHh?Kw%U36wRJ^Hp)~s?q=W1oMkR6hfsiJwOAg-5Y)F~fOKZr
z%t%a4SHvp|;-@UKQ3Q}irfiRCE3<9fNPSHsL7hV-7Uf(#v+4I+F=hpQTJWQsu1=R?
zVd+LpRu&|SmDH|6WC}!w5aB^R8NrQ6r-|epEf}speXPDd)ndL9Gt~tvm{3sR5szGn
zvJFM|^F4hiNJQsW!i{Gtkz<dS>XhVEJYOaINH(}+V}2A5C=PVH#0SeF7=4`9OtKJi
z1qMc^rlD5J4WtuXd%6)19z2Zgt!>LUj3ayCB472or17i8NxiOdYB?2IyV+D-J$vB-
z7nI{DiYzBd#=^javVs(Yz?**U?(aDr4+^Kn<)x@-Y#2jiE#xy4=gTWZYE7zcjASF;
z$My$*@x!=r^Ntl@1=U6L3H!ex2Z9@pEI`4o@j%9+?chO=J+)f(I6B%@Pe)?q#=WRi
z<|C7@>i0yDuI}u#5r>j3F%*GczzMmL!nmQMxsoi&ez47UihN=YIND$DP7yaivZqDF
zsQ}uwK}tv&>_0c@bW%iCpd>5^F372>+w&_an7~P*o`{S>;fWlJ%;>#=B)ekdz~Drr
zm5n3JWBt#11A|UwdO>(jn!MzV-XDxH2D-`WqTb7%O~T!5K5`ks%&5-ZOnB2DyA99=
zjAs~N*bA=SIM!IOH_4rtuFk8jNsl3W)}q{Ie6xA>9b-ORMzgjbd%NpuXFVqJLcGhi
z=M9l{HjcT7j$EGTJSidBU-#7gqfd^c5ImWKS&hrm;?4NRw|@|eH<Lk`r~<MaAommH
zwl7@D6>?FTnXzFNWiVrdadR|FQ~`z?(#(xnd5Yp18UQ_-m=LZCFIf+mw&~RZo9@x@
z^-tsYU^})qpE`Y1b!Nc|Sc+B_f*s25v-Ax606NXtuxMSVKDUJ@DCT)TvN(JMV`n0P
zjGP2>FPKD&HJ_)eb1_A_PRa2U_HUb+yBX!`vZLgj?hSe0%5fCpmo`XXFgWkE1Q%$x
zs$>5o;Sj0(P|Rb{Ly6292qP#8N@yiU;bD5Vi;H^f?>)6LpMByua-y%sZQV<>2hp6Q
zmt~Db8unWKD4slir04d9bCWT@u&iTA{g@SOQ~FuGKy<ofMLdy&+rp{+y&d6SCKeVJ
zb?-uyr{;Bi-s6fx1aK+|XJ_YcM0Mr0Sh)GB@K5wW<5yI>NUL_GHl1If42I^VRJ*K?
zpuMEzO9ljJr`B=3lQ`VpbjnA@jP#I2t%raI%4()9`%0Z(un!dxLK@Fe(%)*{seV#g
zvPT&7enc?8JQu$cU2@tx;ljwM##{J~;jnk!7oDr7eVQCqBtF<9;`YRRlJz}NFU2w`
zTdDI!x=w@*JY_&jY;@9i(?sUGXu`;NQSBkkS|;HGX&u$aE7}8|3wgQb%p=WX!IJF&
zFcK2k1s_&Db`(jx3p~8)x7;u4N#}gp@1`9XWY8(t&_#rpXqY4f<Hqy$g2y_mvCih;
zwJE)CM&p8GDn|JfWqNAdnwTVADCvJky|>ha&LJ!8?rg-~!H(WJ>iJ0oUq*G=i3TD}
z(R?BL5*#7*FZ`5f?wnu>&0JewkAs6FM}6Bt0z;GLGY%%CsD;KsyFB*HKk*zC`#71<
zCX8aVDOafa9P|mejg2B}Acv?O!cpcCN_uE4c#E-!IY;CosYsJS;8O3`xQljE7<-Q?
z-Jm6c(U`{2bK1sko)MmniUwSC^tkYgQ)(ys@g;h3A)Hp5G#;4(sS%mS>>nn0kAi=i
zZ<0fR2V+7c61WbJNYSd|>%#RyJ|h@nq@{0Ze^s1jFqn}JkTJzz$y~-^1Eb#9IBiV0
zdZ{{ElGli+C(?dXa@+RKR;;c)jo<v{ck%XH@5YnW&3L@J9`C;UK%X0eC6S&Nk~P@x
zw5iWC;Y&{P&z$)FoShdYrz$>3cL(o;pF4G*`e);!hK+eTJ0m$k@&kGkjPUJdOJhPL
zB>I!0jbEgWwJt~SXMBnidC4cxebETX#L#{8EEo-uVcCBYeGPJmrvqc8@kUO<=vA-P
zR8QaiMUHd?s$d8`@qSmThlnX;I_3<KV6|G^shrUXv9~+2G<+Qc6^6-<U_lfW5nISy
zI23_X7~0T7GCs)MYM$u*P_LBs*n(#=S1|P8XoK?-1^JlNjh4wKkU5hx6?GuPT3*uJ
zfPV50+QxXp&%q!1je}B>m65N;EguZu4NNicQNQN{PgA+-NK{3K-h(cW^T5`6Daj7Q
z%tV~`+JJeevEVvzLB0SE&vaSCL-+G8#-7yGydU|W-{1!0n6+LITBRO{-XYuTQ9qyF
zN6&@_zXxp?oO5~h<%I$=e8vGxeCwwiFF*VI@ZX@WPdh%VyzH6b@nsZ7{Mzt!U9QWY
zQ6cT&sR%qmDvQRc%9NOd1Z|xUE#5FX-~~&>5Cl9ey3@i<)LzV;eGJ4fL}P435f?y6
zO3Ww`U>4OT#>f#6^eYye7jd+I5Ni(~$HRBukAuw}y$2!Vc_~+C;`@L37xB0M>2KpN
ze()Euw8W;a65<hf)ee|QGIR}y-++)3L4~)dehx}9NO7cdLl&SjgJ;vHKYU-Rk1}8&
zTvV62gAc#QpTl}}AA+`+8utty(&_Y^Tfoh_fG3-8(J7v#*B4iK>a=2WcRL=fKZ)JL
zy*OpVR1^^sc6guRdVx_%V<VWlI>7)T2cZVxf*O+tJ1LqZu<9CKjozzE_jsICC+&f`
zx(*edw^uvkLmRn3`}8c#8)zUrz-Uw2ml5Lxi|--xPJL99^emo8?A-=KP9g3HUJ`Qb
z8I`<x^q<?uV;Ju6nnnOXC}kfn!$$}A;2{m}k!#yyF1YHFuT#88)?ZKU@B&USLr}qE
zh7D;`Y6IYw?!PYA<+}X4D*v{IuOXrOjPm*J`s`x*`W2LC?-??ZVR_m6hxdFbZKJN?
zp)#lw9>5=b7=COXD4vkZN4)XI&G^oDUX9t=itkS{35$MsRMz|gxq|cXHI(lxPT9)=
z4@_M$4@S9S`NQ>#cf)(dzZ{7`8>~3ecU@!g%Ays;)npk@?YtKI9WC71bLT`nuay?{
z;>qwm6lFx3Q5VJ>Qe)vz!h3K+BErXeEK#md77$&=Y8O6#9;B%AJ}%>=zhSSel9f3q
zxKQG<7Z53*o_Um8nxWkTIJ&U-hOguNgvT)aJKc|Arclg9CO@Tm*!!XsqXUCRQXUTK
z5`4xo1zjg%BdjRBQl0bji%}|~@Fp!Gjy5TM5^qONKpB^mixkNTeH3wAO$tnuX6!+P
zvIICf%0y%3eTM`$+Q~-kD4o+@Mkr*O-rs_Bg!VpGb*G)P`0&HW@#|mzCLTW`60I3{
z7dPt1Hh_@oiRX)jqLqfU5&2`NNWzhav9ME)DkRbhW7pKww4-HsFP_=k+gpz4E9$w0
zg@t6`5)L(*4Y%Xgt=n<;?khHku@A<DV6?P+GhTV+8>(x@cuBNMcAx<EzF{aHHE!7q
zPHJkK@Q(<6j9Ba?LJBgXG1%Yh#?5<bPgUcAf(N;CJdTey;=>2;IF)5CJFe@(?SZlt
z#R{AJqhLg_**ztdbzAtXv6mc$eGK_)bOibSOz%!c!jwiTXFMmBATn#hG3wPreQ*gh
zm7J)9wik+1x}1y=-rFXVWK2BCF-kCxi9lgLu92~l+9ezoJW7+ZQJ%V~^~$X77nTSf
z?!V;Ga=8){I-c!erDI&}&WvF=i0GU(J%63jjmWU<m?JK>wl^IeQ!HzqPND=(_>SC(
zT#L~N#q-GcAo@|l>4W!v7ytav|1}zoUG=LYyxZ1i&(4EZ-fJu`&Pl;{6pfREsMilQ
zHXZkoxz%cs-J+*{t6{uv40x1P_gI{XE}$Spxi3m(qxRwPq3TUz#J*}6W09{>Mq+FO
zf0(b8>A5IQEolq{94)#~1hY>KxVo?P#o^I`U<|z@3Q2ttabXWRfg<J?7h`VWW)w<O
z?h8>vM0~L7$%uxG&_2&`jVX~Gom%YgZ$;~*Asi88H15Egb;_A=C!?_(wIW&b*Yg--
z{e;G7Ip*f(Y@`K0W@lF-?^LcRt4Z4j4st*Ai4}=3Uauux8pccZLE$~gFeaSkn%>F1
zR(h*XL_V^1AWb~`CjnQkpZ98Uu>Vvzd==Bv^QO7bbG-a99%E#|xJ1+$_{v-&eHLq1
z*2oJBOEFXBIa1&i^(vIHYM<)I5XD|^I5n)Te(1Ef%-i|71>w(Z<TMw@a#gpzBO1W_
zNR5{%klv0tsxi}8vZeywq?3e>qLil|<K>8Gf#xVYhUgXV_u{_wqgmUDr>pNr?ReMY
z!91dE)W@2hv~r}0ZM9SCTO3&FyBe=g(O<@D7<J1WztH^dwau@3q<Xv*%|0dVr0^xn
z{)+RFB|^+b#lF+>5%mm?=W#$3eDV_*i>a6S$N~(b7YE<Lqv0oO+@y#fb@Y$uh~_;8
zCdMK~@Ti}fGti=O(uL+o!;4Y$SnWsoO&U|$GU$T~u%QqMM`V`Py@UGGJs5AmGrmLT
zvYI;>5r+JXgSMy06=3Z&c76}zq3`1uO!A_KEsgnBJbLtj&R@pT%H0^x&#4cUn4X=t
z69Z9_8MXD`VAu41VR0d<m1)mY96!;QfSVZ4ty54t*b8u?JY!=9bZ{h-41rJw@Bjx4
zXu`q%j_AUXWf&rwpjp{Mk{7@MoM??#6Bxlsu~roBu`ihLPw&Pl0^I>A`jUYPXA5Rc
zUvp11_6(x|2Hw1IQ1B7n5EYtWMY(W2@T_E%W*#$U%uR4gZ9$gMZ>PEx?9b0mqH(-$
z8iMhQecwl*UBb&N8(c-h@@2ILX9m@A#TE(d4UPd^ypg%1@wUt+T1ERYuDcH1kKrCT
z!NW0VkU|w0)2k%Ir3^Jhma;cv(rsM?>2xu4b%fLJzyEIh{`bEX-+VvbeUI<Ii}ycx
z&yF+f#kIY2AR66?RtM)5&G-4GnA1UOKK2PDy)m>IX9`kdP6!W}&lvQOF>$D|9O+0X
z+Qy;Qt0~Y_<UwRr3_he=rv2<eI4OFBTnP4J=;z>mqKYw26>yXg?Z+5|jtBV%+#s3>
zha6-~_FRX5lg1Dk2$`98q0@m^A_uc)XjOgT6FHf>;q`ncwc?><uH7A?l8=*-593z{
z*#H@i`Hij)I?j>c4?ZGK5A}px$6=W<rH#M?#|Zi}N$SJuv}I`c0%MIl2Jgf8HajyD
zGqai>`ptPqf=MQHy~01&DcIJIYmN#84{S`*`;Z-M%vHNEBqDDngB*L{k^VL9FUPfE
z<QPMg4CKtu0iCg2l=M=xpCk1#>Bsm?cn2;(XBf5gtgFy}!)KB{4qYEcWt>G=9}xLh
zwBgZv-HR4r6eB{LwlPtA>KBe)v;sI&t(vR?d<XCH|DRqG_I+|m^**ofWyj0UzU=t)
zlH~tCuJG*e_+tyy|Jv|%U9QWYTM;pa_*2SS)Wdur6A)$ybx%wg1_e1pNJbF=vqES$
z;}YXW_!@d3K;$9dRyygLfBPN;3jGstZq}Re=)s5a{_o$4-PH{%F8-is;rji*_{;cz
z{4f7&{Pbr(jg^%fYFADR7#08+f|Ah-uNat*$^3w}LGDu&^^3%g?*pczkHi?bMgQwa
z>{!3WB}Vle*IfsN>l}J6F*@A`=d(AEp5*>inVGVQEP|^(5nNa}^JadS332eBt_w;I
zmv3pIyS29+Pqv=Me*M7jX3=Zo9-q3M4S<0+%oo_R7nt)4(-!^Kpipt@2j2yA#<e2`
zjwcsP9KjcuIq$dxSu5-$VADd*<243XPw8J66D^J;oTWvw1aqB(xsqandKpv31#c9*
zZLBzvKn)zI1H}R3k6_Vgk{`MeM3u21MFwEpxu<)U`a((ugsH3+7QhUF5W_0hP~2g3
zO<p(ZWdiB|-r#C#Gkru71a@qUjBvo>8pRMBzYf53U9QV@`S(-4I>Xmrebn$Zp#Ptz
zK+FF`%Gaf(sU7KmaRWGm`6V)nxRCf(#MERimX@cZGF|e%QdX+!5id{$bND+3p2TaD
zY(caHDOBL4=6_mTrbT&ZVT=+X?E%FCnZkYv$Op-jGc9~EX2F9nir|sy6wTtd@FL#p
zy_psk?KSJYak4DN@$4*EmZJV<tLaoVICl^YHkL(+NH2T34qsRMBp0Q0#q57c|Fcmn
zO+^+3m-y+3_@ZSn_R$z6GD{57g?ixgD75&5x9Z*tQe}}A2Khi|`A-15y2AUKXZVd`
zY4FJ@D=<+IQW%d^_r+O9@Vrvn@r*~AnM*w1ZSOHq9xD_H<D`<a5*HZ~qZWoQ<SrDI
z={yEM@4Y3t2~S_f15bIBA}G9$j}K#OYu(1Nqr>A^U*Cv#-@PCA?|&HE+dEFfb9~&0
zwe`(-@Zn~xuXW<^a9{VI=;(UPNr`ml_MNzK>rP-GA)O@qE|jOrmT?#x<B4zvIjAMs
z20Q^yz^u_|I6{X#i)g>oim6VfH-@-I1ErJ3f08uM7*$7X<V2xEG#vYvAqS-ed-CQ5
zvr&csT8yJ=gmLh+bD~e-pnhX?LLN&IN+s0|OaxOa0V%C`c>h=N```Yv;N6eKrDgA>
zF-`<oz8Ix4iuQ_LSun2TG8hn${Vy$tu|Empgmcv>Q2}}<(N?5ZLZZenW!Y1*7<<K7
zv3QArEE%Ram|LXvBGr-vNYzc-G>+$Nyq{8((Khypa!S5uifqaH!tY^p2Iah$l%}LZ
zW3L*NR6NHVOLH6}cFTB%kq3o5DH|)*dCd*x+#oZ8&rfRnlMxN$AqKHzOd1s&OS%Vx
z9(V|jpU0cO`}_E(U;cNs`OL;HQdilKQd!V_nxn!`D~koUq;w>WSelotIKTnmf#D0;
zHl-D$ZX!63{eDuMF~%5WFq#pC=e1Rzh_S5b*}V5{GE9U|7(2)FGx}W?3??-9RiFA!
zpD^m~?(W3Jc|T_7P|D70zM)VimAU$;`8l_+9QpE++JaFP*s(@H_GWGljSm#T|LP*T
zokkpKzM-IWv}0NHM&pWNkv0?E<~@tp%S^pC@_Ot9NBwL!54`>X{zMs))@=+!plJa2
zLsL@J*4TJbx)P<wd1U%Ajg62+x9T@EL@=S9j9UXk<0)gW%fdN~w@yc^`9c%}G}`k+
z?`yScq8Y0}YCRN-SL)lj#->FIQ;i+b0n5ubW97!}s7#T95T$-<ArWCc(Py<rcujO>
z$?FXEpCJ{y)5Ys~E4tMVV9ow{DB|HKUa#o<)d=h4mf)sw*Zg5Fk7wW;W%mi3ehxqI
zge)CILkMQ{O=p_5!&rOzP~)8T@x#fG@nhdZ=nF<E6#C7Rnib3J1!zN(#(*`U>4c+h
zfxo^IQLH)X{Q`x@z#4i7Z!C*eU|7MqVdzM4srD0{$aNy$oSxRWP3vek2X&)E=m^F&
zz5i6zipFRRFS~8w0rO-4To&ub0Y2$7&I}rB!-Dq;eq-WW7frznx+FZqILN&kL-wU2
zYO&vM=_h-K50t{jnIRp)ASgV!6tB^D8|qy*^$gy9)l=I}G)#J)M`)aIHAcY2;h1Pn
z>b2}ExPP$c6mZKo?y0R9%T*sZO+Q8F+9yYjJjJ1AWoaehlde+}(I<=<G@O0A{4T#w
z^h7vGNz-H0`kFZ#)F$mnv-;$L_<zr7IEmcEC<z?^woW@J{xmiYEt@i*z$l#hIW8K(
z{*^nz)of%0L$nhy0)I|vN}+@Bp&}_68F$GM<^fK5n8uXvi8lf_4%&{~K$H)Aw_*5Z
zPsnbZwb@g!9w!a<AgyVhrIdsP(frXIZI+y--=^g@aKi`A<MQlSFy4<x4?objvPWay
zPP6;_JB~WVIhw5`oNgIB(O4fH?8lSU)p+!1HQssWy?FX`U43mBuiv}>fzI8xOh6x)
zx0~zh@#dRv#n$Gg=)zX~{&&BR2M<1sy}kX|+}QLd#v|SX4Vj*vwbKh3iK^n=lY$R>
zHnEQ;as%`izHdjLOU+xHRK#midOyY+y-sr0P1$RpSAwHphh77NqtjuF{&OBVuv9L2
zKQ8ofoOilr%}wN!>8Xlfr@1OzVhobOMWZBHfOlaSLe`~T^fAaU$SQUA0#hBxRnRB)
zL8R^QNn{)JUXG{}%%Pv{)*xC-{J(L6!Ao>T=e%F3`h^h@xraT&kk3em$Ztn)seRA}
z`a6trBO?L@eo4Q%HZA#iX%H<Bf5%x28UzeT54)%X85{W=nnrst@C~D_kjbIJ!1+LU
zfWeP<W5lB^fQEkYJkD3=;?qRrLpy07#x)FI!~PFp30$cMd4|tt2jiws-@&Jz8R|mO
zPjOgA8eBJN96sx_spt{4U(X1~xu5Geuh7ndcr~dSF?iDt96P#QjhFk5qssI5>)*5Q
zUsiFG)7MWapYN`(sw7x^_VLjIWxF<fU6<?fXIL;^b;ZcLTEtz5;=(9k!b2}+3N!(R
z%px7)4<m4jAT0)AWH4^LGt3D1(|w7Iz{Hd1@VKs>wNGPfV<SFz>)qIu0CFWE?+=O=
zfXg>;#ozqm7x8l)H*VgPV8@;YVv?%Io+ECD7W85=sj%!$42O35mxPkE@ohRcEdI`a
zoYz+vjL+$Q-!}Nq(OS=ne#4=L_gNqxTz}!igwsZ^!EfKj{nV}V)X6#hC${XPrNs*o
zUaOnyv9CqRIi660qhGIgvUs#Wpsph5QFqGgOD`D<)uMw%Aon78Vi>!S;F1hX$@7Kx
zV#GwS=6=S}W1{<sh9WWyQ47Her344#$@%0>bv_W>P--OQpgWm(xCJG=3Rx(giTTn-
z7W6b2(76fvPB=fFN)=K9NvJ@f$9qs7z}Okrp@H%8JbAM78>3eeJ|q~aEfyenzegJ(
zK>~)(p#((WMZm<UN8Rklg7T8O5I6;r>vCPL%fFxUl^MSN`mg@#kCu}YYw-VH>+-MN
z&QGoRWA9CdMR>roLnr!I7f~){<Myq29gDHJSar&q;axUz=?`9L{-|QUqjN(gB?cB0
zN$@C)HYk3$73DJvV-yZ7HsPbBszEsf|KqwB&03t62Hw4@-0!&Tvj;-W1}$WoN@Yg(
z5TzpCCb<eeWdj$yPRx~vFd~$w+eVlPltLIC;MVY5Hrr1gtD_^N@}h6Km?)x@%&V~%
zk@hV?nOamkk_UJy>daL;&N>I`YsU*{Rikr~2hLH*Xc0X^)CllUYa|~g5OLE`;gt%}
zStt-kMpAUo@EWO#fFtq-p5UaCAq5Tw1Mi!rHo&9(i26T1&Un)2reJ&~xdvGid0<pL
z9z(~0WEhl`C<5@pg};+ddBQLwN`<|+_7C>s@srhf{P=Oa`_2O$tMTyBR&1<q#Lo6!
z?AtpTgH6Qj>_ois%DuQHxsx$NdA6{)<dkOFJYMQ*o&M3XuG)lh3|WYG<AIGq0VNtK
zE02zkVrOT^k!SRo$d-yd`ZYF^XHd#8mdHWWi=w8XIxuWp=o)g1fT?Thx9T_UNyfI^
zs{@575h*BAP<A#N?B&u6QoL13BRPRGKW97ze&A1Bk~**!CkIbr_0c=gY1h@a1xK$G
z**`8@h~mUlOq9TxvLM3P36C@u;{rXAB`CpoCwr|DF+hY89{m_rG2V2AFDTKxAIE7c
z$)lu^8o^+N5%)AO^w0pgKt{hG))pFky+H?a1BK=VbBQR(y4sDxZcKIP&^$^J8Xz8m
z8#u$9?G21&9f6AoF_gO(`adZPg;IjQ;H|b0&4ode$Qky+Ar&Zl$P7sdMlTE+ytmWR
z`@1KSm2uEY#tBEBsVzk6u%_sr*W-7;`6s7HeC?IH@y$2B6Dv!1V{Y+wjOFH3qvoi_
z4kHc`J;l<r>cBgn6j4)A&^+edq}GHUr2RWs&wym$J_^VdvtMl;*Rxvx2o7hO$H>7m
zGfTqHn=v)BqJEG{PP9n_r5Pr83!cxE`lDK3rN~f>B*1)THbr$6r_|07#+o&Qp5##*
z)j5ZOVO(-@SM=r7_{iK!5t(UTGp?jzL&he>9D9*q;K4|pQY?yY=sGFj9OdEm0B{Vc
zDvlZyZ7_Xn3U7{#$9b(g#-Kx{FHnV{aRHjAa~kWjF8yg6pGHLI={bfXqCZGg#<@<X
z?r~>rgRT;zsUs)6XkazQTw+ad7VA$xh}FmUHLl$#sxQE~BYfqp7;mN~XGMD}!kv=m
zF;0l!D$!G?otAI|1!*NFD~pjY&e(B*v9=O?5@Q&283#X&f!an&PohgqThta(ly**Q
zQLFF9$;qMKdu6<F`d1raSY%;<z}P~o1WT=DOrM1_q}y}KP{9=XxU;jZdoTTN=nOQ5
z^oGEmC{Ff!HjU8%UANIoe_p#A5fY!o$iz?PL2jZPMGTu5MTDy;z^$BTEy=vld><e0
zJJlg+*hpVV|KKB+-nR?-MZMq;k$3etK3dl}HdF^`5p{)irTXl3w8lC0S99KmlmYJa
zld&$HyA&8Sw>!tCwQ51?3+>^m_z}kO-dQ~^F}$$w({<>C4QT2Ib#mXudB=E@`U2d!
z#ydn&E=9Xe>uMwLp$}TW>;8+&w%!X2bxQXq9*&dBSj>v<m&>y{%@{*O&!X2pjHbr;
z@R0PnXE9YKY9=c@CW=jSMR*U5jsB?C!=wpjE($QFbvVmtJhZ+gA{l1_>sZE%`Ya32
zo3Xq7I3~*z;%7_3IX#OpUAW0S?yCQc0mgSdpn9RD16<Pg)6;$NuwBtRXwgX2j}C>4
z`+`++R^UEIf~h{sGXoqM!YCco2aFjb9itdmPomX4_MGK95tTS8)a%>A_wCr<-`0ET
z!Y5=Xj0{Aus&2P)EY30Rir13TwR={Jz1_#+MYY)4UNh~PoGb}0;>((^^bZ;Y{xFP0
zB-aiP;vawc%lNy$`}=tJz5DUo-@hFXB@-Oz{Fdg)?|%PIy!qCB@o>o%q9^rwJ)S;Y
zizknss9g=!i~eC;@0p6Z*|}I;Sd5jG73-yzmsUiJ=OVAU4qay6Gbf-+%;AX=QOY>S
zbS(4aB!3d2ljIqUNXV5L(|(@Ev75FNwL>PKWN1mUCI{m^IW;9dKP_05_3mlogq<Qz
zPlb<{Hq=Z>-oWT%Lsa`T>NW8+(Lti4yoaTpC2HvC_*m}{PN~g|FS-%-<s#J~?Zpuq
z!~D?cizvb=!GS`G&Ew;m=?3$iXRQ|$j=;0PYm9&FAIAIO`J7L5Ky%e}Yus~%u{TUq
zFi;UG#&rx%Gc(ngpPx5uni`K3vBh4JIQJk!B|hrC);h=-M2&)n8h^{-DM}h+7E%A`
zvT!z{USI{vg5Efs@m$(_3#S3}M>xHRPOJaG$vP+3Ot1CF{G$OrrB4ojrGNA@o#%JD
zl*SYJ1bJMQ8c*!Z!uP>J8NvudJkRoe#+v)l2jXZ)`?KmB4i?8plJ6SgcY=)_<_C3f
zeA&g_N?$*ve7?IrPZ{>((~d8ze6(J;z_sD)x?Gn((?Wz!yV({4=}3f00v3jpl+H!s
z7Yk4XCX^Evd{d+kjEhA%48~r^{5H`?P)eV^UqTFN6L+_F;yt`+9zBiAGYPOS6w)WW
z{>^X2&wu{&_{mRy8n3<fni(|DJ9QrPP0YYw(!W0>@O;9^1{b;2))9pJ!8uN)|8;Qh
zkkR-e-}R|VeP6Pyr0?{F-`wxTf-j{u`HK#ngE9F>pT(8_1~;bLh)zqt*}to)h1vGO
zUOd$TdhcjII$C5B?Il)dVGe-=xbiY!iGaW&ZphFHeP-iy_RE4{!l-%Oka?n<LD52+
zy{OPR7$}Q(-RWqtv@pE1@Woj3r0dX&P-Y;MWBj6@Ji{j`84x}Y5D<_MvM$AZ8Ee`I
zuHa2W+w9Fr=T#rq!5PvBA@tA(SULS<p@@RSZP!8M7@n6b7{M>@>+2mDz3}Fyon9CY
z`f31c(J(3j8Nq^fQRoNlDYFQakl+Y6!H<3qAH6Qu<+}V?m#@(9^>4h#*Oyp^eB`Sv
ze&ZK<uQ~VN8;711Kg;LG<M!>Pc;k&*vAoP43KPDDQUOH{3V8Sr{O|%rgWd^$qPFC`
zgW({h0A>M-yhH7F{IFbuvL$)UqLd~o2pNRBXaf6u!Rt6jJ&ErpB?gKGj3e-7qQT&O
zC=D=Dz}u}LfM4rb-e)6=+D57^l;A0qoEBx`<CopjV9ziV8|>?YV#>0a_^()coVFYG
z28OA#L44YyXb3S34vnZC?x(ht@D%1b6dh`h`yMtHgr1`f=~0wsDrz560ZKnS=36Jn
z>H~^D!N<l5cr6=*ch%M;L((tC7r5X#Oy5zYIeJ|(DDneJH53o@AGrj$p}cF<*z0Ip
zKXGtD_RRVj+76773rGt`bOR#@OwW7$Aj)fh|0HVl(D#dI)KL_5)!*^BcW*v^@rxhD
zfBc)jiEn)44dF@A<BAN1vIS=&yo8AoAu_++YRA#xQPdj^E2l7i9v&XV-X01~ltD*P
zV-t9_2L;uQ6_jJ_!J@WW!I8Y$N!^1HiAV|zY4igH2z?=n1SCSygf}#Ji4zcb$Fq3r
zbG@UnJ=ov$_)oJj{^YcWgng+RjanQER=c~KvG(*yJbG|H9=!jS?ybkl@=Dygbw_=z
zs*fmkFwRWsT|^l$#uyzW&s|A2B<kp#=)4Pi`(u<P`pJq(#@+@YJeg4zbqynMoc(AD
z#aXqdWPBn;B?elIop>d)UlYzs-ak)cg@Ktp)lh8VxsAaKSTT+^jDm<(e4nf9WSSst
z)4B1*d&sKejD^4$V}dot!Le+~>k*Xw2L}g+G4oz%p?a8$c;2^Fx8|$lMh&RibgsEW
z+QuotgZ-mON5uoe2+6@H(Ah`n*T4FQV1KahfA72T>Z`BE)YNQ@d7cau-G+_+F>l5M
z<DBGO%|VQT82eEEvTxj_8e>DIJ~i9b_Y_46{-G=c=EXwU?<b;x{?5)WSh?xQ2uEzt
zKG7J#odZSrQ29SPh65TjjeVSRk<xFGJ`RIag8dar`LnvlTlYDFhxThM8~eiLea)c`
zcL{d7&U$1p=h*`(SI9;I1DNJ1v@b=tVK5iPV*vHfMmc_CKxiJtg_YnKH?@w@HE^uO
zo^xki<2-A=;htRST?3`M#+|*=8jV9oc!1-K8HVBUOhJ7nnpNYjHAf+@b*I~^s)QTh
zcwTVK=B8qDsv1Neplrq{zp=h5n!N9EU@di_u>eNXGxIiPjpwR{3%H2l^i=Z`V>J;3
z;3`ocI53cYk5r8q{E8D9D|~T?U=2jd#FP4waUMFGnIK(k*+xG0Mrvy=x0)yVY|$mP
z4X4s%G;TK!qjR$FsH0JhEf8UzXGA|TS<-*0-y?#7#y6|+CY>lL^lsn2tu<K5hA(Ip
zDdeHK@DTPV<6wU?u!krkXfBSBqs(u=g$SLD4dN3M81G44nATB517d7&ib>`T26o*K
zO{+}LM!q<wX9Y`~X;C6;9<h%U_{KpL(AjBIxPE3Hm(y5`lhV%VJ~8%aj;l>JcB-@e
z^G=+#j~yv=ekK^PCe?d^oA>S0`+H}iS9*p$5y6EKqE1c2&O`&6y54cBS;i0IlCkrF
z+NJRs&k%*CzKAY(?f_sz2Hd6Ij^+$TKX6R*7hFCAkB1JP%)^Y}IcD4|X#9yjbv^yF
z=roSvU~eamg=a^HyQcYoSg#X3#2{2q8?yS1!aGI(I6@oMtHzv&$Z@ro=qG3d#!D3M
zL_&h|7rj<&Z+)n45p6dUIiA&+V?4xo&3YK4gmoA?pT>(dJAGyRFsy1EL{o4i+1uTX
z)u)f5b9y4U^rI=h$G%5SL8oVdmG}M=O*YTaXJ0(PFjT!veej}DuZia#i{_HnQ}at>
zjB$|^uw;O#)oM;b2M(Y!aJotX27YtVYdSqE(Ne{HM)TlEbM>L<P&jokJde3ttjrnD
z^WxXu4nyOKJh8pC74N-wKYsuF-^bSWc9b+1m&qQK%NqVCO^nvSnrD)J3%LR#7xV$X
zy}Yt)hZy!Ooz(O9?%j>~xp^DdFa{B+heM0?HRv-0Tj(FUGaFWfLquZ1!zz_2rylgX
z^j8q?Gwv8vqL1ePniprx2_hDUvJJcr*(f8rQcU$sPnowdw~-qP?CCXGj+v@pqP8P%
z@+>j~>5y?qDHIE_ytHK4k+vDQ;^^Zv%%W9HP3AMkSLy~%ypvC4V0bddy;igB^wH3I
zWEvv8Fg!s|Q-oA<LISm@2e=}i(bm+rBtOzN4({a>{sDajujvnInCUA~P3W44q(sJJ
zZlG&KmWN&v@rTR<tZkrU?(lxqNn1FGY9sA5vNc98qVwn%x-F);Q;qU|^&j03ZA^Wq
zu9V5)*?Nc%(+9X>`uXw#4>VXU%i18NzA@)Gm;OtIv8Nu!lIL7I_YDreOE)Cgi5Clp
zFq9+X9v-nz>8|maHX^6<zF~P@$Cp!5Yd))dzPrAx@*=Fh%>E)lTpPZw%XRrPE(n^Z
zBAzf3<Rb(z7&ZhsF*1}1q}I{GQiRt`O9#X@d97f~NP;T&W+Z5Ne*it-mvC@&a1@X3
z-;WR9ekaabDQtgO@DzOQjc>&N_&@#+j`Vut^*8J-l9(gHpc*u&7xfg4f<4&lLie~1
z8alX#&*a_5N&OuB5p~2_8@vWwiTctX=lS1X{G_Ym&ll4*9&<@AOy704f9ZJ^t=vlc
zP^_gZD&lE<37I`Du1`8A@pNk=R<-EbJ=#k~FTIoNc-SE*QZ?;FXeFc_&l5*+X$)Z0
zB?&=nbPU`rDCrp(xx>754+|6az2bWE+Cdmk=>q8{OcZ4sJ;1O?pB>$$g$lwJ(P1bH
zST)-#NS#91VE=)hggf40qbFl8e88It0g?BBBa92r<Drc(!ML!nBJ!%Mb13Ub`-D(L
zBY>1Q1Jxhx#W;uu5JKP36Nm6`ge(@dEaV-jrg~`y!UKgq5u(pI1R3y?y-vuBjPOU@
zj1B_p<ppCi9MJ1>U9QWYVfi;Td?f|o=QzHk@@#?jiDkggR6hIOude*DZF&A!S~TJ?
zF+X38Z+`2p7W$;$!f`~hi*AH}@CVOA!9xmP_!ygU!>eta(Y54(+P4t`zJpw2rMd2<
zo!kpw#xcaQj`*q<@Zvd^Q3eZd?!81|D}Koy6evD{KT%02<58^Rz>@f@?iIg68Hlm~
z*^B52cqetUsHcA;S17lWp#jftc%4{ekba6(ZYia)T7PjCCyknwuJnfu%0~&|aY0eg
ziM}H`F2zeRvJ}MYQBF#B=+vWodTN6W8_T26wW3_@LMe&T>++&wWu{Xcsn26$z?W-8
zmmvpA9_e+&XYt5w+8_o0M|PQ*s6?SO<9%R|zeo>8)kGGIWoFbC_F%&BK>ENG!GXfC
zd)AJra>??=1+s`_0hBeQN;;4nwX?hBC=5Jx=Vq65ECh-zqAl2;1|x=0KytuHyz$2E
z`1wzN6hHdW4`X?GQSfH}uePJbh}b%5u*v(OQzc<M^j>1XM-4?z8Ol|bGbBssf|d3t
zYEV8+;a#my(m<mCLb(P0Fs{f-q__g7aPpxJBvG&f`G;$jYQ>5sXJn8p2A<n-NH~jJ
z*gQFo&5cKf3kvpPal$en3Y<>+MAx3go4@}}{Qckkx48e_+cuD)7+YSxWkr3y0F3dF
z*BFi=+e`M6Y({EI3??@0owfvneZj3~#Uaruy3zQpht%2>F+l_bsjN_5l6q@`$TN*0
z%3#z7;0DoGw3$?F4SllzNqK5Um#G&8Rnho?=nE`U+Og!g<g{D>fK!??r08S+qjAj%
z_I1UmMEkS(VvtTSC;S>2O;Kq%QvRu5M1j&b=B{N*wTB253_C<Tf=ejDNrgCFnb#Vp
zq;ow-*0IjO5Zsa6zK^l7UDFs(I0}F?SKgaXxH=*^Skkl(6u=y*52Rf??=>|aPSc!4
zDJi%Y1t<ED=1qz&AnF07pJtT*Frt%&Z+CA~&vgakyvB~SSkMds3&r?(s`p&ua8egt
z8M&|_ggHBgqLnBS)sHa+oW<aj44k9Us&B>i)?>8+Ln7<EbDi6b$B*vEhH$9es_DHL
zFuLlehU;0h49a3?%81`h3cZuYvCb3mm6U~NqFwZDe2o31kBpPMyBiwgp710Yc5!QK
z3dSfgaRf;F&Lu-aQa!1QebHwd(4omK&CzYmiI!k7E<8%k18M(4o<aG?+9a)8Seu|w
z&tyyLGY1iJMd4@B&ralrX=cG4QkTug<kYOLCF4}u4+^+=MQ2)!qXWuzoDnciq97e*
z?A33gelWJM4$2qEYLRv7#beFWF@3L+W)dUmAa!K_g875P00zX2#)&b-v!AtEKX$gB
z#`@a(vA_E`E`>kkyl9SSXt-?{J{U(BI_n_OA<fT<aB^W`(en{yG0JHSUZl93oSIV~
z7t}xY78~;#tJNkd@=WtkYd6uW1b3qIs22r(YAg7Hq2pArCq*3!YWjzB0Q?Xm<n#UD
za;h(75b>T^HndVOmT26H@T)nEw_qe1jUhF?NBz;bjp&}V_t}`ElIyiy<6^hl))<W2
z@JDJ{3@f{PyTZMOBM(zFBXmuCY5YphHezRcRrei4NAsKr93t0>8hiS0oPma`UX0Jl
z@TUTw)tEDnod!|gMo1kP1&wbz&O`%5RD{EfC3wf0m#E`I!JB<8FFfzgI!zmR60gE&
zBpQujPjv&UQn_k8fd@LZvBpEIQ{lA6lX)us%$SbAp!^>63zA_5at7xvdMB~7$$ph5
zqD#}}F{HpH%`3d8P@EB+Dcj)1T&6R~Cdeun=P{rbbx$r&)MQzFE@JiZho1jL$~=Df
zP;+-(_$6p)9<qG`c!h3+xd2|AOIAV=4=uOhQ7}z%R?d0>co#-)oLG><rqu5nH*UxL
z+@kdwL|Cag=3k_YtcmZm+I6=x56+7p^v(r$$pAzu<D8H$Fh{HEqhz_fa7OnM{qp3=
zs*Wdiq(Bz9efN$H!5FhJgy592WCxI&;+@Rje9X?x+bIMC7c_*qGdnx$=fN4;LE2l=
z)Nb$WIO2)Dv6@Zx!bHcx)(j~Pv-KhBFX;qH&Bu549LKQMX_F$5G|uQpFz}^FZ+H^@
zxA7T%CO$<L8^#I;5u$vUZ}1}SMJ{B{VL+Rftgy7SV51JQE=DGdl-`d^{ll4u-xy3v
zB^+If*25qdk}?mvPY$9)FU22(SF{DV!!Kwrkz>fa=xt6(qo}C?jYmEPr{O8cMDCZ`
zhaSs%89@^n6=xuHPCj%UKAencI3y$IVdzd#R!NtnF=6j6{RTI1Y~dV{Ym6ZxGyDua
zCG?lbdLrx@Hw<#L8$%rJ!x;$}lqcE8Go@Ulf9RfQFSMHXGgl}>IgNHDd7mSl=hHey
z75}gA!{6NZV)6VM9E=O_rH}s2GaM>@O3%7}4*%3U1z-Ktd`bA=%tpHA6t?ZIBNbOy
z*Cdl1ST~n+p{f31;rmD1^)D&Z>_hc_(xLx+yzJT+J^tA8FMXZ@He4INuFG}#b1$Ng
z=VEG!nWb=Rggb~b<PK(ll4s~mhoA!kgOS<bq6$QIeLX8C3vs{fUPMDo^Zq;U#Dllr
ziL-jk&khTl#ooMqH-7OSe;)t&KmX^rd-tvwb4r6pgbv0p6Li%DS%>NlnFoJ^Q#R(L
z-z-ql+4M;neoqX{?8DCw%AhXq?Ukyq%&XJSuDJ{th<~LQ`g<~X4KDa19}-Y>*3ar6
zky1SgI44?UtZ#3|{fF<z=KfBcBP6g`6l3Il2sJ2mV1hi4!o<%Dz8G@&j+Z$iMfbGp
zoY8KlE^rS~driFqhK_M;LQEdR8TU9%1qLt)1?d>Kz=m>+g%96>s}(Xj5Fqf_BmA7a
z)+nzK)({j^xV3~7EkL~R(>pLkVK5^FR6{~I0zYj*umg7xyuhcP`f*NMG0t_9aSFi<
zL5}_+aM<{UN1&KILN;}?IHZp((8kpdr-`7wx*vF1QKp50z3e445C*Oa)p5bb*Adj!
zb_sj*KN;UpBznlM%XPUfe}3hwGkpE!AEr9|D#^+hdRJl?e^$jDShSv~Bb3rNZ_Y<$
zngwp+p)4A`AcrUDZg>`{py0dki=h!K@xJ8Dfe?^9Rg*_4*Wpce#8DmoV{xzdCZm}p
z6fXH*>K4P3sE~LMd-JeRx96+cQk|Z)!UZKNN+S3q#rzEqdDV?G3d%4l#i#}UL!oj~
zuSKiXP}_;HK*5;O$zTM!=#%n<DEj2FOMh5U6KN8cXOi(6(Gy>$9+VgKW6X{2InrZ&
zZB_6%4Wem?_`-nTPO%i!2a0u+Zz$@mj2jtK+mO##sG~GMW|RDA1z9t;w>RS8U{8FP
z=z+<|PE6|&j1v<vmd#5(0k%ZMlynZo6Y^e0-*c7^$1@|6bxy1h!y%tBLKY%oCn@&W
z%$+E699u{`l-GBnJEkS4lsz{vh?dK_n3<V~rKM`DtSrXz@<LQ9lQ!^dZ*9iTE_;P^
zEEnZ5nCM9qw`}H4nx8T$nk7@^bRD@0C7F#*l7n#6ayp?(#qtHpE)+p_5EITIA6xk;
zc?#tc3PHT2tz6+vf*T(Dq{yqz%*4|24fUaHWw-#5<Ut+a2<hQ4AUyowU5(Wf;XpDt
zVz_HHQxpVg1BqzCNeLLB#H&=QQ7TvLDPNhMi(;|jF>JPx)ed#<g~yx}bwpdB2*xO6
zxk=}de{p<bo{{PfCH9$~t=Et2-H!pGG+8k&T8T6;uAn?7N&;mph9vMKDXFG(IYsxN
z94{2e?x3H-Blg4rez^&v#s-E0!-BO3Mh+d|-c>g`CkN3wsYf1pm$^jRJ}cUV)nlWP
z=XOCkIYQ}eg0T&1$gG<Cc!LvpK;31*2akT{AE~q1_a>LqeAJvwQGO?SUhSV1eGwdq
z>_CBMTA=ZB8Z{f1FAb*)%`ud<JV!(#bB?G?(!mwW^TK%ykjS&+dWYbaWlty~NH0D2
zkz+Ah5fw)tP~eer6Xh4t7wktxM4;!m-nqNC?RJ8bPFZ+y7R~yxWb%gT6e|*pTtsMG
zp10%hU|VB<;5HC7v%R$*zx&;<bo@Hjo<8>2vZjDuvG%jPtuacGr-Hu@3_4?D(B3oS
zFNOuu4<<z%%3Pf4I#H`1#!0KL@f;CchlXwBeD)_H;shmpimoDpHI23E5IrN^TiQEL
zxUTkN`~xR3Y>$s(bl{q9*FoAsl+o=Lk%cIDX(L8D@R5TX^*?x>9n(B3T9J++j`x#N
z6eErhWY7lPFZzTdCy_#V;Uw^dyt6KW)(%syVRUY54iXi`yk>45IHg<7^pE|oFxFBZ
zkx1=kUBh}|0|N7hu>?uWQ|!^Gb+GEdaP#(?zjBnu{KCAASdNC`aDk@@j;&*jS558I
zyb;o)h(FdEtlc^lt<t=oWZ$QRXQZwK7l?f1KIoJUqZ(Y`>@<U-H_yhvy&2%#`KjQp
zxgvT<^dS5EP@ro>(IkV6o@Jjc(LEbVZP=n7)>{})^LnRnK)B$2p>)pA4BnH5D7|~6
z-;A?PUHz|nPFClZ1REk&p&8jIFebu<$tlqu95oz)rE+yHen#59Bi{#JQeR#?4H`8f
zxgbSHiq44#K(j{2#ecM(zq~jVpKLm6nn)&S2Y87ws(;b8As60>F$26y`}AFUte$G#
zFjg3<i7xj3mcl<AChV-K?-)%nL{3*RLb4xPTJwV2%zN$~6-+RS!W(EM+aM6}g^s{Z
zqB;*MHR}hlwXqhn;v1yj!w^ZTPmKQTl{lJ##%j!k|IACA1~6bk-;q;*7t!AEU~uV5
zW6ZPg#^gBCHEdpe?X_6Ganlj5@LXsphEJmV9QmboIkFG<Yf{+pS%-O@*4xlA_*7mp
zz(jdQV>K0}iM;g->l+(Z;Ilr*&|~_dwju8j^}e~e9>4v~Z^WbbjQ@Gb3jKaE)X<Om
z(SCgJ{s-~TzxtK;_C)S6j}cD*A3<e!7xR>Ph!G6FXvdkJ=3g#3Utk#AKR9sOVxnkx
zcQT@s(_B7;x3UK+h6){mm>TN;(N`fmcn_vY4Du=EAMYaC3*L=m4KfD05bH$LHh3-u
zDE0)yfrb9lZtHGDbJ4L7<wXP(G#Hpr%&YaAA>_TtK+xx|ouzP|8aJ*ueWLy_;DMLi
z&vj%@j9SK3^%Fkkeb#h6;d_Yo<NVe0_<`&2PsR}3rycxGx+@(XZ)CqT?$CFPoJ08p
zy%x_5Bm60}4`(k&I;%3;3`~%TXXj@1TlJV<6l}<2p3_(6i}#q<Z=8&X#N*k${e8(s
zDXJ1<8Ig*}K(rl$8Mv4tjfW0J1A~+PSQAi3GT4cm>Yn5%Ltpt0N~W>)4?c9b@0{kk
z&kjq1NxF^>ZU7%$VZH#L;r-9*9TYq23BQO2=Q~rATX-#Psnr^>wze5t+q=TiM4O)L
zML*h(FRy&Qy8ks>@Wo1k{pUMATHrp{hOg^#UH&~35mpvcM1OU~JTTZm#O%c)28_YM
z3q27EggB>g@j_KhAHhnD*b9D9Xauw4!^3#z_ix7Ici)TFffko9l&qNSH^2Se_{A@N
z7C-*UPvW&V-tcCX2nALch-n}zi!AATivBVK(cjR!T3`9ahYl7$oZvpXVzy*v!FQhX
zJ)HG**QC01m6N`b>hN_wsRDUDfx!nqxz0nb*RM&n`#YD?hwtZmR|0MGv>iJKd+}s*
zHTJcj1h%A=L=Z?}=IImTkiBmtAY-(m1{NF$C&1Kun5iU186LWjXuZW_rSZ@`c!VY4
zCOyxX*;`m}V`F0k4HzqsLs+6G+$v~zTnW+VKKhBVjCv8a`At2J>e9U!%MfT#=HazT
z`X?6a_K?)Q2r(7_CD5_AS5CqUimS9}C4B>7?rJxTUiD*$!n*~5lm0oikLt<l8PatC
zU+{~Ag@sdL>K)GAp1`7}N>GtNWuaCFLL3VU9tO8)OMx8RTD)=10uZ3PF4yI{{JE8{
z%<%QYzx>O8bvgMNUwN^>;J(J?%k((>#3BL?zsP8zOlrgVx#_4*69Gk}H2id!QiTOK
z2ClSt_5xIl10IaR2|kCM0q?bvLA(UTtUWHpQy2l1qfSMqPwH~ywB-@~GQS-#B6S=7
zlTw-yVa7(;$WCF!@9fN+j+DA=sL(*cgq-CCy4nmc#=wJdj`wmO)f0_YJ3fe06!BT2
zk7&zPv~@4_Bt>T0m+Y(qzfaNCC}d9>Cqddi8_WjJ_0HLEYA-f69!KZ2X>ZxGQ{Z6)
z;htf?P)N~FReY7?kK~DtVi|b__KSf{upGfia}rzYtI=wm#>&cVd$^Beh*UrkirgZ3
zG9x(!T`<NLD=j%s;eE)0DDJv(bg&&wjC~hpl7&t!KOqwkDZ)NG7(KiPipK1!j}j_t
zWj1mJ<9*}CQoR29D{<$}ZOb~08`0+*Ya8+C;o~6PH%2Sc|IEzJNLEODt`XJ4J4mI3
zEW;R4ABvKp0Tp?sAlZgvsL%lGhjJ(K6iP2Vrzt46lA?vk4&7()qFB#5o!yO__hOQ?
zRk;b{3~g;S>o!u<8b|)z+ue%w^|g4q`qUoz`+HkaKi(B=+QGi5^Ye>w^Tyry#y7tc
zw{G8y646PyVw5C@O;xJGLBUH6C3>TAQnNDHY1@d7iK51ZXqcoRqYoI?FcOizvRD*e
zjANt$Z%K)mQY)c6XC7XjpQ+vIyYPfa6&qWGFWqiOeJ8RCr55rZ%6mGGY}u3CO4fp|
z4x(n5XBb0?Dgej0k7HOyqkFm+AH4HV@!^BFEi0BZe&ehYKp~X(s+VYil(vw0k210p
z`Qnu7F6%t)7tZRyV3f&75++ik+j9F0h2+%5^UNg_b{IxxXBI_Y7F0*kV~B@4PG5NJ
z+wkTn$z+&7c?n$X<Rv%~)s-(ztNBy92gNZ)0JRy#zZ<T1VaTvN&HiLOLxfV%c(A##
zW-oRO;V6EcqO61wA{!UNXXM;M0ld117HKavZyNQzNc*U@g*(SKL^tXOUh5UI1<B@B
z!QsmI^zOTFIx>%S1&Z_;;UkKAua$%gIE<xzyci3V+)0^;p&_ff*talWoQ|B@M%!5*
zLdRGs>ZjKTM9@J`p+RF~7(jpnbbxWfc+UDI8B46tGali5hXL<ObCMLI6M_d3yD8lk
zO2$kUTqV+n$O_U)iWaG`KEOC<gRF2OLiB_EpR{&C=}j~Ra}@(+`>ZJ(z@Y$VJKd@_
z+F&L+$XF3ggkc}Y1Dpm^bRzq8Y0c3)iw*{;+M(Y`)Dv@q^$7e1gCmN1#<6k2{(dLw
zkM6@br?~*HfVMy<N5`s>EmR^qmg>eJP1K1~@9LF8JH1wOtE>52+lbBehjA>Ppaz=O
zA0NUWgg=5Uw2^mWJi^#(XHCJ8H4-kGMbS3qaZ|Kqq<<=?V{Bm!Cs=t=%K}Nfj3{0`
zlP{3gk%-h$@e@*<Y7M1(pbZzAJKZz(NNRb`Cj$)x9$<=(ihg6T9vQ)CTZ`VQXu4?D
zso=(*UGNUjDMpzOqS2kY;cWT@O)SmoJgIq!+!O7eBAqdYNO-W0XY4~gP@e%-brvJj
zaug@Te{=&zCg!=LMKw<5H0E`adM>MLXK}Q@69@ad#sOgEG`Lr!PSm(N8cDQ8cxr>M
za0@!lo@*twkJPiK7yZ$w5Mc+c)eWSJoSL4Cs_4gbbxAaf%r*&UodS_@)xfg8#Zc4{
zKS%apt&5(5O4-+`*=h-tBast6v;Ps1&bMyeQk^r#72st@3LHm7^DuZamlEznYv{4&
zp!a{nILZ7KKi4~Wrc#}=b3~;w7t7+?i%Tmpzo>qQ&WvT!UXpGPc*6Vyrp#->drUNx
zalkQxe)0Yg4aX>kQ5I*Hl+HA@r%^kJ)hBB<DiM`TTaS-VVrS<do<3cV_dj?L|M(C8
z9RK~l|Fg!Uk>-)=#aIq4+TY!=(F=nYhDP?`x_9rM;5r^_YwPjk$x|EEYW0TbN5|`Z
z<e_#;GEZGJQO{toCBpvT@F=!+c69$?939oHkHFY{e%_Dmo!!{j+7Y~=H$?H%_r&YT
zP=kSHa;mKNFU8#aoO%5JpS}P7wk)~s1J8Kg`<K453RML|kPt<JWH+giG+I)3t6M+K
zn)wgsr=DNh%&gVwmZn=Ra^xtI@Iaw_`poCOf0^_B?0C1T00JOLkRS-WczN>Pdrq9N
zJNDkOT~MkoEi#&!$@;n|ye)llKlF;C%qJxo<qYE)iZc8G-jq=S&#m&#%J;5_ahBys
z!3n&@1B@b)0v(TtW;yq%dzUc|W+*Ju(yMmk#!B$#V#*#3PeNI~Q!t3$!ui&#wG5v!
zf>#!PLOGOR6otYI!f8eGOC}gz+IkDD!u4I*l{Jz3V57<m?25!>U64J=SDtbw6u#;#
z@UHeRWej-c=^rqGr+fPu6ZFuDWRg6%C?!QRdY*ZE^ypEZQF<L!SjK_p8E-M;S?x?;
zjYH9)tTDz&>i1pQ+NG`Nllz0uLDI%*YyVlvavI1aOP+T1-Tl7Sdh)G5&sFQDe>aZx
zU7o9ZlDg_^e^9)bdUtjYYkllyjtk+{<vnkGng(Tkw#*-w>iG}l|54XpoT{<?>(7{e
zRQ*4E{dJ`uE*F$sE55GNb^29F!SNT_q=^k&mWDwB6WJKdEzZ{l!A1^IL?9?G`o`|c
zf~6RNm)W!(xAo`G+xNfoo%Y>7{l=eDdQHvDwXgiWzu*4xKmJGUzx;#WZ(sbv7s6;Q
z_stSv6icik9yZ=VLIa|b(mP_RW`#1i*GW+B7P>`%q#~sVNB*lYhV?>EHlPTqvgD|@
zD{m)gRmau${-1mGRuQhs8LFi0`aZaokGku-d-;AI;=6mW*B-AuZ67}VxIMY@^rD2(
zRE#aAh)_yQ=B^kb<O)+0hQUUKf=1yLmM}fL2qU0{@)&zwV1qVeh{~-xKQKw$%vs)<
zFX0p@sLE12e*`%;S~gr5WdgvnFol>5<2XAbtjA0dV1hx63PF;#U<fccm>BKjHHKNo
zU|`}fY$Y_vd=lu`xZB(rBNjL;C-qes+HMTWx;^7j8<Tua&JZN^!>DJbv(ZX_l&60u
zVd}II++kX3KD5h5A$_)?!5|Ji*}9N3<1<c)oBC%kL(U-!;~01^Ys>3&ovza_Q~F!;
z`1<Qk@Ma<6Uw~v|__vYNdPP=?zICG8H>rfqZ@u+;`_h-*YOAYD*{o02Mp2%klc6DK
z2@6Uk5jGdCbcF`m^b%4cA|-?!2^|jI!jIIs6O1R4oLAA$d7z{rlnSlNw^DB9UTvh)
z9yA9pfWGn+sxzCybd<n1e}v|<vqG;-SGx!m2Ffoqoxq^_R(uvOwW(6Rpj~J+FP-DV
zz5MP*B<6$=_#$of?%2pgDXxT6(oe$SI@c(Fto7%XRzi~(Qtv@~{P@FmaIl@Rxok^|
z4p3ic<C7G1Xn1<3Ef}V{4E#mfMo{MGKv~uKLT7skYe&f;JmFDz4uZcscR!c*&m}8G
zTAvrmnJ?PFa<#0)Q)jPdkiy}m%!B={@T9w;)kn$eoRiE&jmLEOgQ?;f*4Ll6qwqHb
z70$3Rj`$W{47YBtw%709Ys<^aZFgs<lwNqNwY9af26vzJFn*SC4*^MefMjg(q2{K%
zX7DE^h&eBVpAxc!SLUJaHw)?#pd=4i@W~4Q$<rroV|^p@AHF7YrtIRGDK5i@P~1>t
z2z|vHpuA8Rn+ZXeI-SwAv9Z>6cdyP<3tTdWdiwNfdzMRhC<0mconx`Odb@;{_wT>e
zZrr}xW-_Phbe<RZt@G8wHys}Dg)e&578X~^Gh~>cvYDf?v@e?h{SCjkyS<hE9hZ=K
z#+hQn&KPojk|+X!*FZZSRtQ0ghvG{!H<XzaBhCUl47~EVnoB)d2Gy@&GFjm}PYyCp
z`uHMzYWk6}aPE@z(_suuP`2`bB#1g~>rX#ufAYuwXM6DRyQzye1AdZ+&{-hNT1vM3
zH!^1Z&15e1DQz90bW46f^oqwMp=^3RHk|p~Z;zjR92n_BN{J#BOo1`pK=X4e6|T#g
zWa(eg3i2mR<;XbU`BW}pB(r{Und6Zm25I4Ct*gN6r^>vngO_K473X_h)OrJN_6K-q
zg1~~o;0WG+?|k>$ZEJJ0-MF#ZmRD|skDbdHs;{+^tPi0ijkWbh@Y&DH6<y&t>)U$k
z9$50s1uv~G-%4MEhU^|dgs-n<oxJ_l7us8If3dysxi?DzW}Q)pQ#P@pA4<7!6>k?I
z%m}@CE%8*M9Lmk`)C}W-H-igV?DIt8iM7AKS0VG~=bcNl+HqiTl)16L0FPjGXyCaQ
zAfd?9Zh3@u6%G|%rl_ZooX$9m<sJ`JAd{xm+;k(j;D8KJSLDO1r)~4aldRjrjBi&$
z+Xqkp^QNRY(2lVHMV;^T1Luxq?Vs$ntu5yxo}^6cH*=ak3r|>Qk!^3LEn_8MHfB?k
zv+ZDSx4n4rxNy1iy^J@(|6m^lpGc+T3=Zh$mQm);2Y22sFGFqTnTf8AkB_vMsq1NQ
z_lF<6Q|ownc`>+svF6w%1lY=r9`co<ZENdADQJ1i)VfL@uqL!3rP;$rALsdDR+o_a
zya2P-E-5q8@4&OYZ^zLV7SD<73zG+o39Wa0e9)dheb_#J@Im%+*`H-j?GH;BUwGd;
z$Em~Fi7(oR@4pk;^Ks#I>u!H{H*nQq1^kpJpH-1H4Mt{Wf^&pUEMtsLbBqx=#Qxr9
z(PGhN^vN1T^T$SC1vci}*wktp&$w2j)4z$SiS!FRp!HxP5CR$LXT}%J5H^z+lCdgf
zKrV_dN>6JZ{DUv_zxajJF*kp|6kpC71sAh{)rr(=zbm}!Tws2BW+`iy;(Ijf@~{+=
zFP=XQyp3HcH%E)EY==%60Wd`IDU={%9ylIOu+}K=DElc);C4y^p1)^-d28rp`o&X>
zF#>N$^!7Zsfs(e4R!IBp!@B{CwFWk`H_y753@(^S`xz^sW6G}i2(QA!@jPW1PxN7Z
z?M1D#9#$EzlT>rqVW)@HHO{$jeggaZTifmHU;k$N!$11><-KC>{_)2T+V{TqUSQ~P
zd-m)_dGI`a{37GCmoW)_b-v~LdMT+0{zZfVW5!6h!rOOlr#+YL-FM%s^N$!BEG;b;
zZYtatSc1z_ReozFI(iZK-)?Kd9PefAYhB8uNZZ=BZrAD@PoWJD4k#WEL*s<49T?VL
z6u)`r&dt!=JAu`Hb{GTXu80~cqXW2*JjGn`A}_C{yvJY$jFqQW=pJ5!a+Bf-p1>!R
zEF&;y0Q-7^FP+EL4WhhXGNben>`?9vco6d}FCIq-K~W4gHn!`$uAS}edWZ5aUurEV
z&nNx@KP2t%hnG7P8AyMfr8`~<yB>OwQJK2J!v+`Qr-gd1s0#JA_0juhZ2}3n1db%z
zp`^qwdoH{o1!DF2rSmWk+K#Z@ndjoE(?)!Vwa42NIO*@xr_aIzg@+A)B`KTN+@LSY
zBR8tGp48iG?e_fx{@Zw)OYotO;L4@mLBhj#Llu{P^gj7f_fkjC-rpI#TOa=Oc}0i)
z?&VzNskPF7m-no)>Ph&d+=B})&V>gKZrI){_{8fE%BZC!`_=CsPJ?_u>gpMO{3Xla
z=TA;QU9P{b^ke(-GbT`St@yf5*XdU+1>v(a8X6sH^Gi$FJafypLFvR5dMQg+o4qiK
zFGEON=ZNa_7wwyW{HN_ZfAY0<vU||WyGk}tU;W+RYybSe{cqaueD!zQ=idBWTM9uY
z&_R?Sctjm>=?a|BAhbeiSD&43l!B>l=f1z<WwdWVTq@;B%BytM_99S~uak>%_3s5$
zIsF_j-+QWe@?W*%)2yj)kRz#{_j?#+jLKG+&_`iv-~I5t_TXtYRp(*U@)R>dfKo!P
zylPX8u|&*mpvxUQ%%u&hxj@7#R(9IUf*9w9*`X{n4?2vo(oX`ovP8}A62=_vmt~K5
z$r1}5^Jc8O)s0Xz*`%I@u}xi=1x!taVai;v_N6q#WKt>_pMgd7s7<N<UMbmnyk~W@
zf=&J~!zM+$3Ez$|Wa%3QPucPrX^co0U~<t9^AV<*Kq5bi0nk@WSCRLWMKRl%s4kQ$
zfo290qoU8|30x9bvWiz<39j<jH+vJsoKeyZrOlMQJOV1b<aN4E*Xfrb{n9ADe%|TF
zV!^)w^!U>hiLMr*4E5|=JY;}?j}YUc-MYEbzWQ5VX>Y&vdK0RLSHb|_<0p7h75&V{
z)&{-G>jl>T=qCE<LIZ4Q3DOy96n_)igdei8XX!0uTPZ=#om(P2i_k#hjh6;mNQt3m
z6~-_3%V;Gujc~|w^tr;q<=xsiC!zNh+82#Y1GNc#W1XuFLeWsXc}9>z(5Mc)()pz`
zzq(*zczB{s^Sqe7(MHFYOW2;?mC%Raeraj71c8Kz1Ys0t1VhfanVDToU(s{p;arEK
z&`)jYg3ErDXsCa`5}5Urch2_OM<2ac!o2PYHB%l8!tAk#uk%X0?|h<Hhwbbn8~78}
zxhXH>li^7Fd3*G^QIE6DM$JdY0Kae)en#lLdY1m<JB1me3^Y$HpM^3aI9c0xRw2L$
znHJ{z0sCufFTyWereCb#$J(8HuT{hW!ZQMv?VYW*vA!0*$UF=eukq;d<5FA^1c8Cw
z-JQ0%$@_}c^lsbR$7Ai*Su4e_jNyxnPq$Jg>?c@ZeaushGE&(DI23sJF>uLCL%Ucr
zgBxY=nChW>cEj74^ULzOst|Q~r<4nUMP)glpF)IUfVKKmcsN2*o*xteuU_%|>k7M?
z-?R-(&CM=^m&6mM&Bgng<I_wDK@e||-FhZG=lJwOo^^#M!OO@n%WCVepATzYs&{;f
zf{YM9d}8Xx*LD7Rs5x|A<N=|pC~n+0Pdru3DKCF#t2rxYco=MiM?KzZCx@FE|BUy}
z)Arr(e!cDPY*v340@LRgJR-6N@SG{d+9iZ-L*~6JSG4ivMOWOEK$wTg*htNZ(3Rlo
z@k8fPJx@K>ud|TO+hJgCH~h2lF2k*?MSL=^7BEAAXPs4jr8oq?#*v51aVbP7ZAXSW
z%nG5#bI7_pJ9W-g#y8)~Ag;Wq=CXEOWAzSY=fIeYu+*N3CxCqh#nWYY?ez_3<8-e`
zZ5C#2Wo0$_=Ck%D12c1NKHtF?9K^e0YHF_V3}rF{wKv{)v%UG|7u(%?_uJCqO4fvs
zo{K4Cp%gN3lf3X8@1+b~;X&aE3Co=&M5zX64Ucv~J-89Qf>O^~dh+CP;p5x4@3qO`
zDPDBeC#8jTsgGmw7p{O?GA7w)jE@nHjs%u?LogONOgjz=zTp?=cV%}|b1*USDSkE!
z6Il-vQv;>sP}>jO*b@m)X)jY=f#*WWI!}x-C#AZzL(u}qWnE<L938H;-R&2FlcS8s
zOpT5D8P*vOo-fY!p{QAy7dkU*J!5c?{gW_fD=W9#%8l2u-gvjPwk50@ubuMjJuRGR
z97e|VJ6FbWY+|-eXZ+EZSGhNt-(!<=c{Y)G7UKRSczdSZzyG;f$JUqeTwGo$@14=~
zy^Jw3&lLuC=~gLVgu=6JqR7;~P?F}1Gs+I$NY0WeWm4v0xRgXZP@T&d{2O><*dX+w
zd9{CskCgMw+6jEkr~lJ|uW{k^G6vQ|C{?XH9!~aryvKMHYD;;fQnbu0Un#iE<y^{U
zG#N<0MBf5mS%Ywb^^*dVE`kq)9~7RphjAPZyqhmxWN=d%u;e|SokfpRR~rWV!RJ}?
z$C(>&TZRV1yvb(M7VVnK_zBn7Ltw(~_Uh(le&J^58;|Yau#{7BgQAD<KVy}^sZgAR
z{?@{XeDKkLW}dePAAYa>AOHG)Y0sWMYV(UrZE5vJn@)ex;4Fh$cZ<un^Zr7>mJvbl
ze#Tn)=)wsv#o#75J58%KvAIbp%QJUA<3C^P$vrr{ysk<iuTMNyGq0tTbZJxf<aQ?4
z@!?K;^7w=Hr+@OF+ViJR3g0oZg}a5fg$s=*zJqcCtpK~`5)7g<WDtyGE`l%Co;_>p
z&!4sZ<XwCIwEgEl{j>HT{^K9F2M-^o&Wu@T)M4<qNwsFt$Hl<V?b~<S-8*;N&6_vd
z>gsaw6mSM_pXH^D*<$b$IyX6<HNDrq_O<V{Z-4vy8H4k7<7U>|>T1C!*+TJE`Qzc#
zLnShHI4vb3Pxw;QhW1cAz@g6a;ypz%;u+dcHbcQewgCSM!Kr-4xif+1mlwe)SNxw;
za+<(gg+`w4-sA9kDZ6qNl?dLUbfe7Uq4wf=aJevq8Dl)pdGM-n)_$@ed;~hs&lH_2
znp<rRE+NlhOhXZ7dElv>?<M51HA+GB_%Xg{r|M<QfGm^Ud{*X0sfV{D7h%`~mzo<g
zD+;e3Ua_C+rGD$fSz5wAn-BCJK36A&D7xcZGRI_4(%-r%<gzp2j3;H<Si7+roVBzN
zxMB=&8XR<(^*CHc1Qhv{;$6wt<wxlK);3;jt9TmifY0%Ayt}&+GHX0@rf<H>WBy#$
zhTkr5Q3lfKa~>siC7LyFe88w@`rH*<)_Xlj!YlO8KJ9((e{@Plb^7L=y84(8E_Ifo
zF!gz^uX!cyHP2b&;Wc@8dbKk2e}7^6iFtm)Ly!97!M7iC{ZK#ts@Knw{yNHl39l7j
z*XcU_il)^&H{1QU-fAl$QaE*+Y++_t%9T}Tm<gNsVo1X%oP>FJ`2L6OgYUf4_BXns
zWsq!!|H(i7XYIfH7yqLD&hLD+-FfX^ZPaZf2XXg<@k3rZ0__QFlK;a<)gAYXP$4ad
zRnirqNtUY=#wp+Zl+P>udy$$Du71^n&^8~H@&&@|<@oRFs(;#X^}pMSx+=T!=Svr}
zVTE3{(`+W5udlavAAHarK6%*IcGla8O#+KQ1kHOHN1j-$3n}0*L7uaGvN6>b<6?|q
z;<9FuFYmL4^&N9^aiuiUN`$=xVPQHCjt=uu51)bQn9PMKV9BUX8y?I_d1Zy!LtGJe
z^$gl(BPc`wOBm%$X09;w6iJ=wagcQwlx+SnDndh1aFylERoxVA1V)%iGgi#yRa<Is
z(+}-Pr^CG3EQAeD`IK)p$EioX+|zl*DOa7~Pq`RT)1mLikmneIg?<S)r41Mf3ODyF
z<+ly%L>Qxm`PsG*M$Co?96Rse(**ji({;K|e+Sdwy5b97`b(vvQGd1cSAX+|QWf;q
zR@z^j{(?Rao}vY41PVLwz$SPh0DA4-?e=S5daKRNPK6!{`PHd0FXm(pZRo}zui9nk
z2hU|Rr<aW0b&Jqe0!YGL^o>ymp_TX0RKg_2975-!kIF#5ob^Kxp*@88=#~1)1LjER
zr{hM!K(V3y<td@!ypCQG7}(JBwBFj<s=P(BbD=*36GeL`Cqui@T?z+=F053KLgy)6
z@?PrGr}1%TMy$562?~nPJBqVWN{3W<Da0Yql&L&JKZn}%gpkD3r7R;5<5|FZoPvdg
zds#0NB$kD}mJ@(<AxJXk(>V0cmoJan;lXxziJ>+xq~6Rzo0^`>=X}Z_%uIVOGydlX
zq2;`@PI8|Dh@jd=K0MEuaB)|3I&Coa%9#uQ$QxvHbG==h9cBD^SoKg(#)5EzK%#h{
z)PH<((4Mb7Z7<gPxes`fvy0RGev$9a6`HDZsR;S+-n*MN5O$q3inXJY!{(eVAs;DZ
zgwX*fyr~Ey2$=f8>&M}vy27+u3{y(Yl$Ret9IwIRJ(34Mvb?gKGAUy8FWe+PBs>qr
ziSW{dI0R}pZmzbeNg>Gkxo-N%!^C(|*5DuU-TDf4GccvN5IJG>#;p<n3qNTd2^4uT
zsPZ!N$?JyU6^m>NCh&)!q*%mvQm_n<n$yeTBY7H~2JQ(=dDc)2lyD*IgYcH{^_)Rh
z+U0%lMM*L`F63WVkP&=STu9(sf25hY<u)=#K}NXE+o!cz{9+pWD&s)mbAEc<PO~l#
z4>mI1+odR4Uw_hsd%S)7Zu`=&eWfk0-cDWJ0|Ol5Wyb?=W0b8GHm@sDE>91du#r5h
zdMLK87%|qoGANPub~jV+%kZDPAD7am?$xC&cyDXb+?gkeYD!IO#5@Q^-p@5kmUf)-
z4%_+hUfbDPZyRYduRI~N@R>p&js#EOea}wzYku(I`q9TNSL(hv-w$8STQmEy8#n6Q
z7#8<%9B;|<GyJymaCky8DkGen3}2h}FWzjEvn%P}bo!CKRobt=@FHcH{*9#Va3Y~U
zyfIncY#Fnc_7U(XUYZc~%DXM?<&EQSB%l3!GIN;q{R;2R0M7YrDVxAQYnGCJe{Uyq
z-<2MW`{w7DGtc<_j6=qhxLa6v7X1`R!$aB_ymAFY<pGpIKrRS^ouB0#K_04<LBex(
zg{*h34DYfg%~!#r^NvcX9{7C;H}9vcixQ$6FKdQJ#{BG3+Lke!U2em0QOzZTh>qLN
z&i1puC)$!QcC)KBrq)#n`h$b4ZDpbZl&9qxa(0omGt+M0daX?_+)BAK!GDKsYhx{C
zQ^t%2r=6B34kb4sH-!OwXWqa>g_F%1bhZ|S(9p<2@YhP4nH7meon0a5+$6KcOK4&W
zuAvOdo~rP}&9=CDr;Sc#>?h_@o;_S(acsCPt=?)kZoS^-7H+4!`K)JN55e=PUcEv+
z0OCL$zfukz?QaAoGq!<QidLIwihDGy<KEFyX5G4Vr_If;_IU}c@@xd!<;@tlzj^C!
zyLsoWQqY0V>EPPwjJ<xfAv6HJ;VG1{x?tGwtZ<>W*gw`8U4b|CGyI&G6oS$GrpbYu
z%ZsDn=+KI^`66&|m~~2&Elgu6s)oTzYp20q##@-)vB`{ca8ak5xq^`F`Q@+iw?31G
zgHsk3GfopXvu@K?!vEm43VjJkGvc+^EZFr9`k3E!p`F12$6M{+{OkXv{lEVE|7ZF?
z-oEtZue3XN@3$$+zsy??<GWCE-6}kA-n;AuA7B!`Eu+=!<Bfqe_VD3HrG%Q#o>rLB
z`Nh@RuTx@CLd`8Kx9Pc?fuZ^Ip?gSn+z3JfH7|Sl7b){F?>u|@puPM3?-eZExpljg
z7L<g%n0Yig^OC`iwGBtZLu54{y#HbQ=GVX39zTB69zS~AKK$T=_N{Mzv%UN7yX9rQ
z9zN%z2anpu##VFoozT42F^@Ori@~J~Up=Gn5W?CySZ}=XM%K`s3Q2tX_N~H~6jKzg
zJG)z<(@)#?-~Bi|g7d|B<evsVgVV#33;508*Z>OKhhIBBX5f>xmV4&cjF$oItIWGO
zOyV764B&s`X)XC~-s-&bw3T7i;sOH!^lz$gPkG0N#-kz1u28pl#lRKW3EXfP9)dTS
zbH#X8F`~g&F#d%*JI$d??K~i*77)W1l`JM>rT;wP;8y3#QpS<549>@*_!4>*-lzQ3
zcXE_c=7o>Nhrj{g06xZ(P<R=q@_^I_S8&@@^*?kqjWA9=`+2C&H_IAMeU!ucj>Z{}
z=^4f~yxkT8mrLzt=$H_Yj8yQj@C;*~6^hN|uaJXTQ~In-UUHOe)=}}esZYK5rR}Yq
zjNxG!K;fCu7VU$J@HoZ?zi!TWg5d-G2CHOe#mnR>ql)C!U&}!rurN?^!RJ2zT&n#`
zs9`?6Q%2an*P7}ytEXO1EnemMR-gI1f|0=|nxZ}CzPxPFfaH6bwZ#i@GkqTHUle&&
zYNcfP^u`Ze{f++f&!r!Ev!_83{;T?a<dy$^QXzkd^dralFPUm=f6D8JlU0AM__|Kl
z=~pm~gc-m8_FL_9Z@<}Q=I7d32*zm$6=GN62ZKf_t1!>PBW}ES(cb&s_uGfxd$(;q
zUoVSI3FGqC7rxMb@4xu{_TT=Cf6;#TcYe2A(uH-P;Bc5J1dkaQAfb*z{f4UZnP;D-
zesAzGNPZU5&llx%!jtcx$wyp<sC;rS*8uO;#}6UFe$VYs@7KFm?+?hEoBl9Wd#v8p
zp0|&mK59?0Ip5Ak`!pLP%s=4*0gEw1Bni7P35Xpc-7PyXpo9QD3=4}H&&xtL%xkqJ
zgcWm88}ec@K8;aK09ICM$-@%KCJd8=9vjBN`(*`~ufpC`ow=8vGF1OCj^&9K1_1+T
z11Sf_f%Q+-rOq&KghIxZ0#3PgNPIRkvlY6-p{+eeeT6m27!b(xe)V}*2Wu+tX@`yQ
zd>AR#PTIn|j3-yYUSLk1Vt(>2t0m)7OhVq%CYyQfH8%REPvUrY1z#Bb)Qgcb517e?
zFx9hR#HcU#4-V3&4!c!(B_mU#bDgf!b^5!J{#F%Vg*hPIpI7>M*7b8v16p3jEBTjg
zYA*B!y*p}ibCd1Xt)<W#LjG*t%YY<rB`-QAl>i|ZT8duPc~BmpfuTt@%IF>IUNowN
zczG5wkqtNnK{m~KhcO7HmprWbOQCi}%PBtyjnOwWlJEn4a=$CIgsB>@KG?*2*KbA_
zyeOVLeNuv7f&!a&{Y7s_oO^?Y=dFPf4E?nCWx4B&gQL(bN{P7zXI{;vOcu@~p=AW`
zPfIW$X~Srqp^217CrA6a8~$TzCjIPZ!c645uykmo4Zm^7N0PB5c=P^9yF5Q?=coG_
z?`~8=LB<1zw+Q1X$`PIsqZ7;NuJfY^0SS=_L5E6EIW{I#SPzTU&zk6h?w+pdq8OQ&
zZPQZ(U3+bOa;6j&`v=DNqRKHg<>AhQ;)DXzn4h-Yz5NPT#7m2yke3#~Agm><s<WwH
zI`8wilp8|g-MV?Zz4_+bbru`LAN<(;*YDRkB<90eD`o5zzNPpkJW~=lGM9ub1USBf
zwSz;7o>!%`pctd*>hnHU!X1i2?Ipm%KT+rinKjgU_&16(N=WS^tRpbvLC4br-&p5o
zB)72^+KD&NlzPH<QA82y6MP9l#xRLMae0|gl+`vk$;`eAf1f^q?Fv&kJYM|Q-tKn6
zE?(4na<+<f*Uwb)JMGTc4%%MU^nvU0Y8x@%UD)e<pYp8F8si;f9Lk%c6rDo*O@!xj
zhU{1wX;tW^JhA5R&fOd6tU3FJmjkQv*I$3Dt*qY8I4oo=Q%-FF0%d_i>%@iM2E4Cc
z9=82mo(~5Fn^S=wdk@A`U>Cd-?6W|>apP{gcQ5mwvMF#U!^awHR{Pehs0*c_$#_vz
z;6o|>gondJCQEO7%BCRQ+*m8G4xuWcBWH7Ur4!{4<)rZ3X-D4S4F>LcqZ}Trx9@!G
zPb(xGcwzZo=agm6trv<n24U9a**W;-=_NFH*2+xx`lCd^51$+zw8P`@wrLX_wY+kx
z&8^&Pqmzs2U*IFKTVcwA(~K9s9;_(8!k-sT6t-N5tA%{;`;t-0h6)ErIc0y9o@YF}
z7YY8pWn8*~P`x*9+|1f@CSTug*yANv#@-yKAHlzb(dN?HOtqCqANyHB@FH)ijVoM9
zQDOYMqO31PA>Z~&AnMF8?aqK@on`QyAN4Y8y1Ws3f5s>E#~2W1XT8iVw&5`f7=6fm
zTpTI%6a2_XU@YU-hIo+;XP)7j%e>pq1+?aNw|6q%+nK|?5(3X>O_ui^y!$eEX=gpK
za8@{BWofncC4~4-o;+*MvX_Rt2)-$R;n>r&;Iw=PSaA42sWLwPD&xBs7<<unHrG?%
zPKM6d25;wods5*??bqOAg74Y`4^igu9O3x~54+%OeNLX{JjKHrCos*U=QQ=0OCB87
z<iX)y&7-;F5vk4dfl*4W`Gq?H1LFg4T(nogUH!ae#uurZR~LGav7ksp;|KJnpZ91Y
z?ZZM_k9SMSrj5=B9IlX}DffcsVdzHSX>6R~M(|nsrVcP;{FHNX;he6%#>^)?O3BTW
z%=59(f%2{l>{IZ8TXj3{jpCiMx0H7jzJW8N3#Sg5ha|WSp0-9x@m<O!eNdL)>g6eR
znY;(>B<uS7-~C$qU;pL*Cvd#hzWO`A-M;$O-)qY^Zl{jm-VufaDI@f`<ASs&xX)Q=
zytEmv437kV1{UXom+XZpcs~C4<K%tWUb}a<pP6oLrnQ=hCQQ!;W{2l;Pn$Zf?6~=$
zU=sc)_(*+}(88*2Y;3d}t2aXb-pF{3R|vwV;WwP)#F+H)qbKd1@4p)!WFxqlp~PwX
zvp@Y>`;UM8pW2hhkJ`r%KCUyk7>`iSP^bvI2kxyk_>_Er?1N$kKDBOx3JSIuo@}NK
z&M;*NY=Kkgpcj-Lj~+hm3XtvHQl#<Vf*;UbycgPgP)a4?65&WZ<9X=SVR=}|qrVhV
zJ@Sxo5osnf2qugL8k7d{>YAUQt9CPPSz20b3v+XAaY0DZJMHG`YQZmh$&0L?uX<VS
z;*}))q{vCeU?+UccG~Scs($8F=g%l-g!`jVQ;)vl_Xdh3I7_*B9eq@nzUT-3wv0Yf
zPtn)B3-?){LNc2(`~f~jx%d|J(Rr5B)4YOb+TxOPO=sHCq42wDPu?%MPQQ!=zI3KS
z98%_qrT~8k#o5<_HLfp=GxVX;&Dl~48aoOsbeV@5uX(fuKZLfL13Z;}k>w2NEq;g6
zn-`qpI`k2Ly>;d3CCQK9vKA=lN^z8X;9nmGVHv%r{BA6s<&b;w`m`qSbN%Wo1WnWT
zL4v0~o2MsVuHLC-<sIJ#pQ)#CQZA3>xO&#}_DnrrFn6`78{O5j&_c@h<D;|o?D<B)
z?WcABP#RS5qmpX+#}B5T^iI-`eCN-lpDx$mg#?nX6<^osI{m7p;IE1K>GsAK-fp+=
z-wi>TECw2*BMp>a;!6t=vA4C=zW2><w{QHr|Jb&Jm?3ImGH%|y)qejU{Dbyi|2O|_
z`@jB^f6~748^4i_&y6CUwJ8m;v_US(N-8a~a76?V0DnFanIbNp)!q8|j8g0BUj43b
z{re}xsS}k!z9MG%WT81IqOwU>uZQ=kcGY3PbtC^P?~`Az+RHG*K6$~NWJB*fucyzS
zwvQe?XfHO`+Hn~3^DwOE*)({UKn{bTyQNsV5}xQe!GcY{c4DG6qXa40w3cut<9~WA
z+!DruqKg%34}U_yMj&8=Qyx$GE320<6z0l3f*o!Bl*miQHf(vACsMQ$SY%%M#wyQp
zuM_OFCVjMd!hjHvNCgAw55@)4;@trw!HO0GrjFWRPT7dtSP}TKs3+{|3Nu0j<7BSY
z(JhC^iV-&Ym@|SVR%SLR1%JjQV~9!aUS55?Oq+4EF(>>iW;taTQv$NCFdHsG4`GTi
z0fS|+nD@@ZxE&rJx69Lk@@tSfV|Sgd({=hgk^Xi)zCIx+zdGrs>T)N@f5|kU!J%8P
zT%9%%gpZ6~hPI$VL)oy8wYhocM~)V4uyO3l2|~|q)hVo%CJ@pS0xOvE0iEz3I$<-;
zD~9k4?eh+ypfYX32?EeF^o;;|pxnZz5K<82pg#oMz7vF&_x5<-oOiD+RfjIscnBeN
zS{BmIIwHVj1cEkd2dh|C>G&_3c$ac1Ixs=nOBsW&5Tc3FfUs?3__Xb8K5LI2epG_2
zj93X_g?~!(8qX=-Jf%oWe`=J|Fn!i6{Y3vM1RRF1{B9}Vg`*c2ln|St%Y^-eY}3im
z1$P7^&T-;tAxu^YInAH)hY1Pubz~;rr&4a_hG$xp!3c-v93I2i=nLALwxld{TKK2c
zo3Dq*bOs4Q^sI&zUvzY|m-c4<@JP8>H<v{>zC@ccU&0>hv$J<L)?c)}o$VS2N*ug^
zGdNhF56Z;9P&nYz%rjn#0;~&Zy8s_-m0~jkOBh7);5|Y&;f27)K<H!c2+d~G4s*>@
z-uYMr|LVPY^G;h>q#zI~EPW3Q@J1BUk<!J1`*<kEWB6=gpIADBcYGfqC*Du{DQyW+
zcs^`zZ#HL?@Mxh}Q-(Ygf9IKBuv|RS)N*>1wgu*ghc0S<36m(USj`LPZTyBud09-i
zSAi{IiMpZEW${&nd*fpnpHhDDTocM^tMSq+?+Uyk?~SYnLU;<Fm**$7Cm0JXQbH(0
z+bIaGUtVweK;g5xdb2HMjMUG2iO`w?a3p0B5*mB!3!Dv)zHCSPYi)aLEqH=u`TaIB
zu~LFzYoTM8tmCXR{Oro=os8kFHafYG_RJTI@k*@u&OCM{i?DN@=Y>Bk7(6+6aIO!}
z6?|r$OBdc-7)6Q!;mdB{eyvRlOE*Lr9Jn7cH!o^!lG_6)Lc3>AK5qZx|M1_pKlzjY
z)ZTpSt@he$_Y0R0f@!jKPr>7?E1sWS`63J>tNzUA`9<L!uy1_9vb8N7sI|jWtAz6z
zvt+J)L>b*Bf9(;{7hzNhSqVdVql}C#r)~2ESHj7`t)*<JRgk$)AMo$IWozAL%*HZy
z@FkCjxtWCu=N8PDeN*r^VQt{HdrActq;I?ng*3zKJ3oqWzI)Q~+Tz_c(<&NZN`c1@
z)A!?yGleOSD9Umn@%oun9sdjIdU~`QoVQzEedfkmo0?uu_cIO|m$T#D>c8`!T6x2y
zUoV48MC-_WfQ6CK<+OF6#McP>ovgF;mxmRYVW_vgvC(#RC<FFV|5^I(SOA`=M|FR5
z>qV_6;Q*(n(vHB`P}+TPC~C%0@=vs>bW&Kxjg4o4dEu?Pf`u}w6wjGA@OE*s-9C8t
zyX~V7z8ji$l77vmuRL>hODP98)mdbjv+5t?5Np4@3WaJ+S-C%CzcrjSoIY=@J#9||
ze{g9jy)&n1i1iKjfrovUJ(qBcPoF*tE!YiQ1$WL1u{fXdjbS&y2qXnwgCjC8&YjM1
z2V%0<I>=aTKF>UZJH{COOrpwA$|AUqfg%q!#u=eKSIW6kD18EJXh-Ixl&;ol=;TCK
zu1`%7ppORktY_>W2Z!zj4)klhjn6D*T!IU;W+=UTy#-up-(}X6bsO4yew6<AkfWZU
zg$ze{ChAYoO7yFgrN*|GVND1d%flH>qa^$IgLm5h@xTAe_WaqS_WOVEyX`lA<2MQ?
zOom_3{tCgE`gV7p2XF2Nw}+-?9r1R}4!MmA!PzK$gF8#65c&jXW*xyXl%xdwN5S<B
zJ$Pkx`oz#!`>a>W9BY-gZjb(gMj2mo?d-^Z`)~hk```Yze-)TK&bt%s#q;&{Cx7;}
z_RVj7rxZ8tKX~}0eel79_WZ?0>bWRm3+ocRJBx5GG!A`IKd(fJm*wT9QX(;S-QPQ?
zd51p8GvHA5!#@-?_WjC3-_pQ*r;wV+Se&PCyTN~y|BQn1c%y~;!4LW@dpCAvEKBBs
z=fh8R5AA^xMEwIr4m4wI7($@qMfbudYR}Tre1+rV(Z~p4C4J$kL{TMtAJ4M)-g`gm
z=2<DD@QW9huhK^$D0hlRQtb2W+SuF<EbazBfUl05;WFM<s~I0_cslS-`GcOD+YCd!
zPr=y7rSoY;w*$Wn3&7z(X(CY!Q4snx{wrz!YUYJf4L>?HHJ<W?W)&uRx%d)kH}jj}
zs<O-DEo}oI_|oaw@N~&zycqeoj9=%8vJxl9+iJ>kZmKiPX7EDje83Lhj$f{G)Ash0
ze<=LOLinAf;$5s$?KBQx#@LuAV*}54SHIPTcP^z=KKuN2`4!%qGOlO~xJ%uh)%*v>
z%z2k7dZ+XKMenLt`M-l>-7DFbfv-FL8vK=vGFN@3{`&sOXQ!c^M`S#fPjy_qQ>ohj
zX*p#$k}UQI@M@iZmcNeEM%()OW}Sz3A`EvjhTW<1K6`yq){ndV{m{Mq_vh~YxvRJ2
zCnQgQa*n@D`jKt;DXzat`r$H+_O;^cI$ft<u{1n3)D~{vX!qazTw4sG4E3n{5Hn&o
zmfw?MwuVDMSmS)-YhP>ce*IhRc<Zq8ITZO1{=xsH{gZ$8&)Przr~kD5_E*2!-gxcx
z5W@K)N{}`JkuO&WwG*xqA><zgNAYG!eN)=+5Axc0`wN!xe_g&yrHTqcK@=)iWy2sS
z=jywx`EO9RM>Y^WOTG1J{Z)r&h^5jc<-1<C^K2N8!&q)@Z?`8ep0>x&pSIm>0M4?h
zIn8Ag!&1-Y&;|k{))h!S3<ttoo>F0a5Jy%u2y<DzW~{Z_?;d6(jA`D{PC^<RyB?-N
zSd6sO7*ggaFYgo+nKmnjuna>~g0wKl7>Vjfo-1?OA-rKYFcjLxGFKjT843qI*MXtQ
zBfm`p%O4C)N~w0s&#H>`nQ$@QrMx1rJ570qEW*OPXsfv<L@~aGTRSi^7h&iKtn|U}
z0pn_ug$XNRU>IioQ{TlaOi>rilu!{vkODD)#Vn_6;SfDf#WmhGpbm;AXi|m^C5D%^
zcZD6voDwiyoOMNSPi7wATmi;)x=z>W7didh6kosl+dnGQVW$+oDye!4;r|V#PL;3z
z&<p~x;aA}`ZnxK8za5<ZGBm4u?0ljG$fp+7COI^vhbQbbU0ZF0(L;&R!$4SWqm4$Q
zpG7-DQwWX9+c<fJTVf@TmauRofbJHgp*+dctzf$#3f;BwX2_ww6N;0gqjq@6Yp93r
z?SkXHo$nMd%F|CYs|#_uB{X4;O*Gm|5T$-Jj}T^leyL5(P=Gl5r+W|ygT~;6(vAl<
z!GgMZi4kHFFii7!m|hB53+>Msb&s6zU}ICMU&u0+!55{RDQoZ&oOR(DWx&Kl+Mn^P
z9+XlHuh{1(E$?HU_qw21!lRtG^Y8+Mh-iB%YePaIQh>ZVEde#}f??-J)!2AntA^SH
zf$+>qo0wW{3(Gh9c}F~Qa+Pp7`SoFaW2-zU@H_<Egj6iEZ{NP%?%unXIUH_Jo;)p!
zc6?J;R5}Z6p?i^JoSn<VQW`%(V8+|33#xip9m*~949*Bz`k5PxCFm(W@8Y8E@HET3
zl`=H%P!3UQu!eRhJ~%$f-10~lUXcJud!5(g{4Gi^;TLyzw#w59Kj=&)A;pBjR<Cn4
zmX@7?(a-cU*A<R>I5166oSEn$sfDr{N!^RNk2eah8Q9v}+YVgp7pzVbQjQU-dJmsR
z@s%|aewV^3AeuU+T6s*3%;o+>=3=r<hi}q03KwUBh(bUZ$2+eJgTc^2cvecym$goW
zS)vpf9eq^-ejYMqe3f;0dU96bpDJXwxf!QG9Sa{A-Y`pTp~860WG;mLbiT}Zn_pN8
zO!TwrtUWxaP-|d7$VNP-5PRcOymgE{Az}E;@aaN8Zmd5qp49jgnwLVTgz0$59_sX9
ze>eR&OnZ8GCqiAqJ?jwu&}V|_A!i7s-;+~m*PQc<tRv^G5HPOP9A*G(jKTc#r;kcO
zOW-*^<y~G;r|mo6`C9w;|Nf8Ky}+z+V}!!R*Ozd%^TxbUz~S=JYFk{)el5?ORd;Z3
z6daj01SZBP$}?x(qrpBRI3#0M9+mD;e5LQayyO@!-aca&9E(3Lub|+I4j0*%yrgv7
zuJz&F(b4eA_Ci9BhKC=`+U0#xg8Wi&c4GjrZ~sU@f0(>*n*p|Ou?I9)6#CAH6H<=C
z#97Doggv~b`WXgfk2DedI7T5$0oMwbjg1E;oB2=|n7yp=+S%U9T3;{4DQ~ellP2@r
z74sAghi!M;{GX(*%%Q$!AYNU<$HLTS&syH7;08`QO@EJSj+||lit;?|OFRAMfu&E*
z;32%WwkWk;tgjXRg%iN`NN}d}^(q{F@W|52QoH~9>+R0nyBVjU_Vmf4w!ZnIa4kAS
z84RxN74^~h4rdP}oaW)-UPV)JE)rz`ZxTw*8#nK?*~Q!Wp0z#N_eR~b3w#d+|LnGd
zz0GtsaK$q-{pF?j@X>=jZ*67eMuo|vcmh85cjyP)1J`I1`b0s+LvkcEg95sHTo@}t
z-_%y{I$C?HgT2t8bKxWR+w<qoisnsB_V&m_DgNoZV09`mHIy+6eF|JrN*(20d+DM-
zDgd~pXeuh*uiWciSeNP3v-a@eyQzDxukpeSJm0Lrz-r3k-B#~g7rF2i@E)%<G#R)u
z{@e>5I5^k~S`6+PPd@Nc%2F^QC1c>D4Kc0>pOH1xcfBv$```a=`~LU7)xPqTUvGc^
z@Ben@9&MgZT|)QE!+Y8JQjCaS1)m84Io{4syHN+aU-FC4(aV#=%EMrT@ym-BPrGpd
z<-+Dpd;0WAoh4V}m~~rc$T6l!yI%$G>Dy><mom|K`0d%VXTdw)ZD0G^H?Gcu+RL~-
zsq4A(r-D~1^zD?;ge$?j@NGB9fu|Oh7Rtl%wfpzl?OV4BrwwqWa|bDTHn+D!({=;1
zhlM*!Ssoh0`$$+kJX8-K7@E2mT6ndWFFI;19UP{=Cxz!J3BclO_g`yozV&9UlREF}
zv@5LOF7JC6j<!a<*US1;Ap!>ARrBmVyzW!I<q3Z4M!S>sg-;-pp&YZu%lkL@?CH~|
z8UGhq3qmNKlwymgS9x}&z3$`tb_0K%|GNs&oY}#5^7^Of!SmHvEw5DH^}+jim9^lH
z&CQMaO%CHt>(BhdQAZgQZ6KFpa6`ewyNc|@7%*VaW}a{r4z}8GS^X>?HDxhu-q{uc
zv_Ershd~QC!QY@|cxGc!cqwZKp4E2BGvNt4ji)rbT2tw%d4|8L%zSq-DylHhb8~Gv
zb7K8_$DEWrD!9RS_j+48?Gr`Fx+i1O2Jphmt;1SqUG8;dQgB1w!ON5R7>gSBbfRQ(
zxe6!h)F&zHFuzN&1|ss=e|KD?Z?f0D)G>JW>E-#S@6=_c^AD|(+=ILF<p=Xs*;%7k
z-|<9;hbQgPqpaoa-7Nh<pQ>p;aQXO)(x1!oQ@r#SXOZKlsr_#@eRdrGstL8YR(xHj
z>+~y@#%HJ7Yj1z9-M)XnO=r{gst|j(YJ-HwLL^>g^S}0Dt$pv?-)`@J^E>U}`A!?o
z=J(gW{Oj!>{=<Kq>mOEF!!LZ{^F_dI$c0<+SBNQ}jJM&+jxuw<YKOe~cNZjIyY0H)
zasAbkeE5|@@{c7Z!icEISBS5kU%itD{hqq}qfUVOec%1mB%hct%w;H3ML9q?E4g2M
zC#Ba#+Y1BnVl$hBClA}BXCX|xTcz+4quHf2R#`R->aCD1VJ;+AW>-T{LMV&}fx*Gy
zevO7h!3h(1+hInESf@<{R2TwHq8u8W3r1Ol2~sB^i|#61KGvlD{4GolVH}19Bi|Vb
znbc7(@feBBbqUtO6kxDy=!-c>kv5av%d8)CeUaaUb~U!FRq`GtOC1Cfl73NsnOjz?
zUHQ~KwUmq9r`$Wv21?&3;Rdezm^vY-F4I@^9-~u?a^52pF-A43DU5fW@gOwh`K4Vo
zjmhf_51XeRDkk}ox5hH#YmP85HkcSO%((WP9kHzXY%&6${k%@s={o%ark|hU>#JY=
zx$w#G+~0Pp7W@sjqZi-n@P8R}lt2m|x9Q6!KD@@AyDRM<{J~e-7r*qm;PuvcKQNF$
z6KslkfGDi!VsJgWfkxO6cl7O9=NoJ=OW=ma)uxiCw9q&dL7~M(2MGsyNF00sfhcbS
z8-DHbp3QY{Cn055s_759N(n+B;~;egAm}+B0Bs#ur`9<k=__LnG)kLn%qarUP{PF$
zyj-eZm>%aKEVr9?@M~XaE30?Q>w^%B_n7wHx%XyUSbjZpjSy-seB0BuyS>p?Z`^F-
zGk4OLxzx!bmv!%z_w*>`2T!f=5oLMW+cK0k_N`&Y$jA(J@AL8`C?PyyjKL*XBNQf-
z$e1X!gpv3^;dj!Xj2}VCkaJSXf*sFd3|Xo_Zzo6Fq2Uy4>NBq0`nwBk`&C{;DV7qb
zlqP8(q2GfCkK0McY-wqwgfJ_sH`?k|$fg@NZ>6ncCE~h$=T7^zU;mBvT6o<%ckb5N
zG%WAU)7I8{#!;B+v^{Ww-ytZ%+wiQLpI>Y<lnLRFD5+MKS5qI4knkFrUl!gJX2yIu
z_w_@Vr4rVHX>fTZ0H2v*+0DbK3-Zd4=)##?yN!p4uwt4)SX*JXu7oOtu7qSltj*0i
z3xkk-D&vrR;d2O7OKCeoQM1sNmmMlUS3DL)2l##Q;#n!3N*IVw58u=U<1gFAnGlg+
zoZ^R9m{3#rqnY&6y!AQHtIaW9+C-T~!BU~!gt=@J>3h~q*6dLD-_EOAd-%@F@Yk&O
zH`?(rkEo?KcBQ;EKNJuZx^sH5O-wpNMaU<LK3*jgO$aOJBMHA&3Zqfsyqt|=3@Fh8
z+e5Q$BI7!le&KyfnVRv_zaGx)sBMOa-`aYS_FR-V0%NfXVMY;f7CtxQ#N+NMKL63j
z#RC%}Iv2`Zu!@(={qPWBIt2-#X~sVLp7GJ%UqZ7IrUoe}qddumd8y4;dFAyqW<5@~
z$*JrIPWRe_k3YJa=gE{|-Y76BuLwJPSn-A60O1pP4l#x^pK!zC;*C;<w-M+1%!I$*
zYTMP%S8Xa|R?9WxOi4;PWoY~^kFFtnz7S!;nYv36ZOsq2^OL}3>Nv{&2W;5iq(jLo
ztk_22>mYEN@qm~MU!)>=f)kv>xW2j7o@M{IANbxsAe`OIJK@ib74N5E9z*shaMDS@
zgm5>vs=<MoM(Naz0f*9PYj$ROuI8O`iZ_&V<tSFIrwY*t-)6in(oO<!cvKonKZTLo
z+IkY4O4-`atr;1itlbGd5oQhDfa5jdvYj8TwUfimc6!VUb3g4!`4^|zFYLATwI}UK
z;KaP#x&2z3pIuJ-GWNko&N!OQp8C$+*W1#<jXGbi8w{qL)BQFX*jre>)r9WjX}Q0j
zeRbNra^qIPe$fbR&fj1SIxCJYgBQoW*Y3B4C1-Lo{BVYGaMXN$k7OMN7W+I}sF|En
z(p<F77as@4zMpy7FMKQ6$IZ>p_i&#nZ+Kj|T3*2yp&de$ofY1pOr=Bz7w|lM>|J<Z
zw0?Iwu2ywY2=jn@_Vk0`HaMp%pU|t9ug=??z1Mj68tzI?z17a~Jg1m>5twwgT|X0Z
zpb+fdXkFRLbFR}1^ta#V;dOYp(_TD#S_(no7<q{bGl!(XnH1ICqfR;M>=8=98QNFl
zojkjts~>#uz2L01!h5`a7==`b#;dwtW-Jf(*8_hU?cm;kPH^AL^z*nB?9)@@?GOIo
z{}Q_M`HTTCP5t8`#OuzUm$&Y4;f3<T3cWnYIsp%d&VdRp65?_<>!y2b={HZ@?pXtd
zc?>f;efaP}@d`XE8Ao(aadTnpx;$ZbYoopY{=4nncfa4h_r34s`v)20m*Ej^2mj8e
z{NbW+=oq-3b$+LEa=-TIp@G5|zW)0Cf^XA?PCA>5K?+=l7NTMM2S=rl+1}aBSf`HF
z<>2%A@|s#)nhzeEEjZvAR%aSc!Fv?llfh>^l_^h8O1Wh3&SUAd`*%tywHP}1Vr{)W
zf6j}LQ7?H!mt}wxyrlbNEj-fj1`LpR65-PrV^EBAevm?KJoIKdZJKVkZr*6~q2)q0
zQZ6M+@uheL3Ui)lcW&P)9${^b$7j|wJ})>s-Ao!T9!}dUY-tkQ)D0%O;;9sqkZ|&o
z!<3;?_(sZRWtm5OrFT}!GpvUXr079g&`>xAU1X@kh&cdQhBQ2@baEqn4c<$C8S#Jt
z##sa<l=E9#+v!s`veA&D*YzE&&ty(0x4WE2JG2u&OQD&4g}(p*|MW>jK~$+9jDG+C
zr6yX$*h@W%sPn~=zk8^4k2>>0Imv60JPV(3d=-|`H3;oU9z!t*W*p&x@9Ofd-Yzi1
zI2+u7k&=s5ySp6i)8B?6pSiTR<Z#vk;B?s+<niEMUnia4Li;|s@9*ka-<N#$>iMPm
zuqdl|Wn5i#`>L#5gZJtR45-UXczp`z=g-#*?~;k?OO;nSKD~aL^yl*Z6fahmLA!s<
zmH+<2Jb%;IXVZ^*{-drRPAL4f;_EtHr(dNMT)TK<xxN0SFSOM=H?s+0br+@|BZf$Y
zIEr1rv$4@WcsGpFH^0&LpYPN;V1NI&f2;kofA(LsKluGWXkY%dUu$>o-V0I2EC#{A
z;VH*oiXi^ze}#B_A~#9-Z-9&Zre`Vyl|K(Ceu%+-AX6@M;AJxR_~t!tRn&j=cl~+f
zr}umtl;i85*sHwtDW7|3zp~x1QQ6N1Z!HYO<7bcC!^aQXljqOc&OtWen4ILl$c7#x
zfPmRV3kz7n2C>FNRWS?>7Z>sm0mtyD^Eixxur>rA2p}c_GfWWHz0&%5A}q~qD$A-x
z|5GN0u7ps^PniT58eU;P!hB%zJu4Ufymt|%f;F>torj>FlyXfdjFIYtO+d|0@~Rh;
zfq~V3N-|6thDQEUj-@|b)NP0`AU(WEw|*Wl2o#{=T+5HS7|N$4#2^6%VWKMM12@{>
zJq(huG|vQ+eRGg;9Zw&_)D9RLpZe2<3)ukV&%LsoPkzjk`CwI~?gE<nl5+D-ZKA_`
zh)X@6K46%IVmLeLOzNizF!%Sa({;K|zmVzggyO3<V1Jj=&!xGcb64M_GRz1qD8XZh
z9U+#mhQIY&zuw+{`+n#R&xFt`!j;f68%1;i4M0qIM=Nl^M-a-RMao1E&`s|SEIA3W
zY|Pb#)}nPbp%f03O5H%kIW5~|`Avb)1uNYG7!4MVZU5llN?FoFA(pofKI)2AvQA}H
ztc_@DswhEC3Fp&S0%bIsKm{#AA3Gm#6>{(BIQ>drM#kIp+;SV6xDj5$nKiSaix+uz
z+Gc0wLx-JNGoSJ)AZEfFoY(W&yfa$i&TRFJt+C+o;f#gPTCZoIUg8}X#PpwCi6tzg
za6{pH@DUe61sdU3?ffKtKRn1dpr41KnCZLuD1}wVn80%^Z66KIEn(q6$Z5k*@Uc~b
zU_w;h86|*u#oAvUjTkiKGLJfarno5sD?-ckg=KQ_A7cc5-3tmYQHr#TIX>mhH{Wb;
zfBy3of}3)T0trv`{Q1+ivAJG(2~>D>6)!iI-*fXN{KGHcDVzhQ*FyU!pMVCyjZgBD
zpRW4@t9A2kZnc~jB7?0GB=SB9Psq~{%<!DqFM$`!bqdaYR*~`v`Slau2xciMD7grs
zXQr>tyUKlv4~kD$54$-N-e#en;g$Lc=)k>DN`zH}*5#?6<pvHf&JNnc2k*7DwWlSB
zqEMvx5+<#eO#zp=dv(%|b~lsfF#V&9Vcay@rZWD+W0Yg&)|~4%As_|BNO-euc%*Om
zUxL)>xwIoN3{H)?wKh7gKSBqZhwzqx>Ctf^@6;WBbx2eLg39oG1uL$s-_pi_t$feg
z>K=%JuhiWcd+Xg=56qN+KF^D9zpOQTnDuLn;0~}4Z<IG@)~xmmRZb8`kiNIS-+Af0
z>-`E_p7HcD?<hPtub7SXjqsF0mUS5J<K?Wg#Zo*Cm7>Ud6bLWdRMrRq@V$F?OR$SK
zRp&q$NFh$wZXQZlpSJQu+0TXlruYCKUAa8cjt<t_cJ?jufxA*X8=N6?nSQ1IV56Fs
z!~LDi{j;{aw_SS#hGUm!6x!Lt1a2r=C^xOGiQqB{3kG9{2Vh~h6fdRZvv%{o_uw$b
z1mojVZE1NmYj7rg*=^6V-ZY9LadKiV?KOwBN}4cp6v<gInFHB^PlH#CBhLv24!jV3
z4rg9FZVyhUFspf^NFGVqv*nF2oA!boc)HFo3taE-ZM5~ZN9pHT=69jiku#u5@tk@H
z!`UR%Sf)+SpL|pbyE$H;DOW$wj*n}s;4Sb^>CU?nJ?Z;6XA$$DT4^_Lz24>)Zr2*2
z;9g7rcd~|hSg!23vkuSBLOarTA<veUc?otUg|>_Ze>&%|l+QzO(|Mgo1t*|kg!x6+
zGA5KuqZ14Zz!n9ay_CI<xk-b9^B+BUzde5Re&IgC>GBrK{LKe<jZwx1Z`FJZWzA;1
z?5TJFQu3CfD{}{o)Wx$5o`Fcv1mjiv$mFZ_nE5+Ac%FG4Ys(p<`K<Zl&<hGt{Y1aI
zN2d06!M3!s^{~D7{x^#*F3c~r+49a^D!f><DRq>>x#$Ua>)vZTfYD=cy%;=v=kC4C
z!A(O@FklT#OsG?c()0)XRNdNp8Cc%Y^Q=#BGnMj=11nDpf58Xpe|ZV^C<E{k$H5KG
z%@UGQdD-g*c6oDeZ*8@m?TzvRee12yx9R!Ste?58)vP@{N8r>ag&sVaQd58T#A3*@
zpNlu2z4x<cPYXWb8oUP&P3uXUW`ZMc-?`WB-nkpPwp?p;acMc<Z>0~EilNcRhvf<R
z;K4`jyWjp+d52Y`6Uv05yNoBMGM9^OX>qyTx_K*jZ!~x(bDA+?Kmi7LBJnU~I75M9
z55K!Br0r!Xt?)aPAUnC&y`Z*BIfbtAqWawDUat^>Q@pv8Ut1}doKwb&+HcBVIEdFW
zk4au^3`@`sim&OJ?h!_D^YO<IOL<cum;cPKHlRxsTLZ7_Rbl^DSL+^y5njr;cl^)b
zsRjO>+b4XqG8i~qRNvr@qD?6q&Y*bt{M&CAt#_u^;4I0a`I*mBu%u41AFvuIsr6X~
zPPymVLV1jbe+%w_-<^R~=UAnmIAXY!?1IPo>-X>H^KQF&^G4`HH!Oz>@q~=OI`5Np
z8@RAO@eOb_kGgI=;rzHBLR7t!+~>hjFP^WJK|8q#yot`0(Q|MQBWQ}k+qZ58xAoA3
zc=K+^{i<X>&TmsMWgvP-J_NSsX6FMd<2CM#hJ>3|?p|=Kt1GVNYJEI;`mDyEp+XOJ
z*e~P7pc`M{6CMPI#S`UT=NIrM_~U*Qi21jit#t`Z6rlM3il;J%oo3*HeOxJ@a`pFp
z&jXhF^}H7`*e?`6SMLt4D!s1Hey=oW$5q*b&%#|*R)3zJRaP`jXw0WipJ%_}9Mw-@
z#(RTe|1#;%z4B9(@fT;2<FjM%^S}P8>4(dK^{*9Q*XcU_3Z?P6iFWgi*V^4MBg<i!
zSXI_p4U`x8s~F{o)5j0mqYpo9J8PT8X#M{0|CjCm{(t*_wBP@||FXUH=G$Sg!pMXe
zL+ruo9WjMal$&nqDFKFih2DMdAA_I$o*D=5dhhDqRlfS{U$3(F=XEceL8*w6x{?9&
zP;{j>ddiWh=ciDE@v6S7Hn`<!g*6PJVYx#=v$wz39zJ>0zWeSw?VS(bZR^{c?Ig@N
zMU-=_m_G{lfjR1N_9?f9v*F0*Cj?wstj(sg@x&C2r7f&;cx18irV#5J#vaCnlFLSo
zaDbJtd}rA_Ql?=fFv^HGp&G>-Mx^>$Opj79F?ybFVQvWR2v_=M^0Jg~m?KOr;ls;p
zx*S%%pNkNuE0+3LVG1y5<{X1V(Is>V2EzC!XL4OBk&>6tr3+@#r+g>G!oWGaSKIaJ
zEDWtfOO3A$>I7yWbB4(z?8MaZDyzQcI?Wgk1c=^YIY01F!yI9RT;|I8Yvrk!dN5L9
zaf$(c6=v6co6_3UrXHI)^@=U-j57VBFg)BpYR<CzG!?{Mr|Wc`esR;^s^aUH5kvGo
z{q^<ivnBbCE`IhN+DkEjuApNMV4oNtYIkn0w9kM3_4f8#uLp-S$Y7C;euZYH05kyY
zsSR@QzV{hRP|l#QGNA#ram_RIkuU_|FD`@72hGYRT1Z?Qcoyks1G+*9Wn+#$67HzG
zlvANygmMJW4*RDFpg1E$=jly&AVeDaM*%``qLY+f+Kry!v&up_G_VXgLhC%_5k4Fm
zFA>)5?ej!%K3C{(=p&wnV4<wei^?5WXr$Uu4~?}^d`St^g^JtH=Vj`j>Pm%A6kI$O
zb}~*U=|5VizR>iGy|(e<ar$x<esHLJalL%mJ(7gX;jO_lq!e-bg!X1kDZUO4c9Oq`
zDe9g?o&Vy^A&g>mzP6iql(X<o89QDSX(Oe?#N>SX&a<Xl<@fx?g%ahc1itz|I_Zp^
z<u*6J+Ljib+w-c<@Y>(sDj{nLMN%(?MR^#6Pb$R}fl%5unRy}Hq3~Sa-~qH*{0Bt`
zeuMId(2G!|&%s{ut(Wi+9Q6=D8L#je1bEK=FsCK3Nx3ZF@j>q4k9eVgLHrHa0x!Z;
z@xm(QZJw3TzIZzd65sJs-FS#V&>UGc-sjmhIW-$t5vJ0)Ug6oYehQZJ4$py+iOe^}
z(pdM}BIKW2yxC@RO^i+!@1s4GXS~ri)}LjZFDr!E!~|~*VW%m~E^2)a51qGz-4_`v
zUSpeWb923vF6O(|3SKVlmig7o@V8lmylL>ggm1zI>L;NYBe_xSbCy#neh75IM^XvG
z!N_R76KwajIu<yr^P5il`c!ZCFwCf2nqzAKp9@~JC1sa_NV)O_o`y0O;d^J(h7P07
zhs~<MpJ#l8c*DC+%HvG6nYq*<f<V?w#=@S#+RxNie<-{OX?YxF8d43R_i@(BL7lhq
z>Vh(BwL+8X6TY@91t=wsN<maV)5odF@pk*xop$Hey>{c~-SX->zjzh=a!_afmIvUK
zu&+maNSlIl`nwcQ;Lmw~lpV(<KsOFcE0k!rt`r4$`>Qo*QeX9urv%d{sRJ(6x7xP^
z$Kv&k#l-YV#$_oN%l+V%tnKb;2VWBuXPh&Tud>bxR`T09ea`ZmoLNl&Z?&1Z)ixb?
zV-P6+(&9?Wb^e(*r0f;&>#x$DjA0oHrqBJHGmT0;JdVQ0Q()Pr2+zoKYiDOYYiXx&
zk+tJ2K!SKlR1LVPj5724=y0cPZal4dym{kZ)`SrMqwQdCziscpUlhI0nPNaNo-w85
zA8j-9*&|P`q+h(Rf{}v%otI}V2<670(C>qz?F_|53G^upCiQV7aFllOo_qTELEvVm
z+6S(A8y_8PX0QDs^x?EkX1{bwv9{ZB&Ez!YyLAqC+i&itt^GVFN***QvsP=LG00h+
zuSJ0uT6RJy122Gmd$6&#yS3Kd%YOFpqbKc~U;lRd*Z=BYw~dYGZFwd2Or?zT<A8}U
zkXhHM7wl3r*S<9u_yH61i>raF8=0epw2dMs>v)6*o`ql!mGMY_cQZFrDOZ?U?>gt}
z;_|o&oi~&@<%Q=g$X6E!p-Gz=N1-Yo)q50WJkgv*I6Ul}$R2iBeV^h9Wn=`vOU9uL
zX7hc1aiQI~bvL+ay4r2Mpbb2RyP{e<L%Yti_dVSXPTpz{9=@OPI!jw8G7cwYNCW>m
zmuqq1R>~gFy}rk$crC>~C2;UZUaWTTz_UhqL>^{toPTKVOg=D?H9j)!9H$;a4-Y}X
zSzGQT^u~NqX3fs66b|6gE;*kO-5<+3#y8Y?jlrvUABxwx`Ha;{_wubZo4KR-*m&`@
zz5Cuf?Sl{B58MPNq~7Vg181nqIZ_8lq3QWFKeg`;uG`(-DJ8$(OUsLaq2Z#F_TfB`
zgsbE!EgYY|pg+8io;`cf*4KD;g1zut;m6QR9%7EbAO|Rz3=TkpD36{$dtOR1N+4rM
zK~-4FTxTrtJ~vjEO0nn6yREeC*|W7$s*xR(qANTI9<6xGyljl7LWfsZR?C1zIphbF
z3nJ@KI#Dox$}3t!!X?kv+Qqx!=}OU?wLsx=<HpUvZTgUB>+8Z1Zj=I_VvSseB8}2a
zUFZ@#(P?^qCy_l&PEl6Q6fW0xio(PFgTTvCe#2k9qNXZL?ptrZSqkZH{FLWuxBP4A
z_xjq})!0&iS_=%Qgs~0Wf~(+yuIT9sOnmEN(e}yUpOo=ptqdw=XQ$KW=~5KW&(CIF
zMhZ9Gx}~p+sWbdf;Ko^CXkfaKGJBuQuW=}Yo%~iGYHbeCeszLve3ml&xN-AFz5npR
z)ADYD^VNsebj1=`62)#0naKzZ9B4P5`m_8pMX|Hr;dP;KJp(!5%N)R;%Je=sR&UKs
z5*YDZ@>$>O-_?Gh+A_F$Sv?Itd(zj>rv9_%F&4dC@4EsJ)h=!5_vHugN1?yZpREU`
zdiYZEHK|H6=zoRu=id4$iusGP$dT)l@%Z_s!FYc<H$VOL!wL1cR(xHj>+~y>ZoYP>
zz4_%YmABV~jWxs&oPu~@26mlg_3%-9`2GiNJ4BM@)j#@&|ET@f|J6TlU;XW`w)?NK
zz@9I`4uaA#X2?d6`|xOp%_qbnNGJaf&^R1kcdzpItK;^{Ke*hl5C8SlZ>Ky&DPN1p
z2{TjPUxRmw(B>-wW5ZM_wdY^`@?SNkmr)8FeB0HZu!V=&JUm}}-adHvQTzDu$L*OA
zG0t5`J!1qKth@3u%g?e>33G6nO(X$}dn`4iPSmm)$#YCJCRe@s%96u|+9r<xrM#su
zNFmy+v4mL&1uw?74zdm-hv1jRb?)o?Kq*wrK+4dTld=S2y_)Bl3md!=h~z?Wh5x`n
zVGh(!X=GD|5hcK=b{H3pzv9S^i#X5+VKRzQ4x@%Cw!uG4TWv(EFR54ja&1QR2ZVhP
ze#%C@4vbIzqvkQ=$+C-Z%2Z)cFs0gzaivhKdCGucDhZJ=k`#bmD8XWwAq-f5CyZCV
zn@>te{lKgZ7@<<!rGJEuoe{>UQkD@UCT})7M~BWNDFC@n*XcU_;-|k|k1q)B?`Znr
zP21mO8q|cA3{rj1KX}?WqE8e@Q<Ed@t+(#B-~FB6Y+wHJ=iBtuSR=e-RStg><PlPo
zz#}xO1j_Y3A=psxy>kehUrBk>l?tdmoKN}Xxv+CV|CI=d<RMHFst}DuS4v=o9)}Kf
zPZa{@uK3c<qMxaUu?w1HV_p`ip`!##lo!Sry+wy8QP3Gek5XC;6TbBUAYdT)LElP{
zO~{=46{h#)W$^|)$Lc(%i7o(cFHhR$^N%ZBU_Wb&)iuF%c$4Gp_Tu^DwzIR9c4iz>
z2?bn+qn#fL@Ao|S4~yU6Ni#ee{^5AP?e1=*t%s>s-|#DWmNs^!-DT*rFg)5xurXPD
zMV&#@(#OzGw7)AAY%3+a$f*+Y8N!?~9wMA?$8CRmtvQove0(HxJfHWvw-3cv7fd=|
z$9b|ms|XguS7!W$(IWVy+?bl1YiCEh?cu`@s_o!{u&@lcGWM5&0q|AxlD2{qyb>W3
zUZcEy(jVXVc6a-^8CzR<W2ojIkKo?(XU|H3H8VGp`6U!1Tskk<!-olL$<kd}=2%-x
zz>zL@OX<$X;B~flHp4&d29AcxLRx4d*3r)CQC^x@2@h1l=V=~LokuT0U3ed!FW|sD
zDBqkIW8Q8&H%2ly6g|S85l9b@QZ!K}@;I1i<AK@Y`w4+7l-y}sTw3TJDvP&L=S;@X
z+M%$cl%TNN4NM$m-DQnsF7?xz3d@yo#)C07x@h}*{Y)qZU)F*Ua=Zeq#|mqmHsd+_
z`99{^v`|_JJ?5TE>H>717vYm1Jb1snn4Dds@8&u$w6nnN<HsMQ4F`3O(lAelVT#_A
zWAZ6a%|(?f#3Wo1h)u%@o9F7hHsL3&S=UUiz;9r}Jf8*T6+>`1m$@=76kmNEjF#|o
zbc}M60;q=rbuP}$TX)*x>b>%8agSl)&~SL~a|(9hF}r6_g-9F898WKlSJCLiT;?z3
zrG3t~Vf9W3*@fBo=x#Vi7|2Tq?ol4^6bejhiucLV^368CbT8%4t0ZkZs`iyjQ&P=e
z>Nb`Hs4l`|Fy<T;o?a6a=2??{?Cb|}1rHbyiulREP${G`mta8%&#nLw#&7_0<~VpF
z@Gw0&UP_SJ`Q^Z5$_@VM<>A5YRg4ojP8p*-!e-LGnXHYO;JEH_1<rU`7VHP-z+;s3
z@FQ<WN;|j|3?3it)jA{0rf{Pu+S}c1&i7khz8%~XT$Mj-+q!dp-1KbbAoWX>tGs!}
z#wlKmL*{CTg423mZ{PdgH`_P9@s0N2;ltpK;IresHZwa_ILNtP6hY;&n0F3#)+>DG
z>Df`>Xf|U#RC9T9yxTU`o-}8Z?eFck_4OBJAlMZki^&Ie;m_dY>`i&q5y&5vu>t(a
zJ5ku%llJk04@-gL%rtv0N)~w9UMn3a7}WOtoxS$`ciwCN>R<iq_U?P%&A6hEud=t>
zsWwey49Djd+OSZJSr?bWfMyM%OLG~ci5X}6SYrSfZqrv|mcFK)m#6z}e}6l4=0(OU
zls;gC7Cn8|z3pbE=R;T1kIX+s+tci`-+%8rZFBueK*AZG{XFXF3FouHWyU~V;Dzyu
z-{7hiP<~Rrg;oVO+8<wJGI>$huQ~@w$Up`U?N#t=-U=oS&fcyyv$gpwFz<}F!@~Jf
z@cRibLSB0}+rrA7tjpE(UARYZO!0elW!-iojkL-74oz}qopa6bEY6G)zV6YZhr#uM
z<N1{utFx1C$Wa1)VQaI_7@<t1?3sntHZ^-Aa~fQfmNLj$TYsMS7`;pe2Vb@iKYTxP
z{ZaLu90pC}!AHsA%tRiiPaZ#R4^!{PrjVo;p)+%B`N|`ikt#18q3(nv>_#3ueM3tq
zmh%T6+`4_Uyp4EyQRdCePPcon-AkQQHKxw7GT!(sYtDrpP*yod)0vz?4c@<hFYA)Q
zP!BgqPPDlpl;nqX2B&r?2i>M*;njq{<K;%#ROg7@robOA#m{z~O?4jJfR_sF73~Qu
zWX{M6@Lc9uo%%-ca(H-L^>~JM@QOm8ioRv+eP3N!DepAiw|<wBG3yA=cJn4B+Ds{-
zUc6Y(=W!{@@Kr^BGylcgrMxm2p5ZAAuQGg|?8@-q7Cd1ovkpT;G8O<$dnz(Oa6A~~
zea;gO-?+W)+)oPrz4FxB%5OX=ybtH{-ez=#zl6K-l0q#?#?zR?H3npJYbU(Si&CzD
zWy)a6IAINW$>K36i-cD6o^X%G&Ufv`!%%SHZOkLZr?y!uV9dQf$Id_8u5gTC2d`vY
z81Xm@?ZXcqww-+9BhWW6iHD`|;_WQNp}si)^FR#0YD0PEmz;;Z2wpdjc&?KDq&;Mi
z17%fDT>)oKwDptQNdtwf&uVx<iZ^mq-d;}se)Vp$pc9{#FZBm*4c@C<m8E~Lk_Y9a
z51)SbT%XAMgv{J(&z`Q8C!?c2YOJry{Y>jVy>kC5!)Gp!`u9OwKYLY%ANgLtKFL#Q
z@coA`DW8@1=bwJKPe1MTqf)TNwc_hKU8i4(6dW+UINRR-dtYvEed$XTx=<K8h`0-g
zS~d&WL_L4ro__qWZLe>(xiFx=_4j_O{qukR-?iWSo!@IWZ`=w~noW8blc1F%l!Xfm
zuXh0x3%fym{aZ&s*wsPuu4iRc<C}Miz;skUNZv!A>Z&Z=ssqvV9YOg7xrH&VHsyD}
zC5QX{zaJ?3in#bsNvD^mZRcR8J$Uk<eemFe_UPH;wz<30PQpmod)PbnSoW<#)nGm-
zy$~skld>_Blwaz@xDk*Ms+?aD@PSnlD_sg8?(p7aA%hUtnGqR(OiXQTCyaAH_`a-~
z(r2UY!r1E#W=Cv%eYG!@A2Wz?8=L{=U4jQns3NM#T~_v)FG`}CFI~$$!Zv-7U)voX
zi&1mncyG5&ShliZs%-qdi&)p$Twy|NJoB~MY&<Z9`lDQRYS6@0yNY2<zlGS)EB#}Q
zHIeZs<~7gEAO)ul@1S4WERXh=@X_R?Xu`<qU+x=Ej1!odAncj4S?(5D7z$o-gdcgW
zE6S3O=UauA$~a!9>vWxd;nUxm;tSrmPCsobiqpUTKnku#67tF506l!`&3o<l|BK&l
zpMU%HHZ#LJ<s`I7Xk2tQIDDjsl&LUQp})3<#e)RVcMgNHDY!lK1`Sd{f2Vt<pkJO7
z4iGHZbo2UUTtWcqP<JgAe#blV@MaNWh{CEu%2Dow=Ah^96A}qy$Ksk&qJ*+9JMAOT
zp$O6zf=X>=&8jbqIntywxra5}+wEsOpsn5Nbs@Bg=NZou-q&;0-dC50rSy6bKE)U~
z59#tUWgPQ5BY;xQaS1;eo0Jf4Y`Bf4V?rz+9_^L@6t6)L$<qMME4rET<Pp}1;JedX
zN|&>;PNsZw-q;i&b9k3w!oUfNu&FjZBitgv9|IC08ZR(H8LNq!n^)%YY`>izZj`|B
z<>h&eU7c$)Mo`YPVW9XpN<VoS5T^IDi#`#qeg>8f57twbux=D5JyfMRa-NGbxn?rf
z1ZctuZfv{=ztF=L3R7f^N?;ZkAz&g@1BbV7-)<|bt5v>(>j}G9j^Da{tAta8R{gw-
zuF&Kq#vn*Op7wYU$#_;<!(0%;@W3YsmF5?CG~7tJLMI0PGS21Akkmum^uADh)fqeE
z6M<c4w4N6aGZ0{QBPzUT_u3&u8wxyC$jPC>SQ3Qc5izp)e@y#cmJ()fcO&yVoH<==
zBa=cQE@eLFOMp)hJ0a9s;CVFp@%#((EA7Uu*W2>S-OTq~#sCb5FAF?SiX9zxLD|e)
z7n+xs{xC~_Ro7zATxCsCzzn4gmnVVS{j70@Q@K@Rb{rVmY{&Z>?c~7uQ^E(nXpF)p
zvj!*x#wHh%Z#MnEXfK{WDm>!;$T;s3aAaXAP~SU4N5OZnx7N0|)*5e0<q^(y;c1_z
z<SN0nGmdt4Hp4?675o?<${?Nsl#G<m74Ca>DdV%4cddJhLc-dWHnngkedpB#Ph{Tr
zHq*zWyvI0DUcz(AXoB-A1ra{>3buuB<E=s<NWf};1zssl%ZsrTNDBoE&$Czgqd)q?
z_NRaHpGz=#`_5}^(V0ynb16xg`hL_-k0?ucgIxr6D3OI16UvNreQWzW8*TkX)@68a
z3Vw>Vk<swyLPQ2P^N88qT+8#L)DvF1l$hpyKKsawc`4OgS4xTcZtOD$ukubN3C#5K
z&4hn7M{x96vWd8GTrgH4_b3vL3#C|aj*Fs>!p+!GAP7@9p8lJM$(fb(Df5svKY#Wp
z>zU9s>ufIY5<JRF7H*ZyCD<SYW|YX&aem=O_QR`z*{rkE!}7w~%lzB7z{g;l;_xhE
z$opk`y9<TU3kphWo)Xmge`t}=u!OWi?>XP@MfTM1zW4pSw^?(SRa5JPz_|8GSx4JD
zd%-um1;@Od=H_@ijkTThWpD3A)^=z|#t$w4C%jmuoWsO3DC0m8L3zd#c5Zet{a(&I
zP1M<SaPn;E-0JGhy5JFlZD$L0#pi6NSHXYISenk5^6q0~z|*Fb1*vN+I08<78CX8p
z+XyTW!Y5mKdztg>7mxKLcsb*QeiF1-^bz!Siehjv?Gn0<BH(3v^5jwSZ>F5lw!E|y
z+$J<+=LrluqZfGM_;BY+nc0mXOW{;Xr?cuG9D)Y$QnP<N2(B&qopm)aF%w*2-v^dE
zF6ObvYmlPNcrPwca0$Vh^{E`_BQ*#iD%_+9CKOvkJkUPDt#}v)9_YLmg6l3Aw4Mjf
z7CbH5zPxg?P0rp-e(w_a2Zzt1Ydiu2)9HIxlIqvX_TuTI_D6sCZ`;58U;kzM<3Ik-
z>9<gwNA1H8AGCk}@Bc&l`ZvBEy0;scIcOh!^f0jXL3`)>@3jv;__#fM`1liVoUV-F
zZFyFnsLtII+Mg1EqKX1u9~hwU`nhpqrLC+imICPBy*s%_A<lz$q}{xEv)v3`B(I=Q
z;wj2_6t6TBIw8y+rODD_U@th!e8R7kIL4il1|6Whd*(bN;r7A<m8Wv>2l>L&tP^1s
zdGwH>2)8Fxq2oBv+^zJ1oCLl`w^Cxp6WybH?Q$P<*;+@B4|r7_W{u{zHR|XDN?!Es
zcKV~gywMmo@e;=;>~8aLC--37+`XsZC1rLG!pj;{eKf8E&m-~>5efQH99^!UUBa&l
zeJac!;~01YT=K9|FD0QhN7;!MQOxTL8HT!y7cV+<C%maV;J=JsvbL0k7d02go3gF=
zgKCe^+^>@7q7+|GpFXYM@*sIUuB?A_#QL{xe8;Q$-Hn60!2#Okd^t)pG9qJvCxIK4
zi5J4(XeapF-sS;|?-dT)`FlcluGMpMsEopYX*=aDJgC1#MFJNLti20Qnm=U^&OyUR
zmy9cI=`tO2Z`AE0ltorW<_6|m9sbD62C}qnL=Jc5wIn-;{-uBGzdpU^d$A|EdVO7%
z^l4_ifrq~L^jQxw?JWA1aRlp!&cWN<E>Fo_N?D5|4f0p@<?4|1gI7KJ^LIb_lcC1q
zhpznpW3Ni}CI9`1cR#&8$@q6F4d7#N!Lrwiuj_Q3eg)F#)OdUC%b#z*_M2a6x9`7J
zj1~fd8E{?{m;Buj$90H!W2?=F$^Ls^`OWqZ|KK0B-}~L)Yp>mVtxZgIrn?Y`c>ckG
zkbCVj;l&V`N+pQu2vQy!(|YgfZvBsDT;&U!+cV{N{-IxWAuNhSSZ(qVF$4?YQg9JH
zS><*4b3;-FCM3W6mH)J}GUqdWr?4S7*xcP}A3pl1efRzEwkK;(D>Pp*Fu{;QHxeum
zrXkiC0)#p(3DM8?Y@85cJk?<^I#C|Zcov~kZ>s9nRNZ1)9}sNJiSS0Av)Vb%W|y*u
z#U(+UP2qVM7Q|Xv6gU_V0xd$ii(HsB1H_`)#!GDb^6<(Z=0P|{Oo#F@1U9stQAoeS
zWSpeVVtB$_c&>g7h%v_0U_^|+_D+pYRcK2GOfN1h7Lz;4q6x#64Z3_Msizp2z!D}8
zqh`1%<R;TsAs$^_SedcQd-|(h8B0l(UAX!6SHi&n1O*|Xlla%3WBhE;386Xzk}(Vu
zI)J~Lr!c^nL^1Zu`YG)y>)T{jJ_dOZVxU4%q|B4dA1fT-c%81(b^3))KR3nK@BH>}
z|DEBJI|lp-=`YOilT*D8sV5ap{R@*kuKrGczhBHw(Cwn9p#k?_yV<_@1xhc<qij;o
z(Ua3;&IY?IyIp8L`oVC)Mm-yu-0utp0^dE?(PA53xF4NGqsyB)X>cZpjXed4PfrL-
z+$UTtAuYi{l_B&O>s=lQR|nOjHK}_XT?$Qf{td$qv_%=h1FCy|u3K6QZ+CdOUu|)Q
zlcZ1T_MS6WSg14DVDUaPJ0Jd`3)qygu&~lb!mI6WJ#CL4zh5+Nb@f)8D-Q~N<N1A*
zF?d<RCc+d#tK*~XHZehY$KxmTcbp=N5a%G{_NopXFORV@TImYAsj0bA%Am_a%;7U$
zy&Nv#;9S~4p&_K5{KF&RS0*S;C?m$gH#qaG3$aIdb~Fd!>&IE!-&t#0>kr%B_Ve_;
zmDh@pR1{`}hFy3{(0Q2oS2k}aWw`H>`W#kRx1Ei&&3SNB#fR!&4^>C-yOi>$Y91bi
zw_96#oPMV7nKyh1;RwD4kANqtxhlS9zJw}x7W@(+mvBHgR#|<|6@P=*z&GL{GVHY+
zdPru*O~4}XnX7oglmkx6`#t;=VI*Pw!oqUupDYiBm&pqTx{!%s5CKW}a!RrAUE1I|
zA=mlKv)to}LNFaZFLOYEgny$T)`qc+C8Z=-FrKfnhAQlLF5k}si${msZF7@AJ-po{
z<q8GEKxu?e%Bss8Ut})30w8N(W~q%#+zK2n1?~ga8K<e~`8Mi(<qO%CHc!qjr<mEa
zQP{Vcl$ds9tek5m6rXdTcr^`W%)sJV>Uo(p+0wSN6CuwYw5^S2dFQf}KfL02T+OEb
z>Dh%VxS}kaNZXD|h&X+vq^pLP@&%j>lzv@6&ZF`C{4jGC-aU8aEb&r)zsk6t9PPE|
zPad|XPal_Zm?sJEfLib2L$%+$!zJK$Y%FWO6geIDl#(dp*p;{)UN0~A+g9M6=g-XC
zN*f(tP5XHfbsoRBLTdmLW4g?7_6qPJZ$JAFXG2lM3Cl-t$GeHb!@lTg+WUX}@BiQS
zXJ7lX_WJ9ux3@q4#k83+SsT;=F5p*I`A7AAJaE_N+PWNSXUDs3XX8aXKG-V-AH^12
zb)NB>7^jR%-d93_?gq~L`A?&8?Jy5bd&1#%c_yS)x6rS?<l@cetf#%r=Yebch^+Gw
z3T&5g$~rn)Z=35pKeltv{*BThpH)uAGi4gPK2F(XS#yDzSHW!z1B6!Wp)*}C3qKwP
zw(W!9Fdlupwiq%@2RC%Z%6V|@TKX|w;U$GK>}6yRbFkiCJb&1Bwt`zudK3*fImD=*
z*ILUut8=_=+^REgZmixYg{OVb`kL@<>$RUNp?Ug2NbGDIGzngwnF($;Kj|w4mo~V^
z>+JshH-j6TjkZ|(qqViHh0Vu-|GB`p@PgT^fVHe~{h12wf%hgSDQK74%F@kxe@v*w
z;IG4st2H(iyf!(1yL$u9b&sC=_g`;c{`D`nx88ay^ag#KPQNHODBAnph~Y*leFC$s
zfpv=P!?wHqBKt>16`4}FX0QeV(`bf-juxz>uQOSnb8|c_gJZLfC(yUhpM$-f^mVqq
zcJE&3)>7d;q5gRpwV{q{DZe+jUgUGX%>+KjvS0gjU5jD@MlQ2nf^#MpgAdZ)Jg0P6
z%y<|#UU+C6k3p9+|3-&<7~{F%tErjL(UIk>O<q)|ng8dR>+>4xI!iF~o{G}NzGmRn
zsvrKamJ7e8ytJqA5jqrH<=id(9vVX*DZM8%cOtY1H*Cya)))w3x4*m9UIh32+1LN9
z{m~!&QTxW%znS^p$o#!3rO$(pAGRk?UbMa4(7wadw!OXE);G2SZ`+MA$N4$qQS@rD
z6s(j6&dU;3kP!zS170C~FXdkPJ|isP>Qa0C{@wPu&)pB4h32MD#&A4*L%NbY({=ur
z^Q{PIwB;gr=w9aR&YiodEBuT#TchE^o4qW)haz$#u)n^(nKot~)7G{1^@<n+Uhq*z
zhsV|ag9CKGD?h9qd>5q$Wr=g5co_MtbNxcgtwqYIv(rv{c@~vYDBqQBT_6X_BsXA8
z<(#p_g}IE;d_L#fjletEiExcPykER<1{S68UddUE7Xu<nvBJqE0|`FK_*mE4Lm2}X
z83&4TOl`?L!W&Y6QV!yW;1TOeJHb4D!slqlNa7vl9^41d8|N1<HZq5lWnD?B4ceuT
z#eW<0Bx8;5qez1*=I0pAbmM8bk5YU4>Rhzd8!O?BCQC`lLyaO6tmA>?2RC4(_~OtQ
z^MOaI`tjYl$LMP*d=bM8eKQaJoKW(l9$83IcIRnCOTbH0##u+&&UpEv<{g~kztLL8
z^$f4O0`AIiJpCnC!_yfvGMh3=NW01qyxP^<CaK?84et3)4mW_%o~`GVD4F!seSDj-
z_0;gzmEX1Z@OgE4AisRxA7u7XpI7<tgT2fy>$7g3JzsB6pD}6>X#(G<_IhdXaxR8r
zpVsH9XMgVci;_I%?}y*}p{wfr@z)Qx<!|cx>r6kYe4N>};_EtHr(gCIT(WR`wf*{U
z|7Lsr?YBx{;rLA(9Eu1E1KwUcVUYH>vuOx}_?2J(O8dwE@E=!r!`pZ66cH~b5U+NH
zK&s#?JVk&N`qRJH=is@^@1F8z<*Qp)kNe%L`&X2NRa}u=1S}t2;gvh7C_nofY0SXF
zMtjgwtalYze7JksEU&ke(aX8WMwjK+i){El{P@H6{g2;mTVXu2zYFnYRo(>_$=&l~
zvO<itw};w8bcIR6s9~Zp_=H#?>><*KHOq4KR+xkkbX_Zpla!_Xr2s;Y@*N}7&ndw`
z9F~xyyZp1L?N-wSYM2XSWZo1go;L-eeikEm#jGLx7!vtx^jYMvPI5MhO~pwzCy21Y
zDoZ0urL?m%IL6<&=l`@>`Gi0SwD66}Hx9)>r(cuPQ|aqPm=|X;q~Bpy&K$~}m$iY1
z7Dllu%lKFLq<o)c%^U`*1V|ZYbz+)$)S0eoOTJ(#38M&&Snz{G!V*?Q`sZ11kn&&k
zIl07m6CCxJ+L%)?R3VDOeCdOH`X$daC0>}vgD}@82WK?~*XcT4r(f9g^Yi%n?H>{9
z@E0(laD}k{hSQ*i&I9~K>4)>c@=8GAJ6%S@UcTaiMaeVXW@aX8(~UmD(+m{KtGMWa
zGSJS@yS&HhHCbvyP0+^5wP<PHDT9i%5zZ&LxA`vm92&&C7kxuhyjQKtZ^jgdc}B5-
z?x36K1j7qLeEC_x7Y#`{+DTxj3^X2H=RI8B$f>IYe)9yblO@2x6A*k<FVZ$Nv^+Rw
zW-}x&LpS@m36rTuKec1^)}1y!eIq|ZXHR!S<4>DVOn2^mu1(F~X%kcZJfYDMmcJt<
z3>}-CDPbMO5T$?+V-;#{hzI#uc$LXgrcKT)WPgoE2p^TPB{-WSIGvm;LG97ue##wd
zE33EL)ZDG`Qgdm?Xv%QLSm<>6=p2?VM9=t+5}M^(+VQCu7oOs*ZM=Bc)?PerhX*_1
z6S@ZmZ!uw<SYp3=DHI$}8_Ef1ck=ikkRHGg;SXVB;Nc{^M%u^AW^#g;*F?s7solJF
zFJp73%`dK`-&2|A!?v}(8UAdzEw8TT9?!YdnR(M+b6nwu!ejA#7zljK2Zc+8iA+X<
z7J@Ay9X;<xP!zNH7I55Q(wu{X@-j?5^D*G5JS*W+euEc@pvan4fAL>g4`m2-d{|?s
z4B<AxowdNrPs0hDwFA#dAkzgr865^oXW@OOCu?3Y)k366-NPt+Twv4Ox(J%`An?j-
z&b(z%0t=Uc;jOi2ZD(seb8%Vib9PfHG0QU{{Uz_Uli;~^M4@%u#;15NQ0$ubWJsCL
z;yO9rZ+p8tZBHogw25MQYHqQVd&OsC@-vb`#%*so59qS3-niQqZ@k`S7Vg%3T0<ZK
z%ymyfecem@2<kEiX{S(Tr8F_Qm5e=s6+Ze(fkq&hc`i@AgTtzSY-BRH<e~(`1gzj{
z=E@4+n&qwY{`>E>7tf#7e6p4=!EWHUM&W$7JYimif9D-^HAMNBJ;1Ar!X0Ddgx}^s
z+cIAl;GCz5-{$ZlV|8AN0v=tIi2E6@_uu<o+dtUtdx5}uhhOJ12`%MZD1zwo_S$QA
z+w#g%!B;>dp9G9>VEE>-nHqCzlj5=0+xJdqStEFGULCh@-z}cJ!kHiK2e)NkFu9bx
znXmB3Sq{ZdgSQ%UKtQQ@vY#=g{LAmmqk;*jDS<XOp0}yA-CE-{*FC!Ay9|EWZ5!+B
zwYN#5f>*lXT&B448WZ=G4K9VgJUZGAj-bR#QQDB_HP%@>0n^~?^ICu9sW%(EH?b5v
zH=Ff4nW8Q--7oWfBKVJ|#Y|r(;PuscJ3d%%2YcIr)2+6#DTLukg}67@;DMrGdPW#h
z%DcYq3Bn0x-+k}h_Rc%sDSF}jB}xy^N?1NSQ=#MT-hC~2=w{w0@D7a4&L{6`dFs9K
zxi7S(6`oqwH1ELSQifQkLZgl)zp<f=ySU`R0v8EsMybniQWO!PcNqt!Kk$EW>F8MC
zdy29cPMB%SH*d8sfBB2;%{M>S?!WPR+H|iJc<7j9?svC0^13+@mOpTvvPM&%d85p7
z_8#L#YXQyZa~k}WdgnsFSJM9*ckX747c-CcreJ~)zI%cSd4F+s)Sf(kSY<ey%RY>E
z3eOHS)VNhh%FI1aUCVlDnP(o)vkpJzUFbo$Y2ZO={g02mtZ^8d<_XG6)OnW~*NiQ%
z!AT(=(R{cuFihcAW0tkDzqi&tc>mkMNtY>OI_=U1aGy03{K7EfBrr0PGI)gXjC*x)
z+Fm?)(AL&oq&PT#sPM=|)-!sJ4)id<FWch>AGU9N^INrl-QC$~@4oZ>_Rja;ZI2&6
zY9D_1VSD!MdEmzRSg&VIzg}l}^$@QUZ8<RIOu*%p)lzg&y57z>zkdIX_U7l_Y_GrZ
zMxE1yrcufmOLPHk;X(E0n{T$)?gy`CUeGh)?s$6%@wXD(L5@MmLlMGDiIO2eP;`P^
z#N6X!jQbb9@P$&;P-Y2RxVy_Mr}Iyg5ft|!kL}@u$5|Wiw?~h8qitqQF~miyoUh6Y
zo5F<0Iv%9cP<wS=j1*Dpo13+NclNhXcjlMz2pNp9hsTG<SxY>ZXUZ6(<UJbbN}1Ls
zlq59s^72BPUziPSU(^u@&Z?iA<pITGH#jOVz?+R6#k^@F4>Sc(Ofh~^uCd+P-p(9`
zPNzN2`?_=YZYiC*0twybsXvqFqeW|Y&Ep@0GlauCZ!sL07Zt~Or{Tj=y8b)Q?BRn)
z?dg-}<sDR!Y4Qx;QT+)(QtVN(l1EWuQRJCVb6>o0o)vx1o9cIFEpIpBC+D5Jmi3mw
zD%y|lF_&nJbI;Khd<Z!fz8IdvSG%-jAw2EijMGjFDEi1~CIaEX1Ze}s7zO*?JNHT%
zi%%oVDkV>@-oC8knx@o|aV({E@IKta<I6lr+PaYTfqAso7!BV0G=27aR$B2CHB^-!
z9x~ppq8lEhkiqBRdtLdfw>|SaZ5!N&Ir{s#m;3k`Z=<Gg*&y%Wb>)zqf%WzEt@h;U
z^Y&tGBk+BYxh%3)sot&n>Q}Gts+^ykwDkx3E`9o@Z$C{rKTG;-{Xbh>gMR<G>t{`{
z@U`OWI$ft<@-#6w-QM`p7ux54{mX4BoBukDIs^kVgm7Ml$Zv<a*vqDGC>!UyH*dGE
z{^nQPSAXlbiuppMA)JnwA$_PV-xZ$!pggFz|IYowf%)ZMe|UwolDGatNPC`ooc}^Q
zK9dIbuhb!7f;v+)s4G{M*MIp49qL4=?z{3m|NO4P^N(dMMHc1F#!fb}AAit3ejH|1
zIEkYWA&hj!22n%cS*u`NOL$TJMwkgB`fsffdW0gY+_rOJVoGQef-a=ND5m;~P?oh!
zD4lmPJ=%o9!d%$6VrVeigelrxLJ-fwm|#SG*W)THc`+s>;LF&!=b?7@jnHW}<1BAj
zg0qUsy%L(#zuNd?5+_*{XXE0x^1VkWC1x`#Bd+$<x7v_Ws%4B)g1kqpd%~Cq*Ta$;
zqax%Ifsp(9VC*o0hD+b20pn9=m*fHK$QsAIbHvIr?_*RjLD}*aoMNyf3~^qqO$ZC2
zezpiJx(){;#w|=|eigHXNz<0<d+t>^$>%u+x!@|Fn6`<uM<2zIK0obky-wHZI{iYX
zUuwnI4{qrGX48+W37sfh?epsTZ2Ii|&tB4}8<gBBI{n#%KIcQ(uPAF+(9g7+H(25d
zu{R$2Mkyu4pp9}5Wh=B9VM3h|q}b37G?f)?2~Lw(xe}OZHJXQBp^b!$wNXX?@*m+O
zPct;aM%q33^tVE@968J9v;@lxMbP}&*_mqdDSB1q_BP-z2!sgF^le}j`|1_LoUUBq
zArjIRn#v;fM7Gc@@1$L6oAW%E78V;1lRh?GfG=#^?4pCg=W{dVOl9*wztE=3x;^(O
zqDlzJSclL^(Y+sxKSqZs5()lh!w;OdU4qomPT{>~7n}n$myu!^a+r6=a&NVbj4!7U
zX9zLaAvBw4qZ8@-5W%Rrx^QA>h}C;|n-K!XWUPc@<4L;LP*;BC%d6A2v+=YPU&c=x
z@ERBc*(yZV#^biTvsP>yAupbUm2a;T3=x9lGh>4v8k^v;m%3BH<kW0iy>YLt-nt)t
z_f8v~TyA4i(}4lrcPAwr<RM28g&!fTDR1}iOC_kNespD~?|7Pt@zFfLXvfZFNqY&H
z7UmaAXoTOWkk{d@GyuOs(Z-{La)q)H?*aDB1&<r!fgdVCQw<J_=+lC`$qY2!!&yYT
z;Z3);`WYF#)JkX**s8N&^PaU}94OeV5pd_i#}EXbh2L==5AOwvsv-PP!6xOevuXo7
zX<8XDrATMAm1TCud^~XKoGhI3*7}oDObB(Qx%f$*6vHF<)4=%UNjp2-uCQkZ``dYg
zu(|VGHHjuBb9Qlln0G0fc`Hq{8`)pX1*UmHI6sN!RD~{@P*3LeWj_mOap_KAI<Uw?
zXvjHNq8uEg{LZ_T0yJftN9PkctA|o&Bx633ap_(_;4kf|tCvGCO~hEdUe>v=YjaC?
z+vKb;mp9wuLgp@c2s#Nrx3@RiM<2e|K6wB8ZGU$sV@}!K&vOxQ>*8duJVYq>v_W{e
zk)hzVzzbnFFQd^~BS|&weO!bDEk#VhiFI#|t*yM5{Pr8X8hI=VLH5QQpDQoP08|O=
zl|or-kM_kcem*euvaPQ_3%sX(u#*g9#)`*Qd0CCL<2`4JJZ$@!v$13tpPJ9O&!mmx
z72<9rZ54VmjSHNfmN($Uv~Z4neP3k!DP{5ULb_Gx!_-9qI5Gy_tj7VY;mxzh47|Vx
z3>7G;>O8PIA88`+Hj_RDCe!xwtW`X{y-g|iCgv(%4=qPXOTm${gG&w$w%h)}PTtAB
zt&~|EFJ!oD|0K*`c|(EcE4(%I>ZI*tF2O3LN$ndl&l${G2f|?<?r)`!JZg5^#^&>s
zK~VfgwTE&L{_gv$Za4{s(2%{|z4j-6^2hC+?|-M12ycJ>OYO!Dh5}=?UhS873zssC
z7uqOAMpux-uj9e%ZG@8HqU~kQo!!<$f@Y2~Mrg>%@j<Wuf)IYLO_oqS`}d2}CJf+l
z=G~gGUKs~cULU3(y!^`8BA<O8DZqq6>+8Q1=)s9RdmlZ1*dAnGe2{g$n|9Q>kFT=M
zf>R7tjXzqGb#a;elvRu*I{t^FE&NXZhMbu<mbSw&!7swFmOwu64t}hc@4x%Kf~Q(b
z=?kwV`!DpC;&Of=;}y7Ml*rqIhiRQ%SI~c6bJxl%*t)PkJ;?X)3gs1;{=N)uueJBG
z9c-^>e)p;_5(YGcr^)utR@Tf>#xlJ~H`do4weNrbn-$^Y_T4u_M{d?ShQlt-_uJ9l
zTAk}wN`?t_<5y1F|M)-r5AA>XU;kz3g0Q69?fu}3!^6GePs&S)r&oUe=Rf|F_HX{p
zA7))YZjYZlZI2&4>2iSN0~5}hD(|6u=Pk6H-@L7a*JDI8bETkwqbR8<!HvA}AMjll
zm))z~KHfvH?kuNUH&;?_S1b{@*EvmDKi1|&_WJhxXj2)22LDkWxNl9Oh2YWo;M$6Y
z3R(DPfA+QZ-uoYAZBevY!^f%jJUGS~W$UFFqaea>nKuSV5;}<frkKKY=y3n!A<f8x
zM{xMNe6K5TU!f~0jWRaYAVc7>Ox()yV!N5X-M)FF!XuJbFtpHa9;D88Ern2cr19_#
zWD68clJhsq5GO-NRzR^lP!t+7>jS^n<x1L~^2kGy7u_o5&b_;Z2VOjXQS-XCzM1v2
zo3f3?X%phm8D{!r%}QwQ(IJoez-{1zrz05(V*`BU_BLaYgQEWhv%zu3rou>Oek!bC
z@(OQCsm{1U$lWZxQg~Ll!SoydV;<nGh54EC_#($T44>=Vt@7Z@TsR9(-NFp=faB3N
z@HBG`Jd=cD#OoL{@9G~|q=;0v|Fn%E28E#Uvd)A+<lU8yWIi)5nYW7Ilm3-FE$_kq
zgR{lJxc>Pxw&jU`H5abVW2gA4()R)1eRcCH!7Jvv<Ux7A>dQTC_St!Vp2l5H=dVgo
zPR;q%m0S5JoQZcYc=ai3F3L`LTHdY<?C=au^vn5l>l^Tmkd0ZFFY{*V&#Ye6b5%jr
z?>Aif)28(M>YLO{8KfWl_1B*UHT(RVuIpz{R^PSa>pEShU*ZIJ-hBOD`_kY0O1t&i
zy|T!JZ-cRm__Mq^4-(!9VLS;@TAo{MzZRzbE5GrT_Qw7DVYY1Y!;B=~fGP1;o>qmV
z^GUiR4e*^jg##h)LhDz!@hX4rcU0M5_=KpqsVw+D1iO12_43`Wd{>_o9H0EILV6qh
z?*H|B^<w|gE<!91voYP+UTYsd_^5sNy?5G+FtB0!vT^G|k8?_%Tm&<i77W37HjV@=
z2ry<8fs_#MBEY4L!JG+c*h3cDXkacdk=2hpQ?5;#@j>+Ukrk886Az|h2GUNOZFOK;
zdZ<DQICXU^84M(*ma@!v7J<DYl9X^ZS%fl(Yd?=n+2_S@RXD`-2V++jb?J|H^;2D#
zoxw&;n+;`3PoFY5=@-VqSYY5NzJw5Dd1wZ;v(DPW-x?$EV)#viIm`o_o~keH#AIBA
zDbN?$%ylu;n8Pr47!*n9$I5FAF<<#3xWtIDA|jwOcBOF4CIItMa03ogm;Cx5UzL--
zvL-F1Z_25#N>|Jgrjp0nZkW#Om9Eoux=z2C>DuG#uPdQXS6>IIsOEr<{w0&Q(Bhty
z-Af<TQBSJQ;N=kg&%Jr4{nl@NxxM{X_>GyVqLXktywB^w*(-#wtU$|>HuumxgeK^`
z=w%s!phY&gXxliU>Uh<IHV|Bm4D+bz4e<V;J@S`wAs258ZAS}v6tV8-{h@r`LpHEJ
z(balC<rB`7Cowvecl!pqZ-6NmWGn$^+C=zpbbJ-wBn@#N9o9Yq9BoC{39cBE2;bz~
zm*EjYd^A3FjJ^v0qjEGBjU<eeUX|wsfzL(Sr`*~3l{Pv>a6Of#GR!$D%lFw?=KxVE
z5JFQ7_Ko|{C;_=EWeb^?x(KUlK1+GxLd%bvP-xf!{Vk&-N`SNd%oBldSI(5@SMqD_
ztMlWwzV<BBcG+&+yw#SLS28APSL)i_Sa0vX`<=A!xXlamH8GR1BwWbb>Ms69Sp>M5
zFT88`4#E~RUkJR}`PDW&);$|a5S#Z8cGt>FhX)j448Fkp5w`B^Y}Y*F;qU}t2Y-jx
zAuMth${8L9|CBu7qn{z>Q2XN&<`FusEH9VkIHeasWe<5YTJtDOcJYOoj}qz!9*SSf
z<vy5B2P=R5&bY~oC&7;r(5sL2yiW<ggqJFhqs*Cc!Ryp|%5Pp4g#ONPO5yeV)8x1M
z2?Qyzu%V2u0&D3TMMw$8GuKlS6j|M(oDkBO06doSL&H7K@K9jtgy2w{@QXb>dGVX+
z(@@Ihb*P_0Mq%=ogn0~HS6ILjnBy-~Mqrl!U;a8%Z*;nNwo;~~-4_?mrR$*v3*Lq@
zmL>SU%(@ESdUBAudCyI?(ZGJe3OFuC&T92N{eLOU^=XyOtB)tt$oO0yQFKiP_Hc_`
zVGi^vJY&}F>gvt*#_MmCfblSWIXXN@os=beHFpvZ$>X%66zbFI;}C`ML@6PT4)@YN
zXQyaJ>c9)9zOIZgM;`cYj#6Iwc2<2lZ>xbZ9vG)7XMb<E1iCzQC_l_M!LPl=?rv`@
zxU2{r>2`S*R5-1(_Wbe3?ZXe>sd45tVa-qsmB$QNfJZN!lXF;}K6q)N@N4O3ye~30
zcy{**u7z-S22T%#XUsc(=wnS0AJ`6@1vm<337`wtrhW2u++&X=ibdc!uTpaGU>zP)
zCcHNo6LZdcCi^9LZqFHT`?Y@UM=3uUDcHNgzl<Hq;GuBjc_|SeKYCbepAv<#Jg7VG
z@EWE_%w8b-9(&8tvEd2>%d>WRW<KL_TzJ)*XPjp(oY<S4o)u_WvukV5+s^i8DJbsT
zzMJu#4<0{n`v=0b9i`vHW#Ev;<-P3D>@R~OtYMzLUAc19Uc7i#b4d9Bf6vd~$}6u*
zA%?y<!)<PnSMEY79Ea04c-=VIUzM^Vb401WzaLs)E#`g?+iMKd&$ENJx4+ed6SJQb
zR*n+*yWjbC`}pGz16vvMm-dlI%{kSj_&5trvTgzgrNB&oCU^`F4HQvue)gokrd1tp
zqr2&Ma3W7)Asn66Hk3B-X8h=*4{FZ{|Jj3CMigJxQ+dl|Y<S=pd$`fwl`#M>%H~W<
z9yjF~JemIPwrwF8kDamB1@aeX3~0c>%l5tRek*<ZsNK4`TFQ^J%;SRxAGL=MA7p-b
z5@p^4gEf|yr{$ejXOd08qr+Lhn{9jZMds!>xVS3|osR_%@_zoG{-^)B{ZIe%|GPbV
z`m|u-yWjm@c?DG%!psLxx<`+m)VV&!W^rk$=oeVv33Kb#oeCdF`LMMm6zz8Ube4DF
z7JLOmjF;6fbcE3-9P9nPtow~kVPc)Hw3T&voVM|5TrB0y;)3%e$BU-z>}<8CPhX_(
z4~r)8@&(rq9zIIjpBJ4J!VZj=(d=25C#VyQxks_${rBJhuswgVk@_x*j^l++(q?nW
zJF7gMQU`eK?+2^`hb0Ti8)e*bbrnr79MEAWC?j>LkL)CC{FV9XUS}8RV+xx}pM>?B
znF?Lb`#j<(3d?vn^-zL=J#q~=`SzWgZ8^BI3`SB%=flWrC{}06GgE$w9iCanVK(DY
z=LOCM&eI0tfv+*H6g%j>#!_7MbEx{6oBBY3gXhy$<8}}{K-pz%yMZuXk6c5(^IAGW
zD+`ghy|tUU+Ryjdx-2I$3-HW4tN0E|(Z%#<a=gZfEXmkXFqPaWb@D3LX7pKjO|YwO
zVfI`n$Bes!?`DO)9tqr0P<0+voy99>F7&~3?{`JsPTlK#$Vkz`Zfw?NFX)ptt*@<D
z7|H#V2`)s{5p@9Hj>am3Jjgsyr0dh&yUv&s*4UY*nTzxnd<@1-U$lvll5sX}^?7wE
zSE_l)ck~edq@DT%cXhZ!qg<gy`3@#zP{tsAw%od(dh@UUmqxgUCs97wltsB@W!h+5
zL_*rzcebFtL*~Vr4lLON)C-ELkN(|f)%7!`UV{IB_OgC*su4;0dDQi@CEUQZ;_EtH
zr(fbUwKUh>{PLIDTVMKOn_ZZ1uY!<m4lx~&Jg+J-<&VRlg!OE%-MQEP{@?ri?d><;
zD)(fJh6F)}gmj$NG4!YSO*vQhJF?E_;65VZ-c{bKkDjlH#^8oJ>XrI}Nk<Ts6xx+?
z{MVl&DrtO`F(3~0y?^VOv^bx#UZ%I_m)S68lf<I!-4EYwA3u4}4vs^BJq;6o9)_+M
zg)nFs2jviwXanMlU?QG`WUO=5A^-6iCRqIFnZ`<49|%b>J2o~5pZf?gOQ{~e+$KGn
zgfz1YT0;24pj5rdk3qs<BBDjC(+Taep(~~_&v;0cGOjie-Cdqbekq;|L;ho>SD^{R
z1opAWVB~%MDvK}7Xr9$~^%=ulywq6N5FVDLT^J9e+XXSLpEnjpp*EHn$S^|==hO#%
zVku{wjA?C3^Q`70j0Q#@xfiYpQ&eFQ!w3+jbVXrj7)tS%wt>?U>Rh$ieL#f?77l7W
z8-^NdmS4p<kM{AwD5u&E>1lKNgJJ9Q3$CiEDW}hWx?N-KEHay54D4wb!g~2SU8n2x
zi<bUQD8AsWzl*8k#<U$`{3|9s51Q`Z^8?QAcNH6olACk0Q|&i@^Na2CpTA$<Z9>tq
z@{9^SB#^U#MsIAKSro(Zgg70SQ%JDr?X=8hmQa@W3i^d^D~HuGI__QcvUe~v>*72u
zB8=@DVV3afLW>DI3AT88d5^+@pokEtZ@%+p`cc->DM#PXr=qi06vqGPOVOXuE5nH{
zccD@?|H&a77TRlM^hbVeI=={A3N1#13C6l7#c{hl+sV7fd56%3XBZyMcJ8tr9qhEN
z&9(ITxV&aYrkr0xa7W0l@AC7ap3dh$F+wnykCfK=fYeiUr4<ws;geW}k2u3f_SEB?
zuC2|DqO}!{tO*0ff>_wH{q!OCLVs&KG8fuFK=LZ>A~d>t?{n?e9ZH1d+@CF)%=(;g
ze|h;<o8;M&MwOvU>LB#!-eVf-(mq0s@D1Zr<^48P%E3PFJ<rQ_cFfY4=h9A%6>EH9
z>vp!c+w*77n$UY)@PeOmUYofXu6ZUb%S$CNBTxoIrIbouZS~ywAUAH@EWr%^1zbzU
zktY>}sf!>9Pa&13N#;Nq2ZzpW>SrT4`%UPxNlL%Wdxgr%GvOI4OkH><u%T_`@s{6|
zgT;5JJYxm6XJ+QRz(4cd$4TPx?&{;>j5KF*IYaO4{3v&MIL(zZv5)ge8H0_bTyR6E
zS^QtB!?O;L%q36w!jZ{R*q%}ZQRrqKDHAe8fq7xgoF6w*g5mwWgBnkLu|^Bz@-BEf
zJKk=aYfmb~8qcoom0@*c?tyoPb!0d&7<@21s_yQ^;asgTVatSGJ39igSu5r%V^U?i
z^!rLE9~i(t4v$W?@u`J2k+D@~ZB+Wy<wFM;jvcj~T?f!VZ*v*b&wu{Q?atlLmC%%+
zT<D@YQ*Lf4d!^Mjzp$G2q=w8pyk*{6g#qpBet&l@@9hLmx}dBp#lWmNQu$Q_v@vVx
z<*9RfC;=%(#tL@J8b0temGR*b!w3MbGq4QMmX>a|)#W<{OB5cIwlK@sXx7oiQColh
zIIy@|1_sWh>TNXN8M)vTo=y~WF6W$4f;ktBXPvW@gdd%nUdq@lr*5HN?N8v%!C1pX
zQpT{ZD<iYuKZ-_sPx}DMiyHIH$;*r2vgFl!$}9MRQfj!A@B{cYcUR*xL^<`U;HuZH
zhZnW3t!L}amIK_vKSSa1h5S1?It<<$4Src|OG38>_Xv+?4WTFF!8vv2<Iq^Gg`2nU
zSGY}++RyR{?oGiZR8Ldv2fHyOk6pr9ikhtJgRJS~D;yA*G>0#<|6~MOktz<4N}0N~
z)x9v_7x+I-s~%EV!uLhLCMZuAYV5{_LNC%EVG@V42P^oY7<^@HE(3=Pfyd?G<Ka?(
z9tIDC8Fkt(QsSST?1g^pU*Yeyb_~ayQTlby143EK;N_*Ic4Kv=6t*v(K1sf#Oy{e>
z?{3OJ$p!ytAEj4ci<jr{6-5fAxVba`K?K2N=5lx>?Hvxz2$&olY_})R9+i+?2ucbH
z=N^Jb^cS8j92uOsvGJmP@cz3Q*S*lRsdkz+pQdjVbEV`Ats2c3@*3XX->xyIRHKA8
z7RJST|LB8v+K2CdKlM!3d~9xQwDpaw$>1)EPB36?dAA!hPzq<B$Ae1(bGtk1?K|K3
zW_$4QM@5gckJlgE=lrd2eDm9-teQ<9R+n$I>A=wWIYSGfYg6Bb@R?g>&~u|ala@=#
z;fzWK6H80Wf!EPG*Xg5_x4XlGw1;zc<~Ajm^MQ792^WY5IX=vsp7V6e+=Ry9o$xA*
zFmBzt6_}h4%nOtID*f7TAAI;>d-(8Cl|d<m@1az|CpoI%@bEZzY_aGF4<#ODWI*zD
z?;LY;*4EO-2M-=+Zo-S4k+oR+_UUlH{r8oDN8T$LL+*Lj$zyO$eeVXq%5)`r>dKnV
zJ!7lQJOI(1k_FUDS<iV7&GwT~(sb~BVSw;#CBFzhEZUvAD5vmsn_KAoWqa-Zo$^HD
zDccpQl;<ZJRb#Jxlt2_;ykOB)m-#MPOxkHYDZ@(UlSNxfn6!r{a9615gOJk`#XFLt
zEiNron}yOw)16xmf2j}cXK;&`xS2UySy`^U4<0_LHu4}92|ztwfFHGQdUm$jLfJ(L
zg*Jdko?dX2wv^l?YroD-4F2K8CwwKv*g@9T`Ps`-#xvZRotrNBp|HgNRv&`bC}Q>9
zxVYd~%3Nasc9dx@%op6wD-Jw(zm%rIoBGFVkq0eY2cMbi;@2~_#*(5-$VQ4h^-<Ce
z43gvrLlt^4ctsiDOC4Q-<@ey~Jb1sKJh{L{zgGa*pXHvpbwA&#&BedDU)g+KUHxbA
zFN5@HnH`4x?r<6YBIWvidcy1WB-d$pi5?0RzSS2rq8D1F*ZYId{`IrJfAZD&mmmG?
zujBgZle~YK{6GJML%CLbU8n2xOPXv(?tcEw_J!a0^>*{_-S)}`0Zs~o0r5iOuR??`
z!t9(Mh2aU~|K@9Nw9mi&g?2NWc@{;5^a}aqy+ZmOsX^|!+YzJt`7EMR`TXZ;JqU5h
z|2F4WWxbNG-tSrKQ(N*W*<Ta=P+lO~2x>7sxp!4WCw#dOE;qUbs2bJN%c+8gUzMP0
zEe!7a4}~`TussVSeIT^q5Jo~qj9A~OxG*u=sy~P_>qO&)$;J>2zS~e}13`dI0Rpeh
z7$8b1n|p1+v|-8!iK^fAj$KYyp0R{gZz<I>=4JhsR_Qw?gF;Im2ns#xL{_uAvgzWY
zE1cY`_F_2FHq4)muXhODFi9AQvP#K@0O9RI!7wMOr-U^6m3wG1ffH*p7Jp^o7^aPt
zV_ABZ1rec8XN-B%>7Ra;aw`l{Z(_Id)MuMcZDb|N0`e0jSK7mKD>w2+w-l0ilYtM6
zg*h^g%IQ$${z$=zvYd-HV9<;Q3ufoE6$5h4;y<5xALCRC#sIChD6PEN(iUS>8(<sJ
zR9SsK>K2{X={jAfU#N8L@%2|oh02PS4iZfIgXyz(207~<g!!3Nim@M;K9g52G`)Y%
zuX-)`Ia%iBrrMW({mpjc2G4T!^!Eq{;9qBhprd#Z_#V9^khZBT8k%xCP409VP3ek-
zPG5`XNPXxUA*0_VWKnr&Ko?GSi`+T`>}nH@UJx`8z-bF%1Ny?c7k#iv?g}=W{BAML
ziqr2Z|7t_54#pkjDPDO)qemz#hRQHTSQ*|2B}B~g64s)Lx$r#d)^FYr4&*;S-EGgF
zKB`bk1pe|AFH{~Kgb2p#vQ6Y!DNuM~lt+n_HVqLR;6w5~b@eb}p^Imf77RU<msf<@
zYT*r!c~&?>=(PBqR~H$>mjv^vCv+aoH1@&;3AN@TsB(`$vJ7)dAw>~0(Z)wIPbp(^
z+9B+#ZFJH>@ts$inwe@Vt2fG{fl^|0d@k)bR-x75WtE{1+S>);>4$A>DfS3o2SRe~
z^NdjapdIXOHD@Gw4~!g#_oGxIbj1S+X*4r4m$p()cI6VDi*jOVd8w3f=78eL`6k_C
z?6`P5=ihdLWq7%B7ToGKjCc)aWSo_NvRmmpfPbRk)4U362)@#R68H)MxV&<s3;hE-
z;2aO;9HKf~B=cN?*|cMGbFE;w3+p<pfnj`%&}p^iC>^d~u*Rztlz86oKxZe#GYJ=0
zVbv&@Cn%T$qoqhOmf@~4*Z5Fpi(O_tWEcuIN6Q<h!yM()@GFK~$(Qj>_9h(C)HDI=
zOqxU(7(O#?$9sC8AXkVdhEs&(6jv9g`vpg%)?fY@i3!c=EFkaKnxYJ;`Br!O_A317
zS@_(HGyb#qxr}|`hrmvUoA9RQ=JYUM!?TA+6k@S^lBFCd;=(*oKpI2+VLgA=b_4s*
zpFb-vf_wLbyz4?@;p2SHFD$jiT;-vdalfpvk%!e*iYRB`cEuM@oa{L+oK?i?UK>6G
zv#+j{b2V2k@Y@BJ6O%$n2`l!zZLVkD0|(Ba0~fqq@U%ix!CfoMH%s|Nc+3lrvL5X7
z!W14+L<mM^grDFL{EP^Bm$@z>U|@<jj`6hS&oeIDJ8SL1!;jnP@mcWoVw=d?b>^4z
z_ssoi;D~~3hzB76srEYu>?Czwc|T-*<H^Bp38<4N5Kw(&IB=Tvwz(<9=SJW|IJ4=L
zWe=9JGS;ssj;w2YHsg{puy=s(^EG9h1t;#*-oe>ji%XPx^YyIylKH1x6V7Xf*HQW~
zJetqo7$NB?Sj`LAAIYX;EaN&px73Em=QBQOYt}186eW#MV`mHrz}4A3Fcwq(Os$!L
zf?fSQGv?=bf@UqG74Sm76bukx6M)y=DPuPhI5NJIp)=Jt_>Ex$9Qcwl7EQG8%f61M
z(+DLKTKcjSkz>;<`LvE-W{nP4Slp%5FO22o=~;X6{7L3uE3k1~`?1yF+vOW~g8#Y~
z%<{@|yLIPw8%^Ca)3c$CCxN;3^mnS=zI8i!(8uh{GL|F3>!X4B+Jl?Nz&4}7(J}Zw
zYqSg~tiJ)Acf8H3jVBr<Cmd)U3D-6^?;IduR)>OPc}-@Gq-c2K{rBE2eE9nP*9wn<
zX`Xxbq1{uJ!c28wDtJDyYpkYgZYWX5Q_gYL^5gcw2k*4+eCJ#3o8SIc+uq$Z6$MLp
zgzeq!QZ(^~)D}312P#G7yYIi({^$??u>HsX@E_XqXD`~$-fny6op;;Uzwxd1-g_Sw
zd^>-Ow;yAO#l_{K^}JVWtz}JuoteN5Z$5GYV~78+F4or8+eaUN+&=o~K@UF{L^_Z!
zc)uHmFsALt1lTj<J<uF0&Tl-5IT!kF5SrP3-L);;T<YlOn|6k5ENxg=m<t?I<jmC>
zf4rE;3eXW_LqWK+L#bllepqJ|J$ts6@wF!5LNI~nv0kl5JXXhX_S2nD%Kbbu&b8;y
z8h61LT~U&X;N9er@HKcpeYQQ%&-AhWt8(S*GAyv4m-A0v3LQo)BKOo@pIj#N_aN;C
zvkMEe%~@t>&*O(r+S4b`%Aksj$$OM~olbWnXZMv;&&yz`%Snny2oHc481SDIkbSJr
zi~jKvL+>dct#e~%t}9ewe&aQ?x#qL*$xJDJ{C@WIMcY{4tmhP?6kn8D>PFKkr0|(w
zmT|_`)>c<`rw(nPNE*md(C-c-DJyk2%MniI?Jj(z@T)g&tY&`TXWq*~5^E2G8+cfh
zA!|OAEaSo>5RQdA;63wget5&X&>`bUxysuQoG&cQmoj!`WwmIjGVt%#13Wqxygcxf
zw{Fn_{INN3&zKrJFy?o8{$0UZ--n=Edh63g8HF$E->stp9)OaM!daEiN6%B&Rhfgb
zJ+F6NsRvvt#Z|w<XAERq17iotyZW)ax8Jt6_S^dUcC9~aMti|i$uvW22KDsIe^;Nr
z{AJSM74Kht?db=9{Z&$L<Db9&eEQkz`is(!YR6yZ!l7I%zOK`C`ejVPJ+n*m?F+y8
zmG=7EZ@0-1JaL5Gcfe~Q^C0S>5Z2)k%8MXmVenqN`&wPBGYB>Yg+XP((X}rt=PNhr
zKcBXs9$x*f?_t#I6Os7zd!-^C`9px}X8%68f0d<(HDv|GM8BWh2$sER>gvR-myiek
z&u1^o=IZ!FSi{ZsG#l5CpM2aN<l5TbZYN=i34e-F%ccjB>$UcTaF>-xE*o{vwT1W8
zK+z<0i#k|uA+UrVhhf&WX(pRCR@&9pG{!yOwb=&NC&tpoMo1$T;L5-tR{qpyun9TJ
zs;h)=$y0p{!3-0E>8p^0VagHE3VV=t7t<F8fqOd!#^xVGJWdEued@xpiYcDwh&FF6
zd6g?+0Hh-5n5N{#bevIor9G5kytG*NVq(SI#+*Al&LOkK)aJYC`DBK!m|E5@7)I8I
zgYW_l?j`&yg<2R(p;N;1R(pCm+T0lo?0S8s51oNXf77Omz;^dcYZX4S#xrd*vBpl`
znzt~DrbRmm)G#3OnCBiwAq;x}2toJ0J~3RE;>{n1**xjOtm|~0uG23}`juCF^$lMV
z`qIDsRO#xiN<Z$sAHH9|A#`a_?~h7X&wf(5{TbW~7oWGAHx{xfo-UkQLXpD7!VL<c
zCX`keG*Ysl+5Maop~}!18`Qo@Evx8KFm$gP7Nwd~^fZaJEqaOus-v#q9v1g3w5NxH
zVqMyeHwYHcSyu390naQ}se~Ew5zZ1o@DdQpgfc_lYIClPv`_u&^uUI_HoPwxcZ6o9
z4Bk8XSRrJ_of$IOoCPyAH5<A(Q36|hK(#I1Ah@n~^i$#C6By={2T9BLjD$9xJA>z<
zl>tZk_-QD$wDD{=<3!+bp29*W$1<L$d+o`ik2Br}dGEX(hDX}pUC-m=;)6<<ow~}B
zIJ93_oypn7HaWeTF<?>Nm6GO;r&jkuAIW|F87LUkPe^6%&Pou@!(w<e&qt<90YZ>3
zWZ}rfQU-B8?{%fg&<L1v;ZMTvoTnXU`>BKXnsW7h0C8vS)j0)2uJgSL4<@vZIpUo|
zS*0j*#VCjX=l<)TE9Dj1-}y!cLl<pjd8GuH+QneV*&jRvl*h{X{_AhFyLazr9(xEW
z3bhKqm!{|=FOlxK&|ytj%Py>PCLIA4&!)*PY`%53-MDq9thnU?lP9Mv_=SP)R_uBR
z9uM~S+S-e!8DD3A@TBP0=q$~<GP4_DQSRku)mvkv0Lrf+mh6-<Su@4+Q4;073=G8+
zWk-su@ueW!-Q`KMoi<&xezwe5o)c<Ufhp3uVkO><p^TMv(v@Qb-dWQ^ED_Ly;qYpB
zTH`-8GoLn7qKv0(>#<vL<Le0YKK$^#w!Z!%bq<HWn=V-IbFSPJ_9}eo2@kjy9y0Tr
zJm*=jLQ0vJv8>PXOv^f_K;y+Un)kamk#f?9mnU^!XhA!H3hil3GbVU*%E3AVD{z2k
z2ZyCx%LV@otnqg?*V?n^kFpQgueE@`q_Fv9tX{ETKdLe6Fkw#scCBkLdz`V*B<oc_
zGj6F4U+;AaXgCqSO-UdeI2ilz!*|=)zV;t7=Lc<Rd9`o{CAM&9>l-i1OJ#F&y_6}F
zGtN}&b5H58zW%i4*gK`<$$A{mn&mNvXQw>j-S*_kliHswE@a)0Q)Z{3fw!~F`|}qM
z``ib2z<FJvXI#(o&t8k->ayTnC^Y+_!G43M4mhDi@tbmkB8<>bs6}lZ9h**j)t7<G
z`l@-)<$r6%SX<92GVAqrTU@%Cey*nMrGlkbp)*3)odsuH<g?yod~kI(A%)HU-ex;L
z2@IE)9YaL>EpWhlC4ETS?4#hezBg=vSE25@z@LJ;hd<pf{Gj{xWH)c#X*aXBXJ>^Z
zyHz0yg?FXMrWoOs>wSGIMQMfHgfjyJf$j5@seT@a%HDM*Rra*Tvcj%T3V{Z`y5Yyj
zD4ZD_9Q*_~SYPAQE5XA{foB3~p;cR*6}6nPT3C54c-WbU$(Qx#Og3j~Ig^e8jnamf
z^<{vTR|uFDmexM1JS7A7uP%kQ&A0`Q$_sCRZ)$zjEagh26rH?GgBuwZ<VWoVc@{aV
z$yN*<Udwua7<vQ;nD?&GILN#%wm08=EATL0xEKss7Zge3le}azFO7no$M|X6+Sn>O
zPQgbJNRdXd%QME=Mn@F6p*LuQvy`4beO_aE=k~q8`AYTu+4Hsb{P}v@*w}85A3tsD
zYdlh4Wi8ESjAt?@$HAFrZ6@Crgt^VyoX*(K%w+7Rv#%{<j^%de&h6m%?isAj_UCn_
z|JLyKj*y6f&)^&HQ9c>(8>_3KF?Rzy3x$g)RC#*}*Ef@V<vEmj)kjK+8#lZM?kR?{
z-_72g5<;DI#!_HtH1kaHu#__7UtZ=V*h=QZ1C6q(NBbFR@4f$id+^{<+uGbIIN04k
z2u|4yPV9!OBs<0(?X0$>9bf?cDp^Nx7e0i716?5ZP=4WYjVgSa%IgY5z^j?>Qg~hB
zmGY!Ns{8|mLCJ;k+4(p!hfbG{ORfPw_c5mMDqNF-7!Ww$3Ju%d>Po=<z!knxSry(o
z@1zMTt51f$hex4RflD$E=Y1C3WzACdlZ)W*4l*9j%re$^Gd!t2oSpTMqhQ=T8!w(+
zl(F!M{wmj4)!1dsg{yq_^jVed9%Z^SC9QwBlSd|xP&gMa=z<S4m;#2vi?<@hHNJ#b
zo3e`U&0OJM^`}BI=8uvrx04r7LYX;Ka0=J^l;7AJU!HOldA@@!{h}1T2!9MW!o|kO
z`DPSx=3&4un@{Hx>O<kC(9>>27W9_+A(Nsg<4vZ$;Mu!i%lybU;Prj)FxG(s7=89$
z(Pwg^jK|=9=e;x6{yqa!pIrVj5XR{0rCdFqFW#^EuCA&_9r^6?DDd0mT(838JJ0LI
z3tpbQ@hCm<M^|N3ZOQl9`qk;l)u%jvsq~rbzJB`i2Y>x_rcc{CNPly6{dFbO=vwi0
zovza_VHyjey!-Z>?d>oBT3foYYNutpmQO;TzNv$rY#+z-g>nPab|;Mbtvg}V=fnJF
z;{~rl(U5fCc=TU&?pEHdLxn4oyZLwtncF<(xzF6o-TVdN=UZ<3zcUx^_*aCW?$+=6
zUd&GzmfV0`Yg3O$a(O<;n54)eAU;zt@2K7V-S+gwv-aMH@3;3Kyw{#*<9BLvRLns(
z3Hg+tz`%wMK~yB7Rl>Ji+JzvOr<Qj3%S&C5hJX!-X!3<Abz!avv`T;yrbC$(wjvCl
zxYU>&Wh#q>T`>h=7;SFFmv<nna#CouiJ(b4F)iw2b){SiIGZu|FatK%-aE*qTgU_L
za%g`)%sqWcS(s`J#c>#(QjFzy39K-!7|67n0*aNKJ`_V~+|vR?-=>w;(`MDE3;iPz
zD6wwB#1I(O7#8!JT;?w2=mZNc%nfTrvqyl25hl<i2(+2mKOi))>B}qGq!>>dV)dJQ
zg1`EnWZdgnGK>8VruEOJtXqYR6k{n79I;q-mRfmr<+);NtkYie4aUrMA5%i4?qP6b
z<3|&Bovzb$`b9~<l8Ud-VxhtJAJ}|Ax_^oEnKyr8`fR>W>b<&tIC(y}is0uO<Q=rt
z!Trn2bM3~Bg>2Gi%0r2elYoTq)HzgyiD)@I&f=V6sE6AjAR|B{81tTc1Suo!xP%Y~
z`TpsKn{WgTg6jz<@f9U7LHqLDHL!~H9$J8wX-^ld;Dc<4LnE#f5OszOnuCU?ZJra<
zl~5!-Coo3a6o{rM+lHFb%M?)->63878yjmWk;Oj2s<UUNO93=9?My04l7nhLt6|D8
zw9>nMgWd%hyqhJ$+vm?7m$2Tr^f6~S9^PSpqaE(=WL(cHY!u7?QC=FMcLar#6BHUl
z86%2}ZgD?7ztSeAyEnk)<ypp}hddf}mR5z@BB1Zq%fgTGm?)ug{*2pj`k9Qm&|x9B
z2t1R|7^lBIWS+2#>CY!2vIzdK6dR!EB6F7Zyu8?N#|LZe>BA3dOv;)(?dth^|4S+H
z@~G|athM#ECv9gtV;lHeUb<19ImQn2N$G;`a)!_9%B>Pom(VRd0%)C{oeq5R8ksHl
zz(*16lz=+CmXJiBd*jVgoDp_y?`*Zb-L1f1_dasY%Cl!r3nu307Bb)3>3pt}_Wb$t
zw!XezLLR)7Fjun_3R7Ld_A31*ct7Vsaly)2C^oP$UL$#WECk<ro1I}TEd(Myk|MP$
zG|tYC!zT*!^)l@aFE`%L?fPUqhX}I346g&X@O5}i9{)XrC&BqJ<yiPm-hj@}GWTFj
z8#?d!GV_=6UO9J)#d%;WKMA#}E&jm76`b;_s!)2ztn!6Q%X**GXKfpA!-0vc`hrEm
zv%Qol92`YAq4MnfLh2&COSxHpHRsl4*68@yRN6e>#wf>9Cmz#SQhW^!Q4WA#XRGYl
z0@d31)D!Y7Lqmv~zSW*58Laip*+eOT7*64xGbi}hzz!T^9(OX&jNo`zJ$d>tYxzZu
zI{_w-DPy9pfK06^ibaY~d}t|t(wMHCPBT=M{`4@T;3c1_m+~84VJv12$~!Tz0<Y}s
zY_^A458r#|8x^8W$i<nNh17N0_Ob_h@%(A~_~ZBUY`3l6zT1{>+^YF8hL0XSXy5<d
z*Nf+0y>X|F6r7~qA$Tt<_N9<^mx0;C>Mt(^I7mayD>hvptwo{W&Q1<XiBQTLcp-hk
z!&8v?PAMx)AADK+#_UPpmEn>34EKDkx$)pl<Hw_wC*RE6eD(>#wOV@y2@I!=<|%zB
z@579VIUsbLpTChZr}G?4@puSapYCT(9%YTbXlrXv^DKBjuwD2m@A3q4E)=i0&CMs7
z!)JvbE1c!QZfFosCQ94Dbl~q8e#|`fb9tPVv)eY-H`=r8H=WP56<mOJ@FZOdewm$J
z3a;bTaI1|86M1E9ImR<8`zeZcVI+I>jNB_jfxvL!<$$s?YtMO;1gG{#&!0XHJcS-i
z%?4*s>QW9-?Dc(5_u}pQMdkAfGiK-oT+fiec|}uGi-Gar@vF1hcrjAG+Upy`skD1`
zVKsdg`t>mDRXrDhn@2UDlT#C^lh?~`#!L9!)6A>gqOr<cWlh1y`qESOT7g-iCwWrt
z@9h;n=1~hK;9#_c(Lmw%ls`YW6quVU#n0-^)e^c75A(!fut~{rR=ANN0$jke)45;5
zNouEN@ZPaj?7QKXwHIsc!K0_`#oAWp_%w84vwizp-)-Od&O3R2nf5FOCIW+d*0gZA
zJbZbOooCEv+P!<PwVSu^)EL6MyjWh#JqmT}<YjQ%Xy_=TTJ7Rd#nX^yi}^t>cwJHQ
zfoXeXiYwp8@_qG2%6si@TdFXK{k$oP4d)|)XC6zGXfw0n2SUS2u@u@co$<PvF`8fK
zo_hLhPcNOO{X5&d$j;NIu~Np}x(znFM-N4F8T?S9We&^>8uj(Bf1^SmUS3jGcg3X3
zdZ1)65AbF8-s-s61@<UcC<(zt=T+)o<@2qxs=1Rd^7~Crf&TQX=j-s}^GXf@e+Jg|
zkJ1<~#KTIMUS*&+S*9gV=xeJRC>Te5_TBhP<&BkhdDnH%&@02{^ULI6^s(0t_A~$a
zPHBuT;{kgAL=ot8!JK-p%3**N{7c4CpBejsCziP;4<ehvM=&~87ThWfAZ66n=4Q3q
zm<fk@_uj2ia^fWzUs0MF5$!P!aB&%Hhwno3dbsG7nnU&DGw`>P^*i84Daz3vYq^JI
zzVfI_8F-$Wo=Takn(NH@rG9sX8$~U7RVn@Qy9~t8sT%i_nzrJB^Qm9PoFVaU)_oZi
zC*NW4sIjT2J;zxSDZ~8nG_!W>1<Ej>k{0!`@9^Mv*4fqb0o>%jPriS8RgSv)_nyC(
zQ?K;z`J$g-##QiG&#zvR(4ngf>=(`hb9q$nUA_vQc7EW~(2||3TSh*euT1S#e$mFd
zdQqzG-~Et!*`NOY(W#z)Hp|s_AD`a-!C!yfskc*~K5g?~bp356JB4e-*LAv1zl3RK
zd9J<n>%Z1+-@o6cLf9+3V6H+u{_=@09SK4o3L$(Irh0yMq20TCKg_}H5Mfq!awb7k
z^)G+nD6)_FBn|RZzA(&P>E(Nn{1rj+jxuai`gi~GpWl^Frn*zl5LpDaTcdVH3A0ko
zNZIvSed+fq`I5dM^lRCCe)Q<0_Tb5bwz0p__71ag2s5odWUOM}hj?RzE9_8+zq&AT
zMVN{>T@k$ySsT>OC^%zE2%&C)?$GCCN?DkO(^7U7Q<wMT7b*pF#6=xLT+C${L}5x;
z9kD(`P|Nyw(4O?E&WcEvFjGQ0c!se-;N8dQ+N|m_?=C_dZSR)hmn9r4!JUm=+M|37
z8BI^kSLOh7T7tpaK-!3P=EgI9E(R&}DqEc5!Dd1Wrqh0#MrBeUvJjkGoD0)69_Gb?
z_hAl`im^(5p$KXY^I2^P18B@0nr`al!)y^cVbHW~H2H;Lpv-bMnt2~66wM2vBgW8t
zJKUEQ5vEev7(#rCjR;1t%1b%M`sIZ~+>L#=$ht`T>-;zybYqbA_DxdSlnsD3fS3Ip
zO3VW6>vWy2)8Bsjl~R2DM9dS^1uvsrKhK2zsQZUg-O`k+=b@nWyJ$}?G|D+QU;5%3
z?JK|Wg?97iavr0%gsI)*tFF)!^bOr4+#_TvrAu%*nhf`&7e!lguRIy@S7&jB4xwpf
zg$n<NUe$Z48$L%b2y5K)o`#}BXe|YZc6LD!p%TmL@he)*(>e4_e^PHKo8G1jg2V1f
z)cFbmBnpONZi){;Q=N0>9=hha_U!I#HQwUxIVWbIWN{{$b7=@)X2Y8;EG`v2)0WB5
z+(}^qOV~lUKxlc`_I9@`eB#IuA@f|y=g~zV`KqkyDeQ#ObMGj8QH7dWSZ#Alyp5LI
zMEDAVSspGzf{lzXr1`Vq`LruMMY>btm;QB+3H=KlJ=+bhvYoFzG-D|aDh$_H+IHSv
zT^!^^-V~vu!!Ju9f`)hRkUsv>t5-bs!e_iP=R8Js+UCZiwz2lGIm3n5%cF-6`*;Nw
zO6ixj82jFqv9vXO!O?zOUwc*pZHh$aFo7R@B!LT_j3?ph>aFx^q0dWRCwyfAO<`1q
zGkBQ1V?HVr(b8hg4?!luw)#t$v%k}}_qN;SmhfgnX<z1If49nGNl&SBdUD=2*0<W@
zM^D<b=j&~2d#40y+JlF|LlUMl?!oYsQg#e4N732C01@mSWSsZ*HrrI@$oVzSI0MH*
zOM>CkGgj_|WD9{O=NQdhjZFrr4|5VG9xR?@?g-^hGk16%;oRmjHazZlVohiKvL^E$
zzR*=XSN03m)Y$k;#(S!GOP21|NR~yd6Uqo>QZRyf=kjcCcMm-88XN1CH%2MY+IgKl
z^zgy^Srgm2H&&qpo&7_|%rNQr@VNL==XmiFBhar<idoCWBnB2ov#}wJf)_r8G4-#v
z{oUtjV|dK8acIc82wxmLF+4)~8aSsgsWbEJJ;Lv&T;4eo<IX?1sJVuZct1XV{J8Dx
z?3Cb+Ck4FImG}CU5?&^M;1us}4G`GE8*l{0igL!%r&1;k^VFgAp!AaT#hCL#IOrZR
z8RJrR1pX-adEW4f0=JCSc)ctv+-Qr7Hwt#j`yiKR^NY*v=Iz%4vp3Q{ivF|q;G+-Q
zlP8bcTW@{7t=_m@eI3BZ>B)B1PuAVKbEsa{eg`gOR7MFRJf?jkp|kb!@)8`MbeuE+
z-ZF;i3y(HqTH!41wS49xkEZ~UgrO@XADAC=u45p95+(gwSXgPJrI0fxU=|$v)YjmN
z^SzAwLEEV_ZVs}y>SsjRBM`=&pX_C<xAU6^=4N~P<Z0%U!f3YE7H=~3FvK9#-rU$|
z8yVN_oxQdd7<}>KMJWxfz1oXpueZ6i)^_){f<O0zo41<t0^w+aV2QwXmPcONVO{if
zlRAPIcpYwUtp{fIa~-tj&)|Z`*=L;=9H^rkG1!M)WDOk$hW1JcXnxGw-tK1BXvYaW
zAz!^bOdB^c{?0kJCq8LM2U%xj<UlB%4!+7BF5|Ga_Po3R;fqop+ta38is6?p(S*&w
zXXb$S*#x{#@j^*B+@|bXLm%Ma!cC(S8T*q8PiY*`6#G9k2>u-k%=5r{m2xRjUk3ge
za-0X>ti5<v_=3mmTW|kbKHq3_vnfA&Lc;EVN?To7ZMSdVYd2PS-BSK&-%-lo4SN#!
zd|qKlt#KjZoa1H8oU!)o>GSsPd+!!qo?n<xofoCV`m;a%di!_(zyHv_{`K#)ci;V}
zee}_z_VI(qS<6D+9TkjHxG*r?-Vr)3<DPj$)18yn*N5@M$3PuILQa=5!hX05^hHIn
z=kEJl^NP;H^V(htSNra~%X`u}N!Ax70^BNE%uL3a3_>{Gt-#mz)^@EsiZP)K7earE
z_64^vLO?fW(+0F>|KK3w6r7$uQ;ga-Q>0k4lyP&j(?z?`#ogWg;up$bHFIn&yLjxp
z_x^{0myO`+%yst6_Qq*a`rzs^lH!Bl+QPS82_Wfj8Kq{8gMU1ec9zV*{lFc3<{8el
z3{+F5;q4cCj6QXQboKE`!t3IvltXDq$)g>3GUGQe1g<Ofp*I!&aR!W_QKO}Zq#!bG
z_WRY2w26YsJb+K{YGWxhk_Vq^UMX_qQ)j2s!EN~tPK-n6E3J|6ouBz#M_^>VSx@MT
zvMD!j+@P=@FM}Cy!fTKxIB&oC@DIf!pt+$h_>$r?FZH9>j|Xz!T$lVO`KsL+|G}6W
zGrV_=DSVVRXj3<2f&==TYO6j{L=E_pL7$xyC!D16Op3BZ8R8jd&2)npaFY3~nWQWY
z9U>pgGi@OkQ}#jDX`M@#-`a|2E+uBx6HiHPbM^ABjGrqR&#q`yzd7je*q?t2H~p%w
zR|)Luyu8|2a;~(?yw#Py)Rpg^E7#Q%`B_R-=T+Ip(arMQ3p_9m7tZkci}lcswe+*^
zQ3~F#%JS!(YNM;#c=dZwZl%HJk57ZPRGB|AdD=gJKIyaV|M6+i-@nTBvnJH_TJd$A
zuG24D3T{}ub)$X$*Li#0Xd^Z{5Mz)op#Ua8CLuktwXbl8m)Rh(9=&}hjOMEIy$EfR
zr*d{ium1A2et!o4*l!{>d7&ew<gJ^Psh3gv72gK<Sr$LF>FQR|m%NMM)R|Ff7Xeil
zvQz<;yjN{Q;Ob&+@v08<efHvc`}oPD_T<ImwiQPI^fE-XH^1_Hf;<tRd?;Ii2py(|
zry1cF#(Ge{{1^&E*sB<pZZ%^w>kKH&&XrYD`cR<;L)a1f3Mp8O1L2PiJ|SEf7uISR
zr@T-?Ieo~xFO#Q29EQo@c8}>WSl&^-3nO4-hl!9d0fze|8y^A`42^u+YJ7?*HF1b}
z2)%j-%#b#^SJv!hnVGWF5DKJj0d;xR2g1HE?ZsqurPrxGhCv)A^h^6N`xu~N_LAol
zAz~&#m`3G$50kBZm>k3D9o|^QXyi99v%&mg=rAVcm{*rM)E9lLHl)u&HV&A+&g^x;
z+E{I((w5Ad@pWGv=E@kbyu~z)nAf~t%tH!RpSFXKLAcL@-7sh;;>Ta7>vWxd0n#s|
z;tT%zn@T@i_Gf?pW$G;w_A@6n4z5A!w6f@HF0}L2P^Xo-TR55^t_&ZBFWSAkOYQS-
z-75=biYuGZt{ft)3!Ndn$Zv)Wgs;k1M$yaQdK-Ude4t}!4vW`*)`ZRWac!yx_k9l9
ziv2MLMN@Nmr-YQL7axE=U7$H(S_pdwLMrd0<E-=L^IWR-Dl2pt?e0neMib698iY3_
zfa&F;LugfbZJ}=|kFZX7!pU*o&s`C88oF4tF<oCK{F|FE!G$)2nzo@aUOS<oX?t3p
z4fkHg_PEW@Ew>vtUvJZMtKm0{A)$g0RtqWPW@y@S>X>L#;{-;tZDevG_3@5SuabM%
z9O15poE>)ML2!OdL2{D1(x5OL&b~3#uU;Ost&OK`Yx{Z4_tXUIYv%y*Ch+cgJ3HQL
zM@Qku&W`dPPwvSY8{5$C<z+?)n$m|;2Y?^6lcSBcv-PB{tvzVlThGFi_H)*ZeHk*P
zt}eJS_R7+J+uHFqp0=If|4w;^O|;4Bxxj$u;Wxsw5K=Slx_$e#QV@Q^mlQl?1d4|f
zLI{87Ey^$a#6$@S&4VyNyw?2Q+1qL7nMd6aB5`(VCi6d9d<cGwVG_&e9yUvut<$uF
zH(y{obrD=%lmS(@DyOV4UldE3gj!R)OkO;H+#WyrAmg!Ca4ru8&}sNHR>hQ(1g~Ij
zeB9yp(;1t=97^~}e2h-@u#gl|6etBJ$=lzdj4?0jW;BGiBh(evjAb|r`uR3G0d~^Y
z$(c4i*@bNv7wN}|&^|&bi99gdoHZwWsk6GQ7e+}0qZFaMahwrGF#q1W-)`Ui?$>hf
zyxqS0dgkZ#Qhreq<9UUXB!DA8>|qQ$%w7t$mO5W%437`HaE>yw)-Tuytn6++Z0mug
zq2L~2RyWt5*O)Uj>tjSg1`h2B@VmLP_9=3vXIE?fyD%(|(sunIxGn`==4W|zx!t^Z
zJNdfjp7W`MCJU#U>t*tarZAT=Xz9mZg=7@=O315L=iv4AK>6EVQvBvUFqJXI3r|eU
zW{<X1c!L7*(c_0H>s8}*c<c7P_U4<vme2cbe(6SeYVuYbADyi5hi|?8`L?ujtAwJ*
zhX-x_8E?q#Hb1}E-hAux>BDsP9)y!Tr&#81x1Eg_rD&WF9$8Kq_AETiPL74OJ4!y@
zeH4;I<?U$bcnZOLrSQ(Y^zn504u^MzqB*B{2TQhgrML`SUY=2$bVGnTPiA`I3Korx
za+J}_xj5Tt>+28O#^ze51KB(A46=sd$d)<gy$WaLR@!}7VYzsD&MjnZ75=2~1V4nl
z&QD^*FcJI}SP0ICmno}+exq#SeYdr<UU;QEBm?7MY)%-}%!|kuH*ei8FVd;3iR@;R
z=de7-jt(j5z~+9zg>(A8_s+N4qemat9HVX4t$jKAFp~Al6WUx*D1krysL+S7FuD`C
zIN0B4PoI34@^{-raL&uj;pY0&wi_6+o}JwXj<&X5v<DA9$XY+H{=iM0b_Q0nJ}Cze
z_MTVUDg<K@O8dmD_t1|8;nN(Mzz{KWJ|5hpd`2deZUyg|)9tN|jP-uCl{X$7%5bv6
zESi&(-L|#4-WKMUDh%Yv<VsN7OxD9;d3f&bZHJa@R7gMc$yiVb@UZO0qn$4h@{z)m
zht}$iTbZ-a!mQaRkDs)E|L^`o`_{L<U9e!BD0n`6|D*QxZ+yEweYT$VE4$-Yihl2Z
z{_|g~wXh(}AH&gHS?;w@ti3RraMaJIKF_`wUbpuioI_L?HF>)1pd>DZC%ac6FBvic
zo=QFPiLwT=feK9<oCRi_uRgypSBg?05Kq8m8iH@A_p<M%G^K>V<9OHk-#m+X%xr9K
zwnJV@87uU~SxMjppTVn}hcOR3e9HFjUfP>Ie8y=I-p*V<e)6<EeE2y1Ku36GcEizH
zgTbrte3$!#5Am0_me)~!7q1gqs?JKu+Z7kR4_)aMJbLxpq8rFr2C|Bh`#_W>NBT5@
z_kj#W9_4m91%r{RkiAez%5@JN0R#GW@7}$F8{?-Rcs_>O-qTidrH|-fZ~ss!#}*cy
zKLyV-)}g@aMmu$cMDRbwQF&{o9()L6B8uAXwG<qKz61t@u(W=rLoY-GfXBcFp2l-D
zi-9!7ul0Fj^+wT=&5g}c+IA(Z&KWXmzt`7!mi}EIj6XOb+bKCz<~M`e->q?_jP0R&
z7Z&Hsvv4wNO_`m3<1xEZPhU%xlFK~HW9}$(x`J?U2CmTGj3X%2MFS#V>F1?s6FyD-
z&Jp8*N!e#?$Qhks1}5iu--Rc0uPeRur^^HT{B+)|WHuQu@REO5SBD>0ns_xv{XM_?
zmG4=5zSmwO^g=&NDVz2BS^ivJ@76!>U0rUd1MM3;(>_YC?typFgjwF$*s8TxW0m=+
zK3F>b%lrD_^V4>JmR1g~o<7a<=dVG^l9cfWu6pkCGiCi%lM?DmpFjEjnXBsj^w-~L
z`r$G7%U)K{wc_hKU8i5NG&C{VZoctad-DrlXfq4*;l4tMC>Da;EqqpD*p*=Z@U_{9
zgm6(}h<APG?rYiPELUj0WKJ#!xg(p7*I@j7peX|!SN9;55C^{z1fPQpxnG_{S9gnm
z_*x$+s6P7*^q~lE+C~k6u|nXxD=lJ|Y}%Fkr3}k=9zpxrR6Ke9tbO?4qxR&*({>Oh
z`sHvK#gtR~*kr9uMe-x2{b2CUsO4wxLfVChS16KP#;6#q<Y7gGAZsQ|RzeY0R3(H;
zp7CtXs+}nYvsLF|h43r0tg<l#`9okaoHl)>G)jAH;=OxW#MAj#gdzk2tf$k?aB|u?
zP=37((_MAvxlKDp3!^85A@3}ka&`6og-Oh&5tCWoTd7l+6=e&5U?YL~>&zQr6bqg(
z0C`U7g~2gj%E16(J}`zDdGETg_~q^&M$4Q~pbG8dxA8J>j5~`-P2B~DDTmh@W{rzH
zYr#R7#J-6rpa|nf=-B6>zGr*~4BTnP7jIKckpAYLwt{Zw85FZsLZc2Rm|50I@|IFE
zzin=Kp1FsSQ>QsP%?7s=W7U34(nU5CLIwc*>vWy2)8BIX#q{_>+kU~*6;=LgOet~D
z?cmuOgR4qPu#rSy(LH~iL11G)*%lWj+v~62X>$uaOS%xl##31(+z4$eA=lML+{V<#
zkstwX_e`~D8G7Lz!VDX0&0vM^J@lc%5hinmvI{vtC-5Hl4XLbVQ=WwWmG@2fg%S#e
zCKbP-{k7o^J+q+=9W2_YUIMVtVYCg6)lLGtfwDzi=!!n-kN#4c5JXeV5Pk>~rff9I
z*$f0yi%To**4;NsVKp>FIL(Xv#3;03p|XS<9O6-d)?b!4$MoECo-OA-`YMFfMCk8n
zDH7cygdP%h@bb9o<5|}kAcK){4qe5sQ81Q(E6iR$`)jAhk3ga}@_EKPfiR(maO(<*
z$%)z2OOe(UX8xU>Y&4+<55lKO6menD>UsD005DykOVF+Ct@Y=1eh^{m-28OAefLhA
zpBL^*$inH2n{y)O+vF7GPY=7;%L#vSvfVb;9%UZ)%0q^56RkIw=FHsjB3jP)64Ddi
zZf$Qi=N@%lghI6o<_2eglwdla&Id7ngf+nX>60hzS$IyuuKTauuaJtY=?TOsqj(bR
z?jE$w&F%L5*?QaC?+5T3dvm|KvQpu*v_+o?)aDl!E6f}Cs~&~N72@nq{^UQkci;J5
zc(bz-T=Dq2d^OyhLo%=cHy^_7jAa~4K=YCZ2OcQjDINlY#gO6L0#6LA2D~gmTURn=
z9`ZdMO+JP|6a)L=Tb=WE(PpP-)8CoE*{2HAk<qNFz}dw`c-QcUq%r1cYGx_LSmVyG
z8x22fEe&UE!mnnXQTFi=+uG(0_i@|V*=n;h^X=YiZ<ausHwCXNiU+(Pz8e2dsgPf(
z^SJG7J+J5p);&RH@p9q)e7_8TIh64`IzC7{;h(AU;(74V$Hn8}pRHf`rb|tw6hv8A
z3PYc84)2cc!MBk45&3}E;9~Ym!o%Nt?X9+Y^M2a9P<Z0#h(`~vLwl>awzRmKeT4B6
zK5#!{bXp!tUGXreJ1{4-XU1b>Bz)<JbDz357r~+suoTD2Cj15ib;eO(n*wZUd8Lg`
z&ZVrW%w=%ItMJspYb&d_vKPFSeADf4Z@ayC_O$r@H{bkxaLjVeDbF>YkmiFz%{fr3
zt2a|__Je^_dy}K1er{bC78)n}f%CR0#A9#*rIGzd8*)}m_CGJ3sVJWme*J>-<GdXo
z9JIY%hyC}^vlRXm_;~G|o$aiZ&01@W7A`OHIRtkGt|*)?oiAjMa*{H3)83b*K;b<!
z9o$g-{uSP$2o@&LCwvMnmoi`AQa>n>l+8-teuqaLB_9PQFN~?l;P6b#+^i5|lzg_!
z6lwO+*~4U?aWgQwx_Z05^ConEc4ndG!d?hXVmJu@!x<|KOVeK-Oq4wE%uLpZGo|u1
z?dsl?ynuv|+~3_O#Su@kQX~a72vsj%_VuM78|yF1=#k>YUWtdHGq)(GX9IV`;}lle
zXJwE0`s;7BTet7FaZ07Y07WG@7ZUgMXtTziA}?Lccpnw~qFdw!2qv#8N|u{9gR^el
z4c=M`&RZy0v~0AGqQJe`xy6((w52&gfVF+MJUZv|tkz24AN@Jl-E7-C8-)vY_x9V?
z7G)v)aM2z-dQjea)(Q_8>(0K7*Bb?Ir=73bv*h_t|LH%sfA{bHW82=|t###WtM&EG
z)Jy3kG~R4mT3T%@%Qp(=uCA_Tt|@m{N^yxcp>N>GUL7w`a)C37RNom0z;}#AgNM@R
zLbLhr_(({%9vU;ZGv=M%Q+QEqRlhd3+S=NBDKvPF*12|B1Kq25wmg&W+`d)(3@is$
zC?1R<u0>vFq*<@#g=|82#QUQAFbv6BMx*g1lw(2`>I3DE^3gKhIdwi+;NGOEdwFHG
z##3lR9&<v{A@dTtBEi3=5q+>0{4SY_PdBtMZR$#g{;WpFr$Op>ufBZdUfzW>$r;o&
z;JGS`*YnrMT0NfUeyR{Ey5=T0o<R)Rfidaf^C-RhIhwa`->$w>pzdW%$Sv?#JP0W&
zHKY_X`5W+b-s9QrGOiSyjAg(%8f{KW77;q<eO>X*dN27BI#*ziE;~z$_q6q9Ztx*s
z%rFE~W_~DCmE+vlxzq{%oc9G*D3Zz;CS{NTNu4LncqN~_HD)Ob>}pNX*!16T{IX}-
z0++*WE8#oz#c%DYGP8!juAX>jD(ed0)O*(3D&ZIKN#@O(H16h3wyvPcm(=Cm^5Dw&
z!C!+<b-Itn@S?A?#3;N{PUr2(ZPeA<qi^czW7^x$lkec7fA=}~-jmM`WZ>x2?^n<B
z(eHP-Hdc6s8pHfHUj6DkpleXJG4{(!hbM$WJ<9rb4xY8u&;7+i%9C_8Px{r{>9b!y
zd#VC|u*@p=2k-q=ul!fN{!v$@Prm%(CtTLG;_EtHr(d!(yS&(5d;86H_x|g3))&P_
zFm^EpHu3`;Ro*`#`sdu4!vtX9-+J@yvihmm(1o!3yM?&DS@(MCVGSX~5Dj$%siO;?
ze-;n+GV-OrUjMNS9+X`Pa!m#?k11x(>;Ovm5(0R|kPHdqfIufJWQ=y5g-PDr*=!FU
zKWHC6dDzzTeg9-H+hK|`))t755DORqiYFTaLLc##yWoMVJi>Ny#YEY-5vpN0g&Xp2
z39SZWhY03RdHpQ1E||0N)6UT_Bbag<YQisFwDG}^VOq2ubK0%QdV7S%IS(T-8D@(0
z4K~EwXpi^MdYfAe0OC&xWn-vb7O<=p)jd`H!I1GH8?EsbN7*5-h%;t-GVQ{+VcZS7
zHcr^ErcTVggU4;6%G-+o;WCV1$|g*y%`HYLb(o0SAmp2|VJT*_V>6>*UTb67P#GuI
zKl(|a<UBSTcubDa3m7KMhH|uJ9PG)HwoXyT6?2ugCf`ZQ@EwCrsf9`8{*S@4v8{He
zC&01#mp{S{?b8=+7|q;dVCs*Rrlj8v#Wh~W)#k!x8PihA(Tt~Y<hmZPx=z>WI{odY
zUpU3r6(#$HPesFinuIv~gf!R`{Xnk#hF1r#+L#xi4kC4a%kA4s?aRM@zkT`FK9`O6
zYzdMPyS}mQiYGLt!sX^QLU?C)l@)7fJvxoH)ftSrN->c4;cE%qG!G@b3La<iip~-m
z5Y$r+5PGnNMl-7M8e8r4=}R|&8Cc2^c$9ED^hX;%6+BSdbYm1k2<LXm*M((iYyNny
z5keWK9(IOBX*cjt7U2nj=gRWU@M~|im7Djf|NDo#C6IJR%*gmkYU>^$svd5vXj5i+
zV|A|_$_F$`xjj^yv2<U5GM;C`aOp4Nby}gYjt=%p2*diA$HUMohB{p#Ga^i3$|LwZ
zKHSN(%QiVP(<Y|p^G@25K`!sQQ3{ROHZ`*lzM>0ZD1179pbhH3Xs1Un+ScZa@F@p1
zZewY?P+29AAz`g{fg2&;=JNhrc#_t()}GW^S(H?_ZoSqfrj|0!q5XNE;&6JJb^c=B
zrBLYe(%No&_UJq9`|p0UogD9$pmb(NyT@yK%|j1=C>+^X@eM-NITwpC8jlEu@DFva
zPsX$qmv|rZRRYSiG0$H-f7ZVFjjy%m&z`qiH*W`~Z{&V>m<izrC&F*MYzGG?ZD;$S
zy?D0Ko+KaPCpa+wH*ei+fB!dsv;Fp0f49B<#^>Ae@{Km>TruNPz#MqFNc;A8+xNcn
zo%VnJn}6G$s_WM65@0Vau4YYS{=z@;j#!vq46GC0W)3pWJnYUdPST$Ag+lR4iPNvl
z@n{bVMc}267$+VcCwZ^d5f+jsPVVti+T3{7gxo4+NO-3b&SFmG9tvExvvw)+K;YlG
zTa=Vz8H)~+a6#rcJh}NMKo>rYkRKf2C5`FQ;!0awTrTASMUl{f!>;6`aOvkA9R;4A
zwZpyjc5=AYc6K+?U!IXS+SL5LHamYSurgi>F5^wHwY0L@7FX`4uF1B$wUzl64mI$a
zHNX;GKY1}aL#WP?`eY3;COa%8Ve#Mi`tZxZjb{jts?{6!vp(*njXfOc;r?EG@#1O5
zEA2Zu4{Y2iWf?*3$;m+p`1cNk&=V;kyl2*#vF3R;Ji-GYV>V<BX^T&5y(_CLETt#*
zEtxV3DOaW<u!$cg<P5xK9*&OotG`o${|b3IG~I@?&OO7Y-@fx&;TdBpJR0v@`v+sr
z!>kMjGQX6Rl<q>WK6&yqaK90Ln0FXYlbP&!I*(2ed>GtyxVzcTj<(x**3@X=Le!B|
z9z2vjgb&BF6Lf+rcy?y4dj)FaOx7h@(Uq4XH3Tw)k9joh?yQxEV!EDnve0H1Udx={
zYcsRUsUv+HN;zPEOvp0(9v&?ei45K-s)UsjM((0*Z@tL6+N^z=Ft4)<cgst5qC$OU
zAC@*#^gVy}C~&k_%2@P>(%L?PvR~*qin6J!Ih2n=(|K>!{#+?`mP)BLIWbdvmbt9Y
z+1bT(%U-qxE^f9jeBsyI7rywL?X}N+v6SZe%*ZcbnL2jb)5j0n?(TLe-w4zRg5lSN
z#TzMax%$hi>&D8hHb1|T-@Lv~OTauc6Z|*<M+bLL&9r5>aDKjxP6`dU5_kg>6cr~K
zyU^I|owu_#2*icT=Lw2_Qi@&j(hEFEJP)VmZ<K+>cyKcMs!xm<_X7KAWqGa4CeP^9
za(RTl42(U?9{us7_k#<cwdt9DE;Qx!W!Be0=*;Cy%JB2{-S52D{_qd~WBd5yM{R3o
zr@i~$`|X`~zTdXCcLNVuk@j-AtnJ;sgW8vEZ)~@3eEplbzLhz+Xg7k7txfCMxo6H2
zeD#vzP)JQ-=LqFPLng<{TXif12cAggz)RmRGIF&(DJDc?fv09>r|Zl?3K8d!bwPgK
z&)S8D!L$9WeW&mBnZ0oBfvZejVW*|cK#%6;yRwU?a6rA_2%ZsUladCnBauUpD^S9a
zO)v;}{f+zW{u{5Cr=#}3;S_K1_)O-6e8c`5T?V`G^2Wwy)=}0)#+gFSdY}YEQ=dJ3
z(Z2KD@3*H<pVt`RXDA_xudohFsnz*Ae1ZgTOL(&~G|2kE@09GOXv2U{9Po;t{_nn`
zl|S(4DIDv25PHvpze+yuCx7R^GD-$3$)|eJRArC};PL8Q!kOvNWal2v6pT6}4{xVT
z?NoxY$zw_ek@6{I7MB*fGIvIJZHl38@N4{pkM$yBVYHrdv%>WSFpW=L<RNK%KhDB9
zj%3Z453(8jRacCa!YE}^#MB&RT|0BE`Vbfy7;kk&QE<DwW;tVC{ZD&+^shBqd4|{S
ze9B;)2K{C1tKAe$a#zU4^hw|P{8<}a=}xJxeBpdIH@DIr@+fkw<VzowiFVgus3~J#
zQFwxHjRn|I4rN!#*=oH72L^6uQ$GU|ed>v!PL~OR-I7hFE*&#==AbLYIxk&s^zPU1
ze!1(j!%+Y3UeDuqwKZd1f4QgS^~ml1tF9X3@QF1hHK%>Nz0;L?T?Xhk<<`L=BMge|
zgTUBf(R}9xUIf=Dufs6-&YHMt%cpIU<ooQECwb>Dl?FxqQ2J~wKTFcy!MJ_)Qh0Ci
z-;D&DUMs$?({=hKOG~#`8pYT0tuUq`uEmxh1VP}Cw#`<N1;{+a@G^vthtkc}o9%O-
zd$X()5#dj*BCar-pM~L4V*Tz&525<xVgC7sxOBqccfR+VMLc@8x{>=mkKc$(F^ORa
zWh*P{VhlO~fMkmm$|shErPTBF7wx@|KWLA#IoZo5=Q0dB<=D$?{FH@hFYA;rLYR?M
z5}K7hhtR7;D>R=_L%12|2MDtmvR7lE?oJ+)@;(7uSp$c_+aywaRovwK5w>9jI%9*W
z31d`*t_!(3lMn_?A2a60!gHUca7#J3dX_Y-_|)Y+-ghOm$=etYOjbV&q7!59XyHIW
zr!D$sJoU3eO=Zk5Sk<2V^=D<WS?^-VJClyF$+%>~FuvqbrQl0{SqoC&d0%^kN5r5M
zfTc=GLYuoff8nYQgx;7d8#4;<Y<Ts{{Afcdz%T{{bA*_@(=a_LGvmktu#}4C9pg3@
zCMxwyERXez$0B&a6n`=<#voS-Qqw1G&`(SuPeAvCBQSdAM1Jp=RV}4q?hhDpbqm!~
z@u?3x<9nU1({=h=OTRFRufjEdhf>kJ{`JQv8$P&G-|FhO;lV-b-{c8>sOMKyz25VI
z3c%mplX+!jzJ2v~e!cz1Z@kqO=AAtucc)SCtc^OFiB6+^Hk>8EzuH_AzOvX?Za3QK
zG|fg`=plF>y(pTKdMjjAXp+!U+HVtX(_f*u^2cj}a6}&ol6YTHY@kPkA?PeREH6PB
z`i;Jpq9E-hz&hCPN(?lG)p=L0bgyYbg>HS_&!R%}(D^QS$@pE-79j-*Fh^&Xn$TOr
zL!s>#CuLwVJtO240fsT~5-)vcdE!sfsiIS^bnzuVN4PfvAmzJ{?>nwAay#3brMwX4
ztwP<TzVV6KwDGKjo#^y<==acQKij1_L+RCNa~_caoGjtFv!5nQ@LHDoC6q@Wg&M{P
zpqb$>O>O@#S9q_B!}?icF>W4M{tB}=oAQ()pZr$}lKdVWPnqQf_Oi{)&6V;h9AWV#
z&WQpyg!7biJxpCsFE4l6#~*yJz4N_qwf+6As)v<70oCr_PR$uu>qal#+8QrO7!Kz2
z!5j_*)p!J+Nnn9OW@(AiXEJ={QTy<N_uF^A^^GQk8ZVLCckZ@1g8RV2;^Jz%b?aWP
zyKP}^sf5B$o<42s8|wv6JiYF}cCUTmi(hQ_?|-iG&M0M421-N#3MNVp-d@6bz4N_y
z+IPSG-Sq#med*V}*uMCsFVz`krDzQADPn`LfQ0bInQ)Czn=q<)={kFfg76A1YMjRk
zz9uHj*=2Z}gSN5$ylrf5)cg|MkBrQuJjPfVn@i#OW@}v%08$>~i?R#@f8kBTYZKBt
z8;W9-my9se6l+x_yi&?VSLcCU)(dYLhxgyOal7$cv99pK7lh8?my7St+)&~XmT#;-
zOBpBSMaJNZMgKIV!UzvQiZTXX;YkbD11IAvX;aoyN_lyH)D{=IwK>DLlhdO*tH>NX
z->xeOjZbjf<x!nkg#Wcq&<FU8fFI9WXYY-J<AJiQwWED!PHhDSmfF3$uNO=ZqL)Wu
z@CrOc2nY{=WqTC-JEdpMQ3<;Z*dS3nh!SyB<3%XBxO}4tRfu=p*m$0Gx*ND6=ng-8
za!~bmudwOjmnpKV%$J1jr)_FxuFVEs%#%=c+D2JQ@LIy&i~;;=+$qJ(1^mnltRjP~
z@S0g_H*fI5oX^}jqbcj+=&0@QZvOw-`_Cpzvm?(F^Ux9Z##NysQdO1&8V#TzTJO&6
z8fm1>l4iqKd_{hj{xY<Ew`QbKENSc+?o2n(5Cx!6BC9HOW@vGBXvP2U@Arw!DpWyq
z7pi+|;zT}q?|q*09O1{0yC3)CeuDG$Sbq`bd}<nRAdFkWZ)nGuX#E765jZbdwAC~f
z_<{e>SV5sQsdBXz6&^#!r!<c643T)%BmCCfA08$&8}q-Q8(c6W!hAs^&h1rKtK)rz
zBD*5`qw-ZAVI4<Cgma=Rqobm`V{F*Nz^HYDX~KSNY_7+~`l_&CBxW?%@T%iXC-@Ix
zXu(^ITyZ(#^~|B*esDmDE7}bF;0L+Jv!Y#1gYF*+fBK?D7+vueo31U2rp^i`W#Ka+
z!040coaWMlPd<#NPadf}LcNykF*dhwBWg1%f?d^kgAubnyBrf`;kb~hwYM#}5RMe@
zLyYL+Vyv+=$M9tKJ$wHSY0hh`4w@<_4G!=-Hb#Pr<Dw;a28w{Ie;L9L`VM`AhNnSW
zW2d=BxJ)*(^ffO&{P2C@^`T(S8A@zSi1_s{-iiP5fBe71Pk;Io(dB)gJI5K(go$0-
z*od{Y=k|bl_G}~m^MCqBpTqThbv>RwdmaxSJklp;>VZ>f@Y>yL>YYvTHO(olOWE{z
zaEJ#JMjqxCxS%zna1+`LjLyU}_Sw8B7+H1zx46azn`i_h=J=_V*6?X$gVC4~&A%yH
zedo@dxOwA-y~OYi<J?D|+a#Q3&6tXRjCXj~VKjq2VvI@pL^PfGg#~Y1zzgZdjg`1{
z`=;8rAlPw^*@<aCUTO~>Jh1VL%{%ZTXoVxq2=9<r8k&1G(J(gOG%@6Bj^e!wuR>wZ
zSfM9C_Q0!?Im()x>46Ld%(#c)mfz2wK95I_o(O*r)#gE1zkxiH1^{#&td=Rjpw<2S
z;=B4zKH3TY;SA2gkm#gOuk-n{fRaW8aFca6G9d8xr|yBrj1j;6-0$n+p7K#EYz6rM
znUVI<W{h351-S7YFKlptexO$X9vBb3u~whxyP(IsiyVjSNkVqRYm9R64kM&7UVG><
z%$o$~!qZF@sx|pZ-zW>ZMOnZkWkC3Jwp&@;FZu#LQ>G18Y8zuq`ORiqxZP2{A&pad
zW)->&Wb=e-{tsM=#5d#%VCN0ix{zzYP4E#K4sF1w+UazzjIhRAzN^jXc?bu+xw)-t
z%k#32CzE6n`V1_=4eABf7~9hWP90N!DIa;6d~7xY2IQk`jECSRO~&Ykmod+H7yJjV
z<ny=-f7qA+42-*a$;LmyfSyuD0RvLb6L4sND=OpVdxK28$pC01UZlMHqAb$O<$Jtn
zlg8Ef$@^*&^uE>X*hAGuFX0Nt`!n}jdBIf|lo#zW-s?WM`3+7KDer+N-2Vg87i8!5
zA6U}orY|Zd;j1Tu&)?_zmrQVlYs1%dx=vrGBzif!yc8?1+>NRPBg=jRVv~nD=k7}L
z09|7uup?q3>_}?V>TzTFh6OYPdt0mWpOXm|<>-zH7{r=;<Qvv^J_iY}9)9I5@<7-J
zdHeuDVL<HF2AH{pMllQo37853Jv-`Ph$*-$tA6u73y0N>wYdN2L9Fep$I;nw40F~4
z0;~j8&VXRihesG81lgsBa74Hzn5h|}s>VoyFo1&LOag)#G;N~2^kKlD@K7T!MkGGz
z3(VR=h)SX~!;~j5Xr+QN(ockH^l|5U$8Av=2y-a)EVPNy=Q$6jZZ`{Xb`#~Il?4<`
zTFf88L3!AP1}v>S14dLnfI0sm^zdn8n!ZpMLVWfW&o+c1TAt%TNS<_ySGz$Pm5rjv
z#j1BlV(x?`$@j89zy?L61#Gno53v(HXHn%{r+`rPl7JgVHDF@Hm|$uFM|GwEp#CX8
zWg!^h{l++;G&xc}+6t^#C?WVDSfThrP~<%XN0cV~rr#_!Y@AfPEPU&IhZmwuwcR`T
zR0BdU3EZQt2!&kwVOU@Q1P0&@i$P$V-g$y64ZKd*={o&Oq_52I^@F#+_eYhJp<kaR
zCV#BOf%!INe_r`7wSzV}$5npr>E_jyzxCnYd=@2o6N!0(K1EkEoYhg>y}KOW{q8qo
zZmuS}o(2cXKoJK|p`P3vl6=nhB(LDbfZ`S}t!}q#UNTM?FO|VNoc&nv6Iy~{1;tG&
zAcn<b(yM?r!Al7Z1b={k!H3};lv#La(`OX1C@+u;P*~GJlnvC0;S2={d=Y+#0R$x(
zybWa!@6jH3R(fL`>Dj22`7Fj$5rK@8d>R7`3S+#V;{q>m&h6+}!Nq%=`2fG85ogD(
z=ymtB*zO6oCu*l!VVE7qL3ca$`z^JF-TN3j;H%jmz85=_5lSZU1Q?swGZch)2<W~B
z(2A~M6r!m}w)`mApb#A!M<JieQk0!~per}e;1zRvCb>mpO+9#M5JrqpU5pXl80E@b
zl&3hiW-=znr(=weY$GW8U(}b0^11M?mdtUEfd!){p@~qw^rPQv>v<XgP&x{D`kn8i
z@WcaUA1|(hu5gN7$5rE5Pw!zs<v@GnT<`!zAu^XYAMrlp>NDj=>_JtnPDiibjdrUY
zjrxpbGQ9Ti>e<=Zk}Pl-i%ZM#)?43<SMI)QB_sHU0Hi=$ztISfy?Tvz@&Nta-7Omu
zU%7KTzVX&u@!D&z#;dQr77Gi@!cFi}G7jh92wzd2qg+M?B9z>{k3NhiPo7y`oSU7o
z;uOOi9z7T}z$KJVX|zCropZ`Y$+4*zv*8OPBe3<%RpL9ih0Mj-O?Xq~87TF}g2rq{
z?=wb(XkzEJ@O7eOBM4#q@cO}%a<YVCQTQ!9@i}*r>o6dV2|o>2%VEODk#V6PGA?26
zFg9Mq-X7tc*42v3n46imHyQKLN??s&zu(rF=e%d0l*^O$<oe|=ey(xSSWc&54nqUp
zlQ!_E-z1EAgm?oM8I}>b6h&jlbCIz`wg%qd0CPjrO|PE|uhRHMQ5b6&8~TAq)7t8T
z_~89t#X-L(9Ko0ZJvd8HYVM;f$KwTsd!sQcSWF0Jcp$Y6qk3ae_vbxcc!(l<L*u4|
zCprIezDFGB14TWGLg*RBbrkJ*g<)W=h+Z7Ty$|0rZejc+d|gkt+3mJf_L1nL=!VAh
zQm`5wM$SdST8>h6&YnLAxcj|Ua5kC^@!+85@R{(CahRASl%iy7!DDJ-+UE^6G*0Z`
z9}#TqF@+~+xuQ8;(wcI_8!A~Du{PoCC3{4TfPajQK1?@+o5L9SG46`Sj2U>k*i46U
zUi4U(nrGOI<Anr%k^cgN6W(!Yc=tO&$TB`vwg!z21sKNhMje|FPEFKoP_psooE`ZG
zrokASg&i6T1|B*D-#QRZ2yc|3UYm`DrJFWnV?@W(6hkxo=|pQDyp62970Ccj&>DMt
zYE0pMxwa$(UOza)s@dL-c6&!xQ}quEp9MGi0v%Id7=xi<Jg7$F<VbX<xoMim`lnKB
z#O%ysOiV8*A0b4A2Wt0-=4BA>RdW?O0M3tSyzp$bvfMD%ywu!;CcsbdKmb<MkzTMh
zl~arIwD3w}J_KDEo>sfkTb4EwK9zMS<v=U&6dQrYt4=mVa5mbh=Jop@eq^K7nC2XY
z&*x7c#oztM{}BK8|JVOJe(|f1<KRSdU2{X?LEO9dApYSWej2}c=iPWL`nJ9<813~0
z=aLP(d5+SQ;8eG<6+;c15{8FGO9U^xLOCytO$lsT>2~*Z!IO#ofuT<GTVv7f^`qSt
zPcXkn?%+8}dkke`>OXuKZz|!Pej_K$&er4BEeu|_^__W)VRqpCqcT~mVszST!k2qC
zCiMD!%M2JDSOl|C2CpuR7kIc(F87cz_)geHjOuux!k;osUq3o+XvMJlTetTS>t^Nx
zMl{Z5!YIsKMJ_`5j!^~~1y4m}8|WIIL*??cahc5-v>9U$_}l9dF7e2QYR*O5+S)c9
zq)u=I{06CkQDL~qOK78{+3Qd4LC2LXacl5C7x#EhPT%9k;63LRZumYO(;Vbo@Rv{4
z`9$T(+?pAJJA5tALsy*x8bdXd%(~v2n?y7CO&PUX#c*PM4_x6zz>Ko^&Nwix@D-lh
zn5}UmBq8*Nelf<t7TC}ja0R&oyuqLZeB4+1=I0Z^*^PLafsX@l4f=*x{s6X0^n?1S
z16ZU!<5J^!V0>bOE@KGZ@SMEXkEy;iZc$#si+j+k!bnIR^wILBU`6AQcX$t;13qxB
zSQ^AI9CO4*o-0e(MU2x#5<+hRZF4^d{h;qzhaTt?oFr@^ycqa`_rL_a$uq7n`lm6O
zi_-MiDd7Q4&bi9iF!qJsi;T>rgwc@afsJZ<k2=t4l8*~m<frh5IVPOIsLeRgu6!<d
z6?n&`Hds&P-~6CeDeJ>?DT`0uqio(mE?1Ln-0!prW8E>lXd^l{4D2F`6zu$YQ3dt+
z3zfW-5+8W3|9uG!erGDmzRLR9VwH*#|GDW4%lRi1UbOpP;)3g48@{g7b^1Cb5$nZU
zH)G}QEysRGxYb-&8VDy!n*Ppb<_QxM)x<c2_srb97U~F_62eVPp}xwl++K?1cZ8t)
zQ82@TO8ZxSRnk+44b|%Tr8D~OAdkPP7?pEQXbL>3Sh%Ed!psUrrQa}}Au(+j>Y)~K
z+q*mQ;e(Ik`TBDkx;QkHa2hO*sgB*UC@a_*!j3F^2&oM)J``hwCxF3{-^@yjY<8K@
zC-3YQ)8x693@(EmTs(*I6-<=pNCal)7J_rf3<fL+EHL&C7u~~9W#fu!z?jD_Q$qak
z9-%leJi`2eG2`T2LaKwdTR4(X;9b0`6JZd62d0WxNL2_#ZXaVR*s&<3AHW=#626c#
zy>iDqLXeuLciEAGr&=-I45Nh{wZ*ae^_vh6DEBaqAs{3J8CJO{(lC}?c&Ayv+l?a$
zOb+#<OGBDMgUWY^4n{^Wp-t}NK==bD)a$X6kaVv0(SP0<S09Ry5R3`%avSvy%2Bt2
z@RiDgpo+i*t^hX#M1((#xK>&+wix8}ccjHAf)l&;3IWmY>N)*F83s%pa!c>}oHm}T
zKH9!7!Sv!J=hJn%PS@!xPhXwk>wAA(!xz*nDd86<o;wf3@}HSLRfK5URoxI)l8CL9
z75emIS~M?{^6EbH59K>N$r}|!lcCX8^szfui)D;9>_Enwq0{YXA>E8>Z6a>0%!~Jq
zi!b11ELyIzykk~nv*7|Iy{q#k4HoPfZU==1yc_<ShBU$vlOF{S#wHZygcc&?7x__)
zp-NAMjlI70+`Ev)Q&{*Gasito@Ct(;@tgOki}Gn7VZcy|z~g{LdK~92Z}Q@iLmv78
zkA|00hZo88)0;z3fT4&*LC*p_`Bo~TFu=f<Qb9?$ti8SMXtg`4r{wxLv*qIAAP(@{
zBAj!-t#^-A2mRNZrzk9Xl0gVF$C)g7b`c&HDEm=B<LN=2$Q3Behi!C%pR>Uz6~?)n
z73D7%W3RFCdl&*RilLlkfRG_jVDR3F^I(MFeH8PAY*S~{?lZMvQsvZjVPw#I8dvXx
z)i@2|EpcUxp<K%2QrQ=TC_{GQ?1Y|`FhV>ltB`ZNt6Z>wuOnopM+<V%Sd^-bs8ujz
za<)b#W@i_zv}8O|(V50ku;1SE!e?ZI^a@0AmOK9$FBDy^=B^c-7|T*I)f2oA^!-d_
zH{y*q-ilXVechq<@P;DH4GMpUEu^2S6OSlD4^B;v$G86ETk+-_Z+Q3r-0YmnCCb8x
z6;GU1MZ&wJUY!vvBX+je<CBj+j?K+Y;m(ldSBy4zS&TC`R0p`^Ii+cgY-GcdWGUtY
zV@kWyt48p;6fBW#P|gSjr-!XL=(UVTgd1a?)EbK_A7wgl&AE1|XKYG3)m&g)@EEAp
z2{&0$zA3c>97nN)hs$wvJ59^Jj2Q-$p^>WRR$Ly4e&G2;xVy7x?(N#M3$KiFx#s)e
z7+zt8cbMfl^^JMBwYe5+tItJKPypA2?<hl&i}Abwmn1`;A(Q4jVLqs0!E#dLeH6{z
zb<ZI-*8#BYZH&e}!IkkIwjx?XW*QKJ8Qw8?77-To^rWZypJ|Tn2&SV^s}hota=~}W
ztr#&h@5ZMvLX8NnebtX_dm7ztOY-k<H0tw$)1>ZQXg+gp+Np3wxF>uD57PhxJ;sA2
zjnd2?Le>opP3j)wMn8w*MDzaP!;g%kgm+t5<SZ}EdBFi=n+@ee!%!Wx5Bf7cji(5n
zILyuAI6G<^Kc{Te0S}?!oH+x%!}F)%P<a|wjREsC>><|N9Vp9f6j8tRcP!bMGn}CP
z%*m2)oA9V3oJnYdIR<UY7A+U9<6X!YQ7THxA!v)*nc)NvO~aAB@nppqK)t{QxgKvr
z=5*2{_(&m>i!O{1jugW_uuR&3L8vL(y1Ts@oq_Qjn6oxJS3gFy7{=2PI)|~Rf|pKX
z-i8(C5N+aYVV|d`d|s2OkH8zR!m9FB@c0z1gim$F`@mh!hhzK?k1$RUZmc+qQm|K9
zHgYS^@UYsiz9*d>jb3{*2*-%$3dVYH9q(26f<1DD2h7QH&D&g;LI3ccR{0n>1b5bZ
zv;`W6Vts64+O$TQ^*-xAJh(7&aZVyM61@RoG*3L<crw!W5w)2O1+7ji?%)3?ws#17
z+PC4KbNd|9RCvzWdH?bM_rJuCe*9B=WZi%8ApZVGKZ^H0_#j$c@uxGDrEy^{<4Ll;
z-HL8|U$BLLPDZ(e7ZP4kE3veE!}yFboG~D59cQOG6r;wCwHtH|;}@aqFhH)ZuG<)i
z!2`IB2^L<XYJ5{6Ejp%dk-#0$$A%Z(C)68jNxX`QOtP@R`HRb5PvaF`7>M?EcVk!c
zt*!Pq_e4AU2VN86C4;O1t#SkujWOef@e1An-v`(624{Uqd*Lg)yE~?nj1}#4I6WoW
zk5?D;p0(=y!hCw0RSCN({-rUsx0`U1O+owpzV!gieRvl}FO1iFs$)lOhIbKSj-~zc
z=j)n#d(mliZLEV1!z-8*X`sMh0i;sy^QGVLmjodAqala30aqpJwXDHA`pX=;r~Bp~
z`~uX)A3rjOKV7EoabKW%kw=xXaTj?MLkfJ3dqsU1#ONn7mvtQa!w3WX;1|G#;~s!5
z`kq2pgDe0aMb2~hNAX|AtI^PWS8I@Ak!8^bkl+{0A!JADb$@uE|7Z*0K^bFYEbdVk
zuw=d`&O(O<ebaNx&yp2*2LmY2@Gv8^DjS7q8+gon;1vC~w-+x1Bh6#zIlpO#>sA}#
zvET;u1w14lWe{QzIAJWs=!M?Nv`;W)9IcPS2rC?>F7!R{V&)fpzVce5e(Dnps1Mw(
z*K3-CY!1$T0S_D6)IK(bfWydO7+x6{@RNMt6^T#uTWmytCb7N%$GOK?k&Sm#j)UK4
zf8kfw#|id@=b-UI@8!jB))kf%6LwDdY1n=0%5s5S@oviVlutPrO)s@3z`MOK*=#S`
z9mWeDkIqSJ4wsu{6y;`_`c9=N!Kmme=?!iT+T?G8_loQD($#%F$@iJRT)%yZ7k{6*
z_g`9qI$aySuG4k;8YLFkE3e#&<=b~+qRe7Vgj<XO0)@8>METT)f|C?A#Ow%j2>B{&
zVbjoJduB$995J{gY9gK}DfkM%(|5BeBw{W;^mzdxcPW-f@>7?MUb;QFr7tiA@)crH
zmIu?zgCN-njbUh97ea)d^S+*6sGb&wQm;La)y?&2_c}zF(4tI2h?vTmm^U6r8I~q@
z1tM_x94Ik+s$*f7Ztl6u)8Y4&VCpbX;Dg}k`t&`$h7b%=f#&BD2)HoR3~9kaKk@+h
z+|lAQ$^>7|FF<+6g4jFQDM$AZhPdZ2iCU=XZ3{{;Z-Bv~(;i`BwBD_!dJ!%rbg=^n
zB_jqw!%2ccD(_Gr7Vhl0s-q}K!w`ZB43L3A@TY9M*GtIbJB%EGL0{Zg+6~T_>BG1I
zoCk@2FPNQ*dtq3U(C@{S77q3n<h~ar8pDyHI5k`|UNdd6;IF>WC&q_5Xgl!bFm7-Y
zcvA_xR)GP6T>oI-4Y6>iI%tXeu6Bd#D1T5|kbpIbye>=i(ogc>wZ$TXom=1>{a|rJ
zee}bEE`?LC+63Nu$D@^<2tMkw7FQSruhVt9PG5QY8X3O+X$cPcMaeY#4@uBDeefRN
zpu?o(hvF&XCoF_hX}|&XhgK95HjZ->IEw%Sm#=mwx}7}0o*m2+6T@-y=3Fc;)cqZ0
zC5vn}5kOC2C-7ogoOB%?gJF!iP~=|>_}qjVn4VoM+TjT(BH-=tDw65H+J_=5c{;p=
zow9^f$sNDBLzMFb#;oAauvMI2a^QN|`OSG<ER-o1<qsYd4zCAq)V)+Th_~uHN*8!w
zp&TK6mHD;W%O)H66L3WN$zmVx4-}1fZA_IaQK`+wkji7TP9H_0`hX&IYN}?>5Ij#X
z=wYbH(MjbPh6gJ$y$fFpI<*724Vi+A`cv5qY!0K>*|u?t&~Jvjm0Ua%FFwKJN&H(d
zNBQd={3_%8tRLNW(;f#BSYm8M&rr@5f*m#Ky<@=-xNrs-3iaWUlG>wj5%1Rx<)f~n
zIO?qnSGGJ(BpWYB2F94+{GXmfy5Z#l{6|JnbP#$CCEticKTcOLD5zhQt31d=r^k5Q
ziNA|S=hAjG%9T<~O9sXprCy(nvfjl|gaHOm5s4$l74RDzatKD_qhwym5(SPR*J0#9
ziH&j<*?}{+kbBwHU28OamJDa-AiLx_a(Gs#SIWa#p@n<~4lFG##%piB8B5E{-mrxn
z0&Y?M?%tN>PRsHjyT;kwjJL?<)_Od7^w>DEva;-OTU87$rSibAK=|f!<8i;&)ZE+A
zXG;yhOA6Un<ED9JL(Zk-pYEP;f$)=(nGZX$x3g;eI6UIaoTlIxf_43BZh`9<D~=EF
zY-lTAU-gz_VR1?0gR)EII%FloqWL<49GM<3_I^`a$ENDagV!Ji1C%nHvC4t{WtA;B
zY5t_54ZO)Zt!x4p)piW(JKNjw;FFKy>C?w?`}QmG`s;rZ)73eTCyGLc4P>4R&(Dr?
z&R?pJ(2!HfY9pdOc=I&(w0>ys+7P+Bw;c~3KJ>UUukp4(vCUlP{2xMYrtuHC7H<%Y
z786k};Wb#Zk{<7=AsboP(8iq7;A_72TH7|z0S9Oj>k~G@VGLnh+pT7-udmU4&G&NB
zTgkiNNrpHTOb3R?7gx?Z=I!M<=kXp2W~0FwVr*PQY0G&(C>ndh(~m#C7n>XFg5!|r
z1xD{Q>SQQX#_Uq_2jknZ@Uf!4O{s0HXIPs+1GUc3IN{NAc6u1C)^7CnG5j9pywmuA
z+w1GAvG(*)Y;LZ39CD-Fhz)s!`P6WWwp<AJwxhYT=JAfpVaw6bPG~(*6-LL)rcHQ{
z?(K+XX`O<szqh+*!!IyGiEO3+(Y{F0scEEf%XEf*0V80=ScsmUAE-Zvak$^Lk+sw9
z8oy8kqewkJJ+~2pdhES7!g*S{gb!L3v6v8DK2*Ot(bl{01aF`e4gfO@4C#r0k`(?$
z*e<*+@In?GR4z1}_ZH@tMFVaI9-+r4c!BOkOEeZ^I-xA_n1rrlL_W|wJvlnCfp@Zm
z5$n>%4caoMIhNsV@d~7;&}HV3_ylV{4Bdi>KY@zwjZNa2Nw`{IMAAGPhS$L-$OT<j
zVQjJx9pZe)k)b1v+gWJ9)V8D8)!5yC@Uf$S*qE-ei1yJ_{$IWGPW(^*)Bi1g{F7hA
z(`W0kwbj&IU5oXN9hEvFI=K)l%Qt);B6A4*0^c~JYHn^emPOl^^}J39PQj9S0KT%m
zQ|C=9I3oxShgV9g)wXd0nE{@M5vkQ|n{G84)o9ET=5W>?vhY2|XIgM1TqWmpakdxZ
z1zZW?h5?8Ro<K-UHt7JjewWQCdzN!p`$G=T#5=%q*5k+!$O7rbg>t`SK2IN0S&x?&
z_?UVQWE*r3HZn-=0RQ0w<e`uCdR3ozDrud5c%*U8u>2T4(M25V9x?_ahJl(8v*;Hn
zhu_RS=&t9ocn|%d4R|*Z`g3<j<MM1(^LMae7JSiMH3q1B>jCu195apxzW&Wm<uX5T
zZe&G<M*Ky2xS*kFh|pSJ<)*%XE0goN4BZ>t&OE|9{;3i!fm7vB&$;2n_t`l@@6r~|
z2}BM{na3MiCAT0GK{r&OU{&&S=sGaAJgG6xO-gL!Ny9T^h(R40m-h&pOS=d&*lf1M
zS9IU~%kjN3pcH@fM^~ZS08Y$(Z?>XL<zbTp#-R+qDXbTvVe=1i0q-*A&>=SW*hnWF
z1b!I3@IC`Dh2BXZ(zBeusTa|2wXHBh!LPwZ+KK+fv|bGcZ+On!uy-3e7~u&xrY3oA
zpbMex+;>H~Fq*-Wy%`)lp|ASx@Sgf@8Cc~Ox+un}m{SzYJ&a<6{l$=1;0JIj^l#*$
z-QY`>fmbCsOen~d>6uGxw1GZb3f6Pr=cKCKT*HHJBz;#M)RVZ&TZ44<B@=zJoTG68
z?uFcudN6$lZo0>&l@rcw?THVzMH6;gK6mi=<V>*3K4;MaD02kB^K0<w#CI}%>Xz{9
z<s1I~@^A8d`Zww2d&L_sU!>ympPO9f%Y|PRoL@fs=;iBQQ(^{Q8@{g7b^02m@zP}6
zdi`!J-R2;0b_j~>;L;_E<sCEp)=!unjKEAzg4=|cA;Mr)LOk)T5#C^EY50J6CHeCI
z`tM~T4aqi>;eY?9%eRz^?<i9AJ@e>O_e`({M4P+pu!7n0zQY?LMCgwK#I_u|aBEkK
zUM;M+_S$=K#Li~-P5oilwFDG)Ezvg=T)YdDFClb_SyCMW70;8ICgZ1Rxwsee6XUf)
zOpK1EBWMs-F$F=4R)cp0yA(QLz60Zynd@NTkX`e87-Aew2?0azAfod)4P|Pp7LP1A
zSOj1UL-?>Tfgmy{lTQnES_Hyey{j5QRYD2yqZ}4hw3#}*dq-vQF2VtAb{~*u2KT5d
zg%AW73{)e=1-$><e2jXOrLn<422A+`hum5gwR(?sv6v?x3jhp6ypN$3VG;NWV#dW3
za4>wa<1=^I18*ng2XHBXYng<=B>^jqj(Vo|=qKmCp|HfrMgle{6o4CW&M=W_G*n?$
zfFk&#kSc_6m1se4N;r=p*oV66x840Euwt}B!2$kLKVwWtDYsXPY?OM$yg%r(3*F7Q
zPS@!=eRb(yo#CrkaJ{@3Dd_QUOP`x(PzL|MC{5|58}Q)b@5O_`!=gBT4!)beiG~(D
z{pDmj?@!Tr@jUorvEYQJ`kaG<gW#+K7R~Hz#<;~Ni*&-u5e|ln#XB^V4GS}~mH6ww
z{z3e2|2O|#y#CrP)i3@bz6|d|$ywDxc6?&YyZGTxC}+H0q=mF;z+W-ca2Fm24ae}q
zu2uLX3u4oCeW$DJ@Fl#VjZYJK?iR*2pUczbOfD2m+)Kk4N@MCoQE@Dp0z(o867yJ<
zYh{9(17D%<@F3oU_adjT_y=w%PT-Nines8zv4fp<s~A|3>+vq#v8kw5=b|xxBgQ82
z>KRwxF!UXY&rC#p7SD}4F;!l0J26(M8odY4PmgwB%(-GH%#m}b4}~ckUXD*yj_Mh+
z(i6i_EB5xb#NQ4CBaAR98(?5MKGryP_v|%rh$oW`fM?=|k`X0$`Ham=$pzw*gi%Bn
zHH@O!=S8uzm}j|LJUz_lvpS5Yg%OPnC>qC;c5JSH61#gS`jJJ&qxCmBE=dmnhw3>$
z5gby{Z9~FvJx0c=R+tbKU}9?8#zNp_j|+`0@`c9~Sw>@(N?i&dOxvUt)Py}F1RUWN
zr=}&}4!rCvdkXK6y@oC`gcZ2VJ?bn84;WjAmQ){6USVk99tv2(`{3C%ta`^KGb1C6
z43D`Sl+%P$L$QcbgmFWWTW{1gzVk6v!N@Qwd_4%xU*nu9hp$BWBe@3!9Xr1}oivVY
zZEsnYef#b2#;dQrZo?PeUXvxx1)6f-&oqY3=2~oTt*ZW`C~;m8^J`yY+-}FX0E<^4
zA@@)ejtTY{yLz3iD2$(YoPpb=DZyFzgCdL2e+S*2Xg8HlIB)_ks9%I2^8g5^m<JeG
zPlZpk14Zn?J_gYdpV5ZW9Z#Gb8@-0{=M<yBkT*b~(6aIkMW^Nxq0AUNWLb8mV-PqO
zOz;xnT#|2p`}?tS^Yti|8^#F$e0aRC{1^^|8=@~L%`Yzr9ZBdj;fUrsO8ic5C!Rfj
z7;9@!1z*8M`Q{fE?Nwoe8HQfX8(`;ogwk4{!hFGPDoUIerShQ>wZ?pmPA+Kd3Fnt{
z^-?%|c`h2&hu#qOaYEyScOJ6nakK?*!Z^|=Waw%YoGg1@69y7Fb3*V>!xZ=oY_)#4
zN(l0jxz%6RISUIbvAn!ue5S9^LJ~^dzVMlIjG(<;dpT+53RhXffUjCT#NPI*o^=EV
z4DzBOf)(pUycOXWc=TXQA#C4~>SWHBg>y4A^HHCfRf*^E^y&TB-dPt-8ji)q<(OYw
z(YR}V2$zw+r-UnGYDcHJ8O_}d(?|^alVw8dYkmkf2~9XUmf=t7!lm#G*yB-$LjPFx
za|Rh6FNDlI5U#S0!ZXq55*l8N6Kg8q1xFC9L|ZS;kMx-e?<v(uXgoscO^JqL6lT69
zuMjVRju=-(El=TlUCk5KrG7(uFusFx<{^R){1cvpA@yAC0zXl>6Y`ER<(#nfwKe4%
zj$1ch)f#(Aur8Zsf*(hkd(h@n+NdVu^*B1F@fR)&WLAE&W@%~N`}pC<HflqMc?J&R
zT?ig<Uf0RVzQ^s#usJlTHiJQ0k9m{9fIeTcmORlo12bLf+u2D+d_X+z2oJ2o*xOwf
z&3+R5-Ifj6`=TS;+iPy~{G52y{IdEm8c&}*i(kC+%lO&PejdMi_kEQ+A^vqUYPDJ6
zeMRFu7xT(HGdpVo7ltUp$f9QeR|`JRm_q|{&2_3dFfKYgDZG;$BADXMi6I(3j3E^H
zgLUjoqh=!k?MiyC9*MSAYvov6;w-v)ObqC110{S7!_t&!1w4>7H*0&=v1}N{OIB+?
zwFToe@(@Na*2+cLIA}CHi@KOY@QQk^ZjUU^>tZhBVQKF%<xwN`E*pmOEGK9p<7@-E
z@Y(tX<)Kc(+Tj5VKc`L1MPv@X(_hXLq8)grQg%=D3;joW;~*QcNhxI}Nn66hL&14M
z@Sj)NWqYVSe*7#t-F@?_f>y9rSJT`M@R>328aw~_0*g{tf*itotkhYb^S*z&4K|Vt
z${lDW^gMYWnf%9ZUf@n<_508PUbQj5IxTq4#qB#c<L;|>;?C_Gn*W4qOyf&QG>FJX
zz|UoCEc4y;{(>)AhoS~rF4GtSbBu^}@lU>EL`1&j^@_#<gAv9jHgVxS=8dt)E9w{Z
zPzHS*ROmL-NAeVKLv}U|RbO(06&txlD+-th?Z7i=>y*Zu4Is3$)oi(L>cn{JkcxW8
zGP%Z_vH>>b@|SJOaR(piBRU$lxq!PqsLS(0&#ALGqQY%fkv3pqY*L%4!((f{o{4vO
zj@KA6Cj1GSoHA6(+^i3fi!4v;RdA4cNXRG9Y~XEv4y_bz<U4qqQ4;6_@+9*k@r=2`
z*su<PR?_DqeW%>xBlvo5$V>fJ&dnT)9Cz!)JT8OpE<ue04ud+V*Ep+robo(Zeh0qP
zYonLoaVR;Y-R(<;?TY@iC2O^9^kQD{TvZvyMv4Ja=kJ%(XMd9?^S!uy*Uw%~pU(5*
zIp4nM`m)mJmZuvu<FoC0`O<&BK6US3Fd-dW8@{g7b^4klm|<nE5w~7{Ef$uRV@!gP
z4Ip9)Rv3xkOtPSoKn1fx$?x4tVm25PxXNO*je6am?2J~aNwlEo`kw#86ZMTBMH2IL
z(mjMHIDj83v&1BQ&p#K(AH+|W8MK5e7Dnkkqi2%n*?|GaMnOX8yY4|R?%n^$;SF~?
z&EV`VLLMHli;j}b*e|>ohm#le6^bzod)&3sPK#0Rz!amR9vEh^;~YU8Ar{8Y`xte+
z%SKH{P=J|}FT<&bF<ZDjc6$)O#Yru=bFpw_0goaHrCJIFX^6vPi{A+J2n;NcY;@A1
zi8D)5*_OMffDr;S3qJ&4?okec3BnM@IE*)Hu9!8upb%zLL55JkPDb88wUPnlVD@_i
zVN9^WPz7TL9tce~LJkBYyuZ%G*z@cM7LN!e2n{LxvFHQlEI1G_z$F9*ghzxW;KAX$
z2u!{P$|eMLj7Q`v;G2RE!UJu<%^er|k_9(4P5t7z##2K4nCfSzH;H!Ujyc8)1EAn6
z80)n`{yay4GFb3?N5A@w_a6e66_y$YaLo%1)enp(C8%N~%dtloq&)RS?c!o#v)?1s
z2a3Jxbe*o#SD5})8NQ%3Tuvesa0e$5kE{DHCx{9J#b5QEi?^;`%&*YLBA<T`%66$<
z)S)2Kq7GU;SlmK8;VbZy!hn^F!(52M_u+TNj$?Qe-{C{>$Kt}EWuul2TZ8+`gCZ}D
zTgemHT}<74B5QbKg?^)S|C|5j&*J~^@BdwV|9juI5(B=55sHN+J6tg&Adk4+y672n
z-U?R?Bft;Y#S71b6&nvw3PHD_`zSzATo;NM{_%{qv8e-N8XFuagT*po4e38^VJ9w%
z0}N!qmgn|_Rv(gQrM$#hkkJ0*)r9mL@HBtEP#z&K;5|TDUP$YG_^~*M%H_N-cI2a&
zfe)!nLc}@TSbBkB2th$IRl;iu4;hSd?6#HkV#8R5>e=rj3q(|Fc+nLKJY^SrIG+k0
z&)HVUBq&KSo){kRwBd2V4Y`0bJJnb7OgtS>h++s4FV*<jU^G#V$uc2qvj2=PGS%q#
zv?@c13a_VqdJ{&CCzuUST;L|*k6YrA-RO4-)7VtmoNH4T4=sEAQ21gH!n+GkixHn$
zMcI@ATv33W+9)@Q!bM|&5sQ#GW0QGKmn0IyFV8zej$}TefSHt>Ls;un)(xv{ywQe7
z%POO!KBk<N^9Wh**r6n;uR4vvja}L*N@XG|0B?3WPwQE&KC3ar%LjbsJl3+xBQ#~g
zk@-NVx5)|cKF_H`#s)tbFX}`xX0D<%sn%xfIp#C2gmdgdZxT+c-HNTv&FFReu_U?v
zJKz0&%+4*S4;b?%1h<qI@lF6oT6>$()_d&iXA>KEy1TO#o0}WzJ4RW-MR?1B|6`M7
z^%-N;f!f9<pY!NzZga*@wK8KIEkc-eTg^D!@7tJBsm`j+jLoR<8d^h0ab&53*xTK(
zg7*0M%w^$;#u%Y6WS$HSYp$R6RaQsyscYJW;*b8930o*w^o8Fj<9H9R(F{w-c`S2L
zmQ09p7(-8m^N56_oGmpqF=n}ZVsgy$wNz=uM5!iOF(iZU8_t0LV80zt9)293e00wS
z_nAf`zV)r|#G7w^GZq(?y)zv-)nQXLS&_dnb`D*ld}D4SYvTn#NJv66X}s+*Q{XE!
z5QX2Ro@qTbw;0u#1slJti`5*#zz!|IczN^Y?U+*@=vt?f22ki?ZUR6}Gr;LfcE_t+
z<0A!;2K3VOYgl-)5g&i_ajdUzXuh8sCvV-l8+Y%%ZXDg%*b-h|ioP&^Sw|2$aVXlt
znO1XK<Ch*viJuHR1_AiQzG(t=;B``K%taM1%38yQ*ElDf7-!(s;_j=j#f@99Ms@C1
zRBH29Ojq!T(pWM7PLF$n>3MJ#-eh$~b9yvJ1(PY^4{%S9InB!v42&46>oGEsh6?69
zo-o{JtR3!LFv@d<QAT6L(Yo;*PaF(@D7;x`WN5Nc<bKf}&dLH`@aCdloEe8f6}cZA
zWtF7tDoSwBJT?y%eW2c%xuuw#p7;6*kGuW-uGUw=FRcMd;4gfnCw|G=u-$3L{Rj7L
z{DLPeEv?v_6C-ST@svH@D8Ly@yqKp2FX+sXa24E}5HDrT-EDW`(ZffoWyl_WHNmi?
zxj4jmUmBa^!>0K){0U<okx01C1q<-Xn_pSeoFAyHEr-Cgkz$0o+mB9jCoa@iX!_K2
zP5n5C2WmS<3*e=@y0#jRpFXwG@Wzch(GVW&?6l*ZU;H{gx&KgoztlYDY{Plu_0sZE
z%+Jlm!UAXaiI0>xZ;{Oi7~1f-O}QkmG#J94ku5N8pe$!SjzJ2U1H56(@#?`C&bl5$
z+U;96T?a6P#-%sQnP>yvDy(f)oi2E*y=XMI&{N>G4t+PD*S+aM2sZe|*nnm+Mz?O>
zw2`GHTHF?{-~~h(lvSx#JQjFUH5d=Im-&MSFR~JJ$07XHZr0(<H+#V79X@OIx@jlt
zR8q=jY&a2vS6=$WQ42O=Yiz-r#id2peRyyr`l#`itb}m|_@mpP{od56F+(1!3ijxL
zp094iM<3sht?eBZoojgDB~VyDVtI>;-^iCDJCw-<k7W)4FV^WPAMx7+N$<eV)MP#5
z&!zAWkwAI!9x3x@(p}!?pWd@vCAepp$njWQo{#Uo{oVMBzxa#zlW+a0!#pxh{ocL}
zQ>??`1?Cm%2l5>7=Uw`ZzQ#HLjXU@87-GW;q4pS0?s>fRK4DEU4#G?EvIBOOzul9J
zGJ^rp`Uib4df}(Q$nrR_5+qeOIKVvsMhHanKNA(blR)v9VTh-l?G7}k?Qze#pa<z0
zRb14`C$cj0z379!4)P>~*<|f?f$mWr<7cCvV255tKL=$1c*cb`BZHGy<4b|M&!!I>
z#Wc>)C5%)>-_i4Mjwdt=oP+15ER%I&l%jqT{lN2)cc~X+7N4ohqRhla=#$&5XOzLk
zI1tVg-4yenr#XNe&OD($%iN;TMSr;%hfL6a;0+$bLv@QT>kk~JZ)x1h%kqg{xWY-n
z0ouT4Zy)dQcI-9tYKj(P^y1upE0<iG!RB`d|7BU9NqLXlTra&-T!VN2wWNPeYqE!*
zPM-&puPR*|zOK`C`Whvebfr;?<vX{u*qYOC;^cDpI*Y8_9ZmQPltwIYVInXob}!+!
zf2jr0q@GQ)s}`k#7?KT8qFjU-H1|-tI=74O1F{C)H!~klIYIzl{ZLAQ3W7JdF?ftP
zmAkvdxb*2=)M7F)B{Cz79g4x>;q`d!S+sgBEsVsl#E22tStyYgLQ4G@pjd#w#1Lqx
z595u){_&?e&tULkrZ7S?deuuAqY|Pp+`x3byHLV6jL;!v#I*TtL!#b=A){oXnd}ba
z(B)L1q_OKj3kHONa|!lz6=N1kmQ;KYdmp#@u<#7yr7hk~DS<r&<qP!@!E!<ZN4k_J
zGrx35N0>e^qbw_RbfM@(xkWw;E4;&>UlXHO-3T7Q$*>-bA)C6=cm)$LgacqD7NLAB
zG<*&Qk0|VtMFGMJyUBB$97dD(1#9<BBZp!WBNI3`CSj2=0!GMG!(Mb}TUfx*mcm%)
z-A$Jg{`4Cm)9|2vHOvMvjX8!Z^3qfmAtzcWSa3ps5iG{Es72UAYz3#l2^4vR452@{
zuo<-{p#@dtMQ}n0MbU{chEPWLFx;Xv<p*Vw*KYH|1%$Hw{-Nj2b-GU1=_^ZLnc?eu
zZ@>Mw%<kg4B0Txuiv(>Pq$00!UR@+;pfA1;F3KyO^OoQB0yY<c(0LS)@Fgz-Mf>6B
zX{<VQs2?tST<NngSaI1SYjEK`!kH9=>}l$gGb>Q`adE$`-`r=%DV|YXz0<*~OXW~5
zb-~k6ghF%i)_{-L&^1_e(s+y`@RIWMRDA#KZ^rk&_a{-AF8Mw*S}&Ma(L<E6oVNro
zLGgvR0?I+c+mZKF3rlz#OLOnW6i<L66LJn(jggA?Sd_y@;UUNc@O2buC<5R=c=%u-
zKv{ZndKet+jR&yLE}$$uLn%htg!RK%gpzF><H8s_=HcJz1%Gx1??s+SWkR_$?RObh
zHn1>0pa?tzg)A^ap?q+(uey&@siXek^)-r-#AncuP%Q~6`sW3BaUPRyoS*GSx3g!j
z4DgI|P-f?DM17VW_nZrXw~%<I+A~?IM`?OF#wX!*N3p%N;c=UvUy8ZKS7N%lsP<uS
z**A=9)tRW&7xmskl*%(vo)Z5<>8CcvNKN&#3!1Z1)K5G!1XIrRiD3*pB$Ol=SNoEy
zGNhXo_Lq1PZK(d|YWuE+{Y-Vu$MVYEXw0stePzLgGp(9}F}%K^axfsE0OZ0DiY!5>
zI^n!zDh#41x5Vgm!Jjd1NnX_N40i^{Px&L4xbSVPa<<Irc^X%J4};^d#(g}`{y93_
z7YvUaCXjR6fJ=s-L#cu%pYX0sXtSzd*l_!42O|Uym!~S0F;KKKu7p%1j3uG?rZCPR
ze+jOX<*=Ko6Zr?@BnF6jV?pxCTuj*5hW9z%A1I>Ey%QXz>GIO8xc$njF;T9@h~R($
zPFNpDc$%CG$NKHq-CXy(oR7nqV?w29cQ}h?SLI%)o`!I~8WXwzSFg(%HC?bw^GPz~
z=}<Ig7GrL1#r+%_nf4sN)I1|(A>N2nmF1|_SFCj8JP?dvz=aTgy>3&#FZJ$B%q=Wg
zQEP84`U(DA9%=q<#m>%#`hFs~33eF!N{cZ%Ij8&+T92@iNpkdlJDxs&r2M;r=alC?
zVMl2*UL%CDOxf&0^sgKHnjgq_v}b%21E=J6(H0e`=PieGeE#%d-231?8^i9tayM?Q
z+=!asUcyK=g5q4XNOTsM;O)koZSL(z#?*WX(cdG{#Lkx5p|yzc2^e~_m*=r?3!LCw
zFy_DL$IvjL<)+lGJZ}qw8uS6WRuX=JV?Lwp_|)enffwxbhsJV79kfTeH3r8lX7n!e
zO?7eh(g;`52;}6Qt?l^m{SV^F)2BhCgB!PQM}1~Nup{&##-o~KXU(zUq_L&x8MR|l
zv*cLtX^XyqBPYVEigB1SkB>RD|48{pZ7`d#QJyoenE#^t$NkvdS~rXujahqvVbmTP
zt>`(1HO*0^ci{xI0iz(an)Md&BlIL;@;q+B5v{V+ykhRsH{gm<91pu;wQE%GYMyCL
zlRQW~Q}oQ9#9AA$n84_T#}c@Y2QBZXp<Xv!-nj5tbYA0u!H#f)7^uLdq~pg9`#P#V
zkBE+r3wC&#IEqT)H6=XW+ggj~t53|g^1M5Y;TYplmZ}HvBF!CW7xa(w@i6w2AMZ}+
z3_J=i<FN_W+pK>?7Z_*Jyk>L9W8m<9;{R=o;d@QhQI7egWzENi$7xvfSq)R~4n-?=
zy>`XZ4@30vL55ImcbcLh{TNZX?cI%d^zb9kBj~9^zN=2cU4oBO)3cg44dKCLY>J*d
zTV1uW<h}Pkh{sQ#d)`fmrX3xh$AgE@gm3HC7tGGne?l@2*+_t)Yj%c^y}~;iNiZ_8
zxdUFz80Y3r@C}1%;X%S$8ACCfUywWYd;2PX%${2qr^=!`gqOsq_WJ9u#?70{%Byh`
zY?((GgV<=o8W`gi1_eTDo@p#>#1Ud(03<bLW@2e^(II4SEHBz9g2b`7un@Oy-mnp!
z^P=GW9Af~_#K3j))-Ca}l_0FBH==4@QU`d1(Tg=YwA=EYs^UG$<|v7Vu8PW>u9QVL
zp#hAcXg=*f;VjbJjEYFf`+#bEa$GdGn%;8tJaewwAxvLi-_dtWie|uvMN=hjQ9ee-
zr6tkp>FId%_-TCb;XToqrf4xVi!jf~8OWTX{=i1KsDB<m<_hx_cySMy!7Ed@Fy!$;
zFM<)1a?Hzst>>)LpzY8rJ{ec^5InuQuk!up-xwn>V5R;CkG5O4Z^Y|wyl%td5N8ca
zMqmTP!-o&z;iE@-6ZsBU4cQ39SP;1gUQ2oK94bZ^QLonx6WR`7X>00^a%?c1VrWHn
z1%^EHCJ{a7JH|BhHOQK@n|t)zI*mb}$ZY+9|MDJvB5~g`vNKaAn<_FaGIBN{!r8cs
zj)plwn_Q+#%Ga#Z&^_h}Itp;u>jTLs<YjCUujwD-Q5ccnQy2hwXC#Q0z$Q3wn*4?S
zhxR)S+6N9Io6#o2LVYguU#wB=5$$_K#9@=U=&G7ao4{k*R4#$%xq%3SBFAE|iGXL`
z3}ik4+)@umy%@%r!-O+Mr$u=MzNoXVKk+p;+F6&UJ~H17fBjB~WxBdI=$FeO<<Kbc
zJMX2|QSYc+41n+eWUtMw9bLPc-#x7nIO7!D;vGxv`YJq7xy*I?lb618PdQ$Er;@>M
zzWP6I4w8qvc>m=GznNat^PBlze)iA3^j}`TXG)Cy?dkLC`6@2B&b8s|I$fu)Nn!ys
zRjb6(jhituuLY!-5Dbmj_Xy3J^JZkqX#<gWhd`mw1wfef8xqcfi5+?vzhIU;gPA$;
ziioxui9UU+c&PV%)|Y?lIn<nd`81P$uwbDrWTOeR0|CXM@x&MqCJ0prBR$Xpesgy#
zo;-gVJFPwC5rbt5C#9<p3l)R*pa6r(Q!PS68lqr;Zi(;1KrIC5J03s4!wi(Vl@GxK
zrJ@<T$}D!IpkySjJ%T9<G<LZZJ6M1%3Q5<ms|b52_Nr$H9x!>;ueQ=IgqAB0E&aCQ
zkouK}ok1qh%8P-=#uos{Gn9L}@7*n0@F|nuwV_C3Y~eu95&B3Lvd%Ro`i>$1UXOyv
zKv$yg<R=?_puaTJ?LmmprJLNb5RXu8MXWAC-o>1$-3U^Q7iHP-rG*ON5`CX>kif)3
z2IYVo!FYIVvW=-kWQRHm0|aP2(8W$P6qVk+4}2wzvY@jdF1UIu@J>@*2uJJ`1%`N~
zQ3T2#pI4?bc|Ixlu{gEiQ*|<CC=6NbQa6Agjy}JE2k-Gd_~`{PEvI0AW@sM_nU~6Y
zxUagfAYP~Ibe+D6^pzRDzW?@bdVG1Yr3-ohJxZFz1CeD2#UMfFK6PV|STyo}@ov(=
z+*AQwD!iHCJ1i72K4HYdc!i;gvnDW5^|WC2XK@YkTp7GDfME>tA~JVMQl`t*;*v!-
z{Epq3gbH@ugUgFcwW(+ubRW70ZG`^5poJ_r;WGSyev1!@>u}b{*ywrOSedt1Mbcap
z3UJ2sRt8{fUNN83cjz>{45duL@8H)w^Uh>dL$&6QEd0d-37bnF;HxMm#zuHA!`8u%
z8H3!Uf!x62nouw4`Nh~22f*__i+lP<$U$HtuH^g4w?~2yYA7>N<WnzY6=zU+1CjVL
zu%nSI)HxFX<$^;L>W}sSH~271uvA_Q3`iII_J)9OT@u<*ASsmTXQzFiq2qI=$|Z-5
znyf4d4i)p+IOn_^$p|WMXpE3@7?6MieEi6=Ty<tXM#k_C;G8SYIT<J*f!8SE#V{P{
zJ5`=zNFj7_-qU}+F3&OE9q60jXgQ2aeLf~6->zXw{{$l84d;#Qr$I^WLwVJhS&H%U
zym)n8Wlsy9g31NE``L|Lc#&mYl#5r5y<CJ-C`HC6N~(9#c);c*I*peX-XSqm)^G7+
zl}ETaj7PvA%R){B$4=BRHXU)F-^w#IF1h6(I;|azJ3Hv{vN#SxnfYuZ)#bBM^~t6&
zJWlxJH+{9ytBhwtO>HEa0*Wa-)9NugiD6+{{iQz5t0TgMwtSWw#z)>A2Ij+~8aq7x
zCM2=}&m)PoY#z!T)m9d19@J~IQLE3WT^Sw;r4lmA_Vz|>tUrsrolP4c&d)AmY57)^
zr)xH9KYjWzc6TuzjXN|Q;pE1rP(}&2PJpB4h{k$)x)F;Dw_<i~*+xF*SzMgj6NGS&
zDBwpl1~yRYvPawGsG}GROK!iwusd>L?~du(LQG6yJfx3l9GnzPDDyzF^q$H@=H{Fw
zyh~yzad2c54;T!Em+IGM+`s>^`oC{48Q{IUvlIJj%iJ8h^YI{(9M|v0gZrOYNk5_X
zptxn7F{*lp^$vzKjAwZBVR&Sh{Nmzb+`fG~N`iwrA{vI`JiRKRbK~Qju`+EsiE#vc
z!y}N;h$y#Fh7}68q%+{CM&Am3@HlCj5lFav&}-U|VL4iDLP2k3p8k#wk4>LZ(jzBB
zkML?>a}~x#l-wAeY-}<flZyGr8cVcQ;|M*ud1E>5zVeF3f6)qh@C4eYDphVOGHE9Y
zWaj@6UTm75(42$)j+%QeSXWdphFmrSCf&yXz@{*0YZ3`1%{9)L+TGbUzhM&ro6=N7
zDper_O2CbcahA1(2YBcb3XpKQ7(i$vbJ3xzGzLJhpj|`o6Yx;IP)UNf)>nczXN+MG
zfj<#K5-(PaUu=jN85JLZR*YaYB=IhZNsAW@5jMC^y$-_5l2Dgv0K$mMTJK0};nLu2
zI>ta{@UF&tNH~6a(1}Nn?#GkI4>d=JV$p^u&8>;D*AwvCTwjEvsLX2K(Z458o*J()
z*s-?2Yv<sI$RF_I18-t!Huvn^g~#yh{6b7uYvT3oxcA{b&E==&#TY;E=%XCQb5!%?
zptt9Fu)E8A%`+{VS~ov`{@mW9hoU>cuBCgc&!4DI5p%N(QLZ#p7a>ze1G|9vj>eet
z8-MbXe~e%L@?Fh=o)y0sckoz7uh8osXwK~kcMoG`24m0-w+YWJ*5A|9Q=%ucF;lOp
zOh)0th6(s7o)Q=lIM3?Nom=8*bDj?vRPbhEj<WGH*YV-^sDnAlm|3x|5#@ctRqi#L
zv9W<?Y}fJ)n<fc?NL_f0;CV_p$QPE2gs&uwNysPFifE4T6<Gotn`zWdFX6cu9H9%~
zKOT&DP{GfTQJ@bPF821?v9*b3E#8w&k0VAFy!q<NPumwnbMOo!L@|7ud3ka|NLZDj
zdN{ijnGE@gbA>R9mDOJiGOS(k1nYD#<aOfVgU7<9=l)JOL*8RPk>=*bC-e^Q;n};p
z+m6Rio{2Yzj%l3=Tr)JH=BV)6YibypAqqvxPpB9F0XOD7{0L)5vDT&!-XN`K&_eK;
z{LEL&zLccTRQRj6(7VBSI-`<hV7<o$jYUReqjI%cj#uv95}jU9JCVagOT?>BC4cN_
z{xDAE3UUmQ14`~IIAfiI>R^)$#wutW;{;A2zfn2zI~^!=H}r|n+sJin{z85(bO5wB
zjZK3#D5dNRr>Ae3hd(GW&*(XDrC#JmJvQEfmuw7XQv-9v8`lOhx%D@K4|*@~8}BUI
zPaCL>GJ!Ssq4V^UHo8pp0h~>lJ>_;TXel01v>)8H?5ez+1q%P6y*VDV8y!^g6IZB2
zyQ!bLO#k$NvgkAOi!hyRrlCyQg#2x}S7iYr<`af+8+FxRyu&lZBi?yAE{s1q9NJAg
z7=PdofAv|Dz!SWob`2XkKKgytNIeJdyjjq<ax6S1y1;A7D2!^<4UB2i)u-+$OXBaK
z&(w)*%)G+rwXwM!ZPCC(&A}7N`Lu-@ql~_a`uLf@`Oqhg@;x%=XFlUAedDuu_;Sj7
zFUtMR>*e&@o_%SFa(~B#0q|dQLbkXzd|jvO^fgKFtWv!a%eS;}T^NL7fI4#xm=+7E
z6A2X#%cooB6=Dn==sBta@K7QoB0D>A;V)dC3sMGO!QQ1cqW<Fx#Gn6Qp)k*UZX)Q%
z`HW=7$$kF0Jp+LY20c(Lz;L{yPmK0NjBtH>D;_+0Xm2mPsMwWi!v%{UUeNmm)5ge0
z_&e9;Mj+@a4{xasm@dqhci25u(c%uqM|sJBV19UP9lNYFWaPUDrG!YNUW7&B8^ieN
z8v-#NMQXYkD1Bqm2jjH2nHJhM<Qa};gJ$%}khO3oZoHT;0veZ^;I>$R(|yBG$%2l)
z0<YZRnB^a-ZpILy%nMWHqn%v5ix`$_H@xd4G$^mwz49RhQl1v#2ohARWP!=dosdVo
zeG!MiPPl;*G8KvmYXo4za1f4xkV}TM7DlIscxWLU=%Ve6>3ELoFvv4Ng&e~$?jZzm
z`5CPS_q6DxE_U|lUkY~0k6{gCoHMJ4Q{nWel^2O0>Zk=m#uH(T1upPsp=Bkngk~0*
z?0jN}u>~-dMvQOr;SEXsz<~3^9LFD+>6t@1=@;P^@sLERNSQ1?T*!60PS@#=oxU=|
z*AL$Q{-@+*XbQwjl*e@1gv@F1sozN<GB321#TxXO1uw>XJaAB$pd7(4Wn&daEG_s@
zq7e23g|)r0+HKz>56UM(-q>ixH85PUi1uz@)95r>6?H<BO+Q6z&2x0wh{eU9+E`py
zsko=FFZXlMYclwvr|>Z7ZW_Jdq32Pao`}WyMl3JR#Lbn3s8*-NzmhL{bBA~&uz){U
zK_K3Raf%HZC<58!;N8g7rGIQTp}!~(xPg)d)(-z>!E8l^ZlE+HMAYRa8%p5af(887
z@<#TFMLRIz93SA4VG^^C-r%55;D>?_BLwZh_&{H2KME4$6nlk>_p@WwXR3(*@IG~<
zfh#wc&^J7R@RBY<r@^=B`>5JQ{TM|K_P1kye@|n`SxYFrPgSpAFia>OHeMimjMykJ
zS+1%+3_UF4kS$<5^iBL}6s2@}iNp|_KTd*hbq;H0FEZYty#-7;^9;p$8l5iAI1g$|
zyb;5ma*qyp@rCfTvlac`uKLXxVrn1kTICQb59K!S(5OfjU>s<7hF?NH!bpR0E6)MT
zcHq@>>a%l@okmrc+HIp$kxpz(xwxb*l{a43J^C@K?gB%}GVU+s(PxblUO1fZ!vNYq
zCBKk6p{(MEF3z<%Pw$lsi^hG8iSG4!P01B_e$kgR!z|AfD+%U=qoUnuJU|Wr-gsWr
z8Yqq>Z%?9RRhg3K?9DR9_@K}yFXM+-&5qAn!@D0NL4|O96REt$3#7l_icU*-z&Tqg
zf7~I+j)fb2%U6w=`M`T|aw^Z*U?)3@%=NYB@#vEeqbpirL$T)3+&pJ2)oftc-9ce}
z<ag_JwYOZi=g!b@S+X#mK)R-DQLZgTsj{NJ>zQDG*x3=?QJb5as^?g1f@;dP%qPQu
zjZesgqGi37@E1=?&c<R>SVeg>-ou2G68>S_u)$1npkyeN(a6P|w=_PMMzkC&1Zm-q
zaA-(#1P_*z<9*E`&cQnu{B}Km7Uq{UKY$Nj_r17x@BP@>-H62nycRhdOmeH>CKBQK
z3cW%8WiFvq#<1sbkCNepV1iH2IB<B>Ri7^1*U}{CZXJ3zKE_Hm433RY1z|41iLqg|
zN$}_FCBil$-=dg&VJsFMILDK4-}C}`zN#^)2;L6OIjp%M+KnO}oGMROY}n)sDF__A
zg0OcZcr*HJrebasUX9ct`ItGPzM+_&nW_6i$qPP@s19Mc=?QpusCJ$ZZcudb=%5{a
z&2_%xfdmhNZW0n4??h-6?M9JLSXYO3KRy)PRi?&{IfQWr8pu57oGj`%7F~fJ;I%<m
zQM|60r+8Qn2?zKMO@{tqFdi>08V^yFT7gNTJ>9mhJ=MdRjetb+T5xg14$*LoEERiM
zRn-4{&l~@Sn0L%q4Ss0OUu2j+3|B`79n&j3rKT!#4qG@?nXxz9r06unDCeKb*LnaW
zw+W$OekvR}I_k%f@PzQOcv5m^Te&>pxk#wf(NW=^@bhpV9^Y4)qr%B0%{$SE(FxNQ
zyqlJnZg}1uXpFnPJmb=54+=J0TN}cs{g_`^iG}4mnj012`F=cm@<eN~gP5IZ*kBBw
z#E8}JH^p1nG;$QXyW6p`zHa&f?wm;0;2!6t!G{?ucpq{@P3wG&qf=T7Llf~z><i}o
zURz^xrg7hk4?p-w`SCWrp*~<p#pq6WT|rjz2VNZT$GN!~8(dh+Vw?saXNf?wG;d>7
z8UjW%W|LaKUI^Zkv9huhckZmj%8f<iE^-X~ACK(qZS_lh9XW<M4t(Iv3^Q`j*4B>3
zpsTsrjb<~w*1=WQcn#tI(&9qkfwnj>Xr_XD*mH{V2_?(g9d98u-@K4>6t}ks2YPD0
zOy4*v0G<GL;i-j(m_tK~Z+1HU*b!~r+|+y9qLVuqywamH746`6dK)zw4e`MA<~=%O
zt{ues#&$e@{M=`5VKD1<dYVJ(>uDMx@Z4g9EoaGbCfDPK&th$DTkvhi`udLWPk19-
zgO++@llle?nV(+}uCdlVQ~OWilTRMS>e{Ap51NbT7GstQe!)~g&FKR@Ddap)HOn(3
z=hejq42(C73l|%va%1X1C!=?hKkA!4d5>h->*_aU>QXlDr_oI1ie}&q-RJB+@llLl
zcv>QUj9C`Cj9raUtGS~&FW#Xsg5M!KBKwh`3&<aM8QN2CG{=EZsn7-tD@2*`=B+_~
zzN1g*_WG8!k^g8XU8gVBNeQ0FU+P4*S$*}UX}wQI%cS~D<DMD}3?vV8pYqTV05@>?
zP;;-?c!tpk+%Dh@f5M0a{mrp~zaocbxIp+mZ6bH!sf8S515#0s@EUy$?{UEk2&>7s
z^Na`%(2i8*@ou)ALW_1#x5rC;&W&Ls=5NS~z=jPiglUDI@Cte{VC1$4nNs(raZ`T!
z0DNh;H<uZ1TAS!IbxPR>2AcNLcIFo=4E_?I25{$6`Q9wAce1^SZ+v>gg~r>m@YVen
z{qe<f#)urI3&5Fn0}I{N7_gSX5RJiWb89==z5O_{0iAp=+T%9rXO_*UtK)8dd8rM*
zk%||7BMUcv%X|NdlUr_BeDU>p>9;-mGSju;>pEShuTg>-RqBnn@ycy`e2I$KKqC6>
zg`Ll15(9$Kz(no#4@1>^Fg6Tc7!=s`Sk;2CUTfH+1;${4Wd0)J1gYmcMBkr-?*p23
zDW(GX<Nth&!brbiNC*X%gY=a4={<JS!oVrRJ21pxF?wxESig7wUOagAIQkNXj?W}y
zOE98+7B;dJF-+B>FEGp_F?AR$wOg5^M<`=Z7IGdF3txl^7It)r9Wm64al|o|#qbc~
zx>|^`FyNVcA|_5*hLwaY1SmXx2(?7LD8ty<h0x7ADCjVr0Czm-*j0{^1VNESE(>T9
zV`k4GwF$x83Q=7ML$uF|GYJ+59~g4dD3?kz1PfqFofy2vR1b_kLs6X?zLW|3$DoE`
zjCNrJLePWBb8!8c1SQ6uw(^HSf^g)sLUhk(eo3I>)5D~NKD+f19$0vm)Njt+pbmD{
zvO5_Z!voO*8s+LjAqhUQGZ;Y<oI~{Vj$`!$I1w@`1&|a3fjz&$3w9(Us89xqMg$;s
zmm-Yi?sS&-)R*B9=?lt|tMCQ7P)H)!0w2l+R}hjJ2gZ>dOu&qAlHe9GoM|6&4!FzN
zY>W%vL1T935-tm&5LmKM8pwOs={jAfKVJIE3|~L|-uE2p5W4QfCqw{729Z$mVr;N@
z#MlJ=hEG8cSY%?D@qyWGJY(8d#w=fLytdNlWlt^T#~_vq<0JDC%7ccP4vTK(;!@Y3
zPA7dN1s^2czxqyl2dTKnr|&|e`7Ihvf);Y$zXoNv4!#fSG7Y5rx?HF3LpKouP*{BP
ztykjz{D1mC#^3zapNVFQ#$VuNfpU<Y##2@ftMyjyWj~M+IIKL+Y{Cct|Mj99o~o3v
zNS;hGUr+;BNE_D2gj`Fn9CoDhF1t|);Z*QbZzdUH=Q5rp!&Y3fa3=%~ycPol2^esp
z9AJY)@_q3x_0fuR^PE%58;k{QhKIB8=R5FlxGa=1;#ZvYLx?fzrOh^a_VcV!F;sPi
z%@~9flbnMQ>5S0SYNy^I#20|RIH!MS;+=SrsSUu}3igXQJ?^WIl10>4<hVTR3WYek
z$h-OuEKp1c2n-q_+m5>qZ;2uh`3yKP@cqt4?CosCv1Ba_ny06S>ceFmAEFdWuNiwq
zogeEyaKQ^M&)lN#@bw9H`KsM)5;;C1OlIG_5`!d$&-BC?j-&nE*d@H%>A7T(OFdKH
z1jj1_i28ydi@tcHj=h3dx=JoZRupcz(HhfyhXe!8%|KvhdVX;!@8y9$C7amrMSkvs
z7lZ+#B<52{R-20HT0N%9=^aP-FRI~89ONTqH@-}is>;()QG~No+lDp1Ba$5u7OBIA
z3gE3)ooi0-x8gu>1b4s%LRJnB<28VSWi{@9{9!!2|A}Zq)AAN`zun?|w&UO&sD-5)
zmU(h_wD3-Kpd?y*{v;kh_(ZTi6ux6rK9Bm$oZeY*2q?=s$~QVZ9t#W0l5-^6gImMw
z*e2woF3%T?D``MH?rp`!+T(1icV{EJ5q?zix9HG_<`$s6R9p6ZhS;ibH_w-`7oiQA
zz!|`TzsJH;&3EB3h88@IY~0ouT4ut?Jgl+ym@?+bh?hZ#&+_D0+`4t!q4elK-VgKh
zi;}r(Hcp^iMhQn4v3f0)`zVBwccD+-m?RimL96?PJa}|`>Q8%<su5>CM-LBJLNh|s
z8NHOTky+X7R;@PFj`^4_S8Z6p%Yd_anSam*%ab686ZpiYghQ1v85Ke*PS5yUtjRIa
z65$EnY|L9cS}?|yFhZ(7$kO}!J+DhRw=4~{;1o(i!fWSwfeZq3Ee*}c%+N*X<-R?C
zJYERyf&+YJdux3_d-mhQ58jO*{pj!Qg@RFobqX-0;5=ilNd=be!FPDr&JN-Cj@5U`
z%A<JlU?84U-Sic2s@&XRd0OiNjHm&QnN}5Jt2VOXPwm9mJ!Fr~3{8nwCthI~S}}U|
zF{WzFfb-~hMfpp{D~u?d<vTo1$WF$bv;1fS8}gXfzzc7{ZPB?s;V}41J2BQ}Gzw_z
z5c3vVfC2yT!1R)Eo$xn@g%k}MRp0Q6>x%X@n_F@J!9DYCLT_T+W3xxUM|gboN$^6k
zp1d->0DZ2McpF}w)mp`2^={t06Z4v%z>e|7i>Ohb^LR5hlsl)nTAIe#&nk{_KlA%|
z!9aI0{1Eo->66F7`J3=jHqE^M!G|^uF|StFHsYP1zoUGQJRdL^z5o6PvDa(`X9`v-
zb<f%3L%fJSiBCTH#OHc#tgpqB$Is&V+PdW!uO)S{UI)w4(=j7^&}R5MVdq$5?QG*c
zdm@~w2)F9yjquSujRPSUS%ad8Z`7+=qltzImuU|Mq*k*nyibo~yh2#FlAk$DqEEn{
z@P)`)@OyihiihL*gTWe!2Aa-2)^xjjyBasVoz8=h&=|7d?-)a%L(G5bp<X;==@;XN
zA&ta37MkL7>D1S)t?l^uqx<pX$uozJ2G$vt4KEs${dmHryo1q=c||*c2_9P8JA3i;
z*=js}y6SWHcpsc*txEaeI&%`-+u7cYM~|MyqlZrfyM4`PHg)F48Y*MV)nVg2>v}wh
znt}}>*)Z@uefm7sHwe`VtOQTBjpVR!$X>dDwO8BlTVa}Y9=hYAeNTUy59DP7K^kk4
zhNjU=Yh>k3`AF~cjxN4ip44}W<U4YJH^7QWK___x?WH`+x8e=(kzOCU7|+_&hp>TT
zbA3JT-TN?hcDGfIW|wdfV-n>XPlS8ud+^GuRV$*ak}Fi^By&u-#GFr0BGE&=LmQxL
z^ox2N=3BT04iU-{yz+)((HZDEyoY&bITHB`C<#dJugXFuLnmYK<$o$0BMk<z94qPu
z4+^*yvOVRXTY^?($UyS4VGY?DeHI%GL@V_j{TF4>UmJk*n{tpjG1_`O)hAGdF(AYy
z<+`8355fo*<MG1i#bAR4fQAvzPx_1JBRVqcne^Mb8nw|zIFE6TH*H{^QeJ^q%p2gv
zZz|1p1OLPq+K{-Aax>51zx3T5P`P~Yf*|30;EVBI*WknNl8`^qnK5pOUoY^H@0ly}
z+v`?iL!uoZIrrG~*3q2Y*w_~RT-TiKX}p9>u7`5eR@%j@`ZSIYU;wSqKY!25#sALq
zBEK#}|LdDR)z`sQq|bZy+3OEUpM95Edu{l-PS@${l0<kbGxfOn>MJofKQE4laGDE0
zm>)usvuUx60GkFQF*b|@VoDYQ#4MoZB?*iezw)ryQ4#rMApDu}s3<O&kC~jxg=NAA
zk|FVr@(YHb_js=Q{OM=PqYL4Qi^NV$Ne@2L>)j9Eiw_@s9BnOpMzv_+9h#hrR0%z5
zGt87CVf@}rBf%a;6AMuKgdj>qEL@9w#SR~TsuK`H%0`IFa8GA3Dn`cc8pn&)d+Y{d
z=O;!LcH*#u#zq_}rwsa|#^sL3RI;T97YcxcS$bQs=u?Xf4=rKUZv-HAQPBs4KYyy7
zv<dhCUxWpALQzhJ3{2&h+YbB({qW8j<pnO@DThHzWgt`|Y?6mEG$JV!!1%e8m%AwH
z7>kQD;f%(bXSst&Z9D;P*?xo>!UD2uZ&C{j3!SR>h~4@sgR)snsVcXDauMb*l+g}^
zOba$D-$EMw5}bh*VIOnHe4cYg8Q>uOMHxpP@`7^+Ve|<@7557qcc>f*X$WX2Mc7H0
z=fMcxXD0)}TT4m(PEP{~JIX_N92O2Jp9LL{1A=UMs;ue69AI1}RM9sz3?T^NjZiC`
zD>gK$`r)ihBMFh{G_TWjx=vqu`pOJnfBLO&{VfYr=rZ(z3mVVP%_3w8i%Bk&Pd*Q;
z-;YiY#Y5W)#Eur!>|V~ZvR;H6%$<$72!`&$uc5KY+spw>n@qGnn|%KZlAiHY<SG+{
z>vNLt7SHq!d?CLk9VKzce8c%vDvM)%h3hBp$z0G*H9nRW8}YyY-~IRTzxwb0I_Bo;
z;titD=WN*E97gfsQIzOKSQO5uO2r1783@6X{0^@i_!*lQ@N7XLABHy@46I}r@H2QD
zh8q+k7{B1#D39PtcudhelmYgzQr#FDFwo?A4;jMHJAjoh-^<W*C>iilgYV+8L_VL3
zp=VYYn0K+G`atbEHZMXk;`YFwG>(LpBFrJ4B`9N1&W?@{g8D@G&{AJc1J55et+06p
zgB?pV${_rg+k&3zd_S7IgmG$$&s|t1F+Unn8%{d0v%M;O>ZxrQkHmt8=<j~Cn;Z5X
zKrst{#`|k@1cTVV@N&mH`_E1#*I*11+(!a0uzhd7LFspNN=QDBuKSGF2Yth%2@h^I
zqfAWVfyIt`HB@{U#Q|X-?NQcm#-7^E`AhT#n4F+Qy*yP`^=XH*XWDuOyi%T1@mAtr
zF;cwQF4>@@ufr&_QJ|Ah1ZP+}E1R63B$*G!OMd<3xn9VD<V7j2va#3@@{s%H>*|XY
zJL54iHtF%2uHpqD97I8{><0&(*xGm+t-USd9ZGll&RK0Xq*>9fb_~;AHWyJo#tASG
z`3wW8WDUXK>9a?Ib6>K_wB?qQ<Ad1IIKBV=`|;6-AIIwIy1i`hctMfJSzZfEH=|mg
z6V6>~ytiU~^?9s5e;yxxcu&`_V{KzYG)Hu&e-N!sPji8@UN~Q)BKau2Bv4x4x_Kwc
zRXn;RM;Xq*6YQtGz)GFHR2e(ZJ=gGV!WhHc1sCv=N6E{EpAn1>gn{FI;m@St$Gov(
z>YV5iY_h{>Bg{JP$M)8<*jRn6`PPyAd>)+^#<ge0MdRVck#G~^_OaSV^aat=e#f3E
zOG`JRTAPvlm4;wErY0npaqiN%KJltx+`vzdulm6{$6hI-(Rf?}N9Gmu2W35SEVKuM
zgrC_XL-M{u&#^0>bIoRgGta7=Gb5PMw+xMif{jq6qp5&pla%NH2B|TO)1!DDVN}Fu
zlpc@{Yia`!^RX}b(6f;fc@yO<I0fZC5glvpZ3iLmFz^6d&ig_+s2^&F#z})AFevdK
z1{%%@>!?i5Yt!0CeLhw{SaR@5X6OyS31bMoF&?dL#;rTA#_a5(&!fw8zedfEAU^e(
z`IuY0sb|YERh|((iryVzn4!JGdG`m}&pP42v>v)jNLIpyayDSj4HEF_Y2L5KuC5co
z0AtnIxccQWW9%q%+<n~dwlo&d^pfW0l*+0oPsuQ4?p}zdrRQ0C5}uxRG;Tf35$Ltx
zLz$ul&=#N1g@JdN&27xXf%hlXs_or9XytR^cH2BTm6?PrB&1?j<v^eCfa}GB2cHOc
zCgP6ZhVhdzz@X9--}m}J^pADok;iw}^A7r`5>yOpUA*(AtW39;i(~>41`~J<^B!X}
zXI~NO@ly1fu#b4{(GE4s>y>SLWPKbTe=K@(EV?dI^YG!*`2YUhKg0(gJ`|nVj88s!
z9G~2O8r!?w*lQicJ>~uIqeqfS4h6G`*x1-n-jCv=k3WgWk44{F7##XG7Wr%}^&dm#
zl`&ZDWc_60j&Pu$>(I#wt<#|q&>HA3-a5dd+v(b%#3vru_FS1L36@8(v)d9~*oo~;
zj8lXu)cg@|!DDOr#<Dl5vId1md$X=C*0{j8R;x(HXxQ5fqfx!y5KW#E%ry_hzv_+j
ze0=uod91Iq!2&)2e;IT5!KLM8^JVarGqT`A$Ye#I;pc2<B~0b~g66^8f@QM@4<5z+
zPab;AP3v^Ajzu4U%z+mf^MeqIggC@F&wN~6-H0cTpJ>i)sITBQ@(jjN`cjJp^?7b~
zR&#OQ?_i{S{N$<V{6_S8TK}tj>k|ZH;13^4g917pcu;!barPIXQh8^0r>T4B0l;Uy
zC|W|B3&V)M`qR%;aQ>f(u|zM#7ryfzxt*17F3WH8%OAAKe}i}Ubkc|AE#wEwQmo_Q
zy_yXJnVO9SwOUnWik@l?5`Gqg^48XF?Chk0ac{5byujg1_?YK@PN<xe$(Xm^ggl__
z)XTWSmyq3&!PyKztND-ilMlXzo`>Jy#jRVnY=8u}xknka-$okM!EeUkMfgF{c+G``
zD>8Dz!0;Nplg7PM&jH4bi?Yc_*_46jBaQZbz75J@{Q*ugXL9Y31|(pIu7~F?b5LFy
zm9ku9Q{6V-#-ox=FvXD&%ItdbKj4vVLe58rlbgcR@Q3aRz5uSc4vimuG9XoeVJQ3~
zQ77|!MstUaD!`vMF)t~f=fE)AjoguA!F*sYq5LAh8>XMiroLi$W$l&gFuoiAl{dSR
zI?Ob-y)eiDd&;3(#%FzDo>Py<6x`7>j9VDJTB1|Tz3pw$tS+It1yjov9+APN|5Yq7
zq7MJ&^4I*G>E-XgXHp^l<MVT_-#2~UyZ@Z)UrYMDwtdm{*#t&i8@{g7b^4kl82_{c
zy&HFK#f*ffGYJnEVhC#p0Wgu5@NGeLkRKxv%*!4*FkY1Rdd5Nl22vb2E7oD=AX1`-
z$&~q{ARE6RFmN7SzOQ>1x`!YLW5}dD&Nj#kztW#~zKD6EynwOpYoYUOZ8bi)|55C9
z_PiO-yM)9j5r}Dr_?(#-?SNq+OkiN487ykq2}=f)L%4Ud=r!}xJsW#eFS{%eKnOJg
zv$l~?OcYp6P!|h!B^V?-@Tr5HB$Hg!A;ArXZDB)PnFR|<JcM-G!2$6wZps4=36}%|
z3uP-h&GgfQ3PFar>Gpmb%7qeRf+ES!R6p<9>xjNE9+ay-q1dAgcBuPP<+^O71p!O1
zA;;LLM18cw3|sfv0q>o+`s3^{Vn64*#BN3aWoHtS2mJ&$B(z94<D!2EFu)F>%(0KD
z%b-^2`azorqXO(KpsG)Yg13boJ-4AWTjE89z1sAdyT=KY1h$ffy!Zmw5b}*%%ERjj
z$&8&3O;t{ZhY|iJ?5U5?E4dqwe%TPJwgXr2{6hU>S2{32c%v_l!>-@ttyL=OD?$^C
z8bZFHoED5!psKKk*~HX%OqM31T%FQYis`x*{?$p}LqUKrhNAL9V~bwyI$fvh^iNJ-
znc-_`e&KJ~Va%c%kF3Ib>&lBu&wQWVqWbjNTH+UY_Oi%kF^!^zg<7#V^~MGfu7X!y
zC5VYG=oWPBv+1RrgZD|i!wtUYotN`{ww&S_?_WKAxedzUXGJ?+x=8kBQ6BUCyrVxY
zCqxWU%+1eL;?IBht+;t(UJK?tOXBdTFIctp8XIjUm7hf^ijOopLG!&KN3<FbErbL3
zJPS_b4|`*a7f>(W2GH%|P<7}#JRc?fDIxeo*U5{b9@>s<!|rBy5OwCcN@{`LMLESf
z7y~6nOEycOaKbnRe;l8btfcm!1c&#~ZoIe5^E{Tc=O`7IDhHV%`vvbtK1d}22CaM#
z<@cD{if4M?3KNuEcmv?+Fg8&(zhTo1ZK3U)3)Sy7g-?XwIT2r)kPI-UzG7GqZ$9A+
zH^MgIZKM2x<>?`wH7NYqIe%*YgjW%;8yitW2ZkB3PmOhYgt)(|7X<|{Z%g)ScXy*v
zuf-H$3dIk(P+0mv{?e4&NEjw}9`q1SB(8|>AMe|+(d%|RJ{T{+lzy-45On8)8?Y%?
zs*+ddy)&Qjqzo&qt=z`YfNXZ;F+&l}J@A9Dh4cZOvAl79AQ|Tvfktfsm(sv>eu|==
z@O3DR3AaZG!l5`h>{@|u!-CHin}`v@p9yEJz!IM0rNDK5abn}l#@ZvF&x0pRy^$f-
z2w#@+7>Y-LAQ?#QB0L}Gk=P4|Gm%us$XHb{9*Tz#K8S96*Pa02OTX8Nwbj-5<l~Ry
z>62A`Uyq&b4cFZeY;NCqBNi4{gd=#TXbx%Ka8Nypte^b!C-JkN{mi@j*`?fU_hPr%
zijB=p(Iw7}I`Yotdmnz}I&R*&6Y~p8DhJtA;1WIt2B)(_wH-y;Q5@|P?yRZt9t%8x
z2;XJ}5na^yAeT**t9lpjAmQr7WRxdS3?|NFJjHm8;dp{J538So!_h%Awl<%;>~3o(
ze)^Mti1*%o$Hp4Y3L788%TjnNS&Vr_s5M|v6OL3WGa3^VXBmDBg9mE~LJ!%%!G+A}
zv0%M0VUInOi6|#a<w|gt6W)8QbHFz|GN9AYi&_;gI1ISC2I7o0-m5nj^1jMM;n;3B
z1=E&jI=kyr1_S|t4|5JrN(duiGKYo<w+Aj_%%RkzZD*(Z8vmBylxIny3?`%|G!>5{
z&sEANqJ-dymoTB$Cns2=VdO#4oAbkSVPNz<JVeQzVH2T6j2-8;L4SZ7G^|{jwznKg
zUDiLqaC!MwRBN+pbKQn2<}eCEJU203l`FF`GQJq&Qw_CM>w~kt+R7Z1EG_zO*sH8w
zzbTw+xh+I*U`{y1tJ>neYknRdv|@X6)dnHFtWa_viN0ba751n*lHrFLQ}Gk6Yrr8k
zuZ^fI)_@t3Rd{j&e{HMvq8Fm+c$;BNYj<`9QzAKl1FQo|hMNu09s}SuBD%B{yIZSZ
zP>iS_7$AGSUBL>)x%xaXlCog};~VodX$4U`y7oR?(wxL-TJW*9#&n7`0JKbVfsky^
zpFOcbjF6DDYfJRw@#9C)Zne!f9z1*)@BH!?asR=?*xuREv&Zql`|o@Ha^B;i<~Z~L
zW7rz&;t$@Bwe^kIRNhBVp2hp`eH4!$u1e;>W2F}#e)w^`|Ng_+*#$oE8Lg$p@XD)d
z%tqtji1i)g3vP~iedqPR<`cXVua4=8=sS3(#53r3@&z_zh(=;;#)D2UVJ$1P6+Eiq
zWi`t=&1T2t5Y`Y3<sR!U&ih$dm^a_!tSUmVwnVp}W2`lyyYRP4r4+LbLa`1x3?!iz
zS+k?F05?&nV~AQ@oL8L<;qJ8Ynlt|xhh4m|^d3Bz%`n?L+nN)@s(V50oY6aTu`o}_
z*Rt`LxdE+2?qS}rhGwlgH#aAGc46<vIlY5%Y-e{jKDzgbVBPoo@QYNkr}wk<4P%6*
zo3t?=Pa6__<*c>c-KOeGuNd?(#!101%V1+kUMS#qlox{!9$v^4>E(^#Dd%U(WO-2@
z@`#$N?|9#!Q%GYa#(ZQK`U<aCF275g;a$+5{GMgHME%Fbed^-Mw!w#yyZF?7!IeDj
z1n+s8@v?upU0jsvKi@}IR4ZtU+QgcgF`b>A(YVZO3=%GQ)gp7yFXTDGMS_38*&8v`
zCw_xB7)2<Tabpb7!=QU8@-vRe;>hZ3gm6D~74kTJV9crx|Jm>uq8CEeM30j2rk^xT
z^d1<f-OLFayWOVb1D;Fz4h<{%#`+%}1bQUu(?i$k{Kgw(zSKDtx)A}#?PLBxyRBnE
z7o_jtD9@}v(d99seJYna(Ty-h)W!RZ0d(_9uXDot;1ccBBZh1+HhgiU0>hHcS1M0!
zHGT{-()QGWQ9d+)%@%B)3HKN8YhamSPT}imAak9xRagE_qj9k&;x}!wjz|3<@eXxB
z2YOxBR{NIM2`h@u%<n3XyQsDS(;Q#$H^0A}(txWzimoBMc>SO@aAx7&UQ6_&=}+oY
zYt;@s?Gm<5`k<|e6IsvT_SN@)UQ!{yrJw?Lf7kV8r_Zb7^R9o*35>cnd|jvO^fgHk
zoT+Lh7H`~$X$gXO$G|M{aAENTKZ6;<yfWk;42|82ge!oFiBYN?7)Vb7G{yi9<z?3s
z%)?9-e&Hl0qyKa9!#5={27fbCp*;u(5O6U}6LLMJ4D!R^bYY;eQc!iOf;d)NH?}t8
z;q%AQ?Chysx-I6)c~UGwQOr{Y=okP1|MW>jK~z}bru_5+L4yUVLwTtG2yp0i5D>|W
zpo$QP0GPXklt%&$;ed)zg$ELb5F`**V1VR7$VSLu=RS*Ulw2r}$_OwLQV3sB2qQp(
zcQRyS3N{$QylX|z5Tx1NN1I^k7-{e<1MZAj?&u_}o)$T(lX8j$2Fg7cEO0WD1;BbI
zcbVpxA{-QkGvEY#atBQcC-jpx@^^M}ZUKZceMu;0?A>pbg`fzm?8zXZFXMJIbRI#E
z8)1nC%`hGTR~y)r&vk1&EXYU@Wzk4GQ*cyU7*P5w++){zA@sOx{b!h}9|&D2C|J07
zVTDkm3NZe84D~KTCBoR*B}O-lWmy+WkSUG3af`O;9^M!jfXNS>5MqEm@TM%_#W;g6
zX5MNJiygjGA3`a@mpd+)?H?UjfMjPMc~Go^r^IA0YtdMnse1R`^2$=oFVDs7yo5k4
z3Mb1Fj>l9wyTbM1sY038={jAfKO}u+hA+|lzr~n^Vx%y9Vc7D*%ZslU#x3}W7p5$5
zp*<|fC`5Eogu>-Pcm2D#z99L@ryfHC`N;S9PKjA+rcY)6qBMBd<?wR;;J=p=bZzkU
zQ)%$dAVDdIF5&N59O{|Rr5d@6#raCyytNdK*_mLY1M-KBOIonP*HAnc8#CZ*gd&EQ
zT7h6bOvqb|N+*V)l?(9y2|NMVxjNX0VjiF2R|r`IFJfG17X~+!-kI>$OXFiIr^d`5
zyvSDjfJm=L=*gD)ks*9755aTAD_pMH&LW%oP|mS<hW3+>%_@YMvl2)Ao<%!GA(Y0G
z6Vt)jRw!_Qn2jmwD~kBJg&R?>&a0jg;gEQg{y19-g)L{}ATJ&46ZVl^^5QGCxu{fU
zJ#;8J;k9GxX9>@j$+Gz7r0@?<L=GEA@n0P%`zNj7hp(g1A#{zy)Tn<LY>y9*^_#F=
z>i0Mbw)E@~qY@vx8I|dV#!90v-p7U#&N;&SVsxyeo05GbYqIeK!_nwiRc$W?=L->1
z4a3m9%BVG#Vq&VMI#Z?r2MEVkn_G$T$+;MttfisxRPxr@f!>2hO2!dvQQW1s3?6QH
zM~H`u=VDMA9ub@bkJF=`#uslBl;aq<$UmyOP}Gaht8GKW6}^k~MmTwPs4L-N!<6sq
zQs+7I3*$_ywHf<8JV1`Ed`2b$u5<ItZaYR)@c39`I&^{eNh(JQLkylXHl(<GycS38
zE!WyvjYkhYj_1!-V{?5oo;`VH1vp{JdL7A}!eKnZFgkwsyFZL?zy0SivvAV}Qk3zF
z%eUg@-M3<C<wl$xpE>*-%6LNZY;5;pclR)MwmbIbc>H)Z-h2P!_{A^Yi)YW*Vr6+L
z?%sVx@DSc2YpMTgpx~74##0Mf3t4DF^AMSdkd}-EUWFhZI5jec0Va(I9)L?iaWi%p
ze^9ty7@s&x438WPk&~0;mAtOG-Q3-aAN|9Bj<x4cV|sc@c&)w(_i4wTo?$TEKj>NJ
zO$F$L;g6CW{4Wa@$jT_tiAI2N2hXpualG)-dvdbe@Z~d%HO~uK@8tAIWk_BWjsiOh
zL~(9~r3j3A>rciS<0SZs@dLVpaiiJXH6Ossg&q3LCDtHY8*6&DqcK)}&@#z-!vjN2
zT#7!NYo44FRRYB~o>$YJTQjqSq%4WX^aYn4)je#3ICKal9_RPaZ#-<+jF@Lw0eh$+
z%0^w681`A;2o9n*oS8LVYG_`~_)I{;Os=i3c~e=VR`*&E`JJ@{o-u@3G%W-^8b?B$
z(n#jDqg#m9(hoc}d1g;tt)DPTG2eNYc}7S{WPf-9<IMU@QnzJg6{0dw%I4Y(n5+1B
z_~?N{>#?Ei0Jy4r!lY6Up;GC8iprutDgt9Y=dp429Ah$uJRe#E9cxZ&9PkDzmnIWO
zHRm|H3PUS%Ud{B__IrB{`*<qa%K3JzN#GGgi<p_2w|C)`V9mO%FFHmjTGn?!oG_lq
z|7@tkaA};<XKQbl>928uCyn~~`sS8!rYo9Tva#gXzy2ux;YUA@hfkhst-BNN{PO+y
z5C6-L;%7g9H{Sc;V}Jfg&)<FbUOaiS7OU%9@$A`JeDcY|_|>}~#7}?rt9bf+GuF3S
zy1yM;TcS@QlXyZ3rGjvH(8{^_MH_mE{xUm9NW-eexD<^>P2*qF+?-Nhr{lG|cj6mw
zy&lU;gbrm6XuYauYyj!^F%<OeH99*tWByCOm?!pl7wj+|-?*_HuRv4u?o;(^XQydT
zkA?X;&kOh+9-7<(*X)U;Igaw3u>!BF)rzCcFrNtL3rsoV)FBO32J--(UU#U+xma47
z)41h1L*+8_@6z;-v4XF|kNHj+@cg@X@5Y^5x1(0Cdu@jC3p%^8xuLdjURc{>$2<g%
zjd~+iL|@tT0zYNb1;!LY2R?rCL^*~MUZSJWNbm&OZesy30@0KQ*@2LhoFBF?c%ln{
zrXahpUbY;hIf9`I8iR*9INk3{wiFNE*C%t3b-MK(#z1r*@D}D0{0jq)=$YV@1_b5`
z^8|jKn{iA?`Nsn#*0tQH-IUEm(Uk8$o>QmBhw;vhsMMKj=QNlM@(sT8jB<<n{E=u!
zZWMy{ry(Xo@^Nlk*yxY0sqoZ-?y=d4J`xghZf-snh09CAQJzzV4SFgY!z1Hc9G798
zfXWAk$UB8+C7zoYppk>9m;O*beIv3Dcu5-^<wvq6Wgy$9ehgVqebZI+XYj?B`=w8Q
zT0T;H3gZj%E`7kmmNf$SkoqfNkKT<)2a?IG-@-GCa+2<nkNS~yeJ;PonKC_&sy<=P
zI2G+1d@`PcwgMM6qhPGYBNBZc`VvBkgDaH5MkD4NFky}XFE&OHBGGeReWG9ViQjIw
zzFT$>Wu{MDU{CT)!j9qzduhJK=0N5V{mFKwL6ql3JHcJ^3*izxi47BcCmUmpcO>)^
zTH5VtZX6cgZ-fa~{q_O_zXT>W(WzX+(WO%pWozC@mSKFEf4ZS@6Zh2nJ(an)+l+0|
zE<)1dF;5ziyb4@--|f@A0am+xnF>sIk)NBs;Pw}!7q$GCn1J&?<@#f!&(;A|zBYVa
zr|a~!NtX@@IJdYI)AeecNtnXug(42a2n@swntTXpBK!ydsSL?ojxbmYCSqcQb0f?E
z2mX4eGQy3kVOcz2&XlQJNwoA^jDdS(6q(_H?hc60fWbK@@9956)Uk8YMhh)O*S9y~
z(bLDVxwBy-2i`HX2O%I|M#!Q}7!U@h6b4~TMIA6+j7_wUzTsMZuCfpYP{3i(Ldiyg
z=_2?QVFT$u7YrV6F`mQRX|IJ3!IZXOOhKTx_Z9unZy0ysU01w6G9*DnfAo)WpiNfb
zc@Y^lkXZ;+J7KB_6S(^$q@Y|wFtN9i>V{FHNI*G2n=wjR*~+^bM=yG{Dnx*y|FpwG
zlb!(+1R2H|_#q@C5T)`?4bdlcaOf(Zgctzi?yK62!i$T=1q<gwNXAG;SO^3!+6=tF
zH9`q+05ywB3|AI*BsigX&<`z0$MU=2daCxR`b??|p$dVkBtemOl8Hay#6@!v;xY2s
zxTQ9;$QhTQgmDt#Ef;3AL%5}P5O{$-umN6-8SmreW&GBA;5@jLxW<mBapMUdf%F?e
zn&04Q?if=)RF?%y-D76a){*fH0aBXQ!m?V9iY|6gvPjDq^wXN3x?d~D*d)pV!rsu&
z>vWy2)9;zSGQ(GGy7ITVBT+m`w1KPON&ZwGXbZH9?^mCs!Tpz$c?h@p{?hdu>E&`h
zleiCk9}sWf%kRILxSLlB{-2wxD61%gulenB5>IrwT0UK}_{wF}MC)$fT8QPPMl3Ag
zK_K}-JPYM0vPLScFHDE=%4S0aN+Nb;qmZC_ly2i=>EQ+a$LN69GrS(Uj)IbW@O5}9
zN)z}S&u9)_8PI)TJTWl{7szTy^qC#0lv#u{r0wt<+JnNr)!DIvWMqU8e(dBH|Hg=+
zeD=l`FL4+v<R#<=m0?~h-U}~5na*zZ^sJwXYHcp+Gs{t`&c^5@3iMO;vkSChX6{x@
zPOs=)6a<_f)fPxEVtlfqesE?KVW!#5%$a1Rm{wVn<wfz%hWYR5nP92+FrZDK;4jVl
z%qh+dqQ8>^<@YGYEVU0q0*YOBL&J}eAL#4h{-NMB6_q;YCkYOxf}db9HHBgBmU!4w
za0U-&IiZ+F_81z+D970|7(XN{pwJh<9p23zRmG0%Gqs@{)jI8-)m4wFX?AVr<{^xH
z!z1c%U>pF>C(&uHMys{%F-%3iWKW-=g%bNBjt|?a?>t5|W`r&CyBL{<7~3<wqwg5W
zyu?K@gu+hpBJzgA`sICo5@rqWddsHv`XYhLczA6&G#{Q@DD9c+7%CQ*ZrbQIseXYU
z7^r%kwwiV-d<PdWxRuo}JlF^WNxjGh)SGo5_O{{|@BB1=^!GoG_uhR!KDqxOp0BQZ
zmp}M~H{JYPBfj(PZ^fVe#eW-jU;S2$O$&$AXGt9z(^|yntngtW-gxy++`YXLg#F?S
zwWrUwV|9zM)g0*`#a^ouTU*U&wvb~^V}8C7Z@l@c#+UI4jgREE(VE+fmt87`gga9W
zT~pC-V<1?K?VZ(F+jwHxr%_vo5zQ?tQ;mg^LytIjXIJ>tGt7a{G=>9>al&1270^+L
zVN^iY-Pu?TqTDSk&O3BA#$x6^o@<<S$N4{;gEpnINXVO%Qz_S?R-aM+G(4i{ny%E<
zml@TeI!E#J0WSu1tDQp{KZpWhB{`!GV=Z`eq&YyDoP~xG9@(CJ%n@W?WIn>@S^if(
zWJHY3tUoR;@H|CHH!6H-$J**M%j+1YP!uBAqDTZkwSv&a&Uj>8$;;C-g0JSR`hwvG
ze$sC3MRRY<-Zt|KH%vdFO>hFn1@DoPU@_z~@)EB&GwtZ0rT%Hn#(dX2V@}!lYkXtf
zg`)A;UN9JJI4il2>?=CRc{XeYq<@^Lg~u2$@_B8@!<zHR^5D(|o`Ih0lG|C&=w79&
zzK(I$7aJP1M%0{Ty@=u(BU-P!XZd_;Li1swB3$F_yjjten&wl*b7*U8U3DJD%<P=n
zPMc9GUTSXbI|S&{XHTLlx=BO_ybM&acnKkdRkq}Rl*<?&rz;wF!xC7tF-=)T^O{;$
zZfY(w=Bzv57kD%3q2%`q$>MAdfJfmqOf(8XOk+lf%HyL`y}K5>yX%4tv;br9SbTWz
z!+7-Mf%=SrGdBxBW4jn?)NkMb&)C@7j<xkot&@)9{{1KM5C8Csc;^=%`rg)VGaf(N
zh$m0CV{=>hv%c$VZx0XY!`N%4F}v03+G}%Xw-bjaTB}{+p`bO(C5D!&XvK0YF0P2q
z*3_O!d!sHaEXC~njEz?G&E70pe=ID_49>J1iMg4Xc=eUrq8()``8flQji7|agg0z$
z?g-9D_NwANHfgd}JrR#3lpf0bo3~bM_+q`Zv9=ix9uUg(G!_>yyv<qZz50Ar@JRzJ
zn}#;lH(U=2cD%JP#(~@LI(XEPXfEwxeT#uK&+010%uG%6cV77kod(UnP#XxBcB(SB
zG)67qQ+it=AM6XqS#K{duZZr=Tj6bLrL`d&O15{l)W=QZ5bJ!_$mBmiyR>l%!y)Uu
z^k}(s+t)R2oI{pI6Y)8XJG4Sm%rK=7B-X(g*6>cj@P#2O+lKdaZm_i6$a@3h8Nbsj
zXt1^w91k((3ZH1VP|9^V8jRr0+)+V(54lRxsq=Zx>ppx9TFkCiWF}-NDx_}8p>N#c
z8SStvH>d*|WqC`oP8xC~H#j3y5FSZcyvwir{jz-N@4BE}_B7L^@}^jQLbGWz^cGr5
z-+4~oeMXm!E@kJVZpOb_tC~h5-|ge+qkPnc;T2s~VR&*RAHy#VDI|=v(D6JD<JA3w
z&$2mzF(*uBK|7E`N%Woifjz*XERT=MP9x0=*`M#^bszOfLu>$~9Mu7RB443HpsmFi
zn4c&Qc;~SKW-3@8E{E^bpJU0mq#O*MrF@!j=NT8TCG7o0?_=Z})*P?L+}x~Xai4jw
zxdAW0gAZe(!}m^R2y#3%i!<ud5Qbg_Tu<Dlozz7F@1O~mjR)<b&GZv5bT*1GX0+A#
zs${&oYBOcgUSI)^X<oy(j9;{W*mSfjdV|sHKr+(FiN-~9G<8kr*3gBW`jdG>{j~R7
zU(kUuHo$_l1N}qy30`7YZX+LU?wZ!&<<%Empd56W@K5*vBgy!I`#I<om*abN@tZH7
zlX&J&|Nr}>qSXJoQo``dy*``3z-z<Tb-GSpql7>-cS8cN1WtrJ77+*lR_cqvz;Iy(
z>`Y2QD;XE;5`lpQPBJom7bCF|K>uvSfhi$*iI~7Nl+AgQ3FOkA0^tjk4MP)a^D{Fx
zGfI6b4<Z_31MN9IyRdg!v)73y&mPCx_LdfAN6MN47HexWS;`zROfUE(jM_U%f^rH2
zC{rMIxm(P;S;$HMwWxyuvS_#9qcRXuNtsve#%QJAFm5Ze)IJMow2$2}DuX_806XnT
zkFjI#I)O3MhH^;)x&&KbMHm#vD;A9Gb)~{tbg=-yyNlTF^oKJ{yhwJNG87X+5DO<3
zmnipu6J;_E6vS_YNDEzR9}7Yf3vd=AENJNiz(W{A;IN`keKPn|jv2YyY~`Zhcq*Zg
zkwKwn!<LeTJ(Z8**G5fUC@V+Ul_xkM?3_tR0C3<NIF<sSa7@onj*f*>C?*hO+;oqT
z`>M7g^btA}&p^hNHgP3xsx_(?WgKur0Cis&8@-1x%a|f`0b?IDt%VeeZVQp%f!e~@
zGk)M7<+9`6{Zd(0oLUK*2Pgw4>U2G7GYhXH;IH2-V8A!e57kY7ik)`GW_=P~RM7Hx
zG8PUCru_8HJ0-QiL6F3otfB?qbQR@fdLJB7rYF&Lx=z>Wccib(@Kvu?|CTuq&HYsR
z>~E6pi7c20I(<$aZb92VPbmN5^inxVK3^ohe?d|3&#UKyJka3ZXqF0e-a&ft?Tb?8
z;D5iVFO7ORe(-~D#lQWxe-(4{DAL7~Q7niqqu?rb@D^cbp|z%ql11PTR;Y<D!dnYN
z77K6qiIoMS+2XvuKQV>kD6GhZ-%~y`7bP(l`3XA-pJ$<qv8Uio(0pDtpU?uBHkdz(
zK2zoaJFO26yzvBsQ=U=8&RLX7@CES^@wQ@Dsbv$r2tS7xkS@j3rlzK?fE+1tW)6lT
z!W<$m<(Wbg<+@%&F%SI28%FR(Lm7ag9Hln$iOL_ra{*X!FAX}}o16st;<FfrtO!TZ
zkBpR?V8#S*-a$_C1`x>zQ<H=(tErtB+7A6$tIfvZjXN<hy`X9ZZ^={eGR}1xFQNFu
zFh?C~8|P(#ySQE0P?R#%&@f8ybMxgQ1RK0@d@S=ba(LmS0c2P*Lp!>iT`RfRX^ip(
z7@r?=fc@iGUwz~aF&GdgONECEG6{wWLOSZ3HxCtUpgzg57qnfqA=?#a4+RFhuIgli
z4zNzy<`NI3Q#@NZJ4f=H%S*U3t|y1>z~c);2;L7UcEJ}6jpb53@V*)m+`(s*xZp9$
zcf+PQ@5>iBfg+xS2YWA8pMDZg9zC=IwYk@djg3v?E4WH1u)D9^iMPM^cKqOne-XD|
zc~kN_o-xSh$C49(8OB`d7_l*7`Sy+Y_MiMo+`RdUWYY84+1j$PK~1v4xK^)4wTfJ%
zF_|2VS6;bc1#+c=M-ZV1Ul<Cr&3vkS$uA%L>c{aP{y+cEc=y-8jOWju3b#gLabYPY
zG~W)o&FHqbqP@Es&!0Yu2cLWtPoF%rjKyZ3$~0l~AT}Akj&g@bB%@7CN3}fW_a`;}
ziz_S9m|c)eG;bp*Ml%$2%%SP&dQ1ri2=O;Io?eO5)q0E)!jJFMDeopsUWPphLm32D
zxXGX9De^zEq~&DMCTJ3}nvLJ0L5GJKf|NQa2al1x-CcbjvVyc$uUJ{ieBa*Q_PIs<
zUf;54t=@>n44yG;6uh+Z9{hKR_7Tp4!!wHSYM<u`_;w~3?g$U|VtxI2v|CN%8|N$|
zqf##G3q0Xam|K?S+&NxVe^6}VxyN_YRGxb*l{f?MLTf2~vJN6#+ssTo*4Ne?Hg9Hj
z#^KyBQe&9e-rQ7uoI9AC>rfCP<8wx3r`r}7@t^>Y(z_OeaJAMj-2m=*H;#zL5e^=_
zJ2>dZ_Rgm64M(+7cZg0+H04d&KLoAPtx9<^UVH70m|1!)s*OeQj0y3GfmfdB*3Pcr
z)7<l#2g4~mMIFnztIrA5X9c)E@fgE<Y<Fkd@MFEgW&@0us#0@QI8+vY65TvK?nSHJ
z_LvX~5e4y{>e<}f(7Zk{Uch&HdcW827^G#rJ1YA7_dohkJb3shR#tB6UPW-bjE%MJ
zc>MUe+AG@8>BrV~Q)^nyjprMxr>ker;@PtewFl3%LRk*~Dyd&p&zBkTj9WKu#?4!|
z94Zfm_0rOE%+D_v)||OltIild3CB3Cx+~Ka_n-L+K6@=>qbcj5^vuARFh4gFH*TzG
zd?suJV2m2VXV#g71Kix&wxJ3mB(#}z5BRaTxS05Sc;NLEI)$Ud6MN#IOow-`Z)~Wa
z>sBVS36l8_A8HAoInOG+EXHDMYg50Q!fDn}!fD;7t;|!_SSaNW36-k$AqULP&Z>?X
z^?%N^k<e*YuIrET7-JOHo0~go8|%eM;qrmKxFBOIE6X;V&~A7&URw_zK8R;epX$9m
zjXR!~oC_-aF%4mjHYwQRrHz*<p3gLO$X?b>^#`wN+Gc$KMk}?)`VGr=Ij;H1hSGDv
zhcz;CL*kReJ8NF0-hz3Sh6T#=8eFiZ9q=k>1@l@3QoP>5xRAyu+T=}_qGwR-g2vHL
zaGf<Ua}PO)x?H~Iga4d|PxA40_7NRJVZh6G^qlp%p3|o^2xM|ylu^7txCbEo$z_<R
zSH-*jmpTb(FgS{yg$=ZL+0b@2epm;h_i3{YGy2Ut$aZ*}Ql?=)xCg9hr`w}42(^l=
zNVynvDHo4N*7wv0zs9&(7^T<%L8K!>Swi!STk46&Quzym9dp9^g8~eEaX-07KL^(!
zFZbBoiU&Q$Sm-9txevgAHJh5DHP&w#iZ8|_(~En+#JV6D04PT}O7xlWa!%)C?9yNd
z481{2^?-L6q%i!_KJY9JUn$=JGumUEb1K@AE#v<}KLmYFc%<TjHl>WsW)abl<2>^o
zSTe_~b5kJ&zAz?SStoNKWomSk)<H^ig7;xGZgTcukMp*uFTrj&s+<eS+qx~fgszUw
z;26N*4@CXJ*qKI9@<JoPGYn{Wkoo+$?LE=Gp5{7{nCO$rr-AxUE%c{G(!<Pmo@Ec5
z>-S1UpYzS%b&-nC>-1$Nu=?8Yb)Bx$*C>rhz+I4FJHNCf;r>NzW0W>Ft|&o(5u7Yk
zpv(|cgc-6hf{erD@p3^pJJMnZVNWuP1ZOcWEqG9Fkn9D-3u3M$*0(T4Ghl=*m1{<h
z&}5>|XEGnZr(2T2tdMDSTk&-5c|6}-kNqP;HDGqqL%j<FwkHpaUbhfH5KIu9S$xt~
z^1w7Hn7*=jqfXw_P3J|p=RGTh1Sgu!qAEk_DVy@qZoIf$zVgFF&AcV(Qb$Qb67M2Z
zpuFp7k!9~6y=|pPsVpI)2!A7CjI!O?Wb!aigaVWtENEFkv&iLL>Zi{btY{<U6M7JD
zG2lx59!uTxt|Gu=7$f|qFlIqoaPlI<<DJGN`ju@)0h7Y5&t;?G19&okC<75hdBCnF
zEBmydv%$-G8AIV2yMisKNbm%f`pmJG&;qV06N?AN=3I3$7Qli!SeP@qyoWN$N(TvD
z2w(`-K6k)EO5zH;knjqifAl*Ad3FJ@1C_D`d%>)1B?b$S0){DMfzK=qcn4S@Gy`82
z8-!s&@HK4CIPXjNqCD<{`UG5Zfl3^B4(=9SsZs2hCCrUsqW`=gV085^?ec<KcWfBd
zA`E;ac_*ko={v$Ef;T(OCN)<%Q^bdnbDr9Dx=z>Wx23Pl@YSf*ujJ(9V}r!^mwyjZ
zQu{#~<oVP+rB9XfMM-(cmsh@%pLtP}?!36?mp)q)d2}xyCSBF_`AN?Pyy~|ligI!O
zqWpRX1?G3Y{YL!zfB(OZ?|uJUQLa`^L!s>~cG<;=!bQZ{JcGrs!{Ca?KzkjIM|2%8
z50n@vIVihO20@46ee@Gqg*xHGBDq?u9=Z(p4GVS@Qt$!j_3;Uzg;3OTCQkB0yqQq2
zor=!FC%njaJGsW9*Xu@4@<_RiGK^40$(xWnpzE|BKIwMDZ!DMOW)zGKC=@7bralwp
z3QFO^;6wGuTkrt-jADm20Np?$zok5Ne%ckEI7+aJkEQ{MUxZx2D+M0HZ<cPr93G>~
zdEq_Bx=kK<>qHEXrI*a{LEkW{&&)@qaVy5B<}7a@|DbR!O*dj>qM@7^moRX_U*TsO
zH?=Vh2}<mI9cE(`#w_ZEztVmTGVEkN)a~5$n@at|xHwBCCSh*IB{NuAbGaYw=6dYy
zZbpB<E15*_8oo3S2NDPUo!H&o@VF3Ckn@#Zgw#_R!Zq;C1{IA#wol*BRad{G@g<Zo
z3LS7Cd8wuPRyD3I^$Xb-Jd%8N*wc9Rg<G7PaU@*OsJOrMNiv?=*;E^Q{eF7hpdgbR
z0FIRLu$!vc$b}LGr55rX1_U+A_?+dinT4?vqJq2{qwS4{@#}Ygp?Xi_8*jX2Zv=20
z+h1vFJZ|5)6@T^@e-=Odi@%On-uPxz>T|-aaM-_N!3wamfkxG-Na|o5B%4px9h&Q{
zH@^|JS}7*R&Q;z>aPH14uiSFzQNWK!Po+8;Z@lq_Lv0S(ORs1z^B){aJc^5M{HMSB
zpW}c2fBawL{>PtK0mYnQoLhUF@#w)v@!@+vkDvYINAdHY{wUs6zdpS8eypxO6P^%4
zPB7Da1J)N8l9M&g7*s&xlY@3_Z784cXL;pL+<N6Lty}I!efCDo%<F2*N#2%hG*yif
z2D;LGj0tzgrs`@gZB+{|@H}Z-Mu$FFHXK2IRl4ZJ(f+Q+Q8K@9a!B{_#zU^d;6->l
zHl10KCR#NvnC|a$mSZy&qN*3eCUh5?1I(vNC`m7rryU!cTd}#Z6RlP!wzsyTTAhwt
zckV`ov$7^Is_e(w^XH;LXAY6hnO|wB1!vh&exMF$KAj&&ud`<(5k?s33{e5VVM2Ef
zU7~;<SALY3x~KKS@ku{~t~Q7gzH@AXbJ#`&=dSV)CX_lPkvg0wbV2=_EXC*qUU4<S
z=D^-v<*5mWbmJT_JefFC3VE9|z&P^@Ll*F9w$#4;c9bSY1e=oP&8cA3Gp!)Ro_Fcn
z_@!}CpO|+SmnYHgG&Mgkf~JBQ?_DdWMH2}>xZm4V(@tX=uQS0r0vKT+JdJ*5*G3S=
z3a_P0jU#;_>}6GBhq3%<e?N9@NZD0+Cu-l6*Ikbu-M3+n@tvDpNF$8$6SfZ}DC>>U
zp|e<h_AH)1e=2;0p5Q%1sLhMmYi>){7c4a1?D%Ibh_?@(hlJ21RNpUt`L5pG6x=2S
zi{sc>-;H0t_i;RUxEdQ<P2oh#hOb^<v_W)%J`f_5(1V<PQx#0AKD(|eTyMnE!lJGP
zha!FT)w^-)#tj>JDz&N&Uv<rg!We>K6=RXtv4SV-K0Kf>IC6hnu*SFrPIJ~YpOfk{
z;RI`{k2OiV*$zT9vw37s@WVg=eTMf=Oo}h6yt&!=c=fec)y|6XeQkYBWsSNVj9c`D
zu(>Ub^V-@*?Cxb)!RboL^9JJ?yq0w&e2TRfVJms&_KTLV@$~>bL#Jm$R0q#wjS)r)
zj9nON;o&Bq!ee9s#1!gbUCtR;JAyMCP>IZe!K~AT=N&sd<LqqRvIuhrIe<1`*o0pZ
zCb2H~!WYcng}KNP7~6ma=jyGkt*NX$hM?Q&7{B2Acu7M`k``-iDI#6SDc<-vkVSF>
zavH0c``%0`83$R*-h$93)z1Z8Wc}(nqB400IAOHpK6St=FberSW;wi=FKRt&TpQBB
z6Wwu<x|e0EjlTDy{HsqFpf<RyY)fuf7|4XW&l;M#NU+_a{QSaq@{<Oix>p!#eCf|C
zdi~=&V^2SdjZYY)s0%vbjV8uz^g6U(WvQ*a349A$51rLZnvdwH#G8c^ya(LIF=Xl!
zIgqi#qma3Ve2M&t0Ubj#b>gktYPZbm9U@U<gnlMrmG#qR`a^!|dbzj$&3#?8&7Xtc
z;67tcy`Pf3^~^>_wGkM=7wKXCm-V3_qimp_;MHC785S-=ci^chi(XyA4err!w_UJe
zb2moqy}hR9e9}S26B-UK2@71F>ql0nWc~I%J)<1P&-e)f=sP+YQXvcT8-79JJ@Zzz
zpSgt0NuGQL9@921@R0Iew3T-bk4`imhxT^e+v|vzXzm>&e~Nv%QQ1b?#@d35kcjm2
zNcsA`^tL=YgkR#ps`)Fr(CHAeUi`1!i><9a@ebznson**Ui2vkHor33wAq!@PPY}v
zXJ22WqD^0P@&E7p_s>r+@K@!13D?&lfq~bCuj_Q3zD8+CLeJvLa?H#T7A2KQuua0c
zL9%U}&^-%7B8V_@twqh`VLdDk3L}99RSCoF&f;AJ=G>K%MBI#k_w}7iFPJzYoFlMY
z5!`}F`~OUQN6{}aF?xsl@oarH?mu}L?S2<N=65VCD<8%Y1Wew7;UN@z$B&pJ!jly(
zN(i?;D@@Fs{0J3zLJ`^!Z?<Ab35#47c=Xc@8R3gE)K&x@;_MSv<5+?;4S-==fj78F
z2u=txz=<}n3yb#f{*e|H#0;hnhL_;L0)X?xP(UH{0&5muw8;jh!XTo&v<c=*S?qd7
zcmSRV<n#%lfcxyS!{7zH3G<gi1&lidFWN{M)Xxqo?`$H6o)NAOSX%fHJ|HYwamY?p
zm4kprm_l~UFnktv#4;%R;tUw8t>c_&MJPXZDr)g9IAGLccexkD`t8Mp>H<IL$B>m@
zDX4HLJ_|I5w-GE5)-2R%9N0OX?<ZVnGm4@U;VH5Pf(S-SaB*w|;Ta>^*kEiih^d`|
zE6HJ8bYEHBE_N%j7-3P&O&j|33_;N6w*g1;sx5}|89Rtla3xU=-f@P##v_G$wNaNE
zYPfsXkowD67r5Zl9_6wEh22anuxQ87C~@Mo_?nq<m=bWe)9#?BxK7vUI(>2a$_!sK
z^~T?tuQADgM<U0|*O!~1-Y=3DA}?2RrFi(`rO);$e?!|Y<2&DeGydCu^Fz_G7w1=)
zpD7ceb4A)j-?hjVJ%%r`836tS%_l$Yw;@a=!n0C-;KJyX8%Utv`Y^BMIs6kipqPTD
z76vKu89ZXNEKX5i!*5UuaR;RsUK%I>t(a0-?98St<buO}@qf|y2@cE`f3Sx<at&n;
zdd)Y>Iw-Z&4l5kUptjKuWCD1M!=mXs`8fya_@HYZ!EWUtfMaPu3-jKtWSjI@!rOk7
z@O+Yx&QAJzRWMc?@%XtsJ5YOb;|s<!Jby4E_B$;r#qo?Gbe$E*L-bo;CI1j|aA+9i
zGJID2cL)VFN)o}$2A?#}*@$+QatfR(VFn;|Zy%*3-X)9=vYnuNwjW*9je*O%>ea5{
zVe!60Lf>)rR3|VdqPP|(7R-B+vAV*E^yc7fGl$hyyHS|;1P6>_^d4m~A)UC?NVNmS
ze=m;J4$hMTsXWH#2XWYKIaFN#u%~j>Mm<AO-tSW0x;|+G3t!=l;Ev2ICM_I149;>z
z-asDl*<h-Bc!+mfR>U%{D1_nb$Y;})nHZa@7!G*4;T=LhtnAXez;KxJ$?&*Qofus(
z(4C(j$L{vJVfyxWzZXCF%YPH!`{92R-}~-&qBJ=XySuwlDo@5+Z@n3JUwbQRjd|e@
z1`EPe6Bb$W#PMOu4k{72q=A&An$@<Mxs~{~%Kpnge>*t)1`mZ=ts39_#+xxWKVx|Y
z&yBmU+>Pa>JHlNIo+^zAbBR)LFHVozHpH$yeHcIe@!!Yl>Qm)I#+>k(SU>&A&*I1b
z_@j9L{a?qudmqFvfBDOJ@7EtF-)ihO_tcMVje9pH)Mn&V&PZefl*5<_|M$Cl@#N8i
z*xr0@9KcvKH@^~<`n={^RbbT^oO8Yn#udrvQ&U1Jyh6YM=8IMi!h=q8CmyLB<a-pE
zbU~=3-e~OF>ppXfcH8KLccr~9M}_m`N!b+H9FHlKfEaU_yPUZ-GD_$>%|-Bpnwdx{
zADYFW)~Xm9Z^q){lHfZXE6a;9JHKc{4Q;@fwYjwsvnq?QjEs4%g^*F3vAywJIFn&C
z%a!S<&CI(jys}DV3^0UcKNjwBh8W?|@IcM`;H<-%Pm~96Vf>IcGeqPF21am0<Hh{8
zL05D%6NW$0E4AZT>nNd($CY(eyREgJU`hBw!g~6AM&UQ%s!opjqT8upM0tuQoMFwp
zRN1T{n2RTx3(P~pUNZL3RAg(~2u;OE#(7s5jR_;#624%>6H=#va+q`edhz_(LmOIH
zx72EQSq;bT-mc|)HVjU{V?=8SGrF_8VY-BeF6~{{dXlq}@V=`|S2gCtaddJNds@q`
zudSO-VhCa`GnW{1)*MqM!Siz1a~@^-;o(W_wYu@)habg{fBaLeh1<gIW_)<>llVwi
zs|lYF?d5zJeeds|h+hnhsu&stdpt=UN>=zXH@{%dEpW9~ulmd@j9&9|bFnP^S(u;K
zGrWui5aA;{4_v6K|D-}G54hlS_FfUqh0o$q!CC@i3JE@imlilq)D7~}U!uCOrsph7
zV1@VI>`Wt;mX{r#aCv3XXDI=5F51VsWoLWWYiY)LTI)T$K3Vr+ghP@3{Q0UC%)n7B
z-D@BC0UP}AT%(WlgLN;V2q^>aEY@ta1zyZ}QociQYEF<C2k;Hrh}R!5AY?L%@a^p#
zjivBae1iVswSfVQ#Te%*-MW24^KVA|MsCP8HAZr1d9H~wq;b6=e22yeM!A7zO!(Wj
z7aPVn46r?v=*StGFR9<i^)U2<dFl<1dg$=SHs+-ML;q=vv}P9!;rr=*!n!>T4kXse
z{CP4C67}#dZD9`ZN&8a%KvrUG^^P*SF8%SDMlm)Txfr*}*r2kqImP$9cnsI#3vwKJ
z_|AL$**NApGv%D;Gi9uNJKrb0TsC?9svj%Yz^K4|aFIO4`)7g=xS%)nEIrP=rUy3S
z(-;_e#yhO%@yI8N0N(kGDeG#l{{>@Y7rv8^F$1SC(o!#mQRGkNAjdYeMaM9VQ7(1S
zPTFFe6n)HlB2NQ8kZo3?KHBDf<)<(5^V=7jst0+P&%CcT(2hc<#JI9i%6bIgrA+iw
zuTa^b&s3=Y{Cn_SiLr4R8gpm><sy^v-R)E%l*#4<U<xcS;A0FTA`x&`#jYcbMrtcG
z#cg*3vP{aOP8XRt;_-ANi^NTun7W=qM}}NY8}aUAUXaPK(7T0^$p$iAw5KqZV~`?r
zDc+nItQ>xj`V0)sXP3lxUA|}RbsoxtwlUXoV^Vq(=G;qubZ`JIKtG6{jkza!N}F`K
zxR<-8l%=~iU4snzpuaC)MUKzAKAZQqU8LgkpOU_eGQR5cRkRf;=-Tjgovzc@C_$?$
zGxexTkR6*$<r79P7z{ikcOs<=w-tF}a4-m%VxIlLz3`#{X6%q|yr?%|V0<qW{R6sh
zeVYv<FvsE=5HPJF$T28mK-jpbM~Scq&sLws{ilz;V7^dcES`=BOmn(S$c9OMPhlUX
z4tu1{2tqJa7SsC&?BLGbgvf#jDKK95Sqn?r3`2zhvdBTnks+8;nPDZE7E0vdJ%rla
z0WPNR=Cflg86P|82zQ8JLj^Ew7EkslyQ(u4bS#<)#fT6F1bBzM2t>e%T|Nj#HW2Yl
z`ROZdv;d<^(DpbY#8C&X+k8j3!6?V>t`rzDETTcIx+nu-8)1`jfQ^L<^(BR&6wYZE
z<HrI8ShE;GD5ZWI?j$sNM-9TnAnYN#yGFIZVrMKDzYhdk1W@n?p%wTs-iQ>JYDeMu
zg|d$^ve%Q~&w}1N{M7~)alpc7rwI@EjgW+(!q~8*PY_p$?D!k^?o-+gOj!Wo0p?^J
zL%2k-hf)(oWGb)d7vW9_yMiIoyO#t6>OepS&k>%1A`2drMBor1qte*r?5>xqsLo**
zShNd9j3u~9=*#prU|pYHgbWV#p!lOO7LMg<EyCCVDVTyI#2L4N@j6|n>-2f)D>Hn}
zHq!9LBsTr}EtmfL)TRHOuAcuFp0Wo?my!;P9*bT<6W@C4Zv5%D-Y`v^6x}m#I~V<C
z@h`f`1`G3N7^B}i_iq%qHXJEGiVkXFlL+(^7{a@u(ePgkU@#A0X@$6GC+`zljRd_%
z*$S6{x5L*^bP(PI#h-TsqYzeJ<c7neeX3Ah7|^H=WfXkjAa-{5%%k8>+-HLY@&KD=
zP___W7ar$L6z0Qt5YtYS9eAA63h_CWt7`T8G`g*=Xg7Csm(3vJC#Oe}8AR((55?QW
zqg2m+ud9Xrp8Bjd3&uVxYqX;GC!?eH35y2r8uiAF!b3v*k~#%jLLUx|V9Y?NhCD`}
z>A!vs!V}_cvfn2()X7!Y$Q)d?|Ez284#HjGwL=I-3?lt*Q*Gb7f<K#F&dv_w@BlBF
zR&dr89$ZI<89I>uAa9(gk9g4#=8-Cpq6im=2Ltj30>k*2+KCb3sAn$`&NcCwpn^aB
zM}E=B3T&!F{qOJNfzq-f)&hs*1nR?xa&XuYo*Y@eVN5vVhrXaN1rP9W!TW{5v3Tkh
z4SQ!jXJ{4PWt>5Vw22HNs2HDYKpkU8I%muXmdXQ;AEJ!)nMvtwh*tw=BBlJn#u~iu
zP;L<_aSECe3(L1-ymTujO0zM$ycFkx--jQ(FSs7s`{VZA*KGLWd^OI2!w`cplyije
zRz&HC{N(;$YOKx=)DJp^^1B=}3wTYH<KBn&VsCpVUVG!!Sh~4Hht;2vc<t_+v9fYo
zxCtI<WQ2Q1hpl+>@V&VI(XZms$M@p>cYhfleRMB&n(gQx67udaHa45_?D<BlZ|pdv
z+tv;T!gr##k78=b3Q4XB!5>d2WHvUDA&;J)AH~l0Mm&7<ajdOBm3%bebA;J&#Re+8
zJTTH+oZ`vzH1-d>dL}$ZX2qCj1LB3AcVl08Ldft`u8&K8;!IdJpB)H?52D-Ii?(2u
zp%Mu-3T@#6Ui%u?1Ig0+qBSVp*+_>>Hxbj7s&IkuZ>&L1Jf564!&!x_3y`P5E5aIL
zAj3#>p}A+zLE##P#Z%4a!vhY;*P2AMLDRr=31jHU&{6E|tj6<akJQc+E9l2da~j{o
zOD)4p|L{yj4yP_}zJnGCzcfr9F4HXiai%PZJ`*0$^v`|4C<?r|26$rcQLQa5B`<5e
z#G1jI1cW!euEx0ARk;|<hBS|Qf+ZosQ}%_PVHChibjl&Zro6r&q!nuduNgVVOL&UG
z7(6AsVWrxL3S*%4mN&d<3<(noo<gfeht36`L&1)-rH)m<nkZViy|r%R8papqECxk3
z;z3`)E#zh9DVr3xw>G`@@+Jl7hGc%$vutXcn8IMFb&cu*KY$NjZH@YjWbjB&CDD4J
zsmeZy`wt$)-~HX+$A9|IKacm{|0F*7_(42<wieA+Pj#?q;?zoGB71OF9mcF$ts(gy
z;|6CumW0=75b_30wa4pL)50^oUnqlzm6tWtl)aAdY~oBT?g4jbGiM=!AB09{T?*ex
z`pcMUT-2XDcX8NjWsGLvCp-@tjdCA6u)<gCi8H|ur84Uz)=n78@QT79j=_pG6?72K
z#>6q&uC=q)N$~aZbj6{;7zZ{z!t3y)*x74EtJyS;unwXu&IzhiI8$uiSFK7oMrf#D
zq6_6acn*HC&a}ac^{3jK3h^P+YsMQ-y<CTlxKFf;Go1N7ueQ%L>f$lI*xnZ3)Evdw
z#kpfMGn^-xUYbKADB9WRa57l?vgRGt9IpF)#t7aI4d4u^R<q@Gw%3Dt2l@p}@w^1z
z8E?kZ&y<H2GhgWk<DuqT)-g@ec-T0mysSlO56_Wn$b$@}pI+zl-4R@bkX+Pnkx5j$
z$dl`M-*h?lqV;#dU-K8<iH?N2NsjK~x4AOW7p_bLrT&Y@`mH~~$a<ZZ9}LRz`+7de
z={6$q6y^F)kMo(H(C$IM88={W109glzw8t84muC@R`oJQ;64374lDW$p4xCFST;o?
z=s!G%f*cA`<4az2NYHWK#aMnQT!m&a*60}sMT>ze$0YSi)P>$A-yIm4)fRApxmEN}
zFF3E;q~DC2KL>5hPvAPp@6UuKeM~xqe3QBv>mi&ESn<EV`aP#Xxi4VL|C!uoU3?Z7
z^}7Z74|GgJEyiDP7JijFE$U=gJV&TsxBQO27h_)-(rDt#_`t{DYtcNV#93eoT=|{G
zZsZq@xAIv=cGdKQ@=|Bl>-TMN#<1P#?8n|-S95>gV@4vc(`DkDlN*uqg1YivmaP#Q
zz=O4eF13q|$ToN}CpEr!oM!(M?pI|Py2bT5NptNtQ;y4*kzT-fFxG$IRlNIUr2?-%
zThEtyec9>r>bN$1U8n2xbxR`h;}R~X8WNTzq{0kOw19#sup<V=63iARjj(_siEwln
zf)LP6`Nf3UJ%#{5C_#sK5!0|T9yv<%2?7EmC0XPXL*h4!C_TKo#M1mJ*5Vftz#wm2
zhicd6&UV~;ct4(RtOed#ypO1Wrwxh>LQlYS5kl|~qh9w#g1gTO(}EbG(!!_u0+WM*
za-W@r82Mn-ERfkbb|$7rs0Km@ARN(O!aa1nBN>4TC6R<z6eeo5cd5b^)u#N@5^`YB
zu~3X4Ey2tS3l+eh;h>8hlz0l+$ff*dth7<}6h<rOQ(st+(0`OH7)VGwqdv|ONyXow
zF8ZoGR=TOZ2m|yP7}16a7I^xE$<uBGD{J3ScL=`xW>*>uXA7A;;;EmdAjTMYOc69x
z7Q&MIc!H3ir@|3-!4o$cp@uel$FttFan<FpFqaUdF>#yKNBvZ}DUdl#1VW$kvJjz4
z44xMB{l|FbckdoLNl&s-1Y5ml#R@y}RUeC4jBez$Fe&(g-_(J8!uacv7ggSgkHSqI
z_fB;6hxULoC>bb=b2Qkc4{iWk7IGxO0nC9B{YC(1j;WA@7YhO4qYs2_cC54T!jK1U
z<{38WWd=Tz5yKpUHD!45rv)kf-bbNu!Wdqs>vWwylfF8`7am{w{9TumXywc4)4%@M
z>8cff#3b4VALCQiLGB=33-dGa#_P9Ze!gx6Ly650$_@R;BZft=X&sq$7v9Ocx!BBI
z&CqR!&(k|rlEXLk2d#z17d#Y+gEHWcsfZo3YyzDg9cN)p+w`XCF<}x3rId8siqLV+
z%)--Kb%_S!K>*L9EQg)ZGh~LN<HJ;{h<8p-<e~T8&;lJ7KSF7TGK;*(RaXWY<sTbM
zL(2H%RP6Wn<DlOY9N^i6coNT)jCRoP#^%O`7Wp0Vv8gCkW|hAdV-w<w!)#b#gNoYe
zv%oN%h*u7=5lOsB{Tmy>z?I5hj4C6eY%Vzz`w{OmKSjyUC;&6@M)lvGXM#(wvk@EX
z&kPgJiOKU^QSyk79`?)w=?7=%91xmHFlCG;HLjeG1F!NX7xBz7D{uu@jf3S9@!UKg
zjj<8mB=6~g6|)l<YXtuiieWrT^lX10Lt8I8t)1u+j!^PIZDuyA)w<zIUr_*_i+9pL
zJiQ3lhp`QX2%aezfC%T6ZO%}}7}pp#>h~FEk}Hr%XXh6L<H;C85Ce~O&vKu}7ln#`
zs!rgP%Ju1~ag9gf6z#_VVuPz?i}YGUp;wxoj+wc6k9oV*R@oOC>&Y0l;S2deWefJB
zV|sr|_l7Y}GDf7~C{GMq@%_f@ug22SQY<X4NWQ|bsRnWGoANS!2=#~uNlD`{G>iw3
z<i?ZzxNw+EwNrSQM$Zx9)o8r;?l0q$kM2cvaHc4-%v*1KGv5C0528L#s6*k<FiOKo
z!Ko9EAN?|Z_K*KLe)Z1JWA*v-I5<3tc6UEEH}_(%*%R(($T092Z%o7=WYj`&4K8B<
ztW1|;WqCPntlW&**?Hla=FmiL_`$GA{u|3Tg%=g!?zq<(>`X^i!f>*_NtnPxy)z~p
z*>?njsS?I)#s?f{7x}ooJ*JJ9cxU4Mh!GUU7Uu*Vr;)~9kjw>8Hbd<7dhH+#7RGGP
z?=iv=PFcCnTxV_^9_)J_vzGE1DeAX3g{ck5et7dB|Mhx;+v7*Ewze+#9Er|J)>Qkz
z!71UoJ)Arhk%3P|6WT$DM&N`2W^^3G9AiwKDEqWl(fl_(0dJHDCl0>SE;gXqYn9NG
zBY47cfm>+|g|;#<7%kFMl5{F~^gVwrM89YdXXw@Hb+;9-Jn*{Fn2mO)9S<LVqP867
zIICV{ZZ-?zRmGfx&NJVb7s&6Ny*4_*COF_EI;XMrd3d_;W}7Tk9g5H12{yJemV|>F
zmRyVn=doak{LbcsVbScqLy)0>#mI}Y8a%@w1Pz8(gACB2zTnR|;Dt9cGZ%O`(k}Q6
z`CuFEcF#tvCr=)Uc62pP=W*}e{rJ1T`;YPH(c{?O-c_F;$NTSnpwCs+k5?shbv&w-
zG;+?*E!azIeqm8Kffrhax2KPIHM35Fe!*wp3shXl0IUTWGuGjRSZmDG)GyH%=%wfZ
z;REr!f{t;{B3@s3J=JS9uU9?xUQeRjhZl&ysf9K+iGNiqWx;z&@Hw@o3Uw1Q6K`6d
zL8yLUL}YD`u`RvMDmF@SP8T?j!XLiNS`sfmHh=&tJjAFML(j8kYhK${YdKH0wsu4t
zwv8jD(uC*#qRPB=>y`~q)B_!4O=H~Em~h4ro}(DGc;09<Jdc@c@NX2~?x$c0ZKhwG
zHAb17H-;Qxn(MhaC7OR6+grP8Yu`9OrFjKjgQMsM`gn4RF2GN~6XXTfls0sWFR^}Q
z{f%ruy_*}`_F}`pRT!(lCwM$KM;<Hobrr@C8>x{4G&ayEk3XZP{vaoyXg}510N>&r
z>q`a+JfU0+Kzx$;q>jQM<NGR)q}K=a!uOIk@PvAqGsUJ?<r-Z1!8=_3Ir!vVz81eo
zh0)780DceND=wZD->I4L&@cZcS+*M#$TR;Qv>jQmFh;>Mm;+Q!nWkT=8(ailz#pRw
zV-0M9L$US-&)K|c`O&f&9+HHw%+RlRI5FOgJ9QJf(rwViyWjx%pl#OQ0OP?YiFLW#
zsk+Ih&gpJGEBaIPiT3jq+$5K~4t#*&r!HWpG1GU(?dm>o&X@d`iO<0p`riP)`Oir$
z7<{?<cJ)chG4eO#5ADUvi4eSrXL_i6S!)(t7`=)sH>PFX#rVFQOpDBPkkQeHFn9AF
zbBv86d{Q6dM;V2#iHm#4oX~%~!q|L4h*&}#;;EJGD(WrTsk%Pf7H$vP$@lyXT(j?Z
zwIK&G?ilZFRAt8lup7We|9oYATrVZ-Z1~KdznSvk=fUWok_veLQ?D;GeO~#0jO)v+
z>+{NjQeGRruG4k;nkBrZrX(=WO8CU%3S$C0JlLUZMk7W5qw%K%sza1FA_y=nLdn^v
zASMO-#Ha+5Mi_uez!(rD_+;T~G6K`nLWNyMUYwc0=ylykBr10Kn6(aQF}p5yTe}W#
z_;B@!<BY>=SmT?0D<eW7f<8NoVZMatL-1xtC5s*0(-Dd&leSO|$|%~(F4BsaEryTd
z!~N(<c!#kf5VKIBzl8Zf_+bZk3RWcv!CJhDA*ZmsZ||xxEzWXBH@iJk;j^y=Gw^`d
zQy*}ojrwY?NuPlw-38{ri?$(%0e2SC(^ed$QI2{jLlE*J(h55p&Ny3a)C&TPGOS}6
zJ1-z4Ko`5DsFyluJ4z7*WiE_sC~putY|uksB=iXl7oJ=RWrKmSySvI{(alaNyr_=!
zi2>+B-+^(;N*Z~!5f3GlT$BM*p;*ecFloHW9n-nvkoVPA)q&B6-I)kX^aDj5<t6L}
z#vYoe4}=>GkiZ*{EiWp;HNA^4X5mraSy=IYwgm%f3NqX$j0VpTkZDt)NM%7tf4p;z
zwhMkNf`BUmZ($4set-z&3|?F*80HQ^aNLSo_mBF8kArrCGvr}j(4G^)0{9@@VNgvW
zO)y~LHNcl#%uNd}9C}XdasYmvuG4jTDSf4euff;fnf|yl73%O?uHRXCk@ugFAbwm_
za#atD$@7?>t;E0mH-8a-`By)PsnSGno)hwb<pbUk-9>>xI5B7+^fcuSQAGFwf0u<a
z0ObNahJ|n{$$b8S==Gq=y!0Q+4D&TThhG<=_LARWXn?ngK8wB*9*%Q)IHM7M#qM?J
zJIX-brF|^!@nSHK6yJoeH`{yW*(h32=$ZG3rh8GZKA@~+gH9@I;S+>#szqaNA!@TL
zTJYlupnU2lN_Ea_8ZTjZA^cW`5rZdH>Wvr~pHY(~vy4s0L<w0#vcu?PjE+~-m+U`n
zoES&Jq^$amH&^lO?6@7B_Kw;^7)8NzGzSVrKj#f$pwWoM$x%1j?Y6#;*z0O?vTQ{s
zutR*qP%}PRRePu1M`SL>tW>G1-ScWw)%<!?br_DNk}ix)d0rB{m7V>xmG_Ys8CR6Y
zcp@Q`BD08NyG~><LfY+aZ`r_6tIfsI^6jY4EXQ<ZQDxOMhRC0UD8k5?vL3s3fe#)f
z@L0~=q>q<^NvTwi@^oF#IfDx?nBySaTw`t}rl#ivqjSp>C^88pNPmaNCM0L19HyZW
zilLa-v(j`-j!TY|oJ_lE!{K2+_I9>nduP-5F+E*XdlxlUc<xlxUqXH134;d<%5&st
z3T0ChA(I&!3?1AfUBqN*JZ``CM!fp^p9CTEEQ4!);XOy_NC8uAQk(TIcs49N;aoP1
zbQr&m4+u}F`ZOjPT2t^i>Bg^r^-lcs$3KcEPo4-y$MNlNe<yzU!@r7Kci)PkG2t8`
z00-x|?RVGWSMU6Nyz|q4jF0X;kPM7`H01DZTU$+iKer5BuT96o!mQ>#VcP0ZtCTeU
z!V!%-MifG<RcqCF<IT6?t#5oM7ME^mJZdpju0&OHU}5oQ+`M@==H`~vRyJxuFD4~l
zRrEgIEjd2m5PdB|e%nB)xf#R4P2fCUj<Ly-E({|j%Ws_Z#G2vAXSij!Dl1Pl7DMWP
zTwwH(EUbCO8GaaNN>do8pc~9#=*Fduuv=T}R?echL(W8A>vr0)y}cp4PJ<`(OHIo%
zg+tIKA_{OuS+zPVe4dZ#+N|jgc#1)1cYD+FH5;QSU+AK7)Vx-iJ{OlcBib}M&a(_T
zkKB)95JSCkyDn%5iaQL`7_EjzD!MljmuG#gdG>T)FtmZ6(FK2?ThK496EyE9%k>zY
z(EQQd#rq7eDhvj!`!LFPIxW#YXv~G@Q5pcCOJHFdNTAcuO=Q+#jHseVcxsM|{_q?c
z%ou^Ery7rf&SaQZeNpFG+l*<P9nMU+Lt>6^ZEeP?)`##C49d`J3}45DIaPmH2hcVm
zey~xjDmsey6-v{0-g!5E{Nta;<HyfqV`Dph{oaT1lb^g3PoF;5^Fx(!;WMe$*Ec1X
zqYzcS>KvhBG5&GRnGLr$Z^q2bjPNqg)nW}mdr%(ZWkhJbV!g$cM&Gj6anM`lFr8ui
z;CHMu32VstRpd#zKaG2A$U{NvO@3+%(KOO1lu<Q+C-VRUi<SSH>ywkC%BOY7bb4yB
z{(w~QS+9$ZYfiG3q)oSO-4JbEu!kAEoAvJU@{;C2NoBL1X1qduVsk-yU1Qv4;|QK&
zgmfg_Bc2lQcAi5Q@oeNb=ZRsUdiB*;9oCS!hcVdU0`-2U)3vdd?|ADn((q?Cyx2gg
zzT3zxI*fsibI>r1EG;d>?c0PEoljV(gUltK^U3BCV8&((4D5K?I_$paVMlzQ4UWJL
zIfD(Dy?$S13`fo0nnTf6yO<Z~1`ZF8?d|P07X1><@j9M4q(*yPIgnpS^ug<B<)J<H
z^fSH2J8a-F#)eMTaz?X(yrX$&#rz-v4-)vmyZ)}v{L0foj^eQ{sv%j9(qI18>=a%>
zNA#6{SLKmhz7qMl?J?lH?ztcC3+wBYb+Wz4ZuyKha}PMsKX5xM`Al-zdRKqhrz{^>
zf;-G3M~2ZoM+(sRvHoRzfm!n6tP_KnH#VqF-i3~Uqt%LduV@x$tI}VLtk5IsfL<|=
zkQta~7|HDEq;h$NE(ZDw?Ia(`a;Yw4PH1)-=QOv}Hs%~{$?<1C1HtSgDdCi``P8Ln
z)Wx$jaOHjU1HhRaS<(QAdO=kv=|A5dJj-gFuF7V<Th3M8DvQ4IoBnvSfbQ`k?d6lP
zhE~E~(E(-cR5&P)@6^NlniG~^RE~8;y60zC$r<x&Hdu8hT`)Zp{Mb0bMf*J-DhEDJ
zKJzr=kLHHnXZ)cXY_8bdZO5KyjW;Z*KP1}c`lw3Zv&KwU@Nk|Nzw<|aI1Rq%7-k84
z_|n57YoQbR)b;5fpOaoZyQ=#-{XwZ1w=e7ZYyy+64PV#kI(^L&N}MTnNG&dSF^6H{
zkj1|kh>5cPLG&RC2uCm@3lisNR?H*pa1W!47m{KJR|Y49Nj)19lSIJ7$bv%0%n^nw
z1_^^tLFz@aFyMtw#EI(JYwg9OXOH8PCl6w`)r`|i2_X`KVA}Q`I?gk^U`7ZtHpZy^
zW|FF(cG#Or%$!}+v=u>!MKN*5*?|lrFFdXgbPzmE2vr`ts4#BOf5MXxc7d>VEbemm
zN(v}mU`W^^G(;Yfss(}a@Gc7x+J;dG!wSMWaCLYd!H~Lh7mJv>>ZJ_`N+QPzd-L3j
z6a>_7?y<l}0HS<^0`FuREC^6K01px$-oXM)bTKCMld{s-!s0W%-53vjQTe=xVJ3wk
z<RHR>&?gqU?6UEhB>MDOV!&RD1i}LxXrXBDBf%8mhknv`6n-3B&EnOHN7}7|?Nucq
z(ZYk?p)WZG2*_%KaL1uQ)K<pZJLL-DY9Q#+VA?RO{Aaqr2kNFw^3qO+A5^*Y*Mc6M
z)H@jN3>!XGek!@r1CVkMpn(J789mm#uW@0)0`AiXlozxcT;W~1;)NAE%=JEF#=?bX
zAUoh?PSCGPx$1Kpc*a=LM=tP}cUe3$Pk{w_+-$Ym3n*P)aH#ytA;K{n#>FW@>vg(L
z*J+UcRT{qjXbFP<J5s7Ge%JLo5`>8B3lcPnw(x-xR%9x^^R3t7fB1L*Ccg9CZ^iJ~
zNHp6lR^i2>g`&66w>&TMSS?7o0Xj?}&^z-x^DgeEVu7YwVPv`uPb>HXt<b~4yYP*q
z!+1SoMA`Q<ybp$la{~!{0*?sJuj%!>dQW`9b)m%M9IPQ5>!=ImCSFRY1G1wR_RN64
z(=2E^3TOJ|uy*PTp@YeX5oLUQDk`-`Ox9LnbbKL3#t8{Fq3@h2f?~I#a?{gil+6^w
zco}CuM#k{!Ldnb~4h(q*Dqp-<{mXNl7#H!;)1D1(DQD>(4U4nrw$`J$x1(`E;ddCE
zV>3Y*G%+q`lbm#VhL_sD`XTvn{$`Y_i@J}p6y7O*uKXk880GLLDtTiI-fPr6qJDFa
z_6-eTj1*6jT!b8fp@p`fU`=%d@){dR@J>K6I}=mWGwOfg`GNrm?+6rH8VAAW@*L$L
z#(=qKG*)7)Jg+*_8)jsPb5!<~Clw?p(WfRWQLE3|aDhRy-QJaKDcN_b5!1Cr)x}v+
z?DWpfCiR(xm@Kd8p5zsb43gjIlMP728kZsZmu*u|B(GwuLa{FV(HL}FyU}j#2u|m*
zx3?8*>raeN$ePvqyyW1z%3yp6O*b67JF7v+O2!t2FgO@PgqqVOxgsm0{2bTd+&17l
zQjgKea`f6;v9r1AIxcLqkvu8c6VCzWAsdG>bmg!Oo2{LvE(hZUXRl%O>Z-k*gVpP_
zVq<e%eLjzW_wW9Pc>VQn#%PIigK|7c7w6sBSiK()K6*EvKYJR@W>0mLG!`?Sml(X}
z=VxPPrWVVKb8+X+jd=By+i~~Ktyo%|kHv+VSXo+(n=32v#+z@%cfa>`eDC`|j61J>
zL-1J?>?<)T*$IPWt+5>C>QYRU8>)jH|6`gT<tPb9Mkdohim}NN1?&-kF;6%udgF8a
zjEBIE3snX%G-4SK!DS3xY@CB`Fhw|fig1J|)h4DnJB_O*JQfZp1Njb*`mwXK5xrh3
z=lx;d@I>*AEXW#af4`&qzy{e)_^+{;EL8*pJYMH*;77TiUSzbtYELL=#gNMAv|A1<
zNdGv03Oa^?052a5KNw>15(B@-g?D%W(LUtHlodzZ2fVeANAYU2T#3AofmO6)bL~+)
zee$u&h29ZrinSU2#4yGgOq#DME4@NRTT~_nL3l`^@WsH3Y>ihEVe3#n;{C@tXlx|F
z;DMJIXDzWdWRB2YXfboUwYTjv-=O>48q&Pl-q{fzuGvHDQgmu$L^OwX+h8{w>#L6i
zo2JH_IxE8cPOPszHtlR^j?r%nk6XfJB1xc>#yjrgPwvOVM~?;f!)Ud8@!;XJc;^@I
z$NL|Atmpe4V?q@kAD<b%BZ2{VhhZGrRVq~lPtM?)w|642WrG`DNYEyD1I9;($D;54
zB(%Netj3SA;ex;6%|+-u;07(s(0}j^!Ca7mU$K!Annd_MyhDL2>su807;zkeU1N^%
z2SX1$1Vau+Arf_Q52IwgQHjN+SsT7E{Ik|k(V|P^ar5SlG!{*lYy=||BJZ$9Wo#(}
zV;;P-t8qLKzO&YCHd~(a81(r(IyzC^guczXQ5d5uV54MVpn!J6$1&8c+_>R+*lBk)
z9(eR(NIUk$JZ%U*GqW@H(1ULhE)k;`1}S(5?ZjJ*K7qgKp;lJ?X|O<s05+_nfD2w}
z)IBwYcS$L#RYK8XBuP&a)>H6&>kyDzH0Bsg@D9XFY-eXr`7rLDxbGM=G3c^Zg!X`&
zSL<!v0}h3LA`?=K$`vj!&w)RDiVL~{k1S}@)ur}~Xx=gZfeB@SKm4{~iB|^i_(yR$
zWyM7vzJP1cNUyo|-Fg$jigr<ilV8jF{Y~|FP0jn<;=A7W`qMCj*K=_{Kb`20zPfMp
zQ5{UaJa|rd{HOYIEt`FUekDE5vY7)eTYuJB=rWzxJ&Y#c0do+;EHWdsg*thM&D_vU
z=xC!}cgRBWfX|c-o<J8V3!27z{HC2GzNfwc{-JyM&xu|zju_A=4_sp2s$!SJn%)EE
z%!5AD3w=*`0)S8drmoDUkpgam@4Q#wNzs7(;QzeHmKSB^=gY7o4}DHqIdu|O@*&Up
zP_mq<XYeQRn++4BEF<$@4ze@u)^B+H^^`gdE7j());;4ISK@zBAMbjTvgw}apl&lB
zyytf5J#>0SpNqL;L%#Ki<KB?a?y&K9P&Z>$^htS1Za!7=q5D~Xet$W6ZWPaXG^jKC
z?Vr4q&#ta~_9A`iAvZoJefrsF+i{)#<r5fuZTPxQ*Xiq)B%n=5U|d*P)`E;MCKy>n
z=8KCzeZq8n{ufMK1e$w%!ra*Dgu)-j4}-Fj9)UzcJRt$t4TaEwpy3@c6KVYNZYRBO
zCZnbdYBQtb#<3PioaOac3!I0~p2T*0SND@?yDbuI5u*IQ?pdJHUoNIZwDf7_F2M=G
zPLIT3C)@|xkFjNnwkVJLsyvjzVgN&w+fw|g4tB1wTM%Ow$}iq2YjK6310zOohGDZi
z?et_|@N!GiWg4$D!s4+I;6AW}iMl=(1QN#G7nOxTkMV;<*<`fvrQfuRgwcw&+0f^_
zu<U%MLUMtDvWc<MKj5m{`mX=<#pP0>{t@EBLMDq^-NuMSe~?-ch_g?$5#ym266!ZX
zDrZ!=ttTwR1fQ<PS_?|NztV7*xWh%?Sgf<SU<afPYRX|D24#WjJY#39;NY4)9tdD5
zjH*Gtm&UmiZqkUA@Zp2*vFOlG-Oq8e!c2=d8y3|U3~)XNfrYGIa!5&Fr3<AP@7V)O
z_{o?XPjxYdsW{bly#owUZlJIOm#CLETJRksBo5~jOa|xG@D2vF0#|4^3Rje)sUYR-
zv2&NlSlT!#`11}*EBfIbNAy$QQ4rEj3+!qOZ5&k_2si38DAH1Tovzb$`mfFKrSw(I
zerPS%XA{&H-XT_<zeKm#P%vGYiof}*x8u+M>^sqD)I`4tcXt|`al}3sc2Gl;^L$9r
zKhZUKGW32_^csFClIOC_Tg*2w9OQW{D0QLZsf;BQT6(5X7IYH&j8et4Uv!nwNhr|Z
zBZb0(_QEIO74Yjk2NH!3#w!#moPokwCF1|$Qk*w*DV|9g6RMYURa%l+Fhn7fpgiEh
z+ssM^m2G*+yky+UeGC^?D8qO!@j7X#-c!@`VTY~5NYsk$t>-Ggr?!aSOrl)IkU?Aa
zY{bF{bIDmhqim27=RNPm?$(;mzd1c&Gs>}f9eHR!e1!h?`<*xuKc1e(lT70>GObJX
z(gyLaa%ColN9)QvrlyGR*uaKI7{(Wj3E4(uh2cQK%h{djnVq52gaa2CV%RxM8w69y
zAoh4s-%-FLcOc8)u`^z7sBaje2s1e*m?1~8Nl3hMs;2f-BzKJoj-2^|ETVVB`?4*_
zNNEV%+TL^?LakxQnyAhz3+JD)L8K#|JQ|hSyvl9pJ-q&qNi~jRY6;#bLl{idF2NZ^
zz2#EIXkb9WP&PcQ{s^8!Dk}|3$i~MSk9Fa}rs_J2&CPWi08pIH&o8?@C~%PrfX&%S
zKX!LF)K-+m$a81HQOO*Hc)J``4wZ_*)CqY*xJRf-^*v#K5xd)u;?aZmV@G`^L^5Z8
z5gouXkLK7IGL+y)yHAhXadglR!XHkQXLN5U4tmY#wp&&hzwySKapUF<UAN;e|MG95
zTAQ`ZBIHosgz@p=j@rDg&#t|ZILqm^*WZjE{P54?Prv(iy!qB!aq|{l7Yp%?Z+tVp
z`OR;|>#x5Nw_kZJ?!0nW-`|XHeDgc;{qO%p{Po}bckzQi`^&ib>RT~cUeuhzTaZwM
zQ^GUC+~FC7(M9+&fr393Uc-0<4rx4vC(s8%<}nBGWV7d+@SE{u&gZx#j$&vbAHTtM
zjHQzk8cX$Yc5cz0Rm?x$@L64BRehs=-be0(rk#s+v}12)+w&Q`tyJsAW7a<o!>u{M
z`}M}GWj$yaGA7>%w@95mm9e(^OfuVk(g@~+)+C&_C2D0lO8$j~l_*!|gf|oR;$rTy
z-Wndqb4q<s|Id&6lEL?dAB0NN+)#5_Z{gJd3|MEdaV>_JBbrmgXK^BWf?;T5bJZd0
zkU{aHz@VS>B^AARf?k{+7`LHyD5EV0i;mGo&X+=YS?t0uR|GRP0OJf7v>wAAoAvP6
z>~t|0^i?!8<1BV}Hsks8XU1E`4I0h5qSe}qhYvo9ovlsJOTr(rhB!Iu$44LhGWL6|
zsMlwr)|eBWKehMG@iFr7l*+n@k3W78@BaFO=<n0+alv{s-hKCjc=mifda6Ga;+*T3
z^mut?#kh;76$ULfYl1uQ2KWbI^)M!)XlG5}NEd1+^cI;NgAm3j_>DJL;5DM>td(q-
znx2jtAw)GU@I@kx5K@pnUkWD;ziP#N27H6Bab^{v1o6Bj>>lG!*}xTAfnm!0OSIDA
zM8!W@Zy#wLVB>&F!k`8}nGxT;b^C@5K6`tNLC5VNeBs?!@2GvkQ}~+7WDN{YrA`Wg
z#$vS83myaN$Kwgbc*EYnC9ms%5k?=@K6pe6d(6YRI46m<Zbsi2j#|BDPb%;Yydvy9
z9$}P?kz#IcK5nes@Y!g<4kPTo+Oxa6t2~E-T_sjlmi39DjtwlrCgB)zKzf1aMoI2t
zeC)J4h9~oo`jH=acYa~c@Zmn5bA*lM9qus}o=586`ucWk=;{#CvDFlvZ;6-UahG#~
z_j6-y$|#<M0v<~@fx*d2eA=k7W$y3}m)Cq+w}EHAc**aujukahQOtX<<9#o`t7hc^
z%DI+y-a#tz`Q^dIPxAO(eNsiPe+Laza(i7?zVG#`aDq<~Iu+o<Gd_#^u7Y<}0P9yS
zZ{}5h{5R;6KgsU4{pN2{Q7&V^T!UX*XE1=7>8!o7G{5wmH8?nfA(Qz@S-=>409P=+
zgM*AY_4%AS&Fez{LYb~t@6j6TIW+FP!*AL{^4aYEhOC$d7TO8@D>mpNyCPr08_+$u
zU%L3Dht$Dm(cgTmdyFk`*FDCc_Fr9U--~-{fnlc0MKJnGE=N>baDpar(f+Kbz!N>!
zJ!k`&U9X>$^xS16oG|*@LsGEIF{BD*AUqUxI&ap)*kt=vlfKi2EEb&|UwDsp4)TI~
z?#}@3Y7=7%d~I;f!mq}qz(2pEi~P_&K6kd8-n0N*Y$R4bm&c3#PakeerlQX;`u%bm
z+*kXkC-qqQj(+<3?0cVI&hJVtHNOthKdqcEFH!%Ob^Vu^KHCSV+O^^9I$fu)VPdye
zZEh~cBnYs8V&^VG93kdlL?i@l7ILrvGjcH%7^BF81P~U3gndZmJ<rF+31Kr~<q<+^
zM+*Xk;_=ZunA)K{B*>%yMh$Q$UF<$porhYq@3x!q;Mrq`H{9(sRmWw9cThe&t`N*&
zs5Uf6xTTFOE~f@cCPEmb@@QC#VRq&YW#~CGL<qB(vlfK30d9tn#?EIvnqbCOkm=$*
zghGe!&_W001pP(OLcvBtXhmShK!mUjd{Dq~#uPmxo;#jr?3(dT8ZELQs|a0JUSWC%
zxB~~ugW=OQht8u9`pp?6c%6|r;{f2%R+l5;jea1|u=B=arm>-?-i@Qa(szs|EX2&5
z)HsYg2vZhxBrtFjg`dY??K+Xr$ZkUf0M2(SO|Uz4Lc-*+7JA({=pV<S7C*gCFZ!MR
zIMp~`o*^XhN0S)WL0x)>MLD-@c%pUsLAwa&M|;mM&$XDB@TlMDdr%P39)8<Er?NZ*
z`lKH&B5P7VDU%|#_*H)_e9?d4H-IV1JR6}jw!k|V8fhftIm$M0iaH1@Nzvq^jm9OF
zjo}p~2k(M&4nv@tI7ec-!a>}^J<S37Sl|Qa>Cq<Rj^-Ni#Stul1<F)HdNSXv{8E`L
zO6ZT3rzowIj|C-yT`IjYALqWYBiG>|b(>v{EJ&^6xlY&VI(=n^ubDr<@C9A^+9vT1
z2%A}qKKZNHr{m9m_@_SK3)*Y`B_ef*F-NqVGf;{hsHh(BRPw?7&{y$G^G(riloTi&
zy&x8iMBxUXB&=0>#$r@~hD)ANdDIV0<RU@ap~E7zR!kI1V#0612hIkH6xx(35OWaG
zV+<o~h~VrGJhTr*Z`CH%BhDb6tK8z3C}Rjw25*3$l7C{HHmW?q1%B&|A0pOx5}5Cf
zs@dmw;;hE@_L_d2S-FJHerV(@4)=GpxZhCQIEP}O)E|LY97+~y+w%b<M7y)+5Mdag
z;_NtDdwY_<w#BQv-h_0C!a}g(oSDYVg80*rctua`8jV^#4MS<%Kn4^)g*hp|ehlsd
zi}N@+XvWcg$HNS)@QR8H@&0~S@)L?rb_nCyru^z0<y*8E0=^g(#Dga}uc@K2sEb#!
z0g2M}CwQenhdOohQu2xTFovd~k&-e=HoZ8E=H9OB;+!VD$?yo#aO*K`AGW6z_=Qq@
zV!ELRX<%f`CMOALgnUP+B=C^N0UN>S9NsD1r7*XHG%%LXPs4y`r}n5jvv(r(CTuU%
zw*6jH<zU3@#?CHANzsOxdEpoay;1Y*6Q8xUuW{?^{R_*+mhm)B7{|sYW>wy#U<W*?
zS2Ef8p@tPtMd5$}28>Rlx$`)lJh>M=;lkKB_fQV9c?z!%%^|f>*e*F}aBgj-5tGyN
z>axarXk7AfFZMNdcoyDz<<<Dk+ux7xeCK;HH@~F!Bxjs5ma5N&Sd7Rh;PHw&2|^Ry
zxN+NI5#Rp7pT{@9^_{r=$}0{z_KH5g{iolF*I$1tZrpk$mRD}Y(v6$3a^sb_bN9_y
ze)XGCnY}Gszyq!rhr~Y&U1^Zy9uU?XKEyb4qIRksHmY**jLlaEvA4S!?dG=R=`)RQ
z%Eowaou2mf8{=g$p1@P@+NdV^dSpWIXH7Gqyl3{7<9!Tz7;P+AIpsJYi;ii$F<eZS
zIU~8QIh7&dI=!}|10ds7@#rjJ*ihZfCFD!N(Q>KeNEEW&PD?bf8&f6DJDb+n*E}x6
z!=?0k(=}eIOD@$KZGy9&MBA*iMB)^lB3lcex~(nIh&9dkt!TFP1fvnrI?kazPB>`Z
zVYm=wb!f)pgMHBwHVUu?3cb_zbBrh$d@x!tZ|Uzjcq$rlxX%EkKw7`qUW8c1NIs?I
z+NnL#TCH8tp`OYTZsGX@O%y#xhJ|*6`wI&<Vp{93bG@@8nzXmqblWG?&NI;}&S)bd
z2ZjgoS1UEk(I0*EVLW;Cz=kin$_4{G-Zs{sXv`0t4_d#qwd-@Z*4DO!%SW-bu@m<`
zeiV-$udCe{s4k-0JrKR;+_}l9RI2gTTW`f};iWfWh-Q_=0~+;)X*6g2^m^cq)*~7(
z<`ZikWOV2YYbMrR@DL2`coMPk0d(}*Sb4xqz84!S7<cm{^PLQBT5QULcX{n5x`RQE
zHn83!JmS81Bshn$lXU}zE6%d2H|jC3wL8Wm)(SQ}*$a$K8}yU)$GO)f7+Ud)S(u-7
zn|3i0DnFrS@7}%Rkcym>wzauwd|_P!%3?&j&|mh6NEV!9RI-7q)76*{=UH#8W!U73
zR~-b8s4U%H$Ln;~`}Cg+10cq~`Gwh7SX|U^jv3&LrQ_I8-Gn<lR-dct(@bMVupf0B
zS}lx!E#vd0co6&<&$Ri4dAALt2cL`yiM2U2ob^2V2SO-1#GS?&*bwFwqhor9X`MTO
zH+%s;$yiv2ff13lZm%Ef8{5%t^*oo56VQFICbK?4uwwn1atXM~=w~t;C5yfoZ~T6)
z&ok1<x$eOe@~8=L4g;Cj&V%>8#-&{FN}tRfzEidJDEdUE@Uy)6;_@9nNm-`v=oY^Q
z)i|$MA74Cop;g9FjhjOaYCc+~8C<z8&T=j14f@BG1?S~{xZM0+^ovxy<4=u4u`!vx
zF`nqBOuH>p=G;TSMBC80fMb2}E(|6aE>E;eZFLXTA2#e_?4(Tk!Mw^bV%?qRo%4)-
z8n%Nrx!)>JKWx-OZ-u@A{givG=2(ZK1PoZ~Gk58Od5G#^Oeu?ta(oTm;nVN?ciziK
z{KsZ7WTiBa^4%NK`Ch=+S;YZVBtrly(0*jDm-Q70X<1E<cUFNhGy9qb)*=sI^;-$K
zm2!X`dK>V9=Y?@Kc@Mv-@5Pn%B>to!Oqa?-pJH6#KP8u?=g7&-C9a}hY4n1|(k|qA
zLS2seIrE!O<G*kKV?Xn=uXX}==n(BDauPMsAIc!PU5xQdzhC}d^u4%eqnE!ISEk~5
zkv{wT_ey%ly2sBX%KAggE8h9%USDSVyz;qz+p|C9`hC;qz57MiXA@H2wc+bJU8k>M
zf{|Z}S&mA`n<%k+7G(@xQmF)jAmD`pBSR2`>A(c}%`Q6_49WzN7&B$-55!b4=<G{q
z<Xwar4g_Y`72ypKoDmXWh6vS#_tu#f^am0|w)eK<;nT-)Usqc~4O_z5-J!-va5?gU
z*BC%pj3BragewK7R9s=~U_pUfJ3=kWuCjz8-m%e4?e)$^wSk38uiv*2%g#AMGEVAK
zo%9Yao{vi)CeAsYY%C1Ub>F*=P6!u-!c9V%`fY(iaA2n&Frl9q;EMJk#F)VgeuZ$t
zvxN5vl8A*47A?6eSbY-<2j02E1;vkq1@*%M140q+SQy37_Ch$|p3i?!AFt~5*g@nG
zIL|G-C&SNsgkPaw{2JB(SrlNvb+IFjy3fxpwP5c>r`hz*Nxt`5`*Cuhh4o<P7*!O-
z(gW8=TeZ0GE-=O$!5_E@K84`vwkSJ!5%j#U*FE6sU9Rdsr6D2}JAW|BVx(d*$oNq{
zFmzj0K4nus<rv;BK`EPM{9qglLBFWq<D&Oz2a7W=NVO=VZ|neMOu%j41E%b9tkf%m
zvcU!7g%3O+X{X90Z!UyU?4(d!;6*NAvTc-Agc%`EwoNz*Obyq>9pC_Ln6t*eaptZT
z>1rp2!|QaN{wq&knc=HZtN+W%$)XW2(#zj}&*b+9?+s{BCcc06`t;YUPo;cS3r2sy
z%ZB3i?WOqUTX*BlH*UpjqhjR}n{41uoPPr!gZH9Xm=OJj*0Df_Ph;f5sLy5v_!}|?
z$}D?$iN2$xCBJ1Bl~X7i;1hYS7m6&Djf9*+R>12^JYMfXV^Nr(RC6e&QkI46!29q=
zXm-MdMQ~UNf^q=`9DJ7$h*r!<qJU>(FhXgHf(1nj_23~*pUq2@4+Gl7RM~wwIFNjE
zb`+<FJsYWzZ)k%<K1qHW8N%b5^ON8+qj8{n$WdyX@8#kg8KxiGTU%BJ(GC=()e5_x
zQ7j*sSC*$M8Ir41jv=*qzuSy{za?HaWiO55+#M1|nWO!#cpe*ekf~J3kYIdz5IZ|-
z;d9V5X2WX3u%2HiA7^PLUI=#ZMb$Apis#h0*cDz9;!lDPV_=!l#tvj0aj~<$Hy&Ae
zqJHN6(>U30MX%TNcvBaKDbB<~xr;#uFBigdqiE*5)}ayj=2(mlOE$IVoM1hM5>(4r
z8<B)-r+vXTmD+@%%X6+y?KOj;1TQJz>)ThxFce<!Ym9p0$>(f@!Z4e-%e`U2h<4CN
zl=!`FPp@4Jj0h-sB!?+K9xu<IKhe1Et9>JuD={EBl%?<(FQ3tI!V9wWMxX}vlnp#l
z`U(FI_M_X{j)UG#JiPxw>`G?Ft88ImS>guXC}|)CFM(B?FyW`1L6@F$F~qrFz!-Vx
zMB_!M&{2ClP1guFS@#$r4{~k~iag7ox~Hk3w*))o$4hGE#_gy#=50{pY%Y}9go3+y
z`_-6RScwv6c1;TxRVE<|$4kPsag9CwavLy6ArH}C<{J0-&U1dBs2%Ot*;*BTV^m_3
z53&-{4+)v~NOCB~8_svd<8#7oM(K^02qCaWebyQEA`?>wWm5<88=ioKGTX8@7tb(|
z5UL%VadzVmaw;2=M01eSg<InoZZ$5<$ClvC8lzfEI4~CcX060rvJ5O7U=AY-vsOu^
z^8cT`KW)-1$?gL`kI2~fB_ndHwHGu20>f>FV=@kBB$L^gOnM<RnMV3L^F{Px#`HpZ
zIU|#eI3x%XBmn{qpwZ~Ps%y!r+%xt)GIG)Xcl^1T*@fyxqc?Un!Xs}!&vUO|UydI?
zs~`8tiZZJ0CtpU_rH{^Rt7djwu(c+66rD3>B*5pQIFFA56X)wKwX0V*GahS2gG9CV
z#a;}D(S0&kwjOU4twewI%K|^qpcT!4cA}BV&%oTBy=3=JKWCihcZCxfE{?aO%*}@n
z?ibxs7P@V*f^j%ao6Iw^GUrG|0NI8!bRFHC9qC`-G3YD|m*AOZlrfVd<kL^@WL&n|
zY4i){EYUf>^PTUsy}g6Lc0A*L)IRv|c6;x=58LrE8h@p>zCyaEQ#YrGMUS?~oK5@T
z2WO0kBEK1T4-WfWH7>QCZD5B!g9FEr=t0lv7Ub>4#rg0Icwp=@Z_wM(qE2*Q<kqS!
z_~?w8F!+c?7Mx?YA+#MJcoxJ$J2+jy%zfMEnRDe-2_Vo-k$a5qt1Gq;T?=i`JY5hm
zwpic6ZExowaC#h=cU^{_p{?4%;U#KIA4HAWdQ#MjtsyrzH)>HM{Fg+Y)f~)#(z)O|
z1NvZ$cbtJ}F<bqWqbBz_a(1?N+Jgs=$|;8C(DAgvg|1(_zS(ZwycwFY+OA)}R($91
z=%ga2?q!@GK6qG*EW19_opz$D?%%&3-uO5;$~kE~L+1lGs-gD59dqmOaKH58>5N(F
zUTG(X*W*Wz+vBaR!aH;E<TP|QIKRHWR`_=meodEJSy^sZa-XviF6+0k*xTDr8y*(E
ziFPS{9(am=bsgC?;D<3Q9n(#p^{ngLUC*bxcYVCFjPjBvel(4q79G!DvJHz?U)N*n
zd0q9himR_~m`JWs>TT|+q_lc8yg#}ESoj0CJ8i;q@HKMLsP5z&93AF7ey^;ONm8Gy
zmpxR`<&>9gOY&Y{?Ai0Dv{4&+J?iiB7I~r5B#!ZwWeW`!La&|_j|J1(>XxG_Wume0
zAVn7sfQ#lN{h9N$N9BQk*YWkW+LJcuo8NHCLS{5k<d->S%;BWuy}VF;t2RT%YJ1?3
z`{XTkD5q>M11!9g@4cU+{K{~z=SY7-)8R($$uGTx+3-8>6@RbK-sW037MRtz6fEU4
zR<8ce@Lf6eTUj-ZeN6FEsdz(vi|8a%8Yit7>eO>xwk<ZiDBO5i#?Z$4sJz~-8~Lhl
zeLk>zu#=f<-NEVt=mBnIYl|m*M=y&<DJ^An-WeGs^Bm1lo^s%es`M@`?)8q|c|Cph
z_p8*$(C25<=e_dTdjl+otKW0Ix;MJ}^7m)ckGubQ>8H*2>yUm?Z5`nAH(ftZLKQ9@
zUzh1J{RSosMHCUI*JOn2GNwlm$9T%H5ymD8r3*e+qGXAh2~*7P!J)*sgUA#k3gcl6
z*grfdMzkCuWhDyTS%mEp$WwklA7ZM?AnYGS5Osdny@wCmy+`-kC-?8RPaoWCyAiHA
zG$JmRa8qqfyErNEB?<*5|AeuN5QAamk5h`#%)3t_yx3v`lTIV6&-zG^Vl<+#ab^&T
z6<P!{e*}8<iJr2*wtiuT6gkXSvdWvG!+Yg`3IoQ_JJTc>6x&T7yV3}zrI18eFCi`Y
zz>u&=@Tqp`Q(#^zT2ilklqs+WE5e%8^QXUoF{MLA#>E*YCA=kn33TbQe!6ExEBF(v
zYE_c*2&_rUk|;5C-bVV(Gx=QfmON27NgP)H+8*Q}%78uD&!VWE9;I*R5nu*K%OGoX
zS*hMD!PgH$fl*Rc$^|R^EJxJ{j@nDHry!JfR33ebU`L>=h_tk^9AAuS=_feU`++fq
zkq~D5I6{rJ#4tcXEaPd$8-Bv;?wsvTDNWUezRi_1j3b<2Xmh5VRbWxt^XENq*EOrt
zD8SmTEDEGP40D6*Sf44il&}%b`sW_ORa@0JD4g1AY((nGTPwfvSNKu&=H>ALK3Ccl
z2ln(Nb4cM<KP#;|a+xmE-^KJRjQILb=lFUx#fNWyyGf6xe0~ajd3lWz=2OHSzjXQ9
zpI12{bZBcxQhW2wYwf@Lzy2TF|LOnm-?XKLnf5&VR$2HYW1h&Smn+n#rz+ys+6l^I
zZdP=+2(dEQhvqW8<AJu3LWe6NIt}oA;KgNNDtbPdGDQ=4=S6Mx^+cpr>a2+Fyi3tW
z??q|h1?c(}(Pzbv#>&uMXJI|FrBRP;XGlg{y;p|M<W-mQ$Tr6*i*a*$vLA@f`Nq&Y
zH8U4EYq2dZt>;>8$49&E&L{7+ljDOnHzQKGM_D|1@~qvt^J&}P-7Y>boo9=S&M{dE
zf9t$U`#obM_Gx&)1m9nnPd$0ZwrR%d(AHvy`voJ$Z1CaK`sDTxn=R`Y4;b-n(Q7M1
zqj+}wsO{`LYzKMQR*4*#coQew>G?^?EXJ8Z-}p~VrY-Y}-LWw}7k-)XPyVq9#;C!O
zmAf29cyeA6omKarAGNci#~Fje8ozPd8@}L(I&O#ikK5kv!*+DE7d#PZ#4$M878jPw
z(R7q|oS(z7QH}=D8yTy;?R$03mPAgg2)6W3B$h4G^dp_9=$Y|}$c*LmaxM-zf_#+c
zFV3rv6=gkw%SHMw+KJO;JaBXNPcSZZI`bC)Ud%Wzw^@d{^Ct!G`T2fMr!DQy)2F+4
zZ?_+O@7saXmA10FoN<}WnDnR;aH(w|ZRJ;yPm;Ms4TtN2XF&Gg(+}If`PcuVef+@>
z%F%k`#+&VT{{BBsdDk*t8JoyfON%R!zpn<T=WFqncEeTp(if}s@@+Rc8@XjVvQz}T
zbeW_fLW<0hd6RaR110jv^z1^~0KXP0f@C4G{nGMAn-4y7=9+KD-*$x+K{!62an|OP
zUE4u2s;f7`VsN2loRe{M4~|HKAc@8}m+G(`9oXJXBxT?p7?peq$2exj%-8w0u)N+@
zR;~sQoOu;mQF3M9%CtM%Q^#maO3tkIBK1Swq-SZ%`IGMSG-hQ`iyUc8=oS?jRcE=4
z+AoTWtQL5pRc9wBMMFf{O^~HIGdRbeGS>AiujnD<LEF$h3vHR5cV3jJlcnHv<Wjhn
z`6?2#JAWgiKDB)z=c_HOLnqTFJ4T4kD+gLR$HxNy^|rMBR$E$qqb)3D%+V8goASty
z6Un0=wRj)|w#GH@(v@l(&Z}2LQzpvTi{^QUGxo}aC`Ahu1F`glj9I9|S?N5%=N7x!
zrf(|mp0wMaeBAEcyI=EEWFm)a(ZXrxA%<>8*Q@Qb@|-SZd%%Z}9=F@KKW!g<{7L&X
zpSHN?M0pr}@S_htY2W$o`|YET?q{w)jm~jYPKZxFNgWq<@30ua_NM4@Fer_)6=vsV
zjOTR3$j14E`x1vY{palLqCI-N)$XRfN5OGXdSrGozAaQ27w6j5tE=tpw{NzMYqrDV
z{IVV7bi4J&wQ?G94A61ke*3MqzP{R@hZcz(!&?|+IZnVDs@39%i~+~AMlPfdb=n<e
zclR)K`Z#z5=SlE**@)NHyV|szYpM8<sOUO3GS9ZR_5%C6>CblB@4UK+S_rtbywo-~
zH{10a*E2rnb(Wf)I^Y3^7k!YU&lc(Mm{a*Q`D`6l5yHU(I#eyR4E`#c-oojo-E=N9
z2)`A1!C~jDVfqcdXnrpIH+YEmyz|bR?bfa9^`3P1?)~;xfAzg~|K7vGH&NGMTsvlD
zZdv%@oNi^~cN<sN+x6?{S>|5my@)Mmf6>+P8}q~h7=56zkq<br4v&sX|DH0hLvxLf
zGxos9_L$1I!wHzG2adaU6uh<7qe!-+iWbXU%CoMcA!$D0VE!_{Ye190BVeh^udbuh
zDUGNHfmdMYxw@)MWd%;)P=8ein%VQbewB>^GO4In@$xdSds2_@)w<zM((sZrsw-tF
zmn_rocAjYry4@$gHKm;3_#+*iGbb>21y<-=)z&-R;gRofN<aEMOkKT8m9KR6{8_|Y
zkvi44k<%dXs4*PTM2nQlu^ik<)isBW37nIk;}&1w&@bCZF7P(5;K9<uVxMQhPxFsG
zr<_wMhb+h8ZBFVhXLZr(yeq+74khKNYZx#6P=<SOot!7qkfRV7)L&NtrvENn^RRf`
z2nK%VDwyO_X3qoWeSB)v^7$k8a+Sl$*oE%B4!;g-pYUn~)8T$T6=+;a4y(Uj{BVJv
zCZJ#JW9ZsYMsF+m9xUKT#}oLgoIEYO(3bi&>dQ+VVf0gShuq7NNmf>dMQZBpi^r8=
z0lJ7xbEwlq_yOM9d34~sKWDa%>{}x_O8<uKLq}CtdfA?zIsNGSK=LCAoT_K9(qGH-
zBWVQJ&q)K!f6{fCemRN6ed+kROqc05F@=U>8siK)5ez5@1YC^27;p$HrWQ7m`vkwf
z@`f;r{97$Zsj)Rk7wWxlMPpxOv%C<+KIJ}*AU_;(iwRLM504LPE0{YEIJ@q(yN@2U
z2V0NY!`-bi)SO3Q)CNLgF}}2m@Sp&KhSfln^I8d#_Dc8w2Nb1HUd|@%#1Khc7@~wZ
z5^N|r<;Y4s+Q^AScq(U77^ZqKW91Q|3FFFe=&)6k3`P!8SJuFADOiLJ%wPSY-zYu=
zf^Ni{snvGc)wf{k0!&{m+(lQsxnMmg?71P*jo{RIDxsqb!SyR2lo@r`_iBSkvMzW~
zPY5Ce79i{bYZux|m`+`FzDmiqDFB#r5jHN)FCrBBwEEn38Yk`Wz}6?WUg{A~S8NHB
zIwEFNd#a7qkyqE)mEPtW3mtS%o3tVK$~cg5mfhIB;I#1xXOiD@;{bmYub*JnonKb?
zU6ixa`@IUi4vxa#vA_o2MkwvuzGSTORNHM|7-&@6GzxC2i*P&@fwIOX1%Vesfib9Y
z2g6+O+_)JhQ3G%l-WX#BFXLj|OL)wE&PF%^N3`1*XbYpM_7V0(K9rCboMNcbXHH+8
zD`7V8={JaX0Vp`0_FC=8&_nsRN*_F*a7qHoWx7m%_tF=`@l`bX7nOe2l1A_S$m=Mm
z>$Nl{!-&q0Qqds!Dr<P@(<klPwX5yF`|tkG?Z5fw|Fqq@c{PfJ$SYen;bSjqJ0pug
z@R+L9>y+oBi+CYNOJsmN!_PPb7)U(BpYdb#8$IlcZ-%IJtGdzG(9_UKj*_DDq3evg
z=&!up*%LaB?>rOr5}u%byrQ0QM4`Kkc}I<74lVYM2NZMf_7%~L@A;hUkz#wh+ZFwR
zhoHr_|C)^~w6t=y%`a`XnYpW>@z2}g{-f&Sty^!mC0l$gueYV->(NRt+VaY3TU_c9
zI3jgsXQ#^9#7QNRVlsULdrmK_)BCn|foA$Afd%>K^yH+BK>G*%tQcoXfd~HKpnk?d
za!|y-GO(YV?3crAcW0{%`W%vTvkP@D*Zll)Iq&A@r`tqu!PawAvrBD^k$lYdfhpuU
zL;ZQ@sU;g=$WPk&$w7^Qb5NNKMMph7KWO`r`3{fvi(lfgW$aJeIVfIaY@A<-pK~BC
zh$>oej!-`{h>X~sM=t`)wEg&?Z9TqUQ3+wVk;!0Q_#`Jw$}IUSFrJy4&$x~FZ$3Hi
zCK>3{hNF~`MkKE&f9E`j6bgoqbEuLfIFA-{IVURPp0Tnm;n?^n8s^Geo1dM?7+uNp
zD+O~#=Z`=BsD1R|2W@+2J1|@<*_gr2nMZhamnE!zPP>BFJ%NKcM>YsTzBq4x`KN!_
z{-6Kj|D)~fJZiuF_kXYb-aq&!?b`KQdEU<+TAZJ2o153${PJ4i?08_s5m64>$dsbS
zhAoqo%b_+dO77yM<cKRX>GSwPMFvhgTPtNuGfZ6(6$t;QB7bz_A}8I{>{{SwRs3|p
zyth~TIg5<HG|-In=h<<7J}H=X{BUUtT99$7OMZ?U*~na1%CKk`1GoT(y<zTgu#JzC
zd2GcO_~$*2J{Ow62nN?{i^!@gWsiZYHU>um7Y;kdGcw%#0w*44XQxdiPdjt$<@T4U
zxtwkmhdrrnMJ@R2TL+qtC+$cC%Eej9h0gn$hjTN|uZq4E9OA^_%wSM;4xG8^oGY>^
z=h?)Rvbs*CAEJe{@yZl9_b8fv7GK&w!y%dbqS8hYZ4;ppwkj9(Mn6K^oOidne!Z>O
zc_X-I;eaSJIO6QFwe_pzI42|9qH$(!F@1&Cp^@iDp}8CY{mf?DgrY-}d8R3%cR&5~
z)8P2+nx|`_T_TXa^WE=9&)W^e`dP`gzkl!j584mk|FC`d;qCU_@BX0u$3OXt_HX~~
ze`tUFAOE!d)i=N0e)z)=+b5sgZ4VxX{+xpAe9D}w-9qSmCnp>*;}zMqwr2Z2(PLz5
zG{ScGc7EvkoW6r;E%u`u(sLrin{y()=ndu_=ifEXp!9`fXKi&Q^LIKlc(J~t^I+)N
zt(!NB-uDHKk<q88Q^%9`=+WaEf6f@&{v951WQF#I=P|I;kKht`U%P(2UAu8D==D7C
z*(vAp#RZ+B%k;*#M}SV%d3m-pMC0)lxPN#gf8RFo;J`M6wsK4znR=x^WxiLG)h%ZW
zMlT7U@xJI;5oxaNjKxPEeN@wRWhFc&{Sf8x@ZsZ>B?65zm?Mqu$N+62-+18)I%qsa
z-E0N_IU!2FOWry|J9BBJBGMM>H^-oUiNbpD;6eNLx4+x&-FuMdOM&xD;TC$%0lS)U
z#W%sq_}YqbeJ%WR$s&+k3kywDvxO=a8_vyTZU&B=8ysjN=pNjEkbXR>@n2h8skz4)
zu(7e(-gx7US}0-c$O>ed;_IOY9Dmz8I~hxhKX%hDk;t5P7MgVZx^%bP^XYfdmXzIp
zqdo=FO6(SF3TRg8%I0hGRT+3muKGlG^NcP&<QwW3t{$D!uS%(;c$+rWt!hlowf-dE
zD}Ub2ceqvWj2h>Dy-?rNi{Ooi{T*Lr?sT|c=6U8-WF>Q|<`O;C(<*o5Xs93l^|Gpe
z8P}>wJ)Y+W+9jgL-0f{1m63AcFd2nRKo;bXLaX3m&8^H|^_PP%3ErXg9HMwUha8-&
zoqU4Vc!#omS8vsk-)cpJ^+}t{23GXqqIf}<QM(*vEc>|O8GS)?jhoN>)>rV6$dCH1
zU;5&bdN1?Y2P~<NleTrZ4EK|?FTcw{HM-axD&Oc?!R~b&$-M!`UcECaFRAhs%tqx%
zRi*kP8~cO9dDeGIARG5Zx@AA}FPAxC_k@yvi?0s%`#gAAwioJqO7Lt{hPr$KKlD(Y
zC96h;Lw78YKu7h(7<3rF+V&8=7`8EWZ*gHUJqT`Q-1<Uc_NzX|LwePwzn(lBE(zRz
zn)K7;`^kCwvtg`$+IoMH>Ge8aZ^zGa{W8+$)$_Nx{-*SK@4)Cw$Jb@LOuwlqg#LLH
z_qnBowh&<$8y<*%Z50=0gOMd;DV>-(LBv7GRwl}eDG>yUDJLgGKc&jbAp2t<9v;;5
z*(g<|{OoMCPw(ETEe$`p_i5WZ*l$l>JZq;B<j<a*)#q`PHcXNL5ecUo{!=<3nb51<
zwniYJl#oQYNInX(ZCKpHNC}VL!5j&owhQuJIWh8E$h@dGMk2BaGC7mBW6}(D6*-o?
z977BeeN|W&ZndG5p9p7G1~Env)+tuD?s1<XjCch`A*m&7*9t^}QtAL-F!zo&Yj1Zf
z_J}Y-mohv9>*{OrYAK<Cl1TxmmFQ{fu=3k8!`R{QBz-<EVb$1`5|{dGD}l6y5Xb;o
zPNsanNT2H7MaJahxE&oF)EE&qYkS8thN<VNd$P8#bhcFq#Al3wskEcdUrPg10S^Lx
z|4nHa;Z6xd1iCx}<8(lqM4A~ZN!z>P-ElIaR^bRo5N9FSrN9#Ij0GoBZKoJv2tF~6
z`K*<p`E+I%I2ltb(*`&%MqxA7QVEM0EA5sTS84@P#(_ZI-=)xwqP)@u@50TBEJzz{
z86c0m>X?fnWsEp>m9G!g#*yG?MW^vvm|G~}jS$*}>^`>2VuJ>|*9BcpS;`-yntR$q
zfUI%3Oqc2JR{A>;@%4*K=-cax!e@UcmE)_vymoct19DtO1dgxGmG-~=-~O}q&bx1h
zwuSd43&R0ggyy1=w%;oQLTF@&T+zi^9h}Uig69+eQSflq7n+9uQI5$K=r3nKqbI`$
z+Q^_O3WGsY@(#9Oamq#LsJ2v;SQ+U#v7+FI-gds)`=`DtH}v`R<ScZT^G8$w9^Yv@
z{$9MN%E3eW7K;|t*uIJLh2f9EW@>7_%`dK{99zvzwaM`9g?Vz)Y+G7dYZDwXlk>Sh
z*T!F*r&Sllqn!UFLSB@~m9gjLBsw_UNt?Ad@<H0p*u*$sH9H<EqRClBjCE`4H_F(}
zNVB`MQ+#w~WwouZZ<bRH>}`iw(ad?59K;MoHgL|`(aAw%hKaVgywW3i#uw91(I6R<
zGX7`0GG^Kr2n2j*?fCF<bH-Ig%h=v<{8@YU<glF_i`+VByStCeC~uUV*+Uj9*{0;j
zj6rY=4po#&+6})%x-2cN)VMk8>Gb%hWSvMs!JR8@Kd^ST7blsBq>art+UDlXHWAtA
zD0spl#z{A`xDi}%e#}D2uc<7>lP4lmj;CLf;iHj7$+?pp3c1&vPH;H9xlqP$j*#Jg
zHyBkEOw9$qkJE?Ka`gT1{g2w8ef=-m_T#<Ghp)7M^bh~@_SRcp&iIWrTWwt=4}+YQ
z&k?%<cLqDM?S61HxTi=+oAmU2t9|nuUvJ<4{<kX%;vf8@f70H3=iS<(`&n?|EOH*l
z7F-D0rwz`y>jrFFb%`z`W{Kjn7^$~6bL#0q@a8c6i;Pld$m-kEiiWeWhfJx9eapZ8
zPDMY|)_jA5D+O1i99ivkczBWge%4;3Pv^(GnXf&HvJ8NuzNzm;-hc5tbF}IL<H$*w
z7jW5S3^}5mZO5Q*JInsQ_NIR4s~zpOo#2>x!r%?7EKU>2z=>OUT23tQ<sBHGx~@!E
zvClYX47wAQtO*y;DY)L>)jv*VE8)%i;3qkgOr18`u97TfF4<0xqu5rDoM}h<wupOD
z+qThF%+<ShK52J8{UGx~WZQXd-L<gDxhrau(XU6Ex_YE!k8r(`dG+GiMaDUe7#z$^
zH0)tH<jk45xs~qd9Yy3_2_2f4S{ymGdOxkYpP5~1Gs|lQQ#1m+Y;!s_mAQCv)*e56
z(C*y%sAz6qTxD*k(RN~|ZFg_Cee}`owzq#=TVOh`>@WV}TkZe#|NfuapML#YnFkl;
z2>QW$AGAOEqrYh1_{R6!$DcgNygd$_!UHnK&P1D>oJ;?gG8Z_f7xKB3-wUZzq|Vhc
zY}27oX?n+OXfmB{cW<v~vN13h@gX{iPqYO(k=x+|=NMjq-r-HQ0(|GKx9i*}eXj*)
znUCaq4k-&4OP>hr7Z*7A$1Cy-pK(ug(EWS&!;^aCl<gDu_idku=X5;|A2A>4FY_7S
zTeoi3wuOxN98h*HQ765Fvu$N{G0$yXxz?^-+sGI^4gBvG{4Hw3C+QpX2TnZESR$VE
zZ)SF?qDjB}<#(c=Ek*ycMeXA{$19!g{5BBT-7!yg%8`Uu?&Ud$S9dN(Ukp95C8R|v
z7B|vIIN7#$_S?OC58I<h7I;K2v~VIYL$~3$NIA|bkviswC?Pmw9Bf_q;Rhd;(Ry`d
zC3Nu3s@ry*bSizSGw-ImW7c+#7MWV~XgtBt*xAhjzsDoNU@zlD&($xHXZnSX;(PRe
zI@C<?-`Qg5Ed5sG5IN%L@VFxLc6Rp4kj_C^XNd-P$PT5OW$u+e3YXHUx-#!PSEujp
z`CiCL7D`@l(LOkdm*HRfEAQ)$t^RJ_9+f@Zmuh~y7u*{1RqCROM|J&3s(Wfx-srt*
zzt{4YzE!r-i`t43*7Mx!Ge*u(NH<dkIiO^);+uJzGI}vX<^7%dRF2nDR`nx4)WPmi
zi-t1~z`{a|-UbeuE&~o%-Oj7};EVC?PA|C0nGZ&tUfU^06f%1RdTl|GF;u2RPgV?F
zTzzCF<qj|<dl)DAUCM@w#*AE8i%2t%i#}A_x=p7qjLq1lT)aX(%B^yePbaE>ey1J2
zD;xdRtD$!<^YyR&?&x0S^=K%gzJe7RCo;L)P6`gmJ6z<^N(F*ZS-seL--EhW&*YJO
z=|9T3sSonGoB!Zaw>fltu1Ly^Wyk+sXE`}*T>3pVm28)?I$V0)lCr$-KKL4APA`#`
zDy$t7a_J|yCu7n1I)CppJSunrk7^6l<Oc`*xO9!+NG-OnniTi4lRq(aIDKB;;a)+b
zuAcApOBz0V>P5Yt=k@2WfA8t1ZRal{4R!taKL29VP|x4y`gzmmmHGPhdCAPVbbMW=
z%k-O?GVf<r7TS6Q!lekJgq(7;5Cp<dF(nM0Gl^i1*$o6k6qtn3F*uxuA-PYuDF#(S
zf|X8D93!Zg5h+Ukv6YQh^+u?9UdEGT4<lvB;286}6QnK(F@OxrJU=4P*xur-oEV%^
zgc9|smodr;I|^m3$d0gv3H#0oPz?XYSP9QomoOYnItwDqbtVF?mEe@&Vf9H5`|l13
zZDAl&FQG~cODH5nrEOrOt!K$&)uw(Eo`?477{=_Yep(p{#-+qkp7Tl2C?{AlX{+yy
zg@gj|uxfYM$Go5)rJ{_5xlf>wuLKMZqr6)%ODbV4e<dvCKJap)5FC1B*ONNaf%EG8
z)GG1gb{vIjIJ1PHsjV}q+bZX|S{1og3WOEqPt(S7o>f0TKY7$2cfqpiCnz$&<Uafm
zc~{Enu+3r~fV1|>$L?A%PrGNLK-#<CGiM8Uukz*{E5qQNbL-TlZ1@OY{I1n|gtN+>
zcXJs(V{Hp0&Q<kWHELzOf@`c&zBvG%ge&ty+rfmOT;)gD<ACbhQbjmUnQ($KM_2}b
z3Suc=fh{<|A?20em^?4EE#<-u_tWSyZh{DfskZb0kD4dxL-O?T9_xKe8*F{JxWt&p
z;dh=fV%WV*m+9|J`qeqUmY08KIT<zk^+=tTyh@+_o%Hf2Zi0IDPeiysLVKUIh53p0
z_S?7GTW{SAEuAfzMW(O-fUHnaMDq)IhqI_hOj)hjc|CcdJG$`Bx}TCMxM-=m-9J7)
zD%qn}j4CPlMNwGQesY>;DO29!HKB_ZS@<iS68gyT<r!JPJ0jKb34P(rD#L!-?Umt8
zPA|r7dC=P-mC|2IeBW}#88@Oh&|?lh2I#5jg);iQbRLXdbb}whmBLF_Nx8>|+wJg>
zA-D4@yn&$se|9#MXeg`wZ2?H8Sz20+{C1<Qtc@Zz&QHQyE}E_G=2mWoPp-Gc`M@&$
z!}HeGR@>6D?dhC(GufP#B}$D9!nr51g7N*yh3K!tGV0jY&PwXu_B83@xSgGJ=M%V%
zjagLFY8%D9y+@G=_Iuyb=BX>s+w<^IQCiQRp0(}mM<t^$X5&?ilU+U(dCkx`S!Hr?
z;H7i(OR0ag&846C;qKN}$(r-?i|ON9o-YUIoC9%K#?Y(lH`~VbciZ&bwUp(YtAoJf
ztTFyi&#b2%48P=}&ewaQ>Kbe3dmQDHvo$=n<iT#}<J94h;@BA*=QwdcGVasUGKg7x
zQ`w%Mw5<oX+xG5OIo&?`@OHa>`%YV3z1sfSKl^9x_kQmmw#{pARR1#^kvW1Rkz*M8
z3&#Ts&V>8-Kgl>V?o^?XMFYE(`{4)g6+W!3ueHsaw<5Q%1%_AJ;lX}6C`6dS&7))6
zr|p+BtS<_i4jwMmHhbg2Q}bbP;=s9!Cnq(=M2hFnBjaWa?|%Am+R$aDzI8O24NUS`
zjxA-7N&^4m$L)Auq@#HiSdQ|^^Ze;i<jS44_3+d7=_enjjr(;!m|b|MeDD-;$-&f}
z(gW<>OI^_wEGQec;B;o4bGYj4wJX{|UpQ?~qhs_)728JlGeY+EQ(v;!QmKs(PD^#_
zzjyoiKMQ`Ew+|oQZ`)gsN`6a+OHb%gbIArk^=@cL`ZyW6PgF}$z_gnT2Y<=XXw%&6
zO2%|5uRm>{-u^gk*{l4VyC)f|o$bdNyY0w%&cB+exhb--N7xys$oms70^i{AbF@?2
z$1>OU+Ae2r>LQ0a1CHbCVE19^N9N4b>_%wLO61e@*%qRrEkkEGc~6h`+oOl~A{Sen
zN5>JJ$T7XsKK=A|=J8hAF`Hg;jEcnU(LeBqGr>H$cjwdggZJNSfBBbx)&Ale-)bLx
z_(^;Gc(2`m@HqFr-Tv$^zSVYi4y(;**_}`C2ezU+L-%Jm_*UE1jqB}N=EM3`(P-=S
zyXX_yJ+e1P45yB0EA)Yrn{(s&bBhmp)Dh>Cg_-Df>5Qc#p!pW1MXpc%E)HQ1Ged<Z
zfSaue*FzJ{VY*D6`;;-pmujoey}fq0f7Bf!p~=pBvPB{TEty|^+Jkm`M-(UB5q^LZ
z=ZwfL+i03|#&~n{YR&Z?8D<;LzWr~H##+eL&+uAZT`4?Q{`U4x`f%mt{chyXygzhy
z*h%1I+h2T>J~<X#x7&*OvR95K(M$9syw2Ho=!80zLx(&)bH=l|-14kPs~LlZ@Gs6U
z`0@BL+{hSae2>y5G;Dc!spgq_MOVH1?%QoS{Wwg$A~WU#^K#OKesCnJmo7BF(D@<S
zgsybGs_Sr^Ugg+Mf3?3Happ_LxuU(3*Lc&(c6N88UtLt0oK(icg$|X&DtYj1=ijZb
zixA}4J1x8s;RP<>r@vj#_TI>ekqbXpp7xg%kbg8re7SzURBYkft4k__f1Z`&G;M|-
zec@(*E_K@Xqp98>vMA4A-XTx;Q)X`qS!T$n4N3Xg7ZqgY=F^-h2gJ*^qVF9>l5)V^
z+k*z}QJ+IT<?r9o4$g?Olc4dV{4ReA(Vp&5a^-Grdmk)n9tR&dTb-T7o>2=kf21Gk
z>3XdBQ}(5z`#sW&%u_NQeLXnms<{(9G=EZia0RXIG93IWIxvzQdt203%`I7^;F0!q
z9a`D?W?RL`1DP-IOzyHb6^`^aR#{|L<!4-hpr`(&#t8rmCLJ!yFWEA`JA86~G*;d+
z&K*Bq7M7Inc|Lqsrn)(Hv>9&d=a94w{+4479>4bPFfJWO`u*OH*T<~eyYdnEvZIU~
zdx2pulYL8{J8d4KtlaPC=uTIDHaUF*TlS7_U+l03Ph-gTrXu5tZ`xIeqs!uYw6nvx
zk5NS%W(>-<m-&P?DFdIbV?lDym>M&SEeifRI7+YJ@T2MV`$PYKnk%_K=lap~*YgeO
zv%g=Z{?6Y%z23H;=K4ja&#UA0>x+`i{!7Q#Wx7nip=oMyy50QJ+wJC?Z$`jg3Udhq
z4YLRvE{0W3FN`cgCI*F}F-Vm_m-j{3U`!lBoCKI6VG}d%!pC^tqregxo<tY6Wyxfe
z{OPt30pAvM&!P-ew9>h)8lr-QiJ_6o7?`>5R=t(b9syO<mM!BjOHLj{wHRaxD*fE9
z5)RVNip~k6?5ksxGn{YrqD_V|TbWkQEgecZm5-vAcPQ3Yn-O-DOE9+gc)c4qAmxM(
zE0?uF5!xh`AXQ~q@#T%dk=0kWP>`OFMQF;?$Q%CexzcW5!Jsp4u(rZxvV@9q{EP%i
z`3a-)lu)MM<&f-*hHyZsARP2FJ_3&m#=UyA#;8Ktqf9)B;$y`rqsz&$Rp5QADZ+X=
z250)JWQt5}JrJRkVk18<=Cdc@tnMF^^d)UpCZ&_HOCR0WlZb1*PZ^uOmw0Z;5GAg|
zmT@!qmOgPP!bw8-#b~_1B&Akawelx*z}<2P=J|QX2o5QWU<)^`Sn}Np!l14<E5DU!
z-5C0$`Y+X)S!qizvkPI)fU9|uHW|a0`}t>{fNKfgnIjbySc-X{8$C*)<04~>aS?e~
zI1qUFq__~O4L4j<FZk5TVoG1$8E~Zxr1@T}(7-77YE@`(d&X8E=BRmb{z^E$Oqc05
zJpIZXU#rW1S0cXr*J(UDTEwnyhPRO&o?o=3rJ43yzx7u8+HbwnR#vSh?}{+x7|pbe
zF&c<Bp=q^cO1`6YXeQ-aloa}hj(6u*j{@l{Yp04Px{s!np|5xe8KBE877LI^7y{Ax
zPV@0V=lArji`7{M>(FGYzd3QV53fVN(QR^pcNzV)xi37i?NL{<OTGy0t>~uE<*CW(
z@Wp<PCnM?3{%$#G=H?dCZuK)3;Dd~hhh>l#b>WO0JjME0(H74mcT`lxSYTH4KmAC*
zML#U9Y?Nb4RP<!pvaqz;CTCXC-lgVjFHR{@OY@7%k*~Uwj$`WRaJRN=)@Ng4*ueZ`
zwCcV3l>U{(5gy#PaJERpS!1&$*Kr^{f9kBPoyK{=`L=%~k}R@c-ZLi8pZAr~<TN~v
zQM6=`@W+d%&X76Cy@j^CB=X8wMP^B3L~6|~+U9C0<v63m`C?OTdF5)GUD^yjh?ZmQ
z?9tEU)y0*~HoLIV=4LE9I#1siRM!HJ<mEUiXVr1x$jtOyU>{yyPOFZy?ZrtxBfm5_
zw%;Dzzug`^_#}LN6nz+J0_;Sxou;482&=6cBeRP1dj9ON9qw<}*>0lS(ya9VtUb7U
zx11;N7Cx;mueNvJeYbrnmv;BaAQ6Qkp};6|)xt`fTUf61vdA5e9zLr6+h)=kxpzPL
zuzmRccY+IJ?aN>KO1p6@GJDE3W+%so?f$(x?eXKhliwV`!-6JbD9WR@(p<Zq?ysj$
zw%KF+-D*4A_uI!GzSj=+wlmg@<<ANq_xHBi{{CL!vxPeFnOsJW?9p6cD8lQoJ$<rW
zZF{`+u&u4G1{ZTXa4Y9pF7Lq!XAJEh99F*;7FXKz9HYCr&eT9pFdvoeH(5m;t|#r}
zz&2^0w%xs*noHd=aoCR2M&m&i<dB@mI9m|KD6bx})X`yggfq7HNI!Kcm$cXymGvuk
zg8S&uScr^>dx-<Fw#~}CaW0xjv{z-2)zaSdaW8G-L?#!Po{{1i7@u9(;`2OxKL{+k
zk&gp#>(RZu_p~`HieX+<Bgb>?;y~^sRJO?V$GdH7EBKUgnGOwk@!}%5dM9OkP`Z$_
z!lq{zlb0+TxTg+n9eWv($8ipScdDBl3VS<`+WYT)r+sp}J2z*hCo<<n2R&@>z4!f+
zWpCZQRd7VlY#sH?R_Kir>skB3_rKTv;P-#O{qZ0DNxuIuW4{+%I!HVBE8<BWPL+4R
z^lp3OjW=s6#foqZ&fd6rvt7G(E&UdahMuC+-T9%eZtQ$oup`S>#A3$E*<j8J18Z9m
z>Z3)7;EX1chijo&(omkyPQyi@_MI*m+tD|DGU{-f^Q<C?4{YnmvG%l_A-lVK72PEd
zUSUhiat4QXaZ;_Wt)y;rKeQ_SL2KZDMQ_>ek3oJJAB58!T+_K`GRMy=>S}j~bGmPP
zPo~#SG<s+!S|GYXo8WzSjx7bY9NyDu+fnAK2s3A8a@dTQ<7#hrH~1)uE4q2cqKx<H
zEWE@U1Cr3z<?yEYv`zcxg7+60|Lx$^S?Kg}wTs@_w+ozX+gsbAGmlaxea04%wjvA-
zPJgC?Q?_N?*u0wh`?iewv$nn-oaP{;_f3?(!}+H_bS!$GHftw-YCA)H1V=oe&ek63
zZgf&Sc6D_<G%9o8(N=A>C~?kNK!D$aKbmC=G{$iel&fnNNzB){qpc4fJgTjim8n04
z3z;{z`LtcIc|;z7PxSZ^d8iM+jT2qmUtTPD`mZ|_N~_OzvWMGvaDJz4e#>9FTfO4O
zkEXiU|5Z7|J^1Hm*#+uye|VNly~B4I^VuCBT@S6fP;;w0dGrsx^L^;wkcRddH&@R;
z(#w^evOk*&7P;7g)YtR$d;ZJ;PTS3m&B6g^!ifZZ+2&E1@WJ@OD`lB)XmUmHWQ=5!
z7i}=_M1!ucuazumeAPFT_e-x155^nN81fHYyy{7_JlD-?_Xws#LZg*cHlG1rx#v5&
zq7nUGU^i+v9-%1h?=a|}eerb0CJiPxa@4%+E7>TQ=gM){IKOI1Pd)eNq+~R2Y7Ad{
ztKgIReCn6?)lv22mpaWo&5&ofw6|{RM?Tr(I4kjiftI@JS>~ud`CT=n9&(d8?iv;b
z)S_W|^4#1Z*K3C{L3>nII}Bxf&|L6EH;q4(Cwt1@$Ez>MSGM{7JoDBr67WeIjluD8
zWa{(2SUafuQzh+t{qq-<KEJ%--Y>4MUv#_wE+;eN((!efF4J#d8jAwGdgEIA%5VR6
zyLRhVF}Oi#s8v0=N^mO%MKHD+*fs=Rk*e+ba(M?s!k929jLIrPf@(3U<i?amc&$V+
zn2%t4B?=26rM8=kP;+|jAkydEfR#T&CuWWblSQmvwNmT}K{U^tT~sSS@>eUntn|zM
zYE$TcSB9(->5hXcGi_kxv3iU^pdT2oXI2!6%2Bo#)T<m01<Y8K7iA%L!~7@85dmIh
zn2TUkicp>r2Gq$Qr@X#hPui5r$^Ct+r{1Q1P7`6E3zP(D@UF7K5tLJ2SJtLWAtX3d
zq*~fp&qqRsHgE(=oRMX;83f%tcdr`^1F^t`Vyujd2>tedCma&sx)U$$%O9oED$9i^
za1040$k>J=7lUO9%}>&fz{((>okTz^r`u1H()tphBNU}Q5@Du*u86>)lmzf`qwZP3
z3;$*V@3c4d>q`l%xxm6ZUBT<Wy;qTkxmOWfdB=UxU78?rN?&R<Ur?W*3y#`S=bPm|
z94QB8G7w-Xe1tqo5BL)Pq_eb%5yIi?U@cM(p1}!D&cgTPvv>JI;Hls627I;I7<w)$
z0zQ<XF;ZK>H{&O&YY-OUNf~EqY|Yz?5~yp=!rkh7U^_+`&b*}%M(Kc)+R|ZvnJ&|B
zboydAzR<5<ixiqs*N{H@JE?vS?d`nBFLe0)$!S|&S!n<KzxprQfBWD5*X?UxdpEqq
z)+Gl;vl-)UNn?eHtq2*t@R-_4D6|WY!iPouiq5IL_-oO6TMpaK$tqyy4dI)8uYdfu
z_?;+`@HI{=Y=rTxR>0SaRJ^dS630(Og?pcIrEj-Gj$sVt2x+B2%gg8|DG!h6&NGp_
zwuI|O#%C7{i?%21=Miweke3)Vo@Q)0TdZE6one?434a>i!BN)f@^t93_wfp}odalT
zWwR}>T??<BtGfGHTC=HhwXUfd@)N^DWHHA6w5#N<)1$V#^|%~N<>c}53tIyov<LU@
zwf)2Wg4dJ46HLRMQqF91=FrmeO7&G#K0c|h`oRdVZBM;-W*f!R@Gi1feqUJla#+<2
zywE}?C=Dk#8X2IZ+Tt?#@bH!NPb8OUq`iafu%q$d&W$!byBXLn)fo0wypwHuZaID9
z$YLzY@7c5`<v_eEoF-$=);w<8TX#!to0?`c<!pOeQ7!v>yM3%DZC^%i3Lm`^EO>U(
zjt(9r569g}IetF5{eHXi$qyq7oz?j1!?VCrq}|ccZaELj@Scu{CKE+-lsZoH91a>w
z{90s^NV<iL-`j7#)o$LnSy2!b^&MQ~*ktH^_~2e-(xZZfs6bAeR9v!vu~}Q&Xv?b`
zB^!R{TYu5M_06xh>(@5hSAOev+Vz`n)tKS=oN=5$;CB7`8*Od%YB^Ah2}3T&E63NB
zX>jDU0l>&N=ZEe1c()zwY?Y&j;eIObk)uQvD#O;8-oGMBEPbmvQO+rH)owdKVGRGc
zZ9Tr%=7Z<XM5={G=f~yz;)t`A<iu2u@FPoJ-PmYL>DT1c)p}-Zdz)Mms=_GV9fPim
zi_?trX~{+A4P1M2c9eb}1*Q|3vnl6_Gq6HCY+Fd~&N$Hv9^StnxQ?~j5-DS9-iprZ
zWm~A0&&)|@S4{;*91APUOVyu%u$@N-DEbljn)|P!27BAISIaqo&2ff)j@H?tMbCpf
z&YT;o@mgJ3%h&OCe7Kh?PxE`ZUCWqlZr<uLe&$(~O^nU+7oitBTMyEQPXmLqlpETT
z+KyA^VepZme|_V6XoC66Noi}yv@zq^9gpz%tlj?L+wJbX+hyc;+uVDQGVk5Hlm0qC
ziE(o__<Yp9_ucQdz5U&CTENu@4<EHZ{KG$L@4fe7d-Qmxeel5t?dyN~XTizuG+XV>
z%q+HzjT`06woT#O{9>IO<t(dPZ`^9vZ(Pr_l{PyceIl1AV6>TzLO#|9;{|`o*P;lF
zFDDV&^CI(>@zt=+%}v*tQ5<K^n<{xZV_XZvGPd2ZL#OJxqRafL1(NAE9c+lS@-97-
z94?~7f(W`N2U3qPbAIsCz-F<%^Uk~N>ecl;>pVq7l{_mei!Yru97M{~mp9+MmASiI
z<LS4p*nH|MJ!9)JNA9EGhV6Vif6`WbuJ~!%_UQ5BwzIPn`XhpYli8xVt_RYAkB&|&
z0`2(NmXT<jMH?b!MMU?=i5>+NYW>uX9sL|T@VDJ&U%=RTvpMzj>5I0r8@vqeot<`r
zKix^>-}KBxoqJ__!&-<Jo^5Ov=AHB1$8aUM?9*9p-BCAH_uv&;WQ<(qIew}P-)+^1
z&zo=SS3_@RIsN*sEA&q96b(=N*VZ?SXMOb1C+(w;KM4#n|5LV$?pm}nlToygbF%gQ
zJY&cvVchTCyW8&Ey;prXIpI*W@MNU_a_$CC7~!8pPb%CA49(YWTZm4Exur}q9`5?9
zl#lwIJpCHo^uHgzdNuD>8N>b8Qe_+N)pzA|IjesA8$B=Eih)X@=-DurhWx`dJR@Tb
z@W+GLL!PCd@YSV9_1T+NpWTMyNpCZ_^f|>=(Ct2zvC8mba=Mq#p`HDE^!wDmurOOr
zD)g0uc`LXrk-@5t>XZKT?UwNieJq_meXl+u0!cn?*~_VX<9gr1lA~0FAlgw*u;}t^
zLHMpY>>W5#3s5tz`ciOGdM~zqYlHp_{nJ<5k-86#@L4|Z*E_j7tkV9#Yk)ggcUcCl
zt6B<A_08{IFKRw2B>%yy;b;Fg`mUXI<p)sLm(e|Mr95*_-@wc`beQ&Wk?Py<6K*Lv
z?{^$=_40=NzOi$m6MY_L&Z{qZ<&pJOyV$-u%zQWA;8k=hiTzOJ7Rd)%X)F1TPT;7E
zaQd>JSRisyb{X^(KQxx)5lyOy+1x`*@sD&fN!=ry2(N=9bw*+4wmJN&C#vtieq8E(
z|It3Z0Ovp$zkdBJ>Gk)9`un|~T;A*TzFx-XrN8O<7dd?v4zFLotaR!4x=fepH#AK}
zaNl_AjrPXdZ?`4;o97R6$J0b&VbCu|^qJ5n;)gNF{eelAFcf7^>VhzX{CSvIH{KJF
z2vZV65yg{ptKaL)8YA@`10-OaM~U~F0|7_G5DCB>37i++sq?lXd{3l41h53(q5hGQ
zAezg{T22mSRy$Lk{Mtr9wiiCdh7-Zc9*!=GQ#sC(Cz;c)QQ2u#InoF=X<rG&1TxxG
z5<%YiSOj)XjWjcD4Xl#81d3b)6L2aUU0x+96cJB(VDCM!w!#_%FOf5Cc47X68nE%s
z;5aGaEcXZ{br5Lsa5|PymV5*?a4F>}7Z__Ch0OZ|L53pdVNon7Kvvo;L<w1lP&pku
zvC5aDuHc-q>fDF4nv!@H0psZ4IKl>}VS(Z&CwWGFDxoa!gA+qL)6q1$3os>2Mi~Ji
zc_?-I&ah$}>Pj2x9E7wVbc{2+4b)4xtttW&ur~(f@Ql#JATlT@0}j_2X2AoitUZ7O
z1T>CMMh$|jh%>Nem@7dh*V7Cpn2SU+e#ToHVYZco9J}4g8vINh%CSWaye8b&c9Fp!
zt3(Gpl(!479q)`4ys({?dBq6~Cm5o@iUCZU%@?2GX}t8g1j%K_&H379lv4Vl%j}ow
zGX18eFNEXE{QjjS^yzO*Lmt%1HKfn}PO6`;%Ir@myjH1CwEyy-{-gGP`@jER+CTgU
zzY|$uqWJ5<!9npD4wUW`>ML<YVfZZPSMC)}42{F1ik^jjF%(*%i}s<LgVBNEO*wd+
zMJC#W4&nivOx|bsD+54aP+Rnb$MvlU`vK<8UjiNu5qyt9ow3n2$z|-yUyroq01_$2
z&=cM}a>(Jj=qn|j{Gg3iBbSpc?X-2BsGN7c^tHy(v$(J_TIgdjg%z<Iq04J+c0LOJ
zROhV>B$IPXk(E|L!sk;?`1G^%?}cdk)Lq-hg-@OA*SR(foV5ia2Vi8J<HP-0Bvi6U
z`e>`J^V8#!$?!q)5M!<dK<?R6ZFP0CEp6Ov^Gh4yd$TF;EYN)3tdy@vwZK_4-;=;g
z6hy$0asuo0Q9aOsk>9pt-RP2AfrImnmX`ZDVl{@5sYLlKUA+;R;(8mKaMsFcMN5s3
zPc>T>UYXzo3IDtjT$v3_#zcCA6JFtHU?dHneR7bt1b3c2$#cfaleYc%UVHrbLFCry
z8aGh_TaWLyy`9G;CrwYzh3C%}Zas&KfivE^z4fr|?LAI=yV2?-ZQ0w|s%SjUSKCy=
zc4JRQ-~b~pFzmZ*A$KM63;B&f8V+ltNU_Plh$BgK0=d~XV-apDVn~0UzqnEn8MfP+
zn4GQg;V8zK!5JPeu85vmj;uRY(P+0n`e8S=uU&0(3l=0Xx-XSefXueMv|7&?0n0Fy
zKU-GX4zT3M2uQX;(}{8-S~Z<^FV4@mjrH}ma`k$fp0^s`c6PIEKI1<<JJ-gh7lV`O
zZ{U{BCUMTBooO!_NyOk{>VEUBcT)aTjkhQbauAqsPDVor&REdX%8)izV`|$$b4a1e
zB!~G9RxUz{&&)#(spZY|Pvk~uK(dKIbcRu#8xuV3PAN_~eM}pkoV3IJy>|ERr@04j
zr`vIGqi^d-HZop=(=GLsGmN7ybYe2{=d;LzDc?40h4ZP0jM@!<6729188!W$onK2E
zor|<ki<?B$ggTbI#o@P{@l+0@?EE~($y~d-5!$f0p7CBTo#F8KxE4^6<2aeY(Hz+h
zF0049VoY@&-NcpdSYKJ+C|vBYd{NPfaGOIk6tw8YpZw_`wI9C!e)^Oi2Tq(X&W++g
zFn{+BY=3v!9z5J`-}&wj0<#Cn^Q`Uf9kfqAz1#lykN>nB2D^dhr=Q-b?Od;IZk9vn
z=FMB>&?_Tv#-Tg#r`v4ky~r$MQ(MH7Yl--wWtrWD8{ok?4&OP&>RiP9hC67oy!6K3
z`s$Y}y2@OFKUc3_4NhOFh%DuD{+p+CKd|FK@>%*o{wj)UW-9Z|7V{US18~xdKryHA
zsmau5?7sA+ciLCJ^5xnVRzEn|zWk*x7u+5^be3S>%|YthFwT{s7M~d8drmUCfH{s%
zI(rwM(Z4L%uyvt`U%YBzaj`pU@;<#n(fArCWQII=5eoC-S?G~1=2xqq4<0-$=Z(l(
zky&UM=M?x*fbd$zeg@}Wt_VO21A=#fAD)N5m2o+=yXarqW$_0dOF!umWSwTvyR_%|
z(~L*jp*`opPdk9n50$eTc+n4)WBYr~HIZ0_lbHkBjX$8F`U@}ce9lPDfTfkiDo=fM
zbov=uhSz`{=eFOrZsdG5w&>ncXrKhAmERo`<L%Mot#<GJ!-`(p%lJOHf4`!=4v+A^
z=*SoJtKK)4hH-w;!N>@xg85DcFqXyJ0w;Mp9T>gqv;G1s<>XsAPBJER)JkrTz7DDC
z9DVlhx%#b2RmbRA{f)kN-5Fg}V0RRi?mQ&LxM#8@RU})UgCl%4Rw<}(ai~sP$v35w
z*YB>QR}J~kg-Q9uQ@b;&|Mo~ZrDInY{2INVhw4TXIEWX6JL=#ZC5PDZ+Fa5fa+mk~
zHs{S(b_(T?tIXjZ$qOIs`ZC$pLlf4ou2o;Rx129}Qno5(kjvCxi$8-Cb%cSk@(djD
zF8wSRm}t4Dp23ziz+W;4YjJ5VQL5G6yx$i|>-1=h@;?5cUjO;7K4UbDo9{!pbuU+k
z)vK#FN-z5*zssZke3T){eRYtR@%h1SAe%4snJ?-#Zgo4~*<KW5-1^w{^2<I47xKeB
z^vwA6MK9`o)%JoDc-Fmc4`f3zf7mx_!CKxc$5#^Pq)L=o<CyxusPczKD4#RF=t`ce
zUnV%`Ytgfer{A_LEj|;R5WRVjd7dFjon0nDS77Ool-0}o$?3=C8Q%YC%lW+Y^St-t
zuAe6j?|-(;pXK_S(&xSNH(fs|eO~@w#D%k8I=(K`W%^A^QKZ*yUT<Iet*^D4Z{2FM
z5k?4#R<bc3Vdll`A|MGTu9c4yJ=EBRFXD5IE4I4hWWpGcsUB6+1q+G?MypJY6nm@_
zYB62R)p;L`Dg<BxKLZ(N<!miN86y#4s2FINDJIL1GL!t6DW*wr(B8p0@GS2Tg0$ar
z_c$fWVUu=h^Wo7!?nQYS2_wp`XrEIKvM06rd|0JJaMh0rlTTT>Rs9733Y2$?(Wef|
zhg8B#eM_zqas~mm3sc%mkO~xf-h#)AI$sK7FJn_+p}xE{66E@dz`nH-X+haNDyZNx
z(FGIba=>sDa_*$q5)y2+Qe&6j?$>Is<Ri=x2uqnJJOu|RGBZ)k7o!B)DzCPmj9^5l
zDmbQ&@d(qRhmN8|93K2F9Vuypeis_Sod~oXlyE}7C42h^YSq^g((<ohX#hMAENg5s
z25^B8<u~{c#EhRh@<MRI7Jw02QwQNtc|z(3<p6%Y9HaEl0o2vT<SQdVlxZW%7*zP3
zKX|AqC2Z!MQUr2?g4E|hUs=XEO6g~GC1}>RCg~?358iO-!5wY3f)&ndwzI`J?u@^7
zJ<H8$(GIDrR!ybs{8xFcisHO9z64Eeg8%aBBiInuDZ<mU8Rz7OOSU$7iKv(9GW}+y
zzw=uizT!eZO+vLgZSCo^zmw|csLXueSyISTySlO5{`o)ur|qBr*Z-R~7anGNu&|3#
zUY)0b=UMg7359QTXN>bV7{Jk8GDCNCaR9K-m!h3#3=CGaBX%>PU+y#Tq45la<@`#1
zG6>$ukcXFvLhC%8W5RY*R?9O=Jqyk4kpc|IU8!eq>`~mdj$>@Dt<ut_DmV1>d6=9C
zApDfM1f9*z^yi}Z$?*xryDubJTv}_JH{Nc`Yj5ZKt?*G>eT5D)xXv=vg~ybNWt%S%
zK;Z}BEt6ML=M_c*v^_G(;Z8d{-VMtTLC(3xIY<^s+taqQ<L<2CIItZRBkCl>K-&4_
z?4XSH3kxf4<JzsZdEJ&{*Fd|Crw@y^NSj{{-(r*vZ<&~lTyR`<hys|Lv?5pZk3|}X
zRsY14c0DiHJ%4Uxwk^HZGZxeN9dLwaiSS!m-E32H*OPxC`JdG4`srzrKkDiZt}-Z&
zGem`dXB^70Qd`Kic6$6Ua@ym3WAsXU1F9F#kLwJhlZw(h&KMro_HE84nx3&>XD;vT
zx5tm}v^#e`@<O|H>+QC-@kU_o{ErtEAw4m>+{Pvr+tc%dikg9&LsUYfislR)XJ?5_
z;xOYZ<S`I+w7BHVv6b3ZvW*2kX<^}O#?Y2~r)dkk6YVgaGWAIy{6*vd2AlbX<#IC4
zF0Q|<vv4wUk0`OHPdPo#gTrJ&+W=0s7lHrNJe!;A5fsa-t8Fo#wz_nY&n_Zkfbmph
z6z6-AJFI@^069M13m%vU=V>qeG~XBxMbC9Pmf>p(UXJ(3q49}@Tp165ue06YYC4j+
zBZ`pxF@D8%nybO_>lK+caV7IB_sJJ8Ql@(9{15O;y-&{eN{;T)cSGb=cT_#k*gk#A
z32H%7Zq*ijo=M&xaD!tivSa#lu=}VSo{<_0Px~S@c`Q&n3ohHj`?$ttHo6Cq>FD^N
zodidmS36bP1Dc0<r|R^d5lu=T&yIH6&d#I2@1PuE+6BMSwepb_OwzWO@X=0ko;F5r
zF*kR&Yrd536Ifb+hQ5eM?b~_60TF6_+wR>x+ZsO3SlSjb@GdxySAIrCa-fEwJshOY
zWwJHwl`G~^o;^Qp-~Ov_v_Jo||Jcq$!<-AY8~X4U-}pwm{mH%Hlx?=3whur0xc&KG
ze7pVmH@=-d9<<xH@3!~e`?!7Z;hna>XUlHeidqboaSRS~E;t98)9Lz+8||xK{YrcH
zozS=3Bg5h~96TatMOTr7=`EZ>bF<EPijI@#&a`8&CS$KIEmvfns3GIwlVP0PZF@mZ
z7$|aWW4-<MZ-1@kUPU+^ouzCv?rD4U=y7N#-6lFuK8>L01cp;ikvfAWZJ8^5#yNkJ
zO@!O#watpQyP3XRyS`DhVm<T9mT5O{U9Y@guFRF?^f6`Jxhq=eZW-*!_Z&mU0$wn{
z(y=sv6USWH-rfrC9M!w@OFRrNkkjd*<{sV4mgIB^&hwSk<#zSzTC;s3x!;yw&e7r+
z<Y?T?WxgxVb5S@Vru38Zt(VmuQKiR)->j_U{m>e^6+MRoi9Vzs7Kq>{_%-J$U3YDL
zHQ;zwZBVZAYx~Z;k4GBQ$!U&=iQ3+9GIgP4>sPOqe#GgeUmTU%4F)2`@C<a(*vsD|
zw;AN^X2Dr!J73$t+Ttt&AG!#?@m4w#N4+^<OV#ryC++6V8x>8v5<CK9Jlui}e18A%
zxZS<?F!S94m>%h83r9LnU&z3yPjAY5dAIBx!9n<=55^C_==wVu1ctj8I-hy0zTp`?
zuhX);mP>v47Z~~<ig@5r&;RoG@M-XU?LS|p;ki_GkKXMtqdoWXyCbFRuTte3%FKVg
zEpqo=bl{IZYjeqn%9JPn^}OCq847S?NTa$-C$IbZjyJQ*^u-4ptZ+tqUZw7snH+8R
z3I23qkvWRdCUp(+r14t}F2l3p{V9qZ$$3n^(Jc5s89uYNwp`9~b+YSp`3pU!Gxx|M
z_5t&?7AmKTa$W_Ow2=<3ybh`n4&vKr+NeGH2p{2ycU{VpAMGn0JN+}K)U7Ywe$y9J
z54Iuwt3@Ffq2n1p?E{N$ml@fD3{)ksH|DB&9hOPj<$<rpBA;-e;Aq^FJocoQIm(-i
zXluWxjP8WfSG>u)aM?5cGhZu$HsAF}Tg#r7&njP=%Zl@|IUN^kJUg!hive$CDn1Yy
zLUfjPyKmmA3vZMk4w1PPf~UL=6K%scMTcpVd`|*j3m_hE?S$s;*5Z*m*W8%b930g@
zN<({w>$B<e?)@UtFRR?2Ck<`-RkZEbvHzc!P^nAD*JZj)zey>yXMS_Fz4f)PwlC#c
zi;z7LW{-iFFc1bVRfHDh#P)<yA_$-&S2$F#ROMjeR&13pa3#V<Zc}<OGFzvV!yrtH
z0%WTfLNEo}J<Q(9FlS(3bY3R3i0WeKDuE&8V|kdb{1`MLrW_dr&k)jbF63Q;ZFdw!
zu!#`nMJu8i*wEW*Ul@I@>JHPy4z=4CMkK}t4k1Y$710x>%gLafqH-`{1__2W40v#u
ziS!~_GT`V7p{7>&=H4q}PvD4Z?EMG^T_|zC3$$Q17dUZ1Vb<D%vDelvQEZ74Ruh6Z
zgQ0Q=dxWwQy0mE|Xn<A0591F!yYTQL0xIFAI|((50L(Z;Ng}+GncOb}oqg-0fDl3{
z{n|{hQbv1jb^KX7J3gybHfP6HDpx1}mZ`T1to6q@5E2a{=ORIq;Hp1h0cN$$lzyb4
z6yGuc7EIHwv4U#GuR8(tRbFNIgd5;PAU!%cs(gh1lZ;_0KN$mMnFj^u)X9iPnU@d#
zT4mJF80(H+^;8?REyA39CCKKfHgyMYH!N9A2w%EDnK4N@gyj;P)t$18neVkqJm0%Q
z#$ao_`WbKfOyE;T;bh)Zmib79(JyVPF-?1FQ1g#a&yhVFg<@jL`6?nNGDo|w@iJYe
z-=y?~aD1U5zr6HWqWqI;>onq3p4an^$_afPeW|Pd4{bqSIBslB_r|RoWq`xC&}Vtj
zzmUPuRa<e*<$Gvi_=hN3Gztx4OmYSc+2CnC>rpCb9)mXH7sDbWESkxn&d`FUp{e*Y
z`v3)=5t)Hj!>j}ywuSAqH1CJrs#Cl4qKqRcmy?Te=Hhwy!x6{Vun4F}g$*lY)7a`i
zUZLGmci53ZY!M|&VW};wUQITR9{I!bN6wmFME)ORD96{2%Q&q~Wk`vf5-yee&M_l<
z*oNyg@>GvB;K-U^+^ElqiGKEtRs0O3GaM_6tKC5}*H_HW%`LW>dCs1tHa-#B-^dn`
zi@*wh)VERO%lH)wjJiz3X-KAEpy6cfiz^tN$U9>h=k>O-cC$^-aS+Ydn3K^gWEr1Y
zOnD-Gz$mcJ*pbz`3_;!_ds#Gel)8E(I!B2ycQ)6(J0G^4?X9$Etck)KOWu;lf*Txh
z2m5=?nM#<~%IbRCy#7WLg~bWBv;DX^U+7C;`chlHdZUd`uBWf_<q(`;e5<YMMD!fD
z!{gn;MNR~tXD7$)gAYEeadnQ4euK+d#=%yH;II8~MSIU8M?8FZuMEdzJY~QI3y9zf
z`K+SvPS5jh@N#jvU721>k@Jzg=4#Hs)%DFAO~m)v*->y|JMS~vN3M8w+D?x4YMkDF
z=bd)##+w;SPF%9|XmLww-g@+~-M;-%$s3~g;fpiB$i&LGU8Z)q>I|sV%lRexN%{KE
zA=S^YGCnEwMVF^L(n}c|9-o(nI5-0qeeQ_BvdyOLP2nsV4!k0nq@Mk@z4a*b_&E0`
zf)C0TshJ|iICJ+4zMQ1y>Dbud7?wdod(;~_@WMHB`)zl3t8MQ%OK7aN1(oH+GiSQ!
z+jROISts|wRa6hdImb&>l#=VvqV9N8u8ZtBy3%*}2A<}L2%3_si)L^#k)`1`czL%p
zm6zX7k9ON9AAj8L-@jY3IL84T;cTGafGIihH1Zd@&{m?x>Lhc^J0E@g!|r@X|8l|W
zPd@#mqFDT9oMnJy;D7LFtL`6(YPol>{n^*Q-oE{<Z<lOm-ig%u_IKWE4<GFoj2=9C
z)NX%zFL*rFHa4%-`A4=*J->Js`ZL{DSJvD5`qgHe%e6e?Am>muCu}u(^Tzdpsq=Y6
zG#zrv(KP}qx&^tC-T{9_P;qXdujT@$m$6{PCTD99+O7=!hiV)Y$^|EmFm!EsWvSr!
z@X_OnmZF#Kg$6tf-rKIuGc<|Q%iLdIU#~jQBX!Kq&D0i&t1BzvCnB5(y)UY5_*K^v
zZTIP1u5L)b8(2E4%1#TmCj2<#Sw`Q`D18ypN4G`07<Ba`16#Zc+#f!C+zt<qa%Zd-
ze{ApWw*9^RqT}>L5mMK$U#|skXs!q|QKI0Cr@8PZ_c)T&p|8XCe)JK}e)^&4D(4;H
zb;rkeg-Gb<(YphGeNKJ&mGiC!$M{y@$Y5NNWzj1{*x)1Au3axY-%DFPw>7W`DR|nq
zd34U!ayhox21;iOJm~Ar0YhKFW?^wY{meYcJTCeWe6MI=xKuR2IclBP!ySuYJHL${
z6kMctcju#X7+++ZM30SCq@sv3<7*)eo!O#3Q%~QIkI(C@vYp)y>vA$=+_YaC@Gx_)
zq?z;^&jrV#lSOkq>5)|M@udSRzwxJX66AqvaC%j~UZO9WS<mWc>gdj*(er_vj#4=*
zhDQqW+=uT&8a=OP`HmL(0=LRBGUUzY-~{uh58B~zIWIC+<O~T86+hE{2q^VpUtTZU
z{Q7r|P@sGH1t#V?JAxfm(BGnusdv!5r>1+f8Eyr?$#3`~K2UN2Jve2mLwjr`Iy)Ph
zk?D)~IT8RKqaWs$MI7b}Co%n;KJL<<J|DY7i7pL4QgG?rI*3;3g6HLc4vge8m+g3i
zE@_ASrOO95CH<>ib%HOR>4*0eHoy_?s9X6Z%jAo^=#%>;-;T<yEAtTU4-P5$^J-o7
zOF1bQy>aK|5G0MVR|2~M<~1~VrrABeOz<F|x#TUpFz&hKy)NtK{WRHj<Jt&kwas^U
z)9qwzYU#_+hdl4FCL4F#pYl4~GM1@eu=jO*Hvi0rz$k6Q-?XFbaw)gwQ_4`zfN%9)
zz84(RhVCrPyJ<7JvVU;Y9z1y5wzhVPAJ*2T+Fo<6_kC2<-<WzEKQDb=0iV71qwgx~
z_3PIy4Q={Gwe9u3e%|#(NhF3#$Jb@LOus29GkS7mzTNt*FSmEU`qj3&c{NNtN&x;G
zMu=e%J}E975(I%^6%@u@#-|9k6zh<(VoU?G#bq(DVtRT1Ecpo-?hzOW>B{mu?5KoM
zWjVVjFO*_vr5FxnI|>n5DYytGCRh<RVa%8);dmOOj!@;ytTK{e+SR5m_+Zlb5(dv%
z&;^aI>`(xdOVM(<PZ_5ev+aqR!H5Y(=(YYUbBrL7@5-VK)S>%IO!*NbN+B8ueEnbb
z5_Sj$-I)Ux%CO?ebLRk9jVIcRprmZu7wIRVL2(3D0*2oN$DSfo<a-HhQQ|pmD2fj1
zFGEuLMY-?}SSzPHnI@a)AHtoBvBQ}i$|F=-nM#?N&V4Uwt1`+UM}f(o^RM={58B>N
zl&1aD_UwsVKSQe5y1=AcB^;$!97gbnB1aJ^IOdO#qtC=p(NOxzSy@h=;Doku;&vSC
z&ZoMc_NVBKSC1wH3r^X-m4|I0MPzlHHqMNrwe4Z@JJSsOY$3^bWDa;o9;=ROoT@%6
zJu{c`&3h5ws&BQzi2#<eDU-H9Bw*I#l&N0NG)g%&XM)=dt@>!&Md#XhmvR756gZBt
zDkl|FI&9rVD6KOfN{~<4SNe8Gg}bS5Far3TUzlxki;Ou@K9A2b4}ZonyiAwr*E4-F
zBEHb0*ZzKy37tdxUS6a7pPz=OpOXty{E^$Q%FKVo-x#r+lfuD&>(;fl82Zxra9^nL
zG<4?qi}COt(Z!RsScEe{L=s+24ngxmy6PVKOR;4{@yQT|9@Yxf<QI*E1~Y1ixU$+8
z{lm{@X6DQJUyhozp`02QXGLEbu#4Xyz)_0v95h^B^fg)1PWdC7<T)P7c-|djx;|d~
zU~tyWr7rvjT@|6gu+M<%EYmp#K4(EO-akJHzr5Y{ckV}VKMMq(g|~s^vv#z%-9(k3
z*<_0Ilf&G9k-pEi>G_qmu(**nES2L1@4hnSEThGAVLJEH#?;p%;L^7#{j>ETqtVlp
z<s3ANdBQs<bl_=w^x$@T{OEr5fvhtf{&5ALO*x!ysk)5x2M1e`l_uMY?b}x0NT05^
z+1a&{4Qy34x3C(SM}!br!5L7Ibs|fE_xL#HLqAU{kZDiP4%1HCGoKcJvOwzM!q$wt
z!JRHkJ<C;*ZO?3>XX~!+B;9%ZIDGzjZ40@uxYjuQIABgsjrUYrTf5$77DShcY;b1h
zvvOLku3gPNky+Y-_i@6ggHiZA?;p0EoyP@xtH(J#$yy(O^l`~a*EcuH;2b~(A5Yuy
z(O&w5Pa5x(7I?wA$6F6l&UP8=IA(1zd3d<p4)?dJ{bUDt=9x3%IApqWX(Hb_p+s2O
z;&M9uRL^l4*cf~t+`k?89Ml$n^XoS&if(#tHMlYzIk6i6^HRHe@6-03Z-29GJ-Qdn
z?y@ck@%r_*+Vxvs&YYMl+;qMkjGUQ<>(AQWE`!ni%pp-ebO46n<QOB1lfixW2HeqR
zj-p{(z0o4Ll=vd}eRi_j_II~Z-oxB}9vMSFpVoO+oRMUday)0A(FGX4M7WK;IBSOo
zI~BbrKUubKd)eh3W7TTR7}$fOX%0Fede6LAUcDB)SPi}gXET>LR&32EvJeiE+c+}V
zAxV(*7tRm!zs?e@{z)WDvM|GC^ox`+JHHfI&7~gO#?438b>{Cu&A0JMZ3GVq95{r*
zEP6x6n{K0@&PYNN4l}2W?L_A0;?i<tOVN<dRkHosUP|aY2`H-~(=t|!y&Mea>U;0K
zm%Q6;dE@}F1>wK?{eRp3r~l9Y(!Tl4@3h;WeA>SI{rB4Y?|+oKBAbp)w4=<CoxS7W
zZ(2G%pD_uZh1Q{KwvAg^b%vV=9+9s-a)(SxW|g`T^F?ID)3hgZH{e=aSje0?X?O13
zFGn5yikyzF(;2ljoe!;=4sTq|z2LVwkbI@nWImarww=T?Y(vLk!C09rC0AR>fY#Xx
ztvmfih@6)LX^?f%{6QWEN3wRO@pwRQN9h!4^YY?iMQ)*846HlbJE7O7qj`4L?%ugm
zbf~u2O8=F`@lWS)K=u9mkK5jUX!I0XIbMc%XPAj*<LoZy=1Jyb@u@!l_Vz^fiA?A+
zxbvfqGA_1-C%ao%IylcHbsZ=F<HxqD+-nC%76ss&^u?5y`Z;utGLM&67VGS^jjQXy
zC5~{562eQ4Gsn{&b9rN9Ep6-(Vj}2lkIMmHu+R8$E}C8(UfKZHIG3KLogd%+B(!HM
zeX^i~@tae0yd07A4O>m#NdMu4^WZr6ILgoh+piYS4j<PZ&H(&c{}0lKz5S!My}jS=
z-G9`!(pTe(Ud)7cudJ@L#n5^5_xTIF#5SWAvRI%woAwyH-D(Tx35^jxTg>+)v{2cH
zV7RqYc+_=&7ryptuH!G@K;IgiQm)|~E8pQwr->Hja6oh3c3g86X&PO<jLtjVsJ|*8
zKXb3%sTuA@Zs%D&ukZb};T?Zn_wQHF^Xm7@`ul_Y)2Dp)H|w|e`!k=f-uJ!w9X*s>
zHfmh1lp7e8L-1w!y^g$+G^($D4Of4)bliO5prTury(Z(&&H{EEdp<o^t_2qsf%Glu
z$bfiNwY=sI8Ene-&qR_`WBi>9CK?co?G%L8ym9Mh+uYbJzNUTVt~pS)o08u;{Tg-|
zf@8J#H)X;Hb-5?Ox5WkLgWAIRS@_|r#Zt9cCzoeh31`Tlpk4h7ynP1)?}97n4D`->
z`dw{Go$mLcFeX)f)#a;t+<WP4b>DdAzV~WeTy_7|d;UZ3;AbDw{=TG5-Z2hvxE8AG
zS3Y}v>6?D!d-XkiGv})sW0(6?U&@k4$^9CfUTAC}lV|0)%H>^M@T}V-IH|f5&=@L{
z(+m8}MGFr+(<krOqO$bIZ@d%S)v0ah_gQG@lP7Qi|2(X-C(V1?m#Vkg2bc5S-;_Fx
z`}LF4Xui}{Ao<Mos+^xx)@Sp5-u1Jj&nxFAU6<+e6Up|{@pYLl({Df;o0({<w>H~X
z|H1FIul~;8Ya7?El|p#&;%RNMK>*-Hq6`p<te#=qAV};-**HAzXL4}V)OKkR+_59%
z(`v<90h0H<Yo!)}nBqs)RF>$WVywwqD^Jh+>S)Z;s<4V6Iy;NNk<WZ5Oc9=HC36Jn
zau}s744UF9y2kqi5KNLmi}-=nVfqC1S{afyXd~gDfeJI^=wLv@oQqM1G2;WZ1)z6)
zju>`#E_mm>n6vWmJ%&5Nk*yUm{}LWXf`n@jrqx?Qg+a)BWo&w4+ajx|`wH}-o)SVT
z=g|HL3He1x>8liN=b`OA%4bk0l?m=8Se4Nu0!iK{IFz6l-e1Z`+UK3F*meP}*Bzxb
z?Id6_*7Y;qx<FL#=dTpH2ty2<XI9l6Sebm(o}4m}=-khqv{(D|4-6?8VBQs$v@?II
z08d6W3A_ky+TMjWV=q%K-;)=9^rmq-^?m;tP6%5i_@$2`7_DGZZ_R<gT-mjKWyaNc
zApKqs!RIg|guyxG8RssLW}Hg-jKZ8eggeoUggSM-ZY(AAm0+5o<}B?vMKB#tnf;ww
zffs}?151<%_pC~^!ciV&!jZm`o<L~p&hh2|agrZ=?tLPR_qnO>`efYo-JGN_UYQ<m
zr{^c_>3LTIF4JZD4M<;{tqzg=;rfN9*9-nB3$JXWyrCC${k4Q&c4fG}=Z}K#ypfQV
z_RhO+H0L>Veu#fQZ~Ob*xXGBwK;hh>GQ@{|p_zCyrJUiUXlLXA#p91dWUe|b>aaC!
zZL34BnC{LMTNj?59*5rBu7<OtJGqJ%zaWzwjp(>VFO+lqgNoB54%E*Gn%6_CJ5R63
zg^)x%vv>{VJ^djEFuv-uIy#TX^Sk^*1}Q#{roV`6@Zw3^dUU7V{`md&@WGw7x3`=2
z97l#}B~Rhy3nJ?y=hU)}z=pgtG0kwFdsjlEC!Ht7*b-H~BBoqmV%r$2b4NxKbzV`%
zlXiHpTZUz00LD*J=lR)T`g@fAb(rI^c$kGm7F?90V1je<N}HN^5uP`b@mgzR6S>C1
zi>}ySZ#-=w^Mt>QGunb<;C7L+kM}bUN7bH+yb1n1e{xi93HJ#Obi=hR*ch6_otm?C
zoXu)=zXdz^|I*4vTU@?TQCthN85?qE#`Vg~V%p#gD@G`h?|7c`wic@$&z>H&yLWC^
zdyE?%YWIR_Yw|xik1VS#&(1T}J!;>$tt>CJ<)x*9)z;&!wsCc}-MV=_a0m`ngcr8j
zBds}+9L!#k9qG&InQi)>ltX~A-x*==eg8Y{hwpvAjBq`=DB~C0gHL5ZHBs7GfM(H2
z#udBgGzsq2d7LZPYlZy&?!(LxTQ9EX_gdZ&K{yq-Pqp(C(HOgh`#U>Z?fumE!TaBD
zhldBv`DAar@pjv|`A+a?J@q{+N8EmJ58h2gXwdef6I;riMy42T3!1XvnKR{TtHX-C
zGRE4Q7hhi5D-wf&kl`!&Q)Wb~JczuOj&kJMMv+sCENlBJ?E()@A9Ci4^w)W9Cuzsi
zi?ecE!a2rl&P~p)i?hA9&zY9zlQl;<w&qIK5%saSvexOqbGRew#~EVY=ae%4p9b%b
z+s^iO@bjpgU$$@K3{dA(bR%+A$7B7>AaS~laag4-(NBiCwZF?ry<5&TXUY{~6g=n}
z8B0#>;L!NwLM<*@n(vVda2kw^y|ccIvH5l$c<=4+<ejn5Oy^D69xn7TZ4eDp^4^n+
z_Ta&z_NQO}v-Z(Px7+UCLHpo?+wIT)?62B?_@h5*4<2r%{LGEuDZ}6N^juq8=d{h7
z2n;PasEEwTzUU5JE@xNTgf7t4IG4zzoEqpO{eiA=e8`a?>a=eui~flE0jF++<oIH^
zCRg``TtR>u?>dK>%$%}0TF~u^SUNmz`-l65r?!U`l>?3z9)dH&HXgva!WeBnie3}}
zN5|m^<j|2<`{!)km%8Z?o>z3;fo)8kcj$b{F6+}PA6LZKcIlxj83Xt4-MLp$mK)bL
zE4mC_^u8@m1CnxJ-M!<?v$NU)UWAyoainn+N}Ng$A3RPuyX`o1_8?=lYhmEANV2ig
zz0kz7v*>BjqsZ(<mva;B528<PZ|zi+*v<~vL>><e?HmFRJO?LpjM3kgjH_$Qf#qTG
z)~-{~11@SzjFad!3r@kqnM%$Cv&}Fa0ME0y#CY3U(BeyuujTZK1M8Dd?qsg?C|%`(
zbw%FgdJ^7D7vUr`hTzHZbsRi#rm^h=MFn$2Z>L_{9GVlNvgk&i+`bc>eN?yr&*)O3
z$SfS7@0-iuVljk!#v5<85alQ`1bFF}t$|Nd7iY0E73t`^ft|=k@X}AIw&qX$_yA|Q
zg#Xct;Jp5UHJuMX07G~s;p6x_yzcT0`T)a6JRYR;KK&|e^i@I^>R;DIOW&*8xnuC_
z$MC#=4hCsg-N{FPEp?^dx?R8g*V`@MsOWl6yDCF{_q0l?ck!5N|EQeG`0{&)33;I3
z?)~qDyOJue`i46D%hk7x7nt?PpuU|SJyn$caAvylIW3NlonO~&M1GUY?5?72(OdY&
z%IZ>E4-c*=u9Qp1u0>CFUrA>~m0!&q*xcAiD@0_6e`S8U%=wp1*6&=U1HZi2+daTp
zXX?m>)TNzpKwpe^IksfZyS9Dhc!z(U*OmLdwA8^_lzTPa`90*3+$&iFtOnSUt<whm
zQAXiZZ%+M^v=QzZH#q8j-`%hF2kvtEUbd`az@sv9m2F~Zli%8t&vH-(ua(tpVC2`(
z&&V<6Bq#OIc8$V^26&)BQrWoj>7F^v4x|1)R_t6oGK}4Xyk0VOewXcGwCGx!%u~;_
z5uPhUncXHfsw;omQ}(CuZi@#_&z^?%9JEJUp+B`qpdgs`c3B0Ub{Hu`OMjl^{m+)s
z+x9B?{n>keJyo0jhBS=tPr8Qm+25Dx=SybLrQ_=|U8XN^8jnD@_U4Ut<4bS1TVMHd
z`|59htzC_Pe`RW-99~Z%5Qst{d@!P8KnyWGGU-LFnzA>!?QS?<7!w9(4MsF<YatOD
z2-F;3j6(!d%yK?LwY|nEGgkL7ZZHs)^C^n2GOUEOVtQC@5>lReA`C|m9uy*6vKV!q
zVV2rlt4Go<ZNhXY0A<j@M8gXhw<sN+W6Ttu5{~jo81D*0`a$5O#HBqYkK}#ZDKLtZ
zAU6_*iXmg}BSvktHC{(}$<JzI!6JFJ!F@tJApt|?$mvcsD~!7^4c=BR69~L-4C)>4
z%a=U96C%6whBL7rQri`TK8D&fr11!k1PkSN;fE2+isQg3@94j}l;53RGr1yQrEd~h
z!FNIt2Mgm!Dagk6@U$K7ALsKd0&DN@&zwdu(?;@O34vE~l@JE*42cXX+7+0lkCYe+
zwZ8RLL0}KQx)#_bC*^$*%-|Ox%#UQL{1x4oKTc5N%P`TEJjS$s-ki4yZSa(`!ZBEQ
zl{%|E%1eO+dP2VOgO_mMb4sSMbOu8?@bX=oZ6#o3UEy*7U1O~^DNHDTcOv(92boPj
z=jRqG&q>CDP{^ocz6=6lhe;H`D5eBg@0f#x{&Kvf9uUw+zqQ5uvNBtr3cm33VkGch
zrpxrjO<x?x*RNG73i;~m=O&cX$M6|G8(M~rhorR4&9!#@`n6{3AS=l^EUd5>tuPh7
zG%qqH8PP8M41HDz1>ZKiWq3u0Lf_CTvO}l!WDfkeJ1gX~4WUnMU?{T<6#8k^IRhFV
z&xp)nMvmd2sMXt{lYPbMco|<Q)MzU@PswLQ9gM;FjdJi|^i><4ga$JN<MWIN$`M7O
zujCfnb#3qM6o0Aho6{x{KZ{GN;V;ug(-#)3VqR<$GmO$}DPtjZv^-}-)?dbVhUE17
z;;?N$x|8xw+GOA`K4!}`3vW)_S^CwP(Tlc!xZU=4AC{~zF)<VRzFfSWfpmNOLHd8z
z7IJTW{d!wkx!TG}5LFM)yOQ=*bkrDl&$rpxj0N7K{~QscNT-YQopyZixW)$?EdNA1
zOFMQW$L*A3NrV)q+?5w1A&%O~(P8pDtL^EW`9ikBgBKQ7+icn+BCVpg^T|0nHMgAi
zIk{%q)a-a_&H(3Mnv&1(uZ!chyY--*ogQSI`_2UJ-M{n0cJJ<oHJ0u_4+Ax3ws5QV
zzv%q1A_JZAl;;*`aa!317`{Ecf4?~gXkl@#j4UEiY<D&@Bl?=tjT0+;__RCpY&A)4
z1Pg}b=Yh+;yZ74e_Fmv^;mcHwwN>uvcF7rMr+e*B{`g<FAO7H*krxi@?6TwIy)x)6
zEU%@X&Q21|7<`UwA=;=S!^VvRAiM~?w`%;SMWLj><^`F4b#1*}zwu_fapR4^(H4M_
zuMf7`?T_ATpMLUj6M1B7IWh>NF((K-)OPr_zrSB|z_|1+2TP*LB^s*l4&YO?VlZ=r
z0Gy)QxRiOnRBbA`I&Bi!$xyx$e4Cn@%N03hVlib0x1V1GXAaud;|G1aNW(m)KjxGL
zKu_B7-d6kQqxS=gv$ndvmgX2sxXM}Um1pVSZfJ~1vuABK<0Il~%$ZjZxEb@56WR6Y
zg#}=?r(;O#x(B^L#M4+s2Ey?U-=56f<0{{IT@N1KYkLQKsWULiIJ;_^rXgb)o3Zhk
zYM1v#2flcARt_7Ek=wUFDT6#`2_roDk~8S;-TQf-@rW!fBIoueciOkU_5Jq#`yaQD
zZ{KU*``!=p{dW80lLu{UYrj3tb$AfElD3Nc+uXd_u5Mf_+9R@tW43S<%|w^ACUmSD
zq(xJ5pm&*>Q;N<;J|;(2<Y+DtH6;sXZs-dd%20a!`ex?F!b`ecG&Ff(2SX#8&RM2R
zbXhxsW}!bv<&Ys4mqR@HIg-h(qNX;lUTqs2>m6sFo@b24>-?>^-+HSYJ?1Q((Uym{
zFy(X_&TKN*=`<xz=R2BQhHf}u{PU@wyF0e_+pl(W_T0E}GyR=xj~_oS`k`F>N>tL_
zyLW5*$o>7jwtui6_*`U;bmxrz58EhuZY#;{%=KDW5jv^g`g;%^i7_Al((hW#7?iPy
zken|+XA1{VZBKYc|L$^mIUw<(;&~k}<Sm^veZ?oy?1vAxGN<pR9fv8eZ%La5q&~Y+
zOxAdC2&seq20mz>@u6R-tIlDY7bP5Cczn`6x&3L|=d8*Y(RupWk@IbBb)|`T!((f^
z+vF%`(Mb1!FC1YPPr7cb{avTAjhXsw_i51u2h&;F+VKX=H?LhS+%m4((RK6}fwyfZ
zrwabIUF487X123@8u)>!t<m;&4+<9BJDhEyyMeK!{%#YXCw4hOf*YkXCduP_$D8he
z={P@nKFZUt;8@C|uhskh<<r;FVMlofXM$8Ejh;#7UX?X`=E~PTHlzDha3!D0bFY`3
znyaqLk=s>v-S4TMcvv6(%TSL?S=F|*F<%33`lH{yPra?9{Qiui3(fOFrBMZy&u#C*
z?a<m9m*JC6*<o99Q*gyrbWiZr;Gg%p!$ct*U!wW1w5#juZGCOE#>3*BavG&BIIrEJ
zFfGg`W0sy9dW61kq-A=fEtxmIXl2?(o+DG1eJNM9Tc7h#JN3?-GymxD<Pc+6_@IEa
z6HXq4H*(%q%02zmHvRJHHycD%nhS*dS9TiDa<7~@Rc4n3z`nz8G+Lf{m{jx<jZL1i
z2aN93cvXH->A2(GFqT7kQuR~HZ#;oqg#M9f%62hoKbf({B!5cfuq>yfax#WhUdkQH
z^0V6nm4$EEMm%yHdK*}gePw5-(k}PSL-XA`QY{o4mF?2*a#ANhzEb^6Yx>-TA7ptB
z{^xCf@2H|bYXM8>fXd8IW%Nr8dA~k8EPk5w>iyTR(Zeci`0@I8>9fB-KPm4Qk%qQ?
zk<%;q{^Y)VHs3Ge`f-W$cj@@LOqb~kn#LknUVY<c``!QiAGP25-QQ`~Zr*5f5!l1Z
z+j$rcVG@(42w{8#0L-gz6>_EapD%?d%!-hK$zeoggbLwc$oCGWOdzpxl~8;U;iCkz
zF#Dwl%)-<ejpQE=mFFzuXtDy>c?|?wt9dBA7_~YWER>5xG1Mzh8P(R*g*m2|oneNr
zX(tBEpyC<kUQtoXi||lPG{O%zgB9QE(dG)X4%26#N?~m*pM;E;QC<0-+oec!<uZzG
zgcnW`D}2i^l;;d6JufDl-(l#(>L<b)1&(4;&ZQKnJ$@4cz(*oD68_2nq%MZGlv$Bp
zVea5jM!S(<7AV(y!~^{y;7PSjNEBdslt*x-#1Kp<JmAdep*(`4XZlqNCOC{3xRqE0
zN=oKY1kjVC<0##^c8}ZB^YpV<_j9KbsxrPEhEZbEzq*w_Z6~mGpho!Ryeq~<i0_Uk
zxbaHB>w=&axm>95vE#O}BCIl+Dgq8$q4umzMPViQ+Gf$142~k-;j8hj73U-P!>_Q<
zDhGTWxLU^Fj5!>bO<r@$Wvjx%r(EhRd`PM}qpkTXT+EMJ-Io3sBP%kyQ&~ULt_YBX
zcg|?#RG;%#(E({w9ax`x+9DqqSfv(_Rr%lq2If|^CHELJ7~8DGo?n_PW0CU9`IWX^
zrpxpNO_z?ZpDA^^`ts|KO;ls_k7C&kZ6X2i4-SsiRa<l|*LF_Mh{1P6bug^=^B!hH
zmqtd)@Vrurqv)co=n^_T7&|R&!JE3mlA1#&txQE5r)?dFrzXGmM6Foij_*6fA8+nP
z0d#*9WrXNPy+Ok*_CT8%axHdHfo=LEb)%V{qwNfE=qJ9<h{KUq=TGGs{=~?;yu8|$
zSJ!KS$+H(0R7{jZg)?G(<Bhhy`DXga$#W4o%l2t#E+YVAe&F#`<juiIe&PU2TlM+k
zG<>YXP88YUKBwAI-|{&yn3@$qBRZK8`#AN^R>aiw>}G9|D1yhfax;-(R#vaI*`=Gg
zzn1sr+Em7FJkKpSc@@Dmp85{k_Txvz|2Z;5s01qQ(fuE`z3uyHU*MKDJ$rVXcAq6b
zzF~39VdN&~)1)Nhk}l{ohu&6|XB)HW^nW?>=+)F6d2^ft<S1=o6bYZLh#&kcvYALV
z@>%feBKUKDx}W+Vwg>m`=6in!kG5^!C%4~ghle}GNAcI`z(M~fr)^8e023LG(Kv8$
z4pf>K{{JM;7|y}}S@>t*JmU<Ki+n$?_K3KeTigu1mV*c5?eyrNN2moy>s+|VwT~V>
zEPPmBTWd=TD{W<YEjT#W=F91`+Ggjv(+s~oKRIf*)7GaKM;R+oB^-%Q+v4hKVA18Y
zAXyosf?hSY<i>(0KHnIY_8~^$VqUIu-EO>bt6jhOM(Ua`ncdEeCx?4&e}5}5I<4{N
zcwp3$-&S|VVQFbKZA-h(j_VninOo%yt>{8?2As_AZaA!ns&evZ57`UeSUi>c88`D#
zd1SbmnZ8vZCu`3WJgfOkcB2n4xMhSgzbwX?GA}sRo~N7>u->n8XO|b3gQuyl>W!>)
zaS~bdabSNha_N4_b0t3oZckD-nsjA?6Nuw-u4uySbZ|a>wVl?Bl+ovO+RSNH?b8z5
zubuU;$ra3~cKR-QuObniaXfRRXKW@p&72tq7P1|-r<tGk?|s_-@|)jmfB0|zeft;x
z;$OCJeB&?Mesmo&nKP3=y>qX9{ZGHqe(-~j+V0+dyZ`V(zJJ(0{P1pcB3tNLR7Iv7
z%ec*@UT0%1w(Hm5Xz#xJrFM1WdYc#VIALqQt}|(Lr=5}AqrWgf);V^8F&szt$c}^I
znImp#aj_yEMbdCYITMTi<f3bUgD9h~eC6FbE33A)+ZRnrhVCo*WlFI{%QLd|=QsL^
zZda5+E@dt)E>vA~O=pI^_0}70bK`2W_|X=h&hGijm%m(fYG*lWGRM4o?b>F^wfFAb
zZ)Yc`nU8d(v^DdTQB_3ovTggEWgEIo&$7FLtsOb6=2NeEcx~fa=^3H`uWoEq*+<cD
zx`VXq#@b#{ddrIyg|c~VqqgIN?R4Y4-Tk(+dr*DH)694HS6esc-R?M=4K1<&gN$z>
zA>(bwQ}9FQ>3w*qv2mHBaJ062H>N?D(tpkMl%&tbPL$Qji3m9cUxx4T>i?CAv7$N7
zXye$X&y;SSdg#a^>1-Wm8@O^F1V&q1+vQ}it)(q)orlNKx468})-vW-H`XiPQR);;
zSB7!A_;WO{=ZAB6we2{sUR_TeoUj~!ZB+J2#$Yb|F!j%}H7o?C*D5bu(0*{`z_8Wt
za%d>s-O&Q2W92#AG_EyH(Ra_zFVgnC+6HoO_bBfNuhGe(FR-wu(yekCQ&-_>p6gTL
z>L{Ok#<PERK3rex-pH^nf4$JhufFC#_mx*=_LnMknPVvX)u%M%EhorOKBk!qe)Qk@
zneX63p6SklKF`!zAEV^n%lbzDd0%q3-l>1D^7J-nH~QDxTW!jimVHB?MmCcw;k7n-
zzu=<8D$@6SsrC+zm=2Sd@>njk0FKWFAB!KSE{g@&LsnKqV_peg+^ab$npMOgo&+~s
zXdDOXNiAsV&g>~Xt91F1?I`tfyp{b3%+M!IOrtvu9MP`)Mn8Qw_kGtd_${CAsFTyn
zyk(ym96icL0|(sld~lZaaqZ)wY|r7az7059`BK*)^8sSPH}FC?1{zk5ygVk0)_YQZ
zd*3CGXJGEBN6M=!ZSx%6b6=Z>HlPU>25@HMhh<+$CbEzI_tvG1{NMYA7y8wmT7zsZ
z$z}es6IqO1yFHOZv@P|aXW%A3`oY#Tw6prH{qUwf;bJ|lv18Lnvy*Ks^x(Wb4!yD<
zpu-xvy@Jb2_*C2d-`o3olIQcTAD8}m-aZ#zCEx$%_y4Y@ABElL!{YPuxPBSw^XhTI
zx=Y8`Wx7mX(6ksq{VTus+wE(=`}f=C^=s`)gp0E%YG+X(YK6++gi7A-yb5CsQyL8E
zGsE{VEQ|*uqFfMGI1)sNNdyk*lp!ePDicp2)NZV=M+jal!w2Sj7Di<S<cs8CV6fT~
z({(UB-X!v*oI8v;Q8)-Z6b}rQBTm`HAV+*`5Kx}vKE|pol(-V+(+EO2M#|8l9~@b=
z2YTMAa>GQ+Q5B}j@b#QR62+tomAL{J#8BgE#dl!ha}d}h+w>4N92#6_WhAe563jRj
zC|!jAo(7@M_JZ9B!Qn?CAXpg#5kdqi{V^W$5H0|<93p|SzEyprd&;HYl%SmYDK5VE
zD&29f3s3|IzkTnHw^ssOMLLyGlXr|)`o@^3k1mSV(ZOjuKZ!s{2m1vk)xH$|k_f|`
zQFT9Uu_9bwI3V4m1QI?NuJmbGA!MbX1ZIRxzm#b#{O}%Om{3RPt7n06XMFbi?#?df
z(io5PlyA(8qjt!o9;;`~iDv->g?$Fj65#XPs(u1strpK1_Nao38gtPKCH&@H!l62~
zbx;zjjg+(0B|k$Cf!lKr3D3s11kmKMg_1rR?aGrcrL07e08eGZ_1sKbx<X);BxTjS
zR@dYdBVq7hD)V&gS=x2ho<2btF4JZDLZ(Z{*Uywt5!b7aAD7UrGI-=m|K4dbgEu~V
z^TxICNHWA!ZS6HZ<6!d8mC&17Kr+#zu&pXBJ{@}17rS&?h^Q5>%BScje0O{d4@+ga
z+36|X%IJF~&(V94WKYUb!U!ranyOso*R!)8silshyJH-7-555|PR2HM3=3b->p?EU
zE6`B<!+9%=QM0oPZE<n6Ev;;}nZ=E^w6xL|<{ABF!<QG@^x{fIxi|~!>A7tm_sRe;
z9iD>Zi^_35eZrW>aMztFwk2dTonY8qSgG@!@PqU4<fl(xw7HZ!F|(RBEax4D--$LC
zn9tf0aMHGTB2a7vHQ(l@ML|tfG*uh(jI6?F!*RvI5_t#Q0pSIQ!h`ns(Y^HlyiJ5B
zF#JC`J8HYzkJ_`eEkoZ9j`jji21)Rb%#v5}JqvKAEWXKQn?g|o;Pv3a!@zH<t*%~8
z8|UlX)3LFq?SgSW&z)%rFE~t|KH05ywHL_~{$&hycXuly2@WxGiAv(AWK^}v*y5!t
z`EC0zylrKT@ik>mkG2S8vL3U_xUWXWKb_rmu)kenIyb-8);DgpmDSDgNY19o!rg_X
z_4GGa;KuRfOe{t-PNVg!*Mf(0&GuxrmAiTKjdDz^uB4BV<rtbp)-mjo8?Q`9Mo3GX
zL-zLD@8<pMk>%#o&c(u+XHO$H1b@#&%AM{9E<O6;`SV~x00=?%zIZ-|<a<B(Ry#U$
zF5H!FP(C?r&NRDz{mtsH#assmd+px6JGD(5x%9@3H`|+Uf2pk{pYq5{vyqEMNa;K0
z$id!j`q<~8Ia$$%!S^n3ySjW5M9F{TevTFhOc*3DL@qktYqvjsuNJkzCF5kCn%Cqm
z&e9a0BAg4yDY+1teW@+X<yu^B^MUo1iMhyEXKjD?ar$;p5h<cL%K#SK>(O9oO5jlo
zwIauJFpz=O$yv*AH!q@cdMSP43<$n~|3u_QPTLDk@ZeF#qZ`2)-r$Jx$c5cuGd7lY
zCYRgH{8F8*w3>@U`+3GGttokZJpCJB@gnnSf9GCB*Zs+#e!cy{zxl)V^{;;;WAQL@
zpq(QaW5>&3^2zOc?RKuS)4)3ThSu4#X>y`(J?9*!TQ_gC`RGz~FOEA-N?YyGuOzZ6
zl332F;2H;nK>{0R?x8O;>7zNw;j|XLuseUIDyo9xX)T}Sq=65!Q-SwvEnE=Y#K{7O
z(eu^3tG=ym(NlXxgS&CK>l0*3`h*1!7HWBJ(Vp@|blF|NHn<Giy}p@t>((1>a}%$a
zZI2#33f>&GH{X0S`LC9aN=ENHf-rI#r}5zSacyg-E#!616x-w}b1{9?FLlCYaM8yb
zxxmaBUHY`LwvslFl`gxnxe=OU&NyGO%ldQ=`lLl1&h)ZKVq<f?=p08JgZ$|k{?N^;
z7D;H2-7(12_{C6;t@t=t`@+Uo&MONU23?@?&v4K=XH4D3BIB7h(nHkCacxXiS2Isi
zHvJE7=pW}C8f}|a+cCnQwN>W<PL`wc@ZhLmM?VICw8Nr3bKll;ckkW{>W}5!(7XA?
znpgNK$0!(ZMw$<t5ooZj7mGh<PIj6Ryvewu`NlvUcyjT<(15;RH9RZxWqm!oEc}nd
z@n-NF&e*D$Lrq)A5%9qoXml^zFVc}MWI&I|7(Sg@#}T)a=gv<P*#-|wM(F<vH+|~6
z?*p%|`zd?4D!<R9?u7DE{^d)_Im-2!JoM*EDI?vlvWE9xUHyH}%4s3F=c68u@~3=#
z^u5}iI_l;8Ba2j5@*9qnewgp{y3&=0GW;ng&->zGWp;g78x_}^>_4=jx2xKn<Z-=Q
z{MfImrv7``uDW^~;OvlZfTKculIeQhv^h9N&azm@LJa50$)n}nd1Mh(Uu+GZ=YsFH
z)8ynbS2%0QTb$M+mFU6f&SB;rd=OcQ_x1VKcX834veo3$2Isev(ag^-_XQ>yH1n_=
zR@MIWBhLoh=soMulH4n~CV6y}Edj2hb9Czt6ZO4vP)P-r>Tk*jZ7^SYUT{u3v{<Gj
z_1D1Uw><jLZ5SlLG&#@ZDd*#;%*vmnjKOJCxazkne?!Dw)tfTOzaowChrChG@Joxo
zz|6d32V$!m-m{RacF9Ro%T9*3jK=**_+ZIi`OPs_y91F4^Pcku(X^5e%MQUl(jAI}
z-O7yt53BuYlXiH2xTjoBZZH<bzrC|tGy#p$K4n%}qcp(u^?R>Iv!_>iewO6j&%Xb8
z_g<x+uf8vS`l9roWP9oOx=fep3z)7%puYL#ciLBf?|0h9wQFrMLdDZCD+~$Kvl6ga
zXPBfY7mSH8GZ!XR`)r5N5ppQJ7?72i7(C$%Qz`-<LWYrHCzJ+4$wic9MkmkKtvHMD
z?Y$B*Qh%MF6{bhIwC#|s?5so-mD3|}MoceP8JR`~4-EKOm?PnbAVUBs!7ugH=SZNi
zicR~icq7!?f1ZF{TP>yy1ZHi}ZVK<%h=CIVif#FxKi_M0l`C}+BpDAV3+iD+ldo0`
zrO(<+>FXi;SZox*yjKxuxev~%q~NT7B?P2@>Sd$=Gx_Trk-V?J1W3w|`cho-Mkq@e
zgn%@sU{z5;$?MQ&d297i1f5d0I5tK?g90#~;bPdMBx;9tffYCqPzg{BBeh?9#>aLd
zPfnlY^GQ28IBREz5deSQEe?M|QgR7g`7F5R8icn3Zv6w~5&R>tmAI6#AvAf^`<d^4
z7S1wob~x7ad=S1Wq9D(~fwD!(E+;KGClkfe3RDTaD7G~g`Awm>rKDdBr2Sy<?u4TB
z6Ux*<@VEVEkg=3Pj>{5Y(?%;-ODK%e7|g2v=7Fra-<3}DC$p!wKP~rez3(q=FZ|C{
z3R=c)5R}ci!l^6WAPnD<p*jy9>aT4`a+QJ*C8Tdd5alFO|LHR;>(XCfzf70ui<Q2R
zj<10}|FqXHJ}Gll?oTYf8w-liB-LVq@DRp~kfL^aa@?-2FP9+}E$qgB2JkDTB$F)|
zDr{A#kbZV!r&qS<L$4G>X26F^`JTWFBd3gagNDTf_|#Ai!#c-;cA?F+poK9zG!G3_
zHe-G%++((s>P|0?j2<DufQxpEbm-f_;HP*r8HfU}Ec_HrMt3=;(BOr|)waBRwN1>f
zrXLI8JLDEdwjMQH1~cTlSukX4!l!MXQ8}^+C&%f@eyIPmHZeWhre~L{4GilHj2Y=R
zJ-3+m7xUgs_z5S8Efqu8EsQ8*{UDQJc_G+gUfR{qIqCz&;Z_kuwMATaW-!tp?QdsH
z4%)M)qTX%$6PfJEQ9C<1Oy6yRMyBg$zml`2GR9;La?g{er_J_PqSMJWRa4+XZkwB5
zD#NWvJBH%EmG+Csqqo||=FQ40@`)qE)_RM!!kb;GcJFN6Y5Ti(tDf1V&D61)zRslm
z{R|Ih(b&#!ZR1wzVh}qicwF1OQO>gm5AKJ5owuu-w%lAy9%I*W?dgTHZjLfO_~dvy
zKHO=Chud}b((?K{ZD#(hHoI`GBBHv1$N8EZUK|b7V*E4^uRdn1=0q#xT1<cE7s@d)
zIVFmXBc#>&xNsi+*CHJTwBXtN!g8Bmxt{uNRTQ46yS?p4?ezGd*-G>1XfLw*QMJJ~
ziP~T;ot-&D?7JmvT)ld&aGTM2X?d;9%r0dPavYx2nP30z-~C~;72`YaezmQv-^g=Y
z8cx-Cy#L-0+TQL~n<Vc=u5r%M_V!l#$q@A{&vwcfJw2VdmwDBr67yM3FUhA-bP*D4
zCw(4Saj)&}J_-zH+hY1?5s>zq2XO4@Xg4x(Wc<LWpIMf2gTq}`$^4z@aDJZgJ3iRX
zJCUvC>_Xv;RcB2d7pz1%jZZALshPk&eP=vnuqVS9`*b8Q=lB}7Ff;*!+ZRXe;BdD+
ze*7Rf*`uL$ceY9npP6MOB`3ibeMaE)Pv6PLqNy@xMA7YSr7s7CueRp1ONKoAyLZ}O
z{P}-u-~8q`+qb{{{gi#!Z1ec{e*5pYuYK)zYpkC=x1F2Cabs;h^L>5&TCVk68^Pb;
zUBC%H=jX}w{cKsXIEOM?g7z?0(xEhu!FXe1E$?@ShVd}>IFH_W`_1;&TQ|$uu(p=E
z)>gXnX=%Blr#LChYqB7kK!>3($+wU(bboPSuB~5PD?>B~iRd)vevyaGQDra$bLMmC
zlnxnsS=-U4Ey{Z1&0DE|CGA+P{AjL79`l)eZ&!)?_cL$PCy`!gJo>qiIb|!z{k?<0
zgmJe=P8okX!C7!rn;BX4o%4><3r|9GM3!w_6<ON1Cv}gEs|_Naj4@p+*jN0dJ7@a>
z39zlL3R7n^tJ3B0ot+)KlpMAFgX3H$%@%?5B=e>(gy?Okg>b=D@53ENdItR=5BSd2
zIdFC->07sLY-|LEq0vFtu3n&0!9d3LJYQW|&b3s|Em6brU%S4MKD2f$0v7x;UiyK5
zY*i#wXn*SEd>ahdB4rr!T^yOVOGFz)gMsnk(P?N=Xmi>MUht5!91Y`y#Fz0&It@Jx
zPC7U3>ebb1yNGCQC}(u)I}d#9ezBN#gJUiHnCUtbTwpLShqZ;4Pv8w&{j?S_+`s>z
z&IQ9;M3EWi&BEQ4_Avc(Hk$L($V6p7$hhjGF*mk!pAJ*w-RD;6@uTZ?S;}|$-0IFQ
zyfpWG;^Q8S<et(KgQLc1NXa!SBYE=&M>z*ye>NoFd*0rj`dxLqdi(tq-j8r!Ql7dw
z{FUF!2n|X(ulG$EFH?Q*P6qE7o1q`RbjOLS*I##&x3{C0<rC~Ve(1({BO8m{o@;BB
zk+u%~lHWWKu}ilmf04uBlD-zL$$W>;j>TBy1hW97;Mzhny2IQt8KSJ=hiD-Saye%l
zje&;Xu^i-ZNQ2NFeI)0ROU?VziPJvi;;WKzHGW;z>j3e;zKwuZ@G6IO^11`YC7X<J
z(S2hC$1L8iGApm#{eMNn744Gb>1`Q}Wl|saj>BX_xT_rX%V&=Da(k-$`Ruqkx<<eC
zx%!?l$erq+`b$<1T+CTK1r2wpL;KMpQIu?Y%6r+@jE(Q*1Pl#e+aNdKN!}xSE5D-R
zQkMK2nC|N*8tDDPnUtX`Wn)aL-}xMRR$G`Sf64G5mhw7WILOc2qs;dc3n2oQlvZWc
z+anlA9d~`tRkwZ=#y?*gmHQ*<)xDvP*RNl<^m%RkWnN#D1S2mUUzh1JebG{Q>DrBr
z_NBl7JME1xzf(@ICt)@i-+7b@Ober7T$&EU;z)8v0KuA)!Ki{qP%=eQ$+O6K5GHkD
zyNr;AT1^xthMet380ki+lpo=PA)Nu=>N+c$F*=Smt6c~Wb)a?1eHNxG(#!c-gAiVh
zwFuwZhDmY=<Yp1hsgzmFm*R&J=8q7Ci8I<0C#)bU#aH=htF2+CrnI+3h`uPpi8lLf
zb#Mt`xlb6tnDxmuII;+MoErqRR9x*Ks|1?8j=VPr$&^M8H!J7}K@?(w3#N<-m(ZAJ
zWds{j2L(25h|mv4ln%v!yK)IFE(WKrKmvsbs6LjQE#OLl=qp{*t_U*)s|Xv6B$Nkm
z)rfAyVU&s>6Xl2U5xKdYs4wMfn|Cnpaws_n`HAgPa-E#F<HOT-v}a4B?)>?sCGg5$
z!8UC%4rQdvI2(5>jzoiju~lDE{T>N;gsEb*DM$Wt`UFP4a~PG9kuflSFCw&cSOw1l
z8}~|(g$u5Hhd11ueZ?{ZZBNDp%;2G@FWara1$f19Y0S>xbjF&4k{}9y36#bU(omL_
zVVqM;^1y-Ao4*>Pl;hKCIc+yi+F3BjI~2OYm%=aD)A6hua&1AvAe=S^ZyYwx!KrRS
zC3xsVR|bN&!F6ylf54M5!8r`2g#Hp8otNn{{Swk;#MfsN^6)n%nMN6ZLjHOb`h}nL
zt1I;QUEdZtH1F)REiEs$YuB!YHi|Co$}2i<ds*KtCg2o9YZ%nrM`z24GRj-DGO6ff
z@#cQjxh0Vs<n7zz;R*d2+IMj1EFLmJcbFAVNGloB&`0+sM|I&h=wb1TysusIS3g7B
z$p?MA<j&8@DBkJknBu?7<;b{_x=H|xeBg|$YV-75@nu^Ris&LsEG(_InYoR8@A8oa
zKxp;&<Wk@|pDQ$Yf`PzUU6XC9R!TF(<8f(K%6|U*jKv^4N-H^8^7;I@U8EkceEBTA
z^T~-bkRG(-V+K(CEOOi!{@%BSJUc%sqta2}iN~-9GoUh{I1kF9?s$>t0{4q=r2LsV
zk#0Teis4$l92|=)>*?cK`ZLu|j*i;a)>e4pY`gi!n{9dHt#)N{Ep2$(PLFL*Cpv5{
za_VZ@%E2@fd1E_uoK)o1m8q5Vtven#Ca!Q=hF48H2kgpnU}L2?15#fAWNf##`<9B@
zJQduqaI0_I7ryuGELj=e`nE{VpPgoi&)eGSdYhP89@V8hTbZSt)5GHJaIbGO*vAFW
zv@=29@-GlQk9-t4xL`CrpT3L8A4}gKx4r#?a=<Y(IXey>+WN0Y<$0e2Zo7T-!TW`~
zqT$AZo7%r@o6PmiHov$T*e^$hm}(pAYwhN(w*%)}ZDILpTU@fk!B{zm#si0g{oVE-
z{^&on-~au8-By-Y+duv%|9N0%`#_Ej(OswQ5B}iy+jqYGSAkc@`7#ow%_7J2=fQ&q
zg)2)7^KCY5Z*|_6{*Q3561l+qWT3K0#@3`0ZEfRvTUomqU1BjgyCILfDawjddf1Uc
zAA6ZEf+yQ;e;1yZ!+~AK$4dWNRUbHmJqL{~2F-KxZan!4ekp@L?%C7hc66{CT;5CG
zF6&LCo+}e+XXb&nGN#c}&<=ffChEb#exCJR6Uftb)?Kwr`}EJ89&5)3588LX^Ue1C
z?|r|0|NB2^-}~Ns!8J~h<Fx-I_`KKNfA78ajlcXxITL0w_cu1Lw>RE+yKUxE#1{B)
za+H3Q@hzMz64xUYM8;UnZfiIFg%2Di^ft0#8ALM&Yj=py)y2iRc5QR5y>aVCyDE|<
z_rQAN>U#UqyYIAj-+jBb{Ua}1I9O3gsRs;=fvAa{o!wd#$stPK6@>xkjn#Jg!Qcv(
zqJ3;jZ`C~~c5O+QaX@3~G(o*8lbp+mBtLp<yL)pMK1m!1obc+sd*@!!2is$z9}KT{
z39(%}hqwrj?*`60J3AHWw6S@$tp`UrfkeJKgA4A9N@8GVgr&=I&d`t0V7i3GXJ{O~
ziA*i>NM2asEI%@_v!NKy$?Ke1zMC)bx30kc==iif-U>b4IVk6!vwUmpqGwJ9#`Fj9
zH>b4A*qZmYKa?LYK|>k$&D;6;?)<3~xLAA$?yfrLGH}#RxCu735HL85wwOmt%S);K
zd8aeGI~BFY;l<Ikv$NkGKH4g$5WWSM=)|0IV2y`z793_w^})8R#^xmL<k<5Z-mn)O
z9Um26=WNqXd`V;-rv$v==z@D9yf`5q+<#b~$y&4pj)($to}&K35#?Hl3oqo+W)8UH
zWBnK1!$G~3e(Y6Q-m#VJ@^WZ@=+OO#kK4x|e;OP<PP@CYU)e<;=((vA&clmxGU#Oh
z;d#jyCA+`?I5~1e)c5`yOwf@+lGNyWc|Z5+wlaK5nHTjq|M?rrA9U}R8I?aO%T;nq
z;H>_-mu4tuaF}*n8a>PX(q;2Emil^;${rP-=P&zJ_g?16<$nL{_eP(&*_++lP!3@?
z-1q6LTX_cm&`5Yb$V>&I6b0YAb1>hNH~%W1OCQPx6xo5jhC`a1#o@)73(oXwpXvfL
zWr_OZVCrQW)9y%t%WNAW<mD~r>!=<1Ki?zw1joxk6FM)y^1!rmBwzKdu0B>080(X^
z4h}|T$>>?-@n>Ap9&JG%D@VRhgl3ifry3lb%13`T8oRVbZqL2zY1N$H_5Mrvh9=}e
zZ&&50_v^jkUGLPBe9cwx&waVsie7j4qI-@p-~iWNbQw(0Bkw6dyX8L)zcKf<*?5<2
zAux~^e=--rUhU|Zv-!+T_l9F1jIVn|H{o|a;kmJ@_Kxrw4Jf-{^)vmdHsI+NC#Q{N
zpXmJFIfk}1-QM2Kv%I6-U?j8e1%&}-ZoRr!U98`~i1e!5N>%nSKmT@h4DbAHuOWTj
zv!CX=OrM`fx0jBu%XFE(U}-!8-&=q0EA7j_{rB4P`g(g7h0mE@m>foqv2bExxy6t$
zx7@Rx17!e{KRk-i8^K#XtGzIiTJeJUh0SmnS#e{(dWTDPMtI(iP9uQkkFcqpF3ez<
zJ+jGu>JB<T&8KY+9emDkk12Nne=16L6q0(M@x5=aScd<!*GeBN%PBLMbe&<u>5x1O
zRAd2)14fOszlcIEzp_i|&;M3<+D^lnBld`AbZSxJyg`^4+Td)Cnb`<gX#*jM@PujW
z7YC90tA8V9+A3;W+E@WcupqcIY7qi3eB6;zX?weBa}-MqoihT1&&?`x5D>tc0tVJ4
z45eM*s9eH_{#9G&tdO5fKYMghZCRK6n44AE-ghPngAggqDtv8HuNAN4km;18PC{a>
zEVYX2;(1r>q7+gv_qO-j*)fA!D*k0B!VJ72u%^#l7&d}k5VeI6ht_!nxe^v@%qcU#
zXq>u0(2b^6TykIq2AKm@eBe*;^DbqNP_F;rP(ohb1w%rhay-k>bQlKmelynAIHj&Y
zs}wtJ18=y*pkmG$JF8eJF6IRo6Euy1?K|C9Kb+y5@~%F>dxnwPekR{3?v!B$-ZHvn
z%;jT<C}&n02tO&r?!)DUIjdlkKll8GH*kn?In|b6TS9s+xC6fW=y#p>ka-|Ff|C67
za_hs(^u<Y+j<3I#LRpHm`TQ|g{$8fnzJ4~zD}`hg{i|)lLMPB9(Kw8B_wV1UMKtfe
z^Hy6NZ95ePRK_j5$I4JN6K$jDqs_(FLaS}pg=QC>4UM93d%e!=$vyO5X=TU^ZDrsp
zgLUpQmg3d>`+LcY-!U$`PkFYGWE9<i-cEHMEa4|sB0qc1@Qc59qo@wrme~*dKpD=}
zVi2{!!C5Mt8qN?BiSptZhg4rJJ3h&YHD7f0<YYIz^)U3@8AexA?qV)-j^CVf2=d?v
zU=U!$w*_A}M0I17wuX0|?Y7;W2a&rD+t>?>DvaOL_VheF`eeVIo*uP2U+H;x(D`v-
zA$n&&aI{$FD7-IxQ-@x3zM#)X$C1m@Pke38n1mOC)!yDg$pp@UWHc9rba;G_d^2rr
z?Txm$ypj5+y3BZXUPh|Lm5s<ltLb-O;@rsNlg6>Pu(F;mh>F91#@fW#ll+dnQ_(vj
zfO_AZ?Zl8gH@^~;F|OSZ0tcR)i%yG7gjYK&D({SsKZ{&>kiMU!u7h@d;ao6Le<G`v
z+Cr|0sUBUV-Gx6X%-Lj*AKz;`JCEAZ;&Pi#KgP<D7}?|LQG0Ob!<4%d8KFA?CMUtm
z0+isX@fDJxQIP<S4tCn(tw&{)o#sHxSlU*NVTUZcvdp1m>tx$mK1mx!=VJtCL?St_
zjG=yG^G4ge{#ILFy`Hu+<})5y0nPEm!N37HpZDR9dmJ^HPbYhk&-U8k!EXEE`#)^|
z^Z)$+ZGZmf-)y(ue7pV5@BF>YO~$U^ec)@A{P(~2o%Zc-f2-~6?v(NToiBZ*z4_Kx
z+j7Q8yErBo+8HLNr{_|?c{DQ6*UX44=q#*qdbyvm7}uY`u-a<Q&(t<=6O)XP&LeXE
z+(g<xmwC(B#Q`1K@#3uA{q+6zNv?&^in#?2NKR1BL5qQ|6#ityDiV?GG(D9zMYiha
zYGn+bAGF=apSHVqK8j4V8<<<n(-)3C6Pb90?4CAVSxEg087Et6E!S3-7Cymki+<1-
z+b|A&eevQXW7w}BzW1&6`~U0z+W!16zS+L_y${;^KTLi1AC>d#{@wfSlaFtgq5k>v
zv6R2wW}*-1C+Eo1XK8)pPKNoph1tl0p?Q(nZSBTsU`tW*7hR|v2KjD#Soz6sbTtkO
z&VzCc<r&>=ZDpywar1iMW^t5-RO7+5rE*GG$jEudiTUWk!|IPnFSPD(|FG@uiO$Ql
zvmZU~xGn}f(R0pvx_|#+`rjk3w23hkZ3ORw!^orQ*ZC=D#ZlYYw*AyWIlwuy>5IzX
zK;q1BCYVSnvLvUY1z;A5ozpME3nV*^hz_&$A{-Uv^U0_8(!RZtqu+Vw&EV**jN4M$
z^sJmb^iWQ<wbhl(7v~yB#|!S_CG<u5DF++5mhn_yj0s&0-F_N9M_che4p3u3AJbRH
z)pGbp4i=rnd3Eo><M!|Y=PKPPJfxgh_*lk++D2E&^eEjW7g#8d^Rhes`r<?5Hxc^a
zH(Da9LLb4*n26NuPO{0OspcYQVl5I49Pur3{~3pS(caV2DMd@{?H!gw^U&7xN2xcm
z`(WhmcrJ1x?<LQZC$<%w3jOO59p2g94P3*g@ltJqlbmCmiE!Dru6Xy_`f|z>VR}|=
zK<90%*d1>@8WB9qS9<Dzv-(@xsfDJ(Pi^LOeDGkq-Mzb&wuOJ@56&r%W6c<EZf=x(
z0cV_#_rnivx9uHHyHLStIx?LtZFZ&Ibm0jLSMyxowZ#P%J<{-%PJ<6>R~gOoS$bf`
zM$(p&bw<he;d+(yEA`~2H@(|&q3f)pTQ8HkhD+ULvrs^Px$A79VeXfn>6w(W=y`B2
zQ?s|D3Ld@d_ozOf+Ee=Ts4cyeQCU*5%2QuG4@#Dkt>3Aum#2LlM%|H>jJa=t=inIf
zUVoD<BQjO;fj|11w+rs*y@*q13W~b)Y0S($_9Qwn+&71Hc^I=kmYlxbp@ps+B2G<n
zS=6}*elm-BMh~xXX3NM`ut+<V)onSi`aGo9;orLsviahl@1v~r>w;DCSN3|=uez7<
ze3|mw_?0}H`|z-IZ25A3sBVC3moY~Rzx<8L)$ZKt?-jnhs$;lUSAO?)beuLOU}Aw2
z+SchnZ@)R9bLH?#orPyf<}uq3+#^@(HhSS)z0+K1tM}RA(5+fHoP>^}x$4DB>)B}B
zlwA(W&_c8eZYZO+I@gXMvpmihMW?md?l|mCbb|Cg^y;KNe7Kc%gf`chsx|-V|4T@|
zO5l?J`;T8}`b@d`+Ux(@)z#bg^QB)z8JFo*G6OChUzh1JeZkVy;#~XcfA+iWov(bk
zP0h_#G{_L?Q4AuVWk?FY5Ix0Ej^UJG5e8U-Mi^AYn9hg@IQfnd5#%Sr)CjEvaE7NR
z5l#vAgkxJ4rpOXTI1wmon73z~UACoQG<cExR-HQAfnk6UVm0K->RN3XGr_Tx_pR!q
zXc05^@X7!H|MW>jK~(Y>$~baLC`+3$aOFxA3HPipsXfxeXeSAcDYp!25u$7{<2!}k
z_H`U!<q*R+(`K-6CWkUGZIMRW#YpiYnJ{k*zQf5X@i5+87&s*WgRkxO0=sfVq+J|S
zp5x60KSF+fGxkx4w1bew!N^!>#d0}(Mq{M@Vrv<?F33fwNIPl;Ghrf%6DI~CStZ1H
zLWFuaEQ~ui5iF-7Oy~n64EVr>+OIk9bs@^Cd)tqk)#}eDQ7)dI+3HZV*Ks@CwM}G6
z8o#ns!hLWqz$hW91iIiN<0G+9-?Y6}Ucrk9Y9-vJUjRqI1sU)uVKuM;AD3}w+_U-<
zyd{FCI=w_l&5LOS_~cgpD}m7;ya+Cqkp0wl3m3tkE49_9adLKvXB?*LP$rxijES6A
z`cP4hnRCkSD@VI?jUwJ7%z7CVjxzFEVH-S46Y9M*qaA0g;29wkEEoiJ!5rqef?vi{
z9onuPRvg#Ls(gnJ=3rNj%olV5EK+8aWGgzSX41E`^*s2X?#pzUe$CTwkmKtoO}OFN
zPdh15!=H5h=+_^YirtKo@91N_6M|H-L}(G3gBD-gT(3w33Oz%bWDx>GhkRvhP62o(
zgC1r_F)l|{-a#|#9dbix+Mvv%Tb_3#XJ1&*jRvD_Qc@1y<$Em%i9*eZ08$n{P`LB3
z%EK4YWAw2br+YN*OP-Uyp{HfI%nxls!$l>mtzQqFy;{*b&o9m^ib=j|<HX>s5LIm3
zEXIwQ0L*<(E-8gOL+j!sJSnY9dpJUl4<EHpKe=7)WzY#>4(DSCIBSOodu?ZDs~sHf
z7aX6f4=2Yf@T`14d(v0s;x`O|ON$orICsYuG0#fgsx4)7E|+>l)`+rkA4|k{7MEAs
zwd*(AQa;HQw!;J4%}vopH&WknU@@Qa$VOMv$H-1Iiz%%;vN-G}Cvp``r(QZ0o<Ftl
zXt#JyIXjJWIsRU>?Z<Z`D;|}B+xU@XENnVG+6@n~y<}S8mXcETljOa9`{T6jaro3p
zaLV~FGi`m{xmUN^?1D4U2FDsXz*des?O<=a{on`RZ9n+__bQ5HHu7KJdNHz7;P+SG
z{6>3t|8CpZyjDh7cseyb7rAR8W1D^ibfU$cU2v?7*S1V^3s(cT;8|pWFxEDgKCG_a
zY_s$0RqpoIz4qwQ{fueaC;})l9OtC%K^K>n+uFvBHaVR#t}GY+a2)oi^IR_?FFy^Q
zfQNC*sB6UYi^KNA_rBNu;~)RK_Km;z^Y%x7^hfP4|MHt9L;u5n{7>3fzWk-s*Q5U#
zt;nkDtE**D{^;Y6GUoHO72opGYL_`8(`;`&Zu|SYHLb}u(5j;5s^0YNg~e*-E7^N+
z6uGo1?W^zdro7<x{d@P?gZmHiUT~oz<*ua97M?j{?ySaD@y;gdTOBFW+#fjyUp%cd
z$hNi~rc^sXqzntz0*kT0)f~zYWS;jZg7u6A#}9{qElHh~B>GR}8i$<m3p&<)QB9mc
zXaQ&Vg!ZNVFP@*Y2lqb7Jp3^2|F~^EdQ|dJo$0r<7JX&C7RWiT&AC}OZ{2Lyu3c}3
z$Nl^nG|l|uAnL|gjvw0=S_F99j!#9vSve23JzAyHzV3Vii*jBCW@y+D^&@czi12}D
z>K+#Pk%<{Nok4Z5f6(^!_sbE%aLJ*-@aXy0<DGH{iRP?@ed(7vY-xCPeLZEF0~c*`
zV>5Gov*@LFIn7Q^Y|naDdJj1je(Fc9-p_Y+m292?(Qops^QLUeSI(}`Wl>;PGe$7w
zII?Xz8Tj3(?HB1t9Ecx$@KMI^tQHrD_Ohj)Gp;zhKK<lQyL;zOUVC0!87jNBBM%I0
zo5yh_Itx#!vjzjZt5??xcBLx@W~b@Ly?YPJf#y9?O=Nt0iVR&56X6YTNMzK#2M^1+
z$FW&w*G=~`&H8zcp>@F}krt9MD><A@oj!LRgD!!8aOC#!=B&Y6IF&dFMcv{<qQ>yY
zG9G6>;5QlUf+w0tZ{bYSX8cPZ$?V2h1eEi#Iv=u2NY`D;8Jx>};($i$;n&jgLR(qJ
z52k{fqT~9a0g-9iN1rd<DD@jVXDk)YS|pHr=J~J%vHN%_U0K<y8TZw-)r$1$y8QEM
zH|L)6SMr=wkz@4k-3MuZbey!qTt<uVU%cbm^=s`bU-_-#Ti<{0{r1BjewcAPsky?L
z$9bpTa>|rGn6~69dX};6P7l{G7FccP|MKD2`BL|XBM*i$ibtkr{ZqS)Z!ZU&`bAdg
zpZUN4L4Znr)kmdQW%lo{r+OAR_PX)@d{2Me8?p{^Nu5zXdS~d1c8zMwcbDHirG)&4
zJ}2ej%Qn=g(UE)3YC}G|bEv;l_&n+l-IHy{BYGkHE&bbHFS4G&2RKgO^c#KhS@xmu
zAkWYS^avfQh@{AQy;mK6XeL^THo<da$@#*LqZ|=o>_+T179)XIZzXwXH0J3?;eXnr
z0cBU|nUV%r_8UI+zx01_sA~i>c}J!9@#y#KUf=`%Qiq@CV3l^KU0ue+Q@Z2LLon>+
zgH=BZt?W6F9~1^0%cafQHsCP&irx<ODP-t-^<|XTjxC&C70sJ>*gmR{`Rgfoo^}il
zDzDP>y@Sr6oBBnTCO@zdd0!i=-d;rR=~w7~>R|IS=4hAx%kMq3!<bYQSzxEUvc-`9
z@($ZZEg(*jWcr#vV44r_bLfj=<&>|5Gevr|Vd%^7T?rn&{_N*XLs>(4#;uRtD_9Pn
zKV5nWzn>xXx_|cf|5GPzA87Bd({<_ix=fep3zo(sOkaKX&GznZeKiX7dV3M(gau%f
z!}(eilV@QfNG3)}SjT`dy;lN4m{9JOaXtAEgi>n4Y{(9!#Dy`rhlvki(-^cBJ~Lrn
za}h={I6|bArWmLCDL&4Uc<MJt4n;Xk(28j*Pp3<X=nkm4+#|%Qi?YLUg;7%8t&S~Y
zecD$Jr3ej_#YK*<2vj(@6_vI(Jge1cRzcfiJ<OZ2EQ-dnd>;#QwYr-kv9P$<Ci4k4
z6*Wb;$;Am)POn_*8HCJQkrf5L9Brffn6Y*hT+{Zh{6<JgO9?=t>Ien4gv9t&*30Fv
zaIO|(!bmVF;4}L5v(uF2-CBh+5+M2*jg%|GNA7n)tiR{GcL*TL90W0aR9?VT-<=C0
z{}~1O$VzchJ`R>XZ3jD1qE5SG=2x8vO(nplZ|ps#Tm`P$>{*(f?-a^zFu0)T6#ToQ
zCY2zmqrT_zjia*^tSDdRhHVewAtOMovak8TLDmIfmxurnXp8{xuY^*CrVYls<^x=e
z0tS}G#Cdt%EoCqlfvpRA5#r4!IC2qXFoLG<3M20k8VP?^Suy$^93ItJio`PC3Le#_
zXN9}Q7%npKc0tps_=(zn$x8lmmZo2#DdCS5mIa3Zq#`YJU){L}KR9v0pd!CUQE_nN
z()o3nz5wa(M8sF7Khjf|RKH)l>fY%3sgn2l$LRX}zg~F0`I>M3E3-##beVv&K6EXg
zvvX5zeRSY{ceE=(H1SFC%yJ?XtsX1I*y?fgvFLgDvPciKhtcHOvu-4Qf#@f{s2O=g
zl04z?u$n!z3Xfp$LO*RM%Rs?kW<`8=q>P0oGBh5CZrW;!alIRJil>Bbqn-NKWgd$w
zCQ?V*cWO&!hR4~WyR&o7rCLg>l#zBm7Y)yOg4WtHjqx6AtZEHUy}(a$ry`m3CjC0w
zZWm{V;RmO62GP@8<;-JDyb{@9Y9`ODggz?TJ2SV`=9kvX@y1YTK@tN+ojEnT8d+tw
z=z1A0((k^78@^!68$4&SR!HNMPr^&4CTH5>^3}GqBFbo^>XU3QH^02rCg(P4%fZR%
zxxoIS*%o_wb)$`4Sxhd@sxB)=QY(Hbf~%h|mXv;vr{5Wpq{x=#4|yJ$>`~j<zF%XC
z`B+~yHXi<ZcF-O^__Sn0(Lc`Px;Wi$w!*uA?~`_TxL5hRgk-T2{&`XIo$vGW!2#z|
zjoS%eK6tktUYq`u!_6Y3z4q|ld+pArA7yOMD)L8^)#Bnx;n5F&_<r$5zn53n+s!xL
zP8q9#{c*_)&e0mXV!V}S`@5~SxA(AQ&*}NqjIC%T3v<YinH8yb{K`VIlEGfINB2H#
zcT@Lt+P1ppJfAD=U~i}G?(C+XmC?3>gChwZpQi4Ec6zwe9^Aj3w%=}3nM3)P#0XB7
zK56&ve%ikEt-q=*&qM|N)Bp0nX#dUs?tkC@*?;w4x8M5O-)qj>GXCH?8+lV?n)7^?
zmX_O>zx0)M?b=4$dUU_tz57u`&+Vs8{d`l#v$M>j)5tIB*>wJ^?HxHyIjD@k#*nOr
z6d9o+!FnQ}!HdaBa=pmG#o(*6&@wK;rzhtb<ILaL>Arg3`9XdDDLZn?v&f{!2kr1+
zKXTVa-e1TZn9DmE_mjiGH_)1#Zxd5%ZEkVB&1dYU=nlcjqoe(z2OJ$`XbcVNcyN}n
z_@sUO(Z}uUfA;6?kN@aTEAnZ+wrAw@Ic|^c-)VpL^{==0-+wPO;YC}^cr9j(CNd{2
zDk4+SQ#gm5r}f4gx7y0uY8h`Cbo;h#Cz&&mDTCmgzznzO92~2&+tSYEm8EhJa2nIW
zIG1WygTPi^TOanbQ(lz(#c^<Qd{%NDT2fIh`GOv?E#JTYxSR+(wkC|+c*ZFe!A`q4
zSdQ|psGXzGU^)`#jWg&rH#XY4Z@=AES5}MO+9v*L`m-6_5cRpYM~~`GFEkQNZPmw#
z!f?s?Z6QO~Kj@K0|3#ha_0_L_Idf{UJ%04CoHAg?n90F)d~#If+`PG2dLvm_wAtRl
zVQp=9{rX11mtMz-CjzW?qX>=UB)fX`YUyt1bB_cWE18u;mjj9Z22R(mT}>OKdxghv
zc0RcGpgu*JJPCa@W^ltcj%dtA_|}b^*Mnc<)h5wc^e1h!1*mN%mzNd_7f>UW7`r|n
z+#$~c1wE?-?7~6)a53`Y8Fgk+^oiPiB=9OZ9dFEY^1SxrffezbPYxvdfe17>3m<wt
zt=+r-Ang>PXpv}o1LxDOa$;pX!`*_zb8U0;YB`gf{l;m8ciOTH+zelb+m-6$=C$?E
z2~nT$IBf`T$5X+%cnJK;I}G=DqisCV70yVGRFR6>K_}v5WQ14AbnrzIJ^0}MR{FA&
zcH&vlZ(qQ_u2-)H2X5ZHS+LmI-ESY?{<MtX#<ZM++8TI*jW#J;p8<T(0sX97mHKmS
z8m@BO$OCEf*~`JVU%r$yd^eu`QdW)aD3$y`2QS^Q|M^$=ku;R&v+8^8cmJa9y3z0X
zOq<G?muo1eimhtXm%Ly>a&L3pO6ukNTJI=l^e0u8(f3|YzPOpI9BTP22T`8YZNJl}
zvJWNm;AAxLwKz4os&w^?ii0B+@fCejQjR`Y*iZ|LLl@0sdT{kIc;9(NUr=0cq<r{f
zcb=W??eL|&%oonl?(8L3it4YnmE76e02U=f1~y=4jJ1m1tXOTVM!rnd`g+hm;lptC
z=QWyOlE1<ixRTV1963t8>t6n<|7mBR^WBD|ZJg}dTew&Frj7E7Vh4kikv5hsCzo;X
zJO6sB@4YQj$r-(onmPPe?HU%+j<V)+v=}(=xV+|*4T=px(l+)D`RXkE{MEt-<>aqm
z9Js0<o}&R$Z^Da=M_<s!KINwO1}9gIv3sG@wQxP(yB!9Ni@ck1Ys=X@)2@y?W8vc`
zwNT?ApKK@rOWm(VCiVXHAKKC3@w!a%^Q2KFKaz%fpZ)yl)5|gY0;a-&pW_<J_;pLD
z#HHiwGF_%GP+GjY(!TZ&ey^=x-|SAWa03DamW`2EIfH?@PuReC8RIeJa=aA7J1?Vo
z3EFvHj4gr<=0$K-4hDuPc}|h$NTUdG6i_T2GR^o;FtQC&kF>d{mB|-Tyw0OUIKbOG
z#0ZxapjRjZGf}1@==3uQlJZ}z;)`<9x7%QR?`<HQ5Oyd$qNuvFEkbbGhOw&WEPd|h
zSBN+}Z5NcB<R?fSMv1~i?cHwga)t-z8;S5D1PH~aJQYJuU#%!oK4whdCg?Cwfn`O~
z<X*uf?RB3+4lD@<l!Cr0kHCjZ_mynra3lbxtyUHi%xabUsNWbm7|Kt+Bp~%@qkgb7
z;m-G3DV6WqM&O`qYlk%CQOfjG7ZUYbna)5UI9c%tJ{05UQNW%)>B7&!?om5F6dhMZ
z{U=FXz@kv35rMh>lpss4DS;<nOHK|hB@f|=z$t~zdQgt7QvNco#$ebwvHBiZRWZ3|
z%=+r1zT#gUDK15Hm~k9;!8Nc8oC%D!NYb9p)Vm|n3RvTAWt*}iW+u->os-8XZ2olo
z?<=v*JGiFZ1YOE)KO2DI;yjA?aXqub`1Jg=Z~#s!tF{&m0JK+ql=j-*GVQgc3?aR?
zEd-zH%Co*AG<m=W%nBEC5opU$nto|NxGUd1+d$bSOnWocFVkiEbxprvj;|4s`@DoU
z72SFH<K<O~#prV9r%EqhfAwy^UVHX>s^7zdv@<jaJ#`KYo;Ni)9=bN&mKGT(y7S#4
z0sIc{W&pL)lamMivReP}=qNlFZ?x*YuN*BoB=o9CGN+o)E1|ui=NA`edG9FA6p_;%
zS;ZPcyT}1}A7hKM&}sD0J<(Erm%$<Q6FtXjyW=iB3XN7qZTXaZ<RJ0~KEyyjJ!4yD
zjy4foUa$})JSwyqG#F)DZ4oyXI=s{NcecwAuKcN~j7c>z{PyBF{Sv8g(snD)apb7^
zw!U$_%`Y+5E~PB}j%=XL*|oroK{0Yt=rov?VSk!~g)vp%ZEO1?@L){0xIx6&mB?1h
zbuQF&%2cO?6?1b-ZD!$WV7{8Z%;tRtm=TYgO#AE1BSxEy-<WfA@R{`aN_Tomwgx;o
zbY@rhXJFZ7mdK08`)%v-!@%u0{W@%?k;R@}ScG(xF~FO8q`AIZ48<vi53a0g&v-ko
zD6i9!Y4D(hrKL6(`H;L<+e==tC?n&YzSJ2~Dfj4bukG*crv2a|a*FebBPx9H{BgVY
z=?5hn-MDtG-MIC3yMF7-fy-k0j&GlJr`Ou`!0Be4qs18e?EJX!baKWPe53Jh6KVHO
zc<FY@kF}NFnDNd47=#*NGMK!9Ny>lm^dP^UwYBxlwz7UbZJW<~3}1aEJoyd}?pyd-
z++x90YsUxM?bA;_3VbhWJGeTJ)m)35X`U~yh|Ig*-g^7Zax(m*fBcWyfA*jKqxRN2
zUutuU#(OTZUU2)VxpEv?@2up}H{N`!z4g{RZEj|&?eA|z*4xjVwMu_F;}AF}Pt{X$
zi#g3mHFD@?9`?Qs!hr7tMO%ql-S0QJGputIWb9Jr1Q=dPzsH}YzMYEjd;G}Oowk#N
zuidbiDOa#CuWP4+^m%@6A+Qd<X8fJ+q6UWV@kwJcv@7FswAUU!xZl3}z3-G9C32nf
z>Pq@DJ~`GNeEL!QAO6Sxq5aE$^{?CazW4o#F0w5g`f~T~-S*)JKWxsGB7ZT^aYFR%
zOuM6xTrWz2yuHGS82UE9IF~w{FL=;t#YuN?iL_i@S&DvP>uFI7!GjkJojt0jB9cPK
zIJx>ZcbONZzdeuM($AWD_Ou(+(Ji0Gf}C59sm+bxlf?zgWgvX;a4YklL9x?Nawl45
zo}oXsih6QRw|L$b(&m*_TRtwe>({T91MWCw*xJtaedZk6_$X!Gz4suvj#e>zPvp5R
z@}f5_F9r^Md@AxnU1)XaB&XWj@4QiGjnTUv-hWv14gP_z$TjD<z4i9Zit0j-IX*=`
zsrN?O#3{CrJjdY+w?F>0ZEtTEJy^3X;%ew1*))i7|G|T{9r~$m3l?h2uFz}K(A-*C
zSxmdHwB67K4lid}*?QN)uhQc%mCRGJJA?hr8=DoWMpkBYwe=yG+Erv>alW=>^quZy
zyvf(<b@@$xCgYcW0kM=`aLbs7#wm;5Rp;ntEWi!?3P&==o*6Ta=Ze_N{Gf+>&v+_Z
zTPwn@@>$dwx}Un#(YI*qW8%B5APZ)}740|2j0t|$qq;bdGM6$277&O`Tf4g2-g@I!
z@^fB|6`iXyj020JPx+rsLA04DzM^>fZ9Z&nUTX{KX8=<;Y&%Tnc3Gssi3jH9km$q5
zxkP482agsOGTv#gxz9n*=>@z$%pCpTqmMHdeSU!xT2;=t<Q*J8{zo(F`*5#+)bF~I
zzdJ<mY_!?^Q0^+{l~Z7tXT2O_Xzun)8hyN$k}qRPuOmBD-g>V;8q&~?ey=~Td-Z<5
z*ZVNkE%p3q(dcKV19*nH;(OIrbykL<=$D@_?|<gA>UE9kdih$uluny3c)JUpb3D-t
zwWT{&$VuATolVuQ(6Dmy2Cu3>Woc_*-plrm_k-ub+d6_H)tlSS{aRmNFDE=45RtnV
z+>uY2WQ20MNB%Jn&}nl)&r@X8uPY@J2DjmL^(FNdJWD<)%TtZFe1!w`-0zN0FQ4_o
z^3+=elxnNb`kj9n?n=Fs{-fhk$1^;e9Eo0_jUDg$Sb{^zpXvyF&EdkIQ90_VXD`32
zM@o+BSbyHH{;cpIk4nZ(*(tWeK%O*STYBFtp14xH-90z(H+ZaF3LD;oYvo`bwHxiw
z7j=1VTxP@XYX0O;>(N1DTzN+ADtlmfWi4(?z1o3pTBKnd@rT-O79G#$?w-hHTe`Bp
zC^aQ1OdWpKkGk<%`t0wYCk^HQboIO%iywdeqSDV&_D{<9dO5$$>(?=nXfGXKm+3Nn
zQBt(p>u<l&e*5=+w=J%&pk5I!BNT?=i{8SOF%U{1!wZI5QA|;|DxxWQF(i(c?&Rst
z606W&F{Qq;td}=30!831CsPDyc__A4zuF?PuUZ_h6@AB19&G1waUNzyI1eKf!F3uY
zhcSvo;qY2qUJ0Y_P6y0cgbjy`h%v&;R2VpBTSERtKSvAW9HO!Gi!;s^AP!4)J{6;j
zZQ?L!LK)%Bif@jcuEYgK6a>Ol1ooA{k0LlVI}=J6At3LG45B<NMS16-V>syt5OaDF
z8mv~KT!v`}PT*)oHuw{uy0Ad72>|mSMVdgS@0@bAEt4@z|2Qnp^etm1AK{6x#6iYL
zM~KoF=Y_d1ad;50BuW4whr-24!&yUC9=1pXqsinIjg%3ubG)p=t3BVNtTPZ0yli)P
zVvqRKUlT9KNWjrHU}PBU3XD&}C1Xh!G^_u?G0Iv09f6qenLb<n?mfjR3*P1<@b56K
zA9;^(X6)di{ibWRSdv49%`M{qpD0iqYE~_&)7Zgp^TwDlLL^(^D%cj|PdV^ceq|7f
z89ykI$EPP%pB3no-Di1E8wVlW9EP)23lmt&$(JVg_NLC{r!bGbpkzczVf;uNDF=i}
z-BT_cN`XZi%%Scybzu5JmE#OGi2;|>kn?Ol3iI=^XJs&D+qz7b>DMs*${b&7%ParA
zx%{(Szu5FQl$giA>^l8IPS7Yci+r&#&o~ku9U5ZgE*c|J>n!rZ$<a~B7TV){lhGoO
z&?e`FppAGRBU#bw6pZeX9nfyPwhRh+$98ujxqDOu!=SANIXmz}#z6*cP8>!(io4aN
zebJ8dBvN4IFGIwbE3}mCljrCrBfCD46_j6*2I1X|`D70LVc_TFc>eStbn_szHN^1+
zeyt6*F5FJP4vH6(mB<d`fl(P6Cl`{>ws+}g@vy*caivX6lZBSjPrSwI_mR;kazt0r
z(Efa+p4whFFnjv+IJBGW1kScCoG(6C&bIKEsaa=ot>t;=9b_ahnFy@A(?$vaUWD&I
zIqNVfIJ(Z-qldTK!TwHQVv$kqjXh0YPutn?eq@+_E`+h;K)E<OE_jh~7@8hEe%u~B
zycZt$G&0tDTUxzVTLX&fV}xexbtc#Ra^LPx<e2fDiX1t&u-ax8Z?u`Y&5YGj!8It5
za<zLRvdTfbbLYdrct3n<xvj0=YO7b@2)}Z!7)Ksn9Xa4>k6v0`x!TsRz1hYloH-U4
zUlGAKUR!dB>bWw-VB4Ko<bv_^9qu~MjI0r^nCp2v*t^>vKfGJ;xDuF`Oj1s(K>hh)
zV6YeX1TMTz!2@SaIa_FHWur~crtY!DjO}8ZU6{@=S`c=aaXhPh@VvI=)BdsW_Tb;f
z^;>Ou{YJYovy%Iql9>a+o9X#$?dr96+qKQrjL%jXkk>ahGOlyU!+53dPb#YG2jBlr
zdz{a?x#_mKd83JFGymW<hl4S=x^X=)+HA{dCpn&+n8pMRg71+jw5j6+`Exky3#RF3
zU{ldr2FMk7ltY&66S>5|y|`$*TX)*MjM>@wQNeY3%0e?+XHJ*gPo61bU~qMOd@*=p
zyUuRJRLfCx3XXM77hUYm!pJQrTkTK&_>bCu{F6U#|MFk{LHq7^zn{4q9p>VB+uzF^
z44&HJ@!$WuKdNXd4!-&Me&(0$IFIQmnfvq@(MqBMMZR-nJ$&%EZSQmgJ_8+^M)yF+
z%oUO;M;RH5BWN|c1l-_A<e(90N%|w7p-C1BG8V#3+vsg?IgjtGKv@g^&Q41fQ;+t%
zG%BXQ&icB3ZKE7t91>`iGhg+IT=_IK%;#j_!%1jS8)s1M29UYy%%!UvtK|$?US6yV
zjojT8fw<FYQQ*IucHO&ozl@KZZrUs=$DG!O(*II68p#1d&gJMJ11~JhwKw0m)vm6u
z73@W8dFCDOYV*p<QhWQIw+arCrqhP0_H4B3AKizS<k<a#qqeoZU38n#o2*Y41YZVH
z4oht>xjS`>c5}Y2In4oQd|bz(<pN{!wsEB=aTr=0a3wgzS;@F=8#NB42M-<<@1ZNh
z6%HQfs99`4Kd6P2=@Y$8zc_`!-DMz4M@YTstg#%NE$XTF(gzX6_?`Z`-xmY)9Ur;_
z?>raBzDT#gChrxVrS0^esiNgCqI=M%)Bf-$?W4zDyS~|O-nw29ePnpitMF=jYd7Qi
zsOqA>iX1ylosif2=$pu}n>TMk$Ep)=#jhpgV8HW@;GiuhC3t7cHsb;|{p_@peB$dD
zp;4RFf6hl2XDa%Sta46r5W`u{Gtqp;$QXR*J3q+h<92dtSCsB3z=OKZhbDn<H>zr{
zIaw14vG7e}leAkN59+;KLvr75IMmNB?dA3S{r%UIGG9G^J^l5($<^ER`Xk?8dw-~>
zdNBH~*>HqCVCajaJbxAf>AUPzv=_oFhScGZ^m;w7|Nr{^eDte3yZXJfCUwK-p$%xT
z@dYb$u+M!m3MWg4Vb{mIO!d;{ME@phC9gJG?6AM@{J695-G!QS#@?8Vq~(ZL2B(*~
zqF!{ub9g7YUy-e$L6W(v&E`S>&G{eX()3Z^B<(2(8vp$3sl%dv`m8J1miN)BdH^o<
zN`F42(x+2q;ZO4RM_|~?03Y<#`~b5qpUTqP(U0s%+mgA4Fm!ZCFKel~N56)f-p{?^
zlGrtBy!>!4ziVfvTym5HGFQ!=(R*a#qD{G9cx?=e*Yv)FW7!CDSwyCv)tBV$;{eV;
zF~2fbJSsn{e|a(AlP6<ei?H(m{&?n|_wAgZPEogqk^3$5utlx22GNV*({+B1zV^ig
z)u&N<eGGmy{kZ$T-1Oty^UFyk4~)iesQdNrpG`k5?@zk-+3PnonUR-{ugi3qz94B~
zeYw5$wXe3f-+i}@hw(oR)9g-?@nV1+J%kHJaY8@|F0&3XpDYH1!7)}5P#7RYZ;6tT
z-<e6aVxSBrYvmziSfxu4XP`e(cAi;TS%(ZqkSAU+f-rEMA4dtvJ65RH)(FK6pA{2i
zh?<?7kKnUhTbK|7D82}+I;{-D!Mk%VWpd`EO}0<bc3aa~5sCp278t=8+iYQXcz93_
zH}!RaCjvm=f${cx-7zq2RowD&l)sg<W2t^qw#bhJf<a&-R8a_O<x|R1KjDG^K*<7Y
zLMWr3EgzN9;hXkh8l%77)@Oka2Oh_idr_RzCrXx%a*T=4dX{_)W&{w<D{bbWaezA^
zVAw;t21n`n=t66RpAs|(CX|HWLq>)2rO(<W%FKBHC%Mj|e1IvX#p@0eJvrzK?5|7G
zEB*1QpXqKNA9(CWF40^)TCNM1`VQxvXIBDg#)rbF5Oc(5Fu(B8d$pxVU_*&5!%M26
z^jVz*cJjlA6q5J4)5tbn6-5@jbukt(h}0q8d0#`}axfuH?5pIe{NTZ4@=<yTfP?Xg
z@r7~<SCnrx>M&ny_hbHwa5Tv{Uft)+GJb<_2_DXN>+fgIMse=KFh^cLU(8v2M%U^+
z`M|8&7kuV?v+`Ab_X+K_g~3#mb}Q`Dj%QDPeQ@J4U8Yg`)j7UaR{#4VL_foo|9^$)
z7g1-YWL4J7l7{cf&S&+lE8ZLii;)EeUf*Q{+mtO7ud|Y!vnCX$ewXtL9ZhM&7I<hP
zzJ)e<z{o+SDCZZx7aGRcA~CX+lPmPkcFAaTijB<S9T6zu6?w;n2UyWA@=Sf`DZcF)
zx>>wAby3_!de|b(Gqm>nJaSLuAzQZfd%FF!J-^s*hX-4s`JvlsJNo_N1)0g&I2Hme
zlv6=pIDhI)1Lt9lyJo`276Y3ewa)ovtHJI-kia-REpQ~;v~jHtjPbbe5apkq9;96-
z#Y-lp7>t~4rQDQ>2L_I<iHso!Wwge@`+4LmTQl~p*SZ5Q<8^UZ3m?kZ6J9xfh2e@r
zi(#4bX08@Rot+-GG0w99!S;PFQCk=1&dcG*T5d}#tL4BvkG$xNrkUx9cH`#TZEE&b
zyE3_24mqnR@Mw{4;}f<L)c43M<rL&J0SAUIpOeiNh$n~JX?N#y`v*I1Z|^}{Szc)G
zeCc=E<{Q7;7M9;^lQYXHBd|2)WS?=1u56dLn0lA1j$Ur~<H=s<vB_6BQrl5-m>i|u
zchjHlNN}c<Z3`o*6+F&QkAq9<uyPq>BjcX!wc}&X&h6mAQE>9C9q#W$26!46kqs6{
zQA*vpMgn<s|NZu@Z~j@^+J4k#8UF&~WU6PAsguJ@wBGo{TJC`}r<iff&)QORBC_s5
z`?IhA`{2-cyLRnnVB`JZ3gcts)8oUva(*2i?l#+)t*or4j2ReJbBp20yqcMDCQtBj
zdLj3Y+WiN2nlo4|N;3zX|JiZH>X+9SM!Ba>bF$A%5b0Osg{LC^j@z^I!?v?6f@r6V
z-8XN&-PYD_w1vgB(o5hpV_8MK*{ZVoXQDl6yE|L$2jBle`&YmJFWc_+R%EZth>U4_
z9$ET)zrFW^Z@1t7{eRuQ^{wx<J9i$IA&op_ahG$kD&pwuNg3TllU&=pR%e2(t=TfK
z@AANzw6w&r4Ho3HsnIrTcT&cF<fCV0cr8O;{>&5OMMppz80yif>FMFD)<~H2s8f1V
zr$@GoGjGY-oH16k+s#7XZ9TZMx)^zXvK%t%cShC5=GBt>$ggw^+bcSgizDvSPw%vY
z!^6O<!<|EcZlV3?nX{tMSNJe&Z@9K5+H0|Bv#q?*CJs6d7u(V~Pvu}gpJ_K^DS6n|
zb#wFPtOZsVBZDP}yDe!sWulj)za|6P%%P^OFVCKe+*_~<tepd8J3)P`F-kp$2gm78
z=3>ni5rEx^#G&UbT+Y6?-n^AQy{LQIWO0Hv(~USwuC8Y+7Z%&<%1YZv{?da}9yoGx
z*|xCPA3B!%M;!0wcIKk-AiHynama8uEvFn&TBTzoDF<$70|R?40!{nruBKbpzkq;a
zwnz4%Z<Rmwh`8f8B}11kn|yRTWprH<5A1p24IHD(;BECzZ~|V<&rP=#kyfb|UU4v)
zKb&(M2^_DsKIJHYb33+nv=B69?Cl<gu60MGvG?w2E)j6_f4Ycm2KAlefm2qSjjOE@
zYepw8+PAj8R&{EZ_7?9Q=}_zg$0vBA=<kJcXsJsC7>8c%mJ(RQ$HmM!4qfBvY&ZAJ
z+fP2dn>i(#k8U%P7XlA+CI2)j7nuMt?6B`+fU@=9uRfD^^k02)$orZ9v{^Yrd9I$X
za*VFm%JH+(%P)mL(ojbK{_43j{2q9>&*5IZpRZl-N)Lh)=79dVe3!Ep5af&ZIvo1@
zm0quJ^i>~vKl(dEmRev@_eZd*2cG-v_Jh*7)$e(dx&}G}PvLfrU*;6r;?=SprO0wt
z<o${&%$&hv(K+)~#NR8NLQi{CZ0W9%$;zp9uwM~h<}W-UOVC4=r;QxLXp4EH9p#V<
zOv}DgaOrqbup9c3wxM6zIl#SoT{lMG>t=@o``Am_u%3Z=&o$)j@abbJYOTh{{Q#k8
zOWxJyqGxHV{8GK2va0-ghx|D}I%)Vad{)DHQl?A0hPKIAea)p_bF`f1>~5hM-j^58
zs>SiC*Mgmb(@<XESc_`Y5A7=ZNv`hn>c=G%t;&4Y*M-bkyfAOpn0W>=)mKQ6HnE+!
zr`!5z5eQk8vwbJ@-ogga$Y}1_>65yTrr8b?AIfmnyLG(=11U{T8UQq;&t&=)B(VEc
zr_a_o;Kb|SU%!U?!{_H+e^(QVcIo)KOqb~kk~ZJE)xPw#ueGJMmG(43>$7kN3<;MY
zI8r#QG{(qmRX{<*gp1+f7a?{S6ybqlA~Izn_b|fR13m95%YM%GJI6>lW2`W=Lc8h<
zGn|{9FXN4E5zfvj7!gSFzSU{YVjytW%H%LijEcjGv&(8r+s#<nJQC~bIUy;exR`Wj
z;(ZH2%(_QeS!Ee@J3<m;Rx$RxV^tudmU97WTg9}GNKyUE^T<CX1Yy*+I-H48Q%<ia
z3ARf@{JAv<rIdVvK)o9{VCIAZ(Ps)`Y-8+FN1bn${)oT=!`$jZ1c64G$xl(Ltq22e
z0)jpetYoG%5P}F6wKY@jQy%2yq$q(ddCFOq)|ZhiZJ~0hhjLNn=PCH<M{NU9Ld|Gp
zNTdd0RjtO`Kj>qmytIvxabkQrWA&`<ZSS}96RYcfjS_&4Q35*b2zX@x%~%*y#!~&Z
z0+iqhKAxEi#<GgYd+<;yg*4A<wI&=-nsrV{=EjqZw{u}QaXDWr+K?a@JTyMFU1aJP
zc{N1Vm18V0Ccw|;T}ov^JNYP{)05K?+5&sV#^fy{Wy;Mfq|qw#?A#VCBC>i^rmac{
z$`lN9!FGvO_%j&ROv!tcVG01V9UPFrgFzRr)_j8tc@OO1l)l@tix5hA)kfQmsC!tE
zMd)?E>JQxDkFz;ys}>Hh+^58(i<jv#{ko)Iq2uf4*y=Eo7i}20-REBUudaOmg`^br
zvzJu_Hdm)F{rjtH_!e@}=_KAZHd@&p1r{w?UG0XnGJ@b&p)t0bLXQ|-(N)paXd(j~
zx>Yhu?u!hfLX$nn1^6D?HYn@p6y@G(en#HfnvQX@j4$E0-YF%1vS@475xR*NqKRd&
zAV*XmMk{3VUh$7fTR(NVi37=&XY=z*#h;%<&Jo>&M%KAKxzD+g?w2gZp*KD6jI51Z
zYbEbYPvXntDQhx7W7Jn`>U}D@{V3l=_=etwZ+Cbz)Px3y524Zh%qfdFIPB(=Ph>a)
zeB_yO6e-gskqeC7z{{2QBJznt<#)mhEr<%gd|_cmc-VO4hx9)!EV+stb#ii$=g->8
zO2%*XW@M(t;z!_J#t@M@lPhV@LY|3mv1li}ZFV8mk)M)hvdc>N&hpB7@{6j{ZwsPM
zs(*F9X7HdpqolO?`ANpY8JLVV-PrczY`^Uv?4{4fY$AO)4V*<NY(|E=mVBIN-8oZH
zd@s&Y$9_B5d)OY{zf&^6JOkO7#U$r#f9HPt=))hh<6}{I7G~M*J2J+z^Y-zF-)VpN
zZ~oWXj*rY#XON_w*z7EMZ)P?;@S@~}lI1cc@MdpsyB!_wR2#>~CQ8;Cb}oQZLH$v8
zt^|K}+Bg2}kJ`WdSO34-f_rU!rA;T#{?5a;b^mrdc2-q>E1M&;JC~%$G0)BdV>4qc
zGR#q(gR*(;W?Q{_Gce7)lI?<<sgN<>q8xGpJwWtEEnXu#WPH0*w#&)pu>NjuKWsZY
zTPX`}MW!Lc=l&ENcAi^}6^Dw>7|($|NMs~<mrH*+M>6-|+~dcO+s^JzTV7gf>+72Z
zJCf@``v2hm{j@hFqZ66qHAjL6pMLoL_J@D?hwTsk;NQ0Y<$w7X?ct-xrF%H@=i#GA
z?a{;g?c<L>ZlB)0Q_dvLtasmix129<n#^c{-~8M{bQKHqy1|dJw9Y+S5?MtbSqQB4
zrOTNX4L*9b9URyxdF=Z2>zV&x+2<Xnm-A-GP4^!>tTyar9MCp%WS~p6U0rYk4wTU^
z_q3gpiA>KBY5PGAt~cMfQBDiaQuOlr_04wc){Sy}aYEYCk7I*V<IbHs?atkMg=^QY
zZKmJdk>p}PCimUAaibg*=Dtg$oO~P+jLG0b9$Z>nDreWdQKXaX=&SQy&|I>#tInXw
zv(jTC6EjAm<8~JClDyr(+~ID)A$gcVT3&`y5gl*8^=6d^UI+U}?a^cB=Is_tIj-#P
zL5H<)%ZPCHz5Vt(Y5z>|6^2$0H2ST!f(NI`wQJYvd@_r=Y@LP1SxC6M%Yk-L{lr@>
zvNV}-FmuJYa|SZxlY#M`+A1^sHBKT_)Mdt@Bjw1<n5r9JDd)K|lY(9adwfoYf>+=m
znvr9-;F9sBf51bN1YLE%XfWQEGVnM!qbv#U<bb@ou^xO`tcVwF(^k%^#k99O+r|rb
zx3_}3fvbg%aE7kvyf(7CIgT!Ahx*L7nc2=e>9X4U=;6bnU+_!B-@rT30x+bDdr$w3
zd2IooF|J6v+}9ud(>~(~N6<1+bmpM^wKZctEj)#jbbW1u7aXA$SC}&elhkS64ax7~
z<Ef{^?$!O5pD%y-TL0CU`(89U?VvjuEAZ~LVWcxByRobihGItL_*7rjpT76|ujd=S
z_}A;JpZOf<Pu=sc?s=Xog~EYKRer9(x12ZmuFbv;@6dOZ?^*vGy<acplB&ERM{mdH
zr`~V*s=mAeuI>$O(pTU6*c!u(_b4J2Eh`7=Xx!0E^V0mXjp|wEGx<tay=M_p$zmxJ
zZo?&G?_4i7Cb$6iId!|!t1lLXn=j8&Oj~Lvl;j~tn1}G(Sf&|8qlP($w^hnBeX76u
zZk$Jb_iyyvkAiW1m-2o1El=)ru)L`EwUsR9zWUHeX*%;kpKIZD?p0}de|V?pM9O*9
zygrWod;hPuX}Fhn5r}#SANspsp$;|&^WXfg`n>BepXMdLBDxdYv_VovJ!e+|KUc=E
z%iH9ZZfEPqrdZ@V+Z~q4?;hG(3*YiLlr=aYC&s%Ijhtc8@mb`i$6LFR14R9?6Oz+M
zi_=rq`FYBV{s0FDJglBqe?LD}eZ%GVXYajs{i4$^qs+f4{kS?`@8@T)zYFPS={N3r
z>G--#m+1?T#wNzw8()66z4euM+e8FMhJ1{L{%N1^a-3jD3^^m_hk;>geH*hV(Zx;r
z%AInsgi#S3Fft5(GRzJmwRIpT7Qx2K$%7*+ds~E&e%L!F5vNZXwNl<wN*|@=<fu8&
z{&@r#N)7>zV3gK($5#}*`Nic>o{17r%25?YS=$5VJAs8zTU+0xy}oxr$3f5z&7K>T
ztBt8I0=^a79AGIR(g()*wE9U1(KlV@1PVY326GEbWt5{xgiNNteT59>PRV?7A;O9w
zY_i%U;!U5xm2$PRv|Md_c40*y1+Q;qL?C#Q_nploN=tc^imsSg5z49YEX^1%q0arj
zVu=92C>W(mN;{RQj|8N_spC?wc2VvKkhNcS#?ZUQ(0JI|5WFd$rzZ}`&Q;?_*rZei
zeor$lQHG8Vj@#khaS0*6c1e>6L4idLZQ9Qn3TMEAz}baL0%aFuyFoC|3P(oDQOfXM
z7vSMV;GFx$3@();GdOL&h*FdXY|R7XL>Si;t42#`8~vt;b4-~_-otM3Tmqx$0eQ?x
zLT5e8rHvJ3P=a~dk&c;1Rz)eNgvPYV7DnYPOB>zSFG}<2xou4Q)-Z&ArIx`<>A6?<
zuAlmyyUMa^)4YW{B19P_%}G%XL#iz_N9C!{6H#Ce#%ILl{IWHpewHQRGF_(s^y!P?
z_>u+!@Z+vZ`GSD_qEgEG*-Aq>(1K16KAYS@kL6N_h;iqaSh-nSm4rSqwBEdVy=Y(Y
zq|hLab_Th@nBlXO{SaKnkCHzoJFmnym5tUGjjzf&ZAC*x1TYFSu;Zf^a%ive`hp&d
z7Az=<0v{SkNhe#78ORgpVbQzr26VM(ae7wK@u9`DGtS<+(XL***=k$1^x<&d_FLy|
zY58iKS-jPwttOWu*DR#|h%{;cI7ieNhee0ovkT{Go<z}#oD&(WZ_{=W-n$#Vyq9_{
zm|zS|AHCwed=Wu*a=ag!dt64Qi3#T#xfrFLoznSR8Bpp)svKT<(4L;{hbNqb(VgYq
zd1SKPw!gPiFeR%U?mrG4Kd8kP`Y<uI99S?eowuW-<eM04({t-7*A{)Vd2Z`Di<}s2
zw4Vb+eWKNZV<ppQ2fp*9?eBuq(~{SWL2Z?ocAOq@Slmm$oPX69Y#AHyc>ZiZGRkgX
zb(k@Jnm&RRLm?+-=g(x`=aJQn0q3X)p>C8?hB}_Kv!lE1@%<0l&i4JbyZbnOVz7-&
z_#`sf>3(~7|I_y1-X~Qid^<Tkiu`z-@i=JjeebW@zxg-+qJ8ked&O^8msiqn+s<Wt
zXO|*tO$LX%tOy=+3+ruaX0@DirziVuZS`tfUB6NNwvFNO;Xyf@o{=Xr9xt99w*8$4
z?bAEA)3zP~#i_zM^y#M`wGV&zUOAc=gE<8ms>WVM0&)%?N5<X{m`{qAzxk!#ZmU<{
z4GiYn^T6cf<S_8JSca1@@Cl5_GUl58fV=tUTwM#$YB5~K?fi)?$PNnT>+4tB)vMP_
zez8j58ER9>M`mtgBG<s#MF{9~X811JO*ED>;>O0)hlzz6zi)l(o9#dR`~Q$SGWXLj
z&b1HTe?PGPH|-C8{|_>^pSJnn_>0Kb&of^h-n-rY=#T%f{rkN0=+R!ysj=Y6d32ne
zz5RCo{)3wPb(Uc6uS7l(xwWyeUR$=Ul*8R2{<Gz*JUn)$-hRpNV9Id?_Rg;2RCbl)
zAo84T6?b<TSWn9t_2!$mGM-aS<kQr2H|CMCM8!C3synwL%Lm367cZK~7K;ExymHcT
zpc{kZqsZVVJwi$u#(iVMxm{~*wiXbIvI76r+;^_uOgW-$i)fx}mqm>|Lh5<-L7!LG
zR#WDSJQr<fyTi%a^6u)@wbBvj01S@{^Tw|`{5f}wgLAhYJ$jrnZJ9`>PTzvt;~8(X
zf?UiP+0Sk1vZQmm78h-CFCw=0hfW6Pk?3+L1gFTTqFsHLcc~&orc+k<M8@vq_^fuI
z082X|kYOJ@e2{mZwe|FgvyL-nF?gq5aDDXnQRQ<6n20w#gM)Yd>Uz<Ko$bB$cx$U*
ziMHv}TxhfYT7af5yE(v(@yvhYe>L!S&JbrG5=OokL04z4rF;=^PZ*TLKT2k%OVG80
zOCt1!D3|UScoulk8@iK7-~D;!k_x7wjrb*JDT6*3VsPeY@i({-95F^5L}0hDv{3lF
zy|Wh@@wA*;rGqL<n~z1S_ALeR$7|QG)!9iL1={YOsp`{vq8)8L*|+#yELe#q+|T@?
zKPZ2FeXU@27~DNOvsEtVAR|8~Opk`sNBCQ94&J!O$qQFS@NhE2QGF2+250Ed@CCgA
zZ`*Cs<Kc{L8#gwtw&|H}pzn)EyTYMPI*Z>g9UFnYHq@e+w67c9UHC$GG}OJc4{Xi7
zx@UacuRJ|(UB!O;doRl<xOtXKJ8~J$D#JT<=WkB(*Jt|B+cuQu86NJcLSNPar}VRQ
z*4)$w?@DAMIS2eJ9P?R^m6LzT+f5(&4+{3%qvYFA-YYyEF8AyGd=Jo*rROa?&sd=?
z<roePBo`P{PEDm*WNYDw#SZAP^9`L-S~Mm&e-Yd_j$lyRvj#6p9}nH}*_~c~b0A#_
zeY2?hsOF7#<l}_#S+=WG%C1rKujDy>(8rFeeLP?F)m8oL-|GLKJ@5G(rCvs!CQHv(
zzxw_B92}=F)2L2(M!u=?($9ivo)n%I-jVTS&3~hJ>Zh--rB`j!0Ju2R#YSW|nSSQq
zu+YGL^ly~Ca8_H;P_Rc=wBhwsFi|$!Ltu|SXh<*HVr=qhcX+z=rC`i1llO{ly}Y_T
zr0iiCJN0+ky%YLGzC-ia`9d#qDMwVS1qJZDY!_+M(BIEXeGFfH=3D;dt<TT?OG~fn
z|J9^lb$!2zq)$KZ^|K@r?4{%DGF_%$_ml}f7eVz)zw_Je)|+osv=;_}=|~Lh1YpV=
zgFQxPrK8n1ef2YjM92_<MTucl7%#<~ut0%ixGBe3%3wS?i7-l`q68>k8z?CpMz%Y!
z63saPBECdIJ^TN&_upTZCCQ!O=cz(#kr64Xs;d>S;Ovr1{=h^2#B=t)B@fSV&+Z(K
z28JV`1{&RHSt?~@XsuH4`!RnvGOJP5U1&6b>PC3v&G+8DeqETGo11Vm4`co&<qOl(
znbnksaZ5ufICzlU95KtUb1@Wk!&!(|7E?ljPT6@PS<nd^7E!yy<|C}@SFtErAholM
zePNtlYoChp%nn6$yL{^ls-N9^#=+^zv!hMFEzUwfP_ltX9mdI^ls3BeGVNdh!^qfS
z9`>mOilbw~7XyMiy9W>B0NAO%`j`IJ^W5`^ks~}3h`}AaQ%%97n6_Nr#oYCB!yM)r
zW~TBazukB4ccTd7PxnGnA7OupUq2W_F-5+EE5U^E>-FX9cDQ%c_IB)Q?1sXxa{>z}
zC9EYDvG4%yXRL4C?gn1)!9?qeaED+eFI?ow^`?G<ZE%htR50xfK4T)7njhu`q1Kpz
z7iK$IGM6T5j13S55dNiM#bom~ETeGl8$TgE`y8`N67!xh%x_94N*hAp%Zyte+uJoZ
zVLhuK7*BY@6Oe$X?L5*5p=I2Rfa$(D&0tpZJ>R|4jeCXcRqj}X`{`=e1SG3^?WpZS
zb{YpFg>PWdnb*EEjNwf<oXIwsI&Rm!z%=#NnRze2<_L40{yC&yAH&x_RmvPU%m41l
zTX>6i-|lnXoqmn}C&cHI5KEDT2cb;+;2RIK)=`Refm?`R>#6)btl@3cS&@y0wKcuM
zHs?O?UW$42wZbGK{;2~^tIa-n&Du-3i&jz8d&in!{k8k`R_Z}-(PbW0*2#mz{ZiJT
zr!H%zUBO*pDFm1^7^m8Jc&CXOXFx5bJu_W7dw!Vb;lpO;1Ddy?1U2Tyo6|Z|2=9TX
z=q^6tDb5bdfOU3yRGWcr+_*?~Mi`-9j@s$*K@+CAy?Rr-gm2tr+yb-7GI&jnr!2~g
ztQ`zZy-rFK3a9gwXZ^}8R|~Io*>_kw)9m>4u<8UeXKhW*ECn8S!qe^6Ix(I8-MlrE
z_ZU~QevI92<F}mylXc+saQ2mi;|ULOag=_(s<pF=j<4In&a?LP$!GZ<e&Zr_2EM#-
z7!!GyfX8wAe|WGRUSqfI?>j$;hqARQxRZ9=x*;S|-_W5P_`$=d&U^|S#>VqId924g
zM0R$!vwnrQN*U!rk-k%Czj<}mj`lWF$8O^ZvVQM=@N6RQ^s{H+iO`r2-g~b-c=(Mr
zJIl+YD~@=6Pfb$5pQb;L+VbLJ`_{L<89uQ`2DlY?+zf6^&rP;jMvyT)gmDhfaDJ3F
zya~=Pwp-&2iPLR->}G}AJ4kzli4=NJxITFH{s-S`|EK@?|I*gi)@rQqf;?V$Aqi*3
zU}7D%K?YAAa3nE82nV&7cg~8>xOpddSYu=?D3Ffp9lRr6#$3Z2Tf?36W$eJfSzina
zjCTwKhZ(cevy&z~*~0u%8O5BJ)N!sGLU^^`EDtpIBuX~2gi8a>Jy&q4JRAGmz=417
zum8OL-tYdW_Q|J@+Wy{Qg%SLtKl-!wU;gtSv`=%7!j2Kdxjw@9J$WqL;z=12*Vi9r
z93PZHZeeM$3?1YFN_m^tD9_f`*4y1&yvFcyo=>OG;JSZs*fvHUO~vo!55I`t#|xG0
z6WT->%ag`+e2m_-z`;3i(;-X!tSNH`FD-N%ZzUnTyi2hw<Q`)J157D^gZtpCJvPlT
zq~YJp*PA!rWIoQd2M^ZTe8#+FkKi^uMz750lDm#hs$KANIGY!JW%x0!j9-+$+JuH^
zi?QZmrL8vQp>^7Ww>&&LEV@`_q&*bL6>dA<c~daNbQxkiG_P;8!b+l(buLfpTVK~^
zp<jgt<>h*-3{TGK5k}IyX5cKDF}Q7`6kb(((CXqZGmm&4?(+yqTlB9R!%nL|3?Ips
zYd9}5;~2W}>8GEjjb{}}!A1s=J60-h`j)abHZ}{7&?P*-aH@+7%T@o@=4RXD!Ikzf
zgfR-~htP{W&h~})+up9W^VXwm$0NhvE_-yvtozm&G<A^A$}o|!C2Ps2zK(}@=*~;m
zH8QXc%InT^XO^TtHP3UOkwUpPwed1!R5H$Za%1EIW4KuxPb0S|gs~2d(-w8zTVJp8
zoY|&rB^L%i%9xh-cxhR07}kvOc5r$p^x0fA{^mbivgz_9W7P5MMj1UtHDN$!{LoJF
zbIGk1<y>G)X--bznTsBk0zL34nLd~240%G=!Ewe@XY4WhgPX~~OS(KBldt3zuq)X6
zZXA-w{n4}1VPsr9ALJYP^xJ*yc-PqF(>Ps?gL_w(GDn|&*ZZA*c*hXtmHIB#XaD-S
zRQ!#9fZTtLzDvEV;l(Py+NDly>c$QD)IFE1_5GPP^|I>KDp|SxcRf-1+xCs>s5E>{
zw&9)rIV!ADhlBRiyZ)stu=JT3QMe{`{O0w^kkLb4l8w>KrO<4~uYrEz9nb{EkS@dO
zb2rGF>l&qhlfl>8cx+C*wn0S*KROik_59|odC_r+u{rRnzA&EjzVx<h_gk3Om3|G$
zZ`pePdw=Ww0Uq_Z=c%7Q^Ph3*@aXf|J5{#Fx#b!BQwlmU7{IC>{_?4u-d1mo>KnDg
z@6WY$blbh6IqA1~M@ME3y^9;lE&TS%aBoyz&5ev%6_?+VPw++`+%f0q9gGqBksN__
zH%5A~?y08U&S%P3H#jQ?PG}FA0Nvf(+${ro8N7lv_`-n(lo2hztF`sdfBqjz|J!M3
z_g76{CBHmh<r?z-ebdjD`$gBgIQ#3C-tEViUH?$&i|Ts!!o^-2zOK`C`kJSkW4GJg
z58i9v{O#XvOJOD{yM^+=5LpB*BBgYSpp$vsrXs9h@G%g@B*~X|FbIqXJYWb$b`@e4
zgj*p@5R7`rj`P|HxPL%VL9s5>4yF8933?7E<ynM@?FOKWioqW?p#HMtXMOA|HKx<<
zQwyh1FpOTqA(~|zs)gB`U6AlvDX${r*j2*AtP4OBB|Pbi(iyvit0{pa^)p!6krz3l
zY9>Vt9%uDV`Y<_R2VjH-OnvId{4;2&+xzD)3B6&+vS^pEld^dnVIpov=+QnQAMJ*9
zo*LLAq?8O`YoWr7DZ3bql#D?kMB1fYMz6rZSl9`U$LUHS0t2Jv>+&KJR!Azt7^Wia
zQ8%N-`OAyCR|b{j0W0ko6nSt1J}A5d=d>-&Nd6IH6S$llioL!cA^M!0`1Mc5v~WFG
zmG1<uZlHme7PI~wKIp75!Duq3))hEhG3s#6_z+_A+nD8PF=6t-MdPD9$`kk{qg`wA
z5V9sByk7=azz!2g!7w{LS7VD=gkuax1^2X(QdPOagC3n66(g@+^NPZT0gLeG9XJDj
zDW04i2CoQ&+QFbnFg4dP)-%(ymDkw%-4!`~*PP#kT#5#D!7-sn<e3_Iq!s?=vl#N=
z7zMQ(l+I-oPHW(xXC=f(c(+>?PQ5(;bwHf!^c$S64PRd+Wj>qZLqaRR{6aek$f(dr
zK>wLMRmMC2cj3$$p3f{urO=3cyuQ9zo(I;PQYxfeMl5S(t^J`VwSzm4csE~AYNk9i
zw92Z|J?nHGJw}tz(NL@U{W3IInG6~gaxc$`(7XW{+R(tssjT;TPT^bzp`*iUBifEH
zDY_Zj>b%9v@IGY}t23z5j$1>RK5O}D8k2Q3ZR=~!ZLMpsF812a=HryRTeMX@w{B<6
zxjZXRpu_zaZS%#Gc2LjHs-$XfU~qAMT3#G@l^z0`@<+KHrf*(R_&C!seGeV)&o1-c
ze%gAP+QH^5eMl>{i#OQVZ0K~}eQn(mDvJScInUE4=TqgKJ2%mC1~zL{=QX4^ua4S9
z_>xfnl*J&H0eyYcj(484-JPd_>ruPOAagtYzj@I<{pg47$3OXg>Sss@T(eeKUxiAX
znw)JrJDcs(Pk)^C>a?xhebCk)e4{Nb-%Fe3YYf_rS8Z%^rrjQ2%-R^9l94h<=FF-a
zx4NO_^~==njIHz+Z)T0AkTn+PXGdAzDb^`kY%CgYa|^4v*4pIU-N1Aq?PN$DZwrg7
zZFX+CO)uR~eQWK`#6lSbYHiDV6ydy^wzpoix!LLV@ZtS7Gdq{@q)ZLJeR<SQ4|mg!
ztluZX86LH><Gox*dCw-1<YgSXF_E#F$e7uTcF@jGcMC2&t(I2S+v>e<wV8zv+Wbt$
zDzFj(VRmkz_`dVgqb79X@eu>pd3bD|Zg7H^AD%S>{pO;)-?VE8pJ`s+x`8j|1;~T)
zH05mNzHomGAFs*?WaA7N<DDe!VkpE1^s{zSS0*7o`{*eAVfty_nvXmIyN90nFfx3Z
z55cj)c-rw(>dG>nevck~+CKg4N%)@g%z=xxvANd{4$cZ*_wK#d=H?gEhLiT>>5IUX
zVZxb3X-jw;9&bDage43OZgY`o7Ut*M^2%zS-EIem{aj!4h5Uktrp#YhoDYxkvOW6j
zX?eo9H#03PT;~^y=Hx2=F>quY-FzYRUSM?bvW&L8VP<A|hKh87Ch>B*Y@3_g6><(O
zLW3ykg)9`JkWv<{Dqb&g3x0}h6!7IYuZXcWKR=!JFP8BFT)`e5$YXpNq|gM0NX8wq
z#_^F2buX*^>Qw&1;zIR%W8-<JUFolOf)x)Z9z2wa-5|%yrt^&qPGll-ALCkuK|kTu
z<4nBCQM)>P&}m*AJmf??n471>o(cbYoHiX5PrSUm(3XVV3mz+PGw@bk_oR*d79MnQ
zajx>hE%K$!5h95&fN6K})q!0Z%QK$(8WAV&bWbdyF1L4fsx9U|nTavfn39Qj6S%aQ
z_m&W*cxCdGaE{wMyJ_!fyUcHqP8gFIPrZMX^?<jTaq4SN-&{elZZ4fw^p0+{BTLJp
zPle^_pFSCr;-Qnr8YCoOc{XM4P4U>Z-i4pHd5&Syr*_&*=6*$~FrP<d!BKhU<|D5L
zCa+(<Zu`3#x5CZID!<m#teXsV41Qo}ttBfkYz<)#t%vYm+pi2o8DD_)+c|jM^Y0*S
zxs&&&G6$#gIamGL+utjBWqka0;XoNNLx;f&9(A1o+H4a-hYz|+uIU~?T|Pcc8xIeT
z3NPeAtIc6J$LJ-mb$kEtpvFzRcQWP-Y-FL5i!(pW0duR%HjFMVu*@|waJa4v0q&(;
z@Tgz@)Bk=I&tow2U*!v&Do<{7*}Lm2tP@pc+TpkLRzBliWw`RkFgv8Wllxr<GrB2_
z>aDNCyPvy;Oq~{6^CbNNi{2kvkl+iMxa8t~jb5&#{q>o9^f7s>eoNl1U;gthVYYuA
zwPExwcz^>KT*0=U6z^mc20Foj&!|QBAfzGuHD0x5rw?g#%>z6iTs6<hs_@zPn{Rjm
zeK7y@PkR|j%yG1AfQ@qC4IH!i0bMHtO3EbfS0DTZ&*krJt~<R?`S7iN*S~(hlHW26
z-|G|Xz-BboZh=MmFKfYll>Dx7f+zKS&;j+0x8`@xch#DDXY{pycvj#4cDX*yHE?zY
zd^b89i##({^)9@vNnZI~Ls>UQZF0S(2i>C!Z57J7A}P@;1@F*s_~u!ydHL<$fNLEW
za}}PX-@z02jfW71&TX@<qFt|E55`Gy!dqit-t<ac{xkLiEPhpK=<}CdUp4(*@_t$R
zs%88t>iVZjU+1>tRId$R*XcTa&C^(zi1&Z%JMH27ACzD?Ff<k*3m}H!GT$w37!-^&
zVJg(8@@0i7M#CamA&0_9h=-10vI-Fr-N($@p}4cXRlB?_e3%ReL=#*v#M2n-Fw(qh
z4v!D3emmA~S<Ec%c}E*Cw^@zD97GVP&@i)g#uf#D80`4NEIGv_yEOmgSg54)isw%d
zz~os-lQj!w?;jS9k&2>B7>YW(f*_ZR$5}5mVtM*Y5XLxrRX=!q-68l}<n4INLM}86
z21VK04+i>#aR6o^STObkDU6++o|vs}^w6d*AUZ_33+x1@FgvN=r?Ty)#-wBPl&xK0
zR034$AUrURVB!Y`BpCuDaKK1(;|(};CcN+Z!RV=@6n@DoG>LGH+D8djG2M$X>TMJn
z@q+OFe7o&$35`>aevJ~i2gd``>8a_!e7eT>G=e`sGOSWI5H##AGd_eju$I8C8_Ni&
z-4g+v!K!d2V`ITKFXSO~f-88J2UeJPZNR)@4m;!66{i%Y7&drEdC<RSZe(oA;Pm>k
z47l=Yt6kKEH@PP?*nH-dIpZBWo*5{OF~jV9gwPT&(^vR`sf0_ESlZVa(#)T{XIHeg
z!WDwBK2bElf1VZCVG1sI0pE+s%**f-+-j#6+={U6YyhEE%6ObVN;EK+98NCJb-GUf
zMCsb_^_7y<aKr~Bb%Er|E@|{#zRHV|yiNVKwXrKEN^l+t-NIT4>GvvY(D?YR_Ta(&
zqM`EIF;20CM)ft3;;S3~87_E1*|>w3u%4lB3_@ru#Ul@bqQ%y~TqWpd5w^ov85C*W
zLEk;&A;Xi%d0eF^qx?%7c+EIpNFCmxBt(b7g#tx)@1W(WgL1NqNm*kKx7z;xcCD%C
z<n1s#uU`pi*TWW`ogH<L@9-lxLr=ScAoTO{tQ{Tim-5FhdJ4tK@xEJ`p`v(!v<nYZ
z28jx9m$l>Kpl!Z*TH&K}GwlWEG>^Ae)*<{IUWsy$B9upcYoX0q0~vb8Cc1}H_tasK
zFqYxjc3-sPqwsctZLNdpyZYLjvvzW%k2}R{P+r}<ana6?_S*A}Cv9`{S%s;iWTjx7
zpnwctx41Z)_6QSq(0Dzq-+S2RmhOeun`?WiYj1C(JP_`^_s#HuYiTF1D8@*<7`zF;
zbi<V~xM>VWiU`eZXW>&GwNF0&sNn7_C&r`cxurICn+G#v+v~t}Hr<{{UuRRcGst?M
zosBj=-a~NNSmN9^XOZ==k)6NfDRq%CgLmraXXuQ#<>l4za+JTV9R+7T`S?fe$3Oa;
z<mJ(_pL@G)Z%5e3^YC#z)NTh3ywQ8xZ>8;wUVGacH8<AQ@70-SH^-byCRFxO`|Pt%
z3+E`?cXxK%4}S2&_NRaHCmDmiw!E^EYpvZHn@xU-cIWLJx6O@D1OKo=;WJE;Iv+0l
z;Y~{P;FU0muY>zfp7Ol25$JWDu|>H}QN|lx(of;~wzf74FBvO^)_eT;Q5hsM?dqbm
zF7G4nX3TF^SW!3wR|mYpRn&kbMXq=SXD=d0ylCJ5{ts)1|JwTfwy?O;rY7dv{Ra=*
zH^2RzwzhUJ<9(-i8$8MH{Lb&x*-n(?JivGkorEsn_tA**(pg;zuHI~q9zAVan>%US
zLGW29Kw;LM0SE6V%Fsd4j#j{NUI&Xy3&l60N4!qWMe~ox3*!jhiebk*Tw0oMb8|Ck
zS6~`IGl&T_C!`*}vU~G{j-exIn|wl?GgOrvl|GfhD!dbqJ4Uqmg}Kaywcz4Jd2liQ
zl)Mu@j0}hOGER&Uckiy%xIBIOJmnl^KBkPwPohiU<4|Qrn~jZ)&f|wR^1kA2i^e&l
z3f;qB+E_3j*#s=Wmhp+V7h|8$boz(qn@jr^=ADTZZZEi8=jPo+bICenqrSP%1_AOS
zZ;pHS)&s}c@R_gLM#|HN8yO$;&p0@X?C#o1!3)lTH~v}ZN4&ea#z=%d^Kb%N#=4N0
z)Q7Gd2TwPP7Y09uqow8Lw!E}d@+iI(zP^miqiw=SqR$lO{oL!LVCK!hlHvbm^#PB~
zP^x_Gfx5iFR#sOtCZ}cWxpSw}xRM2GEy(&2y5z6q?yN@?(B<V2`0-Q*>+(hno({$z
zwB05Q^9Q}GXIXR5Y<OT}XO}-~2m|xX^prU{jb@Ft)8n)D^zrjD&e|M7wlF944UUzE
zRr*s3<n-HmNyexRkiiSM*kz8)(%?M<2iie~U|=%mcXsv*F7Th>i##DwS_>;Fukh<&
z&xmTH9r>r`Tt2Pi=0UCNc?Oqi<7z&^SLj3Q#&qWRy?gho9ynL$c`{sN-k1;YLqGJB
zt%Mh!cEJ&K(rcjQ<eCamir;`Jso%K31@9Luj1|N!n2y$zN?nH_zwZIrD2?vbZL`Z4
zzX#=hm-PmIN8Roz7QV>~{`Hz)qdH3#t$f4xeo5Ln<kwC~`B!E5%XhjD-#Z^ghJ9;P
z0pp?1{k{Hj)sw&KN!8cO8ZO!Db5w72xsrU^+S}|;|A+gvKId0eSHJ0cGL~u6P*(@X
z2+kvzfSGahoFNtdI2%q#S-K^8yD<z7b;FkgS1Za$3Re`vma#`8;E{f#iRcbmTY8dF
z+rYc78oPJK;HqyupElO#==xk5e&(~Dr7Zmf6Z!=27VJk!J?;-iS#_Zy75yl<E^o(&
z-oGkdvH2_b<w7&+UbUxxb{x!)yl>3ATz^*dvgSeREZva4<r?s)>Kc773<nD^&IQ&3
zTx!#8o>>=+Pr)(muIA-_;hVnJJtgISl{@-9YM(jiUEW=2&u-`okJQWXX<qG{-FhFq
z*K0`fZv7bUn48ss(eK}kRAV^${6*J4bo!z?e^uAlIazJ54PV#kI(?l}7WtJ3OCNma
zJ8gDxp>}!IZj7uVMPFQDj9!L8q&Q+&DW#5uEX+nJw6ky#uDpl2u(o6ki_m_N%OV%S
zD}tSH3F7)6M}RCNLCU~LPem9M-Vn3Vl{0y{lshAL>o75Op24e>pK`h&*4nwxlw((?
zOV}7hu{>SEv=DgILr5AJbPRD1Tf$Y^E}1taxZ3?HgdN3M5BYM@#&6#Vcy82AZiYCE
zcro{>JFJ2xjF_RoDhoFQ-gty7VRA5P6c)T&Ucc@JL2wA!u0M-6!<v3|FbnBmHw=Se
zH}Y}mR|kfEZhk)Hjn$4XePomXuL|>&w(6gEFev)1t@=yw#SrvRigxB<h?FY5DW(s&
zDfGYsLxBMy2w-58kBO;F`3ElurpDRsI*d_Rn_Q_cV{@>(-*&bR+Qq3|xxa=9(^R;Z
z_NTvPGz=Vzu?|eY-uU=DBNSft5I@H6ErV)Asf7I3jfAd(At<M8#z%?~<4I6Aj~IH)
z8`<naqm00$V?c}XOg-RiTzUx5;7&3V4vpRP<WyBnP=<rbg)@|zj7)WQOPa_~1sBSI
zoigFSU9|eiQzAu`zzl!kN8d5r4Q<N9q$<c9Q&tHX`A%Slg9Il+e=*^C4^ut}k>Jqr
zKLSU6?@AAsxi9ROdfUx6rCho;eqE=3qV&(y@MT8+9k0-!{>}G%`O-AXf0fY7fo4id
zct;2iF}zY5cxwI4x1uvyD|rG?DDpPw29{2fyYYzOUP#!tA)CC%6PheNK0Z}>l&b}O
zm$2|S1olqvC<x2@J^X|3eK+l!2=R6Tzv>?C^-f=pg}=gU2t9}nqIneHJZ>3@C@=M;
z)|WSoOX%XO_WI&D!uw9~Ij>)aPdPhor^n8mxCqS*KXb=f2B&S~+0(TDs5T$;9kLV)
zU8$@eT~UrUcW)(LV8_RY)jrA`9Lud+ymnq@eLif*2R+1;(0br<`SP@A{hhJ7v{hKB
z?m^B#hjw>^gilJb(DpYJo`K#CaLF2Xw$q+I`8a*q1Ma$RrLI@6oCE5N7a`mZ3vRsW
z-h`*JDGJ{<H@{TpnWb^1B$=EJzcansW*3&)om5Gov9Pq-rskHyYt1FwWIH?A4vbFP
zz55RXd-vWN;X2-EAIJ>t4u4Z;SsjPh<L!PC7;Ke6V|jJ0%`e?czn9b3^#Aso7O=PR
z$-rT9A@BAx(0Vz_Jxp;YZFl=wd-429YM~svtnuf$jK91*KWtAPeVl%8HX*#{7uVXo
z2OqYz_4mrSHa)%Erf0)*2QN1^c)Xpo$?;olFZmDlwu4*er3hwJ({_fq^RwNw|9RTI
zQJ#KJA3q8Yaigu@7kUv7agw&bY<pW9?LYj7f8SngZMKyaq5ST)dk??W-v8$Bv~PU#
z+XZtx1#~Gz^o_CL#CF?$@kz>|R6473=fYhE57K8ws2dEQ!Q<e>?ZEdS`1A2kf7%{D
zc~WCGpRr;n6W;px@U*;r@Gp1ou7}5;ZXbX2QTxlk_>1IeZLJJ1bLnH}b$Ag{P}W>^
zmR*e_6#_o;y!y}RwZHYWee%(d+w-T7%aHfv$+N0^VSb^EL%e?&w?vK*A`;(@5BbJ7
z-f#cozxbWnSR>K|8RX#LxD@8E^H*M1!EwrdMyeMtc8a%gM(?Ze?~E+>?muWN!IQAO
zWi(TEDQa^$%W3`odRtgrta)e?7~a|WBNY4$6SL*%GF=;dDyjv;M9MumVPHN;ox#P-
zd3+G1Ekno4mv7qM?m-!`7!eK+Ptx|Y3W?}qw7-8p^4t3l+v>_{dDhtp-$O9=Xbcs(
z1HW~4Sw=<%2^*ulyScfU`p+tI0NRAd12+m>JRHNR&4f1HF)H9{MWqlS!@V;2M&6p6
zo3C<rc6r<#)wyR34PanQZJ1jP9x~Xt-$Un-6?y8S%ZyMANxX{!*U&X*INe>l8yvb5
z7%;Z&6}?p^#pcY+Y<vIx_iA$i13jgx)a&cx&JeV~AUPcx6W9xnX0rqjMIo~G@`<*p
z>;66G&oSmrCI3{$hWF8{b};f1VkkTd?ZD6Co2?fNT)cao|K{9DhNIoR{W7XBW-YHQ
z6<!+e`M`*Q$ys~mRCzy5P1pLs`;LK6nT&-DGWv*y;<Fh6oyRp5oPjqsLc%BVGda7<
zp<Q83K*7s{DZ>ysSGmSJAPIhf?M%jEJahCYIGTALy3g=uoEW_Z_`x4|0UxX_=BYv$
z<kUkxvi@j0L!tVe$7z$KO^^(v<`nq`PM7yt=8N@}!OHtqcLr5sgIXI8WgO=-Xq$FA
zM_8ERvCxiw){=S{%owK`6K7`Ti>C0lv$h&Ha=E_OIoj6hwAXxtuRA-3-N>9Wtfksl
zx{S=*zDB1Z)%~uxHxPNfV8S@U@JUIZt`|(*clCEW-0DI9QvZC_*ScLlJhM*ulsxVK
z`zs?#{;u-mAN@n|LHYC?F|8c(`aYDUj;glTJ>0K$rrmy5`R?a=^*iNO*(2J857XA(
z-qGjiyZ7qKU$v=j)}u=94>t;rMrnAja*Xo()c)Gw9K6Ew;I-hZgwHVt20G@9Hb-HU
zQP(`oAi!@4-$jRG^A1D)T=3HS>ZrMqdf|E*-SbC(fll@|_IAR9^56|!>HW~Ycl+24
zntfyH+pEUFHOkiEGyJF<evOjv>KpRsZy3MIH;j*L^)z2bu*{|3=v?7J@n($ikkMF<
z+EPXJ_v-c0*LptO&X;%dp>5?gn|X<zbX-V1qcW?n^-iv#ztwKfM{vkr;m)Xh`U>~W
zSL;}97Ed1v<f*pCeDt1HxLse~N$$aA<0aj`HCF53^NsD=B<r^M*^{#U=%rP8lAj|O
zyF95jd=1mk#&^Gem21fVbx2>AcKl-}T<f*r>pEShuW_0T1N`2%zug|Z|3N98`!0_#
zy`fpPsAXlS9c&R!Fb)=1$|1rCp_W3Z?}#}shN^a{<d4Ffr;MF?Wh@HAV)rzrghik$
z!j<E^urL{6Zj14(MZ7OugvIkPzjhv;X2E<FVE~gtxrIUJHjRm$oSum=9)UUK6_Z3+
z5C&KIedie_Iak_X!9|XJVs7<?;LNy#d8%-K5j?`a=2;fU+&euvs{F#&1uA_Z4N>n0
z@AvR`6*6&pHf_<DsS>0KuzI5(;D>2*Q7T}>Dd_yR<O4YPmZ3!$jpS$G1JkJpz;m;d
zG3ig>f~lY+0RIwHz%;GGqzH?HDH=TN<Sipdn7#Bk^>xFJ&STKQ#Jd)$x}T;PTMDn?
z{%s1vj3J{JgI8WlHc+(K28Ll;UZ>565u!FXcH7xO-$C{3ns^m5^p?>uV*-ZoY}f&2
zOpG@}r8Cir84T=o+Fa;@z=-k7RlA!o@!rkfFozf}<ulI)L3$8M2(5%Pb;2Er6b!q0
zL8v{?J#AqeQ=Zzytv80+L*d)Y!thhlo{rvcrJxgrvm5NvzPxYVD;xe`8Sf;oa(E!Z
zVFIh(F#gdT6EGORoh4>&!I`u3(-Oj!+2_TZ!kzQb1~4Q{m8W9<)LBZ5%n=N8?QTsU
z;WXugP!p5`yx@9Rn_pvux=z22>DSZnWrF`w$*k=H^e=TGG_IeMiq?%3Nj-a~PyOmo
zE8pl|{vZ87HIve}+ISIqL*Yn?`oRYe+v3t(*7}jsDzwDfM=@O?|3X7~UNJynJkH8^
zaGv$e8KR+wDIT30=&CkQx>7oz(b|CiqGdy>`)NyU{&<yjorhHFa2A-|$+sxnQx>{N
zP#4<JGc-*pwJG4W%_BTs&&p6{eXvXa<;7{+-?bC{qLj*%w0Fku)aHiWoy}4{@<`AI
z%Ei}jz>{Jj?WCwaJr2LJA6iVo-NOJ;ETFf}O1;ee8?XC@18cw#x=NTt%Fa7?Y}|3a
z6D4E#B0S8ktO$iyxylPC{X0Kw7vYP#Qt7gt9BsF)t!G(V7|$5mdgvN3di^GAWMD$!
z%4oQ~vyt`fusI8LVrsU{&M&v=`FpL-u~Jv^j*SZkDGXo6ffrifI6g6*=UFdmTw3bf
z3GZ`|`$EPEMQQwvbNc%#JmbM;csb{91x6Wb_XDNCKI_q)jC6|Oai)!T*B`W7<Fm=r
zjawI|`&r)^p0Y;QyGi9OXZ_!6yDvUzThAZ27aLF8$?;*ED(~p<dO@h}iIKd4{YBuE
zWwI@;thLG6_uA~j2W@e6J>xpp#;4Ms@$jzNaUueOb@FvPJlIQ_`&C!Ro8Um;X}&nm
z?1z8zr)_)dY2g6gd1h`tu)1B|Rok1-+w??_qHz26cw1e)+t%0bwUxW~+tTW9x7D?8
z<?}(t0{+^F)s4E0rTaV2Q~qX~o*WAVdYC&(s_C@hR``$}`VcQT9-efv!jXeHUU%j0
zYIzY&8qc|<3jMg4%Z~f<HcURd=HVe@0FSrn>AApRCT$JBnYG^evFJmczc&iA_~y-V
z@QlIZG`!{G_LqPDXYI33KPlr0BZ-jU@|6)KZDtt2PaW*-m%+&#5LWQR4?k$LX|wQb
zJkcoDoeR=&p5bPu<|kaBuxD_>a!gN;ryuj>4S09$UdHEEd;H{i+uYi!`4(2M6!VM*
z&OHVXES-g4la1fDRS={>Q5cP=bSt2g!VM@bNJ;0Y5mFo5Kw3$WmJ;bkIz~&kq=Nw?
zMvl=jHhB4c-e-Tqwe5SIah!`Pn*%d5z7N7FM9Ik^I}7IK{3K){<ULQj=Z}D<*n?ak
z$P3!yNESlH6LT~-`u7oKoAzy4`Tz70exDMXGA=cuywYU7MlOOJm)~qJEF$@hkBV%f
zpOX9h>&u|lNv%)#$NZAnBr%4TrBF|xkY=$nuDrZFDSl{y-I3dtwlOJZ`?0sk$f_0w
zQC2f~R1vMIax=&M{=m(Hl<2$oU@^@sI9rHEy_a}?UWNqmBxzYkX++s4j>jD;K(B<>
zIT$TVX{6pJ@tSSQ=Q5{Jupb+@7-iHYdA>0nVF#ip;NiffMWnF%ZQGHeF5|%vnGrq+
z8O>iKumjt{^xA)srbpT>(ZA&|KRoTfp$~ey2U3k1JC9>_LuU@2SI1K=$+I%+Xw9EP
zWFe2cf7lkM084*sP29Z!JeiD1%AIjZZ}Vm$&nam5O;$U4k+uqRMT7T1gmSo}LRws>
ztB+5yo!vQvkr3RCc-P4xGWYgZcK1gtG?OY#<ISMNSbyC4$<0SXU`HAz@D%o$g1qA|
z)Si%@E!zFq%!W|Qt%K@<9DcUto1aI~xRF`1>NI{Gez-VjVnk9JwXD$n*f=2ppi!fg
z$iS%Z*PPKU!YzrZ!7%T)Hx=T2c`>zp9B0JLUB3y}s&cWNUikrx4tDR8DQ5WIjMe+p
z6ev6S8|!0P5<{;kKb`G%hM(K@*K69J0*~FE+6PvYaob*_c7Ptt;6f4ifLI_4+#;W_
ziv_@G<J8(CB;>)S=_$;nN7ReYbci$;B_R1Fk4oP5BlvmR)I8+G_PAv%?5b4}XLKFA
zx0*e!`>FQi>w<}#8HT#~u(?&Ba$?<pHOv1k9K-^aX<U?_Q3Iae<vngx6WzLS_6qC8
zgnRaC{LMY5p+lIai?zy@{KUWlRsj=&uraGDPjdjgMAZEtznc_=4Ky0yfowyGVCeKJ
z0b+5sz~52G&qF9G^LXr&zvwoc@S9`pYI;`Pz%T}Q9WcE&A*-$iyu{DaY~f+{W~CC)
zpw_30hlRMNN>`bQ^qKiI#hPE#$vM#TJ&B_Hti{Sv+w{P&c>~q<5|O8yE=l&TN%_8G
zZc`XWiw-5yTpFT$PeIHxt2R0;^LFW>T3}K*)bbOtWU^(w&2+3~c7GSB*uH%dm(*6Y
z=T{k6@zwhi=)75bcFQHhkaNLh37@%COwXD4I#JzM%%)73CDpO3@_Zd`piy9!_?+=s
z4?%68tBQu|$<tSuX*bC;`e3~~VGMPfCM~<kwWnWN0%V6{adRZYcYpDhMX2cgW0dmU
zYWp*+%FJ_17X!=}T8}ih+6X-tdmI{dKY-__O@_!P1LJU*;as!XBR`-sgPmKxLY4)y
zaBnQ5g}G}bHZ@Au^Y55&wTw$j*zDfw;i&q22h{`hm*VW!BNg^{KBL@yd+s+|totXN
z-XCYX19xRw7;tJy3tJH3%`%S`T3-LFNG6Vj0m5&GsChi*qHY-w1BUuq%LojsFw0W^
zDiG`zk&OnNC*Kh`*=Y}8nYB@scoZu3O*Az$?ZMwKx`jXeEi^f}<2)sYt;av>QYebO
zI+QTZP)(gf&N)T2Si4%-Xa}_Fck+n|CchwtGLn7}3-Wuo^7P`KoA32M@u>Lc3P!_x
z{|VRyld~@_8>UtuI@){{F(}y}Zj$%w#RTh_J_^{VR(wN*3INgaN$?E+?>>n|^GiN=
zfJj}AhvFG5<*tZeO9U+dMP*Tx!Y~32RYku(h?VP{G9trH@xZ=7d>F0MJ%{DJl`X@h
z?aO+GOB$@c`3vk8ACjHt37s&+VlC{w$!s}(+UIFm@=x;#KLhd}X)m?nx211YR!mlE
z;p?8mh=-DSlf{Akp&~h%?vuJ3o0K#I?RW#HM5Y)eMS=C4aqV;Rjdp_0_2hL<d1d)D
zp6!!fB&~O!2wZ)D<It((CP6y~S^ceB8x7Rv7i~rtRW^klR&c#Z7?(e$Mg7s=&pRng
zR87nk<^KlJS}M!DPW(*`lIstB_Ge5}hiYl~GeZ^^PQOdS<L6K}*=-kDjyr)>b*)Kl
z_lWB6XUBr_qCIFQ|NeDOPx~R9pHt({%mUUJ85Ig$YMjjY>d%6dQwRbYQ(KBn_A<Nw
zJq;nD?3rAFcG$rj+E6UE&nbpK&RmasVe72<=<f}W3D{ZicYoSg|L55zOEMMt4T^5b
zU=6>vy7fzLg`uYQQ@hq!d7h$I-J{RMHE+D4yk+D<i2O!ut8=NKU$>~Z&_6rLN8s>T
zxp|#TC!U_aBSzS@eU;1e!obNJU!(AaD&vu~MRr3sapd&;pbvXd)0dxtF%L4b_O7ud
ztavwo+(A>wZq^H5aIV`hRl9_PaIj7#a)QI2Xklt*tO@_&bKk<Ng}9~`i1PK7a)wtM
zzKlEZ4THiaM`Jk}3X9W$y;D;#Q?rlmR2nz&4H-HZYZkL5-(@*knnm>6?{~fbS>NiE
z5*tO1tnQBNl*NeZ2)Zs!^6K0<Vg={}dqk!oHVFxmT<pv!{*dnn0T+6Sq4_snne;pq
z7g%!>H5qvYe`M@RYP3|_(~~4sL->h;o-5X0(kW`1$ek~q7*nyix4LZ<MoXxf)acA#
zEk^C{-bkYXIl>Nvf?!sr4XI3TWlaEYa7`~_u>T1l`=jHK3$6jW6K&jCxctLw^cisW
z(6v=-q~b~8eg^OD`E(XujyIT`8$9$Mn4*Kw9I~9oY?LM>0DQaZYJjfY$igOFR0<ho
zmYyyesrI=Y>}P%1Pl4fL2~tDO{~Kbv&oMO@R|h?toj?4mw7ukV<Eyx_vYM{5KWT3h
z|59FA`E=~lCS{r$m^Zyods0-L*uvxfz3GRSB4L7VesR$a*+=zdV^WFVc4*VwYeN;{
zbdVPmn5oDq8p27aj8V^FI_E{-QoJ~Rf9R%4w#hnOTMWfT#3tR|vWJ}SM_L!DW{290
zeRyk8J((`S{NZieFMi_g%;;)*DJCo?TwaentWE3_wNTf8zF7ft{izp85UJ4U$Eycr
z_A*@UMW%`^o;U*xQ5EobIs0goPx|uZC~An@Vn0!d39^p1giwB@a<dnacybteR}yP2
z)5WyZ3YEr<OhJIqt@+-Yh2zbG$@=UiW^G9DD}%V0uX+PdoZ4>!%-HeCVJNt>=APx?
z&iLHnlV3v9+V*8aqRA63zHSBVPeqXeY$;-2WIOW@{|-W#nC!T+^yMY)v50;gP~UeD
zI-Ff1Ec0TWFkBgA*>4<RK2DRoFHaF`q}3r_)Gg>?$gtGS&!mclR?D)O0zyj1!K~lG
z2lm>Eys+F_bf5#Gc0J(TTQOum)9FSkwOEvK15;D``etWu+u3|ZUiIgph=wUMt#Hl5
zXEHL%WIwA!*arRVcbFFLl6x4&H8}<ZDRyTB<of|G2acy%qzsobv-@~T*Y;|Q<pL)g
z@_RVi;@`i2JM+p^G>la=eDIfi@lruV)X=QYsA-*S)%H26^-rjlT+HyIeR-{X%5N*X
z)u-}ot6ayMlZB#s$uC+3m(~-?uAio_O}-dsJnQHU^r@`yf0}(uoWKmvGxI2+AYsY%
z^A9F+Jtt`UZ|oQhmiFG(8rL>Lg9lGqtEY(gkHv>;B`X1#wshsp{8GS)$u{VycBsH;
z?9+UyGStPDA{>|JZ;@h_tGJ{_#38=7;MPh_mD^xr*%x~F48p|Gw}%78sPmZFZSHHP
z$4nC8&TJnYx`STYOA6R}op9?+WFW0FDaU^=h@Ah^-2Of+oYfk;EEZQLyxq+dJn&O=
zLKXZ)K_H<bFFFLW9B`7;i^){L=$q=jmY>`rTwgq~grA4<5`A$u)~i~Q0eBfGLE}{&
zmmVO#nUl-Cu|IQZ5m*Sr83Cq815Pv{g@e>I_vFkX4IE<7-;~Y}`r+QF!E%3M31X3>
z_D`V?Eo_26oyYaV*}UY13!*d@VQT-Ha@{9mHasXfC_G}zcd%-a-JaH+k)63U(GTA;
z-Miq^j@-O+xq$PFcwT!sR5^HZ_bl+A%UjFi;S>9@tDH%u>5Hg}K-#S+JTXdyZTbJi
z4yWdq;xmiaqkhPSz{`e6r{&f+jpg4QA#M}Db;WNP*MifG87ZEp(#ZCTkxAH+9x5>t
z5`1A`UFznr9kMqPRF=YJkOmxHru=RRbCMW~Zz-dG{3i+rBG=ZWUcpj5v!2{K`<*`{
ziOPVa@0T{IIA2qv#x61C=(}Iy>WdOOOG!%--oH%{zY~uFAuRD;)YPg1kCm~D@>Ni=
zF59pcjhA)08@T$rFHac2gEHUR(2&l}r_W@Gsf@9Yhuf56#ovX_9O-i#+#|MN`INE=
z)E3FHMm>yXjFU@rpnBO?@BnzORnKbp8;yGpyWC^ckkd4lsCm_wk4`&{!{f8>O%}#w
z#_L@Eb+8tC8XfQ1%@LG;DChu7ee+XaRkbux!B;XE7&e_C?IA+OkF)*Q>mHY~F+VzE
zF%PSsx0qZdlI7d1+@Z6z2=9j7xli`PL!bJi$4_RF3?seqd@C@I%x%l%&%fh0M!|H#
zo2@n$Ue@#)kOyQLM%H1yg9}{e!3XKf!54RpC|XHh&XI@?IcC^J@8W60XKL^%r@Ceg
zR)D(W&eZuFib{SNepwXnbNyPIZETr-iUj!H;4Hu@Qf>oBVd61$Y2~)2pJvV%*%9Ad
z8}wOYyl~vpP`xhx7ti`u!y$jL`os?)Z9}gsT{N^=kMi>Yf0lg;gYIzL!dUx`cji^l
zq+*S3L|U{Du{&QoB_M!3+eKv^>s_&TQRyh??^6ojXjyW*-ni*6%o!Rbf3IV08-H{>
z{eyQn8`UlqHU%(u^mJtWC`j<Y9?}wPDw2Njp8SL|HlN7ae^j0j&RX7l5e6&A7U#~K
zdQm>$c)+ltpG<F+SRsYUFwaO9kM=W}_+9oJX;VWoR*EKw4gR;jWp%)KA;iIpl^McY
z<~SW)ZKEPU)>{E>I$^YW0c~zk8iF8ax2HRUdXVt;X3eEVv)X_EdL)#hb%qPH2YbKn
zb0waLrf1@oJteU=)b~HpF9lWB?BI{rxCn9dU(xJJQSPW8e5+GCMt#`hq6l$>yh!Lc
z>_;@PN;PvDpm5VyMWbi*W^C|la=B&f#Wnsu3Dt-&#Ra-rX7*<ik5D@?yfa(eS6^nS
zD#Q0K>T8HX*{0|LU&&|TPvq8$b)7|>Vd&79mr5)ErX6FWtoNgK(UQ;fCu>Yt$-8wv
zdeu*A(@J;ib4YEJnBe&AEXfOaokCtoXDO=7iVV)%=>9Pt>u?X;nY?Mom|B>)pK-uS
zDdJ_M>VEvr8kn|DN~%Gu-~lg8+TDF}P2EGhk+f&h$ODG4oJn6;)l!CX_YSr*#J%=x
z@mf3VS;l}nS<g@z^eD<ta)xdm1#M5-nFsm1wV^0D6{N(cE#eq2G=>D%K|ZqQ<K2=H
ze9mzv(6{J~u?+LN^ybU~?*-q6Y^KH5YqWda*R2-pOy<`PWzsu~PpG}!BRN5U!q3w7
zA45BNRs*3z`#Dp;p8JsrlUU<6Zeq4kr#B9aS(ah@HF@CN)kynAgZ-{ayPv3ib^|A2
znH;-B*a@7yn#Wzgm!+oe?G1I?*^tCy-5L}L#3lC|qDSDgR6|DhaY=7l{6Ly|@8SMt
z5vE3bP)MYWO&bjzK_!ic=l6s~zd2+1#il4*x3#C5o}KO#MO9;;ivGQVXzPp1GUQPG
zkiIw$KvVD=0?Zp4{7OXEV-hYAKyc`YX0G^X4&;HEyz7`${d@ozo_;z`VIcx;J`}mv
zZV+CyW^advhxek}f%<f!&tU9XU%9`me7lW+(udI@)+Rn7F-*;!x8r(0{~Z(E<p9gs
zWmsay1La@?E<pgit$v2s*TisSup0a3!qXej?uUZraZ;j#-1msv-EyyMz_YtpY<>nu
zelV}WNaZq(#W&5)pJU+FHl0tT%|;XVhvAfa(GQ#A#sA7^7sN>I0=h>=u&28hs`JQg
zG6oQK;#Nk`qaWAd)OIxw@vL|S3Ecx39`~=Wf|FNu0P11ZjC=3f59ar6?0tSKJ#?1l
zNvZm3YHDiF_43`4bFM9tB1?`MLMi?nsd>vC%(z#$aASt?)#B~ZR8Pe&9xm|1{Th5r
z#vZbW6J!@2ABksCNjrl0ZB493ptqr}3xMV0oHZ>fv%LiQ$OsN%Jj0zvXWR9G*1g}#
z|K^phB~Cjj(gkBO_d7f|gp=R2>->Erp}@w+ql_Hja)@ETFhpw<ytnE-+&%u)Yic2r
zGqz?A)cro3iEd1U{1#I*^$HO;YVkiGu7Tlp#PDX^OCh2j1F-O+nf0_+Bym7r%k0Tg
zIYp4EY?o6cN5Ts+h-k3dVvSQMEH7r6gzRZy{_f!lrR%VCxa4X+cChs?@0xUD%J8mw
z0FKt<Y4^cTGOz1#o<p<&<!iMJb$O$L`l(rbWrun3r`R?l|FcUzNCRs#UOC$*e<TF$
z48GcXk(YIrdCB3VCUWf1@0BE=vHd)S(ywC=TjN}(V5oTXa}Zt_J#l$bmv!;N|Jj9r
z4vLe=My&Wfm}>BWo$%juDRm>3^Pd-j^Mc;2TFMkB#}9dh%Y|>g9zXyDS3e%TR^g}1
zD$VV>p7dPjP@rM|2^CL1e(ymG=a;8=DI-cA-HPG?2SZ1?CkjiEH=riJ{@%O5i^;gL
zPMf#WV>e^39v;V+s1NHUrU3a))M#nJXS1xEp)O15nUf|;HP|T=bUwm&N(TP}>apbB
z=zgSDM-(W`Gs>)gdxcv+mD@C5Ke2ehpUpoh^K$^TQ@g~eSE$npZ#B=w8Kg+9WpW1|
zJ~UT{@R~GO*umF-n*s+4G6q-N9X%9Uk@Gi%%J1pKMFadoTzh5(@|3EA8$Ka9`>kzr
zv1MyR{suK>IhDLKWqs_YWtia7J$sWXwCuK<FoA$8;j0a8lZH4HD8)o*tjp2#UZ~uY
zUE@44suUBe7V#&VzK@KQ{#&|u_)I)H<OS!=sn)k(I;{R)r;&nZA<aS77ccUHI`>nC
zPW-hqFa>42{+A;>4=Shf`im9Mf1~MyJ?m?Hg`JTw!6`el%6D#Y7;4_Zv)kaCTu;v;
zcFCxg${a3;6><;{O#OwF|8X^!1cN+6?~L!Mv`}`-*u@K}O2b9ldIKSJ$|@!W!}smN
z65jq>Wta)g<bFbh?zIAc@;6==U$dUcqYR9%--&BG&4eM~i+c+-j`I@*1X=6z#CPXC
zOpkc7&!}(p_B1-?vb(0<#m%EbZ*K#?sQesZ-q+p9p_SEt{gW$0TT3%3xFYVIRpM_R
zxq0#YKKanM_>`p=qx(<f-i5`fTgx)sm(FsluTj=qaW}qGdi*;gY75vw&p`|)f9kpc
z0elYOeEN-r^XMt0V@ytB$xVe>=&n;6t8B|N7wRK~Z$>^0JU13fp&bjL3%(Vm$I1HX
z)`El%LF$xL!pW*l!-}eg-;9KgB)ZMqHJ@vcvnQxoc)NGJLi=dhp%=xbz#KZp#Y(9B
zl#uAMXImBZo5uCHMbIo5ho`7KM9AGfDt^;rwso}GqJuphJzn!p!bz=*+)miv{t1({
zO1hmgkvG5UG)7a)b%d3spRSmk6)<OQ3+SQ17Jt3+`rA@o_4{Yw8)5%nUpjPzghv{#
z=dX#^$bie1rD7VEO@^!vmJ-tE18(5nPtUdIlqtWv<w4B7wUko!xi{-^Gw+|jh4i?D
zsy-A=s9SgyQ?IWn`6Q?1;R=9zY8xgNU=mnbR-655d7G`mu2Z?+GO}@8?0MYZyYi<Z
zqUG!BP>@?d6Vhch>QqETxNa*9nlmxfRDfJN{tuiv{%2gxf>t^dT~qvXGgs2?5awJl
zXf?d;(?(mVd?9a+xNBedEoyo&J?9$Ihp%mDK*h$!z2cPI+ZEy_?{k5JPa+R^2<Cv1
zW+hXU6#M&4{~48b?Mus6IVV$5j=J*l7)k?>aK7$~XNl=bx&Qw26kH8fEUOd5z5nPA
za&s_eap#1cV^5dS<aNnOg{m`rF346Wu8yw#WFFfI<*dI5OWK~rif2hEacbvj`61W9
zGXotzBlxEZj_NOVClARa$)dkZN*)ER&h=AaxK7%gN=wfM$7Nm$hnh({PLsv>PQ=nm
z{T&^u|2f(;-4Vg?Y@a?$bl<&F2f4l#uno=%kSQJd$ix<MYu(BBU+kGB>9euF*$KY9
za?|B<)h(H4Gk*_1rkhx}H}odg?7f602k)r%+KPk39<_?PTV&VIZm0#h>=V`+A5)3N
zE{B9g?vjam@qzrkk_l2?nHbsbZ{qaT9wJ`rzMYBU-VRc$>}-CE8RK8cFJ-sr@-Iv4
zo9fJMSI)5M@DMXPIK&X>CjLnG_{d?+$L~|)^woMhA?1HzktLTgX{j6HB)*n7ZF#K~
zwzqowVhb2I&v|u_U3-mRx~hU79~1`RM^;RHPh+<crG+}#bwHi$<Y0E(P=}ao_bLqy
zcsjOQ@@P=}!1b8YIUf|{)5|KHnw93v8E_r|QAbZdO|M17ZCI|o=Rm%p<&xgT?lojz
zttIT+6#B$j_xw{&K*vDz#fPb_Z>v}>cXs{c(tMn?NRX(^VC|ubG)nat_0yvT6e8v)
zrP3p!U$_f~^Y<35L><xvhCO?`kwbAVcHV!~vtX4tuO&NRPJK0c?0>Ta2apW?@u5QM
z7k{)VwcO~^3O@$$G6Xo|2}Zn9q__BGpu!mp1OVlpnc-O3MvhojA`A@;1wXJa&NwKk
zK346e^!bvtZiCP-&`5VVGTaP5*tY<9+zGrPxs$IxB;iccC{kD77(LT-l?{oGxap}b
znwNO>FwZN?M%AO+CovuOsQIf;!WQxZy#SYckmvedoBwKJ>f8N{YwnnplOVohg3s`z
z3enRDdVv3g%W<2O#-r=3e~AHhe9kmnihz{NOyclK$g%<*Pp@FsQxkGZU-`T&EsMVR
zyAZ+dHzzWFG@O>b>fQnu{YE|tQvcEx(o+ulZyxe^!eKKxUE<yF3nMY9fNd#u@Z?Jz
zEdgv6oIwA>;QbZF03{td$uH(XH}DL<DJBf-FZ6FE9h-1G!Gk70OYR+;02BcH!v=^J
zUnA3jp-#1I(z}hita{b79Y$D53`P_X^(NX(Pv(cu`T03Q9#Z`tJ(A&5h?*gRM?j}B
zwU+614@FASbM)sXUqRh)KNZjvP33r%qCNWJGg(fC-M3A39P4kYAIjy8_D!|ppF*U}
zh=!IM@Kd&n0{ZqHFEFsNGT<WO_p_aez|Vngv(SNbAQ6hQP07UjjeSykm<wFQWgnii
z>P?%UoJ`G^>{amOu<9<2M>sb`zav`cxlw~S+uETq7MG@D;m&of$jZ#(8jK7w^|vU%
zgf)bBQw;V=yTDOlyN^VGSy9cxP6o&N<5g>@q?A=Mp)8}hf5KZ({!`?I7N>oz$@Ow1
zM1j}U6;lTr6eQ@pB+)!UQ|{sO%5&S$)Bf2vXm{~VbE&}Zs>DF1tMAg=iw_Za_`6Zs
zmaAAg@KO!f`Z*?7uBST|)3rgxGMxrBvd6418hK#DEwZgliPP;*E_OgCjdWr@qJBG_
zUo7}!+qXytH<NUxkAHlcZm}xdz52Na^*hczI#54-vZTj>^z`><$U3M(d<J<GbyAo?
zTiHj6<qTyzM2m3MsY*XLbNH3~<Ty@LqkZ$D@)gYq5wY{?<XEK&P+|GZMC+#c_`fa+
z&4Zki?rU@dlKke|jiQkBbYQY(Hi^35*<Bd+Koa64K4uflT|FOe_3Q&z!8a=-4b9NV
zHmI0p<OXTVjSJF0B&3T<wa=+{d`!bFaM8ECVztynN&e?jaNTfm$d_a?9^Y!JWb(+a
zmy8=ORMO*Hs1RMfPc6x?f+aoK-JIecmruNdHz2EoS^weTDkAqXS4AKyU0b5Z*bglR
zg*3e@5KBpRs1alfqv<2;p@pox7JKD<=DkYa4NACm1be}dn-`jmZwXvN0Z(=Kz`b4b
zLqAL0->SOv7irnOJ3XPH+}$0Y&;ESyTP3%I%e3Ud#tYptxIz=pSkBevxQ=}=ZE^fL
zL_TUo2hKOVbJ`2B9^&liU!=8dw-~#>Tu^H&;k(WDueP|!YN>MHzEk7sEV^&Z0Y5Cl
zf=+AZ19J60e-Q8Bf!X{~e>*lORvk<Q7N1JAlCE+EC|W=0n5LnI27aocu0DCFucjL@
zA0(dBO%PY@rA?3*`xE*fSo6)AD>d$64JTq7!e@nt6?^14S%$gw6NKUvZL<%Z{oM9C
zZI2ogYn-I-hSHno7r|*Z!ie1ce#(tH-2TyU=N+%1nq>q2l<(wh&JmzJXg0$bj6|{=
zv1alEhhnuUYfCTOZ-{QVsHnIvr{9Je^GWrq7$`pY&UsI^QY2uKVeb|3LBY)}St*l4
zr6SfQ6tmggVxs%Nn+D5#PP6##Y{qc@W)790h?H)>h|&9DT*yZmCY=XUb}xq~I&t&x
zeAF5MnMeK1xO4d|qry>XO#NTtU~61k;_tFOKb5_M3G0J<>*7h(&lzL3K|7S4df;%2
z0}*D{Y)aAbq#edA-ac}tYZb>;ZwSSRaM_1eIqGp2)rhpFBwv>m3)XZqC;ggIkeizT
zVO*u0==UsiDECW2QGI+*#mUgUk2Azu!uyaDz0J~qin&HH2!Dzi9mimkUQHAoa&GF=
zOiaz`?pDI~*g}`hphE`(rGS`C28GUhc)PIc^N`djX|5^%MD2OPWX!u;^a`85x;J`9
z+OixsN{2#DcPDg^z-+kLs@m=LMnv*eA$!g>6H}iG4hY;x{}wap6_s;_n2e=jpPgB9
zi4vrnUn6|vkL8!%J<QBBcO3HLte=0dW0F8pGk*jywLSFJx|KULlo&X$GuZ3`_fPX&
zsC=OX?h0Ec2RT^3kolC0klEK8)irIsE|-$OwmuhG)R$+odj+Vv7gTRR%(w^Z>vhoM
zeVTJs&~`LZf3iF2`{TFofg8VqWa9^dJMtKx|IMOHjH6+oWAeg~SB<=<vMZ%{>!Kq~
z{Dy1puWJF@kMXQ%2q{de*sWV_-N0InnE;8e12=lpL?3jvX3geRql=+04@JC^=a_95
z*HBY6KNLpupQ$E=V5on`#Pne&$2jWL;uJJ4&-C3(56*Ym@fMd?me(rVGDGL#qOG8W
zsh#?x6_ZQU5-|=7UAnaJ0GDL4?KKc<cPg(AMUmG>*C*}uoW%1QsfI>mvH$|@jhs<d
zaBG{>xL^W7*4?1{i{$opa`H7M(G3GC8(M#1>ixKYjd^U#648zSZ>P87(Z@!fw~vpL
zV&0tF<5Km{YcFp+6Tdd%NI&T+pk^Q4r}PTgwiKLgPG%*u8M&AqV-&9e7?r)><>ACH
z2gk_r{63hxO6jz<xO0w4D=x18219WLS++a~xGH1xqFzrxPZUUA12GsR3<IDOU)6`r
z{I0ool;+8+k=pRkga^Jc(nTM2g}$Dk_wKzgrZz&3R#|}|ZaD_j9?;9n{{>2@Tr-o_
zU4PF%EYanxulGY1W``Rst(y4p+D^wzBMv)n1*9v2164yPe`j5=tZBD>_4V^9AA}yP
zs6Yvu@4~z^6DBR5r!i<8-(#g_u6V4GH+&QOzfL)0elhN5FJ>d@a4bpM4GbsSkMuG$
z0WeBpMvv}4dM`6jJaBNxbu%_LlCj6ZLS2v2?QDGO7Vlo&r6(dxrMQJ?`qxD>@eVw8
zaGEpcbDB{zJl-Q~@}Sm9H#DT&x6xIBN2k0@$Vlb`nMb&~>T~Ne=Bs7ZOj#|mdu9D&
zf<qc0IMyx6U?a;OE}$TP9HvGAcIoq8Q#q_OiB$LqHS-7+l1}!G$-mMg#4WG7s*XL`
zi6aHpy#v5qfi806O$X8JwtSWHE>Ba%gytMyi<^7#%S;#{wdXr??N-M{yo&88Y6a|b
zJG(KOE-w%dgu9?fw^;kfYnCa(c29EWJ%FoWo}c0}v}2z~j66JILO9<{VGjKarxkOP
z^iNA4lde}PbbX>V9|Om%6-EA_+Sc-X!K6X<PLa8FI6Y+ib1U4!mm0eN%$tk-bLo{j
z=GHKNI^oLa?j~P$>7I{8r_ATAZvyVbgi6cr)%n&7ZG6)|KHB?sr+kXB3^?_-XGt=@
zHo9mXZvOiWM-38CPK92ehS*W8#AoSE-pzQ0T5;S=Q|iE^du#cIxow1RKaY7*_E_o8
z(ADv6LW%8`i^%chO|Vwp>_})X5kJAh8Vts)sB~qgspY=1B#StJ)j9Pd1_%-b`;@3B
zArjTW%599TR$QwW(xI3_-ii0e8GGTLoFjU4Y!ePJbDVyqQOPx&vr}t(3GX$)>Ql)g
z*HJO{uaxvuAxTldwB#3jE6mV;!zIZ)4OS3+J9{KYn@N@((lG3|l(O&fBhL5V%^STc
zsF79<mBML3j~$WrNpm+S{6r-Z!28s4gJ#EN#2qAnP=9uA%{HF3Qnzzfl#v(g4zK(;
z89~bnxB5~#vTGs<XdN4KyzY3!NVr>n5+KSeT`}0wT=G1ds3UxF9-}dRy>py@vzN7E
zl=*)*hSsDpX)*39>c)Imgw0t1509!gTU?rXDI9*uL91Ua>vZYQDuwHGMKd_={;f-%
zmH1A6VGzX@M;*zh&J<OoE3v1)Ef{qc|8!(;yeTY<^7aR7T5&nZ)!{ls`n5_U!WvC&
zt4Ce~6c@3F4;FhB>Ottl!GgiP&ZL`3@e(~EfA;b}v{dRFEv=)N9KuD&sDe3q`&NeM
zK0AGCM378ahTY~X;JyHGZ^I*_2n>e;q$*+ezgBhA(_8mzn*!hMkN^1cdHI`H&hvXv
zL~CfwZ^?IH<>1f06eeA_g5QgB^X!wqSN+UQD?g`;r^oZ3`G1L>xRVk{N~4sh6Z{18
z{$Gz(o*Y?!D;Dc(*Oz^zySCo7dGcyfQEZS|;J)0pT%}l6W?g^(GZNJ*=3i4hY!tby
zJu5RuSl@t!7(n|31$CF1rRv-XN(9!Z@%4?#ZC~?S4>?LC{a(mYhtHB2HLjAG$tilO
znn}Ak6j2uJN6~{a$j}{6qa(~w%qrn;gB36w6|=BmoF1Y;d$q5_Axn3;dAYMH!MKPO
zY^n*oPw*N|j<1)UmJ8QFyhzfxCl<%m2|*J~kjcKNUu!g1d0Bdo3zvFI3Ideeuq>KR
zzFS+OA;MvNbV`sBKR&OtE9#b0SVVWIA(_3mGP|$Fm|7|m#?i+ASEglFV%G-pXy+A?
zB>(?k86cxS`%!xGN(^`<30TeIIZUw3+dUQ+a3<?>MSDAtcZYS1+ei1@M9sjyyv$2?
zJV|ynCoPdT#{*)>u?2~JB(5!^Oxmj4&YxQzwaL-VWUmCeGl1{A)%k!rFoz4&#ag7H
z)b57j&|zh&RxQIM%#Q!J$r^?-Nw@587H|C_cjR+A*|lQkxSp>4;K)KT?(2CTZuFJB
zz}?ta5Nb3(bNBu`#$SJ%YFcCIwr?g{KZTXHcqaT-;VaM7`=~uE;U@f~&-gs)tCW0_
z0%>xcIGR@UmXE|8;~{_}UiC+ML-TK8dKb~LGuBEz^rl-r`+3zV2Eo3k0Gdx1#pB~I
zw_T+9IQo#gn%e{SM1(}@!k$`aDvgcgn^_mu*o5MKWwc6eAu5aYBwCy0MBU4qntEzM
z0%yD|rpTTxPrRhw!iGs{l|T3yKLK;#J2RJCE9(N<57xs!uPUbRdGkwlwFczxTTDUn
zQcB$P<Y3CCqsi-1)v8P1?Um1ADvQ+$r6R1_ocngRJd}-DhL?8tAf{^@O%sbu&<UJU
zxI8Z0I?{yh?-T1`X8{0sZ(p;KP9}f4IH@05xrXXMS;Xk06`#%`0sNcf8Q-cMwi+Eu
z;C?CCiE?1C({I8zD{w*K@>~rR=*|n$(g?2fu!G4VRz4uJ>W+N&`<NH|nOW2jcH=L>
zuhG=L9$+B^TKf_gNKnJlT`$wk2VPVLj$fhLFI!JH0<VnV0SNy|4)|#qv6GYFd3n@M
zC|zN4&Bzh6%Q&Bi3q)QW1zuHRBNWfA57b0&Fy=7)yUQ&dMG@s0C@!!RzGJ#15|@+q
z@1nl8ae;$Os^ao&NznbNi`-Sn?FnRibMpenf@!5scKiyuo7|qjK#Scx&l46duRVFO
zAu+<_-Z}RtijPhC;NeZf<(Uq~a<S(+fG`1UQgD)$o$u+U?PbsQpoj+a4TCroqQ#>>
zeCusQC!Jhe^hO=B{H0aZTfDtZ|08b2!-%<Fd7~g-7#$)&%@&t>?5R*q8{ogw;}E8q
zH*D+8rN(F3u5WVS;3H-1NXhfnH-i)h=%=G=1~TeMMq0=yf8ELaU}VkQP+uQ1<X6I}
zU(>Z9j^J!)Mb*CaQp$#CIi&#C?y8oon1=KR$jT<8gW^?WULMCsJXB`nxx*8MK)C(Q
z=qMdx>Har(ZjCSn%p6wuQ^h6pEu^NV3K%DN?Z+a!8~=3Emf!7*>0sd;WTj%uP;RY7
ze5htu);b7u#AKQVVa?Lh1{SolGxNjGp3hhffg@&7+AW>WRmc|CMr7+CRZ#C3pdB8g
z9nQpIa~LeB(x{;)+N9YiTpgzsW%tIKYDDLVp2kGW6cEDqukKB-es7HU^jyN-{^+&m
zFoM-<K>#;MJO%ww`4;nPvC-^aeNSzVyM!gfi))}<KLdVbvyeFRMy~d#5_}%NHDP~F
zS_o(zwwEcQTh<Pc8tYF--Ln>!tVeMfoT-w91rt+s!vJH6JJJZ?lLoZGl7>kCTsqa~
zT%GBq+OWUG-t^ANh9L}c`k=<WR_^6XciEte1l}}1)^z@a-SK?A^`MX!?N?(hZC#4k
zaX<8PkR!b|V<YqM;EwB5)(4Xm$cCP3IRyE=G_X(9J1eJ9`@z)oQA@L$Zp!?kMgFl4
zO90&fe22Kdu(982ce8$HuFZ-3e#c*DGPwkn=>Lz{hU+~5wY1J(9#vTHmdGy~i5k7T
z_}BM_HL>s=?ldRxVC)PXAll9`g|j%_*ghC*#iMYS7S)2?-k3muK%Kmg>Be%xtBePm
zgr0$6L9ks-VUr2kAf$aWU}C;j-(YjaX}=)a6e#tdd2&Pwq{C~r6-l4jE<umU>>`dA
zQ&O<9k&}DfZ}*$NK2}rP&z0Z8jkhbzoX8i)0pVJAW(Q5yPj>^FQcA6)^RCPlWtZl!
z39Jt@kEH^Wj@&Yh&bCKeL=Emt=9muxhb68z)4<ukHfHpW^7q#7C_)BVg%F={Ggk`0
z%Wyg$KV|ML61TIl64xo!i}Au&->pMQPJ}>!wxLBd>hiB;pZsaeLZ2Ehc3=0);<#rS
zkIeeNTgv%2K(mdFglbjG{BQA<J|zLs-*J{F_&#;Z4`*X70ZCg?gPaxrU()!W^!-12
z*Q)}d`(A!VNWbeI+<$zQu4=%(_4Ocz8)dMz^V9cy{&xOOeunBnyW3s&Z0VbWga9?E
zR^)>+iMUi++5OOqT&q4;tN+AbDnQ<5=8Hrn)l~g)^%WCS)=l~<md_rok^<pe4T6a1
z0MOK8vL=bbZPb>d3#rrz3J-J0Co2w^LBvgv-Dmpm6@p!%3V+T&CMuI*`AbfBqLjb)
z=x>0??odMqt9bq@jZS`jJ?ocdpQFRsq6WZ#5O_~PEVu51`(i#R#X5ofmqY!@e!{!U
z|8Xg{%V-%SUYh@l+X6E4$XG4hGa<b@SS!2gNBt%yLE%wN{HMs5L+4gp;*XZiRKHE>
zye>Y!h#)tzPilAsCVzHlD@Il+^}35K=ppyR-R+lTJeCM|BOt)(Z)#*5`<sp3M!~lM
zs}I<uPnWZ<Y-0KOnPZ^+#k7JRzcg>@H{SL2U}|HAM#{*nFOaKquIs(NKNH)XmvTP`
zV+6c~{ghw(>)tPKNldSi`5r(c(!}0m0p0l9R?MLzJTBB$i6rO2-i=d(@VKx4G3RAM
zwpfR0F#5->1<z14-mpe6qbi8jQMofnuiH$^6c1}Hx^f{|1>`$W<lczVMtKTMm{~ba
zW~kQ?FCxVBeol)oeW8{kiSdtwAWe*0=+DF(_CU)`*w-la=Co`-wCiF2JC;_J194>b
zHxP}kZjLB;NBesTg&L?sWywLN0?xZRlX#WG44WsDyk}s~PrnCod*6$(Jui{at3O$Q
zlW_n+?*~no`1vMeTv(gCUFvZKvt&CiOD38osEf$Nloy?{jKj>47d`nBRSQ|h>HshM
z<sTUx8leFXukD0^RTSo4xMyRoa_?pqrIlP;HPZ_Dk7Ru2mRAv?s`r7}ZR(tTde>%)
z_ODrV6k~5z&GMzBno`3@3tZvvF7>h$mPBl6Z6g9UD}97^)DQV8S-*8b*(k{CGdHAG
zk4I#c&RhMPIKhf^?w{yWv+_nKJ0yR?ATt@jP#x$AUPxM``BN=5cy$X0P9Fp(R+~aQ
z+Lh=qt0PV=dGtEOO@9`)E{?1%aD&3lur{-9>nlaq?11Ck`h#7k^n270#KH}J)1<Ff
z-iMRJ)lOc9!nNS07`M#*z2gIzSIz;PY?|0Mnxj4+ZQ4U4ql50l5^A$4jAXtYF4jvD
zW_sM_7x85vkOI|kS43O2TY;ZR`vK?|i41htF|btbn_=?oybP{YR&eVtY*3wZPxsUg
zE2yw<Kkt9g*a|*%ItX)LTo6)e^9ZAE8AO7w8eh8C&m5OQsS8s!f@c3&%P#zLT)G^(
zUc6cvyMoTF5hXZAhDGw0YCxVRZ#=BHPY?ogESD1d;@&Z%b~XN|Q+?=*gfV10-srlS
zc)3I*o(}Y}f`C-Guk8?2mcqQVBJup){M;)?-oncHt@W7tn!cUwBE<J$XVvJcm@RVK
z-(vqBfIOXE`lRV&21Zs)W9djX_PM~ii!Y-EYew5yq@~?bCamlszvUYn48u#?LO5?X
zG8nVfGNsRBg<!k8W_XrHF(u-8ALne^$8&P)y9ZS0s?yN_j(&&5`Q7TbECU^_)&31i
zBqGC-q&Jah6Uo#2A#q24LizraW)}7J4HK24)NGcRGJ)IcR>ZvVQ#aW(<C9BnJEn27
zMdvt3C}@Bj9IDMTGRfn7BikD&`c2+}FC~=ja)2L8kXXpCRbaA@zN_1_upOo6#N?>T
zC~zNi==OKNYIpR$nj75KLo&YV*NvBwa9tiEI83d`zlo=A=od61m|$2NCFJ%4J{J~t
zzC_<TiD$4r>TwwKH|Tvqa0lpz*uq0@DJ0%Mc?`yJa3tQx7i|_G?!97W5q&;soF*wZ
z+3%8W+5(szS*|-83I5PTu3*(qAIfX*Em1%n6`c_kZ#zM^8cd7!46DuUUvKI)?sYV>
zC`k)(oL_!Z*VrFb5(hol+akG$RO43G0IBvvsbAGGFk}vT)pjwzN6arcl^08qS^@|n
zqn63H<94aDe&sKeN_HAbWJwjNoc<RQyx0qM$8q^&kZPOL8Uf2K!KY?P*+TJ(kynFE
zY_nul`G3Nd&t_&RdzUwa$^sOg(pltSd7TL(?{H#w=GJm(8?BVz0Wu}xzD>7bBE=@8
z9fI@y(BEIAs<F|@!<7xj?;I^4B<tn2YgI}+lL^YFR$pXn*{I|OV%BE&euISf(vQm;
zB7*vkHqt?oN89TfMCsm{73~Llt}zh!fX?HD%IILkozA08K8bHdlEoX}J3Ws@8j}v5
zvX}etd+A(u8ZTWI-hun7wE79eY;z0l%zCsp<f~WeI5kS=y;%f&-!Q_RY&$Qey6+Fj
zu~6Ub_S(YA;ubUyTh3f=pl4dRkhWjIZaB_hNflB8Y7$UfI4s@X0<lHEKiLS7C~9jc
zPw_EW+&ueM5`IjgkmM=KgzS4w?LJWRC_WfE_@S-Npvce##yM8T4bVBxkCw=V+~R!o
zyf%#B9DDpSH%(>fdt6Sd2k&)BCX1!$slu6i?n|TEVqW{yC<Q9ofhc))RL9jcanN+f
z>e^y!8&Li+!r<u*pLwG2!}^fu(I!iDo)He36(`~F$ZgG#h1*h|_3!lA;wd*jm2j;m
z8#@{6^Gu^M-9Muea;y{VJ1;6Np3f7Wcu5|nSPPx0(_t*+caB&`N?$20IfIzdp3+^u
z>Jy^>^-c$#J;U$iZV2WEV1((MtOFp*g6-YnM%S|m&~6Y+d3|dO+h7^@e0ktu8!YCp
zqEC8&1<~?!m$P&vFY^Dg6E)8nFsH3JSZR6tS=DIoFUJk1AaoHsfcXxf#>3>T#zaCk
zv-se}eY=E0<{+WhJ;=3&s?XiOv?z4uIteyi%{k=VkEhgua}?Ipmh0^I$%>fA#T4Ov
zmFylj{nA8RWL+QBmj?MR{NHa6NvVIo`-V!ofO%%1r_0Wq4yuk)(ab@cGQ$t<%B#+f
z>cY!1U${R;d}+5(ht~PnH@h$EMSozvMa|%lW42Z^KD);m?H@XFAm(994Y;5+4E2qq
zT(od9Y$?u~8u`LoOJaL_&PJD_I#n(@=+!;5(2u<ji8c#0Ew$9lP6=}NjahkAU)?D^
z{T%0OD<cN+J+Zd(F6tH0{8H+Bl<$@%OvztDsxPxFt;$&Qi#9l!y62m|^Sb0~kxnua
z@B6n`4_$Q&2pC=ug}B#F<Y}4O=C`nvluLkMSueU|6rbiMD@JXuy;s+zuaJ<%5Ipql
z(Sq2Rz`}p~y<lb@N=ooEI|_odFKru>>VWZkvBW+l<VF)1gZg|@qU3J#h&;rIpMTio
z=*d~)7>|adHdxP^p}T18y~s05raia3_P($h(&XHPZW}NK3-zE!%h=J8AR^b}TLUvj
z&h|eMG|$qsZI<3oM``O9kLldnkYPIf{)wXaTp=GO$xH(2!MtB+$o4SAhj7zJw$y5t
zf)O9&Ee31aC0HJs9Qi+QI<l5m<zz-TG&KBA@t+AGtY+7$#bq&K^Qh}q7eDd_hJjn6
ztK@#AW$OG~_^%Rdw9n>P7x1xR=)sHQsGQqc52WbTLicq04c^Y|bJyK|yZL8avG&7G
z-PVmbW%{wnlzT^jU=db81#gkLM_drKxTE-K$53I(?DQMBFDD;NNr{9n{%bMe+zVZ#
zde>l_z5ejg;v4yfz5<djO#%(9zzo!r4~#!#lAznvyN9?F4Rzq6DYf<9gAt4Tw$&*=
z|9lyA@w8h;1Jf919SYStGZyYyAL#+*4GN2ql#d1(e%aSRg^;jixH|)-ax-9wJ>Y6@
zF&MWB_)J0spLs=2c(;LlLd?@72g+qjv9s?M(&)JGd>@~f$u@G@Kri}Ub}qr%v3t7?
zsNEbwN14*=lnq1hl<DYpzYMv8^c_ysiu4A_041qK_3#zLlch7GppiE%GAO}R#lWkw
zb7c!P_r+r^<AyzB$vEU`lp4t0?*<4eQhQ0K;Oer&S)0_dbJM=y-ubj4Yl??^`Wkc3
z6cj+}{(Z40_!zdDoLRhz(<sVHI2}+^6i`MU9-?kC9MZ^@LuW6Mzb08XWfTW^3yP!d
z?2T#b<xj1VeFf+ZFlTh>&SW>dw{hen6{iBM@@B?0>9>!K!wXb8zgd;%8b)gey3OhF
zc>S}aS@TPlO2vx=MW<`L<#j}Tjn^Dz!)P9fy2S1wiHBuBGJX_SrFl^efgtv+asZRZ
zg~8+~uHW|;@?_c!mLQ8|6&3Yo917HxkIPmq0FR^P%|b+#Hr@x|aJZyd<~CME0mL#A
zi8_oFFAQZ`$J%k|_k!F&NkKp5GPPspqhexi2Gmm&Py%D0a>#)Wg_9L&6486ExG9MB
z<Bz%(l!Cm8>o;@llHo2Y8i@@FQplp&i>QEtM+#c3V4%GJabqGTdU-ky(dW^{3K9ps
z;AIwpm?j1B%jt|XMbnVn*YR2XN+-bc4h-CwP5PyHHed#p^bWI`fIZhUe}q2y^LkF+
zC8J&Q*|wYoCII<$R!|f%WE8aI+~S9Gd-$15g7ItC8{faD5_@^kZhR{bhXlRb&tG6)
zUpN2FyklbNo*eWRzW%<xHT~4*q8yA&<c;%uDmtawGhtl%GjSQR2x;0w{rUFR?bd43
zzcfmXukH2=uv3`i@y5?JmUNQl14)7t*gVU3jrEuUchO0X1IEhn#R(DkAxAyEQG4&1
z=6pphZiuGmGW@B7-BvdYfF=l#H)8e~)w3<NnD^UOGTjM>UB|2<TdWyX@2jZaJ*=<l
zjwi_Sod9?y35WDYv}QrGL!A$M3u2f=?`WX;Kmog|&9v7dUH~5Kcm$(4_AhvKw4->K
zd%8w$7g&ZGSTGep{Nz#{8+UZEp_}ht9*7$5t=r!&eqb&|yRVzgtch+2;F~Yhg4qGE
zBz{N948cK_5vOCXQfVctBp%a+H!rr;ilq3V#K&mqXph=PSk6y5(0}go4s=mhQ7_EP
z%HEs^rf;W2EQ5$vo433r6w;i%-Mu>%Efg7#nJ@-O&8EtCz3RW1q?7Swixn9-_{?>H
z39GpnIo&rrOzV#1fcw(6-xUp6qJ$FceTi$5=kesc`;vZ^|HKEZV8|P{k8ve?gQ{Z_
zJC|mIHCF+g-7AtMbc<M`c&xcdQq%18Wy;Fm2x~956WPFR{<#80AA?pT-RU;7RkFlR
z%^n60I(TW;Dk3J5LAikDCerXAAa-oaiTH6=@N2q5!i6eY383$_<e(<=%qCJQM{u8v
zoloJ($=gBYnjM+wM1!hvvU?`xd+;sg*`O^Gc!bcvs>wlQZt2>BS+qD(PiAA<%t%HY
zA9!4YUmW7>++|N~^w<0BoS>g}MTt?6KdOkknm_aRJ8ly?d5LG;2B6AHtJm7KO+iKX
zD<0BxOj;zy&mnk9b9s^P_oX7na~A?3|C}ClGk3p$aAF0NjYTdtF`=P#&#2F54jQ{H
zP)PCRtQ!YAFE}tF^GI;o9QJ1~0PgIySXRbPDSy(|eE4g?ydCeAGjeUHe$}wL-!oe&
z&}zEK1P?r8*=xVNpo?EDKf?JFC2mBA3;RmT#DFxvHa__Km+hO6&CUJ~9iD%2vVMwM
zy2z9Q+z`H^3ndX^WQ+&))Pr4b=;hlPk%d(3nfw4iq4RB*@`$*BV4I-9VB;uIo9v&3
zSUDxZ?*4S445hq>HL+rL;(3hL!=4pAVA442+kZ0hs@4=M_C<XZ3?GU5*w_GIWzf_@
zf>FyS?9T$`RO)VJD;}AwXwp)fxo*O~&`X>DsJyh^+oZ&+q;Y+LREy1Yj!tvR8sCfe
z74x&b>(DMK&%ih)g8cWMPW0C<63!mfHL4B2qej(oc<=3M4=lf;<9Ka1k2GN2+F6X)
zfMuiclH%Jg5z!<M;@WR0HD>NhGBZ*QO}uLn@^h+t$%?U7nwoRnT!A8u`FVKsxAhKZ
z#P-iyLSZoyhi#q{MuWF~#LRph>X*o)oV~>FqrOp$+TW84{1;2tEyhk2cRt2<H#Wbj
z)RX#y#EjJ_eONKC_i@vGFoa9zYI)OmnVODYRwSel&SZ1O^7BdX5b3R>nCcMSdvwJ5
z`EtRy-r&hTg9yA>1;&bLKjUbkl><TUr?EU<aqnA*hgIPg1d}R<>Ex&T1Cw@7vNo!R
zdHg%qA5MOq5x~YK%ati6Z-Ui6ZC>WJ{n12bt`4TS>Nv9waHh*4l^gt=gnM}Pr0RZv
zs^VJTion!@Xqcz7NYL|(VZb00$AcGfAyEjW${X3~VV)&Vw^8AMJk8Q|#q765;uT#g
zePrc#i7_?H?VX>-Z`8^xXFle{?e6+vhTnr})q8LLDfpV7<QQ$LgMPfEP>f%4^ak})
zm<z`3j??XA9WvLH!80G7RaKw!?Bwa(a9WSQoN{5AGaJ5F*TzMZX1%}4UGigYa0_<_
z+b59Tg}dnMN57QT!<ihAHYmm6H1}V5gi|Xg>;suVZWhL`DN$)t4E`~SR)1Q+CZHxr
zO9YZ6-Bh|92ZvW&IqOql1K~GyhwO;#n%l+sqE6<A78$$sC{WHHi)p;I*XY57MNw&u
zY9aD?R|#B`RK+VjBo;K5nwsJAKwksxlRRno|5!Q?Z?@mRjaR8XOG&Dzs=Z<qt)OUW
zDT-RPXQ;hG%-Y{ptlG6}@4YuIwO4HsYSoCujFouuJLmZi@;N8x-1q%{U)Sr^e$!tc
zj=qy<Z0)%&%(LupRf21*%F1q<$wxq2R4EWKW=tYGkU`Y3kZqAj9({y5(^n5y*Jin^
z)A|aIosfjd@tJ2LaXDa@i5xj$cAFsgRzg=ZrpW`VXwmN}=o|0erMC{??+Fjt2GScM
zNH}hU9^H+raNb(C4R^75cfO22pJ{5Z*UOf%<eBMSJmqb1pCNs0Fjp^;t<fhK+uXj7
zUA`Fg8*mgplQo8ijFlTnO!(%3OT9<GtM=-)F9<%h@(nZStzUZ_1kmhFZEbCZ-r`)F
zwi>ZZ6%tmw#p2BC{B;;%Was>${7p@cwN%sl{Q`Fn9)LXVXaj<Z!(U7gSfw#g^s1k6
z_pMn*y74Dzf~31}oK-3SK}kzq9jL!w;00@%AsANC%{T<EslE|`-@Xn;;7=!R>%jK<
zDF$kaar2v&^$mX}URaTAvHij{cSz=-JJOh!uh9xlXgOa)&mP00z53_-omlo1mhoto
zWoL+g^U_Fp)ZOg$fk$@Ka`Iq-poJSra~__Eht%v@naK%Z?&>n~>w9bbX0I0@=&(}H
z>6Py0pyu(<R=y>ko6O9PY5XJSq1}Or)Y(|gcAsEjH|b)RrvF7b3F9H7<wZ|<HoAV}
zSsyS<oPco>pUJEd=@^F9ci%Mt6SUS~row1@!sCa(fd$%K+0uGfb8d(>tlWg2Bi|X*
zF4ylE+jhib=8{6UsqPiD(!{}nBOTR)v(yJZ&q|+jW^7a52)_;-U~wgn3vVXAZuAd%
zJy0~>tF}lWhH|SPN$#)SGJQjQ=Jm3Q2&KY(;Ck&9q44hcu+@S)p4~RT$rXKI%l^GN
zp;YNMX;4OYeaG`boT222zJ-MgvaEG<P06li=+5o@P&fUz=8^Zrg8-#)=?c)dT}G31
zx@?LwG9D?7Z~W?)Y%RbIe=Pc*RMJw%@!EiPR>(lh8;Y)uIuoyKvDVGh02ZQt8HuVN
zmO9S=sNbFTZ&!WGm|p#H**cg@p}F=s(%Vd*By#9u6d)R2{XY5-X=>^q0WYpLw=?09
zFldYVyNb7TbIFJMD^WkW`?8m>SnF3lndt0DJOwWg&*1W&DSOsT2qdm9Pt-;gbY@l=
z4O2B+4RDK}ig`3BBby;*$nw>W%*@xHRN6W(=L#e(oh1$wTc0!Dc!S@vDyb1UCx5c7
zSI2D1=i1Yub-|)}Klc6CeS@JvekuY!y}+<U2gH6#1n(cnhqH&UDGRrVd!1mwU>Vxj
z5la2^2M*V{7pkNIW#yZX>a;{!-WK@j0{vp~W?lM+X4ze3ta4H=OJ5>ekXy)c+0Fc~
zr9a9BvadSVeQqBGUEeIKPe-Zc&qpGM62O;#wpYr~jV%`k>03!!=BC~gFF{H}sof#S
zUG()SH;(RQhs0^A#lOoE@Pz3%x(TO&(;Mfz*s_vp*HU#@Sf(IktBAbp&oXx$iPn}d
zm)X@9yIjV`IXi0Kz^ZCwIgKcF>Ojt*NSsGN>Ic%Mj%uXHR_?x=%I{Y%_bktVMg1cr
z3cJtNTAu%~Ju_P>#O(<Ko7H?oW#WB?O14=#O40qc<Gm6{-Pj%e8mxa(Sw-H<x7jNI
z3+bM+>3@Mk+aKC;ss@1%-Up%#by5%VuSIOvU9V1vq-}n0^b~n%=ZotEbZ#qHJ9(0}
z!Z5jla#~k|YiBMGw`2G9@a51VA?$9+o<JIPWSu-2_dYkBtMx(P+5A<OC|mc59NlJB
z`S1a%ZZVu}$;Cch;%X1vMBDS}_RfTveixqse?)K<k^DC#x}Z%CeEs*s#woTXQZnH7
zhrdEb*8#rz1<rc*WIr<q2>AaMM_b|O#w^)B%DDRs0p*2X8ZiU<qVL_G`Jz}$JgN9i
zN38pjDw^TR%Qfa7QHldk2KnYF0)&l?Y5)<jvF&#$J&kt&ugUX3!I0T<`e;EP)Yl@J
zn)=DZ^i*!@)c7A7MvDheSl>+4aJ41v%@zrjG4O!i5~zFwq@?A{w@JIY8HnQe9U{+b
zJLy;6`7cFt>cZ&*@AvKs;Y6!e3(c7?gti+6BWvRC0UyskAL3HE^*oaA4_NGd#UEEH
zKs>c&<ms|-G-S!&{;4_UT<<c_Kvz28Uel%Zd#<GS(?@&{S#v6g8z}T1j=byzsal9J
z1Kz$%Yoe^R!txrjDql9lCDc*r=bJh&!96jew$P@q9+&21)0f4CpFwyiM7)fzp2@{B
za;qzWB|4olAKDPS5=+OIf(I7$B;<-T#G{#nZ2u5#+=TEBv?Uc^{)`VRjF%Fjpk<1^
zcN@cK4O$4RMn2FK3kDC7UnDX@7bcR#UuG}$rYKU3e3bD^I6r;S?*2u*D#`6_5`?K~
zRZz5Sl!B6{T`LMPs3UwetnuWr?nkEaEqU8$Oo`INUQ@Lc#k`Q6@WeAYTkY+1jwCk;
z-5I^eR>kKPBFHhlJl!?v3!3)BcMM-2$d2$8@ah<#+Nf9ebLxix8O1wxOO>*wrAJry
zp=QPVr&zUap2+Htou3XF{i=g*r|0K6W0rKrb_2-1>oz<|ZB&<0+J9ury#(9m2}?Sm
z)N*7C#H<DJZ>kM_8!2wA>1a_3<_qR;_x(U#YeQfOMu4=aPgxNv9Gy|n6pz$AL6nCJ
zQ&2c4XF%^d(zIb5oM)(ti=)LZAFZ41tv{EVDS{XfqLe27XW#N%%E9i%JKM@W21?jz
zvpj(7c?^-YwS)B>{;-rQLd%caDRDA_37SO8)Z7v0lK~o|mip(w7G$^zhIVpbI~+LB
zb<(A>kCYoTu(6&3Tg$6K2|=qAuIG=PJ2UWq#}1jyWf&d?2|7EDwW`7cnSS}`h{k0e
zzEvD@2>V$%2^n78;DW?<^0R9Jx=K&|odTW&j0G*VvEUubG)5858@kc9r!5^aLo%5u
zyaQJ_?D9(Ar_PDAB9y16Ix37VxV(~3;mj~?nJ=1#-prwT;uZJpZ(qyGuT}?z^p91r
zR4QGr>uN4d^`4Yn+YPtDByOgr^p_Dq`ZZ>ThA!&m*nS9N#IG*xS*0o`VeG%^i;#AH
zXM<|<Ugi<bBa3+}UDgRpZai@-N1?myudl}yewa{)R|fYYKAA5){rSBn5K{~<?L`7f
zqg%InpmoO*Ul$$Zq@HI%$&q6l-#N>xM<B(0|FJd1uuLp>U(?-|;xdA-bCw1MRNDQ0
zT060L>Em&w0ggCY9u;ozz-|x*FLU9y+GGZH=)OkZI$yiqZgr>RZ7)X0obMLLE#ud<
z`7^UPe*WV;ef+wnaW2$lT$@xgTp=ypT|N_W1~(tE{bTXc|9H~@769r239RwT`=1t|
z4mpxIzgU^VvZ9J?>U#w}d2A&p$wdtzolmRTaM>9n@F$p5lfOdT2C(;qF+Y^nj3Rvc
ztNRr4RsiwxxKnN5$>eL5M=JGqS_p~4qF9X3dL@TK5CI&DAiwRu{^z|UhwM7)^>-NY
zJ(3u(3BIk_CR8vzEF|?-4UiX@{vL-q)FEbReXB0#(&-?}^ikLI5>BS<$#81AM8<T%
z^yWBr1y{lvM;jH7rZgDH!B_$L*Vlgo#*g-GA{4s?Z=K$kA<n3aht%`-pLfw9I5_~W
zbwnFv`2UI@EsTzhHx3S1LYg}U08h6b6di+(X${lcb-~+q;kS%0+^a0Y$!wXH6u>T8
z+48Q&%d*_v>m2GT&t*k33l|#lzTRz+x-K-6WC^r+_iWI&|DkM4xLrQ7&xq1F{mHVO
zn>}Q6sF3a?FK=dskzZEa{)#wbiYW5=PNCQ*ZF?5YaCHB5f-S#WV-z;;NWkbzX2$tm
zv(|~Lr6oVk&f<QXu~)eCz5M)1=mwoD;Au=^y)9n^zms}GCe)sBW5sFh*T(J>u%D=p
ze4vL5TU<)U7|feG0du)n&tiLX+$qITB+6YK!S1to#}_ss_U`YO3Ca|Bs;&egOvzsj
zI`Rx)fA{8x9vD(oyzVFcsGot<_@qRt-u4j&8Ho1=inH<eS3tX~yu-lLTP&${BQF!A
zO25`7SuVK_{7N4sbF2hYGr9eed1GS<VIGeh*VvTJWT*`Ve;I6^Di}AEfK)i-JwI`X
zs+*)){a&1&fzl92ndIWiHwubfGRVo^bkEhvA+M|z2+e5Qi|H_LNum<fz6qb`iB(;(
z_<8WF%kkyIs%G!AiFKcDUVh_r<<*HDbSdHeJ^rs<=<Hnv5{DXTFYQ_D7Qq@gB%@K+
zzj$Y9jD~a&Ah9GemZXjP&C#_CG~!}*ea8Jh7-QOM9C(7q2;nM_u-EEk8cU$Is-<Us
z5R#dPgJv<Qd*yYLk21bZ6->zc>tvg43x0G=BziHP$g<kYx;4qlPs&z7BV|MkmwhDi
zaHQnn$ES=RExIHQb35l+V)vz^p>Kc9aHq2lKRSK0WSDdHPxThPM<}pO%m%j>+kO)}
zZoe&~yNwxU(}@bM@>6i|?GBpe-W>KR)FV2Rn=i?vUG7QKUg=yRiH3hPeIRerwW6z~
z;8^4DmX|?H`ypGPyeno8vv_U3WpXyQ!$-&djOIg8tII*p(n%jttKaz)dR&?X?rDtp
zcORdyZhm0_v18}RO0lfm8)`p3S4vVg00m&FW3i81XD0BPW#{eqb=}@$$D}S7)MZ;?
z-<bJ*mHEhO0;izMZdaeA<_ycx6otG{+3-FS?@LeH1Za-d#=`&<(L8w#X(9A?M(+%~
zs!Owr5ZGQ;(#O{OKHq;!{*8S@cGR8Z{@0t23&;02QtEj6pHfLbc}Lk>MorX;R1SP~
zK{X>HFMlr(=%Pa9j#3-4b+>s%A6ha_77cn#@~BYljMMpbJt&ZAgPU(N&=%GV-D8mP
zK-!|p5#bKk4&krw_St_C1(={PLwYr^*~QqhSM9njh_reJQ1J}!s`+7e*He?s##H9p
z0A62fk|bpR_*p!U(g3}^<o)`dn%=-0@jelv&=6~pzMh)sBm(sR^+#!|r`MXCoNN?H
zY2*X4&O9cT7&tJp^gmeOU}gtA_t7?q$uV!Wd}@f0^(UWPL}B>8h@=%I*JP;JFP+@a
zy?yhK4%nOeJM)K6CdBU*aT|ViTyv3cF_PhsA+*>YP~Zd27?dPZEDuMy+G!oZJ<n~^
z;=Z{<Ct}~fKKk$Hhd!3jr^KI7YU=4q&w+Hod(5l(OUVk93~G0k)Un|Q*<s4|$}(7S
zYs3T<)p&uNeQ!;>{^d?@dtZ);TclM9$<Z1cQ6Tw<_^1QSb=HKnT8e)LNR*_zU2xsF
z@!`@nysCwd`#mhMDEI!G25Ha_3vH2p7YcR@jeqK%e0&tEi4uwFr>!g=I!Xsgn0)9_
zQOo9h6CO#Z4|jub5>(z7ahWP!n<LfQT~9-uyH{5wW2VNI9KIgfEVVmLKbFz`D<Q94
zKxg^A!%WHhAI0kz)>K#AV1SnJDPW8tl15HN&~Zj_&iSh*={-m|e_Uw9%Q%<hX(2~$
zt<uHI!lt(K$NOi96%KWO^4mZfjtnyjImYaG?LMBU8J*LlkV5JQuXq~9V7lX)I_DWb
z^3D#%-Yz3o-=tF;NimaeJ-AqTcdZrssj_!IH%>`AD_dM<#bG0{-Z{XE^VA3I)gUNc
z1c5AF9H{M?4Y!HuVJrZxp8iiRi*-0`ol4GWzH)J<DLWH$4P>x#DX(PALrSr^Z~>{H
z_YcLbG$<nsz%#FnO-w%33u1NXHZ+&qATM5L7R8Asy2^L+uSakdtOe<2_;+~vPhZ-O
z?<k$jn1Lr(tTNu8$6R(;3wy6q`1A*Z9XW8Pu1izvE=$io(h;kZia7q>pWj(WpO4h<
zH-X?pD~m=J=f@QnO%cgwQ2zk2nm3)&Gud!N3~TBxnJfR!?CF>x@oI=Fie0NU>YCL#
z*e8EYLAH4%29ikitf)_YmBM__>Di7cqGk7HE)42hv)qO4L^qESBB_-9S^?<sap?8G
z{J7A#z|MM7-GGzT7qSRP0_2jAk-X5t7Z949f!#Pf$Bo1F=FH$i#6vBxcSHb{tR{?V
zFbLDB;23-k37#`^av`jb5a$O&GyjWU?wn<bMM#hDA!c>_jN>$Zd~6Pz$wC&tqCCGB
z88(ev3Af&k)oQN=#y3qvJGzX9$&U_pPY6tfXz^o^zSow2j&;9D(w~ufdlzWicy*m}
zH<&04+qdQreH{Bl3#7!x23ZYHO{YX_UCPT!7}ULT_1O^%A13%wR2}HJzW%1YsI++4
zKid%>vVoEen52~fNpXwSt#tcTgAs7sqgM_a4ts~Bjp-O*eQ)Y+8=uOUgyfyL2%kYJ
zu^tC1vj7)pmlXa0#nCXPu`b!XbWQ*Tr7DJ6mcfNWSVciK(_|n;SIn&qv$vjM2bc_;
z@gK&OT0ALlyYqaGl=qZB-~re7bH1JYIQdGX5?I_uOdexzZ4o}Wb;M{W;X6jCWDxsY
z^o^#z`+BVDhxf|wTL3Dsk{x}pCFF7d>Vhi)*?Ag)0m_WxDs+bmZ{+$e*_sjsCS*cp
zJ{TJrm7UxIs-*7bSP$f{QM|~_!q7MgZV?w0uMPcGGTr-|P%Mg}H7_@%_5`Co>CUff
zk1<%@zBufvYBTsBg!4M&UQlxC!U+p&t0xz#&A29NIjG+_&?DU-h?V6e%5utX6oJG2
zbw>u@;<Q_OYZC@3xVFX~vYI*133;#@?g}ntl<I9n>c4e+DI3!+&&)RQVk$;_C=n6{
zFQPut-2CEjVy|gGFtb)!!Zwh@%bs2%_#=9UHt<hSi!y0sy^=e0D>!wx(MZ;G*n-?L
z{&Up#FCq`dCS(%GvIe<-K7|$7yDoiE`OIJ~V*+0AIW(k8sJgYeKk<8WF1{`G=9+8}
zHI&HmVBqxFmQ>XF-;fVZHL5L1zQC)nSpLn>M*RuRnL%@4WXip{tAd!<(jFla*>|$a
zGfV3&Op&H?u3rJk2GY0#R7p{x6NtDwX~W<4x3_oj2hAZdc5dW*GgDY0YsHF&T9>$2
ztj?^MfNB|U{;djkY&MJoPEfcaqy@2$Y+nw^`BD!K-}$CBKPkc*(sm>$QyjA-@k#e~
zb=gg?`~_}g=g#~@|0I?HyjOX1k4~ML_OMRR5bKn-{cDtUUcz(sN)k}IHtO}Sh9(u<
z{!eYV<r_9eaES!{RI+6Jvu5at4f01`dViKa!IT`_=_*kSgDA%h7!s4VGQ2B8R%r)U
zO=BLv?V^oa0U*lh(5IsPSp}Qr71{S<fr+2<s>B2IjO5T56-00`zDzXrnr#&gH(F(%
zv(A@lXir5M&$EH}b_F|`0Wk1r##DxPHFnIgGc#KhNi1kyj%ziNNm=g7=k?We<a$)@
zBM^t)LwlDYd9y!v#GQYv)>P`Rpz!rN`;>St2>Iu#S9I^Y_}<3yH4gg#=hp2KcStvJ
z8#28t6%;r<Unv(nFWGa-aeQ_CG!y&(lq>CrERHu}QK(gabh5?U_{SF(0>k!9+h`~W
z8`ZW^8odfLwIEM6y_MxdlQhS9<JCb)tK6n1CLEEM&C76|9dRoCxq2&*vSwY)qLFX$
znCeU35`&UhlW>DkX}^sXB2v*EUa?Guv_E#W<(^NTR9HWbp!%LobHd2;rbSuhPJ2tT
z0qLWCi?nC5mrU@y@?TQ6t@A$K5?^DKSYFHgbtvM%fl%WiNiP%psmXpZ&7r0`h5kXS
z#QesDh32Z%EfoR6;rpB*D(4OEP|vK_ULuj=L}#{QflQ#Q>M<r8_`9_FRF*8D#>Png
zn8>l!u(DpnvPF#mkA``oy8K4?r3}RZ#Q`;W_<djx@*=rMca(;wC1>!ds1+SqpP2Wy
z#|o{Dd)Bq5O8@<aX?+&brnN8bUo7eviMHIZIkoGKZ!q^dK4p6m*SYf(YY!0L@B_v;
zPeSRxwhmS>&)qAE@CR1kcR8*UbQJCT3XN0Wf{_lzC{u`}8xt#kqz-iJ&)ZxQ1GSNm
zhq@`Kuh}tLkypVHMSx+G@qlz_HVvmLqE+CL!MY{n6Td?v4Of}eX<y(8asPgnIL}UL
zjZnC^Zy5=Shp)UC?wBu?#!A9%Q^>@go`KQucQh;L48K}utH025(<LT<@sIRZ+;j2g
ztcrd?;*xev!U)q(Bjr?twqlcprffx)Wd))99R@Wgj3~JO=QV45lU~ZxGGr!M^4o08
z^km`VJ6|4)lNN3Z4vq*Pa?|vT45#R@9i@NdZU1$pJdr)^w+lZyD*Yi_y_!2K8r}S|
zc=ije&c#LupRns{4_8dz@up7XNp+t{gueEFqK$>g{?1n0ivOmS?6<rQqYiCo>~31z
zqut8|ZG7Wm;_?)IOR9c;nA|srLj9`O8oB;DH8=odS8aBQSnYa?q@^n}qkIo3=^8ze
zO_Ey=<0%4s19X`+_xE>VEL00ib~K<H+fI9#{C*~C-IF&iPRbLrV=w<>KbPMqy^;XQ
ze`4YFNAcMnKW&#&x!h|50(gQixH~9*HTx~alqPv3oSR9lQ@R!*rS~cXZSAuB7nL=U
z6ag+VVs)j(5#uM}{{l!iKBTomVL>bBSSiPtO<QkXTszFzP}y<x^+gjeDlsRh+nGwt
zH%Q0YlvI7+a?X6%2Gp9~we)P<Q$t36m4TpULo>^N$@3Hn*wmmD1%){F87a7uwTqU~
zKp^8yUR@KFAEWPXJ|qrUJv>y|sBCIUI<rG%7=^7mZs-b@s$ZxE&d<?~^mju3*OxH6
z?xpKO;uUW9Z|CqA-M273A3u);cDbtx941P^{&pLQ&j`3}uqZ(8tW(g@NiEZL-Qv#C
zcvQFR|L9YH<S!2#23k($Qv$(kTDyyjrI`4rkVO}Qu+ri7QsI^$%n3e*JP3_5xNr73
zbq>+Ke|9<M8ra^!B@y;r14IpE3ZB!oM8gHOgaTB@`c1A}X_|<J)8B!;`DV9Lwd@J+
zJ-(JB;fu`iDCpMA-@gGzMA_+y<H|o4Hl4Sgq%LeZYo1&Yk)2!}9Mqv?eH;o%B$)d|
z$D^=PqP@c{x@jmB<rkm1Lqzv74==FUyy^1tR}<7#*n1f>?{H5UM#_h3(QD}23Sf*(
ztE|*O&`|pU9FD(2W3^t{k&}S;DNkL}tXuX@xlTE)d*7>k5rwZs_x#>@ec{J$(otU2
zUk>tH%fp=P{egTnj`Xi*KoRM-s4$BKQRBSz-Q{DFFO`aVkr=7+-}H)QIJmT7#J5Js
zM3au0D<i-`G+S#r$a_!L`Vm)3*X;BD(*6sewRNTTZ1!pGB&YM=yr}VBvJCv`*g_~7
znkJ8H$}!uhZ%p^oUkV`DZraW>$j7xBb-<Q%zsa6(e6!y?EtCuz+p*0?0%4@$Nqs{*
z6H;nYPgB<zj{3JhApf{c^%va30<zS$)!94#9#4BX3ovF+Jc;LFe#gR`y1cjdF=BJs
zw>G_%AG})*r=VLol?7gtj_P~p94#f}htESMO^l8C0e$&NA92?7)2jRAhHD|cvkwb$
zzPXob#Pky3jtxlDWu+X*2U#AyDU|vRBzL=?r`SH5Hvzv3w^>^cuy<7Zy6>9`wr%i{
zg;90&&bu&iZk&xJQI@r4U(1g<t`GJ;82_#C;3A`cPEkg%V?A+j;?toPmBDwhJx80o
zciX=zDT9B<llxR^yZMNFg%B+t8)TU1B1e8FYUQ}<zadaUVSy#2JLcR-hR&;1g+J@;
z<6YTt>>m`$6wVw^J4f`d=&)?hUgAMMu&)=oDLxI|ED7IE3JK3RY&ju@bvsI9+kSPT
zji0Jt$F{%ino8B~T8DINt5BqR^HLxL<utZ)(Mu?!qgNsu4i(}e1La;1fc}zh0Oc0D
zj-eQN#EFAR0h7^|zQXCQbN*_PwmIqZJp}$7(vQXa`>wi%)dqKVhE2U<o$Sq@ggL7?
z{tiLVBh{q)dR)5}R~$D=u#PblD8=Xoqns(IK}0+a+td6M7w(+n$vFGO4lho7ninOS
zsHqeY*?n_YfA6BJCwi)puCqDA8NtbCknH{a*}(2Ct-T44d#AqIghSrvt@EBpM+G{N
zjg4@GpFb7-xL$7|`*22}x=jCmOaAlTDDe9t{LV;;*fBS^)@;VH<EQ|C0d^e)sm|Z&
zncuGT6LuzN6S#m>Od5DYv*lV+#SSkBnZmt7&xXila@OnuxE)ug@vyGCX#2!qUXb~|
zLD%W^G$z2y)b^oia#dG=^Ymm?aLNAsbKHvt<fkCv$^RV@&(s(_XPi;21dJ|PPH2@u
z*Md{&JqVMg=ScKxYG3Fbl#7{5J6bm-Ym<w_qGDy@`1`#x?+8UM`{yGgR(1ws`{cf_
zOX}VjZ2}f}eC^&XY5SLlKW#S+Y>xJO@{X*i3$6?qWRT)=v+-_PgYv{iF@+G#u#j@U
z^Ne75>hwA{fAsNh0#qFk8X|UoHC}P=YlyCZX3vv{MT0;KxC?E&k3M6sH>L<kdq*yA
z?bG^PwI5x6eB1Umv)V4_CNZ%IGB3vHL5jBT0DISzNQbx|p9JbhE~k*rM<(L<4uf}!
zdqnDyUA5#uB`78`o$?0sjN~abMPxNA`YP>TcO3xWsCtmaKB6!tEb})AF3Y!2SfaT_
zpG)perKCfWou;hD_FR22QC=|Gy%sL)fE)qw98w84w|5$tna}f6q^<wqK6?+4?0Y*v
zJjwr_9KA;O;CcZR%iz?$o^J8g=vO$aG30aa<^2P$12_^Wn8{3;_?cD6`DevE1AyN%
zNmwDkx<gVcMxmNyOWf{DvW8m=&?aqcWy<ViU>z-*7*+Q+eYolpBC`26>-U9?Jd4OC
zIy^oLtG9B^5F&*8Ts&L6GDE<=cq(~OIA3Z^&CysYIV~Nb-$ivP`+wh2{2b0uG@1T8
zAPNob%|Ma_=b7@o^@UPID2j<acqrWrn%~aUsB4%!tU@@`t+pfQII5cYhl@6WIeTBl
zNw@zuiF%Cl19pjXj9w%EeCYHq4$Ikho>Di!tBY3yBZMZ#IsXDDeMThYPwOL`suYzm
zPv$JzGt_fMn2QDLxg{ACRf;hRQnc?c^0Y75AqtCygoJqMZpH1)34iq8R#q6`y;)%u
zNTYL<)t6W$wVptAG;XLC{G>F9%xp}|?>zteE7G=YOjO8LBiJ+`*_i<@rRyo^Y=?LG
zw{OU}L?8EI|L0GqVdnd;PtK_3ODa=s-Tt9=2e`A<ntjEeePx$knV&l$6e<!abu}1#
z_qTs1-m1)9M6;95e-DGz@E|HA8VCA3Bj_^7<v8}s?}44FHJ8V?XP{m+y0)KpoXymS
zj;{XJ5S&<a(}R%L-BaEH%Ola9=QA9^-E0RtSv~C*$&K!gCvUI+6)V_Z`-3Sw6gISn
z@fB>+OGe8`*iyDU?1ST!PuKM_YPH|hmd1Ff<@llacH;V=?|<GK%VW#@(r)q!(5~5&
zPh?w$z;5v->Tg$=?JdP+WRO_^s5+Ta_j!MG7-x0$%)rgr{s>f{<8Q_YCPW<=b=a#9
z+FJMK0^cDvCPeJ9uxs4yx%vLpdR)*S!i|8Vk>hrAgOUSlb@G~Uqrl-y{x@i-9xtyy
zxx`?ePBbrTdE9E<9*7>_@H)8quAIgce|JmRR&ti1U?*p3>g-(3mRq)O_hx6*LYHpR
zxY*$aC+06TF1s2!rc+W<Sd|wEaoq1OusYMI1PG&ir+*^o&@f|6FyQJ3$a7=_7&q_C
z>#QJipnV@xAmUjx=U|;wTP%Lk<-Ov%?{$ZHWy_L~+oc~Dy+1wa7bNqd_{(~zbPSEv
zUapD4f<oCbpc=@4Vf9uqxp+Symp7H^{0cv5U@KtW(iri&!J_u+=;A1GQ=#k}WcYsd
zq|RzZsJa|p-IYhxVOh}n9HJt;t1(6vKyPw{`xQO*L+~oBy1M?Zgz6BDF_7*wG$6jp
z=H`iNc(*=nf2tJderpefh{SkSWcP`@A8`@W+IEga6W|P^Gtip`V(+@`_1`l6z>1#E
z*<lr(=8!zHJDWW{&1sV@98GLM0wmHzq70PO1w39B9SyzRjQbO@?d;@~GRWqu1qvsX
zZ$4b1jvJVqwXN~uvn|YV2s@`>HEa+GtP9Z$2K@Z(<rW6X(Fms2ku%{8BWJJcFSz*K
zljdYSE|K=-{SV}px3rwmjmSh_s9YQ9BXhiz^G{XrH8FYC;!=hj`@<xNw*CHgGz&`7
zT0}crREWIuh1u%o@$<IP`XPAl5~97AlRb|*Yd!14q3S=WP@X;o)<po((Hv{uMxU%W
zJEd&P7@=!0?Q@}>t#j$}?l%#9urWjkIp(mhgS1aLD{+f4nW72XwgFbGDft>U+|8`C
zv{0noY&DU5;w-HYLlhvGE&tidWR^6N1T#;oBI4~xw-mBE2kY4z@sqZzvZeM_eF$;n
z<Jyb$#MEtzgM~;~fX{$LWDj3P+KrNO)d>@C((jT~Wk}58;VW;q2j}Kojv?|S$XWN~
zPca<0ahwlFEAnJmVsjrzy4vIE?ZB88s*1d`EZ5O#;HtM+^WTjW;FsT}Exk@$aDQ?4
z)X8vSU3HbfPS&4{Y28(}rEUMhn{(-_kGBb(J%d|5u4JZfq}TZKsW*Vqr!?$y0TdO^
zF3-^LU5m9)Rny{`t%>!AOLS7hb3VbP4NHltt~0U<9lDDF@L_-m>QOcpjbkfSK)aR6
zGif@S@3A6SI{O2{ga5TXe--&FF+^ATF4Vg9+1r1%j^tWioEL)KSl^#^*(HN7N)HLC
z&_7&Ov12sfA#ycl>DIHDPmVaGr4B*jaZg81Iq-BMn^3S0{-6~>b)dWuyyI>f#G4Vs
zn);o}MJ>NhApeJodFKa}I=|=hCG3mQ3d3l>B>4{to)73!o3DLxU}J`1mTM5iAq<!i
z%F+5sD5<RDs%RqUd?N1v|7vgR{JPMi3{5<qtPTA$aeP2YV+~-;BV85$9D=)Sr4c=S
zf6Av9kUXFG-OpJegA#8r(G5nXB5$vYX*0|ZuXZwnm}${?_QOCznXhRdG4eI&|BW6=
z+xn?XIM<;1%F62GzN@w0MD21b|FN|jq3R|VY#DlmsH|YxTb(-V#ywvh*u^M#{L=1>
zGDec<PWh~7VqecHK{!YL+O6y24Z~xPkLws;glf^9h|z!3A065q(q?1_MT95RVi-zP
z+g+W`M{&lMlUc1j7u+8<_8W%2mpC&A!m08~mcXByr9ios9%@EO0n`juZK>L7Z}R6;
zV_XVa;)h*{G+D>R?jmGkAKQx=kVV{n_?2nql*mBO6Ar&K(i25^+B~71m}kNDzpwz0
z$Rx9lJL%A=B)UbrH)ddjLX#zuKom@7C($R3%YP1eId>a?j7B0jegkvYqDPVHqsAyZ
zwe(@qR6!L1H1(O!MD_T>$!0BATOVtY*1fof#rD9#G(@Y98vU{SYh{StmB<icj1ayH
zMjd%4ubEh^DpOdTn#BNK_9qjm-&F<K{C1CC(@funNW*oq5qz9;nVgU5T>y{8668oy
zzmb2W&IM+`BidZAvX9$CO8sBFhnL3DPx`cLtJ%oE>H&t6hl1>yH}0*+|EQPf-}a$c
zanQvl>!`7!rUl*&$xUn(7gWHQ71QpNrjrnEEEG--1SPO?yJE66mUAOJps`(k;*3qh
z@}S%&IpH`$I-)r|??r0G(pj7QSv^T<!&e=vCk+Ez2X{O@3I7lM6c5*q?dioKSnmO$
zyWulkoimvF_VU>qb6Wu%;^T`SY5z2FDU-Q30(Bdr`;I|rSm=Ke<cootZ(E=~jZE!@
za4h)p>d^1>NRyva-7*V{k+Z*x&n?$FQSfuBEUuuf4pA}kDaV7JGO2_8i|l-D<5K(T
zSQC|uV}0uMvx}X^<_#&)N#5>DvHkx1JDUIIhlYawO)7qb5(*4ejRcn_!p;sI5v19a
zIMW|Mi2`Q~j22K2q*fVrcqTrPbmkW=$s1#`?0bifvJ;<PV`$*tpZw*Lm40@7ENSsl
zv~k~isU9hh+Z+rTdVAc>UpQ56z;v0n<5Q%h%Chw5RH?YK8Q|rsQW|Y}P-2?R)_9VS
zJ3scSW;=9hP|d-$pbnkMnxy1y*m_$C5KMpLiKV!pi+N`s>}#NfZ!{vnbBO??T#gGq
z#X&t9GN7h&7tbg5E)!KzWXS92rZX$9F%(2z?1{aB{y?a3b!VeLy1BJ5$_T6PA>6RM
zLFI7bb@8*eTw}P!{GbuFlhcq@S?nAfmtfO8*Pfo<dMW|7B)uMpF+e@B?p+>_6oACU
zXo)t<GWztpiWF&e7~iSlMrtny;2Zi1$Hql$oQD9((hs{1JDgn!d4S+^cY+4h=;{b}
z+;boLe%Wp(3+LJFt7_Ta>2rB(w<bP6H_s71pFt-=>ST-l=co{*9>i|EzmISn*{&rp
z9e5r+C1#J48mJP_1ynaA>Nvu@8^^Wk#gz~*j0JK!>Y+Z^r6c#bBWXmuq!@#lO~nbl
zuYUkY-qyiAFK{C)%Go`rYCPcB45q1*R$L)jD_Gqpv0P#`6K8rZm8%V1M*u6@jxKH2
z<c%QD#00dz>rT43qL?O3EyW>_yUQEri8%q&W*rNMH`)Tz?3|Y=DS#gzmui=@<^8o2
z$S?v_(e)38$AO<MrjxOq!m_g0D2FAH0_JQugpQa&i`z5H#9}7R?K9esY+*%8Rw5bi
z9S4Jqf!>v7)Y1fLPybcz=>kvC{$|dj<4;to?!#q?dq{}CqNRj4l=m?eO<WDsaDmsy
zW8IP<dip>*HR#aFfWdZPwx?hS8(qVhzt6IC*HO^Gp$cInGhvu*hbI}|SwB0;7W=8d
z&k>XFRWTMB^4xmzX2mInzh$ndeDc*LV7KQ#1192HH{TOa1m7n-Wz2ubFm+$K2I?_A
zWD_4fS^9|s-D^nzxx#P5fu5%%*|57&fNnq<0I{`Sr$L8~3d;4C0<39g)$$}gR?Zp=
z(aHe}c4?Wf<?=nm1UcP}^Od-<UZej=hb)2=OB5H6wA?$)NROtT82H_$cPie|(6~Wp
zd<3Y-(^d=(!b_-2`j;$CPD7?98k^Apbt6Au7UXl|n8d+_7rHd~^bEKrwQx{7W^PS`
z=Oi2ArrG(=AJ{(4;%xJt+cTBSbo*fa$Vm1Dn7?((P6aHHCM+Z*nWqdCi1=k&5GD9v
zTG*F_)IPFHVsw}HGNIzJ*V9MSgN190sI_E?eZPgtXmLNW$KOyxU%|`~y}|**fpt$a
z`3Smu(rQY!@8_k;89bi58E1bYi9PNEKpyDVDPXF@qPO|ceRxDt$YJ!St`X4HX*Y|&
zI>Y42RCR!Rt7~>xt(LfErS@e7Sl$%fVUCU&-Yvb2FP1MiF;CiOv;F2+k)Z4^pHsDZ
zxw3&*ZuKR4?sq$-DIRdR=`Z~RlU>WR=^=_7X?Z<&3a&ZtSgmyY;IW6ol}#faA&BE9
zUwA%tuJ*qdHN0l-s*SifemW6v>pP!{$Dy79I<90+Gi?7Wf1VpC)$El$?zeCBS_<e4
zSM<3eO~WsfMkMK<d>JL5Px^}P-_esO(`qf1Ldt1fiUbz&@&k(4K^3VwL1d=KaaB_j
z9f=Q5?%{^2ugos@T$2cL+<ztC0z`pr9;fap9}KKB)e*_M?)an($qv~7>5{YOPYJfv
zvu8OlT4|LHI@QUN9};c~Ajvs|@51GVK<8?Rva+#Xpz<P4p)_^Bj4~`InYeo2VGAoO
zd?eHC$mo1!ugfw~RX;iKV=B)WvG#f=3MsYfK-?3vw_6$vbo2c@79VtuO)XqUn0w)V
zoQ`yOW~~;9hAqz%#iE-~i2WVzZ^-{&TD+~Rho0fcP@S%UhSaSf2ub-1CZDIH^RaND
zmnkzxLg6IjaN_NL59{IOi0z*B-WJ=}R(GBl(wnp&%VGPNQuBuT_So)Dd;n~d=RtGt
zzXbKZexuWZaDl%{b(q;tQs&qCU4fY6F9x(HIdk%4>Pp!bY9c<ehTNH9i@Y9|VI(D$
zkOd$?;Y{Sv?~8^1y?rk4&c0}=5V!wV@?9$EzC$eG@_pt4z}ysdSz1gQ1_~7hT1cWk
zBgqf@Xr2$2Qs28IHg2yoPo5$Vi)`JcF7@y?aeudn5>I*?ZmS|XOuoaj3xQR=`j6og
z7tNO%&-ZR+^h_|?Ig2-(;a__Hy<f8ryUM)&>GrzixE&mH*Su8X(((yKg@Y{VpR~D_
zNElIAH2E`<GAkY3IPoj&sYH#!^M-!^Vz!-=7r1wn0X?);3zbVdv^V6$0F!<+%N1ml
z{1N6=8GVlEI^{<7q<)f*oD=23d~6qdJ!TUt!WpVv481H|>5}P=#ZdD;`ERE<o{*M?
z@eJ@hRL*c0Dg1!Rn1A5<`W^T@RnbUUEDd0|<9A7$(AazV0AP1A3<2?bBb==Hi?5hX
z3g~C#D=p#W(B9Sn%EU$GT`$fexaqG^=z<rqlftIHfZ2<Qve>n9Fo7$-R4s0<`85S*
z{vSymJFx5En2xEu8xT0)Fx*#4j+P#5n~Ebc@h1>N9v1Guj|z?>@Q?3*T)Wz%upMIM
zfr>Qb&otwyHCW}!Zrehlzg<^APenz>8`~WP3oVqpC-3Y79+o}o!`E1b-=BcxJP4BM
zo^{mFqj+eoadv&4<s!7derVS8WjCE?g)^IvL;?Lb`h|inE4hxm=+Z_%0wc|vH>;bc
zFtYw7MTyrHLW2<Slb4WNF51!eoG{Y=myD4!Vx1=GAfqeqjkiWY*#UHCp<j^P&`y`Z
zYgUwns|2cpF^D$i7neyxe`nBKVht3kBCC6Lrhe07gqrr*j(w)+3)%-43dfbQO8a`9
zj!9B$q8jPlQ&GnUHO^rF6*DKW14r=924Tm)0S#y{XH5(bF8Zlz?&ITgH#%s5mac>F
zt&NdRT??c2C>^KjN~G|;4z~%`iJW9H2cc6V8A-@1Bh?DtX!Vix)|9$)+tKf0)UO8m
z)#r*BBC8Jd$f(RsGsCMSK41<y?Uz~O1pYjZYa$S~$Dp@iu8r;3RXW$LOkuU$jtm2H
z+4=ng@_5}Q@_ScW+SMkk1Q4gg)#V1qjm9<dw)wUbH^qU&5}FlZ9~q;$)!WU$*DbhP
z{0zEL0l#s(DkWu>R2vmVu2p{75Ns(n->0)CiPSeRSnj-DH5anxRbIJyL|0fSPEByV
zH8(ak{v}RL(cd>IA7`|;nll>4_I!WhZwqS`COwCncmj3zZIQkwFE+ByBrhrnN+z?0
zN)677zHVpN)YRG5NvmmAN<OnUN}5NfdwIKub1;kgvmog;(4Tiniq<Jd0=}j@Gm(Rn
z2F^d{S!GX(1czZG7@2f6n4mXl(zBJ7Z(9C*R}ftelYsLBm_I&%>t@#%*f#2Hi38vP
z_QT48Et2xgcv+bqKz&040XA`*@$Ff)GTo`q#--hQEq&~^<ZJ%^d&)Wnb5XYx>}kA7
z;cl){zC)7(HIhzu)n}G?IfB18**fWxro~H-rVo{-|3%W##Zs#Hwpg2yTco^w&-q@$
z?7_fPy8~Ac1&5@>a_8!lh>8*EBGb1fFZN1%^MPq2&85-bgUWGOO674#j{Mgfw~rkh
z)(;AKMb$tmbQqI=9QhkR3*1tg!+3$+J<B&qnCNk*7cZby3nRyKy|=S11%czl5<zeW
z8|H>+veYie95>sJ(Y*i`8J2}abNHB_+rQqhsF1T)%S)?eLEqk<Qg9f>{O(SlBpImu
zHXGXSMz+Ed3Q-a-%Xvg#V!!v*p*-ji{gdxQ>EfUc^WU5nK}A+ucrv}%5aO&$*$gmt
zB%$Mhq^~E&P@(B-Z2#n&K&W@L%K~7%fK`x)F&$}s70lliXSUj%D9@gxTKKg3`L6|o
zylpOyq-KM)w7jwxW5h2(!G9cF`bS=niJ;R4$JHfN$jIPkJI*g)ldh}5@OY`(voWs(
zG1FdQ@%$kEEcdSgFprh|#?J{F<-_+^wPl{KUys3@6gG6T3{rf61hrT5PQ~{9i}Z~~
zcM;g>)Xj99qCt-#C0)3hP@9;n0ZOmea;-L__Q03-;M3f4PhS7e;0QwcqzpFcprVxz
zZZ-AX6fnAOGkL#n6Bt_srZRLCE)%pijZD*y%)WIsx1YtBh8<UadJ+{v^xe2_dg3YQ
z&#IqmlKic4`YHiG8vV_K)H>C>jXG>MsIhcOp&>ucA1HO)=_S?*m^^E@&+<7gzQrw)
zGxMv$x&&9u1DVs1O|Z}1;#Uf{-T=SmRE9;-Ej@<gPu3O%SzUA<g>{q$r-j4shVviB
z;me1g^7gx)ng?$4UKIuxR#Sj6Zeh}8&q44Jy<454#)7MZ>68_JmPqC66XbH!wd<*=
zKz7NhXM6D8%d=jrU(x9AsjHn3ovZ#~Y!fm&LxHVpcAOz|HL!SoK6pM$P=*$}GMjD2
z@u>6|R|@)-%<v3WjxHs}eK4PzG8q<6oho(-DEq$ma5$C8;B43<x%jrie75L@_UjzP
zti&U%{Xt>LsU=R<j}z=V3xs7R91w;RA~9Esg75aaqxQ>rVZrWrpAr0z{rP8s+U7xv
z`{+cxJo?s2HfVNUG|>(8|A|PR-MNLzbIdaG6P^6c<G!0LA=WI4eWy=ppWmH!ofM(J
z*Bjr@hilV4DCtNGjQ;Vuf!f%pqa*swLPF|U1+&BY;;=6FD6?e;wOKs_JJ)|zuYUXd
z6D?w|hx1!)>X>CY#$(BMpm)R{8c{pujfVaQ({!HwI}%VF4|&gc9M7ro4o0ma{FU6`
zcmbCW7Y7yM0Gzldf6Owo3Pw|RtKF30%9ePFJy`X){V?RtpY<g9RMk2D^~a$fm*m=O
z7MN2~9h;@|?mzk@*8M><eC1T#@ih8=1^SiMhkW>bv&nQLIl?UAhKl;Y3JeFS<bF8(
zd#py03|*m6&woS>XKNfDs8|Txt0%cR4KavEcsck|z~PxBNt8KhJ4rWenlOcT)+|dV
zP$MW2u%Cp~qBw9Md5t0x`uc9I6qQ3=ePc$Qs#@ay=_Qklc;<Z_^2dd~V<-_XZ#%{+
zmF?lI=FoSjh3Gq|%dE8uFjT$WT<H@apwQy+iS}0ZpYe6LYo5Y(@?&c(lrA-&Mko0#
z;I|y9_FZw#5c(bvSzHABXNa$vfceG8Z2E1zX&FUn9K*w6FbnxyZ=9u1+0BCi0neAP
zG{?q_?xj(X-ul-L#sKSA$#m3wd*dh6hvIk4Na@BDaUKKz)$y^6>Zx(4@wAt=6IHTi
zO+AzK8p(~W<pA8s0>ey}G!lk0{{9|6NOYm@(wyhtl^(_<JDhmx=1kqnyfrni>}q~K
z1rIqql?<rn#cA8_V%zI>G%F*RQb*Qh`Fbm(^S(A(-oxDiIGRtbPm+4En4%<GW_Bpc
z10V%|M%qv?wcohO+{xmxu`F<oi=fZE=dIo{B9Q%IfXUZiwMQ6v#_07<-@z^ifB)r1
z&GAv;rVJmu8J&UX+8^3-kGSG|^VR^ae&(wkFDS<pw5}(QHTS<5q(YXR9Z$3`hz7oZ
zU8$L>b@68Yc=+*kF2#Ju(T&4lA@z;ks&(d?m^x=5)Z;^j4NsBX(!tFq-_qnsWzKKP
zBfJ=T`y02Sl+!ZP?1P3A6VwC0=f{(8khJGMR8k{?jwsy!>Jt^gvN#g~wA4;_tQM(?
zBX;@kp8q;AN7Jqm=dTKe=H<TWPsS##sZ!zd>DiU;zul6oa_W|R?W-j|r0U==PEF<_
zlCE#+&#g9KF~JJm)aEnmP?-Ayv^l#uF&}rzyYG1Q^j$N}lX6D_4`NI4Vy&^q>KQ0!
zD7m&JWU$}@G4j18!_WBzMYrz0iF8RPEO6m}xJs1h?O)MlH2x6X7>skhHTA#o;_5i?
z;xCLl$1ZK345F@sPYh7?s&Qew8xj6sA>SAbc`u|D&Tor3p0+YJ^$=Nj1&jOxdj$&$
z4s0ZVo#tZhhnx}GdoUrR{fdM(H41y&*-&oZzh8{+-YB^>v<k+EbT#XSj;2>!XNay%
z=sWWW7MkPv7<qXH69c`N)efV@K?*tUhH2G)BB9aLGD=7wBjqc|&``d@c&HV|!o<SR
zQmAe~Ai7qM>XoPkNh_<-%TtPBD7z3`e#yRnxqM9$|5;46#4VdQX*+2~Dr{`Rr-NR}
z)h9J|=IxR+`(kKq+PWU!@1qq3CxV3PS&Hv6rL<lLDKlOLR$5VA`QgJoGo9i4RywWY
z2_GrJy}3!a5X-?|BYQgpK9K{GP#e+n%~`AU4BFYA31S^>I|z;RG8*D}Nt#sK7iP4R
zS|}9IDxYt<_Os=Sj};um)2||bN<N-ck<`k&vQkkq9uh{sH=gkKh{q;Zj1(dcSX2yz
zGs4&0yqs?ISNF4x;o**sKE$#aNx#+ENL5u47mq8DA?V$FOvq9R+gfqO@#NyhtoOI8
z7vpO);XrE(1D^i%rKL{;K6L{wE-qCznS*g0WaCeaz2sQ}*60CZe=&2jjIdKyoZW&j
zv8FFLkea;QO#JI}iz=`DOc?9cH;-PCdyT+jOFSkyDvAuDD~jGQtz-~PX!QM^{YaHm
z7F<F9-*}mAftcvFlLvL|Un5{WyIj!68gXa`5%3VE_Az)Gb*yjG@N9mP-nv{ks&zs+
z`etQ<V=r=7&4D*<y<L5Ly#5=NL<*VID?6|g>Bj2;nJu7JN4sD;wlzHDXK_B)z_;GC
zJ`U$9kQHqHm!Z+=Udc$L>3GwZ%Id65vOGm#QH7|DIlD!+aj{wVp{3rLwE$*&w-5<W
z+m5QOwF?c7(1Zlg94<)iix-b-tFmJZ{kv)rP0GuAH$@Ty#|+@Ig1aMn!%n;1C;!2x
z6J@Ka8nB)60VAm)(j^lO_slLoHcb3-TWjr3VBlOZM4J?9!*sSzqy-Ef?(M4Nx9MU9
zx5`s!VXGSbXgn}rw#jD|g}wI2m6;n%oc68|-5x_9nG&yxud-+{ef&{9xBvbAa48Ze
z0OZf<jGfvcXEvxTErRcj-^(GOp7wI`Z-)ONao(l#p0>~yEH5sHL3^&<0btDUJi)P$
zlIm{nUzx{G9Fzz-^7ndt^>*;5jIq~aoJY0ID#IgE<yN0S>EdX-eFkI=wCCUZYNX8F
z)%afqKD9VyVk&1y9q*-e{grgc_*Px-V>b&Jjqb<xynNnTAWu8=+^@2W4*RXl$zWLY
z5no!Y{OtScZLLeQH?5ya|DKphfBtCw;xMTf`oUlbJrstIpC5UDE^%V{XS5kr34Y73
zwKd!HDGcJ}bNESXr+QxweQ~FLl0%XK!8(k;boeb<hD9g&IpUue5e0YA?xs4A%$L%k
z|7S+{pJglmXOBDKr!043k<l1&X1i-!v2(zxrF8DEc~>R+fzlvILfuPf7<pv&{T!Xl
z5cqUS)I|Av`$kNJ5goA(H6bAS>oYT{jSImtEKFrCZmjm;ojO00q{yq$zCq<ZzPm%0
zyY4}rdcnG1gV=#dGQVnUv>_&K$#uX;kW3OY^!7rNL79(!xB%`tga(rSR$7jj(ucPp
z79pkk;JXjkKLoq}xeEvGn%M;^YPtrd<lbLjPwsHgx&pB){ub$>Y)!$_7YQF%o4;7Q
z%vB<ohp}yY!vUkOfe(7`oFqtf1$pO(NA~A;Fs_pZP?SqBfXSKrt}VQJ=FP?IF2C?6
z@^OALFtr$4upd3manmIW<t-gFeBLe0#@GK<iElOYdh*`KQ8mD1hq?7-U3+$UdD;N0
zoKQp)yY5dX=Vn}1|2d)KrBpS;!tGvfp1d}cQR~l2|A(J(w9``h;r4yJSiBnGJ$60g
z`qDtgsSIWp^xc0;3`jHwp{VWEBQxn1qjNn2DScxr!2G<hj-x<ee$)s!Kjo8+G6QbE
z8x)gu=w!*Hz%xD#{X84go+!4ugho(4Pe`3!T8(SS2M<ltI`c(x-!{}^i_N#gn;WQj
zvU0NZStUQ;t$307YEm?A0DKbYVO5sYM(qH7)o+C2ESMKrZE(0tLkXX$VVa?tI9|N0
zh6hq|u}9DI1!VdCPdI)CJie67Jrz|#iTYtvHHNS1)Q2U7+K#sj@Ym_hU2i6L8w!x~
z-HK^NHMu>nSdab>wLnV06n4YkzkmPAh=QJ#fL(N;8`9A<!heNb%DQMhm3Wb$C8c!E
zy3S)lD&7TMoSyHY_Ld)n{@qM{Tcs3te{6C-?W4@6faUoSz9tQwnsO%TSlyS3zP~!l
zI^Df7LXPvE%_Zk;cl&8OOCJzOAqW|g7`b?K?d@!3?YNcqtn;}a+9sjP^NWkM5s2p!
z5RE(QX@03q&n<^fnhQU7D||rSNxOKU-5#IIn-gh=%{RyCCvW?H!2GRhD}LZ**8CTB
z78xFLjFBufc5)o-gu+{C>-QhFsipgYkFbMNZTt>n(yO+xbhpkZyK$3e$yA<)K9AR^
zwdvV@HrkDw$v>qH7vbOb)8C`=nh_4|ENjQ|@@je4<6q8>!+UIR*4a5wq>o#8tD81C
z;bU%Roe&y0>(b<S;C-w5Mscs*w{8i=$Eat`;5mKSo<8}wLO%ZC@Bg><(MLZ?KIcYG
z7SHC)*0a-t_R&v%*naQ#ey{z}AO2DM=|>;8)3eLAyt3M+!q+-G&Ze!4i?e*51^*cD
zmfOSk-Y)}#Gk=_OgO9ZCz(E^M78oYNbHhzWFB@Bo3uXFvaFY>a<M~E;j`2pZE>dvM
z&(D^jhZ2whXQnoPO;-P{)z(-c$V#CZB*1$wEv*zzGc@5_A3nHW#xk4QYNJVTJb)_x
zaw>Jqr4F0m@D{s!J9YMv&}iCubjT>%!@f~6rqa59!U&SG;3lQKHCXs#_{+G)2(q{^
zSB6b(<vC*xmKTpG4_V7k&z%#sSDrG?qGcrI>4g_Az9Vg7JbCo!aT#9l7L@(+_V8|C
zY%^ZxFYq$L9x^(-*xW4H<o^Bj@;c)cYAkq{J$dr1JWuc=bAf-F6<iIcl<_FVa`5Ni
z;5h9%Y+K<Qj`E!$9c|Gcyf?#_dQxr474VYMwY+cw3+3RWs^5XLwwMQacg7KIG3NMe
zycv4s9m!@hvcUWT539c6$XJ4bjZWZ+j&wtc(10^-b#*0t&zrQldsk70PtOQ#-_Os2
zgE|anmCvy0JhhXPZsg=aPXTHZ!EE3^7;qWcwzqc6=!IVLeqx|87CiS#)=9Z{#_(f8
z_c73hMxLGrpI_AZbqFFD;MvF_3?eo&F?90!LchClFfvo%#K4NzJQAiq6TiF+@xP&g
z%E9xM!aq346SDZ`^t0xDaNy)vTO+#!X66!47-ja*p(`opReP}!+RE^l{xZ_uU0p8x
z);C5t=TL2K?bMlajts!l<DD6acxE~)*9J)jFX0lw)VT0=W#qG|F%+xt1C1>{w-S07
z9@RRQZ}1lHOCb#BtxM&-Ngho?^F&ZtSy^l=%L_#_tUVPaA$VhbLi@|Bb7wcQXb+$7
zp83JU)A+(q363#h+eAvXaDFw~AtbDJSPOW6;OTi~?FW9o^K_dC9@&iPIfGl52||<M
zf94_jBa~%U!Hnab(5)UJ<v4A7_UuL52&{xxbfypd2Ltf%MqlBFlwT>UL!f?*{wKGR
ziHGm*cX$rz>RFZBfA{(2vzK}GT2eQJ^fG%X%I)90Q};)|uTp-N!DsZwkd)xB|LbG?
zxm4x!tMZMmo^8~IQC&Xj`{)y`#*e_|zKPXl$6*sI*{jR)mEhK>?q23lcK_7h&-Kk`
z&)x4;rK5L!NrJ6cy&1h%PbzEAQ{R=X9zT<(;5+2^+q0^>Kd(=JxvyR3eW$@avQ9k@
zsL?>bc`SCC(`mJSciPhXRR*xpnAUHf*2cQujbYVBSGBXZ;i`?fUvSic{;gm+^u6-e
zLz{+eT($O)#rsvMztrEW%X&L{RL}bNUQSO{gTB_)?|1m*(&qZ~u(!L$FMo<cE4mk%
z^1J@4dx7;!aq<g&>zq>=%K+{-c{&-&3+rl5_SC;>OO-tO-mhw-EBF2NI>58%?T`E?
z^|ne?{;1q4f5<XQL*IL<j7jeOOvbO2e%bfFdb%2;ui5phPG3~VSH1qO>6d-)A1hrO
zzOK`C`WhuhnY-^lsNJq+CuY2iC3$8?5Tgc#IGSRS@}(HiERKXmyB#qq797fQyT*j}
zz?>7NFy)vIWtL(Yvz4+i-*$^GE-jV^MtPEjabeWKY^cWqjA`L{aS=g+qD1I3tDo{M
z#GTl6f@k;bbcFbs*|`dNSc3oLR6l$-3%fH0v?tPb7=d(`0H8hTm&FzHjmbGaI%tOx
zaD-PRxTLZYObCPnV^GU$7HM$kM!#ISMd-L4m`u;jx7i3fln8{YWGYXxD*M1e+b4Ce
z3`ip(N7$n7?KPfubi)RL<fN1dLKz0O8IKtuJyJe}0L2Cd1tV1kx&lQvk}$-nW4z8Y
zup_YxNnJ_TnF#sHqb_6D87_~)@RVXgXbiE|-<}~rNRb#wtIg#-l=07dwQDdiCP-~>
zMqoKSPX(3zI(=C(27zPk{tY7u$1`LFQvyqkksHP(m*4d)V+sBgB_0h6zq!L3LB1|b
z_g$NABu6rutHFJEnt5Sfde2;^RDur#3d)Ychy0npgr=c>2eP{l=O)I5t%?wyKgxv)
z^~D<^0@ffv*lp-Mu8yM^+%UVL8Rn<D3E*(pd@93N(Gi{poqjM@Q*=-yP^zg<p9$UG
zFX1#W()TjlhwgyUdCEeoC_xAU`alU-3ZwKN?J}PzPPEgR1j3p!s`G+8IykD`x!1<q
z>oiLLXoj!FFExC9Rb1?JS~5rS-{&s(hWo?!pS^}6Mjv08w5I3%d6$nbuH$o8mFWMU
zcfCyoiK|=hJj++YwC6>^ijEB_-$Ucj6P^cm?=H6o4|<3_YgsAADM~^MltWQc{0QMc
zZR{I9dU&lNWR$k!ZERdXM<ym`+Uoj)HoN>V&t^k=#_}Fdf`c|UE0iIl7p1l~qqnE+
zO=u1psHz+4%Kh_`=WT!IMc%#0iXL7k{KuPi5k5j$+qG`<et-PvlTvI7HztG@52?9@
z^){Y*cyln8O;63VsfFb>F}0rh=i2Nv4;Wzpd2!jX9~yZBJ?6Q7RK^K#O7}vmU$?!j
z=k3LdjqqG}C<YLlbSQ-Sd097Z;+-h7d8egsl$o!D20SSqY;t-o_mj8I0Sg>2c2nN-
zCj8s54y2;gdDxB*p4GZlhQ~54fJa6&<-K_RsD1X?qYUHi^l>SC>;3Y6-`?72Pac0%
z`IZ*%W{v7+WLay%x)cv|dn{vdV>+Lc?al;+%Vj$|+HX%DeVjFBFZGQ!QDq!TVU!%(
zTaViR_h0?1_V54we{8?^d;hup;BS7?ws#KOlcyW)lTSXY>+z#U?bFXbYtNo)_hoZ7
z)q{r*+j|e+D=!wCuB<=q<HLF=-`f@1kurH_cfUg8@!VoGfUh&*?HIkB+j?+#SYzG2
zYEo97^JF>*A6Z76tnbEhb#=L|udk+_z%=h#7j?s#Y32i7fniN}Hj2;M{FL=^p7+x9
zY~knDmJpK@rKGn;f?paG-khO#yu5T)>U=zBSzWZP7u)roIe~}4@7V-~U$p+a&r8Uf
z>^^=LT=A&poAsG-mH~^=OUOiqkiA`olf&X~r3yV)^XyG~_I#`DA2@F=bA}?0F+0Wb
zKH?#=v9Z;7xGyZu7yrtzR_BwY&+y8+FEk!}J33;xJ1ApB8U51tmDS)*`ici)K;3+?
zRmL-8%d22{#U{P3VAU1MKycbiIZrc=LIAq2jQRA3k&v-P>TqxtlQWQ<S=v3ByZ}zS
zdwXSs!J{zZy$(;*`FR@vJ3oYva-I@H;pAlZ(qzo%<+XOVhjXlV@$AW~d}Fn@vsaME
z@3?r;@JteZ5U<G7%h_HfM+9-uK64#>7#@`OBJj}8icm44Srg-tTNstn7BUp0^uhjd
z@Di_zuM%?l?8<NpE<ufq|C{n~L$}an=YMT&@~G@yY<S!XB^cP@A$i&jMo{z_EYi?U
zi<2LXS5KX79GYQXm`nI)27ofGOi$J)c~YG=KG`5p_?1TKFXNVWGJNrFHAfg6H#axU
zqB1U<W5O0H|HbBu^7P`-#IVT-X6{)Jctw$uC#R-s&ho(HVdY-oY}OicQM(xu;djjw
z%JG!l`DNV<j^(=nG=^vtgQ$&~<r$Yu<||_*dD>aKa2(!&`^Lsb`gKrqdnqzE10Y-!
z=2JMq8mpR1oz~WT4t|2eROt5tSvGwzU(sr7XFmfh)IadlX7j^bQ!c}%H5J_P-_|AV
z2%4r($`5QZe&~cTgrg-tgvQx=x4XCBo;};lJh%Cjj7sP5u96+kP#>cU|4^%az7Kgi
z?0V|=3WS;WL+Zuxz^R`FM{~K_4(7R}UUDy^^afR*yra+DSC>^q(w5;{t!@45DD`sc
z-FGhOZU2($mftDON3M6D8=Ib~-}Aed)60^w0`$Iw&eYnLxuRcE$(?~KnNPVL*BFD|
zmQx0&WQT8~*Lr)aj!~Za=%3+W>U%ZmXZlj_e8DH9%#gg=c`?tF*)MHI58U$!_swH#
zE1X5=(FlDU)*gI?eAWs04EIXk;mc4LJ|&;kj(l_j82NC3*{E$`pT7>55hmojU-g+j
z)Gz<~P5)Kj^rzsM2WVT#?O<QDye34I*T*)OIZ+k#dv03>s_y(7<r-xf-58Z$IIAta
z`np$_O#W3}%Iu#Ce;f9&UlqcT&Oypo?U$@2s<U?BMf8CO?_eO;PU))dYV<qp8QmXJ
z-O7V{&u4ufG8Qx|^_KQuRZjBP=O`7fBwh9O%Px5;=NF~n>6hpDRV3}WPJg#FjNi{o
zUw61#ajp$t*XcTaP121p{;Ln}wS_Rl7h#}<60;zOk`!}pLA97tXlGe0;hB(S87amf
z46RShhTXT$Z0I{z!w84bwJ2H~g)^x|KZ1pYm7&YQ+Cp5|8HuYRm|9c^X5%7)fHNKJ
z^b8+T#KSH>#wg5=%bLkxpllvw6e+?a*v&gVGgr*Cwajj_t{lIa1)D&UD+@Qq79)jW
z!Uzfhv$yB$iNm~?<u-j4Vx{li3j|Ujp1@_dY8kz(+_e~|obd=Fa}he~tcHaWk{n3P
zh_-*Q7lCj${c;9JHLKubC#E6@)pf2K#Zf)$JIrf0AwfBU1tCKiMZz?uPWVX%fMFLa
z28(jRVf>gbbrT9PFLvP)PK*!c4jhERJ4^Xp0O^Z|Iv62&994+BYH6NJX}B>YAiAr*
zVmvS`ue(w7?DV|tZ0@)769%^H^p#S;%i!}W<HLAZ0%l+<grE5aKX^2mYsF0Gg7-C7
z^Kze$l(gMI)yJGy6y-$n*Rw7RP!yOu=9%$N>A|Br5m--0^Wk>p!AaFcD6jb%T2Rb#
z#+sr+`^w8I&&`Xmu^vu_w`V`Rmas&i?YqvQh&hKr(T3rWZaE44+DvhFdK@|tq5mwO
zj9ipYn|AO!Y&`8Vhs*=z=sQnO#`}U#`drL-@|gqjr9-8xD_o6e<IFwt<n}GhJEehL
zzYKE`%+Dxl`%c~3ogBJ#9KrkI^cs(^Q>A};!&f14ki8@H7hi9`{KE9^Yd?E^rId_)
zw*R~&$1nOT67aV2FS*R~zm;;o1fL;a(pA3E@8Nkw{QSzA9s0#<gyQ)9_wSb?pAj6*
z<mJQIgqASS-wbUmPodB(3O%$@zF}?k9wqkqg>Ygw+syn@o0y&pJ#zjN&!Ly?=wP!v
zD`uzXOF`2^<mGyE+;(0(E(JR!jLkKdm-vIz(89g66W`H2&`-lRT-1h{SH<Vd%usBO
zXT9LndNVk))TW#VHT|I7x??kuy1J47_E`9!n^UQi0zUn`m3qQk^i43K?`ev<jt+N2
zhi!a%-5x*wtZlt`+Q!nZqUGbx7F}+_yYVimFxa=18NTNZFRxnxasEPcon^-R?EByQ
ztH1?~4J^VVlqcH7LHnD({L}WwfAmLfI<(oDF5TGlx*hC1X^$WMB((Ra!$0rzwJbFL
z)meM|*-zTWi_P|(@BU8v_HX@*cK_iwN&)u#*(dGfaJ#Ln-fL?QztiSc-fMFU%Xv0k
zMj>!&<;ex^S&L4$+NVGLVf)@+{dxPtKlrch2S5Dl!1yTr8q3;BX@1dOY&>fJ?Z5t)
zI*)5>Yd7~8u*OS4y?@XRQutBp3a^oc`Ng)jw%%4&R?6TdtYwXL+HP&0pIb;@cvTIN
z1m2XV)YkTH%F9}3eav62(}6V}jv)jbDB3GzTGl#><-Og5^x{q9-32Fwe|!JEhi!Rz
zK5Y@*fOniQa~CP^O|9j%2_>+g7*_|L8E#WH?d=_wL2)nZtMwee53UsOhsQ^CFBP?k
zsj1>?m9-yy=22v$mb0!*;5%7Y%@g>dZpv{6ER(MriPhWr5<A}|{NwEOWctXX1P@7(
z*6|38@?W);QIHoA&l>N;C-TC}m*BWpo=6O0!t60<EG^9kK64q5<+d95RQ)M=eSIyw
z$3%Pl<aztJ&b}g3P)bHVxML3TdO0utgQpr=p^wHAKfivzdxJ3)3?6!TUPdot>D()Z
zAY-$*yx68@ri)j>Q&7O_1014^ESVyG1LtlW;>nZWX$#(j7trzXS>YP~UAg#kMz06=
z*Yf<d!m0ujBhu>fY`b@Fz3{j~-sOEBWHzD+U#H)~XyS8t)9eI~h7CmH<K4()zVNUr
z+7cP(=<qas&74TPgqS=($5Z0D`uR+}dMX5b#;s%uXMmkC(v#!RT}t!Ni{O+ERzlmA
zCuqj!!NdDipZ<H#7^16XsLB{lPK-5YKfx`YL(@|ek+-@6mf@suEqS#+%_{>Rk0YBd
z-FJa~H+~6E$n&uqZf<K=3eqpI>V4{-Pz*7Xp>t)d${gL_-^=_wul5@QJR^SB{4bBf
zj2FDNany!PbH{}@MZXw2AAR<?ZI&l0<^5QBV0oWG2+o?<#tSS;o`Q-u$l9T>dB<m0
z$ah)Uxn<f@W0$({(>x!^24u?Gz-1$A>JbLD_v?1%?_7=VaqwPPLNFu4mf<UH#sBLA
zTxD43=eK1&%Y9>1-kZmwue_;EaB#}FGr++$VMIOa;S{?;67OGO7SG7LUGc9Du&;BM
z82Tpyue8A$iXLul+aSs~5KyELWey5(Dcvp@;qm?GYrEP;sj|M!-&0q_52?x+rQ#8M
z=C6!D!P&uRA@+sm!{3mve;@tdzrMXQx>4_qt|9r_fA)K+spnB|KJ|3yO8u^%=taNz
zv*o?p@5)kT^iTN)1B_2~%47bN7ii(`$f#KDlE1%KeFp>gmFI0gu7363{Z>AE0sVgc
z-%~wtzy7`btp1Mj_>!L;?!kkhZnVN&vL;zq(44v8o!^SIwsp@<{8`FLH|TTVxHb49
zbKBg4a|OpjsMM7@x{*)iZ~I=nNZ_RYdXkrF{i{zo-;P72(bp;`{}nu{u=>~g1BUtQ
zWsTnV&!>E&@4X2BC&^b|WEkE1nIwltRY$&{VFM2J2cz$!y#DWLRC1mT?I{D8%RI~D
zl-6bWp^P-rC;DSNykovw1Le=dtFrX9&$GAfcw4vB+jN!cd3~v?ehs(K!r|4rH)?zR
zpLA7z{dnj4eEQ<M(wD#Ut4_mMT&J&|2AF*r?7mJg)s`>1t_@$;={kLll29)z_wKgE
z2<w;tOt>?rERvXD3gHq6!c<afV=^ps3|ScJ%;_-UW2I~+xDgHsj<qv7mtCb4(6yL`
z(Pc16ArZd&uDLR<Ar7gV65ZNEXvZ8=M3*;6?h*RV&rkDtQ65}&x-oD?4Csq(UbIMu
zYY77#A$oD42}40JDkdt6Fh<^C(hNzpn>mcR`0%Ht=w>*2Ssove;3{kmuM=mBl{ZfA
zRmz%6pkS0E5LbgsX!a52BlQXM!;?+;9XqVLr<&bugqjFODUUGrMjgt|vp}Hw)psn}
zb>72^REUtkup9i|v^yo-Ot*<C2edL0_Fd=|W-|RR!(zT;a0!qYT74r#cLtVM&C7Na
z7}(`_nX-k41HmqYN4N@8Bb<k_%j3&V?1FGWlfHlpg_-(10|TFgZPMq#-cj4#+z;$V
zFuP7)IT>SP20w)7F%|^NJEL9i-BXJfX2v69M6jkDWCSxeF{P9?71AZRZO-)bR=RK~
z1Yj9?QwL$4Fbd!745gSc-h^pqc~SN-#1h;oJ2BT3PJ^&Q_?}8V=z?9Llez4gHfJer
z)G-s`V_FEW{4vgIzy8$x3%+PG<F2{HQ*eKGuibdn!<!x-?UymIc7N|v$U2MStm-E$
zn~PwfU2t9-DIn3Ez6<PiJul%W<(S|4Vh-4O4@S;(AP}kF4tk+-ghVB%^Z44|J-Rl0
zU8f;^4LrV*3b%jJ^^4L-lKi4qhf5-nf6=qy%73GeFTY0Vov(i@U5T>4uy1eMGwMOT
zVorZ9jq=pj`rg|)T<8$Ga~Ar@dt&|GdU+G@s6jK(F0=-XwCmr6b|{C^*g7a*h4srb
zbi3%@OP&`?;bm+vppd=UUY>8a$B#ZLLxG+Dy}f8k*5$LqcCf#ndRpptZpvkyzZJ4r
z8+WE=m)f1N$xyP;Y<y1YrSP1ZUk(4V9DZQFEiKQsISNlNWu3Tji&s@&mqT?kgx$E+
zm+DmX`Yh}IN%gP#6S|M*6BaDpZL@O=$$qoWK%&@~nO<ls>kr!W{JpFbE9tW^cj0wf
z*7sVMQt<2Jc5<*)H2dXc%04+P&-SO!o(B#*x^A|k<D-;+k-CrSj2oUYODn6j_M99Y
zw2Qpsj2^HU8%N70L)&=}k5~9;{W#d$u6w`r+yA1?t*obd=kr)E4oWG$diOyan@Rg`
zOxJbmj?0MmO%M!U7pFVz&;RtlxBvLN|Ec}OU;cIb=;Ke?C!aiS`v=aKY;AAvxcdJ1
z$>TBxQI>!3!3XUd-}pwmcmIA{URf%`ma~2Ct>0_+gFDvEQut(Tf|s?iN@4j?-VEuZ
za`0Hfobim{A;e>hhlzD?raW4l#nV0DDC@wejAp6JSwxHk#^~VSB(<~(mB?81@V(wH
zydb{rU{6TMuILulPPxi~W1<i^r-`S>Ve&FwIoFG~mo;-Nc<am}Jh~9u#!#DuNanp4
z)}y@CdKf(hjHTj9tG`AgaInD&elb?TX`4(cTrwD?Z^eJ$S@WBx(cQJx>PvZB26haH
ztHBM4(L#H9*=@eqDMPM$BwQ(8)cTFb9X9AN;><hiD&yo_C}I6(X2bhb-*~H?w;%rK
zC*>W+K((?Gd135UTUeYe{^!+8XM&PBx=~2DI$m&d3$tlgKTpb;UqbpZC@}^I8)pvU
zi=3+@Tq5Ob!6IX6eldXX>^Ya@m;eC(^hrcPRKs(KZXtP&w}C5^y66xi_{-df11{f{
z3(v?&@Zl^n!sEw}s$TaN(vN;EWUif+K}@^g86&7?&Q3$e`esCJ>a+phC>-@d`pK|k
z!ytognmIBwMs^A<aQ;=mTl}PVY-(bNLYIsS+(ZL;E*WDq%Z5NW%}AMysqa<Q;S90a
z&^TUFC&zu<@u0>J?R43A2j?aD%h*(&Wf_w<uWtmu`+R^6WJ_mo2^)&GQ<B>xiB{vM
z8D_|*6z2MG+!)&Mv<#+~c@F=z*~UUJ2Djavohp|+Y&>hzL)xW3=8&;tEGJj#gSl+J
zyKK6eo1G7RIxVj-9yVk%b(j~%-uR-yU{=IKZ*<5<*5Be~@%wqt`eeQoZl{er6lbQx
z>(}P8&f_zlGISYx^RGfv@)kr`N6BwSFX1`K=Dn`2peHY&Q{-NFF08(>?)cl+4OC9(
zNk%jVH**wB(FsOcs{<YzJyaLDfI*ViBI6#<VIF^stK<fXbOuTF2OHb&-eKS_tS7m2
z)M%g8&a7)-k{{aq?r&*W>HDWNJd-!yD*y1=-@i&jT}cJ|^t(33=2I@XSt&A?yrU-7
z&$rd(OZxq8+v;*}$dnK7lB<W``tFkV>VA^qDreR0$?(pQ=k4dvysD|%(Ce&s{L5Gy
zt4@Dk6%NvEC|h3b@QIF<vCQ{To2uN=XHUNO_d4$R*X!)xdzOALpOw$Ox>M~_{s5Q$
z9SnFcc0R0VUgmg(G0b-~#=N7GGOzJxk~t@T;ZORajG~q1diXLptIwKQ`A77{d)lOa
zU#bl*zribI0qNVz|L-Ju-nm!Z8@@<+q%Bo(zL(Bu7^I}CtNK5Bx9WbIDrfEu--k;Y
zaPjKBTmuY0SC04n8|ArrkndHVa`dyV+^f&PDCNNW@`_9I!5qBkg^EU`AVza*1K18N
z^ZrnO{hxpRUOi6gPwJ|FzK{A+nOstbY`=P&daifUP^gE$AieE>zlPsm_4=~(MR~vM
z`eoA>mG#f%dN=)|=l_)H=e33Cb#3^%PS@#cl5S-2Uw?SNtwngp>|hekvyfp7Fc=sO
z3m?xSOa<m16Ja4Dh`6jqgfNTG_-*0SF4}3GX+WVO3`Z6{jAK?$i~LvwzUdj~5KLD6
zc~JEg6OzIxt}$eGj0!_>nflJo9Q-Rh1#gYy4MWqJ9cOHaFCGak<<CY~5&C8-f;*#=
zwo$@s2PF>U9Ol5;76cgz{%&|`<x!J=X?K|OQ@e60f;^`fk6=fs#K=&KcxMWQaiH`+
zrASJ9Z$<E)Oy3y3hF#<qeTFGQy@TU<^$1%*`ER#q^`F5j-`ykR^c{D0dUv6kz${-0
zr4fwn+MSJHK`<#p83s0WVASpY<I!@%VYTTQhQ)mX4uSvp<hX~C%R7WcJ2eN+JyjH=
zhIzs?cEeZNm&;C5X%Ns#fUCBq-MN>yO32xKw%HCN<fIGN>6c6d6nF>E?6j|Noxw$b
zhxZT_?wFI#94Wjp0`NWaryI9|L-{U5Nf(T}Fd&~fWBklf;|6E!{Jov`%&9Vhm64TD
zN^wUC8Qf1BN`T0_&d!^io>g4UNyeyRg3|`dI7~F91)&zMQOKF67-u_dwa2boM!T`l
z5lV*5jg2yPF;4ULBq*5w47xBZ%?{2ByW~LoP6s4-I0#722<Spk`jWO&xV_9ghun4z
zr-;I33J|o+Gm2jd6FXVU@SgfEFJ85yz08%Pey-1T`e&cM28ORMXRhC-uaYOJvY8EE
zoHAX$()A0g`$f+yeWgsJIzONHi|$G8y}e(e^YFFkBibpH9E1G5`)h4&b-5HR3>D}M
z9tPdQCs^mZ(Y%+$n>{2jv@P_Jk;Xfri<FCa8)Z8z{?)6aw!8JL9qeuukK;U3>+A91
zPJ4MC`j(0ome$(py>En`I;ZL~Yv)emsWLUc78<t{8n)Obr&o)P-?|YR9olvCmh*Rd
z__SALXgV(1tcSJHrwmfz2`;+F{OdP7fc8@MX6Sy_joYD%=@Vs=^QZ7^=(G@l<CE)c
zVrsE1FVD5*rR6p=vzR)a!Nr43xp+Pv{i*ZfG%yLhJl}6SThBv7w^QDJJ38D?`P1#*
zz4zM6%4)4GkDq+pc6Xka5#fUmzS+M0o!@E;;n(n1yE~g%a}L|e-SxKi;J3oVy`Q%5
zkm~e1ovL<^-@aLAe>s=T8Dsisjo97U${O*qP39ScnGHx8owW0^y?FYg_U!Seb!O|u
z*<t(1kAB$x(|`Ie6}FIb#p2R>>RJx$#@o*BQG5LOX?YM)T7UTAH`?mz-SYII=%sYF
z;Rqiltk_C;&DpuRS~s@0w`*e+1A!2^ynq6<8cSYnFvrFU>kM>f41#CU>}+_tthJ@A
zf8C8a2PsR~IGe2SXS_XZ;wZ1NtV#GuUNh_W@0MW)ue!an({{FZOF?ZNUtC<Q(04V?
z=Qh(YO1;cjY_uQ!_$Ot6JPv<py)_Qb4=bfT<H9X@8GaZvx{-n5mQwo3<7dT78cW6{
zyh<68b5BS$2DS29&m>WI8T2xax(m;`;*tXPRq-B-Qw+M+V()0Ta?KB$bDlqY(Vjfr
zO!<BDN`>4Gtc;(vdOG95;A3tI?fCHFz4q|Iy|kGJ`kOKkFroyEnI|bT{PdHj&q|)T
zx3=6?mz~+P)INCcK^ZFc_l@Ua89W#PYCdLs7<zcX*j)74XV2Q+-cdW}l@<8m%XsB6
z;N81-H*{bk_q%r<x<I;N{BRBuUJ7qyyupqE1Yg9HY%%W(d9E+cgaSuKGRC|v7f@Ee
zX`elM)ShMh&@AUoxft;Rn6&*M*M9J_dr|SUeE<Ci<qcN+P0Az-@%G0<p<l}j^JUZn
z-!jYuxWe~sZSPb_y#3&<bGT|uc~6BuO!G@lv*8GSP=xZp)4tBH1>hl?=pp=i_3}J(
zXROY(f@g;Z$8BR{GjQo4tZfW5Ps%`*elf<m;Iey+$KcFckk^!pQK@8`;2#f3w3*DJ
zKK!qDYZGQ<Dc_xO2&N3F&o{Qpvq(6{OuWDWESzHnuji{x!W<gknaEpk)OeY|<Q`!@
zFL)*84jxrI@Rem4gDY7Jlm#Bf7_Uv4?0xNnyLe)70#_RdtufXuu!R!^yVM0&gzSVP
zch^?RkO{xlWfL4Dxv-MS+UH`%*?61;SGBQ+QS6~d8Lb%{)n6&~lT)t<lF!T=V+uE{
z56VMh@#n^#k=O=QUZ~(*Mz7$vjbmiunaHT{2u*OE1`pw#b(4HBG1)_pD}Puo)7!x6
zM%&rRm_&vg6!E24*SzEayW#5Vd%vpXqu<q+q{6>SZ}Yz838a2c8r~Tm`wiwbp7az+
zB5;)FS)Tc;ck_b3Q88EV8_&TA;Td_$cR#BiRb4Nu-ybe@_xj&W?)R5+Wo*<r<n8yr
zI8|G7!P{452<}F&VUw)&Q(x3Q@I}3?qdw$|>u1|Dq~Uq~D?ISEU{kl=KFj~3-|v=<
zj2ru2k4$7YYkoI|-Y&k$+Xbi5Cqr?0tC>3)57m{-mooJYeeL|(&6*43QmN_+{qR|3
zTSuhSJ?K63VRWnht6!sge>;61CY5{i&bv=a--Fr)_E%u6FY4^?jNY!#(Z5j{{qDPI
z7}FuiQMuIp&izWG_WbQql9b~kSB*{GDr_iteEwqe$~e_U57gHOvOGG{YlTPZ>VI!j
z$FCaWO1|eKFJ0YA>d!|Hq<YpfbS(Arofk%Fl>6tVq2ymS4ekHg>mNG(W7Kh-{{HFb
zjnlQ^>pEShuTioKcxiRD%`Yw1svt~<PmBkFH9Tl7qzqhk3cbYmX7Lsp4O1*(rZKeP
zTZ@r8=MfYpxfqTxC=5lldoPTPP-BGW=~;&`cV!Yrq;}tg5h>b{{FpLIC&CQj^WxOb
zv<Ss{rhPZU>?k*)a|vh!4&|<{-z&qHL%C(uemnV&BNQ<B5FCYH=!S%D5G!R@1VcN}
zgju0P5{{6-TRWsi794Gq7{3U51We2mgIJxn5ZFzG*<V;%YI6$_UI}S=M?CtAGa&_!
z+DZDtlLwn#JFkT8$S(|98Q@AN0P6@TVJa8{C`(R{g{<j2Pp4+*N-043(qVO3$~&JF
z`iue$5w*K0^>m{bLHD$j91Mt<8~q;ynY5tb?`$#fy*0*wM%a#U8HT6AI^{VAXAu0r
zS2<u`bp<YjpA8BMVGh3vH*%f6G!<-v6UND+3g7*iGd%vB6#+lX`w6g<uY`bnH;)K8
zrW1u)DSh&&=CphnFLM}f6Z9FmN{GN-CL@K=6;6~Qm7w2o788s<Ob5S9;hX$rFdNZ`
zQdXvoWr)Qj=eb?2l!EY$H|I&_BLf{H@Wm-*$BlM$aL`_CJa6YGr@1Gz=Xr3m_ni?N
z4we^M`opVJeHF7l{h$naoqT9XDcI81F8l~x)rBQ<gLgr0*Yj>HY(-}%F3q*td7Ebu
ze=@hn(+A9b=3eIU_{>Cej>gQwblT9{F)=lsXPIAbUIwOB+;#fRO<$8xhxza4U0*d-
zR{#AYTwm1A&t30i|FZNY&w3Gm`{B>t^J};lG8OueHJ=hPw<!Fb^;NWTO!z1vvUsgH
zbLmES3ukrmSm{cm?jastBJ@WX!3x(D{$qA-F}#iQP-gP}MSJn=NtJQ`-u?FQ!|%4)
z`Fo+kcgnbh@0bz(Zo+v(v$@7=Q_#fN>ss4y-WH~rp{es7FJB(Ey`5*FR~JPKD@1fV
zZ<|j)Edxf;+7P!2<8>R)7TTFVVaoDfovHQg`I9o%OiZt&?F(tkOq&WVb52j4AvNK=
z9R@BiItt%$*w>uYRhyJ5bR!rtr0s{VI%qe-Gcgw0;ZGsTC}vQWSMRlj)rW0r?rwQU
zZ9M%f?R?zU*VoIyH8H&!8Y*<+OzL-*%y=2z+RebD83Si9;<GN=!S+Ub^61m{*=L_s
zSg`SlsTvDw8IPCmfB&!B?(RXGo|z8}($>7iK)AoV(f;MX{C~GU{G&f?o7+3>(UYg`
zqmMpH+fS<<AAa*&?e5xoo1U7@cm)0#o0*yJ)gwF}uMf%|9yj<}YoxW0!j>_o&ZkWq
zts5s9=acYOcuH%FH5H$R&n*RTZDKkEns)Erz4qaU-zuXZKAAxcuXPmukmuE|(0$2o
z{js!H)PSsoyb7JmM0uTsHTX8w9=`XmhUD3^jV3JK#U)<vqSitN25q#4dGB%h|I?2?
zZBL#)uQj>Wbv$6|VQgAoyIbM;DDA=5=9as6*OD*%Y1*;BcUb2<9cC?0hOEc<3TxzP
zc;Mh?c^CD~B<30qv4w^C+62Z>2#<<yt@I`^oJ_r*Pp7}U*n({-4-Xt{E-#Xuot-j{
z9A(U`;m&&M@9<umO0Mv-S;HSZe9*nI(x$bwl`>WhB;fq~ENvE=@;oxnt%4I{4lgwx
zar0T%y=TK1{QAjHGpC+yrX4rh(h|ertuhoT5B~#p#sGhTH_%`0SzBAH%iLzP+}n@b
zk-Eq;nxYK66~$sHnL~f@CdL?#R{VC_r|;;Dal_}#&+$-az<JeqsqiZJBkw!2>B0Sb
zZAn<gz!=Uj9B7}qd7CMBInSMa^*pfT?Zp6!hEeJ>yskuUF;?gZJcgG-&h>Lu!>a}V
zG*o*E_Y3yePMZ-k=klI1%lJ}#VmLs5$RDrDKxX`m5m+${qg%T>3~JqrmK?^IqkJAs
zYj;;NK6Bkj7~Fvy<vkT%u_6Y94)P`veFWWum)e9+B%9%t$(>}>w5UAB=I7?~>`vy{
zPT{N#i=Opx-IF5&w{V8zCD)dxR>llnFfS^i09qSijdxRd><uy$UO7nO>Jd&lEi~4~
z-3EY?q41umv>QgcGQK{O7rfvY*mb3Na4ygE1Fo*D+Ke!rwo(4uXx8aMh2Tv&RafrA
z9apt2@2jiYbjkaTA#JpAG;P82R)dnz1Nn?qd5Q*(@|4_=eAcze@VyMY@Z6XfZg^(C
zrUylHz>ra|44SD^KNvCp|Lpx~lV?eGC+6K;BQy4l$jH5_x~tI*VK$o0KqfQli%6y~
zgO4H83rVkJq-4^FL!(5_V2BZj8G<B0V(;!jS9NW<M#dhI`=Z}-&b^UQSy@XLx~sE+
z@W`9@{_pkWxcl+rXYu2Xc3M23KKiG&E7uO0isgfTq2;y&HTU+B0_cL7NYK8Yzq`OU
zMyR+#_eHmHp2pwf`_KJ}-0y`dYs@2{Ri8Be1>ftNFD(OClFD~XTR;1Em)+0m*!9d^
znPdDzpcAIJLYGgLOMK-Qj^sPMI7~a1e<W18yn`22kMZ4fRewh|gCixo=eH~ht`YZK
zRfh8Ft32v<cz?aziYY#v9e<B+3xB8lJlA`1>l(|=vpS+x49wo4W#|U{t~2Awl;Woh
z^b!1mbmGHBX{Nk#EBsSmEusDOBN_`%soT)i{Zq8Jeve_St9!pyhkB@Y)8{FuYkwc_
zB2>Jw{j0sFR8vN;!c??2K6(kAZOgJp1GuL*87_zM-Kx#MBpm(bx%x%=U}tS(T{uuh
zHqIp;T+s)-Kijb0p1Np$l`F2lB&z(UwA1ggtfSAbg6CJ?K+-sy$Ir$4d=OusP5FN5
zHOBvShBw0XHP><YcHvEJ`K!FXDjXZWj>B<yivmv_qw++Ajgw(wZE<0%K0<@R{EWa7
zMwn0%IXeOc<+FX(BkYEus4egEOn`h4;nGUTVk(lhJhx1U`NQN=a8Lr&Re}?R$4m<+
z5g@H1EangqNm+!%-CcY4+fV#?tAC35ZiWt1&5HzMbNU>m^o92JO8(|b!77!IrQu;<
zBh1yde}U<#)zQg^A%KA=Y;y_64;k19Hb6-Jj8-OI9&a8>p)U4|)Q@2*Lb7uVw9&cL
zY2m_3U|1dIc$}q8wVN}v9z{5NHY&>)sGf#ta3_FNodPSRfUV=Sy_I^rQLJ9XpfYk$
ziY+ZK5A#-^M(_cT_IZ?ilSwmj$%sIM=}RY-pvB-<!eR2lv}@xOS;k3QITqng+G_js
z1>uQ6L&#&;;3cC^2`{z%WYQ4mq$5Z(dN7z#8XW9D9PZ!S9G*WWBp-*rYrts26?k9(
zVonM5R-+1VAHGI%Eqw57h4bmeF{TI=(^xT1@!W*vr^%Q25mc@*Xca8HDFsFP8FTJT
z1LwfNc>QfGKvdf$F3nREo+;%VC9IX`aF3u}c#(1`05Q+<npTuE8pCH>TkU7;I4fmy
zYooSi+Q~RN*xAju{lc}=HR367DMPIMw{^@?=*B`RB`BLIsTr7EWysA%oUx?}KAy3f
z2PXwdDNhQw`z$buNbSRD%qXW_jBiS}b7!5;az6BMZkS(Q7?#&ohIik8clhD|>PN%(
zfAoXl`n%VL8{fM=Ty;Ls%0k+l*Cho?(}?5n_cFW%hA*?vSO59<&B8a1_;(BFL92Sc
z^zU;4qv2K3^?ke$pErbx=D*j%(RcNnt7u*_n-o%%*66|f{Q2SXrR6ddIFG8%^-4_M
z)@YhVH9QoIZ?q7dES|``7utem%*`!@7F??>2v44x9iBfm@7f~kTzG=Z!|9nTDLQ$d
zo=tx3y>R*r&#kk4rNmO~K3oqi3M~n5#;D%jThE8R?a%7`E41@v=={u?m*Kk}lt%(j
zf%!R}H_!5WBmByV;oR)vaN^Wb$_p=Y#<p1JL+5Amo-)(fToz1dhwyqYEmAtjvxN=j
zh~44d?GJ~$cRop;j3{Z1@JtK=42JqBpk>Z?9#!b-xy9k+*_4wUDMn{!7lygTwJ;+*
z#VE1nhlP1<bT+&Z`rYkNXO5*GUOaPF1W%1KX%p=+8rWVw9De>E|9bcz|A+s6_}BmX
z-wyxffBF9%e)hqK<>{m!*6-dM{^~FOeE6^atN(iV*@qtmzRXGCb+pOjM+d_PAAT_W
z)nEO1*bQI88{c-+!faq>C|F%xDc(+A_)q<R{rc6~h7CW4Z`!zjf7sZ#KN<igz9(~C
zcwOhVZDsD-WmL)hVGf~y>wK;a&X=OR1rL6YGMJHvA&~N)0b+eU{arX^dre9{Tc2g#
z4Z7n6&(4G=3*2~Vp8t;%$8HHcgWxS3-q_e4{`O}d4j+GfYgoUZxzxgv@B|dsXEQf5
z4)WaD+1o2)#?z;~TX<fzvXG*2Eqp7_B8ENqY;MLYujV~il+6>464D&UP|0vrMmfAP
z8Q?*C1Q99>IeZ205B!VqLcTbcbk5DyhLrf%u3s50rZ3>Q@g*Nd4MsOcj(chUQl6*%
z7-=pnFP7npF=#${neR~sUSnWlz@1;5t@5=yW8$Q&))(r|D1}E*kNLSa&e^Tt*6sD-
zlTU94@7ELe<#6rVT5Toz=+T49kCBOSR4^uZS+>X1>EY_tYl(lYcwt-M@s6q$@wvAb
zdGqmhpdcG5yS&oz&*;#oxW%XI<MN(LTNq!<;rY*>>vC-+h_?#brY#uT`dm5msf;E0
zu(`ce_{sQ1PQ&|H5NCD%`o>nuXUIr<2NvFN7gv|l2F7K3dx^(0ZEI_*_&a!5{hl_o
z1ty$eWP`U1VwC3iJp7k90vQoL=_~Dnj-ugcm9iP{ioT|fG`0-Ia|^Sf6K&W$bH<!O
zo}uJ^QE@51TWKF$nsH{|$-#EjyeKW~(`RHBIEUVqqP7&Xc0QObBbtSZ+Sa+Tcv0<D
zV<GLyqfgy<XsQpx*tCVEv4o#3xiRgbUuwsIzz@EfPZG~v-+EW$BIWCULEXKB3-F0i
zy^OrM`|Ym2WiYJ#A{Ua&)Atq?ETvu07K_OkihE(#so;6$gp>(xZH89qFMPkYWssYc
z`V*YyE%*mc1IT+1a8O`WW(;SDkWU%@gYV`8W5(91%GV$8fM+Vhm$uT@cyIAr571k6
z@<`M#$|(cDSZ>iF?@}+@nr@K;4%%ZDIADTL#Ve>oK{&zOzx4gl(BJxASLjW@8q-WK
z>gt;ZH+g{%{(!OCOK0S38<g^tcCE-Yhjtl4|DSN=ouH0B)|<RHPt=bQ>_a&7PWj64
zNxIUJ1h2s{h1ctQ7^DZgHiLs1Yc)=DSD)ftlD3P^pMFSy+wie*F<-$_Upig^G`=}n
z^cc0S>9^=z_r4D6Ha=GvZ_0C38wW>>E91)o$)bh1+DO;OHB7Xu@Fn%pj>d2IkD{9%
z;Z=ZH3oMO;geoUE?_TAPjVupkmq(L`pQ~Wsratjo9jZ;fBuJvJm1c@pH~)1e)BCA8
zD$yH4rO1C3yB?-=uM+f|yn~Z|<=ZsYzb;5S_09CGecZKG<yF9;?>*PMCT}nFt;VR#
z#?a+<o|W`-p}%!Fggj|RygL#KuDlwn{Kio5x_`}e_*8j*@%m+kH|6g-4&PmPo7x*?
zIyQVAhvV=T1uHmCF{-Sr3~QG!5B9*vgkZEnJxZ7fPgDjX%(6KM+k_dJR?A_m<WmH`
z{xB1m#omg65NnToOf&<Vya_t(wb9=(cXF;3X}O#Uu@%9!lt36Zv#ZkjpAw3JbJmnC
zFKll}p+Imi@0gLX#rX89K&vMyh#27d;Bn_5*`pkjK=JbI>7zCtU=&lXv`V9RQ2f&;
z&q^p|5MYqNlu<A^H(@5%`Lx}ImDOGeoO()B`K4_rj-09GFnY?NTKO3`tBumGCVtN*
zc|!Sf?T|Q_3+?SJlDWmj+RBXqQX9)hf{D2$bhpP>AC6u|jOT5v(;um3w{MshOd|mb
zgD(aFsu<toJ9z=AgS6mexOkj0C?<OK=ahy36-G7fuzi1f*u1x$_Uj%x4!?wu{;^`%
zc++=vrUGOvAtr*2ey4>@XdypL{`4(<28Rg^kq1U0G>ofw45mQ{<IY7%p^t3~VQfs}
zkjFp?59xD^u%0!J+$s79?v#qP3NNst0m`e@pTReR6{9jQGfE=fUW9TUUb_@&`v=4B
z_SSHa-(?IAPNOqug+dtcoTYO*<C@_Yexn&y?$*{%ft3>37&Z2#+kar&m(Zi&Gr_#h
zDoZ?cK^-WLjNNl*XCi1n4t>}wLpOS24B1A@m~akYZJ!z;-#I|GgeCAh_h3Fk{6ens
z!dz?v+2ZQb5Xw<XC+&S4{$7P&yWy*td4&Hfu9|-n@0(qT|AlwI<g5O^I@I@qrQ!Ly
zP&tnIPvP}+pZ_-A>&m<}3ijSp3Wl7OVwneyxxBXS3Jqd3cTS4AnbN-apU_^b)Tc~~
zw!X9tRnpC_<P9TM=EcmNv$M9S3*U5RHZ+Sj6*`^rPKQQ5b%tqg%XRWJBb4oqABGm*
zD_#VjfqytV6X5D>tb^Ld&1&0EKKZy7(pZGFbm8K#vUWZE&!xb4emHybc`J`*oO8vy
zhS=4LrQzj+nn#M?2%W1Id67EqZ4Eoy8@0&D)`<H%o1w#u7|kE#U-1--C?`%*TxQNc
zO=<r;^F#7D?cLeX^6laFom-f{@DQo9b`F04hUZV6@v>WVm7&jBQqn#-*cm?h=)+-S
zb1Ut_!(97l-V?+6osWnA{6G7@4S(?O{$Tj|N1qP2?zE8zFZ0pQe?I&VfBZiTfAT;5
z>F~kNKN+0UL*aXW{a$&J+`M^n`1I3HL-%KfOP8*OujSQqwv@Kcmz|-sP2cez;q7bw
zvW*?y5RbzkwY9ZbbCrom{mn7GZFL){T;|x?YC7|+In=q9JWF`2n1h|AW4^(^wWrcz
zV3-@Mmj3xiA0I0155n)lC<ZaxEY|#|dy)sQ5?hFwtL!WQ=fG%jSS^rx@;r6EH{9RY
zP8$8)UP-kWB=h#2yLZbN1J0*UD3BQ}<=?!mtrD+Zy<GL@O#;s;{3-q!IP4POT(7-x
zQ3~8=pjhSgk-4~-W3BY^MB0ycWcoZfW3J_SKFJ2gn(@OsN8Pjs4=%&W;uyw-<z>qD
znOae9`$*fTneV~NQ^dK$_@NAz;D7U}yeMsD?o6liiDPcBZ3+XEg<JRpUV^r@`|Q)3
zb^cWuzcT*xH^UOc%Eso-;H*)7!=vf)<&}J&N&0icjT`TzPnU~F()Rn2DJ&{8uJ9tu
z7cLZEbo<tw@+LY{9-Al2pjL|40K<UrG*=l)0$cI^sY7k0R`euvA>V3oUSth?w{Pf$
zt;_XGZSR@>XAHS~d9Az$Erc?z?L1+;suzQ-9UkOr9OB^^v-U&JEVU~cFL^I7&krl%
zTiRH0Iyk~JspSsi!k7V*Gfyd)mzEhFm#RO>RbUO6xhL0=|Fe(|+zgsv1k>)`ZXgMr
z4IO0=Rb9A^A2qfZy%vL0a~UVfVYJ-1zfo-fo~_O8;^lZzGPp2a>1PHs#xH%%qrVjB
zX;)i`s<W$%E*F-oE<DZl0(YO=)e9HE%!m&c7!i5zPDX88Z89oa@YVg)25Gp8&#SYp
ztG*0`<e|t}Y46E^#OMo7%44vpj|NkJyfs6rH0Y19=&sMuMty92;4$G~o!6H(GC#m~
z^@3XrSQe6+xAX~PuYOSoob3FQck%>-tCasoLd|1oMRifXh0sWbUca|O3(BZ3P3lEQ
zupC}o9m@Lyt^}_b^2K3zWX#2b!%KBmW-WG2J<yOcc&5BExCAccm#mO)ye{R5o>Y4!
zUf<2zxmD=*9ef`jj|BbN_oNJ5g`3(q_iF3<J;oU?+4NG>Jm*F_uD((C#0AILwH?NC
zWz+dfR|yY;Hmoqdlh@P-z8%6bMyq(9bN3!z!Ci|{!Cf#YBXRsb%zsRCG?dXW<IL5t
z2YBh=sWKDlJhOtSab$Ym#=lNe_oSD+${xe?*mY03!#u$?wRyiggx5Pi$1-`;PK%Hj
zwnQhZ*1}G-Q{1FeH*kvw@6F%f)_y*MCI9!kv@bZ#J>GX4ziKN<KvDP?d<Taw>REs$
z$gBP(gDLMWsYBJGqt@+JUE8oV85M=fqaJcqvi|e0+B5&A^o^q;cV2lRQl~#G<Fz0U
zW!7{5SGTbYDJq71D()EMSNHx;-pZ~q0w2<r<l;&_rOkzI*A}X&BR-?sO)PCR!dQ8}
z9=!bGhd0(~jN5H8=J{IazJF7WpAVwFA^BlATE@}q7YW}c&2jkZ@O6E0^!w54Z3;N@
zW5d^RI1X=7Fln7zoEz4zUo8Qc64|yk7_Cqc|B9)nJf`d@Ls=%oXPM|2R(!7&aA(_l
z3!`PSEJiH?jJrGuHR2I8tgORqSUpVm5C;QNf?)(+UQ84L4<7WEXqaxoZmkQ$bcQjg
z^aRJiML0hfrh3|X#vE*d#(1}I@~Fz8aIkWY@T^QQpG`ht=StyYn?njEhA&D6%KA1A
zQ6L1iFzZ(MJ<Wtq0izsS4xJ4%e>OrBh4@MYU>;wXRdLJ9>=AD`+m0Nh%^r`&Fcl#X
zSmoTpkE@vfz-*Nug$Tt4q2GS-XHw5oXS8jvin5}v6fxQ-ZHJkjjAv!A)DMYQOm^B#
z-X#zvLNlPfCw9Q-zQ(kqZz&U8`nDJv-;<YBSQye;G3K31st71nV{WZ)54)Rt<NM?A
z=I|{hWnJ*>MUA&7&m!;#r*w;PSjxp@2~Vw=*1!5DT~fFKXG>_w%MwD#J&DUu2}dP1
zW(lj}5lSh2C_8{kX#_{CjAaOdm*Ob{^J@&&%F*Oe+oa@PD?}eY2%cMUYx~d#84I?@
zdQi$4hF1o!2Yb7NZHk=dMaWSXb%pPYUCJ+IZeTz+7@FX|ak8>}A@qilxz8>m=or&y
z0?$Fl83vqU0Q0`MuvkVhu-Ddv5ttZ+OQ{#UJ*X|HC<GYdDVHL&8yA76i~zPkNx$~1
z7b5J>)hbxqQavXqCLc<i9xLaD#S2Tr>e||H>U_rY44Pn(0K?-rzK+9hdic%^UztOG
z?~TlS5#6r|6>oh0vLGLE-kR(4_3HY%k1t*i<F~%Xwn6af!x*LR^6k8Yj?r<<`FI3H
zqdmL`7$qqCD7(w^C3K*-(c%qtBK*Z-(za3sZ7Bu&0gn<Z<ZZpi!>PA`rc8SFXft#9
z!{Re|Q!^M1!?WSh{+(fS<Ib?Zw;eu!A<BIGA~cC1gQ2Y#7H!<WS9G&Hcg`;ki_6!B
z#ih%G^HV5ioiBEN;X-(V)qoZL?9};TCiKU=P=@ytq0hBlB4wE~qc&0|4>5)u+XgAq
z{E&Giv~hQ5C-X&krSJ?FFTXP^ufCTuuZEvl8P3fvq#Si?LsqDNX!yqP^wERiO!9jX
zzUFU#`jg?${`}9=Zl{Ok#l_^Kp`WG9`@<jn!M`2;;xGSVxUjsI_u(zlUk`S+hJX9}
zzd!sx{-6J^;ge5ql}9f{@y_<<u)DQB{Pkb`<?!LpKOB}XEDzs*|F;HbhOJy&OI=S7
z`v-WcXQ9FK!<8%7hf7zkmU7x6oH~yxw4L{cE$`;Tn=y7V)KDVhy?8XOJG&@zk2!(Y
zlX-+uf|8!%9V|hSnr9fK<jETY_Tc}}c08s98|P-u43|>gYWP2V*4=yehnu%<mGMS@
zG89c-$L1QwFop>8r8$s-dX`Zjb4+dFmvYSAc)RAgd!7@|xl)XLySw}4Wp92q_a)=v
zi}fF40iH}8#+RDgGan1i#H#JG&+*s*dwYnRE34ephawvv%Bvf0!1c*c!F%WInNxMH
z)N<OBvb6b0US8*uud`<@!g-Rh)$#(EEHbkto4gm(FITTzsx%bq+L!@{p?Q7%ei`1T
zZT7%!K85#q1ml*WLVoJBz1>^Ry?oI|j%S(MogamVVbHOC+9#jg8-Ds9KO8>0bvJyq
zv!8l<P740B>3>RC@&Q99{$MS2vjr%*<h}R4H>_T~Jlsy&n>X(a_wKEiL4;AUj2CG?
z20)8;!0oDSN<%*wA}nOX6SeX=uTyE|F{I!p$VshS6yR4Y*x{M-dHpdqI8QDkJFi2l
z7H3OdvmK|Xw#T#;B~KX!JX>4B8)fKI%4=-}U-kF=ob$8JWo#`~yP}KQ883&QFC%OE
zSskY~F$VBv@SexMeylnHDL@o&+4g}lFkicNrN+yH&`%`Exm30~G$z%px5{d{MH|fX
z_(~bJB2^zAhG2l>;!1bz>g9@m_wK#IN#hVc@)U&+-~o`aQVX`y28`Y-D=UCZ+}>7^
zp%wpTY~qu}@A?HtQ@(9McX#$O*3e(xVre7om`ai%@rri)a6YWXSZD(|uIO^;8>2K_
z1&_Wz?B0aN1B<Nim*F8x{?)VL7!#N5#!F6nnUAz(;{fCRnW477PoK2W4j&qL($BoW
zjCU|samHj|hC^j!sAuC8jzzYuc~_kpH<drh?rj9me3sV?xja13M)kG;jn^qTgnWvY
zsjvPpXX1zP(E6^&C7du``^+{*?SsH7u0@u{Y%gN9&1U3*;1&8+25tcO))lIM9S$El
z7*qbLZhpS{u4!Mr_urALLNd=>IQ{Oi=LcGo_Zb)U&a2`x+`%<I_GJ6{GU}m|9KNXE
z{tKPXg^w`zOfuYIhC<5g{_#jp!1JqV^kh(xSNCTsSn(T%>hHXpzDetO>ia2O&qbYo
zEq)B{gJ}XqpTG;3z89~qYL7%IIF#GvR9@g3;qY5on5>2vpCk9aJen75s|g3dU?Bjy
zh5xBtG%{8L?UA}L7`Epm`ffqA=QeQ3zvGJMH^W(%p2E^DbqAX#uc{9$;u#HrRi&|S
z3+Vi(o)Y+_{<+KH^)?+NG%hF2G^SkARjP{i>R$h(&%KRv$=9>I({E)w6zNEK^=nKw
z)vMp+|Aw*4?EEVtxN=oYzY`j&F?h&zSfBiBy;0}Y^Om`Uv1>@X(=*X4v=Q>qE3>YI
zWDb%l|8rr2tDqd8>;HOw6^`UJhHk6A#ur~m^``Jm;=Cz*5%w>J&%tvX4%_3GaUBg`
z_x)`R$A+)ta2(#Yka_3)^5Ssi#`WQRgmWHQl)olk9uk<1+Tze^qrB@$y@e<fyLW_3
zXa1DpI<%-(+$BvZh;p~;vXs)fFu(4m-!ox07^Ir9!szx@D_j|j2>qjCf!9=deTCTd
zp}uE(RdrxqxNsqixYcC5K_Wz|R`Q&fC-iYX->sVSTX|O6S;dVRaHf}o?)P`LhZmTw
zl%?!qs$5~fG4c2xjKH+LLopT@gXDicLIR~s2?fa)!)#@%t=snYcZY+$o$?fu?iDlE
z3=SdO850Bn^?&@(7KQfLx1#c82|Tt4Dh4qEkX3U8Bj<B5>dDU}0G{&H34DI@Xu~)u
zlMo18WhhfnFL_io@|)3%VAbAUwt#FVnj%ZykE=Z=;|ZZILSx=ziYY0!BOpE6?`=Jf
z!<)k2b)s&@wQxKAsUP)M^<%!lO+^@Aa0R~dn1t^<BbyNq4p5F|Y!bMoBbYOM6|U!*
zSEO<%=m_%qm|~0(ijtv(m;5s(N@13ADYOXV#Q+B%2*PlM!Iu%c&Hy+#7#<xw$QXNC
ziU;rDF8q|Gv-iMBaY=Y46!EwM17rNOC8IINgA@>@6imDdsju=Mr_NUV*2?0fTe+}O
z<K322?z{>heXSJVw=L1bss|5>q96HY<*A)uXCkOm$d<rG_)NRnI+Z|<rWI`p13!~7
zPgzh~0j;eLE4fadwGx`*(%4I8$Kf|Qe0PSg<qKcrlMhM9H^HMGpI>@ac-!&u3jlvj
zFsCK%s4PNHrZ5ReotytV4!Y&rv!}hyoN)E(S}7b`$;5DeVtAA}mQrti0Uc-$0rbGi
z`$vx-<oWclIKNyp^u(z)#=m&BF?{gTzZ{%VbbiL!SK*aHGfte~F}|1l4g%`)@F-7)
z#}5udPM)PqbMIW{dgp{$<T4EBMxN>9nZ=O3`Qh~0`SLbBdop>2R@GUn2u+DsisoJ3
zG!<=SFrm~t@#0a+wSC#fa1efk!Hc5n)LH!4T+JiHOGY@}<#WUA;`PwN3!#xIGnefj
zLtS(A9I}1?=J0R-?SDP|&5!?T*xa~3{LNqg_3&T+H~-D>@BZ+Q>P)8d=S~mz*Y6I0
z`al2K@cY02`@_HaH~(h%+rRx_*b7f{^RwH-M<0JY{N%?!9{%vZ`=jB{|MIVg-7;i7
z9&X+Iw7j<P9SlzvJKT8p-QmK8wbAf~2R#uUYrf7aTTEM7c!b{9nNmv^(uY@uwY9Y}
zc0DbH?87pY@I=C2F|_dJ;GtkXVf>hzw{4#-44)Jah1OFn<INZ@Cq*#c2=B(Dictun
zG4j3p&W+*y@4Z)sFZtfQeY@(4?o;YnWT8$xnC{)%7`7Q1Yi{PD#|UFlg{^xnYC03#
zdRF<HJ9@6hgP*DS>1Ft?UQm(wd2X(ZG~Y|v6#tBAWsu0^WxB6;h4U6w!9QMQb7hpc
zdgXH3vbWbOrSu$+K1SC*O9dR}Xo_lZQpoc>D`Q16nz6tm_*uCJMsZH^OiJ6iQ+^ju
zTT0XLsL7jA!yNqXyVq+=N&Lo>^gCIB*VXM?w@Z=E&{MlN<U3CXN-_#c3vre+_cIRq
zwz0Jxe7kQtFQc0|{oD+%q4Vj-jISq8hTXk~!-qfrY`A@Ivy6t#1{wRg<p;yB^QF{b
zZ<q1+NwtwaV<5b^b}9X~Km0F${+D$I+k=O_rK52_DQOu`v@d>VZ<nE_&sl;C`lEJV
zP*&Qa$_)<9q_6cm8Y3U=#Q3D0)s45@nX|pk-ud*6_5~k)k4GW7i}7K9??K5-`}+^e
zGs@PNyuNrK+K!T(QwEvPH2BYR&Y~yZ`s}af--;?`Y&ol!Ct}MDsaO8A1)dDg$LPm6
zRNqpUN%p9};1k)z8Bn~p_HE4?`m<WwDK3@aSvzYJ`QQ<oo{}Tah5no=FH8oUT6~p$
zQa5qdF0G|6R+9F?aOdv5ZmY!6w(5#sEYGCm1;?si(v}Qb%GKsU!NN&fUDug}-UeoM
zvz<1$oR{YeIAce-#wU7gtdL9bhj>O~8Qw{YmNI_fA)gz2>A&jl^kI)b1yL6Fdq3b0
zBcwjoe(F)Mgx)xJ?&8G@wFM=9TV2c#3bZXa&o#2Ma*aW4WQ#!kQoDTMaaA^H6(4l1
zy7Ziv;xnE~zDrwadwqgV;ib_j_<`;)zB|iIJHtyf!PvBAso#uj7Kfl^6O7K_O*INe
zaA^-P8cP<g^?7>?hQa^PUGJ)`l7>-w|3PSW@`dl(X>2Y)_f-`Trn{$Bgu3eYVVo&?
z<7VIc`PC)!Dj@$?POtJgj1P`i&r=XLZ;q64<Xc`&mp+^f8%Z*L4?d=?wf$tALL-w&
zzRFYlF`v9|xt+|NdL)Xp9bOrH`a*Z7K4nzuMacTryDDddMN#hh!2BaVSnAtYAAQs9
z1T^nLHr?|lUc)uQ>2HF0<i*st-i;4a9Pq=7g00|6nJLR|Bo-r~FGT}nQ&}~<9=&gx
zY0P-px+wBl!2`bZSHlkG6gZYY1yABOu1_*#r|CBCx>ikdSLb?{XudbjhEAi^xr51F
z9jbkfUP)7Jp#IfYU0;6%Nc~q26)H~J#2ngb`dddG)79gY$Mhy)dQQ1@jbVD4@|d1q
z#ps~t4@1!leGGpCQ=WZQC*#`KpN!Prbr>OW8_W4Z$m@dKy`ofP4@bhA-n|M_n|xjP
z>Ri4)Ozr<=*RL|Xsl2!S`iB_a)DF{iZ1_43$Kfpuvx~FC<#*mG;|=EML<F&#+<eSr
zX3`^cW*FCsL4q>I)x?ic!aNj%k?5^JZX=m_aAIy4zsg_}#so1SRA8_$93{Lw<JI*n
z<&6wS=w}g*vFyyaVRW7k=OTd2TA6o#Hg`%QD{<$-Y+Lo#E8sBwgrH((^1B$5T)pz{
zWii|YL92W)PPSd!+1f0l7Z_}L81|<aC&Dx3HGnep*EVrhA7bdSGqy5vE}FBw)-GKh
zW^46ipRc8zFmcniiDjg!Iwob>OB+$zFv#H%3C+A>9@mPuk)cSu@-|tZbm{h^SWrJ(
zz4U51%wpOvaS0{_Cjnzf`1VZ*+cG#ign$J$iY0L{Y$ZgUo2@=cQy1e*sKPYk8Ze?|
z{7W0^Xl+ppcG_4w-M@2h*t)k(e|H>yl>sxKvGe>1+(@6-Sb=eEge0t%5RowhH}swV
zjHNQ%<Xs8B>14_y3Iak6<%ltfai5!CsQ5vZ+$pqLaD@-XP>-!MHRioHhHQ(32?iUy
zg2&|@8NA`~#k1;3@cT)GVO|it8PSp+uYomr^4=sgQ+PSIZ((tHSXo)knCgSYYnzn7
zSBAP;$!$zMsd41ovT4;O0lWpdKI?-*l!qIS2JO=;np?24r79Skj-9GAIQAbtjPU(n
zI30MFE?fvqb7k~3R%~<j@KNfPXJdNzV6Vo^x!LxJJ$>e6DYGb3jo0&Y6tax*twf@f
zI1ay&;Vm(IdHAZ!|Ng%C<<<2~{yJ`l&!_%k_^NNh(eK~vYDL8qruSbF4x<Eid_!IN
z<-aj>6z>W1%_sEt_HsctE?>SlEH5pE)@3~BeWC=G=jDmsN}8FXH_{sS3c*M4fC*_y
zuvllWw(0x$<By6@GOs%;>c#UX!@11gc$af?i{V`;ozJA~7sIL0pL3yUr_L>hPA}$n
z_?&r0`s5fM0);<m`!s&N&o+mh?RzQbN%#gd)Ru!3H<_QGJ_!ANREo&*ZU~*?;dRnB
zf(%?|YzLNn&rq1=`3x^L_wXX8C>q1lrH=TT6DQlyBs_k&Io!Yd*>JGCJ#5~;J^aCc
z`v=3n`d9y-;U_=&V7PmCWBAEWe>VI$*Y@`A@bIz4F(-!Gch-lW{>Kl74?avgr2Tl~
zI|piazdaLd7l-zv^LP*KWWfX@%++hxik=qNa`N=Zvox(oUu{WtadoBm9ZGxi0Iw?w
z_e*vz2tCH*@cu2{C6}$=cyZw0@GZRPotaY#+2k{ofsdM$(iG<6;L-4qw#Zw*w^3d_
z&!4?4&m~3yi!6F^MZT}!Pdy$F7nYaG%M6|{0`QVxK)G{wz2;8F6-sqyjWO_1z?qjV
zD%(pNwO7&e;Y=A@@z|7dnd#F8$^ugxy>1Ne+_=$(quY1NFko><D?6nN{sy=4iExw=
zjWGak$7^MKYrEPNe^BLxN1z1t+iYba2_BshtmewJ4SdJPmI5;Knl|k-ciW&kKd)Wd
z1FzaYc^GN7fxL0!>hRvX?*^7r#Wz&F^9_HAH`k`+-GmVh1lpfRp84Od4(9TW&COwb
zBmJ<oJM10U*70TV_FT=cj7~MTr#%>+dG{$NO?b%jPh-Tm(m#0EGR7ngT;f4=!8vdj
zS5&i%As>D8ao`C?mX}5^gb`Y~X~EyAGC1+NdiZEOuPx;iUlo{L>VxK}TJOft{Cg`^
zC&LKuyr;!WFii3KQGhzgmzO>IXmO76$T<vxw*BKJ)!Va1t_!{~)|O`*qkYPQ8+bfk
zdA2QNK;!k-TSfASdQ#q2y_H@s0)mt9R2%7kIHC=~sUE)LA;GRKd7~{~SS%PGKY3XB
zfPq}%JDRw&yHo9~<HciyGp;QrP!9$s#z<p^C)m>RV#fSR@l-tS?q!U@A29GjEW>4R
zSY6!V6d1MJ{Gu}&XG-Rhrr=6F8A+vuqv*PRaW2~Z_1@x?p?Nxcx;(nl&Lzhst@G#b
zq4F!kblQh;TpL>GB{YxOa-Pe#R+ck<;>R<F0!u0A;XwWv4VOX}Y;mYf&hk1;9Ai#;
zIPH4+1l;OuM$@M4L5k+-Gh-N^&iGKWX39%8h4036qEy*n=XJ#x>QcV;5euC1MVIxl
z^AOE}7Dk}2;L(@Umap=(=uSUt8)@M@+&~N9A{wIa70#2*xz_5TO?ihtdi<=6ceT|o
zn1koa0rXUZ3gfLGrXPLpIsNlN9$&cXeSepx)Aju@i+)w1>2Kesw4VCjWwaJVex0YZ
zm0#YK@j!o#nNMZ==)Jx`Q}oecT7Q+Mc@^b#>dv?Bt})-b#`thfl$O&OT6&C3>4eHJ
zzZsl+zUVleYlo@5>N(eB)DfmK#~+9AHc)=K`<Fkzee)`Ja}#`OV_WN&N~SAZpli^c
z-A>xvyj*&mj4AkKPC@74w#5PHUA1q(mj>>&@pKAc9^p=ef+6Kj<x5n?D_5SY++2On
zxAC327c9BJQ97dBr+3Ow7i~L*Y9sYe|LPZGv-@PE3-a|Be1DNK+x#hfaJ9@XePx+A
z$-eXTAm8ztf~Tq93X1W$9_#;CVR~PA)RVv?6~3D5^pzc&8ZWes#bOqb7~^H!R_|nU
zn1}z4z%}7-{`I{h9JwZ(JNo;V6Q&%eaI}=8*LN9SwaZsu$Kh8K-qhBAm+PwnDdyPl
zbsUbv+Y~Zp=Pq6tt}%Qq&ez0@d7}U>AuLR5Po54LepU)#%mGG3ps@C2;y3dVdNID{
z(iXHCR49LQ<(rAL8A!^SUZG=B<kF5omu?~>=2`huHiG~Iz{87T-nK&X?&o`DsB)}8
zF#%&J2sL1~0+4{^d13_HOC|M*(Bq(ZTPhyxS>Z^i3&Wi}5VI19^H*nDjSRT$+c0D%
ze@woE@6%{yG+d-4kFZpQs7Nkb71=(BpzUf8D*~fc%eKg{;xn)lChM|-krK{UF%(F)
z*63{%d!-$(HcU`z6o#aXb7_0Ush7AE8v3r39qB9e#ztTgYlTi=dcqL#xK~70dtpGs
z{5E6WjJR!5y1%RnPW+?=dt%f(%vcGV$>-j!_2I#`1Bs8ruRK7lwEwd(p3Z_WV0azj
zhkeu6#mtU<TZXZWMR+kEe4{jDz|_YTE&uhZONj{)^Ua6{cZ>sy67P*ueGNh29j`6v
z;S)s(ydijRZ|@9`9y|=L2fqB~<@Lxa&W90bGZx^vb8#eqUvPG2X12yG!?CRm&sb#}
zoIl4vT}*c~7`>t{6$~tU2m7U9q=c%ic`{z$du=}xnm|dWZt%XgSW0}jYfSPk(1z*&
zeq*q<tx9|hK0_KW(&eQK!^w;bUS8@GFfwi)*5Uv{;^W5;%6L5=ID1uOk2iG9madc+
z3yWwLC5FC0`$hpNZFd}gW5Qct`10RZU;g?r$o2V8N3Q($wbvN;^9jdGn4VFa*S{SN
zN5B0dm+<<#FCQ>1fVhiU|9v6!D++~Nqlc6n(Uy>n8vndlRz_op`L8@xrgB5`7~C0#
zN{KWnFH-2kNBhH_+n){23w-kAVbLSoklBvR9BJ!UO4Qn}YIY@QS9A55q_za)y>4bb
zb%t?^5$o}A`_m7H^}BaMFRUhcQLw#y`C!=Iyfv(^-we&&9k#YNhP}P*;q)oAzG)Xx
z3k|aMn;Mo;|MXnm;UAn)xICOYqgI)}QwMGG<l)xv!Owo0eD4<Bw6Nh6<3@O!m(O>G
zKmT9;$Kj9u@DGQZpMEy1-(Mg8_GcdqyL+Ltv&+f<eCXQAVPSrGxbe=rsq>X2qclGm
zdcIiizW@FYhRc_)mZz?>lAOo7w0vQ>a^+gyT?ilNESsg_@|DZOJMX<ST)40@+`4_c
zyvJ;_HF+-a(xBw3v$Vnwp}~}|`h6+$#bRi!Gefi)r3^)NE9ET?YDF2tnloxHgkO^e
z?%-`0=XfMhdRr`1#<BF*V#+Z`+`qRz+`M&X*tqX3t<-m6aahP)DXuxs*<zo4b~Aml
zU7lF)z4u<~cc~0=w$Z}}Z}035JKMWu%=!NJ->bV_0xaxM4@L)h*`5#{pFI`6ENxjl
zbl^TW(*^<tq`5hZrp^wRF0Gbg|Km?S8}4sxmC;K(*~X6dBJ|~XQgdR`!fU)CLkiEH
z;_1?U=5X!9Ah;A+0RQexD9Xrgzh|YK<ymFxHD|mU8QNK$7&7Ixyu4B#cX$eQqU`1w
zZA@Lec4@eL>0<FB&Oh5&-w5`ZGfxbwD=VqX<(h-<t*@8SO8-&FGJ3%Y@`r6PKfBc)
zb&qYc^xUF2hAUe;9#s2V(1O>%ALu({?*cEW<OB23g4$+1mw;ywuNLE^j<#BME*ql&
zMJyaFBVB%LE4)J)AH$E=Vzi7S^1#CfkINX<#x5QSqvBPf1mhAPRbJ=GLwg&u^6Bjo
zd0mo;hVqbl_uY32N9`Wr{4rw|>#)7MpS)iVOTo?6;E+5wxB6VR=IO_+!*Z2zEotDj
zT_lW;%U9Mi2Iq@*iBBGDm|KY}KeC2h0F;eJ+A?}NZ;FvZTaiVyZHJd-5V>?|r3_YT
z+y)qnq?(4p!`;38!M1#p;u$T4r_$EhY$>xBCuW@17Sb!h$+?0D4cOe;%s2_2qg{a?
z+~u{AOPi~czSYaAMj2}KGeal*birp^rs~&|8Dr>(cHpJN0GdcCGdyYdJaQEppr7@R
z@t)5GJKC-d&Yy2yPk-7fby5g}4Smrzr8pnkL7m}z$=GRod6a=XdBD4qq2;_LQ+(BH
zDPu?7wa<eCi%$Jw=<BUVc_BVZUGUcO1#5Xq=1<w$#5L7FO<aKt&WNp^rQpu*!sXzN
zwCIE~yhlHHdND-GzsgNp8RPmF&R8&_zl|je;FRt%eud4Ajih^A^bjr6j^t3rGw{0(
z_L}CDVKCpp>n^9RPdy$9Qz@Sh%@e4zJo=hK_sJOYsLiH#(<k?S&Nt&VK^~0k+P}ul
z2wU}S16ldie`NG}(l+^)%%=R%0ipY&bu2z==Q4f;KaRravGPi?x*z5>mih8%+#~%w
z!sQN*DSnkV-u>mFo_j&x=$%^igDAe`rp!v?P5t!0H~lSb2*==om<8h)rQw~f{7(Jh
zHN2P%Px3Wp=xgXflz}%9v@dupdVr5brvfK=P?>0@#fEe^U8nYpo9bC`B-HPEg_5?h
zk<nVa%fIq=mxrjis&B_?N?I-AL()&*yM(Tv&~@+M)h4O0d8Lhzj9<;8X^hEeW(>6v
zTwO<cp-kB)xb<(f(};a_n|GO>D;&9d-)%C6@wwi~a14j}<QY9QAC=)hFDHBrSfL4w
z+GJ>BoTq0m)~Ha$`fn`1!u0#~ck(>??e*)I6OQI~H20&IFx|h4FyYB@`1OSEvi*+@
zU&rA%yk)_eG^;nR4VT}!Q7ewa{nmtwp(w^7^ue|!7^2#WArl;CfiP8!4kj&snWk$J
zb$%9QN=>avM*-X`{YGU>;+8;|iQeQdFZoDY3ZC3CPG%wI8Uj2Rl)~u6Q;U!|H#?ua
z3I5I$u&4OR;Yl$<qtHj#3Pa(%FwBhAm2C_wiDC5IvR%j1;Ysd~B8WYW@cZ)lla$rl
zs@SVt-HRa^=XJt<YEsX(C#3Yw%VB12c9>gmw%6&jNi!Q(WAk!(_%H(c<A-TG3_{vT
z-2+3dB4lug0IUARc%&?h#I$F+wwPO(tHYTodI%h0k*cj+ZEzfFTSLkb0-rE3G~fq|
zwjlUTZs}!^!z3n7t?<@o`a1RQI&>ZN5#hH5xhKguFVhC@Rv|KA*yfP8*ON!b?q1(%
zFlOL<#uMR5pSbI9eN3pQyllfUBQgb&^RgIPc{V<8V^Q)k4vkqJUghN{yS%Tw<ZlcF
zZ^b9WJ<YoZbtWD`#R1+KH;K!*w7t2N_raIo3~#Te3|=+HMu9f)Ftmb$f|Zicwm()`
zGbEqQn1_>Ai5jy3BIELT8RCd5r~zTymOKoy0;QcTlAcp|22XjIQjnr!JA1q3?d2>L
zhG#H=r4-U5UK6s>4_>3SwPEg7z02oW+M#eirJv5gHEz%}N<_x#b9p8_J$;t)lLzHa
zM5ok`VAp3H67+3@Xv?BA=g^|$miDxA?l}BLgzwJqb>YJAnWH{;eZ!EUl3=udRrux^
z);#uwtIK)w``6!ny9<od|MgezzdZjb)brub&f(QHKAFGrRC8orB!_A-;}=iB%i_xA
z%f)*L_!Ge#TAn?j3B0E8ALq{7CT}_O<k?|oXRq=%-Y;LdHe9;;y<u_XonhtDwc)~2
zXi@T~5I#LKpIEI3+~2!5Y;SEO{fR33MdC1GJ%9FKxP9|hDdd&!%qd1S-dnsB(7`ij
zLYp%Eqx1L#p6_R7E`)Y1g}%&(CWX$Pz-Ofnp>N?5s?1Q!;nR<P9{RaEJPf_wxPNQ-
z<3Il6;a~pC|6=&?!=DfH^Uhj&R9hGx?A;%3-TY+upa1kvhrjxZzbqx?jqr-^mQm+Q
zc^X~1bg^h{(aUqQWgN87i2;QH#I}0204xJsfN^djPZLHX3#ITx?w2pGCC|%g|J9<+
zAAI=XP-pps)|<~-@nvhgK34{R#Iu-(1#b+>FZojrgWnusUeYIJEDQfcDa-J)6dG@9
zKZ;#jA2JwBc~RJ!n<%I#+IUzn?3gp~nHN@8h8s7om+=f=YfjTY3=2Gp^f99+&p`@z
z-pv%#wpYA+Z@tP~3643l2nyOVZF6I*Jd+r`7{}o8+Qqdpj8TePM6s~AQ0JX72=Ml|
zEh2>~gA#+r&6~GN5yi0K+l}z`@R)ZBzL?RKCk}<ax-rhe@hev^r4QGNY^e8g;Ib0m
z`DZt8-${G&BE)Y7w^KGFg#I$W8E)_oUnzha|4`c$CILPc-cs(b1y7cj7s~)?tH)Y=
zl0W@wzUB3|v;ScD`Ny9QAAS5;dG+xA;i+eA;16!!xi_re-%Og86R^MdWd?@w7_@C@
z=3m>P7k`n<Jk99EuvYDvy4NC?q$L~ZH_BFqEPN3D0-q0`Q3s1=o<4h7Tkv(DVC1cU
zXCYIh;qni!+={@KW8fnED3X-_N>)m`^0=2O^~RrwLxw>E$Ro~rWB6IiIJk28QsEab
zL1%?MdeR;-<S|C|OU@Nd-aHWR+__)IKYassb;l1fDll;3*BGLBwO+e^rQlM3@zkm1
z91>a^KRny2ktd(=bv5Ii0Z3e)jrc;_^chcZRJj%nT&Z)oT7Eza$YXFsI%R4fUK{0E
zk@E3QwmDRHW0rC4VsPL1GNviV8On^qfLQfW7qSYr7ftBCa}F^NSjMWzq3I8o{L~vQ
zz@Jg7Gc16swj~^STl_FXDZ0Xg>_WzYbhQm=cu}z0t{d*)Pt~{U8NM*_@Py_P@M$yk
zEP1bRf%3SGbBxC1IO%*RFTw$RZrfDGY_i(3lJU$-Z9Frk&<k6jf?Hc^4{c_fOL(Gu
zI1Kk1&l>k8{3|7P-X~e|z}I))6x`+E0T+WiXpFKL7erQ1hBRABqAlvgAgV2EVQJ#D
zp|uy#X#-;cJOGt1Uil#HqD^^qf{~FJJv4@WfB4`@@Vh-Y^-&oLBbds+gBrP2&FY%!
zH(mYt^S_1Hzbk)wcl7E!#p!r{zq)E@`@il!Yl~OL5C1nmQ7P&_6R@{w_k1W>O4@uf
z+A4j~g3<U<`S$pcPw55HH-|9hxic>)DvtjOgoLB<3J1pDpAY?8*%B(d76}LU;ZYwm
zK?gL|ZAy{<E3c#ZsfDt-O&izYny8+q-(EDGgrml2AFI&x$M|AQfe-2<Pk&WHeitg_
zze9LZw`5!8SfB!DdlwsvBk5VXZMu&BY1_W0_Ce>!Z}1a6g6~yt>61oTl}0{QsQ+FC
zRZ>5&7$ari%KNIb-&30;e~X)nuI78g(0z=a7~}X${of1A@qoPA&?|FQ8~fBt0aMGo
zjvG@S<VTkao;**07j*uK?``gA4}NRHl`?vvr-4<hZoh7K@LBZmFk>4p2v>~V8mAQ+
z_rE^8g6VAu-wf_=7OH)};yMoBZTNMz|FPlgI2?z!DP;UFMsT|RgZGBj2*`{$6eQC`
zO4z}mn><a<=0DpQ)`VD0N8*`SJFC2F0#jR#<&DW6LoFSFq!^1Z+LY%{P2&FY4Z~}4
zFJ>hvG1$dKBrgI7#)zVPVScfMjhUIzV3fpGSu%#Tv8{nHfrGIo{alm}AR<@+uZ&_T
zi?@d@*d9LE8=gFRK(kfv0d*>ZZwvNeB9bv!7{V}ZPapNFKXp*<QUufs7cY*6Ljokj
zQVZdXYR{ka!QjRC<=%`;3q%-ve*|a@7lDC7fWeO7#2{eXA6vdTufuyrG)Bo<)n^N%
zQ&!*h>c4Ko7Je`@gqw$nUxE_LmPSgs3}2YrYA@}S_Ez6wc+y^H@>@HVU>{g}`w82&
zbUiSO`o^&z>k+uOH}(eJaL3`h4;W^KT>V*F@T89kGTpZr?$H<>oT>fOBODYf4<8ds
z;Gh&pKbMf1dykz{HEtQYGTLjbVtmmB#^2{<ID1wK!8)%i_)8)2X#b$LVthaWnG0-I
zi`O=Pm4~fq7?x(su+(Rgoh|xvHiGquz)G=je$Ey(8S9A)t`ZE5nU`ZZsRu(Bqm&ii
zPoCKZ$e5*k!?+WiU(}fhyL${@84C}bEwNX)438<Eo+i(ip=Fe!v$GMPQf}2TV~e+(
zv966M8z~@dr(=9615H7z7@-dWQwdBN6jr#~8irECSoPM{l2-RqcrGq445w$TLN?x8
zxo{lb>hPTzzE)ROeh&_Ro9mm0uZue&q<CKySG?((;!SD3F6}Q6Dx0q#B(K`wX!-rF
zpYngj>3O$uB=DQ&Gfp2pI;e4e_3EYCV$9e^L!~sAnrBajKG`a8esN`(TfCZi|8mCj
z_2Jr$_lD)g*`&7(-D2q2Legc99L|Qm?hNbqZWjMyZdKli;o)%mvk%I^#QWm$gZ<%?
zk3I<fews2*4=Za|hP6vq0@#V+?%g}X{f)b|J=xl&cZcQG?+;5W?+h2$-W^sit_^du
zp`$ai$(K>BmG>`SSRAks+H5<y%mtx`!;2@w#@&yHKluHBJ^buve>?o-CqEtj_>cdm
z;lKTNe>D92fB(nB!T!U*b!zzd)0_GB<J#WMxnSCDZf;??eEG`o+rRUV%5y6mRd}mD
z>j`htiki^FltU49;lfgFJ;*DCf#%_XGfR2?EYx{Dyq3^lix+0johhS)whzHfKRix*
zt=D{2yv#_~@l57BG~Z%}gXC%M;-$8HVWoIUUQawXC}quw3`3X0-|{Ha>Xcg)?aH>*
zAI~F31@qhK^j&Rln0Cg^Y;5lg_wH^~c^5*vuU)@7T)B3s6!Lf|ycVOFix=AE)Qi!}
zA|!Ld-v0jZ(Z`<*pQJsVMaAR3jDGWT1se~qk3af!xO4kn%6M8zJM$wCk1JQ!3Qk5J
zigfK;i!~@G!_#SVn9Os+7KeB5-49QSFLcJz;&A=i)!->((Vfhd4@+siFn@kny+FCn
z8|z`x&h&X-S1%V|AaA@ggRA_U8)SRNgZ<3!$=f$xAK;o6S@E{QUs@1F31%Bh^)~mW
z+foj1$5Um*p`1JjzikL!Tx)|VrQhlB);qfg!v`OJJpBBlo9P$cT;}%1DLgZU|J}Qr
z!~S0Q*`zTqx1uymXvXfDw8z=A_|{&%Pl1a6=S8!UI`fRux9YgLwKLc@unix!_dK1k
z8$8V)Wqd|UEr9XuvzxawMz^ah-&_7^51~3o?5TW4Pl3Rh(OvUCxy6~hbsy!ZL@9o%
zl-CJpPAfa%G>`RXHI^5HAB==#q4o96;Lq)}NAk}Fe=Vq4y||q5cWT()*&OcPxmRt5
zu<(2^wiuM~9*nt+d-{xbAi9PJsfBMrdWK5!oiWEi1uwOWt@HNwcB^jF_Vr+7gu;KN
znpM8-_==wkPBEa`y0f<{v^^{%2_EW6!O~mUW(?xR!ZSX7`Z#@p=Q>|`qpvIEsS6hZ
z?`+|X`0#%#efH?#lY$dW)d#^z7bQAbtrlQ~zM`}8((Y)HJ|kB-KbZ`J*E@AG{9E9V
zU&*I%D)ont+V?_m##llZz`#)4UX5+g!zWJk3m?f??48Accuevf{L}y7VNha(t+VKY
zqvjK1Jp`}vpSDJ2z=OMtjtuy=U$g}v1E97BUkXk=+hFiG-^BHP+NDGLv?9Ib0QDy;
zf<r#$1K-qzXGd>2mwwDMys<4UJYc*k#dq>VqiZ2+^60tJqO-`PDVyv-9yD*2VKVX}
z`M_L)My64U_N%XLUyXZ<3dCy=+wiI$#$PGY(|^kD_MbwJo9QOJ&iD9z^peNx?@GRz
zevkT&DR`W|`Qgi9T<`jgPZi2IlDm3>6`uPv4skKZeA+0G&!)KG3=gZ;0q6^T24~Qe
zG);c1`?zo=ZBaQ)g;q!=HKuYqybigV!v86L!>${1CnJvET-6r&)Bcrm%u7HAy9VNV
z_PNv5^BA%68S}z3HQzMph<r;KV>_3gfxI{3oOk-0kq2!yp1rI1`J3w2&&u<yJt1ue
zi#9D~PKDPk$F_c@@q4U@AfNtkgx6dV+z;LIjIN@CWjK>8|MjiYd_GhwjCBr{S6$n9
zoLBu#yZT*nAbF8astv&MMl-b{LsHNCBfM4C*uD)x-lskgN6<F?xyu@VRGYb~4@P6w
zcun8RxShIIzvQ=XypP2N>tyuOe^oTx1X7g^Py77qGI|AWay`6#y`8@pMt~1L`aQk-
z)dwZL5v(!%G9l5f$KmY>zwZ9P!5kaDj>B<yOTx+7v%`Bo`mN#JAG|-D3nO|WOa!69
z$`li)a}V4ng^vjlqtFweiMS_Zlcsnk7K}hmy16JU@}dO7CzyOfW+Is+O}J*878XsM
zm}D!<GV!KdoQ*hyg=vM5E!YSwn0R+{ZIu(a#4RR40U<SL-NZ=9uMBjl8-;{=5G-sF
zv|roG?3ba6Vmyf|4WV6nfq-PfClo#0e^?3^!k>8OD6=oFmXLhfHiuwf_)C2#zX^W?
z;}?X?)U%9yX$$eH%>pP!kI-j3I2VDV;GQ`*EJpymbookoyH(t@H(|`G^7FM)sfB!L
z4p%xQZRgpF3<8Mu%j<$idzZ1KRup#oFf5*o5KEasfGvY#@)rn(6ii?&Wk%xLGKRs;
z_A7U9-W?w7MF`UL$KksVY4=(Q7pA?0-CUUMT9KT)@k0><0i0DsVQ26$6!F-CH@yv1
z?!hA-Q3PM((Q0&xr5f*vXM9h_VDC!69+f}5CLbs5b9qJ>m-mZ|sqz#}*^F|hBNzvP
z3bs1WD|j?B8(f}s2Fh%WPs&H*H6;c|mEB6ET1}m?!)Rs2I4}r3#*`J%>S>Eaif061
z8oxXn9H<SR?X9iy7$`cHI#3L{cu?4;#j4sgR*i=mpZ*e0`<>6b^OQ-H6(>r}t?{Ix
zltsBz0$<83nt9q&3%h6*g#iO3;bJD^m%xQqQ0hE;JPI|(;cX4yt;g553@zYJ_piFd
z`Kqhqjk%8ZZxZ1AFC%2sD0aGP6n%Yo1@r46^+_JS%IQ^Z{iYO5NsCVLEO5rkJJ+vg
zUSF)OPVpBMcZ}PVDij&EXFW5sko0FOA7wvzyi@Z&<tqAA2B8;^hON!}p<54Y-Y(A#
zTOvL>sO`HhU%gshR=f#*_P0MN1<vZFE5o_@OX0;XWbR)XE?!(37U$bzcji1plC7^o
z!%hS?G=6C1qqDq*0Uz>gxcTYNhQIykPlw&zouci#yIaGZTb~Yp{TF{Z{K=pE>F~ja
zKOa8*^wZilO}-2fyo~Tx6ujs*rS!${H_J=523#4gUcFYHN4vY*RVQsy<!hgjk10h|
z(lHKEaBAN+Fm04Vyu7v0@GbA|*P;Z<;|C9&pS3=$udi1;p300?yaVte=y1(9dBDTq
zZOm8bexD1q5E$@vdo^F2(5Hdx-FM%uZ3oeFi-0m&mM7BY=0+)AwT;D8yv@G%z4yup
zgs-}He|@-fx6kXMFyhT}>GGvvZgH;op`ZQi!<6M*tH))maPC-b??(}yI`a(V6#;H-
zX&&Y6F>N(CDLENnc<R_T4E|;O58r?Pd&A1|@~{yY@85?zw$E!N6a2FH!n6AEl-k(n
zbBTClU0l5|T)%#;`c0ahovl(bUR({I8h&+tj^e!+TJa#ea`j5hd$&GgoV!<^P58pO
zh2`9BEw@{WE^}Bb(^}c5YzsuhonR5C_|cTh3#iWw>a)<auW<%Xc|Tk9q8)j`ZES20
zAANFbxOZ=Jc>JV|8a#4pe5U>6*;_O6$iwH58@fH<19;}=&IRv!>vI7PNNll+^Mq_i
zoQ%t(ZgbO^VDRFd#*5`#c;4_lft5EP??8O-t=o6f{u?Q$DF!^S(9R+<a)7?qX9Z@;
zt9d?oGBWaJ=q;C>`MFT|2VNw~vvX#^SMp5S$+K@|_*$4RT(RKn?!Ap+J?-<9cepyG
z4SDtP*4YoP8&}(#yTOqsNl#9rSmx;#IFoC6<|N(9@`duMR1QOvzCg2z=A~Wj+%PS?
z+fBZE;Y;8tKED<yMQ)-bE#+`7u!5U0Q{Tf425tRj%T!|nZitINq?Fd*MFSp24yk_Y
ztq2R&jAMCNASAz1f~UGgS3_6oT)@!giI>D9PI4N)3x&szF&sH}&pQS&^wfCv%_6OU
zRO1x>!*x3cz;At{Op7xae;5#z%`l<P#+N(_&#R4FL0sFwr*3#ki!&`q#Q&A7nRh)d
z7zi20daH5y$*UF0@Zn^6xM@pxtsjl8C!r5}yS>OuJ@AmmE?ybk^WGM%+Blh2f50bv
zYb)eicGh?0xt0r!Fb^4n%CwURIJE8LO-7#39z{pZW2u8Q>LV{o^2z9oJ{kZ#d?AH-
zj4`Z?H7UQw=*Y)gB!Q-*M+|38^BCXSz*$>1K8oCsXYYGpxeKs)3`ZV?FQ4~2`2PB(
zU*7cIHKyvn{BBrY<s)e0!w-q)5pKfAlEbSF8$a7PmN<rRp0%-y?4^yxcQ=PJV8YFM
zUzsNRshHfQKYZyHa5`^4npPgdnSYa^qiA>V%w6B8|La#RoE-6{=5=&RKgw#%PZ*n^
zg2yf|WhL!lUMUNnOl21y26v4QIMnePZsP%cC<FOa?+UNl{&PWDDG;us-3<Bi?s+n8
zN7s~mwfEFd^(XHnFB%ef%Ihp+3j7vvGQR192?p0;RVB<1_4E#IlssG?`(3^DF7*;o
z`zZ&U+5^7?M!(GsU@gPDdEqb<d34*!OB;Q@?Wg?vx4L-*i!v*XEA2ZyAAL9FSMSo#
z_$)9o%%N-gN`Sj#m8%`d;@&xv7Co+GGtyqu;_Z$(0rhp^XuNNl<`)lBJwf-Y3R4-=
z^_vuaQJ9a`^H*`bUdAsY{Ic@-rr{|3u4BX3aX1cdNjP!(#ISbb+VF#a{5!)X2Z?7=
zewhi6V)k=PNjL{e226{|dgLUEDeDzw&j)7^m^e+q?t})Dx)sKGRg8@Zvy|A<hYrbK
z9#(DEw3LV@W~&e7p<K*I3l1S5BLk0w)ryt~#>E`vpXs?6m653|VZyiMSM^BU7{_cS
zLHIz7Fj$N>d%HWOgt9f8t<X&1;88b>kbLozCHO=LwUryD9&<`@zqoQCLh@3bO*9h*
z%xY8h&}NSwMvzGSXAz97<aB<R_o>au2%WCFU>qD2TB{B*6JY?AO(EZBVl0-hQY(g%
zzjPF8#h4{6SO}<K7QA=xt8<AkDC(<^Fnz6XX~luIQ@%ayYh`K}(wRKl&WYd{*}h;p
zh%n?(`eLlZBoVv_!d6soZ*C9kx9_L@y6ofdYYge05;9#G18@vuk4HB4!UmV|hhSYy
zcyQi1VRY3ksI>u=LdRJ1o05n@n9&-Oelkb_KW#r`3#Zz`Fqbpq_7m56UW{v$0mhEe
z2_{Og5~6}{@TZh0!Dj{_t7qXH92W=N#=A2*&<D7xPL!7LnsS12lkjYdAo$Evj1qY!
zX&Jx3UaPM|gD7>t+a6yBRmS;qvo$W9C*~}#+PWss<sFyr#zd_s%riq8;eZ0I6%$}W
z8=jTop_u+^dt2a8ME6Qnf><dy5}#3xu+pLgI%zd>E4*IV61D1@G34`ccpJlSh~aB|
z_$35WXooM)bYkuguc9(X|2p+AdigIyuIf`){Z03;_&XxhPnT4kWslRTUj1zfqoOeM
zV_F=rc5!vMdi824FwrFAoD$hWnVGW`L`dh;;dGvp*>JFTXZYz~|9Re{y(bfIH*|v0
zgwk+!Sh~Q2XCXYn&akz$p2Vkzm5Wz~nc2(3iBpTAUk>{}nRlz<Cze7-oN2>|A38)S
zG4$5QwJ0HUsSRAiu(SEm@RPs!pUaTbRed__Zm$pj`G59*AO6kn|F^^4dt1XNpWGh)
z;xGPc_~g@1OS!2|&ffaLZ~a!T_lNhs_s($X(xqDPf(F{E@CQG5fB3E6`ayXgQA#lu
znOAs$-2BW|bnC+&9%}EPl-#xo6i*y9-<hW70sIo<KDu2k7CLNOKfg;EcBZ#_n+$H^
z;@iqYV&3-FJU?tH*E}P7Pib6AuFMS-S3GI!jGgDUbAA+h-FzHAltF4?VQ$#i-WcxO
zzOAc5YoC=T4aE(m`}J$rhwI^o85}0XF2yW3(eqLiCmxt6d~V;lJN)c}k7^EFO5Q6M
zmI4QFrpGmBqWczSY~0@{zLU}H25+&;S5p_>R!@eV?Y&`Zd$-!}@uR0fhttDl+ef~0
zqYSUM5WaQub}6LsVP)hC?6xZG^UcCXW}aGHT(0@jyd<5iC(TpZetUbX=5+JRZ00=1
zNCvs$-!l#<p((^SHaCMKpH?5M2jkAm(dfk(V$hWGvgSl%oH5UtJd6|16Ti>SvH~9r
z3_rb??L^JX6q8_<*XGvtu)p`9&f>at>rS2JWjn#$-S88Sds}5e8F*<=cmU`BP-fP+
zHuk0~-|-jbQ|-V&WHmZDfRc4C_`{H9%f|Ke2LlSYWde6v8A}n#3w<Vda6S2*8gAdd
zoBrEM8GY89@7lnkuZfoip0XJ!$S&~YT-st~b)~k9Y%lR%P@^rJ|A!}{=x01bbIzT^
zPlw+)E1y$^gSJN2N6s<b-eK@by0BA~j>}m31CId)##_dMq7C7R&@%9$8Sn-EID72;
zT==YvCu703my_YaSkuN9_SG3zxfth5{z{#|ubseZ+>xEoLj2ccjFCt4G=094F%D-b
zdR^Y(6O~C8)Q2U<Bpn`AmxCGZfn|3mV`FQpj7`Qny23yOKN+n`-h>*}1}&$Vd)&!J
zJY(hCa{!vzyL>Q8S3d<O&?{R+8qXBK=oX$+b2wXUEp&s28eB1+@srwIebE6li<~51
zMo7;*x~fizV{1)ts<(bpZ@e>lhz5fjl@bTdscj*XRy=KvA0{sv8`@Hv^X}wPs_hsN
z_xJZ|Y?QZF+P==~GDhpNFk+O|jhoa(ut*Hd46{7Uc+0>oo@KpdVt&*Z%g9Pw6pp9w
z>wLhpUCHh#(|AI=o&-jC;LJwhAoC-+6Q2$L7(Ja=Hk+}<i;+AFuJU9H4Ob5Mjd~t#
zJif^B)pzyR0UU?n=ySs7=khNPjSruX`}%Kv{Pm&7VBIUAe#Ru3r;RS!s1_CGDg#&K
zWq04;yE=OIqx*==V(#qzD!C``D)JNxemVwylPjR3(_jY91pB1-sCbD~zjML0G8%g)
zugas_#GF^U7T+?4jEVfH{EFtd`bWL?rLKDJ^z~f%$hYA`7nECaA6bgtBk5`$Nd&T;
zIjR;g2G6I^d5HO{UPt?;__e@jp@H+>Z8d6~wK0)VF>N>Y2SDUhsr{E9#d{<z+Tdjy
z%=Fbb)<sLaPI>h5zU!!cE0S1&r^*@MwLZ(-(mE|NxcW(Je0zrmpl!ZYE@NKFtnS@M
z<3054uz%dg_?2e7>UaLeD)##$&tOAiI!)S8ztly~=9QFaY3BtGtIZQv`!IUZH}x4`
zEt@yqDc~0i1wviQI{N$Q)o^^<YYK1r_GQ;E8oo`M?<%}rx7V*Jyy@GUu3vR{y^Pn(
z|3$7T{dXB&hrvZsJ2reBhvV=Tgt_H~;kW<UKOJs-|9fS$d12BcxDb}YOqj@;@d^W)
zZ#5la+Va7~={|XaVY2NrkFTB%;-1d+lF!rUVQMnL6JSa)PcdQwEn%SZs8uJ)$M#^D
zF=b;wl5ORKd_4{WZSwU@V35}8Jj@~HQa+dw<(S}0U<=cMsVpNvp26rlh5EBF9QzI;
z-rpUbh1td=nfT?OnpgQ0BLqU*xRvt17^CMAQfG$6F#5c`79(WMFD#Z~gkZpcw!5<x
zp?N*R)n4lQG#Mtx)Cuocf-*)w1%dN)dDm#<v}w}r+mHXD!=Fz_P@S){!w9UlJye%R
zj~|o*jxbq_N!~S+9XP9}Qb*ogc_*k(NN_#lq4j~b2TUvb!}x?jJ0D?<7m9P!^o70K
zZGZD53=SoXegStG3BoKo!)bGUbJ$*YAa!Mb9Ddy){a`$lCs*)JKkE5OXeWf~EO`1d
z_-xg&F{A%X($X`Ab=+Q^MwwJ+&sCn!gWEfGNcG<Cb}1<K^4mK&Z!19xPq<ME<lvfo
z;S2%jMLpY|Gc+T(G!tRJltsCmnZu(?c?1SDqH&jCXl3ccRsaOHs#^(WV1H4{S;`FG
z<SicfPoFqdE5#q|?-z~njVJ5rGd#-}=L19VeQseP@6MLNiz2XSQiOr?X9HWtga)Em
zvpTu<z_;R_;p^$6<k{l_v6@!+5?0VaiW>qx+LO0UFC+X#tj)7Szt5I8Cr3DA%xd<0
zIu37N_|6SqX!g;+ZxRgQUrH!??!O*^Zw`U#bD#3Viz&(Bg}A)_?yIjaWaB^Kb1%Pm
z6%BhWOs`&jsfc6TzMI?7C*BRk-(-B-l8T3qZAsCi<}E^p9`Jsq+<h7v$kXB==`w7e
zKM4&wH>_M-3ynEB?CfmU{ERm^J(KIqg^csile3{Er(O(<VCUwS1A^@kDI-$uvx6$<
zWZ<YRUQe_)*^@`x!>6D8yl5FiSU@d4VtDyv_~U>72g4uy!M`3hx3-3#|NNujgP%DY
z>+bNgpM5ZFZSme=#CcqN6eAG&#@hnDT+Q#*+WM@w3QDdizt6>a{|E1v@s9z8*U|2N
zZxOqjdRchD5E1fT{2UmIhe#f1EM7_*ZQBMg0Lwr$zjK3Uw2=W2@3_0SUogQj3To#S
zUAuZU?Gt*Pa`cG>ROm6@(N=;CN)JLyd5l<`(MGo3dbj%Rlm$QLAKNviMaO=nr~xAw
zc*vNeD8(#%;&ElZMB}w*d23}3!vA^C8;h65h46Z5?K-E6ry5TY>G6Emu3jonDqdUb
z_qT=}?Nq$#v)b-a99~xD4vPOfckU#Qo$z?Jdh50;{L6%(j~Gqi5~G{Cn+Nf-(mVf&
z(%nK2{l_y&xi>$%nLgPWmKW#C2uBI4tm>!Co3<~$xO%Y`QrR+YJ-jJipMjGh=#!5>
zEu$mc$17Wi1)q2W!eNU)z~ycZGY-_Rw!b!3Qa{_UF*Mo=(fmuvkN+s+Q}M>9>nyW-
z_ckg|-rtm+7Nga9bx+fOk9$?Qq3KHZj=>AcaQJDQf+f6YV4dngi3cuosD3?tiWdqG
zo3rIPr9I(X@yyAul!qo^xSO;WE-cjcc0A=M?r-0^Q^vSbT!u%67YrP5#aJo2k<{o0
zSSQ0z)m0m&ZEQ2h1BPL^@HIFkKi~0W7Rtemnez-xd7nC6y?P~ZEEn~yGtw9%o;<Iy
z#rS5-_|s3P(uTz+1ptN;a7~Izd{HlCYA+tf$Jx1=jFHDRw#v{PzDT{a5u>EBp?z$l
z$G~2=mT&T(ye#F<Xo7DeUlpJ!uk&n=KQe^=V8qP7N`t<&(MWw~YkcZcIPa34><FKo
zb;n?b|72)04#;BaYmt;9m7`#IsNYrFS+MXAJ)!uo@tp6)e-$lAn}^=0ld)`!<U`?t
zdK-s~)Yq<E9ab`4cr&3jWGv-r1L@}@ml=<YP;ed{(nf;zRDWK6c*PwH!ZHT+zrL^_
z@o<YU{TTdZG(!h@q*<gT81w2`hLGfMzJa&+-dg+>8l>L%X*j7bv^hAmH#mCh%HGOc
zU26;d;CAH|{BTAaT6eH}P;>we)OPkG9cZJ4@ew=?t<jEXC)&T0_A{=)X^TgOZFRCs
zgnE}ZXv)xkb`i1ND8oD4p*S}OP^7C5_{ppC^R4<K^{5MSC3O30vFZQm-RHyW?`k}Z
zVfuFTo#&&!zcz>puDbfw>p_0dQ?{k=NPYAp`Ca=?@9MjEX*a(;ciM(Yo9HLc6YS`(
zXXB@J9!(PiRDaKNeb3#T*5j*nBMo!IJw3yHFT5KcD=*(<`6_;11K8L6sIu$cdDK1S
z80SS+leqIs`jlPedoeb3;cEVB;Vzua*le9emp#GzDrghmJ5TbNa?vr18_<?&%i#Ob
z@#U{z2waE#oNQ~{1b2;RZ)@%duEJg8y@Pg5kX|B>^)qcg!bx83KAGUn)sYV0flJ;`
zGj8is($u?*OA8~xI40^e^(<axGp0GxCY5(z6~XuMJ|;a3evY*+m=C`v^NIFW9Jm)+
zpe`1GfdOsxR$E#CfCkqYUk~$~J3P(5!~751_iI9fI@S5@3Wx3b4Z|-I_qz^nGi<-=
z@XN0Igco0RP4SL~W5d^RI1X<?I2GaP%KPsPzx~ht$*^|iN^Q|cA#JiN0ic*Nf<#aL
zm|mf_E64<hX(mvawg_4nWDL>7kd+}OaWQknc0}l6L=nftX!7*UBwpSWVFC!%gl^vn
zDgr@ZrWDFnm@!t_qOJ2_l*)(t11F}d;cplSpbT;q*W{hMIo59iX&sgvL3r!_`mndP
znMwR%#mD3m1g0rjdth>uk6|TvD8qJt^I_t7%UPjaTO#E>L2~{6y<y}2{RlA+is+?Q
zWzfRtOp`q(rdB{j087GVx(K-u2CVS2)f&O;qO&lPHw6N37lPch;*ug?TVtq7z{9wN
z3&B8&qaAo~mC`QHwF)ZldFe15fCFQU0j%?t5)ac=3bp(zueAt%U1l30F_&}aovC)J
zlu-BYZ4D20A62c6!&?;6A!XnUv&ujNnF$r;y_K<vTZK2qhM}RaOUTMS=wJByBmyXW
zvLa4ij7JX-f}eYPrOc@_@*Yt{=3U0pGX~=1Q3mqlLkKb^Yc+K6hSw7#6G6@PhO-fr
z(GP-fZOfRkj`kGrDc3VRqqO05NSRhR?)@nA5GKW;n4|z>XvzpJm}*<YFzB|~a+V+E
zmT$EUMoO`|xXzy#=?&UYTSR6Y6b$)CC@@CR5BZ-ze>P*GJs%kR(Zcz;R)T;{L(5&8
zmd97Vqi5~4&`LUrz81>Be!7JI-mW$6p1z{2Kr@fSTNS=z!<Ts{KaO7CAXJ*K4O52Z
z_38QbYr?<j_nQJ3U(ch{)vo~a&EWZFq4WFV)unal-)|13PVhZ+BM+V4bItL&@`tV%
z|7yj6&vWL=)k_)Q3mMP68yFid49gcXFE6c>Li_&x&6=YZmsi6RT+iIST1NF}PaX`;
zT&lBR%6p3u>(p@ajPpW6Yfhd9M#FI8Y|7gme)2beR&)#Pm^qWUrx-pS4FCSW`@`Y?
z_<#C84?p|aN5wBuj_z#TA8vm3@$kR=*`L*xe6w@Qp?mFNg=SGy&RY<0zPB;_$N%Ja
zhJW;rey6riM$d}2rJR&r3S)ekZL@SCCFW}Q4;}=()^~PyN+~*dKA>R?XpAol3ly5m
z!)o{-=lf9npFMxBX#R;4r-x^u<#-*+{FRjz6%L%d6&@5%<^A`*|K4!@`qhGk7lZb<
z*k^8bzF@MIpRL%;UDbtY19Ts6gQjPqDx<)eGbclzZGmW>vsmg`c^C0$+ASj`?*nIG
z1<vH(E6{l*oh=^8Sy<aU+r#?3`(<cja630MH(b7Wv6S$)Zr&<h)_es9o<P^GUaoVK
zoDT{6d6KBd_3(zj^}`>umsQ~G`RSk*iCBO_c|?h9!IAlcA?Ic2Kaaw#?aW<?bLv#)
zkSBdEn(Yg3*nTs8!vFxU+&}o>qv5kpKMR~^%iD|b6&&UkMj%`DS!HDr6=l`OAAL65
zxqYkVPm5D-es(K7>*lb$v@kepuG%5ID+8bYwmsVX!ujBAuMAf={4hkp=i@c;nLLTW
zrmX$FgMynURgI_gIo@Dv8}IO_3@W^ucor}m=tr<J3dma@3p@>~znxK+hcYk(ZpJiY
zlZTcq3T@q|T>bv?<+I@1$ufGW6C+c3rWLP{a_|wEU(LDl!cWYcJ3U;yXgkm4TKHqT
z&3o&;?V>&?BSQKgzkT+MGtQi+SDbx-L_$u~)_98<3+ies#rr%WEv!pA{C!m^FiL<g
za8@r~dvPgvJ~vx!#X!LDCh&Ngyq>^d%S53T6KR9=_hghJd)OM$cnw+>U!_e;R!e+a
zK;F1<EqHdm#+Naw4|v-d*LXAA>S>$l{AG&iDgXA|dQxqOk12y#<OTsB370J5fpcg|
z8JZ$DY2z~BC*ZYQ<_z?JG9II(4VwSc&XmD;R&tF6ci~Y3leV6X+`+4eflrxD3n-As
z#an@!5mp}>2k2DoJfMHmRxLL^3BFm_ccu*T+NXB6NM3k0#`kyMdnfolUo`f9<U{SR
zj50vsKhuS_@GL`}D{V0;xXD<ImHHblGCVWXON-yD^a`r>2tKB>@P7>2+MJP%{HK0+
zM7(q|uKLv+RekZSVAF5p(emIZJWl;mrnK75SZ5SPV{PU5VDvmgH_5uKjPI>s8Rv}G
zGQg*8$mQr!8NNzJJQF-!s_*)Q0grJ~VEogcXamD3PfmEk>kX{n=`$FY3NN2M?QIoZ
zn!;S=J$QMMF{*jrc-{5tQIj^C{=6I8<L?rtJiacU!+8Gc-u(fpbmJJO?_g{A`?sLS
z>q=a>?K(7W<)yx8BSOEEzVAs;KPsEtQ(m>ilzAJJ%thewY(67<>0>mYjT_yTu1f7b
zX5&_iF1$NQkK;b(JL0vt;&p=LllK+R2XDQrvKfKA>-S^y+`1ptQJCOTHRHs1sj-pj
zCZ{nG`cTmmx-H}(PvuuxRc7!rbO)ZAj~K4tyLN43jsPC;cHYv>oVAF=8PbtEGar;;
zxW;bli@;nd$29#5-r+y;&Xs3vSsyFT7%S=h`Ji6L7P`hL*H!4bOj*xWTDk%>U&g+v
zp04M<l&N3pcSY^8I&!sIe;cnagnB2}x^>&BLw>ZrPkH#Nmf3Zc`QUM2?yZ;^yBN{i
zgSTLZx9Xa>)i-0Pyxs)E82VF}{kDdqZTVH#uPVH$tY7px4sQ-`YO7z?^?Eoqd>x15
z@D_xX>z9Wg{wM!zc+c5hVMw24lENgM2y-Nqt9K?sOcq87gCW$4rF?7Vr5U7}jQ!==
zBx)kWY*bG9kI+?~Lt$zOQ3TeSTrp!Nb<q%u`q$-6%wJ9N`CTi<i-~FGs0?SAbcrjp
zOF5_r3er$AxK9iw*q=t|!C=`kjWUN&y1TV8JlfwkQ`d^hiP6AB6^y4M_yu0&Vt`Jj
zobp&vm)W^eELqV_2&K%}OnzHin}hQzB6VPo3#KyW1;DAk1ch^F&)0JaDvueYo(zv7
zI6NXKEG`YJ7cUJ<D=p{^=^Jfd3Wx9ur;8~}y$GZu<H$%sSR$l)#{ABl?v<T{dCGsn
zHe*QD2jfFfjet@{uhhL6Xe-ZOB(D9|ZEZpjp(r?2$_lHitkT+A-zvk`lSjdo<M8%{
z;G6T^7^w?Kf>-dz));vxBbcDS2@Cp}BCbYdcyYoj1B@}mDC<D;gZ;f)l?tDwan^`@
zo%2OeQY#B{x7`nqCk9;OCB&kPt22~!!N<9!g;JREGNnl2d1Y)6XyCAwsk{<w?SmFg
zTil>G##62CO9kP%75R%x%ayja(@MRq?B9)Gu8y@TJ@}8F)izJbmv9XZ+uYQ-21O@c
zgyHA)YWt(!dc}We7Mey$N8xesC<1$a8#{!t`B~?2%w!w~<{E#-Bjp=`O1hJS6~bi*
z%ow6bL}Oh%T+yn4RNfA=^Rrbpg(PG1ad>OOugCE9jY7>Tb$zomQ~a+8pZh$;e?3jl
zl}E#2tRt_*`PyqtP)~3A{uS4&!u%c|UkmwC^g;X@*Oj)vOrgeL?i9DYMR>9vWFB8y
zUZ`yV&z-YH*vhc7_Rg@l`pz)7aB;AG6D6;C(YC24PC9!l{L%^A+hmM~CqT1GNqmAw
zRp#+0+rwY|<zIwP*lxwc%jcPcPY(C*em4B8|Kk5W{K=pENo^VXDD?K;-CM&?fAUwu
zzxlU+F#L!A@TZ|wp<AINd%IiZbxVoOE8*#L^*ULe2Je0Edv&H2W%|R=L(0xeYZq(l
zLR*0ChxfU+ey<erJb~~Z6qY+X3@$sBUnyLZBO_8Rd_n8cz_cMl1!MngE{Y?(m$=fP
zp=BgUo@FG<{NWrJtz2Fb3k$_}6wIM-l%8vs*M=*(@LW9Po4&VUO?#D>#Dhn@m?CM-
zD|jN_Ar!}GE?#tH^}=v*bvf;03t4l{sbN3q(D+hbWv<Z2^K%OsV=HA$QqKIs>}Yss
z<v4}xUU))YPfwq`7}nO-lJ81+GOXXfpE}Q$p%ZLZFJCT$LY=D=Ud#N+7zzf8ulA5h
zy9Mr6E>aja9NNXa>|1+mSh+qb!OP%-|7|ZOTM6<6;(>@KyLIPIwTpJPC7-$ETD~)~
z;pO#*bo!KW0B;KyZSTnAi;-pZ!eZ5TH*qMs-hJo#aA7%fa<P_-2)6D$TWwQD*tD_s
z27DPE63-aIBQV&A=S;YawBRSB=Ckgn<RO2!@Zf=Sjqtz5h<0JbvWUfbMHX<GL-kk6
z${2Jv@0O7yZPIn)O?@WwE-$*-v<?3CbhW*C+5E}Cvl@I@yLh46w)kPVl6##o75qCB
zn9AGi?5SaSX+CvdDtI<GHUmFz6pP64KUxz7;rVj97QDb={4nF&QqnRuwlU#l(&CdI
z7!kphHfq9q<AfY!J3$5+xiYve1pm+xIK)^7j~Eqf%V=yeAT{hfZ8A2K(srcqwG2#9
zJAc{<&!Ha}V$q~h{HD!py?E{V)ue2`o`)QngD2-?*lD=hGe}w0w|P&Vcv?oUHo}ym
zIpr&_w(bK{;(K2_->L9bc_%ON;2)!*EfgtE%i~2}!K3cm(ep#wl;ZiRrpfp~Kh&#v
zB*w<Yw4*wbV}ukdo9c78tbP|SuGBVzwk<>(+5ixqD){d=nk6c@wU0h5LrrMLm1|eh
zm$SVvGCU->Z99)2Q^!L3gfd8^ed<h{f@h>N=?rDqev&Z>t{E#01DZB2L`;8rhquJb
z>ciTSG<61xdK*8=Kp%SU>T-JlWMtWt)A3rqR6l7-mVapxsqX59UcI<W9naJlWuUH|
z5%L}#Q3hk51q_T(%F+jD7=xsDzQad`{n}nL?ROz~#voR>lWz=i7B#42^Te440w;W1
z$T*NKgPlbW@WUbmIIMrvskVL1UAbVW`5nRoX1sYCe^roY1N_C%IMMl3Uay5Wz5B`_
zdZn2z|I53d4&Nzn{Ee1Zgd^$N2#lVEzKk@j>Q&(ObEg-i?Wc}o{>YO&>b>$iLK(Jl
zO}ORqT(U>{Tc3L_TApXy<<|S)F$0hV1ICvNtv5a^6kzg(hqM}>56Uk*t?%`wVe;Hb
z>kc=~gXR@D4WHrrWDxUiiu|hn3ED8B!DwEtWX-_nUW*!&uY8PY{0bVQojR9>r{R}g
z`&hs*S9y>D;q7DTMxs)2X>h6@6EvRP#{9-_6~AA1ALYUWc&C2oB$#}t_D|k|@!K#?
z`E~obO21a`a#z3UIhQc~HXlv#>hf*8rr>?u$N&AD;!iTGm&H3-kQMmQA~c&9rY)`U
z9q9K_L#0cfxQ>RYye|u1_x<mCnA-lUuH*2l3r(lq-0oj?ecNzs_&N^9;ny9)=)Ciz
z_lF<;lYcxcNBB1hKZ}rCZuccvwovPCmc~e(#!UH@I2ac~1_987ZNenB_>DR74pAXQ
zxHIBVT1!jFq?kF6(M=xm!mt+;kaQTLRz5@s$ook+3NKcKhaiqAX1K9M3OGTD3BufA
zA}GnH-vxEbRR>%DVaze>6gjo^M3|kYc|IA=dIobJ=G^%}Jm3f;RzX_bi6NwbpqwXg
z+M4iG>UTPE&qjzk7r{yUJPgynv$a`<fj-lVVW5pwwu*BwbbFK#fUFojRRU!jmm*k2
z@Uuk@2FVKZ<%<_1R4mq3l&5E!iPByeJ8e^EASJH6)fK<yy}rW;DUVPl5N@0~Rvtis
zN0S%ho4VD3>zLkClOW9)XTSfnPv9a1V~}~~)N1a)Z*}gyyZ468yW4@MiZ~8$X`q<m
zjYY|3Wt%Zt%yP#5*|Wj%;D|H22=0VTW32?F+~F6)5u+HqF>WyO0*1MSj^GYFn_=u4
zw_tMah?S*93n-p4P|sw1%?F>+frZ7T!Kz_+)@2cbtp2q<&hru!jBi4TK@PudCq<wR
z_ScrG$|smzNLd7PWt8$D@IJ^`q_m(cB{P)*C;3r&Qi?M6pFG{GzZvipk|<RViV@1q
zRKMOE6RmK|ILkX{7!XY<d6J~+4_3l_=R;X$tEIdjz<9E@D;0-u)~j2sWTw>OQG2?^
znRZ`Zyf7@DUnrP))iE$1hqoepXNIrUm6hLX+<O&jF7q!RK6m-6_kC5Q3hrNZeLfgF
zQ>d>~{^K<X*8Seiy@mHY)o){XS~&9gtK%Py`}$Si-23q`ydLjO*A$LCelf^e1aYM6
zx35CgKSF$Hn`p))<@nvqapu{zOP7XaTN^JgXD(h&T;2|uJ2RFS7oFc@K~Q+1Gre8l
zFgzXp@-O~$_?y4^ap=dH^6Gj1?9p(4{pRpTfB1*Pzx<c~YWUgDJ}kq;C!c;+XJ&o!
z$tS~~{n?)nAAkJGU>m#x#;D!BI!|h2b8EP`da2Lhn&ZtuadWl|Srm+=R0-V67nX<L
z`mNteyvwE3{`lih%Miox!AM~7f-Mj~Nf{Q|)S`mS^R}HTMQz@d;wbcpM=!6pR&In|
z=1-Xv(C!x0ls9a?QyMvg$6_K|VA+z<s%stx@+gmy%pdrW{R3yK?w9e4Cjldcc|$z#
z$-C%p{=}t}qKI1vAEfS-bJa;9$rS8XQsY^8dGL}C5|&Y{`;gbNt#B7g;X9u><idpu
zl~->K*PdJIP)4SwFNVF$9e5se+9Ds@rtzlYeM1?}5W2Fm93J*$^0GDFO4@F3`0TUW
z!=3f@wBbSVuXpd<9`4?|S4I%@9{&l^81yE^9}iz;JLhRRd0o21Yi_mLz(OTxi5IoF
zg8{`Hb@SHk+R{=R`<@xS3>~(yRE{kQDaUzoY;9~6jxeY=AC`fW@og<ImvMxrS@_&G
zmMmAf<w=+JHeSr#44p#p|GBhHl^dL3Y@!(FrNrn~o=|Cbc!XEw5r+rp#bOLUz4+!u
zWDb4C`w)*y5zI(XUMI<&=MtV24iv9CcDXdX`s@S&E{3J@7)$=(YY!Y-5zb^TT_}Fl
z*eF~Ju0F^ePU#=41)R8?HlJVMJ$EW|_`{5~?erBxPI!#e)ffQ}<6glOIN&Y(DV)+T
zX$M<H!UM{GacnucvbvnM4;*P1b==7~GX@~<+}unV8sQ>C9^OV@nDZH5;Vu{$2N(r-
z2wJd{R;V_`H%&$q?SwDKQ|VXrrkuoIXx~d~Yh~<!zw(9WywbqiTe!by8G(FI>6*6c
ztLn3~fpV0A=i@CY@J`Zaa9BR#JbKt$oYwY^DGM%8oRZyIG0hm>Y3j_f{CU<Vc%o@h
zm4!rPl*Ofmw9|5p0r+W-VdTM*wQ;OH1S!$sWpA5(ws_T+FME9B7a2LxOWVK7TRx0J
zaCvzt<0RMC&Q5J}X&XLiOP)-f+cVMbCZqCRf2AB{I4i6?RTG~zLS1WNLGp$7El+lO
z{2-&QXK;2qGh)5+2C4o{+n{^q*{WN`X=9*@I<pIH^BaFE4u!Ra8^(I`)MRP-_7;%n
zptj_lhz@8kIBxrIG*=sYhvqh4Y~dGqoe^awb=)snKP{T8EhUqW^Cio05L{8`GUlcH
zrg@=ZX#@F-Yn<?G11p1zzSC~{-`FH?plLp68}RghMe?V|ge0t+JKlW!oiIJSzyAK{
z^{s=Xxn$9Ky~?{j_lv&tXxhTPpF57At-3yAJu7(X`g9hZuELl$UEg<c<)Nd;fJRzG
zQiiP1;#!cH_J!}h=>vTaw<f&Q$M_VB{ur+HM@k=aY#7l5uj-yqKZ`EEQX+R~xL+l2
z_^MVJ`-i`C8JG1X&z)cLsw8<=d4Qt-#H**gQ-`WY#*#XMozbRYJXuEJGWJ%!c^5cq
zEXzN?#l<6_1?1k{&_y^8o&skqC%3xL(SpURy1#<s=<kYOQTrQNu;^ORm>!?5M5sDW
z?J~vJKb=>P->JWf9#8k_lGoASuH+>@p+Fw5{?KhI*2Jg3>Vs|{ZK-TD-B{c|U=Vc1
zJ07lSO<&Cm=uY{Jr5uLO!Ssgf?{j!leP6$h!>>1dmwj+-_&N^9;ny5ymga_c-hY3%
za^rfWw=e>k=m>-+!b!mC$#0y>@*e|=hhnJm4Z~y-#l)Nm<AEWzB?-m4v;+dvIp)Jv
z+0;ayJi>fo=n3W~_FC2Jeaexpj4FhpF!<%3pZC($wtJ5$=dCQWb;xKKGl?f3ldyP%
z3Ce4E3zH!aZ1z$vFBM7{MycdU;4EWY1jci-Gv$TC(1a0mzRzNW*2M)%{0PBHvqH@l
zX{pC-g!+q@Y=bnPdOoWyiZ(ayhk@E3o)Z89pY}SV{b;fSs}*;(QaQ|}_NJ)E(h=Yu
zMj+UaKtO;zpFGzttqs?&UmuntKw}3f!krsJz+%9{bsP$exvIYsT0zlFmcG(1%Bn3y
zv~9{OqZcDXzF~altl*Ad0$yAYL0+At)6V+d`2+gHYQj?|&eXx*+xPcgaazaWtq#TH
z5`qcz8Mnq;t=JE4+hWLAH8$Z!i5Z1QFN?|5?}QAJr}6IDwlY@!TZIdUO1YZnxr`a(
zyvAhmWAGy2FDxw%OG_(dY$aeDr<8=~4;bKV3GKm20mDy0g3+dUR4#l*W2}-icAOoy
zyttIQ^}*$g+4t_=sj-BH?C<Q<&?C&LL>b=}mWHL}mD-|48Ss}(0ABf@Kj$o!-pUAF
zq_`-Lxo4Cu&uVO=Gn5__#DpYp?(A+CjP{L}X}1C8$;lG>8S5z-ltD4VFoy<(%4M7-
zf6BJ}oy_=hHkoj+XPca3<JVgdzB9wu#nsg>7`}ur2u~I1wTRz<7Qd9F{yQ38PdI(|
zKF`YRr_aGqZx8R{cPuzv<GA9v)yFZ;=R+wu#?Yy!Fg^A6!#u`NWxCgU<={^!3z}b<
z$~}5{S+VgiojCXWg~k+5^U_Mz9+MQ+tq^%I*c$JhcW#vD#ObrfH!mj&%k5!%^X{;>
zx0U&MA4`>K`RQ=$vk!;=^Z(`lI{dqT_x~&Y=|$$}JGVa@{_@ZNZ1}@J{O^aK{KpT5
z`|I1eHiuicZkKVZyim@}q>QzqQRhN4mP4nksAsTe^tu$Dz?od&!52`zoBPr7>*48~
zP2(IV%I&{R8S=Khj=C|-QEuDD5C72Ho$)^9QEM>~kARt?=M+h$&`vrF1<)*X+O}X8
z0O8*(T2L?Lqk}~wLkFePr?cl~+gNq36>G}j$+EM%Uvbg@#qdy+(Y$<$&ZSe#CE5XB
zH!0j1P<Vy$yrR^@%hXo8p`$#8C~_~Zt<@Hjl<t&SXe;jm`7(q!Z;OWl1ICF=Wbk1z
za|gP-?aZ!&N5jVE&T!}My~>}lVsX)maND^(3*5VDrw8R3w6=C3d|bng{_pM`3^(uG
zDQ_xX%-YHPP+kvd6NU`9WI+ex3?qbvM5XA;wYV^szPntWd0^FsykAQ3zqn9~vfwJC
z3Qs}#nXmAZJSi@&t)`F84x5|7$IN-=DP?J&=Jis>()=(NoH9>qL-RL7Re5xs8x5Fg
zQ%X@V@^sO+71E!@XTzgh_*cB6K3xpY{O-H&mI2U06ue`dUlu-HTN`J{jk7HSSkkl-
z5D&1jvRdV<54@nr=kYc_+unT2M@LQv{(u~~oQKqgEep2rtKyx`1$p3tx`7uS;qPmn
zO}_Y&gTRHyF2ixI<qHeLd+%MZx^LAva=ae*1CQ;%&AYuI&*DGEk5WVi9@`Gu_E<TL
zwYChjttLLB3=;|Rc6J_bH2w7Wk#p06L&2@tbEh-5`#iRVj6?M}d7>8+K7PzA%;G#_
ziuZeD1LHIK2F&VrW#G@yx>I&_NIe-V%NUrpF+R7qHk0ScGK3k2jKqvl8B)XU&Oy;0
zURJgcG~S8_NuA+4I!dM!v>m?BqCj|IEHiTAd3yWN;6@VFSyGRl9x6(CX;Pl<rf+MT
z)O;`4lOH-eZJ|$2(#}0!v=;<p5_+ZGlx4eFMtx(}c-5zN1<+1#q<Gull)M=ZonuxW
zP{B33SEv`7Hal;Z4dbB~Fd8dpB3gL)%H{F^F@`?=<kO-f#t>YA+m}Os(FE;Mo}w9d
zjF0FL{AJ`}=vRIju~HfQG(6NVcy6#4PLFM@!um1bCya{hd{l2dv;M7xjx{G`3<Vb3
zilWB`mNBW>4i)6pZ(<f|R$FQV9cbahg_ZEW!O7-9(MH}eyv15J78utohCttCj{gOt
zvf6`<vAY*klHbuOa;Z5$=xr=Bc9RcT8(hY#oes=X?WO!Eb63+L3*C%i@S;EF!K$pb
zI8v7i{c{WzsUH)jxI(q#k!yPQU4$y<$n~aTKM#xN#Pk33I(N^;#8Cr>ww#O>ty4G}
zAcS-bLM_L+rudQwQ(pSQ;tu)<&*Jigf``VFxmRDqziR(6z0i0kANO4KvObOXS3hcO
z27l-g%opy?a-NLem3gAeulpIU<wgJSp{h21O4Je8Q!g+ZgX&>HWtEpdMz5j^g|`xp
zV`FSfIE{{3@NElZM%(g4O_^0M04cEGa&?>4Z-3)>liy(<c{j!v%!BG$^iEb4&&M%>
zQ2CDE#<Y3Lz5B59^V{1oNnPIc^XuVgJdq0Sf}qDR`Zwj7|LTjn7`N`|fb;$=hI;s*
zJ$2EDsjHgS=2AjeEdRX<UEV+3@DIKIQ<!k<n_NfZAASC|*PHS<8eaeXs|?46uj6nW
zevQH8eDTJ$;ho=le^?EZ&wD8pz=!%yh{uRvD6G&j$(gik%aMG?M0hA+z4CimOo!Ev
zl)Qx5X26@Ml%hAn4#rkKEl5!oSutW28iqf!X=YDC8-bj{fIw8tOBnAm4uwgw67azT
z`=uY$BwS2Em}M+QZ_n03V5arTm%!&8=Bm;J78PRf3OiEYd)4%rVJ>A7VhG^#R*B}h
z1naYeYloiq_KfEvAYQnzI$T&?8O~s|!pPSO*R-LtkG8ir2V22~D>!T;ZKqt!qZhVU
zDQ~o-DJHAhD(S=UTPaOZF}E;3tX;kwLG0?VxV#+U%f9JbwRM{_B5LJO(i2eBhj2#8
zQ$cb0tA5N6%$`8NrOeTu`CR?Q=&2nkRnDcZlor~q=Y$C0grVf^9YIW8#mmp(;r@f+
z?(OyAV5henIS#)eA$?~Y8-H*FUSoza>BdAGy%}g*SQp?HBa(OojPii2ts){g@rukl
zMkGck-d^YdCBVsyui1<xc;h@0Mlvf)d0y4%O!>xgDdLhR;fOK-ehGvV;}*WdKZ*_C
zjmuD!!prmUGUK-tp}|RK<n3&23@<2OBlsAP7<yYSO7Bp0hH7VxDTi<m9>!ckkO2gK
zE3Z@C8VhJsZ}rngYU8f9JIbGWQjE#l_B-lN@oD8W<wC8rO!*Yd>e7Z}iXBo!>Uc77
zPRtB@TYJO$?R&%az0KiX?wj{F2Ir*MQbxNRhhKa6&J16d*4ApPLl^vOCcU3O7p8%d
z07v{p`@Q|D%S!Re{wqVz1(l+KI7|Yb!)t=;wfy?M|EJ_eydSyO@BVJiuDLvyJ7dT+
zx7Y7_7PpmzovzdO!}xsoEzecSSHHb_^_+ZsuOMCj-*xqXAIaPr+T@inY<^t1uv|tZ
z1{g}ufR*o$h7UjZ>F^)^!+#il{No=FpM83Bc=X`$u)VQ4{7--Kr^8?U)nAtak<$0h
z-8;jb+jmNtIy=8OtgNh+5sN3;l`B_jo4M7M)v7Dn!-MAHMP6o0rJ(U19L<~X%4x+T
zuLz18!PzzzvpjtCusjwRNYENIvlQ9kd(bX?2L&!e8f7xiRdgL)oSU5=7H!`gUI@*z
z$Vgw*;+W7d^i7@QJ)QGN`HVJdFv_Q;#ig_pk6wl^JQHQ>Vi|e_id(S0bL0B(-n;LF
zwx24aA=;Q8$~<J@1(+=^U;yhms8*?m-s4l9Q)=$u8E>mW=U-9aInU8~Ugg;yx-K62
z%roIe`p<#}M#hgm`Y82yP~I>Qf<f;7`o`eAHpV6Gjb>98E1Rc-IQXsdZVLSB3x63&
zC~J{vbE>v`|NZX=t{cTK>MO?6gXF=}0bW&GXHHZO6%u2Q?F`TB+i~J)!z2&4_B=8d
zFkJA`VNj57oiDjCSLO2Xl0QR>?b})@*9IYcWx<wmYc9eUCoYfY)FS<VK7IM9yv5ou
zHI$)Fz00dE^Ca(Z3R2%Jf~xjOU3uCu3@}7drXM69-e~ftaJ1DKMWptzu+8|>Zj3R;
zo_6I0Lb;EpVAy5IG+)|6O_|=i!yO(y#-RmkwNNWVmcjM;%Tbw%w+WDV8S^}xg^%9t
zrQNg<54*Xvw{0%rr*w>$6n?eko~_$bZ{B6_Q!h2&PEL9y=YX+rF?Vyht>VZ<WFGB7
zjsT-Pc@pw`(f|vK#)d5p86Ay_rnj{%UD~>~94&>TGgZhy&GT2h^drMD!w$Z~q9Ae@
z+;oRejB#Kx26$4Efy(Hm^M+G}=RDN)sq>eNJN4y>r<|e}sb9rU$Du35`vvDFnrO?r
z>f_*Rd5vY9ps&8mLw(8^mG|oJj_*{9WMAV_pg68H+EYK4VJPE2WwvL^$>3S|$IudZ
zL{6%+u+pB4F4{wS?To%r+-m<;=9@Qofb!1k-3E*$MlZpbfyY-v8|wVQ%#A$$?uJG%
zX5kHacCMkvk^97zXDxzC{nXi2vS`|gVW<`an2%GAFwH$+oBC{$h3hYER<djIRo`hI
zR41_@xP?}DU~>U}8g7?IOM<k88}+a66OTX1*}m&D9$&@}dc+&F3|!>L)KeUb9Q22U
z1Mslq|Kty5c`}yQ;?lr4dEK?aPdO=itcM@QVR0jQ?k)12!^gPA$PipjKJXPy&@bA{
z`0Fhr=@Z~b<59Q6>!H8bQ()EvM?<yA7>+)Fo6z}x+dQN@dUby4H9dbm@6q^jc)f~a
zjJTscvhq!+IQ>}P;i!J9`1R6%X-93D%g>^xnb+NYGe5vvcXMm$UV>wAAc@Aj`maD&
zXjNf^*LyGOUH`71i7e0QE<fcIFA}^14H>WFbUaBrA9wk7KHUz%d5fZrlfYDTKJ{oB
zaV!Jw2PczTXOlDyW9~^~y!vHVoAOKz?VI8#V2U@j-}rj`mh`#fsPsb_y^h2ejEjP0
ztVF}zG*#qEQ+cUNcwFE7`Bw4Jgf64gh~Kob;^w;-m4|Z2XWysix+>4<GB%afJT=4V
z7^A+qo}|qV!q1qeEMVYi{ED9GlCDpFra$}!hV*LE&;QWFKlC<u9Us2#dOhCjzrW2j
zg`>ZZexKg`D#GjK92>rl!*TdEhEow%ufPA^a3eywLxM9Z+o(jS!mEa1$P8DLYOnHY
zW+c;Pdstxd!%$G>)`Xuw!U@KsR$OKxEe1I=qO_DIrC?9~1hparHA!RETe(Cy>lLO}
zD)!bheiLM^uC;1ZTHYL1=2_`y)}5GzVv+(j2Kwk~w`rvtK?-wUD~$r1`Vbz@<(f$v
zt8%U0wEC37o5FxHzZfITK-HBX5<xiv;?m-Bo$F-_m;g}1$CC&_4kX{--K}lbuxSil
zl;b#q@-Rz1{G60|>X_WquDxAPohK4O3ln<z+V$be^&5lpz4VJin0L3gBjoiqL{-;h
zCSQBcyJJe4nQDfl`=;AN8;e(4*5$eClRm4yqvR*h<~L!n6L-A`C{{}n_z6w(bBn{d
zGv`VvaQpV{;r{K-;o0NS;C38-qrx;Mo2d=*1b5(NkLnY3?wj#M_$B<f;DxbaTnLQ9
zHRdu72<l}xD<R7H36E+_@m7%Ld@lL7QmqBzQpyGw;TC)(1O<>1n8Cz=B;Lf#6PCen
zI&_SHq`ZdtWeM+;unb*~4{XJfdJ%k*moWv;o$qBuFQZ<MyWW0^(nh(IfV>{yJEcR!
zCj+TqYr2>EpPeZq8$%$m27OU?ZDt%(TBCW!4tmB&r;Us`ih??H`Dw~i<co~6C(nod
zo&Dj~$2W&he)h?5=d-)R-P?DE&CT9+j3VqfycOX)GkpE%hd=tgS|uL3Tmrod4xedn
zYPiWWa>7#={G#YE{wmH_T=HVfl}BCmd%Q9qP6{?<*X7y!@iI?JV_eu$$$f&ucr1P(
zu$ZH#_h9u+x~XpMZ9<{B&<gxHTaz&e%-1+&oN;$iIG&raC0?$KG2^Nga9#I9<(~mt
zqfzK*b7PG+&*NRFb2jR!X<q$}Mwfibz;AD8;~1@>s6k^a!a3O69&X>dIsDNd{rAJ~
z|C>J;KK=Oiu)k-#z9@w?`n9mQFkHHPbvQrkY!q8foe$q-JGq%sip<TOA1+_M*w>{?
zbtX|OfSy(VsQ=82?N}d|qSrj{TqJZ-{>|GsFNM*7a`a4S#{RIixmjB@QcUA{w3F@m
z7|a;O@Lt!$V>t6iKcG*_;n9T->MKh7M^BuE>O4c9{mn1nKejeDYk`42u~=yCJfm0}
z<t|=Y8!oP03Ou~Bo+f|%MH{>*((p&lPqPi3ZQhoa`rMZ~htV994oMS+mYQ=~sl2_l
zQ`<JyLV$-1ZEa{UuUx)-sg%-au=7<nHt&~r&I+%riz~yz!rWNa@-TOf(eBak>1VeR
zcQgGF9LiiGKmBU1u-Id%Jeb>?2R^_P>BOO=WiWK6r1~*J?QHLs;j0vR&s(AEj7vN-
z{?vG2WW$GQ@72|dWq|T~XLl#%*m=OhFXNo@K7C91n|3K4IsME?aqr&7aBuy78BZCh
z;H9ks8OEMGc~pkA=ix=Qcb##T9BkcYt2}LJzJmk$2_4wk+Z(pF_mV$l<cVQxXS?QH
zZLpWTD4|WsR>GHZJne&LXb+9{3bWPU?5y)`AD0nAA2WPu7l+|Ldis2DZciJ|jM=jV
zN0JX~tE)!baOck5!m*VL!H48``N~=uB;`c`3$G7(@z9XQyZLHM3SRs{D^Qy!fJ3#|
zC6|7o#51=u>gg-vH_6fmeP)!co~3(u(DH@;LNxFx43TwCSMcca<;#KRdDR7MU_uY|
z+d|;6Z6D($-k$-fWuX&Q_HNqQwxxI&o?`N2%tOD>82292&&rExEBt;fj!FCI{}k4X
zWP;oJL4V+%&=5FYiwpCfEMxrkd7~7|@(7=mI^x05TXf;>-TV38-jyXIq(0i^P`R9P
zCUWXlIGyzR0`B-&TO&SfrKoX=Hkd!iaOl|c7ruE<E=(rr7=2ae37I!e<0nI38UKPy
z^3kr+nqywRBnLjLv9TDOHGlAYt1Zb=cXZE|d(8*-8O7*=G8vESJgw7=Dz*ycS$cNt
zJF-~wlW>s{b#Y;_+J1dKW9;sF84BSM4@2kfoew?K5a_n_+uK{EFsB^msek5d^Tl=c
zR`S8OmK>Z5JlYp*$~AVhpLwh2%)p;_rRdJxZ!%up(@yYfG4w(odPiO}ZqUnQ1I&V}
zg3c216YW4tywi5@jS(16xw)}jThmHc{TrN_&f^0!oSSHnGT^cW2fh8I@p7X0Q{HXr
ztN-OS@#=yxATKc6URFKOh?>jNCe?4jX?y?!zxJsua-V2Fcn%jUtFH}(v#-US;P|T0
z>E3+3DZDA(bbZ?}r9GO@^jufH>35}fneNlOwBuo{+)5s!e<<FJtmfs2A7h{^FqVx!
zI0#4EFl{T%rg4Wx4>(fg;;|}q;kPo~{d*PaO??BK{;zkU`L8n=-%VHRWn9y5nMg0E
zF?QwIAy3nIN&P0=#-9{K$wOU?Z#1dk(N0O@UPhzDZ2qaYca?t`pT!x=Xk4xzGN`N4
z`^(+;iW4~co08NsS7=$+rCxQqDSobwe*~tEc9<W-UcVjd?Xlv?qn{6h&+<zD9L5{t
zPEn@?t_5@2%sXwP-@84?zvR~5irN;n4A90kd>AvXb`tzQR_!>v7OH>$0S7oO{G#nJ
z#s6is-_d+LAH9AR;mh*&=6I|ehvV>T3Nwr6hkyE?{%?jK{gZz(EG#b#55pKZ`^r`s
z7+&WM*aD~ozl>JfWe^517%j+T0?Y`ecsi9yzKk%GU=c8-b3WC?z%ZO(a%F@VP%yh&
z;aJ;#g^4zynyk-+aUsOohK2A*5#a##Ud@LIbHHZ*R#r{Yft}z`Ml8y(2p*>*Q1!NH
zy@eVWFh!<olXz{fnBP_(GfMH$s4djOATqE~8aaOqOnbY#Rd>Ro{3*vBcK#x{QNCgR
z_XCeJEnb{>F+7S8w!N`FJaFzp%H_4@tgsTso;(_!J$_WMQWQ~GU^WR1_QNM+*(#+E
z3%)qK_x=xtm9>i{;OK|vd4F)QSIUTIPaVii@fG19?Kf@vs7)xR(#V0a4GKJIFkgg;
z$6;umL}>K81nrcM0XiF@o;S|SOzM=f^flvzws4rcHdfz-*|`XZOX;&0!}^`|;iI2@
zJZ#?C83M*}_ze#5vlQwPJgvH%l-eir4qh2E@b*NI0}~H-FwYoZ3^OHWkPX_^_^#~(
zGsc{YL)l6}RVzP}wv`{rBl#JZe#(c(7{;6+a<bwu%0+OA;9SB;@+Wx1b>rW*OjZy}
zD?TF^O*Mn;&gRCjmkTYc)vNxJ{`|uHuzYcKIAiQ4j~6Fwx0-Z9#x{it<)(45bYZc!
zgqgNf;<aGRojZHFwqUw*=T^qZ(;CN*iiXSt;+MnyjA2Se^+c~$F03TYxuTnBWhr-`
zW^4|pGZvrJA^2POH;3Dw-5l=US|1KJtu`M8wOq#HvEl2j3A*Gv3IB)x-Tyx50uS0a
zZ2XiG-Q1diG~Fo*r$B%`8qoPSh0Mk^=IhCz9fSCOw($RIamLYFw;(^2YsId)o?sl<
zMrFmJ^1>gTTe6^hnevt|`eOd{+y=2WUV*L3IZ{@KzSqzEhi@>qPi5Ekdf^8sX((>c
z5p!!Rp26`7o<LvOP2Z*pwgQ5!%kMy9Yqjtb&K0y3Sw>>6XSF@q_rL$WVR4!FfNh?h
zmI9G69i3Rc*k=||@>5>l-`EIkd78QT<naFY-y7a}=bbVHeER8U;c@nAj<yBeg{3|p
zX*2O{yQeKEskI%CuI-)eg1HpWp<h79pHS@cvcaQVxpJx6b9>8HU7v-PJTGM?Pl>g)
z)zR4UAaPEH2XemCLMgAewl>Sa#XzvLb5MrQ<qLD6SAC#4dObfoUtYP|_8{=%O>E7F
zm%?#Sn)CRbTU<^a`@`S-)lUcBLaVDwftf-&?H8CW4uE$&*~)7t?MwOO{49!@^2`a%
zTVKClifxJTb7-}+=&^IP%II;D_gA0Y>g+0tR@<@7hgZ3N<?3)TJe0P%clU0!6$8$#
z+joagZ{93}?R(#QXSkGhVF+;c(fM<;<$=e*iXYoZyD;LwciS5>Jem*nAp_>UJ9mb4
zXLwQM1`b|q=FeI|9{ilPvMn#3z^^aYRxg$knDXBCV2m%$*5VDZ6dce-yy<MW#^XwU
zc(Yr#?+y3wcN-Zu0$y@;b)^)1&!6k3(`5u<C@xRrjH~O{->JFs$A9zVg7M<oN^p0f
z3<0+G`|H2@Nnp`;nd8%s%lT&eN&GMzlLuoFJTXQn<ZF>s%732rX6TupqqKffhThrp
z^C|n;aBqEM*x1-EB{ujN5UyOlG_0iE^tY`TuUuNIx${2n!SL&suUstnFRZMlonH)}
z1&258Z)YwI9#T#QAO;d!wZR?m82iS$exaC$Tlf}dh+esLB{+RhXXhbA`jKZD@^|{o
z$<%eR&UG^488L6&zFqvdt<Ke*kr*Fu!P|VwyO_RXFuIxfngNy~e}2xc50Mc97x-+K
zhc>OR-z%7%<;(cO@KN$m;LxW?tG;L4u7zx&e~dnObzY0k+at?(Uj~T8VI)BJ8NF;J
zsqJh5xDdI{_*Fl6`SRtn!Y%ZWK>=;Ch5Lf@P1Dam``OQj4}bn?;xXK{!3}XN(wCk^
zv(XV3x>-i8@K?qmUXWoG9x~7|$oAPUZQ!zfovk_HJ7cVV!OIx~_!)INcitkT)5T9M
zhYpbkd3ot)GFi<RsW(qbd0xA=He9{3mO4Kkws!Wa4#pO*Eaw!G*Su2!{Fu7H6CQ;W
z;%G7YsH1lyuaVi%QMghxCIO!ck2be<(%xwc9i%;y*X()c3(f}bmV*Zi^;&z^nQq%{
zsIkCHzp4ZL!&8&NrZzS9;1vFuahu%TN^Y*hiP8}(Z*Wb9#$JfoMtCwY+OIC!otzA3
z&_*~{bDeoDd1yE9w7uVOY&N(G-@V`5*esaJFp~1L6E91&58l8-3pmi4f~jzVL8XmY
z@ZkRajq(mrFLh>UD_qWHZZ4iAAB`Ui6oS0L%l6iT$MQC}@WjKTuP48o;e+=Nc<XAj
zi67|&uCHH5gEn&=4X^)xQ}`xvgsHr58ee!l&Fe4=NAvSM{dOzEj(aZewO3zxpEfzX
z=aQ<Dj`1yaDR1H6d6nxDA$M(WEY|{p^o?<Do<vuSbr*wf<8p6*PhZh`psoj6Q-8S^
z{YdEA^*hganEEZz^tEqeJ)7sOkh{7_Q${Tx#tL;l9i}!E7hLLzhT#W{dFjl#WJUY}
zT!)v6ueM7X*;-UU-m{RQ;b?=9e9M4TZzrBuKLm=r_Ztl2TkH+rEf%Q{1!mV*`}8&O
znH{9-y7#kur>|9p2j5-NO=<leTf*h7YXr5TRTN3a9PtufRS+>hs_ZduGP0Kz#ykss
z-Vd)+bS3i?o&h~XBZ|MTn4MdL)NsS4w>eC0eY6k0DopW@!<&Qg@eetC8~ix-_&N^9
z;nx%>yB0644)6c=Zx1VLYZ3fyPe`F0CV&Dd)7a?@YRtrmTxK_d8!rWe9TRKfvjt)^
z1|zdj%urA6C8Q<3$sUtoa-^`sh?pfO#?ANg+)A2S4W4(!03<F(g`ta}NEoV>a!+|!
zSk2jlmtq8SMfuW#M6Xl_OMprl&A=4H79o9Nq%cXdVFpX_p18`kP1>Z8F|VrwL7gHz
zsY+;K1W^BHf!lWq9`C_V!T&UI_abN>M1X%1X7Odxt4p0DlDd0V7A040<CuI1B(=&C
zV-^O3FrYmbFRTn}5uRrPR|j=Xdxc5k%|g&=A@hjxE<&jn+Q{^71#7L4kATEe#Y*4`
z#ki&35ppR1ZFOm@6ACe2H`+<xXn)KP#%F3b1_vunt!}iM_uj4h!;=TcR<Gaq5QNd^
zwXzE?W^9$BgRBs=BD~wWicmp$NXUZwrD%X`#!PUUk!kY8I5h*WC}<g;dKGRNvgDg2
z)0Rn$OvXi_aQ;JI7?d#D@|`YXRneF#3$5re#7Hhb0ve^~lSdDS-EHSqJgGCX(0~$p
zl13TM)+km#cx)Wlf~jP4^$&w@{1^j-Lggk(DJj%NgMg92xyDn{@H`WE0T7CeO<N`!
zGumw?^(PpVf+640t<x_<OP;+Pe){8|4nO(JpA0uYyg97jxj!82J{X=qQJbSd10099
zBz$KcU;o?x^gsPQV~H?t&Opm87~w(57(z*hW;0wkD~MskeZlX=@jia@+c`#*ORj~w
zkKc}5@{p%=lp(c1Bjr*`_Pwv#wse-qMSE&lDKB42dl#e6^3rlCU>P^+nc{USKk<2Y
zG02ohq&eJtnlXeX9K9&lj$RbKHQ$F8P3<C|ZkNEm(A)IcZj5ryc61k(SL%$L={IBa
zg%$5s%JYVCWc6aM+%K-K4r`usU%lwMIIQGaUcS(9&o8FDx#7&r>Duo2(i$(V%-7;2
z4THkMBE|iMf)R0|e6k&!ZPgf`cp==nb*r}3tF5B^g|^g|JDJ~k0O3VSQJUe8rZ^Mj
zLFVI9KA#Cc68gp9g=b(yK+Dwc>Xj?y!NQQUaeuSs?04RIcX;o;cWe7(d7)omM$>jP
zk8j-Hs0B5Y{OB5G(Al%6hWWq^c3T8qO#LWa8MnlH9Da<~(^~jRZ4F*4h0mpJz~+ps
zpZ@L7ieAEjXU`s&(cgJX#ZQGcqW!$7dcgrtxo6eK&KsgMXN+UyQiguPzqX>vVucem
zHsGu}F%~ium`k)nof{QcC~=3Bv3`H6;A2o+zrSAd2_8{-j2RTX3}^aLd)b~BOpGSH
zqMSWdb5kY@N@1SE6w{0@79`bqS()STN$9@#w#{QS#9x9F&k5&v*}+|kIsI;-gl)jC
z^76lOwfdA%l!DmYrQVMp_ZeX)PMj_ynXUPZ1M?(a5RWGqGqsgj@~70QIq!jmRS)Vc
zBV*S&oA3e;Yijg<XZIj;X3x9kS3G(xh!L;mHoW=EdN%eM4Yhln<CQk>trUi-dzEPc
zO7f&c^WA)nHMhWw@|`!K@xV}*88vNSG0MsG(LtWUKR<l-*`49`?K_okDKGH^>4Vuh
zhLzrmocCiX<x;5mds?CXB<%rbuBWZF!`-`gOOac~kF@!Pv>`=2nghoB_qT_w&7EO4
zb1>rtd`FkG$<y-A2E+68&EwwUGkugMN<Yq2JGH!HTSeYs=kvW+&A-fi?>F9wk+5;@
zOr=wIZ9-<jW8i)s96TC!GX_fmo%+bT+CFVW$!%+Pc!W0Fe(%yHv}Yx4@U#pu$*p)W
zZSpK_hDKk$bfx-@hhTa8rtOU*-ji)yfIk)v1=qB2{wtX*`NMI~cpG&zM)e2WlYcKX
z=zbBJmZ=ljq!r2~E2bXmN*-hIN;9Ng`Yb^-oBUS!jk2Tio;}gt4{MAX)B1&{Bg2&T
zVQhYwHr75A-{Qd`!kqS0clp<xl{w7!fL3`<+dRVWG`)vAoA2Yl-P$9l)`(hN_NI1F
zRn&}8dsBPY-bK+8v=w{TDoUtL><VIULJ@luv131Z|32U6f4Gn1zOVawo#$CJNRO3`
zqjw4Vsp;_O`aA6Ma=39O`h|p*c2slmP1k^<!f(UFu`h9VA{=~CXWIha`2ed@IpeCV
zLuW|1@_h0)ji2?ar4x&78(J@c{IlAh27k!{tS7{hK+{kG8ue{mkN%ma<f1f67YBu*
zhVT-|_X4=d4f8@2W6>Wl()eexHyvJH0SbCe9u{BjJOiUOMSaiuX)(r4?mbx~D0bM%
z7BBf~(E%8BIjL{EFpw5*L-m@c+(MX$6MfudIniFprQW{e-O93fvVV#Ti8E2OzrCJ8
z|0hYQsh7P35wg5l?BZC;L<sw@t7bN&S`IKeL>(Hug@SldLR4REcS065Bn@)5saAol
zUyCmZ+X%uSiZ@QYdH6WNZFuI1qQ|c;?<7Cku|@;Y!)Mg*1_<q{G*3GQJ`%i3*5p=s
z!*YCh<shxsXHysM>=ZLe81ibQuII&o24?bgHFdtuW&T(>yjGXqs~$9HaTKQLCqH%i
zs&@-=eR`E3GNn?BC$}yd3a;iJxMZKi-s?4tc0XNA7T(dg`)Pk#x*5!KGYy0fD}w`k
zaRgrQ4ez)~dz~LjL0Wz9TMPUU?{Y;RWi98m)*aJ2o4wy_YK)L>DhH3h3%EF*Y;Oma
znY;IkMu~?6-DIrXgpgV2vmS6P+$dJI3!pW8uGnW9NdNDOU86JoUFuSc5bmIS<&!`w
zwD~@bh-bL?ONNJDgyt>D5s;*l+W|r}!b|41NcEW=F;#2W1=c7`+pzORIYYE&A8{6U
zd2_kBDt{A>L;pN_D4H5TgD+E&@PZ`0khHNbSsCcNt+58oR>|WE=WZ{72)?oOAqkq|
z@~D4C;Wf!~gBLng3UMBJ=Hd8FFr!YDGbxn$NtQR>{VsjMQe#@JpclkM^Jj+g*e6{(
z`)AW0Rjo2`U{r%Xh~VrcSwWWG_ntLkep7`Wl@<`Wmm{qh6*c11dU6~a#=#<FIjaH(
zdFvqL8J>}{shFx06V-iUQ8!<-I&18kk3PJHh(OHs7onO%Op($Q(3aoqRJliIT~E3r
z3rml`vlz)&BtF|J=5s1<PAW;7nXn4737UmB=mFL|u|W?78xcUX?CY=94%b}@NZVDi
zHY+GsR2CJn1rUr*Xw;t<iBMTkt<ooZ3xJtIG(YhnX7sBB3(sR;v&8A^%rH=V87Qv-
zp8X)FE^~I_uuxnRzeKmZ6v&Ysk?5B|{pHIls^VBkzAqineS?}7OfSut8Xg}oZAHG-
z01n6rt(EiFC6AisD(aZf%7$bKUHn|;wDoX!&+rXDi2QPIOeZ$2qf{n-WP)P^0DkuF
zpUXc$gItTuHC7-Lv~aV$h;dlDzl!i6d;3-w_c|B(Z@j*Avw632z8v^%XZd&aWVtq%
zbl@<rt=>Mt-F8#uGRt-521s?Jt-V<@cjo}pFwu&%(RuMs(&u#HfSa&>^YCiS8*@_y
z@=_bBbFC`#LzXonRCyTI<f*l?!teX81)42;FqCVgWbij}ha44H7H;J)6hnCm40D1Q
zi^d!bl)4WyhI%>-l-drNx=dSJS1yECvdAgE<z7sN<PUEcGgO+kg08oe+eJ;=YQ_HT
z?*nrGUjKrAU3Mr~31+Z%Tl#&IUopi(ykl;w;Td?a*lOixFI^i+tUYo^d9_)gMed=)
zv0j=K;O`d(gT}&@{h9Mp;S_%qIk<FT7A<0?Z40VyIY4-;K7S=8j@hP=tJYHD+YHS=
z+z6THUTX4!!PPjuPvh(F$rZh%B6iKcD~=NCaevY3TQX}~HKwn>6Lz$1U)sBhk#FfH
zmb>zpU7S~w<^>p3{vveC*&Kc*WuP=VU{E>f`?)@JLJ72TIu}LG6IoHnwIC)Y?zD*R
z*$5VG#1s1UTgka|jh9o2&^4yF_Cc%_g-p?wec|%oIxA<T*rQrDrl}=V%AsbEFs_~&
zc;0$2)oA(FOEayegER68^L4eQf+SABae$7zSy}DqNFx!)6xMgyy{N;FkZljm*4{;W
zoFb=C;}3HkmG9eR=vcJpj4(xtcb6X$iCD}bzCMe)KaBJ~E^b38>nI(wh$Vd5sE*m8
z2NmBI4cae!io<#bi$}63q}4k<A6{L~Q{-A*p@Q2Uc{LdJc)U-mBz*DOyw8HkTt{@X
z$a2vQ^#BC2WRd+{agd~H8IIikBjh%p=fRE@O+5^#M|3pJ;dUlodhUzX@XGxKb1Jv}
zcQ-?GdfrGo39IV}UCTemT@(^FR*!G-%&h;}y>Ib5a6@bk{b3|K_m7*idmyti^&!lL
zXZ17Wb@Y7;M5xENe`xhO<GhsmC}Kpcq7?$nP?Hokmc2ut^9V<kLvUYep|PEj#3l>^
zM~f*zeC2PVx!a=UBgVyU(tWM}aGGjnfBNCMZSJp4ei|WwAUIGU^0iU8g}@-c0KYLb
zxv%R#&O@%=mj?MA5c-;=%J2u<zj%YIk1ZdcI-crUMvtt~dR$Tv16v6&&eEsvUulnI
zJSq#jYPqaOyP%}nb}2$5RX_-7i2{HGftX3t0DUfzU4V&TwadVH;k<R>-#Gyg`Y(7m
z2|QK`wE}fkr+e8uW#vUVqzGBO<i$CF^xxC3?v57?^r!dUb$0-pAEGFfr&wj^m-OH+
zD4MJ_-NwJaf4UpOQzGNQL%Bi)7X3PZSsmN|R0!kL>g90wf$(ErSJH+tYmeProTC?_
zy9f)}e<JD4aiYMe&SD1rm-M0e7SR$m?nr*3kV(rjEBV3`GPAQeSqEL<PRiS-qZG{h
z>^=`(oyMb%QyZnWugI6aZ1SHy;lZxQO*;^uI0#9Q^=3ucgTS45AJYd#V;BCCYoI7u
zVu~lbrh{`)Z@zAhjk=yq&4`n><J_c=@5;NzK*fQrTowmAuD4#WrPZk;c+g^t-=I_2
z&8Jj!6FKpzM!$KsUXi$YCQH1@E)mrLCb<y-sopag&~E+sfSDKk<f3UgxKfNa-lw0?
z&s3v5!)*!+4t2$nGN*kkz1-}SyV-;7eAb7^!-Z)69ckvcPKKYa9@3VM6v_#w#%q}P
zxWo4$#_jK=cf<n2%a7E|bz8(huu$+wz)#Sp+qtDrES?u4uZ|BenZLab;LZC1QiEV=
zAJcSSgj*N(m^^=VqrLPyX(&>;1&$P1FKw@`fF1Eh9Sp_It6mu~6j61NL&tb-AhOff
zC?k${8&=X~%b}-i`8(H)z))TEZP};<W@DH21GaJp#>8~_zrc!xLYhei>*5U>S9zR)
zC9hn10=)s+du$*lleql8S3NHu$U5n1;ld&?<DtMEZ>}sMS#}McmDtiQ!E`Mzjijcj
zYFn)V@ke|c_eI#;2Ytzn`$ZmA`9D*M`-01?ZboL4>uQj`vh2RLi4~Hxf1a1}P{89t
z<35Zjhc@jRn#Kd6kmm)+-LC^yL+`&C!F$U6OrukQ%eJUkLZ954gjL4>-5|7CF8-r>
zz7GF#)34!0LkofR>F(@S!885!=))1#;_&>wX?J_CY;qA}Rpg{%_94$FS09ubYBQj^
z3jVm$ny;(Nl+?(wpbE7#&xKYP<^b0oR@M6l&=Cn7aMhIQ>?8|whIJf1i9oCp4-+}E
zi}uHLPWVCJq^(plyjoD#H-Aw7>(vb93KS2w;z<-hemo5-J=S-x_`r0_66MWN+)5Bd
zA_w_x3mi_1I;K&8vS1>#=@zr4-KF_o!)XgSR54+fzYJ=+wSm^QPl-CxA~TB!VAXQm
zB4Ck9l<etyF4v&>ADi8wTuI<m3`GvzPhql%I}c+9p!Yn{W7fpT8U!1dCpPYh7r{1%
z?Mbf5N`;V#Pg!y(7d=ySd^NC4>r}rIKAEU&5LCl!V8+nL#YcL|H<-;OMX2;L{zISt
zQ!U)@M6k~=P=2NV*y?x@f!Q9BW01XFUbWg?^j)koq!zrfxO<(j{j&To&S^vY-$cC~
zyLop_n+u?RvgLnF081bnjry~?p1E9lSC~`<yWDd}^7@UiJsBvlUX%L1>Ailp@BzH=
zkO}@(`L9&&wy-!bH+x6WIBL7;s60mKVC1ItQr3*HW>Z@TZ4jyS<w2UN{8+3WwNo3+
zZ+TC0X<@hWrubk>bMvX&rmdE5@Ii(+T18N91nA}Ie~`Uf=Sm|G4wk*}u8K(eoqdR(
zjcRPyBjPvbEh3QhDXvecQO#M2`ab9D0)2;lX3}0N%Jxh^5~;1Kxv4tmn}8SF@R*9{
zr2lpLijWRn(LS=QCP$MhGt<QOv(3L(NrSs9XADL`ig9BS9#`?KU&X;ZwF%-Z{+!F}
z;Oer98@pFW_<$nbvTJa$xpbXq{{C3#@ZHHxHluj-`QnxNa)?;)<=l`k1Sb%rOof?q
zjdV#j8~}1dE{(?Y$=o~y3HGCx@;^^g4C~ZZ$+v_P!boI~9|q^urqgVIMFc)h(8bjg
zmOF<D2(E@03(1UlRQ-NAYm8A^Ughe0`<_Qq;?QXM8dH!sa2)+3F<({ACYRGTVBNKl
zkE<piE=`yBU9;zEHq`S6s%v_*K1wVVT?D+xa1Rrvm9%+_P&(=Z2}|QBwPV{$`|{yt
zYtzn1VND?xZ4y?dE%F>bfGlX^v2vV1W@bfbS(C()%MjOA7?G5`w049s#FkMVA?P|e
z-*woL!}a`g{pw*n@#bFxui$ZGCOw?>K1gjWg9BfL(2=h;!|IJGW0tmoQq$yb2zR>u
zy2S@Y55mD5!qqYw7=tc?fu+qs>&R#Lp@J=!t1gSb@$J^3YG=R1D|!3<-kJ`+;Mk^F
z_%JZ4z0Dc0FUJW^o~oNVLVPNL-_-*NChRYX+fszM!;<WCJx2ry-(UBB3HGC$L!neL
z=fUU)tEsZ1N*I!btp|!QuVdic+K@+S+vQS}d51XJj{b^o%gc~uTHVPC6hUs`9!BP)
z^(1<#?`M(wx$n8CSN6b810V9x3X#<1PxY7>WmNT1Gho=NF1h6FxyV%gfup@WwUb<a
z{hUVA5n&t%`MDF5bSA#&TwZ;A`Q3m-K>Wsw{R2JzWiWbeuC~A(vEp)>FxKS2CNb`Z
zkO&s4kLT-5tM=ig;(AH0De{V&5rCnSA@|{j|0C1|jESf?2ceP}q=}<AP`F1l=bpQ(
zXg@ivIiuVbpADp5TOfYM>&*yN(ox#={I&a*0K7lf-fxd!JO1>V6Q6~p8li#uCe+_W
z(#Wf^YJ}usdTWt-jT;}2K-3&+3C&PBLKSpggvy*{bzjCGgBRolqe1>D4b3K>@TfY?
z<5U#}g3F_X8AMQ8!6AGWuJ=k#Q>)cO#YXZb=n!*V=l2)B344yo+JG~=eS#$~;H~?^
zZ00DAQ$l>hsVN{`??vjoNs!`ZVYWw@)46z5*^GE&8-H+;gf3QbC3FHKl_fL0XiA4l
z6*&DH7Q(p8{rXnf0$tt1d`T|b8HD8<`jmT_4=Nl-c;IDE9mMx$)1yAwF&KdC0tlx-
znTLhBs7Lf6Y1me=h+TIt?TKB4<I6ebV8zWS>Xc@^t<byIe_K#(lj1fV*smJ>AGZvS
z&KHACl->H?g(T6K?y$LTrwVVV=4jkYtWQH&3{@Nr=i1wpp!{5d0_a|x29Xh(hN(L6
zn`{Mjni?FX8r)J|j2?z0d#)kFlJ$j7XYce(Xc>kMDj69b4&yot4F2f}<!x*eEmzrc
z>)*`cWdde(>Oqqe{Gubo^o55S=@WMp2m*^6iR+O6>?>e@-Mhxy{1^*6#$dhq%lMlA
zbD}zvlpD-M|Bvaes045b?lOEu>uIEmHVMDZ{*o8nKUdu)@hI1oB94?w*Yb7GE1H6&
zelfWuVRcU+Nvu;1H#?qgxK(PRitOB{YO+WgJ5rH|{*I1JLya7E^uu3AnH}A;3J&;*
z{?l1n*OqMzd9l=_k!H~~f;ce(fC=w;*x;KALEiAX)nw%_*ix$Smb8IuZCMQyan$bz
z3pA)j>YcqvH<P3!0ZqV2_m+0yNAUct@t-njRY=TLzz8Qj)(L#))X+hfk2U=yjz@6+
z3nn4`$$}8?!_R`2z<P3h`=v%9oME&4_8E|;+X+)r8bqEq1l78<d>{8^3ZeoY<xi{x
zIX{ok=lV?Y@Jo4oCk3Ut=N@D3nY(T1XRmSR=g}Hu4**}KCdsfQov5$NcFF%Par<j@
zPDLu&<SOQzuh53HtnM-~@6ppHh#}N^`?!~iHT5?hg(R2TR=6jJW%qFgl+O4;^ViKe
zugzWc_6ZIbJ}HB0GJzzywvaB}+MfJoVfoBuMnwX2T*Qt`nvc=U&Rj=Cv0;Eoxzls#
zoIdm+<5$I!@AeP;(k#rmsUp{L0C<z}aMXiB8xQvD?i;|8ShM{ggLBbaQxB4xIduQ=
zT5fxKP5tW-otLWF1E1{|9AfS%=W*o&viiGc|1Qn3P9H2VPP_l*Tc6q)cZzR1S_Wio
z!IaA9{T(+~G5S#lgP7CyTl)5^psG&RtVhdno_eWrRJ3g@+UAK*N4jPL_r+mHnGrgV
z>8W5(%KX{ji6gWN2{gj%0a*SextNjFM1hox)BN#w&CxL2AlnCdt&iMohI-bHCd@hR
z=CfF=AML=b6V{>6Ubrs|Rc{Qb(HlgQ55%3ef}Dl#GFkZyJH@qL>W&|kRh_QJ`&Gt7
z58{rxa5GiqMFBXo8~oTU;2Ns)?Br9!7gt-?_D$aOGhTnD+l@iX7Dc3+$)w}~xy6s1
zWj+`LcUi`;^NZIWrUh7?`}G@kteTX{C$QexQV!K>;!JJbJdKyr%7Jq=W1j3pA}W{K
z<y_d^l@f}gdSV-zT6I%h8=!T~U4jiXT{qT-nyp;*0YqV!+ZAZ{u(Qr8`cR^<9~AXs
z;xW<<`_SK+KPW96q<i=FDlkStQO`H>l*98O!r`Rt?}Fd9IeIPn0wk6v%<jt>91GV+
zbmI$4t+Zo=;*RSH*25Kq<aC1{z<#!z@9XDEX5y-2`<2J<b5pF)S4%;#jgt9>100_9
zCCMFGFxm>tp~$N1*<3mD*geZeZ_3>@Zb{i*A(Y(hN$vU4x-3xj^9%}mjx!+C<>j=;
z@kYhM%)mr^$)k}Hj>KFJ0k{arO)*_ImE(5MB<@9Ri0gCJF$Pdi!OFx*_N_GpPXL@F
z<;CMA)i&31)c4ky2!)~)xbYF&OZ~zv!N8;%w#`_TB2>Pk<k;siMgDVRJPb!2j4!$N
ziIx~uMTLp72dUd?@@C{y(wv1%yX4YOfh27uZE`AVhYsNZ(31F(_ltGO%nXrTS=3+e
zf3m_~qyuL!(RFlE6TL85MURbzmmA>j)(b3ahPDDPqphf6{Ow;7pR_M02ZYQ;A6@f4
zM_!<VQ3;39V{tQS$8rH7$;yy=Qa7m<@I2-*9!~*7cUqvF#^2afvaGr&>lGtc7?Yp_
zKCZEzZnLmof#Tm&l-2GPTL!E<l5RqbFjC}{FgCZK#IvRX1WUnrckQ>s@YH9KUu8z0
zlO6QWeu=T_D61Ah22U+v%WCr!gL_s9(HL9QTPAZK%s@)crP>n87T-;MKKMua1RmbH
zV4y`)VBO^sR*7efpwTQ?Zzq3<;V15D8+?Zmq8f&@-$3#)dXFHePh&Ok+e4TBBq7F;
z=$tNkSDeaKtiaXMnMq35)jTe1+yK`VM)EIGYeY9CQL8^c2}>qAh)ofCaF#%Pkt)@t
zccPMUU1q<>PX5wJPRv!TWNQZVlXT<V#T7+Nt#Y<W+LFJzYflG(--$r5#x!k*QB*un
zF=FOa%Y&fuXjx4TIMF&}5@b_PpL56|b$Q?-M)U_k3jY=QEk{1Lq*jWdz%@8Y&_{g%
zMJDGzX7f{~6DYqNd7+_|^SdKz+TZTg{33udW;%aK;JJIbT3yiVk@Z}vi3S+C1^q$h
z#<5wHg2xBq%%ymh5a#;+$$^zgiw62E?gAx`apuZSju8E^-n$e0o3|5x&A$RROGp6Y
zbZ-huHx%N8w!Ft%MbAzJsOHZA84h~x@(a;sAHE0tXsG?AILrmkjWp_wRJqW-(>&F?
zEf2Cl*tYh2F-~@*y*(WX;~=Xvep$+5V`?)MWQ*nV-PSr<boXM+;0Rk?(+V9Bxt4qR
zStYal1I7aHC*#p|ZmKSmtvCF3Z{Or_sA)}3=6tH(G}!}i8uZV=AX)FuC&NvADKn*!
zz&&0q<mAU?Nqo-hytl0ptob1u<eTg(wEm$DW99VqV63pKilZ`g*!o}0<(<SecQ?;9
z_G|_^;)$bYZ0q`G)>p6o|5L38cBAgR-C#3l=uMtrYJfq|4|_7_e0%SJgnN)Vp&U|`
zI-UKmLQeRqLgZEer-)xFQ|-5|S8_b;MdDwKLn0-Q%y2s9jyTUwUJe1W6P11z$};=?
zPdb+9+M^$7vGCqLYJx1nS6dtMif=mmekxZydp`QEnZ(knEg9bz8HDG<Rei+YfIPU+
z?Yf>&-yZsBGh5$K?^MUUPHPou{qZ|tlDL;$@ux&{4WQlAz4e5#7re)UR+tQv+6v@9
z!JuJeDM?sA<F66H*wVOq4TWN`UCaPA!3@8dX0~1TM~Pp5xi=$g?FqoHFlPLA-*M=A
zslgsi^b4GHr}Tl1RY1#&k6*^BE?)$EgrWYayVw0_U$af-VDEFPB=*U>DX{iFnK0UX
zN$7S{vd6R$92Vl`e>B>3PjAo@dj&kHh|4p_rdH?P3Owa73_g0p_(5oZEsMA$OHEmu
zouwdAS}7AxNlMiO?|~v$6gl{(j=tRPF~Kh2<FbEzDdhf=)8MC2t$$guPwXJbfyT{N
z#c!Rv-_Yf~=!wVDpvNrhH%%88=&qd}bl1_5;3J7&!FG-0)SQncD)!lUWCE^#BGkxu
z13hCqeD_`Ajb0P6IdyO~znatIVL$#6C5+$P`z-&y`UtBG^Kh1g2t4>GxG`+g4Jjvm
zjk#f^Xg{96^EYh4{Si-+REX*Q$yXWshb{Bnfdo7R-7>#}@e4aYCu8`qe)psGbj1CD
z2lDp&yg)R6cAnQhRF_Lx;pXzXJi*}ZGStw+R`O=+ULsXD`W%uNI)Uy>j7FE^+rDX#
z&oLW^xPHI}0V=c@!XWu=mx<pMGK3V(VBZr)l73hXuJ!rBzrS^WBp*PButc5=>7y?|
z_(j#8w*^bL+8Yb$=;7?M*bgc1T*LhKoUjpI>-n7@m5c^-Kg%OxopPoirsG-(k>SGc
znrKkSSM3<BI>JNW+%SbEZ(2D|&mySaN{(beU?b2F_mf&!B6H}>7rEz^#zb^i3xaZf
zKJ@qhX-x1jtS@S~zz@O$VB0F#yT=g0SwGFscEFRWR@&WxZ-*$%%>Ap%_}9J1_ZA}%
z6eP{n60bijouQF)G~-s`F*Ylp){n&2eiiFfSOGhU2Ce8i;F8R3ar6;Wd;VSKlQx+`
zv6mQyfQ{-PNjUBEf|uvh2dPK&VLR};mWBhGC7UQ|e9A<z3yk#&?T6is3iKjwG{dRO
ze~rs`Wwj<>+Oy=(uDkYP^JNHpCdFw!!y#N-L#{aiMA{e~(4LFr4Bzp5UsC4rWcE{0
zk#|E}sN~)h+P%wK+V8~M^D$9z0Nh5@1lRY_T7k|eS(es>j8oDCuZyOV5?M6+Zd<!P
z@3*|K%qxYSLtpd7b+NCrE@HH?GOqeWDc3>-LLGvE9Yc<B633wd7=q}G%n7^Z$?Gx`
zaEv>j`_E;AfJ_3}Yp0?0a5M8u#bBT5O-wXgXsSmt2vcwo<<`)jN#N#`*}@>5X~G!3
zZiPPTfe<2Qg97djarJK#lZ^$j52tPqJ6liUZ==Kk0z2J#!ET7NwBdyi?Q7A{bZvXF
zz4r^4wf8Tcc>$l}vhY1AVRSvq*&>Td__e|(4C(?0b;lyMpJ(S5N76<NNw2q&9b-uK
zJ>@T7iiZeU1N@vGg~ZR4x`K4U{#@>>to*C(9g4o=xhN{PL-tAbwfl>VC}$@*Pw4iw
zuD_}hrVRjcFhxFbPG*Ao4QH)h6uaba1)q9dN;G!P>PIJbn6Jut0%QFP!kZl4ZW%E|
z9#NR3AX3Wf2|<j_D*RIjMn;qHZL8D71Db7HjC!V%pox8#{c6^~@uT4j;mZ7YK-XGv
zKfZ1&rgCRIIPX$hJ*GsB0q_%3S-5tu<aR3hqVizULM9Nxq5rHi{JV^(f=7wWov;#~
z^Wr<XAB<m&mY^oC1sWp05P_Fj@M87Va5xVoDhkZs{gCP5!_v)W{zs;wu=Uo%a)F|T
zd-xg(4P%@(Ml3s}(TJ~qFeM%ocbT=eyrN5N9UKSa#-gEF{ZVdepNGq++!e7esJur_
zOJ~MzqRVCec37F7T%S11k~0VC6Wy}V$7}AK?xc=--O$6_8E^R=+xdeh;8WGRo-gAA
z$(46g?GF;|c+7gXYcEX&a|Vl0M(v$2r+zNYeNmRkazx_+X1-}~1(PD$;BpW>!zq<1
zLB?-7KykSq{_6)XCSxF5u7B371@ZnDUdlJ-8wV7gMeoaG?H?y$K61u0Kf-idcj{U^
zRqn=)--it(!%&{Wv;GF_Kie<Y1BBa*{)F$Rev?}*r*O10UT55dC}{AMT+B}WHovd)
zAgIhDz+l^}^B|%mr}J^3&Vi{zAY;-ECJ6cG_|H;;o&+GYd6bLlvSMwcwDMy2H=n5;
z#`^2Oz?KtuaL>D#4Yoa!wVw$X@@@NLAWOMpi=@dd#mLiFG51(wEMCya2FL9ld_UVd
z6^j<{lr&Rr2cx&n-9LnwWmqU-e+Qd~h5gS^CQU_sn@ShWc+f7lfW2+$9#9YVp)}Z+
zk>rTpyk-Vf-Sn>-e{(sKd&Ew%3Kn#6r+Q>bp2zOpuD+6H!n5imYE#KAy5w5a>0wjm
zOPO$zcJKAm{K7O>&z`>bID<R;qRz)nZp72WLcZrH^75H~rR|0gQ6PyU@b2i|0NV*!
zpT1U~0`sp>1gIAC`e&#P;Bh$q3y2Hc<m?<*>XAo3ivEZ(d~KQbg^n_tXG|5MQ)LM~
z+P_|K;qv6<9Tpfo6=T(pSCaVQ&xuf(XQ#4v`C?z5IoEX28Gcm{?EI-!C=;tSV6tY5
z^#A6|xxuGy%AFLfV|Oc<*mEyF<mbM=$y{hm-hJPQ_NlLF^gK^Sau^#BDkvS7cOk<x
zUhq3h7{x=C@({t;x`$<|&UUj+4;atu@h18J%rWf^osHh42-Y6Pg*0e|M>p_UpDvNa
zBVzu?$3@Vg#>iBlgJ(}ZN%>yOpQql6dBO?6ZxwCl>e`1dQiZQ8r*0WaIB-C51mEfd
zQ2o=8#`aWrj($4SfH>%?yR&jnf%O4Zbe(f*OmdxCw+kZL^&4jh2_fR^p%S(_qH+04
zM^2}<cL9NqjVOlR?01B;XMA=f<WUzr+poMF0#ch=AP?zl0kDj;5bp>-#X!EEJvfIM
zr?9^2KDfsCMLQe7sBjWWhR~C4R*HkRE6p9};D;<R9WmxCF%5#je$lY7!xk8Jt*gQ`
z=K0~lW5o6Hw!hEqlHjL*j0UUi*E7ej?(T{|x&v9GxOa}%FHW1QFvWuL72WOBw*@o6
z<GzCegBTqhCY57u^CeL7cl%&HABu`DX)UQAA~e1ZT3q5Hce=MnzD~`(yVt=Cn=sQR
zDfnF^>x($uZe`Z6j3(@u$NEu`&hQTgP?+1_X0t`EqK`wYZ87(e47%sy(+aAb6k|m5
zEXUD~;2Eh$(C_E}E`%+OqQKu2jRG7`t=sq-<hRkQD>D8+XO+I4_s7WH`D52sN22#C
z{#yCn<-wJJC8IL%^yEvkQ~l4{g~u5`I*iTDc~>6*IqP8NO*-bT_1_C-?@lJb&bzXU
zJpRDzr#7!#$s225^0KwMjshQWmX{DdQZ4YX26YIS{W{iUC>NG1lmW^+v6MDqN0N&E
zn5(xOALY$XrNZNZdzfEwaiDen=>SXp`Y`NR$$Iz(h#YC_s$n&7wT+tMv4wj@Ka@!R
zNc9)BSQe_bp?E$N#-H7lbiunm+n+i$461XX@7Gnlg>Kf}`zO@RVw>kmu12dD%f4yE
z@B|tS`jYZpQwCEYO1wT?5F7qwoMpd0b#hzPxtGxY_?d)G^aBqfi>XqK)QCM;{NM|R
z+b#e4P!0|htAyU~l!T)OUvq<AC{QgLo7B8pzG9Ig|EJ>d{6^QljYuvfbz`8>O8wni
zBHSt|jf#Z@<Peo{2pfim!Sh^vN`mKoMZQ9xiu%}MxDL`W<TCZ+N@%JRMuv`*b$`3-
z21P3>VHV>iR)(ji>tjQLcI9sxOa7u4ql$X0L~T0#-y5P(3zS*)Wop`v<i}xLp>h0M
z78g<cTLr-rBpab4JNXnlSyeTVw-un5`{=l{kBp#g3^Cl0Gd`qdSvd@PBzIcyo)^|&
z8q`9CC;oi2FP9)?;*u|53a?{LYyOyT_SU`bJ;7pjea)YVcUK!N!gcpU7hkjNQTQz@
za0N`Xc^EVI$s?w3Sqi!*B0`>GJDwO3rG3wja1T$Urc$TrN9kp+RX+@_OMYs5B>$eG
z@#n!Zr47~PlMZkra)5v9orKN&s4)n5{b>9po0o%$4av~`-Lv7_s?K@gm-pQA-PN{r
zZ)~r>+j5(Yi7%$s&q!M|CZw*FBHyZR$8Tmr%+0LNW29(VHj}dnfpN>7+ok|HKC1Kj
z%W>gAA<ZPo!&(%Krnhr-pWLT9d&%Kmk&9+a9_i&oeYz;<IQI}O6TBt#Fe6P&pulT~
z6I``T6|LGh@s*>m5Ra*U{x)?+gJ;moQVGW3dM-R1Fx(lo6i+4H`qRDtl+9T;SlsP9
z2Or7>=<*F)xRDXG+u4fQ{X4^f6a@78Z#a5sVNPsLGM{8Y?n8u7^KRrIU6-eE_0aP)
ze?1P`^iaR%&?WR%jbMmepay#DIJ0XeBus{hZ*`#3{c&wBhC6;qCi-%52@-ZUbN*%z
zyQs0fW4oVCH<CAkY^lE#@p^+9Ff6_JQ3;}NV?W1zU@1L68FE9Qe6gg9CAKHeatgO1
z?rE{?%sn@Djhpc5-YKhLv7gB018=vSNr~CpW)UU^RTlzE&H|GSW(q<RYFn?`7WEe3
zUh!U5b+208&AzJ_2VW;Y<vjHir;H~g|Al@ci#NQow00=$(^s(-E2O!RiF`B@ME%$|
z(abl@{etc<eyV?)Bzxh8wjQwk^t}3hZ}dNSvn(uPbAM!qXFYJZD6Cwd*T2*}o;5_-
z&Ti2A3qGB}Oapz535`=_{){j&w2kkie9Ld$pC;kZ@8;_IY87*8)s^J#brV9qRA{c;
z9yafb6?1G;Nn5^XzD`6uulPSII-%S8)9vm-n0w=gQ`7@GOaJdGk96H1iUtc20-a*_
zq(;a^n4X4XoPYoEIOT54>2T&}dtgb~#i{-70YGD%;0fT!#{6CS=bQ8_j%OdMtsz}&
zk6w_<q`{1hQm@q*zLi@JxML1KO8K8u=@+Y)W7(Bry8h6S*Wqy<2okf0-d9L0MfZeX
zl<Ixvl7=}Y=ZS=q05VpLfFGqS!p{|f=fco+&Ac9xtb`Ylj!!<zs@_tu&JEnXuPCg^
zz5W;cNfeLK#(+Q|!bB*Tx>mD9M#ykNbyWAlm8~EdsYczG-R>dvnzXn`-$Ut1a8<hr
zi+LLNYnK-%+__by*>z*&=0Z}>X@5A>gnZ}otL=6hAJ|pdF^ndiFZ*<;TmOo)G~~Mb
zqt#c2^Bg5kw%L?rn3-28>`bw(soQ;Vp2=72yvtTktQXCy)AveHKljnQ=gT@Dnn|eH
z-sV3|(o(1*S$sez4CQf+zYv^J5+&qYf6@I*M44ru^__9Fph=951JEE*WP?iv+y>HK
z#cGGnmwm<a&-GJm(<U)7>w9qKaDc|F$>T7;>6DoBP)y{}<32%d#dL3)FdAy8?MoWI
z&$<M8KKX>UmQ-0Z8TOf!DNojHwg^$ZpMzy>5;`K!ev0&b^4=WIVx$`p0C<&*Pz}_O
zcdy6c6z-u1H|T7}13=xB&Exlu0gkwnt}#+dm5I#7`*SLRRG!R##g0&!Xjp^fr2o(S
zGg@@=y=5tN?-!HvmuFMxOhoB!-G^P|z05YBd{`yClAP*Uf@BpG7;~ujE8isbP({f4
zA1dQ3ee@owxOlzlT;hn4d9$EtTVNC2_JWB8E6z2NO9}7bOcNL1r^*e!R>`KSr)^V^
z-0*#w;1}RJRM(XL-F3pZuAch+z0p|^g5g`GMHt6S{3!GM`d^rPc{xw_`amjD_sfj$
zw{L6lviUvw_;=TvtjX0H@fbkGKd03w37R};Z#_4!EISC_Rw`6p1R5b025<#1V}V4Z
z3ok~d_Ok!}fMH{gYc}HqegIlLaV8Pz_G9VpWS&%Rut}Qa@j!zYv=9AUw9>aKnK|@e
zqe8mLEjfZ{>2u--cB)XtUhP?^_}rJX`p#nsn_mE9w1t0wp;o7V9Wx3kjrc*#;&hq%
za9Gm3{icAG(g0>w;fqplnZLzd!pfF;p*lLZiboFJX*S8@0Iqb@GT^&oP7^gb(1__t
zRoi@5!uithnHz&QXAxzxmm`-yJ6M|7@~4Sp-|~#AdllM-tNie5h;N}pMQv)z(;?Hb
zSRv34Iu+ve2DgU1ayXdG2RxagblDq>D0dDBkcC;$Ejd%x=`6kGlW*xWv54%ZbHF1!
z@9J{dDDPRGGsfO7A5H7?y+u>IhE67;VvnF4x+)A824r26CNlKPC0pUiXz~uxO*>hA
z;tOQVX2V`N<_7Z#jhL&Oc$SWonu%H;%uR3S8*FZ96Pis3;R^Ya*2g(?ueRM!ODU&h
zo5Q8%(^`UTWTI6;Z;K?(k-co(FBf3I;zA6X%`s@7p9{zPPQP?3&vOON`d4FD`&{ys
zc~mCCJdIfRvs9Tg198%#p)Qz+(EKID&_D2$F!Fbl7=jQ1$xOzOwXK>=s+e>T3mdvR
zGMH9ty`9%l5`iKIF4{3<xjGc|61?671=h6!_|-45Qr?PEi>`o@o`d8F8aK`Z?|?C*
zYYqXy09ekl7$R<?ICmbC;mLTbDc31}VrBe*)WEC!(6%!#aHYIf(*23Kssl#>FSbu-
zZ)hw!z}%}|Xzv)6Wl?Mn-Z6cR#Su-^1}6dnG<&~5<TbU@873_WFTT4Qv$(#?@pf1{
z$<AkPoynzpjWJh9nk+@}N|eY1c^32}wyv|8j;0>jtIwG49JGqrXz@r^_#-z;N<0l~
z7ibOHZ?#a(U{=Hv*bhQdBu`Y>^5#Wx&gj?s)RkSGJFv+qY}N+ZQs9*;yt#riQ|x%~
zw?{{W>OtWI>(NXC%D6?Pd=bIBHTvU7$@BiQ?Zxb58aJ#Ds{6F+V?JHCQ)b|4nVYWv
zL`$hmLbQ16qv8{H<mNz^cs$SI*zQ1A`2o!G*OKytKMgF5t+<K6UhroBZt{_NV{NIY
z;jscb);IK<&^J$NkmAT`q~A&bdB%Jfsgj*N)VNU#@%6<FYn0ceLIZZ-z{z-h!wDs;
z>qQsxO}4%d<?p_DZ}^QYdFW)<{B=4pdUI0A>=a|UxTzdszlMZ+V5JY$_K`J#v?R9z
zn8*AhwBnAVysP7IC(3Mu>7i7o#z}$ItCDBO`B&o+f(L6Bw9LEP<~PMB)sV{puU6JD
z@{QDy8Q1IVu;dvatQ%{3aJ3k+Q@x~@3End=djAV@<3h=~>c78JKJtd91Q4ueu-ZVM
zF6iK!Ar{;jgY@IQjko$W&_EjFnxwYGbaPX@oB2t+C}`qTV<2eas(n{<{VMZ;dCW~}
z$W3vKyKc<?(pO8Kf?5LR?ZOM#&hwpMfwNfkqMo$0pO5m{gH-Wi{|fs<OAH8(NuC2<
z*_mIYQE3#5o^&_*S9+_Ae6;Ox6zKy#Qp)gcrE9Mz>Hnj)l`@0=lVvLNi1Ql3$;sFd
zB_DA4EB)oX8Nssg`d|xLw~oCrTV-+<$MO6b*K+Tx8P%au*``=Yl?CgZ4_q_hN8o2~
z<~1ESx~LU^h)!m&&kVeX&mTa85wuRPd^!5&KJ?T}e_Bo_F(=TBptPa0ePA1<mxfC9
zh}$05{}eHI5p}&^!%m`~sOB)s%JNUy<MWf9tf%uu$*BW|^n-Di)|_0voF8kyc=a&V
zYHoGE-w{EE_p*IDk)JO5N#(McEuSLQ4}$<JAL^6VT7B=nrh$Yfdk%+=mVnQuzs)n6
z;2`99^q4%4d_uR}pw^(j5Xve=gr5}Q0Vl|Yi#HDwp2ziTX2lRwIfF)Y2?(fM9lfh&
z=KW>s9)>Qdo<dr$WoMU|DSfkz4kc_sjK0|y)^T})J2arHmezGnQ_$=pBg;P-Aj`fj
zoaK`y2tX+pAe$xS<>Y@Jb3l2wm3uRRKC0Vgpl_8j%&~PDAUq$*$>nRdifRfPV_QWw
zW~SX0^7QmH>DNyN&@xOoB;>X2e9@nc{-=#NC|zst4-J*@JNoA#YtwgZjcD-{U@?BH
zz4?or>|gTK6dE<<+<m$`gA9uJP5J9=T8`3P<G(UxflH8oQ!8J-<pt5bkVk1fLTCSP
z=<s2{p|bJEc?kS^O6{<X^Y7K-QsKbp`Gi7|7mwcC5ci_4IAK$fAHaDUDVi|44Z#Ia
zp(UA&r`r@UYpCH#I0v=D=;h6+)FY%RSaK_ikKB^8lq=+rY>TI$_q2RwAaf8j)J?|(
zU+f_s-pg!=2|C}W2LSX_`7L-8xe+<I5aTqRRm=Q*cegUOrq}km<R3iPu?V7NKe`W%
z68K@-TcYh?(H>amz_?0*PoKzW{+{m2@HFh|tkueEBETkbz~7JKw8_se1b0o%!={#&
z8qseZt%XF7_e`O2pt}y#Ne=NOM~g8t-2L66$cwz7-E;Bipxw;%?2iKS;$W#MnNM@9
zMIKx}<HU4Ib7mh`>gwkPt|LIluZm}Fm^`K@ZX-uGFs`W+vJf+xDY#}wpO(FO)5V;9
z_g;%%^jV5un<3BD;Q&^0PfTP2UhEY<80E8}<0?9Py<vR`?85LzP_1{;cxzLET&qfs
z(m^U1Z<-039;}U5QseM~6baxeT`tI3n9Ldx3kREF+~RX9NRB~V)D<1zB-&8uGl6xe
zxFH_A$s@e_$kb!n=r<J^45zQ9q}NXw@Gx17M2T+`95dS_d%;su=#kEehaHlQjN3fX
z(>;SGC>$l$AiL@I@;BUJecpRtFX(gam@=adVwE&7^G|Pj_f_zm@yQt`Z|7r!ohiDt
zrdPo=ON}9D!)`Njb6%7ZYDBah2dA<_ezmS#pL&}=5GP+F<eQBPur;!ngFL>;AQiZ6
z5=-K$>FbMjbnFw^saf*i_QGw=0$F=gi{Bx15P#4aa2_Hm(f{$|UDeQ0ArwlVHBJOT
zLK8f)!zo;rN9~yf=vkfzV4#r}bO%iD0nILd)rYOUZ9JaF<k8-HqkKYK5Z8gVm6me7
zUE5wQ%M@A<=0P+Im)ZWC?80>l0>p-fhyhCQ7`74yMogCWync)a<y|!WZ_GCx6UL(s
zYe+dWKb~CCl^a%6IOPyV(QgNTh0KGVP3Ep~ola{~u{IoKf1t}Py-6lc&4{k`%c?H?
z%f~tH1^^ASy}lZy@hp>h1Y9dVB<mehFsV1bHWF|Kt>=&@3`HjZVH0!(dT#tD_zC4L
zwQRd|u3{;@*>VvIsHvW>&zTjT!Dm=dhF@OCFBQg+<AzZ`LoErYnxv|wK17~3E&KIu
zNn{s0vdubCcQ`EYzz3LC;0m@M3)UyhN%UQ_qy~_hTpmV`;eyR%#J;`_JUPwz@rX)K
zJoO9y!dtp~Ut73PW3SEqQC~qzn*sT8t}ls8>M15|7=y<7|GFzHOtcD6O@-Mw1Y$3+
z69=$?O$Xebsf~pl*ak`@T@*6^o7j<Fj(WDYx45^m7=|i`Rqu@#WBRNCIh`QgfmI3D
zgO$Tx={e@sI?L{h-mSI)G5?KCOK;S}CtO-hT#cBrF#UVP%=m_ksTi&e!RWf{VcVFj
zNa)CH9M~G<`dnvS%aMx0yuW<&ilyegT6;-7C$naoUa)F8wS?>R-DSeYXPaTgn4v=l
zEBV~84QNJWB(ANjI^R6%MBYKToSdSX-K&PtTO?F&1zM>Gx?cv}%Het4O!0)w-lSPf
z6;GJmUOuQuLptNnktitcfz}Ju9a~1Wz1Z=Kur^sqCC)D8?O@bW(ZQ+8M0)iCcV1*I
z%%a^vB4gv=`$%dOB66e~`@eVUm93e35LP4nc!s%6gOQoBEuq^Jo6ucKvON>Jd(|C(
z-X@+ES7oGL>qwtJ@ZxWNm<>`SWcm-^18YIRhCUqJq`@Kcbk>V>1y~^{2K2%E*q?5Y
z+`22o@|0Fok{XF5dAFWm{-j^`8c7V$b7DVSyLmB4mJipjF{DTn7$)iR*3F^Ek0e=-
zye1XV%TYf{`0#k*P2<007a&E1#?DvoPd`h*tWoHi({BCj-VUt7oZBgzyi(f{^vyhm
z5{>qidc877LP1b_(L?f=+^ffM(rSs?iA1_CmJ-mfAj>~UW7XBfvfQg<W!ixid>y`L
z$0Fk)OTreq!(#OYjf&2XfB39|nJDVxXtYF?R^cgnU0>eXm$bslV}dnJR<-e%TGOTd
zM|Q26>u#EF5j4<I&j7h13+`Yy6@g}Zk~s0l#|pLiZ7x;?gf$Gmm!j<Le1gcMFTm{)
z|1Ba(DI`oD;vev`n2dBgq{_=Q*0$G}AXU0Vo@R~L0S|?z&yb6o)Uzb>wvhTF-TlO~
z=Pymn)IJUz*Ti(rL3GQ9a+N>V+-z)nl|=I#m@$TK{`cCvI=FkXdg3w$D@kOf2!aOK
z4uV4&Z0toe8OFicLbBOm59I~dfzmWQaFuvJ``$t2WxwJWAv&@YWW#b-mEY}P6WW_>
z?(f25b5N*0qH<0oAHU@J^+^4Aw|@zfU?p%M-YDArfXjt*g3|_O)-G!bZ#1_3wZuwW
z(u;N_St>=BbGm>($Dr$S+ouIPk9<Vr<7vEZ4NNa!&%?~HL}TA=ZtjZE&PNX2^55!5
z4ERTi9PWFCt`PqzdeZPaSNzZPurEU>r|IEq=@w~x80#j5Ym^u!@5fhr!P4W<*Y=#2
zI1SZzp#xqX^Ede_`O;7t|LA0L+;Ow6Zgia9XP<L~Kl;H-Ko&#_-oQD?mnX?=9Br{O
z7g^$&L%U979;L0G<4=H`nMz(qcYUw<z<C2s_8c;LyMj)ig!cYvz1p3ksrVOW^zwH9
zZ&1FOeN5dbFKB3bW5Z+f?C+0!=$&(_KC*yz&*s&*{!#I0>LxCBAr*}29Er&c&lXP6
zE{3pDWo2H10(2Jb@l;-k(5G&6uMQGL>#v!>BTL@jdt=ZOoE7l8<>W`<vF)2R9}rg>
zLu&2aI!`JRh)21Zv8|O;-%w|B>4wZ*$&gpP`k3z^*KOi2><ZHa$Y5oZc(B?O$ybZF
zk;0TILe<#yWp%D1pbY9!FM~C`fk8P4DXwab!`Ll!fYE+IZ&k9I4llgsKR2A~a=jR8
zZTb~$V7KZFv+i9AA2E~pi7=kSzC6t-kD}ci&ZKv}&jL~fNDO)qiU(8)hN2y(81_v~
zTLpg=-}wg(jWmp@hDmZHSZkO~ehHSpAZ2yF8tko?a%o%23iL(Pr-z);nx>Julf^K|
zaSfrV*tS<sX1+Wcy3!d$FAdyFs<+Dfv~+3fXjP4z{(^&U-b4PqW1gPYcj7IQbXrgu
zsH&~$p4hV889nu2k3Yeb#Y=t@_>df<ZaGG`a{O?z9wZ(A2-^eHwSN}{8zr5PbV4{~
zw7>3u;K;K#@<=Baod};fm@qU@YLMT_kY$rwRyiq^fR`<Qu9@tx2M9L{iU~&a8!+7{
z+Hp~JP!}KFQw#c`G8`UGj2TUP07+VrPe~bXC+|<X|NElx$e1iN75+(C&0;wKm<K$e
zqi68WF~}9fD=R+~?0?0V2x5-sC$`0s{7e>DebRUsuvH?aB(O$NBJ_1!m$_3}@SC{W
zlL4R()lqysZ@1^^c5r`EeNFA@&kA<utgV<@4*a-kI8)Yx+5$p7R~pJ3yTL}M^RI;y
zn)Y2I{j8Ap+_g)Fb{+D&Twmw6rds#IY|^rJ`S}*gT$PfhpK*FS9zA%}i|tSoc`Uf#
zl%(^9Ltt;c%ww&oeY$(|;L#D=XI;$3vzf;SFZQ;1n-N3idD{N0rY}YYxR(<<(|pZg
z1^AQe-dnXIKUrzqIv9>#NR`VvBt%(5svK&6Gnf@}6_B7(Zr>#eyA{ce;rrbfoel-L
znua`lRa_q;1HU!50OxP)xTo#!b_x7H*arSs39Kyb9(dlPg<$Zb204^BU(fi}zL{?V
z<^uDxw&#dSpXb2A^MV`MLb3Xc-3ViG8HSkQ{7B(M*_kEmYu5X3@4)1Q%jX4adgseT
z&BMz^N<tXIXqF|F^uYdt)2G|+HeK$kLbgpEy6dsKW-BMN3>Uw*Lx#d^7ECuZh$KAG
zL9O6)hx?E9Z1jdGCkhAg${4YuuToojZ_y&at=~!^8Rd(cWgPjTC$R1I=EQuFUr%zI
z$Jr-LY<^t&0%Fz)l#X8?GWCO7oWVBILr?5_8n!xq5UuLBbP$L9l4DsL2VcGte>a!B
zlS*&iIX@%tuow&5JiPQ_Y73EY>o!1hYKF|6x_13<k#r!DWtPFbc(un$nd)vj@W(IE
zpH8U_V93yos8)diBTWgOw&%73k}~I#mR&w1Q$F8$_D{_a@|Ioo^;f-g=t!H@5vj5M
zI1PSyK!0wqiPcvND3bl3dd;JWClpG>(7#s0xw3V!)J)G;ld%8`0Ep%xM2mo5w@xn)
z$!DVl++1|e@Nl#B1m>(+@L1}GP6=EHPihA~c+Vmei}}7pz-3SN@)sb`F>{i9)rOGU
z>YNFk2j%0#npKH@=3{=QKIQa><Zs$nj()W#_Z|dZ{<>!Z$RU4ZtyVzt<Vy+VFm1so
z$0zDw(wtBGWquhdS2IQ`0sQ$d98Xx4He8Ur=$G{2&tc^hJ05eU#taL?EH$qUq)F=K
zRePdS)0Y3tBwl*?c1kkEb-OH2Zjg=ZXP)m<?fQwu^uk+1=N861vGJ5!c)=lOm_rQY
zh}#AzduS)vmQ|rBU(=2&=TxCWx$bgA1eLd<ZPG#_F1}jf#W1#y0EQVqrPj%4kE>iA
zIU+tJRXq$O;g5orfB`(X8Ftup?lnhu^6pHGu1I}Ppcp8)%ECmyVyRU($4uaRwAIx+
z1eziC$AmqPIhT$we?T=uFQV!AoU<F^s(Swfuq_fD@cQB8NWS7H#NS9OM>i`&Gk}JK
zmz!bVanV$rm5)!|R&+9|zE_Bn%a=zWcHl<|!$AL%gq^zmg28DF@v8h9!+9@y=J9{D
zjojVerJGh3g80?9XV_zU%8l)r^z(yjGA0Wjy}s^&89BIbx^LS~M#NT(dod!ja-=7i
zw!;w+yV_)NIfFS5Q*m8X&sm>){OnJSd`b(%G>A&|i@>lBLeQNA4u}#9T?;yBmHuT0
zHze13J!sCACBRGJ><oCa*Mqqy_t5xrLpANw$l2+$Va%9v?&O@}Rt4fWGZW#nws7(3
zu7}+nU##r?%KasgMKN5gu%nxUcJT%OO(m{;F>yatNjJ4Wb#-xEsWF;C>W69f#0ill
zxwtlPvsbx4j|r#FugL&O+u5LZM8DNi`6<1QH{)`vZ)%R~L$L*{0P#u-9t9GGhlE_+
z5%t!CRu8xp9UM%?u0L%gx`1a)w&wUBA0Yfax5u55-Gkc5!n=cQ9!q|c>;iR?Sn(*B
z`(hOkq^kVUBmt{vr=V>@vfHQ~J#$N_2iA(k9wI8TX*U3TKKm?nhWar?Q}+lNGtT*>
z&`dLRoF11`u$2it;`AB{bZt=1@3SesXn*lG4^Z#Q)}|fg;I%0#FCbB%F%}qC&G{2)
znby2Sa%Vn|AH7))nmE5KpfN8%<FnB962&Jy#rm61zAm{o^6)%Gt`FL}e$MwU#!7{h
zjr{vg$R@Fn|DZM`OJueK$!TqsD$Bk4?y|k!7y#IDOCTWs{{WRhYQJx+{r!ivV8dc5
z?VD_?9`Ku&)r>93O9pUm?6cY)R9<-7k}nE=1|NpIz#ce#GgcS_YHPp1&oj{iJH|)e
zl$%?-b#}>2>g^mlaH*7qU(U|tp-5Jl3y%w@mRA<T!!wr5Rr||3C8yn=)HaQb)8Lys
zAO0Ti&pT|{cw7ixKP+#`&HEc2CwxN5SHU%HrW$yE+niTkJI=Fu2w|(;h1&iw&-&4>
z7TeplZ)Egk)NKW=pfB(nXU{qhZN9!6$7OVRWbseqmPKv)kpYOJ*IQm|7USW^7%IiB
zEzr|W0=|S7sycxe+_tb&4>B5qF5X8S8P0f5p*`&_6j>(ysBh$_9*iPqgPR~{fO?j3
z$>;?y$|y}yoqEH4{Ejh#J}|uCdl(@ZIUzFz_{$gFM$#Lj=%Z#a4vlMg&Dhm+vW@uY
z3Zttr$Eb)7Ofag8g+=-qjV%Rma7X<BJt@@fbkS!==XKy<z^OE)$S0q+5urV6$YbWZ
z8hyDsfK>+MG;y@Mck(B5*I7q_Q9r{7eWo31F;{hI;|Ul9G|Qtp*LY8;G7|ju$~XVy
zshv7i!zND0Z#iB6d7cb##(^;dj+WK6g?W)zR_g>94S5P}*dZnDY24#u^`ZPr!LGf6
zqZTE|LwrWdqV>tE)DL;jC<{cqzcF0x0Q_8Kyh%F<-cP~v(cf<hUlwP2|8?nK4_}qn
zw+%<rtFNGs8i$UCDc&pC{R!1p6E2LIOpmTIoYuYNt1{peerV@Tr=Q`(6nDD1%n6@e
zWBmFnXgaI9_ji|k+TixQ@Y@*i>|&tL=PoO;-MifUQeH0OcQORSJ-VBgxn3Ew4vkxd
zo8A>bb+7Mv@44oU@L9zRrF>~>AxQEQ#$*jS7{=c<%KBf|IhS`GWsFya)>Gxz<0Myp
zgG)vgXN)$*mPh5&nT@$6<o97&NmqDKx|;s|s*K!S>X7$%BxBQ<6~NH_nem%A#<Ou~
zF%TBSnB{fn!e5|q=yv`UNE4jDXqaGZ*x!78oA9Q5j$X&%>%*JcZMwee@Fp0(>iVki
zrucuC>sJ+y4PVFMIDD7kbcFSlrQ!Pbzc(yIFsMV)X^a!B{OeU#J-M2+ihs#PDN?IY
zGg(dqp(f?}{Aw~TckukI6|xaF>s``L!iEXjq-q;PlXiP#jeC0MPrOnXq#O(oVGDys
zPzT4W2|Pk`;_;NSf~DZedyr6ME4wCLe@UtyC1^x|qgbB_6Mj17Vt^=k%A+fjejC4b
zYdbYwQUtbIEg23$9PAF`nH7JO;}k5)V{9sKuLxR{EF~~t;sa-Gl~#2K<DdG0Mcb+4
zY~pAK%%E*X%CHte3Y>&8M!!cR%%s*@F_?6i7WrvkM!U%%cy?535F#nr3BDM-+U79z
zBsefY5^lj?jAG*RPQfe@wi(}WquLpxMKN>!eCk616d{Je!YcOLpWYevw;HbFa2&og
zWKdEvI4_Q(gtyj9LURTJ1tX<dt)|UbFN6D&(eN7lfAN%30Exgv_p0YwMIGEHpuy)!
zXvW~z*pMJdX^hP;FAneh)(?j7=Q=xUtFcE_CQpgGckYyzSs9Vhioj3?C;~TnlfcN|
zMzqk6z*NTV1q!spOBx=MXvj=x0%e%E&KTQVzdt;C@OZemv@)zLUKlPcTo^81xL69M
zNl{iSWzkghD)lm^Q;DLHlqyx1JQw}UGi8pmecpNZhG=yrkQJH7;nx<vQ^Qxfyn_4m
z{8b^tp|A1x6pWSMgiyT4FCnP!=d$|$v-f9hUL@I?*z0661`q@@Io7au_f|`K&W(DX
z&)45fx4Mwd>GSAOw_4r3hpMbhW?~|k;k>`s+=C!0Ab})-s*3dw3!lu*mMt?|z6Kk5
zK~i`#jQ8Gvge1^1N3sjPVgL#mrhH(>H#fX41z{Avc<FqdLSD0UAGyM;8T(o>?3%Gc
z{^02T>Z@yY_>h0hoG=rq%f$IUM`kk^=JxH|(~aAA!yk@K=g)HjtrTA~))_9P!)VNz
zMHV?0S&0$V%%gMX&rkpEfB1)f+ZpSPygwEGWe^d$<IY{P!tDF<ys{KtY6(M2S98vB
zgjwQ{Y+(t)hzrfnvUm20B7ZP2kte_YW-QrCMtQJfCePh2CC$Nu=WnHaHqFP!@i;hh
za+y_oA$eT;`s(z%$gAgrbEsUW+W9qx`P2P2aIleW%z87U<H@rZ(}R3#I4P8iVb&}v
zP7yOKR2dnG(GA#NijftaITJ5kzC2yKdbM?{eCW$DdEw%P8s>dpI9kYM>RcUjU|3FD
zJ;Mz@*K&OKw(oDH%<IADe(sRX%-}Nfg;9$!nNfJm^f8M|S(FL?uRNYK)Tg@hjnFQ>
z|IPQ)-~Il#b!c!J3N<KlUPXpfZyl7DGWc!+a_g9(Je*&a<!<{(o8)k@L>@z!y<vTm
zwck1ywm4|Cg=cN6Yv<1Q`}JZ{w(UAk3}_jH%}&hD5QE1a)tEMBx^34f!%0H^U-}=T
zBfY?kEafNvqNTX#!%@;@@jQ?2(%E!r$01zOdYJN=HS?gjqowa=zoW?m{u^8OYizY_
z;mNe)E)kxZ;V9hL+^IuSdpLG{r2l1t_Mk7RXL-~Cn(}oCsMPgJ>X@w0>1T=0QBOy^
zEF-6v8Bp2*r<^8sKHqzeP>w2cw;2z&6OSYCaoQJBsI%%rX$KI}7tDm<`~&(^+XC$K
zoWqdQ>(oiJ&pFLsbYNYbFlqZ{bBNE0$8pV3iI#11Nh8sF)OI<d@E!e?K2Uihx&bHC
zv!{>iz~IyqcX2jyTH{w+secaUuC5W<({C(g*g=KClUzl&v-G0AX8=)kuC#Bmo4TTZ
zzy=Q_6B*FwIc?6&JdSvLm44!#I+${<U230XBxN1?B$}M4Q`{_5b*L|GhwpIga+=X|
z^Gf>&r>A8}ucp6u>9wR++=4T_wRvfrNS{!rUA8graTw5}I8dGAGun#T6GMNN&j{WZ
zoc9sR=UiJdKmk3fyQ-Un>Rkl}>)|wbG4m?#E$Q~Wv1Sc?uGEpgV-NcP4jv4iAbml4
z`Yy+p{Iv(TXy^K92eE}7vh}KdB;L@4k9WyfgE@RVB+psSvDz74(bMJ0IZJK^`08-3
z4ZYX^g2kCG3+XSj&+ts&P;ogGrPcBWmgAe?ZQ{Hu9F*o>oR;vuus5w&LAv8O&c}8B
zyE#DhhDPXwj{r?-srTLGF>~q8NUN`mcN*sFwhvK;EhU#OUcqZQ^L*TQHd2<>lDGcc
z_U7E<5oLzFmpqo+m8BriCwZ2>Eopr`f8|r0#lD?O-|9Vh4ubf5<sC3&G>+*p@P!Pf
zT}lI_<dE$%xsTVBX8eHvrBR>EHv$VynkRe`7JQaxU~$;JgqP_S=5javrlD+kCxyV1
z`%MqeY-l=p-k(?TtnvNN%cEQ=^U}x3ft+5oY2^*PyH2mlP7gZ^xX#GA&USHnM_9`D
zYB>+#bAZpSPCg_)pJRJF2&Y5G*C8CjM+8d6=}Q-;Yu|r2orz*VSv1SRhtS8>h)PDA
zVL;|o86(#y{FD-(V;|M=@ufgSD5w<kQc$T9AAz;~o#Zw5`;H_UC(VNpOaZN;D!ttE
zUX2p@fD=|IGF6NSEO~V(W|l;uoJC2@JKn7@;Uf=M!nkJviVHyp-h`0n&y4gjm74-(
zl-?Lk_w?C34gZ0W$J<@r%zGz77v<uNB6Qn$nIdjUMN6VlhE8<k=jmzX?CHspfRsQq
zK)*$R>mERPPkEFD{U{VwE+d3VpVAM9I=wilLRZEi=gaJ?B-3<D9I_hP22Y0iE<+P#
zLEi8(TGMlelrG1Se7avp=w1c0R2c#h#gfv3CJ61G+6u{{te=YFe!2#mP<dKk-<)pz
zcw@SIW3#hy4&e}f7Sg7w*r$za=h`ggV(RGFVX$MEVBDx;lKYL9c3uT1NBz4_?yw|=
z2}>C|YR8<|)#q}BPt+lnqrD9O{vZDBbm8))=~4Rr{fGC9!}|K&E;-f>lFNTa+W3k+
z3u@4)(SzfQk#gks<?+NxKhQ6Lej;=fUHaX#^r7pP>bko=eG&A}oLZSKtzD|&b!~O6
z`-&KGeY8vEj1jFVVfaJ%h4g}Sd27y*2me4k3Jl}Gg$oyJq_Mn_KD)8LKG|^k5IzZf
zyrmA!_?n^3e9v(K*J(O0!~EM1+T8nq{xmZAh2iDl&F~hH_9pwf;8Xd^jZu=Hg+Hg$
za{WqT{k|{2zyKC!IljTL^3_*Y$FhD$jqD#z1$=0AZMAV2k<E^}d+$#9oaGEH-+J@L
z%{u4pIrFdvZF`1|{izsgIh#0@u3WiNgS2lwS;4Y>m#<u!u3ft}UA%NLvcRbt8-3r&
zAXl%Lb$UMOZiI(ygtx3ze!%bk^y81+o69nLSFT=Z9*q9ig9}m0ObWB7)*?SD+s^&%
z$TnYg|DM&emSqh8%5?`3BrisE%<RWM{WRS(11EShq;dq9>B9hQzY#OwE?&A+IRojN
zPAX4Xl+TQWyLaw%mV~;&C!aidUW4=f(2vod1Bm>jj_u`S@2V?TF4UlE-z5&1%5%xz
zj05p}Z^%BrC5+yeE<`^5>Z?3E(%E|MG2nh5nHe9xdvCqYp6$?NjFf&9x*LEn_%yz*
z&H1jWE3{jdms&az97bRS=InYD9LADTLYIC@eJL9!5`u6{GF+jD{UA9Xe8aB)bhAtL
zEoEq#=JdPIQR>h4C3h-YlZ`nR$%&UPU#zUSp6`e@sjlT?FEZsE`6uKV!}IZDmgGKB
z`k|*~4dG{I+gfn)-MoAEUiWv?e!W`*e}Y-w%Cu{252lRf+RmWKY!FX5bR*6#*W`We
z#NZg;XR@d-Bxfnd5ZZJ4jPj0|X6k`Fv9$&7XXE{3!+5XU2FzLA!MD>aHY$dG$T_Mn
zXtP`Qce*DYhYk9ppZ3TZz=5bQb0Bg~OVbJ&>O*_7hbS6bM)Jzl%XPj~k4YK26ym0K
zsJt8qL#Lu2puaw-E<1}TN3^x~+M+Ay8)HexV=XH~Ci>Z@Out&>C=OyXN9{GN{Vw+f
z^@lE&HCM0tHQkU9)!8`77ayUQw0-7?|8QD%)?Db%*`s}<J31&ITsh>bYlU9YHj6;p
zQYHcp{C$_iRZq#iv_(UX7wOPfJhQ);cFy@mSA!Rw)a?B3qn5r`eUHwS^z5a@d2WCa
zdTD=Sg@Lh5AimK-SR7b6s!ua2In||aCK}lh-s%&WxyR9~UO4_MTP7{n`lEaXeZtq&
zCz(y<wyccYO(xebzsKk}9QKwP9J9o<8}Cco_oMo3=uB^xT>cPlPCF+#j$Wkgs>1|v
z=k_!ri`F+2DSzHcNR%eicu7ki(+}``<<qCpsx!cHL<{t4THX(RfI3sx_JpbfyzPck
z_@aHz08q=v)hXS6XUR$fT@Cac?FqlMY^`sK*}S8EAGb}By~R;?>V(4+4vX_(Zvd}5
z|4?vHns3TydG@;JZ-Tw=yqeCQw4Bb>yTEuY|K|IuY!^cNR$+f)`M+E_9Y_DGpUtqG
zF7bBJXd55>%1yUk<hol<2#vn`a_e!}fpXz53C~YQdNW4a>S$~5a3eC6I>$njuJ71z
zc5?s?y6DTp`Q>_K-{%Q~ZUOlvv}_5|&@OA2@wiC!r0<V^14ilQIJ@9H_vg5c<Ij7p
zdw+!Gtpy*qrS(c`@S0(i3I5U<0nd5qI2(C4C6@&m<$lvw*E283HJgubJ-Zp(5`Kq2
znt|W%G~U{M){;!x$;hLrzY!MP4ho0xdN_z)A0xaE9F)$73hxRBrLlz15e~u)w>flt
z9l{}eB;e#)Wf1x9y9k=A<mxKI9aR}&VmOw@%q$;FLMK6DaoQNwN-*|8u9J!+QZXjN
ztx;(eCM+kDwDQD^vPk0ji;|U40YCRE1RpnZ<I7<jk!J;hq!yuAJc^8Ggy4=Aq_n~#
z?^@DFd@tBfz^toY2c$_bqDW9?i&O3yq1mNGB0P`fTL|8kmN4Qm0>JimM<m<RD=#5U
z(j5EzP#TGAgoZ^SIdyU+%H`P_oQ_99;B@kQ2rB4D=s+u@ZG(4C2z_9T1X{{EM(LP6
z!l=V|WvHmpJI{jSlPE_uXv{(;-1?h^Bnl)a7o~*~;b^oRW;>Kt3h|@XBf|mO$vfCp
z05r<j>>6dHG*zI4TVi}k!9H~=iubAM7{zbv{&ef+t?BmlJJX}BL$B8%ydP+~lxOXF
z^vAS83iF=PW@&zD^Zw@Z2;)%$h5;yjW-(@Hs~lJ&dCyXv+Vz(wk4;y<`)2z7_rINv
zrtgpa2|1Y`k3Gi1S56$8_l`)=ukpV#vvZ1KFoYI{iuB{~6o$3IYlgDN=(KwJY-fBu
zfAVbFyt^^|>0kd)qqcp$(ut-km#<D2&R?941qaI{GT`d}cphF=<whpWE63nMF~PrD
zrsVCuxOg>VL}x^tJ(I_l*?LgJ5W~PBd{p>oj<5e^c6_CE3~-{!jH{PNZD{_z4%)`v
z2lAS(@=HPXZ-v2IMC#|zaU6ERbwmD*;=ViPe@>^}RFmA0HSn2nXT+6@Zdk)$-mg(F
zef9hAzMWRjoGwqn=PYGRPFPzr!;10YejQJ~&z4Z*`1q%P`q$~F>pz80j9EgbS1hq=
zN#`M3a1@yt_I+fzOO_`LZ`j%zhWBgNu1?>4{ncc^lD)PVsX2Nqb$ns%e0b8OI;?*B
z>8J8cjwQz8FOr5m!|=Orzx}4OwKzh|tk~GxO8Jh|h;CK|<3FBcAW;VZ9Va8c2%mk_
zG#Q1-W*i?3yti)Ot>aEvYHU8Ynsc)W%yv08ojrT5{FEGImi&D)?-R%)c&3>&*KgdY
zv3(4bc-%cH%*fg7%$~vj@Y0o2Cpuf~{Dsx&^2PI!(N=b6%?+dQ)PV$+67>BsgG*UL
zyz*BAvg|)(V8K|%@L3&uc<hZEx2DbQ$Z4q~P6A7CF;tT=8J@|ME1?f33WtxT!06N^
z?DB1r-%I1PW|Hq(<Wg-%+c0CN2L8MlGImDF`3n~-L!u*EkneoM%v!b_p7v7ypZv+>
zWJ2X(JU(}BE$^J19`2Y)W}mKOX?JV+mK~kWg+|)U4}VHKOZsbx%Q^Vd^_$7t_mCV-
z+U0cet(<2;CGW)<0e`dVd^6$4vGZUD%^31W8GO0h5%+Y%#c|-dSw}T+Co|Xd1MS(q
zzOL>4bSnMBY_!$0(OdGpzkT~&9gm!6(w3j48_jT%k6C|4vU~qz%B&yhV`o<_|M^Ah
zmm%FAWk>Qo*H;Zrv8<(;Y1-P>_U0^CkC`fUJm!5fuy6o6guc%KMR(ZTT(1L(quD*}
z3jxhe!Yj0;_4SR?6umjG@e2Kg4yInoMChyEq$=y+Qs>&B{DOM$I9>;xnTSJgI#Q>v
zc2FATclO-b=}gL&R+qM5UpvN44nio+OHn7U?yW=Lcs^#;n2~0s2~IYVs{^GT$86Kl
zuhH8qxQCB8J;#1<)f+<NI>y47^-uIwuL4LjPrVLfGiTLseht0kWN57Kc4_zIFCBdj
z4$Cs}rBmqJnX8rIo~16*U)x^3%=wtL<Q%OyY>P+AT3Im&(BqWvY4jfLfm7Tn9b`ie
zX!TJiba<tHYp+GLmY+Qm$%w(Ebc6qx(G<QPT2@b_t7-A#jn6`%-w5qA^&^kUqN%SA
zNC=J`zHr3}3>?rN&_kJ9qdC$an`G|KCvErDHP!pfS9b3ucrUNd@B3zX4~(?Qq8w73
zRNAdJT4#3q%NT?(-c6Wa_MeVh+d<lv`p_3UKr^6y8g!uFqL=;Uw0AsYIQYi?gh{dO
zfRpT4_tpI(_wjavI!-n1(x`P}ud~JR)xy6Pe(w4E!a?c2ZC<|?PIGPKZ?|G+kl$WO
zaC~FX+;6<MgjZpSG2Wk3XmYt)$}!s_dX=`|_LBC})(eT&>l|OEJHs4jxu<=C-S};H
zzH1{*vGpswMtfTN2%adZ5$2fgJI5bKvU7R~_)(tezxp`-vqaLaYr~OfuSYsfCC72d
zbLWAmrSt9O-Z%<%0M4+)7rAHo2N$rEf85W1aT><Gackf+AUBfna(_+{nMS|IWAeSI
z-6)`(__l9cZkNyK<DJbL9;wga{m25O=cjCMf~B1M!(Rd%gbsZDZ~y@R^hrcPRQLYy
zYr?Oc{)d9U6g)p?*x#NG9bboV2p<v5SUCUHmFb(`{H9CQ_^?qV(k=V(U6lVijB;-%
z(DR2Nf&hV?P@NljN+Hf^goM7A6#6W{5JJJGSc%9e;MkE~Q4V~ND>4zx6Q5vDrmm2a
z+eS1JqPMoUrp*L`urK!WAY5l$a5;ADWEJ5f5snEEaCRh*QQCaWJIcjq#7ag+SQ@$Z
zG=j`<_`%m`SH>Yr9NI@rnFvP2i7u;>vUUUK2;G!gqeU(KS7nd$jx&jJn)~vh98m0(
z#S(=tl3xY2yz87Ji_$ri7^SB`KdEub$U4d+{LDHNJVV==KsnDtcXF^i$cYt3v6U))
zmTj{f4F_TAK?rJ{G0uE3E&B)!<#;G-6?&qCH4bGbik$r`jMO{92y}96Ud;J;>(;I5
z#~*G?8@KLjCWmkc9}M)}8cEV7+a7bQ0zbFk(e__V=hp0nax(p8%-$kT=u34LY1i6i
z*v-nhv(uGtu7%HCXj>o4tsd(Lw~ppbd<yvBZG%4nd!=xs$?glIRTEG4_U$wJm%+3n
zSMvhHvY8LZqI~RZZcR6@-<Z~K-wC66UL(S})$`NEdDODetCm8o;izwiS0$gt>E94s
z<f{mbH)f-ufk6Yx*3pf@Ra_3bZ@>Ab%afVqV6UIj=@32^d^E?G8DDerd<C@q5ym(A
zKyZ6+82{cK^4i{H-Zy~d_(ph(NPCOCq%X7=<b;$m;_Q9;cHz&+Su4P+H%Tvh<LQ?S
z5#$^^%=dKclVZ=Tv1cfr(NV_XE#xb52WJig?04UPH+}!zZ;Jn)e)y?!0t2Ewpe|j$
zGMztvF}%mX6~^}A1h8+DYew6R^?P+(kRiw$ch~PufB54c`;8~B;IHDs!BT*C!mE7m
zIk`B`o-+PMX5pBUCz*m1jB~`ShP896HKvoH$Ts9MvKGEb@FHKK3N~a`27X2sVN_v!
z#{2L?%M^MS&$2{emwt^bwG*DsNo9#PhG{e$`^n}TfX62<264u1P8RzlF^V&S8JuHZ
zv<n;`!9|{yv}V*6pMkl16(!BJln?LZ0IbnV-KE~p$7~%u*i1OH?Ng<#_R?Yd{$%Eh
zrK{DEx`q!KlL1_Pqwm?2f#ZXWJLJVvyWcbQk|@oqJUucTIh9drIQc9|c&fCvx7lxh
z`(2j-<FvD98OIh{!DZd@myt_H+YClINQ1$dOlTQA&Jj-;xU6_Uo;#Osqvhbp>6WeL
zXyJ5J$G30a4Q)ds&ar$i;aEvO`S#nCe{d&XN>^e07^pcAIh4Te=ydu)`$c0HoGOF1
zYvzaka`f1||47=1`jm${CKsq9-&_v1I!Du%EFURNvbTKnDa+fD4b1ZKF2|C3vM(6p
zI;YsVl#5I;oJ)_=-l9X~i>qGjoyU>J`O?{eN#9@<Y147+r)8;0UuecdvI<=Fd&@qm
zW9^MjFl3Qu>Dx&M;gS~Tvt|F_#M$&{hhtQ`pnK)}pFH5rAW4tW*Dby1ef?Y8)nCmd
z(?`i7`nI+$C>Jmj^W@wfbw>D=TO}R$s*9u!I)Eto3DV(Q=<J>#q_(=JTIj1UjoEiY
z?odbYF;mqtl6xI+>d^c2JoUmUth|-w6f|vV@Fz1q$1Edh8oY;|oL<_D`^!os=<gjg
z#9PcJv>c}VIN_C7z)y3l`EcNvNvdodCvec$(10UGo5Y*Xnz{NVx$r@QHo+)N-_vi+
zu2g-uEiIk$7MEng!Mzyar8RI%+M$!~f}bEF{0IF}G>|xOAg_wQ=xEi6wV6a2?LBg6
z>PbI9fA1*ULT9r$<?DWb(wS*I%Gj)SVVn;f;Ju)uU30Emwo_Z~etBsF^4HeL%*rUw
z-Mo?(JgnLRM-G0oFVO^j&{bdI$W^A%cE-Rhd_(%$oAhdw&pW;o(lkJgBM=Y7Q#nud
zW6o#}wq3;c0sMpy0sGVWxwPLl-=9n8U0_K|a_S>z?(PW=<2dRq$NB#PuC>8&KCBVr
zs@h3IuMWa__m-2F<?4TJzu%;s7RRI|?BdZk`rp0fIHNM-F#n(5ag8sOccjf|m)ef+
zOrObT`7P&(U%cc!_l>20xUbEP@9yG28#<VN)$6`G(iqM#=lb%?`^z(B*n2OR#lc^0
zj50Vbh1>(<ZIAb)(FaT1UhTf?X?4~d>3sUqqx7i`AWeT+W@jpc{^(n)j6UVdcTq57
z*7Amfza)=2$R@82*o)J?upbV(_pxC=j;{-E0|&kLp~7DZ{3YS~DPVuS9uyW>uJznQ
z$JZeo!bgG=r%z5_{r<PpSKofy83Pr_(=UDaeC-Gl6tNmD@`#XOR1L?6?n{w-mUx78
z*HzXMaR}f;>7w{h<b6B|xdar?jA&bows?Wjz?2o&1T1-xDTaYyj8>7yuFzdYVJuZ*
zwp9(t`A8E)h9EyL&8Qs7hoUU);RLe*Goc~4N3dRINm#m~>$>NfB3naFM+Xk0jaQ6x
ze8TCK>#^L-sS5jJFR>_YD<^A&V&E9K4rRjo0L{!I<H!>_qby|+I`S>~GH{JNhp6ct
zN>=wdLXXls!U=6AF~}!ypunoLm6;wXZk{enNjwQ15?7fCLirK&;MRIfTJDNZNn;S{
z=G&pKK$#tDysN(#5t4SM+qZ5{w|`onp57mIeF%r}!9k~+gZ5s9C2fD#IFRd(kbN+g
zBOQD}{Jg8rQBID2F`c_|ak})?S5fHAWEjQ*MxCvV&GH01NjjD<Wn2|boE&{yQh44N
zIrqGL>Wn?}X6N=7W=$r}G5s&;-nn&ay7ALb)7=}lr;Xcprib?*w%?pTcfKR;&#s>9
z5>@33d7lz;G(5_+{s(>e!VzmYk{d>$E7Tat`@B3BFNGX-cQU+)A<rm%JmB8FyIq=T
zX2XZ@>%&KNe9fn?Kw5SFYe8=Sed3ki-n)n0Q|<HBNADECa(pAaRir@T?M-_D-^iBN
zVK>Ix2B81tYEEN(nO_Cpc+dNr&qtmNkF&8lJf`2^eCq{H7)}j+pTi4JbIu8|8a}~^
z)R{kzA}@sxFm8AGH3sFxH9&@}(&ZF`qFD&num4nIA|6pWAm4K{p$zKr4&yGvCVs&<
zVmUg-5=Nhk7sDs7T&e@^?%ms+Wx;vCF<{oiv7=;&;TT|$txRDizw#$PxZ`J>?`D8h
zE{gnTuL$?<Z*LD-#_XT)a1MQP&yC0uch@&3OUv4a%#2v&#hc}0$uhDXhmAd9znu9Y
zgDjqo29~SitoY4uzngydyYHuq=Pez{DaRQ$oG7D8%~Tmq1U%d9d;^3m%eLfYPy!>`
zlgFGD_lu)<$VHB3VJt>y<tB#;W46Me3wn~F@N={4o<`QIJbe6U>z>@@yVBOwrjmb&
zjJB1(PMLYJ(&ZK}TwI&J{_4tf>EeaTsrF*CG+o=(!$;HIdmGcn#&(VRV-KuZ_RR5I
z>iYbJwK0n?_u$g}^HUb(MMG`Cl7U-Wmi611?%cUo12(#GM*8lNBMo}8=M?#!F?{(h
zFZ5uX_TF~VQGaCl8VZAu!5R?rYh45f&yJl49dcA(_Sibx{ix2?Xzp82F4tD!fS#3+
zb6=hu6KFzV;Nb9GR3`GRZ~9K&r3;KbFo)wzJ2@Y^qCY31Wf~<cUjyy5WAc3+J$YCA
z8sE~vWjRqfWjT60L%Zc$&J+q)-+Xi(vsLtsgp{2Fg|iuMW`=RxJbE|`j{3hkwOl#5
zTYu$%@-Dp5Tba!aO;MVb*^->M+8f7?_tA)M<_Kdr+aC^|%BD`WFKLRyd8i$6?xL%-
zJy52x45>ac2Gn&0j}(%F=*$Xw9jk*Jf76~;A~!AlLOY}{z{v=nwvXT-F1=Dh`nBw|
z3w)0=(ICUcA*YY4rw$-UI_hI-XJ&EPA5EI2Px_5Eri_Ch8{lYfMEYIyI_)NzX-8&q
zmM*DS<<e%{$7}E2*(g8ZG?lh8bl_z0Q{QHPqMLeL+BbeH!Scb%b7RoMVqAv|UY!4}
z5r%UxjLeScq>-Tp+=m{F_8h$E%xR=;>2KavR(xtP*vkg3hkQKfE;Vh}GcXME&7n9g
zcmRDypTqab()42S>Cff)<zEJC4Lx!x8@qtE2p3K%%S6J#9(ef9nL3n5T^PVnCurKy
z;s|q3y>hPlCUzEO=%}B|hhD9GoW*7w<3ZYwG7a8>hYp(U4Q~sti@V(4e|#OhOP&jO
zXa9T4z4r-u`+$1`=9V=s7MPze?L&-LpqJ8d__MbMjE8g3r<c6v_MY2)GuO`a!2#|i
z<hr-~OL&Ph9@ZPZ*Z$LUZGU{{^!3`F`h54bfQy9Pcb&T=vy8sh{+4rXjhr$DPR#rx
z_q46@v9xPP@3+0lsOjd=`#G0e&b>eKciVSZnHwel`Zr$;fXh<G*9P%sTIiSKowC%@
z(x7sE!k(C9Lf>D!PTyL}QyC@w*LMkTAcxiI6`tRvv-6$lwgM8ylY;kxR{s0fwB8i<
z!v5=D1NP^?KODj@hreZd?k)S<g!h3%$JZeo!bgOawbkjj|L||8i&rm4VPGH`A5y|-
zzIW-!`CvOqpltc@QlzWQB;q3UI_mVx$Qbd7JJwsTFrHKy4mfkXNO>NLN}f~v^GZjO
z&B8_!>ruK!@YZ1#!G>}|AU4VsE<VggGMY_6@b<xWq@WN^7Qvj-Wa$j)aB?vUdBy?8
zfy1z2geC==08WyiusnT22^>Zp<*X1+SqsAGV1^mt3%;jLa15Pkr~*J3Yq^GCW@*C|
zVbC^ZdKTrSPOO=(^0cH6+$aZSq7mlBG1_oZ7<!D*Qe%DI88X!nY>X^ra3R>hPr9nF
zN(BKZ&(Mg$1<j8|ag`0>h=8Mv>d(kM&P4U&ojUX=UT7@4q)JiV-rSzH@7|xDJ{orq
z;SfF=v>gwl9|2z5-46_N$5z50R#&I<moH7{E?%g?i&0VkH;UUHer^2PbovP;V)QG{
zUo#vyq=)fNKhb|ENJg9MT?{akr8=(CcOLFMnEvoj|1|yd#~-GLn_KDA;W<&(DH0bh
zobP@;!s)YT+c)*|F*pGK9V7OXrGiv`?v)MC_wmKW>4mSApCpcW?k_yZULKZo(tkPl
zC^d)htHLLDd}%K~2Wf^M1SG@h^-CemcC__3gf#APT<*>9z7C4pINb%W_kF*2#d}X;
z#bZ9c3`_cb^y+arPo74=GpY`wW%_u6Z!EdSci#*ijxL76!IF+l8=D)`AOHBL$|elx
zjFJq7Tj5=I@7=2bx!>s^fS+*?khhGkw>Q-{-+tR~x7kf(E>17YF++=?k>S|>IOLTJ
z7cSJXM%Lq4v5cLWVbV0?%;<Ygl<vhEekVPS2E3HaVOGX4e3}_$=D;XRxKsG-gWktK
z@h&qVIHb(tzjOCajhR=jUYq{izx#LPbNBAtDL+OR_+GtswFc)OLJJN!26nh`ERdlX
zsm(INi|u)6U;&xa;2cIEaTrP$=L~}`x#-N<(8H3E0fPz<evv(I-MUk{yNAAHEWDVJ
ze`|XuvRUeHHaG*{388Xs2+SyMCXUKiSC+7>3>Mjq(*r&6?=C+WntInYXP~7c&!1Pm
zvz2+tvz#6FcjLs=j>x@cb1_=2=iY+{kBcWGI9#R0AZSl0vw}FUv;*}CZw^aNGE4BO
z=i9gMO!w}scNt4dK{8;Ym%3dBj;VvvGwpK>hB`Ck%d`P)b~tL%F3iZt5g~k+Mn4&$
z-!vg*q9<s3)dl8qpNn2{J^@~wKW0U7Zj5%8?|I5hZzwOzKeK}Fr)_DE2GlLH+u+qb
zol;-oDT{t+1p`YIni0ksYo;T?R(*BmWoX7>2`71a-+;HyGU0SU*R*eCAQNb(b!z80
zoSvam>W7SAPd$BrPEr~L2jz9HKS)#Bbp(eN%B1}>WLhqgq1UYRWLElcBro#3IyO58
zz3>L_!Nnj#^|Q>Dp)0u2L3~M<!<%!#`<~llO#jobq^*89Hw<(;pSbq%Gb>R4R<>s^
z#u9U8=BNv-iVT+OOkcHUY5GH+Ya8<D3^H^|{_09xe8V+*t*j&--6H+Qwfd9jkavc&
zMW1EZ1@57Hmup2o^&5c8GQ%vbt8E;<@-y3uQ<!{};)Rc%E&u2M%jAQ89G&v9Z?VAu
z!+E1!tC{FaN&7_Vu}0=R(>^$*=vvC@*!G<~hc5eN-lrqekF<G)^bxOy`sKUUXE`SH
zs__;ble2y7%cgy5HytoJmxBz5-W;ORhCe+OZP5WpHAVO{7`KMb*VO~Gz~qZ=<C_8(
zcf~}5vCV<|_`sOq%dl?N>DWtJOuT9N&YAtHKGmyPyZX<fr!RcAa`lX>`oq(dr_WMW
z`IEU}fcI>stv;-yc$5G>14_MdI>Vsvn_-FLxIgTL{nvZpp!n|!OB!#R&g<eI^!na7
zJ>MH%fjPZbo{u>D^VtucFOI44&dYau-1q(gP4ZRn%5s`_NBX|Q`p4)a>_V<zW|?^T
zY47@NA&27BXY+Z^YJW-JY`W>M+T7?zbNcT0xSZ$xhWUHu=eQ^2-+V649M3y5Zt`us
zp8>e<hV+BI=XkphMK0TS9;B~rhKJ)@27t7W=}&nMe|=fGe8=%S&%_ZdWxMoOyhGZO
zho|%h%=o?q-U}Vxri^>zzHc~$4+jU)SNLm%gUY%;d<-~rd>z6ed@M+xK6mBH^xJ>?
z_g%(>@IdwQfB@}7?~;J(w3|#vF)<R6<EhGhA`k*x=iA<R5&ip_q&QKGEO+xH0`E|W
z$C4pFxb|@%>Q))dN1TFEr%VDtfs*>@iKPL@@*#`?aI2u5vmx{hqr(tPq9EnNPQWH`
z5D+Lj6hxy)eZ{K)DrL?YRJM19!;2%x0AeCOp@1?7ZzZA7G~Xx^rN^vjbZWT*C$AGH
zR;D#e_C%3786lsL@kHJ)Y>+*a2uPzr{<W;qc3q*r29PKqxo7!9hH6HeMai*@5QUUt
zZ?@T>*$~(olAbzKc8VCZ7+1ieTy<%nNHs5_V0Sd~v1Q+{q?_~zNHq!t52HA1oRUX=
z@YOOX(siJrcj}BWeQP}m^Uhey{SXe}bAn@0y4S8;nl62PZ92PV?}k&|qpnM)q>tOz
zru`&sUjNhnIm2s&Vh9LNVdUuEB1y0OF>RlN**;#33moL9P9)9n1*7P<?yXNZe)wU!
z{nPd7>Emb9`EwW2htsc~Jew|E3@(>1*C;1GB_D4nPYE9j3gv<NqE1}HVDfAK*5`7q
zFH+KmLMDFD4-WViJAfGsW|V#R?YGnKfA_l@-n_@Kh5s^`9Kx>xpVaY%mxopVeE2}{
z!IDcNzA<ltK5e|0p<Qv$wLoS;haUar>sLbJy)*22NgDIn9NrAWVI9tqBO~Y-^rOF;
ziEclfF-V3Z2CvcQiB@KSFs7S{Vi`g_!g6$$1(cqpy(}rkILav@E#GGbNi*RY^)Fw(
zQb+&2d-tXrH*ZcH<eXMD5T88WeG$nLb)biNnALUt#?3lW@IV85%xK@(*vk925|<%!
zC8Z+&83<$+&BNkOHsMIAA?Cs3>G2~@h#@a@M$YzDM<e5Jm%{_U`s%CVu=DW2bo}JW
z>Dskx)7M{JtHa^PAOAFMq|BZD5!uv?4<LWeGE~R!7_ad~jyt#;y}TV<?%v&)?%clH
zW%?|&2hWhK#&-rDvp+UBwx@gRo726!>n(Hl3rk!x!#IKD!C-E-3qvq@OgYWYVGx%W
z*{Sks=vw(U^@R>*baBl2mT(@NK6SbVdPe_+pGz+-qefv4sX?Rd<aKm9cWz}Ey3P8k
z17<As-HqqvoAT%Yz8%^Fs52~epX3YhLL;~_c9DHKJ1YA|w$)ywar*Shwkx=Bdf6L|
zBZi}ijLX5JKWV?(E&6F=iQf9*sPb&_CaZ6Cq^>F_N@M<Bjcer}eR1G0z+fxOp<TRq
zzROdZAwr&K#J}JDstg88#N@;IyLSFuono9v-B0MLa%s!M`NHuF$vS8C3H>khkUw3A
z)1yw^$N(IHaFjfhI8J@5(G5-Rr9E-nDwF=!GDQBZ^QG;43|8Zyu+*@8)sbZ~Ek|nh
z_}1qAq_xq0%Q)Hon?4S2lD~=@ItAICgLXKGrQ_QT|074p7@==yp-wt$EA7NUJI-+J
z^>JhkiKwGxkRN$pd*VdlpwOo*U8)Z3^=6q&j$H2<+~z)Af`jZz<fqlO)z%eWQQaew
zklB%)0hTf^v#w61EpV`Eua@@YWJeeL3ool-6?02dLGLb$n0LBo@T?<o#C9NKp394F
zg3lS8s4OQ>Cf)RVd_=vGzvy~&GJQoFXrzvZlg2(_>POvK;bP3}xP19i9sL-CK4D-a
zz0Dv)4r;Po+sd)gUe!B1IkMqKmcn<<P-T2?(@b7^fj$l&?HApukIsAoeTM;he5c4}
zVEHnMlfULEuDueqXU_SfU)bY{4mg~z-r)>|M`zWB4(*=_Xg6TUwgV*WIIdf+grL_v
zD`Uc`_zoV6Pf1ff%1b?!hommhSz5}c?c)cnv$Q>RP5<={C!h?(_gopv6Wqho@lkb{
zxcavC3{COj(fFZ-d^s?;x5v_@`bf)>=lS3Bm&cbdo{!@};g`aPNaJ1Mee(Rd^8VcO
z1zs2DZQp+zP}hD^TWlCh+IX}Fx{m&(9TnOfoO9ZQOC|KZ-dkWvC*N%Cr*WjI&mOCD
z+4p$?J$vMxHx@jkw;cQ2{{eH0Y#Ix`<KW4Bz58#JbBQO4G9Lu)CTf5!XUIZito!7V
z$W`{RZ2!u`ek<}LSL!#N4XO_&72m8OgX)WxfE`XR{IN5HWicbyJe@A#1H*2)523W!
zA3hWul=ea4uMG|=<KFPOfS!5i_&S6`_y}+`O7OMcd^i2}?|wh6*z`DGxjH%`ko#=<
z=n?1vrCF%qH_xz)D!Z%+Ik*;+0i~m6B0Tucc65pmMQ!H+WpAq=TFMGxmvT0gW=ld0
zL0uXI={kk-ZkN$XK2^AKO`sSGvs}%zvb^9hX6MBkuL<cs?nVL<(8aW;4yDVplr##(
zv7zMUig3?>Lzx){s@b`QLCPihw;UBJhVr~9JtvOKJBmjQ6e)MoJessgXH}3REJ-f~
zNC<AvpVZ)Fq^Wh7l?mPiAo(+{RADu$GWjuhxE8+(O#~M>pnEqiPo9JtGxp#!9DgZO
z;@B?%4XPX_edSajqfcv8P_Go4DBSAb2yl+NbF1g7#ApvS<~)m{6<i2GRoKF+RXu~g
zkwy=<#<DAia0s6bq^+L2d||rs%~#Xv+WG11`L*<q6`iJSo*~6jT>7RXzQCxWU!OXC
zx<*xw^(yG$N&3F!X2w1SW5!kcRCt?yeKdV>XM214)4%@fbnn*f>CyJX$ue1Z$m54k
z%C`(KxNz}e4OaRVe!<Xc*(?sL@=|?To_OHQbMU*WfjXVejg{nF{<NS9afv^aG^20r
zWwSDM_Q=|~8hNgS7K}bS4=mqf@0~;V!0_oEUmp?V_Om%QlD7GVpv^Y&o1rniGA{oN
z#`8C%)pzHgc<&O%>wUK;h2`#_!^<~7j_Ib$=h2Y^48I}Mkav8u&Eznv!t5X4-n)11
zP8%ERHCmW4a(30eH~5BG6OjiZujuQH>uc%boakg7qw2{uzTspG#^HPG>vgP<&Afx}
z;6Y@%?kgDn!uWkFeCt;7y}$jS4lc8h3{o-^jj<j7!;d%u%nq{L7bnnS%p_kKk<ZA_
zS!T8kW2Pk=$zS$oH5-M&*-~0&L2=l;Nd64y4Ce6Us3X7FgOnkMOvLf97<KC)%#RZY
zAFuOcXE?of?mw(Ch{KD+0N#jhwjD!JXPt$Ao{X%7cXOJML!Ujh?4RW&S4%(XZf$Ok
zXQ7iN;LJEtcQulgUk~}LPOO}(afW0w)Xa|QtPt`bSu1jzI%go}fbtDdpX!N|j8l+|
zwYqwGT3a=1mqRSa(`Gk)*>Bd<XO_!k@O{?$(>5F}X`~!=J|+$IE{=k0TLK)1+<Ls}
zcGgtr&5&=2H@I<N<n`pi$@idbMZGGUZxLFG=e*_3Gseftk*O2(?D9$pK26&8_<j<H
zM_S9bja=+=CLP};IGJ5^Ho6FW^O^25Mt(-Rv4kOKz_ETuFJ3mF<BRTVc0agw&zK;O
zmK;3Vll(YM(7sL=-|6I~e~mrMj@HT1!E?dOj3)#8&gYpnOPA1=(Ut>`L0-DzYnOG%
zaE7K`xOe5!rRn_n^ELbr$JW?~NMHEZfBi$3RIHIdw9&ScS!u0&<O6*rShjtmr@qFS
z#c?w<EPZS^JUSaF?MPWpo+8hGk@jZxQ_@R$wLQ+(%J(UofrHwFy3+r>cb~IZe|U~g
z9Lh)99xT^6TC+MF?`<`1=$s!NK)W+DN8RBa($TKZuC6q{A-}vRkKpK2K7$ieDb3rA
z%(m|{r%O9Dk#=>N(8f95)4l7~H>$_L?HcEJcoTd)8;&%x*@^PN!5f0NI!#{MaNBF{
z>T~EQE@v+%%d&SM{K}Wp$LO#6mVK$=N}keJ?5#(Z(k}5HbmUN!;iAU@|D}iKyE0?0
zwMV?**s)nR3|1oddy!`y*c2SmTpy4x94z}-nj|`YqiuL)Dc|DY)9*OfpVl!}RAzvY
z8}rR(r|){dxsTg9Z~v@5=o_USjscgz_k%;vv$jvV`TS1x{)7(hNdDyq=$hv;aDUW;
z8D^u5FWjfk<{2=+0Pp0$496Mohb97gscL+Z<4YS?51@IxEX92MI56`5xbU`liuX2g
z-xP0owl};Ee(w3Zz(~un7vtJ#g0_H<>n}OYy|k*dN#8C_TvzO=_R+uL4C*h(p1Yra
zeiD^F-h*!x=6B`S6mm{ybF$>)Fd~guy^(Y6d$}&obNb6~#Gl`pVYwRR9%(pwu(Wf$
z3+^3oVIWNhI{5a&DZvqPIY*<>m*;ZnKY3I~bKdDJ<;VHvSN=&K46>y~GThDXRTMus
zw0xflEOdX@@S)(KwD#xqzTrc{LFs?W@Fw`V=bsBiDmrw09l{}e3^)-5^XhNEpT12v
z7Dcg33ig8;<tra}4yTR;juhb|N%`r=H6j0_EDYt0ltBogv>lJoV1z8^2|<L@#D>QY
z9@;243S^W-!mJM~0mMg`(IYIXc`P4kVsK|4Md+4J1@sCyMxxD9O?Vn5*yupUYDQ}D
z7KfUA!ufiS;D863+0}g&k|(7^&Xj})4+mc2xXaNc{p8viXJ%j#7M3;BY4Q}JQ!A&Z
zFC*;tMFM)#r0_h;B~TvE2>g^QOBud+{xky0PznhB^6*0x@j#6NGyTwP7<Vkg$iNh=
zUl90)!a-3w8U@MHgmAKK2l7x#%$P%uTu8La@^DIWTs^OnB({;=qh2j>GL#zvAc2zN
zuAbGoyw$C=;BOYoP7O>^i1`x_;SfF(Fx%_uci&7`u3eif8FXUhM2&v>OC7?G9(Jas
zzOlPhPLyuW<I`vC-*u{eneo^1U){Un)ag3e@tSl&gU`z^@G*Ogt>3yc-TLvT=~0yP
z>G|l>loUt2@zm*+>B^PM9Z_z8fxUCaXuPB8&kP~y*Xf(x#JzkYeKPOjVPQm)!70)g
z!7xpEgpmznB1MNn=$Mc^7<VpTxiVe7a&@|N`C^SzlpQlF=7bL6!@x&#eEs7;{^Nhx
zEBwEUfZsMF%x&qQ(9hOe{4~D(c)Suute*`(2m0D85tH`*(BIxT`$4qv@0W#L7&l(+
zLgQ+bzi0go{KkF!$2DW0@2|dy$6UR3rOUgmhBuI*I9k$`!z;#eV|dWxe9t#G*JnBE
zS@_ggN)T^h1bG-ahD@ZN8@+tz?!AP&5UI1tEEYys&J#|lG1JA2f-^Pb-n?}?_Zh;;
zW@dnGO}FpdX*_(&zyZ#qJGp0XM6y90TW8NsYv<R3_lY_I%&2<y>`C~}rRmz&W_GPi
zTla@Eiou+-P9A^w!w<tLw=?$M!N1%$I0_z?@G<+$jGxMiAumHWUgw-cffLRWKaAjJ
z$uV4y24f(A_(P^5^Wo=99aXl?kE4n+1D`y7CUQ{Hzk8QcZ#XF!m6efER(+CN$ysJ<
zdCxvcmOo>twq&IlEDZXL@H;#A`wlq?AD5PIQI~EDUSv3OnD3GDdXC0uWSKnYWHBh?
z=bU>Ro7*+GbCSaUap-dQ?s{p4CJg`mW~7~QBx(m7MwMGb5#KNl1Z5%zasZL*I9E6=
zIZm_%4kzt&aimFGx=R>sgd=RsOl+Mcs7Ex{*2lL<`$uoaw~=T2z%HQY@##X^j~QPb
z<U!X+8beMVLEHAOz3Di5)d7RNc1S+9x<h9(rcOHJE=P3<H_qL~DTuaf=`+$W*iHN6
zRFDTpshL0e5CgjzfgD@f6KB|^OBX`-)wUH*X8EFlSvmGzGn0`+>*med={uXP1G>nq
z+qdiV(s$&8HkHFekH=`1INGNA50aJhTW$u*;H~~SUdRUe2;GJAr2GAK&|l<*yibl<
zW(0C@33`IMu3pl*8-0ianQo>Ymp(_%N1x6-3uyoL#ai$Fahub_$8;CE(UCeQ%v6*<
zx_O^$Vd=%nLn)JY%)B(vhDwzx4KBw^`>{V-`d0d(zQB=Br-+u)dZur2zS?tcWAlFV
zvE<^#3m2Q#Pm#0Kjq+$<@J2s+R_PQR^EbD_VMK1(>9x7mx9pp0#vk3(phIyus^~zS
zEs^@GUK>31nX8ebym!C*vOQ?q#utatosKEbWSaBgc%I6e_1Efc#b=H)o0?q<@)}NZ
z&aUCO<@9o`zCD+OI;!4XiX@fLdN`-XY)tK?`daC75X|}akGz>M<1G(-5xqRs&zJnv
ziLxq-2FU@p3=pMP4hQ>K&Qp8FKeS=IT6*OniR-+&W%8?D96GrVC4GW3m#n?;#o_eQ
z9uq(K*niLf2hGq*G0HjU2^^R6cLQm>Oy~8m#CzNL9~#~^&%N>X;(Jh7;7!lo2HxkL
z{jfLhq4(rw>v(>aU;9hiQSB7kYJ^$+;H6V5q33D4X@mNQc3FPPZZY%M#!UeFViFqX
zeP5rJ=J@vYd(<Fiez_-jWxVD8eBm(9fQI+(?cKf~$ji(1D3|B>)ZkU)+>0Z=yzt2}
zpj&y;cXHExc<PT`ktgLWUc2Q<AIX2X*EW`T`ez3Lg<o`Euco6zc@Hdk?%w!SV3!u3
z0lY6Rt*brZec_k9`!V5j!*eg~_lI8#4$5l@?*oU9uR}P5j{)hQ=dN6uzW@Dir%PX5
zsgVEpMU<L6B2-jp$VZPr+mEdeWrB|<IYR;*LckImojo-RAr<6ve^Cr6fD8ckC!p96
z9E|)UWE=hKL*EZ<LN@}PKf3bIr#|;yR3SKV>_p>G)}BV7p%5_QSaz)M%-?7qcS;wa
z92?C{7)Zt~ck>N{;)F7Cx_babIkt44<;#@eNO1HH!HE)HB{X4VG7BaYqr^wMNGUl6
zkEG4sT>&Jx4uvb!$a$BBc=jat4Ch{jwFqL16HBPmkdl;hqvmV`-zCa|AKEGBqQor*
z1lNoklzzBd#s-Z+av1TtZ^6-y9Hi7!HYiDq|A6BVafV_&lyw`=FUnehV3d5}hYVdl
zNVy(An;vdI433AMuS57eK>t`-Jv)8-oA0OdmoB7FMqz1Rh;q)*lJoS`@PTlc7ejex
z{5o;+R26UiT3gQx<uPYZpYC2V`WdC2@=wvwPw}NE=}Wh+-<)p!>Bs3|cwQ21zpIfT
zeWXSKydv=#?mFT=eUWj~lCTu}8U+)-dF0)&1)Qe+4?otIhoXcmorPc1{rk366i`Y+
z=}@Gk8%B$SQqm0~F+#e9(IMR>-68_gQX^&L7~PB>H3nlejEDR8eg22Nc73k*bson#
zySdXoM*fJmM*~-eeyfI=OJ<jL+_^LNUjDDHPu5B}YdH*!XBi@13+-adJw9!M%?Jc?
zn&aJqN7eQl<=RPaXCtjJg6x##43M?4S0svoSv7(D%C>KsD4Ty!hp2x09C8D5D=cT;
zwWVSnpfHW#rld2ed<sFU1zxNM%=Op%XHdV33|1-`jf)%1GbaooJg|(~<84cwyG2$F
zy{ELnk%XKj$Y6DF21{co%PTW#JVc^z?vEnE%V5v)evshv5!MsmPCta%zc=;wGG-ey
z*Pq>^WFD3x3l)c!_*AXpTRi4qv8r4N^UGjOJF!Jx4fx%58@RruQE<6k_-tf%!pL`4
zl{dmsfED}|mk99$662G@myUXc$8XS0%cwx9{#tN!)MfYMLyL5uCXdxICL=I?J7-G7
z;mQ`_8sY2asqu71|F<byV1WGst#OhdW7;7d!_IEo14_?-I|-3`PMk98^rZ*O?wr<$
z*iXGSr9Tk}jePwk!Npky-E^CS*t-Ew7mMHe4`a{i@aU3ktX!LPu{H(IxM#G>)8DkZ
zoep_iP8b1SFN2%Km%;U#Ek0Bj*xS0AV4C!sVNLM?*ycTbDU`kb;Y6@E&hiAhd}_z+
zlf|zj@|A@{;a=a+Yg<J`!o-{~zOmC2R3=3tzix|QK1}om5j1Kb8_!d!8wUT_XqmRV
zNF!}Z`#l7H0~P;NvRQLhp)0jt``BjW^jm>cr0itQU_vJ!0zAm#8LfRu3r%RzD+>$u
zxm|QA3k9XnwV`7O`#?+2CBF@i)47E6873S<4_^jcz7`#(rB^b$FK=C#7%%Y+Z1r<y
zZtz_wy@UC{_^o-6%nT**T-<xv8e{x_)M?irOr_)f6X%a4Pa0X5h!03R2j&uVj6k)-
zH>o_~<2O$i9s%6-Tn5gTSBDCV;xTw`ofwF;Wii94Ov9I>QXIpk)SRwe`#~TwiFw;(
zfhe-stD!~MM$i5{$93=1tA2)GmVuK^F?8aIJf^J_G}+y$ehc2Vv6L{5DTN=p`$TbD
z4thCfe)j)jr;p+>mD!@sr3_QZ0%kKIOupur?A?9N!0Fo;dUnhp?D6eAD030ST<Sk;
z=3SnGXb<}mV9F{}mr>`28IL<qQT7faU?`xA*22?5J1nY!bRtzPUK+LLud}Sc<RL-h
z`~!c!8THgc1$xWHo)Co;wUzR&38Oz6rCavpQ_73w>|_kZGZHV0XB~tFq%a}rs-N1<
zVBDo~RfqyO&e%J@sYsB-)1A=;hX!_8?D%u5<fWLVxvr+7mBuitw?dtp!h1G^H!!7D
zKn}seW@h&g^)Gk@BTAol^m?$#kqEAhQs6G!d5OD%!kGx9^;z`keA%vs&3~qJiJfS_
zySX(HH3h1)G0ZMRo&A)t!$W%8QN&9*(-N4gN=(iGm}HgBYc_1RYt}y+wY^*-cN+(8
zWV1$_E7Y3$MDaW}ZTFv3ayqBVLHNa`2NHDJ2JWlzG4e;}jQxCq@a^|J;(_|wxFq|<
z^!jnA4CkHjfAUM!%@-2uJ|Fh&B;P`)Fkn0SZY{gY?aT4H%lo(7i~jAe@;-zlJY@Vq
z{QoD*1M~O67j1X5_b2JayJNU+H>>z_9?DmlAxii%KP=OdBPCu`KV^Uv|JCOkbB`ug
z$bBvJcn<&ka)7Kyv^ln~!+7n~PI@F&;?esT+4v07t%@&+gQY`GYBzKr6wz*7ODOhg
zsdF60lG{jn)Ni-rX{=htJ@t*$Qe<lSjaxW>4FOB0lgO|f&fj-S35Fyl$aH^JFT3Dl
z?H-Ti7z2c``l#ej_;NkE{G`MDsf2D!n>ZvOZkFzAHzPgwxFvO6;Z8!9u5CujuFqp(
zAI}fHt9Bd|Vmd(tQd<s>*SGr(7{X))DPrZCtuZE3sS{P66f~O2Q-&{I#n%k*Qc!)Y
z(VSC~m%kvy(H{D^>MQq|=G}8wAp4BsNuf+JF5r1Aj2>%=DRLtYk3Z3*=B~wfEG*2G
zF*4j!R&Fq${(TC-x%1iy(Bb8^(AudR8U)UO76;EJhQMtXbNUI>Dg27{PK5<EkDGZV
z?6E!>pM6uoJ<lNqDwuA9(@^7CXH}W|`uNFZI!0$1d=l#VmOU$V^ZkKsj12MDh;&4E
zLU(Qa^*uv07@pw5Wqd|Rd`S>SIjWdBWV`{#znn~P$oa(AoRY_!hq&z)rSeLF9H*EX
z?eEusv%sxhRsL5AGB*|ei01+PH`Zccn`JL2IlPFb+PwCMXW%1;$mTE>fbVMBIeEXu
zz%v;<XHphEwsr!1iAUH;w#(TQ2J)Uj8fHp&;Wy84BB-en7R*oYft!eK>01eq_Fcfy
z(Fb6;fg+KNfWwfgV`vOve76dZ$zL(kO_M><yQ|avNu;s2U;YLy1EF*d-C4cBb@a&W
zJd^J>E%X7DDge1Kz5WH341@<<pJHyC;i?%troSzNlET|vu{SMHNkH~zjr@BN=S&A^
zL{VGBZmGks#(Gbc0r%&C>!R`E7U+We3n8AORndS;+GvO|k=O@qe+++d_=uIvD;3iu
zJ3aN>gU^p}z3b9FHZ2FcoWU(GO%A#RV;ilw>m83*{wcTNbAyT2b<+XwUaF{V9%t19
zRY>6Y4fR4a;v)rxwK+Lk5&fT;>$1A_=mHIXu`E!zWf*?W33h37zxFyuY3bA9`vgQA
z`bO9HH9wV}Yn3wH|E}EoZ+LLgHh$(Tw$=4SonA&o5)*<4S*n|T0ZiIj8}M;-C?qw?
zF%wVO7;i{3_WbrSv_6V(U(Com5g!Tr1y63r7e&rX$$X*~hK#R7eg&*gaDWM=lN3f(
zKhor6#N|`^hu$}z^5;qkrVlXFHk*-+iTa8C=$CTT^|I_2fb90B0>xB`Bn4?XbwX9w
zm0qedz9XTo5Uf8D1B;mt9H%qC8TND1jc!jBf_hR*e@hfvHKS)kVz!YIlL<0r`A!kr
zS3v=|(*XL=NNVMD)3EalCFal)N3Xl826vsk2N@-ic?HAuct(ta3GOdo-mo~cq9vOL
zu2Y*{M+|}^&|HZWuSPAJB5!Q7b1Z;PAapU!F+>J{Q}<x$_%ZR)9x&~+*OFnLoU4`e
z8!H1y&a`r)sS2h`UzA@E7gtKfI0&}hurj8j$}t>T!lpNk4pdf;-hWvpI~B`#`hsDq
z62i<{b_cb;&3E&Me2>=`oZ31a`t2^1==-}<oXW=v<u7+k)W>!{?h#9S@|6;e$G~Xg
zbUvT1EFmz-5!6+-gSM(*?SSFWoU<<9jO?9v)JB!!ck`>|oVV;<_X-wYTN=*LaKr4^
zL??9N=)+YAtkIuiw}5|XT=QxHU}lQYdPJkcn3mDct~C)HBGhfmsNUbL8H<)|XKi?H
zEmnb6=SgkZ{lGUXKbw82HPmCb+R~shn{MqimOwSb-7H48D4$N>{@~fd0*rgB@EI<O
z4lPRW!LVvyu%lX2U6*)aY9s@HC-wm^q8NtPtK|g|4S#ZGh5~`lvDvEy-e-~C;ohKg
z5c=Zz5sz313wGrQ@dtr*5>6nO?5O{7JSB@qxoNnlP!zjCRS+yQ>y*^AZx)*Wy;z>Y
z3v~*n-B@t$Q!%*uhLVW2gkA-n8?>8q>!WrET4n>1@_B)i0}fIeIcG`!Y)1-kyu#}?
z_n@sd*z}_jSirz;<4%yQ(T3`e!wq+<&J6X=Zr(p7Nx-`&xL}<0liRh?Kg~RWJoh($
zc6ew{WJu^lK=zS%x7@|kNWN9IhB=M;d$0Wetuf}*;O`TzYl}5yC@dCLMxbAl$Y3Ks
z{^@;W!G}vmKXyqG&bN$)ig-PuKplE5&57T~>UQ^z)FNp$x@d}jL6Z(3P)BqnkdWa;
zCi!nWxNvKVd3ch)X96Rmit3#6v*(!*GaaSt>IAX1jFVs4Wr$E>l1I%6;@w@K=uH3A
zZc?ZFpO)Sl6XM0iNuCF_)qOBva=Y($5PN88MEoX04i(|v4ZyxHTT@YUJ-0UlmVd%q
zTtKC(Y)Chp^yV(Ll36XqS81h8^=yA(Vc}rH_HjN<YYJQF%*>fhs%vpD1b?=?<y<IQ
zjXvU*s8~Lu=d*zZR+qJ?vcTa<4m;h%?;=Wns&<XkCdjW?t|2obrtUZ{W~pIrmf8tb
z^}Nbv{<kDj+UFG{r?Rluk{oQz<Dt1qlUKQ?@n59wFI!WxnPLHZ7ao_Dm6K6N6@J<K
zlPL(xmY2vmH|5?VenPxhe!}+@WQx{9qUA3v@oCef%f<u=1<P~`eR8nm*uMdY6-;mE
zjYX^XMn6Vq_JvJIDCU?=@^{2De~#tQ4YR9QyD0{L)QI6QTW*DGvfRrzk&lO9d^da9
zNBb$5@s6Qq(jncNkMRV7f1i<fvKp-q4}bTFFf7i{AFG;%-I(9)2kwp}$Y1<B$`IUe
z7(13#HntbMBEVs%L{gt`IW#OekBz|H?mYQAv4Vd}S}99ZXA&pTBm5nWXx&kCT-u}=
zZ`fI`wN4djrvPkc41AmZZ1DrrhKp*JLg^h%m)jybjD%6?r35{LO;dQ55$$B$z9{G7
zT-(rifG(Y=jbN5*AC(R^JqPfBwT?W8dJt*ZTWiTm9dLQ7KU-E)Z3?YVTW@V_B*luT
zQS#)~6t>ps`MhS!F{GL}sg`O#&}D@WnJ(yFP-COG;t59xH`BbSTtf6OT`rIXV+JB5
zGA4AgR-FAV@$|$GZ>oniBc*xYLCDz>p1@Tm@7WkNAZRUoxmOrXqBQeQC8Jg5t^D0<
z&KYqE-}K@yRQ1L<zKS9MvxP(gMk$7R&_F*OjZO81-p^?ozZ3>6)XqYYwD@jI1+j0%
z)F$;YkMK&r+9V5nci6`NX@o1gPt<{^W6;$0i(xMhKjja%AIXf|4oa*IkcBE^Z+m=J
z4R0NQIDh^FI47J@@nAyYuX|9BirLxX3r23F#lNK9;PcY0p20~gH#he|rj=nYB<+Hs
z-@81CbS-faI9()Z&W3}SE~j&3fBezz$7lcS`*b61*0_kXEw~t=297HorSY(K^|FDn
z^e+EYvFH4j&Ic*Gep%pqW}tQT``XR*>st*!O~#xO!U&tcyUtxhBu5)16b*7L%rg6t
zbVj1iOS^S`XT3#%wv4@*zXFw-G-!*93zm;J(vwUar3`-PwDY_Y`=N(&5M|%L4!~qK
zDP~Yx>z3^ickfzxh>Xd%hXL(+>Ew3(>@xJK*@-5>RPV&Qw?d(NmIo>#3ue|=Q!><j
zzu)=pcn=Lxqf8cM7>sDX3+5P%+ewa*&>xI?uk{!WiKefHwfU#U2!8)+Z9}na#r%zU
zX@rR$8C@JBsLd#_=s%(ZJ4?%|UmD?t1t;@HeDxA;N9@<sOmlF^`2UkQV!Ix2XQ@l^
zQaFE%rY)IqZjh3E@ig*K{;>D1fv7IODYPBwy-zzmxuZ2y1$T*#a+ZArXp}Ck&V5V*
z+g~pU+zG5d5xwsxA(N`~>u)l5Q;(O138C3YT(2b`)-aQoaWu;mWXLksyrP>}$(76h
z!B1^>7+gBWeZFS9dv#1t1<xHyeCT@H!gJW|+pXOxz(i)NKg2<#4)C`ooTCHL%`4^c
z=GwQ6z?Hq*B|PjMW$BPBb>L`4wi;8W%MEwh{l&GXmQNMnf6!%N>zr#ipJnL(dl{oY
z0;4*JrCAR<DK@X0e0lFLHDr~A<r`n8j%<I}CHSOy6_Z@eT!L+hc^BW&E4%NA@^-Rx
z9Overs7TCUyie2<p=hK-glq|Qf~$kSjEV5PIC99dElF`C=}b)TCHg(Xho18=qH;+2
zhQwR|Lr3kkZ#cq0H}W^I3zC@CACS*{xg3kwP+?a~R!V3~02&c-^;ebg<6WRScn*1w
z?*-R$F3*YmdSA6m&`lsmj|z9?1^(wbxcS&c>UO+XQx3oF?abmaX_2H8e$jA3I+bz4
z@+OE(ntK@dm>HoAv{fK|mH92N#_KVfcfX`)(^W)jS1ZEigH%paQ)CZzH6RtuUajl&
zBR)w~6OgQ%%etbGcR8bfVS3h{)alP2k&b=!hDy!m7WC*yq*slCGMd^V;+?ETSWJR%
zTHNDj4BbV=qNG7*T@TXyksH<?1=c)(Xn?C&^LOC{b}0_`IJkM3;$S`uKCi7Lo*I!d
z%=L%YZKRi`ggu7hV%uEdZOG*bTduESMWbZNZJ*dQtnfI5F?<|NBNAA?!<iyAn-dc$
z8xUnC4`4~CE0ae1ChREK3M2#BL(A3k4)MPZ36zeqzdN2He-b&4c#0!W7FK`T4KpV<
zprF7Yzy3+>mQwHJVZn^xzxIgGL=#&){UTBufb}7L^xv(9{<&&=P0kOfsaSr%x38`F
z*sm{H%^=?ggDM$!aD)kSKbn8+M7rry+_V>?9L^V-^+wAV$H&(Q2t73RFs%}%cWaTm
zXCyM}3C)ZV-fwP?VxHh#?g|oiruU0y1q^;3R{nSN1ezx`n{;=$9}~+guR`7PA}{!E
zFGt?6XK+FL1j6(X*wIFM72i&Yqo{t%Yp!a)b-XSxw@!I>Y>^{ktVb4kv~onbad9<o
zul^Wm<D^@2@4CPecIGs+Bup-Ew6Sy5^!u}<MPL^g&sXR{4w#ND`b2{kF5%p5CUrPT
zsD<+H8+Y`3#+gs_*j44jmV8vDtn{%dLMUUQ!*AcT$z^_jyZaU|cl2~4K}Hc<_hvnF
zdQ`L5rDsscMqjdKNx_EZ9Ld)2LVd}3lT`n`794RfK$HGisCVDu(DCbU4e!I1ii4gb
zWd=*3^`QC;%?uSk%-?!D0PUOEA-p4X-;d3TDwx+(B_ls|m1Nb$)(2yJ3JnD>PUoiu
zO@AekE>E~yrt4U-*DZ5hCytBN@R&-PBk1M70+UK((enNg{wS{*MHg)n!r@0amDHa$
zvjHOLgZYzkQxP{Qj8&|bG-kU*VBcxA!RRW<ee0h=_;!~-67F|juXCr118695|LRW<
zRR8I|r{WdX<>kc+Rbp_!J&l>`1aFaAZu4_)^&4wDmsg2@C0LxilIu6<JkqZu=lB4c
zhBzV8Ja<uW=O<kH+bn(Nu%2o>HvYQ-TA-4Z?Y%?Hd%PbsL=7~XWUXg&zN`#Hdh_4t
zr|04r(nwwT_v3lss=Z658M{Hpy7zwNLd7?99=L7WKFt`H*?Ejfv#bilsTMP8)E+Qg
zcWIYF%=#&nPXb3(0rP)Xnmt&6Ww$7&@(!<v&+9!glh&t8S$#4|{)3&E$co!-y#-Mv
ztL(bX6O`dXZJp8Espk7Dh5HX?AF&(&0spx8n%GB&_uOh9gGk*-py;yBM1BA*|7F)x
zDKS#+ODN1i{5`SoutCw93zX`QPy6kHK*?3#I>y+>&n*?kMitbirL@y?siaS?FJ&*d
z`BcTr%~oIWX}}%~gxUBi*!By=Q+S%~vlqosdY2`|%lPewG;bguHpOwrHuStON<>ye
zQ_pUxv-5?KWR-e&yiMX>wkLjhmDfy1t9s%#OeLYIqgm>)B#tui%gAhN9r4D3k{pZx
zl8>q}@8HFEXQE}!(|ks7rP@zMoR`NB{pI}}=5;E~U1@G{;MKhmBU#`oE=Y2G-MYfi
z{Jc>i2_2>6B5!g@laNmZXpw;g$LGkMNEAnjdWbUd4mS9vL{i0w`JnS9>}R}b!=A&l
z<iFxPTiWR?U7+`>Jm_@A!<h21D>IJo*pQWVJ4W=clN#4Zayk+{gysAPV&vKZEt~}&
zsR`MMz#cA6lA^w$r<$1^Wn3GU<p6)<+z!J_e=`!Su-4@F+WBHd?2wV?zKD#j023K2
zteSrARm!20w0If4or)zbo<JCE8^L6M9SWWiNj@npjXm#3_NoLym5@~VmM7Hk=aOP~
zS2SkA`K>=RqAiPvZ~CPQ7~_Y7>)k<e{lg%iBtg1kr}pGgg**Q5!GsoF;SP1rTB0`y
z%)2V_2AV5GmrXqoenh0da+0cK`BmNkjvp1%xrcSE>fOtaQ?Vz(IIhxRqCHx(H3Wgf
zfAVqtcI)ouBSZcNzSQ75WI#0n&rw*&u4v$0o82__`HLZqksjWy=nIs*>IGeN+zJ2H
zS2G5#cM%CrK%A%)l19=JI^p3u;<K{W&iZ8_DWn`hdi3-l?3US_#)6d&F8qjG61{#%
z__RmJ+G1-ns*qy#l~=o9Gds7){U2Mo4;k?<RRpd2D*?9zS<}ME^~?tG5q9w~)LR+P
z0jZ%#_z6EvlNhVx)@k)Q{yQ7f7T_IEbf@*_y-&k_ifUTV)5?capJ@1!Pn@{r+RklK
zwj6xlEKkU0RH8AWwzeN6W>(<fPaRZ>g1kDHv}T^m_EfY;B*W7KehXK85Kkx`#`TfH
zuHXHcXsI`^^ze5C|Ft9gt5ipI-e~1R4{Bm!o>T6Q^0Ab_fEXrQMj_<pznDUUVr_g)
z8fwt>92rG^{#OtDK+CT0?Ow+{4Q^*6n46uXs0!|!-?#I1Vm<`Sxbpj3occxnB#6@}
zOgLoMJw2>zpXx2&KMO0WXVKUK8JXBGZYPla$<x5g4a{eUdKx972bs1OhDKlgzG1qw
zWX;t%*C+||dihR^5N8Q3F4%V`Up!{av*V>}{8%(xsXu4S^nQGI%`82U^H}C9FsIR3
zaZun<?nxZ~hMiB@JT3*FKOEm)C8NsOQMK85`PZw1=PR_*v=IHp&^>eLdH6f^GXe~6
zC`8_MY}%jh51kC9feWXZ3NX_GCAjs?*kHx?g_pKHyY9BK-(l<qr!P@S|0K)d-{QMy
zmAo%IRuMa3^)*_bt_tWnfFCL8uM^6}V71L2ibhI!w}mm`NXd&w#(~lS1{`&9s`gV^
zaojC5!+hs$^uHIRBS^hQxcE<SJBQ65DLs@QK3PQ{)s}9Fl#PZ_uF0oM+9=Gg)KSi7
z?po<(p;tw7$cBFhhz+Y;iV4!>e@y}&rl=i#5x_Gf(Fh{}5*hx9xc(IJyZ4onjn!Wn
z-#s7ew<`EQy2H%{8ItI!5gnbiVDaU4&o+UgU>(9U${v(@zEKpN_*qWX8W=Ng`IRT%
zD(Ks%$ZnkY-}6T!_*Eu0^oi5yhKFg9N5HgfflB?$EICM*4B=c-a0=V&7|uY<5wYc=
zm#u9XU3)V^?h_s7LXh*!!a@tcaO#EGsIPT`u4mJf=daRPG9<5((Dq#2`%H9F%02~h
zJT^Nx<*!X5m-;&)wre7odWy@sZ8qqo(YbNjs_7@oO*iD2@8sbI7-eX!?%GK=jCFH5
z%gz8U^skp}#dgIEG%obu14CB~UYab5`=Sjt=TdUTZJLr$M)ZxFJE?fL%_tdsZgFV*
zIrfxRbRSbGC*wQRN?cfyaA;{>d5_}$ydEzls^F~m#IOnAE@`-w<o#hAFK+0rjhu^W
z80Amd@|pN%1SF&r!}+ziNgoOvxVG}{*K<2J5D^5ZPO<_?%IA&;w&I5>NHVFn=#eiR
zr(e-!e48Q}y!NDuWVj;72B%*=H2tpnIk2AF@w-`pQ!-<F-`XEdb)=Esk>p_f)s_|4
zXumsVT{bMCTmC}d_tQiuxl~+n=p->X&GcRV9=ESd<Rd9{qkA1-B49dTs?dW9)JSCa
z(_4%pAJ(Ipuh?6Pd?bXKL`mvX(s6F)Pp+|`0}{{vqyVQi$w_#%^pTSvLUByZV(RT1
zGRE~(GPWq`Rx2IvO}cNWett&H0UEWX{oNp{|H_0f3Mn4sNE~sr8;Oez9uum3;8Fhr
zMdLdud<_`$P1(JS)pXHz@<HN_p=YmX<2z-^soL?m_aDg4_8I07z7xbzvxHf|;NTF;
zCI$ed($_af)?hpgA`hWFR_^CzXn3B&WlAZ7!00zSUKef{V2|GWqV6S1i}B-Z%N3rb
zw_G=ff<Ij=C+6fUEcqQiheRlx&dp`>M}EiHqe~TD3$MQiW$;i`Qb0Uo9F|w_QtKjx
zz+Xk;H;5h73J(bjIOz>+I4(u{1n%hO>@3{Qefx6s05wP+>mT#jpG*<T1fhnUWtDN(
zWHcQAYU0#6`$&7(d@(LyYeV=+K>6SJd9P?*d+_;-{Py<MNnTH$^go-ZW3rcdLA0@a
zE2w%?DipHEr_@KqP5msiOZlWXtkYWTJA*Sh@tUsKcM--Ey`(q)v(0qk(n<J!x!P;)
zykFGY;NGl#gZP{D;V;+RN2rd(M<8PFLS*?x1v1t(d<tUYoHvpNZHDDr=!c#uDEL29
z569qF*u2fu-Rgz<Btt;6!K)Ttq2rk`<an`~qVkz#eQsW!RtAr$)rC|rOQxNo6hZ8A
zv4V;{pLFuxf2WEFseSY8NgDaVNVM7pcSkp$)sMe%nT-1EuRzb6);jQCC}~&odrCLB
z@Y)%d4G2P|o)HMUY)$PRI(D~MMP(j7j8F_`M@0b*X_;q7Vwulov(ntCI%bnX#o78E
zFw-$A_y$mfN4}2E6aKZC;MSGd`$0I}n=tov4hLs<^ijeG!A(KQ=^%!yhtF~VpCUbf
zvQ%$zvx8Ef37e1epo#-C0{t%v8Xi6Fd;b($k-=>FUgY8MglnqRg`r#VF!Z2Z=<$D^
zy0<WMv9p9+fUwl{yG4d~;JDs6rm|H@(DCo$17z(N6W*<l#+}WzX5sRIMIN8hkI0BE
zZ#QgvQjQp|zb)S<nkHwe7I&>Iw|iHUl!jzDP#Q%ed;snORQKbW?(=~2Yq$m+biUls
z<P}htN(`R1E=g95V;6s<1oFq6q_Rm`u8m(YK{BQ7S%HaDa*H*T8{vycgpXm-y0h&k
zhLP{-&OxR10T*kKZEt+kll-&|2bsU;L$@Zj<aj-Y$jhp%(y`Iw`j9VT8=^;f*KyKs
zMq~n28uCNS#MR-Cw*6_W{9xI5<#T@%Ke`-BEOn^8X5;HpfsJt2dx%;3dk^x`b7_me
zYG=94-5^M@d{EQv+K(ecK0x3Z!~)8(*hJ{!7Om|V#O9n`)Q;%j;YmdiQeO5j7$NjR
zH7@YP7z3<Z6o{{L@gydf+xEAFD}m7TPBMbM&%X8|aa<7(e^Cz35vB{4k4>?2!TD{0
zzb04roxx5EXJWc~s2>V$JOVkhv!h@mbJPfl`k(($O@6B}0grdq{&5FGe9<-+Jk<Ey
zBkAPu#w=^XabYn98wJa{+3IoHPStOZ7-Cj@T(AIRpCh=%rs&T>iHq90<l!SdT$fp^
zxBWt5DWfIUjSnMyid(5YsNvMJr4>wsJ3gNDfDgmy@9F;<$8VvbwX@Wh#+&2)X$(yp
z>bVx(#jwRaiR1vN!<&-~w>Wpl-b*@Bwh059YFmw=wZ-Ywg=)p?U#{)Gp%zD+($(K&
z^t(%;__k+qV12tbD_yiWgrU_TW&xVH+%wM4yG1E(+~IlG09)w-`2Nn3o|iKNchFHR
zY}#+SjiqOWF9cs*pkpi@_35Oucl9yf4{x(Er%3?H$+L#Y)6*(DcL%m(kr)z53BR$g
zy_@A`;^5!@wQhUwht7-Y$=e#r+#;YupQAK=Tv#jIt&?6F#f#W{%8~Q7`LN>LmKxu&
z154VNlof#rUzhBf_qg=Ow(7Fcxetl=xY&4&{kp+Ws=xb&UtS>cGrCPFS0qD(Bx#Q~
zO=5@Q<X~h~Pswhf)Qi{|=Ul9|K?7+dHC^yS2t$sXeZat}{ar&z+a~@%|32wOdie|V
z3u#YPmvL5tEMk#_;hSx8{p~Sc+}Db&z_Gh8Ix0F!rlf;O@V;z*c(KmsG|BxiH-2SE
z+{8lD@DqEv?UH1&`l2J2vRPbhkmvLdCRo8MG!UDCOaG@Lp}eM<<cT>BZwGI@&}Jch
zZ|cE^1Sg3PB}0MYl&O(7tj*)zr4JAppKdZ}qyDb8uPTozGPWChdv~quVzAl2m&M(C
z6A-V~T<CpmpG(|b{LOp4GO~#SRbsM5x+g5xx;}lMni$STxQ(Zk67hmmJASFC8v)9w
z&C{u+lCMI8&o|prch!zWTmv}#V<m}0G}3r&+b*Wj8BRT1tH@E{pm#$B-DI_*xs;x9
z)K5qKT?g>PX;6DPeVJ)<9AdE*vpg=X<mU_@x}M&kIjS7rZbY0pAX~*o{%+oy=6uW9
z*a9od18M8hgVTK8mWu$T*)rU5JQyT>mYUSh(7DY$@ui^SA+OUf*Y@TaV1Ig&5NS?&
zNMlBZhgn3ZMf<hI+N|<P4Jm9_ZSDCAiL@wy0Ra9_kW8cuzh@mMFi7}qK<m}%?Bp@K
zpA!Bv(r+AkmjHdforrqUm+Ul_eU+LS=@8~v;;oG5AOG;t#*jxW$&!{HxNN#ye6sg9
z%+`}WQWPNzNKpJ-^;*3^9gutZge2tqiR6aQ-gk)M1s{%DtD-o(#jkDn`rv(3!Nn)N
z%ox@Ac;azQYk8N1Wc*J2F-0OH#ecWY_`4%rL%4L6Oox>CLr<Q=QV_pQ<~e1>dl={+
z;W)p$j7rm<<R1dMNRd5U{@L){b>uaHW<=dti3+tA6T6p9s;8sh@Vj(9I+Mb2L0rlP
zMit2d(TA$rw(_;8f$pj9=X3oN40||n`D_};5I3o(QNYyYNlMF2Op%Vu7qWrai{Cg*
z)`pLz`xHapLM!AOG+ScV%^3>)pFc}sy<+L45Vn2?nUdYHi%QF_%3Y#O{SV;LDUixH
zlak66s}~3_^w|F5q^{^8c(vnf!~v<n3dg4JdhS*#s%xts#6XxFQroa2$7xV0silQH
zC$mm|)4Thza5ZW?S<lWxW*jT*Q+SR!FXS2x$}d6j)n(rPn?JpG!+Ev3UD_*I3qGbj
zjpOAV?$YypMv@BcNugfrKd1?=Vf?iH*6^>cS&vL<@cD-4#CZO(0m!UnX@g7j^?aSi
z9kTOocCddl^2qFHklwuN%jhe%ij_-6EFEIl)a7nkg4g1_DMiZFjQm!+hm&gieO9XX
zIo7`p(<dc`t}3uQ2);$}UUpF1ydi!7@(gobM(v;Y>rZ-%>y5l**-xb0TuC}!<F;(+
zS$doB=ioDV@Tt=-oO!f61$QLpHIY>g%O)vv+&uplqd3I)t#*<M8U*~);&wTKnQr9#
z7pb|uH(fb7&8ZL225ziudt5_1h8}D_!ga&-4~+p#08pzSVZN<|uin_6%iW%T1i}S^
zhi*d3`+L)zlP&I*;xfj*Bmuy&93&SXI@@}vf20{<D*fYTa3|pQYLutmmctR(OOdZs
z1lBFkM5i{?Jzm?`F-fg!BNLZpZ29~Jp31+O)o5&0=mZc(*)lNjN^PC+GU@k=@ftVX
zxI3$XO~-AzmMHgFO9vnXEVOpr{g4rOv)xeVUwRJsCGl5iY*nqF`9v427k^|+oS!yH
zM{PbeSet}O-AWvR4Vfo1D8X92O`znd@}`^v=;a+=#BdP4Gy7}XocF*2J?NnX*%D~c
zUU{S~P5k7yl3)y|-VH1!my^(hpHPy;7g&TAwE<<30?+W4{3U|Vq-^?P_X8l0_CA3O
zzmY+6<R=W$e_LTN^iEO;i7JiE`zPD!s1&4%jW5S$J%VsYsq8<l%CB8{Q5Sm83<DPT
z%@!Iv95{KE^qNo^UcLd2o?@U0;J_j7RuU&6=0v(YTckRG%i=?xo*wdDzJd+zrwC}%
zM&GF^;zs8VlTYB_FnJ#>p4=|q<Iin;v9JS=If>CO=%0dpg#0RI@}P+~so%r|^OlWr
zU*kICA!-i_&lks1@<hZj@rC&xGPFaqe}b5Ir#OTdj~}0Y$bYR=@ewN6KVKb@)BxDm
zGQyoMaZ3Bp_G;Glag4Up4);aQi}#ooIhlsdL;!<i8;$Z}I9O~!7EcP7GW24q0~;sm
zk#2JoCb5Khp^}tGe8&WbMXhd9-)xeY?`A%!Azm$Iv>kJ7E@bqLIxObho88ltU_k3%
z`9YO5p6S3Q+Qr;$q~ikWtM~E%DUo%><;|2EfAM<7;?E-#LDE+ruPyUV0b+$J@U%7m
zCKEpq%c&C<^|9*}_G0R$t8jEnLVSp{4vi4KZMScUtwzt9Dixf`di*?%8a_eNNEu%P
zV19g27OV0jlRcm`)-m)Q^dYb!U>S5t1~+qCsCq#ffs>A{C<)aVxO4o>abEMJ)9b1x
zG$E^|-6%u+kcg4ytX!t&hOdAd3cRF6heNI!y^0gAU)pLfD0jxi<$FvJXGoi7z2W8g
zFkF6X7C1udqn~STw|~2R9bw-)G3bk>tZ72#=rn)ygNQ2jfJv>IfJhRbsv!O8rgT}j
zz@h^ts|C)6t?qT`@30ta!BAd8Yc4xjne^dV=7%jx&Ozaaowzqfy8*<1V91P!Cuezg
zrM=iX(A$e8+mChsU*7zA^?B=6Hmc0|zo#hV`X#`geG43<=uc_@qK{;{Z2u^$grms8
z2Y8ZSP@t(CE42CPmnBDZ^mzpOh@H_P8mD^>zh5c$n298|%+Gr3){?Y=;{!c^Hi=(a
z2D@1LSfR<MhNq<OogO{tzF5O4&S`E<NG<(*pkir5ai+WWkKbBoBI6O{VZ#@%DM@tX
zCo`bY@Zj2u?hbWut;rXbP|2ga%NSU=+q<2tCwWYa?L}|ib6AF1P?Y<y)=R1c+?O+X
zaJ)K(*vbn+`&CKnhx%<hYlUJa_+yP<*A2f{<p0&p;hGjz*g;%fqCtxMnppCIo!N&2
zuTV)GUuWcb9`|r2yMxlx9;qQJi)eId?BjRD4eQE-qrytB3ZG$16@(-_k`#eSjp`Av
z)QzURiFr{jf1+xtn^8!L^81jq##2HZNbyaMp>b|}1YV}EtT?~X;zUV8p0xd4qa=rX
zY(TJ*=Upp1Yi#^|^4GC*Mk8O+U!X%-X7K|f7+cO*9(Y-EH`Wla0mTS`qPr_^Ud!Ni
z)9l;e{FCeY?k;qD*z+%jzQehq+Uaj#|G_vpsBDb;8EeERX)1;<b~@J=I~j_KqO(>C
zYm;{lOdlDx+@T&@SX>RcP}4$QUaf<IAz$aaESa0*gvs1RZ^((fkNyQFzGlkW!Ml3B
zV|9kJ`8p0nbRa!<KI341;qi&k981f7*sFa_x$;6WI*_=gtW@J(GZsEu+V^>#@95=Q
znceLAarwYB@QV73Yt@)wd#gl-&$;1-NGz)UZ<i-9C8CCK1G-ASSq3obA<)+!^Wd$r
zgizw_j1Op*mpSP3YS=U;wv|8<zrx?vzYn}RR-}KDJMI$LrVhA?)-=zvYPAlN>DxrD
z2pDg_DhB#E_#VzrPA;^#r)dsQTRp`+UZeZs!`M!^T7QNTAPABlxkZnznigxW4Bw%x
z`q?mFY!?hfgQDr|vbt*fXf4I(D-^4qm$hsLvp5b5-lBi|i&%dG4%rL>1XwL;uP&Ro
z1YZn=PQDs_Z5X?phJ!+>o}u;*s#Dkgos?8L!6cp%&^@>D*&Q=WZKBrLa~JDq3cd3J
zJM?~dX0Dm`BreA9IA?Ns>-Ke`rhGy>-vR`m+wwIAk(K`*VR^pn3$Xg)DHt2wYha4_
zvZsn(7ojL-<wHowUZtVjYly>B78Gss_t#(ay;NGikggckR7AP@+frI-UGRj;Mi4rs
zpKSmhZ;gI;>f{i_9k;k$Q9PxIr;X?N^UbLWseNhB6$h>dgR{QXHLPWN&Nk7uW}W(6
zPW)R3H#SmdUCAAPBCg}__<4=llDe+1V1B>b>`$Yp35s5uoQR<VE_bq{_Hc%~M7AUw
z-HL9LddUl-?zPL0vaCV3W}dIbel4E0G<bT1K9CwveKg$Di{gUpMwgzevv$^^sEoW4
zjaLLSp-gh|strW|oxJgkWDyv)@tj94Q*sF*G?PB6N99WrSthnqqb{W#dDt7%-=Cz>
z`88V}CJQXc*39^3fRT-`XeCBOUo0-BS1?|KE4~BjlB*>fl&%rEs$jZROm_N{Br*(+
z1@+WrX8aF#;uqHRl8uAEde%#&ceL-@sjgoSh_*D~CH2#O?6{#l2Rvswb$Afe)VMQK
z41c=oZbuoD$5zvHIze~WCyJS789h#k`nxdo2q$0sO9A3!i>yXyU^INiM<=6Cqh%_J
zIz~;@@yW6|STOl5N%-B;9VJ^i1~M|U_CBTGpzh>DJ`#W*g1b^?;`tzqHD*%f+t9C3
z7xc$U^{~EK;-x_By7Jx%)*&HewBW_!`^9eK2kill7gW9n;TlPBNNMrL{>fyu*W$NO
zx&BL@F-g^HSbmr7pfu@18mfL{6_0=m4qtY*o}dNsIdl3go^&Z)`$c6VPV8Kt1**s#
zbC36*K9N-UnZoHfS6dHQq@2^mTyAD$JOQb_o1htp^~kQOyg!EDwD5fv-#^|$;;ha7
zMwkmC&!32azTY#3^5y4%xtQy0v?TjVTf`IMZ)g0y2;ASo|4T@b5JHL%ZkF#Z{u=?;
zi@OFCIK-rcK|M-Iul*;zK&wx;=?ua&iyv{|hRwzc<UEN;$zj#RW#>&l+1KJFvh-N(
z#go|BU?|LgZmy)hPJ0U6&YZ`UK0t3z@h7Xdde*%m7N+z|22*Q|<td8U2+hq$3A9K{
z&8<{CAjSvVn32YFq^IiFv#IAxnaY(Gdup-orO%rtFvw)aGYaqTzDH9P2UX_0$#q(x
zZv75P3{{83&znezDQ3MBe0RU@?6YYoB*hB&s)AdT>M|(hhTH(lbAWf?4jP6Pc$QgM
z`VKH6m)iW?9j+zn`>8eC^9$os4~;LTLom1K)T&nxI(Epyb955W(yd-opC_u;^W6j)
z3L)BQJEf}*dWPpk17r9@8@`OL982_?bgvS~1XC9!@p)noR5~gIXEoljz00_u0E|UG
z9~^wCaJ80Vce0{~rtMGsZ#Fz}EY}Qz=9y=!=b2TEE5kQ`%cMBmhFHA6p5=w+nIWZv
zD7^d;9#{9V6gCN?SyjbbO!xMg$8|)OSzykL^U{yJH$32pk#yCJll+Gfj*_9>t4j+D
z9@v|32I^Ett(pY=%U_Q~LfWYUsOpk6uO-bY83o>QLuhD1A{<bJ>FHW52XB3=aeC(j
z)9p;)e?@PP?pGC!*_sSSx*mz%|DzB)@6PY!h`?mqWuo?8;Ece)vyl$Toy+~t=N52=
z^3^zA5Y3LL`1i)&*Rb2tBay*Es}m4&J5TA-@bE8Sbx}ybBFdqabAwx2A>h~X)X7nT
zMJk;lbuek<0W;>qmGFit(6Zzql{o5JL!E5E`QKaDoP2W$QPUQ;mIyEObF5v~PUBH&
z<t&dVkzgf*XW7)8Gh%zxI2Y8LXxQ>6LdP#R@8|J?cqo5Mc;W-4-IR`{`U4u@a+ROc
z!4DA#PpVaS&*z@F1ilb0Ag}K$UpuN4@reFdV#!+El0N!=WTa<wR%t}}S_Yvy+A2)X
zd6Pdv0P^oBd?sdX8g%1Ig^VQ4h_x37Ks%@FfdKhnlbG^iDLSe~2L5)bf0o)xYmFON
zzK|4Ghadtnw7>)8d+XWB$?4&YwhVoGobI)eA?3ROu(;J{XiX$Ze{A9N;<$28@ssoR
zI)^Sj+$Y@2ZhL;B3y`G<6U^-~$i?IVs85qOhBUzSJ<+^7>_^hQ9shw!_G95UB6mp)
z9#m|UW$*D`na5bNCbaGnKJV|zX**N(<G)7fY#tSVC_%EGMy0z%FdNMdWVTip3kWor
z7@Z-M&~b~(fvW{H@jNQ2i<T)TeS}<|HkTfuU@`e|?E9;@j>fM|!|Z{(>;H&b+j5%{
zB60$T>=nsO`ycRo)V{B1s&smS7B@7bZ*i_rR}V^In>d>w&U=D@x(P6IR`rr)I3V;n
z%N^Pv3Ax|Z-OI!{MDd;%zg424@t7Z;8l*ITKbu8gQ(PrQggw&%=$2=IzeH(3?V&~d
zq5dN?!TGJbU0v;w8JThV*pxG0$%j?Y9K+quO1i%_eu~yp9@E6j5+_bxDO~}sdSArb
zG&rOzJnT^EWavkXi@~&vVXQ6+r#0$RDRL=0Hmv5xfArCU1!&Fn)fYo@5$kVk4sCoQ
zXw6uN>Fg7?yQU)=sRcegOa3+G62oh7g$s`0*~;R`M#Ow&34r-LJ$cgJwS%{-Us#2e
z$l1TBHgqcH-Ayg#6=@kC+Xl~@qXvN=UfiZ1P1tra#iw@p)yN*ISnMX4?(zHAsL<n%
z%+Q0p!h)}t&J}hVeF|q)1szHAWkBB!EXqd5uaDV1s=18v1iG$=L@|(zh74)M*2jr)
zQ3!pj_tE6b<o<i_3^7zI5*<7X(~dE9HI=<wd2!DOhOWw&p12>?qUQ<FH;pGCVGtrP
z4{j!@bM1!AuBI28E#HI91T2auE<&6!yUQ$F9iq*DU8_a`8ViR_hPyyLg^%zt891eX
zqY4M>Eiv)HIM+OMCpy+*u*aIC7nhi~sg`d4adxh0uN*d+5$OmiC=J{25A600?m)=k
zweh>x<)GA@#RD^h;CKN@%z5vMTL;fgRgauCJP_raCH>is*_nI=@Bh2*si^p%<{bjJ
z*!w!f-J7)GZ>~@D9<e+rjy8P!5N5Z0`A6`b$L3})zgu$mrUylIbtgl2(4#Jc*DD77
z7Qk3^v_P+csw@4+-N}HtNL#to+#)K4%eAfw%SbX$kw-O}e)10}lfDp)0ZG3yO4SLi
z^2yS!`Q($OHBgi7K+|h0EZ>T0c~P5>6n%E#5d4^YK?-D`d#aR5HN>)|VzmDJ#^I}*
zGy#*Ro5JzIijiiSEixU?Du+QuLGe{}aPY(4pDh-BKZDXd*y|NvklAzj4)|E%egHa2
zQe6vvPXO4c#?7;(29@NtLu|XkV6E};LLbEYv1GoSb|2-X7J60fI9N9p@}4RR#@~*p
z=X`e~TOyS=N}V@ph-YuH0V$6OaHNO&7yceJ{$WYYl<h)N=z?#*0#%@BrqnR^pw5bG
z%l>cLUJx?HzPcTX<*gUm4+=z=fI^Okz-DZdjP}3PI%n|2`;VkSmoW{yY`b{g&MUu)
z_shc;|7+H|ymtuqBiZ4aR==J+MLkRnY{!f5&Gf49OW|Uk?ESI`Toa(+zX|Jk5gxrp
z&P&KB^r)#AIEf&XRIXdgwbbz>A|D(+Fp~0QSKqBwb2wV~0{_By6PwjG>uiE@tBVW3
zM?GR3J3j5axs0uuITe$7?v-DPg;UfV383k0F6Jp^q`4xd?p2;?SXtT8{INSvjx95+
zhp>R`q#N^U_CcEi=6Y|hW%F3UDgsZqDd^SfD9$_nKen#^)E<_GA_C8|`pJyJ=Qi3d
zHfw?DI)a(sL;Dz1Ds$~Tg>L$^!i|LkQpQFAuXvH~``g?RW<Gr>M?;j8$hPyc(*a9G
zT|ZP$Y0Tu-aK?{L$w)4tO}sjs6@~5k$Yg5KBWgcFDfZ4@XA^jvxATGl{HG;Tk?OFb
zneNn(1gRvt^ZYHNLK*21FLB7L?8G=KT_s7>fH?#Bp|F+UPX6cEbVaKsx>9{WvIcNI
z=2d6?hP#INnqv@Y7CHT~Gdg!`Q*8do)2+UAzo|Jbf$OVm3Tl|jr0=<UwPaSq-ec8P
zVJx0Av`3|oS?S3_*Yq=-y<BY@HnbJJ?}W-_gw?X%pGdty%QC}7Wvq!rbUg9SlYVcD
zK4V<PhQonxzL3@Ohw{;sizh6m-VM{GP`|)<ZDvdBdFuQnZx0P2A(7SISfFQ3yrh2Q
z5Z<a#Eaq&^2pLUq^>0o_bc<y4D$W6HK9w|SeHlUm+U74;+6$8|3Lq?VkodUTw|1e0
z*gqH9CluIm<cjitd)(lNGhd{V_+p2s8!;1?o#f5fFa{$KTa5R&@CB!gk{{aSa@1mF
zt3Tmh+Qux}(b{~hz;}$o3AyQ>x1QO^d?HFmo+Cq8X<Ks3p^x{6#q{qg4dE}k_nuCr
z&Pl+5t(}sHALmO=57|Pxj)fK;wg9=72jz;PTmNbu24w}F8Wr%hQoVWa49hrT+d`tF
zeQVTf+KU;)kA=v-(Y6*pF)PWVE@i>5nK6aQnnQV!Kg>Atgwphl*9TaOyCB1Ra6%?f
zTp=}syF2s}9b0@Mm-~;TA)F=`kAxu7)_8}r?XJJ^dML(b=b31G*CaS!q_r=Y*_Nzj
zJRq@Yt|?o<U4%zPz`jN!pYT#cVk<|x{G{@Ob}J4+g&5<mX!$n(H!W=Q=y}!p#N;tb
zccAJ^iHu(0zY{kl#<30b$DWM=F^o5$yoF$6DR6$B#5OeYnQZo)h%(oy&cb6<xCjo=
zSu{LpAXY|tIIrt?|D{R89aNEE)ZhJpxq^QDxKYisV`R$MWs!h@@td*F?1gg(N!E<q
zkwf_%AQ1)O&OQ-E%-ZC8Ux7YoPxf~dgCvV+xgFmO?j+<96mz-LxF3y)IuE4!X7U|q
zHC0o=OP~wUu~R73c<{fgT5PNg^@Bh4C&_Hn^9s1n*-*Citv{SYnBWYEfKEzGRyYF#
zQI0tg>QT`-(cM|cUA5gU*+r&ra6&HJoUEP55j(S6e&hVClfotQfAJ)|$9|fi^Yi;s
zij&*(+}k6`+bOq~y}SQz1*QMy+@o$il_V`{F>zunK6z0g1DRUP@=r8|w+v?`dVr~H
zjQ)Zlu|1|XOqV*eY9@sMeR;_(j}I$h&X3$58xOH<Npt9PtdG_mScnZ(cdy^$26hg9
zmC>F4dzn$MZ5oP63(AWKwE_DD^>*Ke<~92VUl;?({C}A|&L`Kj;xM);m5NP-QptU;
z6QvCHb3IhxRBm6}G)lkw*_G;-`ypbDg`qpo@|hI-BXliP+ESKm#5^9YUU0)Nl2!Xp
zPN$df>V3hN-<k||1()M*WGHNgMdIhb5=!j-Z1}K{`W10@clG*XJ;17fUpux-aqi2#
zvhCx0B?PW{bY)#vsVfD`7Sip$n^aS~J~<6D{^?Tu4aX4g5+`xg@#`V)ROMnNQPhuP
z;T-S(Tvucb`o-HCtG>Hz3(~#+_U5OdUkfB>FVX^PsdUs!YGl+Unn(ta`=Vtsw<n#d
zfj5EJYmXE2K-ERFe1@+P=92wN+o}N%sdT~h9Vx^jN9}HxUw}bt#=7ehoqIPlUZFwQ
z{w~bVehWW0$FELP1&}V`*atq!7gHH)A9kmfc?lDq^$_w*bhVY-)j2bAz<-YKVs-(0
zsO1h($FE;SX61bx{;$gG{GlLNHhasBbndPl2Dqc1i~EB@=7Wwm<Q5N2(w|h3z_`tW
zL8UU!7eWKld7MAmi8uj_$)XAST)wjzt`9f2;qKie(GHjrR@i7<LQ|-@Ec-cnlfL8^
z-q)yjwYseNTZdWb08MV{i21K6*0?7`jq-b=m+CD<K$}I@yfnCsiZsKmt`XChpFhkD
z$nWp(n}V}~(G#Tv_vLiep4;SReRHZmKVz(wQ9pME<d1|VUpnjQ393?;4?6FBZ|1+R
zA$?|^YJN!_kjaRnM^8g&*i91mvUptJ8Eq1k_1eHVL|j3S>}8}r9&Y$cE5kSDwmTOS
z?CE$?@!z^d|0PApSqdLO#j?V`h>!y5H}uX438cf@&Dm(=-6xKD46iU*X~eJ)W<-`Q
z$&okiXDAa+mYBuBk`0%$BiQ42Qq=L4X|D%*&S8p~;O*i?0N)A`Vs-tyPP!{YW{T}i
z7!5Xu7f#%u^mtU``Fx{%wIU5phA9phqm2iy%@Ln?P*q8{EQ;gmh+M*)>YFz@!dpE#
zFUz2lHk%F_jOwK{42`7$l*}8+!OZn<^?~Csg&h)K9uP8$pG<w|8A+yB6TEjyZKquF
zj6=lO`E<#C!e4T^-L=g_M}tixPA+&M?R+e6i%%rcX~cHzV&Ww<*IdQ7U-he!V|@Fl
z*wDBZoBYuRyPdI`@mklEv{(a19LfxLxYc1aHI3sTzxo;IX>Y?AT6tBF!qE&GoSOae
zi5_!NR4)R8Kw|$~x7-P#D*SGJb_gtMng&Rdd@+<lG>&+2od+wUmR{OB)H!Jx(v~%V
zTdmvgm5y;(^2<U`0UN9<nQNGj(p5}v)8CQ+@hR%K4mO^vF16wH$M5ykKMj|)j>-Mz
zff4iU{8ZcV-fKP06|m6f*5B)SpfoJ`@&D0u77k5@?b`-XLP|iTOQe+Ukpj{&LQz_z
zB}T&-Flmt(9n#WL(mlGnrAChKW-$2nzQ6bTC+_EZuIoIHqo8U}%-m(zRuY$<gne&?
zUM+_qbSCbWU^mzt2rbQ#-1?sJK~e|N&R%mf4ou8M*(?l6M6IP_SJIpK%lwgWcxQsJ
zTUNVukv?UNs>x4>Wc4`DK6>%dZ(Y0X=CGSbNHf#<={9NLY~)=RBiN=@gGHv-8Pt&I
zdCv+L`O5U}ChK(78l}i!h|;3Z=8o=-1?~tIOrNJ-_4T{ns;~9iM9J9Dy6HRI-p`;9
zZ#azawBcTn9zwfLJs?uL_u*)*=FMZZRr&l2w(FWnD2316DFEu~#{!R@5w8O|kc@>L
zXPqWB_<Du(&h}6^HJSM#SY^7$M><mbP3?Ki75S`_s140m2&}0Pvr<h1o*{Ry5Yb9K
zO!#6epVSO}=d3#I$mtel<1OnX%fiQej?iXBqB6EEs_QU9Mu{V!{*39<ri;tBA+zlZ
zT{m_CEl4SjW`fI}x!VyJkj%dCrq}=R4BMQ<2j-6M4=@b}Uv~#_n$;!s44%DM($=yh
zAV_A2zLY$;ZCv@?NV(hSJ%_|49g~~R?>BdfZCPX5uc?F^gW9W@ZJfNEX|(U~>BU9O
zAQ8#oUow3No<A?)<t%xzefIi9e4b4Xo7q?$^P8x1=6nVw#}0DRIF%t={pV*X%$6oi
zpLb%UL`y!vY3amoSVEF%lv%=Lu1^+1@W}fZxvv7IwioZHL(=4s7k{V42CTQJTa&uD
z|KtDntb@~SW(%0R*-PEm-nCLq@#IkZd8_5Qr6ZdIIAlh9Y-r#S(c_6fI<d`Tf#gna
zEct2GTC(`tqcE%gCscS`@&OFObU`m=o2%tM)eGit=^dLYW+*yyFeCz~11R%;ITc9R
z(AspysI1ne{Xiw>)yz<MzwVugIJNGX-Sm!+Q~K|Se}1emk8T*;DyZ7hTL`&mG2c;(
zsKTUWqV7}RB``^iCa7YH%0kQi`o#|H;SM;NdB2lckP2<a?8N3(di(d1XYg>FhxgwM
zme7DlR=|xtSsb8t5T^F<O&*{sS*1+<8?*%f4?@F^%6VdzQx7<dQU#xu-*w(Dg^z;0
z^b#<?Xy@x#(hF5x=0ktur{-_YjPt{^e~n|J12odRI-zGfJcYd<+|8u-89bRqeLg~8
zdgO@BQ9N*)Na4?uU_x?~iJn(xV)JJbZ6<V=>(w#`flQ!{^qaMsV^lm>Fj6!^*tK;^
zgW~*?AO+ROb2~*vU&Du*#edy|TU+|S<$0>Mc7xANs}x5{GaRCDEBXG$5lWg$c(V28
zBrNseTa!})U&s9eeYokDOz%Ivo^!iGD0VE-u(qNPQ+H1D)VO&YV3J;#O1AjO%#2<*
zPR6+YmtJ5{X31rr^38dJJ~E=+Dt$<^q+k>>EN0s2b6p`T6sQ+*k;v1vV0Gh?`0@BX
z<gIupZ$m-(<CV6ejxVJ~5i@Jid-0if7X2cXiLO1BG}hy$?YUCQbULHUq55T|M#(<?
z8gv|n3+Cme7=*>F$)3i%(-%Hd!br_b;~k>#98&`io~u*iMi=COo4Dkax`myfw^+NL
zPWkSKrGK3fB~9JOgGE-+fi6kiqy|LNGER(&EpyqkhKJ5txy|!GZo%J8xLy4MGs=2z
z9YP~>jB}I-r}F^=qFbdrxD8L;9m7~g<8`Hc`mwi??_5OPXf@TM@?330akHFI&fZHc
zkOKFB!^GMu`>Z>IO`kKWSCN0&3%&`y#e}VH7CcXj$N5jI8)I-ScKgDl$6E5`stykN
z7JwNA8jT>!AJZf>XT-wEL7hd7!vrLV$;qPXCCisHd4!bnxFten`m!X{4#f^EV1axZ
zrg|}pL(q{rFc5#8lO}7<>G)rHxTZiNm0}%@l5Tlg-&@W&h&xt3kGMSN&fX_Ctj4HI
zdC!&i<FoOm4<_Mdz0BXB6&I1E6D^I=9vR;4c4N2<=P@K{goSqAN=fgH$d>m9LzhTL
zO}sjtFY0&F4&Ny>My03$4g5A3F6ftD3Hgw<%lROE>dh6RD?Sv7+iX<h-WE?}_gT$*
z<A0Fsf=~h2lt(n9+Rqv*@$t+N1rRs=iZR@!zY#o{nO^6Kt75b$b>i`J$rJL5W_%u5
zD<=~NvC=AbvI|wlz*M<ObM_9U(GMa5v?{G1V7u6)Iu9uCK}LhgxZFHgHhZrI&%EPG
zP@=o?FX3V)SHz&ib}^_nV%>D`>HI8(PGGn0eNJ(dC#34Zv<I4UdB-&#axXe>3<c|{
z|JsiRB;7WRx*Ys~+E~wLR?2Fp>Dr1CmOCzxwCK|NhDbjmQZyLYh<?!Q8qAI=ufE|A
zv=^H`HYDN2;p!pk=e*#{z&Hr6y>E>F5fa|3iQ$`gK5Sv&fXkw#nSp_)C@Lo{IAnDB
zF=Tmx7W#1U2(O!d1$A(0*Yph%8+G4CRy+#B9)_<E%Qqc)IyST1V;#6uYe){27~+In
z)R$JvNSZ<n=}bO1+*P-1HIzwGY_56}sZx*Ds9pW~TPZeT)?Y3gI>)`YGP&-uc<Vaf
zEU}+|JQyebEK%G?7eg3fd~=uAVJdJ4wf#Hh4QiUBMZ6^O1|2y=Kd+3PxMM6z%*;}|
zYofFH|7pgh&tR}oQHowicr}^xtM<D}eCbYPBLGiH#z#mhf7LnuJ!R|Cjx^7nlDvuy
zCd+R4$9Vc4LODR*&##YB#O(I(0kQBi%a^e=3j}NNgae)L_?S;P#3dx+%Q&iGk#BPQ
zK2N+=b$Gn7PSyr+=Cbaw`J$5irc;}vcqLc;XWjntySde4fmjc&ej5Xnrqar^_Uq9N
zS^lT7d4qBBI410og=}u@qyt;yk4HIO15S0Ua?M}YRJwkbD(kRfaeXQo7&gb6Sf3S5
z4Y+Q8Iz2Z%ExWLyzxC;nj$~ATY6??5m<M(UNj#l?=kCJss8^>B!SbE#dMcgoad|e~
z)G?X2Q!PJR1glKMBVRpB3TZ{XnYIj1rD6Z(!C@qNEU)Jx-Z$c2J$;iN_t!-etvdr9
zCYSo{ix0r{aNq54E%=gbg!LeKtxRUG0OrA<NsEEUdO9v=I4;)<WBeU58Zvi-GQheA
zH{W@9m9UhrSRFEt_U>uhi60a)2Xx}n)ib{wZzuG8;cfEgyUupE%-s%f^5IJ69-1Es
z*^BdieB@A)lr-IzJ35N?(=Ma>x!7xZy!FeMiuGF>U5?Fom@Ej5XxZt_^h<MGZk&`P
zT`gUS6NkV1<4)IWc2!Rp#xznEg>KaPC)IB|vNh<r^VDQ~hR&4IBgPLedW|980|vrb
zgOmRni&*zD`Vv;vCTnTx1SS|Xgn$lP{Rk3@tms1c*`AFDR*Vt!QUdaq&yVp1`aZ|*
z57KOmrs*HNO--e}!98+&Io#lC1&}%@8I$ll+-R|;uXp2bj>ZXR*>gmRnY01kv#obw
zTQRu5L>En_Gn-i@rJP_QL>U+r)UtceKW{G*aKxSBVTUgD7Pmr8$`_B&Q`OmfX$yvK
z%269XTZ54?ds-5rAXeJ1?rFI4Q39@PF66yR8_%m26`P1h2ij{Wsl)ROsfFw76RBxZ
zv=wl5^3A&Btg7sLT!Po4+HOQFQeQPrxdB}>JjeWo^eTMDqvg|Sw>!OH)`QDmcq#?9
z2<CpoT(2Psv(Ho~Xa+}}^OGYzk-O94rpK1JrnZG!P1)sthfG{bWC23dg-`gsSo=CJ
zY(nMz+zLi-yLRk-?9xtnZnWdY+g(7MC`ah=?;-M{D;J*o!z-H*vL#pcNW@-#3Mdg&
z3L+GD{Sf_@o8Ak#n|FB9Q*;^%MVn6w58`CN4Q=f7T`S$9xr}?@dO-2c4I^t&;(|Z^
zG^vJG_>rLW9Lecc>YoEPoYGaBW5<IzwpZ3Y_U{Toj`lP4(r%e7^pc(!iSlj4rMY5t
z!fS@bEu#67_4N{250vI<5^mZ^n$*G^)JiRt^~;gj@>y+7-zIJ~)pqX*wyr~UAWj@^
zL+aU&z(n}hevgA-qa~2H--c%t;hJb@vB>xf&WVB*qTY)Y&q;>dep8v5OTdescU`Nw
zMOAF)dL|daM{V7)1y+F*Mb!YII8m*F?b4+;u=DY5U$hG2d%W}Luv@QBx5^fDJ!@p5
z&2RnVH0Y5D;uU!XBvi?(xs2Jtgd<X@{#TvxChf*?o9e;V*3DZDlI;^O<DJv&(tMv&
zZ4H$Ab`<~E$gH3LaVc?mOpA_tt-z+s;9cowEd%F2<`b~A$xl~NGr@iS3zHo$R!>dk
zT)5O6@XvonJ(BCc-d1x|kP$!Ea+g5qT}!9}&DzU!_1Uw89tdt7=J`Ny(j;yLk!?q~
zN6G7#(dLt2f^LjqN2tBECUe5fIs=e|yYox4=>Ed_NEx#V+aT_;GjR3h4ls?HCoRxR
zce#<$pfR3+xEAA>PP3q{@58E&6&$`6?QJFr5ae+2cIt$*tlZzW$sV<U3_t;<EW>qA
zz(>S{&BU7J;b&!6XBN`;;3e)cs;2P?lAvll5dxpzvnH8)rY3ThWo^+_il$v4--2Ux
zZ!K%OkWV1TTw=vfB`g|-XVi%%YZ4J2+}O4wXzvkF-BLeDdfH%X()hCt6DnX``G}S4
zYFU`;F9h|i5=4xb*CN`?-K@6Q`cW)*yf@I(p%p}UJL$V59EtqD?}Goz#{9tf#Sv7(
z_W=`+FQAsk6wr1iOFE~=Vpq%uslQgI-NrwCR0?!UnhJZQAnTO*#jrr*xlRw>PiBRc
z@lL#SS!Vu@!o7+R$vp0K$s^s<Slj1k__SnVVlCnwZ+`|NnR>Rh<m}@83obXj`Q|q6
z+B?&8ulNRrojCM6_6DOmB>2-mQ-_2c(kI@2tYo30;KrA>%H*`7*MR}DBOWjAC=1}O
z<B>+R<)o!_I9jk;YX$AE$PqA8ap!OWW`<~NO=JZ)M_Im~`AsN=(@Z5uY(3^9;jM!2
zYqKG$#=v1)g(Wk{I!3(CDD48{HC8Z1HACJ|(EZ2<HT`c+W|As^=}1-v&fIv@N5oH}
zEflFB$)_TW9Up!~_x8Ku2kok$9w{ZosAG6^6cq#Vs(o6n8$ZUA=<Lk<NiK>q8ns<{
zI8qfi92WI6GbCw#o-)&Qum=ly52}<J*N=^N&jT`CPydCTazZRims>=au6GJ{HZkRB
z8T3hLNaO<TWE?$-Z>bM%#M!4nmX9mc;wO4N;)2*z@MG&cY(|C_XDs9crmH*oa00z!
zbpOu(6;xL438<tgIUYE!pb;6M=qP6--D>co-Ar7#ruCDsdc@Te;C6KB>=Vh4NdX>a
zcQH3JNF#JYAFg2!W@Nuep9+iL9yFD6-t^r0`2K7VZD#axT{+$2<jMZaIO>~iJJjp!
z_l^ccu%4b|dzfFf<zvBzT0s1ze7V>|h;x}M5?UA7vq;;LAhbRU-(|o#tn#}WEiKhd
z6tUo%0uA5Sx3<h}S#Si#iap<T8vvx>%fv*b&|S@l4zLXUSzW1Q>_+vVq9L)S)+6O<
zwsa$00`IL`jG8|qoY1JqAj24iz8|=PEjBbnv&U9cq7sqXiHnL~sYW=f+%(;BDXbN&
z*I!rj#PZ(g`UzXi)C<vFC4yeFxX)GP8i%S@D?sp_XAED>NQWhY<q4Z%Mi|V;r+mSJ
zPmM&`)Vk)pg?#~>ot?+;t3vyv4r>kr@W7<YtibjLHHKinh|;tWqSk9m>*X!4MK0<}
zX33#fYj+0xgrI%V?(fjU#W_jFK$@A6g@SMU#^zZ6_##-3=!lK@47Wz4G}Q;GHzhv$
zX^Oyxl0iD_qmj<lbuG^tq({dJT;|g*)8w7m9vZYnBYfO*UuS?iPlKr)QX{s+P(Q1a
znjy)PjA|{BN4QkBk>MjYs?xr^WBz)!IrE3XXLbpk2}u;^cLwlDC0O0?!N{QsM5z8P
z)zooe<T&@r@ah}MFdrXLmo4-ZWZ^5q-*I6t@648>AT|T(FFJqNH*xxj)M~ze5aI4V
zlq+3riR<I0<6)VNx6PiWim80>lrRYn*x_sV8Ize`*q>GAW?0T$<Q<x&Nh?Z=5>^QK
zAo#~dBr<H-gL504=&JY*Zsz8IwmE5qX#oi2<9~?xl*9KP&KqBJO`>=Y&h?HBFSpO5
zQdAvAm6V$FuI5XnXY_v##M+^7m7-EY=R3A}mFDZjZ@C?s)u4GTRJnoWSuxUbp4luZ
z#wJU0oU!-YO6Hq1jiN^zIUr(dCZ(rJnJ>5*>ijY`TPC=D!-m#{*ODyHiInYcHSQ-u
zB+~N+!=**WoVSe}K(?N&>Spm0K<_fCY38`Sp5zBG@7!}S#$FdlZ>LhHny&RmVV?0D
z(JnDY0IEjqMrP%U24E;2nYM2tV+&1$2U%Y|RmJf(K!d2Z_tFNPjY!u~U)4?1l(R^3
zPuT@K6@6=m>z7uR)s`|yipxTML{S0;q{f<%4I+9Uo3P2QN5d62Pk+Rs;t^yH*Ll8J
z49K?)p_kJRu8&VA;d#mIk)Z{x>YgE8XDws!4OmoKI-MPyb6WfC{ld=ofZVXn?KeE=
zs+vDT?RcXp;U`bqtU+EB`>-a_{m{U}7UF%NC)A7aGlJgDOSrUU7DRE`A$x^*3|T*f
zdNv@S55_CVTL5q``e#=82S?>VXA_TAKbGHNUN;E`BcQOc>Ja&BWiLyxtQR5Jb{}nX
z94XSg!H?HUhu7BSzJ~lJnRM6YM9}uu_m>akIsYgkiQ(3aXAYlxA?=dx5&>B_X8z9+
zyrIMIR}1LcE-26b7Cb4tYTn-4ceHi-Kcy3gvP{$UMa_=r!!hY3l8%#>j*X6b!|+iN
zaf`e1fbNsIGztdJKnb#A&YOaqspLq7XY5lm{c-iZuK?F=GCoew`Rr|lXi^E5A|(~!
z0a3XfSIw+mvOgO1j5}H*85_k9(PkoYqB%my8P7}|%Mj`!%-x#oV#S{<elIQ$J5!0O
z?0fo_)qDQmX-4|sorqIMEiE>=7VS~@2?emw?EwB)&7#uMNRFnu-(rqLmTznY4*7V#
zN6`_<5ph13Ws!jiB;%>c)mYL15F24U1AnlbMBW~Nx8*Bz8B)K196+^n<H>S3dtGlc
zwKKKXa)PL0!fK`Q<CS~pZMB1_Od7ki2P{Nrwiqo-Ag5_U0t;M1iutR5Bdm1tq5`AW
zV_r=)k{IWT5-~i-8Q?Wn`r9w#_KsFH8I^^vy_q2ik)&4xBhA2K?j*p$bu9$G6S>#j
zalhM<NH1mZHbw+B*%>y;F0nV*XLqouu62tiP*%F!Ah-l~c5Q_0ZMy9h&`w^aAN<Gd
zv_z`4_~VW3`iL~{N>1nB(B->B&t;&U$cU(c1bxyOT$Q$)!DKZRm)7WE2Xzib7d${3
z{niz4*Ip`3Z(1q!uO*56>K%-*VXRU!rD^!3H}N?95qrR`;-qP_R|ScPVlxILOCv$D
z8hLy|XZmo@X!@4@!9E=QSwnox?;IPYHA!7i{v<QI&Ohd*f`w5lfl27C?3GgE5#BNQ
z<}@}o3wmk0L8N%uM;OyTQkJnp-e*BOxa$1v)+AM-UEk~(3Ef+YjiwRh!7*FlyJi@X
zNMf*b+(syW{@~|NB7fg67tw&l7v7$fo3t#pBvXk1I)}dzya8W?+p>?Il$vGHSGK3*
z4))0;4*7k^5Kfep(9_?$OkYT+_E_x*?su_9Y~k;b&&04g0VD}VnTw4Y((YI~Gf06O
z%3trQKhWQdGBH;YnAoZ)?{L8GTCdj>=&x{JIHmD@J@~gEe8Y!`>mK{<^Q;A5?g#y6
zuixvh*ts?Ndic4^Amb%zbR;2Zu3|TEr>Pjix9sfF_hmeZzHq8CVxa}1^37^n{muf=
zVlx1f@PZlVJ8t<){S@QtEqNO9uxde5Pme#A1h1^$zxisxR29j}HZ8{fTO^y#X*dU4
zvkS<8Fw|2!Hjd^tn8HdfBxDd48tvVjKI8Y!ppzzw>HG&%v@q9vOMqx`t-BgcZyGy_
zZFe-+qTiC+*dP0&Sw7a`2}~2;P+mV)k++E7Cu$$lGC(wWo!q%$>K<!dWbF;YxPd-%
zb0(Kp<fB;05}rqw`meTpt0ppP9ru`&(fTCgS>=%s+M?895{9XaY5gQ}VtRmZ&A~zX
zwmc_J<D_S!y>lN9uR!hn#$uhmsdE`qe`6M-tB~`+I6<{6Nzvib3p#F2M>ewFT918S
zHxUY5nXNQK(hOA2Q5|d3HnZ?H8|);V5j;)(;(MrAcw9v1;&*xEY+?;2skoy$stf4W
z_ia<l?2^e1RjT0_h&Oc)le9C9HMPM8`u@vkN9e*NY`3gF{TS`9C2z^UhHFiZ^Xg~{
z4&|O%GPu=*ZSc0@!MM=y?N?}8W>IfaIjIhtyO#lW1vxOA&UHWf;v%3>x4FkFHR=yj
zFSKJBdoOkpCLNNlCK$wMW@qa#0S9ARg4BiG9<DlT+c>K-8aHZs3CnknmOje_>44QL
zk<FZF@-F{6KY43A>cjv}`E9*)O&JbE1HCvmA1Y(ZifV+}@>>H=#`MVxnzeLfg4_<@
zWS%o}X^0T@TdNVbU%bORJv8hvU77+gD?dYrBYLjuW5pYRTndrbfIRzTV6&eZl3B&5
ztP41q*v;;r90J^FxAK|}F;1;63n<el4qiyKHT5hLs`-H^{D?nKEM?v|ZC<IJ+|~@#
zwK)Q=g0Nj_r(P%WEnlgwZ)V=UTlCXQ+B1qKem|zPW0}JXgL9Z32|0F9Z7?5<lU_+f
z2-fFh`35S!w9-KVTt%C%O@@as?18iV|Kj`56HfT$p~2FOKQ$pc_i!_nE!~dn^*zih
zB`+JaR_=F;K2L?7c}}L(xYF13ktt*e=@t)+aM^g+rYy3gk=sQ6UF++3wof|EDjtiV
z4BMh-sp61VL~DO|JEh3sb(&o1C-bQDUvu*NcI?iEZn0-!Eqe$Ga&g0Ovi0QgXe(d+
z;^m1Ku)X8o&y+LC6*+%Px;>QOGiOlS+J(vIO!o)~S~uELT6UUX(Y;R`^BS2GZ5?E;
zt07~xTt~P;A;nkEvO-|*p5*7E_Z8yn@YfTcnsu4I7I+$WY8d({KIG{+xIE;0p;f~V
zrx<~^rqBrL%MdT4iY=R<i>R)=AvSbcyp^J)4g)81qB(?oUfYWEUHlhS^AxvjQ@S_-
zrMaRw#pVoZ#W&zh2E44#&&^du%p)qbUq2Sy8~f#?gq>tnF7Gv{s>OyC{4=|BAksUj
zPSWLj!Cm8P3!73R;|Ytc@oHx~E`K-7z7Fsmsl}~iF;#j0M`m6KjVMi>Tv`w<fOt9r
z4LwEjd#|XYCP_z0Sb)U1J4vXp%NAhxO+p)jdEvlxw#--9xW&7shZhYV8lnjqyj0j{
zlX*J#a09*1oCL}s>>rl?{bld90+Zcm!Io92xjn5?*pXj8|H)=?K3|WrnQghH6;%7A
znSNNQgki9z!EX<n#E0*Xd!gNxB&!$4`eJaumqLZ4SfCuP>I1k!GwGAtCfn9j*ZV<m
z`5AlJ5#n|%XGd%Vdtb`**h_e(J=-}g3JQr@K>}C`shJcfZ=*7wf>_k}7E`k)d`nN!
z<=Fhq0UBfvudMZTMiIUnLOtH6;g-0-Ou$l#YD6E9#WiM$%8LO>3l|=JDFLyv#27nt
zEPhi?X8PSZH`4kASf>HKFOMo8Q{%A0EnX6b0#@w#s)FyvWA3t3sErq{hA$0xPq^hn
zKPQ|}CA<Yi%yA?XllNZCxfT=<QudHXS+L?nMs$d16;<jCyicsijP`|zkr2^s3skNa
zeNfA3c+37Vqu>1zYDNsbeG*N!cTQC~LG?^{BuUE;u-@%8^bVx;oCSbqo>WLEb)sF(
zF5}F`buuk{(}$JAQ_(`N4>-r&c3JN!-8uAzE5QPEos>R#O)2w$cGw&>x2n?YMjbb_
zI%`Zv`n!pF9?{WSkCbiRl1qI_JTRQuek`2J{sRz8WBcKf5HUFPEbDI9R8oxld;YF!
zWBAbLbY^TXpkEs2?$&OCT7uNDk%85aq%xR${a~TGwpQ<k)>!NjfcRS>>3#Z6!EnBm
zGtqBY^930bN$8XPYI#A><Jk>|sxZAzK*~>nax?O0c%%$?GVYhvnwEon8IqP*2Itx0
zA47%+We(zUgr6rzy&gS;BviJ2Z1O?5?K_ujpkAR2Yy%Fbu|rrFY#^v=ZyhVI@^py>
zO9cQy2$75x1Y7an3~=J?K#9UDCCujUPfH*GHiR|nCw3{0#ar+6h*$Fcd>V70$t81@
zk<A(g`<Y55#Db_;q>VD<b88ism>AhlvewIMe~%SOl-xx+#`XG5`Z;H8Fv_-FS2L<o
z+`vVx>G#oiZ}qpT^NSpDSr?_ItfZ_|X$~_a=5I(QgI5r{#_;?d%aJ(HLzzBWM*5u~
z&ffvof`)>v^Yq6W>8@(DV$Qo>d-S}8EYG<qn!#r?={tu@7sms5RXOvqxn4iwk)sXQ
z_Qyr1(XvT$0*Bx8zpngFer$DjHd=AL`ItNDQ_TxWI(l5=Bmty4L=9c<&scplVOUc!
z_M-5C`D!PV7yRCAbVsa!5$69mE#5tm-56z-6H)RsiZIb(l3+T)wM3qjaS-J6tLK+9
zRaB$VYk$hT1h{~ei7M|+0lmjf!;N7FvsKS0u=!UZy+hd8Cz|qBy7A+!FJR0G!cw}t
z2b$^$b@=QghM)9gxsKq7uoZ((4o(8^ON}^ijETfRiJP1iOO|d4%uUOxhpPo~i04I7
zrHp|uj&)rIZLGl(P~<PL^eK(4-R$7es9D#IYqG7vXNW`E)e}i(+ry&pK)&1PWBy&V
ze1|Cw9NPbVM$L=%+0mBQB|-?Qay{kre|PD3U#Q#5p%1A$smng5cx0WyqyS56vH*gD
z<nY@>`z)Ja3L<7EM(6-na1>6kS@L&p$&Skt{K@BNYwd4cENz}4G2;<1Za^&1(d(sI
z5_TI0HwjM<?g)JV`0DznPET~Zm@b~ian_1*sz6uK<uNmk((!Mx_YMB$Noo9(oIfd9
zH_nR$>A`yvqs|+Iz;WQF@rac*q^XA(H#Ai7jx9NiRmO`vneSb_OGu4+dqNH&!YV=k
z<&W&%efFQkTjWokmvUGTJvGb8LsnVCvUR+*h2wjbbq3^)Sh0Q+$iJY`B43#m-=$Ds
zV8P-HdwjoY6W$SX6)7kDTURlNyZ&19vq8mx`=PEXby)n)=NC!lAH6tSUg>_Dev6EM
z#aw>Ej31>Pz{#yt6SJ62^?PTW*FGZilDzv@XE83ZJ7XTG6R!N<^R)WP#z5H49Ws1!
ze*WgPAu<v&d5|hvaBxw<I1^W`e=uDclD0_%3}@Cc-LsjHytR4e*Zo<&|6%xQ8HjnU
z|Fp|4&tz_K94ue64mo064wzdrEG{qXSS6=z$@*O2O`80^Y{pW9YZD1ql$&B64fG`H
z{YysQGzc3|EKSPQqtW|9^$NFxbf5bC3m#M^2xkS<b8=BT5OB0=sEB6z`zdBFeBJKE
z7r6KfwqDP*-#Sy!HflM()f97CP<iF!HYILCW%dg13*knTALfbm$)rv=A#{PI+|hlQ
zg`@tskh91rrTCeESx=}Yj~3yArRB?e<bRJb^l)c-Z7XA+2ec@?(V6b`=2DZV?&a~c
zE_*@oY?%gSXXemmdHch?hu2_o1-wPd$|(kKHF%3JL*RT}<$Ctj)yK<-{c)G!%BdBN
zoP)Nljehe^70R1%SwuvYp(oW<sN~xZJYi|_A>ysIdKs2I)O1mqe?e=iU>bPYofAJ{
zU}yQ$+UUhJ+UpNK(e&h5+8YUCf(HP}k#s4;-fUUD4e@5d?(Lh;tKW&FtP)~U<4H*n
z6>cIo;j4bWt~tF~36dcb%hcq<em!Q=JUulnLII;TNor{TA>vy>eZ-MZ+_44~iE!nn
z(?E&}a#cMfUk*sM99Xe$-Ws+cbwEo0{&_1_64|$7&0jbLCmPx>g;u)s4Gq)bs9l4r
zt_ij#-kBM+IEtLr!J~AU{!~PXyth!Yf8F9Cp}KXt8c%0uGtsi}6-cz48CE;g9-fU}
z&0)7qO*DoIZOm3{kEo2WMyY7>x`~?OJ6gRkX>ciuGCLR_>f)hKt{*U)fw_h5`{Zf@
zqs%V+ju0)E>9sdwSUjikj7!0b2mkcd!B$Lba=a%1opPfFvC^^(6-mLMZ7L!BKqBJJ
zaJq8hEd<&pbcZv~=a}yusnkVSljqUL!%x=p&E7T+!92U}E6;s7XLp+f_gp~En~iL@
zFw5DThu8Es-lWDg4&HI_;v?6UC?xKMjQ56GX!WI^By#JuD`w{P9S4pL6Sd+=TtSv~
zMg?n2U8`=+b$6;|9<R(;>HtgKUjEbDNL7M~ou6*fjg>&3O5!wLxil0Ii$=H#N^O60
z<ErLc5%e68Z{#&(nFoz9<>8sL*;+5{6w6zuM)8>wZ%Ij2h|`I~*Z@q*MOTPE$g?+A
zho8)YCIdEz&MX%Q7v)6x@reAeTN;9-7)8ahLIc%p;E@*sz77>=wrla1r(WBC0`WdL
zs#D$R&h@<e?woP8b#}71H#~^ZVvl+(6l;>e7IDJ4f9C-+)+*S!nk(-F9JFq=9BrNN
zcJ0y$7Am9ee(^Z%&KjPzd_S8Fo^2oPhAg3h*n|w3PcvZ&N5wZ{Q2oR4KPJ*pmAJdX
zP7zA9cB*|U<3U3duS0{wPYt`FS2M60klE|lzjRj{QaaN3iCcs=yo1gURisY{P5N_Y
zTcv$U8XvTdlR}Pe16<&C{|ND<ptFB-&uX(C4uIxRticmr>i=~J9(53ErWvNd(P&UX
z_-vV5E#;GTyo^)1{&t=}J__}@&!tt4joH0)>Z#Zxj$b*=oV@Ud=jffi!k+6>K1w=z
z|AVh1`maH+I{B+(is$%CK_8=D;}Y+sQa}NlZ-xx!dAI1vC@sT|+y3f?QnG(|Oe)y_
z>yRrBIv4XJzM`42d$JJu^k2|ve2LoXd;coW*~`s;^}p8ML`He3OKZe^Ihd9EjB6Ev
zN+5mmx|<6p-%L3b3k2n7dB(X;r}Ca&O*up8k<h?^)1ty0VdN9VSo%SXK{ft1erW6@
z+$O5iMmX{P<HtNtzt5S42M)xnm6UvVI)L?y2V#?zQW^e9$!u2^%!L==ck1qIWtnIB
zQOQY>MZr)pJ4*PRGbtX<(UbU06|MWYWqHoA-*e@A;yhM1|508P!;xf0?cslkA7}m)
zRqK0;^9J;)vrx=z>Uw5%dqcH4Vy3=|YdZ_9L2^hrEbkDenTQW6J%+G2+B<kA_Fx)I
z65<R`KlUA$pL=RrN1cysH@agas+oR;WVqXt5mp2LkxJWdlvq}Cb-v+8*PumXLMy1v
z>2b9r8z1*q(a>q8CByGgsUvrdj2YHDe1j6;hi4}Ku+3{!FZ8bWVNBc`c3f5ya&<&X
zNO!(J&8Kvy%rx54ONw2dXdQ&e?0j@v?!RX?0zvZ<MD>lB0uHg$#i0?=7DR&^{Py9t
z&m~#lmJL4TO>_qEoog!VxZ6aEYR?%A-0l^{=a=?WaR<g#6>$Ca8n=;B=LotY^Da-n
zgT>uy4-BS{K`ka2x^XinXceA*a6v!l<6P*@HpC}4pBlVaAtn)$O=0?z>uN*a=Cqsd
zs|y?)E#vcze@tLws~o(lZJssLT&F(JQy+CjJ><G3cH@G9dO|qMQ;NQMP^HD6*jTV?
z5;sSgmjwEBDB<~aiVZUB+<#)>GIuVMAEimAP`Z~I^*!Qzn_};9V6-EBiK4FjNbk!@
zc%wIu5T(WCiRel!lYn^CoW6uRJQO)*FwT38a1<~h;<5xTTK8`1tL4uT3v4Y~@C+>A
z#ph6^-K^z+_}lE$Td|=PUjZ@dmP%j5^pu!D=#b1Q{c@IqCf$j_%I0ERk3vrODvNOz
zmmaGYDQ22?yG;2+wD`AxI&EwiVQOmJq=$0y&Z0&?R-RmWe(RYo1rfOcPI}N=k;dKY
zB+bwCuVj~uO1v9ZPZNLDj{vwR8gz_aA=LRUJkGaIb`0EDHF@3t+>3}_q%@@q&mexe
zP)94xj=zjAE!HSWI}}L2d{|oyV)+>%B4E{D7%KkFb+l8qK(@H{Vk$-m@WU%1gx`Zh
za?K4_eNx(loxWEHrBZN!8G?flh@BG9@`E`1k0DVD>9BYlk9g9x*AJgG&o3sZq&dVT
zVUFn!c8wlW$sECF7pYUNIS{9*+CDbYiJm8?A&}D#%V|FjSip?%!$@x$wd!ISO{wc&
zmZ6S^f?_%PXZEz3sYDgg7;Cih8_P!Ma%>jqAmcon!ZQyu$6|buK-6~nP3D&Shu{r2
zWkFf4a+Q}0PoY99y+iL05HVGHnbVQ4wZS%`rTkwieEYd=#lcT1j%$(hSBoYikaG|i
z0e#}8?cpB6Flc^L*E%Sel5Iv0cBOqIv{$>)EP<;DF24pprvfTT`_0>oLg?oz+<`DZ
zk>k<S)6d6+lhp4V{YZc48hN)u!z9++z9W+Yt#{3L$#@_CUN@W{u;I<9(oZG}hr#9m
z-HxwC?^%f=Ogjn#Jp-RS0I{3gprQW(*ue-G-!9K0tH#WZ&Xy*ZP>jC0%%2V<=ADi{
zz8$LZUkD_vJ2>n~`bU1*iLi#?e2yY){<(hJhN!#;+mNu#I#(RsGzHiB*F|16v|&c=
zqo<I)BL|$Zk*f~b69juRDP8;s;&>M8f4yL==R>aAD{%AyNMc};>n-EHc=PXPJS0qV
zT)}9SD9hJm;+%qqk*qfy3i2rt$?0|-99HsF$~4+~Vlh+%%A@po$ydWzd$2EQ)@f|<
zyX@_q1_$rWS@0>JT+D^XDC)AEORI@%$8G@okQTc{r5l`@UC}!>@t@U_ozwaQC+w$k
zw0~bS=Mb@!x<!!{1_DoX5$q9)oqyH*3%}AB<$*H!a&neF>8W|TaVARUVEK_LM89);
zkc~GJ3ze3p2s&e;!@ds}PZ`@+iC<0UK~ou^KJrsrW;JVl<!lj-4Pw}St*I0+A|u#B
zl{dC{g>NFiW0ppJ)bitZ^Zv3`spse)$qfV4+wAcoinpcWIW?nc(uT4lmG4Ly1hn|^
zrG7EI$7bE27UlugyCo_Q<Qs+mj+28)uE_g)dssBKf}!)xb}V=LCjbm!sOz1qsG!@$
zVuS5}vNb)x*@95RL>-!(5E#@J_q)ala4HNvRcNBB`OE3zO7gazsgp&RZW(%=h06g+
zaFQ0BS+ljmI4Eh0?w22ST7eIH52#kn2js&&uW&v0cX!{EO8yxECO!s%kZCbGxYKP^
z<Vo+Twij1tY(4$X0rJ@s(A#CwiepxYff1&8+cVa5b?Cv%f4BY2V>Jf=znAKrRSIZ*
zja)Nd6ZhU#`v)cVCe;KzZ2}$09O5ir+<o&Y;%Dn^24(8u$_lb_70>3?PViE%_kS4d
z0}tL$uco^Vedj;h9q;S|g+cDF`*&PWm%-sl;yK+qW0!;y?x+uucMCVQG9=5ZLM$vP
z3YAqYf|vs}<1Q7JMF@sy>esC39kR?*@j{vqr0Cupy8kYt&2;x)ktF}1Fjx=3u1st_
z8{HZI@mIpfK-lSuu2ru4CA5R;y_;L*`AXg(rJa#=q3DXgdJaRMyw!bNJJr#iS2EZL
zM%YT|XC5;6IluT6c1Aa?Z=3lpvC)kHNO!9k>uV<j^xI$xEF5T(*A+Y<N`4ha`{BJu
z6ZmZNd=e-s0SRR(%NeA08GrqEA!E=cJ?<c6AL+8l*4wykYu;XA`y=u6hn}_dHvc@<
z-BvxgP$Lq(M3r#HHEVU9_)ymzgA>?f*<>gp+YDbMLce8DaL6yX{9=h$t(ddUF6KW(
zMRu!@ULkXLTwwclxuAJ0uZcUs_JQ}{-%#E~WaJ7hSmuJ+0a9ozz+&ev?3{5aVeziF
zVxs_EeB}zZ!c9#HHQ#Ax-yuyCQ(;QUEK&Duj;hgcxb2@5*B0|8#$SK;F_V@ym00n<
zQ-5Jn8ml)&sM=bbNU2;uYQDa<a4CAJCRL7n%yCRjVh!J5o|5mG=KQYhmcRZ_a~1dE
zub)rHMQP|)+gD#@z{lb7R~Kqkv0@e>f&E06+mss1e`tHf<3>(>O|&wtxP_Od3ThW*
zAt(fyAmkNGaEQ}*Z$GWgnz4%NtV^s_ysGZ_KB%VZjN^#@t<2gAYiF`T{;-!RL`Mms
z#Ljp5$j6B8gkFp;ECcyZ7Tiu2&u%ly+~LH7v>LEYkrhOoJQ?#4(=>>}mgW0vRCmCZ
z_^__B%ier)<9ZbG>9Hw#!q3<}-uzEPx!(cz1ru`0bQ-VFS@OE0l%#OcGYSMdBLn%H
z5OzCvIkzlcLYvo@IY^~Q(Z}p4^Pd^1r-RJb?XnHHzC9Pk%va&-e3x&-$F3CfyDi5!
z9=z#n0Utx%y*caDr4l|xH3<R1>G5^2AjP!Jlhbs+Q5v9xvW!ZK_+KHJohps<gS*a9
zpY=a@1^26CEr$~$aV5{ugvze{bs)S%YwY?K9HCNEGTc&AMg-f)UPr-qSjtiA>HLGe
zQEkHJLx+rFlSbfx^PxNsg+3kH{3)-E?t$x(PyiI~`z>DR38MH|q#D&A2J*|W&S+JT
zW_*e<!emn1<*nY9W%)4vj$nL~?&%MV*-HJ=jv<{gVzQ&||67`1y|YE_iY{x+!qFb$
z6s&wYmM<I`L7Cu8kB-Ah{hjP5u5c~IZ`D64H5JcKwvMxMwXNKdKOQ##`+uz?)3oyj
z&yk<8CFJ&=6^0S9nC4&&EB<kb^%;640d}$E@Y4RP%#aa-1Z?NZGdDt^cKgB4p5XJE
z@iDvUmJMB2&fQkzSb3$>JpJmlt6s745l2NQ-fA`zvrA=U3G>nfzX|-~L{Y7ySFc6W
zFBRnO#_w_wp3}QqO#vg9ndz$wK~p^uvkFq+T#fkY!mD3ZF!pGq|1XIYV47-G{7A=$
z;N0fnc$|vWw<CX#bCL?#hh=7bVfI~%An}fWB7V(QF=XKPGgY+|O3M@Q^Xu3APQQah
zc?Em($J`~{um6InKK9EhW%IcXlfPQ~O~Iaj`td{!95QZBSA*|02on5xa@msp=F5M&
zTsXN%O(kG3Ft^+uWT<1C6QYb(d9ZlycdDF8hmKjrcT1GJEfiTcWPEY_;nt1sU3vL}
zm~T#%upbH|^$cb7OTIr2mC<~-kwKeIK3uk<L6Z;ntq(`^HE>EU4G-i%0?Feq_LUJS
z8Z#$&lF<}=L`5O|=xf9h1i4h!t>VcGrIRGzqaO5_bB%CbL$owaw%!ji4)T(iQA!_1
zl*1k$y=r^-)fSDLBpM7OxOH`HxkA*3%(`5#QH|0<;q$kOgNZAK#yILcNpoW6Z9`N~
zGD8yx#89JU8L2~ym<z4mO=t2r`nQ^^`|{5|yryb323&HL*?U6^oGKy7b*8GRg_fqh
z1Q{;szGwE!@0u0v)1FcC7;WzhoMPz{FE6^kd=iq)c^Xi;!pj*BIp$&1EHr=XAJ~ym
z>J@d~m$^at#a1z}bi`E56c6vE1`kVByT%+Kk?7y*F}sAv=)Vmd=p7M>MiLRtO&FIr
z9hsi+%6nRprZpNWTzxS?KW8nktIz4-Uaty5_2oAtk$htQbm%dxkPVr{bh^NlV?)h%
zl@-)-w7YE0a~D}E0;->h2211D_?=R;f0bkX85yNh_`H_u`+;Xz%7tl*24F<}UYjL>
zmW7eTB=a`Rus6Z!JrNE|Pi6FgRtke3!6`+n*yxrTL=N;lBT;<o?k^o6P~_Jb<Kgla
z*StL$AVaK~i_mtZH<DI#vgc65^uP!$@WVCN7k%-`*lNP_<Zvl+vnsY)iQ)1{JW0H2
z>PR!)&gpPJD<1i;f_GjMm^%V+7qjad*Z>a(#f>A&BU8ICR-~!*NwVcLY>%u13aD+L
z{<KPiPf!($Cj+F?#=}|u)rhQ9kCg#@UuIZ&B%C_sdTf5&?O<@X8OII){7z~QoPW`r
zP~;$7scwxGtcVlJPiqK%;tm*as5NwpsP4X9NL0vYA;Jrk#wUFe>$fM}NMx*PVwGf@
zdY+ZYX-Vs4_>0dgzm>XVwk3H|kti0>n&{_s>NS`;BXxDeY?n>Bh*S>UsGcX_<?kFK
z4{__I<n5Wr+WitQe5YwR4DrI&aj~b3?<3=gkQN+Z?y!LcZmu_5vuxCgOuHsIZl$08
zO97q_!-djn<};<UH0H&#oHxB97pip;E8~!}vnPLAhsQqP$%V0S$3W(iP^qSl)r$0}
z%-Y<M9UHHxJKPc8NiT32lx8aIXvLpp3Hi_3m6oc}cnyh=E6-<PhDI`SdGm6H`iu+!
zyGePF^ktB;<xM~H(W9%^+0Cq>tcONC2b!sT${!<mqatMS)8@^4HdiX<H`Xon{uwEv
z?qc<=5BIG^PB{yVqQS%wF0f;}PDHZ_rL^^Rd_A1A6+LQVB2InqHSFx$Vr}2JqvE<P
z?S6SudH!uJ6O3&0e#}A<1yVgjLhwRG1GkPsk4D>N>+zws$4NiVOuj#?q~pzo9+x?3
z(R3^e+!@`NDix!*ZZmoR{{!uGciz8q+xpP7Gco(AAg%!D{*uBUzC6#TI=BL0-{1S;
zNXpvQH`ZG<p*knlC=r`v#w1|-WdDsu*sQiEZ-xRO!sVf@PgS*MPET<&YP@}oUsBmq
zpEspyx2q_DC&~Oh6Lv<2LPge!(-dC_Uz@!rC%G!|9YK%Tc?9_}fnV?MMSd?Th2+t9
zlPWKlu<kh&DS}om!~z8JMk53s0on}!#m8QPR}qDVidDj|)7LpALVxL##&n$VuncLv
zNp(DaFW=#pWS5ndO>Ze=tJ-x>I+`hnivjCB=R{UC_y3Dv(5Uk`{8SN_$zmBa9{iO6
z7;x*X2?xu&S1nP!_?B$@i}xpK^9N^l_>^qAXVojBh&y3}vRld5JfW{GM@lGvvIo}^
zp$LECDOhkQZu?7rtPJoIB+yih0xd2jV^qxIZa(xKTR1y@?3(FmBPHo}(20VRcAB6G
z(?qx<SX1&3#5YgHv%j=@kLHcEwyTw7t>-NC;pjnVsUB>O>SVlkNAW&fV8-bknTM^N
zhx^e7TNLb}^+62X`*6)zZ~;j@qviFNy_~PHbN@1{r58xYGI@46z0`sUvYOrYMDEcW
zgLEO?-IUZTLobW}zVsGb(~aoxYCl-0Yuthfk!ByyZe8nKo$VAn7MuDjlG)1RV{emj
zBMQa!bav{JnfFA|UXc*p%toIjQL=CqteZC4kwoylYFcj^RUY;aY;}IW152Zi5_LqR
z``w+)xNVgo&u1j6toDNiiL$L~onAS2eb$@vQ^alK8;Bb`_YAQVq@ybBcSzK*_z^@P
zE$bmcdGEcmmPf<ZA7H*hM*WRSn=;4V5)<YkN`QL5p7s*?*{|>M&4UB1nFwfbsk{}6
zl{b#dhBsbLS9?1;%p<^JOI1^AIcC>of-<%)ze+$QlvfjP!s2@SoUJO;0g(Kv?|lk%
zp=)k46gOOzw9P9mD=-tR^!^TZU_@R4K;>{Hos<PQ%^{6zT%(R4j2!E@-T1rHN44GM
zUj?staTj97xcq|eTdWsJg8AE|;_I*5^TUtH-mFAcau0_1=Wj~Q{#KPafy5Vj#%0V-
z^NUY`Ul9nDy^p%SjJ#SIasH!$1T`bXG+5`nazGc`26S#`%HSO`w<ncban-W6#O!pT
zsuufkL<?4UD7FLjy}i^tTZ+C+?*f@?kA==ix`)XNTH|k%I$PE{1y+u(OzFnP>6)7h
zy(-^czTAoufe7ummZi%;%N(ZR)$^f}Zx}bWZu2vS)?QI8Jg8y(Q=<~+7$`zXhJ?g6
zWgh;0qVd9Fqp7!@%tW#4D+dKFn1zqfPOL+AljUNo<Yo8a5C@t>YUoQ4J1WBWst1v|
zJED{AWa5T%wEaNmvf=6dvUyyErjuV2MjN<Clv`0<B;e9k!*E(!Q0HXBxgJc&xja8N
zb&>4u>UgnVpCsK`QF!2z56VS%$S);x&ex`$#~X^{-b?xYHjT*pI7}t=OY&g0caH0z
z6a$UnWt5@TbQh<$rZtGPWTnLW=GJ=JXujl1ji6%$G8ldxZyOvFi5&S!89Zg18g;qv
zvq&svc(`M0TU^v<e7|ur0hbzt8uwJrcB+VZ%{kpuMewZ&J03<Z@VYLL=9EW>qIRXn
zaC_sBJ9Lauhmo!Mx&_t&XTiORTewAZG|n5MD063ug3x?{!{OP$WtHAIveuK$oJsA&
zcN6p~r0_q!u=TNENXGY88#|Ya63u;piCxx3oQ2C&hSX6l0euz?BC#Hodzrhu5{wIS
z_K8`R_$+*&VQplCb^BW?lkhZ|AId`HjTW$(^J7a#-h}oY>9!Yip_k6|*ipAOkh1ng
z+RZ&Fv~1f-H&gNXJumlBhR@rFja^~nHL}__Wj6HtlT6d7S><(C{7jZ+LdQ&3z`emr
z(!^DkxWOU!6)C+bW*B6|(&H)7Y(<MgWiYB_usC^kOkiLH$l-RK>^bLV1k3Y$5BMD!
zIi~>-Q2ZM|25aPfe@?u1L~n+U(w{Cq`{>6UQm~xd{Uum<o9gOwm|5=N`}crV(juxD
zH<%)si81}4BNdrS$~M9vLA>Ym<yBo>qon$B0hVHa_BDN2w4G@QPauWt%}~m8P}P!P
z#Mf8B2^_v67JlE}e|fY{qVy@<3CO6VC}POT_mD#uE6?!cNXb~hoC@})ev8`r6EWLc
z?K_O8R@(GM<(XeP==s9a6-9Om1WcX1nbL<*TA7-`0QR6TI3Ls9KRM1NEM)2Swc17c
z`4N8yUrB``lUkb=4>!HVYf&WptL~b1V~bLS(4*KK*zdQ#gG}0v-0mE_SE21{H&Gu_
z5S7kiPb5dh7B%(#PQ&#Hf*ud^JYur8P5P@|*v@^z8JMlJjk2HrNq*;ayZYSt3X}0U
z`z~Z0RSs%>fI|$o53ir&#M~PNI?waNdpTJArq1C8ENx>N()79ItNyc~Ma)!P4o$wl
z$b0~!lAt$9J22?Axl$J`@^0%HB2Ie9fB4$UR`2OD`dp)j6aC9PwX;}FWvOYl{RtGc
z19Y!mT~G{oD-%kIO1_YYWyqPN&ZC}oM8Gl0M|X&%L{O`E@26JmDA(%`V|Ug@CU`rI
zjPxt#1EUvoE}V)Zv?mwYWJzb}zml}pXojY^`pz$c*B|hhk&E^YBiOf!Xg4&<ZSQhv
zH9Bm64B54bt%5N7@|s1t9w+`yQzZ4ag>;p{G4YWy4r7H<^qg@LkLjQFQ09ph+7Z5O
z!9TO((RYfCtJyJmZ{7~q-<sY*(zERQLv`iEt%}5B`^94Vep{rCh@c<6Fgb5%3+uv~
zr)D8IWj#_}wr8r$6nGN{iP*IIY24tQSFZ+hh6^~z;mJMbjS4w^SRvU^Ww7W<h`JJ|
zWv^rm?R6rGqE;VtapX3i_zog&jOiXCuZ&-|gotLCl;U&{k)K*|#QP*JGboGW1Y)nd
zjJX(jp(e(f#u}n^k)I<jUhx}|D*gM>u~Gs*aYwp9G(AM4zUGFVCsX2zyt)8wa95dq
ze%vsqm5FEJA|M({G@nJ$f~^@)ZJ?A5J0fCFxta<!$Fq6u;2xe&2ejBf*v3n1xQq|D
z+L<)(Nor&eOaf?e5z@^`v=w=HXT0$Eg@Jbc(#~zQ!FhHv-0XFtGe5rr*G04Jxbnng
z)=)P_OUW<X2P83W{iU8i{dUXr0<YjDk7X8~zOCF%uS-~!xv*-ePz`9-N_3n>>{LsC
znHC9{k?v~{+y<l-WjO1(&-{ibze}V>yK|9RC%(J`6e8=W0mi1Hx`d@W-@F2AeNhqZ
zKMHcp(`@Glc<I$%@7`MJ{Emd%vZbT3@qpTObdHyln!b5l^MxsU)AU2izJK7RuN@>e
z)NM}Pi0rFB={aecPSjip+!06Teo07rJbx9((?E>qWEPo6Y|G<bbB`p!bv|Dbzn&?g
zy*{X&G_gC&$*w#6ut3gUw-?U<gK&?=+$$H`^4@^?82%pstw2)0z82)sqkpAS(&?R-
zFyEfBlK;SAdDl6-T&X8Gw4b9fd~@D+isQ?$7o;)bz(wMN$2p4WzNaItJv>}!y6Qyv
z(G0yfu4eiRbDnr%6YMn|Sv}jc`lZvZX}F)d)o=9u+93Mca2}6=n`z%^yZEer4<~I~
zo8<heW3qULK1CI-3#62J_H(eGPVWOhm&QVy50Tf$hP^m{O8nm^@A1y7@4ajz`_jv0
zYN*H8#(8J(z97Ak=7N*s-tfy|$!k9>_uoB_x5w-C@jgu1{(51d&(aRw6qb00@L^#u
zeZj9uV`(2cz7F9K-T`#SKkJgYarW}1={JA(`)T#US{+54PG3a%AfQt62*(sG%1)%t
zeuEi3C@wCf8PT;0DPgnAbtE1Eph{d6*($LQtr4Gha^JfT9_6}1t0nd*cFRb#k!oiS
z9F2m;z;VWA>_$H3{wxGYpW<f5m>EeAjUtTl#ktnehqMyVf*+ZV1~-b-*?CzUK)5Ar
zC#Ndh&lxnLFiH!)mQmou8Oq?XDh}}DII>qjonG**@E)ZzN|vQl&Y!<9oj7$Obc%u+
z90~VlQ-(Uia!#>kOd6wuhC|3+V%-=ST{s^}<0-2pJBE>Alsh#roJmQ?5-sqeyc77x
zEV7aI$rGoh<56S@_wdD!#p5)TKE;DE#slRtij<R#QDNDi1AQ3(8C@u^oNkmu#`0xZ
z7RK(@#jJGJiAE_36sIUCDnT4|*iFKt1n+Df&iFcn&kGn+)-GL`EVIOrd4Joc<mr?5
zAB3+wO0aj<lj+6N(QX&R;9Ka(9}PMKD}1a5f$$#v5?}puz*oMQP8>hkCC|>R*!Sl|
zjgDq#neoa{ZK)@(Gce)Xj8{9B<T7hFdEvPX3^kl3;6Y(><*RrdBPl}|94vcyZ+#tG
z`&)<4+VRy-<{b9J{`<>y>wC{NSdQ-kd(-&2_ZOvOZ^*y->6_tWHo;xoQljNoS!%&I
z0X}34kNW%_Afi*>YrvD5&Pb)X&9ad1xjx%jtLd}k2(#41BI}V$Dw_ng?uT{!Sa~8t
zBIorOe8iB9w<Lb#iHjG`M{YP>8O&02c!Ux64C7{e7&t(7=rYgA-*d*uogK^eMfOYH
zV<|9vbUx0+r~%tw!eC)XIyk|MF%BWiy^W>j%pQ5xSy*Haa@FHH4#Km7t0e+AH-kg+
z4iS<bhu9dP#Svvtj{H_mpRBX)%vl47>^F3(G7g!E42uWib#R1t4YxsV$U8%ha^9-V
z9r)WL3Z7=Z;EP*ZJKb|erHQY8q_#LHjvo7>aw58Nym2%gJ@#cCbLvld;Hyp-r`gEf
zan8Fm8~2yC2Oq};kl25}{t3IE3GE?IeA^tG)<~!Fr0+`Q!{Cujgi&uv|HYuiNI$UT
zKZkK^$h6~K@8q7kp5OBAwYtD4XX0p(O@FirgE7dVoIu*^$eS#W=FwZy7PP->*RC{9
zW!^dDVZKq*bSmGf1^11O?djI7yPaibHra_2!||0;waq-(8MCJ70m{s9ZeOaa(VaME
z$k&@28_mZzQ2oJc%oJQnTi=;(-@22!-)^uMnZBc~NVkIwPZ-2DXy6-b;EDc%KH3GH
zOnU`G-{FKBGq%R}*UZ52{T{P5IRg_p@6%(X%NYeX-)eHX_H1cQpxxBz8XW4}4v%<H
zr<j>y$LnaT^ELDs`o-Aml)ez6Ca>r1_jQb>-^kiKXrdfwNB)<dHpHo}t~k3i0qKmv
zm1BQ2^<vLb_c$(hlD9S?AM`gb$N)uqHyebfZFp~qBcaM70sKmO`W%Nf<fMV_d%s)N
zskFDmg=g{eEOAQ9yr}+LxWaf|evlyE_>Fhkwq5U@4_L03d*0m-OPW1T*!zBAc|RdJ
zCvM^ew;sKa5Hh!Y8N9miR`K<(I)2i>6lcublF9&s7B$5s{oYemIVP2nj(e%Q9(z9e
z^teFF;OAt%C4utBv>mi<eT*lKx6od{=Gc_pF8?V$%A4nTrrmWVhWyO*Q^uz9<vhzT
zaw~_~!-wgY;kWvTc4j7{8JymI9z7H<Ee#MMuuJdx_&R`FY4Z!=ZPHlU;O7bpEe;BM
zpMNaaoBuD#`{ipd+rmie;Pf;Z?WVJ-UlZqDU~d~7<<Ji1@8t)FgK%E%9rWx&!JntS
z)Y-37&QAl%GQKM>%O?CBpLrYDpXQswA$&O4Pt%V@<3q>SAsoUlgcGYP)79U6H+}W}
zx6|o!=cX@a#e-tglSs~#ZOaJ~NC}ujf#eVy1_03rvtLGuplDE>mirVg%i%og(uQM}
z7RAJJgx&YcM$i$$3AmIY#uUKm^+lAeju1Y3wuY5sQ6w2PFqaDM6sq}MOJs0(Nte-=
z-oq)y@Z0Yy{30w<R*lr;oEbPC$s+>!)EQq-9wv<hMqQ3CqY!<ty3|JEScZ@sgbrq=
z^~06F&UOgC49IJ%t5wFxKOLzWWC-nb6mbeAJ;J&7;9dh$?sK|P@EBMap{O$y_~r}0
z&}}%HY?9sAo8IxGCwjl~R=_&v99|Sn3TH=cCmpo1VY(6D6a<Pd1yMR?oH1;v$1WLS
zMoWN_xE-b4$n1_NO*t8Gy5V}_mVP;EBkDRBr>&qBlfg@&y(2OMaI;U+#=Xty!QqUr
zL-@>qQm+r3JG<JMj}c|2<Kb!cU)fsUsxj!V##x)b%K*R-#t`8gAJNb45oS-n)8Uuj
ze)oM1OK5|iFrpYiYs5NZ)Mj`Veu!7sAQisCI4&*Vn_wA5qu@6;H>Q93Z~v{|kH2yF
zi*S7Hhvoj>XS?@&sP@3CS8ory>Gf(pPud5CPYnm*bWm90_SpY{`T7;eEqC+QxPA**
zzWckdn_myjSKz!LuMF8~I9l}EI!k=ZEAqXVPB7S?S?N+`oy~RTbZ2&)TRk^jx^%JM
zeujMgpCSIcZ@-CLwOYrLS)pVGX^}a|9t__G3+zOus>3953}*tvYxxOwT9z7wb0kYR
znt@I(0^~W*@fQv$djgSfwzsz{Gm-P$H`7bG4P4^5=OEx1+uF9Y+Jo{mGRfMx)uzEn
zb^rdu8rRKKG1GThMsCPXk2|y9vW^BpbVY#hzsjeTZH{!%68Vu)cgT{1@C?6o0Cp+H
z)1|%T-Z(zg0mlbN0KQsh+9-Z$idU{g9;=+1dkpvXQycNz(Lxx3)m>gx$Fr;@`%&k;
z>b(IZO2%#X5PIAN6=U<yErGEajN@oOi&2}Aco@+~9{Bu_%gKS-*vs@@fqj=tyBrs;
z2hEmS%W{m;=X{mERU&+=II}oq%}i1c($S{WHyKqOUb}X6x^nq)<C(F{VZ-4=k2pKe
zR<asIa_zE%mcY~>PM$D;kHhz9=xxSX+QHmU&O{f{#(Xo$?aHVxm@z2tI#W1<QZMq+
zf5_5iI$JHF4zjc*edO$!)s~ISyb^rOnk*e(jQzLTH<OOGf_T2qbO?Qc6IolRGda&W
zjn180D_&+ba%ia|Y3TRbk39ABCNS3z+|f-t9kb|0T@EMG$m7c|t#I+8D?|)^Bz2NF
z+9}$iKVHzi*pk+${AtW;+_~R8&`=td7$sk53T+>%IOnBT=#3um#53p_9PQFu9NP`)
zmuBF>nI6TtZZ4pioMzl<-<^5B+a@i|svbvvXhL^g_5jn~lnG9q0afG@Am+3;dNv<N
z{9%Y5u*7kHZ@dM@{k>_uy!Sd7dF+X^FaAOAx4$+I<<50*ahYPZos1<&qr*p6LyyJP
zwVW+!wyO*6k8_`N?P34T&--i`b?cbY7yagY<M}uu%P9Uh7u+5v{=CEKHS7rD(kYig
z7R&6cC2sZC{Ax!?nfhMZPzPzOSYr2R+8YP50jf`1CIi~m)>fO3nV0&>n8gV<w2}e5
zir8!6=bpa{ye_RlyAeKT_}uXOu=tGj0NyTL^);XVLKt<w<U3@IUk{|cKmGmpoWG>U
z`@$&Ko-pp_d_4PWf~Abh`9ZM6dEc<4aR?t5J{8ShN*#xA2#4@qV3ya_Z@!t18lhK3
zU@RlkB_bj?T1qcnd@$KOA5N>sRX~P_n~oR(cdV;DjGc90C7w?tBY{~$6b%YG#rA0h
zSR*&b$DFWCDS8?q^$EvS-X-Bu!d6aMM(0cwT{9Dg5@z2i&Y2;w`WSa@`&@kaMU<!D
z9)zp7J&K|=;z*M66$K~nk0l<GN8%r+=mkFt;FI7mMujF)(l8Qk%zTOx83CN~2WNN~
zfm!Ak{7ZvHV9zaio7sdORaW!PR+n7J`^t@;iAOmbit?AaH=IS373uX}K>h^#mLvD1
z-<f6+-i>PO46T%3exgvM9gWDX!ke<BG}EIf@ta%L$LD}Fic~$#j=d-ga5S<NZPmXr
znyqHj_lLQcQctH&oSIIbvZ*+u%~0@bj0v5z5A^_F@I~s-5+xlOT%}^n6l6r8l-m~K
z5DwvUf+r83OgDeLK5cDmmX};uJ3oE%)z>xp&YkP80`z&t!L{>iosIA`{g(slVfb&J
zPRrgG$Bv(xEORvWDLGfiK7M`w!Ory4^`EAH{^x(0{`kWW(~m#>)Z-67{5ajdeW!*-
z1}^;^AF~AO@Bi-i)Bp58|4-9bUtJq~^AJ7?^xJ(vXmeH#IT!XfZP$r8zVq`tjoZ89
zH)MeE`&D2!&!2~Z*$D3gduj8w;a%QaU|}Qt7xRGmf;bC3mpJ>wL9nE=H~d+><X(k<
zQQ^Hdo5u2|_wKHD_6Vl{hl}OTIBU#QvCP{^PJqZ5`MT#lGf+l?zUkyaM(G-p6C`Z_
zlt>2f7%*c=LxyL|jTw|Ao_7akGnR~_`H>OEh-pbU4tdS}Ugz&9IWqW|O=2eQ>Y1~}
z9Z36Wa4J3oro2NoUE!#hWz@0s;pxbuV-V80b7#ZbRwMTwAI4BJUh*`PUX7R$wm7|t
zNMK+F?`DxO<Tdh;rK>H&x5#ggBX{A49E00ioC8m1Lu+K=)WM@i&l?^*h^&-rv!>vD
z?8M2)Ur(mncke`Q+w4ppfox`0k=ZJh*CRtV@05k{c{q<|#P)-H`_snAw?y3`n+aoJ
z!%M+TI*akUC3Zch+xpMD%6@d7$&)Be-LdcOzB{KihsK>@@BQIS80}yIIFd2jZgMTn
zmO14`4?063?~^_0FJFCqt+Tz<h4y*>!GjLgX!?ow{KXfcEeF6D$f47Sn<VqBC7AIc
zZ>O!L&hFjY>Z~x%OAcJKpvc#pUh0;NJRB!zGFCoN5RO^K`R;|A4D<!ZFOD9WPM;YL
zPR|S+>Hfe;HyS|7Y_{1bH(iVqtuqkQ_T@oOkq<o8zjo1fn6#Ir7U?f$?|`n}k$2iZ
z=x0e+JGBfaClWoyj4x$CZ}CRj(M!;z?E;OC`k#BF9yxsJ4BOiV$Zbw{?%bKKU%xTk
zxPEK8duJp3Wn<b(yXRom{yD(wywd+l<M52&+?hTJ&^4<^pOyD$LoGtz^IY5Id{*X%
z;bHRa%+27A7W%rt(P!Xc^^T{*0W#`Y-{VxVYDr3!v_{(-?Yep&9+I%Y(nj$NVY%M>
zY}EZYU#|BDu>X4RJFg4-VQ(5s5U&Fq@}mW!hyLS;w{W~#DpGiu^T*nP2y?#>QmvCx
zKMRcqC-A%!@0I^WXFgf_tfO|J&n)-G{qf#H(=HjB{_{$JOM-!!_%HdAGgY6}U-WZZ
z7^pw(Qr#`(UFusKrkCQ4V?`CYb8tS-^Wj*nbG2;^o@hA2UOd5my1oxcZ*Q8*_1=5n
zpgfj1`{RBt@Vc`6ny@#|H-*1`&{p2BO};xUdA(1b>f+Vl_h*n>quLI5_7@2I+v1zr
z<=(izB>XkPLFN1<;X}fQPVZfS<a{{e>ktm%eIc^xrSHC;zWMv#PbbeX1Vk8&Qb74I
za*A+hlo;iJazmKqG~ooIfKW0icN8p26XosX$tWDT#}V?(C@q`LCN2lavIK?EdPDJw
zKpF+d_jFmV5Ddlp$aFl)+uDVT(}nZrYpkGf`8FC6_$W9r@Q~^`tc>vGkh8Q+mp+hw
z(xG6Q{d8jGbeFIh>+O%_YnG3>b>biF-0zIB=kSGV1oRg<64oik(yDM8)F=v^Oq2=3
z$QUWm#^o@wtV}CWXeqEAIhhBPH%4P?>r>`9+d7Ih`I}j`xv?JQaI=P?WK<C#uT&wU
z#MoHg>@y5^C}hv81fw~6bu??z>t^kV$JrQyRUx4G89~W(WywfOz);-0$9Q7oYUz+T
zH4Y?y^r4_yzGdgWeZ5j|iN{FJNoL=&8h3)9x=j{MPv{aZNuOdqmbQqZoa<zf`(I9v
zA3mNo?yOHcTMS9b`VbD`GXnfYAA4~B{&ero-7fpLe*4~7p6agzj2{2(Z+_Ed3@uH?
z5Ht*`V~L`>x9?1==hixEzMVA8c5%>K;`Q#GJJaT7_*MAqxpS-2+W8CRM~rxQ6k`Q`
ziPtl5F;wCk(%KHM{qe^iI}_t?7(V%oF9t3jR5E<eRkK?&I3G&Go@-$~aa_VHFaAQ<
zo7}r77_YoHAKwK=S>HXpFYLa%9QJ+--8&#SybFx@^IMfGBr>k+B+2)A^v&bbsgsf8
zj@j$yL}zl~fudHH%eh%qj%HGs-2oUH$%<06d?EQzf3JMid?G7yE+lU<0-n&^^Q`m>
zf2i?1N6*J2&&D|*|BNzN?rI938B4@5qLVkwt|wa&z|1aTBn4zHa-*3r$G)7)6ETG2
zpa%BDC-X?1%(T9-85wLV_%itpyzNCq4kAwur`DK70!KKw*JY<$mfRyl2ghCKQxcaR
z82L;!&_5+V<_4qh<3~@!FQ0Tk5MY=$C}AZs`Ge5Yeq_=^Tkm;?>^U3)M@oarraG}a
zWh*lwkXXyfe75{|6ls3G2a5ZA?9VY_DcAD61D^6zf9N{~im9>jleD<S$3@c`G+NRb
zVO&dR#L0hcL)`q-k8*m>c-@&>+TD@RgZ%&SQ4Qt>pRKK(>uf$9>gKIm(=7urx>R9s
zd+{vw8C@Z?6gm?m@2Cro+822)5AEZ0WM%Eja%t-uThr#&cK5j@Bcr$Fd&%E5*f)Q(
z7@wxio#-<&PvB49Eu>Hn^X;Tdn3=^u@B6zroj7CVtIe4exV@e4V-PcY#q3Kn)TGya
z>hhe^jZCgi$o&Qf=5cY;#z*@d&b~o;1MN7A=JuLvxTv!)LrZ)>J$1#04q!_kfOqRI
zeWm@M<9j$8ILLkT;V+*$1MOR!w4|dsKS_JlzMnpsoli*{e?V9Ir5xa~ET^i@wIlS<
z2lOxXZnh@pe4WNYYRtGCPO_cwBTf<n{M0|(1of&Np+^Vj<+-}Wvy@?RdL7HPeqCKJ
z<*Qyd!_q$A1WO#DZD-DZd9QtQhP`R+eeXlTa3s$8_J1JP%ekV#r)A9%58TBg&-?C-
zj=C@dxVJpdb^Bx5*pkoU_#Avee8J077`e8MH6*|04TnA|I4|!=9PP_@#-IjrCizmG
zEb+X*1iZHCMK^{In$xR?J;FDM)0<u3(RdzQI}@-va=sP!IRNp3c~I#LrNKUE`S*mk
zf&KA*?%n12-talXr@-rTpv9-dLB31g8tR_L`oUmt9v=#pG+%~;N$BRUQKpXz`}^$v
z>qB@u__gTvOVV6$`n7O-7dUi$9l{~J1?Z*+cGU=uSAO&T^v!R7GaZf6QG-Ag`ynV-
ziH-8Y=|bupg6U9hIfEQ2C>=$Ww2Y?Zd?A1q8qbI>`{Hm&8MPJV-MuI)oGAomX;MIC
zN{}Y}lCDpkq3jvWdvX{T2(CRwxSnz8a)(KiLS>nY#}SyTygXBa2)n_@9u|}+j<2<~
z^HJDNbruq5SC>%;UY4_YY-t=u4o2HNAM2MVttd?k1;h3#B#%*{!-x}XnqP`LojA1;
zrEX=i`7Uw2qv=w{Q^Da>6j(;0B+TbnFzS+U&X8o;utgcGf|{~Y=8|~+o=s2cU>dYn
zN1Xw8?8_>3l!fEDwloSlFr;w8eHo?nN$AnpbIHG(zL$n$Pd3T}g;e^-a$h}5kD_T=
zvP~l?gO@s~b5lD>-6*@ltCL)&tQ49k+`$X(S`ekGb#-JI!6-=H<v_T5>)!O}J||y`
zdkBZ{`9RtvL(bO5*0g?az5CHUtkdhSr?E0}Sw_gfgYMs$zHf$>(dp~!_a+<3U%h&D
zy7twzVJt{U_a6N6>Eq7iV35_v%?3NacCiK}%M;;E3>NsarA@snkFi{=(czD#>(_tk
z@=xXUf1B_bEOqE>wwz(nL#SVjS`FiILF;e4UL6d-q=)S0xhL$0CI5r)*n9mxc>GzO
zzeac8gZPryo?9>B&u+|+`>%U+AjrcD%9h`7b`H>=IZwzWjMryRpPDXRJm3DzD80<&
z@_`qZTtEgiVqd>K6*=nc>X{~A*)Y67yamBJW|EG)87K$LMptr8pluCadDlI%P36r<
z6vNR(e(?Tq?&PvE2>A?;VdOk{>iBg2!iC~=@7}!*qB84;yhY|C_n7%V_JDmbZ6|n_
zgGriro!Kfo_wDJ*Ku`X8ROiFiwn0B;zKmj#XNI$37=cGwJxnPYX&93Ge8wGbg|mC&
zjinif(?U7MOfq{kk#CcJ;u)NUb_OsRU?lD^0Ed4@wu7U5-BU+{yp_s&f1$Vhi`NYJ
zZ;w46&pck@EY-C49pIYGIL`Y&A(8zhjJlh7c@rG(^jA79@4n1$9MNddX@3|7^_j+m
z>%h=a#uAuh>M!$s;qX#d+GQO|x&JWr%`s}}P5Zmm@sS@r0d34=<aF}QVi5N(`CD1&
z3d8ZTl5f_D>272Iv%Iub^<Wu9vji;%DlKV_nR}L5JThIpcwsu1cBg;%_K}e}#x}P%
z`~9rL2<~}iU>Z5SgF=$G8JL`}W=N^mI>(Y1eDx=W`pw`XJv2fOco~#JuGg+R8%9}^
zR|mj_e(0EJfgJN|e@*(;;gY}ohpm)>P6qzykML2x#%ui12c@H2mKZ#FaxA;f@g$Bt
zl+eupyHlq)>Npi?2l+`eaSfhB8|Uz)H*nkuGF49Ls2h1ddKmpI<wa+C)FHFzf=R=m
zKPyw@X4{N9#EZJDUv$i5(6|EJIrZ^3@7ka3>eVZCEYh`_hXexX3zS({&iBGA?KNo@
zCux`7d->i{2P59{&H{VS!SZ}>cq@2KyxsR5ijp+7g@uO&7F=^!`mEp557`Fr0N2AQ
z=KX<N<L`wzO?9r_4Bn80{WTwS5JsI8a-?!E^p@^I%ZB2Vzua$|YTDY*aGt{7sxrQ3
z%V17?ynEdx#z@!EvZ!O3LxT|T(+(tPIY#<mZ#^yd-WCpuzc-!bd~eu$|8s;-iQA{7
z$)~TAy?N_9zXtr$JiPyE7{B={|2`=Re?{2aK9=)O+g?5o*h`Cd3BM*Bl;7KgUkV4M
zxrBFxeP7x`ID|uZ7dW<ZqWfoc)LxVyN{UY&C50noDB27V!x_e5RmCcS5J@37Legl=
zbnOU=MrIC0i4sORk{|_aC{I!HpV}yTyCWZmV)nR7C#b_D`E+DhQ?y*p*a)3)3MhJ2
zCX+rKmbjD^!u7y;F`^J=BQO)#6U=C!&~_AI@+9;=iXzD&XH*!&NS$Xn@2KA-T-Bfg
zdKgr2lh-1MV_C_zc@n~FM9O{2)=mU&%Nm+(#W849Vc$u7VWA_z9c|2v;{dZCT4%|9
zG0Mgf#>mmK=7?6DT<+BXlJil<c>{K3$~}U<XK)w_>B%a$UH&6@j3s}JFijmcedQYU
z%n-A+x!D;VJKOh5AML<<;*HU=mY-0Ed6xV;n$B#&qlr_OpQT=omqzmU-cXh(9(#-=
zhj0j=A(*YY7+drsZT7DT^gp};55`Z;inVe5pMLmZx^?5H@GwhA;T_NNeK=7bRY&;N
zM$Ylod2mDEKBMu@PK|%)#h~EYY|+iF^_+*NB_DY*`dz+!dAfM%!gPYs><~T~;LUpj
zfm+ym?_D4){=o2|;dQwE8hP%e#V-jT5?&<1cM0ScVL2a$V}ZODK9KMInC<eS17psf
zIX&6ui`+t-8cPo{Hrh+oa&D1njvt#&nU!?<O#8KgGK{__(!VR`$Y=fyxxuXUVL&G@
zEiw}sXULk2?L#ggWgS0pbQ;URos6tPelrtl$cSVyhRu~z$Ad@Yj>sJ37_!LL_Ez^5
z!mrI@Q(m~)*A#wIfAsK49VKS(m}R9r_#0sOH`8W)W4lXJ*ANaj0*ulmJ?YfZ5IkD0
z<RuIPf9NG<WXw7x5?4@OH2xyOpP6;1PoJ31B%E7anJ!;kovvJ3o7T>So+l%tqI+b>
zF5Q;*)KfA_zC*Se27Kjx(y%Q1)sV8(NjR_=^^o^>>my0QyZ6|UHI2w^t{I_s>+UrH
zAcgcDNBmg^mVa|?&U*~zLuS{WT7RiWp~=bLVNd!m^O48UC)APq%`d;Un*{Y`7My3C
zPRd~qG_<zolhnzkzTxhXn+<%saN+#)^|foA9j4yw=}R|}{@6e4cn#)Tk$*X%Ed5J{
z=8QSFdNy^oHl0fQ;uL+F02W=NOZX)nPFeeelDS7aANzE5xx&P0zcEuOIGIg#=g!^f
z#?70RyR~IxV8m~I<)ZuKX_KjQ<<!2=2LU-*$?e-)TTOH51+W_Z%W3ty(sL|v-ld3h
zP7gS9cC}7Eb;bd<I63-DLZxr=So&W>5*@Sc!aMS#TL~P^W?^>!y5xrzqwdGPV1je>
zlK-f~&>*3M=h{Z_40xxm=}D<$H2mUA{e!b|v{UgEPWkW?gZwN@dH+5gW+%ZOV-J(x
zvDVwk$THq@ynOk>^zFA-rr-VM>*?B+ORdk*Cq~+(Uhvr)mU{~<v{?dBuj;hCY`*ug
zU_Y*Z2D#lwY0sfv>1)qg&Q5(}!Fy@1ISru4koS~l@QH-bDLL)t*Z1(ACI9)jd!4v@
z>ExDtz``enen0bzd}BHrGWB9+<#P1QKu-DTd(YE0q{4x0#e%UPputz8UDYO$IAN1%
zug@y4Zf<VZS*tG%*?!>Hwls%#g{5qJ!$ELRyd`{Y@Hyl5G2rtk_uMuc2JEeucL}eD
zmwAYG(7VE$!Sg+#homX&!~d@V_R{3-=yM2vi|}*p;%%Pq4RrKFID|uZ1L!0Ner_VP
zo>{vvojZT7OY?Qfy!;4)pv%=%z>U&E35uXeKJ?rOyE?m)u+55xk`@IeN{4g8vu|BT
z&{0|{^io!z*I~jz<6TBo>$7*?2ZrV-c3t8iacpKw5i(*?8box$uJ`0Gz>N^xB{ZV6
zQktqTRnWCu$8ahw1fbMVnBYQbI~f7_XilCc56(CaFU}jY*5D>@xH0_rQLJ-l*_EYf
zCT+<R*pHt$nRDf&uz3)HJam~T4Y@Qstw!{eiz46=oI476dwY8D&`9JbRrDS|urc&@
zaUhs0D`i7oU9&!TJ$OXHebO>n%7^m4b$_e48WC%M2#T`jRiHx~2Ah>A2dB*3N_o^N
zh1g6jGvMysy&FY;qjmHobz!+jGwy8a&8P*JM^B8ROFdbh^6{vDcztnX%nCspiZKO-
zUPBoI?yC>&q7M$?5Dwv!fWCQd?Ocs~9l;w8YY;$}dY$3Mt(#qv5U({0?Dp;3;Zyf(
zNbU^X@GoE~$UArMl+XB%Fuq!T2>-<=7{`1c7ULyDpc%FtZ-?-C0j_*d0CRP}0&f$?
zhfckJNLcdwOM-|I1MfE=Us`@xQoRau-eiPPTLUS5{N}CO(|`M?f1UpL!;jPU{Rfe^
z7@3FBepg;`p1{b+uwMgq`h5rFkX`g~sgf=7^X(@Wx^HRg8bA|)A)HaOM``7JStjkm
zg|+G0)vMFmxpR?IhU1H|ynBzGS*c^rtP%;?U+MbwTg9EDh`}Bh@YDT<Zr|xFFmkGa
zTVx>lGK4dln&m@|so(<R7te~vi4#XV+lC?k|Igl^Jz17z>3!I8zw_X;$CP7bW>s}p
zcU2E4q6ZQNQhcN7O$108{{-<-qRdE203|cQ@PGzre3HQcfiMDs5vB*4@kI}OgiI6x
z%0xl}r0%Zj%I>PlF(Tak&V#$_`+e&k_ZuD_Q)Wz=u|JO!_uO;#*?aBboV(Wf_F9`d
z@-iQi#NpS_!&W+tUy^$dtX>MEr>>4>uKI=!re`koyK&<%Zr;4%`s_*n{^)xj#M|$_
zW$!cy7uFhM@)9BT{IYfd;mCKwE=u>BXM_XD5?<_5K-EW-#eddV{Sx1M?z{C{kX<~5
ztXJ}>#OQ308-4-p&m^uYqT@QaOl|bDb-R_$U6F-cMY-ti=%6Mk$=~^A-YC(eVLxTe
zR-=upJUPsmbCd$5qoz#2n5ge)hlXE>H(X}F+jnFKerNC^!V8R-=JX}u^sJXEU&cJ%
zQ}^#5nFz%T#QH6It;P#c81CK0gK*})szRvY9Rz_zyD~13l?iiPOMgI=Rh}kH@H;w2
zpI%yr-qM(EN&n6spzgG}baiBr4Cw1s$oGl_AdSgY*%j3!IL2M|Kw+QuR4?$=<CV7H
zHyu)t^t3e|W^90bOmr}Al#((bj6GS8F;&SYaL*)B%D6zLjOxMt!*!%@shu5Bf{={5
zM<|<h>3r`wRH$qs<x$-ml`X)yt!x6;Aj4yays?W^vez1<Z)rd5LF+2?zuE&k3p?>_
zNm`&5P+r)m*d)m3FqD!JyNk7=A@&(VJ?-P!N?)+{phHV)BN_$QrEGbRUqwZXUHwhF
ztkY`TuC}Mab@@FHJd`i(>bBK~+TPD)HXY-;i6T9xD=#hjAxVAi$IIu;Iv7E|p5vrZ
zdJ>b$eZIQat9E0&)5i=im*Fv}F{<y3HO9R+r6DcP4)!yT%#UX<Fo~;&Bx#6Xu(!YG
z$r#w|)tY-bq4V-SVvIBC0S{lE9qVA+^1tt|r^I?2ukTmww0`WHf)^=smG`r3uL7^C
z+*gk`wcbyUYHYm9@%dFy$@{qNNx1Hs>$W!u*VVt;z7DRto?pHa>ad{T3Apaot2b)#
z_J{As+wZ?;uP?Mtn@v*Iq$`>PLviE2gBV$!VuWLOV6;d<meC<i*@g(=1g*#lgx<pl
zGZ_WRCBp+^iFK?dQyF#hsq-OxAY>OWLtJv|$}Stih8GB1+q=<`5=%UIzYxG+Rf0JL
zlCFnz;$_Q3N-X&R@}?|gM;S(b!u@S+Z^yy@fy0|%I5Ptnj|L1PJeLS_s7%nr(;&2=
zr8+y`oNdcPh<M<OY0kJrMjJ^OUTP!4d=cJ*8Kta&<$NRuGYPl0cWj80pNS$JhnyRZ
z1~E83Qdt~RtE-*!9G+ciyipH4v>>g#;JWHWTVcS}q|{?*@M46z(A&Hq$QLkam=p5H
zVG&gaE~Xqkk|oP&Ol6cZW7CHIXp}QX_5D=+#)|>^mf4#avy@%I<R3x^tO<t9UOZcU
zm-44UQl;U87huST7umfp@5gdVr~7n4LBUG|_ciMs@}gld9LDFLeIBea$V-q4tldYb
z+3$Sk{n*{z_sc(CRN%p7Zwc;ozqr7=q80o5`}P9KnYsC*(KB`ZiKnfe-+DenlKqQ+
z@h{@%KmWx;FCN|?y!ti1_-~~gT&E*k2hvq(x&BFy?=0iVxvwku>t#F+U-ze@dXcui
zUB&Z~<6Ie$j{Ne~8aw(w{~1oR`cfTx8lVS}ewEz=sLoX<s3O1p7yH`K2h#IA^$0vR
zA%h@m2usF0K)?LfpV!3Fk2x>?d4BYDzw^?cX96B6JQMtqRC*fv1=m<NP~IlKB_#R`
zx)9;ay2`)XBP1NU)IvH<O?ujXY)dD_6P*{==t6brGbW05cjDmSKzdeJ&!9<s`pFk@
z@BTp7;w6R`dslT>_?%z1uR3Rm&WipaSyU~TWBti5+A|C*1RQeRINVd+w*0(8Ry=g@
z5ZaRNm_(oIr2MESZA3ag?cK#BBFRb`TAH$G1IV^^x9ykr=u*1Kb>jsuA>$w<r~}?9
z>_h6{m4Md@BnpHV-bZ+{B&os#Gc<N|FKU>h%YCY{k;h|Ne-C2yT_5PdCcN0#XH#2P
zey*iXt8@9%y2?BA<-3cezpG7Ou@P3y{y^!ZCog+uk^gGrM3%>=Gt5s;=^OW{JwMb(
zM45m*#2W;g0Ugy7sRX~o*JbF+wnwy|eN3G;-It_6zMjS%`s@i|1=X&M-Kwu4KAoS&
zA38B(n(>CGUPF0pZS_5NDH}V8@Pl}dK*AXA5Ijy6addLx-`DC_l~dG-dxN$`H)p~e
z<QkLyk|a)|3VVXGAlzss>;_N9QyR5Y_7CkLy8yZQZd*b30l%wwwv0v-wHM@Hu4|lR
zhwe>}UzH_!xc{lk*ppF+;*f-@Z5zqSH6};&iCjs>RqdS~h#bQ+9739jl1z}fd+)xU
zC25PX0j{Br)B$^iefo#+k<_!=2!+_|QunqA-}^_$mI?9{nXpG0BG`SEUE${^^->v(
zrzA-593w>PhIE*1+bv_YGxs3`<riPxvuEI7FplGsf!Y`H)1n&HZv>y~+d3}B{@QjO
zJl5AfD|qZbVq3>`=O4FSy(SOjM|FeJf0Eb-sw9>8xxSaK<-XL$c)cL~NOzG5*R17n
zalO8fmX7hgoyV&-t>AI!vGid3QJ!&;C;yS>YEOCaJmGo3_(ty}ugcq#XXLrpudRhf
zv40Qfcb>b1&*XO^i=u$sTu1Mdf0ps2Ez@4^S6EMT-T7z5b@@GN`#Si#*FOm_I{)W^
z*ImZ1Dj(N9Y5Qu(>%M+Imsrp1b=|&ccpfr;6YKpV@I1%CdS2_>btuGFLBTcPc3D@=
zQ;mEOjdmm6{oaRh>&~61qcG|s(#dS_%MFZblNP%9FaR-vZ1~7XsSI0gOkN;tu<1Jn
z4C_Rp5c>s)@>${?B;$t#!WXlt*W*Ilw4%$VatO8H7b8+y^X1Dzc{O5tZ`UtgyE3LQ
z0!=h24F(W0VE{4XaJuq*f+%8*FAPNFMsG%uZOX{q-P`dDKMYm~8P))BNWl%tPmeLL
zKPCgv#1zDij#HP5S(mCmWkcpw3bV;3?eij;*@G-Z-B1}Ap^#p@#K5@4AjAk;*uYJX
ziMh$9fsAH|J6_VL{<O2M<CFE#cnN_akHN1VQHiPS^8%(>Z`q)rUl0fblovP57=|ds
zL#xvz98uSGpe>Xfi{AC+<rah=#wnyEB$z$8R6mS(^7AatK5HJdsW4TgDQrNx5%{8`
z^c{PQ4M0C)Rnupj;5y2d1fH3FeD5R{6Y_BKf`WpV2;a53KOwn~?%t21qkDcHuVZ)j
zAa1{P$DRmyDDW&`y}#jj=of-ZJtugM144W8oM6^1&p^_$7(YZ8&*<u9FwZt#^6|p)
z^UuHV8Z2)pUV}BhgnV3AA5R47o;K|T^qYV_!f%(a*Z0fKr+v19)8y_8MM`r8R}MXs
z9F^c#wu;OG)nOwmzV?0fDXi^25v09r!w0Xt_@<pIO&-f1-u#yPOWn)d&VKQz#Jn(v
zlrm|J?jrrjx`xj2^4u?;^#@_jbHqfD&UqtOzOU~2yl$l<Hl)))e04hZ-r|_OKb4M`
zmk?Z>a})yU878oFEpo7~k_mWahnEKHFOA01`Ou$8B;Y;6HH^6=bEJdtvQ>ZRO-Xvy
zlpmq7TF#G`?T`k9Up+WHi2hcO8d!hf9G<a|fV?!tgS85sP5rDV>z`Y*s=vxh!icuw
zMg7ucF(C-=6~cF)uzr~2CZr>irFijv|9GJ1>)2~b9UVzG96>5hOwRDh3tNZ@@^L@W
zwk|sB?w87y92fF%-ltdldCo#t`}*WsbUuHgFLH+Ss~6DdcGL-I*T?hQ8l3*Y4Irh*
zgFe%SuyXG6L>oxCE4W4m>F6`$f~3rTYij(k{uARh$tK1lFTFk1wDB?={Z}Q}o<Z!2
zbb(mim2JWA9pncu$R&7=JFd~dMjpJSAcGJGqPNk69Ifk~Xh%mnHRFuQZG6tvFUa7i
z5JJ9uPw!}6wo|rAB;{rM30d_=SV7L=6@}hSdr&r$MOfF1wccoR>fli9ewj}Rv|aW$
z<I`)x=_oRApE7=igR%8^&heW}xff)X{@gdpjIqgYH08S;B@gxTIMb(lPw7%NU87IF
zS7bS0k{$O4zw`F8G8r!00)o~y4VO_~dYgH|i}qQI4^k6S(f6R94O<Skx~=}MJ1_2s
zf$j}#4xSraNL?M85pN!J3p{#O<B3U!5T4|l;~6rPb=kJ9cig*oWE%u>4v$gD=BfG+
zLYZeE&1(XLytGyKM0G#A9!QZL@kp$v1=ar5>#wF;$D@uv4pn~Dwd<ek<vAa^ti#3k
z`~=+Jp7223*<ZB@<H*Fc{_#OxS%!D=VnNCCI`_|NpZxgKfo#(Zm&gMhS-3CqcO=sp
z$i(hSxavBd0o<3>8sLn3CSb5$<7i0q6xpOIn~S**=5r?F=H!g@gj{8v70A!3OrLo=
z<aI9|&r%L)zv}ip@VI=QR~gR>l{{7bUpM){x@@mfzN>OvZk%(9)%Nv4zOO4@XW1Tw
zr<M6}>As2eei3+;?Z7WuA-)O<u0f?MJ_^lFE8h9&gE+i%J8CjGAe11zP&y%8Fc2|P
z`NwQPj3MrIh$;*KUJ_LIKFTkOC<ZDoI53DHa#$w@LWMAbCcR_;IlZreP&?Cc!Wv2;
zm4P!`FpD)wwuThi{`Rhn%wBY52=OA-Ghs2xthnNl#B5+lnne;wGL8t(hsJGEP36>E
zjc6+mX6u@K;w1})>U_0mxaxAQ`tp4)V-1-r0~uME!Hc17(o4#78n8)FQ8u;Nu#v^f
zk&ensMZAttCL@j)CXjZtOXXdL2OMp*U>}21ZR0x8&N_|~@AlI*%zWfU5k?=zDTY4h
zA-FIoXcJz*<T_@xB)yP9*Fl)Eb`M5xlXR*tFR(CfSx;;<RKKXrZ45BWn|-w@Wn*+Q
zo0}I<br~4U;<QmBBc48Btnl!s-x&w&lQ+f)-!VurHZhuP#2~Bkz%V{lJB($N&qgz~
z3nQYSpx|X9U$$k4RYCzh2n?Cc_~65j;;pytI345{_wwDlcLSn+ILx&~D)76&O!K$y
zyp_)kJ#Y5+_LJl~IdK}E3!a^>=MgU$KmFv>_~}o7TD{MB)9}iPFaHj&wy%Px={D!j
z?}R_A{d{qGbxU0Cy(m})gi%=!mH!OIKa}(d*Iqr19R6I}+)j_#jEkKWs_$;#{3LJZ
zzuKy@IG=~I4Ju+P)2npp27FI_F3BP8RlHj8tj80H`_;4Or7tjfhL_yri!MTbB%q9Z
z!B3$%0mz;{(i<SJ4iEOU9XQ;oJ!O=BdnY|I&^LK;jXu>?TF5uD;pM-*+w}=?G{8H3
zBE6C#1fFSJi+*OkLmMwzsRyCE@fJaEs>m+t#mjaGK8N9x{)8M50aZ9x$hS<JVe$xz
zb}|@A|2%d6=&DsX>gpvh|MAH#dMLUo>3AkWkPu=EqK~@ptVBN}%=M?Ae(w65oG`gy
zqPnOptTRHkK{BA5=gVP`Wv%otMh^Yw6RVfQI<dZSq8b+${mn71CjDyj*wClRCO*|~
zbSrv4BvoF^SardIWa8uEyN>X=`kIF{COjCp_fuDN8I-lk^YSHpjvMlK->GlNB|}gW
zH6g?CX8I&*kic^Dr%z;u452+4pOls3JmTPB-(w0A<OD)R<Arg+2DyelO<yrC`Hl1V
zSltU5<}is+lc4G!`jRpzztctE*ErFbK(Cfq-aIB%HiTTY1{lAEAlDa`4X_#`ZP^dx
zlM~h?2lk16NAY}O{84SP<z7Rd=3cQqp#Eh1L%8wcANeQ;?@7Wvav`Lv+fVzDn{IcF
zsS7fUdsNpXnVT}N{&QOE^4d5J*^J$SG_(ma(-(|)#wkRS?@#5E<4Mm4{XseqIfHU&
zEBcl8^jdx@gS;5a5K>%!|LDl~05)0n4WULQw@SuC%l4f5fxE&J{%nKE_L1F$=#Zt@
zk2~8teqIs6GQ(Qtc;MN@j)_dQ595a8mB$mZyPa0)*%I7~NWrnHeexvNRM%Cg((+mD
zSE!Dyrw8l#vQ>GqJ?T8#b<i3A50Ijg<N;Gmr>pxipKI)MB4w<O<%A}lJ(-qq<NRC~
z^|{*K&widIdKNQA$y{eMA8b-@kc%pSd{?sZTflg#HpXbaAIRgB;cJwh$p{dcxo#Je
zMv3^sy{6}d(&EX;Jvvwa(#Q6AT-BwLoz1el-)ToqU7mZ?_OeiwRgImOS@yG_l51T)
z_N(nxVO@sxtvdHAaaE?P+nb2%+TnHCUN>AP<JIk1aNYH6Z#rIwb|}PGLBUsn?%UQ*
zKi>c7qd2&ID>h|J%p7V@ilok=SK+~vTA-kPltnhwMSwC1NrT~PBUXx{R7Ib|TWYby
z=z;ViZagy&=O#^f5de9^{!)fq##CP^vSiJJManef5rkK7XUF6X9%%?($woevVou&D
zwRmh{WH~QM3SrZP03rlOzZd=fR?ZkyCZ2Jrba<rU{e{6Ot=es1FEqbkNCO7Lm$EUw
zFhIEu#h7#Y55Z*UW1we9k)DiWWTw6F1j7psnP?YOYvk4!%T1V|soE!r!kmqk#sQh8
zF$7`8b-W0|K(LHD&3>!V@(f3ZZBPSxh8t}n;}RpWt9qM!lQ9V?r3P|av9<-CXIzID
zhi9Bh9t>GXv@JZvI$hV#MuZF)jE)R-L_Ji#&SA9FRd?Eo|1e}cjwDC3)^uN|0hGo*
zuO?_eUI2}b2}hLUsi2_X<-*Sq-?-O6_iVe{iT#7Uc>Arl;>L{|ehKySKl^!ndGBrv
zj!#VfSEyc09NajFww@b2mswK-4=`Rr@nV;v{Cw>7^sL&5Km5Z#jDPtr{^g@(_;%y<
z@c4QOu(81BYD-+#=MScMQQ+qniBhU7SU>axQ64(|i1Zv@->RL;fXe?b4ja|yRrvTL
zPCm&XyC_ee)}2;r25@U|O#9rA{N|sYrQU(=#0znB3OqO_W7hRqDvuN;38s8;330&h
zJTJ4+KL}I3x4RcN4iCL{YO__3TQ?6(dO=Dt(FX5PCJwQN79<C{nAhczu0dYBv`5cC
zw?kLt#XjqI`31JVXW7uHsh5J9^)tyMy~E_m>V-Kv6J)~Xrcx_yr7xPu(sk_P^-e^A
z44IcQ;%x)*HyKa;a-BL8#Q<H3wCFTwWtC2ZEa*~r_;a2ziEQ9CYavY3ezZ3-F`<c<
zwhifWgn;9P|6;a8RZ~8aN}uR;mFI*}^&x$ry5{CD-Y)vu|MQv0E_BO0qQBL?ugjwq
z=-*1mu0HkODov)#{}E6}bpdinP`(<@Lx=L3{LH*mAAd4tvR@c$yzI{L!%J)WisN|A
zCAre?b=2le+8UWKqTg(r$QFSZv8Pn8>oH8oPKWK*d48J+=S?|Gf@5sq6-8Uq=bkji
zIMMh+UxuK9B;h4`Pi1l*;tQgdFpIq4XZ^Ur3G49AEi3C=?QCyr%;SlKcbdkW#{8N3
z9pcNRmWiGu6e))><FTa;FxK%Vv&XbPkqzBE!;dcPEk#?nPmo!g$GAS#Otuvu(IA2t
zlhk8J^@X&hpCHj_E83oVAP3++z~=EdQeHakGJJ1)?4`ZKaUE1n*kY1iWl*0kgkLMy
zSR;|LbK4TWk+PClP6Cs$kKJJsT5ZMs0U>@d-t<R(@p5FXLDm=bWIM`I9^^y2u!a-o
z>0{)<W~82^VHhSsEL%kPBl+6%Qsafg<k6J8j0e|4DIvO*t<tXUCBqeVUbNB0At9@p
zY26>IwggxAtNg08>(^X8<~CfVUH=5@=_;%rTR&c*%CCYaC+MG#lHS+kwX}r4aqOJE
zN*1=&Gu6+b)qPCryue$ImmD)rN89@v)l(ay(0AMu7w;5wkdpbwHH;OHIbBPc$VC5}
z_*(U0PP|z7dCD^wf|~n~b*3HlME4$jOIjux;02c6mnu_vWS{FV8-f2Gz}0rV+J@k|
zwtks-)#X15*U9xd$oe|97qYBxZxXK7{ZU($w&2Mq?euc+I<)hpW>HX3@bqYGMn}d_
ztKYSegF#?L5KCK1AhWcZRvcwOSh2Sikp`p`Nfb?tNGq&5hk}Vm5c(rZZ4yzFID@cJ
zmUwqzGoZX$X=U~*3a`pTdCr#wsqCT<qhso)PkH4GMT|Yt=4`Tam)CAGL$K-VD+7nC
z*u{{tQ7FT1W3`wzvm*WCMb}UsB$Gq?@nV9o2Qo|-GJH_3d9Yxlavgc1obyQ?>-DC+
zs!-xNhRj}ATV+K}<x9y&8f3*FrCf|slz6<ZIt^wbDy_<cL|ZPUtdkRMuWLw`Go#Zm
zU?v-c7UUZ*VyYK4tnEdZ6qI;g)bKpU%ZizhUhj)O)iL)Ue@eiNTN@oTl8V!=4F<eE
z@ctt7BK3ms>UI0R#-4x9$GKdTZzEZ9*O(nCdFcn|sq}6Dfi^g^5Y$Pv)OF0FbQqGh
z`c8E%C@6USu&L)JB-P=;Vcfa>mYy+-`10<T8NNvy1o&0tT0dhwhj^*i@k<Jxbr4-V
z!$14<)A;nW&-{V{PYRx!JhcnH9q@0&l}-OWqL2CYU<1{D_5G{ifpVV$$hE#*i*?b~
zH{r3@T>+QyOI&UIRp2M`iE^w!+OLDEZCn=ml(zhyWEbD~bcnIxB<|lo(KgWUIxm(x
zCZ`S$_w4zCenWn3dqXw-yN-^8hddi$3~wGD#Nmwtzr?RMnb4HpGUyBtUia?b_ls+e
z^?O^=JM^rOK6dZ^aeVgqm+{FbpT^xU@A(BfZN`iJ(O~H3%u@OhdIvAwdBM*MYY4B@
zCkfRkJwwm2d_jxOlUs(dKDEaR`X2ds(vddw8^Ua(Yk6&3>8pg7W6dz+B;4K34il(q
z&WF%<gm{JAC-22+dUx@X-ukZcM-Q^zJi`lzhysuo=(N&@)S>8XwN16x){c(hy}%3j
zRHD>O)<fB{?-<`Xf~xVtVV`2Z(#7ae$wTNizci-a+NfXldG<T`0iToZ=9kNB?P5Y{
zC5!cMuIJiSd4}XMxuAB>4ZV2P?$st$uNr%R$y1Ijp?09JXv6KD9ghR+-n=x|MR;cL
zJ3<;rA-uKj+`ePd#w6k*+Z&P!Iq^DzfI|Pb#|ZkU0MTW;Mb{jkoW$L`8Df`qr3}V9
z-XK(rHixu?2%)d>yke|FN{+@%k|Q)A9wv}pGmm@7w7tC@k8^Ac>(HuGMr<HDgBrli
zroPiYU4ZQAF}d;%Q~Mwf_XtE1Ix>C57-Za2Px{3=xHei7`-66+Z){u9SLCBg&}Mj1
zK}a=Z^YNR`HQXPNc-TY8e(7<O<QZByHVbn5o`%HMc+WAYJd{7xL4D_VdT4F9O=w4F
zqOw8uJ$4jjlQU}zb06Z7qho3Z--}GTQy)=ZzMIU@=5e^tHrM^4XAt&b+INJ5%yr*v
zPw<Josz2Ja^8BGp+L}JYLzFgz@FH9#>pio+o<rNKueW-g=&Rl9YO5+b2B@BsmW}bx
zb5i?m2kN)ls<hST#Wm^^J}>W&%WeI-^|aR=ySfjmvago^RluceX}7Vho*jNJ<i>A<
zy`|KJ>I$!&MOr|8u;X)*LE7`Q6}ltde|TNd4?N%L*Car>m$(mj)@Tf?4AOfts_bKv
zk@UH3bcD=$A_w;bp<kV^?jIA(k_B(ViH<{x5<V6WT|$=5RK`N%h{=}J;o=b}RPFaT
zRPDaLy$Y=N_p3rB!*$y_t~>U!@jA#*@F1!_dD8Z6$8~l7DqF#u1GK5)@l{apjA(UQ
zar-;(#ar*cXQOV8foOwPkinEQ6(LqIR(PnQ=%VzdVPzvtildcVowv~;BMC(iMHS+U
zuzgmGG16qv*pQT>dX8}>g_qf{HXvl^R}XH=Krv=ED53gVl5LZhC@mI2?|LR5;?$*J
zQyy~UY)C1|+E{ajFdiu4ocD_v8DNA<+}qoWtt}QOulWVUcsyK%Hk53V6|XCY@{x?x
zo5j>!AIV65er$bJR9oS;Zd-~KE$;3VQrz8(y9Xy&pm^{Wr&w?(#odcLv}kZAIK_&j
zxcg=A^K{2KbBy(nJgux`%*=m&Hufj&9SJ*k<g{140x{4gl%IE~Sl!mgl2AJQeN*W1
z?0cDaXXxCH!_fpo<KICfWgTy)f;fp~unWH2S*Vzc-rFQpGg8o^AOniHCb{@Ex9tI<
z*Oe0$9}Y(E>_Vx}F|Bu8wlau$9GGKz7OnAme=EhvhvqY^EEZ!rlLmEmEk~X0;{)kF
zVmo>8=-J^@?lNNRpUGix$Zn7kXU0VPpGq2A$hxfogpyMObK-1W-74Lfak?b|FFmvI
z&rUy1&|}1D0rlhLM5x4ly+mxz<G-zzpyw|5#fD+eagQ*BbG6Nrea&Y#tHlQbIc@*t
zGIBbl`LLPR4f`p_qu9`XD)AKmEE;&G5ahv_CV&DM+d_B;?s`8ZdD53H;oVt3Z{CNr
z9JN(sKAn4ku_;VM6592#tAUY$lm>B)6N9u#I(2>m&7LNYfrJRS?s7ry>ufnwrn9v>
zKJjO}8!8`@Hyic`fl(a@$Z1D5pnF4Cykk#R{f`Y7fP$7DEi)Vy-SIA$`jq=e+B%cC
zN0ykz!<*#wfD06$u(En4ldEj<tF(N##aFVw#>cBYqx^h1S2{~TeFr-%p~L{mnRe92
zu(W5DT+Q(%7W~#qW7FhJ$1z_H7I5Jh@oeem6l_#3Xd8v4&R}`NUI|pMHI-<BANHWP
z(&G9DH>v?Y^y&BSuQ~DaVM91YF~a+ROGZbz`Lt4j9DfBs;dtj->h>4ON};}u7wq8D
znjtOfLwP2%A#LW_d<5cMYoB=T2Z07RASGTof9jKJ)PoEW^~L^M;?J;%%ZN7_e&ULE
zsQhNhX}4OTl2oECA7*yKzZ7HN9I+<=mP?1@`7JS}@GkGBuN&hrDFj??#FB8o#&ia*
zw7N4puXovyBOA-S>$-Fd1fiuU1$|uX0fvO4%%;0SehY(J_Z2ZO_%<c$#%<zNcn6?i
z;-g&jkL294;O6@FiV5fo1HH?CQ7zm0mn_cqFC^$054S9y5uruo8W>ON@-?p4-gS)4
zQ542>wEHlUSk7cxWtyDd(v0#o&OYV1`WE*xDY;^I&2SCbeFOP6ad-3GrUw+*-fsM)
z0?0^Sz2*Rq<!qlb>gb$^?f5A+#r{K0vO2O_HCx&h_uNGoS(|fkkH17u7E&si#{(Yc
zXLe((5N{lCmOl;`#0MyFvaP&lEsf4*Pp<l!E6hzKeT9zT;~*6}$Vrtca<>0~%Pn+_
zoUdFM+n;{YqG1Z{^@8#{DaI?8G@sPXCALswq1GApOz;ihU3Ux-9U(8t+S`%Kp9+f(
zDAnB^d6|IqyXzCWmL_(8U)s+L#F?rNp1#wc5yL7+R5(g`1_YX=93%O}%mk*Es^Xd^
z?x_!^FhtdKWV<sBiO$pf{7Z%ev8iqH44-B;0f~>CKY5|)n<{s^KO^C9{m@^##k4VD
zxc$z(7kQUh3v02YW5Ls4DkQdXH9N=rLcp_FEX^`F&9?ra+Tw(CVlE;|rC$@ErvsNj
zT$P^)!%lQhgop1{9eZnn73NMe1;c@FKau<MXvD&I7tJOT*L`oD!fx1DZr<$;;riaU
zQ~L%0dJMB%e)tu$eX_ODDLBc0TO14<9r@8Y%*tN6^jrBIA6~(mof!s9`H0yOtfCwM
zO(9xzQa!UkGL;N8?HxJ)2Mu3MMXrTHv=NqGsPjnf{*_JTrS*!3FF)|30V~C=Mj%&=
zA#5I{>`Y;Uan^$D=oSwtgnG-?&kyc6YtEpBz2;z2Gu#i6Uih&^5<55yDtoXg-ZvUA
z!uyc}0~>l$TA=Gn&RU<GO$Y1!>iw3F;5ISlE=6yzr7iLgNh4cWY@2^ePK9qJ{lGP%
zaVz)(S2t;N7<P+Ltd4-<bp9YMyPCE|(nR40EBGZftqxyy_*ZcwF7Ioby=k^SFT2q^
zXJDe=#T%xa4~^DWDQx8}nGB#EZgd#pd;qywwOFsu5ZIr*jBv97cyVbeBo(9ztM(U4
zD}m%JY&bHBZCvee_=sYEdX_3N7ggLWmHYRWmqRz-n<UJ@L)31v_4_05N0o<*8M6UE
z;L~}F-?j2;vy23OHPVD(58aUy__m#RRexzh{E<QY_GR8o%6VXT+LyQoovJ?#J##$m
zbU!>F+^R-0XMb^WApo+SHZBFcx=YyrjRL4E=DU|L9-c(e=PnVf@NWHk#4g};k7puy
zn8jc{X!=UlKcE(GIHFR=Au#A|w!sQMo~UjXCyl)aMu|$r7<@O?uqbX|#N)_2NyL}v
zb6lMkJaIh#d=I(ft%8JAFb4cg8vik7@g7RjQFw6|m=$-4=Ya&Y&u^A=qNJqkUj=kG
zh)0TtaLWK8{<#u(QMuE6hsg8+Sa^&D$hk?y&IbFDg(J9*{_2*)cN}+)dOc@JO}>We
zvE^w^dwxt74U6uIHhhc%ZP>n}ProhxDNhU&*OruLdg#vKa5F4(4B;$|uCd}WloRKf
zU*lj~`hnEJJkj?G%-H>5-sDtrxX6Ulz49UzU1uXpy3~6W8|6A8|El7U;ytJD|D=HU
zzm>)7XKIz^JlgFFQ!8!L4v%&C>%nJ&D9&4G#*i|f8njcsIh*ZRkcU_$m#E)GoocIM
z)<3llXY%K|Pzr0`@!|j8Q@@L_TAp^lzVf0>_4f9L`|eInxvq2vF$7$g_<o+CirOoJ
zlE>tIbMn|zyL@ev5Z>5*5<YyYmeXii*kv)shu?~qSvxG4?+={airwU@`S$TQz^ny-
z+O5mCHD?P+a#;5YBU=gABYDS_J%fZ9La1j?^M|H_Bp%nfu*j6OCdOr{_i(#uc0pGt
zTebwLfs$5%HvK2*Pfe9<*OxF&9J}?7oaFd8TBPgJLH29taQ;z@&vrfl7UCD;_gz?%
z==OAo??=eb;Xv7AcxOpN7eEJqAHI}s2O~4V&53#KgfyV5?}e5palG2@VvJm)=;xr6
zBPh6{FX6IG)w9feINmS@qkcKi;=aMxO7tUIG$e=eR<^3Nn@6~IPl3Qi{fNoz>?7kP
zt2ytA@ciT3?AVeZiq9i6sTl)Xw{!v7NJ4Sv?aR&_VuP$R;KHk-&I&8t?}IzHrByv!
z<GKbC;}O@3Ed0?;m@iQ}LgQzCb9zE*X;d>YU?z?~o&DZ9GdDX)yL8+7SB6hrCMq8_
z9y8l;M11ABo1!Qrb6QkA;<~=THieI*B$_>`&$DX5`Hq~VDqn2ZAPKIk<z$YE>PiD9
za|5~}W)bDF^UXCOHGYqtkoP{p#3{!kzYr|<0<B^>M6(GXk{tc!|FDH}4wvH)2M;f|
zH}|;P8ZalCBaqEEQL&HlpJ9`5a(O-0yZa^IZ*^UO%Y(^r5WuB}!pDFv|K<}LfC8;E
zxGpL@vN?qsZ5UM=DdZi}&_HnYX$2B%e6)SI=jo_I3^nkL9d#Wl9t(&CuZ;;yIEI!$
zmq4AsG)aNG%hx!REJ2e28E!$`aEce~-+A>WTYIbaj9CSRuK2wJ(wkmYc^q1u^vl6f
zxp6~Hvr&&#n4{^~t~NeDNkE4I)UF(rPpa08^`0Mc)*{39+2r!OQs8_=Pvp^swa!)s
z$|?O8*D?>5Znw48ebz86Th&lqdAQx037}LJuq1g{;Q#0*b9+noSP`Snrubs(DXo;q
zybn!sa|mTd$Okztqaf9@WfTp{XF+Ix{TtY@BdH(K;Wmewvuww}d52Huhr4!2<aa!{
zQdsPNSTXp?v)U)Ky1m!W;ga$NayO>mmLJ0+T;J4wx=$=I)-dzQ;1eYke!638e}TK<
zuU1Mr6g%8^IVLLq`rMXK2T@4eFHUr<$rFknutCIEx&lPj0(L1Q-fv#_(vFj5p6-|A
z47@m&K8rq?J@-n)=R94M+&yLFSfER|x?PX`W@qfWd7P)eduQK&9C!6_U%wRe&IFiK
z+GVy%HBc}i%Fbq(qxV-~<7-6@cSU&>Rr7xLs0+CL>~aVqO_z3_4Apt6Hxd8UJbO4P
zbi>gNU42WC%|ym=R~xQqKYvMH!+VPdcc~~UnXOXCdjmbjHFnw~#HH+rSxXyl?G7@3
z#L36N(UP8^T$MPJCApw|!NfCnHPk9Miaz_cZZXP1xIN^um0^8C*7=wn-6~F1ag6(6
z-pY7p;%c29aq80aTEt18tDB8wpyCB>w>tutv#uWMksSs1gIpxolw`{r$A!JPMzO?g
zKr=<vaD5DLSP2%DB?KaM#aEQ$=@-!5CnhY;wRqW|c1@8OopzEe<L5Uc!K(I9vicCG
z_UQ>W(fqGqBu!bZSbeElIF)?RUjG<UQB6;syP|iwOY)3J;+#;?w?B7EDXA5XP%JyH
zp0;!M)QdfMxQG|0;Zmp|o^JiSRJN3MvLzcmc^q}2k8fS5$&<?yFK<<I?3K@j)GF)8
zLmJZ+_Vd7Nsg9Y*0j~bfW@Rq@w@f50a&2ShYJH_uhqSFdJXd`NZftBCaW-q`Edd|`
zT~P?pVzx|g3yunNj~kd_yAii~v<}j*G2ybAOvwK9ugwmC3CgDJCZ$^7kx^+j#1Gb<
z)rd0LFEbFNm!?^4{!Aeurh>CkYoT)EWHM!@-*YK}f@Y+S7)6&!)*q-c8R|^au5liv
z$hqT$CK?}X>-j;zq-~1Q0|TE<3uZdz%=@`K!l?1zk3LkD#t~v6yt5e`ol`I@5jOZ-
zNaVjOLj~3AxGopxvlYAYppw%?P)@r)3|f`d6(@taQT9S%I%es*c^DC7azLhKnG?r!
zU3!Znm2Vtly+vsSDwQztz9dCpPLIpP7TuF6J2Wn=9Te?;j{{LN)_(7^xmw?;j}xxw
zu3oW1{PDZQG#&KyZSoH^bw&){R(@U$Olev|o%2jKUG@HArGH-etM*b=!A(LP6r^jI
z00=tJ55#w)*fH2lvLbMd7YYE1VkJ{^>^hPtRER!0?X+XNQBT4D(L2fnIMZnbNJ~gM
z=u$x!8OgErN48(9|NLS%voPw7t~9E9j46uEp`N<`h$`;xM?vLW0vd#OJ_(zQoy%=M
z3g;&77&hxiA+|02^i0wQ=MFu8>G}YPTY-#gv;2@$Uo%P`K>ip|Hhbb*1-EE*b-b{C
zWKD2z&6h0W`xn|JwS39CI=S(;JQr@(Oeld8XWmK4bo|$FI@ZjdcdQ5WpXjm``MD3U
zDw=VPacyvwW?rbEg^FxAm>lA#r9IG)n_o9_q*)W56moN)*y`V_+e~_@_`bKba`YPn
z+7o<9bfm*VCNI$bRtHK^Wh&dNTP(F=`ph<NI`=*}o-_>RJW)~<17scU-H?2-74o+J
zVtATjlDNb(L?Qf@SX`-Lang0gmvw^;PD{R^L$z-*z9z+M1F`z3XTghh=)ECV+_q)P
zm#LZzO8hdLK`b}=t=O66HqeJ&>gVm^ZvT5Qm=Os7`J%!ER_`++8G#i_Q9ByZT=p8>
zXJMh<EqlfRgI}m9s8JDClU}o#AEgg0nz3qc-q&+$C4jqNe3R+1L%syWJfCzO$(b)>
z_N(zcVgmbreF`3xO~cOgpAkA^qRa+g><mr0V#X)kY#xl}y^NZ=!G6zw2J+zA8&usv
z9uneDH|kqS>Vf-dn(@A3zuf<xoY1WKA~^6t+r1;s*@M(C$ypCGL3hs=(Q5~C(;*O@
z>UE<I6e=_Q?A8jv+-b?`uL5y1rZev&M9RZLg=Za`Rlt2Yq?kR2?zDbLYp8mRx<dWr
z*Tw^udVu+*OlO<^DW()aQsU(@`}*)HKB4In_sElJ@Z%|3{5x3iZ*!9Cqt)nk?_-qg
ztG&>9p+bcdJgR`;SmH^Qa{ESiH$tlWc99oO_@BwVKM!@fZAxl-oBghdMucBAs+9Y@
z@lDZ6gXT3S)mv=mP#~?^RFXK`ECBn*O=Hw_^FC(Bxf47}x2@^}NwLX_$+Eb(_~4&Y
zm5j;Rl?gzMsyIZWp-cG_h>JvRGaAqpWA=2G)WS>|Ia$-uFxJ?0RutG}UScw2Ns4@R
znTB6*c9X>a=v})XM0?NKE%p%HA&lb@&fqN2y?y<A_MUV2^YE;0MZlT+;+yCC=&auY
z)nP54+u14GJd3osxkh2is=$v;uL7{phpf>>zu371Mnm=k=(4W%o&pg4uwfx-dCpzs
zq#}a;CK(pl#W#s`@&^jAScgM^H0{~4=g)5TOlw>9g0FCt?+g}h-#TSTAd$1PV6oS#
zyDD##<hx|P7)ijzuQ6z8X#bKJGue`6W<!bg3AdC$p+stcm%vMh72D2<AI}7Is^CAa
zRk7zLy^s1ngRQ%;7@Y1s32!^eo+K^)Jo!`Stw3cw1<yp2hWEr)+E5Ij3)BoYe4Vrb
zyRKyBF8=`gXq^|I(0_=KPWpG$BqChn8u&(N^fBZSI`MA3enBM>r-uE=<sc?4^H>6h
zV39OX4YT95_0rm6-Jx8N>%?leccI_WLco`1b@7mv>r-=WGrJ$3vF{__asT}Zp!{aG
zdsroZ%Z4`{4NB{bpM_tLUkClo1gC7V4E{XKoD2A~3v)OZlal^=Jl84br@#AHX$+q8
zM3h!<)GyU&>|m8YUUFAB&ev3p>)U6e=o&qnMdex%p6W+60<p{71tje0CERPv0mJH<
z-_Pg%76>b>pWOfI?7c^VeHd6BI*rQCI8YX&pK(;-Z{=j5o?fnq@$A`S7c+a@)>)ql
z`n3Z&-gh-MHWeGV2CZf@Q>E?tW|)EeOizS|&Ke*<ZV9k>jSRVI%0}jbi#pQ5Wy3Ne
z6tQi*o=uz$0<NQomIYLtBbEy^{?qk9VKFhC&CY{|*Xw+Fp_cDBzS2cK9p0-N(V+-_
zWB)`S9vX|%pupO-z{FH@LEVR%$0iuBCCu=U%RlWJ;yZP+I6Dv)Yr<L{m_cV93H@-d
zGh8{rYEc4Db9%J%Xq$4qi7K8kccRUmbud+CK_>dBriC4XHo@L8qv2~fo5|O7Wxcm1
z*g@M{Jff@C4Skapy*r%V_c>+L$xWG?mRX(-hl;>%qy%fL?`^y_GfofRCj)$4ND27(
z&}Yl;_t?m*SI-wpvQj!W>Z<?;BL^8%daG|&Nx!h&(-HP#rr20mt)b1t1KhttuqQZ-
zB0DnXSZ8|bKhTy#qLKg#{FY_>zxV9;jjqlPhJHr;qj#d+wD|1C5pck5-8Ump36>`z
zn!oIc+$q8Ys|ZJY(QMtWU}|t3pj(*DBu&21B%_2ila7%g$X@IjI8m^e-!9eIZfSxN
ztQw^h&lFli?wPrSzrM9LCnO-~Ae!tPIsaqsV&?tgI&=4M4tV&tS=3kXhnQvw^0(`9
zp3=|zWW^@@iTY`Zx}W-airR1BsbC{wGcy<-eA9+a5tIt1oU1*E0%iqGsoyY@VyGjV
z#yx57^f#4-SUcVdJrj;dnzgmmEp51iE3iZ6?$<{S$#l5&AhwIy&>&sV`n+A#QXbj`
zUJ|i$#W@BJ1~9pRFOU7Q8Lhbk<bwT6*}fxJ4FxG8m;49j17(KKMr;IXAE7WZei;mZ
zx?t7erpTjYxe1Qy*nRTmSr&e-MbJ5LId2=R+|ES&eL5>S-bUK?40s9d`jV<qjuT|4
z?Lo#wRP<z*HJgLW^QmYI{;5RCjk8m`L%xz<eW?dYe$bBh@POow>GDO;=aFaTe<t^`
zjLjM7Y`?Y=Br!A_)A-imRPrd9nnGSzw?q18tF^gYSr<TgDL9jZimP<9Etq2$TWCt@
z<yUzXTh4(uhjBP60)BNh<;YP{y(>s+Voj^VowT<6N&apSTym;Cmy))+7lTC^LDJtV
zh(2Kgz#z5rLli6}$1Z1!PA$@w(O5rdnFw<;5K>7tA?XAkk`#rB9@hV@8px<~7r%Jy
zyuV%zy1(~xJs-%E*{}5?b{rFWXuTKfHP<@Ui}w7bx53ZY4(E^-&|m+a8*Gqh0+YXG
z3KoXo+1zw67=LF=(%9jN7Z{g1_`3l2dAza|&`|LHYW)S8*V!^}<x&MMQ+T>+>-Wc6
zx3}tfy1Wm7_c}T+&TD2O{8g_%AgElxE{OY#aYxaEY}sA!NJ(i<hmQ4j?@E5k#MbRi
zsvV%wd*IT)yl@KM;C{tKtd|Ui-1r3YjD2P&QboWLl&=$-b1ceEqLkvlU|3kPiMtDR
zoejBa@`j4P|Kb|Cn%Km0rHQ9@HLv44K(za_jGuozS)KW4rIoPRWi_TNn_aYiVG=BV
zFmXSKg!oXf(YMVo_B~V_oU<)_TmN!I`SnOI^g<tZ2zoT|>{zPdhfd@>MO^yBPfNG{
zlz%|@NMKPZR+#P<KT}jEuQoowuSws)oR_r+JG;B~N#07F*c?zxSn29Zpf<Wbr$6-g
zx%G6KzBh)hU$7^+>Q8`Nk~6k|!OC3~o>gqG5|iDHLheOo{<};p9|<D7VVb}8ftg%h
zCh?!N?EW6<@pKCyKUJ+-9!b1kUb!|SoTyNJ_g^Tc)cXaBS#93RI`#)O=0arHx_K37
zkn%aL(0gC^V#&>s6$EN%s&M_&oE3VB!r<DH3-inZIS#$5ny<E8<-inC1c-jG>;qY2
zCCO{uULsrVugFh(NK;%#s+CZKNrfs;`+KH2Nb{5TAqT?od9sTbq#d{hzz6^QN2OF5
zCX!qd?bRwE#YzXE>B|a$_<cud-sr6dUT@M@8(Z3oxERB;9bbaePfELCvf*xmNnC5G
z?^}EEKRDo;kaV*}dv<zVVOy#X@bDy1TD~t(8%ULr$zP~h3g??RiIkDYbY4;&`{QO$
zj%kJ|3ABV1=kcR4V`^in9fT#~oSYqaWhiy-oP4`l&?1YG%;+3n4k(WCG1zofFi@Q5
zrPslfB^eMIpa^d^Q-t086-MKnXq$HhQcd!2kM(9Y<?@g)lZ1AR{$V7l|G{OXBx6bx
zL(Q+86>v;UwOwm2zZZ39URhTsj*N_4Lq(#|VL*l*0i3}&err2-F1nj>BdBuuqdfqj
z=YO*i@tZ89QUJ#DuYN*2Iw#<Pr~4$3M3i)4t#tPnJ9Afc{y|tx$R(m;tujvRc<u6d
z)S;p)$kb$PYctcU<E6`$Ru7AI+~ZOSKBQY#R~+k?-OrfFklk^3#`(9xxGvtzuM9<r
zg`#fR_=kU#_udo)n&Hxnu0+WG@!_V5edNxF;WWED%lDO2V+9;IKjPCV?PdtdvVc4|
z5XtV+4!g=$c=~TK^G;v5J5;=9jC|`=@SW}t>Nx~3_i;cpi;J#=2f5#XqOJ7b(?mkC
z59qZiYAw_Dkc>F&7?SObP)xwlQ1$~fT6oTi-)sPs6I8bQ$FycwE>gsgS<ti#6V(JX
zy_XOEaENS@8owhnZK|4izsOHCI(R17W*#`<-XQM)g7*&Tm#n68vUls-)PF6H!4T@G
z;w+9AELmkxtfSo5R<!7OFD>73J}xqS-&}=Xpj`g-hfI@Cs-e%w<l3!aJtzxaYZ+tb
zuV`9~mKC9aV4vW))Sp@N1+EjhByRiyD#2#c#-tz;FsuRA?GN}cGTgct^`WWRbf)Bd
z_~P6R$mC}>l%Ac1W2bfscM%0@Q<p@ue4`!E;xc(m>b!5NQ>ER!FM8T#S%A5l$2}iK
zdEvhOTD<0sz5bEZmn=!+qN)&m+nFSF7V_Xp_)QznGIIk~As$di#i(H{&;LZ0o-Mr7
zndT=0yR1R;NSpEsoV+a_Gz=otTeC7dEOb1)<^L`yT5wF80``-?Ch|q5#GbM-RwMTe
zI|=pJ?qa4qj0H!SKph&s_WlNVP?wL#?d7%A0*PfAz)r@pZhz$d@>{>LEU@uG7kKBB
zz0)uwuAS|^_#rf?WXzzVW$ggQPc|_1h<}X}um{4UU(BjtGkZw;1Ci6xg*|o4#>KC&
zL|1+44m@hJB7~hEGzC&rmehRadz|^BX)pAzqd{FsC@(>M?RBN~JaxdHXH8j294PBn
z#|Y(_4VC(Qti0w;2`t_tgL=PqRjFAxRVg4#iFYW;`unz4j_n#NC@lH54r^X*r}Yvk
z{?|;Q4_<j)*Z7rczo)0q(qZN|P|9PQuLSr7FX?pl1GaFwG##uoQ?Abniz^-NYSswk
zJ=X*aOW6AX|I0$;$!~6^bO}1ig1i9COZD$U-tMGazC-JdD7G&kMV=pK`{u~D&&Qsc
zshP$V|5}lReO4nPM9I;NRm1nRs|`~M#&Ilq5tR@v^%{M^JCYH`jLAV@JTFyxIxS6(
zk=MKwnrqh=j>&+YI#56S8%m#yzNU?<i0vNPL86uzrV$cAIvEim@S()4+CVLZIRxuC
ztQOP6_Kr`VC!~J7!nZV=?Xyj2LUYkIbN$o`r*K`=X-g-*h@aooi&rddxYSSxwlvI<
zilvJ4Gek<dQN^B^)U`x!WyowdMS$l=CW#x#@I}<{l+v!;YQ<ebckiy`&nkW;%lv(R
z8{F?Kvl4gZWfyH4mz4ceo$bJqh=A>{9geI4^Ja^7+LxDkPkL0`LaSb(Z(Su=CKzV+
zFd~1djtqY=#I&%!jK<xOg5h?vO#29n2EBr~<~#DD$cMa~SO>Xn%Yt5{f`~WSFET{y
zzdRe}v`s=+3@0(0ax<JsL>HrfypfJ$)Ftj2p0-OQ+R;pPPJHpq%m^PmD$ORmPdv~3
z@out)){|GdIIBnK<5`}gyG5J2c7oyueG@Mo;mxR@FTaq@?(02Qjoi&4fTsGFdUVmb
z)g34BH21^vOf{iB#Xa7(aN@O9FF4?zHo$1WZq!#_d;J<>@;aM5WMcs^HM12&uf(>`
z#5veTWT2}R=6i-EoOLGEt%CDvKY5xkVLq?0s6N*?H`<kQvj~`fei!dF<#G<FNA({e
z+ZOQp5phMCs9MJrg*N?;?$=%Uwjx8GxmFeSegPmEl=T7qeLjwc_=HrC03K5UU!?f+
zziKhyl+^3nTnyc<^q6uhXJTw!tpnvbm45{}r)h)w@Oz``0<5!pxuVwrX(CV5KNb8=
zywR_CsLRJvZ<+SD(G4h8LOl#GX_pVQh!10*JsuB|LKfo|pDt=T3p_A!x<ekt9-W=q
zRWoFfm89?Qk(;ZB*%KlvX6{Q@zh0Ee1yz8=g2ZipFjx#B%StXRE^^oQkDM12SowX#
zLQjiSh#&Z{MgB+}?3vGXb-U}$7Y=%QtRu$rgkm1XIpHme5nyi&xtLa;r!=8*{Q7)C
zG8=aZI<5#}0OagVHc<0p9$HEN0^?mJGJK%5inv(OHktX_l0dPY(x;KS&&A&QlP5h{
z!N`d^PCwLWLe^-xO?*IKO=l*3W+9A0n(C}jtQ6^rs37#Rm8OU0dlb<1{IAqXm#56t
zY?q9@bVDtbF=0_i+q53+jgYQSz0!TN=tTxtHg5`s?tbXi8xW$vycW1X3;E|-<31rJ
zDf7d%CgWcLSa;AtjQfcF;Fs{11{q-5#%3#VyD8ew)O(_1{^7lf0DJENBDaha?bXwR
z<+Zig&Q&D%<KCWZ;<Wq$ph)5AACqtR=^vbs2R9RD?n9&C7rv`mqbl+`kN&UPL9p)k
z3dr<r&0ASrJ^zp5@3UeR$h_^B{90XCqLdKN)O}0)gg1410$paXu_p)bFqP&dW&!32
z$AF2dt3|>6E#13l_q26DZ8qFyi8=4nb<3gZFV9O`Kg!QpCIB}-ylUm3)mr9_13oy|
z3>m)DdsXNTcmq$lEN_An!iU<r#IUN{j|`h!=*xXgs~j6URR}gKa{n0S-k<*&AA>x<
z2qOlvM7UANv`#-N>RAX_^Dc-euLrOf{IZ<g$CsO4Q*ImL+Ph!P#vH%6_zLMUIIrwI
z*WTsJh_i_t%@Xh_n(|@`k-4?{Q1`w~ayO+c763$Zvg|vd(KO^Eb%$U!pea}&k>}~W
zHEOb^8;ybE$n?&0MTju?k29gvR<EwcK6Z*tJoSDKh8hn~lE$iA@{+tbGYFxPMu^U@
z6?Iwlakb<9JW8F|Cm3jf<E5*Z3cOufT8sjx$Sf!Vzf0ntx_xoy_fOdkZ{m@rtowWo
z`5~W~;;7hDgia5v^UyJiQK<H8J;3Owv;$3VF%Br_JEPKf?-DDJfJi7p#rhU<k4RFE
z0CT?)A!Tu)f*sF0t2bvO088N)G2u>8N5<pcM}}FGozujgf!E+31C5-G#W_9l>FyZh
zI|3grmezs7w52uqF2wjw%IWVB9?~sX@m9wh=wg+>jR!U}7m3{umAQ!jOlgc~j3^TW
zB}j-LCyAN6k1vx7k;LP^`j-0h#Y9Ij7U284m^69+$D`w?qaZh3&2HzJAh`1blO5nX
zYpX!2t%RC2exUKJ>ho%CO~~NRr4FO0Ts|oEVi?)~lJ;2C-8o)TPEaf|x=yctZ}Em>
z!tMN8f?3!VEr2#?ZtC!))B+D*IH@<mex1sUo=Hk7S+PbNxUE3Kx62t8GVx=~EP<lc
zWfQ$qvkga>xD6lBKP7a^Vs4pmzPZB>xS(x^4*48RWF+X*DSw(E@F3V0h%%R!x=f5=
za9xn~_~DCZ0UJ%VmNIRl^3^|3gft83WOd;i#Z7w2l79KPuuz#9d3!cO-;CvX0nr9J
z;VN9ctDi1%yR=`toKw(wee6Ewa9+Rgtmk7t^0eXW!CC%;uaZB=JKQI(K>X)$=0wE=
zq$>|k=I+KQYobo;T6-uXEgm6S=QwD;hjC@9ezq&Jxa}j^K9jbWT3OZDIj0YV7BXj3
z$g_xCD4m_Y<=r-Ty$EYx>|3rN^=lWIwh>_hMohEQD$wZf6ZhC?S<OH7^hWbSR%+MC
z9<x1RuJSjd?bmAln01(RNvCR2P$$_416W*cj;^D+8vWbMd%_HpK?Wdzp3d5jana9b
zB9FWst1{YU;OZ(B@kRAzu|$$a!_-L=2Fvg>=MMXsbXeelYlmwE4J^F8Gfqsqv+Q*`
zbXpF~#l~$^TCmkY_sicG?1m=~o>J7Wj6lYBtkXCa&5t{;ombjCdoOLV>2_88%hncX
z7s8>cxMsSU<-XPQp2ZW3cBp70anp5cOTT}sHAM+0`vq`kfB%zRaWtC}IFN0*QW%^5
z7H<9NZ7jnSDRto(7rltA%}35Q+>}Ip{afK9)9p{w%0dWXL($Y)O?bfaVfz!bKxKX^
ziSaOl?{jC4_T5_{emshidjO)X>$&AnSm(AT=UR-j(rj?obEh)L@5x~Gy0yJ$&i5q1
zQ-!*)Q{^`mTzOvf&trrJ<)-_ghi>;lhW(Ay^GD8Qx@NYVgL_{;(_pKL)rv5QH#obe
zwws<V5pdNmhleT`03JN$rM0=jJDfVcad#`R9G-7#G#(-*2kaRB!%cyFGXJvXyL$*h
zh#U{juf>66@)Io0GyYAh#@-2@{|mINuZdBmth8m}mVr*<ZHK*nJ)}ztsZ&Xg$1I+E
zGy8*HC5PKhe6-ploeUEdJUZSx6T{7l_!mRk458k*o#iW=J-p6es~l-bNNC~yQ23Yq
zxGajZ=lZ=c>KR%*A30`X{UW%3e9ifv)3L_6!FUvh#5Jb0=+s{laN^h{UD#9am%1oM
z#uqAn=;4@rcKNbaoi?ZlO2K_)t<h}r*|tTI`_~_b#TEwC3pd4_MGp!q(%De`ZJVFV
zhBh=9;oqb9=Q2gzi+KWE?e}Of>U?ZVVobzM)rWbe2HJoJ%mERjeGAu2La=xZDtb~k
zNQ;z~#b_H~?zw|E_r)wPK7u5i=h7Ob)XZTBs`K^s&5uB>DUvyoj*go=YOYkS4|Cj0
z=~$eLWUy2QlleD=c|O%$R9P{b+mT+Z{}FCCBa4cl5nhg{2a-zMZb`&RJPoP85cM94
zIw{MEg^_FCpI(Ma$n2k4T|OfJp>3^p(w`9p2m-=)d{gh9W*J3QrHg!)9!3RBkq0}(
zz?^2pY=b*uE4U9`GbWHg8Iw|h%JUOxsw_;>sqz;1;B@^$`RC6a$<>jsn%k2ZN<3YM
z;oISE-mmW_uN}?^+fd0Th=C^sgD)dhQa)&bN7ko+IxoUtSXYlGM=UwxejB-zs$^1X
zh&+JFY|Z%9NMoyyb8Cj8bq!@o^byROS-K4fWoT=HBNWCgE^V(oUT_AOqta>53cPHM
zsdJ&>cw)|VX(#w5@yA@NLa@j7Aad||(vJ+Y`DifZ1#u8zxfdNyv!<uRq0zKXUx<ws
zDW^7iovHpLZQKm1E}0WDepJPI20(VJ^$<If9F5QkJ4rYjt?5AD!!boU=QHgkJ;oVt
zRK9owftwC0B?wC>QqKtwUzLxkRtbm6WszY!q(g}6Ju|ZOuC3_CY1km$V+Y{=H<f=|
zs`!wSz0$&tuR7KuZJoqqN(!mx*wg43{zNh-qV^psO9pG2SK%mw>rl<c=lI$FGgkG;
ztl8%J*jBH+I@%K!d_x$1<`-@##x!Xravh7zQ)3D$f0&2IQ0g`Qosd-v)9vn>L1xS9
z{+;mKO6!847A$@NL{lqDWkAuo>oRZ{)h)4YAw>B~R&G<euGMo6r{U`%U6CvKKFJ25
zx6S&TXU^pA{ytxP8Mw6rwJc)ro7?8vX`lSEx$?*^e+dDco%#K=JZ>`OqU|abE%tKi
zl)Ww|zDxS&s0dr%4^u?Ai=pCuSB3)#2sEXtQMlyNyJYlot8ufz@-4(YqQy5*jM8T^
zRW}>4TYaSelE*EPJ`vsLyVz<52H?!z8h#FFuN<jvGjFaw@W|*`Y8Cv<Q9bcICOzj>
zgU@+??&i08!o6eU5&>H5Vw<n&{JuN)AHoJ=x?{g(iE~cx#PWZ-3G9EC@UP^@jH$bb
zz1M;ygpG~)ze9X1)B&u^_Q_czH|fJCP%cms(s99$smR(d7V^T34S7;zFuLmAkaw2I
z6D$)b+oA(um<+aVoFDCIx-KHPqVopc8h`su6Jky4Pc5(I4HCJ+!Le3p1gC2-ZD4bK
zU>)3eW5Z%Gt75aqNxP_n6C+n+%vj*!2NLkQzOtU&`^tC=Nsu&Uo=F%IQlw2X5eD%q
zKy{cFAbpxHq6Onzb$ar+b@zn?Ha&;4%UKJ_Q=cI6+gIOlVC4_6AXd^DMICWi$%8Eq
zsqd(*2YJcYQ9crPs70i^Zz|9Cgl|x(P=JTXM`a@vGM*?gdc&Mb+FIIPvfeqz7*+`&
z=60pyi~RLdBRnRfx?CXqBY^RvUIx~u1^gZ{X&oV*Z`)t~(YY*z{0xOnV7$7r2gE31
zy4^t|YCB)QB!J@FTCljU03a*QBp>(z53!Asaq()K+u|7JAhhHc9)}Amg!3dg4js3{
zi92CDq$;;ZnXyZ+JziEkg?jDPgongPqi$)JguSu)J&TFhb7_P6@NPx$I)j_%c*Va8
zt?>%{_~mROxT)Xj6-h_rokR>lXntY8o!RxFTq_jk*8WjZ_%%aFFHcR}=~*h&-gug8
zs+8iu0AA3}TSzV`dlD>)NZfhV1u|>-BbrLhl$Gw(ixD{g)-oU0H?EhlSh6cz0dpsF
zVRE+|O~S7{ZJ5d0gG3rDAVRoUtst>6p~Q^kewrtM8PoHiYtKAYtBs`%OBnAx^2zLP
z@<`#IUhttImZLSXx)S(f0eyER^Ta_Bho?#~L%gnaRHtz=s@v|a!%WVzDaDp!qM!K4
zSHj7Hp(J6OXY`7h%CO$nQL8zannI$Ate*JRvta6A?dx)W4?oud0mUW&xzCR0ZSeQl
zE*{$a7^RRkgBN@0<6(Es%#d}AyUMBHkHdT(tjk`(cSAd6&M;sCaBnOIRgtR3!ZDFA
zidrF0<9k)0{$pCwFSK)1+?lqrGZ~ZFAjNExdPjS_h>!lbg2^xKzKzFoZ44_mLT}I-
zOP10LyKX9i)oiTkRUyGvZ0)f3KwxSwLkB}_OEDbbfGK5OHGn&)YcWvJ%SzWiAZioO
zMBnV{DIrOwmu-(LbAgY)P7~m?wck<zE)`{>!ZV^c1C<BN^Yr$XJ0K0P23XXnUQR70
z@}+~WlG}rx12IvU$B~TlSeIiQPYnvddq_)}m+el}c6+*B^k*Nt?2LsI#kYLE9HK~#
zhAPEl6oK4xX2d-Rda4F^iQTDmtg_YDCEA$LGxt0fmarOGH3ez@{Z$)`+|kO`7fpor
z=ka?&5wEjtA%0kVM4k>Q-f=>|K#jWV+SLs;fTIw;N;91DV+8r1BzJ_dr_*EoQnifk
zl`bh&%9Ex602JVt0{MZdA1_h;l%amvdO)7j-S+k19M5NU=Za|)(eQiDF~_$38)DjI
z6-aEtynJ^B?ysM96<eJFp10*C-rG0Te5-Lk92uAs4$3-|`_txDy$mF(W^Z3AKYyCN
z`cq^4vi6|Ybb)UP1U`FjMMylkPpz-_x+PBMSn#2dM`DSjy!Uo<bL#EoAuJlTqA@44
zTDb5t=Jxe45y%694}|?<PFvIaJ$Ntjr|?O!am9CW8B0bdjr*KLV|P6A$PESxF#$2y
zEI1Q($qD_gqMb&offrX-TYvsad002h+wk5pqMjX)^;}`a#Z&TsE*DkcUi-V0GxUY7
z0lRQyJoXWC9fGr&<-Kmp`Ud|~Ns0T2x5o!0XGmfnzI>iI*LSWDkX&<?rA+h?Z8tr2
zQgSFsOgT=tfBkao)x6q;SF@EL2+-8HPPa`Po^uddi*Df&o*<GV+(+ZM5~C23UQxv_
zkPc2C5DA*|B_xS$;CP(US~VWK>gTeyPGumUkUD5%4S{HfGw{ToN}pG%VJgKaT>iG?
zjmld>-0UgZ;OOrDV2!OEbs9CQcUA{vmqPYWhn)1t!!bx>xiRNnB2Gq7MO*b~xplF@
zGQdUoLgPl-fC}-;&1>O-%fbS{M{&@Wod4PR%bX)|FIwD-OiWDhf*P_;YB%Nn|Mkar
zn>X{Nla<I+*`7s;M39EzYC}nJiAJGkAKL5Zmb7q<$)N;7RydNOACca_saypsj9{Sd
zs;I4P3gPUGh=|$x_|B(vxg6ZzI-qb47S}7qFm0KiTH1<lA`f_u7Kg=rjT917kAFSb
z&FpPJh5pt@I)fFbo<u*7U;>R2=4++wN3n6(SQ2a4HaadQVl(BKQgSg{)tX@)&)`m(
z5imTe+QUQ_53ic`hS7gRfeK+XVqMmk%%Na#l<1SYBFH&<B3Kk|Xm<7uhs*Krx29MI
zk}kN<|EU*UkIL+z{2r*3X>RQToS@9`L0>*gCF{ow8%UCG0i|lksE)z}Lm0s<iqcT{
zF4657s*q3dRlD~y(EoIYn2F(V%%+C}G>Rvm=$2tR1fh||qBSMMwQ$h4UE_9dUu8&+
zL5W)mK9v;Xu9?GB2@qB(8lRdVZ3=}vT3x??A#F_sB}R*Hi^QMkZV|ma{Qp%aL@(dP
zpR%9tr|bj@7M%y>tB67ce*oOa*cFN5-WQ|fml-j?<L=#XN)<+lni<s~Z$()sOA9JR
zBaYbq@{3!u5P#S`NrWKr@W;(rD1m_w_kKfX5Q8?JC7apciC527q`?EWO{rB}EDovd
zEgDvSNI9l8VVy{5YTU-xxzY{N`AEI?o&J9Ev9_(pt$<Wo$&*kzXYqXj)lcs~)h-_i
z%F^VC&^Xg1^XqcbjM933%3EbvgWB*j#?a=8Ok-de(k^~>{B)bjH6Q}!{k(^})N2u#
z96FrwCFJc4%8r~NZw5R;Lrk$)hbgD2`P$MuNW!724gM!MkiO3nY6jM>!#@g@`>A`a
zr*P}k=fNAb)&64VEvJ~^K8FP0Kw}29RdG*mse1ltd0Tj_)=9K9x^19b2Yp@)pQia?
zJx}P&Uh}@>7+Wfe;%H{2CACdUA<4u+o2X~p6G=y(Ea6=uNb3KnubRLNA_7qk2l6Q|
z7EvO*wU4iNJWFVTJ;=_Yt$uR%#YKGnYAqaJe=)FBxk!!{&(@Jo<EfIK@ynp7A4Aax
zYRjMV79sId3OhqoE4xxP<t-sXW6lJ-c8%7&_&np>hLROP!I_Z|>)7d)c1D|=2WJoo
z>+5k;9Xz?4p2?keN`K{myWAy3N%t3H{&iAE=)+7S#I067owZ1H-fqEzrOUw-HG(qq
z=hqZXW686hwn(8^+2&BcP@9GblETb)_Ki#XhQ-P>X)Muv9Vd|QV>1;Th@UaYK%2zZ
zo3&?X?uM`eOck=KC0m)pX=m^JnZpd26?F~U{5hR!)grf26tNj+6INL`ZXy}!K<gSA
z(Ci|mP4SjVNp6vWk$~<I<-8>iVStZ$tLnf;`4dilD4XFxee4&Dh4w<5JZiFPn52QF
zVi@hVk*Mu*B^H)Em8op!tvS6cfc|&o;+i2NtD}n+8vvp|27@#R*im}24TNzv-bPJ1
z@m|LEr_h$8MMZ!#@T8#HCcT#*98`TpbN-HGOmDId2x~%_d6HB=u&FGF!&oDZMMlE<
zJ@ec~#sp3W<VR^^{z6FS^81$zSAjoU_-<t)$Iz@44TqCdi4o{fDr@4<<f}5XikVvP
z??V)kRCBKfkX^WYm^UAh7q>hI(FQ@mpU$ck`4rjwdb04ExQLsX{%v@Oo?4DGy+(wR
zyvUC0uI|JlPz#mp6(YN}BTPr3VNhwPLcuAE8p<HH588=Vg=jC@b6Pf#RQkQvJ9Xg>
z_{l&fz|i;~uLE22%%N;OGJ2{7Hf<hvU;ddEB~x#LP1tPyE}Nz1^CCOeo1`}#@@c|f
z0=NAd^~eJ(kNJiyaD;|nEsAbc_W99GO@rU#MEr{-ML1H{-ey}DuV#)l4auEJ$={{j
zX?ny-f&cfjetAd!Kc+j`c+d}_CB0kkg5Lr-VHhyfj<g%=F(w$UsGo0Eb!SO#F~w6&
zUKMbL*-=}{f#H-dOP%s;0^tI7S|I->5lO3v)!|Q5Mq}}>tmxr1TQ~x$DPCyG`FCHp
zL6avBMYGSyER?nqZcxvrL?Vbc-yUr8_48MfC0YI=W`-@NY=k%1<yf|2)k{%Kk#U=@
z3VxyW1T*J$%w99c)lELui1go-D(3y{=uDrI2A1ihX}^0xZoXBmQbth!+fq^ppk4C7
zLO3-VsqSLbHl?eo6mD3c(rnow4VP$vhraEm_VV!?Pwf|`rI9VlX7v)3AGHSRs}3TV
zGEbI*)B7-@BR2N;MT=*YRG6&r%}a27lQc&?G1Yk^OTwqWkLt8bH@qj)G;I~e(JT_z
zC-ZBOf|aQZV8t6BOx%2tm7yAtyGpR0_NMgoQ`d2>CPtAet<CpOGFM%1Dm2mu8$({m
zP5-+v{=02v)vR!lZ(eUiecyp3n=n*eUoR*o!AJNSHNnwtBcqWNf-*-AC5Ey$(mZ6#
z?C+0@%r!zZHeU0YjZRyAT0i}6G)~hBTXz;?`t%R0;Ya2J9R?bkHe+%LRt?R3R&d|2
zd#I3#;j48O3XUY`rqO({n_{1Qb%Y%0FwIgq{n>=4crUkfDo1P2nMly1cHGm7M!-K(
z6rrtSLcU)6X_-43#J821A^CdB)uU8+t8@=m3f%FUVhJaIl$KDJf6Wfo&n1l}&GcUW
z#3W-O6DyvlsOa$ITlo?h4Pb1>PAc3p02{!zZN2_BV_YhHD&CnZrT@+@X^KXEE3b1j
z^?>@|6ZtZ^vL_<}0|WV*Iq$1Ja;uQW{!%`wpd-hxUZKB~H_ul33VL}zp`6)#m=cHm
znM{QGU$pZNpjc3rC3P;JJc*2*qhJ0X?OICetLA@LWhHO76)TWV=XjNn{>c?dboiaT
zQa(d<VVK|ho+(R<C5y(JAXcPenbq&XAH4ywI}869-n;Q@*(uS{Xv0UgR5FsNdPM!)
zoN4w%i_T1usb|x*QrAULmbMU+hJp50rsoHTZuL;Zs5RFFUo%ZFG|nHbKca*f`046X
zfl5}UDu|0TK4dcbua970xc9vLYy-~X-^R`c94VyNpb55Wy#RI>!Go?@&*(*I(74JI
z^eupSnpWWQS38)y2l4JU@d3j$f;bs2@`8Co-5a*-c)3rJzxI8%gUUd|Dxc?^n)8=+
zyBm!Xy=M9dfO=!{BYlK9Q1jU#?gL-Z4*6aSOw?Y3(j$t9VQoCZOi%CL7GEXNIRV<)
zx|&h|_>F;fY*QwEQWq$?{P2e!SebgC;?Y`NoRYzb{*MS)BOw<u(+V&Rx|7ms&BA#1
zf0SgG>3^&O=*AN%r%h$jdwh9Sr*Wk}c6Oi0)1fmY4^`Eh)Y!YW@d|C8M3z*pH%7KL
z?PIz*+PV~Hbm2twl5FkB-ktLhDncYhvkX8Lw()d5Seu^Ko3hOS*%TRs5?Ult50dm*
z!}6-#>?v-r-6zZ9v$s6xjiI83<>E_lpGb~ji_WUUEYsa7#Fb@58kGTTw9V)W&A7^&
z&v?{Sz1`)t=|pu@jc~Rz79|Fp@GPQDTRdp@u0dhuuY3cUYU5V1QroGW^>p;zEc&wI
z@<DqqWx1)~t5!6tZ7;1`TxX~e6kHv>M0obCDGxi8uFwgl2osC^?OamBe~_+`&#~NK
zRb>IF-x&I3jc>DrP}!;IVQr7rBQ2kRaF0J}XEabEUW|{MG_qq}qe;P9(BHoN5*iAZ
z&FpEXPVx(4OxoFH=%p8IATpxAd#qqU3pRJ*--v&r%X(WlHGI+{|Mw^&mH$6~kZ7s$
zJn#f=^nCS5w5N>#WD89nk=ic(k|=%6nZkF2-#HY^i<NMt_hS`glIZB?bHA${UdpnF
zDl);KMq(qi3JLr-So82tH4Ir1UuLLf-^Uood)7BKcX|=hJ~5Okan0vhf~$Qg^JVvG
zw{);?rL!{xzfQ~u&Hvc^?%ZsZL2EWX;SOfsBSfPWojS3V!B(V*<URLTn(+FPobd0b
z`HvZ;ti%{L<y}Xtp|Lhy2(YZutGV6Ck9~(8oXQjTB3@`z9bnVW!S&))CqIKyy@{-|
z7eE78=dQlqCb6e^T}EyU%(1|r#M2~AIQAEhJuiww{JLG_PQ~ch=H|_l5h%xkaww_x
z;7{x8LWWzX>*p7H2KUx!W%HJ*jcP{MK%TL$4_{c2c*%iZcumI{4_mjxd7avxp<GeA
z9q)uP2lxD>nYD=4d|~<;|3^XpOKi}7qn?NDXa;shp;EW)cmv4AO45Mr-?Tp&&{xdZ
zFB`uDQ~{gGv-CuUC?lEB2>fnF53!KltpU_Fh{oPziBN3BH&zipN{(G+KXjo{YU>ER
zWqOtS<S8e94+Z!?8~G<(&=p#R03*+unYCPGXUAb0uKwK?MIVW!)F;n&aKsxYMLo3U
zN=M(HwNC`HIB)SUKXR4SLj{%OX#v|ytxBTv8|7er2EA1V4c{0rWUT6GzNx8eIA4`E
z^W?LYYOTEr^H#bFP3K={`iM7uNse+08ytUQB@_jAvK%YkJJdQK3tlDkTK`xgQZFl)
z*poR>7MTve`0XJY)$id@2QIJU2;yY=d)22LjW|&YQ3<^%tTBkkP}(zQNFU$jqKfBe
zzMrfs?VK>k<k;DQAmh6Zqgtl8WBVu?$Wkf{Ff`dAob-q3d6~9p>}h*F+W!9&<o~>q
zZ1KN0da-S>bCTs<21g85fSvpHzr}pfsTI(0qn9cbj@X%D2CCyS_C23_Hhzzy7A+GV
z#?UG+?v}55Kc+XbKEyJVVA4FaW^)%(IPAaskyZ`qR+dGk+G>?ek+lPM;vZ~B*GIig
z9QtDP%Mew^HPp`qBRYG9YYvMySvC&vqx#*JPM-gdkoK)mSejt~((#8ObRKsv=V^4D
znsaTm8Xi9lj?4WvnvkW})3zwCjJKItYXANnrQa-RWTlO?cCMoghLB*#9qJO{ekxgF
zAh|V=#vpJAg7HW8I)#k+n);Nh&cy|mvznS7PHPI`rjk3TFEv!%y?^>yp*~W0S04_S
z^|S&v78De;-0!4u2@1K2^sr8GGL<AV7aA%DDA~y!VB5>%Og%wrr}Uz4c@Z$Kjy@fi
z9+GK^BP7#aP@-L%QO*VNR+zhXgoKU{8>H+i|7aWOHc~P*jQ?7E@c(SxZ2z4<(7zb{
zk}CO1Ol)y$#NmfChfq=zx}aJ6v<s!GCy5*2qKH@V(CNdmM%g$$d9YmlfR?UT{~l|8
z+re8UQSzr`_)jl`f~f{#QrUOL!c?{L3--Q29EPa>kFK{2YP0R4M%zM>;!<3K6e$ix
zio07WuEpIgSSjuh+@ZybJH?AT6b+C9h2SATD6S`czBw~z=6g^6By<1Eb?<A-+H0@P
zCKiqnCsa87;)Ed}l`i%1k1U4BC>;^&DnK3Z@l+t}zQ_lpJsQKV)-~q<jLPriOr2$k
z3=0t-=|Ss{n^5hNbVDfhOy_!Nb8_nMWw*)D98%DN3C?O8%JgKsqsWTbM*;n>wia8m
zh5{o6=f@{hymh3;_l)5+jyY;Xs^h=Ri3GJJ4lxOoOl#DcJGy(uSo^h>T6DtST%a_g
zH%h`<(wI@fi0(W0r^hw2idxSxkkz7=H{GOVD<6$_JB@*V>MUBbzyp?21Ud<_GmuNZ
z)JV@o$-l%?1^dw>W0*|#vvO9Rij-<j>#Ip+3-mhaH2yqk2BtObSJReGnmb`&ZSX%F
z%4u2ge}4OmQi=_SA>8JjA<(u#S9@w3BM{F~f`A@hgF6d9>d$=4%XQ`3uwJqxd7gnd
zrTTIWG4MwJ--`8$+(|orT}M+t=XxEM?jc^1@M{q<xmTWV*7e$PJ%Js~<*(PZK@4yo
z!44y%*^{fHHwJVXpa||Te9vlwh|Kv&t8)tTl!)?Yu+WAj-A>3-W<r+gh&wOYpqo(i
zjnkGRAOsMW@mJ@^P(roI>)0e=P3F4dQi<rC_U*zcl2s2Jw&<$eN@_vM+=lNyfNG1j
zdBjB=IUV}QIup^V6BSlJ?x8=1WOT8b^TGrjC0#g+$%9mTuG)oKlY9GjYBl2nQB)$2
zHz4O-P&mWIl$~fi!m9fAO~`&rQ$=y?7fHv88MQfrTKD#Ps%<68IXsms{-CNey2kUD
zFY0x}28(D<BUN16R>z5cO)+h++8iXyE6NX}(3paZs4vQ4StwIhg_x501M^AeI{V}Q
zcfO5$A|3deSoAf;l<vDbF(O&u(@z6-fWq`l(T6%eI7=PFVzol2+<sq<(pzZ+K><D<
zQ&$fXlkyLBc9E@`kJqho)<T^tt_m8(?q8Sk!H@m5ZgqJ(-qO{T@p>fFmM2@)LO8!H
zGxJfZ-m9Z~d1EE68=k-4EF$jWAN!8b?sEr*=-?-lQA!P7jAH+)*Kkv2LwGADd9-jp
zQ{&}pRCSr8*wuC`Lz^JgSMNP{rpoRGF;}9?Dd_GnynpBqdhsO09AcZ3<mS4J=(bTl
zz~OeSMzw<8We;B2kQn~KVdw|d$gnJ>MAmgr%bjnxjcBB(G~1dx4=<lbps&cISg={J
zHb~#qxG)P>5}BG2s_zL{e(zw<sIDsOAxnbpZ~v=LP4%#)<66y~>#DE5-*pZdI;W&N
zkqYajZ#XG_PM0b8k<Wiv8FhHl@<VlQ^F!ImZFtupLp+bUMz^Ok{-;~0=m=m6AOg=W
z?+5%o{TU`1Zf7<5M+$wVB2uMD;~=@}fBAmE|4GuHp^i^aC(7(gDffL^j#Jp1zDTbR
z!8fb1fxsj-v%@lp2Z<?5(!m3B-lz`@sC0~&j`oX=zz2|0R-ggruDfzp?>uA0;dae=
z^@6vLKTXpLuRK&Ko_RhiS2H*F)4*TM7tv=3_0ZET!zguaI?l&5&8TvoH}RWRX^v{F
zu8Jl8SaTd=StKbaTs+$mUeMGoK*cCt{}J&_U4%gN`_b$jEkAX&QK?VwyXdHJY1AAj
zqbd4vUEs%qoUnk4TNlXPhyqf!{vj*Tjyam{H8c4hFqWV{`4WUZ>M-;6_4>Y7QEzv#
z_{ztP^Sq~aPAf=S;F>yDYiq0bXRs`Q?(G;*U4H<Yid2aiEiiR)9TTQVvJj%_FnN0-
zfWg_-2H>UJ0~}z;U>J-s2kAn>_a`q_0Cjd&qd#y90@$=tPowY?P4U!&RYql2P2ewV
z929OTirmCilK4h-6gdxb&EL&;o~9w1&en@-et-Ypn9KhG0bb;Po(aeKc9Q{onP1nH
zSIO;7nLBG|l_lK&i=Yvo9yN*k8Z#P8{Qa=^h5QsMq4_bO>U5R9i)Jv|Rk?I8Yk)wy
zLeerB9d-V|U~;b-))GC2aZ>Sp#&Zf)+EE|@fokDIIKLAATXhvgN+XxOP#t4EhglcA
zNUut_Cpgy>?>z{}pH?T`)%{~Hu=OrYy70xBzWU-vGXplP{<w=_oB8(c;!~G_fmasw
z^9S;-0c^KMklAA|Ra|MCDaI4Kck#1spET+=Pv_>b3EvL9Rh}$cZGj}Y_)cuoeC05z
zQ9thxqdv(-EzYwkr}iGLnDWo7qEZwZx5(!oW!NtJJaTn1f|z~Cif2v<a4D*xhga}@
ziQW>vMau@11E}8j8=9uuQhQsMm!BcJn)Wj1a=D$FRZZu%lnlS$zh04}MmoJ6JH9zR
z*k>tj-eLTR23eV0)Q|qDo*g9F|D$+%^@{L60>=vh$i;H6AHyXs-cbt*0Se+}srR@~
zBq^?}zW@2DEgnZH|EBf~fBq3qc&471SiSy83=~AKUJS!{ZRc8qlRqyH6j3alf5V7h
zAkInz$e}?cGe1AVDfyd<U`bS0-x8<0KabHP3XdHl*-5$wGD|HrL!48+K<q$&I$EnF
zFse~yA~2sL70BBseFJ(|$<@eoP%buxfr%DliT9>dL@a8Oj;jLDPu^M#fsI@qBdH#6
zV1DHeIo-?D>umL59x|#CJ0Ri5IyQmyu019S1HtM=m^|s6Qvmbr14qk8i1Qa)?KVh)
zb3icsmPmuQA62;|OXc$tnJ1;t{Z~#y0tH|>^;$12O3VC#=z!o>pL~95o>K?5Rmt=-
zjT*!|89{$B&P5cV=HkRsLEV%h9MRwVu8a9TWp7w|=q1z!m2m&?cU9EgJ$J3MTC+1F
zvKfzKxW-~SNG}L`ER6l106N)!2aoO6y4v1C>$@Hx=(7I9z6j;T80zdzJJ{Jm?QQgH
zB5~UTgYYZ761f!p1Ds)hHIO=!RNC|KmnF&Fy<Su)s2V&NuTdIzjGq8S&90&Fg@qU;
z{OdBD5vq87$zw%a#d?Gr8S@A;&oAb0cQz?FHy|v+QV2|Tqu@j4Ht8T4jba?1Xj0k?
zA+Pgyt!Zd{eb{&u6#v~qE)*oFL{A4E7jeXt3NJTg5O4%q@)v_Q&Xx|Zx?1&uJ5M|(
z+E?v`6UKm4#j^mW)P}<!Ub$IUA0SOM`mVmrtv<e%0gi(H-8}<u8jndfu>y>YaP6bp
z^y=EIwc@?U0km{Q%lfJ8N-|c)gMX~>%|73higl}ce4t&7YiYelHy^t-1j4^}dnv-v
zSffWs;XMbu7<Z%I;dzXc3?4HJkIc*?E_Lb}XZB&5CC;o*q3qhGuN#z%v|N|$!iY-?
z8*v<^9Hm1AbO#wt&2&uCQqNKXr;GoW_kZ?Z4$1>S{ex_$B|HpJesFweQJg`yKPNMC
zuaiQ6uL3f|TDr&?a(l<iDjPb)3wEl^{i<nL%B)SZoiYH%#L16h><`<+*85)a=>;Ae
zkutvO=_h;*&?l8rBf?w!nY#I`MCQxm9CgeU!33n%7ayF6Bw0(%fg`h0#-CW~N0yzV
zZ;`33NnMjZOM)0k8*LlcF+}RnJ)XZ;wSW~*G*PwZnuy*ed)*Wmb3o=DDl5ITiUzab
z*WrY@co&1=^2IA!O)O70{RO7gAwHGOAe9p4b4y=GwoBYgOV{8HD}y$O?~gH;-V;x_
zDVO2uej(225R+qCi?;el4Ss1BoBf~JtJ(L6&v#oBTwBOFl%!NBlG{AY6IHJz9MJV$
z3`i!;*>Fh_{;6rD9!x@cCe1DN$;T^EK7YDwm8s+K_XEm-oJc(y**7u-u_e{7$nfSt
zNYeev9IV!nev!rkWPGT%D%kKU0^{4RqZZl{v88H$RS<$@5|M0b`9Hzn|1Jijz+rBK
z=bla1!{gnhR!;60VZH+?GV%T?sbDG1d{_Jq&KUZo>d@t!>D`b$(!p>o*;+Paf;mZt
zX#jQ*Hb~Zq%dQhN*fmB}hyGT^qt)jM(0xyvFE*fpamFgM#z9(5n!aJs`6?7=hJNC^
z#FtEhbt`<X3*iF&px}{IZ0$`_B|MykUl&ZUGzn9)aFp<6x@~hpw}Tk@{A(AD795P+
zQUS&8*V&7snnBf3JMWB(1KCCVkatc1gQuUw1%5QxDxg@~+S6pJOFE}gm|!E~a1l9x
z-zvOj=j(sj(~re2R>atS=EaN|4j6>uq&SGui~xmV$E*nB{AzdPOrC9%Zhve|uv*mh
zv(B#8G6K$v8*q_^+lO~;-+H!DqLQ0>0ZR!QK^Nm_DE$J{zfra`H47*Yy5)d!iH(~V
zFT2F^tU{brJ<?iDhqtsJDm@NIE3wE(<Fj$VZU6t*fMB2cXt;CZj>V<Ndt`9@x2)`r
zNf<Nfx`r*&I{+u(i~bWP7eht#m=hrs!Kgl?ky;|G;>r>I{UDG6Iwll}|7qZxp!B2E
ztK=)K13G6#ys^!T0Ri5wKk{mnFU5cQO6qQ65?>03)rTmCGWG>vX`yj3VldQ+>5ep2
zlEe-5?=%XK(%2@DgeTEpx48LtWwMN$+LRyc$dT`jmO#TX!-=6HsHEo9H|j=`1=%I^
zPG8TM0s&gn8J+<$Qv3MI_(@EoO#P&nZEY3-j*uO~;Uq`>4J<saLllGp7C~fvxuua2
z`T8%b<VR4F_{#4ugY!@v4@?h5Uhtb-KhawPjn*MiYp~y2xqQ0%3JFtvr*mo#!bi|O
zlFG>?o@zC5iJXqj$gPhHe>2E1Lx-j3$4)1ylL65LPipM(nAdQZu4wi!(Jn><LpLKv
z44uC4$W+MJMU;JBd>#Osy;Q1rrCs^I&i?0rD37<D({LZ$2=RzcJ9v4i?(~j#kV^KY
z`3+T(<^~<;Payzw3Q*=^k|tn!j<WgEyl>U`5)c~wYRRb*DJ9MAF7-pW|FMB{;ZwZA
z-13H+|ADAE{F!68Ul6wtVZaTjeCRZxD@)CimumUrzE@8(v!)){3sSS+EPQ(eG<nNO
z9Up~Q$ut{$D89oL6v{<rnu{PcYY?thoN#vDEI&}xFX1iN5*qr&+;dcHk;s5;rcAZ?
zyv6XHAxG407`q`0HLD$^XwK;?nDd3D0zCd|Df^WO*c<5O?0oVjd%<al5KqWt*}>P+
z)>Y%+ii#6I6rNU}l;RGhI)hEZeLH@B#9cWL5{}D0Y|q++%bBEht$uMcyx6<7v4EWu
zudCKn$EObmRY||a;}%`CiED=xuZ1u=-H0;kcDIs0xM++(j@d+!3MBCYfBhX&l3E!T
z=$!k!n^7F=8}JU$%nd&J-_Gx)3s{SIcg|_xhUn8k|Cj3_WpP!-dR1$|uytC4d=m<^
zLw!@ZeXFKa;XfPN(gf;~7M-$xh7y~baCLQKD4~bU{*<vpv0#ze(FsQ%l#85mw4q7J
zxTIt<Jnf>#rgyZF%8+!!f1H*jF;x;qQ+h!jik{Nldo(ooSXM{3w)<XP62zobA{j3!
zy@>g~!Vv>3l&8A&m+NHM?cI5UdK*!`4|7fnlWLM!E5LDg70@z0t+xZC+7tZqz=Cez
zCANnFQxp<U0jlvWt<5dj5%qJ+yl(tVTHpcfXt*Rn)oLsE%uj@5fRdX9t)P9IQs-Vc
zbQqhDfSJ}d9}8bw!Q1X0@TLQ(cQsXD&FxxpthA$vgcPo^KgP4}mFk|yLE<8T!FFhS
zbHKJUYu{3wkV*c)l+7Lo3u!Y>i5Qi9-|X13YakciP)06G^Pu9n42c(?^#<t&Y6+E8
zU<$q|5<S~RJ)0{Zl@lS{n%l&=pswROlgVI|qN7>da|jr(lGgd3iP3)nna->A7I7R!
zJB*CVKG)rz7?BC*wV~Sp#!m_(kQ&_uFZ`nM0uvR{f)pUngn2(b_;cv90jzlMA744C
z>NM;TTOo|0issJTy36#&fS*td4c@aTc7y9w-njs5`Z%|U1HYH#FYL7i$@S_ysp{UE
zv?0qhb1{~)u=;P+taP~EdEU*qxUR#k;QRhPHq0CyYj&)=%RqlSl{2{N9mr*|Ll2~r
z;KPiB0vanfPTIRcF&C5aAJcJ^mhEzY!oD=$9^{XXjs~ZX)tg1A!8kAQHQGPdm3x9r
z#Zu!yYDz3`+V?L8({MEeyJnaM+_BXYEk!AqLdwtfQ^(no#3Qa9_$g*x1mDnFmE6iW
zKO@g(EI%tz+Yi`LGy7nB`0DwJL+dAUHBFAwM$Z7rP-{imx)__AMG1V*W8d4_gDyKf
zMRuZ!7dOjQr(W)}SI77d=y1&UHbrU*WUeK8LyG?c8Pe1QY*Su_hQ27Cbr_twT8|xo
zv{98EPkLb!Yxp~DGNB@V4Hnz+bq7iYHPKuc&KKev7x%Gb#PJ2zY~4Ez7`50CGm^pr
zHWLPOv{P5!fnSED@$jXLYQ|LSwVX-Li4T&~-JfEFzWXj{Fo5*}!<Bas@(W=MbEv)V
z7GJO+m>*+AHkv1f?(7BfuXCns#QY%-%aAop={>1_$)k%7lK_jB)tZXUostA(ldD;^
zS*hS$NVv@y-&3W|+}magrro&U4JTn-nl|+k5i7S|<6G)<5O7H8rkVI#3Y*TcNml?*
z9~b<kRh7@kkJWWF+a!MpPDh=U22WcQIjdjGmWHW@y4`N4{>QEM!%j@vN~=#!u^0HT
zijd9EHb<?x6d>IpiIS{Rl)+FDy$9qi+}(!%5ng;;8P3KO9!7Vz=1p+_+8{+u`qq^v
zH^2$hc4vZ9REr=YwZJ~%P*fOI>{|bU%X-AhR9`@~Bb_s+!30O`S`c!P{^%Y0ytrmE
zT<l8}?*B!zK|bgr_%A+C_~2eaMK&H-a9yZmQM3la!QW$|#*1TGe&`c9*20bV00n@L
z06jo3;2ofichJONT1r<kv@e`}H=J7(OL~<WPs2ipY*6c_UndKjZeh(L+0n5^k&wu6
zJg!kS)SlMDS$v8+8o38nV2b%;z$*?q`Q13O1bNn~vna+;?16bK*KaPpl*kGAM3J_W
z<%vr0$&G&J8gI57dAPDFSH=04%5epfy0FrOk!Z0RFHY;RUoj6g@a0wI8IDo5(aOK`
z#%<W4*sd)7*Zg#=7|2-g9vcc2btjYM&z51(VuB@Yvf6O5tSZ7q^UD-of1^$xyy)_M
zatZXin#hW$qw!qhmsUIr`qoicMY~o89OloxtYx4^%@wOV<fr~<)KNo6)l%1`2AET*
z<b)+i4eBtymaS0ITok)AiK!?N*YM7l(R0|4rL<z&X1@4Iq>4#LWRHGklUI+H^VcLf
zUdh2GSn_MWFgE>Jvi$oW>}Fd3<BpsT|3gn*$v$B&)5v$v$A^iuL3i^-8d^G`7vHD1
zexnkXgvH<nP3zA%iV7rKouNx_VRJE$;skNhNzUbtm|p~P;L1?`qGA$R2Dp{s>-=pO
zYwpU(s1*}2Nmf=79X@-Ng`<coy}g-i_2Kj}r!Xgw_s0AMqqTcGh<!X06mv;2=d0Lx
zlXn`h$K;BOhud}0K~1EQU^pc^<(0EdB#V0yIw*zq^c_3ZKkZv%URw{;U8hd4c<n$A
zSq|JnL=FgAFxzFQ83Fgx?>Dhhsg&`m(y#Kisj5^*f5xn4<%m=lj?tPZ<g#S)nT*hs
zFffuVnj&TEW*>gs4Iz61;y$Jw0Cn5HSy1apFezNNV7$&?B(8e1HBD&6^f5L1HnD5b
z-?Z6P&UDj5&hKC+!&^Z-n&s?t)q!c^jO=EXb0O&!1D7*ipOT2jq$Jw7UJs-7Lz{7G
z!ac>~<k@IyAfQV~o|dIhM<5g6dipEoe*vhj{6I<*&C^JL(z8&o7;64%w>y56uZuzy
zf5~dia`mF#Ao#lkrpQQ+F^C$U&6LkEd)-9I<&E0Qm?6JW!l<{}mq7JJZyG#?DB|sX
z16i337&o=;!5B(%xGyOqIdPN>f|X19_YTEJAl-qXNxxs|ui?k+0UCi-=ZSh&ze8oP
z4Gxbqaiyr$VDA2T;CB&;?(Ean&}X5b4=)1vaTJQPRGE2lp{xQ;3dD};QeDe`)i&r3
z0QMOvD+e}MnJS7higoIBNca)}!a!4cU!C5fNz+Wad82xaf2`?Us9Mf5$mb}K5zm+E
zTwb}TRR|377N!nr)h$VJsD=h=owScj2#e48NR*|1XzdkiF2iNUltN~j>Au9Wubl@X
zqx3$)b*q?>`yptZlN9t#c;`7m{x!Nr5Aw@WU5{lvi}C@Qwu<bxxy`_2_sK!Zps$x2
z7IkGaj6-%r<6lL)dQ*NC5%EWQ6@U__bXCsYjII3gj^}uD<;}Fs*g2p)J%2Y9_P;XN
zuK<DObiJ~duEaDn=ajSiZCiSuVG4+D+I!@%=#hLCI3~fLV=QxL?ws(R81*H%;DD^;
zm27gi+w>d}HW&AjuSG(ceDRL=*V{GqqvIpl@<k6ckiuW`0@L{{H6|Xif<bMF6K^82
z9av=Z-e}OxZ3{oJdhg}Ib1;elZy`tevG#~gEp86V>hE})PVyQTBO_-2w{^j8_?_qh
zv&Sv&HH~|#Uc7P8R7)b2HkXa)+$cdfeCJK|0`9;j;CSPKcR0*4iUheQY_}RyAiHC_
znL3<0Aw*%`3$^{-)#u{HM0HoT4iQL0up1e=pzC1!z=cw)E?=TfA;OFCgFu%}7NyqS
z>AL}j*w(yClWD=};(4bWS(vx*tV7!yv$X=JUm=%XDfb|2e9v(1LFCAEOUr_(%BXOQ
zM1dDDlq18FNRugFJLLSjQo&?H;{dNk@Pd~o!nl^Qgcye=qi*9Q*O)Rfl8KXUZeL{>
z1<L$43U$IQE4Nanc-D&=(_X2ub#>$$k%=Lp4aka8StMg}TTXvWw`r3R-wRas_oL{G
zH#1sWl3R<1!lrSX2$s)}O=g3cy<O!U^Q`}}U{`D)sE7Sp$?TuYJ8iC@nGHO{u%5VS
z&25Q<zg&8D><S;X`9a=x(nZ;M7c~oak?VwHuKm(cW-2t_D@n}7vv>?XMHCVc>*_-;
zWZPRj=D*Am1PQ$ey8s@&I+LND>AK$5ckzplaa|Tx^|!JN&Gq(VdTG&S&Zhi<S^FVx
zEUZx?i$%~()v6fVnXh1+#sDb2`TK)}Xo%WDw(LPnq;l|Qd2QVB@iAcHVZDQlLV+9g
zHSQAIot-9n{kkghADB^#Qa;`nKK(^NX`h51Hzn?)+&w%Rg;V!vfJ{iiA!HDB(0~0_
zV>eWgBKN}o_v8NB^=$7M^q*2d?sjDh#pSJeO?4^u>-fC!^PhK2;EJ4@(Ygq6Qe2#9
zJ<^OJ?GXdgDhO=Yr<XPW0=cYiL@*Yk$=HgM4*}mz@>fKnzzYX0gp)kU)vIs#$rDn4
z8fcO*Dn+j4s(I6F!Mx<UmO~-UT~jW6Ir-{NC-C9R1j+_qH!^rl;=t7uBXfZ=O1j+W
z;MIDNg8W{qSfN$>3K=DKq9rh}`c`#|su(-WSt#~~%<^awX0L-6-(+1e4JFgqecMAL
z(iy;Oh*z}kqR_M!vlbQ<;Aq~IKX#p0;S^mZU**C;5SI2Ux<+^6;%lvnwL@K*;o7!d
z9mIQ7$>57{lAcaRLgtF3^6u{vh&}tl<6b50Fr(~ps6gZwZQ-up6jL1&bZM3{4;FRr
z$E1UWM9sy(XIq_0b?LK8Kl)6|_v%5+VMIi^H8vBl7&T@h3F%IUld?p*5!sO+)9WD}
zB))Ao->p8{8>2TwJU>QkH_kV?*0z=fcHNRm0}t%=>?;Jp3{<d^Qgk{KaSB!Fp2UOS
z@4M|DY-j9ESiJ+c%75lqBX!{xK+b~q>t8ge3s+mgkPQL_&84U{v)x(Qa^d2+4#Oe7
z7LK+^u(4+U>WYeA>TR=^+tzzU$i+v~B(sAF(eP~c4B%;U-u5qkZbZamSI6{5gu)we
zq}bU{d)#Y}b0@f>l7xb9=2unCzqG@cV~WgoyW0LD0?(5rG8&&R13L6=vn)+V$RTng
zS<hk~Msp^KRwDKq-~a6}=3DhiO#6K1wST#|+Q+nR$DGfw@q12P8k<%DnM;T6;U3=v
zpKRP+ZPh)r^pZTS1YZ^oZ9GDPMbN_fhT1(&{@G@iZhfq7!Q@L<d>Wdw{I$>b`@T=<
z*L|)JHclreBVOW3KM@LOQ)lO*yK&zW@Gy9yUkNmR601E3Bw^O_LSvR8e^Wl|P|rCp
zuWG|b#9?G)wC-~&wimUg4H*=1a+q@zrpd`8$5FTN^02C_^xQ!~q?h2zf{llNX)^&f
zj|Jdk?+Qo%7&Nw$UL*_Y$F@vW@WW;!okW(C9&!aOua<Dg(;>A8O|(Tj(FW%EM5~S<
zn2lE_wOL%|oZY*6s(Hj{ffl!s=c46XOY9l0BRNI-EcV(E5Z`nDtcrbks`VcP=xwe_
ziT><qJ{ry|c5SNk99CbQ7X*st?W;@<9p+2LYIAA?1RAjAY-z$zCuoF|I>sGfw@8dL
zQR+NDhq#DK-a$kZY{#}8awM&b-Osm^kc`RN{>vA8-M@+eE<a_Z-L#@B?w~@RxL?XN
zNRg1mMaN6`KJ(po>%O>}H4t47FQCdTV`SvYxkOs?VxjF~D(5gdb7c-pFgF)eBrKIO
z2dvQ@G5u7)^A?3<hC_1wU}%_KYw228P%6?Kr+vHyKQ8<yBbOp)V}Tk80x9$T?)b9@
zm&ImvtyE%Cm1PeyApW6DJlwqG)uj3Q&!5gb>*~-iZe2VW+ME(;=njy%JyZYP<t=h~
z<tp12rAO-TKF}JWy%}f_Z&I8Pjs04v;8I=uGMl$GF)F66s^es_Fs|x=Naq;uC+`k(
za5Z0iASqXCG7ePsl$BcD@9&&w?<2i&yQV>~jW&URLUsLAKWbNA;7BVVC&CG>D0H%{
zamLD*aO?#o!QHF71k9ZB#JRrh&bPhyjK*Sqg#^2%JN^ka6aCDzqc7gxAe%x3Xiw9<
ze>0)I_gqizfsZ$P2=LRbSyJy`-^b~f$OY~HqJm+cq1CL{t91bPcHd6D1n^1yzUpbW
zmz?(VQ^4%LkLB}C28Pg^$Py-PYAdqH*QQ37hjWjrBL;Ym@2G}=yfJTJ9OlNO_~t?N
zrgdx#b*>)YA?l5CW%+fqI2snb6iiYKmNt`Do7~_T!$$enXqvS|{M69W(q`UIX~wyX
z^JoVB)T;hAHwTJN?+#}V+q!jE9^sIN03iNd)LnzN_*K%e%C6jb+nu}R)6cW_PX~-l
zAQR7}!<=DlkAz5+RM$yvSrQ^u2VOhcXmTqCr!Amp%qDktU&pww?Fk@f&Gem)Yd^J%
zfFs)p=NKAd(IxwSOT<(UD9?kH*3t?<)8JWF*HEL{lzW*BbUFJFnJ-}T-r|(NKe^Ff
zgvgZlnRA~KZj{7=yu|3LS5LTLNnpI=63M`?R)U%-#?)IsQSu5Z(T3?joJ}kx1^>BU
z>%bBzQEAXl{PgRa3ePktS_U*ek!}G>nTSJJ(*|VDuO(_YQAWFxqn!r-1p#4tYh<3~
z6i@WR{JdX}s9TSh09pg5(1y0e6x3GYW`!Y|OP=!w)!skb+o0FRYtIZh10Xq(*Tb7i
zcw<qtPI`4f?w7XjwSr8CY&gU^4QpJEHhw_yBSix4-0!B18oIk(i{egqa9?r`WqE_D
z;<J7ai3@ul@5aP(S2Eq2#KFS-4EI+n*?(@g!j$RN9x-8<LiLYp!r}s}u?1C^CU}Iq
zev;`P8mA4a6H?03gmV^b%M&U-_^mkRq}eRCb-XLDmQ!pSl}J9)M~up6bE!s~)9MV8
z2{1acSTacE5KON-3c)&>P3VSsDP9sG>erTi6B-OuK_-fKvf65A=Xza=)xN)kP4Nz1
z^{|fKwe_3bssz7(oRzp$==1;Gdp*Rp<$9kW5#4)-EEuuT!o?QWCTgm;Zw!VFk+j@C
zHq@l8vnsRGU5lo*_^Zzs#n)@S*Q8LS$YY#g(`#N7C$eRH$yvc!nTptt0e|1YaL57W
zJQNA=53>3L2qBpbbA3kp`S9Fm^{9((pW^uxo(f0;CGOe8y3WA2v9fIJ^a$`(sAl$R
zeaL7Z_j7!X@L#!V8ZBHhw%?6(BVQ}OwxTL9`#n6l1#wy^>g}bBYNP9M8Gc0KalUP8
zyZIHOwW}MDgf%!dys%Ro-*#YDiHdejBdKnTW#IZd=iWG_A0R$!NI%($>=Q_$cvGls
zzI4K+f}^d~x45&3s0I$-cXpQRUYu*e*J!b8<vV9^wwPh1W}c*nIq(?G-6*RYey_Pa
zZ0=+VJpCP-#ZY3^-D<w&_~WdR5)a*lHl8<{DViZ(GG4lQN{zLc<5n8S6_0DNmti2c
z1$_FWh?Wk{-4io#a`7WA>*6O1r9nV<K@8V;)Sr@>ARKrx@4h1jhlE<hJI!RNNcza{
zbDAUS<&+5>@j;6-!<BQ^b^|!+RbAyEZ>-dibDWEBvsJCDLi1lN3JIEjM{ka?h(8`6
zWezn?M&5J!{Ig$ITwRUdnW1N2{=fDMHu<J4e8kuDg$ti|nI`mnS-11`lr9$%pc4r<
z?+wPk6Wmz3GMYX15lfze9i5d=#F(CrMmczI34bsp*0sKs5xxTt41mpo4uaNH+!9w3
zB{ybd*tP{zZ5t`Sj8(i^opOtE0>-}aOHbXyh|ynREH99ZMuI-KF|mCDI7{*wA$za}
zccB@1uh>|?#8?!0O!z2SYuH<(5dnKbFOty;2ZWmm8}~QvWcymj%)KxLoH9}!)4n#M
z9;p_dX~8Y7r=4Kte_1IylE{EcI$9@j9o9=8d~)(oxV68ps6e*iIO}J+e9g&%`~Y9v
zGjeKdh6<8AJmg;=yFSU^?~n9S-?h06Gp$Q>umSW=!$Uz8U^?qL7{*MwKXv~$yF4`c
zoSYnUqA4rBoF`kYdJF#%7t{lR4J8av;-2hr$8<oZXEJ#xlPtMKdECSLq`G_sqyf`(
zKTltT1@>qClCStVgzZ{fI>&~Wo?Rd#%t4MF6mq>PE*if!5VW`_|NAp6QLJHCnJ>Ko
zMX9N+3639L9ikKAk~cMsxFIbZnq3RH6wlQtYoN0nwgQKq2C)lympL+)W7EV_O54V#
zyNIW5GQ0$}{4mT<jHx@$7}s2*OFZh+u9~=ZR-oNd88LE@dB(aeje0~o$`oB7K!=uY
zMi@WXO$({7Qqk>dk7Q_>`Sf+>o9#OnZjle(sv2zM=2qX&6NbyM&l2l~+`@iBQk>S*
z1wlENyA-1Bfa`tNg%x>@olk2a@S2`Pr{O_Zc0zj6+&Gz)>bsnT<wIB2jhqBI%D-=s
zWcA^1bBHYG1qyYD!xlVa-53-E^7v!*PLsC6a*+fSl3{j7nG*Tbx0{{FR+sQMXKW@F
zYJPvuERob%xHb<Q6$r%NL6{;#WR09SnHlbQSTiyoz>&F#HX-2kf3+K?dCOyJGIx>-
zH+4EW8cCpRv#|A*z{PbLrV~g?1XEBSJgUJujAK3@RgI~T_tbKQBs*AV1impgtAp_c
zveaPBtc`4*b?A(o_9ptJ6!kQ@8li^iJ~hLXcVs0DB-lEWPov6#abvX*fJUl8qa1={
ze)pnc&z|``F?3uc{#P8nZM7O&a->4G*z}P}nP@uRP5}K^X4P>UErV7gZ+5uucI_d;
z`vd>#-*LRIXi==5l~1bmCK`bug-`c;2=w67!r|b@t4C8R>Wz7VIljosnSYshsNNUl
zbPg6EpF6a>4I0(^v<Dq9dw5MG+wdW-!W@)8jAZZ5da)01jM2qEOACx5OAGlajzv@`
zqny9>IwLZyJIaPHAP_lmGng<iV7Fj!g=t2;<~AOHi}3a!9N27#f3Z|L4hk(Fw-nDG
z0&??yhr0Q(sPLL%a4Ak+&XY%?MusI$L<1NjTrD0HM!OOZ8(Mm)(S2P;IyY|AkEo8b
zulp5O4kr}X#z6srBdM<}U%og2x>5b3*I*WyDU6KQohkN%v<L%&$AB@|Aq`nFWwT5v
zg^dVpEknh+ny@gt6DNgUNSuG1W~4WHe}Y{bq+u~NFvVPkG{IVv#}jnX)4^M01TVv3
zAFl=<X1K@O<|QJU>4`1<jg?|==+R6w<9%|N?aO4zFBuX#<WXj7auWDQ>wKE1MMOkC
zjD)E)rjq*oeV?C{ju%;c`@QT8mayMyy{z*2NNOg~>l)uiS{;4^)-zd6$-T@v5GqWw
zJmG6$y-U*=l^yN9wGo?6lYo{CVe3<mC{n4(XXJW6qvHJl@0)mjD<IRS_zZ_d^msaB
z@)o;gg8lJ4Ng@$xtiD2{QBrj#zFq%itavnTTv$6&|5(qFAqq|xSz?}K13u-nHR^(w
zxVX@zd|xDC>;ymZ4qdRQ*mHl%N%iqehDhYilR?7E{r14C!8Y)-S~@iOmtz+#jGen%
zurcoX&w-1Yuphg^do3wELp@#5Voqcj<>#YS(}?Vk2YKbpOCgoz*R>2}o{x$mTw{D^
zvphkPrv544js%=YHSmWnld0o4V3=R<O*+uXqocRn&Xt>tYZZA*m0Wq52&S!|l3hH0
z$&QQ(+r>`wU-AN^TGCEDsx;xYYJ{-8Ue3zELfe9u8&2Fz%V+ByiPtZX(&@Gc!KXj2
zLr4pLGR@2Tb5q)>a%Jt!^?xt{FMaE`z56d+*~RZ`Xor)Ywt6X$waSoujSUw-5zhsS
z_;%K6i@#mKijf=QES!H+D|J(tIKrfVm`mO>;J!d-T++-{a$Pv?_QgtARLHGDbg98%
z0brBi9{6M$kA?#0vU;Yv<y0^T_To*57|;4{ESe9nJ!M)2sWvrCPkD7(#E+8wgpEYL
zh>S_&wqZ%BU9umX-jw?7e^>M4?_*~}Q!gp6+|g8XoS<&i&aG6@<nVzS|DsP|;5L4I
z6M`|LQiKsS)*-yE(eiqwM7X2Cww3`Fm|UmEgdw%qw6+$zzFyk~U!x|H0cj=pFbKcx
z1jB{z8Q?r7Rzf&O++DM)HIcBAC)ImRTG4q>vKN`mcEnsv7xTyn6P5v1JZzr<lL3Wl
zrm{XcVvqVE%vYyr?cELI2b?p?Dq7~sy4PBm(m<plqVFM!eMD5DjCAEA&^np9QEInA
z{rMiCg_{Ob36m~XYERfswcS{eb1WfV)J=`0lIj*abh3%z%>?;y#du3*-W&tbFs&@}
zHVIsIqmWbG66rZ4ZU&Mp-F&msS^ot@&o7U+jXy0r|1(>hBhvHHiHDJHQ2Ox(;2lt6
z)%gJ4S-p7Sa|EXyzMM8aI&QMj6Lgr%q;52BjrG<xQTn{U1{N!;VldnLW*C7RU|3D0
z(c=NXt$M#=<k0f;`!pJxmT_>wf(=q49f{u+e&-<8gkC$-VbzH;-B4&>KkpRsiBC84
z=KL)&<w&#n$qOeH(1ud9T9HOjd9a$Mc_&Cccc7^*QN$={+8Vlh3kHUcO^dyudYJQB
z(Gp-RuvqS!x32{sdFNS=Oj)6)4yTH>s~=#r+qv?jH;KMs?{;EV6j;$D?GNxgSnZ+b
z9F9Eu)$33#wygKkm5B_?={F@Fk=f^CzaLJ+GoNqW@27e%kS-a4_xPELdJW8@p2nm7
zuR@FcGvev@iZrkop@Z&Tqmx8?Kb-_^L&_HvH#JSYYZq#HF_Z63lWH)nK465N#F*xt
z`cA8h^sgND2t%-Rpj6_Y<OmBh))5s}Xt?n24M-f4WUm%L>+HV^kyOUFBKk&j!x*?>
z@d`2IZ>Cg$9mPQ8#Im+b?U`@JY2~IkqA*j3{S2Fdl0{i|b!ykfiS~G;#&<n2tjz_I
z3hye@O_6&^pnxHzNWM*$PgRu{GFCWyy39Bq$$LdVvP}4*g!CIOrpUfQfPlpU2^lj{
z|6M70>$h8ek^FrZ_;3dHR?p!Y><ZbMtPIk>sxB3%P$)J?`rPJt!ioJ(w3c_9PDh_y
ztr9Dop}>qxMrT=RNyAcrVWao{%(Kw3TJ(DgnbF$~-E;q-u^-VPhp2R8JpIfTyj$k-
zLHm5sl?}H9MX{Ek0+0d75Z?q}?_7*G``R<N@E{9YF-Y1_>Zrkm40Unr%dhmo!I|tN
zU`>lrjh%fJn4&<=UPjk@XJFkrUgS2JciGPKutdPpDfd_EWH{g8zUboK!4`X@Lh<zq
z?Pn?|6m#vr7sV7Z0-bn~wDT{Dy1ve`v*xvg6X>}%mv*|!mA6qtI-}f?U8UtUk@$SL
zNvav#MFU)F@<~h5Nd2b@Bivo7w+myL(-1C!v(!DoPVFJ_BAw<@%dO=rn}wKcAP>w1
z2exNY_dAHuF3J-Udn>I^Sf5P9Y+uwRZg$A`ZO$!VzZ3GCv=(=-Jcb8u$*0quQhv=f
z|99=6<t9nP`ox`~6T<@gj*t2F4Nu0b2%x(@$eMJ*bN~`x{{Wr+$w)3B+)9#dj;FUt
z22URD5pOTc(aiR4a&sIVtHq-+Tl`47fazDo^Kr5Bj~c4wmjw57mL#rWg~W7ZIx&B(
z@wBuN+xuj8|8RX4HxpR>ubJ>C9iUVAn~UYzW{VBS0_R~pA^&=D<FV#?U!j+z=Cg!@
zmhrC*zO*pn!**f^);a1_CJxrl_6%@#ng<CgjX#`KGy+3IukFy7;KL*RnBkg_ldEg4
zQz(+4U{+M2ae8PGnJbXaM>}I7y)U~Xs7DUsDMk#&UPkgx;QQ|JKGUx{WLMuaK(ltA
zW+dy!y0GE6MGqJ5dTkViK6~h0%=*4#P2J0gIzZR_Sv-=-y1!0}T<T-xX2tSqyXw8!
znKr>5+YBSQQT53XyJ$SdC3U~Wbi7DclNIqhquG_!*ltQe%co0c^3i~X5Z(&M53h{E
z?~Ib)gT$Ok1~S;mq(F{jlFct~1OY!7F`7^ip<#Ams4|?$!AYn)@_PRz!k&wJ{BrIf
zkTdIiwQ#<ND;6zC{go=3v#t0Hus~;@6%iqYx)Slpct$*&cms1K^hUuCXfO52{K!&s
z_UiC{`J))(S3e77)d`4_-jiF1$6a1gw*QR)B?tI<O4XV401@b(ifLA!oaPg<07C45
z-lY6h4N8oneo(DY0gfR)Q3Rho-3bjdTbliwP&-8nDNM<Xu+`=HM<zBE{Pa=ueyb<L
z3cO@`-sTIRN-r=GECO#Nehy|^&9C*+IvAb{3yHyM+>|)CJ~$WORd|!M8~ehK2dKN{
z*gdS#!3jg;{HX-!MVN^c);t2Ut~-8=Mpk|@oVcj1)kul*(lOMs^9&U=RonXOOS{!H
z9)gnRGkt5RvG?0GZD+p9Bh9c?ZB=rbWk>t*o(Do?gx`sE=NyA0J_2Q9ez}E{F=kaf
zkyxNAZ~tV!@j+sjh{ytDo+c&6mclOfOTQ`m;+r+?TaAD$)*9T}ZB~|wpFrz?cZeY?
zNbs4);|W6I9`EV53dMczlf)DKW7X4PZ>)y+4KTeZ{`>z)fw88_zWy<aV8(G8jS<-h
z;_A6F>Lu_M2TBN>j)qcwZ@7F76Pqd_o~cRm?KC0M1753^S+aqFv}(iuzWfRNar;9(
z5Za;o-|mrErV%g2U3cPGu_v)J>p?gX<0N2{@WVK@w5y!|lswUlozdint}*M6)@6+n
zluf{TRScN1eRE#%b#b$&Y80TS70<eh=18G<17T>?8BV>xJecry%PlrdjDl)J6QX)&
z%lqt9vDldf52dViMNwMQ+MIFnO))0-m<pM^RX-G_?cuWEI15>cft+-awrvD(wKmTb
zElr}RX}n#{>RM3dWzoOFwumXHwBbQYQjb)O`GgXmbr|wVe;OU_t8%oCuScV#dv$;k
zPqGBH7^onRkDUG00`z~s+qH_}K}Zf*W3jF7Dr<@v;A%x)r-IFILq~Gs`(6vSDc-1<
z?<v}8+gi+js>xek-2E;vH=d}L-u_j3959UM_~V0)=!@>2%(<NXj|#KBASWiM({HI;
z>wr0%>P-18rMif(I9o|RZbE|*+h%#Wr?>MGw|?=FLr~<n$Dah5N4qX6VzGAQZ=+{N
zk%9J|2ChTX)28RGNHW?Lz-8?-^CjMiY&j$eIkxEash>=lc^D1RBI-fMaXnm)SEkOD
zRl<GN8IwqgIv)6;y;pY-%=hdL>Rxkfn(E9j?~-1r9HW!EcNX1S(<*x`Wq?TDn1iT7
z-rnrA+&nqx{}3a0&$jXYWp<shkuQH^UU<`a*+5+tSU(Y6-@yBXJ^Tfem)ffD`Jr`E
zUHVkin%CQa6t`H@Jp-`i$j9lFv6@EO2G+es6b;fwo^&*NXSzFV%7lbpkM=SiK`9oA
zTA_XL+fJK)<%&2B#=1+a9<kHtg^U@bkXrD0(&J3Q!mY+r=acyHMki_SC9Ibw7>f6(
zd%xEEH#lCtsS43f0G(BYq8SA_Kl>-KL)@jR&%vtkC)Aa@ma@NME=kb%5H5eeTKH6R
z{pkA0(t#eplPT@wT4<;??lMfS5EF^v*aDgR>zPpZwhLf>q57F!!;*c+uO!GR#HQ*s
z2YGB-g7o@vv$g)Qd|Q>M=I;+ZbW@Q7WXNs=FOAZTuDB;FY!y&K0*yICnro)jG6+z8
zf;!urRW!-w8Bh8;lmM`}f%beSnWm_+4IKB_r`{aUy?+EBUhk7SYFyW?vL_!-@#VD=
ztNUhI1&xpwS;goB>64@9UUrRNGyw|K-!Pb(Ra-|sH`jW_t8rI7vItp=6FFf*v~+hm
zz8LWaY3;Yv>zb-!9G53eh{hr3t57KQnEH(+?Kb)43>E8RF75F~-%t7=aSXz!f;4vT
zx`|Ss%p>yMEq3ARs?swI=%~LVxko+xoj~Hk$T>ND5rF;e(TPPVGAF){#3~EBAbAx7
zIh|Wh+3mt5LsGecc@dpSUhpYz<da7_YE*=_7)eURY5b^y(Uw}|QQDez=NtyesaeGI
zQ?qAHpRg?`;(%Zn7^axZN3(x2Kc%m(4&7h>aDS#DzDSl+qm0DJSb&zAVlHNA*GkCU
z&n!UaWOJ;QD$`-TKOe?0556U+A1P{V^f###3cB`CqYJtWEiL(Y1w7aZh%DWdHXG^H
z&@^ScMjTE+Y2p^4HwU!yoKw~|)BA@>FAfUC_cHMymNU;=GYg(8$zyg@J4|>3mWr#P
zuM1lu(ebpsE1BMQ1TX6jZC`*VIy<&3PWpnxQ?Dy!dsYLf1cM^YkGb%eMklP;U_btX
zxVJ}i5<m6RqBTu>{w^?Stf{Z7kD*qRrSzfnp%$R0G{49`jg2C|1P~EBOIX+*iZD_V
zSX@tA&1Hk>p#Dw<qO-b=Xk>bCFHnD|8}R5|ZLRAs26>PF`8|YKd~8bBJw5-rW51Sg
z?fbO$<aZwye2TaAgxGj&x@HUBsRITgI~!q7=`ppvXL^s&>oJYTQN7b6wd$m(e||hP
z6XPj1V`CU;^&Xt`Vb>l22|s;Txa&yL6LT<~_;ufvW$|JxdCAKH0b#+^t^R60W3Yil
zECjVEjNHuNS34#tM*4~sMcC}}w^ND>`SSo1B0Auc+uG8S?$PnFd^cU#M{offo=sr_
zdKeiEEppdO^i7)PXJq!R)nDqB%Iw#akr5~I@unOoB9)kbAaavuWM9`WW959&TWM-p
zQ&VGejyOv(hlnl54CD%01fNKH;74-BwDrw)XQ}INygI{>mwN*IpwiNad~QV&T!|jr
zqu!4DF`BCly<w?Iyqt_ZLEfq1`a<MnaH?%1rly;^?quf|i|$~0h?SKMv7?m~Q==pg
z%=DsDETGsN_$NppTbUENLHV>S>^davY*Kt@B-Qbo$L!x+e)4^woD$Q-_j^D(AP>;6
z+tE7fBv1L`BcbGy`!^Y8Bx2)V<*Fz@7Va+7?XX&|HkNcAk1krT&s+_LL&)A7z4=CL
zu$=ojYaW+O%W;^}oZMbQ^uX@ZF!8!s-b^MjnfXda^Lt)rfz%^<oNt85v|+ILmId`i
z7Ye+lG2Y`Ke?jrPrWs2C6C3mY7c5@M484!}?;>!7B(~$y7&!5-pmIQ7AYlund;cX1
ze%zeezicgntR?Dp=H8ZFETDE=>@Zw*C^%mz1}+pM&J3fD<o8U2#9tY^BRmS>?)K@&
z!rx6vyWVcWLh~#-cUPHPVBpZkAnkQMg*`9|<5Drq&A@eThuKITOG-Z2_$x)`@~35w
zcQiMauDU9;lF@$BM~6ZJ<m;bgjm@Q*!bdbO@2z-MQ&cYjR#}9jwKFWN_9LtkCK{i0
zScVzb>fvlRZ#}~OOV;x4{%8Rcs|%k8+NPAk;W{E+^^czi`h=sNE_)wbM|z)nudkp_
zr{G8bd%>rvCw^f#{5t6o`o#S>XLegSWcE-M%>Rho+^(RUeZpSp3;1_LFY8brb=Z+`
zvMYCiXn0-OGLmJE6xSr`y$R?g72F6CnWns8q%5J@RpnetU5(C2lxmdx2|}k-eydzc
z7$84`7>fblh923quFfKvnTdh=lvVSvq06O=7_KY^;!id)YF;>lq~>TJKpLeig&ZYE
zZwtrVj<{p-4JRp7td0GqQ}2b%Dh0HD2nCn@#b<tf+EK|bSiI<};Offe-`Uv^0Oel)
ze3dUGD*Eo5G%#xO4<nkw)teE?i+qJV>7OjFK{oOZe3ut^mdh}e57p{ZU1%%I+q71x
zwlfDnwTlgrE`ql80IuV%3m>o6JElAUvBrn<54Y(`RkQJ}GBom&q<pUVK72V-88{IL
zJF<D6aH^77n?DHo_tj5w;^F9wR&?t?d}WZ&{EAS91A33;`5%tRy=9xA`EV-;%*R<c
zn8@r<deY`lQoI_u17}M!suJtqbcy~oJWh{t2mjKeyw}--7*BM(pAE|+cKPg!CBDIE
zo`$|62X&^?LjGyOPjlQwnsE2<<$>$1the7Zx_Z>bAa0!Rjkz#`DtYStI=){T4B8e3
zQe26@&p_zdfc~w=a{qI1+gN{V&i>D(uhl=Ygj;}U*t)%3uK<^>>#^|spo6Ejt<zz(
zx|N8UIlE@6i~vg%6H)IamHtoENdc4M2Mt*zU>)u4NgYsQN(iwAy&w(qO4V@j;dM1|
zeGm-p`v_JVtG!%Hjjppzzi@bjfz9ISO2@op6*rr13+!oB+m?K&mR!6)U(iVwSW_JJ
z?R-CyUa<#uB%=#3%!*QJ&B~JacxEKzN3CUjBhmX{IV@ABu)Fih>&acr=gl54>I?H~
zpt=<~97ciB(!{p{Tlv{;@+s)a-u**^!k?<`8zytU^?<7|;9Y(jwD<1h(ddFMsEMHW
zDxjA(_@eIK)w-8!OX4=^Nf3eTojg@Qmmiv$T~+=YU^IPRS$Ku4oVk{CM9=~Ps)%Us
zhw--VB<{_f)|Qk0PEA#`L4^I7hWKoK+->VGiN0Y`a|M2b{*{X_Ey=k~=~o)<JqUa&
zmkvTJ=fogHJquqkvO{Wu=esbkaztZCb19b1XKf+)sT7l~IqDegsVK~O<7U5{Mky6d
z+F<1?v}g2ptIlwpGckpKKr65#9n2h$+COl)Dk`s8JdsCFYRZjrht-HQ^!Ad5pE7*D
z8Jqp^`|j*78Fr#fj{nagP4PjpND{_`?w`}18|JvIDwG$Os{E{*&5MLcd{?1(EoMN(
zcPs4VNxAFq;~jFlh(tQ?d2yuA;Ip5w@9N~jM{Yz8M0x!`ExrlptmB81>k`fg^wYao
zvB;Q&(LMJYq%aP&z@D_R+)JST(jfUmdCAc6y98j+D_eUZ!AaqEcu>zjm=5L6%TdAm
z>tpJewzbldzswF7^byg3BPcq5dP{#q4+e4wz820`{q2Kq*F3i1Tr9j#)nnO`kfX%0
zy^}Q<&BU1bGmG!0-5()2svu2ISU`YcSNpl)Df?}3k8Duwh`!}EQn!&&y@#q&e7fvk
zg|GG^=}+cM?f;@J_?hT!r2=8j_hzGmj0<^@R^Kb=Z~|l&IjtV{kn8RSYZdF%k853Y
zdY|8-b+ELgz?Zt9^I^P@k47%gM9k&=lGPs(Ua?mt0grRy5~BP{=crxNgJAC<pmX5?
ze(|~Amcx6PcQ=`gaw=4KewQC_!nbBpufD+cCFGoUx3XdX?EIkQw(jiDi;|&X75o&l
zyGAuS6^t6i*5yQCggJ_ce@tQ%n$SqSj6s}JPnJXOU3W+uW-8~1`iZ0Nj+^sROUjkj
zarm1l6X>De^lRuD;%frgN}#E{>B{<g;?r92S#MG>Y~vOQ(Vg}_;z4^K>z=rT4gcvt
zA)6-HPoicIEo>~;OXfjePu5qH%<k4`M~pW7Bq)s5gf^;Ipj8bPW?2m{?PnbXiXrf<
zUq4}v;@SQjTwMVfTAYH_r_#-mn)S5{-<9%7)`h3{78vEL<`d<EJmvl$lFl+9s<sKk
zDo8idjWkGihtl1RbT>$MmvonOvve*X-5|N-(jdKb!?(WQ|NXP)%$aBAnR~8#cVZ_8
zSemTOK!!j%jFLY^vh5_CH((`Q9*Qy?Q8r^&2}E>mQ}@bNZQ$nFc9S%eku;ZviE%iu
zJ|i8_f`XQ-ubS~9cpL}8aW8-GS__xU`BFB!I*kU_JM#zKcs2PfX$mt`EC%(xFn^x=
zBG#9BtL70-b1dKz<QenJ74ECkTCt?uTD0OtIIc&b2kdF>tl#yR!uYV$#=^jHn~@(I
z*H7WthUaG0@!y<q8N-+x$cjV*Q@l?;MKV6x(GwhH-~25nQM!<wV-xI#(Ci?)6nRsU
zT`XK!6-|><T&#tp6ng!wqPj(B*7oQkFE2+oKgiPk;xc>3fo+eWhmSoRVhVU$w5gNS
z4i5~S==&Vk+?BB`1=Zbi#bQMAtc1M=L4v=U+p}6AtdOhd8nF<trXds^D*sHXH*{Qa
zMK4m_j~UX$0we_Lhzr4az+YGfTw1d6=iv%Lc`=&`14mltC4u?5|5c08vL4@tyL}NJ
z%DE{8-ApXsB-h%n6i<#_J(N~ojhi|R0M@HkEnOCPk5n_*#e<e;blD%srjNI7-#s|(
zHIj?ziTt#xkWO(V(p;NE?k`%M|B~0K6on#lciq-)=Uvb5?*-n)&%QN(H6S*{Gr@$G
z>NT_98%sg>2g&#$<feUe$=VlGJANM$#RR=>Vw+pdrN$koWRb?<F!(|}kbs~-r>jfY
z722<#Z(a8?(_9)f8X9o~37SCW*$ZaJ0r$O9&IdKE+wpJr=sQr3eAt;SO_w27{(sd1
z3(2DDICk465jq_ETUq73=i0;HuKUv-h)=z%C&rA8<h(6Sc~;e*aL-M7OYCHtbwzl^
zAz6yAlTN&`(AhR+Ask<Qd%G%UiIZ$W=(80!0?7x82FRll&Gb~7RgEmn47+4@oPs@F
z&8|RhtSNhv3x6V(KxVmbAMvKbH?XQIZ5hLrJlshs>gN%1Y5JBMeu-kvjyS99m3qH6
zvR1I4z!~sG6FNs-B0_BRaPs}+(eq3RM=M3MMYR^@qM*4eR^FwS>xwVLbryaUX5(M$
ztt6@1qJeG^%j`bS0iH+8^%wa8-C*z#xhHqn@M0vaNsFS6IMNfcA;%UtO_Sh{9hP9+
zPm%iE*iFE(Odi2HMs$RSz!Lio<*wAMj{`}G{V~ktB08A7Ujk`6lB>1N5HHheK$JYx
zOwMa-nM=*f)44tmKXgs|G>?iDtkd)!_19yS;A}m*u@vbwk3BjjoVCVewW&GpBD?7c
z`I{$`M88Y9V@m0qgA`!yDh7T^EK)@nb)0vr&9q03)V##^$hnlp*pKg#sK`9rlJG!@
zIwXAdg(ZIbfb5OkqQCud+TfpP7eYf3aBGCd9ANejvtx+(hk&SF#dfLjcuMcwv#U0@
zi^$o@iSe<krkrP0C0RIDG4Jbq2|&U7?`~1`ijQHQ%=g^;557JS>E7&UN0tlsb%Y*W
zaL5LWc}?x4lAV`N)S>WYzA(q0IYIhY{M$X%(=3n^*}(<XbR9?3FZKxHY(U{y2J!f4
zdXC-%tzpY|N`VabQ4$jsZUJlZMe-3kbb)l=F5?KvZ<~Qt_F>h<UUgF&{$^zD)B5h+
zfL65y=Koi!2)4RTP`?9&c)}&3=16q^dl+n^)6*~d*(ODXpbJzVOlRt2PcdC%Dy*Mq
zF2Tzxo@Q~h*ggRTR=V#~!eQw9J%GKui>FP{eTNL#E&TxGS&5^9PfFsKV)u51@<-2B
zyki)Fa^1(zP)@)e2Z0=1DxZ;=XYbWu`&7|a6%Kk$x)zAKCF3B{x>^~z?2!V&(COq-
zw-?5_-@OX(>9up_&|N`OtO@n!xld6Z5zTb*rtG1+0pzP{HpI$FNNM`?CiD(>fC#Lv
z)+|vQrU6CyeJZQGDxGb5f)dMtSGBPhAw(*U&8PKW%f+Nc)IAp!Y@4$xLGpt&i85E#
zv8B~S+*~qnTVaBplVXM{P!1?;dcJN$;RxnA+lbWcCc&UHLSxub6bNVFS3d&-SL3QS
zZ58PF;tO1V0YD;im`;>NaF(2LG~#iNG#2^sM*h+Scp)AnY}~~KKrX#bD8~5CKrWe(
zu~Ny^jS#<dex36}IzSFB5^&8w$Hfrv@A4htXwY58Wq-+(nAJ3%J$=V~?QNT``+2up
z=>5OI0*gZAAPvgbyZ=tI;7Q#i9lcc_=&VA|*=gs;$r;JO*1uM(wWN`|q>QHqxrjry
zex8=)`%Lo9g3_3;z9#zQ4gAWLcLMHq)uxK)%mAoJtan<I_inOr%O&NYIn_XuB;c|0
zR|5>{hLRULAcVcnx}GFAc)Hmq-UCm6&<D}St>&n|3`Ewbm1+03u&YN?W^%4MfPY(8
zlpvZ>V+X0Qb1f{A_mt|*T4`9$v_jI^M=X_%5X7ONt<v2{F7|O1C96h|13lSHn+N>2
zdI-d-_SZm00%J6_c3!+}JgPiD=0DYxyX!d{$~gjD*AYEWe(fq0Z;GiePz1sPon2lS
zE#zh1=S+zsCX+W5rrSD;Mc0RG0_Nf@uw5*Tm40=t8${siPaUWdN~v;6v6+VB6rh&y
zZtLu^247Ki?^hRajq`4d4_2foixbKkr=olvE&83aUJ|SYOsB}CfUqVMwPSX+qW{xU
z_(MfEicOs)jO9C9<Uixj2CbZ!a@F1cn1|{KrrOcjV6bJ+BZuj*+vgp2u0#yOHi*iL
zjoGi*2Ye3SDIlyR%{3JHfrhrdzv?*n8b>s<zB0R|yZ=$ejmC@Gw^1klM^y<>Bk$G9
zGB%W|jeMUymT~|qp0*Fm1VL2PnPdyi+M_uw3hFAJuON9w)Q63g^vBSpGX0q9WzM#$
zq8ci|`qGS!jDWrX?;uA6uR5QK_95mxLEbYt5qEq1xV)nM;)4Y4V-kTohBrRy;X;cw
z7K$Sa3M{s3Qdtdq_GB8O&?8P^GTX4Ov1g~TRmQxkKOS84>?~&!CRN#-6)}_LIR2VI
zRA+uU`Z=KQ>%W%;^nw6}<-})_p`mDqRs6I_SjE+OyA7=}h<u)SOrUlREiFHwh>9r9
zDqDk;Q6K-8xT$hf5_X|Z-jw{miBZR7nbXH&9j})pkW;Tc1IXl1YuvrJQ?=3A0kppt
z@r4&uc*XHOH=OgF3Q@C#CO=9)a|y$|Dxkoe?YK{ihe8_QHuKkE4EcGAaM!3OymMBv
zv%TK9fBNK!Qw)ZA%Z5(89WE5$Em?#u)sh(t|5Lk9cO}OXi|L##GH_?Kf9$TIiy`w}
z@yw3$!pRNHseZG^;-nP7H$cez-__lHT7OT{9-=<vOMg+s%~q!ebcN=LXS?6%!6Apo
zoK<#f)eT>;mN(snX3v2QT`=m1I9*(d?p+G^<u^xypK!N-Y+6?}5%4fHlf5>#2DuDK
z9J#Y-OtMxsHdZgZ)o#}^F{hk9(b7s8+9L1OUy^41EHiXvlpxNNbq;;U@4B4ArbURm
z+J`)DL!kkzB&$3kiqoB6iEF#*6}VV8dJ;Q1IUkDz&);Ae(b9C^o5isM3r0qc6=g6d
zTjnr7sS<`|`0&HAc~o8{a8KOpZU#}ccfD#>;dDksIA7IxZ111y&8)6RNg@O24LT4B
zQVex=XRPe_*C=dlzG!%i)DCj?lQ2YC=Elu4CxkacuZWez{l%PS!ZD(SO9yWIlH}>O
zO`GZWw(7spQZQBu7Lc4Ol(ig~zcV97612}@f#I&ME^6a2#bx1?R_s>O3~x~PHUnUy
zsST&Z1GM<A*Xti$^t~$S(Z!iv6q5A5>?O(qh669_X)}-JPk(3t$FzDflv9ag;eoLB
z>E&n(9kyuJ72UiGy!X$^EZgHo#Dgy3#)WSGjwwgd%i*KFUxL=qFeD8bAD-DW6y~ms
z7!F|(AsgDa;2_NPw>u{ufX-0JpbImn3E_GIJ~SPX-QOGJR%-gqir}9dYwS|%U%&4P
zv?Undc-{Df4UgVci)wXez45%)$oA6^oaXlZ{A=Iac_*~n)BVNSok^cWZ>dgcsQ*&y
zL3mP|`qcf9d3>>b;V{Pc<t=3IaqXMm^;!^>UJ&roAk5Ai+l*uRIaF>}==}g*3sa(w
z%ZD>h&d5<Kv!``zo3c<-eQ8Co)NUshLW5j9*H|gA%o+geZs=Hc^Wv^PoDWaxu;(rf
z7a-ho`Cq%8N9g~UPV3amgL@2fukwTg3HN@}>VXauR(l9q!2z^ss&$Ggo^C{*Dmt}+
z@-wt~{B++nS8zSOx3*l5t~^7<BsX<nWj?&`&r+vz!Y~irg?jY0^tGsbZZZ_nr3vUO
zfT0O7rT3$~ebxOdV1_$@XB!cxSd5^U+H+Ym!ab7u097Nae!#P2fZ=nGMsh*!EPQ2A
ziT`_YKH%XGdAa|q)kC)KKWwXq3Llc8AO&4gjH*bPLLH$nKo<)Y&)j;}G}Nw)H~%F#
zTS1y*SwoM&$JkJLT(SFm^>}l!rj6pWpJRYXhu%wh^KthSB}2b%%^7P%6TvW^Dzp;I
zWP@sLih=tK*xz*RH>_4x(M(CbfZ<w=_Tre^M<(`M&B$Jp$_)#&yHc)B#M`zz&+|7S
z2@R^XEs}>cX?!f`<bQm9gDf*-cc?$?NoY3<+%kp4_W2+$k7PG@@hU=rtbO!9B_-k!
zE$S#3U6OIn0h*29@QyC^-%uM8|0KNI=Ilag1dPV~SSN=q9Ab#mn=Dm}^P9BdjWc>Z
zq;<Zg->pKccK7+GaVCNlqB0^JOa2Gt$J7q^RFf<u$n6rnP*kU5`%Uf9qO6C+;|xf&
z@vzwm-8|yx=sNxD_B8vEXqL7_K=%AA`htj3j`(nOAXqomSje*(y|#PT4=e9LZW9Wu
z6Z~SBBZW`U)A8EIk6TNm2e1L^d+SrzrrS5exOoc=y1Q_+t6AFJaV)1Y)i<8n!#S5I
z(fkxo_@Z+udSC}LJ~b&*>iJWFB-MhI_}}i{UmluCvv>}bPd>}XE>Bq#t51*ejtybv
z(&}FcxH*SO(;mFGw{U$ifMlLC?Ej`xpY)O6#!~+a!g#{pAQFH_)<p=}o5k}ckN@Wu
zG!IjnZ`RD_M{8MQ;ay=2gXq&=Qurh3Kk*@8NFzo@kGf$c%iy5ZJE-<ijkxW*c(k;b
zI9l@1#I*PCxeEyS@Qax-;i+yy8TZ^tsyyuPn=!{9#G}E|L=+hAd4D`q&&ab=2BUu^
zs#<N^mQ<_uy6%k&Q(q}9_Lq*Ww7Y%=KfHSEbfNcPpQ06fERk0n)}%RyPTR0k$AK!$
z;v{RKXKKRpWH8zMD&XL{7!2=;j-*%V5%E)98R~g5Dn(0Mfm_G5*pSs$Q(Md7<0!7x
zI22GBYEtxa#SJ0Yh^v*u&^^H*L6SBJvHc=4SiO=*n?qqi6+GcYO+}uko~ad%K9G%+
zl;~?kG{QQI9pTGPS|wQ-ll3vUQU&T@`pF*mAvgPbVa;E;qZ3G^AxQ4@muZ6U^Fr1k
zN%*WnIUFwLt*&7X=vNR%Zn=V=e_}LuxbUTD;nt0mpBtTd#Dq(ibo5>JRF&N$rZ}d=
zyA_;s?%7od<ze2Qb{){YK-^A|vF~N0()R9n;!yJCe|jyB0{A027?Fm5*^()5-&swF
z#-&g-*c4q@*yiR(Q)d@c(=%||+KqkQeP?`fLi)9JBfj=$5z%FE3Tkdj-S%!GOO_e{
z?$ziiKDk=<i9vJ}P<svd30SpQR{A)NYBL1@e@5R301z22c<S6O6^7-zua&I<dD?uz
z1~(t7Xi_wX8E1thNcK76L{9Gol{m1!SIOi9Jt)XF&w(+J&WJF3N6Jgq0-Vk~naa5f
zI-2C(!LM3&Rx}hw`%$!n$F7aOz1}bRD7puO*Yk5NPgMl8|1$pu*1cA(pbT<1m-Nm1
z5X|I@-rn`Utvk~%wmoD$k3=G|x@82|^y(PAI5V5JCBB&5bu(sZ)VAy;m!W-wC^Xs)
z$-Pmd0L`vLt}=PuI9eK+tQ&8;yH=B`jTnM~ROZ8`RT^JKl|Bn>x{{1z@9pn9M@oOi
zf#o!`batw1G>uf<<nC#eQEnV=;$uPANr6e08XUrOuv-Q}RQC=6*Z#n#uJ~=peu0^t
z&S5r|$pv_ww>v=T$+Q(~SJ`k3Z9rkNfDDo2Z^Fl96umla(P#R&R|;%nBx~bE^1s)e
zGpnod>oBR!%{c3ZtG3o+M;Z5tTT{tH6mLx>t@EnxdAxR1$F2*<?ZTQhaAKia_y9C+
zVzfUL2#LyZ?|M^r2}q5x;Hfiq7_nk*bLDBEG}i)J9kQzB^5Ig#8Me_>wlSBIgt?Mj
z&0(<Ct&LhK-PnBlf=%bvq3y1D{;iAoJiMnVW?{kQm`j^vbN-wx|B>R7I1qG%I+o=8
z>v%0+)_A%;RfRSGAeJY|iI{KwEvCdq@!XDoa`b}S|A97t2xB`&{@;NGmaq1|Fhm-g
zYk|)GhvozwDu-dM$>e(|URt0dptHr6etdsOuT*`f2ma}l=6trnVRY!tBG;_Md%lvX
zm@>R4VJ6782N|8OWw(jr$>;?bM$<pVeY^bYPH$3whn0xb_BY1P9iK(f8swYdfy_JP
zr+nZ}z=D>uwjVD!*h>9Y|1w+!VYz|xOsf)jfK|o>?`ZI|iFbe}f3L2>eqzgnjTYH)
z^x)y6Mz9fL+e>A4ulm!jf7rHjFqsHlBe3nT+{LGvSJh0*+!HoZKYc8=RHT;r$CmHa
z3Jn+?OK#fK3z2pIt_(O2xpiRiKiIej?LhCo#c~$8P2h|Rb};Da<YgdC%|5uNxM1E;
z_6dy81a=JV35E`jx4b#~n<wUwtLo{G#xSEZs=ARWl2tfylAj63A}~A?q(a%&BE96p
zF@z=Pbwk-wX1TI#vZX1qkDk_w=iFMHtOT!tdpXSm3?&l_qtP0xgf?A2p2y+?J`3xI
z@!4@6({%hi^CzGkmO^tKv7-}*S%Q&yqrOOOl(R-hAdcWv&-ApTn6?YjSH)VUl4@MO
z>{}&imZ<7659wYwmDJl)3lPjZV(2`I@-CuTzL*W32wF;H=g-yV6Vg)J+zS1<eCAeI
z1U$acpSBGuTf3ZRd8d1XXXp1xbrCIuw+|ULEVYcFAx~Cu5s`=#Js2ZiA5^N?k<cB6
zQM8y~p5Pm&aFcjxZqRCXS8)}w^p!sRzIxeKeyn+=nz9a|d6TX9v0N*~sD&)>Sy#NH
zg7g$O)UbPkoQg&v?$SYt*KaoMn%+EE*WmM4FXq%YxEE9O9%X@?;muca|5w`l0mvDh
zJLKQxAP@FK|D}u5p4NX(MIYh|tKtz3XKVcJ9Ob0WpgguWa_}3($Jz5((e0)SY~@u*
z9DU4}NTkW)d%E_~SMVxy*Lx*my^5-%_JfQenROm%*uD1)RLyY^58zNFbB|9rNS0=w
z(j@&Jb#pCS7Qh4@ebxE|Te0r8%juf|T=9nM=QPmH`6CrW($gg<#?s23qt~{DdM{?2
zYY)!uIH7C9Ll4ZN)?VlXq1ua7B3D$KvF`?;W|x71pFhagiimdQ&DW-e7MtdCk4Lq}
z!_nraIQ|!Xrkh4(msDTw!%a(g@86E=>0z!GrOt;O-(JrLJIVZMr)-JPT-*Lwlt>fV
zT562pf0fEm=S#-oOitAGpz9vB+i#moP)^WEww}UYN!dcW5EI5=SY0J(D5l@R&HGAw
zuO$14;x3$Rm{Dd7_aQdI6x%QqCnuIhwVS5ODWdJV%CI3+WeP!P<ZeD7(za&0YB$2n
z>fB)vJ>YJ7P^#zB4}jy&@GXAO44!e;mYd+Ctx0aB9qIc7!wDNzceC<wd`C_Nn#nRc
zuOvM$ZxTk}iXz86<*riKaujrb3JSht^cE1PU$&EMgxM-nvcGkkqowAZ|2<|K%^vm0
zL!3V#`fwB)Rp81*l<UZluPN~WS+-$(2#;d93}otn5LYcQA=@irC!?7-(SLNOEJ^hb
z5u}{NI%lCT>5vCW1Yc0TKWF!wg<Bz?MKGu^0AW+A%cR_u-dSQVKJqux62a;i8a2+!
zXSzwry~M4B2AOdan6o(u!j<dgCd3gO&16>PK7?q0GoHTbnL2&>=BM+9>Z1O!g7m=o
z{7mI}yDWbIA^yMl0uP>dU^ge^`@cmZPgvzkn;9z)Xjx&@abdUyywUW9^hVV81bn{P
zvCwOYbQ3Ds2?|Ebj8F4TF}Uly*SQe9^fVMr0!d|+4TJ3aZ_=Mxh^8r@s#30WporY`
z_+iwNRQflJN{ZC!7P;>SnwuzQGt}%~-Ygl|L2rr;{KI!Kjw=D1-rYv~`H<|m@#`w3
zZjP-_OQx}Yueof!DLZ8=Mvuhj-w<j>e2Ac=-fqtTb^>&gdN=?Z7zdIUg!soF@PVQ<
zZVz3%syZ5B$3s4)VDem?*$@mF)s)dSHZW24``0Sd-A6;FHNCQMeTx!p-peY6MMStk
z5fera7mh=K=ly@4*?O39liJdl&f1-PuuE&O*v9wdwZ(<fGJJ7^?56S(4WdaC7FQX%
z%e@?S{mj>%&)=`bLknUl(2rH!eT3`{e7=Y|+5Iw;aVAt!+$P?$DtOsqg+p`ZjT(fV
z20+cuCb+ZqPzw3Rv=>oTVbNLLim85)&n^B6<~eF!>UPH$d4vvi;5S}IDBel-C-d@k
z0H$}e6_p2)%6sLIOu}!P%NdO|$?cCO(^4BV&3H9d6D~5Et9`#vyi1U@*84(m*J(2?
zP~RAU!R5E1Zg7!OG_|fcJb)LyH@o}&7CCgvq@sgCDMf5ir07JJt+eJMh246@nIw(D
zrlDpzG($ERlR8Ps>?1}3QNp(rTIy8KjoYkQvabui+-X>yku6=LN3v-*qNKB;^D4|c
z5jrNn4AnO1*VSd2ckxVNOT23R)yFGpQE!sRR9|=F%JlV41LYHn=VO|?s*9_5BiV-F
z!LIVBdxE=W3VW=#KwbbACHkq@QxUI3+RXa!ozga*$b)%~FPmFH@;?_Vs2Ah&VE`YU
z(h4*^+Rj+rn%+TvTt6llct8#8L?U1Mvu7wTL)>g=m^BRb{^sVbpF38pava)`qz0sG
zsEbu@NqBy;hQ=Qt9(oM>3d%Udmq<k(_s7@7oB)Q+-Uzw+<pz_($L3<GR~&?N%(bps
z!GDiUPkmwf5Qzll42zQ$CYY558I)H0hy&g=`~<kIJbI{J)ld$TRt_Htfa-=Hj=XD~
zKM~Jb<;(mQu2uT^q*BvV%#%3i{x}S4imfkKcrQO^YL{ahs04(C0h<v7I}h=nm$@p9
zm?K6<%}P{aVnztc97GDn>h(8|uKzm(f&oY(FXMAfn?!MA7V-<5j}%?*bwnbMApha5
zx0WAw+rWBj0HL5GxeRbIojJsfVxG(!m*RrFz4S_l^C+%-+><0$3e8U{Gl3REH-^CY
zT{;2O(&Xbu8Lu1f^`(`SXsF$TLatu51@dIs=jnSL`C&GE6qC?lkTlXPs|)~2-@4bG
zqTYzjl<H7x?mfgGBX0O&sw5Jr3&3{tyY_#EOutwFZ@>b#lKdS$EZx8-uh<A$=TiPm
zH;$;H#fF3APrRe1N_3-#%KkR*j`mzqSPJfdU;<JHWO_^fCA(G=p9At+hxOR5kp908
zH|86l<PmGG!Y?5`rRTVMNX7jfbUb~CC{*L&DMlmprL%eY`5iw!l`C3qf!i0MQlxnf
zSPb5|hS!fmB|=wrRngCK=B`_;DG5?Fk`gz0gI26=WS6-rY$~0u22~3#uL*j@5d&Wp
zrDb(tKE&}u|G7`vUTlrJr^+J5%vx`tQZpW@A+c0i$5_apVAe@0RkH;zQlsq3bYX=;
z`Zz=l!)Q7Z_hIa~9K=K!p9wss0lBxFL0_-4gp7d4;Q-t_5xyV7|H^wuI#0W9rj#;!
z=m-L9!F1Q!(b@KGCPlb|rON|qmw0>YI3RHDSAO1FxSmF8jsB{?lXmg~)_jw3ZT)$O
zSHtX~@xy!oL6}m-cT|>&8jm#P{PNp5Lp%<t_vjN+5EVc+3sgbUZQn-hGmWlv#1dZu
z(646pHM>PmrencksV}E3)cp*pDzs^>@*Ccw>Rxnyc$r{(rpO>w9Coeu=D#$;G7?T4
zhOPBg&)2ZNpTG)jE<Py*KI9EynhDmY@ZhBHCH;Y2`u6j&n_=!;8aCO?*6WF#L>f9g
z=o<F}V=7x7RGM0rAtxI{kvHK*#?bMd(tw3H`*{iQb}0!F_!WNEqbHsa9&FidnEm<v
zMR&087wP@{@>>w>e=?NDdhbWPPrf0&<h3ob*1B&G@IH<k(HpNIe-ZE-IAUhyXn5Ps
z7@Z*nn$1i6P*hXM7voI2Xa$49`>%&gCK&`h&cI#i#fhrPBD<I%Z_NbQKxN}x7HJd<
zpC3NCe_Qxx?B-9lzK-<H4Km?2R^Skd6Dkf7Yw=0Yi>U_gCwt|{!6z$XH?l=zyebks
z<`tmT_^N=tKWhCwa~wqXe_p*{6nS(f*nMdWaKZBExfg^Rz9VVPB9JWetoWi-RPd;5
zB%Rtg*Dk}zeQ7j>JJmz8lUoSaZyY<EqyS=js-zhN)ILI4l>^`!qWXXx3#ghs8AX2&
zIjZRGStk2?^RHi9)DaC*S(?eLfrpa|hi-+SL)}awncn`+sRDT0OnB1o6M59d#YB3C
z+Mi<yG$vBSGjCLN`Gtv5h{Lh=FmyPce^0QGGsR>><{#Ki%6VXk!*M2Ka&EVi8pgbi
zK8{L0&44`b-ha~bL9ihfn`Qg)*9Ra}>WaLu@HK1qm)!`<5VCuyo11}KEIW4!I*Y7L
z+;vtX(-F~bK4X05BP`{$((B)ywkOLwUxWXWE|NcjA?|-&kX|4>8KLj^Xa7-p60Dc_
z&$F6-f<gY8@sd?|JZO%?XT6h4%bl1WU@sPjwYD$o_4WpM27i)ApG}&^JV>hWs;b`R
zU<2^c$*9w%go0J^{7A%vY5`k~J2C?!JeZ&^U_kDXSFW!Sj!~;2sob<u4j0`VUh+kA
zxnQIr{P&O^gk9I`nVSiY$NXHzox%>Nz8zgOqUtlWOMlW*De>Ao;?h~FYN^^>SMk>@
zzpb2~k7<=d1YOAnls?;aqy?;)Cw)n_$Xz{3QCX14&+&jInmxP=p#tU;Vv9QqH(M@w
zayIecQ{g9j>wM@_iNPJ$-sWl<i4+zV$se?vB0OTc5`8H1k5EZX&rf&vsxN1Frqj-n
z3cmXc`V6S@*8mLfTtTZP&!SDEv`clmdr26ap!^HYH*T|TO_6ZMd7kZ>b3fYGXQO*$
z=lexHmGdS$#_c`xbQfs^hmR@fry)P84qc%boU$tM9n4{ZSvnaEgDO)`nEFJiMsR{N
zinGdHINr>U9!JNzL+$98O97$5Y+J&sQBG}m8{teZ)S8R$pZa_Mm}z@Id}S&vyY1i?
z(CVT~4^cnCd>lK|i@5NCQZ}G3$7gkmFajE`ia$-DRG^m~ZTX?GbVwLx9N{raIq?V+
zNfl4FYN$JkiR7>%9>Z7j0yEP5zmGFZ2jnq?H4fJ;kM+&1cuzbI15>@m*Eb}yG&dTC
zpPg!6K^}tpgw6UKsI9L-W)upo>(MPO$iZ@WEDLF#Iu}?4pwq)g-3c9eO?E8thIKnV
znw8{1%R}UdPx`4b@F3sHmasN)z8pCf!Cbz1%HmJ#nXw_oYA?#nhXOwN6{sl_vebsA
z3Bj@Pd;^(}LF@=}R)q?OkOWhd1T>{y#ZJ@Z5wgpen@weTf^5r<r2+<x?pF`JS<-pD
zu=Ik6GOgC+o6qr@5`Kju;to@ID?^R<_s~_==>TUCRXGR4zasZ)llE0k+-ny1lP@x4
zQzFFC%F?QDag3Ml?^5aYJ>l^}%MJvhc|s~Zy}lmTy5TQ^=D2oC*m^KK!yO#u@RjaQ
zkGlUy!-X0V_SzKh#4~H<(dN7j{@F_}60tA&C8hEGjm^tEFbi_8?m)#Gpkg=FH&~D5
zqy}%1-IKhFO5c9&U}<1FR_mLu#;@zL+E?8-Q!)`3R>6Jsk7N2Q9tMlFLJFWcLy=OD
zEABOp7ezjAqvEzY=!*2W`|MCXgTj8XX?LE7h^0Ru5CxdNy$<`8;ha@)^vBfK&8Q7V
zOhvYIt3Yzo$@*);opq$I1T$tC^9rF9J=3ZkYWDu$f~v=!l6L0~yt;Y3g*~HxbpAH?
z5sZz&y*%jOd@j0#s$xB%{fvVb!zbS&m6zR`x9gg!`kNYPAJA_UCvqpygV)t{x@5&!
zxZt4U{JSPc?_FQS^wnM_Xt0kFf!dgp@QF7TKKi#^mMkolbcv`uq5|C@%q(nS;bP#m
zEC2x3WVvniFcWng@<pk4r=*rJ&y%@GnC(R&?ITniPR1ujZzg%VcX=NK!&W%W$Ch8R
zE6x3?-)?)RD~{6r@`AqY@pjXL-?d(Y9~bc>F8?PS8$aLA87B;6dI@r&+?2RW+B8l{
zep??gZn0Xiz6lLb;{$tI(hbxogvQiF*|E~U`S1DJz!}*88Xe8VXm#<-!L7r<<RP(Y
zHG0s1B9i;~#;Q4MM%_o$5YxqpouFXnTjw12Ouz6L5fZslNyke4+|k$0^$SxdNLF<p
zK@Fi@n#|LHvpGgfE<qy&iSZPUEs8QDW?U0_O0)6tq`u_U+J^W~ZG!H#2LKeG_J~0)
z7mk^+dSzGX3+e9TKpmVrj-n?C%^j5<mEPECwwRS0M(p&gZixsjGULWJ@;xF3si4K|
zE7GW8Cp261dx;ivrEF;SoD1vPFp1Vq!6o0{(*_Xc2e@eMh<Ta+pN2uW7;tIC`R`hP
zlPLDc*!dx8f>7to+lc>yro-oRR&Gt3*D(oTuDNTn!@zMc=r(&+<YgO^TBGCpzUYVz
z%@ALaZ6XVARYM+*nZ7+}t7`vg`@-GE4+qp^_G+=i@3QG$c`%oNrxZiGa9nSxWdf{A
z^NJb(Y3PH7jw47xc3Lcciv?GOLxc~))+g%bMhbW3JKxRw4xBtw{_=a>%r5|I4SW0_
zbs_;}{n)dNB4C*WhDMrbdD-^_lCJk{ubnxY+1Rb`hfgR<m1u4*&%y))g}SC(_z+l#
z>UVQt&TCvuml4zIzBkkrEl#+e-b8VaZ~))|*BpLE@^h;b)-OM$y|9Buit)`$1T~hG
z;-%cp!f0~?h}WFj^`OF+U+_puDgV;8=KItw@wf$qjgBrdvX{T0ZJ(ak<C(*|?~H{r
zW(F1NSgiUaS<_<@Sf4d<m0`(6i<#!5mH~QU(ChdVW6Fu;OC~r^8yPQAwnGn}dX>-f
zG8s7)t%NhjH}Zy@L^t^HaL~b0gU3SYWkd=g>Q9BdNysBVf7H&?Fk5~-8Ypw@+XZ4o
zs~*E-Sf6_Zh)UDL*~5`h_<ja`TPxrm+VgNJG;HZ`@ph28rn>t}##+;>S}8t;T5a(+
zR32j3s=XIyEFDfyp+XEYar3SAma9c=HmF*UI$iHt5*Dl*20{HUu3KA+7OeFsDg}qM
z*HT%-*By8y@MRy^tv#d5^ciGA!wSR`bVHQ~OryK4H`yZaG6?4yty5`hpF&aUT2pI-
zs|$%gaefkRZ9lSb{4#1c&0Vs`KUVU}D^sA3Wr9J8=<?H-JR2&LUZn<kRq3%2i|S9}
z(wr%C>Ttxd+mf2wNTk}bqZGzCmmDL+jCvZ|sQ~x}%FQY*+^vf%p7o!COsf^R`3bZO
zLaEHs*miX~SEK#Q%2e5jsZCATvj+2@w}zeX=6tMSs~-XX1fi%fA}~<;0GpmN1U1a4
zR5!pP8uVCGy+agd)zUF*)VjSnG}-2SrFn7!Rh7gE)VBuEt;dgQwZAev^17wF&U!T5
z*twnZK1m5CRi?Kak(hNe-|=@lfLYL_Z-K>6kM{SETPx{L02!hDmyd5d7scLQ)q)$8
z`J{k^`wlc<$UDyEos~G@%8n>E!04ZPCM80V;~UD;M=q^n{aldY>*wX7qe@pL8@P1n
zEJQ;<c(R$1ARjlqK?G4I@O{z&bvo!`YP1-{&?6fwpMI`#!bP*I@w_PJyVah0BcKJB
z`6YiD41`~;wtvipJ10M(9||?Lzopb;67=#ECK@Ebc7Gu`ZwO0I{{Dg8T1jaxd!hi(
z%dymt2K|aT&cpDv7|ZbHMdz%w617qZ9JwppGf*4RW3b)ORY7J6DE)WtNo<u#cu{`x
z^-+mwdztwsUU{BudjWUOfRC2HFHF}*A#z71co#W*fDv7Tgq&xx^2hVwe+@I}N4zT|
zb41A8^Ch@D+u)&Ju&e2VWiS#a7WZ_nhI;?@Fc#?!Lw*zH5Dx~k`X(o)MwMI&4i;Il
z$B&Nk{ybX;@d;Kb<BOYbo>G@=eJXjs)nrh{A$=VJ1xBC0?4$4G-hrb^QeumIjJ)kx
z#`x2dOaSG&;XKv2(u03CDFO?B!}Xe+vyXvZ#oehmx6F781`F4ZzaGC$g9Ge`)0k|t
zG9*GZp%xZXYnkmg!cRH;3ucXU;P73w1cQ*>FJ!@{5X2>7pjo32B+uiMfq}-sNv?@Z
z`5(-8_W?Z$@d&UVeyeI{GIF3~u=#bc^6=ap&P_~yYtB2W`B8a(Ry+mnZWXNNInG0k
zQU1yg`HZ1axu<?tSu+A3f^9cM7Zzwn*b=i~*&z<;Y{INGw7sw9($p|6?n{u&syU5Q
z2h6Yg3|1FMwW_!Q50}EFFUx=kF7pKAFQeIw?Ywd^`~%YP8r-23{+kLiyu4B<(cUK&
zU-U5~mwA(G05Ke`+Ut0$F!KG<CF~8sC~S6mdt&d*e|$O(b+h9W{n+4wEQZ>TKg1Oq
zlww=%`}^QMZmd<uCW+;*b_3(0`N5Wqx5Kv=mczlh&okZOT-C;?J-VBX7u|vAbA*d^
z;pniJo1n^|vACKfvo@Q`#cXSjH)rqb`^Kh5pNt>qXme9j4)VcCOp@p!Kr6)NzxIL?
zr2#uq4r%$Wbk`9c=-+z`!}&zT#MZDxcPcM?SRgi`eJ?b1(mu7*T+l%*=ybpK(UtMF
zBI{^+0&-J%`f5W`(2qBjkHOx>)0{sD(J&uBH0)ZF`}M1HX>-1c631F=tEWo8wb#Hp
zZT!d?(<e-1YZLNR&^cV`f>p!}VGJ%};ygg;eK+t&MX+6qdDcfTMr4S9IkIeEFh)8r
z_Tzi*4tgP|rC1~!S#m){QPA-C6w454e@0!Q-GmCJ>A4vPG0JSh)z+|GFTP|$212Z{
z9hM}U)f^OZKo(B{$w(iz3e?v=3uO_8`IEiwGI_6)Vu@<2-B0*oP2ekfYdF{1hnW0y
z57cw4yJk28u(Z+ZgU8_=biHPtZ%t{H3#~wt7)9Wb+@#nTj#lq)BKLHg@rJ1;+QpU>
zf-xMjB$^QKmjgz&oxR<YiLE_V>iS%Y+4qd$49&8YNS2KYk*VLsr=$XYXX>rb%?k<0
zs39Uh&N2dwv(RX)mxzjs@P@rf_0E3(>E-r0*m0eU{<SsA&iK8kA(b@X<=f-_bbvfM
zDXDB5e#wDSe(_Eb9uon!FbZ?<msq{}EK~~ktgh)+GnjlK?k!*iYX`;r_qETCDBP=i
zHR=E^wF*^*)i|5VW`T;vPTPL165ZyD$W#q4KB4ETQ8MB0qH7p=UvMTEr=oTHtOodW
z@C!B{l)VfY6sy2Z7YR7(*?QC<IdEqdj_xKb{3>jok-VCYIMKA41(AYD%r{v2&<5>9
zDH3SwaT)q+_(uqYD>fUyXe7Sl(zqVMV+_v*Gk3dAdm&TWQtoRgBdp+xM2(8ljqK!J
zP6@XEEE@H*F)Y=_75%GC8a0ZM`9)R@S*F5AfK;jFxIMjQu6_D;w^w_~B~@$k+x6QM
z==yT>1;2Zr>&oFh@VDPqH#+cl7MW|CV^$EC$cM$}HYA?SDeyVWS&{a?R}03rv(uz1
zYZ0`zwR98_@_u$+SzcE<&3$QVbWO0iN!wE~<9+9qE#s1=LVdjVzA3}V5JhmiiO?@!
zl*Yqdq|ox??)$lXq{VOmI+5U#Kk|m_eN|I^=TZD06X45B|G&tmkf=S#UKudCujEdg
zTVFJEuy;i2;4=?Y14K8a{J@E8-GBK>eo4MDf=nnMlV51mg(#h+2Ghndd7&4IgNYDx
zG7`cY8~^FT^+QmE{dE?rq?oSKhz9KTV?0S1gvH7{f_NxU(0R40r__*s2-amN%Xc2e
zR8g%(mALfv*MX@7zU2<#nB$*wpPV>T8s6IA3#>BZ;2Y|2HU#x<QhYt~AOFR;W{A(U
zyJLO05v<w5xoQo6KwazV-u0=-EA`8nQh?&ZkGWOi1`!dxr+Q#^JTqL<EO$7H-~2<u
zxU=xSDWRw0l%g5{v@j=Gnaapb==_RPmCb4+e=WvhpAPI`Y&l47pk1?z2RXE=78$l0
zANcY|qXQvv_rExvzJbbl<kkGvhC6zEotfUEKE0dy{h;RJx|seE50UsUn3taM+wL|s
z*g<0sujh@%@8)TvWJ+-uE+20?&wBw>5{e#*8*5~TBkiTKovY~f`L}ddT`yg%VY*8r
zR#f8<jkRrKN(QQWttOCqGi|K7#HV?|#W2powOgUk3Wy<+aTB7FX#Zx@Tp328DV!UF
z8(Dz_Z(P5(YkpLJY6>WO-1Kc$t!!+F)oEogwtHzcFNnze#;tB&=OQPlYT)UP$3|`C
zyrJrhQ7JeV$sUQjQ@;!MIkM=wFK7l)3{3*}`nM9`8e5S_0tY^baq4<S8XM67!4fAm
zJ=LV?y#))`5$Y9O5b74Qrk%!w+IHhQg>g!1s(B7uONB^iE1M*@f!`S59$#gdeU5Q{
z>9(FV)L!4N*f}8Qbx-73sO#CeYuonU`vVaE{j-4wv4)29B7~xM|3{@9vn}UJMdKT<
zkmvQcl>o@d!H;HOLscy0ub)&cJ#c^CCuk)uCkogDXG$<4h9Rc-ZiQ+i>wkQ|U#^R%
zT|0){ncgD_0||FCT1x^Ovil5dW`z5%2)*DJJI6>QS_o`^zVNqK8>4wNH~x!uB2xse
zrSmd}(B(+SvH(;7Tzj6QEqaV|+Uhq&4?Umu=9;}D;Kv+qjbbh7C*eT0sO+(xXiHaw
zfNirKZKr6&?kmoi!zel8#(A)R9~(^f4_L2Rk6s_AuVe3q{NB+-Z<3f=V>-%h7?aZw
zN)gWTpN}IeF<9`B4T<3&T)=MA?@$`=npjfNNTM&1L?G^V)uLMMjI!n)cL2$2w!LEM
zhRjTu7-+bC3h&JSKr?yumKN&SOYN-_^wGh?wPU&-jJdXI>&$sKW&1zH2Phc5+^#l-
zy|=JsG+425JG8$x!5z~+R1Rbf{|m$NG2`^)Ui%8uLIbpXSP_T6aN3}?2~inRcJ+4I
z2l8!1FDT$qz6&af4zkMrav^JcPan&f)68YODzHUz3_i^(9u+0KF{$e3n3|pCU<A<Y
z)VWa5vO<k;>$PwCuP$R&`PXDNo^RrrY4vVMNP6U1@<i$b$xZC9`)nk=zI>A5trz`}
zd@b+A+lX=j*N-PI-7}5EI8$gj)m%AGWNqoeld9384pv;S;2<F;Jpv<mfBlRW`q^SH
zuqtc|pvm*v9T$PejleGq)TF1%^R)QV=#fDk+KWP9@8deIr>w~eYj2nNhK(q&%FUo@
zuYa6hAL^eYyBZzyo7xxuH!#$2LL6c&(93%egX1f6V|Y2n5=y0+XuP&<SI>S!%U|&R
z)vO=ax_6u$;r>^Vs!8)=yXX8>hD7HK`vxB%reQgwoKsl+2j@jY05#2TsED>rJhx4M
zjuWGYEkQ^6Q*I8XltABom~m~BitNi7>kD-MM9=%RKXIF>CGcMjbNBwNs?X~)P&P*c
zQhI3k@vP4;pR<S7ALGTZH~Z6n2X-8UPzceb^2}d1!G}rq@B%4K!R)NXY+8t@l~mX}
z1r437_;$+JW-Pv@BQpcG#dYg4A<x{{!I*Aac$MIHW_7svk>k$?Zc*lOzArG(?P#qt
z{u`oGt$TTLZnmTPBJV-tSbPHv6fbo)!#BgJI}zL&$8KnqMsNI4SMO1sx7py#gQW4!
z_?WgoM7O=}S@S)EkU?H=7NSBud}#t$EL!Kby?U6phC$wKE44-bfrE8gK|D2cgSU{{
zl_4Bu%rmB8d!256WX(PS=1DH6WruAY6$k$VHm9w|zu7PQPB(`PyXuBF)saJ-)4w-a
zdw~vq*t{)=E&0KwCAf+YR|kO**H2q?-X`$|gARNTv7VZtVTmy^Yj&BQY;n2%kmI{Y
z@Lt4v66;R|+K%{P(&JvbA$+tQm3ofpu({y)dTOE(U~p^$tPlW3>_P+K2N2x(@plMF
z!4;>Fbl(X_3*yi9l@CmF4*iNOk^QMbH9n|h#7x|4Yh6axBbR0i3sYt2bLVe|fWvV>
zk1U>^p0%sbs&O?!A7l{zXQ~E1v?rz!Zu`1AA8{ji!>XK{>(#6owZM<BKr`pIc<Mcf
zCz-8Et+{7s@92d~5Bf@`dF?)RU%5DDh{*Xj@4>VyO3X}9KE#kEPCIix`Kc*gF+Nqv
zi0qh+?2!&r`~#U!i`K5d$bfoE;IlP2Vd!;6E5F3??Q!_}@0^mI8J{p8Zf$zsf*j>y
zAB?$#G$fkm!on6tV#=#S`JwBjpthza>AsFyH{$T5Dcf82{FIERRkdqb_Ed40v76g%
zLEzNQ3G#dOxmRG;F^1O)mc`Bt2WJ)Q5>x0-)BoOR<riE!8>Pz-UN{wOQcx(4?8o@!
zSqy&%2M0!`7Q(NQ^o*znWfXtwWRVKs^6HRc?HMO=?o(NP>&<sqHGf_F0yaty#!|sD
z%+?Q5=y8doAGE*U9gb$b?EM~1!GvP+r)kS~xEfzgF@=D%E{uY)$HaE1U={X8Q6PvZ
zwty9m0uzR~It&lN+WD4<Bfpe;yS}fMQoowGYO&JBhkD_F>g7e`RY>UNto{BlN8$Ki
z8}Ri05oxo9Cyv+gk_Mc5rg3Tb5g^OqWYlJa$Ks^bQ3AZ$c?#sApZfP(+k`kypsGig
zCU85ZtMBNZNN;aG#doDH);obz@R5l`_?e1JSD<3X>d{WuOs`nPiFFxKJ!UE4+`s!>
zN!7UU!Hl~ojh|i{!P9E5J#|(P>~YSWi%tVoz#yb#wd9C9f`V*1{4OUa$}h;ayF}}+
zqyzTAA0arsi_a<p&VrP58p;jQN}-l-C%PeR>^nQF@9_(3H{Rty(gyNSpGwJrj`B-S
zA6tDEq<+A}OfTme*Ml99pW1ubYtB(qPL)PeLG*VnFI^<l(%BbZwh?=N=MKx9<OE3|
z6)h`lcI!ulm<i$Nv`-#QE&ZH^RU;nCSmGOjV>d7x<1-<S!5nIERZ_O_9eVBWp6<28
zx^kM=(xs8kFB?r6os}(OjCF_>Fs9_(+Qq9H-Ct!ApI6<os2e*MF8KZjr%hJEmqJTo
ze(0&x3)R+>rEcwGjWMc7fc9Cn7WenjdVB)X*ZnJKwe`Z~_q+^6Z=sVP$OHwMFB_lC
z7Co~y?c(tI3Qp^|2$}CC2?tK#6U4nxxP#M)gySnV$R1&0wqjBZ@D-Qx0P4n1-6@jE
ziGSPVQDgB^8-z3Z=g)7S>(|Y0(uB*4f`mI>#0Y=D8+N%p2is%N9-3g0IU%|5b;+w~
z+T`L<<6w>~x387h(EDA*2gq-{UZ7XB&&klP&$<2ZPV(-6d;H~?ql86tOr2^UT_EQ%
zHu)oZcPmvFi<s_w2+q>Regz|>9=#tudb4G0!LC`YX1%4LvY4GqK%k<lGx;daV*7+y
zGuGZAj2w%@+r*rps-E>7Tj&0B&NA1;RH7&)A%qrHkGf%!@*z#moM?QrY?^$IEWEhe
zYAaA({S%71cATBBszIG4XSg6ra;%Zl?PAz7>kz6`|F1#uq*-N_zHQf*Ph=lK_M+2<
zaYJLwh<w66$byBdjJwyvNSv__>6IQtRH+WoI)&MH(MamKLgUzhG3DbU+gBI=Y#jR(
zo0(pX8uf2dswBqWi3Y~%ddY7({L<rVaJx+EqPChCdQ`K1GSA+<J(_x7vM7<forW&_
zJCSVfzxU?v+`%6IgqG5tv<dK;5|Thv6RU9K(J#Y$Ixh2RX0x}7Htx5+nxL(a2<x}a
z->dkGNA{!E#wV9297XT-+z3cnlTVs5e87Z92gp3AXcqp45qrIpJntKsbIw1V)xKn3
z<&}+dMI-0Ru(z1G&!bm--p9WgLqowHJwfqrNcC=SqV9^##hJp^Gpnzmb_u}m3B`nM
zjFq4T+Dv2-wX^Hn=f_a=I0Sl}SCfDjzB>drZ;dBBb)CTB<~IuTn{=(?-nvGZuRy<(
z$%y_`P>4rwf41{l+>`EXpXpi6s}cGct%ZHEJ+8Z}{>`6nghQwInoEIPQ8-3|uGJrh
z*pfID2bWM+>GwxRtEo))@DFABFG!MS{_M!dZlS=hv(~cJhO5Ln3!e)&K~CWw;(P<o
z7L41!64d&~I|OH6BKq<ex5>;xi&y{mo8YZ|9#iN;C6K_?u>M^|=-h57YxKH@et6vT
zOn*v3T3R5io|8Kul~@B2|1&o{O;(|ta8efkiHtke;4}R<UkpE_cg7Q!f5@r^Sj+3G
z3BxMDcEy@S#wEz62t(8;B={iiRvm#-TH!Qbvm3T`%?!)D>rU{B)6!{KV^Ol(^EQ>j
zBhnoFTwhza=!wP&_3|<#wZPM-D%bn*_-(gG+UV`96-;`5e^1xN>z&tOH_c1R2lz5r
zM2wAM1nqID!6_i|V5;zOWKH|})2;SxnRoac0yLk5T^!u_D_ortTB<Z0kt2Jnz;@mO
z!uOHC-Z5(`56Wlo$|3jTJ%3YPN(rtrh}KX;z^bOpW?+dG#Q#C}$x`17L;Ot`-i`=n
zCd~4%&Vcfv4wZ6@$+8G_SmI-TnQloIgPjpyuB<3u%(5^8Nc$m6r7)_o<5!V7gO86y
zK6a^hV4h902|wEpoqH7nNj;1nrt)9E?N{iU#j-KF)O&UACU}81C@>6%yYXajtQN@5
zUYT7-I2~*a8smv%aI=_}L#XrMN0=F5C%xq~m#KDVoLXIoRJE+g^Qnw?p)G$Xvw741
z@0f&tJ^2~comqqG!aY8W15WJ$n)sCW5?Jd=c9HeA9(Ux;9%!?#A@Q@lTRZbMr_yLO
zJ_jGCls;O56%oD0FZI|C>H06p*G!gH`#J}>BC_w@de+I8HaH{9qwy{<eyByUdrq+D
zw2GD{AQK3c1eu-gH2M?D8O0#Ft<D}~bOq4!^`!mKZN6SOh*#~>Rm^FQ&Ergz@2S^G
zdKS4c<KBIBJtUNeHES^aqqnOfxIUDbE#YtQ+vk=U49{APYUBt%FL8E<yH0Rcw()pq
z+@5m>#~If5U>mZ!qKdZj>add0eXiB8V+_vQQ8cUu$3=nRybgX$FU5OsxFl`UA#Yzw
ze;71AiQ5s&z!%exKZF?ke)G+uk?`ZWKK(9V0TPj0Z!`sbGR5MtjAon3K1VjyI>8@g
zR*pUWu58-(@r*Y<HfnwsnO-HF9L4)YNB?;rqEGjI26bX*1GZROG1b`qUbmkGe~o1A
z^#?Thh)9ee&6vO#I*c^0OFqs8JQx)N{O?xP&zolFqbE!GZ{z~(Sf}c`5PhDmD0V_3
zdj}hN9wk_8_X&z3TgT?$Ssc+v7;<09Z&54;RV(xpu{0XD5W}*Y@E)twP4*O{<Y9fO
zmi-e>?@Bv(50GKz+IXaCS$@vVxK2H1Rr>m#^+AFu;^HaY4Csrku73QCznc?*;=|a5
zUhj7}bI|6-i|`JwcX{JN$zUc+h5Hy%d&;Vb``=0G@aiFzU`U{@k~GL74PpT4eAI~U
z254W5ca47+)OYT1VA<4!x29FhA4iWr=v8kMMhU}4v!c~=k>Tudv@%Z{Xdz*VKzBid
z@M}!Q`y1{h%~BKq11n~G%W3P`+kBi9DJbwl$S9DEHVDiPp&pY1rj+?N9rJy(=i#<U
z<ERGJAN^t54K|aE-EMn0)_v%Tnt>N6b(0hye#Y5_MtSUKUWPd~#bXAehqE_2%nk7)
zg}sl<6vzKiF&UI{81>iKox)2Xo^pEhTHM~mrXLSZu0?PlVF*`9zge<hk1kJ%c{aJ{
z#cBt%X=mpT*eZ+g>|7__;0t`}hI}BSJVTG$=UPMFW$+=ZbHIhRpq(E^75x3a^!!bV
zHKXu=LvqcjX=*o6JiRsK_k_^7-=e8O3E)D!zEoaKpLLoN$Mz~wiM0L)`alK0z&%5L
zWs7{s<2}JIX#km$yh&*?_ZMp6${qh4_BsG>!GXLUcj-y0+sQfRsMwWZD7pPbw=dbB
z!AB+_B5R{}Y^R2F7W9ROhr4O7Q%}2%YT9Ynq|f!!%acKRe$-8u(iy7Tv0eFgxomxt
z$!Y8W0?kUUJ1%ra=GT$)92VWca(l=Fw66khllGScaN$ERd@a|{htwDRlKQPTg|~XY
zXbOdM4RcuJ>8dt%KYd#C$$S;)`!g)&WjFNrmJMe)d?O)|L>&J|oE$EHjA3kx=-q$L
znOG4=Y{bNG!Y(|~oiQTDg2s%KG2W+VU&A=QgYiP2Q8r~OMOKn=-WV4KvXRFh31h;(
z^L5^g2eK;^lx3QMMfYl$l|^{2xQn|n%=oc}HT<GLE&Em8k~OU1X9Vu3!)c%aUu(pQ
z(0s6){`r6LKTV%}^HnOdx@U{kUKo({VbKQ#9A2d(rODg<N%loSV)jM%tg8>s)CLN+
z+UpPEEGgnoN8>b*5@ci9CI%u_@}#uWY1ufzs-m-Q7o9!T_1tRIQ?-Kfv|)qHpx;d=
zM@OmKJx=HP7_u5@c-BwH&tIma=cj2hm?$Q>V(ih!r<e`GN~|lO!jv1e^!V$~)1Ut7
zAJXpLuJ<WEI)0f>Pmj|?xP6M5$}3{!(kGdZv<gmn+0GH84~7|5PGR_B#n%Q#C-H%g
z^hg)RoJysh>g`Tyv|FiCYj}klZdQA-eImTEF%iQMtGtSH&M-Xj!6tA2@u5~>IJ*ml
zbOLQ6s8VM2)P@wxi&U24>Mag!Koj2>u)t!bBJ>Ac8xP5c6?xQWMRi2IP)3YY;P*-`
z;m!hsje2@!mJh}%%Ey+9UTLQEAji-Le=u}noWjru4={F>wo6`FhGC6zVfeBEkxRk>
zMn;5$dabT-R&ucoV+bsii<NW3VP7Q1_K$3_h>(Hc1P<yK_psMbM=y>g`1BQTJJssd
z)M&L*qto;Xwdv?QJ^lVk`r%)HOuf^Q61`dXP1^_FEcNXG?)$<E+z0M8tYI~%)iJ2C
z)aX{V@#}tzK85>ABt0zQzK&zH-yCj9u@v-bxEFm^;nBk@^5Z(ZKJHa`7yNL(FIXmH
z4Qp7#8^OYBivZ_Asr-5&>&1%rc4s#|c=WNiG2GdCkPaVy<TJZi^;KePamhz{pauZT
zh6kWCeuep-aWCZi20&iFMlfH)`C0(iGv_nRYi4!$ZQ!~lh5qL>p)(}ubC>CCIF&xZ
zq_xd7mP|d=wIMkk9fEen;A2;M8QZ9%1JoqPv)kcbtD64!`>)c$ZYTZi`{(IjzI&QB
z&&R2*WVExL-97Ilae925j!xMAopXp$NhH||kINzf9_n)WmSMT=yv~5)+ob&^0bKCw
z{U+dqw<LbYE4~)1XsP;s*6?fX;}+qq-Y=R$;atN{53UR9^GcQSKuEq{1twU}q4?%3
z!16Odn+Z(%UFZw&^NvmOA`agS#*kt=MCC((?ijsdtcn$2lQi_C`e8bsu&pA7v61|_
zFobcu2fHPmrQvAcaVo|jIhun!JSHVv(T+te`QXxXd_Q?6Eb|~-0esVP+%>G>7YY8#
zt=sn+J}98kE;ihi4R~uK?S1+v{qz6)-=&A2eUdgSvATzJEH}moQj#*Wy~(O7b@J-y
zy)B^>B|a2k_+m8|tC=RAt-%s!6sF-|WMc>Cd68x?;0&phK7RZtRjRer8}!r3*@?q2
zTwn~TN`Wd1H*XvtuxCH~kX}B2s*jrRj;E=6+)Xc^K1;_h&eHiXmQ`ac;T$eX$j5Wl
zR)_7wz4X<e|BycW^2^k0*KKU+o}Q$aN6)=2;3gkGeHi_*Q+yu)sz1*dad}v^#Gd%)
z`p7C@MI{^Zn<|_{YzxTBuIA3J&-5x)s<BmHt)}udl2X|~#HuRqm9Ij2RnxY3>3oB?
zEca#^IIHX`!pZ8Vkv`lQyQ)$`Q8H~5+GKOa5G1@>&RZaE`lH|Now<yJVf69#fyxtO
z8QUM0ZOD`&tT=;V-)R7honr|Q82ngiRZ$r*Qc^Z}g24>KC^*a~_%FN|B@u*3!%Q5?
zgTc%K1_m{SWB6lM4lCR`?T*_cyz?v|@s+c<XhTj@@E>Dh)C256*obz>SzyOU$KGlM
zd4=}V+1W|;76!GFSA0GD@r74?b&u8VmYT#G*06?m2OF$5#p14aH;>O+YSIty@8(6!
z3bdClA}|C-99~WOU&14l#QJBD#mW(bErf!p2nMg?MBg9%Ab%0(QsI#gsOUe6C-BJ)
zM&blS&O&9t=nocuQ;4oVCRgCPLNg%_k1%e~?9@y`xJI+?(5*lCVSl^|*M0FrF+uXa
zz$C3<4QqG{h|WT96Yn{0<SqJ#P46nu*?pKEV(i*KD9-JAoSLncS9M{;U^0+bZmB;&
zo&tye+!N4m;Fn)j`jdADzt%zFqQ+hu7HeV=7ICc#zYVB@NFyC%TeAB`DV_C3>7+YI
zr<|2MVOI`CLJyFfzAgQMt<#aMkEIu#ot}A;-M;jgy`8pnuh>5Q;|IIxci(*ONsupJ
z9!Xc}d-56Ek$>{(Cuz4?bG#p(ousEPPdvy-IXrPt6m6jQa4)EfZ|59VwVi0ccf)ep
zS55az1Gum-ru=%~gtsJq$A+&D1%9o4+HLSw?-xy>aQ@oh=h#oV&%m)j$}(M9h04I#
z*sDS=vT-_@ewOeNlzGFX44&f*Fi9wzu^SNSJpo#GE{s`T(ZzTXlf{KM`%k8<8XNf=
z#aI~QXN*-czQ$M;c)YFRaG?7@->QKpiRTq#tZW`DE#sVMR9?l5TQqtsRJm<zKuc(9
zYgofC4E#3I&eyPp4+`?P<)+44s#nsdzxz7<>3{UkY46bk$+di3A}sdPjWd+CF*vA$
zr_*P9zSz2Ml6pNUKKhU_;D&L6j}RXh3@xna;sXH@7)(l1z&f3F+Uaz>0?c~=3uml!
zN;X(vm}%5%Qo71%qz})JKYW*tUpzO?vGVC2oun5}o~7gGXK6YPBi99|GefiT!Vt9#
z8&bL+eEC`W-Jk!M9(<%S>aX8BO9SEWb-DM1XPPdApL4d#m9l4wOhq3bM?RvgJYs-=
z)n8trq&r1JsbmWgQR4Eob~<VQ;iJ^q*=4{n@hN6iSFKTZzRXq$zCtT^!na~ImUuGN
z$D37ATU#5cTB%tYIP0@|Ym#cBi?bQ<WA9@G4~%0N&QL5F&b6UVc%dJjuvOx8FiO2)
z&&H{$lw9~+@}P*~L66GfRbS!(hA;2Mta4)<t5)kasv(G2(9uO3V6~STf%AYz7|!5R
zY=v0V19xyM52}ZT5ZgCynPv=QY`sIeq39*wD3fT`a~RO*eQ9$sb2$W;IOmI1U#C*g
z;bWs-OO0mJhObJkDo$*sljo=D`+xm0y?k<PI@hp<HN1O>Zw$X&>g6mh@r0XRGn9mx
z=ETEB#FO4oAdsiRi=>XiwJE=B#K9*Uv@VsNyYtZf3T~E}GC<?b13x_Fh=2FXj_rhw
zP&lnx85MVjK$Bz0Pd^^_xhQ<>LEQWt#yw~bjN$)Y_<>)+6aP8`KM+=6c`uN2Uc(yJ
z@W#L^t7~IBZAq@gIMHr*(xZou)BeGOw6p&x?d(1Dw%XezR^yjFIOMn8{heI8=Ygc?
z3h7UHrUxNTepTOQF#9^>Jb!KQHl>c5xQ-$FIGpbdkgLt0ugU_y31WgCGBi3C#!q&h
z*ecf~H&4?^x0hZV9b3L<LS98W##Fip=jHZvO_<atEbPEhm(Eg=UeS^~U8~on>nVJ2
zmLBdl({8(J9g4KW(Ik~5(|_^VC+WdnQ*!(uJvr*9qtiIEg_---qnI2w9qAi(p%%x7
zj+Qv5-QV05Z6~+qyI?u(^1$qM%)d;yvz%hfjd9$@ZsGUa7`~KgtTcTyc-3~W3VyA9
z(tX06^>Ih=ueC;Ar!6k>_&(b(m@O~F<2A@sELQ1q{V&tEs(m<*S+winK*Whia~#Mn
z(f4FO;^4M13SEhBo^Ucv(Q)*hv4QLI+`f*hi*FG?6n1%HXR7gdSd95GZaLAnig6}=
z??z0lo_gZ>*;&_H`Z8YC>vbJ}V}nWGj1$II!Gv+y7zYV_O`RV!2i;e}oDb&}xXFJF
zYj|(qw{YDG*YJyj)<GwI{SSXmzxm_uQe&@^iqoB}thhthU!}tNQ5*7$85J0Kyp`ek
zG#wo?t8bLH6^`)(gAiLEvKKGLu26c|&o`w;y^%WYwzs(B{HNGHaKpwIK2Ru67@iun
zy23V7Umu#6FP^6E>9IdVli@h^j?dD|XD@V}==t0Oh0cTTjBw<S;v$rqwe<OKze!*J
z;diOoZg~5~ey=A5O&>db&@iwu>*HJ>GH7Cj8(aS6l}nr3tbo%eO!=_QpAX!W66cRG
zE8q0dWri(^Vzt#w?aoeWshnGtvXxK5s`Z+A;MpMza_Wwf=`*k7PkAvkp%`E=^lB^d
ziH|68;8Up_E8*hfD!Ql>DW0+PdurnxMkNjzy_j718Gf>giVr)AC&Gfy-4Y&5dC<i2
zaExsT2*uEe@iD`NdS=BJ#?JV#$Ciumh*e+^k03#OSdW1W-tg=&-DwvX`Y_UMY(||!
zC+&f9j-)t)7Xubo@Htkfg?`LgW3<6duL{FRhOjd157W`nk;~s~)P%2@T05=OXtqSl
zWjcG=^;U;ZzklKU*YIwDHaL6D<fai7x9gzD7{w7rtyHggKQmmxaU;SE2*SKaxWTQ4
za!dTzq8Z!@u^QlnZiF|&GVZL~uf+We<VU3A7t_1qWjd^=SPi*aO!<fT#t*>{`PrC7
zMD;K9Qx_`3#s#VK&@5jgAm{*x^HzNFS@egVpcw!+$`R3m?+$Ugpc%jDB5uH4l*Qix
z`E7BB7XXM&nBv&z7U=>zWzz+2<6zKFP$V&iJNcn-MSN3!=v#xw`+_LO8rJZOhx}UQ
zH}@*M@{5pP*BNZ2?FwhYR#KzUNuPfHRXTX^I5i|&?(FQPR;wdfu&VH~<!B?IgD{SQ
zmfSa8f$Wzv%lB1r*YJLUI=EU7i-UiB9oP+G!?W6TSD3|PMVab{byQ5kV^W)S7+u~%
z-7@t>s#a^M-E2yC8Ki&x+jsK2RGT_X<w`Ymy%MaSPS1Lht<lx=eUp6OR69cl8*`4X
z@bcB^caKw}x{<#4<dFzFPcM(oq$g}h_O7IMqnh?x<<xG~rAKb1p>z;r`IF;rdiwG#
z^?V*LTPI4_v3wF|iur3S8qrDIj8$IBm+zSe8Nx??`r|hM-y9OYd7w0aQU%P@FGKhc
zZp_ysycbx0cNfbZVCus8+JKxuJ1QQR!2vhpj+Omy10k`z30&vHtHUx+mth(AUGYB@
zxUrt!0a(nu1>QFA`>gNVAm_WP?|B#Sb8K6gBFYMVtIGVkTv)`hy38x`xRKYl@T^FZ
zvj9bW8w!vIZFN@0rLmw5>$8eec>63e(#Rh-<Cu&G`tlg(*ke$9TZ!Wd{Eic=#7<97
zy&!;z+@n!HO~*rz;St|PH`3_NH;*rS;6bk>j9|X;5ObBzfW>c`^gS0UN_I2v3RtXz
zyW-vf3z(Lmpk*cPYxuyxEoNPlYxw0r(W3W1d62&S$A3tl{N~G4YBnsEvPuZ!3fuD0
zIcLKct9NXml0wDGA!gJa9Ua@qg`(B#_G|>hNX5(v3`ZDGP}s`bQpieiVfd<7t5Vtv
zrAK&dq>-|<rGCA%C1qauu#Mf>>2VtNx?%VljZ^=mn~tBqNJmdk(`1NYN(JG2p>GA-
z<z0JA72G=5OJD!__vy=Tf8(tkd;MM-2{%ij*>(`a*J#j7m&yZODYnzQu%Qk5wyPB>
zaQXmZ%#$)VX6B;uv0|kQK9ou-Q>)>bxm!{aS>e@g@1`n7Jd{zEZhHo%6x0%{vqZL(
zLX3Vs_e%ItSj!4yNj8Jn7#Nww%lT;-lrRqQp=Fy%^GIQ1rDe-VjEz1IOMcmC#HuWe
zWzvk}!>kW9hA6f%+!THr&Y*!6Uel4&QU{PZmIBO5xDDl7uQgo1d0%Cq?5$BsqBGhG
z1~_KiuzkuvG$9Ol%S`cxyxHau>4Q%_1~9g4WE(}kziuPOqdJ#jj$v^_{AW1W6%sDj
zwS^4PF;ZI?oOaWT<72yPs-t$hnf4F&eIHLQ)2Wp0AOH3wJ^kUi>h|Z|eY<?&eCO~k
za8nBx{0Lj%LwVMVrw89uf7E4w9DX+x3s2P>uPM)L@dh$`gN3N_ilO-ew<6P0A|N6J
zZvW_+_TZjF>E&`(P!xGC;szmkgOL~z3);woxOy-YkpqPvx%nAV7?lxX+Asg|zJVXY
z%DYhH5dbDd-h?L7D~^JA$2V5-L$R;m#*cFGY@r>>fpC+V%R&rrEA2IY_&K6GZ>Qh^
ze#lK8V7NlF+=0T6-2A-6kDysKhSA3^;_7~dACZT_2r=g9ZsHWCMC9-71u46d!Wx%D
zX`_7*Cz51-&^`)t8zBTZR`?X<$^39%LVhH(=~4OSyz$oWwf0e@;Zgwc9M@>{qu^-M
zPUM)}e-eH`vym?H!?z;#djT?D!y0}Cz-yIXuxR%AMJwLEJ0Pd$wk0_+&g5z~+v(#^
zzK|^QQEIpL((e9Ysx@1RA~|rIb6vNUD$>yxtb&{T*YI<Ng|;BDY_7iCgIRnls89a&
zT#}B!w&t{fiM}OW=_$kUc`8e%LbpOj=2O>hwIsVs&(gPn&r`Kbx|o>8x3D6ac{Djs
z-To-`htqU?CVi_v;8Ww>JXl3I67Erd;+gJzQ_d$t$@LfMq2%lb(yOE$*<e-?O|{ZS
zs%%bEt6uTeiEPBdR_;<w()Z6!(+^Kys7zDQB-tN=&{3pcphs-$*{jt29n&}1ybuDF
zH@D>&*vl;+YL?%y2mu^ZWpoUC$G0@c;C?-QJV%(W<>2=O%kS>>0ZnS?%>WKCf$(N<
z$-{XKpzA7auMcld`+dSPkAErsr$x(q0z}i>;BE81-tJ!=-evv04IQiCLup&9-U<Bj
z{XXG7@n0*K9;uP27RnZVRD3_eFy*#PxWX74WDS|WCgufk6tfs(8xBV{l1=5lnCg3a
z9^-Kszq}h3V~$>7Hd0(xMGrj&$O<yj^#?ugZe=>iPxNvj50DRru}q@}KCgiZEw1cd
z0@3Y|H{>&ES3`tc|F*8-0|dVkD{C@0fcvQC8#8@NxNn+)H9rFGi)P@haSOr0=O3j%
z{F{GDk3Rb(m77&{^P5(hn33Qmzv`?pdSFy>2P}UKDhzzmv5(ZrBLShP(SdtL1T!C`
zK+(C=ud=FV6T^?(2;EgFW2|Pv@P%Oq!x-V5?WJ4P2OYlhK07%{{qCuhIkvVNr0&sa
zI(qRUoxJF!$v_&FJ_MNwEA3v}8|m<qkJ7h)`BVD*o3B)^X?p(hxzfLI*}NiYJV@u>
zl4EM441){c;XEqN^MW7@L|)mXJJ-1nhu%(=TAf@Jf5Qf@Eh%xErR`L%)l$2?lRBLp
zZxzTEh8U69vJb@-g%G@a?7IDax)dEKfnt^Uit({lZFmMJjKc`YM-yB%DY&FX8O0#C
zDSA<aar;9mobZpbI1j~RJ4)9bbTUi0h|8=L;$U>M!=Ln6@x=fwyg)H^<*AODp~-n~
z{@^Qi{s0pf;}(3vpa`$1Th0RDEHV^Y3|tr|OPghHCkekd6jrO&{84A?$>FfCHYjew
zZ(ZiA%En->>p(cTvxUmp%hPmndYZ<Aq4>3xTAfBZeDEN(ciYN;BlV8D>B--orYAo<
zQyNjb+&FFm*sv7$iq~@ZiufzR@g#8M&x!ucMLpAu@dizBmN)|OMM<0vz+g$3Idd6;
zSxy%#L+nzr7}#eB0jUi!hbkmJPfh_&(i-L55MX41W)@D0kp3bFr|~#IcoiwHfHD1-
zY=uc)&>7KU;J%=WD9*6B3fu<K<)>A+<-XidE8ajLR2VT8=AjarqVQmEIH!pQ<D3pU
zoU_Xu{D`DW{7{>PryzHHN6&;td~i6u)8HBSs6*WPF%8ZhFwIdn5v7Qan_NUVw8Juc
z$~O?>3i{$V_yUDNv%vW%1l)4xu!t=HxBTIsjclfmWUs;x#n3-K*Z2XRBK{SA7y@6k
z>A?4<{GfZu$$;DW5t*=EZ@3r33M`BKFhOA8_Z9eI+Q<{)7Wpx&xHsm9Kf2CRp>M+v
zp;~^&^|YAezCoF;VGX|q;I)c3oo2uIlJOGo7G8zooixb3A^EXdPi0;8W+y#-{7HJO
zi|q?p)zxgYQb}?$6O!m{oYDpYq`1Z0uEU2ixVDBxpjPMg76XiudF3AKL&It;<mKI+
zRysV`@z&?;8qr6-pGt05dzwx|HfH8IGJQUyi!BS;&YS@?{3|83D`myXFC|T<oc-04
z{wBG7B-wvBj5B{Xw@T7YE_^;S=K-S=RVyXsUGn6`W~*&IV^ey=wmwqT?aNe^elxZn
zpl>01fN)GDGrT<MDxbb}3-lh&|6;ORRpqEL5w3*(qc5I5j%CU7LL>cw?=|m=+CLvR
zL*`L_UVhKw&GW#e-vdASHs^1CUvpZ%lXEZze}~^c-2LuC=_0kS1JTBA&fzxPq<UM9
zgq*a7_X83I?+xyfD(^GzWmq1-e;41sH#G{}1s}?{nexXsDDJNd_r<@vU@_lSbYBVQ
zx`WJ_kx4*&z758~I2a%|<C@uj@nW)L#EMm3F-bfoqO-%2{q#MIEf!;CS-02o*qkSf
zv%1XVcLV!*%uaEM&td>3!!zmK&y0_;+tWp143j&60&=;ux+qVSWDRS$1^o87mE2cz
z0Wa_&;$HJ(4j%+RKvp1p^!4ZI&;Rz{rkzIzsZ^~<m|>s|g#`o5I8D^C*6URtUQ7~q
z+`8B!o)u86@Is8nID)~8SqvyO%zD7cq`u!koQXPmI(^RXYDuYYH0u5^^`%siXVY8i
zZOU1EG}M9jPft>}dzyxWK^k;>sds#uPL7V!*~?xU_gRK4LdhdvOMv0)^KXBXzWMW?
z(*EOzqI;Tp!q*kvfpCtcAWlX@(G~mDW2ED4hi;#Br84p1W0oW<k+_&y&-R6E6N2H3
zGT~Rkpr<mh;)}DmFnTql7_wd9L}l4#dq8IEDgz9Ce4xC7N$D{dUZ8=B4{^R%Rmx>W
z&mnl&Xk<fCA&gKzsvEEBDwV~5jD}mn$@yO+@vE2m{k~V@!B2p4Fo0_Qn&zlWpX<d3
zv@ra|L86Kisr*qV4_2OG;KUz4441;@bY2FovM?KkXUasKQ+AAB7^c|D(MChnfdvI2
z!U$I`Rh^D47onGrcs6hm2V)z9@D@PCXZTI~>YsMg+1Y6t^f9n*q;{v7cK14|(QcX6
z!RaVH`TNuKR2K^Vs`jxG+yFerWOp90h8Pb-E6}LyU!dhjL2kF<fHnG8Tt^DN;WRGA
zp)Q<I89AK-!R&Y!;-2c7B7?_%!a@58+;Ep6_@HP;BN>oD9?n)zJUTW>6~`-7lsT(e
zh=bA5$#f^aF`Bq2hjDNhJ+ut5zo5;C`8RIkD8OV=F43{M73FY#N@rrB$B~@cS#2s@
zR0&VurYj;Sa2N6H77AS?kBo>zyG8H=GyX+-Kat?o_(53Ih2l|WN5_AGA4GTNDv$G4
z{Gtv?55$Tg9u*&dp7if1&?S0k7j{AeWiQgNY#;hF&Hj@YH0Sm~ETyp_)c42))5+Vm
z4+>~>B$I>q#wcFfFoqw|!M%bXIgi)khvG1lb~ir)H{M(3hf<pM@aKJrHbz{f7lJGN
zpe-7Ut`+>q`Oo>`fS_4-9K?sy&Iio+5$(13AqS9gzYRb12)T>=xL9v>xP~>{1Mm*t
z^lI@&F1`xH<878Kg=|-ENM6}X4<9~Gk3ar2wL1H$*67H;?Ugok6ikGQ0Tm1s7rf?{
z%>r(JrQ>_Kh7SvJd!j#=OLae<OjCa}N+*(WyMu{jsdCztocrLQlXiDIjzc@+ysj;U
z^#`Lgk_?Tmz=T5N?}~J)m<(}|PEWd??8lj149a=ibUx+6gK>)QVz)mKo`L1;a=9wJ
z+tTSmw__k~r``4r3w)WFlvFNnro+9R)T)=$)}?4vPT*u`jPc~$Ml;SAo9LT2y_7C=
zewq4`-;YmDeC3($Omsw`sg$=<vtCP`cHOJT(3xoU7=WWyFTPjTKR5PK#Mz23KgvT7
zIXxV{o;E@Oc*Vzb?(_cO?ss<;?XL>a#%|8xHr%9oTaH+0S;Kn*i;DLI$ho({edfIk
zp;OFZ{r&rS!E*ill(0yPd$iXijr8<l=yas_7)R0GLdOeTFGgo1$H|}g#M3X)6wZVY
zzE8$r-3~wEa9oVX6cxgqu}K@aZ1CbNeSJ$=V8EqsGTSQldc9)ecwrn<*rn1ixt#cn
zlTI0*gLZJ*54^F>V!)iffOTK}D+jmg#TK!KHQYNyS9*^Pue6cA`2BCwAOHE^q{eP5
zZA#&|#K@qG>99SsKpkkKSx?nU#oO3L2TBL5&R3l$=PW@WE5hh(*>;YVNhoaY{OM@j
zp(~y47-ld^bUGa=H4Q8D7-&%F>y4U}G$}bKY1tqn1+05^lFm+#(y%v5gVVEg^71&H
zzU(S2@L?cX*+;7!TPo8xfB9qjO66>{nn~TgjcB7@H+8$+G#)Yoab#tXZ9t%dj}JoK
zM9LWo8Gu2?E83WKDTQilvy`?gRd4yn_Y6M_U$tg4m8BGN2E*>|PO3G;SK(we6f2ju
zlm|TFEGzg6Ut%>~@Rcx(Pi#>b#w@lTl-j2}Fbu_3gZiLS7fKt31u3{vdaJdHaY1YM
z?9}CB^;M-*mV!$BFtkB1b*ZSv!IqN!Vb6xE60<ICXiEMt^FhZLh~X>e31cyIdL@{E
zc7TD<tI|{kFUKxgE%}7utdyD!&KOV|84kUY47~p4iZ%vwX(zOu5i9LPi&uDo>w<7v
zB!5!QMZ=i0zVxv_J?o}kx9|FINm1_XN-1wOQhB?a#{F@6_TBUJ^t)%GE9f!MzNuhn
z)Xg*)w|$g@+_0Q@_Ct`t-_e|xZ2}JCr%;eZqdjQ9Jj>T{^gz&6h!{>;gcnIep(q}l
zY(@wX;S=>Tj~|f&$T^gH+F21q+#(*&;a<QE=o$Y!*dEb~M82!oDmQ*!t%DZ>>$<W1
z9{t89?NKrK!$<}&IFDOZYkwPSL_bD|`8I#>H`=t*LkQ{l*TsN`GbfBV&|ZW&AK@|0
z;36OSg%6-`Ls3p13ulJl$2ISjaj*I<;9eD1F!NpxUlsp~-@?O+xHr;#Cy)@hIpv%3
zLp57CsZagt!MQEq3#WePFPHALY5IwId>eO)IF8~DKDQ4$BM+kc2frf!8SaRXja@l@
z`~(EeB#uF{9JI)Ta4X&t&EfA>U8F4%74R<j0Yut)oJg~pABF`yIQ7lu_P!ufv4%DL
z;sLpj9>7Lv!}49Qv6U*-Myk{}e{?4u9zIH+e*T4ImHkw!H6-^o(sl_$0+X(EA%p7A
z-@W`U3*sftuX1?uJzT?w2Mn44F&IWYMmqxN`zPwMdxLR0I_ajDCugZU7)giFH$iQ}
z=K)JjXM1(Fj%Rz!wq)LBqwaRYB(H7BV<BV5_TUIcwVGsjWrqx3uho?vIbZr$ukY1Y
zQfR!w5Sg0^jbn5<>0ubK*ui64xcJDRW3-wL=f$~RhkLDbxZh5@t!k?9p;a0tMoyHD
zofbI5_qaPsC%s|n4Vm1>`%n?8%XB8)<Hhl*^t*04>-J)23YDL24{Md})Ra!c$~JT%
z3}Sgz7+c3%78DLoR@~aQh=}e(ey|e#1$~Jb@jk1#T$Y~1_mZt1DT>}PzTd_p*)>1A
z7Wlr$^L+rgZ-Tq?B<;=VZwJbk%eV?|!%eEU<p*-h8r~OJHoPZ59K8+hGw(J0GT}}8
z+}rxo+hA3X#y2^CpMm_1UWVNMZ*J$pqhITHhWK7(z~v{(Ye$4Rio8h6H+G(eJ7Z!o
zVV!X>{9>gS6Id8SWQ;GyPA`kqUyRG1^?IqV?`s~r9ghb3rg!zN9C{p*aZdahbK}nB
z^z(C!Y>Y_;PR1$&f{u3$%k}=Ib-IRk2Ck8HEv(^#g;fo>)M=!z|NO`F+rRuVRq8b<
zJ`#voktF4cP6$H?tBN>xitP@!q_B8!Qeh|)o&l&1n1OHZ%s9Y+G}SX(Acm1dpo}rP
zFf7jM9kw}SyF(NfuVT_42QgQr3{{lBV#j$})3MLp!ie+o#j|uio}@vypN^h9PbV)q
zG%^f77i=cKsvj1uwN5kr_OE|VpMCjdDp$*Cd{G#MPLF+Nm^;_YSW<kUykgLuLMGu4
zBObsirm@P#z&d51Xbh-&rX@3UlXxbDwBBy19I<^(r_)ZYRy%D-X@i%f^C~GJVLL+d
z<io~FGtT|O;B_wg*s6pN5G&)F^`=)vp}-CGvBD^Yu}xX|Y%h$4lmY(b?I;<9VI~M4
zDhyX#e7Iu9=G00gtFNdl?=Q^Zo8F#~lJbFzx~i6|sa|iSiWE;ihAg?}!|H7*rEt?m
z$j>uZR1e_7sE9I7-Le9Z4=3llH5yHC)996e;w6eV3O2?v3>WalhEvfJE8(O(b79n!
zgkj#1pW@P%rc!PPHhi7wgRZiw{nZ-fw0F3dTAhyLk9*_v;)fUM>5nhec4D7q7g6<V
zqYF<PVTh!y5U#AK;G`PXLoon_9}(h-hY)rOi6?#qO6*vLf280p7kO}ZCJUk53IO-~
z(UFSuIEs+SSl7bc&ILD6L<hnv<Bq3UG&>aHKpqM_`EF;VT*De}gxtPqFbmCP_P+Qx
z;_wy$VSKiOA6?;Zx8KRaFa8aSP0{bmEFc{GI@Q9vShylSF&t?Lw;u5U{7hWOi>LS%
z6f~23j?oeAa=4>}+i@~t0VxQw*A2MwpVJ(eUymQewU^`RTe1-A-XW8=hBf@6A>Pq>
ze})Vg&0n&h<+NIAwcDxDY^R5hKi2hW>g?>NTD_G@3S%-6aaq}AY?4Jh$*BP3OaJRa
zM&aFd{~=fyms!IH4mNZmZZi2ob<BBPr~P62@%eE&I_~;eQn%%0WNr*U05UqQjdnp3
zqUpr<fH)Y)VshD3vOD?%MxREbCY`08n$2d~-)pB%t7Sdm^z=;n#7J^Da(y7e-&_9*
zG3kAFv+lj_nRPpKqTNm_eeubo^zd*e9qu;Le!HHeSy&%Lp8yAw1$)D>SF9bMN@p3w
zAlrs$+bC5N8sKPrnfjdDt1woV<@1*>b%!ZT$ZIugsoj)L;;r&aE+0EdP+p94z23kR
z;aK%%E{Jr_a~{Z_iJJ$zom6LA#R|Gs7^_(o7tH6)n(Ipe;x83)4{#gay9a(r`=)f!
zYemns8g9c)s=H{r3&iU+yf-LB%zJ?>uHFRqnfDrg*|69@-rXNwjsH5Zsz+Nke60d<
zvu?*raAV$lQ?7<ZRAjxtKbIZn`7M?O*Zf&`y;OcKyhvpn`sZ6{Z^lIVrqS2x&UcQn
zG{)6P%N546SZVF?G-L#QcbRzNm185uS4Vmt*f<tjEHX|yb)2#Mykd;B!okE5+(6J*
z<UcDHl+DW+&!%<_KVNVwUDwhYK3G`negKrM!!JIz;p@|{zDOJ8ICILg1WVhA6<Zie
z@M9&Ex4_eD;Bej9PJ?~qFV&%*tE2Oleey%$LD9j$K^~-GwG-f#N6H676^dC&`OrsV
z<iK#lEW#4no44Sjla&HA?0c)j?%A==+v=Zm(~BRUr?aD;4PPjO7ZW;om&*C%QgZgy
z&f~-M`+xXzI(YOTwg4Qf6PGeHo($6#9|5V9UNWrCd{dn@DKR>M38NphX2ld`%v&8w
z31ybW=4L6i+nsc&a6T^OdOhv!@5h#r!ccA0Ri?JfhA|MF;1boQWi%Kl9$P%tQ@J8u
zNJ+%-!v_YX4rT4~oEe<tg|R98P-ZzWb09@=Qmg{QfW!<C8;HbP50Z$F;N(NZc7uG-
z+2fy&u1z|sE5cY!z}Z{WH@u)yN}`$iVrC|^c%`56WESW^A3;<n3~iigM(>zEh|Cn>
zfNczPp)9K`iih!xd>N4D+%nD>01su`<_rJ~i%QQ{l(aQwrmzLd`IHYehFH!DGY9>F
z9<m)vx0lYOgpYc$Qm)#lq#cy`ogHuGGM4|-@1La~|NdNUL~)>+rXn;hx3KwFcs+cj
zFwPfnJcmXM7tl}9t_(tQ$*Ltk5E3a2@Lsn_W>@@%GyEU|SWGm35G+660<VSlRbU~_
zn<QVuuMl3P?Y&B4x*CTk`lFezUaAOnh#)`lF3#V>9A8O&x55xE2l8$jEAO%X_)bL{
zehctB&U1ut8E5fi+(tRao!@Z_|2cL<(i46K826lJAbtfo8hHFL4|mW?egN_{Ufkf0
zLG<(U^Ldo}hfMt%*6=F<3<V6pZ%OWGrk&mW)N1Xd_U=J?{K;qO;NU?1v8oHh0<xS}
zbV^17LI#V;X8A?+yIb^0A+*?Y_HsGi`c3;_!2@4KWz+xD9)_cfG?u*I9gNe9<8C@W
z=}CrWBB68-$;rszq_cdlRB@)RzI>9MWA=BnCGcka=nC;IVz3LnYs-`MFdT6fSF2u2
z`+Iw-TCMsFF;84W$J^fWofTl{QLM1yyThcl;b559CLa9*U5S-o4-R(Hr;qm2=Z|+%
zrzv?}M6==xBNbaN^}1chWe&z*DBV%lS$~|4x}$X3lind>*~U>nrkC$I=apP44YZ>x
z)<r8i5On1`MLRw|Nzb3Z2znH|RH}O55`6{z1AUb35^K^u>(XcHb#|-R@}>d|0J6gf
zLx!vb<BJ(%p25$#5$xSU)X8da-Q5aNpKk{AF^e+8GTernRBy|V;OiRR4_F?(CkQ$7
zX1LG1*RY0n0jqknkjLh}tNLp9lJdX(``}nt@W*F5`vl(D?S?PUV$5*I*Dvto_<%!a
zUO_K%C69b(=lyJ?2B0<XOhEcr<s34F%d9^PW{k-(`JHclUJ1=L)wh%}KDJdnJvmOJ
zAzLesbg^sJ#AAoVAsyqZ>`=us<G73qLKn|0Fu1rFlbi6GoHhJB!L55;Mei8yi>c9N
zd{`BBpEPIObgkg0OGKIZ^3T6b-~Pj2(&6KWad1AJpA9n@wxZKwMqH&5t9#sQ>y8qE
zv5T1j7>q82)WLLe?)WgGFvB2b1!Blyg^>^HmXd@qiB(Hw;dN)I_!wE5Qj+4!a3NZy
zWU0d(^x4YpER6;|Jzu1*l$>WjJWFRNg##2Db)NYQ$3+l5o5Fwi=|}0efBIwE+1pQj
z;TR6P>aYi*a~y^gNoQVZ)E^`!Qo;jw_!!hM*ikkcnl9AoXT#Tq%g9*_yL%6mQOxjd
zbXqpJY_s}mTgsKngF;!78`GaZf#BC41|%DyR4J`S-G-tmGccs+UMd<Z?6%>vls4)N
z#g7$W+`Y|*h>cZI7{K`W3|$AL!Ptcniy0XHFuPv$ag&lf#@ME^!C!@h63quwN*^oA
zP!vgmfs>V5oMXl|A;hIVF&wgNnt0GmdII5(&-u~^4x<)Z8}boFxr`6#r4?0g)gl@x
z2a2%*V6bPEFr$>RdtfNfEYY2oOJ4XROxyS-8rUAC+wG}L6A^Hks?BO@?=)@r;?V6$
zf0~|Q`1<~dm0!x>bCB$kTga&IC|s0-E%-vB=fTll52!4FwAl^339dD);nxnlL-+9>
zE{4A@+&AuJJiwygYvNrSzSgjYcMZr37_K%`wb4jBJBO*=+4VMtpMUXH+S@-&o8_va
zlqG}ZRbBbzV4xx7oIJo{*}HJZ?_zvpVu1qk%jG!hH|^&EQDN8WgEO=cro%h#`3b=v
zy=fx-B@9~|UhTB8T}kZ9eJt6w*Bwe9V5`At8c3dJ*0#3+l<d!-5jqI^5Bd`g3OSw}
zOb_`=?ndsXS<$|nzw(Xqc6I5_Dlqg3RwmY}ab9t?T6TNo{9MlZ!r;a#F(xMNNVeYF
z+p(Va^x2Cv7{;y`6X_0n?Rxt5t4HaRN4xSnPhF*DRb#cXoob~m8=ly~1AJ1>(nzBH
z%hP`P?#YXEbk-FQ&__zntE4mr(ai+9k;;y~U_+qvGj<<g(&JgTYr_|OXtY|X)loTW
zjnH?b(=dt7)gV;ttWi?AwzuRV{&_2V>4Bx~jnpVh*AXu~;HR>YFX|z@ffr~E%aDD*
z{kmYi;|9q6z|C-T+Pfey-;6)n=S}!Y_a>#ya2xJ5+y|@;Uu*bKA@`AY_lIu|t9rCW
zS?nfoPnP4|(w7%8!%6STM}L;%JY9Ud3$S-k#>7jEPpg3U%lR*mYMz;%z`e-B<#NSj
z8RP@}B8)grlj9SyNRy3Vj;*xbnMtxoOgfESzxZ}CCOI1SCC~KIpg-^gcP6oO`3y4o
zPbaa<6cgt8zVlsY%SE<uWc(Gp-p)~ZnoejkaEI>vyoP&(qGhaMCB$p7X3)<W=vEiI
zj;&fX{r<oBYx?b<|CBm=yKz2_Iv392io+Gv={P2<e>iswgVniL<Y3g{8AAtqs?+Hq
z1-TPcr$yh+Ohh_4lpmj;#rAhfx2<%n5~|f}Hm=awZEtRSMG&i!CgY)$l?^Lp<FPu;
zK6}$jDH;yb#dw;|nBDi|vvhLQO=Gsq7^@Rp{ar|9DPkY}=F9Zuw_jVaqa((E#!Lm4
z>SBVT@iMEBtpVAx5oC0<?s(yYc*5b%bk0LvlnafB8I3l4?H{CawP8aZMlQ@$82DNp
zwmhsATXDceWu$*d1-G~I5I?1JrIHlQZGHHnlh3OUf)03sp$G${KQM}4t+3rk&2<F-
zS#q4|BsV@Z-Lo@?V}yhM<i!VNDjuQ<vZbSEI*RvJ5>>CG39~Tt;S&vPG04`0Rn-dy
zD}NM4{NzF(DU@(<2`7poWkeO{!!{6aS@Bh)>?+f2^_F;9Eya0gsNmJYIOt{AeE681
zDZYqkw?%kAR-KR7q7WX428>_O<`sqgz7*c$)IFt6RVL9^X_UR<YiFmE>WzAu4lmNP
zA6}#%zkeEA9V*{EptOcHTm`*YZXfR&zzf_5?lr98Zcx9P8{zwabwgjn2LLzs%a(O;
zNH)3tRX`ToO66KJ?HoKzt)0EJw|kI2`|QhfaPUYn6I&QoKq5IN@(>#1%$Q<!@2kM|
z+}EFzpAe9n-8PsgM%$3AO1<JoCF{b->H5X!q99M=LZ3jWReNAU8_i-oy-a6=NqTwG
zOTD3VsBOvV!ix@p?8+ptQ^~fy-Y5+@&(@Q`&_5*T8n1Gf8=2W#3^{vgG|0~)HMg<c
z?qZunrNf9L2fHw^6NS&|Rk@I_Yqh#@Fb223*HQR(IypH@PoKV!96n7Q$<=#1ZME}G
zYBU-?_p96E+>i?oV4X`Z*lSnQS6_ZC`M;cAygc^#WXSpZJFT?etffw4D>c~tLHV)5
z?5scZiZ9AI7*A4vBz=N&jTK%hvnAoC%6%^XiO)Ri51BB?mWac2*6o`vmq)r7d?PQm
zO2jaS;fviR(8I#$g`Ohru+my-(2s~yuLgY`(Z?jnljAe*D8Y8}jd~?<u9z5Yoe5sL
z4agk>|M;1Oel_j201+Cn=mwVKW{5hw8Ey}KHCUzv2)wlGRiOYemUF);?oIHz@z?Nk
zg}V$}*)X;?fc?CH{)MV>9~NG}T3GJQR>7;pc~h|5undb{i<!%RF}}YO3*ioP@H`HF
z`FpbHUIi;@xdIBmnx%kmA)}g+c5#K6{d8xM1peNgOW$0+&B#Yg&S%0q6Wn9_#IcXa
z8BZ{jacl}(EV89yES|uRu}6=0>SFAeG-Ty`8C*mrV_erXy*@1SC-Wrx^SiieZW$uv
zHdvJ^!`-RhGq}`08CCb0`K#d0RrRy0mJbMi$_7N<C|l>2M1qP+yP5vsfA&94U;g&D
zsnUq^wvbRTX0YuYhZAEE;%{SHiiNPy3B)QQ&Z1%k)W|EjPzYEF<&4z%c-uQJmcc3x
zvj9*)YTjalRbNtQ^hz;TfIs|L@kMf0JynDgdRgf;>~(GY8uj}&hV)On>G_iv>GZ`}
znhvL?lYZ_RNbB7D=u!ISFMmvrKKe)uxRgS3l=|IMcbq7#C^>Y#Y}be)iNVdFGEh!B
zRg5g~g-)A8qpkE3McHD&YqdM6)!7vv%4woB+fw+d^|}pvtp0-E({p`rHsNmw9PpIR
zAA=Cvd7z{=6(8l(+Z>{_>ffENGK#aVrm7>=Tbcds*)}lF9Yb+tR;3M<x~MZ7)>yHp
zaArcXEeS?7jB7$09~qS!1*amqw@cf~4CR#-TQSpu?*=RK=<qR`!JiVwNtFZs1^uyn
znQc+Xv!^=ZLsctQy73{hQcBqc6VY>Vv0<8tLpwq`rYvyQ^{kH@O7FRiXngd<gUc{{
zd8HwWzRENjjMDMZaXOV^J??WR8e2_P(!t?Q+SylqD36OtO3#0Mk-q!e59wlj`$vBb
z?+{?~eRX!Pje99UZUU=t+=zQsq7{BOrN0p`unPBUz^%C767CcC8r~bM4PR?m!!q2|
z?=E+Gl2JA!uS)(X*BWVW|6$tQJxIHI2kF7%PtxAOL&-ADw8^BSP2{C5%QFBIm4LO}
z@yh_z14MF60{}8>tf(qvSP!;P{}`3FQ%mwFD~v2xOFl(TZlD86j_&ry>APnq>F-Zo
zrqkX;DvWe0A>Z2GOr2KU*K~ZIULN<-$r)$)a>fP&I^>Fq5)<QM=5^>RmW?HkN9&!h
zCbXOlwHs`O9`ZP0Ojz^eJcYAeqR&~DTz)CJo^1s`{q(VQw(q|GA$|AV50>9+)lzzV
zc#s}G*!RRZ2A|j-aWovKUa#+gsfP~_(x;z1GR_~LzVMk_Y%N$XZKemijr5Vyvb8$~
zD(LQy7}%0NHoZ(^bWX|pYz<#3j8+#Hn_lt78DwWYW{@*jMq1Z^?@ZS0iAT`D_Dk?&
zTfAh&7Uv1qsui&ybXJUA7^j<^R%)o6t!6#7r31GboSn={x2{*P)ub1dw!p3Y2JA?I
zzN0)<2e1@<(RHq8^c)*K&2iMWTn})5b#IgXmf^1cg3#Mz+zjNuNW(Iq6LPJFRXN{<
zaSgvzSR1}RJji{E`!L+a_hb#LVO5X!E+HFM-UYZ$_2(Kb&?27zVHA2B(0A?zMj(Uu
zex)oS69L&!l*6IXor0c`W3oRtzNeIcaZyi*XN;Enc<iHf*jCY7JD(Tlk$E+l+>!~9
zPuL|41KD8E)t%#C_+DRl;yvS<lq2q$9x@I318@gDbI1jUzqiA3%Ik0&hPzY0XK)q%
zWK`W}=09n{ens#`d{(n_|3FG3&<}T>3om|*w+PF@af?X!q55D$FzA7^um0)Z{k!z|
ztIyLVhAn2pN$BTehyoR>aMW1{2MQZoW7yy&;UA+{oaHrkd^$c>M0v%Q6d?>=tSoY;
zB>&QOIUOAANwKOV&bGq16Dz>Btsrd6UEx_uih3m}AI!#!ec?F=YoPm}duGGe;Ix-s
zNSWe{ugL&IiVBs>wFpvr9)9^*`qRJpr_|nUdyBr_>8X@7;ZeCr#8!iYURPy;Mz#>#
z_KZk?Z2{?YQK(Qz-7AZe7^&bR<=@yUrDm&@+B+)Sb}5a77o%6B)s#|LwGwFqn#$`L
zn63-)OO@*V`Q@*k-}$k93C2g1(lQ1ntX%4lz14|mCqMk46D7-^H{rn;#0*X9gu2*d
zkVgFD0}&m+I(qUUjq;~%IU@^YoDYDvO*937+5z<}d|tM!vSU00Ft)LJjM&sMM!VUE
zR(Wv7n2nq&7sfI2vQbR*ah?EeCyaM0gX)@eK1V=2r2fVz&=|VBqEPKZ?S^M&baH^G
zjdn^)HB9-?NeSjK>*J#nwTXe-K?TFt!A|Pz@7VCh*>BIEJWt<$_cWc4*mf~W8XOP6
zhb#LnhOOZhAU7Tum)*d-VB})OWcQ-q+rWL{1_zMwW%xPM4BRK~HT<NoHhit&*ADpw
zxVv!@6QjUp+A7y1i(%|Klw9#3b@mR@!-t=w-TephXR;V8;Uy7yVh^%&Oz_DHHCrj?
zfu;`$O!Ez7zx@EphYM6|UP;B!3H8jRF(z@<YZb9fa<IhZkpgRy;ScxPHhyhhPNh$v
z*KAsT=M3e+_#!=dahCq|$K!O|8>g}O%DKt#zELfu!=0uw*if94^i<dB*}!sp=6=ZW
z(oZDYGq8mGk38?Yo-LUR7k}sr(0Xwxodv^}Wq#^7CX`{oVsaa8mNSIW0hl=U@WDZP
zpbNv#k57NJ-cuFryV7qye*8dXM0RH)q-1goQ<BG@JwKAnK1rW^a+n@H*tbE<h9vX|
z=`-Zrs8>^`Q4?=?hc=~~vC<1&PHLO-X42%*N!P2ps@_ebrgB}1XH3jvMH;$Nyl+g5
z>ko!rnfCJZ%q!%GYtZd{DjcG--EOMHY|k%UN_@7SBz+Vet=WicV^j61SQXLGsBEQ!
zotE^gl)63P?GIAFKa#s|{T9O}6Ev-diEwCQN+c@)d415kXfUB6#p}cc<T|_w?&>oZ
zalI~(_ig!89<J4}D(AZ}uHi!l#-H943UzyJ`1%=v-{6J*h3~?;dTxeQJ>I*7kQILh
zATn;JMc=bq7~!-<TUs4$(RU%AoCX=ikNTDt?GyQ#&#1qpcquN1GJtO`vJ7F2(K43H
zm?Yo!;h=9LnU8y!UU*EAu|(vj(TMN(na41lMsf*xvhs|yj5{OUAzv|mnRycBBX6z+
z`hVHrn)ua}_Ff^k|1~$S0Bqs%Ywgw+0yK=`or{hkm==PB0`))lBfdHY5)mKoA{;cR
zYc_o8q0FqkufI%x{<r^}4nBUI^t#JGRy?sL2pTX@VaQOo;hBJSZx^v`14lM~(HUSM
zVTBhKc=`oq|Mgj&C4bHUV;hWCtDPP{c$hY%tekY!>5RsyB88(QymVaL$%CyyC?yoI
zWg@c&F?w~+PE!9&>7)qtrQ94pKS_fVmcot`=Y_3;Eh#de{^rZ{^<Vy&DpInh>e$#@
zcG%~9Drlv{9HoAbZ4^f;`=)3pNio}U`6(O9s#n0N1N6WK1PV~LIE>ngQ?s7xjh2mf
zD3XnOJyq+qRH;^dVUQU~`ErM?H0QdbfP$Y@Z)}AEEhtmeF^54jBNTpM1S38jIGsD(
zp?Wu^plx1a2*Q{Y%YMDhp^akDH%)`04k!<G&T26H>CAnoy6B?(-abY3iO~`TxZ-UW
zw_ML`TS(c#%Q<BzrL?VV^o%|9W3?Bn<Hjl{10$5pGf=5p(dDv;KhWkq_w@%3aI*a+
zTWGQ^CcMHJ#)pq~!H0=9=;lewgkQRt=`rfDU+^#;A4w@aQ(48HTC<$?4mxRPZ`Z5s
zXlKuVe4d{E@LX*ao{4|)x=;+*Vpw)BLVhMZ$6LJr<Zu(=?*i@{_aa^zHXR*KUAP!z
z@{HOUEEBYhL~spnahKm5Z+KdmnZ;j~aPe8N!Yljbpr9F7_T&dxOt=EkEN<byD$38D
zAA!bxIj9&8DxQFaNHbVMG>fw2|1Mxv_U{Vrfn38|(!4f&t>M=W@dC_c=<C2)KIK%Z
z)FlJeeNO29gU2><9UeYR`v(ssH`gTMv8oG$7qXbJ3SD3fS#I75<m+b$A~E;*uM59k
z!xv^PWN&t5xJYOHfhT1(n$2{0u#+C_@1{~|GoAK^>7*Acj@Tw|uhU3-t!ipXuCAb4
zOk%~-Q1WPhvXPz~57M*az71bvRy~1J`q`%scO|P&y&~-Ci__E_Fq1j<$N;P}AO}mv
zWs7pl@shduhEqkv%0@ct0P;954nsx;_qO43s};w#^NQ<9Zc3l?V6xlZUdNNkHl_Qp
z;{!W<uvOv54-ZA>wrCuAVjB}9nWT5jHi6PtYL)HO=`_8iIjf}51)!1b45!jJs@t2X
zS*s{r7>=b$m>LM}_lN1pi{o^B(zSlTBsRUA(gh^*D?O7jV+(olQ>3DcVYnOWLf`Z*
z5h@S*LBHEmk<L?7dK%@S>~8uhC&oB<RKmI?&#rXMv2@m2c~f=POrHSQKqtRGK1>fO
z`vyC1T%^Hhnw~s=nU0Q*%|G~xUPe7wH<3Q(OBAYD)@QQ}li#*{4-T(}+uqG(Jg*Dn
zdu!gN^`;o)x}5V}7+(kOD%HEnXAQ3kSzmq=aMKsO4URRu9%Q4zn*iUAbyZ!3RXyIj
zgcvJ$by)m1+?96uJMy}DU4_MQht=@f_-Suka<1ZWKKINkv?xmv7a$$wyPjw9EWuzF
zTelb|y_5{mPrcr$jbzA9q-O`KY#8(EZpC4jEEW#1k_@?pg&nb206A;wF;B)QNgwxc
z#~ofN)0MO>`~9rJE$D-y_~!tNOuNd!8`!zzeH+tn3tI`Vh<nk0bsSU$d{G=OieWv)
z<R#pce_>tndrG_o1b5*EGwImsL5TDb3*QTbb&Mr?pKKJsTQ8McrNRS`9u&dU$;tUm
zC<L)DISQMTIBj#cmA?Ap@6+%8`j^yZMxTu$C=H<)*`OjfGwC>ss#2}kpu?&x{9|9$
znC(cu7(wVDScyZ2fwi4@?2`_S?bc4J);RAf_Pd87cWP2P(9JsSj<>a-bHUiefy-sZ
z#eH^sl#Y*HN`W~`gR^ex9-XGs({7sd$6oQp5>`*YTmn|Keg6Az)5Fg`PL*2OMvwtB
z_s;0_F!ZRy(;eg7XgCm#3n@kw8(uIF!6?e;y=B!&vNZ-P!d!k`l}bI;+U>Mmt4ayl
zPV6mT7N2a;+YZB3qfvJ`y+TSIF=d}jrL3w0WxL0U(lA(JL!!z-b&@~ISxrh|quy|R
zbFejrF_dZe#LQ6$BtAwl6kfI@A#Q)z_Y6zU-l{4t2C)&Hxzf`Bn8}HOi*~{pTGSQg
zEp5qP<+t&)=)5T_mBnmE+74x8wkmmh8%WjH7F6h;ZCE%vjgN_yKUJ(h4ob)@Qp&;%
z4L&mX=hcPa=Yy12mf1^w{-6k_4XvEdrL-uU1dfMNfK~s!UN4;<sa^D#DSIj2m(t;b
zy|jC<A4;icVXMO@-#=F#VKhSshi|$?J1nlnXYRnFn|Beqcz^Jx1_<-JxBxe$2bP}=
zD?L}kO}KAFu@vwsSuMwTl{COj@fQ7W$3>`qb#v=h`(VHV2HQ9+=EHyL*>R1G-j}l%
z;+eF(Bd)7iZuF1zpR9<Cr1aW!8ZlWw{V5%51O*cC3FzO$#oH2f9sPa0f2Pz0DtvfP
zbkB#{I|QNy#(<Y`;}`ygr|;PnzqEDm6fp{ZY~VHH$NU>mL{dt{@k4m={w?x@Fv@?G
zAM!O1bmw^`&7@w<k0O%(R`5eX8|wEK`H^u%GXKodMZAntu7Z{s<lt9;#i%#AufS5k
zt7Nqt=T*`GH^p1@e<fT?kyfQy8@|@?>xIRpkDFdA00Y)mNwQj{>g@^-4j!ky2ai*$
zwVOIS`)T*^Ak`Wz%N?O)lS;A;9V+Sl3*d^rGk}Qnhw~UeD-c=BgCegA*ZSjM_}iYh
z3c|p0DVdgS_fC4l^n>Ke<FkR<3ag(EYzS&KtEng1^yO)<*uGD4d3iJKwrc6oemgyq
z-qk3lbo}yp>SD-BrF1&HNH5QZX*gz3W|~gBJ<08x>9=2e<a4tA{^QH^{nMAi%huvS
z4^@W&(&rA__%*UYh`}6WYa&1|b&Sq};mBLdOMZwQ8SpR7<OsqsXi=36_An_8U4@HD
zc#l-J-Q7;w+(y@9wbW%g;H<AsUG59v(=%rpSGFZ{kEC;KdBqzmsQA8z3=(|o_D9m6
zLSd_vx25Mn=VlsT#QTNZa4h~kc_y82Al=3j?>17a-IC6<EkEgT(v>(LtS6mjG70+8
zQS({DOjg7=dD`uY*Qcp3J%LG$oc+7oX{UOFGl(g_GD9buoTsCc)70(u<#CBQEj7ja
zj~}XDAMB{kHm#S5fOIOI_{E8I4xhWG`cfNUwHbVae&kckcul2g)oa$bEVEK)lCdOt
zxn8clM+>n0ZeD$gx?l9mZS+-uTyM^qyuf`^4Dz`~O>w;~WV|<(CFgN-Udw#DjmK+v
zHL$w%hES*tZ-rwGuLrrGy${HG;janc=KXQ>I=oGvb(c)HjQjQC&*kR3;BEU|r#FfM
z&nvX)SqwR1p+W&ld*yq}IJ(@#w?AG*ezb2^eX+xqjb6xAtOR4sSI>-dpHDe{B@AJf
zm2_v3h<7=gD*T)=X;z(2J+2E5ufk`HSLr-MUzr6tF@`gTSLN|CKMI-9xEArfHJJ17
z6yvV}7MX@XiYgesF%LXUAcfZwckM-^<v|sSOR_N!{?ADQZ{aU+6#K>VGR)EyFGIu^
zHU2v5xk%<X*YH5MQb-1NTq!}E9TiWy^QG6t%`HX(Ds7Ry)d|KcV_U{;%p^w`mp7|Z
zE;lbk5A(+)FdHvv=C}<aIIl8rnBrOC5CqqB=f%_0mT=Q2v<`RD@BaD^>FYoJA+`2)
z(nt!^uz0bdz07%A5~gkR(M!y>5|vV=5aN^fo96{*|8?SE>^h%}y$T8c{&3(GM<^Vu
zE^0R0HbkIYF@tX^YG#HH<<oAqJZli;ja5h(y{d}KjK}`jX*xMNN@u!yCnvhPsdv^*
zqcbUjQhs9Jdlz>OD(5HPewlvr=Rc;_?rxf1vix<BrlM;!C{}z)3Gpf@rH6KmcCbEw
z3>ZMCQ^ROA5z{#Mxm2!*rgL?4<Fs9_rE05{O01I7Rj$^}53g7f{j6eR|9Hw7TW{zR
zZnjVCix!kFw>9~pxREz2=g1Gk7JKZ2JGL_sZLU}GjC4xKIbf{Bg5U7W^{<OPZoOrS
z>YNp7_<3tb<%IzuR$K`O6wxlM<OUy5oRz@w02}O7ztkbdD+Y+rv7m`G0C-TQQIb)*
znE{Cr3%XER3CGA6TS;!(SZbq~;!n=UqNn8c#s|W9Js|H?{&1jZvs{^cNe8gXZ`d2A
zqvO+bb}XfH7~9Cyno@4}7_{nG5oY%6^C!>JkAHubE+%@zgv6L(fFbgU1{Ck()u#vz
zn7aVp6O78lVVM6)|DmvW_k#a;fG^`F73IMVT!M^ndE%ZOqOCu3Q2l@$k>2eFK2YDf
z2?BTY6GihiI7h{q8H<Psf9+Cy{B_Yi0WK3|AyK9&g6wB^Bv#vr5c12|DQi$B93J#g
znv6;@=KSCZeEbqW__3jmHOd!$5wXZemwm2!TUepJ6ps#$xJKfL&P7B-`Sd>MG5Eib
zj<|s1@GJ~U8|9H-pm6%Y&BHuB;552BO%cX3xE;^wBi^h4L;xHX=>s>sD`=Ky5i(21
zZ^Y$i=6#hPq$9nZM&hs>5Mj&w&;!5RfSjJPDXq%}9_~eYgYhkLG|~VOH*gapC&B43
zBP`M_g!w^z2+cDG1Gl~}L{x*GwQ08}s3u(?4^&Wd)B93*5T|%ws8zq?l<|$<7x@LF
z4bTH+HpKFd2l_yzL{8zj!8@KKYq=dKV8#!7>RRH5zIO|;?;VQLtziv69mM+)%|71y
z=mWTKZdX#R(M%1=ojZF6>G1HeWQM)8yStw{y9bip*n+LB<b1~>9r$?H<6bmz<JDW1
z-#h!~YZmL>Llo;OM3!^7Epy5c40>z$rLO?X{WbmW<;GSz>5bF(Pma>_muEI^VNlua
zG}7T-C+%Sz!Z<UUq%&`eCw*hoPn(i+_d3n=$)khx$zdmLO!|`DH`A%&oNt!XbfcVJ
zo($7zcaS(Iw^G_lkM=vM(`nm4_WVe4F54kaV$cS;oe6(TN}M7i%O9EFTeC}!M|Yuq
z^G*$3=_DDPj|*};#xU=&Az2)Jd?%RH7^||_{wHwIP|#zlRVMN&uF|qvjWiX}Un`*v
zUw9JXXCFWI)`qpR^u2m1wb6Y<!(bE>)G&_isXV+>{b3mDa91&q=|1w7?ka=gzm#16
z<k<^7iw=>2q3!U&K{_}%aI+pruJ0?}S$`;c(B-7}pv!P60s~*H`Wg<$Va$?$y<Sc2
zb}Ka+)l^kk8jXs#H)OIO@s3W;Ty7B~T?-@=8^8A3(!=y^HlH@P(^%iyzUb;JUmM|6
zmNVsXdV1#ij=`~UDyzIZ?M6C~{z*Qn3iTeyCDe`D5(T*dmfuaD7JZiAIk(YQ1XA9T
zweq1o+!SM(65uY3uL`ShFT?G5E%S{!Tf^G}GU=N@$gyvpehse=Ys1&Og;(j7-X$#e
zS+CQl-4%B^{B<C36^idY!`t@Tp15DQk!>h<p4=?t;TKN@=!-`JbyVDO@<~OeSfmB{
z!V6G@D{re9yI*m95AHB_;l{|tVvw_LSMr7Aonq|yVk((WX^01I(t|6;d6_&zp2#m;
ztVl1u#>zYRpXuQ~F9juHJmVdMTg;pTYq%B6EU^r+*lcs-iGfznmeezYJ)1N}2D29$
z7avPkHeq@I1I^eqAn@#dJM=FBkO5W?9|?t5IT1!W!pYi^i#vY0XmEjtw2p3JH-@2j
zF^aU{zMLz&41u{!L|6%6BQWP$q{2Xk-wbiYV8AN$YMva(!wrz*EZ>1e_jRDW3?E`R
zK%wb;e3<^>-~QY5=|G4MRsaD1^hrcPR5xFyO0#B#ikB8EHgCMoaw4u@O1+HwWBa?<
zCW5Wzq;OqG=_{3O_4aHxy_gh*a11z{&vkry>inAZhK*(@V$lgnajLSqYTKPlqtWn6
zCh%fy=RjyyLvhIRSRF?9=r|ocf0}y7$2NNPm{EB;@)=G(Y&eQJgN^b=`uZRKlz#i?
zKc})3p1zb%w!D~*IE0yP@h0Mycv4vZy%h&#kkXhr360ENWOWvP)k@u4eV}tp#jh>V
zTWPkvrC!xWv8oiwk`&D?@lOh;+_C*()C&R$GZHZ{jZ{XcrXwZ)iuj0e@lu^#?Ad?r
zWGJXX3$0p;tm*`R3{xE3j6n*c*JwPjGQ_M->cJha@G~SEX_Pk!m6!fqK)+eSOMX_E
zsZ+VVSH19OBOK8<o2^!@IgHMDs*c{cgp)z%Y}kwUR-HLG-T8-cjB*lxTM9jGCa>TD
zwwQUUKUSVW3rgIk@NuqM-f9v9n$SlX6m}v0_c45(VE9tKN)ZKTqghV77`}Ep-n$>e
z=+hscr60e4=7|=b4_XX@8)1+Tb>#j)`C!!MeO>H4G7yD}xX{m~sF9SPc`)sZ<o?W>
zPwvI=P+;)Q=^Q7@<`lw%>s)=0gXNVCT?Xh9#IM?bT}l}v{7z%mLYM9o4U9I3m@v^%
z<m+c$nN|#BflYpyMj|?p_~FdB-IWt89Q2=INF#!7MW4r`vz6O%XSkR6p<Dc{VGRqQ
z_CXVg#%v$oX)$g=OYci=(?u%vC%hXxIrR4TMXx{N2!3>F*WMSu<B=apUg>>_B+eth
z<9EL=ai`ZOLi|x$h$Kycut>|@G31x!nhW@$D8y4Anq$5XAlEgl;im(CPvSiYIfDWG
zYN|ClsZwuCKG;u>AAg*7_6}34bKq?aIg=GxiB&y@0*6z}4|&+H9X?pT{ce#*z7`?B
z0{Q-$5Tvh!n+CZlCw)+cEMMIOi*a8Ee(8p;Wq{w5`CN59kiNkRtnYr59x@Cg$#w}@
zd?VGX)zoRT>PdQy<a`V{-9G0`pu_ZCovlVKeI?!NlLxKzaKDw#PEXRy<Fiz2?+JHG
zr@h#EoGPX*G9m7KL%4)vIARiBd=dCgF<8R(Jf55<`Y?Q9z#%<Vk8Z_9L*ko&e!+Hy
z04o@wF;7CH!J)TEg!VuTI6TNB4<@ZK(Qh&uT6bY$TbBEo#muBSCbjMF@1}<bdugv-
zP5V2IbfAl~&oB<1s+^^*i}dN^2kCIXW4!<)-_e=$7SpzEoH2;R#5r`NN$T}S>FliU
zEeM%J_{m2PQ>)$ZmvD%FqAGwT7y&t73<K0ux*YE-7du8wMI&1oHpKg;bY1V3A!UGV
zDO*kHZPL*=hm74#`1rH6Jp9>2XOUi1tx#50cQHBiBK6p11btCF_ZWu_iW|y%kY2t#
zF&)$cXNGKTU3!HXM$Da7RXQ#5hw!^rZieM|lcz<Wwc%?uEc5NIK5z}U!(H545MNh<
zKaT>-p=)?A;4zYWg0<o6{$aV#dYwM)uDEY1^92XM74pYjaC^TCuG{Ib@#nUFzPR56
zi)k&ztK1$##oIvzhN5FKUqmxcj;CkjBFf=0TKO~fX_-iFWHS(XT(d|E4lwX(_PBuV
zjHffU%7mU6Czre<xs3@##;Zt-Z?fxHw|nMsVXr1rTvm}W$%y=jJC;0gdBIkTF<#3!
zEI2cLUWfqg8AwP#{Jbepy0^pdYXA@Ih)ZybDsa1!I~v}l!(k*p#53S$Hwo<{h(_et
zis-l?jUY$Eki&p*tY%mMJ=!Pd<p=%fImaw888=V)*8;unG+trWGc*)%CGpJng8`Mw
zb%1dH=<CnZKmM2hqqP6%LAsPOG?t>liYycze>invkg_FOFCwKyx#2)yTcf2+OnG@R
zNWeFt#b|_*5pCcy4SKzFa(ZfotXg4~U2H``m=9F0NkNo?!NO0r1jNsVui?OXd8-V?
zoeZSV937{Z&!48#=g(7DID04kG&mdCP(<GJKZ{UmRnqVN&0o{EfB1bIB+Y?_-Ci2>
z)wu{ihA%n^whoKVO&10hDP6R?PzGZw5w;-2e1n3766pA99hG`L)tjAE!)RBks5eA$
z+)U+a*$N*sEivABFj5QzCo93+&S<*~VhY%29_5ud+nZ8Ub!W94TNrw~MCC(g3@voH
zHmHd|w0jgtwl-vjVZBzjALn9uD@x%hiKbFn3b4{q4mwFz++l#?wZYiM3{wt#%!V`S
z2m=&~8}BWv*igW6+bAbGE-$0grH%NbA)>ia7i^=#MV&yiXF)ig@SA^XhkEuFENT}d
zM1g0^7W~5KrgUmw&}e>7spL?|sPn0mP!5}B%geLlL7GY-CI3dNoc0cxxxJS-t8Aip
z&!0X|PriR?W0@6Vm@af?G#5nUw3O~%OP63#nRJ}|?8E{+<6Q6t1&=v;1Q(IG0V-6{
zv#o$4?iH6{@nHeom2hstJ;Qtx%}bE+T#3Gz39wA_4Y-R`0eX0Qeym{)x!tUSdB|cb
zT5N9szE@hA1JwV!pD!@H%KIYk+@`ZUL(Dtd^o{S!RhsU`tsgs;%FhBP&oKjEL@oT(
z*QyV8toH<&vNf#XCjx!v*}AzR*`?k}wR$tPJG<$@qsM9Q;8AL|cXUa1lmF&c*@M_2
z#{}@|Fg|Ws*dBVTuDn6f#N$2Vhrs1mjDU!D4HWf;kibM{ih%612=*t!s>Jxe2l(|F
zzSx~$BlU)p^!>Br^!<~UsW*_Wqc(y3U6!6wsgxA+GL<F2vPC)aGCeUXozQ7)Fj8FP
z-u+H3ef#-G>Dw<Kr+RfW{fGblfA9*XQmrMOqG`Dm7!H_4e4Z2}y^8a{goB+AM2}Z;
zNtVX&!~`?W&FZrHNNH^NQaiCRNma$fxO2<CA%D|`*r9<o#Kb&wEmp0ei?OQ9@;8Gj
z$oi7C(P4Uh3~VvEF|Vfb;FQ_|I*ljjsr|8aA*;YR`|H6$Cw(OSazi@fS@$e$Tuf~A
zddybj?IvH}bavL09)Zq<0qa6}#yMxDQrQ#NI8TiUZZBWHOatj{Y^BJmt!A4o;5WV9
z;c(3Mjmdf@+a5w6Ml%M2cqQ0@1EUu^Xw=vaQMnDJGcmZt#6gUDUC{xrDF?h`h1)?#
zeMW_yD)jvnZqActd&54f^xz3bMr@z1b9Apwm9v*#ygao5&a2v_*KpQXr&UjT?Rsj|
zq|a5=FEMFR^#Oe`0c;MqsiSzE7k}sX=ojXx7W=NdAW|;oxR|z}=}Pz_T%%;^o*}RV
z+!SXS7Sr4W;J6vjG9Q-v&)0?5;p@ACx25@ALZ;*<iZ&H+8;6i*ftjqjhI@yZ%<_h?
zTvZtx{y(R-`M!Yiza2+k58j5yclCd-+n+5)4_xn}{_p1K+Zw)>fpTYF7;o;F7V}Ve
zesX*qg&)81MGifI{>=FFpTFLWEUuz@&U1xH#$il5P(@7e#6%yqb!LoKa*SjzuN*^`
zBV4ixW3*GYVALI1$mf^#CC7MNSMn(Gk8s3UW|B3<LyTkI$xF}dawY;iwm^J%JO^l5
z8bl8;^~+_tcOX|-&Ibk;^;PhzV6@^ypaODb`Kh2-4-D=&5trg1aDhm7iLIV_9PT0p
zkKiVXYgnG7<7PY{E_{qT80HYhkxYlFb|uGPY_WwEafn&uUp$d0{0gwk;em-~8UVs&
zq`C@$Bm7;Q8E#VJr(Zh=)7>EejX2!7u{(GH=6;GNJdysx14_w@S!z_%XTSY6{pp|o
zF|~F(X@F3#7m72OYBq2ry<8Is_wmB8vI`}HEh%ahg-bEQU~<;)dHcW(j5%TwFDlH(
z$b!<;KkM3f!lA$z%-Gt4RbRC#hliJ~Xh0Jyrcl80`Cs5<R$fzzN?D3p|LiOsJ$vRA
zUq?@$rQ;V;f=&i$GQ@yFmb}RGmzLRYoyYs>+rRuNee%T@X;b;qQD97>^EjvT&_yRf
zM?-beW-xqBq&QN3w)en6nZV6^LPQK?G^LXGP+=7ttH8>vTq~z-6kt}$m9~7&7l%6A
zPy}+(l()xX=A{))l?~&T4O%t`2|pc_F8qmy0u6A^6?T|<O?gVuvSO+-prrZSA1IU(
z7v&neOr)nQ7@b6f@sKYox8OSlGI;In2#trcvsi|!&QhJHbXzIr4e#+!%OM_TnQf}S
zp<bD#z<`kzaN%Ya!c^tOFjr<fOUj{mD2cqms#5n4E~6_B7K=b#v=8cyva`~R^3kSZ
z=BAKK8HV@%XecT122#e4j?U8QQ8x@<!rf?9)84^O+Ue~0yfsz?zEpd8^8M2^ReM1R
z2S6DdSpftP;SRVJ#~RkKhIa<}J-h3*eARf1@QSbdE62-lUo`tO;yyJfB3*XTuM=ju
zPW`mMJ^Elx*ION~;RA=PKq9BC28Bx&+1{?CW^*Ub=-Pjn4j+G(TAh7QAZoT+l09QP
z>UdxDE4d@TO>Vk4ikC_vMdW)^mRDf;6}utPtDsY!##YE}D<Nk`6uC2h?i+sDhOdH$
zg|LolduiiHveD~bf0XVqVpoDK%do6YViK9p*QFgvE@c8CGV|tENiyrXWo&QbE&YPk
zL?y+0xWAiz_w~o=tIr>$LARU!^S}J(^z6k+IygK?Uw{3D+G#WOdi`|P8|oTKt`EZ(
zXG;0}3e`1+t!kN-Tj*OC>9jXaJ<enn4amw2mUu!LGQQ}*pv9ml?H&1=vT<fFItvCX
z+7AXiPplK|7{{O+qZgCRc<)Ne57C_+FOb1W!z8?Iy-&_lI>3eGcfLLCRy{p9+}FLS
zvQMmERZ82b-K?aClF@fNRp|~&Z@M<qP`X2Jz!s4cH|eVMKUR6M1MsPIz%%p>5sePn
zZZ@o6^)dKyW|-;$wx9%1gWhdI_|W~>F1{*V03FlkoT2ZkLfL+ZbntX2zVWTJu~6yZ
z$47hh^u<TJX{TLx_&_{B$71yuk$!xBoL))~#c0VEi0yV${MeLEa+;1#dMXR&gOG>H
zE*%X67(1DC+Ktq%l~b!$N>v-cq+j}#itl&4UHXE{fqsWSe0Ix}pL?qO@CM+)xnj&F
z<6ea54R3?x&{cHMbLygC#W+sDZ8U@T_M9XA9LF;4%io5@@SDFIH^VaiMYt`mWmt@R
zTRh-R=)6z3jha04coQ&I0=NmhXCUjJ7bo8jyp7?Dwztf)jL*8nuMjBxn{)K%4zF6L
z@4D`9=jc_V%>BcvzC8N{rTp@&yZ$j?0n1!*EO|md<KoLeKNxr!vyRl{6|Y$28Mcdf
zj&DUw=0V=_Zd;OTyhE4d7~J?fh1^W&VQkh5Q>Lu`=iD*;!(hgEcdR^%Z53m?M&dGQ
zXhcqq&-r9@-WL~>f`a!5&Jf-ySD)|1pFGVxg|ce_?8iUEBFH0x;w!*sU0BS~T?w<m
z1m6iVF2##jBqcr1n5+uit|SJn8AP%9%6I4JGbupiD0bRfR`_FFlZ<UV1jTZEB13T$
zEVd4kUm+twn0;sfxU#iiL5L!R%NU490;eKwVHhJW1W_hMh<J{;kj63N1}LtYYy6R%
zTq21e)9e%<Bv672)xpl=gY@O^f1kek?YF7gs(G(<mRPbyN2A&FtT&(8<rhFNoYy;T
zq`H*5rY=?}jid<l^fF;&LGfV=L@djs@m70!DcR}}gALmVq7blp2}2G_2nq&gilJOp
zgtLlLg3(Ah@$ynv%~~~;rD*g{Pt)@!Ptx-rzfUioJWZ!Bj?=iy5>H-CmC3UYb0w+;
zNC|uV)hFrm-+hzzA3Tr(DrHCh=VNBz_0zE53!@E&Fa24OC%x`$cj&Dy3PVhEfPQnl
zV#-Fm(xzx`_~KynO_WEKt6Z&FY1GT&nUkynqeX$*otqmhTca?uQcAJ%L4iam1u!h(
z#?W@5j)xhTD5W+C$&WUPK?x<34iM!IzMgxAz{K=l%HKF>$q=m{sa95cp@Dj$zTi8n
zrhM3<fR2@I9hof(eiTtURA!da9!Q5WFC^|HS(WFyRi5$^K9yAdlnLV!G`ORN7Bo`D
zVP+}HD(#S2mdtRV9grVyGdSFd3kUT{x>(Vtg80tc9q%R;g5m4=ixahnGx1UNvYpg^
zD`|JPo!UD)sorR~Q-AU7MSAl6vvfX!{P{Z!8UP=9j^*t~a-7Auz(S8?$j`3=-~tx$
zy(?Iz^IgJy;r<Y52G;yo!%qe39#xm>4`=d}B^LL+(L+Yx&)Gzlb>yq>F}L>x>t?ow
zcM2i9Bj?5Q3?w5VZ)}xnsomMpwU?T#>^gju4j(*B&6ebV=C0(krsXJPy=(-qTv5nn
zH#OM&(pctrM}S)5d)!vS>}7?-!a0M&BY2`HyyJ!YeZVh&#TUGU@get5&NtH2m%a4&
zr!Uh9yG*Ff0m!NtmoRu){+E2~K|=Mzo`@$m`T%}5KuP9BrbU)7ODB1_-$}p!_KWn%
zM~A6<cAWn9x4#SDIDPrqN9nWAKUMloZxeXZm9C{bx<{`+_Lh2V?MGi<DshGvtCY69
zTFTq;jfSbO{JOoq&zRtRTuZ4U%{om{K1jq9`m*7RK^NYQI+Oa4)z!8z5@Gy8rz1|*
zC6N6IM}B9UdbdF#W#+MU2bHnYYI+4+ymKM{mo_g`qqdbE?6uNfhx4<9PdtUMtfCu^
z*dmgx;H3lBE2%Ep;N3vwVucq2ur@AA_u?A|Ux<rrj}efGc(WGDdoF*Bfvf^M6HhQ=
zu>y|rj7ALl#Gn%Vm`Z=89Bf@#-9AsBJlIL!e)XC1+VsRp%DlZD+uff@zs2y0zCxX%
z3$n^|Bz*mWbR^Xg6G+*v2qPw!bso{wln&f!sLyIu#fy>+Vtmg1T{qz>5%nWlcy`^`
zABb<d+h~Rbr(quX8j!={z@qDI0eHG4`6~jQDjJdqn~mWB%Y-h%tK>}jIgVx8-KQ_W
zV)$~{tH3h;MYt`mWmt@RTRh-R==`bRuJ2<0*1hTX^L9?%4)FOVUcL!AuS3gES%YiC
z*SiHu|2FOL=MB{9&#zuTls@nZm6>MVAzdsGxdw#gyDr9Aa~}z~-1?t+EreO#(%s*W
zB7MA?k`0jS^n7t1TPhOHWE;lhW1N|tw^#{=(JaP}ZTylv!lDosC~yQ26QYoTkRcfp
z{vZDP|Ch5UO7;%HSaa#FqPMRL6v7Ha8bsY>z@MoWMcT}OSdAoq<KvDKfr+?<07YE;
zAdG}y!XJa<UYRA_mRFo9@kqt9GjNu17Y*`LDy8H1!p1qGUv4t+BIQ7Mid=~`yXSa%
zj7#{68k|9tDN;K>=P$SDDyWF0zUSYdt9Xn%KR4laIRrD@^Q6EPGy+%>j^chFu$cJ_
z!ZVj}DbRp#)QKGMa|VU*srG`xdMN?^@z<ZH-~8c^>7&m+N#$D2jh%C>_^>f676TZ<
zIt?EK3@-^78yKj{mDJM<!d7~Dl@*5rqkLiX;+&?b?o)+#yFCg2SdpbjNFk^x9@~O!
zOG#qpoL4RBCFG^8D_@jHaI$I%s;g4!w$-+n5&7bWAJdQD{cAdU_B{1Ydg+|iUMMVz
zFP-WNY*$L@v)_G{KK|;9)Y$1*(ZKqS@-^)BQm=a|MYLxl5gZMJQ(-7lM}c9-OHwiT
z*lG{wS&bC&IKQh>Yo;ynX>(I$QGRT7QL8pwHrJBcl(!>9p;7*HaNr^gK$w8`*!nQ3
z(;1mQj9sLio{wFPD7NsNR2aL^xMI7KP`pXwC9aAOPM+ERgw<B`8suFn%E6yP<JA^&
z?=!1-D8?89IeZb^to*|uHBmiajEWho6`x}UZ&0SCIR;Gy4!1ruH+c@#nX+OH`pKVm
z0T_oAKHj@LW5GO9nZ2*Sa0BFR9_X2NL|HB|o+`|Iq?}Un`@KPW{`|-*zRoA=SGJSd
zQ914Gwsq~MMyrv=!%2Gj^m%&z<BN1boP?1rgJ|YK{gG<c{_Z~<LBvzQ0(h@tb{p*>
z+C3?9jYgO2AA$sGGoaE>^tAv323fHcEZ{FP72vt(7Kljt%uJ*E^diW)QwZO})4x3X
z2`Kt@U6;{8gtq`blPpTHNHfnMiT(p0fI2le-rQj)iHmr6Z|Xym%kJ8n<#m-GK%}?R
zKt1s2^O=@t2G4bVL>yCcCqIgELiZc-qsUiR{0JOx?r?b*Bk1uOKLk<Ex8q09d_VlK
z(xXS?$?;b3V+J{?<INp!%8%KgNyfW|0A57uW)ZIV!LWko;{SNOHhit&7Xp^0%BfVT
z`K+#HYd6)Jo%G<*N9n=g<5X+3J#fz%vdAA2!t^GvqCVaaWIOvC7man{o4f@1r7613
zJ7F=}O)s3gX6|uY3A4eg{J|}5^z}C)eotUN-bVZf419g!7bZ(*5WhKx>OcPNX?k{i
zrluF?5;J&!e2Pwh;fGxmxO=;JrDmlRaw~=}<X!-IH_x_3kEkf@(SzOe&2K(V^~#o4
zX0iLq=H)1T^!P#AKR8Geeg00mgXC?%x69I(F4ME;C+YZ9dX`8<0;O>>36Qg`!bsKg
z*<x&ah>k#!_yWf_J0K>2u_{a%C{JjwFd0wsJoRg17DhYC=4_3wVp>0-+_XW$*p841
zd>FmZFIc(7MLv{us&W#>Dl+s*-aB;2My;Z<Z={`8E$w%jsa)C;zr+*K#7-K}Hjv!T
z3b$HiJMFgX;;Z~9+hAy&k*z0rhvR*aP9!>6iM6H6JGh8m^azRvt`qU&O!`89G*KB<
zbQNgAgiq;RoXN#RPq=ALU4}wAo(xk%`syFPk-q%dL+N298_!-Ivz_6wdX#BuHOlGZ
zhkI#%e=l`M)Aa1<%!W$FL^u<hoh+yWR(+kS&R(9Jc=a4Q7n)zIQBC`M&D3sIZ4AX^
zN0IXT?+Lv2i)Y;Vo_&C2W3+X#8{qvJz6vZe!`D?Bufwb49Q??zO#5x%e(N>ib@0C{
zcoRBrf>re2hX1wWbKMY3^}lLKR^c%2d*VOYy$+wD<8>(cDUC!k+2IWUo-9+I@vRMC
ztAV!hF5lVr3()<`tk?HlCyV*vzOnDZU*WSj3-R(wvJ8aeadv<lthp?vo8^m4AX$)w
zAdXqcNJSV6I><9mSV9H>x5sd&v5G7fDX?;%vH3~*KmVWpkCI?g`mc=R8mLo*j>1S)
zc+G(@ZoWs{P`koGSdQSb6mB}45NvfXggn1w%EVz8@*~f}pFjQ?6f_D6Qrx5P-~*Lh
zCbFQwcmube6-J&6ZY|;lF*Dm(%w>Nr5-2^QuIT~4M~8#vh)~4yUAVI_MD~T#ahG65
z&w@MR2fo6mxaZ%(cQNob+}y9=p4|-D&E4*qV-E4S<hcNmE_g!)%xTme|9FbJfxoa%
z^~M0b2QY0vR3lZmdCb74(-rAw7YtQmdxxJtSiXlc!u!E$sNIJT(&5J+rS{&Aw|^iF
z*?Pvf;uF^^jO35sSPGsEL;AzWfl^}yNzW)cY=4Ns1Vh%j1bhU26u-0HnH57;En!Rp
zr&m)+al}|gJ4QK#&JtUL=#DZHd&!F;2u49s{{1dn+8ukjCjKaTY#qW&Y~mLH!_>h?
z57W-xUMiz>svOy1gVO2kIFzqf7O5OguM2B9Gcz&N*jTo$@(MpZfKR+DoEOH*E)GK9
zlww8uM+syLK#Y2gW<%8%{H6}z3uk6wz@psn(Ux}613yvN=!BpTBTz-kE&0Gp+5{`S
zsQbK)AS=;aXTn2y2IHam;hB_DL{Zk*&XGdLEKK1N?x-^<Tc#=Acj_jTakg75wnr2V
z)G1|Vwg9w~A63EGXc#<|xea#FiAJYv1z5Bq91#cn7(;U(607}G=gO0gHvBN?NlA8_
znY#Smf=FTdkA6#c@}WFlttLJP&o<1*-dW$qExI(`fKp{E)#_F6=MUYq#X)bFhPqTF
zm*4ooZOZjapNC)Wlz5jMs5_N#sknYY?HPUwi@q-KMLJzYdGjAHH1FRmU&j?FX^|?D
zk~QG`ia`AYE}Xi9JA8SB1pNn(pgrGl#NnHY=gG+tU%p09Qamzl_a_Uu%?~{@gcIrV
zU3s`~A`kuJX;hStiypxNzB?SeekKXe1wV>(gz495{C%OZFVgI|gk`_LMc^y*gR%&Z
zV;YZ>>*)r5EYrNq562~ua(f$o@MH3f!RZKF&5wwG6F-*gkhmaK2_xOY9U49)ek_8Y
z^fUKf44(%q`Wgr-oHv7FkTI{dV+UP;U5j)s!LrxlbGRv|9fIr=zd09vui!4+i(WHW
zNyJTw>Fdw`y`SF?+#n~fVGY+|_KvS;(lgl5cfVAT?A1&Mk3LNwfBtp)^z(1hSKt02
zef{kp)8}7)lRo<R^R&1BKr&cIGGJBjGW}4zU-^}GUsK$jpYx<&nm;aHGguY4>T<^w
zmy6|$mjM4;AoC;oK<EWo;eSWk_W`D3Rk`jHDED30I{|&a`0mJsY2vIuPTxH_N<Y3h
zPMjr$jO`s0B(EYnBSRzGv1K4>kX3p1jtUc21L+yL6}einGy@yRq)5$t-w21F7)DMo
zW+}kpA3TNz=^vaMg-(KukKt%rx<p=Og#66ECof8lXD0=3uLo@kr|nyoM_<!@KImce
z4(A)}GFVSxM~PUyMSSZGDkIy}J3eITVi$;(be&G8EqS}<_!wzC!OpmtY`1Nl1^UPr
zBNN*T#)Lo0fPqZzP3be}Qt(5U4b|{U8Nn;ZQMk_<M(<KM_*m)1N-p#icAv1(uh&nf
z(g{wxeV?=S^0=FxK6|PAsSSbPp?<CR=tB4C_u00cGt6QVrsXcx8D|W4`vafV#V$7(
zFdL1A%irZ}Vbu%b+1^ekJ$!%xzES9zDx=E1Ej*=CY**h=*zRsCHJeqhfGcG^N5!>q
zP-WW?A4XC~`}&!9@Jh%fGnwzYke_fx8?*Pd_s}?XIkoQDZxNP~T#1*j47q(S1O6G7
z!`<6DSh(zPNqi6Qg-)`#@~$dER>ZvuIlslMR?+pY;eF(BpYXc$uZoY~trveQ{Ph!c
z`U<d$O8kuH9$*o--AmUpxJ-rnRbd(DeO=4&&UtnVzT4>drNCX~bK7{=JbzCR{p$QJ
z{ZPxX3f_0U<ot4;i_cfC<uIP}wdlVPE{El+7!$o7&fVWxRb|9Mh7DOJ#|wD^`Ao8h
z;xM6!UC9`WNB*eQ8b0&5*=$Re>G%xf|L1@AfBW}07{0D%EQbl?F*4$E#NrkNATU8(
z#EOBQLevrQ5G;qK8~lotans$giidFd2{&!n`STETI5S|D4vZ1Tz4)J{@dI$hS-czy
zk2H}U-1#|T&%?405ceg|4;oZK4CGN!l!rQ#!(4Ltigv(*>zU}Z%G^Gp9hgtXfde#w
z$N#vCyXyvCyCH{iVSnL8kk^B+Xz0bu;Y^i}vsdeq$CiZby^kNI!;e0)k;dn9ITRZc
zhA+AG5QZ4`dXH`Oq-i(+Lx*R6C?0v)Sfdw*{KCL8P2Fxcot~bh;h;aG5B$)MafB5<
zUV16z0K*CjMOg|ZYMfUGDG}+MR@Y=WNWJb!>Yts;&AD8nK?=gfM9Lf*<PvOEFnm2q
zJNtV!exSgRS6osEbz%I12Gm=Kvk^#&6LEqY3Sr`(%EkFz>=V!XNvDDWR+a*WfsM`s
zWs2}H5QSx(GRF!paHDjJE|t~wD4_mq1e70q^ol#t5Hl^MT+=3$2*Y3)s6@NZT2p%J
zL1{4};dUM(*2Xk>m9|uGD97^0k2=FJ1&@JH+SxKh`FW;f$p$@I9`wgy-OQGbEenHZ
z!9Vyxi(!iwlx-Rn#wyAeZ3AOsDCAbmrJ$ntWg`@LM7^kh4>99__KK1WZ@It+PdG1u
zJbm!I{!pgrDZMp`F6xhc_s0X2a~DnRd&@>I&IiCahce67CL<}mBF%Z__s#D}aq&F6
zXmnTnz#MLo6@K=br}uli@VtEKhyQR>n!@ct7Kas=`000?OJVSVtbLos!=2x8{`4FX
zW^~w-u<#~={pL90!tW~G#0lI4`meyv#b1D2j*|Vuy@)r0=6>;<anHk;Rc&{XNdLZq
z=0*Rwc)ZS!!X2@%^26z!<1FsY{0PI`+>ba*{0L9q`3vBgTM=fjB98rX(A;nFis<29
z%@2NFpC1JJR}pw##SeZcqrL5(Uq1(agm8{qH{M<KAoK%{RnSMcUuzhlcW8#+<(|_^
z{D@eD7Y^rzH#5ogM;?B*JHnj6q&Q)unc)_=5=f-9IQj6}ZA4(1;r4rtLA^+##ExUY
zufWS+&`e;C@cOVee68W<3bXfTxq0gy*p|#vkutXb;N$ekm*1w(zxrMJ&9{F}U;p;c
z>66dDNr#U<PMw_t8@o7b6{QYo0c7!h%I$l+lh(0{dq_@P?+)@i%ir|^1q4a{iTF!b
zhQ-thDO^7jc(-i36Y}@#w(FhH6rHhfZTRe_@1C7_i^IW?l}?hCB}ZquI3|6m6-ho-
zpp6+kQ&T>#7};ELEZ0IK-yQi8J$7eN8p;32#*$opzSYYk=}k-=+hmp0D0Nw-qzl=d
zNpLZNF7yawZBGUh4oKl#W#^Ah=R7Hk^5K_fUFaKJ0Qw2~81g^oX0cTxyKcm4E(~C7
zLym5+>9e2P?N+Kc>eeN=^IWaR4i2nVV=F+mOZOSWDxZyA^2Zp5tj=Ue3``Sm?~Z<u
zQeSfUczmurV$~Qr(5N^+j0t%du%5p>Nk_+L#V#XE5|r$ZZb8YYcX%N;hA>z#J>T%Q
zkF4}MJsqY~@edKARz=SYgB#}?Lnj8uDLM&Ns5S;KnV?szNoP^n=_C5Xv5jezt=*`l
z-FDOU>Ds3)xIQmZy;gP}3?`wga5O-i{f0i1Qmaw%`Chx7w)(7gs&h^maq4v&_1Wg}
zOzrHr%l1R##E}A1w-<CIR<$7$U_4cZ0*a2C^0@umFYW+t!WP2=;8~={KVLI|8G^+W
zK#l{xMOeli;j}y2yZ_BwWEvzly&RTt=k*uu6d5P}3J5fY91?g6zbpCP4Y{1JbNv*s
zijLc^MYt_2!*ZO(>ve$?H>PsntKnW18p&@@_qLE}TfUaT{)PKhU>V<iURU47`{&sr
zERG%Mr||t{0i`N>sFjzysRcfm7at5<Z9N|r?XL#Qb@MK-+u*M7SA~^%F5<>bn4SI&
znN#1&S^dgC<iUtDdv9m0xOmo;-HgrucmI$7>%W(@d0nA*Jfs@^^C4$o17XRTD9pjW
z<YJ&hGAA>rO@agE7HtejnXW=0hFsFj22Ds#8Ue`+Ge-cAL45HbfTqRg85TT!MG(HT
zD8Rx4&diiWT&r=*FCKn|@aB}cn6u=5@_(L%MH{n=`gBn7LCSnnHM$KH4D%n#xSnV;
z_FY(6GUIiG#NRBfG0*+1)JM8#KMwYDcC)&Qc3P9*{rIy_)8V6s9w-nBlK4WA<Anh~
z?z~C~6v(oslkzzZWk*qv!lF0`{yqqqtps(kbp*>N!-zo}X4YdURM6@bOcL4==52V9
zFdU_lLZ&=yM3ExH+cadGisR!n=%M^fjQ9LpG;*Qwk#JV%l(wae9e(sEH98%a8G})5
z0m#Mvr{D&MlvxruZGlTZE}#7P!;Mjh{psy4ltJEC%3iLP<yR<*l+A3W)4{4PPAV@e
zX=0#PLx4jH9(mGFxU35ERvSu-!3kV|6(6<~gkEr{459QyeaH5VVE~$lC#ZtuZ4Rnd
zdf~*6yN#1loJr#uls2ZFyL@FSXs#F0MO=W@Y$*7wienZjag%6|zi1~;uLmn*s-^Qw
zJ-aHxSm(33V5iE5a!tt83MR%eav`QtsthQJo()TME9Rj9Q%B^5vP-$6hQ%+{BZe>9
zP+x6rJR&~_sXn$-tx>f?$+nZ{<Ea$#o)q&~Md<!i5o3rV03TmU&qY4@a%}u`;l?+|
z&Ed;{|7{TX#LEKltgVr9;hA8^$KO|m1-FyL^R-xh2wdU14moTN3(s=SX;=uE`4=>Y
z^9o$28CZA*LGgGcrXArk<LxXk;FrUX1$dYIpfk^p4w{E8#Ja){N8@n`B4h!s;lAdy
zV;Q$$=7^9h9?P)E4|S>*)a(`?^zzZq@)G(z%wb#?aue=vA{>h_>cuPNb>};lLyQj`
zg?rX^z-?c-9bUwtzcU{7AVT6ex<d{EV&Df4>2Mc!;|N5iKjZVAIN;?@vF5m=ps~6x
zXvR19T%?Tq+_f6iU#|^cYxwzt+J3oOPnCKn?LGJ;eg4I_>9a3?m%jYw59yoV|0R9)
z#n<WJ@NsH)4pLR|FpO<THj8(i^jy3*#am*xT=u_y(XRkEMTwuetZ*k3_yAm_GcNti
zFMRHXSBv!%!5s|Aed()QuToIQy9-{ql5PgFK|aNJ^5UeI{`H3!Hk{yxY-&9~kwQl(
zWO?#M&PGpA7_u?OA+LCnZ0?C*@<+V3AtG|JKk12!+>cR%iCU~onocj%STZgKpNZ~F
z=4#Z#D8+=iv$JkGIXg8D<a?xJIjsX&j3+Po@yGYZlg30peB@3+(YerPFodz~B6?0t
z#EY_#7hBLbn~l`oX{WvYUD3emE_4Lx49M4#>6v_n;S0l5tPo?T2)?nB`Spz4ACn^4
z>4M2@=tR0t&TXt4jU~JbDc@C$eyqS^)f#7d;qFWS!vNUr_I=jaQ0ds_k*yIaW4;()
zu|ZMvQx0~VK>uV5#A=O+iPT3N?jAJ|llI_0hOW@P;6c=h^-|RddL{-tbPWuI?WTBe
z8QbL3_h7txadhIlG%z(3@kMe7f&r5V!N)=JBk3aOg-jG>;v~k+bI~hr8@l=`Ge$s9
zc2pfJ+k}>OAsRNMLooq!o1IZqm(+p&4ODm0+s(k^2n%rA`;)_LAS>?CFY0+AU6#Sf
z$?l?^5QTVLoVj*RJHFUVBUKW{JTX9i8}ie;0i}u~&DZaQnuE$ud4*NDnD80+lnF0g
z0nsM$446CTe!o`8ykF$c`vP*jG1sMlMLL&h4cT_-?<Xiv2Isv%>rY3?E5IV1@0ivr
zcy%qqD!*SEP(qjUhU+H2yz2E+f_2fK2`wKYEb?ygS_P}(Enlw!tI~1l3+cc3PV+}x
zzTvv#cIx=1M(muT_#@p?d-(7EoBtZa*QOPE5b=?=vW3+OxBiedxD#P9G4c5RM-Xx2
z4us$kX^Ie}3j5iA0gBMqg6r{DV|iV?z`Mcnda#^>`Lyg?=)l*3ctzZWB;gD9=Ym#v
z&T&UwV9=m?s2hY{c`?vN7?&#<fnRs((D4Z~2FJ^P{x_Z?+!sH003<rza$I8#-xX+(
zA-AOfbPo5^C!c?j4h{}%%m5c;v#KT=20SBB{+{WtGBPMYF;D_LLr}^>7&aunZY#bb
zc{{>hUkcfoXZVpJ1`C9KX298i22LSjyE?X91Sv)zlr<Cr@+CfTSOGN|57LHk45V<J
zogSx=?kJTQR4{y;kHR1{=Se`VRZoW>an?gaLCh?a04_z#+bFUs%2C6xg0|@Tq|C|}
zLlksi4BC_uqzBil_iRV`km4hB#zUFd_K<z$z3Qu6wgSb>Oe+Phh*-r1pI8B9qY(wr
zg~AF?P{e%bvGQZKA^EeK$_gY(0z4v($^u5(AOJ3mTV9bT8oj-U=!Zub&B)&dIngs!
zyot}{V$c>IM|(GJ-T4ooPE~F_voAmCP?S;l;16|9zSNP8X{s;BQOck-*#HVU%7+4M
zzKb5*UiGLew=*xhRiTJ4l*S8#$5XZ!K>>@t6GIrZM&oA183P3U)b9?`RLUakfiIOR
zvr$>`RktELl~zf+9Q84L!I4N`u&{W#3X9L#y?kE={PP9w<>!2@3eVRn|3$x7Np}~H
zoBVEiUcN71tNd?@n=iuim7ka6+~mLL_bTb`!f}(|P0!2s<!hDyO>y%@c)s%Ua-5s|
zfqY+th3@c;U^$OPcw^i}x|?vX@-y(-=NFuAkVuldFmeS~@=ZT@(P7Y!@QwDL5Gb9$
zx8R^3Asz*xGv@<0>x2jbHweK|#PgML&x3+CxxpI|B5&O4C)FWG9MXf^{a+E68$ZO)
zeVhG=u4l^@avQgO6)xJK!B70fJut>S-1!P$^}lPw*BX8X5N&sU7f1%ERO|X?x6<Lm
zkG-nv^RNGqe)G+ryt?a?&%RFk4?jxH_MYN2B%hR(M(x@cvMM<2<yS2L7jnk@-B|UP
zVS$>faAUw6h>|Z8Q9yPfUxpyf-M!H(!6JP>H!$~9mQ|O--Tyk`-PbEX#(OKgDr0W}
z%b3zdx`Xrd{gapJ>Cu@DP$IzcDighsbA6td<aK1+$QK<(G$P+iHsyRd%b&>I(q)jl
z0d#=aStEEw+hEcc+rh_bA&e%hdYVXHmmG|dN6+2~LKkWHsy%!0BK`2=kLmdMC=L43
zxrP&e%E$&;83(ko!l_)Ap5bqjWOs!*Unb`1j=seX9GSmS6#eQSx_owVz3TGtjA6@X
zeMzpiaZ0jvJddqs1)j@Pt(0x7<2#0N4xCJS<NL|y5BVQnGl5YLrl-mzLg_G`^j4G^
z`Y_P3RXTAd6X}Daaq9Mm>EyJVPEJnK=~>rjfA!gxuovfbW#c3ISlvc1&My-`rHg<Y
zJr-T7*{q9Bj9$uH{6f#Nex@=|UUVwcuC#2a2tAxl#@S=EU08+j>a;sd&tIO3u90XN
zr4!M_YBdgtr@c_dp2BRrlg`p*OGb=&gR##Z8^S-)Ozon>F=>xoM7E@3v2AO)Qu8x3
zS5%%FyK;yp+q&2?-bPvLlhh||OL<b6(8y0zOYt0l;R)~#F1lh|Lv^7t2F-DSn|eV<
zCXcKKvonczCozq<t7%iTPBu!ZJ3LP(J?ZcmBc<;Tr<bWuo59$rIv<Ibtf)Lbixqt*
zY+28HCWtmfT`)cXee8}Rom&r(>R9gg7u#;|7;t%g1*la1o%_xE;CZYyya|}Ai|k#z
zUIU!O?N;~V`l-P*71z%h7U_J~w7O5wJzx2~$&kMrA39Ku+pfFdr{cmYT3_c{!~28#
z=C^8eseVHqn;XC8`Qt`2ihJ?r{`czj-~KoM^}pYc;Kg@W(e(i?aKtB6AKg2ESl5fa
z+V^#U)XSaGGR)%GaZ`$`{;z`f3S7yAn;w)GvXccz`I}b-;{rE>@Dgt1GNbiup+@nY
zj)ARBgdN2)4&sqRflqE%h}<3+14j|5h#hevHV#LfL&P$PP*UXKI7An%UX3TJe9j5w
zxspDEQwf#=ujTg+0DLndbG|pgLKKV}TlI4K_{%TSw}1X)`smY7)42+MF8|3zY<-8Z
zgqeTLQp>AsJX_H#rea$@3?O_A`%*4Yg6dL?@b{TsrxN_n&Lq@#E%0L;!x)6&E9zv!
z+dniL4GGW~f?_*7UO2YHV<sHR26fr*oq46shQiN|UZ&HRFVnc&OH<|3KkcRN$skQ7
zTr+zx3?c=n%sY<`(;xr&uW9FCUjp_-*U0p7{@CS(E|qgKWb2Bd5}>fM+6W<A%9e$E
z(NMh8v`2UV$Ury0z?ot+N{n0x`53y&qPZ%?N(gk9f>>?Xc;!QlQKE$PT)eXZNYDLY
z%w*gWFE|$sKC%jd)o2(CQ9w}e>cWc>#tJWZjDd%Oc{ZZbfCEDrZM`amh%%FYr1D&-
zyq>8kK7%`Grw*mG38xK7Dkp|J3~O1bL{XZIhu)?Vez4^U&4+WuXcwe~H^f6hn@-2x
zrUj)awnJnU87tJLR{Xt}K73XAArKm3YnEjCoZtDx?Gt`6^O87h>*(_c)HY((m+sWB
z`0w+M<Qz*`MB!$}>)A;!jr&p>l?~NaZI#o(;eOiJ#g>#kwZ~^qo~IKjrRQUnA8M4h
zae=J(rFGIqxzQG=dHM{Z<-P-cdf?hg`3eB>y9kaie?o8>Kv6_>UU(Gw&2i{%zZ_$Z
zBgi#Hazu(59zSz0Xy#Xsv=6WJbdsDR;xFTN--dhUuiusYh<Fi+K;rQn{K$9>flxQ6
za7efWa?JTrz#H)cSfF|4!~HrxE?kEZBT_{CjN9P2fqU+&`;Gj-(}{E^9?`Gz!(RTu
zjRR-ST_DGGe(1`H3?wF^bByN&R|eb}NA_R!yecgF$2ak6JYX>_hi6!hzl?VomT?2w
zZxPpWoZzuSbEF)@b-;ZVpxc7Gs2f)k>90V>!B0-hk0Z~#)EUXm19v<MH+SKWHi7>l
z7-?}U0;3ViPtW$!a}K{ov#{!Jnu}yXM}83C)OXbPnKusbC1%eT2;U1V5VnRjxCzH=
znPG!-ZR)-nt(|oE=%dux*-LwS`)POYFx4Av$&NM2POP8^9RVcBOm-H(=ndeNq>$g3
zYq0E}aa@5LS}d>Ut&RCgp|0YN-kskpE6m|moCQRxSBd#E14v%M;#=XaVix$2^{Mlf
zuz=%=PH+p%a1o!jBq!%AegE_{{pY_w^BG^PE~14Y|K@G%wxz!?=uf7WdnaSbyR2Y2
z_c>OLMl-dVt>o?c&Q7f7utNko4|1;c9A#}?Ksp2~sXDEu4KPe*V*;ViFs8h^>eW)(
z>(tZ!en;=fZaS5m{{0V6Qm=cOhSHP96Xa`*MzK=_I!dRrn|5~EQd_oCuRpMHft6EZ
zmKZ5tbP}JXRnSB}umk;vcFeZ;?EbJ_ij_F%62lROFV3*y46)et0iBRGj$QyS_zbn1
z&D3c(ou)VFNjG6~pL8+h&3DX`^pubF3gy$3zOuj9N;~aFDsA(<Lw7_^ki4%v#S*U=
zf+tMK<LoWU!m2Q6XC)rhgj|nY#UxL!RBN}qO(O=YsgOX24}VwW*o{Q^MHgp-VKC!-
zu9$QfJ7quv{9$F<`9$?vE2Tet^+|fL*R<X`RCxv?CgDxfvGiZ`AW!O)j=`!uzR~Sg
zL*;7OU=&;IqpK-TrNsy|QJq0Y72Q%esTUi*bh*umc62|&>t&3Yg**Uy{n5j4UK!<s
zSJVT>GxS`qZi|Sn7oVl2=Lro(v~7#Q_#@Mm)uu75l~fla>D6o%iJ0M?Oo*qPLN^Y`
z#z?}@?I`bHa+!{MqjX%Hsm7!oc%RQ<W0hZQ-5;w1Ia8bZ6bX6{DyiM5r<Q2&N;&Zn
zT0QBJHhQUYDI$|d$vYcH86((+M+z+K9(}L~rzKE!VYd#cyMy#D^R5!cMcL;N{9Fxz
za}D<ZQ9W0{kMh5%Hr8-Iu-L+vdB5EH?+Sk%h<EkPz;wjRy&8De3+F0WDHFa1yiMvi
zDff*s&}##EzfG-OrGHM>oW9p9o+bS?+yJ+=6*!*9t>$&gJ>1A$k=`-wm%%ZvdO7a@
z;lKJ{Z%9Vm%Gll;KuZkLMAnRTEPj|hGgyG1jWl}BWsA7P(H(}^gAY5t0Fk(dze9L-
z+`^x<3_AO%Na;8@fOv>Jc#<FjgFQ1D|3YaMuoYUwFOrL%Ff0)-#1EoC2Vrde=REc3
z{DUanNk^OlH{swh5&l<D<o$C1sDP%caI-JvqR^Dbyi_zwI;%tRibjbO9FAv~f=I^`
z9{5Wj#UBB-BE{+RZ+?@$_~x5*@X_P6g@H?d<8u@e38fO45wzJ)p5)9loJiQ8NO?o>
z%(D_PLiC1xZ#l;v=ENmETS1_(VE{vUKrnT^(x@aLBS2S+S!*a}Rz{QuN@2r>qpG5z
zV4X_=VfJ1(4bDze_v9pvx@T#MQX(Pw?D#C5zUWzqzLc;He=?wq?0@z+{qZ0FA=NwW
zG#H1m2|?RiBu<Amo}Ei!I+q`f62q!_CuOWIB^2Qv{%%T9fsrT@7zW|<w&;fsn{4rj
zkxR-*oB@U)&dRnem&0e4saz;b<w~5d1s&ai_$GWDRE;9EEnyw}7{yQ;@oT^b*rq>(
ze(C@|q9jsSZz&?YC>1Ep5YMd3f$*d3popSCqjXw16;H~F3q9WILrN<jdW@14DR*JW
zLTO?T{=ULOsYf}m!B9A$i!#6`=s8!Lz&J)7JDu~^+o(pVY+#$)E`K%xiY~Va@eF_R
zR08MYtq|qsZDDA?x~6KktoB1cLutA^4_xw26DeeDe|f65cB-~8^lB~)YMZIC!{ES9
z>g;qxUoG{!z4ZLavvl<GR1I1@M16yoiceW=7!uHbKm(p8pZ-Jl{BH675X}2k1caZQ
z8EEDJxct`5o=K>1dAaYBpPc4H&JRC7;HCGA@L4=vEI=`-x@Z@8@*LDfyg6?9f`_=e
zyKIgTeF;D+`^X{I-p&u*pvY<D$NiP~Q4}EfA(-JNp@Q`IzWI^kMU3nGz(IN6njeZp
zIdB5@Dhlwf_+gw%<ZpwY@D2ZX&Jg|)5>N4E&2g{7q93ITcl^%7GX8mL^Lge9n1^0X
z5qRc4^XGXeu$X??y%-y~$(<pm&-?`z<1fQaIF|EROq+3YE#?E{_=~XUx5}N<FXD}c
zqKj|qBG8^_OO8}Pfp9P!WYq5i7AaW68Ui*-KF_9BYo%s;C+#0TO1pax<lar~-NRIC
zG%X))GXY9+1#Q_AcW6%Zg`g<zd7y%Z#;b5!lUp9tc-<GleRToICK>WOS#;wDWUV>G
zSK?;MZijfg7js_&ddO8CuK+Q8v7W9&POmsPiyQqZYi^#M^wNL+?qzy@Jdj^8xK0G+
zjlqUCh)jzfQ<lt$9LwY{<Z+BK$0w&svz>N!c04((E8PXd1Z{zxDm=kWdV?vH8v_UJ
z9+|z_sHH}uPKf&KvE*?s=?2rW^p1^mAldfOgMH}~73*C15BuHJ>teVWOaJPn)6)~{
zavy#4ar)w`&r++^PNT_1dhzlsy?Akwx{{sw`m^00yF0LY3j+xg;xJ^eQVQBkugYwH
z(P`a5mkklLOZk^8+o_><8l3?{1}nedR|^A1v+jC8f50$;E<~A-{n2NT{i~vZ?cy<9
zRHPep8dar}o>^w48mq9xhruA7cBNmkvx5j@vSXhu5eFlmzcn0BtP^p@mv<sjUK<!e
z)Q@!9Z5tKKWeiALN-JFlIz<a*q;FtH3iKQdGn9czpBQK8doX^%Ya7YbCDldO7a#4X
z2RlubL9&|ad?Fo`bIG8k%Xz}Wr)<;~pl@istto!&c3LX0^c~R=tNNyCRM0_tsEH5U
zZIIMOABa9qOEInDMXfBIPx@r&mdZqRgl@(vHE$h>PAuAqd!e$RC-J&aPfR$x6z^C~
zI8<3FJFDTmbBF4PMQj)f%Z*k#8%@$#PwmdKY4k+22;M&%26abIM&hG>F=!5@m-U8I
z)m4-``dxHA>IA(|dBY>=WM{p;_!jyead@Y}S5}!zcC*rqu>j?V{#Ud^rs`q@MDO(q
zNWKoD2R*i_vUHlYR7J-W4lZbbRt%=(MV)R)0B0Oy+XhzEp)N=@DFWJkK)x3NzGwFu
z-Uqnlty%u7fG|$KKDd{0UvoS7{V&|h5Pj8~!`r0}+9*T5-vrTLzFK)-4Z!zyFIc7t
zh;jU@v@M3;4Xf$7n#a2;`-cz9yj|2q<`ABI%9rcrF8JU7Z~s^C!0<($uY*DH5iBq!
zF;iWo)C&p%1-KLvCe+UIK$^F}Q8;%5Q<~s8zxX+gB9e?J2@%YZv|u2A+EN}2^K5nU
zk*5tOh42Gbqu6jtFFz-pNm3Z@7zlJG*Pw_t1L7O9-$F9H{CM4QBjh$P!}}hAEORl$
zR4YmF4ByNj|H2<UTt~$qXOYls=8tspqz?R&=Yh^ZB!{1LD~3EtCod1kMnp%G#Ai2n
zIV88a*G|9r!|&6_Uwoe0`@5+uL6Ma?Qz>r4@hFTD;FGA0&}~6=G_)}UH!G1a>YzBF
zbdd{2Ae0)Er@rEyN$KmMSV<UW3j~Bt!kH<Fk<9zxOQ0<?b5d>$9N=cgABGA_3LT!U
zHUi&C8c1;*_Ihb}CIwOgGkKi7l;A8O`+`|@1zUh^3AB&C{5XC2+iz69TIx@Rp3#TU
z%xb9Vh?Qc4bgnX7vQN9{w8A7lp(wGZJv>KPkMkqgn>>W}*v^pcNZ8t;q-P1<C?`@p
zAORd0lNu7rE9I*2vsKAN;oDMbE3O}AC{9HW3XKhD!jHlQ?=Ga2Fk29To_0YqLeZxF
zXb(Q8OmsmD?F2<3&s;;1v*A*4nH7g(g+c<b*F7@{QEmq+`&4mJDx39&>Y?PkspE5u
zd$c>1mz8Q5oKP50#Gn-e6=!*2L=+Kf3t{x^N%5wf)GI5<pbuK;TY~<;382uh-#s)i
zo14LdNOSHy-PfqDyd@<npPH}_pq4^MC&!wtb1Ctx);m#qIDU2#+ZYOcrMi`N4m(zM
zt5P&sjW>|;_VW4jbo}x(T~5?%sc-W;W@@ei?~bWl3gj`UUjjUr;x2owh%$qW^QQ0&
zH{rmazTbFO&;cx`gHNiKaL3KC7!QA*Z{kNz_oldM+c+F&Auy*|#=8u}TNTH+SKtM1
z!hK)-u%CYJlOKdbe-uLhddK4X!Xy1c#Nw6mq!VQnH!#LKNnhtzE`pr~IUsr%Jx5fP
zUGde2*tfu2pb?OlzeM&5@4%4p6rTEzboMKdgTez9$|KWP6iHDCcfu>M_-I@L;|b>#
zcZOAIZh}?mfxGYmH{s4{Z^FF_mT?2{XckZ)%6JVcG8hcjhOagJ6d>85(MgTwPHMIH
z(&590X@CDgYPNXZNwsD>aYiZ=i9ESSG7VQ&=#U-VaC1A)9qIyHQ~nm>;JTjT^}z32
z0gHKMynGij(2vbQ{g|&e1AiB;<i3VgecMV<yybQQMAG0ak>cwU^~SCa$Gvg-+moa8
z<fxY>Qm)YnkbSF?_gS5VZi0~oow-`sPMwzY82R({!$@{?(oN5vzf7D})oM0u6pD#=
zVNi)deCZZ!KS(;ck*~dzhQF`LMv``?V;zGPN_`9{lFu(CdsoU^sl`^_l5Kt7nDU)6
zIZ-k>Iu-5Z#j~gByYIe}p0br5ee^hWc6KF)m%POu+X7CcLo{2hw6hlmw5eoqj3MX@
ze3PMvEeP3uj}>0fdx?I5(MxhV;k=`~X2|nfQaF$$n9vvD+ulMDIxi&CqhDDEs)jzu
ziZtD%A8_`T^u9`&%Cb&|5wTX;@;O?&?S|-9ogAO0C(lpPi&M5Rq`ufqLpqL}7=O?W
zC<D4e9(O}G;{lxoqZ%vJFb1L5F#$5`P#F1K2I(NA#i(QBm@ev>m07HcV+B&XRToXw
z^l-109!t-_<{8^hZb}CldfP%)N@4h#NJlyAj?&BHv()Pi+{(QDqv{sJCI-%0HS&g$
z7`!+)&4wSjX=fOdm~aW7;W_#%tFm@GtPI@rN<Z|aTAh`#Y>$s_2rJdDlzt-mhiW(Q
zkkxk>doJbAB*io7)9?)g65A+x3rOi@olZ+vN4yPVo{FI|i5JrphDPCZTXy>hgQK!@
zp6H7`>0)gA2p#D6fL8&EzZk4250e$y6^3?*?u_1uo=M$!42t%qcH}x$rN*uw)T!F0
z;-Vw70+CGEc_jF<?ebx4g}+QT#n6XA7Go{OP4JJ$L+8QD!o&Sms!@M#8`Mzl4|ogt
zjXVBV?`wEppjuwD{Z|5STFv<_eHFMVkJrJm3T{h(SGnFc@3#Y}dzUX5rMX(7RdU$v
zaGj<rayK8uH5hJ|ahHKF+w1e~4#O7+BPgT3A>wsV0;g7jS=s^}ojwCuHcsI|94cLr
zjo*Vp`fw|hfh`0#<3WfQ5)VF+Oi3(I=r<4EWp#`+2q{QcR^UuT5re!Wagb*iT^zUq
zim=<F`~-mDlB*li3<Z5Ijv({eKzWIu!*to_3Id|R4>^L_zReiz|01R9AXH!C6O3@~
z2l#3wIqHQw|0qAN4n;m!T-O1i5yNTc(SG{lKmB9+<nzx`x!LrJ41`HYvcRv48Hvm`
zt1#<N=|-%W8Vrm(j33()HYM;%kc7Xn&Bdj+E9gp?9_x<MhrtZN8vLvj+NxAjX)9)G
zRdfZH;Emd(gjw*0pelb1Ho9Qbw&-Rx)9KNR^zzxWbo$~boxD6s-Io%a`zR?WWt2Xk
zvD;3ceEnrQ{OD2IlwvxN@@7K~d5QTbSL1%ycZuq1`x3Y*sF;B=N*O5^!a8C&<z`TZ
zm0r~Kh7^x_t&!^DA4V?T5foBp#-VJX;KlZVA$X%;p;XdlP@*vqK?jB>ltp;Xc7V`8
z7&9i}OI`eGNGW7n2LyMN0S^j^9viQe4rLbQf*FM<J*<+W{glOLfUP0OANt~u?U*4*
z{l?6<jZ_g1xJP-JDab$*+Z$roBo4+iE?TH(VykTE911K9W3jb{jez8-G=TdlE|t-<
z_mqxudDfoNb1obS*cK573J_kDAH|CFesjn}?TB`XqRi@;iIh>=+39IF9ltpB)`tL!
zReQggcK3Etz14Ib3>9|t;#l#H)5VxJfa#2WP^DJ>ZaFw{7k2}03cBz-yUQ=P-awNB
z<uY#~8DTi==B{V^@yuK22Wk{0T;Lx8@kEG?4Llo%esW?v?W1t02tP-LZ}8Y25VM5a
zK%6<Q8E$U)fm_7dYnCn?fg8N|L>hyfS7;Wv4Fp2Ri!7ANj37LMOfwGC1V7AK^UUGE
z5-tEVK*XE#1IS68U*WgN57L+n9`I4#vn=qj0~+vzA4JKa0VE9{TtL*t95;#NF(XvY
zB2mx;bqoB6pxHfsZ{r6K^O#(&85~#Q7WV=cy>rwIe%6ogML=#%e#{9L6XGU~=!vq&
zH-(wCrqS=5TD%ateK-yg6a)@FQ0j|lTIPt-MH40h5rbRYH^mQRA-7MW=m!+R>xgCW
zGmePh_r>>!gFC|G27lr8r6=_S?l7WKKY@FeM9-9ic)=TmL=^i0mbH8jR}_b^2*=6y
zW%it11@1_R|1$2N&T$I}8s=A|4tIv$+VHi8dxPBSGHj%3y_wqWU44gl(%yr|>A}O#
zQhVpX1MW4+Pu~7o-&bTm;``O$u71G!4swKFQXsczyskInYyI*<99-aAgyr~m0q@r;
z04unH_JEM-Gw`nF9^v&NzOpyvzJ}WXp0BEBk8c;8bu3jEUKq02*6@dCz4YTzFOAM6
zw{BNbtx-=k$^PuZP$}tKELr{d<jm(uZR_#TVJo%jRms4}@+lolCjIf*OP}wAT+TUH
z;Gbe_k;4P<!a-PU`5kQ?*&ZVe@sP1FW;7ZNx2NgkLOQ@$arM54bhj$eeeOwr<Qo&w
zFov-jXf*7nr%!)O-R`ONGbSry2tnt;2)2KCm<}F1NS)n1$?e<LFEEDm2TXFrh$h)y
zx&Vd^Ce$(6j#V)jU&x!)PfTb-w?l8E$$J%*XtZHTO&0xePr4Y}*P$D-ats3(tDP_;
zp)cSESZBdVA_lU$YdjsNX0?<)d9>&4AJ4kI^sj$^lD>a(ocbelCEhdPP<?ouL-}Fc
zf~VUmKYZeQZ{wBpiM)~vUO_8yZDbT4^2PYZB*0oj^{BdG`#tZfF@~OT>}H|&u-&ew
z#}D??#}9VWgZ-9tG$s^E?+_mb!%^yuSaEgXiH#HKE4rl@M`!6J+y0Hi0La-`@S)YL
zrTv{oYB%d{2PQ_emDQ%|Dg$~XhM}=&pr2&oB1S9tx6`7YH&gHQ*f?n`)P=bzzHDt!
zC!B|d9zFuC+Nj!ay{>k`R-9n~#Xtw{spPZRHgcpo+f5%m*h>}FG20>bhm&;N9ois#
zHc+|5S0-)3UskkHw*c=N+bs5a!_)__aN2043u6`S#r03Qh}Z9nKDOM9cM$_CdM$Ms
z2Ig3WXMKY@Rw=yd4*gt7tb2|}%5&s;^|@#YFG+7^RdGXnMaL!{b;F%@%(#Re%#V*A
zbke@+w5Cd|mXT25AZtDaw3Tr2uykL;`vSL)HQU$lGl$$$tb)6m?`>)CAE54CzF?H*
zYKdM~-VT`4^(r#|uNCf-5Agke_rLjHN=4h;yy4L9_;^#_Uaf=Eb>(3T|N5Dq!XLMV
zSbfxac1#4eA_{mUlL&RBWx&>PEV&n9ii1diV&}%;ym3odFdiw9PH&99N4y2MLm&jl
zqCW~KUOY#vo1qB1Ey`QKo1qQ*%@h8Ma)mgfJI~@J?wLYCK_!5CV7ze09Tbi!c`oxb
z^SZeBD|qcc`xAxq;fBwI`JX(CFk^BYJ--&r<KLQ6fm_u|+Bw|w*<PLfJ<nic%Qu8u
z1U~X+l@5|8TN2jFtXhh#?J#Vx6$gSO^fPOZcFQa)1au3OQY<itT}mq<zGrVyHZ@d~
z6yl)-qMY%}3mYrH*g}t)Z=p<KB*7pQE5=IRdZA?_2K(tt5s*SMmM}gZ>MkXSyiFvm
zEN!Hk;<tC0P294e>=}rn)idT4cPzzfIvFbeP&P4cP@io1f#Qk60BtC5C})ntC4czC
zbG_bl8U|o3jGGT7r4Qvb+Cj9Bsq#nZ!l*SZj6B)!gi>I-6(6246VWT9s4wVpdw}B`
zVRRHfpbeuVGufyU?%?&-hjL^5K>;8hMk#P{o*H@;_~>68--;kgV6ocDb&mo;nWJ8P
z(6o7RA>4{<KNJMo1bK$hD>`2;Zn64|`o|9*x(x^c<)jQ;(Cu`(!&B3yN7^Fw&P-HL
zVmuA}C`8m1Gat{>NcG5;8DlBxQ?)rq5zDK!vK5~SMnmWq?PDoEgZ|Jw*ur=QEy6`T
zxUP6FO=IAvu&zJYA3t-q;@b^F_@B9pbj4#1i+7y=;!Ygji63pxZII`JhM+Ctgqu6j
zm!6Ao(z3M+43JX+MH#_egfHTD5eoMeT)0>A19#9Jxba#JCk^Z=+{JT_H^ZIXx}&TU
zDd&$n>ZZ6H-+^||!c1fFM|csBKipSv5Uxw!Gq?SVG>h?Za~CDT1YP(QG+o0DDrgSB
ze0SI~?);1!+@!N#<YC_;PF&Y<+kX*HfjdCY{sl<KJ>t#WhPV@Z?t|;<WwhjxpJ%_|
z=0%g6$0B5QEAq=_Q0WSq^(y1%A`yUJ4vWj!be9*7a2N3jp7AP&aWB$?JIgt<3q<(q
z%K2T1Be&_JJi_hnL+&srM>?c~po?E(%gwLkE4~vvg^XZKIaD|QlC5wA&&boAFLC6?
zycRj&r#9v%Xo-H#Aq8#oI2aFwZwVqa?vw+_&s>1|<bPi{-v`_vC$Hf{f&8Al8Ytdo
zB{kZIY5&2;>EQ9_>5FgvkiPlDU(;8=`*ZsA^Ka6_$DewINu}11%v6c~KyBGumaEnB
zZibu?7{Q4<uI!!#UV%(&z7}D1ir0huPAx)~Nr1&XuI7*X%_)eKt8)M4!6HB31>A(q
zlO}Yh2PPKnkfH9u@H{<zd77S`4$=f&o9)XbbNh@d$@E+5i?=o}QnOyMoX={U>3E!4
z^>W(VWeYou4Y5U@^&iQdmg$j$l?QTYmiK)X<+c1&fDHubEQ-fKJvf*M#zZs^@C&^4
zo#HUzkF&OhW6t1>xFfazA5B!U%~WT#S+nVF{@7w~r&Uk;d!6*)a6cXDdhp=DyIVjP
z#x7)Oj3Fl{r-@Z&=m@q;sZ8i%yq`XctWkIVaFTniUQN3@E%AkI2`{A+hI~&^nQ+IB
z4)q3S7BgXwl}3H*ipo|vVGVi*tGr5Sr&+bphjW;zAN(5P_qOoHq&;+(a~n)NVNiTz
zQYB~kR;z43&n`4k2gGBw7RDU5b_Z;16wR!tBR=P}xi!#lsC=xrYSb#qSGu5NQzjK+
z6l<uA81j$_>{pF5+0e;o7Z|^O{Qf%|YIZv9)K(Z<HIlBOvR8!{-f^}y2AS|<q9?i*
zG{mGwPl}Wq8cEaeCB3<XaZUUvZ>IfbCGEA8ep@lcWZ1H)wu%7^113!i9gwXZk*~oY
ztJ$`sH^x?z=;=N$48yA0lUJ6BU+7(R(Oj!=3;~l~C+Sq};^~XibaFNjo0LZ7U?&tC
z1#PSvr+$BwUcNl>7N0$}r4eVIDJ@%S+UV(bf-U<o>V}buJ9Xr<z*P=(K-!ER{0@|g
ztu8BSKkNj9(TuG*nKa8g!+S*=r43{qQw>s=2zPn6y&90S?Ogu~Cr@~UKU+f5?uRPJ
zh2n0Nt4bSVBCJ@Xt(dv0586k(v~yVg&aB~nV9n(<d<bw`JI%evP1i3Au9qsy<@tIw
zAYRBwuaf4?@>s(ntm5H+`+xtx{Cj;1x2(|1+n(HTGr-}2DmjSng!*vu?eqvlMD&0L
z5&S1N#slJBN<d+FgXPv9=HP3-#bAk_X6fTbn!sr<KM-Kg;uh(0SawGY$6f$E`F|b`
z6zJX<B2S|(0GIq=7H<C&C*mo<Zu|Sn{)GqWkU=sgymb%Yp}=in9k=kAp9q%nDtJwu
z$Yow{TuMM9EGp39u1DV;V84jO9c&Q~$alknbJ2~ni0xl)<e!uJR=AhpN}R<=I2ZF4
zTzwJ{o;`K1v>Ito%Fh0yhh7bXz&nxlF};xR2#55KvW0_W-u6Onj6JOSVx|_tDMk_-
zO@!YwzoZ?FFo0MYiP>b#m_vE9K||P3NH`A)MGM0Nh7oWx`wit1<t1hiVqgm820@wh
zC|0o_xAK-B`1{>%>YZTp7&AeuD5r#c3H&W79=rSdso8G14WN92(}FhJ9x6Snj8HJB
zcN7_f<}e;`7al#^UkU|AEgQoy#t8=o8D`j7ITdacXWEyya=-wlGC?;q5yuLz(y|9T
zJVg-!%tNJ&0oF^Eja5^5U;9ujtt7xhF3}1uX7OP(Gv9Qf0FLkzey=hU4#L4t7-cJ~
z9IyyJFw4?g2PzB&1VT}&J)=}OtmLCO5Sf8j(EtwWnsy7VHqI$7nz2gf{D_yst&k`j
z99BrcC4ZDJ&+HVy3w$Xb@0|IhILyYxpb9_n^Z5X(cLt<w_!2H=zz(F)^tucPutkT_
z^LnPLw8W~~3}?u(RSYzaP{w*guZoSb(Kp0<?8|QaAeB;3cK_<m6?LY@yaXiRpH?2h
z@%H2kAMUge?%aYWV5clOe?915{&HTSYmv_EHw!Ch#?4<523|P=zx=M?&hhd^oFd*;
zcY&MdRk(}vM2g!&ynK<C>x#ReBgfC-`pf)~GguYu@(+KJFbB?`b6TPZRpB4^$eRSJ
zcd7%J2hSh$V?UhK8;BG^)cVdXmpy0<+zu#sW#Ew;KTcHK`7P2pt*%UScx89s<#A4P
z5oR3m!=L|=74%H~hH*(3?*fz)r1as&5qKjHN`{{z+h~fPe<NHMI7u+WO<LnB+!;4`
zfC6{i<2fGUIUdO&yo;Z>&C<>N?BYimId~Q)hZl6r(%3Vuh!AmSk5~P;(_e)j<p7_2
zkjLPwcuC0?^YJC$2-Z)*kEjNoBM$fI^BiD2oO^0_@$7hVkhA}NlZWq-55LB(KQo+%
zrCx?Bz_E&7U>>qIe68V4!0nxe8BfJkz~**pw06_ON1vuozW646_3a<h*T4I7`tqAU
zrjI}SIvqaxEOquCNcLkAQpxg-+i-z{_W&8huK{wx0?e{}OIV`#jexKBD#&;j@y~e$
zyj|{*@}}H>f$%PhDyH3rX7<f?b5lQ#oVig-Bk2+SvE<DS$>6nEEyAiq<WPWJ88&ob
z_}b~TBtKvJ3}UwV+i5pbyHQEiisW>m&PFTrA#@qOF*J+W_u1bw$@2<F#$`eo?cc^<
z$-jgnZ?i%NL(){T`FK1ExmxltaxL`n{!Gs&X`+km;4ogWE&GPjF!>7u*C!u6N}ql9
zar*5yU!^ZT|J2(rqDwV&$AIH4>!hz?w87xT$~7j8k)F7$*y8M5w%|h#>vWplPOe;*
zt|MAmiL~2kq=Vg-<#r}9vdtj|6l#Jq$ViVLx{-Az@gH5K)v7rU!gh2W>@?HiuJkwM
z(QHUxv=K`A>MCtx6x<MAj3As%#w0rO!+<5-RP>pTs*jj}$V5u<Q}jVAyJxhe_wMN0
zQ$B2^2#rj5v<~a?v$M&$<JYS->yySOS~0vZxr{ns6&Z~z>I%K=GX3zw_v!C{`!03Q
z&b-zA{@$K=hrz3&ayEod_eNcLx72<biX%Obed(n)p(pC5zA)-tm69&cM`=?!nu=~7
zZ(MSY)kf+FXOop?RxvSbMY(A+mPdqV%&MjhCJauc8;xzO!br*v85n+8HMX|{O{)Jw
z-ryL5R=e#9vCv^1R&8I=Q-6XE7`i$W5n0vv{Kav4@$x90$PHe0dts*+H6K(i8@X(p
z)St%?#9wdoueQus0XvFx`#tG!UGa3{cY}9;dWKhR(ZYMiq}R4|`c|Xqx`*Gy=iQ;W
z)H{<+p^bXNC>Q5#b9P-!Hl>|nfQ4V)(MG&N`{Z3=FB2vcU8;Qa1q^{m-BcT>q`kde
z8$5Z3sUu}KgRA}g8tx0$TwcS60PA*|0ZzOcH*W4jk&(D=i!+B`b6*^MdbR5=SQW>H
zF9{|8l|Rr#@Fm%RyM<X@`46`~{>1UXR1SBN{2VCw@$iNp;+a)Nb`tj<0qo?Bc_;{8
zhK2k^*J^4CU*9N;a1}yX6hl9hk@A59p<g+o0N}@SM8r?O`q_eK#KYTi8EK3*T}b$J
zI6Tzdd@~+h#4FGdKjVn2a1g>R-|bg87cPF|9xngQ1Mdwe1IiS6d2pE(7S1%6g)#;`
zDyj$RFiE)3;s@IV@JKoZiK_xJj@f@Z4-QjD%1~JX@1)rNF3Kd~9m56Ex1KSY9G{(7
zIE^hVV#XZxh2p|WAKVDFBNQ?TvIvF~3Ad-DlK_k)k3xa5qaj6-1D46htAV8KF_?;i
z<rQBTDI}aiE8Fd1n8FN#kqG6+Lbz~23kN-)9i61(=f`O}V5<a^CJ;vypR>yjA3m^B
z=T%9%d@dDqsywsRkd##p1ZM<p%7Ot0z9KkRDj{SS(a=gPRM#qXg<*)>rh;a9Ks+-Y
zDG$aY@FAR229#%vHld`@io=Knz)uteE2C02F%U-WibqQ4d_)8@<JfYNNobT4eu04l
zlF1u|$=hAPC)LkXJfulrxTCD>g>ECDJcIY5L+O}NY6V4Mye}woDAOo+7~aUAEfpyb
z{I@~U3YyBUGQ?SFp)A8D3~TVsTcDUfs21=b+AnEn10HWz8u(zvM}OqSKp5w~dCSK_
zA+$l3_~5wQv_j)8Ojz~DId967%u7-XIVc;Y8pCB?HO7I;L$w=qtSUQi3zX^-LGBm)
z#M=}u80Y>(`SI?#pCWIUqj1B=BA)v;`l-Tyei@bGkncP`0mkp__yw~g`m&GnBLqLZ
z<H_MUj%miljWF)ZxXF7S9{Bj3$63Z5u*eT?gcmgB_{NPt7dM!|e}2{v4t(S;Tux^o
z+_)EM4u5irKOTvf-SNx?Kk4)ACNBOiY{YS2NGhi(E@BeT@$oCdi!|<Eb$7WOr-<wP
zi?|En{-@aZIfi5T*)((0GY-d@;fZJe;m0@4-r*PSxcIq=9}$U1J^R9~91EJFJlRcr
z0`hYaZZuKu93OlpX%>$w{IB8;f9Q`Q#GfN&H+lLg{Cuech-W~tz?0nzzj33_T=0u8
zo&Z9MC`BB@0xqtrxWg|~lI1zE(;mb%N9@Lras>|jC?nt19FEgJj#Ky*7eB-e+<_zG
zafIUH2HvQ{z~>ux6M7Z5jdF?++{0b;X+-9cr-JAHRig1^vNn9J;q76woSN-}^zfrE
z(<h&Qoxc3dpVHUA`%C)r>))qGAAg?q4j!fU&c5WYrerGy=k+%6E#t1%h)iQyKzF|n
zJS$9}>3e~AQ{E14qAlZH#P9ce25*;pq`WWA{Bq$+6|4lcH?_I(g{16F$%wR5$@bM+
zQ+iU{t4xsL!GpX@9jM8sYI)m+FAPBR=@_HXg({V8Z(Y}DRBTM(%Ql`Ob4!j@dW=if
z;|duaIX2onIsy~pko7U_#A>d1Up7)xI@`g)o>y7X7?6jJPw~-<tcS4uqKH9G=0bx(
z2Ir3zNf-_=AYe?u;Ba<!mQGGiOgm?8#ktMWcNCZ1IN0)!{5TV;)2gL|{f_*q>0&ar
zLF(ZH=~wdOj3ibiu@!x_q&&xiH0bw@XMeYqc6XafR})dvfy6t`j@sXCr#;;}?YfOW
z7@|1$3f+aR^_sPk@Yd7eUNh}?q%$&!O}MKnYe{8fmDHmL`{|*0(rE>q=v$stsCr;q
zc<SMBudO(3Zzb5&Gj+y!VUG?v>ElOx>Cs_FbktM5R`x_e^bTqRJ=X>>PYjekqI_X0
z69w6B9lZ<V3Vl{=L&(k<7wNP(-;32l7=JJl(mudPU18kAa5e0ox{QZ=E%{ZvbtbFM
zC_}6?EBV~8E$OIKYrUHDN;cBaHfqYhv?(_=eKAfo@u#J7uvH&v(K*p|D{8A%;jYN9
zCEQZptmDB)@U~UA7@0U93j-l_k4duEAEo|~i|sF`-nMY4`&j%qJ{_cg{qAY{?uV!8
z=;+KV_>{ytBf7CC8?uT`bW*Qg6{ohvdjucQ6Rmd_ZnuBAX|L#`q@msI?{?JgcD<6Y
zt$J=sFK#r1OXWa!^tocZ14`p`v|+_zbsiICc`x|haJC-0Ejy#+6@KWKq(RognAyYt
zES^$^iOP7U@}Hdb)jw>7A+TIA4rS;xK<K5R=dR&?Va?$+{L<l<)y65!tCi^H{<F}7
z`MBHw)*Xs#6)c9Yy4G+VR`CVH*G9Qq;&Z1*esxDCmSAx!P<M;?mY-i2BKy~Z>q&xW
zW;f@;GIoqFXF)j>!GtXz^guBgs0NYm`tTDeiZ4$kzK}q{gl05^z^Dfgh84UjC_wIs
zEDOIFEi*)nB7}z=<bYX1LxeBjj{n6df-4X26r8Whe4QTxIhaL+@QWXU-W*6tTKpnC
zLR8Qy9+C*>65PZOgh%5myk-#`2lEMalhWSfhw0OAf0G`4{%NW=nl_jbg>CQ<`hYO}
z#B440;6_kHfJBgHMHU7H_=~beEkideg)n@fFmPZltFbVKU=YIS!S*)$yL(>Ug;K!Q
ze<&0vNC=f);bbK<Xa*++7K|Zm1%Yyd01br$;ePt#59vSsfB*aRpZ-7pA$4DLrI|uF
z)JuPr;`#8S$LY~$pQNUQR?3c13f$lu5Bq64j@4+ClVnpHh*-TwTaN)%6d{aQ47f5Q
zkSy35ffZ}*cE<)4-WQi$xs)XI^Bxhm&K7$rPZ-2hR^beVm04r#bB|EYth}+66xY38
z9P`&R*J8yLGdoe7eCY5+7#88X8-mKiDlT{j5Dx|13bD$8QqA*tGBO^NY6N}`1SZUi
zqVlHPlo(|KqZ+HpJSe3<8|lOY=)(|+F^pymo$!-ch0Kt&(n1>H_ewDH5~EsddB_Su
z1`)suF6fw_@f7VBBN?-6X@8#m=u)ZOGJ_Q(mw0c38<eRr4|@YC?`J-{Yj8SL9Y%_6
zDS>S%&AWS2)}>&h2zf7h#eMnYWtxtpysA$tGB6aR$f2w7r+#4bq91M^f$W73ik!!x
zmjJu~Vc>F<0>>h5-?G0hgTr4R_Z7GWcaC}!ZsIy!k?1yl<i~(lp;_UI;3&&D<jBvo
zgR8KJbTQZLk6*Y0N5(XV@VV+2u!v(3<ewQb!y5hOGDJ&WhKv`u8+Uv?uEq-tx8wGT
zOMv%D{5N;`2meK;O!uM^d@CxI8<4*u@d`)0sK`jmzi}4{0(XQ0zPpid*MB4lm^tKB
z1{jjUuZW00Hy-^*a849Y8JvY4XFf$b`xUq?bL2*de}Nas<+P94wiv?iGT;jm-JJRL
z7#RZZa*V(lzsrDcPPN?Lufk&b*<B7-m6}b)0^~L@t1tbZbi0dD?i;QbbqyaFEHoF%
zLG8|7I(+arwRZN??*7BHxBoyg){bN$1~(gC9f7RK;C_*4-VS+ESnfLjhv+%pMq(@o
za_<l#$H?VPAXBtV8D#^O@#eu%Aji0^RJYRwmu^e_%K`KAb&6UrE+_Hha{3%+F_Ppz
z&f{f;*I+7nGqGAoZC7$B1}97ktct<-fxa>vNp_cfy?HtDDxIdTsdOFWXSQC$0K#^M
z5B3^1)UX<<H<+gHpPZ!cetel;937{ApA|UqZs_JVf=+-w;H}%07kVFq_N`XS@-maX
z&`H>mkQGC{UO%0jp832hR%oHKVGLlvUoTQj4$~)ubGbb6P#*>c`}cQR>F{7DRkxIv
zbh_c7=d-_<SjNs3A#=bJ-K@9hO%z>B)Fb_l((mr<q@MDP{77@2zW@IF^q>CIzod4v
zWusN^>{Q_`=`WSk?RBLKo>^Bw-<V1dU~LV00y{qJ?Cqt;AARh8W2`zrzu;`Ey`6e$
zRi*zZZftj`oD|mU4ie`-ol5_!*IVh}(MPJrviLksr{eX|Y2RBlqH{Idac&mI4D!O5
z(U-n}T!W#jES*;5+5iNvUcBt3Gu2O=(H0XqyItvtlBrs)V&xibMhMEH4FeIA57}|T
z1}5o97<<rpX+4$gl-g<s7_XrH=;+A06?J*Azn8xF@^hb~H4>wHLyT<biE4%DTQrVK
z%05m~WxQ2l+AOK;Q?*I(Pg6%_*l$-;ZCkqibR^t;(S~6xr6<o{O7__n4?3x{)1Dc}
zc(=B;t7>P`$yAPNy`pw1om=$!oG<n77t%RT`r~xcqs@=4|FcRAo?y7;sIliqeV<>~
z9gN(cb9S6p!l_;Xcx=5=k8D4Qo{SFI>-D6^QLeZUywEQ%6_G6yS*f_GO2G)rsyK8+
zaPWO_f2w~bNupOn#~C}moORuXnMjGjt;`BN)h%&Zp~YEa#1pEtx3`m8YNzO;OqO&y
zF%1JZXUOqhL(@=oai+RwYf=0#l<v1{sk*6mRlL?)ruvOD&SMfPtIkvo<q_WlISjeO
zS;O0cTh^MLYxo62?lpcIco#2M#*s^Mdx>VUv*n;w`J)?fzg;JO8(5w1FB|SE^P+y1
zfiu1~%H{GKa0_J1$yWJY2Zo>#U^x_?z3?x1B!Hy)+CnG8Hk1d|&Po!Ei8w@Y94Sxo
z@WJx(<Jq&Ei%6c$oKx@xC$a$3faCj#eCCk-0wN?-+`C{I8_&4&{cYenZuqkba*i1?
z<?wPD%Q6t|2C@n{UaJx=XLtn`6BRT&{z4)-T{ZDEO99<^*pBGLnL&}9@N*%3@cAd{
z%Rl@!9e(sEm6(-iMGwPH<cZMAz#$44LjOd<AHpeS2$Ur_6Em!0h8$K91bYj%5|aB8
zZU-YN1H*}g+>wOrvD5DDAEb^1aWp<&JaD6A#_Fbul|gV@z!eP$r&Hl%FqfHVbtzlS
zLSz;l#+GM4K1u)bAOB<e`+xkGbo}CFnn+k@%ml%?v$vZ%2m7fZMXkNlv2w^vzzd8}
zx?}WW<r7&^hvdadAhxi;kc4ptut7*=++y1VU8rqruZL2~$~9<aPjwi)g@UEaMjpk(
zum#_|@4f_Z8_-A(uT&<CThQgREtHQ}q)k)?@q{ppS){?>!vWgRyImH)g@ct&2;0zC
zk%GxK3>@yvYA|MSQYMr=j1Ty6kS}q;i2`tO317o_6<b%Ln1<nu`Wp`WE|a`m1{?5f
zP*V*lB3l(<$V0&#sZ9(pB#IBLY9nuy+~A>jtBgXq*IjwiQi<a&E9HUWLEg}SvP|2e
zj)-H!AM{S9Rx&s!dMc#^JamF1)fvXDqZg;Ce<}rJs#dOSX)o=)M%rhVY`ZDE7zQ`f
z;Ix-sNGY5S)lkkUUG#tMPkB$4054(%L_Kj^^uOvBKK(|EC_oOslf@&D-Lp9OtiYWi
z`XjyqbKLL+ymk|*fb5>($d1KG3xIcEjvizY_cG-006pCUKe+NF2@I&D4OqYo4(@8-
zHl7%Pe(VlxPERB!i?li4>^+C<8#m8J75JM-9`>K(j!3%k>|{6LUWi}ZX1GbC-x>Y#
zr_DRmNaS+<jyyy0|7Y(%*Cb1lGfhyPyK!?G9yEEOs!(VY&}cIIdw1_-FJqqvy|*+M
z(7Qc9rni6e3>pnmD3&CsbK~ZGzV{PVGjk7*$c%`Ltja=&`>~@$R8*9mBXUHE;|()`
z+*DDFvl29c2^v2ob-1UrW|Ph01$Y$~VN;|TQrGcay!dfkF0P?q&XU1B=^xig1*igC
zSJ%Q{Pdld|cqT0FQ=fiKw9pY3y!_N-XCar|u1sLshP02;@XT>!VTd|`_M~8PyEA42
z8D63-kHd^cnV~HYNTv;5q#%s<X!zpS$$SqN@1YU~25cUGFnoQ0Tfonr(tl@RZY5Rf
zt#tqXN2$HOnHsINw7$KQTDtSWe@XHZGLnP>;qoTYGhcE*0qD{r0m$3?5)>TrFW2`C
zWmMeCz;}YzaYfy^6t@>pX<wbXq<B4N@1sus_`rSeyB0M+3#U)HfbzaV<^`kFz4d71
zQpKO@Iyq0NJCe*hW<r<J=~;{6p;BG(>JrQM7-Yl~h6#)=mrNq#JuA5tBTwh(z%p{Z
z(MmgaHqwK;n?93?-8wK<3`BeXu$P`b-%kgJ$2MFrr4-qomFP@n!<fRVAq-iklJi*&
zR;$!fv(@tUbm%RyDn&eLoN;~<`V;aqTgG7^s7ohlO4eOPZxCJLhc?*7q&f5`v8g(_
zlI(t!RyAbN>Nt<9b9|is;UE7w?MpTWVmngF-qJBRTlu3$4^v-dUEf&GlgKX9(cz)b
z$U>*U*!1ARL$4@f+egYbRKCCd&9BpMfBRdn%DT6+of^_NS)oRGnIO4+XFKiOy_fb5
zI_dfDuITi=vg@;tAEf(tcWhKRIT@xECKRd(_xASEv**v!p7=xu+1QNq_Z~h<t@e64
zmM(DI?Wz4#cWW)zHToKp5Hau!MGwB%?SV;^7^PaQxRTz%c~l3-JsT^Tkja*K>~wOd
zHtTke#rJ{@R;)Z)rcKm;hNIKeAB?T@p{p>76JyGm+LF~&>&>cHcAcFK)8YO>I@o=o
zxKY~jDyFt|RSY%DtFis#IXZN&uRe7oz2qX@yStI@s*W19n#;@iSzkYSY<+HLYa{K5
z4u)9z0ERF65W|Nv+E$fjLpm_4>6qwve$E+Z7?voPBm?n_v6MdK6<g+?GZ&NU2pu(y
zjkE#%?NUOH&v{n==u3Y(I3C#W#n?lK#PG*Bpl`$&X7!?Oa$<C3M-sM2gcnw3Vf<w0
z74%+zXV5s~%r{0Ub%E@IPLA;ndeCNzOZp{7WwvA_U%mx&x?Ss^i{gJCT~%)MZlCQ&
z*+idup#69^u^l5@<MVv*PBNmywwA02WX0W3`F1)3>${XeD5+gtP9NP}PZhNp2IFPX
zsV>j^$b{~3N8_ByysA9jfqn<5rg#M8XA`=N>jiHF85CkDZ_Jy|vJVj8`YDz50Y1Qw
z5dO4B_RYaLTiV+PuAhg(d!_E-IPSc4emrn{S$^E`dc#+4Yu}I{xJ?>ee*Wckm|Ij0
z>*#n%5@xR|j&OGff(a`4q|S7Rz^FCPwr%of)hlp(FF5ik0V}tR6}`!BBfJk#9?r+*
z$YB|#+|i-oNp1-zt}^{_CYM<=6g+|1<x}d!qd#zk16rvRhv|mfd_*YV;2HQX;ka+c
z7eOVYJKZNp`f<5uQsjDgi4cClq&_4pnzHcu-W6#|LV=u?t)?MUC;@Z@%8c7-gADA<
zNod>o^rQ6YFMgWVx3|+$W!Xv{0xd=?Rjm~Q1z5nO->{kqp%X*Pa5!{PyhR5Bs1y=-
zK9_*Zj77H3U<*QwG@KW8BB2xG2*M}|lV^WPIAyk=4NwwLTT<@Wf`EEN*hTq-9s(<Z
zH~C>CL7{8b+231Ahtlr7*iA2<JxfOiM`<L*3j^JyglD$ZX-TPAl;XlRCT#J6V0+0P
z_)=Oi>Tst*Y`hWwC-R?5a77VeMVjX-ld=9W;w&w#I*wMNzN%I6p|W5wBn-oyX8;nf
z@=^wrN9vEVumTGOk#J_hof4)yf_p`DFw7vp4=^k$P1z<99!FAWm~n^V)lm7Ng#u38
z@<JHTXlI0UlrEGAj7PLT3K26GFQr5=>uy!;1%JJM&)c1#450ils}tiKV=4ByA6rnr
z;Lt4f8<cIk)i*Jy`Ft$OF1pNi?28`h81HOv5eg79?xYaPP2WO^W^Z@sqHM5=&|4(h
z@D*EG(x*sBe~qmnHP(esf23R(0Z}TMdFZnYR4>EsP>NzV^*UkrqTn`sZPe1{)_ST+
zaYrG=(8X+3R)?MQ#p8nh=7x5D)Rzl8@#rV>QoMkZOcNMzQeL}B#$6YE1Gk?QAWUe0
z<llfvc>c-$5`P=*YZDSCv~yWX0>I&!Ms@=@Cv<Vo#9c$3@lanAZw@nwYYD=68_(ZX
z9|aujAl_NT12gqiT;OoaID{Hk0h2BF92c?E^izl!r_yiWloe4Bn9wv7SD|%1-7HKo
zXK}<yMx3lf$^egnLEPkrsC+LJiwpb$a;c&)bd)rvSCUK-1>nXl?ggp9BqHDSt6Q|A
zIHXOk#R~9na~z17H>a6NpA!&wEv-ECiTsu~l=Q%Cx^ll!eUm&U?r=`P__N{JAEd=M
z4lW}^D4t0@I$S??>nEo#{VYHMGq|pgYw2*$;s{>lz71}~TU@Y%hz#N_<UF1bkzf+5
zxK8hb;p+nggu<CUlg<?eEIuAT_~_%bwS7NT8|}2Yc{ep$tk_>kY>~~u|9m(ML6t73
zagqlh_u_sTUKRdb;cdGV`Mzu3uLjBwyt>@P-_HB{<?P1=uE+1%#QELD+a_GpqU1*P
zUqUfHAy+bK%`zvlsAOwCvUi6ksW&?J{g{<Rx|4>F|7?rLSx|t)Ov~;VD2Qir8nSM;
z)3w29JnSp&Kr(ky+J)3o+J_Gwr0vZu>068GL^3HWNDhuVX?H))Lc*9<S*>}iL?)E+
z5q>1u9o-4po*f!kNrQpF+r+V5ocLka3+jh0=rHQUHr{N<E}csH2P<z@#mK7ir7AFd
z4ThuC(-lS$=?u$@`E04BG?v^x9Cp*;-hTR*fBqNcKk^D8CL^Anj?*_^e{CJ?&i0n`
z{^Ik`Q$zAJa{IB$PF-zoZd#A1qf4kfOrB#!T3536ai_1VBmF}9<mH+8;vBSQI*`uR
z!^pL|EWT?h^I3ZO{F!yM2M_P1Pe1-BZEbB?e}Hewh0Z~Jzu4V(U7=I3r9S!`^|jGn
zbNL=WeU|F2we;wdPf#LMkDRU5OlxaRxBWnQT}p|@K*Opp3}0-a$m*akzkZsYy*Tg$
z!>V{gfBoI>f1mab4pjEf(qH}Z7vjH`P%}=1KOSG$D1`Bg-6%Z4Qum(v(~4+5xVN3|
z-r4ZkWayM<C!=(<|3Z2UtHj3Xt8czZPoF+_ovpRn>E68^>v-sV^_8Xc`KOQc{N3^d
zMfAR&`p#juYkeJKVnc0C&-W(?#iL_UPqaNcFUHl1`p5E$(yP3zY&(_S1J4*l=h(Vl
z&u>=Sv4tY~)`G&29C*I7f(=6{1}hs-#V4z}PNhqA)u%fBv1xMl82yuWvLQ*&KzKr(
zV<j2eKb|ou@k)B3_@tgZ@l!Y$zSzM8<KamCRT`v*@Pg}}6_)^GffaD<N`wB4e#)vi
zjF9wupGl`WnG?^@IPMNy#?#X{V~;x6SYJz3^<j3SIXaRKh!K!_5)<f_$a!p`xwR&|
zO4Y_RzCpz50FXd$zt3cY>{fHt?ZR`QU0Y2Lw_B-MSx$pqPh<Naz1TbQSV#Wy<UM7;
zlc)%Y+{mXqwfxJwQcl`2<p03^0M{Yv(asO>0e+P5r$4gG0KHjK!e1V|Av1$J#lKyC
zp8YcL#}84iKP{LTzGxu7!gAYL2InQFKVEvw^x(sUUx0q`3479}2nI6^Y&?k@7T9N?
z^b26tZI}k&0%oHf`%;>h!Ah#`nZ<CxGokWRM*3pO!{E+IBzLk1qm^IUa$sZ&<>R{p
zGpBQwhRI4a0(1Br6>1?U;tUir`w=vRDzCBao*go9L<|@1n{lMc1mdo|Vk@VdVm47E
zFq%OqU=<P*$l<QTT(>~Maa=ev;47hli+hT9GezLIw92H2RnCUCRWM;>y;85Holiba
z4}bD`YHzQnB`GQ#IE%38Lx3e%A`l{o(=QNUF;Izd3#f#%7diqpGq?QVO>{3VMZ>f6
zSaBpF7lmZVDxeE<!v_*pO|3|IWTqAZEJiO3Je%tqHXNChkl6wep?4r<GG<4{b{pWX
zHQ7_X;i^A4+D`*1SoE7%kt88kg06*4@yHewJt=i7%d1vg;g?Q@k~$dnY~aHv!@k`p
zn;4By7Qjc*beS<EsSL|YQYe`Ls`#-KCxrVIm1B8{1D{uro?T8VI}~m9WLMoW%k2{5
zUCfFr+x(rQ(0QijVydHHE8bf$=KMH|%yp)6+b<pw&@mhm&c-a%0paXlekJXk?H3i2
zPE~ib2W>?eP>!iXdQ8kPz4Vq77{geBhoOpXC#2$7frht%6k>1Tpt>atqZ#!De<%_d
zzt|>|Z9A9=+U*=gJ*l4Q<J2kn@QlOI3GZy52*1#$tuex3=z>>PaiNrk@ik`md9{!F
z2=&M~#8AhKShmx|`b+$&6snjZ*-L|t>OsmCj|aB8Xs%b&#@0q^ueEJx#VFQSf5Gr|
zIyg)7Qr`RwkKrP4Pjx*-Q2tfG1O-q)Od@czMNmP=iKjHc30yxg^4Px>$Vf0t+t~c%
zd`z40e8&+jtLxinD}I6k4etDy4wrs#=fLnrBq4yKX^&a>F7=^%343jQ5HSfV&ph%E
zSA>ERNJVMH$9UjA>%R6IPTx#J31en>hQcei>;`ZGg@+O~gAzOxL_+1hM*9^INoFZe
zX&NA<;D`^*NuLe$367Bz;hTS_2qo<_>;~khLZ;k{$L{>BFG8L31SXXgEj5$cFhSGb
zB&I}9lTO2+??n31|ExFvOaNabgEsj`$fQ1I0I!2H|9Gy^R{2UB8qd^n8|H945jP7n
z<#LLf@d{M{v?DHJUnv}y`sDT|Kpwg|HS&rwot@~!hd$zt`@J`%{4Vf{LVkev4W*H9
zYk=qvU@)i&V0*$P$u}$M(Z`>qyZ0WZYO|fTwzpHG)z*8NGlW<P&*T)|(J`^*CP26J
zw^!tR<Nl^lXueIiBG0$Y_uFCmTzpj?#NW>M`{nG998@K*d=`Ey<XYodCpYKu@Bv?<
zsAR~ww6Wew>+Oc*am)^U*k|_wc1@T|Cuf(^dtxg?w#qx|^iof<q*n_yo04_U{867r
zsdpk!|3?N!CPi10!$uu;xtO~=OPg)!2+MP6_xba*zjr9Pb2EMX*=N#iYGJ5CuM)At
zqi#An>WGK=RIjmv1(WhF{Xrfh75Tywx{oKLsfh63tTT~o(}odn_%J^lF%fajD_xi<
zW?pq~N*8FV+?C3TWp4Dl6K@mFcH;A?(L{Ei6QBF(@NiFZ^I57W-JR_X)tSb-<a>D7
z-QD#nD~uoLDXj3~{G}Du5k@^SM6areH}d`FtKZuow!X2EYK^APNn4ga#e_+=zF%Ir
zOjV`FpmuPuZ+*g(DfMm=J88XDPal2uX=-oYO|7*x<;7~LOVyp)RPQn-0CsxpI&zg-
z(q$LsF4B0=NxM(K(Y@zMk$ia>sJyjC+XnlkmAbVhbc(9#oAyP=rF@98^(yCLNk>u{
zS&_H5ci>5z82*@0*^_>Q9{BLl!}M3b{0pzD8;zqrV|zQ!kUCd6#;GTr2?Ni`3HldX
z;$uWz^$I|??xXHcMtyHL$>hx^PoAb1tYCZbEL|#(&GohP;K9ArYPVFrT3TOg#tJ*-
z)9sJc<^vmTS*6z$ZFFJU7~rI5%7bo3|HNR+xWf44-8b~SXG_81=tO<J@3ZVY=~cAZ
zrWQld;c?&bOo8-v|4PhOk<D7ghPuIM?D2!X&1$N?+JnA@PFj<0N_#nfT`NoIlj=tp
zqI%Nj#~N$m9sQa9i#}g(a6TOMzvMcnFD$avy-IU@+;N{~i%nKM9v>aM?U*DA9&Jc|
z^aH*pD2WZ4Y;Aa`dOGZmQh&(g*}z-hX!{$)p~l|PvGiNUr1GQ>V_rK~T|lN*SxtA=
z*Cj(~OlmxGR7Du##_4Q4@VRPr;XK@GroaB#qx8wW&9rc(F*fL@7rO`P>5F|E1bM$Q
z9??-eh%@lU_0vv4zT=c*iZOM4fGMaIrWN)9KERI<{`5!o<cXH=1qgq6@P^z6?#$C~
z=2<y~tf%});O*tvKVAsK*UGBZM!g;VPU8auvW44^7|QJ@wIJ7lMROs$sY`pmFA!}G
zK9?Ne%s56W*dic)VGuUhCgCN>`~|Q2EuE17Ks2F3IiQ<M<@oStm^cI(rO=3a3pWu9
z*<(RLZqrsB77}J7yC*=BQqKa<ubHj)*?8Q*3=amm@gm=tKJ`)BkN^jA!%+fp$~{{M
z2Zx{JhKv$0aR^Eq%(^l;9F$*wb`zbmDDLY)xh?D|_gOkrz-79Wuw9cteecsx()}+!
zO;rh$a}r=Vvk4^&qXqj`W7zO&EDRKKyN~P7TMuSq6~?au2#6SWPNWR5tsdJUA|xW*
zDv`NE0Q8C@j71W%0UJsr^xoOrOjRj!07j4tgk#ac@HIl%v;hZ&ZZWkRjWB$voJU9d
z5`g<@Wx1lTb{Jj6u2)n^i9jJ`<{W7;yim7n@kjin$`(c#3Gd2>{8+sPj~J*ZFZJMA
zbm9vmE<!UE%U<WK?y9V^6^axlx3|W>l#O{Qs2Hofr@e%BpZ$eVNeTq{GaC=%4y(D?
zl8_ZlJQLt49{6l|qVlTAsAt9|v)<0w{zQKm4cQX#O7+c3D;vrviwce*Nj2@VVJza~
zCuLw|8AdMnp^U6@!+^x9w?!!&<jRV=iWF4JcQPK?h)LN|fNhMT+&0XyO7C2CwB*^S
zq#X>prbj)23s0ntXC0v$+sYe13JhUB>kFk(c+B)OO_!HvlHxPI4IBF7!vKm7hD{VH
z%0}MjQofE3kJ8azH=PbI!tf<UZ((^sihnh2Y_F%bl*xLn?vEY=DMb4(cGIYLn&!Zr
z4AcTQ{o)_Kv(pcyj8n>1eExM1VG+xZyfc{LG9nQMx%rp(IyL!?7wC5QKj-Y9z`aI$
za>prmcH$r=Ttun-a)6S#jikhy_!MI03dnR#-Fc*#VeoXv>%{L+ANJ|bLHZGW<H#M+
z;V!ABX>mtf02urhj)VgLB|4`Ln87u2_LC<L|4hqNQih+9QVJ(_SU~urey4HHXPQ`W
zJMC~6S57c-hQmD4F$>QMhjxVNDg-l3c`Oki%k6p#8bJ8;r#Sa}gF;I(Ad_*B-0r8D
zMvfv-cfEabI-aD^4Q~$2*iP;+`HMTqU4zMF8hCd9$msl1_#EwGWt<6QC&v7eY69R1
zKj`UK1WG@3!e9`h`JVm0MOcgxJ7*xr8FBi{-F=cs{TNF>Wnv8e5?4Gf?W9aOt@tt3
zg2#_KfDK=`Jx&AL(fU2=cEb6?fZ6;2KRlF1zP$kyi=j&|%LU0i4?q4aJ$(4F<O|7H
z>)U#FuInAT>d7d`8|7RsPux*jznjaLxCRkM{Hx(zx?jojZMqftzDwS(hf*J}E<5qJ
z^ZkA~`y&Tc$?bjV+aPMqDfL{WUvloxad&7V79rc4>*<S6AEx_vw-n1x2}A25vCBe5
z@-feO^=tL-qi#R#NtUF)`^+xsA>d+k7>!Q6iV0&06R_CUZB6<UN+a8%&8M@p(NtK2
z304?1o~6BmZd$3-y#=3FlqesJ$DLk3J>NY@{r<?Zc&%2m?2BB>mVvCSL2hSN7`l&j
z2+99U#%oJ|dwBnzSFfPIaL$*vZBxO=lDm=H(Iwd7U{ht^-r7i<3+739lAkfg;D$bO
z^Wo8vWpcLFtFJ72b<%_TcfGwKWiOK%k>8gX`<G`fJ4P>zg`G|}J%9GhTlZn4=^Y<=
z)z_N%L+8O*xVX6NE%97`(#P1Uek9$Di5Qey45a0SIq4q@sZm|A5$|Xa=P%#c+O(17
zxHnFtb9AB0yxQ$l`rusJY}eEFS}j%hz;VH@7(AQiRVOED#3ab&YGO(sE3gLAQOK{N
zwzZC{dcr`|sIq#MzQeY7r`BW9FS#(@!0)+qC3GHE!*TX9{YEN@%g8g)hA!!arwSYl
zr2~)A+1NW12}*4*9Y+OrU0$7^nC}hM851AR6?Qi2r%RQ|?NG0&{@PxpM7?9kJv{ED
zp2{@0ARUxdTB@5WI*xevmiFrZ=uFFtOqRT|9(8s(m-hA#)US`dI&oS38pA354L;}9
zE|dcv7L*tD(d$cZ#xR9WCe~S5hdi;p-tu{A#SqNyD~ubdP}Wa9j^suUwDC&)V|}fW
znvJT)-CXMShN;VzkgAfA`uA9MOh0APC0j~Xs}<KH=YO$X<>Aq>&u->iyL)%H)o;gY
zmlO5NR_Meh>?FhS0cyY6s@h}`;n4k?wiu2tQfJ8IS@dsq3z1H(#|3dc_but@S+7Mu
z9%-Dh%_4d=`tyo-Szp^oE2>`%d#u(hlSMJ`tt?6#KkcVSTaENrzxXtLbbr%^V|G|!
zT+`p_2N*+#(x(UN_pF+up7@%=_X2;Li$|CFn8Lr>H^FPcqP^<s1AKrVEBxt?>@q-a
zmXz?92k-RT$Q1u}`FZ*!;>Q5rRi1slp7dt$b~5yL2Mk{e5`O*`rhVy&>_a_EA-K<$
z2RA`E-oLZ)uZ0_k-yt3Oc!rM{L&$Z$4&%-?5Fyk>@XRZBmbTqF4#WaDbDUp9az|{&
znfQU=BkCyla9t@S_fVqcR~#!*hzWi&-x6#w-jV+#y}eP_!m_Y4p-b7EAEXGUtj-{C
zd<EXruPgJCBZI~)ZLSE(ZrzL9Nw}Bmravw|Ft`l-IZSy3b2gN~gUcmw?m5nUafhyc
zFn*X07ctPvMgphfPMq@y9M?1964I-UdRn`)omyLLS;L;<5^^Q*lSf4g2U|R#N|Fx2
z@*E?I8VSV^{t&1!hH!VAanXu;F5zTbz46J&t8(JQsnkyzSv)OCxMuYbGq(7k!azj<
z@&4f1;KUXVgqQQZpo8%Z1%s7C<bmSpeZ7y5rQon_2!}t1L5chj&OIZ{+l9a*DzFqu
zDR?N5R5J<?<$(67>Hs|U$Y)=D3~1i252KIVY{|gPH~69`D03KnSm{N(dPSVlGs_J+
z%%TIArenPuZd!snc<{;=A+!@=%t*uV7{;P_fMK9Q(WTw|5mR9c$_wKf3J!`7hBK5y
zn4%o4{Gx7n7Ev|~X8`-ugGV}M>ybYVOlimyqvfKMACzzmmKY1YHKKT=tnfs;U^v2f
z7UK%u!x+UZSeIAI05ss0K1V*(7w3_cnX9xnN;yVDfWC)Oi@sU<dV#_#G(^%d)T6pa
zj0KcN(H``NX(XkPtq(yaA9z`+Ectwd6<ruK7o{+I)!0z|ldTS=dAMKCj-`B=2>*)1
z?QhL9aTB~r+8Bdzg&zm?1VW6>g2%n!Mm(L8tb(2quHi*|2@BeBL3<`|Nbxs`kBb|$
z;UmKEMmqadK$X(|91WG=TW$y9CZ2Y(H~i&3q2=2QPuY*-)%B5S(%1y#cp+wY#<MSn
z`A;rI%*F-J5CTe>!);vZ*lECvI|Q5_Ck;V;lz2A`01t%A>7U?<pSWdw?)i#1{z?+1
zF>U;E=e*?0VgB)bf`c2LjWfOU6Lh)EqcP$Sd{Fpv-NwKGOB6}q_-K!4mq}g@Cpf~5
z<K%&BJki0;JAl7JCvb9GMo#uRAcq&)vpg7+djvS%j|0UJUU+c=E{ChXnj~S_A35pA
zzk#0Xra<5mM+Wu<Q@EXuJ2#Ik5FPG6?(nC-T)GI8Q%}82hH{lSZh@aYzD3^L@pDC8
z1bT!KFp1}nBHBHRW6#82f?PlIxjxEIj?{bNKN;nRfxpf75AdF0(kHUpjeQg5n5{3_
zWu=ksee_xSi@*Ml=@)<b@6zW#`-}A8qc76d&cjq^5>ch<cR$-k`}|kZ>F!=HvR*yt
z{KzXVM!=MZ*Inh={I>A6=}Z3a2c~6zb9ld;{b3>GR~)an-T*91Q_p(PBJ=V&gx4qc
zb2&Jh=%_bLhdt>B{c$=x>ZQZup757aoAa`&tFCB1=pP+-)84^RIvI`BzcBzn^U4#k
zV60KAcsn^(qm1*%Yjh0s1I`}8P=%jKaJ;M7`2d{*d_Hmy2BUPae_;8S?ZmAwD06ft
zzsr=@m1N(T^rjevvqJ>(`nh;OR<<6ZJlGM#^00LewtTOq2IohKzZ%;NvITryZe6P@
zaefj8v?J+A$86aqev$R5vyt=!R!#M~Oi<*!GA67oTQB3>CyUX#_k}YY4!p%7XDY!X
z#vF9FTCE|OUGg;Lk$%MaS(J%g9af}6vekNpRW{;pX>ragh5EWU1Bq?&PA{dyEG$b-
zUlrf1%xWrMwX0~=)lRH#tFFYEOBh*bU(V-(4#oo}O=2Wu%ep1i{m%Ag+PSl7W7zTG
zq4ffG7r}_hg<+-HtVMgSE~$<bR^c4C)zqx5N`DxnufKWXtpq76`lxr~P#<7RNOUb$
zejObhda@;3T~f}P($kcj!A#p>gkkj`HFI~Xown9$X`@kbSw{Vi>U)@4;$vg2?KWcK
zAuGY~d$rxhdRkkLe#-WQ-i^a!WZwO(#spFIflED)(LK?bI1dg38NvrT-HG&X>Zvz8
zO)rkR>FXB<X;<T*)@bSZzGK~WS^epo$4}C~{PV9=SL3w*>`7WtUe#6gwZ*H{Zmgs`
z8_m>&chyr(b;j9O^lh~4CFNC7Kji$fRaSPvCu2~ZXUqz{GmK5rh2dQF$abC>_-N0K
zwMJS~->p`dNP1~QBRc!&^ipGYke=@!r)PUda<gO0Njm8C(u-aB)nCyS=cJonNpGwv
zeX9|4&~4H0Dbqjy>UTcJnTesSCZl~$)GtrQXYTjM$Nl7Uzj%%)&Za1phjZXCI`VwM
zm`$G}+`FO7U8Z%l%iX&>X-)kTgLGGY<#;gmN;Xy$awgizDMsfdm9eEV*7VG&r>)JV
z`T-Mm#jENJLm#`vcuZ@Y@K)8pbp73Y3zX;G2l&%{KYoBW2kT%Szq-EXo^g%#_j&!R
z>cQWwtfg*d;if+zo{m2jxUF2X@NMybjPPIn^Z)vv7M7M4|JGeu%=39+T7>Ujm{RyK
zNTE`OfkOWgMo97VW8pY*KuC@OiYOg)jG`Zep$N;kIX+y|--*j<l}4Vazgz@FoIjKr
zBlD*~LYTcmyzvQfnh@{_Bb@?}%bdQ@o|Mn!2t2zfqiG5Ud5#}I^c3iD@FwnS7dSKN
zA)zb6a0g!TmvnfzyPjQtpyWT|e9z&TUoMPRu5ZU>H-F_$xbcI1iW2z+FToE)xtEJP
zpg9ej<=Z?u2{;Z4+BkBY@u49F27|>Mh5!j#bqUy;+uNxn0h=%kF08Uba7KAyuWrs<
zLg`{}cMMRXWJL_6iES+~NMb~weNht4p9DA*ID}1b0q}WVk)By{2;3-u>>pp18@>?t
z;1wec3J&#0{;V2`1B2u2r=}FGt+fq*a2XDI=|oCppSzS#A8&`kD#e4Dgx$`u&zu^L
z2I=(tM7c|}l>o_kRPf366aWG*!mSNaQgl#gF(RVau_Xp~w2n9fjWWW^l4t%gYYRRV
zY~eMw?32(hekijQ7MjB4N4&?#MDxg}FlK8Og_kl@XM};xYBBl=^^btfxZuMRkA8&k
zFhXHu60Y*0tvDlqnVVLURbPc~w(p>PtII3y8?-mZPMDzHp~Z|hDgt8{3Icw3CSMFz
z2;a<rMR{b^omXB_m++-Jp^u>O0-V*h%8EV_pnQ~tGP}I0AM#|i7h#kQMFw2_HlC?Z
zP{%g1Da?`ZsxmOv+&4ux#+(!sDZfMY1<$5qIw6(hJk(Yz>eEuzP%xlLT8v)fp_Cf5
zl+qpmcODeiym;un5ipB7i-*T;tt)QfnBqsM1MCbpPJno$sL_q9e^`k?{5cF1`zJwx
zpM!IN9XT}c93<Kf$h2`d95*-pxlW5)mlG;ixk>FX{=m_Xqe&Zye9iJmoy21){MYbH
zfSqv<>bmE=ohKo38ztu-{<sv#fB5+mj(o~Er_TO@?<BsXZt9Qt_}34F!kwLgTkgb#
zM_1Gpb>qJr8F;wy%I*GG(h#ScY0u;pQMyWG!OQQ*l9$~?;LNmzV<Hh9Zu>#XffHv*
zuiUe{)M@0QaQR(x8Lz~b@fPWbn&mBrgUf~6EHK`6!>50{BeAcNF5D3pVTJb8uUmG<
zJ!r!z`2=lcPNc8E)tx`WgkWskdPoq4Gp>v~@q?b*2REd3FZtM+{ec?+5%w~jDC6}o
zgZ;wKe-3gHqFmEUf5wY+lRBPB%ilDGI)-l0(T}f585KJ^8GkT*{Ye9@4Df&u=vlY4
zR7uN~M!Nse=joUK_J2vg_~pM(pZ(+)>EXwpr|mlrQnkLO_hmJ4e%GAd*_IjIPeqIF
z<UW-B(WL}IpSW<(exxh+n~-^$+<!=br&sayhlM|0&b|lu9-A5&+#s4|vQ!wKDrrG7
z?qx!!U~e}*&!gX5c<Z*oVCYGCbX=dkyRe*2_((qvBbO&gN&mrcV0~Wc>B}L5UnEw)
zumX#%0%Nk<dD=hVyr~y9Zea+6HpZC_+Zc*JwjpFS)Ng<L+qAp8XCo44SZ!^tr&gn3
z*_|CWkZUh5#fQ?P>nuxGvN1z4EP9d8_o~-*HEh^I|L}I;=ord})iW4oVr7_QYUu;9
zEqPpC<)u<W8#%bt3pxjS6LK*|ue#_n*{$0fq%XhvI{ouM|2qBt_rFhjdxu_y!$<w&
zZqIc+9G#_3f9Ta~eaX}qT9E78YwM{}sfvf>Z37PukJ9tqUGY1g+Ux7m4fKpryJ0lK
z5XXu&jBixfigZF~SE{StzOE{{opN<LopdRk1D#Pxr4!D1mDc7)JFPYAX-S2<M9-5V
zh<-x-QMSMPAAg_z%m4PjrDso{y6n_9Mm&sa=ty4iqY88#i9UudbU<&3uX^E8R;@@U
zraP&P>h-3Lr>tn>U5ai>+rj_Q(UIEyAdPy*Y5wXgEsF$aLM_g*nvPX+^Qp43=yu=P
z-bm}Mri!d~Qu!_~RDMq+i2m)#g{l$Sk?r^nj#Q7wUFlpRsyaqxLg%8A(QRW$Tv``q
zhZGEbtma}{!4n}bs?OKcF6*0XX=80o?IS%~d?O9C8yFauQ?GO6)n=6?>653!R8;}i
z8dVzx8DrKvX;fWo`-r|r-fV$s{a5sCkW_z!Kl%e_+3oHgc=Z<ivP~x2WwL^<tn#9N
z(_W<QsxNUa94n<L2W?2dfj;s&lQ8&(P*>leJu#@Y+j?)Q%~*BDicPj^WHnc<R`-^m
zOaSGaJNmn7S7QuaUG1zg(|#DLF%+_63VJf~FXNuR$Q^x|c3}l1V~>1#U9|<_=%}o~
zWBe*ls)z3m`}=#=Z8;0CMn9%MH|mMA?Ksc8qJBi1=z%5uRCS_yq5rd*5gN{rfo&%}
z;R664YT^R`=s)ihK8(5V1Ky+#ujAfc|M5Khu>)n};#+6&e_v2&zv_Bd!0<I!sZ?Is
z5@ccnmI@d9=!XzxP$~zhv165!oo-e>WDysk#p9Up&mv|l!^C?dJcF}Q6asw#o*lXF
zZOGv7dU7WaD8OOYA`Le~tTIdsAPBc<hfiRl`M1OcrI(&(Ga10Eo%+eDiNm8ZBa*O^
z{Ul!CauTj9ocb+wTbvos>1aFrIUF?P5uns1K)hwnDd9=e1cw<BS1te{tYR~okS7=r
z%RZc>m$(@r8Oa#@$Q4^l6+iA=Vg(dqlKoUF4heb75_sEpw$r^&KT4bTchZF<W>!04
z{z16(N-ha;<`n@~LLR~{)Y)2qZ9Xt`Fr$!FJLG{uLo%z<Fw5)I=XwnWeV-47(L@M|
z&FxS&IsCZQYD<`|J3kCF%+dmEc#;r|Q3VBcD8UzlPQ6<7i*7|iZ@bY<ds3pl`R1F{
zIXuv{FTwTHLMpTLFwUS1vsD2#2T!cpYD&piQhMl78n(9}&9ZP=t>vxx#0!cWE8b8n
zDI2TCC^rHw1{@58t`oHkXM6d~3xs7pqKn6}y30lyW@n;U1t+RUR(44V5}w0F-yf4z
zb`;NZ@!)nN45dVU10KL7t+yByuhc)Y1W`0DCA71W%Ze`bEMeQL#>h{6aCQJ~N}ph>
zhjHF+0p)K=Wum=?CqwadCZ)0}MWNyP;M_E}<Uk?eOfjxym6d)LTdzpycdMZourGVh
zWdjGr1ZA2Tk<g+%@CIMBIlNLn!l8GrbYVQDi{@=0b@vK2wG#$Y8@$xV$m0~{ZlJzF
zJ4mq`_eY*>+3WPDXM9OfU6kThZLLb#TTiX^HI=!hK7Ex&Qr7!Yj!yb&^Rt`JtAOj?
zXu%l&#4s`g7#9WLpJC%TwXq*9@;C+IIld(JbHW~LI(PzKKR$a3e@^GP;&!^gEp9(Y
z6yTRzX2SjnZCw+FA`!nO2oHY7OW;r8x(}@AvuYB<wKDOOY2r3D`GarqC`WcWEbb1m
zn|jXw!sE9<@y*u#Gz^X+O-ZY}evF$LH0^63{{#tN?h7~Ya!X`dx+eIPg)-S=U-Xg4
zE$$Inl3lwi;iW2PQa>f!_kx9A?i|l^)=sC3GG#72=1MBdAA#n(@wo$p(GISca2J@Q
z&vEo)2k43^0^tr&92wx^2I{2#bU7-Qy#!CW2yX`C@XK?r<P8+`92B>}4txhFC>#+p
zLC=u!a|AxGl65HlT#qGDuDggS^_c5VKaSIn!yV?|9N<`oml}`Y2rsl9;~=-3yQrUr
zGo;uavf!GQXxpzJvc!E2azNno{15+>N`b3%aC9^q@V?<CRzJY|2KxJw-kqyyN$>T0
z4?a$xe)eany}p?$m3pcme<^NmfseZJVKJTq<<f6%HQq&lKa?5=e}qlO5&c@7m@9A`
zn(w#4AF4}C%lC(YY2Dp~KQw3W6TIT<dxB#4LRPFuXW-*9vOH%O4NlHdf5bVk=cz9l
z@whwimSgN<z>1A#v+7kFoS#&y*3y~eZnpcIJ3mbi?rnNoM4p|{V71Kt{y{qKbiIlN
zLl_1oR&g<2*ogw84+b^1He6bo7ftcQO04Q?YS&g0t6^CA#CB)@@|)kJPPdc(?ce-$
z`sq*q47Et<hc?Q2J2c%#qC@`6Di7s5Jq>x6kM?ZAhw%;ho5_X9!Wio?AXxW6#?~Kl
zH5YX*QCpyAf&Az-u_c~#Iq6s!dg5JD-f|egF08+?EuObRS3c-x7zofMki`)R+N#51
z;9_!MY)>!UL{wNMM?UC(7y+R-9`)1C_Imp4laKtoxs)!m-l{pjZ=UYMb!xGCO?nCE
zRdL?d;{2tzQRF<Pc|BJ!8m+9x#K_n-QqL{53Exm&Jb#*=J$<bDucVFj^>k-@)4E7s
zvi?8((?6%LA3sh{p6{ly>hs_KhyR%V`mcVezR+~L4#f*+RZ#=b@&vtebQkGK(mn5N
zujxT|o(>K;lWob{Th?nBJ6PSr+0dsx7w*g7{a$FN(ow7FC!c?A-Ee<@UpncMw}GVG
z-Ci$!`S^+G%-aakRewO|ZAdT1Ac_3LQ35N%g%0PQvE|@Mddg96oO<I6)j=g~iihO|
z=}yw$=wGWVY+XM}91Y-cq31g?#c({d-o471X2PWoIlqd|$9Y}s?1AAGy^+<5oWX|f
zwX&-Er(a0tU1_w_a<!E%RM*TEL|3D3&|_KI#8@B?R_n3Ci#q9QOdK9|(!S~p!zO)X
zYjZ8FX{@qRi$2WRc0S+BZ8{g*2}&<xRo6&k&u83;7P=?j4ZNyS^}{=c?*pugq>#;e
z-9}Ulq@1_cSKf;Y>?opVg6fYmx;oMk8AIru7|)gz=dJb`Cs(n}{!o3O<cSViRlXi;
zDzDo{w0v}ea8~8t)-z;P{k+@l`aEn5gr9%BlkRTTrDN)uA%51{&9teu-%z=$>Id|r
zj{3l_zury1eY~HJ`qHsg{`+?}(#jHNxW)O^7)?nZ6Jb9<0YCY_zgNi*2Jd31p})UN
znZGq}$dkt(8N5rK-7J%3(i`Mv>-C-TdPDfO_AK9R`F{7~gu?gDIwky7WkP?*PQg@%
zn1#?aer#}aY?ZQ=@Y~D!e&8lAZwfx+t6Hs2sbt>$VAjj^J45c|T$67`H$mYDsAl?R
zCEc1P0^`GS25@d94d9cC{xtB7l4-?<*XWETSs?+R4pV?X$Oz&r4Lc|C!ET9<d!aFf
znGo>9<zhN!5++{EoVX%R#G~eBfq2s9vgDBRQ8e6A4a^H=%lQ6Pyvq=?q6u{Q70xfY
zIW2e&E7VKsz?(3B6DDX={D6#nEnT7e3VJ0<Kd%2t-Q$k9fN;1zxFk$3SC`ZJ&YiUL
z(Zke|av&);onKw}krK1ape3s6!wBRUuDoTT();YL6Sl>;k|4Njp%+1vJWG3!9^tz$
z0d&lM*iul~H=HxBI2@W;Q*0+_Ly81Y6tZ~0lc$sp;bBzadkP;_Fd(t@0Ne1f4Psr2
z&5LKx)8nrnOA*;iyWf1BdRC0YzXbM0W^IZu`eV#`lMt`4H7T^KQex2B$Szi&t=O<~
zAwid|{@{f?5aP{?giagpBrqemQ|2(ro!Vfxyuh}Q2<;dQ*$R;vZ`7pfLHQ7dvPi%3
z79*+$W{VMrvOsN<Kc&Y21s@8M&@M$nm+M44p%k#1i?k6RAF80owv8xjrlYvnQt(Wb
z#}<UF{>r>rVPOkOW)@y~k8(I2pA0>F4|>dw<72?dXe<S-BL(18ihDIxrHoQ1q${hx
z+@C~&zHxdIX8^!&Y=tBRNW9?(H|ivwN5p#zODT6y@ZrgdrQQ9dR&+xMK@}E0=cS1^
zLOgoaCs5LcgK;|1b$$~4^JK)V(0&?r#!{@H%Zj`CRI%Z!EhW9}m4@`|li^7k9<gHZ
zG>l|aoVg+U$a@yts}+n<5_go#tVSA#$*o&_;Knh*bR6JuNZpkbavLWQ?Y<<PlSI0p
znePtKuk)Bfi6hiFa@cQ=+&O)Yl_wDL&2j$7Zr<cDP#>8;A-GxyAq_m*AHdm`#Q7cI
z0R4kw3Yqo{K0vrJ<e1PVHv%$Hm-G?}P?AsavpdJn0=_T`fm?VE{8C0pM0m#26;OQK
zff9cmtH90BD6t)ih?rU0<Q%jBT;Yy#2t(e2ZxZ1&9_|trbRryZyqzTlkW=yRxM}Kg
z532aWJ&kn&w<ACsVtnUs7D~NN5@u&il@MG*zlk`T_$Daa@!2~I6y|1!nn0#U^sTfr
zLEQ?E1tZ6VUmpC0g}~vr)Kf_tu2L`Y1`Xe|HshD87_$HuVdjJIM80v~jg~2~xfv7?
zAsI8f1GkX#Xp{%k2`BbXQ|%odaQmTvj6cA81)5X6j~V28w^yo-w0-AZ`sDL3bibc!
z^|r!mlBYP2lm<~hkUJ*WC~Z6->@R&Jj1}rZ?kP$kJmbzZXY?u3sn5#I6*#~Y?`0@B
zZws%Z{W>W5zH9z(2jyw;Du|kY9lUM1-xc1CukXX~70<LQ$?etDkPg9_P>71GQj%cl
zkMbkQ&}=()BAJj0UTmE<Iytw23z_&ra=ur}2<WGL$UZ(gPG^#X*?#TwPd~CTX^4Sq
z#Ccic)awsz7(t4?yp&X`F;c;JAm|T9l0{?15=J7%5i5?^%I@c1JW5}DaxYb+Yp@f;
zNHQ{K1l4M*{^-wEa9z%w8lCyvBOA5E8|6foSl}Ed)f*Gum_*j;4$KQ%wBbJ;k5z_b
zeJ6IR&=Wv1Hrur!vo9}mjupl%)x8iYFCX&RrVm36ANdJGcVI=?g6b1v+6B6Z?r}yj
z#=fyvbfH_Ib6`Nj5XOo$wrgh<%!=9sUNNSy8Vdu|xo8ZOj=sZ|ha2sB+Sy)94b>fM
z&T7k?Ked;B^}EMuYx_?6%fI-!w}!j8IMaKQ-6WV4$@cAP%M*ASr&XyTC*y(ZvC(X$
zRp}>8M8psS{L?@FWBRZE`+rT39^Fryn;R<Ia%#5P>Nl6^sr1Ib|DV57S%zt?)kr&c
z@22+pR;o6n8`PVg<cHx3qeWl(&$AaVtULYR{=?s-hxd2V*Iz$NyZha=bMH>t+*<c}
zXH1IR-9Pfmrm@m}{>h{C@ZOH!!IX_{CO5Y>(wg*-rur;Kp@UA>#!~!#9;gk`i<rE}
z7L@2q?CnusUG~<3l<f3kA&oBPQfCnH+pTJ<udpp<th%DU(e*IkoY3z%<4XM(9h$8Z
z=_7sd2Oe#4e0ZdOn!Ft!eVK9`?Cz${(UIF<N)*~w+Ev&;{mbvt_+mcYdH6~C`CtB(
z%74##Iql2tDnt6B+>8+>Qq&tw^#_d)6%C^x=j5Hy52b5iobu`_wV4fP;+d@#Tk0Da
zwpdZf_MGaK)~ngNbgZxuMm@DTV-4kQVPzHaHd4`7sXxYmC%q~!bicavd$uoS0xiZW
z`q7#2d1awzah?}-&sfB;b=)7M{lkv?EMI#t0Me)VHW2za=alt40h85;+p4>rtxfMz
z!-~Xnm1}#enI7F+lYXtSC%#xsMtv?SpN8swz0J5>PA`sz={JuL(sT7E#{62VE<QA7
z)Yv>fs!Gu79;f|-L+jE^GNzs50x10~L}&@WJ@z`zj}aJGe+c$1)0FD_fx$A?JCyld
zj5_a{_q)Km@cqXO-&Uu5e>TpK77Fh-+mG;9m5VZGr(mi>%mNbFj}KnembaJj{ldHU
zD;vJ*_4+MTis6JP3YJmy0=J7FzJ3UY*lFSAhbbo^eV~>PseHh2bbMn3+YVyw<W};~
zZ_tg1TkBVd;FdV>amH&ks7zfM3a-<>6niTK{YV%&kSeE>cgizf6E;@)$V?nG2y>Y-
z@+1vGs+erD5kMX6NE1H&1tW$C&yE6xAb~2N4LZOiJcA3L6HeL~2*kqkO)%E<M-XZ8
z&EjA47gWH3`kASNf*g5H`vQF?D959MB$pD>d66vDR??;f{yPuvr3SMbB_LAQd}3zC
zBs0t~L}5(ftSgiqgih*%m0{-+=Aeu)hGN9ZDOM{nLyDC?Y`xc$aL87J%qBx&Ve33T
zpkjEkLbk+NUdt8$sY;AO(8us*Lz#qPpHC&F%3E_RE}1^UJLg`pr9@B4Dyx!y_b<On
z-~8rx>7+Y$Vdo{yW9-5+w><CJRMlp~!hfSJfxIdqSBe{ovipSE5Ca#2?F#$vlUC)Y
zObF2WiE&4tusR|^*ma~jcjYP_WkJwa&754u%)!{cfqlnC&w0f5eBOqH_ECA6QRh0)
zABr>ERPgbWGCEutpeRu17~Bx*5zu+QuqEMDo(;B$aY%fCkAjcF!rBwI$Uw<Qh>uPq
zhE*QtuRqhhh(nw!YHMcCLIXvYd}&{{i=Z!HxMZAJ8CE}`Ebxor?Seytp{vVh6EG7`
zX=pQ)8)nRffmC%SzULCgKI$XdooxuICyak=dk7D-4gG;UDFd`QXANcOOo|Z)ji2f|
z8=w2bz@R5(xHn3t1B|9BtFl{=^02J_-d=BbhHpg*&WdQ9jm}cP(-*Gln>M_{#1>=7
zqeglA=HCET=H%F@N?;a&KsqxZTp}k%FhOw{MIMLxM|4SwGca<B_$61MJSptsN8U2-
zfrt%P!JLKx;o}!~gwOi*N57aR2#_D1sS@HQ(N37_%m~EjE*ds^$!mP$fhE7A>5bqJ
ztn@;1(gMcMKsfa1Izi|bFy+a;xXN$TQ326{o}E+7fI^=zxy2OilDoo`#~f`hG>CUW
z3E)PZ=0v{d1O-75WzK08T(Ao&`P$)EkUZ0*GZ3j5M>&BZat}Z5*~GYJYpA$Nc(mqq
zh+HzBpLYeDfN&e~t%SIqXZV`N6jzDM2@MhB^pkktb?E{#xSz!gn5~b9&6LZJJC_S)
zf2lKb0so}~6mU9)IZaM=vv0aT<gnavpj@iMF&V|_vp;}9BEy+OoHo)2uM>2+0xcuU
zA92&rX{jQF)Vl{C3}1iTFm2CSn6~2#%t<cN`<oBA8=E`n)6ajJcJ4mZ`>v5Hl6RQd
z?yutR%hP6`?I*yrtB0d_Cq0UH_{+J!26&g>hF>U8^L{fZG~a}mS@^bm-;A%>GS2E5
zHz8sei?8B4r@i?+nSj#e=gAKX;>X#(j<0;r@aBdt2tY^Ir2OTc>8t-ydy7k}X?trc
zZ8TSXX3&XbV?L0xrQ4ZMnLmcS%yaqlBBid+;6WCa46ny0D`PPHFnP^8CoBkW6suoY
zt+l<qWdngHm?>ZM6C237q~V+;=`od6$@p=;EJhE`m}OEK6TtYWkMZa3c02vzryr+#
z+v}po)_5tg69yBv&~I3Qa(FyQyGO&c@9o)=4F=1VifOUZi&ZB+J4^C0hALJtaSs_;
z39{UcOwMFJWN!S({umtQ<wrL`pP(L370$(3N8af}b+9D*wR*)o^1;3)ooQuRa<=4n
z&O>6`!>(j&R^(s^V|5beCh<aIlHKKH$oS}qZ2w02c+1pk7~tyOg@Z|mHl~PY41x9P
za%xpDykDjLqn>D7q<dR>ezoe-i!MZe(FPg}SDf**cM#__HI?r0#pCq!$zwh1w$kSI
z9UBZ)P(SM)fAv-RFaO{FoL=lcOAjACv_ACY@e{8!`d|Lze@JU<TWL5LdL`7}-d=kC
z;vjwV<hlD!17q0Mdisk$|7rT^PyWotQw$;ZcQ$-JQ>UYPP`!=Rzs{t4wNz(S)hnx-
zSV7lT-`NlidMSFadZGJmt=UqS*0b|UdXMxM@?!NBTd|`La*o<S{SkvEyPj|+R7E=K
zD#n`zTSi``zVbaem+Y+eM6X!U<GiXiLPn&YR>TMGPgRW6uP~4u9S>Cx2dc+qpMAEB
zrAxqhYpkf7OVY)ZKNCv_qft5;45S~;dmBhpWQ?4wSo=T!?tiBL{lES9^hC5i`uxw*
z-~79OuRhRJeoV5crYpVyh%S6`v0W}J_f}~?^;_DORhRpRJ+G?bU4+ibw}VBs|HUcB
z(M$8u(zPo6mkE($)qi(*miDFRA0BsXw57hGy<BA*LMDH5w1C>=gk5JWFnjD=NH1M)
zVw`H&AV@!;k9q~A?p~><afA`eXNxOOwqx8qI5MrPOX;|3Q`(5tZ%ouB&EEc@a=1$C
zoZr=m?J}!M&sLGFP^_>$Bt}W`x5V>9Ecq<*ZZ}<t-kP3Styay3%)QP@+SRx}?2N=S
zI`h1B+hw)S-HkQt)=zeKZ9Jq+ILA%}R+Go1QTWIY#(V+1twT1QLd099HI%eJK5z?v
zPpn@vtcd0{Sw&H=dzb&oz|C_15y3kfzFt$d`1Y4Q8D?dc_i1!9Pp_%+KWwji4P)O5
z-^Mw#Uqzp^-<JO7bM40orBA+(=ga#Z%Qk#9nvJVD4HiC$`(u{N!HlqILRcvG36#JL
zVPH|<Y$b?!qdBakmfOz!8N>hKj+bD;<qpKXkOha2^(YqlcZg{b#9u~?&?yA<*^q*B
z6BD=|7VKkSARj!l7{NLBpcQfO7C92ihwn^BKY=xcNFBdG$vN=A$D~ty87AobL}=ud
zF?{1kS5(vuATGj*p4Oj(?2hM}ejF!vu6H_y|J=Sg0Z`IHHh%2S*$8sjwYcIIEeA!u
z)E&$e0Rb#>yQEV}h+CB)z5U>R+IjFG)z(`{I!l_fM8&Ej3|p-J5#DlaD}XTqVGP3u
z2BK(>P;}U*8)F1Br<nQ2e%*Wz48`&|bvqp^8N5%}Pn*MZ5zg6u2jh-cdnppS(5y<K
z#IS>a3(q!oNois~b+)arF-c(@L_HSWZl{<2>F@tO{pKJ4A@%qA@m!=XsAtjk9{E)%
zS!|)vYNhQv+o`$UHV$W-g@KIKMp9H&W8}xh2cBZsVI>l^;~9<$r#?@mX`L#5iR~0B
zQWC@)S)ZuxP>4_nn01MfD2z-3wH*pH2E(#VAj%`F-ip!)EX_;F6>rScLy2Zf2kORK
zdPo`Ig-lbT=)xC@q*qG`hx{lULjU-LtsSJO%8$}aze8wdHYF?A=2Ui8w?%L$ssM&h
z`YXKB*ch((aK?%+@Yz0s{6=a6_y)(uN+~b28|6cpLx_M^>P-&VbC?Zr)T8ntozEjv
z-+>Ma3;AK(LU^}wuM%JY3uB+sh(5FS=#v<|&ZHEaj80>=WbY)6r0}5(Q90COwI&67
zqwZOsRVk`cMbd>7o3ZdPd|k5DVJ>@ddE$bB)-LXu*DgVdideO)>T8n3kP&<d<TY*g
z1)fR@D)<?N{yqgK5C}h~Er`X)RGul0Cx69*0Va^K{Y%|WB5+)zok1Sjr|=QzK{cp{
zufR0Nn=nli5Gk+Uff8S0%5TK`Q74$ij<}q2u0Ze#jzPgLt_hU<GA=&10il^g`TaLb
z#0WV}W@Hkt;MqKbR$RE9%7M8o1wk-z&a_<)(4Y7*;ZhR;Iq4*EqzO;H%|J=d-M_{+
z?cxW`VY-<BZuuq=!58Hv2)u~{8k0Z)c)m>#A`p;sRtSqwc;mT(lc!f1QQ?l34Im$U
zb9zBxbmp1oKJT=YC*JL@j|@@n96()0;RXDVN9G^Uv&T__T<xN!yBxj~I05M7ddVS#
zMY>tM0;Uiy<7BURj<_iVmP=w+B%P+GOwk?@1_|d`@Uqi!`sH2WdkHDDbhC^{q<&uP
z??<v92At~$coQhiGh21^;eh$YRew}omHe@`v6Vjh_|tUf-Xp!^+G&+-3>Vq{jMdUi
z)+yBZEk61c_mYo7+&=@t$n`EjKD@(kYx?W?z8ygKHrnqBZ_eAb=bY>Gb_mL^<69UW
zudjh@aR0#}e^h=Y_&GRJt{X*sy+U4St89K?;x1vxeey{LtX8Ym2RK6rxwchZOtn?@
zZ^@W^{2oe{8`?l{$tQiKVPf5U8l7I)c*6uKwsk{}@3VSjG)_yB!7+F+aSVNht@W5}
z=k4q&=pKt+5yJbYr!=g5x)7VJx`9LV7$$IGWTJ@3&8MehZz+CpK1w^At@QE32cjn$
zl6|kFYoH5t`~7rqa3sBIA@xUW$0nIw{89KTWd^UTCZnXR$l+WdcFW*RagLlOS{N44
z!PMz2OC#r_zpzsSx)Joyage9cFFH&{6iv#)Ia_?}Uyx3<w8})gMd>WkXQaEZN@jH#
z<HWqTU*xQ&9@~i{Yl}7$6KN}SfiubM-R>X_#$(Cd=boHseL;CbvyMTHt>{)(#G~H+
zlHd7oza-iGOgym)iR}m39`WARTH0t~_?VM_HEpz6rCLc(Uc5-Z{>|@Q=XdXHq`3?6
zq5NvKMq1z6N^B{JA&ZG{U;qAh>7W1Me~9m4YBZblC~u9wwXvQ)e)LGXTTI^j=;McU
z%Jk*$ze<mvJy#!_PhWiTS$g<j$LEOg{6o*9mD`O<YD<qeI_{;fp1w%OW62|`zgBfM
zHCC7O%x%~hiyBScNdfY<fsH25=Smu=jySXG=%{Ocyp6t^4dV&BP%!C{v#rMBpRMpP
zIyR~+sacP0AJIiQ_X?v8jYl7DVpNn~jE?2D*R$Ktdhsxje$(lxZb#^2@zhzAF3gHH
z%fGZQ=N=O#KIlgnE9n~;e{0rVljtw09cnfR{lh>0D*gAr`yVzGszZtHW%@|<`PYB*
z@6+b}2cFc5A&dSwKDqGt!U&4dmaPm~3C2oH403cIjK=Lc2BRgPi?)Ao;BO4`D*q~a
zW>tE%-fb6ii)nnuc8{$1J99dWxIUX)`Z-&Ko}bNm>|>b1;LU0;_^*lo^=3`9R#orR
zf%>HS3eN=akN_!%cP-Ia;X4L>5koFxoXM3K3)k11D%032`)C6uYGVAQ|FAnszuR%!
zKfJq@cD6TD4WrzZ^!$^7a*eXGLX!!h)s;Dy`9$rq*BPhh2ZOZV;hemY$f5hrrzN$^
zR-2<B=F_oP0iN0z=+%GX<LctXI2Pw<Po9R8_YP7Yy-Y3#fy8h81t`GRj}QD-_?}q5
zX50|XYwsvUz3yH90Y1Qc0T0dZZ)}FnCwjVnm);y(m%mN^Q2(|ZT)1$)1m7inJlB5g
zP@Xe?i08}e<?wF@8@^i2mfkJ#!NH9thu*2#$47ErN(y!X<yQRkWdt&EO!?RoIOPTy
zRgk^cVFvFexQpu_<$(MOoiYGQI&PCV<H7utsME@8Jo7n?Bb+p@M|qrQP7sxw2bx|k
z+_?%bH|rjZ2*QNOZ%Hi|X{>Z91ll>+;J8Q&Z!T}d5J8Gu%@Jxmr?FRn<9eA3M4$AJ
ze~mG52#mnR%)4;=rVOrqEWMT{yaCx*n&YCcO!<#wB`nJ3^I(X~Wrp)C;y66q7K5Pf
zJAIGz=V}L5cU4>UbpI31_qv-_Bs^mfL%^bnQEI%Eh5QJJ&|=mfvj@*FnAs;v6iRmv
zg2hP1Uf0Z=qz=3?OKA~MJIBZAV1LgB5@t76RwWoOOW2n1w#dpVN&YB}s1_rXN#)yW
zH6`%ZBp6>v*gdtO1!WgQlDG2{?TXU!fu?)ZNnif<x9Ok%{(q)tzyBtk4Nw|qKm*Y>
z7xsm(N@#3V)4hB5()yiE3E>M;_AoFl*oZ?zp;&X~)uIG+ms@4<9^$H0*Pw7HI>PCN
z+5san#;7GJ;;Spn%3P7cjo^GDx{Eeyfls;EN`(5PjS!%rjbgZ}`ano!1}F87kqDuk
z)@0?==wxJJ+1mlC{N#bqjgo_L$A%#;3M#;WNO_ordX7;_d{WQIcQ!OB55iHDFe0)=
zgH&-Tj3bqYI#9jn0uQ6x`T2>J%4#KqaJFtB9e7q^MKhF!vGP`#P$)>tim|g3m%}UG
z1PJuNg%kq%4{`CKpLTMu7LROCDcYG=Z>7R)JoPJYxydTGGq##Mi`8CI@Xn=(oU$d&
z_$&>3gEZ=j7ru-L<`(Btr6J`<3Uzxu4s@n9FHbIQG#mGXD>vf=a4OX^oT_>A51<(8
z(0K8S!8n5TM?&IpFbG8>MripB2%0Lp?z2V*T%-seza`B~D1n4TgkUxYqNnLS?&*9A
z3QkIFT)4}MBge@N@9yJ@n?MdP<%`IP!@Ss4;ti!fM2j%j2jM{GWrj|nVV`+|k7<%X
zg@4L~n?Mk9!uS!87F6=hamFQX))PETK)6n$pb2^g{QNjy14sl+Df!=s^5^>UwtxzA
zup>o3cPf{iAN~8CI^!n+*pn(#7<8Z)0fFi;0)%+Q@FG5gONSp`gis7#=9@U<kgr?{
z!d(c&uY;A2hbQ>Cf9NlSV%!1Klm+k=ijf!KG8d5B<0RK3xDGDFd@qjdwKMY}x5J3z
z2Pk3GhxZb6f+%u<99)8~^n9HpcRR>Oyzz3Mh+G0EatP8nVbqEnHe->t(7+>ip>5m@
zIE;ixEt%cdY2%C>Oxpp0S2z#3NN&Xio-s>eqtB^6Jg(%)k6EXFzPFcsKL`ZFAK*1m
zn&VYfrT66WYN|KaQ?s$2);71&g9jg{jjf$j<Dh(GB)zLSzl0BaF)=3k9)0lL4Owpr
zl;eZpYZij?>-ZLi9|Es|AIk707O%J6xAmVJ<+@pnR~6Y%Xsc|Khi4E7s*;N(cVcM4
z2!!#Aj{v+tL%-|`bLV<DUL;mXu$|xO<w6>rUZvjnG#w9EMZw2yjRMJ!(4#nfFy_NF
zaxrHQv6AFma&5cSNEONPN5|cCaBytn!{+9ibQFxooLj|*eO76m*!XqS>3afM+1Y`W
zP{^(rO@x*%B|Edai(;&|*Q6_nN6GVZSFxo!+l*r9`0Dpxd&SYl)}8d}r=MEBKl22-
zOYwJ_I;<p<j7`5HA6A5URTuKHN=F!mG;{~r3pp7bWQ<WydKZQliTje#Pppe^=2oo8
zlD>m1FC4ag!&t_4cW08NPew!4&y~x<c7kkWX!@$fp6Zot-A*L)2dM4TUeqD+Rn={+
z-tdYzj4hSQlDCj#m5vQFG_QEXFm-rvWP=ZK1;!V)_QT-CIaKS?1zU}Vbf9tC-#_#=
zbQsW>T*+34sBLjR;ey+<*{FC0&NsjRU3&cWmuYKz%LYqU`>^#sypdiShIJRV$&B4f
zR9}Okbum^PQ7-g8b|qo^L+UVHo~2R0lb$_$kvhYZbUI&4ossl4X-{j-N?KoEcl)EW
zQQ7d0VQ)|U==ts*>W+<Y)bW|h(-NQbw_2r2k55Az0L3E)EYVz4-E&46+Yq9E_Xj=e
zflbvb+w;fPetD&m`U1OXl(VSNjTj5$$@zAyreku%qUsi%)9(@VXxc~lR92*iidPIP
z^y{ZD_SNs0M0%N;)s@t4*5vP{Ubm-w;*93w<DSpH8>?Pf6}M1rrmda3>Ce9S#49l|
za6Nmm=d<xJuCm+4lIjzKbhod5C?4DOO1isI_m-gEW>Mpiv)tHfaqq>hXbioInQcvx
zZSHQxiZ#B0ELW?kroO@X&gW;+<)r%el_y&~qJJ{Mu_b+<NwU^cRTfr<(Z|@H5<e44
z)dg}ruv?09v8<*lQh2Z1PrG~js<Szb%a5!N)Atv=E6kFO&z$w=$*>sg#ACg(;xSXz
zeSZEj4SF3L|LXOso+ElrsJ_L5A8Hu7pX_$h*So!R)JI>(U>>{0@SUTnGVp8}DZRIg
zRM{HpH(nv9=Lly9!kf1R72ke>mUl;fN2t`I7h(JsANBcTg(#<;?+hOdUmxI)4%7u>
ztZMwu_Q$sX%l+B?UEzb_>-|G{&io;sFV|)IHwPQO)><u7DcuXt)Q}}muwzQVZRJEE
zCHyrwM3vxK_ia$(;fa?7e!naP<-JJR11GwYKceM0+JZFkBgj!RUV~}a6XoT17AM>&
z+;9>Yj=(*6L5Ft=@$tq%pkK<J-*lypr%6EIj$*HJ2sb_;>6#Q`3Xz^qfl=PbBO~Qo
zV3?>YEnL|TS#C;WS~(cNkJknWGQi`?k2)zInGGcjcT}LDgo!WmF<<Z)+&D*qfE*We
zoFe=R=YuG9OV*%yVCm%!2D1pI^`+`sGd=qJv$VFok*?$?oIZ;&hyB4xvZ!((Boof6
zC$>RA*tYRRLL;-K$e9^rs|epHaT3f!fmxAabDj<l4}As~h6=VH^vW$M7WG=)h9?A4
z_UVVu<8Ig6BWgTJ*k3c>te#roLywdfpMRx1QF1C0q^Xf7Uw)JR`~UyHrN_VdGMx`^
z7`<j-PSWkdG6(c7r|q4sw7GL9RXr1L#npxZ4`H6NVz@!6<U?Uuy@bIBg$8Aj1PGHh
zCQ0dHdj~X-m6fUlRSpy8pk(&>2M0xzX1cUN4JC#CaUvxL0}f?Dne_hjrp=axp*#a&
z_`>)k#RB14VXPvnOX$D6#MmT7Re~}rPEmaM>WgA~Duo;1Ll;6hDQQ#6N}YSFMYZ9Q
z+QbSpit>2uK12Q(wb<TrJd|QDJPsAEu1MKb8=!zKsocy`WTl@E1V)h-FBt4l!ssW1
z5wjE7Mo?VSzf=)c_QeSG%@~0{{abmV-=K^|*@Yt>F_6)hR@HB)Z}&g_Aq<?UPmX-P
z*Lk`)mEyz9zrl%>6&thI?*@?0x>{XGn|IbzdvnA6)e5{C?tFA^BN=Ceu^uj`$aO{U
zmV*Z5E<dgqrShtO*q!S&M>?%Ba#Eto@ASbCzd?&yFa9Z}LCT9qzw(;U&A&HOgs~@r
z%(n)WKXy-WW}g&BvO8$QNyO4zHl_wv{>3p#fJcAXoA1)+BL7LKJx*VYP{b6f0df<i
zgnm@=Ab=masVORuq>&`j_@ympMwQb)(a(VJ8jR|^qoy+n_aFSu(?swaLjhB~NbjJ4
z>>%1@wSN*T0GaY;oCvuEa%>@zk(?in$xqNW9pT1%=Vfqwz5{kcOAhXlMZ6oRH~XOF
z{0o#GU}S#nk9sCF$Kv-#gp$?-10No7DUMjXBiaxvUmfcBDaaQI%Fk<|q%W9}8xXXC
z5+C>3-vHd9R7AM~!sR-Of;-gxQefYt%}PlEUgE(h88^q#cNtsLkcbZ|g+h+a^cmN)
zU|;5Y0zqaPR30;#@0o@d6y%hi?e3NoKLiv8KENws+BN~S+hSVOySF8oh%LIB?e%p3
z;V0?t-G`E2s(ODilb%)Eu@anjEbmJC+HH*+ee!k9ek1>P0zbvg%kA*4^r8Dcy4APU
z(aSJf#_6;1WeCcz<69Wlp3AqukHqkG?W3Y#@{WEbfG?cMvlsa}Xxh*vS+F}8xPI?$
zZ>D>9HYE>-QQ|xs4o=SI(s^1+{nLxo8=a<}WW?TZEcsS?&EkT`Ol5W1D`l8a$H(ja
zy?sy6L#AfO3AQpkJnBl89@{vvCVhzy<QSTm6t=vw<cs1kA&@hG5ox`OL^yE1va*<#
z7bKT2N#0zN+^;ycOq;u4OYS4d;B0BP>Pdc`<4*eZ@4oV?j$i)eU!+ey{)jK%(#NEO
zNR(&A)pKvj_abpt5$Arb=$TNqJXa-&mrLtMe9UL{8NtNcD3%RlmDP&f$l1vM-hDyo
zSmn|m3@tlj-016K+dX#v;M}kEwRYOvT=&O&Z*M3-FR&c0@^Zdwtni}bOk9h8g5jqv
zJ*!@0qNCbUOwH+ehE7L?A+z%_pA|QL*D0QzDn^_S-S4F{)h!eG&?A^I*=V)({A=s^
zHs{rJeXl?}u}ojDN>5mtOKs)VKiXH9I7ts4K1wTUW6tP8r#Znug1#j7F!(IfrqVOe
zWe0tgS8+BRDU20iY~_c3x{98pGI6#P#+l27<uo{3NY7pzsjPEe(FR@g-J_%9)ax9(
zfByXEKl7@g7rT4jE|9urQUdx1D+#w$r|nkL{cK-p(W4Zl`l_T&)%iAcj*(C0?@KqP
zozXFurMFVoY+p#fq<+rD_u^u#>S7#Vyv2yr>yOimgX7d_ucdlZWl^7_kF&$aTB~Lw
zW3w*3T>3Gql9)Vrpmdxi%~{VPnT`(k({ZQcdS-<hx_)2v)$MfC$*6B**+EBbrndO#
z!QJ%P#~*8~s!qfc1{=o7f_TNSGm!sCxQCsdjcWI|+cxkun@#nlSg}RjbCy?Eec<u4
zXX)va=dSy|__Ht4!-sd%D*dolPxGqF3(-W6=8=IBim}2ZT#V%G_Q9CyOD`OqFo{*;
zTVtPd<`^riFk?H)&Goi)S?Qi4!dCz6NJE`|%^Bi{M`?3wGkyN~ryj!?^PWC`;YlWJ
zv&mMa8UlKj)cibQyTT*2CC~V+^=7K7lbxN7)!oA|#&(FT9%OsO-smhnesP?B|Kvbz
z)R&mHBn<UQPsEiztj@rMUpg#rb98(7uCe+w)2dbV9J*8;tFABk&JsgV*fFNU@C8-E
z_$zX7g#8%7PnYkB{SSt(5Aa6^>Vot-uFz564d}yf(u;Xdz1;vST)z&W`zqSsCH)U1
z`@Ux=&zV2O^X0ls|K?!B*T&kqsN_QE!Zqn;CBQd~3niWcPgD8s!0b@K#Azn&(=DgS
z2_nWKBLRl;5fk7A0q(SN%S-QeFy9#=Xa`((Pbn0#rtcteLxvG8P`Lc@!o{6~Ujij4
zG*n&-$PSIP1p8$-DMyAt${6wMFu)$&CP@N0>P~P2&yeXE<TlsNGXk%n+LMWJgJ)k6
zK=kxuaKMxjFfChz&L9d%HwjTRm{Iou#WkfTP#}K#jCprp(B=hyCE<NZ!e(P@Bi(!S
zFf}*UyaMbDD}Wd$2MjO>^H$g-cyTy1LMW@0r0W<5AsZ!-E$Ub?#7ZV`F;3Y~<aDQL
zDB+mx9&FSRPs{#*)3(vX+J_3sHVz1Q9SPgdo<EZyUQc&-b}U2^&h`mxbLjoCB|(!P
zlIB1{_iujntMvE(?Z2mSr_4MInATF1o&bzr9LQX4SJS2x<_!tkD-|h$Dg(wIgl&vC
zHm;$>DI6%}!8nAQS!;Y;Lh)l3YNVH9$5sx?jDw_^J&6)Q97>woS;Fjs@Hw!W565h+
z2y@YA=B(UsF`JP+{i$aJU6djo0c;n6!AZ<0ALXI+%v6jSf>f`R#!F`Up*V!W3k87L
zf_!X6&}MZV?Myz*{-g;oG{GN&Gl~i;(P&?2s+dw<F>J0{QD*Bz6g-4{1ab^$!y((b
z#OgI{Rv4$Oc+<S93zTZYN;zF`;tQo5V;;sW8wd#(KJDdap4yDML&3v~vJM}P7yX!}
z$r%XBPkr=cIQBmIz0OeXlXQ6sEsX>5w6r>();C*eP5qrMT{v5gcl)W*_&hM-&arGJ
zcO0?lIw3FIzB3?AQNK9D8y-L7!W+mq1ailyqU<yAhDeuP@snxUQ;?l7C&E1;0EanB
zX=wl87=PRXf|mIx;Z9T1mo)YTy&!6TgoiWAk?{=#W<2n5BA4VOg~Sw!2i%+{2VJAB
z{LR;-J_w831ZHX9%x9*~Oqa-lb1g2HS3opC7`{LccTT7)z$9?wpU7j@6=1*1kOMLV
z9#A4oNX1p^a897)Txeg*O|<b(v4BFLU;!lzDAZ;#rfH`>{fZgtC-8uv6Md$n1;Ev9
znphx@@UEDQQqobl+zE|d(qxz<70d*On><Tn38GrB6NoV&fbSd|X+X|*cydHUQI^u4
z1pCepZj#xdE4wEw$Wzi3xaPlsXTUKbAW$6NunRv<Jd<vka0Zl~H_vql<We8Au?`^s
z$Mj~}Pj_ETygvZkR+@ixz&j9#ey5Q!X`E6$a_hag%tV^CZOJJ+sny<0ckVt&_aA)Z
zt-ZW*OBbp5s3>7YeJJ;L{i0`tYf6*e@m&B}p@8#wC*WP5owq>L#dqWDWe7Rq{dB80
z=PP&#C~bTj1m)N9EsV(n;T7;>S@HEPaGmz`dLxPp84%;kTD#?yORVDRNT=9YYosqe
zx$m>6u2}IPIT9mEr#DQ!(Rn)PNgf<Y_MKmn47en@_A+q}7&~CcxZ}gJWMawMQi0ON
zX`JtcA?H-`Wv4q#M@K!c$U>gRs2~y5@@PdeI;)q~T8%JzNWNtS3|rax!@FJ}obSW7
zhHS~sc6rJ))uhX?VuLLN$AeBf9}Uv|yIX1d&Yg7JAN!*^2Al`??<c0g4M%5678Bvl
z&Q8QrFC8+m>y+)>BwL@H+3@C%W4h3@@Y4pYtg6+k{z%T`HUwVdiYE++E0W)tREAuA
zDmrYv&bDzq>ueakqytDUM?YI%Yr3p#vxZ@XS!2yc!^S_(Z5oc*x;!Z?3}fhgtaQUj
za$kBD20x4<hiV(Pp_GcE_6nT@e0I%XG92YOWS5Y`BOBt-Inc+vn~3TFy|LAj{-`t<
z<p#n**Wi3(8?q$AU!JNBdxz<x2lvw5M~{3K7iVvc)fZU(C)Bhd9gr>d^wd-WRtvG`
zMKLanoj==9HtO}Xz1dDRcvKy-WniDW5^i7Z^)J8vz0V!vyr^M+Ae|STPVZ;&Hm5O)
zvE-*e`NAuVo;`b-zIpOgeT&JWtKys0*H>w0XDhYp4ePfrc6XI;S9;N$^4(D1xa%zf
zIg9S_=*VeUdAF+m#}<<7Y6A>b=wPf6WF_MrrNda;5&ysX%~$ESUp~>3l0I8=o5CM$
zxxLX$pMSKYx~Rp9uVuDCl%Bb)KB4@5-n9AzHW+y8OQ%|vzP#3MtG=W^D&A+ctuA>5
zB3smRcGs{!Vm)1|cv5BZxmb(tXY#0>IZyxtB<IF)UR(5mSjow%G};P%5xtd1;f0=s
zY;#CI7^!YXqrP}hf85;k_N9l1$EmF`wcf1T*oiKBC8ZAycC6xd)lMgDe<~VBou10N
znC|av`YDBR5Iwk}j<Y-`{ruuYZNXW1aULeC3@a6lAttn{PImVX(*OL&U!^BcpL$Eo
zfB*0QU8*-*8Z*+<r9%$Xx4Y6w+0u~JlAGdbv$g824Oulh7>>Lx;+E>1Gr&l*e>_f4
z_Im00Vb^IfkI}yw7u0DO4(C*l;!9)oSkHiw#{Zn^l4s1Cw?kAPR(r7JKfM?O9`Bg=
z_7FSs06h1mQUW!=_2UD-h2Dbon-eiUzz6uF1odrtD%ZOjzGBS3DZK5o0lDhiQofFM
zDfR2%2a<jNgyE*#Jqxok_-#O18@@Kz*X52EOn?&R_RBZ7GqKPtSg5T^CIB%$)Yvye
z#*qEsmLS0Nvco7j-j{yp9`lhwP8Yv%BT)>9OkUw~%@jIjZv%u!bOD@J9>V3YX-btO
z)0vbBx59#=?z*|e;7t<-VZrqM<dpdyP(py0V*?2DOt|cx#iL$|zkrQ}xvl`a^y761
z_C#v`s^M87HLVXhuGE@Bosduv*-uaeM1@~ziF4;GbyA>UWkRkfZ~`yVO`}WF8P=~o
z7Dxq>qa}P_ipOfRnYKT=pEh>xq*|+)t|TmvB{T;L^Ms%v;b~qPf-*y>W(FC8C}%vO
zD4|4o`vB!{rO+Q@q7E_epp1A$k`$@2(xd#rsueB^xD_jA{=g%el&gDmoF4!FcYe@s
zZf>RZjZF)kbqVPh!A7Ek@K=>!ez3Qfe*3FmrQb;iJ$iBY(z9@;I!YC(aAe2IuZ7fD
ztEDX|ob7ea!l+s?L*eB>YGyll`wWbESaL)Rddw`N%#^`~FO?DDn1it~0x_!)qZjRp
zam&UtrDuynj8ImvFjR>?IQ&5u9Ng3|hBXuq1aI8bJ-8UG*lq*E5QZZP6|*W)vhc_3
z%S#NH3PWL~ztBD?y7)i`hgEqPj^K;gtF#w+VaP;*$6!f5UY(_iwGmW`8vVfrJ87+q
zTgr;kz-8kb^`vmx67cp9coc_0@m#ctw=%2zy)}y!@~8{XKy_o~7DHIi<BWF0P)6Oj
z-wBt(ok;0MSw5HIJsym#EccK4Y1rco1|@V_3|~?bHyf$F(M~IDYav?nnR97Ir~T7(
zf#Hik29o}%Ssb}2x>J|A#Y&(cb9xPko`%n0WZEvIX%s?%6S<Wz!Hn~IMmhcLpce#p
z*TW)V2`g^W=lcyBgn<>fz)V_#O5=tnoRc`*`mvjzS-6>Ql0Z(mOIm|?AX=BB6znmT
zc`l_4PH{(AKyl?zl`mE{5fm{&o1XwbhZRR(pVARTCR30<-iZn%Jfj9$1Wc2JYs#rY
zU64tXA+j!>9PJ;61Sr7rh0<qC0ThF0b>rYGr^_6MXYv~<!Uw&H%VE<r^n6jIK7=$W
zfqo`*91cVrRHhIi<$Y?*Npv{)C8g>5FVo3%y}~KimGjNa<{OBnU8-{e4u}x&>?#f}
zxn0+KSBC+P=Lkr-)%_zNsyIp@w5JhMP_X+?2C#VS_{l6!PHniB6pzz}*W@!4Jc4G^
zm}O`xQE7vc%yoDzp*)4>8JN_y;VM0O{(S0v|8Sj|f8}t!btD7LOI}!7No$gOws&??
zyS<Ti?mb9%?|+o4HOXn}ReZp-F>0nJ{cO~`I@%kfS7}mOI}pS5LpF7R=&P@Sc&5KC
zU;mm7U#~0kJJ!dI<k#CTB%@aYxZgehSM`mM6P26#HD@HXba9>zawEno&d6dV#YU?x
zIg$_cj3D$4&YBu}o3+O;4$`lG_jRf@n(5E~&A&<O>uYIzc9FV+aq4sj-d=AY`I^;{
z?PeowtTiR)BG;cMR&5MMXX!|?Bp>KmUBn6y3=iH~PV%|88C1TT8*6EMd&35k{r#hK
z)EkNug|D?zySglXdg;a9Vd`{xi7nwO(tVnZT52ec_=oqm)BXE*B)g8%<0sEt#=3Yu
zIloE=M_uuP@rzYORqIV>3g-hpALUu01x<7iR_Cy#9P<7-1`f3YwTm(4<YeTn<k|X=
z^0E4fE#5JHF~N~la2P1OmAUe#oIGaW6~hZ#7GmJ2tSot(K2`wnah`1fq2sMp^~Ym~
zE#sJ2H;~Q<&#bUv%Sd>`Nc8yme(Fg+N7tnOF!GQu`D2`6g)8Thp)0Wx3j-g9FY74!
z1DChhcr-{SgMqiWq<^$qRXUWnIBZmw)8G8{U#7*Cs#iPp2dXEPoym)>S|v4Vi)pP{
zbDgb7Ph%Avb$oJqp}IOv$HzxDG_hS{rB+MFs^_7NsEcWEa-sU~rrj4WT&F9Z1i0$$
z)|nLOE#Ai?wZoYWI;`&DY%z2gw$H=}13u>w4~JdrSM6rYI+n{T9nN~3+JRkD&_~%3
zggQa@;rYO7BXrh9)dy#N(f&#xz3$w`I@$%@^u@tZdis3VtDjJ(Fh(+2k+Z@6?9)f-
z=bt}PdtG?N6X&uWA0DLpJ6q}D-EHgT`-eyAOwXeErB&C>_$;=)KiA-7@?hvu>=uHN
z^1^W^>Ux*rx4LkVmKV;w^<Pc-EiA+t*ci4jfMLv`FOm-vW;sKSReek%#RzLXTxGye
zdT_|fwUKx~wXt=ry)GIHX;)+5@zZD4kGG`T)i9vx;(d6iF}{1yOGEfp+l*jL`Es@v
z=XF)pu8b{K`c*XUAF}eSUD24d^^~rZzWnN|^!&*a;cuk&`d0e-+4J;t_aHrd^ilfc
zlaEcK)@rBz=sZ1nair(XH>&fF^;-B?uUFEY^+wuWui4Ouyxvvc!qC`mG`uzAQCH6_
z-7(hh9(2>eanIw4tuvVf%US6d-!Y`NG@jVTa`&*8x~dOW<*{oE-v-#y(c71bah{L7
zW8w@rR)zBRx3H_wA_In0@M8qOh2Dbo4~DM~@S_0r?b1_UCqKUqNMBO&ZhNzyPs+E+
zS5Firdwt5+(T2-!<20UQAK?3eoBfNv@;U%N3||`?mM{xCWTr+JI$SigKU#o=myS3?
zKR#0fqzIV0f_8|pz!YOjAHaorj8nf;v_c@j?%4n!kl`s}ZpIUSr7o8T5HC{E2#nyt
zAGuwLA<)K;V1`V|He$$ShEF>brdt^7O4*{E0lNDoF35E8;h)xzP(VWp51(-n9N^zr
zk>DkWHVTJ}ll=vC;?9oiaWC@`;M#FqVCmPMQh)L5Fl9JRAJnKu>d8qA#D~wW>~)pN
zT^<YOU4A!%`UBzTssmPe-TmaF)ZAQ8t5U$Ir_+lw(<p`>R^iylA;EjtA4+JH;3eX*
zodvX5^>rcH{37xo7D3fH>sM)E>s;wkI1t{v_qmus=|W&+^$vM+pfr16AMNd@uYdbn
z&lbec)t2B_Yqo8i^7a-gLrV%O`G5V}->3iifBYZm;LGRE@jHW9Q+8>&vZ#7!rA-O>
zjh2+R+*cUXC>Iz__y}deUf^v#U|;-O(Prh9+@6i8wCs`3sxb>z6da?H{22F8?0A?!
zw``q=v5>r=OM3GqKU;cYbRvNIN7=^B;nfJZs%|Mps%N420Tuu9NBwiKw3jI>kTS)L
zFa5b(7oMp%kP;6*`<i3;LJ`8?#WRb($ILnl(4s;6QZpE-*#Di6ZuAKh6Sh==URiZU
zo|J_>^GR%Dn9Aw)6;BxU;F;2)@Sx<-{?LQBC^ZG*SL21gqIk*$-9;2F6m#l{I?{DI
z=Ipss8@^7Yn6e$l@nJU&JNdjaXo}8CP0G?%E4A0!YFFBdwRfU@dX-Kve4U7YjMwC=
z{5{%Wi1auW&B!9iBU%4qEFuY|{&TPaL^Tp@W(-4=p5J&L@Ed7BluJb7<Gcgo=5Y9O
zkeoMD26DuUw)v9N3>8AS9UI;poDnCqi6*MhoP}tW5+3QMa6R_r_z2;*VDU@X%XEPp
zLb)CGGVSO~3f2G3x-cl-RjD86L6p&iMZ*|X0lB`XH7bZKCU@#*QfEOl$LB!1!x`?J
zz!eKuaW}f-_>T{_WlY$>1K6M4z6qa%yweK^*J+HI^O}G{g`abr{02}Dq>0pFpprAZ
z*Fj|rJVgZ=<3KIS5pmPRM9OVk{bbq`8vYIJo8TlkNKTp@5Jo}xxDF=o=j1Fzo->Jc
z%Q%iIbi|SU@CU93_k{Ie280|4$Q`)AY<>DM#IM~`nC4|N4w>F(A@Cx$#7)D3(rmh5
zNk5Z*q)X)vnF(C?cDbi1!&T@_;qkMlv)=a&F!KQ>p!baAsl}yAT3g>qckX_iTI-u>
zYkMc{+`sQt)qWQtu<)+Zue+r#SEPTY%||PHf(`NRB-82+-w)Z;1sG#fK5hc#n=Qke
zK|DWa%lZAmoAWhW&M4RI5R_lXw=gEpnl}Z}bkeu16D50Hd(jVH52XF>dA}+zLiy<3
zs5jH*)<&u|RxQ`E9rv7M$93r~^%{nP^E5h(NnF*%t2F3z(wAR7Nh8UEUFj|V=Rf>I
z;+(Ai$N%&Ho__YTpCq3zG?eUmb|KkX^dytAeK+zR+sk>i$3n>eLP@9RS6;z2!kFPz
zR;&_>Ek!X@Va!4}ymxmaHKd!c^&V$BVf?DEE~by~ZCP%8{CqDxdw!4x;}aXAt|T+B
zt}LcU5ALQ%_czmetLoJoL($w-`mxk8w5h@H)aP3H3@p;Dat2l1MwU2Z>&z=emYGb*
zijd{lsR9`u*_^zXc-ZUrrN{McV4%t^b8yMcgfa|VOlo9h9l94rD)%E@tZ*Z5R?^UK
z2}aLg)mN+4NIQ2}nY3-1FLw7+w>yx&P_?1vxO=R23&YBT?e%nZPI*QyA1iriQ)r=k
zpj$wLZ5&x06chi(p4divocmm~+&PpEUZ~?%m7Opyb+N78iO*o#*jN`|JMIU22M1#H
zG_}=V>dTALW28e|o_hi&lM68Z&B;%Fbvkjb6~?rqqazzlh8K(JtNm^|><*RVymX}t
zuZTkz*}1bRy@s~vSyy8P6RXG2`7l15pRwgSyOD@#i3sQ_v=1wYgd$w}6%z;1b(jcB
zyGYL#ZFV4`fsj?Hb&O5u+~^{V8|s4X<IBz;XVT$4Ay71g=Kh0$mVBbkRM*0nDrIBj
zj`aVZefF_eTG8&q-m&zke){au-So>}{F&R76?F%_kyoj);tozQsW3K1XIJW1!!XJy
zH=nnsvK>jE8uWW<tF@Yb@%e+a-Cp(i(IfSJzAvadX-tjNAx1UT*~Z$s^s==yr#{E3
zOc=aW;Wz_~{=cvBFc@-XHs^?Q208tJtvf|m;~(SK-r=$PJS!4cRCm-j;~%3Z{eisc
zQ%%mnl<vnwOmurDbIz%MuGOpQ;Z8gK^5-9?pMG>l{c29*_A>P~#{S{&{?6YpHn#7j
zt-B9WTB(RX>9k^odg-W-9dsGD-SqUuk+*R~N3BTDtgSAk`pQCDYc8iRKE9jo-&qr_
zzS5p~b=!&5`~9PS8py#o?(|O6H&6G|^Sz!Y*s>b5Rt;k|XSg*~KI-^beUY=<Fzy|z
zea`d@<a+~cO&fFk!KHL``uyfv7};43NjuRmfEddRrXL^pE%X+we=vOgYlL^{<A1>Z
zlG%QhjPyFrc#jhAwAuY;K%Q??Z+fD9YqHnTj;GS=;ij(r0p0~}_AmO*>i~QVUki<T
z{cp?cC4{#S;<$o{jbMs&obeEH3Qh=nzGs|pXTP5CC;-MGtl;w}|5RYqrN8*t9XR^Q
z@p|JqJmLutdUED){JQ1v*|hO$G%wCm;kxtcA-!TGP(T-4!k|ZX@PI#Plna{05I}t4
z*B5k1Lpbr7KODv<Z(ZbR9`jZ5(9Z;qyH}XuhFd;}iH=ulWnBD@%=c-y8n?s|&7Xq_
zu<x2n+{qJVi~NIMm0~7;$`s|IoKB@U`}5^6+|&iQjwg(>wlXcn!*8jhQqS^a9=J_$
zQCh$u&Ws3FYSq+WpKb|OtoFj#LP88u;IIM+V*|n=+a1_YW2K0F>LH>uC?XO%O;$gA
zprizT&g#eIu#k(WMIgdBE5R-#Ajj-J_P1X~$x*tI1nkbie(LNWq%#Ss-nLH4#BkV`
z@^C7_FAO)V$YN#K)5qVW{U<LZsK-jR?*zI0t`3I8vg%NG3zbp|n0W@z2+(X*0I#g_
z;*e-oo5dMi6qo~{7hNY&W>!1NCBn>>q`Vlsm_<mtQCAqaphunZ;9!dpXq`%#ypU2x
z8=0PH`=dOMf7dI9p*Y72rN-^8Shgc%+d*$3s6X~xpDI7nN=!Za1JtNP@<wYSe=aLY
z^84&G@$K^fq<j#ELc$$I#R`oo3PTq8a-le)&|)xSb|;Fd`>pEM?Id2P7yUaOK)e+n
zI8(X!SV5l{seZwS(cp`;D1Qs8LkarSiQ-IG1R2UAVMUHMxsY;-62f8GtS~%by9g;q
zRHMfjbSlf<VunM*F*wB7Q{$r`N`ZsV$rz=MUH1xuaXH++igUQqa;I_?6SOG1aTHDz
z72#NJqzhgq_)_{K4=#lf<9tbF0#1)#=;0<@>G2025e};Tk(>PB$!EM0hMUl!0lwoi
zE^!gg9RL%Gfgd5pu&exqKjI>Mnj(0TkG$-fO)DH{#s^*EfmsY<l!~-bkf2v+DJ=fL
zQ8aZ&+7fQuoSvA{2K<0WJh)uh?ROaUGzlxToiNAeI3ny0S}w~3A0iZ><T=ed=+ZGs
z0~(0@5{`?Xpf|yvxlGe2^4p#HA})uSZpP;iKZ(MpKc^}EHHRA~@|DB*$|h**&omT5
zm_y`n`iLdcG=dE8_`q{GcW^W9k|O+2jWGU9ggkYRIt{n+iXS`xzb@j9L$KZ4m4hA>
zmH~8mQjWtx@sBR$oP%9nc}kqqxQyV0-+#O!34?r&GfmTgXX0k-qogN<`pa}276~H{
z-6!cXR!&#)DQ)9#;#_uwbCSp84ug0p7pPDSJRBtDo!#yA|C54$DEQla|C0sk`1P=u
z=9j8^AGOlj)=v85^Pi`m{QQ^cr@#2O=`a5BZ_;Of_Vcu^@P#F|^rhYT$gZ~Z`y%&`
z(og&scYfVJ-1r3|JQT^<@b?Y8puZagyWqdg^*-U<>dfQp6)>e-TsJSE+}{+G?%R@1
z9w4u}fZM{$^(`<#n{A`nbb)|O2XJ?N>CT7sIb==6uU6}+!M17=<XCOuog$<&ux%XQ
zwyS*PUO~qQJp;qR+~TUw$+^&2VJpIo^)=}hy|nlIdFpjL>FD4{_b$StKmN1QixoM`
zOe&Ks%ZeG~QVapm<XkFbYm5+VF~_Pi;*den6KoLT5Uxwf#Te0$(-FUsl}|Z~YazAU
zP07%b^`*P8s*`Qo;UC%gQgSP@FRQ{j-M(biWq%N71(xO3<uFQdel9ChPDjIZaWYPv
zEWEV96GC;Ta$l)VSgGV45BdYU&qW{UiY?b!&BNq5%iq$!2wSa4|5AOTtANi0M^;8*
z<YT9aZcp`fbm()dd<L58ZZI6EXjfj5cK6;L8}8U@4V@N)58KKe9S_njE8P51zMlqz
zv5g}b-bl-{?ecP7^&_1~a<|`6Dgy>C&gdPg46L||bB{4Lp4o7v6sfwrn6}pH>F)Mg
zy0_Izch)4+&!4Hx9ktc6cL`}Uo8Ar<1KL>WSC?1Q-OY7f?bNQRex>)dTP@csJrUjx
z4v*7ur<=~i12PXrOx~F<RBjAj;|uk<#pTo_-74)<OPgC;qES`7?0faspvN|RgLFDZ
zSCmd9y^XV+1ESBPv(Vn5BcTh1PG@~i_0D<D+&MFDsI=%a=w=~zq8noT!qBC(a!=mp
zWPvivJ3{Rl-;B6oG9udv(uYGQLm%TTHRxQWt)2VQi00L{W06puX{;@+R#R1V$oG=|
zn8})KH@f2U##n_!Kca3iOr3fKTQ#-X>xpx{*!psBcUSe%N$agfYBkxuP_o8Qy52e4
z56-9O2i^4e`GIsl@v^vLqZa+JI~=Lc45gbddtxMYL)kDobB^Uog)Ik{%rAyoaQIe(
z5ta65wHRfi?byP!s%Hp1%`2SMglw0}ia7L043!v$8OP|j7#rCxka3UB>3*PR$b)-3
zstffM`K$6%58yJ;S(V7vs+^_9>M>TavHc?3?_)e(Yv~!bsyfxU#c0{!8B!NNJmXX_
zmr>(!evZkVbFtl{$_`IHB0_x#<8T~<a3&tuK?Xe?!|1&FVeEjyc8KWU>@;(J$vZ#v
zD(~Q;{=<<Ao`|b*BE)*f1Zr0Q^bz+TrzkLWy&AyzE_gt}eUmib0yjx!viWu^yzagH
z0sfRBdia~c&A#%(_8nv)1KHk^_3OBypF3H;|8T&2>f0{(%DmljMY&(ab%}dl*9V{t
zz777%fBs+p)7;&y9lbGT2G&fvYfUK6l&BiHG_e~zLJ2n|`Ge#P1>N`{OiIAfIi8U4
z7wnu)MELmP4EVY`fymwb2o)bQK@g-#8My?0ag|(5Ih+o~fsioeEXRwQXO!dDi+0kX
za~R+MEU4K(A!sW8Hb@EuhhoG%(i+F{l07mblV2l8oVm&6%Ri>a55B@%fFlZ8=3f+?
zJeSLf94co<%*+KkcOy8wY=;O+0n<|;^%03?ppYagm$j6gb}{{2KA^1ph=7b}&ZhT_
z>O1-vXM8ocHqy=~4^ve_;Bs}<=U&mi2-BoP*krqf6NK>5Sd|>YGdW>shz~Gf_`+C%
zpvVfLVzA*KqY8ZsOq3V)&PSm_=w`(fvlCHdm*(OS+rt;TY46$7)H^y#lGH4aTM1cN
zO!f6vTHoAC^=3PiKnZ!fPadb;ub-x~A^X|C0$gbV*b+$I3*A81YJD}WZ>^<eDZdDl
zOU!f>KPX_#;EOH%q<pDN=m}ObP>3<uiHC-G;U4uXLD>Q&Gepltvm(W?X2lhQn$zkJ
zlNd@2co^AeA%tO%%NPu<OX`%mV^$}#QBQrg8fRjKG9KFoUN}9*Ay#csMieR(Q1*Vu
z&$b&VjR@IRATC*{g%J&5I?`685Ywf=!#F4MQD5RIKH>&z@ti<Wjah>T<_P6hPF3aw
zw%wp|;Y|twWa5LJ(km_X!&V+dFxwU5ApHk}-nsbV3<B2?^pzGxgtNs^urRQByelo~
zz+?8|!qt4N_!@~1DQDj5u-i{vDc<A$i9afk--48cTC<YYHk#2c;?@0DG_KB6_JNe2
zQwq%oXC+bjaOEMGAsm$5L_w6hd{B09#<Ab^B|kXEk2p|B!85?<C5l9jaq3Lz^$?g-
z+J&UaW#)}DMKL%&-){rnF^-K2cY2TblE?{je0DmIh|s;b0&qu2#s>;5I~g0nYlP`X
zw9SVIPu?Lpy10P|n+>@M3NIBi15rl1bQQ`kaXI6^8AHfySmytRE8w~xUx}Ia4|{S%
z!5}<`Q6Jpw)lW`xEw9|r)Azy*%y^@%qJ9HYkqkORAZAkG)>WFi(3ynE5$@bx4x9FX
zGG1~i19x(F&bk-kKwy^qrX=N%LQh`^P@gevpxl9ySL70$5I@be05La;9a#E_dcF=0
z(~o5t5F<dBdN0T`>0BSU@J#Y9Fq;8C9$hZ;q(}l%0kbIqaLl=xpo~+95|^jsY-~`O
z)<?iMPoK;tf8TJkC_gHo?xvNcS}c7&!}Xe;1-%<3^KiatwXvD*-uozRZtkSk#zxxS
zxhJ`1P4ZQ||7aVxT^Rs^X}(a#RdOh9(gn=M+=LSMenI@tcH(aZaNm}**}C|?0ou3K
z;;Z3WUzlmXn=qmL-dKNKovN=`PJgqqeOK^vT6ss0)HTT~^2??gyZ8ui9I#b)pdyQV
zTRT1LG5+9Y1r<i6yIbq2-K?e+$!^ukQrcRtrfum4E0PsiDZ)yw-biw+(y%ojtAQ@0
zbNv1P_`CF9|LcFVVTElEnQVpeiml2%{`3=#zlOIJJUZ^BzH|}Jk>YGBwg5x`=A(Uo
zc$!8hYyr+pspOSG$HyJf!H|Vqex9B^-%I;PN-w^v^_q<qNVAU~?4<Q}Jslh=P4~pI
z?bRjQ>Gf<BVmrV_qpI>>`lENdS@TLNy3na)><i_Oab|z_AdQE8pYzkGttQUz>J3iP
zll@*gRy^<hbCn%kd~t!T@T9M>x=1=6mx8@5BHO+pQ)4V+Vjt(?c2t%VKAx*Rh=Uhe
z3Bwz*`4Zc(B9lwEs4CyB&Gq!~?q<5P(J)=kE_$)oO;4Wgr^DlJ8V$!@p~QQN$%<zu
zC+XyL=v7&qbB0msQt~@`4F&~_GJKe)JDg$Ul5D;#{R-nH_0gofl?9cXZ6YhuGtScS
z;hxWug68-_I!2`-nsJ73O=+28hY`m*(D^7emgmHG!-hwU1VZE0t#<FH-NUZSu+eT<
zN5CM+`Bul>uIl(K4Nj!<s76>Ngki6*c2iZSUcZ|TcAwiw5e6^mqpCmnuZTv`i81Oh
z?FGG(T{NgmbQ^=}5B<65Qq()T6K?_)$j=gX&WQIz$W-W+1;8gi>XG{N=o3zPpD+oD
zzJR_*-$#GM07p7jKW%N_NuPc4sr!59_{fIGE$PO;{P}0;?shv3q)TB$i?PSHl4?Ko
zi^J|XJ$`<i_SEO;Q=CcFRKG>n91nV_vUrs~{pfDmQ2TK%-Pgk1KT>`6*g@vh#@M#%
zyQZ;R5#5!=IUA8U4{v*WOKs5-k1<h{Z3;W$L#kdnm0m|VF$S^vj@4tVxZ~_?2F<1V
z28LbqV4f$eOk-6P<yuh~+l+!U9F6>(ITJ6`5oet>DvRmPdQ&`0U#u>v-z}!ijhde~
z9A|)WYb;*a>4onlXO|1<cyN)P>lwwWucO{cdb-=y80f2w7t<&A*3-xL)~xHYwdKZI
zH9gX}!JtOJpoQjDZm-M~UHa?kIrT6~Uq9JPhw5jG>PP7NolaNnxRN&Y%xl;6%u(7)
zm9MLMdAfV#GuRdsZ$sG$D<z}uX#1MVvZiNDOJj$z#!A6CjkBvuCi5~*>1VK%pU}Eb
zfe-=l6M%Ck&{g_kky8kb^p;&TCIdHvoDn~p$M*xyMPz4ky9SJr5Ad%I-nFlhmVPta
zcYgSO6z`HLyp99j8Slp7+t%65db$l>Py0qOZ@~Mk_mc73l>1*7{D1!6|F5~ZyLawN
zgi~dyvrUwr(mr@`=R3lWSJ?34DM-Er(zP(*M6w9;%c;aWM!)h}*HC!G2TU-D=P%xB
zB8i*3!Y9ZemDE9(kKO)wI~xifu4}Z25Xj-zqx?=0|4vs>3b=|KAG`$J2rCdPBC_W;
zBETdZ2axN);iM;$Yf2!pin5j%|LT9Xj=1~b5y=FZ_B22!3X#_znE}02!v#ILR3T0!
z;fe_2O4$P<_v_`)X}Qnj6=H9|ghWBhVI}<1b&4l=rY<3gc(NcRttFxI&c_eassvI#
z$gmv(@h7Je1SJ@w>>ymTq6md$K}rvdBa|~Ut(^S@E`})Dof(5DTwckfjE1AZq^u}c
zCm6mYXhH+Q80ClAn;0j0ole?+`ZVo6f1b{zs8Uk4vBLN@FF|!tN*mjIv^Tb-v{Xb(
zO4iqp)A5V_bSZ)RB@ja>0tosD_HGC1R0zWuLFOf7v&~<<-Bh{5*o0Ar+Co{yn1W%4
zJ^KeLe<*>%VFp>PmR73uw6?zCL!&7VGr-si5QXbR^~1I!;L&_mK-@Sg4;M;@{A{m*
zGDEz#`a=VQb@PHTG2+n9IFOoc1u>*@mYR)hLT9xYIN*{FLlknIx9*Sw1J}i=+Dr97
zUaa~Ws($GlUXiA7_RBX9QoNvtLcy~P#mjw8^jM);lv1`hBp=wtD8|cpjA71oL@_U1
zcN7vH!uTbcy1TFFPMf1#vn>hzi800MxLjeT>9WgD8v-cMd?+CgX8Of8AW|5mh>VzZ
z$c$1cHiNE|(#|lQDC{zKjX5b6^%mPnHLSF+XbhmZc@>**&W5UgDYh2Sb<q##)8qvY
z<#~bu6BWUHW-yz2SVp?*KpcEe(od2yDkC3-*zW<z{|s%1%X=GbfI0#lTA<SQP@Mio
zI()O~gO9?)b#O`{1nrwV2b47BF$2P+M8+q5gggIC#BT6%+;!ScO#}&49lByjI4J6e
zc-*@1;><zNDQ*LifWv}ziMFeFOl0N`%(?*=GzbQ2L{8cz3Pgy*a5|iF$QAWL;A|Z^
zTJcdx@QLO3CQL(dX=v%@ZTfOZ+=G`A8uv(Ln5n->{&ExKKJD^3#5cc}p<{Q%#E*cu
z5}xk>Zjy&LB7+am@$(<g7d}vtZc;ZH3O?iSdVK^=)V+R5=c?0PgfhDUvq>gyIgF#g
zjB9>pc$_4dO=nPm(-TpmLCA>y&;C*&j^byw6K5z+P~$d%h$<X~&t(=Ze^ky1JU+u|
z3~*eYmmtEe@DpG7i<tEF<HzCoVc=E{e=JbO81)z9Hd^5JdY9a?T$L<P_v#3?F5KGL
zNzLX)+Ss_8TJ1GY-ryzb-39V^C{R{+0VViV4OOy#^U5tLUPbwBfckh7PQZP0&fh;!
zCvR3Sv*o`29={F~%I~Y#ei*)rxcY8+Bz@1|9Oob&d0Q{abyB!-z3?ov4^u*XnB^n(
zIqR6_mcwu&xsHik9SjfsAt_TsdH`qEV3b&1UPyN~TRx-c;hk2JB$tl5qjcEor!&cE
zOsuO*#;-{>XDhw0zx*m4?(MnWzx?X=X?J%weewBc>8C&cndSPS^a!>X+}qo=%sIcv
z`LoNZsd3tDuZ!iyw13p`&JSq1ozB3kp;*!J@ZKHCUYF_XZ=R;d&-VRsoWabl5$Hkd
z(lfsJ?32`Jtx3jC;%PBeR~LM4(brG+gmdC;O<938FIlptviSqG<ko(#BiS8;&9>zE
zi_|^praKq~)}(_p#kX{oU;XZh&vD|3amvT|3-pk=G*Vu?CD6&%*4k-Z7w6mZZ1D;s
z={lov7zBocv3K_<I?9~#lxmQw7@w+fVW?n|<O-{PL}wWsPBT+|(_T-W?WLzG|FP=E
zAKm?&#He!`AL5l4x(<NOg05k=WO-JxU^qb+wLT>oz20b~_4RhzSa14lBerX+t>}4q
zHc~wt_?)qZ!XG|-lr9&RQ>Q;x`PKiVUtCC+U6sy)(XQX?r58^gr|pea`t*}W(#<eP
zNtPEbZPn|?8B44ZTvHp-4y*{G>vE<SXDK~?^1|nnb-KOO?Qyo!g;xl2w$-5DRsHou
z??io`iGi2aKWi1v9OJn@)t{J*=;HwthAw1%NW5U2!?@vHIq0Q+Kv1=z^B~-p51$@F
zIPK*}dC>TiIu))E&<Ty~UZA#8gCYi?+sCdNK_8%<Iaj;YY`JX1VbAsb$-{d-Yl^mL
zYwWQVV!OGT?%lha=A<(p91qhsYV#L+M>alL&y#LX|De7ZqpQ-v*=q1e^|pU_=zhoQ
zEO>Mu9BF)GEEFH;)2&u3ee&p``g}|J+>vy-p7LyHOfIE;<+pbj=k5*0XWnj<=LIXm
zd_J7=r){h1Lm1)cOBh(Y!;`dsa9{&tTjB6WzhlJ}df=+YTtzA<hAz(KTGDvN$b5d%
zPe1+ae){;)j_MB`oAb;tD9!6RBpq^TB@KpW;%lI1&nO-APSf-KzWn{vAC7HE`}E<q
z>R~N$Hr_z}1mpJR)_Q8H@2)OM_tiM}#BsIXNY4TrYuQ4vf0~{ib`>7qTXuI3RQ9gh
zsx2NXl@*B@8i7&>*!qy|AemIm*<#e`?*6e2e&>8g5q%8w7>+Tt^K4vQk}j=h@THVs
zCZi&YV*tb`rjDWh%v5QC`<Xxu6cW&U^ugTs3IeVPuHQj+_&pHP>J-WZiT45IG|B!3
zd@y|dYlL^{YbD)mzi|%>=MM<i<Z3xz#RXWm$)9h=F;r&q`_Awxu4du&w5JnVZv);E
z?+-#^d7E;-ZK@w9VE9^SH5z~GkGyX4cqv%0(PahJcl^W$o;#-R+#(KK`vcMGgNBjh
zo}uOZ3Xb1M0tBfXP6y#%aV()GUEzm#UkF}7IOjKwBMd6I{b9**kda$&NB_wUO(MuQ
zBlzaL^cVi=y|~F2Crmou%(vrw6<(qY5zb9vb`Vcm{7@@2u7{g1U6F^9fd~JJk1ZDi
z*Y)P~;f{KY^i#f|oq5RDY?xg6FUs!-z4T4Pq56e3Zl{&MT>OErd+<(3&?b+-&3TnL
zunTU%$zi3uVBsdh%j6{#Ri%hE+AX_DD+ISO2gkB39zzKqb`Z2h)vIHmj6jc&XrqIY
zAz&_fr4GRMeXCML8s3K<A)f65JiEwC`8mjzEdZokVJu=>0<{my7W-$TIE)7a=P4Oj
z4|ypIlFU|X9N5h2oY)r~H-@k7p#)+nYhuc|8(@SoOTb@J**FNgDh0CMuBVnR4#}=H
zYc`@)p(RBNTn-Ax@Pbf0kYdx50MF_)3}2_LCX#YQ-5`82GaF@!m0#2oTUx*qw5WH^
zkBXUmJVGVz(H5#-&(hQ78F{Klwkhx|J%w4(6mDrJujW!Xht)C*4>}l=s2l3GR;k%I
z=4}<Fz>p4l;FKAK^bb*Ul>iu|FjP`sD6p#t{G#t!x5^8J#qBDfo%pWjzNJ4_`(Y>p
z&$9(Z!?Qpobd%1rCoLpXhR74e5qykv`2ErrJ=cNa;Qd_vfYzgaY@AcyGS`w?y&VWN
zoHn*t^p+>0<}Dj^_hHg!?4!>%Gm4kOz=}j0vd-zhPVVuJl451uZLWNvA>Oefu+<^o
zYhv`krd+DiC=26=F*FlLI<UFP?Xl*%%6H}GwE78J;Bt>KDOb>QT+j!Hx(s}VWPa}Q
z=dielocas?G>&UFUdhd0<X2Fxr|}(z3tzYhcOj-1ao5u1`?M_i;%~O>`O?pXMu{iv
zT6n$`IO7K;<4tHALvHH842wq8N5SLH-{cvN&%xR6JE4(|;t~+$i7<l9n{cd6;Wi#Q
z=+9jNT%sb*v?p=Knz#Y;ue%;`ktgvbuHfs>e3&<4f*<-D3M#orMBx`4ytBN-MdZ|-
z<D6gyHzKls>Lz0Ljx*XyX-oMW5%=IlZmJ8nQbc+4Uy0XkQuY$>@VJPi&L?3RCkXQt
zC~+}POWT_za7((m*zU)8Wqb)qJ_5tfpVAOk-qn#GwDp_Q**%L-p8E4Y;krySmt8*7
z1HpHNkq2l%sjJ!k=}6~|GxEb9VG-^$x+eK?&vDMz=sZ7ic|kLVt_f|f`E=aveAw#n
zI|B8_Z%7>By9VlOOp2+sHd1@@UV8M&&(cr+{FmwTpZzj@_LE<vM<0KYHt*a^%ayv?
z!9CS~G;g$FKxxkc#YMiiT~o~O4xvPIk)MC<{s0B2kGHRr+4j5%??(AMz-x&sGMfE8
zfCx{%IPz@t6e|0&B&YvNKktzDkmHb>>43<PgVW3Obibco9QIO&586jvZ`~(h%|<9b
zjh_hbxHnD*$9(uc7adkUq;!0AoO->Eix=CGElYl!w-E>f^PTO@^!X<bY^2)W+VIC=
zc0gbS$f;!gUXQc7FlLOTSM<}fXV23={nNjsufBQgkMV33M#*|Mf;>y#Jbohnj+EVH
zT9Zz}DlBAiK8zn7AE$%EL!Warl1|gNPSJPVx#aWtt5X}dI1kF(mx(|09pq}x{;4cs
zG*P{X-k^7!wl-R6W4-MaMCeOQzq&X-NlSC*skXXogVCM!M*8f*R{CgXE#2L0`Qvu1
zG7TNahMe3MtXiAs3esONtgS*v`c{n<R;@<b-dxkYmTJ|Nbbn_nJ-WZ0w%4UcwJNEh
zFt$y^pop%5PGOxv<S`neOJPu8TXA$4#=@d>K=`jzD{6Q2E$K1J3*$ty+4L$oCZut(
zon(7WW!`8zfBFn(M6+F7=WstA@9(;9+HH7Tw?V{m|9I##nGSYeq{HWr)8fTAEnJ?a
zk?QbdII<zCF8$QsL)6yjp==9D-@B6T!4`}sDn_ijx$qfEPoC}DfN*f!@mT<uKxn^H
ztkOg0pq<#tuXlXlZ5C<EOIFaRm%8uk-O9G{7@K0Zi`e$w?_*){2OBoB`dtv<cwh49
z8O;rM$RG;RA7MbOm{fd8&hR2E5{G;#`XJDEZR`_Gqh}(Z!Yg_m+gsAt=`-k>d;5p!
z+3ukYzMP}QmY2_-t8O*c!Z>qbgY{54JqDW|TU$zZ-rYY+PoKR=Uw`u?ef8B7r`gw7
zVP_Z4Qft*KX}#T0zl{}@qw!gKanRAFaWYUnDsR43z#IDbh3d9lTTJzpxl|Qz^i}Ka
zYIjUGp)ZHL!-OhteW+@Hj_X!s=DP#s8>qd<L+SN4xaqEbgFy?rjoMPGS2&Mr)%~CE
z0vM%vMldEA<1zVhF`d(AT!UBAgH_MewerHChhELfKy-Kb<at-scyHIGb4!oL2-jsr
zq?lafS?IQ)y+W_%SPq_^$O!0w%NjfF+KP>H)R)g4SNYgFb#QWKqhc6M&rA=aI9q`-
zzA$1EkI{Ur=PSlS>bR%zz&4HUP2yRd-($e5(v^<;lAgKnZ&_D7m>+W*GLSz|(|`Ek
zH@v>04@Q{tke~iIMR<Skx-4GzUj6_d;MFkOZ(i5K-XBcg4UD)M_l-+G#%GMeX8086
z`}+oRdAFR*_bcK3mBO664$J+`uGjPZrv?8%wmN)p?_RtGbDPXK4c0LQs++j<=ODca
z%FODBn^s<VVGuG2!{_w&WFQpWpi{tMo_)tFG-v_w3wETN!fg7I=ahDFWy-pyKx}*v
zn$n;wI9-mLNs3>Q2wm4X$uuOud8SCF!4Kj90*2{Dkitq@kRokxg3wb!CQ<(3R?<!D
zxA<pJxyv!d4VdsW0eQd+IdoFqd=q4jM9i~kAUXrr=s8hIIla?QtVlYk!U{e(8O>2h
zwK*<k)exFXsmfcDP}91zl{TaRu1H~<lb|K06pQeCrbL%E-k``}{Nm#tMual^#~<TW
z287<&_7L6>Kj*#Dh*eN@_+Gy+;S<4542cFakh+~?T@qv^WLqgv-1*75m4?2Q#bah<
zNg&0D0W3>#S{1*GQj(S{tEtgyEAPuR6ph2Z-PCz;lrELtMYJ?e5!FW|tSwcR(ke3(
zYn4=~uS&s4;^#uNB#=tTW#(e}7|C^|GWwh`3Hd!{ke#rt-=vbA2ZosywhgH_Q*&)S
zRm2<lVqn7%wkWzNb}E>KLWE?5U~d!YrU=6sb;g#1+!2<&eG0WLBpYmK0}04WQh-rF
z*zS;~q<&EvNk<*Bg$ON2J)dGgRNY-jp+orfiZbf={49if+8zm=@{0b|g;K(l2(xtz
z0yo<&u*W|u#n@($z1(R&wt}Fo*>(kk(~=Ze#ycy$m@!KAd%VMq%0ju}N!}3TCD}8U
zi!scY!lzsXCkz-WKXtDEFv_W)DPPLuRb&E;jGPUJQI(mb^cA)_JQ=Z4?<@_a*z}IZ
z@}EQ*l^NSG)}_p>ZLCT0RiEeFh7~C%YR{|lxpY1{PtuPC@hozA#00BeA;1b<!r<Mj
zZK42_HmVc6z<~AA6&)WR81Y0r^-YLpLx*b;<{w=|F~$NTrvm~ta1DXMI1hT0p!_2q
zCvn9OjY1)($#GF4`5awvrZ6pQhDlyP_FN15ZV+@!Lhj`wO)0%NbH?x&ApkDdha8bI
z|LR5?u=2IECous*!*!sH9X_EAJ%!~grZ`HG`Qv6BuuFwFjM#uld7OWsgo7~d_#?rj
ze`Fw<pJ{yrm=bAp_ixU(yy1iUKygxlAmJGwWL&yk&$4)2O1lCxQq4lHiR@Fn|9SV%
z;Ag%HI4}I%Cr%P(q!Kri#sRLY961~B@ax>!U4AlMdjoY^A3$M<cGdHUC>~Xo!wPNP
zb1g@Z-O!}pfF@VLB^;U)x(-%&_;YALL?|!vw$pLkC;cD?lr%ZiFiYq9cp<mqFSj7#
zvWPb1OZq4ij+DN9{B;TbAs{pO#|zOmFSAJ{5n6i26p%A&l2cl#(cbX6FL!qCc`M|4
zV?$S*rD{3X9g-&#?V<Mzei4gzC!f%Ra`<N<a*pf!f!Q)fcHbB6=xbBB=D7uk|2CXI
z1iX2DU$4_wr=3uKU&ZD*^Gf3CzgbWDz5#u~EC2OEn7fSrPMvrHlXzk^6I<_bju0PO
zFRm8SP_pIj{(-jv>q!nfJRW!*81kQGKgoDL%SUn_;|LknGQQ+Rb}ZOjYot5v70GWy
ze}rdy#0KXu30GoxT9v%d$YQ0?j&RxjjMX9EeDgT{%dh@fGJoGY5pWI=Mr&kJ8;UL$
z(%=8nZ__s~_R_}oowRfBzQ^+ZK_`9n)njjqhw<m5`*+eWfBuuS(OQ*m(@Bq|ll<=M
zr|HH1p=8!e8+!ocXXM<9^oMFidQx>o<yc53l51C1*;=n+ooh7er=899bVsr(aw|p^
zR=RB}jB~4Y_YTsN=ev^02We}q<}=LBuUJ9GHfDMbtx#9<qI;S8W6t~<r%tbLLm?|}
zhLV%n(hzwYgADb^BtbszQ@7}E$jIm(*4Mhdw6(R7?rm?XKCi5|@X=oyiulC<caesJ
zfprHSBz$D&jH9vI5+e$`cA$6B59ssM1I9c2=wa1r)#n;{>%T?m7ShwI()%i_($l0j
zjD`d8(DnIQ{a#u9Q&0DIc2c|1G)?*!+ZUpb{PuTWrml3)Mr9>E*twIobg>%kQgyVr
zv?|>}ZLq>Ngwk<Uj&5(LK6su+Y$tgV1}__3q<<Yt58U5BvcVEr65|5K26PV2E9)E|
zrm^_Jfau{y9~VQUMOUn<FJgRP)e-s>E5y*pIKvA=Q^+4oDpaABo-lN9bRg@3=urwI
zF5Vp&M|n%c`@?wxQ3>H-?BSQ&<1bJAakSFFALfff&*KxCx-+SjaR=W%Qb6tNT|?BD
zeEyVmX3mUTO3h|NWn=ZF^w9+_#u?{?jZ~)t>xCFA$*-oiu^|?{Sbdz8jP$!Rm5psg
zrI4xa(LIOOMK?F}?i6iSJvKE??rycx7Y{eRbt%2*RL>309UF`<(qM3w_Kte#+3rC)
z><rUbyz-8vFOwguoXDTD57qCw7&^~lD@x8NLr-lrsv5s_je+@8)wQMP#3POCdaa&%
z-J_H)M&hBJT6%`CQ_1PovigDA&d*sUv4%m4Z5Y`S^KfwLZ6H~J$QY<BFQkUjwrXMU
z?Fw;^tq{fIz4cmJuf@>@w5yHAY6G@6?G8`VF<ZBa$GXNSM+iLM@5vuB|8cLE_PRr_
zSUgv`&otI(%Z>F``s}0oHr5??JL$`>o~eG0{Cs8HVxXs=F}-<J&mzWq9T|qYQhl?E
zwWWJ)Njm#E-xXrAav0|0TLr6XdEQ=XOwhS(5>wsyM(<O3ayN7UejCbu03b^@%`M=5
ze-my7&8EEt{R8~iz&$dYZv+TRr8CT?$NdNH7-H*HTojId7eajR;fVoE%M-5nWwa8{
zGbrF~g~V5JH4EfX@_8BjZhIw^uWPr%>r#Fjm~*rKc2r)^_p7N!z5lqtTOB^UccEo+
zi~s=t^hrcPRR0>^c>EeuiOL6(C>bA#@RzVCvb!7-!UW<a=th`Oj(C!U>Kq}NM2?Nl
z9PR>H$N_YV4|qV%U;mRVz|FtGL4guF^GiB^1r96}`xsiqPn=!TdLS-5L<Hb1{26lY
z>O_+dk4oc`y@Xq!(7PTNL)ng*Tfv#8!R6421849N*-^4+PvaQmnY62%@<zYU^`;+l
z?XqV64Wyff6q4L$T|q}~1(^r^l$a^Z;&TVNa2r9s>&G8eBy6roX=vTqOs$=*v@D^V
zbE|ClQZ!rG0i3-R6ljAu{1?H0X&%E#tO(+=@k_!FZG|$*iXk5gt$bpuh$|am=oze_
zV#ZfrN(_n;0uw^;r36S6&2tH*eF<r!!N|rFW=>(a;v6kjeXU3kMghK*(lM66$F?QM
zFLr&FlUdcEKsZN<z^Jt-A&%`bn03h37K;+%P{@}Qh7uj^C1GDY)2^%@8cAS3K0MOZ
zOJ^eqanw2E(s$^qj1oYXrKr}KrKtK_kfyefm2!-9YzbjL#Z=7pQap8_w5(b~S%%$#
zgnWclRtNDVg1Vr^r3UF&sjaC{-HKP<80k1<8^f7rVJVJ%-p?hFdbXn!D#`-8HV7fu
zifNRYqGUlA;Wf5P6s=q?*M-7SUf~IVU)qxyjLf)0fJe}!Ur;`H#?Z$0h+g@?LrFBQ
z*gG&-H=tl(<fI%YXHo>ze(o<~6cL*7437xKo*k(?0r;S7@Xj-{oKr?#v;)R8`VeOc
zur(zH)A106beIO6k+(WTd4T>h%5Gyhwbz<Hw0hoc#x|IX{y=la8DDI5NJR_ywaZnZ
za4OX>9mRy>c02f9aKcycsCCBB6ff=uKR-&k4629g&zXQJmS`(oNopfZ_6JN8mmqk9
zTgN!fY#e}Zrfl#akNZVVM?$+wMnW)MA{~_dID@Eh?9Q2**NB;+ZHHZf6X8=LG=kDW
z3n;<z;&iB!7EW_$i6l?xPLK<P$7u@fZ-JLc6nJICYapkYlxqU`i3BD!WQdU&@g>Vi
zU5Sp%D7T+IC2rb4f=r3J%E82V=*pF2UjZdHXz}|pekM8tl<#^-z?53V#wVcgm(or1
zn9ZqByG|j-LHHwXT5}On=!L7GQvD8h$BT54K43Peq?rwH^Z8vD@s7R;IiY1ByGa)`
z3P0hGFw>Y*jX3l0Px%LOn0PL-5q^`fD<EpX)bqH?nP&j=tFOPh9{7G?%G@6*@V<C?
zOctQ3;+<Zrvr=I#RqO4vwzZQs?>tDYjUBHfuS;HGHaZ_rd9N}qLvB@j<OfmY5LHhr
z0kK-(Hi)zZ|HlRN(N{NnJhR?Uw|HH?-wtnI-{c$Zd^=1izn@}@ID}tCT>blPziHe8
zvvuqz4Sqeyd3L#9s~*6xQA2V$;%-+m+URU99raJsi~Vjo?)D}BNj5#la3=kuvZC=-
zvu@Grbp3%CKNFt#D9j2c-W3Ds3VdwcXvfxFXBTt+SdSbl4J6f9=6ycYf}W*|a~J8(
zRy(y@O_hHx4Tel|J4i=+yQ!-D+G`tLJ<%Ty(#E#*qvm@0^>4pU|MdHB()rv%s@Izy
z(}&UxIG4*SsZ_Fxc>CnRR{F^&+v4{mee+~5{pQQ3>6>SJdUwn#&utrbdOel1bDa3j
z&ojY0MMyU0ZOxmH$#cl$&{%6X)7JXBbc6Xc==al5Y42=prcXY7<edsQ%LiGwx{9%C
zNxD@SYcQ^`ZQAPMRch9j{Xw7ABIA>j^!UYb`t_I3(>Kre#mm?e06kex>Dfi1*{a!i
z#Km^t7!5eXY#{j>BgYnJsbMUVOg?5};ygOeTsmM?(6i@m7glb?>Y>Xt8kKF;S?v}2
zko87BpzERzViH@d3}dAi#wfPe!{EUdimbe2o5JJcqx9sPuhZfFZfdD+s4I5iSXq$X
zg+3)+XM1x)Wm&M{Wo>iQtHxHA)vg<xsXLHvDfxbBxuX2X%70|th%GiR)K6J))R*oM
zAM&LOs1IXwU^~p?!;TGGCm5vkM;Xvry*0eTucZIP)`hAPwE<n8KFoW$)Cql#vND0t
zhOZNJT#gQa21f*BT}%0*?_f+sf3|+imB$#*3=ds7JkInV69yeXjCCdSJQ(7SwD1yA
zG`g0P84o=QJ<F366^bF&I+poaO7p58($8N}S8OjI(gf{0ul7f;Lf@y~(vIuvYhJkr
zed^}u=ve(=<UZEWdy#s<Kh|}6GE9B-ZT#MeM)_e_q>L@~H|nik=UlhtbeHqSe)?&8
zxU;Ul$~Or5^tp|=7$q6^d&h(HWcN57_eU!0ocUraK#b#zf#s!`e9G3AY{N;qk)Ahf
z6^bF0X9Gt0LAR4Q@%+Kv&GgZO9Z^z06tAq%tju9xI!z0!P35U)8s8dJp=jeg4=z=I
z!|}Pd%{<n#?YKL3e_UFcOLe|SsQy+&2jkOjHjK4v^J%ADv4M`~AMJqgnTe}CJ>Q-k
z_N8N=r#9OY*6N}=u;H{(t)$-gBz^h(C_Q`8u>n^_^I2%L($;z-eeua%8|>)cUq9Q|
z*gy1nbr{gjQ~<V_^om&Oi{lwE%2zQ`R?=3hlJ2haTxCm2+EeNDT&1u1_z2y34zOF#
zmBs>FCaM7LOSJO@NUV!?1$?KwX5Z}l<NE^g@GI~;!c@N(nBAGgihA9<`~&>RK}}s+
z_hu_wp7G)lU^W)_AGl9szHj3qMq3$k^kpt2CPGSiY^oN-Z@ew=75oT)bAU?Z{Azd?
z`jq3<@Gi!6b8gn(p2q9oCT~1vcn|Y@agDr(XW{z-Z*|yiw*J<JA3Z>Mh`NheAW@iW
zv$%7-Yd8ELaCVPB9=j5#e17=nXExly*&i=l0-<AC6Pzi&XUAn4floO2!54SalfyyA
zjdYQZaN#c!Dm=pYEAB~{X~bQLC?ctRLBp$s^hN?#KM_xuamqb9_Jn7r2|S0vNeOof
zaE+iC*L=joi}6Sk?u?e*;koTzuE@L8Q9LMIM$?pgLU+n@)M>_Xx{O1AAdSLNcL*aO
z?z6a5OHNZR<ry^d9fk^y`<d?iDSPJvJ<{OLcy>zgT9p!p(PmYO<N{j_xSbT41#upH
z%;-W;XNDr_P|^@25O8@YSXh=oj_}C63|JHZ8*`MO&!JMVqx6H32II@6geO*iol0Oj
z8=WZqxeaO<AxuL!7?C&}y26%(;<YAarz)Y(d%v@iN##42!f<vr4r2|Py!f<XO3L-B
zgtrjL)fOtAx6#AUA)%84Y7xY#M~pyoEa#CT%&az_6V;PoIXX*o7clSts7GaCr9kDP
z4X6DJ<pLuhHAf`}z$5jCQHON22P=#`yHA+RcEgw?CY%ak)Kj@c4?mYFs_xvs#0ZK3
zZRJ%|)Pd3yM*U+jVwN7|CN1p}Gx(<M=v89Er>;?w&5OcOWWZO}QCH#@ni$X6&IE87
z^haG=`BNIwd9Qd&x^SyM#gi5uyrrOxSy5(Yi?U4!eFR*TJ^C+wk2K(xGreM4MpnM5
z-PAw4t)lCby7vfjpM(}?hB4WUnS3V_@>%)C3^;cS6>4EwijUfU6-7vW$ULb(PzQ5Y
zQbf*3Z;Jl&7}K9v%pv+yeN!ebo@0P<{2P8s9WIqU-vOiRavb;Uc1rv6ov_@mxCor^
z!WeZp&TtpM@BBHfy!;V1p_Rk3p8)Qrqi|Jd;FK_e!OdwLNxc(p9Qw0L=znlXFAs6t
z2sLgQClL<{L>xxqOgAU>PvCKPK)f?dO5qNH54jw!Khu)m87tQ5z}25<v2BR!z&z=0
zd|>(uKe!@3<MAhwqF}n*ireA*kjnIUCKP<(%Hbs3jYga<{P=W_;1GD6maE_?49rP*
zzVxGuGJybI4tF};aLJ!Ji+^1czI_oD`P*+|xZye1pVLJdCiRvtqvtr+fsv-P1E=Jt
z8-J5H!sCuVm<9tLqs&13UiVF*V28`}agm2(u7%EA)8CRln3{>hpQCb6c=cm%4vRVo
zztct|Mtbh5B=h00DNpeq?nxb(_LLvOz$x{3C4V$g$+O_hxLvPeg69=sPM>KSlUo!5
z{74+`<Iabz4!;xV9aL*<rp@g~>F$G%)8{|?H|gj9=HI0+e)<>b;m4n+J9i(Z*4mcd
zF?GME{4Q|&P|>-{CzW4-Ho{K_;(8U7bhByyG{9BzM+xs%-_h2$!@E)bc8H$zZT;!(
zpvd}AxS0db9U3&K^X`^>#zd{9)g+<b=T3d|{2=}Q@oswhqMNr-%PSZp3wolM<UO_|
zZ?;;Avu&9$$C*o<8-)BX6kU?nCHJv=M6b)ob5_vIrRCMCS6}k%L)PTOcy~BTN0R9e
zdPARg!!~0*CPiuNt%xT0m3k}HT5ah>^|ackdtw`_SQgl>Pw~}8Lu0!gD_at)qSz|E
zYPp(^_xlI?o<zpVsf!C%L`n8m9OnpcZmfITwDoq~x(CKER(&vujTLQ7xWh2Tq`(np
zDod~7!}eJB0b9EF;yfd^n>*@s{h^*M2-xl%*|?8UNBYX8-w&!6$<4>zUOL!6kgR_e
zhOE`4w1TKF-YU|GHrLzcsU~@tiHWt^N?LEJ%va}WcW*zv*xhrz9?E?x9_CdZ3{q^@
zwxR}zNk%F42E+91>2vYbvq9mUNsKBVd|B6%{(x+S9_H0Or(>nPNEONBHLvJGZW1Mh
zU7oW;13OAcFHmX5!(lo)>{!oOQQtvFp#D}@>&kDC_71v!CbLR!)$LJF8yg$x(MKPp
z`wt$Z^|dvxuA@JS0bQ$JiN+R{zyI><^yN2CqW>Knrvvpt-wzJGYL6`}U%c2&d#VR^
z0U3_t95nP+woE)z`*@3a=^^OJtkzpqpP?^t{v0d5&{F`6U*5(M!-&$vP9MoSC?kk|
z%w@SxX&7GCp%jKEhFj=U;Sv;&puA&9Ykf}_D=O)K^v{qpVoV}!7X941Ecs;`Za49O
zjtVCI2z@?u%+T3FhD8Sq2GH@*CDEgq%-S0c{0#v^&2TtY`yHo_`Yh+WF~(V82cN9U
zY&F|yO=FJLW!@Dg7mB>RC1Iv_q<+6IIv9a8RMPltE_KxB4?AOzZzjod{@k+kISep0
zjls1>#ajbzv>U12U?t?Ljj0%myj9}zlKL8}%~ba)3m+u;uJG{gR=Ts<Op9~pstYCn
zt8LjKMrB+P{e`&;pZkR_ik?cDFxo|3E6m>;LYkx9F?4x)=jFM^yPi)r7^%ip*Yt8$
z51yzmk;aCtD(8`5AeJsHB527N&e&eF+t;|CU$TzByfmkJThf?qriT1BNK%eVTA1$w
z7v7HY@bJj1GdXXZ-CMj`R&=>&2l^lvTdgvOoxHhDG){S@@~jz3m+tBL!fH_rb)CW3
z&rl3}82T{Ku{w{*#f&%35ab)v;_|AWNhi|D{d}h%=5~&E2<;uuk!bnv51{uZH0jf?
z1Vz0z{sa7YfuWe4z^$#!cnwFs1GxXdeIjQMcQ>zy(s;YZ<IDUKQZC*wWi-Bw`6f^(
zzAtz^ZA$RAV9wto>$ma7v*pM5eDPL?A3u7;*ENq3WC9C@#h*LVEa*KqtpwxqgF7O4
zJvg2)IigGA0muqr#;zEb@iW14xPAi^!aw{F%1g(jba%cH=G1sG1n`xgJP^Q~82q>q
z!oO(|V0!VgCy!FzC}Z$vJRqV1aytGdVFeTwuSn$w4!8!n3FL~*2_)1mX7<lJ9RA@?
zf(Q)AYsg80AN|N3DWfbCxBg0-1b@_x?g(yXfTV~R5;njE2hc58C4G)nK_)BKj1af@
zF>(gyYFGT;0O~=v$SrUuG~v(waJx0)tG`pgyZ63cxJs?fjkLD2ovIscZ+nO_>|D4g
zN#_zekxe+O3j-1cFzB#vH>+z9@~O@<338#7vXud|sU&DCA^ixY$HpZI|7?xNc6{bb
zf;8Llp;#Op9;Q*R>sfnb$V?^_2G8(RICac6his`=Yc@S25XBhf0VCOo1mlC<=jq`2
zGu`*oxI47*0|7A7teOXeS5{Opn`v=XN&qvz79~8ch7gG0i9qjd0x)_>iRG+-L3iZY
zSA5W>eq7DESY@;*fpY=cQZ7~{=rNORUJ4;wO3X{4<$&E9%Cq8pP_qqBVa%a9P)x2=
zX14o4p<;GkMf6Z~P#7>wQHS2DPzg@Yq%@q<z9^Lnmte_xRczyDp;`40O$<_`qa7Hx
zR$QcbVw@shC`9K}8+mMB<#uP4pUS1QAhH)amyLoN#9oC(qYH=Ghx8MP@&;xrqNK3p
z3Fn4UrzjNkBhoIZj<}OGN|aZosm=lFgqggsgt3!y#7B1pC~w*urHXLq@iBpCf`xSy
zS@=K2FsHV|;5B3=pcEx3%l)IF^P!GJK+4Nvsy0^A+IlP18#Oh86dhI*E^-!~+JA^a
zmV7mGbQ2f4RAX}Imh%Itj5zAjRe2K#qzuN&qf<Y$-R!t65UvQDhT@;*6=@1?!J17|
zVheV_tZNF>bT{!LwPN(|cSQWFI5+tdxSTj5wA2MYE|e%>;s;$0)AJ8$9HVgVQ}DYe
zBa^dvF_jFCgQLP%#0Bm0%p;cZOOGLpBvTsTGTsOY8iiKQtCSh90nFJpj1Wk2C!mA{
zM0!F{ajB0&E5eBp4o)h^Zt{z}yg7AHn&z8xH%`PFs3T=Si6S3<hzCvX39jz)JH~hc
zuHqb!vHa^CGhS35=}p3kn6Ojw3mP77rolxL^Q$sOxuP6_loJTdKA3(kMVffu>Pf?W
z1}6OA_MOC2+QlDn@l)Ir{L(LSiYa8Q!rLV8k{XEmi_{SoeN=aOEvxA6{ucju9G`vm
zM_ENnL9C|sLp5h(5`+my58?>Z&!m4C%^{Ju!_@w`xkwPP?zMpfN~F!y^qbP}zxuKS
z{t!Txf8-E#XP4fI4apyCYa6N3Xs6AM`)PaUQL4ArQl;MX%xqSs^Ffm&7aHhAHh20=
zhYy)N(|Es4``a|1X|=qndJ1mA|1m-Cf3JY3tM98@Kx|6ub@_ffynTI>@6GqxER>j6
zV!UH@(BEE7<n7A^lse`2Whx5MKe^+$)L7_Au0E9<cBZkxc{zujZW;|wCF>)%o!bzA
zEQlc~KFrRi)>_N=gS`Wv#lvbZR#b2n&z!~@1|tkZ!~RI}sK%C_YsmhsHs?aEiB8DF
z^K*-mEs=wfIpO#!)kKFgxf-=qwUKxbePrfFeMNf6s^8UTr;O(XJ$@Ix6=Q!i_SrbB
z_Uaz@r5i~0yp$}9ajHM?*;4oKY^0xl@hEMsSN(zipfgIly9dtm&gPm|aUDso`0B}l
zWpmC8IzA4=0ER9Md~CDG$8&TPbS*wuqsJikBWtte7*CLz^p{5u?xoGGb#D)M+#5RG
z#+sf<l56?k{ruU}wAQGmzx?yh(xZD@l4H+R??dsecZ%d*R>|}R<8-X_toq<%J}bO_
z_ZUOgfQC+E=|du$nyQxv_qWsiyIbj_hYzJkE~T#8^7-CT+B-N_oh_vC=v4Ga{@{<X
zuUZdd1ZTKmuweT^R{CHJSQSxLN}ZmlE%Q6R^jUO`Ikgv)0$CYIA6pS`m%<+m2dUHT
zi|>K@$6{(USQ*sNGrF4A)ut`=AtoWBqqLfJ)p<kxhFwioQfG9Lo;*KDzy9@a((@P3
zy;}wP5_&Ct3_SuP3i=Mu=-CN{=yn(*Sdj}qT<9n0cNk1KOO8(O*R{&QTUSoeICES2
z9J5AK_0vz?ZpV}QFqY6SId2NR2%|5+B*KZFC3j3RjQ0!Ru_d?1VGKnB&l`AUWfpaV
z?i5=Kq6<nF5?`z)qpv}LF#>)}LPR8s_mlNg#iRdWkfVLV$j4CzE26pNG!e#1Cyo)A
zcYBS{8<jRD-7uLn#vbAxe6cf%S4&sbUTxl|dXK7qz!!#=zWPt6)A81pr)TJ(p%*ef
zYtnmnwzt#zdfR;r9;?#pnw9yqy{Rzqh2fHMqD)e?Qqz07Wf`^XWP{##!Z!zv1`O;y
zjp?I4<Bp>j=F?9eZKh9kL7(ReV*>SG%Ar@DRiroeyN61xQqHfWt3_5!uBaa{X;=EU
z@)cqC6}CcTyHN~37?0S6Wr=>T@|^JOQ<I)#qj#mUlpfwuf7O^7pC}Ld#QBN(`%vY%
zNZrA?#<=QRJa26@QoF^9OpQxc!Kx1)c8BTlZa2Ny?|Zw)qwXN>=^4<~vjyY(Cm-EQ
z_jV-1tgiZ-M^|IMr#x6Shw+a#qpw-#*Z5mmSxPJFKNqS?CX<2z(5_zjr*iULty0lg
zU#q9hcHM?-+S$fZXkF=<r8?k@!!wm}NqvO6JXb$qGVSFht3LVuK;JAff)FO}&2pdV
z<lhl2I%oHH0e_<?ptR*3XE00G<AeHjoGILh`&S97gV{Q~2DyH&)!82iG8?aAg%VuN
zX`mVci?ERELw|Z@F^B{}DDGLfneaAZvovPl6~R-GJQROGyjQ2#xQl;EA@q{jFrEq}
zr*DTTX3&^Yz6}ZwZv*k3d=*#+eoLUdG1lLn#_QlFZ#-vs=aQ#u<Q;hbp}|`ne)h@7
zUi1*glFU3ed40=nfI^0o!m6U3hQd40csN}LB}`CKkw&WZwEcxQ`(e#k5G$1!0aK?z
zd6Pi@K{w+XGG8~saZS@tp<qP0B8~AZ{18rq2h~EsSb^$>cYeMHxGXu1!!j<kf^P5!
z-sE$n%VU?*fE)Y}r(1+No_mA@uQp6MNI(8pOlGj~6?hYrfGMen3?ehmk|4rM|8S^(
zGHv~sf84t4(vOoFLHzM@Hen)*6MEr^S`ZMFaSHJ0ChG+WnXQe@)V#Bqs%s7VCH*Qb
z#vchIQW|XBK}eO5#(7c*ILAj_=dsFO>=NwQ#sEc)v!i&dvPFOoG1ea|q7cv!Xiy$d
z1o&Wwf^jy+u+>RN`}?VTbnGqanEA+zN@h`txCL|Se5tbHnQ}h-T6C!&1X2X=Inf&Q
zJ8A#<^Yr5BlhoNgN>>uHZ1`DSlz<o8cX)-E(k~#Gs$6Ux0u6*kctv@(qM-}knBB)L
z(xHS$j43uEN$COHCe&@FgPU{5)+E4(VQtm3LzzXoEZ$j7=L4RVR~%p*iUmzWd!SIn
z)`H}t@?pS1!BIvWs>?Pgv<r$kb<f99+8m=5v(RW81{M!XD^LiRDjPGxP=eWlV61jR
zsUXj)jeQGNvXrU$VO*Y^jNC@#&$bf?uFU%LDltKvKZXK=qQp!<gn6|){Hq;M?mRnD
znII4|gVc2;pl+c@yl0aMjXnl4%93ZAW&TmD+<qz_+i$w26z8;}7%zx;rhrSC+;$S&
zQLs)&=V{pOrQ^eH>K_kOE(C0GB7SW6sxPJX28VvvMPJId+Gt5ii;aq76iv~)kU;Mo
z)K*ju$>D_NW>66nY?<%LI8z|~(>%!?#v*(8bsM=nlW<p;FP`1ua<m}Q5H$PoJd+YX
z^a_Fd675eIEks!0nuHysgL&m&2mp#Jw5Qq0A(|eq_Q_9KaAmx>kuE2iq|NvaSA4$R
zNNAV_>nGFUmR*xD;&Vw2gpw>ET;y58xPx;YPcR7xj0s-(Q2<Wzo3>|A4zFVMUpR4C
zAu<DQe-|tR$>t->*<7oeNE}V#n!84_xM!#dN@+#U>2;O-@bm975GUX!HxnL=m2rZq
z94;(jaerCp%5AE97;_>o_hq@S$2ls4JPL~OK-uvGHc^pV2J;r_2`l;JTi`{sZpLsp
zZ4v!B(+2@?0zvlU#gHAC@l()4W{L!v6fXVam;_|Ve*H%Z(q=~vu}~}KuGdFSFiXl%
z0`pyIa2Y1}ayZa6G{w%2$gYG25Ek`87#?U-U-%;^XiskGw_pCgIDZJpjJylnWcS-h
z(q|$~^zm2Y&&@OYc_}rT8)<X<UaB`+X>ENyZQt3kdv&Fu_eL!(=)DoogP>guTdv+L
zpXONKujg6ZH_P!|n&G?P{IP-hw(I?NC~foo0QpSc88_iIY}^9y_5SMn+bB<;FRy@?
z5c0Fo!Tfx8-e1=~x4|s-q?SAmc&<uD)E#5jaetilB#&{9mHz#~A3X*!m(@Y%(h>Mj
z%_{sxqsI4Y%YdBAOB|~{E+vQCD4<Lf<w<rYl2h5{5TgKFe<S0FY+9*sh7PheAHJ)p
z%{jcR@?TQC5cmMjjuT6Y=Xs61ixF*gWg%_0Drsl4WnQ}Cn^iy(b<@1^!;Ep<8Tfpy
z&T&_1C<iP>ou8eIr30n(`A0kHum1eA)NZ%Zu4K?X&WDl=zR|9ycC(uH4~OZ?XGh+x
zf-M-4q4)O=<0EyipL+c`{}`Q!ae*F%;j1d$0={^jV`$^tD<*M${@Eus%rgn`WIS?N
z>(W1tj*ruqUwuuJq`&&b7wKo8-IsobUWV~v)pRk|3{TJek(}+}SXpv#*hx>H@2Tt^
z)8m6RXHv6mBp>IuH`mi&{Nm5kot+)AGM6|ri8`V(_*g$2Nav9LhCYKqYF*DxR`(2~
zTd@L#GTNY{Iyo5)W8#|V^ZfKjaFzX3bxHg3aU7!oI;OW1RDZw_6EYBeP&$NkAPlBV
zh+I({U#fk3y`FTOQyT?#?rf*6t*x{yz3fb~{a|#O_V*85KMP8Wad9A?Y>W~w^iA|0
z^eFhE&(O-`#f~2E=kt@LORegnufjK5#<3L~?FnGOjme+tlggvssHUcL+=1Gc`e7v>
zbkXrV0dgpP6MY$72O6qWm(_mUp(n;iCin??fxhRk4D@?edtu-~KSbY)=W9HJucUin
zV4_bFj-GE?q7f^WLN5-ztr)$?UwP|K8k%tc!i1<g3!U1AJmGSF9y%g*24~LRfDVT)
z9q&wdYpL&UZftnGFxghgQ>WMQYQ^Db;PGvPqzGY<<Qqeo3`u!dMaB79O+7o>t(NPP
ze$!WdvJ$J=YWtyqad|i%y1l6Xx^(29|JfHB?`zT%hiR;ShVj>Ze5>tkMe9}R^y2;W
zbd-9Xp3|}-Ydlc@8unG!aXuf<kd2LX^TRWa?H(~$Ll47`{uAezkuO^wA|J4Vi*xoc
zJYo#2=(*KytfW=-;n4^~*Esbw9yyPUN@S8P+nF*k7GrB{of(EC>X6l3tj;=SEO*Cg
z|G1x??Hzl2!%lDHtz7Tl+xGU2829<6!-Rv=lQY#b1~86FkOC~;(CN9?ImY2i@p{g7
zb*YNdit<@gei)b+7uCY*my8j{ABIW3JMg{X#AAsw!mD0grw69`j@nV}iII`MGtXAu
z@{jc_!=R7h+-H=jyz~R!AqJHZoyMyL?GLGnK17{6$iD%+7H9ppK_Ok>n&C^ag>L8q
zf^sA8Nn+%@muXF8?>|{6WqVC6tKOoOZiA?w&|!hs(3!xu)#rDiaSMnnDPPO!w87;M
z%qiz`LBkI@`b`Nh6sKt;ANp+8$8s9~(zn$Ixq?Q}imzOmN+xOd4CLVJ0D0Y}sk!K{
zgz&-+e<=_CDV=z313dE$jx2Nl0wTq{-T^?w#F)5F?=~oO-Udn@x5I?zw}9aJU3hyv
zyhK@z9DOI?U2>j0vwk4(R)-sF>wn9$lS0IYQa*AcY$K58i!fq&c!Ebq=g)4X#ZMd!
zn`sq1Ih>hq6TJMF-Tp9an&u$Vf}iszEic`;>oPCd?eIuL0D>#`_)s6=g}?Fvx?I=F
zH;Lzhgzm7c0#5UG&N*GiC2iysc(BbcdFszY*7-w2cUKX_Ox$;9DbnGQKY@YoaOcMZ
z#<>$T69+#M5;1BJWE#TB>HSCkx(a?=@|s8Rz|UdXEe3Rl4dS`aq{W@DQV-L*lbiIU
zg_iuvME9T*{=jz`p<(=)^b;KFM-MZ_aqd;M)ku{_-G&*jM$?1P58??Re1yQwLlUD9
z!VH2Q1{dnYhfXU#W`{`t)tyyYWqSnxp_Q>{p5+gDAXe#&`Vuy!)H2gf+)9{TO3PC2
z*t>m20;kW(s5f+HMzsVd2~;+U=)x$2!Gz+X6ft8CC5_^GYeVkhk=a!!A2!}es7Hn=
z#tH;;ggz8;X4Apb$q41~L}iN^NPPYm*D4(^qx^7JU0@($7L;c~Nk~Ls#PC9WvxSMb
zF;P9S7e8f(mXnJd{eU__skcleH-;szwnMofl0f2csnW=k_OKG7c<R6_??ek_2t$^Y
zL#3hr@}U>{C`*uoCfkgl^w5@60QCnQ=#C}$!hZ;?96B8{*J3-t;3xVNFG^@fp7am;
z9SVtZH9sgNt0;WByP}l>We(aXag@t_lXNKJq8UcFC>Q;Se3iSaAj(b;qnAKU09`6n
z;e>f+qIg0vMwvK~0PX$A6-EN`Lm*#Ti9X5)1n12q-WTS?E3`z1T%bu=Jytz-gNA-1
z>p;Nk@VH7?xZUTKF8o~bm%9j-=Zby;M}Ho#5G7yZ!hY0!;OJK3a{5A-c+$r61Mf6k
zNy0s$Jrn0L;t!m-<OetYou@{W;~2!;DH^fm83SFX%Kj(|A$Xyd@f;p_ChHhE!V|c7
zjUVwYlR~CtD`^}n&IFW!&JhSX=b8KocP|R?T|b3Z$#-%&On2uIfrXA~P26_I#h=?e
z(<W^W5PSSa8iz;vNJtE5rjYU6aeeWDi9h>&*MIKU4sj4=jkcMzZ-(Fiyd2KI?#`2x
z5qS;AXyDi#I3+yZN7?T*+!?W%4skdM#R0E!-1;}6Q~Wos*)V9>9WBe<`OMNbJ;M3R
zWp|hZ4b%E4<-<So=dg0+IPP2;{KY@zA>I+;rhbg!KfrZIemR708X-<)7ir^1AO`J<
zK=+qG%0W1}=WyH(E^b#kEfN8f`pI!o=4fZ!het<0vaJp&-*<;7YrKbD=eg_wa`W+N
zVJWrNx6;R-{h6+xrO$u*m+9w!@i*z?&wr72?thZn8+THz(N0UNRpCn>nW;k{?$jsm
z%;{h7TCQ({oA|d~e|k{r?)$l-jc<c!e{jAl6q-LgI9L6=wrXya({d(Zj`Pp0@amZF
z1~+TButW9ov3IO%czTicj{E7wK__(wr#^Sa=gBZhN_fbQ$aRQ>$i&FYp>tf=&Dg;p
z#-uNH9~fhVk<7_tDda@Xi9I`&P9nE;Y{_y=_+qjfy9XfuF+R~(I6G%$nK8}^l#_J0
zzn`AJ*!Rchp;v#1f62z@r=zqWo;uxL`rYroPQTHG@oQme)z63F5ZUs?dGisQkF@~u
zFT8d-$7#?zPRr69m>`I;%rdd`Ayz}IDF0r6kT^4o57vCJUzHBVsx#zwwv7*6fYoWy
zN~{vWKs4<2T?ZkLGg*zT5&Jd{kUv|EHQ9M0*N0cC=)y3>c7_$@!|Ida3Hl1#fuDM!
z93P-rok4wZZW+eCqrBpV{LpV;g}UU!b6+?e8?P=?wN~|N9agT{kRpA9ywJP4$A_NW
zczk%6_ILNZat~uu=eXllWj)og^)B@X+&#{H9u89<Bb4&TKr|eVT(9s(-@&-&kKZaE
z{eww@HonNuK*8XGv7oPZ8jeJ_Kl0ge$K9dIrF*BRtD`&y)?eA?9Al543rbHrp}VoA
zezU!n);BhN?i<F_dadC}l-{L6?~4_^2V$i!dZhSLn~60K4*1iL+ky}OoGFYEfp(#f
zpkp~s{dKL`O3hj=)vKHd$?C)y!^l{n57BSwM{0y%Q$I1*%(ug$uK<KuAEIugbD?)(
z?6o0FZtGp5YXT&8nlL!J-(WBkACL_FpXv?y1|1Fh*3Y0Jux^Gf28f*N5JQf};i~Ej
zTHd}i>vc>3b$KMfGM?E!k`-R8e8gB~!=iYwUMBo<tO0Tgd_#i@;M)VMBU!z~O474S
z_0M^BVyXBY&8o%|U6*t8dyOONiS05Mm*rM}Ro)nach(!}-ex0hwyPSGSL!o;_ak;K
zVO3(i8s~D69}{5bm7WQe<1vOEjNzU1^yxQgr2g7fKW2MOCIq)yP1S>Nl@A8`(FrRT
zFBK*o6r(75dJ>(1#@VQ+ezKez>MM-1eT|zhzy8|W3eGQLw8xOCXUFlbaQm)%Z~Lmg
z%IY=72Jx%Q^QocdjL-I~S5m7{%PVD9Tz{+i8JI5F?le}*RhH({>Jlq3=bUb%Sxt3?
zv*Hq3_qS^4?p7;pu4z29SZP*IE%55guDAQ!Z5zQT6O*64QcTZ;Rrz@?sJPx@b%is|
zDvPPQD&G0dqi6Ce=fjb|(r_kfv)NWX)ckG3eNSVX=V44vqyF8?-2ZdeH8_~|wR`&e
zCNMjOUWOc}n`OQU^;e{M9jQMlxL(kiCUJef+zL1R6YsZI*evba;LTfn3Z#5p>L^KI
z-Rvd~M;X^MkmskL$?CesR=6G^8hha;OhdxYKb~Uq^Os58-#E~K3_m<i$!D5x2AsyR
z&oJYB3F4^@<mYoy7$>!fyHJ?TDO{0S*EB*9VNfX4Bc$*;MJ}?+_Yc!NZiC=r!u=av
z?;Bpq^G1hyDgF(v_f^}k0=7C_*jQil8DAmv#0Lcg2D@RaxOI17!F3S$qCz--f#x8f
zJV5v>k<godyf9Msb0@5Xb0<yQvs3AVNBarDq5Oqq{)5$=Ci8pkBFuP^FRDJpj=XQ+
z=F509ZikBwt2^Xz`E=6-4c+aRn`n58$VtAW$8R@(@I(n5P82x875=m=`Cj6{bvokW
znzRpYry-Vf!3S~I@^pNrqia$Ja+ApT4#yjDgqJc*{zzZ^b_<s#EPi}Ws}9cH;n33s
zeF=o7t?<Y%rxAEND&czJhZMn@1o2e~D_4?A_=s;MR04)&N&z#hm<@zt;s-f&l%APx
z{9!b*;mZ$b376n0KgzBI{OeJRpsim#`u9U!X)#7HL#mH*Bc+7xCX`$+1SuZO+N;J)
zx>YF(xY-I%a;@{{+zbnWx)6%V$V##3u@8OBB0@mMkR~A;p&7-7mx_3^Qjp7pqJr=X
z4U{|fAjbfNLW|;I!-(i0)3dD!c)kGL5csJ)Zz-rUpon4sqYi9rV>`m-Rr77bj{ayP
z6mi;-ts*g?m2D`v0`e9UD8DK<<t#!zN+SK)dvzng3x}-|ZLCsC3{La`_UAqoE;I9N
zAR|AO6(0RzLG{ST1u~%inF*)r_m&(eF5tU(`oj$!3`g)2E4-xSDUKI73JT2wK8gv#
zILZnEC~-tmp7_uq#5|s(eszgH5!|fgNx7hXm>EW0vSo*vQ(EZoLbC#-wuyYBkDasi
zj1=UPF-pmax7763h2~3rS3)+1FV2eKnp6LvPn0v)FnU?oCW+`V6ezpL7R-?sx#OnY
zbt4{!|1u7>5q^qZ+@g{1F}C9}#w?xLE0>vYQx#7oE}m%_Pk*k{;-9(nbL~F4+>}#4
zS$TYrm43`Cd_voJ5uX{#S8<Y_IMQ%an-KwR^X!=Tvpc7e3u*+I4`)8bW6TF1lr`g!
zj!b64NHcks+8<OX6khY<vcMcr?!kkx6<^%G0ZI%@_Hs9U2j94^^B31_nC>PApFx+f
zxa;B%eE!5*q{R(VyrYWW$@o>pDadtZztht@qim59oQQ{ppZ%aZKIg|@4s@#(zMU6w
z!3(U=4uPj|;uRNE6f7Kh?K5rg`73n%m+`oSlm90;Byw5#EBE3qmvhXtXTr^&?vAmW
zdW-utTw~;m_#8K>V}(z~mfXewPZy=NpK>}ZXa~<k#hu?m(?0OPw>#hU=e}>4`U@SG
zQg{9)b)vt5YuXMsro;S>C&Y1?7&!P9XKpDQcsV@EkkY}?;h+2&U-5CS)Vk_}vw9X*
zQ@y#FKKcCT>5HHKrLUj<`Cq3`zxeaCbN^9lwl{rd3m-Vy)=EFF)7*1s>T!BS#lJ1Q
zYx*B2ynDUf28HJIayD-SCG8&uzCGvZQ$!v7+B3z4E$rXCvMlj75P9e~Cl7LxZIQ7g
zQ(i7CrOx;)?H={h-ciqIPw{b@kI5Ku;-d~Sy6`32BYPpYBY$HQTa^xge2ZL+%!*9L
zoj7m(SE*WF#3(|(KBHC7G<J1hORp8ldC1#<C(|MS%6*QkJRGNk{iF1JcR%$9BgBYQ
zSDM<2Wa?!-3zcsF;3)m`zx+P^?stz<cObcXep#~WQW{FX=t^I~pg`SXkg*{~em>@t
z2gaY#pr4M8_=rE2j4YaZ#xg036<R!-F`h`$iVt42f%0U9%4j4VLUm;0AI2u0*<uiR
zor!p?U_!oS`#0RQHgyAEedS4gVfZ5rys0ogKWTeoP5PAdCGn_&C_jw@H6Qgo8lCu^
zfPuos2%c-=spun|>5uHsR)LLX+r}VPyJ9$FGFyeSwe)<%7{#g(R(ZfbD|rCUeA+)a
zOndtWsUseVXVnzCBgO>ICt`&fll~|(?^q@xvVEitTH=*eAFN_Wcf#Pr-Fk!FA+yq-
z&^y?51Dz}Sl`i4Z6)=#}Q_-2Yv$Z*TVCZ~dT%-Q1XR5u>-!Lq=K9v`Jtj0vlM#F|I
z@}W;qS8Ue?|L6tO6}oG^(e(Lr%d7NB&RkwGT{;QdQgTihTjiaO`)SlaR(p=T8gP4i
z(|EKiJYa;wh!DoP=+pFd#-OTPVGc6Pj_1&cvJsW7?!_Mlu(E2*+tfiv1ZXe%1G<%d
zT&Lb8Mfuamz|m+GRi2+1<KT<chthGWW7<LCmN4K!|4f>sPhrGpHk+bHn}i<f1`%EI
zM-Qd_C@=k*Z7k8hy~;~`qGMttsW<EDQ))Y<cl)~!#H3Qf^dgsTNBdXAS812zO)06b
zTCM6WP}^&3ULA!YlD>cemKBcYVuA^_7>hCZ)YNVmZ7#$UYVW!3)FrFU;2Y!q83Hwi
zHicn~rw+#B)AZZlewDuZ=4&5+P**<md3ML)tUI>qWP&c^9itoPhVhMLNIzu775!Aa
zFupPVzS!GO`+K`-(Cen<c|M}ELYA#nduey~g~EG&rc&4L75b-*DSTbys--ev0ISzl
zR3?o#&N-`d`~%Na<-xg!$K7<;>G}CwQ(xU$YY1mGHI=SWTlQILjT%O@W?FAIQf+xb
z&$w>t9UrE;p0zuhjZ{&U!(Y9+lp2-gR9{_4YxR}XmVZg%=X{^IJhPFlF66qNM@x(G
z?V*EldZ04tk#)JaqIRSWS>dW@z4Ud?R7>j1l#g#js&$t$0EgN?|EjgGxo(3=(i<>4
zeu4&=y4=5V82`8@{f_vEx_14mfl}Bad&IM80<NEO75_}QvE8Dt1+YlSlPD9oc9qtS
z-#3Rs={6{Fv+<KK{lJw=koynMzGd}oo*VRW4Ck!uu2on4eU9ywH`g0!t6kODT233S
zs^VAmoMY?Om?X}(Mao3EZ}#MpH!z90jxvF1%>c70OWj9&!HaqZW^n}*X9+KDQCejB
z2NbOL2k)2qj}@lv_rrh<Uz?j7f6Ggj2^Ag#a(iIl=FUx?_z)6t{)m^65FH5T4tB)r
zZZ~PJrw50JM~UZ!43cqpLFis+mXC))PkHiB0P)1hWnA3QAtuw9<u~pM7aqI^4Ui+;
z-ibSC5hlz@xDtUBe*T1)NjTp393DJGUOo@Q`8qam^B;-)H{lt#aPd1oWy@XvqGG&>
zTe$EPZgWCZToGVT<Tn$Wqkt(dkRlFxqDZ*Bb`u*>cAK{T$vggXcqyZCgGMP6fs=5h
z!;q+g<v2JgkZ~tI{HkXO9&ByMYOf_dh{>BMzqWK=WM6PSv9F}@AW%{7D4RBhh)&Gz
zQlBG@F5=0XuoEc{?C;LjddzMh4tx|D&Y1FcQ+lW~!|V9)AaxIqQlEXiCG>?*k8wgm
zKC8y0Kzr*z-7)%D2~s}H1S3C;S_sW>jxxy1L~o5DJcL(P^_Xv67+A!HjdTdoo^d9@
z8DaZe%5WIKP7`O5q5M!5s)L*N@n=<+Z~@pPoVuV*=2cf#z7%o^9SN`SfPwD+V(-6}
zWKEJhPt1acdpP%Chb61ZEIwV+J0l~q0B&~Y1!nOIfZzf477(ey1zrMgfd_!#61#xR
zqNlfubyX!fB)EGx4IY92@2_U=eokgqlFH2P>Q0OJ{Ik!9nwgrKnX0J~j?oAKELN@|
zc=Bx_NO)(b8_F!0)I#GN0mx)F_zt+-r521cc|{3?Af$qbfreR@aIFkEFA7iQxHAEq
zfd^Kcfpb=qQYa<?NFQP4Rtz!%pM^G++X9+!&IE%Mc*^(sS{8_@WBnOF`a>Apkl+9v
z8eb~Eti}y#g9EQPlmo{QUL!;c26X{0$!n=gCI$5JH&`~V1;>4i(k4BG1K`2<u4IK3
zb>Ny%WLDzAM>#Fk!8J#DTu`@yC7BKUtgLg9Ig=_}x{FZ9v6#vSv`sz{3@M+LoRlZp
zSd=|Xq-(;DbVakyi1WakiBbN`A28zE0f1NH;HqzZ$xCm;zz=tP!F%)tJUA9R;b&px
zMgC-yp9?#OX?uwySO1eZurtHPxBO$WD8mb<h?ikFu;K6uc=0&kEG_q~B~HnszUKgG
zK$gGZBOi-y`q4iR)*KG}&)oG8-dXVP+#Of@5gGqHCgwjmgbjDYUdSi*1sHb1j$5|R
z{AKXBAAg25cPruei*MqCH2nRqdg3?i@q|Z#6S#=)C}#x3d7iiI#&Q$A!Fvf`@Il#(
z1-lVGyXWtjow)iV?}C)mA%HPt*ssOSZ$q#bZ}1TPmz?>Zg)48H*>~6(55N!Keo9y#
zgPQ-9Kv9Rh-*k`fnZMIp$PeMdHB9~vSlM6yWjr$kf<!3&ff)a17w|(of0myatcV})
z2%^98tvBrp*MJgs{sMsiOA7Ju(hq-jpTUm4zxZK`zAAsq&s7Mn4Li4aN5LO_rF7Q&
zm!|llfNvU&!y2>Ly;0d}rMKRBKmG9ipQRuE_!sHNKmXhG-UmNRyN}*VwR+c+FN6(j
zfg|WbcH<Fy;5H-jjpr{h-vo@mJ~PMVH)Wy~eA;}~H2!(lhq*HIUvHW(%J--JJ`ZD`
zX&gObc%)G$!#zLTZ#%raNiUDjedi4K*sxHF9Xlv9*mYtBnfO5XE>3rVk7vRg9-Lhw
zOi=sIB=LbUsm#%3;{EjI1PCU1*-?S4L4O8M%Z>&nOW|$rAK3NJ?y%rF>HoZ&3B}cD
zk}fZ<lr9tDjnu4H(>4>mb@9HEjo9(@;q!xZaMTlT&asI#^NPdKFdd&9dp8n%ID8^K
z17*iBZg3PKN;DJ*Bk`XDrG>(dqu1asuazfwnr^%0-Cw>37hYXu6H>%Ciw6@_a6}rG
zvy^%G-;ha!JAK<tpFYh%;gNAr0>I0{Z__`)%E9w<yyM=s(%agy0!i6+wQ+Ya3-HXN
z>j5tH01OldC@Y%HmhjzBf7nc|RwuPP-L&^;Kke?v(Sa?FFsyPkCCap*1{7e(Dl}(y
zaMX$4Xr^kt;d_ragm>zT46?h13Hf@BNx`ZWnw;{141%JB9dpnWi{X$3%wr%E=wbIA
zw1ENyS_D6ggO<1C26-Z@fGGTYnuQ+n$l&Wmo5DXdiJXhPf`2)J4cbF?uuN;k*p(Mh
zB0F&Oex+LRQFe_+%}N#SRD-5GVPC<$ZlxD+Nn^@UZkJcy)x~js$nEROI{HB-_}4l9
zPI4Z~yQ*-;y>iey^o%S^9SfLcJ<8y*$bDuQ#g@B$OdHgO$0!yDdb}#ZjjY)lTICO#
z0<SC(MSerRqdXh~Z*amXNA{!es#ZCPCGR4Vdn}5KJH=cdG9-aq%^R8rpU9!$8oERo
zM|S83kke6iQ&!yQ2T;;6XMijU4(SiTH;TYQ2jH4?$#29_y(u1kTrEJ~Lfvy;_;fl>
z7m}6ZG?z%bqIyC+JU|hChN7Cg%9m18b=Ph-jWCpOAmC1MrW1B_4N||~Pv^staD~!t
z)%45Vef6!ncXl10oT$!Y7b%J`jvqu|-=v?SyebRzPya=q#NBc0n-#ZDc19k)I&h!W
zQ9pq=cYAg1-ILStNOTbM2wcH1nQ%<#t=lnk0@TN?`jGx$m<|pOl!q%TtS&A_;<c|+
zr?HuK)z@uz)D}gX*Mc#<Vu$3V?+2vsW5?xac<#2*X*W~1qrO=AX3nP9JF}v8bxCQ1
zhhE=(W~;WD9&WDEXmFM;gd^zq{%VpQuBK^YMePy=tm<1pQ)Od0HMhinZN&TyUD%c6
zZSGm)p25kKg#;`fzE>So($l?G+*zo8g0Z1aJ5&F}0&Ln2yXMw1?b4^XpTr1%v<T{%
zCrtTz-8BsDh4-R+hUFK~;Jo867M>Rq;c)+#ZB)K8F)%Z8d)NXC#TBd-6a7>4H^sH!
zjGCH5`s!w$-q$n4`~0}YH$uGjC0tM%=KL8`8I#y`{piuI#!Plc))kJ@J&50E`KgQN
zXt!&rixPR0xhE9Yt7&JuYXv`Zp^X3ZnX#}{*wTNbUeeY7%&lM(gFT+vF3Ic<%@!E1
z#{o{H!<<<7DWp4Zd&^||Vl;gh?~qE%AK$OT|K`msW8a$v^KE1OO=HE^PPh9j3vl@f
z#u&yMCMDY;kZi{}-1R_|`7wdYG2+eL8I(Oyg(u$;bLJn+keo=xK^FbW;}~=M;#+@=
zAJ}JPEC;g5PfQAtB))o_7u;~f-4N_f2>UrpB#^`9IKT>c226+uYd69OhQsH0d`F=&
zJX^z0?o5b#qBQ5DOa?lN-T3DZc-gLZ=?t8v-av@-U(bjC^0)r;TVY6N(NBNm8Jsy>
z`E%HaXN&Wf-PO5~PuxoxvHKN2{sw6I<{fuW@x-@Hc?%A!3Hh@d@%7^9J?^BcsDnkp
z<O}uX<zwx{P<a=P@kkSjbDLqtVJu*h>5wtF!DLwAkROT?n7wke9ts!lG}52rC@y(~
zfgF!Useg7R#z~Bln7Ie}b94|pkYL_exy741o`6qROr9*%XkY<mIW7T#7zGmda$Qcw
zW`No4wSm%!uwukLkZ2HrL5^J|$PmGZlUG)9-HI{8?ly;;!pS)wWf<WwW~`JUFy#7>
z8&BW@!33p`@~k)b^bR&LwVd}3XpHlN&?Tl`nKHu;vq!?fjiWRq7z-x(#K7!=0}{jC
z$7<jwoU$_tVZ!hffV;<_S+Yz%5Yk}^%dsBpY+?l<f*o}i^+ehAtAFxORuLdgk8-qv
zNO`co;E@OS4SEQNnF}haPx7wED-D%bhE(c}B#e7y4#<WNzM`0>5z1<X7x9f7h1Z{P
zAb&0m1}?OZTR@T_Jj^__z)IW&h2@eM@HLK=5#x-Y0G%71O2<cIQetSs!NDB4<rVEe
zoUjuKzo-ki50{HS+~u~|FGe3m6uu3X5Bts9y)wc##_rBo4x>N!W&9cL!dp?8xCZF!
zZu${y+_FRXMMmN~V!6mTy=HLZMV|9V$$xz7*RT0|Ha5tgSpIfc2X`1OydqoJv4p>W
zvRi!5!_8rQ)^|MoIK2Mi-ZJu&)3CR`U1hix{PK$LnH?YzA*Wyb;tdelWRC~Od3gw<
zcTOYROqO<=|6&};{t$?A<nZxT{36Y84O7BI5T&ie=VO?j0AxS=+cA48EU?SYFUSn)
zB|wCq$90)rgB|__W(h-rw(BEkkZ&eX@SOY44EZBHze@Sy4}iEkyz7ua4raSUIe?x0
zj`GE0XWOv^Y-85>3O|)dFA$)Axkp(PmiB_1<6^<i2mM@p&)dFVau;-a;SMIl-;M^5
zUxtGvaN$pV``mc)fS=<yR1TBF^5?fbA}@B+uiJP*6>j*=`z{_}hkukg@Fjcpi*U>7
zw0HJbrT8M>eycn#(`Opo;Tz)UBX)K@dHenJ{*Ql|-uvL^=?6dhMSAD`pQJ~R-%j;5
z%C1eJhRH0bgs31l;FxCo=KI?;Q6vBIOyDkP`_pCoEVsYRxZ=J6^G)aLn=|AJ9_h1<
zd};W9#t@g-<@vIvr0cgHj^E}OL6pVE=NIXZFOSptfD;77$LfJU=TwF`=6cDzF!!x8
znGf#?zj=Lq>7%c9cDGZr*;LaDeiQ!Ai@wC)att~O9PaxXa~zrAGSR+K*|72fxbU7V
zfb%`I$QAm#)th7cn3RUj>h*eh`eZNd@3m9Cx|UWR?!1G8$=BoFFrD>>Lfwk`i-wQ#
zV^LjC@(OZ`Phn747K@?K@Zz(q7~u%Boi56vn$lr`7nAsLA5^DROZ8ect&7)W(wXA`
zk-6A$G@4Em_gJm2am3zMC;&2Y;ZI|{in`=`Rk9aK3HRVkq-P}&JB_?ERpGfah}uQL
zP?em=4mx<^lhdA${sZTndcb5dJ5iW`zqn#o5V8<QO>S861+ILYVU-<UwY07BZEtrx
zDNkH>!dzZ3S<bP9Q`bSYx+(e)-+s+0ASy2kutpuC@q)BF$!Odg23~nk{y|kJ&5$7|
ze}j7!+a2$kLV<zYM;K(PP=raY3qn+Ulvl_DOO`vK8|ulA;K9;Et;j-gT7~LPagd4l
zM)oQ)7_?%4<QC-%cjHs(TS4O;MI3hs?aEE~r+%x#Wo@fw1r2S3oo>CeQ!5e1qk(tq
zT;^l^7{Y6+k8Zn}*rCXdCMG>mwjjSOuSovoI9<^P@W6+6HBtte5OSIFgNz8BDLK;z
zbb<mDItoQpe3Ng&GESCbco*9~Kk#oXJoF+_r4jenh2jn67ce(AM8E5sR#>1M#g7G)
z(X5eMLqSOagZ5&<B?_JkmHqPS!paz*xS~i@9}n#SL*1fe;Z31p6sF9oGr^&pg!LSN
z)8igM+I=W+S;Xlbd}@=_J+#|yHU#6ryE}(tb|l`pZn;wmxV{^!uJ)=lFO)ZSc=m^v
zX>f6!E@-==C+gwW_!9nv6OLv4`@jGD^xNP5KAjz(rVX`ejws~FPx>G3v)WTX1}@oI
zdU$l=g{5x)YFBHyE}%W`7f0d6d`eyI1X||4ZSKmd*XqtUbwzvQ4m2NcSmFAImfC(L
zUEiyHRpN-u%d6XTaNJ9S;ncgh#wsrh1kTiVDT}G8a?($)viMp32(%rFRCb8oxQ{$K
zI!!NL9jDX&NPen+r56i_k-2ZJtlp}s|4|=?avC~D;ScU#y*x-SU%gV;d(o2MD{T}4
z>af$j+WE@Ty^mUCmm3m2ecpJ?k)!Iz(~XZ>tg4x#EZbbUPtDp!+U+*2TyE+?xroxP
zW2M1X+SPl%BllWGZHQfdLga(;sIrnz$OoE%hRoCPp*9k)pgH;kIbrj+4RLZjEdJ$N
zbr0PUoHz3pcm!n-iWt20#OHTnT&^!^obRt|+~K<P!WsR|>#)cNPM{t<*jEmOi?F<f
z|F9_n0r<#uW9u9y3qXn-g!hckcRKWC;a337V}5l*+|P?26cE0oN0Rj6EIWZ}n73>;
z>uGn7V{uzvG_t+ZRr_2`ql>FFy%Zm^y6y!ajMXd*nqJ-e-an2gg^X&At<>$dy)z8{
zz%y#<pXdX@C%gp<Xnc|tiOYRi?IphNl^?Y^1%&q5xWKWegxg@Tx7wWNca0}et7slu
z;7%zh1NM<SO5~MgdW2}hzGj61fp)SGKI*_<vg<ed;ci(t`%$lkjjzLeQ|6n5^)JW>
z$3df?{VW>3Uh@y<uVk$F+TGc)Ku_1`N08`XonQt|<1G^@k(k`*gh>^DT-2Z)P7iLh
zGEWd=cZ^8IN88vfr{%cC7l-(cN1sk6@-usYy%5$hbADutyz<U|vv8!XU^EdkRB9yd
znB2f3oF~JGR`Lf}(hG0HS+p0p(V-Q$g?Hq@cBhry1y=c32<sdGhhI#%5g_uI!<DeO
z6OwPrk;9gB$y32o44AT&ZzrG_u4k;^A^eFiiypbl4nCYG?D#os`1vpUnP%)9-=%}s
zzaGJx!NEP!GCbmxczOc`SYneM1}+2!Z<Q_NVuBDPAJoak<i8jq4eT&Ni^>*OSHg^1
zS%W|zW(lT=)eR^>V4mpQVn@=A^9n2r2w4ay3<3fpE3Ob)hn(j<IQMaM)SZu(kbtvZ
zVaHe8$%V2?xv?E(1p<xBuku;ZsXq&#R(y@U3JC$5vJ%z|v<!Jbs3IPSAuDFQ)nVg=
z;>!#Wm{J@RWEP0@+>0wEyfExPb_~IA!@LLZfX~1{K*HgMz<aNJY^Y2ekpZ)Ze_3Ja
z_`p^D1RmL8rPv5H66{=G3Lm@%<yG0;#v)DX7KRkMgCT|SblnhEWdLthgek0zV8Q6e
z%0~%xvXj=D21V)&#UK6%#NZXgxE~TcgeUTEW>x|SM^Z5OSr}Bkfiq|(Xdska2BAOP
zDR00e8}e0LvK6$B4<62<VNrLKo4C|NxGR+MLriFt@F=XHS-S~`R)~q4n2x6sv?Tmo
zQzwxoX+ifJdN_u~xCWm}Yt~-Llft{r>C5ehJYv(^yY?hF5|BZiHW%$~{x;6A+RiK9
z0Vlo#uC1<H{P>u?$uw4iyZeazsj4GPHVPHSVFPA-XS-qLIN7g!`^_(Yz{X7vG_GXi
z7HLK_yEtrolQMR~hcQg>s6Xm1-ne1Qp%u;PWEWD3G^2jPHDJ&3!I#4XY=8SLH@U~x
zf=?upcN6K+kK0oC+5wzK1j6$~80=(zUTzuO953wg8Ez%d@eNSQQ@+bf?y@+q`R#NZ
z-=Svtp_Jgv*@ceqjuaGD-x19&h7n$~?}C4O<p<C3v(;}x70>W{J<e;e!Q;$dZ#!b=
z9pAQR6XEpmhrhqWG4P6=vIJgccvr5XeEPwn#~2{YLR|jBEr-wHu$CWQfC;$SIlJeN
zh*4nUz>mXIhjJt7aMWLh?=L;GyvQEJp3@3C#(}&A>>Q3?zO%cH^A5@76zu0x%jhG+
z=H<9{21dY%DESYE2oQM5;S5^<fKl3BsZ+uiSf^)yU5YOZ<pEa=?JkTBZ<Jltv{`MW
ztwtxk_3r!W$3OeG>HVMnoAkpEev#gO??-9((c7uf?rMDA@-7v|LU12ogcBoCHjN#o
zFZ{wKzrQU5{BJvre=1YT_UZDO9>1}2LlBE;z4?3<^L65USB7$$_&>wY4-|S<sLTT`
znpv~?k`H_q{YS`kykmC|J3!bSiPGqg&yUmj=-NAXxWDN7@-pv|VnTn#yI0`lX<x_@
z@T%}}?D*N++ezK+uHiEgOP=B9iLV3|X3KGqA0|wgijQ5}SoiJ|CgI^<W0L*Cd?@^>
zCt1ZuUZ5mXTFpi?b=vICP(D^4($#d5dcE`1A5GHv_{x*Um)CcmVCHTdc-r&xo|P6L
zjJ}hFW+=VL8~iYeFTPPAF|pfiw^OUtH4G-CnSk2eX{RTTc0EB$e2$1i$wj$FlemA1
z;~+8M6~3CX!E?Irh96V@8JLz=+09Z*%?{~wQoGwtjaDl)B-5cZtm|E=Zm4gQ+$3H%
zlrmM}^Ddp7oTZm9kJ8EMsd@fJgWXEpMV8W~%EAt{TV#nF$tOzlPH{*ZT!!4Y>hRP*
z@_;AVs4wCQS19y4otBkAC~JK5;XM=P;?Ge~2{y+`f=I$sw{dS5i{UuF65gGKYr?7O
zka{F|S!}0>`X;v_Lm)@_#0m`pC_vaX=2I>dmU!L?r@V)vNck6?pd>_j<8i#;mG*&M
zPsm{6#Yj*8YwMv{q&(!4BjGqQo?S#JL)po9aXB?zGLb*@9cJfgG8vghQ8gi#VW6YQ
zXkc7ycROjj(<SM|{YZS%j!-~y1SL2{8AW+$cgSWa25E&ToV?RYeyi+amL26u>_*#i
z9Mhh{AwPo~^A?m%Z^}k@V7^oa{JKsBCyo?kF2If6<12O)VML-#DDcPwCxQ^K^atQC
z7MvDbD&MLjpQfRBw8dC(DxfmHCV=*o>l$P*NlRRekGogB0Vj_8<X)}l+i077qdfFI
zR_a#*N9_2-?@IL+N?#VVu%nXEoH-44GoideUZ~f%g!^`Cced3g>**F{;?kOTVe#BZ
zu4Gpf3aXCU8b=uVbP}b7lAWDckR<Jre6P#Ty8+uB^@H3Cr~WC+$@Cx4{Ml(Q4F*H+
z3hnv6yMbwhqY*o5*C_2s??N){^4g~Q#qHEm{?f`u8cx+GO|H}F`6T_}qvt*jk-Pfn
zL%2`z#jBI_3Wb}V{`pX8Z>s;<asP04$BsMhRF54hBtqR_r|nWd!_m~o4~|67(~GHS
z^%4$9b<NJa4bK@sXASjp74<7969<aJafy>l(X8l@#YVAfaaHYk;>QYa)fsg7AV0Sw
zwH+qxIUbQ`S$W~Ww`N25Q=O{ntt5vKSv0V|zLY8(N?&<sRMt{U{dmpCGH#{1`X$Q1
zL%+x5KSxvYwp=ONGS3LztH<N6fBfi^eGHw(GU|Y`(pNxpRyHUP?oad=a1@s}F0#WT
z3asJVLVM8XcV@`Smo&7~uWO>6>&*xFpo7H-kM5K84X$XpgbA#ZIeq0k62&cB7W)Vw
zG75IE>#(Aa;Klh4nHcxd&tTktIeZxNIwg^>VIqakGwQ=gPi?#8(|KQA@q!uiAZo9i
z@N@xRAU77K+^G*a8(gHxwc6OS+Lma3pt1hr!_#y&x>6YwpIyc3!<fsXX!McvpX`8j
zza}1v<4W0i+3hr}{Ds~*7IAB<5^fs9H`W>F*HgP$OC5zFe5YMck9IZ2=*{#acki(S
zoNjz|mAO;(1FAR1@y!hu9W14tZqrBZvLmy*-PBk>pT#(od^9ol_OUx{73G-n8@tND
zP_7N?NZ2uL;O{!4F8%)vnr{}`zZ3&jKJR(c*xmK@+5TU~p!iz)>5qO4s}Ke%cH<P&
zpdY*Ah9MMI{EL&GGKs=hX~LPDoWZcfD<bgDmiX6C36Q^KA{Pp$m&o49M*<Od_L(DO
z(v8o>4D$0@7`<^)nCN!0|15#>p4$p+1{XRs8cu|m$G}!%@x7P@aOC@%{4U{T?5Xgx
zuqB7D8@%VP`Kth#!3YQYx(?@Qm7;<6l2(owRvFv$Q*z+W0Y_kC$-!O9Hv!_O9>uQN
z2~)mNb~;LSvwL~t?=Knp9Yv;dvakN{^xQMyFJ__pcrVq|LEeB%^=dH21oBEaGX$(I
zh-3L+0${=sMo<L#_&^Ud9Jj%JN|!!o9^sb!!bJI8Z!tiZ*C-&w*uu0b94n9py?*K)
z9;V*$iG)Xv7l5Iucy-@aOP#Gc`b~A}D0)~`)NHqW`~^o)z~~a*N-(AWp!`mTqjdc0
zRXTd{iFc4u4(_V5Vg@E$aIDZF%R&gWx2g~776s6Q1_0i2RsK1rJC0GgyAd`$whQMB
z*ecx4&bE(MJUKoxzQZ>ZJOqN-BO$T+snV>c{kPvrPu_jU!l+kI=<)H0nF;o)E8!4I
zF>|kc5Z|lxfTKjN#CRfD6UU6F$_Ar>VheomAh=NVijwms=1k#X`YAR$;$UJqVj)&~
zZmK*e`4A>pF?sD>awscf<=mC(J>nx&A)p{=aMS}wS#n%QLrgt_lvmIRXI^b5KcoQU
z7MMm_PaJQ-L_`Q(-j%0#4Di$eJL>3}7=T?@-~@asE;unT<(pM}J{m)Uo&{5t5t=}7
zLOSDE%Ik~kbbflC&d-MFYI5m2zPPxX!G5zYrdUs{UbP^*EM~zI3@RV=;CtNI1*Zn~
zaIZSj<DN}^F+k|8AVx1pHUUy|lrp;!CO({J%11!jY%mFq<QT*Z6K)RYFvUG=A)H2^
zfI&%1ID+C8G0R(j4o@249wuP>J3G$&72f@+UHDZvz71UUKp=VHYyYqmFmjTO4^YKt
zIL0r7pCc#@v{4G{Fhs#waSA*v*zuqJ%sj~t?}E9UM}ONJFaGsVhk6s;e#H`DZOu?h
z#BiH|nI-FQf5;a9kq<dJui0)7OtI&JL>~32!p>6U$3{7&Uz!2oH{11JvW*MI5U6|m
z8+dl(Q?{Hj$8qEWB%t2QUNGj6_9iq&K5_`(Foxsz+$C)B8!-dgqCHG;A#{OV;OU($
z5h@H#UFe7~`iXM!5#ea@yeS)b$G8k(bsDViehm-yiTH+>-Iob#eBhDsf!!vT$$6fa
zX_hD1v&+ne)sTr(;AT5M@$G-7hhU(b>d&zZmq6@REOp{v{RD(;fG4>5<5OYDvtc^j
za4&740I^R7Q^5V^_rHq>-;K#R`TAy|I5PUuShlJ$X=9_7DqFR*z4OR-bhTUT8tA0D
zc#G9_j?L1ztA`yOj+yb~KGyi46AQ@T62~P8sNj>D;_oj*RKSeqvy9UJ3sd;sd>(GY
zM7_dad>KPs{L|IPn<a_7&kYlJpKa(Hprud8!Qk>)?q6i!*(#fyU=ZzulNv6?W4$~)
zO`p6vNz>~G@mMWSLJ!Udsdw7*;<ixaY>4+|krb2J+`Cgr94p4Ijo>p^(iN+rr_-3Y
zMz*-Pw1TN99yoIFT=D}vFL|k1+0*a@H%h3{U|?kgd}5<jPj$&5>^zwY&g$xhc-Cra
zY^|qF@qV}W9L;{2Ml6)NyjFT;2g1FTc0y1(@1GeayeB-Td2gW&+`*e7A9;}+iU56s
zmYtn$>Td5@W`WKg#A83&ZKn_3eU_TyKarJ~@TyhW)$@=(`uJ7)(}yp_J2MG(pDvVF
zlrijN<2W)_g)^DY*a%;aGKpO{?4oMa^+=9uHEOB4vF7xvRVII9*VP3R%gT4NzLjnj
z_xbae>E)|KmzSfM-hTEdJ>J{)B1mw`&O4Nj1Mb|pyiM1V_t1eM_n~lLhZA)L4w#%{
zV$90P;aKEwm$o??P3~5liT>cv*H<2rLP-YU*f-%A<>ZxQ2FVIc{*T93mOD5e5R&Ei
z`=MkyFM1Ol!$Sgxz8ZcBT*qQX(Lb^QSk`!o!Uo0H1_}-RAzOL!eQni~{V1J~O_9Zr
z3%M8B3OB(8K4s#bVB<mIIl%<QS-sv6oi-$&tXkF~o)w=8!^AlgjmlXX2nW;^Cuxu$
zL^4_$_$NJ-hDkV@dbi?b+TYzyKYsrQzBh@TD(sX4Z7<Z8UK||z=s8Xc*%X~nZj>9~
z4p|FaP#26x)D7ub;Y3+Pt0)gC2X=4@o{*_xfgq=|#Br32dp?;%@>wXOnG0Z-SM0JQ
zo|S#_hvta`?%_K&H)yw8Zp+}V^pdyA3wWYEBFp(i71^!K)SHo%@_?J*<>Ku=OL?Fi
z&;zn1N;dpZB2&Lab3LOpRy_I&@`$2{`2!TOmfh9nnAAnyrQcc6*v}3o>ZDRxPrKVK
z?}U7~yG|APx73Hw-hg#|O`B6cbb%sVv^eNX=D$c=n=5X^KY07G@BbRApBY(+)fZi>
zk5aqX63ogPO5qjn41~{Ma-2589+4_af6>alm3VrTN7_1l56G=^ENHFfat<Z4kEUuj
zSCoahf`>F#yJa^QeOCW`lnxG$+-B=;^O9>TTj^T#7k3ZddXkzu{(k!YJL$c*_tVYg
zNOi1q*qx|)hTc&6a~va*1@-*=<#9SyAB*y+Az67}{TaIvp<yA&eK>t8C-9(jr41h+
zous4FGvV{W_=Jv7PPLoWba&11h$-#ubkq-Ry8W>u6~d%$O>^p3Dd%ZF+7NA+9dOG_
zcb+#28Wye44~izBd=@p3C-97-m!Xh)o9Lk&<dOLXcGc1Uk0zoE<qtvyKN$Kdc9W4O
z=#BcGPOs9GK38?ZZZ;HTL*bj_Euq;Bm7l)dyB!6N`vMuO?j9tez+bvtBm<3qw(+i+
zuW8&?XfD}&6<D;duVUOj1%K8?^;YTR!5l>J9?TEMB4qk!`UzEqyzVS?(7))V$GbK4
zO@D#(V|V%&ML@ZHuYTUAwHUr$GVGKD9yGh8G>J(*_?s)<^aZSWw6A6qmw5Drq=kbQ
zS7X!Le3bIZ4zBCmP8ml~c-@E|>6@08R?|Sd!==Uyjx_Zm2DPsv^$kbIJ&oHakHE#P
zXiK!Ga_y<#f9LUT(y;ACIP@>@C0_VV-z7R@Aq{;B{1mxlGXRT;P?pjD*et-wM(_pX
zle!>(l!?0Felg}ZPt`WTFJ)j#hPH8hawdKOrSF!NY3y91A7vhno#LJgR6F4Q(8~*s
z1HQK#{WI-nG`aFF+L79ekF*wQ+^@qI(wnF*8C$<roqVB*asTrT^%r&cz4>wmTK}pT
z`@jBQ{*OybKmYMhRnJ8<=2!2<T8AcRT-lFA+?j-zt8y{v48g`7)?%*gVF+_~2+rX$
zq}%}9DTu+&q2pH{^fA$4%zZG#rb`JZkyG9Voskor55|t!7H_-pZ^Bp1^7dcAj5PBL
zNBz8s_!5?P<N-@S$whN~zhxo5LlQ<-gH<62A3ixO?lQ)YT>P6om?Sdr*Fy8wBX?|`
zNYZ=eg^?W?M8z58I>$LL5f)E<ly98$1fH|On>-UXJ_DwssynfB2|W+PZ+u~7v!j0O
z7#$*F#J%2|4GDky+i62WAnXnuo{z>*aF|q9^l;Y~N*Io_F!LZr{aOqMj5GI8v6}`#
zfG`~E0CNoUO@auG4G9Z$pfIMKa}5)55A!RbZ_qza=f@{$;4`DS--~1H5ab%EF2SkF
zg~zP+fw59Y8|R60_f+g`LfAyHg%EKIlP~78e|D12j*in%ZxmCkx<ZM<DkVB<41+Cb
zp`73xi;reh-k}kW&>;Wh5#jSj46qx5%J0Enb_NNUuc4Sd46D&VF-#_%zL_2dbHQ8{
zBfS0iQF`>21YQk<tiW6n2Z-V;+Pz>3pG9zHASYianyhS681m;GQ}Td01a=4%3@kA~
zdj*~bKf*X}G*s6I49SDwUs<qcvS7)|v(5Dl=h2Ec6h}%Qrl0-;`0S7acNX3S3u*=)
z0-IOVDm;@J7EqKhFi?(C-(0@S8SCJkaB(CJ49`mB%giV3Rp_N`DmC>oYkPVW3iL!f
z5W@fyK=~IHhA>|<JX&B?S`rjS{ZSg6iGdv{FTx9BG0fJcn1EWVB6{L1e&JHMVMm_J
zpuZ&)U&<TxyoBOQf|6^OI_1f2((RZu<135`gR|SZzh}W*Bohf{A&|W(CQjIK%kgrt
z!cBj#rAAu%BnTm}^WnFQm4yYUP4uF`Fkd2<_?iXC>6LVHpbW3T%@K2(z(J-2)epgd
zkpJ!B@`P3I$cIfh%>3T`=50G=C3AWkcK+pEaI}C7f7@T*kr!KL?h(hfaQElJH&*=w
z9%c^lDxU%pn98Mb9<fA#QbvL<gz}fe5I%zqw`?N%%mQt{9!~B;sK-e?0X2RxK|2wl
zC{P?L;~?z%@%XY>jgd#ZVmBpe+e4;A<b_|j#dsSfbJ>Zfw_ONkdjt>foR`@zZw$W0
zlzad{pT(=#an?KF>7D)Yj^s0aD-ivM%L)#2r~sia9C#U){^)ca2cuuy@szy?K;gTP
zpsc`(P?76MTi<r%gTM~xyb>+D7%uoPu?g?|=%?_IbNJyNAO2EU1kk9%@W3l#xn6Uc
zGrtVWMzChE{gcwKfA`y&_jh0lD!!)4nHC06W^Gn$X{+8&dyk){r%&Ea&30G(LnCd~
zn1HV+w8j^Z!T+q3VF71821G>QRjZg8&E!zv%Wm;E1M#&CkiV$-)a+-OFN?OnEQkNH
zjcU+s=*!{~^G}D*H**<zpPM<QewKMN->-i!;-NtKbV=o-c7ei}Bj->muZbsQ5!6V0
zKa+`P{Yg3<T%?<O?y8A9csN#YG!&04KGmm2h{qz0F*^vvyRo9b({07hi)!pNU<bf-
z!h~(y33e;q9zK^m;nwT*d>oe-fvJhGD6Y}s{wcvw8k6xj9UUH8Syr!A(%#;7YIa)U
zOPL^#Bg58K9#T!bAK|Z%9j;KmjMHc|Qc>Yy;YXR+QW+(i49?H<ZZ9VLQT*izD8<EC
z7NKl{%>p(S@_7Yy6$Q+O<b{eAMH~tC?D0<e!8?zw%;31eb+N--bblmS=+&#!bftVU
zk<^oH0e{QnKJpKIE&MpUdl+LbnA8?e?}@R>rowIdh{a}g)02-J2gxe$jm=f_>nJr)
z^WWWY+QvmXI_Z02czb6%J>Ko6hVYLPg|z_4AWW_u9{19#!;^G*BVJqGBJvRRM_GB8
zoZ~(c?l(d~bSYYpxRyH2y5xcl?_?pLC?YtHtWjC^_wC(^C-d1M!h$sL-xvM3yvhg1
z#36tAXv*oOcaxwvp#Ec*(~9y4Ns9khJ)i_(Ap`Y69hBoUkxk&An50K(3LOp3`+DDc
zR~*MtVvq>EyGS&P{KgTP-fg9PS+S-12A?Q<fe+nZh=$qGbv_sv-%lRzTiJ<xN*V7}
zH%tsN+0H^wcDa%7R->+TD^^CGo}5_u@px}1z4c@-)$5`WrL$hCnl@e@oTQVJp7G;d
zSSka@Ml$hg#TR82oq;!$Nlg4xW|T|hZ8~9>C5o@uX^Dcs@|O&K28u7}0-29GM|Pu~
zA^$OnyS`D4aw|Wp>#N2Iiyo1`QRE_GfY;5<>P!x#Jt8l11Q&`o9>N2cHVfWpo0gT8
zZ+IXkn`u|(heUUbuSA$Fj8t6k&crWySzeK>CORN(A1Nw-a7{TlX6(_UJuA|2hxRYk
z?zszVqA*N8)0;rM?2zDosLjo_)UMTiR2gabh<%N}<7rInLdVFoYCLIk?ICp=mGpFf
z+saC45QXubWJQ!v)CJ1TP0_qS8=e5F9z?6^KiF-t#C$`pmrKz%h(uvWufvWhlv~Tt
zH1jdbtLc1nkzTwwG%cXC*--xr&qEu5{!qMKTuj{e#ICI^_tVe{?S<nwE9&od)F14$
zR6nbCX-nTMd}TLOquFxUqtg@huYLCkZShasJ<jez+6~HN=3r#?@op%RM(S(9F?|{K
zv$J#eqs%j~kQL=T6aAa(YJ=<;Q91W^J87?@HlTJ+Wb(;QE$WfHvdfII1O?N^#%gM}
zYH36DaesI1xd@b0+-Y{BK30yZJC1f;6KwU%P>$Tl@1^>1l&{PIaVI1M=J7=BoAyKA
zPFawlHwrdru<SbS_s{J=9IO8l&7;6*YkZli%^jchmF)+&<1N(#ieU<ZqHZi2yS`E1
zN#CT0c?X{uq<j>P5n3iZU(j@X%giiT6<YZTnYV-vpYryc2TwMiulRb?J=)vcJz&IF
z*zKzy%T9dD?mWqEWv=QYj?!bvP3*EQ;~aB79MegE2kr2n{GwlF$0!S-Ln#_@nFfOH
zQ5KEVPDdjaU9bQ_DawyAihhZ{VNG#R#BvOuX@Rz?%#k<xTPwkw=Zx)AkYdN=GKF7Q
za{+S@;ucTYlwafm@_D77Yts)4P##iMw8-(wm90uz(Rg=obgH&G5iUf}l1CV`4~~vC
zj-Dw_lpjTHqgf4wE=O|SUZ!qyOJjG(yR~Z7)zqqsZ&0}LSUkY>g~r!S^-o*o6TmYQ
zMqC%eysPKyP^_y>fHSV8p*}bPYp6C!d+4eSv6Bm0W&sJ;+<<$=s<Fo48}*6MJqm7)
zKsENjtKw4Mj9Zjt%8pmznd@<ASMW*7f!f7QIzAbw?XvmdI!*Mju)QqMWsZ@DK9xMm
z5^tNp`sbSx?z7D2$w<Zgxwx1w^8eRoUh9*-qzT2>Pk-uoQJpcM`-``HN;O9yehCw<
z@ugq)I^ke9d+W#btnen9Xz(^}xY1yXTXb%YB_fY^F%cV=9A75fa=7dTg8;+=D<>O9
zcDv{A$P>;vfAWiRk)ue>U$cDa&EyWza(Dx!?D{q|St&B!v;GSYV4FFU-QV_!5B<|A
zg<quWq-J6G(Elvn0tne6ulj-ymIsfsl*91mhkWHJ62Z=aa)4Q0^-;<`H`&dAax}^l
zpDHT={BO6SG*sU!2T#GPeAyE?bYXZJX8{!0bk;xt7ly%VXmw&R0}mRsDq@1_V$`{R
z>QYR}iUxlxk^)g;ia25;c6nK-6FeAY7(Jict$__D5*mQngPA3L7+2C}<zB32jNMal
zp)8Dm1Oo|390f7rtohzaD85vv<>fW6bZSUYt#5ZztJ5}9Nm)6j0s-?{3^VZ9{_9I_
z!54W_KF6bB>Yp5^!P!~5QF;ur>#GQ)>Uh1Qiw;;}^)N;nXJJvEm4+Bp{s3tvNiZ24
zZY7K=su`+#1{gjRrgRckGU)$Kg9G^H10@QOm&9B*w%ciY|8d%3hh4j228*<iM1c|a
zZK2#!d|I9A+sD6Xa6xE;K71@6bRsvlLL-EdQQR!3Dn7&nL*aYS2q$<qH`I&5sE!bF
zylPVM5H1m@fsH^Jim$+p5g<e9I7Sf05<8T^$|vyP6(9)sq8WC60S6%qKQ4m?&IVhg
zaaXzsW(X9}pvx`;EwRdtI8e6p=`uxKK#SxJLtG29DjW3g@=_0SH$LTtG7YBqS_Aw<
z4ET9(n8xQ9=}vC07lGvXmqusH0yauj-zg_XfSrjbJ(noH@&xzFD*vU2P<DC6kMcvk
zQorCL(4-%K1it(s4|~~Uw`}HpLb2s|IRgD7ZvKsC?T`UiWrRK)G8=_A)9d*Dv56S^
zlfQw0RC_2jg;$X>970B45iSJ=0c0oOqMPg`*F2Hiuw{fq2;1dm?`)r0GYkihIQGug
zut)fLYD$T|%kdPDEaMizK#YycCGQ;Fc1!_1H{gKW6Ih=HOa0=mZ#VJ9SlQ9<FxUt|
zo*d6W@R-LdVT!H9&e_%X>!~anido{9c*Wv6%pNm8nK#q&H*mmtMvC42V~^MH<p>6h
zU%ZSbxkVV26}#T?D@VhQ;22{2V+JCABMh---Zr+s;ovOH9OkA3v%fz0w#nZ%vWI_3
zYay@L?9UH@<qr{=7K)LFKlGz;2yNsKr+`%23np+=(sH`~ks-o7nC;Gs?GZjigYbT>
z9<sp=&q8^mUozuli=W)0j&UbaHo5K`IY+g#?Zpw7hx8A>{pX?h@;Il*yvt%)W=8(2
z60~cLRBv|D_RfBK{N$a~-rh@%R>zZdOvGvQa#$XNSty<BJZyRjZG5r$+)Te=9PaZ#
zDl=c4+UG{j&!02I{c`2`*J(bFuKsDvH<_10*NZS2D`wOnejM7E`8phZ5WR?gO`H6|
zqsdPZH^s+QH`h`HK2Q7}Ywi1!%k=!^Q9A37#s4cL8WZuHC?Md&yeo#qPRb9HRP;Bo
z08Vm%(yCTAQdcqvcN}#(aR(0*lI#*<@{q9#`GN@-@`$|A>z!E{;$1-E1=wkXoYCFx
zc$G4f+b^EKw4#cgNRRfqzQc>j^TG7mI{`RSYw7Mr{P1ns*sMyPNS?4nQNZFXCiCE*
z$uo>5%8%)EBG?n*ktw<`?w^nckavVrWD_e8;Ik!bkf!h2f`?N)_}+G_ox0st+TZP@
zo!wpIc9|8~mGzJ{By&u}yPoz3l2w*`ToNgwOrVyKvsSqqNH|7;fD(e!KS&n^)H_d~
zq)k=BxnwI<ly_G=+Ha;U$#P5zqwrx;x;MN?r{ay7ROHCVx1a2$t#!$17Z<MojY{mS
z>G%8TH^2RD8Vp8~%NokdigCq)xxKwz@zRpbR6fFkFZgD5n6c;%+3`wz_i6v!yXBtl
zchV2v*|(f#WtRAB-+LmOnOxm^_X_1?2M$Lf-ip5;4moD+I^C<hOjsi;AbVmcGl~&a
zgz$DJd|U`F_qTE1QS4MgDTo}!B26Y_$s<ak!THd1=6cpRfs7RfTp~L${xI1dJMU0j
zL4Q1|OwnA;N(1CjR?)LFjR{8|@5ltc@`bzu&9FF<MWH?h^FcI+0%ChRbvqp|9-NLR
z!sn`Y>XA<-QXfBgqH(R^-A3HE^vXMWt|hOmr%jG9T;KG4Y=eF;U0qBpd!t-wHCt9B
zke-iO6y35j4dQ_wE24LB5?uqd8!{n%1@ansq+LPV?lFX8=o)--bRun;wtxbeqvvP~
zD7mUzH6P^<9T1kCQ*E_vlskRVBD+4}p{}Us(O7ul9_~0I)H1Kq2j}!tMRo)iWIXOL
zW8z#bL3Khom{>;PLEfs>s&{lF8)BdvcFna!*Bldu9Dj4c-GC^|mYx3T*_jnQ+_{Ew
zhj|s0n(RE~C_ofy(87)SAx;EfAt|(j@~Bo>Nwv*YD|@-mi~IvO6pE5{m4C_$J)$h4
ze?l?M<ngNLnj`g?)TY0^7CxX^4D%#sy^-qr*mEweZZ~Z;T4}2MfAZouef05j(~@^l
zsm$QzlA|?M9?D3&qdt)Pz57Wtg3{bqM~L>3P?Sye1<Sr0Ypb&A-Jd87IJE^l(mx-I
zo@-l`^xoS~JQ1!&lDIyBog(x<)D;VmS=>P1hf;@hN9QBcyHD9s9_aho{X)A%p^g%e
zy5V65*Sg%cYN|W#2}Dr{E(LF(wuH!!xXPH*YE)8VYs2>rtqT9N&+F@Z_m$)aIwc=!
ztm$To(rxPg2UxU0>Y$^x2EN#3OFzsw)gMmM!NFnBx9S^ZHcAP0AaiFU<z%jk9hQWF
zroj;k8g|wI$H#usMxaTRfp&$G8U>l6ioPFQ9@;T89PDahVMoEEM_KX%0;xN}W3CQ=
zc0Yxj!J{B<v-+W3<nU?`*@`<nuHJT&T^?@5#EZzr3r9t+DbJqHAeus1F=0r|;gB}{
zIP);1P5m+tB-+{9^pSg8st?8u{OOIM2hUY0|Hd!SD4!DW{F*@Dhz>5)-}4DRX%Bb$
zxVXAYQ;o~)K1SI%nX;%L6x8(h+_A^F9rJX6it(-~+H5sBK9q5qwnY5P^q@L-U!w4g
z3y_A(2X2%Hc{o4D2zfM6@NS#jG%g6f=hgJAtyg>u`dEF)%d>&<dh0gKF2*Kvq-s<s
z_UNrpesaX;Mx~begQ<6L9~_^0ry#kf?`7U{r`t}C;Q?CeH#e8k#Yp|5a86sGOjWf_
z#)KO-8i+3GXE<JvF@oKa)FB%P%E?^5dP5~)F#>Ik1rW@eu_%c$vXgOrJ+2~&MG!1T
z4?YF{YCN3!7-Y)D{2qKkT+O0yrOVtayP!uJk6GmOp!oEco-Y%Qm{-MpEFNomp>#FA
zKd1q0tciD#2*!Nr_>vvuEF_6t+4TLiBk#Wavj+dq3zO^YbIlj||2f#-n=fmA^?&|9
zEG_-wXFsFQ5z5`|821)N;y)XZO{c<}1{TP7M?@g~5Iei#pwM#wxw*!#2-}K{;+c)V
zW{tUg`^zR$4ojpN!QC&&V&;clL!D*Gmvcp0v|{2IeniI4U37uRx4p<8rNJfV)Be12
zDV=D>NBoDuIfun9LWjc~AGrA|`^d!^&L0>7G4u?q{020>a^PZBULwc&)3)qM-3Wg8
z%8h@0h%X2i$`ZDCxolW+O8y~C`18AnlY);ZXa3cX2O;ATfB)NfcV7-`u=;WR;fVp3
zK42QX3rXL15^^f7dTMf(z9gkfF{miOPyzspL2uP7D>NX82I#PXNvFxSUc*#JcqQf)
z1_b62CIf#uK^Os6)WIl_E(`=6F-i%}!yc<c9*c<^_j;*!a+<EitYDUw#9XQauD4pL
zxzkNMkM_*u!8G59v8hxm?x;s%zBeVXup-k8rBJh`&YugGd#5MqQViFf7*Oh*a4=+C
zFzgk4Dl3B`<LH9~5w_N$xMDSu85wXZzYuC+LSUc}5EwL64h?QrH0bf4an3(>8U`k>
znvjwFnbuW?rUcz5?>tNGoozB}22Rxxc>$j=_WUUy+!Mg6F9sHvGIrnHi^$kcZ(<X4
zj?#dUp?(mG$Wxrn?sZSj69eOoV6iI?4glpNt8riiP<*i}&PN@nT%5BlE;#Z+zV%p1
zrayLBnL*R=%@G3@>Qs*mq^uNSpyi0eDKKPUOkf;fb}63~j;ep?118)%&6JLX9{G7<
zSwy-F`h*k5ReHFQ7K{voH9|h+@WhAm61(3ZZgMQf0=$bZFU2^G#fT3i+)u=S+=y8L
zPfLPRsja7)m<t3U1m8_o$cpwU!^Y}{m5^MV4jhRJe!SB{NjQ6==M^f|C?;L<?&NL2
zn?JK7zqnxvV~T^_+%V;)HxuB{tCbUS0|&(fB;wFEgU04=#NoU27#MviKJfSrKjPxx
zz&S4t{F<4=z)pXIDK5Mt(qg!o5zG>vxFxvLw};@#;<zd&Wy!_QCZH0R{?yJHX2cKU
z@N#z=#%BTIkOYwl|B@p+&+=EyLYxu|E4U$lu5r8yB&VoAvgPz}qfF2)_xWXS`_178
z3{3n2KYx^W`S3rE7`f%L;YNMVYQF@LDMgRG&mcK&<Z~Vdcheb;GPAH4EWmIg`f%FV
z1BRWJ11osZx6gB=JSuUIoC-dN3yk>k_x#7kKKSP?)cD|rAte;gui+!x7sFxkcMeAQ
z*ZgH>-?CtL2<MB18hZ2C@FqmSE2tp6@tYHi#V&*^?F*BI@SG8TjsN(cV`>2(^N=}w
zd52TrvUtow6q7yUUtduNJd2_5<2U>jPPmSK2Im}K_J8=zZx*9{+a|L0nTDJ~^Uk2Z
z8Aw+(7S`$w3Es8TXtvYSXYZ!{M^EL}Qa@go&{<J+pzO++HlF9kc**q~wEOLuH<JF8
zDfRo+4H3)pS%!qm^S${z<39d#aroVu$jdilEVsQ04}LWp=)TaVF)+h}`h?ukuqv2|
zEO@YW@t$4ojA%CGzm!gTgY@yiX?l5jo(7|vbj{i6j3|1+_qlng`Q4hAWZrbu3&J>x
zoFn(t`J}atHSfl1)hmf3@EVP+^yu+pw_PT_IA-cx{97Ldka*LWG{5wOIg`BfvE;3x
zeC<ev;n-@9D?B_rbQ~tfJFTXV@Z0Iu%$FXW4wa`njng;j{Ol}UPA)uw&OWH%e`6sT
zyTF(%rVuECV&PT3cM6_YVHqzdYb-vAz6M1L-%KVE9=_duRb|tBE=msW{Mg&yOPn^q
zDH=?)OUx4Xg0H^2zVal~v1Ax_{8ZGZwc4FjYqa7t4waiU#-sD}hkyKiIypH`^=2z=
z@9n0Yot^YVuse-fIv<SEaD3@qVEa1_Pq4AWO&Jj`uF|3E?&NHgIKhK^Ja*f4Vc{;F
zfo~B<yS;103tZ`-cbbllj$IEN702W|$C^EPx}Sdf;}3j<oR1PjMvyE9p4e%G%mGe9
zk$IcAPpw{EOD9K%>FDsl@&l7#EM8dMl$^M-mb#s8stR8$u!IIU(hquJhYJ(vDAHm#
z027+x4)jnz(1aDLR%UVEPt9_Wk0X=}#C^n!*LBG|+&jjRO%vtQM@BOF=Y?@G@yvn=
zPAehr$biVIO!#s9846X7Y-fj1OY$w_5VFWXv^p9OO$$ssmPLBBHRy%%Fo8*276&57
z?n)lysJWZVYwwg`(!JmBS((;o*3;hOeJgWO7M-4POcshWCYd+VR=sHj&e7p3(E@i`
zUpp<|?WG6WBws8NMScqT>Rx^*%}_ANPi04$L|7(W!4+;rCbN7de+=!3^3dk&r!sMz
z9UTcf5Kw+?iPkBX`wW$(tv20lx9D)Jzzn6OXjig1%4^C=-n=7CX<9zi!}qo7gq?O?
zjH&q1ry(ahJIXV=l{jKivZu<!DKYHoVYd@XBb0N<nJC0Lb%r)Wx5Q367R5j>htT1{
zK^ln;M4pCC-q|^dqJ(zHPKeQ1?L_?tv{4gXG}UKurw$4g%FFyi47KpS!n4vpJUmJt
zee~RP-)uF~<HwKEyKg;FeX4F024ztkm&tsI>QuO;9JJZflb&$}PPzB*;}@^e%R{w4
z;f~2^j&r3xrt$-i?5Oo_G!=j{vw)Ml!-lf3m3|0C9t)qTyH2~NesSN*cII8K1c%+M
z%ni`LktVR7?YGmD$9t)@y`AdqwyJu?b>>rGln&zqWv0L9ItCQ}?!V;bUBSW+$1T!^
znP2hlT9iH-7mO|$`V`MG$PEPs{60I_s2>(BbDUsX_<p?CO^sUYc7a||7!Ezr&QXPe
z2CdS@)rWcE0yHhS>^8)YJM*A56koOaminA}IzBsBzjl(^z87)J{8*fFqrOVt93M+L
z86TlZXai+F^e|NaO1kWHTTutjA`0f9!1ZX1Vt3-UR;|?CUx5d91W||F<u+EntZ){T
zP?U#4m$4Rw*p2dVKAebtGzy_`RQjA?#KKP$!Yiw>%hkvyCw?(s5{odHTZ*>J0%z`8
z)3@3x-_$!w6rXUQZ*ax9$f6VPgcBWb2NgS2sWXlVeYm?x9GADd-PU-vo*Eig=mY7G
zP)ea3R-&qd2iGHh^n+_c3T49kN_{iCCr4a8p?J_Sryel|&<==CnN)m@&0LwG{!%nN
zL7}C5QLpq5)9H<Ig%Wy;bU4>nbdA6B`Vd!P(08*C2IUuuJd|waON2{$1Lkq`ue>d*
zy|FV*Sx{5<ZZeHU&^bq4vZIN7AByi8UO)K0K6Xy-cQv*v550araXfKV^uiIP^a9Wp
z%IfE@POZRX{)>g;&{kD6)2eNyz3pb&-)^KwyDcxqVQ~n%=(t;}UfWP!*$t|;sXVa2
zpDRe<nYaUwE{6r_%zZJ(=60xG^Cg^s<i0ZcX^)?(N5)L}2imw77AgFlcp+#Pc~t0A
zl8hs<aAtRV0G2>$zsG$Li$Yk?@E|#yom~@+R~L+z!cRr{S`q#!7rVdU-B<|KI~T7d
z`rhd@(+}R>Pum&?j{C#(;?<d8+-VG}c)VtD2IEwb$v)eVQ=YGCfcfX(V!kfy@6DpI
z;_GjI{)>mjNlTkt2M`$jPyRH@%Nz5i8^S*}?4nD%L@!rj$VN|L6AXe^X6wuG`JVX&
zBwQRK&eDsJB~#frmiz#1XW$S&cb@U!7LoF+IM3c#)z>KRktW}UEt4TQP^3c$=gdBY
zh)JVxDdD3WIOE6nVjPUV<Yx%;==2KuDKNt+U}jT36oH){AlxxBe&=CdhyA9nO|jZD
z;~cmp^8_GFL=HeQ0i}e2K_af)Tuyv~e(*2wz%6+to6c&E!w;QPOwi@RXmFuE9yV4}
zRh@TT1KE-~?`sVfFe(h<UP+@)7kXo0vH+t&AA<mUBc_zWf^^MHhyl5f0A-~Y9j)*j
z=ViA9v*I!!U@_3%2{wZfa9|+LkB?KIZNS$OI+Zaxa#l&!By=^zaCIciZPjZ^FKGbL
z;39#M!EY$0A0aCSbqxSQ2!r08I_1e|Bqrg?3@ED$VOlT5jIhfL;R7MplaoRV10RDN
zj3R?4=|dk3dMN3<QbhJ^4eob(SoPph+jx$_#ljbE#%!RO-<YPBO(W9^%(-y=<h^&&
z&ZGUbj1bB|EZp!%n?<oj4Cg^WSUqE<g@U1Ay!KXqF)a*6Fe@sXQ}O6+b~=O+yD27R
z65LpkmOL@Q#mgwzSY3sXKsY8ds5b^h+&LP8m4d7YSrcx&B2x7aV@nyyBh2fS7+UaS
zCPU??otPe>G4QLf#D#(7EMgcLXr}C7BX0;czyJrFGi>39e5hVoC8r=}>d6a+V=GAy
zLB{bEfOfztBb2SId_nl7tdqgm_k4}dv*JsC(9A|nj2ZVki1Fgs1Xhv3P<xdelcuy4
zI4&i*l2+3)-(cEx40*G9O?lvhGV71Ji@u$=ot@sy;6RyWDEqw6CLGWg3=N!rzxWX(
zYk>lGuq=y4{x}g2zX%$i0>W!%3|n-r;$@JLzKRH)*^~%)EEt>}J;#}u9Nzx?K?_dT
zko}JE3xLDxQ)wEu{USiXv~{6O*;{6YM!UraAPMHO=Qt&N!JTz15lGKG&@8w5jJ!qO
z?dqtG6jkYH{MaQN{L&Bq_OhqLEoFq`;xP#0m)roD;gs=vRye}QTB6JxOpiRuUG^)u
zUc~yRO-^G552Fvc71|_e{6iO|ADX~x5$qAh1tkFX{3$EJqfX}Wg6@GBxW?|h;_Ni^
z_$TsD9XNIYAus0#*mev^yk8sFDMQXkIAsrq_iJWKw1TsF_(j+Y?S(f%7GU5&8GtaW
z7G9>oNE5#Z7%vRthu?Z|iWDPc(63y=V`g&HnJHe2!GK%JV;{d}rLj|~Cw^To`v3di
z{_|0M#iKrbZ6np2-PCMvr|sSS^yJxFsk^=F$rV;Lv-pLHSC?1=9r$<qoSFHD@j_=O
z^C=!(h72wC&3}n;z5N*j{=c4aF7xwQhJ?%Wz4<&-`l5dh=9{9q&oXZ33-B}f=W%)w
z1L_&%$poEuy@;Pec~IY~Sa|~<e?FR~SI7Nya60g0kXYn&aluY1CedTk&AWZz#l)+{
zE{q5B-jnfEX^VdoO!zJ)tsZWt;@8j9{@!l-+kgKbd_2|3Y2Wuwoa%v(!!Rx(Ym~`x
z%EPf+ole_tR(kh)ee?Pp(Z=L<x6?|UR!uxD_sQP)xH9<v@lfy&j>O+9zborjDA8xY
zi#lJDF-VV{I2@ZuKf%r><OU|EP^6f@7Op5Se6!^&l^YmLXw&mjc1}uQa;Vd8tFKy1
z9DxU4yZ>m<{5MC>v2cu?Ybd@rs{QojD7||5Jbm)<hvuJu`2Gj!o%eo_THPJv^5p0s
zy?p**`Y->V|H<F){qX(t?tAa2cBhlxc}MbvaKa7)c8n3fT2+}O?=_X~#@ez^*XR$1
zzQ2c^Og@snRuP|l;iD?YS9cogBp*u7U{Z~RgeVBuiS@ycevsOd%}_9&pZC(g`T381
zv>ZFNRN><7IcD!FjfCT^t-A5bD*j59<LH(}D-*#xP3I`n4qv2?Km4P^qcmx>+UebQ
zf9P^DS?E>%EILyj5W8;R1w@ySGWE$MJoU}Q8H?h)2u^=Y6hjjnTZi0=yvo8rXk=}5
zNpvBascv~vkljV3&(Taw%(AG`?pLf_zp!!vdDTbW%Mbct*W8_CF(&(2l!@HT4oB`z
zW(STZRyEF`7(fZJxw)=tVsRqMFXcsT1x}RvuXnoJma)JsJHb#2^}SPT>V<ioPB-oC
z@28e<!vf56)f*fGyPe4Q8AsEdohpAEv3Bb=w^glMnaE_Rm8A;LBqTC8^lD{|!lCHB
zRar@&c;pRP3powET4^YMlr$*vHoeFc{Il?oaTHlE7RNFf&o^zKegXv_yArxSBCaDp
zCf9+p;&}p&Kqb8bgZMs<Q|Vz)e1T6-e#;HTiI1YBFfz>PqsU=Gmv%-!g=_~6aKxq}
zSb>Q`gPl4)K2mJ}`IbAdwuC?KuNhB9%ExWcmuL!wi|?gVK9MPD+uk8|a^i_y%fbw+
z$^iWk?HicBN<if%i`)~<E-dcdBJU_MAMNj^r;qm?9|isellh`=_!UmLfnJeoZ`3DE
zC!r7oK1ZsdsAfKfd$cN*4Y@OSGI9OU2QiV2vVWqokTrH<z!$ks(>JHFFo8`SL%ZDf
z_4Lthda}Q5B?db)P!OT$U?(ODcNW2_#@zmCn>(UG79>+2C`Z7v7mA_)laN52LK`fm
zjon+ZOAnlpKa^IF_O^u&%602qG0Y`$#~(WaY0KQ7SKHcf_~YJr8mhkrN9(Ew?_Awc
ze=a!dE22MW_3~PM0?H@RC4GcwhB*Q6u2NmJ1rHhjg83=6OOCPSgqL==ZM?9<2-<0=
zT%j<TpYnr3mE#W~H`+Zk@0~>oMBUO(!8Pz{AMCKCu36APTW9Aa?IPyRVh0{GCqk2G
zGwzr}vCDi6%1n;Hg+`gf@eWM^pIqFeD>Vm{h2V!YfQ?K*pTfPCD4cv0B|RVQ2*AQC
zX`=X|&RJjqo`zH5UihhQq7-Ej^mS@&t*6}<J2{i;P<@p0M1S2Ij#6KAP*uOp&OGMR
z>KaG3HJ+}izNvHQk2Xo)sRrP3DSKYbQP+5>Pql{_?@14uo=&gRp?C*&+tDuAmFa$4
zfl-pt<~aJ5vZ8E*rY;qa<A`ZH<bOP3u2H;5@635avBqw@Ewv#O?s8K8G&ZYWV=j}o
z=Xsdh6n?mJg*H!nWuZ9zhVRr=n*<KWYhGy#?4OU*jd00Wcp`dyesC<_<UH-|?WDK&
z*-75@ydFoBUfo&cYM$s=VGhl^aD+0aWyL}awG#EGC}&wnu5>gOE4kF<7{i9f5q-C7
z8)>&wQy;E=LhTWHgul9{uT#4Y1r}o?<CVr3)&E+)OOU<_`h=ID{n7t2+so-Jj4dJW
zpg4_t16f!?e}w{(MJXF<?B1;n&eaFO`&?gVrL}m6JJHa1tZ`RFye^)BF_MKF-JQ1T
z<J#lo((;;jM1Jz}Aocn~)4~tmeUkPy4xms!A54VHaT<;_=Blp&#v*U}*#`eF3iDZp
z^78yS^L1f=Zx)ReU;oYD{2NuJ5H5T|YX$YVg*~PtAVi~FjmtDSGZE6EaL1p3*!4gs
z_V+{v#A+xa@)+nCa2(CP7z7;I!!a9tnJgzfDP>RlE3AOoC5Ml^I<)@7k%_#}orPQE
zE6i*)aekGke#6AP;e;s@M8#yV*z~6l*`x7BEFbX@-Y%Ove}~my4(t2_hnG1Pz2$~8
zkpj}(&365QL;K?{Yq%Hl`d2Y?xc-#!R~YKBxMvH$3n6TZo81fY<w#DQr~F+aMqT5F
zLol6mfVg=wRRN<6*<ZhRvM*^cL=jVEyYhwxHVtGj(=e!)5)c@qfXzU%crLgV6=rgR
z;pK>hxCoRZGQ^N59`uMnup_2@t6uj2%nBGY$YOl3v%Q|nr~9X8X{12~VIX!9DxV5h
z(Lm7<gHmHPju=;%SQuv%Y1FTeS`b|F#>x`P=F$o7JuAhO50q?|lQ_5kTAePPI7}yu
zEleflgE8@TUh*b*K*6ABK$xXG2owlIC{%pkm&!+|`<qZIQC<d<LV-bprXn0Yu&yB3
z!2G@UPI~n0sRyxp4TLaM#P<q1{duJzbgg_@fu#80oAb0;onc|w!i>^Xq8kj{76yg;
zdsb4xJS@9xhO7Kpa1ksn7oSeYrjd%^z!Z746+(%^383qnx@UzV^yA49g`o{pH&~Up
zW;}U!mj*_JPVUJfFBlkcm>dKd24k2N@B;G)^Nx}cSmet%1is$ng+Fl2H$pc|4}vL7
zIp7g4VpXZ~t#pA25R|hBK5ME&5C+VfO9=<(!_+(Or*U5lj06h$v^CNC+NPKSJ)0F6
zA_>;YC-h34Ahftw5hEl1T8xEggHBHA3vKpu9m6sBJL=lCt#7Y<!h>HbZ_&sr+N<IK
zZ|<B8KH28>8U~PtMSQ*RA^zewKV=*!20!<;AY1|R%O-*bg%((!HW4A`IDhe1g6H21
zpeP#nnwyAaQ#>M?&0;=^dnrZ+XZP@TZ!Z%8$QOtKHMpRpK>@j#4)JF2um>Q+c0hj<
zMs9YL0}*&-n)YW7mjNhdD5+7Tg$rY8xDjiP>%bT9z#|6FOdN(C?j(;Z!xU*69`F2Z
zki`MJto)GgP`c_<PWCMpf-6%ln!+wPXJahNWFG=5oQq!KW0ig8B#Yk!*Bev(Ff;os
zLUB;}u&d7EQ+ZzSE9l2yb$(?jSPSm}Sb}C$MBE(Nz6Km|$~y;@MK)mT&oIJObr*Gr
zvHODOYwoWBd&6Wu`q+Zmz-d^Ezkl>!Xw2sI!uT5(+7&L|QK)zDYXPdz8>WPf27#+x
z?4KcrT^|`e_Ap(_fjs~HZ+=~Zeix?XLZ&23iB{68#u_F<o6SyYwzkuwr|+cw$4@<>
zM?cMq@8IX6FdnP;QXfHCXj2%*#YKaEF~Li|Zh-M-Kl*h|sm{+crJBFGaW3=oStjFU
zhSTrOXPKzyuU!e>sQD&nuD~d?UyR3wSy|OKvK&+vsJOe}p}!F?w8^dzjxJgk?;{`h
zQAlab1ncn9$K;KtH;H?q*vY{pIlLG=9K5Ml#*24@SA>VV@X=@!^Pn8P#gTBVIKR3W
zr<0>k($01#{p4pq6A#u*&tD$;4i)Z_V6q-Q?^e7qaeX|F>?oe#D>*_8B^6}@J`-_X
z8Lhm-dosbx>SreUm?T88advu|jt^d02?xtd!8k&kg;h)z)|+)tE-~@OPKq+Q%rSzP
z;qcr$bda&=KZ;yb417F_FEEOX<gs0NM$v*2gZ_x)!MC@!)6VXWS8bDrYOUhMf=vGR
zPLI<c|M2_t;^Pm++p{A;d8$;??xV+^Q2pJn|6YCkm5+L2(jB>N@A2dG<DdL2z5o6P
zinC+|2@_Su$EJ9F<zrPT)vFs`xsUww{Ns<)-~Yp}CEwhn4}S1Ysw<t*)qOgj-lmiO
zP$f{gRbLzjcl7F|zV6f0$9w6C(qh7p3Bp!`348J7@ZajQ*_DOThMgaKdwZU=efjDj
zaUT(V_TkZqWj%Hdu;Tq{I!u51<3Faq`@8>|e)GHEr@co{(%<~@Z+*Xzh)AeGHo5n%
z5tO1n>QCu#G6^(6Js^X)9)(NDoQX|j6qKLTJM_i`dXL>y7n9WOG}GSBwpY+Yt1Mc)
zx+RTE$)pqCA$NFqkj~C}R(iy}C?T5_#mB8rI=Qx7iee#@KTI6Q-E`0Z^b3u!3xtVB
z+8A--bPn#&URND9QoGsIyRPCcSuqOj_XmSC5Z)Et3bIH0kJ9$;uH~P&E1UMlLcv?n
z^2oahz&mYWL->iET^nxC7!($SL2hFR^eG~OMW#6MgNc6QLh!unM=~?JGvhd{nE0pt
zvFMUE9rB*~54|W83ni}=o|8(D|Da>ypzN({%x34KWm=^{omn1Ky@I#66Ko^!s=CxG
zaqlv+-&kcsX$~U57wwvQCLHxbT5e+==gE`Cb>$B}q3h!$+e)8ZVbtB4l>o@h+)smI
z5gJCOW}=+>1Gmrt$9DNBOoavh$;pZOO7&lAhseg~{l1mI&<YAo`V3A_@eVZgdsdtZ
zC*X$umV3ldl)_^LN20z#<=_rL7W2|?LN_QMs1L4)ppH0Ba839zZ3z~NW9oz5qs*gF
z7t{-FnB9ovi+-GZtgVHz9=PD0qt0LsuS6&5{#JBXOOGGzsSVfZ{N3jx2cuwv&Q%3g
zKC-(L{Bk$BX;^iK(h*#sG)I}njzkteK*!Dcmfg<IPJN^#_5IGXN9h+o{UB|3P{>LK
zSDujqMjAs-P8C=EL{)tx{T<4U%Uk9*P(*ROtz>xhyBxP$S6{v^MtD<g#R~gYOK<KU
zzqb;Sapz2Z7m5My<YFEHI!5t<->vA=JE4>gN5*>B-ukL1`JpM=vfGgQGh}?`U6{!C
zq6pCs^vZqg%n4D)j4zxNM7^Rkf(zw@A^K&?O?!&@m}T#xVY0uj_Rh{A=>M#LE|^QA
zQFe!^O+p8%8|GjxZ`9^EidVP^UJ->8b5@d6lsEbYc4cy8EMuFG307WLb6fE}o60XZ
zJUu(txIOVVw2B+_z0;|sXS>3a$`<{>z3_FJhN7>j(q)GlyGhx3$4=sUMYtOFJ;w%~
znA2gLSXVx2-_#v=;)qIL{~?UHKTvw$i?W>d28knX_a5!KFD4ATu^6L>!ab49`EWch
zJMWOO?^RyfwUSK3^HDlGK1+SIJrwI(^sn?EB3RlS_kgk38U;B0ANfQ1M$1DHH6E*<
zSNWLx>$EwVcFT%EluR2N;$0M<K45w&3RN4rTUtv;r-Sqlzy59d!yo^U-g@#V{Zh1y
z0_;LGIu^{)#dUgh)JuQ*NaMyib8Pfe>IYPA7Wu=gbW~PwL;s7CkNds2_l{$II~qsX
zLAKke8^??<oLYp^PGq9GrC!vB(?4CPt*Z<uz8J%{G^PU=`atmsKf{zLiz!;_TY?Yb
z<f}C+S(!6MS@+`Aq5JuU!f*_6r&(3`;onF<l*{+GjD4J(MW212KK%GqIyyN^k2zAA
z{z5dnyVF%)CAv`FF0WK3;e0fjs@>dqQ4Gp;;*+<j^8TBD{5YMQaet|5hW13I{#k?n
z7lrvO6ZH4z&DVwfy;(F?eEqw>`DI`doZH5agbI2p8@=t9Ctmzn{KD8htYHWv2P?1?
zFJfW>_B=EJNlVu374hOdizY*uaEidZ`Q~AIcMd==8~oo)Ie(ESSxY>J;y;7H%<@1?
z@;LLrk2u9Y8++@AhOuBbqvI5mnmgBs=lFUI$seU;eCRWW#Sk~=GMgxS`L&B)5i?wz
zhAn!s$IL&6b^h(oPZ-Kpz9K4ED!G~){CLb$Dh{s*iM#&PVa6nOhAEGdjN$5oNii5J
z{G$9h5bm-vh<F`cbu+d1#QdwnePFfWO00eW4y%k{q&QyAD=6d+0}uWHt5M1W3kEGe
zTYblzVXHwfR=YiT6&y!zu!@JgvT_rD7=bHRBdSw{aT}kDNfASiFv$Q(ReSK%pjYLn
zL>Q@d+X6G0E5c-12}KF5xKdsjw3UDgiSp|v-DvQ<)If;feJy4l#++mR=)e(n7-(TK
z8R$qim}hVf)`e{0EmrA)vDnhhx#nITCUXZI<%4{IR~Q}Br2eAz<N<>vMv`4cOJWB0
z-hP%Izx7OvxddrYsDv5xeO9mRFyQ3J_3vcJjl6_=>K}s;2BXDjR@2}E?Xh|XI2%d}
zCICmu%D}s%ft6iK<kc${$tPJ82&e^y#R86m6(9Mix`io#@nc0C3QBOk)7kM#BN#jw
zlks$%mIR9<CyWExVYFF^NK^AFJi$N+B)>3qUYSCfm7apg2xBTQ$~!Cnlqbq>CRL0a
zN6jESl)Jc0qY6a*<L^~|xVsKjE-^5m{`TrFjWxKRi5VI6#fan;imPI5)-_0PYB1)w
z9kywsNMrW^0wek3Xf^%n5$p%rgzD39v618ao7<K9WV_{BH(tYzjc<GV6_#*{IQ$0X
zMX&<HAA<em7y|{);ec2s$4h$gi(k0<S3jO8$M{8Pg~)*cXKOBJaRbKd<tU~=FDBXn
z-{rTM5<4Kh2|;fqWY-x${ltF`;a<+!D!9W03bG$onF8X=yfHKT44-l<p2aAT-Sxco
z&f$#1;*nq3%NBzsu*(^~#h<q6zFdE_&v;`LhKR4hc3>=@G7Ht_@FhyYdSGepJ~IwG
z$7<AK*h=`BaehkdfIEj5Dd7-h(&IAQGfW0pz&Ou15ylZBup!eHu;IJ_$L`o27rT4o
znSV(fxP+i=;CCKCMqvo{X7&hA+}H9Q8OR?puYoGK`ZGoWa+V@4CWG|n0)RO_<XEZ%
z2Ro`P%8TRh30(aVHedtPJ`RsH^rnuS{jdu#u=MwLzy5~;?z=F~0b|_K^1Axs4e#pe
zboWxLyPfv-pXzxm!L;c~4TM-fr7)D$`Ns+_!iwK#6#^ax<FWHQFthx8j-h(Ko*`m+
zKFg4BdA>KFXJTCa+LiE?O{DkD8~Q&?f&Zy=M2Hb0I*_1$+Kw1X@n;Wu7-cw0blJRg
zD6hO2>`uHdlUriwSS>vmU-@pRiQe$^-i?3DBv|l!R(y$fV{#2nnAP#qi?R5&rPOAp
zjrjSkjrFvoI3IoZ`@~%#zxX$Qlh!I*>E)}VbaplrAIn4{Clp-xf#+u;7JiQjG_OG3
zWLE<_J+eHI-Qf?hv*QH$g;nZ%J9WtqHQz6B$mR1K8z&irNkSi6C^Uc(CZU<2V&d_<
zf9{<EOn3sH9bfRN@WPkM6Zo`DB&Hm>C??Kh0h<+H3P8RIW4T5J;uHg~*cLv))!w7M
zw7<9KNmLdOQcEZS*hN4(r>DoBEZ^ohdZo=o<#2GGM*Y)-!s*?2pT%)c%G;wSPtt~P
z$O1MLJ%nZF8@TWUtP05SX_fVrw9{^QrTp`cK2HDh|M$P7^YecC`H$X9zx>6IQngx7
z<LkTh;$)BxRTu20tv71vO1%Ez!9f~IUgLs(R(5kA3HOyT@w+9wagPW%VD}MbyW4fz
zzx~}m`oe!E)R>%yzkgdY&^vEE^-i;E(Z_Fp`<wLt{-6G5PsF|b>~Z?Zk3R5GV=t9v
zUo@0fH?2J5h`(-U+qASHd5RrbUUVti1b=0h*;(&g@?q>qK-NP(<*ps>K&eZXWY-!;
zd{GaaFv7S(I20ey^l*qw%g&Ce6-u0HVg-WhT!ni;DI0PP$_G#2sd<5aCZ%a>$aK^_
z{x#*x6Z^tDN+@Ux8Jk6$9Lf3a+ixlFn^ssNXCm94_4;Wz8v33W+GDfTl1#mu;YhYp
zTSNB6lu2dA!)Svjhi+Xy+671AajevcV-Y0-QWoz-S&t*>8ud7)5ZWtB7zzr_kY8}Z
zq%v_(D1rkf#8D7KtN2}CajYD?RVd5ug=-oDN)D8_HPszy(Pp4a<V5akqisVw;kTr=
zD?CdkL#{)~Vg;PwVJx31KJoB#TbEziNrb#h|3@Fj@rV2S`{~)!ebHi5?SjdDc3HB6
zMfmd3c=w)|-P=~*q;jzf==ijkPW!`j)}L6RfU*q56T3E0q^^s8IfCq?k3UYw$H%7e
z$4?%oW|Lzbsr$ei@LAxC6C`uZ9Eb3#=A(xlGVD@gLS8slS?EKdNfaj>eaSI%$n@;)
z>h%Y%zv*~nMIF~jP)8^!!4pR_BD13mV8<Zs9~^)i)0x7wTij8l_NI1${LcJ<dp`GD
zq6aaDEw%M$PoJrMzNNM;oGBi4LS0f0=7T5zxTn2vj34(Gy*xNdhsUR;6$KUz$Bq$>
zGo>$X33osF@dxQ|e)gmE<neamxIHU_l|09yvBQg<LvDXe_zN#>^(oID@3_8BRIX2+
zze+D(9H{M#Q&Vl_ttY$b?I(LyTtVZs3y%}RHg_XJE9)K~$kIc)Rbu4({^3Ub&Q{tG
zzX^_+mtsdBM+xq#uZ5PDyl|Jfi#scT-1gPR0mgBiK5|vCfY0t6`ZTWap!}puJ3{HU
zzJ}7%6Fu&y=&#uQ<vY<ui|6A@FX*PvhDc=fxB*?0779#g9YrGa$sz;tNx$Mb8<mUQ
zu^b=hyAb8ZPU@!m6cnZ$;l~axl#=Xlr0-@I9uIwrk`n%wt@z>?Snk~<uiP^^RC$IY
z@*MY~q4;WSuB03F*X#&{riB!>AXTFBLti))t)f_D2RA#J*{Q}Ii0}&34fx?IiH7K*
zt+At~ItDM$5O^h@9Hq;+zPsJAVukypmP8LIuRs!4q=;Zue;WIgXK>-g7D^wyfp3(E
zR>~@C7jlDlq0XSK35!|O_B!p3?-xZ;<`b)wK8v`aCH`oiVskVOj?#zEkJ3laUsxgf
z<M)3co}#Tbdz~g%YP;ZH<@@->p^rCYH`Dz?DC(#W<_|qK)v9W@HKo}|?fRyk4e#(;
zTfR%1qLZC&+q4Q_LffL9L3iXE8lVq87VmXD5MRXF8s(4Ox+qzpA7}&F4&^R$o+wGV
zk4?^|d+71$qaBTXb?^B8^>2Rf#U0z~``&rF=bdJCj-GZs-FhtE5Z|`l+EV+iSb_KI
z;5a>h@gnypE%ztYjTP0?hU%!2CfA|^(J=fF^#F|_SEA%)LC3+-S^D_psdvUfKa8i0
zS&KYtS)5qrW1+`*zN#SwQ|X^EUl;cGX3?PdT6w&`|0_2gsMxqyBcL(qjhi3)&nuaL
z#Z9jTlN}saFZ(UPF<i%t0oGw0D8K+d$9wa^#0wDx6R`Y|(+iK7R3Hs35!WfsVd=+2
z>~Fb~7lP&VF}yrUoS$$hcH3l+hz{e(va1mRTaLka5<uk+yMsGjEP)S#0!K}XFvJs#
zh)F=;5kCB8IOO4vu*Dxg{&1KT3VU(O`2b&rL)`FVr656gV{!NZ5HRx>v{A~3FuIsG
z;wn4w4L<BgR}taMpIz`%n27FQ|HZ|<gz-zya#~kMwW)z@T|$$3Cl4M7obCqH3Goy-
z;10!-^CeU2fWep}AXt?(kzmB`F9bhtTUO^uXAB&6BEdK^=%K`rqvJ9v89p7P@<=|-
z(34-e)Autn)5)@WZIyk&a*I9^W{g)TYEWVU^G<^iJH1eHS<xjXpA}j+m+bhuv~Yr<
zOej38G+QD)<%w;+&JQaj#PkSW2vp9h1duojp0bk<%H+K0(}TbBg&{`Z&jJy!#oTg?
zBnq!h4c4n-P7rEYF9VEN$rD06i~~XwgB^JT1~fuhd9cIeBGeea!a0Kr;fy<x3GTq5
zuD~lyi4{;n8u@~`@(L;HUVbpg2yJVEi*SxRyWPMa0tEt(SIx+7=0#=a5?cgg*PYS{
znv1lJ3i6-_rqctV%HlBeAqoH<SwZDhR5_m8g2MX90WqBfEU>6MzEF}uGeMUz;Q(V!
z9#CXbSr)3q+_A#(3N9hXB@SETgU*OAJb>Sf4+eb{TI7%84C5}Z`8U}&&FaBk+5!{G
zb|-A{qkP#O{X%Y=xg9&6{Ro3!(2UI+{;o${W-ua<!}ym^*>ZZ3kNF$_9BOWN*zD>+
z6e3<WvXj08=n*KJl7`>WwfP}Xp#vVr(+_`P!p>L3)x(=Vqa@<Y-V5akl;u9mCLBvT
z`IYTCjz3BnaKj=%{><O`J+m81{?74j#S)bvJj}yKZG?ZgVGVl@;>T|Ob5Ogv-WCEy
zoNNz>{MbAEBXYcWV_T@K@>;Nlk|W~sP>1t;hFeBSxMQDtJAl)1UD&~H<(bpt{5!1k
zFyS6=hZcd&yqv_`-y~?OAx4Y9?mXg_zstj)oLndP5sG>OEx4J+B4$|l&f#p&?+6}#
zcIPQ&acDnwH|)TN%b)Yi&pd1xxy)?wMVRbHc)biRKK<c#rYBcR{#MvbfB41A9!g(t
z`ISD&&SYKg&M!FNZx%P(XQsF>r^A!u|C}Aab+f#>`K$HKdfMH8mVWTT&(k|U{Biop
zPk)(y{>y)#-v80h)AsI@v{`LP7_FoS_0b+<omI6#;~d_I1x&dt3mw;D>53_3E}_14
z^I4f#Fkj&P7Z@Oy=X>*6243Skq=bJ$BO1tb&*(|Hk#Cp)6G!UR;<rm{DXk#~tS1gM
z;PQ4RA({N*h-xO+ZpA-wj|KORFloUACHy|TATVR+0e5@EeKII|P#mzkr?yo|kM_6I
z-flP5ByT+2UZvZsajI`^rXT<0gVgGD5|^U08oRQ&m8x5HPsFn7oe4yEGVCZ+VsRN|
zQvacT&5PDlHWpmLo3jgq$>7)(ww&7SMyl5~EvLLXI7x%SDELOjA#WkCAPcORFGrac
zll&~?gTH62Qs1g{z#o(SKGnkb3_2n|(GQ8uyt@-_g)GG>EAuh%&vetC%u$?MrOQqo
z@MJ}T_)zl5M6wskN#0Qh>77VS;v$<dS*d`j(O^=u;`&=zWuj6%|H_(osk}TmwgQpK
zMie1dq6jCW=~X)JqfG0i^YdXkIzIF6y21EDQ8yF2KAxb6+wEAH=7k{-vD=CX%>A8i
zYPaet&w9GJyh**mB=v{F6RXdaE{ZKCUriTE^YG|6?kZBAIkM}lH%N!d+wtkCl~U)!
zp?8Tmobtpml-wc2t`qLydh4yHs=H=7Iy_3h``zz-#~4afaL(i-6VB{fnU05v`^1<G
zr#_L1Pz+En)G2sJ4nn@6f{^z(qK$eyIzCPxe)LiL<i$&+bL^cZ)GhSeYO!Oe<As|j
zWjL~s$yn+SInz6vvfc9&qD5$vU0#-Ll|BhVGi7%aN=+0@?0!PNvhqf94i7tzI^B+C
zFL2I00CF_6M_Hh6=$gf%Ozr{_xr_%IGtF3rS+{J1?8M|JI|q<0QKqn?5OU=Bf69-H
z#m*Qee4%xeM(o_e4H(d+<yWN-ozm7x-@8$uRehsG!`)X4s6I%GO7h~Pj04Bl11yfE
z<K8mUu-Y0pn}-d>HF12jTTyt)3pt!Thca@kJWFyGjZ(+-J>YGt%5Kt4wPBT0^^R-@
z9#G724B$xdxQmT?Mv-(PT703pM_Ggn8oNcsYbZ}BXi?m-yN3)=ze`K%ldjnTD%x5B
zC+sS$rft>Rd+)ug{%u<&SQ4KS`9i@#J}47PJ$BNf0ByF~;%jPZ<B~O1zwR@X3W`j2
zcE%3DP|&&Gfwtt&ahi1H@OjWEc%<ws#)Zb}wVD+rz=TGS_XDrdSE1mc-LSK>-DVL&
z!^&ip#tH`}y1CB|*%^foeJJvOxn=+v+5MdGGnrnf{hXwaKYpG*`s77={^FI7c04{g
zO~=Ql>Ex{MT{HKhV`!QFhokdO!Mpl}BlS-MwP!#tQ|GEH+P}W6$nnvlN_%zb-eW+8
zW9-=dByyn7VTVquqdrLw_o99L@;DtH_pBsmr-`dgdFB2=?)>v^s5{{rnV$6MPX&z0
zc<u^g>~UR*2EZS52hEYEd-0h;o*E~VPw1F%94E>B>D<@F4kFquZI=26SJYA5e@Hy8
zQ>eJkFtp{q`eAUtN*<LqIPLe()mOzFQ`9-*+NI~U*fq#KW*nC%>XR?OD0?(Es=cUM
ztQ=#Z0{tL6gg8bL#S-~v{)Po3KH0|QK#`<kEvb(b-mhq5;s@^3@2{zB8^ZURaL2p|
z@EC`we-zp9<S6c{q9GPs_?TGeSU6%n20BC$3BitrBb9&IJDkbemg;kBb2U}f=W!$-
z{SLdE7(1wU(qb1DJ19}!J}j?!tYcRoiahANrGA+sJyG^}N3+@i$KbmCD}O%0M|DJA
z7+dcZ|LR`#urB<xx6^vHl`d|VH2zIZm+(9S;&JqzomzLx-sSaQfBRwjUw-|^^!$Lk
zx|V$JCF6l=M&p|LW8s#z62}X!r<&@#qj7R?yOlbPsuy-}Of-riD-|EEQ$>Bp=E{9q
zU%5-2W+m-toU`Izc?S1Wc3UO!3(Cih>W#6VesgbU$MjC!uxy-u+kJz|NZTe)vY8H{
zZFY%qU+2RL3iK846zhxsVXVElMtKV#vuuU^c9+csDwpUTC47H0O$SG(Zo{-Y#!dQ=
zxL;B@RK2s4c13;RZKhA?1e)=}b<qh5X%<1S+Zz6{FP`C(7l&31l5Uwhd(+VV!a(DH
zi2=^1;eT(wdGr7JfBrxI^U}Zn<-hd^6{~AY<uPzBRD|69Ey9crEStIaXPe@$zX&2Y
zcH}EO0U7c6W(2_G^fO#1E0bR`VS6Ks89Mok^h&A@6Au2F7n5}Kj*-IUfBkZLg5)sS
zgilV%(exAlc2i%vU=)!+df|tOaK&VsU>FuwoJf;SfOJX_{Y9#LL>WmcM+>7kW%ooF
zLhH+k1_YeK7NIi4H;f}FPZ7o@XM|s@q(9GexW~%>_;z09ZJf^H;t;iLCs%Cv&&%(~
z!KmY>PI|3gO<Q8v*6Qj+RcOx8W#y;-Gdw}mNa|f*8t{0`)C#Dvq)r#cmsNCUVjNjf
z#)=Z(lcbK>9VCpi1}2y@1WJxXL7?S0IaW-;Omim{=P$$5;m2K8xG{*5eHb3Lg*Vu?
zu7Qv+;28#=^1-mf^!d_H1Odgrg<%#W;N4m%K49j=__A`&coaPbvfzzE$yF3W>a+FM
zK&j%B7OT(vh5^yw$We!8R5iFT*jPALJjzmzd;<=y#uW9SH+8mFT~9mDo}~S^o~5>!
zWEgXHE<qC&4VvH=A;R~eDI?$wkUTIgoKe22fr%p?5rkksSC#*=vJ^Mc_BMAtFuy*>
z{SGEL6paYuF!`)<!LVvERzk%Q9JEvhtx_fca<>-(Kq$E6gtIINADo|0ICgy@$e`Ff
z9|ZGO%HTRvUSVh`qd=)lDD}jMX8MNiz=O-HZ|c%YIl=bv9107=g+M_5StUiCK<>n+
zjarF@qDp?;$#yZiN@rru&wC@~IoiXT2IEQ%MhPJ-CTh4hjN=bsE?CiaFUDg@3?Q)9
z#WaXDjWR<w&YjvxNjT?s!4l5=*hb#;V{Zar3HR)6R|j#JnX@gV8=mnBYMX_y$=+|6
z;!@}_OvDs>q)aVBDgBavu{+a*r%q%iBjH(G^*euK_+2bFa9->0aF73N=CLDGhUui{
z@gjT)Rbnr~)z@PFiJiTp7$uGj9@(wX_ou?x%wYhOFtcx0icH1}F#|{890AM6q5)96
zidift);Spc=Df}@tZ#?r2Zz}QFZ>kF?$itYncn)IA0ez?9I(a97T01XG)JF>k7UZn
zERKxhIDEWm8KBdC3KtMEJRJ2i!y)kT5K+=@31nCfm<@J01ikaEK>m;PB0zre`G!&V
z!msTS;B|Y@iy#TE^I{m}_sy^}3hQw*ySQvQKQ;tcUR?(Lnyd^TLpaB9TE)(f$7tLA
zk>BBmdjVhE@{di}7IXU=Ru(~hrin6O@sA!tFQxzGfBmmmz7z8||KY!T*xG8Q-My#j
z$y@JQDCY<r3*_KNc?+E8tS&__hQf<%dr=5@lW$Bm<H$rE)=~^&hKaDy>D;Xt`{Xpg
zDKpE$=NPj4^$Zcq^I3+3%k#bYJX89B@5+3mwDqN7(Z@gs3ZHpZ+6WUOtYp8uQ5#$3
zc&BBb-oV7RckGMzVh2d8#qOAx5E%@{K1PZQ%-Lb$<FJ(vA1efpFP@(XXn0~^p&Y2y
zDruLywwhbw-<HJ}a?-=l_k=us`q&e|T$p}#K24)BI}yaM>zfH!_)FR(lY}Uu=r5Ug
zE0b^Br-HJLNpiw4$xVJJ?fUvs+TLb&P}3Ob_Xa-hiMUKqjfVO@J~a=(PE$|FtDMLY
z{j)Pq7Ggj%@ZcNaW!Wjgg9np`7%Sj2KW(`H-d1uG{HyU!Sml@ghfufT#la1HBrrJ+
z$af*rg{f|kbJ%%?GJ}=u$W|!*z$^SJ6MNtWej2%fRoP6aZd5kK6W^r~yM(SUQ>WER
z>w<wI3gy`tWrE7R%tAzkLy>f)y5{IH?)>@X&)-i!czaLv$^AEX>3npV4$sD(P<*o6
z@+k>SUj6CgSLve{N2;SM<&mAZ@bRK8!DaW^v&WAtdvO$;C;U~%OvbT$kFroE?%OJg
z62?mMgA#!i{?ylwWYf1LD>1QoaD*c7)O5)tGLxI$og{gvCAqNG?tnPOLixJk^0HG4
z#e#R}h$g`~cXo{iL(4f#)NxV?G{M9i3OWq*#NA*!J00($q4H6*^c0RAH?E5*N7!Au
z%+!M?{FpFTeUc83nH%K@r5iFNhKb<qogL*_a77mEY#5HXkNR5mx|SY4-b>G(J@xoU
zTN_W<aUp!UO(M_64l<5ygAQnGv<YY^#?B?nMJUubGS7++$w=Uk9YXD9%XEr@!>6<`
zX)9S9We<~{$mqy{q=oSg6XDxOj(P&~$~#tqe(yrQLm8+vp-&WQ6b!kTT{Gp*C6BwR
z9~2Jk$Yb&rdbNxynBbTG1X^Ne*k%RgA9Nho6o6yOjl2imnB+$Jz|m~<QyB0Me!bXM
zHQ`Bkg#~Bi?dCe#GdLq{9_ku-m40kP<!Fdj*k!U+SyNeJk*wuU)gKf3>H@8dVv-xV
znNvW>Gn3z>OC5L@lyJ|EJro-#cqtoYt%z>HCuQ(r0==m-XodD6fm86+o<+k<lCv;?
z{*!z%KSG}viodvD3Pm5~N3jQO(jSoyeHS|iX=e~Hbl`b~OSK7h-ii*I4R+CkD`kf=
zO2&@+>0*&9iUp1eY_-{?hk}9rorR$&E9pBXzJu}As{`*IDRU38D;EU?J8)NwXYO8d
z`@i7$TG6`dV5dtSw)E7!>xNw??(<Xz?lb}(G{(+K<~KNz1zcZW-MU>|3J!A)gkd+>
zd4G`BHKuUH9!I~iyA9aQMosmyVVsQACvaRPr`9m9#au@0go>*#Sft9Xr-pY6G1s6z
ze`PIQpqw^si<arH1jEWz)#XTaGE|+QEC!jJwo_Al6gn(yroU%C#WSFc3rce=nxKz{
zK2S)YfUItD=N$7(*H-$F7y1rx&h9g1*9w2iyCi&|cw<p9^-JH5;w~+(sSRihQ+niQ
z%q}qDic@f6yyv(^jtH(g%$S8OC_hyHF~=1$BkA~F&n@QigahG~HV&z}PTZzrrx%|n
z*G`U4L{H1<@!q!DQ5?YwjyE`7Q?O$URQ^?#vBscVwMmZRt0*0mm@JM!feHOJHLh==
zBvN|hn_X37cmm;>MGZC8FT2KlzZY|v^g+-bZAh7<zfn~0jDz+|+0Q4}dOxJ$<*o0T
zVpk{SX{pY;-A3xL`%CnVGW6AP&%3-nIXKn#ZCY2IvVf;m-%L;U+Ue~l+o>y_1`}f+
zi^(OoP(N~`5B5D;8b|x1i*$I}_i?FbCr26|chg%>_C#MSVrh600LsTp^|Ksr$?=->
zivWU_NVg`sX4e(tJx3fK_Xg?J$x!fPr#oXf^+x+)vCO%~56U+c9ay=lyitG95j$4d
z5yzcg?7SpD+`*^XRGqR@Tzv-k5`8j8RMlSI+HG2q$FYpT6ls(ey)NIgpybv&0O>bk
zQO>6PRqx7|NK51HeL6Vqr;nZ=idNL0yP87jN~1B)20W0sU}$IIk2^K|RSh-4^XCk3
zzYgs0&7$#9hyUT<{w+00%?YO&sj~A)Mc~43G|xqsxxq*6)xvSO+a<>cKPKBu8o=hQ
zr})QLHui^r@~0l+V+*_9nfPXIcxMy&mXYIY?*1yF^^WKPr-bDj`{FwX<Tql*v{HPR
zh%z{2V_ySMiLsW*nAyid1es5}2nsM1o>wt+l%>fTPlWTAbCd0JNF{1C%WUNBiE-yK
zyW34L992I!8aHnFddq>l;9cPJa^!6E?Yw610+{H0M2A73{H=#y2h0=ovxy#@dBjMq
zim9q}8fjI6%AGnub~nME1D6Mem3+cjdBu%7T$nxzV<wbUCam-m)6+Zedvz3L<+u=^
zORY3mkwH((fX8YE6euu^?8M;+I|ftIVCy}MG3UxoG$2wL<e6eI;J{FF%mBNkaEE!X
zR^l>Pgo;ZsuD~VYgYtDPW`q@hR~Hgcln%!oqV)2r2qP{ij(@)Bq<MS5w6nfo29%Fm
zE4o6MMIpj!2(Pe$F;o5li7Jlq0H%h%+WoYRm=}M9!;%Ko?PrhDlOMd3cJ}u?P{4SA
zdsgHzfH5(^&NQ$164M1!L%v|xH^h)LF$i<S3ZGl$fn8Ue-OC`sAP+Sm=phs{`M@zV
zFdI-SN0OlULZC+R1p{j#P-TYcB3-sKQ%@AeD+rYb=#-VQ<S!WaIMNVin)onLO?I-W
z-dVNNAN0(Gz^D?3{GvERK;#=?g24{P5rz|*^@<(k(n1?Zm0@QWX~CRS#SF5$44kvY
zoU*&JRBl+XLbxbR&2GwRD(3v;@FWd-BPxlCO)FwHHmmE(3){0f=0gmYaIwKT$-+6f
zHViRi-0!unXO(-tfGOi7We3l)MqNXzN-*5WNAwLr3w$di#O$C3fGzCtG5d{$0`}Ec
zxbju{y>Kt_7DJV9L$P}?;iKOGTpS$AcvIe-y<83lSjd|VPQ0i)I}le75vc?F^Q}*C
zNC-Z2dCU0XPr?$wCi04#BW0J_M=_2^S;;GYQEtaoB}5GU<*2;kH^T9bvSABY&{8;L
zxcZBC&P)N8GZ)Yb)cn>D<0oI(9bXpw_`zqfa0aPQ6sHas&NIWeL%WQ!>y@MUx7eM6
z0{eRoe(s;$BA$clQ695<4(<R>%I-u^8V;=goQD9a06BhkB|x#*YD^P1fY~QZ!LCC(
zJu6;I5C<<<m%BVXrV&H#^I&r*_(nvB@~{18(TZyZA3-CA9AU_vS2@_th~cgO9Eq1r
zZsWzn){GC_oExVXFiIXVL@6cnw{5v35nN%&clbNl40e16t>Uo==}kjGdL6y_FYy-)
z@o?fXIJtz|Jm7!%uYdPOtZ&c!pZ>@H+XF{yt*jy_WkINvJzI@u8C)W_xVd-rU(~ZI
zDCaubqI?Q+a~#YJR)+tk!T<C8zK$`Zl8rZwJU`Ftiw%&!D4&6c@6GFm`h^Uln!gB#
z->msYY3of`e12ASyKf<U{!}0)ny%Djm}I$@=hY2&txVE*C|>bC?j+%Ejdr)|dnr(C
zaU>4M1QL#w#_;#>m`px<!4W)^_$P3Cb91GH?o(Xazn+-;VuFQxEtp7w$7Rx!NfNQw
z>FBth4o_G-$FXVL|0`ZfydmSYkNlAVKKvr(h8N~;7AEm-#RnpLFj>R|2`GSX1%9hl
z6JNreQx^$E2wg0b+bCt8fBaDz^ao=7S(LZw%i5WIWI^8f*_rv$GN}$sCd4Q+{wTgY
ziK%jX7p_9kcN9e!d<FA9R(vVHjAbZfm<VNpiV`8OF!@87P?EtPS>_Tf<fS?j!rdJ!
z<Zwq`LID9x7Pc{o%<dc?o2c#(Wf+UO;AxpSMaj4+o|;Kb?&#szKjedp33n!O%-OXk
zbK$M2ZzhP}dFyd{_vyBeU}6IJ_U0k|@%c#_PA}6_@7UY+q}uaWr|IQUFZESlD4CE2
zz_|n@@x8a{-M62acmLh*f0wvGe`P&(4bX41Yljv7K8fK<GNAs}6xJ)mnT!(+fqU{@
zlZ@JFHR6;E?+m!|ok}RcpyaA}_U?8!J%04qcS*0VRa75ql1XD%4`U`gy(jL)?^73E
ztay88TnvXJAKU5U4B7P|9D@_;wjz23Hai!PaR|#UCzL`yjYIJ*Z>er6JG8-g1U-^}
z(&0{K;33l>LsHj`X3M*6n6&5kDHOUa61=-biMZ^0(YVjc?Ns?BUrdrSaZh?y2q=H+
zC?Z4$$j~SoNY{%q1q-E7EE1GVqkK_zj#gupK8ixz$<qZifba+ZE*bzHbnGMQL|e#+
z+#$ngs}fk=L-B<o2_+^nA-gRoKX?J(l#xCJnU8i(zRH9qJ8EbP&;|+=6e1`zEJrIY
z@-X=3NJr!{7Rs^%hV()?8M|_56U+(F_ka)FDL0e5+@;4Z8+rm{P++5kqwe9kXvbEf
z2rpicDR*|<p#Z+Vnn=FA_ihx9{A<?g-nj#Q=!Z~*pb+Kg!+yVK`P#>M3I=p-rG(mN
zCGVW*^@r(L{U$qatf*Jn=p)!wNV^za2%jn=aj83wMq{3&vcX;2;AkZ^=tJ3Ap>k73
zD^itj<{wBml<P~%Pd{;FDU6-_3&BLO4gQ%k0Y|a$n2GXGxLe64T(E<NHbO<SYZjV>
zK4^<f%ro~wzUbRom<%pi=y0n#p?)~+2O5D6s3+zo=wE^kLzzjPu*jFTPe0gd0#|h}
zH_p9TSAPXwW0xU)!B%?fSsaCmlKK4n%t~uZ_sTc1*nxa;sq)FbwyHj>wwbCM%878f
zB$z1GIZhEJ4dWI&)JPxYiFY0ew~STnutT|EdRH6dzNL1P`G)A9x#NpnRSY2P%wn#B
z`e3dA<t55A>VmQ`(csx7)dR9JciP7(K|z0vV=Qc@KVS#%Kz%3+FHjaRE-<f#($jZ5
z`7X$A+HN<EBivLmX-#;dUxOAHgM9au@&LSxt2o+{`yKC>)<`j}JS=;6?u)}yxibc`
zb57$XyK_}n^nV;RNPhq=u``tUK9qkbz^WClk_Z_jb`o(ikopVdIh1U%gOstBqlGut
zV{r-e!?*%&*BK|}?*#**!Mb3xNP_Y6;_|_D2^_BFQ1i1w1v*|*8^6~$d8M(4{)hCy
zDf4B%6HWP&lllS1e`qk2;T(hLeoKANjq-gyzDg(O8rvt==~(Zh-pI-}+GSHb0gFxO
zr%<*Z9G+;5J5D{}>{hrl4J!;|(RQbvTDA4`)_x~F+HR_xjC;a~@aP37)r!i$=0zX>
z_^0RTLMS~r8>Byc^ul+Bazx^P_;-Js_V;%^_s7wyS8C6fS9fVdAFm3c?K9WPosU)Z
zbKnyE(J$3lpsoI#-Dks#Ti^L50#M$W;JOz+IObM-is*nHl|8l3P)x_2mn;sTZ)TA~
zRc(WNy`JoKRYz){qAwP)cvrMwa2)hN?FMCav$m1;x=rucCx46=^gJsn`@QOl@tk|#
zP`=UD>8C~PZm&nDlk|rVxf_gmL6HH>k?1T8{E^y*+EN)Oiuqpg_19xU@%7*QJB~!B
z=AcCpW+0!AuT+F>SY<(2k1AF}SBJ$%C^K?};FiM{H(PltF7L9Mm=EH5g+T;xXeL%7
zj6N|D)CA{G`Qa^kb{t2OSqN_Me#_CIR5psoM3o5_yHXAt2DY|9z)vpnGCk)K<7#-=
z@Wd6^vJ$)a<7i064X&7+Qhbyz+w7FxZ2_KuGFaJeSN-ZcLi0p^!<PLq{wS`+6knRQ
z?BIaR2`NPZ=4M>-jmduMhoj7TC?`L&Q2LH?hhgBw6qg9aZ%%g>KirKQeZ)Ir!TdP{
zl`GE&b#C18wJyPsw|Ywrt9jKCaTGaLSFuV%jGP8I7*bYO&<6#hC}C0^Yrl8y^Xa)m
zi}I|Ai6HEzxAu#{6zy}r6?b&Otixc!ShM1h-B=9t<Pw}PupyN3&DL<Q>{92+R(KzQ
zp}~hi2*nyoHJAt#H5U?yh|8${picdg6*Yr?nhXZ%GOwtkY7o?c2R6)@sI1_L6?QO5
zFe7|Z5Z*9<FqumlWEgM}fLK9utAQWkyePy<e@z{*%4Kd?fv2862jHCIy|*8yCqH~I
zb@z6K+jTP$mtqE3J>i(tk8<eMbjlYiKw)waN)U{INqt)RqrnU2ia5|UJD1j12!{gg
zHcwLAScrhpMWF+;<YNaFmOg@=elQpeZlR!JWKucQNMTA~nh_L12m%vB7K7aMLa>!S
z=!&!Q+0rjzO>UtK1FvH81rLEB@U48BQO|f+AQW!6BM9jw+(L&8rV60EQhu27Zl`Oe
zhBN9%65eTRFjl|-j6ie0<+uUpmm>?^j#Pj5FvViv`^UZ16XS7vfsms#g`4#X0>*}!
z9Rwi;@l7$<2yqxNvnpn5O@Av~QVR2^UzAUjSF#bQ_=mQNa*Mp<$8WS{+~i_Medo4_
zVV4^<SrpO?4^RlPA_6oJc3_p5w$aGqBL^i1e3sbya(?tCQ+6X5-+YA|=v{<?)hJgu
zM{{tX0V}>u#q#Hm{1deRp?w790#0t^O93KT?2#wp*dG8gc5#>}XZ%K9fb2mgY<ejT
zdlL(qx2tFcuShQv6qJ&<!j*J0Xd|3<!>UVQ#N?HXajUR;xqgI4$0X5N{^gm0l|17=
zvlCoD4rj2$#hPL3cNSK^#%(TQ;5XviL=+qH1S*`xY+S|#Xapu*zGn5KAJPbW2`7U!
z!?IOBF@}Ys+n)?yVasC|4m{wz*oiBL=uslP@n>7kpZ*DYM;bU|WN`=zrN_>)Vskl(
zraU`u;o&$AzhE3D5|oLE^vlklzrxiZGA!o}k7DE(BpLb0xxv|BL0txy{o`HiSQha-
zXXeAr*(-U^z#QIIAUI~!^9&ayJnn)?od^(R2sh%XoceQkjKdj7<iW5BjvYhT_*#f%
z!`FZLyYEi%^*{dS|GVUMjR!NtsWi&7$1;}<Q|u97p5u8oDb2#CO#Uc=U-SD4=1s8w
zY|I=Bf8PA_%>B1#g699J%y*ggYiVUVEiF9!zy)7&qdlpe(O>u&DY5o9;%Hb2eXYJ?
zIGT78m)KSDR1NN*XmKQ4O}yblI_nM7;lYt7g*++A<PN-`csV#3E55{=GI_}(b}EV0
z_T(9ME=cb*Yl%Cq&POrH28@;U&2)M?OwV5)r1SHU_(qOPBhwm56c7H6doh@pq_^^p
zF8wii9LgYehjA2FOajwQphVhSUr9~wa1u`%lep|G62B-O8fM|})vI*y>ZN$zTTjw(
z2|g3R$QwQuO%J$W@`-fO4ZxowdvJmR_Zq=hqo69gCcp*!DSQPJNQ{T%he@8*l?};0
zafb_w!@RpjZxmSY#;oo~&hpV>Di1s~fl+25OTjb4Q(K{?u-td+1!W3{LX3rdR5tLq
z*M$3lP<&YwQRZ$X8!-XQ4h|Lx_0G=H!SPAz3FjyUd<>rAFiFYIyk=!3z4v50J$|&C
z4$j8u!{-N<w{~_Vv#5+`ig$2)p3cV?k^>~ONfu;sZhyOFC1+JKCA9a4KYo}_dwn0F
z1s%a3fKSR{`qZEC%p|4Cg?m{T2LSTVBpYR_)hos+G>F`VBB#-;r%k2B&Lk%F*4Bh8
zChR0rBKvTxrFS9;$IvZ`mYr_fxZvAIf%VyWF|uMScGPf`-<9zEVES(~TBbK9@{tG0
zGkHh$qE4VYCc}||tPq11CAW|!ld;e=ZtTFqptPyg8dlt3yt_)w3+;)q9C%*nDH=g$
zhHsuu$1abRmB>b_L(6cYSFhw(7%xOb4z|)k@}&M)012%z+3877$>d(Vs5tC0Asx6j
z%ae)=&QS>Q<`fndc#;N3T_QU}D@>|G|LlsStpE%9v{FX+p*=vOu_%(fsQmXVQdGLj
zp(LmNP@0iP@1#Vgw|rl>lBC`3xc^~y!Bq7(QGdZM4C;%8rzn?L1nGOLg%5V^P=1tv
zz=xmr5ot=(JFw(WouE7gr?eB_-K26*b`)r|Zzk?<IYv%?5)+*-A5W(?(NI6sl#JM}
zS5@W}(aK0LZqoMlw)#wV`>8Za<3jbsNg^B%#_rhj(S`btrQ9|*tyG}R)2BW^JWeO4
z=Uxne(%SbQs{G&|SzN-n!;?2?j6M$vWKy4<Ae4PP8Am@XdKmNtYdls6wIOv{%(M8e
zLzNFDA3GiiPrt!D0A=U+ew4;mHoDzD_-MKjiwQ(;)IYnJf>wl2)m=s9W7imp0NN)s
z$HY6=4AAztn-1kOi(oyEq#SXKBYC8bXnWM}E=S*~P4YcazNRRk)u%&C^~P5E#ZNy-
zAH4f4)izdzkBN^_<Q}nsaC&id7xyc%V;T8e`J^9PQ^onZ3zd(ufHUeE8bKjPo8>AD
z=2y5!epPkOQFBlBcT=ZbPdgpaU}Y^esvDxks4phqna4n0q%Lo6RX?IPcvX5e@E=ME
z#ygHLB`x(0^m(Eq=60x4(zF4d{7|Y)E^pIde3?$p$LW>mf!%H$^=sShhTz0b!Y7Y+
zQ+0jW_q(CEWTA6QVOU5)zezjyIOKMR!kJx8D`}!(gyZediEvCH%AT=scX_G)Q8Y(e
zVvJ#KV5IT?RDCP+TjZNEk+ydf=kZ3tJYNMqX)CmwRpo`S?Be4HLk(kYLuKA-OMN%H
zS+|6nSOl|_RGq2Ys(VKx<7j_)DV~5u7`JL?8!CTA;gkoZ32Y@k>pwYqxjzUR2R7}F
z9bd$yXQuBVUnrl}g&!84->NN4MJq#gl}&~3>2(?kIUH{&@^Jsfj<RERWvM+5hC%x*
zilJPLXS4yv$L&r-HGZ46)Ni#Lu>)*tV<oN0t)_Ca8;yA{==2YNdY(oXcj-^u5BBoN
zM<D+DU;ZNf^d}#<?H(NWln?b?%J_xo6B-5IO_Ys-h0J#=T4PQOz5wMTyVI(h8b?J}
z*Ejd+LXp|EKNU^WkI=XJ&PKtYof3zB7{x8e5Z;IuPELECyQSUL)u(k;CKSw&Cu8G=
z+ERr+XIXUstqSLtC|zaf`*^I_)ccNcgHc4VR#gKP;T&F26=Qm*eg5W;uhO4BJ`&vs
z*~&P74)B_pf4`{!qyLEK>zFswdDDDdxZj&a<2%0Idh+B~G(Z><I(xHw`nJ-Air^oA
z5fwy$b9jkh=qY6!Hiu(^*ml44r!bB$$UrC)1$0g&uEX%BNCAt+LW~F-;loWo4$n7X
zSv?S8@QS!4KcoQ^OvEYQ<yV3DmAl=5Lm0gc!)X9tz6*YY^}qbEgj0T#Yr^R<JmTX@
zc)Jla;2=a4xHG&No_}ZI^pn%!na7V0UI<d*7e5ikdC&PEV3d#Y@DQH78G!wapNv=g
z5m(s?Tq_uHOZglo9^<LF5yr{Q({$X~Q{okSagV%|JQP^oU89a|S)DmbFec;}ND2c5
z!(fFCu8NDW!{9=mVO(I`SmA+^%EuIlS(yzydeHVWprTMfu>eyK3_48W!VDoykSb}g
zs)3akFq`BRhL$*7oJgLQl{U;Dt89pY8;m%NTjWQ~hVtWGXv)t(O!=jlsB76-#TSYj
zI#XlA3K9k<jqd?Zc%fc>>>+s4AMR=hW&+6XEd#z71&+jcPzO#q%9+~?^5DtBF5CeI
zL?|veFHSHCt=(?g-q}mlTFnEl8BhhFPCdDqJN+`~Q->(oyt+{CtctR5q7O3zDm!=v
zmZIyE6@PM5JmP~lgl>c-(t>b|d+I?jEzBu>@W|sj2adv1c2<AD03m1v4q)g62U=2S
za9{A_vH(|cc@u)k2AFoC()a=y{DEi6==5Y~g$ebG0Uy+n3T1zoSFfVfV+Fn7BP_FW
z(5oQDJV7_W5GI^&>H<X-yAr3Psobb1K&q}iI3qOj)<7=$ibW{&AAvUzO#h7n417U1
z?$6y;2}<4r2UzEi+_D|xHxCG~ThIh{K7r-M3$sUVeirSnHd{*Cb_1Lopq&7QTe0&F
zJK<;Xvpd-dKOW(Xw*p1qLOO&E*raL53Ik>%eZt`nKJ?9Bc5`@DwbQbDNhj?5Kx0lH
z$k{y}+4-BnBwgNu?Qa8i9NzgAVfANR>M#G#!|@kZ!HGOOjQtEb{B!&FztVELwgW5h
z!T0Ow&-2K4<g>(!a2TgqU}GbmJxGyU0XOKO#Nl1aWGC#|+e(ofCepzgVeRHPgt3eL
zaaXBaN8uIWiXU#I7h{H;oM+{eH{be;@)4Y`j1OG(4eY=T{^1sBMS}iYh)Wyg8-gI-
ztbPlD1|caw@mb>AD~EL%6+Y@zZX_kR_ERwDf6>lwL<Jft<hsI7e}*YXbqfB*>YsR=
zp5O5}Yfdh^>)Wn561WUg@G*lri$ej!9#8q3e>-@%>f7xVciZD_yq7qj!R|Q8FT49Y
z;~<RO{1@in=;%AU<LkfukN<n|$-!?F!?=($nN8%|@7JC%K_;K=_vYtUHj4X~ME~|o
zDcBc1-<y9eCTiO~k!?0~N@`E{;*+?$;x?^#qU1`=?H+f9Wde_h4fuO@L$z9UANhmA
zNeoxwm^FBOCQISR*#*OqR7`51prGF&Ze<hY2`j`m6T7YWUREc|<WsZ0<w;}sHaJY~
zr&?awNF0^-iXHi$B;^>aTW6V-;H2~9>65g(v!id68ceXuQ!r3w4F)mEYo&(r0FR1N
ziSC0**jTthpSq&->-Oi~9+Y0(7sU!~<doo%Su7^}h?itn1hNEJgzsYkp8H1uhF=9&
z@W7UFB%`3nfS2TfKVTsqk9V0UUn?t|t{^|bGY?Cy1eXc!h8~V+vl2t)^R6xNyzUOw
zu`xNuP9KiLW77KivM3QyP;k7R(pTOYU*MfXQN|9M+jM$zk`6eA?5vk2qoHJo8}DM_
zo;7w9_&${@?gm0x5xYO22eHSJqpp<aht&1G9~|ez7zw@b#7+f{a^uo{R;bf|Q{Qht
z*-zb0O+5F+lU(4QNq2U4GvN$Ujcf3#GE*nMmq{`T_zgZkCS6f%apWEH2e?5V1TAG^
zm-?lONuLQr6eGQ!>i6I{9UdLIj*pLzQoq;tPDT_*D8Y~+sK<a6@yh8LJRHHu!cOXn
zx<Fxp`~sdx)AFI}naARi3Pl}I7a=z+^2bUik8Q;bgY3yJJ7gzzppY&{r*TXfbVyp@
z1A2EjeIw@-+M%xF*m-yhy+xxO{}_u)L%Br0IOzkKi+CugI35w%kx6{Yrwqw{<2H`6
zmp6+Ap>K9{v6F{4`J=p)j~zhWPD|~4D{XC6tYAhti!uf!5el{jI}UT5(oUF^XGhEC
zrot)A=6W2{h_ZQ|JShx(BXS>k!>}8N`_<S52;HzylY8jA`$=$_$VH*W&N}dca=a+{
zP|{FO$hNWjh@E;D{-y<@yZ{gEjAUXPIS{2-UFo5WVK>H(`kyWJhse*AZ+oXJw;d~4
z*%?aSXv3B?Aw)fdrT^myPL8L$xtgYm>}}zQ<6^-J6U|ND>J!-!%dwiBPCM=Iam?it
z^}TJyiCq;;LdU!b?Grq5)Lk4aw!C!j{+dOtDCoH(h*MWM-VYcJ^}#4eP)>1oABvKy
zaD>7gI6_d+r*O@J0(Ldg-+(vIB|uwx^ME(s(<6V}nFrxsJ@OGdO$1K#!D8n|z2UlN
zyzru7g}1^-b?9A?f&qP@kb?%T5Es2dzgt_}ImXU2=4LcbQOERoqCdu*NBjG(7k21C
z4;+=sm~tgNUth;=HeYSQ+?4t^?x^JcL3XB&-FK?4lwJC+q56b@Xr4ZdCdz!)>Dj68
z96UZcOTGTceG7R(sqOJWZ3!h+f2clrEErSLU<d1s`xFg>Dkd0xNZeV)PAcfx^9RNQ
zeS7SrW1yq2hGsYp^XTL(y?Aw$PEPxYqcx-7Tuo!a=O|znsi5HSot#R7#TxY4v`daW
zq(7$(QO3aQlJ9&(v44DerZm_Ujk1R$5Y_Kpaql5YBeet3G=1zfw8f4trGKrmGHpf}
z?o1;O>?XUAo!wd7*~Yz&9DRGz8z_yQkJlScE{*e({@BmUqtiGt@a17TJn0M0T^fs)
z&dw)(4o-*Z)#*6BI-8`P@CuKzw5q;-c|F}}?7S5YSggZiMUUc6*kOk9QDwZ;aHwjK
zTS}*cE5kePDPJh~t{;3~)`jxR&T{a%B)AXahqzOe9Z1|Y$IeXdp7YKqm7OuU-EO5v
zdpqgz{*LePW4A2IuQ;lCqPKV^)hBaD>^k$EW#U_?W9nd2{ZOmfaUVgy%Y(9y#UYJG
z><WYC8DHo_Fz`tDZ@1OYRoQhH3VHfF&j;!aZ$v-BZfF*`pj4!vWh`L<3JW;QPpKU+
zcCcHIovP3!^OWR+evZDcUS~nWit3z&>{l8i>GQ9QV`ztCMb`uy+Thq=+VJ`LAiX*`
z@ZuTLV$4N}N}r09>3a*+AK%_BnO@jw$nJ1pt1+oL!3n)-7#Cw~3={bH-u(3#?)X}H
z>*=#!sYlW$40Cw!O($hH{o2`X;X<c1a~HAk1tDQPQ8I&<eP-|gqwp|ua+|s9pEwaM
z+;9%J{O!mvam2%#)6h#<nEBygw*cyR1x^V+gDn?_!A(C5i1BR)zioAyfNy6();DSD
zO&I{>Nz@#``O{x%5(Rd^D`7MkW<PxqK_-V0fWS>WwwLrW+^7XTG-5mz#{NzRYcLDp
z$A9=0yhoD!2nPP~Ln5=Z{BO5#&rb8VVdOu@*T3>p-r=N9kg&Gpw8Q2&vv@ccgS%d4
z+|~Ikt3zECQ@OmZj#?fFR6$!9b;TM?VW1hL5vXA9W8l_+D52(xl~|Lh8Ojh^V-*cy
zP-5|*WMD<5g#rtP!Y;zEI%#%)VMq}s*jsy*H^R_4vZ|3mnbln|M!+S12uI)_VUJZz
zadd@*7BRusFdGtXM;gpfeBs9G9ylB77(;v*W$<9ZT{wmDkyT;h&)+H|VJ$UeL`L4l
zR484PD;T`k9$YY|u~=oWkDvHcI?8`V%xG84TSEfns_=>e1^_6Gd|aDA5}NWe*icTG
zP78X1VSJEp1q>yXqdOkV4hlJP<yEE`Ho_09YQQ<=Mxk^oe0g`5aY>j+17_9b5e(o!
z+c<!0@MjvqUFCHi5Ug^(!G{?xJ;Ww%(8HR(F|-L*?FI~JD`1dLq`^1oIUdY8nE^KL
zg735xmwH1<h-VQjecubTqB#6HjLLVZK^f&I`3u~}>N~H})C1jt?`5SIXP85a2rf!n
zao9lzY^7j2Buik(>DVrJEP4c5FWduf(GDXVK}p|#b|yd3ugv`Y#y^KyjH5psqc4au
zhZCVx*7%H{+=m$F{)7TCxXPwD(0>HZF{ul^XSB;ZKTe&u;zS!2L`63a0F!axd<y{4
zoSyr^2yXy3q~S=0>vn12*a=sB^y|-TUwD&K$MeIhym_L0JdWY~kS2Bl&Fc$X&hlLz
z{_?wo(@%uQYCEr>g$NmL1%_S2lYhQEK_B+`o}+}Y+3ts*Oc(htd(57wO$W#(VDL6x
zD04VO{bYNIKpii_jlAL#u9$ceIcO+{w3F?Q6CTA*8BITY&%*K6UoNlR<e5w02!78H
z0^TFC-Sr#oi8{|F`x&nPWY2N^T6h<j_9ugO%YJqz3&4mp?clemF$x4pVV8@&{4fUd
zondD)hogU<unHU@dFzLF@%N*jOt$tb_VN%%Zh<p7$kXw10DkADeDj|B71(x{9UtJi
zzHD$`POlCRzca<x|L(v4AG|;e!+UOr{^Ps2eXjWe@4wW1%izB^-y564$9yf5qef4S
zp>2y#)aQ-ZL{A<qt$C-{rFcZ{0Wy#6KItx1H&?wHnWK$XR+zwr&u7w{1yroiMk!WJ
z+uI!<iv?fBP8QlG_s(usxmRT~vBH>zS+ExH$m}M74P}y&$wT;QCe@gfW2O7q*||@Q
zVCM+%Si#POP^;DQBxj>uw-N#c7k9V}h67K$aL>}M_(<dmChfd42j!N^0bl9K_Eje0
z$rCH~nXoxeFFtv0g%EM!6`7<d$}0@V7_xgIlq2xqYnEH+B>4lcO#Z@4t_tp!?`Gj1
zq*ddD6A)rySX2o7HoPXVyjw<PgolRDM)4NM9X>eSp~_Y6ewldJQ5?BPn&1t7jY%s`
zDPRJw+))GH=0%5kGttFyY8)*k2H2C;EO4Vd+^aU3j6Ct);;5^eTTgDWdfy6*+Lq)z
zrJ=g#VaFKv^H6@22h<sNeK8r#!$cQ65b_B!7{@O?eY~5xtu5dAG8}RO$Z}e*R6U8y
zjv{0#8|95WF`>^c7-Svf5yH4`6c2@m@0fxnBrCD&3mE}rhj+!1uF#AuitJ)pO}OGn
z$B`9XqcojPEK{K5uzbO+w{VI)i(-v|mHgB6s3%X#>4WiyyrHCv9a?G|DCk5h4x^+(
z*10wAp)V#7kx{89Y9D%VdlanFKIh{aauG5bi~VBpBZgBmP!6Fu;FODIt7+v1atD*C
zB*YPk;D!lN<dop0#fu0QM?x||OWD|M!Q?n?f&4*F$mwP0Ao)iwqFta6<~}Ie8p>y`
z4+t6#*$^B;|KJdLi6cbYZ4_P{t66t`C_gl`-D#&zr{UcgE8gW8^CL`#q6GHoDk`((
zd+K$oZe<x|jAQ;e22-@6^uPy44MOYpC^l{`6^>k~PDK~oKg8}1@Zia1l^aDC6W3NY
zyWP^CaRg#j>9Et1d=YQeN}~PUt|z3+eN)X=!#jX>nSky#y)%hBp=gWj(ji~8HQE+C
zEXI?#ONm{q&=K^)gfxm^cDbPV;s`Y6J6Na*$B2B(Jya|RXgBMr(Wto{&{wTP7s3}x
z9h5#C{Yu*a@9exoX-&IF$wRr^E}%c*f%bv|%vVgnL_XXa{?0baG<HKRTQSvaw^eW4
zEu?x<9Pb!X{^&nArH4Mg;XC1)QNM!6ZZ774Xsh&P?9dtw)Mu&h<{q~3XsUjK`vQe)
z^-<6{N0Kr>M4JXVw0RcI#`qx^qCaSoaK1`G6c@)Gs*h7$qR=_-kGvD=_~guOttxyr
zYwLP8QoB)A*;=XD+_IA5y?39bx8HhV<z=g>GONu(f9w*XkMNFSr9+=c9fn*FT{DJN
zm0t4Er^lxwD`D6z#eHq`5%BLQ^Lyvx^y26|ee&`o9h~-6X7%BsT@2-<T|r;a$LU$$
zN|hlyV<#LL7rTV%zkM&JqOq${BemYp)t%cr3nIKILpZ08Md5>@=T3GMVE1~jt}<?M
z3}IF6i+kUhu3&d4w20Ctb`3+5^d0o`DB-TJIS!E>af%?^KocD6*lDY6h)xAv^+Rc(
zA6W#zN=*7XInftTM$fy+AKb72+b0(ZE(#PBixcV|1*a8Mp}1oA&lrW7@`qyT;AALV
z4$?_q{m|sbNB8x{+$%Xu2fcAR<mpe->0p`$<7?l|HNMigCURoO-}vew4KD7}tCLCk
z_-LH|`070U@#R_iP!Go$4pcW2g%5?H#twC}9M#KRqQ|7cj;oQz9F1-CNnUW%Z1_0B
z$NSr9NBtN)4ebjhE;yi%q(3Dr<Lu_TaHx7*kB-TGgz>BLGVU{2umEUz=}uvnd|Y5h
zW8|iI75au1wNn&l?6ltRs%>tmUR4i_!6?3<cfyg5?>4KhduJItvS=&hZ*v2_Nss8x
z^QwXwb8E_w=~j3{2@D-xXe?L#soK@fH+(cDyxCP&X7<ieKu*({aDL8FkE-7otJna+
zqIuDj;(MG>8JSBKv8sQFBV-?BM>#rD+c^=87l&u5*W+F<dQMFIzHV?^bQht0Z~m%G
zR(w7E71bC6uZB9k8Sto31PNEGKBKy6Fbe1IXgp?)_`?ni``g3uc}Ex-{PisQ+3qx8
zw2Bkqv4orB5ia}l!#mu<uf&Nz<-UA7T$BiSSV(c^2W(vM3-5UJ<5>RZPYs90nLUf2
zUo#llPGnqiIDcpNPd@|&%h>hK?w;)B&9^%x+<9Zm-}*CLJ>=i<aC4l<gU1By`O|6I
zPj=$+CpQ-@l6S!Ta{O@1uKF)&%+kW{K;a(|<nChO&x<#VcK$B7#-9IN7UB>}WhURA
z=vSw?sg8R^9VfecENBr!kzwyp9K=@TxTq3FF*<TF(F|^4lwr)wP>X@_#kT775Fm)h
z;7_NBuu#qprvoJ~WGCzzL^asMR64ggubu?(zuZ*>4YBUXDGtmDel8<S0L;d9D88mc
z2|Hp;uG#IS4xSxgOB%2c;FXM-D;N?6rE&={aPdb_cLFdFvQjQqZem~vp&j-x-KkSA
zE1w)*8TL*l8A!m+GL$E0(%<U{1&{X94(HK};pBo+m>V<H$`_OQ76@<`AnZV~l1(tY
z-CQ_Dux9li#`Pia#1A(HCe@Q*0JjJ`v!gFm4=~+v#&jJ2!D=KQlOZ=VpQP<46pg@!
z(WMk9cD$O<ID`3#!JWL~?|M@h7)mA$VAeUBqpW@fXY9TsUE&nDpeY3C5b!O8()Q#B
z{$hJ|D8dk)O&&@c<_m#|$AR?_4&1}2fV=BUj*k%|B4Lh*;Lv#!gYKPPN{=$Lx|D$%
zKX3@4aWo_X1rs<HR`n(g@G|ccBEIRtdDf$EXWe1q9sPp*$~STS!JAiZyaCzYJnoe!
zpX1=>Kz=(;Mxej*Uwk{t%wGI4R^Y*~5I4KUF1`rox7wEf9A7!7oW#-F`5=&uY-AD}
zae2pw9&kn+?DlfKx`c98E*#FV1lM-L;p|ZEk>XDt%97ojAW;YmY~z3@!t!<+@t<Mx
zLo$o*ltpjXO$KGV{;=yO!V$i>7h87Ydj>Q902KZp20P`zt}g>K;vI$`a2;g};~@C~
zbc6v0Cc`OT3xRmC(FtQ0`T379*(`>Syhm7m@QVBy%VEHgU*XPQc5rx)rHl2U$8U_o
zDBu2$xHFszp3}!Kzj+MLalu&FWeo#7Cp8ZrkG&k$arhuOKX%WhaX{Yw@HbB0;a<we
z%YPA`x6_DE!Bbe;hH;1^ZVTZY2<QB#?+7}_O9{tAy@%ucRor+HrJVK&Mv8lQQre7P
zfAbGq?EJ(#=sdzW-?_cmO%^vj|E2WuJ5+rAZ~tQ`zWnF><@Z-HUorSsHQzG$@6Gq-
z>zgv}Wox`SMrp|_!&&KkDIVrdBQ2}M#bLRhtW_)d?hYnCHa+oG=OTNt8Eeb;R)jU=
z$A#zYVDK?t^_nMpkT&4keB2EaFRUmQAH_r-d?QDgUvejp_)d<zL2jTE<{k@H08cNN
zFuqFMUBoJNjw=Kn6T#digS<c=zyuSM@>~e-$@$^HG6y^&lcL-Y)@m_<8*$)0eVm!{
z#ERw(CIl50rP=ApSvoyENrUr2O!_EK9KGeqa`Akmhw_O#Rg{2G8jJqCvnKPl#w>k5
zUkYEX)l<FMRQc9Dp~<8wlR+pk%)_n;pPo==H<{!g{lPDD6d(TFIXoGU(gnL}BtNk$
zYpjP|N$g-^GSL&Rs$asBK5~e6Ok4|hjx$JCOde0BX*f9d9U>g_g}laC18>icl2)tX
z5^#(kuo}XdES|V$f)pII+fBJwQd9NaYH+Omn$l*G-F+Gi*d1|UnnGS;!g|b#aK(+h
zk0fhcnO54Jh9|^gM?okUQ8byqN9Is@C<Ao^uZ;qSG8S0{#fOh)6YfK4DSYdnI;L)K
zC690y43oLOD@2iR1eaYX#DO-bU+-cOj$_ve<GXODFr<SUVL~2KeJGFIVdR}lq$7NA
zlpb|o!daoRNnM5v5cI|*F$ykZ67s3B6xl%rI;Fm$Eo3j6sqG4eA8+8ZKomMbsfgS}
znc`URCFhB|-oR7j8+@`rkHvXWKR^bzq>s#vf)1sbk2ka;jN|jvMua;uj#7yo3mj+1
zn;nJZ$2$Q;)07i~QiS`3T2_8=H%vu%i6al84Hgh`gd3EJObHAWRM0@#WdOb?rzg!V
zuNkzm0Y+S{&KKoG@qhx2yrYcqt}Wq*;}+q4p~K_j<8*X<WICy<Pof_~QEpL&-9Pt|
z_w`=&Zm5m4v(xm^WIJ8Gx7$|M5s$l@s1GLncea~Wn)v8Al?@yrZ_`gwE@%_Fjk~X|
ztmNUyy_W}vsn_RDL}(%E6a}qiOVug$h7yBg5I2<``84f-hX)ivD1c&z;aL5^S-+P~
zk5Ao(x%&>K?Ra!<CAg34blt1pQNPVSZPW?4;@+o`+WwOI9iL#LB%pilH3a9-G)k`}
zrFnaM?fAR9+o`Me-t7n{>KmBA-&k8tHTC)IOnbDylb$`<PrEw`r!d`4$GiM|)r9B|
zI`n*x>V*E3d!N|3NWX>h8v0;I9PI}8vZFHSR2WuU3`Hn&N+^F-AB-KsA-FdGF2tkM
z=I%yzYjT&^iTd!DuMX1T;j#Lzsrs<BRM}z|pvs{1*zt$r82R11(k`#8h(RI8ya_nu
zu5rdB+R$ugpwgwi)#}2lp7o8Y(qBoJv^n>|&_m=8B_}(a4vvn~X|Ly_SBEUj7_ex8
zI~O0+2Gm|Q*W8bB<l{L<N6K%Ye)>vd^X)z3L$nPDjU(u=GGDRV-A->keWdX=j-Z2P
zp;yK>(l}Gv>{jBKL@FDFH)D$NsWFSb(7Q<WX569uGY0s$KIY|AUT7TrLaX%08yfh#
zooFjQE>iAZ6l{AaudOIkVCo7wTeD&&`a^KYc;lXswyH8R$KXD^iek5^azIC2wZB+E
zKwrg?k=!#Hiz(PycAL&s*X+bPJk^-pS9#Bc+u?=WFH?VfAsk+%;Y2hcJ_2RgQGb#?
zIqIiR4$jic<G$jsv+dT`im>Q@o!#Z46FGSYAc`Lpt&9QSjW&cbxb5;QkHU9TeFWFK
zplD@$hBx4DGL%%bk*3DSj-EzkHPzOaQez{Ix>OUQ5>rp@cs!iAtngLvSm|ElvbvYu
zuExhpwQaRo?!>&g7^ijdT%Cr-UMr#C*6w{Iq?KQ)`$l!m^c1^E>F*dH=yxJ4JlmSS
zw-jbi@E`4}9BRk0^M1{9q_j~rSMvwA>R(otZmnc%*W(!1E%pD5M~thqLHN0?%DNSA
zlx^wZ#&>vm*T3p=OL%4B2D|scExhJ3JN49VCYLC_?$Qt?r|6%BD)2wh9r=FUg!}jA
zuiIGh_3pE`EzlRmiH%;^<8hA{ethfC6Fz)LpC3Q@;pGV<yW_;qJPhqE-tiq?{Dyl;
zr})VdpV{B;_KOZ|HVKry`N>XU!qG2V<r)6r5HEgU%lOWZ{wRlDIj<3}#GOsh7$o*w
zZvO0#kKT?Dc_kbOU^*2JsNfM;q%r@_kK*#D4`SF2J7Da`plsOTm%Wvb{Ov46T;B0F
zH`r&Bdzpdd4|ucivd3GG(*&N<U3dpBA{;_WaRX3+i9d#S#x!5Ey!z8Fth$ng0ZIcB
zlmDw?O4qkEz-thLF+~x?$~fX(OAv*zfpKNE54(iS=<}xm7y*=3NGLIU?zjXX3o}ZI
z)oR}Dr2)}QpaxHNJ3;aY5k>@zhB|B%ZVdJ?Z!l1#4YLU>xB_-k!Ms_hpg7<G{3s=U
zD@q5rH&YE}Vn{~4e!3BJu?Fr`V3ZxcI93Apr3Ol$5iX(WN{j<r;W^TT^bv?qez8QB
z0g*w%N5BXVX2w)zZ@VYF@bbVGIYtqN`kX(?PM$DwvJF?_qvUGtbkke!y_fc$JW6#j
zovd~%!Wj%1!YZqGV9H1rg%Pw1{$N(X!;KhaSuD^(3n;s8l~)uJtad};rQCYpGRhR6
zd~lqFj~dYzEA2dqE`RC*!HxUEU|7}>0x6oxM!vlh&9USrnBV~cA10Ezj8%1_E#aCK
zOVmC0IKb3UdKg_`kQaVPgS=8ka#k2lyn$^7Sl?D?DN&R&ghR3DRh`0@1!>X{P2O?=
zs_G8fCLiDlhK=;1K7}*Yo0WPuDzktx8N(40Fuu;a{$Spy58s_Yii%I&JcxeS@x@9#
zE4`FZ^bPS2{E|`s#&=lan?HLmv|HYGkB5MU#h=3vF<`q7R305nZ!E?Oxs*Nf5RVN(
z02t8XZwAW=g&kOwU5t~D<^-stXSj~1;5n~Ok-88pw?)1QfX!gB<L_U?vx|PH8-Inx
zVRAsdoMFF<vtj9>j8-TD#9?K}Pd@&`oww@*TlR5W*)x0~&%(+tr}R2La3c@caVOmz
z(q32`Ry5>rd`8%OnP;RE!EB2#0U(YJ<~tnYFT#aC;jqOof54+pDd8M9yaO*-ag$v|
z&E?IxF6AYh(~Ec|ZTn#-BHywEJ_w?p63^c}1|DFIZ#@QA0{8={;g<bu<E<z77l)XI
zFOS`AeccZ$4)!N;yOEEu^Dc4B2VmhpNK{7U%Jx|rE^~Z?IN;^~46oD`?}&qmdX_uo
zpCyGqzkK7K1IySi|HVIkJOSU~^*B4c?WW9TH|LKh$BY29)o_{TEr;PdfJ9T{+wRz-
z`OWAFzlFyrB9NY!hX>!8;_JWt@BY)TNW0V{_OEOx(6?y5W$@pd?@b<4zk*Tl5{DMe
zgp>bh!W!36Pz<J5X?!i8l{kuz$ur;G)!6do{T!>~J8wL>v642|mQsClHFZ#S$-O2w
zCSp*2#Y8cD<&EFoX(0ZIJ0O;sd{JH)ci@?)lekw&th?edxwo7K;?0kHqcoh}q)Yf~
z@#8B??9yOI2a|wI*jLS?dMAK*_+##%>h(N14}Xko!h|>vcUbt&8KsK?h)Lq>o6E#~
zEbPF7cN+`_R(PGC4}By%3O4vwx;YkPu`^?1qhkJdIu7L%yb$9^EM(FfK3jYbJT<Zk
z6L#>POeiv04-d%%BNM(XvO^|_-8wOmzqS@fi=iY#CSsD(e5PQs+w1xg1tE(DnG}xQ
zL!8h+y6l#5pDdVVF`SQh?473G*>M^QH^XzuJJ-k$H)-tM7XweW!uN9Q5xhE++B=dx
z*d4&Cf5tC%jWM|!M@L<omYDov2SvBVJti&J$;CzNtYXFfU_A9?Gsj^eM{%6edH>uK
z_2AzMIB-aoMGLUsmgzE;#=<k<P%<SGY)aI-A3~-REvp{j=e_GobcVkb3F7B(RsT@C
z7hej8u~3gmY2gZm5ps`Y1mrnn2xtk}Q^~n(j4zf$l^JLNrR;_2feAs<3>=UzKUv<z
zpZr*nD;Wa0DwH5`q#DYokX7KHP!g)7dP673Au$OoY!ukayN&7q*~fm$6h{HBimsqX
zCS#FZkiVE%q^?+zkFv*#O3@kx0T<@u1%p4*LjGa5(`E(QhW0{{)ND3=hZD*yCN7B=
z^15)Uw4v9y`)bW?klk`9F7_YordmCeUbH{LBhMl8gEMw=vb!l1chM7Zyx;NRi*$1M
zQuW5}GURB{fzpJAp{qFNYhC$7ajrZH20J!6HnH95q>dgY>`8FI<oDTG8mS*BJ1J0t
z!Rk!LL)GQf>0Dh-(*4a;g4<Q9Zi>b?mqp7Q9o9}yAMdH(VDcHonEET_1zNaOpTT4{
z?UAGM=xdP2Jx3tAfWA<8`A9|VdPDc@KswF4eN-^hOB^e<skXc7qw4S@F7#jS$)T;I
z)B-N)0~6VsT{<iTKv_kb@qz>tduqE@Mhh3<ia0z}I|^eIYnRhu>YW^lUIqVRoVMF-
zg$bn*0a5U?8;@NiE8cmcIufo>n)AL84wvq)(^f@&e{(DC314qN-A|7n?|R~%1qm#Y
zpx=eIY=nQ(K=}&&xL;JerheU~^jP`TjNM`)Ww#CbHE5ka6uLCM%Z;|cZodoRp8MQT
z_MxDNyTf7^9g047IZZCFy#uH}zDlD@jW-%gZtj*;U$<7!qhMkI0`1k;aHt;XU!YDF
zTJlGpQS_1r-&LeO8rmi;Xc>3UrO4r7N%X>a;c)cf^g%1B)?k-0yX8WuZ>5U*McOX+
zWbU-LeE$>WL;j=t>nY!_PE=odvp9fJ6D3Y>Fj3pS5e+eavX<(?d8e_Fw&{zzEiXjj
z$Y76mqRr{my~@sUbb%-LJAxH+D9diE^o!i}OFLmFFFUqTi~xtZ8~P_JuoZyz$#`9>
z%HRDhiYR3msQM1xTy~WkWhO_QdgpMft};R+6}4gT$xhUaBee-uVMhAg6UZW(YHEr1
zXtS#g7|IhghBB5p856hbM|~w@1GrPIF;1}?i6bW2i3I+r|GwTwr-O8K*7v<!j7!uf
zO1#~jj?2sutl($Dk+!NMIu$Rnfu>dO;G4F~B53nN!V&!%xTKFn=I7{A-|yE_or)i7
zZEb|oOU<FadY9T;;*%<>yA{>w-F5o-Pk&6m`^|4PE-OFctG1ONde4gN4|4nH4}VDg
zvlAwnQvc*A-CYe+m;1sr{#M0XZ76+|k=#GGsrqUNKKIU{sHg4Gr$VgMin8sywI0-e
zsX;5;a@z0K)sE|mhtlg>?P+ZOgtBe7Yw8o5-p$rlA5#&201KR=Xsau&R!w8?!&O>V
z-+(f$y1|&C{!Xci8%b-bSZHIS_IJ`7t6s#1<@R{mpQug-ZpWkxkEJY{j)-4)*YC~0
z9Am}T+iyLa9pU1Zugd0aH$8Mxo>*Z3hJbPyUiRb5cHZF^cH#vbhst&{9fXU2e}^A0
z5t`mHSw#GB!{Uim{%j97UKq6nyD?$KMTS;b5jLmA=j?H+cw-A#cC)9_5HQE#Pj7XH
zh7*%v#3Ns_Gom|M2p~=?$B8ExOx#JsjGyBX39Ccnl*7*-MTmGx&T%7x{yl2wjSGMN
zoZE>@j!8o<UJWE023UqqJsK8yB5(P-*zsrB&2BEI9^+}@4PXWLPrxHC?`$VcVCmas
zu>1TWEm;DlLdQG1>6zW#rjB(PK|sRoy}aW-FTs{Oom5fyFfeZA!HmRXbx#NlFcvV$
zFu*W`cY4#gag+cT)e;xRrXnT>;e~Xq#1dTP#0m+YTduSb_?YzbU0dpftsGGY%uyr^
z<OnUSL}PUx%ogXEZ-@!vI0a>1Lf<I$k5AL^te37eK(bp0C5h7zqX09;!+_2%o_h_X
zl!e_}_E)&;t2=knR~NAg^-7F|xASUn;_YyHQ1+o1a$)&(CmyuHU{w6<jP%FKjb<(F
zK6#w>pFXw#Lb(*ed4LHfECPX7%nCn-DO}*;Gum;}A0&dXh@uIh7NHFR2*HFE8Yt#q
zILuU1zXC)4nK*zU^V#}KvD1nNfebg8B!Ra8jXP<@%AQbQQ4Z(`r4jWN^+q|vkX7nr
z&h((%3Z{xJ@PwuQV^y#T+-XEvd4R%?1VZYC^fJAWf9Ndmt}vouATan-$3%-VXympJ
zxK|qmoZ|Z~ItdgU-yxtX6U;ZvtI9@ORy1f0nnOULeUWYm>ga`0@SzX{ezbkHpxpNK
zk-z<qTYULDp6th8_9rgxGq;6rW2b!c#<=a5KfX(tzy}WW4>pb$?L6#)={^KEoD?p+
zf?+0%%1&8J-}G9)!k^sY{yCN3ZnLEPL#|7i^rIKwIbROQKNaF+?B<}^pYbWeU?+a|
z$L@>;UTkn9pZe1)oUlZM=vR3YE<c}V=ZWzRd!)mIvLwPrjZ=R8+HL;i;0x|#lDG4k
z@ogY}@s2!)m7g3&KM^O!2uE>z;Kmy_(sEw;Fb;5oX2Xr|k^tezb+J1(aU+o9cx=Ka
zg6YvWupEXrX_qh&AJ0$#&fVsZz2Eo-$07XljXhAE!_4ya>2Die`LR=R=ADTm>Ob7(
z_?p84m$R|4xKpQ&7XD$^JJ;b%p^5`xO#5%Zz@1DuT=~v+to-E1ehR7-=6?>3jbGBi
zSN2i|di-)8f<6oE`LpnCC{Qfrb;R(O6=NDF4cthBKl}5Ja=bkF?i64DoB#el{fakv
zz>qs<d*S&)^Q8iPee*4Y|K9v7Ho-f7bp!82T=Ob!HSlY!5a#%sP4{g~KC+U$<2$*w
zQmw|ZWL5EKOstFN7VpFave==tp#iV5wwx7N_vVKQgJO!waE^*%G91N1Sq;v_KbP$@
zdCIEgOP=Y(N&>nWCO6gV`)DXGjYo09y&;#}S;7%}${&0l3!0eVoKB`*)jvd0brJV$
zVVL|!mSD1ywCTU#2id7Y9OMvW1da=1XAAPk(cwutJ2~?tHh%EF>~w*b05|Z=J`PMc
zj9ph#+u?Wd*N@*JUn)L$63@sHjMd6k;NN|ykOvc`Z^9M@0{8c#d?0NmOPSQ)WYxFw
z9x_xM?G-#WtLAfB(1YOO=DUwnUKZ9du@Anlt<1WxlIiA#yQtzWG8B1)fmg;LV^OBj
zh#hXXmY;w{Z)S>9nZPT%i=Y$YF(F!!JjfAB?4sk|W$K5C)-quYEOzO|*ckUL!Na4}
zL2f`P!)p2~!DJ$zJaIf5c^nQ0&M(SICeSTEP+lf$H5PG1)9ph{Hkz-Ozm*A!4^1=a
z=1FX1Nrgx0G_^vC$;^-~z?GFz!m$-YDxVkc3AgOxs;pB7k};H5-ve|(e&dcWCh0it
z59J3F<~-92=ow{2K8j6s0}VnC(Co||c~tlUPZSm<L9^X*`=PBsJF9Cf2!xL1CS39$
zvsfC2mZ%d?$TJCyk`AQ>$8T|L9!d|6%M95_AA*a#wPrcSr&y?tX{*!)6OwQN;Fx+b
zPQkbE!*PFBd@4N0A|j8nLy~ylVl+{kh92aP5&(IUGBc@7J-Mw4FU$i#P0SCFU+8MA
zI)3%ydHTcef1UpHhu^2;lfyI|^xZE!sQ&J*C+YTToR)4c)z{pmR=b{_K6#Y(cDGY=
zcHG~l<=MaiBqF&_`KQfse4v$@(5{+GQ7VvECcL*Q8|l$**NQGC`#DCAJFeKZgmU)m
z^h9~M5{_we!jbZHF&U?Rua|ngzWWp(?In6ySDS)vQF^gEuCIQnKNzIp2nF0tS{FTW
z)G50@frVm=d69v_G6{d$i?Hl!xp4m(v_$<CUKrUQn)ba#<2VwK`@o(&+D(u5cGBK<
zC+&3Gp2TO@+9q2{Zm-h)4Sj%MU!mw@5rOile)v!S_<JiE*wt}0Wv2zl5uT;V@HCws
zK39L;PfK^#>G{XMPrvzJ|9|Pl^N-WZ7ax0v%P)TVBk%HhthCwvNxjFtKWY!^XKBye
zGe}<=y8~<PC)g#(8=A+RxdfL<xS+k%w(6qm*q!Lzil!^!Uu_%NpEfj|s=bMJI6jf_
zBb3rN9vi?3xUvGb(ex3FD8e>HBcw}z!V!IIE23}kc{xq154YZ>q7Fed$Q|-(pUfwz
zo_uwN(&k<~%RqW)JB%?9G3jEFGsNe%6AEtT31VRd$8$EDEh}lc3#!|NemN3Tw5zmy
zMBV0!m6kCV6UsFDt}!$sx<ElUo?iN%tPejqNFTpAOsAr;mm0gcZ;E!!JQ{6}{)zSs
zF4^tJq6_lW(|E~oil@C{Iv-A~_(K86Yz@aA(!E+55ZO?ED?e2(`YbDT)i<DQ<wP9X
zHT@<Hi#maJQ5LaSf+Hsh&u$=;U(7AFnzhtco9(Lnv|airlvpUQp$(7M3RmWk;+R0@
z3*%Z8`eYP{(KqHkMld+$lK9LWvfvoS7x!*)<RoQ7xy6ny7OpdPA=7MV>_R!X-P}s=
zJljiev$%YV@hcXkKu0XtSI=*{qy5u2GiL>oZc)w)7JUs0ojCHBI%MGwM`~`R#@0sK
z6#kY)TMz2LmTuKvMYrsPV&|LZPE|%00)anvsH0rwUOg1~@B#E|#=X*mUQi5kKi$0^
zc9Ee-<tRZG5ODPgL{}72DC4+$b9A1MhNdsM;V#bD8Acoyb-)p#*yHXm>XY`wj;~GO
znWGl>yRGzOe>>Hc7a>=55<G*z%U@wscJJ`Izfu42p}L#P^pqWl>Lc9Gi>^4flkt<b
zbf*-#H+FP&Cm!$8cfK7Qou;Fc=xbMLW2VAI!)Q?qKCk<JZ~ppBD8AV7#TTQqDwU2K
za&{HFvxTviodMAP5r&@#8v~<b@SoknKf=r>2njWZgAETlbY_`6QKrX7ILXN&Z6}<*
zvwu8%lMeqdWN%KO#CvffO|Sfr=LkQC7yfe9AMOeh7W*-kY&Y2*j~-fW!r!vb@|@#2
zU5AsqAgXu~){gOxAn^y#Y|C-<=K%A#L~sZKIX{FUy~tO=6^|!LNyY9tjQm|X-VUoT
z-X)&i01p_l;-0@1aAA^F@#zl;r$0e<_!2fAbs&V}S#Xa!BrgFg9(^0E>={-#h=rtc
zRVTNk!RA4g5JH>-sIx>M)qtYBF&J%V07Ho7+p82cczH#Qg$8y0>NpuR<9G@&jOysk
zm}ubQ9fMym8NgK%76xwADYA>mxKhd%RP@BkseIf4gQp&LLRAEBQ$hnu2v!_$cFR3N
zn+CusXS$04n6Rx@`GRpruxBNuS8i$0w7>u-Pz;|M$gK#IpO3*n`K5Rk%JtwLelbx1
z3=3?=V}@and|=U2l91Di1j)u;Cp~)WS=xL0I5os1Y-qq?khF3|<-aAb8swmH@)OJl
z15t#r(npC{_!Qn?JQxs(OWG)X5Q0b+ff|O16=UE@xixGUKqe!|HuVdBpcz)fc$bsx
z2%FauRACTcRH=7>GEiGtsXUvB5iIXCqb`*JCSW3e&=`3JN3p67y5a~*3^09GJK>cj
z47BOVALR=n5Js3g5S(Y3fW@SO!s#2v3z`WNgDGhOz-?0wdPTmdH^oQs1tUV4sjnq5
zNj~CG0awg0uUf#2^1+C)di37$E%b@nd^S7jkO`$l>jR;YzMbZ8{kd-;C||TgZ22ep
zk76evzl6gH5B&HKH=)rE_VeQq`P)4Y3#%eL`OEFeSyY&UpD-nU<k1CE0)|33ht~tX
z^|rh0e&@0y{5r01>^vxbyooM5Z$joD+%l%Zjqvu;xBdbLB!dl@q!!^FpFh2U;W&8N
zt}Hl+9U@##+j;R{WSjT=(Xaob97M-GJLVTZ@ohLxgSTGBNB9Nqu*)yw-7fl$I?;~|
z?heBrKxTH_BfQ=8DlzO7R+|V<fbhVM)&97LV~&^WjBmU1A6NyCEHH4WMStuOW#$&+
zEp-xUI1EnufdRHZ_y}T>d!!LhcAI^bn$X)o;&bMY-PYnVx7$DG9}JN`AGpoK$Da*u
z_ScW{S0);wyfWk&2Y&@*`^$ayheK4x{M-NW7azb0x8m~}P2z7$@s$OK49j(Ie}6bU
zRxePIi+`1X+YkS8pYaGhB}{(ol3~Pm_`K#u5Nysz*zNDKr}X0B)jz4?i+Tw>#6#Zx
ztoZ`(ztnuo;J-KjO3c?)a=m5@LO<>U^shDy|5wR71eoOQwBr~oc4e_LoBMb;?griq
z{uN~fCmL`c$L-y<CpcDauhQxRd~d8;hs9c5-xRO8p04%5L?q!U1AP!j*Kj8l_fg%k
zgW+l-{+P*ZR(HemZ;78;m;YrtIXzF4iR1(285o1!dE%a~(P-+U_gIb09YY*vh$0KV
z7GBVkh>`)AD1|3Sp~3<tCIR8+;EUl!xYMTFZKYbH?r?+iIC2W!G48Ee3t2<uL(v4E
z;S&RdH<VsX;=psm>%lu)ndQDmJuBn89&pVLl-->@Pr6%CrZ7yxGg(g^urQBFS(I<^
zk|-Q5E+-m$9#TVjL)ie2Jsq>li~G?~l(CzJqo05QZxzaq4Y@I~E=Jh=FtAX3-Kxyy
z&&8)INtFpcIr!t?PpKc-<w=~$bnJ<HfiX|ZBo{P*VhS1}o=@&j8Yqr<C@eUSdz&Oz
z2p;jFC(6i#D)j>n72RRL1AIMl0gmwZ6V(Ab?O5f{ab3uY$cY>wh^%*u+=HB{6gg1@
zC5z<+<q!FiIzbL#l9GvdWXWmf<4K!IIPdg9At<_{Jn-qB>=b?tUl}1hG)#Sy1~QZv
zMIz$~m*g2)kI7NOf>Ll-?i-UNfgobqhdw;vFS?9fc%l!%rOrGdE;*HS@PqcCVUCAu
zwmUJAYlUW2ZX7>Ymz%;XF0d&}Ot6Q1gM5WNM_#}^<u~0b58xaH88R}Hsj<T=CSQxZ
z3M}#u435$p`KVBK<WYaiZVQfA;3}@NVHix*qkN@3(O+QVNHgR_^%-VD9(*(>3tpKV
zCyk?{Ln~A07p%mpZn_?cb8~Yc+CNLDhaab-mw!s5!Kw1UnVvp+l=gSGg%9OJ9SDjW
zcoFIcnUV?ioo&g`dpka&kBRY64pvizqwT6FV7Mo&>b~Hex1XkGPo9`2YxRa`cGG>v
zS+Adl%9ro)+3s2)%MqN^7fc&UhDM{6+AZ$&swIy8Lhhxlq9A3W{Yv;`vUg21O#RW;
ziI0MPFdV9{>!qXPGapIU>yOg;=pu2i&RKtuPL)1%NuT1oc$79uPmX>)RlJw4j(lep
zM~k6Q09in$zW_Hvxk9k(gUM~xoBHa9TXv@2`gp^$lVkO}*J-EQasScp_tN3P%k=Sw
z>?k=@8|<Y|KKy+;Irt=<9lw+$d!M?^dRkq&P2*wDcZs36c>jm*rl0=g1MifnRyW*_
zpy=ukMpn+mae*j`QG{^JEwoc{JE&-k0SC}<$o9+$pfE%cejROF{mC*j_)3rd>{{at
zN64{b52crPy{a6gozW>EyPN(514Vy~4bUP>J#N)Mt^1BNb}`i}tKOYl5srLB+*(Ro
zwl7-&vb)nxb<q)uVftqJ8S+B?V#<PF8b5s#^Gql+pkMktk+90IenMqiTB48P1e#kb
z${CC32T{tRIN%<px@g&Bp(>ENw(>&xg+^%?w82Z_Uof}By@a>E3+a!a9Hb9l9H-|8
zr|G0W5xuD`3kQcsz4YqnRQSIT-7H(dOS|gz2I-_XN+)N-^h$Jha;7#edb+&2RoH9S
z8_G&{6*7!%k{|UQv@0^ryx5h((%(`h>I(xc6{Rf7CikJrhm}{kk7D;X3uV}KNSmUp
z)E~GAg_d{@l^uD9ewhCCQhgC`c2C<d#wslGC}TMbK$!ny41oUFT{#|U3{yU09xF#9
z4HXCl&#q|p*`scHvfW5e_t>qt>tiCBC!&r#hNw-ko2n^V;D|=(i!n?u0|Pg}^iFW$
zmwu?quJ@{*%9`j(aa48`IU0o1!}Vp_l+%iM3dVJgR0Nmi3q;?vNBA%lTf2MvR`f={
z!9s$y)b4iE?*6{~#V^o@TPeozwX3QFl+9y}%Qr$DJGz9+j3XAiOgPS1eXfX6`DUlK
z+JJW!s$RvL#QbY4l%NfP*Q&lZAwsn~c5`u@Bm4rp1v$QzaK-_=pvttWHs(b(!r%7p
zp8CU<#sqjs@&{j|aS7f;^+9``Uf+pc$LUz~$C1d?7ktl_aLE0zj2o|;T#+{Tz3KnG
z`D-&)e7*D5Te0Jde?KtNbm-w=;a-$Rb;smVanr*eb~=Sc_iU%L;0-bP&+c|M9D>9<
z?D1BZ0(bF`Zz8Mnjc^9yMy3zh=pe$)@n>m}x3D{Z{E;vITrT}NQ@HacV#L9hR_5q#
zZN$QjutX_wc+)`ic2c~t*~vd8EV(1R8<^hlIE4TBp2JmYgu|!2b9wmopg@k)0dv$I
zZjMJt<u?WmF^<?>4&oBoA@YkiZu}AkYt(;C5{7>XJ98&J{ei9UoWCQTVsjMP^cR+}
z+dbUarDxZO<S-JERMm1<C%6owtWNM=<plQ@oD_mB&Ik`E1Q>*1h!O0_KOH2DH6cBa
z<&`-wE-(@7WMU^7Ti;m$XC)OYIvhWyJ`qHD5ZXY$cL6D_4KW)iTZmH=Gh2}WjgW#c
zqhI|=Xys46?S+Kqvy+q5V^yaZGdL3zEF5_l!eJ;bP}=aojPbnQ;l5$)t)E#qF>s8v
zj!U`aW}M<>Ob3(Zu<>nU>r&bf<Jx`pIKB7dpQLww@LuZd>}XJuKq0)t2vRLm;SFZS
zN*0&}Xio0jdzQfnN7SM4C!{l(4)aCbctxoON?-JBrDNb2W|38|P=j%r(<&>MZdEAK
zs!QlbxTS0er7$<Z;8+a=aVCp7vW8<XDYp9_rG2Y#>=5)sfx^OoaZG{T^_U)17oMCD
zBL#c}HQHt{B9u$81QePh&3IHe7g7aRAPbxdM;Q<xK|d>ns3RDNSP{t1w-E4v3zKc-
znm(+sQ&<aLisx0#awi`vVs7Hdjo6_EvR5@oQ$OU1ItPEy-7060n@OS`qJljYP4=)+
z6#qrt@wmmy;%|H)bYdzFRv2&m%KP<)7`W3W<89b9I$%VgaJL_Sa?%sYL_BiA8#jjo
zy?pDBFizXHxu3s9Yx-G88vsFj#Kld;@SAY_IX(Tm+}Y3f6_7Zv!%^RXqktWDGU$HS
z0VB@=n@`Feey*2rG(5x6A9ng<=R=j@xO^D~IpwSTgo$rrC@K|*jep|XFMi^sx9b`!
zzcIN)zOySCnfvQu!A8!R^N;(?U&emQGqEVI$6Mf$UBcr|7>AEA@d+%f0WWO8!H-Ph
z?||8@SR+hvaBO?z@Z_JeI}bTu5n<t1?}hl07i@Nlx`{x!4A=oo{&_`s+3?YC#*qE8
zMUP|R;D6!>w&UU*?|6z^xR)pLL><XJd(GVKXV-Z6541VY1<zq3Jh^lm$4?f%3oIiE
zJGkT9PsFx6pTcLv$1T1iSi~!s^Ed9qbM(?i!`*g;&()Lj<p_Rq&I0AdUf}9&qj1zk
zgbzdX_z9f7c=a7BzW(EX`V|=g1JP2WoY>3rg{BnbTQc7=`0vfX026%RmoNldASfn2
zi$igqy8t&}^QKksFjmk{1BwX|jtk<x7L;DByk;dl$6LXrNB44{)>j@<Wpyd7t8ehJ
zIO2y!gMPZXnD}_Zirkq@W5F4d`0Pw!LXL?rCcogD;KkwN$O*jEt@tbWEsh6bl4o~k
zKUJ!Ybak83<nlJL!|SZa-CSewf+%jT&1bSej(&#mpULg%WSlN07vgL0EH|J$Y6%{@
ztC%n%jCV+hkBeRB@RQ23c-Lm5p1R$R?<!_8&H0626>Jm{tg^SlM{bsHay`MfqG;e)
z#<Iwd#d+|8Oo%c0Z@EL|<v}sP<ooV!S9p0)IWCQRCjHoTP~!V;F_mFB7^GhBRCu~d
z?67IKyUH^=;KZAXKc@`H02~X>1UQrZ$Vg&d(~8_t&@kb7G2z&*EAz%{@C#I^(&s2B
z(#MpAX;x-1ai}s_VWpy8Nj|e5eX8oviWDnUI8rheDl+bLaR31S^hrcPR6-)a06dEf
z!w8nrpj^(ka7Ue)kM|BW>K!?TW0ICVc}H04h;qSK1E0EOr0H~2mnx(E6%SfsqS>ci
z2r#>RPzJH^(F!l~-^eT>i*mOVibv8w-l0vPxIhMCtV1p!9Q8*%V31X9RG-C=H`2#I
zFP>Oe*_mhs-{2_Z(^w2h9>6Phlw(W~LmS8=O!#xB5cEpD$1#E^#Mm)$?Oig^3FX@9
zZkvvfzfr8x#$p#<95?34UX=lPi^)^&fAeIs+6rw3Tv9%;Qtl{b$0D#>?69bZEXTyD
zie;LQiDIQF*u0U;kP~@KGbj**Bjk9LqwM-1o#@LbISN3@)e3`R3)zh~`9l6$U*B||
zjb`;n;D*UxcC|EHo%H0bchl2%ewf<b-PG#ts10;{Ul{UWug_gp=aSV`)=PFmjnlng
zk$>8QC%)A;K_BeQ@I6Ip>%>P{Z21%=hH)wU372G>$$ReHOG~S1sy;{A@$QiRVCbXh
z2m|io7|}Ssjy4Q^?(A%*ZS^H6bF7S*UMkL|;-m10_QixUeI#w1yn>5f?<~Fi<V89?
zIZ4Ce(Dlx48ScP4SKr5xtuJ4`OfOy@q?ZTBsdsj6Kk~M`UWvs8HxHKCsWK+8sb~6K
zlosr2L@`5;ubMReP+qfWfchank00%&W|L#_x>nSPQB<F}<#t2EW{1){s*4~0<QM7j
zlc(v?-k!p?)8qZU^phWbkluQt_;rr`6K#opQAihB@cfPOC3hzGm8bEL<7`z1m7jKl
zVv4el=<7vCmbvxdPk-aK3>}F^d^LvW3)scV8)cU0nz;wny%$N-7s`Z^1w~hke)QGs
zn&U1;;6lGK2Ng#vGH#&=X2Az-isSRPyA2;T$8J{g0VMh-A8{yuj>Gegoh8N<ji*~x
zk7rg$XiW4yZE_>brqTfC>;ODGI#rz>rcYkHO0N!%<i@VXd&OB#wW`{p%7Db`1r17=
zw$A-ov7?cD=Act{YOi<<q1-4GP&5uTcAl_6MEHdMP&$I&!=s~gptLS8af=;azH3qC
z#|@fcca+<({<$s#8bayHj>U(2a1x4g=!A0BWZzH***S!=hy^0CLoJly^dpp+{t~dX
zbNY*^=x;a@9VpKtGKGb1S$s{Ob(yAmplKE+c-}^3rC()H4!jCG8*3<y<VLzE5r-=8
zU^Mpe#OyRg;mlY-y`bbm!N!~W%Gkk<Qj=YE99xJR{WAFU+@c!GR(0LSO?I0VzqxCR
zHqZPQ<0)mfY^r>t*rASb<3W*!@(V>Y_Xb)KqzA<iis+qIE$#2LQ%n8QgXsKNb#pl$
zr47bW#n(um_IGzuSA2u%8Gc9Yn}dW_gkSE)x??<-dqsGljWX8KZlG!UaQaf_j@a#0
z#(VnnEwyWQ@3XjKOJj}|y6Vs9x8gqd755P+D1$GE9ha1waRM5p4RTa5i+`@w$J|WC
zhm8k{!wzpzuK1sp#;xwoo?xz};pEasRu4y4sn30%>a(bC6rla#G`%`LPiOs+z8}0h
zu3blYtTw44hBUfp_*#6vjCnK8;`6=vr#4o6z4!Lpzk*2<t%=}N>9W9}!91Fw(wpik
zc78(f5Dk$Z$QKxnX@3X^CQNVr1WfE1c7gBk{FZpxy|^hE>~0W-C|ALYc=#`-sm_qF
zddCDFcHr?14BWyl=N#s<9v98!6ac^UoBz~&j3@caO85Y#ZyBep#~S*a2Oi*%Pygk#
z?M7a5oRXhNQ#d8heB-CTS-BJ+w}K<5N8$v{-(<(Z^hdaW!OMXnjGw?ChIdTYM%YLu
z(k<m${C0Sk0g(C=J{ZKX%bDMJ6OYi8)v@I0GKSsX8MgjOH6B)KiLqH$hqR<eJ-By%
zAyAf;Zx%+(T-^#^Yy%f#NynM<4m^Ypjz57ZVf*d1200W?RxGHqhB1Xn^!d_s-s(u5
zC15z8%8K%C;Zg%8JG<7!Tv<`Zg!#Rg95Jld8q9Ar=nwi5Bu~y#|CDp>2WhN__jI7n
zS_35l?3D&i^6$K}H5%bud1N55@(Zr=FE#Y^e!F7gp-|K;mFq%2q>2W`$M3(Je(<xO
zrgz_aPXkCNlqw7;ib9?cieVtAUsjP?i6lFNHh6|HVl^I0K`Sc-I#$GhPg<^w!WX6-
znnW;xDTMK~U?k>^Jj0~ih|z@U-&hmO9Nx+>)jNX~bOpR%HaUY_jDhmS3RCJCss@K1
zOv#J#;8j5?0}L+kC=2jlXgygWM%hdVdBwOG`cXJP`a;3>|7Y*dnxsv#`#$jGzD35C
zd%j!k)zw|Spg{tZfglJ_MiVt8F{1Pl^cAGJ(M8YA=!58TY|LaM8=H$UQ8G%L5daMc
zG<vD(THbn>d~=V8jL3b}|Mx%sL}nFg>!k`Xpgi-5#~<JQxcl+r=lF4_Tr7BquP)?1
z1hav<63K8W^<=TlPAsB(Cy@MIe-#Si3^x`eWMhG=Qa~#)FGgo^On&rH+Cx}YV*;TU
z3sDHYBpNm5^iXH|Xw^Q9HT?La58o)R$GiJ9V^m@EHg4oi=6=$SAIQeACOv(VD#_aB
zaHK;xg^@9Nf!#!s{{6D6-WjNhXFFx#&uL~HBJl?%F;bBq`8uuQM=<J}ZpEKJzDa^V
z(!A^EzgvHK+W>X{5tri2LL9pn7!C2C{)k7~6%ND!SmK^#!i~OC(j*)f;zGQd;NwZy
zW;<kD!eks$CQM0FKKS|HFst6>ud=HGF6W&`&PB*ZEAUBdTl_O@AA)A^J8|=&oOa6s
z+wOAnhkJR3{it%+->`qA2bTQzOFxD=&gtlKQXt$DB>dYU<D{W~IG(~d9mh?3zU|@f
z44853;;{L+77pUT4ojKAW5fF3=2x~A0W-e6iaj5Pb9)e0ZyaoQG{5t?jl+YZ%R=3q
z2=R)4eyBWlD&IT_Jff#PU9pmeL-1qwjD=Up&S`tjwxb@u=n?rY|7l8iKPpI$kF=-1
z1K>Igm(J;)Rl=Tb_-0tXO9Av@)KTJ8;q2!=oPOy#@-+*-{>_glK#mrfEZE15ygto&
zXg=3d-X;ELJ^Az5<_o;PF@Fz@X@p<GRQO=nmt?8%=B;buM$>wH5eGlHS@T?U3xXV&
zMN>1=c4zlC+BD6^ycv1Se;1;YPC3Hc8?{35Mq+MqG-fA8Vj6_O@m`06I1$av{Sh3e
z)lz&ID3g=3D#N&s74jXpq9@Vp1xcE*A`+UN{2YM<lQ}i0DD9}$1L2G1Uu?7UyyGL7
z`kcUUk-NyD(O}Y2UNoQ(LO_?JaRhS}20oe?2obpR2js@?XBapzt19gl9t2}(f3(}J
zm=x^>9m?7Rm{4i*0*B@i=6UZ5YXWbI6U7L92%XG&!JEAV$JB6x&J(OzkBB}*s{}f6
zzFrG1sUIJ^G2;)09s)ed!~>&VOBOS*V7Q`<0mBcNwilLE=9*yuJvy%S7U_HAn#u^%
z%t8oFYOGM8VWFux(&$KS1QQyjF3eJ3=Ki1uHyC0to?%Qvn_EC8cXpMtNrJdM?puP%
zrSgK7hn_~GiT=VGgE#3ycVl-CRM`-`0M^-Bp147e1kU8i(ODeB$1FtJFdcbQCK#3U
z8<*?)GB`F7#v=K%!ykBihrHH}Fs-4z8%?xL=Dc2HO#q{hywFZ#9fweBM(aw-e{_80
zHPqz9l;cwF6!x%2D#9=7q=&UAW#c`seurjY9+jurR{r!8=q}a<xlYXWFXNBAoF3~4
z!H2$zAPUAVeGbMX>s&OB>Sz}+E(H;o%`nF~(vvP}4NUbT`pK!%$UCxN>S`^eB?ydW
z+AdF2DW_8?9RxaF2T>m-<fF#iKeb*`z6kLq1#7fGzy-viZ8Dwg`>f}1qwJ(zX*seU
z;y6d{RO0BY!#?-i9a~t6FpM=M{QyCmI+Xj2g=F;qQv?ub3!KCun<F*&Dd&8>8S@PU
z(T%toofIBXoo>cCU@S(}R<pIH#!Fo=S9_`MM}ux`JpU-3KYJ=%(N}JX_ZVLY!t<!r
zNehcvFVla#=2RKj9LgqL`VV)1b-MlNb_YK0t=rE#X7+ok*NMiwaN3Oe7UlC%T`D&M
z0t6oj{3oZ_cuN~2Y`Kb_(%9bXD7^y<n1CUeg|(vvqw1UV5$+<~-`k7*ogKpvoWQD4
z9m9BwHsG$k6NG3-r%rcgw-Y<Uj~&?$PdVoA+(J18SG-a6kL%bM-rU~ZGals#PK2|x
zy|r@Wf35mVPS3^C@@lNCEc+-&@F&9lR(mN{*B?fEX)UJanlW8#3GR!rxO_iai|cWP
z#*%OaeaIVAH6P5+Pg{G6O%&D)R{w$5fB*@6%la0a#4)F6-J+e_?RM42(bzinQI-hl
z&*je;s`xF|s8k-{Vc@}-rC)*fz^iDD<wgsQG4&6GsjTl=mowfNdkDA@F1OkX#`g%Z
zsBfcLSNYJIq<jeb(5OL6m32I0Bu~EK?!w8j>$p5Sb{Xk+2ujelp>7Da&~#%C<jyLN
zoAU+|^*xUBWW0kz5oT~~9k>8ZDukP8#O>_&e3WlbeT@wsQ=ZE-FDUO^cO_4X&IJzv
z_@Y^a5D1(|e*%v-nPb#Py`fiQ93j?#F)Ca_TcJe+eD>6@IX;+s7kzz$dK`QbZ~?+6
z1}*r^JPvS#%6xS*>WwHzCkS!D67X>}uPV&^tmf8f1bzsH(B8{A1OZ?jVW`GmZpzM$
z&a@qLitklax@ZkC4#4H`qQDs%!p6N`)9B53sTtP5wr)Fl6YxS~AO(JUG)5}TI`BV7
z2SY1xH)*>*t2V0XL9>0{cTu8^g+>v51rX=mdTjPUAUqi@$D5Aktj51+AdcMwK#)Ob
zc1_ET`HFFkkcJ05P#u$saQ3j@i_Xq=boO@RRQU3Yety{3TJ0hhTJw5eTJv)A`KC4M
zxTEXz=+I+;V=1{Kaa8FET?tkfYERWyZ40E(nmWH!d4%8T|DKbW`}78cv=#V`8JRJN
z5EMFWLimdLgbfJBS;~ufg^fQ5(XXx)R(N^P=MGS`p;PDvXGYCo8jvzS{pfLAsgK{N
zo+H9t9IrT{F?T9Fa|S^(J}nqP*A3$7iyieLmHq58I>N8qwI|RSj^L$(BeV<8+}_Qa
z#+~|<Wf){zjQfR5#RJL*m*Vl6hCrY9d{O*w%vWs?e2x6GfAUY%#<vZUOojYPn+%Pv
zS9XT%0RA?zgb9+SYEt{9^)_CKbJtWXQ8UMezH*b5I8}!$u6N?=OZhsSzN!${8gO2;
zDUQ>|Q@$BAze_X|+i7uBy&OXJA#20+b9#JG7To+47&ssDEP?PW?lyKy52q>Qfo2vS
zx4%PQhj$WAP9N!gYf4%+w%+;1RN)<8MuRrPWf&$mFAgamVR(_5{>>aMX;B^)2@Yzz
z6UlGqN!)xB1Azf~sCqVd`JN$1%WyXSP7ANfhs6|oO0iKs{J78SM#z4QMVBW3n4O9n
zAui6GMVJLq!lcR06?XJYY4T_{n<m@=AgQpx;H+yjtq?XHaBPXbPeh<GVbz<hsB<}O
zz2Su*3mMwS>hYRTtv#bjg+&bnt+1wbxxD+Q5a}F=!8e2wZ+4bIkW)4)lE)5;pa4M&
zu_M?3R-Cb$f-d3}LRk-Pzbv!jmooG}x1A<zO~&i5z7(&%^;X<}^eE;u@zI9tP;xo+
zSY=%G0_K+@jNEnvq7F_<>1c5P!NEeub<$)ms5v``7HY!+*b5Z8iu|Z6i<1e(<wy>Q
zH+FWgz@cxkh@h10vVy=NK7uOBV&Y7x@J&^`09D?HWfF%Na$d5N7c>N_>+JB#9X3fA
zaUQhG%OXH^s&)$dh{n<el%F~yv_QazKNa_mCB>zzEJ`UK?ZzLIEBy~`E$#y#EDLi8
zT_$tVgb=I*hY%0cfrZSVa}e9xJF4fU84tcxS&Q+JafMTuDUF5kDV1L^VT>_u+1-YC
zadvV}M3ekSR0k0dh7<Krri@|As1L)n8rAuc<pu_BQ&$}qg>f92%3VK6j>%Ww85FlP
z)SGx|DpwgN%v)}icd?{XhRt%_re_z&Wf7&fh#dNyld^xt^lyn&#m%DQUZqYrf7<yr
zjWqd<3uRE8QdZn4q*}`%Ds%glW#(Ps)2bH!&Y@5e;w=klV+hVCR^0MGgW6pwm2%)w
zUb2)>hK*fhCx9aj112u{IJ_tsL)c#qUdo^UGL$>#1j>(FX2Cc1DpaP%e-+t2>1}6g
z{glkWqqkLeOa)Ch$H5P4Nk5J3WLnuBSTCGyEBV~DkPbm{lW`612KKis{NM4)`99r;
zj#xhVZIkVrHn|dPSdsj#(!PbhU5BrHD;4FLpEzLjuB;e734qnG$?%x`(y#JPuo=PH
zq+c4_i^E-am_(Ji4J$Wg$lKTgk17~-EKV4lDbw8`X(2Z0*u}<S98o`xLx0LJyqGCS
z-n#)acEXT2^dWh=uEqEN_rL#{Q~Ns1fA!z~cVt}P|JlYc{tQ#K;b)o8i;MYn<9}oR
zHW~HcFX&6Z3G-$17>3j9ZZ{i_*D~@<z}ZH?WEjC{@`f-rrYRirBuA;SvzcRu28Vgv
z9y@;#LcoMD8i595T<byTz&$pjdYd7sj885ut&t^yENW=YOy$u_6t!Z)9*AD#NE>4y
z5>?vJ-q75>7fG}@`M_Lb7xVze`4P-y(P^p$46r;QVbsifA3LTs1iQ##`f)-8<!ARk
zG$r|)$t^k%fe~5`FiVNgv1Ty9xFf@Nn9ZW;GG}JtrRv0;OfXf^9AeGFF>4$rm>T|~
zwN>8>ttU7F{YG%eV+d0l0frJn1)o$QI*|GzxU$e|dOF{tH#u5tf9IL#$0N~>^P-W%
zaqd5O6!VRi>vEwsWaqwCH_`9)yhES05)58zC+ST+q5Yv<5qz-I`daH380(y@Kv_>t
zQ&R}0Ic)-il(@djM*YESi0cd45e!`#_P~I;rdEN<J;8Bh2ud!@xTRkq{DH9sQ<&2@
z=$qKN<H_Zg9p)M^uI>ndfe1~%*lxu_b6$C9?bknwz5SkQc@x|bMcI3ULo=|UMYwz9
zh{wWF?hYGsTr_0Z%}@V>VTZ6}XJ^mlr;jf#E?6_J(oCZtaAe#-aX1EzKD3~9CBmiZ
zE-v6PfiO&U@x3Ejiy%~EO~85y;hKeUg3H8I3fGpFmORYpD=_mBO!(9c<;hWQ^hb=>
zSOrEFIKp_8KgZRPAG_#*BU%%zwdx4J8uN~8!9b_0elc*Hudc4f%Ib=I)Hk$N?i~z#
z+6QYUGrraLX>7Eh`UuRF2W>SqGiO0J0*GpSP$nMAKv}P@&b0<Xiv`B%xXO5Lm;yVm
z=~)jkfEmgvfwgTECpS3CPw-)Evp%=rQ+8{Y>tQ`Ku5|OrQI5IH@pP@eY0ZxdjvYNw
z9~p^OqhY}{_X>i4fKwi`#t9&&hPkzD)NjvEIeM`Z2i;vg`&uhrcswJ7V60E-Sy{Uu
zZ-4){V{v&^>(@LMm9-qW<wWgvM4PED)*`Y%Ms-yU(LiLp7+(lx*cpEu9Y{MOq&Yu3
z_7P!=i!Jw0ggFS8&~kv;e9%9R&Fx*cBkN%cX~2`JCopH7NnfEKvbL@^{j!=-Agxh3
z(kz+Gv=z9Q{^fBkSO}9^z=0r^%?!Yw{zMyqqb*=P$$QZ_LXSQG-UC;Gqqsw`-Cl_0
z<t43g)mJq>7`HU=#l^My&$-7Wf*Fp(17|EWThVScRXnZn&knVYKaI}bP8{|2<KbF6
ze)k`IFYe!4w}u}ZCnl!U4pTF>?`ph1d$FtcQ4EewtwD64{sphWn>7S9@D*!xHM8r?
zcu`f1XDQf>`UT}T-$roaUTp+ljDhSUxp|hW27nt>DL4om!`)GJkFl9(qUkfwH5yuv
zD=vl&4B&hEE*eJ@qt{Uv4x5_Nyf{Yh&kYuvJNvP*bD-t!bc|2e)#ltAg~n29BO#3O
zW&yP&d;!{acW+l?x#zy$tRsM9vqj=hj!;F*ZEt_yf-vyRqxF?|`^`u3#!D;Fn5A)Y
z-tl~{F+@BB9w%2Lad^tHtXJw|xxPSf1AI64@@T`Y-F?lg$AMPXYcD^HNB7s&FOK8+
zi|sf*r=03Zf-lGDfr}Wk7N&#8)xQ>+b9sEF`VeCfe1Hbp4D^EfJ%Y5^Sv1*3V@hp<
zU<3iwv0<5lUKPoL4vtYAXpX0iReWnKa(^A^G8U{Ms5YYyQ<fRw4D)@!FKW-6Uo@90
zUHEh0(!3%?dC8Bld!vOdeS~)BcxD85w8@OnFIrXzuHaL^zhPWYP9j`^F3D4l5Ij<U
zYljJE`uJYOo0(F&?o_Ie#s|F=f!ByCLcgU@%=Pp-`YXb`7hBu$!G|A3XJ<?ODJ(#3
zs1G7EMJ&~rpN*GZdf@TAr~YwzeiMtWR@CR%kRp7nvd<_yuwYCb3ol<>Lw~96Dl11`
zPT~h3fxOC-dI|9EuRZ6PKM4L&Tfw{I=u67#JJAGZ<~izXEzCpaOtn4mQ(ZU`wYR@*
z{5UZ&<9-EA$MKD=`mEZSyI1!$A1(`Cn*RkK1f(Y#t~<Ts=$&2&_oF$<eXeZMKQ*p`
zFUj1+*ha9pq&~*Rf}QShY;1QeyySYlIl-N_wvbDI7~=(ciQH}t^BFL}@6!x%GyZ3r
zgyA>lZ_fPazxgjmM*irZ{8P2A|J~?x5;r~$d|>pVAMoX!0rZj*9t-cv=7!Y^O&A|<
zV>H1#lOy`ajKbL0Vbjp*`~~vhTThin^_9tF-sx85q|WJ}>27~{+f8qKVF9_6FaPSN
z%EMElJAQ$O?YNf^{5U*ml>A7Wax;lm;VM)7LBjI(Cy2j_=Ld5ql2fSs4I{IKR4m+O
zyA#Jd=ImB-*8i{!<%jUpi|8P&Dte}!4q2NtZfB%V7_>z)q4bv(Z`Z+&PC3g~wZHx#
z4v3jyGCp4>R^{lbI3A&>A4!=cvk((O*EmF!7LOq$c&)kFS~^U23TFhU$zh;LAL69d
zZhAKm1QUx1CR!7kngsKFeioX8I1zDrDZl2zvd=WP8nq@r@+BS@S8^N#XC7ZbxGFqn
zz+a1arQGboI@Lq^2-UoBB`+ad)VSHD#|}y+N)R#daZQCZA;K?3BwjG_X;PK5iBWb!
zK%6l-B*1<HCPDu3^l($Nj!jKObA2gZ`S!Qt(W|d$A=iv6+7oJ8kZ?KFVObEPfd$Md
zBZL486cY&wiHHjVkx89}KZ_28lc!t^s!6_L7BE@faMzh8*3`yQKEOqOyfa@Bag!_t
z@>mcS*a{f4njG0x!wwmCs}UYGXyzpPOSMbT3uh&X5Q(;7(d0a2Sa<+m3+-gssRd($
zF9;hJG-wQRTp@16Po}I&!UV7Id0Yy@J++nU;$1_f4wTPYXZrN+Q^iME27zg2hw99N
zj~#53*8(_%kc!~`uZbL3ABi~K+1iQIBX*9Ci6|?_j8HbgWn4rS0yT&>@0t>EM|xv7
zV^N!~iC`5GHL5rWo+%8K0h-Fnlep?bd!#!i%TF0#9uT%f$@s+Muf)L*D+cd0imPA1
zMQDE!PcKF`LaRz(K8#`?!s1@RKvq8Wl&tabpOJ-JTso}GFGf#Q+6t8&qD026GMP5P
zG9&~};!{kfssP3SN&MtD&uuzxYdhVB$V7>!_b2?a4#lnlSdb)pWwI=pgi}xi;@AUY
zyW9%wNAZ{alWA1O!K*~87yX@ox!+CJb3g&Oe@PoZ^~(x&e=Y6uwvnYONtTbY@um7x
zi4wTV7boBcbZwVomVqd031wT=E#Z7y4jbG^TTU*iB35<_jtJ_+h?n+KBPEYR@s^Y0
z;gz2bP8rhRS(5N~rb2{aSQ*kH$}rq*1exliDqBedcPzuaWMr-2VYuCmTWtK`>gd=%
znM~y_D^4CuwkvP_R$)EwmVQ4B%NLeA#>D{sGRa)JO`E*I&1t88gu2ZmE&0j~3YaQ^
z^c=!NKUE0{Nf6xJ{v;;HEOXX}5RO;O|M$=SB%Qwo^Nrx^8}pkrjGeE(!9FXG&&*rS
zjBmPHj?&w36;>ex>kR}5&?W4+uFd4pJM4l!I7y~HyQN`19iPE$6}}f815*u+Fp-0z
zVMcw=2^ZOO(fWnxH)!H>ks&84!?9?r8__ka>(D?sQ`#q|2wX1AbccxrEsk&q#v%;(
z>V6gkEzsK3(|4h&07nC~Mh=}8G^UTA6CDeK1V%IWU$CPenhi|DX1V_0z|3awAao$}
zBg{tVN$7o;J>(5zG>>MQ@y>7VNSR{?|J0;t+bat@V7Ms(zqvc?;NU>(0CpoQZNY%k
zC|Fl8jwz3|gP1=A@uWZVWD3eb9rJFqnvW5pj;F^5vH9e~*m(M$@5)$OT8r_i*;ra#
z4~_wXX~|J);98Epqn>QAg0^RkM4e$ma#S14)l1Q^90P@Lh|B3=bU`Ou>qc;a!B1S;
z2WB{0bs|R9a$))$8dWi@Ex@`8*dHAo3U<H@ni<BM;sJBkT7o6VB|_M<b_A|7svEE{
z(@$_|YMs(xqoUGB>w||4X=t!;>>do~m4$}V9gD%qnYA1^3Xb&{415^j=NDIA-*J>z
zPi?TX*KvPYUTDTbv+niSUgsdT_V(SUU<~#T4r6b3Uw!b{ba=hdRNlE^k2XCR9JxJD
z<qz{1fnK9gcb`X_iZvo_$GBkqz`7A8BLXq{AevR&C&aKM3~l=0{)1>Qa3>Oi&g?gA
zuCzv!)|Cir%#2lEpr6xTXg#52Nv}a@N!`7!gtV7I@PlxLyXFvnv5w<TA~tNI5knc5
zmKWpR+KT$ZgvOcbrMjIlC{;Ez4S+3Lh@_8Z4NEr|!VaK}QUIs*y6oJ0Lz^RXIzjtQ
zWkEQDfNDx>6$^*d=gAu(6n&nt33Hb<6?esOCs&R&1QQ7OQtJjSHry<@mOH1Ez+k1#
zXiqeRSQ{~(I7-cHPW4v$F`EffU|dcr5gbf^mNhTq-D@uH%n@wqvy6ebK8>@(ZtU-F
zSi=G)>%#JC{NC^XQM~c3?*zwcF~)!!u%|C`KRIpe<IS`#^W8IQNAN8D2Z0enMaDF1
zMd~7}+W9mlRQQ#prk(|lCm80|ECdIE6ZOqeYsQBft10BqV^?wO?sqL*L_kYFfr-sJ
zop7wr+5G8Et3XC@q%D9oLQ2MB!UEy>gyE4I9^fC1PmNv1EpSHob;%gTU2fy57wz0?
zw_{~x)q-FI<!n4g8xr9V@CLuI5rIt%2tdI>iwcWikv2R*2&cX?=yv0<yBp2voA}{(
zUWsqL{<5_@ReG<35k}IN2P*%@)_!yj&NPp}L*S$r81vL8H$x!QMo0r5&$YeQ(t<zZ
z(b|#1xn8F*hv<Q)H={Aby=pM^-9+>c#hK7r9AOOt4DjZh=0P-gW)O<0uT4)2M++x;
z?XCP+V<YhG9iCbvYV77R9<D9<C{Xx9GgBi`o0~RM{=+9*vD4AmxX@Tu+0m#%V~P2o
zQJ=FA8oWxn;L5)G!Pe$>bbCFwD*_dSUuf6?PsTmCV|%X~onGG?Vb|5S-g;>%UU{&f
zHbDzi<5y6f(ipzEW{$gzvyq87znM@wa76D#@@iCwt82km@H$YLHn#RGaOCJe1TNot
z<$l~>TlBGW&o=jCzjtVX3Y{s1yBu>orhbXg|03qtP|&W$+EUAPOWqy0OK;(8`a1%q
zg;p(=g`?3{;)qbRBpKTba*nw~5Crb#E;0+p1((x{ydRA5!JT4g`ygxu#>_p&M;;gI
zv(|o7<%BaBThJE=8b=5gp-rsWs(Hj(xwU!UvB%My+?6(`elV-{3&HJJ>7Sh(Tktin
z_FuA4Z_0BE<I9_j1#RyAqdOre0*6nE4zPx@`Vj2{Z9|{A5Psy!6z~l~Ks3x5J8b-U
z=YwbQ_{odVNVL%L)t4XmSi^b2k$wk$gC=2qWUO#^qY9}yk4L*v(_E`@ukitnXG1=l
z{SoTls1JZMecF)l7VVI2A@J}BUkH3IvrvfefZ9Rzy1G_>g?3Xe(0JB91;+?9XzF&m
zhr%=T^%;+qapiw_(DD3<M%~V~=2^k%!TPdrM_ZHDh&PR}5oE#wxQOHP8};||*gZIo
z-GSy7wH0(!ZFVwh8gtB17w3#;`T^S1V^Q~p0`+B;ZLf0}o66@%c$&s$ynEw?(nuaf
z&M6+ZLjyfZYo-0K8%8G2=b0-0H|B53Aov=4^`%EYGMcAXYM@y)RCW{k>X0A-livEb
zpT8AeZ^GaD<HwY!j-pR~2;yB9N{jwuVir%*ub%Sl!20KL+W7dL&UW`#^>$3Z^S6X|
z+3$Qi4d?2Kk$RDDRSwtH-zuO6n4>rx3n|`&p-$L&Ic2Bge4Tz)fdJJX#!lt8GQcAf
z$~;T-l8!>_oz~m*AqIeKWw*D(Wn_Zl<A{{g{-y2wIQE@yhsd&E^o9vJOu`m>=wETE
zoIK2g8hVirwscFl%gyjn%{1ZIE;om);ttE_R1}TE-`cCNqCMhDdEW?O-e`in5%Ls1
zEhdpFnl3B`K+bsv`FQEO1O!PXydjQ#iLekmJ2(+$kpZgJ)0zYzf)QG<fP*baIVM5+
znlL$5p|i6adt2MFzq9Adt&g+_Iu)^oy8N{!QuDow00QPVnm7=Q5L#T)3_*w=3mDR6
z(ZVE3Et1gI#HMfW;QKUu`YRd1fOaeG`Y|I|))(7x@0FKh{lUGcw;Cc=(QZ=|jR7yj
zX#)sBEhw2hyqFP;U|djU7H<f`AWjKOdy^4id8j`-{oEe<qkfcyqeZ;SQFVo=goq*^
z1ZV^p_7%lLSjMrE92LRNT8MH8B6jC^fur<Tknm6@k5d^h4E1DwFfv#`AqZ6)Xr^W%
zO<&3K;;zM`a)yv4Z)<_+Ve+G1Oq}#FYxzNFDy(<IsvRpqDMBrb7pL?!`I8oeGLx_u
zRZ5R^83U971IsrV^Gv+#aO1uKb_@zOh5`MFT|O{o*y*H@N{__~@qhtwv(FYrg8ErC
zuJoTzX;c0PU%4}mZ%sPlp03#^GCY2GlS2A>EDzlr%->itJR!kDxu|rw62@uxJ8@H1
zq<A`h^_lpB;4XhDuhZ~@o62c7r$n>g@x#J@mX-RFCm&UL(jTXi*YI1v?!U!Ov@G^8
z5Waj@dEKV%myN^dFKrpN^0bTXyzP1WE`I!ZC}X!jZx_wq>E?fjwde4amIPa6FL7|=
z+s^s`VcJft62`XjTe<76noDuR%do-fzI<!)%ds;+6$J-B+5Q$<7Wd3I+k#)mmRm)a
z*rg=G7jPglu&3Yqldz>w&Vk^BvPJHdTlEb8?3Mnu5SDT%T|P5#madc;xBTNf!(*y^
zd1IVjsjuz0;WYfpvR81&*v}Qr=hhE<$-i>flm0kiuip4!P5W&L(mg%%HFVPNZT#x(
z7`JYOv5QhPNNGoX^I!Sfp6$d(m4;)g{B#%F366WQun_>qDr}~!d`K%D<Jrc>|J8xN
z2IHd+zl5P^pJ_hV^9#Ly{rKOQui4z<g1@bX@$_+1jYW;QM3b%s)S1UCOu{KP;D~mf
zow0UQyD=9v(Le~kMEIGp!#y}?cT9*lW(V}VnSrA*#T{O3OTK}2M{6bNIcPlu5?3R#
zPuJr7W;_PRm(l4R#f$BZnVDQj4pR#)6f-Y<jNG+pSA-J?oe;pl(5&bzXmMyfgfwVl
zz(|5=j^G4A7jzrP)=kfHbQZ^&algxjX?B?LM+3AjI4K~Kp7$e~Sac{fF~Tl(7n@nH
zv=Df4Od|6mv@^$|?QHMF-rk-CAjc|muhWfAr{m3ZFp5Yg>B*#7VLHvQvsrLKIOJXF
zHIBY(xQ@_*Fw_tNahKQJ++>`e^ke(^lj!&MqrJEoFTeIyEU!O^m3t3-%osR-e{VZB
zH=ag^J3fvE%vP%ZWD0?lE&{65%wesfy2;P@R&`-#Hap2V0t#joZZH-tyb?`K8)lyo
zMXQIo8o@<%+?0<GQ{UjBKY3TZ+DFy5fw=}vi_iycAQ+MrLmJ^8FiGt^(cdnU)+5uv
zfqQahgCoCuM~!;w>@@3?(cm~GnAvEw0N6anek2BJQy7OZU}0#pu@5F?OK@0SYMYTw
zdvhcm3{UP^f+@;SVGTk5fFVm=s2dDQgo?Sbj$<4}%d4E#+LQ{@ZxG&DGfDL#9KtaA
zBje3!Q$GQOc;w5`ab8OmW?RC1RPdhkW=%Hg0XMYZI1UZXg*=LFGUnLv&*n|_OValy
z%~LjMs!XiSTuzNg^;sV?rnZIwtV|Rd4LLObwD!6@)wG+N8=<ez=7X6DO~RN#%cWKc
z096U<$g!6Qt4<j|>f<Lzc{Cw{FCMgy(4Ztu`V8er18o8va$T8NhazPkG*3Z&CV4YF
z#LF?BLPfz<4ae9#KVyB)@szn1V~wYjji1nHf)7yLQ-&4-W129uFO2v+!to|JZogKm
zN4?RCW_!T`<VI^z@NMLsV6sy##_~aDFM8eG=xLmFb~j^xdqeR@V|8WG$J4c%^WK<?
zpbO0i#uE2efiI@zmTT4IVbDHw9fb6(d;7w7M|vZ8Kx-2L?%rNUZo;7g8e^7z!)Y^=
z-y1j8ZonVhM|}}QdQD3o)VQUukPmky5#M7}58u#QxygF8-17r&OB@6~Viio^rS7yp
z+L#FTxEGv$F*~pN3SP~2TVrg_xP34X{#8EkTE^5zH853njhzeS>vC~Sop6xKZ7nse
zqd5kbFm7Xw8-!{cngg!}M*ZIIPHb<!h@H)6(c9lsnJF^DIQ3VxUro3G!30__b8|UI
zp`|iAHxnESi+~kA5E}syRD06`ctLG~<{EewUITdpFN6-%lko!1LwH`V&qWiwI6bEE
zF>c{IV+74A1V6_oSFzt2#9nV;Hw2X&0l2$&;JYP_H`JHu+u&fdzp&@#)-m@<HXrPl
zHn-MR(PE<w@T0F-<9JqWFsb?rX5cs;!qBb=*%8+Js7cieHw6Einx7CpB8+agTB<vD
z;^i3Os8-@L=Wu!r_bM^QX&3%Dg0q1pH~3%UAN-wshHJx|BlOsCK)ax|W<00|1HT2{
z2^(R&nOOao{tiqVXv-m}m`K4hb0hu#Qu!i~pV1ry;|(|Nw>vBv;DbZiPY}k#Pf39x
z$L=AdrylTQpjmu0rqV#0D)9?K8U#3r|BU0=ysh~1wmb6~?ar~2Y?N<kJU8o_6O|sD
ze5i|Ir2gyt)YqLZa2Zwq7EYV;#%_+b<o+tr32I+>W5C85YO)h<TK)7!ZLqt!slNTf
zn{W<$op}EANqqRhNAdW>4`X+G(;N9|$F-HEc;nTV<Mme_#luGrt=)Y+J{8B8BXMwc
z9sT1Ajf+X)t_797CXj+@RZ|s^y9JL=uPwZ!edx>JC~puDtf2+ygKW&iO~jqb369ZR
zB0Mu9d@-$d@)4}6J9vh%$HsEn3c(l}q0obzq=jZ2aH46cE9oOp)ZkXX8i`Adbp+=3
zAFRbftFAFH;?3{06`R|8-2=~`2naa>vQh3wYxusdnOJT$d`B90<sm3tYR~(aO!_>1
zj`<H-@aT{YJ<Mwch5NJDqU9TN*N)*kY379A#|`eEi163Ur^WmGZK}RlnZLd!`OO+1
zb@<Q!@Q-M388tYaT7`h&im7&FtG^ZoMKOdYXnx%`H*sY%88Zwlb0>T{V=GLLJC#m9
zx}Mp={sM{O=9Gs8(c`?bqhfHT+<a9#z@4%)&3L3mIE-JnKdP%z(`1w8Hr<kd$|J-j
zD?1(234g-zs=}sM)?dGb#SjHMbe!UOVo;PMEbOSD%0po?J?cscD6hgpcoNhhF|d;{
zPFRk-^=WGcOn4F6spL2DoOTr)3-!+Wr%Qn~za^lPv5%_YZ^Gh2IWjE7D294wD*AJD
zVr4ouC6U`=lOP_sV6x_c9>~;-arp^xj%WfLofTm%B9eAs2PMdhg#*V>fM}UqXE^pk
zZzf}qE_F*JcNhm+EOUN!mj$SZ4;EchBFKE!w}`nUfFRzv7wj}fH4&Xc%!<h19NY`V
zg&20cLI@#vW1+&11}X#+wKfxF6QKwEAuJ(6Vf=B%Kd@p^aLI%re<s~y5l8Gsfmky<
zej`Q|@g$}G1$^%$JE6(p-fNm9-+n6=HDPjm9*b9wRx$CbvH?>jHS8?NAjIf*5Gd4-
za^@m|9XBaBf|#b!QMLCmfaIl#hJ_6T1aW~Mf;EVcs$a2-1tAjgXup~WY!)#o5Yz+<
zqe1b#2$UQB$HcoH(oR8~%Byl#T6VyeNg99hV|NTNr=BdLD2wlfU?HTmStPIsq<mS&
zT%Z#c>ZR&)UVy1iS)bh92!V{i>C$Nl9`d8_5j7X`NnA16vvY|&NylSFFkt5leaE|_
zwZLI#&+$?2j-t;%Q2E{j*#+iYAc#P=@DC6wJB~K2jyj|)eBTPYXf5zaj!XI~Er@$|
z4&2o4f*;{M4k?&@6iU`xJk1B)N%hkon3|Mw+~R-R(F{MO!yCU0EmO(_t|Dy$q^6NQ
zWhcc9M?9C=IESaaaMCLmMd_3iC?(U|jC0<Z7lDm$W#>23PWQw^b}AMEVE9m}9gaZM
zla~UNxDMld86Uh8)@jd|;bZsA!@)5QmU!PrAmA>xj4IE{T`#9lzB6@xF|rY*Buogx
zI=`xpdhG4+#p1LwCCXrD+JLw?mC*cxyRhNwIZ*GU6@(X9VCWlH%0wGyOiae}H|5Ar
zXP0xSocMOg{J@SOxaZ{b$+$8On-MasezLF1;yj86b>|QKhZkW>8qQPxgm53^PkzqH
z0r4lE+&>wSP|l<H;idO5Tm875>|Z<y#6QmZ!ep-5XMijr`Av7iX56F^<l+E#6Dj_-
zx#Y5E8(@aM3ZHdMw`@b)69-UECkea=R8p|jrlhSO-qgJ+V-+al<)^2d;*eoTuS$`3
zzKk#Q&Yh16QTiUwwP0Vh3;DTUR?&&XhhG2wPriPQug^7vsDSY)#zp@eLpyw)`J(v0
zj`^azzcIhONqqYCn4J5*ydiv!txv?`Q_s*(9mDHE(f6aHFdjwEz$BlZ^bTe-6h!|{
ziOvG|Lz5qJl-t>P_!uD3b1*xglh}QKDH@D^e0a)5<A*{RW1_$7PMcl$m&%g6U3v${
z(di!gqI(}(Bs>8l5XK;H%655qW@Z#jD0crtf1+gpqnMq;&}ay)C=bkb=yjNa^YeKu
z8%Lu-8!Bm+3)&Y!4U8BJbR=402n!4r!vigTgf!z`W8{WAv>{T%0Xh!`9?WzY>(Kh-
zPgL&MB5ec#9Dm1{L+i|11j+$iSet<%mq)On(Zl>Tt#1Sw2q;K@Oz?wA$5E0aBK|g?
zKaQ=9r^dba?mdi!)dw*#HRladbG3QXmfg;N>})=d-R&1~ba0?DPsQ4UN3nAMfnZnn
zItr~Xj_cz{Ei@<^^EJbidc(+_p5zED=xMdHVA(xD3yPiEeZ$vl7}W_n8BIR!Vad4?
zx|j`W376D93BxdHPBxdp*k)5)b>vjS7$H%0F@C0jwwV0Y-)IMpg<EVjVr_Xr^tH0O
zNUg7#nYrln261pGIKwd4Lp>0f9CZ7!zuz@&-fTBi-b=Z4qptF<Ew=^d36*gq-v8)%
zJbv;_Ei>Z#w&>p+O^Fs0$IVfR)D&O`e)jt-t%bY+Q+|wrTsx_i^k%)pI+OlNT@kDT
z4|oOiY1UX^4B(%zM9Trqw)qrJs(di7=|?axfdlbbbHc2L@ts-?9EqkmRr#_O;Ydw1
z3|R*=CZ;B*V_~5cEy1l`ubGhx6WtFCSk<lGn4=+M(A#xCW!VLPB+Jt9ItC^+@Z>l>
z1iA=Wc-S1udQMI3GJpqJ!=d$6-3vxvr+ufVxSKAuu1JsdUaq0g{6j#R0)eC<pqJ8E
z=%s#RA(-;ekKn+1j`g?Jii%Gf*8C#8V9B}*|D*|7pV5yHE+Ak{O)>f%hBjwhB6LN=
z5TSsX`Wg?c4KL46H5P=AG%hZVdvS8urSrwS@DRt{Arz%QG@A?2US83f5kUj%RPsf@
z1=CyUDPPvggwG?+LSyOFf-W@i=xg8*pGczq4*uXo5j2+&AXpfKaFaf!`Xbz<ZU`a~
z=0HCpL_*MrKnyqmPwGy;vrtg^6W{K711E%!N+w~VIs*$f2&b@kT5+yjK6pXRMlBZD
zJS#hW79kk|D&PtnuY}aSv6qHWU$7wclItPzC{gHO+W!2U<6sY?-#?K3!gWI{525=w
zW!2b5BLzl!v#xn(QouStj2SJ9uTBLo^}+epaxARev!ER9H249R>K7b?*y|5sXJ=pU
zLu=a1*ISwcX5(CKf^dYnpyFYe4<fi2=Ll}#HDE@6W6m+Y(l>Je$Hh&IT*chXxZ1mJ
z%@j1zy8ROiWR8zDkEm?Sdo|$}@DqKv-{)?l1GVW$ENHxO7gTeeO%!8pFC`*4PAZ+_
z*y$X_uJBf`&ylLkuV;b}!qd9v9?HU;h7b-g_Xo$mca44o-UBBgXhA^39gqk%&`#Xh
z=_uTe=9KIB_N(jh4}b8jxVO@d(W^796PQ0Wp9+tSPjO%6LYxWhZSVChprAeot<ZGh
zKBbM#PV96w->JWW$yS!Q=dAAW*6E+d&c5KUxCeqieGvg)Tkt{vjIiF?c1jZg#{H#+
zH66i!^qpPd>ut?79Iv^y(u{lZo1UGrplh#t6um?63c_FEVZj^z8~Dn6A>$795xGrR
z=%lfsbSZ!0Jo+u|OMh3%1wim7eTO#i<_)>iN12yBPfei>DmqAGa$2xWx<vg#6)+wW
z?C_hL*7#Hj&QAt0c6}b}>np~$n)Q@FG|`QZ_Cs3_p|uJbb@ii~`oPg(5RX6nNPdU0
zwzlRn@9*we1D7+Fz(3Y5M3~vu{HJnxQ!5%_YNrGBy;HSuU-;xuaT=PdmNZu9XUBcK
zChI>mtDu#RMUQQ6?Zm3!`h#!37OnYNwVUQz#<1!%7_d=6bcbkw!=r(P?`UDosZV*s
zfXa<F8#E6Fya`>$rVMD7YP<taa5=cEqc%M`Jh2wnq~`0J(Fu){^Eem&;m)?Dy5<0j
zKy<&psnHm8I|{pJ+;FUZzq{LuMyuuHQ{jn08!?tI)Sv^Uht?nS9C&JFc~R{!p}wwk
zRenwkc>ZEHo^9+ZEaO@Q5+pF$OFrMwKQW(9+vE8~seD?zzu(4m^&R8+{3|p*>hPcc
z;XgCa%q_})>6k-ioQ$o*!oZI3YuL>80gK?LSH&o^v!1MoZ25t+O|_7yT!uEiu#=Ms
zoJ`7w=uU>@N^p`<C?Ah^*WhoLTZ_YDbzb~C9bA~AG3lPhIr*<@oKpK+qqMV1#4iyk
zOv$H8HDhK7;bVTy{4>Lez$BXy{Dy^?fak^AkrZC3=-cU5Nn@(`PAkjoe?3`dTk_lR
zC?&&7#)J=+k~XEXLj}9aUx*(kr&wT;{vw(jOR>qs&icddOw*6FlR2um{NwG=DMZ4~
z?w?DQ?~;=<G^vbeaXmga6E`Aw5H28upd1KK*ja&q#>Wz9LSr|DcZ6s%=2!s;ZKV6h
z1tE4caiGOTX(u2KwJ>CNr57ZcfGH5dv7W;1Z*NDx)78Yr1SjG`^<a_4E<p%T2=r@K
z>V^#y9<|bRf|LK4P=)CDrmlK?EP#lKY5@na%h}65i}q8I=`U|Q7Sj5uh9eGN|L)uI
z@|&;4j0h!;hw*VR@?)nf3wQT1!5P9KnUg9%RnZSSY&jl=z6eu*$&Ljcs^+vG#PkJ(
zykG=z0D)#AQenNL1;{8Mf*XLsM0%`=A6P*Ya6UB)TQA}S7vN}SgbIOB5;w|$uoI$>
ziPc(l5bWv;5G*VZA;h6#=!3blO-aa&MFbNsb#h$AAt+_V&BT(!t3Q$iuCiO`rnJ~W
zN_`bsgkSEQgYX1Oq*W2k*IEeC+f&Po^Psu#nneJ4i&z%SO(+2y5oIuDxYQQnh6N-*
z@(wWdQ`Eva`=29mz=<$$eD|0BSa>3Us7<Rc`=~=LsDLv4QV;Nkk<6b*o2x?!7^|*Y
z)(?xmGZgi4b?wgr2fUHrC7u(?bTc)DlOLnlxS`-VthND6hZjusL#*;YG)@~!32Vcv
z1i}fgbSrt4U{zX)tE*DUb|*fTl0*7uMbgOXY)gn?6n$k3%x!$^M5ik_{GIw?7*D`~
zcfsBzE-CA+@D?1Y4<-SIm)=gGgfR?o({l$NS<XAA_!~YYZWV@HD`Oug>40{rvUDhY
zRCb5`*l+pKN514v<%_A399984fr*^9Dtht9FUyQ^9|M;9PT29IAbKZMOHI;R(jW=y
zY9E<&&9=aj0qMIr-`aN=heeU`1m)0QddoHptLS{o$Dg-98Lq^|#T!@jS-kkwQzgRB
zokS~RCw>S^K6a{N`GeF+n?8i!Z2*35+sh$crs7iS1X(1sqwl-*z~4>|gJYFW@%or?
zy+{|h86!HCG?|F2R;A%@+x%5b=6&meO)i902F}BFxw->n2&aPucU!5~b^QB3`QJ$H
zYcbylzP>TPtocpRb6>{focHBT&fywIdVee~Ry}!t6j~H8c~TQZl#e>fSg=dj3|G-h
z-~t#P;21XcK+{5BLFaNbQGIUKbThmCd!lL1FRm>BL-571IULIdb5kUfXuI<`lLbZ_
z+EE;72%RRfNM+4UPCh14w4fQ9mlrN43_XMbFnx}QCpsUY1E)N|Aco;WU7@9F^D}1n
zLuW$H5uQ7)pvS2X+6xHAnAg~l##+G}q4Z7qY&e1u3%$oTS~q-CZ-i_dErvE1+GuFL
zak>Qf3Pvo#3pQT$d)?>{`lbu(&4w8_9ErzzWlrz_(lCD!CQ&BtdqVgFy-OVg@aT26
z<9Kiodpj@Uy?1^hIGx4!e)v1_o$vpy(wg-yZkW~3x9pJSctV5^A~NI5Oz<mf1feBl
z*b7d;)LlV!0iN@ME5f`nm5pN+*|9CiM{RaG+Kq-8f!y!HJuW+<S%Ecl474~q+}SzK
zMl#xoH3iHuggr2*>2s71#KV0lv=4L3nAR1vFYshdM*c9zQ>d6!6#W3Y8RjJY%#35L
zV;-!vW3g3pe?aSmqu}<tJu|ks>qcO8dA#Nt9LB!diQVKJ)ycdGlXp!1FwADPCglF1
z$4@tWe+@GZ$A!U6M-z)9rcO@i_bRj2H^B3Vb&wgllS*^mYXIt-8WPo-B9u0ob9KYY
zno-J^_45R4Ox4{Q12F1XFKJD~I*Iz{IwJe%;=-by1q)*?wVuEaCL{feGWcE^#b>?D
zy4NS3sEn*N@{Sy>>*6A{b<__$tE+66*P{*(=Hh6qxH>m89ZfHUY-l8r_vG}Pg@5D?
z6aKJ2@a9sEP~*O-yju?;aiuMXK={mSK;V*n0){D@2RUwR+Ayea5<;syQ6rh9v^~dk
z@viWdna8p!iebyq_NGUbCc;+WYb_`F8_&57SbyT4_>VkUYoaAY2l2h$>Jtd35QdFS
zOuNpU+5vMOfsTcl%9p#U(A*jyxzM<{@DXkI*VkkH-b(h*bB-3%TG6mY1BbQ0#+X9G
zP!{OSzks2Oz>2;I^PaUI`~Ywu0&)bF;8ciX)_&9l20V4AZU|xM7p##vuCVGe^x;Yq
z3t=>P5vDwKN3eom9lXcME5MjO4h^1t8r-jO#}T0jkyLNez<`U;j$-2j%y*6#YpU+n
z?f_PT3j#ClGDGl89b;tNIH}t^@ERAvfI#PQFjv!9xDw72jzX&mMn83AJr0gQ@L-J(
zMW7E|UY=U8J~cTK(@JBpsqr*D5+{RR90->ql&#I9X{3IqzX^mKH)GM&nCE!AQ{gF2
zkb#%eT38aUn2!r!sp?HkPOJSH0|;?BVMc3t!<)^_XiEZHV4?X-@VmMYOoflH)%RX~
z`K5T}wb!lbhQI+~On-3d4FcW_g0NKcDS~>mguq{^iFXmLxd{s!)E(Wn93jfsV!lN%
z0sNUKX@9im5Z<AYs@AuF6CN1h!Clj8Gu->ib9Zk)_V@Qy)?<$ifl|0|+?yKa=bIi=
z)V1I53HM#ay_HtH`SSgEX?-CcENkAGnTXv^&)U}W&4rkoZ%6L{As*vKupQ^FMbQMp
z;oG}iYiAGW<AT%N%#`m;WPBk292{wWRJnIMXu5GM?3u<HeE`kJB?}GF;^D5bNx5_9
z*Xr_;g~>-p{doWV4`M@e4Vq*xKU|B|wPlT$>-gx|rux=Rw3==8!I?Nwe?zkl{LNa3
zd4zFD4|o05e&A@|<t{^i(ZA4!@U!{=c#{4H{z90<ah4z%1aqJl#s_}H*6gN^1OG@K
ziUX~}n>Izea8|GxQJ<X{RXFvTrKLrUCj?*PqNCuUjmCj!0pANKcnFfQ(3<!0e;vWS
z(@`4>NKN7TJStJ`tWc`1aVhi{_nrZ3$|=kjI|pYLKw=gISD}R1)I2>9jz+_(F*g~n
zsf{^)u%o`YwbxhpzHrld+*@0W@4fYUEUDc%+LiHxP<X$4s5y*bJf?O$^!$l{|G|3R
z-N)Twn_GJpw1QV#^@cSI(W>O=PL6J5{%0c*aE6{c=pCs)X#P9DvS77A`)RJ95k9GD
zJS@~EH7+LO?066xPoJvK?7LkU!wZXxQE${tm!$Sx!)>@Pc>Uz3ztCLNi?`o?D_+r<
zgdRcocPSjy?VrVmk6&m^9tkelCoteXkRZRGXH+$tZv<bzIpcZwj-ideO5>vrUwh@1
zA2ImIoXMPt6a#|5R82|QybIZdWH^*J@%&a~*&*676ut7JJK=+W`eNrd{i|4nG7w66
zAlmu!KXD0U7duq8633oHdzE(n>yNi9mvPcP%aI`qqnBUuw>?wI@|QSD=<Z+s+rxQa
zO>^f%FGIj#GM@hoWPq+5WKCxE$CDsjmWvOiLU6wF(_skUba~rBj}{)-aVMCyJ@CR{
z@wfblr%%GV*8a8w0c~fofIolfpW*rUH-F__eA17X>S*Xy@$$z3@bh=pDE-Ty&)gQ0
z1<{UblJYSIYOYEMLoDL)4N(F?>_ssIf+iaTA1qFpTrY%JSxm5yU_qRBa*5#9q5)!-
zNfE+3&xJkr!gTZ5v)J3(ic5u8s!oeH1Tk>V2~Y*16iV9*KD{q5Ij2|1ScHXVQI!#b
zR8yfQX%R>TVN&L5N9vg-7=$*zssR?11h`;s;qxg*xPIyWTHJr-B`vy}nsCq}%6g>0
zM{S{oU;zbeAv8I<kIBlZQfIm6u01Y(74a{^0xdY&4R_9}CO!)u@1QbVlDKA(NE@+G
zVc`RFfH(c0uoEf^u<?c3BA|#vTI~GfgI$wmfG9s2fkh(=B($a&UzCZwK%OSxMSy3q
z)t^-E+~JmOkbN7^5fhg58IMJUV*zX+%8wW6RfIo;95ANWvS{<!=PEl%Q7<nzReqO=
zextt2I~gd7DLV|C;UZ4SvY;j>+JOE^8uT;Z#V$mMZH|$^Zown{R{b5oAuuD2nQ@LI
zfGNX{Q_2r<#yiKa{}rCHWB-_;@N!gWSqmPz6DDyNA=1&zbrLG>n9PIkd{>6=9Osq4
zBc`2qv12l<OHu50P5)s%GA!?5x~`{PRaX7zA*`cSepUGL#_CX2eF@~ei~Y{SFNS`4
z*Vn)0!x*$m*H6OQf)!_w<#!kdQC|8lVAM~+gZ9M`7QYOgcH0d9VLfH{UwN09`XS}Y
z?$Q-|x;ZUZsJLUZqb>F~JZ!%WQ~6b49qdk+Tf9ygu~Wx<<Hnyq{1lksRp4is87R1=
zUljsV<&z(JvQdf)uqpMHgWJ}2+<B`c%t@r-Z#{%y9J)EGaOqjS4ZC4ldXxq|g>M{&
zf*YF*pMmqqS6YaVaTSSI-1GJEKLR^H>@t}yZj}q~>W@!;bB?U~2gd%@w;q?C3KaKJ
zV!rf?opRlpyK&3U$BmqDH3%H8f`>o+ch}^X3!t<OiP9#wAxqb^TZT-#VJ-X9jpu*s
zjyn8YQ&sFU%olini2%Q%5%|BVr2l>!4}h=8&|+WIYIpss@ESiV7IX*<(>!)6=@;xi
zHgIw(YPFeYHS5vl-rUB#nf_=2;pd&j!i5Mr_PV{;-Rs5HUN;U7Pb_TU9#iN*n5Qt5
z2gj$e)9J^~!EyACui`+@$)#}21V?6#`)H!U5zNR7?>vU-2LqV0ikvdt1%nobDcU65
z1%Y4)p#hhwbDTJgc4&R*Ni<%d<1sL}>Wz9V_^y+h>NBs06CNfUhfQa!3!pK%BbYiv
zR<W52<}maeOl6LL^1WSpIQ|L&)bvblyrP|tRF<9X?bzAbR@q9q1cTIAs#!~hW3<o?
zgT7;JTUTC-%S&eBAPj-lqt0eVUz|J7g~bIulpW#2SezcQHu*5#{mGxl`#<@M*naj=
z++3c;>dI<7eDzx~HQzQIITDIv#JjzY?_CjD852`8YG2XmV``VN$v``<qxM5k$B~W*
z=~$m|^cPx19N)+><KSD$%h7QM3eQp8loZB4<%L1a=EADYn0MIi4x<jbRZfBfS|d~5
zm}YukA$=Nz;O=#PnP6ydXyQwUqaVP~g(glruwi{*#{LYa8;r+=`U`gljj62k*YP4m
zNd~+cb0oFN5V+E&?9AsVKLmL59C@aGjpi1#3YuBm?ZW-Sl$~~iaSH9uQSTgCN83Zj
zSMwKTrS9JFr+CobXewAsNckclq)t33^g$3i>U9&ex;`8RCOc~aA1el}uQ2o<1Q*tb
zqWtPt9M8u(mhC492J&v@T*o3D;CMZ>EpcNF=40DbKgRIA)nyB4wzr?h){CcV%YN+d
z?L>EfSAbXB7hFKzz#m+Jz_!(BcwE3Nh4~LWNE;?H0#WkBZg{YXPxWD)WPz{ZHX1GA
z5AKztZc~D{U?Vs}sQ?qNp#@aN8-0N?P#+#>Ew8_2_u8w1qt|j;%WLg?ae)x$R4`OO
z5e`T}5i|(vxlG&N>SkV(seVuqtlRbPcFXk{>sm?$GaUbLe>ghko-)>&TCaj%6rRnR
zo$j6)vIhus)URQtUS4ags4eNs^h0nU8i-TVa~exyaU^_)2Fu2at=QPuh-XhficP&Y
zUu?$4^NrZr-i*%fj>a0}vadeNajP6tsqx7`mZ1-`5t6kq?Z*0#Iw8myYFerA2%s2y
z-XJRYv2IMQO2%5**eU=}cZ7}P<F-?>Xmm`^plPbHp)`RN>u=+B<;&egj317jL+Hib
zPV|pSrHz)a`-j?*zD!?*@qVlQie?;IikxPGaFO+|Jk&t+5t!M`H}q%PA8i-<FZpr&
z9d$*>F;{OzZNB9@#Fmy8jAPk|z(#;)&!5IePaek$!FF$VFSzTBI&j?NVyhnYIRtlX
zoX{Mhw3#E&5aIYbwD&kJQ(0@2v-W3$F9OI@Gx3qnoDBBku(#*Cm8kp5>RK!=ti<Kb
zXdLu8j#BfK@GAHYEk1;D-sFtnT=Q5@xOs5cSK1doqHTS3(ZXx<W5Ls-lfEym)9q@F
ze;TvG|I14)g<BrtD+Eq?FC)jn&bZH_@yvW=K9|}T;Tkxbhs_k=eeejdu;5@;uvWh0
ziAD_klymht@}xZx$fHHIwZmqNauSWk?wDYMP?9#ZRvy3-thl>tUh^3+p=|au4`)Pp
zcUp7VyxW}3<cxL3Jn<0Tw3-~HI~^Bt<2XeNlH{jisQvo!LT!pB6nr8&{qD|Ayz{f4
z#z3${Q_LDxY8&6pC+dQJ)lz$+nT~)6&9+G)GPLgInd4NC$qCg5{)+KJwl{(-1cB4?
zL$J$H&9hVEp?}SNSDI@y)=v*((AkdN&8M*=T*|Zg^l|jNyK#0XI%RJ&`a3WD9##Kg
zRHBXA?Q~V{J<p}gxd<gmv#$DcY@|10fT-k(0GFdxIa-=QIWwnruFZ>nx{RLatr(p!
zywE&lo?w3O9bZ}lude>EywFg3^cCh*m0RNx-sHoF4?KQP&aZ;Qh0d>qYc-x`=ITN&
zQ_(*ZZO}W4{r+iecMjv}i*9_h)rrS1@_Di)I%My_8=AO7?cn5E^xb)Mjz{A7a$M<-
z#KF;pweb!_+nryDuF-QaIE|jh#Nny><;{4UUX4fh=rZ~$vr!wk3}mK<HczekNzWB%
zku_T_3%NMv8M=-mf0+~5G{BvF2!Qvtchw*E<KewkjpbTAf3_L#zW<@dlITz6j}T9w
zC@|*XSGea(7)KeJD*kOcx9`s~UoM}&z2=j3uAW;1Tz{G8uc!6Z!ib{|$6kN+)gM*a
zVX_=5P|4fwID(8z!T4k1w440!F@p=lQNj}?t%R%G%0nRAGd|d+@?%27$=^caerIq!
z8J;Iid6)2+PI2OmJAZa7ZzE^?6Un8}kMB%3?f7BQzbm94fAWW0hNa$BogI8=&oueT
zWHLO)-&mb)TAVMRHrRlJe`SYwN%u?#t3B=J9L*)fx#T0K40{`H1VNa-A&5#4y@g;L
zB*PjseETW!Ny~Oea9J{ZN!M;wyWGaBo(x^!A>;gU)lbDNBpi195YXx08X?z0=3|;^
zCZ@FrU^j!p5{ez9q>D*n8Nn9&szxk;;TRScvI=V95=Zf|s7OKpAr&!M__I4H7f>wn
zNS0lXCL9j5col&_e}quvuBmZNG7x;!hjZ2;CJ|Wq{AnR(;`^9DrH)X-b|!TZ80;2-
z$VA|V9m7Nf5e)$k@%d#;HSSD37L$-DZCaksG$I7n9^H?1O?*?D5ZT@2g6p3J0RkTg
z0EB>~!=k|gBh{0g6Q>BI5bEfUMUr>UC_al(@0igai!}t5w5um}!4x}-YJ@bouxA$}
zJ7rQZ=rey+HsTWYQm{k#W!NZhmk+`n0g+$;K?j3_osq7VCP*(vMBMqPL)C%ZBlH7y
z7m_EFs{4f!tJ;G)vp{A6N<Z>Wj_l-aANon&5s>+(Ad1D8f~cH!(>Lk4UV4*`g;h$I
zM!D8Pkg`!GFDeCd;Ki;tFIM%qUmAw$uYxm+PT<ADo_^+qu)>iyMC6zVJIZX$JN?n=
zIKGC*Fm#Ds4L#XcQutjxxa$vG##7vk_mqdfiR}TZe@Dn~{bqWVdj$(T?3UlW`QbhE
zqm#<w@a0VaO6WPnut#W_ad`+?q7^5(70F2Ei(|TDxhvFghE5L$>Xi%YDgZC1UgcTr
zRLc4IaRoEHvrYGWr(ez)Rk#X%d}i4E#;*i%nBor`Np|B#y)#V0GsEPQ=@vNY4?H<c
z`woDrc>MSedwv#s<yPLMajIW?6`TJQ$bSk~qUBHi>ruH%{#71BA%+8_ql1*N*&aNV
z_icC!XR$kcy5o`cDt$pegve00eRB9t=b`a8zG)en%nRq+v>ZlH_DAf_ET3ZB4R9EG
zXPeu##LriLXa69q!a0TX=O@FFNc!C7mvId<y{hU^UQBrn-xx-c=h5n0fwMf+OaFRG
zbL&6<_^)7#Jv%ujVU8Q^0pt@VS)Qt%39}62x)Oz5D#g1BXFsg=DCO`MrN)-O{Ac*e
zip8<YTQ(*C^!e8(`1;5H?B5LO<j;Gu(4YQf{XhG0^*_%j$mf~T#J|$>jrn5pWofkE
zg!%luKG|Y-!=^p8vr9NRF*Zq(w}_4ypXUCO*_f`)``D+6$r;laFcT0S<Zj(r3&3Dr
zneI|N9}6WKj*I2Fzl6);VW45yp%0U#Hp@{$qM1eiAPDnuPnV)0V;q)|@}IOe1qgs$
z#wVmBS{r6Q8eYDL9&Lx?WA7}6N!UMt$uh9E4HwR{TOJreOF_eX0~U0@%F0paFikne
z&BwW^4r8M*bAid2=*kPFIndKLZ3=zQDGt;hp~CZL8)lYr7Z<cx#gvC0yimPU*p?gU
z@(!=v*xA{&b_TFO3x@}WF?XCGXlpODg8Nl`G@{DcYysy+v|!$-tk8W3KQ2Z0k}iTd
zv?>;t+YU26av7UXKZqay`JcqQKmCi?dj2#HyL%45ytW>%z4hH#d-#ghO_LTLB8WL;
zXZGQt%U&_BeWHV4$1!E>H22X&YCbc=<qz#QJvV1T9LF0%gF^qSfNCQcwp|O-`mB$X
z4s#WnBL@ikM0Y1Mbt1T2zdG%K-a-2y_+b9CkjHK?z4Hh;Pikdesf>?dQrZXsQ+Sbd
zG)Go32f^fCSy@nCXgH+cf)hMAiu_D)Aq=!R^_bE6j{9>t;;^SSqfjv3n{~mKGOFC%
zpL23@<Ri4`i)cQ<plzuQQ)^%(ID(1(WIly*g(gDC2y+{u8FW0uf&L8>8-Ww$fI-d7
z3@lhnd0nW=qBX&hcfh3@^OK6h*rZ=`pOOVnZS^0$5oWOtK==UjdZ@KSzKWac1M(6?
z1P`yFF3#d`(2d=#r_tGc5qsOuqkq`*SO@0szWej|#XCRs9aV?@ZVV1OaeUY}<Ny5p
zM8%{}kD0+u`e+W(2UxQq1Q@;MWDB*4U_qO*fspW7zZBe|9e^-{G0HqoI=~d641fe~
z-ejrAnl)ZSDL!LXPmV!7deaBo4^%%3;TSWjBlU#w33J(N&9a^&9`Jz2fT7(emrv^8
zEqGRIAZRB-nMa@>n#t?-n_LSr7W}O|xC7|?^hAB-#F~BJs=fVA?CkF9-HrXOaKmoT
zT5t%v&dx72=8lV2TtB*<UUa*RrvvjGd?aBJu5qe~`;^y{SqFD4^&{VbC%F0O!8z~;
zN0K5aNvhAAbEzeRi7}u$T62oEuRE>!n8u@z15{gCyNk7@!m)0h^xQC`wn~5U25(et
zQjHh1)WD(KQ^lQCCl(MQOh3>(LH`EVUdBMUhA`YqM%lO*j^kY60kF<R_`)4$snw;m
zxYGB=Q2LM7d1wLU9ZHiHRLp67S{qE`oMTLBGaqfIw5Mi;R~MIJb$Kyn6!zr!FrGhu
z9`C>Zeth)!L-l_)Gam~V0Plh7koS&Z5a5oQk8~9rk49M#awBM)Q{QK9L&(r-wX8|g
zs5hhDlzn+sc=}$9PtC{R_`;h&wstyh|6`3&j?+73z5rjU32E1k+I(LRn?h(-?ma>n
zzqr`8z=$J)(Uh5;5nfiAe3b0TnfWc?2FmQatu#;K&OC^wa!WXIW>WY;bwmr3%^mKu
z8ZR7?iSQo*s(CN!FSNJsFjSi0&Zg$ehUR@VmUe{?X}^){OZC%^`>=((!guCZs4~<E
zAunyzsG}jwrrWSa8FwZEud#93G5Z|%fi5pLW3efmCcNF&JjgMbZG;)VGm8xuQ>Ha;
z&=S13R-WhX_h<#psC`?l1>-LID_W6Ps?WUW4YcW~H+R(0N0~d0H3y!b9LCnh^LXdy
zKaE|r|Niby?C<Rw*8ywp`(o_7|58V_nR$1YY<j;@pA()_z0L-`c=u<274HdOZasSv
z?<u`M{nI~*|NYPZJl^@)&*I7B4>TS+7AP$*Ektc*%9|^uH16O@A@l`z(-+WeL}PZ6
zai+PcCfHl6dm;tY2&@^e2;_|a=;dl-=1Rsh0($sbdAe1?>stN)?20?Pt~8#we|6Sl
zog+lyo%;S$;RDjI&(EuT+y{r&^{CpJxfMDf?|wVtNMFHruR9Q(cNRy24Z_de{evM+
zW6W>w?#K4tK^z>O$Kd2r@Vin!gWoB9u1wF)uj4@U9fDx`5@X#*+o~+!|M^;7{U&K4
zj>T-Ke?Y%6?#*XY1)yg*5|TbXCz_5;IT+vTDmn!LDgDsAJ=sr0BXT6_z4hgonQO#{
z&vxPiJr)S6JRH4ha>J=$K=wSCTi4Gwg!{CoO6TvQ`Lz00&uzVP-SNvjA2*)1^fO$)
z{Dw^mzFv9#N9y;=o{S*oJY{-;WI}*9q49$_5GrtaK$;993ds+Cc-Vn|63sY7@ZZqC
z*lp#f<SQ#~3dE0Jlquc(GyE;lWE590AC}b~`H(7a6P~t`3q_D0mMR@W%90oHmO7<}
znZ>0vj)yznq(t4EPI33{d4J!+Fa2)g@*r%ZpJe&&#3!6SGp=mK&E+GY3zF`RsJEG|
z`19#x6_PR!<MQS=!%<-Vsx(MD^TUaE^^|X?UBdgj_z~M><c%AaOy7ApZ1GoEtXQ&K
zDkYA0un(_HC%>;1o}&%NS#XHBwjfMIEFlpzk-C==q8j+S)Fgs|IA@UqG3&*Wf<U;j
zQ;Fav;+XUuT7<gNXQvDkWmON39HAU0ylu}NXCPL}@7-rG0_4vGN&dL;&K+Z!UlLTg
z5JCXwU1x;RqyXW_+4~S9l&MtvQw?>+-{mvRsCW9^HJSRSdZkhPDNU|ud#&DE_hOVu
zkoKp{2o$wuqWSI+_@pK{swxmVW&oHNhd|7-QE!d|AgmxG*vpkfWMCFpm?T5JEr=8$
zm^(Sy`H28cizNtI!55;Jwj-X8x?o2i095(N(`{+E0e8VoekS15_ej&b6jeU@rjG?I
zWgsr~r#&md8#N;?1TKU&4P`=5ZZ5CtswgfGHAU!|LNKKxJH!UYM=YvbDvtfg(+gwy
zd-7H~JSJ!iYkfN};7vc`I2aadz!w2+63ZN;i64t+G`YxgR1M?>o7^djcl-(Nz}|$q
z!nrTI6xliqK7T2W-KicGQ~vx_eLNq=JKn@7wJG02yZ(mwjsB?g(l6cY?lF!%)8l9O
zHw@#5{Fj{g$j2{#mpA>X_@!H^lfR2SJ^a44<3X=19|}(|<-?!Y`Lut!<vYJCdsXJD
zKH0tbA>|Sd4~ON`0n*PFY*jcrq_g64ub0!df2Nq>@^J+FXSBO`*Y3RROI!NGcGs<(
z88-}zr{mn26V*jZ6#F@dtt!`@KmHAChcDmGL;snWADoL-c7S`ELAt9A9F1>1r5{#)
ze*3$sC)p&N(~UQddi3cWid*%zz^95_m6<3ulqx^y8EGXhVKQEE^62qLv1h)-(im2`
zogaU4c0KH0zR6b>#<|ljfBEkK4wr!w));>d|M<os{V*=OzKd@@<$E|T@4zX`lr_G!
z=O6i2ea0{QD@1<ip$@8VTJ4|0UJ~P5Z<sN*^HI{4wSwm`?c(lW`pNM2uwVZ2uAEC9
zh7K8rbTLn#Kl@uH__FylkLvj%<IeUu<_psJGFg0MeqHlLjr8>xumwh=!SCDFzkMgJ
zM0hc3#=+^um6<!-g>W-A9#_ypiq4UO{r+KaNq*96&@<3W`WJlxX~Vq|Tf5!Z+})48
z-l3TpXcR!VakogLK5uQ9x!K8>n&4<EcEyXfR3evGBZUdVCF*$(4U9KVJ1`BXGMF}o
zwiAuWan>itM;1b{TOFECM3d>B8RZ3o2w_08+4M74OQ97z%_m?|&CXd^1QW5<ZV3P0
z#9(j`gZ_~jFr9q_X#+E7NRy33U>TUWXkcu<*s?$iVH^xA%Il-!fSu|9U5Ws%*X#Ni
zDPRX(%Vs#3IxvlARK`}TX^ky*C!<w_kmC|2q1GrI<J4--o8b$+2C$*Ss!d)#CUSmO
zX^zCvpc~IWdOv>r|NqnY;N72UeKQfSz5aSUdi1i=UX1DacFZ&u&GhBiCJPf)-V?zN
z+8ag@j4n2odBd6L%j2Vg*Df4s2n@NqiL$2va9ZVHopf&cmGx15PUW{21V?*an<)k(
z4@Mj?M+m{(LwMF;t1C-pydmiGMlJO>!$s`}qm^X^bTtAt@+2PgFLW`}LbU<TG(Typ
zJoIenYI;0<4~@IEmBqNfzHC^gP?4SSBaz1`B8(MnuZZ+T%78XuU;O|&8?6B5xq;e}
zHr&|UQk!-}d#C0bnkO9NiBQ1WS*izGY-k0}%;Yg@(DaNkuk=bkhT=kCKz-3{vCs@U
zKPOO?ztH*W`#!o~;Wg0ghqfg3HjO_918_|6%iZp+R@>TB9F;{_3my&*)CW7!<@n4#
z+Ce8OlR&`uR>BBT(1Jp1=*7m9*xT9Ecsh&r!cyFO@W|yv@B(}pGaOrnzzdkuHxa<r
z)R*UGr=zX>DGN9P;WJ|w!555Sa0cr;3r^H0sH2bZQ#dpQuN4*=g1(-cN2xRYMTJ#4
zz2Q$`Xm8^m+MNE7&@sF&Su-*|N*@9K*+;Z8W!<ef*eiF(B_qNT!4~~1Px_u=tS4by
z?1CE`3Rx>TuG7fPnN?Yb!Znu>0fOL7|5e`fFYYQtNYDL*yTUa)!kPQs1L2Gl9|hL!
z_8f-($y$(pL%s;F5aw9$Q`uD(n8XN-j320n2_Pz>^}NRF)J$I0z%h&om|Ux(v4y~U
zdYU8jG;Y+l5U3e1sI1^jZ+;+*a3+FP?mI`zWOf#OG37C79*4rG)&+e*V_4}SJV$Vc
z5a-y(oN5iO%IqQ3>Gaeuse31O_qu`w_c}3Or1lT6q`b@@2y<A*RvI_78#FV=Yc90w
zv9`JpZS_s=(i#!GXVovQk%~}-`AO|yEv7T!X61nZx*`9E_gCZo{q<<K+tzYHIKzER
z;MK{A(dZp?tQGd`*|XSq{@j9#_k>@cKlvzno&D%^_B2=QM!&ZoC&HHqG9Es-7t71b
z7TC|u)uY~8h}n8eB=A)9`ll9pA@JfjLfQd5g@A*$23Hyvs6D|EXy<U>9x&kEDVk1T
z5In{$2vcz_WM62t)Q)rBL}EUJ47?8R4&xrfeYF7(nouY@wdyr*xL`v!_Z0$1HU<ON
z9B15_&OC6T^puv$$g#uBJ%j^392+<_TQd&g*jfZ<+($?L^uyH199Nmj(VW7!8pY;S
zc*8N{MfE`ojCl__f_VbEU|zU{Hbes~af18=8?TepM`=HVY0xDI1`)U%$&NsWjWZm<
ziEwswjQ+`Hg!#C?w(5RK8rHnMI9D6Z7+;U5UfipBc-Xg)?CFya<DH-WRlN7k&qNcb
zpQ7<9_;L3b_>{VE%pclb+~xP|$>VtX;k&W*{IUDUk<y_b9d!4#2H9}GZ!avw{Ra=?
z%{SkOZ++{{SW+M1n9-J=#l<C!#o2)C*k~_Cog@7`-qc1aI|68g*uVgsXQ95@1VJI~
z0sMeBywD4^7lJu9(1S}6XwxN5)PB7q&26H~m`m7XG9`GR#fJv!*d%jr&9H$tIidEr
z6rM-$%O)Y(o=qX(E$FS>ygp`_a97-kXdZBnqML3*c<aqIYM*_L?R_>9^%LiDoGJI|
zar7;8V^?nkW^4*zqkx8j*UlhL@}Yk4O)KqPW>D}mxI(?n>36#Ww<+^3T3&1(0zS;Y
zXdyC(D?!s>2<hpQ9MO1u5%FSQc<NYl#|T=ho*lrR#ke$Swd<|VuWqXJ{;rwNgU@Z9
zhjZ<(Y^wM?zj;_%@b$*)ul>l34NW>`-0&0v&E>}(f0qM)ULdUEuLm|zzU2lJ#A)0<
zcCXT|{_rb*AbaL=h*}KEI(`3<AFN9J^euNYi3qPq)h>I!OCkLthU3YDv{OrnpwzXL
zNp6H4Rhj+f$N!mI){%HsSiL<t>z%F{jz8j;a2X}T<L1CFK7N!7YxPe5%B@)Jo>sk;
zxZfmWN7Y9^8MpXhCkrQ`Z`UQ$a~#Uepf6!@OXQPfG^4)qb543YUy_l%#D(_5&6;xj
zF{u&H{$yj<j7Nc}cX^WsS>kEtnOezd6Bal68r}-a@vCTjjUfGJVWqISD54C702RiE
zAY(#jA|O6HOxRW6yfukZPbLNoNR~yb&$`x(PdtP_Uid0KkS)7gtpTP9#ac@$9D)?;
z#UuoAjN$BYj`V{_S4m8KK~yk~EChq1P<c`HM<9VeWp-O?Qo9u4V!{I<4J82_pJsp$
za8UtOdzYF2FEWN>T8Cyt6Ip$!84q556pJg%BD|PXlORHvb#<*-qQLZ8gb@=QlQ3md
zNE827G1rQU$&uZJOjzuqrER^-jylqwlv@@SRtRjUC+UzE3uo$I2_Pp$Tn*cS$jSmg
z3l7z<5J)pITf<9vvSS9AL-dn}cjwU7DqG?Z0HmB0AGlK%ph5p(vF1r!f7FNU_<@M?
zy<jRU{TDZ)IX%4%lk(sPA;Oz_K*f@GrbmBM>gpG?8Fk{w6c#oRbx8y#!^k^Sl%99A
zx=nKt2NQ)|#T*#|QDDKW`=atuIxZ4rrY!_4$^bDlIVNI2L?N1q)ZgRHb;<GpD_Qef
zKmH?$;^r}IAq=1V=+`gYu>(WGX4`%&jMK<4PP^DWu!`MJ@k_TH&&00m_Nx9$p8Vla
z<v9$;kH6E-AODzyk!x>%eyT8jlePmAC(A60<MFQC>{`lUH_A!Jz}HTMDSv$1oi8F*
zKDL&4m`q1M!*CfEI|a8_x|KBiFF(V;euT+V`^WCUTKLD`m0!juy~NqKbK<ZKDTn+>
zDASM=Ke!l<w&kDwvAKu&ZGSzLD-QHL?Bqit%pcgP53xSkwtg2kto&q=(#=lx&vM^}
z)jJa|;fg!P!IYcB5d=RH{iNNcPI@A9O2_gd&9SUVe(R|OO!-X+>c?;TS7l9yoUaWB
z9D%97Wo29{NQO_?=cm5qhJ)!63g{?ACN}voW=fh4B462EX9p{37k|TnFgQAmkNXkd
zA@CbMC5V11*i^8>BV&}Y%OCiIKb##LyEDx1Dx6D}<xMvk%4}Et0T0|t9D}0rFCI^x
zeZ7LOV*PyMvi)iXxMl@E<M{&fC4zio{*IWBcZteYx#n1gX>fXW9eu6-d5+Jn;>gUw
zYco(_b|Hv@In8r$c&u>T{c>h{`*LI=28zG4&rvt~@zL`g-}S)`;7bt#FxA+74>Qa6
zyNF(gez{N{&<QXy*^t9-d^Rdsn*&-!v>c2bXhL?6gWq6UvE!Xx;PW-;I?;C=O(Yr>
zdLGRi?f|*};C|e@zZy#jl=M9Tqh9nNf(!1+KvReGtZg(gVOk0y7fc<PvD_o$yQox7
za2AYbj`BvM3?^r<(~0wwWDJ97&?w0bY`K15j8YFYJdO_Ym_%qiACoNuZ3i=|2J^9A
zk6IDxpq+z8h>zRS+MteRoF3?JcHsjr<5*4Js&<TAoyXBpH?}sO#P;U1xH#@x7{*at
z-~Ij%;yd5{tysPHa!exZ8ky2ufG`QB8L)-vdgC@hb8l&3(Hb|#wW<O1G&{-Bh+AA*
z6dYTE*R8e%?S<wEbOiAbXf7|zM{{n{LN|`^gOSO4XL54d!We`jKDrJ738w7$=lCrc
z+-BelcC4KU&l|Ys<}>Jc<t><HFIEEj2WBw7pz)bMfGbIxfv<i|n{bi`8FExvQ~7ax
z8H`TKO;s5Kl%IQQ)DykFBQAXpnw&C2?<0iRf$4Y52@@JGXczW-v9`()l+!7^fhIpk
z0C1ytwP-gH6w)8mSTH4B$Ls?TI^@N1hiGBIBu9v5CbHnG(y7?2vsfqaunefy7wT_<
z%ca)|xlTG&{%9}I*HRG3I!k_9KeC=h6Az|0cWd>!9p7)lT~b#n2Lf*xvmBdtb<I&(
z%~)Q&AMJ%z5hYXBYEu?5UvJC(UM#KLi-o0?SXfz&Msrd9Vm9WcRR^^ZS}-(%*S_j!
zXQ#);G%)iKH_<-sLuzyS2IFy{G16Cg271syL|Dgq34sdh-`td#b*H|jBTO0qqh!>B
zCjgR#!g>Q+GlJER7(YD3(}EA^r_m>UXh@f6jmkqm^mtX@vVhBdE2Ds~EZ`z63HVVu
za-)BG4X1LlhIAy;U({LnZyF6fG~(zJ$-K4(7eaGuadDp!ePTxN<2X(AHT5U*pieXI
z=p*!_X^oK-prdsuw=xc3_)~Y6OL?=-WTODEqaRZz){5NShZe;g+PwNEZNh=KV00rK
zrEJ`(2vNoOq<$7+3$9)_0y~u#(0F|h9#;7fYWk!X^=<kD8jy<%O>0;!EVW&<1K~OD
zaAJc2_X`e=I5JkaO|U^zwzJ>yF}2<}s_?AGDG#U3uxXZcEzJ0KyB@1ctyo)X#!_=G
zW+oK)hD{Ja7#u(0`<^%^kfZ1jZX-CRzo5MZ{#g-zyMOOqJbd^t*6%%t)pgAu>-S=P
zec3#o&i-D!|BH9y@rUon#>Ufl`s{Ih^!WXF_T+=uc>Xv({NSDV@V%eM^QVtvZFwPn
z>-%rV%3|B_As_H98x5)}G!PI0N3Yq{U#yj>@e9v`zD9op2Y?q=SC?b0Xsl8XV2^g(
z@$tFG8vG_UULY8xF24VYF`~Rx0OiFdk<?H+);MD$0&Ob*W?!3`idD^%<k4={)c3$a
zYD?NyZHgcgjc0@y^iecrxC;_(A9x5W%gdTSYSyA|YYwrNmHNXWcTx?yD$`tWFFrVl
z<9+wL1C{;E;|d`c^9{lZA3ZFB1DpgdMIeDtUL)CfmE$0{cRRi>a(k~AU5?K~>-$t=
zNcAJlUjI0@cRKNWYe(S_BoE>MP4aW$OX0Wi2{hK~7C>>=6&kBIm*=kY<b?XF`qN7f
z?+3?8j&sruTB_<x2vX5>JwG{)Y4yF4^J8_fqnH}GjP|_nxxSx#@NPW)=mT${SXo()
zS6+E3zWvr4>PKsuBUa*DZ+$zy_r34Mx4!e8c=XaM9)I*ta1_EB#&Jx{geL2l7M>lO
z(Y!dWxnKU%YGbtfMufXgMN`02Wd3GuAU-$;I!@J8T{SnIUR$e>;{*A=5Kcd0Ly>4Z
z=0woa`DNmLWe`Uv7qPL8aC)G5dn%UO4d1)g5bbf;KhT`9ANN)ltc8arqy=AU69lnn
z<FT;;;l6pDv=}&7{l`K#^-=IR{f9AZ{+#*|{rw1<O8u!vzhwT>_(0orM>t_~uN&K)
zLG0?;5<SVguQqm@>ruCM*(b)QKtAdpCmt8jQV7634n(1{>7BWrd4X|<5LN(F87HHs
zIM23~c3=6PC{4hjG>veEzn?U?e!9;$2J?CFVWvpqx%f8(N0Ra!UY_5CDT1#zeguL_
zMgx;wX5`5aM6FJ4e}8AjE&$f@c1$xuRDIlJzx_+MOq19kUH>{h`DHrx%x~U&`YEZn
zT=wz1Dof?zPs*#e!{b)ct^h_|?5~IXt9lIU=(6$Fp9`B|@{f4=uH0P)KfIh``6fLD
z<HvB#zlsT$j|%6yVlVM=C%oPkAeFHCxnn0>rtNUVcfIm$|H@BxS!FMI4eL)B`4LV}
zwyz#ibGUTRaJZLzZ6|)ZIW5<qxXGU{+?}EQ`Ob2thh6;VzG4@`R^@Q_u~6;`?G`3$
z5h$A5JfSnG637CK(-R9qAeceG5M&@nh+8wLm5yjay;+FlA{@d1Vu1+)BEp(8EW{`W
zaTLxMdusv3Vi2Oy=VU`jh|pt!4&{!pf;=mt(Xi2Eg|?WF2~^$?(JX!-K26*UF|%la
zV50mm0a(CPyLdj;2pB95)1I#{V}R@@gO5#(MOzc~vL@_Weco_Z?t%rdN`V)4fJQkl
zDIc>C{(=(%rXww~&{89PCT3s_vB~Z}m<;4v5y;fbM7iQ3=s~DRnk;~*Hw(htX{z>B
z+c3HMqFrUl;t9e}O_Fd^8Qd;P7uY&4y$xseO@v64o#RY^D~oBLEv+^nFNjl!?i}mc
z_ld72Q(h3;BhG`mS9lLXqUzU7?z9_ed2tGZL6708Jb-l`Cvf4u#=?Mw2z?AybLxQb
zigL10WRb%TMen{6jLF)<HH9NxUqG%lV1W)i8H*W4d8*uA?9u1s=3Uxy=D5z;DegUB
zp-&&oIue)R!$*!qtn%=;UnU}W<3L^Hlb`&_jc{qtFb;>i|728_9OD_zhBR%bf3c9w
zFBjVOk{9Ey{5gbQ4w?4K$3J|hOYtf{`mXH$_lNVX{)!Whw{QDrxb)XI?Qbvp;P|mH
z!)p06-FPa&2d08e`Xzp2z*XhJgM18&s&z`_%$E?w*rFfe7W=Rb^%(AU1D;jk@XGuE
zc8S3hU~fm|FQfV=+kpVtUX+_KC1i<{2{<fm=(w47=4YF~c^Q_&DrUUI`+7h?xaC*I
zx2-ZIJw3O6e59ZM?R2NTviPzy&NB?29xMKI!17da<w{wKTeg2zZU|k)cMyHw&X?qc
zojS>ndZsH;i{Bl0-14J*rvDIrerGuHa|j3E$uZ*p;&hk(;Ai+r-Ek=a^NlUdZ93_1
zHzwHPmQQ}mqUYA`IK|E3967!1&tDa!`pb`m8<vtDuM*8ahBWo&8-rVki#_9F)sMjV
zcTMp1>zcp8tiLf|yD50Ytq$JLcRmOQV2a+v;P~9f_l$^EV|Q_*-tbOi?(|^CJR8M0
zf)5QU7+f%E(2_!{>OwRH7xV9R2GKb<7CncC!CcJD&6yF$eH={~`=aekLz%W=lTOk;
z&|lEc(AXB3Sh!U)?TH_BA(JWoKB>XGcolY<urtwUa4%Nf^fR<Lnq3_E2EB+j630Jv
zI$a-8$K5Gt@>c>Dm`8)bk^H)`ySpE~9s;jZ$UEl?4uPqT1_TVCK3WS$N1~<jSZ?r5
zMf*eJo1PR+2Mt`^^FtnH!cU4O)t_ly(f=^JpjDyeIKqrOJ-DZZqXo&6V<0&?3wrXv
z!hvJu2Mkhx$nkP#r-!2H&b(W_)tHU-wdHv9=uzBz_);t`-HVCoMx0!PnP4#fxWCJc
zZ+5xs;Rq*CDGvf3j-^BMhWlk;ep`T|hn|CmoKF>)5UngXXjs-%X!5mN4Ktdi$FHMa
zn>0fj!53{fUvG-eO<ETE-f=}IlMg)*=C5f!!5hXrJ=S!)Y3OSUa}WTSwuh0XGQ#{N
zyz#vrXk{OVC_iX*`nTsdeKY4GSl~D-HAZk$7Du9S<Rn5DcG4rXYc;rgsKitM@UiV`
zXWF->F~M<9C!*20_l!2KtKKjHj|ZJtSX}Z^pr;x`2Yv2z>ZzYyTJQx^ctquZPN)4U
zI8avl1^s4bE;VlGKe*=xKZF-KqvyC~$)i=C(qY`f^vn}O5VWDCbE>|~{Xhe2A-bO;
zgezmmN7E~9Qe>?G6S~{&d$;}I@Wc$q!y|+qY9GKqGHJp0h1MOTqX=}=b~nDcMlex1
z)t6K;$30ZsGu0jK3>p*0D90;Sbc|2lkd@;e(S~Gfu}-EeFhJ?sw9nZIN0cHgKC)(M
zbq5})5Qis)QV6mVei-@`LK_T_^4L|UW*mFHR^Z2-JQYOw!`bvgWz;?}`?GDz221@r
z3BlV(89G|_6T&-n)|`y18IFS$oKzNYK!qc69hOX5?wsQOKLj86S#U?X^gD%Nt(!+y
zrZ!H6e@VmbsI?y9C?haHpvJ~ggvE@xycbRFMZUsIJQ#2a>r&R^;BRp3tnfSK27bAD
zl{<6L=0PCMk(sm+Yk2ij^-1z~d6gCqz_7qi`J-W&T7NS!qx?BqjeCfem)ddv-l~sv
z<G48Z0tX0A)JG96oBkChKp+U;1iTH7X53f#^bZyOC<Um#Ls9UUKmg$*>D51k=P!k?
zW+z5sp*1fYlt-b_KM;I1=cg<bYt^QsHiJgmtnct@C>-+%NA@vZ&<FwlqdCQ<iN#hk
zmfCI2VGZ+gHaDJWeD`8yc_Fw%i@U_wn9%F&#rD=loE{AVZAf*8=<M&PoE+1ZLLcfe
zC%mzUP@&zlFo^y^J<$}V&Wvg1NX8AD!;D*-b!$JN)dkK$Blk$-6rmv-b`jz;hr{O@
zSH9G(n&0R@;6^quaQ_)vv+$aDgD=^5i?I3n{7mWR5sIw^G>}@+Y%RomqoqDo^Lz#0
zg^kV#shF<@{nSj1>r3_3T8yFjr*@=`Ij%IXDAC+@awvQ>i2L`~Jhxw8j~ZX}G}nQ%
zfB{EQjtG~~AK=0C)fNc45P%_cHx5u4LU`y>@nJWCm&|ix95II;omis{%~axAvq-qO
z$3`9Ds7`kfn>%|7b09p2mUEa-WZv7f`lach6VVd$v9QpJwdJLF{nc0F;REius;TEC
zA{f^*slLZik#nL$Ru`M@bIgGVxn6tyjd=C7*JJ(u1I@upvAVhz4<0^>^?Ua{PjF}c
z<P_SKv=PT}PH3K*R5{OLv(u06;4IFB4xp<nga<zKpNlikmuOglW55>zh~{)Qw6HPe
zI?gmCxd&@&|4?=9#h&OHaQa0NaM3T=G{f=39gf&6M>alwx)DEr?@4TIZL6Q33OA2O
zeMWPI#uZw!X!otHwDWGXu^WwzMfWST_NT{hER<bb5FMsI4K4uxay%{H?N-Zl19(z1
zy4qimRGlb4npLzpxCeRyp%m=0<Fh>a8oH7;^rJq5fDX8DtR-$6n>(?&y=`IGh;S>k
zC*zm?%{V|n%Y!yEo3C)AuW`TWp=hG5j^^qk^_|PnxbWNn4hA2ksoozqxPH3NZ^l&V
ze^t*{Z9Xp_dO;fZf!|nps<eJn=@r4(o3Fk8BO@p|@asQ)gQIa&cgb;B+;L+vvp4=C
zy3<a~G`8E(A9v#OsTB1#qeymwc#_2iLd~H1$At-=U*hIdyosET49TY-X%jBId!)q7
zKe^#zf6DA|B^Nx)cd_cHnz%J+a3iDPZw9Ku<Kj8ZWz90+Vc6M9xKcL4@Fo`N+b<vc
z^M}2>`Lo?`Y*`N$2g5j|S-zo99j}Tj;K-Wr7`iJB{g95+Ds4hI-dL(|Lp%QIf}`CG
zBfry*e}<`kv&7i3rTg7D_~0s|j3HP-QCA9TKAaX@W19TY_!71?5$b%I*@RRfo{vQo
zd3UCqAgmxLAn+iJSZqP45{Gi2Z3B^l8%Hm&%y91byW+A6U?6yiPZqT7v>K`$D;$Cc
zF3*MF$%Qe+QF<y9MI&Dm&PtaB$(RTeDuvL+J2F&O!Xqfc&AT%x#@ApbG(oIBd=PC-
z{8JFzsv~8&6cIzYEa(xDK(nUCmnLcvFc+s6dAuOIf<!>DcxCZd35Fn0m=H}|i(r9>
zt`>5Jy(eaR02d~+!{}ULwCEIpHYZ}9FfakCoh}f4)tzk~OG=9uZRC6iqp%1VeS`=h
z1piC5FYQV^j=QulOqI!d#qvl=+J^-`1crs7if3U{c})B(?JKPe-8RaYK9LIp<)bi6
z@VNj}jZ`Me=JKiCNt2z%5U1b|c7l-(FhV_>{DBbyEqc83&sCiJ0dq79D+osfhk3bd
z5*!R6=gT-CAI7cHqs+vg9v9)OMG}H9U`jjE?|~0b+WF<d(Z4bZSXBU#KyJT6#zFua
zCtyQN6THOC|9r5k9OTCbZiz(Fq;vUJT9r?RPk$^<xqOpK+VwV$q<s3fJ>BJ!{`$tV
zyh>FFR9qb(Gj@FbieDAqeT;v9-`aU+_3gl~9?~jsD8H5AE2H_x&u#gOpWlSwuLLXZ
z6h6QCrCh+3FyzW#iGv${PP=-t9*${yhNIIOAIZ*0LFh!C98%>XP2%BZchYda#3K}m
z$S;%dw_D4J6P5>Cjzb9e3{%3Aj>?Hcy5q*Lor;UzxVe2W4qVDZ3H@2Vv6nb<8K!Ly
z%8fk>>oggkCB8j*$(ywJ#y>k#dOJNE`C-LPf0<4BbHFSYhT&J4Do#eA?&bTAao#cx
zt8l)1LLUysyV&t}7>qxN>k{BFY;U{RqIaQf>{{H?{k9Kgaf^RC=ePd+QNAlTJE@#i
zy6Il5{MaeWb6Yz*eeB_fcLmF;kCLuY2G7tWe6u&D$+v#_SDcTYKK(}Ub;tY-X8n!%
z62|lNZ^-2MVLY)y$@!4vZ-l7L7@OmmBhh%GH8>Kk)vCq)^`&^IhkG_=rpKeEr`@Q<
zVw3wgxVs1~B(!IyeHr}h_=p*f^P*c|(1F`whH(VYa(h13ms)13L9@doXO}O`3+{D*
z`3b`V`UUN#=KO3lR3>&1Clei3TFxVs9qsl$%n)m39hq@M8tnK+o68JzeZ%zX>~lH%
z{!rkB;DMvB7M2#G(W;xqIz2(diX-d>N^hX_xyws+n&r+iw1`Gbe{-B7v=%}Jn8&75
zVb<vljmWV&J_2mM=JFmL9ja_crY9+PZhS+Sfu;nsKO3mn?38rwq3VPn1{!cy^!F0?
zo3tCI`8W=!e{`a$WX$(@aWo)8CgxgpTytC%3^q25ZM}FFPoF;5dgmt2ug6ulGYd4h
zJBqt_=H?Y&Wke7KgL=L;Z$>M3lKBWg)dj{8b@oO6<D>H9E|Uujj5sQ=UY`pDkkF>=
znzt4o_OTFcITBN9H;%<RI=NIoLt`q10cer{$Kye29iaIDUCc3FkhW-=!_-A{uVNIZ
zFe7PS+^)5r5a<j`v|=bT>kBPMl+S!L+buJ6VWRtZ`nj4l&0zH3T#L>Y&5Txocb21J
zqQTH|-=2&5tk!W#kMY61VKC!J+nRrJ<48B!xY?NZk(7&qJKBbuFP_KX;2_rTt*d>S
z!4X_&v+VVH9)GD-Gh;ZzpgTpXgby^jyEtn-jZ^if-0a4>0RP0)Ftv01$_}7pl6GNS
z^CFjA8(CXYF!CBn6#;H+c*CE~pa@+kBgfPs1m%7ln4*rWz7GSQu`=i%q$ZZyi7^5L
zmSYqV5~d&pX62=Y5eT_B@)HIl?z97Ix_-aswHAGz`T}!cNqMNu<dn)oTg%UkZNZpw
zvTkBM<ThaaD?i%Ff@R?<*6#>D5$^Df;FSL5HrLqAH(CG)XxS*Kw8;}(B3M=Uu=;~y
zPFMnOeFOK>F1hiv9JNaAOy96clH`-=&NHL+Dfh^5v>F>{Ir@?^!(>h&1cF+Pd&Ofd
z2BW>vZ22Cd83eWy97U*cEf_Oi>TV<YO-+4nK9&#~sn4lGu8*;y`cID8Dfb(gy|tRg
zcRj~jt~-UJ^guXyeJc2HL@9g=@Q;NpDkt>=_W>KUV4L+uEG)KTb#*D$)>cvwC%iF(
z<}OFjsVohR8-xT1g(<V|j%wCpWw8}&dRBz@mlhgIYhL+{88>h3cC4*~;0Daj#s>VC
z78ZO|9|C#iAp|6?c2h7?KNU`3EFi=|xUit{%n^h)SHf9iH!;PDKl(dU9q2@81dqm5
zboaIe?`zFZqj7n96h|CI*x!%S!){z24`OQcDwZ`5u-Sq>hA@Pq>^^+&z4+*Z_qF~W
zQQ5c~ZB}J#TS&suhRZ7JzxYr7Mf}kp{!x7A+ux0s9z8H#UR_#^S6+T4PBmt@50AT|
zjt0G$MF4km851K{g3W^J&@`U_+~%WkH6PNR2n|~*4^Pc=ECQri^H;WZ_WVWw!BL8g
zZ-m@OC&y}IPB4<&rN*4bEGUyrAmC@(nRF4pb6+GvNCc3pON+7CLfFrJQDcgC?4#z+
zG){a(U>RE+57<zfvYCWad>BvQ3I+%?H>c_#Xgb#Xo10mt1WV05N{2oVKTQ8|cHEEs
zt>-a2IU29L^dL8M2<FUfXlV_OG~PA0a}o_&zz7t%dyS#UJm7l~8GEV^<)Ti6Lx8xn
zz+GS(f2zl|=36$i(`Ohv93#l_%D|EFjgTEJuot^s;b-BttFgEmortTE3D1}3TAyBA
zjmEKX^!fEz%+wk&S8v7j=v18E$S)>kpU{|8-xds)mIX(>TkQqeS7PPfJ>TU<oN0~a
zk+G?`B+byg0)E%{vzEEng~B^ylW}-*6)*M<<LP!cdWRfAsP+_m8X9W|mBv*ccwJ~p
z!877$LU1VeZ866oaOO^~GtFxVhMsTt;)5rfu`T-NTy9O(eNl53f>VTM2#t=^*C#YC
zfDQNT?H?RPPjmts6QGlrYiEU@YZ{A-n)BQ0D`?GMU#q=$w|x}ky|u+yUv8_coPdP}
z>ZBgg0%{|0W!?vtn%dA^2*2o)=FQ39I9#|8Lw=VUtMJXhv15W8cnX}u1HHi*M@Z)*
zVpaZ*`s+5DVcWtf`x*<%@9g3xo{Q#tvC|PvAQ+FSZB-xWFy<PB(DX?*k#FtwV`Jw)
z?F|0cLk=pNaVt}SbNn=e>!<ttW=xjlt9!m`^Xd6`&Xse%zpk?U>LvwWufOpl_Xh^7
zI|qfA13kDZzrI~D99%h2BaNuf(PW-w#c^OR<{Jn3@STo!%P@+|w-X><<);UXNP6^H
z{P@G&f0U7ek+;bj2TDRfPnAY_NQ!z5{q*N3PP=^FeXCv&Y5K^p`lqJ(k3Yfk+iBY^
zA1BJUUS`l2cj}IPST>M+74MU8J+eAZ{`u2wVFn)Felo0ctm@))U0Of(lV{qi6y-+!
z9Ja*6=Ko@MSkl2UT`-DbH~y%LBMn`uKgFe7_!|%z%K6EX?kWIfDB(!gLDa6;ujJ1X
z;93iWF-?TyT67}-8)24JY7m|5B!IYM!eI9q6RjMK_}D=OA?>@9$Pc7T93e>-0M4_-
z#hWxK4oI8{oe7%r`3-Zqdts}(FiEqxI}%bx00fcC`R68v*)gLuogaoOI$sg~Lil;k
zx`=!hE&Qph>~w@Uwb0`$FqEW}CXGI(MZ~SY`f}74+A*%Ov&dwD@5QLfM;%3P#;D+h
z<`idmADZbf@S>KTQ|#vCXc!jr?AoDU&WjM>TxZ}sE@GUC4q^_tA{+$Jz#oB`g-fb3
zy~XDUD}T3-%HUN~SpZyWvI91TDea9v5oNcEv*3?li*cC3GA3gPWaX*xZUzSU%<Y|v
zMan^PELb2^n4BwtUoOOfnCfo9p3>5%k3Ul+h%y||ZX;bUVoQ9=Oj;~d*r7*Omx3qd
zp$=ZeXi?{*augpt<Fd2RbosJRFn-7%AtQ@J%0yWqByvYj_BrE7#iPHo5Ted3NP+z{
zyRB6Q#=LhFxlR;4>v4Bs%~;GnODJHKHr#To@|AJ1lrQ_IcZSXR;Eo#wD&cR#*wb!)
z^KG1#Zg;~CpW^T2c<aCXlwTQ#y=!+(hW30DUaqb>f#o*LyTs$Kyep(Mj9>h)<L$Q$
z`QV;M&ZQdxGMN4-H~FM<hAAJ|^C|AdEq&9is)yS`_7t+}v44iKzawQB$E)1)HFO7%
zPLp4~vOn;xw{f)tJ3W_OpXr;A>uVd|`ctHQS78*%F7$^IhQKO_LlRCuPG5iIUzM@)
zcbV)@87q6nA;9e){`PdPm@2H@%a`K};l?{>LpD}q=+bun#~&A^cn8+92p;)u2j0b`
zwS+BpKOPfUtM7cAezDWfOlvs=;qq-8{@6=c|6#4ll^=dr?d?Xa>cyx2%HYAV0O~e=
zwsG1C<3Go9B6<^{{6crhGhK?QELB__6wWph;#;QDe$1~D$*;<lRMXHO_)PaItiDQm
zezWk=*EfQ%kDI^2tiLf|vti6bUxLvYan~ap^OWy^w$KCfB=15%JK)*^jM0%R(MNfV
zU~P8Nf+!dzFccYkYzpG&9Uo)G@q(g-Ycn&Z`#4IC`=KT$My$02MGg~vd#@WCFLw0q
zc(?l2_HMk`+>ZVIJ?~bBwGT#tmgIgLn32$D<O@v)eGG$rU+HuZd?_5exA*pY*4X0S
z4KDKM*eDCb%w$JU!iE0xzH?=HWyMTe7`>)j(JVPUE*rJb%tBi&8SBt&2!Nnbd9$W~
zPW1*i=zY;tFnkd@t*xy^vw>E?g_)zAR^U@6M6W^D!91<aaUag4>05+d(72vw0I2eU
zF@(m}gGcv#OxE>H9=~|1@IJ<DnzhHM*A@7|97LlIW)1?8k3Re`UTi#5nxg9_Yijp#
zZ{lOb!?=WrjD{YFlU?;*kEndm^U#*yOoSe2o<W<EE=+B&leEr4d!x}54KLb~G9zRG
z=3LZInb;)`GnF?2&fQKw1{S)(<Qi9hQqHQgg)4)8oSvVW9%fyRMi6!I+Dz?}Yo-)D
zP%hK9ss~ItV8`5xz=5(cpW%)m7s1pVj4~L%>YFhAVaTFQMH{f$i_L`l>PI{K9dFo!
zvA58yt6tFQ>^SE>8Z?fQDNWo%^^t>q-?Ti8Ycxt=NY-ZA^*$Ou{_%hCczEr#*8<Hr
zG$-kYtlyyL(fr_Hjzhx(xS&Nyf1zLTRL7gMD;?O=&!{`=l<Xp`Z)hjc>{=7xPW@q0
zvqnhzKi3M3dl-iZtJqXY8?YX>W}-EOrUWD9p)g*bC;)4_3zhj??Q*4Z!7OC_u+Cw9
z;(h?*QfadeM)1fw8qKY$KOo43$%%$co=S14^#~_$Oxg|%8Ef<hYi+4s-tah%w$zOJ
z=e*mAGEw&w+~wv;3~4cjV1To}LKrcpK3cCgqTOEbQEO;QRqHUX=k!!Gv0<jT=3$Mn
z(Guo*(nm_&u)|I~Gh;PIxKk#X{R#n$1wY{C(_Ro(SyOLDa8{o5Mfx)JRyqhC5X7O4
zh23qbhgBp($X0tv_+up&7FF(MGioY(-rGey@Q~VB594HBW0>)Oa&qi3K>et{g{2A$
zj^OCH{HE_yC)I~M6(7MI40`SoL@Ng-IBh^*TWq(ZP5!`M;Z57>vEWX50ux}sc*LD}
zv=M0_uyttqmukj1=LkWxR4&w)_B$Qn{jGSmv1JVonClkExqURBD7I2}e>*ur<LcOX
zkv_uW>FM!k3Ab{jnuePCDEB&YU*eg^6nKe#$vnq#p3FfZ{CItF8l&nzY{CHNFxEz{
zFSPzuJx-6E&h=&DuwMU2W#Z1GIrZm>xH!|?Fc9t)o<v}b;AC;36|cVXa{TV^{%-v6
zhrb=`_aACLS&jSm*5kETUy1L$_3il98*f-c=-$1RxWBp-FFm{$FWp~@rFJ9g8Y}be
z+c5Bteew^-3?ks-F2@J!E5iSavAocX1)g@(IDq+dYje{$lRFgGmfNW@R^wPp+5ur@
zYD$Agd#VrZ1YU>t0)N1By!widRYee_zBDh~g{Drc)l#3C^igw%hnhEqv)Gt`<|Q~7
z?L-7%Xvs1cd#-?opngxE0<Uy>eQz8==!Hfs3Pv1{$he);xK?+GzUB$Ehwf>PLjZ?H
zAlhWeq&Yser!)@w2gV}^#t`zMF@!)B;V|<b8!@akPTy2tyVlsGt+|T|SQ2)j-BS8X
z*4Sf{hUZDm2?!_W)sJd*)pw>I*JINv=UAL*isa6>lMCTsC3k+Uc_zj!;2N8pRoI$n
zrJ1;ynDw#F2(xBtb<q<m8nZ2Hz;c}8mgc+XJDvF8>5F*($woXCt+OMyjqQ$)?K?b5
ztv>Gl<5*h8#g@W9(YScNz3+LxCU|g+^Vs!yoE`RJPID5QhUkaTOC8})H1h{ai+j2d
zSoQ>WHWe@ja=++amm>kyZ-wJqtwyY%nXfq+Sb^K&Ez$mKOD(nOw0WE-8Z!t^**t>=
zV@t4KY|LrAU3)VUf<|!Glp1mbJfQi0(MOD`9W`FinncKas__fGq@u*p>ACOnLg0;N
zGHplw@>tag3jk}zox;KLgU1KTLQ`zcnqmzf^LhXN{dn-;eypqtPYQqK9FIUcHycn-
z1Z>PxXp<tq-W83B_B?mQf&+c3l;93MiN^7v8vV(OPIUWu*O@mXDWuIE%h38UlX&&3
zGhb!eUx)d0_~aaGTY3I=8w6h?|N3A2pCBQ!$SyOBDgDWjZ+8(mB$|BeE4w4uT`$?0
z;D*NWieLW7-0`<5?jTxOnI8S?0#tr!O}FylSEiMdD*@6k8AutI-zqIn^cf}dk{!RS
zBOm<pqhEcOpf;p%_q&9=HFrL4t)x|~pGd6Aw{jpDU$~~ph#97o$K@(sypx8@5W_a{
zYiMqRlSlq$q$I@cgOd}}Lk#}ymY+6cEXM@1?34v}6J9FYRlac~3ceL1FHNbhkR|y+
zh??*$aUdWq3L=i|gcVui{6rzf6l&H&*4IMH2*l2r^o0mPbS!Q;V;_PFqMRZjfN3wZ
zJ*kef*jF+jWfs2d>hw-CO;Cp%0ns10dYao!h(Tb3xUGdWJ?HQ&idiHdD{OafFAlWu
zy->aoHV8N%NU6DLFij2+#-jIJ4-*lp2!tL)J_O8_%E96YfyjXS0W^`g4}L`kFjlFL
zDF3CG?#Ej{{C+f-7DO;{EP@u?0vr=2#1zCVy5^%IV9!NRa<L^lW{-|UC<<m?m?}}o
z6Nv9|5nFW;jcpP4?C@d-BLtn>QG}2e9)c}lfzOT7MySJi(i}NCrHPGW0oVzOP?C!;
zS(I341tUjx@;o~`5nNe3BLsr5)?!Z+F2^)l;7Ezoe^_ur2-6@ep4pMf<je7l)}Uh6
zRyZbNFA9|o`O$tu945QlOz9GqI`}vU5u$k{A&Z7w1bJ7TBAdyi37@e9449O0rydZc
zbVT)9!8(Oez=#E-?_+RS!Ap~N#jH3#IaB}atIXwMZ4qhYKgxneuu=z9Rve9!M+#lB
z14m(}M@10TML^8VyWH@qT`THhC^42xudoMpS*t#%xBc~5ULGTN9cbG$_DyRsw=VSm
z${}Omp16$ff+z4OreG=CN@%|w`(xqKRQM=E$vr>I_b`-ys}f{tS%wm;^yj<j$mS3E
z$&)nw?N{;RX9$n-EkjTYUw3jVzPNpQ=qiiMTi@Wx5~}jd&{dil-md&Bj=v4l6bhuj
z+)I7#!p14w1}blU2(Onb&1Xrs3aD3=PMTq&GMSfaOd=(Tp}zrr8}s7<%BwiJ3^))6
z<9gk(XM%R)>zcOGUo{_Q2)X!^@?`W+z8%mpi#?+&LA!iPSbbMvWUBTuJRP6!^sC@x
zPz)0(?haG(#iT<Tf|XBR`6-($^d-{qe2I}g>9PD~KCe=}&8L|1!K>;6x9?j+F^7G|
z{=EG;G?g3XuKn(ZV>+?oY<IaTZN4$LX;d!h<}y~EGA_Ct2$%l;jsNL?`FD3heYNKQ
z@_+m9ZVUMth5#<w--L1IUs%L%%&%c`oc-!BuEyza#3(IBgQ$w&&Mxlmx>V;In}AWv
zj$4FSFtX7+7kzMcZe|yoom!%~mPNO5Tpc^Ek3?^9uM?Yk7<bTK&@UXt44uV2U<jiw
zMB89eBOi_U8`0F4qG`FuBN?YKVxc`?kg_9wc1Coj-VBWHUeEVftgWxcYp*^O?Kf+t
z+Ka8-*xKPp!k%|0pL>Td8X4TJ%TZ&KW`t07XhqyOCJZJ$5A-#3AB;}lQ9)hhMty-f
z$B)AxXT8Clv<TjyD>Dpvz+{ApjP}%&8OSgjVU_}r{oUQz+1eCM)-&Cwgv{VU5VNqj
zY{ouE(2+Ov8}VR3PiYN;(4;L`FD|u|_M{n>Fvg%~VX`aJWF$h*UR@~eR4laWF{`+H
z+b`mspZ`40MAzQG|1eq$D}n*X&~b7?wgEyR81^uT(1s&@uqaGzwCxZy0mq%~UBiW=
zE)nq1MwgdonzF8Aaj_ZA#+>1YCLO!{^XR$@*I{j`9dEw877LB(*zFv~v#q|5BSM_H
z)M{A!1DYC5ypNu5Sn$L!nVOz4(~2^p?XcKxnr7v=yS=@A3nwqs)-ZEf7r~^Xhgkrr
zHXj|EFw>8=0mdBn#=6T#D5<}qN#=Wg)W_7--9OYdqN{b<L3cOqtuDt8zxzf!Tx<In
zC-kFw2S?sD&hb$QCJ=nR5Nx>L4uKC0zuMGDoE>#zes&_>dFQ9`vv+<G|LFJrAl`WE
z`#utpi~6~^pIz_^tvL(JxbyDC)_#2WWIMKZyB54r7}g<-2^jco?n=|z3+5!kUTaio
zTp=t&7<Y19#-Y|E7z<QrL)Dpy??io!wG)g%j@9B+3xsM2@z^nskkSDK7Z`%)7YK__
zEbO|vIM=|<O?&idV9UCO@dLvgW@QR<xi@J(mKKE{xNA=<@?EWcclNp#o;<q0Zh_dz
zIsFVF4+7r{GjCZd*5~G<-EJ^M;`IDleH-R;3O~Um2p;$M_Z=3klHA;fRw|lOjJp}*
zBp>yzcou>h4k=J!Oe1upzo2~u7Oa?`V8aw5@=n28wf^RK%e1ozlCh>R^kBgQf#JmT
zY|PB2MiyFIl!=ukC>uO$cnZb{R=|I9^0-3~1xU-n-+3T(Yqnak;QNNgi$)ynd*sar
zb+j~ugSp3vWA<1V5*eZY{$9rdT%YoxbXd7DCa6Dx(;4AC1W5=mEJ#v)r_~=2&eImt
z)8O0GY^}}AdP6Gx4_I*2Do5CHXA*b{xFk;mCNkqLPb49InC7aAmJ`Ay!ck~_ppA;Q
zF7V4Sdun{c7|o+-=PjJ`27n9hdP<6v^{3mPZmqs)jSrQt-JBI(Kp?6*A+S>ag|wwl
zfHyhvmT|+$Iq)Vf&(E}`y^0AfY~dfEvB(`q2!Lm2#(gZ^B)FV*RYHfFYq(PgEkibR
zAmpM!pKol&`yW1WfBxb3-;P&bedIBqgsip6oJL=p7`rm_A1y=Lgd;mI)bD4e&^nq^
z@M(>&3*o(<aukee^|;bp)7N-LYnkd=AfWc4odgC8K?d{%!I$>l+1WF^5LVn@Tk>6L
z%mD~x;d?L_&Z-^N-Ms07&DWPo>ricoHWha{QC{Y5@B~7qMh$_R#-VUJRP*NMo^Vp%
z_e4=9-~XpK$6nG8(X!)?uxBr}W8=l1+Vs-6bY@QVRvFmv41Ql+T#SeJ)--?21$U}3
zH{56r+1=botvAgBXa&Fe+AG$`-9I>q-QKbK-Dz|V@|fRdy=Fn@>>MZZAOJ<U%EpH)
z!Tv({_$0c*(VN22yeTU<0O4>OP0D6n<5YN0{oz1q!lOc9bfj=X>7EzqkH8DwAHq}c
zGjM?hK%fYmELc=uL6CiUnHs1!vNNV=k28${j$}n^3V2fv`V;NQ*jj40vK|_bqrxf2
zgPz9QbhH|EwUJ;3u27$cj+ocHg|H9Jq>GayjfpGYU57R~+IZY;cC7TRMW*&V7S4ua
zgeK^M=4v!28NW|AMTc~bV$5TDJg(G-MsI{8M=q82Xli0EE_i$~HugH**wmavKLNf|
zlOyrQ%MUbX&ui?9{$YFy2JlfA(2UkRy7BC}@V@Z3P!ht-b>Vx2o@jHv5RJCeJM}Sw
zFSfb6?^11-G#^^2v=jIjz8E7P9;}8%3IqiYVG@r8XgBB+<~nc>IEsGAQO}Hb;ACN&
z@PVRhyp8*)VrZtWXeZ{hdG&Gn!a{vEmPD7R=yA~Ny3ak`MqrqH!NfqK1C@@E3r4gk
z{o}72TtD6C%Nz12`TR8lp8qP%=jHvIFrN;eTN%w)ZPeSCkbh0{qyOf=92xoJfAKGq
zk04`~nL7uQo8Ell;d+%oGOlRx_xG(;hL4=cZ;ca3tL(R5bja+Uj_-8!ln#m;KK^om
z%495lZ6`_zV4q>w3{|B@vUyw=;mS|-F^rU<ZHoVGazoqDI0$8Cr?Wm>{M#6J3?8=(
zXkxmow&7gG&vM#76Vrc{u)@e6@B9VMw`pWu^^Jo~@o^d%1EyZWpoDY8J1G#{@ns=T
zn<`yA`3EZ|I;lbsmUvbDtJ)C?WXPhEhxBnHFvPJ5JK51nH7P2$Q5HTz`cs<hG)u*m
z0$vKfAW4WAE}w>ghB#&?8$>Yu739k<7YJ|EcPa0z2wE;d23fM0hfuTTk|vGwGff)8
z@En&xI0y_AiYhmR`I!h5i0dQSj}Nt=Rz4ib;R#ek5|f4(5{dy*J`th@nR~&lI<qK)
zxP*{IV+X>Y9kAY!Dk8><=&#94YJzz1<{R<a+i%6Z;Dg|V<4jmosJh<ifq?6R-7Bi|
zncl$oP=qoU(H;s05F0ik?wI6I#hw-M)LvfpZVhY835N79j*;Lf40ftrt9`v_fVj|v
zM&AJb9LohG!;1_NMO>Opn^{<-e&V*)B+3Nr-9To@<N^jl+6;e{-8+`c1P%j)G}w*F
zA_0PpcA?`TP^cC})HxR%>UU}@jQ55B3;BEUmOo(_bJ@;#r=2>`pCHoFP(vWgu1gpq
z^l=tgu8RnGFUVC6a^cqt0+oe$X1a_{80M6t+BwUe6D(#C-?F$u0LCu8aS;I{db1$*
zF`N(^B1{l`iEN10{6aM5nijCBDmfsXGj-FW0Y}+wFFwWzC5&BF?rQX=@W-J_S|vVj
zzw}e(RSb5M8pFIXyzho8rgEx+If{PB)Bn{QC;iGR?JC_}d`g+|wz~uHpJDZ2K88>I
zoA_&3SBFX$8{*rVAuzdd0Q1Svchg9dF*0^#J{bm+X=?#5kE(1RE03K%8OuKM%HO-+
z)wr;c7b&^wcam_Dwq#CKeih6zv2;to@-B%wp)7Thm$X9$yK5X<ajO24f~hP^8s(};
zbVX-j_jJpscOW;+8#|wje}z?;8%-XbQ}8NjxE@tGl^*U{;;K&k<QMK$Sd7Elf_)gn
zA&aH*@KgDwDPb_w17jzCN?@!Wb6EAJaG)3REpf8ESoC}%s{I_yA^eSRe(6_!#V7rS
zVM|Sl|7{ZbQG!*K_;I^ntsF`?hjHe1u*otzV8$sA!8}4Ul5A<Jn!De~-QUz5PX)`6
zx4)h-6wX+;_!IWu{=2VF@bzDR3BecUPBp(?L+MK2`Nn)T=59a#RZKOue?7x^(Z~@U
z#SuZL2s{QS(H)$M&Ki%28SX&gt`9xa904;O9P<Wa3nm*!0wGj^@xpH3!(-7Jdagxy
z!(4@4gKk4e((fOb!8ie<QuG}o78*V`H-+dE7*z<WE=7lodm|MLbZ8ad)1p9ES9#R)
znP}~;7dtWN^<sW*DjwcjiboIDEEIyCMDVcjVq10L6o>OT7R`qC0KyKmJhINoxVFFt
zZ3Mk3Cp*{=RX;YfK{HY~3k_5b8_FS?j${2`(!&H#9Fl(=PXs*$BMR-YrKKgU7fxe;
zcQ5(}2ZBT2!YAlZ)+*4xFtnCdRxKnWPt%=B6J`}V&^bB`L0SKx7Z2{O$Nh)v6vI2Y
zVa&}=Pa2j81%M6Zf~h<`Jr);dC-K3%Ka0m7z9044e0=*mZ^!Drm%L6vn*=(R&3>$F
zNK;@C%sATdFgS{e%~xPeAF1UwK8fwE9mN}o`n=jv@nOy(1Y{@ulG=bHwT?yqqYa@X
zR2I?TGPK`uzZ1I~AH?}lHx?FGW45srS0j^AZ?t_k4LiwU2C@@-XTPg8R-TAKe>am=
zG;O0XA1lj?)<R+Z1A`y!KaSL6CpT?q;f>nJ!U(~b9pA&_xzy)q&%FC((i`f~4S;Fe
zoSlqG&7WvR_*gjg{j>8Fj67OjjMwfj#LPHCLWIN-JDo${tu~G}l9|#Y@l1X5@zdv`
zw=d(hM{DuY`a&!;r{nSC_v8Qm|M>5HG}gcTSN|rKR_|#&acKs*1*}@zO(}Dja2)Bi
zvDJ-d8|vqSBejnjM&-0NmC84%bqb7Mv_62%#`BGMvAyH=M@tET3xX)x7J-Bu?Z&ZH
ztgG1djzG=p48ey!&KQAsKnQ^lnROa{mo+hWxt*RIdo5<ckn&88DCnJnkMwW&DqrmQ
zv2KFly1u>?FTeD_ZHr(FVI@M9)s<zz4dGD43`|PJL8}A&LfH|>De;0^&diu6D}<Sh
z5ylh8s0}n`5h&0<sH?RX8x4<n#uMvL)?t?_J7b;lA*|zM3)XQKG6`4ZIG1Os5B2go
z58;l&K(mvEH&+T4&X4qfoYrCpHfA+0xc3O=K79y5bP8Rwzp+F-L+eQJqz*pX6nrT^
z80Wy5IN%f719-xV08da3+6S#AXqY+SWHjf{s7(P<;soG?W(R9SHI&mkSD%7`jsVR^
z4z`<KOCpdkF4I`&ZX4rm#X;bn>t65(JdVVNUVp0KF!2$5!5C+~DX_cDtV0c3J*3MR
zMKDPK1iZdGkd2lqhmuk|bBE9g!1jDV7&ac$zo;t$+1&WarqIL%hnlC3Ef5|v-e}h~
zR_8goPxx0&$khd`W%m&zsf|WO<FmF$`xYJwZ3V6#SDCJbm%#TNYw8UW3d@miXfM@-
zTbcLJrg`#w%Q&bh9K}(IlnE`Sz1@z>_VP>jExhP-Ix3g&dX4+fa<j(N_(<$-Zz$i<
zSX;ZNHbq<bDh|3k76v!kOB%}v&rjmL_dnFQ--vq;9(c^Y_x=a5-|4x}FR!k~4}R;n
zqh4Q#zTkc!*fN%Z`Ptc3>}sAt0Lo^Rr8Y;dqAkNl-qRQyoyqS;Y37}o;1TMBR?yyl
zKb}3`7S7q#I7=<AN%aNzU-#D*<H7wUA595NH?}(QLeB~ODUF{C^(lBGXnZd(HlnRL
zi+T5h$Is)r(mVyPiGE;Cr2jCknR6Ex+L|xw7AB33sxN6Q@9%8I*-<Z6TD5qvwxIFJ
zQK?5US8v9Adre@SixYSh>JyY7!G`bX9Q0#CFj!t(^4JGoo~Ulz(X`(`jm@0{3&6PF
z3t=mQC5{Q^9=2Aqrn*j8z=UuUO<#@;W)lH;_)7VKmk=%^@MN<Ad>Hz-iXa>SZa_$;
zKCW`<BhX}|PqUHdLiS_T`BZ6fbgVV%HP)Ed;OQV}LSt(}M9!2d#rNtG$9guCZ>BaT
z+(N5iAur|k{f5Vf7Q!vcolPmsS!_JwzOI4t1vN3|nD2lSg7U;}Dv$az#|(4KADVgl
znm^Cf?rUm~wdJPfEADxlvM`uU)BC$SO8+{hX6wR}+)sEOY-U18&(XoxYX5EF^`{?w
z6w8YHgYUj!x?pQ-PdHBS7)?3}9E1MmOsy6D)9d(XV?Um69%v2)pCWY46POSrnkOoI
zZd@5tzlAP@2IYuaja&6Q!&ZHpdshejW8;&$=Eb`3KJ&Txk{X)`?h($8b6SDMBKQX$
zqE9`cBgqIoh@KUWQXJ;EUG;?*JJ1y=M5p{<<m&g<;PYwbv)sR|`D8wK&99sG7n#q?
z`-@B!w<_Z&zbo_C>%chs_459*Ml+TA+OKO;@b#bmD`<K8By*VS<!bOD$E*^L8uXeu
znLH~#27bj$FBW9j`OUb*a~Pb8+t3s@hfx9s9*N!w$7k+em%W6~Fa4$^J@|xJaC#Xh
z%UpsvE@3ji{LtgTvM3GefTN3048+#3UPPdrey2YkE@OE+jmnRAl~{>4Oh=~F+;5D7
z+jg5onkwX7bNiL$$#2(BiR35lc2@Sc^+J*k$(Qk{ca{q~uD0@13gdWAGs{}FQ<hr~
z0UY-h?)(C|Y>`~3Qm7mtB{+5-1)@A0b?6-knwTa;5Ke1i6&8dV(jrR}uO`I{E!570
zm=VUH9mGzzGZEJas1QI*X%cS>p|=)Mw-%ADG^aUd{w8;L(w;2*i37pOVvETWwQQ;~
zqlFMhW?1l`ML!ECt%|&xlZlea5zQQgAuKxB!Jr>k0JyWb=4eB9F(Sa=taD6i2Hm&}
zdBHD>JucE#nqQM?h*)_2x4s|iuRM$y5zzDkCSwQ>ngJo=;ou;SIM-PdG%z~X<e5x`
zffrn87BO+<;$MU|#GxjqCC&<8To%z?OF~KQn3`K0A2VrfJs<6;^e_<UFdrap!87D<
zff#WV*IHc)Lz^ND1MUErT`N8gMsN4W5|{R%9>?sEVi$}Y>5D$fLHU^o$4%kri{9a;
zZ;lc`i;J{L4<RUN(oQ_pWS%2XQHGOyriB|sH?U!`z%HEeX?ABpnCA{?b`Uf1v+%H%
zS`p&0F!2J6z9b@sJH;pyb%XdgMB_^Y5xcL^CX<)@8g&}w7*N3<Z7(y0M3`Kkvh!y=
z77%>RH)L12A)?e?#3MfEMROkl<;qc-@MBSjNuR9i?N_mtan9IHoREHeRDulxI=tMf
zbQt<J_{r1xR^k2aBnnQ*_!U0o18tsRaK9-tsV8<JJ;L2G<1=4ODZ%hNO%*O}!|(EP
z$z33O$}W!#n86GzJa1uh*B>)<FDC0PlhKUx9EP+?!#W>hLYX`H+c<$c4vt)`c%}8$
zV*k<t3Q(1_YL#L55{~>FinQ*e;dqWO$Nbl)zGNXPKMwEse9F(^)7|MgB57BF<$}FL
zzAcZwadtey+iL&(P0tLEC?w-y8#W21%I9NlvZxHipmeT-{GF=Z`5>@DP}Yp3f0zZl
zhb{_%A-&?DpY)X>UAt##@yYNXH_pn;$zrO?%D|+llFn4B6spwH?bcOaiUUq|QL|zt
zT82-js;;uD0K+jWlW>tCTos=*i%<2=AepnjandI!CgJO7Hkkxez@7A~JWG3EDqJxX
zDBpU?B0t+JhuacizD~i{fA!yfIl<Sk?4L9$>6c32D=+}b<L_^(0Dnh}VfZ;fO#d%$
zs?fKcI9>7^ET_H>?IU;78^%XM9hyL{MTa2t;=To#`5ftUE&po7krFg3!Wwqc!*m_=
zMRSVwf#EfSuwY_Bv=Pi&HrGswew&M#xoK}Of=SFSdo;%8=j&#CasL+cU*1Q>-K=BE
zA1xK7DVoi~3DNJ|sR|vI8W~A@!65csCEPJI=(~2@skFGb5G%_|7MgW?-8kqClm^V3
zh}jy)=ggXZMyTV>Zn8s{fy1o%AzBIg{8Y3bv?=94s|a`z4?2!In@P+gy5CGy=w3(&
zmCZZD>+?P?5dj|?sJ36c@Fo2)k7)zyD#Fdo95m%tS65<XZCU9hy~ZwbnDz)Z4iCGr
zySMMt4A3G%*aIV+-QA~x7kAollpb@1NdOgNRPA(qc^)rbJc;+-eK$@{E~34#5Dy>T
zkF|UEVsg5sykWeCg(}`b4Fgm4gVv<ZbF*Bm55qiZPajJu8XXwsi3hlj2#zoa4+X<(
z`RDOz2xV$>Q$C6cMiemab`Rn&|HGfhrC>WVq4bUqMC+z7Wo7L_y!D+Q2xe<$y!%)s
zJ-sekD|^viY{lBzx;MRXBoci9Z6}y{zSrdB-1XSrg(17Gw$~a%?Zi~VID*MYU*_&2
zgkZB%+@Hd+d+c5p%tZfkJRMB+g~nVgw(4F}b+vv$_`15hY-U<xY9ywv4`MRTe5?`z
ztWN(tHupF_6PkTY{afol!Sczot=QV$GV}TEHy^}uV>+Hcdm8`!|MP!{KmUs#$8Y`i
zKaBtQpZ}+-+cIoHGZ$gLclLLqySJsbo{m;~F(#&JajCJ$y<aeMIZpEW`XWw_23n^G
z&dP6AeFCjB7`5DogTQKY>xI(mN3+?Cb+viD-V|)PFNyp2<}8#v);L~Wzvs0Gf+e@9
z#sX!5nMi!v*1|7@7h2a{To8`+7Q&z$2Ul7gFb;sT1r&lAb1Q2n`YI1&^3|6g#&2ug
z$kC?HEU4R+-;D2wq7QSd+ML$wZ2Uy<hG3VxSkGYJ*+1~HrQ}l+egSV-sH}cM_+tb@
z>QiXAz*wgo+$Dxq8^&5iYFBVkjzfklbb<Pa;0K<ANy@R82>cOZfdepTZJ;6Jb(W0j
zhD^ZphS%|`JHpi&m5t*9!EFMA*P5*Le9xHrAZs!1M+2v#dBwx}AFajC{=Uk1Xneyu
zFYiP``$p}iK8QvWV-2la@FE&`2xiunIg)I~%yu-fI2H~@H84Q<#yZg>R=5XYC)6%l
zP~^?pm4X5@AAP7gu^E&p2)+mcFOD}wqmVquH11d{a$*QLoCiS)c*`fU@MorY9>dSL
zVxvHJe=l~oU&#H+T4fyn_u&3L;~Mw}lxak`>yTskG&b23z=jA5Sb4xP6B@TFPrnGh
zm}@xpj`sZC_rLEk*cbfht8;U8uX$NxgP#!sa$GM0QXU_3t9%d$Oq)+|qc+yOAQ*CY
z6xy<kl?maC-R>aX`{;%87QB_l^~I6a#g`Vq&&|$jp3!<;dGGG-#^&aW*xue&8=l9@
zFFlMmUVlZnye=G=ZOgbrAp7bouc)7m$H76*LW`!xJ9kID^NaW6#j_3fMes>`amjcQ
z0XB76S-l^l6EkjOG_;wk5U3!Kc=2L8wl=pdoVmBA`EaQjb<NLg*6s;rXhE}~n(?!^
z&{UtCGLs*{-hQVaTf6(Qv9%w2-G1~m$1oqWVS?io@2xLJU40mBwf)Y(=?#RJ(EeiV
z(8p$HCSzUWd3|MBusSJPHEeRePCTG~O#fqp2f|YN$Jxn|$Nl;7Ah;j$RC8ua^Ww^K
zQ|ot)l>_1Rjg9!=@v}I+9*uhsUl#72x3=T@+KT$2>Zh>@Pr2P{t8a=vP>#?U9QVhO
zb=;BFRe3o=nfZY^2BFBxQakRiEc)2s)N0e5roOY^>kGGQJoZy-41pxPJld4ndgHL#
zTZ1@mG}qQ<h5P4boF^Jw3&Q);YTKsh4)S4hI&=0h+Cyq5PIyu?2o4B*=#T1)vbWVw
z)>oRIHyBB5069Ds?K8#^#^a_DEVNaBFkftx4fzx5Uz{!lzDjzFIi0ZJ*Ex7bDm6Sg
z;K-eCttMLIGoI(TYZ68RT5#YI=#DYQ8T80$@;R^0Phv`V^yG3hj!rK<eh_NW-|ydF
zjc<SJmFRW$;_2hZ@ydhMc>R_8#ywlxdxrOzXotZ;KmPR3{xp92lb>kLUymRD&hN$C
z{9<gYFLD3Yf!Ych)SD5&J<0<?J@d9VC8%y3F^$$6o7}H1*a)GqFMP$O4f+#;H380j
zkvSgBhGv}+fIiUlWd3Za?<}>OvAV*Bo7Bn$uXKAo(`;z%anvNIN_jJr@DKBZAEjwr
z${%yUH|7&2A!LBteDeFV4D-|11<!2DU)Nah^`HH#f2ow1poHe=jC7J=-DQ!7yQ1D)
zq^rIuX9d<j!Lahy%a)OA5hcVS91eI8NcOTpKLO?R@Gl7zV;IWG(m1`<cQc4TOTbLd
z$tE+04+4}BPDHN6Ri1XklydM3ftP7!9FA?!Fd>wT6E-4t0y5-LlC|9?gV|EKv53qZ
zc+%C?CZ6Gmm;5qw{THA(e6hLUO4At<NN<AcZH5&=Goar3;8m=&AzuC^oIQDG8L6-9
zthhu6x$3jj<BqBF!jv>wh-3=+rs(bqd=tRw6u%M%{Dtt#X_%v(%)|_%Nr<YdV{k#_
zs1PG}dQE9U9n<7~MWPU3dJQ08HEDq8j)e5lzB<zajzuSnQ+9)*;>%^N2){T+h2!B|
zFHLk1T3!SQIWo~9=z$>QjBXzpp@|-~Yn8&pGQu>}<$W)c5VKX)mCiMjn9@vwlSLdF
zL?Xyl9fUzNEy6K~ZeeS~-bA|W5Kt^iAWU;N+m{r4Q3t^Fs~W|>CqnK!|L})V7tw%l
z5dxmwZz`C|hd@sYCrz>koo*aMJP96`BEs2u%l%$gsuO}JHG$%*(KIl*sCh<kZZEII
z!qT#L%&`E6$cD(U29tngoT2g~_<>1_@X10T<x4MthDQ*_t|tPr<BB$60mEW~T{4w$
z3E?6W8jCgG1!e)$h=}0Sw5k{z<iTZ}go9baVuQX&e{+0|F&017&MSmVjDKsp=^+ib
zvHGBqt<qzcCBi1}o>QOjIq@nF_q<U_(gEpM8%<fU!-(CIgs+Z|V3!nigP;TkzTjB}
z8>8o`T;xkXJ!WUH%FIqs%H}pweSsxB7z>D%20~kJNDxIfIXV%onur2+_^5BXZFztY
z1)_#&arumY{m+;#Zl%7qS8jR|O_gw!hs?*q_!r!gxWeD*A%mqv8rtuKkrglxaUN#q
zB{xqVeB*CR+WbR6!eC46R9tCpzg?$c+F4`DX+P|K-~ML8_z)<=XIceXbS5=MnI7!w
zIQqA({G=-jZGzv1!J`C89EDM0B}D!LkP?`DDgLk?1{N><@P8Xr>G?+=7{+(TwYd{6
zBc;7GZ+^<+`9a^6Rvjh<EP6A?gUfNly9{vZnB~hDS!yilkTGz}nbt6KS?!@<nmd(9
z{^3su%0iIhRv{CcO!S)h!+8);ah#nCW+w|zOIRG-k{VAW<2)QS!%+89YPn_p1jMC+
zgwvuk$bq=oxOzK%Ju=?x;g598loZGXA_U*~;gr5kszkdh%*r%m<nB`9=g|7ajeg^T
z6hovE#gWQ~<2nt#Go8xUF6Bec#ett<Y*hfm$2PsQUX>v@b2a#g`e!LK<S>pbyh{cN
zKmFVa9A#t~hW7{QsDG(|CL!!Q8vo<J`=2wZufhB>g0D|26p`;X(bsMO{GA#=|K2r2
z!Q(gPGmYthyFmPeN&hczs?gaIw{i0u`a~~{BjY^2g?BIICJHD<rl-t|fj&TEr8=Ss
zVGtTvoNT}jV>T^8YjST2eG{4rK?1Z1j7c*VClGpZuSeb=Q*9n<G&q*1CA#k_dfmR6
z&}cd^7B7?@Or>g^n+_5Umm{7v5Wk{<M5l0c6AWY?7`8AzV9G;lvBQ^q=H{l&Ov4=^
z#0^9adGS^mVGdD$qfVtwS<v!2)*1kMnsTFEQRn0V)eRaBIvzm@&0=AT!g&Li=x*jv
z;;=D^<I#Cwz;fK*_SRM$_4}grla}<xDuhF#ZR@Q@+<S1(0y~)XFk-1IZ4XV&@s8|H
zX2(36;Mhn<>(Lg_rZB>Zix!@+rD;cw9PDnth<ASWSFyXd6HCi$@$jYBeOwf~#%F5{
z!DKA@{UetTjVb`)qX0!CioEnr{N<%ZYkP1K!Pz-NkORT&+DDLaR3uC-?=WXm+nMNQ
z6^ymek>G^Z3U^{HEpQAXo5L<*dvhm#{Fi@e?UXlPeH8UM^^-~1{a0p&j!(|S_2{$+
z>M7Ch5q<SDwA^TS;0cY(u}#pe6KY2^D-iC{M`4N~a5y?Xi;nu%Zs#D*L=UsBKv0Pm
z499sP9AX&EsO&H)>BBHN57qu0F9@R<X4r#!tL|$rUc89i-TmMM4`4lZaTIIyk@(IV
zk5s0nVR>|V6)$!UqIZ=2nZ5_Dzq!2^&sApx9p8EVe!RTaGQ;?T#~;Q2_NRXqfBBQ2
z#hc%HJO1FG{$aFKKl&g795$(KZ*HjleiEJiJqy2@3rnKS8%k?BdcAIJY<v_u+nZ+W
zEi5g?!$*(Q#;-)PwWu%}cdFy**-`Y>r_qFa`sA5~A*?Mi-+%k<c=L_d+{UfNC5`zT
zt%*5~UU7PP^df6y-|?og=nYzG6WWnu6**qtnu!Q5IaZGKkk@`dm!Xy#e&7Qip{_Lw
zQv`P{{nmHC9p8TACEut0;q&bnXpG{H;Hx3Lf%YP6ZPrGtw^-|t_VMXO?5b^Fyx372
zBhW@0rR{!5TAl8Jg<nUUrlS7JT8PF@O@UnhK}L}tLKx_rYTlwx(RWkGix#TZRBChj
zWpykj^b>+WHg9^Z>G3YVT1{clj^K_sgg@hk5$oGlE05z87>q~A?O1R|1B$W6y4Hdx
zl?|-|w3GV1zCx)lDm>$rK7}?ec#=9GeD?ZHYc%i)S_@+uV=p~ii`QOx==?3%7cK#B
z0|x}sXi_3Ti<>d8k0~Q`E1IF!s#Tl8RHod0!H}a)(fnl7==|)w+k^gqR<5<iR1V6H
z=F*jLFrXwIZ>q2zA>x>E)8S!1-ucN-<Hg4FcyNC`UVrTs;emQI=jY9A_wj`1C#wIB
zkAi*v^l1!)o4~7R^U-gYmzLwzS6;G4Bm!PEiYU*haQ;V6KJt9^-S7WUY1QLgZOdJS
z>#J+(TXUWxj@?&<$5n4M-w?ixj_2toV>g%4?`ph0`Oxcp1ab%o8uN4U=-~t5yIO3k
z?|rn<5qw5cGj)2@TAbWh$Q|&EAq28$W+CuKpx;&<5$dvugMP(W91MEaYWmB+{0sMw
z-~C6wYoR{#0(ceN48tEbS8h}hepDYsGiPsaFE*Y(i%$PgxN<61R@NMM6x?@)aOXns
z(ilGIyZmu;9{1Om;<c9_81HiIEJB?A5kk?6;Pg0f^n&~l+A{y1o@;(SJW(Aw@tk9W
zxAuiQIR%M3pJu!<fIHCOSu~pMxE8+dsZSu7PoejCaHL~>j;j(>A5j<x5;*~h%^dJ2
zdMe+}&c6FK^ne9p8iTa$8?U~sHltsmm8Jq-oknkOD^AoWiTB~t=kdWuPvf0;f8jFy
z{_p?e_|N}~|2*oAmgoxZ{z7w^fzO5<j?Cj|Nwk74)UVH?*FQG=I3|*QO8=Z?L(c4^
z>Nf994+vt9<^M#u=7T4jaePMGCjMhXGGpiPsIR%{CR*ylP0b}&7lU}HxqNzNUg0i<
z6Es#ewwm)ZhBHTMGX4<6!MmYs@ch8b2!=SGmSS@U>O!mL2Tj3ag{9nslS{WBI2}GE
zf*Lm1R9A3tha>_V@F#udpxf0PFFdJ!$2<=25&Gps@PQ8Ecui=v92?x7%BE@I0pThH
z;-i8KT589~qQ%rs2-qn9c9-Lc_k<sgttkjxYSUw})SQWCT{M~C&JmXDD+`7XM+P1U
zFM|XrLs#Yc#V_8Cci;O(JbS(w_wK(G-~FxMj;Wg3Zy>lHv$0?OLUYMLbp<|b#)1a~
z&x~VB(YoU}Luf&8gzu))7z3|^^O&2c3;2D&W`Z7C*~h}MC&t|fXO|a*A6DA2w%Aa9
z;I%`;h4~r|8uem>*EMOWU&G^LcFVPR8t}RKC*PP)7$AkIVDriE&oZL;zczRZOB8zL
z*D)4+{hNRFuRoSl3PcLS0xA<XNLg8NXGvpN+==w@YU6>6Q=q6g4#V7j$tHj~3w;8+
zsvssA@K9MaCgq;7E1}Pjo#G$9?8$GMbjUD-cH;ieVB|Xtk*3O@Ft*?|jN@PZIW_%3
z+T+GknG&J+Idmy+l^+H6r~UoOZ~nVr@^gR;RRX6e5zHVeg#}3@%y2JZ^kbf+h`jXE
zPZh?nr1S7?7yXnE3|6O6%EzzW>?kt~AcI|xD^ca2@F4@k$loxmgm@^s$^>*OCJX2Y
z!HDQOBW;d19Myz2E<`yYg7ij>e<7rGrU|wZ9C$aA5T$p&v6$4v0&|~5r+4qMD_~_g
z7PK(&y;E9fAsFMxl1(j&IJU6b1u`iDmECpqTHQn~eIM8%xL|=|q7dY%@-l&1t4u_@
z2cQ0dKLj3HKfnMY0xu>}&bS7?lno*QW&%Q%>PQlHse;hI$gooc;Ie=v&o5_aj1pLJ
zr|NU@+IPPbZ)gF7P?7VXy%18l*cIv>!6N!N&heB<OA`)+0)j6Vsiz{UAeyKf%TTn9
zvhQedFeiek)s7_*bE|sL0J~!0%z^}e1XyZEh_)0G(dxiAcXcsY!jy6SNL%HF__oHE
zVGA=y3pKSff=>u13X(f$Xdhq%182lSP1=z#`qbp8(&~Z0)u&QdU!-kCC{JiYr|aU6
z8_=R}Qby9ift|AOU0A3<?6E_Iv!O|xc&?zrAplEFpk!9j&&h*hLA>J!1_{SR$`gWy
z^U|pcWdfhFQ`-?;K9~_OhEhY0Fc2n;88jE&9~7LOG?OB@V8}oyji`O5MyEy4HDi9J
zCSoXeJu<w#!_O_J+yt{6vjs-7)2_FM5Bw`b0=#jP#a_h8NV2830G2>$zwPogK1u&7
zj5AUwhb3>PkwGiwe$F{El*cLZr847>iwLV>8urUX$O4?0X%M3LV3D!i+;YlPIg~fw
zc=4le=k&2aCBX2vnBj+FIvxE~b;V7;nF6IvQ>BQpzvGpVM5q1rP2JosY0JQsQ6+(&
zqxu^s+3dhiy5CKkbndo8#syxtki=v?GF-M%=0A*zyKTh*KmV$YWVub5{O~X3r(x9Z
za!it5hO2(F6fz|!_F;L`M=w0}@A}&+qnQ4NxGibuH|uVje$qYr2jLyaX*dtktm>M<
z<wPJXl$F?-hsTS*Zfj~Sb0?PHB`KFH;VXj!M&HD9NLg;{n5p?q+>#!j2^%bWJATF?
zN$`QotZ$-_O2P*lt|}YlDE`B`D3a6RGktBv#!tct|Bw4QIED|8#nPU#m86{|9~OL-
zevt+Xad^qjzn&zR$s^OtJky2$%48`rc`8j^i1Tm15qy1l1HkXppkg3a<{QD+7n^Df
zeU`Z$(_hwnTKF;sN>*&9(bVpX*^f`J;($x)&o4zYYc5B4pofmdcwm7K-G<=n{4{qE
zvs0K2MF^!}l5!NEcm9hmAuo2E&(5UQ2KWQIm0h`wdM&s}9>xMk()h@_%M^&5oE$|S
zi+5-SH1x7H`b~SWF5zC%+{p}W#xYnh#n|n=w2*gWOhEfxor%T{Gc^!eSfFnPFU+1R
zA1QaJdZ2v)uJDo16UpfHolv6VE9N)M?j{V`xjB~uh7oa~#V$qD`FQdEQE;>#3}|Te
zdZTV;>!j8izzv2iJNu!DIhq?<kor(IPBK7?iHq1dZVVVBY~bD{;Lz=MOw+L=|Ngxd
zmGLGz-JTgoFt!o!aTFI9%ERaaW@x@`z1WD49{(bCH(yvN@aA{kjz()S4*I9@_{lTx
zGGAW1XDtDQeAEf1DzxQLyG8wj9>C^3@H-l1+>e1)0s=rZk$e<6f-m4AOzNWqeIf#z
z(9qZ$3GUKhHr?Objdy<fbKgP3(Or)oJ&FabA(q><s7-UE8N!80!>V(@adubMKqPO3
z1e6D@fM+m#_jlv_-}{c@*9>!nRxnV{VZt_=u^?C?<m?<A$3W{k7OC7#h8Ekw!J)Mn
z5D3tYiyUL$tQ&T{em~GC5<q;k8O&h(sNeIg9c#O=e!Dsz#Dj&I_?;hqM{u2Uf7w$0
zXBX;gqX+_2Czy|-%ez_|zWc$m*xBp(h{W%{@lbWS@S1IV{~&(;-s5=sY{R&dW780d
zF%zw>u0^Z05FFw67k~a|@!q@dC?Adp<rv6@*Enczu;#4K&&Lmb@V)rfx4z|iZNAub
zdwu8ax1*-=!;t3aLypSY+d~5q?Z_Jo9A16-f%<sELPn0PV@#<i7VhlrcY?cxEIgD!
zkOz~Lb<2s?5v-kA=g`lOkMph|)~UJnL4a6o9MsrA7&r@)RekIF(#*#1y!o;PLj4ns
zfAurU<fB6cPtri^=KS<n?OC%%5CU^>(vkAER?1FS@I4el<E|wb#8b}S8dIHQ_?{>}
zYcIyD=OaCA8pP!KDmS9C_Td|xW33B?XKhBE!H1Nysj<XSgBS$=Y@THOjV1y(Vs=*T
zUYk=to>Co7eNPc%iep8AF-Nspvqf=e1Mm&PIq2AWecnQFG?=*04M8&Zdg0EIp1_SV
zdF?E`39O0F5rYDWwIfEfMn+rZ(fw6x%^_&YV^t4KJKbBq7cal^QV7kc{|T?Ep#cQ~
zG<X_gBi#8_b6-S=$kBWB0XBIeuxR_fGz4@L#*_46_!OLq(`=~jQxmEWZn=>ZVZ!;@
zp>War@mD|o^Z4+C_bsIR)|=ms?|%3DIcI1LKDf6QFF$<XIpTx&-!UJi#=U4WbAlho
z>k21@*6E6i;LYE{Gr;IbYg#mYn1^T&G?#BiRCm>nO-&pDIj?aDBcEeW(GIKdH5(LI
z$8+=?_-tbIDhAy>GrGBAfqS(P^e?L2*Vot79wYJi>5F*wq8pbZlQE-lxW2X=_tzJE
z)bEwsQhiGOk+KaA4^^&d!xVV1N!fQ!f!p`C1s^uohQ{kg{KZdxDtzB}IaV|$eD6E2
z#zW!Bx%rl0eQfPdPO6#K_<tZ+Ew8LvJNw=DKZvIvJ&w2Dd?VH#y&li^2JvKLPh}rb
zItxBpk$HrE#@%NxJy=)YnDaeWJ6!}n2v5({_BXNM`>y8I)@&dcQ$ITwzUZm{9Q%0D
z9s+UA7u13K&FXX0vC7=Ex*W5$dC#v0sxtyb{L%PjQ!YYPSPYCc+6f$!f-lYYf)fG;
z#%<2|=dueo%kKGE<FbZ;qOEy#R&kXkye2eSIokI9cYhH-f9Izbn*Q;B_Q&equd1$B
z9wThXK;!p9d3hu9Bs9u|=Xx}QNdufPt+9*dA$%41dTfH{6dknT4pcslxr0XP_Ksq2
zci-cX`I0=(FOFl3@vHjZUt1Lpfv2Q-SpAPXD>?2G{D|yseB>%x!Vj;%_OjY=-Sw(*
z^rK+%?CH~Zrg;D%I{YdG)Jw~1KaDACYpN_UqO#whft4F>;Qv|6S_oiHxE#JGGz5Jb
z`Uv`9Ao`=j5t{P9Cmf1kh&Jaa>eF+@6y5+=d(*n$js`y(PIFVEv7j+Mt#ZIKy$LqB
z+;~2-@K9ms4{Xvg`%UAH`|#?sdH?;P;00f*p?Pq9sSz(fSd5c?FP=YnY{BRKhmSlr
zBe?uu|NXy@&CPAk*|p|ky!yK4)%J?{eBc`ROx$nR5&d$EU{&qJeOzdynXjihfS)<G
zm40_Dya$Z{ZLqeytg$y1n>%~)!IKx(P~_N3Hm)wHufOqdRqeJQn5)gycMml00iz}1
z`Fg#f{xRzN?6x%j3#`VCo;4Wf#n82L-T96AxT#=N!RC|SpJmi9zczTP_upyDU)7}G
z>tFvdskwUoRVN!VgFG>1)vNrlS3f0~Ka;oUkp8zv6~Nz)^r<ERSK?x-crv#@E&fg`
zt#|V<TS>^o?~=~lp#CMSjMLP^7k27qpHfzbmt)o^{nA$a^p!(OuJ)Au8hdfW-*y@0
z=_2ZT2t&O~Q8Gd7#qai`ysL0BF1G$kZPRWye^fvIC2W;nzF3fEUF|+BD-Kwd8Xm55
z`sFcWxj4!M`Op4QG50dFDh7t%@>L3e;LFKsQDb-uE15BdpC(#XzhhdwOloq{?5PF)
zm<S|M8A669TFCekTuo32XRP_7iIQD*EM{+D0*Ihv0Zw{TTG+3wt%v|`c>+Jwgxx<7
z(Y?DDI~(le99TdC0n0*@GrfI8h{}a9WOAHE_=L)uPzE7G)aK4s&YER0ap?*5=2DYT
z1!To%Vum?@9}5mZ%!CIK#exhX6BT3@%<OjkGA6Z%1WMKaOAmTrjUl*zu=sE-UjOdf
zasQ=<-uZQQokA#xM(#-C*hmPr3&HbTlOwx*QacVm)#XeiDwCEf#O|58CPRoGh+cMY
z&1%wJ)#7~h{zKpS1rd6BDI!;s9)uGDDle$i{xD#?OHXC9@JsCvJT25x+2{@|Y=B`!
zn4^7$AdLl<eq24$WCt648o>{X9(Jvwxdx*IVhtt=M3F~}!Z{zM2NCaGeR78IvL=-Z
zgpd=a3Boyc@8!aBM8qaXmH;z^UoZ{WVWRM=1HHn6GV+jL9-XLetoSfvAWFU96rn=e
zoW;%}gE~Oisau&iVw|zipninsXbm2YmqgfQp|KVOEF!#9Tj{b(6JakJIAa=%Q{y5q
zYK@qmuKB1#U;_@PAG-f~{87`pU4$>a{YoExi>$~_zzI`*XD|Zj-w`QaIuV!`307%i
zGPn}pKLg-83`0qWp&d9gOE<o&x5AM|RTlqbT%1afQWgimQU$XiRriVf$Z&ou9Y;|J
zy$Qi*b)QsaGJg6EjXqLfK)h0(Dv|WQEpvG%JoIb33`D6uTsq)aA|$@Gm*Wz`5#(Qz
z`dIkez%uTRSzgC2&i*Lwj$ht%$P!wndXuJJE~wmw@5&2LjBL0Udllezj;%^kg|AE%
z79)H9>ql-?Jw9gchEK3n1`v>!%cTD*Pyaf(#M2p=@)8CQ(gU|-S?y#y0Mr|An-aHr
zTkvKeJI>I17`~Xw9V0&nS4M6#`{@U3Wv^ZptcQj&^KH+3mo%InE+u&O4jH=>FQL!l
zu7;n?os{$@Wd34cJ9+oA0@%mS8QSSo5h=Ls39Ag3X_1@jMG$>rwSQ&D3kzTV_eXx)
zHw|f-jw%iQjmFn6VF2)#s=+trvrNu)pAE!ceo^+x@OSz;<4B|LQg0Za+ns|L9E%2#
z8=6mNqDNprc}#0iL2E%LKzm(46JH4{87+yX5>bU_4EMjl)P+d_6BHT=GbNg&-QbSf
zX*0lJ0Gl~^bmDurI-P!Dm_SZK538ISgqNaQ&=#@~N;DcYBQzFx2O0?`5C$Ad`8dj{
zt^8M)xrDtPT-sinL-3}s@;|HSd!;cmJ0EDX44{EjM>bAD3v;gyWlsiq*|a2o=u&7K
z{3t_eG30TGW_%xVJRr1P(s|H;?4V}OhH;!49VzUhUciPHAPwMwuxovNH5$#jkMu(*
z1YFSe==FNalcNb6@$mjyEG@On)S`So>QHc>7`uu6osGCSIW*Jw`NoU*=<!F<+21zf
z>jyvhLHtMm`1eIWH{+f6K8&sH{TLb3S_EOA9<&EwEWw<mZj`YS@FApQa~&J{FdXy6
z`U99j=fVJ|Ehr=5p~GS7`&cK@?l47R#B=l|U6}Ul?Ds5at2dh7r4G|~QDGbNv%$SE
z8_%D_&dz4c)|%=RYG;*iRBIy`^$7hCym15^Ylv^Z`C6=?Wuc|ak=6|yrL@-_n2tpd
zI<Gu9+G@Ak^SXjQ(rVP*mb;y9?5Xc_Bpyd0wN-}2g$2O{<~jGlY0aYb3QYNSb3WRw
zrq+xX>bIM5rZ$DiU7s3>_0@%#o1arZN!Y<qgDD7oJU=%Zjry$SxvMxlCC@>e9u3S;
z{OH5?VsU9P9=-HxbO*=28))Oj3u|$F^wEd0zqb|hwfT7Y)i>fF{DXfO%j(xhhkf-q
zw17s{cC+e5F#OOqnzqpC*^?*Md}E_vyS?NxbHv`2X!gTXG+eZPf-$Xi;*Ma2U~5YB
z{_;XoZ8R>wEI&t+E-bb*9=Y==@9ATlG9G*fGr|p}#qp0YBk7MEamcy~EhyGjtPvQS
zz?1O<{BvU_ciLT96J&X@8Q=T%>psGb<HTT6BlttG#_s(6{jS#7TGO6s{2uoN%bED$
z55E_W9^ALa-ict^>!0{89**gv|02j3Q~zfDg}|<>c1D|#H9mcm^}cB@p&D=Y6CR{r
zRk#E}o!3=nzAL;cZWP71Rh?iyFD);}%F2q@aWINc&)M+GdTLBKaMJ5DaK%A)FUGD<
z<AKIE`~cQrj9D1$&}Um)+wrq^-top+;(Yt<x78mWsf;74`+@3qC|IUO702wg8}k+d
zp!valLDs5NUqoXQEl>+sW~Z$&m&ct=sQ(I6t1R#fSRd0*TkUpCOwR>Ja^^8;Xziil
z%01SM;k>g5Z5uQz5Zt4w!8&zyWzl1ZHL=_JLhv!+q;d*>pBx^<#>R7(3jt<le<wct
z;1|Br>Tn=@Cby-fmH55i`zP^_{-fWIxq97e^LdqfTsZUO2yNR#3yN0P@5iL@ocojV
z7-1X@ICAqu?a#4}a~5PVAI;CtTObKeJ4WE4CGY5j+ChB{;b~`gC$=`9$JWk{a7*0+
z9*&`V^kCg{4)+{^o6k=N*0KawJ$&@Cg>Qz5*28EpL6xJ8x4nB1TWX{3A@hiEoNy3Y
zVaz@79++pjLyYyhwTRU};U%!a84V-G&T;=BR+rl9A4|qFA3fWQzxvs`7F66@X~)CW
zT5Jj*K7R5%KK$@~Yi#}DAN)bA-+SO=YFF0oslMDfb{2a(92<EZjivP%oQ+sev)4UT
zxkr5zEShG$?!L-$tTNVOMfj1KXlqY+u-nf%s;$A_Y@|R#3*i^@TxY)%&o(x#DLXqa
zyfCRY(U<{0aHQr^b2jS2-N`=?&KABw@I|{Zu1pT7zpH>sU-Kt)yV`}ty~h`0P<@{9
z&z*Ffl0~1Qzh#^;Ydj+y8rOVMo0+n}385}p)@+Eoyx>UUbGH%w8TgND-WU^n#wMmy
z7|J6Ymi&;M>(6E9NW2S`@k(nhjuJ%hfshg{Zv^On{vUoE@4o*eCPh267nXgp4&#OS
znaw&U$0~>NMDV=2wh}KrdZfJ7?&miaE}=2W9AGUh;aiSz><GVe*V)@|y>8(W_>eoh
zI(xhEMDsVAldrz=C|-W~q46zv4cJ;hsPaNvP$P~sJWx5n!!6BU<H950!>;fmb2LIi
za4~onm~kAZj|~;B?;W1T{sDJ3o_P+ORwteo&fwmx8R53M8THq3!B()^(VV-xCmek}
z7D5}EOH^N#hmA~>-D`1`jm=KDQNG8IpPC+l&-T{q594>f|9aFk_pu>>jVyyh&9BM}
z&DE=mylMoELhhF99UQA{qSFqK6-IqhV~akuv#;^50>E=xR$bWq!cmQE_MpEZ@B^(2
z%<2o8gM}j+q7Tv3LKF1851yC?V)Mv}<`s@LU2D(Al;|k#&(q+IH^22Q&-qin18hQL
z{5+mM-?g?KeU^E~Ji|MN-VF@CF`qOHeoO_MPkw)v$@%u{G0gu}xqfwH!PkHB$Ny)i
zMRE+hs>1xbI&%9s0|iqtLa?a=3wb$6i9~uC19u$LBa<hFfF&nf>HSr{w+x+I1EMFe
z!x!&--G;>-22D0VmKGlui+ELmaKz0o#VE|$92%fo90;G^5ItsI`itQDrTn;;bkeuj
z?Uo^jZg*@$=<$+M71uG!hhHU~ghv`z4G#@w%EV_z0e-SO0PYUxclzs>f1(mbc9V$d
zSrV?`n{m^#N=J6bqt2xcvcn$DTIO#Czm1nbl)Z|BwRmYhP!e*L!;L1(Q6X^zUub=4
zR?6M1sub<Q;tjzUggr#gg%COC8$%?a5pykqj>*Twuku8N))&-jGC{R@Z*wy`99_~=
z_+!lk1rKr_?#jxtiCYK>jux0x91AN%d|V58Gm#)*@kPEuyeychBSbk9xfu+yTlk=Q
zFd;x#`%VFsk#e$g55X4<R|ucpLD!RniG|X)KF>g;I8=qe3k=ioiw)=WzV!B+@#<S|
zMMDJSq2PuP24cdy^aRsOm5bdo2*EDcMXL$WI|5l$u##2Q6C#*6;t+-gmw&UPa!-pg
z+Htxu8;e?Cp!HSlX3^A^ggHVU!HvbrjN(ElnTXe7fb;7imgtujEU674z<jTk-Vj(P
zB3zh|S+rAUU<?Q#SjH!EVxy0;+l)TX?q-g|wBV5l^dq$+Z9$z8x^jOUyAJ7(0~QJ*
z;vgjX=3X;0aC!_ZjM|AsIGS;2OtIj^AYg-0LK$F6usa%N1VTOpUi2H%p$~DaB>e!Q
z7u9ceApsv2^aw+#vxRj^lO59tMvp}#nb;AGxQmX(5Oqc<On;*b&@Z7_$F8*4o1BZ5
z2;TXbISaz*bE6{)>w2(|)LUs&KY*EtpAU~voBZN^*De1tDDHRSIdH){9^e@dcR3e}
z-ZJ24hqtM4m|Lk}k&+&%8E>RXw<<iwrMcsxH<^_PymF*gM&Xr+i#&`&lv1aH6f#G`
zVz0v8mZ^%zx2tM<x?yfZq#^B`KT2SmDpX0NL~!24{;6h|I3aUUVtC?&arA^+7C7^y
zd{vBLSex50L*uZO!Ct{ae|q1B=Hs>nvZ6#Ma7h7=;_mR3-PZhe3ij8J;${CJJ}Jvp
zo_-U=Cf#wd57F#;Tb7}}(s5E%zi@+TF;&V}iZHYfEiNbNv49uubu{`%$y;t!gsiWv
za!V`0@g|0|#8hGMbso2$SxCy7fvYk(PZE$fPFin@Mo_8lxP8(%*{pQ+ReojUZZDY%
zc3H>_lj-HJYYHwZE_dNHET4Vvyj-r*&m2kd{Vr*gaCS%+rKesjtIOkG+<?5ktMH}1
zY1dN{BFz8zKYt_m`m*M;%JoSz6bk>nG>nliU;kew?57%ah$;<8%hYP%V??wpyM;%h
zdvxl1E(T`YLHD59kXm5YLRX{a@mfhUK`(IZ5;O@90uLA|Fgeh;g4X4Z3K$U3DU7GN
zX_)G@SX*8QE)-|iFvsG&*xHGm?R|xZ#tAcRp`oDHeAJfcWN15RIy2LcL|ckpqg)vB
z<_I$2O8KCNVBEu0<DznyaR^?x>kHuzyT`f9fjUl!R)pz-K#k+ep|Lqgk<08UKfM`7
zINB^bJ9}pOT9Bu*kuS`0%HiGVR~(@xT1aW28NxUqtT)q%?&aPsHhaP7f%YYB1Wp`>
zhOn^RZi?nzumI^`(2s4=y=bT)P@^p1t%hjeM-NtGZKdsvO)zG>DdL)gDX-&1YmwgG
zPTX5vj%QCE$G`hu{#U2}tv6qfmtJ`#=39#~S8uue&t7bMw=+VF`P!`S6oYP~U7;6`
z4$&TgZWMiZax9t{1|{<a^day?2*uqR9ECln_JVN<b5n3oS*GNVCI&)I7*A+Vz;NY=
zKKdPZ%5eM_JH%;tm}@YlVA8jnwP@*m(Cfti^l$$^@mD|oi+JtT*W&dz-;7!VM!ovj
z6!bAih;`!NpdZcoTD<Ynx|#bhye?H#n6qeX0e_fV2=>_B4&#}2tJi9=u-FzjluoxV
zxNH4(d12U2tK2hsbIdpM4or3COOB4R@K)vZraqX*b5oi(FJqu|54i_d?aCc9%pVBg
z5KJ%+oecKk!M){p?aepL2u0uvBlNJ(9V<@-gRxlG8fRj1#yi<}_6}5s!#F?cScrfy
z<Kas$#~W|F5shZk>jfCK2(3A?iMyL%ilQmUMf{^$lN>1R{r&xz(^`N{n6-L6rsq-{
zqjPX*rXqDaIXu$5dL1pn9=h*9_y7hf!Y}&d>>S6=v4+c|8)4kH7h2YEVNJ5v>AKH$
zl_u~`#&xc38EY8q7F25pabL>*?!E;Zz_lj)@yern@#dSaDWq^xuV;Y<T2-8yf`Fsb
z8@NASo*x>PGm|%pb3a-wwc+TL+WIt}J$vqcI9t=Y2n`{QWS^W>S(GPQ4UBiKm(k*4
z9i?XRcmQ{RXVCmXD+(q*+C-EW<~cYIK`U!4#sbm6kKDDzTJ_$&^=P)V&evKP!+MV~
z>ytDDhwDqlJ?_TQ!Cq`_JoBd7mE~o}MWDktMQdkLWj;Mqc;QSm7SP_i(mHK(XD9Z$
zJwO-SWrJV^d<)~6J9nrbc!Hz;z(*7iK7;YZO&(pyQJ!cofR8yAjpM?g^=4-0l<tV@
zh>!{%NN%FldRc8l-&t%gSeOKEPt9Dkq9$Te<L&8_U&OmV|FPl9abs*Qc=qgR?C$Lf
zZ*D1j!D1xue!CX_a|e)lxV{$u^bh_p{=skmPTZ(({NlsM(G~teGYNPuFRus}EU4cI
zZ!1593p?9e?uYYit`Mv=`v~UKS_4nHUi5wHNF4AZ_g!TjXJbeG?Y;LtR5&!hYH=aF
z!x4uyg<oyY$II(W8iV8N!+mSFao0DduDlw;TW43f=>f($$JMc(XFdgo9RN4M1;G~h
z=<tZUk7flo!A!VBC{Xw>g#k0O2xQqrGOjr1>eqeqa=aOYIaaufBSygw@D5fNoAKt$
ztI?R7h>d5D<Ny8t`XA!`_uh@~zWv?!r~m9fj<tJ_EP!HD0d1n{Iz4bTn_4&~ZlHbv
zPGKHjqMxcj9A63#00GCm;DPXpqgY?;bk(n>V^Q-FLWYSEjbqUcXvz8*VZmU(a}Xas
zeyXw`xbMIlL%9C-TW`cdv#z=Sgx;@lHEI0P)m%T&Tf@uv87(F@H*gP?1u~jT>069J
zU`$`;m_$EHoBj+>0+^$H2M>jLX|er(u=nTRktJDvCw8#|4mcc+eF?;p5gEBvWoFmX
zUCo{*C6U9?9Q`74INV1%8vPm4IpX|F91YnV?&4vS&F-r1%BtKWV?|(x18{f1u|wa_
z_rB*qL}pf2R#jGaAC>{npFP``Yi8H(wR??2E)!n=Sf6=NEs9Fr-pkWyY_7!y%vZtt
z(CB#7re<SIFoilsm^GK0E0FGCZG(x_JUodfYuoX7WixhEW~7uB<}^k%7wqgbJjXG{
zV79}UgZ6?1Ch8u6Kjt%Vm3@_iwTbp+-bUT*gy5QWmUd=+L1L}0d3AGp!|HEM>rpu@
z9j$g=nx7Y3PHM~?Xf5io+EVVaWB+q_auywRG3#h+f6o%$%k$IjTh?D_dI!oIX4dXL
zDiKrq5Il{3Bz+iVzyou2gZUn5NsSq#L+hhj`!shzW%Hb&{-<7N+Dz>$&(5j(JO-kn
zHhTVi!!#wz_v)QRO9`smj5k3qg|5aOp1|9&+-B5gOLH36Y6Iqo=W1^xsbB{7o*!G)
zvFEqi1jHxfV{OgHj5F2{?(hU>+t|v(Er(i5VVcd%sQ#W$huudoE}7Gi!i63Uj!{*6
zqrQ^$8rmduFxE3@z!z$V%7A8Vm6x_XQ>E)#XHi=^rnP2sdpCai$#b<SDqNwBu?qa<
zGRX(|^D;T_dobI){QYHyaAnvuB_HNjZ@%9C?8{8S`1)H)sZ316wk)$Iz!LF*^v*Bn
zLRp{ONEgHk{!}2_NKHRkc}G{gjAl3duN59Q9=kYlzAzxQL~%+k46kc(_4o@v{WIwh
zIZIR+l1%UXuwRBt7v;^S0h=cFlGsfV%QQ&QQ5~jmRan3Ln{VfP70Hh2p2lg)q5H;D
z?M#RaUWV3N0r<icE@^NtgJB3-TDqnPqsN;;W$1R+go|N=d_BHRnTZQUO|AtYv1{Sn
zeb*X@lG1K&=OU8dibu_ShZn&qgDj^)`p>nvBel|}$-P&TA*=nF!d)mG7$T&dr%Wz=
zVu}r?RrA5r&_u`rWc7QR)M19<hj{srkQa0ZbFyQ5V>Ncx)}x~dSo>)MqNQgvsiSUg
zc6KJ}n!rK4KrYw}a!LUBWHF`eAow8I)DOR5xq}3NR3SNbs7(UM8;Fh7(g-o&6e1@n
zkp5zG%qE!z0AU3|M7Oz5v-}K0)4C~XdkhWuHBFBe(3%i~H@^FBEZ+tJ8!*xX0>J59
zPM36qymeGYPEA?`q!zIY+E;Zu(*}&q1_p+UCuYGtn<eUrq!JxHI5wR3d<hxgbZ8I$
zoZVqFr3KAQCH@p0*rR=YdRPALKNxXpXWY5`nRccxQ$+)p6KjE2csBLq%LWf342T}B
z!-C5REjG+BzJLpFmgL6<fp!PMCx6s5f>c3xVZ1j5S7uz1Ckw1#lT#|3tYOZhEP8p4
z;L>(8>(tleN%}d46ILk`Qwimyjgr8hVNg*)8k`DZkw;|*h!kZ5k%U<aOnI?a9w24y
zeN+k^DnB9Oa)Sr}oS_P|x1aIG>1-}R)pl5A8xewCAA|8FL|K2-K&@=TN6z9!dG$qG
zdvv-I{?~W<=h!XWfjc#v!BC$0jfcL8?{^WH7shUSrKjU&@a(V5vk25R)F;z${x<qx
zYAzGR@9HZpIM|e)I5KiGuQJ9}P}%&EU$)C9W3a+qxx1YB$e#7rqr$<YC}X5&;m3o)
znU{V@qgx*bS1cuV&8bL&^c)KpzpwvZx$!PtvK-}CerFxB_8B_+p(rzDXBjBVEo9kc
zEW%|BzVz$wG-Y36AQTZKUAiHhw!iI70Q}O;ZRN1?(aShO=~4#854UbPN<&!TACBY5
z4oCP5neT2N`9-)g4S5uKr2CaKcE5^Fa@o@6Um^j5`IWh0D1poAvXDp-{&Ob7I{~L%
zgf9%=SLGzV;w@*_k~rO@@yvuWl`@{b2*)1@IJ{%a;F@J6akn+TMWR>zP<YrB>6E6l
zX27l~KeHT#2d^@d(#_x3#+F{TDG%HV*CKR=E&aMq^31Qy4TpdE@BVEW^&2<8VSN49
zU|xdzuP|RH?8^q&c62&L7LR7(2z;{H7jkTsJv%MoKxe4hMhy|njJ$(|6aL@|9P~nx
zB=OcH%y4Si8U>VCscwpcHsBwvc1zEJasjDPe+#!78n_f5JR0@!n&O>BW54C+@DQy8
zFinKJLO29B1a}t>2L1~E25txboRa<Q5#=vG)LTR3@<B4y#{?G2rZRAl2VyA8&~9TV
zxZ4B?i1F%JR7OWERW&nFcU}-`X(mg3DSqpqW9EqOe(9W;fd|2cZ4USQkZwuLr#&EI
zQZ8^Ha5m;5)TKi_hv)}B44w(Yj??Jinj^y%OItv+pHNvKnu8B+L(JLP5l(un@x*;Y
z;FH{8HWD+aL7ba3v5?dI5X!h*f4{LE&!0Ya{oZ`zb>ZBH@yW*@1**tixqCOptaeKM
z!XY}f2!vVgsOnL<r)Oqk5@MP1fl1VC!VH1oMp=dHo^ilSc(eKkO$Qhx^-xcQgqrV;
z5{^u}0~a7uNDe`iud2_`gz@CjXJ%~733sng&6ofSQ4zIl5T=^DyS{sk{-h45_M)yM
zgZ<Iu?y8l?Ccc65qC&8LsG@Su$HGG1xpkyIY6-`_JXczK4Z|+#<ze#l_6=%Zr#cD7
zPr1CVe=tyY1;Yu+7w#0Ao10d+eiQU5|LMt5tUP@Z$A<^;)|+p{!s0@3!koL8A%t?^
z3|wChR)NS}CVg=lt7}{FXl*CfHg+uG@!IRJ32sMXWo1qM^IYM2<Lz(Vb$Xqn(|EeN
z8EcyjOA10{@3b57^x*@w%V8|vyd4Y6w+&BYBLlHy4I`88bH*iiO|kFh<T~dDk>FzN
zpmtINM|({qHAYp$Q}+R|$ev`ZT2l>sVqg6WqX<bK#vZWC-m6*_e5<V?P6Gq%C6P!u
z797CvgCNO12BIeO2~t@Q2T2p@7l@*i3F)LWwU1BU>wy8z+(2CK@if+e8Sf7d+iLSv
zNjk5+<hQ@|EsckX*xK3F-l?UrHmG`snRKY1>hHaX!T!rQYH!D9AAb-#TkB>Tp^A!e
z#`t32!~J!OH<n{=;fD9?NcRj5R9xR=WU0Lr7AZB_nLYTB#x+u5NKTN38Fex+gWBpB
z7-rrBsz9kmEawC?l%S@RXH0OPhb@s>YR2~FdaONv8qc0SiuKiJan#&XUkqrBjK|c}
zv|xEAUVrU={ML8A7em@t!H{I$@ZErd&mNWWlLrst;gcuAtNScr25kXG=8feW9vcwh
znLm*@WuE510F1o@wI5PP;Mwf!ya!eJU4QyX@u(y7DDx(a2AHYr`<N4<k0B`wqno*s
z!>TY4S(oY><MUJ1=<hj-kAC{2*jitWJ<UZejpq+O_#ifRwv@(&_S@PgY9c;Xe={Fb
zUYKe3@7{~Q`ul%P<6$WtJ$)Kao;`K>=4WT)_MO{tYx$N5(%fOSqVe<O*^1_cnfTV*
zZ^zW!g72X`i++_0Z7i%4!r-+oooJr*ed?;OHnGMxIM86`Y4bpHGcd39L-j@_A9py_
z$A)91zgKI;R{Y5ie<WDjkH7kRe<g0-TDEGu9nFa=8x6PD?3Cv1n)bg+7j>x|faYE^
z_RzFBl8Kw!4NLgI2%MXpQ5y~lh7Qyh$&h5Pp^4K6Kww-V0m-_UYYXF#dYpL7Xc~;c
z-gD0%leOVEL89}~W5eIL*04*ZJ)&`ad8+wuXG3G|Bx+h$V4Tu#J^flYwQoP~(Hy7w
z0C>7M)w-s&gFB=)Ha5)gTfQ;x^&08O@%l{kv3{z+v=xk6=vt);TGl-7mSRkx_U+Gp
z{9`jEzxC$p@y^@d_WTROocb`no^Nc&lht*ttH%s|)k*L#d|*;A#Ck?qnOgyr5*IMd
zLTwMtqomg{XE3+Wm(Z8M2F6DRyk<bB;K28GW6u&`6~P%2Vceqy%pt|ayr=G0nwmQ`
zr=tRv0|&s_sm3x)l$CX@Ihw=Utic+?gMzDm7^#A{q2WBx%skATP^y<5VI9z%Uyc)0
z8gk&El!Qdub#%NMqryL^Cz4Ua8ka~okIAjA@%Z@JYCKz8GxJ&1GShLE12Kv>H&u)2
z`k2<qw$k3wI;e4Wq`9BHHTZ;p1P0BWn@g5(t|`ud%EZB@&Pk8gTn?i{_u&q#io#bb
zT1RLXwK3{%_gcc)WbXx!I&5iv-Pbs3`;M~PcW;}PD)3P=48^|M1_mkXDs%I=*14&P
z(HQK@xfe;S%^jpe58Q^VVH`9;a+yOh-~$}~-{oLNqb+=7R|wc~%uUu}LiGmbtTs_N
z(DV%R+@j_Tu9ZP`Z97_8zd8I4)BEhA&#(f6i}4A}AOTF>Ib_3{$bFsk3pnV4(i;`L
z$HhsYPJ4S}Gd9-NjcYB<gOh2E?60W4T0=Fym{)qvJMrw<dOUctt$9q}#r*h}&HRS(
z^;PEU?a#i<B;)IE{Vh{=SVRFX7g}kv7@y19?F?SswzzU*Lw#LZDSZl$vhwAp4Hh0v
zmVRl|rvcy!k*?D&NgfbI(xxD`mBG51>5Ub`<X<}D>Nx!5Cm)Vnrb%Kr=PzEk%U7Rv
zfnlSUCEDbUJ%6!eI(F+82>&94eoCT}W65sD+%=Qs!$MiJROIMzP9y`ARmcz|oN+-^
z>`MOWETcRLMtTYaH=V#ta7+jU0@-jJ!WI6Re>V{sJhNpe5BYUniZFI7>@bGPauCvK
zxgg3x?btQSn+foe(-B%=dNjfIXtSmjP}>t2T?(6kn9Kkr!$ix59hDxOEM=pBL<@|9
zGZtGRAzbE6z96J5WJd>_T3e5U-Tmm%;>PBQeq^%_0tdp!DXA$TA>-9?BLQsUVJLt&
zATn>pjLLyD5_Mw(jbsj+d|x~%1dbDEs88XZ02L?BR2EJ=P)3NNobY0?MA~OhNOfap
z*WcaKy<cLotzWc#T{EJ^?iC?BcV2rfDphT;Phli!Go!X-LwCx?hI_rd@lS~_q~|WQ
z5obJ>CrR1F8Ge)|utU*}9I74r)Lz5dkWJ2O^M_gqAuCAk!Fb9yl5z<fX26uyDf*aC
zhzKygU;?qRb-o0XyHEb<Z5`C=4<zw`0gz8lCb5BY`Sh5nrWAco82Ji$fw02{4<@Ao
z<i=F+#vK&Y1$B*VDoi*MmQcbiZonSy%q!HmoTCf}N*`Ac2`a=LRIa1of-+D(RQ5o4
zpe!)yh|b*>oO)(+MeV7d)pZC_rd=6pqziEd{nhjn8wZebior$*q=XZ_eQXW}VSK6V
zD#nPA=!x+OAr4ii)2EFGr-Ue1&IKx|(qPQGrBtAD-l2ZkF~z&EUb$eiCB5-eeEhEB
zW~eenf9YRY=#_3=L*9CoAqhks%JwIHjOTB{l9XYv>tQ?Yj_+Tf^2%;cy&TT2*o&Gd
z-7J^Bt~_mb-ldx@MOo#6pS`oN8JIYZr(E?WT};L#7@lknio6R$Sf`;M!a&?}l6sii
zi(K`KKfPBZTp0Y*g>sQc8G`Vh+41q8U0wbxqsOoUW!mLOj%`QAM|LxI?3H0M8v+qb
zr7Fv85B%lp_;$sTaSH#^!A!F5r~I-oPTS>72-Tl#3Qy0HghNmc-bH$(O%hk>r5pKm
z-RUo<sLdO1`O(It%~-*ZcZSbMhP|R3*>0rAcadZnR_^|jBTi|^FH^?NY3q;g?1%K=
zeN~qtHtu@tLcVq)Z<ST<7bVUlcf(0`kZc*gi0Say685qlHf315q?^m3B0IM1?$`1I
z*Zj}Wg&T2kGi(%TIzVZcFIM9P*zH8!sfS?4b(0?+8Kfw`4CgIm@4}xy^@~4+mH8L{
z_Ft#v*I<6b`1-HG<e2$+&DRP0vSFSr=T7j8A#L_iLB)w*h%E!!*rT##r%?9?0ghAW
z<JDmkj8LHm;qX`qagPTQCETHa8Xi>GaSsT_C$xoIz$hvS5DZFA75n7#_KqdNre|m4
z@BjwEc5LnJ__R0@Mu|IU{!kq75b!6MKrp``Ffkumb(nAx-<MP!H4YB37U>5FN#M)p
z=O@9PQ!s-d5Sth&oURE@tTND)5Us$?Mn{KZmb<_vs;UR-*N)>zcoHY|cXm@<P~J1b
zdZjrQf}G|qn0?@T;H=ah+!KPb@pa)`loiQ;)B;+PMtCnLsxj<egu2D>$e<<6CO8nM
zHilrzo`iHEq=6Sgkb>Y0eFmahWpv1GLfOfWwFbiIvnNmE$3Oa^4`ux~|IP24LHNN(
zAF4i`c=eUrQJbhM&yhIffSJnAzJQD5!8u|6p%#o;6IE)B-KOuAf#3wboB495&Or35
z{xB(E7IA7E)k5|9MB>pN^!iahLhOeaJOpD-I5hW*aWB>V*ItkD$!RkYS2uQIuLT1T
z$sg_zInZ1U{>&XEO*0N&ef4fE&P}T=kwWQMg&=pVpdn*zeKS^{ug275Eox&U-fPj{
zsHoy_;>PMTGa9(xV|H%F%(;Q#G3^zqaUy*6NN^0Ca(5c^)Vk6^#U5%X+xxp_wp}W{
zg{2#D_x7^#9nL;g*@1Nq+R^_<hiKV|80tNbscOQ&YU3b2`s`UeezqQy>f>*{d0*ui
zjHgeYS%uT`jk%bePRSXFut;@rH_Gl_Q)yk8k@ocIr}5DT@0oG<?QefO7H`}V%#V97
znUa^O;yf<dvjShNbF7m{48Wj+VcKH?wHX12Efb5+RVRq;+y#VcDbz6L-8U({z~N8I
zz#M>@Hl!(#P=UCOL>wyYVQ|ycZH*6z_3CU(D^6H#B=_Pa0hzf3>8xF~B~mM_!R#x5
zk33wOducEv_GJ){VGICD%Qu$Hz@5<8B)-)bs-Am|rrs^*Pq`So#|QiI=)nik+TRlR
zUBuM%biDP}JF&R9?78Q?pZ-~_tUU3!xc}N)@y>UC5K}XY?nfAfw1G9K2;P~mP;&_*
z1Lm2pE1;BWOZHGmLL!O7_+Y+3A`*ruaF2u$dnTlLAkZTC)oJa-qmSQ@kAC{6vA4G)
z|Irwom{faB#nQrJ+!u_$`TFY~OP0)1Zp@=dDM7ygMuwFT7@o_j2W+fYc~KS!??{W*
z$H&b)U|(W}qspXiRy%iMYhzRJt1y~>XVf;VCn<r-oGn;Uii`oq5PMkih8cQxq4`b&
z+tOrud`Pm*Ua{?Ygfvm@d1Fp{!T!@Y*xS-xU;eHAc=q&3eDMAUv9Y;nc!II4CW^;$
zM}qR^&08@yI~O-@ENh(KkLkGu&qrw7+1%KQjjc`RH>CQXogF13TXRQcv=+UCl{i+D
z)9k3M<GydJW>qDil^nMAgkzp2Lt67c(x}W0NR)9G90xr(bUUbtm@6wz)Xp$wP!aib
z<#{}K@W{-6m{o7TbuVt;Ty$H(_(C$8HiOv$;|tZQsL-XHNZc?_9;*#tzQag?Nm8kd
z26stunBw7+l~`Hb3e=<`4TVbJ8Rf&+quwx>xf5<)ZX6sKtqfYa<y3PyY9En^nVr>G
z)B8|s)aK@fnfk!wSgjVblhs&z@@f3=PyR?{8jO3d-isT{x7EJ$QJt93{5Pn%TBR4P
zbLgP0hUA_mq(awLo`y;vZ`^;yl33MRJ&rF2<N3y3Y-rpeRm!2=;epFoo<nl9>h+Mi
zpe}WHe_v~-U_<?ewi(nt#^s6D=Ze<Fq4@CQkK(=eKh*kv6x??|J2Pvwi_@ArfTdUN
z+=>~MlRHHzAM*+ZI|oNNkb+7?=puu|73HltPX)M?dvDK~@D<IA>Sq}8PuI5M#~*yG
zF`IWhB9XPQFsF58%<#OWxomBHCz>h?nsrDw4@YQDlN(Giq;^3VdV9{zfEpUA#KB=o
ze!{e4&Gx*d`f?B=X)R1mNVB=<hG!(rz$s9f%bt)!_trL~JO>5yRn6nRvrlsd>nQVh
z-g^(dQE;MgLxNo%m|h%WU|xlxhV(M6WIB?<F;4-5v(r;?Pwjc<*0RbusPQg8m06V$
z>>tO3@VnWG(U_TFUKrEh@6#MFc+i@F)F$+*cAL8`FJnUMD|ZCW&Io?ACp2A51wBxj
zdkT-FKKGb{6S5wn3KQwV)1E;~eVwYGdKB-BYe`hTn$|HNJl7Zkw*dF(fqAWpaL8h7
zC+B<AhoXiWsjycx27mCqZ^!K$Q>y2w)=c3k&6e>&#t;GptixuAX)M6x1Hb5~jJ?!b
zFi#&K;aXFk+|>y#1Fi_1-dtMn*gEbsV{};Sl#s5iy^hj4(-_n^)SSg(q_wrp;NDK=
z;M{Ni<ukuwe0`Pqdi%34GiH4K-M{s>lvWZKC~9uSvMMgVehTtLVZU<6wfwfx2Ly7L
z&6O|Q^i!tmQGjut>7J=;rZ#~F&ub~`zp&+-#GN(|XPFeoJ{jHVlB&XFcst2n61e!~
zi)3UYTLmgpOdFB(!YB+b{9I3uH^s2M8z<kTo7;gZIpVc&h3U45Old29vy54n(gjl_
z)qQvCX}65we-(myxw+(*^~e$=Gt5qS@Z&TvWm(f)mBG$VGsC#5c1t^-2Kd51-x(f>
zFD7=%0znf3@1+(bVE~r&;<?l$3ZvDGFLG5{=PVvtT-kJ5#2=(k3l^f(E|nG)7zQA|
zT=<Ga8>e5%hzmkbI)_5|QYS6bIU8g)i9i6%5mXqULz>fFoM==Lji^<$P{3@1xd0Q5
zMUob<Iy(KaSz>{J5y$CS$^v2yGR}sVO$vko8kh~bFJj!+g1)-y3uuuJ`le0KKyS?7
zx)E=@`%ctnX0)l~(#d8VY4Jo#?N|>gG}t7dV$i4mw9(?&h$Zd7p!uh-)Xs((41Mai
zereC4zNpkjV@iloeRfvgLxEHk1O(cbc7r)&6$Btoif4w9l3?tzc)}R6^qdywLm?1m
zcB$Wi0YC~BU!}wyr#a~x$JY;+4<oIH3LYHUc#;oOK{8`i2jvel5-C6!RxrIl(n!bB
zU<wO?h49Jf(?c5o142B2114Ph8Ri{;!Xiaj_oeaxfp@-2pZ*6{aUT|P2@!~L7$&sY
zg-Aj<Va%~{@x5ecz`4>hVyR7%;D~;L7^0}&Jit7Ip$KEPUkCuIJlKGtv${{uaNm%S
zAt9?m>bdued%o~<=1SI#70R6*=z5d(b>nd*$8roil&XkZ`pd=f3qN_-4F{!t6^{T#
zz`~FOtyqM%$uH7K<904T@YQ#wmGQB-d}R#hp>RdU`C^vAUbs09zkH=JgNAp}Hkf?L
zU6u?>$*|{d`658J2R}k=9Z~ioziwE*)DM)$L2L^DuB|Aa>?I+z8@X^Vzsm4luZj-I
zxYt9K;nGG+6a~z_FI=5d8BOn8i)Af_IVQcz7=+Oy3t97>5pZyjeDUcYJ7pw)W=eL)
zgAWJ5Tz%N4H2tTH7{{y&lpy=usT3x^X=0qq^~R>CGsMpB2{75q7EVJh<e|6IC=w#D
z+aq(tWNI=_SNXW!db;5hl6n`$IF;Va6U?1OaWaTJiexh3qP^0+@GSgFrdIY@5mZj3
z?lNRB_g~3e#l^oaLx#aPUV5azU1gk}j2yBqvRq~4@t<+zM~4v65wPmd{`qBReiR<Y
z7HK$7xj6{G`Y*u3{-saGQ|^jD)vVev9sPS{!(7I{_&0w68DIQ-QAm0GTnT(b=1a@<
zk||;8zfJ?({jB=G2yPkni^AvF^gM@uH%NwOgxOC&FiIfaTP2Th1sFwr!Y>8=Ca~=6
zwM-1+5^{)8sG>o#100l7`ptGHcAM>JwUCSwZmRLw-*;)LqjJ6<7#NO0;RMI07d~AJ
z;SW_ZPoF-E<IYLke&wzSR;%kfvAMBpbyAkJ7jD9-UQ5t8ojkCU8V`hf3!h@_C-Vyl
zFYe7s=`RQ$;NoW1Az?H&sxV;*AWp->Wgi(K9K^#s_(OAod%<8rRT4y1+5;vN#4hk2
z$_}%N5^-7^HRuq$AO@i#47?vgy;X`SJ(?K?m^A?Sz<|~h;h$$}I|^j!|IyLlp5Kbf
zIIJ>(m-((8rJam87<ekH_9@iIgf->ET{cs7?o8|V9s#_L`JTf~FpB=o|Ng(nqt713
z-~8|XW-Kq?j0X=NsXzCv3J?5qSZzKq$lX$?lG5G+{7-gpLkizIJhTe1t?iv??zdH3
z)NzeiG7iSWxbSWmh^WCn);_0o&~kernZrRoG&_|082X3%VQGtw@XFPdr*Y8Q6-?D)
zZuySpT=3R&Grzc31cEaXB-9fI<)ywiH@D1axp{LrUVHTw-=AfiU2fp<;MJ&G0)Iz(
z?%9)vvAwb8`rLc{&6t2Wq;R`CoAHw$|5>a&dlU!lR@Cd$ar@P`w1-%Vp~|@6w$Dse
z2&?SHkTh%TY{$nRzHfZ^*4_ItGruVKP=BZ|7#B!9f#0)N1d!-Im_KL72T|!iS9$M7
zwLTT=JFWQ9PanjF+I30ow6wGU;uX#2epJUtO?ZR(3>psc5&|8n3b}(yMGx-OIcV;v
z{C{G}qxps9n3<W6NyS6OzVB-hj8rPzUv;5zbZ$lonnS?vMn*<8KIsqjwP1vMia2;!
zpPCj<Kaz~E1MV8qIF<bZ$r-(2*0NVQ);<`fHup#&VTdZcgQHV3%jjFT1N&w6I~oJr
z-#1?6uHiwYacY7s_p?A~+}msF-SRvET*1g;UfA7jc$^|h!CsE>4s+2e`vsX0#ozw+
zo63JC_SA0M+YR4+wxc$Jp=>EN!RiF}4-NIk`s(9&{P1J-&za_e3AORGrQ~*Z*W>;7
z|19>@w?iY<_}06>9dEt!y_lJs^WNqnG|y-rq}`wkFb~wKRgKYMk3k*``5_^KYC0HR
zFk_K?VxB<af<4#K5o#k3H7*aMwYwD$KmJLqK7OG3oX6aaoAK(K@5bEXE!Az#j6&vO
z<}8?3R-dW7xIgineVQ79aYetua3x<<-9v<BkA<30)PxBbV{UpP7U#7u*7)<CQ=1!_
zD-NyRbzXUmS1YRFxuvX_FG+`OA`B0h;PfSP3vG;a)7W@jeK~4r4pfOje_&2D<9$$b
zmDUE*93IdNd3qQRKYTAf`QXRVt9cdXAq?+ZH*cxWVSXRQ%_WsV^(7r7@b0{FPic&M
z4dQ;f9_=YvMIhcYhp_iWlMNCYG(Ua7dUl|G-`LXH(QJ7yTh@g&4){Xa2BsbozKmJw
zkCY_iTZPxW0}TSnw@h#U;2?HXzqb1E>B@Snu5D<YICFn2%uL7qyEoKVb&XNMrTXCT
zi2EixhI=Gd7G@`{){i+G+7R<GYSK`Lii%3+N9J1?k-N=ieDdINeDvw#XsR60fM^q#
zk*sAfk}avExyJ{m)s9FlGxyKW&#7J6f1@t3Cq|%UDX)F*tW!MFKGxVr&ET8&ZbqlM
z8=pP+IDYcpdzNf{<IOkXJKz1TC15pCtBnWZ+1jS!9qL<asM13sr_z5B&BjLT3wDtp
zt4=Oio!v&GW!Puksc0RVt*hLEs{ySoXM#oO6UV2TPc+^w1ul4c`uH=esD()bqjsy&
zh=%94zF1Iw7G}8uWXy9R%+Xu7G{?;?#);M@7|W@yI1hdUn;-#0LOeK30JBp=!+np0
zrocthF_PObrjKLq@I1B!yD(>9Vgb9%3ov()W<(`rc~A!&;Pj;9xrDi~GFsJIFzRs#
zZp4}leQr~8&&ujXw4159h(3o&$-REy0&H4L8&sdM&Q^v;ECDvCvi7LW$(Qv8mB)K~
zyFLuWc!iEKs5MSYqYp4c|6|<4h&@)hRZGDU2am(^=CS4%>cO1p`2o6)-blrPC#j3P
zmID9W&&GVl+|F3!;5?FbHx_1n7u#?j2N}4BjrQYU0{92_FmlJ5($pk@6y)OEl={1-
zHQ`*ad~BEojyRyiJ(A2p942RM><fp0;ngSDK*E;xCw(MPnP)0$5854R!vWPB1{h3e
z)Kw#peW*U$Z5^9I3=<4#a}LE!XnlG0&VuJT;y-=1Vw{w9pZnJs>zXjM7Iw_QZ0sLs
zP35ovhnWxxK&#V*8PsE%YcI8rh0e>psBgdZTHL+07|os4n4KPv8#fjJe?Wl0g+m?0
zho3x(_4Vx-tJa0P^s1kkpZdJ^UuPhx*nerv^>Y9EP1#aq8D9SWGE;`TYH48ftK<D5
z=BxH+Kc6w<>+k<x|F*_eE>5K}y~2Nds6+T6L(O9R;d?zi;R@H{U2OV_1k(TNw=_wt
zu`T~Kz-@@kOo5}n^JSut;xL((j_E|sBt`yL<u1ec#3Fg&pH45DvQU{w8O@1bvRU$<
zjDBba+sklQ6@9@JF>D;P2<ED}o@GMULl&l+Hg@?dQ>tKRe!4~h9Qm4~8>EPev-^#2
z{eigHG2LA=X6^heew?7}dhnB*)5?P5<Fv4{k<2LC@adgd%dc!}yAe>~6o!pCo8)t0
z37101FNJ7dYJxpixiSuAQvVXAD27ioX^}(UGt4g*Z#Lw)u~pdwCWr~cDDehY6F!TN
zC1bSEp(?H|1OpX*z$^?75WoR#t_HP0jB8Vix<gLkkU#M_VF=TM6O2aoR30RYC=-Yf
zMWYN>)5iu{6^9|hpyc#RxskV6u{IW*NZ@qox{x<6d}dSi%^8(W4LCMkk6W+48q0U@
z#L(!l5e!tsbc6sMA?2k7=ujIvHuo%g5HXPWqL(md9guto#VpL~ODZof`iJscVod{)
zyV5E`wkCxv)rBZRJn&{on?Fu_07G1;3<Czn5=<dB#t=q8w9L#>yIJB){m24I9`rR2
zc>sotd5|p-5}0zFo}{Y8F+wN9q7Nd@rKa>Bn<X=+v^aw(;l^T2ejp~))Y4lplE7Tm
zk1*vxn9N919k~b|m_mgj$Q%f7p5n>202W|6F%DdE=b0Ndq`yG2VYY#M6NC7?O&m~O
z%&jbVATl8KJ{=@iRJWjl62t%lVS&w!{|dw#WC(F`U>D-YQ2%gDj84R;5Vqb6!TF_-
zbdDG5-+e|x6qNpSSyg%};5Mg^W#_?8P$J1fJUcrLr{WMfW((s$j+LLKm-EAN<wrYX
zFF>m6oI$RIllh#D*Y$9``03`Om#p1noJkQ0H`j@}>jh9@@YWAfd*z&0nzZWYOUmr_
zNUwbGSNdMdx&=uu7lGj9?|>OrfRh2zNiRIG6b=|CL%FkWx?u?_J8v6%%1s3?dig2|
z2QK2Ji+zjUw)<7xiwr!m>#a<jkHZx)2+O}SCp;f%=ND5}A+y#G5i_=ai#U7{u!zOj
zO8?TOs7V>OTSk4ApN=oX2X^|oh*P!%?MxSA^<fN|#5oA3FkL(D_O|ip@HcHx*!4qS
z*qH)5PPU5~KC<PVT=?UjX=5@Q#n%ff7D@v+n)#`+xK6fbK4olTmG*oSR$rx~ewb4n
zR&KpY*!f{7mp=G5v0CAB<l}5d+4F73!k#TfiguLAIMmnmQ(HT03~@DvVEA6SQimeG
z%5c5Soo*Sd__T3(un^AC(v)HJ{K3EZSA6|i%+JgC`pasvfBpg(72{XgslfTqZv_2-
z-L<x`t^18Xa9xbK^ySOHY4mm&h0@!qwfWLoZ|N7sgK#zug9xIuaHR3^s&EV~(h4h#
z(<h%zKoL%Ia^4e7ByT#0z6S=01qe^zF9(OIlCA||Zl@6t$3bJbfFDVG>JAZpeqm8K
zZ7tTf_v7L7O-r+&f~vnCbx8vzf+I~tT4%yvIebI?s3Ob<h@6xO0v0$JX{PFB>XvQ6
zHw1DRc|4A*dRv-JxE>Nr^;*pm8$!g4PlH->{|5Eq;1enfAtH0CA4bET%0wF@K~qZp
zP<Q4?@LAO2S?!<jK5$i}UbyGSD!akIh0}uDP&N)9fpgNXl*xFX((>KDmNY}1*rf?X
z5D8J6hT0*B;7G((Mg~lXQdBcM_I7to?7DgDW=u@g<5Kw!tIbE02Pc(xHrC_c|J#3U
zm5HcH+uqu;3d4%pfIFN{g}&A%rVLv!GAC=JZnI<c_mPSXBX4Jydr_Lg13SidD=yDi
zjOmT~D_|8S5q)`~{l{>n5+kGB9h(G4*I(@i!v@@uzU5$+iJ=e(xgSO_(CVZbHDlUJ
zl)aY2DmJPnfmtL2I2jL96nuGhN_(b$;oH5a%H*Kih}wBj`_z8HapL;G;)g$dFaGm?
z{P%H7?SvG`!~}$URB6?;|D{;+hyVcp^hrcPRJw?j$DhUzfAm8W(iWF)#njBa(t;qV
zeSu0uSfqm?(q1Ut+T@Jf2Q2y5f?%eG92*<58mU`Lv#8uK0TLB_oX|(L+NGYw*?KI_
zPe<>-a6DMqiT6K#q&kL~H{;_KGs+lKFcrC1E#aR#qlblWClh*4{h^MKKl_X0c>4HZ
zeDvXm>Yr9JrDvvNad9z_c!U95SN{$Sj+%kCvAG-j8pjj$NlTRpOspMbePbg&`}8Bv
z5x~Q6bs~BPN7M)GL)1TNclJn>jeW;S=QvRJ$3DtT2lXe+CL{u2e9X;Gt8Md6ARopX
z5L^f@fqmvC2-nOT5TTKDfnf&27=|GXc*Zn$cma3d!|Yu+(8riQK1uaqgqfL|ihK8N
zsZCL1HmEWk1oyh6%0rb=b%hCm6f5&wySc0V?~3-H4c{qqcHGu}rzsd!|FD;{=~4gA
z#hrWiEe$g?Too+!dC!JwL|~P}l&D&S$;kemJ(L+SiVFezNNoyZ9cc&pw7tI_+v`ta
zYvXzBY;VNY=9;B`E>E?$RGf+VrTG5u{I$6A>YHi{m?LNE$4SBJlIJ<XST!qi)OkvI
z(6>mD?W>JoNWnDnA=>E~l?j!495dBiDWnCdAo>OPd5-$bwKa{;h^fiC?+IcopPw9f
zoYB8X!7vYf_Sq-#=;3FA?W)?lH$ML46V226aqG@2@yfl|W2`n8$7j77kJ`tpo-q75
zmODJ$ue6$GG_I{Yj>n&UYWQT{W39P=`%cWx%&7klqF?2^F*hCG|L%9=t#5zZl0?92
zQ*aR)%jZ2qO8X)XR5pm}Fk4a4isTxMDb@{EJ=CbdyjywxELK-n)wjo16Ny9>^9ho(
z6~Xo6=WFp_e)MC_<8@0*F3eAv7Q+LQ=ebPb_mu9sB}bqEs6W+q+z(i<3O-Md;{2!`
zckkYc8#k9d=M#p#D*aE{*-OJLXB@*EVXw{lxUsS6vAn!6ACu!Mv(_`Dbp{4{EMfA|
zCl6ys?aAGB(|X1fA7%gt6BsMINPp_t*x8HKbyR(}J?Ekhb8dDjrbsi+;^?sDHH`U^
z@jIcup4E7%C?8bDu|7P1_Bh`A@sBLM_E-MO4`Nd7#agzdwEx2&{%5cG^K%PP(_AyI
zwP#{f{lWOty7X*qJ*pG4z7q=?0C&|_)i+#CaeS<?q&0n5ea+m(9MP}6&au@4qxuhZ
znMfUKjnEv4WEBiqnARNFm>2BBP#hcSi!s3xs#SqY4ho?D_3_hH_1A!4So5;rXF}~<
zA0O7dHxrYS1v5pE2onOjlA2g$I<<7s)_yylZtO+pRCta03&s$AXRSHvdnAA=NTI1u
zw92c`&tq8A7KhYfh9Duu__De@)fGwjjqM#vsl$w7yq^j_xI1)SZ8D_w8rl%dQPvaY
zsSw_QCMlS(nrN+#a%xI_JvyrO;KWS-);_9s^X^9^s7E<CpheX8>Z$E~PpHaE|DtxY
z)ofbbB$C|VLmZ?T;IK1{7#ZUyw7Kj&(4m;0!DX1+nBSolF~_hY<G@O5e^+%W?#)tK
zsP$V~ocDT#G&)zVu<r!UYg)h5+L})__X>{HJTWsl5moi!>Cu5{SsbYO=?5Qc9X*Vh
z>8W`4-FM8~_1*ef=ja0@msi$ztWJ^xHQb{K7H1~8x(Jf1?`G6P-$GApH1;*Fw%nIX
z3o~(NaoPtAK6$(z&sMk0XyX9ExZoJcKl<a@+EzSS-LX1t7>M-Efa=dRAkeGG&+Gri
zY23cC5Wn}mZ$)(&JoHpJ<hbBqz^XhSJz3LQ)z<tBLlA)m8P%OOHUgBN`3AYVW>~|1
z5%Y37FPmR9-Ji$&`7*|Sw)rNC_QiWQ)`_dkvOj)ypS*06@%6X=2bLT=kYG0nO=@Mn
zxI2V?uClzcVF=;4C50kC@>~BdZ0Y4N`R!DSq@2q&KbN6&m*FeJ^IL@Bhfmd_n>PmM
z44ipm<L$8iEKE9-&Q#eo>4qV`ei>Edam{29;=bsP!R<x=qV#2Z;Lj(I!{u+jv)-8e
z(wk_6Ej-FL%+v8jC?}yGfI$iS)9$h}nUI-2OHvjIW*DqWha`FOCsQR(MshkfPM`nc
zyL7k7`Z_q#^aEPTX4SP4l7ni)UQL2$LJChg`6fUk9}jF6!6FR|iv)``%!oYEnWq|9
zya)r5$b#;Zc0w{Zg}|vgPHD3Ff_Va>4>IaOqVP=TY7L8Gj|;JxnV2-Q1SSc{v`>|2
zLT1Cy0!8^K4G1tNe(a}-AI4279Yx>3{PGD8A>N$MgW1JN!d-3B+AU6RZpNM#3fkhE
zFupTQ<+~89&T1og`~H3F?ic3cwqvu7j_)(I$(c4wAX>BwOgu!x5tsM9X)52D>OebV
zXcM2NkO2uYVZ_La9Mqo`ZM3H5XJcx1*1D3>&&{1+mcCOM_q+O*%{PlIOda|Gctiya
zI=0W$7bikm*jSJ@gbg-##AXAJKNAulLc3E>vPZ=w2oJEshMdzreLlUadR?B%o~K+%
z4`he5IAKH`JtmZXuNG2EYpE>wL4+w072lyDe@+=?H4;vJ`KsWW6JsD6&R_X}!~h?h
zgrcAGwD5=!;FL6^Zd@cxd4TQG07oVWmu;_bh_<6I83P=eh^8RGHN+h1g=)#6!C`G~
z(G9Il;aJ^_ul@`5*(Lo(4XKIDwK48KkvqQFu-k*Tt+JDuoeFo16L{HNPv^?`yNFuE
z?z)u>1bdT@9r;M>^$0H`8)Xoi7d~Y4O-YitUO3n#19eNPe`i}5`xfb3+0%jFZWHL8
zX%&&(PFLxtQ7-<JYf-nduHASUAWdQAMfuY5LM~~jlikZ&5cXPRzOSWQB;&TWU-A0l
zK77$Q_cB}w2W0}JmkZ@plErU^yXG!S`V<xnj%B!O?NjDsFFCmE*cH+B^@o1Sj>F$p
z!lqj`QHJq6tH*7s#Q5~;QlxzqKkW|2PxrfUR7CCDGTkDML*=I*$aGFCfAq+>KG-u=
zyXrghR2kfc$vkpa*NlV6Ks-*Pcaip0P<@kL+6teHS$^B)Dni$!Q<0Y2Bjc6D&#xCv
zIS#wk!W8+JCVQ#~t8nC<Ve>l!`$y(_E>~{MtB%K)GII|3;Sgnx|KeZ$hVk{QGB3}u
zf1c)x;Pn?ZUyt$idE=sV%LBOSDPfY)$xC<y%xUlda0>7r@DtjU6TYY*gZTjA1sY$k
z20rSNxH#UDS&C=$N)JpA?sp+s2rTX6Q=c|Rm2+(@Ref;?r?JNg^aJ4&!WV>3fg_t)
zCOig0i3xvIB#MM1pyui5Bv#fNvAKH?Cqg?=FSE71YxPFNg?Zw(a(jWx3<M|HDF=tS
zz{Pxjs{SAdLr6zW9R#8B@DjK<Ed-9k!vO%)zadEi{=%UpREUF5!gP|8iQ?RWA!Nun
z7sR-p-T~qCr@~J;QO>Dp;o8cFbRY~-Ch#R_Y~V@s0VlU<|FcV_Bb>~P0@71xh<`96
z=m!qAL13{&8s!oWMOueykJdq3IQy|GkgDJEZWxspTvl(S)0+FuSXo&y@o$(D%44H3
zQLjc_Z9QHQ-rI8??cMG8PygXR#N$Vw#qa*^Ux~5tT0HpVQ`N1{5*tIKW6?8EQ3v-c
zjg%}wEg%))?h;hofh%4JkvGn&{vq%2ap85U3u@mx9n>8`?5Df90A4t=@XA5;HPQ?~
zs36e9-(i)VzGZI$!HT{-6zl-lv~}|kLg1MhOdPa8+K+>RCVnX`h*99d04a<TRA-@0
z<KnCny%(LBo0*K;w{H7xFBq5Gg3Z<esw6umx*zLdtgJm>F?`&*eLKGQz3-ZkJ`m^H
zmuzWov|(Zvr}vSRfVhbygZEc@Ahe<uY-$$uW~YJrK_sH+dkELu4>DV?sSX-D!mT+y
z4+Db(ZD>y5E}7AhWWXMs_Qr$9Yw=9u0X)7wIWAZjQhn40Dj@v|tRpeRY4;J0JETB>
zH6OCm8&ziz<4-!x*xTNWCyyRldTMcLNo}PzRi6F*NWF}xF0An|j|P1Q@>G31YE{%X
z9SP14e0R|w{ipvJAAS6>CE0G>c{M84TJ#9U#;Oy6G?weBu>;W$;xf`Pz`xr<^8<$~
zk(3~Ph@-yCihYgxo;zz8XO`Te4r*7_prLw-u>unm$vaCUj0_t-8fsspnP4VB`k;Ln
z`|ZO6!zAaSP`kRkG^4)18K@~`-a7201kQf5tupP#-u|9X`kVNwJ$&y4u&w##6sgXA
z^$n^L_hVoC(!*BMVSp8g_lyr1ji`Z}n4F4QeL6<f7Z8E(-+xVQQ&oQ@L7KZ_=&$YV
z9oG@Y7l#G5)}O`WPd_x14S09jj4t*{(+jb*d@q)6-Zj$^=_2;SOG}GN|6bH;;|i<3
zgW+|C`ox1cP+QR!NT1*ba}eS+cn@kKVVE<gK*0B1If6gtBJKr5)my@{iAk1-85&fd
z_FYB`2Ho0om5e#~JQ{*2=A745&$+n;(=eVqeI&SjEZ8~{%q|EnC)J0lw<dwj&GlG+
z{#5O`6YXZhxbH-*W|gPRfu}0lq+sKF@4jPq=3gX1hL!*1L{0rSZb`42?97_<<KDqZ
z>??g>?Be3gbAcra)xJ&bO+7wTM((Sd6g*=`73tpYj^`}XXf#K`2pAk0HB%qed~KzP
zbPy7O<d2kU-oF%9dDzhSZyyORHMg-gS?%2DP@HvI@!<U*N5A&<zy1C1#{9y(*9)jK
zz&=cz&BngwAMS%cRzHoTl&Q*%8bpZmFvTuZ@6%(=e`hBu!&ofNO<FbQj^eGXt_OEu
z&Cg7!J}SG`TdUmb(Y$+j9F1ltws+g{e0|H3eLhqH^G9<TMERPTWk}GRn~q><BIdHg
zj^L-Iwf@4=x~MA!41Cl@b1pO(ODbwkc<}IXtZo9wDY-qOVK7nA{I7iOzww&dbKHkM
z_V)HQ#~!GjT8GeDG&rL1hMLm@&1+iQw8o){Xb?Jw8rRUTG%qxtpX*_aF_&;qgz=9C
zBi2f!Rk>RWiMd*J%rpe%+l}o;G<Nqi7eNE#z=QArRbqa2N@HeD<<L5$exM96gtoUe
zuWxL5E?^F9sD5q1RsYba+KszGxqoj$^X-V-D#jH!Tmr*|!v#pHs$r}t>B-X-mlLTg
zq!BMw{tgoCDzlo}u*1D#2koZHc5F3{sGnusr%h=q@RX^^2{XE&(E-D#B?Wf@CxNC1
z9S3G7(#viznNh9Hz`Z)lb02s1qaxS`bkrWKsmv=d%e-EzKrj|LECT+(+?})@a4e=Y
z>}r0MCvmC7OURpb*IJm0gZ7kxazTqkI&xZNXI?^T4TdtbC8Wntm3^RjbZzCi=kIB)
zh1{=(YD*4Z`%caN#0w5IC!>1y?ww`dzlOA!8E2YD4^Mkyr+KXQRXZp@@HTKq7`{k+
z(e4}ur_9jXm`nHeTJiAdb2HrUETn|j`p#i|{Af)uv+q9R?nflUkP17y?A6-49c!E0
zR^tpzGLHjio=cTCRi`{Kir;?ym3a5fS7blYc<9l*Kjyi#v8OV1Qj(gsWlbU;1{Xg3
zE=;+{q22#78sOtEs`)00_Kh;W{=xt7cM955Cgu~yCHNHSWj_2QS>yBBP?&r>l3#Yf
z;#f(^NH?cI1Qog)6r>x-^+)D|r+(9o`2NB#0XM_sWRm3~?sb#l0T?G?w=8R>mPlQe
zFzdyGfhg6D<2N>c=@-&WrkyUOx%w)qSA6JScBe-W>^Sq?jpGl0(V32`NLeyl?U}W_
zax0rA%X-xY`CGg-(DO~&vT35}29Uu@O!9rXM&)N&m$Lj_BZo46*<fW5-q_C7m`qNd
z?)$Z=)&kzE3HfZ$C~&qV@syS&Z4{PGe93rW6U~ML<_rr93?=Fc9DvB6>V(Y!i#7;P
zO&d)k@LDh+47h9x?KHKy0WoIB$SIhL7N<EakQE_GW?rzl&>{$90>gsHLd9GOa;899
zSPLeRl0DNRL+bc(N`W%-5T6sPh)(Zokwn+>=EkNbdLiInG@gG2ChvXH0x~ibi+66v
z@|`={2#vd2P+QV$Xj8ep9VbHYS%^^?Nt=2@O#6c13K2rJ4a5fbHG~|POnTFoG#!kq
z3oVvTSk5E0v6x<%i^;h;EwG$4LS2kJ)E+D{oN$I&!ygMeeWs4_hMLoGETX6a;Sy=>
zFfuRoo2Agyj=VXE=YCUTQ<9X{Axv&2V1vP!WGuK3w0X#dG{y4I)bAbF5k_A@SkmKq
zs*DhMe5V_KbgUloV`I$893=Zn<i(=?a!0DmYIjuD%m?*_s3sAdDj$#)5Kxd5kV4|(
z#wLRLrYg@fAxf&3k_OhOJ4hM%!}#J3ITM5ww^v9eh={CyC^K+Ar1Xaeh6T$bQ6HNS
zf;<reLe$mzu4Q*rE6tL@?eCWiZGtKN2v>d>7BK2_tY)T^{#gAk9Ew*N#2ATTzKC0t
z1CyheH-TSFJMDTD9-p$@xcRN;`WxdSIJKhA_PzRbI<j@czvRX{%kp9h_QQg&-WfK1
zAjY|R_Lqx13#$wv?W^|4S<4N>w;rdJiFN&4eY<7J(q4q;r!X#0QJ75IuKJX{NF<{Y
zMz*YgjbYp|GQkX&c1+eS%Uzo5?Qvx?4VSZwFDJ(@0%qEUMW&P_Ed5JUcz2y&2%_@3
zs&dR~6d&nTS~Efdl&yI%q%<1SxTgm{g==a1B2%=dZ5f$w)~t-f@AVY0W3ojsj9u*6
z=4q~C6!G=41R3eeZYCS?Y<EECnc<0{aOK+OM(MuvQ$*sQbR31W2~*TZA1{t!{}527
z>`%gEq8`EyLisZO3r2o;W9+5Jb&-`*yhy)g7D`+BzA#>#(=}6sy(->S2q%GE|9|kW
zf5Z6tRhiHA=bxjY-+xu+mt}kzI*Yb3?hFJ9o_!b!90XjY7paR=t6M_t3hH-|kN{r*
z=bVHwGJ&*Ho|1;JiJCki0GbOSAPv}$io<{q8YbNJGLBVrR4;L=ADj$g%%JcK)XBhn
zP~yUMktpvKes^ruSm2zL7S)NULPMIRG7QnWCmMV0Slei%ltQazf(9rRcnieA99)`Z
zG_Sx&V(5e0m{{0_540jA5j}|%G>6&)d{PBZDJxW64VoYb(Fg{LCG~~-fZO<Rlj49g
zRk?Qu22>x?VTx+BQu_j-K%BBi3Jsz*c#@e_0}%fy8;8t<rz#+fKMrG|^6ZR$QBfhJ
zpuz_xMye-dQD7zs6KEl}a+;g^9INkO(x5h*zDAk@!W?)pa}9lHobIR-`+F^y30!S#
zjMMI;zPNs3OgQhwNo+oU8Xx`S$MK*4(;u2ZfA7^-t={a}(`Q!ubNBvhF+D%8b{QA0
zJ`#h%)gVMtUhdg|Xg4ApesOxz#I|kJQ0?p_BWHFliPESD60D)-5{&>I*<qGZaZAvt
zzd0p8q`g+Zp7Rq9lnF-`?4SyY!(tpt>+P>ZWB<_iSkZR06ViFKDbj2l8bdM*LKCpZ
z{s@T>)MY}%<$jwU^~>bMxYd5m%}g2pMV%wWVN{jvK`2$HLNrIxr`_C%+jo}Z{{8z(
zKYWots@XWHPai_ahDgeN9ed4IG<NsB2jIRq+K+w3;n9($Y>@7O$v{7V`=3H7S3NIL
zlcb&H;_QT_7f>e!voJMJC{NX$^mkfnN7_yO%V9^x3DPH^*xZHJu&T9_7#iU&5eTyy
zqe_Fl68n}ZR80<R&x#s7?XQq<qI`qei$a7){nq*>jGc~&g{b^Qig1j4R1oeQ+FW1J
z9&smDR-VRB1UE2%k^wOt^{J`&&Ub%1Zr!|Nct1YUJRsPfTf7mY<8{G8pBX#{N`G}@
zThCT>v>$=ciDb?=<4)}YjGL+DIVPpPV0<%vsve(75C!Kz>VUZbW=E@i7-oJk7?nTL
zf{YRN-|W+oEQNUkBZYC!UI{4&;LObAj`lC=r-sH1#BZbxm=nj;)-VZBJ9l<YS@vUh
zYdzYHjo8zhHfgnXbKIK|RvJbx^Z9^PLsr#JFeI7d83!=vIo!z{#y<Vx1SU2_X-qkP
zX$}}1i^bb-#9QzFc8u3%R8G}NeFw1`hSs?1gQO5`2h(^$@HLH8lhQ=W1jZh0WnfRu
z8Qd4nSOrJoE~UNJA*W;unnWIefNsIu&Km|9d+k(nJK{NdZfY!gPY+{r?QwKechs0o
zPu7)2#d~Pl7O6MnFh6_p%&Pxd`)y0UjEzr3qp=-N9)239N6lDXT#VcIUKgA!d;D^@
z*S6-RUbW?&+qVQS?fB&3gIM2K(;PJ(^Ru&_=TYtVy2d4H;LL<>Xs$cC5Ij~b*+lyY
zz&-!_7;G;RpBxhCQ(J3I+G)hr#)jJR(6BK%UbA!=eX+T{?RoP^aKZdKH$SVPbZSOX
zRb{BE4loCs2Wnr<G3X<JLECEXtDgJO*xFG2K8vIFe*EBjzZI|Dy{j^xX$~EVVZj0A
zgh_Xxxe$gfYi7Hn^|c4O4102|JL(_y`W&K00}JzyrEMq1VnTBqhZtB(>2nSuj97Ic
z%n8jVnpX}wXM$s3TJ`YSsq)wJet7of93JTt5L8FFTP)Wa+ExE%2tg}=X?UVB;JHg>
z;l9O7nDqi#qy}OBz^M7`*?K&9{9It$6Zh}jiiuI~l1hy>Fh&^LR0vhGq<O0NF#n)6
z4N{iVPCQ**6TBX2+>ga%ZPNY7oo1IXyp%353XE5@MxYk&;NV1V17zuPv##1-^<(`-
z;+q5JP)%XBrc`pv>kwmpQm|H4pL6FLjQA%{o@icsrhYt&5v_NO*~KNK&xNz7PtXX&
zJis7Am7c&PhSmN|SscuOi46muI}BmCF>b&)DySva9Kw7=+nRnOe8ubY{+{O0hQ}ee
z2sj2>h)|tMo-mkDPdPm^ZTtY71<5U#j@;43ffDFL9E3MRPW4KaooFS>yY$-arsgeh
zH(>{Qz-?f#0cWT+M4hbfZIqpRFo9FgJFN4JAtnuGTy;7yU`aM`8nrC<D_T-cZOwcG
zUcg*xDN^+_QIUwQ)h9HMRud-_d;%B1AdGluTC_P2?x+VvaujvZWgeFr0}EOkrvz8E
z;q#cTDjmTtDiHfe#x1!G%mc*0?Uryy4#>2RG|rmbL78h{jl)%Hqiy-09JOO%a>&f&
zjopKI@_fgPdFJRo)dxvz4lc~h&CnW_Zd_U2RG5zPL3)?D0m;-+GklR`J~(Jab-3SZ
zLKo*Igd++1)Q0pKI5(%ol7S)k0T%SuOMO92zu0`?{`KcH*O@L~WL{49tHOUB^Bcw&
zhH-S2S+@7r>q|4f{^8&G9~DdgP&L>wkSGsU{doGeoBoURix*B;c8B3hZ_0}&1~>Hd
z;phm3NBP3LNF^(jOd5zXxpc_aaq^q6Ab~DokzHDdlOeA1AP(t5{LP>a-i?eYNm%yL
z@5=*5UrtvqrCX-+B4IJzy7ebKR++LWrE&PeuLznyaw|`&W3o*O_jG!}IH*G@ufm@g
z4$rG_^GgBTs&s=Qx;<S8#g!Yj%ug<bJD&t|{KAx8X$YW~!;}H_gIn6J>MjHckT~O8
zi!4m9^C2O}THr2O<*0xf400IKJG33TWtq6OF8CXUmN#Bne7I<lg_gpz$e1x8KeA?n
zjz~ZLh|wYq!zP^5DV)@TfdSHfDI}q#MUtg}1$%6WlR65kcn~#M{Ky=pN!C@HDt)tX
z=7}=&d+X5~Ni&cZ7%sFCa7@{Ew$XdOqw+R=neUpARqp!wCQXkJuCbZPSiW~x8|*nR
z04$uSAK*k?yRm1af)iO_4V2k!B7dLCQ#El1nPTDfnB-mbEt63)pb9dq{6^|RtQHrn
z>cL{(Y!J0@(XYPSO9(z2LZrPo5l7pg9)-39F##-LUV@N<jB+xPegu)>P62urW+aFU
zVUhY{^K~JF(VGJG8B$<0K55f;s1Aal#C;8Hwm{-MS_Cr?U@+9lbqCRKo#am*l)b2{
zPqnCyFkV4&SeQ$K4=^n!c_)JpM=wr(3O6N08_<?+Q@rE`GcO56Fw{!8QL_mFV6<Y)
ztQ6u8?AuReai65O+iqmIB1A|VtiC}ryQXW?F*#QEH;gZ#;toqWDY24y!PuEHDS*f8
z3xl3U$ZtFGqi@PmmKO`{P)>aq7NcB{GA!<_VGiz}gdP7VKzGi;fw08GUnzrRr(MR8
z^pb<!ifAs2-DM$fJ7xgGFV1`xCc_qKU<!Y}$+2|5ii3?n>bb%((?~b^Gusz;+EKnP
zRPoNx_P=h$q^}%tClNOort5PRylmudIAFF6S%h<$fW=~bW`kq1U-@OVj)l9+EL+#;
zBYld#b=0C=3*+|dhRd*5sTaTYECONN9>r$?Z^W>1*tC}B1t;hKqUrj}VmlvYpu$cp
z`(f9QTZMkqpX0hrMI;Br)wNcfqGcQ~y$K37L5K`Pa%IhnwDii6uRo<x!eF{2gz=Z(
zMe+)Fp+3v8px90qj~ufZ2TPX8C2*QDO8+V~a#kL)UJsuEu$ARW_;4a*<iLd~oYE+K
zk(kFcdEsxr%+XC&xE0oXyDnGXUC-jxEj1rS*wV+J@}~3|ksu07TN`$I;t&4Szf6x`
zgZT~P>(^(1uV37}82M}0%VQ(s@SCunkSc5-Q2D~$S`dED?Z<4;tBv@P3EU71Sf_mu
zN@=lPBW+O`8H(W{a2}XnIfsB#!U#iE4NMqRXG4I2iP3>@qWq9MhoEJgM)(BytR<r4
z2JzttwPEMn@zSd&8A8+<<_?7J;o&^Ub$Hl`^^NV=+S)bIjymSx;eIK~1<{|wKZ>t9
zfUEdWPC<l(Fav>xun@z*ZKxZ0VW-JVun}$r^Q2lsh2waPs9wYY2Ln&%Fb}GHP;mtA
z&V5R#?c<a=1Ynq8;GtaT4$<r2;Mn=2%>bexcR(QZMj4RmKN5~aeYn4jGSed_SPu=y
zSQQmJRTD13nINQse{$*`<`54!9+C-2^!rc}^CBJZzE*mQNV$>r=s;vVgAtaR1vuFc
zv8dDDi{pd+Xg0Rv(Zf&U*`r6k%O-I<n0$)UuRV`oW_oTme&=_8V2PE{(P|tVUqq8r
z%gPU89=L2(>CV=x0=Uz7?<YTrPd@%cZBvbz*%{%DN>4aypWxv9vR`TU`;A1A3Bi5I
zARUJ(pnVSnGYH@Psy_^>i*xQrmVIg_I>(o>)o8`;e#>o082T#vOSrY_KQc0CDG=@;
z0*BVWC2#;@!W^^@+@8SjsKSCDL%?e69r}JcRA}z*?rX1fsQR3%|6p{%Bms`{z8xfE
zASkk@s!vYE{KA6DKBB##@MF#Ov_E(BSVg_+2JXMHwdwsE(sIkUmMqzE%$-`T19fyt
z7WG|1+*TbbBOZ$<Di_2z+Jlqp^Gmn27p-VKv!_W392n3L9U=0cpZCUgqiH4<jn>Pd
zP2=YVGnczIHW)o;CxVTZ;luZ;$s1<GWzS$VTOAWatqBC8rouA5Q04jT>0|8?*E}cG
z>ow04?dD#rudnIZG!g#4{crzfeD}M*6Qh+e)%ze?YNy)7r1n^<Q@`d6<$+}2;n8Vq
zG_>DTTQf(|Up2K2ds+;KO<?}A7oq*QH?C!=zaBI7fEP;@Dh}!r*`L7-g85_yv(kjA
znd6B0h?z;_Uv=@_GQc_`)pL+qi0Msv($43?HIFntfnCNLdr!}e3dW&GU`$!ZNU!$)
z2hCW0{v;kfdJvBvJ&cWwb-`H6?UoufI2_m`SY+>{mNsLH9_fqO`Ndeic~jv~L)KRO
zM{y|lOjTIW3-qX8C#L4&-ka~L|L?@!e%lfp)R7qx^>EaaxsCF{1evN;<L>Pxt3N~?
zTL<P9bF|uq0|ZB!TamDxu1~lvnTKF>m@c72C@*u}f!Z4toiG4tr_qs$;0?)Ajn`4l
ztrwj*Xf`~bpcQXSeUAD`)&&*Mk`g_GqtUCr1t){)1R<FRLjTt0R-6jHXS5HmP0nhK
zy0B`;m8XxiX+Dga)}U8z-}1qS_doh5HZ4`Av97k=+S>5pq;G%gt$6*lH`J3o@#M*h
z8KAG+yBD|a+=&6ryD+CtI-THtCk>Anf?*7Uesd#!`e#3mPe1y=j9l>h<;8{g!SDRQ
zl35(oV2&AvJ^&3xV{CI{BPO&)OwCMLGY!(BN2lCvs5wjdY;JAE!_OXQjBIP&XloUS
zc<0-1#l73hak#%5&sSF52Q!OHzW0iIsHSJ;+*YX8JJwu%a0H|NB${oFd!@;Nj$FgE
zrU>?6z7DD#k=~>pFv3vFR?)gSBzOE_ps>~;F~eHSJ#xUYiS#hD$0~vq7<FiCNtKg;
z8gR)_fQAIE1L-Q%QX+lEJ$o!K!-L9S@XDT_c0uhfxb4cuX1xF5N1CfTKHPG1aW3jw
zQ(0eupOSt;`mEg?4nVTf2XZtfo~Yg4`^o#RJ9GWq!lLHET7>GrTG3P;wsy7!y9Y6<
zGQem!LKM7r&@c{egQ`__-^_Cw60R@OaJ0qYQQmLAr}Dx0Vr@lT;DF{+))V?0n1oiu
zT8#S0`c&OvxKkb}+zYj#)h4TBVPe5}Brn#1E%p80epB_+#Lze3`XNC_Tfs0Jsi@6m
zhqeWC9-0>8i@Q)!`D_{=^O5=mtvIy_Bn&mj4zcEI?xJ5b@3Ni<5Ild#NpR4k^5q?<
ztVz(Npd*6QCH)2Zk6K0W;sfF?7YB3H?#e5r(U8;@EFY<lH7`0p@UtQfO8`<vt;OI6
z%%h}Bnbg646X!y32Y$u;A!nvkCUT7(=G8Jim}bzq`V2c&;XA`fY7L+wmpbWBbz)8=
zJZm*;4CCoS^R||+sE=I6Y^@SE=BMNC-McX>SV0XUhh&<&+gdYdo_?z+T;1G_vlxus
zRvtbX26j~@RD`lt%ySQy#{0SYjr-bMH<cCUSX(#%YY)<CNQI#(>)G>-c)GSDd>8y0
z#&IZ(x}_UYdCEZnq{63bBQZCvF)kQ_>BXJ4s34`ynW$+SaCJ9fHq#4+GF~@s(?s;X
zmU%gym(9y*epUFdV?I}h;{6xFfL~mhU$kT|w%r%^4}E@>eunv?a9?f;#@F9ftc(bd
zYDlroFff_eB`evOfWupVr45tYCkCfK(^p;wNQdD`&x}!;%&qI_AUVkMLl61uRi=c&
z#|$nkr71k^PD9v}l9nmymw!39yk#eux<hzgSE23pf+0@nL*7c8Xhl9YWfX_Qz3}5(
zuk=ngelb~=bS?e5<|<HO3V-r+=7p=QCbSmDeU`TT=UaWj;$jCGhbWWi{@_}+AOHEy
zgnN~|+^)fajX&h3^zEYf4xu+Ey&YAH#JMaWN<EtN`m_-}5qzQMutyW8*<$i%)5m6!
zGQi|GE9@ZGsD&Vmb*-}?>JL=a8$m53Y+T6;B!I<@$(Btl$OI=uIF;iKy3zt^0dZ^Z
z?MI*DPH3^HYcU-c5{FJ>kTy;S!UTc2z=F?0M4cfHkS-$0Db}0|JID=}G<rc&NjO!<
z<)y6_BGFADdE4>)`I^4lw8%GR`qj1zuilQuJGX?8je7F}(c@T~hePFku7w<i3ipti
z38eC~5n&U7u5>lH7irqbaOPI!DeZE%k)5y*0P5pWpP7zvAw{EVOI9y7Jx0uc3(^uS
zz>I<+NxlY0kOA6Du)xL;rYCxqNg9=loVo-FA`ZkGU>yPvn|Y8w2s4a9;09HNAhZyx
zK~&X+N(LfEK331DKWYFRKpX&JK+230q-^R*3wSeP0SqclHrw8O8wn|@t#FhH7;{^>
zlC&F$a^495B4l_`8W4gM%xS>zLh8)0Ek`y+W<DN+AZ7jdBR>7q(k2~PB~6HF6ab`)
zcmvwdRjSn(8PsNtcD(G5sqyKkYt!8?1XgI8>I3XjJn95eMGH_78OEbqga*%gD_L37
zHLI(5*W%zKZ~J4)wDViP<O5s)KJMf)ZniCHd9mWbkalS&ES-$WD1Mh#n<7Y=V`(eD
zX>*rB-&gZA9z{gMfs8*@7FF2ga%6rpois#F=dwjy_M$z~-8R1qFn2T?>@sE6pa0M&
z)*_sp3wx%a5B0SJIbP9r-3+c7oE*1nv6W~VfU<&EGZwNYFN7=XnfB*R*0Tt&|NLeq
zxN3X+i&TpDmEoJ8th)SF59*rjRwiiAET6u$Z_SA5oN1Kd`Qo2A3Y%^?5y~F6Wk~tk
zk|kpuD&r8w;Y&j~{MM6^^_InN#g>(_6_)OsGRvY*r3-vzhW@vWV7&R3CFdqb!Bt7v
z;%vHQkRd%s#;(VQJJgr{cDS;h`O81v)Dwa_uHCWBN4h8+0rQ)du-(y%IO$ciTj5<M
zA`8WG|D+EV`<3Aw3@80u^|#)xOA*e$gw6Q=b|x93WIDQ^Xjg|&7<S}En1B8o#@DaR
zeA(Fl^Dw~IFKWv1@$;H)Rt8!@SqX!NSGf^DYb$WdNogL4MyP0Mbs%n}YGe~I61){$
zI3q+_;RvJ)<ib>83_x54Ujcr7;(G{!#gK_xoS;R!0NMgztZ`=&hlpUZaEhDp4}s~R
za~iD!2oD_-yC`S(4lAcEyjOTCC-G752L6irBZ%$I7HYr2nO)1`=}}(bVKB>vO^l*0
z5Un_z15r2$hKXarSTatfI+t(*o{T@tw$af-O`z}+jPcK-j)_W0B0>1-pb|~(2o4Pc
zaA;&Orl+cLYpx!n+9N<T+1%cZt;W7^Ch*!pODRE|tjG?dkaT?C)_z;<1?~i1rg>60
z8>+^*hp0MQH7*4X2YzD>2WnrKHsG2Ntx!?K0jOTBwMa-oB!hr-rang+j55-H+!qBd
z1`~%vWM`+H=;)1#v9aO77*n1!2jRCi)_tcCILM&Nj8q9yHZa?!XC^Jd#T<vk&Z%|`
z%eU{vox87@a44)i1QyY5?#Cyed=fi5+iKfEGi@4;P50Fls^7R1YP6<J&sbE(C$$e5
zR9mX=wbwy~;+f!<eG!-XPgI9vZfe|&L+*h=n(jztI#s)%!i#&&&eRUH-SEhe36K+G
z+_#b%4j{B1w1tlf?uIJWKs`8z{vb#K`<5sYZ2N$kV2_LR*VeaUV|y=E7pkvdy7XOY
z@5SLawI}eySU~y?T$_G@s5z_l0fq?AokvK;w5&O!r9Pz1lE_<8Bh60MOnkmnSo)a7
zwTHA%^X}jE)#q9>gg*=JkcNOU!u`WY_yN01%eQ0x#!Zca-1E>r^cg4iqfc#hb{V^>
zi`DhH%#7s_?0BICsaHRqMr~B%wErTG4q@s-faZ{<#;MwsJs9IUB@ns4tS|QV+OfK}
zuJleU{r2j;yP5;Wv^VI)v!{>Z{hz|{{J?F$uy7-8-n<oe@7;^Z$w@PU);Bg)zQL$Z
zAw4*yI*oYVK*EFh0p=TRJ1V$Dl_+fuQ5<!lJV=e8YHDM>5lFRhD3Lt{r}bg-ET|6#
zH4b5Bv)@MTB#dXqwV9izn!(f%%uQwt2zJb%RQqrTnc+!wX3X`doI`^Zj{^w&zzpen
zzXb7_c}4K9*)>qh$Q;JL`Ps9lv7!Fk*jS0(oh?f_AVqnmv5I5|`|9>V%X1x4a`SUH
zOyGwAySBa>kDokIU!iIiEhHnF=KA7Po08G-nk7IW!XBuv=tu54^nOlx11rqWv>#Gi
zNJTBp&-#uX?0qV%iT{d^>PPlzgPQ-4?41!jPIzo{I2M9<v5sgy1rF#pm`4L>=olXH
z96d6kIaRoMpW15>W{liYlSq%Hkr*?LMl;qnnz5}pzgP3%l=8W?I2|+7b<eT%59%+G
zDr>dd-g`5Cw$yefg2`KpbH11FLT!(VJ`Pp*FzAK$*z#Xpc@is6pT_L;wC16^hKG&k
ztFhN;#PsZREZtgGJ6%QthN;HGk>-%oKJ9HeG(4I_^6k}?*x7g<2dzDo2no{Rn3<XN
z9au;)!9akiebhOKAN}c%qpkLU@jX8`BN(4opSMix-q5_?Y^t7`^UPEn8;x7H?!+s1
zUWwbcmg2a*AFEG4^B8&m<4-)EMk|$g?f$Fr%I({N$pOQ`g2r{<&}f{fu1JJ&nBwRJ
zT9L}`J-GTo^HxpsOGSMOLjxvkb(Fpl?4O*Qk&P4~YanY1Ohyi%Si)23F$IixOlVEi
zoB-sQo*`Ibu7qJbJyDJ7xXPjSW*zXlr*#$D04h_NYX=90%v|<CWsR$Y1NHURhT7^#
z`3-1psGAu9?SlCq<_U8UGz^%hF!+&bg;4_&pYhicjsoZ*ZQ46H6wQOvSXtYR_3cJ%
z3K!Vk-c%h9yjK9B&jUTwfwo{BAN86>`bHQuE}*4>7qG6tOkrJu`2xcSW-bgy@REJ?
zCG!z;3Wpd_!HV=XW53fr5R9~fJIpck-Sq5~`fOfh9Woq13!<$VBdjwVu%KO;A2|3z
zJHROAPR{0j%U41$zDVCjWi``TutEON1bq;ixn1iI2de2)+7G+}dJu=#$28ty{wsOS
z4~JeefL$2p%u{`Oz=4>Tp%to|<LpRyh{7S&3O%w<V-LfePunCdJo^?}B`O)Y7woXZ
zVUJvg%#;Fult2CJ{TuVACG8R?rC)8QUQihoM}d{D;tKwHRYvMe8QtbpwV&EH=d(-$
z>F@Sl!|NGqDbmNo1DflEb1ckh?BBX+^{A+Bj?;r^tUZkRsZp&_cU*sjs(Xj3vAK5?
z&)2s#?zvy^Jf3YeEZxkQVEl1d=GM(c#T{4vmx8NHjsG*@FyLCphWSyg2Yb7%_~_wk
ztnaBER1Xe_y|OeD;~IY}YnvW3;4PIQ)mboz`bF}yggbY-P%S--Lt3S<xX#|IzihG@
z?EmX9-()^MWd9e&XnrV6*7@s~?dRLOsb8<vJNF-7(9b`^nDO<G{?6Z3)FPr1WDZQJ
zh5WX&9*5wE05;b$*MoD0Gx!xm4?TK2Mv-=zSm|#k=7j=}#G4;d;k$?}2m9!C<)f17
zf&7;4w&-!%vhWEJ*+#D%2P&aGNRK!r`HnQ0Y_7_W1p=<!L5hm->5;E&c)tB%XWR4N
zdD_jTk<EGA!FJb%0EE}X<~ME`!ev#F6jrV-xcsgXm4CYBN4}_e)<UM+Zu)T=>E8`3
zLvdUNoC`l&F|rg%$f^&%v6GGiWq>k2Hr=#2WyG6izBJ>jCoZ4^Xz}Y2B7TZpe!W8S
zI4cGdV?b@p<V_pH4B)hc5q2%IR?DY13>uKa)K?EOPkRtvb=GPH0f0quKu8&kmLnnM
z5ERgd%Vw8yviO+cA>^P}>5XZFKB<i^cPOxVgkeCL5vlb?UvJ8XT0NkE1)scGkYN^i
zK~&jUOgP!aNe0T=IpAbMI~v-wt*vNs7t-rH1imSw_8c4=j@v>2r-VcZQ;0UmkPu-`
z*l_v{^#W{gFF4(!KA|nB0}MP6E|3exKbt%Ho&IA(kc0zZpI((C4@kQ5t8$Shn~5>*
zn;J@3QXe4xt<>E=AmT7;U=V^N(uUlVz==P=9<>;UDl5oRN=TtzhE1~2R((SxB3VlU
zs!uu5hgukr3imI_9bt-c(jQ!)PJGk}ovK3gJtt2<K=RHqkW2_5Dx><-HKXpp69^1_
zukxC)s&uG5ZR69`@?+x%JYK3yAg&OefE6};X2h!8oa%!a%jsP<WT~=|6QFFEouS8(
z^??hn88?g{`U4`%p;dT598%tDE82`Vr=37BDx+iGbTLV&Rii#WDWs*Yu_R;^NM?K}
z8#*-S1&wy*v!brFX=!q1laKy!?@ldx{H-T#DvKVcNF<p|QSOSA=_n;vP<8;K$g^w8
zc!aTkemhV$j9=H|cpjF6)8<>h_IKP2@7Y}MGVcssB<nY6WEk2$<H`_6ud6kHkoe<~
zY4Ah-Swr5mW>E&dvFIn~)3S`%^)BjbLpuB@?QStKPCMO_Ay>G|6itXN+t-=#rErEJ
zBI(O%W6C-ZNN>5@s>jO@!O34%(n~mzqZ7y!oHnjF+eN=FoIkRy_4cPT=*GaMFjsk$
zb}lM*I}^_TxMtSZ^X}$F+GYS!PkmEvyX)VTk(0(~QH0E+NYf7dDskG3xyCLe;5c$8
z4Gs@GVup7rbebHAM}MRtfz08`=sOGMbjy0BkLD9w(%<-mY}|iCpk(mK-SLWGvSs}V
zSUM68t6H7xWT($EZ3l8%lq0{KQ2tVW63YZ|kyUUiEA>zJe7PQk(WXL9@+Xi}mt85_
z&I`!LUgT!Su4nPe-^`s7Wek41<3<rnAA%FZDW)esG8xB4c{o+Y`)B|153v4P%x@T9
zzdlod=r3e`VaAt?*IrLTGJ1nsqCyA)q);T`lS45wI;!uB*cV>6v)4A$1%?1P0umaf
znmh<D5Gs5CMr%E5I|p7!-s^7hpbBevq(5raYD`X|GOXhD9U_76*Ac$LftL#o`v`|a
zRW?i^$|MwlW!-5)@St2A&;qxDaRcFFbadDfRYGXuShy!dHzYQg2Xnj;N7J9{o~nR+
zNK9ogo-DlN00s%nvToaIyvlwi{0u@9_!AO5oaW{ps#*<#E-Leq*xLs0ygV~OnEPZ#
zgbz_SPDP_>fCCxB!-`)S@+JP20nI&XzrFodeD?5htZi<&-REZK-Cp2isHKD9^6=q<
zc=Gs}a6;jjaz~N?!U_z6ciwr|_v1jIF7Nq8tHAEgj@3ePsLF@1RB!Mt(5PB%EN(2{
zFarivZ>Z`6rw1-jC&s-+s4a(az-f7i)RVRG_|Cg;$E{n-CZ0hQ-Pu}?y`4?9f%;1A
zwX-WYQTv0tB3*EPb`(z@J+?Z$p1z@Y<BhlC&fQl{xI~g^S~xWlCd`x&BhBnN5DtID
zfv<}|eGr$?L(o6(9f-k-;HO?UF$sd!fzmlJA@EqZpzz?sR#fG8^X5&hJ$;tK0(Z5P
znfi*uZ{VP2N}+;GeKjGN1Rp<D+W^1&Y70~kqPD6wF(Kg5Tr9YuA2<le%neZ(f+euH
z@@y@htgib4df%lrG47MxXfasd-Z4S3JOl;-5_N+RH3x@aIG_T3!25+^q%KqsR3JiN
zg*iJjH5s>W%$jhwwX+q^l@BTy4V!A)LyZ%dtS~G08k=UEK->pkzWds1aqrc8vHJ9>
z$J6ir{$Gomx9^(CIj(wFD<dAmND&IrHUpHt1NvEgIvPXV@6yy9z<s<L_Axaz7W1>y
zCd{FlZEttm1Wu$TTEgKiO`*D2b^4I@1Ov$=J?L;xA9rXC1o!@MUk+=0L$LF~`#+BV
z@}K`>tUZ6Gwiz+w2STvNtp28E=3{(fT5$Jj+`a#X({8B_d+ihTy<lB!%bh5wMT6*^
z`{XmVaequ`9J2E6?Ciz!wGA_vAtJNVA{n?iH|2L<|7HBiAOBG__x56Wc{$$v*4rkk
zGUt&NjAAO{RhhfyRL$+3Mr?0ws*UzyZFNQOR!mRN#?0(u3@9#h3G2V_!P1@);yTiW
z<N>jlxr_(qwu&#<(;On$pI=yv+1Xjwi}8ze2k^>%vcGRobIpmR^R{=Jv7tGGxeEpp
z60$J5py4o<H1()o*o&3(nT-5s5B>^lM0JDMJ*09BX<UtHPEa2RJ}eQW{z8Jz1ZvGW
z?0YmdX|C;4dykl*a-hBdgcw(*NoZb0;tR$A_mCs?U@1n0WzB;z3Ooziz29aI#IMuZ
zS37RTcB35~6~0e>IX$U9n$=!Ua}SI^q!wY)dcM~D!yc3Smr$uYtUj97yvBW6v^(WF
z*4%6fPR&OU-%kp4s4${fmw?}$t?gKSzNWeQRDCz4aZyR;l-jpXFx)>n5j}cC+^?ce
zO>i~br-~>~`VrzZ_nL2PZU<`bAkL!zxj;t0^7p|fAIHZJ9>(I$JA&Ptap&fe;Cj%~
zX-N2N2!3D?z4PvOd}p3AFl_a3m{X0+Ty}ZdiLLeLv9h+RrR^k!G^cXU7IYrwW(e;S
zQ!^SLlQA%&`AG54gwv|^Jclq=k|}wizSP>IenTxC(u<S@X<@>m&X4(C9c|i>C5^Pk
zvo5SXUpJgi)N5{U)*$E*+^fibq0!j0`aJH+nw>-CBXpEPyN&t)ItObXV|!?LRPBsZ
zBs8joJ@y3LJ*jR|nYm+%gUZmeP|>G3sMR!nz+IjkOaN+OexB(8#+lE!|BtzgaW|+u
zG>u1NZ{HH1d+N(3_o->zBgMu#=lQ;rRumkd78W`KI7Z(fchU{njv8a6MA7gAEsl0V
zH60SZl!5v)e<Ov&oR0J{`xGR!t<~uwndLBQQCZ5pj}w!1uWjHAW~ynvBs};7^gq@z
z4gzzC2B|I%nnU+9<6QkoTY$4bTR~bIe2#g^QmASh`j+{E@_XIaxM3dyd~=|I{7}iu
zIyI?2MMa>QeBe3iTgDA``kKQ*%%QA99LAw69L(SlH5o7_m4)h&Ofbr6hNf^hm~_zP
zIJ^zx9Cy>BlrC!_^iCE9;xk8acpwdjnlRKx^OWEzG+tnUF;@_QxsY}+^GodyLoMlh
z!vCOGstsU%`9S~Rh~@;1U9GhoN<xJreKy6xpUPldoV4SzvmbA~G9Q2ScfS>{+?=zz
z%#+K3*lM1{`c}(~sY{Kkt^EV7BRjFWv7`B4eI!7eo2tj{<!QBN)oUm70ds$9eLDBL
z#9fqo?W0)T;h==Zf?x&4S5-JM_l9i>-@3CrAB*$#ZrfR!7N(chW#%fq3L`hSja$dK
z$>)swKI8p5%r}`&cdzmn!d$JTS?8}`wx4hB_F~`VzV!?G`DYk2zW%WpUnG>jO4Vgf
zJI<g2-4A^NPz2zcw=yMPzKfJ_cTw<@JpqtGT7(5YGl0XV)xUZbfUM7~AA}rDUb3fi
zQ670Z7sB)H2>HUKZxI@IUQRm$6sb6PM!T{VDLXvnaypt+OaQa5LZw&nzMk$2Kl;sd
z)3<OX9nvdZv!eM@-;-u(ob0u4jQw%f!{o@@<-n8pwgE&qDlfwE#YCB|S8p0tCJCqh
zxBFF1OM_2lR{E7Bw$gOlO;5gME86c&^YWz@=N=(PXF|v>K^mE`^=3gd<Hd-y+*oY1
zh~ma6Ko%<yG&3=jCW{Gn7A462Yy-%O6RT|Es4yoyP??9gvM>x!=$y#O%fx%U;X2mB
zqg`E$s2meople#hiN}V?$alf`;2UNL3n~i|n>4a4%Vm*UJ)9(A0R*XmB8v$2zBYMV
zLLi^7tZLJ^>k0oGGiu7=(cxIurfxzDGG%6gwR#!V_1NOgTDVzEwGg|%sFx5GHm@KJ
z0HMd4+dzHf2cn9?sJ$%>rMF>9ZCOFh;fywX)iEQK3SmaARpYSmDjnXO>S0mNe$plY
z0t@2UN`w2ZK*HKW<T%Cac<M_KQ%>djLUer#LX=*gVImn{sV5s<%&4-YE+7h&nSSK7
zRYypYH+(8HhB4rcy4nHc&P+J<KSW&<uT&ldSKK_22tooZLnLs!P!@%ud@#3QW<fYX
z2RbL*$OSk7CI<zlTt3Q4G8k(R6qJNFd2GDSK%&%;zz$^rVP`V}(+<Ker({9$O5zfc
zhlP03ww$=7oj^({3q<f<!Cb9ci+Xi3hWgk*33g#JP;r?qaZ`PQ8{?8h%TZg7>$KC?
zIX$GCLEa`q<aef_H-UNU$<S%aIGA#t%wmvN5n3;KF;#Tuu`*tnR`;986iP0%QR$y)
zlxb&LMbb9iw2L=B*XxBX>mpYrV#bx;PN(o=OgN0)$SY&z#EmK9`|TL{O`&9UG@Bw`
z@$UAEUm5=@tX!1O^)g@gdm)|D?)1t$(qPLKN47GU{$6MgEUwGT`Q#dmOOZnQF_kHS
z3|*Kq?Xphx%Q#o&>Z|L+i~O~Lb$Vrd%5!a*ad>;lCO`KN5B@yBVNr+D-`;M+glU&q
z-|6QR9UfD*wNp?T{3abd^=(Fx4UwHd5grR`S*Fu9d^)3Sg{!toJH~Ns&$vadMOn(q
zD}0f(zvSl>m3A^MlN@#xyGT<N&hp`gk&OvX(2(o5j6BLP*(PauVSLFL3h>nUE6dU~
zd^;0a)JEjrP3xL*kgFe@GOYf}GM9OlaWgHxO8g_!`QQG<Zx~;{Dswen|NM<%>=!gY
z-{$FyO@c0Ph`lF{&U>Px<RB=Xo}xCV$I?b)qp0Re!Y20}zyyP61Tqd&A6$tv@(wLV
zW#UekTEowZHQ(#1(gT-73JgggPMSl+8m>Sj68<P$(I@Mb_kr?g3r7Oy;4l`4h`@7v
zAW66bI2wC_B-Rf29*D&EAtv<2tnf7O69o;JFA#4$ceosx&YAK9cZCpRyjSskhtMH}
z=c58Y5niNt_=AU#H}xfdPKeiP$^60~Wd_rUcH<r<2*=|U)H3(RDJOQ-E|uYd7$4;{
zwy+fy7s&w_d>l-J$nnW1AIFb>{9fF>doRB8o$o~FIPbM;?Q@C!X?*tC1FORV?_)A3
zaixVD3-Q~(`+ITo)*a!Os)ORRTYF|+aTg73ge1b`<cyg=7&V>x0%;Bic#U}S_;GA(
zY$#3CTB*Lmy{Z$_LSWU77m0V?TwaPhw{FDD%#`{*rK>n0-)rLUsmcYx>BM(n9H`D5
zjDoo`)PJEp!PEHo!Ak5l+g7&)<L<D%uYJT(yz{MZ#rMAdeKVPm$U$1{ppejlfNLmM
zd(pON%|K#`{sM?0($QC)qXWU*ah$Xpv9rD&)6;Wt|BW}}LgDBaGzGvIh3NvJ8XS~%
zLS@vL_CvJG9?ZY?kqM8SQilkPG-C}W77P`|WhDXs1C}6ave$wj%H3frYg@6tx#RHx
zq5sbEqH*k1xovLmSfYUkW+4PrfOwFxP_3$uQdJ{-+xj>c^w(oR_%zJM+SriQL2?HV
zjHUfnCt48J)ZW9rg5l1A4^o=xtA2U*=o7WU^Z4lF597)6wYYQlPAtyN#RngL7;nG(
zTk+Tb#{b!4jYD!st{ry{V^=UXIgKWZDXZ$U6pjWqgxS5uZtU#r2=3YE3|cB>ZUJ@g
zv)aE3mp|Awp}bz7)*g4rccUO#*AaZ4o-oc}8tbh+3dHTbeHf9as^_@j2WAjz&~~=g
zVsrIz?CosC(<e`2eeHQP_ZuGTQ!}&PpWL~9KmNww{O@D=&fVx}qrKVa_?{h98Cr@>
z{aDwyL~WkHMRhoj=?OHP3|qzBd++_!DjiXqDIik&A|atY2Dr!NQT)e$_pf4iV>MoT
z{q^|k|K0z^u>V+N1B^o%cpe+CPDD-c)iXRU_|&|mytg->$4`FzBTH?)e*cYl?ag;$
zaAe%lE~pdh>({>L^jKp=dpyAuT07V$9v!u#rE!476|hyGoHS7sm2F5P<=&2Y2pU44
ziT2zFME~>*R(v2~eG|q_!%Rx%VoTsEeeW$foWPz*<A6DXH3Nb_Cih;shS;C>Qhk^c
zxZ906Vti!KOxIU#&BsJl^O^brng!B+?Af^I32Bduek5&1Xd;h^0l_uQeCAS^yknZ{
zsk34${|mt`5?}1)$p74XM;HcY%Dd6n*IIL+y)}1k!JMrK_K`Z_vWDJhs=YgESNDnX
z_Muay!@2``e^iA;DyZDMdTr1eXJ(7$4Dedo62?)l#!E{uwxV(J;NerNZG{=IytHTu
zxI-8g!0siCPTH%->lf={O)zHlZj3_?604j@{K7DUIn&&0#Gn1}PqemP#_ZzlSiZ3s
zH|A#Ipt)<t*W%)$>PWfH&Gh2FG1k}T8+);}dtk}m(GkrLBWllHwU_EO)T_1m^w4)C
zo~bQR0Zdy0`^+KGkznqiHkX0JT~i|!!J(#`Gqw9b-$k(G4l12KwHbFpanRyiYq{^*
z)O-s3!0h0_$;Q@RJcFJin46oOQkgXl)vhpo=}Uo#N&y2)Fu;1PxFhOs>IF>EHmq|4
z`bLU!NNvR3hRm@r1vng?QeoU-hFaLXqJuddM(fl>Rc+9(vUlXiAp!2AOZ9}EW5F^^
z8rA?6XrIG5T3=WlIS2!DkT!t{vAy3`o2LKn9-4``zY6t!(6FFAFitrbd<K9gzMx5-
zHQ#g%r31qSIAhIaZe(nY*Xo}CftiWgxDNyPu1VFKH3B%`{z(psuWjzCv2wnrjgc5;
zK1KB*>(KVjcC<BSnpT_JP+9l9Ca@o{w6NeGDKVJJ&~jLZd|$lUjxuqe$8;-TQso53
zIVcS-1RW7vjPT&Uj0*;<RR(K#AYEqZI;G>QDb!}BBdMO?Z7}Airs^>@IpG5@;9Qi2
zJQUCeW1tCwPtoQ+5F%LQx)V#A3MLJg^fk;p)<ETh6fkw9F5cJCu!0q!in-Q1Bnn7*
zd-Tw#m)JG7!6#J)#y@?L>uU$Vbln7en#Tl3eM*<PPIVm-3{BUk+&`@UFgKAt=CJmu
zaEMzA_4x09|6B3){W;@zeFG!WXrISMg99Dv)7}B!r?-LB*VevaYJD5kzlTb<5;rtH
zxC4@QV_dUVZ|$@^w=d04tE?lj+diOHF{-jJPIF&q-S>}vs&!^`xHn#V^_JmgZFR$U
zzMTs<=RiK=E%&|IKKa(KUL_47U9K7Y+@4?G{G#zNzf9aOGQVMb5r;m`e*Sq)GQR%D
zzvoyb=yVIs?$eI>Hshxo&JOw@oE%D%p82B!`Lpyd{7lU8QRBi1Sd~Wx<DK#H%gLls
zLjKNJwr2=lBv$yp7%BgYLJ*9f>nU_!Bw-I6GDP}a3$360B5*-Qq+CVgCsB;ByyfOd
zSY>hM`7&uL3zI?k)GI?^aHpL4#$@R0?`|%IDP3fv?Tdg}H2LAB2{QliW8%m#l+SUx
zsg)72UfHjH=`?_pv`W`9ckEgq`BGe}&gphE?)I>dXv1=>MU>McK4~am=X3{~X-+lz
z5`0b0XIivDTtGNEslbAfj4v%vTCg~Q$fC=J)Jz-dMp^Z+fU%ha88}iLkWena^rEWp
zauAYn6z4+VdSxGBvDQM(Nk4Q5dj;2m0^<YtBw2_JEUat}Ss<t{(owV<_aK0juz_L0
z;)EoNBzH6I?d`|9HqRRyn?4N#1LYetD(9e(pv6~iX#=JWy2^(xUth?nHw!V6UN8XZ
zQhJO<)prrFpsS)FW-KU%1rP<k>2LbdNP!+j&c#?r%-C?uYcpG|*Nu1~6-T>)pwhoA
zqE-<g<c`y8M54_UHS;4~Jur}ra2<D49wAcdGYXmaH3*qd1p1$0LmPAAhdaJF4Gl4Y
zG3b0{r;kB44Lf>BgA6%g#HRHOBuMEK9^}BMWCSzxHOL0AhN?^4OdwL7LBvd0P(mK>
zDl>l|gD^fZgzr<Gk?;cffoM&bA;BxBZxqF4=aOnpJG&jIzuqt;*$9&~Og`Lz`x7B9
z)SGeAuLAmHlQyLwUZtdAZV-D7TeVRkV%n&Wu<3&FrTi(k%c?G>UKoG-UH;}9<1gj}
z{ZZyJG`*G~UNjlkmaCVOE7GJ=q@Zud*XBVrzH)PfA`dd8T?#wlDT@>2Td!<xEI2tN
z4*t$p`tj)$iadyyp=Hr`k!qR@=AxGI0egEC<xeADz8$$tp@@tryu5@vp1!}JjM=cI
zoBRbU*)A#_UUuh)8nEe*V;MIy@4nI^Yte4mvY4X%RJNihrAzVqqUpN1p^7ktJ<HES
z4tBybUG>AzD5eA(+%bgfHdEIXgJX%ZQob@~Y0}5W<yXnOK{6e?b;Dhir3{wsp%8XR
zJ9AgUT!wNS{A9tN#xD9xw{pDOhct`6!}%&hg(q$M`14u{rIDR{?aZ5R*~@QzP(G5s
z>O+F1A1238)|$LCdis%uTU7sLj4};E8g7fwthbpbAEmu0KV`Tcj#P@cX=E>)$~xLR
zeTq>0sum8XP)f(C7U@t<hs)@;kYiz$TY9+6vX;J%tz^L=|Jfh>3n%gGFu!4Z{rZeS
z?gew*;qzUdkX6`=dUU^w_sic`23RCsK1g|Rj|W7o6L}m77h!YH{Ui`2`V<H?Y&GFD
zT&&N%9uSy0?acb^d!dvLxB$2WxTgsiT70P&_!ES<l%6MCukcwtTEmfUfZ#mL30xDh
zdbIX)LN=vwh{L*XRfN=9X^|(mCvAWVE~^4m**NqC@ei#8gMF7VJ2@V6(^Do~LR5t*
zldS<h4e_<t(&hVmElZR^gevb^g+R5p&;1#w$4MeCWk6a3)rM9x4Q{RWfUr!QR`bwA
zULo``ULA{x@rosPDnh?|)s7HCX^@5KdeoE-#CS@xy}1*Q9zBf5pM5Gk?9?io9zS|)
zNu}?9|NEYMQE>*t0(C{3!mE!v2g2P)eVB-Q!8mv{syrtrRo~I7GE^U{tx!)?tBw24
znOE-IjpduSVtgW1g`w8yMEEDF@jm+K!}#>$kK@tf$7;W06D?7jb?=RD$18X5tBw`*
zc}H!46iCHt1R=^GT>(>PNH{pS8T|u(#@w}|a_u%-YX6E)lMfB{nAo+|Xve<V_PEmy
z?)8D$F{1eMGjqb>Q8$-N(4C##SX0};oM8)bbPBPK{1Ud$)mErr1V$MHq}|-#GgIN>
zn3M76ap#rSqCP!mcpn=di`sat0Q|(YX@3Z&9l;!u4ONK0f(wYGs5F9shZ?_@_7lJ&
zxIF|!q*gdNKCZSPe@<otyU{mfBL8M%&yp@2Hk?-7A$FsF3w3oc^nhPf+d+IKuVeOD
z0GetBp%wL0Rt2iI1OFWugh1aHlj9?LM$Is!P4<q?EcFE;PeGzbaDWsFs+tBaPomw}
ziVr^cY5d_I{})RHRmaCQejqen#9#gW|3l2pEvhUX?JbYvgP*<^Po6xEK_n$cxKn1#
z#A)z)GYAF0jM*9W6GT_kWzEme#r@Y_k16eiP6U&vQlnkDo9PU~q~gOcU|gVPQacdW
zv)NV~sc(@Y$pf9|9%EC2)yev}U?!y@Npo}kc|82|gLv}rQ>*ifIMW!c#=F1uo%qi8
ze>-YZi*easRX&VcBn&US7h+z3$%(2$(pIy?QF|}`<iGr3{NaE64<^2jR_oEP`i@p7
z4BM!r+ueE=>noqd`B5{fwaK`5{|z%~o;_PpIxrPA2MI35G|pxhZ$xc&UhbMVw8z@t
z-HE49ADLnFTkn1+zW0OQkBR!c%9aGyV~tm<j1$}v2jc7WR9(YW^RHm`RDHponjOpY
z=g(D_^Z3CJeh@d8Z<@hLTO!S}wG9(;S7{7c3QonddQj#V%7(TMq-tP90#CpY;N^q0
zY9E+D?8|6((-ia#JVW?KjU^IA>?L8yGsYks-?}juZ@zZhjIR-;3p3vm34&whrIz;F
z6PkY}>M+W+A01$vgb#NzKXR8Cho51tTnIK!&QUd|wE;k;92$+o?z2oEXX?s+7K4gO
z7-WQF-Z)izrP?&6i++ujVH4LmILU4v2{jll$$a2mHKgsTUYl6621hu|t2tP0!aa*f
zT+&ZS&oOq{uc{&%^GFVj_--vEId`>QaOVz0co?=wjvT7pk)oNHLVewWrDv;^VeLr;
za|cRqYtwZ<Q+hDiDcAnquJ`HFGgD@$P(PT7Fm|8~pr(*}rL+_G+QVE^t0FNn6oVjy
zXR5o#<?*G;t+oPIVBC&rJ*ADT4o~YeD(eO{mfM{(Gq6Z^M3V(`$xMCJl55P}v@By2
zn#Q@}aHyAg81;n@9zBnz&z=ij&SGI^M)0_xb)=>`Us%cun5A6Yd&t35RPya>J>-pK
z^C;3~nq%m{%GkKuiTm0Pn)_NKn(=&nCmudolU-}K=02nxxkC>64YZnBwb6~m8MOiP
zZ*Qu{9OghQOg?a!q$Q~N1Y0mw!C#Q%g>erq-ls7lc<r1<Lu=2L%DkcZXKQQ6YwPjJ
zk>UbKiNnx8&~;$`u+I2!iE$W#jM@p+u^f_sNm;3IrzVHadxdXI#LU!$54<3uoeVAR
zx)g4uGSMHKJ9|o34k|2bC{k3cJssgZ927@#6iH*ojn(8)7RB||69sJs8X$NFw4Z*Z
zuL?rq^Z3?>%S?F29mez?^`#k9@+Ula7;6^?Zh&3ZOBmR!rOfw^syL}h3VewoGczG^
zhSc47%*@THujW*Se$Q*DDrRki&IJ97J36N)C(U#M7X+tl3RZSFM6tbNh8VDl%H@u5
z5lj0Cj(xaE;S*OJ%<@BzQh&qf%6cOa$^B!(+h}X~Ax*cpr?IQK7QE?PINuqxRMnF*
zSqqo!9HXhJ{M5nP0F);Od1_i)QAY_w`0ZC0;@kIT;`YM0>Npjh^8p{Q0EXyS=B3s3
z?O0h?UG|S+zjdm)gY~TA`zz;!=fR+5|JOb|i_Kl;(oT$P+(F-iF2`De3PjdD)Go5-
zH``6Ei(9gf2`8HO0SE4?MPmRw2~5&gz=G&YJIT<-+28h2n660&_rX`1Ytb+-r1Le+
zFBA8R%-2cpSDW(}nXAV9e6hYP_cBiUfB7Etx;yh%VU@`!Ul#9+4UDh8x8M5qKXC&l
z43Vr7Rze`IawiY`H8mKxGE*V*8e|#IAu_B05hvn14v+jDqWA{dAmpSy3ExEs|K?kN
z3S+{j!|_i6i*)%T-So<DS%gFQJ3R`2-VRe3ml3~f4`~yZf#i4TN4%>y7XP{Yq-9qb
zN0L3?Y8tmo`At~Yvv}OJ{5g%Z+hQAkxd|!1(_HyGO~0`cQeW5Iir`K^t!_(C^W|jW
z_ESz2hHv5*kAC>?e!GjiY4caU9mrvAiZ}HpeLP6sL|bj{&G^zr8C8ccMJ%~Oe8pj*
zWfD$F7%eD*Xig(A7)oYDmN_MMki2122(ko|gmfq<8!}X3fZSt1VpyzTQo$r)!5dHq
zRoJv?aTu0=Mei|AYsd~y)=VHZ6APnHdTKFAy}Q6pp6JDx;X=S`2E>AegT)dA1HwZ?
zNO|Xw4PL=u`lbwRt)iU?k>KPR$o$y}cNB1sfsjURD$h=Yun55hVMrp3B6#-T5dtNI
zjSI;^fVpSOhC1mFsN~Q6t;evTMSrY1=Eauu@?z0E=}ez7-mE@H$TIB+!h~LRE(m50
z85UxP+7^fvY^vG#K|Jt@V%lC03!){i)Q3Kmqs(3QVuJw51eh=~qco|GCsB<JhX*d?
z5Aw$88|tH=4x4=p<27NY93PHOPWlG<hPXqwClgF?Kp1{x>@K^J4#5@aQa0e2G67el
z!RaYBc@O|dJsF9d3Zox@d9^2XSD85-=1Yx1R+T610AY!|LA;E#DNi=FzWYmsfMGa<
zI;J6*j6>eQAq`U0gre$KZ(DQRdc4c7CQ7@%i<hm}!xi7WXdgFK`nyGxUm4DB1n^rw
zdi=tHH6U#n$Ki9XNIUWHS6>$Eh5gC^WG;J=Lg7}t9YlZ5j4_KFYnNWQ*)Q{C=sFDr
z$~=oQJ4}Z6yZE*#EOv90tTV3pzFv=XbNIp!$L^Y#F0zzsq5~H0MOfx{XH>YkLNcV`
z!xwm6Ul-Hcc6pZVXMcwwd{M?Sj>Fm~-~KP&MVP|H#<?n8+Ef9$Me4@oyLgxBWMG*x
zJ<8>j@YF9(SQRki7Y_c<Hoh`h@4_?v^q=Lzy+}Ko%|Bf~`&0kxd1iZb{S~qF%e2cf
z)9m_@CF`7ax+tysA*?*R{tj<I>Ds%@s{~5-kK0o(L2M?|{mVYYCF_8jXJNbXdr=@h
z{wmW3KFHVbkn!D@xI1h%j$y}n7Je#JITq|zHjKma@Ts@c&b+$e`Eh=PA(&iTHit@o
zyJrBu3J2_N2HUVCz~q`+hAm7H?$ZYk{+Gh#8#llAH~#iNDXhOS-?S+}zz?wNfs&?#
zozfUTG<;u~D_`J8kaF!QO~%o;;0rW`J2ToxXR)()7#od)*l8a}tE2TALcDM&2m(AX
zp*p^NA7UK$yl~fRo-pR13aZaYn_+<3H_C^EIs_~57~r;Kkh89XlW>R&1`E&OAzA}a
z4=5bz02M^wM90Pz82{iH;71%HLQQny7T~8G;z3#lV(F-vC*}P)Ft&`psyxIS=(`jS
zEL;Ix0{lnqB5xCKQ~grnSunLw>jlwm*bKQLtAK*pgVYU78cW_xRrQPq_lbd9Lrexw
zg3*MO36cg7X5-{21}{%zw4XaA&Rhmm8}9FH#>XFhXrkfz`etlzZ3$=Iid#3A<L>P{
zF*`l$Q`xj3_waH^4WbuPV+RNAn46i2w_ble>J#<g9+@ME!|G2iyx&+~jh(es6Rc6Q
z_2wII#Or$IW@jwj0Dc97sL|Mp_0`pQ|0nNRf(+?8h<JDJ-H&g-`&;qO_kKHW-F-dA
zC#Pb3yc*ToxN%o7W9rd&$vs!#f8120+0g{JRIdfI$|sc7Hts*Q3$TO)f{JUr4haL)
zb4}0A#YBBdfIDD9;{{S<dMpWd2!o9K(%Nxxa;UwF`d@9kz4k2L`{5tOpZ?K*h!1}J
z$MNvv_u{jU^nCb}Sb6lR>h)B8dJ)xXEzp$EIc%w{YNB)CSNNz3VVDw*$vqpK-aZiQ
zu{PAkMzvS!QypRW4#h-m)J!`tD%J~#^{6`nevnv%Agzd2Z-*oSevB>lOqP_|gRq)}
z;d*`C_j657ad*?W8HU^ig9ZuS;L^00RaLRq5*+tyF9A~)JRbGs5ITpnpJa+S)cD<P
z9>&_%ZfxxA$DZl{e%-JBnW|L<uP}25V)@36m{B`Jbe@`-^VsDcpH21u(<hJO*~-({
z+*mix&s{Z0{V`6KmKI}Sem-VZzUj$&+`V%v-Vof~y15ifb5p*{XI%B_>*4OQu;kv#
z(?{{)`|rh5^@D~`%&V?*dZs4FO&CY|4|QUexD$MH*Dp+Tm^Pej-`1FD?5iGvrO6rP
zzjPxO=Wkf8C3{o$7_D|QR@YYI*^|#=_4zZ82{S^}*T<cf_8qA7WRH;&K_eR15IdWD
zYA3-CcQT!yDG!B1UFnUbB{T6*P1@eyj^^%`VFdMRn`(>3_NFDZP=$$lM$$SzJ2qTy
zZ#*~5Z9IP*M~6+#d1sn?Jk5IEg8<!b?+G@vC)RvvJdFK|_R8!9P~SPPHb%`S<9S;B
z19RFEh(o9$s|m)&)&G})I?sms8x?>MQxA_$HFxZq>8t6mU`q9>o?GgSPV`<J#kty}
z@A5?BMlh>6=uGWo=CEKAJ9X^2IEx9*YYWrWn41`mneoAx92<<f+~$!+&|EjEzC$JJ
zrS`hudfbNw;T{5XA9pSf49BF#3QYYI^(l7|_FkOC*l=IW)GEF|4_LfFEuHq++NqiW
z)odQ9zguR=q`JzK++nTR<L)G_OZ#o^r9L%tfc1!zKpZL<Q@=8oz-UDp89(|B#zIGP
z(1FTv**BzxX+#Ac)x<O6Z)g@8p<ZWpH21EpuPG0~kk-Y4iePkVCi*Mmad_Grt)ok;
z)7;nCYH9vzYyIP{OWY2QFZ8W+E_&ij@OrMaVZb(3Zq(H=&o=k=wbxfytDRu%ji_Ha
zSb~bm{*h{2YTpV|&HFIT&09MSm1je9P;cB^m{xlZ3yxH87?6S`nBEZ1Ex`k`_CjO!
z=sfmXxrPD<Fy6QeyH>ARf^$%7h?*lbU8v2ed#@$as<m1S3;tPYPlObJ*E0XGW<$3?
zZ7EW_NPoh(92M-tJRGfzc`U-rU0GR;XU{j>CMRb-Zj;^iY3wv<rwgl1WuD6e83UT5
zj@_^9^}!F2E^jN&)?Ozzw)f4Lg4r=UHRbuJrG7yTB(UQ3L=!K0Gb3Y1wJA@Z=5FqW
zV|}PqIH<#QAJD{Lc%5ioLp7mbUhvNS`Ou1xNJl~q>Blp{jI{!3d@|QVOMs4sq&Wvc
zkkH~z_qx`qy4J+WI`=Y;3Z^wbD{r#8u`n4o7iWDq0r=z~3kPeSZ@|3RHT=M=W6hYT
zX&ulSfa*JN9p*sjXd_xrStk}3W@BMt&UesFsGp!IVXS(PgBz?z`s3d;JEiMG65u}?
zpx{5?ISv3$rg*@Q4}Gi7*g5!;8g`Vo+<-yGGmM>-s3u()c*$789Y#4!m{di~U1W!X
z^LFes)OVV{VQlg2qFz#CYj>|9e1L=etnDz`6E9=^g2n|dR~u(t9CseTpy^Y3P`f&a
zdcqO4r^Zcnd?Lox7PJQkdN^RjgSunZD`;svB#!#rQ_b%j>M-U@GteCJqrJg{U><T+
zOiEs(_7n|R!)E>>Mc5L|93Jd>okwDlLo2>BR&56^g48E>;Bwf)`!=;X%zAKI>fvi-
z)HkfRhr%(yfqW07AWYM_*D{YcJ<Jz9nxkj5PI730JJb%D_mmb)Q|Q)7JEjt8s8?yg
zU8tYyNM-eOzj5N{dW0`A*Y^xxWM1}vnR)Y7@qd>2W#WF$D6L;%YM(QdwS2M(uG%51
z@%h}#wkqp+J&hM1d;iMJzxy};^ndC3pZ?z8Pr6DD2mQ0X;)_X-!r`pU?T1G(!LYN#
z3n#l|e8QVpXBXbs2xW_ZeELzgBBivGhk_?j2$R1>+hiEt1S-p8hvJo<`heJ8G`>D@
z!s1sMfsZ15a)lM%0kC5Vce%NYvf@vT3QG}8e%Q-s7<)UUgXR0`w@9(;pJm0Z>(Vuj
zUpmR#S?iqyDs5cso~=wUfTWuQ;r(_f+aao5_1m?ug(=g)l%+Hr;Og`;1Yy!v_Hbbe
z52xXVQ5p5p-Y+itHBo6((62@ALK|*QZ2JU;{w><e0>oxki!LWMC}u7`oY=VV=>R5E
zm=l~XV9^;6;znjL%TPB*`@nEO%$kcVX;<!Eq6Vae#0>oknlPkA3=FJCll-6-uMsWi
zFjh)t1ml232;`JaA!WtSmE)xXGMO`3n~9`!C_AUDP_Mu#CF-=bxvh=#L!U@P-JmZ;
z{w59OQ+YX!hiU~70GLXsmbeh&hIATj=t+l%@Ots4ebL=a+a3ui;iM0M#v(%>Fg9tU
zu91CE2-T!FC*vlzaw^!T%Df5CW<ZNpRhtzyF)XYQd0=p{;GwU6OdD&MLJ%4_Eyqbg
z!ciU=MdS?w6mS|)osex}p@ql<vPR*JIB4bM{xA9p2;~nIgP3E8OQ_dSU$M-3m6ahL
z`8aX8;hKc941(Q+Aa-tQt7GNsyT25Vv6iEjUm+&i>5@0ni_C6rUxkDCQvoTLADb%5
z4U$W?K0&37umeZJHH_q-B9YUOoU-MlH|?2aOC+eDQ*1>&ed1PSAsid5+E^`SCT3zx
zn}~jGKD6iW)+;N^m%G&^x8=Iwvk2xEDZa_e?s&;9tzFL|ysUm>D^0dg5!L3(w}@1l
zBAtwbxt4I+Y_8I}N*g~{g?<ryXBgRKNq0LapyJpq9g0^s4gaPk-(@@*w-<AO?P<z*
zt~VX%F#0b703miG0A74zHT)Dt7Ow#c<2+=_cvrue^h>vVJHd3e$tj9Dmgx|VwB<@r
zJ&a@Wfyr+;?Zz)HSqk~cT-owfq)l8P-SLYOmOt`_;eo#mNfgp-{?w~%cMM4sk!2jY
zOy;_4*M8mLmPrw($fKKfy7}!FWFJaJ{faVUPm6<H#m!HCN~iQzd^y;y%+K!nbbuo5
zbS^ydMZPXiQ9leXt12|?l;#Sivkf&*DZ$LG^vPc*B|j(I^~-lzN2ifdx+Z<~ro5#e
zX%zwawSQ^hWqXET{A8XO<&e?zon^$|?OeolBn<K8M_Mjd38(%p!W(wv;JjY+F9Nt+
z|Jxt@^D^~s-26u3>(9whu4`p8j(g46y9*_0UI?M^e3x#d<VOySEUPBe$KmmL>>nJN
zka~P_krFFfZ^03KfJA=~jyb^#!-e!9z;oB8rOt(GAQ8i6o|IDxn^U&u#bKb+%$V{p
zoaM(S1pM+I2k?3rPI=dsa3_d63Qss<z$gQ!APkHx!ax8r&Z7ZhHOzykzET{oQxFI6
zQ(2i`AVSL>!W1X-2ZvCV%$*$&wnBIqQt84|gd>8#4hUz5=s?|2jR(;IwTlpIEKMPN
z2DMYDrQvk2ORCKJdWFBJjbIw=H}`zc%lT0!7N;j;VX78mm6V*JFO^h${L#np+2bcM
zJHHeoqj_R{WMm*-yLUIng(E)r^g;aO!w=lo^;*q)0n{Kt@Zyja_tGpbEc&$iPu_n&
z9zJ~N!*86j_T}^nOIxCqU~+cWgi!FZiHQjl@K`j7`&4Ckq6dQPtvh#O`PS{2Us{f_
z@tWrRem^L%;w~W$wQ(2d$<cuc=175XPuuLwq*XmaFrvFSb=_*`CH|;U%(hpSDtoFG
zJThS^nUy7;h$IXIc){8EaWq@pD+a+Bm795>ivFeVxyx*Ob3Oil|Nh^^BUC=J7YR;u
zlZo(FncklB_}1I+#1DS&uc<DR$^(YzkflPv^@oQ>VnqE6VGuQ8yDg0YwK0sDrTJOo
z3EZay9?a?Xz5Tsd--K~->2X7Q!1P?4n~s}{GiCrZcDXxh+hgkR1f2gw?a(oE7kqXK
z$uCqyPEQ#Q>C>j_XoC1~#l*8_8$z4#eBsHAG4>L`=$ZBi5b4>|jc5!^R4aix^fm<f
zHY%i!)CSz0GZ5n=Jy9QuxH(&oVKvFoK`W3Jd-7~0K7R1XL^c?h$J)0%`Rub;TU)Ut
z6??#i#YIcT%+4Xn^=5qg?RR5pYT9EO?GZ>5F_x*%VQW9Ox3?63O)%GTny3TgPMS|X
z`6Qkyzoo^c_?_SRLHyQtzZ;9ow>2oaXQ;=0`|#0IjT7$UIaNJSZK-;gz@2yDG2hJ3
zO~ul}jPUz0!N!SC!mmDm96$NfKh_@eW9@M|p0h^BCSq9Qh`X0CsNkwkP8-IMa^p^-
z@#?t7g!ahV8^L6L_Vkg<F*kohFn?2hQqkD#iM5r_;*bC64`Xv<4WLzOf?>^h5L|EF
zxEVFQkr1H`m=D?iAq~cSLVGPO-(;KZ`GoPpT)4KjYN@wg?M;U`_^AG_s?0aJL+s5r
zV|n?8=Pj7k^fCKTh}R|pD{rLlPSh@q=0R-iG}UfRO9LS>0Sv*YN6G|An09+#WBQ@`
zy{&Plv8O#Ek_2NKtElLL>0*fr6%sg8qq*N8uup4G%l%LC6uiM`Ik!{`glX=9;@|)p
zV$d!#HacRdD%$Kq<K$8gcWS|efbp}jv1R5l2ML&anVVUA4z<>xo(_Vs2AB4I=h1@b
zucn6Kn`#*ek93&cY6E(Kc~)5~70m-mcS!aDU{PU^&V}}XlnqQh)Z=lT0b^=dZHH7J
z3>d~biP5&K4M?`NG+%FTZwR)J&D^O^X)if38A!P9754(dq=jj6U{$u-r?Td0zJUqN
zJ{-x8#f3RDDPcnTo^hD=8hc38aR21I{Hx<*Y6JG~c@V%10ho^C!|~ef`KW09`QVdB
zv9i7scW*AlTd%*OJo-GBG}~>D^FGaEz#xq60kuDjsomC5?Cu?y!C!^ZI9c-?!+Hh`
z!HhypTVwy&V~j(ms6<5_Xmt#=yINP359*9*8|V;77oy1rDVeh~!Lhy>4weoa;+lmp
zlkVZ;XYus;Mw}?kan+MU3_~NM>L{eFwEk)Z8PmMPooltRiWxkcTia%!BJBcfGRN&V
z+wthx^H|^BGZW;^S8jQp<A6q6>lxz%2E^<%%%{574&aKlngcAXFP2_XfBBwzr3qE-
zNN~_RU~NuGThiJUyr72l<n+>viiRcmG#4?CvN#DmxJm*h5DcUd<?BN*W20(UO|F7F
zu6DCT+Nst-)eF_eFsml&lNw{Ev4un#2f#H)vo>)sm$Agb2UG@f=pu0?4gejwJz-jz
z{-+PfW-uS=EAT8NnYOmKyHc`@VXq6|X2vUoM=2jABbbG@C>Yh636m3;W4@wa!N16t
zFy&sDH{nT#`36QB<BjoZrWyz7HP>0?wAuxR7->KQENPmgVYP+IlX7vW|6FT1BN+xH
zFvDT+QO(mZf7ug(4<ISV!SnJCMx;l<rC`K@Pk=X?sV6*-^<Pb&^gZFG<Vl$+0(3^&
z-Aqh9Xi@@KqTU$75Efb>?Sh>{J1}y&PDHI93p3L`i~}CR0jI%>P7GbN;`{Hu7Vo_K
zj+yU|pKipq<^|fKhSsMT(04mFP<hyB$I3=Su+-EX*&jFNwRTNSc#hcF%Y!l8iN;(+
z!^5PW5bm@%KP|g(XyGx~VfrHZzpgnHe2K9{J5c89#{0w4yx{jzSlMYeO^aV?Ue4oX
z1I~Wk7>2HyuZsV(%r6u7WkXuOy3{^zt`{}i;Hn+69-q%W_bB+&8$Zji-~U;i_4R7@
zPyXNk&z_$D+u!>K2D>s@C(&&r29bQ|T)9!aE8x0nNZi=ukr1v9bFG{P4twH=9e+X<
zY4GDH`XLw%l~z8m`pId`t`6fb+Rnjq&||WZ%5c}h+C_%ET^7}du%!o9r;%mzx9u`=
zcO9`8{Y2@zX<jpgDO1T%nCq_TlBwm!wDiw5OhY}oImkQbhwO)RR$3f(z^ILjqNiWc
zzOElm@_Zp}J-C(Mc2NN5QG8;}RIWm0Ow4sRM=iS7Kj|hbM8ZoU9Orua(B&_r{X%6m
zVP6w+N*<{adPoE57bfyE79vfMYy=6zsT8NlwoZ#VrvX6<skGtC8$Hqza?2)3lQ@e#
zlQ?dic!0sxqXhxxN)K+bqf;BnFPLJSpruWc$#dZa&CC<3Lq38v778!OS`bkg2-D4r
zhtfk78w4LWbnCW2%r#j*d-~L(<UW1z>oB?Bb(vjm?Cv)8CgX*U@wbdUWmHYzJDEWR
zWXkTLtMDKlhDRrvku0&LJ{_N~$5?GbZPIU~1!N2LJ>J-8Lx9>Dt2(6J=mSNkjT2Ew
z0LHFgBbxL(8)P=O^gl!-Hn!GF?2UgSED&sxp+mj0LWFhURZ%lC^#`LAgsGH5!<&<0
zAT^Yghe&*9I-I2LQ{VS14-h3_61d~kcfu^pIS5e9*&zJrOC}Fs*aR`kE7)Vx!O3qE
zr}W3BfqTo6!CRd8Q5`txN!dUkklNyOESp4&^D7>VJI0AmAE|7lPn-BOkotvt10cwa
zj*Mx;HWL%tNWwrffm(fT;+@L&;%HV~DX<-~t$<^@+ms>ccf#pWUJjw3(ih`o$O~WW
zIC^fzRs=3179aU{y*wEu=RHh@B~{-1;%2|hu>30AinQ~?_QJ)V8AWlNkbcTK`HKKN
zILVE0dNZBnTH!$Xk{@9RPHf^jzP`FaWufGWYG<e{fPM8x9<ox0A{~tF-H#$jTDxIN
zg9=&KNA_&RA}z}15HI-H&M)gI`2)wLA(ZfXrB~*rUlJ?Y+5Uxl*8nS&Uw&oU*fO^w
zZGzy}O|}S6x$$#l9Z0Wq$Id6lQHpohjW9Njq~D@!SJ7WGnU~*1+jyK4UXQA9Ww2+m
zgqD>6>E?VjRZ!#?n1@9I*>*B^!vVL#&bwRg!gRwaCA?f1oQr(*_S)jFB9RwNv2n3Y
zrto$$PP=HCYtb|V7Vs$34PRRFRhqIKybIv)cQLTjUzzcC7`w`!;Ii|$dlfv}AU9Yp
zu{`bPz@$wez3qX!jQR-tq3qN$ZNK#X|Ng;0D?@(c<~NM5KPN-Eeu1);fC0p^?ip@m
zfIZwSQLXjcL=EAQ2S|H052FnsPirj~yn~bYG(I?k*F}mVTmT$EWn%3G@8q5pv=5vK
zKY(!KdtQ{TwGL>G1t&u7HU7k73_w5vc_3BReWdVvgx`RJ0O#NVT$JxSYnic?uhsl0
z4NS>wQ})WIKPtk*AlOnTi1Of9#fWhkP)A2QQYXrRKZklyrG;b{r-UJ98Glkx;z0mJ
zHB;yCAR3K^aTLnT>1Xg0Gy)(AHa<2I)yj|wy{HE|>Kw%S+In#J62wCAuQ9dHD>s%b
zC9=7_A-7ZEjurRCs=~rlxpnJyRH`+r?}6yZJy{AEj~+dWhtHl_#TIxy2gpWMR|snm
zpt)-a;t=iK+H0BtQ5_wR>cqGiU)(iDotwhRIlM#~dCGhuiQtuCGc11cllM(n`PSRt
zj(5KEotRr(i2k99`VYwp)Iau_vGRO#-^?o1rysUk(cIs4+oCFcaklRJf2Nc-xE;iy
zy%u->oI0N|we^s?l*3%~Q>zVO>nujp79%60o{N2F)!x3s2=`OlefZIb@!=;QnISkf
zR&zg79~gQI3wiI()5nkE@Sy4X06QEkYaen5ucLN6Gl6mG=IvNmx)tNKDYXGA%V4x|
zVqGvnUd+!0Y_S&_8t8F5ad#}tJBA@}hD!2ht2?pTIJ9I5cPz2DA?)j~F2~KqIZIbT
zIE9hC*Xjs<_N|7J_TZi}4S=W?r~E5|9ZtY=AS_R-54gQeY-YcvvO(m_6ZXAU<pt3e
zb%7lCQnzTVu;1f^JirUStyj-!bXq&ncX=AG-M$$UmC@+CIEm50h@ZayLHwsb{^O|6
z%*MTY_hV^s(Q3tEApOao{84=H@rRa9gLwbj-}`?2)^~n8UcLXiPu6d4ZpFigj|5k3
zl}&g&`211FYVB79N9+M1#*dCweLxL0eMk{9j&7*G{*V8szaIb1U;S$_H9K#@Ju3Yk
zeDZ0$|KTU<6I2G?jK0B29G+gP4^iK_AG>?I>YEF-S1sn}rlZPf`~j`x=g0B<(Z})Q
zKl|f&^!PJNW1(s;XF7T_=dHf}l%9ie!2Wo&Iv$hLvuf{AGqJZf*Gwq)Jv3klYD0*g
zFioF4dgz0Rn4#fO-@P<9H)koVUCj%tYpd!H;#N(}-Pvgf9**M9EBE5|-Fuo>szw!I
zU~R6i#s}~JBpO>APLF*w%<daEZpFLrz8gRI-QSC)<vW^R&LZ#f+l!OKWSE>@_7v&^
zCoGM$*X$GoRKeNqUTkk|doRU(GTfK5+t~EIbccuB8Q5`aFb_f(fB^=>#nU=v6L3Sk
zW=x=J4JnGT%9#C7P0y79zHeG_VYDNmGB%8M4b;3bZ(eGS8xXvtB+Phy(#&<r!}{aH
z4{DQM&BfefjS4ZQWat>D=bES0H_V5<zysw``y&zLLkNQRQsRU@g_zF%m$3`R#K2^J
zVI5(&h-yb9Z#iJl2Ps-@%>EE&2kLvzv@hk(!L_X&!Pl|x*cnw{x~?Z3!znNc($+sX
zEEv}MsPdkx{8bGe<}vonZ5Y!U&&>JMg~Ji;X4`FiYk4V_m*$lQs#8-k8Y$AfeN;V0
zOixY5^z@|rhJEy;+6_h}_q&Z%`s2HA-HLj3D1P|f!}##Qst;GZ^VU7zD@eU*7bI!+
z4o=)wtUvXMG1X<zelWg}kR9yrae0~ZA-=O`J~+CF=UdI#*lwFyPP?(+ub@^^Yv8En
z=&I&97>697hR#5no~bUZ>qzNQpN={TI)(3x*L-Zod+R7+QuRF2cfZCR48){0oG5%J
zxLY3e!!uJzjExyL=x7aqiLx|15!03k95yp>m%}FNvuPx)x!3C$30=)MDgy^y=r<&=
zIJm%G5RAewuJkE0xCd($YpN=uw&K9ch}JGmf~pV9;?JI}Xg=Qz?lbFEc{Qb~{OxF|
zoTyr+zgT0y4LGFC!4B|-h4~rHBcp~t)R{Il7n~oXc6iwC!<r8|n!mSq4`OGxEtuut
zN;T?Qi^kN>gPI$yUeL4<!LZ;BDPG2@8d?3KHCynsjRd6TdT21vQIPPXF04B|;9ShP
z;3OOv@pT!B%a~Ttjq_kJJxlSyioE6-kAt=Y<Ezjp1noxQ68BRY`h%8*`cV2Fm8BdO
zA{=mt8*lGx6dp-DXf=!v_P{V8NsF|BBl5In0KK7w`Ea=4jCm7jIqKy)Y97K3NoUXF
z8ndixillV&z&&#!_xE>%Z^0ZST+Z>yh-GgGzDOGbGb)Ia@?58WF+3;?k3Ez!YXgU=
z87EWq@wjtKIGbP*b-{h-hw)r^#{5)0e)k7&$D-Ex{noM8yzSUk9|4cFi-un`_K!5b
zv|@E**LQzW7St8a%})Bz3$O|^eWS4_d~jED2~uGkRvC=j%kx$x3mgw?4xJXvfD6M|
zCT!xitWk!PuAv1=^MW7E*Zsb7XYc-1=H)zIHZL>2ewO*Fv|ct}H~h~sr1h&y?epe(
zQGu`Py^;0!eC`Q%_<I%m8AhYfX@1^(S=v9#{D1ym|8q~z|NZa(!)}ULvVq7b6U#Em
zUS@3Pmm85UwlMi-a!sQ+q@({NP}{frPC`)<PRba3YiwtzBFh&I6Nq1F2qQbRo1(}z
zT_P;@B0_q1U2T8CI9!ns^>u`zJ+AxPt@!o71Im&$p)7>!)=Cc8%9v#d9Jz4KFna6L
zgeaAvn;kLq;r7C>tPhh`5!c^ka2eZ+xP-TF`OPF;Soz6h`Qf_^hoPk801#w2Oi{ed
zwlI|0ZGO!iH<(-(n#eD;DekXoa<5k6T<&=RCCm*@Bqg{1Op6`FSSD-ySh3A`&>u_-
zO3eUZ66M4L$SYZy0p_*<!S8bz^v{oI@qoF2`k6k(A5h`Wv`BK{B&X~-WrCCtNTNl#
zg&dTpM5r&#?8Qmt2SHB{l|ybc1!=QrXG5o+a5kE#r{P`%7GZSBKYsMY`jgL8uV1U7
zkUUwTG}rB=QPu2T`q3U)d9^JG$jC3H-mV2?ffm9!ss(LCNDlXUagn%B&M6vwG^#}o
z<gUE$%lB=`xM)fSB%Y0}Ps}JTn|(Y=LKFQ$-{Mbr&m$@W8+G@&?1mqO12H2WqUtP;
zNb3>(P@7VaOT&#8b_fG>E~k75M;}v9pZ=i?$_Lo!71E4oKga~g5d;@h6It9}h#-lB
zfS7qj{@R#x(ia19prdkFyxR;g%E<iyh<J1QQE<fMN<*j&z(&Hj7Iftg{Db&rE6C7J
z`XIQ)_Cat|U-C`n?u9n)rz-bAOx7l2ral|PsN7&cQX%D4Y?EwA*L|Qr#+Qv9>0^8f
zSEA~Za6Xmp59S2DjV}={=aY<)HX@a7InT@Pw26}MZfL&Ca4w?a+Mf_4RhF?x`voI6
zC;E9`O5k|;pKi8vQ?~tcrgYZhx!5UP#j=aa3Syh@E0eC7L>k#B9|%%L%y(&MPvE%t
zOiSiw<Akoh6|`&rd^nlX3nRCpY2=*k>o?!zTLiW%ZLiPIwT#P$?JDl|NxnNDm}x=q
zGKJC<UwS2@&MDZnOfg+BMOga#dOh`W_^W_6j_ve{bYv}xYB&B1Taich4d30I30Bkv
zlkHW0X9>&LhFe+Pg~}|mDg29imuWIsU0?lOYc846LhLNXl^e;GspPvfe36dtz%1jG
zL3rE9uV|0%FR;Z+-;@FOvaI%S96Yi<w%W@f6trxEG-+e=sFu`62}dOZwup?)-`Qre
zUC<7c=!<HWznM2alvy^%<(-}ei6SDe|Mj2$hVk{wn1r=IfAcfTA;0Wo!GSCp^vlDO
zmI)V*M;yUhVTV9=Dx4CvEvRVYunMQKAud43Lj4c-mH5z!a5r#{l$K8>m~jWKyDqQR
zTh>(42X6$YU=4-<!=V<+fJ8g%J~$$H?pd+MQ-0z>tneRm1#7?DaubLfxBzRta7{hr
zZN`|&3XW^sSH(ELfEb%ZJEXH<{y>OPLM{h{A<7RvM>$btGXTMy306-s*an2p3=4M|
zba_gg2BM!(4Kt6Bhypj^gg1m4`U8UMh~h(B<bJD5m8H466PufBKB-(CugC2Cyd|tq
znTOgU+HrGp(?n7vjv%@+#!>I=dwBGKk3p#6P8}o<IQ@>~5GtaO@EKBlAvlvKl0vBF
z;*$6&)DVtVv&|rCt1jT9%E0Y<0FhGpqcRObC-?lK0(?mI;C>!Xq;pv7m3wzBjYNJ{
zZza46$sBrQeQPf^8ZE;TeY?NA>vDqI()QfDvNThV8}pNvAfPWH%%KSa9CdU!RZb0P
zQbWBG3=Qs|Ny&%<Grf>zfzSn!e{^&_9zT5+AAb5EP_u;Qgz?(A@W?aY3q{?=$7}B2
zk%1l)sumY!1OSdealad9v0dDK1i`g)q&8IFz@%MRydl`E2)MN8(RW96L}G?MMHLc~
zE(>$Bf;0A3z#eMMfzedq#~R-}IEkIT12bkI+U~3DsCk;Gj>Rh~!;OV$OMY+?A61}R
z+dJx~O^>@%wdJYs1;!$OfIJ{Isy^j@A^QJB^~0aN76<B5^9K99UiE282KY{8^`ELP
zxXyd5xSL5SM`LSUFxya@^;^2>m7B}auQs_nKJdMAJB>zs_T+KQt1Tv{QF}S8yoaOR
z+Ko>>{xE*{qaVc|{?UK7<l7tfUyGYJZ@cXwuCJ}G#vlLaPc>h(O(;Za?c3ja%c>Az
z)=_>W31P&+U~lYf$I6q3@$AWyIB2$%=VC0~ycttdvsSst>3&q<z5o9Ef`LxdRi3%U
zWj!~eZ>VA^jApB4CMlAJNLxWf<o>cD^%3K4eQhm1`Rsuia@*V6hUdcr!Ihp<^*yl8
zyp_Zz_Mv@-gN4N<GpSdfKam?{SDBfk1RK>_)$P950M0nzr}<-~5_j+2^PPK0?mU0?
z9F@Zwqx;?uBZ)#D?4@9YjZ|xj%6==QpE?KovAdJ^@{u12kJo16?RVad-+Jerc>C>l
zV|sSMYUmn*ubsx8nQNP>6Zf}G&(Fm2(hU=;cXswQ-zv|6;aFLD9vf?`IX`QpQs#EM
znR_iJN*T|s;6s`>m|G!QUaDNI4b&M}^npdy9pbN5w`mW>oCRZ(`IPwtrZqDlhZ<q}
zlq%IGG;VIFZ<m*049;pU8V>HMQgU98lyMAe-hsh$rv7kTd8X7!9@IQjn0~<`3^?W@
zBy5p1;m$P@Ium>`o|(5`8bLsZ`O?^FxSmJ|GG=cq%*Fi7RB&e%`NEXtPA(*tU_`Gs
z_Tx}(M|keF?Cl%Y9B>k-$sE%BG+CceI)m!hj+yYxaTn*unxl~lQM(8hIDkx_qf!!v
z)Sh4)ddESlsrhiQkmjifW=6bUWe%kck!+#-laqDNDNVu7vgWL}?%z?no@=c<jHTJC
z)q$>Wwd1`H9+_tI<{Pi7UQ_C0q>?W*2cafVYqjbDQx)~Q<OyTO+b!TxYZJIQ(uUoG
z&1z2$x*Q#|Uc*@AP$@X4I!bNauQ;?D^T`0rMkN9BfP*Jy&}cpbUqCf7>)e6hj5!V!
zh|FzB+n{EW!x-%0xogufBv|BO%{oOo^Ae_^`c`Z5=xEJqK4G|xjSR$%xoO|Yi4<G2
zrLiUa0cH{RmeIdRE0%L1JPg)vqz+Ljh}vD|f27H{KZ^N-y792CqYkgGhr7*S5<U3r
zVSMoMCuS<Zw16f9!wW_j>mM`{)@I+Gs0jzAKI(q&-+xu_anrk4A5M;-KY$ZO3<(}o
zQngXu3)<7e{H!^vI_h~GxP_U_;Q??g4mN-bpdy&`_IDd<*Mr#B*yo-(7_1zI;II$s
zSdq$RNO4aVYXR*H%>?=zQd8#|qb2jp1!9!zZTgW6GlJ(i`RDw=TBEj8e?#x#0JaZ0
zpCh405-_PyA3fxGojIO(^egj?<Ll3dFH~=6P0*6^p2wtDa=#+;lNoi#F!I0$lQBi9
zU~V(#(5B3_%tsW>^eKIt=Bf1Q4-S5CZN@S4y6VAwU%*5!_@Vj$rYBOFz#5D;#u;Tw
zqeN*Ik2df&>R-5Pt_Bx_-U-8)wfxO{%kj=@H)3?)GJf=<pTzT()wp$YCVuC;Z^y#I
ztm<<V`}^%^3fDvm7u?|Vyf=26$FaJxuV+VP?Wp{NS{ExZJynb8$(q+(aO0;d8}WE$
z&5UsxVoG?^y;qh!hgCIqph_A%mBTk20x^E7F-(7A$~UDiYmzQkrU)W`dPxRYC<3S7
z%Q4cFX@8mdI*hNE%~z)V75-nxkk+p*wJ$Pd%VVzF%X)lK?qAkhFPn0Ve%^d}+Fx#x
z@%0b>ky7k}11aez2VA$|^^sPH7G+pHcB4QUu1oyr8r+FX43G?eVw1gy!%vZf2_p`q
zr}S`wd|XdisR`gn2fNEeQ1VHqbSuK9h46(bzO$0WtMI%|IO>{fVMrbW;$wAAlvy4{
zw&~&d<yr(6jx_Y0|2BB}JH1O6n-^*)iw9T{_9|We)17Y<m~ii!!lN*J@#|tZD1O)4
zMaJ!5=d`#b^zAGuNY|A9S8;ezb>0|M-)a7fb0Hj;Ls~SoFb@bBJk#P0;{pZ=ixit4
z79^{I(;rAJ6QUUgt~Z;^UN2N7L4&)FV0bY}mt~<K4zEraHlt-Q2@Z(#3kl(}<4aT<
zD1Fkinm8?XbXj?#5TqAlh5|6gTY13<VF6^(F_NUGluSb$oTdS(gJ{Af;Y`w<L&VsV
z@x}cGEIz*q(`{e>7IptcO*U6{T$v30p?38cqwR|}RnG{aP@rM~Ws^`5l8H)1E;xnp
zm1611;Nv9H*wlpe^|LX^jFK=lsJ;UU072nomQOM(-6Ss*Md63C$if)oqHqihGTNHU
zI!Vh6AOb2Pb5IE<i2kz)V`1jr2hsp0AR0jc0mfW2;>+UAz1$&qVOVVQ&aS-B8Yx9B
zw<aEO3XbCCe8Hv#BEq0IhU9&GB;;Nj6sti{a(Tjs@_+=Zt}08v5EZIq1WGUo(F`IL
z6-lHCyV^^E97c7aj7H2<X6oa1*Em6iC1U=Jhx&M3Nb`aaTQ*+Kz~xIO7&D3s?lMq%
zfBFZv?uYU#KJ`<ED{YGE;Q3LwUXM=zS+36?80z79q==iZ{56qX6BDH>yk)SJ<xG2N
z$}pu%F&p_i{W2}eRib>|DUv{Z6dq2)5i{?r_+=f6v~YE}48`Vyw@vZNOfqa)6Q^N&
z>3;R)ayyu8=`CyK=QWMbbo6f-{KfUQh*-9tJTmY6Ekp8^Wx<rCx|V}GUEi)>noN;*
z5mSZiGPv|p#Qa=Zn2gJJhPQ8-wm!0)GF05dyYS;%R>yLHbR%v7T18^^ki8oyf3F&{
zcw-U{u3XaAEn!(0KE5crO?vwGx}QHXlAXKlS0;GPIG?LD$`CGc5f691v*k;p)<|zf
zPl5=aWi<MP`enH?V1_T_6{+dR?J7r{6M~&ex)Tm>{UqZ{ezIr3$<OeS1II2!TbJq@
zGyinUG)cpb8G%<fZH08&q{#=K#h`S&G9|*;Vh8g0r@xW-`lU>F9R4{Q6Cl3C(81Zs
z{3Z-^hq_E@Eo+kVLlQE<-54XhjX&z~og>;{x3oEj{MjqK1o%YyhQl$(N2rB?$O2&k
zLg$IrZ^jw~P9z1uH8LO8VhC3-gJ61)@0r$Ya8Bk?h!+t2j4R4LamQpTIjnG>l1NM*
z7>_NLh|RrK@yMTpE+ui*_tz>vq!tJX{(!Vu1qmkMViTy2?(ef?69h)?Vc_6bo@j<Z
zHWZWfYSe^_b2kz=7<h>f{s=FD5CyXgX$U01Fc7~WxRN166(sLb1!er|{3JT<Ry4LY
z<A~Gf$`^w8G=$&Ugw>8)JA?A(&>46a1XJ1rsrw<-XMhvj;6!Sh-d@$0Oqma83zY+j
z2h=c;@2D^H4-;2m7SIp06)M7LN5X@5a>5<8+8}nQ@1eaofe*om6Z!|OmZexAz{0>o
zE!D`_m>Hp{?H(B&6#jE=g3aNfU_p44z#^V+Y)7+wtg?VZo?5bmc7PxTlg2742YO;^
zd?;q?qcJ+1Ogh>Ccwydx=>`)3)nX9mQNzaFIzRsLPvWBo52HtYJ-cutZrr{n{CYBW
z_8|Z^!7|<d5YRBF7(DJIE@{GV-(}ptdm~<d^`_NL0naeMj?~sLhR!<8=rnf}HtR(l
zApAG>+k&&c7^+m<_VwzpCFx+sRVrgqpPp9zgi9))&Oy_v@qE|GP(^z@!Nb6?38vfI
z>XUu7hsMa{#7NX@d50Ba0|Fd-BGfN&chzOzu=1;#xO%R3=5#kq&0%m|b^EE{9ykHN
z^4g((W{9Ac6p2ijY^6pBU>eCoqy$=fjo9B=GXu6hIT80>eLbqRY58lsoOa^0({kH*
z%nFaMsNXT{1-RozMKD1YM#>NW?9Z%9Yi@eR<GjiJEQc+@)p2a>?8N%^wqR$-cT(Md
z^{yqMY7@#!FjV2r=wa=5m4{}YxHvvgnRczB>qzhi0S^^%6BBjA4a_Rk%ppO@0Xyy*
z;{LJe*+sn>OOt}>5w{oj5+T7-tE99ga}2P+-B6p``&K<gI>(3m8k;<O9xq93g^}Bi
z#`c!lVJz<7do7-?t;NS5ec*br7h-?0cw<rRuW{0D`~EXjYTmo|s)p!=_c*m0Ou)Js
z8^AU5Wm9<|g}^uS3Wp)tC+#;6HRmERbfPZe9vtqw*)T(kzNDVt|Nig9yWjp+%uLOw
ze%vW@83)G^TYD`NzL~3G2Gx+L(%iMPy%XCSix82KsM&*otp0%zjKoEOgm_M)JPAiK
zY{o88SgJqushJ&mfI%C<nZZqS5%V{5D~B1aaYSQ<NfcN)QhiQNQSqpKs9<tnuoBg3
zEhg&qm_nWF$Y6{yzk;U;o*~*ZPJ!_eXcuxD98ms(**@jT{Wukk>we;B?9r!#%6DYA
z5@VxP)qT{=Y4)<GYU4xA8#|jDn%`S0lFB4EB^hfQTPA{k`;FIQZgyH@aNPOQUP#QK
z;&G3=dQQ$`tcD7}VXHRX)YwHU4a|2AB-E!TJ?A}N*%Dl~)SRqO$xMJzgG3w$2)KW+
zGSnAUl^xo^?*6{>VyrW7)8<xPtG;8;3Y~?0%U*Nea|(W_!(EsfkKcOpRl!u%a7SLi
z)}Y#Tr_oZqw#`_)e|Irz<0FFQv)J6()58@KDFFoyK~15#XViD@F*cxKpc;|pXCLo*
zL=OzVlA!>N12?2gX_sS-1AQ{tgf*$K{%BomcaS2}dg^uLERLDGG@i(l@cvdaF~5RS
zXfT<cK%ZIS8+wDvLtmpFldByVGqee76n%|^C^$hyaK}Ax!~GhsE$*#5F+(KpwFeJK
zHF>B@WlYoesL`C(+K$Ri6+TcyxvoBi`DK-!3Z)|YKBlVHgW-OaMSaIrBaHKHwdJ<L
zc2rM>g{DK-mvwbYbLmWd+y^Yy1$V&QomcM0!orgJ?;wW9#?_86uPScSLCx(*Zw2t7
zHIVtB2jz7tKu7I)to5AvkFg1E(5JG1qkwyCZ8hRRb0FglnB~4mq=P($)UsA(%E1Dq
z$=cuP99vrVP~~oE-a8Nu<9ndg20W~D?t2VlTYs!&#^Ka<cErdHJdw4*>$$#x1?YVA
zFAS|be35iO#(vWDz)K5Ua8Pr;nNUdhK^ugILmB#f0K7In;r^q6>1P;}NC6^I2`!Xd
zSQF}#bu+t7w-L<bJ(sDy3OHgcu?Fx4)_{Mjqt-eLPD7ezdI|PeCxH>lN}j-AITp;2
zw*i8vKXguYh{CA)!V`z$&AUtSgLm%5&ACadIR5C<&ouw+$GzK&@%Ec{G#?9xrhWPc
zqtQBzC(ku!2){nN7>sS<1Kj7vdPhrv3xIcWSLpoAWX!6bM;dnzpRUD|XRDU-1s>|c
z8*a?c#@tMfW0>)%;^hDqQ;}N5rM_;cNm-kc!TCkTe)M(mojCUwhflvRH>S(~g61pJ
z{tExEV@T^)m)aK@)`u@PKdXIzK9l3?i_9;Qwi#dlkAL`&iX>?mY?(;9dHR+q66fC#
zp01dT!5~Q+-kw;l`TAc2ogg3q<U@!eibL^xy=z>kjOx&7%U|2<p&$L*S3i7qjS}z)
zp8^3gafZ)~{eme%=ufs63uTj5{VI*Cs>`CV{-kbL2%<l%fPwuSjk;1EI*@oc7x|Gm
z-+DVd3Ayr)=rUt12Z6nvp1%rr($+7FE*{;6P=19q!<UASG7N_JeCy#w5>8tdO|(pc
zynSK*wRB!gr5mwmXGKw(r?Q+e#|~=Q8y${*Ayr6xU1-8L<4couqNK^7NJKRU(XciZ
zELuKoq0J(vepqnKxX>hy<P8fH#0d!gNYt=MVdvxt>SNAWe6(O*vTK#!fYO42LKAbE
z&1&fMpv#@jF_R)~$eYuGxk#b@BN<=lb$4l1MhF~cz-aOHhF*&;8-G-AKtO1=xLkWZ
z9zS`i3A?TS`SQu(7cp*o*X3)ZTXa*`@Ljr<?)qkdr{8IZtH~YsWzpyKh&DGuu*QXG
zA?c(D+OX3`AXgA6)TTmC#)Nq1Mb)SRN~RBp5ySvaF{2g|RWTrz5FqFq7HIdI-X<o<
z9r1QH)SP5t<6&v7Gd8}QfL6ZB(<i*#pLybj%?pe*tD_YBda=}w1LTc5P#)@<OQ0$y
zcZ}~tlRr!)5Ds9IdQ+#=b=@Dwic4M)ZQLg6TM(4ImjfhA2nGyO)dz+yr*G7f`eTr0
zUU{-85i)c@gZQvnKnk&hCkC|%6iS~H!bG(SPq{fi<qtCsHJhj+qD^6Z%}&k5sE{zV
zM&?J}nheT$gzDld=VDB&{fci}><{8dnJI3@BUFYjA}Y<2h;3862!glgdg400tiPkn
zxEZkUbIlQGcD%xaSoy^}O=-(8m@-tE+*NFTof%~;f|ubicE-CKBhx7gl7V@Z5lV;B
zvuNMal)h=Za>qFBqCIRJm$&|DM~ApFMH4%f(mmbstBmegBzq}X?v^Lr(kN|xcH6^G
z=2g06Q0iXz>4lB?nKE!{(rW^a*JIh5cH$aGusdO|>XYlAEcz*1InyRDxs<S38ijSJ
z680Qdoucn<mNM8?G-sDzyd2oemtBgmbSe2_GNfbMh8>?QS&@!S;i{L{2EV(``MzdK
z=QM@q^*DO6Uf7Ge=*dXf<mT|a%W(L~6n5GxODPA(E8GjW&moFVM-TZa{FV3hnx?<g
zkiXM$^2M0MFFkC}-29pCtcSZhiIl0NUmCs>HZy$^QL1{=cRqcmw|$GWF-2+oeZ?Fj
zkIVR9e{;vzFJ-=LGx6&-U&Z*MubCUV<t@g)t;HK3`<7v`YaQkUF@!+yH?$9MsWvK~
zTJ0kfSItxre#j|$7$vAUG+w7Sr+vX4U}BLE>oo*W?3Bff7Ucy|4+(H^9NZv0kS66Y
zZlXV{n=wvvDL0sLNT+K}S4pWqcoE^jDJZ-T;OKGNC_@M*dD2_B@E~|92eSGsmBd{s
z6BAXdK^s+BQ2)jqD0w%K$|*b;h7T$S8BNu(N=$LT)!2xMa4=$cPK8gkP(`IWquOXh
zxJ6@kC-xfq(Wkn>WFS8OFndsK$eCf<t|@#M$sY(`bMp&PpRAd<hk7VZYm=rKWx_}H
z+lSFP=(wy9W>Fi46c_kS=eVP`6^>6m!Ew}gv_9=|a(EC!{oEllXr>+yD#f{1s9$AO
zBPss5nO>+4M4EzmW?%^EFAmn7S*ig-8Vrb2m5USn+#L&1YHWPmgyA8j&z`~urjCxi
zJe&(wAn@(%w)EL6{OmHUrb}%yR!Qaq2jE~-qH#bw5#e8{ZVWYDw{I?)G0CB=lw|9u
z{rcVhz{d9WPW<Vg{WL!K<Z+y6LgYTOitxCez9ADZJKDdHC-8v^HHd4K%BbEWs>7L2
z$6uZw$IRqdEHBQ)%^M3bt~#NvZ&&$Q)!}qCMzo!23Ey7dSdU|+OPf*8b_W7y#O&-`
z)T*Oa5l1~zI*vPdvRz^J?KgJ3UxJwj@gHIq_b!c$O-AeRJk~aLqJ5yY5u78PRu}${
zI!z9e0f(rigHTr)trqI>hLyHpUt{W2<=EcaRXV^yj$_6ROkk@OgLv#ckDx+jr|(dQ
zbfL1g_jfG0bkN@OeOhmR>m5sU`L>YGK~z+C)Ns+Es47Pik^K#OIN$+EuESO{YBlX!
z#wRSn!CXKgIGK<17x$YT3zy$%aKF~7r5~1WEV&<91Bb_I>NDUo?;yU=m^wW^RKA0;
zI6rTxMD|3~3pI1vki=$VFW&#@`(`RaxE>SiQqNBxK2%$73VugoTK)3Mox3qnpN@eM
z?vNX?RLX?fmhkL}*l()Hf<Xwv+GnxdQBJ4oBM6`^!2x&r!1P4wlD#PmEGC#8-|KXs
z$lRTFqA^f069cAKv$+@CYxUM!Z^eSfHshOk?2RS`r^7Hzfq(TEOmh0z>O?K|HV|#)
z%RMg~2&Jvt8pn?wKaP(+{?r6fR3UOlANOi82XaRmu-Ma&6q5GTvct^gPQIa`KEdXW
z%DWXERPAvOm>Qlj#N5xE%6tu6uwOKRTHo^0Tl>As+*4tIA2YL*r&WD2(P9Y9BNyNU
zW&^Ya?!~L9P8>jhs9Z{1!9=<288EGYbw;y*=8I1BLW|IT6XqIHK)@_z(y(9~s<b_`
zSt_hr9kV3gn8qo7<UOPIn4eqpJxR<t+`D9nP|feG7tlkH8k?D_Yut}}{Jj6+$LfQ#
zxPSMK=1C-kkcJ#F1B2@ZW~S@vpB(ovQdn1#5!*K7w8<TL@@F31)jEk(9c?f<p>heJ
zd@w?DBIRA0os5xwjUUrx&dgj!4gaqCbmjS``n}`zRnc?*uXbX6Olg{A*7zWY8U*!%
zlNxud1y$est=+mJGERpEE0)A&4!Ct=DsC)Jna%)nZGCgwl1i-Whnn9lS=)CROEU_$
zH0QOVse0`0Grw_#0f$!A0m>XnEZ__S?EwZV-$)p5H5#$M2NOEia_AEWXmL?Hng*nB
z^egZUliPG6tyf;JRZnYWVeL|%Foz&<$-Ds!n)t8z!m8@z#^Dc`fH0NW_tTcpIZ*$4
zej@nR9D)QhI7W&2&@Oe&1rw8+4^$`OfI6T?a7t@0QhCq_I1r08Gz<}F7T^S^B4&&W
zNPPf<@N?6XrX|2UI#;>4>x_PKnN|1NII3GmJ@-C(^w8~r+TE$?88hXOOhsiUa{%+7
zAm9Bx*ng3Xuf74r3Bi@xLhBd@RKNwOGmN%<&G{Q!ThZ7;I*2u8SaZyT;e#?GSw}z6
zuEtG}Zc}-|6TIJ4yRtqockIC+VJ>9dP=A;vBv`VFP_=>A4t?|HATo2ubsQ8N3uZSA
ze!_tZu(mVLkQLu2w2Rs=8Asr4^bB}i(x38fK2I$I6b?LT6%vfF-mz}`0JdNUSb$c=
z8jR#4hrm?=GeQ~n92}xwxV8c*Ear97OqDQdBRRZJn?he?U2SQ8FL6y}>Uqp6vNd*~
zg;5~t#QZ|PmQrBUxmR<T%dK*Ge8`{sy(X)}ac^lR?%$fX6#KE(GZ=N;!3gZ!zA>lq
ztFfxS8C09?wT@$L>mUx#gpZxT4Cfv><_#Fy%3tk5yL0df{BLGvT5Z!8E2|svd~H+V
zj!e5nLjZMz;XMN*{#0<RwZ!8~f2AqulR0Cj$vS?qq3!T3LuA-54*#>#!F(0t>#O2l
zH(#08*GTJSL)>3H_RnLk#?Q+p%l*sb&bTXk_C@oHq+Kw+{_zVbX4~eZaAg=Qy!es3
zCN~u&vy+>>{EO?ADeXjqIIC!Ud7?_<pF-$mIz~pf{|jk%{W3MX^9$Uklhe1Au<|RJ
zaJV}ver3#SZdYYazapSKoKOZXOj)YZDJM^6Hp0rWh(kxX?Fe|S%$Y}i<L9p;7?w;T
z+o;UGGz2O0&59SsVXw+xX4!qqC(|h+D`9d@*R-Xzj9>WS#zbCNUoZ-aJ3g6Ox|JsL
zO<#hY3a>n+$l84N3JK{Q7BV1YEma+Id8Rg0&MWUi;9-G5lpYmsEZ(+jbAvP$n|c;9
z{`!Szu{pM$_hNB7(}Zt&nkM@bA!ta_uw6%F9W@_5xut?71C7(8xglc#@hJ!uo<$7u
zd`kQvx|~c33=32eAbM4P)D37s)}lg<DK};5XaPRaB)>02^69f@@$fVB!J{Mp%O{5~
zGiB3W?}u#P7fl!LzQV95(=*+Lp1uK&OAHXB47@3zot3T^Ds?)<77%6H*e8f(A0E%m
z0E7phpfgelLMY@xi}xrZ_1bK*z@xk0aG|^bITmyb>J>Pp$LTR|B-D3Q(1acOOm83)
zM1y<P99jP`lyC$2fgwfhVSeF8I!Fbw(XdFl%4<YO{`4j7#lp@#6`Zm&v#e+nkWW<Y
z5C-^Taih%{@1zF;!qUrKTrfvF1>(=W2^H0sa+r`M8we`M9h(EXCpQVG1Q2Yi93W<t
z6(${JED<T6(9w71OM1m=akX#Hr4UIWCm_!-3Sr>+H0Lp@6SNtv)?<2VMw@N*OON1|
zNky<px{Mpr^yrkubB?|X7rVKjHgYQ_BwC<5wsA^de%wyPDCY%x+L{)^5hm@0w-PJo
zk0J~E+1-BK*(Kd&id^*MyL887$kGt|ddL^UUZwbXKfV~IE;<#?1Z;REgp`i>U**!R
zgsf#-;ZQi~jRzkY&Ut5}rBepV?;<Eq_K{=TT%{#fMJ(fHSO>{4WdwU<_%w=V>-7Mb
ztM4pzk&di<mA<9Le&xrtv;npB)4Ll{KMs<NPy5S48SVYD>BcJzVfk}?i}oOL!DzWM
zMC6rWY}G>u`mTg6TryVad;OihWw_GKdFn5Rl75tj)6h?uSH>(174+E9AKpb9<#;K|
zZ<~|wJ3}zL$fX!Fm@<^Zx#4(}4%4k8@vVOlalVT%W}Xonn8t%-9U6c6`2$BH5hZ(+
zwkq2hgIoEFn{7p3cuY92OeW)ywroh;@Wzb*enIR}+QiqN#{f|*!ImQ|9STCb3?;$M
z8*95<#y|ays>5H({5<%=H*eI#9#FEq+?#U736N&(^%_#bHIRclfrxG|jT_-0mM+&@
zn5FO$;U&T!k;Gx$K0fAdl8y;!?e>w0T;+i)m|ZZ3kp8k%5_o`c2A`7EpAXrv{<>_y
zVm{!AFu%ZOC_@sjAhN>bO1u%gFqvpd3w#1xQ~n$h0`Kt2V~qo!?4@GhB4xd(r|NG9
z`I)Gux*+KSv2=23A|`58;UfdqGGNtt%8%fd%!2r;KL`pCNl2>#v!P#gQnFm^KUo{K
zswC8K4XJ2@%CoV(9e?_#KaB0ot+;vTcGRaPjXQCw8I1~~BRs<<*3#D8oq}XmpZrie
z)zfqC`&|x>AacN5>-Wj?UE#vqlQJ<<alWVoL%kIE7{oc+9H}huYjA4Xc38L^cl|(^
zM<t_z1t+~>7(qCknwyWlp^?~E-}+$Gh5A!C_=WOAa*L)s*M8zy`5wS5Yag0mrxq}i
z42D;wGGcrVJPQp0sH*}tDNA#|8GDUg^_%e6!C@1hP-o3v0acU00K__oS-=2<9Ww`2
z*6FG7xV1E62{4!;>(5t(t94>_cE*e$2z%=r+wtJxvv|I_7YD-QC~rTwuhOIpNDNRf
zaBzrn+(QI05<7TsOMTusIuP8PTK(P4<)yg$%58_=+28k^yEs*i#n~yrgZfr=+1_p#
zzaG)Ptkv4{US@h~HU@_)Ccps?sIojhY`e`+tF^nktG+vpxmoV<5^j5n8ojpfD4JWm
z6&F1tv9;H+R0(STVA7(7t1>KDfkCKoY34_Z`>`&}9D%sUnBXuQ`EKv-1sC5ZJ`V81
z!0EAiJk&LUKcgP7TC3;y()UqoU;RY=&jf#})7W^7)lk=nYAEeh&QH|-W3jX_r@V?q
zQ|Uu+-D@=B;lqdV$)}&j*vLq{`Q{t3w0zqPCUvF9Giw+GY}Bej^lR>I$64nv)rQq;
zUa63R>le<AG=rH4`$~IjEgpRGk@9K8!u))^cK_A5qp`%e<*qZNxPZwW^#yma%}=XN
z8iP&s*^{S_<Gr8$G=BQwM}exk+c$5<^z@AO7`3QR&e%RR3DaZ3d+AR;`9v^s9ye~?
ziP5pVV~qL(bL?fAdl-+X%%W*mp0BAN4$Ro;xzzmBX-4nGk?p8}W2_<-xwEsa{2P`o
znyOF5y;tw6eQucvz`kpCZZ77Q7S-k|$7zolT?1-j7%wo?p;17%hH2c`S05_REy2-F
zBUPBee1<S<7-!z1zctsgzl5l7sRp$bdn)dLI&AN0K0Q*uK;TU&1|mT(V1DvCpm`FS
z0bZVqmFC3-cOSA>*C+ERgkHUw553=#dtx>jKh^GMSttX0!~xIqBic7|m>^ZW`5>a&
znX<qvhOuy_d5V-gmn$BJ#u(I;k2!|)IWRy0t>T|NwN5d&C1b27>M-I3$HSFTmyI<E
zRcA0lk^X?;vN$g|fLS`!YxQt26WI&SYn<mjd+J}!P3#L_xpgy^7w0^lnG=zyV$4%d
zRIK$Yy|(7+=HW@~wU5<b>Z|sF*9fF8VD2(+a*rMB)$&|D>SLM<HIH&=b8KudChFtS
zY#qh3XKPli$ja#hs|ruw!-P>!c-%6!SU<Q+NLjcn%%?E&CN-wW&yqlD<I8jPmm*AQ
z4Bx%8B!8GN7iw#*0h-G;8~d@f!+meunRw!ZwMz@LaYOSi`8<8T728|fo0xFHbFO{|
zpY*z`eqe5*FON?yJboE-E#@($2fd0z5FO3=*2<zZnMYt|n2|AzbXm@A9DD%=eLzKZ
zXRtAsoq0`B8ZhwH_q024V5%}tLsMbxfjI`vg87h)C>Mutkq86(@Zn#@=PoB;5vf)B
z(lDwAi74*if`JHgDCsBq6C6`t)29Tdj0-RW?xy6fMHp*+y=n)|p~eSBgp0_oE^z%=
zKh@~2n+mM5L8BQ^ol%v`+RdRKaHwROZfPDzqRdQFjbjBf9H=1fQ!{Ij+|;wRy`?p2
z$4o`a0b_k;YC5VgXI1|L&9yLnX(Jo;8Ep#l7!|o3@<37xo?1s`X(}udP{0B-7w9?3
zSj@xe)Pp|E_J%GbJfTYi1IL2Vh)PA_1cKG1yFizL@r3%oL)#N~?R6PRk8w<V`iv|R
zP?S$mZs49nE5yUjxMM8`uLSRe@d$p3q0I-CKeS00Wr<6&4kcrYJ6O>~1&s#DIAE(}
z_5nN4!hi?U<s38fG(Ae~adyfaN_g;2a2RmHJV4}j(frCAoN#7*G!~}D;>PS~+`chm
ziO=<o&EVe7<;8icCOvNNsNR=Rt80#l(b#Do#ohsT`!Q~yi*k77%y>A=b{LeXD;}?o
zD9r`ecVlZ;<5)P*VIDeyZV0YAF)kcU>xm%8>q$ai8n-=dTzq3Z?_#L<ONRDfz9|i1
zucvq2B;Nhyrkl?f8F~Fo|Ch{Hh5b5dy=;C~9e*|E3mUpGzf8GH<2mOG%rBC*8DIbN
zKmMOGMShh=**s-N=`27~1`II!gEW$%!^j0E{2c>Rf*?SXu0>24hTl~lS@4YOd~h#*
zG3m|^`Rhv`PD|f}&roH!BD7vAsuu~nI1lpjul=b4RvVnkZ%qQ`;>v-r)Bwq)Oi9M}
z!m6ixDw{7m@XExB80pRzh3eXQQx92bhjg{c9F!)(9gRS;^Ie4K8mwhn#JT2|-?n5J
z5-3bM*)h{*?y$A^z6x5zRcH!S#>tR2*wQ5<mgN8nw7kbD>Pxgq?yYE2)ueaP$4T20
zPt=I{YZdUq!$Qo23=#m73WS)68_6vC1qKtF4~PX!_#jl21tg4x#fu{8C>+S7W^|=}
zsf{2AB^Q+nX)NNEOdS{poYLYk0*gD3u+MTrr927W`cpX|cKEce9yaGDoES;T1&9fo
z37t!RpFVvSPoF&XMwyoTDs$DQKcC5dO2!;;<P$Al5}TkPVrBWS992)guScWmS^#}|
zg`G0JsmwUpcMd^-Z!L%e71Zo-hu5IfLCVQG`L$SAYE^B}Cas3YZK{V8NFaSa#iIT=
zQ@f$VnvEM}=u!JpCNqHaP%k!u9?SAKgGga`b0Qlz`qL+-Sge&E@BpEyM~g6iv<KDX
zqIJ}_B!UBSLzVJ`j{BNg0*h)x5CZb|JVW1;x4t>`>JzFUtBOZE(C3sB@q13b9tt^z
zfeT{7<^%+dy9jce7iAZMO!-S>7g)sZ6LCU<3|~r@4VG0LVpll&fKxdjPqZKEfA9kt
zI@X2-wTBbh7*9{4W1b6x+0-IM=Pq?VZeSk&t2`dpGAduTm0TI``r{wp1dU>x`v(^%
zP=v`4ZfUG;PFWmG@k%Sso=dP?H&^q9D(bM#s3=qUi78&`P$q0sreUk&GO=Y4?06K;
zX-(s>MHrj(&9))DOJ(B_MR@A<|7Y({x+G1q^gQsIxy^&B&1OD(gnLBfTsc$$ec0Rn
z1-<SA76_96pjYT~kXZ4D^sadb6bW=MxB#&saR34ZqFI$085t4o9`0_lsj8`(+H7uo
zpZAKWse5`%m049;tP-<NKlx;qD_5>uxmSikT9nzA0;54%ii<$sg&;71!LFAA*073h
zJfr5gxh90iE{7a@Ycyh%$n=*)U3u0JLuOFG5>$34n~4tmOsr*Y$}JDdjHym<8J9wS
zj8E8IGrko!sedM3vvJzC<ky?V2@UM{lJ@Y}oAA)LJe-H4B)rlj%y6svFsD)cFs6Na
zsD5?J%y<*Gd=E_4hX1Tl#s}!N9WqLmBa_OnnrwB))OY%(4QsZ0Nd+%efQeS}E;wa)
z*V#6iLMt)EE7k!F2uM*=0fmP`$WAza+lc1c{iem7m-EI}0va$feo;1BTQT&VY5RZ3
z=!>ulQhB67SlkJ(^b=>Q18!+eljSxD^mh61bGb(1fB66XOML!H%-=M{SK{dJzp>D=
zl$Ewmo5p77w%;s}8S})S6{6q~LJV*w`C`_>BRssYkcYdjn834Z<`4nM>6Hb`jH&RF
zjDsjV*dfgw7zjtWi<O;2C%*Ti2oJ<VFcvd*!*3*_m0$4LHcBfTR=Pe0QROq-^(2i!
zDI(rg2BD~roYA-s9#`Y^319NUcw(-K8wRq&f58D^;7cbaSvfov4e?hi%ZqvJkIHa#
zbZlitF0ukNS(b^3TjE*rz9lYSr(VDjMUwy*^Ww|bmz%M+vKWhvnGlB;oxPp--~Qu&
zj9>op7e2!6XMgy!SXy3ESx_QQQAqJkBQZIZLMM131Yqn;VyBFEK3$&0xp<R$F}!MM
zJ{SS+tlD#anv3F7Fj$3APy~2quiEZ*%+Gl@!2C?K_qMELS%IgVZHVteiFFpsYilvr
zTvT2oJ{E9iZ$I`A4&0{ZCzYyyRb~`(=W5#&n2*HtZ0;mMa0ogv|3JZH0Vc;@iH|!L
zJdgWa&i^{t=|n4$HY}qT$%!w6pXE5X#YGes!UN}s04$YXGn2wnrHuk%e{autO?ZwH
zyo3k#-5AGKda02uG({mn_vmwc(_z=;g7=$iG^4q&XyGRO+=b>JC@8!G1)->TT$FfE
zK6pPKJiMneB2+sTWb>FPcH@xeVf)DSoe>YuQHUG?$nkgJAxBNo#_XJ7M+&upKjm&L
zb}C$*o%rskk?Zq#@%(eO`)PdtM?a2@+waB2b>9E8xV&t^p?PV+h%^vRD!JgkDt1Mw
zYpQ&?01Sbzc_85?N;Y^}=7sbxIPJ^Rew_Dv!u8j}351*pWrUF3PTNYlW@A1^C#K`{
zax})Ko5FWxO&upjVs1)6S9%B<QKqsh?U%p!MVt$V=H};IXI8<ltu9-sgd&FWY;4^2
zoku6<C$YJ;>!ZWK$(1GWCU-%+ew{l!!QV^x?GaYZ-x8jjX#V#+UcGuHdp{N$bMc<q
zWNE2sdVp~K*%vR}cRu>?Nj!P{p5St<e722qZQ;wWe*LNDkuNO-Tw7hWvT$~GF1SMw
zoJ9cuj(qXObJMmDzx(}o^5J)5eql*84W;Oz@DK&gsQS%ZtgYV;?uA47_-m!TtG<49
zp|Z60gdY>}z3+Y$E2|sEO@x%-*6!YJeEjh*Vs~fHbej8Q4uzBSeeUIY|HBX62S-$I
z1kEU>*yTd|(H~BQyRGhNw0fes!f|$mpfo^v-EJL<cJ&1NREW$9522f$Cuy$eW5AC3
zaVq>qv7||n=>?B>ahW>m5e-V20Ch#+Tjo^i?_dhb9(L{#9>Z?06i{nkiVQ(+&`nIX
zWquHvZ_<<nLSp*P*u@(~BFEf&9>(}V85NgBxp{OQ>GDJLRIXSU&74kYq=1<@wS!nD
z;XIv(EFJ?y69g}IZ=nc7VZFSvYy~&FnmBrKZ+ADgcXnKk@iE4zJc5$EHKN3$2lwK8
z-+7|>;f&}m3nE_!#|)kv_hYfyRDWHGYuZX6Iyt+J&e2I6bh^e96mclnW;Ir@Q1wE6
zu`x65qZ`>3#SwMW`d(dIQC&u2b8|QL_B-kexoZo7_x4^Z+TEVpjXn=vklxzrlE?10
z`c+3bg@T;jkqwPS?9im$jzueIOO!Z|9^O$O+KA;QM^K7JtDm;JC-Ho%6|Z($aVQ+)
zek~ODPafTk`*$}~zP|dP`fj@y9E-~_d?=n#a6q4ZqL1J~J97sZyOqw*Sv;M)ZNXpU
zJSgY9>r3Cv`w36GFfP)+7~g>pJGr2vjNOa@Ii9OU)Uc$BeB$bwz9IjS<Qx149tx+u
z=thraR0_k%GAO=$x43YE^nnNXfie^x07W<WNxQS#*7u7mfA3~f{pin)`KIq>7m69%
zh|fZC>I1ES{?UVGg;O)rV^(@B%+GpULwU})%$UshMBQ0*!qL#|)*98=G&|Q2&at!l
zGG1-(sty;ixX@5sjRVRH#TR+C+nw0n-m%ilW7o;C#}&pZ6g`Z`+|7vMi?NJ2Ea*Xz
zn+m7wd+vX#-<<lytY|pA334m$fa7RH%jZOQ^T<DVk4wkTo|Z4mn$#||7yT1zVp<}&
z`bbISTk5lx(J)Ra4cdw^obW`a-7R~PyX7`6AF`oBbMy_+m4wqyx+FL%U3Ox!YnHSq
zFFYwqFqOu#G-w*~o}@o1N5jx3;Sp&A#sm0IXfMY-f~Uy3)O=Q&($}1Sp*xh7BQ9Nk
z{dl4KnD`Bap{Vc&loi;rvwux(*wolCHhLA0?ybcSKDZYbCx@}UyBE!7BkpaiYMedR
zSkn@`nlZn$;YDY?{#kHWBfGs&eD;rX;gu?(zSuF2uw$QvEZi#!UF!CZqph}PQ5t*%
z@>^)!!Yp@qs+^|OG?@y>TQ9dJmSP6Bx7?h!+-_X?%rF1G<yTw>riR4~+rQKd1#!y5
z-`^zukomgs9%F{g+u^@WZr`?Szc1#Sl>49j<9|3Z@{j%>{~+O2!t-L9m&1OU@IkEy
z{z<&k&JVwtK+V&1M32Ljv|XS=V4Oy6BwkPuya-<n_7Ye0!+zz{@okZD(d1G6e$$k4
z@zg-PF|t#REZdEnO$qDZ@~_TLs{&vsP^nrm6>>P@nSO(^miKQM0?E9Y*09OCB<Ki~
z-;|CVp$GJ9TyEep)Q0jKq^+M6rlvz83{Ne2%}f4o8C(c6FvX#!UGo<I6X&vAj*7Hu
zvX1ggoJw%*5ceCP#NvW#4<rl}{OJv$=kZvfKhk}D0U?GcD?JS+Fsy_{$we5Bp*Ry0
z4YIJxlgJ24P?~Y!F-O2LV8DzbIG7sGGkhs80|<i#1094HCX*Gy2nt}liCa4q%ruM^
z1F<8^FqwxLLJ7j4M>)D;)c5xe;^oViVgfhCh%wmyZcVliOcg7SG&K~Uk}04;(v%{)
zi7<wn-c`tA1I`S6Iud<S9%j1KW_4BFh!|?8FaHq@Laa#f7I-mzTu9AoWCrosg$1t=
zUe?1TlQsi(7ch|;oM*(`u$l|Fbq?EN9K-}DF30?lKY|HX;xMqoIMPPck8;!I-mN9v
zp_DM3w2_sP)Qbs|Cc8{#U;t5U(Jz=>!n~3fxCYFaR2vKQPhE)*UK!U2r8GQo0tp3c
zj&i7qFF?ZC@T9@9FKod_0mxAvl#k1Csi*Om`YR8Pg8-L1JuYSj|7jcj2yb8#DL!yW
zJl8i&)K#`CX;2O(S*-S=ysQu;VXso>=o&HY?DSe+(gb;KL49B{@@_A;o?0{o+%&Iy
zy3zx;vwm)3zlq|44;vLWX-1kF{7sXtU*(riNxN86P@idM-kC}*0`RahfDpxP5Em=<
zff@3^rcaW$-S7Nx6e9BJy6J6u4dIq4ZG^K+WcN(8xbstdNYkJC^^O^NI==F$;rvTL
zC!lykcGtrVq))wT@h}6wfg8U#=C8j<FB{kO$(F#>_nU2#r5-X$KdUAmSu$KHqa&sT
zzjVm>37a&7IJP;G(k(dw4_t@S&X?16tdfw+F|f-*9J$rLn4X!l!6JaAFdWyLAh;0+
zMGoNxWhXkOeDY(MWfF=@0<seqKY0)olm2OUn0!%9er3;ohEIwyT(e>$Y<^c)Xw4wK
zVsgx>DG%9X0Up`DZyLMlC#$IX@-4gJkU?w+m++~syx&f(nu2sTPS}rYR^t12C<%Q1
zE*{Vz_yNjM;cGu;-SV?r35)ThPY(HqGm&}g%hr+jlmG9(z~!&R{GH<KcW<&EQ_d`_
zaoUXxK8P347&kMM_hG;TAncx*7}xlJ9-VGq6XsJZ#aNAgg|J`#b8|C3rt4J5#;Wz5
zL-7|!C*mjE!!(Yh(gS&E@-gDyj?bdkJJGm(<}`iTJ!5St8DlUViQ}*kR#0RFSr(cq
z1mi#aD*O#$FA*BZ&Yu_V=%Fto@Uy^+bW{eAKKYGOe<f@_82r}b9qWPtLQV_N#23J0
z%r$0GDA<_ws&4Al>z??I4g^Z@rU>KU3*lWxf?e*p8v+3x!vj4GP7ndt>S9xHy(PZr
zSbVX<$-RBBXMsI5?}tD7Q9OL~NOcmhMo>YB*Haxi#!88cXA_U;4xGE@j?b^-O!YVu
ztl;f)@f3<L<tm<zV+|KL1>lzIsCGVA+aRo=%_qmFgJb)+N94uJ7vdLBqd7kl@7=#0
z-PS>z^bdnOZ&3X7PsIZ&&h2~mV|um`hyAm7xwRLcKYI}zy$IY8TCuBuc>pwpdt7K6
zfWW<FhPlecT*Svgia#`OFN0u{_MkyLFJRZ8qQbY)9t(}8l@YWj%B0oR<(Qvu1iKJ=
z2tUQ!G9Likv^@gk4wv@#QB1;r=A9^2xXGPU>^?&ohoH;HfOk95-am*_c3vn9i_GTB
z-AbmCN}u=$%;*=Ke1X6ZerZYNdHC?IkG~RU>zxb77MOBM!kFSvrk=`pE|^Zno%Q9o
zv$o<Lcq~lppIo?|L4WEEtw*TI?ze-TSJ7?n#_rb3`1G?+l=g|*;a)s?@55*|mjtg{
z@&2RtVtIwTN+tz&@vnj-M+mYriXDmV?y>tQg5bJZ9b;v&+BJ`k$4zBF?!{4OKPH8O
z>&pu<Et&vMvvY$*X51Hbt*{6HxohNVbV`i$T=;0MlYSf@wPJ7ci#YE0;{Ep@$J+Xa
zaTx{a?DVYWDJvGL9tiI_VvJ)$!F2AvVOQ72`nm=993|CK8=fB@`S`!);)-x!LU=Z9
z9HGtGfr9Y8DIR=Q^+hSb5pf7$uQZ<=91(gUdLz95_y6Hfo#ub`fBF0I_x|4Bi&4P`
zg&1?LanS&d2t+Bst_LpyRi7p;6ieU-3TBS#M7d2oc75G_@F?cyoAP7F!>Oj-qw?#N
z9g@?rw76vY^{bCRj;*b&SiimDom#*E<=WcXTKwSqKZu|F^e1Zn*P6p_$7i2?rnVbX
znj8&z=%XG{wt&YdO&&aa6u0l(ce~J^kGe+|hA*&FbaFbnC$IHfh>j**V@|`a($-$f
zeUB4CIDG@!3Qn@4vM)HJ6homqqJHKDs(NaFt@Bt;(C{2kc5oH?2eaXMk^A{;@gIa$
ze^(%$0S(ua^gj7Sh#J4V^T|=kG`=h}^q3$=(lsxuG@xZ93@OAU|3+O%J5$8ue3W@-
zjgRt<5c)I2rgUU-Y)Bu{LSRlDXajwUc{j>$=mK|90S}6gpOD1;oc;r>=<DpdU0Yd*
z58i(mA3k}kK0hT~Ig76F0)+=~rC%O$+~MhE^wlqW8aq)Uv51phKzG&_Vr8)*c#D1s
zzLW=LtChy0|I5qE-l4(HjaS<TailpZ?ZQ21t;1vQ`0_nO$_H7@+}xDnaP01^<tBT=
z2g-mDd~Ib(<C5sEz=0wh84A0V7!Mvl+K7j@7h`%{W4Uk~Me2U{G+ym=W4}Wm>?!<d
ztS-#QcRqSA?rki)9{VlfxXOE~^ibAA>r=)6zagB_o4jZSj$5;mN*|OTTF&AFXoBxk
zGJQwZkP9U$uAullQD23o!Ea!gdqAgYcjzEv2DnDMk~fM8#%mUAlO}nYPY@2^&UnM(
z-g(iJY4s&99Jewjl^66&j^^|Ip9%{NweqN$*Gv!y9-Lj_<Kv<}?C@0mxziNhh+TUu
zMqq5iq_Rxmg%eHj87Qw9N5EBfQPE!*pOCX~<Rpu{S?oNoG10iK`hYvoW{#Jf5RGM*
zBJvk@4ReGl3v}RzP;91*B^N&U(I2uT7Js0ugYO|P76pLQ3SzjxGtl?g`Nr;N`a<$O
z;@7}`wF?Wa!Ef+^opSJS^@u;m7u{h;B+Um6@<I#MfqqDO@Jwk+dkH2etAH<gs*!y(
zu$6bzhkmQJrEf#ye59b<P_Tgm^ked3F*IeNtr+vkC-F&rRu6A_82O=egr<2Bk=n-m
zg1%80!-F7yVo^VQ07oXWI~N>;fIvUUleC<7;Wycl?wxFOF!`}t4&@YbPr|@&!mA;(
zy1lv-cQr=9GokQ$v^pJ+?=J<r#aV#OvDuv1)H_W6Q-g5KFD}RQOv6VVb~OHR94z-!
zve<#e0`Ps@(YL+dHt)?2;nk%jkF)F$=59#vlCsQBX^a)GKW-UN(s9#U+p<2uwi?>0
zSjwyV@lG78=571U*PDK6hQsn?8LFv$`d!BE@Ai79`LcAXdD}n7t{d}p@xRIp!}4#E
z$vS^^+5gEu{$mth|9L`7VJ$GJA2tP$O3bVP&UK|PLJvNj5zM{bbR>89bVtaMDXPPj
zcm1}HUXI5A>achb_9k4~2%oVs6z=k|L&hbZ<CkBjYol)w!gR(}grIn3O>WXQ-qdI^
zq)2J1u{UKMxXDF{y)NbFJK@21mM`52mSLO*@qrWm>0K-;pTj6D^>N^mR*9J@<gZJC
zlk?Q02u`MWByBh6q`Y9vN-FNGgyju9lrIm#W!}iig^<DD>7<{_pRbHtBJ=IeFFSC#
z5Kdf*NxBllXyw<GXS-s{u06qkxEnVdj{&0)HA<&)vO~xWors?mP$JyEyGw&0JF8x6
zpk`MK4DPuYs!{+eN*HiCCzi<qg9QfWmB|um*&U`r`7n^c`0>Cfz~C|n!Qe8fg|Se|
zIWTE(yVRgcSs8RGAD6;1$ig_()x=B=Uc7i2uU@{=Ag8owi{H7)c5{c<Kcd}hG>a}R
z&zq(+(>r{MA%I~bi{9C`iQ{IwQ6y<F91%P@3wU~FIu^vBB5Y@pGAcMVnv0rPu9*qu
ztZ6U>CJERAKemc92Z2#$<q{J++Khpld>J@dG3>S>ADR><5a!5}fm8}H0rPGt<vpp1
z6vP_b%VQ1MYj~|TrmZL9M3X<1qm1d`04ABzgrAuxr33oFa2eNJmUFc=TgC-b`i=58
zvV%v3IpW)RD`)VN{=_b|DZz(2-x>wiz$@XK!ZT^@Xu@iSR_RdRGcnPWmv7R7!G_6T
zwJnCD7EqYMl(AZlx}s#`7#Z>f=QuuyJ~}-y9m~xXO;p5qh#?#`@2x=Uyeg^wuxB5o
zKid=<19Qx=IE1eQDZhhYj-7NPUF8Bi_;PsSr@iE0H~vQzCG$&|TbY4fuNxXl*b+x@
zAZ^3eiDvmJJ83&9nVKJQoCG$d;V^{7Uw<&4j#HC&IKKVkcgE+J_-RV6HddYZBV9A<
zO?;P)Pkm=K5>{Ef^!JO;l1_QqZ`iZe(0Ll>_Rv$m36}Nr?;vj(2PLf<C(|~ZoRr`p
zxMmsrX_Fx{p6yuaG<*@*;qk#_WQVK2gqJ<(c!3dtZ1B`qjV>Qnhb9Cv{>{4>Sv@{c
z9=%B|@$&8P`F0UqDOJ*T`S3|R+0_z!<hQL#Q~2&3Ome`<DJhZ_U+*%RmBXT(p{Dtj
z9%(CqGBLd~KWdQSDPh*!!Q>Bi61=9BwZ)_>2A7HnH|`mpx1ml(%2S{f{gB-N&__nn
zn{c=T0B=B$zt{Y>vmV+ny_CD$X-lA%&LoJX1oFcxzj@WN)wr%h309+LwDfmdR!74L
zYewLg^7&1{NRuGx9sm6A6klIzD((J#HO3D;Hw~C=n>GYD;v?qgf(!pW_KDx(PN+GJ
zcO3Y_G15Ex``&SMdt)`m#1CAGm*{K4k1%3&ZC&HzRJ2ddW5078-TtX}AF-+#LEUTh
zclcm<LGEwhj*#w=!g;cNCjLWx1YX1C6z3*8k9iaEArzGbD&S>RD#5QJ=<sS?$IYIO
zP=b6af5O6>+;JqnxTK4pg%WBu_(W(kd<d(&eaDP=u9;cx#NwE<c^@mqF1U;H>u7b@
zk#P{*L4^>DvaoA{qYy;}%qv?tqcZqHbkV!f>njTq)>oU+oSBIiPoKueAOCqQEiK0T
zPd-q7;;F^Ua<o^Yu@D@?#p+wdR9f(EoZ7%*92R>epMv5A-k2SB7xHK6xL7LlL6iky
z+Qj&1G-f7aVRq6-#o^B``rW-f^U(9tvu=;Wqi(#~-cmUVpDtMVh$-hUSgAaUXn_&R
z6z%~&Xdn8X2o!(_aS)cVn~P(!ScK%AMCb64D!1q&yKy=ll!fi^(T6CHR38LP;2pdz
ztIwxU(oIf!97M?gUqZTh#2pllw&Q4`N%M5<67Z=H;^|O?FhB6(I6r-_#?E6DRR!iD
zY7^qJnjb|S0z~)*;Lk1>j#NWPMR}&hPjeau_u>F|63KZI^O0-g1hC-#I%xcq`Um%n
zHRstK2A`<%3m0aU7fOUP@x>@RXt$Xu;ls*e+`G+UwMn@lXr?{EQN<NMdZaq+>>fm~
zd+5_TF3%MALVUIOZSKFCRQr+7_RD8+*g3F3`_X&v`R=d>58sak3&Rvf{4qlDgBGj!
zPpx!eUcf4KRls~YCp4T`>9ZTHy)ExP_~7xAc=+%^aBnaIe&W4+_0sp2&ChaI)Lpj+
z<>byE=3w1!+d^K1ZDZq8@%Y1!Vy?NQzB-~X=W*1}c4Cg=dy0g&2o+iSwzso`Bv5q{
zY}F^4i{P4It2W{2C4~0$W#%HQt83n!&=wx<?YE3ev?YQ|lrJm@WTyml=AhM4+fv`-
z`0Vq~EUbKZ|6cs$4}Pe3-Wf(;MX<<tekPou|1k`LS6uRsV4cN2EFk2lQ<RN+JG=48
zuRo2?)t~pci)}XV>Vwwp@3nlK80nfW$sIvBwDRu5C&HQV9bjj|hmF<6c>MTbeDC|;
z^R9>Xq1r*P)Feb{?E84gebL{8g99(@<fyU-kDkO&e)jicW&NIaF1A%i=0nT_*>!Mm
zcoYYnzW5T&$<!9mbcD0Vy(25T*?GjS3?BL@Lfl;Z2ffvtN@a!~KyO({$?hupG&Gpq
zZ{R*|0LjWz83gU5V?j5F`Idq1GKaT<TX~q(mB|<3GY-9+`6zM#T$HWT%o91tk~tp&
z>0EFC6-Tx~AH5RH>R5oWuw4%ZcgjF`=OlO1P9qol$yb@=To)lfJ1<gDPki!4APy~N
ztl{1{jvQpcsgDlT*nq;oaJW46PR0N1@Bb{m|Gf_dQ;uvD4N$xHG)C_3ck~=aUw!jb
z<@eEqDi4a_mBm@#A-E=-pYy^6jcKCG)C64Qm^RABLSvNLPhak;y^mE!;5wr5m$n)e
z9xGk7H|;t%KO_3wNct@NMv2l9ZnM}0K|8xp!SPcsm_=d1ZX}fSi}CQzvc{5I!q-E`
z9Z~=6pWlkk(Ybekwc6?<!iTk`Mtt!8{n%JrQB_7gZdv}M#F*cstYE$^@Ok{CuOi1e
zIp%ma!9o7eBKkRxWjweGx4=Kzh3`w%3q0TfpWp?^AD)3@ItRxZgEzJ(Zvjm5PC9T$
zaRn#RBrH3$Jx^0x-?}wn#nPf^sPA?X?O-gS{2mKrn45FWgq0SeA>hKw%8HMsL-EP|
z<!30A)$gq67A-QKQg`L+9pR#7q?d9z@RPhp6n`XUITqMTc6OVld<2@&QeS6mB|Zp~
zatiP=cf>)HQ7A&Ydfkq3t}6Aw@B!QkjIaY+>cMWVMbTaIq+C8~Q26G>L-al613!?o
zf&N+3$+SGsdZ7WzB(<HMggbqoU75)D;7<~tYyYfesk^+KmSG`2$_gixK58?igX}Ny
zsO+#T^>KXhTF^RZmOPawihuQ2`GFU7clxc5oYmN^0QLi)a-ojOBRG;T?F(#C^h3Yl
z1%P4FBGDS+P)C$}Dig{?=z{0az)4||sj|zI#bk1}pCLqZD}9FpiYhB~3EqQU#_vCT
z=-rbSC&HV<z1Ucui|>E*C{`9`tO!3BUU8%^_zrBwITCwzPIb7BS37%tPpU0d<>2^n
z7OAiUa7+Bei!FGHW5HGYrf3c01&Zq96BhI+J+;fksPI8!FZ6(NRRf$cNk@Ov*uVOT
z4w?3rPkjyS7~{lHIzwJIZ}{2%c8GVILAu{8xXo~0?=o-ugLlK`>*9Zvc{}EBlh$9Q
z$#Q;mS*`f`FaJSNd|^O*#YshrI-Jnojx`-ZDq*KRWUl}p1S1jSmfj{7_~;o(G5l}q
zZyLu^6#bUqj>k)X<b_{Wkby%l7t_A_EANtlQzLB6TP8!;BIg<hQv%pzn37Fe^oHrh
zIF78mv1B^L;@d&|s^!R^{K)(XM?AuiR+c%_#FP*G>UYV*mBt&ZeqBA6FY8h~@`tz6
zF5&ctQ}%sB%5O5q&28yCYkE%6ZbY<Du@+QKYXM<~#@j{92zYBCRhpOL(b)NQtpRy-
zN(`wcZKMnFg;;y*tp<7q4>u(^qd^Pd2P>Ia838Ub7$8(&u;-G)OEE&+-NhuXUUJFg
zlvR%?HyD)iQb>KPatz+AZeT#>2nQxIOmGOxfXAv2uOci14I@5w8`YouA?lQ!@|oG7
zekunmDLJ2*$tr^t6G<jOoE`uC`E%8cSODiQ&kWiDSpQa8C7hg??I>9o;7MBpzZlPi
z;iy-g&5Mzq)SDHnta#;kKZG=_h@8;Gn}PT2f(ZzWpco#(kincm8n`h51eR_$JpwIA
zpn*6Cf0$aD!Ayn<VBwM&A6A*tPSYCveR*mb@BuTEASQS)1ZD!%9@K|4fD-sddxG?y
zlqg+xhYgrf{h@Rt-Bg+)G)aLLLLnHfWP;D!s!VPX;E)EZjR=pj3kH>)hrkPF<@!2d
z07HyI4xFe*f_TEDs2sOM8JHOOI0Kc%tD#C6QHTK}Ph2%Y$bKQFTW&m)W0SEkx1>qR
zTFgz&n(yNXMA8Nxw$P9=bZbn-7r#h^VEW3PIq8|dC4#=N;YJAljo;~DbE7zrqcY0)
zNiQ6%{^Y++O|{0vVm}oy-QFw>?xt<Up5gV4w|)k8yqy>EGp$+<V$>J58lErGt{wyD
zAzPXfhF`x3oNoMD0r5r@m&!2+X2q_9mw<1CttRVe7#ABwb5P<L4kcm@!cP_`jonIY
z#bJymYz8l$nc^F!jtMnz7PEY3YHz&Uf9-b@Vi2I%j9vpHcYEEmO&ZyT>%`0NH{Awh
z!|0XxP(AJ7Fx11e(ayYUoS}S%j8nxK3&!DWNthT0`JIMu9(|XWjng1b{rGZIeTSo`
zz0^q`=79!27-~}O<x4?{pXDzpsl0CW;&xMIxho)enqT!dA!Oka-@ug@G;-iuOogG|
z+~H~6IsSYPhm&#23QO6tF2uzr!@G>>TKtKhzv+nOI{x{;{5!?hml@;d@2T;)`?hgQ
z;?64HDGSz`%{d?Oa{>=&;f}^Rjg{jXv)RRSuJIo}j#cmDjM=ZxV`p<K+6M=5cVjK?
z-n%Qlq!EY5XVK}4Z`2scoh6+^R(+?y0Y1d{015dyj)nUbe4)L@T`NeC2o>A}r0{=v
zhery8;JXpHuu#UK6bK<WysQOE8n;nQqymR~8gf@TyaQ?HQ8+2QOg<DIA$tsELx6xV
zae8_@78d4Xd1XQGIq^x?ULn7`hmajX<*~|(@<p%{FE%$n8`I*w*fCOBIpU0}V_qoi
z#c?m1vlDT5eK}YQ{PCZE>|IumKl(7<d-74-QXBa09hH;2nt;nuKkrWwd#eXJqEhM~
z-g!!O<1R7y6;><5XH22&f(LsozMgN@V{v{i<~jEKN<5GFAj-uKdUp3Cm_m4g0GtI;
zhrM3hQvDG`B2b%~TM%r<qUH8Gk8$ytbKC*KN@v=xk3bB#GxLBa;~pG#dx>XKn;_&v
z@Wy-tJ|CWw)%)D1WZ|J=aSR?uIwCMc$cOMS@1V*BZmj0#?i^r@0EU(SBOFzwa(8=O
z^RfuPS)GqyiF6S<vjYu*9l}bKEC^B&a)B4LJ%9vO+-a1@8X*WBRXStlrLC-hzN#%}
z#Pc;Mn_vSTS-5|ud5K_T9(i<J<vNd!;0LcfH8JYr%@AI*avC8LxWPSaTzt>U>%6B&
zuu&gcU!IS3;n|eR#hq&>n!B8wtM6T2$FX>L6pVZOZQ<Xk?{L}JV3*iT9Q6*(vvREB
z=b!u{{`J55A1w%D4z;kf9FL!Tn8yY#Eeif4Uh#j>?yB8Rqt$Bp=sYiqgAW$J%sAz_
zRllpedY<P!c<?aZfBe`x5UJ;@%~y`Qv9TVH-+SVV^4V2*bfmO|tA@GeiyXCfJvO7U
zX)M<7-V<_6S{dHy90{LPZe@nDfSpjZ5j)6G60%s1okrXj&7wS%@${!r<uN}y8Mjq;
z+8N~nJ1#ihk#_X1xV;_c%}y+Yze}sDF*!RI+xuNB7&@vu$6mJEEw#V!TXc;3QXbvi
zh?S*=P~uW_Pxz*J1Z~TF>h|po_b-kXV>d+}dpISyU&M)Ud2esm_ewu|_FQ@Fie4-!
z-Pt%e=)}`!FI1kM<{_zkc>KYW_}Nc>8t*-NKPKmwMIWKRnonttzP{2_KVc_qH>PIh
z)FzXrRhQ={dQRhDe_wU%1V?pkZEeQKAAhVl;%q$m&JSa8^{$W0I}+?r?4p1Mw_2UH
z=f>Q{M<3C|PIG3_ThU~8OmNQ}Lit>1Wr}M$3yo!N#KOxeD5r-*i!tmRLct1sMK(ZN
zqD+R4GS{-=LFo|3Y3ZpnwhBvgyhPaUPg!xVkBs8e&b7ZHJViNyP@M7-y&m6(f9{M)
z#T|BPb7SD6ghVpro1Bdvxn+6ssr6=#%h+cHpu9NogGIe5)YkYS8b7W%A7zJYVkgSI
z`?ur4ec`#>UcPu4+dF$!P#hnr{l~83CqMdb{LxSUK=VR&6*BIOMfd15p1s(O9pNT-
z@<8K>1I=T%&;5IA-u1;!UDFTKe2zk6VFoAUpnT6cD>VAWPCNG7C^U1Y)}?T6Vrp9S
zY|dezz3e`^b7#$pnRC$-Xg<3=P<}#xx&LcM_=4iA>#{)OL~~Xb<Ae9_#e>@`>We4Q
zJ=pVD8|vd%Ba@=#x0H^?d(jw<LFNeGmF1@Uf@M5v$CULT*U5#YjQ_sROYdIyDE3>>
zrZdsJ3%f(dQ<0iuJG)__Rk`zo#kh>+8e@@-2rjf8f6!pYpHxJEhoSoCH!!j6L*c<+
zAA_m1a3gG<I+S5Z7hZ*94N+#XP=ff-ILn>jHx!PtfRErF_~Hd1C_|eKk0%%~jPm-J
zXXFpcCy$ODF|XyF&q|X8$0^fdQ43>T?kGg5iu@rJHllZ3;g;GjWjvyp3TXHiIFLRH
zL;4~+u;@F+efd#VFCvz4zjQg$jzNBcvTa4Ql{#A~sX9|e7MOVP5NR-GB8SnN{>59#
zyH8Q>%Fl7d1iXv;8Ztk2q>110u5`hKwz9HIcJLJbjK}3s6w;=C7=Gz<DCnR=7)tLv
zmAT_VCrC?g+6!8QT+$1n6^Agu3f_$!;K0{qQFib%<~#JJy(ka))88CUdB8W3wo7DM
zqqbEz!lOtv%@Q51JJ&oOvSV23dh-G4;sO1+#q}AvNXv_vOusm`Rs5&$=kdckac6x+
zxXU=-kGm^#vA&4n>q`A}ET%NBv(UtgzJ(h-@u=+XCCwf2<s1_{Jv|p|YqvFKuc=*U
zL?e3f^u<nm@#;W4%Zcb4_r}hfFEszHagYV3s;J9b6<@Uts!7_J2Fw}$#$@>V(1y6C
zu^<0Kj&G;`wxQpDU6^50^RV5SrTxpyw@Le5=C_SMY~F_D@2kl=e|1@{`1-H@!9Rcm
z=o6AxL=5^tFyZLe#m&~Vk+}Vw34!gb4w7%aE1)0(IK!?~?=-mcRdTIq<K_tZ;Z51>
zm&RTWqL9T|Ma(paYX{si-7I2px3NQ72X?(lY*2wW4fU<1t?+OZ-ug}7w3D8JU;Hyx
zrc3fwafTW8AN*z@f2C3SB|Ru%x)O3ISBD>tgUL8mA!nm-gt{^HE7P-o$=|UYk#AbB
zm;$5Rp+Tth;ddw}8}IlMV|AvsycBaL>L{jH1H0acwjj)a<4VjugdGJ3Or+1d7QutC
zFzB9WLW}Z?v)GTt{IELdLIWN<gy!d9{JF$)%FF}IF+vCiTpo^LvBFDul*cF}v0}0g
zqru>Vvc*h=!ZJb0-90&RCe1on!PKyt0%BYTXa*MCeQ~%Z?B1e&1;c%=NyvVD`sruh
z^~EaIzf5Bku#rVsz1<ey$Ydj?Awj;?1}LwlG(fRR(~Onif;$tu#$q!XVn`NO7B#tD
z^5hIg9e8=Aq3t=~v$~YQHdpSBh`Huo4NX)88)wHzF9Xbhtyfm*fk~v#)84?_FqRuD
znekxNB1}grxXw*$m~66whkTDu`^;P1{xDJ4QL5#ttYpBr7@3`H)DNZs`UP_VbBA&j
z#=>nc0GYvI+qM`f-1Cfg7z=g<va*fA`Pxb;cB0kQTRuub?l69o3tTjF%LLAC##~U7
zL6yOGcgfDNIp7zJ5x7bmCge;?d?_yBl(r{D>{^jKJIog57BqR?i23Px!BtEk045KE
zm>(Lq_WR<_kDb#$!@^`RK49$dmdSL|RC{@sJkyX4?rCsF{mn{1@6H;88%|F)2!z7m
zoYwq?LV+%1;ouoWRz9#Ogggm1Z2VPzvhF1T|KEhQ!mq#sSF)w;Gd+GYo>llIiFC*4
zm;dTs-awJhbgCZt%B(0sakI%V<ZZjk>M}avjC}(i`49vT!rPOw*-}0U=b#RP$sBJS
zt9x4b$@tZVv0XpO4`46#&g32XUGOW}yPUX^ufy^=_^zhfT)JYi%`sVA{n#@@WO3wL
z+I;{^g2_t1r_8itEd!<|QQ?PciTRd2!w-x;2QJu^wo5BJajfvv#{fPUEO=%C(hS<Q
z=8)z0xNZl1p-du(BV+K$(A0~t8GASlebszON>-AjD-z1%cS)|~C`$=n;}O6SiGzOz
zN$+&8<;pZ2%kJ(k4yX^Av}KqYEc+COg!C#N4omsclHX}{k)S4j@-O~}jQUq#{!a1r
z7dMPg7{)FPWlNs*#t+2wjv$S32-G-&&x$V;A8PQUldIU%1o_pg9p7p6(R+8|_FB_v
z!8<ZGbD{mH#@54w{pfYtzL%%TJv?&_3%ZX^FTCquw|x|?UO!IHHTEihgs@hADIbnu
z>7S%Pg`@6r>`lJWc_|-+E-1EK1A&>kR9?}-4)cIjU{qFPD=rER6j=x=9md2Aeix+H
zBVLBQ;4ero`B4NM6F%B-VR1ehf{8|a3+7PdaGa3&1<?oim6_?;n46suED@Sso8Lhw
zJU`72o>TK4h-bJYoO`E6*}<{jik;n-`qfx$+`StM%d2rqJPo`Nd~9E^U?&`_@wwXy
z-VY&}m0#j*5R~6uUx}54hIr|bIPD+#2tAHF;x3gf@xkoqXoyc=TUl0JFXN=wi3y<t
zJ182>W-Ko)TDil~hU|X17H`#<pY_px=E;<fS|ZwqT$<kt?(i5Fk2g8Z%JW+m5TT@l
zn!r==Hjb%$@QG)qCswp@1RKI51a~}yp)KK2r?^K#{A|;^XAmw6_6oy%z<j8hk;~gT
zsty67aphQPi=90<?8ffyew-c)K46BMHu0Q4+K4>xgMW4TtlU9h3XWX5{_N0a7bAAx
z4K_3Ba^%9OYYU2+^K(AQ0(@f)_+hUbtu{(0g;Ca&6TuXUAM%|RpxK?o-Du0I-?Yk4
z9V}2+IoZ*5tiHhx31E*x?B3c^tTt!e2Ri+W=&GLFUBz)#2z7hh!_n!Q51&CvJwFrN
zA11UDKYJEiuU^E9XP?FH_RIKB|LIQ!CzK%BkGR7JWh7~^^AbT9xMJZb3%C@HzM_l_
zFXEuAL*O;0zVqPTz4-3;zaPzoWtAO%IgBSWGxKg2flaXIo-)k`<cYH7Q2nIWN7+5<
zT~NRVL9M%;>VrZHJVdC-aiRMMEeobCTt)e%m<Sp<PL+OtsX1%zU}QHJi!o1BW_C}l
z`R*l@@*LyGy=%-xW@l!N6D`4PYkOCCcq*Jq#Tqn-74FlT!?G)L^qK{r&>rT*YEx9Q
z3d8+nckbMF-n+ZorbjEQD{<%chUz?_`O&x)3zy1=qxMlwakL}KGlZw)%Qr9-)r%FC
z_3rwr@M1Yu)^Eqev|uRuK|N7S?(S^IVYe$j1};~4b|~D?d;_}LWA~TZWnM6A3-{R#
zcCEIYoSI8|A=-rUd`9zgV44E<Jo=0_fNq@iyS}rHMU3?O^DE|h^pT{gHuNFo107`k
zhhfYxP4N5+h5DrMPxDUsAsfJtwqUM!34S0LR^PL5Sm`pi19o&#BY=#R_)Y<%xAU)$
zNwZA$XYjEqJBU^3H*^Sp+^pbW{1B~E2)kqQP5%BY=H>%hQS&7%ebn*@8R|k%=sa^y
za<WqIHFrg!$UzxKvy8;zVk18M@JVc}uf?(2;`xg$-_OSm*K5(6_aEMgKltHyd|V+%
zgf5Ey&dx8ytDS>*{&HLRskt@nE?AzQpJ?0~jSbO+WsRro7-q*Q3jlmv8Zgm=@_?}l
z_#d<mV@GuQpxw_qdpbQU(I_8^FJv|r=3k!0%F0sQ6JGl`&Tdcr^jP6IB3FS$pwz_b
zqvPU%N8M+SMDNChhmRlKji3DR!+7uhT4<In{OXHO;BLZ+7@Jc1uNfBvBjE}2Uyf}<
z3<b>8G%l+Zc$%j(zqSHEnF0M1!R+PcUcB1c7wyaWHVab0V|JpPie7@tjPbtfOSB5%
zJ`DsNL;gW!8Ou=&q2M5G^FE?CmU9UwfdOfl4-vc>^GPFROjI1cLuD>}1Uu50F9Sbz
z%uok#7r6=xMCfbu70?r!LEj=C`Q}v;;3(ry@?E(5$X~F@xXJjTu}b|Aytb2m2!{fZ
zafp6IKP4P;Gj^C!M%*b&Dz%6PHfNgp^smza#&V{t84miA2ES@2EQY?vScgIvr62T(
zv{6tY3qdJqSt2qV!MXC86&CPZw69#;H~iK&;gI)@3vWFZD{tte6+@!?xzHu+j3KU1
z+aVtO0p(B`b0=jUC+++NXD^;N+!T>^q{6BhvL(Xf&aQHH)3dmS`a!euR4WuvDPy}-
zK*k?-_`*{qU#hz55Bb&rp88|qkrimr29*JRh%_u8h1XFDfZ3&R&O2>2Hi5&`5!}}R
zt@x9UqXt%EKwH4WtP5pijAy4i3xojaj5^d3e7eSqx#_WZu(1>idY_$iJl;GLzs~sn
zy^lU{IbLpV#?H=8+*x0bAOHCKUa-Q#wx=)m;*)1P@p4BzNdL_Jy~({|;&+$}+)`iZ
zbvj=72L4u)Fit$dUf+KFm8Sa1rYGxDjqT-g*!^weG>7ArcQxSK+vcm%d)q+kht0Q%
z`*r5GP3!-JimyKw)<BR7qBx%HxR&?~u=N-Jic-bJHqyv<+NvQ8MAbe+afd_WR@26S
z$Aj?2Z4es28ZKWht8iKG;^CJZOP%n(Y4~JTs!uK3AT2-XKAe{W4D8iIPsMd7kf~J!
z&M4RkUmRp0Q;>R!>oSv`;@k*&GJcjLzte@Fbb5SnD+%K7cs0-JhOOo!pAya~5Vi!y
z&G324twh%+X)y87w;kjGvzHFe(~+<Y+(eM&bguG>w8pg<j1f(cM^S!3uo--d!h^x<
zO2injv9gYxQ=W8d0P2d^_cReX(u4uwJ_9b{A;K`btYl#V>s1>t{tO-(=)6-)>9NZO
z=1dM|oO3|YrydwJ{HePKEtSgyrP44A^e_lO{A-=OrMXNfGg%N|#Dio<lSax%a4s6(
z+ue(&&!77Yb&jF?%QP;lEqc^&bujpHQzB&JE6&iHsPGA@2Moa!D~L4-hGDxX%ne~@
zG$^g!-H4@?RSTH7AUZGP*2F{utQio6g^9(GEGtM@QH2tW;|^G9HRaW@tjbMh3iwlg
z+S#k8H3-A3Uy4CDbI1VWvfePNOb);g+DMq1V9*~-7!z2px@M%7M;)MOE0k|maxrn#
zgn~TCo3Jo3!z+=%12cERefJG9TC8}2kp~}9^nxkiE@^T;wx&K-+QJyIBaVKLAcndo
zM5u<+RK8}6^$pyp9_2>a3uEGY=_aRDe8q=pR^K5!`BY+`D|A@vCI)4?5x3XwSOLa2
z1!f-MpiYHfe)Q{ND9!T5o%;D`LY(Z7G4i`mAAR8B7~m1#CB2f0{nPM$^8!%Bv~fb2
z9>y`9M|!0NH(4E~I22ReZ~XN$s8<@SH?fL`!esdnmemqxI4rbf3huG!Db@^g^J|y<
zw@apU69$vvZoUYU_B7y^!^n|3;A?CE4ystjw?`T(D^u&ulSfHrI`D>u2Z2c2b-gjQ
z1cP+>DiMo1FVn{9=BKoget@q5l#j=*0nGiQ-vS92guan!INh{nvV^33xa-H|0vQ<-
z(>LF)GhdjJ_8<hV^!)0dKA4Kvjv;%6wH2<s^+(w<f5N6)nhZgCON;O|@ZpEI;a}_-
zGi{~h{;lA`>!xk;O$gUrg{|MV_@y8EyvP=mi}GU#Lk`r>WitG0{rRo%yD<*WTaGv7
zQ5;~6LouXMnyoq%Bg>$j^(tXV*CH$1Z`=pQ08}WKorh4EA-i}5^BGtvli}fLPCIcr
zYfXCp>|g$!;_J&y#gX3!Q~9SVkib+o#tW~I)`Xv<Y|MLUjAd0f7uzGCc&)Lue?Ahs
zt)tl5X+^)=jfd;=@tybA#XC%-5K3btW8d7Q(h|?W?gx&(?jLo%>w%r<N2gcO)pIpA
z75$5=IP9GSN4Z^zH-m3cL45RCuIxs@fG`K%LL;cgdW2aBs}YL$o>0bX`0;|m@Q&so
z3(w&3Tj4DnBfQ!KMGt(1=Lo9%@MJk*k-NOC092VQoD&=xjae%gSW({-PmBPL_=HX2
z9Y;D7SA4MEBjUNZKbNwja2gxA@+k@MNvr_J-agok{r$Z-K0c4h>3JVhcTGK28R~+9
z&UF*j$yK{7SWCYX)d}Gyilo`8v3MXpaJf0Bke5ETnY)WRTt=_9Ij4@-$+y9yot+)K
zjPtV`<2W9T1@VsT+EZJzvX>nO6N1C)Qd8|Vqx^`+v0G=+?O(*v2`kwVK5=iCN~dyh
z4~y>?Q~PlwmW6j}AI7CA!4iBz$jEL2+K)v@tR6>UKz%q$a#Hn}68?}E!X|`JtTYGa
z9J#}4_D*X*Ze5?p!s3GAH9pgfy>{OM(pKllJgeJWd5|CYIWNALc4Y1|YCOtg{44|$
zUki^qF(G_cJ1sUED#KV@oT6yto-zdZ%rn9Y2ky***0MVSA?BRwfYRaR_O{E-BC4IO
zO&{NgqKf<OIGW4~FU@z@^+g>y9uIs&I5s)Wy;$RFQv_ndf6ZN{<-f8p>*GN?-Ba)2
z;YcWsTV%H-JTAL0To3GOQxtPc&4zbou|RQ3?RI{86u<h_p9@#^jVs*k;??ScOY5K=
zQ!}%%zJ5DaS2vXYSoHcQ=5Lvs@r|817z4`L3CHjq8Gkp{H)5W9s>Y|{LP6M}2`|i1
zh11jU?IY@cAzV<~sqj4n*5|LSpgO-eRiEc>m9gMr{sqAtyYKkgZ#yi<cCkwZAt*ap
z5ORVio=dtHs13clA&(J7`7$>@AM;9!He=TVb49|^=T4Q*{?3l)geU>wZ|LhBp@^W<
ztMP@a7GR&F3`-@Z=O0R61y=s*%bJU!m}Gv1Ld&?Lyyxc{v7&j{!qQ4i%{J5~*T#W^
z-EGmc?Dvex2p?D1Rt;b9d0b56Vq-p<i;J3{pfDddy|R)R#iR;xs&C>_pN9IWS|j#%
z_v4TqJDQWQvL8V(?c{zZoER5Pnr5dMcqJOZQS2z)yc17-$#)<fwPW<^+%T|UpZN<k
zHHQZE>*NJY=LCCb9rHx!D9Q&Ej#e0J{=d*%F#e-pzf{}i94h&Zx<C>7Z5pcw18Fm#
zq36cnr^jLHBjZ91hr?9<D&3i5CT7$$Yxt~CrS*(=)$1m2QsIv@^N%u6j9TY9W?Jzj
z+DBa&W30FrKBeHFv4_PPqjCH8dVKKYiD})|HuR~R!fw$Pc5eOX``<Bb*xBB4o8P<t
zQ1kr7*gZIm?cJ98wC0_nSqZL7^}DH97EPWL4VV!PnV(U6h$-W4sDpOLJ3`qpxwNpT
z_Pi8bJdBpcfP+>y_LOEzbn00B8-m8EFbM7`E3|)oZQ0}7x%wyzPbMcuW3f4JS<BYe
zP8@dG{mU++DGTe5^c}Cy<I(-q_`^T=Ui|QT-%<D(^$V1%qQ94-ZDN>DG~S<_ssAWX
zc9G90U3N_RxMTG(>c&{h7+UAu&=B|#ly#pydnMSld`zR;R&98xdZNVZ2%ankL5{&O
zu(U7z7#<ZIv@A$?L*L_GJu6G~R7I{A!N^X#0)H#O6wgQ2(HHb~Se1n_4*aGR^m)=J
z9_heyq@qkTWu)>KNq4-s7h1t$W0hsxb)<Z$w1fUzo}tGWt~9K))teKLD4%x>DSwpw
z@^s!=-rUVd+SD0_A8ht~RpAUw>I?eFCzVdg`(<BHIAuy5EmM-Iy#YfS;5g$Q<2dn=
z(=cvBZ{eXRH;Ws97x~pPV|Q-q6W|$b47^doKrgL4Q#iM|`Xl+X!;LZEP<U$k1P^#A
z+NXA)T!e*BMft@^1f&67$_2{e2{o3xys8_0ly-rB<CZ*K@*2M0L*oH-U_t$b-Fx%m
zGpL`-Al!n-^O2k?7vn2D3u7t^aI9#jYEWY84SZbpRFZQv<do<<^-@zP9DF6P5}ZUU
zQl^!3k+$Rw|4g!s@hr??aRs}Y$;*%7pRf{=2_M;sr+z#!?S)k=2$>uoiPh$GeE0F4
zcywn)b?nFIUu+qkb2DRBZlf^U-`$PlURz`DY&>}QFqYRgqTM@-=bP<#vD1xLJDupM
ze{$UKLSrtLQGkg?aK|fmq(V>K#$RH91vK7mruU7ZZTQZ(X}>Z49&&n@|FFw%`oFtO
zZ-@Ib^S0va>&!Pz7xQhxzRSD`$KPL*_5A9xqWHS?vmgKX=MeM^uOFU5aER{;>z^mB
z+4wvI51aZ1-EMrHw!hN&EB|dQ@Z#GJ*z-F><ae%AC#7`HMj$13eTPY(M0S4>k7)Iq
zUx>3D`5V}I7h@mNs_E<H^aKy5$4lQ9@X2I3C>O-N)-=<woww6ZcPFSf_&o^AuN<-%
z`cj2(%`EcmP)>$#$50xyQHCj={-z}PM{xb}FK?GFKTA@4>v6s%EN?$0J$~(6!ex+}
zU8aepB#gIR{G9{omEM=PR9+_Ln*8~$1Gxh`h!;D%5ZH5k4Q1yNX9zrJyYFso#n$te
zvGa0EL{Chz!c30mDhtA*2w)QA!9{<pl4Z59iJ~5qS~VY-M+R>g2+E0JB7r{xEW2mS
zG^%be91f%kFaasmlOF>+f69w^1kuQyM>T2^$l#&4zWAF7xCTRxaNttwI#~Vg4JFHV
z04g@$G^>nS+Ot+NVdsxQIwP9gXu#K?jq`{GJO=f-g@stZch~3b=L$On6oQ4ADHtLc
zF5|j?1U13WyR}$hq_oMBcrYTs&=(jhH&(28HH+-t#tp-#2(%&Yv^N7W3?OapRWgc5
z+7#M~G0N=ryw0N(nMf)gI0WNmC7Yh?w^^3NkHi(P-cz|*vBjz&1~nK2GkHpri5}@u
zHeZ+uQ>nTjydw|XxxAVQ9k`lD7-;e?hJr~!p2Iz6M%VWXh*`H%PxS+aFz1xdfZ~y<
z42XA}DXpAXtK5Qt%Pcz~%}k26%}&kw7)W8O9LXJ%?>FfZdEl4yqlV$9hBB1$-!Az#
zWVc_sIK3Jc51VvwdMcm$Q3f+T<(+BBy?m#lgS6{A!`PlL|Iml|t7Waf7?R+ty7*nw
zEB0CrO6A1@DnIpgC~&tLS!*7=iECdwQVu?{4`!_Vy@|zdrk(a7AN^1);Ggu9Z~NH8
zPP{Wj=H*8zxNY@zc-c#({lyP%{<5{a(&Bg&0vDV}20!f1PexTTZE-l6j9OkpAJyL<
zH#y`>Z>No&&-A}mvbH&H2B_u7_;vGXL-_pWH=j4$25C`dxsz87ijUrdR;qP$96pD?
z)4w7JKKRjoJfu<6HOg1JUi`WmL*Lbgxv^26#0d9EdzA9ZkcP{mcX5(imZX03&WB&d
zQ6gjwiJZ7x64A?Jp3_LT+TSqFPfxnkk9_k=oW!D6O}X@!0q!bm!$v=YzNH`M#>Y^4
z`BM#UgeuOmf*(%X?PxDTIfGyR>R10XUH<aT|Mma&fB$*0{+;=jCi~xr$6X6V^qA+;
znCKllV^cnckYkbBy_0Bl`?0f+Fu&zzr`7Y_Fui`Ryhh-FLIlMalW&X9lq|b`xHsrZ
z$T2lNCD%Nj>ZE_<G4|OPFXDgy<*&UXf!#!$7Qmt@lpY+p<dxYT`&DL&>B93v`Uv<@
zWU&AT;X^JufydzZVM61Efsa4{!Q)gY9Pg5rzkvsKm5=c72UeQtsY<VzX+5*5YlHiy
zI5L_e2m9tNr>4eZu{j&di}SIxIOn76*m=cnp6SWan3H{OY9boblfE0_i>IH(fBg6V
zE{-~F@x`ODDc<@QAAe#c(oAzTE=DH=`(qz9!yRYvC_M0R7+`|3ggc|A)Ml`={gabG
zxdXpJ)!BuHfc1-)JF%_yJ7^!pDLb<!X2dI@4B%KMcqRI8x#;{-@l=NWc2{lQb(~)R
zJTByiGK*F8@OLaKfLDSC@-73xa6~*EtCC^!*bM|dV_w8@X(&iWt^_BQd40JNtAf#t
z5ar^iV}Uyg(v9V2EH|g4A=plis(p{!aqIP&h4H=if#w|h7SOPRaDQhrwqHDtSI>SO
zmnR(`5ymMKjroQJ`~73?{u;4Dr!hAjiwlidTUqq|a~y}av9{!u(X3vkU8knT)J|iu
zw2*DsoE?wr^CK0oFGN0y{mo~}e@EpUkNFAV(lvspGkJC6aDOMR&W~e}T~<m9p%$pc
zePHJ&NA`RE`6u!5pa1^_o|Cw@z8tfvZ@07WV~7wOvuhCr?_sB-a<ic9O0bHUK>^2I
zVl$KCnXi5M{I2THt_BMW1?SUa;UM`a9r#@MaCTUpXx`?XNASAhjlpG<ez!L^VtRT;
zI5Zk}@7<3lAAA(w`Of!!m)u9+`CfeZ(MR##dyixNHVQDoOK;+=tgXfZM|#aRH20Vg
zc*XaNmjyR`OyOZ)_32w^NB`s~Pj<*5bUZpfjRX02x~ik_ny1_22u*ezjl|sSd@M8<
zT`uN(9KrPR)wYiW<XAQEoB0lccCT(%niekOp%j>Fpd3N~&Iuo>jNs_WOO*))U&npb
zcfIvblzvNTx8qoF2Zw<V^91v*&;r5Gr<$C}pZTZz9CMKrdb8*dAugn2Y(g~i7CThK
zihJ<ykcF15!`Ry0i|xGwALoito84~a-PkcB7`|3NA01OW>YMo*I~!3jo~Um^JGpy~
z4#k@zPEi7K_Y)^v&;}d{+hv!j`!H$gQG2WM7AkW@=!V*EeSO0;ja_&3{w|K$JL(=<
z`Na+|0n2k9gxRd{_fdiB>z>c6Y?P0;GB=&|eBIxWP<Rg70>uTwaJ7-=V+itdep%_U
z!wNsDaGf-Gpu51{W75FTlhh{UUD`(fBx2vMGDtj__li>a3zNdoOo#UNTjivl^b-r}
zMUNH2VX3?QRW{>?%b7>FQBJ<Sb69xEJLBn59M!DMSJ_>3X~8||a8IO<t<{{erh6$E
zp$NAyT60$LhP!q~gl|(Q^ig=JK8zLIGshega55(Y-rOm7d!?aqXUcP56kIR24x*)T
znq$!#9IGeV#bONlgzr!UznRy%eNd1xC+BF!Y3~AMKFfk(`Y5u1rjJ*=6XOcAcfe7o
zJ=X)A1#bQ0<Jj6!x%XR|gLBL;JA!5fn_O7J{lUk_7goBl2!jO)D0<B^u<}{;W4F+;
zaFkt7EJ~oQeV?Dw;0RIRfUJOS&*B(%1GD&`*_ib`n(!I)Pw&!Jxqy8hSvzi;;Wk!Z
zMZQG2k%8e(JJN2n2k>M^4ho&>?zz4IZpEf>8am-{;(6HV#qqo3MR?xq8ba~j@WKiD
zDT_7oSX~wwv*4I96P|>fQ*+{d=&#=GUT}`egJ&)SV~St_&cPGae7rcNd{cJg6XCPl
z2m^}+n!}qg#&O<xG7sNx7zNATOuRJ4>A^#v<<XDYfj*9c$cj6)TOI`oUO}JfgYYUT
z7ovQalnv#ZIx(IhTY^5bFo1E8#X2awSWs|&AzqV(1;AeAav7*QeNnK2_9ELs=|!pO
zZ?qE+<)aK<<m5OBE836tVetWx;3dI7@I4imDx=(eye|s4uEu&j*irJ={_h3Zz#Z5)
z_k@?BNXGPA;oz*sUUrosYeEsujz4Iq6@kJL%CWGp7)#5`(Og(i`KAD`aS#3%St?;^
zJ7wlNr3{L?Uk-5dN@EO4x?{nzqw)Cg=v1&${<BT-rpvLa^tlp+a!pK(S<$_{y%Yb-
zzxlWE>(8G>TVoZwtfBF<!cFR!ZNzT6<U^rZuf>;f1S)O(E|W0GH}$A*9!&9mry0Zr
zo?j7eIFXumecT#fl}RyQ7UrA89X8)4{I^Z}ZS&j4|En@t&aW=(zy4Q$@=r(ptN;9u
zuU{)LL77~@&m;ljmxi>}4V%9iw%XI^UBZEC_9@@EGbQmvMV+S_Qn?97#OhZfVf69@
zl=49n^pU0d663VeTH~e#XA;6%zOk2g8l{YvIvD;AEMU;dDK;I^LHVIP4o`#(PPs5C
zyh`$KlgZqCW9&dU?B$WYyp)&HcU{u}JMy$YdD@L%h2gu{6EaT0g(e=fG*@HrU;O!(
z6FFo}FtQV@rl&Wd(@&<>zZm=olL?iOnW)`an{pur@Ipv+smatWO_({FWK0Z2UAc07
zdF_E_L`46!Cdh1M-QL`cjwT_T``yt*;YgF0*CMKzrIn?aZOq4|1_~HcR!MNiIZ7E6
zV$)(mU?9veieQ__D;#z5Kq11<U_tpHuD(!IVHunllu{|CKLx=}0p0EpfAaH68<mB!
zGq}N6S^<OI1F<Hia^t*wCY4<H`Re6M6M43%|7DqMDVm|QP<ef+$>61}%HtoH6*&Sg
z>MQ1yK?b-q1p6^2!>a3~n6F6<P>Y(FvSI^f44A;EaijzTm{-D}G*prtC0R-+IOOU-
zuS8N>JQM=~3hfN+VHSZautwPge1W}J^~fE(<{P^eN~$-Lc^GFVKlQ>-uVg9ApPL;9
zPZrvecBL)`nfi*ENIjI1y2CW#mclEgrL?(23<E>MrKs%sqx|3&lQ&i#62}+PYOrRf
zo3gR7QcK~k5!T3<77`!kNBGQ+K^POSaA&|*d{x-xh4H4B!kog;fX{}9IyXu!;7xsj
zKRXLw3)keLiOHQ<Wd-Q?jOqk#t9B4H#iv%ZIe$spC}?7myR*?%u7(djvSbjsITQh8
zoPz#t-uB~#o1+aP8rBe1@~p;j^zE4bNxuPom17NE=^{I0JgyX!%{YQfZVPc*85c{Y
zThpnf#JJR1$YRps-%<|KSPZbk@aY7x%dI%uki8u;PR$FKggH^m7wOBhWMktn<R?e6
zaKw^ImOj(1VZ<)mrvxmfwnT9siZ)1xe`j1vn{leCaI;Zf>8D4N$@Ju%PJBDvQg`Dv
zZA_XaUx&+3`lZwvo!|PMKMG-Z`=c1SQCH<dd|+JtGab2=D4A>VsqIFf0#7C^cBRY}
z!1x!d4Pi8Hxw-9~o&(%dytEN5>*z7M81pNGPZzd?mXW1qTeHmSl8D}NDY+yr;anmU
z&Q>QNQ_C~JIX+5w!j>}O2L-0-(!rj#CQJ#gdf-+v$@DUe{rpasK{ABbQ(=Ioc@Elt
zD8is@ylNiBwWd8V1?cuuIP$PR`H*ZWANl-u|Kh(Vp}!LIcRRkmp?N#og1r3JBGJTC
zQ!^I8ydJp~ueSH%)2A<DXSeOK(?(%58k)W{1}8r^=6Ea1OL6b^N-WQh$LJ+@ehG^Z
z>}nk3B6@`1E6v%MRL3|wY(;B#Czdt&ZXcb*|MWlqyLh?19}7z3{U;yA?Yj?rPZWwT
zvS!B+cYGZ(f$#OLkb^oQnDEid;_tjWL1Qc{pv^ZT2tg1af8Y*J0uO_^#9d=MZxA?z
ze!^vV5rr|YG%4N~0hH>Q0vnF2Vb`{JTGeHCZrW)f=sQ2XP}*ZLZ=oEk)5p9E1b87h
zzY@>EPIHcbrb6s4XO-;#`fvYT{OXq<$K5-3;)4&LL~~(D{q$CxtIjCDP&TrXot5|;
zG1*Z17K$rj7IbwG4E2v(H+JdxsAmxr_*yUgx@4i(u@`x<Vjo@vz8Ri~vLTp)mq9??
zTxfd1nvZFwUJCD>2;$Q)2u7Bg%~)M&co*5cV8D@#d0&s(;OxS}N;*HvN`#3*5Fe4p
zf+d6~tlZ`(F7BpT6Yt8=Snd72c=p++UU5vjAsC`P;J-SZ!)UkmJfB!!U$-!dU0&cS
zZNe@c6r~?Mc`u&4ch7D9T6vydPsHXyFLw5O2|wZV(qi5ZhOo)Uaf?rzVxbcS77U<a
z9MQ*a0gf;Nw)=bAvAex#zM0j{C@|O8)-5>d9=7A@(_hEV-mdt~>)^Pjk3RY+9zJ+r
zCFs@lsFf7YpFNM~pMNTNA<Vpr&p!P;8qJ0Hqd)waac*a43zjlAZr_f(_a0bDfza7)
zp_V{l2+s{}%*-!DUvS>uYgv)PonahH2;a_ee6yO%p#<}eNwoxSi)vTemV0_W{p@M<
zd)-*!2z0@jqxnAm)yJ{Ayb^cr-HAK5?*w<i%}lY=4@IxiQ(tF1K)6Ubk5!gKwL$0b
z*oqi<TMG+MPCRd97ZrUGq3?WSHk#@u2&SPN`<-626z_<mC<Pnf#9e0eON7<rOI@sx
zLD-29PyWmaC<S=V@rleQ>6bnZO8pywJoCJxqn?lcV@C>nJ!x7XFMACO!GmT&1HoP9
z5T55KUr>SLJn3KNcN7*uE4v-xJt%{Z)j}b}!cm8zpXEs!^exAM7719Dhmw=iL(z-^
z-SZtiEG|^xa(GqREXunQZZcFePrjx3GrJg8msjKN-8-U}ONK{Vb69qKp=d!dPro5N
z^9bfX>|8iixjBXoK`AT$p=k($54%Tkp!jXg=YiKL%2MGyFkzl{O<h#BaVshijKihm
zuCZKXkasC!NYezGn(7TcdsmFo19z4eH9u8<>xurL)U9EtE_TC7b5rCyq?L0!%6aAX
zLZF+*ofIGr?6gsqBg5F${x^o4$Va(3U%a73$<b%njHCLhuS`y=j|&HJL&y#uF<&Kw
zbI}j!QN~mnbN6%IsfbV+<rIom%1*tQyCUmIT8prp`*0T2AMf10r#><*cow+}VMMQI
zr^e&Q-~Ui!!g~Dj<4@wxe(|yLn~0UQ4d0bVTV9LqdbbI?^sJb@FnxaS{(5}xgS)Z1
zFcUq|Ms{}m@{<?7Z<6E59zTBQ-C-!tH#NTh`m-<M`ODqt_BDnH@4%&TwK;QfPVIT`
z(L?1o5qo=k-kHYzfyDo#pZ&;6wm<*nC-LzopZgv%U<V(7BCXwRtNqSnMKt%r4<5!3
zKYAFCAKeixniIZVD$jlK7O!G+XHT@1;}P%ph(yLWVAmFowmX8A#xm&Y(jq%RIkHiG
zM(xDf)7^b`p$Lah&o%#McjcV%2O7n_lC3s78QE!;#~{;g?1V+B;vH22IOv1?i4F%Y
zL1&rs6PB?Ar56i`+2z8R!QFZ+{$`i#3FE0S13m#dv#`KDpH20bJZT7pCNz`dR7r<j
zJ^KfHmOHi7_uP-bC1nrp6E2TQ<X*UImGeZfNd5(d5PbqXc0WY+Af%$t)4!o9;8`A<
zoJt;)sNf}d=l;fz>VzDQOsOAZA2JYV7s#FTHXmpfQJ6_kPRc}{!+nh1%~fGY_*`&N
z`ncemvJmJV;Y8y_r&ypfDY}(PW8$c<tDNLRKeVjZ23ZgbXp;Y`<%5O;Tf)%~Q$D1$
z^jPkvJVnKXCuLzYG^UgZ1s^;B{6Lk>P#<8=xC^{H8b4S#A+J<aDeS!Z4z!UYc3a{L
z+TEVhB!3>#pf35S?gBRT%6qRcgw4D6n@h{yf$ZIJDyIm7!e?7i|EiFrjaW3sj!YJS
zfYZq0I2x2bKtASuD}L|*Z}3TFWhW?jn@V1lQM7+{;yTtB8u7v7NAcj^ZS~<<g;g1^
z#9JSB#G8B;|L))Zuj-#KJ#wKCd;k3p;|D+ZNi>$kf2oLveU2T^U5PAt8F_uCH;PD=
zUe&=MYskeYY4=+N8^)bC>~FODn|_%3?f1ZI$gP;S<6%5TzhU0Q;jc393OB4Td6)U7
zX=A=k*ms%VtlRIW$$EZuS-s=yCqMr2&k5{~RN@Rr_G(^z`wMqJ`iFoy953EV0Gg2r
z4E$<%zsW>zhvyS_8+Xh6#sV_)vJZdBNaPm=Uu^QxD{Xp6yZqIBG9<n@QRbROS`cPA
zEFP{?iJ^ouoIjlwnfbGRV}{-?i^JHWM^X6ZO^{4czqLN;>Tqcryz@uKQSu`@JL%On
z$+Y>d<#8E!u<O@RFmkXU!PW{Fe*Ehd)vg~eobgR?IqeKpzIle+F!^0lRXhZ0mm?x+
zFzp(!QBK_waf7*qkaDCU=Y$_<QgzbP1hCzX7tfx>%csv{=jCR!b`IjKrvV%8!pY`j
zgmyE7sJ3LlF$1f?ipdj`Mh2N&VZk8AHw2ymQN_Yf@gdF-TI?A<nQrV58{WAx1vg-l
z@hApL7>>PRFk<%(t_*^FC&R<8ATjM?NMKlc8n|+BBlhp!xDE6}^VxipYzX^yILDVG
z<peH*pP79=1%sJ}m^}v2aoNQnF0QP`^4e-N<Zi)(;UktnZzg3ZabRLdL(v&q?GJ3u
zm~;t_FzhI4VAN2C62iq$JTo?>ooNHdzSg9PHllobXm`${mxq}(u$xJR@kALiz%x;%
zJt@RBd3n$$Eya-?=EXRm0bDKUa$%@Lu7pZ_B`kdoWhsoeu*>aDy5z@k60jaj*r*p1
zQMZfYQa)a!1%qSQz=WyZcu;N_2_BN94a`94&wWx(oY|ak&55*%V*D(bBurEm_d@j<
znr?amp~j3EpHzHh8)2O4HL%Pi+<$SxZYLVP+WEmhziL=Bwss;lhZ3$hME<3d-qaAk
zK`8rWI#{sV<EFlfGbjt;MzhYim^PQJ`2}mzFX7<FZX{6s_2#qG3wxHzLF{JZkNQa*
zQp#^Z*xs3kn||Qqzw}I(w7SmtgRdom+zUK#%i;{4`r|x`D{Y27f9>lpK3t{}t@>eb
ztubm^)t|Sqzy;KIDP#J`CeMUb*5CML588wr@#9A!ct|Xx`uisB%*W;Mw;r`)Dx7$#
zDJhZ`VPs0l^{#o>A0DNuGG`kSO&0fo;V@}3jSOR-!4K~mC&LXsNK}buy0qO*1!+61
zhPv$vml^1Xh-o$aa8kHZ1O4i)APz-5oJ*JtTSI2+VI0IidTxr5#&$x=gFyTfrkWBl
zVUHhS$-UN_cVZBqF7v=ny!^)90EIw$zgg&IE7iuH5eT85w3nYWq>+{(=Ntc<w5c~?
zGLuXfKe;<;*~^=}GTuo1@>d`Kt>5wG@#U}C)G;m_Guxa;4)(+DRqGlfSs8kA#-w>N
zIF9JW)_(lz(--k-XFrZlSX_c|f^k@LWQ~LD`axg~8whKH;2#A8$8c@$bu3K#V!P$z
z=Gb&|cFrQC3-f*m^cyoApEx02bTk&0mScH!#R9}nKl?mB`Sqv4u``S8QdnJ#xfzbj
z*2oTigLCu@6+Vv;gkJ>KrO)85v7E8hN9MsB$sJyh+9zMCI_r;dA6~^m0VR$Qnq!YS
z+R68CvHPsSMgC3kQ!^G0T$~=oVVgU7y0Ns>wBQonb4vWpt+<GOuN_zCr>f&baJShc
z0u1;m0SbW>i?29Rtji_+EFhD8X1*DB?!One@7#$E@opR!hXB!nHN~41ufR^Fm8FIS
zhaA1fJdaiQ>>xqOL0iyH{Ude^<^2#Cgl43}@o$Kr5Q4y0+x@t2p}!#mEaaX-tP#b(
zvLb`KOf76u+0XPwpw7-21V8LB>78E1VgEdiR0fVuJW)K3)=H5G4`Y*Hpa(t`h0bDA
zby;2zZ>Dmbp1K{m59x#V9>$1Z^{4;wKg9p?Z~l$w$S<rMM5)QKJm3d~<<6Vc_1l8k
z!?<_<QGE2B@5ZA?k1V`Gpue}@if)hl%#OrwiPt=3k=}{$2mb3^bDN!b@%&k|_IIOm
zup9kuTl|}F_~KOGd$G5@8C$P5V}D<B9l@!&xE%K%J&yMue;6Nr_+i{u`=E?^s(I1o
z_O_L~&E=(7T3r_oO~=meL2PNBgo28lb=+sev7gL+*4J0#_T4+Nu(0H#@6IpOm%t6$
zL3qTGhaBU`G2P%ILVDl*A-J<x5P==zw3;KXMmXB?TJ3Na$7<h+=~*Aq#0qhapJc}o
zM<61Y<!GkO?cMn7v#0UJtF7p;)2e?KlQZ*z<=wdV;F034c&8_Ve|B6VFywfr<NmSx
z8z(HF09RHjclLcvB!y4(Ddv0#<WTJ7U1fO;xcbz@q;a03)sB?sYsG>3)II@?qL0Gk
za}o!~4g&iWRx>wA;V*(>1bi043QstOAEh$F$o`SqTMszPoPhFro&en_yg}Znm3_m)
zP~`{TP><Rpj^~4-Cpotfj8#77O8A*imW4S94@x0m!kx+Rk<5RYANd$A<_${Ue7)xn
zS#A_$)WeU;f^eJTO;1h)XZ1(EnPYGV--T!@vnOy0I>XV0;KtI@qK}(fTU(B$MRvI~
z++HZZ$OkyElg-C+DG%Utu(ub_gs;z@J&l(yUc~P1j$qbxIZ(vVcC-~-2=&0?-7v~?
zN~kOeM*sl-^hrcPRP+a&%7ul5khkg!E#YA-asBiLH-RDcoLA*8C~(AMgkVBjc>X4f
z$Eadxi&T6ieMT|CBL{^ST>{VWQ#j9Ylmg}Of?9u9I-lwDeUpCDd$*tZ6!maADCYGI
zO+=Z=oROV`OUo-UJJ(Qtxlp1SQnNBp<)EK1k0V`RM%{ozX3Y=sigE9*Y&7NtLyjxU
zMVaibWibeJl_HYAcNbYnGUYo1IYM=5ao+gF&L{5JVizFyG2UHY@$r_(nAmLvec?WH
zluj%xKq)mpH?4Fyf|DaJS)hW_NcEeZl$-jg#$V_>>2kLoutH#t%*J#_<)k01t}ZLR
zhUpM@dHL8=^~XCK>%zem*SRW_I1(GhpOvgC7rT_XAFS0u!HU9^-Ax)PHC`QQoZ=3|
z{nnwwGTz_aSk@Z_7E0|)-$}>L6Be{^$D*i&=)<=AGGk=LE8<%Qr1GO!sdp!F_tEY3
zOp}Eu9MQJ2ylAC7?a$nv-EPQ?xV`{Id_MHi3*nQ{Pkn<$6L;@!#QnSLKK_wA)i{oo
z9bPEr=$FfCue*ZDebM(j8*A#P3vPS*Ch$fP(VWerBu!V8xbe;iqkbp4Mf=fzQudYj
zj#3YLqXHYRIMpV{amH%3*SPwqXar-1VN$~KhS#y&MRgccALf|TX^j)~Ytp7INQ*E@
z!?W+XZ{V-8r96Wqlm6l7egth*7+^iC`h&aRl=Gv%(znC|Sk|gEh)4Tlw_H$oO&!61
zXrB9p-lR=AXd{$VxsbjgcnhA4jnE^;6=)tbi+r4ZmJiYmOsN+;t7segH02}?cGBbC
zM8XgU*fS3BO}|Dl2c2Mivf}aLTxDj_lg3lgH0nY<xhtM=T{$NW1Yu-{?_{h-(ZzAM
zEY`q2Eqw8=P^IBTRmwY$d@iR9seiE`ye0a*EgF4rAYSH>V`y2Rfs9OfKsi;%q?K83
z$Rq}LIAmxKHrkZ7rhVp=H}sMn*_%79Xo&~l$Yjbhr!u0*pA#<%?GpxykJ*b?;%{Dx
zU)Vo5^f<zuw0DI6b2D7n3C-7ol2l_J*=8|{r9LhXV}d>@piF%aj6O@KL7W=IVX+tY
z`eYg<&^t`c<0eKvMEQ;UhYWu8d6)UBxR|&7hsQX~*To+;Ul;c6y1nbce49M~GEC<G
z)n#DE*R7xa_$QjstJo#1x^vnQI*<s&gm0KJJgU3H@`GD>%Yv}NyquC$NICge4}O`3
zl})y#p)-{vhteh<oh`rYQqr<8&i{NnK#7;&cRu8gQw9~mk{eZXeH4>bRfHp+`z|SE
zxiC2U8;eZljbn+Al{9g8TzxybBRZ@Fd?lXCW7rbjDH0fq({>E9!$CjQw<K5G6$Xpz
zP#rO{IIrr&FQw1CoIh@E!!$KqMkC{r=1|%WjlG!4kQ-*T%;k^&Oh$Ps96K+ZwumNY
z+l$ljAk3=KV@(wIclP3o=TGCs)2FeswH>>goAL75tLSP%dC7tDsec$4(uCME5pljM
z1A_+x5_d%9N;DJ$IS4Z#u<F%ELW*dkFkq0fyTURM=Sn)5*;Jx1z#8x}9x#*+`FQf9
z!4gKnwRKrE$&eetnUs(o0}2Becb74tfQe#~f$)LB>YJJ&kbgttUR#U)%?%lqc6%c-
zFvFt|Qs6hqwtgN4Sxti0Zm)S2wU4;~eu4>1S2Dm{xSA6)aEpN)STb2)btOU#m|I|o
z0Z(b$y4u@qsE6=%Cl?AO%#0W=R`XCi+J<~#47fnLG77*N<_3SSW`sc_f`DY^Pu03q
zTe0HO6B;oC)FE*TD-4EEP=l=*K9$`L_A3k|54BSnDJOrFn~9r}G9Ivclhsh%EyuAq
z)TyqxK~YCJnH;kM4n~aKhu|P(&+QrDxu}N9K;66>OyR(5aGy9lE|)5og0@<>Ycr{H
zVnEpeXJwK6+$T#6%2_^4tN1RT3QPFJgN(yZK9%OBQ*|d6q3|DeD&cLlOB#pqkK3Sr
zSNcIn|5iVKsKdaiS{QHaP)f_2$g;~qcn4J*5zIewbbqk}WRZ3(VX+Xm`Z>k4+ntEc
zM{b01@QhP}*q;m>3J3o4r5`7dA975$d@_vwC`+Q3%Tq#?Um4;#t|MeT!-fzy;jv+8
z$^6Z4?2e;{y8F)#@}ernZGzR5@|TxgYGhKRtoFC3-3Q)0HaH}Va!j;U4`FkR*MIrO
zhRHJLSJq%)D?!{gCH_r2m0FUI@Q8f<jd4clMlv<NU*)l#EUMk!2@l{xk{P(>TfP;F
zdn+@3tw)wSZS}Xh<5S(qFuy1_R*zd5)Kzhod}Hcc##VXo_HPYnH~ZfF;%0c@!nZwS
zvXrTC%F0theq9gOC(BXe@G`IGkAx47#NppzYr0tNo=?7T*PnynknLnV2IhR!46C1h
z2#&%2xBaNYw<kJ(Ri@^9(^u0JjPak1{CyXK#vqgu2kmZbZXf8`kFCA7Cc6iE_M_cp
z=Z4Z$nCw^dO#NxxWM{RHztWQe4)KxV6;Qeybb8U&<hsYI(AT5F8+eS7Jhny?c*bck
z)5pXpZG^fA22qHi%z3rB9osuQnxJ>g3&2aTa0Y__AD)nL7sV0^CXR06NG6{wAUA{v
z@CJl||L{1XxCn4?_aYqez3@Z6V**$pyqOgb#ly<$nJJ|`8IAdwSZL10xaK7<pMMrl
zKl@d54-bN4nI1m4XF&`{HXiK2Cx7On{7{rN=jRdmsQf52F1<6U-#_#HBpfsLTCik?
zR&z1${<$@g3I_x(7JiDK11|6r2rhw5bAHwWZ^{kdi=dIV@jbfYZ%_&Vzqa7)J5p6&
z^K0U1QrMrXu~YD#$3g;^y4Z)EYwcDm_V-fp&^kD<&;{iOe%!-^fYWy@=}micmlH=x
z&CJfny!cO6y`wyVpGWD!5sV3c@W^Ns6XNeD^!93Vl@Y-pJQ~6s6tL`ALy`34gAd~m
z{@^F^-S7S|-hc98Jb3sdHtyUPzq_EP5!1?(`;Wj!?tVHxJ<mI%F2tXn3vYm(V2Cp0
z<*S!o$<6&QPoF)vz!iM<TtoP!DS-u^2#-0AZ{*fw%rzEbzPT84^9!oWd^8))m~YGp
z)+unBotuxl_aDTE-~C>^_xMSytgb69gp5<Mu(TFqlXG#V{&RdHo>%TGt4s0W2T$Th
zKm1YLyQg|6&Bo$#tgLM4S&K#S?TbsxF(Y0*1^nu3?#m;Fjql!dyNM?kIL|J`PYYHE
zk*i`FK{joVz?ft5=<9ub@9wo?XTR;c#TpAMnn%qDM(VF8@WRSR<>R<Kl+$Vz^;eGT
z<G3g8&Qm{C84$R#+X#XE*%|qqdGR2EZH|6IXbTTb-|{hOf)9dN?i}Oj!<Og__+vVy
z2fSczNS}niHqXx7P2S)w0(4ffGgm_pjiG;2C!ajhX&YaUdOgouFu;p+5x`pDEIb5)
z&>AbwE(w>b@5#gP)jxR?j`<F63l)W%R$k#J$Lvp{dE)z1Q9}GtD{xUxP!D)yqq^Md
zFzGZ&z$WJ<sbB-{Sq|@<r5{0igxBouTV7p@)s^+QdslR(xghxFi6zXt6s`1`<2X=X
z<-Y2jo!vN4pXClQ?%?jrkDZHM7P9W|TbN3G`aST@xp%ff_QA0tT#f9%YCA9JQzeMw
zopCBdo$EV{`g-n)g|0Fd<UB9OqA~|WP%RpT6N*7h!k+nG8lVo0NQZe<Dg;D96$XVA
zf^z(vtN>e~`L;oL4P42;@^Wc1266@FW8{Aep>>fFq>>=<fi{OwAq;NJH{GrXc)=Nj
z$jBhbgS1#=L7lD8V$P`kX<92dqUZr{8E?2}8+k;X8^WPcXX0_6HcDQU!|dV$S4o3;
zIEu~t_wM<qQI5d#E{oCE(U{|2uLTS65$Mwvo{Q3tH0Ch3pAdN2kuy81zK8-rV;MM2
z`+$Grj5*x*R@xZQ<#CPd63HFY$OPDRv!VQ#mlxG@)b~UW7!!c;%F43GZVclK{U7|!
z5ngVpA1AP&h~E{SqJU-m8x=mxi@qX|NBM!`F0L;$#^oXpjy#5bp!A`v>>%ZCOctkh
zI<5HSFMr|jiE$geLrKQboh-^YD&rJoM4>F~^+_-6d=_xb69@#}g-AHyPajq5yZvcv
zS!IgC+VX}F{H9`kb=it9j>_HGSX2F1)mIm+Tmes56b${NOzae*9iTh3FM0b&U6t9m
zE%XG2;GcJE4)9jcGd?l~q}_A{qi^~Yv;u_%_z1rM-Jl$c8c$gKfQ*5$&-b$`zbUmd
zVfY5;sV8uyKeB_$_0%`{lP3VkER*(Gv7{$y8RKU5-|R>144&4wH7`6*1qyv0H~1Oz
zSKyTDlXHfIGjtkbT4;dFP3b|q!DaY5pSDBZ3QHRBTa3euWAr1AzjdEMIi+?4{){2i
z2SZ;)Q8wd6OeiW*nvpi;;waYaJM;_UGnSAZJ3VPL>@0pEed=E62W>(=^i$)S@8~;m
z-vwYO4dK&J-lcL>INfZBr{E;4Nu_t;1tnY`1Dx5d$s!KuG-D~W8(Pg+PWmho@qnU#
zQ4rNE$4@V&5IqKN&@W(2U1%d{1;-wmo+&(V<w!^Hg&o)Iw&W;Pcp><6j#4$Qi0&gp
zynTlq=yzg%b}nA4{_G4}S(uBv>x;3mvS@`JeezP}1{Rpoq<#Z0kUZV^$#x$$Z+_RX
zNxL1Rd^;ViZ-wXotJ6#~>~>=V@A66S^m$kOH_S~mKVKXEE%R-{f7`U*Wk}(7k<M>p
zGXJkGgB4#t`^is!4)WF^9wg_B$ac|>zerFIZ@bITKL%m*RleyEdh-o&8`vum%I}Pn
zZ~ql9J7K|{O#BT(CZW6umT40&OP%%%fIai`A1lr6uQzT=HwBjZx1IXr+pDm!@MP?K
z^CC;V`0$_K#m#={mi{<VW*NSfU#)-r9oYOTE+tI)FD66s14t;k?bJ&4fg6spIGxg2
zEJ&suN}Oz?G-=DQ#mZMHLrLF`c6B;~I@p=_(04Y8h^ClsSrFFa&wT)!o3G;OuYMKJ
zKKUeGeEuwUSP{4-0^B-`qr;=P=xZX#<j9qA_5&l=ge6{@M8ZTvU=ix0RA9o+)_ev(
zCPP*<h%hslv85iSkcos9Tympk!%R6$6p4FfgT5JXTzfGi2y4wK5QfKPOE)V$c#9x2
za6yD&!g58$sRoxruZYV*_nVows{RcO^#MZPz<iTH3PY>wr+mxKDz53tm|bYb;>wC9
z&I^LG7+fajf+>sxu!OOquO@@TL|iaceDElHsB8?bdb4`X=OhbW?BwzQE+&EYAq<Qh
zj3%qdaHstkfKh_cE)2d1V#uFC7v_rUjUgaYo}|mlN3XC^Tu&T`FXoi|Elg8f<B8iH
zrUa&1aTFq%Ere+~p(SqZODO3$0)*Xbd?%C0<R<aQF!PJ>**2gK$4B~j4duh+&h>}k
zRfx(+=u<dOKXWndMw--_^Pfo@oMQzUK|Dd0Km8Jg8+BmyDa@yjKTuyF4n%?%36r1y
zNLx;YSYYuppqs@h?F!>G9AjW=EQc#$>`;^Rcb1t?hm)JU?VnMcgX(GTjHjP09f9(d
zw5K@nEtp|R@BH&Op0?i9g@1?Qje7>bF~d_g|D{P=y4=9aPW&slYoLF=Zj9SaF$QTn
zbSbB-bxd?P;BWT~Rm0F<a$K_Ozz_--7H&AYe+iHMCT&bDGd`r3rn+Pp-sxAure}UR
zNWQat=~;s45w6-VAG<m!f26Tr`KJA49ki*RiWBxoyG*)fnX4OSIQRN%o8o6Ss*eDR
zolo*&b!dV`_EO;VFM<4>_H--tDt{S;Fe(v0#bop6w|@zU(QAMjyo&P>l>NAL4wDeE
z@4%CG84g>1^~yMUYE!4zo5pUyI|J7^{AL|8btRFWer4%0*E9~PpXyTKz&CbJ`>H=4
zz}`;!Q$(B+esXmCr$f3?#|)d1GQ7j^T{5;mody4)zhC?<-|=NQ{WTc!9nKq@d#Wrn
z8FsJY)%Y_ysfl)s$E%&g_|>n!h*w*CvG04F5d3TWvuqO{vG##9brmvvUR`mV3rkX1
z##rKGel*^}gP|Ou0E`C+(z$Pd9XtpInbdPsns-iUTt#rm_=s?QuY)pfH(tDY=^Z>r
z{bMUdFz}2Blx}&K>Sb`>3HOe$q8nb*LO=5$@F(NqM^#Q$V!}tnpeWK9EWU-EUGO$3
zIGnH$lU)x};uj_eLwv^(+?3DN#c6O6|A$ZBkMDf=*!ImAU&Md<5C1;?>_7c!wD)#n
zg&i1k^X4~4Reuyb@H-q+!$th;4w~RDAi;y90liXPyfs47-ccUUba>RKF)f$@dI(>*
zw}`uGS*$~SfS37>%X2T#L6AZ?gf(8EqxcBleB6!r!&|Bcd=+B~f@^rU6x!0)SqvsV
znfQ9(r)X!7S*ip4B78FZ53A`>Zmp~?d*=zl)LHU?#}~}Gdu7fFL+;t)=sLM+UZCO{
zk5CfKz?;TZoS*c?Yn_UBI`eTpuQ*NOdNfw=+>0mQ`F?z)=kEPSv3}=%EN|Qs134A#
z{zV+J1Cc5UmMCChTiJcq?w-Wx_)LsV!v9W3V{u8}Q*m-N8vQfzcO&C*@mjp9(pg;H
zh_&1IVz#-Yl8k8{#9c}kad|zia1$!Wt>~#AfV04cyJx5u^`z|t%s@HG(NqGFU=Dn!
zr+D143Ez*lt@^OKn%z+sf+2Ssb9c{!ySIgBv;_*LTUOMubL(7o6<+u&91)zR!9VpY
zx*|L1xW9?EnVM8vDlL?I2)ulOJ@~6O1;{9>I7U##j*f7NJA!!HoufEbJ%JY}!tM=r
zr)}@-MN4x%juGV^x8qa6xOW`;t)B0v+HG}&n<!FmDWBYhW`11>A$X(@)8DOllbd&-
z2;LmS$lY<d({tPca_^#?%e#XR1TzPz$`?*&fM<69R=>pLkzp4WgtB`AK{xY{zUFS6
zP6AGOH;dW>ejUXFG>G&_gFa$<12AOB3)s;&m=h4zcrKH3N^l?gk-MkZ*<wYH%88<c
zJ5qeFmvSa=jwST5nQBewVciX6zFBrgb7ol7AElWE#`5=3iE8V~>A9GmN1?hH%gbxA
zEWBLw@qnUJf-`gq;Vg4l6u0aEXzSVA-;Z6<o2||5*xKHTt*!0Y-rlxy3k9hc@u|){
z-m&2NqmWF60zzrZl!}fi;T(4r<{XUjPoU6NS((?Ji>4ry^)ZZ^%YpCs(H01m*-5~h
zhO|+H0XGzu;0x`>@qg?xArA9F&o>xX^yHks2*34T-3bev##AnZ{3wl>$I<s_pA1L%
z<Xr|2e|8^o)06GYe2no6x|g_SMvDshrMa)>eJB%A^s`IWb6&Z#cmdkUI0tT}a*N|M
z!3*^}B>`-5zDxaiv&b~l08YeXTwqM39>Cs;2-%Hm3X3otI*-z(t9In5K5%Gpu^H=Y
zt3HkmdK!#-q7S&UQwrtQ-0Y0TLBSdtB|JVk>c&yGqjs9`SmnFDM4wMYJiEtd?w{a*
z75W-GIU$C5I|?>nj-r6MJi95u_vNJ}-&cjA2<6wY`ab1W@l~%&(@65ba4i5lJ$=&i
zXu%q#e2cpY_fX89T2682_Ij)?FDk!Lk6(=O%=uBoW_@P7*kf*%>k3A#NI-#r();<d
zXVKxl$9`8Zq+jRmFCU+3nk1OAI6$}$Y-w8*0kmD-_kL<x!aS5+)a<ZZSeW;5g^L`m
z=$+oPKH8B*C+>?=qRo>mn4K{1U`3<w1G&<{!G65j+*IG(i^IdN@c`omv{x7ACmLlz
z4Ll2dlwD=u4~io1R8zl1R+fq%>YQx`z2d>$JBc*zpdj~7CY8lFuDpPOcU~!+k9k#F
zq5y;E7!%GXo~jSRpC~+hkv?cJ)g#MGpyVx*-=ghv_ZRdnc|UNsD#b|?7=XuB$%WD!
z1sDn+@Q7nGQNA)xsi^7$*-t$0bD!nSu$p`aG#6!3uZNtj8$IqI6wT+TP0MM7CyZaf
zo%RIBsFP`ro}@vk1k0;1$RlY+_BG|NH+`C%;gfiPB6*<1;;2*VMcJT_^o1(dfZyU&
zf|Ijj(^tw1%zOfn>Q6ZMF#O1ev4|rxlV5o4oy{CI%Z3AX{?^@y@UvdnE!Z+v2;c&4
z;x>zP+(xt=Wnv727Ah~o(k?7MCok#>?asP#jfwh}9%hE#Q7(FnoU9|*?ux(OQ=dFN
zyVQ7;ay!PKJGa;3$3Oah{NOuJ;_-vqK7RH4AH63YcTW95cr|uy`5pAZyEzLq!NY9L
zVsPM*Yi(!DupRSN-vhhagXhNY%`~g|%Cvai5of5NHt!Deo93H_tLEE;|F&tr%aFqF
zBAwsHWd2`W1}nb)@Fze0IpGLTq>RU#hS%Hu?<OIv7M$pTp!q%r{}bj0IiNGxLDBMI
z_l&A=c1y$}6uwS_?+oPdOoz#d5b09lmya5U$7Sa2zZwHK|K9lVkq&++ijCZ0Sg}}W
zgFpY};UnGr>j%4;&*BEa?5A|9zyI{%df}8lSp3+J{CF4lw5P=>0_+k1cjt*ySssmj
zV9H0r35GwRXlK_O<Jjp&64gaNLy&X)O0<EEefYqySo~X@9lpeY+?(i$d3gEiMLhfL
zvv~RW=dt^IJGz?aoQo+!kk4IT>`+NUQ_^$h<>4i`GWeTVGGWnx2t$C93x=5Oy$pur
z&wXDAB}gn4UJOK|CfGp245rd$;3K><A*}of3(YIpnn@=-<#CbZ50lS8M;kK`nUQCJ
z)gbAWN2)X1h<Us+NQ~vTQG9(<;{jHG^bY2-S;D`7sg)hPWueaCoWJ!yF%mOk1eez|
zX%b`1B!kt1T<q@FHbVfD!9Wl5u0d7lFzEwtU>G=+5Lo4^8CKk;&`Lw<K<UM*KC;do
zQA&$=8jP8wV_v^z#TEm0PRh+7C^r;IM1*PL@^dRBRTlE4v9mv>JK^0ggd+^-Fv}=Y
zsVGb$jExyUm5F-6fKxBpo_LhW<yT$mB;Cw`;E{|Y%o+`n?Lm5GR_vk)iP8fGR@4Do
z{lPo{FW1Ysj+@F$zQiOlRdJf&9gJ7{fwRCFxKzOrWko4GJ2e|KVsiP$Q2wNCiT_yg
z%ht3FEP7EuefW0}zM2~w#z7q*-IFe6z8F$Th{(v<$E)51w>=$7K76ND4|taT{#KY8
z4;Nd>N*3Jhrgy1!x|0TY;4u18VbWgmtKn#8M;bEKU%w?cS<`m&<llfc>*pp!zS2M2
zF#lw;H=g-zyYaLdxtDThIQ|EK8S)3dW$eO^Ur~@bjQ?&-{Tg;pdx@J4gS7dO)o(YQ
z9z#l&`e$Fp^|bjnYs-@%GHw0v?{eB9>&!QHdpiKXeugn1UGYxL$nrUO`Vvo1;-G!(
z?{fPqztb;Gx|3Rk{Y}W5I_N1sZ=_*sPV%Z>mPt?k=Xc>z9IX6i+W7?yNViNQOHxDT
zUw%|uaVyS#+g{w9XW9oas!#c;_+nrGFfp~oZ)_Egd<}39cfLy>#|dATIY0cJ_12qw
zil5yL3V$SC5yqZ3;oKU88H!t<Uwr(xPw`bJ$kqPWY0O8JGSs$#{=johOwXh;;QVzw
zf3+9C{``d}z8ty6@k`(hy#-op{#11$CEhuPra;-_sNCom@NX78YwY7FX#^3Bo6rr)
zf$*B6=UD0OW4PLf77THO(Lv|PcXnMW_^J4`iOJa*8Jmh@@t!E(h`ZlHShgFj)<GP$
zJAQBPY{#pYFXF`)U&Q|Yo&_0f>>Mmu5$eNh=8;a6Z`9u$_r%e9<9et3kR2dbF+Igm
zGiR!~c$@PR^?{36(Bytv{1tZ?@NBFu#{=#`+E|Om+;r@0ZN{Je>3@iiKmKK$og69r
zeB4=I_pwm$McloS$8QUUx5NYALQpng{$zLWAYQz96<gf*A$V~50E#b^4BI=q(diz#
zEl@HbFy!tm9@h`ShblNRm8<i8tRuor@qQ==*!_=Cnq#)$-B9pxL{HwWk%Dx>!7C!D
z1V*facRKLx2w@f$yu#l{X3dLNQ5@nS*hcV%vJinAGznga#bkLjT^^4$IRT#BigDRd
zHUba22iOF^IeHC`1YWXZN%Ilak^3V!RDxwg9Lt1a>qPUA{@G=8k2PQEpT)Ug$<dIr
z^9|!Euu=ALcyt_FyZf=<=|)R<#gUsFrNuoleG*k#v;*>gR$QwSc&F0y7cb+pr_bZX
ztL@m{J&2Cd;dnwQ96~Bq&vQrC;^Kk@uJD0Zlv;1_p0?l^GPq@gQ{CQ4eE$4pJo{o(
zed@K~JMG0-2*A!xPYi2>tLz-+Xi&<8AQ~kb$~lzI2yF#a@rUeKgMS@YT9?7`q6o_8
z)JCk(X2%smdiZ;U#vS#Ip86MelyL+m$37m&jW_V%=tBgB><&ZVNO`HJD&=;d|I-gq
ze03H6pmP|TTf6acYbW-V&Y9W)IsmWEt}legD7>g2<zY7_W$y`w9OJpWw=dl09w~OJ
zG51hb8k6AJ*&Svjj{Mnq;iEX^PTw%Kf{&)LlqPt}aemYt<@j-5;q|UNLE(e=1_sap
z<{wo_f)Jeisz4Upr0+04;C(~4RWHSbhC$~LIHT-2IXhKpG)GXmtq>6I&?g!7Qg}<b
znTIgXad8Q)<QR*Y0W!Z;Kh@k9fiyzsN%8*dAe?V5m1j}7rM|Yv@p|)yQ}^&NS_k`{
zr*=gP`u!89M_IuI+8tO?FO(5;ibuah$jpP#oc2ZNO5NBMIwySQZaE(b&)iLQVRb+A
zB5>G3TEQv>&$MNJ10Ciud4fs`yMY^XH0ZvM5c50{fqfpKW;iHs(<$mMx&h2p0p(rb
zuXd{QALe3%If)5{9;IOgn8GlZA}!CUY=<@yPB>NkEc{hnpf$OW79|(L>g=0*6E5c{
zEViWm^k`mec|mE<oNID?pJ*<-hI|C&sTC!c@-zObAF)`{`IdI^BF#a$!86$nJ$BGd
z8-5iA@-?1uOdd)B6k6aj_)YpG!VbSSyRo;ny!#o&-Qv=c6*SOSk|!S)NI;92YBJWL
z_+mE(N4NHShf3<oJD8!@^ow)hISU02MI#uOX=`@EP&VkmxcCF$$T$Ga<EY()bBOjR
zoXUSh{pCpQ&dyFW?w@#^p&re~f|VX9sm4@J$_B1+Bp=FSjy>-7IQF#fQ&Lz!!EPlK
z1}p-hJk*Q%3U#F1D9Y(O6arl0zPrZUoEJ^(i5KYid&Z%L#x;~!zygI4VWA)F)a9N*
z6ojP7IK=qG0t6NgaDOJp@2;&b`u;=4J`{Z%XS}|;;v<OJVaHfjk5}cm+au~B7$5aI
z>XUu>t3AYD%*{^vzE2h@@XaC}?rvLJ5G}J3O}vzFlm`VBS5wfg%gP%i$TSKvE0!k(
zfBG)R(5Zim@9+Ww)rm2=Dw3d`;3RnJoky4FNl%5x^hJKELYBV3f(qJ_yEBp1W#3j=
zjZ5k~;4}SE@OGclgMvJjZK*t`jR}JcrKU|EVo{Tt!uh}}`i?A>nK6{RCQXOk0$_Q{
z_sHLTr0{`uq@P`~s7J7_3VarH9E-<5DGgoEu|jRBGD3?gt#DsdBlAs}IL6XDDp6*`
zn+g^@^mUF}243`cU<qv5oe6wM#|pC2=lM1-Dflo+0#E6~;JNuC+C;EH$&Gu~fibhv
z&b%XxJK`vdcXJh*LVw;^TlO)W;5xX=xTXfkUPk8wW^T_MkKoxFji&R3=8*^e3^+oI
z+-ZR_fL7U|$Ivk3p^O{3`)$%=d0TuFd^Wqxyi-s0F+EXT7d4*TSzq-|O!_75XPKJX
zms1MZMGq~p3{Uw~lkJNWZa41vePh1tyN=D(_l<Dm<?Q5EGWxP~-ZoD2OZ+hDcjNz6
zVKHCl|F-!y;lFL#?=qzDyGZA^F`55YmjT7s$bbFE{|KHb4dK<ti3>rYKU_Nvejpub
zciM>$Hs#ZX@S?r?#s#9Iw<0*Mt+J(436XCMZ;Ddm=3jAhgbZoCCk&o0Oopxn%aZCh
zlX=sWXilkGO1w0(D`KXnPdbJEs;`5mHS02L<cXWz9asNY9VtI<zGX=XwStwh%Ghqd
z{KgF83>)X2@djR4+?FZ>v|9E`pz(x&I?y13tR?LXQQNL&$Vavz1P~$*k*I?SCfmvu
zs*+7K`dX99JgQ=^a}b}td>XG_zKq^s*8wtV#f0xtnlNXqPP)>B2j<^NQhrMeGAzvG
zn20pzxkC_9l&~FnQOwMW7zb}<7X@YT$jLHH6)SPYlrsTS`V1T-2O~-$U?$UyC_KBx
zCOp70XjP()q9&CmDyqKmXHaBy0|U&Z^0P8W-yC(qQ8OIX!UV9_J&ISGuVQ;^Tj8kC
zw==W}hB6PE#1GntFIs>W%gI%mw_^+$cc$`GLppr(CT*u*%IiEvVtRQ#?tJimJbL^j
zmR6U|Xpk0+%e5Mkt>5I2pbEyC1ZfZ3#P9=iaEl`iV6I^_7-VxNRthY<GR|2Gf5ZT=
z%j#Inf*Dk$IW|I@py)~b^CVEFwxj#0@}d1;B<NY-EfXu!sx%Eh+6!Ep7|T^NfbNoV
zDqkix9(3i-H%y&ZUFn-am$CtIPv~$Lo{>9sK!^dJQikIbCfd0RkOBi%;J}JTc6-%H
z7O+F0!#9i!I7U9K_A#DmB0|9lo9D)><Z2)8M`LFkFlSXY42M_0T_NbWj```iSYKR^
zmF9{SiWGu+X7ePY1JjQi<ugj*r=~Y7vEw1T;}+Yn1vt1r`&*`_GxXt5PFOa&QVE+b
zLt$m3EDl=DO#mCPvi4fcz*`Yr4}A<rc`Le&-5g$zG6OF)53+~hK^f&&>A39`I3-Tj
zD$~xe_IjsrRQqLstUqb+k#=yA{Aj}x%<sVwyZ)8~dDDOnQ_GTmu8*u0)~*ym^Oa%n
z*B`c;uHE#5Tl%L-dxoqA8@LBNvK}&&k#DSeG8JccW0y~c$aHB#+On3xA)QVVDC#nR
zcZpleT}}NqmBHDso3i9r#&ZCd*&~eI)LWcSf-_U&-^OXw`1%FT_QM5J#%;cEaM~_{
z-_;o_#^Gd)TfF7LP7S;1aaDPzlk-c99Ar-%Bb7{mU<s+rBTH2wQr$AwK}n%QxVXa9
zNA9wcr^_sVLX^mUVQPxxjV;?ww8&-4v@tbQ`7hsk>&JQM$$D0w@{3>o*nk4ugKPyc
z|1M`~A2*&;jsNyf{&^<yS7831|IPoql<u$3xKI2Z8MlibjYuAwyh|BHgeJ7Fug9X(
zJB^o{`|<3>X0+QqFBlnWb)jLwPk1c^3Lf7yzUw>FPN6hoA-HcLW8u#bdNBTj%88GJ
zL7=9$9HFD^lDH+_goQAye1<PvTIOz}*_hVY4PSu*f|alcl`qc`OkBjwl*VAC(?Ln3
z3A&GQQCSecAOM(}npLLCweXP$BRH<gW4GMR55jLCsMh#R9S|Us|APk)R95a0A%Dtv
z78@Mx<gnraR3C&?9HV5)ru4Xv?8WAm`@q`bf)!)$J$x8T;(Jk^uqwXQ>c#;=ff1Eq
z3<2}G@83b_4{re<V?h*ruF66G5h%qs3nq(;R%WqNo_kR9=pF6=x>Vh!#A7#Oes;!6
zFWQII?g%x3-JxJ~dd$5G>{39vB|cqmr>|cNO2wa$zlBSxCp?*g+aQcW0OOrWf~AGP
zEE+pGS02DIc{S=oI};IQ+U?tGDmV9Mj9X}e@@7hW7dv|p_OQYj>|*y6b!Jxwca)*{
zJM5i!XB=&X;Bj(h#sbz}|IE9F;PVkqp-ABVF8KFTmC1LhsomhMP*i|F?9}0oT3}+O
zjM@~Q4@?KI;l&WF9Uu2p2aX04k0|_WsvT*+)g`?b7EIGl&$*Y1c2Yl3*{3yknV*{!
zk1YO@d5H2u5Xa8MUw!gfeE$4}mAK93Vyv&O%l|44+b!{hXK`nJDSrHeCwecd&2IT9
z#J$7g*xqehc!Yw3@-$|~;=Au}#0T%+i^jsd=WN`|1rD<#5n-h7&(Yk?yWQl+T{$QS
zg#zNO^Q4R1iIsf5`>e`}vVy*bkdBUupXr71kfX!K+z;8o!K!<9pTMin&rZhewH5D(
zV9rVZV~2*1Srk0c7qSwcg@xc$?$#2YEk6q;)vgH6EF4vt*zJJ8lp`q--f^tl)XbE3
zCtBbrKL#mwkU?8mC4YE$C|o=?j#fnjWltVEtbjq72dzTzi@OD)3dima1b@t97`4DN
z=n`dtZo$*nzQKGIVLEdHaKooGh{hlQW;|wYfZ+S$96{ol(iA;ouB1l@=5UlN1*SQF
zfX>X$&3Jyo@t`PADKm5woF)wQ>>jqQ^g_W0V?ujyPYlXl+MT%s^c5le3_E?)mhf6A
zX~08{$pSZW=S$A7P(IP$pyxU7%6YQe@r=9t@(4sPN>tr)@xW-zPKzch4`?ZE2#i>~
zz%DxIJaazsN2x}5U`d@l=T><*<_;x1<)`mLe>iF$0V_K(Jf0|D%H*9ficg<Qo(ggf
ztz~Y>7=srrR7WaOyD|8I+q^CK6~m%xn68_qL-SQ0%FSHSb7T2Yh8iC`Z|HlSi`9Aa
zAUDdSHf4t^ZAy63t$4_jX`u`-W2Et^A^O-*9hgHZB_)YJxWz(=)s<y`Z*6VH-rm0B
zTM?zOzzZ0$1L^Vm@5OgNd_Oi;n!<(LCB<%F;O1j{g%i+h(_ExISEteKw5>Q<TU+z4
z<6VvUy9Y<29h`h~?6RQ9K%vTx<C9Y$rg1~*Fa|(hCs-(=2^q=~6x{F*&>N2UToirE
zUA@9%;pg0(=!539eT`Kdy^A9Lm?KpM3s^_`RPLw|O_`gE_aEK&PFVWBfTuRRj;82q
zL-dAxQHCIAvMdUCjxg^RKCqBiY3%LoSV8PXBBFh5l=QN2=OH_P=Ne0HCm)@sHrqcq
z@b1XIcfFmMPO#&OqxygYeIk#K1(#n7rW39sbPrm(zrP=6C%~Me@!D2?p^Slcx-SZz
z_wU||m6a7QBw_6FSg*J!2a$n6<54#39kfNWPla!}_=NC`<&1YMG@$)0W6@X&&e3n_
z+u$#TJ3~)d*dbFF8H4dGG@vhXy3@7r$UGQ`Cj%W|7x;ww$|y1mrA42nzgm_i{HFi1
z%bx!GTKxcAga%Hje9$@LyvhRIU`G~sNFPlalX4O(M?~wO9qe}ViBD<=j-BMRCGY~}
zZW{Wz;t5t>%;LVvlXYX<1Mi?q$W=^hg%{*eP3^}R`isYWl@Yj;HgZQ61uSY@M;QuF
z!p>pZjvfx3g^sZ^jZVQ>v%R|;92<+Q(nt7;@ABPl!V|_z(xnWHImo9N|IXAVC&Dch
zyU4ECb%zWy<w<?zn>;VPD@r-;?pziB`}pDQxOaP1^S7(m-aCk`y^in8?H-+queo&E
zrd?cBqBN)jyb81><(#KQ(V62Vye(s~@j}Kl6dI;{;U8BumM=EMTd3_?q7g@;BjDQl
z%2GVMe<$9%w;JmUGwN%!zu*oXB`u}H_)NV<N5<k;pFWK*UTnp=@@AJaCh1?&QVett
zI#|Ck*iC1L%-jCeIBc<F(k~xiQB5if)71FG<*4SHrJd&OxNn=`Jio2^y0pK}d{umW
ze%thaSLSV)zRUcRfBeTIBmeb3{6{%M0Z;Y5DSx%dTORn)>^A}+M{#(6LJWxFL~df+
z$j_M!+=qUN?1W2H-qlY(xhsI!D#MMBe@%c&Kn4+FW@U&%SS<Q3(d<&9kZiSMoHljy
z-;G-WQhth%RZDx8>82yp^e9FOY4STwmfdbP`Q@bZGrwznc#;<%iX*!skYo*4Q}l0o
z_(jA)+QpJ#?CUTN&IdpGsI@LR*bY;YZJp^E7U^Yo%IDAy;~%*ZOH7Ai_cdt0+}?^`
zefCLgZEu>WvmKtcg4I(`R*_)3a!^F^#fm%_GbT<DH|h#Y$5D8!JY~>l&}4NEE0{R0
zfdK->DE&16ir6z?Q6?rp48#z`JabvZS_7TSpmdm|apWE=$nZ~wR#EfEW#LVH>U$}I
z3^T#xhDi*A5llKq2l(Prj^tocpn;&(YRAi$o3XpIqk$yh^KA?zcvmbFaS?xmQ00Wr
z$iAtc*R&vIE(SNhWyt{VP85STPxa@m?@<j%3wPJ!(RV(I2an&=1ZBoK02?z0OblhG
z9a*Wu<Q)bAK@~?guu==yARr?S@L>gt1zOGr!BjFCm2d%Hws@obp`B3Nco&q?<^C`j
z7Z^A1u2P<?Bs@_*J{w+nkRB5*-~`jj<QPNy!l;06RL{yewJq72AySz<i7bBJdannD
zkU^N;UwE=w1!l|rT@K&@Tf67QsDKZY9p(X;APC80P?!WVxWkm^afHN&xja)DVIt^<
ztX5M*<65>e%zmym%Jx>tDW_V2RW+nT{uU$(mN0Gz0cIv<<Mz^etS_&{q~Hgh02?Y`
zx{8fBc=(eaYLkAZue=qe(j51DyQRIlrH+MoC4j6Yihl9~XFLn_!y^M>>KLGx;|!S$
zU1O*Jo8gE^h=TVnTuHkYGJ_1o!KpZwulhTvqhOWM!H?>nX}LZnZKoh>hL9N~IdHKV
z!0at!Czlnwa-c4xp*E~#CJb*^tGHlKSAEfj#kgK_PY~Qdum0L#eGBHi^CKN=I+zkG
z!^n_U+N*#0CSHxE_v;F`0W4}xZ>F1W{Ne)Kv$W|VyUWB|acWW+eK3|dMS=`6FE59^
zi9jhyF@Mu1TQJK{8Xosbm})2wCgTsAOo3SWe#<y;N!$KuPjD~}+G6`4<=S7W*C5Q0
z$spA#!$(P5F1TfG#b7CbGYP;rqCN682*dZFG=@3lB){K;Oh&1Fp(NtBj~B^*I}@wM
z%Q(fZ$L?g12rmB&JMb%S#+iilNc^{d^3O^NfBEM3LGgvT=|_KchCGM!Wz6unQQW39
z1~b7PL)f5kid|mYJMDP(#g^~$Vn+;PQu64em+eYZW3%GRV;Dk5(q&Og3Jz=IlOGd)
zJQ1KX{yTrY&3h0?dFObY!WZHr*g!ZzJOm~15MD8^aT)<3E3+F-F65pTj-$>+042hx
z{M!io5%!6nVbaO*JC_=hQJ`>ts{*O?@FAoNzrkww6h<SwK%kGHn!-^I%D~Ed((?|A
zIq^{{;|#pT`H6U11R~taG2(ptr{}I$yK~4;;1$9v3yoM_m>0i0WyMhI@HjSiTd}p%
z^6`=EIHK+-T8_QLjn%^J1kqE3_3+?6?r54F6w3SJ%zPI78N38#M#zS;17!??8dit<
z-Y@ap7L@WREQ$#DqNI-q*2QPfz(*tOPN6ONT6hDTG(UiT&^8E4%-2t#2w_1R3P-_8
z`6FmM;;yoG*E`h^y3*F{qT|@ZwY6oJ36z*seb{N{-Cx3!bG6-(_$lfEAIR~42>6bb
z9zxP<1kEF?kVl}ON=)A;#!e&E%jK8b@$p%-IG#rjf+X6W`2`oq^M)^AC&2O1kp-{Z
zV}%k3L95|CHEn(qJWnAtg5TL#SX@ZKo*bqG{~4vF?38{Ul_VBX@j~;MapNV*1a>HJ
zWM%uX7q4Dz>Df^okIf50XSwT*um~1U`~7Ih{`k?Ic>mGu;23z?iXCZN>X)xJcjBOR
zs5#21j}dyf(TFGS-HrSA?^s|*3n3KrP9JbaZ3rycJ#}<)9-VIA3a~TvbB-K@m!-XO
zWjV(mUZ`#9pRC%R5RB)9L-3!%6?kIRk~s8Vb{nwE%6KR|f=}mE2ks1_f3T|rJOX&c
z2j8G2dCyo`Y<ScczH(;@3b0({2sUyWg>g`Py61&69NjhR{=l3BAs#fx$H1wbXzxSr
zsX922y<g@CDs7=#7-$QN{-AdVmOW>HzQ|YzczkSOJAIx0fFE-q>g`AKTMYPxV);V&
z<8r9cxsP{wdBsP`Q{PLqHwxe5BlSbgrLiLf1`a+x(86Ahwi|ca%w?fdR?shStQ>dK
zO&Q(@9hsMnjZM1Wu@knhblI7Q5{mf^^El?m2>zLa_;@quy66~6MDH9?JF<Jh%5k+d
zyYXDrVmR8*a}?Fv!eFIs#TRZCFrQlx3`Bioz0>8G!Ws38#aLQi5^bAS-1FGq+qY2o
z=&0xXSe&{l*g!i_K#(>H25<)4Bwu!m(AFph*v-H$z@GX9ZOJ^6auJ@s4F1(|#%-uV
z-L$mYTIprkg@0-<j~}8BIcH>^#kc}`;Fdd0v)*~PA@w)DtB&x8!@^f9MfFWOsrVvq
z`Z9_@3c`4kb68{>j2q;QARky@_$G~{wR!(DwA${9Pk9*`sVDVk7wFo`Qrx+7+wnIy
zx4dwH_NMQ-yy}<Cak<m-y$5&VdmlZH`}giRt?lhC(ZhW&fZ%wtjrBE^O>|o{M5Ncd
zz9>He^39zC3)N#(_%|}C@l11Y=Bny@YKS;MDYnlYk!M~^Kzq&V%>oPZX16J0Rjbp9
zm#;QW|9$V0Xg|CmyOLF?m{xyAv5VX#cj}y{P+kx1O?iM9_=3W0QE*&hx0836!3Ruh
zJe~7#xfuEbFvtc$!8xL`Oh!}X;@&K1Hsk)@?r!|z<6kI0^@Vfs3WDSM#{F1S|7CXy
zWoORM4s!0l+uPfVp2|Sk;CC!P5N({BZIY2`F1vpk8sC?emQCNFL$5_6ro@kojlxHu
z3_dYEv(iLu%lO6)uKN$}yPuzH3<IA9nm7|J`r`Ss_~f(Cy~~(tF(y%WaNyp(d)~$Q
z`by)t=n`=-NBtu!MR!%^x(Jw~98pY7sNCQr^I>+Lb-3qG{Szhp%p7-H&8t}8FZZ#t
zJBg!a^{+MqhkSIP%2gM{fX~3@RJaFiLNUmWJ=%^up^b{{F%h{NZAtx*12Q&wSE~9m
ziZk+n7X&8k9*2(5|9m8``y&ckXe7Jiv+si^EZojMn*H}ow1KzN0|yvClozxQIg8!&
z=E0cmt6YRfc?xgEu2>fJcz2g@6g*@>546E!obr~VVM6&hZWP?Zo%(?hDDt_>%)7gI
zln!{sm;{c(&!G6?x(#MUjC$1T7$uw9kbc4rIPM)Y52kX@Pfz*?!FwAk!fExdcFQ~6
zxhwae(|14RXk+k{a)49tZPbN2(JtIerXkJr*m$P#E*BWEn4H~e>L;RM9N`ErG%1?2
z)Z|{cW{fBu^Gbab!D>g*=!bV#;|CwT7q{1%>R+PMug}bz!BgxVv{YC4>`~K<{r$H2
zIcNxmwy5odF|7vqim|(%>R;cOVLOIzKjr(z;9h8TF>flNzB<in-j#OxzwMgl?X=!D
z-$oJkb!mN_`KtK%{I==;uFO}#vl<j%w|@51pZpwT^>$x5rqYIQf+R;nVuR#lH+>v9
zD7GgO*u@y|%8xX~>H_KCXj%OEg+^Apt@?3$>fr^$#G5i=To#p;aypD2h=&<eC6)q8
z;-o*!H(}H_Wb{+wIV?{2*@74K$S??U96|+l-u@>KIS}68>0ZlLKK0<J61gIo4$0Q2
z7>6ll;3f+X+2EK(#le<h!&2N_RP3_a**Oh5gmVT~B|V#J_q){5PT6*b0T$BcPu^8v
z;kLjJce&l<XJ0)cND7KO<$)j{^*QSANQ0ava+=h$<($EQ?dA-^5OOBIFmD)+<zNsZ
zJ|%*vIlr<Z$4AOA!PR?eB9BCYsKdB3IKlKYv2pnvGt19_1;dKqpA?z!7<SY_4+N7k
zFrh$^;`$4(YPm=VFi@WQRDnkxT_A$3Z!?gZz%no(Okh%mQY?>x`1XeKzio_%ltn+r
zD?9pEro4&hhxU=hjZ^|xzhP6qtN)N0of?ak+Z(Zd?`||SIRQSTi897?LxYfPgLY#A
zhH{a#u(SHn7iY>pTZPG0!h<%(4Of^AltU<jm=vKbq;Kbz@3PuWX;72I$xMrhZZa=0
z@Trs}tQuQ`D`~QdMT4^JG&zH}Cn2<>-1r7RFqF$USK#J4^6;m)ia|kLXHSa6@KHYE
z!x(`NE+_cmGJ*%W<IHYiq*DMSCQ@*;KuBR=cqk`zz>j>%FokyDUY5rzBxSE@`;2n}
z(Eb!qCh3vl%4tS@q^Svm?>)m!UIe4R^Nn#d|71+(o0j@tlP>=CZ66vjvwzFae-2_7
zz0z~wZa03?j>#bDs3iTHsI6McGZk=@y|hIMP02GD3D}lCvU>~|ge89Y!kr)_Kv<<+
z;|#v-uka2;_$&{3rwczNZw%q_8#vsU3~Q^+z&fyG+YgM>DppytO|VlAauFO150^Rv
z@n8Jy#2;3>(@=KKv~Y9D)6wO&)gHxQsbOlJ9kTqT&HlB3GPc=Cz9dR9?PGK)DUoib
zUH#HIzYJsImoPOJE_Rb&LMvTr7+%=$AB2~kWJw|MCd=(`{*{{oq^<nQiEfr5GrufP
z8r+PD{?ET^@XOt6PIlm9yVH;bZ*W;{SmEUR$ucCKH*fu?EqyW7cH@#JzxmCYV3f2C
zY51-0ilenmH6{ia)q2}YVfo5nNj)(3QbGj{i<<-K8~9-K>wvs6phLI->5+Xg6X2D7
zUGz<MJ3B4lgB3@PI{f9|>N~#vI*f5wPmasuUeU=e=TVJ)OrZDL{doF(J3f8(Dz>)|
zEL?Y!gVz9mK!Cq$Gml*(@Ln1_89z1lk``PU<L0HtKqk-#d8_cD=9@x_{00H+rg5Km
z2!r7$SRCRNryiSyG0Jl=o;y78aaZ{@{<8WT!5d>ZaoDv6hjBeBK67R^<{As;c@~#f
zy%S<mZqsvIAifw2OUrszVul?O8fRatP8z-ySK~LroAId_mHU;DmYrGbDB5dj^3Lw8
zj`(7IA92?aiZqptqXQBAz|W3M&cthuEF7B*E&%6pf0O`?*%>Rn*um54oyL~nz0*F5
z{wYFy_z#5O!0*h*Kfw=k8Uk%(LH6{d1@{~ygpgD~4UWZO*Bh{zP&=}Vik(vMzVJ`*
zuk4WW<?2VL;d`16dvebs-{3`nGs6pl7x-R;JP6m>Vb$*)#mR|yGPR$A#ijCMmkDKT
z_*kVG3qbO|sZr%U5etpkXv}a_GlC;_a;PoXoz_H9IUAEI`t;bS@2BEerl~0uewuTr
zJSQk^L^v*0A9gTtBp$~*QU>N1rX{qI>IYv)-6xediX)DPLI6uco~sV<ugp1^v#_fP
z0UC>qz)c>~Kmdj?Esu4Za65u4EQCVAg_{+!C=)EanzEp_(@CK|ik?Z;Wp3hD%ubBN
z`1N^=zdl#~y*Lx>;qf>YaQEOaHg{T9Jh4N7`H+<!2nWR%vtxppFv=A8NZOD2m1jHx
zJaaGjT~^rlP>`uTVXZlei^XW2u3&VEu<MqdRB8f0>J6{j5d0Ao(>_hL8F$hwEzHI8
z;=GlyD~pXhs&RhKLb}Aq<deB42pEFf2qrlNB87Y?u@UqsSNed;0`CiK*-e3RmoAF}
zk)0-$*S7M3u56sjF{-u&b>;{q$^f1t+-)@3={e{8N&EQpSh3g{k-Ln@qs~1577Ac?
zF4A`J_3SX=A?@7x$+3UxQ;G{N!QcDnFX0nqqaPvc#JwuFnEzntFZfv*E<-vTb%`Q+
zd1Wb9*Vo;)oamrJ3WhB3%X{<?s<LZ?vbY@0`B-0H^Wx82<IvJu7Lrn5{15=MJBJ+?
z;5U7ZV^_T!R^?`qD1vZ~cLRsLLrLW%e~zf4{wP3!8N4a*W$q450lp|T5kiB9Ubv_<
z!8K^Xg!&O>VT_njA4XW;n45Q58!8vDrxC#g6rauJY^<%U_y{=f0#$r=rT}AbnbS3v
zMF-ivl=G=8(=c!X+U~Z8maFV}Tq;UCgxt`B>@Su#=)rD0Qk^Mt9WN+6?;JaLs7$(f
z9w<9>vMP!Y0wdt1Esf9Qg}@kuA~m(O7eugd6B?n1@rO6?W?n|Qowwr@eoH^(&&~)G
zc<c~bTv#+7kT?Cmj%%*B@PT#!W|RTJH;YSAz%ix*E9S+Ahi%Wt+2IS$Q3m)t7A2Ez
za#^FMb1BRg&cd6?4Y<?SQc){B6aE2v)0@llU<Vj?_n?U4c;R1v`cyo?juq!9w%Cng
zSXe%j3nzTEsveKosw;}?oYR8$qcOu`X7xj01D^~HL-B)Rs4F_iQL8=zR_*I?$Kw<W
zYZ-46kHH}n%^Y>Q-)i~z-@QF{L$`e&A^ZkB3GxOMYPeDUPNyG-8aIy5Q3lfH?6~Fb
zE_QPx@6mXtw%*y<iZ8x+7SEo37O!4@5$$#>_I7rnqxbI4X7oBO(UWV#r=kAG0uqiR
zoE1FQSC>pTSC*ii4Y%W!@&WdYr7VQmSX+rl5AMaCjdk@IXbx?+ESR!0j5;O`u>j-v
zgq=Pp#3GJ+9hc|RUw;~(=;4lij!k}{vhHlZQk>CP6>r0Wk?X5-k3lF0w-o0WfBwI#
z?Ouw$3dd1!wf5uuw66*aw*5}*ZNG>Y&pwIIKK-*e*m)VZu2Eu)_}FF?b(g0{anx;j
zM<Dm7!RtfAS**f>1~0zRAkM;cAMJU5D*7-LYpN?dx@k|#M1&jk2Ndv4(Oh=Jv3LPG
z=pCR6OP>c{=o8Rh<bb||N^MX7!&IKoaus1Qauo0mnooPO>kGG3_AsA;n@svkfAtuq
z^q_|*0x1uQ%2B}#8k2l*DyL9bK|@jM9d<fO^U4Y+j=<#TNEUi{L5uKupj=nEpi9(;
z`ZI3PSEy6U6zRVg#tZJcLg5N7LYF~A`U?FRnlK`~@ffO-L(f@Y0bdIYP!O^^iv<DD
zFO-Sbs{69WZIsF^n4oU(iYVKu7i|F@0k`trTxiOL`xJZ`3v1|?z#Li)o#)t0WVBXv
zt6Y?ozCJP{K0xVv(Usop#6MG+*u^z1z5-rivB^EU@VTNFf-k%ibclOCske`wG?7eG
zX>A%PR5qJwH>S|Z!S_4Nu%A;SyuYx5Gnh;{KZ|`3?aj1ndc)~re=}4yHLZ82@nz=g
z{D;lA3HvVdRq^rpZPWi<nJmXwmnDxn{D=P~Jq56r;2?-!*&fAY<7cQ0tuG5Jik(2k
zV9(I`RYVpJH%MQebWDfSx01>5a;U4Ti9pPXxM>Jg+|=QrPk2?p^XqhKILfE=$Sb=9
zzk1v2?a=w>!r(?Qf)s-vacis_AZP51kl_XfXT8{AQS)`^5+&2L_2!Lx`EZiep$5{A
zOOgKqB$fn37GV%jChZiJ_97jUaukA;3Nk)afl`e+uoAG8ALiHfE}kZ|SoPyo6MDba
zV0I}+i^;`)M@+{4t|l?bgs~HbLZ%=&g%nnfoQN3rj^p_7IFAp(+llc)NP%JtMp#U=
z9vB2x04+2ZJOFaE1A`$K7sH_45;MT+Pz=9jEGZ*p<{JeB4s!M?SCxkqZLF>&FRvPO
zJ@Z_5VlluGpE@ugGmtV+!u+D>fnj5H4@Vrb!jvt&Tbo<)^5rY9Z2F5Bz(9F1*eUaC
zszL+W$5>@gy9@`$=$pD{`v{o$*Qmp$hUX2;;m>)?VNwIp;~#xD?mu}y8X6$S#K6H|
zAjD#|Aa)W&!Nc462xnL&0t3Mj7pw-=C@pU>kQ#{1&`>tgkTcAIcVg)ST4Q|F9|}sM
zc~T0)Ee|F;`1yi#xv7F`!&Hu)iGe!Spiccf;dA>TOw&Y3ZEEI44-7-DCSvy%1MHY4
z`7pGUJtxr>XS_2@--O8tIhXaqxM^}GW`z}8+!4k_)!-;C0@DpM1x)IMn8^++9)TJ5
zX)y{+D4000ecUT)RDX820Xvu$R#yT~Pt+6#@t!AS&_0zFhKNeI`BYa66~!n(Z(fUm
zUTv<#!`t^`W^4*z3;fxSfnU923;6qowjdd&_7;NJN&adyr>!^bXE!m+1a}$IseBWF
z3~pX&WV2fd$iJ+7)jaG&xtvJ}R}41OQ~cxw&SNb-;VnbLSyqxu7sXX1fw9mR#Uwrb
zx$dPC5e6H$m}QkI;WA83`({jZMG8FZpBDe7zaIPRE4}oR#X<Fx6|uFJV_^6MCFNOz
zk}n=v?h=TyVs{$ljg>m<FXN`57~}LZa)}~)eh!&TXeha2YK*MEe}Nk>{b3orOWo3x
zGGhw-Yw8YP47NctmB+p%2vtq)B8}skb|+1;3xz2L|C_J~&6qF@=c->9UhmQ$>}HtO
zGS{$acgd-9apgT+Cj1BZ@rJoct12-4!cA`;3_R9w&IY4C(xu>l0uQI3QA+}YkeIA>
zO}x4bWUgW@>!=(UgMs3p?eO(0%RETFSnB6+UVK&GL48Ud<mZZ%5LE$K!xJv!WPZ*Y
zQ!@Df{Imb=Cab?>^VhlK%l+24@!J>SEf9XSG0-!Ao+`vfxx&g#lviK~LZ()$6TAC|
z(eC!G;Nkc^3(^?8Nk(HSb?~AS$4i47V<q7jTj2q`D2Dvx=Qxb-8vDHHgLL%A&IEW1
z1bPS=YP}GySU_FJcqZw3s>$@&-JtPz2El9!3rQ88korM8yub>fn|Om1t|~pnCm;F>
z0t*aptnjc17TH}6pNk`7Htrk+(`g;%0w|5ylz|Do6<Ok2S#iF&FdM7O;$^*(8|8#}
z2Jw?VRtn*w_y<->v$F~QrGIiBt?sd((~A^zP<Kj*B4JEp4Bzx3FQ%c+O2>j=rR$Z_
z#~fLPawm_BoKoHhGM2^fvBLdQ?)&@j?=ABu@M20ldaPWA?@&Dum?K!EoEDrHI8a{r
zF85~DnI4Rg3jy6o*|9(0n2m>z9>m%ztK?ZF%uX2bZDUGTb)?)V(9X^;y))sk+fn`;
z?bnoF3gO{BSyj#n5C{s-F2rwwceD+BDT+~jgEQ=Cp`OMum5Fp&Wlb1%Vxg>}EfKKv
zkB+0;>Dq3bQ{7ov&Prv<MSq|U2y2-Wp=9GYa|_S~cZ9#hO97ka1}X#Jz)3~1@Cn{^
za(pya7H4B=e#S@qjHrGHq2WyrT5a#JnweWr{*&%oU6qUD7ZGeT_Yla9l-zH!y0m1R
zV&1{Za^^*At4nIrDGOUUQjt~EUu^FBZZLKgqR?7doDc4JF^{ZzFt3^yjKO_ig&Ts}
z#m0<zRu=ZXW`f{8naWa>T;K#sUheqf{xgo(>S#U$Z;voC1%@niRG%_mpGrEnvG`oz
z!|p4TR=^Pc7v&sqf@d}z66{8m7jTw?K&kOt>ExUO;ibapn>@isgoofKb3<sAyp0<p
znlCdyNWp6!fqANOQzq~p*d~opT53bnB*g?KC=F0h(~qr8=p6bUYw+DPOCebWU-^F1
zDPv&LA0C9dlm#K<+WM;L5N(Pwf?Z$KyC?kfk?86aC_lhm((9|;C>Ld+FD$RDm^QUL
zEzdnsh>-?I)<L^CCXrnR2yMB8iQOH2-zUXx2zEL`w>TCKq2iV4#w5F2^oF*RC&~+G
zPwv3eTt{gEYv9arXI05!=z0#4G#k3ZJ#*e&3GGt5A#jIYwN&1xpM5I$wXJaG7`lfK
zAIAKAvj{qKM;V~^{cS2M!qHCaAlmJHwUhdf@oHMMlDo3bTz(WRD9&huyjLsxG;}?M
z;V5fTp-Vp_j>jq^2iTFu{7Ctuz3A^e_EQ@YFWp(Ns2-uT$8I6$RWkAipQ$JF#%%K<
z%tyE`+>&F`ZTTZR0ngd>Ntxi%6u|r!b6(ncb!}aA4m{~co1-{IyTDWMisQf#I-^7d
ze-94!-Nv*L=>vOYCfvt`|MK!uJbLuL#<&^prb0-M>>`Ef%wG{wqsW*S4dG6prusiZ
z`x%u%?W^*6Y*&5GPWz_m?3#iWk54l8j>Oj1Ui|vkpIi0;Ugw=n3&Q!iIMWz(baoZ3
zc2{(r<NWB4D7XYi!DM`VM173Asv4$k_>-T|S!rKI@90eQMg2}?w^Bs#VQhnLb4>7=
zV8|R9;W-Li#_8Ts&tn-!$}!ibsiBc~Z)=WxZ`}fSdMW*w`yd%ltT@rXiWHNy731nC
z?am1{uhkaa`1IGmia-1FKQjzj5VO3r;`U$+p#O-(sDGm@Z^Zrk54_U~#R9wr%1USy
z^iCiVtkgHv7r6%s-i7gaV`J640kj2p(J$!Z(4|u!pO=dUyn|G6xbu^pYHjtC-Q6AU
z&fO8L*g4)7pT=FKRO`n-`eFRhAO6IOmUE2@9<Kyrj-zAj$=%6I?l*_sp3<h?*NVR@
z{{t`DXdUdD7W~0aekwfK5Uka<XP5Ex^Uveiv!}7Lx*882K2m=W9S|N;ZpKS+99Vnt
zjSTda^pUN2Ox9?~Jzv6p(<k~T@FFh?4^0wso6a!5Ll(pXj8K*`zIeQlU-Bd@q)s|c
zzc>A3hc+}?^bs0unF`%fZ~p-#dZ6<4F|_JK-uX4bJ#xU9-kH158QUlyyPZ+K9`?Fk
zTymlQgo1x#V?%vNup|TJ1wHqXj7krD>#`6<?aFb<^g9MR6oMY>#Zv+sa0L3r5xrh~
zuCbQ-F}l#tK~(C^4lB|CM)c`prGxL7X)z}>h|{<{i%FHAqm~yn{&UwV_2bCS+|7-$
z9Jro%Y-95R3Swv&H0MHm7BE;+eB#n)p{blOpkgQ-JNr&lc4!ofCK#{F!Vks<m4U?o
z@C{aqj=qkK<@xyjll$@b{<;;Y(8JuF%aOKcmJfga^m)A6*>hj99Ly=^T)|_k++2Vm
zQ|A#j!*-crJBIJ-M!af%BOC;wnDnd9n`SuPZ=|8`VV5lXjd?r#+a~ezo0+dm>+8%{
z#mDEjP5*agvK(JsmOSe4zd-R-gJ$RP2OSWn^hl?)$_|oMRxaZVzV(S96rUWr^@p46
z7z*ZC`r(bASqWJNVe!C8VN<}F;fUaL^y47>*dfE(zxcs;n8;wazpeJixWpO70S0#c
zmJ<4@wZsh4A$+y#vBHS3dS@||IxxVUuoiTtyCYCz`Q_^6;zW?Q%WT_?aTy7xKc`KA
z5_b@|5Ws5cm*G=B{K8R|noK1jHYMCmTj<AtEp0>xElv(PD6{H?yPnhQ3vZb}IcIgj
zp(b!gVn|r2aBQJDJQMIjpm24<L{p4cyBEh|+%Losk%q{bK&3oYSO`80Eldo|72B9O
zvwCS^(M$wK*3?0iu&hX9H5&vN=A6P>a00Ou!C<g}fd>%!72!s)0aK1*WXwzyg9!>6
zm;nei69EQn1Q;;kKE}&oR97+m93jEP5`R_~u_~718g@A6nXS-9^fwK8C-lDFWE;}@
zeCT;cL`Y}q#Pq|5+eUR%U1?*EyLh7<gvGq85cp}3T3la_@BHKsV)fqL7!y-LJAk_+
z#Hx=eF#-&HSFgof2nG}rSfaea`0`s#>VON3nVEhg0DQBU5)JrgV)!*In}LA2q1Z6E
zOd4o2YR!OaW=8Ruq_e#|g%9{MDLju$4Zau@QZOc$Di8OBX|`d;i<PIID9fF=UPY?%
zgTqW@fU(<MelXgeq$8{#1No6ZxHCJ=>QUjBU_d;=u&NqZczZp^JBr~T9$+Ld26EgH
z#F+`9+;&GW!zIsxA9;?k(p&I^5d?1pYb(Qq*OU)9AiOfO2Cz%|rVD~%U43~i=45GZ
zG48G1iCIl>gyB>|p%m{L0UvhKp>k<bsiN=vPd8<Oy*N7seyai1@Us(do`QR(7*0Tr
z{5xQNI=K7>^n-LV&>(qyoUeY8&Q=Fo%HvwN9+?+@#L1wx>%nW_=b!u-q)k*g5zKb*
zlyvn$T*Irxxk)E2Nku3_zWoghFrd2OQPM79F`1l8ub15U$7)wPYNniFhWu-p@!>_?
z{guBMr>Rf+yMyRw7#{p)dl&y<ctAw-?=)m;DKJ@9xdD%Ku$$9J7rT)}`6_kNkK(XK
z!+FR({fFzAZZ~5Jes-~MEk(LyJP)r0zBLxUH-^5Pxq)vr9j8q`#SIwNm^Gb2NW-ao
z8Q_lQ_H%5#J*GHxO#@qXsivf--<#IO-=%rOjrPd2YKoXrW|`{BCn5bU2g5pl2v!_y
z$pDN2>ET}|fWeP+_$*EXQ<ExfN7{a8uIcSKlq1tgBYW|5IrZZ*>aUbI{irzKE}#D0
zs5O+{?u$5;Z`uE*DZWgbhRmS9RrtSUzAXHg88^q57+{v@a1$s$lStLEq~=lkHuF2L
zh}VNc386E>ZFU%O7Z6AMuyXpS%h;|lm3)NrjB8$*NqdX;r~*I?EtrcoZpL%Qdd4uu
zUfgQkd6PGVaXRA52oK+oM@^~M5ncf(Cqh1i6bJ^GaOb7v@C2kIydhmZEM}?YO8$Y8
zBS6eu9p0VLKu|ZMwAcxe0uzqOL#Tu>m%33dC8Kd){HY#b22bG0zS5;m2y!^yh0D3Q
zABP1olm#L5+R{R-X~Ms}$lWY68vnTm6(J7%B6osJTR4F5(7Tq-Sj3aUX5dX*u!G2p
zwOk&KpXQDTjt`qw*-~jnA-&V)l6$mLQQ6u*Fi%c7*`>rvT<+v(B5Ye-j3x2c2!^;r
zBJZ9MeG>1{KI$vpsg+C!FW~d24?Gj)qV5%sfD!OV5tF<RM^~qy559eQb;-vTF3!)!
z^aMg1J>q>PXcN_+yFE}YAY}7S4Z%W15j?m<IFIB(7{1VCXA8S|@}vdYm^-p=-2%T-
z@in9Lr&U%2-$#N0tNPhV<&p|ELT;sdY91f{3|<o6%)9!<*P$4Nw*%J@{aPX7wqS>W
z_*&r<N*j)_az7AWu)BsGEqOsZcS7Zzv+%+!@Ed)7E`Be3jMlaCIu}g4-ILhw^i;l{
z=1r&K`zDp<l;WlGlT##^FClP8aRbgW-&t%f#_Gzd;4vls_5%Hb@IbgBqRkGz+iR;<
ze(Y(Uwe6i_90B)QcrxXk1@u7_3>@`@!h(4ayAN2zhvJE&7ug-)-GQnb0(sJZ^zcC}
zD_<0i2z%j85%{7c<hV=}+wk!4_Jl|H$=r;tcy-Axl-wD}Zao26xGbDPaI63){ZlK?
z5ZbcZot<fzG4tij71)8r-6rD}=vkqr{zZR>Dp;^49EGoD_bEF?yhB1|z0#cHxZg88
z5u{CwbFU2iv+zdcP;{k{v<4jUjzqzcmHq4}L^;iUU>pnTBe+!0E43dAuhZkcV5o9Z
zCe?@G3%Ug!Qg7_w6bm=kH&#@B(Nwjq`;&0xL}{NP09IHOT<oMFkKSR&M_JP6X6HE?
zan-v}QDlQxly^pW#x6JV1rJz!M|Y$wv>(b&U<*y-ct?)e^1anc!;9urce(>gN#E(D
zewB*|RUF|wbOdDsb7EjaU&b)EM5rmsXF7F^B29QOD!8yym7QJ~%81Yuq3pA#&wMu&
z3W`S$Zu<x}j!Q()i~PWXRHe^u8|c`vVEx6@Pei-6Ro`QmZGGd8U^HWV21hn-Z+K_P
z=H?62TgC-L9-J0@eYEbayu-}%Jke14EBNH`#WX9GpH-Ov4WbV~4@k=jZlw#B5e_)}
z2vx=&;0K)%ESLvRC@=a7I0U@B%R(_xf{`CMOMWPn&r~ON!$L0;SRPH?-rRxKm}|;D
z;W;g|fJKvDgsaYu(s;%FA6SUKAMAT>t!6X+aIX;QAoM3cloA^ocilG-FoR$0ibdI2
z`z*>%cB^m%;r#5l@BF%ZXDt>QGZsd393RKsdH#8=I-K`~+vgVQudm(_o{h!JSFhsp
z&p)#w=HbIf@&1#KqOrId{qxuA*L{tPjEy~&f%)&H$1?B_JL#+~uS8S0&W@C3W71>L
z?%rX%+}e-bmc~>s4w#Nbb_1!rv<v;2cC?aKbw=5Gsd}-10;Ljk3Hn4kZmccE`w#DU
zaR56ZfdR+3nxA)nP}?az?j552MD>-XU^6wLLY_wZU^BjW_PN76e*B*AQJWO~LfJ#>
zF=oJr(B9B1cJ(w>R%jq`kjx-M$gwkzS>&ie@PE<Af3mx4-Z(?Q$fH|x*CY6C1undo
z`X+TJZ}7Bph_Xxd6kPi!7nVnH=jTCdUvs9{@#x-0{NOu}<L<_?;LY*MN0wiW3x+72
zI0~Baj9s@VM=jerMiJ5}3W>|;Di89$bN7yTm?hH`{Py>DV{7xJ?Z`5?0}Lf2#1Gbq
zo&$YaUtf)T4;~7~7JanhFMjc7@x}A!Rv>=wyFZA<rDfskxp&^7@CDbAwEzS9HRVQm
z1XY0sa@VHml)gCzm-6BcUxlj`UxGEf=at4a+9hRt>f3rqlQ9t*josxHj=*~`ep%tA
zK2F(bzv(HI%IwC~xGy|oM^dZZv7)dq`sZCvqBF>&*y)8_@mM?;{PyBPBkqV+F(yL0
z8K2<ufG38qjQJeX3BOephu%3UnxQJH4Ok$+(W>xOxjA4;yaX_TmjY&c!uPiN7qDQ4
zNk0+TG?pBi&jJ@$#T%duA62?l4wd61JOF$Ud<^sw*%rKL){WD**!|2RaPZMQsp5iv
zj46kr@dCH~=fvZo<YIvZiZwbpG?y`#*)S)JuP!y>(e0I37Ja$85TBuZ;q4mYmAR+)
z<*Tju^G}~fx6jeGil8>*SjDO|qL0wd!%8H((4f3k-+{fnG4HZ}JH?l6<@wF<nBPc4
z-@h3O_nWkb&AZYZ7<qlA|B(5n;fBp$Bu+J7R?ojK^Yj1Le=&0F4}bR4pJ$;c90&s$
zmJ5!hzyp7{-RcGc@mB))%XY<AIHzL=ob=YiM3aZ`1Gi%Lqd)Nu#XY}o-u6&F!{PN+
z?SpTe2$gR+%aUcqGsEJiPwL|+#jn`-v7NA#IUV><Ba5}jcG0g0kl!@`aSSUyN~rp+
zu<o3W&x_ykW)S49Ct;_@@Jf6A^S7SCU$qywJOAS5aJI`@-D}$=Ff$*1sfVNFHzs3J
zFkZx=Y`~SsRHpdZz$y?-?MOO=6>(!X3}KikiiE~B6C}zYcLoq<tE~9TT`v$=L5QUD
z$N4d-@`^Q0@F27dM64cUg%jI{@uMylu!vYQ5rRQvK%-0yB37bM9+f3mNMz9y7Nx9(
zpus@EI3d7Eg=RiLE)@?5K~zq;F^IASm;n=_4g<;HLD?B3ApRcAGRrp%`TvE?mqc_H
zSt;jl<0_TR-!8+K6`-UsrNL@VlaY;kcVk|RAuC_Ky4QWGqNjfh4}=B`dQL*lUdbkS
zF=+udoKa6q{9s^VJXlHMV+adw(PkJXpiIIrv;&NtTYxaCh!tHhimH>#0Imp698ce=
zl!A#tV8$RW9H4#ilRL~h_PR2ZiN1wzYFn6HCK7g2`jru2#T^U~d8Cj|1GeCSI}<_`
z)O7_v01y*s^7SN1-;|5A5e|8>q;x4C`4W%xz*D^hM;J2t0u%V0AQ={=JnkoMfK0o#
zZBAV2S9vr9cL#VEBxL|6)USY{e(AD?0fn`nKY>Yc!!CD!VJGbHLm2&*Wb>_W8{+AM
z^7?Ji6FviF9>uM?^MJ2J$S;@EU&TA|T8|y8Aq`&oA;aQNx(;rav=~fS2YIHw#2Tst
zY09Q@)-o1Xmyvh%=WCFU-kHX5-udwJMm{|Gm1!APrN%hR*f#@~?~H+Q8pN>+bttLx
zuQ=J>X~)!dO(TC43^i>#mz0Xz%^2zr0yqTLEM?}G-n=thjc>)K-q^G6W;nhp-c~o)
zQE-*(P@OW3EOz!?yAI05li{(aQ-<M5qi@2Lu<1dVtb29m&wh5xw)3Zd?VMkFC+gPt
zgAgt|?}59m#ohnuPH_{>36J4j(@EI!Z$2Xv(Z5oG=J>O^@hHO3w=DSw-qt_-oNhJF
z-GBbKe*&>2<p)1M!^Yp~k-J7SK;lD*@ARt+=54>?=I?}c{Z8Ac{>r!OOh5-VOtAm?
zFMsiW>Z1;`Ex)<>>d-*huYpOIbRe(fi3$#AJoIsd8aGD8M~&#Y6ahy7vAy5*k!vqr
z?O6D9c-YrOmPI2PcNu3C*E|@LZ!1&kn8|TZDOg}k&9N37(W7D{4(iW+n|JxvAG~07
zV_Y|1$M~ms^hI!jG1-rJ0IC9ikzpAB89!N(k5D@WlDOCUW_k0drh24~?~a|G)(|;T
zgxstqKQ^Dr0<P|1H`@EHXt&y)<RdiT*hJzW=zu3ch>x8W-{gyuh9iEOjag0do3Xai
zjN5uRJ`Oe+p(qb!V(}EhkaNN4`1m4PouhcQy%!vHhtL5*1Y<qpg?Cx8dkbL!rw<55
z2)I#buv=$BiKBoaj2H_aUv%6TU#fWUBk<wC#5|n%DqzDMDC|T*L4|Mz#T>`qaEuzO
z!P!ZJ5&)qnJdl-{l%F<L8<HjhA%u1aSC*C+V|8^o*49>Ih0D?fU+z!gXe7#q;)0{@
zXunkMla3ga>nr;;8_k#$FT<TGr^@R}M8kr5@j4dbBi!q;lKWU`goXUvG0QFrI7Tja
z$8dxmg89=kr6XPyfgVD^o!!0I-r1F%-BiF2fi;VW;NO_Xz<;s3?v@weEiI`&%S$fr
z?%s~(0LQV=oD=SFXBBZp!(N}o<#9KzPI|Gf_P)EZDwqfd)z(u=ch-WlDGT_IRi@t2
zd2H{t)vxyA0ELi`Y3zH&wJ(rI8NsOmYD<vZ<DPH@J~4N&T)D3S>*m6OaCj~z)h9T0
zfxgGQ34RxWu2-<@oBoL+gCnCjeu!Oso=+)V+LGN5rz%@leIEhZ?tVLVclTpwXHWIs
zx3G@>jN*@R*?o=ll?FTc5Cku?GF*7W(T~8z_kbbXR~`5!Kj6kOcQfGYsCaNzh||x7
z%NAk_CpnT5VIytJ%Jw`iT>VP*sK*bYc!Y=ceKdM=TqBA#gpns_98H&ZgY`OX-%G^K
zjl~A{{IKIu{bhPw&!p@#7I1UtA9?!tJ>`qxC^!@k^c4*K1)kkJynuml68z@gD$2x>
ze?AIJuw{25igouTm3dP1fH?xRf;+-?wstHyM!A{u2-<<;>vC5E8A0dRT`|{am=+uz
zvwJhk0v%_*i4v$P71@o#9YfF*l<I^@rJ>3rcmqFnn{u}jN*WfL(05QcEG?-2>WxAT
zH;yS>Ry(m9#XByQ2K}2hZqAEds?Eqdh02=m3D?1QlrWTsT`MRv6R%Nhqi~~7%^Gf2
z@TxC$IxW*E=o$Tv-KqV4FX_74j62G{_~L0C?C(aWwWqYYqASbs`0@L(wAgT4fjcP8
z_jWg9Z|7CC54PgyuoWjt=S2PT{0xEqvC~J;KR@5J9DsPK%*lCH&h=URp)o>j$RbB}
z@StF%ZP@`hCK$3fgdJ!;vX*a+oAd|d3cP7I7EIPTByF3Dd3MC6%mG^Nc9%(63*7|T
znsNr(FOM8cIYcVym}k>hL|4>K)ThovJx}#0#x7*$xH*614keV-&^pF1jtB(y=c1R6
zt8^{@(1VPFahr7LhumX@ARpR~0)QiQp)Z`8!cNZ@FSbN8&V^62K8llZj~&MBe!dj_
z+Spi&+j8fQKhNd`1M+8g4+=vB;#3?(kFw1%QS_i8{9+dvG#Pw4(f9}+?H{lUP~$KP
zXOw_~|I`#^<7i9KA^Mzfo1Lw+gQ}f0eH8k^aePgUO-rT&D9OkR`h?P<qp^WqY3#CQ
zC;PF+pNp#zFZ|$pYI-*A-hUK#A3lzi)!VUt`(A9^z8@QR9>n^_y;xbltA4ty3@75~
zg!?CRj7%PZov0(Kw{TuG7CwrWx)v`5&qcNO_7CIv=1#oW+>Nd6{pg7oVOQ8et7Cc$
zU2cg7>#+O#O!zLCwAy_=z1Z4o$L7|q=yc0tiLVl2*@o&3UPG^GOO8<FcxQH;vD2Jm
zo1y#M&&tldm&$)%{T5n+8_F;>jmE88dB>T`v8Qoib9*PYm0wT&pHp!rCe<${G@dAr
z$*Eb@Rqz%chGNaULt{aGZb3N8f)L@Gij;~f79G<^a?FLtQGs)jh}txd44n0tF{`#_
zyg>=n;}}5EP{w4&251lp_T1rhV*2Wng6LoBSL`yhQcG#VtI{VK-&3{&tN;<?CH+u+
zSa!=1p^>sPzHygg(sI#b_#&l8pTKQtaY;A_@0SZgObe7R3PBdXAooOGcB%g59aYK`
zS^yws#T)quLZzF>A5M9kq#s+pD468<mvn_m6m@HfA7#fUi<VN}32a$(eJXrnvCOD>
zGVY^2k(>LR?C=CEzCc07A|)0~F!n7d|Ai(;HEJwagwAtE9^+Q-yoN>rOlSr)8v0Co
zsLHa_kFs9ygY!xYUKl=u_O${}0D=GZ9lC;1Dp$M0Hx#!h#5qxmzM)5buGJHsQ-7t&
zSY@Q41OS#s^;5r#OSS85+t7D4_8a<k+rQ;MFtpQ``Ms5YHShX<(<RN@;i~yEcN=+q
zrT>umrs0OoUnEX7#QlrM{@t5@{jdJ}e|qbWe)_YYyI@ro;UZGdEG#P)uvz$I7rfIi
zzj9zE0ShnO2QoHs(yxYf=m8+=ZyLhum&puHe&>@81Q99@xmht{D?bj9k7on<b{-`h
zKYsDXnNDMWeGl99X9XPJ<&`Pn&%gZrTfbdgzq5uJ9y9oLhCJ0#u_>ngGES{m{R5T*
zp|k!R%Km)UpOQE3ywho5YB-3WKT7->FhNt@^jkpph8cXz;s}{n8Z705xPoI9@@JW3
zhoO^~$X{G)@XA$Xl!w)}8W14L5Zhx--cL0kDGW*Sa+V6mVB=jiVx$mEz<@GnK|~oq
zQ3{YBgQ^t`dYlINQ8uST@hBS}IYGeBZLjaT(<?77CknoJ<se9byqW>VTi;bEnJYfo
zPJHcwmDMp2d(vWnYl$I)N&Ra#G{Uzqx0sBwBKPjyXzD#7rjZGX1*)pICj??xFU#bV
zE`u`1OocL|je!XRp$A0`$|%Onb;?wCgciiNl8AOJaCd_NcHl1Nz!MQU1ACZo7<NL|
z!IyH;CO&>he}t#aNEb$h^3?WsI>ZGJu_+<ka|K`y(lA7$VnE3ip%x4bOyU?yCFiXC
zl{U->Th|pa&N<%D_@RlrOObfOgJFON-MqU@*}}9?KChG!^TPxQ;arvp7y|E{G$j)Q
zZOXb`GBMGgYfJJ}NQOAnlRU=6z|M)`oZ<*i)g=RFU3sX7f3c_6z@%@rDBlvqZsncH
zL7VbDaN~#a=$+ri*JaEymw%f<+PtB8RNrZkwm<N7*tGLSU|XtF8YjYkI`TV+jyum#
zk_ihJ%*O8W;9bjQm-Mg$Y1_~CjEgUiYgfYhmoWAyttC5g?WL%uErXb8O_xj=xB38Q
z=b0Hh%{RmGn|4x3UkCI%>m&D4-|A~Se){3LsayFYPz_iN?s%kEO&cerT`MT#_`GW(
z`GHry`S&aBE*a_CG2i*%?*9y1pEnwltTVjda?J!~{DC2y;6WV20T0qhTRy_OEJ*%2
z{hRNbB>5|L!qB<-C-1>yUx&kPWu)ylaq+KjhbvDS+{lqK$~6;ld@BSTB173mx;5<^
zPyZ0LzBA1Xn*r=OWGnb)TXAtmm!1C`PuDxszA;W0H(vHDWwN1P4?NPR+VxH=b+iKk
z`1g~T9Dm~!U)j%ZOv3%k&DZ(=hN%F0`|&GbRz8hN^ildqEfouqJU(fxV<!<SsZY<Z
zqIZ1Z9Vjn0_v4FKyYX^!FLn+(zF&jgLkPw!2rxpFex;n4j5Ch%#*OjXtK&7+F*bQT
z*HgpuCXL~7nul+XlX`$*3gnaz-GP|;Aq?Xtd0{tiqWoMg{n=k-@i?mSk+4o*{`JT(
z1YgUm3&Bn=R<csQ<D+9Aom^Mqy6op1k#j13l_SK?%`<qxl=$2fD2lHUPl!;i%dVaN
zk-~A$X+NeXZ$)EvGVZJ|>fwHmc?<fvOG7|W`4ff+)_R<TphwC19x&j|jvRPl1lFuD
z=0V6~K1ck>!t9g<)7-=3qZF0b1VtNRCm6dL#)JE_I7)_No1oRKsz<PbU~rn9GkUxe
zUU&f9P^uv4_XY7Pqxmd7@JH}H2=F)p2q6bbK`ZOT(*r*Q6To{CMS}QtRynUNAv~Fn
zm8IF39ar04vGeI7#?_8vDi4CtOWIg=lpDsQu?e-`r0-ktE-m3l3WCOiJGa{H!`MGS
z@Oh@P-11#i2%jKW?9O9R5;)xzF7F-oW50VGT<FeyF6=gFw>!bzF9+r5Ep`EMaXh;M
zI2!Bd<V5+Z4V4%00P%nUN*e_3pM3Jm*xY<+0Y7a5U#oy|b=r-Ct><yLzwM*65FF7~
z><ZzY50r=O9y;t_`p&l9)?sYz?gz(pCC{6RQ1jKmL~#XZ<1V-i9>W(RU`Odp(U_0W
zb{ws=fiQQ0BfJoH3l0-%cf}zs+K0Y_@Qk^S+p{P`*y-pyp17;dLa`LC(LW($<O9bK
zJ{$^GyZf!+E;R(z9Gk}+2qg#in1MI+9|W>pc8Lm)xA$7Hy|ZtDDvGq8+`$zTPcy=a
z@e$Sil6$DaiXsGz3&Nju_0M~^SK_`Nj&Wr5I?7r4HthoLp~TExOzdJoNv(EMKV`t8
z?g-`b=)G~@DYx95joWKW>W6BR$+4Im8;KdcXP^P+C%(T1C5m}`^-Z5pAzT;{{20T4
zrRR473`#^4=Iqi~SX#0|2q9ymp|{!^Tx8B9aLUfShJMAONsdQ6i@n|5*xlX|{-yGj
zxrb?9!8w%OeDj&gJ$51{U8K)YHfS<*iHC)k>>xEBBNP@aI2Mp2wcIb&e+*we91+Hy
zJd~;1=|uZrUogECKBo{FT0z;_v4W5np=A{i^JYgVqX|OS%~vRyTAmLxpJo>g3WS#W
zS{_^4_fea)5q+6E5xEbBMHbM??VZhd@%(ea;#Kr}?KnGTXWpTe8;l(*tLw3G=eBp-
z9k%ylZ|jB2{pjJNSYBSY^1It@Gec9ek43Mu7rUEJqpx@WsH3MPc=r@Xuva{xo)uXr
zOjx*jdZIbJ@RlP=HQN&2q`*5BTdA->p@Og+#TRd64;*)hLfJdwn1_iz)G-V=@nD!E
zvk;WsA1LAE;{FUS`H?@yN@2m-aun6WW31W^Tqg|U9Jl~p<}YPcIQ=F4$aLttl$$s_
zo|7u=Tcf}n1z{?~{HOd;uyX_)N_KF-$HfVs&b?rOou|To(W2Rz@mNy-K{<+|_|_PA
zT8@iv2-Q#FR5$Jc{OE(nar@3)rH=clh2|T#*J56Mu-|RT^>s{%9wHw?9^k$zSR#z~
z*s5=M1;$72N<vA^eN>F`^b_#QiqzgwDth~(-@<+RJ<4fz^FUwNvB|C-6hH{s0h<>t
zD=b4V^?(+vtt_a&&KR#*^n!AUyAk^c<%P$LH7t_YQGac<kK*E59s1Ukk8f-rqKFcm
z>l}+Vistpt;-GU9JFO$}O}*$Gv70Y<+OQaag&Qa5*BaX*PSji{7bBtJKz!G0jR*aB
zwWGZDJ8_^ik2nfZ4|jxa%bgu>2OW)LqMwJo+=;`^aLURU!qJ91J3HQ)hcY@BB`5v>
z8~P*T;K{kh>RXdmc(H&4UZbya_WGw*u<fwGM`I&OFL)4k;C3{2Leq~$Z{QVKFv9L|
zz8MSPT}G7V_|%;G$Gl)U9%3hzxA0kYzBSIS(J8@SG(+Pacecd{yGS`MS2PITOMMZA
z(1mb@orLo<lX)MS#!~1h^ow?6oY~yoj_s}OXzlNN45SZ&XV9ON7ol8&KIU#E`XaK)
z9RD;Hz*EVtPNuRd4*da|t9}g3Z0LLPbD6*^rOSN8i>#p|!Xb7h!9RgN9Mef3_AaR|
zM{J&&Ph!CrbQ;A!%3^vibOQWl!H7=~P+8cG4L`@5-Qvj4SO9?%2>t`!6}}Cg1AK>%
zp{-jj+N+~>fcHcu2TjqVdZU12w=s*CtZY}+*uBf5FK8Sz#tW@Qk637evUpy7)4P6E
zQp&}`6KD%cDt3gECNP5jLrW<?M<#PDFk>Qhqc34BuT#G4pa%~*dbck2;^<G>R6Sd?
zK<Q~b$VEX{(O3WIf@i{Y@9G!LfZu|anYNO@?Aca43Dat9CtN89<~RI@-WZ3}cXb)|
zBQ1Zu5wAXJ-j)6vMs9DrzB8Q6m$}=>>nr_-%r^};Z2lr~sv+)QJofM2Sn>6<pZ?L$
zVcvv>a>54k(BNsu*iG2%FzUoqgw9e0OoYgA@)M?j_)dg)1hgS60~o21SGqYNUif2i
zaNf2Iee31EMc#?zIP&K#aMufG%7tSU?m4ZJp2MoQxDGfB%9LRcD$7#*N<5q>v))`>
zooUxJYMkO%+%i(8<CSoPaazF0x!GP}PI}mgGk_BhZhY!-f30=WE9;#P;e5Xjc@*bN
zBiqZR^0&Ry-}P{?Vun7PmmZ2(4SvPngoaNDX9){~f?}i!8)zlhTkfcRt{g{<oQN9?
z4aaD3f$td;6fqYZAIL?eqK3}iT$1vuD*{=8$)$)Wx>!+}6QVr3nw3ltYX~`h<j=sx
zju+f6VdgXdSc#x;l-<f2;3I$XVenuwkY{igfs6u@4&eX_lM@6WtVkUQ1z-qak`>M?
zD>S)au%QfjNi3`J#NfM9e~m`f89eVY!+vVFDKV@|x7TA?&kWm?)wZ-LZEvPP10QcD
zeK6o`I|qxoWSIfff`Nh$IdRVkJVF5rrj!k7U25Qi8E2qHxdg)kTn$sT5$#k5O_+0z
z008E=BXFVp<t)FP%#*Os5LR0<S+koAgRSw=X{C_QjE)TXb5Dc&2xWGM19ibygaMRG
z6IU@G;0_ZBE{~_4KI2_~&`6SIa_H4@%9nZ?$CQWNsEgbvligKD>c$Z)&=JaHbd(*O
z_POUO1CXU`Ua_aN5IAsTB8;q?tmN&<4DBR4aAilv^aLy2#2ks)gf?-^VQpwo+*XBi
z-GL8pJ7yRR>G7HF4r0Rs0eX7zbQ)M7Q<)Z_^zAg$Km#03VSr`zxCxQ&G=W{P5B^HH
zwB#3MCoda(Gh&vdQe*N=H#_slxZRu??n8bqn4MjB%Hs0khKDU`+XO(y<*$Up&B<aZ
zCd*^5!9&mtLi_2<8P->p8NWf<2y9OR`CsAeQGU}qKlH=D97+{soTE%hTaU*{eYmWp
zp&gW8`56@Y;gxB~E_nM72OdY$PlbmZve%I||CFJ^8h7GmUbYu^zcYZ%jm?;wR+qI5
z55SwRQVvrU-ZgFgW2@PwE4G{Or0wZ8{HBiH?d!VAg?>mrHaT8rsq{yFC7l|g8uF;+
ztDb~%yUEQ?L+~s?N?EI^d8J!ExasXc*qxK@4#3YKk&<?C;z?z^9=$nQ(723C31hn+
z3y)|!g5oBNAGZ|_dKAI&2m>5=(+}-W02~H|$*{I%#B`;GK(&M_@v3jeDP`32H+{#~
z4KBUge4YQdF%^aw+tfz1D-dUi2CIGRD&`}U1MOby?sZ~ouN9j+2k~-q*8*yU9VeQ6
zda;V0lA3;#Pl9|TUyu)i0q^QQg(qT1zj!?HVjPR50E(Gdz#IDKD!bodSE_m}bk~F3
zX(GH)G{WR~$HPi`?p@G$uJ{y$KjN6K7E}2;`g95o^3;IUmBm<GTCi~2J0&`YKKh51
z%Lq778Zq(b86V?#9TXQc79Jq@L;#I&nq4^vt-S!M+Yjy);ZpF+i?bM4A8XD{#l4N?
zSX)+Jqp#zF3&T&2+?QAoc8!qs<Rb4`(A<7>oE<tGv7Ia8;TL>2#1zNIO~lILTr4ll
z`|c$aJ*&%$R(RAf2rCcvTNZ)|kQUG)IAj$ud;&*hArM69JvTERi?b84pr<i~f&hLO
zh25#(j^K!u$!fpX;%(hF$yew7EF1wOxT`%lstUde#oG)!y4ayJiC}3WIMR>ZGt(9f
zX&xYc=%g<=vm1xJSe4CNd^ATonqOBrS%uEghTPu;Pe_?L%4b~d!n}ct{Sjbu{8VFh
z%JuOIVxbty+oS$j?6tb_YPS{Jd+pd4es_B(!s9C+01Z!i44x{_T!e&BnOzp=Za;Qe
zU`NoxPB`WYC@gxtTuqLUJOy5<0HnPsAFI;|i%_jG-&99oCB1k*wKd1BUA#^OC%gH!
zclKh7)$%=#F-$zI{G5-jRynP(k_Fn05FQ~eN`slX`B++6^_?sPq@Va+uf?X?Mm(k3
zig^wKDHN?JU=VEAx*<Hwqu=18)vpvD!86?v9ANc$ofl!)Jvlu+8&kr4;CwC^vkRfs
z=~@Z4f6$5j)?w^=7u{aG+S-X%+q>Qs03G43oxQyS`R&KH+__Vx;~hzT3zt!pu?qqP
z4W}1?2f%{)-x302J?y%fR{x_c;2{E8VCQ2()#p=TfbxK&rY0<`<G4od)tjGFzZY(*
z8+*4UyO~g&sDhp+aD@9UXozrK$*2sXk^cvIfAZv6vgP+-9+?^QJVy>SzIs!?p&LMV
z1E87M8n}@eR}eQ^5unj%YyuRtkZA*%;Xw;Y|AZD2B|s9fp#A|Emok$<K%=|gT(7Ef
zj(MIVGxhm=&%M8h%vV`M4@J@B5%+if`s~M#A3uJM55UE?(C(^!d1{r%(logp1&|B#
z-tjX}A5&O4n%#DWw{N&h@ATyQ7&~!JT+8uUJi^&J*!IpfoaaiPthaLqg@e-2Q)p5I
z!t{|WptF*c)&3b}7qC;%>Jufrb6?)+tF*J>vQq~7<_;X^S<X+5=k5rFrasb};0|?A
zap2K*3d?&%qWDs~>YOh@n<=AdK6(1AJ*T}@K;{$|${j^GWP8XRl+tPHtHQ;!Z6D*;
zk*KAFaKV6g<p>MweRQx7zC3G>AAXWMifrGdrA2Ulkua~a(^H}O_~a~UylkI*`kVIg
zZ$1Y9+&Q=3jt@55$-y>b7W9rbTwGd;B4+L0J2@)RBSPor7b(wh;A33`Ym|P}FT&nY
zcSTOp-&8Oc92H<2&@{vMI1U15cWnu8=exFz^w7utWV=gKqhOPE(55vO)gqAy#)Er~
zZ{nMa8(fSaEmh&>t~}?hq(+{?E;!(L$8Vg4r|AGc>812pd@}u7xGSl)Pr1{VvZYNF
zXxA}rH`A*Hsn8$FV>_xu-QA-PtMma+9OH93${lx!D)=u&g7Fl5yA$y4+FEE+jxeR{
z?$(u7N~=Ep_}4j>@!tJ4(s9S%Y0|V@9^<PFNBGMFSa9v`m~{NjNh`@qo>N&A^`_3^
zoOC2ReZVKjbaw&?9OVo03BgP7*t;1WclM}@^jtyx>dJC<^tywm<U+MTpiaB-lphs=
z_J_;tz!LvdI(fh0@yU4<L>>X{PAl&sbN7lnva(x?Cw&wAu$|rQL}#GB-ThI(I=k(K
zJMCtBOJ{Q>1MarL!G7^3D!Ei3tDKYf&@K(AF{6|?o}yHJw(<nzse}E-`=30TPeq>N
zu6MIR0n>+a&y)B?c}f;>2iQR~dCTYvKk2mLER$`PU8KtABI6q-BV}za2$Gjj5${gf
zQp}yx*Hxgp6IaET3PqJ`ey2{*Lc^=9lJ2SOsPWld+Kl<o`Rq<gUviPr9`tK3$|TdZ
ze3pB%+mYpBjwp%17mY9w$E5Wx6p_|CPTR)9&v76N!KjBcI*WlwPdt>zF;3A2dI{si
z#o0;53&$P%qzkZB7&{)h+ff`lVtf(~!alu=elHH#S5%~%k95MmDUJocyrW?2!Z|KG
z4k{}%tTfZ{K>W^*O4>N}BMss1w5t^8jzuN=kuIpvlJ`}HW?l4mho%cj`W?yUBOY0v
zPy}d-=?CV(*rPPvS!g?u75&k^Dx9QO%4qAv0Oe=>aoo0JXdn2f!rHfW5GGJE@DDm;
z|KiCmet0*Py=Rw0QGl73V*;QHJ?QBl*3_+Nmru3vO>^_xc$%ATz5R>geYxLP3#;+_
zbr;<_%Nss##(mA;HAv|Tf@!`mov&!VP8y$A$8XM<^Xrd_uc80^KmE@^mqNHQ!Jhoq
z{aV;8IFE?c*S8NHOCQe9dg+uj{mk(^e&HFn_*Ed&(>FebGp`#N`5Q%6LgbbB*Qu9}
zouE%6rVC@r$wwP*+WOxlmI4^3!gv;~=DG~Vk4yO*v;NCDQDaL8<M`P{iC-qQU3v*F
z(c@e1gg2jfXn!q9%51#U4B~3}Os@-5K@u;@CK)dsoy15!Ot0oCW%T1oS|lKW8Ki;f
zz`%c>y6MEB^7L(bDU&cSL7<FPF^Z&YL-6ke$@DLYaNd{{T|y{f0$(Np7*!bAZ6-cf
z4u-A(W8gjBJB)HFgMvjcE7BWgg6J`O4oq&bU0GgX0`ID;(c0$BgeI#_VVY#nB;+Xs
z`8YZ8?WC1g%BYuRtCe>$91esG_Y!vpX^49XNErB3bV!h`8;^>MFbLGw@J<FL;{CA~
zD)Yh|Iv~4ZlmTyJYpXqZ^b~i>@ZW_I6a@}Cb@#rB7dMUJFvARfOAqe0`yafY9baQp
zrSu82%)69ICI|B*JZM`Ph#1-!{)wbsg-KAv=49pt43bQ~+o)fOqun5S$)wqIJ@&wM
zwfz+;cowq)0t;I>@lw<#U&Wucg?4<g=%0x$e^JD}>X?3+e?@zwA3+9#X(pniYaEqj
zG8N*o%CZd1VH{p%<(|h@sNgbuCTgT3^I-j*bYA5c4p+F8Fhn;Q9OKJii91eoR8pC?
z%4zTT@Xj|GXJI6R>tv+wLmsbUouXX3HPTLn7iE``kSUholgV|*ob@wb8Dtq98JIgv
zW@LCSU>HX4jJCTAYi(_LHA>>5MKV7kfyQ^&=&u?3K0bDOztR9=K%KvyUFhzo_b;1a
zq)mljhUubKIN$!*j;5t!JSx?4<A-`*-|y3h=~MV^Q$EHwZc(%Zk4d6k656l*D+qpA
zQiOp?=(17PB;f}?#!ZClAZ6E>Pl6XbLT9>wwtMemX8pIOLTmV?)5M<o>qMMn=|^wZ
zou3XS<yWTYlkdJPmCPVsX9R^N<82qJ4`X^Mdmny~Po-DiWsDoQE?$>Z8Qw1p>v#8#
zwJT#xm$p7-wf-GEfPJOTx8>0_IN^7nq!$0{U19h)fqeJjX!le)gIBrd-3RRQOk}=O
zmkJk?Aa91{i?CIeaw+UeDlxtN+O6soJ3o2H^Jcs*1a9Fq24%hu@wrA^HD(9XE?KOk
z-9@{G8)@?qSQ^B@nuP1~(XaE*UpxoK5~t05`UDcB4`*hb)r{j#dItBMFL((3@lILG
z8ox4zE0>DX)vNRSzx~U9>E}0M{-Sq$-Rz%#-KKBn&o%H-kjKE0p`>}-+X<7&3qCx;
z@Wo4&Aay)etrk`w>OvLAQt_o)EA2;nJJzIMnV9V&FGwDaWmwZlRd>c%_?`iMumnjf
z*8V5y8Yk)EWEjWX2u2xi{q|0z?5<$UOB$3T0tm+K2o55Yx&_~$;=+kEWh0!6tK?Bd
zQ^$G7<vfh*-h;KcIM-(9>i(JBeFZP-u{$cWrr>*~JuYxu;k#vbw~oRyk4zd4%-n(G
zD(46^<TuDa$MsRBtZapaVT5nfV|Usd>Aioib|gH0^0*xz9JG6@D{W?BslB{4*7o2T
zJ<7!+kEDC<woxeH;wgn&@-ew*Xxtqz6K#3c9q6~(*3Kck=0yY<9!;YV^@6f&ZSS|u
zt?kez?>|X>rY9oke+f^%y}6MicZMnJ>=gVp^&Fvo@;O(O`b2*ArMs)Vzf`_KZc2E%
zBk(E$J{37LQ2>%BaQ8Slg^A&rzy*?nN4XWjj>?B&9OTs$8Yw(EI&eo-9q*-LDeW45
zUg5vOK3C^c10EF#rYAKBKd8cmG!%vy*JC90>sJY%p1*|Gy5Lc8Ry)@86KgF;lpyfB
zDq^#$Ux69Cr+n+IN7MmpLwj7BxRs-Rcogj9=yuvh!JM6H#Bz~N1oea$f9B^Gf(InR
zzrAlLJ8se+0M6;<%dC!HfAYLN10QyFm=h^1CoTJuJFdL9LxG|=oLwl!#W;6SP%4k?
zT`Wt>OTcTY9qjG2mltR4{@v9ar#bZMisW9V4TO(v<w|kyxp5``F>q4heR-v@Tv&s@
zw6(&oQPRGIPZj@4>7h{fm_GF?3IOL$-fLvpY`<O+<8f~CrPf#l+Ziym*!D?}L``49
zZeOWIgYHK27&&m8_E0I>?><zCD?aygXpgy4Kq;MgDT0BE3eTKloYU5!NVJ~<xPX~(
zDF<*B?h%xOEJM`Q;}lhvTIb#UeFE2_F-IDAbcIJZ3l)~~=kVF~N%NO3W$q_04_;a4
zaTO!ZFYeUw{obkLTrhM>Iz~N(nTkf~tb%usD;%GgZ0<7jULJ*|;m|2}cHU417Y+zp
z;g`jswm<N-Pbr{Op-m&?u4Mb8HIP;mjVy(Vf<EK-iw`s<%=Xx+QlzLb6-I)!7mC_G
zD#Uz?hxP+^C8!wiUKn>+I_L37fA5od{``5{-Pxh+mw}OSoJUPg&8D7~M|yOj;`QjD
z9Z_c$tI|&Cvv7>EXSn*Ccf4Gl@3&7s{%QO4Hy_is!)<bAu6_GEzti6P@LTkor|rqp
z=d{HIG!FPv&!ZFXa8jNCu3nzD<NXcld`kOx^zKxfUs`PoEBD&W{3`XFYcs%5Vf{H^
z@d#9RAt<muI^0VdD#FJmXWH!i3OL{qYQPzIc*mbOY5NM(?CgcMi;FZz@JQIJOc7@c
zYd@)-vGgg%*%vCwfu}p^6l5DW$Do-8i74CAUD<|djE+r&-smr_I6oJUz>kh@nAYvv
zE(Wzv)_Cdom>q4@KV?kXqC0&IIx9Wsh5GgnX_)xv!d=ILQf#R-tKB2gIgh=BbI2l!
z_ush>y;;uU3Eix_f@*2l3}fut>SFq}yHZ?$AkB8ihp>72<WYP6>`8lvG!7a6o~`fD
zrRiTIQ&Cz-158lmy}Oz{8gymBV^L>QZjT$>*z|bVtt@I#xg>vUFAyH~193tHtGjU4
zR+e*5yvmj*S!i%X`!Z%xXBV;FU0rBD{QgJny?5?~X4;2UG7T}8hAY8t9rkj>aqcFh
zk46DXzY*W%4OB9cb#Ov{B=<@w|G07|-37GlW$HCQ=i-FIKX(VYSY*_rUZwYxTfu+N
zU%0E;ezottTIcjF;9=WJ*BrlUv5||tN7~}TLXH!bZd$+R9+A8cE>jk1LeE!uT8}ar
z<0?z@&U}@^?oiFq&BXT@%P6~)ml0RljVBy=2c-(XqsY0u)BGZMs<P0XlM+s8nLC1n
zt%`rYwX2*|VO~lm<!|zkp}N$?9egLvyAo{=Wg7A;j)U$B70&h{#|w9TdE}}KA*9>e
zz$ZJ?D33A_`-`$e#|(E^XMqFm*}VtvBHgx)dikmNq>P26Leq}HEfr7FEcq_!-Y|4Y
z8tPH2476=wcB<XKf0uD;BK%%0#2$)L$FW_WNVwVORfM`A!2S`XF69$Hg<G#Q%2;Q3
zWuw5(vI&3twRLu0C_kh^Q@DGdpYk9Jtz1mJ&t?mrEMSp-D6^w%#xq@Mg^FW$Ok3wf
z#?)e^i8f3bgs1fk-$Z-gh7a?eJ!Ptqak^k<|2T_XR7hIJ>1h`oji-H8JVv$#4Y049
zUhp+Mvq!eN*uupda7BbEoP=ih(0PhKV7y~IlgG(okcrVY{OYm|x63wuhde2h*EI;Q
zGS!atU2eT!Tn6E=1NVX78*j~<_Uo_D3D<p7@`1fKZ%eZ`Z>HTF>+yN!^U`?D;Qggx
zZkVqVwl`ll{jZ<HUyb=Z82;0L^7n^^{);HS$}m=>fFkJ4@3;*LkEe(gdzskzNwhu^
zMuE4T)X%J(PTcY|Snh|Yuf$Eb`16&-4Bma~t23{p)BjeaKG_&kv$}MbeEHl2{lMP8
zpcJN?g%{Wr+8fO@OPEj=m0~jAiH1F;zTuy+5>VehI45*@^OD^o6FH8=)6vdXfh*nX
zGFzysZ-(%SqDcc8nzv>0i%u+oOf>(Xtb_2HK6vb183QE3m1F$7f?|&&Z@x`eO5Hi>
zCv3>Dyr_?>K6NW3x7;tX!|M{FbWWN_FrViRx)5LohkXX&O&H{JCiyD7JQm?(|A=u!
zfkkDLRXn8S1k%CS36KoMyytf_u;%DF{G3Edyqz><wE;xZL0x#r$W_IxhJz~Y$iQk{
zoH$$7Y`q;2LEzEB*Zg(Qgo8&I%lye`WTg=G$z%_r@5)&fV!GF=7VHc#het<keSHJw
z{8<KqzY0^O{F){W#d{F>jTCMq>Hs-^Z>`<?*86Sw-rY7iJJW74xe-1+v)aQ~@t5#X
z&;-&Yo!}vDo(pnKlm@-hpKo{fc@DQqFO_8uf(kSwP_7hn@Dt7!CX;ER3u6aTo5umL
z+6jZ3{Y$38`nf93_H`1V5;BuG!;p^cD4xprDRhZohs0mNFh&^G<s1A+UgDiw<wx5c
zm<?CN^*G{R{s|ia9D}w;*SMlqhF0N<{YHGVep#)=Bf}+VgjcOJtJQ$!8zqai0dCeV
zE8c*yxGX+dMOUplA#`<<jG-`0IbjH8hy=B8G@cWt^Ai|R81?1Z#rELt1NzHkM{}tG
z#4cqw;Y8%=`+S_rJ*BZs7H?p>uoa{WS^@k`|M#uCNH<6qcVOEa(-uD}W|vqpa(ofC
zx=1mL0(K#I_qO`dXzPD9@yQSVUD-4l$3pWux6~s60}sQOfTi*``BIvhPv<tsM>h@M
zUD#NOn>d=hI(TU9YZ3?F1rI;t)RiHx&L?e9zHzy(PnT}I36(OJn?Kr8U;HXY7eAgn
zDPxy5ucCX`PJRVHOuL-Q@45*)w=QsRk`8Uw!O$;zbS2maEsJGzyfulgB7<^t!C!m#
zCf}W7uq6rfj=uTEnvw(K_?A<Q!D3JN^6&jLT@l`J<nbHN;CQ~vvBGzee#@;Av96ll
zrM@t4TGIYq*e<Vu>0sKYsHu=e3;jsiT^W3(u0*Q0VWh2zR3OJBbi$N>xzumignXG|
z@8i2{T~>o~b)PYVw$<eA4^QB(p{@j|FcmxXsDD1?A+JpzOY<-O<$rzi^RLzX_f7Ft
z)<Ck_C1ZlCbW2&Wm1AnonUn)x#tbK*S;fq_sc^#a!*Qt=iKz6sopCzC6@^};X-oC>
zlTJA9I6j%D<F(^kuL$Z(^xMYJ^_YfZtRZ{;LV=w|p6ilvP+p)n+2PN)9f6&8;uu$f
zYYFElALXzv#p|^QFq}Z=UJG|O%}v8k&XPAgDHC{CX*=G!J9>U$s;w>0w556PZSeRe
z6;NjpB6zoitK;Q+vU=IFd4d5j5-yakUGlo%Ti)>O<e<$?+-VC_<85|wy#4Iwzsfyn
zu7v%a?|ql@kG3P)V4rl(o#<09JEw{r1wabdtw-+Snf8+Zspd|Vp_iB0Ew#C|N4c-y
zk*9K;phrR3ez`A$GK5!Sa&!BRe3JH|HV)jpJX|4$Jka>)tu_IlHwB-eQ1uWV_QjC!
zCoH^m6l!BCeCp08@#w%y&MEKsg#MH9bJtGp?I92Op&T_x`<RAL#(Zfq_3>l`k4b6~
zeo+Q_c<QduEhjVZcR^cb^)~RB#y>(Ncx~^aP%vY==Ti3(c|-C)J8v61hdH*;-Cp9o
zyqCKq!p~#BQWyncGtM+^5<w&B%C9MGlXsg{DK<SB{N3K#CLj1>VC%*<1-Si5017_c
z!8SKN)x76RW#-2A7GU#Ou7&WdFDS$A;Zc-p9`_ezS?Y0_Rp3W^d+qq}fH}lTaNRZ&
zCtdCDt_kmpa&AM>DUVQrFCyLWrPQB_rF<`e)8WCcf!aKAW~RK~)xOkP9UED7R#2z_
z@|<#b0*3;81x!b@#qNQ}gt^NRxRPgf^!P_#RDkRK8}5b_A8!LUcOv9ot*)=y_tWPn
zL*`$!we1)BqVj|^ozt1Nc&K1dVVsIY=NGPecURgJ&n$g(-lI8F|J%ve-9VBZdEQq}
z-f0gNq8VTLzoXKPGD-bxZx_@>NlMvORC!`VuT&Ja_Ty*I*V`wbJxae*$#Z&q9JzpV
zkg<u0HZ?O#y`U}9GTLQoaymyrI*%}nbj~`-V+${ZaUMDK@X_P&|MS@i=6x?J#%@Q!
z_3Y_-a}k?IA-XDFL22&B0#;FIfaB-z`aUXStzUKo2y<Zv{&+mC$_(cQ!ofL>JD+O4
z0|fAw)@8*#_)#=jdQ9DDC(~7cYTQ!Lc;C+i@RDx7%!1Ct{eviS9T#%1ALX?T-Qj5a
zc^s$rky&;X8qQA?mWrd<6+u{)Huf=TwKz=b&`bLMk#`LqCcK4EsW>s=BG(eWdI#Fl
z;!J4M)6ah0)}KEK{KdKZ58lhuADO5ULM5SsUh}?GNpopgPTI-gHt@P^v(vNmhv6vc
z_UU({<J0Z#+WqX3%}{oGB+9L6=1lH5+}L>D_I9>N<2?7Z-FxsZFqk4gmF-nmX^F6P
ztWnwKJvX`q{*=}7+U6>h?DN9Iyj1o{$4py1%8~iv4sBsS$!>@U|D6k}Ocgeh(0}^^
zu*6MzcYb<I-90MqAmd)>E@{|)>4zQ9XMgDxw$iMOS-9I5B2>3cJLQ`5VR6NJM(GD`
zi)Z4Aam>Trx-YL>@Nq_2hT77?Y`eF%)Vv?h1p~8l3#qGgN2QlzM;44`(S*n2Ro&z%
z+y(jS)m8J(KzECZ!_rK5^(mZpG3P54+~nm#P-(C`cpSUEo6V;4c)sTw+wIx<7IgMJ
ze2q$oz=IIdCF#)uv}AdC9ye(6-adJ5W-M_5ZYc%MGTtvQ&PO@mahUdrd-v`^H|Ill
zcZj!6Ja@%Q({n!+>AT343W4j;Lttp#a?~qzwJ(Vqmshnw#xf~47y&NQsF6GFD!QGc
zQWuw(fY)T`unS)L1qb#s$2?DGaED!FQNZrX1-$ZRq-*<@5=*7=Jbi8yygUs(+CSRQ
z*dwoHdjrPMgOo?yWGtAOnhCv{WW1l49LxPq9%BjlWt_Xi_))S36&L&T-+lN1ce#1A
zt@t9-e+PP`f?6ef?$tUuZddYUlsAh{$WJ~pJRa~Ars9V5%{~*jQP1;>)9|(K1dI|?
zr3!JCf8_W|!it+7Q+@aDT9k*z%YBa0Q_7=)*|u@Fm-ysvzM^rBHc@c|Olc$gOOGQt
zx|DYo@=#9~I6Ds-8F3_h0quU;9zFUrJ2rpt{qMCu`lElp%`Yy7Mmo1KJ?WA3(0=0)
zn#w)5wszXnryCh#FW?u2w~w&54Z@d%W<t+_jpL2St@d&g!&oM9DtXa@;*(>!{7vXQ
zxCih9WBYCtUyl9MN1nLGVfq{J^Eg89-7=muLwrz)FF$w$j6Evz5V|6dGDLeV=AP0u
z_{*u#ahjJhI<`vBRfO9UU%|^lkEP4h-*RNPIALlbn0#8@$>!MOB0TvD>5A=ce+@1W
zF1&|)n>z<d*(1A0;A3aU>BsQXV_C0Jc?v^o5Sptk-(?1V1G_Afw7KlR9rkrMjo-b?
z=lStNF`&tm|EBreus6)jbZ?rz9&egA<Nuby`%A;zFkdI^+sqfG@wWPX6Xx?^sN(DP
zAN<Wf_zAwW7Xz84^T&c(k+KfH`zJ~D--MUJy#CXxTzYpC(#<&i(}_0AZaTV?mac}^
z&F}g$jaZU*rJK((SSrOnJdcCDc&**Ev}n(t_*r?!x$`q^Xm&nRH<<>@q=#_Pz57d<
z<Di|i$gg)#y1wILhpd?O5VW#Rn%Chv|IRHTEq@2MszZgVf!xn}_r@RsK)6aK17W^X
zbm3vU@Wv@7v58b=uJ5GTmCtHh`MC9A`ZDt$V}4|2o_*cxtFm-+$qN!XXJB>$tB^~g
zDnsT#7~;p^a|YA=g2}Xuka4Sy5IqQGzpBm2fvX`L@Lg3SvoE72LoI_9qH38eZ&ywS
z1X|SuLsTtmR{P=a#M?=WgFz{|EW7cTR54J<C`K6)2GR)+Fk$d;KvY4ZASQwzo_g1v
z5Lq3EpQ|gp0DYZFmCRZy_E%sm<<~X=YWIHA=x*IR|B=b@_TYo}+S-HrZEAKFzit(^
z@Jn!isdENACw#8V9AiLpP;(WJOtga=umlMlU}W5dlT5PAyo0XEL@-4<Q5eD;);Y<9
zXRx!Mo2Pk{vXb<Hz0Da#8~#!HfDe6G1s9f8Wu&W_a<qqS9K2%SC6?)g`GN_vjDuB2
z9d3qMhncedonU&`Sr}QEGs8!S#yfWf;4bBnnat#kvS-B*OhP8vFc>nHy&$jzE3Uv(
zph$Yc#nO|W6A(>>p_<^Q^%S=POCDk9xA-Sqvcu7`l9u%s=S&apf-UvbZ_bP1X(M5l
zUXnEZ(sY(CkNoAk^RgpjN!w!;)_`g7t#AGK(o$&~U~sHq@1A}s#44@{Ro$^08++V4
zdji(i;I~ZgVYzg#-!3u}gYrwBI{3qPSBt^}%aXs&V5wr1OAm|AvHvgklv{iGS+AJ*
zX(1mh=GIvg*U4MeD_<2R_VN#4vGvL6U%^w7c<0-%xYv`9&LK8_qV(bs7dqea_n}Se
z&$cRCyt}_XTved^?)@lSpluYrV(-HAj~<D7?U%qj89R02%QA`$!Y|`!Dt_(n&&@Q;
zt~=qy%WFV&{`6tq3Dkv62!1NxE?gHwMDg9jstootgO#uL$e+Oz)-XDEu&cy)3}u)u
z4j=j_m87N7z1$3)zb?MCr*NUdWbE(!^H5j+{1~|9hhUi0+pIAiRI&8%$gdG9Z=Lw5
zcV|zyviI$7U94~KKj_c(pX~CjZ``lnd2p}%v1)FF@y`UV<MsaaiW&HsPJJgl=Ffim
z-|wRi|H_SE>7L@ieDlG$;!$y@@K@eH@nmDKZR{SmgHw-E@y-q&#x4b}C1i8rKN=yY
z3ygBF<*++TY=Ij4e3Eud<yGy-mt#}Lo)T*HczpfbG_)(_^N~j|Z+Bj+c+zfKSB!z)
zO<>s)gM9Q0zek_wm0t1<_Ms{#)$VBtZ}|oVi<8rnS^2I&%a!BP(<S^?=%1stD5|UT
zy;oswdZf)y54VMx(Kg8p+ZDS<M;`g+>ek!M`$=Y)fREm~Y*Vlga~_{Iiu=T98-*8=
z$503pz5t$Kc4nfDQ8!nRfBebQcJ^}At@X_d@4bY6YwsX!l%s*DcaDNNIg5}(UN!e7
zP;SdkkO&r@Jlk%cJ$l}@wsxa{bM?FJ=^ZC~yZhKrYn43VJqi#|HIM33(3!g=;0bQK
zk{$S~8dT|a1=D^Gs4rgF&XncsxLq6{kjASv0UXAt+lWFiVBij-9Fqtv^m8$ji`u;R
z>Ue**UBc@b-msHIx&vQyaeCB_5AcIGQut=uE0j?vo&{`#^O!)JQLfE7f;$a#Q^BNg
zB@2pxi^4PK0t)18Zxyr`OyW*mQ$*QKPobH0l~*evyo;>rxJm`h3X~Q0XdfN9RYKe8
z@#ft$Q&TE8NSBh#f)m2u-TNx2NFpoF$A@q69w8oSQ)YM1je%2M+^=#+VXepfee&o@
zd$zHYJ1X4arch~neWUGfZvzVj%$MQe6mE&P3ao4H;k>1UNA{@*k`#2xSMo%$N_qBo
zw%g|Vvv#!Wj>Xd`Ok5GJ@|9j4cpM*}XP1L3^;H07Oni0KZoN2fGZS?T;4tvCAK4j%
zb5_T90<0*xc<lQgfi-0N(Qby9AB{5Y%pEI-?eJi~d54jTWZnCnO1Amb<W%msnO&G^
zIhJ!=g$lSm<^56af|zN`i*xPn@_bw6JvTi;+ue$?U}tN)Z9IPtp6#|np1tjLaBZud
zAMP<RP-$JtJ%YE};KvZK@)-fg0BCq>Pj&E^vJq&TtNs;ad+egC*B?K5)}B0m(suX2
z;}c-(ZaDfw?TjtFdO_bXp7+y~lCoCIdz`1V_Yl}005`zbCMKub%-mcA;R@GXn76mP
z&Ajk=^UkU$K6!EnSMCGskA{VI2shgu+9!;K)BM6h?hu<_tQ}XYYs>BK-4*PMQIJkg
zsg!fq#aQq@LQnE`p5v~E#ijYSvb>aCuF|>@+EIVoWrV)sT+5@+o;_P{JDXe0qg!)|
zhFIB=?+A>A8Fkp)SjWB_`l#a7`BCAo^F!uh^iu`?C&&BkU~jwa?Qeq{d(4#&GZ!Ra
z37ers!fJAQCQ4<JNd1q2iTD3_tXmsWkue&4e)+QQ0#Y&HF^#>@-?(ErhI6`}$(gpW
zwAxnhzC)dtnXgVz7REnt$F}gUHSZAn>^C2S15Z0RtHigBsSDpK0v<f@NW}Rla<b@^
zHnHtX!ApI4pF-O%FVDck66$ByD`k@QC`_0BM8QN`l)|zUN(%HN$Kcz$cI*enQT~u|
zjdW}y=b+taoAKG*Lhg!l4w$k4%LsG{6@_fRrHi_O{^F5#$5m*mcTzj&R`FyX&4N+z
zx^HL4E$I(GgZXrLWzw+UE`Wn7o;<d0&o(-yJs`d+gQRCQ#?~EZj@#~F5wAVg*WG{a
zcu}FDvib$>IXO*z7Qka@0VK6ef(BQohi!s!^__cbw2M0qo$J43{5zq)GL{^>sMM2B
z7$2bb0?1=m9XrK|&^X%1v0Wj(IAxutCdcRp&=%n5&Tr*Q)6A16re>+*ZSd+M#}uw_
zg0Ij^`+@YsKJJmJ9tmfADdfs5A}|9^;CtV1Z6|TtHt3J^G?<U=XnV*fIA&XB@3q=L
zR2c()x=7;SsCIg}Fan%aq4J9Gwa}wG29-8awh>%OUe&%TJ>+W~M{@rh_+y@t3-A;l
zRXmd()voeYF5;~+D_^jk-6=jbC0{g=9ZqvR-W};8bdOG+Uzl!-&^Yfblm56HxQzK1
z{@4KKD_wAbgV_o*ziki4kH~rqi!p!el6uh|;=lAN_mM*1s_wOeS~}oZrCdcNYj$!I
zrdRYTpOGh$ZrMk~6T`U6O4*(~YWNPm<vZjNRDL?<T$~f{Xt$jl@*M26%hQ8)u(Q#&
zH=l+UeD6EoZSTDMZWba(CWD)&9Gst?qz>{K<HJLYVdELcre}&aE7wf_=aEJmR|Ti1
z0f`fRyDB$w0n_=JM@!0kP*>Vj+!J@D0pg-OohScf@pnfzRXWyDf-VFvIE#BKmE^m!
zlU|`WuKL?I>)1qj3a@Y}<&<Mxc7-#Bixbndzb6N%6rG=wFQOmQCr0kru0y#tq!zmr
z{&^3s<xrWX!o99XvCk;;a;$LC2DHJxLtfG{>miPckG5;%XCAdpx*j*~DGYT_t=LjV
zyK#E^wdr`BICgVbOnwL9Uw><E+BIM3SE<E(op>?0z3KY7;V^G=9~iv9G|Ub2b;7>Q
zd{G+u{pEA`i<!?W*MIx3{+oYx`}cqU_oMi7GU3F>U?NB%2_Jd;jnzLLlU9Y*l$N*L
zo#TdpmRBdNs2^7pSM>7L?w7Gcc{}&s%`fxSQot+sO0!bw)2lE22-Ev@zp)uE*5376
zfK!5*vX`~sVde>)Z@nrVo$^x!FDo9E#C4edyB`$ej!9Zj{R%5=u~!Ch)1?00ca<}?
za_hro<S;Bb5CLoANkp_IXNZpB`}pzFZgk(&vC3C&CX)OtrDf8W@4&+_q-U54+~>!y
z7X5XnFuqJzcKMR;6PT+Dh^~W`%CDNVLEM;tc)yq1>R&(*omnIaa+K`H2BB5>qCmqz
zWm1NTLBmx6G8Hn8uEI5}^(iF_<1Bdyn-F;t?FBoQ-!f+~Ait~{<`Hl*RMt%5ZXz-;
z5h4-a0n|JdLOA(#&`?1mL3X#7X}W@v@hXFoOv)Y%i;n}wS25=8^H*m21bqbO=NaK;
zV<+RGHnlL{-uvjIwy?aMm7K!#1(Rc=3)>PNs1!6w83qLk#Z)Qe@dyjI!aL!=qJ0(m
zx=L3jRAr&7@^qIGqLIUBXO%1X;c8afC6he?2~5MZTL$77uhR7Ro$&}3jO&UmajLJZ
z6I+F4W)S8HX9rUKWNK`yC`IuIUeW$x9K-{{*q`fO0GUgVvFI7umPtmPE$-y21aq#!
z5k>+btMyKfnFM)+KrwnfQzy(ufnC2E%m^yVoD5lCk3*0$h_4z!6agQfmO1^0cTKoF
zT0sR`t%ea-U-*dQ)&YO%f=3_BGNJG|#!wy7;+G%cYEZ*fACIEnxTYxV4VQ;+UrD-L
z{fUoWy5%W*e3Dj|HhC0X=C2~eo&xrEW5te1g?SVAI^BdQKhyP3Yww;S^3%nsFnZ<J
z(D@$FJpXWbEsZ>8j9(>HFzem?>%vwZ*nG#k^E3Z0Ts;E|uH@|7#4D6ZT-$AnxM}A<
zVe1`-JoSn@*2=oC(Y0sdcIEcpx+DZno$KHyO`^>(iC5t&zHj~rWE<+ymnZ0BfV{hS
zNu*q~C#>JyBW|ZGapq4vi5qY8OkE6Hfy^UzUwOJ;z2n%GBj4D{yUO3$yHv`KdtZ+-
zrs%iUu55lAza06qN5ms;$&06lx^5NT_+7fD)wv}*`di+(WwMz_U7WJSBz#<qhok?z
z{j2}39ZZ{74Cz(3pr#3{oBowg;hUxH;`U|9lK`E+rs5~>MCUC~Vl&gSRoQgO*PxvY
zc{8ofk}Ucsj;7sxhu(E|{>yY0Un#eEQck>w+MoXP&%Zgv*Z<`2{iMpBmKu~JpYfo4
zH`^no++)L9Uatj+=~Mh|pKB8+cL6Fqh74_9A#L7$0$L-S#Ub~0G0f~8p0%yr<F>Vb
zlFQZ=;=)*wrj2lm@yDGbciKpVa6IEBh*Mx<dMY0j5I6=$_$7D7c-0^AW1xN!i{qN_
z0tneKeh!pcJm#4<2oOHl@i~H!viCfq@lDIIw;%gl0cu>w=^P)%+noVX(v4Z(TAX5@
zrft2fk4GC#xe}ep`52St7Z<=`cLRR&tS!#ZG>=yEt_J-_?!eO!=FV-#;};54UbMr*
z)3&~S(Dn{gBsh-WYLlbzp2L(C-d2T=W4MAN^R>Jsz^L6N5o$92dUVb{ZE^7mKI8Ui
zgz5+IzNhdEb>E85?I;giB;<-_6?XC*9<}39UBbbgFCNRdf8d4kb^j8XM_?r%p;F9y
zlXiD^vXf20nfEccYt4Ja?vS^NkIPGV32-4tLjiJkjO@91>*NAHLcR#TMBc>}`2=v4
zKOgNj?cz#eciMO~v)uYi+{K9t()8#XWyYf;3e0ArU{HA~=7cZ2xQq~Mer`SjNBLz1
zL87qyi|J&=@`bzQ_L&=;wE0Q!lDtO2rI+xCCx^ISl9odC!@aBq_SmDIPmfYdq6BPI
zoJk+uB{2%<ZM&(lJGpahdK?^jMH`+Sx1p;u@_5;n=O)|S)C9bf9n8Kbe1w-f_~mik
zDRV|yhT1rpcsyj?eP$bsq~GoCAGgPkpSMpQK5oDM<g@mCbBCl#5bAMgDymcfUXe~6
z19chxSw7MFXV;DH)KSHjN)eA&I#Q{`Tg8$J1O@m9`#aeUV7eYb7C4fQWp&qpJKjdY
zKX-~`w*_rrdrVGDQ6Acz_60N%_RY{2=B6jJt5A66Qu`1v`c?WKeaJ<K<JE4?$%6Cr
zc?IYyG^3~kR>I49S8tb>6-O0#yVBfy_1sP1v7O!pq!7=e?H1;z;aNSh)79(HFzTSf
za1UH}C!&facjS2Vn7akNFYVUwIJi9$rROEM>-}l&VB4{exVsg%Gm!F%wmG|a)sE<k
z3Y<+>7%Ak-{V>2M#|6^QhDXQS_@si}#kO{LwXLqMwmHg}#dhg0)mE0#ek$%+E(O;L
z3SG^w5=Euo`uci%^5iM+=WR=6QIuZZWux#Ox)K2;bqXD|?a2URn$ot}+1a+bw$hfC
z7TVPGc;MqbZcB@^Q3|XrFXp~Kh0_Z4rzS_EkR1o%6~4M#!5tj^T~9gnK^owE1$c@R
z9`AQZd3+RrUI7<zDYTbzI03N#dTNKeF8Y%->>r`K4g^L3IO4R*4;3r+mmF2w%}u0<
zuH^R`0~I_ju#irKR?z0bALFRhRq-m`AF3#G_nS1i?#}kggv_;0jsi243z;X<-^Ks*
zDd%~Z92@Jd9v2tT4<{#Q+sur|!A&?4MX{q|Nm+yQu)SUH3_6Ii)g2q}y?Z~f+~3=3
zd)u4sh(33AyicDyN*(Qc>3`I{#=uVL0!%DdDRk#^q@ncOT_@x8HQU2S1(8pbzjN+n
z>J^xd#f)JS_^VWa=6HgL^het3F4XDCso;&{MnCpOzCsx&LdM0+Z8NNfW5_rmtSqZ@
z=g@hdv{B)m<ELe{{oR$J@Y;MYpt-_Uh0)#Bl_(iJ($1q|RXC}{v907OTtwmm0`vC{
ztY^=+Ne5UnCWa>9?ukL0o14(<TT!6gU0DF1=Oau%h2E)nSmk>e+*e5u7?9u2{$cJC
zJai1DKZ{o?P^C!#1pK*(LPD6>SMB4L%i{u-QOK7}t84{+eO>)tq`#|}pJ0rVF4{MZ
z;|{Sc$JPCVjF-mq2w~^9;;XdJ-E{pi(aIr&ZxkT(VU<WO9B`ZuUqsz~SBn%MXS*2J
zcF1mf#}pb-Kk-wUgmf{xGl6?>p2CP}DiUmO+Y4eC8lnQsaU_dvlV9~^$06%n3RV|J
zSO((8UFB@aL+sAxUotW5gw<U>#j=P4?l4n<9)1hE$KmC8KAtQj9x3@sls&-IyHRJL
zXDaGF8gT+R3L77f*v%pc>L!n1o0w;4EwGgKdNgGGfoa0?l=4~miEt`;oeP&$G|gra
zi+7Ky*fNA;dhr9M`4T?w>hgrQ`0egL7am9p4|4yUigo!&cONb;&I42N>au<Nqi?s5
zcwAH<EmGO!F{R!&rgGh#;okFl#P{u2SAoaa_;C8eEcC!*hdoMLMV!0+#5Mbsi@?oO
zxnlSY#uWJv+e#YY7_hp!%(Il^jNP3cnJ8@{4VG?NhbZIHzoCUZ@<p~s77M{oiVNaS
zcn<JQ1)<78ce$yrRL&(2AbgdFsiafM<__!y^0n+P^14Ocq=D|TbMaMn)x%5D3W1x8
zwLDr}oK%LC1rg*|$}QqsZrUtL;u=K;eIYyqG`~M~(Q>Nzni;3RlDA<UQ^eEQY}oFr
z>}9l@S8sa1{KoI~u)VoyPq@yidCku<`0f8qXH0!xJH8(8_3z$x?ezArpKI`bt(0FD
z{L9k&3g(N__^rBj?{C6<Ua@k=*FXHv{usQG;N``Cfvgt{IseoJKH&`$S3_WRpcMg&
zn8f!+guziZddArw90(IzJdG2ldWSZ_AjRH?PcTduFIFdGmaih}-g)>_X?OX=!(jQk
z_BTl4^yMqwj7Pt!yPq-)2^I&_;;lo{?ciZvl)YF$%hSOKLg2fyz}9^`8w+C4nXQ8i
zm}%STWBg2B*u&4@XLvsRO=@~4LRVI-iQe1!=~&V@RJ?9*)aVe;Yh#|)MbDJfdevtI
z&5x5atS;ziMO5`y*l#^v-hS0ia6e!$*=G>BfH-=VxP!M0*C~_t!`=P1{`hIzeDtK9
zA46=JAXvvf3l&R~FoFw<3z_s!GgvtAWCaxyH_wyJ>Iu?SfNH+3rm`Rsd>KOrX9X2b
zz$BV7A;LXJK2RxFWjT-+!&-_G(>IMANdob;%!V<%MB710B5XJ(3o^(WnX1zh34g61
ze)8l=`{XyDW<Zbu>g)DpMo@pXkY8p<$;9f1;)6ugA#F<;V}yrDOGAz_$ld?&gDAeH
z7~DKlz+HPz9xf@rjGE^pyUJReX6Opz=Rj%})_<V5;*lxyNDN_k#(;VV^C&|ggR5Y|
zy8=At+(}mySKw0GjRAHHhAK><h?gRQ!<1WQPOOYF6u&d_bH|vBj(sVEDubma%n*MU
zw2{Xe%9uI0y0S!O!n+Kdfb`qVWr~sx`C8Z1(7sANdWJJ)0(Ny_IB>9T49f2Cl4((>
zZQIqPPlf^-#disxWd*2%JqBjnqYh*!Wl9vXy#)U~9>}?i!W!|&FdCTv%a^u*xwiSl
zNtrvq!?zO|07yPg@`hYd!!tcP-QIujUYnnpr!BzYm-dn#yNu~mWy`yJy@snHEIMD)
zk3OL7GoU~EThBz*A2<AjZD@)y`03NAScJhZ;L&ax@#9arbHhb@pROipB|lrLbFH9*
zkfxm)6QHX`@1bjsI1=Wjk$m{B_?6-ye11&Tm*Jp)Sd)F+x{tJWAH_5~(^;xc`s0Ts
zxnfC3%VL-cSfO}VT742-c+!r$jw#O#|JdUj#|UTm)+PBEKBhwDbC4e&)NN3vxSF4~
z4v3n>$$zZ<cW;rYwTuJ*?)OHFGKqs>uutA8N9V|QQVU%5i)%$UZTy8cT`X?e3vVkd
z=#LExZfWP(@zw-L{Rc*`4yHN}vP^yHR~7p;qffaTuii1nzy88>Ap?)(P;R{XP^l?a
zTS{wi5=;3T$1%TfOU%>7N0a&+&N}(-+Ni2nWh)ryeO-cU)8|oX*O92E6Q?BCrxW}Y
z_YLxbAN{bJSL!E<@RvNb@v1henxu^k-}}c_?{dekr5caN4-IeJs$bq}|KeZ%-}U`Q
z%>Vj-_g_&w%UERzUuto%J1$i{^wXpKQbT}Z-xDrry8x|B>ib#^bh$QzYR8az^4B;H
z0)ObeLJFyyF`d-OV>o7bj7oNLF~)h9m&z{hipUC25(&>~9LGDJtnzg{!`bmFLNUTd
z5XLwrf(iEsa&XI7!`lLxmgBH)pJ%c_iLu^s)eOWV%OAm96(;#<GEVT$9U_I)sb}$D
z3b@IObQEluremM{o<a@dDs;$7Z<3Rb^WN92h8?}r&W;Y+#*>HbaBmx4^O!If?fXCY
zPWvDH`+t;G=N{GLZY&iHSp+jS(M~QL+ug--$@?OF!Su{TTUwZD50<Cf)aXz<RpCIM
zFcVRFRGYNifpK_vLc7UF&_=*_|L`cg#ys}O$-GKLfv7Mif**Mv_@!Eq<*NFL@GQ2g
z#&K4Go%cj|w3sV&uWS?CEwhWf6l7T5@I3Gofj#bZf09C_spd-Ov^!-Y8{nx>tngK3
zocyY*mdA!Y`tVL0x&1P`sz!OwP7b$ecy0x<E=rSk+4l%0c&@SOxpvt`+VSaC>f7(I
zvQZTjx!db#f4dEX%ggZemPM-Ojzo{YQ3$Gl^%#Co;m^nre9`GKbmpRc_};seV>or*
z+Vswn{pLvu@{2C&`QU>OavYU9i$m&v4E}qJqP+SMZK{$`9!({kcSX!hPPS#*(eJ(O
zt#+`F|Jg}!W}dPsoOCC`$k;U0{}#BQ5=S8!=~7=6Ztg~!pZ1QW(RL-`QofUu7wzfB
zR(t$xlXC8aCw4)cE14CLDQHt^r7&L@;}E#px5RbP-rklzo>nhh_M<33_1&9>=;2G#
zp0u~Ts(r@pmHXDfjg{r)z{h?l94w1>%q%TTMH#5FcS?R3`)%;?>b7?ty#mk9+q12m
zw!N=1r^dnT0Hv)wCec0-o|OJ^LR+cikPj7ya{N2>s*2DHD3iNCJUVb`X`a02nn!ll
zj=728vBK&sP9vWj^(M?o;}XDJ0v`qJ&gE22NI*+rS_)McP#*EvcD9Ld6;#qE-JEx3
z@gj~2=ftH8U=cwnVXjCsa{~pNd@Ee`cr$TTrIB?|N#rq~9_wY_>suu|D5;yfS=?o3
zqu5H8b!U6KZLP1jqr(I0r$DmWT6h4+ylrFodCRJwji-=Y;jhP^&CX9}o_h=pdGXR6
zkS;hJ&JKt7?%#{DQkvloh$!8iOW4OLgZ5M4rsDm8dD!s*^yifMaKftiJEzf|JG*np
zod+k$#~obq&PA6MW<s;5vwdE-%mc+s%BpbPzJYV-SKXg1zJSy8Z()85yz&??+af@#
zpQ<-~M}fO^-^G*eoO<E>n=~AI6s)?_(RQ*8#TB2#<y*XVPH4Gwj{;pdnlJtZTWaID
zKwDXkl7YPBTg3$ZUM14>tXZ}_ciS8sW=vUGS!mz+_P0WRJ>u-)r=KwA-=j>I%qtla
z<`==|)zEQallyDH5Br!3OY`<r5$`B+5rFes6$|D)?6@clDXZgPjiUu~zs2X=3xu<~
zR7MG|l4xUNlRDgK@4xpTJNO*G9zJ@K9YXd;Wf$iPze&UTJC4*G+2NeFr0fdS-8D<*
z<fs8(R7eum<Ig-2HHuR5k!I}f?zK~BY3}+Wjyrt54_v%AOC{9P=bO;B<IKNRGTdET
zY~TIX{r1rZ52C0zKE7<9Jbc~`Nz48z44p@>EG<x;QszxFMt$_*yKP~1B6x7Px7(gS
z{Ve0--3K24!-?SX`qmz4ZMG-RcQSssxFow!XhUg&WjsArNh;l@U(*NNp&9<@5`0(Q
zL4R7DpUW|%;@tuL)3H;zjtU*+8Y&O=_x9TM?rs*?s$f%wp)BD3z14RA?ppeR<A?WY
zZh1W1(FygT4*+M6+!Vg1H8wg<Q{O5Y1D%xCxj066+3vNt)&;Gld~ye*aV$^xZs>%&
ztd*$<565t2Pdl`Wb(9|2_O|)$A(hMWck*YB@%Axs&fX-33A0jgnYZn$9K*hBeQbZr
zP~(Y<ix~51Q+N3+yVDPP=slgbqqJLDhxco_Yc28&WfA5@+tksTv<Ht1hol7*hVm8f
zXlX2aY`Dpf5QsJeexxI>V?8}e)YAaGi)@~Dldp3Dlzft7hP<k<v%mF9j^PpcKoy9$
zfw{X+uQE>9uiXLbf*E&pd2FY=gyTzi7sj^S?Fc{Qv5oRNu7f!_K4`!C`Jc9zr1d+$
z^CSBG8vY*rXq(f1;E_jgdPnJ@$3Q|yM{m8%F`z1uqF^WO@rh~h$uYBVRG8Qccfq6Q
z8#__1x|8zuE%K#4DgafO`FOOne5Q&tciRus&L@=Hqmy0yEPiDngGw{_fA1+%j-?EB
zl1I2Z=6RgxsrZ7QbO;nn+^Z}Oa2GRUvouG(h0=*9;*-1oJR;iN?ywEvA?4>ccMj=~
zTl8=8rR~hy@RNkKJ#ybF_#t25Jzwr5vmP$~PMgtQb?;iuW1^IIyy?m>z06%>6T|J>
z_ZM?~cyUSeD)300&REs8y~gj(l0MO!-aYS|_8afs-2B#l)Au#Mj3>I~H}>At?zU_5
zW*)t%{%~z>*5#&oTl_ao!hK25ubS4|%-4zkX8pct^OOJ0e=>CY5C7l~e=<N8q0>#!
zEt?EaXL;Skx%~1nMm~$l@>|6fP9Q>k_rQ-G_fl4M{`mFGRRyY0Ri-XZ7bXg^xClVG
z1;cl4UHtgQt?vYJ=HZ9%_3hgZQ+3s;>RIjyV;OHbOvP893fMoo=ww>u5W9{^T5IQC
ztCexQ;imK?O0>9_t3<!z^OW^9lh5wQ6<gZVFA_Me0*24-@w<X|>D}<t;f9Iun#3%6
zB6Q}atpqr_3@Vtf{F-kf69&73>p8^h*b6;hF|g|ok#k~CLSec+(%}$B;b4bJAcLZV
zZXzZf80!qUb)T1nzqdOOc%(sXO|FT(t2a!yRs=Z-EM-M8w}lJ2i-~e5K<Skny~3`m
zQQS<Yk7JF^Pv$R^QOa143_EFCWa_Cg4JUn0s8p!QjO^|0z}S?q=&PE7>0g1lnNaNF
zd(|z46*hib8O#z)+&dqB*v1*4MXwAh_zSbDJMb|Mb-B4KD-#VEy_`1;^I=`nmLP<J
zGL?i*Og$^!<99r>TjiI7nT*<%gBz31>{xR0L>nmR@+d&3Otz7Owkx1aH|g`|<i<&s
zS;ZY@R)*6yx6K@M%|k}lw@NR|YFO*!V4Ky(=0*LT3}r<Y;jEvRMa#t15ssd#Uc0=E
zFT*Y~Dfo1gx$9SvIU%(#xbje4WAK$c#7XnF{N`tUoUA4f(yLWf#rT?zVeD_#R=fF%
zUp1i=kDbVbvQQtw+fIfI{V+bbJ9Gyob*A9Q;9KqQW50OQ;I3C}^;aLht3M3FX!k!f
zteoOzVN6>9R1iJt8GQS`w$*=qR~G(QlN+IuHjl;bie7&FYN?Nb>HR9H3KRQlrr;VQ
zA?-nuuwCwP(?;0T&CiPEsAQ1(7+CUOl(Z5|I`I~U1(z7?y<frh+N*4R*n~<JRh)bk
z?6s$0e4C!%X@6bHbiw2ANBLFQG@sFpU$3$!yzc~#bLFT1wRz(cQ!ceY!!O@m+TE|l
z{+sWVSqH*oOw`x#UD*thzq0r3ann>0yD;>Dq>iIJLZ6P&(xVc%!XyHALnY(hiNC=P
zqb^KL=TLBoOH75Vyb`E?87NPESB10_lD?d`pPgI1`gHv0${efjf|{leli&WQ{$I;7
zCYJh2o9F9B-reuO^a<RwcJQcTcCm$-uef14{IFkK=W!h%_B@*U>c6YqyZ&N$wM)H;
z6}KC(`Ll0u$Jf92AN;+aRDG_E-_r7)&Zhj`ld!s5WSwm5w4`7C)4%j9)47)6!|yIh
z9`nM*55}8PzA+w=?v)eRb9gN#tcRzUZTH};?eOd#o@Gaa7f~y;H$}^=i7-sT2YgKg
zq&y1WT)3jp&J~tE3N!6e#!C5!uif#_@hl1y#~%Ip9=yB9ajcY89iPV2hcDhN5t4x^
zm(r^f0ODpC6<p!}#@#96%J&JzW6S+2t4Te+(6I7dCaeJ3`@1~WYhiw>O^n`Y+Z#{Y
zFaP}Cv}d1w49~emmRIea`)lokcOOI%_T<sSw!O6(<<rW_N*f)YYE0XK8@vwiwv67f
zH9tGm7H20StapdY&fam`Iyi0nCl~Dq_<L8^!7*@m*9Sap@u}kxnCYev)1#BZ=TLox
zPYN;=Fw9UlcWlhhyOU!fy98_-`6qXP9UL6Coo#n=)zL?~3#Yyvv@2*2Uh*A<7VYaQ
zc#oEnuae(VXn%Hk-n_qI{n_&z9p_yqTie^sW9{@)kZM^Ko+vOpyL{D-&*0PO<9A$e
z22U{zoJ~`~#5w$%3W4?Y%^c-+;W2l|q_Io7&mXl%!2Q?1`17{2xgLR;c;xP-jrHfi
z{?oR#{;X~DJb&^eiUI|!@=Xe}U2HWwH3i<kXwNCHN31y)Q1~@HJDdA=6v7f9CoJsj
z>;SLR@P8g_cy)f9{D1yuf6{iho@b}l%*1#EJL0;#?v9Snb1#_3ysd9|f6HcsFvXLB
zvv)i;j{ZV@6`D>?xQcu_N-+gJ_PyiN7j1k0C`aZg<dIqld)vkPK-{6F^5ezD8U43-
zz}%T9zK4$$|K($E%oVzRWZ>#k(v_km{wcsJ#YsK(FYm9CU$wmAf<kI}Mukc8i0&Y<
zFV-DDwYV@!QDE)ragVl2ABCt{huSgd(S@F@;2kk9$!DmotuD2-74UR&tlL#o3MYla
zIVv33$@liV6UKn~4E=I8$Fo)0gpFnPA*<k$)mPx!U+N#FE^xfvDcpuAvqy9qL40<n
zmO|&};NEt2fYi<{k9xGP?e6d6zLU0fSBg8jO2N|!dr3)Q`P{;M`rYEnVp~~V&Vow?
z%+6WXpFM34A3kbNpFYh!L>>W|lO-aYt(|Szc_F<5M;zZh&19dxx&tmLEO*zFN0kw%
zjf?^B8S!^6$_01Z+Gg_U-p%Gt8ShK-nVTL1QO??<ho7`(PoB2iDI~qSs2xqVm5*~Y
z>57W3+{Hvb;+A*acwFGd=1zP3h(7jgJqi`WN*|_X$ZwYZ3~Ut+d#sxa6=|H%76p)w
z3%S3FIZej1q^}>iD*%1`>6o@CqiA~-KK31z$fF|)s8v+LgAm`i!Fdu;aGE^mx9)Jf
zyf|e(s35)+GlWk1mZxajZTxB;YCRN?T35>#dP&>nP9+r#Q7+(yy}n-z=%jrocj?W{
zgPZQilAd^!<Rat9c;=wDypx79y7=L&J0AEY6J!@UxU)fUNRLz^s2I&|ILaUnxP!}M
zDBW$D_~51R$&tRmM_h5IM(P5bJ3yu{)G?Js|J~IjWIfu@yv#$K(zwG$TISqW+{kbx
zT@WW6Mk$B=-W?o{rDl;u7beJ$;cO?5O4Dw;lmbQNEBG-v&3Na%NGb_xk?IxjHh<en
zJQxRtwz>4eN9CVN9sO)`;cS-H(0=Ex<gKl}_W1ESbvpv*r*rzt&dx4yxg8qfi7t+f
z_Dz)oDlOtSe5*~l_<}l1qpWxC?53{PU6=`P`_&X<Xuqo{yNz&j$C2r{Ys}*WPr-F%
z3B8iVMHMQKX?S>zsqhkW9_2STlbt#$^;E{Im^U5UWS%~ht2p|3yKPJPRWdlg&vCCe
zWEJAP{VM&Zdkf#L!7|V9ce#k$GML%XM0(QdC~2Y7WEVI`>C9Nvaa0)Fmy+;re&>Wy
zKH+h3p;e|yTFS_z&7sfWyY#>|la@)>Y{!GcLwF<TCy%^`xafE<ui%0b`JC`)w1*t2
z_dyLaOxi}4(_Qr*jp#jF-gBrD&^D4k5hp!fxK0fMcWgxRx6D*F>3XM>&m`lF<DKwP
zj;J!uzThGp#~P!E!{UqbC6Dn1`i!-}_v#{wu6f$gomuX@b!XK1`3ZEH@%R?)&s#XE
zNK#qmEXL!ST^zCT{Av3c{KGGP`ls1(=}xZ)58ee%L+wa@3K%$Md5_qGwY40}43MGU
zBY}at)Ewj7q3^E$BlyMbJ?|pjZ@Y|Hd+=M?<rkcRr;>()H!7hgfth2pW0Cz>#nIuR
zi>Mfn=@;@O#b?&yrul{0?8sUIFV~hA+G-Y=@a^uf85No(%ary+`6<pariVvoY}E*d
z+`Yvkub^_#UEtp3s`AyN997QkQ4V>Rd3f;U`AOhj3olfv`(%NL3Ov%Vjm#r-g2s{q
z3az3^D{pYZFog9H4@Kkj7ydGrp#C0lI5Rn#g;@p>$$e`zPVbiA`T&<eXur@_f8F=B
zJ9ZH#@ou=i#U5kWyvjXq-EM??GfsWK-M=@VAMdx#+uYwY3HQ~4e&saZG+!s)Uo?-u
zD&vl?+kgA_|L`YSvEKnF&Dq%{-@Hne5vT4?$HehfZu#%reL{SBgHCZ8D8jm5{lVW#
z+dy&Ww-TwpSp6xaOd9+|9@0A}To*R!bWtkj1k*&Al+13{E58-qG$3G>CGhCN#+Vd^
zO)ww*eY(H;_lMv5B#kOeUmxo$!AdaFjy?XBd{QbWOob!R4gdN{nfXs0D$xqs{h5*O
z$;0rv8igMpqa?PGFoU%H?c97_yLZO=RAk%@t8Z-e)+u)0UEO5j&KPViAh0hatPBd4
zUm}*G^IR$nU@~&Bv)5i;I3Xwxvy%wcQ4tJmFqE}o#?`ySVPM8%cjc5y6w1J(B0?ob
zw)`sCfx!1noV8K_BjJo-KzINw8NK9dp2UlSO9qNRp&Lb-L>PY+95Uz$K%A`Fg+Y-m
zmx+?VOYH0DL$}Dou);LRSXmB@S^alnC{GHNij>o!;e-eETzl``_Tatu+6aT3jD_&%
z&ms20QU^-m5k}dEZzt)lEDIy<?k(O~i3`(|6-_)>*d3sAc0B{3ei<0~wrzasT>Ead
zwn$9-DO9Pg^0YJUn#nux)?~s!7>#*^nZw_z*~SVdWR`Q>0cEtRR$t=;&pJ7&5f6$1
z@Tdc~%$L&36?0CqWTZ5u@T$p(tJj2w6Dn5}dttV4keTf#4yIubv5c1IT0j_p2X9;z
znRJ62uBdd<%4C%CdC9xQ4MUvA<`U0Iqf4vsyzsaBc$YpA&6k#HBlHRHRG6@CPCyKo
z2nptkepzpUB+W=y=O^?d&K01@n)&3j?EV_hJ}4gPslET>JM>rg?mPY+jf#JKG@V=Q
z-CKX%%-eRgj(s_N#SJ@dm6z`ISN{c*giS)&`!L4UnLlCRAk&BQuRglh4S$o-A@=-s
zZi#1IJDQ>s->x820ouF85Vqi7?+CK-_q&1>zWGsU#}T`V`J$n^^=*{639m(m@~=-F
zy0GTcxuni1Gk*r`;8DVbYxjboVdGuK@Ck9#SsNL|ksmVa)5y3QzplOsWVi}m<8ejP
zA^x4cypmGVD)+8syKkK;4MU|Ad6$31_1`*&htTMR!}VuZ&P3rSZk_4j5f~YOR~I){
zgLJkE;qSHSf|E>z)WO(q!X^QI#jU<|HtoKgwtb)eH5`RUUkcOJP3J*~n55d(CI9#*
zME&zrKCzQX509j(-9$P(@k?AWK&7F37n}b+=W?mPxF#J-TP}Tg>4b0Hd;A_49B!8D
zn>_09@BRmW?<dK(K4V@|R?CyIL!oovOv!;wCg-W6aR_Bc&SOk}>Ph$;_AWq~F2m}h
zCg%tLJdV2xb`$`6w492qy%W#SziP*4FWSi&WAc@|NEBWuP_fCnwjvS7Dc_El8G9Lb
zvl`a1jCby)%-96cBC@fpU5_h{Esmk?B+0mDx?O!WmbGX|wRMCTn3RFCS6ze^{^BC(
z)`Zv;s=UTAMaQ?S3MDU(z;UOD_gy`E@-V`TwYAkKzbI_tdMt|*^zrdwU_{<mr)_)v
zN&EO0f8GvvHX~4Wd{u#V@BY2C-^Ti8+uYbn8Ri$3b2p54MJY^kBJNR&3T+l=r`po|
z6q9_9*B*lBxoEpbm+k0M@yhM&kU2c@c)3fUTl}l{Q@Hweb7MO@!@L8^T^cIbystuq
z&C1GRySug=p_;-Y@3XRfT;;4H$z#$i<M!58+u!v_9aq)YN?qH~y2}^ISIJM5@Jb#)
z#fwU{$u{MVklS@Mbc9;8@xks+lzj(#3bX2nGml>LC_?jhH;+3pZjVf~SMU*+jDJ^c
zxV=*80572+^Nc(d1fPO;u6BR;@UynF@w~k}J8YN7d+pi7-?UFZ{#o08`lwx<9%Xg^
zz17v=md68a5ccx?q@5k@NBOwFvy<Ie_wU`Ot>)m<RGLgg@FHGBnQ?rYqxM$r-fQ#o
zi_IhHte3+4=j$85YYSdf9{d%26!=PCJ>K_@p&f8yq%AKkM<A{6`ZUMh(QkOaXz+T}
zYn1k1yZ@j)c=x?FGdmwV5WT!J<A^fXG1=AL>pO?-G5Kxn9#Wr^tnxm`-}#S<SXY-H
zdLI?>&d*M=gFpe9k0$(`7Q2MqEwhQv!G|_<E|WeJx8nW$PT!*Mct>0MoBas8!m}Kc
z)%7(~%+3e+bOpQN(@&0prFVdof=*nr@0*u*l(`@a1V|tCC`{)AuKX5v6-Lj_PPXOw
znKolTy93yOv#akFM%y2~2WCzoaq9Vs_8V)<;`!Yb+I50)S0yd<2m0hnb&ulQ+}vo}
z&Ic8y(q|M%o}Jb%E(NYCt~Lncy~*|kcm65BeE!TM1lMz{XI88Of9E78;Bp@7oI9u@
zRP_i{>OmcItlTVh;IThSJ6x6T?k?$q$`jKSryoChlwCF|09Eith)vCe&7iGl8|Qo;
z&$b61ZxYu#$xbML-%rKi%xl2kiRsxk2@Q}AsKlL}p2;eHsfj{vl|SxKJv%vQThOrm
z-CbzP1m#_XR?N4h<z-+BJfR^f&J_UL$A?G8%!K;d?(SM8EpRnQy)jNm{@qD7GdB}u
zpKar4EL~Bc>Re1ZFO9cWSzxGA#=?V5nR|c_y~%t<MFsiR&LYM`6)Dh(kr7X2QCM2@
z4>Naci5$Rm7{j9UB!745)O~m0q&sEEtDidz<~_8Zuy#x^P8tE@{L3fvAxxAA8pr9b
z-0rp{De9v@*D{C?_79K0ylO|%7W*qOke<5WF*_%r?ZVNuBkr0RrY=Kwcy8CO5Er<L
zW8Qma8@a&KIe$)o5$~w8aB&P%;dFjh3NUfpIj^V22z&dbcxBkYhCb!E<2b0F^IYei
zi}Um8i`G}z<!DRV-#FgQL(PbLCp*k+zf>PMI2J^nGUnpp5<&|<ph-I54dHz|4i|6W
z!V2Pzje(mgi|P~{=eq$tcySCJIRah^{(~oEl?AH&sc2LIZU2_Wc>lI;&Ra(*!`Rpq
z{+H>u9+$YdxD<GRxDmW-caHAef_I?tmeivyj~ExFqryO?pYot6RKbnxB*!m`RLX8!
zEAO$b^XO;2XkTG1UGT_J?>VJ$!7;`;Xp##TROsAdvQQWzOotdhZU@$LvvcXUD(K`#
zJUTApx{E<*2Whek5ab;^k}v!ra23Ax6Bp;GbP(3Uv3A}`?hDsTwn6!oF#w0qvlH;@
znEojYY1q(VkHn<@(za5r+4`lhj?xxbt8_H|moGb63*%CDA#3LmkGz*Fd;{f4KjR6F
zm{DotU5wR##if$L$d`?flYK*lTnip9E<lTjXFnEybiaMe`|tPL8RK>>jsRzYl}F~u
ztIfiT$@c(=lwaA9?P2@7Xqgcx3jo~tD$N!jv)H4197A0M-~yJ~3177Tls+XLan7aO
zNhPLv$tzecl~F2+CV=A>e9)i#$-iws|Jl#mXP?#rgDu9It&I)FmnUuG>0@a6R?7W@
zAN-*G=!d`CR@d%<6C?BopkZ6nx7-0ZMP1z4K15xXmgn2N;|nlx0SFBO&esAp7fGE$
zlT{*Gc8^!}q$ri_!nqc9yF;$*j&0H>mBTwb9(T;R2A$DeT-OB8?DW9w4)s>?HU<7!
z$I(&g8T?Skd)US*9o_jTPPt>dr(5=a`>?x#-3h5&&-U>YB^MKg)<6(9Hh06@4gu%8
zOOtJNX@+j<?mB6Ky{1JrYClY|q{qI9?-Tk<Wu-IWL##Kp@I`)Tc$;5#I727gu|GLl
zNB9<BRAYpue`|irp8t9e+&a_S-WFf;wr}5W`e}YUe12ose6={g9Y*su_cu+#eYK!p
zIn6iC*NOKR&Ev1i<fy}c^vC}&Y@3*<1)Pj<LsKZLy?<*HHEm?S_yHk%!AOF2LG!6y
z9>jNIRoMFC!987|bBklgt7z||mxso%_3DjQQ;wZiU*@WZj`^$**d^+|bxxozY<bqJ
zFJt*x=I#^EfkNy?Jv4niD#<|v=!6{-i%;^oVfwIrIj$Y+Uq3#Rwypta*TLh&PZBr8
zb)ECpJEmeAstn<}k1p1A;JA0+H>I>X;pSJulR)D3j$|5`6(B1N1oD`{Xa9nU;47GJ
z2-Hg&(?Kn(JRo3D_qM&U)qeA{U$p1H{;a)V5a^9b-G;$%m5xlkgHwO5?*vSTXDerd
zBEv56q$~<xJf_4+pNw3PiF72MeLLhB8NOwDjpug}R0iGDCI2yg>L+s;DDp3%b|A_z
z0uIhFtb#$~>eX|ZS(qhPImqxj>D^-D^YF7rQHE){Z2o`4*dGK8?KKKh@SPw3uzmQw
z@3vRK#p6B%qbpVeH%gtsD;0kIu}&r<^W~9(GFiI0>c%Wy!Wb!N@rW2#Ex!PEQ5gCF
zPnBkwn3BHC_l&C`X?IuM)QUkE91?PHG^0#<D2I%+%0}DMeo<|sP@-tZ5QC>Ur@Q6L
zDm$N!>37?GlwdGYP9WT+=S0W&u8uWsR%=n82w-4rqWB|#Omvi>l*u}~a#Uf0#<ZN!
z$z-b>6Q8{=?G8+d>18!N@TkcjeS>;Oa7BLNs%7fynDg$5W7V_6ki4yrc?*5(m9Ccw
zA>p+X#@$_a=$oUrMqx%)+v40ZOgWPy8Kaj(8W7uRL0!sTKS|Q`$RKnu-~A6S-OOB%
zM8fzT(}lJ$ozu-cs=$~&9zOjsbLD1!d|7GDAmR<9SMQG980A#u>a)=FWV;Jc@7^E0
z66V(qubV!KKYb3S{ac~Ihc0B9-a1Ht*d53%kEYtK4>AZCle~0_(T-pK>z&YPY`+QJ
zR}!bb8AAIi<e<!5kS-2TCSJaE8ierGx$^4UE-p#4OkhoXg%f@LeHxi;CX(SROahpm
z?y<xr=DN+CwpBWnZ^f&$uEWQWcFdc2XWY)}J4!`e`L1-kk3NhUlj&=Dl4dd9rl--(
z7w;R+*4t0rEr#~$i3vPNqfezmbbbT(D$}*SY}LnN9kf%@OMSZ>bR5LaNB3LxD0}(p
z<78yev~lDk-+cJCE#ub5uebi4PxnobdX*`UGM#%LUx&bnv@6Tk%ce>61nqOVkzC?x
z$yfgW`G5V*9d-D>`ak?%C=v+5BkUp&bOH&D&cYCQ4r#p#r0ht*-HEESJ1hLVHp_5<
zyM3c}3rNEiQgnO<!QyY+{1KH}s_a$}>sVp>5f+Y&CcXWmvv$B_*-5m!lRWC`#fza#
zp0CIwf(_oTq;>q(t#%EoEKwlEpBZ5WlM=p@Hy*Z=uKhxnOjtWV>*QEe<*l;cG+D6G
zB`z;z`U?LP%IIEos|j$@=G$>mfr(`_j62KRMKV1*ouiB<#)sR+(<kj$KmTd_F1*Tj
zzxTbix4+kRwsulCg<+<Dc(_Lzj>$)DZ)>X^AMUsJ-hDU1VwEQfjPE^oKz`og<qG|)
z97E;K5(eKAB#^hty{QR><Q|X5#CuHmzaq|Aj(6HWqO4TJF<M1}?dsyI3DWUC(|yXO
z;NGL(6e6vxEH#e{0{X!v>+K43;iK?W!HxA->9x7J6@E#D<cllIW09&2vig<s*;4Wf
z+FgmRP)HtFC71$!FP~SyWGlGi%=B_JGyJMM-y-A&UJ6Pn1TdH9D<O_NkMQtLCIuwU
zpIn6Ij=RZmh1K8@?R9FI(_S8#bkt~~c7JI$tADrG*W>1jZFjyY@U#tFE&ObAGrZEw
z<al<1xnpNxVYz+#+uv(Tw2$ytn5SS<-1qL71C?`!M-jlRt**3%IgciT=T_ha+_n`6
z9i1kw^_ZEN%8s!k+D65NiYal<J}fS}aL)UfJa))FA^YcDSM!;7jD~OB1h*96?(V9T
z*$4LEz@3o@^<R>fLbuxOW6uz`6=o@XJ8mB3ZT}NDg^yue-EK(Te9Sk(cwm^mgs1J5
z#=t~C<l7DsQ{V9=e;P0!_DuV!G|7cr;+8z6_+&ofgXJ?X)3z)g?RIx<wJj~qX91x?
z%hi>|+#RK`F*{D^)(V2%<>dUvBbZ$HwV=Qf_-W3ludCRpO6CmqLyxQcH~;!yw}%fO
zw!8Q5wRhitKW%T<GQHCa`ss=HpwLHw?IC?sfv;sBrjNbsj=PieDvb81Lyt~Wp<^22
ztrZUpQy<%<c>4&s2`{g2y>(YeYJYa+{NloVc2+4IHji3>=MGoffH?_u-P_%#|JL0t
zo@ODuhdg>}6#QpC>JBZ9?P7lvuX4W%^Foyxx$B4Y(lWG*&6GuO;NIAzX?PMtDG^r|
z=i2f9ZhQF2Z*p&-b%^o}xZ0ns$2jv|k9t!8?Ltr$i_1&%QP@pR(MK33>U5D;?FiU=
z7vM2{!M^Fq7L<U#!yL>VanclE<XEJT*qw+fK`pa5Xx>&hZ~?v<`*5hm38g%8QK93U
zw9CHaj*-Bg_KKp97`i*J>J=#kmpXzk%AnyoVdLz~9kR#ST~UiJq$i$)A#CX{z&3*U
z<Xz)T#;z!(?LY2{an7lrzXZD(f2o7*rX^z$WwkuQ%6x+_ng8o9?zxb*V6S{ZMNxCd
zSMZ|7GspQO=6Wi&Gd~kYg(YzJTO4ve7li`p*dL~6ywhktG_XIG&i*IuFNJ08!Vs6F
zThc{$HLDOZUl%91V^GC{JB*C)J&w+Q=fNL$Vo3KsefYF(?lAU2L|)KwJ-S!=E9~u8
z?zr&?J>eR~6!b|sMzvSXT`hy@+fKIa(&A!!|J?_~9m}rZXB+FGx6)>nGImDCg-O!A
zLtUeE<B_hJucz^>t}JKKig9ee`MHI5_wL=K^Xp&!Dr3|SzyEFOc*K~Zg5-7-g3^9r
zrxHeiyRhHi+t0CjU|z=TwEOTFJkPDr6KTs8?IhX;VZZ?Cg=pia_fe|kS0NXAhTV1*
zZ|rNfqdVtZII^}vyG@N39i>g>r-ZpOBNwWS(MIn28XE_;w40f1?;W)#n~eL68TJ?J
zFKriJ%+Iz~QRX{i7j58iyei;SoC|x)Q{zeTRrWXggf6D3ENp3cse|jHuG!TZp27R@
zNarQ<e0TJUgZ(&A{FV4ub(StE({LBC{DYji`GEvc)GW-q>s4N$ly^}U(k_N^L2`D~
zG2nYo<B~`J%4<;0Pd@o9iht!uS1&xCcq0l<anO4bmFvk{*bl`q`>1_}f7)TVjgYr|
zzdHk!r^yFqCl~GG7$e_SJcma!-eMf`bf>~wX_kFbrKR^pt}LtES6S&Xm}6-N6^imm
zD!JXg^ow8oERXz<{csYV&&4kz6Vs91y@0N|<9~U1v8}G%ZL@QW=}Ynj&*2Z;8Rt&!
z`)l--yNm7Yh`vcd7nT;_2f<7D2FD+Fm!4k2iyfc0J$QHVRmD&i%)^t}Zr=Cmf@|l*
zF1pE%V(?MEO@G50$F>_Eb1}+zo088{NeNzz(s$eyb>;Dm#Fw^tOr~R<3oG0~Eo|+F
zrHEy$1_$N8Hn;b~E4iRUnAu+T@$_?$Pu|w{zPr2tTo-a`z-OO5hA3aQAAj%t_T6uP
z5M|z{k2l-1Eyfh!=HhVMspgh74|1;J7!*F>OuQ6dEh_vB<#p^92aT7wJmRr;oGp0A
z<=Cq>^2(#49V1A!x=QF;ZyYmvBTcJsF}%Ke_rd!OH_c7E=H|C~4Zd&MZ<yE9s{C)7
zH{;waf0_FJytHqcxB18Yi=z(Am*w#`^L66=cK!bP%;%NspZ=47Ff{a!{`en(-jbXy
zK>q^gl#SM>2rB~i-!^bw@fy54tH0%3jv~5tG795W9RRm3!GZA9Ss?l9rpc@Ssi2gp
z(n@$u7dB}V*KTFnw@5p8gIPP%P5dgiQB$78)D^>r^;JNY&&0LlygK*5qw~K`h#&2m
zfwlXNNuV3Qx>VgwR|QS%_Vx0_f-SJ=(n+Qjr%&VaP1VY-r5K_FWfeE+QK#$jcW+F<
z(|uoicky_%%kui}5y?2ogYUSIwpsXLzAs)~G3cMP9hli8?5`$9+Dn-h23?s-@8%j}
zBD3-MY5UdBe%3ZWdE8z(yXp%-B_znxObiy7pm=$!m%aK-O;14VJ%8C9I<+#&d}U_+
zwoJt|0}BX^3~YAxh|^@yGY~3sEWN}uXS16QaZ-({X5t&p`=E^H01<(alK~h-CndI_
z!mS-9EfRPKZ<(882JC$pke%%vCVr1$KADVvMPv3oDE``wZ3qlz8N|N#qaU^VAAZ;_
zROB#d3X(b=L1mB%9)VcQxzK#UU@SwWpUj<%hY)mkm*=s|ZOQ~?yS|JjIB^18WHkEY
zlw`sp*a5Ci9-X9m)WqBZ3<K?{U?P)K@-};!%-fm7F)0f(cNw9cC314&YMv`OUc>X<
zmoF>R!70n>ZYnD$J~+v61(xZF12R_Pge!JUFOw|lB_lFQd<8K5q}oiAFZrE<+tvqc
zBOiY02Qr5_DvHK*E4p|4<>F=DhBN&VJ~BZNZ^B@~_&6z}-mVVMgp2$vlPg|KQ^xxW
zH}SK_72nqLRkthY4wE&2)<#Hcd483N^-7zbn86QU%4U6gS}fk1LScXX-h$wTcAA*8
z2JwVHVV1l!SSr0rOFE`qH?r{8A{E_E9-T2fVf9P+BvJYI#o<-Pum<ehGG1T?<A)#R
zUFl&fB}%<x42oxpU>RvT^D=>?sej`6k<!0bX5Di<Zx_aP><Z0yrPCKWDGuB<d}ESd
zpI$=UD8V4Eej0pn>%;UWVGZP?(KChCR2n+n9Iui>506BQW0jBpYlDA+cIkEDEki!L
z`X(H|=21R|2cU-R{Ea1^xY(h<P&wB)qN$pdz01e+aEKcTRf=V|{Pi-iYpC?8<rSyv
zRO^>--Fd1FD}>{F@e3H-(+}#O?<#-tGi(y-%V&D{8?Y~vVd5{&=v?;l!AP_Qp)}X_
z9v(h7V#K#Xm8rbDaOSTsZq+3GC7?fs&~%=iwfl)lo|Y-)toW6%zF4fIMedBweX4%h
zu9}{|@SRkEsRgO8o8>Q;e8j;Y@Pm9R-HvGRq)uh){K_O+T)VPh`M>@@zDdQ`zj)>S
z296ybE203%)xVA@t^gjv{Uvn#f(fJdySxCl3RqQ$xopM-50?D`+MP)`<3%3etUyGZ
zQTU+HR{@kXKf8z>KdEKtynZC&E)!RtJCS9eA>6C1R#zBZJ19zcU_MTeO~+%19V6vU
zT(sl(k;M~^VVFUCixYX#J~*1c<GtgZMqbNDf5$)LXB_PO96z1dWGvvxN>1`g|LTn6
zSSe}<M`nKa!fMBVg`L8{$A09U1QYOl9<g?Ku-o=_x8Nc6b3f3JfA{wyocZ+APjd;m
zmx#NwM?S;*Ae<CiUss?nE-ts<{TqLy&A?9z$N7b&C_D}hPvOlRgT+I5SIe6<r|z5@
zj=*Pjda5lgdfbjeTz5vCw5@IOID0{vJ;q5P-Fb6&zkEt)0Wd$@KgjBTd7EJc+%q!~
z{w*)g!3V$0kz!sZZr;6e%+<d_OJVc@Jf-|VPP&ly07uNr<Cx{O<OPIzcCbu%MBP*y
zb1|4oS#U<-n5)v|ch@&I+UC|?_$~#CCknccJwD6(!A5h`n(bc;brdqF7=rgx5KP|k
z*7DNwOH=MpnVX8xUICN$%80YBZWlf?<G0(pEAwp(6utB)EZRXqy+=}6UwP+8kDtK9
z+-mn9ywkR~*4ys3N-&jbGi_ycEy{^pNKf6y$0vj1yGN(FgG^k@-4V<ICdlN(opGBA
zZT8yk{xLW&ZWgatyWOUN$(`n|gni&Fti1!R>Mrhix7e$+p#nj;&ES2MAn?J$OT~+P
zs{JfS31-KFM=W_lgng1c317STOK7Z8P_lSiUJPST0mRi;$lHD;FKGWXZssj~JC4|Q
z2#_#|MELYOk;8}fwX<g7l<vegO?kBvWYS&=v2iP*XPqSAV_vT6mRDAQyXelLrIiT8
z#b=L%T3ncJOY;+Le%9S2Du`5MowuExec;So2m0cjL+%6-hL$}-Ioio`s@!WYFWaB}
z$-kw~?zA8L;D>GP?%g&Dp1y=9Jvn{R_Kz>xvyHvBL3(x1ufq2d$X%QsgP(_4G$&n1
z*{PF90jk`!oS~Va8Ng5YKo5mW;VbPbo{*O?R-t1U5FznYd@2M~s4ZSANZi}o&5kbT
zCzs%gFtaS-zwbcn$zK7n_x}kG1*f$--a0b|^m83|DcFak|I#(`aqO@TreWE9JC-WQ
zfBAxWal6dXphwWJwY9Z2JEyRA652OK`tEdOt^}=DP;0*C;XQNH^!HisVw#z)`UAhv
z2XMn9AHA!}vA~@%3$rtn<#rU59#OV;=$(isnRnRlGhgB96|l~4oy*z}+@a_lR>Den
zxS$|M64GA`YaenKri%kr7OAMy)Pg?7B>F<iP2Ujj#AwSYJT>R1hv1jbDf!Wdpc$EG
z2}jyY0e)x@Fg3ljdDn03Ss8Qu>&lYFnzTpXXY5y*`(w8)ppexMrK9OoDk#7U9`VN=
zcXB)0g{xtNi*21T%D!r>^v!;3tbT4^6z0|w;>QCH;IGol-Bzi*d4e-4Gv^nV+U)Ec
z{Ro`ET?K`>XPv}V@f1Rf|Lu1F{tA6{G4<TrJA@uxwq^S2!rXL(@DCq8Yo9%NPTN^l
z#;S}lKJ+K&r_^sew97@D!t#=Fhk3b*Eb68aPbn_njEewl3fy^c|6aQbUaH7=`s`VD
zX69a~u3c|Y9u>h+B7!5etF>1msdO8T0?n!)9vlLX);{|1gZANv??$2SqLFXE_W+tR
z-p(kq{GkgS+}#_cO#19l7BbwrHNyDuyghmPEO<LNGtxf#)(35AWewcEY#X~r*->+P
zMjOztZ3~6$9+fD+=0XT@w&<1ZB&<9p@`5_KV;9~GIwt?Hx;)z++*?gOx1$u0Ur^b1
z6~));@?tkmdH=ceX}Dc7@!8lp%Hj^kMHQH%^c``}9o5nd@kQQ6g`Ry%Ug+$Ue(CXX
zd+;$TA1IUPC5^Y=M9E5ewwcK1SnOS7t1C<G?%lPF;mWEkpUUm<6Vx$%kjM5k?;Ixz
z-H`{i?}&4@i#&ySY3w`pHOnZTc*2rL``Wj~OLtVOFbwZ8J`y@QJ2#VYUg^T_&Tcz^
zp1L6X^70h)Rn~Ks<LTsq9Gm5%<juTe)rB_7du$Kk>|&dh)qC`h8TjejfsOe|k5vXa
z&R9>!B>SPe?cDuldF7Q2D{hR8-cH%(=Vx+Ur}!`K*}qi^Z)|P{X6|ZI8Rmj|`?zi5
zan5H9#^z<?<ydV`3;@SN?}4d1q&Bv9+73L>$<YD4&sh8Zw;!|zcNep8-!{ATs+3yh
zJrDnR>-K0ng^$^W_uSe&X#0o9l$BZlS9v7i=favO2Ei-GE@8C?pR-4Jj|w%MIOABT
zeAHct9{s4y2oQsdE~viBQRgRTzz&>NmbS!L;T>&`LDpM*c31Ws{mL|V_fN8*%VR91
zlcp(uu6)adQ!WNMJ=|+^QzPw1-+#Zod;e}0-#vZ$9DZ!LefL}M(=R5ngKeEYaCRY`
zqei4va|U^(nzOhF$zAf+-(wpiKNaUgOK*p^`q)=A(rWQ*f%NCc8BbrFwV{_P3gcb*
zfDJuOmeBV0N~>^Nm+LUyw~4$K)^X!HtvA0l&S7qT^YQvO=JoHIJKZ#Ie!opIbW@4<
zdFIV@dK350_j;>&TUu{3Unl(A>i5@SJ`ZO9^q>6wp`ri9AO9noQlb}@*^q%z7zi99
zC<57jyQszyQ7UXgn82F>O$3wlyR?lKW7@o4ixLMtI#*wO=EaYQmkDz`$rr~yXbg7U
zgl#27={o-^V*+%Zg~+L-zK%*gpdys%(k|~r(`52ezU5$;_#0Pfl<BKcN$_Ez$itVe
z2^YI2rHfndH1N`Np1tta*olwrW?jpvtCx|i2ZiW-Dw!Jv(;Bx*q0%;wE^H<4R~HWW
zByXJogv_<=YRzTh4QgyJ^gLvM-8*M85A$#bqRdP`lR+=19lkZ(&LON%KKZ0Q{P{22
z{`1}BF)*YXf(|pdI5!vNm#by_?d4%yncPbdRbn_P$uWg8<q}pIP6{&yBj(`kfNr8P
zJC;uZ>R$$W(#i@j;>uv#i4Ehx&x^r|L01J$H^|oxqMCfymVX8-1{GIAIf3XW$TF_3
z$a?tM!?v}xl?g<(%U{4y{9s(7{{_={n<4F6)N6?W?7KhuLA&?vyX};ET{3C4juDEC
z3&UEqq)^Dof&I{xKPNm{T|+wALJj;C9>{PiTu{(*1%oxjAe_}ke4jGN9fkpvX;#Qk
zLd09NqwSHlo?mFAFd7PPZo@drygA`gspN#o^S{MInQkxEb|sdRA>ovjh2)i$?#3bQ
zLD#0Bws#5KZf6CM7exn0Xn%!_3a?yQ8o?&ar8p!L<wPqJW*9V^K}NZhZ^dNUUz}*?
zd}Q3DA}Y9CkrV+%m``_^!ALTI%h0%m+_H#M$4pS%o}U$=<ZC7}c9z-xFak3q{-th&
z7ay%#`W<DizS#9+-<L1iZRpbS>~dRMW->jkz>P^%heEEFH0|_>-p#jk!}#Sicn|0)
z4rIlfP`r(=aTQTg!%go2{V1ZD@>K?943lvoPF=kE@iXzOY=-o^a>`d1$1i`2Dd}9|
zhrM$c)TB$sGNc%$FHP31OSJn+n7k~Xu3cH`|3<ol$cC@CQO!C|zOV;&#^4=4-3{M;
z_i=PfPCERSDR^u2PI)>L=l(P4lw1GSWQZ#pPfA~~N_Z^(Q&=cvmD!-b70&kJ!?sA+
z$|Fv(l`)4zDBr}ZKz^6|p#Hj0u|9TU;9d&63hQtA@Qa(tn0sGNlkWXBsbQCPhkj-H
zAUJ64bA2s7i6-5fk@32Qw;u+GZQy5ljMIneJi5AJDWDAE1r(SL5||2CcHOlRDajP9
z{pnImnZpn9emy3p4`(@f+xC5mz3G*+@-D{qt3nelAAIo+Z82a{?0?HeIOtpCwS1C7
zB@}vOwZmo&-2J(3mC8Am$};W7#L_~<Uj9{jLi730zd^;<KlmsAr&sQX@c29@?$b<e
z$H$@3*cFmf(j05zJ*CHtMMpfh7*i&hgnJpiLM!X)E-lBmE5;VbE5{kfHDTm<<`W?Y
z={Zq#ywf;gb0z;C6XlJ)queK@0_;`p9#Tj}CeU#O7`Af>KhBw4mS9KW;C1`RLo+H#
z_u+?(W7=X@6kV7I7vupbqhnTtjEpxLd8jmic&EKH;h%WK^`doU%-cK+TeQzUBH!i8
za*_ca#`B~24ae8IpJ;dpxOzO##c6x`<WYP0@Y5(EzW4p_x4-$f{$^WPTuNC~5N&U6
zv`>Hi%l6Bk{j^=|?|FGvo1UL<D{FV#<kT!<_*evt3L_%yqFhHur{t+XeJnhM0>5F}
z)Z!@gqBw1mvUog>!V`*2yPdbXCx~(ndjxZh;|c{-sK_h20LCL|Y-izRo69rgq=K2L
ztP)qSzq#?8vhOFpO3SSDy*-@HqHwBKAJblz(RTI`&I)5m(}|tC;pCTG6sA%@zSuh(
z++pPkWp}G=tZ%dp@6F2U;yS6o(+L#dDx8dPi}Dp-Uf9UtgTiOYcZ`6S^8U0@zhh=?
z)gud+sn?z6(R5KP(CDUNzgU@{YIBn=P?BFc3q7%IhrzpF{^FNyoieRGc&E+6=Y8_Y
zXAx-F$0B6EJw{ru+9>_ZBlEmJ#RYRGStac8K<;c*Kzz<T;-qb^KWmR4J!u<TDo`jt
zxr$593+iZ`i3kPlM+&uERWGmZ9Y9gc62o#ApI<A!ZNmt3DX;amFJ@PlZ~>PTO1aWn
z{@6a8RoK8zq2#MqDz<scqXA#}OnZd+i1R8X%P5Se&r%orR|)hLx7&}a2g#o#Ciz<q
z7dF|COE^?zv%Hzx@K(@<Yxq?Aki4rVG=MlNT#RR3<vp#VyYS`~7TesM{X^lUO2={L
zUXyJxim!3_(UAy^y|?Q5=5~Zt3gC=yn%=b{{fHuhI=YKRrKR_pd35IT$};$AA9B9N
z{O{r_0yB?@JaEUQ^o8<Ci^LZdC9i0!%oo9V>)}(o^gLFtdpzX>d>3l+<L;*MNGtDh
zae(kn6@}+V@YTXP^9RzAm$!{3##C+(w@L8NT`MYH+-+%_*6vCbBtX<M@OV7tIeyZB
z963t;ZAZha2yjv4FnFhsRKd6cS?}U;x61y`cH7?CpezT#iGBh+>{~m#+wBu*m*L<0
z;5*Q%5@OHOH&)kHL(6~T4vNj4&<o)b<^JJ8+S@*=0%e%=#wY4N9Cu#W5Y}G-?<>k_
z8xB)fg{&ux5w%mX=2gLK+lD?O-FL3$E))gGqF?43;w|<%@k9lH$`aC$X5i;tUhb5T
z7Fi}~zQSqSMY)3GLXA;{+nH}mWT8VUm=%uO@0{Z(gtlD$m|@<jCv|k}65nf{-Oc-G
zY?HknW30Y&y~@L~IObM+*+Zab_95djH#KbCZ$q46#wI*eU}X$`Rb#8TZP`*D()E##
zaUMmW;B_5S$aqaY$_2zr5+fhSH-+x@UwgD;uSwqlULLnO8(O7;((zY;y2l8LPsX4Y
z^r_+Y&V#${JKz3R7K9uepMZxaZGL*Zee3;u;L})J-`H-y`SeNK-8&8s;bIIGDcMzY
zbeOgfPAczJtUAxM^wLAeR@*&F5BiCGiG<sJ=JARi4eWeeW!n1YR*oNaH%xX+y-)yM
z$}Zs$1tHNhKb1`&6YrT6=WXNTqkTm}?SuE<0p~vmUAVh6)84<oNLrWKy>@<CM+Zk1
z15Bh>?f_e3O!sc3$B&=oKFG1LTkWH7z1QA*=Rsh-?H#B4wX0F3mXCet_{cj7YY|4T
zWE3ZaiS)?!a`A)xS_PGNn0c4b55E7M#Bs64##Y@Scme*8kB_t^`l#|LcgHHf2p>lu
zJZFrmyB<%3B6#T7tsLVjeb~B$FQ6Q@xySiVP0fUkxXX0MBQt5c@MiWqafWflG7DE1
z)|jxnbUo79qgp4#Bb7+NSzL2m5&hgTq|Cs!bvKwtMXFfyNI?5hFV`s}HVlQX559uC
zx6*dr1vWZTIGP3EhNJ&dmC$E-6ln#$Cw);lh4Tz)i!i#swnSUKgPpN~vN&c;058*a
zoD^68{Lg=$oog!TXMo9vAAPI6|K7K<P)xe(!WQXD?ige2IOHvTQQqgtP}9&+^Am5P
z;DbJ?K$ag@`Qy>Y0>&}><n*E))8-x_D2*5%hMTva2WIdJ<Y^i0f9`s>t;j6nN8~=>
zj0#lYsq97?HZwKazW?2CWd|c6vw*|##ohPr;M7=$9r^j4gWUCX1aBq}I0QY2Y|4cl
z@TJ1Qz9=svp1NZ{_tKqHHqx-K^+$0_--SE9k9^>NC*Vz}i;I2sJUQUvMH_)%jgpMI
zI97U}=`nte8Oq$u({gw(WGzUC-nq!nV=ZY<$HpDyneea6OVjNK-}$h;3r&&_Q+Y10
zyf`-%{5UvqN2r~@croi?8prT(_DzpDJvcsZyZhA1`()z=ObL_43*cPp4lYY;qf7@L
zBaq#x+b`P0ofpgzC|^foL-)<ipBTr1`lvBMdaY}H_a@)c<%GEof784cz7Lb%8-Bds
zjHkJFt1`Z6Zo0iqG4y##oSWwJ($&QM^S$0`-j>$e%-0G3w)*{bn9qZmim%&$`w#x`
zCspWrVudE;i=PtuX^I$LTXm<4;WP*#-gLq8)|iN46D|%FzcVEa!IUQkhd!`R_j#SR
z3{buMa@VKturfO3$>dvm=a7(Qn9l_0UgamFlNIGR()A~4cX<mB8_po9yY@bG{IyrS
zn~(nG?ni9p=gTb3yFOzoZ7inWhws9qF?IK=d;3=kN+$pb<PU$9R+-MC(4$M!cZ9>H
zbK{+ue)sRr&2nlFBNIPA`2p4i&!Kk3TgAzVD<xqluPE*vCMQmUWTjM^TtVP=p0Bs3
zpFM2*FlrgZZx~msdMul(VI0h4w!90$BOPQGWE!1}n5N&tgi_#+4+Ph8N;Jc~4i)n(
z!L_Ja4V!@lqUtIl2Wi5ld=OR#zc8l+&MqYE6}|=}8F16eiY^`*sZ1V-=KyJZS5EZI
zt4H0OFsSL@74mBtVcIL?{IiZ_FJ?Mv_K$|5r=nV@Ys^jmH{Vv3XAH(^ZgIX%%*;@~
zZW0chJ;u-a6?5q1gn`E9W(+ZZC#F$a2`B9&=#!(Bh~orC{E!LBN;sKbkBNwYf+7l2
z%S@Ut73PUtw43b{0S^pJ{BU=5suLfDNY(Cm1?~z<C<AF}`iYS3JIVkrlViC}w-`_>
zO#mJlbp<@;W&5c#t(8`VKmG}<Vh&~Cz>`dDWR6}02Ehs9Ub)K7dWvhID=AFAtwbi4
zRJNZ$8MgOVpys!A%&`K5Gkp_DJj!i7z~<nRPx49W3O<^A+o1Rbo|MbgRj%@L)m8~z
zz$1$!3<Rz(>Kb?cf?s~)724Ye^{Jn7sHDoC$Nzf!?JGM$`p88{dAgVJH3Pf(6Q>J4
zFe#fp<(FC~3_tkArz>xs`00orc_xlw@$2o@(ei|Dc5dZfabveo221`JA}1aH4HL)u
zu5#6rG_-g1)6YP8C%$R>uJ7L77hMy$#r^u}+oI2_H;G@C3ftR#5wI_Zu~J6UDu?T|
za7&ave(K&Q;`_Rs{dd_bRpI$s{e1fJ_P=`h=O^j=s(3sUBB-gGcHKIEosHAed2IQQ
zV{dL)HRWAt_g>~P0ELE(%*)UT-~aM5Zp@85`jqmnK=tZ;f`VEsIhI%xw)@ta&$5_D
zg-MK<s(0t!{Ulr;P~R?aITf5@bjZU`=NNpAW5M6FZ-#+tmLI>GqE!abzxqzXCY|eW
zhVX0Pm7j9eZW(2_jNijNX&EP9orym`T|xyD+l|j8Rr%Cd$AdBazx(q)`^N70dhc)j
zhd*&~h(c3gkaO}GYb=dpWCWx9xme<oGP|QefoTM2z|e`W6L~L9&*a+4x8qgnM10b?
z4gG%QSczR{kNU|GE~6^jp#Qk3{J71;l@N><3i2F-iq~+3ymZRlXBi6tdhU24j_3P3
z*;VN>KR4S}SL=8j?@Z|LPxA6?@6a-w<7>a0q7)uJv;m0*Kd={0nvV~Kv#nDOaE2%T
z76M}!O}SM%PJ|VdRoa!VoSnNzESGmbc(=`?Pk-4S{pOeL`IFDG6XU&i-)VpA_y3^X
zd*_`J1nlj#M-M-3&z?MP8_%Al&iCJa(B6ChgZ9z)f2V!;o$s~z<$G;(YPMaqJ51;=
z+s@8j+u1GQw){ceGXZQCX0p>|fr<MP@3lo>wdkeYGw`17GI3=#{K_3KP#@2@>H?pA
zFML<9aITW>D!h*{k++Oc0+`Ot&$b1Rle<0AcDHugv!~CqD!Ps~bb*cv$%18$Vmd#K
z&|=WXaL?Nx;3s~$LRx{T!XDif;>lmB_!`IW_~@Yh=GPwso9E$0J+|)x-dll{by3*l
zQFqg%A#R7~qa776xZub1UFAIo&UiGO0=`A+FgrJ$9bYrk3R~{vE(v)%l__<<t-Yc}
zc@+f<aA~*Qg)`Pp&OCC^`_x9jM|YqIqnS1}J=ey7m2L{c+)<-I>+t9(N<vZ#@1`J7
zp{SRtyQ1EE#oX!Th4C+G-{CQZtCP$foIi-m?trte!s`LwG5UjdsMSuPR}t{JP)Xsp
zN-K93%Co8*E8ZS}5Z6AIe)gjHuPegQ66tHc75*FFdsOVh0y6l{mp|sEqS20G6KYJS
z(sa&HbBOT27{mAPax;?Y)qKRhRsF^|#d9jKQwVAPNef%<XoF{sq8Xgg?y7xvQP>Zy
zo;ao=@f_GjhM#cYxB_B#hR7e2r$SwKKhDi9@XSZRdItWhZ1h+y!_MG7H8T(2J{AGj
zKK;!jot~_N^BX&Dd-pK&A$O{}lV%hgv^=h&S6D0UaZ#8jO{h?_57~w}F#(z+@Ep^0
zpPcsCwu!*gol`2y>^JrSalP(Ma<|?k=?@2=Jt9;5KL&sHp&gr>>uvqnQ~EGC2`yH^
zEv_#u7kyNbV?6txLclEcJMNS|rc3+xY1;Oa%vo2KmfHOXciY{SrM3j^7#|*L=g0f)
z^6Y>y@shr{Pgzgf!t7MLe|I&zR~2A;$MEpjOgm<3lq1`K>o&Ok=<(C`@X^z@w|79j
zE})C{Gx{-Taf|W5`(<9h<ef49FwU{ct|(6EC&cmY7zLKzspsxY?@HU**=dJ|2hfN^
z`qKgayOhh5I8LL~3Oq9xs{Wx}LHX&~aRk1OfsSKYEWqPi{1Qi<k6jYhJMD7UkGsM`
zTfleT&c#emcq%CNNG|c#xr&2+Xq=|H0N;#-;IB%hC{)3P(BkwZ`$QKWlkZGU`?0AM
zozj3Pk<*_H*o|dU4Z=BYSoRuA9fMVN@g|7rSfq??iDA%iI>g}DK4?7We(|yleg2Mp
zl{Ud}m(KH<IJG0Wzmw>|9m}+biyjo#2ZpqV$B>mm<|Kk|aaX0m7UPGD8eH_@eR%Hr
znr0sF^#Cp&aTk*CR?%ctouh`10e9a~T2NLM$ywkT{-)X}3JdF#LfcpIBMqsYN+9JW
zCTbee&@1UE0j#h1AbpYsWdSH@L^eYFlB1}UdC^pyZ*A{1cYGPgog(%Hh45y!yXVnw
z_A$mIXpr>11pc^*ci{aH<H+q{PweriNa>%7*{djRbB`MR#lG2NNzquxIAD|=TF^Ln
zEZfBPv$Y&Q9f#AV;I7BVt**?6x3Ud9dQbWvoTgs3wfN~+p_1LB+mvs(D@w(riZ}Cg
zJTIE>E;x^^QyJyXG}<3nMD{>HVei{^a3QsAr4jo|0pM<4c}VDV;0!EP0x=${j8u7~
z;?8y#KUJ=|GfY0u-D$M|!7<SIwukgyIgAhrN(7Gm7~&7L9N@pkAcfn?!c_iP*IHx{
zB~W0`cw#@h<sxOD;$K~C>_Q87bS*5dGG5HZ4X`91#}WBa`^w?|PJ0EdyuaqI<R$P<
zS|D$08`4gcU0f6X)}v&4HO9M;p%yOp_YRf<MO=|a$a#B2`0l|;7M1N$_PxX7whhf(
z-=uGC?Y1p=HFq}(S9y5vmwcw|@|pKh?m$!B9VN{K&uBOC?EwD7J1d2~io#3aAfN8Z
zLNWF!jBV>8wH)IMj;P3USJys|xW5Z;<MEY;&;;99MXSe5+y6qJNXCAr64kNJu}Ru$
zSa;#AE-ysUXrG#9oK!C7y|gJaIP938dtVv%x3{*lc;DkF-F@ZGV&^OFWZQH{8;@h2
zjpSIWT<)B5o6ZRB<X9%3DBtNlcP<FBU7X`Me#leFV~Jwg<?Di@>7PY!?4c^at9@#o
zU_ZD0taB|4a98nI#?&l2q|ZCYDx~!#?Tr<q{ib_259z*pGx$!J8}2vkn*Mtb{)YJ~
z@haxmi`$#Gx%cJ^<9%_|-!fm8*PG_+g#Wr}|7FbQ<@MkGtN-So-TuQr{M$bf*(7En
zWFM?1S>>RgVLD6P!|L;|iEx9z`b@|Qm!ICXGklwyzj{Rw;$zxc!Anz*-G`reJMiLe
zC#u+G5~lm^E7`qn)S*I|YTUZ;y6NA4mr&84&X;%T$@oiJiBg7lCLM7goSyy*3+?HM
z{_|(ubV?W=Kl)$39c&WD&;H>jaeIgQ$-6%JjbzBLgzW6ySMS$_>xEzWt6;ouxQ8HE
zk1A)r6Rv{?|3E)+Dm`9$>)}6-b#{fe_aU9Y%pNijUdTv{$h0uvkmgGW+$Bu=?&fBD
z{_s)Tdh)DY9F+p3H!3A$XdJ9v!4$>>21R8+F=igC=ipz$qh2v(*&LWHO4>Edi;uO|
zv2&*!zO(WWf4>dmckqie93+k70OBzfefw%x>CrRrfdS!#hs?8#jk}m+WMq&OHspAe
zZW3K9!@iyg%q>i!b_|g9uztSdPdf$YtOtXVZWt9&759yNaSvSZa||}U;qlQnzqr_@
z8N3v_%WPRs!$bhaV;!C7<+#Bh3+WQ(iosXKjf@r*%^>C^%8OWYd>;c{COaAAJnEtt
zKmnI@OiBjN3o#j3d-f{rqrdr5FIOu@_{X;}w{UsITK%xM2R8zvTVc!sf6Gf;<BLm9
zq-q7M%ut5|47%cxJGq2G;Rj3-_A0Y|p-W#lfGfI1`AIx?RAuELcC+)x=5Zxu>Ph{?
zE6ZMK$~?<pzG}n7vpmAtw0tbT%$#^<dBia%8advP^u+PNf^-7|Cn%L|rAs*To`jj4
z8lMb<n1~c6J3dUq0tRx#$)7<gOq}y{cGFEhy?f`ESJ8&F!oa=!QbtYQMa2W2ftv<P
zPcXaB`T_1;TK*=F<dv`TqQd!>WTlLic17bSb@kCfQ~CRw{**BOCqhD)R&V<9cduBA
z+F<2%UFUMR&YP!z(&eWE=0@7xAX;Jc?!%f6W$xqnu3-LZ%BSGqC+~!>`1M_(>)Rjg
zgHjCKDs6uGPW$#{&P$+HSYP5+!MSrtKD_zq%Wt>s-SijqTFt}$V}T0h+Hv*W8PU<N
zv`e`pocZdKX#VjXFedR>_%k`v^Rw!nv=iBgalyUMvkQ}${8I2LBZkVPYhd(GG8L?z
z&hpy}W>Ukr#W_~e4Nxjn`Cs?7q^rIA9fa}Mbo`68FHpHtqAsM5zxCCn+`0DX)|ZP%
z@0dPB@7#ZP`E~YdKJOY{ed-;wu;63ed-(VMMlgu!nU2mCpbYjb6zJUY)@nfg;!M7#
zqXCY8`e)yq;_FBMqyOk9Do9+VsuIbu^OP}t=YWYm6MTifCl{~U_5r-j;R%ypSE0M$
zfN_;^!aBREaeZ^Aoxm44`5t2uoE_<oQ9zWS(D-VWRNFC%aJo6hdS?IF@GWTgXpXm%
zAIdnKiM+$EyCYuWu5fxh#~;m4kG0t;#`XEB=6yv9!W4SsPD{r;@^Sod3{yCm)s%#B
z*Hi8};W2&3lwQD4LXr}6I)-L)+cm#}LdX8JSr^uLnatjHkqK#z43qyz7B#uk$Q9gG
zuF*C{8NI7ylp$(+<9Ykl&;G1Ee)#Klu)C8n`<-{*X+QqmAHy5oX}|i#&)Tnk`ls!)
zU;P5U!6Uy$+x`1%?Z-d*QTxG<e%L<z_D4BJ&r<^21>r6ol@{*GaKVzRcURYz+x@%C
z?ZMh&ySp-<i@6tPC)>izcw3#DYIg&}YWMJ-0NT61JR<Sp?5cUW_~u64tD&GEyW17?
zIj+lZ*shG{9t}7<JDbMc-r8y#&!6QOM91+6Y~Xj@=_T(bjw`fRFsv!w%XT0fb`?bt
zR02a+HoIFuKEiviRMv=33P=<-JaL!YXAj#+x6)l9WRw7;H8nHSX6NSG43BqJl;DfD
zr>zv4s)Vv_6f}F(ol39e<+%s|S9s4%xpG+no&xNbq4(zPt^{$-3;*5u!hi78u-n^P
z?dkK490hgx;?_W@=}|`tp`^VkUR6BKOsmisiV$9bj!NDfO$lGUcLc2AHN{U?I=ho@
z2EKJ+aiPu4&u5{Wa}LX+5zgNCDt~zUw!+!sEra{yDLlegcHtv<44<Yj23kfKNl*!b
zX;tuCy9!Ff=3BnZog;)d-J)6c=AxG+_?0d~Ed>^iwuz5CpidNMrX>=*@t9su7sM0$
zl;)*J$@5g0lrrwMqmqZPDyT-~{h&7^qZ734M4OqJ1=cgzy<=#SFcT5dxwvp<b~*x5
zal&#t=E#2w1r;jZ>mh8{R+ii1lE)^AkKhTo=y8c;-rQlO@a}N`sO>}hwzl>n$nz*c
z@1F8L8~dVjBZb6SVUE35I=Ilx9V+flP|&PkSbzIqXb{g8?W$tu*|VqZ>9gli>X?@&
zZMgHrU6|5hg^SM1R2u9k;Pco|`rr=p(8K**=*wRE<tFY5ITa$S(DO(wkBjpd#zV@h
z;>dEl%SoQx`<Xm?&HgqsHP%+<XCeTc0sq}ereN+Y$IN-$DKrOI&n>RBkpQzmOuxyw
zHoLgi7T4~#iK#jI1LMh=ihvi46Gss?3a8D@?e>&8@-}eu&L$OvDu#1k6nH+%<0^go
zj|=h?AbM(qiV_uA&cUMigPw@M(viXom3Vcuy28x!Gunc=nu_vz_OnauG72zH?KrOI
zjC9Z4B`9rDQKuqmEc8O03=IRf9gFI|H0ON^_uZLQ%E%~m^rP)O8KhT$MQPFXuV69n
z*u{6s>@H>Vj1ZKuQ+%X*(hrL(DZBwM5}ta3C)E!fi|pIfo3f>ElWz4FSqb}pjkS(*
z_MWgYykiJ=oldz*-xHtEksaV~7-LEIv-q(JM=)uLkA1vTz>t>VO(C)c@)M@Q=Ujnw
z6jep9l4E{Uh|~_22&KUf1=a4Heg6D8?{#R9$NtqK2p2eJhYNM<7ilVfuvk@|Qi2Hz
z;pLcIbkk!1#e3VNqiuOgzS1jka^xuO?veZ1Sx(<Hj>bh4DyyX%p>M*3C-9(8WE|^g
zuDfdX7z+=f7mgz;yuFLb$2iv0-60UN?1FSops?IJjxweT9O<I{F+((V6^Zg&(x2nw
zv-Ep~{U_K}iaGwf@KgoK+#F;6(rk7W%M+xDslSRnPZv^<KQeL)zF;b45)ZvgQk?eQ
zB^P%HS8>dR7M8<uD+97``S!l>)EE9u9Fmui&Z@Lmq3=m9*+ncT$N&KU^hrcPR4pYA
znciR=QFf-1%eoc6mbMM->f8S2w>V<ovQMjQD}^?Xv^EQkRY2FS-<}S5bg8_A$5UEg
z%iiN<3>ifJLH<46&@FroZt}z~>~TDicc}Rq<qAhz{E}(eFLTWV`8jS~(O{>*?(Fha
z?tn56>#0)9{JjdryT$I^dysqJRD`-X-upFAsf&EYn0W&`cVJ0_7ook23*ZgyIn3A}
zoDU9LjOw2y`=CG29=Wp<yz~A}l~gL0l@qx;*`0pJhaS6HyH)!|?on0($2u;O_J@YK
z&`BQ1xb~|uH8#nk*rz-WI7%tpY&VY}6h|DNg{N&Dcv2tfNn}6vQQo<0p1QarS~$6L
z%sWwqfxMij0)*CsYo5MjKXd_<yFL5+FqJ1cUob0o%DD^JMJ~#e+{yXw{nfSv-|KPh
z?k4k!mgA&ZcY6s37ptq-6o(vZ<O}6(r6HcUWZ${F>ixAUR%?OvIb)GW^13)mMP$*!
zqAQNeF3>n*?02kJY3O(xT%laSHM)a%CvKRgira~a8f(Yh_7&x`3y^H&!rAUs@q>Gv
z!%d?bPeOF|IOn%aAJ2EsxAot>d7EAH+ut=l^ybUMzy95ux4GXmUlb;_{*olVN*-@B
zUpMYwCa*6tmF{n+9mUt*{sR?Xm`EnNK(x<5IH7Y)$d@z&(|PdH&2R<GSVGwIZ8~x0
z$8VmbnLxfxrSh$0D-ZuO`1I-Jr?(S^L{hAB!jMYx=)ZYu@rSULR-E}3fsLU<<sBP8
z34*8ZYlfq*UxmIl9Xu?v33h4q0V{TW>!0uXug5rvSKocwaWcGhGzg#7>Eg>@mk&mN
ztqBvCfe(IVOAyn`+k%z758vAo2LG;BiP*n24A(|2GkBN37YkCvJdm~AadCKY+KwT3
z%xyEUO9b8h<4UDnCcj(H*4yUO=j{Y0ztiOmjDxZSRmRC9RU~#Q!5lDZ@RF%=1&3He
z`#OkcV6#XyiLKp0HyGOf!OgG{CV&{$-fCn;SRr@$JJ6eE7-<N$u;ahN^n+3{js6s~
zUJRL+BD(^_ZS~n91yOfJhC&A!$}cm5w+C@=`bj{=uC}ngmQ6&kCOtD_9TPTfW8C~C
z=D-j)#~taFD#4~OFwHJ8aLvyKC|N;{m+t);^}@wfpDLh)uZ((xDKdf#@-me&jS5|k
zk6<2{Jh{SH!9&i>XAt%1B+G4l8$U<9T*2(YK-7*eU`cUB3)|d5&-0!==AtGZcoJ8e
zPwK`Kp&L)Rl19S1B1R?`=36(M98{gqCtdL9y6x-O0Vi;l$(49A9$_p6HYO_)bpy`X
z`PIQYySHTE2xGdI&U}rpo490oGU@6j2f-`+OSx8fDNbs~xyox@`At6fIiZrdb%IfG
zsxA(yrsZx%>k<YmiCA8DEh->V_&E(sZM&*Z%H}U=Y4;I!$yO7G@~fg)zBs)e2K&IN
z^S2FSw8vt7y0lCr|E7nfi>P1HzwwS^{glBPA$;!?`e6aep=gy1cVB>;N$@j>*hlHh
zWZWu>L3HcgZ@jQnc*AK+8iO$Pr*GUVP@ay2l-+mdnY{W8`|!LIhDep3Cb476jbC3B
z+B(<Hy3W@!`kgNf=?_{QzXa98Cwcj!W7kG}_F(c`WA2?(EV1V&7}Oc_?wrNHm`bl4
zyVSZcI@?acfq^kh2bX+H%PNhGQC-;72Wzhw^v{2v9;eNqb4ngnjPk2w3QM#O#!Sjx
zQRABMecFR^_}AML>;_DGcyvFh8;|*v@mr_VJbp%w;-~YkcV9ApyRh*S25~Nv^t+I;
zSD{kAI6@cnC67!)6C@t5nM%3x_P65pU#VhDA0q#8tJ2vXg=IZZa32KKj>S60&3Bhu
z09Uou=$)Dy*$=*|O2&ypoWIE(Uw`9&{P%t`&SZLqN!R?$1QRaL#y@K(=Z-7TZD{cE
z`Kz|If0A?S<v$dd5D@a~9S`m{m}s}VRgjtFDKueRa>DBhNyA470ex0zkoJa-Ilc&h
z+Ymi>)(kPx>sMlDj8aGwWm4_xn)hfp1xJ+AJCH{16wFT<pIs%cK>C0Qe~uP`FR6t(
z{e+&$Ib(WuH%Aaztx~&U9fv$-7+#G0GKPoGvOOGs<%@{Jmt&Ika$!jnTpstu<lj3W
z7UpJ4SWMZ%J5rxB_=J<A{dRh|+xE8B+mnZ%w9kI?tG2cNq+OkP50!!~@+M7J!+-km
zuiBsgyZ=vn^4X_tmOB6LkAK*{``vH1AN=qK?fnlvZ1YQtZP6VY(^E7R6K%#z6>F~8
z9v!}w$@$c{M^WC+j;;rH7uwzBxi$q~Gmd{9PkElZ#N4sq*i54~hWK`(V1&u~(a~Ak
z*;b&xmEB&#u@{iIE5nuWj@t?dT&a4+_w)7V?fK&;39q6=66qN3m>j_u@OA7~NUqV{
z-6q-nGI&PY@YpagbSKt)@Iyh3yh)BCf@gLGyKVVQWx&o3ehS`-PgdBcA}~UhJMs!1
z(^djmd7<p8r_EH>sZ8pR#mO#5aBzGK9&kkb^qvh@@z#asw(0crcodi37cmZwcueKE
zd=}5n?q1v2+@g-ubz~BJ^!TZ>@QK?y``Nvq09Az*O&z?I?;06)9xw$Qr{RC<NI8Yy
z!gR*FEhb&XPJO^9VdqX21wjhePfw3&L-5=q!${i|$DwlKow!W{gy-V19>StH$iio7
zkN(A%83$k9Me7J7Um1Q*XgD9hkM{8C=~dAz9;5Q=h2-+8_WuZ#6W5$8PxB5B-<3gA
z{2{nP8{L-2q&@XkkZgJEkLi<y(=W<Y1&y~y+Qig!TUcCfE30?gB0Rpw0IDQhSzV^x
z;P)}|+V}6>Yae~|ttdHFXvj;t0L*q$fua!l;loGmGy0voh=hwu^bztB_J>DDZ96*-
z_S@dhUUovdJIY;=5#DveLyuka$SuR|?$*6pek)KGUfG#tn~=BtNdDcv;ha#OUD~3s
z(OpXl{kOKZBSce(o@TXmnBOQ!Q;B6ho0}VLV`IH-Z*Eeqo!oKvB0G)52ZhG;>r?MB
zIc%Gxr_#y$vE04ke892I{x}Z4W#{AkT$`L4OMg;fzq&NvR)DFyb6<e>nGf*@7Zs1*
zTkO4X=S+%^FNSiz+2+ncd-Qapef-J8_UO@ijz&|#=IVU~=I6wx>TR6%TU=Tp--R|$
zzCM%F(;0Vc7nM}f0Oy}Cs7K%5rI*fiL#M$D`*-F$;E8~WGK#*h!ld7^_@YxrML_9%
z1A8e!YA1noOCxS(&P86%VT4i9E*A`{<T1R;E#sRmkK;p+)ApytFSqoW;C`iL`lj3C
zy?w=TDN)659!=$IKeE+pj53aRRQ*<aKaQ5dFa6f>jl^{C7agQunx}Dj{7c;|qiFy~
z`)~TR)xb~<9`e?m{H%*<RGFwp5Rz}6MDo9jTS{}wT_*3qq~hxz8Zm%F&GUNYo#CR;
zrQf-uPeJnvW1r)W3nA=7-fQHJO?eI#G=XE%1^(2@x5nKwF6wcohUOZkQWw94Ls3WR
zFj1LLPNq-lp1bW*u8Jp3$by%Egtkn+;zYrrXj$$4F}@0#L;9OSd12=9rYa4T=O`bD
z5T7#H7A`7wSIQOPDi3G|<4wOC#DzQNcXlSdk<ap2O#Am~bFqcA$m3LVG%@{b#Ja0U
z;mJ-FDkm*<EVZvKEkZBnJhs#@bv&tc$yhI4;}O1&sovo<KR=&xSodmU=m%{f?J`Z3
zUeoYs?$n-|Bwfcb#xU=wa=}6PtMCqE6^60cc#h4MOEe1(0yo3Ca7g7w;2Yew?@%uK
zw2F&Tb{sX2T?`DMX;F|!m(wP;2mQdYSN;x|O8b@jm``xq{CHCt;3rL{UbLj+k;?Z~
z#+Lb6#traII_Ux(+g3SA=nbK5`^cMasqlf%<sF=au@Xn@yW*34j0omVD0+0tl4A|Q
zaqH;O#UAghOhaXz<;nf=&_H?I+^xU6AAZB#O4;#BA5objeO1vVKe7YQQ40g0ZxqX)
z@DGIbbOF;%AE55(_g!De7#=#KT&7?Z@Ps}H!_n}4F5vh0!R5tS@WK1bF2GlnP^BET
z?}__PbsY1i-CaLB%lKRi8oc+<#S$u2bDSbL@0d7~F$mslvhJtMjy~|ig-|o#m2D!w
zzX@-eg%gZrDn9L3(oXT!g+$`7{QeX?-8*-e+alwayp{M6B`M_+r(I|!oQVuPfCcy=
zPioof4mk0$>g6uGEEWN-j`fzy9cRu<RFamA%>{wQ%QzmnK;J%Bv=;gR%xh7$wAlsf
z^2oR8r`ENO_PoV-baAHqxOh5#{7+l;Z(ZtLrmW@Ghl@)WChwc}FL-O-X6NJeZ%lo^
zB|PTKzI*dF_nYR6!o=@OlK3ilyv=;wxPO_vzQ~AApJ$@@`a3GVz?leiy8P?aAw}2e
z^W}dRD8F&yPl^!-!qO@HG_h<y#3O0PRc0cuGB_peM8KGy6d+;xbmNFqeaBmWO)NM3
zJLLB(-U*_=U&*Yq8@7y26}vMDa-$CARNwjSL*ke849lQ#)1BXnU+^d|+=YkXlWwJT
z9qamiv&?yQKY3wZs~=zecZ_l4QsMQzHl5pz;F^R@Je~n8ujAI2-w9F1IvBj7Bu)Kg
zCC_1ej$o+Ic!xmBNI~Qkwrs6$wmr88Z|}C_-Gl4^d##A%AT(t<J*VBX!rh{sEv_oS
z6y(4Z6hj74lpr$L_^r|IY7dEJj0MKef!BDM98xA%*vN=D*h_dFY%^&j-}n=@j!~=u
z+rcaXE135y21D<MaFUUgY`lvBa<aqYU?L+F=HQ9})Rl%XrTG2DO+mhXc?bRuws!7v
z4Z|W6prD7rL-+_5*o_+zy9eVS7^wNf1H05FT}{Y<oA#t(Gl59GJ@2>_E;7?H0K%f*
zT5UeU&->Q0TM4E}p`4RRxUf96>&r=*E8%7AO7SDpaRpQAZZC`G)@w4x2BQe$=LFBm
zrh}yTmY)z32j_0}An9ZUF7>VnR}I?Ykt?!l;y_yiBiq&qp*SJaQiHF|9_gp-GS@ZH
z5YGJsLk82MJxVx48A(5koOx1TC+9UOEXL6KhvCIV6KIoGO3vFbPKJboI9H6S@TpZ<
zLb)bB#Wef)?Y4B&6*mi~3;)w6#D}3;A*6y)W4qWNrp6T#O;+8-9X*MoMUa|bKJr`n
zl^sJcErT$LR`$*%Ka~!TRqfn$i_jaF>p0yXpB1L;#aB8`+z2N!`ej(HH`3OjQnf(7
zb+15%(V94MyEgiD5Qs_IUAm!xT1+qPpw(0jWPm=L|Nc}z(<{;ZO8`GiE1?u^m7gCz
zG2QpILu>;#|GTuVp;tcx#Q{&{tDBFua!Kj)&o}OsuBPJnj$6khRC;B!2}Ruu9b1>c
zb*K&&7(*4xbmbxJ%{=-xHk_uyNbRd`8vP8DcyYRRk4M=GO=7R_-c(u@t5B(rSEet2
zOlOZ<?35z*xCZVOgr7dq&ZM5fOX|qW??FlOs<b+nxb@jo&(=-}bxVMPbN-X5Ph~2M
z67fbM`VjS1Ch5QC(xqtugDV<n$RNz?q3fG=#<A+7lll8LZ$HaJ*ba)sPr<!w_t+A9
zU{6?mlV0ERuIryC_2j>Tb%ywS*WIFJDs7KC{L^nt@%116&;G$r=Ev`}g^7_iH{lB1
zIu6Gbs#(;-7~s)^?goOOM{waWVNQCTG`jo69XgI53NIXM6x2RpQu^7W$60wQf8fFq
zm0m7jafSbtV;M;*Dk`WmJ_$#RW4*N8385=rT@mRjdIhmgUXPAl_3YXHr#X7%=-7J*
zYT~bODXTfp6hhYua*rKR3E@3ZrTD6GPXUJ@&yH{Ms@2+#ITa?4?F6lKOm#f7zVd9|
zX)rhEYFv-Xn*~-c+Q#~mw*K^Cd-nLV_UzH8ZR5#jZS(nKc*DnS8@~<gyW1O-<4}o6
z1JvMvg0ri3ba;?Ea<-p8jS%U_zxOxV@BXddZ>#t2wW+zeHas?w^4cCz009d)l{Rww
zRl7LeZ%;n^b=!UZs2%P+=UH#JUY@sinee}Rcd3n0w@cv>MTvs>qqcW+(oQa3w9|`M
zdAx7r@W>V9Cv9_cH^;+xtdI3~%no*n<G@FSfWj<=RyLx$N!Fh{ZHEW@;h6v~{j3Bi
zj{g~NZEM`cad{+#4z9pgKq&tquj04qs+7_#0u8W2p@%%OJb}VO+fu<!j@SWK*GGhP
zchJDA<Q~@WvG8`Zxx9pe36*Lq%S&x#WvK*ll+pIKiq=zQgvvkTTVI9Zy#Ree!475E
z+ud!?pFE+iFWV7td$zHc#Y+lx6m<D0VA%)88(Vu(GPojNcvyb1{L((fW0`}5@>9GN
zdI7lzF2&v4hokU!Z?7GYuXhRCH($N<SUK-wQNiI+Lmv6;%JbrxveKAx$S=7%yLjB1
zlZ2llR}I#{ESdY!wX?>ylBOA6{|Lxz*A5?i>m|=--;md=wD6~oXB^Pq#PTCNtH+ly
z3^*dMYJZ167QaHM@e6+!ycmbRjAI|;O*`{;C4Y`O<n4XQw&}gQt31nXc~KZ%wP#Nr
z0{4R)e|rDHy$G|szeKziFXV@VBj<H=mjc8+=!$oId2HeSfk)zbPm4nEQQC_>a^QU;
zDuF$g?<_p5?PWQI2PhuFTI~|4{whqSI%Tex^(#<w4(L&yvntDZPcavWV4kYG;$9qA
z*crZAXwZk8KNN2a4ET0;mh%*qpejVg1?MgxYtby=pWOz`0fngaRT`32^UO^m0HkbV
z6X5Q`Tw7jV$Z?<Utd%x-G^;xj6xLo`g3sXa;ps&?e?i;ZZ-;RQKF^r@e)`$t_UO^m
z_Vnp`j?CKJ*osj0*gMY1FZ_Mx5VaUnh1ufDN?WAegon6hoy3nEk4N3z1>x=v?-aDH
z#7ki;{h0uMSulv3^Dl+`E_SegxU0w=S@Z(BhRUj%k3>EpT<nL;lils%cvUn=TvO04
zO{xWSSDEWLZ*i`ol0jt`b+D01UH3xlO4D(~hsfzOX*bKvW8cw81Je(U-~DEu)#Xv`
z?a`+-KhF-CF0Jq`q@}y#RD|%Pt6M);GM!Q`7d?^IMgZHT7ak>Pn5^knn}lAIf%*E@
zST2I3KEf>t^4-_1ysAy?t9<cT#{^1VQBE44KhsK@{MFpk{^!DkSAo4_zNxyn#J*si
zlruOFc9F;3wYzO?Z7uEMj!G4J;*o-N$C)VL%!@iWexyzm={8Tz4JjLCuo|Y_pI+kJ
z-87X8Sbt$+T$MKBn)T}8YNZV$uM=9A`r39?uj=!_lsp`p#BJMOI#WjpdcR!V;WR!v
z9(px3K1}_nBkAkDwY}dqw|3jHcLJ$EAWe6*cu$@?-n~!C#Wlh<3pwZuSwKP^bBYgT
zawoLPbCn?d0wDP!mEGC#0(}x(j)~q|<St|t7dbMTd>m^t*Cs8iE#2@)NEPGnJXnL?
zE<nplIUeP+_=sJ2$*)D`01SdFJdR5)U=kN?Cl%kqT-ew)Q8oZ8;pGBk;|OfS>sR<y
z=@ot#{@8Xs4fkXo(^MI4Ka!3p$GW-8EVxR3jyK|n?c`|axLmx7I9I%3`Z~q1KJ)X7
zZEk)pavS>+(HuCFwtR|o*hLR2>RlAAlFHq~r?jUF$=snU?8Rr>!yUXwoibG3)%?i3
zoxu|vLNhnFciZD<o9*$_jrNT9Q)%j!N4Tocyr3;F;29aC_7CANc=lajvz^5mjvel%
z@&p8T&C5Hvs8~MOu_6lr=m^$Zm<ZdV3mqM!Uxa1=55I*;WJI<FX)3dF2etUP4Dai)
zgHzzbEqGv0ZEzu$^|_>d<oz9M(m$x8Z;u0(C(177J$EvD9PL5slKBE{=bXZ0c^%VN
zRu;41%_9x7(-Zm-{8sTtyuCdamRyxWj((zYN=2IY$sNE~t2~uI@J`Cm8E{@)bhp~d
z>QY<6Z)REr9&H=EaD3?MOIYFK0(bA|+N1Ao?jL0=bna#!u)Y)I>+#srQ=WF=BKkWy
zy}~?Z>02Jpm@)hgxEA^*&dUcor?B5Z%bZ)NpdP-}F3xBt#<j>1>`0<p_b&Pko$J0e
zH{FWfmg&RfebfF0Z_V56e7ydRsqeRh$9&oMP4h*5nlFv}#nHbs{Vy_KH|}31uP-uh
zOScqXfB1KP@(O<g<!Gb&h*tiV)SS+2bWtS9<fFHDjSJB$At{gJK7S^YbVSHlyA(SA
z{1HdLxagHIewADQiZU&qk|6CmVl!W@-CzClL%M@X=Ph&5`E_aODgo<160Cj_z8wAc
zr=EJrpy3^N!<9Fra1f^4v>3*J@+?Pw?4afLn$fL`Z!PoN2k8rp=^3*4>1-YHT~EPN
zr;3x-i=UsJz4MP7@q0yF<?G|Cg8@&KzS6F*3R5Y?Qg-54RM~)Y2D(=eMhAQ-EQ5(h
zOu*#xoH5`WIWan_?fgkTrgHFKc#cnuhCzs+h)Jw(84!tb2rsx9Cfc`;D@0{3_)C&u
z3do+pO<Z%*FGnW><^Z1K6mc6Ho5;!<ncp5x**ViO5r*&ca6%a7K8%Bm+t7e{FuZ~*
zC)x@GG%_ACtsd3kM6fT@*D<zL>S|jFoUT10DAdT@_;Qf$naeP@dZrV1_6~PrzhOGC
z8)w`BXkj6OyPEt6Q{gTAOAr&roABA?1RP|ngzFGlWpV{9oQNria2xhZ>aB1gj5%Rl
z=`jlIt<M$lh0hBZ8H%2KCTF_KKhJ;{`71!oDrd@+J65Q-f)`h8$*@@`%jF=PiKFMq
zlTUCQ&qA@fWZW(EvfhGD#$2Y`yAH&stcDRrGC1IoxLgxSnYxLTIhhexEDKx+Yb?ug
z#-kBW!sNp^WfDu24C3UkJ(C0Rkb=cd7#S|@wrD?5v^*Xgs=w~(%Uv83lHrg!vCQJQ
zxTc*-<DaakCo%Isl+{sQ6kF|=kO@EN*+xxH*sUl3ML+a65o730YFfLeVpJaa)<M_g
z(WmHV*DpI$ar(D$wbgeQA^z?P=rc?WO_#^<LaJ3b&CiX*_0K!B)5kLV_P^Y&)9s_y
zcmDD&Z~e5JRN&#Sad9a3f=BE;8vQE`%IG@*x_5=>-eoEe{jD!?Qts5Yi)Vm}uRZSZ
z(>Xu&##Ed24-Nwul)b`N`D)D7gI^m?zXUU!&e9IU`zlQFqteJn?45OB%DW2}TNf@K
zucu?(Ek|{&M2rbPLcl>-Lss+z2_K+Oimko-He>=BL;j=e`UDi(_)ZYv5&zyTzJ4T}
z^^aRVe34lpoVHlY|Mj$E`m_mINmy_5xQ-LUJ9YFYZq@GPZMY=W)uVg&rkoPh^3+#i
z$7cO2tNa8m3E>a6lvy|GnQGmP7h7+-uf*@`%3t^0hcw-?_rJb-{~O`D_e~=dst%n$
zKjCjM9mvb@wtRp34et2*pZ~}I=qF>0Ma;FBh`HlM;h!t&Yw>|XRt1c?3kD#lu(Hf5
z$6Sn4#qW8)5+*6UaNNmCP9}(sr?!trce#V%;>t7q>nNO@dGEL)t~2Ji3)^ut_Z{Hw
zxZn{jDgeCK;(&?&4inwI19zY)Fc1$2NBWLg;b-_3j*de)f|#*eMS!iOzk*E#Mj3yp
zqvMV38(|z{k_$}afpl|xbA0x=Bq!b8BQ`%j%Y=8lc}JNFug4ERX+Qh3KWU%*`WJ2e
z$)mQnz0nSLw%X~@evT^itp1ns2v=vR&k)Zm#(7t`){%0TOqeIz{d;%Y@BNM6Yk%}V
z{G;~6-}}9`eD7Wayz&MLsTH;Z@*D|s2c~3%cAI9xJjSp$HFi74=*>@!XE)q5^_$@B
z{VKou)vw#rr%!1|kM^6*J#Du}C*Whp+VS~|w!K$^e%nVOgm4yLUB#MvsU|1FBY6am
z3%0x;NCB99$=+`5OmUZq3sOSw<UzxKC>*3Mge&8?3Mp|YyhY$G3k?s@!Ff~x!i1~%
zN8mHvQF0aKl!7OD26;B)djY?Kew85J(=|Fe9{iF&Ea7AcQY(M*BW>%VfMa!arFrL>
z$^jK#;@RxvXpWe1N0#L&WyKJDP2rzMrS0!-x6P+dniri<j!y+|kHyEMi`=ClAL!WV
zQEV>Y@t(A?Nf$QF0t@;B4JWVb?m(4W;+6{Hjm@3*@X@2TvAz!9R|2VH@A-0v408zu
z`Chhu!1u|?ap2nvJxifbbue95Nvk|4#S6T<3Ow!b?h5dWr!B#(a4hP|FKKmPOFyzd
z*)|Hj6av|XB81+d^YW2|Gkgj0ZNGwdDZl8`Idy^1Nl(GCvF(-i*K*4Pt|+KBtdu$B
zFdy>BVlH4e>fL6*JiGbC4|(Q}XIB8{jy!h)-eI0J;SNmi3gKHp*2>yyTU)!Ebm|T>
zl|SAaBTV;7I2n8|VW%tcJu1oFt=_Np4BYk@LzO^Fi%aPbQTl^lFJ7EyoV6S|77O~2
zhkiCeKlBLD#l^*T@9tWA@ZbUQ7ebrt$MPi3{cK13hh>WJlsTj;{8jYV9J25}ynELU
zR3z<e3yPmAcT_k#e-oxf|1i8m?<jj6YfpZ(lc$bE*^}eIpz)O7d7^^vlq++{Ix#}L
zQLv~WcJJ_{T~H)?IW+ubb{>5E>tDBD{Nh*O@Z+|*u^l{B0Vm#6|1(ePS#zA6c0!+a
z)&I==T!hOzySwdhf1kG4YX`iYV>#av*K2<0+=6}yOnj`f^Y1Lu(+EGpJJ$%U2Mz*C
z_|=@lFY0Dm;<WX4?#;YKI7GSHeOqo#<}aZG3d*Z4{XE+Cl0FRPsT$lgp+CvH_><Iu
zw=lJznP$ylJ7tRHOPq3dti0;lJ~R@%s5;eRjLK8tsbh1|SIg(OWpe&#xIPb$g0~&*
zpSqi0aD;e$88rHtu6|hz%iA=nKKh7D&QW!*uobt`N;#-wC<J&Xo|&)hS9HKs0vX~L
zT+G9IWzi^q#!ol{@Rabm+C$ndP4Spoc@dA7!zpbl?ux@Mijg)r9?Yl=0zUSyv=!+K
zN8w{VRICXzGqem*kl<%OQTZues!*{EhE3fumRVScCxO*~c0yNaFd-9MB~6WUQg`l}
zj=V}1Jm95rL*>Z$=<T-T{hN#P&D~Psn2HVin|FlmxWkm^l(E7waz1xZs?6|6Xy_1Q
z<0Q0Ze%76@?%tYh3-qfQ#slH1@;EzxsE^7o#}pO(=HZDoDni_u?(uS2R6ri?gkD)*
zZSxBYj5Xdh=&s}mDoNvjf6~{rm4&vlG*3R$b)>BDr|hAx<Ry*>@7j6qBI1kSh>LdY
zyYAZC+1ZYw$P;tKC1qms(1oR?r7SKI#>&N1#<^omxkm6&B;h4aL}nFuyBJ#<0WPWd
z5ACyEsysceYSKoug*Lm1G{N>M+O5pY@uTjE%pwFAAb=ayhbcqwr4(+~t&TY@8th3(
zDww>B%yA<-f1#f)rcmiA_ISrv?j#4N#H*se@RN*hF66Lnat{`?6f%qZWZ*IB?l{^)
z9?5U{Ea5GUt~=jU(!UD-y|ldC)>c=;ugM2XUqe?NBdj?;#utYJkl?aoOu@*pqrg@D
zH1O={w_`AUTWohXp$os{J*H_}cR`5@rY&r3FUM8OsSL}pZ3o<bw!W2JTPnQV0c;&q
z#=2N2u!FC3hnhIA;>y0M63aSRUKLH^Y7}^sufGr2u{Zo+haa}R^-uvNkLMUCy&nb6
z*GCL?_odR_wrv*rJ8uA<@>sq-3UQk@dA_yRHg@*g`p!X)F?D_+n2hUOLIu6Yw2F83
zMU}`)i!P3%PiFCpCsxo`>GLZ~b5R=JU0q0j%2AdK7T#%i0bZ#L7si1<^+{XVB6(N4
z^ya2po^l?9$@`}L3*MTy+4*?=8&ltJ36FX6yU^`*x8A(Xy*FRvAHOe-{-x=Ek@>oD
z|1x=fk$GFX*E_!WNFe&K;I0TAQh-SjmJ~(}I{&o*rNU|wJz?`1B+Q5Zz1s~>ZIyQK
zUQgV*?Bd)T-F4H-lktToZ~yuT$+!ab`RVV!W>8LFoxMU>rv9%y<MUb?6(;YbQGb<2
zh4R0%;S4jAckf)5K6s^3-{r)2{2h%3juoc&d(EY{b*B61eF`Q8>vD=M{uQQk?!L;U
zOz&RM#MXIMjS3a=+ZD>ZRG50^vv+%CM1@iHc*8x2+=1JQU7dN%AP?bw0ii6D$o0lZ
z#B+2B48};b{Y>l_AbdNJIH-<?Au9&c71x5xiHvsfM~ZHFB>L3`wOYk2gof^Ji<hxC
zjWGB!kHFl4r-N9BO4glkKS<ZXU53rU&q2^h!xe54ad&^!#JpDiC`gi_%F!zfSUm&%
zbxh^$+w2SK8m)fE%{OT~AjLz_i$58lD~@ou?HSV~G`MTba0cnjgv=9@&?pIsTS?R^
zB_};@H5R@l9B?Hhupv-(Tfs;<_<0$xt9mYYUop881~L)~Xv`|-O2hmK&iHk4VHto~
z4UA!qPJtug#Q_-}%jBfU@G6;H&E^D1rn=fzK}098AT1zLprE|gxnH59qXV!40~tB-
z$oMj(*3+=AP|D;8h9L^FF^@l}NsM){-r|}Q>j(}LwpR3pF|aLof|E`vh%2D0y<#0O
z55~Q+%!DOGmq$&Ma12w_x4YP5y@ZpKg<;BVc$sIzIQh?!l$OJX@R?i?F{{L<Wg0q0
z&OGu?anjaW1}^&f5f7d0Q}x|NvoHygp!}qZ;1aaz!*nKLc&C<$RBq|7rtRM#FP=f#
z`jkl+-@OB01)0vzuzlPF3>sq!mSvCgjbGo1&CArfK$=)14j6!8khx_}g56X9c@n1k
z?%~m=UHQe@y}E$i%QglT%DpoMoz8Um`+Fm`m^ahvi<%V3egJp>cn5a*2CJ-#{&6Zx
ziG~t7F2*-(zLGePxY=3zc-=?Y;#VL34pJQW7i{Y|wel;E?%}8M?cDvo@nxhwXyLV{
z>RX@X*87-Pj#1FyXdCM1Z(Mcb6}-&n&H9!-KKxX<eBv?)lehTMxesu-3)ZKtn4pJ8
z{*B6C;KAdOo>iyv&!g$ws=Q^-U+3>Dew}-tM{Ianvg)J0ap>b|?Gmw!{%HpIPh0y+
zO0iVR`S0STuCe>*?~@)7m;U$KTz?y?FKZ9m{F<K6b$0FFeS>#={ZId+|MVy49>?n#
z1MT&Q9qXu2arn*{<EdjJG@NmRDSsQs-N~-wgsYPk05~yK$ucglfxG10#XU~I6=o>B
za?Htj{Y-LQ<@xCG)8>&rj#I+kv82Yk+ijRM;BGPrG_ArFS3WbIQr3RlP&wyHVpmG4
z4Ddc0-<Hpnm)=FAfT%y}N9B@2lsvOjEtjHG#$k8|$GBIoYK5icRnR>(K9-~HJaW(-
zH5X?m?O<=aJ$v$5d-&<E+v86^j<W01kAK-Vo<43DCr9n&<wYjc@4fqO`{=_DBlPmV
z7Zn&3yyImqqd5ES_rBGB{G%W6{ILDqfA8<KAO7e^?Y;LuX!jmG2p<ulg5{J~1QuhI
z>($kH+u3;5c05|}(ZlweU;VoM=HuUF%$=GT2W?+Qz;!`BW7Bi+tasXp(rS5=QI8rL
zr*pdM+WS0SGI>ABogq1f2{<c!Q{c6*v=BJus39KV=Qz8-1U|>cFxDO(9ki{@&318o
z9Q<``Rv7F3K-06c@U^b|y}=9c$XnTV=|{A)xG&EtA5zK#a4vr@<0k%;AR8V8zSEsF
z?)DPzviOXirEo(5hyo3TN=vKD?e5xI^GL|t8#FtOeW}gPyOZn|@HuX0Cx?OQ()?7r
zw>TYzp~vD(jNWOpz`)hI6X1mkVIKh8yz9b+LhjNTx;5O6FRr54bp^M4N*khmm?(I3
zl|oQ=BzZiNNAKNRTW&Mtu`uiEU-}kEtuWAgCmugpZ_hV(+VPn@w!%B=N?H@N`{bCE
z^kw>rM>$C*6#V7^)(QqAbfV0o;JXds9Zo89qU?g_SD6zbp|}Ss^uj|GA@Wkv98DA}
zwo97DcIj8fTQA}yTG2os-Ga|udxqzP2FSlD#Fo!fuqS`2sWgn6Pu_O!5*&y7?9#OW
z10@A<Jz7fP>I6KV#~0e3v-As^3SRt(@Go)<mwdFx9Jy*9FanRMHa<B=v4??&N_clg
zxjVoeMJhA*phbI8L`C^WN}v&CsJWGRtKf2DYcqYsvahZzwRi5Xv<GYS5!%xG;HGCL
zBP<^0QAs>IH{I^uU1{&WbFY2-!}r>UJPN*g0etTKBc4hp$7u7702jPbfbQ{RQ_wG0
zysKcAx0Qz%)*h*TdVJVUkGf+nDWiQ~eqMk1auN;PlP>N)eye2c{d#3%1i1usH^T1j
zL0jM0YCF{Rz#WH2?nWqqs>(C_wSCMZS>5sY>8HPGPaZvr@KV8MDb_lLmpDL~eXDSv
z^;or)<+i-C+Lo4Az}d-mNIo9X=*|Tf^?7I58TInMEtSnGBRsCkIx1j1KX>7!JKT5!
z2a(yfqvF9ih1KGp7(`PA7FAajKa4~8yO1MB;B>Fho22FL1ZlMjw3_#d7t%TvcjBg9
zM1=MJIp4v5XhYh-Z^tMe{XFK%;3*H`G7rKHFEPCG<)f}tC}q3oZ+*pw_!BSmE$N$A
z<tN=sc%JHG*Zt8sD^W#c+$q2Pk|*;nV|H(iWeF&JTdoc#q?0LY`6iBZAV#M+Ti(Jq
z%R@zo+tCd0z<MhzH?H_$`Ktcfoeh+cE}5QnFjD5a;x#y7S)5lZbC3plZ0z14xFJu1
zpKx)3O$pacOZYL?(1z}$aLjihP3||6{{V07E6%&cL6t|AC--<0N5pYApo)=`(>h*G
zn5!&yUL6<|NK!BBZ+%}eo|?a&!3Q4O&9Ybr7ihSMrH54%9F$4rhWx0zQoW0Fae*-w
zdLk@xbo1eHcEP9!*`vR>dq$Z~?ebDqq%vs)`p>uv{hwhB_4v7W?kz{)?gAz2Ck|Uz
z`5R>r)}vRRSymSSoZ;@m#D0g)6ts3$ML+%5zT<e%E4DoLbbhYx4iihnbd?Vt>$<bO
zT`~;ELE8tO&vDytaZH+Ny28rcQl<2lm-Hl=0<Jr9viplV%Ga!TymJ<eC?^qr#3MtB
z?<HH1eoC{$CxS!vrNVphw%RIo(F}X~5^3~#3Ul#RSfy{70A-X8oj{vh1a7}onVUHL
zOJi!GdC`_A4~cI$%P4NkKSwDCOrzu_uDCZh4{zeQ0ng(?Fc-vl<g5IZaGk{cop<ju
zhP>bI-*e~FQWj6ly4c6DpSF{SRmtmk;;yJg;(K44e3N|1{$4FAFkclZ0!_J)I3#@z
zJaMy3c0cPRfP{^1^5gauk6^5m4pgd6r+uVRE&`YKy4x_PD=@~oTU|QjVlm^JsC^*a
zJ1}*(^0o4?WQw#6V@t|JSo^m-Blk&jA0AcQKRAT{h6adBo&vDEI2XQ0xf)4U-aMu)
zPJ0JD1+hZFhcRe+vg)M+^%jrizoM9Qhc|f~cGtPMLqB(&(dN!wRLZ$?Z|{IUNgnQ6
zlTW*X(bzvcZ94~)%~KPwyHoFz&mLhv$b880etKpibpGj+NA2TZ{<LjBe**vcs(t%|
zciMZj?Ig^M@Rl#Jy)L4V?vJk)n|jz?!?6F>*p5DBukgKn@V;s9%-6N^@%lHWzTXla
z^X9ka^W7}}^*Hf)=56i+L*ssFn9ng^CC=AP_b*b8uVUVou8OarfBeV)WqPtJm?X>o
zi}0WafVDGpKn59IfvroD^M@zC64#1aauCzgy@D73OQn;v%5@+!$%~&#BR>_Q^XvWl
z73rqn+jJ9-QaAxjz+|JV&!Dt11&8ah#qg25`5WZjrQNx-mjuw3oa;iQJdmJ-uOG`)
zFN>L70R9;?VQ#z)PzB9rZ~BJ18AUS)mGp(RZJ|3q#7X(_3w&Y<9)1e;C?*p5b$DpH
zEUQSGE}aVRQy%@7Wt1dV_zIQKCd5C(MmvGY-h=TzdwB)%J*_j<8RXBR^g4xj?zeLY
zW+ft&#5M<58BrJsCNL|jYcRx2ULn>F<gV0oRfbHA0v9JvnP~FoARK`ZgMt$d?N&_1
zgoCYlMCpQCRy>8l_j1u`+-tzjC8G=`4pM$=dM4Poy|T+krH%?NnQ56LkF@iciG2vT
zgOr5c!8BX7nGinrJ~Lq8wqlhse_<nt0a@QvU&+`m!Wo#T<WNwjV$U+0ULH*~MO(PP
z+TQ!_ciQaYQX6JKcIT7#Kq%M|-VqWLKlgv(H-cMW;3Nx}I6y~m$H3?r$x-$JldNWE
z5@c&TIg&}Yt(`#ScmXDf7pE{gz)pJdk~G~db&IxMTw83*ch}k!4A8IxAAe3Vy>wZC
zShispwM-;nrcxRH>r8Mj6Qv%(X%$n1rJco@81aPi<R}F2PWXDnix=(699mulTH;kE
zZA4Mn=k7L5F^eTIaUha8xgy`(C&9NcySf4wx+4ozI;lKVaGE0qfR9a-$vuBzyvQdz
z8>x>?=TR-CAgn%ZIn39U*W#ywCRb4PWtG7a*KKc)5_Of6U4L!)UVHc6yD*AO%!i08
zEu)VVE$Q$?0ALDQq^R3X!*}muegSm%D);&bJ<zR7uY&Pc&J`>#9Z0jH#oZ_}+Bzh>
z^b1F+#7&bR7=@Ujr#1?XeO>ZSsCr-92wg!cUrdVWw?Ca*0`TEyPdS|sSssnv+5*O|
z0{QFR`RVItMS_ER#Lcoor#tvq&x9j@X~d1E&$rx!pEfe=f0HIQ-hsJ>w56x7I(Q^;
z-j$wt=w`Tjr+8RnDrg5vLtE~0FCR^n(PZ^Y7+m`@CRF$1t45>B!QaHeQlZPpYgC!F
zSIZfq#xC)p@=IB}_!{9+4bVY~0EW*`8d-PaK*tH=XHq1wK{|f&ddW}s8Q69y#@#-W
zw2eoIsx$ADwOsOV8vGQ<aUgQsh#r^3@6*y9|3daYP|S5){R$59At^^Cp65F4YrADC
zhx)OE^-(1)qb-db<I6fHk-W+y@W;ssvSrr1B!Am9{}miV+WH$rbkv2P2#?Dbx3ZTh
zIP|5vX)E{MmpEinbf|a3hV}paFaCK&_y)}X_W$^=UU^}-e1I!9=ck#7L$e+8>VBd+
zu0<h<6LVKw=V%*Ilfrt8-sH$on;d2g)}1!Ld{H}s6`;r~c=U{Ir=UMax4|p9OJIF-
ztKGZ5*6uDZwY9a?+#liY6?ap3SC3#}2w+SZYL`40FBIxCA!clGBI<-ai$54MoHUh^
z#TDK1N5g>E_B6iZDidwS3@5u6l*b+Hj_+k?L&8L`XnWWQj7ctZQ6RLrUKi|NU7lwZ
z?K=<dXELk%FMj&dwza*%*j0R(kCS~BUJ8|j;}vC+zff?ztk5^BPbb?0>b|n#T|~8O
z$vZndqQ<&d78k}$PQXv#_Q`KPj<8~4ay(_7o|tTBjOE@Pa=~~$2kho%r^A;VF)kic
zwOb>!mAoHy)?LA$`~}I;Ccu+&2}dtvSFq=OY{Jbxq0&UYwTGqSy+SC3+Fl6mF-j^h
zUHqj`QK6GyQ_$^hFX1S!qyS9VL>NW64O4=b;#E}ODByCX@VG09fvbF+ZJ)(kq$BTT
zuTl`9(#s>%glR8;QaR#XsVYB~mKWPNFtR_Hw*nXyXqIcp9bWJ$-dohb0gs<qg7=yl
zg};V|Is+($z#Z^N9<BuM2M34k;iD&QbL)V1tUOQ7F0<3iKI>gGd8Vd9H&m8*?~0YZ
zO8MZe+>JEc+K+$dqxSBDmF(!ru~xgs?bn|?ZBN%XNNXrUC0Bzh#JacUaZ-13OwyyL
zTWw==Hw&h`S596{p^E7^TM*}59CUHzYVl*r;IT%OADDaekNu)}RRzT2dC@&-WrRdR
ztnUNyx7-Q=nSX$f#iQxQclgl~I_J`S@XK&29mEm+1cYfA$MzSNDk+Rpc_^GLnrptf
z_&w<YB;uP!?kD2i-z%b0Q@SuZcDp_L?Bn*UKmXJ4j6eMG@3r6i{XZc7ND0QmYf@h7
zow<QRLg@kJusyA_b0K#Zoq(g_nL94T!-7%alzmyA*%kld+tlP#d+=bjeemwxwz@PM
z`mnPLoX@Y?$oND%0}d+o#os#~JxRVEjd?^KPoHmP)xY&rfT-ZmyrLupwk1GSf#2PU
zC@)QV{XJf;BKOX{<3svfjud77VL2(ge7~j_<Vp|Y27VAkGz<*nt<xspS(F?0RpR?L
zjDme(5M?p+OPum?u5oy{7eSsvO^=z{raTHg%~O8f`KWD0GJ&6MsM6Q8<`kCB%x2DO
zT3!v{ym6cQI4>BX&I<5dIM54ceKL~rpE83lnL~8&vfSPqld|(C?U<ciNF5YXAG+I?
z_IF-b`05<EcKRLTpZQ5KKWS%i#d7!a;mij!7g2ayN-u$AU#3L#JC*sD(D$0d2vtq$
zj*)JruNVP0X`k@1E)l-+enCI*EeByaN@-a;kj&4xx@BzZ>TNNiU?i=`2UHA0)3R8W
zG)j4Q7C33_!!ES#`=$QoX&Fo}N^#x*IC=JqBqJZ798t*HCn%-$5iWhdR{^0htPN#3
z3m2U8ImWm<T$@L`)~+5+;BCEm-?S^Nj{=%>ZEuxdD!nI|*NrlkMG*>p^U4Qz?Y{rs
zdr=@buF->l?NIun_m0!iBj|RVcD&kl2Nh%1G5u3G$ZITlXBhQy91|b+=vUTbW_m7)
z2*-|#3(}^V?ktVs7kDZsQR!@(T37p3T@g_y<QOOSh^DSJ*3Hh&XY994#l7@(@(?{#
z;D7k;Qv3bi`yO-Ih0u?!y_5FwXHVOo{_NN7*PlLyrXK~5SC<x`Klj?bRgVaTzo5U|
z0*_UAEG{mE21^?p8(c)<;$)A@bBDZ5A#HTWymWG6Y%<5lx!A*nAj))lnSt%=U4rh8
zbO)^WNx9=C$|-PbZhE|Z_gnA43*E#2Do6TmQ$}~ec=x99BWDulgp29X&Ymb>+qnZ)
zxOoqvZR1XJX|{`5gpD$Q!jD_<2BYv@i=qAAKl=39dKC2Di*<Z*kdt__Glz274^3CR
zOFE3D!*^WV>$u+IXs?V_CRTTXMOi{0u)GwsYdgb>$KWDvS&Sl`nVZj8Fg!9wJjXJX
zfhAuvtTLp;>B?h2vR=~sY4FGeS@It4n7g~W65jekzE(LJWpL-Rec8V2ZZ9Vsy@KeH
z{G9W-<6NAFY{gi9cZ#XBalB9oWm~u#EqAX$*WI1w4!}x_a*>X2<w0bZ_R&~x`+;!G
zl^oPrzQo;y!p*vd?*Q+-o7bZ<zw_a{(C$<C<mYW19)4`3cDJv>FHgbCP;78fc@bmD
zF?`(qaoc2m;UWtaN^)P`-8qsSTgoyin=piiIN!?HLfhHS&L`Z_cyDb9cv3I=caHCd
zPxCYY`Fh7rm8tG@bm9L4@ceha{UN;fc-tkM#}=NPUP2T0v-s=u#6?G!?Toa83v^$h
z;DWbl=0sVvPFmr2Xb16Y8n~=1%t2=+;LDWpT{4d?MW;gEsj-oEw7=6Hd*m`SLq2_J
zWvzYhyWayoW7*ZVva-@X`u0bmU7PkF%XUa#+}@1x3oK1uX_2(cb^G++8e!v8_6pzI
z2k)Eq&b-au({0U7chh+78&mndB|PTM@7~<ZLsRM0INaCeP4l+2ZW_m^uWP<ix?ds8
zP4i~`{;JH|VDnG^$=@Fu`p5t9zofg6aYq;|Lj-%oiJ(q*dXxjY#fvb5jyzEMWQ^7i
z=A^5DUnL;Kj$4;>nYi)ee-h+Jx50M-qG(Iy8Y}LGDPwtb`kb_7(n!b<5i;pkUdgJ@
zuaDhXuhZ5ZpY9Wnz=L-$gww8@cw^YtjUjLocCYzYUazN(EAT7E(r}3oV|mJz-wqxX
z)8BII%WOUdcr9(73SFiIS@m@@?X(#mpNp(rls?~tFnHi#!~0>ocj@4E0Wmy0KW+Qx
z$L+uYgbBz2%>OwP0gnJY+&gHe`wU!!Gl{IgO3I}1veKJ|`MC$9AVVd?<zTY9=*rqz
zm<_k?GAP5uW_5&U2=UL3AQ*_6oXVWa^up{qxZ|HUX{iM9$pFhh7d$3Lga^h<e9IzZ
zP)xZ3s$4S6Ag1EV!BRnu%$dxj44R4!Cp-m{p$ssd6)m&q5jX4Wn;C4=kl&aQuD%L4
zgj041<vxaDIQt+}Ic8#hrrrD2`|YmG^4tuQWEdR6<nAs8sjk)2M{Bo}ij~uL%H%E+
zUl}U~MvpkWpuSG#oh+%mkx7q`Kt_+ZFjjDMLRf`A7Ywp5U4p7YfOH)=ZJwFsxwZ)7
zGfmlKtUYGI9a$>NUeGok^<?=X6tJ$8E$IX6FaYH1#K_5+s~r{aM99SeD+BB*O^-3m
zDo5hjmX^tZU0m`$I42tlA{>;Rm?&q@QB$NPOdaSIrVN^wFowGVXI!myo{0g>8s!qF
zWX>Ykh@y`&2bS2oev$q`I4AbP%(9AuNglWvUS*sjOJQL?68cOqXdvT<Uh$T}0h8O(
z{Bpa$dcVy(@ukc|FG<_Mfkabo;Ff-E1M{7>j(c*&Ni3p_dSKG!jjNr%&=~&X-kAYB
zaI{^_A^(Iaq56$)gK5Dbf4Cb%<F}^L?%af%Z$C_%?+R6M%Me1Gz$V?w*b}=x<(II1
zLZ&fDv`in@u-5_ot@`L9NQAp-C6w=IkBn!4%KOc<2a$badZyanKD}$d&MFWrr&hl;
zS}JOp?pr6*(R4|7emw=RRQj~Lq;btp{InQOraMWM6K~w9P#-K#dTHTDyCgke8$>Ge
z`d1&f`^~GMTW}<A-YJ9c6t==y-YS)_?ZQ_9l3JxKjo09oIKI0Q7^nBUPAN8H#t%!6
zr+NGVsL2Ln$V$)l)&Z9)_{~)ETT#l*AoQlV_!{(|#45YVR#M$Z1;F%WEq{Ls;{5jE
z`d@x}-|q7|u5T*R+hYW(E|wo_{QJswrqZhz`e_%R%GCD&W|&HmU){QHQ6n!;(ykD`
zdz-(Nw$@6ox0}v4srdSj|JlEI<@oRF(dBt~5qO7n#`Vu0Kgmi<6-JH$(p3eSE~N3u
zyMEQ^^wfA;ottb6)9#Q_h@-HN;sXRH;#Z;H)DHwNfrH2AUbgM+y|%HjMcQZW`LoBF
z>`u=tWIQrId64gX^ieLfRxrx6F6l!3+a-kBG4wLR$~yYayJb8Y$P3=(q3qSxX=!Pp
zEiW%K7LH{#sQk-*b~2nspkuqZN?ie|6MF@}8iljo5p=-hba!W)iReyy_nrG$UAup<
z2fw%9wzoE$_geM44pclSXp_J4=pa|@x_b9J-}zp9_k;H$oF!0%c?we<w+-*k9`Uj6
zuB1+UNDp@RcEX#uqenhMdMBK+KxODolmLW4d%uVZ`Q#gObR^@s6Z)r{yKQ~<2-v%_
zc%m)LPq#^U0Qs2pE$>A*W=t=-7hclx@Mz>KtdFQ@`4pgM(Fu<N9(khu{k`yjVzmB_
z)$*6NwPnvDs|WyviwdmT@#A<sB)`SC0`<|+JK4G8@sTss(Q-uL1RU%i3c^%&DVS2o
zuEJ|~Z!g~^R2H9H@jlt+;H|xbLmq5zf4A-K!kfcah);9#vr$UTOpUhXg&E#6si*$J
zcl6FH;B&i8Wyhc^nOlzHc|t$%jtrG1;@86`Tj2D0I}RV^{i{P+U43_Dsm)GJW`}|9
z&o;!1!=x|F-8leb*5;=t+u!|z-)%qm{zt(Dm1uyhZSNkmEsr&#zOGVMSS5Vi0kE<(
z9hj`|9JS4z1Nx&!E_q){?E-v3J?u~RgYd1u*;U61cib%?tW>C~V3CJSgLL#qBX4FL
zVQF8GceDT5o)j>Anq_g$62UdmEO+=Z-gQK}ydKy@{_=G8-AWyF&v*i^*<Z<Dp{_il
z-wLoIWU7$C#XI1)M|8W8FAKSJgU+Z-bvLI+eZ6{d+Wz(b`~TI}pRBj<|IQEEhu{8Q
zn`91U8uI_1G9dr$f;{_;#B_EB8a2y23%E%~<eRg)pT|Xq3S?Dg$aAWg8XcQRxm*#R
zITw8_Z4dsdL{{i$EgwI668NgXH7#j{_WSR?(^gj2z(eTV_HK@4bq8RMSOad(A5@aL
zKu$Ohj%y>YDCp>eCFHw;rqPCkJv=^YCr5|vlKDXt>iDUgk$zbh+vC<P@)(~ifv<hq
zegv$Af%8h+#eOXA+Q%~v(Py1I*nh-lh4<SV>rwg#Cgfu|UGx@(O5&T63UkU#`vjK>
zmm><XPttEz@7~Q022cYS)e&IcWoTbkC~3Z)PNGop%-uEnJH)>LuFgS<W@Ij5;mOW&
z)LgZcZ63oWy%zUXcr7k1XLqHzySuj=p}&Mx`eY$IqCImuXr_wTC|7v;IhEzH4#D$&
zo-KV7R^m6~m}81cIwhk$trY)_Q;Opr^QC)rjIQ}qTi#HRb&tk5XD=kP3^tbO^mBFb
zN}+4!PPn^Z)Bc(?(oLbyY5$_VrM#s~luJB!JapVhS%{x<kTwzMFV3-wcAK|wjY7?N
zXG%m{GtLUr9*+u6)TtY<Z2K7C6oo3^fdk>C{f3X6fpS?n^YGDbIALeCGlqdnDWCJ?
zt`Cfjj<*Hc)_R<tJ2$?}BFoj4l{N#Oxy$JA=$JB(l;E7cJU{F1+UeBmz}+GTCv9VM
zt8MS>!Z%zs@27ll|6Y_u9{Kld{dwEO-|^I=4#((2&Vy~=v-6X{dWJE^WB9C>NBp|5
zLfkQ5`=Is6Zbfj}MWND6jZjm0zqq{8zWw2c8NX{BbJsF_lrRIgJo4;E-}|6_@Xp=D
zdj9Mg<=SqWjO9Q7&GQ_yCam3cy|%L4Cg`&VjK%x#AmYFGAuTP;wz=tvD8@WOQGU#(
zb{uf`jJt|=;eWQa58#I`vgrE$ns+Eo(Jx)(v)Rt*=R>zgpf}!|N4wFk`|eD}FS|Qw
z6P0c5hW`*ccb9xm8H08h!>p5cnW-SpT-{|Ew6SA)lseQ?-pEBk$|dZ(Q8*}<AWU$M
zpY-+Mi(}IebbqFO>;3oX;|rAYv^{&i**3Peq9{<2?Lu^MtZ=FDN>jX@im$>+|3dp2
zqoQQRO}?ZSOQ`Um-%!pdTBNP?sZu`Ex01GtrQMwl?O~h{H>DFr3u<?&<+mP|CG<qv
zlQFs$Gt2Mf4u8JGKd-LRri&I9IGi$eT~H2i13YnOlRQiosY`>wXYU1#e1~#~TQ0Dc
z)+xV}ul4#7`7;-o2}Ai%7t%<-w!rZnaPO3MdgZwa;;;RcKNXag**fc<ehSUdpUA0`
z^g_N)`Z7u#=4NJcD#82UtP6T}wn?97k#Tw%JU!UkBIOtD!w<ie^xR$L9i2nBM%y9f
zc3gY~FPnuE(kB&qX(#3nlri`7rXH57PMQ)%(g%+?R(Uu(GeJ#&H|-}+?A>7-@ayaJ
z7w^e4EwAHvKRd7=F#f)Z^4CRlDiU3!RAc*DJAhxakGN>U^gxmD&>ppEI<|8yJQpwN
zi|`ch++T$@Fh_ZH)sFY~;k_vHIddxNI}Ok2E=+g7WidQstBXqBfB!w)Xg~NT$51z(
zoH8GgCpcv8@#wQp+y2fLBbPG1!lz#0Auj4t?=t$`^zVK8FnQm!cjj&On@XFT?xykD
zH>UD^OL)wi-@Un+ho;h@?P>qME^nH*rFGN%|Fz=lfB9ehW9DGhfr<oTC`gKeLkJ*N
zSDXQ9(TEOM?{I{%cgFBKl`}?5oXZeDfeoY~;el+hi(66nF;bULm9C3}Jq((zI_cGU
zSI9nG#c+bDUMah_l2Kw-<KUh!360$k(k{VRoG^pLW9`23^l(bGu=o}>HsqkHhT}Kk
z{q4e4LgibL>s9dOm+#cObFYy8@-cX8J|7;&iCgEViFeXY%Jt^kFOeku0FMgYm)S<@
z;>(}Jk)OCZfwTc3emFUBoiP|4z|`!Y9k(5rcPG}~<8aO(e-5!dg0Z@SP-jUdNvMoT
z>JZlCMS5;^UuKX~$m03aUJki1KM(UU#l&MI1B{Of3kN+1AQ_rWU{V3<NckL;WDsir
zuT^P&XOie(#oPE{E(7GMr-Ma5snyxME@9BJ;ss*vge1%Z_4O^0SMcC=e;HvJ;B5vJ
zkHzpVv!~Bt#-43t64B@JjTzgkM-1Rp6Kp5o@h#JN+myE_N7~AJ@3i;7_uV$@u_G|f
zj+S?Tp}{o5R+xxCPX5kw<VH=DvWi6JM#h8z(u~|G`U04WG7;+07Um`MboElfiYwkp
zhr#y+cHNzDWUD?wC+8;H(%MR!UtVlpFl{}YP^e6EaE=m`(#Z%#xn{dEfLl+GM09fR
z1ks7MlXs8d(PT##?zXYag32%B%7huuJS^)N1Fk!;beCCjHQg9t#m`J&cw~gKg2luD
zrsbC*DEv5sIVk~`@E>IYAkGPL^KgP9Ju2Y{j8+td4&<=IPOhvkgT3`IdE;4LcXowY
zB;PP%^re2)afCNEvfYZZxI%4idbX`C-D|7MYi-mE?r9fSMP|ngVWnNBT?TuQgbZ!C
zcwF1_$!}3WY@Ju^{_qic%kzphGMs;vCQrggji$Bx_T7iUq)d7hjjDtbsM0n9uX64~
zkzSWZCDVoTze_8hdP>^5Sl$1?R9STEJOq5~^)9MvTrGr3S{Mmo2&JY=dtfT>YD50C
z<C-Kj#Ld`hdhzJPCT8WsD*}cTFiy!O{(g7yu8jezeKo0X0N{WBu1MykSG9>k<$5P&
zJh~(>S`5rTWv(63om(DL?gDoON<DC@uL>ay3vhg0d*oY_vgS=+(!qLbEAKKS+(*sZ
zAMpJqe(;I4BAHH&{q>!CTK@XPU>U?Ezqsp`|KyQa_?vD_DpqM$naiX9y{<<-N&aTI
zt}J?5waTEJk{-^TTOY?y$L-`-ud>J8c!bLbp_8-$3eK8qH|#tKnlDWk$8_^eu2{_1
zzl1gb#vIdzF%_=|T@vkHeP})Vy7xF%b;hoD!fARN;YoCmMsMi?|5x9j;_Dy$@Bg2a
z-rZVE-l5~0&z`j>Jjd_^qa#dEp`!|h+$|-X<O5V(<PI*!9w+kCV@z_Vm;g`Ym>!Q6
zarcm8fr7hC+8Hl0elnI_!D}eMQ7}C+?BWgzLcP8JBFZT8^f)aif9`T{Cy&Rx+`D@>
zt5?TJ-<9cC#8D{Y1o`Zw%BVm$f_NCQ?3kvUjx{QUJU+!swOw`Uq~D3A_c7TLL`zxa
zscbxj0PYZCuxmg6>7TcM_iz6V<Co(y<8fCP1tpG69+$Yfw%UI1gWrjAEcI8A&Dbif
zaZFP=bBppg5%-8A6&Ti+EZg|-P#Yg{oOVGIBR-F1b*H_&+6?^Rr1w8qK3NPW>rC49
z8@{6=(bcQ2WFKw&$LCDEPuk9be9LK+BMZ|L?e22zXmRDY0s)U}DNwtT)}7jpt#xFz
zkGzQEe{d5cZE<&uN-xKLl_3hQT=1fBN!STT%M>9oapXzmUDM})RmO7a=*}MRlXKOs
z_Y--n&pY?;CY<+Xox>-n%u#XbDroNm+1uTX5=s8S4r#jXc2F>rT`KPWpiEw9zPY*4
zT#>I}LgC0-jy_z5*H%yk9|Y`j(LDHLIuoO}+w}Nwb}Rw}(l~6p@Vu8WX|pqP5tu(;
z-);v-rR;KbzXB8EF3!(I_%c8B|0VBFf+S6{JU>i*Uu<`C^CdhYa?h%)s;sV_#&lQr
z02qM`L?8gj14$7ccpBabViw{;2m}d%;LbZlNC2Fn2rs<ggcD>2Gng5mr@AUDt8$5q
z$k;r5-^|T?U-19^pQ@QfR7Pc1u_iB2_s>54bl0a(^*w!0ms#;YW9e9Pfpzi32w7Rp
z&Z^#%yl@KT*!%C?SDbpPHyc(oRN&c!a~wCyxC`F`uapFr`uudN&QEFXahbOEjx7KM
zcToT>G^#E$N;2+XIu`zeyC{gnjrvXsguy7kQ1*rrIs{g<6~Y1IIzZRki;ZLlr)gj2
z3e270FX=z@PudpYB=|=@;4~v8d|BMB61Y=@mLB7jCVhl@DG9rwc%hHN=W@gy#(Aj>
zv+S(kZZzRL{f6<Kva$P-i{wvFj?$wCzfQYbTdCb%(z7f+Uh^ZxpI033h#ce-eFgsg
zT5#e>FocfgZv_J@kx@FU++GQ<F{ja}c?Tvt<q+f^>>rrNL@`C@J5|5!9`)U(7K*En
zLC<_qJwj~tV-_?v+ilJF>J~7v3+2*v)&u`4*ibkY7xpbASBV8DV8>jNxd%KveM8M;
zd<TyBL*H0!&sgm+syF=(7*U_OIl)i-Fary92S319lJVS7W0D1klt<|S2e0UF)(zvs
zz1{Tm@nhdxO<6U3WE{~y#S{+k5z2BDlI$v-^t6|9v0H)NH22?l%RBVgNwc%FXE-sY
zqww>wgrX0Wm%1_UJUZI<t`lKk;GD{Z64mXfyeTJRDYS!nQI|@kZiO~&Pn)AS;b=(4
zU3TgnYOG?+GCr%V&=Qdz^=1xZTosHl#0LkNcMxVA;j(B9w1~0mTJvQ3hUZ#JALXS=
zW;Y9Jl>l}WbIfN*n{uEi1B6}(Ab0vm2+&1!^Mhfm!fu6BIJusbg}J5UDSbRW_M!9^
z-jHty!Xlp2pgd})XkU%*jsvYmL4y(v8G!s`CoJ@Xc^GxakRSLo*jc8BzK=YBzUwrV
zW$YrQOks%bg^nsW;lW+n7&tKgA!nc-%taZqi;@eO1mnfrtmvk2u_zxIgAm|T#%6QD
zy9U8|9%v=U7<w0w+IVMo-?9oaWk=E5Z?2_x-n=ImU8ENpk5Te)R2uxmk;ZmlwY=0x
z_wV1g42E63&z?Q^v8IeoWw$ACr9U#yML@lF=Z?nvcI*V-6<#Y(794XNAZen2IX#Oz
zDA^&-P9eia;~2X$-@1P<{rHc5Xa)H5=P%Np`WtZb-BKFY7n^hGjn&1}ZnER-!V7LL
zt~G|vG}Gq3$|&CKOk+@8eSE3Y@!Xy}^0s%j($V3*aP=b9D>Lcts`!!bzmvZA&Kqg5
z)eu;2Mf2I6dYMi*;<$V6-Nq<Q7`wP@Xj*ORoop%)XsBkfq5}HSKf6l1N8Plye`qDV
z(N=+2DD{ST9uzbu>fgcFT)O|kmoc`fDM3ur0mkq-(O?#fv$%-5(EiXBa1;eRa0f34
zL!S`>n3rQ$?%g}f9_KlNbysEK4zylBjz@)OfS!OmFm&`wLtglb0gtS>&&Dzc;UWzO
z9^2n|7jYQ(DT`07ao?m5g4>J_D5OK~=Kdgf;6`JTE8qzrMf;nF)>uqkpf4PkJTCjW
z;7zD`@h=ODi*m1<?}pdrcx-443KH&r@`(%3cg2Htu*eA|C3lCRR5{B^9`Ka0lLZ+Z
zBg-A3C`pks5^j5Y+ei0uf&x4g<!1o^<0NAP^@q<d^|rFdV?>6XaLIgx{8%KiqW<va
zy%qJhyXue1L-baFPEGN5vy+6qaGxZNUw`^}dh&EV&1vjjyZ?q4D`01Hz=`POiN*+a
zBlc83c(hC54-0~Tdt4;~ZXj0^-Bmg2;%zyux+)%$-E<sJPB~FNz1U?(nfM>2cO+WQ
zNfz(EwU*Xa+a4=Ws6jtbBzjy_`yPpBVy7It@fokcB^ExxyMco&ssTp?I>QQpHyf4I
zS*%;p$3l+N^K133t~U(KsJsj6f0cR4CEk*Sa%xvLIxq$@Cc_t?V5Tiq5aGjEI^5qE
zkFz0qD;^-20KuDa9<<_nNINjlfS?26Uwld!zenxayv{zVv?=M@$aypz4lnu3yzKv*
z-^IMD7%FrO{ChRcVqTZlsQF*B;%nlkfBI)1DFn?T<c_!DVSBv8eON$*%vgDFv#<TI
zfmmu55{dueGZL0Cxbfqlb_@T3pTFX_RKY-HKq0I;lRr!NY#Dx2R+U8#lsW(Fv9H7N
zGsqhU5*0DXUi_BO{x5asC;W4`Y&VP;Rb}i%TJ{$l`3`G5PK)r^vIj4adf;#GY$6=L
zf%;*$KjFiH|1gnBB%9rNFoSq9QK#Y~<8;Cep<pSWlfzw^gH(K%wj%)1E*NY@xY!xf
z*T8;!c<9TK`w*7C7@UhBc#$A+dC4Ubm!s?=jO^+`*?<6tBTYE+jume~fYsqO0K;?^
zf?AS5*yPn`a$^8w;^37=V!*8^Qydt~vic528sg<_HTV+l3Zf>WZKhIxUR9@x!ld$e
zKvz8>GPqNJ6MHeBlm+GlMwD<6g)UpFHECfI@tqq$S~`#FSOn(S?T6uA9)WUJgrAjT
zOk&4awI@T_y^4@}!#Kg9i?QpAA@TXsazhZ|NsS0oFaWH`5gau@!-V)a7{Ql5;>il|
zH7o-s><-+m00Q#3Yw|cV;gx3`2Lild=pn3B0vrJ^NQe8x$j>Wp#c-5Wdf3UpWrJ~Z
zMIjt0o-WO-dlgkK8<TlfyFpl({DV`}8)gV46HFcqIZP96PTuqfRwo%3a(&4LI}9@E
zBT%s*L2-HPuX=<6D#MxZ%AY)#SXnuw_+?^1`kw3xXMty_mlARrz**Xfi8FpyqUcTf
zFmFs8YL#jzzLZwvtAfZ5RP-VeS%sUu>=^$Jhr9Z}+M<jdw#lC~OBgFz9Zhg`dL_P$
zJ)$LSu#kWZY2eLw4sE04!VP~~k#7a$Z98E{-1s1*?Rs2R;K8Q~WCa0V_VdeO!;Mmh
zgUc-UsH1_Y5V%EL3~3j;1M}gaycJe)!dZ^Q)nD9c6_4Et+;QhGDm;?5><$|#*qtZ8
z{6zBc@KSiAj=qccpe_6t#>tj&B@exUQz*)cTlpb4fdjt7l){v<QH5ds{E<V3J%5y!
zvnk;zxD`CIXRu)l7${-#z&Bj&Z4-E>2+oVz=Eyv_DcwPt>^C&H*o(0FH&U@rcF>Qa
z8IF9qy>Jfy(R%Vr4kav(0h{o{M*nx3Ja&)ActAQOJ;OTeOa|=+ZsoAV%3pZJ7h&v&
zW4Q5e{N;~;c=?s>`ZLdz+Z<=7gN1DMS70B#DIiQXj#0u6eldy>7MV!9^fTNYQb1yn
zQi*|~&P=)@4D}C4QvQ!W`Q>+Z$Jf1|{BJ($_s)Fe$BWIKw6VP_UPqHXO};tW$2*ZN
z0M!44#%PV-#zVzLSiqu;!)`aNzt~C-zj~e?K3-27TjC2&MayoPOpd3s%iGjHzf6ZG
z=bBjR>0hMY#dVs}m|btSEC@xR1WZ`riLjZY@3<p}9aE3KdYu0IfAl}3fAl~6^YpV{
ze4>f-i?p%9>SK=XJvH4#K+8&EAHO6}nwF@nurkne9{Lyet{@2C-r7pfpFA=DM!a#+
zJLv4?<*DzPxjgHqUiUCPdibSxeoal>re;+WRW&5XtFYplV;X<>gAdYA{^VavKl<^9
z>794q(FC}yi7t0Fw8Udf3zo`Xw400VncQ-x2RnRNaf;vqT!9X=`-vSx2p~9q5SX&l
zhsp8m%)HV_X}{a|YVDq22|PpS7(B+!*n|~E*Md1KYKh0aLzS7CRGFRh(K_sS;+P%Z
z`5K64z785pzQBgn^6*FSmk0w029JOMx!3Dj=?AR`Us=!sFXE)YdliJoJSl=x@t5#W
z+*`n;AM9qed!s&=7F)G+b9rWkR^ZsUh4cuISmbhccA1U@BZLzOWZ<h1qEaqj@IF`Z
z4#dfsxa$gj7e$}%WKmd-j9X~8(n_b9?kqLZ+Hxatp9XbB&_>-Y;8nit`a<Bh{`^H+
z-`Gs`Ml-Ff-bwY=V(JNJA3uKX9bhNo1%+4^TCJ?CSQtxBV7CgZeGx91x36%-Tg`kO
z<wdvwPsIvx1S5L~he~h9NB<zK<JhK0kDjKl9zRn*+VHNJ$+1hf6#`y(Mg)W?1;HU!
z+)t|s>lKb6nsr-pbf^6huBk6Ee{dVXYY7igRJc7=ev}5&GjqO&tyXWQ>b&?S!4}>P
zz6qR|LfPZ8vw%)GMLHg%^e8`gE-MYI+?7}L4mK1I2rkK&u^Aqb`38b3BZ;6z8sNng
zX{-MO6L!_1-1WVH7SN&i;yw;eDG;wM9+3J}RNw8bjr8o<Qy*<PIXUCi-G@gU?R1<D
z4-dTKiX-$6IEt@->~UbR-Sm759HIQcoE-@po7?HDhmX@2556?)Y2(9V;kj@RMI<-~
zyg51w<<0&+I~_Tl0E8F*gY&>dbpvJyw^5j&#|nY2`jlWnJ=i(6w;RWUva1he6L;tV
zH+Wa(ZO}gVC8c@8pZX+o1>nZG=6M}(XWl0ma<rpisj+&lQcLr-rcV`#<Ngqg_C23q
z&d9D=P7k4-X-|Zj%ssd}XL+Tg@~5=3y{$g*Je?f()hAF+vj~ozcktxQMVZ$`KekX<
z_+lZe=)){ZCymu_yzxf5`^J5b1JBo=r!B!0fgb!j^BGpox7%$iF{q!9B|PS6&=U*o
zeKLf3eq|%Wa8CQ1j^e0qXtw9Zas#%w#oR=FTJGe9l8v$F6a^(m|A|h;Jc&6(2)P}H
zNHV5tsuOf?M&m6zIS_0su;(}2H-&<fITz6=5A|l(1#@@W!R3*s6+faUd|TNF-ns0|
znMxby3d%5sM7mKH!sXzmS@;d-*kBp;r=JG@K)f<nawI(^t$5MzA{`>Q?)o!aT^7PI
z-vcgzcccYu2^;w+4ULnG4~1s}A2HO)Y39fB1($_(16M++?Rlu<L|o_|?g|`jqIgyS
z!V3^qFy+k{q9BGPJ8C$I1{{V?G1l;nq6Hd%Cfa_6GC_Hx&_W5#{FeKIykK}{mSUL(
zkS0g#LKiqDm^&s>c7od;Hz^}yCvB@bLnBoZ(XYD3evV?E@VLwPhSG+6mQdcoUV@gM
zqw5d&GVbu*6K{cnm|apF?<{YRVelkH5d*Kl!V1#ongSHabF<^=-reQ&_FHebPF`>#
zIH442FR~MjW1^=LJ4U#xk9#AZJlpg_7k00rpg20*OFNtE;xjmEkDZdqN)wJ?J=VB)
zadzQd>`U!NYKYgFo6-2LF_(D-Fk}oyG1+XY59hn2So|_R#_mheBBjYfrb!;baSECu
zcr&Nxo=5tBr848YnU<FpOxsu(%q|`Jn0XM<W9}IZ#i-f>dZBgz=Qyzj`4bC=7+Zl8
z_!e@jP~!VONx_Xb?Z~*w0$byi((n#-rAdER%lHpFVX2?{wAvh)F{aI#w#5!wf2;n$
zl7%abpU`s_tuvlN=TUT1XCE1=2YiSzS2*Q*fD6fED|C;Yh17$EDu*nz5G_OLA}7ZI
zSNDV;o8l2r$XK~6Tms*&gh%Wk0`K90Oh>)2Ba~zG8;;|Jr{<_V#!OBufWA^k;J~6M
zdAP6BZ|SeV$x1vK(k?PD#)o0wXKWLXq4wvb1b9twkwpO<Kg#`-><A<ol#)UskEu&5
zt5!C*T1~fq97j59+3L>Te%chhK)HHw1YPL+gd&cCEPV+&-Vv_1o9csg(NGjf!c*oC
z+>OcItP72b%Aj|v;iExYDhJ19s#gk5lWAErZ*^IGy7*A;TZQk2w^$UvcxSogo%ZaO
z<OoA{v@I=)hm`x;iug(OjirT}`2PikYbY=A*>miUML~Yyg)yAI<m1~rEtMx|2yH>X
zgU^AlqAu(lzf_+?rUhEi2jG>#arkT&OE4zl8RKTqgD`4?fn8s=hFi2zm=ebCQF}I{
z{`4gbJkzL&FvBNO&yV~@)5CnzcQK>xri<AgeCF%T>(Uz-RpT4OyvF=4@xCeS?^}j%
zGOtUQJH94<`lmnr2xLUifvm_z;nmdzm>f~B0g7P9$`225vX|evEeLPq!!07XAhz0M
zi#lSY6`v6_K6trYcFT{-<aGFm0Dfb|JIY)>co#dL3P*ln<v0H0<?oluthD%!umKYb
zpYV>a2xtGCc7_K|r6`yYtKPCXUEAzY8i`-(>~sj3{cv{>?073L|Ak{N8y@)`|M6Wu
z>>dwgbGn8HPms3=7f;{=AKA%i`%@n!P*|DR0IN$4Kx~I*BEi=BT`>s#?s2;8Uy7-Y
z)nn#%^dc@o1hc3RS?uy+wE`>hVnPwSx|mczSRtTRZi(@M_(P;j$VE9|!aU(IfmWJw
zCmh5NqFoqy7(<Ax88sQI$pA}vyb?_e0|UReXDfYQ8mu}oxH9lq`6GKUnlQeyQwtGC
z!Vw=sxtLh7`wM2ob^ljk+zk5hlV9{h-fA;o1T!0WMLDxyd5V#p3IDu^Wm^nqeW79E
z!>T#p<~(x!eRPHD8Aln4p`zX*P^=imUKC|k3c;1gSXiKT0Ow+|z-rb|4kDbg;Oz{>
znkHS!$7NGkjz;20K_>Cw$)wr_<}g~0c8CcNlNxZ4dQujYu;4iDK>CyioK|Rk$BIKA
z<pE5DJF$vb<)GYh13&0jFf>dUXg`=!GlRL^OZowtm?Yun35m*zVhtt?n89=f)5{9?
zj5jI|eG&Yj&XnH@Ud83e1AIuIdefxjL9G?cZ}0X}nOy+E3SZpa=9G<is+VlWfq5|-
zwW=6#F=L*bk-2>NjUUp^77SJ;h?B<)>~_jy!RUjrJwC#j-x5hbP5{@yC#UIeARUpS
zNdAg%!sQRs1KB)IW3xxLM>D)^2&X4f&taT~(sKBmcIgxVuekBUcX``{pR*^i;y#Fw
z&4@M1pnt;VtzoXCew>cnfd{`i=P-oBNAD3=nU}+qc>WvtE-)I6>$c4f0q1aYsL}9-
zhrV1k{P>W~WiC&Y*$@B4U4NAOYmJ>9c#syR)KSpMhKxxmn-X4sE;lyf<}f^Vx5d5z
zZ#(JF^--xJayIrcZe(mEYIgU_Hpf&xE+b{&ttW6XKKM3{VR6(_{u~B3ha-#wguDKo
zZZ2DnE{7-=pYe`Ad%2YHm8<CRI5>Vu1Yl;Xy_J3}d{EpHKHGCJ8~cT4JjEz(gvTCl
zO35!}8kUNQVX<R)7mSD&eklz_jJWwH6848**;gZ>el>CDEstOP^1D-f-TleG{}H=E
zIHCptJ3>>0_6Utwxz7ZZRrMC;YtqKiRZUH<nLPVQF!39V-KU}@r=oEur>wR_sJxpV
zKUq)D);D~X|LMhzcqs8I<1=Y&axPt`$#gTJNukE@;ALh!;d}CIJsoS}k5G$Mq9~9W
zD#y9X$Pq0FJkC$g)7I8@diwN9`t>KDq@VrEpQT3+9;By_pQKMe{dM~M)6deQM_;8U
zkDsQ8Uq00HWjf?8);{-jo-6)Y`s&f6^oyVW+`GAc{mYNjr@#88FP=W_9g43zu%NR{
z;D7wX@29``>7S+#zW1Jm9j#V7H5c0H&9~l6Z@v3=T3lMPU>UlHVuy)1f+BWzp}Z=`
z8d_AV_MM*8w`eW9(NK61hYQ@PGsiYkPmZjCM`0B!#FUlUTpmyTQ9^O#*;)TM&8Yvg
z@*SEVtDjHIH?X2K1bX$v@o*?fW~at{0X|13a>NctxDXEI&$N&Fn-R=8Hn3`cj??2|
zrR?<NSejB^zB9DS@jG+I74BQ1ZgD?T@DK<{;WvCIi})${B<k60R=qk_QN&A~2^!ZH
zq<O)N_@HW~CIF2m?)%u=Ljc54W$G28@9btL->UEqK}!h#qu+a1pUTD_40!B@+GwS-
zknXK@ROah+cD$dORp3z-{Kv$D&<5;g5-%b?=2mTukeoX&D)pw1ier}!yS$z}d7k$6
zyH+yL#t6PBGrQNh<M;97N9oB|57Y5c+&k9m_tKH(1@!Dz69LMCg*6C`+|Ft%gp(Y(
z#STUK4=c5|l-K@2+@XYWp-~h31t`kJihGVwytFb!Fh!6r*n-ROSsd?$B1;^T`Ehn3
zz<a`XdQlTRsd+8NB?P#^$C(eLy%>GeUhu965NSJBO?#1%3_^V71kPV|gy(~I^c@-q
z#MHj<kJN>6!pGOi9{gkM#-R)dugnVy?mkvja3MVScXV)&*4Nh)M+2I7R@s3QV+%Nk
zQ0?ma!gHX-&XUIql%jFmn(#p3P>`bhWCsx|-MQGGlQ~d;QC{Fd8&Mv{yu*Vd@7nMR
zZ;dJ3LxVC)jja&uAe-}XYV7dfI8yI&RX>2_vb&4BS?DVW?>siPo2k=Tw7`v>4eTxi
z*VrM^=a@ov{;ADSv@#FlVg3idN}*^QV8|VA?7~0^$N~WFk0UR3(b8tL7fJyOS5ee3
zf2?s_dev^;K_h%aDd2n2+T1VH@tsB$jS19)bP*tO6fE`G7u-=oQ-*~m_2>wH<LFsA
zncw#UN!ri(hSF=Cd92FJk+kg0rCsU2?79R8xw{bIFEj*t!uZJ;!Etc(nZCv`cG_{g
zqmfw;c|;u<F=hD9a|IX1?Jb-FPc?`2f;i7vf(FyZ%zqFbvNO&Lf>cRx#EUwO0pMg_
zFeuz4PlVq-K2c*0`0P8ptT59Uq25K-j70_C3L{eHV8JO+T;(Avcy6UJ^`O`oy{Q{;
zcD=H*BREa@6UILgmi8%W+vwYKCs{2RRXo4VpG4X`lrfSdq-T^4%N#A-B2CYi$WI11
z;2K|HkU3cKU0f9acmq2g6n#-2k{`m?Ip~{mP+sH{F=x+Z(_b8C3LI%0#vt`)Cq*1+
zBm(+4y@b92t+Db_`I9E&0`1Q3R(5GqOUgrgkP2f53xpXf*wsZso~NqKeK(Zw1H~ls
zS9Y-SRG>4$SC>~~6N-I!9q_Tt{aMuDyR((I-BceGHcKnZR*aEP;D|~ANMwMA7Yb=y
zI@I{j!s8_t;;PU1?m1|;La`u#evML<ov-x$$~?*{(V&@Wjjz$4fCr0M=+7Ld)=@ua
zQ3fyr9T?M|Z)~R*TU*L}+&kP>R~FM^Q)7i_eW62JJA3KH=C<hFj%bnSP500|%u-ew
za90!Zs#vIn!Z#FMi|s}6M{VOdb(@~y6d88Esht?B)P7c=2n>w<D4@B&oMRnPG@=A2
zCQ2E`3XjohTLkYYz2KYRF@O&{=#h_kx3&AXXq)1&_{Q=E<sZkgGj7D98{-eW7JN#U
zt)Y~HcVQ8L+gfc4Ur|w6{06S_gEq}59OEf+I+R}U9w<B+M;RX%S_@Y8^Nli+dYgu3
zr7EyOZpIkmx*0zM&%F>x{_s4MgRz=%gq>U{zjj5JP?Dlts#NR3hu9fJoxu~}#&NQg
zzbsgRH^h&{6}|3}6~Mq7Wia>!oT06x30x^Fw1RlZx|pyqRw1XcqFTH;VT8#_pK&P0
zGhl+k({fAo3t$Cr#qP_7=zx!lR(;_g*m3#mPk)vE$v^!c)3*2}j{F4H%p-ssJT`rv
z9g>utai0Fey;|ILiLwY<7*~F%94ZTV#jd%QaA<x$#svB~IKmFAiLsmDk$hhyJBY_q
zU370={f8qu*_q#Jinni7MFbd2XbZ(r_%Ye%M3-w77U7^!qt0=(G<3exQQND{JMDT}
zTB!S&&^s#~rL|y~Q0KG0<~zNEG&y!{o_bciHS>w<i&O95V!VU5XsQmJ7_dlsf(7^V
z(y!=W^h4SY*nmgC-`K5pm>A=WVH_&-5BEW+9LDcadp0lmXA|<3p}~*76=6ohXFKn(
z4$NqJqxP4~D`7ld;`U1TS7X06y%F<m!hD_iUE+OH*x$Dd-(+5wt`%Q@{?ng+#9%)V
zlIiUyconlO9<KdA;ve3&J1+mYyWv4D^`+wGa30qqmOeQCige>IhjUoo4rr5o2T!D<
zLFVN^dU6<EDzr*vH-svF^Vk}tAVhpeJPWai?kDWUlrWCqc%+??FS~?SsbPt)ckT#I
zL(ibIMDhFsynt0{0~_qP4blo9eNp>t516`6Iea`}cR0S47s4X=3I;@C4}&@d|CQ1Y
zj05D{;@^(oM&?!QhLvL$XqQV3Smhohj#J^RTrL+!=@q%;2w3!pRdhTCIfy3Zg>iz2
z&uKE?)zq5Mcygn`A3|)HK=dIj8uY!xi^P;JD_DYw&MQ_50TO|N$n#A+6G0el#euP>
zI4B#eu;8J1Dh$t+7znRoQl+h6k)3-cm>6Snt^CboniUE_jY%;plpv<8sEd`ye|QsZ
zOkHxL6hlMV$bHag*aL@#Cfo56jH<1A>cE(?Dn`VONst8zB33@mLd=L)T8Kb-G9`C}
z3rt+C^a5T?v@{WNo_Yj1mlI|_&~>=F2&SKj1WKIiOFSc-E%?ENOvU{i?An8|BOdwk
z_HhfUA2_Fk)Q^IZ41q$Dzb4OWf7+dIw>R-%5`+WF(S{XS0D(URu=18RgvrDmyz>M_
z*+A%np%x>iFfeQo-+_Qcc3=bU5GRC|!abNxwH5gT2PVp9?nqPRF(a$A<*g^mk1;sp
zZsyhS1gwIp@&Uwz18Z?3uJgcIKPZ%eg`gt5pA}=$2>jO{0X$CoPigAMxKTbZe9B%v
z`KCNEbm(W`Awz}~h&NB97LUH;kMc);n4wFKM)dF^I;9LF&*APE{&Of^*=kqv(KluE
zxK0tnE7H~v?(8m!s3kH1@YkcrB`^Q!?J^S-=UnIbj&a>C;q36ZN1g4E-NGJe^Pj`m
zFGB_UAP)6ccs<kspAz12^0%YeKU*+*JC6NHBj}9r1%tE85h=$XapKJ{aEcFog=0=r
zKYZ}72BLh$WU%QO#u=6o2V3ojK(^#IF#e<u1Y9XY$uq!4T`;-a4j-<riW7)7#)h%8
zDjvB-TMo+{t~s@A3wyC)oToFhC0z5TzlL7=kG|zU;Zmtf34{Bf?j!Ehj~{>PhY<Xh
zFi~&P^RJQIZc%u*=b+Ryj4ixLW7G}1e&k--IOm#w<BeO6?QhD*ceE!DFZsIQVWRF4
zC%$<Kd@;qwLtNREmR<F>EdZehGqjg7^8H;ZzE(f{cRmUMR?r+Kk?<LS0J;)()m-DO
zcL8W3haeL^hNF3)6%Hr(L$9D4)P`gHII0BU3%kh>z&u;u(tF1{p142ah@*$(_hMr&
z?H_UZv#?@nHZ|MJX+|^=afXlKS!||v-g+~={l@+D=G`?5W?3=(!F%t>ubpO<KPy>R
zIxDHENi}qb-5g(h{vd6=SdVk}x3|;dhmX?3FTO~hefn8?^5mI!)*y`G7)FTSy*o?k
zFaGpT(qH@IKT1p7p)o(}2{THn-R-UP>F1xPM_)Zj-P5boXm`?^Z@-;Z?yQP_BOnQm
z&qP#h%nogik78vu6Ye?qH2Gbf_tWgeO{&h0r^+-JWKXAx;D|7}d%`_9@E)S8;sp?B
z-Jm!U?>Q#=c(}Why1U!yU}rO(9qpy{M-S4tCe(M9mjwH9m6Hp#L*WIT6wBZ~YQd!@
z;WymJ$tu#TGfn=v|0{NYq157{e3UKRmoPUw>7y6t_2$^5MpbQ6*Q9@;;fXo|KaRR{
z8I_*}LwX~8V8t`aFw$eyZQaMfw9@kOf`x9E=e@MIvyr%`q+AkTuhzt8T&MLH+k)>?
z-=}qQa%LU|Api?oz$=6@@KZdjc1P%k;G?RCJHl9D%pG29!i_sCEx}b`(v^>(S!}Vh
zP`E6<;)G+>dWUJ2_$n(ak7+tC&gYmKE?;NGF}qdR<%3cY{u!R6&rYHB&9ouBIu>4m
zJD1{(Svk)tZFW~3_fO=%>0L*xfJbS9u#a};m_m3R1hL=;!cW>M?p_tonFTc{EKw%#
z&An2-uPAYRr~0UH4FTBp?m^nw%R2)te5#25iS$uqdv(0Z#j!ywT0<ZQPX|AWQW8NJ
zd^|@;k{(Jklt%*vk;+9MXEnD4&?*<h6oOxbc_<eU)N$+(4{;Dmp~OT84ByA`UaXe4
z@KKNFDG1hF&db=r_2h|_Xr0BSw7PcB3LFL?=0FJ2*!9U+OuyzE#o7%MGmXbn8ZW%U
zUinZ)-p~b)tjyWezu8e>!CJ4UG3mrSyys4e7Ydb7IL)fe2tH9_Aw-*#8^<Vdcgfmv
z)5_C%`o%PYv>C7VUT)XZ$`UKZ7i4dGwKzKrR~H+pp}ylI5D}aS&+1AaC3Ho&%8`gH
zyaUMQb>XY^jH@r!)sE~MrLQr*-CbR>aE)Uc8F!fzp>Qg@xPTYRFZzf@!1{+TN1)36
zc5V6^_nWCstbA{`YH3O3TWE%W9>pI!**G?Bsk5lQx?}~FcjG8AXah$ALbKUjhG2eX
zz9#t1c`Rrxa132dVYw6S*veblo<*L9tICOh)3~WQkCk5CBOi%%gD{x>Oq<abdMHzA
z7Yz6Y{o(F66rQvl-}WYKfYIR@BY-3GHRG^47jE9op*W5w_+g;O^cC->V-_KI(jZ>k
z8N}T(oOTjNqQ)4+e4BY#2-Y!H8Vg30(==XbFg|&wi5?8|7xJc___^Hr1HP0$p3Ezd
zA^v(i=jS_;bqxI3f#<oT++(3?1h7&|&g8|n(*w3d)`#a=`Xe0z>eK$zCF-xR!3h!F
z{+^fVA&ru!$6@ky`cYOAb=fJ44E1zyl?kCed1DAmeDdOmT6U@c7t)Ibmc#T>VwCpw
zk;2Q%!nLZ~2xSV#lyUSOC852D!`Py1qZrDIHUox?ZR`w<J9=(BCNd5(-WkrqS(GYN
zq$+&o*g2E{EJ$tC*)d!*9-%y7?u;MT1;GE16-T9^kh;6JY9#;)w#jkk^r0NHtOeZE
z_=8}3fBzsIb-R`W#W9XI-g(6_#+Lo4jR7|FobiZuK4!<aXonYt$e_T6?y*Cwk{7Nl
zurNvG_pVs=moxEI>=p+GD90)@!rjC@U9t1cS63*#ogL`Op^uvb=2m<)s^$|ozIjGv
zp#7mK>|Ez~Wsa6*F$*$)E5--`-KW(k{uKQ~fEmyWJ&ehW<=%CsvO~j>A8`a@?BZha
zF30>KdkHuC9*bGH8YFh^3f@ZBautQAjltdMD|##cGEP7bScHMH%knmjiS-)1M>{Po
zuZYiEwi2tYaTdkAcWDc5#v$V`G$wX2u?UNWE6_6JBHerxF8950Y~zW>J!mPkh9~Y|
zWd{##EV_!FQ{Vy%G?1N<PFaXTzo1?yN1=zuqE{Sa$zmpG9rXd1Q2OnQZ{FQIP}_-j
zl^Z;#m5XXm#ufPC1@DLg7W8Kpo*>6m|DvxdKl)}drpw}kSbT-zx1v6UY$?W0<ce{e
zr|FH-hMu&z6Ik#99w^QTYd%D<f`>WnAA8KC&$5u^pZ|-0k^b>N{io^a)8|%lfg6mE
z>`3!sA>lz)JU6s<t<y;LxuCTz;WV_6-CWd-ML_UGUJRD|6mVfoWE}qL%LnP>kAE)y
z{z!Ek^Ma9?nF%lcYKq5&2ZI-f_xVCR{r1+TARwIc2}xHT{d{b((qjw*KYd)H`d(L$
zK&|vx5GHtuhLaYHD^PGX#Lu_Y9~u?G3;t64HN0s{`Ev5Y<e1=pL;uovCO;Hv$RZg(
zxWjg5d)K(ZF^QKN&q&uJGHu2`Z<{>kW`n(a+iu%1yx*htY+mxu#(9mH@cYgDin0CW
zC@=X_H#yioYJS5#+ZoG&G2g{|H1=rPuQT5!%-5OUCEhoM{e8>uP3Cp!TJiN4fA;4e
z@j;Nnj5v6H16bnQpMhi45GFg>pQpsfU?EAjoQyq+e|Yj+;*h@w8^SwN+zC_si_v%f
zu=~Ix!WF}S^_!KqW3~tK@>QN8tE0!|@aMq47^m&7FEjAKaaaayWHtOARxQF|p}Q7G
z-Y~btFFU&tup2`5D@C)nf8(RTI$R<ZcMgAz#FbkFl?(oQ{BqvmPg<ppqkcIHn-JdV
zkq4~extJ9gB^V=CWUwXtObm!dUWXsMMfDc;QRAaUMC47;^p9X3;?2rCgh=eZg4lSc
zkC2YCKr9d}m=RTjfS;b|Yf^0{P6IL%8w61>YaB-bQM3R`abW<+4CXEvebRsd3<U#>
zI?RY4O{`;;CyJi1<Bp*|Fc%P0m`1ks!+Zt92yu>MKp?mV=O_SwFXPq-s|@v~?1Of8
z<3(E)c&b4B;39*M?878vh?CMJD>jxoY30rx3ly0Az+iaQtm^9(!wQEW!>fPQ&Rj6c
z<<1-}<2x!0AA}P*Pa`EOP=<ncx0T=z+?gzfl1Slm5xFLbdm{7`Nr(W#piYQLT46?g
zl@RS|oKWH2PO^ayw0kiyWSsd74&d)N!UZ!dDuZ${GXN7!;UM5wDi3wQPFuooh1<2m
zvQ0Zy5XITfgryAB3!w?ZhcY?A<X#szX~r{fjJQ0s7lICOlKiZ=lN({ctum<swj|0Q
z|1qhG$)YPlhWZBwm`IT?WiJZ3=}A`KHpP6fLY2m#bb_amM=9YJj($4~2{SHX^LNh5
zztlP0{paz3Z$gDTso9@!;p|b5yvWvJ?U8-;DkZ>9x+OuUPyKMqCVM%O%ixckFR&^Z
z@UO=!-m()W${v2<?3%`Rb_)}AD-OJfY0wP}xH*jH@(k?!jO1m9?EZ=grYgdg0O0%}
zl8q}I-76cc1aqA1?ihogaCFN2>nZSXnOsL%9H!VAv-RyCeB$N2%7<Ush3|ozERk@8
z9WfCkeo;aa+eYC>!#ae1GW7GWz7>w7@hK&>62=xVOcwqfro8PraC5HuD0zo7Z7N$1
zgD>TaZ+=RsVhcl%fJeAucYn!&oQGgt&@bv3<+dR##ladb@y_YlWY2=L<vXmToeOIJ
z0<obRrr?WVC3m~pk8eFWTm-|O;}tvpj5FaZ`_Mn9W`k|mj`AU%o)WqkN6&tdtIDmi
zW!oUUgV~N5`%V>K&3FFIkG%6+592lWS#YNkyAwEOtgNhMBJUOSd0fX0K7er?dc(2J
z2%s73xojUk1X_gPo8!L5H2x#}LXbha2yPI#a!EPIB!2beX?py0J#FuD%*nAP!hH*1
z$e$wt+1bI4-d20j0t}AiLlCmEw3t>`SKVy8J3BtIE<g?qPHAzWnZEztyXo%Qn&|^8
zlo4KdCxv)0E=}(&FQt3;@22<Oeam-EF}a7gdHC>A`r?ZRD&KSOrmEJO>HQymnBIQ>
z`>Eb+r*YMdRnSGy5@+_u(O*6WW-1mdK{u%{NAy7JV#V)8;!Yio$D{CEOwNM<VNA5|
z^i1&C<7lFdw6*af?PzlT`KO<FGJI#bliqss&Gg2-yJ=N+c1yr(urmY!JM@vZf`8#&
zva56MT)0T%H`nQ;dzkijw*;r26)5lzD67nqR#1Aea#lR59)u7G)VM643(Q+}j``vE
zyEtvYLVZ1TP3jjKHLqX=rj2GTEiEmi<)v0yY;zn8tCHE}(Gwn>rMKRABegrr>I00w
zV`+DHKRtc6o}NA1@CtMUG{MJCSpbS)2!VOv9(P%B+$2ZY)cv$tRSQO32IbYP&!*a(
z@T0;`bn(%-x4}!C2*x|xTS|K}EibRCo>ePnxd(<_icg+zq_4hus<ONgoZ>_SctnIy
z2f_h_6DX}%Acg{ig==(lgq}CT1%zzyOdNxbU~6V-%y)=c0VddzFa6((uDpYb9s2Pu
z3&`Nf;j@4N0!&~Hj!bB<Mv010?&9J`Wjagy2gg3@kGO8bIgVbMRVr~BfCb2gp~juT
zb7P(lcoqRPWeQ#?;AEk-aF=^USe=bv7KIQzsmD7RMa=^j$8?7<4|sv&q+vmwcvg<`
zYc!iqgB@Pv$!d0v;+s()YA-IP#l@vGi?D50bpVg__(&%BN~OioTHsq(b?vJSeaeRF
z$7pJ0!UXf9SU3hRYUKiay6Y#qaG!Y!f^Ht_Tv6RPYSl+XHR8@SlsF3uD1Vl%{N^4y
zR)3Q>tF}>mGWM~9(L(XeGjnVq{4Z|66+tKpAhM=BD1ld2IzGOYmF~1_9AT#Zl*`Rb
zb3%B6pu5#-Sa{yjn8mK!j>^UEQ|=&RM_Aku5<BcjhdHDdJj$Rz<=7+?YbeB7b<Zvn
z6#WSF+6ySX8uA<STnAc)vJGV`yAruuhuvS?M}jf|<N&rP3c0teD!8%IpE@kBta!CL
z?QCJL`gfV1LK~<Rc)(c9nAbl(@{w!oSiLMqIVsd&{-8IwQ>Y{DXMB?T<do=(#wrA{
z4x=_kxxoB~G?@P~1~ZR}s4BSeLGHjwWv1M$%=i6->PH@DuusIejS@0&Bz7B;7In8$
zRsQ5v=3$KQ0dwV}KI_GU#y7H64S1`J^f|^a2ct}~xORk*e}o(1C~%O0R|dcMCJp)c
z0k;h&`mNG)lI{lz$A=!nhi|+xEd3pU`H*Y)4;H#cp*;6;8407ub2a1wic1{Qj=5^|
zGnb0UkwC=L+hOd+BNzY&#vkA$C*&6h&^b1i1rQ#W=+lY=jEz5fTfU*N&`ow9t*$Pq
zEvxot97CZCFF?PekH-QXcI~iJiSZc0c+%K8?=m9GV0>k6P8?(zz#o6=)AiUCM-QUJ
zFV`9{M$ji~^-4l#jRK1?#`kIKxl!8m0VQE23F8+EQpTE)vE=c{@Czj{bfw$v(s~-#
zXGIIH<tE-sWda7!TKXxAGuR#K>n()Oz?3l;WeRB$jzu6WeyD4VVAm;%9L5HYUxk-~
zzW@#@jrhA`kJaE2$_kENrCnLvOqrOQbBr$h0XXW#Euw8GfKX&#DPQjUWWfWAT$-Z8
zvnm5Hp<dpFGpjzM^no+VM)1MAc)ZIUSr$htL(if<ma7Q$z@AXzs6A;X9`!Hk!r07s
z$gUt3+t3H0k0{uoe>`4XZaWJD*fon{v2Gev)%ZnN`hQu7L7#?q4aFGrGn8Z%;SX_H
z1i(Ud#sQQxTN|4`#&u(3!^gv-IN>f&Xc`K>=r`axG9;9&L92j0^hy9#eIZHg&=!vi
zk0CzI{YrU|4-et4)V9zY7MYtzWMM)7L~rpiyL;kO#5X;C<`WPOM040}+Y{eG-p~<t
zqM<ORJ$$#AaHHnMP2$hR|Dn|6WCi#-cs|BR`XoE@s)`412rTJeEI3;*?Z|wW-n6~t
zIjSFvU^uQ3g`ed`>YoQHFAIWL(87aK_p7hIN`Lv6f0;i1`OjT1FPvzgRE*uz^TLna
z?Tz#=|MGuIpMLr)!<FO3*?G;vw%A!Gdaf~=c?-im_p?q3pI8iy@{=+A!Ixj8pZ)Bw
z)F)X`#u3Bk-mL{Mad9OcS$t-d`!H437aLn?Q|(B90LR$@S`*EnKVkP>jcQYPLGG06
z3iqL3YEp&(_oXw{dr}U-lLaoo8s53a;w)(OEV~Y4{6moo9LLlb;psU&34DN_)8|N=
zg*u)O7@w3d|GaIoTQ-9*#cta$yx*htY+mxuCS){26Mny$Uop179OWf{%&XtUgzN{)
zE02CkoPkp|uS;vxe48*|XMUG>-xT)uEyFjN*QFafzW(&5ADN@W=pC{UMbH!En?Z=T
zzU3Kd71uBksyqx?qyCC#sS!+Cn4!Q$E_NIVKMeGeKo3I*7=f3BWctV{*%Vu(6$}d#
zz_0b^9TjrA!!iJVK!Lw!<9IoS?8TQTxY{K={S_h0JHlcHX_v5i6DCS0`v4w<30Hm%
zu~CL867siAwuHZa6tToEVIrUKj~uWws1)BYPG3*F9K&`qiN=fgCIYXFD7M3(@P*pJ
z6FW^3dSZz7_xF8{{G|p|QETOy8C&9L5M>}H9&lt(MDPKV#-IvA#$bv-gig<@LKF?b
zjEk}MuAai!is67U7t@U)PACyrbw)lAZ6=q&jKaiwWt-yCS^&d#y~_jxH(;b`y4)yN
ztmyNR5ej1lTfwLg1ev2!u;XWejTjP`K<ecCI1Y~N?h5}Qj2l5l^`ZWrz+$iqE`cv(
zrZ--L@O)>tf#D+>V_Hl==k98{cmJLzY&f!_h}9zG0rSKJ6h_jQFyG*p;gM4U4z#1|
z3f|~Nd!gK9s8I=BHVYEe@TA4covX{hjT^nWXG_?kzzR#dX`;&tX&5k8mLaq#0yFAL
zp7a+c|H@p9aWL%o5stQnNrk~i@C6Qef}vaxzOfSu+(5}!&V%OLU{rcAmgj8o)*ra(
zRhcR~6A*T2kuDV`K6l%}a4^}V-<*l@IAc|c>M}28F^-UkGFm~Vc$C{a4h<iw9g{T}
zVfPE=6B9;NkIteuWus2i5x7vcGGTK6B#i!Gr1?{kT-Qde>D`p<j3W%Wny?au%(HPh
zD3dJY!Jq$a=N)z|e22IGl$Mne`eW>MK>J5He;E;SeeqW?d~!&=WBhaf^jMQ$IVj-<
zWp!O6EsTZ#<mG@7Axb0I<TBVr0q~B`u#~im`%Uqe$w`GN?s|Ll%N_$~rA_(xep!h!
z3{%Pirn@erAqyz)ddoOZxf>D;#CF61e>@`^WiT8uekmRNave*pv!{P!)W*j>(r&oe
zrbv`2(#C{OEOaQTg|ENt5OGG~QCbDJaEdU+xPG=73Xzu!jO$3ioSxw-cYHE5{51&e
zaz{MB^aDORNqrWWkK!}%CBNcNx{+SEMdV1=)>0T5zDEt8xMj?CNb@VM4}Nl|jq#_?
zIh=k0iozI9cH%qovx!2+V@G{sHG<~v3|Rk;=Hr0by_oVr9z&~A;U||h{|xI=(sORM
z4@yNkBmk_-cdj201*dT%=}0eN%M)RS_V|rIl88J?PM8S&i(h{HohiQV|K#8Oh;iQI
zwkCBbxV*?EufS!}N|`u*dB6{OMJ~!KE^$9|T!b?0&|vp66XjVR(UW-us;!15f!vFs
z2{o&cHR)qwZ{fejc`vB2og)*k5_flTJmSXIUfSO8ri<%|)IYyXhdq_=^43SCEH5pl
zx+bX5Oqxf)*HAg;RmlsQAm3SCN-N7ul&h53D`b1U6Vd*9dgrbC>Bm3*e)_==-%sEB
zK+g}ppYGhd@1wfluh^}|CG7|jL`Bp6x8F<eegDI>xN<j5SE}Bf#nFDOo;^9cO2;SX
z-sN+8%5jgEatno>6*i(<-TrCX5}&kn&`Uc<{nXdwe{!~xW~&Xul*#m2zn3mF!S8n!
zZsSE-Uw>gC2k=7Zu)2CD{m~!)B)$K=cP)g`bU*c#m+!{1(lQp=#9|#3D7W6VRiB?q
zwOO7S!@@g>1Otwof&K$J_(qhKEXbiv;89F}l?K9*IX#tWwg2@+I@;e(`|LRDcl|vh
zQp<5KOYNqG6TTx%6MWyBru47R`o1r1XM4v&nnt6Y>Us{3dg;ZB?ez5dhA#r|YvPZf
z&$~m!t8ok?yUC!_;2#S;#Qs@$u*e+?i_NsqtXgr1vTs^^0z23AQr{9TaoN4{y%K+N
zAbuUe1j<nEc&XMG1piv<on5E(t%LN%!)NKiS8>PFQMZ>)j!%q}@L;4X?!$a9P=)tK
zNQO`l#g}<@!B5PH{>6jIjRn^9A%x-8IRs3y3KsC<+%aZ>F?R7j@Jbx9$B7O_=}1^s
z{-bCEKTs^xt4de(BM)|!U1@wgSO4^)G<a5}!v*m^8G<>2GBqD5o(!cd^9Y1rC^q2#
z7_b<}5VRriMbKw~q52Q~mavqUF@W-dGdw7bQL=D~g^!IAzeU=FrJcZcVCB&eoPuSO
z9|CCN$x&E!>pbf0(4VOY?@-dIr>QL)t)}`nJHGU<_A0x@7-NqP_I;%3<42Fu0Y}Yq
ze4yGHWheqicu@qptmL-hLX08tO8LwygOiLoDyQ#sq7LF+xpNFbDEzKTmS99%Gv4{G
z5zY0+CT4^Hv*}jx)b-LHO2^|<_5TZvSBGi$@YrL)-l6dQpzkAJclP$IjAxF4vX-O6
z5Mq}dp_h!a=I2rNqR3Idp<N&lDnsF=>8G{oOyY<p9_}{**Q%@}uhAAPlvF>Vy9i%V
z)Nw}%f?(nzfTf?HJVP0Fr973X_~j|}qY%)|ss0G=QQC0?A@f6&2<+-WiNH<(?qoyh
zP_I|g(lQFmmQvRkcN{ytnDbBsb_M$+6U}dF?~`8FJ5{^;I~vyx)lNOt{oFi1co@p1
ztN_gP9>b1D=9^RuKhFdBlM`(N`vHF1C>`1m<+FF13CDc*9761B)yh=v@2jH3t%GNE
zKg_*J!@F@9KUMFTYeXL81&xb&uJFcXLFpqL5nj=DUJ$AHF_%$X$JDRl;a=v?mK9_i
zEOTwwS;lWYWxmdLnJ@A9JG;?u!D*uK6Dfc)0SlMIHU}7_;l3~9vMUV>*yZlsrpl9i
zhjqwhlD~!b^7H8^q$796FsCOwc{x7$&_=vTLl0q`j!d|-(?jKy8+ZsF-?_7DAw6wv
zzC*C0F3=_xGjcZ$_>Hm!n%5EiXDl-96aDmkd1`O-E+`)q7sV5N2XjdH6=24$Eyh{K
z1Ptv)|6>sk=~p?{&;8)UIL-LV0y=hKGuC$&7t#`U47KX+$K<8VTqp33`7V8j-Cf{6
zinEw&<~C5Cl#RJE;~_0q78o&JB1>Rr6pI-SIC@lZ5Sr6Zq0MLKF&;B!qXY(iN5?($
zDlFD<UsAhxr>^>bO%Hg&QPD@ep62iu&X*(m4h2h&gRRa@r?zP6%2Lw;d-@LLL~+5<
zuf(Bz&_k42j6Kw!oz}E5{1S9)xg#Dxw3|gZDBj(tA;?ON^bB8iripe!hfrW3`(W&1
zArfPi<z0$rrLJHogz)Y<6oF?NE9o-CXJJ4`JV2X;HR3&5-a%eB9@J1c$ZjPMiU;9A
zp<AXULRsW8?07zg=hFBL4uB_RcUV~f3}4HDhC&~iABrv(zc8jVrj}iN824}2l;7wx
zAO>}TZqr|4d{WxTOY{!98OrCN#e#$IhGP}sb6J$ZJyXET3Pi!d3PJJ~P9moPcWHZe
z;GujvIXm@ELf;9dzJ#I--k#n3EJB)|P&u#8gsWF-&v7pjg1<nS1+QKg%%BJC@@F?8
z{gYz_mpZMqyioVOsmq;)+)#qTPatPh{Y)=~FKQ<gUht<U!j10nnd}#;PgDHs_tU%Y
zyyG-k9Q5UvU#15SAEoCnHqz$iMtb)2Y5Mi2pQV5HKmCjJkN?R(OF#ekm+8|_KTR*5
zKNpZL)Ne#1CpZD&OntqdZbkb=P5tJcUheorfwjab2!i!L{-^(A`bYoxpQJAzd@6SA
zG`;oKTWPLZPtP~^M6)TQ`nTwMqqUe;*Y2jpP8%}fae-qyIq`uL4_bBjXHFFgB^YwK
z-sxrPorv#Je`EoVdbaoi(R3jxJP-7i<HgzG)jK{AzKQ0k4Z#rl#gv*JK9xlv;8QHJ
z;?$`#^~-(Zr;#;=@WB`w@1O<6U?24_KBI2>8QQaX)y_ltzRq2?aFBV`f5gjX)F1Qe
zcQLPXC&A*!?`Wvk#~GS$6Xxs8?-K8u!v4Nx_$KqZbWwbb{k{Lo-x0lX6EY6@Wpt1m
zL<)m9JeHH;hYttVrxMS>!e3w6$G;`fh#T(+QT$&uxHHiV|6$r;E2$NCOw{ky@FncX
zhfPjlP?ndyU$qZ`hmgFKDvZPC9}7{j;-R0CE~cbcj8n;;gD}~t#K(;0UE-L}%FYCf
z>CrG;lqG-4I4QmzFVZRShgq;ugm8)Qe(T3*F=9f08HtmIQ&S)fr;djR|3nk1BQZid
z`@8A!<HzZ_o(oNete{XJ?}pM~2;qc5;vNYzx6TztkK<(KJh9{4e>3DF#0WJwTez;l
z7;QfWrB<whkzI^boLLO<gb@_8Dk8(m1_VqHPbN_?i7*lP5r>jPOkFqX?3&0G=2Ydp
zzD4nHZo&-#Wu>>z>}OzB7znJHND)kgQ)U$UtmuUBUU_o#^x5<DOi!F${JWWeN^Zx2
zaY6MME8HY7h|~iDPy54Ym5GqsJ`g~D5Qu2KVXDDYfL!&Zc6#?GKT03`=!a=uM4t1#
zeeN*=C)J6ggkWU24}cYqtfG!bEvS4w;Zhk8h%l*!$&WIs{gt&Roz9O*T(li4ep#h+
zE&}Pv4)>g>EJc|{*qQ2dYArAL9+SElk2x{kBuxAI0$bIUiLbZL>j4%_7MX9rDB0+P
z2_{M_3hSdglm<8f4&RtT3Ty^;FyP=V!aHymMv4-lgk^OutH)rV2?sIu9cO0nZoK^)
z9J>;pAatNTd<TITaqttIXEMTF7r+rW(xt37VhR|1sl&yUa8Knb%px%Ht}lQpfWqV%
zr^WCqf|&4#%cKwP*YHu=D88-`hECVg{k8k4v(OO}GOIYjaFU)UV&DdqWgLl@vy#pK
z4#z9|6*t}yE}P<)U-r|F?ZsVX^B9!F;!GVQ)9hfs62QK8E8YW>gXEVikv4YR+!t&H
zA#DA+Y(_$aaky|D48Nr;qozb0BrvditpWE2#nRhZWK*o<5<9ZgrX}rHXh4kU^efWK
z3Hk%KT+Wv=d&S@`fBWWo2_pvZD8{+CEZISZsG$rC`-qq056bLx!V;y$7C07e#gqs`
zTMiPgL{P#d?Li(kN+;KI;6xtzi(jWfxRNrCe(49C;ZyM)O9`{{)Ub<W^G^=%Q1Mml
zC>8j$zy3%4OQ{^rwlF2^tM);8u)0qAa)LRGJ<2B^@ojhrFhQGLlPI?Q{3_v{R&mrr
z9(wH0w_!;b|2Pdt$=^d0bql|tuggptb|J0&=G$$^JA5%gjY036mX&(cp@iU#Te$x>
z|DS*8VBdxL_x^+b=WX0Uz|m+NA95PJjLS&LRVvj-?1zlXuX+zB@(ALY;Kdz}7P!Nk
zKnFC&b0iKENWwu=5G+`D2JH|pfg-A`f{w*EA$&89W4wos2};C8se&+<a&Tlx3qkGF
zcxtucz5p&5557_*)&%&9MIL##0z#Yp!@iG(LEzo#up4Pn)f-E@N5^S*zpF5D>@Q6R
zFToK(Q!{g()RtqK=rX~F#j#A@jdl{plCVQguAb~7SVMReD@Lo{^~N1b2>79G&`c|o
zW~W3OMdQ>NCzbJOuWw;RU%Vp<Km^Qd%iJ#nJURX-F3_hA-9Bxm$@wI-5V{<^pOsiD
zACN=9%?ebSZboq@#_2Uz<|TkXLUriJ&G=Nh5Z<1P9&_9hLO*y!lu8xBoLx)o>|*8h
z&p-Y}`s(R()f6|5K&qw>fBd1!T2I|$rF(vv`r;Q5RI$+K@`{z~<H9ReeZwEmiWh-j
zW&w}tq;yfjp|GL-C?^V3_+RRLd69*ns)s<E>a|&~lm+K31X3Gttj^r5;3)h;ImXVC
z{rx>zuEby0(}n7S;$mF&6zI}nx0m*hSZUvP8=`RK?hxOHr8w-kgFm93>7~B2gSMZ6
zXBQqKC`TYKJ_lY*sA2q_nVn2a3r!zqv$eTx!T#OVw(wv^`Cq8rLRke4pw#n;2`URm
zS8*iP;zBENmkapB{bKN<6~+DjdvB*7eD57APA&v9gh<CHmsSuT3V%^R(|6)DiCCdN
ztNwyOA3Qic;f^Wp`(RbEU@F)nR6&qvrPTDCamz|Ix{%@kTkwD}0saXA^2*AR1!KU5
zqZEDDobW&HQ(}?Vxfchb_<|oq`Hc{hu@byQ=xE`v8i{!ZJLec9D7&&ulfnh&L?}QQ
zQt5ME2q<`+sl3~p&(r$TC+68hk*BnTCuIdc^<uU8h2}LhG{Sicft4mac^P}ad3H|B
z31`{G!92>mvi!M!mbRupA#6N9?WgYHzHs8oE5$3U5LbtCpHh9O3%jB?GL@Zr<b#rE
zX0GNVQq#@3@piFU^|4{_(1%C;w6C`7pKw1K3Iy<sc^Gg!)AwXzcN|Ju1Z=<`{=UtH
z`}H{1ZAW9-(Q#kxr!jZ7qP{C${VbGdD8CtN*<C?NfhBVY1iZkOe#f|r0)0k#Q-`t}
zksV;nS<W~TQn-RblKF%cta=cJdzXpED~_OC>@2!(9jQO{j!%pSv<EO{6+8WKduPkX
zfsLt;Old5juQ0bHeGtyVV6`W0-~n66!1!<BuEIdO*|p^pPZXay7o&0PqCnZJc;Kzd
zRE{)5V9WzvkQO^CNE=*Yu4V<TV89%u+dWE02ZzGNP?*xE%`dn!sHJ76-H2zQyxP@w
zZmZk@PtBE?I}0x6Occ*ivoY;(8cLJAm9G^^ax05Ftt85Z_&@<Via%p9X_!7aJY$pm
z6c1WOS&7ekG%wPi{{tNSFyMvXf+M&}Kf;f`LtKJ+Oat#6QfZJDZltRRr4{w_PHvUK
zG6dQZe}|2jQ48{ld9jsbdPn_~fXdDeA?}m>!4E!2ch~Nwot^#ke0@{2q^tIc{={xD
zj-x}lva`D>+&pz(>~xll?<^3Z9Z)pw?Cfgn-0~RIRNr1*Uh&S9f^+6qn5(KR;Q7w(
zuK0p|FY1JTAoyNhS@w9vacqp8>_Dd6$Y(eT^zPb<6;Wqr;wg4^l&<=P`Z0VH_yS#k
zj-ZgCAA@U0hwN_Oh~w$h$9<)M#uxk;Ls@jfxG|?b31`cgetL4Oapy91nzi))JNMHY
zZ`@ayPyoU|0Q+;*lZ7^ny|fX>9a1*>G<}BSNY5{Bd@Sz9_Kxauo>mv@>AvU^$G%-%
z#9{%IH|@nG6@Oa1$E}b1WggF1&s|B#fuJv(RK)InWCz!xXGh{gQ1HWZb=t90h(5)g
zjpj4*Vjva=oQW=5;iZ0~42>OSM-lW76i1QBj<;Qvm1BnceU2Z#6o0~4F=3eq3r^aC
zEp5?WWM^H&bgZ0Az<sl%75zcE=*{>AEhCKE6*?sxhqnV2t)x?%M}L5xUz`51xEn=e
zjF~FKut({aYN#Gc2VC{y@f^nC)W*<MFHncJ>YK3xI!IeHrqjO}3oaB7oP@t%`~ufd
zBypE1i->NOHai0uPboh#IZ8pFq_3K1#)Xb)A2p@y<bi^bMHBGN$ha=fX)pLowWsg}
zeuiCvlkfp58)-o|XC~QEH<j~ZF$4Fu3HQ`yv_JWAiouciK4=;^18)M2^}UTM8@rd-
z?d($n)DJ6jN_RYtoBs4ypQNAt)nBD&&z~Dc(0+JU;?YOAhZEXE8!|po7vDX(5WBRv
zH<<CAoq04ki|*h>zWnkFhhs-MyUXs~SxbNX;rr>Y|H+Tj4?g^freCw_3y0zX`>q4@
z3cQ@0xJ`3nnBmK3Rc|k>1~)kNP!F)sh%FqYd{<ud(iP3~2?X=>yHIlT&GE)8(n7)b
z{ORNL+2@}-{FN8^Tqo|gYt>^{SsV#}m7cuVHjlW_Y8sCiPrSfF<)B4rPfWCNjHAVb
za-DGDUwj7mz<bo5&FkzXFHEG%_iu?e{InVMr#Fm#j~d3XuQ$W`37jt_|GKnB&Fj<s
z?dIF0UCcLy9fjc^(v)_7)%>Ub@!uXB`@8?a-wD|SAY&xsEeKLrI+n)WC^&A>GWwA_
z04JKiWb%i@$cR1-{G&yA!X?6xhQIvFE7I39lE%Pfx14?UQyC~T_MA3l9#mIuq?w~e
zxbmR}j~Ft=O%5>FIVnB@F>)H=oQLvU(NbUMrI$Z}n?4*qM;rtT*Z3VNGdBD@C=*%^
z!xSSNuTtIOTGHjiq3l5ZId)Flzx>PJal_N@SZ#`hFcExU344S_E<rgkl$F10xuL8$
z74bXl9;S_r7wM~qU!{%p=O%`+qD4%P20{xoref001O~x=;Fy|4L=6JO3f_fQJJt2}
zaXT8UQG!)rdLVW}Y6e{HPcWlB6-on&1=GgLNtitR%7iZ%%2<5{V;eAlfu+LA-bEk<
z5l0c(fE9u%E{Epa<!dtrOopulQ9J|_F!R0uR&{5^5NGPb<T1J7NSbGwgtA2#IQ$_E
zFatg5;8FL?&_e_TW9kE(<YYhBSB4rI*k+RWk5!)7+jm#e_kQx@^!9sin{dZ&rL%M*
z#s)@>d%W0Q3}fI)gy6wSQkWVF!y6`t$remtnZ(czD5od{`9Y{4;*`aV6O+hOR)4GP
zf)lGe&5RL3|5NkRY4J`cE!|!9R(upa2u@*)c~G`pcqLD)2r&buNjSnSZ~0c<B!_U2
zmA&8y4IFu(|3Ps^om4)h2NC568=p<iDizh$LMYXZomnC52}2s3+YJ+M!JBZ*6E4-1
zcHkZ_3@b%hSxmiYw=0G3X_Czu{v1CN6CavEDeIp}2+Wk>EO$SKt-e)W_dAt|<0HUv
zPsn)mN4sN3@#VB(L@ut=^u%1cv$B@%bXHPjwj#KxedyyfzTiQ+;wd5x@=&H_geVG*
zVki%Nq}T_Z*^Tcqp3?J3bKq4h3WOD*F~+~HTO=3Xvf<Cao(LNw4{v*Aw=i}HYfjgV
zkN+UtFnolFU)R-M*+1K(=K653>}OY%gF^9L;_!C3TsQqj$YIPd!>>V&N7BNFVwaj0
zL-HIKg1>X%+aBRZdGISr2g3{(y&WKj!vLo65APf(9I@F-1is`}yi1H6ZqP=y6GSER
zcL|5pi45b&6q`c{&uz<5imzdZvpzFa^zhcV-3b@pVW@NDPaiDs2$&UkL>S%?qZnXq
zXxVL$>L7snNA`D^To=2P)e|b-jq)Cb!J;1rm5cs~H~3U|=PJ9~n@I8q#Zg3a9sN%P
zIq2K3Y<Kwl?N7Tzyl}H`gt2>swaM`rIh;3sxY!=>pkNVdXu>BVI7|*7E|~ZoWaJ<*
z4m%$IvU|iTg_a3W@f~ePILyGBVfkNwhl;Pi{h$B;ZsRTm=$&`~gai6UAb?=pf@IMh
z@8H$AOBqqHpsaxYgpy}E6f2q>GXBRz^~yX00vUwpO!nACZThWzgAWfuBEkmf7?VBk
zC|5qvR)bCxegsKCGvKbVA3CFfm({okg4^sU)0;bVD%`^}Ga<fe%6BVqmjtw?r*fQJ
zh%TM6V1pIRH)(hOFl}xkU^&P_OGrIDR48L+d<->5-yooX-+)d+e={#47;!WZJ5Asv
z5D-Dz<ms}{%+71_%h7EWD_{_6_!v6n0lvU@nI98OCp6*a8`yv&D16v~zP8-*T|(R2
z2=i{!@}0Yh{U0c8xEJBP-!m=b$USyhOo?9l2sDneVM(ddf};DLp()YKaq)U*r)lfO
zW_tM5<FtQpl&0r69;ubMGYI~MmAIFer>Tzc7~!~J2v0jBew92qHp#*o(RhwhWTBT=
zW2eb9Uu|)Fh<CYhkK0V8o?44b7C3Rg3NZF+XYmVOd?UWaJd4K)j&J5072=KgXZS2H
z8k=E9`(&!kBaoO7UR_xzNt?_oPGwqngn&r=N>@DBvHU8Pn%rj6#`dm-ezX@%KC9yi
z!%lp7plR{mD56+Y2ERE85{rjKab?^WOvX`^<lO=st3#RK@mWocAPQx}LZg!IE-zT|
z#qpis{{6el>L*iaXO}$BG>^Dc`B7l9OJ^o=%pw9lc*DDEOKEwrE}pUPJMp%6cN2FZ
zu{!#PAH1JF{NTORXjH|!UZwsy_oAT0R38=Z%8oCNKeQZA#Ye%!T{zqyam*2o2+Vq2
z!3*Vb;O>p^khZDRT2`jQTbA*PHwq*c@FCCx{}IS}_on79D18w8vdf1U@MI{lif|D(
z<|FW&z>mJo4uwXe>3Xx{AAv993_BRCkW=5GeBb~p-x+bOT<)EueOB9Vy+~&WX~8u0
zQI1BO7w-CaE5fS`R^F@3%vng2J_Vk66~Ee;yy%>aCS&6W`$8c{n;_U?F2w_m!RH_B
zZhNO0N)dL4Ep$5R-rc)trPE3rqe&Oz7|*NoUV8fIq1tjcz4PAt>CJb(XB;1c$GyHt
z6Su<Ge%HG>yFJ15M(L{%vw|t;f3UB7_I7<-*g{+L!&Xc692fp74~;`Cj;z%i7EFRO
z>l>TKJ&tPPxxN<VVGN##U1QK8;7)z18#uDBe%(`_<mh!+5-UK}eyrH<aJQH0)UIg^
zon?Fl_S}ioPh09cn_D~X`zRWiC!$1P7wQK;{Gs8-9D^ON>{LW(&rylbo<B?bsuStZ
zFKGjg;zS?>ZGc9AZbB{RhoSG|5LhUqu+)KL;HJhWgA90d(ELN?XRM4}8KF$2?YL_a
zB?t=#m=l3hDDv2a!W@IVfDJfK*+OX*JBJwC$((V(a}C*<|De=_*GF&-Z2~_$-m31<
z4$`o~Re<#z3j&pEL4RW&R_2T(isBJkf{CgNC*_Wnv?G7QD!%$fDKB(0=pX41P2g?7
zu+UQSh!mlv2AskuE-;oMEp}`|&k1SzPJX6e(77lxajApnXw-vm(o@p39mX`<c~QNn
zH}D%8%8g<@ms5FrPE4QI*oh+Rz4zWzo7F{=4y_<!2McC?ZdSOqlJ4ER<KsFvH@D1F
z&>tB~z!{XJ^f%zdt{m=~qkhO*R+i()QsBW|R@8~(%s4WXabkOSFYTcG>K&^daiks!
zU3T6uk32uS6nzuEDsAL1E6Yow=kva=bANZ!<>#)b8PS{RsX6D%*iTtlsKV|l+z!?M
zUTktC-LY{B-ehcSQvFGEM12NqV8H-<0x&*5XE%uE{i3siuy>oiac?c%UG1dR<)-=)
zN7`x}5nsS!6Bcej1KD-W*vIk8*P?ga$poHyF#|^&tG}ZJvC@Ox*`{Z=R<^_ffpN8s
z#s~N=wJVB6j@5&IqHi(wv#W)LG9uuDOB}}s-vPhUs?S<cbE@%aSNz5H-l4`k(I@qH
z-~g`zZSS)fL~uZX%DBMpnuS(XI4PRtac4Je?V{k<n4&g_d%%>2{>D)(C_i>{ajYc#
zRVb}Q0}Ko3p?DI(0fmyXHGKnb@uR#PE9t%oha{Z!0x*qhEQn@2fbM};&_)dM7%Scd
zFL)kcMSo}PqyLxPhQw2X?vs=W869IJ{n`t6MRz?G3f}B&H~*_Vpnvpv6zblcCS0Sx
zaD*Sb{!qSf{9U&z{-)2O9TwBjhv1u`{i+M~VSHD)pucevgUU*Il{I|@xg)!_Xe$<E
zaR2D$=4M)7-%5wahk|!kIKu9Ac)wW4-(Kv9SLmqDYCGW+yPeKdZx%uT$GCUyOl~Z`
z5M1>De}<ZhcqVp{HR`JKtoYCg@tAvC=@&o$S^E5oFTBX)-M8LSALr<4;k3%mqH1s-
zdIInEbbZr&md6a$lZQn>z%B3rI$06SlX#Y%(?F>XE&R9t&A&+h_TT!~)Bg6R@HVBj
zyYKjjW6dPIu#L819GQV16*iz$o)%39)+}H`xyL+!-GuaKl&>sWd8RV0i!OrG@Mz3G
z*vLS?VlHqlUIg0t@yDN}Pd@!DZR{Q;wXheXy(zx%{kPY|?=5J&SxDF8GwGlI%U`Np
zdup3S?`i~ZM#p;N4sVc|TKgRaJShqh+>1{M<M*gNo7dS#l{S$s-@hf^@Y810pWZY2
zJ!(u3zt#-v$Fz7f)a%k3HLp+ix0`R1b}`=+b`*wxNK@MRRfFPd;xB&69bb_#`9~oU
zsOQ3DH@*hS6#d$__~khM#qMzN#@%low=7mM81{-dk=HPt5<b{croQ;$H1x;L*Qh-n
zRW>~EiZTzvjM(KDE_(aJws>WOdqGAzRfOe7WsWcV`QaNUyZLwUP5B+(;p`OYV9jwY
zz|i;bXs`$y44I?Y04s-!lG%;mw#f;&k|uWC3OsBN=Ll=N^5Ct<JQj}SogZ*#@I_#Z
zUWAG8N=5F#S*)H{j4>$f?d*BWGb<?BB?P04u)_j4Jzo7G1_1^tGB5BASVKsB1dt{?
z7KDh%G4X=<!5Er25=W7kun-QWABNqlFvK(v$15pa=`jma3T$vUq35w1z$0KVv6LkS
zQV8UQnL!hO)d5C_avO)`#y6|>Okne*n4}=|td>Gx)Yqi6FuH$8<Ff0y5@SSrP-a$R
zx@j^@F@v_w-?k|j=n+PqYKI1Qfvl}*l3%g#*2*3+DYW0M@<7PJDsd9?N=+DlxjP5)
z7IVS9U%=2A>VdKMB!{<{8Qf5OVM2%}#(+42k8w#+(aDtw29KhqSuuR=r6n;3n%IiD
zVfCZW=2ti_PKTNDL_-O9WuENd3qlxRQG|(1mSAjvmnT}zkO|FgltmLjCOL7J2MkwC
zu1JGP1Nkx;gMnj0==#eKA&<*!SnE-lRb#aY6TrYVPmq9-;&W65%ownN@dA%{z)R{(
z9-crcMM_IL3=gEwe<^pYY>de#^)pWM5KnK?gE0W69G5pYBj#C?U6gt%sKfb(vKb!&
z|NY26GNn$=4qM>9e(dIO;Z9JxbEFXVu*sCXY{etSjBwC@PTOHfoBHx6j}Y8Dio9^c
zj;{=G#sMoY08hHvEmD<n-DGnZr<~s<K2Ls?xX~0gcEO#0KM|(1HG%CeAUbutRraVW
z{)COU4e_$oli!hV2LzT_?V8i}J3<W=fDs})8$NchH@>6-Jn+Ro-h8_~^Y^&hlk{+t
zqaOd+U#|S?W4Jq(Qc>=(I~Z>w7#_J?`YY{2B3LPd{qj399eLX$Jnfw0lNLd2!;dmM
z1STRI9{A-rC9PN_li`7r{UTN|1(y8eR<t`Wmzi|@4rjw7zumuaB#)9PVg2SeB1I|7
zpFiyHzw%Zf*IDm~rF!9~kBH+)4qxIHw-~D;Y*wG-ei7jUK5~eH4AZrL4re=#{?B$j
zO333h;cyL0ga^J6(gw&3(r`dM)CWL<L>UQ&G5qN?xf9B%WiP+TBP{MOKmQJoI{ff&
z{vSS~4y?>%#U(=I5E@oAo=%IVpri>NL1^G{H|ngq2rf*p%|qx7eT44ejvGP>1PK;G
zFcCZtL5C;m;y0Kad#o?W;72qCIwm++0H<^q;}HgsulX&JqKoTWO-Q*T<0S1JbgjhN
z7H_txi7Lu0c0_Re7^^H#E;I?h8B3b0r}OKvbaZl_w)fb*-_-<Lj^evG5_o35s(j|W
z;DQy{w=87AuDY|z9G=fg1{7!K`8eNP=O{WfG3ft|^5H%MgaIfrI3AnT<*Z0Y(6)D=
zvUSP-+;|F{Ii_*WLaI1o2W1||Dj}#qc*L%(E0yWMM?oI@f_Vf1<|h@`7iDYGjxc_D
z0z6b37}i0T=_@FA;2l^*v&FGJM<_#t_q40p;zqp2?e&?^SbP^s2rCW5XDD5cVwy*p
zr1%isD!c3yZhLdf3WsKE!2*SOPB3UJIP3(pSZVt%Fpkb?a`y;#ZB#65fJXqGz$40s
za&B^5cBMO+uG5U{tPq}`zD+Y@7inB&yJ6810>OSa9UbhZ<0HXFxP^dXQ?T9K-b)8Z
zr^1P&^kQo}?Fx4I`F=UI9Rd?pd$V{5MIlNr>gY?|h4<hZ?O~n>{#;?$eFLf@JcKXe
zo)>nW!Mh+ZN2t_kH+>Y)?9{m7i%^Q)SMc$@{zW?Wkx`+DL0L?n1IO7p1uo5Vj1)%^
zsg2xjDldJ8olz{H5{MiZq8#`s*ojx2oS=;P$Ugds+6kc(cVn<4VRQXO+I;aM?W(=^
z_IAC)1>qU}a$%w4qu<m>N-z2$cnseKf5;Ah`W<~1-jjOqc;&cY!Ongw4)8nedxi7l
z7@|a#{gh+jca`V1m4pa9xnGwR)JnkZjdGUo+(XA@_}~?|@0|t+M+F;p5A5&lDIbph
zQ5@9~rE;}a_o6`rd+Z`WP)$2psVbObXfri~+Y{jx%23)Ffj{j6TWf(nfmBcCQlw9P
z5jZn1@QzlLp5BpB^^vHHi;HQlGUxcjMS1z;=`#!5n=SPLl?6T;r3m=VVyhQwgO5M?
z#N(1MT>XARI5OtN1_yil=J|j4!3XKxci&c=TOT72z9Q6{Wru(s@?uveib0Ny1%{_;
zgA2j_=K4a+^+jq4HxZU{HxP<9AEhdP5D3^pM-j|jsUM-#<`_A~1uK5kezZA#97Q|F
zZEkGs_(36oQjsxxeqQjHt9U%*^bd9^9%w$ep*lW$@+7T4dzOwg=8-0iz#PImQYSS=
zO+_0q&btBvebADi2h0t?ITT(j{N&y@lpN4T+R8YpF$E<QctZbVUTr*;KVu1kYXp}Z
zjfe7>U4ke!z)9w`2t_?_r|nIvlsCPAJXKD{1@Oc9@OW%ztfFku4?{`M;!5U*q=)Jv
z%I-OhG0-ly^8g2@Q{1R$_T)twTyN#gF4%AsY}D?mnae@=vf#(N;xhgcm-1i?N8I$r
zl=&8^M0#0v0sT^6w7W7xaZ0E`nN?q!NZH!ozj_e@IAbF{9@iuLyIzrk%V3+_ik;sm
zA8FipR}H*^=X}Bo_#_pN)SVZ*nC`4Bn--$LbovU**uzdnmE1fBvI*KDcCkaJ^I{a`
z%Yv3mg*gx&LytMdhW15f!k7tcL$(q4%pHtWhQ$%cU}z80BR)su!OMVWz3!pfl?9-a
z>E7B>;)pnQr%+d9M(oT+sl-kWc1&6ECqL5l&P0ugUht~A@&gV7L-GT+SyYO$0eX*u
z0;MwLpzMrkjCt&;H~uLeeHz{k__7;|u^xE<V+QnvBci!0(qk`pEVzKn-faUeDF5h-
zQ66MHj9uqveJ`M4M+{>Krvq>wT7ff1&$8Q^wAr1^9ib?v7)xmb7U&=coKzaL1xiK6
zkE5ewjV%YhFCMuAJHVEf+Ns@c`RGJo2EX9Ds#LF#fz4@LL<SbhYZfnCW&v|(1Y^vJ
zqnf$bkTO#)lxrw-eE*kV9x{{QErB^y)i4n(fFt~X7gnhKp#zMQ=9Ltf#Us=U+RdAG
zgAWF7p5tqLN2$k{2VH>%f#2XY<pr+ZEv)$95`0=Yd4WX@EcoEgt|P&PevXpNcT@>3
zzC)_U{i_wbqbNfzNIviuz@1|$;k)1~Q4m6t;JM%%z!lmhXy9$?D^C_%Q4SP;G&~Eh
zSa4LCXSZ?N@=6x&KnK`ud@R0%y3>C->7Y74ojM$u*>0rHLL)6L)Y4+BCVNfxdRkDw
zn3)iKL}O6oEp<9+d3nWm)!x0ck{0#NE>2EA>MSj#d-v`q?i-_Q@Fc4%E9uQQ-$-xX
zzwaH&$m`S;ak7GF7<hximE*nPF~Kp~j<B7Drj@;qzI>RTK6xsd#ZjXBUVsyFs%R(h
zA=>CxeELrM<da{g2VZ>Q<FfZeqhG9VrY{~mO#ky={w)2AfA-JP7r*{EU5F1wdC0Na
zESA5yK2N<P?#S#~DL5;hqqA~1HQLLmv%H$#ef#b77k~EG)8G2H{$_gb-S^YfY$YA}
z7<Uxq+{vakRG;@)s0VcBF@nkl?!{-k!!2;qUwRLGM#GNUv-$S+;WuW)JM5zlGNbV^
zuYMQvI(PXD{P-OW_4+tN^KHU>o%vnjePdYp|9++T?dEm)#!-j=*S{NdL}4t8kUhdP
zv_Wv?j(<4mKN5>yLWC`R_zpkZ%QwbpIN=<^qh6%uRPd8ii7!snla`3W4+C|C85n&k
zShhQ0*!6PC1XlnAkzRdHdf^#{{BgED{KGTCglwh+4l{%kVTi93$%6<b4b{zX$%&H!
zY>UH$PuK`zn}dy{E%$H^Fk#_G0W1X0aa<RfocqWZW)%8-g=>L>z4R)f?5j*mj1Usy
z8u4r!DGc>-BH^v+n<g4!0**8(Ti@77Up;!1HaFJOsRn+SMvafodmvmzv7tCbhrP9t
zJ1bNmPAyIRo6VLOkfsOgnurm{%lTe|swT}M>Xa461BM4e#$=WvlJ2w!6>b)|>cNk#
zz7R5$Q(oaAVhfQCrhv;&J<al94x<TkW5Lr6La0kmfZ6?Z1tX^i1{7n)MueBu%-s1A
z+rnct)AQ%+>A{09?N7RYNCVW0#y*l3Y}L_#f!oGUz?(nd;((5VT`xK&Fajs^zWc$u
z>BB$%qtxbzBQc4mng~)q2sanCax5WR{=K?fY55o%aVQqHin^M)5mRghkNyxOkv9SY
z7)GyVg9?iQXR_-fA_YeaSrm`9VIqHV9>+*pdZ(gJ&QGSrwdHjGoww8CQb#z#q<z9G
zvpDVnWfG~91`Iw&Qj!lxJF!xQ2@M9jSI<MhWyerXnD-Mgm*6Tpz1R*OtK>vbl}@Z;
zLYSgFRHq9~w9CY^Fmq;-gf~n+5l8|r@QTTj1*XDBm^UUN;FXn93TH@Z@_fyGX(}d+
zFE|EXGRYuK7*)QlfXojQ8rmD!Qx+zslm`aIN?rPz>IYm%=kgp8<z#BNmeShFYFel@
z<sacxgust6F{O@0RrMWU8ydV`a-;P0VB-4ICdQRW#|HeMaxb~7%yM#UjSY6yH=%6T
zPfT0`QHcqFaCtdxec@G{NC2a6kI#<psDoGzTQYK(oMQgYFMgeN*y7jKP<_JWyk*O=
za0z$497jL6+~|k56XxFu#b*gq>^9j>yE|=VR3`F#$5<Z0qK1U^uie5#c_Ix=hV@8B
zw&{nsFNg8RXsGOu3owh{7*fK&yo$je{20V3=_)huY><lLdDXYwUP{8>k+y8kEL?Ca
z;p2-m!UjBwPic`bd=W14^oK%^rh|!4m{-z}v*VYz{)m@NG{7J=;?ZV|zWh>u=jCt_
zCQK=_tpiKsDhC4yH@*184JSo%J#2E?wuJ${<Zrl_43xM1RSEkO$L%2}Jou5HnnNb1
zq<B#^{bazE-<)Ll>ESp0%1<^p;xH_*vWfhQyJJUUE{_eL3c}H9E?0K=um9oydzjF_
zV)K9cPyWB^LQzr|hF`;v9RVTSX4e-hr4eo*#9>7%lYXibf=2ZObp_}O!fh)w7?%~_
zbV0Nb`sJ})G{(k_5JEuc7D7zu6BBU@u9O!-e}FW@YEFdZf*b;Luc{SIu#m%fiw-Lt
zdZU4=v=K0J7YN7VAiO|Gjt~g}5W8dGd5CYiLR{XIL3U`oSG&W@seTshaI7!1Nq?jP
zO}et6C1@ltP$Cvs+OBYgFH+L_kzEh<V?vIyh5JNMp0t`(uZX3Nv;_Hc{9#SJ0IR~u
zf9o*r42p3QVY%uc7**z{ec#pcLM6>lv4AC|S;3K2-w65G;l@g9s7$~*ggOYlq1U*L
zjgd#(BZP2>opLvtz+0fMyl)hSOV$z6QzeAvT<+aFJV@*7&%HALkN(=fmO88V(y4fy
zeeuTJMS?&HK81QBe4<|PCQxfujI&dlRm0{lgcsaN)|lgX#R=uZ5mtv*+#Kxfq?6uZ
zx;Q<t0IZ?5^l_XjG5K&uL#4Lp)wG+tyJ>&#NHw{&kd)QnOPvJ^KrCoGJx_hXjbnKj
z8N6e0nw9gk>$rt1zy+QTL6WL&UWen{s3$N-DR5_bA>F^*akHZAfNyHoE9uTsJ<Uwq
z`VNV0m4nOm5%TXI^wQp8&-@Pp$9A)pI*W_$ThG?l(~;Vi-CS#{i!Lj>0Csm#<{qS@
zBh^QEw$N&%wYy6`b`WJq+zEpcGB6*$iCzUhqHH}l-1kl|g#N5PhBpHjS;4IYgoo2<
zv9pq9D;%+ok~a&@6`wR1+c{zoK^$>e$&HYlK7(MFU4TXSP5<Fewae3P8i(KGTW!P<
zmnc8)-g_&(_0Bt~srIn4KzMQ{9(QYNU1i)gKEW%m-hDH5me<r46INCt4Bgq<_FXgx
zmB3L1dF=E;SnIJuxXg56Uh^o@W>~mJX{i3mSOIUz_~1Lt=3`d_?QMah;$y}#-^>_6
z*bDV{<^(KmM0r``NJPPjlQ=?vNebeb&qMf)5ECWO+${Ia(FfVJ#vMMIv#GsV4Sue2
z9(Ioe|JhWjH-rmG<LZUr%Dq?|Ejq1!+!S6gR;f>3-llF3g}Zouc3No6K%hM{qYMOh
z>VDkqT8PScH9a?{F>EX?ciQex(`sKiSU`Dna1^jmKklEP^g6c?7)3P-IPTa(0motl
zlt_$IzSm~&Q0}qYg|_z{SOsQkV`zbo7o*>xR03C2F-~l-!d>Nz@lZt(&QS+;Vxe55
z%q)mwZUTOJj-rO-kG{=OZ4*;7-W4YYjr|;@cy8Jf^Bn|(tj>?Q5JGnrC!o}FzDmE$
z8LYTfotanC{vqt=3V})Y?~6-~=L!Qpd-15^(FcIBDYM?OP!d5eN|dtWi&HZ+mWeL;
zK4;az^ixlat?G~Rhh7*4N*4o6yt`RMhO!ctbf6hlzEMB-N4Y6Ya5d;GX+m3~-_f4n
zcgYjl%RACi`SlPFSWu4WuaGJbST)nPAx?ayP5<#iM>m;%^iTPUQ4$^>BHf|;ZyM4J
z83c*&3OcR6BbW=JT^{h~7MRLTyKp>VquEwi?zCc82jk8~YF1{`AOHA=@~@=L?X7g&
zW05L~uXCS1LwT6n(jMRl3K<0W-~#yUJ4RbA_dSkSrr)Sk8WTigy$CcGAuzu_65U~O
zfX8Z$87xShuY{5v0W*Dp(_Fv}?m$`<{hgb-_2P^l{@{I$jkBgpyBhBwKHc<@YS5YG
zrM49o(2&^qb!wRqav<_XP|i4ltOa^Z`rsk3q&<QUK^f1{xXJhHz5m`j>HT-#QaLof
z2o@ZZ%&t+6j}&_ONVo0%L*q%ghige?sLg3SVJr(YFphJ75`Btz>>@i=9iV$#8h_WH
zeU+|G4^n-8R{1V@r|_l5=st_Qxn&933A|P2rKQDmcWFT|Vb|D2da?dOwCLEhfL%5d
zUUb0yWH*X);vHQa*~>V?;^y{3BmMAuZ+o{iyHSA&N9i5#h)$k~Z@5){@Qah;86bMd
zBw&r`lip1&+62!*{gA!+iU!63RSr51ty2|!Wc1<jS?Y>j()K``ekU|FzOo33#RvUf
zU-%Tq=$@+me3F6qQ)Fl;^XT97UGR{$ibYWBPZ~?;%PdY}XA=fqn1xh~7xWGK9Ag;D
zRpK#ro41v*a$gU->9~Vy)kj}X3di6fILdV@otz<`JkvYwLIdaBH`V6E$6Ce*AFn95
zLA%Hs#USP9#3Sl%MVZnM<tcYmA^SSFqRYn^iU+(M6OXI0oHp=j4<dMTf;Go>gC8g>
z5#Rs|@PTr80gdwOsbB5y@2E^1wOA89sHm?UrUzeqnjSrVs<t>W-G`Q;=<FZUk5AJ5
zd-u}2@4o9DRp9r*?q2$<U;JEY9i-O6LRx5d)HW!%uf3xR`g*9gh0j3&TbsE}i!~NR
zkEbJz9o#MZVsj@=qm){`liG{R!keo5sc}GXIX*f_Up@RHovO_AQRWQm8{#j8tK8{Z
z74H4y2XCc6`N@yd(&~y>s8E{C%ueLploe7^UlN}xWZUIVQ1LUsl^x^2w#yNy+xzMH
zi=A|=dJ0sA6MYh!$8v0LKT5_pJ#-5be&y3)^Sc<^aevh<n^8N)^5vJzNL;-J_KfTL
ziSKMi(^HU_zDG@|%Ws-tex5qn7Up$njhd1!dH*)^ZPLP&P~Rqe$@A5azX{Xd-vllw
zFLmx>_l~c>!HzHGgYZ#4`d4un8EGTEU4~o^ed(KUZX*9h9G+L*V9BuasUKOGcSJt?
z^AmCMmjIz}EcO#aWV@|;h)KHPUhH@>dB{O!GGbFlyr~O@7vXTre*<HyJ+h(fyhuBj
zS*FA(!T5F!<nWT43&4OJ`4|#8Oo5Xh|3@AUsdw13!7j%FKl|I44sTC|kH(A;4(0r@
zhf6NKet<7N_N9G@Z!bbPRM^T7%|Ti&Gk$JB+u}cBhA>ZrO2`FK-`?Fe5kEfaIt`4O
zIfaI}wU;_EVG;9>z%5pHM?eT5t4x@%!HB_hmz6{&GGc%rjxdIV^{P{dh0-5ogmLo<
zCA~Sugq0`|KbTf(O>V4YH4%n*sv;137$4GxD4N;Sx8b1}FkQYwiTKzRo^97Y{=wm6
zVuHI@){5}6>WK@>5pMlqO`b4N8`Z&Ogc<}zuz!3D_Ob>JWkY*{XvNq{yzw;OsHR4H
zL6fLC!;~FgFp`mzn0Ud;t0Pq(6jyKyq%Uw-;pL-{M8uuEl4Qby)omkqv-*gmTwq|V
z0Hch+Rxq=WP<08$s4zA7bJXEvs<*h9wUMe~qAOyqnY^Bev4**JFr`(>O95b9$O1o4
z?7$0^1EnWQP~R&eA{<PZm>ZR;D7+{e?TPV<62+w=o)l3mB^v#QNl7UB$iu5O^`=^O
zhlLeWfuC1Qs%%9;W!#35BaS9BFz<|4Ox|L`7Gs?omq&?)5)6ixz5>%oxl|6p?ivZC
zJdER->^K@F#ZBn~SCXApUbXpH&8rj?hES0OunOfwxbo|Fd}Mbl;fEQN5i@uT@D3AC
z`B8KaAyGEv?ePSA4jCRpm;4jX*-syV(>yLPtwE%L8GPmNYMrnXQBm9<^!ShZV#m$j
z4kBw#8-K$I-<(c1#CHS(Axxx$A(dRg2$^joKm0p0OoSl1AAt38fH2vavhhdh9Nuq#
zxeW0{SlmicoNjjaL-Kg%z%S7#|B@yUjBn_RvBsZ~Pk=>#PCFjE+nU|naTF7)-fqK4
z!s$hIb3WOP|9D7S@6kwuUxPwsG*3SH9&Otko^QQ~Q~CyOQL1<w&vQ8HZd2TF#0QJK
zZZ+v7o}Kjzn7QlAaq+i1CPGBLaysSFkBfoL@Aw;q2YybCxQ>zC9iha?ZuZxc>xMl$
z=%o^t^6^Un9UxxjUtDNl7mg#L<whCs$8JykXE!=D|N3NXEEA0gmNRlV`SR9>U&Y;y
z#LU5pJ89|9?yl2q`o(v+<Ll4=z5nw^<j)vN89Y7<pG&>GQeXT9yoy)biiW^%gs?Hj
zcjy&kGyM!9V-XwzFUE7ywXhK$Mi22#AK+bbnh2#)G!PHkLTN}ZXbA%8SYY8@$|{qE
zc_gk(2K)l`7(C*21kX#64tCifh~jPf4oxOLu@Oedg3uHGhBRG91O$QsE5}(R13gAq
z!GabqWI?biI*5=W%0ZcU1uqCM=L^cKqQ|^N+r7}KQA_nYLbzsXFElO8;K(Hkf$*3k
zm3;ruEUS={C)fzko^S1^&7FM<?NB}<XoB_vA80ss1pO>>!Jm4u+==6X&Q4Jl>McKn
zZqRg&azv0%8?YPh_>^6F99uZ)om`FToOfrjL&r*knb}mUaI_r9-7Tb*mDSYJo8#B!
zs|q8U%khG}e&0JS5X^Bu7)MUF8}sR`caSbmdlm@H!zU5o`ci;DPdAsRY4Y|eRj0;N
zt1^@7igVUKN}vDwm+9kQ{5);0KTBP9b{!s>=U}xr!nWIS!Dv#nex{Z#ZYR=VpVj~3
z=~VZ6y=EZ^N6EbV)*I>W>XOPor+Oevhrj0T7}cfGun-&=A(TKtKs$4413WLvR|H-N
z_u;>IIQEX?bgnPBLk^*i%0yq#;DF)-g#yAMR{MgJ90!E3jysqvoRLAH*%E9~a&d(C
znc6OPc_4h8R6ET0&6tL=(-cN|z`MeednciA0VM_s3b(D&pnVbK2QL*oCwOiA7JtSr
zG|I)UA^He{GIoY>3?=n64rV-IXNA){k8j+|&PR^F;{FpW#grEOGdv)7-H|^^xf$Wh
z!qRfO_trb<-n}=iJfi=i9B0AQ(cxbD^w*!Hjg57|=_V~LbDY_m7Br*8qr5(eLG`6C
z_;^+Ex7AAQI;1Zm7~}Xe@@0n>^+1{E@_SArei9y6I1Jvivjl-S{s@<SCzjfW;hhnI
z{MaGHCHZ}&3l9xGaaRsIz8Ip(tLJ>!fyUT3A@>+>P*@^FrF{?<?f|qvOTQm<Ev)Tv
z4CJM-VOsOQ`G)%Vtin&E8)a+?K)VVzIG&6<KNz<rre;H!FFd$WKSBA9kdpi5n5V%{
zlMhFRakMCR2||}p#P05FrzxemwAfbLTzbLg13f5MnSX%uEC>KZ+;OzA5j!K<0lKAe
zW^;4bJ10q(qfJptq8ws~qr50i5W0X5%!hm3t`)`zG&u?i#Sli|l3}clFTZ5U_-nzi
z%EFk5klA){pIxu)@{;@5cqmvyF-E%3D+J9UY_@=zHdcbb1URzr4mv}h!q0**{WE^V
z?nq8ti5-CRjBC8rzB34@l_z0IpHecnDoUtO_9!FB8}2H9D5K;`ouZ=t4tMCH0?Xh>
z|CX2R@m<{YYdd@g!fL<4Ejy0*IZeANUd-<bEW>qZ9E-Fao>t?tq{UBsM{MdVx(Gc}
zy^9~m*`qWxUmz1{1Pmh9>x|0odJ_Ya^Hup>2S4gJ`Xf&;gFH~q4#qNe_M%Yqc-vQd
zO{c|nQ{@ZA5QedYJW$#iuhn)e&?J58PMXZQ8Ot&F({8vkCKAVOq;}zGIp4P~S`EJe
zTp9m;pP~9Xc|aqL58_osd{kC;4_^Ct($$qED_btlPc*(yqy^D;cJH!yAa;?(5r^zJ
zirt&=9NeRYl3aWcxK973-Ia&$vSa*KSt$!N3uOR!%b0|ctE)6o#Bo$$WxlGk1XqQH
zmqL~RT|zm-&LXZ$;DjE=e(sH*p^quexUNI<Y}F0<4hmv;93`7h1!v?~EV@`kVJ2RM
zIwJqj)ZFlIa@=5}nwsKu%%iCfQ*RAgR@9+50S3=S$59-G9BkH$SU7^RA)bf_#RIrc
z9|rF5mndok4AGg`{lpGnFbet!EobLY%L~XU7Yl#Joj$alvM~k#D|k%cK|MJ(n#B?<
zMxY$vqnHtudrV>Ad+OrNT0VZWuDU_r;q|~x+J>>pIILKV3*Pw&e=K;i1CO1sDEB!s
zlQAsDH}Sh_XLe1YobwJa_jAV47;E(J_6|F8IPt221#K@hHLcy<-Azxv`YLU1Y`CAY
zpo^n!!6z^Bn&&RBX|>a|>(>^4xu{3fDV_GjKaIr#Bk+@lwnuRVEN^cv(*EvF+I;am
z?eA@;gWavP{`^V$^wVFZUw-mQdj9O0`5^i%R7-?K4<`eV_F?KD?WVEoGv6J@QQNG~
zm=R2;H9sKSg>aAs7sP|7WD(Nt&Yl-I(JsL7NN)Rm;bEFd*JCr{E2~~m)m~cD;63X*
zWXDB2n~SSywzeSNuaa7wwX}HWewwQ_)6&X4eXpj@y|>fS+FNO1`EFXd`$l^6o%hr9
zTs`$J$I@JLCAAffW4JlxVQgaBCqq2j+)K}12)^q(X?<(YeD|Da^}V;=7ay~h&aTGO
zLH|PeHq&BfDYZMD^x=;_Oh5VKzm`_--u02|%Zohh%4=Ft&OL2veD^o{G)nnf(Q*{#
z%q_lr^f-O>__=rq6!`*!;84bQzXx{4{h%XdT#azWpQqS^Zj^6KN$WS<vKh5wUj4=l
zoMajJ>lN0Q!+)FfFt2kbz2e93XsFl68Jf{Bzm-mXi?jmjVZKe?#k?BwcZvH=W|-Cx
zU)b?A@fSb+i;ozvaT;Vt0FZ_kKVnoJ31Wk{fAyAI*a(vC*?}*ZU#A5~MKBSD*vKi8
zi##F>FU2CSp_9`Y{)QQ(ZF~Gs14>tfb}p%zrwC}9%g;AM67S;cDtb3kIA?dg-Ix@H
zf4@pTc!V+Z27dBY5wH`4((sUutM7;iD`b^}g-jW)1LF_5(Jn;Fo}?ZhFNOA()0VNX
zz4gO?whzjzAE$y{77q>%Q%qshq6Q)4a>OVzcKthz9JaWj1nTy>R`l#_Zl-gHul!)9
zQ8Ez*=7r-gU{+ae3DbBZcM>A4=~<5A;CP5wacDu17<^X1AyhCi6mf>tW<@aHg~^2(
z@LBI7uw3x#iLt^kIVKL1hY&nPc!(zxUo(_^!@($J(zy~bg%LC@k{jVXNru6apLaIN
zKUU(&kJ8JZOO9bSeFun`y~BgUbZ~H>Fu@T2L5<-C_5okYDF-)FY105IZ492dwcsA~
z?=aYlkv|6E9V=T|EsBtcmD&&<-}gk_VfJtn!vzCq%u-sw5Mp&BMv!oxWSHQS6)O)z
z`Qu573QM~ps33iU!RP_EQ0(Z@<Wb2OZ^pzxR+}R1OUtR%Zl!9yVuq1+u#ippT2ZO;
zAXKpO5+NgXCO^QT9~5CQRp60V7KwoXCNNJh>A1z2(UdLl8T9wcCe^|9(&LGg$_R7s
zyw&dbbMdV4NReZb!u={R&LN25*c%H76vivcVFJN#R%|LhX#;B%snJLQN8>hikqhxr
zEaK<!0UNM(xp1Sd(U#PKpIEgBe+}-&1hp<^Gwu-x{2?O2ogg;8!<n3L3_slBE5EZt
zj+|4m@t4z)jWYU;t6)JK$BR^)Q89!m{=D=}s{Sl)dPpxUQ9|nNy8A!=_>ZqBS9s?z
zdR!*^$YejeXD3_)JqI3?*`e%YlV5{yp(rCAewET7T)4U1{th=MMA#h9dFQx<@s6Qk
zFo(h2CR_Xp&ur%>>cJD367dG*GVJw~czoN>pHUCH;o#cY-)ZR|!8uX+@Cy@OgC~A-
zI_d+S2!w~>g@UN$i3u7UFmT$Xj3oq)a`Tsd945yvZhnyl{>331(hXD6Alxwi00z-z
z>>6)5*e$+?>0rWjP*fnKG%)dt8K#4m%V&5ft(+MCPAC6He+eJbCamMziA08B__PPk
z{>vs?{3><Ri?mz@J2OGcu~eo~c866Q(+~Hth;OfK_ZNNJ-#A4aJL?Hs)M@aQ>s#=M
zkOj|%JK=KLHnM4aBWuG;KLr29$KRde>)k*3_di0IfUx=s-oU&7!XQe|gqty2Wrn}d
z_=-Ri{sX}ZlX!#)^O|rYz~|C)3pMh?f*!(qmy>7>N-b8DB2;2UCl3lR=n_H-Xa^tx
zR?vkhO~|3I92dh*XcQGb@&<v91vUsvLx};;WJR39#)G?fD?L71OJVG<vO}+39_pjA
z6+xJH9-#bEjm;B?<|Fh+=wd<o2#la<Zhe*0LRY<^l~zWnPFB|G4d3CTdBp#56w5rj
zEHvqDHtMNX<4%Hc??7Xx5%(mZpo32!ACwX&=U1tBdX<iQr=A>BCX_Gm1Ok=AqWCyO
z$lg0kd)<??cXXPLPsPJsvx~_n=C)*>f>4@ru?va$5qE>^cTdxi@`fgJZxM>8X?LU$
zx*@c>LZ}Vjpn4*#M>$ui*3-h`Qfe-C(zxI(3M{%m;rq7gN{1ut1hmww&Zhp+Ub+>J
zg+Qu%xS#fSx6<+cZt5NEq{HovbaJ?xsx#wh!FQ*eriTx{NS}TFnd)$x-g@ip^rIhr
znBIH;`|0kzchY=gIo(dxld7pYizmHQ?Wot(j0IWsDwl=}|0c)t{a<Q()rZT=8Qr<x
z6~QX)$Gt%pXB>;xH=hO%1pd>;2qocl;e{*+M1dqX1UP3W$B8>k5JrJ3D7(@%N86oR
zu#TVuVU;X`8Td8k^m)K-b^%dVaM`<E6xO&YKL%+FaOY3}OuFs3zsWm(<OWa0t^kxQ
zKDsgUhxB0{@P@4EKi<)%@^~Sg+T2Gk2~G$DS*&FNwDR;x4@$=ZV&MV(iu>uf<1LO4
zjQ)cG?_w;C$M=PaFJu&(nwd|PW;?YycT!VNt-0u<eh~WjXi3$Zy0em>mEIibw9r{j
zZ@m3py7%VW;$bUR{xF9^k-}m$3w*`b)#`N%6zSjKKlk&Vi%;fs2GEt_L|H;%VY;lo
z>Eps&SCthWm%fXi+eZD@W0d@E;w~soe&9{Jq9o<0J?2`p3;Z!{gy3p1A34PecTVnL
zd}3Vmt}(%wHswC9p4y3HF8gO!>d&Es=RPd%g{Cbll^Kl(HKkdz;FQ($7uPHxn6RL^
zF{g4~sSlo?sGc1E7&|e2_l?rpKR)x(r|gPg4B-NO%DdQX2*+kDB>n7*hiT`amy~Tf
z7cNdr&#C`U`k2_~>)Tdxaa`=-VF>8Cv(NJY9_7W3CKdpogkUEeyFodc^5nSdotW$b
zrSGt4O<)Q<rWK%}a;G_H6MCTMS<ElN8OEfH+rG0&^{-ayJ}z~3t`fK^7=d5B3qFV9
zZK#k5g%Uf^V*CVW%3?plBY36{q3Gvi749^wHyWwlY^H_wqUK+VsnO)bm08d@>Z~z>
z@_YBNaJSUSxDSt_vKxQghlG2%EEv-QyQ@@sNC)G%GJF~~{N*=(^h@Xxb%Cz9US!9g
z{V7MJE2l`yKaqaPhd1SMU&wBlVvqPafow*@S#c!ed7*8C>>O?w@^yytMjBzFetrNj
zbx=KBhKwJK3A8hF#ZY=_4y^u-BI)9iBQ~$o0ywQPlle92b3zUULs<-sro8Y4%)`Ub
zud|GUaw{LofW6T2@+LjTUHUN#KA1PdI{*uICwhEU9^Cm==C6z!Cvs!w35!%v)Ui<G
zaDUr^@H<Ow@gS<N%FevlcZjNuxf7^Xn-gBbPjO5?{0%aoDM1I^3I4@%Y4`~DEAGca
zG1C;@Hmq1-Oy|x=XuXd=orktVM<#sqVzY)4h&!}!XG|JTXO}nPDS&Y({W*z+Q-Jn3
z9(sobCWl^7fLy8UfMpzIq02k(yzQN9<^}bzlPB&{6;GgcfCog8!h8q?rHbrb?ZA|A
zYyHL3w7s?AzJ*c?B@;WH*~!PQcVGbCUy7!q(B~*>>c<^u2m43f8OE`{z2l>Fe8A#_
zBaO+NuES!CeYMT86?5>Wv?q7c6aSpW9!eAX%Pw;6F}=~)M!mTw6m;hb2o!PN9WKm)
z#zKQw=*@A&EJ}hVAZuZKVcZ(X^1N6{_~P*ig&T{w;4M8fR2iX<9774;#A2wRWgJf&
z<Afghs|UD$FxJq|qrbDTNqm6vhUOoOSK)5ft(^mn*@DMO$e6$}c7xSLPilH9!l_wk
z?d?@M*xO3m8&A`<`W1K6p{POubB#h&Fx=nSNKYPrnLhvYS9(8`+d(?-AEn)0j<`KA
zoaWWnPy`|$+Y|q`zq6fQJb#*=Km96w_2_|~FV!yH)KD0B8{|323xS_cmD0ogx?V^m
zSh5T7(W58n`Sa(-^|kvbs@BqMV=2ux?}(3BN^`Y#8k?@BgWgqovD-^qho`A~dXsjK
zRfgVm+U}mGO}RbaJx=R8eSKe~?fzBT$4~CP3zb3be|&bGp6?u|FE<X;y3)B6ue!2!
zH{HMgrpEuJbTdAqK5?6RXTp)2iFBs^gxl?8C0&irrry;=IzGQmv-M`W|HfPCZBA%f
zUUk_{uf&gC-J}_HfclL3s9=s^%(zxNo?hH|w<3H1ibhUppzj0MM~@!M{WRo(84obv
zf<EH0HO9_jAAMt9^~d(AzkYtxEt^q0=GAY^z)kPpjQiW;e_eW**SV8k@#A+i)a&C6
z&9@1I`BrJ=H|E=>i+MHV?-KW$%rLFYkH}=j*H69UYnZzRciCfj3Fo3r;^EhVJN})J
zSrOTDkijeB;2!=a<3!5-`qe*fxR7QrOcB<h2HEnHt+r*4_;7fC7ANB8Z;bV8`NL(&
zfs0ayy4o9OUJe>%8l-I>{P>7e955<1@E?>JYko%;4TDpZMUUUulvlXeR?-F@Bgu`J
z60d}fdK6R2?0SWR(o)zGMvPOyny7@0Txdx%JjCL^1eXP7!alh`gu@Vq2Qmcne}v+T
zV^>6MEx>oU0%H+3G25J1%xXX;$)p-dQpSl`sVGKeFbNg0<nnZ%>8eQ+%p6BLz<600
zB;x56Z)Pqg&5*Nt4Lg$>uap)eLpr6-5J`6BFtGq8OjLQpJj1+DIm5w~730SQM9dtM
z5Ew?75$XUV=v_Jt#Bv|gfbLZ|^5ZOY{5j+Lh&v|G?qrz#p^PUrxlWYd*q}FmL*$b?
zb;FGU6zRR=jTFeise@oQFSv<#Pl_?~ib)Y<gk2oZf$;B?Nh5BmE6N-u-6X|Jz_0*m
zIJpsgnN+JSU<%3BFjRV6^3BmVOzg}ki-F?Ezc?OhDAb!1v$wReoL29yC0381oMNj#
z<%0QV@^yM9e4;ElSK7})J3W*K;tF1{%7{zjDW4fM!N~$R;Sum80m{k5)O89(a64Pz
zB6!05`>b{00p*}>FnRPJgg%tP0vV+nX(|s9a`I)eLBgz}qf43LQ`rzqAUvc#o`jLP
zOtfp1gNZ7NMc@^a6Ur9&>U-T3(u^T#64`Fb!*<H8_$nWaX06&tRrNLcP4rC%@}HRs
zIS?;Op-<eY8R5!fH<G3t_Ra6G-`E~u9D#4WNHfA_XI$g~%7*eMj8iEEz~E2NY%dA9
zWF(wDN_q}l;_K5xSoxMIUBvl}#`9Np=ig;RIgq348wd~bERTJQ+c3TH!2*v6Gw7ow
zetg*&202~{g5g!l6GmT7yTr+le%uWmKhiZAzzdvkv7^1i>bG(=qVSDt@r!&)nBq>J
z<jn{0H^xOEKUf2YaUeWt#J7*Twc8-LLxQOhFX|PZ1j?|c0dpFGCl2gz*j>&@DBNTx
z8fNGmVPZ@yZv5oH`o|sXpv=m5kXC+-OBmA86HUxtxX~wVu=8Ku5#nWdghw{G$2)%o
zJp6?zDNv_jyyEt9zB#Q(&|y4W_)CT$@e`l<O@N4wZ#eQ>#-#YlWs)^of;Q4Ojmj&-
z%yx&5P!6fjfCpbBLF`w`Eh{eJA12a~vH#e2s`y%Z_h0{ryIELG+tY*^<qAS4=p#a2
z=$9w176QmF+6FD4^o;S)7<d!-7gn5Bt9taV*XI*=U7+YF^o0jIJEP6#@R*+oWg3t`
zpmil0ieL%aMj4<>n*KX{wNmx&=sJQ?1PlmWL^or{JS+KwH}Z}F^G4(5eL?`J2N>%c
z;W<0+p=I8krSh8xQ<{Vix+Fe;vM7Hqa3Brfa)WX}qY-Web!bI!L+ShYDusiNLi@av
zM!XaBoW_UtBm9RxBINI~BSdd2UZ8>37intz+5!v&z23bbI?d{41X#?g$HzsVMaz+?
zaq|~~7!i0)+_{VPO#BQ--&}e^jzWq{=&74`H3&w&6GZKZFb^2Bg1X8bVW1+5br3j;
zFQkIdc<RMn6T8Bd!=oeNg!n|&k2b|0#R$sCR=t)M7r2+KlE$VuYHdvIJ0(8z%mOz0
z&Z93Mqz9jUnhtihd_M|DH&xXJ%S(%?QLC!`<5YrndojKDy&tB#Z@!%xt;IAoTT}gR
z(u?iG^nCjubx+PM;OZ<bq}Am`9~sGRm`jz-JPg8?)3da<zn>0{j@2%-{e;SX<>QJF
zE+Q<n;7&NB;?X>A?`Z$BC@d7g7!)UJVdDtxLcc(<#a(Rkvy*DilMr1ApN(g<5pYmg
zAN`{T;U)?R_;Yv`F{k!N$jPo3@F)cL2<;J|3H~ZKc6Re4RD<8Ku=gSaYLu_=gYG|S
zQ^s#AmFG3@Kq%*ahf;<)Mkr@c($TM|9*fxErAUjiFjw);ITQr-s{5Y%7xNFjQ7o{V
zu_b?mm%wsLeHlKJK8`ZG;=5iDn6V06b>Z%>vl9z0uhc#@rOz=%wR%e>7*8B)#VYWw
z+KV}d;o|%d5_9Cc#w6jUm9a(HqCNl|-6qPL{LL4@1M2bkBRAewz~QHdKPpZd9HYqm
z2E1XH%OpD@W>r7ct+}w^dw75cb!;|O=N4m-#thYyBeY^C26aWTEgsZ&b+E`0#hGBN
z@}txNc4W$~BzC5;`;0jl{Rkl_bHHhZL$S=gdMdVf2F3-HlDCQDOX=6=YAavvukn(H
zJ8T$t$)BU2R#xu_UbAUyXE#0k>Zujtw-dAJM&kfVSn5x^doh4A0k;@y_;xvDXFh4=
zFJoB0r@o};1cetzXL8hASs)l=JG73!!sj!-nU|N{NF_u60p4-)2#X2nYk|YGA@H)I
zRX9!*@&I2LBWNRKqjJ-4qU}hP{zAJm4v`n}C=+SXF5J(?*v_1qy77(RnX!YpAmJF3
zDHq2jvgnYb=BYQhz%DlCf@)@^8*>oXQ4cgw&dN6ERMa=}FAoOy;uob3gB!FCFE|g^
zgLdKGA>%V#lwj-pnZIRWdX7ViKwcTy-TC1m6YUUTc*tj9a@i~cDc|`$FeQBU3*Igs
z@-KD!dP6uvCYuOP9f_tdxkXx05$Y53I=hKb6gnG0LTQ1$S!>34rtturHGNTk@@_xI
zzJh-)Kk0;>u@@y6{)}_%Dx$wLH)ZD*?M>S=2PPlJH*kdG5q<opU{9OC=uy-Q6bEW^
z`Y(4K?riU-?d@&fANFEnOL%*c?%iEY3ypaz0l-1-K7y{_T%D@y({?is>4APSrpyUX
z4P%wl3rzKGoRFPwFF2}HG|q_5YD`7}6$^IQ=_Ohwn0TJgj7fcqy0VM1UT<ogZJM^Q
zGaTg!<L#-+&o1%q@k!c0?53UlqqMDY13tvZ53@5%d>uQYP;jePs$U5nDkJ<MZOLxe
zuK33*!4$rLae_X>_y`@VXq;cWyXL<7^vRR-;PWri*4Bn%I<|QkXgBu~QU?VLxtn*-
zqA0mk9)h`k;B8{}7yXkX8Lu3U`_w28^+Q1ejDRagX0kB$RAu9VW<opreU5j%NC(|x
zxm}45<iwqd>F7x>7GHSBmD-`a!O`Fc7+=_-#~6hSEO0G&1{NoUEDipOg){I{$OX%o
z0j*#MJ;yR)XZ(Q|rkxr4ptC6Af-kLj_Yiy=<8zD;9EE$84psKO{X@4UeB`;@g3l4I
z3SX&TxsPjd8X8m;Y^PFRG^E=*^F63vJzG!fn>#)hn`5J=#b=F+c1%o(*0opCoqKPk
zpZxfb(og^F&(gc^zGHku5y_$?ONGS0m_HHTEsB=2Qx3%!i@Cbsmtf9f5ALvnZgo#C
z)8@X)eR!Js=i+1M>s}}`Hd9UA(`(Txjs!k+x!FDK1s*7BFK@jNjKvxo+~>QypSr?<
zQ`%GQcZpJ0W#U-DWATP4*-Rq^b7W$>!kN9Jp5|~<nw!MpGZiQ8><QmF(pNA+p334A
z@5qy#o!rn|V9Jhf>c%xS(0KY2JJL=sHI^tJ?jdZ}=&O@fwgO`ot?e9~q$led8}_<t
zdyaFhrrWEF^wp!MX-9PcbBstH4`a+MGTm_JmB-3#AAMt9^~d(AzkYtxEt{9@JR{%5
z4BYho&A2aze>MKsrH6T)JLwfaen&&SKF-ht1Z?@O=G&z;G~XtCF|UUFUE+R|874J^
zlNDco{^x)GQO2l20gO_7$ejrogxrRg?6C?sJ~4zRc6dWpY{N^2ukdqRUiq8HOak^n
z88pN@FrWHze)u?m#!Y$ppFdl}WVgueWs|$eAdG*p=e!3}Hs`083jph35JU_-h^OCh
z$?o|tDnR`FUF^2vKm0E6!{pcC9pU)KC4Na8+whNI86M?(@XkgSSivDY0)oRZ0T0*E
zFjY8zW6$~F=5PP{@sEiYCSZgc3=>4_=(y|S{W!*jT|q$vBfMd#2&@D$ag$qk!C2mE
zunXk?t6`Y<z(hgVP=vKwEl>W;oQu$Lu6eA0Mu-y=02nSN*`x>K1>y5fFA+*-of#*d
zs7qmJ^%mpoNr?)Ck|*GjC&!ryWTJ1TQG^M??7}M!?u6rb6a-@2iG^~4Rjyq0$dMW-
zbN<lAC8rq;e>>RCWd}d-FcyRxXARqrc@IuJqefn^!YJ_A6~$zTi#wruz=YL|KBhzU
zMFGR)hY4RWF@gu25Csy`0~75}B@1j(47u^iLl5B*C>3TqRtob315O;6(ut{w)LPU8
z_0CG_EO&hO8>>sp>Pic4H0dHy<Ga!(J+C~30oNalJUBp|xTiuvWS9hoKm~v5VS%IY
zig#JL3<y2p6E1-_q)pq0kR|ZNRVHH@<B|UOras^cFmQMR$ldvfS)jbEss>L2uL72C
zHu-ralk$XVEt4Ixi82J-C_^#yFDvbwm)sTBec0v9@o-blFuW+fsK1qSaw};N5dkJ?
zM0mN$-NPa#yZM{+{f#~R;$!IMp-OKLvtf;&91jJVcly{_ZzHuo{2gTqCw>N5<h=Qa
z5R@5br-CiQ5Z`&1@38Zt*7-9Y$~170|A@Os4gB#aX=XoLaN?yOoE-*>Ke92h@XpTV
zEfcUF!UMa5=5!1q-1UtW=ZLJI;@|?vu83;wD02yqGXNt|8?~VxN8@dmcmQl9#g85r
zWE38eHdqM^q2mrRlD5O~=*O@8=4e0?cX&J4E7A$;;HRX`H))eXfk!y%U*CNC(GPWT
z*|8G`JHchFq#`^I;pB#2eDRkZ26p`tC?jw1217&nFr*YQ@^6l2J7oY`12_H2?vQrS
zn5UNw?#`;zeVB)HkYC}Mh)WyfFV~fX4IkV6AO6|N;}rgOf4S8oZMivYeB(nr9BKQK
z7M@|pJ;N~&CSdQvfPn-a8sgwW?Hnh6`JasaR;)<GY1!ob{Z=6A@m(Hu_`N^<-+qL!
z$VZ>4;0VkSka@H>U10S)<2{QwLI{dI+{Ek1KgMv;CXqssFi!xh6;14`vtD&9_DA#w
z0rqsJf3YZmgN{Pb!BI2_br4Dm1}X@`1Lzd=^cH%hKk$XPOsais-7Gu3#!|I5Z((#z
z@3{)RRS3gK3!yOd9C{i;E*4lsc}-(%r+gteMR?9)9%vkG&C!2WKEPWkJ(>!kt@)<%
zP*w|Y)vkI&>muLSHSeQ?lpY454hjL%LfOT|^T3kt8I?iVrP+z=)SRD8cREc?q&Yrl
zGPSC6X|aiN0lJ1#BUzz<P_0#)OUsKb^Ng@#<PUy|<wzV)#w6TF>?scU&!A`&+`+?1
z=X;x`lInEUkJZ1syXo=cC+XR<XX5kr5=Ywsrp?{G^x(@c)1yZZ)4?vQrVkRw%aIfl
zd5)}M(GFp|r<dvEn$_jkJ{f==Rq&ha1ZS5$cS=ktB0C1Tf2G~-_-IIQiB<3{1iEwg
zUaB>k8VsgW|4iu|o%q-|j=N#EU}v$N?yYgJr}(IL!$<dUSJdVfcfj-obMdWb7wM>f
zVj;xMnE0ZpS@Ta;e2Q=4h^2*QMQ}yfruJ2t5a`7nal#2T2)KmM2|*OcRifZUXvNW6
z@SgBID84whk{xmgguIJLd>$9#`<^cZ`Q~BR;X}T4^ImEq;UNl|SUpZYUesoR;jH=*
z$`W?Yhd`X&mFyHm0E*%WfgQ$aD1G-?(P9hJ;az3OgT^lccIKMQDJC_CU@n3{2Ocu`
zL3ZE7t{M~t2oKrS#8Fqi|3l?tp&rVpxvFrJvZ}ohUS40F3Fo=D<l4thO^l15lzmDx
zW_tX}JA@8}Pn+v6d{6S>{z2*q|5-@Lkw8}$*TNHac%TSk+)$gq^JQL^9U#oVPJ3}=
z7=4nt6+0Y&YY3(lUTGB`n05wsRvH*aq)U0DpA;b_E7Iws2%dm{rPfHTb~`ODEv1FU
zMYlWc$O9b6j~$TQ6|}vzsd{kqVc$E&;Ncl_0+$6(!HTyR#vue%{oun(z6j-*W8J##
z)NfTT#^iH79M{M$IqHUxu6M3@*Wl@tm2aKJhH$^>oh3`1mimx*Uxnq^(dQdGzGG;8
zb3YyS&eR9kxfpZ43k3a|&oQUrh|Z}g?uTLwW<H~-k@BXG4i>JlQ-Hdis$a=YyihbU
z7YRDS7|5~kq>qE~4iD9-n1LJq8fSqKdBqqRr)yj&eRh^{)a3CII1=N`sq$uSgaQnO
z6lpMS7Ud1JfH9jnPnma+x97_vbLkK&a*W-%#$S%L-1W1+w=4YI^{xfllHHrk3%T2i
za?-b$Q&2z3iebkO^B$*&J7X{Plc8*wppQY@i(fH7Y1GM!WM}8x$3po>yr9eF4osK_
z=%Q)0;**PEke^o#F2%KkE2hLLZl!!KtKQVv{^6+pB^apxxKJbM{HDRLJP~w|j(%P0
z5?^VJD&>Ov4jVyYE;nIC$eHk(GO=h_5%r$-PCZ~npCSkTe4ibP*_;RU@d6fS%lH__
zDzbZ*e#o(|;2jHKd%dn@8z?S}LrM?6g?O}o6cq&scpG8J6TIO5DDKAMVSz=bvykqr
zv@P#}H)zaHr?us#Xy8SrmB3VmQkw7#v!WR_(M0YZ1D{AU7Epw;gE5-&1Z*_lR$O=&
z@R~X@zOndVW(IzSySI|*Wl!{&F_f`-c1Gi(>Vr&zg)u0GSkT3BuH46Xm6QiCW^s+i
z?}e6VK!fWOQaToYz;S!*iuFC8$0rJba$qc-AiEIEdI$UIK>dhaWAsJ9MIT`?M5nWq
z78ly?R~((Ux$z>cKYuP>;wWuxZls-^Ey3bi^(gH=Y2K7E5~a?$;ELi3B?!x*tn6SX
z*RAS|;;I@7FHRglj>Z^IoAgkY9rx1C?m^nxKhpSkVWmV@Z|+{&KkTV)CszJ-k9t;Q
zGCt7W?2@-EO!(xT(W1Gu4edgk`7TM}0Xy%Jk#Q`xcRDeC8n=Q^_NhM_H^5WcmGKJY
zXGOdnW#%qPWK-Y@@;#Jr@Ve04iOeIj8x6T3yPvJR7as!dqo6@4czlKu@GR}K^Ivu1
zZqZ}a^XgV@I$04eD9q_~Iyw<Q)xT8u8_^WOa$I9hb0saTi04{Z^-ibLE78!|daBgh
zUWBx~gq#W*Q%m($Cp8v2X>sLlTD$jVy2mc4x8F_o?%kJv-LQZUW2e)(6;YQi6L)wW
zUyP;R<+zVCoUOJ~rL~l<#-`JYox}9i`cB&3*;iX|r`n177ITw;muaHf#0wI#e2xwc
zjrYz-;g=W5h+py}cXqs0#iwy^8FY@m!Vbhc%k9(<AJ*%i3r-sAHI~kcMwBD6YvQ|T
z8;*eF!Cs{;lr9T|SOmuILwIX=Zg7<OT?M=o?*v~)-FNqU;+uQv>5IMe^!c{h<H(EH
z`r?(JKYJm3VT{tuMvuow?DXzDZpo5g(cg=`ykF}+@{O?__gCGr8MOy}41;@l2IiIT
zu#fo5ycAyM)$d|n=T3UXkKfTyuaEOJ<{QGknpUBIqwc?N-mixIoyz?!O~Droim$Q1
z{~!MS;0UOP-X47QBc5SsqK;8Q6>8v>;|%;|C^M`8p+=l@oNy&=hcNI;Y9*Fz`ti|f
zINJfs(2P2j*Z^w~o|iuOc31={*yK=vYCoLp7H?dgjy`ykhIcLGZ^~8D#W<zG4@|5-
z2Km}x8F9dB_+lsHoEkQ#ZMUHrLNU8j4qJxtu=-0Md}9wt+EN@w0tJKV5GAgDaWNby
z2rtT#D~+GS%iud;_z0_T94~h)hWp9-^YrDH57L$fb0#?mKQIX6VMtK`;l^asY@5=D
zXp%O}NJYeei4+rFR&Q{B!jdNU?e;<@)Uon|Nd(L^#LQdcA+92rq)%F&NQo$#DaJ)*
zv@l7Gg_%1z9XSV~W(5rlj)(`7D-mf{;V?mgh$xVWIg<?tHWRiG+CXGD;sQdC5-0|G
zU*IezZ*y}yJ$dp}lZAuGZ^XdFi9K{4zpqge!3^Gj69`YEEg%kVB^pzo(T27ttRCWE
zm}uy85C)@gGnJXtS-X?&++9ocMom#P(NaB6&rj0%$*C1zmtt06hOfQa3`R00%=8zp
zPE<Z54Rc1@bH`dtrr3dnqAOr)qO5OU4z4tr)N{883?+9p%!}DsSzS%{-h9L5snq5T
zYgRJBP{LSxMXLzD?-&!@VNhvL;1DngB`Cz+tHOj6#tj%gd7tI@tZIh|1|MJ?FbHOR
z{=Lc*^-;Zq^Mv8<0+>Y~t4M1Kt_-k%+CoXy&*xn$8u$o<IwemFG+7}dX0P0%FrjaR
z3tmMFH!DN`GhPEvXC9Mg7(!1dm9QsxGT=V_2pDs$f=UoLr+OI#`g7kC^E-Ysk?yYE
z)wAXmx07O;i$I@WU}nO`hcm|F4?lKMc#o@k2lnBpFtWU=Tp5fv_ABYcE24xSHWwp)
zoFgVaDU&SZSsZPlev}{^yE)h(EqtQ{C12u_o?a>dr61IyB=k}``WV0?+)J=gn@wbz
zqeqSbCp*Z<P4FPULE0q^3MqRjH}>!*PNd<Gax?sG&uPoe-zq0*8}GAwF(tmk<zGx`
z%Lrc*(rb`5d68~T<0W`RV7Cp%5yIHzrLsFhc(_4Is`ktg@;4?2br}1Fd#PW^imZ!$
zB(~oOQ-Y3#TQQ??T!koLeDfLBVS<B%376p|GZJlRN?3bP#~i}$k#dRSWaLyBNj_X=
zeFko36VYr6=toV-W0;BzAEeDoKZA0-#*{J;*YNe1jLKXVG&uz03Z4pYIO>lvJ@`3-
zlj$pYz4D21iqI<3CVn{DL|Jgp;p{pB4|)7o|L`A#^LJqWkN?p>xMi&87!t-+3o10W
zLNAzTL;v7`7-OLgA$WwoAuvH0h@t?&GPHn&Amd)eTH{zhO#<Pg7~9PYk(Sb7+=tg-
zf(0LAC6y9D;c)2*<5BAO5x8FH3Bf)Bd1xO3A$F3rI2KrW9UUJh?%Qa#>Pmy%JQqG*
zj3Z(Cy|@R7o!m^s5uDMUtmL#}Lgg`^C-_qjXcBH=;>aWf4+x&&rx0XxMW4Cf2!S<1
z4B8Ps!tJcM2&Sws5e-Hu5R-C*QTj$`fg)>ePH?;wKNY)LSP{+fC|Bn_A1}kMuD9R3
zmsVGo6@Nx>M@Y|Zk6T~Jy}7gJBhA>s)oxY=3wGO8(uLaNu*(YBzA?*>+LGf3*^R_X
zbMGRX4w`-~SaWO{cRaB|8~CIAqkf-#{&{-x_>ucdXK^vT@x~jeR%`l5r)T1YX)hL-
zz_+g5yK8|ABzb&t#zLsw!;>^IJukN_@s8VRR$y&Z=2E9wPxI507Am8tgLmbwC=_#~
zfq<I2a5*_Ul2E}S=sN25@-b}ULB)?&>h;7fq?y@R2~XGElmEj<Pt*Fwmf(8r^zozH
z!dtT&%|{uD_k!<1po9VyVI;@pAsqCvbi$dVqhqgx_bw<^jIpbtc0(`%4#6`cjAu1E
zJCWYFvy@hrnu?=#?VTD2*hK~ofD7OlyTK5^$tw35^V!_D#a&k5Qz$2c&!SIoTw|>!
zc%Gh$*FLmA!b9$Z0rz%ykJ8Rw*FqL{dNF3fC$ZWatX2~XdSEi+TnJ{b#JhQSfbv1v
z1P^HCn$krP?Vb7VBM48~C5EsKA>6F***Q|*JUr-nX8?I{1k{P}2W1ojWN^H-*tYNs
z9v&qfi@La@iunR}uyNNJ`OyCvSD5E;&n>4GjEQH1heT0Do3f0Mev7cniXy=x<`xKj
zRbutavNOc|w<3Xu3`mUMh1O7a_j7%-+tGqF6jaQO)b<$iWcMdWQ88chTYMzr0EY3J
zyvP&$;Rv?v_4Ty3v#s_}J=A7J0S8Xd_u1{m*vDAOLNEmPjI+K6OMTwEkd!AvPj<}#
zXY>5Z1Ns5%xm$;Oy4VSaVi3hJJ1*Ge*jZdq8-rUKn=~fussHcn9J+mQ12-5`f!Xb?
z+Fs)hyIRi9G)BmcJ8`BcGYTAyU*H8eX$75Wg32O%frm%Y=UuRZHT=0xgit+!D|I4_
zX{E{y?14K_Gj7s8v<)tV;a_jY5uvHa-HC}vi+>B>G_J<HM{TV5M8l8y7~y?yj^`iX
z3p*=A_|M$rk|X5>a|E62qC-d=iv!gMlqTVzbsS?$cm-DC;0g0z`lMi`aD>5xFjaFs
z^)2Ri;0)#FI7;SQZezKz!-L9!cklq@<9f;!{KS}Vlfm8a&F_e(^3JlWklj2LE8tN0
z(O-ZWbtN8fq1?d59@!KdIOMk;aw~Zj6ZyUD?(*a^V(_CKM$2oZvA%&(32%4YM&gny
zb&1?@4?pn2aM7d0B5l%edeD025Yg5St9FH6gH2X!slFUhd3Wtz&<Tw<%PY%PE*~Bq
z2qv^m9F51~hqKcQ=NyYQ1w&=WbrG}|N{LWZ$2@sZe1!qO$=iw~!c*_@SWE*g1}F6^
zd<fj9PT&am=KM7Go|~CW@4R^@{q>*xFx|bgs$_+emuDW=sq5a}ZaO%IPR<J7CcJxv
zI&+8NwCL4|_?SJ8vQr&cBnqBbzM_Zx$Z32`ebaIhcCk;0K54wI&839~JDD{GsV)~P
zEBurSB^owkd6l}jcs}Gj8tXN0gG6yO9?b#obfd7)?5ciP^uVIrGmQh>S-7*ir+S>H
zDfk}MT@{E$9ZDC)4ei9PaCTaw7=*{Fh$mcKz2lvzDA@M*4%34N57WaZPt)^t(QB25
zvAEmawc=;7vm9e^rIA{##Z(o&Ck%HK?u$12h(FP37Gr@ijJfQ3Wsy)*{Lpfz<?)m;
z;8=X%asOQ7;d$Qi75ppo5d$yA4)$^t1H3O~pgq|Y#1XXgGx%iV)ERh$LLJ3o@B;9c
z<qkM6tO5xX2A<Ti3&sdqS9t<6`thvBc6QTow>z{Jx;LN`DyPR>#!e-5<$kVq5<Vj1
z;Sna-AHE#iK)%GA?~qkpJ3KsA$aUa@L(WG$+KjOVo&vs=`|sK9#2Ak9|6p%lxP6tD
z+6(FJ`^)K#)lOPlZ5z(uBV!%B1aZMz>P8<!p?h+2thVY}`O96b=MJwj32!G=SMJpU
zX9XS8VUEGv+1pFoyZebdPl1Dy9P~qvIRZq;Cqto)yH7O|9e{5{4rZnJ6bq_m&1bSW
ztRdWEZsNI(>Q<evst$9B9hv(_YESip7@--G7+(c%!5-QWWeIvuJ)wob4?e_;m}tMe
z07zM-wZ&?B|BaPYn^S*2WG5cG%+AuY?QVLy!49>4YR*rkib!27AW>$5iN`_mmp$H@
zevY_%PANmZ{RRfvA5-jhANj`Ej{B=_*^Jsvd$SpJ9~im)X51)KF{A#}<7?BAv7Gm7
z%`gw9XrrNCm)6%BrEfUC#=M#q_<<R9AFbK9$-CtDYRFNz{-I343Cw@`AOG#Ki9i4I
zzwwa?EZHkVXQ?-xjV~q%cE=5*;*Ww6$8Y?VXbu<1nf)-bTY1EzPZM%%s&&MVa86tH
zQVxfaacH-wbG94Cwj9jKsGubeo+!5-8_FHwO1bz2@D}J0H@il-T*i{#XdFA_Mz%kH
zlsdxzXaD9h=W>;>jv0~Fz{JacIZVNCKIQD6XrK6`{1&=7Y=)O3Mf>ncdW0t|L9yfR
z2SYgA^*yMM>l|J<@JAkAeH{gMdKjl}?;tj%Q$j4c0C;`lg#{ll#G)LIUj}^?8_wG?
zl(sV@zCvj*<xKI;V#hnYxN8dg^3sY1&KQItv{qm+Sr<X2j0~*6fPBhIX~J-CM8K|;
zSK;_prv-+tv+QO>ML!^Vu|h5w-YYQ}p>$vpaBW7BxYWx^8-?-8JAIo#susQw9)c=@
z#BQ&BO~es^MLmb6Tqax?>)*@J{<l%T+!C&j3*t&pU)*uS&p^m_IYS|fAx<tMb%DtG
zSUu`0dwG2Hh8SSr#PNDGJw%uG1@>_qhuT6!2gWlPeoX+ytgs@LJg75Fu(!1fR`kv*
zF%GOChi`!?fpAVtkEi;=LTYsu6Dylhev%Kw7ABnI4Np1VgR(l!z%9c?FAHMyXFO3Z
zfO<^0!3rU8Nd;o+U?u>Di=$YaPA1^2=4Isy6Mgz8$MLYTnQ!tW@9W!3F-<V?j9+B0
zNu<ii<eqfNl*t#&1f?P!PkxjK<%8h_b6NEaZUL9VjDeJd_X>M(oZ2`{OnNL#rEK^p
zJ=&h|2<A$eX-g*5Ho#VR41)qgSb?$OxER6;@QO~;$qp_@TK*=3{5F<*(7>iY;}?b)
zvWJ&n{P3n<78CCJ9Yza#1RSL853DxiBsbMpSQSq0?=sXe=NIJ(TlR{q^q83XUw=-5
zu#uLW?GBFQkLWg-Y`4GCz&l%w`{5RCA7KoIfqnQ=ISH!|jy`nq+0X9&3pYH&uHOO=
z+1-5|1PkrM%aF0b&y|yfK2>1hKj)(_yHWNC=d?!P;TYq&Z8cG}18yz{7VNyDJ`pv-
z5rXucj)N-<+4v*!^ApwaPdE+XSJE4l-$0Hc2K;Pu*is~i#ns<o4;wI2_za-%9{OYP
z7fF(redMASQ3fINx1nPv(v>f;&rb0~jGUc|Zkp+|F(cuDfkJu{E+l~Tz-hY=>*lcj
zbC6svJ+c_K{wC7PX`4sDUyhEaZ{EPe;q0jV^(GGK62SKO$bT{_mw6=OF@`EF2azE}
z;YXryjuA1PUjCzQ{5cWcyrKe*P8@gS+<)vyJM%@-geW%r#?sHf!=nyA{5SshA94RE
zyQH{B%exC`I~m3nFSOC*l3hyd%nF<j!H;Pr<GsdugdGU);S*Q|iSYlZcalytab#t*
zMrhHJF|Ro00&aGLaD2|<Vly=w^O~G$f?Asw&RnJ!PakWNek}UH4(+)3oi~%%`I+(5
ztk0wcJ+*o98e`Yq>C>#wrAD0{GBasbbcKiVBDjZcu`--Dya_ZM=tHZZ)zC4oBvo1A
ze@x@mhJ;1=XW@OTVIc^xfM$883_^18atN)N0CO+M0+*sMwo^y6eRX9ett{)k+>w1D
zEiJWD4W)~510ja$$Ih7B^VC1sOwS*EmX3Be(!Dz?>Fu}Pvf>ALBDC7sJxT|MJ+GR*
z63%j0jd!LXG)Hlvbhy8vf4G+}&w9R8eDd}xO{I%8H+hp<6~RxiJLw*ztrzQQWBpm$
z-P!g68w428QeXtXa3*-1pg=sk@UAg-(6C$NQamn44js$BxV)0?zxkGrsq=9#DraYD
zHI2{Ar>%oyD`c*(HQtR+n(v{!WhWjuhOht_QEnE=Ab6yGIKB$S4M<45ak-}}@nxqy
zJ;&*Wx=mcC>a@}vX9t=3oZ1QDAom>Y9`&psM5#3=xNw@n5^bS2U`G~0M}&XeSAmcW
zJRE~J7B0Znup2H3Zv8%dzUqI;s_$Fl=0bhmZMe9gbcBoC8O9~<cUC%SNq7wI0Y6~L
zV+Dv{XP%2y_`*kaiB;#L%@9l?G>5N;2cz$^V=R=zC>P*0Z+*nv&enR`-QG~!_Jz0S
zD({8IFO()TdQbu|&xkof^k;Uanb#v9!QDH_^hRNdkgMFyhZ2PyO|=G!Oy(lO1C^bf
zaV!igccP(u>UEFO`tzr0XM4--h9Zo91OLVyGOf;1>a5&J^VNpRRZG=oD>WC|D#N_m
zl^r)#3&Id$USFxcSIjj~aI@ord5I8C_)U5&++x-D$w|*U2@qamd{m?Ai((T#k6nYc
zTEn{oYxRcj1Dn&s5m6|X2J}{Ga+d~tJwi)h5-?+&5Z_p<r$$qJd3(`E8v<M6BIshL
znRj!Xo_OZ~M=)(|@1~6xn=0R_>%rI&tNwWqw4UlgIeKD2SqS%;XT^Mj_8wC?IEG9g
zF^(z?6jcmR>^5b`Ao#p@&`W#!!b#=B(XbvDS(u={hah-kTl3TH1I<xcS-z?H>Veut
zc~WQP!&ryHE0mOKGw+^?BmGdov-7FjMX7(3dh+KuL*_#$gi(0WKdpQ+9Si(K;g8_c
z^&yP>gC2$<yyq%B=ED_J4<>jiwFNx7o{4eVA{Ic90<@53BQADm5$W<|@+41aPgcgE
zkaYSe>TqMOL0@7{g)sI){fHy&kJXl4*B|(JXP&}<9?&Q+T(xpD&yjhYzw%bbUZAHs
z;}!;d)tB<J09p??Ox<|%jlbuM`7i26f{_C46L5(KJYd{lQGx438F-g0vl%G8fI09n
zz(_VHtT%Mbi#mkoJmH$NCj2Yrl`JAGZIZ(>_v3pIkV47N?eo%8+L6DUpWitx{R<=P
zMo1^9v<BfFCdNcUkv?qz99WRc?jIBuheut@BG`$<7`n3DPOEp8ys+5wV}*x~qr9e{
z8x|_R`2hXV#w>C*euAfziZO^n$k3-b>Y09xa)cuhjZ@$ae2vnF#zp&pCmQSLV%G<E
z=$&4SrNf?Z5BZSl0R2R9Vxjpw#|uxSnrH$FG<c8;g`=F6DksxuoYq+E@mYKW$FnuW
z=QL}v`@6HiF|!)`80VD+JQ@p+;4jKfFkr{nc_mzj*I>~QN&$AjpzxlX;a=?NROjBI
zIq^vX*_CK4N2|_PG=^4dt`iG74*OTBd#bu$ql}m~PJ{DGMs(@O_jE#ImseL)rQS^Q
z8k0E^m$7Bz#YWoL*zi61D2Pz>be323+(~n_R=T)lR}*(@Y1~&?_jb3_m!E%<9zOV0
znoO7Jd+*#&@4R_ed>C_k=o&N)o^-~$$iWxxwuDz=+@;&Ww{Wx^cm!s-FKE1Eu@!tA
z^wB%2puym@+Md20cT#0uP_P7dkk0|%X^x<kBM-D2g&xNb`?zh<BkDCbE85PkEYTBo
zNBKl2(R*YQUaTRwF(y$@#$MBjzUTzkZJZ+iiCw|oNyb<vykyEscSnANf-V+?p{!%<
zWz3236?sz`hbcFF0e#UjK*n<YQP(Naj-a8LhgErj2Qn7Qa=5>*agrT-N}HY{xN-NS
z0Ro;Y53x@vc648hr<xT$O(-Aew(?@v;FS3Jso8WhG2>%)uhkap4D^mYl|N`*jLXOy
zDI@hj*<5yU!B0W2%?r64u?P-Wf(I;?@!h!<7L;)k#BAy`D`~A$PxqD@X|-MV!mzsd
zHF&m`>Nu?q($(CAJ1@O}O?lGR(D%L`;KbN%r6%Ky=<%h>#^S)f_)PGFv8JZJfl`>`
zxo6d9p$*(6XF7Q;eg$6byq`FtoL%$qywErwFRC{7I0W9Qeaa~4ac|`93i}l}SM23I
z2q*KB`^Y!OcHCce%jVnI2S#qc8Rq41ug3e@bYx!VPI|?U-_cO7kMlL=8^V5@a{az}
zzZ&v)D)+ZEqiN=&4*$kS>Y;W34H)?HN4D{sKP>zRX2m{4bAociPk(y(JN}7>AHho6
z#3u{}|9FKVef>sp9LL`!!00!2obq4fJ2a(iu5c8<;q2#M8G8*Q4ty!I|H941$!@m#
zr+mqf2QOS?C&@wV_#c?;>L^9wW}%eJp>I5#9)6ioa`<o`rXA&i$q73i478*d&%p3)
zcsPk%AOG77-p(s~<a`-uiqOohjz#|=?6QN)2*E)xu)n|Ot-B|^zKIV+2z0Rk9%fm8
z6u{6#0PHkm<i=fHt9S0Cd-v|9#ZE_)-l`^ptVn6d4}uDl6NE92?Fc0VgqjH%!W-;f
zMTno$j@5;+q6#8sWd!9H;URAeq%@%8XN8sQ5D%0q92o-fBurc&tBI>pV(@3R2KDvI
z5D}*`d10j>!VQ#NySw{7zJ~aKjYcr+vw;A+UH|_uhOH7co=`VEPLYg22Ah&cG)8=7
zKN?jIw1wMT4{+nyB$QZ_3J=VHOIhVdeSnWX%rqLtMIa$!=v9-#59((Etu$B(#kdix
zB7p}PlXc`tdB7DjCVqs=Q(|^SFxyL=w9sBq`%asYKEDj28?Y9GEm#ADVEQB7k`~6t
z=*fLd=~;jRGY_0Z*v&YIxuYBxSD`>Nqa)@3MV`wmH}C@@PPr`n6SGAcbBY_vJrsah
zSteX`jhL9J-UzTf5uz_CAn`0r;9G5i8!L{$S(q`710XIiBF*TRR52K6+8`9NR4eN1
zrq#<6YVx8HaVCuFA8CtO7w%FA>I{QHe;V`;z~op4XW)k324BVxoXbd#i8tdC8R_4F
z?dVqs{&V>3ZsYKBh-hKwRetoi*5v8<@fY@N266P~eEdl$+sT6B*ez#=yX%I@acCL-
z!)_1z**k{?lU#2n5++LJ&=G|<ZgvbG(l$u+ZEy;X#h<^TcnBET=CI-KSF|a1{_NmT
z{&PHk<Ag=ulv&nN&EgPk6Rw2x6H#oB%=E8tSPSfIawhzef>R-EcJPbuNIo3&cnHGI
z%ip=qgRnrAy4xD*V&v)iWySr$0095=Nkl<ZW#hDM(K9Fu>0}SPk#<hYCep@_aK=S7
zzIhn?4a+=;!#979L)2))QZR=zU>uN;k*@z;I6~Msn4QQV+y}l`$OB8Z8$bA@d?hKz
zBR+2ArB7T4XP>BxgW4wyV<A3{r|$rnKxV&a`(X&Xg;ge}p==1|{=tXCN2rmg4(lMr
zO_rQne)O9^T#sx9MW9Sru=6RqcR|`I>X^NpD&JwS<VR`z{NrDKXNs>M|IPp1M=W}n
z*Mt$lePvFQe0^(pwxFJ4K5(nevzivAkH%EdjVmVP8n2<t(4HBTXL<4;bd%lkF$tuf
z11?cA(Hv+ewB|A<xC+ZLTPU?qdOdyoIDP!_FVfb=3k!TY?Uu?FLLl!XQyJR}n#>}k
zpGAm|00hBSJTszYzW2(zaoL52fIH|Oi;CD?1y6;N1$eXj1G~e)i?LF)!tQ&7E7T+6
zBg}%XOihUoQ~F%yj^YF*-a@;PmRH(o6`|Y8Qd(MSTT#X_dbK(ina_ISJu@L1h|u}u
z7)g=d-E_3O>1TguGxbk;zA*cp_rISiji!ZVhuxF3v9+7Fw)ZS3;bLqAUaO1EwAh&U
z5r2*9Y?_{A<?%^6=^u&4UZxrNNR93s#RFV9ZqB<#wztyy^Ji&qXV<iIv9pxgd3Cj6
zM0)_QYTXJ8{9CQI+P12;p7I?*)AGA}|E+ZI{yiUo1N;!uqF|y;*?GnN8H(Wc@x4P$
z;L-5WJSb|o$B0E-C`P?I<>JQoJzR<5Mi6v(bm+2M&~wQh4#Agjj}gbva5oRQ#G<1R
z=mxK1K3Sg}aRcs>@5=IWT4*)Zeh7IK-^w2N6a*CFjfBf6D|)ADlS`FDWkR`eDjY@d
z16&qbE#n1BBj3%U_$U`EvM;vkK5~-F&Ed07PcM9@25GVLg!uuBsptdnR0v<;^HDCb
zBL4VrUpS?F=yzbH#sK;bEA=_<4#jNua8GTvk<R<wbaQc<&QAN@afa|7p(z4VV2ywd
zVO}HOVSylMLT!(LpOw;H;B|R!{HSvepKuiA3h*l{!r3W=a0a2Wl_9eCy4<(XOUHV1
z_X#V~*|`OLd>__~`pc{ywKcmGCnhzPR1x;nTyF%%z#<m3X&g}9JO`nEvWFlrjvDjM
zxZ~q=-0P;p1CCTbbX|S#oa%wF%t~tUtMH7Btq8+dM2MgfAzt)f^>x*aJ{&?ol;GfZ
z9NCCJ;~6_fIqtDuNAbl@FOC{x_4rsQOi;|aPlbXK1=f|yKsgvUh&Ml1Q5<%MGDadG
z1iM&ts6ME+XJ-(*Q%IXQp@`r_4}^tkXTgGnhtP@`b7wua5YPT92mOh%Gfpw?vh$In
zxfmZd^&A`?3x4dBWp2lgC603AUPf@`#P#qUJt_}08KIl`cV!AMz?%hY@Z%WIZwSL2
zL*KYEcd&^d4r>LBzKKuWQAB&pmm7RKmIy*iDKF->N{1bX?B-z{m;i46R$Hk(V)p^}
zE1?Kyo{{GoCX=KQ<#NVkBsb=4UKpo8&$Hy_IfFvcFPU3`Kb}7kML(p2z?OQJ#da8y
zEO21HLVF?ng?8DUwEaZdD9d;=&%r1;8!J?J6b_z-Ka6=_MK+IxD=>0?F}GxnTIMg|
zCm-eQys$eOKT(D#OFWn;V%T|6PwI$Elwjy@LtNf2!_WjDpYcTSiE!jY1xMZr!)`e5
z)&yR9kCunmAZ6RB1lD+mi(p4^M~Ql4{MO?y64FCEM87!KIB}fz_kupM2xVnuIW4xE
zRur)l14R%5^w`A{r|3i-l%CRIEHX}D1k>Q-!L`WC#mQy>9*xm%Yx#pWlqXKfIk((|
zd9-mw{wF8r-T}3HaF}{07wJl)JPRCXN0j!Ak>Eb}HZn%*M|oarPH|-#Q-lA~_};3|
zrgmd4wd-?U__L_*R+ak`HQt-H!3PNbDxMe1s^BaVVC*>Nh**u+&~WG(bPI(lr?arY
zSy{WR?DnLr>;&f?zf+CDEWF^z;NJPQXrpNHna1oZXcWgIa-?-tu((bq3d?vkFFL?o
zk~10;E48M^)OKn%7rcn!$<xOk(+}94r|}b|6}$sGk5FVan$5Jby5@xpoIJDs{IM1F
zi{f+I3r&yNEcW3v0pIB=evYx6aUG=sct;a)bRD~JQOz(W0N;?W!EdtDHjWMCVXTF}
zBu#b^vU4uRju_8SvcY#W#LLy;JK-TUhO;oCrt!Qbnz4vHr(IVW*`X$0M?3+iJE0t*
zjp4PBw}f(2<1%zij~7XZ?y`8|Ol<*O#m~E%X+MntA!}igOe{QO!5OrNBNcs86KSdq
z)q);ZDIeu33ODLSKcEiKd-zgzt`}vI@1PW2L;lC^`+Xlx-BW6+zu?7Uq@X)29t(KV
zUr-3KxD6il;$}RZT-=JEyYjm$Jb{0`5xr-82T!~R>Kfkv!V5>ZI)-C2IYj`)GUJOE
z@+d5M(ogAdF&?Td1rOXPufxq2KAm~TwD`K|)K*_u5?|92Pu!?+#~Z!(*8Ly8<3MG7
zy0M#{Z0x0-uJ{`9IxG^|-sN6V;kw{x7%Dv!oh%w>H?(m^@ShVLn5!%{E9uTsGu`co
zzid@fV;(=|LgVJ4_EaZMPGIp!U1eDozup!Pijs^z&H@_lw2LuD^`e+E!|`(9kY9{R
zFZ+$UVP0|{`Nr6e`>Sr*{B}D}CTQ6CFJas)@*9bR8BGK8>US}(b0^i}$M0yU*T?x9
z^9^CYO}T#Gyk8CZyToPW`X=KrN|QIwKjNsvzxQ|kZ}i*Zi@&ULGGU8C@j!g=gdN8v
z$SdLFZ6mAWMk2Q3<sihd-5)k)X8akfUJ-E8c}X#ly!hGb^6@UI;HT1s2adc5Co6Fn
zm~Dt-zet-h=HM<L_DGy}xR*>`%3D6f3@?a&*h;$bi@{xXf67&#u89Njqe04^!}JU{
z1$4NP=pM{?yUg|_vGByA^oC_5%uR&J#`!o6*~o(^;pP0u%S<b-E?@*NX(01<cg~Sq
z+RUV4xL4~^5%x6+|NP4@(l0;$Bprz;cygdW_&^M42H%wHAmd;CFqqHH*VFso`+oZ1
zgYWswcoZP42!Y5$_z|!$`Gmj)!{sD`=xHLrn+bDFRDvk8vNKks#fl~f0+VG3BGF;G
z@CPmsSSFan<ESDSM8d)Nu!;%7Nqu~tt_ErpUmRiL6(Xve6++^SSYdUbiSUEZAEfp5
z7jh4wo)tO@n89d6dF)04zk{Ldj^TvWFKI`mg+X}W4y*6L2iXC<Uck;}g<+sANQlXW
z{~|dIX*02c@HaUdyfTw+Cd5Rmt_VP`#B72$2+>>z{aNt`bC{VC*k%%hnSmLA*<!^z
zD>Y%j3L{Mkfh8+i&bc^LZmi&$5+S_*&fDqj_ufs*%S&!QR_~mjGil`_Y3>W)?gLGB
zRo}qT5UL<7Kv0089@8TJC{H=QU{Z|541zmN;t?n^A*I}u2LjH-A4b56P`8Z}Hhf4s
zCWagxAirE^U<gABBjU-l%Fn7GCizS}z!Th2a3K@|KYf3I>Oz{p3jAYrFz5Eu4qo9Y
zW`cS%@|L_gR>*=<l_yrwQGMDsa6DG*WxSFzIN(W#a1%iY@f+1fx_9S(T3cRIpJKJI
z+L08U3^mI}jm)@Xg#~}|AYbgX|L}|7FyS)viTEWHu0vD43Mx|XfOpuzO8ji(;qd&B
zo<Fc-aYds{dSbI9fgECB95jCX?fiI;1T8__Bjgy+$N_ND+hP2vKVUTCHnfBx4@{{S
z{vH9Nal%c0hL8SVDwFIE5nl3BJG@+gFmiL6cBCipmSG$qJts<-3`76QkAL|^JO?2Y
z!w##zu|;&fWQm54MpIUC%wf3MUv|E**_4lAun57w(~^f?#xU|O$RK0i@K(06C_G-d
zT6)3_fB#T!15W=$w4pwc=y05MZjL)t3WZ<P6Fcs9in>M&zm+)u@{V}^#o2%1W_J@$
zxy5I=73-+UWzLDps__gbS3bwjrdWxRfBYS>2(CN^b(eA4PBel#%?xDHCJeu>Q{+|B
zE^X^5x#Wx!&{IB3J2*aJM3-?aeIk4U{*jdI_&KnBoTuX($fV6n-#KsGZ*OEDxP|$D
z{e$oBsKdYa|M+hO3iydh(Fl#RDie2=a#1!DTULfL*_$_i!fv1tw(cJu`Dh+aJ@Beq
zg@adNd<G7{l}eZn!h`8S!DgYcXbA8@Sj|yx9LvLfU?)dM>EXkNY5m2P=t@7`zjx2O
zrVv`sX_C(BVPNK+Gn$BMjww*HT2p+FKnN}2jto{ibJ;XU*ZKYj@hKdEhOmZXZV*Iq
zR3Hi#$|8a!UXB&A2s03nA;3n6N4rElnDBGN%h=78cUIk81zmsk?78wf_p!wbt+seQ
z6gsoUA9gH(Pm|+n`0Mj@(LE6VxbMk2e9D(!ewmIB57V8sJE_rHNK=(sYPCZN<=r-C
z9NVZ#^xmN-&<mQ#FE7@-;|K*6i%7Un4aEn%2`f_1MYB;9!JDzNJXQ~%32axMa6^OP
z11aaiLMQIf67BX8JuJKttwu0`kPIOV@P*zZY(Ob>F4lupttgC|;+<;MiuvKH_>P5o
z#TSY{e72c3ce{c&LYQf_Z#6Bno30ZAK}K-|F>`YW`>xV~V89(*2;IhnAJ^x7!RyRI
zBvwtMAYiqzcQL6Ah=ag_V=h_wkJ1I9CYRiA?Hq_-XJ^c%>)2lGD4%)d2rn5s#Oerx
zls~(@*tK#dJm3hzKF4ZKa$m-n1yTskSfGPoJXRi~$eT*_>Wuruq;L@&;hr?1qv~~&
zj{6sB=diE(AE)Cp0SjJ7xHGHp+(C!(M9n3fKzTZqwl`m-CyySAA7uwz$9xYfqp!!N
zQdc}9_{z~evq`-8MK?{e%T47$frKJz8=;ZlhC*^|Qt+uYQ)g)<;HHegQFiw2?(C{P
z5zJor;rPm><xXk}@8|*?kx8FshZ6nXE4|gOLSysajJqfznJZXWsyDnXd?Irb!oVk1
z#aAybaf(6Q*{15JBegGg#c}Ky^BwNjWw#OczOg8a@scvp7wM<qv&VV6qc9_%+z~Mv
zE1BQ$i9n8V3Bjo^oNqJ)gPM6`49d~CpOADA@T#(^m4%@+1pl(5G(-qGJp-Q2n17}p
zaLgq9EQ<kJny;|}9K~E51Fx|_#X@M;JM0=S5&nbV%o)HVj&jrt!u&2f{ZM`(3}ok4
zIWmo1b-)m%EW+K6@E*Y%I}hVHJJAlXOL5pO!OC>tLiv5dgYrVyMxQ5L-|5pk9^6&M
zjt`7^1Rm3!7^A~oZK&}B*mzMVgvx#~$0_v6?JRtUC#T$`VP0I{jPa%+nkRUdfnpek
zz>oYWVum;KDy7XB>s?;z6T)e5p8yE75vC(>hF<iA=hWl;JdULcnnE8M&M`?7!8S@*
z%86pqb3vsSJVVq0D`mBkEIz^~3n586=5L}Y;EU>jyYE$2Ji?Jaad9Va=2Ns)7~F*d
zDi3zj0<K^|ltF%)ldxbCC&GF`rC?a#NqUSK92YvP`5m;Awk>_c?PEiFipMu;2_Gmw
zv>+De#ypWcXq%$=veJbz<DvXqGTHG%PC)9EBMdT!SlnXepZ-csfDOT1cD+N$Jd7LV
zj5^`sw;tyOEKLb8_RZnLkP=U%6mi`SIWD{gw3<1naamy)pO?iieDJ+@)0_8}($Zp^
z(x;8>y|e=jRexhlizDmUNg78Wv*3hXtv=RP{+69kRx#{Wt}1bO3-A<M%wJLLsm|&j
zR&>iSCa|agI_vTZZulV+U~EL`y|Az#zGf}mU0qC->9I6Fb!)*m;{wp2Ozfg%XW}va
zSM-{4F&8hpq}j0*v^tJ_<6f>w6z_9&;gxs`^;dWS@Q!hiva_Jz@UUlA^G>Ip?)u)b
zDf2-4UDe+Vl-h*x5#^is6?RzPa&IFzH!gbJTryo^H#oGKwyB67F0qq(PV@eg!^BaJ
zC?c+IC&Vjg%o6QpM=Qrt_mB6}&gOHa34eKHctK-DK~rmK&AWmayW;e$8-+hej~@Ig
zZ9IGA1x>zZ>dw8iqNmnqr^(rBx>8>{I==8dO}!pFr=aC53^~>Kd>P6LjitaB1sOb^
zkwv+pOhWF)QIGUNOKg-kJCckW8aL^y9FN>UnbN5Elmf;?cq4E}NN+_`+}ldux)pw%
z`;NsPM=PGNsOCat>zk)#Hx*+wV+&VXgpx~mrgXSx6FSB?W<?WYsm1}0#HG+)<YD}G
z1w0PZ2tqaYZ5DPhX7Vt8&_2Y2e*<5wtkXk$_@)!lV`!OoGm6HYpoG)oV;0r78J}mx
zTh+vS0|SoWW-$heANUut2RFO`Nn<zo&NuyxyW=P(I}1?|Qet=+Xezs=j>T_rKPEfg
zz;P-8{e@S9pJVX^X_YaUI%<@OdZ>OZ{)j~rf|1<Wx!BZP0-ki9UHocCMr0Mud?5XF
zcmF6o+uTdr!j-ZElkt*v;ffP@08VV^blP#ZuG)ZdQw{;q{2F*TD;^U5u{x)^;jZ@K
zGz*W@N|WOgxAzXxt$3W}#!PzSPAj#AJL25Z?(up0<*y&4FCIQM{=<hbPc$u3ZXR!B
z_;&l*oo{%s@B_|%%dg-9rr7O1@{O?__gCGr8MQ}Q!i4+i6XxZ(<!5L{{i(<3_o#tp
zeZ3i$l`(5H)a%mvTBGz0$KSz_zH=ZS#&Nx&2QT?y&+k`F$?sMFQMi7S`8H)HOu+@r
zfA)X>+hb#Y|L^`sj!SuS`CSsSj_l+2HoFl3Mm4+Yk<BR#tu#OhpF@1J$?0NBcs?Qf
zE?ahUR{C%rVGU!%5b%|;!(i+*_>w6Nm2DAakdEKBKxBpu_cfLbD?*j{*=7$ygc-5s
z5W|%1Cu@$GU4=Zj$-!>nIPeem97C^Cy(mwFDW=51Fle|V<nXVQ!pP=tedZuGfrL@&
zNKL*-CdrI|#1rgcB01CeSD$^BKKbP@y%GXLbR(UjK!Ac}i?T34xwUvJ%zUMh-v9mw
z=|?~MVOm=gf!AovF)XYUgegMt1wnRPF>eqm$_KMbMx@3n1k#1M3+08FAs7_p10ja+
z!L)JlFXbi=GjU?nQOaEBT|Z3DI2wW@4zJ`EJG{i?hyeST6%iSfVXXWxQ>nTkfPpxp
z*xB0JNe>=8@RoH59t7J=qWu4`24R`f)dTZE6TzA=)3h=}-?Sce0@SS91aDEoV1Q-8
zAts%K;h~gP0*Q!KM2uLO<JH5!L~RWwSScd{N8QbI3Whe+Q-8Dwgorlr<V|+q=Ho7N
zJ1}u#NHRkq=He2?m;SC*pW0$Gz4yTn(i?BSnX2`QC(|5n$m(5IwKBm&xZvYZlosWt
z^T7D9nvA?)zS&7QGana?Q?GG|vY4M)F)H8*6KW=Evr};d4+(jaqDq2azzbz30yGbw
z;Je~jQEFJk#MgMJxYUJqhJi%MMO>7qz-*4POinwVcMGcgKH5d$5vBkG7~UXs(H@>i
z(S;NycDF?tFq93jyKm`XqFN>x7*Ce{AP_(xbD{Ka#hg`r&lk$CxfwA$`g4j(+xWmo
zvDmJc-p&tKzKcix1-2MmoepD;t#+|(XcT)8CZ|)p$;j_vT27ntIB0m-swcvP-=OTg
zD67(PKK8&dVmX+d^i4e5vU^T|IK$A{$Z9&JH^Pm98zCJn+@fx_7AL==KgI{%)GH^K
zKa|o)_$YvWc!x_k*)77BaDzPc_KP2lJrPUB2^YA_>TnU!KH+V9fO8lZKgO>dEy6j>
zh#CFBPnq0=#xsL`5ZkXCB{~|#<>IYp7&K!?_QaiZadUi!!9;}8&JJgrOb!!iMmjn8
zFi>`;e-qALgK!a;-<as*2W$$a$jFkj9$F%r%`kLHG{5v2!2{Oe6QL<f;D^5P)JI7t
zqDCBInNPB7PHC8!?FAkszMbum!xod>oZm=18{F(abRYC9Ifjx<#_b~P83NG%w!Vaq
z8gn#e;U4$ik)D32G%q;>|2GmJqi_8FtN-R77MFkJ=D+wa|2u}w#KqVMO5i;h*N?i~
zpK#$ZmdSEC9*#+@h3=Zj_cYEUlx32NV9<2%Mrnxl5SHU)sGtRUdcZMg13NpIuri5d
z5?!s%Dql^OMFUSy&P2CDaL2A=1i&0+$S#iQ*--W{8Ah;S;T=Xa0Mg422qvMdW=6pP
zd^sk}$07-^=b|$gmt2${#|_2#{VE%@T=hoT1)cN}yrQKX^UK|p(9E&hE4LZP6W-ov
z@~#Q{&dy%?tH1hL+S=Sqt%ZfO*zQ<Cgp#h-Y?)r6EI<*5qK*~Z2*wa#afj4rpM9RV
z6XU}Vf0U|?W*Sr8aX-_zCgySUiWirN7h^Y3n-#|#<<eO&T}JtU01N>q!X@|@c9_7I
zu}G;X2jEA}&rVe56AMKUzOf1!s?Sc75Q3bh{r!DU#=YWJ>3R3c6tF>n!J?LF!T8!c
za8CL??<!)E4Lf<bN1-~c^u#;#&Mqxn;PUu6!IV|%td?fAJ-9_VCnn<Fq%(vYN)!Ig
z<LZTQ&al1`&Y>iUdoPmmrjGEywBLc?x3hngxN{?fpwq^iZcn_2@}MnS3(eGOG~AZP
z74dkoTEO8W4}~Yks>^4ee(L)u=I5(v<qr3)*3~CD;x2@HzVlQ(j0z>#HPY%*N4$l2
zAhp-bw0I;jD$2n;&+gt~>U(8;N(a63bO!#ZT;eLd(z+vlb5>>ZA}7_Ag>)y!`(8oJ
zrR%QWrTU3*d}|kl;DzvNA+0ShrB-dm0>()imA)juox`q=x#Nh(Gv!;WH`ATFchl@#
z#r+KB7<X%NM_AnXB)kwTxZ{>|E0vn@faBZ|j)Dg$iohM}?%_%K!A}>#IgFW)^QJ9P
z%z$S?5>|EBS^eHw5-e+OXO7O>-`h=xD#O|7x$zby5;%*{k8y+5+w@iQp1?|W@*>^X
zwS=<n+<96MrpKl8cpCke1yJz8>>lv0uX@b_G_PhC9;4Vj7d#lBpcC|M1m_%+<lU;m
zb9O_)w|CUO%*`yoRvR7c?YTTlOUqV7cLgVq0Yxw?-Z}ccdvNS@D+mZ{8ml-GOZY<n
zC2xeB%<-&To0>KMN}V|_4gQv6nB(Zi*~BIKJ~m2yo5gPz^k?-kgr`SpZv?z3K@jM&
zld-O`(fpUvL6FUEMZ$KEk2H30e_vjNra+J-&yg7KGH>tsvkc`xQNz50H^vGal@q_9
ziF))#K?1!no=^`x@bmOzD~UBn^NwG}SK=#wM|W0mV48=iA8@A|un2yh5mqqN`03-f
zn8&DW>{j!UVEwM)!I(;sp&u>_^^u{S84DOMF4UIL5Da0h=%l`iLlI&=p*B$Y<EXos
z>ryu2d+rA;^iXc^NF%)5sk31zcPp*PpUFx#z|-?}`)llVABlAAr!n4e*PrL|@}Nx2
z)eTR@!63*+7|!k)@EiO`f%{V1&^$^TL)yp%@ORtj9qktLC+IZr1oqJ$vXQsaG3;fJ
zIKV_{5>OeCH*H0`7Qrm$N;nYhA88G60Z+f99+b;@$?y&sFsDbLAMuHsJG70_WW)^O
z>ETUV)0Y_If=`M0IJArL`R?jcx_@^iy>WldxV5<z_hOO_Me+SU>Pv^+UOMh;e(Iew
zC>Eic$VSxm;01NkpK><-*k=9!g-R$Y7<U5iuG6{3SWeL)6St?p2o6vo1oa#r$d0u4
z-)3><LYlZeOE;SDvlE9My6_&z0Csma)6=I<L<4rkXY6=!1Z}d!*d$uZ!kZY+IDKVG
z<GIHDJDT_2e={wwu89YUab1*F&_hNu7RR({{9<RylIU1x0eW-m@d8Bz{1b}v8R{o~
z!{gz#_@tAbm1$G6Rgb@*9Xo+gZox0G!+2$}nQF6Kl@JRISUf?08CQF#VGKv)IWv=b
z#0#AX=Ibvu)8`K!rRVFL9>d{%I3DrN-FvC2agBulSLbJG<LSfn$zT1?X>;R6T2uYr
zeEVJRB5N(Jrdn%Jyx&CH)%d%++ZA8XPY1LqW1!j_8JEz)@(PqCvFne;LeZWqWT8C3
z7B+(&<j^r13|a(^$1b)R@l4aHDcA#l(x7iKzSEat>|_yE@R!G;5#8fcA5R^3&8aL3
z3m!8r8LMRXqAleilonl>GA;D6h|~0Yj%lQ?DvW3?xK3X1I;IuSF4@6-`UHQZ%NWUV
zeC&GT8=h}o;n^uiKc>%JiRtH9zit-=>#^y+7eBclG8U=XRPPyf?kNv;Iy3IFvlU(k
z{AQ5{umX-&)aeO1riz{81834j;m4h9AzR`!2#xE)^Vs42zfkw5J+dZOo*!tj?-A}1
z9(&v^Z)TFrBw5{6UF>32v073C4M~6m=&dj>=Jf!AfL3o31Pyzn?;v2%)#z@00gdKu
zOEdjKYN>XTnau6xU1MK7Ji<M~V~_d&{>RMSZ`^xXNur9-Sj6LxJ+>Y@cFgS9IkvFK
zTRfZMy3Z4wD61GlfuBANL@@7)aUO+NEVLU|-7$6#58R0U^6|fvQ0ZR^SEJ(TScril
z>cz&c#%tk4`LjUGe5J}kUr9q{m)v}<nyTW_XotgseR7OrF}VkIG|lj6zA~=%18q}x
z^dHDe;ZtYDb8xI7V?D~x3zd1e|0XTZ)3#X@cAdU>zGcPN*6yL{jWLt@p#CUh>0d3M
z!(w;7spIhbKcC`D_U@-muRrDJ{q8mY|DgEdUB<7DANKsdYf651{d@8Hml-G?Q)1jr
z)8G0iG=V|!HSl{s{QV#EL0*3HpWXC?lB-)70+B>FM7$zy{>%Y7>14l>cCS5*Yy^@i
zY5QB*M%bPL@a{D^08<SH+u~-Sq#AVz5EVi9fIVFKC0!oJmWS&v`vph^HsbsBhAB>+
z{3Dlam%pr$R+v)O*Ng-44n`2W1zuRncYfQ6H`e?v?*8s^v$ZEsfvaQ}7P+e6uffa{
z0kSLMZ0Wd^xO~C@wKU1!+Sy6FJKO3cW0j4K2Z@XwH85nMq~VUqVIoB?^W><*#l=NW
zsJumYoTD*fAc~cSaTdLI*W}3q4--WOUM4AY5)ir>UNsHElRg838>T*Das$%^BdWnz
z@H1HXND(n9Fm{CH3EEHsdJ~=c@L;Y%0cH{}GkbYZ=aTq(qmdeVIE%P2yromXco25V
zUAJG>0F%Koe3Uh?>?w=Bi&BAK8^QnsWpZSEb0UHfqT-Lv1ZLcm9SwF@!U=q#J6|Wo
zz@v~#-vbay@*~4!g}R__5zs)i;Mc2Blpc6TL1zJS#s^83Fy5sl=Rpn3bMp&n?f!i$
zKVb@(*qxin<emv8q+%Fz{<zUTZedDQCmyURjo|RED8b7EQ;i$&nh6ur4KqhNF<~K2
zC{uyc<rgyru3UVD14m^N<(-IPLMSsk?Fh?Y4=uvbGjV{?XJrv#Nrwp=MQ0@{dHW1%
zl@ErGvM@2BHia$2;uV|96Z&Q1WBQgGVLbtJJe4S?i!l=<cth#Mgjs3euZKIPU?e!^
zaC#D@a?}^FktT*R#+#q8=P!qeHkAVtT3)!B5wRa`<>~ri59TD}Q~4UB5y(#ohbiUZ
zvxKqTt{7;LFEEzF5XWKlFT3gALHNNT@^du*gg3uYyIm9BxH+)nIUFz3cRZkvw1F3|
zUSn{0N}T*%0+KdiB4juNv+-Ec(+_UF-#K{rIbHk2OOMO$O!?N^amfqMu;UmYm9MVB
z1$;!{45PzlgR|e^82%YXIhC|}qu^N3hEL>*HRpu~$W@^D4kwr0@2KfWyAwMc_@Hg{
z{5gCM6fay!Q*Xj%U>rS%v%$ZUXvt4d7zljAucU1wJ1w<1QYXbb(&neQbkpW#{|rF*
zbmQ9HaAJ)&ZrLf^WCBKvZ1`hUW%Q&Z)BOrO#&;)O#_DX3^!1I!Z>%|f1aVMI<Rw4c
z<HL{RgAYd!f4p(>j!eQ4)V|qjKP(tr=*x1x#C9UT-aEej;NSa?f6RR+@CHmO`!(4_
zAO+OwQZ;rVbPYk!&Fz&Ym<Sl@ht>aU%tCPMm8+U;d-2RgTV-Tt4FWTjO^Nhrtl=GZ
z${Z`6cnggW8YNU-gly{n)vvR{l?gArQQXhN(R(P}nBexOK_QR31yJ}LA8{00%Y4d!
zU`8o|5(}jV<I-T?jg>{Wf{g_iH#Z2(Ii8FQrEfI$Ab?@V#88?T9q|OWGBfErFh)lP
z(wGW5F*2CAe`97+eh57#hSRjZ`;=DeNIa1ELU#L9rzX>ccs5pn(oPUmSn-F_NVw44
zJdFuyjJYUKS_lD7SfHe|E3;O#Fh;ZbnAM~LOE?8)L{~SSXfIVKJ?Xv@&w|njArZm~
z?qWFD-?Lz0M?C7w=P!JG47;FFSdPFKi1r!38I#y`hazWtXE(jtSWoL4n?7oWBkd5H
zbAdlflhmhiA1cuJ&h9dVGQOWeWw;RzQ2<tFtKwnDEnH()$?o=scbQB~Op1<%4D0yl
zU|N_JUMGZO;dMxLi1Mjl^$r}Y0){6+5yc%0;G5lJ2-8psqVya=Nx@jJKa>rZY9k0}
zR7CGUWOa3e75ry{O*{ZR^!V7MS1EJ<39w$YTj{KY;QY*Y{P@^qC9v`8W%`pp{-gBb
z+1DcWi&Rtnb0-9}b1D3?LvQ`Xv-I@I7b@SY)UUdoMd2qN1O5V<9P-Wp_`Ivcv2rZ<
zf!ARX61+V;604w5N^-9X0v7l+1d(FvywmI6+Ol_jp^WAJtfx=D)bpA0yGqLo>|Usu
zM(EFcq$fL*grm{XiNsx6<KlNFC#K{!t}wCq%6TdS?pnLP1qQVpm6=uP9EW#)9{2Jg
zghEKg;-#`G+zYzkLj^CpQ`oIGH#_U&EEgBpDK?XQu{%4wYI8oan>)jN?3C!3BP>z4
zH4ZtZ@Kp41F5JgLI~JLtT!Bv|EJrR)O-;p7g?fBno{t@@q?u{uBYz)J2LG+U(pUO;
zE767fEx{Rr$5>U)BCQbq1%Ga(o#N0x)0c28824@<<Qg66PiM_~YBlyO1f5q|5OR$T
zD&EyuIz2k@PD56{(+4u|p^v{3?%0jQt}qm7)JHic6+tWAc_>vx^TMy^bb=Wr3c_mY
zkYlJOr^ZaH(2IFK7f<jqj_C8^;wH7!Hrb^@dtsL_^D>lV<zx!-r@uw8$l?JOK>Daq
z$}f1~O_^Ju1OPC2d9RokEFlo)xG^5_{J_YuVbnhipqW7MM^x|*-;Kad;{g4;++v)K
zu{o4X<d^G)H*h*V-tgu5+mF5#mm!*ZBq=Y-L^vt>r7xo#%w@qRb1VWynIe2=&P6(E
zqbOBrkHE)+q0H1l%wyvCFv8O>hhj?e2#3_K(j#34Y!+IA57(a_xBE<cQEsKJx64LZ
z1(T2clO0bh00jbRyIz%E3Fmobu7{XE1a1Rl(p237tLBYlWkI{>+=;1=0^`74jNy@?
z%(ycT>-cB96?@F9VxGa+KwQVikFYj`k2v9nclbxy;TY}cI$~}}e({d5ow(UVJ}SHF
zx-2S+a1j>M4IL)xS*49cO<xGeX$J_dk#AtYsgKSa81^`Oad~D1D$4D+6A3=2&pUt^
z=h#iGz)qj`!6FIx5c(VFBybqz1&z#8nGZ+#gagGREd3RIV6{4{wj6gmu>+OkeL2by
zg&edV^-Sq*thh$00B;pX$HtL_jOFai=E%82^)HO``}+s>Q)IDcYLB85c3$lr9Hi~-
zUH7r<J~<Olvb(({zTm(MLfNH)LW0HB9Q8{-j`9aoL-+6(@FyA_VCv*}Wqb{(AGp+5
zs_w>0E);g)9NuT(`pkS_jh(6D(->38i#*u9<#A0kN<JJNjdBTGGAChd;V8t(@hQQj
z@kV18{7+ls=94c!OF#egF9iFZ$NeRZHS=?G>Hg}n#=u#PJ9FvVAAXoVc>l4+qN){R
z-X$fThF#FmJbec3BV--$E_nx({6#>*dpszCX;+j9#R&L_`<!A|G&F`IV-#sJPSUTo
zy^D**Tc@In=wnV#+CCwHlMq<Ygfir$d6JrXABs;rVkfog1^OdBXq0i1MH<{m$u8xY
znTp;Vvm0_j_w9l`#&Y_<=*P;N2Rc$QZM<Mn4+=8M;Bj5!KYc1ip=@R0#L>~Qk1J+O
zL74{qLHi@Z13Xtb9+7rHn={`aoS-a%7CEL6r5Ouv*!_=!jh+)6v#5(>A36S+(*S6n
zEa0lwkIZKt2~RA7W(RNNjXW}`E*K+XL4fN@2IZ^ARnadG$CR_%kz+lBUlgw>{&MhE
z^>uNYW<}d51p9}_AX?*(Hg}|Y*xXS)v{1T<#8n5O<mP^|KJ%K@nW?lmKPMVfJt%F)
zJ17ADwd$iqCo~5bOcn78+?9(892w~T(MdYc)4sSi?=Y>o%9NVd)Q~VDn7EUb-C!um
z_70EI*U#6}#`eA!LzR9I6a8cEe>>l{<ky>iy>6J-+<U$;w&Q-+Et_9#?|s6&5vKcT
z)9a79`(4bt+)1$b@!K2f-ErPxewnb@kiKJ;xIJlf{IKWuyUbnxUj}Yy5A#ME`E8?*
z-ZcNs|NKAx(}Cap;qU!e-LyV~8FK&)(tI0uK8rp7;i#c3{v3e#gp1!2HvYqv@cuUg
z(6x8N@vU^7mO84=H}A09=*^!8#q1W}*m8OWI$6!c=D(Z}tZBet*Ni&99Lixzetw75
z(f#&Q^1>e*aY`8fgs&aq=}n7(hFMPx@Hpu$I5duS;mP6svUh|r?lN*1yh<H*3}e2%
z`QekdBl_Vla6tY_l8yy;dxv`o=YP9}KMED|VLUR%r+VdMF&x{PjI#2EfzILe3p?pr
zR~~fOHVPlGDK>eFiG}fb|AY6_!-o&W7*%}?3Bnz3*;R)g%rp}K4c0Kkz$f6%EI2J5
zXQnV>(A-&4=E&QOtDazFgPFDB%L67%H^LYOP3#Ps;f^4Pl8_JIOcq!v%<&^$X^5aG
zGZMU69kR2#n~v0h^9@ks^@e$i)30KzJfK2UT<R1Q#KW)LR4!IZf{&;RyZYB5vARC^
zgf95ToncykFtHkdsCUkFX2l8kbue%QBdAE=2L_2GAvM>5+aKS`IpZ4~awmX~2*h1^
zB5;Cf@Rojs8PVWfou5zh3kyzzZQf@bTOhh3Ju;*oLogs(6Z1m)*9ae#R+J}b+yX;+
zGjU>NE9vtG!v{S=8>ESX$rC=gyUbRE#=r}cMTt$<Ic?wqj*b#fpQ1$B5u6a0<ECIJ
zV<n=G#}JKpGN*|H$BcMOzZe@O<N9%uq$60Iy3z%&B`?06H#9F7=#;p)+1+q4p`={}
z-5E4JFq6(lY2gOL#xWs2DnJuU-?>Nn`YUZkKaqZbWn=gJj>jJzcS717n|`q6uLzsn
zY5T=Q_y`mKQGVPZMNGWxn*GXmyd%x9Q^4?l9eAhbx4()-w%$NEXS<B;#jD3J#xKfH
z{DC2GN5FX7YWMgHyB%}*&N%3%ll^3O$~|Er0kBdOi~M)|vNiI-gumVR%WwZX?d+zX
zt}+9+^5ita*=`+E!r(8U@UcIixH>G~frEk%9(-Q6^B;EVg>rQ2n-5zY+?EI#4!!uv
z-W~t=)k9*o$So4F-A^`Q)sFxud4p?e#qs#IYk2bOB=iR6VyH)ZC=rMBx4mRKmTdSY
z94Uky(}7W@q)iyyyVRyXSty%}r+|ef_zP+(u`&AAr@N5sZb6FU>f34iv-2DGB9eao
zCdEiV&Y4~nx0|m&<r(P_%~AQmi9g~wyqr1#@GAyC5(&H0vKt|KpAbU*`A>d*im(6R
zKl#7>7*58MarNn}42Ks%iDPAl#}Z)Xj!?Z+KN=Z4{<qc7$A!<VaE;>w*^Pj303kr*
zu$2xDkA2Kgz0vehBfQy_<6~f$I4eGaD+IU*N-U5X3SP>KAriZ3gfCSBt7$n3hg~pd
z8VeB+vLkFtll2+J<IXRHS_pE6282t&%Gktik=C&$zV-cdC_b)vR8Pl^y6@F$*6ZoC
zd6?SA$LZ$0?G>5`bp~!OQ~&iv8XN9Y+8R&fX!oS_*p;nudsuuxWqLX-*5=Yet(umX
z7Sr0=YFf~HYKmPLw@Ujc9qsR>?alSHx3g{f<Q}AnsTp4YkI)$&usS`J=4UIK)K8>w
zCi%l7R+0>>JhjSnT9BWQ6%>6~(4g@Vo@7X6<mfmn22?rljtH-(SYS261U@B>Bj%{N
z`o7A{?w`i7SGyw+;L>{E+ci3wW@cu+a{TJ1KOL(sPsGcxYX%(qzK=Hi;I(moCK$#v
z#&VAb3KmukHx3U~|1I+_2$c{#vgVwl+o)UaJm84zz5PS+nP(!YEAP5O;WerE#4tx#
z_NP&mi*kseGS7n|$%17)u~2Np3N2P%bMy}PRdEat$J4MPo6G-aE8|{)esOw~HeNhQ
zPd@)Nt-pMh4)%6LpG}2}OXuM!P`1u5Ec$MdTCL`^eN;4*M_(k`V+Yx^>U3UpU9DCu
zbcN?sMVRlPeZ#Ydz%+!}2vj)6iX-k&bifxP?4c1s=cRoy4?t-`ezcX%t?krk94f7m
z^ytxh>HhtD7AV29q0E8LL?Gl{972<O6O{vAkRvKl1kx^JF;T3dX2tiIaD)(&U7L;%
z&kL=o{L@OGJ24O}A!K3)CjC^u_-oodG%#D4G@VaRPr7`R+cY<TQdje&=1FR(j*pL;
zqA_;kwagC_j$`9#$H40zgg#d8^3*?Rz+*$lf=_J%fhqzx@WfG4lz}i-9GSNjEeJo{
z7so=vVT9layoFPgu3{uD+>EpbMSr3%)hhzMP~uIwEHk2&{#%5dZS%*~xmodyecsi0
zsJ@}0valFnN^Rp-<>a_jdI2kpgh!4G<8HRvY&9(|)GWC4d`fc%_<j}z00WA!5y1um
z;GfwMfr1VNH^NN5eFSUk%)4n(7_y_3oqF`89KA(fMw<uU&<w|GL1VNB`fC;mP_NLd
z6*;0!3l>EOz~FhRaS}=y+Gx<O{-_uIc&xK=MR|^aBeaAmbspn(^wIRvxdqsbbYd<L
z?Z9^B8w#Xw4|s4>UFLX<LHM&!lm3Bu1Z^g!VbF}`MS3U?@Bss`@CDq=J;09_8Y<qE
z@Wnh1d|)Vlui}cGpPs*{jGpVsI4vGGZs9<26wxpO8}&|GMp*7L=#PGh_GjGkD1FL8
zJRXt)1p1PK`v@D#thj#;fiLmY*c_hujq1(qJcPkK;h7&^f}Ob+Zw&W4V+sv;Zj=M_
zi*HGrr(-g#emd@fV~oR-7j>X?>4SXique6zsGrEC#1{@^N*cY!bq9`ssFP<8JiV~V
zlr;D)KGdVfJ*7ikTOp_rD4wgcbFPbq#yMcGRA$sZhD=wqyIav7b7qw7j7{7XRd#UE
z_n?gTdu%-V9FCGz9O{XF1t6dsw{eY$&?No!jPhY8En@`A8O9TiPMg*E&F)O@9-FDe
zy-(00JWQX`zY*LV4|k@tIGWjX1bwLPksp!fJ@FRb{rJ1-H-6)9ruW|emUnhlDho<)
zKCP}jO7A`XP<%xt9f;3*`s`);{L8P>tMykJyHC_mRcBWm=g2+VYE$q<8cR6kXGn|<
zeFwWUS%^y+!3jz#>KNQo-^};9Uvngfr;O86Q&5T8Y4r2-bLZ{zw6%SZ*0<_uzi}eo
zQ0@3eZCv<^<6UFS6w*{*H$L9=;lq305xupwkv{$OXK8bD)4PQ)&rSue_(YYV;iK_d
z!rP!1g3O4=AsrTXu$aTtp}K}v7}sLqNgSKX0u=B~5EK!<*H(TgdQkK+?s44q)THWP
zV<dN8(GS6Yc-+;v&3MMHC3bUh%rbBDyv=<4tivL^Q5_1mZP8kBXJIq$^i9wseE@YI
z{du$n`VdaOVbKP%6O?7}QyfL;Q<+o`r5$*Ihuk=-&~OP3`02L!1KQ}V>z1ZTdM)7%
zC0v6e%GD-NV0jURV4+WA!G;%`%uJil0WV$vBRT{ZDC(??6AtOWJq8P2`v2H5#m>XH
z#}a;r<Im|AJ+_Hw2d+43IUX<Uk{|sY^)xBo16tx<KKPKSNoY-v>cvNaigy_w8A#I;
zLw-0Slm$l@r^l&vxR;hRmanZXnU|rRa?E(_5NG%Ox$Hs&?rI0PLnrhH?8I7{ucrGe
zH9w29)6R>&g*Ff;1F!&`1=^}#;T*dPmf9yx%^`Nh*K_ykNtzlNOmoxX+xxE4iRKu4
z+ncG~YPxNF{p3md`Ild(-FnkXGTOTUHI5+4PXE^P?tK{N2v_{dyVx<Wx%YfyY{&hs
zTQ<Mg-ur}kBTV<xrq>^H_q&*Pxszb=<M(c;cgMS9zG=8@zHz$Q{G#x`GR!h%jC&)E
z{I=0YZ<;8+`u@S+|A%ye`co24g^o8|!h{WmDpJtD!^uU7m3T!!VvTxFCbr29o1<Zq
zBOOV}qqt*mab60<YDXspJYt4$cLAIq0Z9XYh@HV9zwva$$T#wnH5<n%rhGVP#37G2
z;FC4V3ImBt_Sb(&+o1(3jGL2k2zeR+{oBn!vL(mW8#huU?I@c=7?v>M;NPA&YUx<B
zlTuJ*;zwx0$mTR~M3F&6+{2%+_=Y3@`s)-Z!;eSjjV%-+7>AcLLWbS_z4Y0aU!>1I
z`NXR?B7}Ykpa;g3T{>Jw4TGUkOtBf*xEo<SEw8PmAN}Y@>En++_R0<}frg2Vi3q{7
zB1P3~94l=GMAD_cU?{;Ce`c5&aMTH3X<!D?X8!fQ(qPC$2R8;BpFyv12o6|f>A{9D
zVs2n?xHrOst_Fnj3mC$4Z=Yw7XF%ohOs|R%9BfJ6-Q7*=>svnN5ayAx!w~&-7-|H@
zfC`No(jWLPbx4!Q*76o_Wq$`T;ek*FJmTS&+meYSA~ARmj}67KhZ^{qP~vxa6)RK0
zEsOyGF~Hw)43FUQgiayBGoCOQthND%)C2LTKPJ8i0x#Oax0uMu%4Ax7^e{bo{DJZs
zOl>hl7wjTbDgwnc%8K}F<xirl*7j<f6Yk9t&M^WVw8Pl+X@DLR!@$ZR`nP^DD=;xI
z;Uo%OaC8BKIP?b7VVY1}&W`3DEYpMX0CCVBN7P&jCQq0Z2OOdFa@iCXg%T467-Szq
zBX<UH+5pT9#~`t?9i|a?w_Bx2S%A^IxyWDmf$pIp;1Zz=2?{Sbz$Gxc-f0`6(NH)-
z`(o@A2R~M=QpenXfFDFq;hwLx2WhS{m&QgWth9rvhYpmKi~&+f{SiiaLJ_!-ZkW;=
z$fmg6KBxF#&wUf$Ihx5U`*n;yqivHGaj<|7;u5DDKVmt8e!*Wioo@e~-SooN<I=VF
zI7T}0$@qiem-LExBaQ4(+$p-j<c%w^*s}{`52&5A2<PznbvoGe(;K#=86o+|X=gVZ
zy<86cK`K_r*xf0HJzE{MH*QYL1>nIi0_AU_g?o3Lh?h-xIjG^YpJEhu{eV})r7!xT
z2omKXnB(T|2Uo|B=x^X7dv}XgGEyMha#+%alI>_-fd~;P#}G^<xA-omn`e(tsbo9#
z;F<3nhLr8ZX9m#0U;B;6(K>Ja=5jb1U9<2JG)-@cKjH9_n~RgfXAkGipW*~RmR%Gs
zz${@)?u5o#Qjp1((gT!`5jv;SlQ!_No11_o9KTN4Vf|zOfT83?TC(FPLy9H5W8*CU
zoJ{BEpZ}|0-yL87?SKFOzJ(9qe1C?r!2wq3hETpG9>6;$6iz*?`)maDOn&jF4Cm+9
z>Bz#`Gp~STMK$~ie24`?2s+%?he9nLlxzt52E0o0_V(O6o!IHa#CJqv8bTm;n&Hom
zFUBDhEF3e2Ab`3XQhz@pvY8Yt)6*LBMn<d{GX6DAqiAUzHE94EJD?GcI68N^INk?A
z%9&U3#*uLFJqX8Z^Ya!ebJQ8a?8omtw1UP04ULlr^?G{s>Q!nsPb`#LRCvZ*##eUt
zBfQ!Yue7nbkzPE1mJav#J^6ib|3UirH~vPNtj=m;j-X3$sU-7r6-~MqHD1k#&QvnS
zH^GeZfH4<g1Q(MdP;H8r*x1<-p3l?j(o&ifob)>$Gc~R=wsN$V6<Wd<N63IL?$lrh
z4z)R{^u}1h4PT=%cwlJMa5DqAR{6L9ysok#JmQGL+lyv8QW?0rY;1fgT?>!gw>CRh
zwSeRFxRox>PBs3^o&HVrjez&b^B3kL-h24K<L>Uxo_NSTkL}B~YI<~kCC%4nE%-o~
zOnU+!jN_CafekzlxEl~&5Jp_6(VPiJFKQ7y@M0VTXQ5N(!8mFP0bl)KFMayyCssJl
z*XGTKbBDyl_=M}CTC17qf}dcuJ*)21&1E`o)vd@I8J`hfRkd(cJ)xkxHXnr|2Hpi>
z3OM6<O|e2MBlQ4p20uuBh)I(h%BRNZbvigYO)a$%+QM9AI<YGd9v!|C-j>}Q9GiC_
zxY=<wBidqD9c=<(3ovqr+Tg&gS0y7DLeWcGIlZ_^M<=SgCIY%xolU<FpNbOC!hV&N
zU6&z56JMW?+(S`hB?^iF<pV#rySJNm_V&FC5aAt4o)yvC<M-ZEosU`}gix9L)S4#>
zgW|h&lFn5gDx0<q|4Um&G07fM>XG9eIf5Jh4fqg7SqY*1i(rj!a6Y6yf!$dl5RA6x
zBTCzG{|a-n7*OdhRXemDCCPjSz7hc`w8gQMD2UkQg|KsCVpwkCn;};5<lL#-WWJ-c
z5yTxIHhh%n)Fka>B2}t$;&-Q1KR2n>)SO6g04vHg_*!;3A{?KtR0UIVItcn`yFgYJ
zh!sAY{^4A8#wu{1Y9QP|2M7o$6FUPCm?2n$Pvy8NAMd2Tvrl~|yE18C>=@zB8I<BP
z-pRq-rBC&B?!Mzt<=8(sP}^_0zM)BQey#Sx8~}qLu$=M$u(Uw>G4I-uJ97c}aK1e@
zI!qYo6(zUdf}8Px_JjhF@xe#*4TwL_#*5h$gkDcSvXdA6a>PT(O5e?{Ot*Ojwt|s#
zgx_oR>2d5OcfO$vCt98(f+x`^?ScM|T?OFLW4F?V_lB<6J&0n2K7fS;_#s?wdA@>x
zo;r#71qRr>z))_qdEzohWS5#3mI|-LV+T+%rA$1O8JaIRig`ct5coFXMDCi_n11lX
zT!|JqKFd3DR6djmC}f}`VDZ>W|1LXWqAuxY0{4`ec|Q4)zUz{PA!CJHHo!qx40s?u
z2A7Z-P#Zla!%h81yzGmEOi8P}!y|j%G0+(O4ls~TfwLH=(Mcyv2XAu64$jy$L*K?u
zN5%_q%p!q@4<DuF<>jCO&B5=jE~W=-%c7mLw7pvwA91ce3SMWz{RDz?cEEBEVemYr
zX14+8sRiF5KEXTgL~r!jEDYx46UH3s3EJSe!rI)NkJ&|$z>XvqrbK#bdnzM40GS|=
z7Q044JIW;bUfMg9F)mtWk;1k543zNuJ6q}Uteut?7gVlQ@5JT^z5V^ew6}X;x~<I2
zcmc&>{UEKse34$hdaiNo$ct%E%yXp!JKxt<mecn?{$84{*2IhUdqG20yazky;e&h+
zxzeZKq|RAPLEjE+EQ;WGYWTMy#yyP%XBvB098Mp`QE$-L^H=-ni>I4Dj+k1kRHo9x
z!fcwInNk``TipU>q|EfqBYNP!KKc2l>5u>TkJEGU4%`7nAIEXU=VvErdwa)i|AWWx
zr@!@^znRuFhL6ou(%{IXhLJJBorIIN_|di(N5NmYUxLO26G|m3kwg<Xq0C`-I{Yqo
zj!v*6oieG<0%msD)(=r|iGMmd@tu@Vs27W%C>6{cai`HH>P>wQ3pZ#>^!*%*%g!hi
zOy*ey3u6dx0i?8mPxRzBeWUssjSHpkDeZ`MDU!3i1$u(sdllRS3)!sF1P%obg%s^x
z)TDZ0(S(&g^i{HZ$Fa&!zk*_rHiNQ@aXY8%zAa!7?b5HvhlOhJWR?@co2f5iQ4)((
zX2jEg7viA^q8x)LjMX+6>)^XlT(Ya(a|v~P9-C1vp?I4Dj&Uy}(J<$<-M^^~KpX4|
zgeSkcLfP0<eDMkz2WJ%TJ0E|?JH*+2cy^|?sk*-r-Y7qd?6|}AsL6>{N3Ium6Yu)g
zykWjNmF8wRT9?J+jNhT8<4AXS5O^>5(LBP!0MkG$zxa^KD;nlBhfR*_eDXXAqv9!7
z(!<r7_^5s>Y;QyZ&<Be!UTtlqr<?m}zjc)^Z#k(c#y%DVsfKOGFwcp%zV#G$_suFx
z$IiRQAJYkod(St<cHHl}Wz%bS-r0nE?-S;YxZO{iUVkX4_q*57X5Mak%B}?NhI&_e
zZy2Rd{@>Jm({u^t=Plt&es_I(@%opUG7^>;chmH@K4|mB_^89*{o(Kb*b_o$1|h^?
zNfem7c!v*ucEKM<eoD$Yoc(AOdQlM=6R-XGlaWWsBLd3Z9(+bP8)fFKM94wQCBB2@
zFFvr5M|gJNr?9fJdO5(1e8XOt6XHi)r%7D@mvCXHlca%1x&*{-Z~f%F;)|p^IPiz9
zPAz|y)b;j@AEt&8hCtA{iM^y@_wrIW+&bRM+xf}IZjl4QSU-pfh5}E1BV5TB6FAFX
zP9q-b(?TiSWgN<3anW0TCR8_^7zCKi_`W6$zBYo7FokK1l}|FV!+`UNfZ^ezhv~x)
zK1|DEIHo4Jcp0Wa6Fc5aNX5V*)Ms+Ugv~-CR)%HRqhyg!Jfs~1sL{iZ{DIRe{qzkZ
z<iQT+Di0tiG7zd<$(>0UlLsbSW)9^?9x&DnY;@KdLk%AcKS$vl9v;PZXHAA-_7vQC
z{Z$%Zz!<00k-zn;S6K}g6XsaG2Q4wk<L{@_Cc=SN{K!x}CWJ1`ZZ}E~Jh94(2{*zy
z(j-FZSimbdyu69yGr$SBk;QN@2nP*BTHq0S@I*y8_i+ibbJXF47&*4LvpUrjC>X5x
zl+_9%<xBY~Zzz*OXaI8*@CMV&<R_SP#dkf$Bnc+cyS|i$R};z|JUTC#SWOl*O8HE4
znf3?+oZuK{9vWmqLVcTAQM%qju6U+FJ>)|kOa_=hvQi2~m9mr_W($QOILi}irCsVW
z*NpQ8<z@2Ip=|{K4$21HE=P_;T~q&N{#;+t4w-OsY|q#zm&tS2g7S)@8-JupI_?`v
z#~|afmbCGUB;#9;{p99wxcNtqzMLMPj)tAzh~psqDuVORu&7{8kFdRIhnoyxb2@(X
z6S?pqT=Ja(=G7f3;^DXCPrP^wuQ@Im75AKY*n55rDi*`l6V{a@yI;kf58FC!9$#dO
zUwagk2FwcYJP5CEhrt26-*V4+dG~I3g~1{(f>Cy-$G5|l0AY-GV31Ql7QW!4gG+fD
zhpwlLO-CsCDx9n$Ifr#n|LL9cHC~FF9^x1H{Ebn-TLD_agnPhGvIVtv4^JDT*MAWp
zxR$P)VUp=Up#M@HD_=@EaEhRva`PE)reO+~!<fFn2W2OM(}^4%%MtBSUbz6dJUHpQ
zj1#29x7#ibN{_MJ^kF!|pUi+cLXlqg6@X<2G9hmMH69|A--g;@^^msyqRcs*yo-B@
z1q|jLWhcIUsC3>r-xz1`!#TU@FUkyD5!`809^3W!g<Hfy%#r_doZIy0Kl$0OZO7O2
z!{7Nat8x#U=c#dgnocxnWV}LPg2Ldud7QZDoFhAq59>bKi(MQ{=uo<_%W9M(yELvL
zj7Jd5v4IGzIqnc)EGwxIZcI;%d56WUCYh|>yt&i_wy~GK`s@=upQVi#&(qP~UOG8y
zq)S%#*7s85U@sjV9Hh-x8>zmp_xgHz`t@@k>$ktR?;S*kM@L@B!?=q;oK>yta9V%)
zO3!*~YZ8uNe{OExE5V_2ga@oJWpyfj&`GPAcDA>a=3Y9`v%kBWHaFI-JVAlRQ9~#J
zxX<dW&BfeJ3&SSI#(W$PJG@v44&UINO@i~~ix;UW*brDU-m}Y_omlW!>>6Vt&E%OQ
z5jirkUZ;GUX?KTJq;Y%^O_qswTj3G>Y;0^O-RCM}!()393N)0@q1+Js?AIIKHN+&J
z<80WZ(ywuFWMVwc&ezf$a8&2gC@Y~Sr_;E2QiOJpI0Eu>1V*CBSlDvrdtKY-n%a*{
zq<(}x!day<o!B9WIE39n95+;*nYQqa#a`^wYqnL+oBq_k?n@2P*ZGzB7ESs|OWFAT
z1q9Iu+2CzhA$_cJz-v;E^clcFUBMS1=zIG7Mf%y#er6mrMJF6%iL&tQlp~_J2dI_y
z_O{Zi7tef*A<6^p&id}hA9=O>((;P(oKjgv)mGIzTJa)2PI<wramo44)_VFvG+7t@
zS0*Pc6hr`r;s~YQdApU`t(NjXOfA71JR5u{JPUn*_zi{iST2K5oO`0!WypdiaK&R`
zirNVL6Z{TJ9ge_ZKEN?q<EkHc9_WL6ELeSyG8tYH!hnZE`3sNeosPm)@IA5nhxB87
zQ{N<51fJ}`BzjYs*`2`ggxtqO*W*Y2@OaD_5V-OtA68Pci=Le>@NcvM=0em-iydad
z+oj-wA490drTFa9f=1X`Qv`ze!_Tpk0l|0KL4(rT%1@QcEBHm%@ZXdHUK8GtBdZY9
zaU>*yI<M5vtNBrIu?r7(1Yqn&SeTnlbJZEOO;+^B&cezxeGvDDsLiX7+u7Riv0x~#
zm<KTyaik~8y$K&zHf6zJ?4AszAnl&}(zu83MCGQ9A>2gBM%_@S+)r1X9Q6^Y2+BtK
z)Xw{_(+n%tl|KSLcG0n_9@@L2UqRp|8nBR0u%KXr$EQtt4na8C5h5a3<ybRzMVvO9
zqKmBLpgpSod`zhNJmx6~v{68D{GRC)VWMdEO7p-FB%+WzQ`zX>;yZTp=%Jy?c)kG5
z2{+WEjR+=aK0ZBX$@38Sd-5*xC-W9OVAy_|mxv~!?oqtQ5sQqS%(HmYA9)TWexCZJ
z4hCHo#aCJ=|0pZ8T<)=fp6E9y8;ToXWMl?L;3IG5ZOoGpdY7FbQG`sJf{}J&9#C&Q
z8C$>yAGB-WqfaVt+CA}khzE?o#BpvY?FI*-Z>}(i9ZJ-b@vONL^K|e{+Fld`-h~5%
z*z^muH_vgLI&CcKf+Oz?UT)u{>k}@*(a;!pe|N*U9&qb1uep(z>ecgXo9v#$l$eH#
zKi)b0*A3ry-Fi$8&)ki)9KS><S@)P;81V)_;Lg#6kEow+UC|dzsg4<wSpdS0*E#hI
zj9&=Ikz=5kK-nC}esZrR%1st_j=SIBE<|>2af~6muI4z_)lV%|s_cH9aX&|$KyxQ2
zZSMp^xR1b{U9jw~iGKUqi+SiPPTJ>*g+k||0qWHKoAM_e7IAoYjK*Qc(EYl2lY?XP
zXy@1LIagjced+X!UH9}sm(J@*VH?Ni3MUv&FFlT{TTx%a-G66ZxcvO<FVm#@qx*|<
z>7w0C?PkM@G4JFOy|Q?N`#?F8R)}*O_MK7cYZyye5TMS$?c_-P=CQ_LbTI5-W9~(L
zu(*Vy<LC#$@8J=PjM(Xa8s$>?8pkJTuOS}lOn6D`#8R79+pdW>dHCQ_di3yt?=vG^
zl!=TV@H`yjjM5FI;{HKBy?iAetzNg1m;5-o+Pj*?lhS`MhJaJzdpDB$UUtaFz53A>
znDJbQCt<+=Cl;`K45bzdm0=W;S^2?nnJjvlgpU$W!Et_bl}U{c8heC0?n7nVg?`vM
z#*TFGurOaqD@!#!^J#HjG^DurQ4aF90#yAlbSjbt?KHulTj7cY#L!l>Goebjhi8aB
zHe@1Yu^Ho2JS;M~a(pYW9anv;(g`ISYo3sIg#rvk5dAL4IG%{NDGEFAL;ngLxV<SZ
zd=TUM#KeddnG3a9m8<5l8G7H?*iP%4+i7R_z{ioZn;GSA2PYc8IR#*KnPV&$)BTl&
z*m1R7OZS$mX>FmZaZ~)lwRmJvhpIqzH{|gfJ_w$fvQLP2<*4w=?5q{AC=KCDz>Rk&
zYW$|3ho6C7>8ItTv3Vq|EX=A6@Q5#;VOJM2S9Z)_d2tdnuWA=>)n^_T9+O`CMmRnZ
zufT#2=v<ZI`-A($8xM|6idi2<Lzzaz+g5AyhK+@0)!KsEZhzW4Y?;TS&cJQ(h~0ke
zwYQ$v-QNsT?3mZwd%iKY<9^pIn_q12eZssEru%8r>yNqnUCg`ONwE0wdpFd(<J~de
zG+Z`+S-SZ5KEKHP$}r26;qQ$!^4mrqy=nTW!yo?6@BLV=89<y%0Dbdk$8ID{xa$qm
zpkC7<+<U5n<PUxbSEuQZFb-HOGHP96Gn?Wx?cy}^xAL%`d|XT%&8+B+oi~QOY>&tJ
z65QXQr`JC^^5WJBQPS)9Vajj3DO^VwWmo=4&!6!qF$D>C2crZLnt%=~ySmuO6F0dq
znQ}aUA%woYg4!SaM+XpTW6$mm6Y1i_ZzpXT!k0XFB7VVbI09)7?({+ch(Df02v4Ua
zW1$@HP9*hX>dZKzj)@h{7feu5elbAFfsn+4*&-e^10x61I<CR<{{08(!Grs$rp}Ln
zoRzi+Eqv687%j>klZKd7SP?;e;HH5MH*&}Y_5cPPtZd{RtM({yz|Mf_m1ydm%{Uh(
zR-VKy3?gO49upx}Jh7^R3n+=F2{@}Z1t)_jOb085kB^QOrtR?F))1KfD>bGMF@k<*
zOZtXEh$4ytOc>-1E&>7uW$=UDg9#qRF5_QE3-@s7yaGoWIQuvvNw|l3wn9e(CiQ8i
zfq@^cM6uXW1zixfw8<+m6)@QBUV@Qf0?#Cs6_yOT17b49L=RIO5g?|S;}xi1U;^JT
zI0#8#zOA5CI>5o|?kh3OrYU9b36b)G23ZNq1c(VIMdq#sCXsCA=bH&JaE0)I?dhN<
z=!ca}D2Bi>@R2^{4F!|RBRW9Q$I&mWGy+y*kF=mo=v5(IZYIpiivVS(B1%=#aGL^$
zqBT~f``AXu;nHXFrR{)M7z~9|=KfY%u09^z<pw@*zzI6HB?kg8q8QUG!dA`^9~DDH
zXj%|&lZNaVV%Zf$*%qd_3u*Yu8V_a7p0>NK=e+H%di2UGS?L!rwiDL55+2+WUJv<N
z0fKu89q>8r4y2?fdrsSa8RjzSwIH2;63P#C4V)-zi0AOK;f-->*!>F2-F%%C76TP_
z`(p2L!$jQp4lmlOzNr7~7ba}fNr26d!YGoza~x-$GqYQd3B?6I{?I$VOZhtEON8(*
z-BsklkDYVe@Gr`U;?@b0AHGT3c7KFDezOfXKLIoFxSaOvM!H*`>{GmZ-nmU=w*rLY
zI{*ee`(@+2^G9~-q!DT2Pkg}^2*))Wx$((kKReolvtxD)MWT--2|qcLPPk6s2;j6c
zurYQ<0GFo#t2fqw4TlJ2TF&1#_U5OXj_cpHpaIh#PcZ^8aF09x`pcgkc=(sy;NXV8
zja;w`h@jmZ9>XgKj8`-|tey5pg5;0Y4l;f#75mxtYdz|4c=@+}+-h@t;AP^zBzBKn
zo@)Ypc4EORM~yIfA5q1j45ZvXhK}*#;!J&cC~!uH`{O7QO>SAy+jq_GiCc{u>?&rQ
zKsj-onvK0w-`h+to_(2KeEnH^_55oq*ti2|MC0k)Ox1S;u=<m`tD53Fm}qkBn0FDV
zZ)ft(ihB#;!I<I=YtqfGB~}D;hrw~<P%z$l^2<FN+)**4NiY*&lw?fiSxG!QJLfx!
zP)OD0YU$CVM`>+sEpeX_0xym`Qc<nQA&<eKQIG2g6A;ECfSJs~FRBnejByee;j75|
z{l|~f;}1Vfi_1%CO5X^~xW5AB$Kk=AS7`4D2S5GkPtzx#e43s=f0dr21XLNkU`fQW
zx4)l0|NQgx)mLAom(O3Ot!<9OYgmcK5!@S_n`s?o+pCTA^!amtBVZa8zc~Sn!t12o
zD7!RprxOJDr{d$_^E|m%eJU?_Ms|$BQ?eT-7De@0ImUfX2n`XGaWo!6G6VqZ4k!y>
zIHrw@=8srqD>^zlxl9Mgr|JAgBlFOh9`O<4odz^6qkQ6yCkyhG2J#qIS?})cr^DkT
z3!vZ|SXqx^@8!mZ{Se?Iv_{c3CtR{i4_=whg<b5_Eef^OmDTi}kG`YlBQHAR{wpiD
z;EAC~y?vy&_$%JjE9GS+GRmZMb!NpEJUhIL`C}AoJHoZ<=ZnujOP_x7)3p9-Jsn&5
zASzMb(3p7>8Z^Ix4JFIhUwxkb`2YB~>FD@Se9?J2*xR)N^j5TjFdn{-9dy{KcUD!?
z3!zBn2uOH!Qf0xE?|IQj^u2wCpF7h~{GFT}TM2mDI`XLvgt0<EFu|X4zZa|X#Z8({
z&qXJ!-j1V3V}TX*$!-eho3=5=k(P?zY$CiwF%g0w6j%tspbPj}R#E#1N<HjS92?=t
zM-*}D14J*N1$=Of7dvOly<RAnLir&cQ}Du*vLlXNpeU7T#|R_6vRrh?omd=?gz^hv
zBs;k#<<Bvd<jbyv;XzJ2NaArJ<+F5k-b^=_g8$?w9jFeUe)UD#R+~Y%MSDk34*xsU
z&(Vfzr|KJKlrIVwA9*F(p=~3iM)(Nd$_}?Yo6>hIIvg2PU!i%#t?1`ye<wBechhCN
zmDClcg*o+abF;$fgn3jHU*H42kvUEg+QrVov8>DiS8?G#?HK_*baAHk+frZ0vEtM}
zN18T#r%c2ANh}P6R)87Bqi~yfXyzN92O4G+jm$A<_gM*}1Sm_HD}<m`yftZ}*u_y$
z_)qRiT-^%FtPr9Ot50Tb%RmRs65hKD^)L@A%1-*^qVyzB=pLFxk?Vzdf}Oj^xHE~}
zwJ7Mp59xH~%CfVFl>XSeGnFAUY-Ln5LD62BOUhIK>;UGQIT7<6+$a}wIocCxMtGG?
zo}NG36ucC>UeIe!zobQY(&oM?=!d$%9sJWza+-)2pBmT9b@X;R%!M>|D4gj@4F<Zi
zY=JDDox&-fGPf<p<#GQLWgxwXL@iNfxtsCGkNhbMCg4{%;sB@17UBFScjt@0<42gt
zBc5>WHQlgpH^rX}fg{^7{?7JnBI}^1r~~HXC=G(117_yPD40TNOz0uuKd^Ue{$#dN
z$wKSc&4iFY#+tyN<sN!a3ec}`PZdfb=G5Nlr?lt?kO8=EMRV{a)Cv6?{wR3Z-NJ&I
zdcC1`DITVAl#j<_2N8YssrzB@$_{281og4QnB6tk-kAn}c5$P&J20YWI9;h;IChww
zDU7AH#f3CiTSz0C?`nXx!izo~0lP+KwaM``K2=E-?o6DTPTzj-L3(d(#XD2D$CSR0
zetuYG21lIC0xXmj1$+IVZkf-R_#YN^oC&U@<74qDFVnN9&wUq}=jGIyXb8N`Ofe1&
zs-F}8rF=;bnl@jdG;Z`9x5T^D#Va)1!h?8E)feSPfl?Jrt1D~i@#7CYet-L;k5$ga
z#NAXJ<GZl9q%bvS$%#x)zWP$*@H35h$0|ecG{frnQR<*zV#o6M7>f>+c4f*t(54x4
zRA$;LaGg*usxx+b(>_sR5r?Di7*jYAh#m0rbK)DTvTKavo+%c7k$x-^k6ldM<H<33
zEYe`db~(atRy+#|HIH8k!?D-&X&kAEe9=b)s;{OsnZAYBE73@ihqz1YX=}6z<5heZ
z{Yl7RM1k_4ABGq5Vr67C>icL9axzX(?6Rnbv9~C`;GcQe$q1k9w3Qybr0?cBX}ka8
zBnHOXislB4%e1w<y}EZm?H?S9?@;?wopMS8z_Js}O28TmQuNI0sa4aWp1F#6tWovz
zL#lhm?0!_CS81^C+KMjlSP>m@thkRq);KUdqw#v0<A;a6BhQM|{kj*2omk=A(i3;C
zapZPWd=c^{7QjIR3)RWAIIFsu5O1P9;qRcoQ}H$`p2qJ!)uUh%-k8U*Ac}=(jI;1|
zeX1idd*)U8)V^*<CeoQm`&f9RPvJxdE1iV{jxuGJ_xjeJ=>6Duq2d@@N`Fdyl?t`Z
z{$2NY_uAuK?3mZwd%iKY<Nlf(W1B-}_jlPb#VyS1>B!#wwCVN7-2E=*UG5}U{P?{a
z>fQ0~m~R>`o4+hw{7Z<xEakiY-x%g>dP8G>BeooR)67R5{v)+G{P1N!WYA_%iUiF6
z$U>kJm3QZx1baOlo{l31UJSxV3szn;4rm)8LpXvncHH!%Ah<;+dnr5rWCuOj8t}_P
zAX)e&fSv48{Bf66X>_8?V!IlynHZ&E_wG+8?<kpFaL6yE#d$*{lD#Z6lu?rnhlmvE
zVByCFjXXU$#=XaMDvPAD^>urWpB<g0obp$|%g0*{A;PnSx8WrRmy-nj9lm6TaXv7i
z4i)aK3WXtLwH+NlM}V-B_}TMkX-l0vf&%bGCqi@!g$U0qUJkR1%jtK1=XcT%fA@FN
z!$*%bA=Bhaj17!ClX?b92P881hB5F5@Z35aUodt61tZLWNLUO^9s?c&6y>&pNO{Lu
z*5}64^%av@#rHspaz+EB+(_$+m41p3bInAXc<drVk>UDR0&FXO^=czlSF=@E%m>US
zU_>4L)f#H0ZxGsn%9v1!hV<~qw^5~sH`v3#_G0i4L+C;YIh-SbBjJjPjqxp}0x#Lg
zZ*)=<P7RDOxiES#Y^gscE8vN8TPQF#VaBQ$>ibg6E|X8TwUhh6;E0&1mX9f<x`vf*
zg;ir>+7?!pJkiF3x;Q&)3-@iMbfbY->8d_Bf(k~4<C$PYg25O3VnV1xcHa?ya9Hq)
z+3FKBGBGWttTyXOvl$kp0TO(1x!`5O2a{vrgUZX+esBUEGvPEH>l-G4L7p^V=)l#P
z(!vjhALbUjk0TQ7Fw1V&a$|BAFrT{K!c8?K_)UL;AH^Mpi7i4L>Mciix|9=|B`t@;
zQ^t4&e!;0MaZs6p3U0+vXzF5R*PIxm`)dzURm_oC7gs%Tc+*BO#0tJFzRL>}x#1*V
zbqxM4{$0=Bm)(Ki>EY;bB`^l)gGPDt@Ed9KD;KAsaI!L9;cFke_of};Z7p#zC2(|R
zwwt~?Vf7>9D5OmgTZ>1$!s0+OIj*fe_^^qx1Y9LwOoS*fQzn~SkNlEq#0?XIib%VZ
z8=HUS=YHEJ@^s>)p_k)0LN{Is-3u45uw~y8NI{&l62mEj8Ac9;HXX2=h)nV5IE0g4
zI7jaej6DhV#?Q`%n!HOe`|Cqt32s`)QjEXD>g0&7ie|hGIKwzSzHuasU3+m`Y~hoG
z#gFl(H)9TPWPcldbi($S@Go_klkpR986S?6eL4UZlViErlW%s)rhJA|_~V8H@1S)Y
z<p9!DaS;}GpvYyHqrdb4T!iV+Fx~~%yuz~+#%0uxOUXYDIiIMD2x<P1pyHt%Mrre%
z&u}!b&@>~ro$=NGAO5TVAsl`U<`4eU|L!&fqX-yd)jg|y>AOcX1~1RgrPamxG(Ev;
z=r|*vT?rhAgV2s~0HHTUpns?DWw-hI`i7oOD=b<X4;zOFXj>jTlP0w+=3r$mcj8p$
zYH49%#X=+|{<E6oGES_ozf4?YJwI1<`Uu}S+80F$$J-sB#%f~j#F!i(w_stCW6W5|
zd(9Cn8b5?jcp3N`cq!L`?B$Lls6ivY1^<N};PWh$0uPs(L?1WJ6CuQ5?Bm{@#$oLA
zsm#u#_a59&Gt;arj|DEQh$Rh<|2Y+mo+vZk4h{J@w7J@%c!Pm-)D$n(Y9>~&!ryIe
zZl$C8p@sPfPnhH{E-j>m`I-e;yF2^o$+Kta)ywtNY@f<~K7H?_@4$ShY%IKC_rAt^
z6=GCzS!l&EZQOyxJw+VFO}rLISZg9qU4S?4$ztqh*Ngd;b9hZ9C|Zcs!bZR7W6--)
zxbpzz6UW0LY(Qbe#q8YKgff6#XRK)F$T#37HvCarIAx*D;$kh$R=8j5)^~TDpSCSn
zTom6jBRab{KNUVMMIQ*UudMVu*yp~Zy4<hQT$MXisup4(bV2ZlVgiL0LUE3r^j!}F
zBi@ZfJ+R`K-98*k$f7M2ya@B)MSzbscy%i}5D&~Tc++D;>85?`)xNAIrYpHreY4^h
zrRw3~p_Pk&@spoe88<aKmDcX9rH?=UI6Z#+ewwOO%;%5lxlo#iiu1{*KTrSD|MdS#
zoE(8d6yZ9;wAIzsbnn5#w7jDC+I`X3aN6G9_H%yGPP0>!Y6qhhpmK!e?7~8-%`b>P
zCY?UC4F9;hyPvjqchiASW>7o5h@Atq+1d2|`+{k@B3!gnqj{E2Mep!IR${<|slK@*
ziTd$wF}cC>Ay7kzd?cQqBlD<7=#n<-yTA|(3AfORkEOikw1`lq5f|mIkLgre@QUQm
zjz986S;2zBP^Jj~YSUI2yRW=eJq%g6R*o~Huk_LEsxb@qWvAV6R2@oUc6M3`q&C}9
z+iD)8{Jyr(59RquOZBNZq|MG%_{ML48>Q%~U>Py4SoP1nQ!A>6Wj)BK*s0*7z7%zq
zU2=k%Hch|cqbgNS=6$~N3qD)<p&)(w_19@rZF)-Wc~<q!?vZ(pMU|Zf3k`aw#?eV?
zfEVEhK_U1<2<PztURr#(6>lnQ9(6oVqrIM;a(RBg`F(`u^sDgh2(sCk!ZDjDI^jQP
zAK-|^jVN)iuFgF_;C?guIv*jZaL@tu!Q*f7+e21lv4aVLI7(RNlFW+`%&J`oD0X1c
z&jC01pnf<_#`Q`6Cwd=NpF(?8BJO{^$PQkfK3z1-T$Qvb6Feircg6+gV(iXh?!w*2
zXKK3~seY#VitvnILz=H?{8IUwqE&Y0p~R;z!O-v1KSMk88$J<2Z)gkHpgnN!xe&NE
z9l}PaoX}-1GqeocT_vc>Cb%%rqm_4hBR7a}DxczcSES%6{=kAVkvDfmyA7fU)nDXA
zTH+6Y*?wv>s)G<_hk^?bftz|jQRrQBf{(ZyvkyF^Wu-Y`_4Zs!A6BR__K22QMC-OJ
zJSdpFmG2eQ#$Vtr6k}?A=5ORLH#42tK$j)1AG`2m0{|HE=Y?Okcl?}!OvhQ@<e_k$
zzX1bIIlN^dGWP4F1$l`cMo@yIV3@M7d}d}cJ<z;+ZFNcYKkV`8gk4f<+X&B*2~3E`
z0MFo(-RiV;@XQ$3R{zFbO2=7I1I$3-1)_R<44mqiaL5W`F(`eV=KkKb$l~cy@i4#|
z^KT_BJ`2TJC?_h@DE24SMh2ZWy&bzfeXOF=XF)UgxKZ1sPHFSpJ2^i$Yh?xt9XR@M
zb89PYX#6-hXsDfpvH}JF_;7#v&V!}&gYSQs<}0J=wAB!=Fro2kHrltw_I~w$C|}^K
zP}1!mHI1u&ji)zi&+gw<PIv_rO?$gr>9kpo-9HO+;!Ebl6RlcFbb7|Z+_7|i-RC=A
z57jPF=yAL<(j1PRqc5|PNW36Qq|)|i%P36X19<9(2kC$NU;o$i=YRI6>R)bslqL77
zLJ-fNJxyPH`DJ2PKKGQa-Mg0_JbFJZt#WtRT$<Dvih_xDPd$r5O#d8ji82y?1sMzd
z2}&;(C2*Ie6+5aoXv_E)Y>Zdz806?g%Qe)7geudB6{t!(lwtIB952bZMnCBh*WZ+n
z<8D!Evfu=M6I@s+s<GBb`KsTh-+{)F)zRjMG$x`%XR!%s#{vNnk;1S*9av}@EWlxb
zM7N(1&4>mHyzrORN-WyCReduKv3r{_0DOuvAwI!NJC9vy$ebWbXl_hABT8ZL2p@HD
z$o*lAu^NAcj2jIIj2qM&_otz3gpaT3?HzK%>i_8*1wVZhayT_R;Y>V;-W*}hU3833
z%s03zm?PEM?a1i`7uTXg<*_gKy5;~Xmf?Vwyg*L)8J~))OXz<Y(@;3F3!IZF#>C4G
zseIHmJLD)Ic(SZZg&cq*6l}4>TJ<NSGVh@IK|hR7DCH*LC-vPp=R|^2<Ci;%S?mM8
z7;CFj<FcQmr%#@zdZQ)$iwcxi^a0UlmA>n)TgTmXoWFbg-}S~;zA?7rZny3`{QTRo
zN9@;4Z+Og`Y3Gk@di^o(%kH?@5cO^DHUTI7{oN4nj&sMT4$1qQm~Wb931!%_{hOFl
zuiwP{BACh7{!Xr&CPy9i{lmZik3H@h;keQp0BFuUe|G$12*c#UbRvf7eD=6^-6D3R
zeJ7%fUHIuSs*f1JdxC~F(g65wl?6;W7yn0^@f}6Vjx2MQJ$vzQ_xu_8>7%3_`E-p1
zTm*{$2vy=3;BbQp3Qv1Pe%X|Go%p=@$fgs`Ul`f)N0hubE|3-*;beCQ(34IlZp7=D
zjsrgqZR5HC<{a4RQim>k<X=KN3>nZ-$$F)M=!n&)FJ7k4zWCCYg?sX#$rkrnxDzJ>
z4HlT%@>=@Q-}!^|qaXc&?-ig!W`!I0VlcAug!C1@42a|t5D+Nrp+FVP45AEPg!h1`
zo;ceXl%0k;Z5R`TAZ{cY@PHcz8KxV?9;S>*p%oDtV154=X(&CE77QXW*<$<tNqYAD
zMOuHg;fY~egR?J){OdHJTz@XJ@ac(F_Ask!f)lR5os8>H#^{oXuh+i=%}`Sa!8qdO
z>V`?7LW!OsX&C0=aaN*^d8H|W6e2T;pv-2-G)cCQkQI&c^yET=JzK3$nOL5lMLL?S
z9dbOBm^~&&6V>TdU06tS3-i7ogB7yC?-gp$(0SVmp*T{6Nt7nQFw#noU4}4E<b^`2
zC@Yn4z#u{$R(exa4eYZOGbglB%1aAkLc~PPM>%K$NgmWSlUR<+0aKL0*ipJ>1ZWTP
zvCv8k1@r{d#oZH3k`QFTXu|-q;*GGe(vGq)>0;$(?4DzCActU}WT(v3frWy~hgHl1
z*5wPqN8lCu!|n+W6{v?dOci$89DZi*1UG#b%CDL6ne^zv<FqhabJ_cE!4sIGgh4IX
zOhwMm$vFfs`x^{GC>9^~mjNEX@XOy`ARs*UxM6p+954HXABG?*v;VN;<?yzZ|D0{K
z1KSli+`I7GT|ZW^I2<ncX4A2Te+MqRcT#dZ+Lwj)#h*7Or)>u(5YBeeN5KbmV0#YY
zv^$B(5S=$avO7Fhrx&o<JwELbY2d)8zSJS~;vdfrp0jggUJ}COv<aJ&1J|9jd&<-`
z;a~7cF#mMYj<|(B>}+fJ5tMK_m+;3e`}5Itlko>Xm~bVg(=36CcQ*JJ<tHsTU@b;&
zIgdy?TlC)xb4N*20{K6_2!UVN)lBvHJ~O?GTSSYv@f(l7BRM_cXKdi6kLY>AC*nn5
z!biOPfv^ANlHFe~r~LGZUyk745~9G$6TaC_T1IR{E-}LiIKwUChMk;^JBO29f!UqL
z-48a#L&PCxVwIHNW(phx;q}k|)xV0UzXtPv{m1|8)=DA90Zo)y*@-e^V`IzxFG5?C
zPYbitX@0)q$uc|)$I>7)Ly&iVewxnOCq4#lOnwM$pFMk?zWVY@P2!tg_{8`(H#?VB
zR#y`{JP-h}b7N#|%*TT;exZCKe}tnR=dLf(9IG?urahSq{sdSVcl*5>`slcw&NODy
z$1}E#YeG9YqA^b0ec$bsS9>B*;GQA`_ud6HGOF<}Pn0#5QWoy}V1ms3G6)?Il1vCj
z6af$e^-_-JYPS(^sodfRMwC~D-6-PYn8dS#0|681dI1s22Y4opoeb&|lQU^_V$%0y
zK^db6F%US4*4n2hUSu>PH}VgHW=u<|C0G%<eJ#F|Rke)o_wV1cvglg8XGOdUNBAH-
z14d32KzIdpfm^XA!qatXtG=$_F@!_%y=5{l-sVL0$FYtG0lcz#rkeVc4l7q#gn}Y;
zNOOxgI!f>emJ{)fCvA>E6fY!rysBPfI>!c4eir5+?2T392Ua#zE7kPy!Gkn6HzT;N
ze2n7G)~@LPCM_>><jsOc?<@1HC=$eRr_)wTcscR0k^8&s4s2VPNgIUUK=Fbg9O3-q
z_a3Fk@4uI7bMu1lIz3<C@(N^xS1SuMX{k12VJPj5du`2AX)Hmw(YlC>&{-{Bo0&{A
zV}qh;)t|~>g$n$L>iqcd&@^=*-v6gR`y_4c?79J4ktkYNURq8ceEZvJ_1>y4;pf=O
zZ9T8n*VC6@exCOCwo{{CH*bevW^L`BSGWV?{d@QPuy|^BFYXpTZ8fcIWZuDQcN7}m
z`u2CyhkE80RuZelhezR`ShX%ZwazT4Y_+)$>&&~RP|i+IP0Rf<9W>7@L^@TypakSD
z3}_$aBZ@SX8}MleOi}bf)3g8-L-1}0eYm#^`WX|i4EqS(G*AdP4z2V(63ncSMwkgb
zp)5mzcq_W|j<G(3c*ALOdQxp-(gH}15oJ7vm-CxkMHldqp`b)Z*XP}M98o%{eu-nK
zpo>>8UwG$eeSa@C4)--zd2YoQ+zhh_=mVOh4a39I*R8Csr4PUJk;*e>dZvw!jEtwH
zmDRMkwBp?q(9n_k8E62(^8D;ns)#mN@eXfGTAnw+YYYFL2k6<~tEZoT_C?whZCg-C
z|6qK@@o6a0xo3(UQWuxERz$+*b6*;YcKR3ib><)7g8bQufMCx`HsL>%lA3D>4irKt
zMZh^nRS|cJzFPev$E#V9sr~>M%OW>+7NJ;eqkL8$#SykFbYt!kV;%Tsp&Vt6c`G|x
zU1qsiAgXlO^+un-xac{A@_<KUrvZJdEb5bn9gqH$_NoBtqoSUfH_;YYRj*H%8Ko$B
z)5b$t1FktD6y+8tO3<&WT{bzPB<2~)&UHkZC=pqCZ-s%6a75w6NhYE_)ibcf?jDXp
zS6t>%j4y)U^{4`oXAvAjjm%3(N5NEn+KkGl2cY>TJoE}q2*;bW%5jG5rn68Vo?mGJ
zCpa=cWCY4hc{xoP!q5(Z$@3qzGmbJuIYxPzKLH#5C`XFoi!?pwQl2P0q75m1*QmgC
zc|u|4`}Qc1a7{g1aVCF{8^D=ig;pZA%1(I|q$m>c$B<^!SGYkB;R=ch&(t;H6Q+}f
z!i6Osr=vW`Gj<w=w<^Og{4n6Y*Yx+7b`C!D*2A2=GE+$_%Qf%Xq5pyZ;N%u|dT`7g
zV?W~-*}$JvDheljgA3fbuls}@SQ^)<H|ibOI&DGq9!DY*1{%0_eT<Btm`3TsSaYG&
z1;1c+T$R!Nlk)Dry|fZ&93{{w3ZJ|fNpYZQ`e5>*-=i(j?pe$+tA4a59*w>UT*N(p
z&9t$#r)MYaa8xt>x$1GDGLe4!yYHv(ytkYthS|kRYZo}Ge-;<9^LBiEQf~d?8E#Wk
z?dbUA()UNQi-yI$C;{LD`!(;~+kBa})}N;9^Ok5zeXRI|#}8S=v6N0V=1)(}r0@Uc
zZ)&VsPR&z}ITXLBKB;d2zAp)1>VH)B*tt3AF>6QtCkp#P^%1mp6j9Uz3$J%}SRk~o
zz7B<l_$;;Wy`7Ep;?+xy$p^+Q`~kc9E0x)F@BRbd2R6^~l$DAXjIfKDQvl%m;5(^%
z>JS2hZryLGZmcv>ok53ZUNCTOr5B1Kz1-iFg(4`a8QVA_5akl`A(UT`G)hiLm3HYq
zKp!ZGSfGO9i!!rVgtUEZuG<haA22{C>`sHXgP!1xeEg#F@J?5`b3y}m146^(>z%Qb
zMU9oQ9Yq!CAhUs|<f!i(jl<mMT5q8IJ#}BgG36ZRh<pkYyPehKgdfs^cchJzSJ}zP
z4l5Kr^cT=IaH5=J(F?oZQ2Ki395Sj&PUB&xmHJP?M;`|P4zlA*W$e3A`w&mdal}GG
zdj4`f{q*NwB#vYs8bgU)Nw<R|YO|NAA$n}H)yn<1!FAun?m3PshDPDhIbxl=Fc+%a
zt%`C>eJ1>=>hntYW+C5c>m*H$4W(K&j{YRM3-N^P+`1Ot+NvvfDs}0?#dR7McvGr}
zA?0<dx`&6P9>6iO!lR>B+TPw3UQU%CM-!i>o6O_*QTp!h_~-5K7!$*;{-xMeD8?W;
z#&>>Wdc&7Dwl~8>zSPf~eql;}cg@}Y^^N#<qV@WB%)8Q7`+B>{c!H#|OEGUv|83@-
zX@|)qWAlyElp&;Hi_dRl-i6y=+R&H0B`GVu{=t9okK_ljXy7D}JCSZd-POx21a$Jq
z3B`+06iS1V>|rAv5|AH(BGK3E#aaY;_Q%z3b|4jpgN(9M4)&ICAXZN=!(d4CUCP3D
zc)%d&Mfh;(8iLp{m&2c=jU|W92?l(!$A8D*j!vxum;S=u3BZ5O%Rjav3?i(@lx%WV
z5!CNaY8{(iVN+U!<0t&>+cDuQ7yccmm~MLZESbvm!ffO32<Oyu@RFdFPI_C}CkIyf
zF~MZjCo6y1>dkR8UftX}c4w-d(F4W3_ufx`@9+J+^n)M#Z3}&TF|Imk9~G#KJTU6d
z69x^I<VP+rm7%mSIEKqKBR7~SE=YyW%(MWq2T?HrXBxoeXTgKuAUp#k>A^T%dje2a
zuf+-unDizqk7-c)<H!^QMx=p)g-b@CK7E>Az1Z+1igFSEuhSUQAU!|ez{ZTW-ober
ze*d_RXzxP39^=gcIddNYP%x&Y4dY)rbBl0Qx9*tq435OkFV2K!g3K-{(jpvj;&=mA
zH##XXe4!v@&^L36k_rY#@KFa>7m6q5r7}O4=9U-r76UddhMF{#rWHIINUhM)MD!LM
zD?KK*Oh#b}uf<gOvU-()a0t?v@NqmN_0cZ|XnMAqrmM4QOic42LIp8)Fh8t#WTJ||
zg2^7MU*Kfu`|xLlIt785Aq;&UlOhZZE-+k(!GuPKV6<VVRColDp6D}CgkeKDb|IQm
zI;I^xgCI$0jSE)ugxmO+qw+$)1x{g<z#nx4KHRx09oq>@9F>8I41FE>v+{;8_NV@-
zBQfe!=j6n6dbs*9Ez}m%*f1-OqplUwb!J0nn@I@w8~p?Y$;Dx)^IZ4PMkgc&KZP$&
zYKtX=)Ab9w4O;6&lW~*?sI&!u{6b+BzD0@QmyEwUACVrvE-!^Cu<5^BhMb&?%c~E=
z;(P*j91USletk;d#yeCZ&TB?)5mJUcowq;u33xi6;g1!1DjtE6@Gg1yTR(&izg)Il
zuDkfKA)uasRUX+7XRM^%2`@WVpmFYoJG&@QSAiC<5}KF&yB_wzblmj9Mm^Zi>2|*z
zDm-wEx0R$gc0z=6_U=tHj2vleC5(-5q>oJf@{fq%I&(cv#MNYfJqq6$PkW5Nb2y_$
zPw++A-Er_B06)=a@a~v9X;b=wIfqliw7`Obf}Qw{&<^NyvQ-bJ;K60FHCr4Y$1V}_
zyLgwh!?^@12H)(X2wmT7iv;|Dk5al&%;!gLy?zewmn;I9mE8Ohu9v25B1pE%;xyUu
z_3Kl7{b&D={~@|i8R!WC8p4g28(ZnHeysjH7AY{f=WZHyEU?O!9XCUosGgs-e0R~o
z&X$Gw-c5kuM}0lxccofMHO1k6tGU{|1ri7w5kPxvQr_&!DJyQT<Y1v4{k{lhL}Tyt
z#E=zllbWy&2%mkoEZR`KHkZz`npt57;+_^?+^u+H!vhu+;KpiA?i$$CB$V^}5fmV3
zLc!P8<atbE#njY{Ctx6cW@aKS&dplkH;!<Oaav`e?r4M8{3!n`P0$(Nsn@bA8vaUy
zoN?y5CXRq*;tJhR&nWH~pS@UvvTDp!F-?O^#t|@c#3Diw?z{>m)0rmcH+pOQ*PFY=
z5KbUSV0H1##Hjd{VU5q^18l)3QlG#ftKAtx&Bop+9q`v|3I8W54g!(^3p?03hZ2S(
zUZC3>*1u9wN&vo<$v^jqoof=$@k)ckTyTGt&Q)&ef*}pzI!F1z7qbeOOc6Rgy#K(;
zFND#tSnI~hJ}%wo&Z)7Hv2=9Q5D&%CO2fiG0v-g@@Jh)$T%a*<$4(z#I4{^Z78B(F
zm;b;2=s|k_y@y`)ye}F?**Pep7#Z%j5{M&rCR9FAksbubtncklr@|}r0BsE+{89eU
zQXFx3n#-p4A^7$V#QQf-OoLa~s&~1=m(gAj$gi!gq(_2*<CG8<Q5P2Qou8zm#z9(t
z{xt1utUGK~w8jz8TU(o{dDyV9vtJogcN{5&fQhnqguF_RyQfxG*3!NE554<qYC2Xk
z)7FNE$Gv;)RD36^%29^aD$~O4pcRl?y9aU?C88J=EwbonWQ@B{f={BopvXcPNO=(e
zvqJmj(<f<rYs*U6(^GZ^wv8tQw%jwdwY_O2AIgRza34~=k^h)zbxyFc^MZ1zN}TWG
z4?alW{odbHnMRZ^;2=aCjP`X!`#SSVYWgI0J~0PF2+BPn6^=*}y`pGoY5oEo?e1)*
zFF*ez{p8R8IJN4#zGJyst*K3kmmix*BkUZK+w^oLEi5dF=H>+FK=eW4-Pxf89ReqU
z_H)6?Jv#8Rj4{)bqiMD>nda0VjnRhHwxC4}JIzp9Az<FyYk2VhO4l2;X#~zbUQ+lS
z<A^BbfgpB9Z2=)Y^9Jq)<fuZ9lO(KH%&V;QlL%I!Z`uXzm*XU<6HbJnuZi6Rp$uae
zDf0vFmtv<BeGqs<i9<gYcJ+;_AMQD7svU8+3-bvSN)(hi1^)2nh1W*N4xi5MxN@W*
zZG!fN&=cMscyMR8mlb8A73zSu@A6W7=$F2QxiIY-MFDk3T-u5ixr$qqHwdQbLz&0I
z|N9Of;FlY7Bk*EDvHn=ZM_a#Cn?#7r+?JgwEGl4!8{>me5DE**q_N0`_yXMN<Tquv
zfO!z5UCi4N0z=#M8PF=}5Y{|_@(%^5=v(=E$B$^(IAJ_c82Ski#G5pz2g*!4qc0>3
z2AV?P&3H&%7-qS79-|M-V3e2pF47aONVh1yFeJs?tjtY|K>^ErkUV{#xBP*JU0>7>
zM^;ic^S+`1%0n8oYv&_am@~34k}}gbVVLid4`C=r_!M8@>B4|1J|0`|__@G-=a<VI
zWwObZ5-&#XVxaM;pHO@a+t>4C;gmjkZnlyprnqBrT;u(e#-tg=8A?8?x30cjZBePj
zF{^o@&XCe%@q>3WtG@terbmq>mN5uG1n9H}D<e%SD4k>1PsmM}D+`a{RO372BnvQD
zAjATXi}RLpRXr+Q(x$GU!=Mk9HMd18WR$LVP^v8rXe{GSRqozgT3k$vEUvB2vV<h9
zZ)~S8zkZoE#Q!i}-fFB}TdJnN{WrdyK6-yGjSaFxle$BOq;@Pi7}cB^rH)!%x>V}T
z_I27lXr{gTk$2xT)n+-8aJH)P|MtT7u^#O2ny-o@;Z9Qj&3UTtZm3^4Prv)Oe>?r)
z2R}%wEBDjvTrKq}{d1KE8i6#=&C8v*tc=B5Tf1p<XWxA<9UgUo+=U}s!5M8z04twU
zADz9k@gnW0@88`OA9`4KTR~1iOil^lelCu5eD48whb?%43P&F=EG~P8HaJDrF(!Vd
zua5;k$#sbm3Y=J;ral^_(y{vY7Q2ts$FQRbg%V00=$_qdDD02{LYuTRPD%h4`Yp?K
zpbMpaX?$O)|AYUHML3)azygK17f^7k-zhuZP}s188bu(y?U3+7|3qKRF2y*RX<hZb
zVf+jY_PH+D-SwUCeA`Dx!gK8%h=18{sC<3tMDW(xsnr%vMVs^|@Lg4nt+Uk`FE9az
z9ODcw*v$x!(`>e_^g_|d_yLY+cbo(OJSdK;;-jk?D}CP^JhVap^Z59nkBDS32>ms5
z4ZQ3)YpCt*ZocwjF6J7HkMQj4+xw~Bx=4r33)2M0eTriB$2f)@6y<2_LIZx}s_eLj
zFX!0f>FMz_C0>C=aom}TGE<cSX(e_$vq0+UvlrqOX3_`mKT1mrHPb(R4*Uwd!KrX$
z`4#odd_{Qh&brYF)2xq)7oJ#9$8qZd&TXxAbSNJGM39P<^<dne<u?|;{jK)hv5V@9
zU3kLJf%(nunBMT%dcQkw9=Z?PyLn)G?IpjvCi3o@H{#!k*6ZIf?@F6Wc&kx)6o+Bv
zZiu(0{}%J^w2h9}-rqD$3?U6$e10SIF5LdoMs*>Qf5U|0>+k>LkTDo}c9+RTWI)%8
zi9Y@wI~_(A{ZkwehBKzet0#^F+RiInvjw9Mck4i?IQI9C!k4JM20wTjTO&;a=5+Em
zCIaZ~p9o{4FiM*_c_|O`R=$ZKQqRm8;c&w>Qs%S7u_<-}${x-hq9Pz=upa?sWDjSh
zsPck2rwokx_H-?hvWfKauaoJfL-Q&;8h)6Pb}^()*>Ev>!#Nxb18yaKK_fSR=QxoT
z>_u_#e}pA+1nRgK_(_9r{0Mi&;B|3s#+?<8tjObH?uI6>tS)9p%!5borN8@k|Go6_
zcfYGiY0WDTAYAH#v|!N4kn|9s;0Uu>7-bxQ(~P)7(qF?wiGf$)VYEpJhKk856C)3<
z>YQP`St(?kDn1ivS=@0mNizchqoy)2Si{^Ruwt@_vIB;gPMnM6*`~{pE{ZN70;dM%
zuhD>Q%15VfoCABo5q=5-2!SUi9dzij2}W=Q9eL#qlO(|mPGEeA<Z`GCd=u7mL#j$(
zNE6DD2~9S|ND`Ncp5c%m11u|#Sas)!)P)9a@C&1BL7^BJuMQLLTvrM<AjV~3c_~%J
zlunDu=2CG40-ji)fCMas1@ByBf1^4<h{o#qi?$faGnl=Ye85<<;*6bo*XLrSRsI1n
zMcf_65j!ZpM#QYM(iR~Fc!CjtS$11cIlz<4FIuESjyvVT-Mh&|2ZV!hN4N(q!I&{I
z;#ifK_}&^m+CZ7KG123`mJsrsEAPOY=^6S$fIxmKI|zj4VN7B2fyHGL-T;L3sPky#
z5r>tasy(}zE}Xu~MO_Ql8!;YukB^S0`)d!=()^O_6KRmuf`rvCVU?WU@u4?AcF@aT
zv@6-6i%x_R)Rck^>uYq_NtYDx!~nlS<S(b~Z@rv~JZ$HkLq(R6kKJ?F-n6lzs6vqv
zv7B1?-R8DH3dC2$92LjT+b{k2-Qynlkb8!~HvIjoR}NT=J`J<1fMl7ejPVzPP=<>z
z`o>OtrQ>hDa3-KFx!y5)IUa99(?2?v+MV-}y`-%`?mtO_9eDUUFG4B4BDf9Yv?E;3
ziul>hCVR$b4iNv{aC`@{I&X*9Pn4DXfZ0J^hDhTM%=r(vqGW`&OT?v9lbwHjIzLXE
zK8Ezd39oFmp^bR~ipwL195?cD0K3tn2EAjqkAHb$yp^l`%~#kLQ^JSgU23P8D2w7M
zV5A$q*;N+O^@k1V@e`%T=r;oU)%kHiy)u5YhvOIm(g3%>tZ#x7B4Eg2{N_7ybAP0-
zAXzfhH}RYSG5OE~c@iX>LJJ*Q%MK0`z{##!GFdx>-Ey@28~Od=zx?$bb@-qC_+L1C
z#(-FGpoyMfWJN1VgYBJt@iwgFm7n@sgqH}3=H@DCt~#A&5k70YJUeNmufKYdo;`i$
ze*E6ORWJ16NH=yj!1p*K`Q7M2u)|o&q>`ic5CBjc9utKp#uJ362zADW`%-0cG);{v
zj@s|F#v&9BhsWZ#8V#qx(K&)p=!;dU<D-mGLl&Sy1BBh%KTOws?06pcSc(9PmB*J?
z(T_0pXb|y=)SAjQ&+#mp2(tqVzN)2h4M8M=Xdd__cHwYIJwiixgmWwY@(O0fW7RM4
zc!7y|j~fdy*mdpj3db&SCgO~%?5IGI#F!3FPESJNrO`8;YOHOuh=$!4;7@T7cCj*+
z<6-*yZ!An;*G}wak3}Z*4VNk)!fAG0wT_R(x3F_3cFwUo=t9V2m3gI772hy!#UaNc
zh47Upc{j>Iz2PH^xV)UZK2Q*_qs*&fg&UuIpgiG4IT8>ah<vKGTDt$>zJ*UM)y2Vn
z-8iYu&sk7`0tG=u<LD&q9@Nb*a2fgxD~88Mld$Xl4dEdIRs>O2s3~u7v9q<F*4JN&
zPq|3biu2xsduerTHIzA&Q8;d~D^WOuZfgi*5b3GDp%d;-;;uLpMXV%7*lvNe@`4xQ
zzEx#r-tOeI<s<t>#wNulPO5%F0e}!0!6^%$P^?W&jeCa<yC$cmriDxRS(Ky-uQXmf
z`6?ZVr{1sc3${}$zPPgq{?hj~HBmgn3x)8L(G0<r7s%0?<mN?K!XvOEz`eJ2FV$G=
zGcl9Kre@R3+<aPGsHOQ@)im<z#f$W_Pd*h-zbly4R)ljD6YNAn(6l0)4RbW6kd1Jc
zdulk!y4l!I&%XN1Jn4f650yXk&@^sZY6I&pUue#A<a_dlRrc+j0GvQ$zpb>Zc?10}
zG|wOH0mb+H!lD&z+yhdZUrg_R_)%I|TuZZpm80k2<Gl)7d?n>&6+L%&l*L}`D&S74
zzS~P5Tg1_p&<pgyeR+q6>azszpz`FH$AzV}G&4J&hR3JTwprQ15#|W_IP#MnoX|Gy
z8(L*I0LO!3$n#X?r*7$+xtFDe&|GjL)I#Y(8)r8i?HZVw`_<J(IGXaPactV7|1e*z
zG`JTHK|0E9`YGms&>KQZ(zD=P;b=z)KT&?MJB<5kVjg(w)$=I25fCyDU<V?suNU8m
zmN=Rbp&-I(?kKXLT(l3_&|cUrM_V{PK6RTe_x(Yi%-G3?K0ijVXa}~6E`SNatiY3#
z+CFmtm5q5P_thz+=SLi&Hm>@n{~wdR9MeZ1e|0H%g<Iw*q=mwdI%fbH=tuCYFruwc
z+$a~(I_)1g*zw2CFCIBLZ+1*DcNL^^QC}xqut0$NXsfN!525&?&vp>~piHz<ghbq~
z*sY~G1Z5X4JeCNK!GYKXNm$%m2aI9zD~d1XeJmt(S|msP>5sV`a~|gBa@1Ro@*#cC
zgH-qIw8HO7xGVWu$)$$^W<i$Qc<jcDZ_+0Xj0LiS16m>;Wn+9HKhh=55|%pPVg7_7
z2RVhucX@MMB{Tz#U_cdfx~?LRdPbR(=}94IGvviM9Q~!r;c}8f<cR?WOo5~DVQ)1h
zjM9y|hkuJU(Q67_c5}ffj>>Pvkml};Gn5P5TDhQdQ%00t2+A2}=4-R+Cl}M?)QsEw
z_U^tHG@}eb`7kc}okoGHHmHQvcd?5c1rOtba4LK-hoyf3-qL1JzA_dbHICAe`p0Y0
z0<;D!^kwX}WTCNlAJcv`W|#U0epR^Z9o`&#TCwT=i+rFTg%bj3Pb><d4We{dTAWYs
zJ${%TK71%#45b&Zw$hVlFT7xb{(NxYHhu7LCH>BCeJ_1*U;Q!n9H}pXcW0M3N+|Zl
zkk{zsbUIVB+1@`+2S;u3Qtfo2F%DQbnz2@$G(8-O7vq>k7VF$xu~_#sZNB_6J^%XW
zX}Iq?tu8I3#}6K)@BhYcr1!t|oisH&XMU1}K=3QbG|oh0RzRu`eYLTjn(F(J51gEY
zjEdtH`_&%dF*zz#K#OHNOE(%P&rT21*5<49@&&tnkKLDAp{{<K1;uO2%jw_!gCC{U
zl@%{qf4RP{K6EPm2mj%Jn5r7L81GPG(08!lfq5%@6UPh3edTH^s!No9N9s3Qr)t-#
zU+=^cjP&p9$n(x<J@nMHSCnq$?nxB(Rv6Q-(KbcCq<Jb>iDw>Eb;_az=oW<;G|F+*
z++oRL9FFN^rx$%4@{w{4$L{`qdinB|#+?KAdHvoMdtycJZ~yiW(zieQ$cnq|orCmZ
zV>j&`o_ZG(JI@%C*!9XTLl#uAxMN2A3vTqC^}{2L!SFqC$0v1x?1#lD$fRhl-~iql
zMe$;-k{0J?QgudcY>b5)v|pNa9Je_(g4{`ci{kF@9i->apQq=~UWzZ=*En!tK4D^7
zyr}Ah9oT1=w`sS2lJ*)c^X#;bSd_qG4$|r~j%QgQCtNWW!ynN{alhz@`vZ7oly7RQ
zv?F~Fif3*%kJ9H~Jx_J<7?U#^+XX+z6Qa1A&|HB2SzVgwDK9D!4uJ7O^l>iyuzTIS
zH0@q?6nGr<jvR{<GWx{5Hya0P59e;L81O)!#T$#?{*Jz`*o(5H^n(t}ZzoK=G1&RK
z>*qI5#=qe#?%DL(OP<9<zFpHB787aak8JLSv3Z-D3}t(}q0D;T=-=Ow{#(qu(+;Dg
z{b0UnnixVFw)p%;<}O};Wuy8~7yX9m9(724DMS%41-MuOU>FB^;-T+w?-*G_P!#Tt
z&!;_!DO+(Y>9}R<U-6uFjzAJ5<&Me_J9hC7H~xu|VeUlgHF5*;D2spm?YHtLc*oD>
z%K7QvM-ycj3#bu0%0g&|cRcyvPh%r3!$Bz0Bfi}A5Cor+28l*mIe)!E;6@UiP>vBq
zLwW>@Sm7yG9P#(t$wqQ|+mxU(cI)|Ys`z)pl2^9l=kNt4*>n5|5-?)i)D^KLgR!7p
z2?TunYv5Hve8cQB5HP_(0p{~=H9#&dEu|m)_HS#_|1eEyVv4|rbb!<hv@-}G=%md!
z!pzCf%rY>Mr-SjvU2ij)IO41URD&PNEOrKUjyF*JIP;x}dkjzrEw99kpPjdr20N5u
z`!+b@*aKFe!l)NT2A5<%*JSyb1|_!cmOKM>CRW4stC&dR?FK*U(jC6=#S@c<h{Hn|
z49ps)D5r!#d%9l0vHo1wdSI|H2x&;giZCZc*eKB;1L`OPYD|0>%*zTKPj-}#2VBw-
z<HNW>eX!~o8n9wfBc&%>8l<_eY<#Ag7S~oakk5*d;V2JQqlgY*el)T0BtrjIe2M8~
zHLdGXb-@aH>WGOU%1c(2py0W@5c4F4z=|&o?xRy;;wu%O+dQfXCFfdmF+0o?Tisa+
z8A4ZOrM#Gc!MMQ?u#*d)q7Vfbm^LOc9CL^O2RByRat=De8JKrgs-h?b{y{O=;N#@9
zX#pE_<cW#mA(-IU94pksEEL5Vh80l=aOkVNT1w0r@vP`{8xbtRKX9`9P9DY;%rD4i
zVD?HwF^$kPyS@-0K3I8>mX}u2)VL<sFeqYF_=`3RttfxNP<kb3pwoHyAG*Qc+4&cH
z_{1+R9w%JRFvUGgPEV#2t0btmvkR*o-RJ46gtf0<1lB*{{f#@8F!Zkuq90;*!s@T!
z*^YSP*ZGVDW1I-sUMq>*oo30Hcf@v{#q?8OhRLzW)$aWA^!DYErShcyA_7ju!!`%c
zWsGDA5pX)JVTrcyC<;V7z%N3^Z|{pA(ltI5K41v{Zn&^GK2{uapePgg2<I@sq?g?M
z4y(gDM)r*F@Wb8TIZih`^el&-bV^=Pw)~MjO55s`g9qazA}s0hFVYV@Wz&I3cnEw@
zhY>D2cVNR3buEYZv`3^J?xr8T{p&Kw28<<~9(rOTcgm4XPOj@BONP!LhSk87P`R#2
zyAw%1wwIDcxcH4!B4ODKYxo)_hr$pGe9D%q0<V&n({#DiFUc!X(I?@YjALW5gUc>A
z-;;_vb^_z>)DUPyo=~zqWd879{2F(B{rC?%*Z}NezZL@8;v3*&_UaAsG#pPOY#|sM
z(<ELK!-^)y^K+F{)8w}Q`aEr|ze-zM+Zxvf(~|g}8BHP)#0`#eKiIHWM_L%kSfHVt
zNj^Jlm{1}#99KA25Kn6yV<!j48=}<V=p66}PeRx@66Pv#x%(L_uazbv7Dv{BUh@Iw
zu^1y*t$O7<icoZM`MXzvd!m}WK*a(JrHfF2acW{r@YiP2z2!OI>4V^P|KKPcHjgb(
z<A@l>Rm#DRDK67yocG;08n0T%$HLdK;6M;SSw_?J#JDHqBg$4owZh6`dbkdO1IjRk
zi#yB^q;gCWg3wGS8vCe^bB*`xjzJm5#oY+tG~lH{(HM6Q)aItsye8Yz!XGqC`RYd;
z547)vOMAP!UR8_WtX5+u%y>FNpmB1R#x&taS$SW43JQe*r3<}Y-VXQ<qFvv!)kyWj
zMmiK-fH%fp2tCFO;a!=c2tX)ig_39mK?uqTgbhcHM%vlkQQ2<N{QP2CU0W6$)x@qk
z_+C~KBaB7R!|{F`orxlA93hwReWE&I_3b4*o?v4Hf!^3{h~f)HBrt>bIq_Pn%Zq7B
zFrb{NEAI=@8TT!5cirIadAem~y6A<K#p6>mD#t)-il*RsC=Ydj@RFVD@Hl2a1SdPO
zIKGgbNF14kG6&)Qg?JB9g4+?ij^l_Hfss1CJU?}LSh?NU-?t!iV{<({fB7uEc=;lY
z3P%qfJ}|96ef}(c_2h|_w-46V(v09cJZhw^?X7fptg?tF?3z3lAQT-$sWX@+ro?kD
zFQ>;JyqDg8{GMpJVt$I9NdM;l`~OK_D{Ta==Bq@5GqbblgKvF1&Cbo4Kjp|g@L3n#
z931SWwY8P>gWvjFK0*q5geOH%wJE%R_0<=u!;$pEAO4=(-XH(TAGuCAwh?9MmruS*
zjl;vVAUyrSkN!@2{NcCV_PNib!BI)7cjiV@YFqH+>_p<YY=nRCi_kH1mXp&{D^w2-
z4%GHrqT@mD)<lSf!j~Ozk3alSw57JI`PRtrXgU?2d?fll^T`XMN%5NO<Q!B7kK!2~
z)<+$RcIRemqJv4nT#g03cH5>xp_73@l+rg=xSyV$cs{@?@tSak5(}P^*1*mpc7N^c
z?x(}UBh9(!jYK>8yRlL_?#G4yrq5!59Bm!JwNG`>!|4g|*cP6uY#b5Gorts*?kHhr
zCT)T;oTz-P#wQ<mcIKWa<l&7`JaJDNJFG+i!m04gF_FR<O4>d8HuWXUi5R~IRZi#x
zp88Vlg*M<FU7{5$X7q=0%JVXn30jQ>e)LWB0it)o%TZI)YHPsFf;IB<JV)(fbK_;Y
z(A@L=$B%qJpcmE&%|eIfP)bQiL2(b_$%`>Kj`8AX^-J}mD!=A9C@-KF?_AP6ivF89
zND)BOMmVOGxd_S|%0it|4wOT&TZbJv%%4l28;S|fqlE)@nS_9q(qaUY;(7t3#u2(e
zAJ@nnO>v9|1tm1}K$@<3rR8!l4-$SOCG}rQ13|9wtg;#R5Cv(FxDvx3+yaMpMas?L
zoC<ZIddcyK3Vh668H1RQI=yTt2W3VP4G+%v2<()b{9>M4$}3z5r_8+rci@L3&DbFZ
zT+I7`DHL*)D;Cs3C&qyD#Pq`8{z~6noCaJN%4G$gX{nn=l<oC$fLCY_TI?F?2l}Lb
z<Jdlg`GmP*&a5O*d=G0dMhOc^(!N;q!H(g|?6k{;p#PwLWX1QOXpk|J#VOp)P5;LZ
z8I;;AQeYtl?E)nZI8+hTKXBi7+!u;6sM$L9oSrgRQK^0eWf!;&UY^oKxu7d#4xaa_
z4(XqfPk?70EAHtR)Yo(GCPxmQG}Vp|_SL4_o?ml(?ql`)4<Foj+?N~M>DA`8?-spO
zTY_JBq(1%oAAgV@-(N`!)rmAY#sbt^m-)CA#}BeQmBl#q=2_a>J5KwJQy+WCQKAF=
zH$DlVqP7U!j8BxC-KgUuf;Fjs*?68dUpz^JYGd5{#g5v?j~`p1p=M+G3P%%DX8L#6
zvk<txv6*(oYfMd7)8x#Y#ypPDRhw3GLfODkspL-|2>!?~6dAF5ZXbSSpIvST>Dkv`
zn4kZ5{?6~GAN~ICsh_WUaR7Yb{reBBpy%$%Q}MX$Fyi<;>Jgq*s8g3Ex;O2|j=%aL
zNBlL7BcI-+veQR-N0|Ekb2NqO=ZOzrH_Z-Z^}`kQ+3eh6Hx_LgTBXiWu+m?--&Wlq
z<DvekNBAh@H_$LVGPDVP*j?A+{z>8>hjHH#r!&M|tb?-0csC*5g#MqMz3^_k`wf-7
zovxJ@ixl9`Id&UmS7kbO=rVq=V;O}e?SL_o1sXKLqF`rH82vg6uBJ7nRA)67YCN3R
z+<@;<F6$EFiNk^g#}S_lJBMQVXFvTU{o6nK)Aa1ci!?nxmcI4D`|19JM`>VWA|0Qp
zKT|vQvERbem3S*SKKgr?YX})tpT<f0-ZQ6(B6e+Q-s2thK^sGcM!h+2@e|YwV+1h4
zJHuCSFYTG?m_=|nPmWTa*eOO@s(LR3+}z&xSmT8O!oj8HNu0Pa3y-S$VSzg*SL}*6
z+}PSRpNedx)u?+M0tO!XO7~y<yC35%ciX%DUPtkzcPMeX-@R{)!<6?Ocm23?346(>
zYwmvcDrr0h^_aKD?Qy#6-!boSk8-`uL^)`;cSF20k2~hwX@@Zhg!LPzi6Nw6i_c$Z
z{wfq+<Yjslw0?<(?wU}1{TKg4{`x0RLY8+n{sbVU1|iZ&%*N$L3c~1sLr4;tMR2|`
z1ft-PD1b4z*sh0Q78>R>BN$(Nla64wi!4O2iNb_HHo_6GlMAr%c3ipHANR0C+BvSi
zBMw%6Fbb2cC7lRag6Y@!g@C-^g8;=3Cj?}C23)dW_-5?RFh+0((swXuBxOUC>=1%g
zzO$pl=O8`d9hbD#(J(Lvn-K2y%JDizAEe>9VGDn(y#ej7M?b)8#wxpYfAyyS<_!$$
z%-o>VfkHP7LS+)*j#iy&+49bjK<|I>L0VW`^eR6FLc)?wVN{6+{diS`Xb9$!H)#<D
z#3JN}VPn?}*z}QZiitlH5*RNAW|#mNG6q01tP0QB#0(5{{&e_E0$5>$&;>gaOja_(
zJOmTW>Jv?>l?6vSe*N{6^z!Ap2HCdKC0FGOJ`gl@YT}nOlo>-Q-et(ggMspbs@yT<
z2qhlwt}_`X17MS21i+6r>2zq%a;AlOiUu<IaA1rh#3w&``o)~E`)hQ3R7}IDnP?ah
zPk=Ru69LBcN_c0*FSuppACp8Mb0Dq`{0@%|r@57tw79yOCN#<ODsT(*#N4Y6P&isZ
zYUW1_j8}o3q$|}6lRQ>vaaV|sI@ILMyUNa0?@FJCGWLrJ7}nrBT}AOVorX17kBUKO
z(mAC!lQdQdq0}prm<z!!0@B0<=AYeG7h;-ET{aY4FgHx@5kk;BVL(QQ<7h;1PQEU?
zCY<0H;~igu6=j$wg7Rld1kO3bpS)qRm}o$6FpW&ICX_alaLUibhsTox>QAr{2E5ZI
z6w!88*^@0aY37p&5Ogg2$k0eyURq0Q%lA@sW;P9pQ8t;VAc#a+1P~CDT+kNuf<JH9
z2lU}@2XbDvmLF)Ogzb1>2pHi@|HMO%PMf^tXS&V)PTJ8j9E8E2FJ2uJY2##C&T$;y
zuKrZqoDRm{*&<_qaLlLN!!>>hSc)342`n3V=uLmePO473&)%5)ipUr;Dz2{=;<dD$
zyb2RLw7?I+_4Ze|86VU!egu#k_3R`B3Voto!^`1GBUhp9;3zOc5KMsdIxfMR!x@P1
zx5;_>-{B%1V#hcae%-h{0Tp<Qydnng2p69vEAlQ2MGE}gZ2e>77akBkaTPDH<cJ*`
zZWwCHv?haZ{6q}XnA6~^nA}G2FDctILh$Rp7=Ps*Y3renRh`&hPuC)=!thzr4ttoQ
zSnZAH5O(7b3E?GIP=u1N0`g_!ppiy#0ymU9#tgy}I?^L@1S2i*L)=_W(vgi2j<2}k
zpVKBSyLAm=96kqex+GmHk93KrKOl%4x`z6;ve2od^v`~cim(6t|Ne($Y?yj8Wfu?S
zMi_=bwRyrlRqBsf-Kf5pqf1!HIyo_tCPp*?JZYsPP41XvqFAz^RbvH;G7Tu+&4VzC
zoebP*0B?lgmt8{)#T)}uoz}#;GU44_lNvL}l?O+ofS{Ap3$Mtmd$qA9s2aP>^C&W{
zoPNmySoaA*VSF?Gp0v0J;KFcl$-LnoWVba!ZG^?_Sl~Fk0k6uP)5LvsiQ|Zw@Lqd&
z4?-UBF|0DjT^;O{YKJhu_dRGL&DbM2ILja56N<3ui3z7S$G9v$VxW&BsxO_FlG9|p
zUoZ)$nzbVs0DJ@>7L3CC=qbBDE|f=GWp4@>z{JW@DvdP8R0kZ*IIa4aVPTD+g>TqD
zJWO9b`PvWR;UN$zq8y2x>>S^8E1u%YN)v=-^RpF=pK@m`7A;)&4W(n{*;2U>&~Q(I
z+qyE39TW&m8H?p+YJzWKSIMM>+Te@2=Kd#kcCfI89W@-uw03VLafy5p{-H!-H#|H(
zJ7BJbi!r4&F8&7L3U`UY`m!qB0$19NaK&nO6e|dh`&GyAHr|aRo_bQe3<4VXVQ7ZM
zPwXNa9k@>YH=KeHJhd<H7k$gs<)$!JKnR9Vtf-#!W+Kb6=Fgrzvrq`(K0Du<;+3Z=
zHSzo33Bi$I8Xp}}eQQo2`a(#$@$!Y*$eDN|6s)7Fvx)Td>67&NS6`%)X43*`1o|9V
zI3?V%lZXEA;r#~|%pM%n)B5H{+Su7td7G{S>V#cr@NUcnhJ_a&ai_M&&U=nNN69%~
zn@<Z1%fkIwdZs+LHeMzb$Mr`$L3oDXjpLT!0V%srkB}ej>Am+JTS#aHjTHr#X?J%g
z9oFl%1K$T9d?<+e()#+k%F^#$cL)PN|KjIXlCcByqmRCqKK|bK($dnZVcb-C{^)=I
zx4sh)+|s*p6zaXT2kHLWy)-__-DIuw<yT*NXBqgTJuv@5Vaehq;JPRNdakyR21dry
z^2)L=(`OgT<}SxO*41WEj&MD`+KeW#)HJKX5n>^{W;e%skKgxETqsx1)ZQ2;Z#4M8
zcfxmD>8|Ibd6JHs$0~D0Wtuh(7DXh-AtFHL8^ssLO5KXC*nK!7T8B61Xe$J>?C7MQ
zVy*yQ=>SmpuxO4W`k>)ZhQKF_Zl@;GtlB4nRQi;1iN25FVkN%%Guk*pOXh2C572}1
zq%Dqd+|$IUg_7(RrQOpnaOplfKY;^!^L$oyL40_59(a2cI~YcB<R)J5Mqh@T!uuo)
z`GU*wu?fwc#ylVK(W#;-aMEtI(%$a2+cQUDZEbB>Sd9X6WMo3msN<e`apsNSu8JlQ
zJl1`uHiC8LPAEy^&Ym&%H{{Q}fbkewW>EliO+SW^m~p_1B~U1;%|L75gN1?gKlItO
zQ}RN2Ltk+vdVvOl-{~64;sq2IP7fIuC?JcnGJ>mh%EDYm6)U^?a^*wb^qVD3lncN@
zxERy$hxTL$B0mdD6(6Ms<Ck}dIewHG`fyui3>2+0pF@!Ve3l~!C(h5^CGra9u+h$e
zj(Vo7&<+OoP>2U#1pL6nc!+|LylCe@fFK%pyhBNb@zlyiCG2)8)5*VM^zAm>_47hT
z+3Aa+Kgw6)@FpCv!(#$}Cq21GE)gzd2BA1&=M21vcb-pa?kyNmDiDUcM;Sm_*eyyM
zVl1AWn@aQZvuZPw9`_Fp8=}vX#68=<4QL*(RsZw_=c27rZ<uJ$S6&no?ItUaQF!Y8
zIu<}dKyTOuw|C!&rhLSv#^_M4GJnnGR{o|J<-=GuTbc6FY=cUedpbFxhkl$g&T2e=
z_~2fu%uJ<2@lV@(?1(*d-vaM5tillg!M&yQXl)^V>(OdjU7SnPQxoY>Z4%h%&rqPA
zoZWbr(5szBdbxFww$-m5dZEdU+cz}e{#EqA?i-XS+*dZKG4SMY-;1MWXJ>@>VJ|46
z-Z<iMSmQH0d{{J08QE34P@DB`Ef%v(tB+s1|DKQT-D{|?IyzN-u*3D*N(Yqekisxy
zz4EgxLunzm;27yfofC^b)AK}SZVMl4>En;Sqj6&?&CSjG;hs|18sFn|dYP3xltt}L
zG-0K^>WH>RJ+mY2pnfdcXo(+#|GRbHM?2!KEc#!eBK4`gLSx)>HKV?}CR}6ZxIcCt
zMV-^mP_S`=4hy=Fzloki=c;kHZSX?dCk%>I9(EJA+wjuxrYgV6FhSq0x^o`lRgsCs
zczx}@nlYL5Le_Gj;TPUqyuA8icB!-AiK7uOH0FV4AL)F2?DBFXBYDyHLjTB<IF@!w
zd`x9ZeDmzI7p}qc4Bj$EHpBxp-S#S?L;5Yt-NcWebo}X0f0q9Bzy4p+&wl=ec#gsJ
zqu=>K`uG3d-%Sh4YvNl^(pJ6clL&_4`@oS~3H=<qel_lhm*$9Tjx1-lJG>dE2TTjk
z9Q6s#*ty8@;j{zV1M?92CFVEaUdbq2^lK-=2jd~U6LiahJNiA^E(@Fv7$e)4-VrvV
zhu!iVyLc-atBCLBxM@ZY6x9&qzUYJFMfcSnTjG6vwT$Z1eK~&+E&sgn5qx3CFG@>0
z#}qm#_F}qm^E=z$2*ab|guUd`HND>@&teKJ#k@6ckJDZMj(LZ>W$AA<Q4U!AyCL40
z#~t(Tw8K!tJeY5qCWerPEk1vtxr@(V#gLcCStgHlnHGAD6<`18AN&*g3?q$zfY==#
zdknxU#Sn}{NAm`lC7t5oMw3(1PsE7tuoaWPI`L%*=sTw5+KpJ!*4y9wlyoS(ixvJ<
zD&M?tjUZw44q+0G`p(H14qM-ik4Q5@h(!u_1qIcHOnx+7@X^6tPRePQOmY~rGWs@b
zIULE_n!lYbZ$A+)a8)410h7yH0viZ>bWA6_!Z=U+2aH7XcLvB|m4=<d*QVIqZkz`e
zvt;3^Z?D49pwD2$U<(64{S<$VFJjs&>EWYCX<m#81H3yOh!%fyhQSb%5289Vb{ZTE
z3wf3_`{buVAkt?bQwRn$ciK!yLQ%lc7r~%j=7}pz7mR8MX7FQcJ$HJsI*#JtR<`V;
zSYg!-XD*{a<LJgGnvilyvsbamP&^F%A>eczet`jdm~RlYV+x%i^rVJh#wmX&105pX
zxG7=k5(dD|ojMH%m@$|Z20@3h=>fi2iR3;-al|CB`cWfx8rDRX6*sK(hIytAp#~-}
zOy=A!1up_HE|z5ybS+NGOyTrYTF^v%c5yLHde<Kl1fZa<5x!}-RD4#0U1+k0;?b*A
z^o`Jgm5wkRR;2MJysFL<C^3VqgrV%itdg!yE01X(GcqK4<IT|r+*ignb_QLr1S1q@
zvR^B0<)ySRFaa@<5sbnIlSJU;xIh$E4BpTZ!YuqbmXQ@R6bHr^6mq<XX-GJ*f=bW<
z^F{l_<wUr}+2mzH7AxAILlk@0UM)*wWfI08LN)51_5i$uB`+oe#(^H{g0h)c5k6R5
zK3`k3;;UBGq+meIG3^*a(c^aP_C@;PO#M+P+ii<lA~YuV79Bs{rQgFIA$#afmeS{h
zzwGaD!y(+V1ta$^Ea97j`i<$eV|t=?>=;9y9g1Ix+l5+oy-Jz<QY5>DHB1C`fRYBr
zVR~FT>5_(B!bCVsZ)}-PMaX^!Q4ZzLQVX5b6-72#i0bHyM>I#}2jc{IMZa9eD0>sh
ze)+p&BCk-0@tO0sT~7{7{(Kk6igA|m1I{qv+ZmU84QVTEC$Am@D%4auFk?!2G0t9o
z&c|-DchZr?j$JsSEcuafI@m$XT~nfWOom)PrM(0`vzuPDMcecWg_Pd*D{%D$&URw(
z@Kp+CFQ*g!bju}7guq0(9JCmD((jT$p@j~BP9W@M+_&9-chZqd{AWKWgk8Ukzd6ZH
z8cMf=TU_w-YM}6=GYNcvp^{TXFplikm;Twm{52}R{+pus3RB)J46uMH!w!EDLqwT0
zW#O47&K?QjJv1&*M+JDsNK$}@gja%yNuPe6NxcO`!U^|Xac9kpCXXm^d`3Ury<#$P
zr=M?%_t-gXr5#O%S!IhNg|U#;?+EuWp*(TMu78f3VQeO>1~n_{5JI2?;m8^4hSiu!
z6KED40mCf2r8KE!feXTK1VzWE=N80rqz?jPR>>nIM(}{Z3mh>qN6-ci$eZ*~NKKB8
zSmDLa6BRpgj1u?g#4$F5-qnUc!n~OLDFaHlGZwUH+{1VwjS7Z>10l=N@riNAs#kUo
zq41g*=`+p{X4Mb&)Ar`J_>dFtpyDW*=1I!})=+kVA7!TpiZ6CA1`*HG4=nJCjH~db
z!xQNi!HjBGqpZf(k;FYo>?lWg2t9-_<vg`C?sNARw1W}{p;Xkp{KXeoC@bEA70lC>
zsWdY+B?w~oAHp+~V2tg@N(12>^*^ld8y^LAnQFD!w6HKQ-fYCY6N3MVks*bnzE2JF
zl{!&&ZSC&Y)AsIOIz8bos$r*(kQ>8Jr@5K2v^-z+y$>jj8qIb(Igj{OHr{YYDQ#0_
z85%GT<zuH*U*3ger2$IS^R&LXnVvp-t}@N0zww*D<s%2#wKzQ{Uh($QIAQhl_U4w4
zl<K>NziX$(#W^bkX<sOg5TdTXdgZpm@#yy+K2W)?ghPZWqC5H1ir~$?%jfLUf-;WX
zB3;DTzV~Bn+-(a**}>s~a3%bXjVZqsuWqM4ZiP?oTH08DnVvoUTC~y>-=;J~$K(&6
zhJv_xbd;*o(_W2@5}ae*IkxBM;6Qn|)7XUSQLwYfugWoi;yIg*qqM%Uo;Fop@@KWP
zk7pO|IX;VBWs8eTKH3l=9Q4W#F}DMiuR2#v-~Zm<NGq%N1?sI=hd=rHD}{R{|7+jR
zh2P@RT3T6IP4^!@6zx=uWA5mg7md^e4|TL#KUTldu#ge~-lh0e>SRJZ{fg>;o?V?R
zK2uqNi5+d1Y7b|kCFr3gJf8|L?6O3`!mg`p;bCugCk<xhF|-}4;l+oG?!8-5IOoVM
z@QR`qVekF>E6QU&Aw_Mn)6xP(cALWIv+Ie*C|cH>4Vppt3olJS#@vt{O0~J!G+V8B
zR~rItw?D;GlM{^WY*N0>6VYbu`a08t0GE5#s?{kUd&zN){lX)HL*^_T6?@u7sUAXH
zb#g8XhB*NbIHGT};7+tiz9^Q#CC5v0)FpRJb7}<qI!BgO=4$DqkAK5E4_Qr*pc@4P
zb7r*+wflB@`TR-R-PuZ0qRYu~^);ds<_+9Y%$-H|R#wu53b(OaPrG|{E3g=cP>``G
zkv2V{zI#G6k5HZd0|gYy!*aYj_~4$$qvNLBLzxII(ht!uP!^x6!Q4Z}^E$=jDT3!>
zuoIp#qWqv;F<uY{MJ4ke@I-yZ97p4UaE9>$4f<8nxtcWnD{0B!W0vX;*kT+}o&wT*
zg~B3ZAYTis)kb3n9YT4=GWuY~Xr-+OL!FbZck&UByqpiT0t}Ri{wz2vV7t<swTz>j
z5E9D0n4cjW20uPo2Kd!y6<p9Kg7>nM891P43P8EKMqw46{21R2O9#H<hflWiZ?`an
zH(s*AulSP=#^a5~82U`+f$W&2PpYVI<_#XKsDQ>$5UBnbYxZ{btbi33y<@mmn^&8y
zsxH`xeB_;*v`J250S0&<0Hl1(Z=oMA)&&3Zrw#*`Q6Gd8^cm*hl)_%hq260CbH@|3
z4NaBBDO8oqZ3Yt@<npRdVzJKJiu%Ll1@R%tJG(cxw!JHv1*U7O%T|`7xZub-6x|%_
z*A{<(oM=e-aa<pZX4aN!>B9#rR)EdSsGpwZZnHS{_UyVpwJ-Y8_Tfod-)Xo#Zfx(T
zgGR%3P2UZ@3p~-xfES98wsEX6dQ|j^Vq{)@?$Y9{(q*UJfa#qbr(Vb-gMw%lIflk%
z#xr)xPqDjI{Rzj*vG9l8U#A-TkX0bp&<jfT!dw<yu}imcLLaE{nRYFjZ5<z`_2*xw
zXHPy;A9AdIV>m4>EvJt^`p645=&RTPe|^($C5U(OgLu^q?S{obC`8~-SuD<RwujAA
zFA8({Rj1Sr?drUJ;YAV}?DRimCDhpHP?}O%IQnl?V^2}$L>7Wxg<xSnya#d+^D?46
z<aq2rCO!B##uJpPXX1&W9WQEzpVGL**yLlh<w0BJ1SgIyhX$$N$`tg@!Yy`ZBDV}a
zm3yy7gl{W?)%TyC-K66a@fM2j<9_v~Z$qX<-#pBMaMkIQ@}g`kR3rYmcrcC|uJ3Gn
zd_ftuxVV@qvvHK<&h}n<{!0ADix=tXlczp`U~F{U3ngkA^FRFXeeeDoLcTOOY+ekW
z5`K~$)s&N?oms2`t;6fKjv79N;X=3q_C1aNtE;Qwc8&}x>3hv>Y;X`>n=+{HR$n#J
zf8(n}*ui(AIvi3SEYzS}v@huAP`GXihBNVz{mNsZR!J+1HQ{g|?KVz)Vgj(S<G(I`
z8>KTxsnQVYje4r@?WA_5ZAEhbLyrYKz5QDmQ*n&`K{Va*lbieLZg@;LZhmL`Ywp<&
zamIJar)zq@OP<9PSc-XT+#aX9{vGoUck40UYN8x;d3Qs+Gml?j-j)yTl?U_7q)AA_
z7N5V++{NdwV#td=A87r`w9sp;`1;5HkR4wN#xDSNN*4t+;IgQlbiQr!M-<7!l;6OP
zJ>qm+^&9DgrNpqsY1@xay&YN(VOK%;^tZn3nF|#j9r)~(4R(i-aT?hn;)e+&Wtd6Z
zjuGpd;G<+v-2LN-xM>j3FP)#$#u;g||Mg9Jkx4{TfQZVV@6@`mMzHduPeS0RN|h6?
zj@k*=d3Bv^$vEV@n9c_<_27eV0LnmQP;=R1(9VS*Wbtqyb*uU$J>mnS-1H4Y198Vr
zEu|ql-wbF|Gc)Pod+)jPA5mxOtTWLXPkb8`R%WnXKxx1yk%x>Y9VA5lW>#Sq13m@S
zzyM><P8tSW*)>SP=raMb^2;j#nQ&4z2E({}idA!L^YzgL8c<=-fD0i@OM|NiOuc=-
z*VkXC!$wQ#6j*@<3(T_K9UKMBf0f2`rZ;tnz=VPZedR1|0iMXc3xVp-AJmtA#fVXN
zKEA@PL2X82t_MEB0cG>f3q^EX@*e6-<6_`gWjM+m2&zIDTjO6$P+{I-*qPX|dXSaT
zw+c8}o$<tPrdHFSI++H=Fu)`t0J{<HVWPmlcXz37T%W3^8!@$Z)8LKJ%k?ezVTfSH
zFV1CW^}CoN;2%VgHZzu{s&i>-b~X)b@I`pRAnx5-V!T+L2<e(xzmApeFznD7f-!Jw
z8WcKA!-9*QRWK&t@0xH*fMac#;IX0xy7e}8xidLp)hVl;Ij+e9LFh{5K=B1B!>CzF
zro7xX<w%*_F60kPZezkFaEe{gH|-5K^*x$zc;Y9xJQ-8@f!Y&tH-zw?t<0s>rF&^{
zcG1jm>`?5|1$~n~l{WYYinb?&r8jPY34#T+@JlK=5)tg~1pwvCarJ>wIQKV>6|p;c
z$V9C0_kT1~o9wEzu*puI#3&y6iN35G*2eHhybhQSS3+b+<QDOAcwBQrfe*P5MKE<>
zb~yZl-ZAkjLlE9L*dn7ahTr*zGf8!W$R>wShTw*)Llv*cHotAMcLC1j%YH7i-}>o<
zS6O5?9QF)n?6!*jNXo-lPaNn2H<8U-VU7=e^j*qFUUG9f5x5(iH@*sMyYWhZ;(6z7
zAKCM_;qN3C*4{iEhRVurkyKBhVwD^CfVKlKIZ1vnwuf6MUEcDDp`L!<qclNn94xFd
z84e7NJmjN4r%9i#1cH7Di@QtOF}BMMd}ZJqE{xt$fG_|;u?b2(1Oq^uoCfLZWk7h7
zM7;DwSmMDOJ6Zf4K)$j&?GjP}+%L$(^_dZ6F!9F_*RKGQm=5d20v1QZuNRRD^N0W9
zU*_0<{pLUaZ~nP+j*{hH6}$-htfZssh3ga0ya~s+U=v6RD}VZi2XcHBkTfiaqJET?
z!>KQ~(2X~FXlzEPz%C2~7$}XP2lBS?Za|aq!I9L|WPJaK<7rx+M0){<@+)~_*pcOX
zvNRT;7-9TmtmNLI(J_u(WQRu_1%t3>L>W>?>@Y$Rg(8Rg;b_Ob2A8>W^j`8LdSQ-B
z@F(Rxpvm}|#(D0_;IeLZLL)rl-Y<kd*K%jv<Nhg*nu-Z@T@z#!GQx<JaAV}JhcZLN
z-c2MMwQzQI>kGWWt`{xo4`Cw04_5n<CU=olCNxnO+*fBUkFN;mo`_%j>Z_+|cx23r
zY#3J=+c8uZs1H6eXcAnzzVt_Zz!wsTT1Az4+ZRVJaSR~*-q0}jqfCjHnM##v#X>p@
z2{hr4)yUkZ(_Psff>Vx$y0IXYT_CJ52G<B5d`y|jMV+xi8~&6N6M)^ysv(YO<5<aK
z3(8hjmc5&d^jTfZiuI`pcDad<Fux?e<ScdwZtd>-m`cb4p(nfcK>`Bt@v)(_T&t#~
zxr*wk-@E8`74}HwKqzG8i1Hs-9F&Zd6I!EwQK)b!{Lb!P+S=Oj%4m2rgdZI9`QX8Q
zAE!9df8`^kz$L<?<D+Kcex~Zga5_CY5I&N342?~&8)w3G&1&f8anqd<d?-i14G#%t
z;^Clo@B)pq5DMi9yB|e{R^pC{-$U>f$Cx3&I=500x`Yp7w-+20^uta%?sR(j@_G95
z%P*AgnH8D{7)=L?4{vvPaNwP3k5ndvVBAMGtP-`G9DgMH*ti#2A*7`&2&WF~hv~)h
z7e21)xOJ@j#Q!lDxrzA){eR4nxFcmat*)%4`)g}nsgAOZ<1yKJh4O@*bl{FtDp2fl
ztQShikx|-WzxZ;&_TKwxY;rms93EK#<vV;-cj{46o1+fZ{`Ta@@r}%pSTWA-(k1R8
zQJ(bs2pFkf(&K0^?t4SPx_@|V;pPCxiYY9LhWg%K+E?9!TiOPF847v<EE?4O>+CdL
ziEn4uB8pK3P(Ly1oh0{G7SqCFP4zvRI369kL%GJnLKe8W?-M{Qc;jv$79T<r-~imP
z0FQeXE3?%!tGNY>@t7yMjjJ3iXyXX1<D)o+wW)UCBg7O2AvC)#YHGKm!+q&sf5-C+
z?pK?do)x`Zsc(x@CE(dD<P^OzcR-lU4oeorQ7_CN4(ogA)yt=8{nZP#`=%Aego}k~
zLuqkoC4Kwj?<ek9qkVXH0)2?;92(>fnu~Lei#vANs52J-p+H+%T1@v>SJIr|r9E$Q
z%v$|OeLo5+&5eX7c70H10UJAC)n}-0VV(s1^zk(Qrf3^HA0xy>d7<>EbLK-nvXH(<
zY0;1IU{GQJFKvJ^q+^tC@M6qm;utdDBc=YEg_TyI>CFxw`a$!`uAdn1dKF((m=#9M
zz4YgJ<S7_R9_BS<C?|cM(^4JK53q9=g-Hla0ZnNrPs(5=SB{He9F7x97=PJa7BDj6
zAjBp0P@IuBi#H;C+0n*BKN82`Ul~@u&CBKTG4H}o8992*#`w~^FgzHg53D>f<|qvn
zfHVZD`>ry!^G7(!S8x#Zq&~#sEM;fsE4Wr#;6nJJJyHhmkkjL157ozxslNgjz>m_T
zR#QK%GO}}(g#w&%Lt4<(Y_*!^YYSGMv2&G03hbgq2>=d2FSv(RXh$fh8N={!`HcT~
zfLo+*_==(Kj9c7zgg@jQL+tia8}TkW#ex3JD=CgJMojmgd{fKxoA=h1tpFoE#-)wT
zO_imk_M>^Q+EFO*#LuAI6#kI`u~V3Zv*3gFK)*OMI+#{!6Y1e<C4Kw-duo$IY2(!^
z#TiJIxy5ulIF{<o^K{s3r)}{^yK0k+g^UHv#ZlH!KduWIl%XvA;f^Yf$F8Z*U7DZq
z9Z)PJIy+;$J9Xb_1;5%G%5N40sX3?9<3`%u-F98hEi8B^9=p2Z$WQ8Dd8s{dWatD7
zax@OGIESO54>Ufqpyf=oP76aB(>T~spY<$lZoX3gdYC3Ot}iOxwR`u|gNKjO#B{}Z
zus{Q019xD>ZnP_pnZV5ADHay&9SXL+x@aS0Y4FGxl<OdryEdzHvuQ@W5~K~ESrk<&
zXgX~+gJ!M;Tkbb~yrO8EcGyvTDSsATaO^TiA;L5Dvy0d?9Xra{`3gl)R*(6Lhiu68
zp>fuVuW5y2tRC#QoCetwJQKX<xcpHVH(KZEpdnta%>rb#NfjyDwO|tt;cG`m#Jh?Y
zW0xIA4)5>pr|qqc^!)iVjg7ljKz{u3x6;G=t1b(7pRTWOr>D<eq*t3;sUiLbnp<67
zPTzX}ajLLuTH`vrGz-t*?^slFuJXqHo>%VchS<T0f=^>03Ot`;B={I34^UodtmTMO
zcs+O#D_k|Eu-h4B8A?9*2ZXHVh_7M+&!OUQ3ITVvvMUT_B@-WxzGSE5#?G$C{HftV
zwTXvmRb$iH#YK9)zL#EX#sZkx>YU~#O)H(Fe&D%~jR`>2FU^6r>AqT7(`RAv+uzZT
z;fNhW8^^fcgMhNV_l>cCewXjg`!%<Cia(~;jwyL!3LL%WweYgujeqwWLr?HFcMRjj
z+fBgPJ8TqV_iX>B=56^z&m87m@V(6t@|(i&%cR?DdSj7Bbf+<*xHGw9vf}H%3`zqy
zy^3%GQxfRFlPgzA6HvB@=NQ>3kgp_5ditZ-4igSx94`ha+z5}Qd@DetS3>EduqU{O
z@J&8B?rW$ec*OE=_{c@@bB9+KX50c!hp>fj!st5}s4Eof!sLI$9N&Z~@%`(c5P#sp
zD$y>?0aNE9eaDDzjH40)f8IGv$-4|1*;STY27*Qs_R9hCN0<)715=<5i}(>XI+?&n
zBof)=Z@Jr9ZjPmo$lD=0`2@uJ=)fHKuqVbJj;WV^<xGb^IW?UgzW3PMs+sU3L=F5Z
zJz#Qh#Zy>6?vzP{!CcI}9@1k1!YWS~O1`;xlVljwILeT)VyWeSX{8m4FV2tmWD!P8
zj1IV91q6yQm?#uw2yd9o(%I1|Uum#p+dn&b_V@PF^B2!FsoC=C5-OcOJjx8VfKhk=
zW(pR7|J9m|D;NftS#V|gmMbMzK^)9&SV4m>!;fP!99QKE;bOo@4AWZ<*@0XBFw!K|
zo#;e8LF7ZMaM1u-nW+etP$IJO4V<w<%R7{m7x?8ebW-hyQ4-_1xV)ryHH~QU%MnP<
zlloNMuu{6Dalwi~P5f@yLaqAo1VVMFB8%~hmF?GJW-d^!Aq)_6a;>nSpN@T0Oj&h)
zK2=#wD`p4<6$}F9inIOMxy2m=DBMt}u~I9SUuk$UBR}%<AT9i`>IqDdJ{SFt=|PCa
z>NprYw#E;LA)w6AB9l%OQlw4$g{fjf#buwQ1B1sle+)TOMsSIO4Q2?TlZAc4PlN#u
z3aAe~{R2$CL-9o$i>>_7pu&nd?Ux%&>(pc=tu5Y5%L~hCbR?KUNcA<Nx{bcWImoU6
z5I}Y_jYUn~=|2pLf5Tl*c-vt0U=_AVTYt7GeDndD5R(}8oOT4njd#T1KmX+49EG>j
z6O>LbqIJVZALXR<Q$o;ggJ9(Aus!B3%t~9W!fqWf`N`p;?<Zo9k%Pwr^2+4`kN(c@
zFcI0VhF^Yml3R{vxN{ycrVuKABL;s!8}5;3RB}OS#4Vw6SpB#dc9D_2jA72k!AqQO
z+$eLnm2iO<|JZKS(x&j|r=%7Cks@L6wy{fCOWOX0p!qJC>v;R8jIlXp#}pTSjceB-
zc(Fsc3mz~CSWd<6c9BJna_Hd0F|u)B|7QaYi$2_L2un-BpZv%-@PSA0Nf8Enq#efe
z0T6l@H~7@x3Bt|k@eO^*rFY12d5n1fcltE{;-+uYm@DhXJxZh+ZQPJ}kO*7_&l35V
zhJh8BaZ!AVi!nG6G^a-R9M5p(fEZT*_=xlZcf^yury>1X@A%3)zJQ|BZup@*8Hbpt
zhw@~cWP_bA;~{_h9Vscs>&3C4OQw^Cc{I|zRvL0sLIFzd`8|FQ3U7lxiiSHD5T1Kd
zFBsMRlZb^0tPEz?5JF`HC;dv3bhw*~@oR91F)W0m_zz`zK)ArTO;X-j)oi8a3743+
zENnmkhTx1bkrmk#35CYwBs&f`QicgNe2TpECLO4Z$?;LE?MXQogClf8U}+wJKopcW
zV1UO<#!h&xelJ|191|$>ILcW0c~?REMC1OscNA&xwEv~XUG8Ce^7YfSv9ax=)W#>K
z6o1TVV?rSbV!?GhnC$KE5Uv3eVfm)M<f-!UO^x&q3J)4zS(VE@J{%_ny96(avQEg>
zgkNJhLLBfVe1t+U?%v|)DTFi#)ZroO2X!wL0uL<qAv+X<2v{!8IAZF|!i%!|fcl+w
z8u1X8Ivo{HHZ0f&`q_~_lKRB^)SD-1d$*pr8;#=#;V(G45aA}Pz=xFwtDxEKhcMdr
zd#Mh{3;4X)N&s;L5j+gIf+y!54|bDr*A(|)or@NS#A9%u(zG5{h2M$?zc_6fR*w6j
zPT4_HnVv|G*5=dLpkTeda6XqeqIA)C+;cT48XB+?b9Zk~ba*Csgg5T8QklRDi=((_
ziq-gwi#79jEVx48g|Hd{63Q{(;iD#SELvu_7Hx%j0fJU`+i}kgcW`mX7Q2p6O7KQ`
z$*#MjL+)&8i8i1cjtUww&X;_7eZ@4x(M||d5xALulN)Ub-jLl~Zd;cu#JflwkqY1I
zxs7P|^rU4)(Sv(y={LUraavhg$bz|`9VL~{;EjcQfh}J6LVfv_@4;o49>V4kwSz~R
zV~vbWTgXX!9e_SfgD5q`n<@<i*DTgTv2c5HDSnj|>}n?{E|h27`y+l`^*5rrRF>{*
zj#WO6*}8>x)PC7XQ9n5J@rdxQ<WU5a^c@JgxW@!um>pf<0Yxr-o09O#eC|uTr!rS&
zxp(bW<v_VY8*xV<T+^P|xz$Sbdc*hmq09mg2u`VAjw3fuuRf361PFeu$WeJn%d7QI
zW4BYifs#`52=%R$aa?VeW5K8nR_f2!=F(hs&O1$yg+pk-3((X}p&antGAR_Sv@OVi
zyXijv^7FKRuy1AS)bvc6=D5OP^*JgB#TXqM*E8X^ua=|w=g#B_xs6#V$DM&U*B1(Z
z?LK)@ZH0aufofHK1q%&OO0b&$Na2r9M4N4tOL0^e3JB&;K3?p!9e1;+Z(+wA;|NEn
zGpe5ozZ|Q}u9qXxD#xJFPg@WwIwKIY?75x{FKC8&3?`10M}Q1}pneg%a7=gv^3182
zbD+p)+(ej-0t}kC(VQeF8}mis)x3<5OP;hDb|x_{c!8j3+>23#6Dtr@jv}-x{Wl5`
z)9aAxMSTUv<4&d>@4S)W+j0fm<<9~@D}EFo_{#1yhtr=GEP8MQF9_nHDIWS>@(8@p
zruEii<yS5babhP_%q1;n5FG=z%b;-Jfcz*C@4^o-&j>-Q%HcYRZ}SfdLv(hz0Rw#p
zaFV8v0#rE4O#1_$ZfJU3w<;IB2@7h9@SL#=#TIE+HK*rpFXpKnrOBOx&=&OoEF6(I
zYvskHV88NWPj)A@MN1riNc*K+^vRrlLO;xWmhl8R0eMpd|A3=HbCito`5}NXpgN5{
zOK!qFbqDU;XBh`h$T$uqq5PfKm3M!>|K7uNZ*A4-@9iI?t<5dV3D}KG`$UGsq7Bz8
z{e<wvE;e?CqTnG9#+3fsOSPrGw7OVHmC2#hcYUt@Vanqe3vL-x`UcfU4zM8U%6Ejq
z->_haUC1b;S){{Q8RZRS9`~EGL%mj=NOLnBlgGW_mnue|`h{`JBv8oFrr9~ej$(GK
zKYRAVZF)@c;ZHcGc2a%G)VO#)wfAY2%{#qR9v0)U3m9bwJ8fHMH|fOf|I&DdSD^oI
zov0sB*{?K?G`TwyMWONrZjKFn|MBDW-S2(uYf3nF650aK?1UUtdEn&^8ci$X!5uWw
zQeW%)-o>9n=NuiqdT%w&&DGR*->M%HU8(Ib&a+5>-N!8GVL=S-1iEIdWc-6JIO>>j
zg#HRTn3|pzU#I?5_0FA^qiU0r!lxHNXk124*U*>&9^%g2s&`dGHtbgBcw6ofWnspb
zUwoFneDc(bW7gKz(s#fA-E`AGlAf*armfvZI`_#r@XG8~8&jXnF^B_coZVjHMUlxJ
zHTKgV|H+@EPd@uR?QHM({!PNKF0Z5q5AJ!aqAr1*;{v%;ZB+GKn_o~pRlU1wPGjI4
zCoPCC^63bQ?><!Z!2O)9_O;5Semjp#@XqWo<DN>!K*mIPCFTQOL?C+eF^o!QbX0ty
z_yf{(KPme5(bH<H8vL!W+u7ds4zZQh74vZ@<qi(m#i)K-ye{|It?%x;Zs$0`>%p2A
zp0J?h`OB^J<k@rA+tSja+^2jJ1jiVsYw;GN{b@|F(6>N)r9X3f=byI>?(yb3!eMCd
zoiJ|iy>a4=!S0xM{QQ2+EuP|!>9u1@AA>1y^qSYg%X&Bd-ETQ$%iB#E+uk&<!`W+a
z|HY|>ao?NG+w#G?1bJ6pZ!?7a%V79b(k;g6hxJXPI#IXwhUqH4WQsolEocD~Hvp^w
z9uM3d3M4zE0Y<pR+u`*eehL%O^;S3>?H6Gp4?_9V77GmG-)_7eJi9rJ-690+K;#@@
z_=)fEk6(;z<gdhvo68|f#FIO3Ovy}cIP!8lzatKL@EPH9j*kB(KF9~wCTw8}<hHwj
zvgp%RwH;^eaN%Bzt%S*zP8xbjc)uMnjD=iE)G$Zz$j26c+F*@_OPQ6n3q(M@Z2*Hn
ze7}`FBZ;_Cr;(6d!z-fb!}W}DTqwwv{K>DvkcMoy!w<1-_WpwhVyu?k;XAN8L5GcC
zhOIDwfn~>*TdbzS4+E1@7+>Cg#E`=fz=X1j6$TV#78696G&9orW)Nmj@NO^|awekQ
z@e@bqdBD{Ga-~6;RUjObv$M6Go;`n__I3|FkOBoTdv%hHpfyeequ~tKze+=A)$KUS
zTW!$4$`FfSpq{}QSmn=X5_3eHGGI_MtZTzKn%@8&sfAHkJY1GsAUaIt!3F1|Gh)~{
zuUL!=6MTXYoRz5H9Ol~-7vYj~qgf%!`Nk-tCN)^XG_h-nGTjRIFw0lxOm@!2n2BK)
zlgf%ZB_OagKr76-zR8M#8vz@%$4W%b%5P~<=6H%+tTZTwV5T;kW*3&6AIE<Tict~g
zDP{o%krlm^%M>lTP~M^k7*rT(a0FcdPdRg&Go5|xf(CHv-V=;WN5q2(Jip+`Llmqc
zZ>0-!2lEdTg7HMo3}39!qj3NQ1!J-mimxnak`Fr%VFYc2E9#et5#{U?{2YM<GtDX_
zgbo%!5LS4{aJ1pr@OWBTT21#>?uiMSh%}U`a~2LuTO&<7VJ&v@(c4FU<#6Fu;^lz(
z+x<%?UXC2mY+j22O?SMrMV?NZiY8CevzrVZiA*Qi{I*-pKjQo40MSo}aZtHMI090l
zl7OL?CHwc7aK+^G?2Su_kG&fYH-F`f37?y%<L6+Gg*Ag>hpu-CglDAPb<YXQ6xT=_
z2P}56qd&o+ow})Xj6NL=d$y1c>BV<%-eEd+1MOtx7NHzZhH=1)P_XOCzH*N+fsbz5
zq=6;M5Lxq6;J1SV@-DdMS1A!DfQUe{k_TyJ5M*%Bt2Z6T4|nH}!5**9J0LG%-;R$?
zSY;S3U43sZD<+B%AKloBbvHA;4YPiFLiyMIy7H&5E@_23Jwkj(tnhW-IUQX5DVuR4
zJC@)cx~8NZ;FO$cdhz8ZNV)Xq6g{+bQwV>Y<eMwCcpHA4^b-ms#&&#qIS5B5<zKtv
z3!cGH22gJ?z#Z>mOq<!*-V?Rg?`;NN1XIF!ER?+y&u_+XFcWUZk;V<p#WhxNY#xFs
zjS+;C8>^$a8;YG-<CENVFm6FA2lT-w!P}Sz()h#VAAu)IFBVC#K!Y6>2vZT5v*H#z
z$IJMLAB}4$aeT**9*$b!qVn48q=jJ|=i|B)zePPWailCLQ8+q`RjBQAU=X0>sRx)@
z;K4XfUI=^O+it{1Ktmi+gn)x1f97VUEN~bZ)L3~H$0>ocG4U5Ct)}nJJUVKm_4SRk
z#cET*$Z>UErAuDH4A}&Zp+BD7Hh9~eGSIJNy3`w*==ue1v-9f~A%ezUFUHY3TEAjx
z+#MWcMXq2TA4?pGj1UeO5D>F-oHuDB@IhF@ad0f0;ixkNOz;U;mk6{GNF92I4uV&d
z6DXH_ROLXQ>v&##6+2&02Av5PL*lKL7OLsq>ave~=J=YOy}IiG9)|XJe0-XAclWJu
zYo9db-Y>dgC%~8mUMP#8CzNyWBkb1O->-W`J3IseV9J9)hg~@cenm7Y143wa9>hIH
zC(WaDE83c#m^ADf_`zbDnHf*N^}P?%<A;KANc5*P4i1}EvVdQX0i3DKx*jbAJQqGx
zEzl1%X?(&vo}VP{D&sPD?o&b0u_HLSydEVQf<Jh`e$@#|-1=c7u_KR#Q_uy%Aj-ns
zaVS_U*i&7jEMZ3tw26Wc7_7Kay|I#-qxM!8YiVtHK2>KX%?lzd;*x!YRR{s$Qz08*
zB5jT<LJ-W(7~@%W2+!PZpSlj<`Z$IUWemH^W<(1p<(QYiXVWf+#lxKnS0|zu1XtXh
zSE-<Q744|qr@kSzGx5c0PjG`Qh=MoteWofSJSK_=F2qMk#JmW_n-ztkC3dW_65f2S
z{^);DqOmLM+(JtPyzsdk5jkMK_)6)~KQzsYw^aU%G%330qJG--;nA@LqSWOT%2RlG
z{c$wkvg&PdY0-)+^RlWF=*vnF)eGPBaaJt)m^JW=G9i>-@X?fu@*w2&5n7|{&>2?y
zhgYYsVKL*ij}qfJC*Y(m=}*~V$*xh3KIBdv6rmV!$n--1C~ftZXvc1^e1z3`yOpYy
zIn%+x{(<kCQ-e-b(JTrDcIT>taQkpSrFPC7v#GXlbA2xOhO9&w9iK2f+uOUQiN*Pv
zm1fMVC^v%bX@%uDO8SaR)y1jscY1~blu>~>(1mw!p%6s5hr)q7YPcg3fjK)NQSiA<
z(th-ygls5Z`g{7dn4c&$;evdNVuc=$x}^-zF>@3vCK<bMpl{T7D18)PgZ<EuXj^l9
zXbX4fBIbgZN?)}O9qNf)RwO~+3YuAz$oz_X2<5LjcRYoIZ*=-uXy^VokDu(kL&*re
zls=mDp)b-d#{Ib6LrKK0RPN@1zL5g}YbaJC9BoDxVu~)=ofkW%Ojqjnng5X=urjYR
zo-!VI=$}HdBc4KG+=lf47o{k>-GGqrgz?;rhY--U{ZU+@oC>Z>FWDit7*DLAQ#dOQ
zSr8x?0|w@`icbXo^?-Z&0on`;SNaqdT4OBXsKDjLg|x6xQ)=v@JuywuccFMdK}dfV
zcQbK!B6s7dpMf{fL-~z6)faPh<N!tS<?&QNSZ1Pl(3x>5nr9rO%=jB_*;rZu?xHUU
zB_nmim`E55*s$VDc)-m&B-I}xT<5sBM-T6(2lwxp7LOavw5vA85xMMS0Vga*uwq1I
zVGM;Pa648Tg0C`8g|orF8{sFV35`*6mGSiG-eQ_j`x3*UwsvMEjt2Xbd=&Dql?IF#
z=V$aEsvDI<1@<D@=#QWajoV|RJ~D84fV;T+y<^q#1<FHP5d`NKst3j_`mILGM@}<_
zTwh<LoAc8&(s!NariN0D6K+%v7WuFf{aE?e5A&Uq;+s%f4h)T_xrJpf9O1}lWHW_E
zQINn7UJHMlo9pR|&%a22{uh6dUcPvyF=`+^e*9kgTfhB-^zrw;?_+Rjqtq0%=LIwx
zD^FD3HvP0{rhUTcOVlTMut-3Bgy@Dbm~iYKVo?$};YeJL+;;mDjX^I~7K?VE4N;V7
z&rM%&dsnZ*1A~xcr5QXAG|R%E`C3hNI;p(q7e=fYnpeLLekt?0(xq=<hgVy?*NbOQ
z(%!+o@KW(^x7D?KseO5yUhV9sy}IgA-4u&*P#}4Co${X?A4+3GEJ$jngS{OqzW(&j
z{yhEs%O@KDnrV4qKK<Zt{ibN}p=lgtA8mt$B+vrwhSL|2i;j$p3+^$u2NsHyg=FmV
zFZgYWCptR4R=o<>ZSf0oV~3aTjvGgrl~k_S5eJXJIB32`9@HxfLWD}U6VhXG_Nj2z
zYSh!{FuVDJ2VY%VQ(KsH{e1ex)AZRFU#XuKX5lGB*O#}V7y5U>46lIFoL$505It-h
zsXd)*PB7_R{?IZC)pT=~M#ZxO2W^>#<+e>5=9_=s{-z%&-@yl9-1n6*5ESoDoOolf
zJLVlfzh86frd>=4iy^;Wb2lueH+*jyin;qOhirMf`BfEP-^9e&@pcovY4`n2)5H*x
z@V_WVNw=7Hm-8(KTFKTojOs)~+#3eP*T9c{|3^RW79<d8KmZnIBD-}6PLJLEfqQ=Y
zMVLs-?j)o_U@e0(z+fQj7;t&~rAK}Q$>|wrHAKT#-1N(5`Njl%Ic-Z(vX9-vkI>;o
zPC3m;ExYjww}3~l;--ca`QjJhDQlERzkZRD+~i%5jS?)<l)u7~Mz<p5Xa8*1pFi>o
zU_;4gr-+AN{>zZ(UxLdP5d+sI9^Nil6btv9hyBX0<Avdyydns{M2uJXcFWhn2ktqI
zc$IHGei1Leql3apKaL)CPKVFg{oIGZQFH-op(h>!>v5b&tpkYMaB%(d>F=(enOO~%
zFrruL{2APe@QOh+wv5A^vpNbUEEHQ_fu(Q^xCnzt5hj?`ldKMcX*kh@;y@FfeRb?8
zIAGrbM!l&+#|l`m7pAPZ{mQ1BY{xj|OuK<+gfDVcxt+V)B7YIq;RO%j7-#}EJ%&5z
z8y8}fur|&&hxJDWguF?Mdegrr40`Zo&;}+?ScEqpMJT*7NO3d`13bb9wr8_q23VPp
zQ6~(-zRyZ^;MJmHBrh+{&6qIZiaOPVM1Q~y)sQ#odp92pCM!|(haXU9j5vQ*>u^sQ
zbb=4ei`~RvGU){$X3kZ97%eU=MUe^<fg*{)og)lAS;+KEcve5g%tHT)%PK8aXQ41<
zwG%Yua>*Y70r|twxpgb8pmYA{f`S=_9%&a|t*3Aunx#;=JnowuNiMlQ43~^?q`*x0
zz;WDB)+felTuc(ek(jh8Lcl>J#U*)g+o1@<hE<`kl>B9nSp53K{^Xux=$kii<+O0b
zjc`sk?BvUvvXLKQa4sfinJqFojII0;zQj{MXg7dy7~T=i7CnKXh-1hcPVZ6x`xzL!
z`^~@J@i><M!!O&3M~bhx**SdVCUi#_{nGLrPL9}1-_Bon<!{1dV=F(|2`jiPCc+ZO
zU&=G%BRs?1AeMCP6CtnypzU%on6l^M+D)(cDM9h%UBde#{Pd=*PC`FC;mq4MKJ;mT
z?2=RK7~%&VkxoxMy$rwp9S3_@oOi^@VREd9Z%er4??NlIyRbVDKT5|h8`_)8VS7N@
z394_BjT8gcaObDI$TQv@nDg6ppWVwZ?Kj}}SN;vZD6HPZE$I}4m&)Q0`YEBKO=F0Y
zZL}AEgrA*db^k%y`YW+)6rg-Vv*1S|GcD*RKM|&zCh5e7+-U9*(ZR!+{7V0$Kluml
zZ#D7$=|9Ex>o7m~ga7EqVJl?_=*tuRub|3r8`(Xk;#TZEhVSg<rwGb=pS#9x{J-Yc
zwMRVo9RvUf{rWX#A;6BES3c5YcraF?B0geV(-^{Cr3lv%LIMYP*MQ_P_EHm9juLzI
za^1(1AneNCo+MuqPB^&Ir1$Ln+QJRm!|<3U?paYncq=GSZt3yPk34Djn5O^;(-BHf
ziU(jtYdJc~aW!VLQvv0HhFeeQ5#DkM`h0aV%}$T|h?>D$P3#2?t5gw;AVf#dzO%iX
zo;`ozdp+27!?27%kFmZ7eHnk@AlvKTdIvAbH`J|PzH>MW7;|`ka}=VU+v{^bR~IPk
z+TP*B?j6Q$?v6s~$WAp?iEzQm1L6_RPE(5=Gy8jKduvlTKTg~OHZ4AhqYvAsEi1c@
z8+GxZq2MSyEIbkdVUFNt*A4geuo4=<EUS#+LALh~t>n767LTUK!V`sKN7Dt$7u5+o
zF01-2w5T7N&qa8F;E&^rz#}XF5vn8DCC^5q5pG9K!3aH_8CM7sITnd~v%I?w{9Fcq
zMxDYtAy@>fm*Sh+;_na|!iRBJ844?QL9x=`*jIet8#Op&C9jH*`UW_z2LUC5agH*J
z9Zd-JP>v#4;mAz*dUmtyAGOk9>pUHwT_p}`;G_hE<KAVjx&dxhO+%|FtE?2!gW~P9
ztu`Tifj4+A6fzu}IAmUS#>e@vV-G%16{B`H=%b$yuAvO!h(dNyppc47_EB_=cy&5E
zO*k%)``Hjad-s~`eFMYkOnfZIif-=<&qpU#GI6gH$_o70vB$1YzE1@k2Nm@ViJu;1
z#W%vkpk0J>2vF(E**(Ra2xS~NgRiA+5)}Gi1^Hnk1biq{$fvD`Rr4)&dYzu7I+xlD
z9_|S{XvC3+C<WnbQP^<&C366;G*{j}!gF*~ZGGHIRO+cHp;4w1hn4r#BMaS7tRv86
zrvW?c*dYcVZ2={GzS<%DZ1Cp@G>4rw_x5t+;fQFb@XnlSfTFo5#aN|3E&5_dV|9i*
zj@185j0pymW59y|)q{>+AyD;Vfq_s6Q5NoTVaFAMTuuQ2f9&GoSSWVQOiowQOtqSZ
z^-X!9Gw{HjdpB2Ywc-8r;>o9JXXEQMKCC%O|4llotL<;RNY9^qp4!LtR26)yoQR=5
z0RG>~6tz?CkpWH?3YL1LZoogaO-FsHeo!yZX@GNfX59D6P$%>kghK$!?iR+3FdEnB
zQxvhxgQIT)4>2Y)|G83p?|>?OA$6fZjQ!9TeK>(*rv~$D(S`D$KV!a8=AN+oo<n(V
z6Y<E$xF=J&xsC-hbVUd3vC8hzf7p*Va~MU5(=l+5T}_22ayi_OW2fJahq8K3$-GK-
z60sm1sc<e2@VHDp#?F+Ta*~GXg7Cyg0ZZCZ4Db}`v)~s0z-i>=KfzV-Y+gqZU3&7Q
zTuxYJCPCqW#U}KVv>VS`<qX<@k$GR)g$^#k4dF&d<0wP=OX!rkp#2G0=kGbA(k*qs
z!<?G3uqb10ZZ1_T757==8i&W!UfAhP9iN_ZBxkfw#sv2#ELKGsBmXE5Fa<r5kNZ&9
zA7e6op30*1SOm?kN904%-$z*Z3(FK#enBPJM5k&CKE1|uK)prz{Vlpfk!pE@+6M~V
zmUo%Pjz*N~^w-cI3kXnpdpEDf2oyv|8rR|u!8qO&rJb*7P*`e?Z0Qr*-1&5h@{N8*
z<F(R7DYCV-l|K9YGaq%j_3C+g`SoY1vG+1voz&A%I<wNO|LQc2_FtyS;gn{^`cq|M
zFwIP;eTau}J6F9Sd!Tua4)%Em`|R{YnyXBD5d^!ZQHIQDOj%Js@!tJ=-r4r8#}Cu}
zm8Dd1zrlhV)zxH0IOM3)Xv4c(JL$!<7nV^>Djj%3>XCW}$M6p*%)$R`Z9c86tfYtc
z?ptBBx$!c6`svTplc!I;bB=m|#z{+w>pyrmFN_j?SrEm-oIw;M+*hdak|TNfMghd4
z6%;p&t@LH=EMxZ__gJAA1MY!gg&&zr{UfTgVb$FfM<p(%+Tu!@Us_R`vuW7;#*F$j
zm3N$lUb2r(IzMEMepvJ(+|k$5XYK6nrq93lJbm)%r|HlC{4dhypMUP1ec%1~qxAjn
zew=26L*FH@hh2Hp?}Xlyv;)QgF>#|K{o+FzdBvlM_olC7@yRmRxM=L?6HeZLxR!qS
z+kaDOemi~Z!^i3Vy?beSdD+KpPET=1E=LrGVhSA6+wdTJL|Me5?@`f0pYnmf<p@LM
zYIIYK9pfW?qLGxw)TD=Q&o!nb?;PaV%()qM2ghO=X!u-wKlkx*2iV^3UfSK>@sXWh
zfBCrv{C2vxyx_6e$}iFW*6u-iwR@btc)p!pY>OW}yG<t-YJcKnhes!c?{V4t&40i%
z9X7;&ikIZ>S@QJaYta?BU|urVr<j77eilP}^pova{4U>}ckTo8_uIpC-x!C>@A6%Y
zZTQ=*n_e-mh5ZG7n7iM-Whmy&Fn7N1m^Z?A%$u+l^X@pWo4+g^<4-Z)G))X43IEMw
z{Br4I-Wi&7-s1Uh{^$SkpAP)~@BYD$Ul-<K=fzJvG`R53Vf1!k2r6RD!TjU56QHp9
z+x8g5-H~E<yl~Hf{IZL~`KRL#;^V9Q$|wy6P;@(_SA_i~9(IrUE?&4t6?W5L01bQ2
zLlFpO_f8x!Wc*~ef{z%y!z~_Xgs*+Di)HYi%iiP8yWpc^Hv~D23`bbWDdL11KlyEo
z|KkU{9+$=L{^`I^K71DeGcf0rU7wWOLG|Y}GT;SI4kQ!$v=TP#IfSgT;TIm|<tGI2
zIc@&rhMma(JARm?dZM8LGQ!DM1CqnarQ~Bfeu4)kv3z1k(?azK1Q%pR)XD}iKwe3$
zHxbMbWg!HsOJK~r8q+J0g86lLz;J?^h>2O^I1mIlOrSZ&rFj_JxO<8%&{Vf72Dmfc
z?Jn5;$e!W-<&7d6KX*-}&C13OZ3s6ozQ(wKA<LjRx$z-uX_G+{GL*H%*AwBfl15m=
z5_J@Yg6d13u~nL5Z!SG4f^lNyD)R^*X`um>`oJBAn>*bmU_iuF43B9NEH?~zhvDO4
z*B6sI3*y9dGI-u9Emkx75@CIpNmDMZ@<GW4<Bf7twa4V6z%3>U=2$Z%#S@1hXI8r5
zKcLYVW|iX#!3i*POrpaebyGBs4gn6Ba8^r_H?#sBK{vn?s~|I-$<CS6Oay{C58CsD
z54`J%BRg<|-Y5ZmT3D$gCd#f?>QQb%$645hGk@~8pk8lZsO^bg)J<tC7O)Cdr5F=B
zm~>4lCMH>NJ!VD?TDRg}e+3RDhD{c%g{DbZCW7Tx_aoaJKHl*|e-wSJEb_5`d^keK
zE@Nwc>`tR9PT3t7<3J8L0dV-`v;#q5?Ed=0%MnU=f5#8b7%0nSmV+U1xbP&5+f4S4
zw|#Pa+yYneDR<Wi|6%Xq18>~QN0iMG2~g~vS2#fR{?-p){CC?f|L#p?>E`ohy5T|j
zoQ5pC?VA1K6*veG@eex2ZX>(P(DBzFw5RXDZ{V_YDaKu)W7&Jh7Yx>j)f1lY@bH6g
z>Dw}DZ3<p-vwv73lmkV`l2rLAxbdg0vJo!|-Rog9pyax7NL#;hu#*$FM+smz*I1^D
zPO`j3<K&~J8Bv^eq*Z?SqYn6vxRIyFbf=x&94gmOCOWLd@uwd1W3qOPKlmx>5XLbb
zFH*^Ak+!YaN!0&#b6h^?u*lRO{F07C+r5n8y=4S95gO4#*bOs(3jSw5`N^+O@%4B9
z;~yjB0Ll)(@rD7yfcRH4?@Bj}^X31Y`G=pO<gLf|e+i^P)g9x3I)05mnt*D|Wb)o_
zHPcCxOThK^v1l==N5BK0bJ9Gq@`EEun&Nd3Uc?Tp5DXy@Ar68GgzA9C_`!HTrdc=2
zERK*qZXBn^VZ*yxP_l(kQT47jN3Ajbuo@QuxCSi2&G?-8HSuBsT9fsGRGl763$xQ{
zb!k4$Rj1SR)R+|tGpvBsxXqoJo9nOA=Eg>P`C{F|929o@^}2WR0JjZv<Z(3cRB&qC
z<v02JGy!ntFpQhDBhbq?V|&=+O_^-78|@;5z>N3spI1J9jt4%I`<!mCFVfYS@P2ZX
zT1WNN*xO1+2fJ2kwT=$baeYttYKj->w=e{y<;7_;9UUH|)8>(JL;b>6BTR?)%OF@N
zSK1_aM2IpgJ`1G)N~>Mf>E3?btK3<sk8q0JWxxsl%qsm*nw<)7d3QVf7lJ6!0dzuK
z_`IUvK#1uhm{o3+DCEoAurTV2Ht2`giEhD%1=4Y62ZBs?V6i(LzK~~3{2<2@aZHgF
z5mrX<9<`v%in}ZLdyUTGbx@pelp5J_qz}1Tz%DnAS;HWVVRxI4C=$QJ(N7jg3zrD1
z5GGMS2wXY#YGMju&Y03+#r#B?o|#D%j*L>8z+}bF=%{%o1hpKQMEMU6_S4qJX4>A~
zO2<b>rXAlYrv4ZK9||>iT4;r`vty`Qt@t>-N#(<Rc(88FhfvZY!1pmwXYjP_G}8mC
zsB)tyyH*G7W9sl%li~g?lxKW%_b_*Qu-g92I}OwhgdFNif!r2Q4uD(kP@^m;6!+_m
z^y0;4+S=Mv8r&V#5bh3)!`3OseyJSbT5z1{hh29Zo5h$z3xG$4ZlED{cL=e{gT8_L
z-~xY?9c85jqIowEJJVP(PJKfU2%G795bklDCUqNvb?7nNpf&Th`t$BKcEn9jnto=c
z*}cOm`C-4=!7+|N-Q`uis_#L0#c^QlfUr{1ib?e`N{9RA*o8&7G4(^&m#2w4aE^}l
zRmL;7OK2H7qV&`myK88dOfxKmr;W&s3;t1r4)k58#{N$F?2|uBKl{larDspRP+LDy
zyBKu){`%{$($3bV1>C>!y^qs}AAI1W$`AyXqao>w5U{g@&G&cFzC^{$KfsF>SDJf(
zf8Pn1@x?uj?8b$zIIfPPv4D^AaX%Sz6X*j0e|eY_Unwox25;tK<)}g)=4l8v-LDA`
zKoUo41zs%l#GiQ!{UbOTU`J8Rzm%QbphX-kTf#$6>0>)?ILpcIvU_JC?F8kCANqr!
zLC4b{brJ44y_^Q$(bpIGfck6smGOr@Jm@*1fOGRwI5>w;j@jh&ik&pTcg*Xuf`&Bp
z$NUt3E0pxsultBnR&pbC!>;mC9zVvL+O6#L2|hZ~eH^=^z#S&fg$<|52Qax!5|4h8
zelL`>^rh5+@`G3v3Rp4BcdcZUp=}{3ryU{)=Quywi0`#i`N2PmeHK@M4`|SHeQ3pW
zLEVE}!AQrZFs0s5##!m5a#;Z;Hy+=Ur>E2-<t7}k6Q|T=<UxKhCgV<6+8zE@WSah>
z9u&fS3yVFDj*na}ju`X~GvQ)z2t_@8PaL~@+CKHsjg;M2WY8Xv8>u|h0mm5gX4h*A
zg<0!F{?}=0rea0W$moRoMeuTQuCy8lX>xQR)ljnQJw2|x`$cc!1#Yj}sWLg@9htO!
z_yu;kbCj_LDM23ef-)hLg)EFvUrN6s9)UiqHe0o_1tkhNM45Das&Sxsp4zIDn|@?$
z^ii>c7usMP<4#vd2plo4GB!XF_4<MN6Y%T%vy?YH1IOJ`oiEnceOKD%=5~7a>{<Hs
z=bxmno_uA6-}C1$tW4V7-L;I(W4&-pJLO~_<}1)`)VpX0cfC2nI_NW$V$ew}kf7d!
zhJ2rv(qo*U4WRgDe1(h=voOYz4+{{u(+v3{$}7eO7MX;ild@j9jDPm0f13W6fBmo1
zv**vmuO12!H{zLQtl%Th?Va89YJJl<Sz4Soo~Kn$ljAH}8PNEmb|GG6iX#Tc<K!TA
z7cML8+UiPLURq4|^q$oi1OIj@T7WO1ZdvFA|8-d3_Z^$$5sN@Xi>f2|YZly`DlErG
zGXB>a&9uI?o7OjX<xbx&|C46AI&Z6gZVMAvX}Iqu&4^E)txl=j8dH@Ya;*!ETaEf*
zdQN?P^0V~GPk)lWR6TwD<(J;2HZ30H!M%HFY+^FK+S*T_J$de9I#14T(thL2i-!7C
zFYK6SaT53A)!0p^cqr&SPf$DY;wRbNCWUX>JpBZ9pkY(|)`iC$+L{8nr4zQivG`rS
zJ8y;Q`F?wt?i=H9`CYz?u^oTAb<@N2+A+WEcW)UmZ}?;0=GJ508pq~!SbNR8<J>Xd
zG@Q&gPE&@Eg#YF-h?7Hq!$|K82Aps4MDf-4PyXS5sbL`ki0s^9(6RE3I|2rN{KChB
zGE4-WBh~<u$Na8B#tIV|J35HdiQxu^NE_i+xz{SvwgsF0aWK;Kb|?}eBriF1vcaxD
zb%P;T%jr0lBVvdb5zB9U5-K8)e^CS>fO$7Qu;*|EM#AbTg(z`=$te&HZ~cNh3}ug_
z5d#9Jn*HLX-~3I0Qf|NOuJE#yaurzYlf(Os19tn!WD9UaoyZ{7!o-Ku^=G^s#r5T{
zQ1JAojf3sD;mL;r5RaNM-N=m&7Ypv#ykbU3o~~5VgNF~(B6oagumd;X3Wkk)M;s5B
z99@2LvrvMR7|0z`f9BWaMt&N!7{FQKgK~r;N;s#V9YhEWe58ju=8Fq18D|n2->fd?
zI35NOR=JS6nP{cS#DTM(xx0W>7JHfma`tWGh%?v&J5fH@w85|oaA%ydq9|;JnAb0D
ztY~pDyLp0l3qBY?ln3y_XBr3-vxAxWa{h9LO~ei@z(jyE>LWv@Xrf@Hi89A^_=XX4
z9oacO43iV*!5{}SAm(dE6E;su6^6l&)n9#Lc0558<2^k)Yk|VB7>GeJF-+LZT(VM3
zOcvYKsT=P1x?=LBy5#(63t+?uFrnZmC72va$0P><4Z^##w!*YyvJ9OVcZAY=OpM8t
z@||GS534*iDTbNj{A-Scf(f+p4W?Q+fa1uDE%z{fFk>*(Fs3ly9REc6Avio$m2*tN
zh;ahrd@gvvqmNe-u25c4g;qAn9hzi9X~th}Ru(9LCv8lusdU9en8buiv5a;r@pApf
z@eX;?qkrNOC?;4u)PtB!dB(AjLz)E7rj_NDv^2jc#)y@B@>e_rkibbE7%1By;3|g7
zeY32lXb0?u^tXQKGu^N1ZD<K2*!)rmj!EQAT9^=iJ20+x%OCOSb`+5*vXvybxGl&r
zM<F?ax)0U|FoF`tb(|c~Y2iS)c<~eTu6TH757KivfL|`gKNk5>MjOLdjNBqZq>!_)
zbG$ll^21)nFo$)3cjLqTV+S8Of`eo?$J86E{On@bg)5^ghckSUzl|;7P8cVf%V(HK
zD@XN9KM|(55ijz?)9}kgma;o?#Icc^ow4dEY4c#G`@-ZJc6e-#gOz}hq`z}X{n1S-
zc!UnD`st)+lWiUN2&8`nx(!X70dU$SsGd&#`eHl-)=nOH730AE33nkY1j0(Ogbi;$
z;IhC+o+V5H&H3oTKnDai{`CNJ+>ynD11adO?}CHyi##x;*m&7Pwx9}3sX-aVi^VwD
zNekE*?PP3^Fa)5l0Cl9&ZP%ubPCWd=oiI5kr%OBv;-vf|AH#|LpZyw-I{d%=*Z)s)
zwTXV=ZALH)R{TpCx3&EJwpeBS?oG4T-fN=o@0t>>^XAX_2%mH!xH)~?p$9!M1U-3L
zfRN+ZjUR=B-->iq0|@9WFyfnZorBVf@-epQC;REmrz)wxAHFGec_RcPAJRqGjNpvj
zHz+{jK9P}htFilBWxP<j2*MGHFphcED?6vIVuFm&mR&Oldg1Lz7nrz9=<BbZq-Rf`
zdF)4E=|v{WgO$qA69$|xhWj(3+E6OS+V}=%;E4x!eS?4gbAII8QGUrU+{51?90X1y
z$JIOX41Yr)*gA&rB^`)dUW5a1->oFPF@D2?o?9rzIM42Qjy-ehlYj7O)HCVBk8s3t
zpYXu0DJ~r6GV}G#t@LVR(*o^<xw-VwcRuia46OVe9#lN>Fnb3_X=8KG_pTvufrmzk
z0aO-LQ7?i4eto7=O>=Xz;^#&Z_n?%W?L~-eg_dBX9xS93Z%y5^Ljl^lf~SKQQdz;Z
zd6=R8G&M1j-hZ%~K6r06)n>-6bO5IpSGTEsjxg^$9k<$ET}*%OqwL1H+oR9Q5>UyG
zN0b@}^;j&my~}+Sd!of#m3cbN&Q1pn3r7|(DGoam5j25Y_;mzm2%Om2f*=Z|35f$Y
zd>F_2wN5WmOFZ9&c(~(c(~9Er(-YrM%L;20TFc_GCZ?31XdN7@w8m%j$~UQ2olc9j
zs)bXe3m(~R#=TR=tuybq^Kp{R6Vn1pqDpl}We~qQHewud)Db&;P%7aD?+TyHoQ5`o
za)+Hz?3!eyeJGdcyJA<*mH1@txM`i9`CgXotv&AqWW~0RcGLrp3q2y(<G!{*;WreX
z;EJQ+a(@9m8Xw1|9negkC~!D3i8KfXIWs<(KNXG@Ozz@w)#efKvf~T-^L#-Mch{n5
zA~I#N(gU7Z^hA@Q&!Fs;sR=(mI$de`z9ltT+6DFO<Hbaq?0!RV$^C#VFoS4;pE)4A
z7@LBh#f}`+bkROZTbnP_)32YnOv}q_>4(4jqxAUwZy7f%0zf#7@DEr`ThO)ga(^g1
zo}V_;`pf6()1UoCT3>%|m}h1xX?|`#EzH+kUpqTHUJUxZ?|$sN+2|ur&aP6Ub*}kl
z2-ZW9Ht5~26QgP~$_s%g_+m#V{o}b{tREhwz5OG(ow+?iXOHeLTOho5P)~K$GxV}B
zH?4UA$EU>|M7!z(IBL-=_B~g*a-WQec?5mC+Oz5nL;p@0=^KEJoprSPxMPgF8_PY+
z2&?msp3yPW9s*M8-u<T@lqDGEI-VEh@6Oyt2AF7*Adoo~Z3@%r1LaRAN)-T;6)mcJ
z6d|<TVvq;WcIdAhS8?dANzV!q3*-^9=k)PaostIgKNZ0B?C~(i;{iY3-KB8&0~7t7
z$0wH)nus)`OaZqmiAQc=wkY$$9cRW_J#s@4fFcNg-(y9&oGwB1qkLGDi~NIn3MG38
z)HzC%-BlQH%>0xE4IC?LURF3|Cl|XS%N!NFu#0kSbum4>x9UYFjJ*eqmgrXWq4C3X
zs5$Y8`mLiQ6z(n25LZcnCzV6M*{(9sMu5H48E=YdEPziQ^^aC?2v)Z#{c$1;3SD4^
zzL=*|Zt(6Mv*1hLv|-D7<WC(VB%<$ks<+f{zju4p)ITGG2qg>3BzB43q*kk`GBlLO
zx%*Yp0#5o>!i<ifsGs(@%I-`0sT=bu{i!xPlUA2%R-V(p5A>^VXdn9sxe1Oa)tGdn
zHc{FQ3lkWVtx!_=Sq#9~!vYn?(AvVH`h@w^KQxvaO^r1w4*ebrjld1Z-cslNf*bw^
zzKg|4KIW7@R4}vC6&{Q}961ZS{(WyCac?xfGfpVq{^1E$Png#@X&t7as}s}Ai_QJC
zzOkM5_cSgv_C?QU=__`7ZEdN-piwAReUtj%N%iq4DxrUl4XspbX<=zOJ-GKE&DG{p
zMfe_{oD>hCen;UL2T>4eX!jxxw|&tp3vQqb`h1Sg1z%prMGRS>P3Qtz_o8FP=l*`i
zGX-`3&dx24r5w_P1RjzFS@aJrjnORb+S%MlUw!_Wj}(6N=uuj`f8UBN;$bGIrqXra
zP`VJ!%;W_s6XT<TShS=}ydX&Q;e|VDnj;>UpfczhoIs8oPkE7!>IZ3OYa`Y7cP-aq
z;UV={smys-{P5^R8lDiJFsS-hLttEKop4kz^g$a}KSEm+ZMc6?9_;pmZ=j2Uwx-4z
z7sh0d`@pz&RKU58=(kvi_QjW9r~l)R{v^G8@xphK!2@zP<=V<h`qsC<Be;grq3G;X
z{1yvn7=wnD-q{5^u`h-5fi$W5n-RUzH@4bW?gtJITdAHGyU{k?N6JudK0288CVp;Y
zP_SQ|ismldM;kuXf3*2%|5*I?xA>%vJ;ou@#5j)9E~XnU-WZ3;@A6%Y?fBb`_nT=I
zQ}V?0hAZZ7_}(&L-tfn~&8^3{FMq2kFu!Tu<$l+E({PwyDlCt}|DqTry=Wm32jlO*
zEGq?i6Z_eyUPMZ7n7HHXAN^O>aWKzvF1F~cMn#-XL=z)}eg4!_{4IEsM-hG?(6F(4
zwvwI-XdnHAa}Fm{(urX;(j!#FBLuG5-O;e~?xv05x0`DBD+hJz-88b16YktOqW?}{
z2P+X}u=Ylaq>5_<jB>;$uOcrClPwOTyb8<;^BT?+lNVE=j*&Ho2-lu)C2GkyehV&p
zObORx?Bqa^9tn0nF&QDw&gn(CNW0{Mkvsnxpom3s`juOR!A&-o+rCbo`ePt@_}+VI
zett<E>xjLbXO8H`L)iG#I|g>I#*(9t6VRaIob_f$k(nIq8vF|5V<t#%9+WDqdho6x
zb>ggOv*Jr`&V!_s21nL00rTnwP3jrAP=4)c0{v=z!<UL)h<WLNOL-O9FuPOi!I8`@
zGcGx!&?`Q_tm#2b*C;;*ALk-Zlsv9JOb_x<gt)!PlRp?tpafRnwlJ1QLFnMA2g6Ll
zJUJa-Nb7=tr*LQ7!JFiL>;x+zS!pSzhofp@_0C}8jBl^>)L=}9K2@n$`N=Vdl!1we
z@5a(3@j5G!*yfGm>sE}7CkGl#nIL*aiQr`gvG39lz`zZo&v82^VivAoqW*vE{aKT>
zOLpf8t(d1B5j*CHcOK4sPNt-kTBWMWRCQxu)^t<ca!GRqeTRWrK}$xH^$~z849bj#
zFTfRMMS;;}T_u&2QjRC{oIJxj4>3<W=8^yJzufoU@eU_5m8xtm@V%es@nQM$<@0i1
zKB%*1paH?eiB6J1|A+?04TN9nyoW}IVvNBuW_`J28m5gEVP_0p>`wz8l^>L+D_}fe
zu&~gJa+LHkP-oC2=0$d`-5taEZUPNta`2US%w#K042nD`go%@4;0n#;iw6p}75&nC
z&?IJ<yvqqYa+5Nbj3A7EZ5L7Chbg3<V4LNDf&O4LM#nAmx}iZljC^p3yb{UrP4A$W
zfi%RB`qXr*^pgrr$K4a)wHy{}@&RR3Qb`%{ke#%}S}Fl+yJ;aAR=^lZ`hhs(YWp>S
zLEG(ZamhC-{X<L&Wzx#-HWoL&iNeZY&{OIFHU(4rTXpqqx~U_e(v$H|`x1>V&$7eP
zFdf{eLX9j4KW*(^UP-Ioa(6fs621hS;IApH4d(DtUGc9XX3Qml?Pe&cLY$Kr_<W_3
zoxK#l##eAunBttOo5o_7eymG<1;DPl>08sW@)8M&1$M<B)-n@JRi>NvxVWWC0OPYx
zs?cVawxyIUikt0=ro!00DA_q5&=z|@(RL?`%Zdz+fwY*e;46BA<x6(tOFGoN11?Jh
z2B1howI7D*<T-l#EO--~ml0U+V5BfwUto)M+84OuWV0>!%d#D3Tj*cPW>uMo&NW%<
zZCXpd7tehD&;Ql$PVn`g|M~y1P)PdHhrCHaEAGU<&{^uBAKP426<38vwe_!faG&=E
zmAFe7#V>8V3~e{lZCLsZ@kQ2Sqr^98#a-#~r@&A(I=E5BTaD;V8&d#@iwPC=SawcQ
zihupRpo4NiSaPHfMZmya7Z*&5sSih(!s^taq5fcp2D@-}Hdf;J=s@-jFH~VgI}-}@
zvm9Nty}cD{Yn$rt_k}ZsW-j&6)01T=kDY?Z(wAD#ztD$KI5VyF$e;Dqjd&Oa|D<Kd
zPfbJfPbht|udQxbrin8@mf^IOVACJ4c+}HXy3J?EAN+%%5G3j8AK)&PK~GT1!hO2H
z1f#rAdgfv#ft+9X9w3gb;h3t;?d^E_{CRM%5<BhQxpz0-efOU4qT(WaCYfI@ugBx3
zEAe7=(?Sp?`nYc@C!j?q{I$a4j@HS^saTkw6RuJ7*!zO_WF(WYt^-U`=gNC{BPLe|
zL?b(UM#e_P_gsj-h6fkFqjd4nX_w-+^vpEJ<L>S2ap%T-+|$D`n1{zH0ALmS+Qx2d
z`qK6@^MGTcV=*)$-c7t+9&yI~JH0WaG-4s!s>;aP+FJC9{^_YkEH32nLfj>SAnsi0
z#=R^E``8I{%<5`*f7wAFX$)V-WGVN*FyTlMITSnsuf5|#AB)9_5a%kJhx>bRu)nW(
zainyvV)uhElZmQ+@=R$Ydbo^!X}%eY^V8BJ)QmV04IEj?ZmZ>0mGSj0Pbf3-Gb)~y
zqwl7srj-X{;?p_CamPn=Q9iwkQfa_0A<D6LdJMsTvdc1sflLBgh^b5>o(L|fElflr
zL_v^tz@1$dT(H_-aXLhRk!6(nOoL4NJC6~1=?8z$gfBurghyN4DI@$$P)>^1pPQYt
z0F0W1aFP0)b=F136L|1r?9xP#L*6ktNV?_ZH}y2z7<OBb-#Mwzv4b4v+82#s)r&Lk
zg<*FFJ1oz=lYsKUgd%y&q@o37N*4rQdGynfN+(B49meCw50&QYj<fGcky~qL(<}Bl
z=J58dJMq2m{fF`H2OkM%kA+^8hiqZkk(di8xet0%csABnV|DputgXCMe0ySkaUtgC
z7d%Fye9$-V?rbZchK!qI%h-jo+d7Jk-GeyNkYTUn3H64f37g~G*`=|I$_!~wxkYf5
zLgmv~-`a^6FE=d2W+It-e*5N9OixdHK_0>_@GQ(u#K>T;+S4OXtgmx)rOFcb_POm*
zdLz`JJ<H<_nPjF-B~<;$M`a#)IWjcjI)WnphlhI}_fZZR@8ph2c5)7TY(v?jZApO<
zxRi#n2QOdT2=7Xts<<SK%7Dra`JZFkY)^qnGY|?_F!io4+9LQOfWb{Xz?J@Fk3T*9
zTsBot8Mk>%Z<GpKei$E$5RY3Ctkp0H3uT-LfU919XzLs&#fPc~Ja{v1z&l17V-r>i
z(el{F?s1*mFnf&2E-Q8-vdga{_yV3f13t#XJUH-k+o<@kQzmgjt>Uiuc!!blLk=p3
zBQdG=kt5;APx9CIu_+vOAzxdZ_c3}L0lTxSc1L8f6KqUv5B=oY##XGTz1>zh;$A`0
zmgpgbp?^XMlXV(8sC5_aH|TFNU+5<@51^I0%AyJcY0Sf%9<wl&c++N)RtOw&SI`;4
z?vuJ3G3SJ62N2AV1_S*=F(`h6bQ)pTj_9CYWrCj*P}rG@VBPnWsXxmTYsf#2hE^I0
zuTs)!JlW)k+o_mP-J)%VZt60-wGk}QR`|Z9IE_B_Yb-oqXV6FywqK~7%e$Ubhm;Qc
z^c%_p>Iedl3o**HTRaFvj+K`j|LV3_G<i3)+6=;hU!onR&tj+V_RfJ9hfu6Js&WjW
z)X0c;=W#FU{7ln3=1<%=AA8~L0DZagiS~?pQzwU1UPaI6Up|VBjqP~yn0xJCcxTb`
z<>h$tLhV_<>a^P5elD0&o_Mi<(gDFNf_?IZ`!mNB_R{=(+`Mr^?cZWd%`C>)L^HVa
zbU^LkMNb~ro4cZ^59}c2F0BEl6@9hSjJl#<-C2rI1>T2wx_!lE*LQodP)Bh)*pFki
zb*G0dr8Vtm%ct<Ht8aQK`VU1HyX5}p4}L%1|KNid7d>9=BN!IbOi#}uWmFo%dnwO)
zREFqNRK^g<QO7uX*nO1xpAkKy;#CkdvU`rcl)Tm7eJOXFuDyI2t=4wdtuf_?+W+R%
zObm}r#>r)W9LOK{uV#CGrZ`EjxaE<>*)9w#FGtjelkeQ=I;{L0MW{NW@D(<KJJK)9
zi~1qqLLl}0`Lp=ZkA4!LfAM8Zs9xW@u@pD2sZU@bjr!)PnK|FvMgQNcGRP5`?6{-<
zkhRm9x<OmSJQa7c?d~7P7I(w0?}*3T@d*#qOZW>YcFUs#;CS!x(Y_c{_uLm}N=w=+
zI3#j8+e1V1<)4j{zS%~m{_jul<u7i<EQ9o~<EpBLiRy-{>Q(seIG|p0N4?3eOa1?X
z;Op0gs|slRe3P$l!bzF=M$ZCxJ@t!HzR=#isx0{WXaDTKkekUX0jTf^Sqd!DEXq1z
zW8_;bYA2sxEO1gqlfF*eR9M$|V5QA4(%pL8?LYC^XqaLR4F*D0O?A{qUtn@_6)P%{
zH67>@p3MrUkYi<7^w_6WTiLqnTlH-}McKYM*7Ts0H%J>);a5Fj5ijfYZ+pf<y5eR$
z`hv6gv})^|P}ZcS^@%%ITk8>wv0*MS-RWVgYKH|jShhV>1)j9Bt>EpySlG41-E_;|
z4|uUtXXdiyzX8<jU8+-=iu?E8i77R>3}6xgA3>QOaPvxo8AEk?(#T#6D0(->m|-rH
zsf0-}^DhPk0T%=JGchj+R_N5&y3WK2Za%uf!Y~GYG;T~B<cc#GEzWL-0b)SAwX+q=
zFVxvTe(FgVXr&oWvI@bnQ;T1;7*cs+PcS5z(~3{Jv~_<MMUv49h-zUWO=RYQgf4yF
zJtw^kN+nxcRgK&d2UdEg#)`(*c5<iFHOUHy!2;&Qn~DyByR&-501R@R;rO_jNmiJh
zieW(L1tW_e2JbMAIS4|aC5B0lkEc}nvpNT6fz@5U-^&bWuIA!lbrn0SSP2S37;czU
zm_k;gopRYT19FNCv=#a}Gn?ZTx!}Dq3DYG;P)ta3y6II{2<u?Jk2C-lVP;72OPDay
zj<54mWsXTTrG@W+(*v`?5l#$Pj$o8U2MOuHnUaum#Jv(wb__f`@uj~fV8lF8_F?`B
zf->usL4+k3k{$-Xt?V@!3<@nYGPop;*t_k}lY>{K066%`y|<UMsYhdWYEA>~+hT~v
zRYoW@!Xh}M2QA-3gHH5G!{dUCz3eCwrF;QqP~*vRZXC8vKNZ5N1Y%l+z{6Rz`gPh5
zUR@X*5GEAy#I~A?GO%z|IHHEO!h#Zehu^^scv{suns5KHm&w4=;miFXbSimGGmcIV
zU{ayQ4dfjhXxi8lMlY-}I9<VQgY19=p13<JvE;$Nz^6&J`9>88yI)DCk{)egk(_Xa
z0LMr0R%qL9si%LTKDGUz?edZa+`tWUtFO)JAGcy^e^Bkb>b8^K{fsTT9NKz4awiU$
zGFtXz2->Mv)5FS+G83l$B&&lj5y82(8Q3{}OE?LW9`FJJ!wZSqd4KJY%Km|n0#ACi
zpLW#Jgr>alR`9<9pJgejqrSNA=>eJ@Us+!g54MIY{;S4%jOYQ;Q%Q6hO5Y9>jVx^j
z)^J6mkNx4;c0((J(>|@Oq%l3Dhm0HxY`@Tqf2%G(_?J6j)_nL~?)XZ<R|eN6?^Pvx
z_v`#Bkjr~DzwWWZwc({I|IlsMt**d!g++g3B{SMA6{&#9hl%2Zv)Nif6&%&ADwkF8
zDV&@T>~@zeM!ZT>$Tm#&hx*zN<($b(z~sRwaZ}%h9fNrNnJBKp8v0pwX{69X{iXO6
zF58BGgGcaj%EVo`%-gANJUM1EF^RAMNMRvfA{>V&`%EKwDsa-7P84(kRPzvuatBq%
z-F>u_YZ!XtOXb}KQjRZQ!Y_KOf+um{KWR!uo^{jF<xeUn_~l2m;fp$eUncx7#Ci65
zg*y}fBTB=8xRjm=Q6>UdU^S>T<+wJEg={p&#iN|Y%G#<YhM=2;XYZE#KyEC}#<0?U
zV|zayKVOX(E1R*!eO4!@f@M+?-ucA$fH5h>QG4)j3v;tEB|e_KU<Ea&AowUmCfuN#
z_^ZQ!=ONETb&QJ^CjN&9hoi-vWZYK<UOnv4Vd9cW9gYCJe`_iJgKxbX*B9nuU;Oy`
z)}D`U+S%8$#{xDc`biziO-_=UU!zTtEi0LsxMnhDb92qd2{F&wY&K(IW;*7k8{&Ni
zgL`I~;5kry5%_UT6h~N{9E+bkI`IT36N;3A5%IMM@5aX(!OjpC%^kq&$uE=Jln-`b
za9<hY23FsD@>vRj8A4s(c~;s9Dnw&!IOeA&V|IEXrki8l<#u+wA1f<s@p5%5Hn+%&
zbMG7*RoF~W_9?%a^y;|~k0^dM?`cxJ>3_sqGFipFDNMd8aa;yjKt^~R&%{nS{7`?q
zqFj6;JSni}=gEiqR1C$1Nkt}{$$J(IvZIKd5R98REyF9kmDki;CK4IjQGU56j{CVN
zp&n090095=Nkl<Z=1ibZHK!~9WA~MU_1J<uV+V$!;x8-8=?@Tm9UPpv+z$;i!JCuS
z)ER_rBpt$EFDgUGsWju>mEQA{*jj%P8*AJ>bEtAiTfk1aF(1)LeHojW@(B(c|H{28
zOjfg_j&aC-YdfAkd!qb4^*u2oV@;QjL8aB`5O=q59AeX}-f1h@^+z3nwgZh<T6?<|
zEFqNN+{jh_J6l^Whl>k~@y@&Nd*YXR+-kLyW0$e|;+g8sX8hqF{!#qixBnm}#x*8V
z8@ay4-F&;TdvF^2M|w1#AgvJ?EX*}yady&2S~052U5E%kR0dQOwssEU>5J8Psk)3%
zmvVRG#**s#yxPi3*9qDuj*>+ny1Tm{tLxhev*o(Qq$c%?33l*#Os29x`cuZ-)|S47
zGQ_S7CUM!RN?CHb&=~6YG<Quh5zUUU5tUsg#TlzHZekqnaU+b8{?1qGFr`@r#xFUp
zCp_|<6r-Ba4)rKqRc7%=nUFnt#)ja7PIkF4S&MxYsDL}^&q;Q-+t>;g_w0`|9T9Aq
zKHLQ2c?Pvnq@8zV$(@j)%i)P0LdcK=PaIAaSCuvI`a(!0S4UTPJ#h;!q#u7TT#z01
zgz3H!IDVKBUa8$-+>0$czRnTsC>)SFoz?DfuN`LcfwZPR6Dy~I@OT_t+M^-WN7@F;
ziN~J`4}nlq<8zK}qYt5Z;mBf+Nyaa`eMc0Ik4sj(xFc^)ZT95EsD%qFYwKR1fB=T$
zQR(ZcC*-FH@Oald(9{lbOeyJQA(h-IPo!sUf9M-nNI?0g%y==U+({n<#H11Ri~RHn
zKkiHPs2yh~D~o9_SyZ7g*r7Z&(s18^z<h9s-7wUjbGJpDRKt<0lnKHnT^Uzb+@$A$
z1rgb%Be<q7<ng^#N`H~!^5wC&{4t-PCdxbII2LadftB(Rd-Bys@Tu)1-CF7|&Mtb@
zj&Qw(`ds&I?1Ub1+e4pDAL)Aq5w3_t!t&gR!h<#h9RPnofWNf;#7l+8X+*oqbq=En
zi@Tq=D~vXLdmCZUiNepFwFtmi;7lJh)lmAxSv-I8$nE9&>Z$@eiGH;okDfk>FCRV<
z&%v(a%NSAt!4Jn9ZftJHi2A9Sxj71SeErqeF`&f%)`#!Mts6HiC_-S=Cmw*KAvu1y
zSNiPN5K?DZAT24c9B0jraViLH0sIJQf-uebrM7)zYtu(uKYQ`qbqOKwf!er3l^b@$
zA(UmO+rIdpL-8w9GqW+Fe(L6}Tk+9HAIIG6yvpN<+fl-RXQC~0`yhhw%c7i<!<O)F
ztE{AOnIm^8HwZd8QalBz%<Yh$>I314*rg~Qfp(Uor^gV0D*f3dxWBU#Lj$UVDwC%@
z{c)m)BaG>**a@eq<#wGkp^bt+W~Z``SsdiPO7)*^FKK_6le~;m`k~!D-wXO|c~$k0
zMNcn$pBHo>=)R_U{hg2Ai*J4Mq0&PAp$3Cq{G+sHF@@)4^}XmFjKkAQ^>ga?)y(*Q
zPnDT1b_J`Pory+vy<ulzt%@A;U6eQ4650%QlOix|4EMyuh{~br2PZi&FKMEsXgl@J
zudS=S;?~wP2B1h+bGC_9b>Q-gGEDhabNh>1G0PzR>$p(eLN8P|Tvgrh-Sk)Ws@Y0*
zw8?L9vHEXY@b&BB>Q?x`-#7XERd`90w<3_^>#1Lq@`WPxsw#r7KTEs<sN~tAD+_qj
zczDuD1r)gK*weKGDP=>#(s`xcnj}Jc`;cc_`(<PGD%EA_e>!5@tAW#{Dy2!0P10P<
zMu5kH!88KZaq!!?^s~E?4Zp<=$h1$)7Sdn~^J9I&pnzx#k7$W-=i9C9oOF5;3=94h
zzrpn?xI2?&iThPQcyK|nKfDnSaHn6r{YR@$>iL$}mGl4x`b*;3Pj#17uhLn6D_AiU
z3=0spHVSBkPVkVH34yyIrJ=#boEWfoKzbN}xs-UoeE0r4F+Dr$)l2CPa5myhj+ibf
zIv5yD%nn<2$>@stkOB7;LL`GsC&?}?CQaOtsiU+241tCQqesWAWK26Br>Bm93)?FL
zQE9}UE#nO6SC&`e$)hK+uE7j~F~}9$<zfO`=fs3Clp8E{Qhcf^bfT&Gnc&sa-%%+G
ze6g?ai1L!BiM8E8xXde09KO;AD?JWbFB*0jM>A_oP8kXqehjOU$wLnwm^tELq01|Z
z@HGmGVGYO{8t5YYVyiS>5je5>s9!zbgzVWC4-<{hh@Dz6S3NM4%183YlLrT4UiS9Q
z@RBd}zATRO@pgAHY&~KY7=$r7am0$w@;$FBZX9n0F@hKK(W3`CM%m%T3aAM&Y4fu&
zsX@C}0IQRxJP{6Q2osirK(V2Uw^(8iKs}iwoG`fP3Clt*{mwMNf$=u;EB_u)h>SYm
zLbW?kh1Jg_mYMj>Jp6eJzx=|4dmtx<Dw$<6O@0-x&T2sAA%h~Bp~H+he)P0=%mF2T
z{k>u`dpYiKD(>I8uYt{^XiL~0G2Qm3Fq~@T#Rt0#tJFb-V?Vgax>rFs6h)MnVUIE#
z-!!J4Z+YR0KQ}4-^d%TO>+!4f0cnd>`ork@rx4pTO95K_14jr6FTDd!cMl#bVDVl3
zRs}t6o8rH1Wq*DZb0@Aac8NRvLSKCYl=1B-`$MAuJ89roJ4|2_F!8u9V#gcMM382G
zivnlS;+E;5T#{vmlXO=X+u&yS4yMG~!ENTX3*M^ZoT?yke2YHa^-|!J51Xq$ef{aY
z3Ow3`hi2oeM55;lNGVLU|Fl~HlGgC{n<i<Bo5KQ9Kl#Iq6)6lXdYfHGrE%DLj602}
zcKO1YG{&6&@)Cif%{-I6CT`;awd$%vq<t_QxFs51&LinDJFfr$#yC>h4Qt>DM|PRc
zG8<M{^d(+c6Q%6Q4ai$A1}yqi)N52V^EF<nERh)&eNCm!_JrlI6UMQy77B`uG?^aq
ziDhxA>R<d01z-QOKkM4@<x2c^MZ9Vnkf<H=>;5WS8(ym7@3uoZW&FSD-u<;+e<s)}
zna(xC;7(d*|6SvfP6aqihIMHV{d{aFU`21gw!q~6u+mWN-A7CBjFJ_TU1(~*=l)av
zN~tJ*4XQBHua#)dNo$f+e(;Yk;YeJRB|=Fz4^HI<|AMFN6GV?_&h+4M$;2d3@YS%A
z{>nq9Rk|g8l`bQ(Jy?hbc81|6oK<C*g$CoGe0SV2n*yP8Thk81u0`Md?ec5vOy-JT
zK;YFs9DRBad~shYLJ=0TaI7CYd?*(z2;zR)miR1BHmf4f&P>M#@7#?K-@Rj@3ll7B
zTP-~tUBjJDc~lkYdv?kKF^<N{6A;8^#N_yd?|I;EEOtJvt#4QW=7mQHfs~)QFh?|7
z*vSs0!B|>ajGH%aIwmh(zKqSyE%R)Y2qq|1YJ80Kt)<!c<8Qwg@7xotyuVo6h$k;M
zVsrN>4vv)`;(Lh)tPXczus|vWhM86fVwe=;P9);Y&L}R1hi{y2PR8Weu+nKD=I5tl
zZhk)c#i#mSDTF#qR<l4(dX9bD<@g^@^l+clV2qEAd-4=PDU;gVw?R5%hj6ShIx4=r
z$E(bDxW8@xQ02APlcNZM;ML6A37>_%GIKoRRC7G$rW!HWb7sL2#~rb18$niM0-@1>
z1rr<FOu)6gK&&Z0w{I-P%v4i&vi*V2Lx{J!z8Smwtilh+g)m1r5^!?d0!FW*SKhFK
z8y=U*(@R8@p1{?ZKxG_W5`O&REb9n@Jj#NP)zf%t|6pGaCnTKuF0J|5nHV2$nCCU0
zEO>UF!QYZT>^gDP5sl={&=9+FIC_wKb~({N{Ikbihn~o#A7{*9{+Ns{Z4~(kzs)Wx
z<_XwzaB+GVM|(Tb+Sw4Ft2%Xl950?diRaIs$Kxl@VrptGmag4W+Hl9=UfjKXD?a@2
zL(##VZUa89jq%8k%FSq_sWK+I5!eih#y$~!eyTFQ5np`qMLd4|BzAW9H4N@iUWbpu
z-&S5zKP==BsUt(fKJM%0&D(ycE6hh<YV1sz98|pbb~od2e_MHYGp^sbt~eYA_bCmI
zO~>;3UTp3hL`&nDmdfD1$^w(rOA9lCpO2|=ji-9p(ZX&Ck*ju_F$n^>=c+4DS2kT<
z$j{l?>9~1)E=H74+$YGrW6g;%wVepP_bl+H?o*bT%p@H=)+Ij_AM!*s)bJ?!74G-)
zM6~P>xR4)=TM&jHF}_w@XhUeLDfhG)9CeF8c>i!;W$DE2W8%!V0zG&TW+S+u?9umt
zBilo_6!O<;mp#s+Ur}B#Dem&7HUxo!$B43{EN44Q+fCgof-ipXDZ-Cm6zzy{%Z_^N
z{)I=1e9L1*Q}Bf~uz*7|!Y6n)+^E~sFZ_6Mr_vC?8PNrnF`aqDq7Yj55O;lRnz%mm
z6E56nL!CZyN621}yR8ZVMz(uupX86Wj9opnYd+;hap$-=$|85$c`QCM>NbWvXO0AH
z<fkX;S;VR~0|5<lE1ZVH@rLZ&B0lW=B4rTFTw9oqvBr=EL>!&Fyt=LOr17uZ!N~#)
z@1$i>sw{jApV|O~Az7C*-sA&ih>3V#U2uZHi+D1oR-U=9RZBry3lD88ZIcCFiXRVG
zTv+I*^gcgFSSdK+9T^^ToH?qHJ`#t0j%F22zI)yaa*3<zfzkt7*kyy@i10Yxk@k;x
zkhatj6blJRGxD4rt=@5`IMC)&Vrk#lmC6n_guB#J=DrYWv0F!|?T5NYyH6n@FKHvk
zhWnLALoq!G?W2mXXjWUWwX>(TZqMbM{PgjGP7B&l(nIA##fVnjNBQc<u44q+%WGS)
zy1MD`*zrc%v+x%|4soGxY>o}afPD9=U*@>J-JQ*tndT0~hT^WeFPzWShac_lt1TIi
zjSb;kSq<)iV_J&2v!{<Adtt|a@4xdO#qFCnR347wmtQ}OUw-u<R_Nmtj}zr*Lv6#@
z*o5hz4PhtVp~~&{mfA<P*XG&CY<=lRZ(hIdJE`b%9zJ*w4<0>qo*<N+m{eb>eg<KC
zQ+YkAd?Am|RF0;b&G`8Dz8#Y@GwSozPb<yfr+mj7m09^idvL^}7kC|+$J95oQ}67o
z9KnlVP300n-_G_HtFgq#oyXMlw9;PbFWwBn<-^AhV|Dq3a74WK&i(l4<4@EdOvRYe
zk2yE)`D*R7T&7Q9AVdpCEYeP@v-j>~3xMTLn??Cn>cqGgG$kLwBAq;y>d^E&ee%S2
zuRVJDT;?^mUsH-JyM3=;TZpFoEzL~EoM_<q#tD^o_`+w;U&J%@m2-1*-pxGJ7#F{z
z@_yc99%8q380%ZRv9i7$+v-zPU0k22iY&~bUikh|#fwF3@KP<s??gOM-^GcKVstrI
zzsv$!$|8-Y)7)BGkzz)ta-834pJv+PHjKirs^c%eDD?LIikrW<*_=;xLv`!j_IIiF
zyq%PyVRhB-RWs^MX8dN~ch#+`|G(htb>(5{TNL@CYWtdM3%*R2{@?(PhxVk6)p|iy
z069Wyf@GvJ0NGdr>!7bZ*aVF9Ze|NgD>=FoH&8M!wzfTf(%>QiUc*UmhjjC&y4!%B
zKkym8Au>W~hqibD1#!R$keu8Syv$hg0VWx(GJLM6=xFT^H>`<YGCcNJ14)+>o~gxD
z;0;ya>AKU4X-JoJli469UOIhh*d>Ux_b0<kI{<;h-c-m?LN_gG1{ZJzgOuFb{)=rf
zSIYb~5jI=_t;%7TKnwyAjD5>KsYx7k5clri_jckui<W<7JeBThR1xHWz;QOBqy<<y
zK^k1_EU1tN3&74{MARXRF|oi)j5(`Sxfcp%C>gs$cV1=&#TbBtv)?%*m;pNs0(1b+
zV0nve!Y^LNqlb?*kktUzr9#xw=$ovh(-~8u?yu|^#5qj>gm#&&BM*cJ{coj|TuwJC
zy^IK{2%0eYL6!#tvNXf!pX?wRK%ycX2$Jl^z(7{1l75(`R9XHhk{rq4ydqw*BOLMz
zbeIRk_zX1~KF*LM1=;O2#Gt=179*nwz=k!DL|_M9r_LWHE*KSh<u6RD@`n{W2*qH0
zEsRnCbmSgzDEMA67-*cwVtQcWSaE!IoC6orC7hIvI-nxHy<+~j<IB6hxc6axE@tMY
z#S{;VqRz@OIwkxtxPXbJ%pfrJQJ12DyB~b#m%=8G*_uxpdBvmR1Pw5_q#%q7igfef
zNbvaeM4tRWA9?20c`(c5o5C<06@vJX-cEZ4RKTfDxPw%hcu=JHNx|@X@T9QL7%;f9
z5C$v1*p((mdUP`G-@YGH6HTGiK*|MzM{fnwuIqY{j@hPD2MtmF0Ax9{N=JX(bfN4h
zm}-$mxfeVB-C_AxaOg?OrDU04S$FK3-V@OJLSMn{I)J&@r(WP_+PLJ)X8V*XIIJsf
z1s>>Xk1nlKS3qp(q?9TjLuY-j(DWvq{>4GJ(*sQTscm-kWqZG+iiVG4>BNCveQR1v
z;S=ThQ1Mjobb4cYbYhggY~)i=U`dI!!QtYX%C^NR<CghY>2x@>n})P^X$u1DF$zw0
z9qW<`Z3}&i7G-S$=q20=$L*Z$Gd9-h0avXPX%BAvWFB!0GJvF|ee##N290u`1Tq^s
z?hd{<CC$8{C_PskOiCES>7j42CtmhZEF6rDS?Z0P#{s1(KBT}^->GnMNdu3~sVZbK
zIGj|{IH0=xrO*aUmziB|#-e;ISfsK{`&eJx1($Bp*iZU3wft+KPIt^{Bx+7yDg2}R
zT?)Sb^FRMTIq(#2Tve5>LTcNLN_}zfvPaomtlsEW6ng*ozG83xB|*B~t@pQ-kZMnV
zp+xl;vp>~eEs+WD#3kRb#7~7a4mnX@WNDw)r@wYs`r5vPNBL^wfUiift~Bj(=Pd(2
zCQ40@!c5#~AdLZYcl^@7VTnuemx88lIKCC9%CzAM4Fz7<ji2~qmx+;h;2k>QB|e;M
zf1Y;ux5@wf3f_J&Vy^MVXrL}dI7y@ZcB!N_?W(_|xzzGY_<=t0(Iet(M#O7z2iHPe
zzj;08#nUhWw!O70KDHG*t(JM<Gw}onK<?hU8SlMwE3Pffh>z&?{aM?s!`R&3kF6c<
zIm$^;+9>W>^$~7NMjq{(4`Y&K@!GZE$U`QAxEp0_Ys;r1z=y-%n0_I#pj7GKm>7%u
zcW=in(YL+Ril<MWTPOz*0t|R9j)9sGeILJfD?a()z6EnT!ujRH=ke8(<yhO=_v9kH
z6}(Ga<{eq*mmE!)d=qRD6X4^M<KjCTrf0vkFJAK4lfq0WG6^^}Ip*DK9H-ge$09I#
zwd6ygmzDKQ&g~za$KeT!pt#qj->dc&yvs9_WWMJ~v>^CmC&k3*P&9_%Z#jmEg?Hhj
zsaT;cu1@@99<k2CFeWOw!$wS-Cz;2G*u_BK(Gwh72wzLtWu-h5$}iX03ASTClga68
zi*s>PPh)hz6RRBA1;6Iw64$wJNBpSxV)AjAiENJe9_cp^&Q3n?_=rP&Y1i4k0}l91
z%DVaBJ|^T{F_b0<Bj8mpEEq%p#3Di9Ir@sDBuNK^Q%qLZoiU^>wUtRS>I=dqR;?3P
zc118bocEUWdlJ)^>#JN<SSHG$ncV>F@?tWY$?3hl-PqmT(Rg9kE9pH^$)loiads5H
z{QR@{>CZlOIi=pf7c+^Cps6`E9nI!ceDJ{sKEjiJuUB|A3{<|d;0?jxVZ40tG#)+r
zO3&Bv_19mO74&j9MioKoq~b@~O-;|o;=*;|za4XnH)Fgx>*K8`lT1{ygA1PCcd`u0
z@|065_7#V%_~p-k7E?3R@tr^Z#~PzFW4Cn@t-~|bOVuycO`09<?^{}!iphr3^*E1U
z9G{r7;E>&E$ETOEx_uDO*LL*XSH7OQ+_A%}+31faU;Qi=nq%?pPd*mO1F^EX9}l0e
z#p>ps@A#p8q21t^El$p$d@`nB5}!JU5R#o^oYKJpMe-H_6X{CXKxn~a><Qx=;Uaw~
z6LYh(G1Na8I~;|%yY2C}+iSV=qO^g7aSS`#sN*TfKzOD+Q#r~n@F`fJ-S(t#8B0;d
zy(4l^bxq?jgb^GIrz&l_yy%s=0mU!d=X_8(u2j29TZNtC=<%Av6?JmMFLMQuK%HRc
z5TJ=iUW_AmRIwun8fXLAWykK16wv4KkHnL<iZbVL6$T3!5CoCVBnt7$qs|n*@(@3?
zBa9&(AZa7N;LI^03l1`TPFrCj8bK`MW8#kxPO)?N>^|k#!r7@Q*H^f=gZ(3wQRO=Y
zP5Dlpr_N1`4SA;!^)CBGCcJZ(EPW(96|c`vsV$vQJsxm-gJ5QD`_M;ql9#k&zQa#p
zct=yAoxTTw%b>~-=>_fBGp=X6P2Wk|K~j;<<Q?S_61mTe{w#Or&~9fr@D4NSQ|OoT
zKj(61Q3XfVA_!PmT!`7`xKf|vRS&&$m+%p$a3|^U;l62OCl6%^H}Z-S%|a0b{yWOo
z-Q7L!8qM^jJ<UA!?lbjglpn`S4`1h@@|k0QXQswtakddnwaFX-m;Gf5vRDk^en4?%
zcOQD%M~<BY_twUC?CtL<?MGs1ab9h}gkZR*l6*WAjQSMS)eG6FT6@<H$Juc_9Xo=^
zAA+O2P_HN(%%6~UltF}Tvs2?1mRz2nsC{jzZJUVux2~(*QX8qVxwXR)cl*LI;Bv-d
z<&l2M<$;ezeDvTeR;I)|_wV^me|8Cd@#Vw#H-Ghm`1;w4*iyevd1KcoZ4pJ9MLitv
ze6F-+AwVFwRH~y?MU-dFxO?Yz%+F58u+nmUby@j;rnYt|Zr!=7_Iz4xk8;Tc<|5XY
zm*e@fCvorAjd)kQ2g0RI<=gO>@?PzcYK;3c>N1P5Er3)Tia>{Y*i;|O@sGReXZBef
zp*C-2WjP)`d=R^8%jTOCF{e6o_x`)8zv}dz2bZzGzZ1I~tD^BTKK}UQSh{g1PL;Q$
zLI2R0%FAVNFa4?d7k0KW?=v^W&hJU{DjVwKDYw*P;z^s!Vu^;@Qu^uj&8=7wUuS{T
z>SjEA{5+n#c%eQc7mKr(^22*K<2%3iN!(ZZFD}f+=x|?5EB+kIoME%b;UKm)H)B(L
z*woxYEUMq?85|a!)VU$I(Zt!uo5CZ)4{?va(&_w+R#fdi<x4ygb(}Vty2pu4=SK*>
zjzqWG9LlodKprxlbz6>sw#}C0FJ`{oZnTBThoG`8PCHcfgX;J%-=cpFhQ~A%eYG#D
zTaW4%P(e?YMk#y03j3Ry-=t{&-mI?1rCa?vcu{}fa8&_~pKtQ@O*oyV-{iTS=x<it
zWuhtsUwz;G?w@|2&fFPpYw43(eXL2s3-JuCQeUVk_V%Q|(AXV41%nO&BNlwD$GQ5e
z7DeC2U97Da(!?t}G<1v@3aPxYCLTm3eI=c!nl{Fax3n(!Qg0kK$dZ4-Qv<=x$s|MK
z;SDa{0>MwW*#u+8E%ogHWr+@DJa$gIc4gvYH!&I*Y)&}d8DBlv6c|X(x?w;o<8=V#
zE9Hs8;-`Y!Jxxh8N}uY-UeYSxvZ!<c;zusffDV1W)KKzCL%I`x=b?E|6FoXM=5y+~
z>=p)&a2>DGz~LPXbb|QHSNRBE;w6`uxJYKtjGsCVf7S4M)s`MJJZgktC|F^}>Q(H{
zWQTB!K^1rj9|jJF&rFNlS^dPIe0_CY%=)ewf-Dn)Z+}fvmohTz%*fg%nWgk4ssXY2
z??DL-<)C8AFW8l1Y6#d7hFu+iWZLm6s^krJnWCxBvXExuk{LZVxM<j4V6bEI6M+ts
zFr=gWz%YwhWThX2ZSHd8a&AvBs3Ujh$PomIR7~)tI$<%XOw7QXlLum$%*@*zMqHgI
zAqh8Bcv6aol4V+0ok*vPA3?~@>Fh8Br>FW+2?5;O`v*1nt}AiLf0!r=1}Ggz(SlGE
zffxgMXl6hGBMWVyf-z^SK6HBEAO?W+U~rDW2{gV~TMP_mQ?nf(OX3NA)C24obg`0>
zy2Rj*x??+)W$4WnaSWE^O*9f$5Rlh3KS*0^^>SRD7Qp0+I?`mYe<&sz8l-4&<~m4z
z;>CrTGLg6(KdEf1($G#<DPi@OPx=9qjdBxwTFJmGwW$arzx-V%ti_V2D1xMn7Ol&R
z01KJ6M`&kjJ#Cf9MN7KaBs+f8%QkG6o|Q-yFRD8&GH&3i&grj1cPjS@B$_SLw)@@r
z6x(99YU`yz&&PIcha{n-2Z@m`PU)+v?Ki!=N(;uCk4}#Qd)1Q^Mjpx#8vg$1=Xhq=
zP9OcP*V94RQ7&Qp@upt~RyrqSnt2>E|C2sv1wQO(gG>d3+-*a8U`&%;s~<vh$*f)q
z4#S()^h-SLPlf{Omua>EGcbZLG~22=Rx4h3Di*f8qKb|D8Fyl|Ifb^VN{?6+F{^Rq
zDKw`|`(vEa8>7AnB2&@Vyh_iGt<9JSkMNU%^6=M_kZI-x-6$s|*3x8|LXp~zSUT#d
z{%z9B^ne5h{OKXwf;JU#)bEEs{>eWtHvit$AARqie!rUtPlECFJ5|~h``7dx)ouP&
zZkg(eKl<Y4ZvlH%-woqmm%07rpT;gY_?U#Ru=I;o#SbqgPz)ot3S0FFO5gGdE<NO6
z>eaWQuYP!Bu6{e;bW2^@qY6m*!egK>xPUZ%J>B%l?rIul9IV<f)qj1GPH9r_r@-lP
zi9}!g_^U9nuQZ|fcFZ0Be6Z)S!R}JB0H!|GPi;%mu3Hr|Fb<ozDt_WuF2&0-!FP7X
z(a<NJJY<0xcWp4Kn5kEJ4)HMJ1DISJ6W;@$!9<rB3%kQhoWcJdGRedJFYw8PoprQ5
z@Ry#I)nnr4St-v>9|UUn7t3NgeR_lAB)&zw4@cE8!N%QO-1)>LAZ3q<>hXz&7ZWkr
z!~`OIIsy;wdE>Ht_?~^yw7btyQ1#gITzFKEqoz?o;KfRAXci~vNya(}&2ii;w&N}v
zR=N|uC^QZxcXCGplYfv84>lA-!=o{(xN)Zm^s|GGqqmq?q?LoOBYBtz!fk7N+dI8j
z)RhyQ<32i%Bkq`>MNq~rFgUj{;hkzuc{2M{VV^4xeXoglGL8XbGB_uUaz_CIyHSK=
zjZv@q_7QUfxoe6Yb3VmF=|J1cPB`u?V}e;Z;+5hHbJI$DaIlkr2{p!SDnpbBXMgVI
zK=4I4`EZ8_7w8ZAo*H%_42UM~i{TEo1}7*C<sD*!{Rqj(d+x85UU|v!itHR29~ZsC
z>-$MWGXgB1R6seRcq$Ruy+FJvM@;l0$bvQmW4%2}Z;p05J5YH)ip$f3c=GU<@##-~
z7+-w$X}oyyFwTz;O%p;Tj;P{(7bdwE7nkC_ciz=^!IP$3=s!C<6E9ypkDvei)8L*X
z)ZSjp_w#OVZN&QW%UD@{7N5!e-~O9_9slmH{%w5q<(C%jlBP_QQg?iW+C(#M+<7N%
z-g!Un-up1#dH-8+{nq^$9GQrt(=7k&#2FlEC~w$7*01y!6RpXAv$G8Q<<*sV@Zh1#
z5Q_v67*QuY5w5h$<Fshs80VbC*2b!6eChGWgz$~Yk0)@2|5EjZU5b=Tj-_KFwqH!{
z?#9d5-(HW!g}ir--7f4hIXF~3SBa+1`Q9Y`Xk|>F#xY*aukSGBMrDyZ_A>9tGujaD
zz7bt2e;$>nPGq|gio>w)L>gCngCL6?VGMoRV}5}FYGJGBZjV*6P`XVaeCm(|2MRBR
zLr|%o;Q-C?QI3&vOav~Euk;Xz7lA7M80XQ>F@c|Y${HyE&BQ$Md3*$2!iA+>^-2N8
zklR<<Hrhhyq@8tmvO^d_8%!Po&sdP1o22jHAo<B1bfaox@~BAkw1s(plfJS81|bz+
zg!q&rb|oQfWLFF3iSY@yzu-X8hKXS7Nh?k{;dsHMcStD%4A%o3TIfUD=iPObE7cp?
z-@3EuNNpPRki`z{b|KH%iQ*kkq@}P?mq)#eHU&5w4M$n?{YMC9)%Hwk>_6QckD1AF
zl^w>Q;quQNMGfxU5}c1>rA-k$I})b_+nj7D4T&R9BO2;&m`fm?Xg{)Skmt&8JuLbp
zt+Aucr`(W#2%x;kQ+0)W&9<&z_|(1&CvcRF?0Z-g&CWQ&lAG|gC;Da2(QAZ3U7@{(
zA2?Q-X3mHse@Qdu0W55lzbfQqCljaC_+BmLuZ0@QGw_QrrLFqYQTl{Q#T5Y@eM|B%
z)GdyhrH>)+DC1*l=h%UT@M}Wt^MKOcyI!|9lm}|_InGz@^ay>P(v~!U4(Mj*@3r}v
zxV|_S3)4zx<=^O_`aHMcS%yg?jsoVMN{-uYs62ZY;E>9;@_I_;aYnR~hR{kIPri*v
z&k=(cqIY+D$3j9M39SBzbR*wrcj&kK6h8#>{R)dkhcyrB`<P>4ZiXFUl<_gq;T>0M
z&tATKDcqax8x|Jk)t=AE-*D_6o_I&w)0b=U=*jc=?6Y6SvS9Q0#M4;cSdX9o;<I@C
z<Viev_Ck18yj$vUZ#Rzjchxour^X%7M0><eJ$4UoY;JmY`=0uTz3r_yIoyxAsYx#&
z;Mxcg9$UM+K0%7(FYl|L`S`<k<K26=y*mwo8|7nfXE)2pF$*BfD0xTa1UoEM&pF1J
zyC<QCU6%BN+!x1!6$H2I8{6^MKm1AjH~;Pj@#CL<8o&JI*S72J7hlkvi}`EU<Nm$-
z@dw}fIR4}hf8TUWaSX8@>e%tYL9DH9#Dhmq;`!=E?Cz;A?H!2Og=;atcuj53fXX=w
zztwIl4{6sa2g)Y5irEfl+mZrH+HsXn)gAL;lubdx3k>zqI<QbcJP6IF@Sw`a8vn4k
z;$G~kTd}L(sz<#^-~JZ;YcMv~aId<fuE6tjn~NSE>s5bNuiCxA{03iF)En_yy-sUY
zZ})Rm{k_AL2Zf(+^7Simexv7ae%^%pU;n@V@BcgnUxttkJ)bm{OcRW@c&2l+Q9oyJ
z`DrYmcRs+j{lN`jiy{9-kGnbxEUfQ5nA3~Q_^;;TS6vgjREE=+N>=PW$>@oZ)boQ5
z@LAx4o}WDei#=a!i<>>@DU6PO*2%8zhP}UX@i$>Q&iLYkA0BD5R-`*xep&PN@6c>V
z-0TG20$bcmNX1`ukut9IQzx7ZZy2ef$3R}zmGZb^?QryJr-y8^L(fRHC*Pzu2rnxe
z+UA7OAHVzzY>*7Y`izBHBy~Qls9}{Iov9fXxg#in7VGUve|3WV7@El_gkUkz7yvLs
zfZhTP58(6_<9#MZqdp9r%#iDAVT$~i*@YPqY%k25I$M}tgj=~qo-OSsvA(_$E30c}
zWC#jM_)!j1K}*t|k(CjW&=g}P2K3pdJN6vB81T1J@cWqpNgJWq9#g5Sexs374H1$&
z$lxbK^_MsRs9u0zr5P(?261sZUg9WK(6)DhF|m+(F`Z(xVJO+@RR?V_@T3)kBl5xN
zt}g?2Ru95}v*Qb4&zbDVAHv8BZ8NVZGa0hNmS5=SxImaX^3^-}&gGtemvhR82VsJx
z>e!7z4QD2cChq(~@I<~-?rJ%7nSfR^C<w3QpLmjo@^2v%spPP7l^PQ}SmZ#5f&#wu
z<5DVoGC;hUY+xBcBP2uBu$YV@ZCN$ui8jSAkHN@X$<^@+U$}@n(Q<lNCq3@qBHGvq
zuAo*}gC0C^?*5R{u9KmTNKtsq^mQWoO9{y`l3+3wFC=eTrbYKIGJo2SUVQCt;yMa9
zW6!sZd<*V0XADh8mwg95-76j}iG%$Bt1m&@jC<9kp_tN~aZiWTlkW+XK^H@LJN2cp
zjR4TM;cT{xRiUofrF+}1{UxozTJT6?x3)p<`j9Kd#dhdwdy2hn^snA34Ej!+ktzjR
zX!4T+LBSiKLP?jNJ2q9{0g)B0hg6pxsyiJ@@CoBbFTRSH`y;GeR;<%g>;N+jnLag+
zIPfW!1yA})cs<l(+)x)vLq{@=#kbx{nPO*US{xhOCvAcSB{US=iC1RAwXP^>+8C<e
z`YtpZUQVDwIoftBRN+q=KACUXzEo9B_GZB1>NGHhbd+-3D((Vjocfa-D2G#!k{-pY
z6-^uP?ISNiQUPqlQ0=hDG-`wMu`@4Oq#@DhLE$Es_|flD@b%q)`hD`#Pq9N`F1mcT
z&8e#SioLWoO{?mLt3SuHn5$dSUcuM)TYuHu{pJ6PJ^d64Gkr$YSJlxcPE58d{YCog
zdMcbAyQ_b~uQX}a*Sr{3-EgmbEjX(%cJLpIgz+Ps`npKeuY*bZtGK{}9RTT;cns2p
zt@d63qi@#r^k0Fi-h6}Ls$bxfCjKRibV>V?m&MiJ%wOEnwkrGWqV<ZBvHG{MxW5Lc
z1j&V7@J$GA5LTV29byO4m?t`&YD~JS(wY|*4+8&yVn+uUshjW@+^aMyUd?wkoSu27
z7dyTPhzVIHz_^qh!P*4(RxlAJo_<1n9+R(3jyvxiErqNG5uT8p{}dA@RB>Z>6BFW|
zgcNUrU+(upxC4Fr2jT_A2lfvPnRgszRlIm2j-X@W5kAn9@qH=iV&WMA!m#2zqOk>&
z(a-_h*l-FCyjw-Iu!7%{&iXP54Q~b~$BOvFlZ$8_p2kj#dz%h@j|~E^WATRjd;77u
zxf5HP+p)Xbiv8A}VD@5H-~EFF?|j<XX&HtU+gmbo+!q(sBV4Gv>UQ?IA1nm~#Gl1p
zyIbpV(AtS1@u5smK^yJhv3Nvwei4uHv0>lSGSh5CbAn06MvONQQVCvh<Max4?XUxl
z724*fg?pg?BIc&YV`+XWxZ8|-S(wn{$h#CwrI2N0Y{YygVUhn#f({P!rjScCHHMXb
zaTZ(aD{->F8|Nn~YX_}3-rtVPGj_D>ix<5#?+%{|UXJMEE;A-98AtHImoVO>ZkbPY
z0nCRo&+blyR{MK9KBj8FwH12^qgG$W%jb{d#j}U;?D5y}*{45=FFyY?HrG}ar@mO6
zpNYk}IUj|^r0LG~j`AA*o}<@QwnoQQ7Mb8?mr`Fmef%VT@{=FO*AKpqCr=*7g9l&5
z>dWV`xymkq$FcJASv-F9FdphzUtf3qn4epS`|rFT@4fe7+`adH+`jXEEL^`AbJy<1
zSaUIY2gcQba&+c}$`Ivc(`ALG#ygP|?m!=DGpIT+EbHDly;Rx^kBi1h`+5BMQEaTQ
zsjeJ(F(4DR2%N^oDaTo#>6Td}$8nC!FP^Ck_o%*)xE<V6nz6gHwiP5T<KmIQ{+O8<
zi<K8oz4K~vYC48S8*!@f61!N~sgOrgGS<sB$>SuC>6r9P!6Xx+-19_zN<lJu)lI?E
zzGQypSb@BzEYD0&SwIdRCcUAGa#jUgcrhO<c*Z)^Yajmy4&@m<DD)Vg#-a$E@I?DU
z{bdZqxRY_0@fBKp+?@P)JSMo>7JFAJJ51%y7>hBcj}A2bLhLv|4ln637FrQ*@uh7Y
z8POP5X~;c62(`R3Qhqq9k?;^svEz>g0tnBE3yUOqh(i^?r@)VPNMk;YV+nRxV`v0D
z2t&9tn|2mh7Ca*Ox3Eg|vKx}IH9`W)0AaT4vGgn)U^1TFA_&w-19na^=HBDDQng{e
zk4SZnHlM}9)ZttZaOkq*#hog<m&!VJ_8~CaQyYnZhea~bMfu>4yRW|drFTi5b2`eU
z(pYp3s|{gtzd1H!;p?-<U&Ul&D6TKe#*M}Kn3)=PTSi+ktairt8L53jFvgehv4vRV
zfzpY&9Oey<)Ss{r0wG==QCoKQiRPN`PFre>vr_r5_B?mHa2zLPoVf!Q3UD$Icj_If
zJ<mI{*a^Gm9nE{OwY^(*xGF5gTNO|=rBE&DqAz8xOX0{a1>@>tM0XY1B8(#s$a~^(
zcyt!q!ne7rhvP;MbH^loA4gLjpIxdwSD6(37lffQG*ARlW6Dz&Pk_ru*DBwR)Ni`K
zP`HfqWo^M4X-Dk6AVF>Wn9_AZ{RT&-&NTXCT-B*L+HV0G`O;LoJ)(L}Iix;N=iOfn
z^vGRx^ZY1Un`?2@+K!{W?RfI&t5{Zhxwg8hw)Bbmvuz*ccdT;On3z@@JQI`6>6m7h
zt&bz*zVYE$n46FLckji!_wUE;TX$mV`n9-q^Jd(;dpGXiyBF`!*1r2*ysw8v3o|pb
zF*?=|uBo`Ov=n#bXJ&RGcJ_|r$;wtdez_UXS9fDoX-PkOcygw?Gwvf0NA;x-;I35e
z;o?3tj&~)`MpREOm8S<gn=zm=%kHXEm0zCq_0@R%^tq2q9+LgE`pIdP)w{QD#5?!j
z^=>wHZEtR_$CJnO8_Thy{)&b6YB|-eTskcd_xF@PJF&j<GM+zu5}O+<USxp)kA7}<
zXWQ3%Ak?Qn<_Ja>+3d<~S?T)e=U>NP|K#WKmp}e_JX_gN8V<$c;<dOh8gATD+RQJ-
zlFaYkyBi<9|Blk;x)+<X2tvRv`v@WTlrEd%M=k~%>Z&H==v-8)t?5^}934>^V4;iR
zh`^pQO**yY=U8c`;^6Y0yK41Soo68o+%)CHi?P*?ve>5Q@-%!GsoIB5+sT*uL2(pC
zJYeOIgRRpZ%Wk{+t$Ng(^zCoazXoG-v46!K)oso>l#Xh<+3Qs|RJYyR++Vl7s@~);
z)$6oY^>#m3)!!!^Fx}6uRKOMOtEz4P>yh`GdQ<oazIy&w|Lo7jJ9If(O9wAa(KA_-
zLJ$2)#}5x+XY>Z6qp(eyrHc3Rk_ihuRz3E{Eo1S79&4krp30MqjTsMp(+PVNZ<&N3
z8Iyu69cApI*%n4oR9hoeJ?O!@7&{=+SWkNBZDY~`!Xzy{dr+(=!>M%Qp0p4jd?s|h
zJ2cA%f6~(k7!G|p+1~!8$38Xr;hQfRY)%Km7i&UBRcGsMFJ<rq#82^1d=!O2Fcn{V
zl)3_x)&?zaclvths;ez$p>vKjjMMDwY}~$m%S<<`o=7VTVT@Pw2(C_Iq!0icGRvOO
zEabu+a0dKh4o=0iTIit;|AdQm**OIB0CTH{dzTQ1Fo=LDIxIpKn8rRa@-ROLw{Sxk
z!P(N=J3H~<s|WG)=`%4?XGxoPn02RYM?q<pOKFOeQ&(o%c*rM?xeAH%qH|09Z>0za
z<W(sQI!y)utiZsXvel=b@19b4FhUIAfF&-Fo0O&yC(Ss6TyGH#Lp`^!5ZfBWZEHY}
zlULjU!pmWbp%Qc3rw)<4;JCy2rNwB@&S*eA7Q-;K>PYkKc9=%dB7Yofc%aUnt<fik
z2XW3s3wa=h8}u-cXBuFW_Aml;xGoW*9cBY2oemXxS*3g|X7yAZ^yvv5uVUbMGTB4h
zqE1(gTMr$!29TV$JiRy<*KS_dz<^7uVTiNLPzK2_3j&1wOyQgyv(op}`NyDyvcge~
z<RkIN&$*aX3kNx0K>kQ;;$bYJnaMWOuW}1cGr<}#5q7RTlwaY2S>?-djkpoVdfbdB
zu2j~^Ef4rquV^>3w&crsi$9FSTW}E3*E1ldS9vo#A4~Js#rRIT@VZUUs1j*?GA!O$
z@rLd!A*oB(;%*ufN4p%6hM2I&1~<bLcY_x<`@#u7Y45M`N@I6$$ex*b(d_mUIKdcL
z3T?q|E5hrTQs+-K7JAxl-C%@;0T8_^rWzbxJ$Yljfvl3wVu#uK;waU|#a}QbW`g70
z6^`H7R6hxmXp67(lJ*tO0VWKpfOKH=YKJO2;+d7D0Ft))74qAv8@H*k-!^u`)6S)9
zvB$|i4W4#>Qx{}$6)ek20;Xl94do0kxT?oOkK<76l1GphWvgl~F#4b~t$Mg)D6JDW
z%1&)0)n8K0{|ca0g~4CafC2?NRx=tq?zWN1Ql(T^)1%<>zoe!8C1gg1I3*ldfJ=Aj
zY+Z2po$<568YS~p+<s-+b%9^efYbksLDhj0GY{pXlWwTQOBy32%b`<s@4z{^3<@wd
zHH_cwQHTH2|8?H+MI358uDq|rtz*7oFKs6tueeFkPoQ4)+lJGp+mGSfP94}Q{`hyB
zrQQ?{{b*ImQ>A^o^mM~^xr57Y{+F=%1Bc!Q!b^W?*X^EqRGY3Y{OIae8vUsaqDvN(
zUzZQt=R+l|`1p3>TI~hbg};7rw_g4$|8BYsn@@s+v%;4!i#_~=bp`~^CM@#H+wdK`
z4h<^ilyf_E;jXG^D{ep3NqeaF>EfDH;K)Ddg%*xIoM_I+t$QED%-preqkG_!dV=E$
zne=9X5xi8hIpK#LL`;$}*}+6Myb>#%S^TxVv*+Wfdc>!UHrQRj3hXiO1fuQn1n<yb
zaC9Af+uGWCtgf!d?oP|QwBXa2^rdYl?O5s0N^@4LbEFz@_??|@nrB~ITh+5E%n^%=
z^KtL?b-9h3=j6ygj%ng3KPD$Rl86i8eLR<_@otnsc4y_THPpo9cuY1YWnYd(6mJa8
zgG@w<cPeZZEU8lx{*9IOt-U=GSbS^lR$}!&LM|?lhfa1hk(cmSeFOc{fS-F>5O^KP
z>xBg{gIr>-^kH`p6VfRR)L5Xe*E=JW8R9MXV^E2?v>?80aUq&h6Xx^QHa0yOig1sW
z_S4EI*mrhZz+1u-pPuDTgZ0fFrw4gS*(dE8XE15KFgq2C^K;TGey2=EWn3-zq#x=-
za2ByI{(IkJp*&(rUX@1uDi_1vQFNs8!{i=E-NAKpOw*C#I@xS0y*M7M&w{t{iAm+#
z-B_5L3$fmDC|;KdSp;(IE;{uwNsJ@dWyHjG<^$!~M{5c{X^k+^w5Z*J_ojRuAMNQm
zjD!6>3sN~Ii!wQ-^0PQUAJ?xfS%7#Zy1a@*`2<%<9bht&NygUBX8ibXe-K}N^@YbP
z97Bgt^plT1ia+_|KaPg@Vg!lA>GLnXj29~_F~7JNiwoCOo~D%U98=jB9D%uwkW}$H
zAx;-e1ooGC40dT8YD{qIyIXGDyb<&Bvnosdo>-(F!0)r_o*fa4PdOs&*^>tvXRWK+
zLv_|`cKa+{yAd~U-BbSNoy6Rk^76^U`1D7AqjWfo8@F$&E;lWh-Pu2kQPDX(I;pf(
z-5zUL@GvpfA3y!uzlvXc`s4WI+uw>m`G@}`PBkWYvCQ#j8sAX<$uE@wCSAE7Dsw6u
zEq3KGA<jfGZEkHtfuO!|)G2oaI;|Pc=s}QsTXDZ}ZP975wzj3Pc2(w6NO8j0T;p|*
zn^Ybt>lS7!-zX2X8`Nj+PDU{a&$#JK{_?4d69p@H^tGLYMg$MQBS6QEvH!urp2FtJ
z1KI<GUTPDl%jU@t@*n_AVTH$VqJeQW0u)dKPy0DBG47pn-Z4ZORb6Ih2f`TgjxZ<X
z-e`<x+i;+^K@a5?RR#2n&ud&MH?%9XWuq)k6b=iW1Wy{b5W1`E(>|yL5*)j7xcess
zWV8vC<=UmC;6sgu6ltD!(rsIK#c_a~oI_eOH^GiUV0~Pf^2YsyymGYM&Fc#>H92YA
zl#R`;U6)%L2`_$7`8q#U`@j6mJ7Ty~i6d_hRaa>TSlB=s|MVw6iBEt2i}<5I{s-}o
z{?T`1VR6ZAA`|Q!yUG!YEPO!l23^FR-CE=@b1}5T93Q!(bl6r~yw5J+!`y}IK2?c_
zP>IF5-fc(TsB94@<_$<&$}?q|yT0b<W-Q1hzAo>j3{iGTKm=oyBNl-$zPEQ36OrZg
z@?GPyrM>X3FO@GZKA~+TB<E?CP4Aj?SsZbnL0O_N@ySG@aa8yw$A@Ennw?*{<Filw
zKf90yl@aXjbziH%Y1R6%XMuw3DMtwET1V6=%4m;w<qWGo;Ep&1d+Y+<S073ma%>_y
zrr7Dl{eK8Bs2dz(IWVMj<H$SVM3`X#0Ocf<-xnwG^wGok@ehBf_WwZf>y78j%PI?M
zL)8`^s%%`FnTq${y&u<Z-i(>qS+}PMVd<;kC!Q~_$CmPrvQ`Ddv?aS-aiKVH=NfJ1
z`1rWnaz)PhL7CX2Ul(2MI;F2aRDbg5>9csbz8x)<o7Vnu94Rkpm8dn-lZ|-i&W*UX
zsJ2&W&-@>K97k1dZSN@0PaQ_jsrod<ZEmI+OH1==a}|HZojW3b`Sl~!m*cp((2VPI
zQ|hBmVqA6l?!EhYjN|!9Y;11Cr$75VR@c_dV{sHE!sdmAMK35~&WL%U9mSh|#S6n_
zw<kV>a_o33%@@T3%uG+)yej^Sx{i>V`?mT9l~z-;F)N%1x!5tz@rm5c%Ke|D%bd#2
zt?P3!KQp1WKpCw38CFIk&{9cOoj>&vxaWP`fvWPLG^c!zjSa=@<fwQIgy^}z;z(s?
z?+~7b9iXZ|Dgi96*_VHg?xhWazo1SWsy$g--Ebd(VEE*4PwA8kFsMV+WmFI2ZGRaj
zW5?IEs(h=tx~sYww!joTsA`A0iUZ%Ne+`C5X;$^!?x=3F%Y9oVteaJPz3LCuZTB|!
z*KMyV+QK)h*J-WlZT7FKzfZVdCobK}@lxVdKcpk-iupG|*rjSY=~h1K@O$6=-uG=H
zLLhelyh`xcd#Nvx>2waB2$5c&;)aj3Z<~snC-G!evDC>QEIgqlmH5&p;Z!VmboEi<
zq#ORbaTPPf$TEu`kgwo2bkVm9x1H_ZxX>$|5<?9Eq);FXgeAWOWjFkFg<)k!o-`ZA
z;em%gdTZ>KHc4H2HG~;~Zrn~Ap#UrHz-6a{MTTV;>}*b~3_^_EpaoysPVZum7wgM6
zJLh!gWsFzCky|DN#&j?5Vs*uoKj}bdq-u+Ske*JC!8I2ya!f}Gyp-my=!Fgu2ENqG
z4no0M;_6t~4kZRf4X=d=dc5PNtoW&n%dwAuVZc)vT!a|XmkA&XP}K3^hp_16nH*Tt
zz<p(TRf8lAU=69aY0C+jq$%TPI&3FpD`2J3r_xRSkOsunOppfO#m8@_6il0p`W5EH
zongCCP?$Bub|WFX*Sw0->W|eb#F0r57)O|H($slIT#B2qO7EeE(vuEuL<22fTFRFF
zGEjj@N!$uWFkVqCoF^Laksb(h5IS+WtanC<C?-nC8}u;nY?JruG6X)%9!Pn@N1ZYG
zOh*h}%!HL~gT4xuv?p8`_kIn=hBf#co8&V1X|JqfVub+-$71H>7V!%+r3}P5R>=N^
z6G1AI3NS{%Km++o>X9xkw+=MRy#<A$16-6p1h%9dOg->9=p*eYBMh#DPc%54#gI}v
ztS@dT$H#Vh`0BBeyX!gbg<diTEcs^AXV7tm8S8U?);t9Sm@FHFl)h4*{<C~xXS3s(
zX47De>`J?V59`usH{0l+I1?tnB}~7dw7^uqw!p7^@F~U$;QChG^h*5TkuIJ5c&AGK
zixMnmBP$QM2v7PdRs<jZ%gbRHXSFI{v>A4q6SoY)*X~=48(2B$k5$K?@CwLg{8l^6
zsZBfb-o|QSgC1U2>$lUx7R3!jvLk%QTYr}=v~tXT8cX~DtEVeHYM2g+M>tf9^vm)g
z(rmI##eph*m~`Xo^e9dyqn%Lt@W@!O0AiKtj$@LSIw|}cEU75ot-~ha5-9y=+$kS6
z;aM7+u&}CnxtT!y9fp+jXlpxvMXwLZg{?~u_OHYls`!JpgS*(t3YGq-pSG4tJQad>
z7Nx{(XY^>QS@3Rx+8ELyP56@@&qkT{e)IwpzfPpfx47|d`zv}re&LKizBW=qGK5+u
zlXz@!k<usmAZf&t{^-X){oU{Q!r;g!O|{Lfs@vSw>tml}HpVyj@vnS|FM4<tyxCs;
zXH{3=fn!pqsA^aJRlEAF{;Jtug=0NNKE{`BGT80PBU%h^=U4QF&W^nvzQue6w&<Z>
zru1vjw6}s5U<;napY*lqMvq%b2ht7rP8j-<mXx0>w0GcvK~W=JoQq%T3UGSkz4YfL
z{0dXPsQf|~!RggyUy6+J7*|)g`K?d-ugW-IHDh5qvaD3q(HGj=CQdXE4}5Z9-Zpxf
z7&`KusqEHZmlhNAOlY!jhY2t)l1IS6O7^{0PFmA;kB^P|m>iDm8XDm^E+)f<av>YM
zHawBy1}_BPjo^aGGcjML+0B?5osmYv6Y}%(b6)s1i6BZ(V|>gLdGLHp+z}oEFcOee
z?w;rF6|XlynCxB{h=~R#MsVa6!YI)rzLFL1(8ua}CIy+a;Vu`BH)1jp!5cgyM?3mh
zCh>moWVpi{c}I)%?EYY)7U2-PC@hrNZTY?_;>ZFeCbDyv%b+KFnY8nfbm$cyj*#N6
zD;9W>zDx#jG}GR0?yT5HP^9!`B8%N&2=F$xw!JdFr*FWQ?hg!&#)aa0bbRg!u^o1j
zDQ}qYJ`|2);W#@!@Ucjf4dsi{4q+A(v*dGsoCg=-_x4<f=05LA;dn_OttY#~)|MwM
znbbVH;4U!)hup)(E(MO=%3g~eUd1GiPA}p>nQ?G@5&K7H%GZ9yHS>^(|CyO-3+two
z$CHzdn3@=ix#?ys&CkSy(gnei1tBUHXJ;HO34ed-d>9_iBhEN>451m5wMTN>Ke>pb
zGxZQEzkNfE7;a3*=)|la^nJq<F*rIU{(Uiq8nZDlHXSiI9zB9R?@<{dzcv0CZA@mG
zB0w0W%!x)$93ed#<8k}Ot+;plj)ggt*@gLq_z(X0PrQnLVexugzj-TW=NIGp?fY@>
z-H$b{nU0;-vGVL7R=IO+WmDg6Pogr8RKwx6hdS@0$(Up<tM-|!ryMbkVK)Y2m=l$K
z1h@#6IR^7w{ru2~=pLWZGv#s9?(TLx{Q8&i^64Y*3}mO>u=0srRM(a60|ULj$L90T
zei2`M@yqz-FF%Vf1i!wqti0ZdjpY|{ae5fD(~~Z<?E3iPi_hc1!v`@pKOc*Wi{`~>
z3)+329u`m122w}Zal<4kMT0R40-Jh@2je>yM>2WMU206;4i2N|=*2u{k9voIeM04i
zU1i|(B2eLB;U2p}DFCKGS-*1#mk^R$_=NybWzu6k1x%ZVaDX-k8mQBz)x3~;(=vvk
zZr25-w1-{<Dh0fcWmI^mPXAG&EIbzE*#N%81x>ayeS`rmIW7bqJC8|4Gx%$~=-+8W
z*h$5>6hFYT0G06`?K5%Wo3s#uR04V~g0=#|5pxnP1XkNy{?e`>v|?8vdfF8x(Am*}
z@P+vh(;!-Do7t_JJ5&*x9h9*vf*{p$(%$hR=$s$%y<mc=R}j!Hva6H@qCRp^?H=_9
zp&UEiXcJgqal-DAfzjw07}J1#*uoa_nj;GloIuCc){c+*yMFy<%*@S-h7@YkPLN)-
zMNG0IAX#4Bh}G52Sl`<6PFn8NLeS-%S&Hi(0)xXEpS&i){W@u@p9NqDib`1|T*@U!
zDh`e0(Vz|H0rYqRUuoduRuS&%seJ_Y-~vDCMxTn{ZI^ca5aG7ssdB=OJ9grFr=H6E
zxW1EWBPZ4GvRjkBfg@zmv-3*mDTuCn?C>HjDMKTpDYWCrKz6;24)^DdE7ieK`5Wj}
zd0|JW@{+XU7{!Y|_4OBhv2%PGtGg$$vV9WUd*`vIcAMSIdxzY;cA!3I&%4ZMcep2w
zx=-K9akah5qn@+l80<NZ{)?mNKR=3IwH>X^)p+p5FT{)gIR2Zz{CDvefBnPw@h^TE
zPnTDHtl-AZZmh0v#@gnVBGeai%GZxS`8e+0e<zw{msZc<NVHBOHV!Ug?{pv(N}Tiz
z#d+_DXjdQKKjP=CN9C!PyyDcLk>I#v+L*J8UiAq|F1eoze($6w)^<;0W#=ebiZeTH
zhejG|dq&l6<nBt2@w*VNlk>jVJvfQ=9hJYm)7aiWj`iKcSW|keZnk1|Yd1FcRpv$0
zrOGeIIxZ~D$JET6@@i6j%v3Zcn_jek|Nb5I2}^39o8DQ+F~iIcv51>~oqmw(TsYF0
z-GbbiMc=-@y&KP0HeyX-!EX(yyl_k~0_Z2tm*b0vPvgn*daMGsycSPht}367)SvaM
z?`p<^>dW1CKZtif_#|%KxgQf#vvG9U=XmTKoQA55%Fm!^haYK#y3p8F`0QHu5z4B+
z^t%X_>1%1bsVwFb)lR@~9dcKfU?`ZBDeQCCqUr#1rfMo^BM|J<j=(#@@--VnF+L)m
z>f*$3EHL57Mt0HC29=iW6_ssl=PTV6^Huw+Fsa_8Z-0yaiaig}wH~^PzNl_J>Z)79
zTjBi`SkyP+LSgs1?N#+Af2lfgc~!mLPb!D|D_~H+E*#3RMgQxliuYANSIoZ&!Y=hD
zntjyaKmP7N`F^@Ffb!dV`Os4eKnazp34RHoAB=@T6P`~F4|&%gz;-D5`Vww^Ql?hm
zq|#W-1n)2kKJZvy3R;&ZYs;&UWqN^6g^lzTA4a?LO>LS92gu@<fL-o#sWkHl(y*w4
z8)Y{==|*_CRou2JE>&Sle}>bq#)EvZG6<GnwL$(1?h1^#t(U@$kDk}lL#n0+BU&lr
zlACw<S#MtnE?9?a5b%i6!$@4(Ku}P0(%2nGENscEct{jx#8?5rc4t<V()i-X5AgPq
zDeWKLjf%MD7sZ>#7h`2aU_xOoc`P&$lo=^8eK5W$6cckLcvgI3$3)DjT6+c~Frv`F
zs;nI`mdh`fV`F2(gI5UgL<%8`pvu%@fN3;aU5PDm@YhCZ<djxQV;EY|0R6v}D){}?
z<C;P^pv9~@oit)DwCll!%RJM}H*>ziRu}8Rq4Gi+vm=glMp>AmyfZ^c^_ExrQHSZ{
zGME(50H+~_k4Y1bnIX-bZl(N@-p9fTQ_dEB80&NC&ox+t`R$&p;us3zhmfd!v<-tL
z)0|<#c;<kD{AF-NJSt(*3!MijVj4M;aezUanBP$`3DX)t`UnvL<sgOCS9sKeojP|P
z`Jix^Bq5a$#xX!Jeu5!9#h>Ae^I7pk*yTZ(7%&ov!3cCJQe=<^bqts&|LjO3HXcC9
z-ji^`NgPvXCKUSSpeteB5j%YHLy(x|Qp*-e%@GTfJ*-R@J{4ZS>{+qY7#(xEWLj30
zFg0?i<4qXdcBx!{yURQd#cVWs*_d9Xh)0$^Fxko~fiMG~4ur{5^wqlSi>iVwaHEK8
z?ue0IR`@K&R3Mav=|BrLHn(HderKqt`jx^HwAw1|?GtO8!G%S-T3gK3q1usl)d>sq
z>4szKJ5(uV1t3dDQP#9gZI=#&zM0?!OfSAf8<UN8XusPgG+i`zh1#J(ai~Y+)?lgv
z!v6fy<H%rVO<PxCT$ZpyZ(AGXmUiT@LFB31-SB!9n=6VEUL7i4V6(0y-D(yy?CNe?
zDSb>|#hqk#Xtt^7{3kxC=jr5A#<N}9s=M)~FWaHh*CiP<1~mE`m%vg$hy7JmFxx3(
zWs}TmbK31iO^<e{2@f5Ggwlkhd&yVoR2xD2Nx-V(n-wl`x0P_)y-joaGwu?<+O}la
zdL^>o=~0K@`N#hQFA%G3Xz4{#UlXOMwpkyyAK7K4zDogD?YjL`<-P`US8`QlzpsP~
zY`1;(RmF~9weNQ4uYL!28}y_gswny|9{eaC=&PVgoKkv0>Dz|GNSK1NyY14g+x<0x
z7vG{1E^&19>25##3f+{T>Q-^0+Wvq;&z~@qo)xcYwCd;$T4;G~r=nn7Vr55I&YCWN
zhy~DKU;Sm+sZV^i6qf4mY8YMGqyuv)16Sd?^yQ0!zM3WuLElQ}#RWnw@xSo7N*5*<
zM}|hcI(cD!K8A-m#z_2$_&E3~5Hp#_YF#EY_V)MXU;K^SIWCU7uHd`C$wVV}At8kE
zkx@*793VK`i(NejW%tX`A*U^Hmlb#2O!!_Ucn-o63zo;}3^5tQZVx82#VZXBd$qTB
zSupX<?gjY0GLfA}t+8l_ooP9_1+PV1;DIuQdOUfEJLv<x9RJ2-Zd@wPJ^1N`UG-Ri
ziEwr^as1+7pDdMjO!Bf@3qFrHG1&@lNZaP!^Ms}L%7Qi7=kAG93&#kT)%@^snAsTy
zRp4Tm(%#-~9PAxf5XKQ*O!N|u!C|F`(ubAcOr8@Pcx(lci}Un&VqWRcCq9;iWWF0n
z;rS>q+Fa!`{HE_76C4wO92dz^Y@`Wg!}+wc?cHYFhcz}LdKDOg4S00#Qe)zGgu8`?
zl_%`h>y4umjg>?Uf{D%~@~C$K_6~4V67)~zQJ8uVJ~9q*ycJLGqU(je-t(9o<DMCg
z)8gJLCX^LNvBn5IkMyu(0dua|=YBM~$%2V=V9wQN5VrxP8An;c!!Iq&#%z<jU{qEp
z370wP&O$fFe%x6^ezQ|xO6Bp|(qi1ZcQ4+3=e@Xn_g+kj=95!SlAvrbVJ`Z{CoN>V
z=pT!%)@eL`@-m)1U-PQ{om~W4DjSNMcgKjoC+(?QlmY4`!Y}fTBhc9afFPCQ*+@eK
zu}rMHeNj4bbW|>U>yM!^j^~<+`GsrIoSn14nVod()>-$-8=Ib3<i4coCc7#6<Ke@H
z@zvK~Dx8+e;-TVt8Z(oVNND2u)5p=;+m7QS>US%iJ$)2UpFCFiJ@<|qc4ROCPWhtT
zQYWZu#DzQ8n8as{gkY3DoHm8Fqwb`tU*F}du^$Uu#~PzP!jxSb>_X)3CIny>49PF7
z04<hJtWY}GLg~Q6xP|t>_1fkXP_qlp#}hKnFKxHmKMIf(JO_es1ksdV!oi&}4(&%R
z(+Jg4@I-Kgkzs+MK5c;(^+}KE5gp)T;Vk8vHo?2q7{97)F&6cLM%TTHGwXgm<%71~
zl#6EQ_4tdp+e&c}9{Mu$z@MlLvFMaGmWg%3OQkqNKVJ0(p7Auhb~x&iwz`_DVC<j>
zO6nYm$JA=0$P*^Gp_{m8A4y!YZ6$pjN7C8xQ=6i+Y^d(jZqYZ?ytCj_Z2@JReuE?K
z8WU5>6ZHdAvoX<B`!G8nJ!&)7w_EY}*-AWnxvp~E^3ikD`+C$r$Ca+Fuf@jtrUhTr
zJ9gBv<CS{jc^u-RG$0*4rgr-WEkZ!rD$iIn!JTQ;%OTOn@sG?!v2%G+Ws^I#I0lw`
z%Gg;u+2EMg6yi73o^THt{R4LhBCw}C(zjBt*@fgC$4V>G5Fr{n$rk3Cu`oLsi~h<y
z(}-(xYIo)*V@c+@sWIQJMYssK_|8vH#P#{9SezLTcHgi|hxYP<(`JtLVo>3-YmB=C
zxAu=?b@wn<b`E1*ZPVVd%9+YBWtN2`>_emV?Dga;b1KY>@SLd~Mv&QKAslr*%OF{v
z^CvsooALD*U&LoW|9O1%#h3AK|Kcy=r$75Np1fSKuxMbU5%0YBUVQS&x8ly7JMrO%
z@5hq(idnTq|L7n5aeVi?-;EDG`Y>ka7QD#dRMf2P?92aFJYU<5&F#I|6dralwGa?-
zl%DdmSLru2lt+46Xnf3Zfw}9L#Q`r@H)DBqE7mqzdUn0&fHK5!!jwIB#!jg(V7DQ8
zjNOU)*OvNc7MQHAZ^_S=^t%?`u=AL94_<(Kvgpg;KWJ-@R2KUN#$tBidfdEmGj80t
zraogyahi|0#cSSm#ZJrYli?}gRYqcRdM@t1^KQKJ!H4nQ$DhQ<-}=3H|AUWw9O3NT
zVvJ8Tl^5C8aIfE`+B%N3q&^`GzIX3V+`e@qZr!>W?<$Soeec7#apztvT)U}yvlxR)
zPZpzWDE$zku^5Hj<TLDeUYLv7nQ6z#i)GY5!3+9M#=e1gwz3^xJX*0ZY)obEKs?`u
z>eZIY(eipLUMPK6H+N%aUp&pxWgMKS4Kl4NTWV8|)CL`?|LDCuiBXOtRUa|buXvPQ
zhbY42WGLD#RB76}zOI;xeYP#s)7t*!7yIsS+Fyl9^(KA$Tl81#-=;6>ik{~!ZrwCj
z^{dP%?69RqzV4dpP5x4K;_|9`yPs6Q%Dt*z7Y+sNuMchNufU^!&HrzLuuHv(?!4pc
zdw=?UJ1WKN1Vbi24W8H!HK<HZr1vB-ClmF>pZ*l<v;ry{O5kiK|Jd@fvtOwyQM_Sv
z_)_e#ss1n<4<`P4DwP>dnlSt2cvM5RlDjln{m`ddu3S$AmI6=^=<%m~E2PY27?}zj
zglTOTK53jjvT+?oD@(%&-a*-8@nO*PZ&2R)yKj=lzv7<u)+M5%=iP~$FrZiC>#rVA
z;acOJbmz}HeJafu&=*$BZnmd9v~7i3X5&o@Yq}>Harf@Ag4bZyVwVP(W>yMLP348o
zIgt;Da~~ixl|@@{)G#nfZ|7tTi8Bn^8O)>^;xwZWq_hPnf@5&XQzrqc5CXq2h>ocU
zrK8<e2gz|OJ3G5^#t2L_fE7CEkU3f)S8i#LK&G3H^lt^t()QXDbQ4b^DiCPL>hDvP
zV5<`QXJ*EEM!N7Q<O=4vnid!%3KlN)Nrio-8CA;`syzrXb4mUfW7`eE5*>O|gD#Fh
zVW43=<d1&j4U9aiWH{E)yP%H6@S3Sl0b4BxOdyqD&7656flgfEke3LXV8+dO$^!#_
z52T5)7*jYJ2FWmV_~}<?-K&8S69+KC!>rg9lbXjNAlS>YMH#^ldUm-%GkWR@1L;DZ
z8E*yzlm}17IdgNM<rTClFSM_$=wg*MOl2Jufy>8OXh21r!Al+S;KTT9o+-U$PeQ=B
zW}4$72!%)Ov7l3Uh$Cs{x*!M!0NFn1FME7ZZwY%sgIy+~7}RCG%35g5_(9KCf1xE<
z`!yaJJYlWB9f-F+v}0IdmtV>VMyN!a_IR<4-TX4$sijQnYp)3dZEflj+(EXBZ&4=0
zxQK%d>0G^vW@EA&4hfr%`KhW)>+n%YS#i_LX2LK*(wpW~nLNCTyPpbR7w1F9O*VMI
zxxy+k&`EtZG6z<a+%r8gSR<%WlHyye^wrNm2~c6GvOD3SclkjBAojK_Uqb>WOqqr%
zNeOD(`X!udlXeJ|&|R@5Aj2hHD56O4SU9@XgTa!Xc3PvVyEHXDtblV~r@pPAx!|ec
zWcZj$nD*?JAnj~KUot?R(}gz*9I41pR1E{PuE#QTC`{h~V3Vc_jDOO#qpEv_$uh36
z{LMT-rO#9*!$|xx#;U6dGjTMn1!n=0BHrm{lG^rF3m)m@DKQGR3nqb7cI`hQaCdy|
zO?FsQ-s~3-3e#u{B0e(z=qJBR!PkG9CnK0RKKZ4usda?*FWb_N+287?sBZhJN44$s
zeHDf<s(rPiU#ZIQ=5uhhP3MafCN$Xa1r8%$+u|LlVlQQ9#jh&6>6ctXDFe_@@dy@s
zEa+!sAsE>dI!#;G)9x2>2E6#S4tsQhF|M{<{St2aciGXGU(Ju|$1vzsuF?+GfkQ!~
zX$4*icw8#&RgXV&d~&ymul?p1bfiI++=8pP0h7GEefnL3Pj{)LC-sIYTGx|mkV^QD
z8C7v)c%;o$cVQRZnGf(K$44Ajc5F0+%SwG8r86__Rl)FP+#lo}QsS%Nbr66Z93A^e
zA&$iH5lQ+o>AtzK9^0GivDa#OLZ7yW<i-!XleSx|fae&T^WgrH6Y)m~Q<z|EHpeY|
z+T7gmeHRG5xVW7OId;1o9-YRS;=^u|ysxI*)g?Pt!=H&qf-i%wVl}>bE43e<ycItO
zZw?>C1T+FER`3tX4gQi{5k7)&|1dZ%>+Do~n+UJt02T%yFk!+IK~Q6aqq&EYKN}bn
z4+*@|P+`L_nFp1g)aUps^LKjS?{H)H3lskhr7^p$*a^Vy0<y>TNAY0s7v($YXVMwI
zmWk)=M-V^}PbOSBYLN-Kf#F6BinnG5Q@_mcx(K`wTu~Z*JezPaad&VGuP8Hkl`a-q
zGTA4d7(v;b_+BPg$A|laV<0_VG%ko#KByp%$eu|p3J`ffTEXl34w^IUR8xBT;#3_x
z$Nn)P$9^+v5PW$GM%neW?_&~A4w$?>kBJGzYn0;xBX(70*xkYXRveSZ@s}L8MLBlJ
z;9+OWVNTk^<FW%_d}7=?Nhntw>$EgC84FX5nB-V>l`YB#f;G}+V|_#Uzi;~41vja2
zD7!Hjr}PdEX<RU>vezHG`)9G$5)H=+Q*qkbJC5h8+wt|2m3Z`YCDu2#l{6Yhs0>hc
zs4oaRYTIX_i+981jwg6o%19kEFuwHUK6Fq{RK}@$+?6-p;GwLDF6Au(OLhy8=fh*o
z7)M|*Gbi|Q;ZptD-}NM<P^%pDDg9c}+S^OPhRPNKfw7U1n4W6J_V%XA{XvYY9=>?-
zJRUswGFsd#qlb9R&d&LGO2XwxGTH>vm>oU{t=ZW(K8`SH$m5f26A4%4#p90TK`b~O
z%6p86mv^n`%T7+pKRaG{7(bHtw2;&z%D!}^i-4z0;gJ_iDh%p2leef8d@;U9u#W(Y
zalV~Au4A`Nb85=-4AeXPBKSy}*qJu!yrE9mpYjtyn_{8}q321V6+%egdqy214^@^p
z8Zi5-JfacdM-^^T-uYJH5^WcCiZV<)B#pypkISU8c4?d{cnkC#Cv2#L2<1pyx!XUs
ziYMtK8lj7FM7<}yz)70`6ga>K4YXNa3?Nv-rhOp~`ejGDAP^*N*ad~zchLza@RSKP
z#BM)mKbaTV*0Z(KQhimMN<X2x&4LMbbkWwb7y^M1eMjry)D!Fb79<QsquEqA<8i2V
zhzWM?x59mA@5llWcDRuT-j$`YMEY=w3(B<&3q8a;kC0W{NgK*I)@e$7RVL+5o`Ea#
z8P9x47sl<U^3!9X6T&ZcYH`;XcNz9woFHzG@sWNXJ2*c(X(1T*<V~oaB3NOEGG*C2
zw3Kc<2$IG|hoU(?q_$`*7G}mR;F@a=$9!`%7N<vIerhD9M%6a;oyDN~ou2b!A18^B
zYivMm_~ntpKaTzF^?3Q@>v-_R=kfgUgXlRsQC(IY9T|%QwYw|3NAYs|FrII;VtGsH
zuk<C~*nKlS))(VL>O<JMrS_J)8WGw}O^n9Oxab#sEIv?m36AaC-(xr7PCR@5Jh)@<
z*|TS{xxN{zE9)xL=fRzoO|?DO?z|JX?|%?~@SQ)5-~YXD$G1QJDDL065qECfh+8+V
z$L;Ia<J+Hn5;yPMRsS?<0ZU8eVomgY{dhSZy;${8g!DC31@e{lp5q}|m|(#>{j2f?
z-i@+Exo3A2JCs>$z>$8G6?Xlx<LHnbv+5g;PQiQOJ`(}EPqd*8R#~9zdPl496NXtW
zH$XC%dPm8j%-c`zfXeMzZ0^t(D4&!;P31dnBk8<%c%mvXV4)sD-r=!{7@wMr#^h}D
z509B=7*!uLI<EYpuU5GqYvkge`GuvJTeucWH*Uq!%{vxoEi7G&ckbVfZ+-MZeCy-)
z<D(DXi+A3+7q@TU7VMJ!vdDuw-ssyA7Pb(CD-DKK2B+0VPb+V^OOJUq?xQ<Ve>|+R
zIi<E6p&EVJ=a1Lo;qy&bh3RI)Y0dG-)X@#ab!~Gumi4??-i*i3*W>Zab>&@4ZLRua
z)jsat+*4n2e!Ls1CNZwEKE)14r32~A5v9w^D{dpGV^qg>JC~ngR{Q669*b_6wCMg?
z$M*_MsyFG|-=e=_|BAl4Rn={-WBIDCz}ovaxJ$jsTy1K<@Fw*(d%64_{G|F-?p6J|
za42AZedyQIvs<}o{!Ng*qTWQe1z&&qy?^q3Dti(uApeu_Hu>1eosOcdK<G!rb^EnG
z-EmH>9n@$5QuM`5v9`Nm3XZm8vFlh#Pf8eGUlkJU3{ZNne4}z8T44>EfTFhpTjBXq
z<$xu`jB8dvGN#4>z|^Ll9}qZl*6HhiYGubHkD)r2e*Ce0)u)rq{-m1?ZS?6zx}jn6
z!CEYn$Fx6;iQ5`K8INMmlZFm%xHqMZ^OFi#U~mF3!Hd1V#%CA6q$e-=$pTb3)HpzA
zW*!PG9R(Ac7gLJy5pz|KSWCn}2wMKl@?}P*H80HmryHGCGIp%Af&oSV#x9()d&v5-
zeO{d_UnXF3M-WV_9++3*WB|<M039!bqwUS@SknNXPM0&=9cMi-ilha*b6~hDjV6sa
zV|3j02+nDvmnZq8H~giiv5gOy-%8=VKB)2~jZ%1!X=E~!1k>dBNS{wFOx0XmpxOD{
zQ5oH|gBj;*tEyz-u%$)7IkUK+0TZhclL5tnL4yTJ>byBF2f-rSqED6f^@zjnNvI41
zfffub!aOnjDo4~?=cV$SRc);3%skDBh;y0i;4IsL4F6J$UJhU`Vo-yz=Im6=OwWip
z8E|>=@ewdk`oZ+GTGz~(m<cHHgp3#&$_^`9C~qKUaE9UtL&D+y2>g@arW4vc*-?ag
zf+3IcjBqh|!iP|=-VZ>Tv@lKnprt*j10I!M@{w3EprRac|5p`6R^=kVi!_lR@bWmF
zltv9!m5M21@B|#;BwfM<atypN7ro1mf$ZXsABCOYgf$rj#x5n!&Qas<)DK$$>UZ0q
zPyVbbJ^mN}X;V}}@>po+Iww0@rb}uGS8bZZ&9)|86U?}+DYp8iU0N|H#~PI&wzM5^
zwlBD`w~h3C{VruXEsA}4IX%#pCzW*i*SOX&NKUJ6+*at)#vt}6b7rWu9bA}k(_evG
zFIBS2uvV`S=JHU!RdvjXr|Rw2eiy{NOpm@LB3QS*@EE4hirHXx=yt9q<Ec5>hOc3u
zOp{&Hu6$4~57r32+B>|W^^1HjmWCrVIV4j3yTjrIOm&kh^g)B2vD8z;qlUC;1+G~0
zD*CK9uAw${R4F$FuD*@KV7w}=1rLhw`Pg2M+X`EmPDo2LFQC#fv87D~vEab*k->KQ
z*&$(Y;7bsuL<Y>YSlLIhGeSPHNc#WV5qvqn^wcu$HoJUXzl*9cw5e4Q+ctdp@vq9T
z)vPb(E8$4njl;0@(crU<leb{}GpwE(ZvDD{NEj6TOwndi(_t3#Ra^xJT0Z3)>Nn}1
zE3B129W!o(lVwD9_J=)bRAEw|VF7Pj`*6XqJ+uOsw4$wHqv{t2_meWW?ZBBxt*0!}
z|F%`ZU;Pkn@=DNM=&28?jaROQk(KE27dJFE=qaAk3t`yAiz;D#(%Jb!pvNL@92J)M
zlai=}MXwTPo7Yr{Gpqop%?FE5;r`UUy{*`9Z7K^-gWW>h2Qw-jfMds*?BR|W1OZIi
zFp0sv47qDskC1uyz+v8}f-nQ#%a`_x&w(o6zXET9kDh+-2pU3YGd3K}@sXHl3|k~}
zB>$(!hYEMVEA6@83H}6O8$u8!bC^_OG8X=bNhu~&IqHwQfpD+FBKQv6nFMB%2fmBC
zq?A>AbLn_mNG3ewuXpB%ZcZ@ZVRDuUB5<C_4Llsz$0Q+PFnQk;k2I!qVS*Fhlla1q
z0gHKPa41HZ^b~Ih&zI>&d194xS?nZSzJH9gW^$RhDa{AbD+~l>EX0DpJ61Zwi^5Z~
zvyF*Xj=Q1-MVQ4sXWXxGY+jkgcnHf_*u})%ASF(D-E#b=$H!uNa@-TL@Y&p}f?$Uu
z4o64FOf&xTzA@2zbP)Ty+tJSi;!uAK_g%zLpZHI?sYK|@ZaVSv#NPq|j>sCZK!hDa
ztomjZI6_B`RpOo!=sh^%exsA%4zWY=o~$D0cr<um^XUj`O1)yzlOw&xNBceTJU-kT
zlVgK1*%*pR%8}yKP?;i5-1jsxMA|BEL>sI05x^;EA<7+L+>!S5#Y&u<i?3FyIPBg5
zm4(aL**o?M_h-xNvAn+R3Dw=blh|n;#LKnKc=T*Jp2+X=>ZT|5DNo#4gD?nT!iCBj
zb%lCBda&}Iod%3?IaZXNN9+n19#Q@(ui4p$P}X<VXsn5V&n+cm8+L2RKa;E8VFZ4~
zgEDlgGy^~N6~PyGnz6e7=8bDEpMB&*qZxDa3l{9`@8@xN2$<RFgOKap_uh|x{N3-G
zzLn*b*xKG!+8kIA&CWE^es*q7VT?I!#z|Fpi13QChyaUmicgXdadkr3J0uhr#?b7{
zV22ZeVD3y}oJU*XzDNa@$z?C_Q@Qi5D8=17H=T~8!9XrbRNmIzjMU-W9mw&4IfnN*
zQaBkWxy?}pV1XclL-MCNHR*fb5Rj4&#2XyO#r%W9BhnsgRsY!OOI~E2<hav2i9~M|
zFy&%F<^ojPcti(vH}9}ke(9mTWY??(bHrWBeLrPJcxk6-8w`I&oo7cC=|pVQqKXdE
z-)SUzEX<+aDQxcXBApPlrqDd;)l2lIP?2;YVF-tKBgCAZnX<5Wj5bAiT7_O?KE6=x
zzsyxYN!ZMJ&;^)I#>Xlr2S>S+WK-p5b8|cP)Mg`4rwk$RYPDLiv$JENDvJ#)5JO;w
zU{?OvjY9iJK2dfN%yJASf>qB8iGIo+b!}2~O*JP?yM@U^DVXzaSA8)<KSJ7Udqpn<
z>rWev5SF$XH;(w!Urs77=k$y0*2;yp<cr&O#hbk1cu@q)EQVN^nT)ySXpE@|K09p1
z#qqv0M=>^p(5%n9wx%XVVsUmnuFX!w(sUzcCx*SCfVK%?*)Y4S^r%Ty8>upUauhA4
z-_e2Ur@~r!`7|DU`B^-A@I}0M{>;LI2M->_mtTDq%P*eCsp{#Z+CA!_?>AIG%ovuP
z$;5}dn}&PNVnXz=Lydgrh~7>0m*k1-oyy3N(zW+oZE|ZX9zA>zzx?deXl-r8!Omvv
zY_7)gvnR2(^3td9P)8^?H?A$kcfRwT`1sr3j$3!`#1i*Q&CSNb^i<4nw6E#~cSW6C
z^u<}vaO@smdKcRBwH>#c2&)iMar7TYSkecQr|i1rs9T6VQTd`?vyfRGhx&jlx9$to
zu8@}qpqc06ey)AB>9jlM@$jNNqb;IMU_6owNXTr2^en(Y7)@KmDLwds$IJN`gw^!7
zxvQEvo!se;u$SW?Uo5Z1ix)4WrS=+u7=7pFj(8FErAH?$kjTD_BNo}|N}F|gd78U3
zh0FJta$GU+N~gX-wWs4M)6JQfo}Ke?v)t`DKRe@{f-II{=R3Q6*)>U>r~J|$`aVCE
zr!lp|@C8l!ccmr#%Go(WK-D!pO30X^yr|8lzO8TV#e*j+@oZ%)w)aJs(xdKbcim7J
zq<0mHb3`v?XJcni^;r2vJFa;5sGp&oA)OBPTjnY1QK8iJM^B%}^4hw~oVixv0xx;d
z`KqnL4HJxcfw>BYH6zBHvr6;TFX~PD_P6M-*!$t1`s!9yxA~jg()%~KOTEclZ4~R*
z+^c$<J^bI>oZjvy)vt1|>eq!s0sHGicW@JTRP}4~Z-VFz>P>W8@bx_eUpP{hq;hf&
z4VjOPv!YnpqPX+uC3wbVsV`qYRiB@>yDjpQ`nI8C-?r0J&E2@Gs%}MJK|GpGjX_|4
zOgn!p(j%I-!o51xcy`;_J}rv{11Ke$!@`=0Tm#sxJqZH5wSrUQVOU!;P&PoKv_da~
zBT{UD=Mg^turD5})+T9eF8V|PylqOpIzR&%manZ7Z^C9gtjGAO9u0rlRe0OkQQ_Ju
z?F{2^i?bfPqeKdN9#-fe5aGOkm=+%|KvEfhstf}LInDkm0dsN!MC39<JN+0e`4Pie
zn9VaWkacAg9T=TYGLtEIK(K;J!A+)L3Qy>uI8NX&w#ATeLHDsbX&UAvMPZ9h`dOGj
z?7)CZmez#okDVSm3F79$LO61B1%jc)O?tp_{%ur};IFmC?6|q$l|XnzOQpHv%>+s#
z0P>;a2ljkHQSsu;)3zw6nXoZ)E~d)&QK|foYrrw9&U~aX>NLnyymWrRxWROCeBg=9
zXBsSF<|q{MiN|4+_8JVs$Y+}6E-Hxdz(@Jbz>xt4dBR`}x;PrqJM}0=lotln>a0C5
z6{FQBW_VD8=1DPbb8`zm6Ip?ZDHAPX2#7QSFIK})W;kaZVHgt~2&PDbT&?TioPyo~
zrzlb~`2%YqlkmEHiXqE#$uO7rlsM_ficw%lL+01fdxB42{1bl|eFi!pw{S4WPZ>AJ
zAb#M~F9$rNw@f&qO1X5pLb?hvyS*kffaYk53P(LA-Mib7q*6+Dz}5Pi`uZeHadG(8
z^D-Ln>d_A>EsRX`QrIVL9o&4Y=NfLcNxL>f<X=H1yQ)un{p^=7;TC;`v117((aMm1
z(5Dgj0)xK5U2RX(t~mIMwZp?MJzufYr>%$+-qMaYKu#6wGYm`}+<x>!W%}8vz}ECH
z`YznIDyGDSB5g3G85*l&O(RrVXQ-JEvdfQD!c`#%F!(ht{P>I3E{R@RKXz}&PA}>7
zr*7HJb~dEy(zo#zRn0XWvb<T_Ru~h`m^)70c1es?ey}K({6QeZHGX{27Ngu~OuKwJ
z6YQGq1(09nzbx9$9sR4ya6nXCsI;=G<rP)G8Gh=!aq}rL(n~tgT=hBWjcYpDTN`(N
zT?UM}!1;Yu<t`e&&<wndQU$5Lj!XWNjEYYdKl~jYb@&JW=s)otU)^czel*GHe7h9x
zwA0XBzim5k>BqN?hfmcj7`xekRsO0Uth>L~)5{b;-Cz5z?&>>aup9R76VTIfl*t17
zwH~+H@1&jmmHyK-V)g?}PBvnu<feUrm8yMS)(=}r2?Jq+)Ox==VF9Pdscw*j^{2Y)
zrI@6->uLKhX4@6|{FHcglzsTk<cGsFL6x@RuPDc{ln3I3v+IMDLDNYeAZy;GP2#FF
zqrhKuRZ~>E%t1#<2tH{mSm5m^{dE=5ge&N>E<uI9coF}5af-01Wx?0h#(K1Nc2dyO
z&u)l8^WsdfF!AS!#i5*}hgX5G@R4W3!^wZimh$CAW#lt3^6S<0!ox%?{@MA$Zu+UI
z@tB^F9^sJqFIJhu6A})Sc}#w?!-Q}Up0NU%Nif1<0+xkr@Qtn3zE?7HpAvjuPDm?k
zxA5vK^nw@lF@VHD^m4D2@1yY~5EHB7qeLrZf@8mS*|p#axpQCI&gAuo_`{LWAz{>b
zLE{PTtYBi9_K?Xnj@e>@|LBkfTkLwf4DKppl9kChcvDaG?(X_No;}ffbaEsp#Yyo&
z;Kh+l6U|BUpv@`n1ex@19Oxs?OiHtooBO+R@-%mBaX$wWf8-sLeoW@p$uM$)A^;s6
zBRSvXuB}Xa(w$vhJ))0^I*t?Lh@}B`6e+#f-Alfl?C-_?-cF2<4#xz?dzA^>5%Hev
zxMA|2Fc4}WI6PFIoQRKSN7lvJu_u&)7*+a{W+$R~|L8<?pGHgH-J>(FJm1<oQ9O>#
zw^O#*>BRj{EW~0m4*}QYxY9}S9#Qzy6C*J*rLrVmot<I`G7+A@7ZZQdh710mt!&2f
z#&+y7F{&{bg4nH{eNW(VoD_n?-TmWO+uVug+@~Vg?Y)E8-aRnfp4^zsJXHDG-8(ew
zp>Q!iB;xGqn4X%7*|}K@3wjk#CPUe=)W-y(zVP-4H`#5(ggknLSqQb*VMAGE;Ut2o
zjZK6HJF%^>S;bGeMi4{WP}@(B1u*N^+cT&%3&*!HIvO{wFU8#Kw923A>%eHt&o9RP
zySH72*VkTp!gcQ;@0H^&<KO%C@5lQeex&*^884TYV{3Cm`FQAEU|XA;apT5KALYeF
zG`lzvaI?@3fdTcChk8jp_i^<q|3?~skwyrvS@FNyYQ^q$9!tm3j!edKng{sU&2uK&
z5g-i@kHi2+;;EhC=uGbpQ#s%WdZNy5IF3rB-LZgE592Dzy(h_)Ca@@O7HRL)7Qivm
zH_<lr^j<ps!OKyf7LKUyBS;%o`p|YTdCtQajEPpttQ#lc@wihzh2$S&JlYlCm#49l
zkMI{g?RFM=<kQ(!+DFexP*;_P^mQzng+6Fz{KpPNV0<K|#=joh=?4iG42p(q%SDSU
zEzD6`abk{-8!f_rPw)#T{jP#hI7zh}WUu@q-}-yh1`eyO$@XSqf;-n57JmB#9)-hL
zn#pYF@M2))7lPB{<3qQ@b-u-ECVdJG7*ML+rT@?ao*bJwWt<ieNBVdSthiYCulm9S
z{<PYuS@jVJ<Y>FegB1Ky&P2Cp&mD;LQ#ro1fK)KK2!Vb=?SODnrhM!xZpts}D`kkC
zQggF2u{b~Ldm<5%(`F+$BOOWOGxd|~$V7<8+!5~0@zLNO#WA&=+?{y5w;L~>J&uQ8
zeI6St&th}+dGuTy#bA%xhm*ZHZLLSo@vhq5M$ArYY)u{tcfZ<K?qsCR>Mv#3!g9s&
zK<WGF>o4QQ^CynyNMl0nmdbGtCnQb8O?FdVTT+=DQ5$=z{zUD?+FI;vtj5;Li`ZCx
z9!I+y(H!ZGyVn-Ht0?-0;!sJ={ZORInCj4w+J5fvT7B^>KKt2E<L5vBG}c#FVydBj
zQhm(+&UzfNi&S-$UBAuAiMV&?R@}RLC*FDYzS@M-INaNhi{qpC^0S}E*I#_DhUi4?
z@|2He-95RCCo9|WXn8Bv5q6z2r!lB9HmN?IBO`}nazf=!eIGlMlts=X`doxV^r;A+
zRM5)8hGEx@1GOL2cb6}vAA%fU57gG=A`!||)+<VV@}>yQ*lnEmZZXfqqGs;=&vMKO
zJ1VCtThtSHR(1fA?<{gg@IxJ9!2$PiaTn+6y80FAsl$qz`<oZ)Z#K5}?CukylpdZl
z6n^?P?sD|epCfti82&h_7rLR1wh_9Zf^uuVOXcxGeTp(s?V8Gg(ucWD_#2L59#z^k
zm1i^T-c!1=!;t<P#jd!m%`KJhOYi(c2>syMN<4VF;)NUsN;hB;dL{3res_#^QFZn5
zO!bBOIO(H<eYfkV`Z)DlOsBCcY;!mJDcUpIq->8k68bovK7Sq$9zR!@2Tsov?&m{(
zxIML5Vd=>uIORh7#aDK~<@aso*T1OO+`9GMUp@7!SKAHMtw&uoUxjr5e12{JX|6Ud
zp|7}C^=A85!hc156RvOa_p16$VSjySsfwFVx5wWI$(z-i=<W!<WN?R6fOtcC7Qb#)
z><pK%oZr!g!wQbWPO`0feG?|lwz8($(g{~Hu2QZO6*G3Yr^!lAPG0?K&yu8)ndPZ=
z-H#{VXes28Pr~!>qS4f)84`%0!?Nl?)kHO{;V{^jzb*ikIQ0IZO@F{yYcJ_X)j!Eg
z$b=>R<VF(t#ou^4a8^c^`qUPC*?D5As3b{lMySWMW+Q~PeM#e>3J%A`b`?hOHiIVJ
zn83=Y&|J%GkL8t~ib``$8qQW`CDzpRv>8YXe{iU2txxJw!6<>klOuXemww4dVaJL`
znbpB~*N-|t^j=-2hL=v0)us4>2pCcHDacfZqA(ahz;v?e^IU^m2E&^h8?mvzl}wt5
zsw|{hbA_p|=pg@4yfX$#tG5&p5vMiw(n@1!EYR3qKRe>^TdIPyqO#GQOtB$++>5!w
zp`3@Jm2^aCm7dF2@pTNyCq6C;SspU)fQ3O9lPU%eMx9lAFt6h<*BTrSkBu5XE8&y@
zO5c4wdtyWm;#eK%nL2%TXK}X|400Y_cxn=$>CBT385bB829UU&DQ_5Xfj3wFu`5d9
zi%wPuUBZl0PKCpHBSP8XHOva*5e;6%P)>6u_57SWZ_a8*xJCQumAs1V1$38z3-M$H
z9E{PCXoVhDJ_1J=d^3GZn0oYb`k|*Rvdawtl^J?Hq@|Bb!3{t153Q7Mo1OkS;Dmu9
zeCQzVly5?_g%ot>07!o1&nux&a)#kx00B8TQyPI!!Plhgr3Y9=weh0yM{YC#jvwZ%
z_29xCg^gZm1XMq~RDZQy0tM;hpT3T>i&O_0Z$ExP7oPa13vV3LJ@KIL;7z>&OSG`}
z7aGy#%iko8ShcO%sz9~hR@UlOzI+^R7jEOi%r~{R&tEZ1SziqJ&3-E_+{)m`{<_>T
z0cKn5^J~kZs=pdi+b_7WvYTM6FAw#l8q=E%)fv^%>s?@+hZqY`qL-oARr*S)Nd*41
zHEm@G!P~q2DFaIp$O4>UIa?jReb~Kz+x9hFB4azLV#mkn&KJszTV_7COE^?VSi`HN
z5%sA+wVht^+bPqs(I?TQTU*%@AH+*~MXr#T%6dHO-_{$EVbf<f){Z{E)m&`#L6;9<
zrUhnt##;8^(<&UQ6R+|vBsi}sls>5gN2z#Sih5p^;{sSl3QhWz9r;@vZTGtre07dG
z6a|W8y9`P1HeCw4YKB6}!K<*{s=_&Lufq9Hd#%2&xbZd5Q`{Lu;=Kye^6BXHN<Ba8
z4czvNohJj@4<EsmFw#ARRkhEtTN__KvP@XPfzNfOxK(@UUc=*W#g9?+Rl3qGaaBAu
zZWQwZATMBUTL>3vrT7_teNci@IHXtVWnZ170p8&XAM~WP%+~W{-k|JmA~H{`pas2Y
zELJrDS-3Rt6><27#_G3?LkfMj8I@)|)lX`rFLc&$-lAT`*HQRKz(<(e*ShaJg0{9c
zg8P1$?DIs7_$H1VWHmWo3wFetFsZ?WXYxAA$J4Xmh%Z(R*Lzx>23>R`G-FcFe4_jg
zizl3!Y{cZ~uy+q3Fe7VNZOlY+-m^01-6l+=kiKI3EUaO244kAVf`P+>Bk_=VbRiR9
z@Cfj52!WWUMkvI+S@2ryX7>&T@qP%txYLS>bo5L@!as6H6MlR@kLZOaj!GIq5+wc;
z{_Lns>@z9KBppXKdbgM!#t=UGPwB#G0~8sbY@qRk%c*eY1h(iAod_Bcii}T;MRStf
zJ7X49P3oB%8;WLQFlLojv&}{f^a}5B9{t7rMhAP_adx<8VbCaiw&K-b;z~FPYiD~q
zTDx2Fa}Wbc4+Ljt;<e9}hogcKfw8r*CVu<KI}tc;(!xmLf=5K4M1HV4ZNHT}5)h&=
zv8-q*z0U*T8)?sk;Lc7f^6r$4Slir=m9;GkJP<07pM=k@3U+?^$V{aVcXb^epIU&o
zcR<-=VHl;8@-pDvTDY_8YHfo%S#l@VfyQV&hw{&Hjh+-#j&pS2`c_L}Z<{s+5j~U#
zjR(jpcBK$LV-v~>WqoUFH#S8#WuLyBiE{UOlouw?6@OOSd!*jWB)Y~p<bl$F2LUA~
zKln~M@z0EB84pko5N5dDaa)csRc(!MyG>OY^NAEH8`LcpTRI(gcf3o6$<r3Q8-3T_
zzJ*0+lrO<@ha3t455mPh)#--v-xI`24~~mt68ZY#LcDwbj`FoX*4I`%S&Q1*-qIN4
z!uR}LyLR2XzMemOs&d$fci(x(v>*(5M|uPs+sfM`CYM#_Ilj;1XO(MLMk)6h8ovOG
zFxWy6(Fr`0)r_mD2XIOx8k5>}2N09eOzcu0NY|a6jaXiO7OShzV^v@75WK#0EoNtC
zMH7Oi5yf-ByE=2CwA3lpadr@RJS9A&F%!w`rsMcMCYBkqIn5CUiC$F{w@uW4S($e7
zg>nEM@|yPB<5_TueCHeOxk6?)n(J>lDi(TZo7^XYLuQv1%BR{FCcvqio|o{C6ih)H
zVWEKG$&^*JJWk9p8!y6ldh1I)#SI}q?wn)yoxy}7VQZV?u{miZJv-BAuP9HX1vF9a
zNlOHyl)+uK9|*b*%bj4phmgs61O|dn-i0UA#PJf{jAJR!^u5HLuw6$ic;b#XJte(7
z!LNLTHrPbci1w5E%wpsAJ&_15s2|jWBX-Y_M#?YuQz`s63~p)@h(GlZLFcI2QWhB?
zjP^nd)oJb{n`}<1UJP5P#ZFz?;9>PKGhQSxseX!46$=>{+aq{jj!5aoot0{*T1RKG
zv#&Nw>5M&Pi+cqZ=B7f;soK@U=({+J>E?J$8aIM2<(FEj&5hOg`s=Ub^Iv`uzxe#i
zc=-59yi^_B*xYg(@vU$Fe*A+!`EK00eJ>`O(_R#TAp6#>TXFa9?O0k`is_l@VD~6(
z#%EuA5kLCzPlfZVc=`Of1!4$qSRnFZd0Bb6={A0x9bjs2)>dA`7oYtizWn?%8IR-E
zjT`a7d+)@I!l4ge-`G(5sWwE6I{om{^`*Fd=cX4@u$XvPZ4-5P<>gBJtAF_~<GI3|
znVj&BV#>g_+Akj`t9%*gJC7TSGv3LBK+|`ojSt77(uw2o&Q6cKD1?RPt$lXcW;@8v
zZraQ#rQh__MD+C_c;X0L^|cCvt1M`9DTkCd$}_@Q^xknvA1pXZC~ZOR^vd}p3&ZuL
zj#IZ0dRpkKyr!<I#?!dEPoX>tofQ2x!a4H)%(PHG$$NN@P4zq6k;PGvD{BbBw%xyh
z19<wTBbD2J^${-nO3OSQND$n^s`5g6$bue@P-G`Ef+qMD>NiwT$EdTkwFvB3h=UN8
zBQg)w*B`2%=AKkG7SN__Ztn=59rP#hY<WZZs&YZyP+6E&IlsL$9ShTqxUn!5@87;I
zo<r?w-=+F9grqDm;5b;dxf7}@`oc%iMrNP97u%whBT0SN<^{XZPh<JzYCL-K+<i85
za+(*)^N{*kU+t6$*tji0&$pU^{zmibU({=E-TLmYo;2rE-B8_n)K#~ny{(GV+x)+Z
z3x!>^rCEOy4C>AH(3)-h+i<^0z0Lk@>YLzk!M**Y^;^{ySUrg#-K<_U{}#&lUr`9Y
zdj9Os{x>9<EqK#uWfzRWHux@XdP&3NC@|O-tLj6l+a|x=@3x(AGiZe{!t92%YTMgT
zzFo@Tz$XoD-1=l((`>f_o7#fQc6?J)jctc6x3(fZpt<1STPJQo--%0A6|&&3?zWVh
z^w^oolGe2#<OGwEV(5Hv^dCkBjr`bq;<o?pFbr3{G5j;I;P_SueetH3z18sK;<w;|
z!H_gto2tvu6)axMtGE~X_*IqtN3R*6B{AIBZ(dX9Jt}Y-#o}B3JUOFA9%T&D8<#pl
z$65B3*`^a?J2qcB?IQ#!>Xhl|Sb@l(6FsYzIPQhc&`hWqQ3f8Y=Fq@HOcR|i10Yr|
zEx&vjUqARNHZ*|qwrkOCdI-1BMn?cGDAS2GoTnqBm(xn-uC@}b(?QB$sZGna;qR!r
z!eQ{=%Ml{_Q3kw13)-X-Eb|^Wr@!DWa8e|^JsiiPSk=)Mloxvj86YC(lI%Q&kJ$_u
zic1Cxy_Yd}?ONP_|NWR<n3KI22{8%W=e4u5B?gL%FU4S~GxZ%`dYDKt9ZF*c2p8o@
zB?4if0kZ(TzNlM+J@Svq58~!kS{fjd5x%?O7(DrqX9Z34b976y5u<FeXGQGwYjH<R
zCG;hOL|LIc9m6Gga74N=kUACPH>k3Z$2~C7LvY|u0UMJggu<X2=8#Du3#x3F@xv1X
z7AvJ`6}Ta&qx`cP)Xa<g!o*#c@=aR9ymNsyIDE_y<=I4lioDAK7a7&AGx~YuoEbMI
z0v<>Uk&NBa{I$4s^NtwI8D*bv%Prd#YxN}nFzQE{wgIfzp$G>*HrPIEG+^o1@bYV0
zy=;rw_G$LydxbA5O<CvdPZpFo+sJ^s-Ksa?Xlo5u+`&y5(AQ?Ef+dwSSv{)n)X{8O
zFLz)ahhmwy@hBaiG}*;gX;82-6bs|fV<$AyS<rz_xKrp+=x);tR3fvB{>H7R1T8Bd
z7!*r~XDd7=F1MR_O)J;W9&j6me+kq6z^6y<H4Z34gP`EGoy~^!kD(o5*$psY+wrlr
z^+|Y159ukfX^$uA?7{zHkG@iAO0cz!t)xtfHR<lYOGbw)jq{<Xw39nDfI{YW|5D%`
z1aAEF#aQi7i5&o0>W>K;Yq)7HN`W|d!KSO5PunvYW-VysM2&4L<CfWf+VtYPfV9)v
zA?qu5z*C=K4iP)ODB_H_Y9F7};1^Tsvt`PDBg+8T(v^Se>o>y_xUm?-;0bFdhnEJz
zF%1b}<u5P6xDAl1e249@(Dz^bSO2$k{2i$O$^Y{I;9!Y6w7XnDebJ+ASKLwbLEY|E
z5rbSLq+dx-!vI#&hM)UqTbHn?pK@a|4ZG@QoF$B$fGp;AIwyYTcQ$J6vX4Fa&+%jC
z>lKFSE<ZBIf*vX8ta>X@8#duhm&3{YblamJyp^#DcU6WnJ|?CTCwBe`OipqWO>zSd
zlccz}>nw32JqTNNw83`c&tp3AlV7JcUiGSQRY3y^Y~V^Iyx@Rl$FW0G4HF_V4y2uI
z)%RMJFzm(`-~Lzp9hwt7zbN!+?$}j@e|b&6c28KFA+Ad$j#S<0UX}d?X%8+i5mnNf
z1wQ>TG}4HX#%zo?=VM}eG5Q8aJ)y`>4i*%VF8r8K<jz)hdoZcZq}9nOtEf*!YOjw-
z<VZJ;8Z!^mKSZh-$K}OoOb++Po$K>)^V+;8RXFYrp@%pVPaIB<kC@eFmH6QayM$86
zM43iVL>4je!d+0@!OH|0lX37o>=K!snTq+DW^fc6tD4!p!{jOHH6*?d;SIY+;3GMf
zjgm&b`f~SeCa!iJ#>gN?TQ%eMjd|b6GC0^v5_sZ`BLWW&88hT0KLRf%zjBuj;|q4i
zh@Mf7-x^XFj8#Z++EnTsE8zP)q189o=Sg8EhvC5)Q81Bpe%Oiu(acV@=EQ``+ePf|
z?;CDtpwGMY;B^NFM@%oL7;rBS_~C=Of9>pO-@C--*iAOxh~q;h6!*MB9zJ)crMwee
zF2)LLYI@SE`^h)T2(-Y1uC6XCJqO~W58sc`@rl?IU%k~ji_O+aw8$UDdyk_5Sp|NO
z`N6~^lfCzD&c`S3-iq0Y5l^PB?Ht6qo|Z4oCk+v%h|gF4l7|*vC~WQrCN1|hhVyED
z#ffx75XaGs?A&2;uAdz{%1Z>Ct-Tb~Qa_ncALE`jCcZUPIahoU_At?j0F<h@-)hC`
z`nt++%aia8m0bi~b<)%AjlzV_CT>i)Q)j#rM*Kf@lCdm{#s-x3@a-fVJUI*dTsBDu
z(K$TA4y{X#A2jZioyRJI@h$}6V-lY#O#e;Y%3Tcu)^ju&@g5s%Xw0cHBpmEoVn@g9
z<XDVr$T2xOAoqTi`HOhIvaP({i`&=d;{E&A<Dj)3|DS*TU&sIBU;Zl#CJ|^34h+Yw
zJKP;}$2&fL{^?I+M&<qEZ+#MvA3s%@Y{Y-}KmMn20Lef$zw_>Up(OQgH1df&VwV(0
zlR+PR1e3uWWyn#Id#Y>X!5KnHr8#3HGL$r@-jUA7a_6YKL%A{b869Q!9)jorrSp2M
ztu4p$^0HudV`g?SzV*o;MMHJKJHAFW22>g|M(1v+b<xEnG)JD*alglA)W_o^<=46E
zK7umtiW+7z*OS~?&JcVv>5W(pnh?k#$aWt_yoBqFv7qdXm$It13?aOC@eOkyTO$TF
zzT%`1c2+Q#z>W*jiuhTWDtsvrXVP7DT=~ctmUN-aam*}jGiB|NT~)N#wABb*C>NMT
zFXfRZag?2Rs{07MX;UbBPJ5+=doiUI<5%*EFXhU-6%*&BBN)<{l6e>wt)855&y(5+
zl}(?%qVkRa7r_qWVdhXWUb)i}9)k9capgJPr{WD#3-y&>gqwGY4Z5yScfGS%P~;VL
z7x!W1EBAQeha(G_n^{|1Q~qu_P6#DD2O-*Pnv$O68S^#>89Y`l=}w+D8r(&vbane6
zx0)7eTfEC_KzS%LiNq1F?1rA38uxB6dCB@t0xI0Tezi{uoAgE@@SxA|82s!ohI%jJ
z#@tlQstsik1?>a#1T0cu=i~auR=jx0v7%e?@X_N~UR{e`)j>5J3ZqwLXu|z!vpE&l
zuU+?oT;>Q6(j&N2X1m=VQ+qy08zEoJC2VbN#iOTB<K@bV=TA6-mhwrzyska~A^M=o
z(6z<In4D-hU5OWEXK7&}KKb|q3#Dk2sE<6`Y7Y@Eg+iH_Y{cTiY)p=eA5i(GPq?_~
ziLbwU7=QhPAGyxG`_6r(+l1R3<}rE)hhtm)<Cgm9`8l;IAAO{LW7=&WyObuz)HbWW
zpPgOC%eAd|@?tYytnI|==3cCAw%iZzs~@3`-np?9@4tK73p17lyRy-8UEoMRj%X&Y
zcU5OCbW$3#h=L+WUEow67Oqh5sWaqr3aHh7FxNyJU6*n(GkrFKoGs-Sc+B@Gtq1#6
zmWQ~Xo}HN~ET--6@4a+e$nmM?2t?IJh#;pe3q9DGeDlVl(qPKv5T1h_RvgVsTSq(Q
z#Q?55>NC_IZf>hQtG}Z^XIJK!`jrvY589xX>gAEr7e!tVsr+&uT(8O+eLM4Zm*Te$
z6s?2vUZtbz@e$>ec8gtSoU}#zH{^T3=BFC5r*cBR%_*PNg~XKl0Cu8pu^W1eg%>9w
zHpl!J_nGZ-Jfg}zWUHVl9nO6m@`d{RQ}Hav`#aubKnhXU=x5<LDaY`EDDn*DHmrQn
z_*>0D<L(D|#=hx}N<ZE3sj9yH?bdgH^`tqU>V|q%-!WlM+O6c~HrCp=`+rq+<D{MM
zj#sy$k9b`{tDC{M;(n8QoBi9=H^Jlbo8WtMJpTQt*94PN|NPJXv!0$m`?LSjMX}_U
zK4?l}@+}TdSiX8Wn`BX8P<8_decQeuO&>BPZrkgVl^|U!WNL7S1^|GUYA0;k)ycsE
z+Rk`v6B0%t^`896CdHk6_0*nNv`SxoSA1=YRIjE7+kP>7u+&9Q=PNh)HwwaPJK_(k
z#E4>ohmO?I&6nd-@feG(3l2LKdJ24jGA`hp9<r?I*acyq?F_J<KW=~)IDN2b(+yty
z3K!UD%#2m`nXVODX5uC*yWuw>U9nAGG2_nVxl1>$#r5mgy=sMuZ*2H47~N47TnJur
z0){Z;1~aBk3p+Di>I8D|0K=UODooX}7^1@yR?3Ni)5GcqR;@BHBYhaK!XSF!D!;jE
z<S2I3@jrg@I37QI?12fi*TxB2GL3}-0g+P0bcrQ!eGqL787=TCfKD4j<x}<bo3HZK
z)3*OjRrU`GF9)n9O%|jlbTP~)MQMnfvRIRz0<^G7aU&C$2uV@^P$#Q2t}EA4;i@!&
zF;h5>w(!F^dzF_ucSf6Z3JW)`$9o@s5L2^L#&dFb80Q-F?eDcTU{Ob|JhD(y`EY)g
zM=QX&Adn&a<5P~Pyf6dJS<|e*B7T%B7+{zj29e~W2Q11<1eK@92t*Yh(ZfIhyrxsp
z?ibTB+MI|9wsC9l$cm<gxml;nnf!C^_6Y-WF?yt%2i7WMJf75$8K#@nZwv_Ju_V3X
zE;qt49x(>}qLm{JaLr4P8Q?07MLU3`Bu5LKDlFoFnZY?*-F>`|EbvnWf*fhVfCD)1
zk`oTX1ry=IRP>4IB0O8j55>SM3PlS7G?x8POi3S(Dw&^|x8Q4WZc%;*?FI(CTa_>6
z(g~OvUI}D#sl!<DmA#Up*y$&;??I5c?H7}@n4A(0v|uP8{#2xX2}i+OC-`s{FP^w%
zWpEp+rdskSB<bxev)f^5NF%%ZQ(HPzdh{EM;<4v4%co;hk)i2I4;1ynzGYwOP8al@
zP_Rv1=8a8V>0y6u&mFs}uW$P!t&85|$gyBRlL`-f0N1bGm)eX*U(&b0VTqP>$_d;K
zu1X2YxH?9|7`oyYhNLzXw(gfj=&N}F-5CKVL_t*C@}nm$-7nZuFQ+-G;uC%mMU&hK
zy&!YqB(x#HV;7kS7lUzk#mim@Q*}%JGiJpO<YXY3hX*6eFw$6H4X2kZOYp`EHiu_A
z5_iJlPMB$LWXhj*UtC}krG2FxI2?d%_}B#?a4R(gn_zh6TM_A>9@1JrjMkU%I__X3
zzBU5scB`t&kGLh$-$4M(Uw{cKUxG`@6Op5+Udw<||Kk7he@)HrKvlt4ZDZPTE#FK%
z(l5;@^skyz^SXaxng7nW9agmipMJ^vq^fS#;qJ1Qu<dX0i-vx(ZC}15B4PO}vxQ?a
zIQ-(b8@{S)&ySpeF3jXp#f7T&)r_iow}+;=o3>6IT(;Ww8CJ(H-+b&=?Hw0zP)3=Y
zmVK9-zuoY_0&)eHTZy!HdKdfhwFB?`SYx<i0bEIiEAajhXA-H}NwsMKPd9u~w&P#m
zOq-r=e{WOJb=6r0g^wQ|sN^A?Y*vNDwx@SUywP}!P0YmD<g9psQSr*+uf(6RlAmN_
zClC{K><Ga>lg{iCU?Tcbyaa-=yu&ElPOrxH1flpm@pt`~XE8QB6xXjU#77@}7&9}o
zf??MYJHf{MJv$elbS{2L;lYofyn{si(aD)87cT=pPPyfnIPQ?yZS9G-2@43u$46pn
zlB3av#J_XYAS=ZYu$*{;Yjl`N;$FAi@R00u7vrQbFY?0t9qwD%^TgZybTgLbCS$yj
z_s;Zt(UEA7U#Fq?JSON@R#%mNeK9;d7DwVY;R{Ju?{sL4c%lxPEf`a}xjmHGJKlr|
zJ=VOA_V?qUwe9dZYH&_8utV<faL+sy6T0Nb(aE{e^fLDMxjSS(g@odPTYCpicXrQR
zTbPL%c8n>$@T<GqThZFyiHR}MH9F|<*S9kLMi8nE<$|lh0Z!sjUWm`(PLr{r-ne~p
zF=l6GR0c+4`#}D7*?k3{i?Bz$z4C|Sp;+|g#d4~Py_YAx=jPh9Xb=y7qIe&Zx5_K-
z3gNC71dw_3AOgia>M9qRX?$kBJ%vGe*Nn>7@QB9)&?i>b@Ca!4G`2vfJu*akv$KQa
z=r|TFSGkW4k4ldKQ#gh9L}9&LS&g;zZPyuclrhew;ty`}n>0bO3)08a2|wkS-7l0a
z?$@D=_6;Ls8&lciX_)Wd-)B|)j`Kn&Jr1#8Lv(Tkrt6K|@~D2*Spq8K2LuRt)S(he
zv?4HK)jjDz`C%M1uQAI<@85}cZp_9&59KP42i@EeevPkYrbgr4+e@J&h|R5y_<#PZ
z|5tqW*=O<I`ya%>(3p3C4Jl37;q$Yf{Y2y42g)Oj14RbS=5>t;zWoP(9PfYht(cu#
zGH!O%<WYaQgD7_qX?!F2-2H%nB*!_#lieZbqH9uZ#?<7v>mOx>I<~&OtMa1qqjV*G
zm=x|GAb&2r(~F&20|UdVqr=V{cJR#4P3I08jgOe5-%@?$s7!=z+$qLbkmLCHlD-IP
zXxlw5QrePkzSD{^6yr0|gFulx$9wzAcuRGf2+?<(NsmAaA-|7dqb*e2nCK;syklfA
z@0y~1JKmxTf7!OE42f=pJ6=pl*;YG*JL6?9WMpisc7vljIc5*R7vsxX27m=Gb>0gs
zWyg3O;m?3@<ehSvAGGYX|D~L`tr1<spT~J?JP4ZwhY+Ncd+6~PSYdG_9OH7Of;c;H
zlJ_~kf<NX6QUI0n9@MGRQ~EjZXZXY!A(tn>6)s~}j-4GD$+lEP$e)wi={TZ-%8`q+
zv(rAxlrizMm&>uSu_Ju!3eMv}3CpLkD164u2)?|NRAq^|9|U#82chD~XbLv-t|Zb-
z$<a%kSezi5WC5M*{u)!6=D0uZL_#peJQ4R;QkD=_uz-Lz2$~TD_>M)=Lt%0REl18x
zjtwbVYKwMM@8yqlTVLOZFTZ*aKm5s0V|jHwCZ=X$Zs}HxjL*ce+E)Y~eeOG1*qd`G
z{i=twbqH<PB}aNuH}#JJ(XVRa`N%pCX1+@C_O4yg{^a4;@q-`ywNEXYot}=lsit8l
z3yX{F;JT|i*>s+ssvQH*q4HaSTL{YCgv=Sx-;?N1o;-4Xu`B8G&%cOIfBLD)dtZF-
zPydPgz#+w-BOU4Mrq!n(E2!1Ab?@q0TAYubz9Dh)Ju%&!P@4>YpuRxmWVdw?+bwo_
zsSi6mi`A`GJYU&}HRU<!K0Y=YcW+&bMU@GJm-Ns3Dw`}kqAi3*?$tU{ePM^X@9v~M
zVZKB8L_O!YL?XdXW6CVY>5{JWuPWKT{}3TLb9nTrtD8Hqq5R|6R*EO}bC~`{g^<3g
z&NGn*q)E<4(MM1>q>&jr+Q~_^c?%13u^{@{@iwi#IPdCI+n~%GVzEofN9qg$bIMin
zEb7N)&%7l^$FhqI!94f}g&QI5&Bd9R7vI!SI&)VZyCrGY_tg%w(~P#3vPC_FHzF=(
z*r}h=KY08szW(~L@E>>={3jpXk9Y3fGGDN|zU6#qiFac6UntLddIuFopZWr|Q%Y|*
zN6I7j#%`><jHA6>+cExy$3lmm`be7CL}M?785`U4FYfuxG<3V9r(>UrdB>`**k8e+
zZ@0ett0&F*R5w(&zRvqr1wCygH`ThLz3QKb{H*3~W%r_|yu2DlRc{x3y#@X)_NZ@i
zL;WiEZUy9T1pmf({QFT?@~5a2eEk>yC37XpE|@9+sRQa5Z6{SMaZhr|q%>isV@p6O
z>6y4N5-HrQlNBH}ZH$s{+7^4gTrfJgyIutWFZ8nOwl43i!01XLD`2MFnuJTcROnH5
zu|v}1APRiOF?~3Q1atvK*RjJ^zpilXPa4vcZ$fn`TL{)T+A;UL($JmO1zlSacUH%h
z;H?bX)+8-ewcV`Jt-97YKudXLcx`}cAddP}WU_06=z*pyd01djIjJGN{)wYU1MBfv
zym2jV+*IeUK@1AtI9mzOlM3Ka`8o#&cKVvBQWwfjA3A)}ivclPrp+9}bh7$DdOEWs
zm`Hu;;86&^vXj$5M4cE6CY>dNHdeT@Mf>IR<#_n$QLL@572%1=w{mo(sLU>BL`v?6
zJ}3wL9FdZqR@SEp@RBqJv#OxV*W4P+0E_sxA$~mt-^<6HLNdVM@}#5`DTf=4RvR}o
z>#+cf^pqPNle%G(#S3Pe`1$TGa4Qk)U9^Iaa_k0Nowz$Hb?6Acz_W1kX1w#kyWXzP
z#k~g_pr5M4gt6ur1rH#_T(T34JJ?_p$<w40MucNK&fRI}ks@M#MI-r4(}Q9_>U+Xu
z=iPbwGr%wdEV^^Wnqov6GiYPbKi!Px{JhHHjhLUGF@ExZ0RfX1oJCKWVSwy1qH^Tb
zdIYF+s&{q4jMY_i2w+KLmjm)op#7vrY4{mXFqo&Tvto|&?6R%+7?;8!ZHeEY2Jaln
z2oCHi!!R*0=^kuIj~~KQN|m;PHVVO4uC8H9#fc<*JkE0%tPApxphWv6m$*;G&FeSg
z`n8*4#>6DLjN?at!4ZaSXm^SwsEh;~41}Mr49+g30Xn@Cg&nJ3;EX}9^Wq1EpQ=v?
zy@{Xzs|%Pj*i2{ajn^<$p-URvuG-jvPd7f&S5HY?G0WX?PIXoIQPR_x3Z9}2BTsx>
zw`xQy6vpbIgbfZG3mhL=TXpNJ%2-z1#jorTGT7*suy)g2{P-((cv+0bU97Fmgi6v1
zZX9bY(>32U?3xM{x_UE84AR&NEi|_wrEte6YvZxL9fH3U$?kNun>3w#Ezl)jlA1yl
z9*Rm^Yf@LzD}85sX~l;C^{?$r0k6OghhW9`8~kd2PRPRNL`-;I<&#SGE?@14>09W}
zawBWG+tn~>27%FvFa*8;yq2WkGd<GVzTiY_!1MqQ&_=DNgyuipN0O$O=}17o0WqDa
zx1#SN4B0uq2*2v>TZRfxyltbm`zh?(p{D}B5MamacPaS#&;A$x`wX^ZK517~jR)Uq
z@2}%k^tSBMOQCPWclphJTB-})?(Rd1x!6;8jK^gw`(eA)FFx#E&2RV;C+Q*ALNDy_
zQwzF2-S+-gyf*Sm{tG+uIHO)*pxdzN7ghD<;h4-Vu+=WVG8<R7TYgcp6>eaQJAXfU
z?*Rt?__p&6c+xCsMR~%n=s|*;OP;OxDs_pQEfSwyy1vBu3f)&!;w)M{iHkaF;~Vg$
zODdcB>uXzmlZAMlmRD8!?ZBmyJ_)vcwozG*?ABJY6VDZW!!bC9aBDt>Mkd7{=fW`f
z5O_K!E~loZVze>tw$XwL_$zfN@Fu<76CpT`bY}6;=@CMVW3e9MZ%9`?2!s|F7h+*{
z#`jd*zB%t5DjX}wQAzCf8X6fl&*O=wp@HD|I__;kuwxvfDQtlMf)6`BhHK49!_kq!
zm>BMlK{d`CpR~TdBU&##5$ef(@`AdgHj9Nx=6{s}OrS9-zr4H}J6kPJFf}I{;k&`)
zmt%lNhPcn7-#eO^SmR!+7tf!2;*I+|IQoa3Oic1Jp*lF&C;n|emR~#<?rp<x;>Epp
z-ir~X;UWAT@(mRk;)~$uO#aW#SbTL5M|)dwwzuouJ=YfIV{l|7#wMquS3Kgu(V50A
z8UvBm<dgD(^kLGF;lKnFxswg=h9leA&2g?g;kc=ZQSqQkNAeEg$3bh?ZcL1ij3`}&
zzi&w2!QtpRKa7D(cII(Rnt0fuMhuNlN00bL@5q2JRXW0RGLC{jWwF-A##$T;@6qmN
zJbn7m;UKu`8(>G)gwnO?H0T=|Rk;&yKfs+}11^hhrlf;^X5!bv1t#|u4|Xju0orOE
zS?GcJvIqWnAjXwX2)g=u*&(KU8W{?9jKD9mdyKk0s`50bvbwRk8!uNjaz_`t3`INX
zPkCT>3cKydcg77YGNY}bt)O}%RAHwR<HsZAIc+m#f{9%&`ezJd0gTE8u#8<9r*ORD
z_U>-dtvq9whL5Wg?E{KG!Vdf)Xt}r`El*wlEkscK&*1w-3qqBT-n$+D@H-#G%w+Bu
zVnULMQ^r5TgS|1=RJk1#ekI$pr%&Ur{`#-t=B?Xt@7)h$?+8J1=Ku4jkK?PazKBN;
zzK$nP9>-2=TPYtgG}?@R@DINm|H*&)KaBB-8R6ErOZ3r}kr8!~q}vn`$#{z~9OLMf
z<>h$({6%bSZ(2y$-*@R<Fifh?FWt;tAQu|riH>894ZTy0@s{GtF^L?*#w4@zLxHd}
ziDS>^W;Ax2YO0*{tNe3soW}WzxA7pDQd*-Z+u+0OqYXtVdBQPcOnNhhWT$EhIfq?e
zd)3x3IqnH!h3WoQKB0lK<HfZ~e|AP8*r4nK;FC!7C0{&Qt;g+_@`z*gSSY|4lQ|?N
zj4d2Qu&cCZVjIELsM=K)EgqdxW?0NgIAt!QgiG31VHM*%!k`>@_b6?z>yAhgE#9?e
zz4|jf{8D&ES=HE-a;LZhW888}-kh?{cr+JV5<zTPfLa$8qnz8!37GHGUw(6bAj3pY
ze5t?GWoY+KbJ58?T+rO#kKm7bt#(+ri2u08=Qo$;;>O~<+q9?8U#iaRxbAa}0p*YJ
zyyGm~wEZkv@V$^KFQh4Hh9XU9o7f%1v4>-g5%mEaAqPzmMeMojXh7}ZaEz<17#~sl
zF84jvf#nx36y90PPECrY1q*HvG;&{I>+n=<0-;=5fHtIhHl)6fd%U<i)5jNXY^sgg
zw?Ko%7aU_ZGBF*alXJ0uav58#0}H-Lb_BcJLx}*C9ZuPXQSOz`iua(tlw0m$8dN$F
ze+%+ej#!}0LV6ZhAXM4k-Hr!eeI1+Yo5C~T#S$E~$PUMag@tG`AEWkeV{P4S5^WE6
z03wuajE(zlC&GI2?0Nj)M?dxg7mh+45q)2M{UA10*VJEMi@SI4x($2h-tD-yG^?_C
zoPuSwol5+;P#xh)3uT^#%k)KqEJ#uIaHrz#cFT)CMn;vcgCnuCe-h7EH{;pzW^AcV
zW)28}%%pI#=!A5Hf1?a>Paoxy<6UXPus_p7n_wX<Wl-h;>ZZz$^KAdX<&d3J+#xwP
zGwDSVvdr?MxNPq3#mf4Y!{0xkFW@NQA?FL_j-7tAX9$iEh<PDHnSWxZ9OX|0x;{{5
zNq_FJo0u4hxoO(3Ngv@jH8JM<GaD*%v`cEF6dQO&7O7-;X$d!iCqMfNN9_W0pJVdF
zafORB6G|J>fiyvgKc=!XsPdfxRC;9k?1O_tAL*Nm{Lf-wcr?y?hT`eV^?13u?YI(;
zg_-gA_J?=l_O0v2v8T8ms+=FIep%=xY^34x%9;h6kj>(dql2B;T3a>0M;}EyVjhq@
zP#)4o)VAgCBKWFqRk@uhs@v@<4zRqzj<23H=kqE|Roz0V4t!BElUu313dfWD^&8xZ
z9!2{6W`ZwNhu&`WCcCP>$qn_Z+`AQ!zY+Xf<MH>e-X{3!``(}a>GvVp&JeG>!JI$r
z^b|Yt&2Ht<%bk0g6V#ZBSv5WF>Pu!UY+GOwA_o~2&bX}ruuHn(--%Z2$~T{4rw<Jb
zR)&Cv0>Y>Lb~tIw@Vepb6Nge*i#uPmR;gi<nLn(6ZSa;lD`O(8OqTXjq1sPc+BS9A
z`E8B;B}sOta+v+lpS-K~X_fhzX8aca1;}Q*K@{c!ml*9{C*Tq;-7+O?V;JLWn@bL$
zNh{l7nc?AzQkcp^S8hEt@I&h0C&c`+3TyyIjeJHCe$i)}k)2#^Ko~ffnN#PN6B#mN
zr_L6J4tMXW5%Y7V2A+X7jH!=16hn+2co;mdymTijCQ6+hOdU+qt_CD)Ya6k-x#^Cq
zhVAIoD4E1Uv*Sj?j5DtSSQ7)jkuqHUqGZIiXcN?WF->&l!e=(MhW#6;3>JS$SB3y(
znFn5-hn>=nw9Rgoa$!ocKKUu(7aE`kd@!28VaAX10Z&ry(2BbSSBeV*ZjJ)sm;*XQ
zghRPvE*Wj7z08#vCvV9&9}OeC3{t_%z#uR2%^hEmWm+oT*vSva1u`JXG6Fs(wIGfZ
zP#j<$5qzoq%`D8v!h9aNLq_F5fpdughZqE<Iq)Dy2*b>~;Ru=mCHa*?tsJ01i?Ee)
z3hm&6VWK>tU@~Fcm>8l=W+$Hbebfuwu84sC94siGgbNs#59z#W4;uk_lFen1IAWJ|
zgEWMeN{{ax5Ij%`3kHH>qG&WG$D3Xe>2{(hXdr#_4|csBNA&3*JvwX%%;_v`3XM!9
zz7nF8mFi->tbwQ&iO)91;o^yX+9X)}#e<97Xg>@i|FTAB>uxAJm#=MX%x_xd*V^jd
z`6hsSfz*SUuquoVrGN92FgD}g;3W@;TTPGr7`m;ho&4E9K9N?n(<4>7@dLAP3#O9V
z&Pl=oQlEmh_$kU^6jS??kXZ(#WN$hh4?**>wSF=?I>5=2ul*N75-wl<+D+8qP?Td(
z6QU@`y?9Q4rEa2tRiMS$LD{9e3<jOnr=#1zYL-v?m9`TqIQ5W+e6UhD_D3AkcNe^r
zfeNtb$+jb5@tXo;L+Byhgc~QB{Wu=dR4CJIhcu)08!PFmHClb!c`S2=nINeMQ9u`!
ze(+O(umi57uU|zuT<}&p%9F3QB3!z+>g)&f(sbg4eMzyB4ylSB7=`DfH!)k0W`IK*
z0GNB?hd=q<-SPGNfAZh|e!CAWs_LDFe7oKCd6T|vCuMkkuex=?quTCWuqgCxH-9k~
zef<_c*4g})FMWie(wzNe_1pPY_xeq(a-BX$e#qA=u-$kP*sAK2pQF@N_YR!D)}}eX
z)r}HU%=CrD4|LdGg(vayO*g*XVenA2rL$eNx87f=($D%=-K@tkzb>27P#0_135R{^
z+wQm(5b8zZ%x}8aGM>tM{`z8TIO$yPe;w69^_ohI(87<}tFJ4oL$VtRUV%wx?t{un
z`GKs9;-y#!2A{-amnVJImYIi!_h#|faZU>PXgNJl?Unu`V=>y8h%qj;pPGts@&1$r
zCXqZzEWU#~hK9xGSy-lUrkYbuzk@w?FtMtf<E^<s{UkV*VVlVsgfPA0i9~1e_e|dA
zL^eW?0gkBBk4ZA_7va9F<^*?FaO~Qk`6-UR%GKjNp8VvZ_?hvMXbk1*d9!1Bkr(i$
zO#J#n`lCD=ipe-1Nu@ZjGlz+~!`<!p;a~sT_|f0|WqkIFpT-aW_HW_`fAhET)q^Lo
zf50gLDIhyn`d>z$(~ik*1XG+gfUtvIC3`ztv9bCxF3*l)VRqK5|CvziSNzT|`(kI0
zU0&=S>5uWzk(d_ki?b84Fg+ek#gQF!Odhi%;_&1=cJ_~AeS0s~w)SFqZ95)4eIAeX
zumFww#hMdMPw25T$^7c&x$lTMXzi$&<eeR))z0=#{Pkb`W&Gv8`LE;0fBRSQ#b=+!
z$zdyoG`wJCJAy&@TH?5~cNiNx2Nn*&2d}TMN59I&%^TNzvHr>7f#SItyBn*qv-UDh
z_qL+<^dN@&&ONbr`x+C@4No9ueNmpP4ECwq4)^ydUq=jgYjIj}?vK{S%Xs|N=U$0^
zZE3-RD1-?6J__;36Un3v6Wmkd!(J`#zD{|};=R1%Emz5tuh7iyp}qaR;Jz*Lim?pk
znxokeR#7$(^zH4nJcexP%h-jZE;-_l$w*fEv)h0&O`b63WP+5VrN+vUiHwVzlgb0Z
zGigs+vr8&>&v5i%zr#9X0#o%2q4JQ*-J}I8dDJ5HiX9QTLyjZjk{?I-^7Vs<%J;RH
zUsw_yN-Nbrj=^e-DgAEVjyrem#r119;^wWpdhX=WeTRo}@811rPE9NBxoeGGHq>1v
znGyJMyq}K&R5QZRjibNNbL=ViJ8y4o#V>yL<M`#LKaM9)9;l7jk12M(%+2{YGR8^X
z5vOz}jgQ#vAU*NS#jZJ?%?V=0g<0ov64foF^NBQ-v*AcH+5z$kILb8wcaA1wHw|@P
z#Yl9k?y1c|SclM^BP!7!(2fZJevJi}?A*#D{#4JE7qlOI$)h?EYu=AMNQ!NTh1;ZZ
zETWIJn`*{{zU<;-Y->;Ydl4n$S}Ee=(`$&A+csg<*X^?fVhG1hlqdQ!@1e9}@|*UW
za*{%+>|-d49!Cl<<toco?%E=}nzocR1dP5*8o`R}h!e*#vU3XI1v`Qq2JI(lF`zo2
zv>qNEvs*)DnTJKq2oovG)D;U!l{p^YGs!Q!q$PMq%CUsdfuNWjcF@M6RCW@vOSfNb
zxrI2&KLk1KPGvVwlY6w(=5Vwe$8{nYottgOydLh{T9`|_`T1!JuCg6byoj69$h*D<
zXxAt=$@fAdLR5C|UaFm?u6XW<LQUHvzbukHIowkLIud>CNM*OI(wp)&s4_h`78iYL
zr~60Za!76B#9R!FPRH@(K<pl##{QAgL7|aG%!P4p7Xl}aBqTlvMy4l6eGk+8#7NAl
zt!j=AL~~@oM=Q3rHsa?$`$>HE=}*;muB+0~XP?CLCr{(?<Hw@^O#SRsjOythX~ekN
zms!=f=FF7ho&s3P0E<}YSJ+L&d<<;~<>O~R|3&=mPkyEjd?c=^J(!tajN57--v8i(
z_`^T={rICl_`TqY2+IG~=0*%FU9T<7so$P(|HR_mGxcq|`^R3;K-nMSs7B@E=_!k!
zMX!?6HKlLB!q8#mGshIO(1pbg91}aCyk^H8>Rjc9`c1x#kB=&!r=7>sQ_M3BTkwd`
zmb$?_&hUV6U#KrVL+G_1E34~1a`Nf37x8j=Rc-Yi#n5q{no|3#et$xJ1A<WUo4QAy
zBXC8ih)|X|a337yka9Mo^2yP!;PH#HrLxQ7j#2SFESw--lo>CUDE+b9Fn+S9tdo9}
zO?WPjF&q~i)01P~*~PB&CB<_|Wtsk+JJM!U?#D*h1<6TPEZ#p+oKq;uahT`F`>~^Z
z*%dE>J4N?UZR+~wwvR1jvCZV<L|j{(a~wFH@$mRS{k6)08VdTg_HNY+xvA~nZSAXE
zu|x5~0$zAT7R+#DqUYVn9MXX{On=8vG}Q;y{Y7Up|10hZPdW9kxK-6{ev`ia?bdgH
z^`tqUS7EB^HoqoRk(**$wB2x5>3z+u=u!CZcJEeivzN;o-QEuW&2Fe)<=(A;{Egs$
z1K8hIy(uj2`0D%9@BK&Lmn#ryt8eN`Cfb}6dvu72C}9j?v$HGhS?+8lv#ofgHuc56
z?FVaBhXk;V-!4<>-iE77sjU(+t5P+~F%43Q%ge9Z5d45C;n+IuD8q><Ve<=J1=gnO
zH({vD1xXGVakm0fnG6-)1d}fhYGg8wLYMu^L>kE5>FY;WY|bbcz5Su*jWH?aZH#tL
zo5Ufz8jKO3&u{w4WNtgvw>$w>@uCAJ-K7c+EE1ypiWx=hQ^&gwv1OtRVNe(*O-#Xv
zm_1fOal`<tw$f)3ZdkwYB0r8$V)VdU%m$MyX5{q1=+fYy!61tv$&Q8X&bfOA#u1?e
ze%<)%O9zz9Bf<!dK*9df->t3fSY3V@t)1P%a4Y;IEfrrUA_YheGS{+TSQJKf)esLS
zQn4=zMw0ZpVv8ZsIL^@d8!Lho4(5MMnhb=J>6_s>M#2rP*r!W*n;_f@IkJX1ta$)Z
ze8C)g@|J+?6(&;5p_mSocYy*a1TauA#jIw6DI~8B5BA(ik_SHf{OCYwUsjh9hV4ZM
zNg$$NI#{t;SGV=CVpnk93B{wlXSbM%mO?1x)h?pR=_s;<(q)?6Ufff_F^BVui(=3l
z#!opVb}m1njk16c4F=eZiQ+-I!ybGv7{r<K$Do9`5I%I};46h*q@g|NPrPiDJzR(d
zRJI|lk`ZSWpN~Nl*|oevm+O}j-UAhxszBHi^72P~tDbbPbsQK3f)<pK7E%l@T>hX}
zcF;H`#&}8tTLvg7mrxwQ$*-g>dj7)5$7bk5Z-@354C%>bMP}MB!fg8ym}yNYoAYJp
zwx7hB@7hLID3oDAjNbSXw`~fnNU+&@;Q88nwE-}{ZdbvRPWD*c69zikMkyiB(Db%N
zc}pb#8vR6B+bo4v-kcZG<HXmr@dIrEO5ebwK9%fBdZa3v^iu5gD%Pg37`xMhcjjTo
z2Gx!Wet59+3o{xlyL?tc{56n`hA0VU7Xv%i`T|k?)lX{h4lLk6Cpx_-jKtf<N@e=9
zDgPaKDbp=w7`tNu<^oaNvGX($cx4AZ2ZswZclpK6I@7?fxEmfj>RGn}T1xKZV>M@Z
zwG<Tu)&p0{N6{CjiYei(h?h);DY*49c<~2pdQUfCNerhAhLWZZgEoq<?D5wIse1j2
zbH1!Wub9gLno%`vGClH(?+&I^1zUz)@TH(GT`~XY$G<zl*YEwwfBgNTReX|)N=-Fe
zNo)IJ)$LZ>ZhmgB_3Kt`ysvyw=wGwPu6*?>FxCC-@NIkgW&U)m6^4{|Uu4Kvz9^+V
zdK+bL9N^8bbF{kKD$V)O-=qV=5&ydE?GGK_qUW3KNQEu-s6uzQzr-a)-zI;ygGQ-N
zIOVIKQCrn8)4iay9erEcr+Yp)@y+xg-ib<fc21bYpK<7>6FYt#bC;@cxEH6|mSRNL
zZd<=Wy@t1~(y#FpU8nBpTeLM+tdd6H!Rl>xGx`1{wS8U`14V*|=RshFBDH)35WF-(
zEAe>sPF>19JU=V@eI%Ut3wB-2EzHO5TX)3Mh@WB7YiKBDrkZhfydPWZ>*n*wQWiT6
zBIJ=CUY-d`j>BTN+29bnx0rNIzL4EjKH>;Po@Q*^j`YCCiT^?n1P+8axG~8jE>7*0
z;KfIBcM8W<AwX)33<f)Rh%<r=1Y}H@(_X@djf^q*IvjoGZ%1Qhy6K5kCW1ECSA-J^
z&V5e}M`KM)FT}#NJ2A6(GkOO`Vr{GCV+MCx`?0yX9h)0lUU0|6J3<{+QZos-Ff$V$
zfAnG8yL;ak%`-trc{n?T9~J)$ZzsB$P#hiX^)teHLXHU+57{TanF&00O)M|3#giAS
z@!;vpc)q;uJ4N6j7R3wSxqUk(CYy06TDNyN!ZH{BjVgVayxUMb*u{WwVs&LTo;`aU
zCrA6rk5T2-nNM**Ahx`+5??=j6w7OyieD=(`-Wn2W<IW6zZrM$-iiBqe(&4gif?`E
zlel&3M$FI7D2~&yG(YRRMh2B;t<6=HgO_n}e4splmseVfmh-aviAlP+IEnoorPt1S
zynOOi{PNSkjn97aH!-4g;I24N!cLC6Pl3<v?d@0Ivm<NJlbcgxgR!G<Hr8H>hAHnr
z9G@7ChQ_Dth(I{YnC0kj-$xd5U(Sxkh3w$iYwdZ=QcrPU3_=>@A~nVY)FI06pm!7y
zCc7;rC&qnW6?y1!k-}m;JlYuZ0z4lHsB(mWhH;4njrwwzR0<L}=8>@oZC375I66F1
z8VyFX$?={8rh`*W2%pJWgvpE#5Yqhg(@*2y{Ka1=uGeCEdRFE7&^rkb;!@wZ+lw+V
zGq(`8Zr_bNckX5#((nGgcT}$teC08M9Bat&VeGbH*AhZb1lG_*xrARDR=Jr{I2>t8
z{bcO7wYla=@b^Fb(1I`8Yj&(4`1M%c<N3Tpje5ooo17f(vyg=Q-6+$HlMuX+HWo|^
z59!WvgdW#}M|ne9kta-+ldj|gcfVQi&A5(orMAI>4$*-SkFtNpxKecsUJZd4PaRXT
za1#Z62#df&8lz~2>4H*N#nEUSjY%3GI*fLor6LSYXkwC|CZ2JccS#gH?6C6jiU_}$
zgqJ%R<@pQgiJt{zltY)nJPv<EWzI+Zkxqp+zJ$|I+S3NJ_>!X)IjWAb;>mZ7)42nV
zI)HGAcA1@62(QL19GFrWZF(UqWtT9R#{nm1{Bv|H!W{C2ug9#6OV!fSZq;#W&JA$?
zooJPR9~G*0y`k{AYnpTz>c3Q*Iu?ubGr`W=QRo-^)U;@soKTq`iwV`8Y1Q+&nW>na
zK}e=}_FwoOqbZKhRa-dT7?S@nrPGPps4WY0&s9DU5^x;fIBmV^mHrk$(C!j9gp(}V
zh~7ch5svz#|6ykq@fuXR`WRK^A^jBBU$B#MZR;R5_fF&JqCYMN)jke1;`nkfj?Otv
z2f-L=q_kF9r2RyI(buE=JVD?~`+pH*ibrF(HyX6*N{8W|vlx{9fG{A~;K<q^{LK%%
z0D=0$BHn4G2?d8;vyF*nOseip&o9Jeb4JgE@@`!D9<jT-<GT%+-{8uLFTVKF$MsS#
zX+Jj94{hwUV(G^1_=E5KaokkjGd(+}dZ+$aZ4C=Twl>$}@uNrTuUbk+gsCh9?sGXD
zA6FZv{5(3ljDwR4-~Gpe0@`2Na*o0EQI3s?7*Ks?Hi7cY;uCgj(?`=Mvv@`o*kzDA
z^|<fMcjl?Qy1Wa|06c-p>ZJ08vS2|I<wSLSrpfNAh6P!aCtr!7d}nv*!NIZmxm^pp
zDHj~y%3Wr&&2cZ@@X?dXW6zhULL%&9(ahxdu!Wo~o<Z1p>-u8cxW1sW$Iea`ePr7N
z4IF#hRGM>q=iJ;(Osg+v0l^sji_(&MiJ*?tk1XsLeFLf+EY9G5PRhXKNPjfde~gdn
ztA4GixU#5&U44%p<_=xzKkeAw-fpa~iuX|aFs8nMxlSL2O209zayB`od|YrlPnrTV
zJ2e(J7pLRK(rip|lrj0?qeP_Co^a&s^2&NVeEcjPJ%8b2ky*@feM$YLXklT@j(7{^
zP?asoQnvY3=B5}ItIl^lDC^6&+fTwU8m=%M`>NigfAx#9Z<}BJ7PmrMRd3T<y<*Ne
z!=iu1-#6jx#)AKD_iptzd%3*P?d|a2?1uVP?%fK=-w6IUfc<UNo5Diy)$`|n{=ZdP
zm+V5P1~dgQ(mNM$D29rJ4{mMu`q5hrvHevC0a@UK(Hv*mrhD5Um1YNF;EqZt3P`6S
zp-FSXUiFbE>V!|e^W|t1JzxB%x!_KZNh*3ZC{>VwkJK8HDqQd+3~6VJ;!b2T%6e54
zMLFCWj&UTNR&6s-Qq`fIDaC(r@WgD1SUTg@r7DvACmt}g<E1w^tmB=N*vXF7aMC@Y
zQ`ts;$Il=@pfKZ>{^|g{8{R)Pj4-M*>WG(aT=%_CL*?>o<I*1@95v|~G@B6wnJi)U
zI8PY{5a9xY430%$aGtw}&b^A{@Q9TwIq3FT@u%u^IX=OI0i}boF$FNm<eH&W5L}e{
z<muCR_~4NTIK%@+lIT_1C<6Xsu5q>=J2ZS{a9X41g*iW%VK_-+aAVbWL)GsaRISuC
zecz~Ru!LiH2u^qg#}2A^Hv_64w)~tAR%xQM2a#Pp$01uK9>)=+;45)*Hn@-1BVpwU
z9}GZP^)f03d2VSwZr-}(-CrCTxV^n@MveG$PZjqLU?zT)Yj!G;caX#Od>>80S?^*n
zS$W6ghL6$FfB+_@z4}=6oQpB#-Y^$MDcs0Hn5rJpJt5|}sX^G>wWYXri#z5<+<_4`
zD@mES;d0JA=1}!Qc-(Od`LW6ss~m|qLM#|5^b9^O80e8g(2cczE*ip{gQDIPq{-cP
zq4kLSB>tj{yTAmN=_iFB??@C*kYMk$5u<{@jj~1f)B!xgm}Sm6z|=t-TGcrZ)D&(O
zOD21iSH>a#=fW%67U!<Td-p#O6F!;Y$(=Ak0xo|12?;zG(1nEqDPJ7;3O;@Et}5M}
zE;e6v%ND4*cHX90KjM|);gEQ6=~kH1Uky9mZPqK>ytFf<?b2M~oqdj5$3H4<Fk_FZ
z`ut}4%D>Yt{bdu7dZ#z;lr>uw7{9NEMfzpj!Iy6J-ARx1Wd&|%eie<LwpDpmJ2cRj
zDVedYcv1nNzk$ddL$p=;k>3PMeX&E;!Mht<1IfJ$1zN1zG#B_bjNPgu&PCs`l9IuA
z)586m^bRgvtw?u!t9}heyH`qG!ACpDq_05?gP+8gmbS2e%-!v9yHBpF!dkCXF%81q
zw#!e&OUg09j6Dkn_yLmOdKov4;4MZp5K5ty8&<U0OuATMkAEQnpTXrZon;3ueG5t)
zyLyY7lSx}u=^GCfCp6J6*tVS}`K}88R4je$pjV}<9XKknJ7nX>z4(**T^@D#KmMQp
z?-Ui$<Me71*p45{ZY8m+eYZPp!mIG9DIfg5!5@0}e}vrau7{74u8sRDjBUEiSai(b
zEN04BDIaY(e(Do|u6_(>O@7S}xg6Vr?fozMN|(PjD~K8f`Dyp!zbN-Z1y6bTTX7Z>
z!P;MjZS4M<9`b>3oFBk+#nooMsX90+jYeGDcz5W#>L2=5Dy5}#W*FpE!d6xNqMTQi
zCfB!a_lj4_Fjp0B#Yf^PajL3298@ifsk(4WW!uy3p0zr|xN4u_;0hLax?%BMV==~#
z@Zm=Ea3R0=m!6)17#I{^Hau($98YG!mv}DHhe=|1tK5Y!D80<$h1hw)#OdtptoV&_
z;ZZw$aS{v7p}4m=9(`vA;>#Mbe|#Bxr<XA_%<(-#DWvMV42~*<yyLU*WF(XD2!9Zu
zG3m=Vfyo*su{q9elB2Ce4-@uGwl>Cwf>R)NHaATt6ICo|;s`>79ZU{0xyM8~llhxF
ztyo*%h=zEhv$L~UU0sct=0sdun9ZGQq!T=##wJf6KZ>W19tTGt-ne-?)^_*f%O}r#
z@6wUtqSO%W@OMnMv#MG&i>KwzA}02!3j^#t8$$py;v<pfIUZ?hLPOKD*xBB+KyGFV
zKCNGLU&I0YFPxe9IQ`8RGtqvmw1L+f7W~l2u=v`Oc=UWZzI^;LHbobAHnFQ?a%{{i
z$!}hp4_2J7Z){rtM|=otX>QgN&+J0mKVT=*smiP96hF=d{OnjdJmO;ho!Hsqj;{*~
z`q+s_91%()5FGTRawxnJPcAdD$t3jir%yB{*mAxhV0!xGaV#${M{94-abiM!$dl^h
zieKLA#f0sF`0>3~%aeust=%{f{LIvJ{Ez;>?EUAj=392(iLE+0=Ty0F)y=8<b~l@$
z$!1d&MN(mA40vGJhB5FP!!T_453t0c#;^gyHvHZi7!pQfL$WxQ#^dn}M`0-DX7}xz
z`{s(L>g1fGzn|~g`+1&os_OQ=B-NuAUeELU?X>pVdH>elYlnaT-;Zy8^IP6^axDIX
zxWGyus;<L}bBr02iUU0zv9Z1!AARtF?@zgW^_jSG<5}@&U9q`;><R2G3nkYr#9Q0g
z@`Nd48|pRX#pEcH_e?HQuc<dQ2I7a%qN}^Zlke;#1ecvNxkD`P&f}Q&Bk5UT?;Q}T
z$K(zDEOl;iVJX%(Hr##?46t+H!iD45*w~Cs#b<waH`-N~Zd{v+ufK9DZd{qN`~1R6
zZ0#JX?hMDpapgS|$*LRw&)@mS@&EWo|2V$)y<dr6`?cSUgVT1^O^!G_a-5h{W=xDA
zmNqpuI;uL$Jvf_j^~%+lnz?9U7srRP)A0BNS0&k<eD;sfV%HA4rWi+gyerz&-~HWn
zrE??L!Ey7}^D%w#vV~>b)wQ&|={tXRieQZ~JADQds<aozIX$Xt2=n>}`<2(UQ+9@E
zJgEF)JkQ-YOlC8g?-M)}uE$?$%RWM`F9n-pqXWw4F4gH%=Oy<9vWt{symt4QL{?ob
zVWA?#q3oELr+?%AHg?Rwqh^_NJf+5tsvj%@Kp2I<i^+L^Q`M(}N8f_b$^8p`qx>@|
zPhY7Z{Vh3=A1@wIpJV}}%EOcViYxsUf9`WbkU+XU=0KRDvSWt~i%f^vSv#z9S2z?+
zg7Odng50UY+)cQ-zUF=#e%QOWgx}wnI}^DF5Mes$MA*wQZnR~_%!Eb!e3Ym1D|h>q
zV;-4LQM@u-y-pNY@}megJ+@U`I1#3&yVnZ{EC5zMBOGPtX`lL*9#8yVh>-y%{SglJ
zxQ=m*EJyFEa#+BC(344ggfUE1%+5ZDKmF4`iuw747#$snThBfh&p!LS1x^d9*Bcu<
zu9J#b>}_vYm~@?czNe?$H&a==MVI3R+1>erAH5s14;Nx`@?zY)aXoHczpQ*>@hQ7b
z^QcAAzpeH_+u%MX`oUF&v!s57yDGWM9>G}d`e8TOf$I);at-!!WS-KXaM_)SecnmP
z!UvU!zIokk`kpF(yHroOlj^h2KZ}n(yBk{yvtQq!i!;-4d1fZYhPX$My3n5SQh9h#
zAoZAf#MK{r+v;B*y1!-@*nskOWm)~j*0y(?F=T5~e>OIKDcd_IA*o$_`uXkn;hR5-
zt5>hYcYgk7l}8t1e)ewkDBoGcK%IJc|89Kp;RmYoNAXMF{chZP>E+noJC28otL|3@
zl%AoUcJInXAb<J#GjS?@V!QOW)E9(~M44mumR2@md1X`Oa1w3mH$&g}6P4q>`t)tN
z)5mP@Zo9lD$H!xQY|H{%svSpy(!P!k_hWwUk@~D{Y8|b>3r-dm*QBnh?>&hTwH1!n
zeD=nrxH>Hws)LjzJEIrZIP#FZJMiw{arLPzKA`?n@5Yq>2*lV`MtwoZ_h4b&ca?E;
zBYiymH~mPST%@u^*s69m+}G(l7~4+NS0HRxJ%T65dqKNZ=i7bkrK*6B7yRJk&*H~=
zcGTvkC#Pa+YSNns28T!VJ?d0Biz(7GHaws{ppWLFHpvmb*=JK7`uhh{K&QU!h=on+
zXSwr?MM?X|Dw{pk<(&hi=S2JieT}{+OKV;*^Kgkia^LA5R>V6_RX0|by#pA4m!$qQ
z=Cbr_6@1yIx?xvU3cTnOr-@r}fQHJ)tiWS#RmCTI<sEj_9%$kcS!)fpuWo5!Jz#nz
zo&Y-f>{hzKlfnV`He#+e=<Pp^4XA#|OY+Y8;y2;EkA&8+&cheE0YAz8JTzkRW&9PC
zuNuZzg{1#@7{24{Z~t$8i!5w7yQ1nx1M$R-JfkGw)~sO0k;;P}#ENTU(Rh3(H3wSI
zRsqUGB?~UtIjN?<!o-faRM%EBJ`?weF_*r?8I*Kt#1(6RI6Y=pUT4$Lk{>3--<W=r
z8^#bptBo}3ZEeQYxhA_pIza~fi6!s{v&I?KE5PN?R>fa);`1zk)yTU-qlT3Jil6Eg
zIxWcf&otKfJLr5>f=Q^*Qk-$JRnZuP<bTbN;v}3(iV1RTv4=6@gFG!J<MOp@K6;^(
z%c~WS%+~^rFZg1Q^wG${xwr=yV0G#+(-wZo-mAl8=3X!KFo}B>e6f`mhD(oBR%ji;
zJhH+`y*~^kgF`-4UX>-4kKluQ_ry3o(!d!3M+yv_MUBV_NDSrTRS_+YDA@DhiB8g!
z2-xe<M~8&Me3BzD^muYRDQJ%Y{s5tP65+J$_Nie$4Zw9&$XO9D<wjeh9Eg@lt9;H(
z+|=-jhPRLZAbw5@18mh_b;xwe3|85J)uxX*!L;e)3I2(g0>bVTQ#C4PYDQ(wV1;sF
z>opf`vm%&_&si~ad~oErF+j-j&N}MVj!Z;|E`lGYOUxg6$V88sZ#`aJr;j^~`2O~u
z+=y#NnQTT#)UG&=O;5(fD_8VfjWG@85zd(rN3y4f36Z?GI?v2zumx?}39Aqh+%YI&
z%Q!{AprzI`^2A3UDqQRk%))rVh>{o3A#FZJMRtTs{bSO_p5*3%j1187$6X1ib<l!L
zm1AwcqzP{b9if?casy95++}e6&wM2ej&ne0DGJIM3jky{K0F@J-g-d;!U-|gIe3S-
z%R~q+4<Dh2C3;82;e(v$GBxZoj+nvW8(>^Hf{UT{J;c$~t40u<f?qIFUjvk-`<vpQ
zH|fB+m`uCGApKy^-+7ZbsCGhU+92Qt5YOtaH}*6rRHGF9;$~~C>?<Kdch6PggHg7`
zy||SaCX_O=t!Wo<+NzHP>l<-SGx-RZdPc@3YmBCtwA!0)#Yt$5Gm$CeMVge^(zHE^
z3o4bz|CV$q9nf)5cFVMzHfHP{0`(g`;n`=>mYUGi89*R|<PRS<f+Ig9y@5@f($RK=
zr*WfHkZ>z1j(%l&+(v97pY#`AjlT(^G)x(1bk6H^1gq#;^wtoD;w;k6TLC?@y@DV_
zm$5x${^R6)#7+8|YQoM~2&|+{hKEJLG4aeNK0Owwk#@ONek{EcuI$q(!3h@**3R>!
zl)eN7*WYvu`s3nw;7^9!vCjiO2pKzDK-d{TJQKy}Sedry$WML?GN=D<fB!$h<);Gw
zlmGSqML`oPQArgzqbEPlx%qKyo4DuwSZ#6Rm*t#x^y6}oIcth>><_zA&fxOLt@s0~
zzN$&9#lD6wJ_tGu_m;FgiH7z5SG=ZOg|t^``5RW;TL5!?^h!Iw5;qI?Y&`0>tlD`O
zzVw+lH7}$$NpPzFij(&CE!EZf3|@_A#kEVZtGF$y@`*5M2hDOyw+gLh$FG)MHtrVp
z7QdxjTH!2Ra5Ih-9t-yjjkEE}c1v2*T;V*83a$2NFSKzyb5FGQ3`Ec1NDK^5#L)1V
zc#obqQbVR(nABvq2fK^BN*Z1bK8;)u5!8Alpm!zo_67H??C<Zy?$$;OcAdsw`_Aj}
z@~z7XwJnx6cH{GhOR==F?ww1cgF`;Lj#cZOo#pg^J@Ej%KtsP_@HyhUVELG^%cH}>
z6VSwKu!jjD?(Y>}7~;Lu2_GNs#Ho0_lfymltYSCYf%1<Dq^_Pma?fOm|NcG`&>U0U
zrm+Rb5FW?KU|+=1eth!zr?JI2LcHhP!n`MlJKMX&OWlgAH=c>T)6RIbx)lq{>#?<^
z@KuJe5sY2D6GQQ)Z)6f5djy9p1|u;{NHLK;IW}V97ZadNTJCPGS%2%s4NrK|12bv3
zDSs#81G(R1b#22tXb`~l_p-~0Nwhv6J+-!@vCjHVEU$0Jw$inyvCjVPmM0=P+KAQO
zUic+`7=a=8t}zKt9400?6@c;<e}8%;-W<W5^7le#>>r)_-kWXi1YswV@(Mm$3}{ZG
zi_W8m_v6v*{TS8Q@A}ot77DT01il-A9r>IK*^Wy2?xqlIW7Cu3xXmrjE38A6Sw{>G
z4aTKQmt$gVJo;Fvestm;RudEBF>~o+bSur=dBuhQ`%2U5%4)1FEyd-_)A8c-&x#j5
zj;-}Ik3o7=?tMcev3GnSHusKW=ipf3X{;!f;~H01H&xaMcG=COD(!C?-&XPKsy`g%
zhH!<+Y3^p?$U~0M<4(ANzOLYoG9QB`{yPN>2viUna~#-iY;7P++400EV;%$!Om;4=
ztf`8$$H>S?jEynL+N-vH5?h-)7K#yfCLwQLo{Z~PCSqi`PvyTIvvVt+Xuff2QhCR2
zxAyqmfAPEVpZvpr=r;Ht{0INjc<#m5Vp;WbZEM^2#3A@Xkn!-5+TAk3$cU+_iI^H6
zQkfrUd@vH%uHT5A-6M^c*JD-nh@%b}KQkU-oJjxA#6HJ9vK#G4V|eVi-;1Ndrl&`J
zwAT3em<1E$2Lh+1r47fMU1&Q@(u){lrX!ExHNNxsDZCKUJBzw|Jb~;Tp#8n{F7mtM
zosBGFWsFbx@N%9jPZ2O9faKUogt07|T~>Xdzhk^bTVhh3ek_knWk*&Dzq~u6%p*`Q
zIr{QQcyw>{cchUp8Pg%mL(f=|@Nr`-i;&pD`O;r9XTc&y7I@Mh&=-)8#D|9-pYMzp
z=gRDpPb_H7aX6g@3#LvqHY1I+3D_!t_|l#cgt5SYI}y3>&O$4tfwUtCVWBAB4#KC8
zhAefJJCHaokVOgzcGeIstDcfK%;R(jhdB&{0-LIf+o}h-7&X(+TnkMEp#%bWkIj`{
zB~RZ0eUV6W2S*rYpN&wBqd__Tl>RjZUW$jx%6DGLkH^I#if}ORb;|K90%jHrFa}|P
zss$ZutKRuGeesgw+ZAuW`+nSiIPd&gU0#XV*?Y0Mwj9sixDwy^+Uu%sr?Ent-BTZ=
zbg*OehUz-~+N!=q`-kip>{tDrw7{PoPD-BZDME#Z^GmU?u<Usy?#e_^b0D{U)m!Qk
zySpaTmk&zyjzRUWq;GVv+q=`)Sxfm*SC}VZegGV%J$xUb>gjuLzZD;S_+tyfUwHNF
z-bFk#!Y;gi`R|HT`P<)BTUNjG;L*c)@BR1V-o4w=FSqMgF305PSWHchE8p8=VQJO*
zI60>8xXK5CC2e4CelDhzXYBG~S1R)~%gf90um9+e<L<rL_=WHMQoQi|bLxlKfu|@Q
z?6|HY*kdQq{DXV(qd)uOc<t4f;=AAbmAKH|7Yplq-r<aJr<Z&?kbakY+ag9LCSzb^
zG7eOS4-ZehprWrA!6XY$xG!*D?Rq~pcGRcnJIB42>zjKr@B3~l7HVwo?5O^9S_s9_
zha<|59_D74V`4r~^^SQd7Fr+(zjEbj4312s08HOX>OX=g?n;~)<v3*ZvC8|+-2;ca
zf4~A~j`u@wgz#EzQ2ILlAs_fQFFxX~&%NjwnvBzq{#a7ISy7+HeU3s6-{t6s9bFwh
z&hp}v`gN5*eZ}Uw>W&b-yWGb@svW7FvP1Fnk3Wsy`~5$R`;Qjlo8SI+{NgYCLQGCg
zsx2iCLA!##*xT85y_r_OjZm38)DXTQEZbMyXm9NF;=a;$^)TEYPn=kg&9~ckp=}*_
z!3X^W#~~hxCzxGai@S5{`W|hnAK39xie0BFccgL(!(FYw^OQ9F0`yDN|7w+UlBr@v
zd;TK97f|Wd@xYfAd{zHIlcr6)47OMJHU6z}ukotTO0Pl<uR^Q2Rqt<1i@7zNR;Xc8
zCsRP5<@%(sTj6{>TH%Y_fS=_4B*4d4@pnEi|FYp}`BU!r`sLrILS(G5t`(sI2C4u!
zxS<xKYep7c`hil>H_;j|M+00e_&1(v#;;ukA`UmHK&nma8;|I+a-wUx;-eVTtbhrk
znvJ1Mx88K}<6a!C7+;Urhk{dF&=V&z6{JStWkac9J8$d+K1!woy^Z{wA=%hHsb+jt
zZX+ckR^S@sp_=*GJAhvl>x+95vz~uTgdFE24-I2G?u705+mJR^i`Q!HV&MllzY-zK
z5toDpZLF(5sie=|1TZN>L&GsVI-1NRX~YfOf}erH_`yY|9y`ot2HnGhD?N-cLN9&b
zQoseX00R!Q$PtH3`H{zo&kh@osNi9R3XB`8EI9vnb#+<H_lj4G5+u<=^@Igc0I)Pb
zP)e(PBLVi5W}pS9=7-Bf4|>8d4(P=T0GbEK$|I97BB-f85%6EDs%s1uS<Wz<9G!uX
z#C6eRC@*P*R-;>ifgDUS9k9#7Q0gl>{1bnSWOE}u(985emjTPZKHf0;2#$7@*X+pi
zQH3!6qD{Tv0|~S4@)f|0C7jrv)m==SP{%mh(7ORk89Fg;FF81oH9?obL?)FTW5XI~
zF(974c+o53*gBu(K)q5r5rQdv2A0H!L=so$DKU{fX56#QArNB{@q+7fuLj>nls^VI
z(03jw0tnd%tC>E<SC8{go`|D$vdfEqmH)WQfifVi(m21U$AnEBTt~!6mFP&NP2ouv
zE(_9y8NL730>p>%@+zZ#&SmdYxMh_fxZt4{;<hF|sY^VCV>PvQ0mu_OJ5t52;hvg8
zw&a0S$E&#cXKe$F&x0S^6fk~;G`Rj{=rYzglV>uw#JNJ!1~VH_n(fk}=7W-SQ>}f3
zogbqO1`4!M($^UjnVbjEwh3)w*tfI|d&~ulVmH5PNOow}C;`qO=ie5YGidR6KF$S<
zU*aZwl#GAcyKbRRtE?}^0_<YiM%52&yUGK)idpMf)50E-uI&=RzUhlHASu|0DiWm<
zPR2-&{gp7%De0wGTU2cYXr*P18wMzrcngw*+>NJ~Ea*-%&S`E2gqpg<%|6QAuW%D1
zT`O+G2o$TD4^~s_cZ06<4&E-9OPmXS=8fzrXInJid<$xj6&;<aIz9NsAN^Lb2St#2
z*);-6X!x_6-XPXcsznQtf@%ymTUiurLFU(+KgFXCU;o)(_fd!bmKCs`f2&&UQR}bO
z?0z?Ksy~~p$w@7snj3i76#O%0pvAu0JxTR*o=#DnyL?Mp;#d8hRg>qcIK{ff{c*Js
zW-<GFfqk)aU9i5!0~I*y&#LWgVKwQHXT>j26l2%?ENWfUgTGS6&OyU7zHNFe4n;Lc
zTj?bB7lXsAXpaZpgr7`G@Dq@}z&V3{9HD+=o({@4H9qPo!U-lycXu~pZ)+`fHkM;w
zdUgWQH;8KyztV0Y4o9~QaGbf1GU^cxCU!eK`OK~hCL~#@&&BubF4|aIi-WygPo|Ae
zOn6e%Wv4$!|2>*tio3J(aesa}9xblN%KA?1vV*FFJB-@ROEGEc<B1eboA?<f#yh%t
zl!i9%s6n_ezql9+OUtpkz7eZiJMmyn=H)eC_|9n$E32!fXQAEcQJiWF(T8wHRQ9$u
ze5XsV_(+b;BI}qe;^^$bvB?;mn3g&dJwsz&#eUS*6`eiesaRn>JmO=H5F9Ys&vAD{
z%AeuUVVfr>xU*v*h6j7aLy{(TZXL$jn&P>>8QZ%DvCENBdm5|k9>va~5_F+cbVq#z
z6hfUX1WigPlilo|pkLk-4IXyqK!ZtI?=b7@@q{?LI}pOa|1rq|zsYe_OwuzM#wiUW
ziYF6w`v<$Ry}$2CFpig`Y*=`fw3&#O#f1)K3jAkI<ReVtzL!I>t3ru`6Us(;BQulJ
zckkYgM~`NGj35)Q>>%VQwTn!=Ub`MQZ{3U+UwkQEd-c_L{q?WK%P+qYufG0TeCu1^
z)blg(+_TSGfWpqN#f3%h99vvoiur}PSYBI+Pe1)M{)_+O{~f>gd;ih`ICeyAZLG#0
z{Of-e@4fqu?*wAS{K2m3hT_4ZH^w~F9Tul?)FpSrO^yx5)c9bGbN9}0zmK(=93NGA
z4SGTwffvUjT6Cfgjzx1!UbFM5-eJtdDFQ(zLAQ7Jy{aC8&$8;@GIere-9jYBdmNp{
zt|%53GC9rUW}nJxOzq?H)JRM#F6;_GNWc+a2nsmjjdtjx_mq!0PC2r0h)LAW_EWd7
zh55O7|HmJxUJk^w&pjVQs=wYnB;Sk~*@eRxfZ+!By{xTo$vwyI>|F6eLE601yS=mT
z9aHSG+EhE;Rl1H8Uv--9J?LNP?^Doo=v|0Rnj$RO-li_G8zy&DQ3u&c;4zZ?_!vQr
zp%6X;^uwI4fk2gez?kf2p(}ZWfEXd@fbye%fIGZ04M~F?r|ihf{sBA2{w!8pUS5oM
z-}zCz_2&2Evya}5+n;=(e&CLGzj$oKxJzZnWHH~AJVFtH8g*TbR0=x|8RORPoE&VA
zb3KlcBmesP5YTZPB|Creh(!<)g0Q=-9<xZB@?y#|#sdpt<O2fMZndcarIk69(Xlb#
z7sllDknC7^NqiBSB6wuy8~N>pzM{RihY%kD{HcYa<HLQ?-L7=)u0<T}#PCp0jExK`
zKf8p3&@oTWA%6!{$0#4}C!(w=PXs9RBM8(ibaTIz^DDkGKxIWfkndb8!2N2}<6hqf
zmV#gJP-@fISoNM0QK&x%*9TRWgbOc;u!1&57#ste$Lg^#mi~&qOZALCNBPQ5zR8L4
zxOC}K%uG!>4B7#>9C3U9_Gj_Qhwu6B#P!u>3oz+U5B2@(>QjH(KN`J*V=*~>DMlt|
zqHlO2+Pep0Q|*-n40B8C`gZKuysmG`^4g|%O)jjg$GpBz^Xi8dmsev=eG<EBS%^?~
z0n$D>+VIlUXk45aipk;b80+uUH$`!035m+Cn{rl}?CX28IRDT)liz*&?fBQf_pjm)
z|L6}q_W$``_)dK7jc=&jhf^r1{6a9uagBTjX-kWX3qA^QL*F#+<odav`?eP$6EAje
zaRe`YJG%pEPmlC%JXQS}9-oMb>5ED;yLoe`Bi{k)?N#;5Kl?M^vVd(*`F3e$Dn>=)
z;loGqJHPWg@kf8~`_YfERrLn}-K6@-%a>++(uq1T)g{$?mDAwRi0bXA>eyjCxPLq5
z9^AKZkE1SqY^i>fJ4aEnD-Pi&JD{io`n<Sa=8oATr<Z!gaf{?H^^dt6=1};aA#mNM
z?yD?@M@Hi6)r&DbITnL`ow2esr#iZ&l0S&yJ_zr{5B}^=;t&4t_u|35I}Y>Wq`vL_
z2=x%G(FZFoJ?f7;*|FE5JXbpp^+6r%)b5V1w$r%%>BsS@>evB0T(?$ZY_L17UmTBT
zE>FZWmpIawcE@fswZ~n>O>K(fV!7LHjk_-O4ZE*z<L@g!|FeJgU&ep_|NU>`op;_<
zJVW1}Q46vuH-z%ciLonhX>nQo*{b=RL-p7CEQ>BD1Bw69x0yZ~K|MPs;fCpZkB?O@
ziap<R7Ahiu_dR3EclsWL$n+cD#jCyp+L%v{cB5T=RtJkYOaD~Ceg0!+c!Fy7JMC?5
z;-^~8=j|U?fv4iM_y<yd)()`wF}#E);+}&q!ucw06&m(@L#XV<@HG4iU*!H5@%uEu
z$5-+9=LzGhz|-jasKdYgw|`sN$)}WR(fF(;bCMRam#v62RsrK)FcYO9N=^I3Giv?t
zqs0!R9t)$;7k^br5Dl3#;zrWD#P1AV!r8c>A+E%^TA<310oo2GCY_U*^{JI6sb#t`
zL(f(Pu$P@WiiV#?KshPnRnfw!{vnVd+P7C&v->H95n^@{-Eh9ZYU-+Q^!b*lXsp-6
z+bgV#U4`my9KAA#CS@~SHO~Me%B7f1PwA~h9cwvaOjPM1VQo!M^-M+;e0zXcI>7`^
zi6Oas?Xr&v>Cnf}i2~8Mij#~mvm7-CePV5)gy=GWfN4!ZKwdz~W!kK8;bZ9MkRw;H
z&!Zy{k{oj0{c#F0Sk2=dISPo6BFAg+!C<g7H#hGi6Id-z9>IhXCh;)T^bpC+l+**H
zfjq<BK8$RoCFeY_1~AdDd4-qKW8OfXoi@?c&>!w!%#h+Ooa+AzfncDc2UZF{C;Cwv
z`KUP_v;`>^@UR-sk;a9b&NtwlRc1z277ib}8Gtc`*5(gn(nQDuK6)4qqe?~K(<$zN
zV<TY55LDK(WP3OR5Tly(1(lD=M&-}I8Rpgue>)#z<%O6VCOI5K<vV4{>dzz5aJ5r;
z*JEBlf{B9RB)<J(=r3Nq>I+(Du3Yv}C`U>Jb>KjQdR7N9pkttVET*3NfnbhRG7S8Q
zGXrA=KE#<iX(5*I2*(UK1*+7zKO&0yk+V67(x5;LqX+oPH&z*w2N{0^xJrZSHmgE?
zX|mD^eFhi^37scost8k}lqdO1K9R5B5-0DB6CD_2{7`1t*LdPT^FRBpQ!zz-eS>lB
z()GA>@tOu4L$c3s&URE-ptzvquE%y+zwDt}NyFeD;7vGL%hGI;>1^QJW20Vn$1gZ~
zC0f3r?ch=MBX3Dp9wC@{j-9>RfBLJt#Z1jec9gwnl*nWnfJ&w4aW`$kkXA2TsUM{!
z$WnPwP)jKGYy5&!zfquIPnw-CAUlk7Fo^<KIgZ5*AJ*zA9`r8Z*e`=S)yqbH6l&62
zDD9t(_GwDsNg++?+O$iM9reoo1yUQkha5)wslE+WV;ZtgkC{Kxr;p-OuSVG=$<!3L
zj9Z}%uIpRVukA^@5W%&7=t=KB(E8#4S$AdVGd)-(HFAq~AsvLB*HT?T0uT$>A9;X5
z_E5?pK*l3+lZL}bZ%?Q=5?A<wwDz~sco>o%dN3VbNZOz5(St#H^f+m_*fXk4vIjry
zT_<Fho3P|p(V~4)2cgvn41DrZ4q2aUk$U`ABzs7;dW8->EqKJAm4@l3;5H&CY&qBj
zt{$h+ol%;dRE#Q&n(qJjAN=%=I{Y91ul~=D9+3H&U_ns<?EUa7I1QB_xcrNqg<Q5u
zdor!MRcQF<yFwm0ytDM3KP~1~eN_qTd|aEV9=ki&pXlFD4VV8JH~ba1Vh1$(!*lL{
zg&5W-T<fdfj013<9w`34npUwYv~5u>1(jx{S5?Br%{54NS(hq)#>Fk;3*u2#fINkQ
z>vxGut2@wYcDS_+DL8Zu_ZGh`WmVxUU2vOm1Re|b42`q#qF<}<n7?!83Tdy0dMG{Q
z-FMhg%@sbjy(8K?`(t2uD*A^eqO*HIyk1ZA^!0mUT@1hY7x6ga#SsL+r!e`;L@?>u
z+*ptGRgT-*jjoQ<c=@^O@%k&z#mw|ruoDXo7J<!!N6T?%b|IEGIKoW9Dr|Nev9g<E
z2RTZPT}`l`2o<(=cg6oG-ooXGF!(EW)3Bq8GJ%gelp~YY!$bWsF*@uWVCF%EaisCW
z>e777J-qLUyrBW^BW{na?VWh|=uv$Ajc>%n<W#hC8iCT<DgJ3;b<@WiZEfv(A`2c8
zUX@kzbkgwfQ~@SunQt<WESd;RxYMJ*x7&gZcKg7;E-bFZ9LFUo%~V^CNMm<PPj?52
zik|k<xI8r)m!?N!V4&BNs~g+fu1nl=a(baNHn;aIWPqn6@8SPCI&-lU^XeQcw<~^|
z$ytt*V`7_KVq!$a>u#HW9N})5fgZ)3Baqm&acp5A`8J?9Q#aTN;vFCG;VK(eiZkiv
z<9ii0M{Y53wy%dg=H9M;WlMjT!qiuzOW{(Uj1^K~*5e5{cz?cQtayL;;6dEIcP|ze
z7F~BZDPi{Etam~%al6A#3;1XGAq;kqkoQdZUA=N8{=I+qzZYM7?KN0t-;49l{_}q-
z-uob?#z&)Hc|JZmuJBY2;(veXSAI3FKl7|wP4an6wtLa!f$EgfccgT26c&@I2l`g-
zC|&G6WbDA00iK<F=gw$^l1v^mhCzVEE*HXan^OMfQT@3PleRV3-yfGR&bV%~P;hB&
z!#ij?<#(vJD@OXM`|TDUxMhnDcjlZP?Rillf?w`qVgi`^$=Ge?$=WWBQ8cFbqd)o6
z`1}9q|2<~s7UNg{%CE;?{j0woT?0e0wz;J|M5xo@bg!&$#e=yePvRe`JQ1*6oF0#f
zv0)3{j-3bW_}R}LhSd4e2YIZfa-dIQmxp)r36Gsp2t?>}5t@vRkHyt%GjWX*K~!hi
zO~ek*RgPa|W|lFi3X|~JLFXeLRTmCK(^X4_$!Q=%!virjrMOK@sDI%ur&HmnPN=;y
zu45j6v1vWV)bl~gGlbha+iS72wWhKoUc`)jj$KZhcX$#F7J<qSJFTdn-tnX~f}aaI
zxw}hsS$6Dh$VIhT7k$T$>LS7$>arJ7QufB7?d2SX?>JLE@Qy<L5Smec36J<@nCz70
zPON^{liq&rD%03cw1~qpNBJQr-^!z2>8P1>rvBj%VN-8!r|(4SYCDO!`=7-pAH5%2
zn``cexNmM^awcXjUGqI~BO{|48`2i>&-lJQ);6|dZfVs5Q6*60`V?q+SC{f)X=Npr
z)OS%<<PG(JJF&(m##L_y%3=h8@*?XmyZKa4R7OWiJLRIT--`<-M%A|~u08E}Jn68)
zVd9>+a@?b@vQRytZIb`wuYyzm(yscpW8uX+@4geCeD<mGgPm>RYb+29PK-~)H@^AJ
zxcU5x(V_nPq^&y+Pt?b=!$<XNpr<Q_NAfO7>h-qzQjX;0SXauIzLIYa$B4p9Ahg&}
zU%j%lqPn-`{=U=o4S_31L~@K{PfSk?#mvNrcY5`zt7XRw3zC64=W=>{pl{=b^N+Fn
zXP<o*fAmLx=rsQNZ~R8Q^wP^dnoiMh9pcD3c7CoY?-4?M`uXQ^=l-nHvKu$9U5z(h
zdqw^AV9YPfi`J*A^V=HhU-H6%%TlRhJL<F8G0nUP3rY~vIWEeyRnh(Q<4=^{{&+**
zJ{BD_p5#8lHpL$y*xfsyt1f*IufP0aT)%QD{`G(RFXLc$Gk)W*|3=JQzZGjcN3pWW
zQGi+hNAv;c?@-+LHawQ%uDDL+H9W&1^A;*|`ZB(u)GdYRJHysDcVm@D^^ipb2<54V
z2non5`enX_ohly$rwGgV{?i|{T{wx60euJKDDL0A7ytVAe?R6ImttmWBEJ6G%W>=G
zE$_;tqU<U^*Vop(tCn`iJ%ucixHK~zqxvSUEG?=Htg7EZ_&1{ZwxzbW8k_P%zvsKM
z)aUbUVDZW4x9`R6&u_=eFTd<O?Cl@&v4PusH<Vr$oLo>@9&B&M{tidKF30!Z{89Yr
z58rjVzw@m(;&1-iuf!F7%XY-axGz*Y@74E;g*D7oay;bGX}j;GWAO&*;M>K>EO%-1
zoj6oohxh1mpTxIspr0L=pwoxtE=*;X`%e1y#g(mCUfqs)Ju5OJFg{V+YZHC?CfaiL
z{L<1_xos!pTp)d4nsZ#L{#SL&I+;!5j9U{2sPxias%Tbd*qv8f{M+;-ywqZ7#J@G}
zHC`2(^qNrVRcJN0>TCa0p*5UVs9{nT^!e74&rb-u70$<_6~4#~_(|?h0(^WGf9Lb^
zFB_hgKY7$)L-3Umr%BZ4VC#Wvu}Swd!DR4CyL1C;Woi1WzEjtVcNV85BwHnBhGT6L
z(#iHM+J-mV$>NNi4GN*enHPnm`eS?B32J<R)YUi>yQUqGF)m<-Oa)MyaEZTs*o(~Q
zQ(IpG`AI}6&`5n%&@{rbJNlv)cLV8YkkZT6*rnd6wkdG{?3Q#KN9RZ4LZ6gWx@lix
z=a;SY@>`H`tnnvYY?|TeZ7117Qxy{iZ2HngA2-L28&_5uH9rx9UQXaA^T&@kGD*`o
z0svK=f{({g=gWYW9Yq$R=!2ZXFfl}8C@nY;&YlJklszkdIEI4sLknhwOFenm?X|YL
z;xoRDAZF|UK?vLh5d%mM3$bWp*kwXf->{P<Dzl12UP?Cl%oir7Nf*N$)~JArIfbIq
zJ0FFoj}*mY!&d=^SL-A(a4y<xRh_BWE9VWXx^T@h%ttwGJirs_`X!DYF%#r{PWm8N
z;&=psK18e<$&Z|TF?e=7SRI->pS+z~Zl>irsB-qoHdY971dMV(_}t$GBTrafeF8%u
zTHf&_`Y;GEXw(<($2wpOH^L~n6D=k)7-%sF23Sec)*&Wsaw4wWxE>cTUyU&hka{=^
zTKNP+%)M~b2bTc(t~jtg9tN2QAr!m85UOzu4Dp7^qOAA^!vs5j6({nSbQ7N(n6Z+v
z!*TMlhDtxo9Qeeaqk$}>BHaqx?Sx4R4SZ+^Fuxq}0&(&u%PLZECM|i&qz2^zZQfpC
z>B01o+oj7wcwS8<1mjAB8^^<RclBvddpT}gy%}R8V?xhzrrLV77Pz9@XyXKe4Ju`s
zreasxb8)UfdErc{fXh?;xv3OnQ$vq}y`U%xmw3qJ`U7U_ZCA|1Faytc;oEkC6fjsT
zIO|TDtaD&C6s7t`pSsKq+D}2kGj>)g-Qu^j6(IeUw~#WAFg0nLPSH2sgoEJ>ZCS}+
z-_;qrqNcaBu~~2XbzGAU_Sk}9v)rgB^cnnV4>k7op?@Qf@L=0YUnV)E6Liod^-L>x
zGM7kG4h?}6U~Ba)vcs)-w|!Ia0v+t2iHS#G%M7dHB{CYlj7p!IY4RkriC1te9miUx
z={?PkGOOFrmpCVe>}{9yz$XHv6fLA0@oBNhJRqIvmk^T#VI^(M1m)I*9$Ux~PNkv0
zk%}IF)<DNWW}jwT;}D)5GHhrY9}GJrfQuO`yz#QO*juSI(yij2hlYCzKZ8Lp7b*5;
zJ3tzTjV{Xz|2Y50KgFXCtKci^SVH!%O}ko7)to|_7WaY=G;mt`f)+aFP0#p)ZsJ;q
zRotY>{1Og!tu*-Mzu*dL9BSME)Ae)Ke^o0j+wm`MeDCxI>kNKX`ALIG<TJeL4x}Ex
z)$M6GXi}dzX_sa@<Wr#fskoJHL)d0jPsuaatZE0o(sVv9=T*RZdr914ID-SUnrm4Z
z&uy&Ly~S@Sr)qAk4+-Gba9e3Kaa-%#8QKDR^a^L`COqA)J{3gi94-^!93|D>(Hq@;
zgV8@U9zKe!r_X}0w$6U>5aK7q*Yx!E#n8}*#{mc{)>l`9J5_z3kJ88SV^=TFh!>fN
zNhSw}L_>TILM$e<RyTLz;oMTpKA86eD?C+iZ&wWU_ggqMGLk!R5WK9et;Ndnngt?0
zLQi~Qo(MsEYl{xWfx9=@SumuB#YR4gdz}f}Jqvd@>T7ypOnoTRfTtGbaco`}lUhnU
z!i_!gX8VfY))qn}?w(=twZoJ3yDZ2P&xas{@`DfL7_oNourv`SNAqNgqv#U9$7TF8
zO8eN@kSD?y7FIpk${j1DiQRJS>|nyRr%U-MzP7)oEv80?ykqQG`LVUVYySAOt<%DT
zgCll+=_{f%!=v}~=~3D`6;JL++Yo=d!=(E@LaPhG{XUZ;{c=B69JZXt2q_S>z?1In
z9tKCsQ44$|S$|KQ?APPv8M830c)`a4u5WHLjLHdo<q3i??j=(My`yPtNO|8YefMb`
z?ry8RmQ+?u*u%H83nq^`eD~dV<I_(+iOtP*3zqu(27KHX&&2q+>RNx?zW>mZzHyoh
z>6mbS>BUzpg#GxFPrO@XW_miVT$+x%cR!C$KKU><w>K20ljsr8{>#7ki}B09^o!18
zn$nf)w_<Q))b)-&!a_qP3l(>@uH@U5lmk^M?`Bi^AnZg~LK#pumRHujD}-<OiSj>B
zb6|%ZM_aOMrA_&wy3II4X(cS^kByAP#hI}f8CLsI-ZK_iTAYvDpMDw-XYa-O%2JFD
z_r))M=jY><mtKr^C71=D-qCY-YQZ5py%4IM9PP(ox8i)Xr#xtP+x*en@5Vp+C;ucq
zzxyD*{>C@s*MIYG#Q6A>%8Yx8C?Af!I*8@fO&_(mrZz!6W?X{6lXl7u9(FS_!Ad<w
zD2&jT@d|YffrobinE~NQtpo4OTSqYCT}TH`+vwO}TvnM}nx3+7U{~QXslBzG_c*be
zh&IDGY-MHD4?CF{OL<{~qMO3ft{6}sFo<xNoe)fx)7}d(7Nakr4IqeMSfdWXLWU#d
z5sMi~$IdpB$~zXYF+V{4XB>u50{<4yD&F*A2;Jx>y-Q2=mA)7Or^nbl!uRC8{LmjD
zaA21!eGv7tpZl!zAk@s`A(_nIH7;~nAV68NV3;vpj-e4OBFOHD-k!`G`cF@ME6+G;
zjrzv|Uh;=|6XM>j{ORdwkB$pRvA4Yzj~?8Mxrei=V?*)UE3bIh*aO)iWMOwA$HCpa
zc`IIi@x{1$<+67bF`h(FzPz^Oowpo0hfs*!K;1pLi<fvTFRfT`k1*W#vnemwH8eIl
z=A#pdH)Gij^<Shz*(Tcbg(@!$D%3W$3k0<Y;#nA=itJ+(>Ep-L-*LB8yLfW?PZp$Z
zZmC`|KBv4C&&Byi@wwVPeF#UsabFz*;lADh?<yS|8;kLY$!J#{URML=Uaob0U-tG>
zkme4WcE$XG%7?OMT%Ef$mAAqf=<kY|iP1b7Pwi`ahkCo`g$oE<5I7(gBT1dggYKSA
z#dlwIVAFF=y(-g7Gn2~a8|v4G)T_vc{2l0<xV^ch_P8B)?%avl`wzUk^y^=HJ+5EB
z?!IGPZjTlg<My4~-d**H`UdLibI(4fGJ7e8#%7{}<6D*XUe(p^&I={ajy>;0eItFt
z1B2eRhfqG}T`qWs?uz>EyZ7$KpMC%Ps+T)4bLnb~j84Q$ufC?fWgzxxFUou7V<txj
zqVwb+KKkI@_~^&)$1~Tj=zD$?fBgHu7eDjOZ^Uo@t-l!;y82>vX)6|1cVlgnBmGXI
zj|C8&${XSG%^4pZj2?YY5N6R2?Vo6TqVHHoJN2c_yRzSW>n+#m$*BnoXuJb*a7f>$
zL4T#$iHC5Fewg@mdl#|XRY%!j*5<>RLuI|GI=U>oqnJ?L;66kaIKw|6gg@Ndi_MKS
z^#f~euYKxMXcz24r7m&TCH)zGXrqe@^KtLaUH6k}23|1m@$Cn3s&5AOLvkPDgZs1Y
zw`OLh+&41k)35wTKPf)rxo2+Z`*bk|h6a2u*B$XWpWmC0uHOE*CAVj;UiPjvzDN5i
z{|(i7c7h+NjWOTJ_nZYdd-|3Sh%USFhTs*H$3uhaZ+bd>A13#fuB)%#QXQjxO;3*K
z8{Fo81Hn6YfN_`EsLF)AT3BMI<A&<ny2{{D%+JrovA*~8ae#SHRJCbUDVJ3Rz6H4R
zq#2wRJN_+hSfm>?(w_PbR6L*xzs~w?QB9}DqlAU7RZT4h#~-L+0q2D^O?pkJ^eVKP
zTlMtkt>)Hno&>F7KPlYva6TTb@I`LGPjY_};Nz?KJD-<-+3>VHvf%6Q{O$jc>@QZX
zsy4||9;FrowmI2Eg=pgC3ov5&Cq=)DwO=ME1svi;AOfR+YTAi*7K%lUR8lT)z`m^J
zyMjz5?$)-rRD4u)t<abRD`brc9fh54>Fa!c07Q>f)M!vxMeWHetco5Ieqj$?%+S-1
zFdQCMfZR<Ktsb;i!K(D^6G*%gnr+jyX;*!trfce2-H4~nSmA1?#BIEYH}Ti+ES0Kn
zu~jGRysCDML|jk@-WwM$U5QKVxDg|i4M(^%3UKhN4<bh|c=E)m9X#>E?ycMzlguWA
z1<tf(M;FYgC;Y|ivjYr9%7ZcCnE@6(gaO<M#YE7NKHOZ6xwN$GbB)(l*JFEM0~Q7<
z8qAU{-Z3UT5B7-$plFdW+8Ndc>}tEiuG(kvQU#=UTAf@OTHOTBvtuR+n0hYme;&~H
z2*FB%Df!`p1H<X+uKcNXU|O$oBduA|Ny9=AOhRgOUN{O^$zl4khUrD%#lWIn4<2Cn
zyTs(Oby_88X3=%t+X;oCusj(^m>C)8H|EOR!|=;PmZ8Gpqsqhqd5#?ePz!m)pqiAh
z2!|DLM@)9ehI`0HG(f(3>t<XMW5K!DU923_hx_1&yWu!uBxw^57-x<tVW7<fO^qKy
zQzmK%n-xWLwxo+eJ@<W)PfSq2IQw`8j_{C5-DQ`d^HXU-ct$;jfkW`*U3Cs4%Z4&C
z$c{-9PmUCKnv4f=9=s_p6b}Y+loT`xpY)TD3{-N^mT}GYMs`wJ6_(l&;d3v=)c8!?
zy79d4OK>LEu<S=ED`tQunKgt^>sz2WrdBr6lNN^~Js~3CGCuy%Bm$az+YZWUE<rfm
z7>cbR9c-gN@givKJjo{u%&k6ZNDY9dEfa@YpYgKJaW-AKRliBEH1zZFw8=h;Rq9LB
z^KmA~PWm&nrOB*lhSUHjEUCJ$CT;pP&>I1yJ^v<c+-L!)1<&+=rmd|}WpEx8eQHcK
z`(fOxb%7+8HoQwrZ3c`lE8>I``D?RYjWky;vP(R@u*P1FDp465NU7q;zf>mruJoKJ
z`kw{9CjzJ1Xo=)7@Y|r>RE!Fa<1B;GGB)VbN2bS6=@hdSBUJ<di*(EKOuctwp)ctI
zN;a6R+YT0Tw6pN+jzMbr(_@sx%b=8VlpSEJ^yl`Boe31E^rg7=%8Ag7XBS>3VtrCD
zk{&$ecgy%Pv&+2^XM48Jd{Xh5Z@ijJq39Qw*g+%W@BM>+i0w}W{zw1I|5HXy&?-=H
z9v5g8NTK>ExUKdT?9T6;T`S-}?NBo>D-Npj2(#*P^@;g0JpHiqANTU4JJYZ-*Y<G6
z4PV7v`~qi%dae4ieq3zAHGS-{qrWP+Dfq+wEDfNcFaA<n@z1&OuXchay~J+_XRPXJ
z#J^Y>M2o)B7H8x0G!;@xVCvJ*bSuzm&ip8NrJXh1Tl}7@H?4V=0B#Mpl|~b{wVs}#
zE%@^YzJRo^Pla3*li$gVS9+(tv#UjXi6=zFId=BP(CAbQAnfYxi@~AcoHX+!?~Zx%
zT36v)n4}u)AF%LcW@^ln`>bdeLz5HD-FZ*d($ad|x%VLE#k(UYq3?u$V-k3LY$Qg8
zhD?w9e%L9(WH39wNDrR~CNkN1kOCRSQ~W8D!+Kfpz&#^ea6dCS9v7!ZV@P^<GIsc|
zf}0iB026OiEsg@(RT#d*M!Xygn*fhN;0HaP5TfFkyqNe-@pT*<#-b+Q^CLcvBcg@}
z`+TPo3j+`Lk1TXxLW&7-7UHpEXrQmp_dR*{O>d`n^9*!dh|Tr2SXo{3itqMLb}&%G
znRI87-@f_RwyyRV5bw+GGVU;90+Tz9anHMJPkp2lEAkmfG5Olo*69;E`g>Ux&ZH~H
z9vz5(JoNoiJ#9yE@AJ=;t~T-66W)!0AY^@WqcM@#*U#>W-q=055KHUgtvOCkc{D!Q
z9U}wnv9qxn@4ow^c>70hc|2AH4#Xjkv}4g&#PIM~Ja_APAA`tb^v33PtZ(c@SKpB9
z7sn9xA&e1UOSwFJH0Pr&EdWqG+u2%=*?XVG*2a?D_dV96267_Bm5Vd+(hJYa{!ru8
ztMU9xuSDO#kjFjj#OvqiFO6v^tGweVEcE2qO@t<SWV-4D^<<abFKgv4wIkp6;~jAD
z;fe{9o!k}FukjI!I1!4l<B2-McnM+b@W_CLx=hZpvu$H@(}JP3)#bSN=?5{`*AZ{L
z`jU4Wtgf!ab1%Fc*KRzkeD1Mu0XnD0`=MGIfA*(;>^%MXpZ|FahCaLfdHkb)@;mY2
zC!fWwo6p6s{o1d`rK>kBB;~#%gr-~U=2HCG-GflZljX`U?+#KOVnUgTSn3AprjD{(
zhDm>qQPhsQ*`0=9L}Nlu<k;NUb{$<&+oWBxD3QG9xKDOcv|Z>>I(K~k4);ZIlqW~J
zv119L+J?%KqcQQz*pG*CAbk??V!VjJhy|0hL&kVMc3x)sN<aEBJ|lnhw=k74B)j$y
zd@*UxuCjx}J*nJNrgo!#f%aZ^v(d*P$VLdJ>f?4!9it5(cxFuIv8#YK&XIasY7gE?
zDO}w8`uctR-1yj-C#u=`K^zDlK^o&{=#U-+^Cz;qd*`-~^dvtKA|cFNURt17Vtisc
z#`I9W^sST^{gM|>2KObc#@)|8jO9o7V?cTO{PQoy3opFn@M&uo6o1+o{Skr|c3UA7
z8DNp9?6<e*OLyb6tt)m^6*f4oQuTy1lfT4~yrsWdMCiN0QH$(C(Rg0&2$ctwW)>@S
zs@%J~xqq<NJHZ%dl84kgCi3~D&{s2lrX0yLcCBy+9><MxtXz*e{_Tymc>le3;`=}N
zVcdQ2!2P_BW>g+rxik}Ryz%w8cKy2Y^Ta}2PW5SP@3b&)VsbKeWRK8mPx-~pSA-j?
zP9DQ^G%0nEy2Txd96<^%gb<eFBRO)BV^D8gnTgRM)lc;&-1m4SItVqk)|TVp{m)`?
ze%^U=<;o?u8@@dV3ORn0KA&$*PY(k13$AzfZr@d!-cujB5ySeHz54RYja|pP>Q^7m
zJ&NzY`NMeUgZE=o^@<}UIac<KZ+s*E#$W$U<9BxTStx>lnQxL8IV0fZbemo7)LV3(
zU%h&@2;IuDoyrINy!Y{^@!>}wxt(0Pd@U~DxTPAz2|(>Gp9`v+@OlVNMh3d$lMmjG
z_ulzYEYCl3AIqIrr0DfmUXH)@@BKUR!Yi-E-brUHuI|R%vf8k||H`E3Ry`Q$X^)|9
zj^8}>H?4nQB#y*OZ0)NL+T}zX=KlDGsjpWXB97`ze4OK*yZ6+O4#dx@U%Pnqh6VP}
zrtjkPAovyLk@84i+Adw#snbYbMe`kGC#~vXAKy24CAoWnwfcq)zNyM9IyjEvb$wU;
z;mBZhH~v2P#3{A0x_F!Uns?rPFaFEl{a3zMPF~Evuw&AKGSz!_VqKh^@*RpBYaI1^
zsJwb5zV`JuVpR6rm-$G1&f4ZdY{PS?%)p`UaHQ*q>H{5*?`PaU)VELH3Y8-ZOExyQ
z+~!$a;=9<Czk|I9xDY(|$I2QzxY$9vp^APWyFSk&v2cQ;i+g)I&4Wyghz}U(6<^m8
zD;s<9(Wm#~t#>|%mE{%ZA$@$d)8u#fPV!C1z5;Hw-%T?(Eq44{-6}1sX<vmz6%Y9H
z2)@qvM_-N7lMbL&Ak~=fEYPIagjSlsV__6X-m__L4d+SF8upXIJrC#O(F$MW2K*%V
zCjmabiof%D`Iik(%cB&0{T+6EHGz({QRLOdiID+wjcW@E&iG1m!LzRT6SPnjJ~&M?
z(9Ew&srtkwsZ}bTa57T0tM08PaAa4B*uK77=~a{xThngRs&sKoa!m>q63-}l(S%c9
zF&g?-uOwI%-yp)C43g18Av9-|1i7J4yVQMgew@|Yy=wDiIEw?IuV_s?$}BDAUgO`e
z!y0@38P5c8)XNv3bF#a=wpj>r<;u0VbmdBL*(&~xC;H?!MuY$`{|pp88R+j26Fp)`
z#E`+HGO>YBYg-?EjtTH_2KpGX@(D)OLXTrHQNl$40ON8b6+4)62Foyo+xnQ$iNi>&
zi5d5$sn8J)f}s;WxJ5wG0OT(!bTea$6~covaL(uoQtikuJvQ<x={GotmNIO0ul~Oz
z&<A8u@^Y>5554L8Rf%RGloXz_Skuie5YnH1izzLGkn$s^51&v^RZs}3+F>p@vP1&{
z!sBDleGZ;LQ31m=vir^#-|CkG6H!rw`4Awm!HuJ6d_<y|4~6Lniv(d3M}9EJqU&Q4
zD1Z5dDX{=l;aC9WRjd&m`jC%|jmMQ6*L>FkM>_O~31?76ycl@aI_RS@#SrajAiAvq
zB^Px2a$)L$9(EJ5Vu}F|$zl-1V3>&$1{>_+qTUmC1Y#UxKo%nWN?}(Dr84e{r~Ei<
zKIF0^?m3x6T1h8**)zBkz6V?yG||zMH!ySzpkOvw^$8OM(@vT!L{wh69g=TESY<|B
zi4(^hBKT?(Gdnsm9xpumVoXoWxNH(v_1sy{t3FWVXN>^t)RWhVkt)tSq=P<BSzsV+
zz;5L!PST_=XrXmolp)<~J4_$gnHDHEq>FtOIu)$btJWWIEPQErvd;D;b;aCtA&^Gg
z3qiZNyplF{1}RCqNvn8Ka+v(Dsc~*wcOH^<2@`)(&=Xebi>5$zH*PVQw)J`hwOnhM
z**!JWk%gH4jA8HezOt?syg^H?Z5*sK!B%#I3c?SY^*D^$mr7frVtpei#Xpelyb5wr
z=&R^^F@Wy32xJ>pT(PXd;74js))w-3wJrKo>mn80!2=wpqId9^i$Cn0cXGfj@f3%K
zXB!Pf1zapjaE7{&#dwW~>ZO2;W;1Rz)E1&r8tE=^NPWQ|9vA>fW3N{pK}xwKE#hk|
z&}~lt4LfN{R!ZND2RbMK6_>JwyzR4M88FwXg(bT?((hS1Ha8NGO1huM-~0RjX%p?=
zFjT=8b*4gXH!g2}sKpKdw+e`W^Y~SNMs1-FG~J;idRCkE(kPEh*iVWB`lhkrR^tUd
zXQIt#{Wox3P-#BzR^!!b?_V)XqX+*jb_JSgIZFeu-hSjp1Z_<BR13bJrW;V{RsGY2
zUYJ#riZgEXYL&ha?N(`@R}E-OI*Q?}yP(zVq_~WQSK4CJy~S^-2i4sCR#yPGhTBS`
ziF>v#osAd0WreeJoA4xi>@wOl{-h{-vMR)+Pr#O%MV}0rMpVRSw6}>@fw#GkM}(PA
z5zod>2zKi9^bAB>SGV}Rp6Ko#6hFg#KV4qgKQ%sVG8>y4u_K-yzK7$?xQ~=o_Ydcn
zVsVk<$#;}~#X#ZqmYp9=QnN6Qg-eVF;I)W<JvJL&wXtGZd=xxBetH!rc42U|8^WbQ
z@rg`Ig3DcCo-p!`lLIey;;xN}u~G542rw8w>@y3lT*`@fcw2`}CW)Cs(84@46QoSO
zF@egd6!5|bmbgoVW4+i_gAkBOG{!l;uV;5(<-jC!cl0ZMqfB-wu1p@XOJHS@9RQzt
z@t#;#=NI>;Fmc`2+pl!C`}idU5<~qx_Dc?MYR3B3w#Pu^68DHqjE#75^Yc4*<I_(*
zwfU77pS5se=b$Y%_fLIPB0NZMXS*k$Z{NNhpM3n0c*+AGNtVasO~mLRcfEBfLS1pv
z(HlF5?Y?XuUYH{rJ1%gv*K*vs{aL*A){ms#5x<+l#RTe)GEiyn?Bot7j>Y0Mi0&91
z9*^$cftZ}S6qhbvj`7i9?~3b|8-gB=8)PC7;SCdq?9k}zJdXF@{*(CdqaVk*+}N3y
z1G^9}pGV<Q9taQ+E^vR;xXSqA)Rdl?xOj0UM#m;po{V$EjVMnMj8XTP6z3i}>dNVb
zjyOD4eLrcB17++1V-@rKDHx($1~m?DS02EJ(|8oL6vi7G>na~<Q;eh7r?KJV*0z*i
z7osx`;@-#a#_a7M$JY9CEUzrZ<;z#&H-7W4$JgKZmhvg@PvXwF?X9)=XaDp+i`C_&
z_-lXduf@%qHx>2Hm|t3oKl{Ot;s-zcVe~2wzw=AK5?8L?u;7OK;<nk{r!cp7_bkLf
zSk1031On_*q{aIP&S8$*9P|Q1+CJ?d7eFYBI#sV#&&IjmPW6?EbjIw<Trse^9%~!y
z_++BF-8<Y6Fd%GLUfWREYS+pWcJ^`X9zs<0FrNIU-K3Bx^T>kj5+<+g&)^)gGa|nk
zV-h!x)FZDEJPs;u<WoI81c4gu(qlg1Fy^BkQ3n}wG0yFx?osxHLAA_K-G@-V#KJ-`
z`a1O!zSoSo3XOfeNSCpkywVr3V~+dNMn}igru*IZQ3;?;oGtWH+hKhB;Qsx%clUGQ
zkHi(#b=o(AlD$2IrW(VZw)r#@c0sb!2O0=exz`PUi;E9qn|Dq@0095=Nkl<ZX=x#P
z2ZrLsmtKu)H*ShXm*-SC>Ty*4=Ea#Q*O^|8U)l9~c*xZODXjbSvybCX{`mKmjsx|L
zlU_hNs&O@S00CW>`%i?9-Ksy_0k;>Mdf4q^VWINBL-pZ6^p2FL>|{g`@Xc?2Q|)Wq
zJ7y8Q?rKcT1boilsE+mYx<B#BDi>t#&byzEb~fGLy!*~OaYuP`L2Z{KO0QnI8Z*;V
zF{Ad%oo}n_>+$LB2eGuesq&%d^B6o9SRhEF?IY+w0P8$cKg$9K1i+LNRj5zt8&G}e
z^O29-pOk_{<{Q|7OI@OW;c5f~Q7W()?Cn;Ue`J5`q9L#8_h~<>0V*%$2mMT^+Ovr6
z?rg<JAAJ<R`@8=<mKGM`wO3w@*Is)yMpWK&^YiijAO0Zj+`F&1?RoJ9^H_X;FV0+y
z?|%1t@!eng<(O8Q>4y-gQ~zimoBDR#dpH+AeCzG_{I=@PKySS8%=LKw<}<!)a#d|{
zc5co)k{BOcyLLUUJafZ!6B_K;=6F{G_Xzme(Z<4G$_G9Gq4<N@`$}(T%uJ5N^1^KV
z-ar3cwfCX;+N-a{mg0Ny+KqVS^*5qta8z}JV|Fv|)SRMUbzrEcEqdG4ryL);Ua;}B
zTi^MU3*E7)@?bYC!Y|(`sr;o};49#5S@1G3DZW7889o!f*OD?+J>@&d4l+K$hr)$Z
zp_8Hy?NFX~bf_Ot{<pCRLT!`sCtc=KgiYRYeBs`v`m2>S=l_=K#<BPr=9s)NS#c(>
zM%f8Q+g9Jm@sNv)%kEFS8(27(R5ri(^FI^M-MX$gxCVGmh_p;hOvR;3SE65i0r|sS
zay%P5-0ykhzHn3Db^0bJgFa3__$c@m1k@L2Cgbu&)p_v|-XtJ@>dY0E!ch3&d;Un@
z^*+@P7PqvWu)|e-g!mfn>g1kIc%8A~{<t_UUP$GFp!4C}YP|cwM{)b^tovt<Y$mS(
z`bE-`bAs8|R=~~U@`7x=`_f{^-w%FT@zM;?Z&dpV<|$Dd_~%s9s_{twtu(B!M(Jth
zK&wEC^t3{gUK3hr0*{4JfO^$xZVl&2&>HrW!aWb?<IxIV<OcjC_a^~9zKXx|dHI(O
zPs<|<zJBZP{9W2u1BeRd8YS!XQvXI(u?=<<nz7{Q3RQ3cul!R(#7|m+17u}N)+7Cx
zI65KMJS#Q^7?^^YAg$wLZIe#M3cJKjQ?)~<cY4x4{iSW1nm)08l50ZR6&z=ZTorM$
zgZSg>pY-^vp3_fSWRjsQr68rit3uLFe1ZLEJhFdl!NpAu85L8sb<O)0%t9f3qpO;R
zS?>mITk7S1`UG%KD%DTb%Ow1ojOvyyX~K!dojbl%Db-PRt8?azS_UW<I)J4F(I}|4
z4Bolml+`k1vcr}ggMJ>|VA5c48F+FW0h9d*K47F^EMYWRec|0*Vhk8`5*ME@u8${l
zyb?-otf1pk<cIp`Z*FW7VFj2gHVzLEP~>VdZviKMdP?#v(C|QC%+Zqoe86Q>A3<%N
zN>&ppaTR!?6=dia-B$tk0ZK=fZ<c6M`eJGt?%A))mC9GqDL&aT$V*G=4)!Vhp?_dS
z7Xl{l#!^`${Nk93o-PJ$92JsBPi${**bg5^2F*}}x#IkEXy)KW_Tq^6s5>sgcTL1V
zJg6ho59*8ZABOYTtAY^tz({0j8Hflqx-__)n40tvn}cHX7=$yJWW_52EaCwJMR{{n
z0*vc{m^I%ACwlDA@)3tHIl{?76sy7zL@Dmn1;W%rIy=NnBU}QH6~w5y0*}>$V(zFE
z!etj2aYt~)hnxYO2U$+4az>G_A9xrlj#43?S*h$kS!CE%Hl%?J8R?aadJue3|5+hu
zCd8STwZW~v)D0i6=mB?|27try^7AhM-#{S0`N%`!oJU(x&x;?$xlYt$rtb51?64OZ
zs*Hoet2;Xk^_|v48;f*K18FgiUf6(CT(VmtY;0^#9KbVX(c8Cm0Az8m7{#}p^JASZ
z>ARXL_IcN|!LBuA=YO>!UYKRD8MTx#3P2qEw!eh5DQYzcObYK<H1sWbM3E(C@zQYh
zJN1bHG)#JB>}97w948Ymw$&&t<(io3$8kb$l__J?mT;gU&(+3463jA~N3bu@YUG7k
z31X5ywS<;*Qg1~7zBHM+!bP9{u(P%G{O!T|R>;ZP5=Y!j&sK)C)3->Pq>)6?gKEmQ
z)~gYZ3>14)#?Gi&3IIJ?SsQ2j^r|*wgrr?u3mtIu1{W9G6$F8W;`TwB<iC0ZtoPsg
z8gC2?nC#DKN^m{R){LpV6ThUykmNH>4ZnI?Fj}IJK25viN7KJF`lsJQ&uHnZ+8Z4x
zxc}Fm;*PKX;s4_Q*bKfVN>KHvO+il6j_=WVeE-hbSuOo5X8NRty;b}`zv;J?Hs8or
zx>orY^LctaHNG;)KZn9J@vGWMQzNXiVWjq4yjoS`$i2{^Zo&hi0*PC23Z%cLTe_oG
zJ5+4-t*v(a3w~2=hFR0*uT;rrz+R1?0L^Axl2&y;uNu&{;Paeu7qprk4~H+jY=|xH
zEq+_-0?=B13Fyzpt>T<1>z1^(;*$@pQ0-gcNp^-Bw-W#I#fQ4WQ$3PDCR51mT4&P>
zag$y=5=QtCcnbI!GKS+3#a&1vUPk;z-vCEZaYxUn1sOS!G!h(Lwky7gT{}K@T4Avp
zfyq`TR##S6eYX*de8>xS()f-jJ@9TEEyj)^_^*tY_%iWU$0wZpa4h~00bK6>hxug!
zU1Xd#CR(|}gT+#HCkd05YZ@!8Z>;-N0CHn=crY$qn)dN;-RyFKr*z(ASu%0W*r!YU
zFky&GECP^Rh{q`v9FfFIYj$IChf}xsVQ9k-vg(}sW7u^vG%^<b;`hi^c5^URVb=;f
ze3*<pI@t9QO!sH!V(!tL?|ot78e!Je8#iNYaz;F4r+CSoI5|{)o*aq(y6+TO)Wgo2
zpZWIB#PiQRr|(mn(|Y&LU0=3;<>Ex#x_Lc1x(75C>x@InkIX7OJ7X_SL?FDS`MKEM
zT#K3Uu^1a3@{X53{o$Lw2W@ofQjCmGiBIl~t&P=qaQE}LfA90y+1fBqO`Tx{I!`WW
zJ8})~QoWd%ysSDh<Zs64*rd|ZXJJftcVCQign19VF{dyfNbL2IX53M<tvbMtI4<PB
zu5^r#^v0drpT@uXZ~tZd_@fU!R%{cq%L2Eo>G`fey}LL)?N#~glAs+h5zDcKj3wA<
zc<JJd>)IcG|NHUrC!c!f-KERdV{GDL^t0=+r_aY_a%`mUan=}tT`xJI&!Z8TDh=^q
z!g_9D!4s#$?4pz3rRDXQpI;KseiUPUUD1BL8;|b(INtuz_kDDu$oOcn-}ueH70=v!
z-s2}Gg+Ko2{rG?V&OeTa4<C95%(uV&ZSQs&8X9$dL;&{w2Oq|xhx765b1%eeU;A3L
zb#?_uA2P|zWUD7$H+ExfZ9CQxJne9-VO}S2LG_aM$T+53bzCGpF2;>tM`sFhFHVj7
zPPHDDJL41REG=(ZC`tQ70EKX?qrFq%5+-+5X}qsK#=Ce1ly}glM-y%D_(F=Te3P-!
z{^_ec{%68mab%(%eA+bl2#e@1*ljdAGUz&kU&f0FjS(!+S23YXzvVdq`YyRMUStO&
z^B2U+!Wp^Oxej(LF){1~vkq&|M_DEwixM+E?ruvRpbpW$FfN6@cZ;dqn4qR^A*5uH
zBRkK0l%L{)@W%5TjQ5Um+y`Gk9*t;>%Pu)~D9JLGmzPww56HC`7#fP6!4Zwk2UJe&
zDqD{3Q~#+v=~KBN#HX(q8SMAbc-vd+vAFQaWxcwx81KLLmfGrV_g9nyyBwLnVb>PB
zlei9mxL>?<Erv%Y)QZR#je(Uu(oI}9R#%+g+>JCbF&VGE`f5y0PI>p_k(gox4i+pi
zpQF4~#`%~->Hz&~Twvd?3Y@XDa8E9@TadD^PYb)b7&~{V4RNF^`MJfF7E4OU@<wd#
z?5mtZ`Q7e26DP*iF4c`8c$#EKjLMe0+~lrC@>FrAY<*Or!tU?uR9){?xLjXwXdF(A
zVQkNiDe4#tcM+^`L?i8)orT<$##kMJD8eJg{s<8Hb{(m|TvmVi!TayV+i$<^dVTZ8
zjd<hrS7Tc3l>YG1++6(rpZswwudS&+829cu;;GDtXKy|iFTVVG^s8K5uk{_7Qo6WP
z>{R8ny0R9xXCKDw;u0TX?}Qsue2#Xv)y~w1EG)(Y^$jfYy>aDQy!P7bF+MY`KI+uR
zERx2d(Xp7EQlBv}sD64g7Ut)LzaJyAKVX-p(lI*F8`Gmb(c5|IKIy$5y&Zk(msd76
zVrgYHCiE?R;iXsN%C(!`CEBex(D!vIPuf*aj}KK3_jY1?eKqdfyXRv9NA<nh+C7LJ
zeRmP4dJ%!jnSOLge1O}c+6{dNZQu8*sh-fEviN{|&Jf(QbBu2;-8yp`gK8Tr{6I+5
zu5aMn{oC=U%HIz^{@C+^2)LOV;d^xR>J|0<Qx=3SZ|ua%)?RF@esbR<yR5hebBb?|
z;^Ey}Dq93m93{vuPt|Ib%f8#!t?QTKrDt!%^z@{6%jUg+EVQ6rgpYVcxVOH2ps;ph
zL-qAYZT(dBTEXj^f0zq76h{{H&{quf_bW|p%F7E%55i}5q95l@?!6<|UyePb51D=R
zP;p!ppV1vJ-ntsEKL3o`Yj@mv@Gx%Qn^pVd`_pMb@XREOdIrTu%=>uUNArtnb<Eo_
zXO?qdbYtjg=kx=#N6$GGz>W0rq}jbjn)R=^>E4QmJ^f45Un~4&3BEG^a4!ipJm9>r
zrb({}tu%qh!YDvL(rRuE=Sk2S_LIUr59j023SZ;~{3Q1$0Y1Kpzw>$dmkm$LqZEAo
z)_<h0cNR7uX{u{2aH7S)pb@{~0E5{8>9dMpSw7w^0C6T|k}k+XFaauE^$TA5$A$`!
z*iuEX#>i$=p`Hy#hGLmiDrQsPprG(ggE$zu>E<^I`H{v1<DqHAEVONCv5^HPPujo$
z6hd{P&-y~E=~XXBN)L`DEtCiLsic3%83k{k>!Ey5(_RMu<pWDY+SR;o6hXR}wh0h-
z+h<G}NZGwQV+O}|>C=ADK~xM%#x((aWo7Z7%mN=DBS@jM?C$N0%U7<7DZ1ztd!}zZ
zsNy+yb@8Exkn5xDP-UqP14p^(gCj<WPM-nZq0i^e;{jmeG%9dR_=scY4o44L!ub$D
zibld}nfb+epBp@{4?hesgM@akqS_Y&CtSsYk3TeEHl4Tv7WOucUkoHqdC;bOL{Ngi
zWW=Q@NzaTePlf9CWy5I&@|66SLYLs8ARUhMlL&%e%R@#`6u+#A0^$#fSWavi&|osE
zUQ8TtSvO}yw~0YywIl<62J0}Hd~_My?1@3-taKPs2JNiwQ9!~GUPlgESuq1x_$B6?
zP?)Fy&!7OAq=PB(ktSj|U1tQ86B9A`J1X^wm2?v`GjZ`so{`+?fsPVb;>`e?K|AHn
zsz2gI9b@o9oGl2G8-gVT;5>2;Dl8Ao6%Nc448M<i5KY8*Ox`eXA#QuzgCK^MKq&{W
zBvhCPb~*bTdK^EnCngR0#49I>z*jmM0Kim>N#vN7oT$mW42A}W91q$?4hT~~2NRyc
zhb&jeTeNEYN!zLLE}UvG*E1L|KL1i&y?8}TK0;}MklZ$TN*~}}LV4%D8h6`^9@eHR
zEQi;uJI0gFeT?z2^xyudpXhBb9c7Ezep2H$44nk&6WM8Ncy?zewV^n+>CZIKr^m*(
z!D%ac{j5tA<JWXMl3C@2p4hgp?;1JbR~i)~>DKmwPHF?+)`bY(Yo00YgbCC*@NVE2
zGqlKA(jmzFkQ>I5Hn(}~{V#fA;S9<6X4c`aCH%xrS_UKGSi-YU`*l#p5xCzF$f5B~
z#0ZYc){{Omx7G(o5jwJSR59ZR8w^DcK)u#KY+Qzb6xTgD_?SIefx`<F*DDo+>6f3W
zVK%WeJ2>N8NgR+;mv)Jo^lgikexl(zYde>8x=3*ldO^YXC0(Y-_rP{(&ibD2B`80p
zWqrnj{4_>MZyu+_zr+c1X-`S6xH~e^*d^T@Zkp7ttr0!hre(&Tf|JJiVecT8*x+tN
zJ&KQ=s<Bn4kEeaen6Z`yYjl*4Z7D<Y0e=_b@BM?H-W^|6@Ky64aDC%f>>A&1LA7gD
z`Bi<Z3Y_;>p@}QndR&js`A@f|o%Q@%?6RLvGvK(ExMv>gyV8OW7{~fTv!R~Fq0YK)
z$*R@8J}yAH*LYWG#kDG08JVUXhE_augb5HPPZJ;9bgQcCs?AbCfl0&sgmUt)#Vp?~
z`lbpr^e81P3^iP!)nDz82se{h^44vl#T`4WZ7Svht$HIlyjHvOI5L-Xo`tMSt#IC7
zg(un9{CqN;%>$s~)^|-?O(pn=l4klVcdg<RdwPaqXk<DDM<&HfF^N79!(*etN_p<;
z+S%F>Uvnhhk~@rg&F5@xZpHT2ws|WHOWN|-EhggOlZX!!jPP$ML}Bq9i{v<JvG7_<
zFn4n27ZXOOUPQ-2Cnl;1n|nC87iMs9#23knne~J^yc&-MXyV5xR*t1R6>kUssN8X#
zWO52Y!`Q^QC#2aWh8_6kuY{8e`ZNX+uRPG-9lhe6;Stw2wqt3H`-XO1Md9`O2L?Qu
zhp=jOebeDjPmIRZ%hU4L;YEB)%Nu&uoYt<coU~=PnU50FSZ8BxIgWPLVrry6p1F1{
zCNEy{t|cZY5KJIU6JbTL-9jUd6x`Zei{0IA?{w?y;b<@Mj`|>Qe3$uU(TDeC;<gs?
zkji0iYa{;nkN+@!_h0<3g$nHohfDN9fd8Z%PT}u0MjRX%iV^2eFS%!4oHFVbohk9A
z*KfTXqmvgEcI3tT^9w$;gGqeu@L?Q5b*d*TAOvNCePEzBdX)#g%yJV?)x+W8zF3^Q
z7ynZ6`ux*Ryvm&&Kue;#tURPHc--f{;Y9h<6W@H}8}ai$`*Yq^L!DY(U5&-1C7%#+
z^Tu^g;_oXj{_v0gD1QGB{vaOBFL@^lyLzUkrsHd0dn10~JKs?~pYTK~lbPfp;rO%>
z#T#J=!lnnakG#r%Vrs@|W1N#hioH<nh&JU}=Y^xV^VtXS&;OhMPi(9&#j7v97(e@S
z--*{>|5gl-j%$phvGUGF{P?}M;;pyewr~@HBX#=amtPg$*L_zP0^QjM4}7QE>#w~Q
z<C9ad!|oe)LG2w_5W#VnYwH|UxM6{)kH=Jba>@j`#Q29XkB@UxIRNZXB=^{XGCegG
zS7#<;l)K9mZziI5_n64tQMy?4mq!Y=w|B|U1%<osBQd#aOEp3DQf;xnKiC13<B2R8
z@{gW|w&d}g#*Mf@4`DfD`CP<Wg$?v0qNg|w4-I(7j)knM?<|V-P7$^S2{=)M@g!~L
zNcG9>S$4e|-%_`rX~DYOsLx4*JBIcY?zZa)-`m{fCEnM~J^KZ^P9}cy1dimrXv-eM
z%AQ4}oyt!HUrdnu=tO-(=*Kw5Gk2FzM@3h9{Ns;_f3A{1Fkr!%;?MWir;)hRVWBCD
z8_6RU2($Z2bx-vV8VJlrImU8;I~F^X2GX(_8ym~9wz?Es8|$&Tz8<S9%d%e&?yR~z
zGab)8`%HZH`DbzO?j4_~!cHXkM9Oz-cR$9)C*q}7Ue!4BMc*l>$f(@Pe279maJtpE
z^!KShZVPo~s^7{_<s(9C1WQDpIRte$E`zD*OWyU#E}XTk-BLcCZs#3hV##a16Wn`5
zU8DT5LvY!r^wFl**48W>Vpk{WVDSO79O|4DUVcBRFB}z1dqPl&KZG_ya+#2a+uM8I
z6-iyNo8p-B3M@vY|3ILSySpe4j_jPBos9?g@5SxUK2zSV#^mIL+TZK(+;h*m|C)RB
zFt*j!#>dB854-f;LrBQCjm)^vp+37^eb)M3+@D*r(1o@(IMAd1Y#_#{+lnJQ#Mn_s
z9qv#b9&B&Ky2^5ROl|V&RrQ5KzSnAbanWt|h&v)VHcxrerS{gV{sI9pyV$n%Ex50~
zVt8cK_79i0VohnE9O{afo|%eU*C(P&?R)nABVQ##fA!<{-;Fna^tR&M9y6D(dI3CX
z9Z($|P@m7dn|B2&&o-79<Fh+=<I&<u?5R#{ZXbAO?waz8BQ#a@y_@g3ThGKxFFY4F
zrH%}bsqEE;FQ}dLAeheMLAfiRzGVyHv&s;@@PzNX(n|Sw_qXz6O<{ij&wdbhAI!Sl
zohqN`j}d5n{gsz2@LSnYA9yOBqO&jByLv5<J5;$34{=BOxP?huDys)`3%=ix#VUM{
zUwZa>T)%dinywEM_s{P}pXxWlNxt*&5G=H2aYb)$zsEgmn|racxf|R1UZ1ExY*U?U
z@90rG&z;unI;XFpf1)3<pilPPC$^~X;fliNo=t?T^oBE26Y>1BH{^dm9zA-ba@|n-
zXp0wQ_u9+PMUUFr^6H9rJF>%WWj*gvWbw%6+G^atf7c5L=r>qA1fM{Es}!jkI%T$_
z|M)j4_&P7}s?hXT)CR2*7Iv+w{=~VOtG<Q@oEO$)oF0ctw?eD=Np5Ft&cc%fL{AI1
z!uc>-;fvfVd>Ndl!B0xhd3aL#{xaZcd4%Ar{kz}&E5Gen>H&=pj?-LKJvMqIk_PKs
ztI<`(w$|4wQ<G1+$c#GVF-{`d2K#Dfm9Ws1NK`kFGmo%Kdz;bAPqp()f76AE?Ne`D
zr!@6gHo{GWVs|bq>D&OyQ{$ZOjg&Rs_$jkzsW+gexnYGJnJEirsh*m(#oA?3T^eZu
z3mg;5r_yaqU}2Vv^#;uHS&S%^cA``@F%u*0(2@Q|**IN^Sz>H63<UMfwl$#QUjFIk
zGHCaQo9)1?NXm)(8gde8GK9q2k#!tw2h#|rD1c#unPITVju{xDeGi^k{gwW`8ioYx
zLBNpAM>3T#9SmGK2IJAJ7z=$=JpdOjM=jFfa}1n?IQke9OXstowws#FGf)%!1faol
z>;udzC@@0lkc#P@2i%@y|5ZRV9Dh{38st=PnfukIsdrkfW#~yQ<iUZrb+TztQqhWG
z!QS3rUd3vOeon-k$gGg`!E*VDw(Az*)94&Atdkzpp@A9Ul~+Tmt}vKk7a8Hez-Jvu
zfe>_$Iaiy(iU{y=W|JX|BS$83zW^)USdqv;z^m!R6tS{0<5C6>qC?#v4sL6Tg9l3p
zz791wP`&Ua>B1qFR)xo<&oCgQm%%^4dCs_5ASHj$rq08l<fICN7Rcm420<+YwVHnH
zF<^%2hJhx}U=&b0#4y$VfP6d_lMWr8DmW(m1|8`$P4XhBWnpKH#>ne2F|tfNc$XSK
zm<;?n7710CZ8mArRNYnR2l+<4N?3TZzv{WzCvMeLmAaO6k|s}jBwtcj>>Rc*rMmmU
zQ?a+wu&NZK5-AL6A*e0|L?~(Nk7v_rwi_>MD98lej8ncH=k(|wz}qJ9@hp#aDTJJ2
z!boE?TohxvMqo&j4GB#w90@~?_=O%?J)ql4s=ZpDaj14S+rX>oCqaEMglNxAOY5C8
zEmrm@NLtmH@0u2)*Ow4$Jc~_2(X<9<GPbCcGR}#THr9(+rrR&$n7m&J=`Y>wx2PpW
zdQ8dhs)uOied1I<)>@Ah*fL-)0fUh-mQjjzs77nzCI*o4Z%Lehu%V3w{SH<}!qFtE
z7XmS!@iWe;!?yTw+?xIYd$pRnRHd^bv;0&VZuek|27Z8w?|dryrY`*^-Kw*7CKFQF
zIpNsJvPn4JeCwURTTK7P;ni>b+Ha%QZxY~dLou5_sro0V{uMLOaHIVc>+}AA6e6Tw
z=Lc#!G}KgD)PmPAw?L~sKZE0*-?-|2CT&)l@wZ=|@^d_@JO6?wjh>9Zzb&>BzC2W%
zR8{)q1JL3IRBbCw*E5I+T;I+B@yM`%if7&(8Ym^KW;m$l=>VeMz-iKY99sQ8iB5}O
zIj64R73=e~pBB<rh9^Zw#2dk(+ME{W(a)O;`WNH{@T5Si-v$M&43$>3_1&sWd#W6{
zFrC%&2th8G@9Pvo-jT{gG<*%bPaj904G(+5nsPSJLmd`99(Lj!dXkr8$=KPSJHWUI
zAO1Thnv)M@Y~ck)@<{n})HdS?1Xx%?o5}0VjV%k=I9Abj&4>;=iJ;dZdW=JuWantH
zZG?2Ydlp*w4yw&f-)S_zuoP=+>#?%D8XN1IIUy&S(B&vYc13MVW!J?1zW6z%V^4hT
z=FXuf*_ga#Qk@CBV|B{Zd>;+A&z(~HPCKi&NuP!<Uh$6L;y5~d$69+Q_jR=U7^#7w
z;ka`3YCLoEW=zanjGn$>A2~^Bu#;?UW8IUvOb}CM<PmrH0tZUx+`|X)@kbxVC!c&A
zcR&9$-hStuc;}tB<HsL;6rX;6JLZ(u`NdV=!}if9pW2>P|EJ=);jt;}V~xQOIB-PO
z;P9v?Ql}<oVrJ%w^y4u!JQmlkKNDa3+Bf2jZ+ttRd-1gx9GMXR-yR!V%3tNJcUh3|
zQpJ3~lxQMYqpln(-YjxF(pUiD-}KbD>O@a`^4XpE<kQd0laEhK`8Ya+hrj%tUx>?>
zF8Z_tjxD2(u-FWN41LMs;$l2}@W2zC)D8H4cDk_R1%UyAO@y2rm&xuiCRY#kcVcsW
zC3ZI0adsTs>!YEdcWzO)c6T{aorQ7xp8O=<yX;`04jij2RXsWKjlAMcodMN*gcAMT
zoiRGpAKZ(xwz{Y~x)$I5nV*U8e)qeM1CzRnbZ~Utt!J*qYp=f+H*Ve%-M#qnhd++{
z_wL8k^tAJuyS3PTbL0ATl}Dd>f9e<c!!8vBAndH@>Sh-iNBuFK>d%SmDeVto4CTT2
zIoqTVME8R8FHgo`_Y1pVXj>d_h!Dww3e_`&!71cn;i3wZ_DH){eaW^-`<J~QpR~bv
ziZKdtP+oGX3+1?{yd-?w*tNxO9UoDsusr@&+kqa!IfP&A{9=)&7hiU=c#%o-oKUtO
zrJcNYo+v)dJusI*z|J4JyPr_n^Z2tACLwgDU!dL)Z-ldMpURJVL>XgS3$>I7^ik|w
z$z#MwBa1SXAH<3Ap2H_koMflP0vxq%`a}x_`%=%sQ^F#z7;{jT_+xC$QKuYB==3TK
z1R6sF^54sjB*yJI-t=+#>T~haudy==mT%s=70*8Rd|bbNE2d{I$MnU^7U&Wt<JwM@
zF9IZl^9WNWCdM5;a{yx_!>YTzditHV+$pu|F*B3)?2chK5rQ;?U2(E6^SXDNe*W2~
zasU2<xO?YLEG^6h_hL;?&$xf*ZZ6O7D6hGvth4RZ3kbMNi5)OKN@tg<$v{tMObqwO
z)M$T<4t85G#_m)El#KCr6#l-#J~`M^(A%n~$I;t)A$sMx2SJ>we5d*^7Dns}j~%Uh
zyPL|d!{}jtLh0BPjm@=XAKBL}KO8Yf+!hw*<Nlq`<HHZ$jkn(XVf^R^-;a0S{*n3-
zrCI$x`FH8krFiM(mpw<ue8B9(+2CF)j&FVG#TVj*7oPX-9t4rp=NmV!$EC|xT+Xf&
zR0<y2gxVP2GEPMDv64cdEVs7uo;lhZ_w8T2d^xxq5kVfu-?7_okex_9{oZAT&~jgX
zSJmG=oSpS<J&xMk+}`&+d>ibfik^5lzY@#K%f3^JU6lQO^o1j8<6X`x=27S)UwP@J
z__=R?Gp=2ncK>y8Y9eN)CzY-Nw=Wj2bC1;(_4}{C@y&Ss>u<yhFTEVkJ^P&cswsbS
z>657^9IxA^GMZBVH#srkuxNXS?7mYUfPf3ZX}{`0PnX)i%5Q&nJMO6*?tJ>Og{c;*
zZmxT$*t5?*AHVoZzZ5_Fb3doP@@wAt_S!42#^ltL@6uY<Hz$-Y9Bs>vR0K==>hm}n
zwnufkzi&|Au@Unb<Qqrca?dDtd9ss{U1#fCyEH;CN?;fG+WJP^RlUA*_rCHkcS3R$
z>?itWe)Q2Ns$W|%JV={SKcKXbh5^;-;enpGG(F~LOzz4`)lEJs>I)Hys~WI7mL0TR
zsw-3u%4BeW`{Y=hz+J1yv9h-5I?E2l$;okj3kKaUovIIFQO)f;_f>Zn^{vVU8O&e7
z*ZC(sA71y#=%k{nX25pkX_%`Y{w?-ce6jg)wZdcW_Dee;j%k0+hyQX&V}Vw>PYdHb
zd=cK0+@4G&;YndU$*vW?$o(hz1-=Z<(*VC;#rIz%tS^G6(fz%D@lXHmg}?K6{v*#}
z&^LldM@h7EK+6YGW*ig4@6-c{SrDYhC5cpQn+;7rxV!E(te}&$(+}!dUApkLcX_FH
z6(^y_w_%^br9ayx_SsMxzV)Qu_({lEsS!5BT5+t)Hvn$tw^ZjtnjK!lYx*?#mI#`(
z@f(N&kZdbOKNB~i0Niu5a0B1^gl5=D`>a4JY4cA#wk__$wVz^>W+{p+VWdmq+pmpH
zy5NzIC5iH5afm+32rhCJ=GE(0W9HI~&DEW-O;0~oJn?~cn9ze510_D<VlG)(2!q2h
zBik@YVgfiSgux91QU+)+p=R7f2Zs1a4=b=(%>{Fg0E4aB_aEGg&p!Svwl!$<;6_ZF
z<Dfx?fqqjz*(lt!XYeiFHxrmiKdXNd$Q?!=CKiUBIC1P^!&8g@pG4{_0mM)gwuMY0
z4C9Q@!ruhBF_<9Dh8l43N3(^Y!bm;H4<GpfBVCZ~h%hFN<lzM|oiOW^u?H9W0F%EA
z5@7IQG7rU&iI|yuR~QWtO<S2u9iYCk-MXVqA7Rm@3<!e<K@tNCE`z3QT{lI8)i7Yf
z2(t2&K}u(Lua7qz76U#ahOVc-M+}DsGaAr%=Yi6%0`bM+)C-sf$A|KkJp(!JK!8a<
zRJ<EUJiuf~B|Qv|h=V6)lqL)4q>|TUV73V|DC;xvH^>coIj|$$#GmNMPP7;tkX{D!
zOcW5F2ca4Sv7!z}k+RE)98rg!>OS#zTO@2MD?#%TdGK+!09biW8%N07)gi|3nHS>f
z<?AujJ1Ck;gWEO_`CGIokWzrTD?|opfs2E~O4NFvP(=f*7ucB2sT7A(l}REIQ*v55
z(XpBi=}v;t5{9(to*G9*_7y*=6mPa|U}w4ng`91RR?wtiC^W&ai~mh(l!j_Rb+x<8
zMOsrWfw^APK+vkL)OaLzI-{%9Y8W<w%wLu!V@>)qonlWo*``2GKj{JV?Bk%<h@@UL
zhx`U)Sk;4_^@diB5xuwRHb8p*&P-vXu|QJHP}9~?7Aj#jkW2imv6&E*4(h3XxeG|S
zY;-81He-v0Nyy-QF5v#ke_jQfAgst_LQMOCV+sx%Q#yh}JEh)J&ax>9fCjpXTfd=p
zPFsWfQ{iVlY37gJ6-1i#Q+slsh8;hGSAa>is<L$3My+N%IHd`0mpBt9{(!X8U-s$G
zdYlS?Yrs-rI6wTCg*-QLMN2fCe_#m*lBQRPRx^<CE_8~Xx?*81dOwMUpR`ZxzxPu-
z>hOQ?KmR{ugamv;fXu6+@|W6b=U*#MiwaoJ|C}528BC%&%P4!(IvXzTXZ-S)dd9A(
z^?L|Zf6xV>>l~F`BW(>g)U#pxv0eJ1j4NN3_)R?gCe5Z>g$8a3i^QG{*SaR2RMO!*
zX`y-64{1A_Pnn0ZqYXYrAAPouYIojE?=%P=!$B$GVW?pPt+>S7dWG-!IG%DRPOa|P
zVQo_}7iiTR$>FuyoyU>6=4~su6Kd9-3YgDX8Cvm(SB)ERo`!77(}X95^CYl+=_Fj=
z{f1}aDyZ)0?jMWM33eh~kG}q4@%Kz7ClAD=DtAYXjShGmu(-7172$j0-SX(`V+%Z}
ziwLEd$YIi%#Z#CG6(Iv-2PRn=8@TVS<tf#I9+eZ{5XKXEPnUe^VZs>U2`3bo$JGPx
z2Tx175Zck_!p9zoKP7d<o3~E^Q2Lkz<Sr_9X5b%xC*sE#XK?q(V1Hi>j|@2+Ca8#`
zcXfz1{4^6<@P8c5w!FM1bvrm}X=ZvX#zzOEOKuziIJdYQn_EhscuB`u{Hoi49t05F
z_rfGH6UF2g`M@z^9nvt##tsDpR2+}QU2CiBYn~`XV8<AS-CZ1^_weDP-~xURER-*l
zS%3df^0dka=HAFR1Rva`a<IEiX~*Q$jK+jFqi<l)la?Fn8=}1v?c&=fXD-F18SX6U
z6%Wp;|6P~!`qqwxUX%?2NEWambjWgI6+g$t^_U08%vgawYkX`tUViar+_-in?%ci`
zAAI<+CoWltxwbSHJ1dXkTd%(mufBLIRu<>tgCBnoTRR$OL`P0^mWg?U02VB1j6?a7
zKkU?Er_IdFOx$|rrf~XVcaKTcP47tC*jQ8g^UgJ)`uM7zfngsh$ZiaF*<HPQEv{aB
zMrG9<i)&l4u(GMK>4}fOqCS&f>{=Tg?u(JZp6Ks5iSYrA1v*dSy?5S>|LVW~uiQ5N
z#^3yR<L1p<s=vx~YTe0!zUe0llLd0&d-49@_kZ8VS3Udui}B0f`!$sn_vu~mPD+j@
zM7VmOae)^Xs_N4=;Zqq?v&(B!WwE-x6)USsANO&ou5srU3reZWj7iu<$@q)$7)LsC
zXVmz}U`(jp4h?YcQNPDB&}8=@$86DN02Z}Ux3|^i4-b?FiVt-ug_<d#LZwO4@6qPz
z3lKgZ9342mv@=c@0cgv_iMSzH-rUOL6&Yi)>x*<Vj-!nd9|TymX~wMhLD<E()w{bG
zCsNNC+bVr^=NI{c&>JC6&P(L+gY-8!R%6bazKHJ{wi@e^UO6DxVt#~OZG0bmOuyQa
z@48kxNGPYHFb<`zQ}3wfq=|OLF3d)rD9-eC#NRs~75_X6u2bdADJ1N2p?{)p%f*(o
ze<scm{;-RY`4e^~u~QKN6ZMw+u-e*lQSrvcuKO4Dlrc2e8@<Zr?X`vIX+KdI1D@c2
z$Wg27o30ZG0XNrIErgx9c-eOuZEUQo-<Wk8y~9)SqzmT0$FA-^wfi3B^+C+dFIsTZ
z)7$SOV4uBlIbMC~ISUTfH`IUcD-Fs+%8EO~nQQT>D^fZB4PnQQ`VEnH`=P9v2U=JC
z;{GHS2(bH-Zv#g>B3PC;`$y1qtop=}v3y6l-;JF|^NVWJszYj0o=fOef5UtPJ5agv
z@cC!1TL`i)9Qx+jd-vo1?0x%ThcfdTmoLu5i!ZzorwYdl5w2W{%a^ZMs9bpw1ZUj}
zgL=g7N801`%(NE@D1x4Ep^M$yIfw=I_4gkwI85$BBd^Fqj#g#Y7Dom0ZDHpb{UZWt
z@|VRH?8-uzKCCv$y=<S{osEw^zaO)6OEEa2uoR!co{o6!#b@Ki7hj1#`qMY#2XDO}
z+iIJ=L!4Nn_I0u!FFkuDzW&lPadm1iMtW5jWrraA?!5=`(FY%RN9`-Gyb#lq6RxWW
z1zA+WE~ZY!v43dPyQx&s=#JdRvJEot!&Mo(zGv&Lx8HMHd{wkwe(`w=o}k5dWp79Q
z(#o3Z_Nn_rzOlq%UVOrv?|m5G|G^KVPwnLwf9_}FnHyJ~uk6x&Mt$ZLj;mEUa@-(&
zE_H%MBk&gFi|Yz~joev`LjBp^V<%{cW|w!_O^yv(aC!Ic{g~r4DYZj{i!6$ut#TCM
z2)o24hSit%tL~~Us?BZdThXuj$?mg%^?QF1pL~8Np1XNHe&H9s71wWE^S6Tg$~x3O
z*qzSKNCfxvm9(=B^~wAD^xO2=>UWfH2MT*l-_y-4<*(aWhpI9Cf$FioF)T*uQGJ~n
zABrjQ4)hPl<g@y*dk<#gw!V`H;<8Srpr1-_-n{f_=&jQ3q|(<_@Ek9=deRI|+JRhg
z)4dfBdttWtvw8-<2~B_1p8j6AqEqA8>H~ANZ-p8jKzdps<McS3r_pMDlG|CEvv594
z;A!DjI3GqUe34s)FN5<m_(|zG4^K+pUj{rak9^eOfB0MfTUA{GR*}2Z?bZUB2s7%%
z-K%3TIDyj9A(xCQRmblH+}#D@(VZ0lcFo+h%6C#ssID15%yL01r8zm$#Il~RozWb#
z;*MJF3J@nbRXVL0XD43LKXJ`;w1$^Ze5Ed(3p8EOSEa_gNdai3UEcbeLJQ5Lhecx5
zO4f*H+8WCq$q(BUAJtBO;@b4lG}|}G&?ax3C|eX2MS9W69bm;4Pqr>;D`8^|8sLd>
zIv#b_?PBa%*>_nVcg_lCJ8|oTeNs{Q>abz@%;XXu{mh_=@#KtaCiW2m&>6Erl5?P8
z#9-7o^3bczEDV8p&J|d6^gfzI10a~v9WlqtVq`ex@V=fS4J<i>+JY|_Y()j8Cu2l1
z_12$Nr|iBlR7ZgHR{eRvP0gWQgAfZtU`F*k4ZeufpBqXX5>ytl@br($jCy_sum<1S
ztRDb6CJ*?S)IP?n8O|y%00xeuW)OUFoM5{?)Lv26*W-ExbKbAPKC5O~8OwKxRXQm|
zB030#G|*95o9Tp^Z%6Qz<wu{1;D<oB5k~FqsjRp$K|PLex!f0K#eT(%vWlliOl}X`
z=~<OIIG}o{dZ)p>SI{|~Ct}z*ijX5a#K;$BM=YfW7YsJ2Uocrdwn7Y$C)DMav@$4x
znIuzSYGC5+D7vmMWgwTk4aq0vBLS25s!NAr*1&N&$Q>2N!xIw-3zaq0N9sR}pI3c~
z;pe>YZsjp;2)Y(v3egjfqHkduVG@*n&`lglUY0Uv*{aUVpN0%^<Lb@0e&v}M9~xJ<
zN{ijIjyQVwbJXRtR1&}CMj0hwqc{i$6o*EKfVEXH@sM_O8}?~qhGiVwiKCR_V!Q@^
zn()@pSGouuD5H#Ly5kRIQk~(*2ytneX#pvzCGCROu&Z%Mo4QznC&NiU=>sUViA$m4
zc%`Et?UH80Xft}_VYWsTGSVVPu40vTNzpb<9sY_7(9Dm5hq>sBR|9_VWp|qe*6>oe
zEw~Nq29n;Q3668ZZiIsZU1QkJY6G|FMwC)7JE8Cm!OUa4)Z|O~(Bvx!Q7Yl-r<|5v
zHYE(#D|;3=5U?|jm9G7lJT(qBMM)y5&oV6G6@2#zHsjSU1s&SLaTk}k*ymGXZND|n
zCH{pDT?-kQ)CFp*xMe%ewr;w(K^wDapvad!LJz%ci|{Ue5H~y91Qer_Uofz4Iuxzr
zncbObJG-!Vex$q2eiK?j7O01_uB6oVgwyl^J@q~fqW|3pzN%mdAU%yVq1T|g8RK-T
z?m#PUK|5<FXr*tEedZlyly2Zw`qd1*{FWq~r(2=YH!j}nj;4h^9~aaHjq)wBz8SV{
zq~@1e{#*Q2{A>e+Mcqud%1Z-P1<r>n-C1}1TBFBxRvP}sNxWtrr78`th8bR~g6TOf
z&A3(jgr+~>ad&||R-C8fH~J){G3$AWzY6z^J9b#xRLlih^+s}dt#;>eWG;Dn2Hb|8
z?*E+C864tO<5uA;jilQOPYUCSVC$rweL$^0ILW@fE4l|J;_CJ1<I>ezaY1HwH8GKM
zBBo|^WXMPRv4VMfYcsfDpNV27_SpfzBs7mF!^FR_^9f$4r>iIWd)Xb)<95N(eF!v|
zH21=%<8p6|cqde1L_K2lycg9ej<l<qcL=SS!1IK&aJ^Hl2*h096b6&>!nU8B2ojF)
zIEty4T`tle!E5Sa?8A<GcvV(lj|`8Bm+p4i;x5nTT?dEqSUy&2bFUV=#tsjTqPMTh
z0-jNLUAeJpn<Ec5w|1Pb2xr(O0rO3MvqQsoi+Qq__O0|Vxutlp3yED3V}t!_f(Nm(
zvJ@-J%dw&SW|cTcVR4Kf<ErK5RgI}uJvpZoxV+3mD^J;(=bbm=j}fwQcTn5uk>h{u
z`jwcubV+5`85`Ssv9Pcdb8}1<%3VA=JDfOf>(Shbk6-fQN9Dhd*HiqxBd5F5WklY2
zAsOk?cmP2H6LpL?cDHjU;pD_neEn;$$He$VJX%_bNAt_7AN$^g)pe>d#=%xxoEVHN
z)1%76o?w^Lh3-KgrRICk78aCsZ81JM;bSKeM9e>Y7$1N9QGERQC*G0q`m0}4{Tz*z
zwN;n<^UvOjD_1TDM@jZ}_r|)?@#fp_#+})R%CDZ-S6Z=u{ngjw`ZKEQ9o^n-v9A2s
z*xdKsL<@@RuF9iVc{epa9OEN>F)`2?mlW6W;a=s(R{X&q{bBszhd=T?UEluJw|zGq
ztNj^6tgo$D$k#WZJXO7;{N8!*z4-W(&*JBP{<|8N-|*3e7OZN##snjiyx1|>&Lp`f
zp;fOL*RYELVKm3PSs1y#;~fwPBDfEX@?mVku1yvKB6#W1*lSpIX>xohW+plLWkB^p
zbwp*yWG&qncpB}*+=Kh^@X<pHBt}LiVqkb&;S;ZRxl(;qht%L`!-pEXapV~Ff_laz
z^|snG?ScBhL^Z%<FTx?(CBg{CkF<9LI!q$7bBc5$Sn~a4idWr%X2F*55or0?M1|X^
zXBv+&sZ4t!Pb~OSpM!wRydk&>-+hKuPY^2(Ec#@k6<dz3<6fYFfxN%aM-+-C?UZ)U
zoCQ}q?5LkZxCd>2-|2%G?<$_`nDcHm)ma2zgh$AviQ^#=ruV7e>+DJ%4Znv6yWWks
zr+%=nTlHP>fcBBn$@p(<Y&<SszN$D4#J=LeZdk^+zGF;fMn1R?(U`iw=R%D1D-Oz&
zJ=KwIj)GHJa9Yao!CoxR&&JZ?eD0D_|FFEY5I=hJ2Qfc4>!U-Tx$%s5!bHbFoVNEU
zECe6=4k&E;RQfNDksTlI^D%MMNBa0J^+zY7$xb%z0;3A~jyL+RA^_?xyZE@%sY~X&
z_wUEYpWb#IM<7I;7;m$Ck~^Dd*XwKR>NmE0wBsWLbcz9Yf-(1Se{Mnjk-}9T)9+h2
zqr4#PufO<0d`o={yJmLOZ*b3``yqw3p|-KCw%ps*tu&0r#hK|CQC~92QMe<+p1*pa
zxXsPaxsGrYUM}`V7{O6<UGc(;&v}>KuIICM6?f&c!a?x7xU!~xXI*_wyZVUXn9?_d
z9b}B%_w@Z_euX;SrEehW>Cr(f%{}sQc0&mL1_xtdbuHd}|C3nXQNJ`a=DO3aeC!DI
zpQ9u3;`7h@&aof7^<K=*>)Wz>==#Q7&*b=E%&33j2;D2=12Hu=7!%Wzo`>;~d3SEB
zJ?wk;(aF(%9PJ^j+E=|;QaSg=^_$Pe(D=0HIM}h|{#V~i*EPm17ur>32k~fO*#c2z
zQ49@nzZ|>E+C#NBKL7Z~@%CG9xu2yiA^0ON??0N4cR#t~ooXZMyRToq6wltgBEQ4Y
z-qRB=zWAbeic88zj(0tZy`wxD_(b314i-`<eLh7<^?-g#pTAgM+l<fe&Bd&~S)`+1
z@txK;hHo(aA9r%?iVi|3c672UnD)wCWoO%ozq{01=6pQ&q)NY~zM5Uq3yX_!<Kjeo
z<Mrp`#<k1QKhUR;6(7YwjoNL4+Q(6}i;JuJ#_zhn9USOaIPBhz*mj-U^$zL%J=LSq
zk0D?`7DL32uM>S2Df<z%E53vDealPB-eI}Ez9A~v-&$3;aBSpftKcgKswnAJA^DwZ
zhbpsG8yj?+?&;S&lEX@9*tg*G(8pDCt14aM78*_agvap{{{MonuLeIUJ?G)S3&B_W
zFaPqd{&o&4N@SZaS)-~F6+806So%Vy9>86;boN;2<%+L&s=TP^r81T$kx}zWQ^uiE
zKyRI1nK`%=WRq)X1jR3}YL6lRwnMG7aZ58V^yrFPQC;|sH?EmSm@-1?qxwTF!KG~>
zl@&;j%;ZRQ&oHDTk1H~Eg$O2xf!+q}N}htI{NRVJ5uE=ee+(_EiBx^l4sE`Rnp*1}
zZ+9Tr*Z5T^c4z8`6s!0pE~)oJSfELporptF8rV|b6GOwpDS!~uf@@Ahpi4$F2fs>t
zjfNidb`xV^0hm5MKI%{(aXM*sfWalPTFmso7cRhu%DZ*M*f5xasb%2CeM)?6clEI%
zd{((J07WQ8etYK<s?q>M&Q@Q5w0MGn_<b30B1`;<+mj|I{!Ia4(;s93N#Ud8oYwnE
z@lfiFis+h$wwyZB4J~(cm7Ykro|+UY8o4VU5ilV<v0e}5f`G@n=~OmeeJf@lkB~5R
z;jq$;a!w%;<PdaKhQlp_4o-LsfMF7;3xw-9=|j#va0p5<cZ<p3OzaK~2$+}%7OR+T
z5R=BLH-!xY)2)FsA6XboCfXT5Fp!}xc+f27nt>H(pc5zJnjd(@rNooq9ftIjGaq!0
zhl82P)tWg-V(^cD@`1XHJ9Ufnp{L$v{uUvk^dhM+SQW`&=TuB+cJgLmHHhsZuJT8o
zfX}bBzIg>8<><Ik+gXvB@rNPL33gU{_Q!DluvIRsCK|+qSB(>O0f#L6bR&NG65wDD
zsWO3$BR&nqlY&nWUeY#px|5-p8}?Wgx`w2ero=6w8O8XR#H?2zjx&nVYMexBLQ7b@
zTTMo6jf-CX#@W9O<Vp+3e=BW4rPR=gh$r@ov5*%@JYq<dpL0-FiIZ*xy+y5Y7T9a%
zQIncX&JWr$meQ4chBlA_AA7}Q9EWKH2H6GCfIENEfts{3aOy2urADgBR_6*OelBAi
zt2_85TuAV$^#B9<nuhv)F%U*7<z4I>0fQhr{pr`xw+jA>Q-bZ4B0DRdq@&@Q@%F#r
zA=7CZ@>6`LQ$fTHdWEFS)sRXJf%*xpOq&riKR~2!^t6v$_>SPwZiH6+mk0NZDgK)Q
zona91E99n4`be@4QC1lr+cxZ|3sSM?NrVLB6vxIp<960A)x<11SCrz?2n*wdpXyPE
zFMsn_f1BLPJ}1F?N3m(w?4O9QwGC*v6}9@Sswb@kt@J2go6g}@u$jNqM)|Zrt6lZm
z>Zh8)qJFuC;73kPzfC;HrnpuApfy8DsDAC{r<$Ao8fvY#=!uJPn)t*m<Cj)Klxn}l
zPlYDk(#P9OnPt8Lb&v(E#LxV#et`@-+d<R(xY~qPzeOVe@4O%L2pKy!vb3*H6V95x
zIyA!OU-XS|zasIBV*<MK-do+j3^d~N1b>x&De=<hq06#5KHLwEiG#0V)i$fQyEKMi
zL6uAWg7_No>9AcK1vD`>5|iU&F+4nI!Pwa7P>e{=?yBDYUhz8O`3ld&!YCFGF=@@i
z3To`Sy`YE-^LxDly|Y8SBW?<WiRhfnmkK|{PtmCL82+xET`nw!k_xRna*V}Y;-};n
zKk&C4$Fz^YYJ=TVOt3N$&m(?QJSP+L+~>qDH?g7Kb%UTO@AqH@Hxv5@(c9e?(_@1%
zH98nWy$ZjbNpN=CbVfImkr%|%p6pxL#A<Smmtuzo$AGc=n;jT^2)=sSV^C_3{FA1A
z1xProeCPfZ1P1-R12H}}9y8O^aq-gSn3}#AQ&X4X;^pfxI&m?^CNIUcXJ3lJk(t=p
zJ+eTABeXd33f`WbX9z8aM<?Rq#Vex0j;=``<%Do$X>B_m&U0ths)ZRG>C~?{9f|i|
z+f?4Lv%`F=@&le(8N);#rz5yk(ysNu<K`rEm-7KT`UQj*YpWX;(CxFsLuEwWBX2pr
zs;_S#M#rb)=#*ogHaywOjyR6knw*}AshOF$eD%80Ib|V5UvF0o^>e?A>@QA99g7#9
zy%{e)|6GiWjVTXLV^8BvPNE=vF4OJZ7#JLmtIs?WSFYcRE6>~#&0EU9%W-g|aM?+-
z!)YZN7nDT^jBy#mawI>y3YiS2tT=kjcLs5vk;bo{;8t4h-My!@?t8aRm+CQfmU~@h
zA3pL}3gHNM$<QuVS64jF7#|(in3r9D*+;5^E1z>>lNQNQi!36QvnQc9H#Ss1w_>2b
zCq|ig)|j4$w!}^=1UZBY??M_Zpz=<o+)0Vh0(Zu9DNN3BHscbGpL^@gKa2O?`BB`v
z^Qm{sjVo<47iZk=h=cnRX?7Hc`MC$NxG<~!Vo&u`?H+-j%AGM4J1>~%<(@E9ggPE$
z(Z4X4z(g?PyE2w!Cl^BS4W*AXas;XUa{Q?9yh~K(JevN%$6<0*9l|KWqjGvOS8-De
zttZbBbd$fNk36m4AlGNcc{b)wF9d=1EFeCdIAVd9`-mOqV=ng1_%beJVJ2q!3k1=)
zXBhNL94Fc3evdNr4lI=ed@V=&-TVAw<?|io!AXoyOvY8!@ylx8lan*fOZqW{0JK3K
zc0qcqsdyi|T^$6c<}k^>PdTfNa|A0p29MR2NE3oX;>8ZsyAS4E&)HSNv4^*wxe?#^
z`s?w%@BK==_WB!YUzek;o1^wlV}1J|Hr1Ex>~o}R-Z!_xy@&F@s&=xry&pS=N{7NZ
zJndAwzF?s-f-iQNv0L=!&1ZbX=ZMlaI6UNjoJG$+{NWFLzt*nm5a!!=Z%e%!_aEGk
zk3Rlb&qs0R?%jAW`!F`t_O?{+d_zuD2RFBK_m7h1Hq+bJ8zZA5D&G#(`L^in?u$*u
zZ%<WXWL$m6m7CE$JRK*U++8^s(=(UUr%Xqe`W)!7V1pf+Tgu=23J(F+p&G=6&fYjU
z?TFoz?$|o&itQs+4B4Gt=<?BdTgs!A?c-S9+>fn8I)wgc?;VN@QjZlE76db&0gu1E
zt$L(#+EZY}%ey&MpAd2~kI;?~{-h(ePda1wv?n?TCgZ5BJ08w0#K)h09G~C0ACKmi
zRhEZwd1lNz)TmoNZAlryoDRapEtSc=+4;CVyA-pFo3SCYcYZ3}or+sq&!CS;{PgoX
z@zz`K#_a4|`CfMU+qbm1=*1`;=}0{zo_voE6mL~4_19`+!tYSMJZ|fB{n*|=R=qy4
zP;72)-p2~Ft8YW?W<~AzSb4mmw0HNZf6zCpcWBhR4zFIj9<RLiT0H;!^D#0qrgC6G
zNSo^0VJxg}#@xzw%<EZP-ik*HE6R%{A76;@n|T{{9&c=Hs_wJPQ~iPJ1V<Y`oLlsQ
z1$Gtk&CcDh*{{=wvruPcc`at=m((5>;@(5G3q9o7?FUP-vU8;R+~K_A*u;aq6r!^@
zg#{e+(j2k4vwy1Z=APQaf$F_@1Pu@APy2hjJ<o^0i$0=D-*t9B(#i5Ys`G2~SJV%7
zk#iI<J7BSAw<bI3J(uQhL3T)~rumFr(N{rM1<xNgaEm|N<EN<A+@K4M>L=kWUQ-3m
z>oJ*^tWULX(N}2I17GB(Ui%4lUyTOl$0;=7%g{d$U*!Ii`~qJF=V^f7uj2bJ64n>N
z)9CuB!{7R?|D6{E8>ro`_<CU}#>4X$*&5QRywzHSZb#eEpc=`=KP0SAA1$wf)uf%a
z37#+!Sb?@Q&}K-HlQnX1e55mN>DB2uOu$5OlZ%Nosb{o`yUe&a?e>#+kainudP^wg
zLf&SqoVnKcWz`c^X{iSVyVMe(`srL54Njp)yr7*sn$juVc<GTjS5en|!@Y)*lYaRU
zIDphP05dd-_Kb&`pS6E#Af=P@$cFM@=8?V;mo!#y#l4i7{!I9DG*!i7br8pFOiWGs
zoM=|Szz9|puvi7o2m~gAi1_EYsz#InPGwFR2<@<aT?0dyLe8RQV8hwexidzOaA5NI
zK<yXC5HqX1=9d=Zju^eUhx4{0(RsP1LSqo6yrHb~$U_FIQuVPQ=MmnpL+sBDHCuj~
zP}1W;gNzKQ3^g8S3(`BNN-JzyTP;kzpj4wGAmmAGPejSg1=bv^k^>wOz=OQ9@`~dv
zxLA|pE6Go)4RNDRAoxO%$x1zrFyVMe%9Viyc<l(O6d@Q0c4JW=h#rFrCPKKQjn!{V
zkoZ0?sXa`Z3=QUyi%gU#oGvk0#F^D<9Pz-)V3;nBH{|?tgiu_9ysbfv2N{afu?8Fr
zwmC|{=d-KAps#~FR`9X<l<>jZKR$4M%Dm5qJo1V`*>i*fE42{dTIeM=A74QLxj;t0
zIjM2fiFp_o1}hA@lq0#@OM^=;k?!u}$ej+i4+fR)CG^i@b7)V*S0$?ap==l&5g!X9
zMf(Db5_MaK$<gV!e(jmKJat8R((SN0BBk~`ImJl3tQA7Ggr%3GnrKa+9ot+j>EVQ=
zKe~p$h9CEg4zYF)p7eMBi52$gzu~8%lD~v!3I(Sq;AYOJ+5N0^;9?=K#YJz35ijd)
zRWS>Oz#7@2f{xLUSS9M2wp#b?N3l?1?60^rzKKRrI3;{Sr2KuvUlK|r$T33z5b;Wz
zFs-5PWE@&)`vomj=T(-b>Dft{j3*3OIu0o0784Vgnt<OF$XMqmz3Sy~0Q5FyI`pIN
zS+$cXn?}Z2I;k0=NY!)`7qrHuMiRaW=wDUvrGl6dNjE{GEjfuv-w{AHF~XLDyM3U7
zmwpY#m${MmSQ*29(l%qB=1gPS%4{-vIi2_|W?RXS2~2;s)L%%}`@xC@0XXOny4*YH
zD_k|e!BP~Yx3iu^PM3s&mzkAT0(6d3ddg5`rY)jo9Te2a+elcp6@G@9db!A<)eNMW
zG~t$IX<H#c$7x3;hz6$cWbxD7@%7*TpZ)J0RhLVFM!6So`kcm8t6i%KSkE8aVqb+`
zK&v089n~~Tzl*c;r{F0bt$tedRkgpO2h6`no8~5c;5<E-c#D~z{ilVijmyRZ=T?82
zF1ow(LcQ~9GcIXJeX3c=rH>TYt>D+bLxI~|)6hURjbbOLWtQ|&$<wB}LW8#b*7&sg
zEBN+Y>D$nBE6}7>(jsSvTiQ<(zxdTZ)7<KIR&S)Igar^DaK>F)n;mYoe=?18Rt+4e
z*0`OACY`hDdHStz#!cEry8TWQ??yQ(UE*G%w{IjSXRgH9<V*|>k45jmu<rrs?ddlU
z)Qf;bJY}DFJ$OKlM?>Inu%8p#@I355U~<a$bTLV-`U_vg1UZY5sJKVsC5fX?Hc+3#
zH_u~^LyaXEPl!)T-js!RJn(HAKPa|L)-Va^;{?4roZ}F4_XG71AyrPW@8v=~@pDWd
zve-(2nvZNu)~H{EUqfKl*FWI9uKEW!_ORD<Im(Uj5zI`D^~a6NQ!zb0Vj3Lb4*$qR
zSG)M@3yg^r#z`BCyAY715RnCG&_STVWLI~mcv|t~>>xW-=6f=ciCe|nWhH)+v6Z<%
z>8W!}Y;*jRk0IRBcuZrflM9`)KZ(VK#n{=}@O>{=FU>fO4`v^E{KYN=F57Px(>XHG
z71yrLcqh`@=0VIaF30N1s^WhvessXRIJ=!bf3O(K>pLzV`lBkuU>tzp#sWz?p90`8
z35y^GKoG<!A-ftka!(qo?}xa9#sV6Slk4&hw6Wpg7~m-MlbslB-;UA2j+mUj7?T&T
z`Zz|0gdAzc0!ns+a9@rE9;eEuW0kE1mU}TgHW__GV;;|}t*$9=cHB3TZ=++Qaq054
zIPK_(b>XfkU5m@>v9!Dq8yljhxVI^97_T5;I}(4szPT5xYnv*H<G4CA5zk$ljNu-&
zjqY|I8;DBl+1S{Mx88a)KKbaQxO(kIyz$MSiOI>SSXf+$KlzhCR6Mrg)~#nf$@|eK
zAI10o^iM3<_$$BitMUAcue$xPP!HiU<2c{deB9<d+flpX)Co^0Z)~{!aEB8+iMY#)
zs>}|TRi%NQhU*(U-u*>cSr8$A#FAZqQxn|hHWmYe{V9~O&`ouV`ggD!pMCU!cL=Sn
zt+-yj^2*oZXTJR%(L40cGL8>}=Gw|qEHBN+!QM`cjf^PG*J5OJLUB8ewY81dQh9LC
z9y_Mk0mk^1T_lXr_=fF?26cn_!UQl!RMK7<3nJj6j#1~D<6e!SL^~H}vJ-~|x0F{3
z^Jt^=1M-sJL&jc86YY&Yf<A{aYA$l5uTi^)w($94_>;vWE9y02dod?%*ly~trE!~9
zxIjZ#sXohL(XT0uEHv)(ju&>25hvmEw4L~9%>j0I>}|!u+^ml+92uLo;A?Pj(7V4n
zRA1SR!frU?h=7A5PPf?6r+$GnBKRUr2xEsl7ClyY8apZ&cD`{+3dg1%9&9Tw4?HeE
zL?Cw3uD)P9x|J_eYUfiEL*8A+t{w!xE2=vW=a*w?WmEO;NVT3ZtMXUzMzEjXh+OPU
z__Qte4XTsv>N|!Nhifxa-d#vP!;UVFbz5FiKccqq&f9PMzA$!XUcGW7PSj7cgNHh=
z%=R3Ccl9XE+=a)<K-7=tp1tKdf)IOQfnx;`dhEu<ix=bK)$9IVaZlk3&%da=>vbPV
zeL?uKraWF%`#e%VG^qHo!wX>pb%XENpxU~EbUN61#l2VTa9&;8^6pO`J$kSg{XHFC
z)X}Frr(E)AJJnIzAa%icapHN6KH+w&y!TX>*H_nLOKpihjN={$)W007Ke(VU4(0Dq
zaX+}w6&rgeF~73m-AuIe?alQ#+TY64ZdC7XUYm~ZeEVy0O>J#w?>Iiab3c|>)?Jqe
z)t*%-G5g@5`t4QEb&L#hjlwncd6)EU?DjYL_;^3|)|cbXCqIszmF0Nlc@_*@_gn_T
zxty2NcTsJrcW^j*)c5xEa@?l+L(x1?Jm?n@RFP)J2=D<Mr|RPl)!))D(|6EUuw#ow
z98BFEAskg1u=ACA!d+~6T2H6GcL=7AEihYNQ@?S*J)7Y^9bxy8zTaG_;@!LQ&rBLe
z5mQGIjv}<C4C#vzgwf|Q&qZ6MjdH{!yWLpKLm$K_h(4DF&)AN>mG2a7h3_pV6OQ!h
z8#L0det%5g`stXO8VQa%W>L*?TX!t2@5a(<E)3b<-%;Toi67vEqd}Qf_M}zm;~TfC
za^lEfgnP`KQ7@Sb8Xi)c*Y||DZ>?{{C!c)c1tRo8PK}}oz*~^E>}S+JVZxCHD5&@q
z*r8X2Uul-VqN1t&YPuJ?B@FXJ1#I6!JMj~$zM&R6@Q?D8cGBC9x>E3fRy}Yg%#x;7
z$he(_^EAGQKIX@%HsQ;L*$U^wXoW9wtMFxTo(4ZDJ?G&`>HEuor{$4%e0}eG?D!Ir
zYOB*qr1daS(ufZZ{EK@U_#~%Z*hLQ%+wxMKN*BGHnbm(b4?n!IPfF<_O_*u^#t>j*
zC22X%MlLSGR9fvsPoJoEvpa67OR!Zn>rEVrqbK*QPj|-yjV<-?ugEQNHjQG7E`K)I
zjr?$NfC|AzBb0;V+{7=sYOvYD2V2v@$cFH+z|L_-Us7c~c7=xgI4}LDV<lC^tjDG_
z7W?#Td%aA@cBG@Ve&my37fahY7WNO=iZmK8yaJ3nmbjFf<2H=I3x86m3v^0knzP10
zs$tf|(CCB0M;AM<<Pt-|Wwt(=L=krO78(pDSp~(eFh~*rX0M161HY<)$efrK2EjEC
zHPW^41g2FZN(;SC83F2JM06NbvYJvs$O=C-rC$*s#`DvHM!i!$TVL`Sa9nud-b@cR
z!tz@sUu^V9>-;Tl`1P6EonkzdcN~XtL1x0SSzl3nXCb6eht;!rF>(%I5q$Z9krf&%
zav7krDhoyzpD?#D+AugEd7#NCfWPX6)1$bM6#Vt{FqrC5&4Dpe{ZM#4Vk#K~u&RjF
zk<J$e0m8vM^@st5KYTKCA#5IZk@jA-1~a01#K#}Th4c_d<BPc=u?VDK#C`Op>LIJ_
zy2ap7&n)m#9c189&D<O2!4+XEkMZj~SKJVe5k;85UDbDum*kH@KEgY8WYIPi|Drk~
zyzzvYztn7*$}(~UAr~|#A7_F5J1>*wQ2&sR4e&&Q5b#OwV<w`|K_#SW({L%y4Lj<c
zG}(e0&JMS#)$WYOX%W!Jq4IK{nf7je`Z-SE*eoq*jh``F!f3`BKh6~lr5#%VSC;%T
zy@{U~iPNB0^nmfJ8P#?U#(+@KaW-vG(joP!)>p>T+N75+TUCMuquM5Idq~<Dk8GvT
zYq;Z@XnvaJG+{;yK5@29^^imw{(<aI`0M2iu&v!(aUGG2Pihqxq0uKeDYj>DFMS$i
z?-UppGf{F%3Jzd<LG_0KH4zC5eZf$mj+P1Axj?44=2eEI2R-3dL}=-OqWiH1?fPIY
z&`5tnNwA^CrlRR(-}0Acz4cH>Y;D&>uI|`4O0vlIF1^aYUNNLq;#pJDrxz-=1*HL4
zH}L&Ra8e7Tb@C!iHkfX4$gnHjOiPm=OUh|7dP5IXJy?z4q`^K(H2uE$_D^}o7x|mf
z&;GGM&39D)TFj5BsZCrxOoU~*COpZH(|gVyu%3U5TdTgR{#CR4bu?#j(o%2w{xkNe
z);egAby7{hg>rq8KdS}vY3>H;O@}^0_%+QLDzvS%`1QNg;X+6H6q28Z;x2GHWUT2)
z)fP4bXW}Kdv*stkPjYVw&u@S*i}nfd1x{ZI&Xm`A_;V0%jsMfUH^THcp8loyk%GQH
zt<Y+3H)o~t9IZPxR228NQdPdHBTR&EtS#j|OdE^7OJ;k0F?Kgse9RRq@8KB{V!$IZ
z*}lB8;T6mC^GmU~xD@MaYq7VpZN3hb`+<FQ7`rISqBt+YQhIp0ITBF3oF~LuCwPdP
zRaj7LsqrNbq<R6C@ZoJ4Z!q!8F^NZ_kFX2#;lWOvh(D}CGrpC1#3+0rl#@56zcAma
zaSXy1#xC&5AhFvYA(&Xi5F2S>6|D)Lty6h<xV;g}^K-Gu9XN`|9<!b8aG1x@*?t-w
zr{X#Hc4A+6+ncM(&kgT<;3%<NK$yHQ_id4vb1R#%u(lP8o4YZ$wCdIHiyM0}yRa7b
zA1%hc2M^=k{aH^2lUAoeaq1Dj&+%Z>Q&W6FyjYQvW%u5p(#Ek<JNw5L7;s_#P=9xf
z4fR>*H8R-eaTn#w@l&k)cb1danNg}QEP_jcCFR7fJJcp$;g|XQ{CK9`f=i7lm}K-$
zEOz#&POzGuJ8;-ZHP9m-U*SKRUySvQjTrJ|bgxO0=O@t_M;rjA{NfmXj%1vB_$VIE
zK8j`T^-`X492NI|UAlBBk24z`_A!-=6EbE3#fdvvm^9tq-SZJ$+<UXHH1Da-Y+DGE
zcW5z*JFI-UGOe++=rB3Y#3lK`(4s^6$gaFcv$OH`+wWM=Fg|%P+PnK=_R+(5@BR1U
z?%jLw+3h>=-iIH>=lAZZ+#dLdLU!g24i3Bidg0#wq3R*VTx{f>MF^8s4vc|2I&#Mx
z6Sh3=J4?BsR^HLyJ#aMkKu}3L^TIsRjh)=Q6T$*TJx8Z8I5H9A)0dTR*J6BpGKL0*
zaz}>n$%n1&U5($>hSk=O4o{Q^2+!E%(B{cw1d{|#+!<5(WDz}#$NYHwrbqR~<5lX6
zadRAmxfN)S<S_!Sta~{obX(yH2Gu9>=YYP1cB=eC=t=s?3urU$LcnK1t@}-rwnAUR
z9Ea(s%n`gXdC!hE^zKLG4oPB0J!kP{JvpT=#3XK{hcpsb?@pypGO%Mr?g*|GZ^oqv
z#RfFqBfq;eZf2JvJ2TfewtWm5LQKZq9A!Av*QNT=Biaa(5rncBGs_d<Qdh@GtgS3q
z7>VF}mb!KO6V=6gG5g?-k2z$qC&Ju*wIli_@{1jd2qEaZ4%HspPj;j2U{ihAn);9M
zy-5@7<QwU;pnQOnT7`44y&C&lE3u<*!Rq3Jn0xT4!uxUD`Rs#ubpLj2Z>SBfF2~0o
zeGqTH`DgLNAN*PT=!ZYF(ChZ?+seb9c=_d5<CWLnh>O>5#^}_gnBY-+;U2Qdi75-a
zE!0<fM`gbG+zT;t>2i#E7b{mVFfXBgZeSz^My8{0>`Jutj>Y<J#QfTx`plh}U*3p^
zi>t9FI^6Bo-q9P!7dnM=<XvkUd&hdz*CJ#&Y1eqBCywO4rMz8OQJ=fM9UD7)Zi5`h
z$=!Ipy<O4Y#}Tz%7Tyf^_gI+Lr|(uz4?Fy_e$zK{<RpUgwQWv2qJL4|>YJ*D<NL-C
z@H|@Gi3h8@@nA*zW%Xn83+~VOHl7^ryC38$NZ;xC4K-eNrBUaZGvk;;?yb5qGZB|(
z#@&Y^Ktfo)rgENJSoWQTgvIyb?)?S%SNWeH=;nKI;&~aqyJO>1UL4Q&xAl%Y3!Ye@
zytScnU5%}cwK!GW275dG%^>}p-qXu*p##Iw-LL+7bTT-u@j`chbRjrX9VdS*XzXV`
zsK*cd1A<S-M-Sih+_{}QC7J8#>QWBqyU5~%Rh7xI`st<BHS0IkAJJE{OAVnFLUmfI
zzcus^s0x6iP!Tq<`;p^f4-m%j(0L2!r}8AEJaOkhWyDUkWfnec9mL+T`VQ3}eZk$=
z(R=V6f!BZ&frns0z^>{gB<U}_Yg~Qy;LxDxUa)`;K^@;Pc2Uxo5?$u;_|DAfJF&X5
ztTZ4RbJmv^iVap%U+jv$Rd56(KXn!tUHU0%HJ_n-j(&>@JW1bBtKa%yS5)9>df<!P
zfG=bBMKmx!PN4~3hW>f@BKM!<7x*$bPXqjZ72ki6u)YYMMwcC5?Z1NHi&kQwLSn!}
zu`Hf>GpMxAR;Y<z*nEF|n@OgkHM9+TlZQ4Oj;&ym($!8Tl(;}66JaiFgPp>)MvpwX
zVd<ZfveK9kc<H~GiUt1`6=d6IoErG3Hj_FqvClVt952F2>wNK-AAIm9&do-jjUC`!
z^_h}$_Wl=p6u*cC8C>keRHTj!YtptG|5P#_1+#i}-nT+an2BQ~JE6}q!wrlUd;Kb$
zF<H}e%ZW=td10a|c}3muum?QsifK@AJrgF3AHA<s7`L1}lU`=@FjvB`J6nZWxyMI_
z0i&5(q314|9Au&<X_)Ffrh$(Og9y&f<qTT}l4fv~J~PE)Cix)Z?@%9vntUK>p~-ls
zaKT!dMorDXrJ#tGI$!f7L8Sl_P8m4`*rQKCHG@vIvjJOy6N``ri(I9~F>3j(Cogjq
zEp^DpN(j$`jWQ@D4wSXS5e^y{1j-FJuZAg%u<t?OJZiXYp?XK2>VnHuANV5%8Tx`V
zkb?1tS>k+XcB19rLrlKPwoB#85q?|-%8EZGNw`bFOa=oW;wb-lzn8*sJ<3&eJnSyx
zSPSek?dgxmQ@3D(d&NL`6_4r$0~wf?ZZS@TK{AOCZulh~2cicv#}S;APqynih%AGA
zf$OdqC73hHgSNw66&#)6{!z>fLbjYZ$^7(7<?YEC$O04;zG*C!ox_n8oV%Dk%5z8z
z<ERFh9h}V%b|r^~zYIB`O|zY>(OY)tJkgHM?unG|vnq5;b829@KSgb^PpYYRzW{wj
zyxIdqr-4Zv`GH!^)|a~NxYM`Drr<bDHrumtjgO-taRF%XBpGlrF2<Fm6%3|{K~G{7
z&%DyvDhRgHt6?WoMG&O9I%gOrV~aw&f*f#j`?Hx>dM0@WX=q>MXXZIp&;co*5=y$u
z?C%$-R`P3P6anGc$!=Mf9hS@)rp(}CFcJ>*p$!Cs>u>%|k9OsB8YS^?qxV3?G1Zz5
zo1~~M*r-=a8V!Wmpn^MpBd4wWI2szH+D88>5Oa>&EW{Q+#3I#vFgU#+W<*M`O+4O}
zLggL&E^&5P*vO3uK#V2_uJq1hwOKHUGuXDbpTtcMIM`65VyJk<CG}}2_Glm|OqpG$
zY*!)eE4VEbJ#O`F+C|UX-*EvkuJ{+!|2PcvX``pOD&RNYe)p#)_<HFZf5i*L9J@vo
zYD^Ox_hR=%RZp6WT~!NC^49|TY!s>Z0W;H&uK01kjdQD-&sjTb(tlC&J90KI`uJGu
zFa_snP(Dq4GYn~p{v1B`1uyGK)82JeQ0pvDTIIK@H7vk-{w*|`kYOZFO<%#&lOdgh
zR_}(#ES?Bw!Y}$K(tiTfGy|!V|1a?T<>=BrDrTygf2p>b4Or>1yDfmK!yys)rG1N=
zd-`f_x;zGF(n<W%sx|#z2#TEJOW$e>00RCSSb3i(28b6sI*6^U)mUAgkJY7xSe$?4
z-5!ez^RcnIB%Gt@@8bx%9{Chswhw;ZN5uqRy~m3(g>NFV6fHdBfp{z*hc0&?X~x9S
zK~7>b5uK~ISv<!iHxt|NoQ2;KEl)HvSw(*gUA0+HddV;0vQW?bsvZPcj9Z8!{eY8~
z86j1?9BH+y;wXE12qyfQygJ}65QPhm+}qdV$*h6?KJl|0EgsRwBr*&16vl&l_u}n$
z-i{AH{wN;J&&9^(dMqz0zZMrX23n4XkLKg<g9mZ@&i%M||Dkt8AXs6tn>$M$%yTT+
ze9SGa=Y3P#yRovi5sS+!^1m7j8bfiP%;q`*h)qw>JO9M1GX6?|$gy|5afi>y=&*Td
zO1pnxFos8lop(&OT{u4UPKH78=Y2{GE7X|`XTp*0k7FLmhq_>k^fT#C*&Pv<G7vu7
z0s)l+I|nG^ItCyf=mD}i`$d@<pXdkw@8k1T?i{^%X==>7CD^sc<ly4+vUkyq4h?yt
zmr2JH@#7p*#zIc+iQ!l^#wz{&11gIg|FFWpL+PM=S-6SdY;<@?cHJsxg{;Ew(s&p=
zCSbiw2f={qaqbLA9zMqo%1_Zj*fq{BnekEI&&H`0q-u|e)q@k!=?#u$`_n)BzT-VL
zb20iA?*Y}7{{DW~b0%LoZf$gI(&MK2xkuh<$L<~kOynbV+;`7VjtbjHqcK~m{PIpP
zROO+^ZV0Y~%UA{>9y!ez2SGG`_4nhKc@@$`yVE#Mb(_gr>NXR@>>y-^1@#>+lcO2g
zZ9|?T7-R<*$F|X?*=^&UMM{$&>Vm?h>=}bn_sC!B1G|?zwq!D!W8hR@Jl<pMFFoTQ
z%8dF09%EJVfyb3jP#16FhhCDtgopaY<C94A&_*c}`{HFB@`~|l?wqNo)6l<gsz_C{
zpP>z9UxNVJ^;daM{{tO`tTGVc5`L;#zZv`1u`qc+x;fUgOLgh^a62|v7c`dLvXFs^
z@LjcC%8lJPjQhxU7Fx1Pk;TQtM}c~0=+NMR>o9%U9)fc3MmrGy#$9Qh7L;skvIAwy
zyZ$%=k#g(m>PrE(%6EKx!g)-8!M)n#Nw4DKHmE!}I@nd;bvu6i<M(4>ej)AHDZ8^7
zE6WS<X!d^Gzk4Sh+`k+1>eo0LZ+Ur1^>;;m(W-Zvax|px=6W<6AAb0Oo)2U8;j9JN
z91}}b92p&rXKy|0olBJO>Z-miY6H|QcEBOvU)@x=8lI0$&BXN^w_<dZJFj|NhTIds
zqdGe@F&+IQ({b845c|g+v9hrni%S2xXt48`9-Q4*zUo8mkRuT}LXzWn2h_&7`>s#r
zGblZ6dU8VTx|e?EKylr6-9?bZf`;M#?wA}KR{I}}@e%c*L;W$-*Dv~c-xKW;At<{j
zd-N^vy^ZeYs2|KF99vktsdlrgxNob!c(l41^J}}Yx~+D%f8;#ks6o=mjyD!VaDSG^
z4&<ErJ?3AQSJssOi$3m;c{Tcdc0IDombn(+ujM(B4yup(r$eWSw#_cKrB!`9Hg~+M
zYkYb-CMKuUFZQ}0U}skE;7CkN&&1^9gm*8w?YVBZ`Fji<@8ncGyE?hgy-#(YI}_XV
ztvZP1^*uei7Cv%>?%dLv`a|wO+fcoxpVGLhvnSf=qxE$4DU9xZH6Z%;cGu&D#U=MQ
zz^J|plozcG0j4?uWd+9zBV;Du=}U5shW@3aBab-t_lNeW@^N2I|0tC{ioVqg@)07d
zp3}B17#DeP_*xR);n9WI-b0AE<z3e7u;qA2_zl`W<v}+?z3J=gcK?Ks5dOt;aB}2G
zV~*!#-V?#L1=k9jI)*TeHodK~UtM1G+$u%w`B9+se`?CvpJ6Um$pfXKIR{?sg@l^=
zq*u+Y_Nlhoop)=hz|-vfvz;E)r{GyF`lid1@mk@_*nJTV%#TxO!k3|c9=^!^C;0`w
z49?R4zhA}oUnHzAf~V27;Om#arvZJ9XyZjkCe?ra>f}YkPESE$uv7!cX)h3@+Aj4C
zdz<xFCyBEs(Cr44$}iOo#rni==xv4`F+|0zr!b4jQV@2bjlFT8jdlJSld!yto^U<+
z>wktJB67!XVn76xL_$;C3QkQ4PWFLY(;@+VRTV!n61*gdT{<_+rjt0J8)~`?>241^
zDAj17p!8Q9P1<hI!;VKUqmXaaBTwd|U*u&0=P_|HF#xv`w_*p5y-Ek3Y1<7o{WEao
za!XdCv1$&X1H!6|10Mvb`Z#+9vhXWC>?{z|L+8!PNtnL_7+nTevO@@x%%_-2eE`ik
z>N%2`RTkd)RTyFhpvz)z<}_et)eGU7fz`*<=h&AERO{-sWGEFCBBUpc<%gjF&Mn4&
zJAtC$M;$0o%~)BWB2A`4ujHZemVT@$kAe1~NBPH&DugfiqwX;vrG3doIQ3WpxstEU
zx!I5t4t%gZ7?zv<sw>GnPzD?&!@2td!y_qp(#QBn<>(JNM5q%QH1Y9A-~{8tfSLg$
z%v1-H0b<7465gi)9|A9Eo3S9yM1#0#5CVhmIwksq1q07O#1k9}=SWODgO!~8R{N4?
z59}D6DIdH_P%48Yj=M1vM7}AWC|q96swoB$$446Ii9W}^FyJ#2<h-Iy@MJc_B*0Wv
z=7@S{d*wOIBzX&%K~i247FkQ?Qa|zM@_<2u;SlKt;{mSo9f~|mz_g3O9MZsiOz}tX
zRU%?%^huw%qDiNQCLz%peFSMwH+(956cCC%3jM4F3z+q)sUWYD05+%?i>~5f0F7Tw
zzx^bjGa3CFbjy=8tg&mcY<L$r$EzMUNVJni69+%`jbrMHJ-SSj0b)ij6dL$AwZ;s$
z)E1@1E&Ui<e(;^)>A_Q>!<(|pCX5ZfGAv%?NABoS`J<XNG|)>;FR5T{cMMXbPac7^
z)wsyP_UIK3_O_FrIHAhjB<!~pOey_iN+dnaFz{RLDjm1M5@-7ZMUZ;op~}*<%NwMM
zhed-)m+HEXRyu<d%0pj+Te_!z2;l|_X-qc<Nb|)mZ5n_-Y@PZNE|Z+r>6PxNm3F2g
zzcc7f*f{oMRX_YDiv1Bberp~yUP7`RW<8l+!m>A;c(8DJnJC6W6C-p>oc(rP0oRxU
z$KGxYE5RAs4YlA_L?a|^gEWppLHKD3zFz*uuZmAD20`n$o(V0--T5i|;$$;;HWz(E
zRZA>pz%|?W<s*x|-BDy;f7sbcFU5y)NHj_kja0|8=nIW(hv?GYZUxU_tKUbT;isR3
z#Pv7Pv@#C%t8!?BnTp-`orN@~|Ae#p7TS;bEkIg%odMGRIDWOk&FzlJQlag?(x_@;
z+msp7^pmEdrgOEozWQzGtGVe1z<eg2az3BVW|&1SejM%Mo>3`oY1=TTU7C~5(;(ez
zSdW{5rd^J0cw}451qF-7)0!Iec9t%a{diiR3J#<B;MLek5a&Wu{guZH{QZ&xyr1~A
z-R;=kSc;9+h1gtMjuY<r861e|amE_`o{WV?PQc?2S_m|H%rA26HvFp?BTv{{(8Yw%
z?oRCO?3x#>JfVeVj5qYl<C@^fk{&y-Quw6@{?cj6yo5L7YfKt*H8<`&B|W?uC(EBb
zm`<lB)tHE8)p=(plh~-N=4UZjcMJ~n#Kh=eOpFc2&_Gx8ceMvoc0D3YAF#N%81H@X
zzVE47T3m>EjeFSrw!F3xtJ{aMxV{q`Qg;s{jxN9}_l7bgwhvEYVQn{Nmp5X5bvriM
zZFP7Y+)J{$yqrQrgo7L#xWmq*E%D|1S?|?3_~?11**q!fQ^n?}HFkNxr+4-ATEH?s
zHWCx#qn`L}LulF4rF7&XH+WadBn1#0566o1ysN0QGX-3XGZ4fu4n)|%5pM`3xK|9}
zkr$!K&EpICM+L}p=$p64ALU|t9q!);d)njKYZq02`Yeb!JW<_H1dfzu@}7FbV!9eG
zLMax>a6H-g#8ixoBIFy5@$oUAa$<bt3-yY6(9_!^bM6#lR}gi<`BEmh2M2m%Y-Aut
zRkj1t_d35+m!)=g^te&<D$mBpIhGM&g|k>=UM827E~)!RhjHiL9Uo^1Z+hv{OiWBn
z#N@=Zj~u&s<C%Ewg_q;n)$4Kf(&ga3n)ly-*LPoY?A7q_km7pkawd<d<K$V6DH!Ld
zzNkFvLPF}Okle;d6ZuVjI3kVm&n_{(JJk6qSds_6jE*mYT}p=WAnlBuK`d%yJnu44
z;t)VG*_}eK9lLRY1&t8j5chaTdFeZA6fXoSv=w%FAOK_Tp^G-fI8$-qIDLMM2@y_M
z&_!FujL=Bw!VmUt*TSKGVuvt}ah}Jzv|IfZ0U|<dgqveb_iu_JZ6XC+IhJL7nd8i~
zvtV3(jRoVS{dm$`bxejdD}Qn>ME3N7Qpq!aSEVO>w-cv31*jZ7*wL>1+ggdcpS>3k
zZ-1clM+}aPdPmglPu`DpsY5FF@iBy{<XyMhXx{)r^%V3Xpd^10yzXwT#p=={sq4XA
zjEuX7M6a`FAdWAnj&`W7_N#u5PQ=LM#h9Go9>7a6KE+iR6BbAk7Dr_=E+#ML)nCmo
zELxa8uYN|Y%4y+#wH2Ay*H*od0paJd#;sj_{V_Z;5o6=rIe0a$-MAUouRo*ueL1El
zr@c#%V>_vnJ$(Z)GB)n}1vyrgc6wRuV0w}}1_yk7#=64b&PCVVwvIS%Q=RS}jDCeN
zIzDBAb-VI;TR1G9UR+(5^F++tcs>S|w)XCR3sX){RX=)q)OH6f#JV&y7MG{T;^Op_
z;yUW^rY45s`n3ALL4ALYcB4(-t>K=wxQgA_fchC0dn3$`7#-=4DfI^($NRCffnaab
z_r0B}jqLB{9Z<867URx?N3pW9ZlMqhD>~Zr#8I5Eb5L<+N3Vrxy9cqddlbvN$5PpG
zcH%qSXxrQ5uW<8p8|KouUg6mFf&P}fp`z37nEOCz!jX6@>zfwJaR1xNDo2beO&Vl!
z)a{h|ouLuc1@&S5>fif%`rJa8lS9ab0FHYVyZeS!o|C?wVr%~>whssk%#-M6hnDK1
zceAM<RA%exic>x8(ArkNx4EM<t?k6z+)8|YZ$3V`^C&*OI~R9m7kz9df-ml^+}Pez
z9ovb8m5o?fUW;Xw<Lc&itgml7{~kV^kK1=2#3!HKiQ9J`Sop<)74F7_r{LZ+CR5n)
z%0dE;9qjAp$YtW!Cwq>d%<mBW2K^ucN4HzIQ`#&1h<d-w(9)I>aI4m<Poj_U9JTrl
z=9!q|q3tuDxvqRhaK$cY_(0ku-!g>c>{RYodmmK0XR?jCQWg*O^r*j9IjV;?KEhuX
zM<AGG_ZfFja&#}k?8E(Chg%o%@YJ2IPkKlJY8_9buWIr@?gi7_u$PX%-Ktt}9@ERD
zxU@pG2O2mHyGp-2*fiWy-@t6KZ_Eb>GVhzvOvmH!80~ZB#{it7Jr7U9e<F<ne_pp%
z__BCU0{ng@->NKsk?@`dPortU*Y|$;uX;caP6N#m&z}cV3~)WN&5Gs^w(V{3s|J!*
z)m-rlT6J}+C}#c+Pq;Snk+5YxRd2QUL9&h?G7KyGLYMEfmEH)%q1cKB(|>6!J{-1m
zjYzfN7kv{aZ5<!$l9o-xuVFxo579JPEYTOW1+V5?TF4CrXX1)#TjC(SaT}HmzZtg-
zk9V4ji&fGA%h0sXw_|8$<CT21WyN!x6I?C=hfv~|o)g`9>v+NB5@vA#-|@4-b})dP
zRnPWiK7KwLkFY@zjhqZY@0w=+bo9m2KX>fv^nILw1t6@XNq5rA$Akf@S87THe3S)d
z`GwKsXv3xDl~`O@@VVoTr#e)QVdMklqxxj#qvRt+M7lEi2oLz!J6^=HC{6<Uf6F16
z^DIAJ{*b8*QCX!^i{9l4r&!bH1WL6T3u*zDrJUJ_r4K7^7Sw1!LtbXy<v=rifX~1o
zKj?J=j2%=Sm=<>iMbOGYb9mJdE7v&kfwPK<8{tZ0MgXRO!3P(SbN3Y!9sXbwCB=)a
z^1Z5athCBY%oRW3!+5j8jFd32@a{o{i5WpyUH!{ogjJ6`zK_B2Bzzb}qCpzdj&v|k
zVr3WUv)^ORm1n{ucY-m%VGzjfF(1__KMb-eqvN_JJ^4WS$i{V_)j*_|yurN@4*@Fz
zonEI$Kk2Er6hfAX9;x_Ee>f3;%AoWW0?r5@)L;d{SC85R^wS5Fvd5_l3DRC+xL)uo
zV6#Ip4)v-!n*wG%E^i|jJ;>-~h1SMmS2gqld3PZ3$bjk(R}|B?*+$b&V-we5B#IEy
z9dfoyvh`(16hcosP!ckvv`>BdFBpY3c7&Nw@C5j?z0ne}pclK;RJZ0sZ1CHHFEFMI
zO{_wr6hh+K#=43`|DI`A8nLTbjZ6~A-o}zis;U@x3FsG#wuFNlsN`D3!^Qb*eNly!
zcA3LX7Zm$YoRSvSndWrI4!`)YIU{d8?5$RdGv<ny`Ixk!idkCdU?+v|TzXAEHT?9V
zx2%yJC(nd!Q4+lc{KYJ~bWAGN$VMiUNITr<6)6x~xzD70vW%dSsCx0Ij0;^`6aWXh
zY10TX!vkLkMQ1ncfwa{R0tD!cu2|{AnD$itryG#&HN_RujlMhMU(BNIzNVOqZBrHT
zH{baw?)ZA+ulQsH@)~HU{95c9@jnNNQgKtYuJ|noPM;~roRkvZx{7-avYvro{1`8v
z>ZdiF8bj2KMR9j|+V~i#HT#`#7AfmPE48l(HExE-DV)#KG-r5Ef@%xiS)Z*y-9l~B
zu(2hq8g4WEC&W#5xVCsRJUPxy$TCa7jr>P_9IvR3Z^K}7g$5<u^`sg1;!^CZTl!3T
zren0CH=q_QaPXU^d|OHVlf!OcHvG4y7iiii=Hr09+^t1T|L}vUDN6C5w9@*5co=$0
z?%iO8|N6?Z_{8n#>*?~O8!O`3iGzTP9vB{qT}_^VQaSM5@`N&cs3+(TP1C%q+=&T#
zhe?R=X!=nT93PDf5Moq;7k=sq1jG;CFDF!zZ$pUDKRD>|P-hprfmoo$!?8{phsdt8
zTfA&<pYVo!4;4a)5ve1?gWl;gGdU776GJ`@aHt1iS59yqDBW8t909c+%d2a#$nJ%`
zL-EG_qTMZClT!-B19uO`!D)x@E;+fNu)8%T6V9p9vA)SMPCGsdiMw`~>_+H8I@o#P
zoomV~CPmqmg7DD-K;;p60RIoZC#sZrq}_M@is$B@H11sTB%tz%<FVMaFgDy9+*by4
z_pqCWom^Q5Sm4F3lssOZJ5t17>w%|*+a<5zp%I)gHh^!ou%l1u?C7)`<%fU+cgA%t
zbIM+PHE!rx^yS4~iZ2uK(~}c1H8o*=oe9&`t$pQ}%2a&((7=ENR7@`C5ru@UI^rnj
z1SLW^j&MW3+S`}+SdoW>eSzIzMM!j{yd*yy4=kuJN*i^SMWUlB^P!;um;Io^9Z<T4
z`VnY!#mHc<cZ#rMhSS0*zi3mul=tKlWidDRC}tl#u)oVQGp5nk*JD8%6SwT98W|dv
zU0>Y0e<wcw{4?+N;M51)mX?-dRbyt#58<T+e$;W&tM*BqX3W*4aSsAr(!iL)W1NHB
zLC3LMb;o4w`w)D!Yb@q^;&!gmq&d00oJN=IbGIk%Dm%9`!d5=1+^BN^<8m)vRXh+L
z!&z|Bh6TimFDJMl1SQP~z+8Vk-bZ*!eJ%0VV}UXg>)qr_?({-f1`Wy&yBwz}f7M>O
zCyAXcd6!@2CGD3o?`EN+(udHNitDke^2UoL6@Tv>RT@~hNq%8Z-jEa2g&ZH_hkmHp
z7rBaw9(2hM=O20J{Oa`BlT&pD1_y%smIep%Xt%xXwfOAEZ^mcuzZpC0%Q2|_gt*?l
z`+3~{^uyrJ^ojAw7$2XKJv+fVf_t#Z3-XfQ%z{UD3VDYelkSA8Jn2+=dk55)bjJRX
z(t66c8=)YA3EDFCo^fb*^r+wS(Uy9;dipdrAB=gmuMa=F7oXpqjRy}GV`F<KE?&AE
zKlAOMi(5CJi;=MjKisu<<@$}ddi9#0Y1QY+sp*)J{q)qNcTlNF-7j&~!O+OKckv92
zO~>Tr8!<k8BSt1K#mto(apTr=@$B<2ir&q*p!lzBs(l};T!lX}p|(GBQSE=qdB;xU
z-A(nUbF*>p?rjU_4wRn$(Ww}oxE#F$Ls_p7Op6x%Hv*oKk>Qw}9QTe~>MeZ-dE2LY
zI6Tnn;|Uq_b4(unH2ng*K-op6?iWeFX|ns3er{oTHE!Q~pl8;Cu19mrv9P4Rer;2I
z{kjE*tE(%quQICRk$rUx3v#=Ce?IO%%0&<C3_YzooupEaSlEHkYj<-?RqZ(IA@%b_
z9|^^sw%~aOk)>ZjNRQgCGDi5%0^aNwPJ_Gg*ipQ=v}B=5m%cB3%2(!KC?f=j+>=Io
zV!?sG%c|msN;E=l?^;v5=*ReS(%1I)tN+wDg6|JF^mho$EC5p+I3ks!IUgbby#Gk)
zeH8Z}F2tjSrC3tFts$6O-|+FMsF;_Q*Srhx{_Mke@IcRf>F+;`*||9%lZ-%-JHosJ
zT<JjgMV_9hpC!JGOIe`8j^6Au)UOnwIsKB;r#zys@&aeKWq1x{qW-nb^X)+I{y=8>
zN9g$*rgrEs@j{V0N5b)1MPSG7wH<b-bD!Cc`d!6Q^tve<l^1;_yE41<O<}hjeK5jC
zj_#yAa~GofrbC6Tc5`~fPA|S+8IB_3$GlE4H=+R!YH_bPt*ZMdn+s0EUON7Et7^e{
zOy9(-@FaWV7k~JdCZFO~A@Pjyn0<v-J#gOK3g`WO5goudMSmWihW|A9^SZUdm&JP$
z;P<Qg{)>e7G<X`##!-hx00+fdUaHD6c#6lC6dtTStqpymH2lDe_?JuxCbttio~JFd
zb%lWaq|Ur-C2U2V01h&HA_sj@a4vzMI|Gnl0u499F@6T)K&(1K<;j4PsFZZZF7>SP
zHZx&&EO8?cjG~7TYnvfg3E?0`FQ?16#jSyZn!1E^bIj2e6l`GaO#?UHW^_t6;8r>(
z-Wd;Y(EDXI%Q5NXX9PEV5x<7VRDo(|sA|UDwCpB&P6lQ=w#0NCD_w_~b}-GXiWnIi
zF=GakmiXYp_z6o|I(B<?oUp*0ABx<~MF$Vll{;}<PK7DtqvE3~=)e)0$j*$n8O&p!
z<G!}O7Avc(vA()t|1i~FT_MK812)AI#)OXy?w$aZfjIFYFq7&@A)-iHiCxC?-+EAS
zkgp6ZTwc7eCjDY8oDAh5g9qoS9W!ZU1LzD2tG|*f)*FNi6T^`m3_<`i@}lQ+=oKz`
z41?%gOwPgdsYAqzkG%bsaA3YY2-e5B^_Uuxf=96pjsYwD7<{lo57AV%N2w}b1YZpD
z1{vI`4l)R5Fx^F*U=Ao()fXNFgB^-5`M_gAD}pL92*;{R7BK1OdIN2@2gOEpgn=OK
z$aP6pFnrwSMNE{du7}5>1!D*^!fr2^De4u+fOaaarY(BxJY-;>S#&DD2WBRzmtI}W
zL<`|6k8*Ity$!Nw<t*hxdlvBaj=VBR!Y1o4ZB7Few-apjlNlUV$&E2!U?PTcQ24kf
z$*k{aZHJ!Vij9(zdJ>Ia@&~+99u*Sb!zTai0YyM8OZ2H8Ne9JsQnYf*n12AVNVU?%
zC3UG!a1>;KwlS=c&T(`M?4B=DmT?EnoECd%TlnUGy7ML<4AP`e@&m@U(k86L1(?XG
zaZg+F19SF`yee)9uEbgH_NEtMH8PAi*0-Fb<To+F7{tM8!4W(8H%W-tN-w#V^1>}$
zK(eMqm0#hKhK69u09Eki=3p8Q#zjb!`m<^yPD)?nJ0`kLLlQD%x>An@J#+#33=5%%
zk12v!JPMLl(k7Sj>-0&boRW?;Y02ODk?szQmz}{adK^(?qzs*Y=%lsd#bYDZnE8oD
z@u??coB)z+)Zm@N4pTbJMV|l_5Gw_lq$hnmX=9((C;f^E*tg%VM%CWsT700F!Rgmy
z_m;HRcT&fMTZKv|?em=~;5h1)X#!8Km}RVf{K)R7ebgc8J`Yc_18iTSnkeL3s!xM#
zJ7?iB`lff*&PMz?-^fkE0x{j62H;gNe$Cz|Kr8NfaGdS@MOa?|#9hkywEtvihDO@Y
zLn}?YKP7Tc@cI-;0yQk^W-FYJ!%sFBO@Rj8hQs;vsRc-}UQVbDKYk}P?OPw9>A%{v
zIHgkjrhg#o)mgiWSM;eEDjqR!!S(}wD~$@(PwUfwRy$PFDdrZyKC>&^Ky_xqZtq|>
z)>fBdb#2LafpBaVJBmg|$6{b$5HFsD_l_kd&^)0f-VELn{ue$@0AJ=wYly2uD#1A}
zb&}l+{-k0@Yk&vzT^{g$NA;eM;!%$$r^SEv4|3%DaCG;u<BSRP0ojfCt{v|d68|+i
z&Mt$on3x!gnVG4$d}%T+U!069QzJ1w)*n6X8V_x)#@hVDc=X_IEH7{b*u!}E;Grk&
znCx8L6wlbsF;0UDcOWiwD7@Ygg*h0f?VaL{bD|!VJHWWm-}k?WZ%w|Ii~8F=`Agb%
zcQ{^tS3Ge}0(w#n0iW|l_-*1rMVCoC@90t4F!n+KFr>WaKHi0eMemkaT3U+DwKb*V
zRQx6j);h#9=R_RB2oqJFvN(++*8E+N4Z@0U7Oe>fKAf)1yM-tteIF?^CTSUGvg-vh
z@VxN)FsQ@<@OX|=^EWolj6V?GAjss-v1#S~z(8NDZ0^PE+-l6P?)VNQ?jjo<?h`)5
z^9VqM8VJl7vs3|=h>FjHdLF|_&$qAi?Q<cU+?X8oomEOVV^xF!r>alpbVY~7dHuaz
zJ_3y+s=B+{qgP{vp@H5Q8SeKEgw76)EtJPhWbQMeEt=##lk4MSWAfkU-F|bA9;jS8
zVtjPSZj|TF&c1hSaI9T-uj+}?zP`$FaqDsU^3}L}<+^b8Vt#Sncj4`;%%R<>F&@G?
zCYzZw&BchMk+Bq$s3}0EeUUCEs(qA~RD`DB(1yq>9`Dpsd6Hi5&Qf`ID_!2*p}ItU
z;O-uTb3JN52tuiEKFUz}!MGDMdF>rzidzcc$Y&LW>KOHldPEv6^ddaQf_j)>hW6lK
zpLcNN{eB3(h!e-qrI4#%bpat5ei+AM$I+G^U-qXkyI<oc#+tp-BM2jXfRgOG3ZLcq
z5309J)M3WR)ITQ7E$E|O;w%r0?=TzWiSZu-C4_6tRrqLs)l~~vRn`L*mh17cmMjKU
zTTydbTYeay|M-Wox$ppKp@lVl>Vp^ylNShQ**(?S)$2QpIQDdnV@x-<^5{Q}S2-SY
zb89Eo*0-eYtFKkvJ&f4iJBpQ!9Umdbjvpr0y-Suy<<EU!7M3d$ymL~+PL8-_rw{i&
z-F>tYkCwJ#eOvWI3_SO0UB7-KE?vChezJF9B>IOZVt8!Ig1PbW@#t54nHS)wOLjkP
zt}n;@+-%Iw-c|p<rm=8O4311j&(LIa4NgYK0MB^z4vojq=!6A-J$*xQ+|e6zYpUbx
z`W9SJ9uAJG-yM&M@nIir*LJcON4rYb>OyR7t~>rc>Z?bmuSWO4sM^AS+Pz9>e^X^g
zAD{Q$)j9@nC)DiRl7&;Gn?)Jja;I<P2-)SWgV;K1x8RMtoEF#h4clf>!KvFXecHOr
z50}>B(ej4!dRuv~zT{YKNAYJN#faKEN6B&|;4oK;NJVxwx3UwT&#uPZxwTkRI2_N4
z@Nsx_Bzje65O{5>&2zln_U4AF81*oPH3-yQY2B46O&!b=ox09&#4e}9P}dN!Fi+xd
zvG5SUl27bPT3+3VHGPLH&{ADSh-blbY11qOq3#Wd*5vr8+dDgvXv5&rry%$mU`JJV
z=I6HZVoi0F;}Y4uMgL11BHuYqkwqIjs*@ajO5d1iNg>`53pZH6p}NeR;KtUb%4b90
zJ)ZU0-pTu}_^y+1*IC7dvf9}@j4eI;>_$_+xFy_ejsx9=9y?U^T~xov0%_(>=x^KA
zk6~xQpy-fC`OTqCyAyDK=*2g+lOP}QuWuNOwdv#N8?bXfq5P+xV2+7zHV;P@(x3Pw
zEzxJr2;o;AQJHsj5#GMqAx9>1bnvdyu`9j#6n2{(?TZ3)fWqV@e23XEV^t{GzUtkp
zwV2PkrM}Q=b?4u}Yr%O;-^8oXYJME2(k~A-Qd{xR`6+h5)AYc3b1R(p_eFF7;}rdQ
zcpCoG;Lq#U3SSoQNr2z4>iaJe-qYY|G;_xncYKvei-uqcJqoHl)9}*G6Gi9?(C8A6
zX(MaZP|}W0r~Zt_CtfOM{_@1X;1_eDZD0Q0S8DmPN^}7X(sP#(2D=%GuHZZBP6ujn
zbySKU(x1)(5KJ_~ao%7H0ujPsheBA|hMLxZY13h$w_R~ClyD8LdV#!TTxBn%F{viN
zK7Z9r4qMmqq`me2X9raC%*hhXjFVjwjEgC4nCNT`JKc#~8jS0IQJn|Hz2*m)sQ*9e
z{)9`C+{p6;5$@sleaD@dBr|zd70IsInclbW?d$65p0C9ZH`CLzRb9N3$;^z%IQ>4%
z-OT(B_x=C=0;sB)n@2>FeAzW(rV2nH5C~2MAPT4>%?R<p$v};_kmlJ&uYuz*g`G((
zDs~<Sboe;*XAoOW9MzF~LaB$VC#pPL^q}CfA`In(I&_ZXh%>a`dh*6#ivc<hxpt@H
z^PBhekVkRmcqInA&wR{?29$ogkim*7Fv;`4S#gsHWbxKyz=7q()G4F>-9xU<R*aCp
z$gOlJ7fDDB&i7&ugZ50_qTD?JbvZTaqy9W_{lRSy<aq#VV8ff2D8e){Eal2W|4IWD
z2FCE=A?^V@Z^|$U;zu01DAUMuWQs{NI8U?`&j@djX<*=qwgySWi4uu3_lGsO<oFGC
zd$HgVZ_h9}invK6@xzlr5#{?<6S<R94bpg{O?Aq3NQ3rE4ceH@kv9xNQI<IyRd1LG
z#$<&_uhNNK^|DOv$U+8}3~WljL>=aQbL#XcZ(T}GPkwZf2KrZLJvSwv0nX{kx#HDC
zOoK}DiNPfUZPGzy<w!?lA>Ukg<z-pL4?YaEno2R!O9>*mTb=N@rZj<DT3FKHRC%cI
z1_z07wE<IIatDfgO@j^b;fg!~4rM>^KxUC40=j<zA5P;5M;gLa^9G_O>Jl?2(O5eT
z&USl;#dQ~J<+7zmPQ0Bjfe5BKP&^m0$-gjy(lq{uw{78HaCXU3?7b9{B?2~Z`7S(B
zVTUA>RhA-aga}%P!EnN4vr*+%(2B713C;9#75@mBdDtDmhq7~ZszL6GO|x)azp@`_
z(}gSrA}=GfZQRt>qudEETLctq>ODldFmf78QVV^!jWER6BnOjx_AX(J%)zrGqYNgg
zBM4J}%SqkS&;4U~3h=jQ*mL260m7xb{^dDtKcfo2TwdUc@G=UILLW-BIzQ<(oqqpt
z;ik<8ZNw1aBP0>XVH)Ve9sO1gqlA+0QLGS)KlscKfM3Xv<+sUQJjo+@I<5@lU6n<I
z%=S>wVn&{Pr#rs#H@*Om0Wg~~XB^P9*HCZ@mD5r2kx6<$+H63HqvDzK0;uDG3VSvS
z`aRH$U$-9K{ZQl>2o4(l@1W{|x{r)#qdvVSo*NK#ryo34z|<AbH=`?P#@+8}p8-sJ
z!(=+g--rKJZq?)L1YfS=4hWjs?`Hn=g!4V1;EP}J^~j{c^@eXUFg^x)6*R*E5f%&f
zzyYSqSCmC~Hb~wH!M7ooIrN6E{dG07MNW@{@WWexV&?Ful{VMkq}Oj=`7LjDeDJLh
zzAME?^cmLE4rmXI-!ytNe=k03eC6-sjIoPFV=DDsmmaT1ugdt=BPa229ajGrJ2_Zj
z(|Gb)<3E{<0RhHgjOnPK(F{@8gv^Y6=jP|r635U<zO}X0w03XRZ^mE0J=0kEG%d|d
zr;i`6r3cINX>{l^?QU<RFFyZc`qQ8O>?4_8ZEUCAgSK~Som^a{uEwheM=VxfB)<8?
z!nBnl8Z(ZwFi(7n>)z^W<cG$senC(d`Oc0hzeJ{a5&*0~Q@_U^ot_-0qr+C}Do=R<
z?TY!A(in5F{A|f-1gzBXm^i;csLYOCH<O;59jQEg9UZ0PE;|gaJyvJ2l*LN&oZUBk
zyNa<kN+a(&gD>+8nW3Z#92ezcg_BJDn0s&xAL;O%L>J>@#^;cEUMxBwy2D(5lm>Qw
zA-{K3Ehx`5hv9fIzFWj1F8PE4W_p5MLZN&+=Os%PV383;jO(Q0qCQ-nyLuuYi;jnF
zPAs{WY@^P%bJ3isoY+xDIdNnm*~&tw7g5EFc?-vIvVe@zmboh50GnWObvPDiS(xSG
z3B#O%T^Vz;(`j*GA+2w0q_4mF%H_!JB^Jkd!472<J9yZUvb40~g?7HXG%x?}{oqq8
zQcz~H!+>2J%$L}K1`qF8P`jYsGY(+7;5x2!#^NsdNuNO3*a||XS`Yao`6zc3SCkFk
z1LFI97n*~e9k&(lwHJ?vG)v_e!nx_mG&eKlU0JyO{VtV*=S-q<KB*2x9-T{OzW;Tq
zi|<Iq;xBENBic~dG58?gup>7|lTmN{5~1gEs!P<JyP-^_58(w~j;b7!3|5Zu(xuuk
zb!~!U`S}LcOw2b?j>UYAG?VstIn>`5Q+k=VQU`oYs@$=2jeJtu(dGV*u;7{Bigs*N
z?Fa>%=C#ayQFbxMWr3U>UL3FI-F%~2(aU)5>M|X-chk<Br>UdzAJrW1(MLZ>KltHK
z)02-rNvroBrn!aHG(I_JrNPe5z7^B0R?FXm;&dDq_L&b;Uyi%HT)+`=E$=!xY<HB`
z>MK}xaX<JrovY51?=hFAZO|Xm)@Tz~J__$Z;n&wU)Alak)#7_tC?~F69_{X_`tXx<
ze6I4m9FknT)PRCm^-T>91vv^^zAc4P0VM>-_@M||Tz#14R~{;z^C^v}&$$yv)A4P2
ze-lsnLLSrB_S+{`p0FEjW_DI>WhTwfOr}{BEh=B{0#uogPR^!@>BTg+axaa|EUF;+
z)>A0BF3yhCCvPiFr|IOJm#j}DlhQ1CX*X>Yhn-N6P$DQJsE_Q_;%L58zkeEYZLdST
zy+}jqTPN<`NsH7cc81}nyrbQ)8+37TAw7C<FFkquAU$|^U(b&vwaYmxUA?PLjqgx$
z>>qaBC-6e>i0UXuukyU1ZnX~f)Aq)C+TY#wF?Xepq(AntdotZ-c+}C!#b05Ey(5{U
zTp!%O?_GuE1#_OG<R$Hzom)HuDOc(NeW7=bsgm=OI&F^~Q{`B|6{X?vgH`pbchv`U
zmA~=*e&!%D+}}~}QAT&HkUvP9+Xw!Z8qbblzVkJ%_9<TMqNX3Fy+t3PypufGy)#Sl
zkvNj&_No4Z^5ocKuOW@{B7n+71n!4L%aaZz7|$ryWc+<WF4etwF&pI`O0UQ>>V~59
z^Mba+(T?OJS17Rb?1F%N4MA+ueqv`9>zpW}SX)y2r+g$6eM+oB(Uv$KmX|Ls6(1+|
zOev34U(}zol*#uhCsgk_J{2X`iQ3VL%EP-Z)eo{ZN`HTO&Q7n~k1L(XLY|bq%Y9LW
z&t}Gj+3T+Dg-7nsjFYK<teXXdUxW)XXnJ`zVG!P78*%`*(`6Vy3pDLOzq<)Hc)x=W
zV47_2hui3HgTK_T2_H%~2nhRaL;p4L-3GVu%!;p=&_hE*&aaDu%8)A}3mGOLQF+8d
z0=>Wy_;72`C!FInjZFC254SoD2AI&9SKLiTIGg;rL1!n;EhsBLdxt&W0c_xHlNn(^
zsPS4<Q`mHxY(yH3!rwGCt?k8D`9-=NCcI^*A>*LCRE%^)_P`Ge^btOX*01mZo9S$o
z<e4^m39=ssNXkQr0tFFe0k;ys{K-=xO<MeOZaQF)3S_C=ksr_pKJ$Qy(iqaqs(75g
znI=c>I1SL^WgKqdjnkw?4;u!93z|gm7J~@|9+YKQn1Q#Uj@ek*c^IJVI@LhOABWb1
zkcTLW1HY}QPM-(qrA%)Z_TWnw52<rLaNSKUj{o4C=))ti)`OG@gB}zX(F0K)QZadp
zG~``8l7sX_8N@@Yh-o`d!~Z9Mi-&aOTIiycO&7^=`b5M-SI9_X6luOuT9M`;HBxmg
zcrtLO%6NM`m&(G5AqEf7k(aP_JHaAT`OU+RBTsmkQybt#830W7n9MLh2)XqrCN7m1
z11cuoUR2-#s<^R_pzzX!!sV&K4rS~&XN2RZ4&tFsGMMt9Lh*2UqRwEz^@(F#G*IC!
zN==-2YwfKDf`p@56P6^vi=AeeKFUBDKzhmmg){=*?UvKT2iG~X+(#?whf)|iv9oD;
zFm;hu%7b&fPmeXIhmZI=4`j-yE(F5#t~td|c);JrP?@>hNFyLS`RB4mZg_xK`lmH$
zUR+p8lj9Rk2NDIR$bZM9D--}Y<s!VU=<7oc=&itVM@*Oqn|<ULbPlJR+2wEs@D5*L
zM2pR#{fj_>ac>8TpLvC}zTxG__{(9L%|R}Q$acV-g4TR+$jEM&i`<B28vEw3rU@kU
zrLU;uuK3Bq-}XR-pJ*J3>=-i51CTv<XScv<$#=>$Ao7*ATGD06-ZhIrI6W%o=*GY)
zjOtlDHMoSMtT5enp&|}p@FIlM2`<+!K=mr}ILoJiIctn@IJ=5H3fim!Z7eB?`Vmw{
zbI<WdfS@!O2LgZmKtLw>BtP`=Y%`%rBQ!V+(n?rerlB1|6ReAgB!qQ71wtN$FsCUy
zOF%~py-guYxJ(e_*{--PCn1zIVb9^fL2Z<KWi0tF)A*Rc(5E_tWs09=CqJOG@PJI>
zG)~scqu7a`myRIE%ihfdI^O_u&$is>gDIo}2g{egpC^mcnEh&{mu>>e;k(@N^`l?>
zr%LfvplXMMKwkJ4nlNsscUH(7vkEhaeh|ty-Aa6YG_?%99zFc=bL7R@3{$Q?ke_+j
zJN%<x`}R;m6`2O|xq<4HDZ=#Pfc|v%0)EuvaJ?Ujd|}7vm$kT?@nSFdFo?5`w?9mn
zK?7h^{A3h-Gf#qFrq9<+X!tkifgX1;=->NsRhVLe%Xs>Pj4M#{9drSMa5a?hg?~SN
z8K1jzpwICmCx_9GIt3-;8+MtrJMGkJ9i;YA%WvlM{i5j^j$WEdV-qZrv1@{tvqroK
z=P{$kR-9bGA~L&9P<XMZ7s@Qf%bb8{1zx<dXt_2cR?bJ<jJX&ad)&sDO&2`4@Mmnw
zVmXS4SudJTq}f?^mJFx!PAi=qAEw!<(X>1_lkU7dOW!>GI{nMP{7d@s>GO2ZJx%9n
zB%R+GO=njc6Az81w?iB^cAbtduG0Ax3I_?Naq;NHc$!;WNUQhm>00&oqgWW@+b=uY
z8|h$o+eajxpNQw#iR=gIV1LKBj`Hrf&ByQ0tcZy@Umh_~*Leym4hn-Ybs1aB)BeoQ
zC{8axDlf<rFL)|1Sd`u8&HZ-UV{gXx>|pV?iI}7L;maI=<Ld}Rev__fW8^DiUN4v{
zP8M4|uMpnH+)c4#g2hu7srkn5UCYVb#HT0d@?xSggI)M!7mL4?JEtyS&Zt~aG@UD7
zIsUPAbnNsk&Wxmm*(v+6tA)jSb{J6pERgcD8VjN<Dj&8EeQE={y-v=kn=C$3G?GJc
zi`;pEV(n1Zk@CCU>3D&aMN=iu3$7>_P&TmO&m!yisASX}haF8k@w+GNg5k~mq<Gt!
z`yQoxYaDSno1Q*>ntt>9-}<P3c2mr;NIo%PIoNeJG(2hrEJw33KjG!K#l<Dj%t*$u
zbcaj%KQ%F?@|kqrQ^u5oPtp+2n482Lh_unR_$CiK@KDb9*uZhx9J`iSoK}6Hasbp7
zrE!06C++QSs{U>J+TGbqJ6ms3S9Nl3X42nE;<(EZr5Pm{3+c3d&*8+1U0+tv<OOLf
z_Eav&!wZ9I&Mq4i5{IpJI#Ahkl;7-N3uSd(zX``0Ld=y=?rCnPc>R*B>Q2?~C|=j=
zJXnlstGsbvV3bY2u!tO4Vdh16_dzHCqo1IzTS3O6z5Lh_!#A`hr)GTgDP=M}J8LBX
zyU;?}g<_c_QpXrn`FKe$w5u($>vebIWqR}MOQ+@jlkcS;|J{E`_aA?(K5$m?t8W>Z
z@bP-==IXY)O7Er1(mQDwqbcp46Z4`mb+USyGg(nEr8Yu4p}kSQUJoE&B4eVxTIq#i
zsH5`TOMBZpR=(ioctY}*vOytr)ILc&EskA1P3@D5bfNGlb2wTQ<q}FR>W+_fR6U-W
zT}TtNi)nm%Spv;?_tc@<Ds77XX%Z#%u=?+_P8w6Y=IFh**CVd`^!0pq%iq<TRo^pC
zJq{(uF^a6ytJJ>WJ5qPdmwuHw35xkk@#EWL#6!PJ+e67m6=FB1cN|@?PLZtCEuZ?J
z{qF1?y4~^8FW)zu5ME-Y>+|C@nl96%(l@X2ot@G1Nn`JE)rWb>$un$mekwg&TS#lG
zOKNw+?tj>o$5FMrs?!|LN<Tus!LBKkX(wmLZrAifd^2r(Yb&+(cU6{rk1G`ZlsRi?
zDhdPj89(|v*3;spa`kPBj~5>o=V#N(@}ia2-g$dRbvP??sT{+qKeP+l6>WthLTN*k
zAG@nC$us&n6m3iMlj#RfR(zx{>xMoqRP~27Kf+C`58)eD@JIQ1pnAd%#e+ljb>c@~
zN#<LD=U+UF$p@Mt3VRe|>@>3yPJJ(Fb$UrF?OW{{IU_%!?p|JzC%h)k94G>mHAcSx
zz&eV;vvW0`Q`CK4mJdZ4iYe+7vbjk}rg$+Kg<QQ>1eCEKim#$vBmMMq<=ccjf^iS&
znNR)Z<p<(sXK;BrU~FV4EzV7)#f4e7XTIyTe|W5Z?@a08rE~g1)p0#1&Q&*<uf-`w
z<XNmQQ6B&d@L?ZhMm`sJzh7M029AHG0SaAjxG?S0pauHrfST4W|6&Y@GyK9DHtzl5
z3pDLOzq<+j;oijuFip1i!)^3~;5K{{{#G=%0b##w=)Weu+u$~yan#{|;;6%nQDgA{
zR;|dwXOBn-l|St6IP@!ukxZo@K3HXUOU5CZK-wB^hG5X@;sKiZ_#zUA6;38j@ZQyx
z?Lnqnp+hF(1m^%DhpnKvIc^+5A1>s;892McO8`FtAtl~sWSse#9^$wntSBp)1L%;P
zua-5_6INHq3@Y&8_=Ve;pfykI{1uPg$Xxhl9@$PfWHyl3$m}o~<y%c39@xQ&7N-GR
zwgT8mBGydAo%s0EH`j5?=BR_t&joy-(~XA>TciD!ne65vBI_UtVwr7NOf!H3Pb)n<
z7<4kwy3j+LK?QFxfxFbhgf|&@DDe<tAm#@VycHDx6FpG<QRe-Yk2>Vd6FPh*R}7q>
z<pJkO5p6*aA~Hp}T|Mk!#zRf)WkoiDXgsKe4#7VVoZPT}w_sUGoBlK}@?H-Az~$dH
zo{{~&pmXYUD^-Yk?mBJKE8dlLzWG6~LD8iyPksrfd%XR}z=w%Y982K|k_IOXEJ}SZ
z$B^U*_<I23K{T@H#vp+Sj?a2#K%nyRfJ5cSBw<VgBOlwLi4hYF1~%S#CcKYBbo{)9
zqk+5R^?*fui5umr&pKAUVG?w9dXmmHC}ZNzz^zQ^T?LTE6POs_$1XDJQXIj-0E$#A
zs+c4xQfPQ{j)5nGOD5<HUZ{f-0H*$zomx~YpOMdmS>ecUXO(z+Aw>=bSY#LZ3wW@`
z3J)^MpY$V_Cv=V{?BpARVdQf?hF$~usmbZIys)f+Q!E&`P0A~l(QzPvtS&IQv)z1b
z7Z3oUWWgG~0P)D4ozUYC&j=fo;9wod9NHe4o9w~44ncdx6%@KmQ`@m&RBW@`R)JaD
zqN^YcLS!7J>99R=z$bp-jgxn?OUUFI!cnrp`5OXdeuA7IAycFYpYRJ?;A#kdU^6e@
zp|xaTH%*w)#KOlof>KAoWrTk@K4^oJ5gAX9lIcqQ5YV0qEqB1(P4+kBxSatxPFsay
z@yTBLRr2drLK8QG^zhLYnHZi%L7bWCA=C7tmDuS55{DAEe-THdCv!J(hD;DB-R)V@
zh^%#baTnjP0dfU0z;;4qdjS!JqD%DhEOvl<7Tr){t#b;qmt8=Fw@($8{D`nPBA)O-
zUmcU4%$MLnR&#P2CmeX}y?Fp0;=(pOL<wIralne+XrB2Qz?bpjcY4$z-}n-gXMHXq
zL-qWBFDR5qp|1Fpu-Po>1_trHJudLQe(#gZX(%MnI&;H$6W+lq2YnxuzGV;;pJu+k
zA2Kb0E7pN)_G5z}{Gq)WpTm^!4VioFzSs2{zuGt4&HQhMCG;S^x0P4$$_|0slW^rS
zG}zzp6V&dTeL<5BXoi)|#7!JfyAiD517|ejhd|aKg#EEr;Q?cO#fuVy^P!@HhOp(z
z<l&6I$$W$-Pm7%PXAC787M<EhhiQLjJGJ(AG^RS!IFJ{i;<zxzTai=2gCEKe7L+;4
ziTR5aTuP_MvXCkQ_*XktSfqyMxOnotA6@vdTZZqT5D#z3vm3*^G$yCgr1<j<+o>6j
zSK}Ko@M7F|so~&R8Xp}_vy)?TU-`6(m#^2;({J9Sjonsydq=!SCzB>Ik|)NX<P~x8
z?HqRQbDY=w(o$MnTK4>l;~rTozt-6Eu(hAwyndBlJ%8@q3f=Be9#fzCNW+tpPQ3KR
zcv$0JjeZqF&evQdW$aGgFcuEwSF{E1q#+z*W(Ka#vLP>AYk=y-eokT_KX~pO9ku;Z
zB5S9!nkO#K&!h>ZpD`*5Oy+ltt2Ny9B6&Gdin!S2!y-S&>aobkc$zUTGO-hfh0bOf
zm*e7CkfofxyN|_6<QENNdUmIwNMNCnBe~c?#iB0CF_fSvoR(NrRQ(tq9hcvb;*MP@
zJi?%1Ck#7g*a5=smM!Ho${6Bf_XhOjD|yDZeUN|uu#@%<4pXbuNgSue&L#4aokuKe
z0?Z@$R@$+6vzvyw28t!-)}%dl@3A}Q%)9FN?#t@RlF~4gUOs=JI`k~9Z*AItetyBb
zKuA0CpqN=^_sGnw>p2P+zT?Cq_4q`5E1A4;{(1DLvR8hw3xks_Hr~8S2YXxcAM(p{
z?0Q1k#rz9p=FAjzQ~0bL@w{3Y#)38JA)lBZonNp(Pre<vKC*lB%P;;UnxnM5xS;xQ
zp3Y8=z1t*?(ns0Gi-@$f5tly~bq__DEJ`B^D|TrOD}OFjE*#qz-~H)!RNqlrC=I;i
z=*4b@^QkeCjYa5FpDx1-w$y9d0{s_#0!ogcWp@;Oc}bTYKh#so%g5fUUPS*C{h{lj
z`wfmwoKn6_r)infO4Fn+l#A@%n3<kP{w|ctX>Dyy?RdfU+Vfe}P3kNBD0}9uCu+Cs
zg7QN4^+np-c%Hue!*A68ZFu+6{YM|Cdk-J`dseDhse5v+d}ycry_RLCov_>0J6^?`
z9XaGB`QY6vs!ss(T6UkYQ_9EdsV!5tW9OabvyzE<>X_ONv}6EBusOd~fBmg?)j6*R
zhz_N<cfY9)ib;}JI+ko_iW8ZrZ}bn;eU4#7*~4*d`>n2zNo=*xd>ypA-VsEZp%6fU
z&bJ5;5B5|p=jpEc&5O4~sdK71u)C({(CFW;@4Pi%c4KYqw9>1MJ++nnw6%X^StOPf
zzcXm4#FJ+VyWB=br)1Li4f(mkQK_si(MR&qE4ygtXQzDe?LWR7c1QJWUgi5>QT-h|
zbJWG8tBcfDog}V9cI_%1D1hAeDSwa?g%9-*CI7iH<dEYhkCZ3sZ?<>${aq`*2X=VW
zQ5!mSzq+@#pIQe;?vFakpLY8wovK~&JR|SD`_#%B^_{w?tCS=CBW)P~$GX9#Dm-~h
zTRcBif2Oe1Uq9>Eb)|a4MF#@E5|%ndeLxX+qPB~&>PTrs@kP5i7O(58^Ymb4Ha&TC
z&qui`xr(1ULHiCRKfCbQElK?kr5eX=s{C6yok};~O+yjJivm7sQ~3om^HTdSig(&y
zk<V*OYLoOM<TVN~{N2nIiC?VHMGI!#Mb*Xm=kGis1hOi9R$?m-@+gjXWvzw&ME!;$
zQV@xvIxpGjZ;6N6M^!ZM+-^uO&$#G=RNl(>IObJU5xu6jw4Hc1up@3}l5ZH&jyMsC
z-GSq2N%d<)eLpV%piJJ?vxTD|DK`{fyx8*g@=WzqbwRx0k9@R;SQkg$9B;PYG)uU5
z`C-dA{+R}7(iOLVnE7f#ldi|qRnib36FazqDYO+$K-2v$zkcWs_bxtwX|lZ^ZlfOr
zx8a-cx1zZX2>WeA|26U52DkCdim(6jD^ZAw&RAGaSoq~-L)=bCn8E~kW(uC>X18(C
zC~UVi@VTuNUs;NcfpQLyFSMY!*&eupG6R$9W(2$o4W7C%z-L}OAdM-!n|FFJS(Ho&
zf-eaFv4aMW#f$|xZ*apC`EhfPAOg}w??~8;FIpY`#(~V2-QgekQE1^R69I*De&FU#
zc*_*-h(^DH!s5KaH}i2Ah%2JXFyJr05Dhzjg_PC#Y<u8+vj?&oVPKM&V-s*!_np&g
zpP_3jBY3-v0TP{U$c#T0?lNJo>nf*%fu)b7<KdwJCkiXxIOO5UAQHtFlcLje2ACRD
zDoRU6SUqHTNbz=)&+^p+@Q61THK1mZnGc5YR*GX+Uh^ST2I;(=2Jj%^q2##q(Ba_|
zACC$-3CCSj9@LnR<+jJaPoQWdql*=X{?0S3$`jH;A}KRSAdegV&_*d&Z>LZ`P>a~5
zPx9m09}jd5PM9>7`oY79HyfR|<XPrz1&^YJ20J_%Trl_`8t6T^p}ZM%smw8X>+C$g
z$>%q}H0Yk1<y#4}KHqvmlPiuq<ZUh{UA$TAbSlBr5w>^pF>@4N3@Vs(F&O7)3?>cd
z3|JHg1N}JGg^3i2ypF*KgA(D$E8eDI;$vAAM+_X5ZqZOHQO<(%<Vbaxx@@@{gC6IA
znDK^dyonllcdCI~R|DG%4ODouhyt@RkU^PE@`o_wqht~f(g-kcVt@#cXCAa=8f2yn
zohS4)^jBif_a^wT^`s`b$Y5d;naj)7Bm-ywFG|^V6>QWMNpF;HCL~W3mr+9GG6$#|
zK=8<1GY>Z_?C=F`PbkAy^OtFq{9MM^3w@Aa53a>5VRg&#)ev_0I4)$g%#?~kW|_f2
z%e}%Ov?&F-qL)~M61y(>1#e`ADxl`WEyCGw7zgtWlx4xocYFA6ocZ98>4Z2ogJTXN
z|0#UxEIi|eS$G#z@h1$!^Z-w8%#6OjqS8Hbi=<W5co95E1cbvOO5=i08E}UHg$9sG
z+8aD-DRX{8Bg1k9)1bjBqsnUMOIVRJe&K5`UA8yl%QygQ4sMA21c<FTFg3a7mkD@c
zIByi#bi()Mhd5aF?4M()%Ew`B5mc(HjKU6veu%bJ<IBBCTSr_t*S;2zw3}n02!G3!
z^R_?^ReX&o+>F8PJjmgk$>DE0^TQ=-Sxf!{|1%U{1zZ+=P|Am^lnM}}!t~(>Lgw2W
zKHD2XZ^kcoMDSj4ohk8PI<t#u^Z<AV;~ey_f%k>4`__Qkse0JwS3}e}`xo*1Nb8`Y
zy^T!|KooU9;%(&1Kn67q`jDtVZ#gv6(VwQ;UF7Vgt^6@+ufSJf1DHh^o!x+WV20D+
zu#IOBP{otM;f!lu;G2Fm;3CDn*Z|{h@anySr)-VOU=z-Td%jE%V<a!!$PEQA9if2X
z&1Q~u+TGgp!YK==jKjP@uFldg18KbPH@MXSv9p7PeioX2e3{0P?2z&C+8R%W(o5qo
z7SZIMn4Gbq0HqU)*o>vdC#SqX$3itYzKb$FH|rw?XJk%Jb9CC6<{o^PYShOPopcV<
z^QT{@*RNlvz1`jP=JlKO{MCBeX?4=c<#2jCta%EH(e9fCy3vP3{bKjcxa8pbD|6l%
zf|7v~F_;S-9vq~dt*x}R@kaTul{)RV(tNJaJfW$V@QFN8E-D_z>UCMf^9Bf9n2bSH
zc{Dd*3=ARiFvf&W;Wc!JZ^y*9mr?jIZl6=WF0rdf^RKSP-rbW^zirR%3%+lK(uN&-
zEY=d=IlEVOx2z1|s5|CgEGTl!AdB`aprZUTPV+xx<JdXAOXb~OS#HM89{1`;`SBv8
z7k}l)c$A$hd^3kdOjQoeWeCgIS@T(y6UqpV5k#p*dKtjoQF>U6_80=Wlx~jsy1pJ(
zJ(%#^g_qbkVi1KVaxrcff#(OL;aKwyc5a-XGvSQgX~fC*!zRW!>TbllwHV`)kH;s>
zk5L{le~S5l7phsjRXpq(Vo~$_w3F^iOy(k+J3DE8b1SW{Z&<mBqHB41*=0Z#V$QU(
zvg&$Ic)lOAv%TwiH1blPS=@hn#Ugyn#hDwiVE@JEzfXVq<8RaU_NKpo!}!BhLa9M^
z%9L;R#BK`mEnZNK9a_if@aV|p$hVMCuJH}6)s;2n^{mp`Nt>JNX?JHUJ%9Ezy?pUZ
zJTymQ&USvL7-+NXdSK2M%1p{1*-+YW5#{Iz3;HS_x&5s_wF@hFna8Q0;5*DHy;whB
zcT22AP)wx4)xmuPa))2^9jrriRWDKOFrP$`#vGNjc|NH)2oolY;L(<tXOZ`OpJ{Pn
z(MQ&iM|U)CJmcv7ldg9SqFh+Jci(l6`i)WzesQehvGVWCiVhaaN7Nq2*;#bBo&Nlr
z|2u8I_{wSL`+~!F$J6x?eb%(v%5~b?J4m~G``-0N|3_Oc%6t?6jKwZgmp#wr1x=-w
zd`01jB9XEr56H((r(<O+{R(;Ox~)8QUMn4)_L0-#-J{~^mpmya)m?UhVn>0?v3%4q
zEA>=X=SnLtFz@Z}r;Uw`w7H=^Y)gIP#&&wOv6D7;57U0DtvV;ZlGlCr%ye2(8=f4y
zn_7E&O5=gr<Zx=AT&C^4Ln~g|D9O4f=~(%Uq7|j#QRgIW?j5O}9Hv%V^+R=l{vD<5
z{Itxmp)@+A9_#|8oXT*7U6Uy3IZAKbO3QJzZ<JN^Q5;pcG(VG8m*>*L>|~l8A4#Lb
zs)Ol5yoS??`ns`U@w;TllKRFwBk7=hW~JMktsU?Fq?}Qt94h~I4q9oSouQqxw0C%H
zWjWtoBORU2sp<kpEplQ_><~TfqF6$yOIuZYqmD~<`ux~cIIcQD|JDTcCu+PXEZJE}
zMWr7iZ(Wh}r#|qq1d4s~4h1Fhqr#zGQnu73uOkRJqLL#@+1(~?N*4;uBd>=bFE6Rn
zS1ry?rAH4}((Lr8<@C{!XBP^~_xL8gLoU7%cGvBMH4fxQAxL^TauUTCyXj(`FxGBp
zx2|uJ$7?34!zjU6C!zoJvp{a*@fwV9<v3NIXz?r~1H?yRX#?a3>5Bd#=FsM+%WYHg
z@C=|op}j|c5wl_Bh&7y$$1=c|?pxu@jx@k0S((2sW7Oy9j*B11vo6feq?N^aE5L}G
zwJZ7;>fg-FgvxA4`Eu?Zfo;_j#ylJ=%S%B??b>TelEEWh=LhYan2Kz0!)uUv8%^*G
zJ3J$8H))%0`7?6OG#*<R-=wcipu!cqo$%m!Y-FHp&9sg1z#x7F`rS?F5BDxUfN8S5
zA8w-`1h?Uv@VBD54G8;fL;p4L-3GVuwBqZZ{^`H`3NcNin=7;jY5Xy1y*Nx@vsioV
z7@`YO;i*{#FB80j2P$ZcQ;dWQr2N%nGaXkn=winM8W2HS$wqGjPHvp=r}2kdSF{uJ
zAUrf!eCif<8bq)#c(9L0mM=^iO_2x)V#npcCJVIf;AIlOEhZ0o6BIl$>6ZiZtNewj
zCK>S4&&cAJLjXb<iU^r7uIv|(-MT`o5<+Grr!p*~?G$){ihqUIj|?P?skkJ0v6>>-
zU}wNfXAS{_x>-&1c!QFOgpZ*U-W_?C13u-=z|2q`bG+HY02D<B4>0mA9%h<Q=*O-c
zCUv|Kha!Ra{kEG1Fq}orR^hH_c%$&-B;Hgi3JM1CV<Vc7a#VmOAw1+P51@w}eEeo1
zDA6-e1CX_pkNx!cG322CZXvSO;3lGkxUQ+35S{cuQkQvLu9MmBwC9I(d4Lyx%FhD~
zh3BCPQ|LXwP+l|9VZa`FoWm0qB7qe~GI=mkM^K(wQRM+L6Bs5Jst+2_@KE;=gt`dB
z0M2Ln>*6Q@Y6ov(QP#Wx=!q@^4HhaGkSPpDdBoeil1u#J4Yc^q0f}%OLt*5kaAPn>
zKN*wJA{S}#gh_)9E2pZCAqGpXw{qdnn@*k(2*I0helrY38X=Sq9+>4ftk`+RM53#K
z+*^MTJL)czHs=xbnEX*b4)LaN$t&k|uH(W{CJfFT2|&K%lpp@OXS^mx<5PYMbb4}H
zlP7lCSu+30gVKiuv#2lJsX<PK(jNI2G_F~o`KkYq%Rm-6w0VJbsB9K0*aY6ug6Hor
zxC87|5(wo!9v}aM6K2>U%<^SdKy9<)y+IL44VSyWf-@#U2Jafkvl1J$^jW!wFfL5W
zuGj_U9rP7s98l0DKe(0r2tclIm`4s^RAin}ICByW3=7>0tA5aA3qyM=Tm*$T7~7!E
zww&%Pbio!N9=s~H_NpwhMIP1l3O`fZif_0YxEz<`!LAETU>zt3ivfw9;ed{&kOLIX
zzQP`|ggtx;-ts7Lkt2S16>5jhY33L8(IN-Z6x4;Nr^O|}Uw_pKWBJ1!vRr(I!iK)o
zI5USwcH0({BZszs6H2(u7eXBT!DauL#Ia~aKZ3iC*9%tARQGUW7F+n?(#1b}INdn-
z%MO})m%W9P1acr-T|ZL#^>?iJqMU&!#ejlqLh<+WpoEdp#1|;fxjMB01Npua7v)VX
zgP@jUu)jRA<u-VSzr1>-#~r`$csK9JgP_4)@~O_74*)8MfJtpYJnIS+S&Yl#G~?|D
zhs@!7_*B-Yhure8#upRvaO1039iZ4Ho7|?;RYSNSt?9#F$DN<cgjZUd_yUE#@~gdU
z_2sdOXpAfb+1Ez4b4k#eUoXy(4WQ747b&D1Ag=uY?#=WgPtzYL?k27msOxZvD`>$N
z&N-aNG4KO}Efze2BJ)?hXHJ1~1Z7I=U^jI;hw0+tR9-Po0~kBeL>Q0ydnPP;PS1GP
z6vwsmeH9i}89OqcVONX43o<sI#(hj*?BYUMHa&~73;UG9Piqdut}l*Q<V|x$a(yXD
z)z@BK$o<yeb~)}GrLVvGGChC(JRP-<Q}^UDonH>6%j>&VHZX9Yej%^FC#Er9^w;!3
zBvVfn#-kc*^L;M9$HUQ9t^NIU<YOO?JPu`t4mFgw*U2A`8L2awPvm3dA$iHYl$95h
zlwDOn#|=D&c*MA~%m;#s@g_7VUr<!=Vi_+na!l6T!lK5#qh7dVR|X5N?2=%I$H^%R
zzVVG_j$x$CcDA<^|7DtG$B)MTyj9P!Sw7y5T~f-MOBNnaV<DVxzp&FTlp#v1(j4=}
zc%GHx_yAs-WPHueBrB(KdVLh4e#p!5QL|HHl7VlioqNGurRVw}a>dEZd(<Q1;^;&a
z8|;?S99Z=PCF%^{5}H=LC^?*Qv4|Qw1jsw)yBf=DPB6xTtn%cp2${pM!-^etB2A~7
zvoP18TO>W~I$^%Uf;Ea&j=o~Y)aQTvJ<TMIO^kVwxpjcTW7kR(c0Vu=X2$}H+3@nt
zF3nNk?Qbe+-m<&9?IQ=r#G4&K06Wt-=8<odvB-PW?WQ;D>*>v#*FH{dYjZsv9qgs^
z)2{O5TKPBP-CsVU((R79m*)6+p6~B=%gt^e=ieRDp|-cUthO_gwzjs@(=Wde|5u)~
zp<tkpnX6OZQFJ`Izv?_j$>cQz@jaLPob<tt6MqLtdBl9}T=TrBo6M<LjAnkA>#z7y
zA1ELe*Zqwk)gkHxvb*hZX)wd$H1&<072fHA;!HAmy@G{knaJZKQ)#!DxqXn~o~x?P
zQ%KB1Pfk%N#Bp_Z?viJ-D!WN5w(NIuqIghtp&X0dnao*vIh8&M<;=#5uhP@c|3|ty
zYb$w@e_Zu%Vm93!pV8oM!bjt=>y7YnR2^+M=IS(BUJMO6l?M{UZAWDu$E-y@ArCU3
z@H#v^R3F1j1syB6z;l#g_vkRS<;NmBYc}Ku9SHoy$aNV-8~v!iT{gvWv>cllN*VI4
z?LMI8V?(>0ZYXt5RA(eN<8jI^cKr>jE>F0P^35aUV@Dgir@HD(j!+V*pYV&(;u~co
zc~pid8d#yY$a%pFptEx`X?1BKt*y*!PJER*?L);irv7+1U8GU*NLJLci`6@MRkxl0
zs+a7*SzDY=50_`sy`>rN>_U%(LKEdE%9)nh&Ms-P^5`m^UEWDM`+T#p?Ojmx)f|~f
zpQ=n!T6wn7w+zegEFCFdQC_lJ5@j>}6we|0Gsf7|adzqXg$~M;9gVIhu#2@3JY0`4
zArj!vS_JI@1*pT)!w9cp=darxE3MjU^Xx36{9WGc8diHdyEt<{2oLJF`&G(U7kwQx
zrd2yWRo~9BiyURRwz80x=h@9Ur1CyWn_IhS|48*lX`@dniZ9we>pQ%-&dyR^Ua%5R
zc|`rfqz~apOV?rLTkM`=Jt^8tS*r?v(yV%j@=Iw$Zg#KHFZ$&e$?Tn@!+~>N2%I;H
zkC)`5UkIMGd7fwV2gDV-prc=LHtNQ67Ylg{Bl;5B6>DardN#xwQ}jddi*+*56IfTw
zqgZDp-}vs@vUentmu-z7P?(ly847VL8PyJsmGA7@V+VDa=UFE1@FiAmvR(UychO`&
zQ(=n0<ALlp4&TazZ%uC-ZvH*A&G7X3wV!R3U!ku}pu&a2c3j|<J?{)t;N5<{+ujfT
z;oijuFip1i!)^5aFbFpYK7{T=c?|->e%sK0O?<b(Z9MaiuYdlP7i_2r0-6)#p~#`p
z^#*oR%nekW<&T|48cthig7ym<F!pjBwq+_}Gf56B$c-i_%B4GUSV)<c3A%9GRTx0#
z2Y~GArGv8IF~d?8YXiqEI+n~j=LeXO5nO+MWKxBUu)$ll2JeuonIHa})Q~OQ$qzeq
zRsQBx+yxnCi5K4W;$P6(SzV=}7k3t<aSxfZ+qYWS;7weCw#lvJfA+6v<c1$;u)fLR
zjtf5?bQ-ks_81dB9#XtnU{ZJBaK+)fA1*8g^5Eb>&5keLl#4gAF8pw2pv*&r$sHe1
zz0|-4x%dIRO@(rShbjXm-r$Niu9#dZKFNgg!*9DVan%3;B@4mh?X?)tTDii^LcqhP
zJj^0#%4f;%2N@MGqW<>_$XCf2<$^P3yUH+^M9#-pcp<rzRg|?K&`v6TO26``0@V+8
z+EISazdWFb!ITELijOx*$y3KeDJ!pumH`$fd5I<Bj=>U9I!2Wj`Au0fd2xN<Z9KVo
z+m6ZL<jjmG2E%?si8q#FFwFox1``&ul8IqE@rBY$6K)2Wyg7&%l7%=pnt^)g!A35B
z3+s#^pEMv4-rtxLAMzy>y9}i8QhLZM#TVcAQXJyTV3Q+K7^uO|gC>~{pUZ}H@jbA!
zuIjz=7M?CQWeNi`4<?mA<dGG6yfr6zOa}+-{Dg>sjc|@j%q2fRCnMvART)?un4Ou^
zgn>!0+JuRbMpiSH98?-x6%L%3I6m0P?H<Vf-0;UqY}oT<LpQU>fd#{%!<x;Sw;Z}b
z5o~>nc!H|th1@xQbA(w0!V(5J-C}Vu(+Dz$pfZoz0~VE^?m>f9_DlxCJ_=j%C(y)$
zJ8%&?U;K>@I}mbbKfD56+ri61IQT~Xg2M!fL7eib0^^pM!#SIDsfIUbm{?^izlu-9
z3yynn5nWN_7-Y8yeNaL~FuClw<D(6oR{EOa2Q~F8Y{8GK07A<*Y8AABHsk>OX!k+U
z6F8fY22GV`6^I5N8AWu33;d!#flo6`?IJH7ixGj1u-vlDa=GqhHqoB&#vk7jBB;Zu
z{-M}2RC!=qfcfg@&^ZG8i7a!g?)ca)0@ow63Qm5G(~sFDI(Nvzl>m0(a3I@#m$Q}z
zzP=YS(}Z{7k!>c#CLYjb*}qH0*N=W)f8&d?XqtfS{0I1j_}qH&O^EP0&vEBG=%LTn
zn><7<RF0wY?>CzeeFJ5eV-(H1=;3=S$g9W=#H73l?+7a(Y={9}NqImeUxubT>`>?D
zSa9m2DOUs6BW={nfTkT+4ZS>xv$$p2uVI#pL5y(Cu!c<MU-mDT@eO<J&TTBaNq3>c
zS?mQWzx?yxZd+9yOWU_ceL*d=gmwR0+!d}7H@|Ez5bo?KAUAmWAhUp=pD4SVSd*4Z
z{(f%B(Yy*2IjYOpQdYS2^UI4-U&_U^4TCwv<@vG3SBL4i!}piEl_JPIWKcU9YcjTD
z?8+h_i)-_<vuSaD&I|Z%SNgFqh(g21=8eYU>-fZInw*|U3-h91oKMqJ6L1#eE6q=i
zQ>(S-xH&G0W2f4OEiVqUD9ku+TH~<ssaXx~CNwr4k{s%V#elcwk&cCR6i6OxOJ2t3
zp==2yz;TD8q>j8WN_=sGfyT}XL;B%ic~tHmbCNgud7PVlxP#*deYE#z^Nd;Ts|<BF
zqFdxgjCnm@GK7)<KZ2upV9d;Txh#V7jWx#jUdZim6!?Y5$QO#Db8?=#ol_rUMPAKJ
zPb$orze{!{9vltFjxG(;6W`7%yL4E{=i4#N0~w2R+Q_)dFLu~zysWtavLhx+9KQR)
z+@ri*&pg$Oq%3-hhM6Ga@6nNAzg<7aQCvz-Tk>%<8Zr`QY68BRKce`Md@MwxbQl}u
z6bh7YEZlPZ8459t;}s`kcOS!^$MC#?Z$*j9oNQ*1l_S>po)3P^eUXVpf2OGHDQ9t#
z`Rf@A`dQgSI#Bi<?C+=F{q8sE^_!QfkMn75b<Ike{Z=dO>};it&5g9RxtTiMcG}(B
zOZ)qKR$g#48p<r>BHu1fyXl)}U#2&&pQ|)E;*H%tDV<8rOSQMLiRm;uzvLZ<C=ob9
zaCc|NipmYey|Mn<i~f`u#}xWoYRWU_WYimW@KMLSOVjljMKuZz&C7;H(%izbmC)mg
z*Dt*)e?R&7<Mi>zPtyF{Y?_&zNb@r)W7RR*1H0>(zfsq&)Fx;vipaYlQAQ#oJBrw~
zgyMqTBUW07Kjj1u(nr0ZZQ)mB=7mLnznS{10Wiue%FIejw+-q`na@!Nq76~!%D3wP
z+A9}3W=_>UL#d_yNW4%6Pbj^U({q~6jVLcp)vk}dJ2H04go2HA5-JMuP_H@G@<M&v
z%cp-#`<pKv@7?j~w7B{>-FxzbG&#GRPR_2?@3C8smpvm7*fr>52?b>bEOm-ZpiMg^
z$`k6N7ygwt%Faqhl+K#p`UO(zzAk^G2*t*6$Mb9Yt57yjf0ak#2@iJn(RZ>IL)kE&
z4c;90kMa`b0Y{uhUx%FRJRDPd;<(hwNnSc&HzB(O$3po;9Z|Ub19o?HtRzFxMaek4
z;$Zy;g#q72Vn-Ko)Bae)Agz=)?QVARZkn4?T~fbuc+^ejS0ky_xlsF1AEy3>GN%5}
zzp&eEbaW`qPLHM4#hJ7+H=bt4#7o&abbT59ugVJr*M8?LZFBV8x$2Jki3@dj2S+DX
zq|+xYEiI*$wUxBIx}tWqApYzi9;Z0f7ph-ux6^^z>z>*xiYNLz>J@z!Wp}3fa-#M|
zzLA;nLImq*l9{}xoq<#Np&+CjX)p9K48Wu2#*52<qSd7zjY2@5jlz*QiQ9_bGhWu{
zTA_8^?W){PDQ5F<ttYRP?_40xREN(wx#>dn`TRtvYxk}5vlD5Ox}bW0q`vjd=6>4V
z*|&m_*!eyu&tuAQOm&{`xT8=<Vd@u5WT%{|FGblzTD_A^VTROZS>N$mlInW2UG=1*
zC*8Dp)>z<2oA)zCS;H<zl!`n<Vy#Mf=x0XciQ6-kjKs;0XBl;uw(8wslA$Qapd~9~
zm$l-@PZzmN`5-6HLGn51bV)#uI>=N^-CZ2w$r$7Q{gw3if%-|w%`U*rtt~6x=x-P1
z)OV`Qh+#TZ-`_quaNVUlGB=A%>M?z{8;@{@zwyeY@RzOf>!m9^xJrG)*-IDxHLl;U
zHi6sxxYu;SgKP;o0*fW2Zqf!c-Ti*RARN#i?nC$t=5#xKKMcYRf)AnlP+o(8u-`WH
zUlZSLa2wA`@x_vsX%JW@lOh(aoB+9dg4;ng!JjH5GunW|VR5qccV&`07CtkbZT4`(
zGGPu^+@2d5Cnv2M(>!u$CO?vyyY1OOwB*?N>qi(}(IFsTc-t4*`DYv#c5p>zL-u45
zEck&mQJ5wOL%#m;54qKO*c~|Rg*PNdn?|6--<BWMof{ooPG__8aeC~JUD3<SoU<J-
zj_`BL5P%1X)&3yhEgUW<4V*=b`~U|K2hm{?F6W1B_04|GJYxu^{?r3wj@(#4@ui0p
zZ)1?ZRtShDKMHlR@J|OG4<J7T^e_npmj+hCp_Jeu8A>n|AUsessbL@`4$$yGW-B*}
zE*@gM9l`*Y`r@|$<>xnpB!^|8ESzW_NLMk~lgXgW4@=wOuFFp|r7`CnwX#zFK=SV$
zN>&3@1VMTD`r$``2^bhqmK0_rE0-<#B#bAqE-ZT*syyNeoOyD&>^yjmyo83Rsbdi)
z0DFlogv^s5O)y-M4D#1M>O~$Tpls8mYiw#f%`YzcEn5#jH83fPRu4`HExD;PUdT`#
zVo=Ok><q|#%z*}Ke)A2QJaA`MnPRYP9;{Hu^g}M>gRhScpk7H%pU16w6>{f%L}ujy
z69Oh`q;gaPeI{1dOqfWI!jl~=uDC4JkjWFjEu;ZE>2$taF{q6*!^>Mbo~$sLW-zG%
zARp898+4KbU?7hYuqfBKPzJj^h!h$b(0O;})T{<jvp!OQoFffx=91hwi5))F3DALv
zzD-tBDL`4Y#sFs^KKOx`N5%>Y4l6nAkc~SS!UiV#H`(Ex?LcreQRVU-Srs?@?Pa`U
zG?So%nf^)$d_j`#p(3*#5kA72yyEFR%Hg<0`f?l@Z`xoNAbWZKL>e=nDnO10D`@Vd
z0~=OMd4?i}Gz#bZC=x<Y!bX^$$e|;B*r0<CJRC@O(SjeMsswda&J}x^CGWtQPj<#l
zkiQ*hxoW|K1@$t^MEr<@H@+r?nJK}9KhF=!(~$X?LN~{aiJvUkOJ@}!L=R*AOTOXI
zQ^~#r0ON295cY7CzP~!@G0Tk^Qefvs95#ZUaHc4{LF(C)u_TImXJWN+%8<b3Djf7+
z2v)*oPBuxGP3K9*;B8m1bqM6i^@EN$@>W*)+m`8K{30B2+lR{>jF;hZzzrCcZ;_9g
zDE<M(0?zp%FcT84a9B;F8~)HCiSYmSyYEo(^@~t^m3GmDK{&wA#}M9Wa9_($^aWpq
z%?8*HjiVPQdl`4!0!{j2uW&YN|C{g*KY2xJ0;;rCVcx|z!j=&AU8ZE}&BJCa0fTVC
zEx5$wyD$DhsEQ5oW$Ud_2!aW<yD3LgW&`eQ1+WK=y)qcv2zS%o>&{j~ge~}*PLRFi
z<6i;Wg4pf=!Z+NqGrsHR=eb4!yz=E3D%wI5<B<w&y6i5v41tGW@CC<mZ(fAyk?~e2
zVVYp69R*`7?Ci8+$JZ$zct34xtflLix1f(T=F%ME66FO8axyiBV|WV###}su4v*U2
zQN{CRhHrL_4X4@Zab1(%6~jU-Z(C1mEXv~6{@!kUhoseVSYCqS4RZQf#$?FFDFCBm
zlj+XzSZW`0RJ+EcYVR!4vmoyUJITX%(u>kc2V=dmuvm||b;+cB%2*K@V&Rj7Fs4<U
z@?h*ozJm4mK5V)o&w!q~PyzXgf6$m~?q|wEU@0%+3<XDUEIUc;FH#m=St#e2H1gYv
zt$H@_t(z8a`;%0SlP8p~92<5=_N~ngD`(=HX)}JAk+HAkQCJoltq3~HN1!R6P;Rhb
z%A%$scPg0+#<#r6oDd}kFOyL&eADLAJD-42mHmv$bYUiTYq3MCOF4B<oWC<uV`*}H
z)Omk+sQS@9@eYz9jro}q@ND2XLduf7BF!9EXd21PNf0dJva^U}uqe%U-dLdL$SS_q
z;(4yd`6s8E!<@4CF4<IvMwDliC*OKH>~_;ZtL@!g>{6pn@|~lt&27yU_|91JF={9V
zSvcotO^&-;(mV<tn;V<nb@<}d^R&MHI&oAex%KMhH|e*(`M31u&5LxvPA<*Mne+I&
zJ4zD^^HbA&+F&V7&oL*POQYjcDueU1sq)&~c$1Ef_LZiKG<Nrn(mUjQBz@+=cY-Fz
zy}(S}V4?g{={;pWraWTT*Zqf2yi3r>(<}ctB6M<sxz`<~aX0<?-~Ki2?QEynx#=`J
zI}>HAi+K_|{X)rtf@D&5&70Itt#pyyCvl*_hE{ci%irV5?GIk#<0v;!5~;jKHP=$P
zdRK^B0L6@Q2<1x5MM+ngpD`!(QIGQXcL-^l3Qz4qF=OSB>Qv0lg%?XJf7nT;vhvZ1
za+^kV(uySJrJ@V@?xwfrowWY^^YrrBpEQR(O%wC0>G3B&Nsqqw<21LpCZ5CMK^TtF
zj9smKn=ljt9M8{MNX)a{MkO1~lt{v^I}z$)tp>TQypZh3!eW1U!4#9W9{ov#W$sK{
zP*l!`$YZ(X2Q5l^UTWrrgfo?|*L=hWfA0*V4Xe%b&9YETu)}H83RTLR`7vpuZ$MFX
zru<^vhn+^#Q&Vb3Gggw&FQZ6jZHjUw{p2f0)cJdE%-63jl}Di{qK$TZRAO6gi=A@x
zE41w>Z^DnNK1{3rF3!Z!h)Lm(l|LNg%C56^r{lihg72%HU0S&_skoWf%R}vv@Ai$P
z`GxtkcK=>lTv~ElAp8+K*E-$Q<)}k;cy-iQ>N!B)$ylCx#l^Y>>lye_9;`>uf0MU7
z8<Z$r*e~jG&KG`uuDB{w+Vth+nLv?8*;}FDG7RNcA!jYa>pkMZGZTdmagtB0L*U1A
zj|&|2fw~S2VR+0;sIBm=s;i4@mCHn$nNxo+I(Gf-?j5O5RoX7uJvtukp7yCUq6`b=
z^jMmopGymB|9l^ha-omnJ66$VL+MHVqk&oJ<@OZoJ0zYI6DPZ2y^~k64|@l=`r)Z@
zE8mF|@Op~EdS1$p`YoUzfLrX=iZw7U^2u{6(UaG-696}HyUhqkTckYQN5E6^6dEA<
zh%kwtc<7gSKHghhQC~K!dV1h*zFoe(<Kve3lm+eS_3Kxzi(c;&FR^xdxqxD3j~n==
z`EGk$m79O%7xo_iY_I)l6S&Qfdrc=(cF&b<4=lLaT5JJLcfTJn2nY0s`w%{ZIo(d*
z4})-n;6vy>l-D32?6(d5*Ti=l+{Uvy>d-EPH622$)u=tuT-CJlf=d)$knk|Vh(?DW
z11zi#Ybq0De8FYFnmYsk?X6qpVI(*xgzyi1_N!>Jfqz&P$Z_H@1M>`k5PSBo+_vcg
zUvPmhGC4rzYoBt<W@NA(26$O+qYDpRnWWAmpr3c9uS~r02|Kj$GqHbWk1ewbki&OV
zD7(YtZa?$Jg}Vd-2Tiv38K!oah=|VBvXw|}WRCSMo3^ncufgPmcYAR}2;W>PlQSNE
zmK6u$>w3$>m4|{laex681CO{+cx6+ADjwJj{=6fHhlYaiK)BSyvz*(_fNE-L)@NoD
zH-6pYjt7@K@pz~>UHtrjil3#I!2^m1HxDL1#-LRID0t9WP9?yLhomE(UH?7-`Qlle
zLFpjzPK~@&E)49vpi6_-Wq(<nKmLiD^po8YETFJy;sg%QpFC1tGI{ehaSdGk#+%|}
za6vvp%fQiZoym`C#6Xa@@#4)d4aRhN(0~l^=V8wqkt!d~-RC$>&V(Ok5a|IRXYMl?
z<ZZ<XO|oJK6a5?u5~=|%PZH!okLnaV$cTf%Ad?C&L}<{&8@<jS$;M#DgFX2LFzHcz
z8VHfL_-23xU~-db)J6DPi7Q#j$#{b*@`c2Zw<x`saQSVw>zI7-cHfyM)h45Cpi@4P
zUUH!*+~NSfKtaE(^-y$i93+!4(wK|Dv<!~Ccq3VYS?MEmi3mR9W10x4&2TIWleX}u
z?-vP-a029oErFy7_@=TBJATCgQF$Bcs+cKEB0K*e&p<qqUhzQ_eGvG@GliQxgLlwJ
zltzRdc?pXbXmG~sa`-UA&yMVEZoa6pK)e_ccEIs-IY5NU3<a{2)tTN#CJp*N6N@tn
zJCKR-H-x`aFTCa8lFXGIqbH1jgo(f}xmbSD+m3h;%N|wHd2LoDkvTtr@X64$mOPP3
z`zk?u7fyxbGAhT3k%^J*Du)A1fm@b>CjJm7rxzdn%L^iPy&{jm3(d`#fN<ea`-Nvt
z9XBrGRU|QZCdY$}8ex1U<AzMc4X)5Q@{9*6TDt@er#Lf>CzNs&f9NCbY}ehta2P@g
z0Fw)UnN9f&A%77PKAGBh^OnW=k;~gYjj|^_WHToHwfPxBvJzkIBMkhqyWmTeuuo23
z6{HVLt2n>QZ+umXFQ-UmGY<fgMQTdfkk^?8Gk$jY-tS+4&;#Sh)*f62P~>#U!zK)(
z2TB|^94VH;v;pt%lNZmy0woWsINr%OqAwwAt69}`p#jE;l(aSdd+kj#;%|5>iiYDI
z&<xXb5Lb_%pkLM=H}u(Bptyrq?K{ZcG{YVy*5VgX{QB^L#;3w(R1erK|B9wz*8FPz
zHBNw)%h@eI(?OWswsH%6plkcPYl%Q~F+|K^@>Kw9|J954Q%L5`u=2Hk*r;5LO)oT#
zJhSqP?;#y`57Wu<k;ZNvx#PGgzqPFk#mwQ+q2Fd_(f$1NSbVND22&i@mwsD)b!pxU
zswipvXkwgpN8>JznbKemc}6u}^J1LFxGY$+5X%u&C~9_gQBZ98n_TQ}VqtxRoh4VA
z_ngKNMdxS7mfho0o*SAMU7W{@fW+gpDPF|ldqIS~y5zf8>}En)R_|zHT<hFWJM$tO
zveM?nL)`7h|7IUk@|FuDn}XmK?Y0^#yIcZL=&-v=lpgc)q7@5rC(2hA|6PU}XJ2X@
zi~<nl4vUU_BW7xRBrVR*di?y&ix=tj`nt!zi%ZM?E}OrxqdY|sM*fiJEN+q}^4#Ne
zc25b6msH~{wtDWsF<LBcs{GhpAz+-%e3KU{$sWF+G&{){U*#+Q>@4C~LdN*R!`D`f
zQNEq-S=u=`N=F=5t~e-X6j#oF@!;EGDAhP}j&Bt0?;j+Nuwx;Y-6PDOE2Sk1@F<Fu
ze(DJ?b@D!(ziZP`o@p)&T~V}L$<a~1aWwqG++6zf``=4X9z9Gg#qspn)6_aVbX{|q
z>d!ov#o2pnYw44ZKS}rR-<PbT=~(s6JGM5~)9&uBj~d+CT2K2soBmcDS8IRQc{?{d
zrE6a0c30)aoSwRcQfW%_vqc{lN?YOB&6h8}Nw1ziRhoC)K4KAj(nl`Nay+Q=pM_eE
z=t4>5Bh36lpz>gHO7cC_JZ0M7{Ce^HS^DzJKcuH$f04GgHq*N5!^WF8sjGHE-7EQG
z#fX(2d}A$?9ihPDs&_FF#>ef0SDEqcFy?_)WK%BrC^2?2v3mwZr|~lRMjg8gQCyOy
zX~~G3<+!m)%|&C^VSFPHlMDG+xMwcM6$%ApkG5mQisT_o(5U?~XO=gz^KBn?F42d*
ze*Wk5?2BKgv!hm;SX6m__V?+d@Bb`~PA#UZYn1l1Nu^IRdSPEC^Ge!**H%=|S#OCK
zGMPKZ{MAYlD@&*=vEvlw3Oh<YXNJBpStZ}O{Gup|1J)UwzpOXNl$&y;S0OBlGx))m
z_?a)u!}XIFSS_>iiX$OCr&oTV%$ibNnVp?Zb91vUU)B(OWT@u6Lqj80j?Kz%d1)c7
zuB@cx#d-D9(`j;wvW)NMS?MQv*mXsl@y=1!0Mxg#gO2>?8%aBRhu%dN-!Nm1fOZ{6
zvrbP>xZjxL#RnD4sp`iO-!&6Ij_Kn?)-AOmc8T2?8CTvb9ct4Yp~%|Bg(g5K=_qsh
z2l5RC5xco|cXs_vEOv9!=TH|p>UC*(AuTN}#<jepyqI#H$2Y2?{<6;D^5glVazbt^
zYFxfbgUVe&l+BccExw}sq7X5m;TB~VO46dVDP>C^4-GTH!m+G>*k3?CSZPf?bN?$!
z5wb?XGiO@mMH{7WJL(>*oJYNe$WGF$JEMu?f7uCqN6*^{6zBByv;|~v8LIB6-k^Y_
zUQF|SO_>Wm{!;Bw?cPT&s?4CboT9g~3E6eSUe*@r2i!(gUc5lCu%JE!g&GQ5r3r->
zYb&Io<XKTxg&lsRF+UUK#*e<;d1*R{;jqdhXq_vJgYr`t!N{G=gScX?iZ;jECr8c3
zE=!cqcdqGEl}^=%lk>Og?>U0<%yI2(Zt5BEOnHwils@rtAs4Q&XN#|XeA9flJ+8{l
zzw!%v)8BcP-312OZ<r0cUD)l07e>^`uyXH(ChmRyP3RBzA$$gNx}Cls2H^(5htPc}
zuR%cAZyWlriSIVJjc29!@*B=9MyPVxKy1R4a4^VjC#u->i#L+h;Bd=P`C;O(0ktRO
zLU<v{AVlE<U@5b9<3xuE&f&tDDTK-h<kpY75&AnSh`<-V7<FW_2hNve&VDwG#G;_G
z<FpLKWNW$k@AxcxX+Fl7gpR1D!Ni~ZkSRng%;6jU5Eq$i`XC8^97PW3WN}vUg9u(I
zYbs1lmvL}y;=u%^E%FvsATwYo1D0Ga+%SpKe01f!s`E13Rsfkj@v4*21BG2Ti;Ig^
zN-)6jq6u{fLh`JDI&SA7gB+G(J<vk&#cr2)D}_mtO+C~uc%w)idqhpD^b-#hY96eK
z3j-D21mdVg-VTC)_n32)&q96;qGFPV!w8!(Fk`Z#$&v;c41Sze9uT@hkrM;FNP+yE
z9y|VeKokt(wC!&I@a$n#J8VOCSumnbL9a{zL!RWTViMKlz2Guc9+S;t5EBCp@X1H=
zmq<OJW$@wXpr=jnwx;}X3n<eVgfd}|$!6L0ha!h^XF@Q^$GkbqUjrKyQKgPriK;;s
z=f_irfVUbX#2`CNzeRhQ2mHENEAY1fJm82oY8YrR$f4d~_h3i7Iai%G`WXB%ab_^i
zq=GQ;_8^L|l9_hzfew6yj{#(~MLrnJ1c4GjhRzY^r1K%#0+L4qYK~kcZ)gJ^xGGL7
z(}c64mI;F5w{j5KiC=iWsX$t&`!RTpfnp4PLV?D)=JT_Q9^?@p7wyY-XNBxGv$x#t
z7vpaL?F0!5IXBW3*aF4n7F>eJX<~PCRlg!L`cpd`5*s)7Yo)&Gir`>%4_1}kBykHr
z2f-Iage!N*s6)^ce=ZwYM!@-_KSdg8Kvo&p%|k{5kF~m>gUEoHWieRlY+}gFp5%Fn
zgOuZEIdoB)fe(zU8b3KAS7hODHzrir`316>*Cxg!4ts}GL5g2Ca}djjeAy2<AdY&X
z{uHEd0Vc2b5R|^5k-qRUP#@?Ah{IQ4eT6;nB|Ln?O$`TOq=RHK{-VK^(}gLET}rSO
zhV~9$I>6vdVB*PkGlMw7*zIuGb6EF#Cb73kvLj;0tf;d+@ZlHndAybV9cLq4;AGm8
z!$YGot#H}DH{WYF26e`G-1*k@hk^6W_6mu`wxEnVSi%N{?p6JYI4W>jfbfU^cdhs;
z?V$<%xG0l=;)fY2L~f&Q1-yf{h7XGuP{@9KDT{0`P{o1aNNd@ebT#BM&wf=rGC+A8
zAm|AzYaMsf-jFFv9`~EF7bsy$yv=+bL|5S&`N8<1k{fSi=O=rikFd7&;BoZQe0#U;
zz6UhzK%oIjm?~VuuZDtyuAc_`0DPeF3Ae!hx^BeL=eA3@lB30!aLusrE;yqJU`xKt
zAner-|4iS{uRsY=Y!#-W9ptB{UgI(}NZ{8SAN5?9$6ofeyk%F^#rd(1b~`;jO2<e0
zy7p50U?;V9H&bhGD|HX|(`om}3*d|$*`W}N;3Hlbe|vG74)?dx?&d3pIqe*!>$m4=
zWcb<(o$2COJa^OK?s_`w9HdUBee>dJ+I;m~VK#){Pj63JB0WfFZOL%BojM1b=|tf!
zPQ>rLo33<S9JkVW_rSQzvv#_^=m@V%_RG_wbk+_3%d?PSRAWfKamylnjHi`uJ(s*N
z>gN-49_1PMQVF6k1ttIRC-0Qs^}xqa1rd}y_Z#zEjFB0uippbd&GCGSfn;Ia!p@d+
zUP2Q)zVAeQEEKadiMbcwismi;I6_O~ZjJk2zj>XWee=wV#7iqnX?b}$&CSiFxtSU7
z#AAWnU>R8qABu4?V{(`IwRaf6*9sN^-+^M+0N>$4;e;{(r4-+zA%DEHU}9XdXXT~n
zQEZ{4#AFu<V{5)G%}$RK%_ZR-3KzZ+Q+6K2aa<hXwYAIJ`K{FM9(#eClJMfF<{`{m
zVo{SZe^R<xXrHpo)MH)(Wqxy^I>^z4BO~LgYa`;oe0Df-4A$cZ_tJy=tEsiW=lnwP
z<@S)<3rdfJR?CWl|M(yO-|3@|zn9k5?xm&WHQ{D8KR{tKW<>*OMhSd$)Jn%4)iu=}
zc8fiH@W4t5zDsm=%F%twH(mbr(Bw>-U0Ct=x=<vpZ>*<pzImEnym*!l_O`8<wL(Xi
zk2#ZA90@rwri(mP-to<{sVS6L6KQRADLuS@U-SHt^!%CH&DLi6*-!s2{r%tnT=Sn%
zE35g28Am;KB<DFt2&&x;-+7xxRRxs<=RZnM%7eKZb(#9W(1f-TN?DX%ES8UXF}d4r
zrQiMbf2801`d`w{*1Gx#6z235ZOtoBD*30UIU1GkM4^<Kvg|CJvmlS6EEeEVJZZkB
z0V6v}=4WS|zp+Csc8ZNpPWY%Y6pJW3=4PkUym(D0ti*GA?p&U#KiNsozx+db_54ef
z$J?}c?@9Xj2Y;8=9{nKQ8R7e2s_&`~RuU+GV)q$}Fq9L#NKC&M-y@3ekcpnTGkEWO
zRsE7ZUOpCH^@VzeGK_hsdjQ>N=T>epzm(}0T4i^B%8rTKJKkvHInVi~81)ev6b*h!
zR^>u{XY>yJFy*1VrSC$iu`oZA=9Rzmb2C<AG2=yni=v6}ljGx7q|VPzr{$&jv?w!<
zLX0v;X+piCUbCx{Zv>vATvAfTRY@jQ|M)f?-QwN>JH6UgdND<zAK_?4l<-T7vo05Q
z7ILZ#iXC<+p_rf^^MXXTjk1}eT|<dVUFoPUafIRBv5C|<A%9ht?n9JL)q75RVU2?F
z;bmBIg8ZZHFD%Zd`}bDU!$%L&<45<?+RBRS6@3Q4S^(wdqi$7hp~z90QobrzuVE>3
zT^5Ry^pO9}aVa?B3#F^dMfHYVlYS{#{zyb0L!7h~e>y}otS<oYgC^F|sC)5(0N?+j
zK9$dvQ1--&TqvYaVv0EJ9vt~7+0(O2m3{2~Lb*AO(lWmcLjAGQT5jYet*sVET((n(
zmkw0HQQ~t%@jQw!+TY};FZxe*>6JB@qC6vgjCGNl^`%%-p&y-0tIP9l*X#)Pt}lh-
z*<&T1?8xByuPf?sC{;bO7EeD9$qPNdN?Y)pRzI)*aOJ!sZ=-)EzaxDfxv1R8E6Nx<
z>uD%^{k=;)OZN9$>8Nv}dNz@k7iQgmvBP(JXD7XU`8pjO9Ew-0$;SK>-i|fLo8xbk
zV=;T|)SHa!y&|0D#w{BrmO=xBS%-sHal0S3yU@MEY}oC>uDq&nGEHM^#x<ZH_da)l
zCjK^i6TlDl864(Aaf9GP>6-9u!a)0NWB5yCdLP`zGmbj^r+@xcG@>#v<!M5IsGume
zdV`0{jZ7WC7s6yG2v+cpu+(qa0}m?m*1zZCW;!w~=qr2y{N+(6rWwXo*>j{BQIR@*
zxFvz<2os(Jsp8FP^?)6B=3&}ERq4VSJi{YoCiR6tX~o(jGn5Rj<#wkMxe{hr9ao(J
z0m6eKY{;TVfdr4d4@%>8!O1pTuw~wY;0?OaAvceeEVVln416L@xYh0A8}8ESWIhoQ
zc9~`h?+9q|xG7A?%mV#gPd?bH?GNWtCWM2pnJT$9j8rGiAPXf)D6d%f_f~9m?Ci#2
z@Woq3-sPnSB{DHkg#!;K^9IZ_Iy}yfW)RNXR4ndai1#@US<yhwK*xhLJ$PIuqG6y#
zSjVf2hhA_5B($6i^2G|NgAq{7zZMjQpQSvA#q@8+mNMcJ<l1F`zXu%3lz7na(ACcm
zjY4S*LiwV6>kR)85Us-a?KTZ+7@%BUMjfP{_^1Q%y<~7l#tJBV%cFb%hZHdYiz7l%
zIu#|=)ft1`i!`l)&A0|OKEGW9TFMcHDaR%7q1a1JtW>BjTNF}fK3iIKinn<g)N8<d
zsR?fRrWUDZP!ftRzKZ}vsi1IUatbfx;+$#cnd%b*K+^2DtQa^+eor1GGvPfUa6Xqe
zyTsGUrfmqTSw9~oZL_8SG?Z`v>7gDvKPV33P{8=!SF|t6%*W^Wr{r4p+_rN0a=eA>
zy6ir(5xGT78BA!P#CCYTLFSGUfQABak;@`JekO!VXzqZ_g2c!j3vw5TZY1#L7YQ*y
z?;vyx*o89*S1mv$kD6I<m`*>xjFQ+9Uc5N6&>W!&r0lii<QAxitI*VLcq5bhMg1&i
z4v(TH29MAaOODHK!iU@Zxv&>prn8Z)IWqt4Payr{xvEQM8CW7i(1+dL$VSBW!x{1L
zGa?(7nOU#u8}w!HUCNCZB06v~9X2c01THU3d>j^_Pt5VjlW?%3f(H+9$cZD&emgEd
z8xR8@>|n)+JfMC=+%mZ!a3nz&xR9rEV>h3I*Ja^k0f%3%?n9Xup0;P1fFU>3;75B3
zd-xW<hMd@lPc*n=m5m@dZ^idcv`gMBwG946eKJ4!6`tTlqAO@ga}_r$h1e-K^8qps
zUs;ZZ+v$RbWs<MkorCEw+Cp51FbBwSlzLE<)1ZOC$}}Pq8zO$^CUFVE4IoRHAVXID
zWdFC{{^7e-d=-g8pc>@7fonkB&y}>)^u;U`6+shjXWbLNXWVfMH07*hD1m~>El_AG
zNvm|ZJZ$TQg0J}Z0pam4RIsb^DjWS!$$yJ~o!0=gb=<vv0X0ri*baTM1GSxdKM$Nu
z9QNDXwH*_8Tur!1+lwpE^p`7N;Ctzc-wix&;{z8|K~r$#M04>XMqmQFOHS3$3|ld5
zP`Dn@ub8IB2MF(<Yz;oOz4pHeH9hv*FWnow4SghzL;msHa&Ig*JsM{2f3G!OdwZ6y
zFEn;MJxZtDRyye%r0&sv>KyK-!@ccvbhz&$r@HN<ba8r|PEXotXX|x(`sE+fAAkF=
z>FYoJA-(+O%k<*e*D{}`t=BKo#_Mm=#;dQ>=BsCEck5L;*nXY1*I%UVjTfo4vyo23
zA7$3%Nh_VB^y<cSs<_Wj4z29Ey68%-V~N*Myo&p*E!?q{Vc;%L+ky9WsH>CS%8zfG
zp<pO){&SooV{KkAq`xlD2zCs3j6vaZtMam#pmLsh-0RVa+if%7vHK+}jXA>|b&WBz
zh+{!lV^hZ2ED*9#>^Yif%R(y)>&JZGh=pN|r`e&xPKxcF-83~hl~z`ky&LKIi|0Pt
ziE+1&GF19Ttk4>_!U?6BkJZz~co^jwi^;s;2rat|P+<A!D_tn4R+lt~otyEZAWDHt
z<{k2*4@NP8EX<w!!kgwc$jgy~>>{ypj9p_&|9PGVu*<`{$uvJfLE@c6lF!Ev-npAj
z&M&>of%JN@nz&e`mV6v<*lr)E-Tgy<zYQ7u-MP70zhp`oaBMq^ntTt6dQHBg2w?7m
za__;tRWF3Lm4AHyg@w{6UzMNY;dnVtd-&x){-fsZtBG%uO-;?Ex%q`Ozp$8AR@c&#
zkJu6RK=afQFUqqU<0n7<QTpt&AEZUu%VMtQb<8z1&v=`LC9Uuy6KQ&OAuX+}rs)}W
z?Opp@MY}uOY5k4b!N!}k$?mdkxz#4vP1bJF#`qqd>b%<I=~*}JtL+^g?59cP@kfsz
zDZZpSu#+A>e4Kv%^IxX#fBy&R!K24%S#<;D$ED_Y9I^QF#k2Iyvp=V;jo0aLe^2dI
zZALu#ej4B9qYP0>P=B~krr@T$P=wGo@NJyuYR`ZE;*V+l)$_E!yJfy#eE$3N^_O3y
zgT38!CRw{}wOh5jOSRSZ;a=*rbg92+w|3L9{O||AyR)uzzw{A>4<D?hRmqGpj_=M*
zO;5xwleww1z>c~Z&28_f9w^+w-lp<lEA4H)NzcFfQ+oZ)SL&bIX=!mOegD&+rjK<k
zt$eJ!AN32r%tbu}WiF^vXYS}@(x_kZlf?}*+OK;;*Rgo%GhX}z=->Uqs@Kn8s5UP<
zYbRdJXR$vkyEwLwxv5{al$$#0Iw3PL2hKPY6VzqC*~NMh?FL>91469mGkB>LB_Q=|
ziJe=@6BLJ}iS|tWXPt*QncK#0ICh4ijNyCl@%?vpmm(huFw#o9=9octU@EZkel*Ra
zTp1ZME#Ds8+i#18(}dL4)t|77Y-M>d-CJF<a*b^3bUNv<)lSFCZ}^WYU)n0CgPl#k
zaImzp;<TbXZSC(QjyXMlE1K?^zwL$6y6h69-}KI?tZ1XHp=@sJ>Z;z-NAV&o$~$`B
zGvT`Jj*rsa-QQ39x}f#<%hi|Ahj@KW{UdGNNApQ$O40cxc|-9+*{Yn}Z-+w5dBu(=
zwOPMNpn9W6yu#DQRZ|E*#LMRDYrRud4N}Yr?-*2$^!w~$;~C)>*(JXfsjTx!9_Xl-
z)Gw5g6H_zk^x{f={Ylz8=%nr41C>KZyobD_dU1Z<3Q*$b<pk;{<;Y9#RA<U;cmF8u
z>?xe`K@qAi7}dpd<&G84D9xkYhurKco@7@T3Ow<l{?YG7{#~oS467~9rh6+3K1#5(
zi%a>@zxyH(c}M+ITZkRTwSre92Kr~e*r2O=mO30I677(76yuV3#zy-g4`~<79kZz)
z@p@fJ<sgyM{=t#I<+Y@~<Fik{@7-V9JA3KHi#P5IOiNVG^@tyc_MgqBJ)4!CdeU^;
zUdB>^D|TGQ=^u3>Am}npzKXl{f4>=d21Hm=BnvKM!;WtqrU^})?R_MJ!T|l@%J}D9
z^n-l{$MK=KLGYn;P53rpp#8Qn{3SBI4{qa`6<_~F6Z#@-uIRdl5=cLKG5f^8w;xPh
zMIcwKYKTFNr+H#6iYYIM*pJYT2&@e*Y_{iHxUd!)!f>(JW|DjhzD|oSQ`=r-582EQ
zdLwZWPXC<N0K)iDjz`FBKj>wuaTk6C9MXJk!=Znssq#a7?H#-whiPOtjjY)puvK>#
zqW(c*31rFXD&gSmyFiwvbY2iFS&iHGW*nH<hzl8Epi5@Ri&ga4jc?e2EMKwOG%uyg
zev}k%z>TJ*<>j=nxR6Hm5Fs1Ah$nV**2olfhzFJ)1|EcQv5g)@SM0vw?G$!+T{1z!
ze!=92i3(>1^MK<F_Vajn=wjf4l7qn&+s7F|aRxJ!Djy-B2QP1c_$?ao3&l%TzUbjj
z5<Dp290}w_QYH=z7WCls`PY7Ok(Qj%k)W`D3xK>_wI$fWF>}2YGWq*Kgf1&g8~mt~
zIgS#UlEq-3d<4iV7AP4kGf1L*obQ?Z+W@VAuyLeBC}wzEQ)kLK-;k7n1ilQmc>62L
zMKY@rAu|FoX!4H0Gs;B)CFflYmd5mOAJYIDWfI@t;z7h)Iiunq=lJu+5&|h5s;kJu
zTfTk^inn%D=Xi+5TXBkCgG%I}J~4RVSQF|UHG=^$v>v33KKcr=c6}ypqEUSjPaoeP
znHdZ@EySUTC~p`=EhBH=M*C2B23p6O<en>i<O`RZ78dzi`KXI@GHG+$BY!oJAMzkj
zG;+fWx%?IwJcx>XG{2AqIy{v^P4FhuyatqGW12|AU$QwanKndMD>(wZa1O@p**4$l
z?+8<<NvFdaSA>xvtMdSQ(?(ch>CYU%6o#;czhmSUObupR6PH_rvaGh~7GX>(XU51>
z3}Rw@^R8PK*>VLk%U}mKWX2h8lVmTKhiQy;-Jwn+O4S!<fC?!2;XhXQ!;yzT-7N<$
zJH?YWgqwR#4;P*#i~_624ZCl;=XBw=3|VIT2dnIFj6ev&1aTF%;(!}J^EFYB0<uIo
zLl(b?K5%YMLLgRQ<amX%Op$T+17os$FLBrfJ@JS(dbZ+QsCD<h#H_n55s%xZpo9vU
zEs66m1O<X3`{>V=W%eaIC~SUqWO|c@KZ-CT<YjqekLP3nR9rQ08<xp-<FVCY09;@T
zn7WYW%Lca`Hl7f5SWJs4ut-5Hd@pxsg0Etc!#)iXdu2yPDL<l&2nQ~BMB2Y=#TNoN
z9|bo|e%TIG{YeJ=B0qZZfb%Rv(Da4R)_$H%Am$Q#zic<cWx_hIF-4P&!H_-rc-zQ9
zZ-bJi+aSW$G*v#zsO%X^T8tmSwn>i_S3wcB*FO_9`1G`?2JQQxpY~mD?4?IB_Lg`C
z#XpE|jjK&ifiD-(#M@bW^#L@NP-aC^!Wn`Ww~~_L+bdH6;Y3(u3YsQu6}q~bKmT)>
z+o14z*HvJkpROv+T>SDPQ1M3&x^*re6{u$bPs{71dB~;a4i}sfaN>n+?+R=0r}p7a
z+TDDa)?YqN8?T<H-L1{kITHS;m5vVg(qU`Qi=Bu2yRqobPLVV5;f?&W6Z5@u#dweT
z4Q5;%Tg2GV<3q{70<H#j`QlhZO#GN>YV65e<?2FLEN}`T`6E2vF5|mD=O>c;<j~(W
zV`m!*v1`Q(ZxlC-ad`sNZ5mUV;VU5WGmmE(OA1%03j|F~=dr!yiG}kBgYt`{K*Mv%
zJ4rNNX57xs2ErY8kG+W7K5D1KPFHi>*v-`Kbkefsm`@%*N__vR)9t36o$a*0yJH1P
z>tNscz>8@tJhK~*T^v*Sn^268PfyRxU(r|zHLUS8FR5{qB?>HdiY&}c`dBjD%o{uf
zRGnd01`G0x;~2~P_$!Uc8LNB8$pl9t@;w<bS0b5Hq3i<xR%OB5fZcaK>P~T6A*eiH
z4TUi~Cr(r^vgG{Z+eJHj`)PA~&u{gkRG6KeN{e&T4#VR6&hCEdwE6x|?9B0tgkJa^
zbAA#3wd#q#t0sP><fbO4(@%f;v-FFf|I#~+tSoVPDYqrNQXn~4oL@@wi_4aq1=Y3H
zm9(<N3v<k`l*dyOX=ZZL@lC1>c>&csK_n|(`<2E&e9LZW<z8C7_b@FiuGr1IoATJ(
z+qN=k{q@VVvHmJ;ZM{i5+nZ@?^R>(kE54ro`SbMb>pxo&g`#Qw^(&V-b0@z4gkoZD
zZZWN{-cL)*t3C>odU|kh;G-MA`R42N&9g7fpZ3|#%BaiB6IQ!b2T#-0<(aYuB?*dH
zCXO6!Ib!7$?P+&s!-_E!0?buWGIqP|^h$le*I#~~UOfAAdi~<7^u-^3lfL@=H<I}e
zlKJ!W{OK3z#W!E1SI@sn8?T?Go!8%}f7r5ejk)Su&28Ua(l@jfz3RJUI#+*jxVI@D
zzfZsZAO9_V^{3ya_2*wJPu8vQnVOzYpZwq_=|?~PWm>)eQ5v0?OXtc@6g|91$=V3@
zAU+epyqQX@`LXf@Wd^%-Jb#qi^VI14TyJTER%mKI%<d=@xG1=oyOXEX4U`@11}b>(
zbWuMN^BncBX2zX)YaAapGCrCnR8OX7XZ%@=xw-kocer9bAe0I0z8M`Ku@Z1~S>?I1
zkQNs>j(;Sbon5&8wpHiZ!AQGgcM^HXyqI=R&E^|?dukKAdo3%4-}+_0IDU{9Ofhj2
zhPHvSaZd5Uuq(fo=HeHZ>KirJpI7@~U2ADs{la|grlG&+9P_>KQ^&{7tJ$eB@95<P
zho!lh^x(lg<;k#jNVX0+wh$%vWm?~oT)PL}8HwV3Vq(&1+1lFDwd1n&cLr4_P80`g
zSLEO1_=Nkc!`4xH{c1hEeDylLS>I6G=UBv+cRP9qAN?!O2NAnSX@I&=g1Is3qGa4B
zE3aLCl$YF$rC2AU-(ekQa*7>O)5f!&!3z>DXN4o(@ZlTg(8AB_C@Ks0YxE&>j3T06
zf~QaO5iiy&0MZ2weJ09${JY)LbaKX#uqSDAXFqMI-}BBpl|A`owYug_l=<n|nRl5E
zvpZFOM0TR`J5jmt4R`uSjs)dsP>!AD3B;XsA$D3$tK2Dj`Y!r5UJ@pqw4V!}tKx`K
z^gg=>XG6X-`LV;0)<k`ETAZ!1#zcD>7Ju3a&r_Z`@=|%kGm(6XwXMi+%9IO;GErHO
zhm?uw<Jkax0rJqo3CRD6i6|p>9pd-+(Zh6qbt!FcZKoG6UWzK>B}l{-?ZLLDIRMvh
zzs)Zo(_jO~zlWv{<9@=I>WH9+F8fvP3^#F+r^1BO{4r$O)uio(ikI@GAKn)RsKU#9
z7yV$L!Et;jZV-GZT@$`d7-+w341bAC?}OWTX2sWk{*@~hqGCd9H<iq-(8OUx=futQ
z&AY?mmq891nCl958Lqa_<rNV8A}k#&4VLB~Ji`Z@aGB2hx$#Fd`*TGH9rl3eZUN#6
zy5L>Y7+{p4T1?9F0TD9pVJmS(tdY6k@rg84ka=^5uF8*0MSc|Cfezjgs_?Rp0Uouh
z@D|Q8%TdV^DUGn%mIF3r4!<zX%~nxFhG&pZrY}+0Sr+I_%8gE&M@1&e3>;Tf0(I1P
ze7xZ5+NuWRbAIsgFmv3>n<6)EFVG6d!@^1r6dHQK@Rpz729X^F0}3yWAc2NhB5X|Z
ztOODMEpKM%LC^6Ld?cE;hj=TG!3GmK?>HKbH~P97@I#9nF;Sr$oIDQ*xSU?e$Djte
z08h>&t7TC9k$zDs{q8W0`!FEncLkJJDee+)kEEW+DKC_U==j6rp~<8aJh(`6hHsHp
zz`TfEV_HA-mFLDQuP>Fy42B|a=_S~9UIt;xbA@F*;h~B=;Y68pEXp-+EsE`^CR+?j
zCugS3S0NNe!po2I&>5tW7>@kodkx+_sQSai;EchgOm=^fu26Pyesk1G5B@ZXfFF}V
z2BJP%1V72aWS%;}fRUJ;Ho2+aqz9hVacIZ~CbUd8sC#5S_>y)7H-A22dwv@A&H!mK
zb79iwF(HdOj&hEjXJjz>O<2-`Vvz~8Tx6o+hd@Cc{*G|Qt6(O^Ii7S$gX$@f5J!}e
zg9vCJ@C*xWI6T4>ePz!sS!=)IBb%M!qHH5%rf?&S{vrAs`QvVc32rqMd+@<kILb!)
z6s{T^VmfW`g#f#A-d5ce7qmFC%(mN-|Jk%WlwbHZ=^`{&jz_d1f^Nj;3%qXf2vt?k
znG3WLu;OI~f=X`YE;*d*kXg~E8l~bhC%Z$|8glY{D?YFcvXIsB5W6h#gNA@y^pp65
zMMjyX*NuRd4QI_CupA&t{Lm6FPUE;`GX#)5$8Il~@z0R+hxE(z3l{u@j4<4Tr|If|
zrq6-Gl06+Y<%&DY0QZWE6K-xb4}&bt2<#D8&rWuDiY_SQZZBx;;3oe}UTE-$)J6D=
zGC3|<tGUKMV}obJ77U{dGV@}00CBgOofSmdBZThxc^MFS4T3Nw9e~|fbVCZ{E;z>%
zWCF%a#@dd_Zac<W-}O<4&Mia`^tTyc5H6sB7Yip%X)0x2<ydwXJZ970Z`SsbGZ9ab
z<2HGmjtE-Yp|xrB4anpJ<f#2_gHoG;60RRg8VZCRTGDAp4><k$72Zv`nsL<Dny%>&
zz$aS-O(|;V<r#MK3Y)J$hrM>!P~tA(xJo*q4ZefzmD}WDHZS9#fnlcQ!fIy)O?-G1
z{{9tv@$)?(+F=dQ_w#B(;aS=9RYQd53cs4JAB3&%_73;Ay-FPY_yW!ND>Gl7ANm*I
z@n&u}JWe6U9O5Mf6jfI!tTaD4LxI(`B8weVL8E?CCdXy@3=r{guV%lPN0Hx*udU<|
zJ$}#-uE-fzcH0-hpb0uJC<5>=^pNDs?H-<tNBzAqc74U-I!A6@USOVTjEbTSc^UuS
z&E2?T6I9#`bv(!(W6p-RA!sC@o)@IkFE+AJsByCw^%M@8*o`7u#@wQbvH6u1CMdt)
zd)(<-snhCmWc_8D)EtvJ9b<NOApP|3{yu&0d!KkW2#b>_q9(>hy^zjJZ_x2AG!z!I
z(_`t;gH`VWLlJ-i?WogMO0T`(%6FxZ=~82Sc4keFkE977eKw{!ExR5jB@<~XV?WJR
zG{)xJX=F9KPQbH}{#J9>*1?_^ppnH21f`vwPA{K-<6TyBb8*BSN1WY}KSwySpw4bJ
zc62Zo@|;2O^X;-rjcHMO9ULB|gVv$z+sx!xnwM<k$*JbRd^?QAZ{{qB$Q*`x%=grg
z`^lq6>G9(yX-sqE%gZ}nv}cFGr=Nb7KKtx5@1|qSOl|WVLuJLVhvU~ddW>&(Ew8So
znduoTBwxRJoemEVB<tNYr+Gi$wPM#BiZG@_%u&c2loiK(e@t_nJ3}LBYGy7yeE1}N
z{Jl@p1DTX1pGP1am<tOFX?||b-!Ph$$%zz?9zC$)gEaw;GJO8cbALOm(>Y3SUcXMS
zU%&EEiX82AaM02m{;icd_wL_M-~Yk)(~5Yp+k;&njPu^Sd71w7$KR!|**W#{dD@Ws
z)$?aozP*0+ENyRZq_%i77i1-a;}1`cTk?DEx#W`Oog5AM`0*o`!Ijz<N^ib}*4o{Y
zeNT0)mAc(l{)XV873Sxs98cO6-xDjscegjw>le?{`s)`;$2aL4<azbAaZkVeBK`6A
zzfNEN>5txJ#+nFuw79&Qo_z9I`sCA}rj`32r@Lb_>Gb^CM{IUgw@%M!-<&vt0#x%^
z$xoX@Nz1Ws(>^|uw!~bUwj6DhHtRaAeC9>w@d>_-HtBCbv$ld_!QYU|w6x_=XmMmH
zyS{j_les5;ssQu@stYL1INFq5O|x_JX<>0OEiWxvnM7OXlL)k3<e|OHFYtY?IM$D&
zI%zBPqkOue)sAl$p~xftG0B9|1mzcXonzm2_x97q#!lK-J2}62>+ggyCq}u{>8O3E
zEjj&4W28TYf?IY@b!m0Z)ZXqWpJvjW>LSPOu{({v+DF}<oqC;$H4TnYr2LoWroBV=
ztIvO%F7-TEQ2MT};#3>fq|)$c+CMx_JNs>w)4A&>?Tg)(OvSgi`Ig?6+VL5^sNx#V
zN@jM0&8E5ed6zv3d%n9D-v`t)NP(dx9h3|G2K^9yc<?7(;24dnx1&MoV$K+5)D5ql
zsIL2D9WJGhJf=^kZ-OcP8MNh;B>K*>hQpc;;Zb0+&f&VHGKKCf>nN<TDXuuJDa%0x
z5m4vYeKk8fBm7m`+i#_ITluWV*0K7IoxN7t+}=&=>sx70WqYu%Homv*w9+R~rt~j>
zzxOU2d7{UP`*g~V`Zlh5#@b@6?eTn8yH@{8*>hwiYauAJ=>yM{R?3LFqOA2d9+wy9
z%-c$3@kRL+Ylq^-&_GEdoqkRz?V<D&4QbUMHz}qpJmw-F?2dd0P}#d3#jY{h6FUZd
zWT@<fiFLJztAylb3fd~YcbnrLd7i4jc>c{d%CiGMXJhTDK-3r8n&tppzr7z<%&=$w
z;>W*<w>@yaGQJ6Aov=w;;5NG)WN?j>S@~u=dDetM{0j`i0Zn`}{JZD}`wWibLve%P
zL+P6EZNfnNZDaULWO^Um#xoRO|M`E{gdPZCeHBZ=G2&J84?enO0t^p;je-_Srr}OU
zY8#lSq!mIItRb8(#X(~%_-qX?{fNgt7R4B-GT5`C%^2T<4iD`5nJ#=m!#F4c7n8ff
zyFush0b%0Kyqg?MTU+eoTaMcRZFXb>rI9BqMik%rUcyE`)G<cUkk42O1W}b|jjNWq
z7b?69G$_ovksm#A7+q*(=KL`x&;f+kpM_c!T?}BBmRB^WjknMo&K*AGpF8+$@`MP5
zf*!DXI@$C<O%pRclx)`^g@<q`65_$>onU%!@vw~#lA<VM0+l>KU}D6>!4DY4&%m39
z25I$?8u_R?a@9HJCl<2E5dV01$F}TC#qD#qHTdxHdJ5-!aI(1}6&c_Z{pA4DzYr`>
zSbKr&$0S`cuv4U#R7rM|&4Ed~+b|Dl=#?I~`J9J5(Oqg(NWy#zfSiHvkOm73o_Ls3
zr->sJm$9qQ3QkQ<d1I1+ECHizAXhmmtho3{6wYyHrxrA0nrJO>&b0=fAUINyK^gVY
z=cqIAl8<*6Djtq2WY-uUFhAvpgi{T0g`=Lt&bJsq0v?>vZq<I-QcnHxxy~{fOud!7
z)N=-HF#%vOLpY0~Y*Topkq=vYLZLK}2Y8b(w>giI7#t3zGs)69(uC!lL75<zQAk61
z=&r#y#mB_L$}QfmV}M0E<?lfn=_kL)XZ<|T1n)9+tRb85D7hKzEiW#oDNS-p9f(*7
zo#m_r5n4P8&fr#HOsQk`X$G@r`1_`OwUK`nn5K5S55%bpYvY!~<+w9lV6r=I&MyQ0
z&V%qOv{)>c@;1^NblErG0el4vYz>88rpO87e9=Eb$FBf5z!(W&@fuj1;mRosCK(TX
zIB-W;;Vq|pGa}sh;}##cUCKH0bojcwOf97Ccm_Kh4ugjm?!?zrrVU=<iBX#_J2?TP
za5%fNpMYl}L4qRRUC@}w99E|vw<(CfhER$M@4sWoc8eiXKiu}<Mq3C@;pPgOu!m`<
z({6-779s5fZJr8O5cwXQkt6#dpJgPFok8LB1U=No+mFT`V7kf{v{f3x5y%(;H;<XH
zXZf*0CO_PeID``$G=<V!>}jVV`1+p1G&}{Cn_yXv+)V{l3Flw%iu46U9SE0G6&Wd8
zPJbr;&W}2j+$D!^fJM?x-jQ~vD*u&}4u)sW)tm=5v(<Q;bxBr$XCJB;#J4It{cKmT
z0cNqACkAf*O?QT%w|Cqm1BUBFPD_7$y>?p#weZE?Q0&kKR~-H~lUwXI36t~Baq!E}
z<3SK%_!s^Hc<3sAnIQX%5`S#)&sVs)q8`G}=rRU^fuY!Y!}fqsy|Rc0<6(z^AD3Om
z1s>vf<l`v3d1jXjo`flWaS?zZ{rK=GV>q0(J-ZvcWDV0{YJ3?p^y2LTVSA0o!M}D_
z7kyN=_ws7s^>?j<#O~g!)X4W<6+JF+l~1_j2T4^UlrP5R5CN1#onOXhFPv0K@R%O=
zseHN2%QO&|_8LLC9A|Nu2FQqlYg}1E+rzu~Au}Ul8UA77MSS^^+tE2l!Gi?PnBc9I
zLhJ%!Hyh({s;L)H^>f^=Gj$zm_Hj9>rlg+4sTlJ0b3y*<80jm05`xB?9)pX9u_|L`
zFFW#*r6MgaZ~0q2qUD=YGcz+Dm!D`J$PNe=q`l*6bVB3sD<6Bv`1_-eo}`Bl9;MaQ
z)%5V;qx9hMBOlXu#|xOmfjn1fc4i_ydAOFA7U!&-VT_Cdj5zsbE{pt=d@opYKGLjI
zTTw8=`2G4aO^)96H<?g?kk&KJ1$gtHquWRi8OHdVjAHI`#xa=3$37O5xhzMv0eqM0
z>DON+j>^2eICGx!-5|avH9g5SZsivWFMpFsvQBAk%{4YQ;XGsx#Bq(}Jv**Y7O=Q$
zzdUa?XiUp4lS|EM`JJAfrXT+B$LU8u`B_?6TuqZxvzn8Pr*^07#axblL?PfE9TKeD
z?f6?+te1K>gvtr!$F$~K;}g>gcbwMOH`ALp>)v%|r5p2P1t@DCSC<-t%I<jyZN;YM
zg6tw1nUFay`jIrOa-CaRP4^yrobEr=_2`rIK-b!XCuvTy&de>P#pV0LeeB(0!<@u1
zGUnZ4rzfqnzqjk76JI=knw~!WGChC(b$a#cn{=!+%t_vlo;<Mv?9s!=X>nn}bHN=e
zsrZK9Zra{tciEe?wf-t?y!j@*Bs|~WlY2+_r(galJ^kwQw7tEa+O31Mr!a@gn~uuy
z__(8Sfa<sM{ls~ox~dxFenLO)tVwvhV<#S1uKXP``1uG!@jhU8+tzm4-QMv|N76{z
zSClVnkG`MQ9)3T4^u3>?#nne?e0olGcO)Gv54v5B%j`zoV6LmSNBf{GImV6ejV-J0
zuPo1{rG+`~P#U2uBnCTStN=h^q4F)e5ow3)q(f<iLK;O8bsZ%WZIgK~M|O@wH#uQ?
zcF^#>Xp~vhrP;Zew6L_0R#umNBG3{FulYG2?>8Yj@0wJ-;wZc2g?aB%xxcoY9_U(<
zo1HnQrx#Z09=4C&?&yCAV|nDyi_IuMP<ZX`?0KP|Z|JdZK)azmq10k7LtEo}Qlx|3
zF&w3LcFK4B)K*j{4i%0g5Qm3(X?jfk(vVMK;V4p+<tV<!M((89sR{46T%1!H)o#13
zo%FkZ{qO1dH_y_j=G1rYPNvZ*<@egdG%`M&w)Wd;edjQpt4)qgOsRjHOY=$}dEQYS
z<>*h=y;#E{AK6uC<s*5dGNupN-``6+su$2xzO+Y_=O~@&FI`?Lr!m@=%8@c*eotQc
zTj8#J>Wk^B-Cxn)@)B~qa2|^C?5T?veEA$ieCsWaeT9aq8_&C-q;HR%g=fOWCoZf2
zSN-sMp83ZM#e|449sJ>^=Yztdl;M*VyqLf#OYF+yn8{P6o1I%IJ=@*3m0~-)2Wew-
zH%+L{K7Mr1F9!H~{_J#Af5&$b)#W8t`c5z2=3_rma;a~({8!!uh77C=vF^yr^VBmE
z#S0+uEK=X4d=`=NkTp*F$2elv-_PSQsy^S(N7iD5S3=AKT&&Ru=Lwm-WZ*7i1!!PT
z-ILXJ9wC^tEneE^Xh)QLaRoohz%zSj$w$gfKNNf9A>XXysK>p1b_wq~?|aA5Az!(S
zK^D71uKJ1xz3sH2Y%jQ`SpX-L#y7mo!~C+JOp;$n`uX*nK#e0mLssX4(cIwOpySGR
z0DE<1cO^iEpdTEjK(Y6m{orufUcfl$`#gfj4eADd(A5ut8_X-9#2+-@5(t;A?}Cc|
zUl7HIz#!SHqYlGM0^~}sYT^cniwZ{~uz~k5Mb4lUEj5X7YK2tQD|tG+2{6Myv&eBo
zroh|Zs<NjbHy8)wCIaGPIjVRBTrz{U0095=Nkl<ZE!c3|0#0%`6A8mjfJ7EecKVeF
zV=kgqBTyLg&OXt+8j!<6g^h5gP&^@527;SKwjhUXqP87Zky&8Rl9xdlbaYGPN02x%
zw&1PUBJvz>5QRkcFhs+)mu}*P*ANXJOT<Tq>F?!K5P?il;)HkR75L(YM<qer;3*^W
z8bt6YDIz`|3_O^5JBBxT7-S&}VO(G2?+Jl95k{BXe&dgaqB?EPBlm+<9X1Q~C>Q*&
z<gEogUARzk@$j#<@bh5O#m7?Fvd@5li2)NR;^Sdz*_9*)4-8^>$ev%GIo?PqvWZR)
zVBYRP`Q>x4HCf~BmN7j<7)bD;V`Q=H4n<z*smGJA0x|~t2Lj|6z8Ujg2$|mlnIFZ?
zyb?PyS8-rQBJxF=WDza0l!u82BY4Zrn@*lUYS0#eJ{)*+u$*bnq=&b{eEgsahWzHO
z%+t<^%Xfr<o@jXhyA7!lArHq7V)uuzPfpTX4Q?1bjBD^Yx4e`lre}QIAA@nqnt`s5
zSVRHFfJAbP<hqR#lgZ(wF6s|k)2%3CkV1G(I2j;Wp=2HmHsZ}i4}$U=pIm;!O#>1x
zzvU?8mGX+gB6XEK^d!%3?65OTJXH6hPRGQAwnLUOap9<vx0=)kVHi<vv^eDnk&-h?
zisM5@enL}*DCc-v$Zx;#%WVW!9)J;CZU)WNU3iX<G2vKNo=wR!m%V<}AKEYd3q{YQ
z!F^Dc3xDTBMp;2Zg{yFdu<*<{=u8oJhvpB%j44(p7q9@Cx|q61H-zvHCV{tXIl0wc
zMu`y!Hpqf}SP4}7LrQs<JTOnu^NapXWldUSI(>xa5|!z&!lVHFWSA_wV{stD)M2x~
zY<#CI@{B+6;A5NYCK6eW%lsO0;IAvXC<x@FA1>kOBeTQ7;X`WO#t9@-|B#^2kz2&)
z5D|}_Hz9SPEFw4YA*6A5gEv&v_A_2)HgqWz0fY#gx%i4_Iv}j#h&*s^xPJ&qjf^bd
z`3Od+@RiFfa1~tfk1!DfIQ(Q&R+Q|_5ekSq^T?=JV2jW|gzmdU(u^x?mBhZAhMxsR
z-pq-c0y{o%GE4o(&)(sTGL$E7?xZn$$&T!H8p1zA<_j)pY|FgmR|sXXzi!Z1{`f@(
zS5QSIwAVR}Kq6WE=jMCZ1^=_Y@zoEdZ5MCbZR!Tj8ImW*{Bs7GDy%mk>P??#b{76M
z(9cACh4&!n4@W*$@es@m>|wK}&*{r%peft%4_s|8X0Zot;EHs$pP=9~jluS?VyUgg
zzsOR2vR{tCeyT^VgVga%SBDA0%B?G8z-=#5+Z&c&v73&|kyt*$*Z9oDJ_T3%+sAgE
zEA_Ka7C%dZCY#`N)#11zOh#2u=(uuL2nvpHy#Wesxth@LXKyuR=ZUPu)r2N(9Vb^w
zBYA~;5bs`I2=5msoBokO*<JH%UZyLgePBO9)35L@SMd$X+8+eCCav*~^JcmNvYF|W
zW7*7fQFWrsXu#A}rz_lrm-$ugy720SDoxN<GzzP``SUBh8elhg#(jP^m2l<9)1I-c
zV$dKVzWo*7#5G9HW}S=0y1>!LmvLB;kh^JdK!*u`rLWAz2;-l|(;SiJ7siw?*!`W}
zyJH$RUwT2*N7v2HdQp>wJI@a_W*HhD^+NRK=8nJd#O|9JzIipx4h~+N8*`SjE2MSU
z@kjspW)5%l^X(rN^_dG!PEK3#Gd4aM(;D)FveA5@nmAc%<|w?7#@IZ6$7f<p`7`Vz
zwU|qI*M-N|chbdM`lX>XgEC5Sot&~8q3a{dm?x|(uP_OT<NP+>c<#xL6?P*{Av=qh
z%CB+7!Mqk;>?*TT!3wSk@7Q9F!n}lX)Kow`Py}e6B>5<Bo)@I&N{K^G^6==W<Kx2a
zt=&%#A3aGAbj>fWD9or81zX$O>8N!`GMr9!7<D>_lAj|wHE&iuAxBVJP0z3+YdP^<
zZxjdYwtDmCO?vg}RodR!PJDmK3*DMaqf}wn&f>ycnwyQCZsZqpd6YKo?n&yXyv{B(
zcT)cG4J2mhSDM@NoiKLau;c9PQn3p+uCiNNSxd{SYns1I2*<8Ll#QI~a+3J&-cf5m
z?QU<S_4U_jLvroy@1?!nJ<0dd$0M>+j3WwB22v-iXw`V}oFhmtPF>!|?N-{|*;M#f
zO52OHr*xgOOVai6%-_@5+1W}v%D4SJj!x`&-NJQ8ndqknnQ40lyJfmhpgnQ)J#OWd
z(l9zPZDrs6M<1mJk3UIIKKY^CAEo7ck5pD`s_V;XYIY$_NM70q$1u`vy<T$3QT!;e
z*x5!IO*!BFU8=REw7k5K=4Mc;#3K4#)sHx~k(Y)!3NCg=aU?%;+jyyxwm`lkq^iB@
z912-zy(4LwU0ieCp|`lWm=>27{0%tFRgS}4UA3F_A$CV0FX`q5V~+4cp-tOnUW#&x
zZ#Pk^$z#@C4pe9Pc2P?vN7^x;MHuQQ`Nj*Q?10=+*|rbcs&{+_W7wy_a4aFa$51M>
zBXm}MM0{_nqx!n(Z{8ir&ODBJ^M%?m#}A&Kst;0LqR8QxL=;|pH;evqZfYzo&rhZW
zlpyR98o879x7O3||LtE?KQ`0-2M^P|haaW+<p*i?{^K+@IVagq()K}H-Aa58<o^Bp
zJ}Q*F<LFW95$gpg3VDH>G9*t>_!OlR3f!ZkL-iwFmo@prPFa*w)Fr=MpdaN%xzG=~
z43!u1be|IIaH?jwWd~<B?V3)r4=8fh!o2n-8Qce7T$;XIQFjXus&^=SyoRE5AS*Jm
z_7+Mr`q-#U@OB#9o<w)~R{dD)a*O_fdL1$n2YnbXuOFX^UiDyQX)b;I(IeMC*6L_W
ztmB~o>vT9S@!a1~+-n`B?cH4~sY%25*j?rSvcDa7cbGo%G94bZef(s*dmL+viU8I?
zDZkHI;=ENJ>zQXIt?IsC49IN-nn*Wo1%*BeQxul$L=43)`M`5CcJb=jcbDg%@burp
zJ9&OaJDHB3RI`R<K5C<WMyO2TNx$xADsj6USx@Cf2wtM#IXdd+(y{6m%2fA6!9K?q
z&-2=iC7xf!U5|aTTXB~0fZf>ydT~MHpG-eW!xi`jjeSg8!ph8Y0Zp0$#qM#40E5eY
ztz;Y&#soHUT=@b@U2H<6y@noM8TxU9Xn+!D6CCFN^8R=ONK=%>;ENj5gns&de_?Nt
z^+RA#dIrI5Y5ZHkpfp(V^-urlzx=8UGy|yqWC9t3A(P1t^7%#s!Z(15<?!~$I|>DU
zg{A=h@CF|llPSbA1Z;+oJ^l_6jj5m&2PRkU2r^;tF^U0ri7Seg_`q;CQRW%4fKnLa
z`R8Be7x-{HPW>}V2I0b8S@?6AHZBaCS>|X<AyI^5d0JpQIDEpdHmeYMo{}rCnFB_K
z;rt-S6+*-rmG8m3$`1z#8k?nYIUJs%QrQwsp{Y9%mq=jow@9MbMH(Ha_@FFc(8C~@
zZS9^UkqGz9CNgi_D7O5N<iSaYtpNsS>GHOe&(zgJhPN>Mz?4ZCpR20@00SjHJPBJX
zBQzjZM4BXUKJSE;3{{*&h+P&2c^(KzF5bvuJ3K_ZWr8f;$>V{99yg>-bRJ&upktSk
za46&afoMbIOXQKA%A0to1vt^~h8z@VSl<i%oO5aRLe2ufbrJc(86G#p%K?-Iw@0Ui
zWoCs^9#Ut5P1Lz}LXOYi;0mGeynRLfSdpXwz@_q%i#PAeAP!-NL<cZ1V9-Ex3@DI;
zNe2T&-Xa{<pmRop>$#;RPXOSL!iTrm7!+7(D&8Dj!r;<pL96WfSnP3^<7qUJl|QmD
zfH5!-blulDTz0~HV1bw7_uD}X&{T&#!O(z&^SOOiy5eV&Kv?fe6df7hw?T8B`NO`>
z8@bg!m{{;ZU)lt9@>qklw@mH|<a{Jdv~$TKZ_YrEiAlsqB_I?7Nl$#$CKybXfv%%d
zjgE1y>b>sn2d*`+8lCV_A{=uP^;ZmuON`yeI4peP9wxHz!&M*)3&0{<ET4JUrCWr{
z=nTac0i9mCB5u&xB81&|2xx`Pe!NI%Vx<9y`tKWLwqpiu*Z|*k3v$8-&Ul--57P|}
zj@EoLI4m;Aid}xRB@cSlmK+ezKHfPgh8#n_t2~8?2?>}izwWM!8DHY6+NcGSU3exq
zmI#~-ry=KyLqm#bMBqcy)5yz$&){%`a6r+yj#Si@8B7HgB4|o~Wtj_+nUebKn*|GB
z?!knkk-|@IP)27rHpVEBbICIsIkX^$k9?3DXHZo%;S!LsnZ{JmDm<aHc*YZpe?*;`
zI2zmHZZ^7=`jW{E9%mJdj2N?Jj2Jl;wRiaF?>Grt<jerR2tsTjA9e>alsXd*{p-x2
zMLB(Tm<<G&V1gI7=s#@L6$D{7MFynBF9?et@dpBTP&mzy3H9$q@ddbUhrj~>JoO>c
z<@_o3Y!+y``{T1;UlcV@AUtbaHv8fLnmFv1T`168{b_NF>j_g#0$>R@0Iu+_gVeGW
z&V>f(4TIU67Fnt&Oq=aMe^}cIE+eB26+YYRtBJ!FWyMvX-#*9<2D6#I<|RyZ1rDkj
zz|qbJ$ymU4wW*vtzUPSROW@?oUSCEFnz*<MevscFe>cDaH|R3R-Jq(WqV46&Rru8Y
zy;sveXoT&-85Pt$_<Rk(7y1IA`u*PpH7%|>#Z4alS4Bpc-0tke4|{eup@&Dt73lFB
za0LywxNdY)nc<DF4dNhSJi+{ig+^WsJoh||l8(8y#^z8mC!mhHo<=<g6C0c&%@Iz{
z45iO8h4V%2!E*tWUyP52z39u4UMNYt$f|KZiXvWUWU<`4tj>@NG47@-$-s^)zMX?&
ziG^u)bBy}CTrme=QTw3PN{2F0uAH25T=`osY@<v;(Z)B27)LW+XAFKQeusyLsl%eP
z#{Vp^CUyWw?vvA#baZ&2xx<Oy{zn;wqK3s*7Avi^@eTuC&b&^env-&D-;`gTJk<c^
zGOaEzr6-Rbqy>&TRN5GO`=!9VaEkKi>o5Q07XncbF$ZEkLB4T(W4y!}ii-)&yT``v
z>T$35RbJQP;cqsv0BbutJ;*y`WOvwPY9AeW0Uc!oa~Brl$EW7f<n)YpJhTp5lHpZ)
z_3DM^-+VWVa`3mo@-BkCot@Mc9XlmZm^^&&AU%BeP<b*V+@*J49Lj%V<5ha|`gz*k
ze3QmU?xctL*gC$GM*W~Z@Zu{+A(BQP&8Pfh#|Lw9k1<p~crmf0o4RF1g_Wfoy?8G@
zc=$N2J$RUwSJspt<DR3%Isr#lo~PEqL3;iAMf&F1*WMAvF^7A5J60wzFXB5!iwlch
zuqMskF_-Hs6^=z^E-MO*R-GLlmBII<NC-zT#!H4)obpuV!OD696ih-$ELVZdo(1#a
zk?}OY@*sWk>Ce(<Kl=~qv!DJlJ^tiJO4}pNZRcE%XctWFXq&Xhnb{eY;hgei)C=nP
zlSk94lQYvCzc-!c=lFh+>fqc|nv<IyA;`dsnUvpQ>p=6*Yn2J@ZcOQ>zJ~%Wc5o^g
zDg*iq=I5Bq5m|g^fgQySZD2+wJ9W5bW~LNxC^uPiq3&F$EZB)g`#nY}(LGk%?4}mS
zjnX!c+Ns^)60X%r`>l4`Jvd6+dng|FRVNQqmpQx2m3o097+F}{KRi@8@qj<`Y2=w#
zzAP=ws%=T0>G3p~6$hkmdwVxMfBq_M$z(?n>kKGBP!=&?W@ndA22wnc=6K0;YW!}x
zw=|s|uP&t3nXxoEs=`n{t-pAl{`J5AUy^et{rD$8OF#SjU#e`PP2ascn!0Cd4=3u&
zG|ygMT1xlT1~{4(I8?pd-`!QaJM&SJ^mU|<^$FS=;N7W8FG?W#MXyomvLcD{IzMy&
zcho*qyKdQTg`)HO)Vx_Yp-inDQ@!>J-U?41(*M$bV;=;8kG|gPbqeSD;rQ8E$T}W%
zC3vxx!b`rC1?{(1deI)E(!-Mrw}0>zFTSzJOA4%+u%1WTf(P}8Z@1AV>F0=_ILF3E
zl&8yXx2(_c92xgI#$CmC<?=l{Jx}Lq)9m^>;0Q%M2WWFslWG^Uw2N_-Z5)BPcW|iw
zO@04yyyPB`O&O&BT<yaz-m6aggdp*XU2`bQM5J`FHbb9}GM#4{yK$*!C6AHQyLtWm
zW4%V~S*yu$lA9&JSch|euQc)^1;_Hz*6Wjr;1_d$uS+Q&o-1&nZlF;2(>lnr43Yn}
z+ZJE__@-H|;trFJF3Jfrlzfs0`B+OUaoHa2Ht1_U#x&f6%mRb(*rOkIY!$9RGfa&Q
zXwm_Zj$Y`e0d9u7NdtUHoCOBs75*0bfT%MCN;%Zf>kqt}-@9Q@dIrH?6UQKMywoe3
zgL&tp4*&UAOr)aA5{MQX`GV*KB|J7dpQa7_wfssJvV5jTpZp>mei61{GSYQM7z<}x
z@KfuAD(-<V35b{p4mM0TWBwUuIuF1Jlj-fk9<IRRa%cht3afrz+{yfWv$ybWkU6~L
z2yY-$hgbEh5S0v(A0%EF5cXDgQ}WulB7cd)fpW}tRs4LpkiixD9A5UI5D;;NA$@^!
ze0In#xBEyT%P-`N{2*8^HUoL>-c+XXnLZ5NOeZezgJzhskZ1jN4{!A_NG-hJ;|^Wn
zxl@C9aPt7+EeqTAQ1X^%J<Kk7p!z|{+aC-*z+L#Q9^T9<VS^_Rq8J>HPq5o-Oz|pW
z=Lvv(qSK#;1Rsh-F~&n5B^uxSVzN`t9OmdcPaX&<`U}}n$na3}z(<2R9yn(j7{RtI
zWHcX;5p!{WI5hMBHo&`|a|u%6Gos)TkqfYx*xw<Z(<CA<*fSs^uXTB7q&%l>8gAhc
zwJXyAvJ#&B98uo!wiOR>-gxoAQhD#)b>y|&ZsRH^XrM>IgDjNsl?EETb$BT{lw&Bn
zW;iloe%?ndz>|S}*-hs5p!#9Oss@4#1_^&=<=L6>_*1r4#*M@qIF3)ElSJwo4{_c|
zMH%RNqB_A~0y~4-3EslgAd0Q%WDkQ62B_3C_*F_ZaV#GzB`guJ+l*GmpzW9sx*oIW
z(8yc_DcPyKiM*to8gC7${Xs+>8d4hEe#IQQ9hSVnN|XxE&wUl;sS(zwCS8k~eDek@
z`A{iFY#@v12?t+?hpp_$1IqqFiluDM1L%TEzOtHC1PQ>7k==yT&0!r)pbgO3O=SGV
zFiI!`^bnAkaON@;9cE5DjHCXVI4DVvQGs_n8JyRJ99kSsOz;lRptD(ig^9xwCnoqz
zXR{Ew7#z+~7~=rJCudNdVk5&ONF=O;P?!=n!r@^GMN;x0D3n)DkNzO=@-4^ffWa^5
z!Eg^-4k8d=_r=66Jb@h`(p~BjfWJNCPa&CG#24X^(cwS=$WR<vWB{C>X`&=%ho=JG
zp)Ut7s4O*9sgYkKzfNC?waAm%%U#kNF$Fyyq!C_$0fMLOptuSin{mM_XMKRg4qD%;
z`0cg<N5z=saQ`7&+<{syuHaiR;TI6Na7TFCMdrF8Kt_D37>L{cB0y;5=W^!fKn3Ik
zL%4njYWX;E!Nd7sR~-fu{I9=b#aHBMqz8Kw`rW-im<ac6U|?LC&5ihdzdepZ-wfBo
zqnBN{#g6BD0h2U?<O<<_q!D%1uLz1g;;QY10{bADFbcX!(@)#)E>QcaMm6XvdqZAf
zh09jki@VrM*a%;s_|>>NedTIK)5Pn416T4^E<Aiy;j`Oxb_S2Q3LSohF7SmOd~v@I
z-}JlbALxos<y%o#G#S%yS9mUa-Sp!M80^Ot3571Z4dJ#;Mon*21%a#UGf>k8J!bat
z)$3nh0YMX&?)Gbzc!8S6P)LM=Rn70afo3}18F!IQm*-QSN70^?1m<PTy)=g9D89Fs
zx){%v@fCDA&F%txX}rle)A45B0`5V=JTd%PTz2?O3q)UX$4V3wV9>GfdwhIsKaWfG
zE4x~FyB}r1+p9b2R5+A5EK;+}g=Y-M1#;B67Z1hxSaZkj@o8#z*ag;22d$&DyML5A
z9nArm2Z%S1H2AaV+-e<qVLra&qPd42N-QWMF#afYm_P76>q*TEQ7(}d7N`CF;Iqrr
z>7IBYpE3HvoaQN-uXs+Nd|-zS-><s2b}!8@E^6*NX&J~HUeIf`4${u<j*m8DUNpv>
zh6p%qA}F0s=A)c$QRdu)cTS4dr(I}%!h*gRUWa&5F1~@rE;M#F#qNNybbc9YJABJ%
zUUMpxKbspH>6>q!rS<i7=k3CL?Bt;=C==!^e1C0gXV*J!P`W&P{4hOv^4Q<{X1?t5
z7B6IGx7XV0nq?s$s87lY(L}rTqPZ@VoTGf#O!Fq@izqBmGA=F7`$E~r_sUrGC$CY$
z^3AY&_aCLtKKoJn<dg3!Z`ozEXhqoU+)^B;s5z5QFJYc0-ndci^3Ad9B)h5s$CsfT
zQF4Wkxf`c`3=MNy$ygfUJL1#qX1bTw?mtQ2S9~9T@3Zvq(I;tX<w2UBTTF9{%V~Cg
zDRC4em5SkUDK*-OO`NJz>^6M#_~Z2BpZ-Jo`7i%p>Bm3&MSAq)dkVj-G97hUqFn5B
zT503Wi?p-#CY`J7*e%4)AznD<2)4-y6kah0MG3~vM0UOH?QW)hU88r0e7qxRroEq@
zo~Ql2eU~xEHL{T3Iga{D{bO!Uxlq;|lZWX~0jNBgSE9_~CDSs7U~Wx&<!DU4%hx_q
z`#C(c0vSazI}JI}1Vuzg?W!Z1j_BFJ#S5x*tyTi4eyRx4nQkbHXhZBoWakrYlQ|yg
z<%Lt^vvLPTiNc|9LxHz2Kd(AGtM;a{8XpyXtofi|+uq(!Tibh9gflTjIY+(_hI+>L
z!%+(2$L>k;basYM0nDcR%X8`R%6wXxQ(by{nszqU)0cn#GX3tizwt4fKl<@c(?9&;
zmukxs-dRX`QJi;`X6E3;#c_@(tohd8=H{l4Eu@d)h*gx(tY^{Zp$z7mSA0tkWjE^<
zC~4_)=%>odx|9jK<Y=EL)>$KCXDLT-vQx}2iYtt-3zWCD@(Gjjq$*M;Y#9Kip&hGS
zSOdcDs;nQqhS!BMp87!9(C0;9`qWU8QFcJZBSUQ!AHPI^tc1Dp%LItVaer3Ms@}2F
z@=R?4`Jq8UMjhSVRej#tO{eE?RX^D+8{g}SBXT*;b5!ziOl`=-`s3Kxu#W(xKjwvf
zKBd9&tsIw#Sl6uc$Ru6xqK`a9d8@j6bi`4N@gfG#2fp#hOAp8&&(v6hL2=1?me-bY
zdnr7To9BTe15dw*X-6o(xSU6pL3LkYP})&nX@i;6aWPlrc|jk^GY|zi#~Je)f63=k
zPDEPDA<8slYvL-q>R7D9WhR2~k1zmsU$(<faus@L`H7D>g?|R~$!2kX8xz1z*~sE}
z3a)7a*lSln(;w(}10RALL<4~D_jwoXO;Gq-=mT6^3_(|yNkHunyqnkif$$%~;jfLS
zAA;YzdB%>ffBLVQ&^IcdASy4mm{1iTOfG0k!h$Sp+&StHGpe>gPb*h8bD}as4h_6{
z)XM~CS~W_ZZn#Y@UueVAd~D6Gf(thVSt#%WU=*^{n9QLv!Y4w8FSx)K4A!71si_zk
zS11ElrKj@Aa+zu2Q0RrM@*~IL@c5xh2wssNx`HC(@sXL?6y6!DU$`O<WtWKp&-oFV
zYJZzy2#`mH-EaT}F6{P#m_Mfc3J<_>*eR?6I4$bTQ8dhJP|RSLK{hzT((!vBt|vo0
zY<0=aLz;_+N+>*d$YrHh!$k3g!U81~0~{Uz-q9uSaff=K@eqwI_dMhn3~?kKd6Mg6
zmQzUn9u#tv9?)(d>1yjir3Zr^)Z{R6^YA4wZ*+uW4#m%~(m3zIJ%bi+eHQpl7ak%c
zPY~(+3!%>H4}wDO<b}1DNA{>hD*Yk&5~$dV8;de6s-x^mpRzc>c_>GPcn`o65HltZ
zBg#7l7k=2Qj#$y9ANk89#)1;J%5Z{@Bx};fIsFXqs6%C&J!Bk-=$%#?08Gx#nw|mi
z1m}~>kB?LPp{zP~b&-`{3d>;7gFgmTil5Cb-qkl6M>n8M@|!fI9l)ycMPAn_Mb1tz
z>MeDTfe7_!LK9c)1Yp-9b(f0gkNj$I;;;->#hWw}x7iENkAW6%Q=XmhMw=2?vY0^;
zLV%M&+YY6Ues}o>LA)_b9Y;ncjFwm7NE2-q<(hX#Qdh-?l(|jHPF)`v9`z(~YHG&(
z9CZ+J6OSSclH-AQqyND0J)lrfcc2r4xCC$rzY-$yBU^=T`0I+kEXV3wfT>8d8#u{=
zOY~QipZl*0qThnl@(@@M)ZlPLXds#i0c0qk!)7~qL2&M#ERc^Ffs`QOE<k#s|BEnH
zYBNn3_RO}BA!q_0n1JxllJj@{A^o%;{Yj*YSv2<7P4P<thv5lfnee0T^TQ2S-*i<V
z&bjQN7iomgI9%nK&dq#mw=2+MSg7Go`T_I7&R;xounH^-AP?llzkaU8g=|3<d_bBm
z2Sx&?)2;w;)g;ftV#hQ!SrrhOhb8dlX`8*d2PW9qv|V>D92p0|1vCk>QE-j|Kgy-_
z3BKEI-rOR9+IHZfY4R~--T*ke%nbQ64jwj*FeOmZP~~Z61g(G3C`?10icpy!v=I2O
zbX3&_D&h(H5;okpY~^ly)VFYb*E_y2>)aa*wam3yza&9>3w#p>;Q;Xc@tL+y1fyK{
zg1^Jr-OIb!o45kCY<0L&j~Fi&a7v<*+GQNH4W7Mr?oC-5{B@aAmbIABV26AVUxzKE
zg2Joc4m7NN_@MLElMY*rtLejD`Pp9c0tznh#jMMtT(#IWu5ud}yfUhysp$scYInIB
z+0%n>!fkk<>DTLD$LTwM&}12!<qN(az!X0y`zi2iUj5MTU-xymo2dqRX`_18@_D=k
zzRE32hiA`~L-byKG@ycx{;Br@FYgL!nQ%4bsjqj!x1s$2cqb3Mbd|UmCl2`$MBSE!
zML)lQ$oHf;=I>Nv>64V?zQaPM=;;4~UUw&#@mGwM?K1CD7Ro{SG1tP)bC|I>3)+nB
zPdTRU;>=@a<_MG>yB9j0w*7cMGS+2$%vgG6cFuDE79v@!_pn*l)Wnn(LNA{`_ihRn
zwqxOuZ+gA;POhWQiH|{Kw+Kg@Asa^pvS7$Jyy&Bj#g`pr%qvmWuz<-?i|n3afs>sH
zC~77(=e?`+u%OMNBtpb-{1@Ima&jEUEY41&2pLXo6ndx1Q^o1USmv}IvpUDTV_{})
zUimPp@F>AriSLDtvEZ+Cb37dj{mg|uS5QEX*BaM+X@tBb>xE`+O1fAawL+1_Ql(*P
zdLqruPpdpHd;$jPW%m%}LHSH;9<Z{qmKK+mJRg4b`el0l&GXbgJoFBR`}ZDNxidO8
zovtM#3&|)IcK3G8k2si5EH1C4g(Y3fYiV`up}#3Ms(DgboTYjDr9<TzN;&3vtR-;N
z87F9<v_qk1MTgQir8LY<Pp18y&GhHbf0s7bU!>#ikrkyg)AQoZ;`l<ESDYxp9zXeB
z`skA%qz8|`r}&?EN7eG$M`?*`^-)@0eVA5dE-c-TT~%Wol{xM2FD;74_kQqc`sDkc
zrH4;G5w9PJ*JtTRKmLdG<DdQ_efp!nOP_rDll0_!KTL~jkJH5TQo1`flSU`z()iRu
znq64)x1E-C%`L1df0x8pX?KHD9oBr8Z+CtAqo1Y!&;R$I($D|H|C;VgW|XRLUVW3c
zH(#gCH{Yb^&;Bg;v((z#Qd_z3?val_{$6_W@yFhw$Gpx*7Ahb2c6ZXo#v6Yt?Ck6~
zot|{9z((P+y0V%U7ndA2ioMpsQQF$vN;^9{{vOt0>(I)ev(pPJ8>myXwNO6BG5*vw
zl%3QGlqe`@*oD>6WhIyDDZ49M2QBg3x55i$^Qr1E^IO<Po2PFH{?RU!-wX3;VJ`2C
zQ5~S~_xb>Hbn1!b*7&irY<5;<r20UQ!cIpNLwvVyWogkny{PBp2|Q3jaNHefJ3Q>B
zF0x#rJR9{(+DqyimKW#K)XbFA3_b0UJfpqNs=h7EOr+(x$+SExoa$n?y`TR0hu^0E
z^S}L{>GLoC>|LHe{@KsdKm76^)7sja{9Ea8f6vE(vLleS50r3Ix=^05tCajf`NtXu
zM?_9cPC9?t!G_X^K7}x>gB=~Apl^E}3?(u^`)8d4rMLSw<&oFSyc3BR#`*p@{Y2yg
zCl%S>>ut0><{L2<qqA_Iq4bw$K|e$vUO!Vs6?Xazn_SV~Ab*?|6bdHda#;PsUOywK
z14Ng_Fy6bTbYZeX5Je>QZ+d3ZyG40^P=_Uz(!xs#e79}F%IvM3!}Rj?cG^AAGedbv
z-KGB^?W`rTCgNSPdOjdK^%8zjxA<1tl>Nw1+LYI-NRzHmGRMwNzMIH%gBK_0qxp_l
zD9sg*o@wk%L|KW#mhT?=g>A_k`On$`>uSnJ<pD}%ui?n;K3gU^>@1O;=cCuGz$;I<
zphq5dH)4|CynNw25gldXU5rZm*-1ywc5uW;S(Y-c%i3nsp3Q!H(|yy-G{qejye%*G
z@DCH{p~r+S131#$ghB^2d4xsvNtdky=m6}sE1>2Rb#o95!uR9e6}I2!X4u=~mhU_K
zfXJg>nx?(Z!*}!fFo4I0aQR!r)r8>pZl0Cmiv~;=6k)jKOmb%0npv4Z4-%Pp=%NnM
zT`?0SQrmOAv=GJ>I~H9IFM3tr$e!@Ct;t(%z=}3H&Fp2k!<$v~Js@%8=a}N}#u4%q
zWt`_lSagdnqu7fLP{&b4Qg8q==$_LC6n53(DlS{;l(6RX`FmQ#h^H{5U6=wzf|3-V
z3Q>656KwVak1XX|T}5W|4!h|x)5x1lYl>P1Tp_-1A|pWfyHClyBhT;yRl<ZH;F#?V
zPM-D?RwjdlB@L2M3^7o}3*Jz8HvoBNestW?dB?FDc@U+JoriUF@Vtq^!+{4C1M?F-
zG&r+4I(qV*x05(`nX}*-Fr)M;=lh0`Q99*8LRu_~F8IW5BDVbB(SWU;IKBcPjeGzT
zdLJVPA75%gLLv`av8!ET3d3X(e&h=SioXO9S&>Ax2I`$qXqvQ*s%yZ{yPk9*rvVEd
z6++NtlI?!LS}36u-scJ1B@n*<VRS2$xG>2_4KPqxVP}BHWS5FT9pYh4BB=X(7@0g}
z>pv491`2-QE6>>h$N-BlOqj;>;OFDD6B@MeRuGdszUzY`?MwszQ@>@VOXbEn-Fy`G
zOasriDobc6Maqx?C<8VIC)7W7{&`25CdyGq8Bn6o!p&P$ac(k`vr$jd;>{hm9VM5#
z-~pyw3_PiKez<{`{A3FC)dN5c-gq;MH%u?iQ0C@*GdS~f386n_O8(g)V|b(<JW_`k
zSdVG|7?a7s$A_ks$1GM*wz_QhM1+a)hzH+HczHVx#%})xT1A)?5q5aV6>ixh#XuD(
zz(NHMdcfuD1{q-rZ}&@>;KkTR2_HdC>^O1DuPlavGf`w~mPawFEn`CFI_)1c6$P#c
zN}RGf-oQExK-07zI1!i^g(PpA+27<?Fkr+52v3aVxfS7!bbdq_JA$Wv5fYk=GZ{ZQ
z!Y6p-@UC0%&AtwgcNU88b@H>vm2Z}j;ATN`92b5vD?6UqDOU~suAKP*#2GA+o!oLB
z(<&6VvnP{=oY|r9_R9WNj8$@BK!*h|L74V;D$U<?pl?9XVB#%lR5o_{`D7p2je`YO
z5!Pl2?;wFH;CLfU)i!LB6FIz$pbclpAMP+}HzGJcLFuxO^CxH=J_9nAwpEHQ4^M;w
z5&7zp0@suy+o3DNjvNI@S>cU6P_{Ucm8$~kQWN1c2tsCE_6u2pcMrMk|1%U{q`kDD
zCPdw|Enh&GA$Km@21I_}geJa(sc;PzgTjLWUnY0JZSerbuZ|~NO(>xYt?wCGjk#^F
zC>!>Iuha1^y5er2`|x#G#$C0$>DLUSI@F+R+JPQ?BOQgL&<9=9j$fa@lh~vwW}Tmv
zpYZ_&S1y+Y{`smx7yp{w06NiB=%Ba5P{&dE-GXoW_4)@6Gp<aB9iyTNU|ds&)n%%H
z9(?0!!n^zm^!VHC!)0i?!zgylI#Xk@>1MccHM744(d$3RugU8s+`_Fd&VK0e>!}bo
zp@;SsXwnR%#mV@=JvIJ;ig6E}d0f{Vf9Q)HU@XFFtgLw)bTI}M2aTa&fYj9Wh^x5G
z-L#Gb3dY0Y<Aq6fshlz2JF_3-amLWR^w{MnL;3k_dg2wH#m<$L6(9YC;)~<cPP(U(
ziKF+&HCAVV5M|f77wG5b=B<=r@e;W>e(hBA02Wu-=|NsIfAF`t#GA!j7Ei&mFgiOo
zW4_?M@G2hsyt~2qa3@WSu{%a?$;UI;eW(&nYI$jrMfFq7ACEd+rAg`a+(Ppj#W8e8
zb58klgde-vW@Z?Z^SvL*aeSP%w>Hwo#)jrH-TZAPj#nRXK{3W>>^`YEGBWVuUyKix
z$2M8)A5mUSq}kai?_@!db}n8g$8F8u7yP*Z=32<W=?_bb3s#scEi5^|Ha6GO`s;P&
z=UL+WUMT3USQ1tqvtw<PT~+!~Hk`mRAs#~`EJ%-OqQdvJnfFo7u>*tJ#^SQ`i2A3w
z5DWFlIm&l!Jg;J|F1OO=`lx{l-&^BbUQ<(J>Cxjy>G9*on)^+;OiuU?nBurIgo01`
zH#09fjtZPk!=sauMe^L8(EMmZam;x4)|9SE(N2kWX?ZO@eDrbp;g5cj{{A0+kv{qU
zr|JHKkF2;`x%bG*tofyTX>4LnbJ;1C@mM-JzfP^AP^O-osV?U{8Bv~0O)sSB*~PTH
zdf&>p`wt$a#l;oTu%<BV_}1<{RGk@5Uw{2|`tSe!|44uQ{qNG7*Duq-!EWL^WV5p~
z>En+-PM`ho$LZrwK259l9w<GO6~!Ct8@$-bQFk0^$B}SLON&;*Jbd^lJ^AQk?<&01
zeCXh?qdK&mUcG#sHrF?;_~iv+6aZZGkL-%0JgDEaRp!IA!LkFB^%6|hLO6+s(^L3N
zM720ieWPtqo@!4l$fIaNDNFmJescUGN(PQ2ytm4C;%41<P=@Sqqd!2I%+aOubMdXN
z)s-dX-Gb-)@Ig^ME&IH}avUc+8L5BBgS;rqQ9QFd=lq<RJ~b|mSB>KkXVZc%cKe`M
zLt#OlA}{lK+R(_|J84pL?8y<;yKXBzfBNV2+h6}{`v3jwzoh3cU#Nb~rhoXwe@H+1
z+0Xrbd&&U?B<WvOdCxDa{qPA8$sbA+<)!ir87QNB_wIS!Eb=*awxT?Dc^)06R%_qi
zrE>dK{*<;%k4k%IhbwIa9535ixuQ6%5G4)bMV>_2kOu*wfC_t*SJ-a@)+#uyhCa>o
z0{R3lT;(e39xi|UlqUK|*8`cXcVJfb>RAwgtneW${zOIpO`Tw!hNCq3?jz-oFz~S=
zRh<KU()FFuba>SD&ad6Qeb)o}d^l5o+383<Ls`!H67m^;a_pFAXQosq7G%cv=%}Mp
z>ib7j2T-=6G$Y?p#L;i@(!5{15Z|*4j%-C%j+D&LHQ9Ndk%stITO8HNMLx17z{y5&
zO3`p#4CJ5u$Oke6T5lAfj+(FcsybhJR`G11P6a(RUG;W?BT&`no^<49zF$~J**Lc|
z(q_}1&3=2+ebdY|#jSrC7kF-3`G;A;RrK#L0n=tT&%^>1O~b!GTmkI0OKu)28Jcwc
zFvzbTcU$=WIDroh8&IcfF!YXTp^vKoEh^yMd_D}|@gZFP7I8Hp_`RE_6<`1Ppa1Kx
zNHdMVqWT|!y+A^Kz^qD$?mQcU1gbDSg416YolT9$2#?4N6Sz{EwJt;P17UJ*YRJa&
zpard;ErAO#)<6I0C@S7UWPnc=CSyyOh`~w4NxYdg`0Q(3gezcd@O6I3v|Rd`Fz^`@
zavLJ;L62KEc-sIGm%Vi7&FI1#y70#tG%*ngh};6=?!^z$Ij;jHry2S(&V1@{EFe-B
z0^}!aP;=AWEOpEIfk<T`Eoivu;f84f@m9yd+g(gRc!P$AN;M%SZVBSNmz~Z#I{mC%
z@MP!0Z=PN10l`Cri?@`#wL1^4F5kwagolX-Sbn(Ekq>!-Frh&p4|q?g5R3REwjcB|
z{icIuq7#;%=*j>Cc~GJqAD=L_Gd-6RpA*auACz5eC1-NRg9mTYX2p;mE)L5B_*;Tw
z(2WNJWy$Y-(4=VEkqI7r=yUD<JwFJ^M=#FP!9!fqhP_+`NA}r|(h4s_@fRX+J-#=g
zcwts#Q4dil5}G$(JYbQFGQH4%J&v4=ZQV?IcnfMsY4C0q@`mzLIy^`e6O^p88Z=IE
zcDWw@Oip=I%7XxCTvs%hXON&mEpOrQ=8qMyu2a!w*`CkNEFn~uj?Tb0<eUugqfR3u
zb(K2FpbLeozb&V*B5`_=FY0Ja&<zUfBFKChNb)uv+#x~X)oE*De|q8kaioANGmcz`
zRoB8D`Qe>(ObRs+W8lQcP<i{4H^SKc7;=*i2F8-dX;wZHsnSk*IL2{NliN8DD921+
z`XC?%x>0`IT{|pfn5J|fa6%z-5EyVZ`*wWH!(<VUg9tF(r{SQG`8ymw!dsRMSw@G?
zcl&aKE<5bxu78d#?0Co-e!A075JdT$y=*!Z4&p;LcvF8&UuX;864IQEw@o+KH!lAK
zkRyb6mZs;MdEh56-H0VH!PVh~07uW`KHCC1uy6z<40Lka8v*cE*(*-EgeWHm@y$j?
zRTf7g9tsb}G#SFKJ8{6?a_98w9%;1=67%sxFN=$$1@H!GTA5iIxii0-)Zv;GgYMyn
zDI0w4g)72VVA(T9H_J)p2%N7#&Zm&2NDa@B)yV=$KFNTEP2EbE{LeQa`UA|!w*bpx
zeAwlU#Rg!0@DhSx;GM<>--=#Dpi5X}1-NIa0y05_A`Q{k*)7v`NI&y34{(GK$c(g$
zLS}_0sO-j-n85ibexL}K4d)#Fn|;4q#n+F2`A=4S0hO$Sz(N;)QRV`ObO%22rx%=;
zx8Qr?gBQ0RAS^Fu^!haGOh5F;)r_a<RnbK<vR{;SL?*l2fE=bV(r&_*@J)O1D|TY7
zoPuKav%Ci%4)emaE4!;eH60N4lCEBSaaVX5#_Q)0vPZ~({xD8=RBT@%Z(IR&yv@sb
z*1+p#x&UzDSD?^wMPAkJf`g%4J@^1)Yg_|wzl^W^4e#=U1u@n+Q~B!(S;8K8lhk0Z
z?B{rHLgjxWaIatg<&LFfM&t4i{|rH|a-uJBxPpr-7W`^=Gp#_6pZyEa7HU8H3z`%+
z{RfBXhklx#w%zB~bLrRHe~5qobqf?89t+@BpI>u=vkgHktWz)Ic@<;ytE+fX@%oD0
zQt|e^pW7aD4nwHqGd78MC_nRuMs`@S=vl^kG%IY3m;EB7u1s!vUN+;!L!bPCKXb;5
zOON$e*H}28^TO!q>09qIkyo0USxn<ov+3-DZ@&zsv(v8TXs2myWg#ujOeIc-D5pHI
zSjpS@EKo9*<|R6g&ZjQ1P>q~ZlVfRNZaPh9e9a>DWg6D_dsNS@A<rxLmN4HJn;siU
z<KoN0E#q(&pjo7k-A(bL8oS0$HRfm6z%k}Ye0xk&2C~ZY3GX=29ASQT#&aR^b-#6#
zzWDPO>Caz$o{kP%Rt(I{#J7~F+sp;VCBr3Yz2dmEanEbmsfJQPQKzZ#ku*Adop!g@
z)5-CX7e~DdLOdAxTwbcIBm@iU9M8zY{POB*n%6wP-8o7xUp!6cr>C0pFQ=)gY00c|
zQ(wt<s@T!Rw~W}W#%q5Z_2%!SoulBBJu4E4)Adhs@q!_;F(2e&w+9u1xeFBqMI!l#
zoqXdMvw6Oi^x#o?^zey~sVm1?lBXPPhEfw+YBF`G?0y4BUR-f}rQ}m|c3;ZG55)*a
zKcdi><M?-_j{z%XgQ9YG|0r#2@20Jtec3q<v*Y8=s1H$(S<t7PHJ4Od;FwIl`Q^MB
z=XlGtwD#~LU5{0E)2Z9(rp=A5^zzl~^z`Z1R%~)y-Q4_A`pG~1GX3=D|Cql2*-z7h
zN1vpb`Q<b`I_@1yDEG*>n9I>VcoCC2I-S<;-A~KQE8ZP;cWgpwyOZ`0I_c%>jr8i(
zy5iVM2YW5=E+fx;!ii+|vrU=h_R8N4o_RA1`W)YhGKn=1)@;~$q;ZMr7WIVrH+3Pt
zX*83jrn#n!XJ-=1uF1*KG&P}XEEGx86T_-^e6wy`obIH<_HpW{?Xt6tZ}g#bSYDV-
zYs(x($hW#;cMJK8oU|w6rXOJ*Th?(<L@@W)&FyDW^K(8)Kv`SqDwv#7yP1un2jR!=
zOuhkya*#aX$UE|MSbg1?s@~!5M*7Xa{@>|;==$obKYN!P_2Yl}$Nx3`;+MZvzE7sJ
zlauu3^=lucIHmGiQ8`R&PEGxZ7qxkTTm6Ib1;vULDtFZ{kQUM;fTnYF=%W)`hx=-0
zC(0SxORQN@BJBEN{fzlON*(-YTmDYH<fi<fr9HD-mF5IJZHv0%e%BYrUPc`<<J&;|
zpm)6xxSnS_2K5G6NE_=5R{YS3394tA2t(P$F2cOaNp35_;H|LysMEXz;54WWv93pq
zD2;eonpoN8%g!&pp~a^yh@UulVPIy4wV*NQ18wuDd#Vs|%p`fujz6xq>aS23%}jCn
zlj@Gj-rs;cd+YkhZo6suv7?ToF4;Xr|Bo_hmX`;FV;3#E!;lrZ=swsj=^b~fL)99S
zF5H3Rc-P2(-E1#q#nF~lROY^qZ@=SD9z}LXUm5Yxm&P{-d0Bw^hhmO?lC}UXed4Iv
z5j(Y$p4Tz9qkZSLZaa)i+m{t=L}s(!-gMtIGmUKJ)E&o&pYeg?Sqd#s=uI=gEbuP>
z$}aPr{sQc^>vp;(d?@~1bp1Xb8aAL#*G-7F(hInoxZDN>?~dog03IL0<!=!e;P`TV
zeILY*umAjCf8}SFuMoD>Z(Pv;<WhqOx2@sNO`StTiX*@@HiKRyfzNqs`OM9RJ&Ixo
ztbI%y;S1I_8HF-HCh~}fCA2fBOHl?TSm>}<+$ujTxFW6gvfEzREi;fE&;*6Lglxm%
z{E$_*AW_4{jWFQww2v-#_%bu^;DKMb0Ofv!v543+7E_nQ<h$ds4HW5(vMUrNU#hUk
z3N3h$nYNcG+xg><$N(0j?SXhQB0Wucxie{4)&Q6tUp$1X2PFJ33n;uF1U%UE5cXhA
z4{!z-C?~uP8`F~-4Yc&Y=j>+iesE|I&w#e<$Uz1k>O7=)3y;pfya{c|od0n}IxLcX
z<E_uw#>~-s$ia5~7!=_ba`RAOV8HqId}x>v1UuOL(Py!CwH7Z_B`8-M-M0j@(4|M5
z9Z{c39lQyJeh~1GT)lTM{rHDJOrQSXQ@<6%o09C9BRirbYaqL*gGRfe%U(<@!t;<N
z9Lha%Qi;z(iX?6~0d(=ggukwmUzEK)^oJcgzxZyN3R`XlVMFoFER`z*0CL#7sqz58
z%2iny^v*2A8(^av0P#^{Um7s6&EId4YOr)k`o$MJ19%4Qq}}yVaZ+y>z%#+ZZe@f;
zb>gE=b14BRk2rGRQgxMyHucYgZ^emBerrdx_%k>m4wCJ@)~2(-wBeB9Jh;>(=vehk
z^_|Ee4lx8_iNOis<R+Zz2?God9##K2B8GH?J@N>8`U(bj42VcC?FaT0E)zCSuICrh
z?DVV#PwcA9t=lpXcf=oe;q%4aNzmmGlr3^)g5ZM#H)iAo5YGfI25ALbVhAucmqjH9
zhH&N^f8P{yxGh7bwOh7|huqjDix+${+Hwa6i%D*ohY<~#p~?8bA-m}1fV6l7C!FCA
z&oaTFE=M~Q6}-hS^3J5D)gPPrIsp!2l>H+#3sSP#BV<Ijh?{saULL9?wsS?D&S|QG
zXNv;`odLr#hQyGVK7-xzi)TXt;tWa#xY=5T!;L@gs+?>FWOAa-e0q5UV2fEU{P49X
zPG1h|AlyhtrZpOzT$R7>MIN9EuA3UDMKRz{{n6c4{IOs_P0Zko%0?BN_~1uid&o<E
z!i6*PtLY8cURZL=o~2ZL7tS(b7OX7PA?%TtvfGn5Gae!XG%|F{q!onB!4EIuvqhI>
zFTCx+4O;lf%=M#=rvP~2G9Ua)dbs}Wx4-}H6kosi5XD#EtF&6q0DAV2zBg>4?T045
zgc(Fz<H$dO+~A@O+rv=knmqb(4(GUnM?ljK6nAfV#iS#k9^Zk_NJE6JuLu*|pbdPX
z3&5VO4G@+94Y~-QbISJcin<qcO?xl&;AO}a){Jk4X}Wv-BXUFSU)u}xguz|OTjPM6
zxNO(etVb11HlwQyck}Aao6`5nuK~rcc@+pZHuABiZNg1l%_pvm&d}=@R}<dl*AMpZ
zqi<fi=`vxMSnbrkQFi@c|FGS-{4>A80DKKq+}ZviP{O>6{zGUCct?e}74W@{ON5NO
zV(X39V@>&G{^n8QGm1V!V@i!BuikP328*$mLNiv;xQCb3WX1v*^qGgl2T%O$FMBKo
zM%?i9Sc&mDbi&c+F>muc(PLPS6XY2D^Tbi(4<4+grR5dbQM`;RGI+(fogFMFBv^3e
zEq;zxW1K#!F|xl4G&|$PdFBl47U8>9C_&h1#EBvtd2fZ7(&HT~y!C%~EM3b_laX|H
zMDqs8*=Zl9llD;>6CXw8`7As+qK`$>@iAV`REm_ZEW-L2GF{{hbCZ*kGk;^r-#udP
z5ZUB84T_lAxrH>hu;@H^{_N}Yr$7B6y?*t=c~3e}0`guQyTMxfyJ=@<+xf=4JQM`V
zN9FC^^j2w4X@7q^{a^p>|48fWZ=B}V(V^!poMJ)bluJ=Mfn%=4?jyb-w7IcvUTdqX
zX?l7_Nl@;1KEaq5g%aOrI!*2FaXRj@^M+kF%;i<q;^(=yWQsXEyYHAU#l<||yLB}G
zC*0}jSw1?Fmt)2E_*8jxmQGQoX->m@kQWyxCZ{BC?0TS%Fi&GnPrU(Oj@svC#VO5o
z*zv*17|biFFNn@!vXACdK2U!sqqi#mj^<{37jAoRKW*>qrTv3eI^+mCrH|c)qzUCD
z3+&S~)6Nrir%X&tS?R(qI`Rm`@X!d#O6BF;qQXoI&C8%eX-x93tUX8%AAgdbeDBlr
z=#$UV^!#$Vx-;VM7M+|%9v~})6Xk;9DHh97N>K(gO52FyIFiiUdk1M_Ygc8wopxn!
z^W8e-8?BD^hH}BDgGfFU2gpe~iFstqWmOqmkKA4rx8s#RN`W{YZaOV5E~KTUdDVx-
zv@l27@l!reuq)|qnweDjj1Fm@f0){>opgAxou<_$_;wq4&aNBEi~2U9x%|9%&`wb@
zd+x4s<eTg+ck;ms&iHOO3OP*bF0xYJQ7+BRz=Kmv_@*7-3p;k(K^ZVD*-&mUo*_@E
ztF%8WDBz>MY((S!;p>ak+Feh7{O!M`KmPtVX?t^Bai6L#jeCvdfBWD5V_H?2ZEo$P
z7cX9={Z>0&vU77Rc0;0wjiXp&4F&}V{P}j@`uaxN-DOu_+b?<ZJv7=Z3T<BarhSp$
zWgg|WA`_(^?TqrF9O7jG`d``~&+kyC#L;}pKgEkbYiM++lp~DX&&!jtZH_r^Lab|W
zL1lWF;-k2bgF1wqez8{a!MDg+E+Tgw%410C5poh9-T-v04+5s4{}Bz(0u*<YHFbp-
zhrQd6Z@jCVY2Umc?eE>Gk0<T4brfQ}xImrYg>YVWhX+bAlxgfF<(OAqw7$6D`)ces
zJ{5lyRPjwo00pj3AflfUZ|XMR)kEnxC6l(m%MHA6O?@RD&3uadj&zg$*!iaZUvWUf
z^MDt(Q6BTXH=Z+9{`jITmYsX>RDUaz^-ut%7tb*ILj2tpWYT=Euat(j=ZYIj;wP_>
zCHh`u!s}W$*9oxpYr1oN8DOu?Qdec?-|v^p9lL)Ge`66gL-8xnq-~l&WtSOt%z~=v
z0NhQ?Kt6pi2tSB#6L6Qj7{v8`;d`N^>n4;kq-_OYznxc2-@LvZ+?J+(cz2pU9D>h>
z@e0M)KmXTXnOrvqMqo2h1q6-TwVgqZG9(Z1qXFp3G-PTv!b7CIEpA)|au}h83_t8@
zcw`Y+^TdRS$$Tq+^Y$Cb{_}$}y9+Il{b~F3-$CQMJ<%D!t3a_vrvrt|j5J^K&S}Sv
zpbXrzA2g&VM?egYGx7`5NCSp>+d??<Il@PN6hFe6Bp}CaZpN7?uI%QhDhF`b@IxhF
zadmUP8XH$wgpa697iqz*n+D(xOE@D#2F&0jfL?wk<K{fHD}ty1P&%<?n>X%aa2M$a
zR?repEX2Z@-BWbfdNO#Biz2}aCOz!@&{K!at}h=aC^wyl3p>{({5^oygOG;;yS*5c
z)0r39yiglaS3o*&^7sYd#sH5=1rLijK2ZZB@rSS9*g{^m-6KnWm~-)q#o+8V<HGqb
z1H^>`i-asx`+*4m4EXmYsRntlRhn({fA@EPpMLgtKTG%S-*<j6ap27^9^hVw7cEr?
z)A3++oq{tipqQjjh6|H2HFabI9(I>>moOP}9FeH(R+xCBCr@k@-n<y-$a05=yUR`!
zLr+k~{m`eZ*x7{OC=mRuEV)M*C}?2I#CKeiU3NP0mLr2?6qS5vm+xkg{_Zg!*KK=1
z$$8g^6ay3uE;GM7cO|6$;20qL%^7G!%Xi4Epwr|c6kpU$KIF<#hw)u4vcqpNNoMLU
z$0ahU@<4;mN)T^diKk5OM59gVccFoC`v@f^d12BT$UBz<*`%CT-3N$|a=21k_P4|s
zDA5jx3%(*CzXF0u49G%h5)v`EysJsTj0STvewzw;fT({#7k6D*4tQ7XO;<%3sBGvi
z54XhvkSH8D`1jOu(gq&cbVcys>^?N8Oyh9I1w|Gm@~6~gc;JhfQ3g!tKq?^F<vyuF
zqnz?RAWUikgYqB=oCl#oaJ;7Dc7A8xwnce|8-K_px0ztdBIFfm(dA}he%yi+_Z|@5
zbQvPOIbM^NFd}qp0#Ar!BOZ*3XHKUAk_Iw2Lg|9mez;lCD&YceyJ-BQPKC{BB3Vvb
zwi6i6*vTg=0_g|HFZ_d~28SW82H?y@#Y%jZ$B~3gQQxvGy7<@J<$$kozR3U=rum7N
zv!eJLCqq}%J?tg@I0`6)@d22negrr9;gmaQ!~vxmCVpV>^Ht}YDp2qxFizit27l9K
zz9Cl*i_-+zBn}ycOQ8?mPK~JyILL%d?Wi-K;6XeAA(q^v71MOGg@?lwn?t&95=Mmd
z_l37X^&VON?Kl5(6km}Kksk%Z9x3PtQ{{S7!rTlS(6ko-N7xEPLVn!@(e#OM8^4<X
zuEZ1efXM$EZrWqqT<qfEcB?i^n$@p?E7u?Z&vTMA1zq%9b(p}__NXheRqpVs!<1|&
z_@IMtlV9l}3r*1X^Qr*5t}2UfaaWh?x=hDcUjgsNm&+`AZ~*in|Dbd<dG-cx@~CZ1
z`a#f(hb~()W8D2b`rS3uc&;k%nn0Mo>()H(zXSt$0`l*b2~%#&TLS@gxc;!c*I*|2
z+y=pOP&t+hXu5A+h;&oF8Y+BLrW;_tzN?oqhu1#<OHjy1cgb!4k_NE2O91r+jF&IZ
zl8-<<@22YuU6&{6j>h61+gD=;fmnNq69+OkSlu0F0iJO_IK?Gz?`nZAL~<CcMUj9%
zN+OOs<fyT!>DU><j)3V276W;q?XJJWgUTn}ozPr@1^4N+ur!w@$A{CXqMVr+NwZVn
zCp7=Oj-4m;ZyMJ-4(6p^WWVt7hwP$3aOMqO4AmTj=XCd|l{$y}X><Kedhz0$)RCKA
zahK;D!+4aATKj2lXVc$IV>cDw5Ml?`1PU0X?Bx6+9d=Gq_f&bRzL}vumlsNvj)|%1
zw7Rn5;}|=Khw1g}SLyYu=V@<mNArmb<zb}PY1VuNzunDsl?&ep8&bUStu%H!pm<{O
z{g0o2p8lZgPk;Jj+TPw$e@&mw^E{4M8zF<p2T6#cW^Q)Qil|NT<=91buCVKDdNO}s
z*}K}<O~S%4^J*5W&wZ>UN<{KA=1wdkYu+YV83ZuL_5!!k%Puh#8{{`HZXT;VxKP}*
zT8HW2fbU5i`KUh13S}AdxUmV%8#Q;2b%VE7WLjAw9@G;>tuhUT5xX^fG^8${5`yB#
z;o{jw8aOs^O!A@p;#)q!>r66*QZ073P%lFfWd)nUu<*{>0t@2iKOSkN-cp|INYh`!
zjB4H~{K!NapPWq-(~D_xdcofdJG)G&eN3E+hrAKFkA2id9PB>fgcIr?t-?E<L_|I5
zc2CpZKKavCo*i3Or$;h*7nUv*fLtS*JEO=M&T>;XCO8TXMV>ColsL&NzMq9ciJgxe
zQ@FM~msXeO)6yb1cFHKfhLv}!^H=H%(uL~Vh34U{t+es#Y5L~tFT`ggJ$mvu@lALu
zKQ-o~Jg7e?5y*481j>77XD_|k*!B*)))BkwP>P-_|H*r2pp^vVFFSD9r9+!S2|<RT
z$k^N8x1y3dGOl<Q=V#KqF7l9Xnvp+UPW;iFoZ}V8Ro|y|jb5Lpo%Ltw^WXlf<o+^s
zJBQ{;-^4um;e&@hO7ePWB<<{XQ|C;5)!oT-cYH>9IT7`rc|I-s&amp@g}?3f^5rWl
z$#!;j6*py+>+``uUW1|?qZn~d8THZQu2?6jb_tE}Z70=Dj!GpS(?wrFo$w3cqAO!I
z^0YiJP(rw01#1r?3ILCP6Z{NgPAaJ8$^vM?veOjhjk|5({3#Qj3u;HKIryar5qiuh
zKHkkG*^mzf5OzLIz`W4(ZmXh|tdx^qd{#WnziEe*7kvxsFYKP1omIXoyF_MX7Hy5X
zigL<|KJh^*y}S@F0IWb$zd4}TM%mx#p18j$ZIu@XcxL%U4COy-jjf~ObktU#?0C;4
z2QMxRdu{OWU{}wx%QUV&Pc1vm&rVyR%Gx5&Nld_bhC8RXyqwHwJG^8OySG%fiW9{h
zN7C{+7M?#=lnL+k9bM!Dyp%n<X!(JGH~Gq15P3&^r{AI<PpZqLjdaCxT^Dr$TQ+eA
z+SqIcn5J<BuC61MUkO+7kx87z9~61vd7gbC4O;3_AS)C2Ae>D4?x>c9_suZ1E1>Bw
z0eto1`e_2*M}zxjs2V;bjv7c$ufN+}#)a9u`l0q4Ok2#sfG>GVSYLsDFBJLyTBv1y
zKjfnh|My=hR7`Fe$mGB^3-H;U6BI=cZB%MUstb&pZqzJL;m-H6C<Bj}OoIzViPwvJ
zSnL5lWXk0Bi}>tqvf`}Kai=qLvchRgl|ePq)Sh^Q*3V(^q2XcTCM@J7pX?xvazgf;
zLXZ{M1v$4MFf{X_GE1ahND7}S-zs+x8<~=W$_*wXBTVF>X>B1*;J`#WDZiR~{D3fZ
zN5&v9gm~E@u+1cyR)|n^s#egELoc-7me50okKp<&ayolDwML_aiyvT&$Uw=@J5PAP
zS@FfTV3ZZ-vg_dpz5Vpy6CHtsaX1FW3?%5shlknW!<!v^$TS`hq`~PxPKUGACYO0x
zE(_v24;tPA!Svf^XR(zZ`Th2f9)di;gFeE=w}Y^TOd(&O{!%ExL%LA#d?<`*K>Fm#
zlk^Y&@b~G{Pd`hmt1EtZ!jy;m?#_-T9eQAsFN%P4l>DIsb=;A(xUdHnf(mC!xv3NN
zf*BrFpa85Y4@2?tl`YCYg2zNB@{Wi5$e0GJ!c#YfH9++nOYA^W*${W>>^%w8z?N^j
zp{zmi#Xyg@gDzE%IELd?1EZ7U*v02|B%VGy{34WIh~oMsJaG>zZU$KF;-Wnvw(%OE
zxXu$sNZznvN1|$>2kpFxIibli4+bl8{Z>^BM9CK`!dy26&ZigzVRbnu9C^vv(4C{M
z>vsi2OF)7#$e{3W*o_fwi>iXmBTVR&5BR(NAcwBVw+K$WR>Wvv3ODZ>)8K284{CEB
z{sbTSjc9YK#V%@oK#m+!hI1<ZrUXoDt88SN35^vxXn^p_ZW+bLlmu~p5hAxd)0c4i
zlPK9SiQ9GAM1Vp?)w10e{`?9ZOo8mi>6nPX<e8TIvR}ld+zq^b<Ue43VCd&cf1u0d
zBR}eZyA72`@DSLD_^1Q$i+}*~$Zi>MMZajgNrF0a3RkA9q40$cJ|QazhjSVV6gy!u
zK}1|ff~B~FSdis3L^(zB$TQ;17x^MH>VDWUofj@YOt|3}Jj|7w`DN2cL9v(008d9;
z2pZ(Jn6s=k9g^Fon`N>;H~jFe!PU30ZrTl^&f$CUfDqsmhS+%il;EXg;yE8+(TYok
z^DfI?AnJRz7rBvHR{VmivHE4+y4la3y=kktnM_xX4StyAX5sLQ@KFyl&VJAY1&}cr
zB9`K|+2aVKylC@X@A#_Pq#*LJq@qBiJ7;$fc(E));bSw?HhC9l;!0eD>A0g1Z<P6a
z_!a2K;qHk`)<HbOs`zAyUx5-9lRBWEA1X!%eGi{vukq$qll0IOv+&7J%?j7Fa}h1*
z;#cs75HU}>164p&;h)1+VFb6ixpFcIhKKp|`k7&c;|l1f18&lVJ<KNb(CKQHVUJ%R
zEtk!L%W#7>LqERh4qQX(OjF<{Z4Lc?<%)uVAg&w!b#xVgG8+H)U1AEYxYc(JW(M;Q
znsPPeFm3U%JIsKf_hkb!u82FXCN%j5ADK`^!UkU${m_&@`}gxU|3RIVpofNQ0D!H1
zimR7@1P>EM$u&pP^LF_a#}V?K9!|>Oo5-j3<5|U<^dz7>`(sh|Qur8a6OCgsU;KGl
zEXJ}iA5=KCX~xy~F}`B%$k-b^^92@N+4aQ2Jx3PuJs0pOtgJlYMYxj_FCua@Rw!&V
z-_XUkuGp=^c#X^7_|ZIsBMedfaO5K2N!oAm{hh;fe0rwD^ZjX##G3JO;X{)3h{f#I
zL3;V>W%?Y&(jR`8e*2qWr?0;HD*fU2zfaGeJ+o4V<ETb7zqpiaZOzsEQl~P2MRVpK
z_>)KMPVlZL%|Dp`%xT`Syt?KW|BgCG>G|_#>HkmMe>K~-WchvAIA?qBkH6kj-Kwtc
zGF{aM0Gk8>G^oHE#Rxv~NeX=hJvCmW2#Ewk0tqrCks<Q|hzLC>6rzHp0I-2>cKNF7
z?>+wPeYW?0AN_y-kvZpFYwvSSHPAiQ%9)uXM~)n6XPTLrpMLVO#(HZ`3w8j^&CPle
zfA#sZ`1LP-<~zc;n{9q}+9y0rDqR+r7CiRdlC0|+>+$T#(^!7?OtdSpyR)M)6uaM)
zo*L`ti5naPna3ZJClBs_9uFRT7DtDB;=|$(i&~7Sa?+Q{XeNmVINDIT$H%%JKVp%Y
zT@~KdA^zk%M-GyYEFd9QPK<McM(+0TaeIdv-?8|l^k63s>9!-qP7wOnql!C6K#q-Z
zuM~by{lOx64lTRt_yCf52j1OGgtPF)n3-`TckZyj1m25!6D$^uW*SIdj$-qKGX+rT
zQ9t<24>_rGOjZ+T7Wam{@Wmn(lkt=j?xJF+2xBzzyIqijPmVX)(L(tjP`M)wxsPpZ
zBo%o|c^u$_V#OEW!%SqMo6Tb3?%uH%hTuy&ddG_DgFn<%U0~-?E^JMB=ghd`$b#|Q
z+)T{POR0=3sQj=Sk~CmJiUlRg&eGys+`O?6H<xC86yfB=h$`uE?Cx%=tS!4Na(v^~
z=6XDN^dPo3H{#i|$0}QoVt#Hee((4HFz&tjn$mMjBC``OuL__p?<sHA*4JZsc{LtC
zR#|+$=HordXWo0HDfMz>lwDnUv?Aq^ydmWotFd6ZwYeQD^80joB{nxURfaE={v7E)
z88;T^RW?T57Hn>@+bWL}oEjgAnF;Y7z80qkoALOw58}gL{WP9F`aF(R4|43L`Z_uh
zvkNz4>DHYXpPY`t;c>-bB1R`?V`Q9TNvGW2&~CA_XhY$zK3`QDaLnX#tgJluN4b<s
zcB-+vk$ETP6}&K{x<MJBFU)v}vgW!{>H`Z*?Brsx#%-pYsiW;~pwf1z3{#J=yUze)
z{cGEs?Q3U|%XP9B@PY$0w9!C!;mc6u&2lTA#6jqkrPPxWX6{O(4a?JVXrFU8Djx{r
zo*x<DC+isP33VVBe&9iX%)^j>p086HGM_=4H$0RdAzLWX()Q2H&Uhyk?<f{sxigMM
zG06CcnbW1<L4Blb`QCQs7F9=B&?Aqwc6Q_W>P9?YTaSIU`MyR%^032fc51?PaZBwd
zyND-Pd{r1Mx-p04#VVy8ZxAnhIscT8v?r30cF1iRyACA_<-+B>v@4`FX^U(%uSn8d
zcv3!=w4=?UT)WK_9Xs#XKyY2}*THMKdl0C3pDl8+rvY}hV@O>l@wN_KQPtk+^s|$X
z?NZ2EX|UTJC-JL`W?lE<xmbI3;A(R#k4l)#y?p!qw6=t<J5c?1p`Ql0?C-Muvac4t
z%zrPmaRjnnwSK<XZ!cV?{c^uhe4Q{V{3d>`g5>knczNOL@BX`g{3j~yvPv~Vkg;k|
zWX}OC&|zwP#gqgGXS)D9c$@Q4U8_#V&14mWz1Yl-N?U%W)vtO=x@;s95@)P%kd=B^
zz{Op9rIV>di^kqq8XZNOrpMt|N^69MR=L=cSX*e!D?VMCG%50AD9{r^($uGUE9}Hr
zg2oLVnLWv;V6pOy7c>x!9UZ$d=9K>Q;wpipTa7a`Y0E#4;`TxvBwK47(`Q0f{2F!=
z$PB$z;VQ2~YSb4O7Tk%Z!Sy5q5kePl@~gfisjL`1fRW;b#*3Opo{l=ZaCjm5#V5P(
z%i^mM%!DY<C#ShXfSB=8pu@ups=go`S7g=CF-2fwEU|UFgU^g}ocQX3P6DEHtO6HZ
zp0Nn07jLFHpg7?Lv3Ma%PAeiMMm+!ZfC6z@6<1{ZQn(hkZ{LdF{_Wq1ciw$B78V!W
zkr<>CEuf>%K!)QfS{WIB$X4>i5zJ_0z<_bcBISr<PqbfL<gpN$<^}AihlMYGfDTFJ
z<#gk(G62*CA_lGXi@;IkEnlQDuqeqa?_w?%<O>UalbnCA4itA^QErY7xU5-9dB}y(
zbY=*ivc=$s&ZXl`2dsom`C*}r&PDQ0{<c)%315Q?$_1TF<fEfC;sH@^FQ@F1Hk{22
z(9wW5DV4%2EJiWl<)2Ov3lly*@qi;%4&%j{BWeGlPBYeubK_Rs9Tk^TmsN$$k(BwB
zmEwk1@`?^AZsOs1%k722ta--;;668dnhqkJICa#7Cn)EW!=^4+b5fYO18-^Ur#y3l
zU+Ik1IB1}6@#twtd<)olrfxS#ZO(z7If4LhL9xpRsv$O2qgf`$A3C8OA(Adr0$he^
zgX_0yc!PI7s)(3IM*D_b0IwQ0^u@IS%5BCw@s7KhWQ7CkaA;HPmUbL!dzpaGRQ6P<
ziBlr%3M3t%c&AX7(P6<Z$xFUg3(AOMNpQZ}jRp{Nx;u2tc0*;D4qI_CPGLH3{M>-b
zQaZ)Z`T?CD84kqByEY}@O=dco(6K<8zN{xvIy=gC_oI4FqaIi@RZUUC5hR&$28RO;
z4*6_|#K;&(u#3*JI(?GL7>6yB0wIS}VptcyXhv0o{5Jx$W|I`%sPtsWGy;n$@rkSS
z6<48B$yA997>I*(<1#)NyJWcg%Wxs=T4-vVcvvftW>e_jN@$+=F$UXT@f}~C9bW)x
z=r*kf(vF-dYcUIL$qTLr`dJ9Bq63%tRzM`ZDOYmtC4T#%AJ?^a!!lk*-V2!f>AH~a
z(rqd3g~aBr_@+(Amz#Omc!{QhN&4J++-2;EOM!;2L1Cx8$u`Fr)3p8?KLuKQ8?lP3
zbaufg)wt;gzH;i)S7>NE_%EVq-O#buxS-Mn=<!qK=;EX*^y5*ar8KJ+aBIUersD^d
zfR4S_{VM3XcU4X6tK9OJZhDyRi=g`RWDj+@<E7z=>N;EjNrzlDS37=-N_@%=Tz3Bg
zkaIu%7pea6`{UMy#{CPx9J~Aq7?0|>b=?yhD%%pW?WE7bT|M3j#=_T``nKmB7f*j&
zGKq#(tuZ5}*mT42L?;yD#o_?{Zpi4r(Z?km`uuz_d3bPGj<4dwH;xTt67j^7l=O{;
zm51zhQYSu+j?ZFkeLFTbw$*nZE(>Gq0%F(5Sj<mP#N6ye%ubER?BsYX&Q8UG>@!m%
zF)(yJPBaeSXtbT3{k)fG|Hz9We5F4<m6P%Fi;F&XeqeaSlcrC($LaBtc(%M6%d2bg
zSoY5!Jc!3np2YI<N__m$C-IA4{34z_dKd@0+Zwy<DCTDx<Farv><`-L4e=ox$NjP6
zY+z(O7H;10Ngs6R*4Ed&(DvZ-&*RzCCyMu3+_`-t*je_gU;Q#Z{P2CnYbIW~bK6H2
z65rXG*?8@hSK?JEKH?)EIQE}~Ar=&uSDs47hq1f89lJZ*iu1PcXFi60VPVl@iH|?|
zAU^u&*YV`xgLwM*fx><mYpc&<eeGFnZ>`3`-lk+Z)tK}}3=Obbf{ASEro?vt0hR65
zh2lj!!$-AsLCWJ<jhnoyrH*Sv(@p{_JtPymCN4OMLwN3!;yAL=k%73bNHLE9aDNz+
z=A7_C`C(Cx-9Zz|I~IN>IKqr$DaUeQXl8266Xom#WU++>BQG9l%*osX$7-T8PGz?p
zyEITa!OcPiS1Ry9V(u_wab<Rj#V^TDnkX%ZYwpTm&f!$!FVaKyi$hs=<8XH)_IK7|
zPwM1oCr*#{d{p4Z+Vj{~nP*HlGCHofm3!k<HYp?26M(cr7Hss%8N0H>hKczZ#chrq
zP&4ByyX=w~i80B^VlWGSlt1L2kUpil!QxkoXVF$QO7-gg{ZHe=58jUtKl<25D;}R|
zd_Kkk5J$|-#?0)T%FIOEd-ctjU%DNe+xzi+bt|@{uWjtb6V;0+Pgmm6v*mdDblHnV
zdx{6sKtn1AEY4AOXH_?6r^%~aFd>eIEcCGxP-(HVx9`O)jvQoxnh#2^OUNnZ@A&XF
z=`2Ey#oob5tZ(jk=Mcx!%}<X@Del)!<7oSNeDv#|#;<?z<5*QWW_PB6ay2;<w_bTG
zUVG;|v3UEn7#p9CNyU$ykSxCCT*XBk?C-{k>h)vQ`IY76SXZ6j+*tRqhCa<m<&m;X
z+r=DM&QtKC-Q!(bZ0HJ?c?6d)7Go%5gDj--;rEF0wX}ogSw=J}+qw9hiz57}7npH#
z8(nGjW8T2#;)cq$IQ_93l5%8!QZ7s#0N;R&jr;{X_aNQ86)x>Ji#uMpR6WZNmviyb
zckStqe7Ts%Vqot4@|}!I2jn71PKRRQkOkA-y#s$-%1&91LuH5H{^5zzZ@^`JQe~Ef
zDKCCaPpZ7LdvVNVe(3t3@<I8=Zo*UMaRz)8=kz2WRTFoPtvpc~J5-rHjP1Q+mpj@4
zK9;zB^F}Pp&*>e<l@&Yk!-v~yS6DRG`-KHL79oAis^U96H4)RQN6H?@k8*IVHpz?o
z$RIn5+028U$xLKv1N}}^eh`2Do?mjlvTLm@uB(jju?8Q1T<HCKaeC;5IkHo^sfY5>
z`6FY|i=*urQctMn+TK=stEW8J?OSx|R{2O3H!q$_HF%!92+~;aHWaPc1y`9HKGL&{
zb@8nZU~X**U3Z}04ty2dRrbriTKF>my-?FJ!MywIUlqO%Z3TRMtuSbBeuHrO0Uot%
zeN`wszW)B75PlUR8G*ozRv1^B{2LEJEs)0Hj*FfANOOZzVTN8%G#5gft)Q&*>4#>Q
zPLB&~{@w7|V0wOOHi%ZCnCOK%QKrX366GX>q65bsyd3C4<AVt%@X*2+TK<5^1-UaV
z>A~d$k4)x`1_S=0FvFAslA1Pkf)l9m?J6FoXmHq}OL?l@bY)h|KN-C!<p{L06w-X)
zZrVa`bRjaKQ$psN_cTPZlUa}yx;i0}G>L0oifY_w&=)y&K`$@A^o6TmH>Ai(Pp3}(
zS()jKDc8Y?n-{&*38-7R{RPzNv2Z7P`Q=O8jlN{%Yd^lsc=3W5BNtnO$JGho#oq!u
zSTgOays7g<cQq0&<3%bE?>we}ue5m49+n+LKsMCv0L(99D+XRj=0_>a_VRN0GWQDN
zSL@Q3!LT~s-}uHiyxVJGZeDg40ZAkB^t#ViUwy7l^8pt-D=nQiIk}G5mZ=6+&5fM}
zoQhDM$!~pe0U7@SkhuJ$n;kJ$g31#F|I(Q$;~01L%+Kjc=Sm$Nbt>4=cU>JGyQw_*
zC!Hh0IAV~aP0;B`9Wr1$I*>XlM^WgduK*}xEQGQ6m&aDp@ybP3%J`rLK6DTf4Dsn)
zusG~8C^;EavLNN7EnHsdDCR<48!tK`d;yeSWC7$MU+$T&!pVV<!lDDCju!QW_?|LY
z(EyzRI~{q~9~hHYPQMz6(oD~$UJCLlB1k2tzaDh`7M=LhNgiZTPiAn$n>yu1Kf+Qv
z<3<OXg)fdgq%OE~EB8`L6ql@bplK@L<+@#Dp(cT)y3v+64iFogkLnk5`snZnX^@>H
zFGlmQ$BIBNe#sl@bd;fU`7k|y6DAZRzHm2<_z@%7<eql$j+2Sagd|9JP<BbqPJa^;
z*K|w&g`ND@_eLeHX=OiPu$y&;Nv-rznEIKckeD1d=Mz*wQe^d#iNC#=x=CX)1I`&w
z+Sp`>Gf%(MzyX#mz14s-UkkFNhddPq5F}|K#z0k3cqY)(zU_t%Kg)@k@?<-46Qu|&
zirlL|<Ema9vvrXfzZfe&*}*jVN8FLwev`=VdYY5tQ{*d5l7qeYk#6Z)%|$Z%$Ap`G
zJ6!!FV5K)d!YCPHI!x&S6NVzYz%tulyE!_2(w%gue$s!%$tJ7$)O5garLiBWBC~;j
zid@y2lrorEDPM#QsRJtRRtb$=f9DNGd#IL_!2sTJX1JEoq!6dX8O~qt!dJ?JoT&q8
z?8HS-Q7u(3*njcT<yE1JBgcTtbfgJ&rPzw0OV{8!lJwK1-iFbHcS(BiMQ8!El?Tu<
z_TiJxxZ=Ivy+Y={^t9OA1!&!FUJ)MvbG4>9OL{dYT&5|uV#i$MsBW^#r9<2Lt*V<g
zLNv<Z8;8|gQpX*1pSeO84{!Qc6^H7Ui>_*+i!UfZW?}HymDu)6xB}__a(G?n;#>D>
zhN72qKW$Z4!6h2AO8%(;x$*0xg}gC~3a^b%yLH^LG|=MuRTr+pRRE1!NjWQijjj0Y
zD!X;eUku4JmB5X)6#s&H8Fa&RSmj;#Wl%!61U<ZpYV3AzY3K~_E!p*J^flSdpnfM~
zzBBIJWulicfXcl36n-<6vJZID-YDVZ&JW4?NUSb;IWW_*7{ub&K~8|~?#BM!uJ82O
z+}MhZwRKNAZ>+CtY<C<Sq0R&~<5Cp!s6HxOVX+ejnVDqd%Y1)aH$LWx(wXTA*>mD`
zRO3!|?yzHtcyf0a{pQ;L=gynKvGJIgo{yP@n=v^vACuE_apTtQxOw|tOwG)Babk6C
zBOX0^5})6H;6<w^j~~VI@{`!v-tcZY_i5Evr%wfc_@13>ETSZ3@k>~Z!zbg;-McZj
zu&6W{h^_6Ncr0EVgUGI}t<9}?@bE!AdHOUCG!8r#U+*44M#<_sm1bw-mSlVN)z`cL
z#;&*N>1mJ0*Veed==1pCgZJapPd<s2)s@)U+K!`Rju_-Px@F6~F1j_jEiXTdr%xWo
zqlcfx!v|8{0rXh;u%bLziS5mG9~-y3y{UXzbzJ#i(c^oKQD3NhuwXT=^ko8<3G(hc
zFeme|=MESau6)0a@`ZA9rn2CZK19oSeV-kEDhr3m%eYWskOpUpn?^b@&0^H-RLspz
z#_SAtzD>jw_XjD@xjT*gr=0HYZYwWNy&y-Lv9n`hj71ratQ^b5718<*IF$pB4=LFL
z8vCi7V8<T|MjQjmZWR{m*!f3!<%3OjN?u%?XxzRXPal68pMUhrxc}M5zRPW0<#%_7
zvc0c#+*G=6#`ZQx_#e4Ua9rNjHb?stPbTpBs5o~5ktY5iQuT)P^THx!Tk&N6hDmeY
zH5@B>e8P1T=T<Bd;-7lxaj}4fPh@7nfiX5aDc3i)V@Kg$TpLt5UyPCQ887;>AT+Eo
z`~b%}4vs4iF5=0v)p*Fx!slx$ONz((mhU3ll<XWO$M~H3$<cK43-fX7#*MgfW6?X%
zri9}oWyY}_Q@6Rf<6{ij@yEghJLpcONO$sy1w0>lIy9iMKf8L^Id~o$O1E9f%fcu5
zd}D4h#s<&h`4g4b&puF@{Wy;#l^+Tockxk|UwPx*c<rrkNUobPH9Z^Cv-4`;hGKJP
zFLqc2SKd%x9zA*-D=W(`Gwe{L9uYR9G0HBg9s4l0%Q8;9$sa3~ZPMJkyi`{Q203za
zMCFH_O{%v(8Zvh-F_*?T*mEN}Xl=%`0ClG>fC2e2CGLzjX=l}z6#D81NZII*F|U_2
z9zVAv2K&KpF%w>`v$hNGHXJ%xC{v^dZ7CnkbGNt`w$vW6=!AT@)BagTcqpQ@S9zQ(
zZC9=P+?#gc1^dmdEiYzL7PfcyT(;J^+iqvuY0L+{%!jyas4Q_m-Q3KqKLB7s3fV|Y
zuF7ENDo1BhKG+SMcbr{Q89t9~l|7DFW!^^OslAwrMU~Yl(b8V!d=qUg$J}zv;58q|
z%MrCKq%X})t6a~<yxJV@eq=|i-#yeZr+JRg`9PZXpE)dm^5KP11SKsEnP+@7&cZ%Z
z8Y~L4OZ4pcAWqngTNb_?!891i(?o+$JBHL{Qr-hSb50$-(xV!h22cJMcW~yb!fGq{
zgsvI7s$;5dJv4$|JSx*9$FW!Y!`#{uy6!-~9r!A^tL&G3weV&Bdx3OIck}&PetY5T
z&{n|5*Yc-g$y|E+sb%!*L;C-E@_OOx@BH1r=K($4qar?4q?q(9Jkh}VXUCGbp7cmp
z<0w^jPkTag%&}~8p%9{()KAFal{!G3%{o=^;EI3$KvY$DyJ%wvVLfRL5}+|#JbMyG
zFPIWQ>z)mCPW;))ZVn$h;oz14uKI+Vy%>kx6KFPpOM@}xVV~dyn5#dBW4|&>Gq%zd
zLro9<bbM<_rmdz<_>IhkA?_uFG&MLqZB9G9c_k;S43l4m56IS$Nv)w@zSHU;bH5i~
z(0F?-jkYp1eOBsxvHFEa<IGnKn1%J%E?!7rF@sJP3k!U?!WQt{f{Yn@I&gI4cwr(B
zUr_M%1ZNz>t;U5S>X)SCbuh&(U5P2WFnm#kO?3WAcV|TMau)O9kzN>jVbO^I5BMua
zI;QFjww;O+NYTC&QqZpe$VOa>^$I}do40Pnx4!-Dc=Pqwy#NGMvFK>K<FA)61HgR^
z3{y~2ftOO&Q+2m8tV?obyfqrJ27Pws69@-?IdFw<Wq>X_NnD|zdO)7>X<f=MbQXkE
z5b5;fKyo-2Rp|imV)vrxh43RWp+O0qW4?yW!S6|)Nw3a<ceqHg!{<!p>{$89F^6;_
z_%e!)GeHv{FG%qf6@wL-i6>uHB6B@Qn+_O@X<7b|UD2cCrgBgCwUY`^u8|R$YjWim
zTSJ+m@=MvFgU+#;ETno-O`W9kGucTu(gOiWo=(D%jM)x~6Ln2#MMn`}GJ$~~Ul>v!
zNmJ7Ax(4;62W6J{Gx<Q;xLk{dj2}?vorSMi4akv|^kk9&e|3=}1A&}y(~6<CV=fTT
zW3|YBlrW&clMLcVd4b$)AS+H7DLA!m*~!REh2$X@hlO6LjBV;=?fj>1QWK-|-k~F!
zud|GsgLk?o4H#-kE^yEZ%?>I2n6NRi2%V;4RK%A`dY%qX=?G-oB6`_V_6$e9T}{A~
zcZpR{&f6?JB%fs@5WwLj9bl8?7M@61nhu4+O0U8=J_0i<VaZIL184sj(?LoWlePTF
z?`k%fSUC8|<QO}S&{kQKFOYGmSoyPug2SKvBzdD7hplxXAJcDQ8z1JKe2fAM0ZxUo
z8$z1eXW}xBEq$s=S6%#c7?n?&tP5c?4#rm<-o?MSYs(~}x|mieE4|qRVptbH;b{)~
zr=&@r98g)#6WS1=tg=-|3CO#`$LS^ojkWY?4$dEwm6My(!qZkqR<y#qZQ|#4CW)M9
zxMk5UXm}Rh$>pzQ;mh$Z;PO`NX|}28#6t#&>cL*38hQ}Dev`IBL)QV?2XiV;_D5pr
zPg_Mb>~SkR8eWk0;gXJoPWpya1~0G_yyH^vs5Ga&LTgr@UWso3obi&mF?aDcw+RBK
zNykFd&=nbN?%;8${W5U9@zHy?6tJ$JYHsbc{oV9wdAGC$RiMRZ7&v3M{9TpghAzI}
z-o;fO2^HVccZyJhtEymS&Nx}Nf^X^4T#%PkagpDi;wq==zf$x<i*M<=s=9U6SE92J
zt^JGq_?_=x!Ca#2#;Di5&#8fyrX}e5L$%P0ud0=(59HRUetTDa5tyVmJfpwBUq5|6
z<t6t%)rCy+mqVpShtBjb=z}wcI+o)2GG5>+`|3M0xvWs|;(n7CzXU#kKw;aif7W}^
zb>zXFKDWn{>fdwP!@({K9*42HwH>SL8?nB%E433x$A@}%oyX{4#P|sJ+m3s}HT!?+
z`-=bZ>3QsM3WF3Mr{%Hh>H|Wge)i<(aNJy&jvGsPEF4E?u}jGJ6DgjjlJ)T9Ja&&X
zZoL@rF=)5$yc%!5{cgPb&2Po4Z@d*ZZoL{4lXG4W8y*|=v0J-)yN>&tufOV@SR4n&
zn1{Q=*4I`%M!qf?&m@Mr=bq?4;0~yZ>oGCC;6=6h`9)8Vvf#6_vf`uTxEF`lD~p25
ztIuQk`SVy?Tk|5->hl%fFSNb08?&<uaqsS{v9xr<>B>UU$;nY{Z*RqO`C;*8hkMNS
zcVlC7BUYY2i#72(*gvq`jE~MV{yAY$N9lZACV=t3zV_V5F0MR#6ptT%77y-!9G`#s
zems8gd8|Bn7%NX7#^&m?Sbp{(jt{nDW^zJtnRA(776JLGisUtAfQfQ1-mnvCisQ>B
zJXU0k!(t41x4(Z9hfIR+?J0iSiqD~U@Gv3FPKfI-PQ81H`>MGAhlMPTC7d4T?lt8n
z^BNixvIA~KX|?uzId-?!VsBSzy1yCwdmC|buq(MvyudO)&t&$D7r*!jl^tJsCs*#U
zVnUjBKHvBGjzZpR*UsfW8gnx<F*7?Q+>z4glX&{%zS8wYEZ(>ybywv`<Fi4P*U>4Z
z$Z(vTzVHIj+S-QijM`%2e{bK%_j%!q`4p8`@<mCfaUNqsPPk!QJuJ*M(NTuDzewdw
z>9v1w8hhMlBz3GZe0-9-43DKab%&jlERK=JEFumJaYxyNn2*Mbff4Z?SKnpG{0Aw&
z=hxy$VW1=Lq5K`IOy~VqN&^-!MaxI79A7v+J?nX;+qajzWAVlkNAPpsH;cX(F3Y=n
zhq1$5c#0=;MzlnXdr4;&66U6-VoCWqJ)yDrMb0Oks?M+hba!9%`tZbU4m*05<|krw
z;5?o@_&DDG)lcH%58wBZjVFwg6_1gL=~%dVFJ5`$8*%%!cVb#&<Z<Qgw94nu$ark-
z9>($-#~H3mty=Bu?zpZ3ZFyxk;E3l6$P31fepgc#u5&+=+?ySa$|ITixSU;vIgw2K
z$~|nO+-El79eC`ZgdsdVEOk4JP8(Y7Mu<n+nQzE3XO2hlD^Xn5OxpqG^^mc_i%CC?
z-wlO#(<+04skx^crpk-3ne)Ie<<j(WBkuHl{c*C&Ejm7F|7i;{El9U(Uffw!o#ReN
z=Iv<T*!jz`r|j%ySJw8<p56gFZu@stzQxiqAPjd7`uYkl@E*s=zzfxT7N-Y2Z^E&s
z?7-wMFOGd3Q661a{_)WP_wKR4#z{TY2|7;X)tK5N=cN=Mp3}x|Zmz36R2zAI8e=M>
zHy5Yk-pyIn_uN5ATgM{w*f{Okxbl{TbdDnB=)&usi^-3!6$}aGbmm8%^6{_grQ*D|
zcc8k#j$Phcs{gceXGikGd=%*`KP7`|AF<|Ljsw!sxIVq1wLROew1c-)X>aH{^px?6
zH%XUQ)srV+_o7#Og_JWP@u_SUi>v$q*t?jzsMm#lKfqVPeOXu)GX1|=7=Zc6slUDf
z>2|fhm%-Pi2k`NA{AIc#?`4_44rG{LPu^Vk`a6GL1N!uA?@lN*N{126FKwj(^rQzB
zD)CvG&@&>q+sTt0>13mjE|mV~&#Vd`S5^F$$sT^zWi^dyK~)@uJVP}tZU!n*n(WQ=
zQqUTY9IgQt-IIEDhf6ZX36<15(o(8#{Y#aw;9*<F13mdAq<7+5(tu!a#q2)>#hhA6
z4;jSFW>}Q)GB3cHk<BHL#9<eB5~=(OD}yC;BJzhe{hN1*i}S{A>1<mSsGw_l)GtG@
z4k~SVA<xgx$L!o(E<mwpA(bE^@u99dBpa9@28XT}u3j3R>}MfEI*Ux4&Fr(`FY=2m
z&fWD62s+eW^w7&mF8~%I*xJvR9(g=n%7}8Ry+DqdA_R?2y7JWnvVlyM>5*n$q%0WV
z&evG(RH^ggd^VxDOP({mP!tI(er4?T<p51pgntPXjzs6hPWi2q_9@`y0O$mL`&-|R
zZ@u@P^T7>{Wilr{G9O7Z&VFa`;d~H9ExSBRs@Vp#&ore`T;+!}@{XkxG|QXZ;pb!~
zG1S2|J$4fa>{fJCVO{GBoebpV1!>-5MmxpVIUbaXma;;b;0wBOb!_=cha)%W!!Y2b
z<HJH$rk!G`PWhPzVZH;vWs4nL`K1vHtpww(;w*gewUGw4lw;zN$7;}7P`s3;=juT5
z^-gvc+3}aV=km+YTzJCnbX6MB0q4sk;y`EAUxvz!1zh4np5!21W!r-bz6ujh$?wjV
z(jymH>`gkRHg9X0#0?y|ik(iGj}ei7e=Vktp%>)HHz{PN?m3d8cSlz;Am?;h_;Pxe
z1?r~5;xJMI@OTI8f)E1Zkw*{c>;&F=(oj`EAg@WFH%@lwG7z-^)cfR-4y_D+*dR?F
zGI-h*^umXq{ZOuo-{L33q`W30WU`YUxH%n_9+(s0k}~HGKzxfY7@KQ)q>gfU_V1K1
z0xOb$aoAG!hYpnqd6cmA$i;v@kX!N=;Egm8!kh^QcjJqcrscyCLP^IaZW&N9XIMs2
zUJJmPG%=qlkF@$8Nq^{Q$N~ZB&cDzLY=XI!Ek>qP`Gb9iNB*Oyy@?mT>_>`622*4+
zF$NR++XkaR!o!>(upO`s8e5aHrAV9I@?m=~9VUS+H9Z^#^k`rRS7uyYN2F&zDd`HS
zAz38P>c>3H3|g5?oc?e$n?mI$I1JdKFI+&`tH9x)OZjzK+#7&_(lhzUNEnzKqH0M8
zLNN!z6Z5t|mSwP^PZFE$-oaHlP=A$oe0}Tt|8CEYFCf#5e9HFE@f2hn)8uY!V^45i
zX9Ux!nw?A?!U7f7=zaWv)=F7y`FC}c>(NWx6Vkmx<Ib-|3!qC=FMTlxz@tclYObnD
zE6iyu_=Kz{)t^n$$^2?hZ>}Q+di~dMUT!wXn7CrKxs=9+HuW^y)|eBpxAfJmL<@&T
z)iP6kH+~vS2PajbAD8A9pU~1&Rl@@+t{2RtQH^w_MeFl}xv;Emtvh;)>v!*(EA5y0
z`7*kSOX#Jos-=}pTGtL#Q{peXLs4;!YJ3&IUni}*Q2o3NC~iHxrLz6*0RB6w>r5S>
z%vQf5C}h?Co5}T6;7fg{Af2#~1zmRiQg&2nL-TJsmNMKIcR#Q6UFomVAJ_ON`*H&F
zO&<Bql)3at^In+5*S;KmZ$6mA>@nr>aUA%5C;HbKb839Kv$f@&MteIuo@Bjs`*zIC
z&WZ<L^J}ai{l)bm$u=M!8gq(wev~H}rOr-{y?car1LKb&jYT<mgGo<zg7H!4V429}
zPOihFQy;m-F?8qZ)0|zXPz;D}O#SVJn{o5bYjOA0*EEJ(P=@o-S;XzTcjC8x>$l?d
z*IxHY9`rlc*Va6aBi;OwUWq&L;^?^ToqfqV7?YDTv2^pc;&mtP+_@94z4lt%d*z;Z
zz8W{~+>7bi1<UdI{rmCjU;R2h`Q+pH;G+-Y<BvY_ZWW)HadPIQ=Lof@Pak{Hi*b~X
z43q;qWjHSH;)0{~*xA6sTFrmy&Ne@!1ajlUJdVC(C)@E6JG`V0_JX6qHaAye?fJ7<
ze)1@`HrHZI<CZ(OZpQ6fw;WeKR%8NtKrda6@M4@cK9cwK%uG#r;ftMD<H`>&I5Cb>
zc-~Q^aUbI@7EYLu_Q#6Kqp7Lsn3|sUE-l}GC7jEY<Q_n7WrXJ<<d^XXWtxSa?cH5p
zA;4X694n|y^I`;i7U$<;PBINiPL8VNZk?fl^H2=Ry<AG07fM^@CwIYNzJ84(EeB(4
zgbys0p1a$zvHo1?vKOP{6Y<7d--vI%_uW`nx}%_2AianK#-l2mECjJLi?Jg|wRty|
z=T-Pfn1x7G9_v{9o-W6#BiKnB*O7xG9}mg`Kz5t?xHOM}93549$(;Am<rtF1;~k{~
zC-Yw)9F`n;x0qPRuz0geW@&NWlld%Yu`6*_<4f)bTi_0{ImVi^G0$$BSr(6Hd}Q3*
z{A?`BAAT2Q=NLMF2us;fKAbUzmw=~=7xAUY<()~nm^V6_$0E*BCR83EK&~Uzk%Qy&
z*gsNyPnGs61G(TZ97Ayw4?g)I{_KDJ|HbDYf1q^Vb=?{rn~GVb`JGqaj(cx>Q+4WY
zuv?Tl2-Uo}HZ&5uho`w??fIJO=Z1G3QRZl`Xj_<HuzZ%0`<@0RQ?_}?Mit4nlSD_C
z5}tC%eQ@l&98(#n{ZR6QGS0#XNA(d0rXOf8bKFh-l(925`5}!`olE7ZBBW)eY05#c
zo<G7DUOSL&EQ?JAA;`Mk0^wAJOH<y6EB31J2fp|L^0-IJ)UlQQ37_)i$L^l95kBW%
z(32zSI6giaOADNKG~_uTjtV?d`DZ?ndcXoO^}@#=Ztr?W?C$QK+BuHI)%#%YAl5cE
zV{Kjex2OD5Ss5Rn^lrr$N_XY~KmF`c+<*Agi)S1u%AJ3s3eWEY@h0E*4h|G1?S;};
zd7%OlgDU%jdM7AxRW9-x6&CC6-C2x<*-0P&>IJ-!QH41XKE6`C*{R6R(>c{6cA>Hm
zN1gOe=~5Q_)2qabcnXfZr0lWi&2DEF+6FF8<U|`&(%Wf{N_$nHti#i~Or<>dy(?&F
zi@DVcpK8Zk@s_RMJ@uq1xQ-6kz3A0mp_ZK%M|vn6z^Kc90Q)Ispx+(1%Fk8sWpvoD
z^7&QyRcP_u_$Iu}&lkaO<OlUN{QV|jTm{MJ>+#AvzW&}H|4AdX6a&th9f{Y&g&Vhm
z(BD)pU0I>hspGSjG2Ch;6rT=-Kn&7*!RzTj5->ACQY^_s&P_PYBRkLqVaTvB=x-+)
z_$q56i53k4{M<1vv>6Um`jrK^;>-9#l<w%n6`V-XcsTl|Nic8IOO6)`6I@9T%$A=&
z7C~wy><%Ar&}1&@m(HzO0VU2zD>UoXo`k6juBULar)(864Djj4ZtkpQIs(R<K=iy!
z!AUU)n3+*0X6Z&uvQ1tu04{ojnZ4G06)<s|?>gyfn7u2dbYR#n&n_AkneuW^78Y_a
z<Xtm*>9Ir1Tex}oi3g24ixfT=c!;~5=%OMcFI)m5)Dk|zG&1EQ<A+5wb|b&^N-2wC
zTo6genB6#h)y5zIaIVTLN9>*J#Yj&>Qe}?=a9C-4DUke-7cW<VI~cqyMCAAY$XjLZ
zz})<N{O%8a;DHSbFcid01D+XgI#Br9+}zL$f76`+(@-Alq9|aZ6OJFJNjD_;p{Gow
z1Nl$K*_0*BEGs#+?t~!I@lu(g_k(d6rUxqW)9`V5BTp2T@<RyaJquSaVn`i>*@cCe
zpPTiTf64?4l#T=8qU4t{<4d7c#`1_mb^PFq5)EHP`s`}eRTexMFcUrlT`#<tRq~;{
zveVCemFL)fR0!ohW8t=fw8`C%EPAOkf`6Bh?6lE|BR$-SJ0*Y%b;S9qQk`kZ!2pUc
zN9&i1$ZAGT7YCFRu9#`$U<R*ZrTo@53;Bo-GWo)0cCH~0@ep5CHEGl<v(O-e0pzsO
zbXe@^+N=CHQiLN=UL!B{*b6`ci-r?CsQ1uFCf9e6fcY1@Y5X@Zh);U?bDdILObh>P
zkFreIYWJ?5kK$M$-$``VmB+|$&p3#tniWP94rweP4=sKH%M5SJO&yUP8HMuXExd7q
z7929z4JIxTSS=a%4De2X&O8d<erlkIDvdOE9ORU>K=hWy(%>VbVpD=$3m>~ffZ1(+
z;*&mQva`{8n76=@APjIIie@^Ql8A@E$n<V4pI}WPLrDgMlc@=``pEiTvBt~5`;xo?
zGA(UN@206fq-Vl+ApIpM1o=;A!YbrNx4YfUUC*@H321m%H>mhky2P6c;h^E#Fi2d+
zT}lSD^k;+3=0=$?oiL_FrF^ET>EW=_1RX`N@lf!o3)0Iav;n(RPx2KOmd|cA%+dsi
z*xdwUATD&a7k^d9Xnv%J>R~ZgH#nQxR$?WC%fBi9N*BJ~`+j@WVJ&Y>ddu#*C*`)N
z8xNkQM{oTC_UBj@=oo=6Er7irS0S;K-^=b@+*RRK_tX>F^^@@^reZ{g2IY7cbH7{K
zX{(5<ZTT!TX_nq63(zqddfhtwEmtvj@W2<E`GeBK!Mgw#GnJ;|O`C3Mq(YI`8J5Qy
zI3#hQlV0#GkE#n_DX{}a7k?e7xGw%OT}#vJhMBZaKQ@3ZG`+M5z4+<|kIX~k?S!7n
z)u3ZnCq;Tu8L`yeXSGSFHXQS8!fK(Prlxxj6zgT$ekD>1+WPOO>F1sJjQb@Jd9VNU
z+l%XsQ@>kANB%BH{SvK3??CG&`7%X2O5hiaRIz>$bi=t4I_~{6SNUkm3MpFRR|s3T
zE-xj7<t=XhB{<LMe5#*IUl(BP#o|vd58>5^t?5C3+DC<hV~p<~{ke<MhaUDsGyQGG
z1wQFPX8O|5oeReV>+b$x93G#kZ_G}NVfE=>__#Ih|JvT>$aHq4oW;r6MI327$cLAn
zcojd!q(`S0vAKT|YogueD|;V%b)o(n3qw+pWkmgWWI5M3mPIh{cu|;4m~+q2!SRLc
z;(L8a{eF&xoAg52(TT<@iZ@5nO;524XfdW{XOzenaj>@?TkETFdVJss(UCEZ;=3+h
zEXo|l?%t97UiUp?@4WMFeEZwqinrf-C+@xWR@{E&&A4^vwYYiv6^$R~yvv9MoF`8n
z$MW*ic=+%^Jbdslo-IG|@sJ$V#?CNycb%ObOCpVr)c@z0JC2pFA0sls?6C^q`a<4y
zR4uc$9yU&Tc+AfvHz%xI#OtrU8bA2n_hOF4$cs}iL~JO&$47e_dmP8n!CvfcZ-~`-
z+*q2Ah56YS<2bvKVK0=hi-$X)^60^Gf6xf`>9E6zB`A&chJ;soGr>JPIxhTx@>He6
zWtg!8!YQwLQbZ}LDr?+RH8DLGGjmHZw{Y9{Ro%FCHx?If3ODP8q4noa<FikGE%mE-
z_UMy%`sm|$`tZ|OUwsxk?3`P978|Rp@$BjS_~?V5$EP3tI#!h4E6<+9gZrO)R~doc
zd+m*Q>z(h!-Ft7v*yLQCsq^JA8?sTxG``~~&QZxfN<2roZ(8+2ZqyyBmv=!@zodKd
zPI%@72IQCVFL}T`8TE^F;sg~IR2JqYW09jDXVcDoL=)pmM-6w`8N)G`?5trI*NY3K
znF5;_9g693jeTF7#QfB_7q5Jb-N=x4yXF06;ypXz-F5RbGcr$mp^zizW~Vq}GVfgC
z9wzSIqCC(}AS3fkEN&2Y#s^-s5x8f@xSP5lcNQd=4`C5tPjz@tb@xPd`+`MHsd0rd
zH!~%9)JB}_#Ql$d89)B>e;Lo8KJw9;cW>W`ci#P0y#DsLV)6EyF*$!HMyHn4MvO!p
ztIaq#h>M}II6A-X`^>n@Yinb}3y35r$LWoaac>~^It@!Uc5rD-t@eq9W7;3o#f!S@
zD$75MHp~MsM+5jiIq?|cjw$6OAIe@==_QU+D&vz%Q_36lz>7g{YgInTYlQ<4U(!nn
zX*_1qleUs^C^+o==<9e9MUd~)1X@@A6`FRrB{8n5uHQP2>jD_)>in!__Z-ey9I$w#
zvgK2G%6kDFWlm~Ga(7miW#$h!8H%F_S?uH_C{AEuhv5l}W~4Loh$YUnS1e@h?CyKV
zAjcKb9&^9eK1X2g9LD<AUhJuTW`T<jAJ%t{<LTO7JY3$0r^@f`oh>!PXDVmcU8Y%h
zW3GktKbCBVN9W>y6x;jat#`zr%F5)V%K6-!+cfIM%^M3bq5Nb1A?L3Y54~UNA~y43
zxX(<CsGdxyyyosc?<!VX<|8jvR+(#}-m*wq>ppUG<RLr0h6ngiK&>17RsC{2Tz*Rb
zZ0AazcFkAlwQK;rZsH}~ChV9?-s0wek9*MzRP!ahYwpljXmJ^SG1f<qU$=?2weZC}
z6Rx5GzK9NU*Ohuz;j8kiQ1Pkv(!9*i7r}4j2lX}l{U%{t1y{xS%ODrN{w{ZXl>#VL
zhI8(N*o0fQCQ>w5c2FU~iaS;ncI^0;UMG^^nS4#!fL}rdX*6M=nFxHE3m5(2i#h$;
zgnMzu-BYlTCg<X2$yz^pW=Fj8ARZk)dhl~}vDra1Qf+#G!Bo5e#jX5}C|=E@4Ie0O
z6+z+(;a~9tittREnDWd#z=rCl@ry~8qyh-Jq({Ld?`liH517WEh}6F`uEiU{->Rku
zw3W)b@z?#O7rVWdmKMFOm%)muMDOsCqlA?K2=ABMxf^6&!hC(=;|$f2U;*rmg#{XG
zI-AnHJI5WHG}Cz_w_rImOy`ITBh{%;1IxR$`?5s_pn19@i&3)i0sw`X9}Y!kQ`3PO
zV6lh`j1Pk&8@$z8?)Y=3!M#u7!9U~ez))WXt(cd=GQoy0NQ-`G1K=yFciwp?zWI%B
zc(<3Mk9_=;)>u*Ab)+<+L$OmXD6VN`8vInx{u6qA!IkhF`J^GfRZalE8Ct1t_{C9a
z-VmW@AGHP3V6YU9F9<GKW%=xkFD=yxpgi~`TGW6#edE(p8tlw_@s-_HbY_rYgwCZ|
zDh4HNfAy(+yiiBQooJP*x@f}}S#(%f9HG;|E-UA;oh64X$m4cVBOj6gvlnZGqH|4I
z$sJ=k5C^hDsZNYKFqCUVAxz3Y(IgCHL>_nG6;Dv(>M+yUMJAu^tZ*1$ldAYh9ZIFO
z!m_L~Gcn_kH83(>GJ7W)0~=BJ3rBU1_)3$(5B13l&<Z<u6w&$AAbXH)`V;Cz=Fz48
z%FG-n<CrLuncGz+OhD$MY<ieV3~@>SLK0sqj4nmRjxSqGr|8O@WdU$KEpe8E{nuSk
z)i2Ma$<R{NLMwBIgFH$f6%5D4ej$XWmRW@*bmmo22qZTTAOQKRal{N4`1v?8_*T6G
z(3m}DdjnH&qJ;}TQafdFitCqt^MO_jf^_4jSjj>WHOSDWDRDJBmk*oqfiq#SiXmL}
zo3hl9LGY+HJxz#=)(SuAa`HQc^uS^~a-^Gyq>7XL;VLr^nVepx&G^{0c+PSM6I7!4
z0Hgx;grsZij%3BQ;MkP#aK+s*Ziv#|E|sj9c|Z&J+4xCkr$_PYpP9lNbEQhX<5lLQ
zFaL(N(o~sg7gZ3AHx6JhFC%ri;>^ncrnLc>GmlG{#uYQ^0!j)_7~QVegqC{s*EQJ#
zHY4C(tc@Wx6t0zT<Lq=RAj6M;`d7H{^@G3tC!Np@mPaa6RT5J#za}*K0iRO9+}}uJ
z6GlIDWwAScOIzH~!_yetvaGbcI(*S<*v56u7>e0w`D>zy1MNdA4Rnpa5=OOGbBn`&
zpPM;#Y4RP$Bkh&8ln!vllm6)rn8y7!{{4QH7E~aiN^{lV8=C6J|1_Jlp{qFYsp(SO
ztayi{L-rA@AmagbKD#{IcD>qD>ESTCc7E2$^P+^14#$?ZSOtD_n6C1ukYk|Auf?}K
zx_Hc)Cl#0ORgNqCba9GDewm)C>y~+anMYS;_%%!nt(jjBbgd0-8@5%4$Rw3lQN`s7
z+{>WX{mcD0W7~q*s8&moHeF2L38&&4_a2<B)?4>3ZI;mr72mI326FYYQ9b^k09EX!
zX{mvhw$ca?Ne14H^yewr?gyW%PwKw%xyn$=O>N0OfqGl&lbO!6^j+0Qq>m>%<1+yM
z@MB!a5rM#&c+%%}pAJ9b=8w|!-egC)kHOp7iM`!D9}%^^yC17-o3XO87SC5#V`F2}
zM<uRH=N_R0(ICU1#-P`SN8|A5%=Z*6FR#Sf>bl(bya>du0qzpw2s6fNqhn(+roKJ>
zdv<=%PiIWaF<I;&BFwdoE$=|^ZWe{HCtiH8$j${8Dd4fRxZuT!-R+%taR2jo`siVB
zSJn8$s23Snz#xwFPuUGbs&G0BJKETFQ;$iT))?xQyLaP_H(rZ7w{Cg7u(iGx9MQJB
zw<8uW{9)nD^t8gAQ1}C0xO1{=EX8q^Cr3w$|4|&51wX%MP;Lwhbc_b<R(iNB6a*S3
z`Q^?j#!n9)-j5GI`XE01;QjdM!(V$*=<%cb@#x`$IFyXjQxlT?hIjZ-PH2^xF!udN
z#8c_PA_OPWumEO7KFE$6X~A)FoQmO6rt-vxw(gxs9@3IsL{wjnpxoNrRG#cAzIlw}
zHO7ISc;AnmjWr*E`1t<E@sq#!m+`Zo{#ktT(fje>^G|)x+0!SFyg<TngioJ5jVDi^
z_};b8l!qL{$-)fBr;*mLz42DO{mwVz-Yaj%^xTr<V8@8Ym;ksyN572EB>SY|J;RA7
z${W&Zn&T0rCO9fh@$yl5N&~-bRL-e;<R5WmzGi%4%oFF=uI1bSV<Hw<nZTc%n)I=8
z+&#oj6Bb4(8>9`p+c;KnZ=WL&RbP>5V8{!OEX2&uP6v0LaoqqD@k8tcV?zJ<I1UdE
zeb*H`UKkfUt(9lg1JZeDn1wAV<;Ccb>>6&dvyc4DJJXo!8FZPZj&Ot>ylcT>Vt-iv
z@IN9y9I5IPgjC1axyEk1*{ShZoS%xDbJH<7dM$Q0p2dg1{7L-kr$3IP{k?eW&9~zF
zKlt7F_HX|nUVHmnF)?>5PGiVBKxs3EnI9l5)%`Q!*HvGiFE6_;vQR^sjw=twRo@tI
zQ<ixjN^|-0H4Pl`s`^vgB<hRXG^Gde$rEAnqt)S|QS)Vf<AjCg!<;J{8hjCx<0DF^
z+#$$S3GngahKeo=WSmMwnEC!EJ($CCiLh?vJfMw8`H;7^)0IZ(1Q=ttInx|B`zv;r
zGr!AhG0uRu%BO&LtD>Af`L2e9wv+Nw=T2NsO8NjPKca9y7<FR63uPSZN?9LKeIHQ!
z%DB;sU8D`Z6jt)#BWC6v&(wBNj+M5eVPTt{i`P}oSb*Ywu$`SfAD6nle-gXW57@OU
ze;jWJAAb}qoiaK!q;`visnKEYctplC<um-*P04Owb|`XRJ?{e+b~#d&1v_HShuO5H
z)GronJx?k;y!m*7h2!aIjscxgI1{S#oOq;qD%{-kgzGB{$Q;|tZZQ^+S(qD)V=sL9
zmFg@hd93sRJig{H-KC@CoBP)S%1?#r1~hb}f%Vp3np?g6o!>Tqy~L{%cGbJ)23N~i
z_L)3uBAs7$vwG=)bngQpx->?>r3cqflkjCU*t?j<ChaPIz?b9Edud+grwf&)U;ioq
z<=6Hnvab`yWpJog$>!_vY8Jk-D`+<EY{Jby^UE6H!fQN5zl9pEY12cGJ5aIS=|t`7
zCjilw_daLBqfr(fISLl4T2FC@E-R(Fa};>H;D&A&{p3R>A3eW(@AAXHH9_*WBXlKR
zwPnb-ro!A>(?X`E+0UfGfLWq?QV(U`i3gx=!Y!KZ1@fq>^p8V-dQ?chX?I+eKAv!F
z(S>)34+!|7HN0uiuqWP6|HaWXV5@&P;g}{mDRiV4mzMm+7@bVwix<Br0jFiv3Qn2_
zA9_*LuQ>7^uOq$W*xsCj7Yf4}gB1^g=zI?NOAua&85_N1M)cy)3y#46VK@YwihJ&i
z%R=Y8up3qoX)cA%p6xCr=^O{_UNn#!UpCR%<!c;Xw#31UC&a@q(tJS~FjoFu07o%K
zCp?T~00A{h1QuB@XcHvu%V2SFF~0lV@5Sxgchup^uN$kUjGEFxVdVMD`IR0=9j>lD
z_e=G<0;$>Mu<@#}YwhBh9Uq8<DKAQ7&2&ta>6@yNw`2-@3>9Ai2XKWY>E`|y9U(71
zsna5P>Ch4Uh2+l;@<0r!6FfCNBQ@oCQ&#BsKuafqbG5lhHqV&n_=fB(;_h;%xKoaK
zq0_k>6b}}LywD^684br!VJi(J7hjC#j<M`0dQnG&`Q@Da(W!#J+??j>jPW&|2cab$
z>ep4|3*mcFg>uf9S&HPaI$fhG(>@YG@r3uOI<J&{*CPaz!asbmDnLDestYN%A$6vF
zY=`84*71~_M8-QQJ%E)TCweaEV5H2CjgET&#ocCfFbKo-R*Jl!e7M7^csi9}BZo}0
zwjIc}Kz1|tQ}HOeu(fQV6O~Gb8+H)rRW;D@0}*wdiB8hUp`n|`1Px>w3Xf*r5b~!G
zVzBg{Fff1OhmsR*q9->+k!c~ngdv}WVGrP98s`tZlV=j9%mnK^ES@dD5RRGtUMw+l
zRIwL);*yqUQV4Lv*^w<8;if^F^mHn`v7#r>nkNmBam835WflmZ@*oA4A_qlEU>lw*
zy$p!q_oHon^o1_n1kmc|v@)UaRTwi$qoMdU;6@no*LroT3JAN6xeYkb;5v;eU@xYU
zro}2q`Fj#VfX%#%j7}R2^4|%u#bdUw;!nS7x3rR{22rU%PzIB4`t!iVvde~1KmVoD
zAU>t8X8FRadkxlGuutK$QN|3WO95n1g<oQ<yZv`_m+{D2Rk>ulk!JL2Pk;4SL+1Av
ze}xNQ-}*rwb?A^Y;Zm+E%x&(q(7Ig)cosv;r<(Z{zg@kW8?OcDkc;kqnh0E00OMr_
z-)ByZ&?;$PxK_4UC5{HFpiPNWufKkCwc{g==*sD$jKpj_1TyPgV!!&?Zd&{%eVXxC
z{9w19D$@dEi%Z^eY0L=@r*TX2YPQk?e$}68O^3P2X_e&G8zsDOInk7=<7nQ-CBvf7
zQm?ogPNqf6t4TxiDgNv&)2q_i-85BoXbWCi<6fcK6PK!n397}n(BsE#J(E~19{T$8
z-#8%My2|en`;*mxIqlV7+B@;ZFHV`&#l55Ks5C<;lc9H%c%_j?Lf3x_*fU%Qb`_XU
zQ3V=46wt1twPJ4YsrV64wPoZ$Q8v?TWjRQ*EE84LFdc3Zekmkv!X^LcML9&}Q~pfd
zqq=@u9#EpEmiW|NX5@}5<J}i$-2J7#Jd?u?2g3B{#VtCiloO%6_@e$a<1kO^mOeRs
zQ~J*A;3Dtc&6OV}PN|!O2o00e1L||*&c|^a55<H!y8&`?pL?mcV`F1OYTb8W?QHLe
z=2+n<jG<xi<(R9B*xlKS_2=ucwXrFG2XT0E7F&CV@qAtWH|6K}*dInQd0Fp%s`n8w
zW}(-~G3p#O$0TV^iqqGWY-J*Ne}B*8P~;yUV{u|u>F^>}SDwYw$DhaA^T%;`uq}Sq
zVsu3PevP9@-V60@7&mcz*7Ve9OpXrbq^tEYA8-E3y*shAu;BFdWG;6JX*_iE=8bs!
z?YE`A5pPO=<<-}Mdw?0M&}}<co-vlNrRFD!46^Q=k`7+R3gDGT#zBGtO8)HZ?8M_I
zkK_JB7P;=nlgAHZb#*1yrIw#PiH+6gargGE_}=gQj^$=-<Gd$DJWk0u1Qd&Mq#xrm
zKKx^{nH@zlQ+baLd6E;*b#bMPhbU|CR9Rtx?-{$X9()$dPans|+H$NtdmP&vD^|--
zK9BpKyf5|Z_~fHs$MW)HrTH~aLQhOg#l)1xt5dTvDK$5LBgSPPo0yF0nfX|_c{gs{
zeIxF^`cB-q^;+DL`L);IikZ1Zr3D`<N*LuKA8E3v<y~3h%1e#WxL1zy#A0gQt;NC?
zlke>4VhlJomJ3}>m~$81?98<MjQYLH*z<%Pd(=0^&x-q~$_5kdBZDeO(pjK5Il1u8
znB6_@0n_-}r(Na*KD$t+r;&w)F4{<?nHP3*2Ml$^<<NP<VgNf**pZVvAUQ^o@-d`(
zz|Ju5*wXlQNI0^BbmZet>K(_-ZEfzw_U?hk^c>%KO=XdCta>HC92Ge^KIENeECNxt
zCfEryF%k>2<1sfi5_4liaqV<Jo_zj({PfTMc|7>^BlGy(@BeQ6&A;_W@y&1lc1$hY
zh@I1c*w{VtBJ`;0>+k^OA!2B9Hm(nkd$D}&`HGL~B+g^1lhd=aUWlYPvm26~NV!PN
z=|(4Ue#(w8%9i4$bcN8m4AgZJ2wp6HvY5#n0=vtO4>&%NkFXDeD;8XaRc@IN6~eo|
zSmYRS`e%E_;)&Zwl_LLSuJ2g4;YH=RUV$Stc_(?ifCeMvl0Tr<3%ASA3z9~9dB0-!
zH5lSy{srf`hGMRO89EeqmtprpmBOQ~&s|<JE512zBALWSSuUxF2a8QC2yxZLfZE1`
z*xBRARTgB)`*P}z%AglYkg2vq(uo4QdCwFUx<3e(;zMV4*a6(zbs*)mWIn32$>;36
ze8J)tqb1eSYv*yezZr*nn|gO0i9N>=5|;Q$7Uq`-mK~)W!FYJWUGZlzs{G<kL5@Im
zzezMF=P!c0$@p-By3gAV`7<5PoYyR}jj9~ej?Ybv#q8u*%+HLwO`B&o<tz)=Q!#w~
zJdXCaf+JOPoTvbr?n1Gn(dNt0pFc0x+wxKM?DwX7zZ-wY&jdv;P|YbrubX*FZ*dt8
zIym<PpaECyJWbQ}CxEUPfm&uO5PlE6px<8M%jmH8`^>qdtN8&}^60@~zD)B~pwhME
zU(p}xk{n+Q{ptGkAj4|I{QA7gQHTGYk2-ApyL{6WI0%e3m4EYa+W-y$K&ryGk+(zQ
zr6_FKAlW2yaW{FT!k(r?RC6|ti7q&sgii*>WMfxs!uxDw+!HAyjT(AEz*YZl;yF(I
z5Vqhs>afeZdN*Bqkb)rHj86hXYMEXJ)1}`R;%#~yDfFU(-Z*};7{ebw#cU`>rl&A<
zLP)&b3vme#YhoOrO%+@gi#y?_UsT$=Jb>ImkW>|2oIEi<x1bK2I-D$kd3Tf?d67Vi
z-lfM47s`i^`%(i=XXgc9LmVIZ$OPYYgh#!m`BKH(nt2HcLqi8W9WxfXxY#$h`Htag
zf5_?>JK)kd>R)xxCE|wEl&YW#n!LzPbCxYS{&XCVB^zJ#aPhDQSE8+pcfJQnIL93E
zaYO;wD@Zt(#o~sMg@gRkMWk+Ys~d)N(?ey{*n&$U<a+DPH+=^Dj5@OTktB9=z_Rm7
z_Tq#X;191{{K;AECmLiK0lOrZu|&ZC%=-rDpa|nK;b2Od69;uJ=^T>|?%#+YUn0_#
zz$PzGGp!1!I?g4YP7&mGdB`xy6VAyVRtI;~U%Si%=eaY$bqDDz7e(plus}tJlLfCs
zzAjT4g9hQUGnGdGkY=PY^0P36&H&W;YIjK@ITW}%Jq+;hf9Az12H|wf(=P){(%(E)
zZhUN_WOKfV8GOx@xM5d(I2ZatauYx55Ce4<(Yy%9K#nw32kPR5<aIhLh7<{z^9x1N
zR!`vvWTykkQ4bSJml1V}*@ebeesrF>yUhHQM=sx#bH&MvGnR@phc^TAk=#wm!~pRZ
zWLTNM$dK6L&s}So#mn`HA8`byxZx9r49bu46bNZ|*nl)q#w&?%&rq;ZL~4iHad3D>
z@-%n?1wVw1x$ra(^3*@YMre`fk#ch+Y-K|lkzOv>_%uOvgMX>><gc8PV+}=R1&ab`
z-if9zWgD=DN2wAfX3&-iro|x_x=IKbdC2H+d03uwZx#UU2hu7-daY@bfDR-6dSG_~
zz(D*cCI#Iw=ucQB-i6#w6^cB7>#!pvAI2Mt@5Y<o3xz&G?n-M%p?I(UP%RHYe2^91
z)|1Ne+jFt&+3`{A)eJed<Xv4b*@)7kz7x|Xg&CYN39W192!AJCG2>#ed?4g#p6D2p
zqCU~nHJ@(B^`rCwS;i$dVW3=IOFfYn=Z~3`_PbaU?2tx5a_*X2TaS*WFbM;3ErZdO
z4(MW9b5(xk-#mC253ZR1su#YT_9W0tAj^t!!tz;Kn|nbnPV@xRRvKy5+=Caq6mL&y
zGFQ1(z44O-X=?oT@sY7wp>O<3zshW?#;gz(C%fwhRN6|@()RMH<}OZ#<YJsP*$Jc2
z^@L&L6=p0Yor_z8w@TNfOMEK(kgnQb3-L;Gx=E*|U@-zb_LioH)@ECJ>=o78rS$W^
zYfq{UUAntP?1our@M67*dcm7ls>V-IEgpX0ioO1fOng=L)n~UK+i`2%E7ZPh0;bk~
zrIV?(^DlTH&8hljZE;;Z=)$W3nf>Bu749Z+5v%rA)x_+@C3JD_&laa%d?NepGNfu`
zkg~ZtrCW8cX4`o}Z#*Eq7+tPW#ZSNAR_U3V^_E{(wNTx=DjizY;yQZ2>OimUGPGe;
zTD7wDc|8WCUzur6pVzs>$F-PBUjjc><=Ck5Q{PQ}Up(_mzgW`^I*uM=R|;bY$}aCv
z#s=&(V6lng(V6gMx5dH1zQz(L?tMOp6O94j!I-3u$rux!93A^EvBMMY7CVpKoqg}{
zSy^3;t*vc;^-ml)ii-ZHKaf+uzm7i!2J!?7FTi+Aq4?^Z&i!FizUybqyV5o`IkNsN
zMnpF?#h3j9vAeY%pMCbR#&;`mu)iBSTWfJ~d=S?)&gN*uo$d8FmwiYbJMPAsp3*o$
z;|0bZxm$)shtXKNal;=~l6K20>`+>F+&HG|&h0z#+Usw`TW`G`ckbME8ne^lgm@X8
z(y3*#2pWSxX42~#ljx;GZGfgK(vk^p%7MlllApz!Z-4W>c=Po)y=x7b7v^T;-rc*(
z$Lr3U1Am-)sD7*bopEGhemKj*z_Id?V<`C`(kEz?zCGhRPiTvhiE?%dviLHr^i*l^
zk%|Y3AE%|Tn7{(ltdA;N@NTOSr5QV=4n()MwjAqgc@NsXS6_`k{2PB1zx}-*#cOZA
z7q7hjZrr@}S}ZKxjm4YyVs7D9%rD)FxurWcPt7jI_2CI6%V-Rbj4B^kgyW-L#b0I3
z_nK*(H#n$#l58C7$igW}S^JliZ5EAK2;hiE-?6nY6Sr>8$L*!LxM?--Bh$vjgPl<v
z#mS^Q3s%=P&ScT~{CF?6*Pq4O@+0qH+uzythoh8_iLzr3<((x`>^>pQkb&blSs*(&
z)ENAf@uGOEd=HPX<B{DkoM<v8oZ_jprtaa-3luW@jxvq2*-^uS1G~ERk4|HKV<(=k
zZCbIwL46$?9g0aN|A()|*x(DrDPn4TI3`AwhwQph`mum^al&rA<v85kP+T_Rq3X!b
zfBfh1%b))^b~e{lhF*!k@rQpSe)PLPj8|T{7vmFClHr=dKXv&YA5~guOwPjbfaJco
zHXM8VN3p%R?GJ0otBJ|!n3`Mg&ckz+lkKg|*xzMmV9r0Bo}gIF&-<Aud*pA{^C40|
zxcmT@BP(ZS=e;1!c%8Wn+8FogWv9-P4~+M{cp^UY)7%ZkF_hrsNBNKoUutJGSjcfc
zZELBY0*_-EqcSH@>aFK<s22(km41{%rOgELgW0rJeo32cFlk2p5>n;Yp3^-^8oz|&
zwinuh#v%L9w2!%{-V=sYS|um#CFMZ*!dwMyHYYS49CB15yXq7#WW}H3RpXXnD<t#u
zLY`<LWdY3hxU2194u+#DDQlkhAt7aEeuQ=!`g$ijyZdOH*?q?Sk9;t(t~Pp2<&`kl
zok_W5N8+%e%RH#>GowvX9LcL=e^8-1t}ss(7CX{79+$e|9h?Jdr7k$4H?M`sd%W^^
zUB-jVjnbC+WGu<XJ9K|%%R6Jq2No9{c~ptJRWVmwg=)S^&+>pd%LO=(O^d2HbpNGD
zM;pKv9jIpe?dT4ngR_-iU)b5*^8<843-p_-&G^>I(50<#6%OdfUuEz2`7&M=zKo|Z
zMYZr{JS$wy>&w9*e#I~<d|mopAb<Zd{>#GG-}~dg|0gsQFjKu(eP9AafEk?4nMo*5
z9+kk^!R5&l<E4uaw3Xgwfu~wP>*oekOmXp~w+d%^F(%FR2MQ3KU0sgB<4?K-&!P<)
zkMOr<*-WdSjr!G!2V`SeagfRkibT-gFA~#AA->Jna8ENvTg$H^%o|}H1YM?d!GwgE
z6M;=cM8aB;*(!<9@x{-75H&q`U<Q>GW(C4dm<36eA*TztR`-OR;_Yz3OJ@7^BmE`c
z3<qE`66d;ziKBlq(+Q<>JU2V%jyEqk@FWYqC@1D(nUsgl3jywG&{^2>R_rsj_HxPX
zp<m#1uEDXhtGwOv<s~yOa5||BR(KJhv&G&_i)=LhnFgScJ2mH#^>m}WI-dN~yMf7W
z$6ATO7bJ9q{nentKky>Tp@U;bQoj82<y_6MGGKsbgGivfDTs8F%KE4{N>3E|kTP0)
z@(h)j#Oc8cpAP>w-}{z2YOi?k#!JBE89Bt%b^-uwmr{p>t@7I%fYCYN{xrayBFmco
zYx&@rFDw38M$$#;R9I*IlGOH}od<icN)49PoCQsg0VJSXz#2m4B0skS$vg(txgJ+%
zesYGd=cZi_eY_x)qN@YAGYu3Nw7SD7l}BdgLMr9UJwbvaObLhXd)?%hj)TvtSCGVQ
zz?}t=5i$9rBgVkkJY{FmiO%Z~$7QIj3=NHX;fpfs#Vzsk03Lp&vxR&f(9x+=nz6v@
zc&N<L5zHeB8T2L~1NKVCE;kmsEWc!Cv5QVQcX*9zz_T#FAU?F+FFaVOolD38=a+PU
zRZ1sGHu9N)2Qo0=X7J6xVVHrv%3FFSAu@-YhD0h+%731jmTAg|gA~pkFUpOHLE?vp
zEZCtfp0e%EI#0SMk4lzpiW%XauCh|GY|@`pcG8%Cc7EYmc|eb>_@?YAet0V_NakG<
z<L>aJoCMH0EEzyZBOKV{Tq=p<3;1bAhlJ<c&546$l7iN_%p>JW5my!r{Owf+7hOAA
z_s?XBOHGgJ&sb=SO2$n!guq8g=np^guP*lD0qRTx8)}pb$iyGUm6BpnD2(O_uPPrj
zP^6>1B~Fw@D9gale&k|DyXYsJQAQYAlgte*fFw!jFe^LChtk1iz+}aPa84dEVX?8q
zy5Zr6gO`f~O27}cY`&6rVt{IwsbONQo+)2JR8t9$UmLyyHb9u=HG`?t$&E;w4}sYL
zPVNp|ZqofL&WSJYe@wC`#|+ObfV4LB={M0yT`36zT(&WdE3}PY^y;JgEHq+@_SdoF
zt3t*zp_(muif6qpE;sJdTXTyCzSInTlM20%;hG?6I%L&9di8H}7jI+cYU9_&e!7>K
zt*5!gr5WhK7c)@pUA%SdjTya^wLZF}!kqRrLsRW7eeP07H{<nd2ioB4`>&cSZP$!m
zlz+7CEzSr7wFR7>Jg_x%$*<+ry7kf(dqdaa(r=L+l$2|W&FHunBR>pHJUVWm?T)$9
zrRt{zWNUwRNbJ@0iBC7eM%jj=%t_ycmPe|R>m3|4g%6KQ?x|XQO>1oU>G(6EC__Ku
zvK_zKo=M4ERSW5w_@qf{Aj?EmU7D^kSntyHnp+%z886c9r{L<(bgB2Ngv<KXbXUVw
zbX|Mby}_4vJpk2JG*$IO#lyj4LdGb5r>Q<)=RP9R8B}?n7r+nwRy@}CAkOrAv+s*e
zKbl3QA+gE)qA$(v7k|827$O$;^3rc-Ov>(q)3XzeSx;=v-B(Pmo~W<R{aT|wMs0%Q
z<IFpc6VI_L_xGO0?%tvA4%^w!MS=YT7SJ?4@faZGXS~V61^s+~P^a;P?_g5dIy-;i
zMHlYuV*EoGTibiSmy6v|Ly8X*u8$r+iH)rd@0dF~*p1<9=P_`86ptQ!9M7ISjP>Ww
zV{>gic0_Y{u;=&_CgYm@y@Oa?-;{hO-pRuv1;;)vudc@0`g%Ni^f(?qeB`4cIZBLi
z4@bza$(`|JX2=VtFXI*W&kK&UPQ9oXjH@y)6I@byf1Q5I;^K|?cmD0a9e?9*{#)_e
zzw=!mNyp+G^0O$pvHm=sKYwPOxNWYl#@5DK9PID;v<{!rpdX=oAx82O9`g{#$1IAm
zP{8qJEHZ_6%$*$W$A|C#I^KW(gLv|E#m7)m5(gRYF(Iz<xb6Gk)}>Z`Bqix^>&`3j
zNB_>>i|>B__hNSOUJQ@TD=()N+-RI%8;R5N>v4K97-ykz;<ceTQ=gf;;*QD=5AvVI
zq$!oDqvQQpUtN}rhZ=Vt$KKwycl~j+C6ft^McE0)5ttk|$RgI%#7K-we(tXtRq+{k
zajf!ip!n>>`N@7<oE*f9)1$a{e&TzIWQhye>oJX=e)x-c^w|fo{N#b}9NXO3a6M$V
z*!;}2cc_p)xj@K{8g@6Gco){D%H)>H%&yAz;lW9n)IafJEyu=jCm0{!vS4&3JPW30
zjMqiSLYH@yp;QJ>G&DUqyB53q$FaF{5L;3lLC0~8^D`{!O~t~D#=X~6H;;DX`im3C
zd&KpIac{)-`f|Mgi=V{LfAX*5qYr)+AAk6M{Q6hFbiHBz;CtWu-T1)|{~%s@^$qVD
zW8BPI2D{6c2N=~jdv0o6@f?Zs>niW(*L?IHabP}Se0(A%rl(_ka@u)tt~P?Z_x5(T
zT_4YAH>97lleH{JkY2P^EP4;B9CCam3)|cY1|Q1${KCA)?R*r@2ej<$8&$ocouaKl
z_G2Xw^9zeJ6LII}tQS04Xas+B%v>5DVA2jz-?BZ-2i6yE%RDv~vUy6ltaI(Vl`7nQ
z*XFn>%>el1lkF{aKj$2h#_l#H;J>!Zd2)?skj(>H6#XvGtyxiT*|7v|^2Kk0{K(94
zrr_AUn7iFj>`YYIu4y2d(=YtXJ3-;&4_D90h>r3|y#mR}JKm+M^yVW6%OpOOd-#yP
z+zrMZVJ}_`#^l6g%+4*u_|$Y<^R8I(Ol>;jG?h;^$~I3b4_PohRetRsavY-aOl>Lm
zDYCHT`A3DvTqko&?Czr+a*SxcV`w`#5s3RD)rN8WqT-8rTlHi^?KpP`?rd)c_k0oG
z8fQJ6ay6~1ne?%q#&j=ky>z|0c@}f=;54c<q?4V}TJUM<;p1O5_v#ghgD^HXcxh>_
zcxyfWfXc6$8%u$!=wAk3MvJ}Q_ZRU4zKo|ZMYZr{JS$wy>&w9*e#I~<d|mopP`dr4
z{5MA({;}`)5;v>lsfQ}H&Bhm8c{X^8ugw&=)X{(@TSb7N7PgwjTlG1erQB4=;L2b-
zI!#RiZsu71o60EJk-b=?EJty#T(UyTK)B{!sLf9XAPI|GMasV#GJ`ZG{Fp2@EdZG#
zf@853NBCOF9TMDdbvj@(S=ws{&h+|adZ4Eykf#5HN@uL2Z-gDpznYVx<3DY<q?84O
zPPvk{hz8O=?37;93<u2m)qGS!$@ztM+|13*$Ks8pn4FySBtPj^7jFz=QWs%t<xv97
zi#>T^=WZ`L&hBt}kdq5v-jQQ|Jd&#yfMiaw=`?VRn_d95Y(s1LYuxh*I3CiCfy|`I
zw`xQfUd=)dJZ(mkf4vm>f`^Nec`3n@m!uap)B&Qy>O?Ex#-8$2>_83QAPaWAHw}cF
zFXS9_c@iGS>Pfuo<UVJ4fAgE)^!;A5vvclbWKx&#EO)-l!MVdP3xhORk;V*`4m5b<
zMV_4qIn>NfkAliC%M`9WP38cl=uo(@NaA9In=~2xv@G_FA5bM2Loy`+q>C;y$DrJh
zY>A?S!eC=u9q;M6xfq+6@BqQy*$E}yS$^nDAGrhM#Z~D(PEcmPx*`q9mmx2nQjT&V
ziNP;l&?)CFW*#--OvR#n7H{+5D=KzjWnR$9BV8$r<O8zcg>-=TfI29oVwRQMDTb`s
zc~rdUAPzCO#jFmTIz$YpoCJz{c3Ly<Sm30y#Mec9705yzyTE**uLfg2)=)g?93!9o
zh#%#Xa_j{;nc)Y2FWez3<wgG7PSFv?EcZMjF$dw46BKwWh3(QI`N%rJZ7EqToWsme
za7YD+)>&QrSgte*Sn#3jC%qRw<e1XnAH^=v;w~K*N_espp0J%aauf5cSF#EPy}7*5
z0E99vPj$h_d8C0t@+y2u`^vjRXWstFE+#Ji<|VWIn4TwoiVSu$zWD9PZ#wKSLQsL-
z@GiV0vz-qJ96n*$Z-xO6hi^PT{MvK!C^$pGBS!<Ixp2lfGcd^qv{Ja$a8c$458R-^
zt!lOlh_gM|QdE)|WI$f*H4jYU`k!_*td0Zxl7&6i_~4$bu_pn~lo<mLK{|DyLl0xe
zM|vsOmW^<V0FH}D5|1zFQ}<#GN?>;zF!PXI5lQ)Nv48U`bQKB<qu`PDmcV`rpBjIa
zF6WtPjZYdl)DKc9e>FYQU4P=4h}3bx;IucDbsW>VxTnz+RmSv@I2%~NZEzJDA4Lax
z%?77`it*QS)M1GuaJ5<Twdh9d);-Hkg)irq=4#K_Cv@=uZjD`9<EKN{rD<pd7#n<T
zumSKq3iR-@C;n~OskrK1{r3C&DrU-M)~&Sn@GW$PK72Y|Wlq9gIwKl$>()!#m|gyg
zy~DTC;N-z~u>xJ1hOC1tm=3PrF4aSm^%hm3p%bd&Evw(v*sFWeR7l!Zb$PUI#0g~L
z+xS)b)-9ozM^&!Ln58vsVpYWrRom<)o_0!_#?B8jc+;jkQ1KO&Zhh`09~*ztdh8&I
z0%bgkw%YD3n4eWQUsBm<y<dR`Ds4hnt{w$V4{5d|sIGsY+Iq2<q4G|%H35cRm2@&*
z;ns!Z)#5Afd}mw@y&Nt>m9gPfKp69-AIV*`oDu?s$NKyzqaCn+^I)t)zr-nsi|dFU
z?rSq{WE{b%CC+QP`R*<Ctvv~9I)!0BkkIedSU~;u+UEz@UBE}1+{ps{<oHBPO-*{o
z2Opv3LKr)5#%*UTz!;GUPIeqIS$c4EZ2wHq@*y1o`U5xh@40)51ti8M?5?3-%VHag
zPTVavG&(66hvHoFUR)cD1CD{)*s$8%+LC?W3tKE~%+F10oH!lty!lqV@%kIC>uaKW
z_G~#OHICu~R*ux$IXsQ6{nOY#(O4yhHRu?Nfw9?`m|cpI$ytSc5xe_GQYW#wyXPHj
zoP@G1wZFIPbh&f;c6{fzz7scZ-0+8@jAd@$c_m(X?ajD(`<{2QFlo!gGUF(6$q8M+
z?N9;bVlm}-jCljbeTz$rarf@+c>Q&DP%Xve)O3tbOvcFAq!$vH@MkB)vu7*u=+TpS
z_~>CgeflU?SDwVq_J)|9c$W-!B(dvl+zX$|Z~DQ=B*oZ=c?taAfA~07R@Z!=)8@{u
z7r<B`L|<E5k8R<|Q|?BdU$_z9`u6X{_rL$cn44Qtk>o^@oF|}+dhUQSjLyhL=1he=
zdUHZ~O`cGGI5lEb{GTqbdI550_b`@MWmmk|MM#<1-dc-;y{$Mt*p4lBX{|i6zVhr*
zJbw6DeE9y)<EMY|XYuo&{YCuZXFrzxm+{g2zlu*ke&6?wJ$d{no;`i6^m!=z{a9aH
zlUy%iW^O*_7jMVx{Ee7hSc+N2g}al81NlPTqC9gX8w+0Rn>(?woeNv+RAJmo9U|T=
z4k9mVV|~MO5Bmq~qGK_S#YWXFmxT+BUsdd=175rucG}hZk*L$dqr<*`Y=#|9s#o-D
z+1=zL50!o+?3Ppg;Yh*dC->vWfBrAxXFvJ#c>Lh=*j#%a+k1OnTzccpcjNm%{87B~
z?t2;+55>Of&zbUIbMH7FK3$EK^=<FEVO-D0``i1+v9`6Ji*F|^I<f#d<V7hJ7JpRC
zczSndEB3b5l{b<}b{0~73@dF2?NUj*cAS*$5^_&YO~t~(V$97i7<Y8cPA}yJ_c~2Z
z#^}h9;;i<-yS-Qx%M*xhFHOg5cNf%V441`^Tu5ZGyxtjxY~+QjT4~oB87x!k%3#+)
z+O#?^<Qhr77jpY6=A=m~<BMO<iAbjmU^kBvCbZyir@d=<D6R>#?-`DN61(B80%E=p
z2=GDN2!Ia<XutDg1;w3(Wae#{TNKW*G#icbEsL^F>Bw#>e+&T=;b`Z52bkxN^xhFq
z7C2{TX1oB&PDyr=o~iy(&iQzOqinZSu6I-i*eQ5=!6{YOln>OA6Sv)zWfraIese77
zy*syKetuSEOlhQaV19=%S5{X1aXKGZ?CrA1edxtzjxhA`rW|Lgy2uf#ymOF=g|UgT
zK`)Lnzqh@icFE=N*x`{}`umY)rMb<<7M*mLqK{JWwLR{oOFhjkZPjIfb?V^w+icgw
zb^HN#*RR|DSNDE8{GuAXv=&d9to9bLcQJPFB3v#ow8(xQUAT&W7y82kUd|7ECq4x(
z(<faQUe2=@uIBaS;1IuJ7*)KlOn)Vi4#jBr`n+=C>yQ6uf8v5%DtwWxmJWh3rYo#S
zlL!+0*{%i)8rz#vf<vd!WTR%YXLpjuVN`lrn{4Oo&k84pRp_kgA-A3hEaB`7U0Rb)
zf1H8hqzXg_qsU^0g@u%*G;hqZ6^CvbVCvcnBQwePV4$%l3{nIp0yUBj0TQgTS2N(C
zOqv**OL`<8gXo&{NIPNjvwy2<5)`BHHp`@RcDP1LhQgahFv)RDcbf%Ga3ZijL(81T
znZ|bM#-cjY^ByqxOIw5JU|wj=7d~R@!3!_`a~8GqQt%=|E>O_8XQyAplnv?lr2%yQ
z!BM(jj?x)`(BU6ar^sLUnJ&{DbK#&LPoa1UkLC$;%(xXDyy>x8&zGu2Q3U*QR5)~m
z`2vWSXy$cxD%@cc5idad!JGV>dkF$?cqqN2jdOSfa8lYe9R-Iko@QN2mL9yM=hZ2C
z@4ffD@OAUXEx$<9Jp<w?E)J>*1f+~~3RvtR49<(^tEFS*1JXk26mxcF3MLPM9jhH?
zQZ|#ym@|L)V#j63cBnF7`>XK)NCW%t*DHl|(K{WL?ry)taFjvX!n_AQ<CEj=Ad^;{
zNA7c_<<kpa>X@FYTzS!p#U>44Y$idK29#MZOcIXrm^^PgU-|OQjuxG)Vy7dea!O|^
z^Ezdv40?b%qJb(MZw*xI&MXEq#2VUqYzLtdmKQXLW3eMY9d;JBYATRllrg@{LSBDK
zNj&-LN$KK6Jf$1uGh>rJp;wJ73vziBBz~}aAx%K|s*$B`$|rd7a`_v~O!2V{3KxGk
zg0&({Juu_nVY{x7mLSPu_f}=;fZS@GKm)R^fewlB1D2SOHd&TI*_-)gI5n(_OZR+a
z-lm`QZC_dLtRpkw%9S+4RX94l<|}*hlLr~#R;q=8v~xLeVJ|q8J*AnTAWvR}2saZI
zd_lJ~BES=M&BV#UIMZM)o&d|5K2s%4E+nO2$Q{q(juWM_CI$Hh688)*ee$e%<%f7U
zXfv!d(y=-`aVNdNNx9D8#~1-L_)9H8W`_Zff=zdYtvqzFCM&KrJqlD4(%Wo)tvi4<
z%ShtQ8>4w+qyA<5s|@LEkG2@w;hoe8DN^DqF}Q}8c~gEeJ@8tofoe{AsU#_Q<V&FZ
zLF2X%jtx=zNt1pZS?ot`NQRqfgu8j5n+{yj{lG2bTJZqXJ#>H!#v4=IU3Q5B;T1eT
zXl>1OaQN7vF%AvN#L*`FRW5ve`*;6Nv*XKYSQ6gm0<FxLAy6ER=$aFfHdVFv=#F=5
z2YTGlW$*gynynWe6)HZ_g~tBSZB2W1qmC6Ywh{@tVPJIohu?}vFXUan=sh$A`tddw
zXUw!0*x}X1r=IC6gYAAR$eh;7x4Kn5A;W=dp~GyFwD+2&unLfepC{Uj8^HS!6B1Q?
zBfDuT&IEd(!J{f;q)WLjcl=en=n~Rip*45$ZVhZtnT@i7#2CWmhkH?#sIg=1;LtC{
ztGXH2;;X_A%JOvDm`dk%2xB%(dZ1@oq9=7pd(#$OU^m;*z@>5+CXp__(3wWRhNh&m
z0XO>rmwF-_X9)l!6Ba}@lhy^gcIQ{)QlNxkvW}h%FGE9@?tW!8P%Y`6`O@Lpr}BMw
z8Lp!3LW`%IR`*^7D&1r)a4EDT!r&kJ)v6cT@Qc!qdUTz>q5AbqCaaU?{&GE?Aj2tO
zK?-mE&y)^+g^R8p^Msir))>1o7Gg|5U!CLgxZ??W_<C$<3CN4mw(IW+ed~FolurUV
zI8x9WhsZy7DUFUxA0La^nW<Q0ys|j&j}_Ssz!-tCg2xMdRHg{hKj$7_b{_FzC3q&f
zeRtB7#ym6AF+V%wiBWdEu+ZRpmgt@ijmPlBbj;2z#?4!|6&8yv*SxrZVsVapxPR}5
zzZZY-qaVc^Z@lS=^;xOunK{$5XmqYIHM?09db}7I7w_wFaB?k9UW`O2jG?h<jlFKf
zYj3<0ufO?D+`fI+eE0}<N^#@kJ>q{y_<MKmT86c?RmFGIJG6eF@ZWpyx8lvW-iZ;#
zV}spThx<xg#$+TlkD^OuRu-HH7{9HoK95iDe-@uh?d)s^cZ2Wl?RkQqg}|v9rSt4U
z%r7j((#>13c;lvIUh+io>B(L^{QRT%<ilUaqt8Ey_2*CJe>XTf3>gNmU3fx!a&pX*
z+`c>f`d~~>&4}M(%*@Tj{L&5aydBdl3K53J-#4UQfBnsP=bd-s-S@r~?|$>!N~1TG
zwzD43=h1`Qy*1*+8uD_Aog2#6DdjPX7^EA!1=&qCJ2mB<TPz-NRNCr>@^Sw-_6|?u
z;OJcWPx<4BJQhk+p02AL9`E{|v1gAzi<PIJ$I7z@v9kO$*4LOoUyoCEcj+DKq*YnI
zP%ghvyh&y5&!XHT2S+%*^4e>@<Lb?~zZnZRUx~S;yD`6bCnlAy!y`%;)g28Gy-RI-
z_b@iL_N0*Uz-8ErI4a{z$P-7#ro@HGc*dBF*_pKWj>56gq??{pU72+Hu*+?jdMP=n
zS35hqULa!}%pA<{2=|KRDLh9S=X2yByX6K32{W&JV8_)*zy4|b`M>;U@%}G=7JHlP
zF+4V*yt@@|zw^!b{Xh6y@q-`zq3>E_!H%71FBGSp!;5&dx)WPU@9P}LI5HaBdq?r;
z*=jsqUW+}|IYx`LzuXmfd~%{Pd_<GvF+cMThr8Qxa(IyQR5Ur1WYN?9(5_OhnKPt+
zOq)S_<1wh?#?GQ~<M$5Wai+E^@8e_9Ye?gA?xthGk9LcNFWSdj3ls6`9hEnwE$MN1
zl*df6_(OZe;!W;ey>Pom8r63MQskeuvPdO&+E~x85P$)_niHVw0<9%%{r>hVG#9qK
zAVr%Ez5dKy;L~9!-ST~tss-_qyq1%3WhRv6_f-(I>GBUF=sdsTwz>Ed2jYu=u8P1N
zocY*~Ep~@byZ9AxuJ!<W?gr#&TRu+kVyDua1t!WWy94o?*HtioDQ77bP}etE9N&(O
z?OpGHWVi3p3CFE+%q2T)NAxaGeV=0iS8c1>9TvGsG<GWT9^n4Ct*x!t+u2gvbnatH
z7v{P1jrpAXh?~3L&W?}Wh7D<)IjuHkX?|AGI&=L$I#By2shl6qSHd8FO8=$+`G7;s
zPpJ9ZTya<EY3}$#H>%goXxnR8na>pmn6ASMsOCm*%+MtdAW?~L?8RIm`L|#`9R%eb
zs9^zJb3gREU1jgV$nr9NS@tixUB+SW#l4K@%b+FyYW^JJR}7=}`Tuo7vio{GS@^p4
zAO9!+n5y1*&Ft;6!j#yY{47PF@s)Z~keO;gBn@LxlN4yVROrKCROiC?mB`GC*i())
zJSNq{FK6krbGFiy)UbhW2=-(hFYu}I;KMRvR-tMp<Wdu|!_mgHXQs*Qf6YWJB~1!X
z%ssjuPLE<ww`xl;KR=T$J<&i*mpKRbG8S*u=gWAfzap2+yi%(<sqO4IfN9Y7R%jhA
zG*}Yf4@oB_8alCyOAB%D?!A~(Cy9xAc6wD`=rl(rCl8h2e);Hy;o}X}`C;(FIoo^<
zcBBphM-d$9#X<wk%O+KRq2XO7eu;^r{wzANHU9cFjyoKcYA58$Z_7_)(8Lcp0N&l{
zj!h58-$=U~WW-ke*i@XMW&pw6SPvgPis!4#UNqWOr|i&|j%rZmube2~$pNp(FQgj;
zwmb7h61#JtVUnFrsJFuF#VInq^vFjf*UOhhHNMa(z@1J@>Uh8X_Pg<0-~M*oRL7oy
z5TVr<eNI42cZWa{6njcj+=SoLVBw>WKa8LL^r!Lg{(}r(G}K7IyrjrWW8foyHE%OC
zcHVFtpD1vnL*?`&zSKF0TYez;YR|T!8LDWEGlBGCd_Df}1?e(85OW%&-G1$rxOMkV
zj7{*h7kPUv#?)c-1-MGfeIKc?Eq%vZ$LW+Dkw?dL{&P{0A74^wWRhn2^|{RokIo*8
zUA|LJZY=Q7E_&;{IyFap8K$!Awo-zVuY3tNra|hMI`e$3N%9j{c8IYEMVUV1t2E&l
zY}46eAa|+`s!DxsW!J!i@<G|4Ge!AC9dKEz${y)SdSh_CpksET0YO<zW4F>dL`kSH
zhFsRz*+#s5Jcr6qwuL1JkxO+E9_r${Q^Ju8BkI&aPo7R|;CE~BW=v^tJ)*(ZAd7BF
z6EA$xsTL1+j>QnU=)haXl0L*I%MI}<I6Y|}geu_LnEoUm?v#V9p0?TDCbK9!2sWM5
zU1s`2F8>NYNt3|3&@5IV8y^-Rm&=<xXIsNphl-ctL&+fTs4ul+nD0kY8=msy@RPqx
z`Xfb}{Y!FF+kywFOo{ZDmlG#PXsS!L+t3$EWG?IzUtN;s5>^j%6drlMq^wG7-ihgw
zGD=~0h)$FQ=V6jr9utx>*71vb4Ig`2-3}=Yvpp!`+Q>stgv$_%lX=LL7TXCDu!|X|
zCBsM^1mVF8Gqg!q$O+r7nZ{|CElC7vPg9{YDmga(?3SDWl8K^&WAYM(%}LerOp`z0
zk-*r&dRHVaNhVao$Z;zzdipKCFjF>(I5>MTPLIOd6CC>4N3t_Cv<<G%>q$17g;io9
zOA{a(DT->t+vOzby5O|aqlRM&$%IrbKLG|I%*$z$w4NL{lZgNFzy5zT<o~AOfBwJt
z-#E}LA4N}NgG(UYYiyJ%l?FRi8MhMoUVqdlmzV}{&-Cu4&38*e#rL?&Oc=)L?~e=$
zzCv}YX5>Y6X#nh&!GF;-WlIRckj!1bj#(47a6;60;g)F8$%45gMS};bTWimBtoEwI
zu<cV7oC2(IM>XkyQ)91wTCf=K1K+Z0(;5PtN}q-Jrmlis=}GC1e-~PZgDi<Mm{&<d
zxm}$GZ5%4N{1<!upbp6NZOcyUx5!!>Z3jn96?9&(j$;WX>mxi1LQvlgm0tpWTluOi
zPFt8&??T(y#KW$Qy@{Xvo1^U+PRGfBob)YPzbS>o=->>Oa6B^3t<2W!j@-#Q-GGJ;
zW=?}Fht*#K@qvMjX|B*Q*7DL}Z|H_sJ@m@~^pYV<i@;p-tp`lf%LjWu@5FWUwilAV
z%e>!z3(2~=*L-{lz0mz4_rg`OR_M=1`fYrarM~INSPYCz#=z*bIOKSW_+>vW!Q(yo
zVZxbX?O0skm_t}H!AXCcdyP0sjQTY_HRXi}#$!yh^Fbh!#f&*~(w=*B^7u3No2BZw
zi~c<RZ{NA2vGJXloSstOS$%KeIU<Yua_J+VsK3udITQBuHF=dYF--VJ8ecFD=I&+c
zCObg<WxvLFK5|@r_d$8!&f@25YhG;NXgrU?6jycxEY3~Eou%10JKBps{pbIE{OO<m
zbB!N1G;X{ZzyG5j#_#>`hjH)Tt6u0rKKlKn>wwaO-8E-wbJ(r0qcO+szQ*Q<r?I)c
zA6pvhvTH%Zq?l7Y=jNw9@z2qF+*?GNaRlGa#+vVw+27rcjpxg;y!<o{HSSuvc{A?4
z`i2)5Nt4ypmH6<(_hXAA898bXa2wYFRt+lDGUAFOT|B0grtE}cae@W88#nL9;?h#g
z%+7h&OfG6Fe=d~n$A^;TLHtjD_NVc|`@dFLr(!=6Q`6I4pn`s8dM;*UzIlrUv)gi;
zkIBht^{ob!7Kcs)cDtRNT;xt5l{Mx+Sfm-{-X1A-{UD>ug%ro6aqkkl_ZXApq60fp
zAiL&;3wDrE-WWqpDLfYO*xj|Uu@&o^Tgr>Q;7+;IV@?`T8>aYk*B9wHJ2MeC7CC;9
zW77sS&fQTN+gI4<!A>R8heZZ<4YKo&GUwBEaF;BD{&<)jO(?lj$D~+*!Od||`O4}?
zzn}4bj)mF5vaNE$5s@r-(Jy7^3cF9Z_iI%7!YhW|TF}2ZV{DZ>UO1A^N5x4t#;4px
zP5nJodT?Ldp5ni?yJtIfkaVLCFfL`zgR;j?D)_NjIi`HzXwSpltyq5gAU^r{*YW87
zr?I!S83V(^%G0~?);r(y$HKR7-S#2_i*frZ-|MPJ>=HcJfbig08KZblPEIL(uE+M~
zW;|P7_QJ=Y#`aTFb6%_)k{jVOzi@UY88mKZ&T4COE%vrIa^Xwv)Oq)3kyQLC#uQ)L
za>_$~$jZm5#BEgR#?D2Bt+=w#$KuAx!LI7|P|T}MQSvrz2D?Q!H+P(4@4h-4zx|Ed
z8t;$B^L4ck8@sW0a31?dXR*7heA?QMJ<00%7s<m1tv+>$IVP0_+8f5*=aLy6@O2`T
zSGqq!;g#~|`8nyByU@1PIbKHLpLq!KnTX@&aA-eOW~f6o4J?`yC({f5w7q;3?=Z!m
zaUOAJkt$Evp}kjmENwD%i+mH4^so3Km;1-!>l1k_m;8zk;9YQ(qf)lWB+4u8`zbm+
zM8S?w793}$Ce>cd$E};UTu)dS+uYbxdLMeus4j;3C{OXGJ`omkC!;F!vs2@0r>Dhd
zBtLAJ(R)m8M`~ASV>y=f$ahH|xSY>SO~kFGxws+4E-~WA(S^IJAJ5mhM{L(0La^hP
zk1k%hcPI9h-yf)bWq#FVt@IB)HdH#e4KgC@cJ}SSXWCm&WA?`cJ#K}zH5WcAoZvDW
z>-(2>qlTMy`~c}c!EXB1@PO)G^>j;B<3t}G_fq=&bg?Mr3dz3(^Xw%tRK7qB_bTXj
z>)O9+SQWZqcHwgP*n4r8X}%a*@_v7$V5VXh9O73D!*dh=TA<~VL4Q5I|L}kPA6>il
zzyDAEk1sr!wO<!SI}*n%pv7EtoM|l5TIn%m@Nz0p3_`FK7(4EO7q@7|Dv6(PN>hC%
zrlsqZMFtR!U=`#8tNG(r$ckWo>X8Xn-O^J1WvJy@27okB#fu<uB9?!!_GnCkrIwqX
zg}0%Q<>XCr%s8fxN{QXPt)~r0+QPeF(!IrWd1sWn^p&#uOTbH44*p3YZ1GsB(^}F-
zCgT+lFAa_qxOwN6JBJJ$DCpH6#7#JUm*^ptFEG9AG>|w^=Y)HWIIo(^U^(iLfd@KA
z3DQ~cZWa0QJzua@(c;BtnUq&tCOL4O^R>C(i<eiHJpv~?+%v%L#K~AVj+yz1!)UYe
zT^7r&d?YIqo6zotV1ob53r0_$K8eRq9{F7FjkR@k()J{)I(F(!S$6ozFRuc^!Or3h
z3x&7u-tw1Xlqm)c9x%xN*7mk{V8O#@j&t<I3w0*tpS#RB{%(G8KHh!vop|@%cjLy5
z8(yHO($Z;kM}&B()o<D##Ftj%pR&+tdlBpFtMTI>|0I6)v!BGy?v4m@GQ1fe<JSr5
z6ckUTGmMQ=cv2XVfftgS06j9O;!^B%v|x*FNVW;+9eEp2=bnWx%dW?O1`yNp^D%!@
zgBzvAh&lmuhUpxPs^dw2hXHC`Og&WR;Y1zQGj*I!)k)->Z8K9CbVz-Kh&t%x4KX24
z7{CrHeduIm2QNEl?D9HON11!W=-5~!`6J$R>^OFCNb>QeCOfPgXQ^TF11Ph6nM0YT
z<HA>f>~uWRfCk!e2HomV(E;S%0KWR9Fp?a6bwrt95%laN>jrs*I#KyK=hHjMOZr+D
zL(<)YESCi?JZ7PoFBoN~_pEwWC;TY8#FN1@1CSA)?XLlocu;Q{wB1;^8H+QEs`HaE
zLPwTB>|O1U_>im&z(Ll@2g(>fFM^=(f*(C_^JJT>lr4UC=Z9Rx113q3{z?mt(R4z(
zOk%e?08Sogmk7UgFfW`eI$;RNb4;bwy0e@3tV4AW;Kf(-l0q{h9MYce7Z5<!7ju$-
z$4!v!0`X3}W2k?z3TnE@Sde}uSsT9G9D?~J;L=zG*s*3>8fAo=GuP>Y5Wv={02<y>
z&NDgNzw=3D16sR5S3FigUfGQ)ex1;O1NGl@#+rsu#kr)NW0QY5XWVeI5B!S{xQ4I&
zR8tG~hkuhLZ9`dPgl7rSHuE-#Q$kSemNB`sdi9rp&0!Iz9+yH}JvVN$C=QkcKPqUJ
z2Wu(grYii3O~y1`DnFPLRcNuxl-^4b#R4QnrjN{0VC-s~wZs&+tSWMoZX;OdMgB^{
zHIV+HRb9j#yJ(CqJh5c9%WlS*2SA}P&42M<|KE${-#qj$d;vBWYZp&jMBN6O^s>37
zuhDQ^i{E}7z+CgE!At9>sc@CK(fM6PE6pXD_DFdEzMszN4+^cg@gy2vSSG&Cv*Lzs
zJO7qngSH=;OFC4%&89DQFM74Kw1EJ{Ksvuh%ba*q<H>p}zsbzFF&nV=AUcq=RSx%H
z3Msr>e`$7DCE7(Gx8nwowG31y6<{kjdQ{L=zG%rSaa}u5aR4oDk3Z^AaZgb)>=rUz
zvZ7X<Tvv^fy#!$=J#D~P(-UwXQs7s(a7|o0P-z=}MHR*v@>lI5gtponynuKB6@^>2
z8Li)B-tkv(*eVV_S#FR&^9^XjOd9MBAA#E*K}(0<YH#T<x87QO!;5yT;6Y{R1sq13
z#tl?^BSY&8+S*4+kh$=&xtKblw1D5fvLTG0m%ylW?ZPElUjj9qE|<j?c>i?7Zb3N5
z^9#bNr5659VUc=FL1%w(Jch;>VqlnECEUG|?MO)(`7LzP{n6pbNQ|;$MH$Rkj$JO?
zqr}(qOklIXw6wIOKKwvzZt>xq#&t{#7a6G&_+v6Ok0#6y7MVD84CKykjk)>7n46o6
ziAnAtW-@!i$62w{hm$EXUm344naj8+$F@u+pK6@J-9npNyS@{N1%=VE5%;-=#D8IK
z%De{Ehi50jvz5(QS=-R~Esx;$j;p2lxHUf=+Z)UA`KKSm^Ofb;+S=4OVn60J27T?7
zdvW{b&6t{=jmha5ja$a!#?4#ao#SXIeb`lVpz%pA^t_0}lZ)6p)EHNy)21CWAuo5v
zNi3W&7cf3P;)S_u7fPRl-PqsWh?V6>@vC3{B0jl)-|0A|an3k9T9jA2>`2?&@faA!
zV=VMdO?&*non8C;hZ+yd4XBG?E=S}kg@<&b+>y^pA9if8lVWlzZYqAvR~Yas{_K`H
zk=$E5+sZSJ2;7RD%{7y71Ux6mu;4%*E-c-Q8@H5Bckacjuf83Nx9*6|u<}J?^_~4V
zVl1dKz{2U6@@H~%NNLY0A0uiCltRi&+ET_4<oWP0i%ha8Z5d~IK|qRm561qxDr<v^
z|Jc~HkEz?-K8W3e!#L3RHSZo{49Ehk%A?9#_K_|;-ekwf;^J)FQohejj*9=e(tb;F
zZuu$$?h9g3fU?8{JZ<Cj^mN=)IhdPO{ZYI$tW<uh9%$^z_?B`_znnIWv|{0nqdB?5
zh%r8SN&M=f4GSB5m`hy8#*~f<gPljK8(VRxd}RFOBOEnm_py>J;+9iIIF^z{X6RAu
zm|;OM$G{qYst!>%rzSNnRUUJ%SdP0jCet`}e|ICc)|TVhqX+TXCm;A8JMx*u=R0@q
zD*xW_E=3l<SkTx}zO$&nQJZ5^^Km{n;yad<rOnuvxv2HkRUaG5m~e_6X>;>F%98jp
zF99#EJs>^#7<>KsN^EVedy$p#I%V7o9+I8-pLs5UHcCmTw1o$|?6@vultuWm00ckU
z9TrbGHt}e0J7y+E<Ic@RrJeE$-fBZoKAuta=ib6tymfCjre|hi|M()FJy+fzYdm`~
zptj+_?HLQG?ZTJxm3a-RoKN6ge_<I}{DL>_EDKB2M^9+WEB@=EmHm-{;zv&AJ3wY<
zD|*fg*}pK<G4hOg2*P771fQf4ai?6_Ck*VawpQU(Ihmu#c9_M|ff!c0@}V*QnM>iL
z1nP)9TXyK&Hp>t3LJ0GAs4B0NUHQdM8He5<{7OEj1>x3t7sb`5-^f347+0IQpgPBK
zu5;{mmEY4-=3_*w@D8XidZ+BE9ovzaWT$-@Q~Bk1)w$W3n3<hYyRN##J4Jp`Di-2D
z<g(!fb+swf>G>IzlQHfJyQVrY<o1Tt-cj9p@_aKMJY9`PPga%32Qj7kH*|3vCr5iF
zEpm?3QOC|t^7=L8v1cb;k|_23_uyJ{C$G#q^V)u-!#een|LM)R^jrC~dP;&wy>37?
zr(3Ej18Ii4%VIa49sfXv(SqH3Neq=Qa1{^A(^dE`VDDn=-b>O0-7ve*&o`O$;x5y?
z44}*LUqZh>QZQ4o3s8O4F#h!{e37?*IX=1Y^^g8DjcKz{0+erJ+#5w)(q2u<TxsnD
ziXS%9O{SRfG&Kxy0E<>811m4{sb)tg-6^o;AyuF{i_Vj84U|Dx@q-3qM@zO$CpxI;
z0@*!S?-pY0{Q7w3fTJ|x0`wY%j6VE9WVlILV-H&93j$1g7Oqk$&BR6lhLRG-cX(8~
z4t+KDqB|xTiBF0KC)scY%v(_0>`ji2ssI$Pqx1)|dV~jf{4;dm3m;+i;@}+QrKQEV
zefMt6FD&@W9l+@!-ZZ9gavt+y7LOM!9_Yovi-jF7yqwta#SRu0yqLJ>i!nNY%|)Q{
z<1$6S@Q_AFk{2V5^@)0EEP4${)%@o|Pl5$3g)5+5c0{=FK`!_@jc}IkRI;A#2GXRd
zFT{?N$c`*8LKcZupFfYsPoBoZhxdI3H%I9a$DHgJ6#q3EQxlW%jrYFkBMP~+6eJ4{
zbnf_KgoT<t4RrY0iS5n1vcrR~{@A|HK#GfeZ|Oz5Fu!2D<#k{rHyu$HMd%oG<CJ-l
zWuSD%l~26z8Mu7%>8J6NAO9pi{NzK~Pwbcb#Vo&h^Du|<0#jPobSZHJcseW@&`B@j
z<;Q7E$3v{>yj+4LO(^j(@KPo&$ux%aLZ4CEP0h{3s5+uE3v)j2cubv@VLB_~eRNcI
z5FYNAGNtmQJgeK$NfQ<qVLRmWvh#~LIyS_uc}c6><S`^%bjuMW$}0v1>X38BJ{em(
zbw1uuopU<4-nK91@TaqjJG;a1&mCiwW73{Z4f&3|*f|DeM6!^7+)c*USkyyyqPdeQ
z`T1J(Sbiuonb+h&PlCC_?!hQ|p@GHV_3Ugi@b}TL5==?uaH!Mx_SdZP!?;Wd?FBpr
z0NjB<+!>%vPtL0IwGeYt8qkkTSbi42JRt!u<PdL{kB&^Lb^?_E>C~jGe^?4l3A1*Z
z@awuRXR0pFW><L@8uN8V^+LB!C<z=-gEf0X3M|%BK0)DyY?x|Wmh~ANw(747!rN=|
zO5T)tESSloq;b3fh1I@)q}yZRFSkYp*CXiCz52I2jwg2FlUU5jCH))l+JsxYbm`LI
z_MuISN+*!PG6?7Vh8~NEZ7lp!d~nDEL-5ivQ7}nmxar>p`^m5`>js%^B)kmVW&*5M
zm%+lY_)0!hLcr|{Qgz~sXaCI{f^<(q%1|&RMrk5!_`zFYINbDJwWJn*?C>#0D0?Yx
z$qy%CFhB=U;q5RBQPSy2mjdtlCHG>+==hW{^)tY`(A9LcwbTiW*-=dzw34(LcJW=f
z*?_s59!-K;esE|>=-`sD@WY&Ohrsz+2pW#bz<~B&{#So>3t#`i|MGvUs!{?Mbj^)S
zq+`K*mlF9!{76%xBP)aEDzs_){JQQ6XukosnZ=Apo8hJOAzkdJaI8EkTxG6)Y_9Io
zobnd9g?|55E(B@MI+u3x3UcFxHmw~B^ceM1nh-bWx?o!eXZH@z3g+K~&(JX1E<Dm*
z=zWfkTk7eSAajkfa9#dD*N!<uur2MSGB_Ru*R6Z*1IgM1CrE$b;Facl2nu0?>v+6`
z6YDBp)-@ZAvmBirUHP8N`h!ev6HD5sU(8Tg>JnJSr1I;6UCV+bbU^(Gga)EEtpH)$
zs)y49fM=y4I5hJ@Zex4GtY&0(lOU)#G`wV1RL9hcXh~#nqRF}X&F~s$<1<~lkm*&;
zRmPewSTI||WJ9GlNVqP~)>sY6tLvpg^+((*cw)1|k9g640xDglc0DV$lntJl4-HMh
zSLk6aH_8RNZM~$IpjWO6CEp7)G?i3(H-3O=gtDwxu@p4%k*aaz>3-rh_0Na-T7F!8
zXO3Xwh;YWM>Nk;oO!%pw`dD!FrRbZFjEwmsLU^$l;CsD9!&r!Z@(qn|=jUd9%+&Me
ztNx0fiAhpuXqZKVd<ipY%j7x}uHFGA0c*R@4y3WsaeoZQ_yWa8ZA%<^D8)oKJ5(57
zu;|aYk+E0qP7nho$<?NDgyQbruERM$W}JB7uxDnceUH|H#)OPrcD8q7<@ri1Ki}{|
z74mAp5F^8b-XTDn&*>MaYcEbck-oXU7MmLz!44&MJiYSDJ;^mLzbxhqIt?_!h?SL9
z$vCV~2jk%AB!)FE@Ue!H(f5rV9f^JJei~-H!N-c%U&M&yVh7az{&sx)@rUv0XP<d-
zV{U$4>A_Ar@jE#1SeNlXGR|mR`?vn!Z^rAdzpl6*dMteZ{^#-QU;j#D?rp^#P`_Mb
z8kZf)jT;Nhq&#_}Ja;+EGEH_t&q*nN`1`*f?|u8*KGKc}{MF~n@#|myJU-RfcVj*8
zWJ-{K(XgO5GCC1=@4XuDeB)d3`WtV@;*C2>gK_VQ+uh5>6FzigLBhMgII)9%yF{W~
z`2$P2kB#IhBIFGd>8B?eQ?hU(IauV$U4VQToD<;Z7b>9&pK_J+A&lMGiNYkg#%n4c
z{1`iu&m0lBG(Q<P=BMN0^dNru^Pk3}hYvLtRN0n{9uq2$*ulo)&8)`IckkVexw!=;
z;h-1tb`Orc;D|h4NKl#du9Y+CjGvV*>>8u~u!E7KJh_vK4@m1Gf`X8Y%#rNI^Odz&
zUE6Tk#*N))Qz~N<%6}GCha@)(xX4D?_vAl|4QI+XWdUW0xN%1w<#>9M`>j=nm7aVg
z%*V%%AKs4#pMM(nKl?cLceY*6Zr)l_e&33Fuf7^@zx|HN@?30gZpXt%Ph(|uL+O6*
zMQQGXJ6D-Ok>A`M$6`45C?8VqR0bus>+g*li!mX0J|;bTaXq&7Phw3xnZuzi+1=by
z8G7#YqD0b%<HJ}^e;HA`MBB&WCFw|^WO0S#G$$rW#|h=jgp2Y1{;u1e?d@%sc^1)z
zRBq;`CgRSmMWw|=mK&7;>YR#(Wtbkl7W30X!F_dOO0V7hv)JAf-T5_@t<%`r+4Dj<
z3->I9P%cU9dS5y92}Qfgd<BbJ`3}j2DdO!pFyV8j7;|0BA1I^nOPr;Xag3Z@_nP)T
z`TE0K;p&lsc=8AB!qFb^fp?uhK^Eq0D2L345Z^j42oDq=1JIVTSVi4nK9L=>*^Wwf
z#_%ftgx!2_ZIQ%Z&eU5L$B7@JSKgFo#{bR2p#5k1Buy?(v12iJ^Kzu^Y)nnh2si2l
zIpV=Pi1*3?a#3%UZ=?+$YEu^$7WF=un^eoHy0^3Ij~!U}mH08FIzk)Fj?S6Mk(gBb
zF+6llGQs18>JSV39FaO4M@pOL>pSt;gJ*I7@$=YHyB;TdY6+Pq&9tIjbh`1g-N~;s
zs`eIdoxkae{gU4GN4pEFVy6)>sUE_0%R(MGoMxcQlOOP}YvBv%<=+c-tA0z~7nh<J
znhsp1P3YP&_qrySrq7}Eca@K>dj;&hxRxg4{4%J0mzJ)}tBdaiQ(cmy7aI33Ci{v&
z-u~qndEtwr4zV=RDjA0z_Ek@)7@)XN;iV-F4GB-&q~u`gX0*Zs%4{?XqQ{tkB|8W-
zo7EY;CCvtvVnm~xNM)Yo>G+X>0V-4`(E~RgP?6mpG*rj(tTeDr4Qj;?-X&g@r)e8V
zTZwt`Rb(vSLT09-L1e;^-9Gea%3uME@aQ5lhTw%O=r$h&4Vq{AwMj<OwbB_!npJn0
z(BO3%3GjfwDrFJRkeww@#K=Dr^LOvwi(Be+Pp~^hJYYZ+i93AZ=dvZF(}Rh4`FH33
z^vH_`?f|HR;Nu?jlJ$#Cc3$#K>em-nz9_{)W>17#7S8&{U2bJjP9Zvd0qlsCk(Uhe
zhrQ;JaX4FM@?5DPzh1Wl?{aL<N$2#i)iU9Omo<x2&z7I5Q~z;1dGgd9F}@IF@I&KV
zk0>M!cqz}$&czRZ^uzf6_rI@B>Wtm!gt)Vy^dZhidP%aQj80FzFqy5_d=<ob?<|1%
z$O09fs*us4h!6QxU(8izIuW%IVSvJ!&)Fg81r+KKZpH^6ydVGSU;e9DU0)TC(rG|0
z1AdDce+n}v>4c*UcYtt^xeG31(sR%*8R@*(r9id760b$>2Jj4Ce1{Zwc6v?D&cx8v
zm=^>2`jKN5xU-9L#o%sFoktc{FI<+?c~OUh1vlo2+<}n{#2tP^L&F+il$s1uaYNSH
zF{J~;f-Rkk6Ak=0Q~sE*tV$aAOHy^>MkNDBRb=O%-H|LV=9fb(_R;|&{ax#njw~`A
zXkf>Vv_TD)ra8ty@~}wEE-T6daUbC;F!7|r%wX_DgA)cM-WjRE50IT3>FUrDKk%e;
z9uGkrS)f*!%5#oB^kSN1;phpzgrZZKapB7#>7wWB*kN_5sJkq9j1BW8+@w0!3$dV1
z9v69Y<e@L^mAu5WrXlg9oWQGQuf%g+Q$CHg13E4$vx)<#I^mO89(u}RHz__yaGenl
zXOK+j%X=Y6V%K|lq%F@fCBArordAFCR=bH7-YID6lwC>#M6FZOrnEd$R<|R?-Ou93
zftPj>nSk_J9>P@~fGPBFy*0)p1w>gbjiU^OPsRCNXq<hJU_9lVNSk?(=F+=yF_jX6
zB&tbQdD%Of5N&WRZ*=VVE7@9H^JywMUemy_4sKxNDCEN))0jT+Q~kH{rzrt@r3KQZ
zALrOvXt@(+!6x7)O>Wh9<zv9uWhW@)JR9fgH_g?*Z25P(*}cTg@hEYwJnXKgF=ml`
zYU(<DFym|#wxq3TXgf6N+(2>$2v-_xYMhIISY-M@YZrlR9X_D3yUrCFQx|*fYZw7x
z^;_NTn*RD6?I-g-ss4+<;)O5tCVz^a<-N)&D8}p$p!;bZsi`$xdfK!u^zah&(8}zz
zD^T5xuY9CkMFTYS{O}{?a))laTB^SIbj^VM75<e+T8c6|8Th;WCa!oXkd`D%vq?I3
z06wsAZsUeM#qF3=2KU7Z-puU3(8-*WP1)bbFp5GeV87TK|E*(d&Ty>To^E!RKbPQQ
zM?UB(P1Uomcl;HwY_e8djVoZB`4_wU3qU>_K1HKLlem1>R=TEtCdmFQevugy`Jc4-
zb?v7d8`Z`u7s)~P$`53ICG@*h|J`_Du+g3>KXmL4wt>VI2V2sya43Gk7JRi=X0Cha
z#a`v$VRPaef1SV+mucGKTRH76bh^}dv|y_n(3Tdugrx8HV=%7tVJZk!X^A`Ihz6=`
zm6x<jsM^Q5Dm+0iRyNU52UF<Q^-6iKfXj5nE)5fHS2wKzVnvwt+w0Uowt0ouC16**
z%0BlUsc$+cH8i4e=>Q90?4ZJ(w7DMK2SY!YFX@?#9vc~Rzm~qywL$fJm0t8^ImYVM
zSMIql{NT}(c)q&kem7%2CQ5x|;^?qL_vsSqzvr=7^vBPAnhE{xJUWmM=dw>tADv0k
znVIRBo1OPgGmcf{$gHVp^`+Hr@lo6Q#=1X<8yw~iGRA&mJ_?Qx-!{1KX?@+t5HcpX
zF+US`Z!L&l-fOhCud(&{b&plHINEIgz>5%s&(U&I6Js$uJr;3#==-fc|NPTfS$P(Z
zA3yS<8d-3pu@k`VDId?Lw40q-@Ui*}H*TtUU-w-}+zqD2*5jAs!+niqc9fQ<`9a^o
zeyl6q_I7vj<3feIySp7vpFPvK_R!(@*gewRi%q4hULT3y{;l8fU0HW--|~3w7r*#<
zeDcXB8uJ`FY{o%H2kd0%g35<?YmLRi{E`<hcQ!W^w<q$)NhT4~vrF;2-~T~;_xnGL
zn|EIE_=Xc`&Q1>Fm%sd3{Ih@d-^ZiJ4-^HL7z$0N*lFyku^o4zz4PvS@!jwJFy46k
z8$O<FZ~xGXR-|+8h@vbruR<SJWlrhAeO>I<U~z0%`OMgkojO~)xf_II6j@-+omL88
zEGRqvxK%uIY(*L3ZXXsOvW=n)uuE=~4?-tKV`e&chMXQB#KZgdV|#1I3&KZdEYxz3
zon+9sjbrs`Z1R{mKEOOxJd}qj=RV?7V^uGTu-I@`KJrx9eAvm(CqB}gRz2ZPGIl88
z&SD7VpQ8y64o_lxcQ>{-x0P2MBbvu5a;M&u#<|>cH#s%oxeY|9A8}GH;03UgZe(P@
zJH%#GS0=|slnZCEyS*8Y9(*1jfAnGe^4GtNhY#*+{CTE4zZGx1`DVQS=3Bm_jH4?@
zN5?&v^7zri*xJ}snjgjW;fWZToR6aurHRsoa?bH^98=2OZj6_Q2bAA5#^U?p{8(u=
z;+>L==UIq;wzd<G*7lTt2f+y`0}}b%<NMR>4=E;o->M8yvblnQk4&k<9#7K-DO`@~
zqkSX2Ngw7b_I9>?r`i4?3*Lv`LCJ@vb2F21b8*ghu#r|=x4@#|(7=$}p9{&!afsJ0
z4prU;yg++;F{pfH;f4G-^1|2AF+0@Q$;IM&?pkDSjzZ@-O92biUTl}l$Vc9@(8D{3
z@<2YHu^=S9w)<$r^@TEYP<6*MQ+ARsm0#jb8^uRT%)8`ql|#u#?##Dj+sbjY`N1(j
zd%*%Zb7Rei<t0ogSb*guC3ch1PP1@JxkX0a0q~|^5s&g}Hq>}|o<@FoH*j(e{_3tY
z$5UY<d)Dh@fSK_tZFaWrG7ssUqBybRc6@^NamKq7$=CdVTKTJbYq{i~JY=`p{Oowl
zPS8%!E?h|dlQ=v+k7H#CbCMiyJ2A?!uG|SWDtT4bls^~h`%NmX*a18>!D&a=l)q}r
zK7JhUfACT4?`)`ClIDnw+@(#j?r3Cs$dc|=FZBFztM(op9(J*+c+W$TE5+TpB;8b%
zh6r?^<q2K#N=&a?g=TThWi|1{1DaYYl00o)9%*X9ZuZMM1Zj1(Y+Z0T<bR)A*N*wJ
z&qCAZ&>v<$^oNrSdT}ican5*O1xbf`In7mIUzg<Qg{$KBn?sg^zYOG#umAWTlT#iX
zz>b2#i`0e0dI?00Hj^K`W^n<U1twZGqvTVK2K*YJvG<=0ho6y9rjr|u#HLYhg)az9
z`Aj0ShK+RlwL$-D>J(h4kirzKOMD5#w5F6fIoXZ@a5yED5~h%yXcQ*4>p>8TG^k#X
z|B9;tK-+5hP<ojic+r8%#W|2rg6PB>0&@UtNoXQbxmOxMzSTOuP$Os_4p96=g0YzZ
z?fSt;+k_i0clm4bhxjHZ2a5rlk3mx-&n336sZ%kpjwPKa)A5k6?pL3L!D~kz6;8f1
z&h&!IFJ7eREVIxhJA)0rSYe^UU(1M=j*u5Bcqu6Y<fp^IjeeoZ0AB_1HIvJL0!3a=
zQtMCS2wG1hlWxiUq!^PZ&H^c<2Ipo}1fZRMG~!qUUMyjov-E7-^Qa8&_<Hp4aXfya
z0n+p5QY(6K?TMe_tn{FBLHK;>F|2{scfa@j_~DO!H<rrz<Ah3CMqYlbW3VfUuW)$T
zawHxdUtVgwKv{$-0!h~T<%A`6y7@;!_@8b(n|u~ecN`=W4Q_rhbQX`FK8YXy#b3mS
z>P+saqexuM1pcCNI*Ko{&>``zExo{#xl{p@^Ad4bAhL5onvtgv=fWB?7rump)fw{=
zAPl}FRGLlA&&Bk@TnwobFr*hhUq3RK9MV9F0od*?Tl_UhQkkN&$w1WG-)Xay$^KGK
z!qwvjeIJ)PIdn3RD|cGG(7-ddvjYrBNIQ-*@DU8hCmtA(Z+zkBj*|v8?EG?PO0qKG
zsS98JicLJkSjnb{3P-8q?k>J|q*FPn&OMzz@`_^)kb}WGi%$#^N%LcMU|IMgd^%8k
zp~%;ABoB*O{5aNwJZ3SO@<Hbk`N$V{!b+U!wEAm4brQYsC9iyS#o!(qpYg8nomcWZ
ztPZwvAf_i~<Ho{L%=2{<gL@K4B$OlSIOKG02phll4-4o7ZhI+@pzTB(f!tZoT<7#4
zE*QaQJ%pEVO!iUq$V=hjoAp2`O1ffJ5?@>|ya1ew=0!8{DJYSaf@c|AXdAwM(ze16
zzEbtwQSarlT-c!|&R~d#e#9-!g*`z-ZGQ<Z4Fvdw4_t&%I@vE}7v6Tnk2IO>E1B9o
zfpjvN!CoN|!n~xA%u-?DLA+=e2oKMRH5Gr?RQQ{5%4+`=3U88<aNLQM)$Mjy7@6_Q
zuUDuNm!xvI8AkD1XzP0)haTz<4?9BEiAw_f^s`_4l6%Y3$!DVsHLmc;B2gXj)jU)F
zE<T}}oUc?6=tP!%9$ZrnBf&IYOPrzAFUvLlYA{LPm`o@?(&_|v!%WD0|HBDiHOvYy
z<)1$)vRTN#2Z_qeti#ZzN0$MmM-!GH`&#;m5S>$qES7)JlYj~KC3>oVMk;YBo24)g
zWR3Gb`LF(OM*men|H2pGJS5)5+yFB6(D9@K5Z}Z%?tQcsdj0phSG(J%;*<*-3FMK}
zyE+Q``BgW}#31VfX`L$bx7S>un$1K1l%r;j{yqLxe3YrYcL0*B@h`0;Uqf5%`2mCY
zSZm$7bZI6mh|*r6(h)h+fRv}MpSsWoH29=!>=k;=ZTM*iVWZ8c<coXSD{W7B*lkNl
zx)uk8i;SRnYx&rBYj5!#{>BL-RCQ0lT>YhjCi93iVX$A*OuA{3pYeK!*ZgS~QN#dx
z`zd*C+A6pJd!whZc&KhwW<i5b9jG*xNxz0)Vb+a%)lG-R^+T%2V?Q{cgi6+?6iMoh
zEDf0G3VaJSJ$h+mcKs-k5~O=egWWiX-QufoZWjwc*77GN>L*~*qoy~pw%xW~aGxPf
zMQ@>3?gT=BPwQ0dl$E{^uJYdx{WPsS{ZPwRYW-k7Ul1Ny&TNG`oqE>}$KVJbpiOx<
z$HldwIQGY3InhgBnB4}H>)|1eoSN|XfH3IuA`5wN=k`tC-@@@=_a8izzU9ej+TZ$7
z81(rOAjkG|H2sk$wwqlEh@`$WAEVKCr@u}*at9Lk0dXhssK4r;_np}{G^V+^xF8wB
zJIz)%HoZ%LT_`tiEXB=*dC5Ge@yA}Qu5ZNiRk<G?YV2@5CfM~b%w1RPJm5ICvHX}+
z{c(<{;{!WVnPM|3xB02D7$3RrG56-iYTUp7z!S#Ag~f_P`Qby()wN~gXQ!s)JHP$i
z`2G)m6k}5}v9){X1rCpYBojLxMux6={I#{Y5o@cfUZmp#J?_vN8z1+=+RE~>!ur%l
z>@$A#4x3Sqf}Dz5ckYO8QsHr2<ER(vjt=)?eQi~89(VzeqYrm?_nn^HQO2=|&~o(K
zp~gm-Z!F%3`PmuyVKHty)>l^)k7u#G{48E5Jr)+1;>N8z;y3OEj_K)1$+jDxeELZ&
zuRQg*$lxNU1gmYI3#i8QL!-W{jQgVA`_^~jowwiBSaHmEU9D~I#E#_iwGJ0_gfwPX
z`f$wr^u%z?PL9R&_^>C$8BZRm?CmNKxA&x!rtBcg34ZFeN{jrnh?>G&xcst+#;zyQ
zhxq^&9QSuNV{84n(rP7c-o6pPt-Knan2xP2<s%DGN*iQi?9PHNJFO^#j8EA?wxh7w
zy~oZJPBb~K3k(XAxd-kJW1*ID_KfnAI)FrHN>A>f@?xCnC8f+PviUR{f1n9I7wK}5
zNqIgtRzBb@V>ZH3@(wF3?n7f-J~qOF8s&F4R##SH?fG)7K3`EBpT~pypZne|^5l&-
zl^1Wm6>}PsE^^1)jhjlN(Ky`SjZZ)RIG#RRj)CFPn4VpX;n7LY8EItgF+cf2U17mu
zd`xw6bSNfBSGgT2&A5W%<an<fEjsMQ<R?#_#pcdloLw7^Vbx!Fu#0b0dBJYB5%Hm}
zA03~?q4JVNQO5qnl{SdBoA&2KX}8Cs=kBiZQ2bPebFPBL&-_UI#l?y9YjS+bWqN60
zHfHA}i^?tO>kn=fH`)PqiE%{h&~=rm?N!I?wb$QPdQT}0k7HwNKi0Q)V@GA}?38y6
z?{_}3%>_=HSjto0Eyg{B+=-qGvGnb7froMaiBD<bzH;)teHSY4X<$U>g+rHHst=1~
zDh~>Wa!h%sb6M=Jqpfs%sj|-E6YaBigh`QJ)OB`#`lwB%BWXh&^*d1hhE*px)|4wN
z_+Wq!$4MK{kEo8(2KpUB=v3hfgvpOOAs!=2M|kEUAfS9QkL1HVOBYgR@{~m_`YDMA
zM_#i?>^T(0H9wfpyGhDr!1a*hPgyMEJv5~JWk>K_EX?w8`>2moeX_C|D=TY?)2?`i
zKgwWdCySwH7tBAsh+B*E@y5LydOu9W+WJ=f>5qREAAj~RKd7csDZNQ5z$gBApqF;1
zha8=jq(vb_uX7KLz2nx>w&qm0i$604vN%b=gHfC@q`!oIGe6R<LUr%br8{8}PQioI
zKeK?a(bLxDQK%cR8-MZ2aRNMB=!RS2Djd+oUlkTm!|OtSnEh}$JnX%=mWDW&f_Md7
z=JnOUA%2P9tK#;X1Nr-xW0pI<{?R{n-6x4Xd7;E1Yw}fPqylPe4KzW^-EJxzNyAgF
z(rj`^UNq7@sRy2kM`n|;fFcaN;0T$JD(^y(eq>2}Vv)I+nueuNClt?Ms=wM%Z*d(U
z0J&I1H3$t9KIrx%0#AtPU*m>_%&^T)FoD2i($e6#nMs8fZyRo)oF3`Ze(VZhvIkEh
zq36VKra?dN+NeW_8&)t$R(L^&JN6=T7ObR7nm!;6@FP053G=nu-P@dbeLE(mr)Ys7
zDqfZ9=$0k%yx=%1SOy6EV&`S<mnoehF2(grKF_aa5}s|m3@&&fdtwI}^@lB!>Bz%Z
z8G4sVr%lI^G;k1tOphjNq?v=y#U-1h*;M`NXE44bcmp6sK>!{(&=F)Bn6UE<VqVIe
z70$xfqsLFw!CqD;WJ?{NBX#D~xsZQ4fTaDKZ@nGg`@#3)%{N~6S5LfLQg#~r6sZB>
z9j^XP1{%l|$AmR)s<^3<$9m$ey2+C1!J_~KvJ+o+C+(`U@cHMT$4`Fp(|GXUv$7LU
z9R_(J-S{C(2}))q20VJ&YzHEGg~v+{Vh732wCwEBLBQe6ossGzxl(IghzdTj9S}=8
zOylZ&PcLxP;Y19Nk8032?yt2tK7n0>91n1yPVc$O7abQbqDrScxh%*I)CC?oSiF>J
znV|E)NAl7sp<_m};f=2@k2Of3GrXscF9TiT$pFF~9Cg|T#E&n2M%2OK9CE(sqO8+t
zV;epbkCfk>fK=yLcshkFZgIvpg8@n4g)as<EJS%Bjl1k52zQJfXaISxPNx^FJQyd>
z#ZR(mc%O8tucUEyu<&0NNl9DPBfeNAuK1^82jEVJn%!sqDweQiCr^-e@R|l97kqs*
zp$^)OSki!fob$ECn?amAxTTYrUlLMZm2$W#zT~MwQ8@6E?lj0Zf)m&pvLvNJM&!yG
zmvyz_r3Wk=E$QY852<8mmI6v|f){|($^Pxje)O<sA<IgKo%L{NIGdZX(qwi80}8%c
z0r3Ny0veONg{|)u9FmGOvNiqLO+dU!#*%O8hC%jZmG!Tom%(NnJI==GVL#Plp($3!
zO=f6a*NvB@kWxQf|H(@zgezW6Z&M+!&nh=@<d;-!Uet_hFp}T46j(A8bB%ukg$^Tr
zI&O^%;d<e-q;cU<Q{3_82^}VfgMlzpH#g8SgTfAPJJ)9t%L%ti0Tg@#jbv$-Ty4Js
zkeNi6%uS~FN!bU4bNb+3>^7v{N>;JpWtJbOft`wfL%O>;7MQ-KQEF1OQxL#8F+95=
zSvtIC02uLQw+b%8`ZXM?vja*xTW!j)GWqq8{Qu<t^4GBA3yD&+OOSQ2Kry;**i7s(
zV@bz`UK(bXr-~DF&A?S|$a4wrf6=e-19bh%lDY6Dw7``-q}Lz+%u^X^+tq=tf1uwE
zTndBtSmBeCDcy8ZNHV9r%L{XQ$_cbITP}2PLRB{4tEkDU(n&{=H@rY9k)V|!!EzM-
z3CQKQSl8Q?{`#T8b)3w*^FBy;iQlfjuHAxQK{m`lN(WsHH`6=qX>{nR3z1X^(?Uzz
z8()CFJASgSQFsYzc<I=e2j~qYkml;Ay7l|1bO5y2ns<k7Jf-XJvYN)yp5iCtRcVt3
zbA>kEy?CJCE)4yoE6MaI!Y3c-Qg34gqzo-oI-6w9eiZM;W^%~RZMpA)-__~;3b^c$
zv|W6+OuUqlW$JI!u7aNNb{Ev&`Xawy$k$F8cG9VzqH)qS`rqTTF*-3D*9ONz-D&p=
z2M6*fI^?C#$4L=G>f3OF2uGFE-{*KMCi<C7ozNKV`RaNsudH~dhbK{$t{e}@NfAum
zqUeKrm&}RAwHh~h9HG82JLmHwFv%qv`r2e2edYlcMKq@1ZZZ~ZW~RpD_R@U3`TE_s
zb^B(VsQ4{EUyt>TZKu(~%v9XDIqxt08E+gOU3kY36Qmp4yIyedE&}nIo?sCyWpf{1
zfihO*loDWad^l#O#-t`=RAT`u`<dcK+}Sn3g5TcWj>ffXKGJ`EeM4z66?2O>#CIYN
zPdQdh;~)LlB{MfYt?}1z931R=;fpZ}cT^K5ldbf3SZF}DYcGx^0SkyMQjQAGPNH#N
zoxojF2Z}gV(|3Mzs>Rt+Y?fmP;dgj&tiGMbI^vf*s)ntzz(9Fm5sqW=m{1-cWpP9C
ze{mW++v~BtxfaJ7uXDWOit>~@oDL7!5hjmxwqKlOl_ZmtmIlmK_aZ7vX2okk>GAjf
z_}`1a^+$g@2FIpiV~a(}t=Ky_ioL@##Zn_O#&H@?jcGhMJ2@H)(_=9vnYOpq<Jq%k
zitkBW9~_CP>3JW&$NfNidq;76bn1l!7F;}*l$noGw-kqyW2NuWL2PcU#me&2INIG*
z*&S3KjmH~ry%RtBgTLjSO`BT>zL$+w+2dKIx$lb;zZV*Z9?Fd!YwUy~zbFH4!xcTo
z+8m?Cf)G1Ryr`piP_7xDvhc~mCX?%wv8}D0c>3(A7m;q?xDm6{lddDwOV8=J{3}g8
zmI(1URsQA&)r2uE{ugJ*zNc$#btN{|S7Uc)({m{-q_WeIqYRx#*N5ZP*ItirzV};R
z=wdEqRBB{m#^rQtV?BQPlfO{<Jdb;?y%n=J?#0pBb?@@x7&?waWMPIKT3&>a;x4|?
zVRreP%HK{rU)z$rFJg+jl_n=t2X^Aw@^de)aHJhqez1VYz1__93@Hs}xCc!2R0Yzz
zE%y#M!HBW^pm?j@QOThVVX=$5!!}gUm>XamPN63)n0H|@>Wur}rFdVBk58)J&&Mt9
zHlLYP-C;!hBF2V>+(t1ku(qapcXX&c8<3w<)stPP@9kT+^KLUG+lu1!Y;{|8g&ktb
z3zaG2>zz)bWk8Up+~nPJ%x$sQ#GD9umLE6q0sBet0egNVOl>ahFLP9i10Tt9lq>YK
zDfOsA@;}?^<mL8LY0o?bV+rh8?vxkGr`mqu*9(M7d(6zKh=tP=UQ3IMz6Jz07VO!@
z#=IhF@2gcLzu!@Kz>gHhV$C*~BMK!mj^5=(oLSISzVI<a|H2n#8lET?zbdXv<+(yB
z($&eM^kKn~`4tx1@*R>#-p;E1f8*ZGc<c4sF+V${G`om(wI|CQ7rMUbwwwH%@FMM~
z-ZRv}gBVp_y{mV|_kQb*c>3f~{Mk=_5u4ll4#$gE3K$syWC7nqg~C-*!V6QFpw++j
zF&e*(-k}p=wVR&5?Y#3Sp-E?2yA^YRelx$OZzHJgT{_$|3~;SGdFvjO-3U1&+2zr4
zNw6D#`N^?D#-&0x+zMCWfG+;3uz(s~7y85Ohs)t%@5Qw=ZM?67%e=lCIK(gUdsW<i
zb08o8a?Dux8u;Ts{s(`eI)HQ;2}(&BfPl#`Y-)ZLJaHo0M9Ia(S<o8VLS83I@}IOx
zr_42tA9l+?({5;twZm<yq|1MX?SCmOvR7XGT=20w&`O(NuEe?eF`!dWsvRhFgb$Af
zhc8uV;bEU%Y9&i(c<XPjCbR*^RDya}N-IawT1F=9j4;|@Z*sS;2NdQFUqUMU8jqsk
zRpJ2)K^q6U5Lr*K$-Hb;0{}I4c!Q?%HlP>!jE^|va@y-d<!cUj6E=aUlVBOro7@VH
zmw_5-8fv~0VK);mE-y;x#p0bJOybk&5-tZY3RN9T^T9@gtrsFVj+o<oY{$?IRteI=
zAUElhaX>PI?3o^Et}<I*P*S8RcB6^`dS@L!*{MVMrHnne5zfxxxW8l~Z8*k|?fm0o
z>WB$HI>PzX!q3jd4Rr`PYn(HuS#U95hvgK^0O2cqwOdY8Vy|ZW*b0di01fQ0XZEG<
zbRkW7>gN|A7P(2MgF!qvQu0I{2EO|85gEJ;{LA|0tPu_g&CCe+B!86U10`GHl}GH<
zw2@9(@g`tlkdBh`%Yqk&mw@A>a5+9>d{SwuGRdXN<OyF@G2mirJPTlSxahc0o>)A>
zFJ&z|+v<2Iz8oz`r?f6W(rrN&I%A{>NIJiC+TlZI-Cwb(qey&mcUKNDsUr-==+Mwv
z%k(8qoWafFE**V3Qt*;4*}&t@0-VDWQ-3k4pF4eYsFW`7BpvB!l8h46bm)Xn8j7_$
zjS!q6uYTklGBD6~$A(A}An|g?Fn6I*k8o!o&TcT`kNi;QVqnG}nW=+gLnH2(4kIIU
z1nv;c-H^5fn;?-qmI^tD2zf3%ixd?I2O3Yjn^s7XfyaxzC1J4%%v&H-4Hv&ADfWgD
ze|t;6#kKGgfJUxjEC@X_J$TYa!msOI`9N2325U)gwt9eaJ;09aDNuS7I0YaO-r*v9
z<pUJA#8tqpxC?BuHy(u#X7enLnDTG4#VE26WHAb=9|1_I+D0W`f{@H6;(^R)nQn<q
zy#c!?7P)GABnCrCa|;3~c)&#IUg8B<7C(v~L%Nx!@Jo<Mev{nKqEbtfXly8W?0!0T
z5@kqix~2<HQ+3H27k&!UZi%yw98G$V#o*G8yVKLDgnRO-e$#|*Uj?dPv*BqwPwRT9
z)|Y(JKNH_>B_pKJ2?J<(+sQv+EhBWO>Y6}4cmM`YqB{>=uyMq_N(>&wzt{oc{`|*3
z`-f@yHxJ+W;otj{6s5+gn31hOwKuq4O*;ABOP>b*emL$GdUy%0GB^IZ0IIqxzZ$dj
ztIb8(mQdY*m;3AJ{EHjEtn<BefL*)x>ZhnKu9};2RP5$b_!lw(W|bJ@(%j;&;#=Os
zHdjB%+kfmBihsLRe6@G{cFn24C^H*5Te>bETRZMw6;}{FDvJ*0)-7qTkU=1u>@}Sr
zPj_B-mCl+~p_94NC5^fG=YR2sxvGW-fCaz&3T>BPRb9Nz{-y4J4{dSj;1kzpw|i9u
zPyMi8wRgo=*FF7Khs1XwX%m0RPxa(yuX{gTW4B&fzJjZGlycJa)t+*vIm<Ko1m4f2
z?Nd!S*wBkgNdH}x@4+q<H~v;#|1DHpb^ju|OYYVxuJY<t72o5pf~`q|?l14vq2hx}
z7NIzniwSN{U%);4ZCM5wbJF?3jSt0`lw};p1UNe=7)Q|m9#mhKiAu(N{%}fxxsNJd
zyoaes$i;gO9`vKLZ^2|gox9FS3qu+kz|-Re`uBX$#;F?@UVPZx+=|upHSfw`_tv7u
ziudl`ip9AZjg`*h@Nhr2cQ~$$d!>eBa%wteXJ@>~!-RK!;h!JxG5%rq5IcI9)MZkB
za$NoPVcJAVuCWmJEV1ite0)^=UwGFRi(yROb068gdw0FNX?EdO3=EIQ&i-+%tZm2I
z#!hT(?Z%dmWIIwgr&3IYYYhK_`2u#;T@UV3Vz&`G_cw9G-qPZdPmQ^A_g>7-Ed=*X
z4GxWZ5oT+T#e!q+p4#>OVH(pNY3#&M=Q=xVrekDsTH}S`I6OU%&7EBMI}!mO!JcYd
zb|U<Bg<p4sk%mnE-@A7&UVr74c=z3R<GuI39dEq;M%=yoiWg2eE^}vR%V}qzfYk8-
zg^U8Ge-^Z2W*mHRs{FYf@4fqu#`iPIgAsqc$W9y<_81>h{&Fmr^BEM0W0gJ9jCC~1
z(9@@n;+H@DpW@l0`*D7J5T{4GvA44p+ndkh(SuLp(dQq>qx&Dn`@j5IJh=a<7lyXB
zH@yQ1S-zq0xklj+{_qdu8{d2{re|jrkm`lX>o5yqW5Y2$$^BYmo-}`<yxH63el_Kv
z7iZ2a8**~$2zLju6OKieX_W;QggJ(dyoR2!C*_&lCoE2K#}+cN8!P4Ci=Fl7u{b*!
zZ@h9V0{B4r0095=Nkl<ZW?48J&*R$0201qWT1<`(`gq3qDamto5TAbdtN0)OyZ=x8
zi+}dt#ZUj@&*SGm`*D2s$*<$-lZUajxv6pYp(MN>qbe`AZr_PJckadAJFmozrJIs>
z+DC@6fcE(QC-LJy`%}*W{Lc4&5O2Nv?YK5DrgTzxN|(90S=V_^TA>Z_xL4zAvU^wM
z?(x&-@%Y)Q;(8bZWpQ>*>Gb&NipwH*nz2aB&O6sL%Dl=A3)`F`Lt60>EDOUmZP^vZ
zJOc|xyE{AHLAs~1?t8!GPW!--eUlTa3o`p2xEDEBK`~QKdXa=VCfX3%9_}>f7(905
zkr)1GR_zFJ;7&W{o7iQ<C%xo9JN8(-*;gFcp~}1lK)e~hGtT#EGzyEkAj)#n-ntn>
ziqh7~PtJ)HceBd!#v(K)#qdEed?@#Ip3kJNSMpotD3}lNyHJezSsv3UNaYP_gTmd5
zQOrpdxjZKzBa3z{Fq8f;NO6!Q+j<m!ic1f`i3KXp5s4<pD+!LCykvWyerwoh=4lyn
z{-m{64*1o3Kur8GzZ7+cg+Mj{)H1>%DIXhftnTx*4Yft=YJK6n<}NSlF1tLLe_|&h
z_doJs1a0f;+Gadme(v3T&(~Qf-f`My{{kT;M3FVqBXO1yTt@+kx1neq{3ZReKkG_3
z^ef#2q!m!@fa&}B73ktHrM;?Rx3TpP^qX5-*H1#yb}1^Z7p}qqT|DNis7%x6@G^cC
zF1usz#eEqM;H%KQ3>@NDk@GhXq7QoW@i&40;s5(j|KYWN^q>CcR7Z8ls1UBp0-Ax6
zK|Ffe1S)#!^;t=*tI}ngS3S|$li9e6Ek10ibm)bs18xM>@uE6ZN`r*W37OJl^*CZx
z3<@qOO5>Ko%f^>ouz^l)h1mWG&0c7bQYJ?M8p1WE#KDkqq!m1hH-Z4(m<u^1O<ZaQ
zVrl$0q|%Gb6=4HLrAH+fs>ZLTNA+NLTlKJ|x|q6B7${7M7nJm{c*Sc@ptL^F8EA1V
zz-@IlUQx%9bH~A>XwVHL9FybvnVph?=00Dh`Sk=}uN<l~!WLY1db!h3@Ao=U2Z4c#
zXbCU31Lq<DGJA`#Ui9$hsJZ=vBjI^b4R|qtmxOcFks%JmHRa-EV-Uhgl#i6W==KF#
znPzVV>-l&gvN&a0;e#^Oya?^0pOD0pFaP!q_WdHGW6FYu&zINWgvG=8#f6xeoKa^h
zJHnZl$SGqoGdLW$^CvvxpwZL2M+D%D{}#w~G@g}j<}2}SoOpoJ&n%Vpd{y<?=by#L
zpL`O}o;(Y-yR&FXyvR@f=3hk5v?Tyad-AT}_JDPmlBjk?7v3_Jw1$J-g!GqTW$a<V
zL9`9X&*bcM%r7Yom*!*X)=hWNI7VZ8n@e=_&bSM94k|u)r!4vIER~ZyN;CXr&#*f9
zbQT$ykevP+Ny3ooyqH<MV&ROAtIM(c(^00wN85o;06xxy1tU7Qk|lR(rQq(kNzutZ
zR0jt>Ui=bY7V5kUOh1k#7#_`CU8J-Z)QZYG#712|C|g{3e9FS41|j4fG<DY`Uq6C7
zrNb;f9P8*EUXl%&Tn?O&xp3wkmZg3mH($eX@ib)<`~}B}DjnT9rR<zv_b#uc`K7oy
zf5Tteu}e&WnJ;z1{;(y2hcns5)Y1uXw!yCDDI^*^3eM%!j0EsPekqK394BtzKnQQ$
z)P;_+27++OHE8gGKpCE$?Jg8{tUPf}iy@`pS%{3Fmcm&kaL#v}2{&<|9ah1Cblor*
zk;bG<n(C~itshAw$RSx3U5{e791aJ2GN6d&J4BB@H!nI+-a$*X?O8H3FMG*!Oz$Ry
zn{~>B?@;q!c4*UkrU#GeuUK(UT6FtJq(Lso1q!?&DN_@DL!LOO5Z!!Clp4j_<w8(h
z;GL){ZNd8|vlv^aNZ0frFvkN*@G>Tq?fAwGyu4LT=`15Qz;Oa+0g}FQXvs4@s-KK&
zO^-xWw`xeS^f0YBfKO&=u)9D3HhyJlW1rCGw_Hox=I}vOJjQfNvkb_G&HPMnAJ(C*
z@8p89tAiq~Evzw_9*}hamw)nq{=fEN|INXF_`mw!5u{|X*0ha4#q~qQTZ29g1=nvb
z&^31OUuJI7w%RL=G`m%6zq^2UHO-UX*?n}t%gn$_=}MT!WfnKk&!fW2cx6#2F|2xh
zCbF1Q2h!Z7$uiW^7i6QSTjdTU3q2%2!>>5k_}QEuGoB6Lhcc2x_KsdX3A*xD=*L|}
z1B(`2WxEXEQ>1in==#WC9%irJOGmh=w@_tiake$q4!(z1LeqZqxm}hgp`yQfnqJP2
zG|<L0h$b7sWt&^w5(iDH78>t{nhssacI=X0BfH@zT=rk#YP#ApG6)BjcTPV&P+7Yz
z?}1)>(($M&p%tyR+7*a*F*a}sA)3p$F1*a$`0CWWA6n@3*AU8>Qsz4aI_&C052<fG
zHsxc^UJQ(CEIFojU2f|CGEq7?IjJ_|y8FYk(=&17#*+H^C*Bow?b={2%&2dSJG&nk
zyQa=qfVqSGxP!hvG(KU1v4X}SjA!XP)Awdq0ml;N5m)7u2=&)<mk0EW4Oq0|G>jMC
zmBJ2>TZ?nCxG)=Y^Nbk>WB>3Zc6RnWo}iyTGczUH36E(`jyVy6@${L-EsRCaeBT#s
z;<)da8dUYjyTUd$HeyQs_g7!J>v0fc@FA72ndzw*(s+XpOpi47;m)xA;|njiZE{rF
z&VKA^?7<=zZ6C+0O;1k5!rV;m@==<SM%>H84z0`!p4S}?ixViuBpj>8P9xNzWM*8y
zySL{%%(!dHc_`sYUlt$fqev`|N%(-4yTsU;0-n2&DFbJcnYn`+4&Uewsgg{N4aef_
zRE#MNs59)8VXSj<cBHZ6W;}iJAb$GOAIC3${<Anb*b)UTMAoh}D-zi;EY`Dl>&-Xf
zKlu0m{rLUg`@<L>pNgZC3yu91_v6#pJvj9df`=+A)F9?6NCq!ds2q$cJ}g#_N}knc
zkK+IJ|NL*_XFvUkkFFbLClTdA;j;^j`_*pUy6t%w>@1273~}71;y6C)Bhsc-2E4;W
zvawLWYw48nt5jWFU}1yBo9%6m#@qK|C1uThKbKLBH<e&4h~{Dg3&a{XD-J%XM{NmZ
zjw1ot>BpUA#|OKy{Pd~Qgt9QMa>!9^+>!O_-CJ?z<_(p(b1#yz%L!f_eaWKigZua6
z^ZTF1Gx>k=>}jm6Jjrx6h@k}VWIb(|@)j7I;MmRY#CN{)J8}EgT`zobWajGXv-s2h
z@n6KxfBF}``}^PeqrVfczx}N^KD(woP)bT}$_3??qcO=#c7SpOB<a9$pgSsi%r6lp
zcbKsgXaDFZHnw(rY~l3OOf1ZD?4io7${hnNq(V{0hg8tURj)@io@W7eU2Vwr?oOPk
zJh<R08jRstyx=%XUsE7|EVgq>j88%m-<_Qe)rDP$!EP+#&Z0cKnz(C>9eYzOCQfj*
zg7S^>&pe1^UwOXn-9Zy_pHq3CpBnKIiQM(}d~+`zJYJ9I8@q9&cA~!Xa{ho11z12D
zR2iG_;wNQ;dhEN!y;#P=lYFTS;SPH0n%g|es#x(^E@g=_GNn9bA&)k~cW)i$-CKMB
zPkNE3#2p|#IpKx{$eQ2aDDS=tY?$L0&tu>1!oKs*ZLGqlSkU(R!*kJlp-pMc=`+Y$
z+iDgVYkQ2J9RKJI;kG;HS%`bqA=()}pkS_skDt%dOeFDD^cIMB&h<EbglTDHS3JSe
zj)A2pcwD4Ei_p1ajCga`*|nHfJZ`AnzjtRTUb}b8X5MShmRIBH)8}qunB6*5x%GW;
zDz|*NusAmn<AdiKFFGe2kNQ6JLKtBZsO?=nX&XC#>2~R%L(eub?HO+%$9ck#ezcp|
zyMAqL!vR{J=owza=L)*kjW98{0Hr`$zqI5jwK3hC1DANTwzPK<H6C5)`Um>$UHGD~
zD)fg5TorB?E{BP|7uV8cysv`Ge7-t3#4quGRos4a=q)F|3HYePfApXJ6O{u01z8zj
zOWn*YvcYzRXoD+mF*hKizs?6mNi$?cH>Xm`3^F2oHWq{}`X&rNpb{|QOl6U76h!#7
zqoMJWhSZbHT+j_-P^AH|<&*fLtHrs<_!DBrkAL+AJ{zYtE}}QW4`UK!uXuEndCI89
zo&ufWCKQp=OT*?W-ua>{3>#=ET0adNp!yMW$ELBThMx54Ud_&b@v04V%AGH1-1(Yh
zX=y3$-nkod3-j){L(7Dx`y(>vVusRAcI9Hat3$y=Kf6j81h6Q<S9B-bza)3>IN_@s
z>3o^Tq5=z8-fcq1O=$=nUp<WKrA;T~OfRp)LplyzB+AQ)yNoyvK{5~x$GwSFi5Gl~
zG?uvAitJ!0Gu^ZVZ`xFT5+8b!&euX&G}B8<w0KD-ItV9<J3HJ6jTalvzOPH2!@9lO
z9+hI16}qzxvlv=NGuBI2!5eR-aO@W-0)QpHM#=2<Kj|==x7|QMYc5VxT-dQhXJLD5
zTQ68TqR(P|b=~%@?JW;7Ok3nLGXz8#nGW=2zz_`gPI>1ZIPqpsh?x#RJ=Kn6P}fZs
z;z*e(`Bfp=8dxBD^PRV1Vs<(X&yL;MA@4Yo{#@n9ccO_8i%M+ChbJ9F%99sJ)$t*0
z23(XUI#(WeDhw~!C|nj}Ng48p$dK*uVh~q5;FL?B`7bNwlU-5_LWbC)uMX(2I%;H_
z2QP|`2Wcw%XDaV>ZoskVMJJAdkk9Dn7!bbLl$~<Ix(=~HR?J>>qtmFe#GPdfkXiJi
zp5;*j`Q?~*p3(UbAL2}jU~$SFWbq<T;KO<1*|8%dRCmR9SOZ0l%izLX+HKtF^fCCk
zI3*#%1G9x0?jyJvljD>A@{Eon2}Qm+?<jjDfGlL6(~VM5+>xo)Rks_`Yetiw;9Rue
z0ghkgmu!++2HGK;i+cu+kEF2)OTnaFeoFmI%EAwzn%g5`LQq_iSqr!c=~{#gqO-W>
ziLR&YL>;t4E;N)ij={+fE~@-8l{yY2cPT$WWO`*7NR0s}hh`IC7aQe*lLaXBMJB(C
z@h3vNRNfUQO~u8uJgAod<q9BwOs|JomHZ`-oGh6IsRYjJ5;mw@_)8f`Kj)!QDF@(p
zN};jcW`J;T@7VoII~>flFv_;Q@U=f<Eo&xF;Vp+YKa#%pDP=EmH5okYEos0tlo(|$
z?#Z^1!cZwI#xfMFV_-Hi7lO143Sam^huMBBBSXf|Ud+HmHg{z<KqsTA<X-Eg7x-}y
zv5%}KeV~zQSXzORDsnko`Go{<SW>P_HB1*H<io+n!V?2}30uhj?Em!Fx9~+Qsu`%}
zjB|&sqt`o4E2O#OF27xOoG#NRe37}P@FkjJz3f+E+Yy~A`$e<~)r{UBXP}zho)ve%
z_(Jw&egRxf*YSg~;uBq-)_U^==(i{R7t{2HqkQZ0WBOvPQgoaR_^Zfdk8HLI@YU<#
z%jK>(NktuR+&eFa0$snDODg*bG$!a;bjsARSLlt)WrB3KsptvGz8`m4UITJoW_>BV
ztHV!RH>|7uy(FzVq}XeEWxW8Gc7EWncG<UK*(Hqy(kVIiy8#(~zn{x=z-7E-F|IdW
zVA_vpu7>P;y*#Xh%D)TdF2(qybxX3VXnWy{!+W{A@ztrmJUY<puewuLa@<a5Idg%L
zfN<0y`uv0Hr;kp?=;XZm$z#D?KP-SThF}4N58UYU-@0|n_v!3x@5cJ(mc~5io{-H`
z8HPN8P5_KSIT?fTYyAi=$9~zTK9?MfcX9!P^yRgu@$?AyG>yfy`q12&Jv}uZqZ$+P
zVJIKB(bdYM<PO!RKT&kf-6!X?kh^#9#KQc%`exT+b8Ev32^_t~7{tfDX&gdZd~|f`
zMI<J|eKmmk@ctlB@^G)0Cy);h9bb+Ko1L3>e~(2gCgI04-u9^#Qs>tO<B$b3)KQKz
z@<a)ZtK^AA9PVUd;Ys8khj9;(kCGqCQ%$(@i3Fle;OMb>0)}1;PE%-oZ<@wq^vjvp
zW>*SvV^;&|!T1Mte9RmI@zVHCiAq?E8M*6+u@{mtS$}+ZAU@nbrF>QBqe^#<8{{q^
zcE9i;@9VF>7B?1WV{UrPALQPDa6dlz^i$u}y}Gs<PoF%FwY4>keNI#jsC${uRu?DE
zmsekTCEokyH{zXlzTurnbMrT1V0bK!PBku->@0jC9}5TE>9n`S5qc`4l6_-yHMVy)
zV`p<c9)6w+Kt9S(5-;Am9pC-#_v3ed=X-JECLeg_v5@q^U9KgBPy3K8jOCnP!m^u*
z-60;6$?;4gpD28G-R<um#Wu%d?jE}Q0v<EUuiw$CC-A{+nVF+u!9!&b9APtFWo(Ig
zSb0A-#652;BFQ|$0!H%V4!a3H_MIGw)s^M=Z~o~&jZZ)OH2Ek@%HQTDyVf=pXLer_
zyy8d95Jf+aK_x%q+!@uEZ~xYJ;v3)mcHDdARi*W<n3|XhmJ^>oc^d!ifBXNAKmC_~
z8gsML@elsNKZx&r|3@mj!%BbEL-C@lI_p%PUtAlAqw{M{ht>7%*xldv)d*bsz>$j_
zJIYag)W40*9hbpD;xjqzHiKQ1ESRwvc<e<XwFM_fs=r6BXXB$2F)}`(^yB`t-8k4g
zaM<jC%KC67Ik<8_b!~jyd9cq>s^UX^+TY)e-Mwv<e-@8(k%zpVQd>1WIitGH^$w#k
zH$y!gaGm4i8;&hyx9FDIE<S`EQM)ofGZ}YoE_&{SyW}3P?8N#OyD7>47jC;~ue=LM
ze6X{~%z`Q(xr@Ye9sBa@t4r{swvF=4yoT@lVm?Ck16TSiEPPE%)|sg(=Osr$?kg_L
zc~FPN()2EK0(2I>#zx20zUBfgZ7GZSb@2;I!lT@<Q_72IN?%SXVh)G4%-2jvxsIsp
z@UEc!B}~slDjmhv_-w&cE+~&K*X-^cl3<D-Zo(r2Kf*v>?3Hg9uqF@cT;cC9(oExu
z&5In0h$xdzzYFH0h2-d4fHwNN!ebGY#m7507UR8l-;{bi##O&KdiU|uwb)RZ+EIPp
z+2NkJy*N|Zo1Pd76$^2v#Op`;ma$Z(byb<>tzO-WoxgOubnt0zY15o@0;Hqy*1@IO
zZi;_3w|-l5!&lFL6ANGHX)3fyS-l%-Ji1W*SNJk}g>HBiOnU|SRpEBwav0cqaV<^8
z`zm;utX~Bj;+Ob;S^R!8ARqs7%yQxD|L~vZyn`6JI%4kdwNW@;?HOJv7->{xlwJ8l
z9nML<N?}BjWdOE=sdgr)JjjL?m;4*D0ZCi>!IU@`P9bQ9uqRa(SqmafO_q!_Q|Y4%
zm9w$tlkSO2>D?&khkW4EwSdS?cU1=14Y??^3g3ZM8G$AQ8Rb}bfTi<mdcbMa<!{FP
z8)(eP>IBM|*xY;yp1|%JHPIPzCqFRT4?kIRg3ky)z@-tXNmDxTd^x2Tm|kA@?!6K>
z)$wMRj2A7;L-cx({!RY!mvZM{iXA2OJYZga$9`eu3zrVPJEk5uC@sw^UqUpL@Wkvp
zl(-CaaM+z@UVOZu`C^5yYVtTc2W`GYJL5vUc!@Ibu+`vXNdo!HbV*w+LHR5s)}3CI
zGdndo>oQ);D7Y*uHhK3LW;gl*{L<kisw_}J3p)>7O~NC~htpH;_DVc(5z2V#Vv#YQ
z6|f}_vnyt^QaYNJm$wjvUCT8xLBy`5fq~p*$59`<yF32+j&q(j)j56o<XJq?K!8Os
z0Pk6UrQnI=5}4gp+*jobkB5idnV{pgx4-AR(D)h=(V&-DewmRblEkuF9?~c|TZixT
z^mM%Qjc>$T@4T%J)_FX5`Z$h!u`UZU`9%*MF5=+D90q#~B3Lz2M?A&Mv|;d?#~gA#
zeDbHP0q`TmSrn^zMERtXi<oqN=yV-=*Wn8dq^3N09#A}~FAM_lO9#eZkBNip0s}_m
zL|!_KhsQ4e4EpF`GBEc$(6Z#k*~A_`3_hIi$as-oRD!3ognp*D=jFu;UxUp-7T!cl
z*e>B}C$e4%r}QCSSwBlhmYrb=7r(>mJaeBJa*$r6r3bqj$k8?9zOm_vnYhCd6LsNB
za*(c0WSQv0Vag*A<$^`2=kg*H)rs(!Aa~XaC{Lt2G=6tLB#$|PFTEj@f=?R2<TH&O
zNSY*fDj2GdTK-#)g<lh{FvyCF<Y^W>(iN<IBxFh(V<dd^mmkMa=GJc^0UQr$${)Aq
zRnGrEdw=$9S(2T1Vjef{JWnxYWM);?fGVLdl9V+dW6RX0Hj^4elX}%)lAdLnp7jJi
zOD$sPZlEc(^e^Zipv)vwlOP+SNRc2&R27hwm6<W!d7dId-|zd}_ul8+b0Z@Q&FUAe
zbM|)k<;#~ZUp{Ofw)^Im;%GTT1ynkj1U1hDOFJZi(;g*`gd25bf{64tU-{V^T&2_B
zT9=bI(wg#}J~^9SxCjv&K-_S2Krblc4_jp9W?ztUI}|iEcnjO{04LAJ4|!ZP>=TFe
zfwzAZNDGE;?3nAG@1_f9$09+@PY7^?H{AeNoq{G&z%q)+bSCIR8(~-Ttjy3s(q3t?
zLn(jj2q%0^%s+G{g8&O*9i+uUhREi$kS19fObO?3f);P+qhCD9@6g~7)^x$}T%`?2
zobV98P6m^7v_O|n0>r5#)!qUG-{@?&HCVR~LQwRL(d`fZ{{OX)<li#1g)hLeVDhVW
zpwmOOkDHl@z8$98bP=}E+pwi4zbkAH;P~Yq|Bepa#PH<LkT>>bjU~n;J7&v<7A}4O
zldXQ6weiu9JO*#FKB(<HLx(`Pj->JcY;WOPSmPo?(Ae?34Ur48BFv=ns9>Ai=2e+C
zJyq)5Taac!mO(li&zL<v(E5(}jHY#*vhlGL*s0JXfNavM1r7#)cAV0!w($^Fel1W?
z9rS4l?@4*og{=X5Suq>4$Tv(U0BNXV88jV!ytu5~0WF?Bh8`Ceeo0??6wQwi$kW_|
zaMkE|8sDw~one$mqZ>gP!flkchwM@xg>gMkg76!r3!Lwr{0+wA?%X5jIPQ`Yh+q6C
z$1x0ho6oy=1HJz!-R*F=@8U%spcf?6#d9>Jz5-<eo)t&GdS=nz|GGcBT<%SCWcQ+9
ztkeL=v9L?w=?9Ms-=o^wKP<cZ`<`gc9LSU>VL4)u$zdj9JprqWN!7Lp#%?M4<xCJW
z{)92*o0rsl0)u$3tge>_+YigfA3ZKl9&UN@i?)hBBi|Cf(m3HvV+@WzWQP{VqOq{^
zV1r$0Ti(&Z(R0sV?E3pYj4hUzG)|tKDn2%l?+{;bqJwILE{{DGKR70gtrN*fI#$<K
zeQZ1z$D`rr-4O~yjbe8SV*tLf<0ISAwJLD5j)I|QGMsOC^J433Pgrxb9Q;`LV(|*T
z$ilIC?DFD8S9mZ+=GZpIPAmejkcY0(gLR>}z2nFWG3Q>q<HbMp1u&)(bB*U%;!=7Q
z7srAz?_mCtm3%=5jLrN#r}??Et}*4ezxh%5m0$dh7XU96&(ZOT7abV`pPU?f?EC7~
z%W|x7FFSI00rX5`*^{H)a&o*cT)uU-p|q|puQ*><H@4LG<#@|^r)B@}P~&J8<1WgD
z!t)X+<85R_9~;|SK22w9Yr8Cq2W4_~b*(&k@}#V<ZCGY>jQmcrZp2C1@{7NW-_%*M
z9yOkvl@3{aqAp#@@09WCg>-o#Q+mH(;YN8t-ZDqR!a>S^t1rAzoN|qq7Z!A3hk*GV
z=vZu+dduSMbvZpglKWoy=)(`Zi;@Mj`B@DFug=SV@&EdNmjB`({=;&5c2Yk7{3oPf
z0MmXU^O5xb-S2$6JbwI0<??m;mw)=F<+IOzTDG?zlwbX|UoXG#3*Rf>_~ctY9urwv
zY+-I`=f(5#kN?phmw)t+|Ja{D`0xJv|9<&z{pN3)j?GfkMf&-#DT5kEqc4v8d$o5`
zp6wo}4ylYtpX{L8Twg8^x7NzW8lO*~Tv1Lb+i%OI%GBZUS=l>er`)yb=4CnI8(1nQ
z?6~2(;_T3RqdaBj5TCi2on3I;lt0=N%ADsCln0oM{VDG(aB&18bc+iMGP6s{FN&)S
zx6dT-X@Mn;xjCM6b!9>Il^sq~N*l+tQl3}b`W>naqZ{&PL3N6`81mv=a=+L=F9#==
zE_=)uAQQSocIvq2?v%OYHGX^>4n3TmUdWFHJIYAXz&wQK_jJ*w_+@wrFs-t`w!%@F
zEU4d;zND*j=}qbM9Ega9?L6@@=fDdHw2k=Ph#zTNkUqf8s*C_E!qE<KdX!(9R~?{E
zuyYH&Q&-xBhudQDY71XJ9ZLL(lW@p~9$0uIFK8oa`+87%MToS7^NBs;F$A0}dbK(i
zoJ{>O!N3FEak0o|Ui{>dFwECLNDjsi1bO~wYoq-2U;bA4>%a1S=l_dWd*!9d?(X4<
zo((5{MqEn2v-kP#8@t%3+oa1JyLbaV4m&hN&wPyW`16-Px-rIHJ&zVPcF8TbF>WA&
zBYxwd<v$WOFvNvJ@&g#9pX8Sdd-A8oRiRnI_1GGAh^R1W0cyw(809~KNrk)PydCCU
ze1K7Z?C;{?k{0V2Bz=?pz6u=c`^fojlCJ@ydHJ(IUHJOLw{1Wk>14v#dZ3CWl5`sD
z#F%}!HCr=vxzZ+#*{Z?oC*0&GuReS(1|+IV@u;4z+<*c&oj5U}uJLAJK=$fKyp97h
z>}n_~7eU9d@*_aQ={^uFR&f)d(=dRJhddk`VFc!cr)P62($F*L!$wL+r#KSrP<muX
z2y~90+;UnLH-0!U!m;Pe4VkO#ICVANYHD>!&t2HakB{{aT>=<Md=7_<;PDcYw;Xuj
zJbLu7e86t6^)=IXc_<pU|2CBeeqiF%!-WTo9!NYucqq2Fv{<-ecbBO2&u$VPN;GIZ
zbi%vpK{wtQ%g&rEQqUPYyWni(OLZjKF_SZrofW?EE>xTF2giemhb(MOWd0<;li5;L
z$|d@&UZFBnrxDR%c=$mDy<A=SF>#f>agf6ys*||d1bV}V-XZ&N0v9n`(%{DjnPsDQ
z(J`=x?470h#g0Lv=LfLD<@u&tqiHf$H|epiq#W{<Jmgr7=g(i1AOHBr>WqHoP8n|~
zUaI2(3p&OOCRWtpU83`^4i5`e3-b%gI~Hng{BZu_^UwVOT*A{)@M4n~C?=;V<%dkc
zBE|wrEDKA&@V#FsfBkR#H_P_J?Xs)E$e;b_hh=B)m4Y#d<L$b$LJ493MPSMyTk_$4
zPcrzyt`02xS)}4vKRVnlGwQI=NhICr5b39g`IcAJpa*Y$qepyrlkY+U8Pqht#3J94
z;vhW7dCV%!>6lYiIZh5jCX(r-xwFUsQap&8fg$Ix1I%I537b-9$Hx{*AuJYChq6O1
zZ$q(I>~c?@!bLbbV_rm?s|$`w54#IdAOmpxqEdA-<xj_(j-xTs31!R!8O1@Tv324c
zpDvd>%I=$~vb?Zdo<4d~R+m=Gj5=`?opjrjP8D3<3Pf|r1r{8Uw|WE86Q#0d!izRZ
z96-np8KxIbWH%2`xUtKD+o_ksCAbnrBe5F?gkgnBDAV`mG!WX{TpyY|&Oh0Vue4q`
z%D;wa(0C*rU0Xp!`a|FDfO(3SO^SZCo2vzgiy+Nqo`4;_)kV!j5R(`9v@kNO+Y%bt
zAsQRwyR;jhoCR7qrw)H;t7#2k#{s47f{wl-;+=FlB2&jpZpR-2<4kX8xXK5f%~^Sc
zj}bP}q~#i5eec5Q-{LX!VG*PYgT)}8!dLa$jlGYh;lXu80uxnt<g(008YOpgB^5y2
zl}1MLQ2r3aF+$VAVQVrCQZe9LajK{28>cayJ(LM0cbtacbhpK!u(+;HHJOx7^MI^{
zmEG3<CLnCphwlWiRcsRA5B~oD1<cO{h6`UnE3-)0*$2u%##?v{+iTk=Yf#mH0t_d=
zLpSd3A^n-)Y+7YJx2XLA^Q&4L@;2yM0Q6z3bw58E^fH9WP=?+<iM|JL$=Q^JArDAL
z+ydNH<Qn!iObSyiF~&l5FJ25FVR}0EHw2~<Iy`b&jg)7(di_?NSYN``N)xt8r^~LL
zeO^J+%t`4OVtW7Z83(6p+)F_w@omuCF`InYD_+Jh?T!7d@*^0Jr$LX8-bd-+F0PZp
z3f1z#4&MiKL+c=XCu7Am`EfQ(O=<2E;s)2+sMZ&8po;6mz$;}T{rKR9x*WC#L+EiW
zjRV{m`K0hbub&3ck=9ZAyTeC5YGZ@c4?t<CW!4~)#5b9D%-B=bu}^3VE3^r#aeRQj
zYoE4OcNz@i3XJn0EOEz$cjM4Zmy7&Gtv>O+8T}X3cV8<@s~cruX+z_I6)#3G5xBm#
zTpm8y7VW)q#15H*gK~Ppj)m-!SyX?YF%FA6#NkCRjgy$*^$TShXRv6*t}nls2~KWa
zPUEdGUIwHO?nN|>AJ#dNOmS|ktd!;DC5;JYyxWGOtO!Hgdwcuk<cK%>uM}>kY;CTV
z`GrL<j_vI6&7c#7Q{Q|=GHD=jddepO&dY_y8%;NuOgu;4Oc;By1X&gqyc>!MXyj!w
zo)amczpwEV+GO__J42Ak6T765G0kg@17(u-w%|t^SRiDl4&yhUVu2hgS0Z)%OrBp|
zU1Y&P`lHM-=i-+DRSw=V-m;V`w;CU0j)1t)snWn?z4YOi38gRd5S|c}8=5Z4o+tfe
zdKpjnj67slI#?8V{ODo%&bK}(-}vUomX(Fh-Q69JPuDj#M0&3rpBzd5hu#^)dd}OM
zQy<Ou7eD?{dH(EKx$seLSvc@mM%R>dK<T^|KjO<Eu9S_<2jy3O^{<uhfA0rnb#u$#
zkzz*?N7|DY9;-<<77LlrU<}Rf9d?0uZdtlxQI6vvS)8JrDH|PjP2=u0jl&s}lQ$go
zb}pLBYnA-_Gi6R=a!&WSmU1px3=$sumGbtQ@v+-`#&8<f%Qk&q;h0F5ep%RLG3M~_
zuzdRIPs+1rUzFoR7XL1l#&gx7h4SzIxBsS(<U7;&_;7!>{8;1kpM3h4E+dpR77*dg
zBsq0}BPaQu+{aI!s!VUWTs(XJtnBaZmT!IY8|B~qn}4%B`rxS~$}yJAOYto?7N!5}
z&;E7!XaD?Pl*6O*vi0CedHTsG<;jz$WmA0C*EdxDmaRL=?Ee0-u0!Qh7O~k0M_+zU
zx?Y%_Dr+lCWNO*mSoe3pSfG|Vq@TCtNcHd4?m^i*JT51v7phzAv^n;i7G;*5T9+E@
zv$#T8R1r`fvRf%H#j>!xthT~qYW*(GnZHu~Iz3Sx<+#ps*Ez~J^0N3v`6s<T8qteV
z(x1u{JL%Zn$@kUfm^YxjslG2Q&X=|IRh5Am&l&I?ITlSgmXbwF<|Fui-tNg&IXvO0
z(o2u?`9|QP+9Bo`*kMT%MH;SoDOdF`i_m;dfqG6kLoU@ll?P;^4*NtWn@YQAygN;0
zkmFj316p2KKRY{ddMKBvpY@sP+-2^9<4Ko&)TZl_(odP@XjtX}S^T0M<(Nh9>Vy}{
zR%qg(&XFI~TYq{&?FjL>jz~r?>bZQQw%Qgbe$y+RES%AX(Wd0(0OV4|qB3AM5MF=9
zr%Sr2EX^k}fZ!RBXB$Ua(q_3$5Fd8wrH$yHTiC(zgmM_SW1*RuzE>V^ua;l_{&&hR
ze(yWp`N!h_p~@gT7GJ!4)h*1=O!3r`P8i#Tbr)BI-k!AEhAhqp5*2rQ`MS~HHWSuu
zEPT{Xg7TcfkXHW|PlKVq<LvB#@RqyM0j(bam{v53GExuxjH&J+&j$7zAX_?yFv{<?
zdk90nN%0v#qMxz2+JfU<FdFY@JYNTT@><}&1`zbC;<NwP6X_i=o}XVIyzuqAzxRi(
zsU&%npBzHEz}q7g+Gtrq8X5zIng#=UWumIH(L+^+OEsO*i5I@$h^G0w!7!eieT_)N
zfm3E?RKg3titY=0rMESWAXlSK2mwY5Y=BNcIdn4MlaTi7=mey};mrVx;#NT8<!<s4
zft>?cgSF{e+GrZibcq#Vh64`c`@of%Z>ES%rpe{1GWYyKg`SKFGOETQxf>_zP?x$m
z4#v?Abo!ot@PXe%^0&RZbV90-YM`NQTY;G%r{l{*i#JNz1U;R2I=5`UW-+3rk!}ET
z(IMcWpY8dQMvVavPcKfm!DK;W&Vthkpp(xagtLqf4v+khKrc+VyRjx6y5Nn-NbN=w
zOVt6sl$W5ATfMF;scxtn6z+<(0j`o7YHYFv%fxPryM@Q3ox@IA@Xr##fD>g$VW}I*
z6BkZF9#WQ6=?%Z+UcwlWFrZ|D3PmC{&Q@d%np;OlR3utDgs!~}LY~qwV#n6A7ca`*
z&W?`<phKB~lLk5JaMN*RhoH}8R!4#hz4Op#yZ>jO|FrzefAy#3vrj)&RLU=OK)kY(
zCf>%0pa`i4zpT8)$A^f2_1Ax`eDC|;ce(sR9pgX$@ej+hS1%<SI}_=|UsZaQqx;gI
zI-#VOMX0G6$w0fMICwzkjQv^HeU2lbj1kW5375(;Z-&ufvd&aK*`CfDNu-qlDT59=
zbhHCrC{kKE%3yJUqf>H30NeZNMDZbM2E#eSo3{qf%DFneEO61uMHce>K1W8Vb9Zk_
zolX5XVvyrhc(c?4R=FwbEG%c&7im&jYZ0MyIIqD6JUK=IzEmH@oWT_3O7vb3Qy6%;
zyr3sK-MScTQ6D+Fg7}<2u2Z7Nt#GUA(0};&!?L!-N4M{LfJ7$^kaWpc#bw@>ExC;8
z4xl>f>HNvfV9TXJ;VFnvRhdTXUM=I`?S{rq;Zu&{&mZkU6}<D%VFVBn2u$TyExL4<
zKjp|+TWZLT7N|IZ<HcQRBV7wedgEmhRD>5HgtY8JL0~@61ruli0s9H$Y=cwt5tEKz
zb(6+wR61ByK7@x7r+GF@^_SJh87tpNm-PW!DAx~%5mcn}h%nG$Hm^#Eop2VpL2Gv%
zHaTTNLraT5H_PD2bq&&BfozY?&Bru$5sKg@<!-FymrX|WQPXBO&V=a7OQi=V3mCXS
zVSD_sR5#F*3d|)ynuIsTGJ^t9WzZdVDKN6=Z#j$%OVeYoB)K&Z(nQGb;Nhu^4DYHB
zZus)|!^nsV;b%(m2`x}9W*VANT&xaLd35RJ3PwSz=0@U#pqDv7x>H&H;6MBM?fCj1
z{xAObQu$D}9vU9qjMH>}4UZwaRU@-eQ*Y6E>hCC|Pt?QYIJo-3`y@zw@gMRS#%YCC
zzea=6?8**9ADrjLsC>YXKQy_<@qvcR;=%xg9;Y9=vAcbL2MsXp23lN`^bMG|1aPU8
zeVl3%Pr#GDc)5byE)PK0M&I&NeY9M+(*%I)B^fRHcgX|X@`KL+H|~Gi-SQ>)R(9j=
z<>(hndbkz0;4WOBw!maM;uyuF>W<dJ#XYK_<V_zkd3%>E)i&mjJ?>lK>@|uD%@|#U
z+wq3>HhvQA0OU2|uS8o%6ZwqhkIQNWPf*<vKTO|OUb`o;sL*7{7_5%HCP3pm+zB_S
zt~Ut2U3Wq4WYzi?81ke~#_{Ire@@T(`(JCDPs+;L!?Lup;qUXXYliP;EiNs4Ou`X=
zOyslk>O>}Wo1?>+AomUpb+Soub~O<X<1|E2TdgqyV*th(EEsWAIkbKb=*NUKW8|6p
z_sTN6HddC)`r3+*{9-bh#(0)p5I1kjvBtZ{N2lgFJIyhTiyn9ILa2{M6rNog%j%yq
zUcOLxCcqhKBV!grII@tjIJ-#Dg=DJvgFjZ}Ve&k594n|XjOslq^@84O#v!<~i;Q%T
z=cL`7bmRG*7gsgb@>pcfV|>Tyc>-R%Ltk}FBe~>omU*13@b?&3O|ei2pIPgiu@z$(
zjxs#h-Stiz#$jG?klcO|QgLJ~sq#*`nOo51qbL*sOJD%c(a3+#M<G6ZuvtF%;DfRt
zc{xUs9X%{^uB~m#T=0S;i$Jf>Ps;VhQQ2HwDvuv-JI=#{Jui^HV0^5wo^+QUUdq19
zae3z&$9bXTx_tAKPs(rq*559__UnJ6yq3b*b>yQn<&R#_v3KBHG5M`>@|y2Kvk-Z#
zd^s)qyD!U+e)OZVulP_IFMpFR76~{K^CP8!mmmG2@cFsQ!KKUI!-tQ{U;k@=qkQX=
zZ+lUZBmB-TZp!gFCJPtRiF89*<U3tg=f`F5)tBYfi)Ur$<@56V`SWsgc<6G*q7A#(
zN)EWIFjbb<Hp-Uj1@W`E^X0P_$`8??d*vrb(Rl%m#SC35t1IQ3AAML>q~Fg!|I#PQ
zY;A9sU;N(p$}j%HC*HC2=YRQW`SD-;rOM*D$L{ABugmjSyXEDpeZ@Xi*0!FMrPZx+
ze{Mnf%?_s}@z^XIEKaLVNdSM-jPKM@j+gm9*2;=^cugrk_Fp~s@utY)BT>bN<008K
z$<cEAhtm6%>Z-=7)NvNi*y%+LrmVt~BmXFaltnKpOOEB`)v~>{?VXg2&6%eF*mcQV
z%I?mtk2LhWMBmw#nlhKbyaIKN1u1k+-QY-3WMAPMTncxiGS4^Um*i&A4m~0VJ47k>
zE6cnH%VOANIXpQlJBJrC^PN8AUQrvu@u{>MeCv$w0xC(aUz9<&8@z<8`bfI`%`2yq
zx|Vv3jJToa<#y(T$OCr%Q6^~@IOdSWbjmns0S7<E1ep_1J2#z0J$88kUhGr+@ZjZS
z&+W(!{**Zs@l3xo^MzixvMlVV<mf|Q0M5dj7r#X7ei@CF47XQ`KfH6~Dqhr^Ru=)j
z&qq7ba;QC{Xu~<}2E6N`0{i476ewUJ&~u?~lW1qb69etBYBp`0Dbwa5v&)m15jTD=
zOEOvbe*9pgeEZ`M$~QlLQZ`nlTa^Vq{qpP?r!jp|&Q6Xc2=o#rA_!_c2v?!EBftD4
zizBIkzkimur@!qce6^Ei<zc5M^B`Ti59o&`U>-H{QUB4nfG)0q9Ewu$HCde#BVn-j
zuF7wS5Hx;681ewd-9va+e526h9MGemcfeiozDu4-5PIj#JD~SBcf!;<@>O7(uO@p>
zKu>=)W?A_9-9MD{N))98sP;BlmITYJ1{N2lbkTYmhtEUSxX$gWR;dz_%8TOg32mtI
z7C$$IlS#PNYGe|xGQt~sy~o6(+E5JPDmWz0O6Hg`jjkSqQNw2m{h*0-_U<5lMm4;n
z6)lCpvUW((#NE>9o2>B-e`Ja0z#OJkf<_ZfX6eb{ghNNhVLF7<lyEU?$lT3~V2zf*
zqzPAKvFhBgga%36$U7~0dxAG+9zJ?lwzjtP5UU@JbO%d7c#|g#c(%^dVWt!7#{QZ&
z!*~lUZ-L}o-isQLXn3>hzTZk|mBzbW)G+WC1s}AeW7KvA`8!!Muk|p2*V*ZLIXXJ@
zF#_kZ^YDd<`fL%f%9yyKLy)M&nzus8&$y!#P$mq{w&1Nd*<hJ40kA>Dzs1v2ROlnH
zvq3}k^+k60MNPqT3B2QolEDLfz{g=6K5mDFM~48&W73%iP0E*`>Y+wqXl1F1%m9{$
z4ht^m{^F8BmLBH%kvGus7TJP2nLPN-QZjI?9{I`fgLF7{HBkBg{>y(^{^>vcr|MjA
z3<VEKIz1v+nmiyvfLa#ey{%5!ul?$;m0$e9FO^MoX36L08eIJFCqF9BUc8hX>S(FM
zKu8u^VKyT@@rF;CN~e&H%^S5hJ_1oO@-42pc@~B46~1dmXL3#*zZs4~5ifMfLZ8n#
z_i-Cn4$B~ua!C2kfKh1^Zk~==?Udcu0IlsFd#gG@=f~wjN#gBBcF&z29hKAL6UW6H
zpS*3!x3^f3$|9Hss0!<2KGaF({CNgOEXI)!bZR-aA>}*yih~Z??4kw}OB%GR&N4`x
zU*bd3l9{vc^`nk?kV@{8x#(UzSm<KmnBzk!W#|k&K;pE>a;3o$8MLO(>4%R$E~`r`
zWrnwB-%@WyNIJ<(E&?DM9Q;pt1C1qb6X6B{)cWI}Nm5>AnnE%VM<WlFE#apd&z6S9
z;@Tv)dCKbNi}<)|bWj;o6B}w3S-=rW7YgKLnGXN(2EduT>Z#s3Spm~KKxJ&F+^{jB
zEGnaEukq`t9k)zeCPhGeg)|K}+4zWVWJKbgEP<;&VvRlsgUQWwA%-gX#)2>QE~EgO
zgy-gGrBP!v{DtVTLewywzJN}T+#JCE<_)?+Yu}PS2M}}!nGPt3pXC#D-nxPt(v3(b
zyY9}X&;vEvnwCJMulEG3AbvuCsw9S%-jFupc;XIk+f;_(VXuJv8YQiz<5Q+MU4)a_
z%4;}40FQ(sJSfUN_V)cPmq+|e3yEZu+kU3E4+O#4#5VK~kUICLK$@nnFZRV0b31{i
zRNaEBKA<fZ?xf-Of6fbE)oSS5rohmRE}DA;j~%DXCirmjc^o_tAHev;^=%p2<ZArV
zu8h)k{WQU?@U>0l8v6A-eIM|RU;G<Q%u#@5RCZue91Vs#G|2icdf*n1uW$oHeT>K9
zr;wkAJ#G>Y;466xM)K2jVvxoteZU9%kS}Jlw_FO`Rfg`8HN54=-!!_!uHl9-8h;?7
zjuIwqV>mm<ad(%4AtiKu3Ij=Jy8?u@HiU{#^^-9S51nJKNQbLohtTUX=IuInxBM*(
zJX>9w1moob2;ash(DNOI;2ZbbH1QAK_KrjrDQdi3X=%Q-2vzZ;)Rm`T&}tjN_q>(<
zbo>krKd!$vvDEOR5>QTW>k4Qx+=d=MN;8yYoM-a`&+~~1SKL!&c7D06ZagX*4?ijo
zo_y%z>iiO!#tAIA(649Q!DK5tJ&uo$y|A*Rem0ZPybNa=;owOhoA3Pc%`6s&m~>_F
z>|Fh44RO4agmD2Up0EJL+xyp-7iCuTi*s{c_+p`jaK|Sn8hdEWGQ%-xt*@_VfW|lI
zp0Nx&$k0Q3k&Q(MFEDt*T4N!FX%oThnxOB^7>3DRE-!A09}6zD`<!%f!`MT3E+(yg
z@&uE|k~=T2aw-h6Scy_Ed7N<~x)7t3IrRxlA$J=WIVVK#MJefuNq5E@<Pk63`GrZ*
zG8xUH7v%?@Tufr~JtP)g+IPa4YoH97w`j9C$A~bpu$aNb_>6Rde2z-M%Y>P$U?QIv
z39r4-#SxV37HW%UyvRAFIz=A6(Ku>K`kz)`eQAECJbm)8eD}NGDi0Z(YAosrew7pS
z$9Kwh4o=GFFLugjUpz0zryM_d=EaG>{@2*?^?fgJa$Mrc(Ls6f{CPPz*z@i_#<=VT
zqWxmgi!`(23K^X*$`AB?3n+WachZgi{cShrLn-r%%jMTKmj3ttJO5sJ^!SO(#?70l
za(aGK&NRm5RFfN}@8<fvyn6mc`8R*|r{%Ms{6#rGC2mgF$tfiPaxwbxg`TD3_0_fV
zoBz&llwbbA56X+3z48zL@t>5JdxxrH8jmimmaT`6Rc}^&8qA!=s4R#++*&V-Dg*4Y
zLw650H_G<LYI$>cT>kP$|GIqs#g`t>atz|k++z9c*>3sqr(cwdD~(lWm&$9U<LZVN
zI{8NRa@pSADqCBds=plD&zz9Ts`8pPc~0e@Z@o$D(k;hxUTM6%T%LaT*e{~O=S+NG
z?i`d)fAV>G{_>Ud&5NUa14?yJN%l9yRrYwn(`6k^payyRd)ulbQ=(m8-z-}j$T3Ip
zmzL1o8}Ah3Xv4j|U9~4i2`@Zz4a_s-1=jZB>XgU(%tz3svGB*70pFLS9<t*TV29Pi
z2V2tHni`nbUfd$Kwbd2n$&%#1=S9$$JI7`B;JEA`o_Rr(`Lgx3Rqy<wm{1p}f6V3a
znG5Qh=jD`s>cb6le5zZ>Lz_rD;Gg{b{XSi^Z7f##yMxpnrGYeaJR#*4{SelR=cL^A
zj6*N8Fv@~H<-=`^=&0AsNoH4<lBaO&uBGiFE`V0W3uMHlHkG!Dm!3(B`ABZ+0dlql
zE%5O6!k6f{ShQl{ENP)lK*F?V4ZO&MNm%KIwhUgz$&UpoWCcTc$YNjBmE%lXM|hyM
znaBY@7A$E8p<&LAJW#;0x-?%NZS#Hg4IgW{t$MSyzFPKocFK=`{6jCiQ^sKH@tOJe
zvJOpnx3X(oLkn>GXF02TU>};TKVh9HQcTjFFyT8MrkMbC=5EGu<8lxCEkjvIC%v%X
zh&g%M0DJGM{DufY<2QsM4`AFqgm=X^3Qf)d{h<Fe3GRybUGhwVuclA=jW&Nq{$DT1
zJ7Bat{4C&wufO-Zf9QD+0~J*E;chGj6vh!@M<PFm=`*sKevm2<MYeKqAXolMgf&Gm
z{$YhncA>zA7+iBHRs!Ny<i<j*cFHtLgeP>mIOOi^P>afB71;;@d(^l{So6Z^uz>{S
z7TT6ZV;Y$qLmCIV1LQ}%M32+aJEitD!ccjHt^OoBRPAo4{sB4E0qc3=wuPJQ(I>Hy
zu5bxWaL5IX{M0EREi{I_q4N0AV;?KWE*IXK0LXV_fe#N5+w`ZS=?_inA;sdwjqGfH
zZVxwhy!b)lj@xTRo_N^J!=NE=n$hXaM_Fn3`34hj26^F&2SR<b0)9LQyxsXcXCv?K
z?0Zl`+{l+aqyy}vDy(iaIKzApzz@J}eD%QIko?e9yBlQ0MqG_saCSK?`o$ruJbHT^
z$(L})ZZ!T7$s$n=et;@(a_j;A(r0~eL(?Y32|tqkHNq~B$p_*@m%^vBF^I^JFdm3n
z{?H`fZ8cl+0t?T4^3e}x@rQ=Iq2r(krE&&3JkU8ZfitEr$`60|!}2G8@+W0Soj$&G
zh1}#1@*tlVz*K$!j&1nfcfVVH>6d@GeE9T3FO;GOjyU{_pZ-LHgXiWcA#`#6^Mz&S
z?L8J;86eW(BV2Zkk{&*qT0eqHuTSY<FvyXE$WPiB(9yADQ7ntBXX-esGpjVPut!<S
z+l(rox)^A}$Ac~Ot+Y^v;mbQjP8W4R9oRE<NDue-%hBOcVLLwO!p|>gkXm<pQSSLJ
z5$AEcV|S@==k-I??gXlnPv`oaMJV*8OFD6O3h3;y*u}t&V^dZ&h+bZqcSn(LK(S*k
zZ;w%)B{OA7^11V+bfqln(hnSEil&Hs;e)m;lB_MSmJc6(sLu9knPsu?tvaNRUF8Hl
z>jzJFYvFGnR3s`}^oN8%)y>WI0la?p!mj~gZ2E2)I|AY0-N-ZPU+F4OM{NSSu{|wh
zoxh#BTQD3>i8-}rjcKc&N#*MCp>{kj2Lz-Zh@cXgUU#QZG_hlqDH`(@m>)r^qxrP_
zApH&tkpOX;Ob#e;SV5DkL372w>Z2jKkXvR@_;s<@#Wk7ov|mVd3(i==+Mio^SCOhm
zb;#}@K*im9zGJDpJ1OiSq}-K3T;QRBFLK5P&LLaen2nc+hKr4uF8B5gbiv~Czmvs5
zgM%n?+9z`20)f>5aS+;w1Tz5;eb^<s1RGGaoBn1kKhjA1>^JMGpM8WUj@H+RyLysl
zV|54B<TO;f{)t0Y(_jhJZSu-*9G@7BLQmuV*4=m*aE=iAx8)E1=;wFT;rPN=gXCkN
z*~ypAKJKr-g&yGDXMP8?N#h271`nW*L#=9UQzyYtmPrd>=o{H?$y}kMQyw%sW+Mr%
zrw4jmk8fcc4Bg{?z)+un?PFx`;t71Edn^v~mpc<uf$=a6zKS<+JCD=n>Ol799yeth
zF~QS!>bxhwq{S7W<0+%FE7Q}j8)ewXA877jr5i7ch9U0(1lO+HG<J!(`;^tI#ns0X
zkD)%H;Q}+{36EjErk$z&GFmubTowp!*O76W@o?D5*Tmfbnqi$8LeF>9?{3-#o#evb
z&I8kqkc)0Y+c4xwUt~zr$H4Polr~dXqc|Y(>T=*wKVaCFLEGB%2EB{^E3yoE+~x<2
z%kWh&LNDvY@OA7=p}99bUl!LNl@CAph4QU$f8RS-829FwEhbc@4-I7KhwmTk`Lq#^
zBWJ9<xHylpvT$%idf25fT^1JS)E{TG#Ujh=a(r^Gaq78WXuDv1!T4C?0(Js1na)J|
zGCMcc*UB;rUm7zXXsp2a_FOca8o}ZXiv)~a*e%Ami-j+Cp7{uK@nj6<u{fH(H=U!l
zL`Ocb`(t@&#p5>03lqXDtawZyMUiH{my_>giI>J2e%Y~(xA{glJ52z}OzNn{4vgV-
zF_uqXg!&u3x5;LTLB<&>Yr3YTKXmI;D!d57k&9W#qmIFs@eMm(IChb-6Q{kLpPq`>
zTQ4XO6>|!dX(s1cD9e#@jBQf(*g<32@K^e>6NJUgdZeSqckCYF$T<qc-27tM*xV|c
zES}8Id(wV?=T+Hz_0lv`8sp4}->mX>Yjdr9?|a`V-}%-z$|4IDl4oIF=}@`h<;9)7
z!}617FUn88cv)WU?w8Zk6EB*qtS*)9jScVIVWH{MAOCsz^7Efcj<Zbavw$Gozvg9O
z#<F#nS_deznKXJ<M5S?_0MHEW<o}#^g*|@!q&$4g_qf(Q);y9<y{LP1s7rDlACwDr
zwz1gdI;t@=WkWVEkW#FPPw}(h!0yNoKm54-;CsL5U30Uu3+0QKJLTM)y)?d+`{bOL
zH)(T=bgVMFwB(nn7Z){-R(|sNfD6@+w{IlR?#ps=_|g;mKls5fl?M+WtE|734v)&G
z(&eW=eNhgMPAxkp4pEL)S60iDM-R)xN874<i!Ni-sX5h^EG8*`MEimrY%jhPJ-d?D
z%i8*;cW?6Y>Z$5C-_2v8^7)IGKG_Am@e(#=3vK#XPvxQ8f2xs`;W90~dvRF!1&;dD
z*q-@AXlVy{L6-Fl7E$@$)z0p&ckNM+sYfi7Qg4|zV19xHVahBgzO;oVzQYZkdZG+a
zy;M0-IRp1_d%Jw}!GkhCGgWr?kG;bSKI~-N+*nb*-6%f3TX$0SkIvK{otf|4+>96T
zsl&8k%sCN`g?BznLCIn6gG^_>fO!lrm{UGg2HH!k$VZz6PxDaSp<XlhLtE$1om}z)
zd|r@;7wPsKq0IZzmFuU{;@x6Ot9NQCEP7zRusvhhLF+mqGtV>Svt*_$@=WIiW7;&*
zlzENBL43>;Hn;OmG!Cnac^ld)(!mSQEROTpoo<}%TkL7)($>I>;(!-r5q`)^+4Rc|
z;ze504&Lt9kuJB5$jgGA!lFawKmCkArz$sma%Nj)mECW1v-iuR2YOCXrk{Q8jR0Br
zLSA^1FOe18&;(Ch)fziO$sdZ)WnR`KHIK*nATuyr_>yOEnEV^9{Lm8s&X<2ZO*4ma
z16@2-4gi1BiT-=qwD|z{&W_ng#$d>^LC-HR?jFL>5BobPYg_~R`mDSQ?uz$a@=SvF
zqbF}_{&0PbI86KXM0y7d^Ydqe!BK}nxI)O@OiOD7xR^w0|FkD2Mh<WgkhKK0mr<|}
z9!v^1b{ZF^)oNIHH@l$c0=W552r3Zx%16YOD<La&a7II7T!lu8RpFf$Ou0!8&|)+#
z(b4egm&&EeDWkm^eH(hL#S~aKn7SIy`T$=OOL2GOS!IFDVRTE7(DR^#AZbS~LM9G_
z0}ZT=#`*}SWLt}+hZLp+wkWpAbX>8|>4ExzIz@a~m2=E#c=M1@7*$f!N;f=QoHoTv
z2am>?4iR{}(J2kKIx%#jy*-)<do}*Nwaf#C1p~gbG$VUD!D{Gbr*WVYi=Bs7Diw0G
zJz9f^*X(wI=WBObpFMx!GoN{zB)fi+#<$)AA}5y{c@h~uf%r#vnf{;*8kA1|;8p8E
zbV_K^m=z{<gHZ4YU86K7c<TorzJX$#fCq5L0=xV(p-g@P;lZE)lMZq6Ni<4(;{*y3
zP2D2xOiEMmj!*hR4|v@kNb?|+oHcK8LklaI@xb<g7M{o|TH@tlmVrkkqEo}*i-8%j
zr}(Q=<A<Ie&@W$>Kl?ZTx_tWSr+!1lZzC#eV9mMayqN_f&cy!s<B!TO{on`Xqo*Hw
z0hTgF%J<bN|LGT>muD}ZmE&XHY?5pmj4(JuhNY!B>5_#nzURe&hc}k;wjpU`v6XYA
z{jetu)F0_5Tkr|S(FK%e-eO|M6(3W6>od$52+<+o_z@OrNuTS8ctcNz4qdU+j5p30
zL{jE+bOmoKs?+K7`j6c?U*nBYb;ze@pqEUl19U#U!;9lFq)$3dIp$FLq6k@J_Kq)^
zyjjRXBVGVnC(TQ%8oX_+`FwHoLGDaVX8^#09*QEZ3~mD|r{ZCM<-?75w~h+>a3@tV
zP)}I!V&Q8|gNC{Lq%phX%%d(a(vmxl%Dc0h&KCWU7B^+7);uaiON(?1RSN=j+OUA*
z2-f+KMHAUEAmCQZC6^a+Av9E?c*qo;4AN&*y#u7j(L$y+r7Z~8z`Dnu_Ry`2<H6m_
zu1Zt6RWP2AXn`=|6m||Ha@PrH95R2`W5SxWnRvw4GI+6@Q1%SJ&<NgvGsxa#ahp^<
zk|yjSZqJ$otnpQoz`;r%1k<kr@pM@+?||c@xYRgC<ag-Mh6{HB=r2B%#9=Lg>|MI#
z(VkJzv>1)DO`rfNdk&qzGCBf+R(jCwDPo>HUtHd+tSt;cxKJ7rzDi(uWXDm3by<cV
z>4--{s}NLY;*oz;A2Y!TN4pHxUcJoFgu*zWMlYOirnd=HrWh&9Ae-F!*RYNwsCpBC
zs76CU-#O%i4xC0YadK<EL&8OwOMN5Gd*%23@c-O}{<jVPqyOdqL64}Kh%s=P9l+4-
zdR`N6*=1T!V>FY%^?3lJ@sq~`w-rPk5AH7hlj7<o_$g1Z8NMnU06temR0w@T#y*)x
z@-$|ZqYKyg1kmG%TXxC;_(6~n8erTFG~7hL8V_Y0o!WR0p1?$YW4Pq&5I0O$OuGjp
z@$V61EIXJO{|NuU&^<8>IMQAV#IWqS4C@*%pTO<DqX1SzS}!VqtnLcnXg?a-@r;Cx
zIl`+#qif*_J5Gme=mNms)(N~C&|mAJ(Z!r3!+6@840lB_$^-lDd;-8(-ibh%(Re5E
z0ETpv+OqhM<6F!_393Abr|A=LS}Wg7+KlqT?=IO?n%?0Thy-6v*WhkF+#SBj$ON?A
zx~!FZGi7;wt9<7>zgWKgy<aLT8xJ&&;n=Y`PvXLo7c5_D48wvHiyM!hJSyAUn|@)9
zemlEEI4+EOL!X|P<2Z4FzBJ=iz7NFW%^8#2iqvCf`tceg+{k1?b8~&QJbAcL=Gmp8
z_+Gu*FJC-=Sx!!`$~}#VS@2=;2DurZFivBv!NfjeYQ_pV3hcU^YwSQDpY-xf<y}9<
zFCKGCHlZv7<04*)W0IPa9{}!tu}*z|#xRVpSSZ3yK#vbG)&KWcolCg($pLggSXW$)
zpXj?Vp^GdY%ZP>+IN_($M^>bMd6Gpvjq{^VCd*j}%6P}~2O3{-B%v1v6sO0h3QRtL
zXQ9IUiCTJRlALcV;U+&h%9111SPVEhJ}w8G9CLmy|LiWAo#j-J1<ChTd2muL+4UrU
zc6E8-M|i%6#!HLqYis4>51*EY+gs9@#!V_s4<9@zPaZueOAGhQ{sD_$yBe?V`CCzp
zS=oKV%aANYa_SA^LQYb7{rbu~ZG8Oa>&tR-xL3~4j+FlFHq*^{V?`<5F1;?ix{+@r
zketI098TH0ml~xD0x(iRM@-U7n;v&VN_pj&Iu?caKGgNqMUMD9m)==Kb>3VkJm3C$
z>vH(fhaZ-2fAXF3gJ1g9^5DUfGPkhg;}iFf&&%cODKE~l@Nr50d13gq>JBf45-+Eq
z9F+^{cz1WNyyQ4VzMrKr_Xm$3mf!e|-zeYx{x7MoI9I;-@>Ti#`Kz*baHKrBRFRSt
z5_(p=*wyrKbEQ0a{6O|)=Yp!C)RY&jy`xa+r>>FDT!+f{)3X4{KsLWCkLA}C25tKD
z1UFY2znAP%(m0&OV|G@0?#W|q)faRr{bC2wrr73hyFt$hEqgn=<y7&|)-b2RZcY}q
zXB5H9S35qnhFx3NY7AHqryO$v&4S7`<rH8ciMa#%aMVxNf8`BNI*v^w9`Fy;4t?j7
zr)7I>Svoo{Uq0V;JF~RBR5mu2m0xeXsEMqzb4yMS>4Puzmg5ll4jAR9eb#|C$?XPn
z9Kw6SiA68wMe2N>UpS{cQvAf*IPql1Cgq8_78bO?a}k#E;1i+96UwQ=dErO-g)H7N
zw!Cb<Ec#G}*-7O&2}#IKFUugkP`@eD%pr1IC3>TcrCzsok8;5C3jHFl+g<F67k_{L
zLU9t7c8vN!T62n3=0CXn+&~8Bf4kHgkbVJW!aL+d125VV+DX#N&Q!k~t+vo{D{KSk
z!$Ba6+tb>{kq_S8D!ubeVHY2}6lu#<p}qU`?e&?OSxy&nW8n0f5%(&ab<o)dCVx!i
zk%>?IhxE6@0>g!`F3usHopCq40pNV`vl(c#!?;@<K;r1>q<S2-0}UVa7`}HJ;y;G0
zFboTfyN7_EIZdE3u0zWbfn65g2kwgZUGfa!edz+^4YmC1hU?*d)AephnfR+fUHJN=
zTH3fNVJKh!ltBu+qY46pn;`u{Rcr~7p=a;_Hr(Chlb{YyyTaWSaus4vZV*Tf4Hk1`
z*2A_IXT8HyWN^cz!_+1uO<;({aX22i2^Va2!Vg@nEVg$H1eQ!cLg0g+E)3X$LkTXd
z045EV3>xc7ZyHdzs|+nfRW($R759K84+=sy4<w>)Gr_9n(cwCO+sr9W(7MG4WvrFT
z=?~u?2+acqoWi14rG*CoZ(MC{ZkLZfdRjKNw%jQqJ-oGG^4hUV{_qfh1)bQqTPK+g
z9gR8<B)@$|M-h6NOxmM679;4Sw@x?@o3`UC4?o^8io5_aJEh`F9Le|M?b?UxSaPoR
zflMBD_``$7({F=_MQe1?Z_@|-n%2N12GU40y5S`=em%sCGEEu4-h~uEP27V}H9rjK
zyZ%f9%2pg$L}!CEC>G>aSRTOExzo;ro`HoP-hSw-vu@oDN;Bn(dgf^1Ne2VQ@~d`C
z!jP?y`wbSwPe&vT54t8iw9QPIR^}a~^2d+-c4tC<&L@18mwxk6NIGh4@&EL*Ps?Xt
zJX3xhDc>`2xn!Y7>0H);<U@6)zw@ncm#0sjXz-u`3cQt8%E_4q2wyz^vV8jaXX@Y`
zcrgq-0|&PM`^ZJf$D4U9_Rv}IuCFMg_((w(!X!L0lU8<S<?SWjvLSDylS_5{Ii`d(
zQy+Nq2>)qzBg)P$tvMezq5-D@<eX~_MrEh7eNXA~pq6|RzY~r)PzRs47g^L~R~CcT
z>A7hiU%(Ncr|j@j2lPhqq7Mdtghfs|PAn|nywN~K16+5SFXXTIcq0@#^gOdLUA8oM
z-r8LE;Ed{YujxV#*Hxt*GjFP1E0{XG>TtVLsSdOCAa;2%K&JSsXq3g-h4S>_2W5MG
zyDX{$?XvCW5M8SNi&~t86W%J1T<99NP3wJFPvS=JSyXBDy-~;?s+xyJ>V~`H5&}E|
zk_UNBM`+hWoQ+1I2vp+>UC;)C$cP**7I0vyyRj0)nm}eL|2!Y#1Il3)XQ~h3b;YAI
z2@4%{JZ*U+(53vcJb(#>YhkNlr@~E1X6w2^qmvn#c)kF+H+FI8vTmx=_$qg9ogbE@
zDF}F5X2Li_W#Ggr%U$}gPYY9hD^~a#&TR3+TQ=pdb=0M^h7GCR=4~4Mj53vJLTcXd
zrH-|<Hn5ELui!KQRW5nBe<FNqH=#$kh!tGI=Z@?Yf6GlhBVUn|XQkYhqlQTtu0VWE
zKl+;oaXQ|}NR)&Q4ER*H{TxJrYg+K<h77RsB3lhneM4ufZ1%!&2aPTI%Y;u#T0_d<
z7UdW$g=uAoLF4iJfApW5`sV>Gd{wI<eX|}$>e2uj9<%A%X{+Ru!rjFy&**V~acyqr
zKZQQdaXv#gc28=!oedC{JQ=4;xebi_4NU<1z$uUexzbq}frL{D7@-G14ANEMRy-}C
z19+c&8Rs(&4d*iMdOAuoDGcyFbhl`@jO+C?k<S44K75C8iyyyt@ax?%yS%IQt}DZ1
z;PN>FxAPEi86AVrB}ZEbaKuHKY%qzRLsUGRC&$TsRL8WheaY@&7{0e-re8D(LN_Vg
zZ5S`3UHzCKOw;|)>o)J^es{hCx8fG`vXA-&PkY<95zv5d?SCJ~xO*}T;~x*(n}Kov
z{+ZLm<oG7xdiY8i@voOlq?GR<(!>1HTKVwf@04$R`xnZp`nXHWD`i2~wEDeF-1D6;
zcBd$z-tD!#vZOHulf?9+Ih`X5J3b!$es)>0D8)-<w6m92oLIsbS<%rRt88$h1iL_(
z$Ym$d<A>|z(e`>-nwu*}$0y~p&tH_iqm%UUStzO#s@HNeK4+}R0?N|jg2#e<zv<}s
zv>a)i!FYr+&BU>fI~4zEg|)V6`}FhpUve@gpV1hB1vq~nS$-MsS#6&5_6Zp3r!dhx
zr?J_67MV1HVhqQ4g@rec<MX(|O3l+B{wzX3=WiZsEW`Jby#wyz(u-;=`m{-P^oPDI
zDpH~==UHAhWTB27OiVEQg;^1z7j|wiR%4Nn@e(pKsgG`Xkr197r^b<S?6703czAf^
zZ>XUs<{{Y4M|h9FG^SGql{b=!V<{Ocx!j;TRb9rfENX0SY?Rd%j%Z^+Yqos!!K3os
zPo8*@?d<Z#ixE4!$7TQctQ_k)IXm@&S0?ap6c;Zya%J&sVMh6Ljb08lZrm;VubxRi
z2i9iy!Xj-4^8xM%W>J&;B0nu7B-WYiE^q|W&AZ6(1))o2YI}RTeDcY+$~Qj#hR5A2
z8k?RRAC^D=(T~cX{pcs<<?exGkuKg|m069Qr!~G_S<+bj=?CST-}ppfzgxcbo$r^q
zr4{9?@{|Qwm2*BBKpI(WKt7Ipyt-fq-$D7&5C2X1@sEF8IM$Id^~wt0+FC0c8(ZaD
z-}+{G^!TZUUUMq@7iItOyd10S<#;<5N9M|c@_K!3v21TFmyHz`39>8m=;XYdUsDg2
zH03w#3gwA%P5I!PX3ux_%E94j$q|Emue-QBJbJWQK4jOJ>I!<Mj!~znPt<*0;-0U1
zW)6XIGG&kX6L5_6`99d^UwlzszIfpmvdJgrxmH#d%QrsyusnM3P~&`#9^7-eW2e~#
z3+AGmlZ;CXq(x(K>5lSCxuv{OC$peJ&E#Ync0S=(%IfM;`Tn;)EFXXHP`s|n7tddn
zS9{0hQu#m|w6ri&jt*ZHj;#Fn!;k#!{wt;>R4$mSJ2*U+Jo^f(vA^m-7J--_s>gi_
z$4N_+0qD8NS1*JM{DQH{A@bgMzAF7n%d55rH;Yb$NjazOQPpx8RvSh+tYN6ngk#|h
z9CWUil@Dzmb<^KJmyVEya4dXns2(EI!Ty0?96)c(CwdM@fSWdo7oB-2AKv&=E^@pr
zJ4Caiwk>v1=13EDi5df7C%-(GLJ1RJbeT2}eN6jej)?StHV<ejopqwiwU7c>ff6f)
zD4**>zGQb;$w3)lrzFQ)hUrC_DHr9w+MT>`4zK=s-uQy+?E{lPCcI>3p{ef^^yVN;
z1!%L_(dcCtE;<F*(~g9r&Oj>w-+cNwDlb9g2{c*Y@A*69hW-OPSI19g?`4SJ7_!2U
z4>0Z?0)FN+fyTHFEsMt0n9<o+!(H*dOP(RTFI@m_-Y$z@2i`Yb?}pKG^Rs{#zJBlb
zS@?1hsYbN}<8Ff=6E^XW3|Mqm#&whmmj$U7J~o0f(UeREW2~UM#ju~Q<etXPP8x!W
z$MC>izsh+;At5KSxUo3fsKB^nxN3A^WFoU{P*`?*$<@))IO1EwJFGngqHu$S9sLL_
zw-EeU2&rjK`hp|1p#AoH<%ihd-6;``^(uRNfGQf{(2Lm!5=Y__Z4DC$&t`OVvb5AD
zI~*gP)f>6!pi4#}cwl_^(Z}WC!v}uQ!3X)=pw}=?DmdbYRc*li)&?x)_Ct{en;L04
zJNUakR9;xwfkQ|Az8>Io9CCafZ#krKrULPEr20pW{tB=9HkTgMJ9~TO)yo%t6ORWg
z4`6heg&}m>${Vp;19G8X_{h&?8H)V)BZu=l9u){rI49kF0EfXYp{%F4pzQ9whi()V
z`{<s3rL2w$aid>?k>^|?uzuAkLstwaEU*0YK!&Cc8embGMt{vrhrneVq)ozOP1?N(
z1}DlSZj}`lPr&gICXNJ8z5~g7dQ`|m%*dDrnBS(+1AI@Nl9#Vuc=r<vLmclw7|Pw&
z=0<t+@NwDR+$#Q<vBJ^8AO(a!zrHBXl-HmB^s};iaNv)JA_ScsCXVOk)j^u&%{}6i
zd{P4C3|Xm!*+x!z6H4-tRvyDCn+!Qfn+9ic!;`@iZ-sHpfZrTbII_?m!d2&#bK39U
z%SWr*Hhx&cn?Vyh9bJB<65jsW_W<I|gEo!=n4424a{;>saO!k1Fuu}&fleL#*xf~>
z)DvXV-?H-dqX$msl2IMGZs7~rl>SW(p0^)tm3eh|mCa?!$0?HEyuBr|H`1$g&f9Mc
zyy>LT!D1ns<14)o>V%NT(vSG_-Lkpqx$<!HVR^Xq$ZsAg^~&_*G32lBDWNWZa+7xm
zETAm7-jN3UDqs8Q*E&Pcfq^QE6`>x1ipJZ6Ve#Q-q>Xqse>kN5gCcU6#+Yiz8Cflb
z(E!xVW~)e=+y?CCC9n=Evs(ULSO)3vA`CS41B+Yty0{j0n319j$<20IlCbI#yq#fD
z@i4L$W%m5=ObT1zxGj&v2M4pzCq-R+aSh_Cyk&P@xUmQDtak9tUikzQIpS0KVq+(j
z%4gF+M?;zfamx-L^S%Q@UXALga(DVw_;9G{#NT%5&bq)!M5343FQ6%0<pVtq^9eF_
zz;*{UN5a%t1Tm>M?L`-d;n_!cg6GylwTGX0@tlLUv9t#bL}D2RHJGl+Q1xp0amlR!
z#s*aZE)5~O<A@$iH}D=(qB6U%<Zs?iuuSW=;}cywth{TG^80`IU*E!41D=CJ=;|iQ
zs_q8Z`MIw*hMxot&+~!jBW4F<Jfcwa6r9cKJ)|26Q(=fhp>49g;fLvI(A?wZIPh$a
zY`y=e&fs0Gf}YP<m;uqGxWMsiFezUq!MpiVSE38R{u9_t;vF2?OcCX|HVTyS(QtmM
z4NzTt)*qlSW9b4W#Sh$t2k`T6Kr@NH$6?YoM4WNjTW+pf{BEZiFb=dB=hugA6+Gqg
zj<5pz^Bdzo<OK}v<}GK7V?2+B?%Vn2paXu*9L+-jnn`u2L3oe)jnLXUnzkl00GhHB
z+}(V}<G-cPTF!3kxoY>0`0;DzcpO8(Z&Drx?h4!UwiNH+-vcyv%kot`?BAX@HGcYs
z?qAEvgyZtcn)=XB$^%_%8xPCM#-lRLi(XUHg+(qFssQ@zj0-#wtA4ab9_rK26plOO
z*fLIXppQ@A+7r2LETxDvMDawoy5{Be)n#5}ySWzKblF&2E{`@>HD)*}|KeZ$>vDE}
zRn|2|=X*(IYOe6A+U3>ja>Zhs;)N%>LHJG>{gm^It8#X6Q7$|=oXKyFP+ZnHg9QTe
ziHUv2%Z#BZzuxswW#-5}cBafsYs@48J^6ZhRgMmh%IWD@nVDNCOY58E-pqVCIlU~0
zN5^IF@TeRfa%A4I7l|_d*H}V&;wVOc??^7jWY>OyFE2ErJC0D}h%!%P>mrXNk}jFb
zD|SOMX-{1u=ji)nf?jfpx4*CKlVkWc7`xdtmJtu0Pu@YHA0>?ZWsJsx*V590ci=G2
zWPy;=FyIY6#h9Zh8NYEfBK<+WbIh2KiEzfhIdy|?dvV$f3oBC^e@2G2l|_F8>>D3F
zDG#^TT&7suI8i=cTql266k#!CUphNc-t*lm?;=tfXO;M+h1s&QqH*V&i*mI4MLF1c
zUQVUgoxOvyUnlliC{s;voKGG<_IIT&7}H6wOxz!^sHbtKkF)eZP<byuzEwt7o$($E
zQXhZxary3dzEd{VIf8MeEVJ0FvOvE6@YB!BXJ5Rqj58{<%k%fk+<hrXaqsW#c!%M`
zhmXpy{n}qEAAR@@Dp;A;_?9xx5t%3Qr)+)r^l5qZ@>O~E<(K8zvoFhE{NyJ^S}QXP
zt18=D<)e?jS)P3Gp^r{v%&t+m$|rL?H`3i3>FwGtJF{Rhr*gNpLjEn3)#WVs?n~BB
zzkJ~%{U{&ICvXht(%gNQdybZ50cZE{q`cUrte&~No}Qg}A(e01b5hFk{B&8Aym_Hk
z`Nf!Ab&N${UT&QxZx)yQNdfqi=afs1!emb4i!YveXDH>l?JnZE!Q$`s)^_>egQpso
z^Bq5SBWfHkx)WXOGGiwX^^}Ebj!w+NR~DBDj~*E-!xx#DFCd?Jf%cmpJ}lq;#s@x4
zW=G}Y^DlRl)>GFl@_K%DO1hdaPakcShYz+@YO+vA3&>Hd&tB}6-MvHa&f|q;<^U|u
z(vla3Sg2-^I(0KIWFup%W0WJ;W$FXvPZxBqf0R3gqb+j#OFJl&x(M)+Fy#h%F4EW*
za@(%bq@8w}`6d?7)pn?U-Y8D$0d-(&bE~YZt}3rieVioU-Gd+HBa2^Z_oOq|cfzS1
z;%P@*@aMt~q;Jdg26G~!ReZxfI{v)ukakAtqR|j|zD)8#OBr%Hm^*@*FOn)+{*@o5
zFWIdFV{-~m7QbdZU*ngQ<?n@im6Mr!H)Xb5s6Gp?c(XXf6+T1~Gk&%LG93?BMFB(F
zxQD!Ajx0tg&YoU&_~CEsfNo%Kb3FifY@@W5m!QSf<gk7eXS7!1YOY2*umiUDPDA|1
zkQIhufpPZ`@H3|gG{$vkS@aykjLyCq?uz$a@(kg9=>q7GC*;=+*Teg!>)p`u^k;`|
z;md_ZS3hw>1tY`a<6<MUCxPVeXt{yIZj10D^x&;}J0pdx)fG<z27`Z^ln&Wwbam+x
z%D!e(Iam4IhS14@*#0pMrnSZYgvgEW>;N&6Myv#<x&gRb)v<Cg?GMPuINgM+>}07S
z@Z>MQG*+49545z^Kuu;F3E%iRwVfRclOIe>XQ3}|2s(sWbQ47h0NVOOFW}sKn@|@}
zX##RMd?Vl@waEWQ9l(|H!IQ`8czjeAm$U8r+7F`SwdEwOxbm-V3^njPJi<;70d^v>
zpq7UOspeso_K|L{Ko0>r4|%W(<A)o%<E@bhjkwss^JqrbJS3Tb2jGp4c#s|*9lArs
zx1%@%+>@q$&}Oj6LRdlxgh);(UOdI82O}yfCbT%Ce*7C48=K;D(CQwA@T`A1<yPT#
z2{tbL$?7bOVmGDnJWTB(E#!wtNSC)wEB&O6<0S~|ybvB;nG-lnDwAleW1R4WAMRS-
z5;MFVtMw~~ap`C!PC6jS;*Yq>+YcepxsxFGjXDT)rr-^pBP-zJjtm0`ndBKCvOVUA
z2X)eUBae;&=X9@auIoXsj;zAN4}(P?%7_Ac;p^q@tMa)z&Id<F<w|l8pWg(#$KXSq
zR?ZFA;0~pt0eFypc7U<qN++o4kut_o9_e*QO0qc(>ip3qgCECY%*@VPj--j>HnJt3
zbW_TlcaoPk0O{m$tO8r^>C8G`R7cKFPQ5dcV>j4g$y;SC)X@o^;iyA8W)h5rEWb^4
zl{~f{^yAnFY#g<Kj0}9vHR!z7zyR3`YmtJfnR{hpYqe}W+Em@1FZa}Y^#YwcUjpe7
ze*W$lM*=8crzHy!Q66%9qwFY2>2h77T<Mybnkkzrd}Q@eSyz6s=ttWGXX-OM5nJm_
z94<$;!<}-+Ww!MOP!2;MDr}*vD4R%)waM1`1C((pQHQ_|G&j)v69;CEshR-1ZE+b%
zBo!RIhLfHui-fCmF89K_4hLno<1oq&5+ir?gcDY|Ro5Uab{qC595-FWL45jIANB-~
zmt(TskjPEnaFG+-fJ|5TWuhQl2fVTAZ>u38Y!0jzO%6+8yljrU5}1%Xb^&De3$Dd$
zI_o@E1Wq4CX2rKOv?Ox$Dc{P&j)%$O2;k;Qb?U(kcos$OrbCm630O07M(FY%oGA<)
zO{%SZtY?1kFajP;AMrP>+j==%wD{4M6F_cgAcCTnJ^NH1aR^KkG<NEW-Qp8{B~9@Q
zzC$#ol}Bf{ziu^u>=c-{<<r&0V<W+A$6pqLH!|#a`6e()JZMLtBWr&DKmGX~b@)I2
zU;RhQ?k3PE+%=io=!fB4W^|3?-UpzAa)e%1yIIi0InLM4;BNVi0XX_h?hj4yhtDJ)
zck!yW&TMV{IMB@96`$_|zt~4<$HTUlDH=@5pYbra-Qkh?JDHvu_-;Nao|9?X6_`Xb
z@b%SXYH(M%1IGC^7~;q6!x({49*N&661r8qJM(k`ejFF~7*4-&p5y_vEp8jPXyVo5
z6J`jwM=>>=^w`qVAi56qI3M6P&XP}z^DgN9Fk`=qegrS8+d(ejn-9-}V8`eML-(C=
zyhl9&f*aS_`-Xl0G7Qfl{p~!89Jk?JJX^RSjJokRMp1Y5U(ZI$r59R+V>0=k`nof7
zi{AOQvHeu$gR-%??L{#bT1Xlb=}c~3-tbKvzGMB?J5e|;Zf%WYshE^k%9KvV4IH7x
zg!PR~<_;L=tcqr3VbP0IjGO&^8jU{ImuAc1-cI=k|M!1TKKtqC<>`kXm5)F9c6spl
zBQJn4>3V+kx*W4$cExTcjs(k#ed&w$^%psr7@nWcqMW=vR`le%#sQ3byfZ=`EPQ#v
z&Wm?UTvHz3`g_C2hsPSbDV>UoMXyT@JJ{jE_pH3gqA`!Zv7_+|JHMDP_AV&_FEIj~
zkinROmj}_i7n-DR7LS}a%&EBCV3(<jMK3RYvEWq4K}f~8F%!$2!okak934OBg%|Rr
zjngtt&Mq!?mR($&q5!{W`S{7>va+&bdM2=$+-EW8^i=7%zVb;J?55+p;}jQ9KFh))
zA)j<OK0eA3_9w@++@$sr?%~#2`Nq?S<-?~>%KG}c(mU;iFTNdwY|#2UTPJ7b@QCl2
zan#{kjZ;}DoG#0Y(`9>op*&bs`d(j_!;_=(+2>!DpZ@fD`SSTo<vlyq-gvQt&jWnp
zqmMnF_AV9W?epg^%F9=;%IW#37aA$u)8w(n!YpQ@ckjYcSy+`09z1whHaE9aP8Kz8
zl&n|R<xFMk=FL5gwOL%9FWYNOx+@Rg-gp7)v!8zEmyS86@6m$?%KJ-y(`tTU)(bb&
zvOoRcgYttP{9@VJ*(<xdhd#29BRppp*USC+73Cdef^TTEKs>AQG2i84k(T2uvwM#h
zq$me`6K_U1pYC#BX;K={;q#q?@=E29JZ4dV?;&lis}3$^$CJOir8K?RKPu054wYxe
z8jl})SK9nyj=!YLo*wSIZ1A$HB6q#=Vl4|`iksyV$_vL_vdGOMDBo{8P&gKS&sA<N
zF6-hC3zL$MaWQ3Zj@>$HKk|LRCAAIAA-pX|sza0?AB(uO=rWD&{HX)xHUxb4%5xDa
zceEeuZlf(>;f0-=<jXgoJ}M8kHoYMB<*WVj?B#wrJU;c}*?pCh&GqH-?N6ST4<2nh
zUpNMode4iq`v+&``OBTMr#eo1;Pw#SDksdzycHd71@dMtLFL%()s1-c^C2t*BOe!S
zna4TQ7|Ngcaq+!*OzI2uo0pcMp-#{q@FFgN?p=3<^DZ>`Q;I3a0CR%0HEjWmmx&)g
zer)-ulSitX=#X%fE9Rh(i8}7ZFX@k_hB(l9TS)g<t;$MNq}gp0<<QSz#OCFJ%+YuO
zu|CU?8NE1ds%uWK+d10Z`m(tR`Kz6VPT=@J%O48AG=SW+1*!|w0}(5&l!w<|>|9gX
zUY>bfW>|<OkmAlLDtySJnBs{a)(VcNUJZJ8=o@ZmzSqJRU}x-@BX)=39_allAHhgG
z16kB`T2Hk93D@*qc?|4;?Y+|w|1o5RVOU_?Jp}yBX#$OL9a<Iym-+u+_<9$lT>Moa
z3txZmM^bkh3fJMm%!Bg*m+?t@4}?fzgWTCGB6nonITVG%!6?+Ow))!>z*#GOkA~6=
zZ4<-zKw#hcgSnv5HanaUp>HJOXQ(*9VpX=D+`(p6v73^A=NBInF~G4sE+B5Xhy&&l
z+nq`pt$`4!pP@@=F4;T&fKq`Dg-$wpolx0HU&3=u0CLpiH64VGVQ5JTj|$cL76FBC
z<$yOio~q;bL>-g0wRJcC@N5ly#1KYiCc0#IJA<7+`4NA|FX4Frcya<Haw<+9EJW>+
zpugI4+etSaP9DBA^0>`Q;B*j!Oi=!?sPv5-ltV9g=pn%337zg;b@28N_Vqx!QO78Y
z8<b(P2RYm1pJfQrW>n#u>GDT@1TaQE4UjK5{8Iiy*ZeR|uN&TN@Bs-nhp$)gGP?+(
zllVZRJ38g(&Vl^Z!0;eO9}Zf>!mpJ{rxPojpz*EOMe1=EP`VerE@Eea!y$#|;f)``
zIiA9uIOLF>g&OB8-twn&<u{z<xKSr09X`dSvd|VWDHfcqe!~LEjRi9e9Uk6_^FW4<
zkIXmMUQ)8(Te%5;p$^XO;a>Uj#dCj@SFxoWs53+7k~hT`)hVBurGpfCNiQ8O50KR1
z_kcq<!g1^e9XkZ!tZnp)R_M%8Na^Ua<BPHhFgWwTROx2%ibb@08W15{bgfuL!&^@*
zO!0;o`eag@a!Lo60Wvc1RwZv8vha0(F6Ey*VwV?h7p8+LT?ogUd35mTOh=Czh{(-b
zXcwm%=%~!QUP=Ud=Bpb^W&6Rp2I-4sM)ANJJutBkPddvSFL|zk6LpBekl&cp;M)su
z<f%G!XoEZvx$}c`uPyQ&g2(DaZ<HAg<fKvSj9ozY7_<sfEy#OZcDR~Y2A4hTz9l`<
z9%Y0pIbcfp$yRAxFYJJ)KvvV(RN@I+y^V*a+PJ67Y-ey5F4XW3HX(dR=ISzZ^f3}d
zH%J|^wPn+!a+#m)+_;PZVf~mz^EX48{!?FY;wg(0i5V1r_yC@07Xor4?FT*XiZCQZ
zjnypc&2AtRfcO!gFlkHRD7Q&1omk6kGUKkMSsJQEB({m4sVuxOoy-ZB<jLQb8kR8F
zMN9ei;+O*oV1sFI^v*w003plW(mP%|g_fnJ58lC78k~5OJ{;f#+#7dekO`xvrK7Jj
zx#Q1`bPRm_4_{!&+mS=;e`w-g-C~I8yJ(HC1eJpWw6KkHzsMy_;~|emWlLjbe8;U1
zhqdL89UyB!g$pWo>t(=&1xeG&sO$uBgr=7m!G8DW_>C_XzN*#vI0Qc{?|??z>VjQ*
z=2)>6e;3?}XW+(mJFME{in8C(PW#u{Z^H<0*;BWx-G!&o<ZAT8aKN}57@zEJ!1Jqz
zJ}VpjyLiXmt{%q2j4y%#!~B{o@3{XcP4joX9l$uRui_oPy)GL}a{KvJ<(R~`!ChtX
zt}sJ<?ClD4eN#hZZ8<fG=V+exFr;;OxyLD?xs6R8wpPMz(yc6|-1p%o$raovKJg7p
zUcj{a6X&F|dz(LW;CkA|dlK9Z54>MocI%HGn%oH$9ff{Em@#O$q2Hvkga2s#p9BLQ
zm1bP0fyweU!0)yUa@UIHu#<xhpr@TIL*>&!v-dd5VVZUUjayI0;)L-`R%$#szpzv`
zx3<);{-A6<`mn5TKJe}o`tMU3yD%Zin1cQ_$Ca@ggoP#+d_10!ACuJwC+Fo*{dSIS
zV*++o<A(=Z8|C5VMp=S>iXA!EWm#jGn~T%(&;Qx~r~I>j{?9$$SzcW)Pe1yG=(iPz
z;+>f*ucuh_Vc|f1AmmqkJbT$Svbeb59Z)ywoL{Q{&dYO*=e(Osa`HU(G3{5^p2Vep
zp9x3Cgt}K<W}Ka)&bXM2zgBuKHP*S%SjFEP(zwXq^VZmjv@$;7;)DyvXDkvi_VWv_
zOadzca?x=hw&ZH##5_y&W6UXk!f}*6qT$Z`gy_WsoOx<|#|w(Ha|>l|VbKe03|&1j
zE%_NQbBrP<aj<}bjErqqOjur7&BVQUGIqPVyi$A`4~m9`!i9zG2B9*tIKl$sgY9iE
zobkP@{r$c2;>9Z;O?g2%@L2tgMn-ST{LFpReem>Y+1`FA8Rwk7HqoB(*gfefJLfp6
z5|xlg%L{X5Yi+(f-k2+^3%r1MQ}`Cs-r+_0^5uTn=R4Qp<y~*mX~v_|egT=6gwM{-
z%duqgG5kss<41OqZEvoX2fCDee^(2+DdUu{+4=0&;yZC13rM+OAw(s@iwE8jC>{$k
z>6=q7*x^UR%_8pQ<(bC7S7m8&&hk8Z{K(&+V!`Iwb3S$OOuTQ(>G_q9L&PMHJa?e7
z$b1lEYj&OSGAuie7UrkR`s#dHnpJt_duS?8EIuEeU6q}~6Yp5!+h3FgzE6gXD+;r$
zarLxhWMOD+bF0iPt(Ak*3m*f>QHOlW;H}CKi=`}9@#&0{qeJIC3oTh}QU2621pj%B
zkvYPWvAYyn_ILNn&d!eVPjX3q<_^42OPN(ZurNm3fFEU)?}abwnp0V2@oQmzQS@&L
zM?NakTpp<l@MB@xr^qN`PKjZs8UE}}J(JF9Kah>|Z7eU82OBG8ZF$M{n0DsNm%Cn=
z=2RWzSWsEs+E^<eJbCEdZG>Z?oVk}9wGW4)eYtl~PEJ(MS#(oclw|853uG+*(8f^~
zIC}7nU+zUd7Us|oyTJScvEuX5hcZ2`mz>mHE^veaC~KbAlR2d_NIhd0W{Xn=kH6bf
z%22+S#nGAB-H9*lE^)D#&ykWWO5R9*c72_xUQw=CJYum3eX)~|op!XxSxnD!5Pm*p
zQTaq-UcaVI%bW%?2yH%#x{bWbC_g9AQ%hfH+<IBA27W<U=^-9p!ZjTL2@|08BVFi?
zb^@Kyo&&r%?w1K9$JXj%*;-b+r)LHsx&<QJ!#gH%rmYOLwllIAGcb{*4>#(c=Y!ld
zu9%K9{3C<#{Kvu#O@6(-_X8R~po^=f%ls8@^c4BsC<{9N13O@Q?=-}J3@HFV4BP7=
zFtj%?r%8Bl9Xk3MAnt@oyoT^q{te!jE|Bs@h_4r}hgKf`YA~w*p9NU>y8pNS)_?dr
z6fG7!kjn|vFC7f*Dm>o0(F&P;<p&;n|I6P+ueyh3AUr8BF`J+54k|YtT{QwATOvIX
zW5UpPvEUXhe#n&xK7KvKPe03ROu}_nOIxid4ev-&u;r6*4qgj@u8>e*`$A#7<)@*>
zg|4a}dgI~;UtO}qElS%#*`>R~S3C6XZK@)NSQq?+GLP7~A{h`)cFE%A0p-`Q2Fsd@
z{KHCxrenCZwdt+b8=IT%fWe8-BivzyUC53#oFa(b_$3O`p~hVgMjnnd^fcToFeJ;-
z31Q?<0O8ZwX7U_)tw;#XI0N%MMD$#-+z7$LhfZlaOuPxi8+Yg4WkUxg<(v2=CoG7W
zd_@S`<L~lZ<MN%jZAa2ru|+mH;5AY{;u}BNWN;xb*bXcMCW*i#4YF`G!U$S{2dJYG
zWZ}dQR2q5Vm@MEy+W=m-E_IlApj$uE6;~Tb7;iIrq$5Jdh@DF5xG{ja_5gsz9*%tA
z?I16{)Q7zFPgr$ye1^IP3_c1%^n4(cj+D2E*Fgpsofh6sByp*e(UZD3)*Hv}AvYZr
z;!IjZUgINvwyz(u+e`L03Ui}4<m-+moo5CN3~0zp$~13{wU0uB=dDIMdvw%T^r9S4
z?pWBPqeABn{Swv#3eiwT$tdb4W!mRA!%H182AnRZ`t#9a7Ve;B@EN(eZY(bY5cEYc
zq|?cVv;9_^WCjOq25B0E-j|#j6w?V~FyK;1oRWoNgbvXY`NNJfL_h{Ejy9z8&%(AB
zvd}Vk$%<ZDCylU_3y#ffX>+|H4$7Elh_@}66OO^m(t<kRbBoet(_bRyPMuP0u?Ha<
znQ$VJcIzfT4BI<XS4eGl+i%Yv9j9xqfqI>2OlfE)Jn(I7V~L~Dw@}tghXv8okU3R!
zdxVW{HGMY0TNU=Kv>|Ks#$i@Fcc)w7s;}HBYoRl$0*)vRy*6p{Y>-ctg~vHyEDXA<
zVf9lTunhHTG(8SyaH*%VN-WxL`GyZNSGP%Yvs}%NsWKIjsXM`^Z8nQ)#vN3Uc4Im%
zN@K@2ZvIaFuzVnoEy4L2&F(LP7;jO;M|Z;2cwz^iutCBh(E1hVyOj^~@zv}JXB=kZ
zAthn$-Oymi)n7p8=RbN%U3GboY2A0>b03zQRt_3XleYDlkW+?&*k*@AR{#D7N;L88
zO<nP|%4~0O!ox;oEon_t$qg+uq)C@95@iKeK3Mt5{m1|1|NHM&)W3E3m4D~I_dC_v
z$=1N`PP*GLr04Eyh^zJ?4tSTl@zBlgK($j|n_b{==&|N^#{>MBJy$K0%A?pDO|ICn
z<e7k-=Rm^`ue)%<-ojTFvNyB$8@h+~Tyc-%$K81f+)dNqF8@hxrM__;!CTh@cW|R&
zU5CbS?{Wuvzv1;xKSeVMpZH^UXybab$Hi@GEV?XtwYRJJ4=;ngdz{18c+=K5y6|aq
ziE%Wn!_}}uT%!ese(@jY5g6jQIu6whTiL3A=;PgRjaTy@)|H-Scum&Pq;yWgRk!OU
z5vJYmXzgEVj0dof<8HfwhBLk@V@NafA70hh*C@U+=>YsFpN6+z9P#rrweqNR!*G4r
zaj3jHnsFJ%VKSeDAJ$=T4Ns3@fL9!EH5Ph(eO}Iv_shZlZaF$SC}*b{gIt{HnRip>
zH4b3HoyCBqMf&vY4$4vDUPMt}n(;Q1#q{A>U||Oh-$UXULyo@Vh%vs~bFMK0iy$0h
z$Rftp#+G;UJbwI0<AfO>jmJXE%V%Gd-B&METHciPwI#p2cEz`jHP&G4Mk@G@Fk^t5
z*Y~`0hn+Eu&%F4>*hpg##vzPPcrlN7+3Cc$hELBf%fZp9kB-~jKQ6C!_RI4ZJLSdB
zPTAW(@;Ahw<=a1u^U!q0SGvTTF;?nK$}@2~-#yMiI_Z||(}RJq9?$wAQ7_rskG>GP
zW5H==Cf`+Nk{W%m=$LN+u{bqfW|Urbv$RQZWMkau1ucyo{o*EZN(Q21lHNPKIBrmR
z!C2PgBahu~%H{c)^7`1PV;mhFmSbHkj+~vHh~~BOo9~6Kcw$|GWGtu@`Pr%Q5er9$
z$0yR?p0C~g122?t<p2KuVL6o^S(GAPcA+qV&yjwV9d@TN*5x`=p76~niCZo-w#LqP
z*8FlN>6H#xP(+u~vgqFW`%x@jo$9){WWpTTve2`*xa<YLr6sNvGOuu)A>SioN7j+@
z|MW~@RVG+G;HbM3ne0wGI6Cupzjj4?cBy=uo-gx@t7Tz%wJa{LmF2aKvLL?T?#qww
zNwGWe^0mg<Z)VEz`E@xug+B`q<a_djF)te=x#l$PouAcMb6$08=1p0ceqE;DUY4oX
z=VhKFIH&J<5raIC)w^FFY%KX`!_(t~^5V<S%kjZJIiWH#TVAvHtGM`P7xj<vKjUx8
z2|7PwPC;qpSUJj?7ZN1{W8568Ij4HIQXW3oEDyHU%lhiF>pk-nlrhsf4>DiC+yY~F
zmuBDm7vwp?SfBVLjKf09@siXV_<2r+xd4Yy0#%1RCn32pQ>RqV-1cyk<5F4S_&bmH
zvp~Y46zvN05E-lIJA<vfu|SLd{Zg|c>e2v@+72QBb9(sq^G+-hF>AxxU*>k0Q}WOB
zsqe^a7V?8n7LCFOpnpIZ^B^x3SY+{2TjIP>TS`8$h`+eNXIS{Y8H-<hh9aMgSX}12
zeJmo*sa>1$;v@@2#CN5-LR$QtK`!)1+n_XMuBXN*S&$onsk`VH*=aAbFqU=<3PM22
zLY;BC{6e)Zw{y_)gNOj8uNG45DC5T3(%c|EHLN__+zck;s<cnVb36-t+Cvp4EF+(<
z8eh!F(uvUA_2+*_U)yHmnmvAPE&%B1F~hHO^B?gaHT|<0=>2M$6STNmx@<_d(G}sL
z?KRi&7}_yo9fA1Y_5k0)HW<3c{Qz^Cga_B58{gY><1mTW5WdR4!TZvU=#cBZ!;Qkv
z8g3MB>;LP*|KlJ0-~R6X-}<e;tpPpEd`G%4cID)e0N_Jpy}t{c^npPK(GwlHnG%R8
zBDZuP!lk{*ZoBT?L`wpge;X7!0(r!%NfCd?4Mswmi^Je&VsWiB=4VRNW3L_vV*@ZA
zv&l20t@SM=+-!lOIxB4hqgx|4xbSkE5LobBw#Gq|2!vg9<!6MEMp<s`an_J_fULrr
zP<}Bw@PZWXM9>W5m<*(G=Ha%wvRocNdR)H+w16jb>Ef!JS>y+MV;v7n;<le6Z}fl-
zCB_3@(kz4<ZCU)h1S;>=khjTp!a7ee8{aA2W%!}WLsz{79*$m6aO2N{$7?Uvu&8pT
z&LVHYkPaU7@OSzhC#GVqyq!nZCuvN)rV{_W#gCo?;SmbL@RLOsHq6-kuXNTIvc_V3
zj>fdF4c|$Z(-)kXaN#9-ztumu+G0xAiNGzpI}`Xfo(`Wyucj{^;&gJ@9feyr%Wxz8
z<xLV6syJs|Zgq@Uz@qc-#VvJ~@Mn<)p0DXR)s8xY3KqRuSgtH~VQ0ZjmjxoNZ(IjZ
zQ(Ndt+SLKFy{1F<2BeBRestP6!<pkn*tz80S@0triq9Rq+CiY4&#05{1t4{Zk~6hT
z>gw`Tz;=J$yh1;eVcss{D1s~kMK*wrlaJJ(^GPR{_|)-$6AQf1ljkhNF#v7HbFdT4
zTk_>b78c_c)R~**%{e+Qq!}5iuJUnI25`I`srsTY{N!alpdT+vYCuZ|&YepG3wz?f
zFrzEG)u>O%#UdVLESMo<^0Vq0eTjySJ9=rI9_Ub$xl1*((kUNw=FPL|sl4^)z*ePi
z*1fSEF8blt>PPN^iVH>KRp}FJbxTJzQ1b<BM+W#df1BX!Gz~Splhtg)+wG;G8m2W=
znm*1Nwz+G7&L5MT;;1O|jvf5Ku8VtQirwL{`ELw}<`E};iZGJnuqDZtStX22@U<*F
zbK-|5?A#V%f;C;mSDF|>^m!B0JNs~%&IFwwc2B}r9XPBnXpJzwlLa%B_Om}>J3oPW
zB~o}*q|@xm0Lpyh*QKe`M%9O)QS@xEG&=Cuo3^+h3H`8~ggf`Rxm91&+s_Dy<CAn?
z4_~awr(AtUUncyTpKdKpglznSZ_wfcRNrm({-M?d5#=XU+i^O{0?R_)K?xqDZ#BaK
z;&wzbGcME>oOv68S?w-lvA1x<`POBm#eo@`zSB_3AODk|-@@0g{>Fd5Tli|=^fW*`
z<ZQB57@?^!#Mw?69YEvL;9YL+ox9;Wzd<<47S<Y7hZl6=<I06wm&uK4{#c=}CCK>r
zH9kGB_rnd%*mK2guYq50ANm7B8sIKB7k!BiH$i@`A+EtKKg>?nTyP$`o6<*d0j$BL
z!nvPo`yh^bg$q>>no<1F-L3?<jn9Q&!Zov94SF0d%FigxyWH=@4P??SxMpwoipy2~
zM{t!+X{=ti{Omu#HyHW{e>;ELmB+{Au;cF5cNpRVn7M|2HP#Uvm$9`R^yL9uqe=Nl
zc!7fCnVPFgVOqD}#?{!Wo2$WB;VeVtSADA)7}B<96fnf+svZuXYv^y>@PQ|py%5s*
z$-u;ETdfCD#|WQJF#_D<*BHLHwzwK#pL`AD<MLQRK)?C==CYg~AC`l?S7mqSdD-82
zRrX~+z}(--*hFK{Lobw^loNKuXq+P^?prfq%-Dg8<KdXxC6OLu=;B*XXV<UG$%V#Q
zW!k$-9zA}j>v7rKepHq>9+tU<WsfCzsc?N|(YrX9tY=)v_=LqEj@5guhxF;i>#}=z
zR(1~98FS`|Z>1`W3$ypV%Z#J47~`?{!;T|%`}q4c>hF7Ot#J%5<Xv5!d4Y@<_}Dqc
zLJgDaS69MaX$)xlr56JjKQcz)1=ozjV|v`5OjWYwmfd3;F5$qVi=pl_;XhNRXXiC$
zS}JpMi$3mg#*02IcFmW$dGNB&YOKdZJ!4eKok{teC;>mlXz*jvjPVtd;Ed-yPMjAX
zjmvOTE4*Vz<G9!8TH`iK2L+HZ8sps@>&C9Ixw5#tR2G&bllUryo(LzcEG97Tf*#pr
z!U7Asu=Wp*R9^Vj)JZu#)i_%EVkZha>O7{ty!K)jdA56STK104RhBfCyg%bf{Dt}Z
zJ`QtriOKJ&vNShU7FCqib$htJQl36};DrLxh@L6enR8Lc&HXYK6E&7o`C^yM^3uF_
zDDr|R<7M=ni!pRt1X8(?>7(;CmSoY$3yv}wYZC`MU|4Km;g&K)+HZ8RyX)}esvI4&
zlj+9e+EvB*@BzEeHp)YGJ#KH5jg8f^wmheDsj@lywoJbf{n>swIov6SyDya&7xG_J
zK5EQ<bynu@N&jyyO1a!G$9vDq7eD>9{N-Q#W%=^ii*j#zp)9U#m-E-tW%r0hL>4$z
zW~*#0G#{Ph+uO&gtJELTLYeg3iSmqc!%N&t%G=e|W$(h{DAIWez_-I#Xu+Rw3oLr0
zJIYqQd_<=7!j3P--n6%#qtNe0{Ml7RJLsbw{c^7IiTD+YddJwl)dSaC>HWTFnC~Gk
zz4)Q9tCEjz%dr54n>hf=DQRR%g*^0vyt1CMOZoDSve#;JT-Gjg>W%9jFFmRJOJMw&
zE8_DOnM-;t7H@qt>#XV#^Eb>5vD3>hJwu}iEi-i*xi#Rx&GXg*s66D^m-46>9S0Y6
zl^2jRpTT0N(x`gLZnu?{Wu@!B+aeZ+i8IH|PPyzZvH(avseCLf^6j;^%4^9aUajqM
zn?o4qF?oWlk<SY&mDc56`G*~yAQxqVXD#8J62kEe)6e`VD~iK)Uvis9{!TM@`QHUY
zKls%sDxQVLv{Nj$;YE#R2cH+d<N)V1n3~58X=+!}=BuNL%s5)yEv`{>++U{C)ZEae
zeX*PWUEwNyv)fhufN}q%N%_>!9c?cgFk<JbZkxSV2l2n{0e;8}7`n&(0CSpz2iKt+
z-`jNKFp1X?zRJJByXk7->L6}Q_@4a7;b#pu3b*zDb%BMidw=ks{C$@hN)DD<l)%)q
zsuC+AF!s^SooFF&yR(p~bW^;2LnU(67p+6s*Bl#@3g*LnHN2`{=zLdrOkH-u8yR5$
zKP2S}?~FNQQ^UjC8i3PcJP(Ob0}gF3BJbS^CLFq;A+sFe|JDt2ODFNjKMsLtM2iy~
zRUC$92`}&p%V!uYMIsMh-TVeVEOq5m4OrxM9FarLDzi*@Lu@}GbZZk4Ra(&|e(cBu
z#E0-npVeg_{Aj$fsZm$QWqEPgAGmz-<Z)SDTQ!cO1oE&^KnKo40UXyGCy-d2P=g*%
zp~IAn)`ta&jO2|M+K|Z)4s*8>=`2j^A>qz0;pEBBaf*a=;6-FCNbq(Kc7I2RIQ5XG
zGf4-Wi#Md6J$qJOzS@-^Thck^frmtUDsaK0mIrKjp<lT@p^Jd|*C5qlokZu>t<xF5
zq!Su}Or^McP<k;%!sMYSaz&2>+|g;?SVAwNhcBHdbf$ze$jU-l{9*3S7#(qS3>>$4
zMrQ^dWavrsJmgtqY1^&QGvVm)5e|{PphCw(VQ!?$Zm|j3WrufUp%X}MljbTbX?cq*
zieLWDFCMI<Uv6PVODxDKdFTYW99DhkvTO?H?|(TS(kZ!c^CP;NP&$#Y<H(N}8e|@x
zo~T26>5twrkf8&1p~2UT26OJbs&hE4jsYEc@Agu^+yfcwR9s1$<fWs6Hc)W}d9Q0H
zvA&@xIT<{0K0mVfc!GsR;|YXZydlLI;B;c&s58mnj{%~OJJcW;EwG5lK#743sp6=F
zrR5cOMkz~_XVlL?lEp>dTIB3@(w2oSm3IwD>14V?s{!hT_@5l}CS#69L{${D`Gx7S
zzP;iPVoysS_c-E_0gHGg9u_lSmrKb==iccTf8HLPSK3-x#AHBBT3epaYmoP3i{mRE
zYM?bQ`KD8*b!*QCD^oz3hqvWQQIHocxboU`gpmh+T+}P9t!_bwH^>m%FbEkw&;^~k
zn7nK<K<7FRs`6FN$zJ2)Zc-=$q=68Wx1Kj);+toez`SKNUUD><2pHb7*b{5q4qef?
z1(I$D4B)gOoxxl2w1vo!T1w#v8=Un4Zvqn9ju>WUnVUYi(LyGzHPsMRPYk&U4=yM<
zqXS5~Gy)xOnaI{;j_)X~pyw?Rovd<m4^5|!MnoThYDe9Lu7+vxG;mvPOZB&0;Dn`6
zR(1yb@IlBL4qC8;O&<t1;Ye3#A{sZ^uJACj(pHQBJwS_xEKSa81!X()5!msTENzg|
z+I|!1PaB$e2<(522b21ec+rXTp!KVSv>YZ(*cL9hYFY=?VS>(LutxdA+g{OApwiWJ
z+FK~oRjMjokfc-AE{R+rE5G+2{}HaA3;a+1*Z;>FE3$M}!?@cZl-omu$60BNvu!jS
zZL43#kU0+H{NLrq-dxzDE-<>AF243_vejp5ui(HqU%+Li^(88d^BsptIH2K&zFj=F
zkB4n(2*f`0AE(FNWQ`_(Ax+mWY9O!MJSMwGwIzI~N3*e~ag9jUQQd{U=O4n!@%7<w
z_w{}<@C>*G!i|I>I>|G{P2w4uC(}%fM{qZv9-v9RucEYTV%Xq1RStQ{&;Gq<N8jLG
zJjY?kyNeg23Omf_!0q^eyZ8g6e%N_NkLm{)=W!R$;77wF?j-z>N3-|1PKmi=^yEF5
zEPNfd(zMV+7{)gVI(@@tNL%B$9X{?ZUek@!2y-i+0^|Jh6zXyDpB%QqZJESpP$x0z
zZ?mI^i8;P8#o`B(lXHtJWpQam<B)~yR9IQ@X9PHYYi)hq6Wsou(G*7z-Ye=}mrM1*
znarK0pH6zkpMDg(xEAN`mE}2&3q5wx7)}@c!8ccD8f!4N;3&6Qf2U|_N@Es{bvS;V
z1+<gXGan^)a&c46F0aj#z6#^<xw-pgaek(3tSy&ijYT-2<LHbDaE_nLH;SNV!3TXY
zKX7%%w~P*bB${_1v4Eg44+~u$^VX}WH>bXEX?$is+LVlC0_v|gAGw(LW|CUzWQ<au
zndqH~WYTE;&T-@!FEXmbB>gzFZl2UmVey!SG1!|M78Sw+@M4(KgRE}4HFn9kNHpTb
zA_WU$mP6yPoK|wJvE8L)e(l9dF|&+9c<i^Vyk6g2FB_X1o|vaE%@On`8lRrjg~p64
z^8zVjUZf$7jGg9`7QU&qtgwsA3zjA0bjH@`DSdtd&CT5RZledA%VlLjc|JSkT|Rto
z>_)P3biGocapQ7X(Ae@!dAfgeR-V7wEqi-Mk|m2EjQeLbE?k_SEvqZ@UNBi);N%gF
zW0f?HuzdZRyw{lXic?y$8;vyZ+N$Ldk1XIyXEIp;a(+-RBx@D{@-irSjvh!8xx~&r
zRLMBqFQwK65Ej<nytyhn&%Y>t@#8-)FJC-!zC7IC_II%;1B;9EJ{pvTm$!`7r>8Uy
zJ}p1|@sG+s`xpPJ>}WhZy|7m9&n<~o=Ip%c0`nj@3ePvdd1;fKkn@srDhrN`gQaui
zVXk0CGJEj_wJUxfd&q7oRnFH}oZ7<T+fsS-VB2N$RQcWN!=>W6)Oehiu-Tc!A`^?@
z94E;FICDOPL1(VXiiddy(ntC6P9ynTDI<>gX5PgGS+BC6<2JpkiTqYL1{HI1FD);Y
zwZ*xzwlwG6pS+k#Ie(#i|LGSm%l`hE#{ixuDI5!!lXqCmTvk}>67wG98?sRjSoER1
zunUx(qUecwK|O$GWofDK?K|cYSe#|2Dhpq!xA#1!NPVS(@<K0o$~@`!%fM1D@^~Id
zZHrICp?t|7pgb^t$B~eHQ*C3NMbR~v9gZm6+1XXRr+!K?hr&+V0Hr`$zhzzI%jU+a
z<eM)iXJ_S$=euQBW#8?L66o@e?p5YkJoC5Ml{WGpMN@vMgUs=8G$QRK^EO;fHn(Lm
z!%~HO4E(H+Nm?oUPP>S^`iveOmYfX??ErCj?8)UaIPaH@=Trs=|3>ACqdZsUUY9vN
zPN<WnRah?1Ps!H!!o&6&FLT0n^KUdZV+TxE{X9=AfKP))-xj`PsYYmqrqeT!scEgz
z0F6(J3mqzu?Z9nXAn69UkJ8BAJ39!?Z4dCn@W_0N-x!$FBs{ndJ<!otz$9Kn_$vPf
z@1{eu*6s)p;(hs#!?0ZZ>@ceTp9TDluRr)t{!dC>66lU{V`_&j{&jAdP~;APY`6;4
zSUXgh9WFkdKRnbRS2ScD46FhghZ;X@;1!F!z*Tp5ui*IQ-&U053m=MAH|ZuX8Uj8R
z36;b925(u+JMqB5H#JNq(cR5c9Jio`G$iJ*i%pF_9+n@AZq|A4cJdU!PQWrNq~%hw
zuvgkJ_g~iB0c>qRVGIZ>0LM6xPpNbiDmv&7^uYe$$zylO*>Pg9ObkRsNg<Zm8f-W8
zj!#JbH6DlXf>?9cyTZ^(R}-&1Naq(G0HQE?=WXru^3l(mbdE@iV(cQojV%vMJ(%!C
z7RkjT1`A?5yx4I>=jG+Amt}Wn-*36hvGB!Wi5RrXMEZ~_&foyzuuGQFG7YujGfosn
zH^1ClISGlWD~<;pdZ2>>0e*@fJ_0%}H+s+$7v7#6LpelaJ3$l%y&#Vp7SXD6!P_@}
z&<W#*s~@Oxw}*ttVa3DT12j;gcgIyY^23wG(AGGRh21~c;g?1qad^K;KXuwt1dv6h
z?39nb9O_pet|>2E_~cCJ`@E?lesq*bX9h#j4FgDcnqJDoZ%V{|z)OBQ>U73P6vu+>
z9PF3<)1z|1padH6;Vm)-J5wCrAzI2-I;?YLfrYPGYfF)2;DzkS>W&u?D<1E9Q`+hH
zaC8J6vzb{9j3kfCpu(m@l$}(_#~V1j0f(;95uG}A!ZFC9)9UYNX%Ln2pGA7&LcaUz
zu(9|{9Qbh_KHuJ=Q#Qvr;Of*O3vs`>so(ihCzORTI@<8Q&|vlAjGbgk8(E<=%`Hxs
z^{r)h6p6zN0pz9ll4j}}vT`&EgBq$9+F8&*oo^XHYlKDBP2mYsR9uxa<?;5T^5nrI
zbu^bW7*M{{G7Tx^fi${2z{boPpo8ibF~NelgQnSWV#TO+(G(qCw;B&J1Z_=sb+IR=
z9s-fE_i46zZ{+xhx1;UQvURxdX})n|r5;rJDznk;201t6wmkrCYa>9LV~boj$5q1{
zHNxAH#jlacZeQK-<FSb!c7g&9Ug?ZC4x)jEIvL{EoL!t_xIld2*XyIw8=|?6G0@Qk
zZs`M0fvOOW_))&)bWBx<=*dL(3-mUKS~|kHQmU;X3hFj)ch-eY{Uy9iTf=K4J(VQ-
zZShv$8piq<gb&+Lvj~rWYnw*vjzElQ7b`RC25T_7^B>4r)#h--oH&~atqp@7!Icgp
zJkn1QrV_ee=b-SGzcay<S#ZO%hHcQLp~_{FU>q*(Md(ao`QDb_`-A_iXYy|uh6`T}
zlHWayyAksi4jAC@Pr7=uK|lEy80R&Kcf;94|J6>vBKV%?Eq?Y=m~lCp9Ygjk67yV7
zo&c1`5cDw4zri@({k3ZEGIT(L^|m~{yTLf$+wOPEh!i7XWW9@LUp@olbnbhh$La=#
zG_E)G8p7Rf%-(ODuE8X{Lr$U>&iGn)t)DTn(fEh%QF{aT1t!34J}HORpHP_4)FL<n
z!v%~+Hwpt?*0S69+8E3C+u?f{_iJ)CxD^(saqYTIV|i*lnG|;~->BcXE+>Trf^Kq#
zW|X$#q{pV(W`1>kjq|E*quv#EvLAqdi+>c9hmx_2&qTB{ck}8YG{drT7mV{uE$wmm
zhh{YIC&By3W*wVMe2vkAR1Y~X2F%QOA&QB1zB{$5@%GBvMwwq)E(=TQo6jtidwicu
z{c<LGS%{kVf&gQW8I5h;mg6&yn7=I3Do6L<-jo@Qx7Sw}%A<!{8aFJKy~C4obaJ9@
z3lr%ae>GQDR@XchJ3G55r|cfOep@bhAy4C*bK#iCKD)T|n2*J_jTOnZvLbosmF6jr
zsR`=|a%eQBT3u#=MSdE89ULC}_&y)W$il$|yLi|^l9%#^&kH7*S>S0e7y_vismsXl
z)-O=9D8%G?#%8S##QcHvqp`e?^PA?SM0RK8#a6~<9zTkXqx!FuuIuZZK!Th8TaH5%
zo#OSxyY!0L7QQlRM?Gd{LVg+>F(^`UEd#Q+?5QpF*jePBC>O9mzzd3tOMF{srL3(k
zdvcm|v5V&DSYxz0v2H#}Gja3$CpNy_MA|rpjU7XziFjOBq$8(a^7-XQh4t7~dg4e+
z#*L4*H_9hZAA0xElsfn9wmLdt2i~>Ey^Q}@z&MZ`99`%+GV${QsK%hP8oROpwxBUB
zx)2lllTH?1Sp;I5hT{m?k@)(}J(nZ;$>_nSoH$>Z>rj4Lx6}#de(0k}X40;U^1}{8
z7V5~4H_BIuCi$Y9*%{HlzAT^p#h;gd`X~RmeEI5CdHBIc<$M2)UnvitJe5Cb<TMt^
zCi&+zHlC9XIKuKz|NO`0pZ@6&%h9RE<f>Q90kI2_c>Vo7bS^1bK%joCtgZPQeJmuA
zKKLRRyM|_^2dP{2hBB!<mz#NpljGxZe6Z*7{j&1=(SxnBwXyC+Kjt1dT9WUOQ(jrf
z=4D}3fWkNM_74ud3zRy;SenIW1yjCeL6k*37QWcsMEPb{5@Y%k)idTKe98<9g_4nl
z9`xm%bCPXsb)~GYXzZ@+<J;kz>#JVSr0)Loi{0|YGsz4sa~Z1d?Dk^U;kES6H|)r7
z76<bwhAeE&tA3G3<mH9Z>xE0|lw?A77Ry(bvjd9-J%2OnJQwAgX3%w37Y>(W*_h8^
zCmVq5p5M@q`odg<ce%;lfI3ONUshgiZmyT@?QPGE97~t%nmoHW^E@DBjfKml#aY#f
z_44tD56Y_a!nfQ$fA-S5Z!c*BDQhYSq)iETx>(fmr!$xcah#N~S?4j|8*^PgJ5}0P
z{7PB%LY8QUPdn;o>k9yW*^Qs@(4!WByns(y$ixRtv-^_W^&Olp>Lz7lS?%<q%CH)@
zGJEf)EZo2HGeOnbfF4}$ls2G~M~3ZU82+?lAaQd+pSXwVXl7fCN?Y0E`*2FDP{VLT
zJ|0IC8F%3#-6(zhBR^n`I9yNHVCe36R6F6!V>HMhj=(ri;4a)SEr2;qplx8k&b#Hu
z-SiE<lIK^z`_M$Q*6uAJ^!xD$jHdIi3Mqeo74X8>f9!>?2F*$a)PyS%qM>?a(gJ@f
zDgt1S9}p)lx#b@$_G%g`H&_XU%UvA2d6)}R*$QYBfF9Q-xlA663tV#n-CzaT5z<UT
z!yT?PY^~GZ@z&LIsrI3gMX#=q;-?uFwf=BK=25T)ptKKZjXt0pjRg^!h{h5J79}BX
ziGyo)g~o0&`y{NaaaA62StfZ@4m4qrLngY8Ot=B!Oe4pH{l@xQdHCp2+1lK6LmCQk
zG|L2XaQL7JwbpcaoRPR?9}cLnTaFs1?U0)Ri;sHfsj*DM5joPa(t}@iJ}zpR9Zq+`
zxg6lsjW~GFsN;p(Z*3`$c=LpZBu5*vbA{bChld;;aO}=G58V6rd7DUrC2<zec|^P@
zDSt4&nauXk_TW&lAp;g}Ja9}(Lms@#yDkX$z^oSxYM7)aG4kMhtxhUGz#U_*N@KCH
zp$KEdLw@c&^YArC#ixfpaWq|7N7%&xe%R5uJN$a^Q-4^9!bCpeuoUW~`N6NCIpP7C
z<8Q4wUM{n?Uk0Ly*AG%PE#xgUrm!q5B;}1RaLAf;fx`w`bnt|?dZDe8bfXRjaWZgX
zafw*2rL(=`!?JgJR8HuqNml6mMjeZ4lq(jMlt$i+VlXkMi_Y@Ylyt5RL-dRsk(tgN
zWiy=<b-+CsVNk_RvkV3-qb`odV6h6FGMMvrcy<2d>*GDxeMm=&POt5}`K1fF=yVg_
zWmvjKkKXyH^pWqp8ODw-(mTJX0hHf3;z$I^>r49A#qoxG4EJ1}&CB!aa(-F|oXHA>
zn=Y%{=_t<3v7?Ogq4fG_Lh6$`ZQ^&M&bp7Vl-@|ooCdzUkxD1ac?U!3(ruQ)6Ze!H
z+nZbEgNIMb+VZLf3F7J8LO+nTGC>)ZhGos7KH9dSTP=6;k6V`j+0bK~V5iJ?1o7l=
z0NY%|Z!!Tg5Dp*L1N(-ul7x5lGnl|J+ext`KXBPg_uki12?o5~UI`CM39eOEAYT)V
z^0suMPguN%yu&xHdLuk>;C4b=CA6n5netD1ZH<?}ylqE5dC1ZHEvHNy+yH28N0u<L
z%*F-7qDN#7amDHagRC_UVX%aEc-bNGBOi5Lb~B%>59nJ=H3|q^)<YwJ%=(DEOB`+i
z&w!>6a3-~P9Fb3e9a_q|+5wo+2Dnp?Of_7`n8pJxX4Y`TtvZ7>m{{<JOp#Gm+B1I$
zv^B`Wiw)g#t3qtLcL@PRjMM3@ZYpa_5JpFCLYvlA!&kYTr-VZ`V{SvwA!dWl-KDDs
z>t5y{7ZZO6=x>xKb@T829Cv*EPyaW+t7eNF8-<k1+tBdIcUya%s$kv~4jA`KIqTt;
z|G+(&FSx|nd?~*T`Y;ks`7_S1<pp+1vus^>^Avf@h@L;To)0kO+aP_z9vq*{#t#^V
zYw?Z4czok=_3qH!?H@p@A<p(8iQLx9SU3P$u=rz+^9KA3ga`IuZqbdpu~*6tpWK)f
zHZV>P3}bA5Jo|?<)gRmdJ){|>4~*l+b=Y87_MoyvEiP1Qe8mj`r*O~@0lHQ<TZoQE
z2*(HC7RLw#F-cCKO>~U|)P!mA*i<}W)3ER2J;|@HQ{#|6|0wiyz2B&Af!lfDCVU)M
z&G9yFNH<BY2KF2JM<KSRzc&gkKA`ahce{;kIt51gyh|QnQoO)uIP!4R9shUp?&<D^
z@pu|f&>2@{k;LOQ;ps!L0K<aK47-+Q7u+9brxv?j*tNk<jag-2cEZf-QGUN1UN9kj
zQ}}w5zj>to{Koon`Q(#t_)`T;=<}_pgM*`TtnmZO5R5aJaA&cEBjGRAZ|4X@CcBT1
z>+Tf!F`><IfP61$5g8U&y73NU4#w7)_ceA}U0y0%TkGC2L+bK9DrDn(R2P2vl%vP^
z7Icoj*SJDqX(yd__mz15q`orwi_Sce{Fcdo#vv?xH9g7{-F@kR@U(}FQ9O4b89Cz6
zN5Lbn{LwNCR*Y$GuK9Kpb(^>~&eZr5Iu>bIB%`n6Xk;f2PWvH~coyfy#S0AzuN61n
z8De1|`3Mo?Sbx7vt{W!k6^=zK#*FA@d1b*15lc%;3ODVAD0ZkI4{2heKRStC+6gb_
zJ1cplAL<EZhsDp?X?F36#^vp;7xoSg4$AJ{zQ41^LdTQsjq>3K56Y9r+sebmvb(!q
zUcTJ%A_SkYSXx=}ceajC&g;UIbnb;T^*J@ZW1((dWzM?=6)$p<Pn2Pfvtx0I9bqg6
z@LlU}@ed{H4;|?vZ}9MrCS@6OB4nNC79E$|_e3Y&#7*<7BH~4d%ggg}c5+zGHMT$A
z-zlf(r)6nvvpo9n<Fc~0mAoeZiCBNe`&XCjB75!KRG&S2RbK2JD~%iz$xE<|!_T~!
zLOwD7LH%Q4c#Us$J$z6eKG-UoYs*<6w486fus~k2Ai^=FEErQ)m{;PvY<%DCTJcJR
zvLKzTuPm4K)g|$r@xm92TYh;~31<$97iBqClU-rYp1&xENAj0E=#2RZ<^q_zLg%<$
zPL=0`oAaUz^BdGbj+JB~N##Vsl7B2rGj?B+9NU{~KJt=zf?19KpQ<cuZLBH{*NS&X
zIy$r7+HSGKL%z{>ro2=+6hDqT^f&vIX4(z#KK4_xL+@{bp+{ZJ53vA8dqw#|HuQCP
zc<7yu$i*BTdF3NTRgS5<lpB>Sm4&pg@xN~xo`LYDJs}>-U#r8+)iI~DpmMsgx#<@O
z;6=Tn9YWr`bo^R+f8#~qCyzGDw?Fx~Y;A6ogTo^)e!bc~avoh?W)6yWgj&t1JXdb7
zGT%g;=$Y?^u~V1&adCN}^c`ewNAZ!z*NRV~b)Y2p=6)}ZT2@_VWq~z%%_6qyxlHQ1
z!$BcBQ_8HY46ci`70M?T+1FG?+1)nt_PWd!+Nw+EA;5TxD{W?r8%woYth6^qSO1u4
zms;NR`0#c8uF(4vPvS!t!*D}B<X7l)+1Aqnu_JrrHdLOr-l7eR*nR8c#clse658qz
z48u0K3pY#)U``Wg8`!V&j(!G+JK-yNeg(V_jSGRV9tib*VMk#!oqttG`S_~<3t#vD
z)^Gpq-^s*zRXRK?Um|Ex^hFVaN|6T}nH4h^ocs{ezpDfG^R3e9=F-~pk43!rna+d_
zV-mm<LNvrig~mUJ0iy=RGT|n)sdQ_&*lX=}SpOXu|Ne>xMyn*?<*s;FKlEUHaKo^D
zv8nX-!^thsFIpr_{>>j7VS0B`6u>6QI{%@MV;2HklgXFeq3QE5m6plu)nz^K>!KGA
zwe}!#=Ug;gfbX)WQKnOfpX~h9*)d(E=Sp9-Hu&oF>v0f&d8E5oY3X1Jnv4;U0g;?s
zvNQR{&m;h>JU|i_Ir&Kz7BKR*#<?F#ysbhf$J>_aO-Y-1*wRVO;v3PW`0A$&CXR>>
zSHy`98wA(g^b-aectOza_O3j16T9%lg8JkS-ks;_=cgce!h)6(c<|Hl21gq5;HPuQ
zqK6*JGW{lz(oXm+)QF}o3KE{TaxlHPBVKJm3wL+sIkNH4#b1097h&kgOE$zaUHr))
zWU~Gpmmb8TLm#a}j=$>+rY_2UE0;-f;`M{g%rXdo3*n&wZIHq&M&+HdoNz3{(n&ef
z0OwqtTej%?4LVUed@W4q9Q`3bi)pj!$bi2=N2F8!lvh4>OL*!BogF$ue0$5gy6!Uw
zWRN7DV$M+mlqsY|Hr^WYAc(hz#E^VpR~Fy);*C6XNCzD-t-{k`w4%h%yB?J`;&0#o
zB9GA-i)p-RX};tI>5!{+5ZgD`cx#Je5dg|LKxcAJonbnq@PJ?9lfKcbbYv~5u9|O6
ziwBjalZ?)Gl?@ou22_)Jlp`VLHR#G)VaSzoP=WXaX+3d+;%0RG{cmP2+g({|S;F4H
zVaNwN^h=)f5W7(o4HZ@-qf9n4&{<;D)#M3lm+rXnk-Mb<l-;6$rn4DmUtjU9&JG3+
zS+h%77SJ^gTNM^J1`xTcqodDdl{GK&0lbi)EVGI1FkQ9Aw7sRH_sgxC=4oVbf{t2v
z`|u1N_{0X@ce@R^31fWo0S7Nxt3JGV2D+9ayqbPH9YG+)*bS|(>c<biVL1ILza|vq
znO*VV{EuHN&glaQt8lG6#GUk-UjN2R9vyxB#M>F9Yd~+C!x(KllOTTnRaL;N(HSAL
z$<W-?7j|@7{^Xu@XU2mI16?%`rqba^9R3+s-Su)X>=4Ve9b-suy;28o29Mo-_1E!{
zm|4g*B0u4qn(+DKfAS|k*Kd3^nSscKEHTHSqZ^u7Yne%21e&cg<u~Lz;>KQa4Vr($
z4b5h^#v0x??h$5$4tpobs9$gI^Kgi7Zi3s*&o~NePgAK34Oi9D)6{DmH*^mneK+x`
zm9ndoA{)}jZEJ@E8cj!6W$$QYq+il#ZS80}I)UvKhcxj=p6b`(WQO4Y7dz109c}dr
zua>^iu)|AOU*q(>f8*KH1i_W^($P8#Wv^`-FycSvo`h=w(PN7Xd~_7rh656=$t`!2
zVQ^JKZgCI%jJ9`;4;p^8tndSuu%dDH)<6|T!dAZ4jr|5VU1065W}xE6u0VJPhud5$
zuHk!}ZXFF*>nHd8aK=3vw#SEv%Se2Cyr84YzoT=l!cX}hYv`)?7%rC^YvcR07$0md
zaQr$y$25RO57aOfhCCB?6vqAAHR`8lX=7_I-6~BG$@|eO09oTV35N0YVS9QOTSzVW
z=)BbUM}sj>$e*8`c#{A0c(0ru?v&$$S7m?aS=oL5d3o{0XXW`9KP_K;{*&_AXP=hc
zmwZR;W!c-`FUP0n<(LJmbB!scX8b*$)ul|@vonR=IM=$`cfI(o(eB|%dAWN~p1s^H
zJ3D*j==i9do#$I!*BUc(v^oo19L>h0JjaJI5zpcnV@bwWj4#-!!qI(M3}gJnu84eh
zh6NLjb!1YOog!Q;lK6#33C9?`^-oxIp&ienGku-7)3i4orcAY+V(2dmU^QQBM}hGb
zivqj=dvzgsH6~<?%Zs7VwQ`;KvKZlUo@ASotP4v^ele1T50A?ffN?P2^z!&kyzG80
zQ#={hv8bkq#aQEKxf$PG3&+A;#?g!^GcnEtH{&(Z&)AU#PQJU$*vZG}OP=g><Kz&f
zm(xmETw?r18kLtE|2I9w0@T~GFmum4?^yWUWI}#^PU*X^{E-}r7k#g;ES1fbMd_<(
zjC@@#uQj%mJn}9#8o!~hi*t@e<fY8Za(a4^1yPRcQ{Ix_?7%ubyA<xC9G;w)gX8mZ
zbaquPlvWLwm8Rk&G>=ri4&=Xgcv_BBHZR!8qV#ek;G)K$voq+M#XSEd!&JFcTCc=J
z*`cxXv=?({?kPUigD*b&%kt{g^Rl$CSibw+@0V|U>pNwAeM9_8nHA0a)Ej^Mi(P8y
zcK_(Syx2V{FZNH%35yx``K-ZQS(;<~K9jL8dR07p18!w`rEG1km&Xq`%hN|2<>L>w
z%14jZ%HwtUEl#VSepQwiXMJoQVw3kS6RNk=1!ZuV6E7AZn0q>xj+lF*e6e$iXT-V2
z<<t`vk=O-wBVAet^5p~=7HXJ3VcgGx3S()?JUpmZ%mJ_v!vZpSA#~<oL_-;AbnT^Y
z>M!Np-wRZ^IaB-`MM}Q1@H#)si>YM5Thp*8#%|Q`WFc%yG?Igzr}}w8QE8<;;dwwA
zRm>inP!{c`Xev(6yNEAyCX`utdgrI>jp~ao=rV7l_}u5Gu9N?ymBqBo-Q-1vTp6Ru
zs*Bipcze!|vZH#I-K7c-Zzd`rReeED;$w#+WtsUbcE_>+$Syg`Om=-Kyl5g{(%?BB
zzy|G1+HQ7au5t{i$^@Q-A)b)Q=JqWY<-&DZJTrHyv_hww+X4AON4d|AGliEM$(x;a
z(m2st2rfs(!kOAn&)d<Kv5?G*0ctzchI;%}>DtxYvGn%%_s{xfPQn3gJQMejccaIw
zc>W{wF=LMz+ikk$-|TW!qv76R=aPHew53Amd&Dq&gLmPEe%L0_Hn3mk-O_b8eS@#$
z`4#X!G?RdEKTAl*_p?Pg0fzbivjN}uy8l~$>+k#y4UJ`kAC1F6P?GB8x{~#n1Zag?
zwYCW(GvPzWv$=V>hGVyh2meX{eYFA}M2IenSkUJ(PY6t0(G_n;i=*1{vkY<qv|TL@
z0BcKc!!zOHrY6Y~p0p3skG@*NW0Hh*fFR;d+!>S`-HS)!h(!+ba-&45LlEAgfvg7d
z1eU|$7NYXCu8cR4O&VT2{c|M3Gy4U|$ptz$oP0b}9j1--vbnkG4|V2YOltMufGrI*
z#IULlg0}Ew@79o`WdG|AT=nnXUI0hqFQ14IIujyyr(13g;2#iEco`ju067zve&XF0
zYk0`gDRC#Aw|rO-P-hPv@o>p*BQzk;%{r_0X4Fz=usW+>lcDCEJwhw_orL)MBLAw@
zdazoN10;TuTSz^=Y*!NGlW-0sJjX)<@NtI$KH$|k;31wjwdjCnQH5g<O{WK;7kQxJ
z!3;Yy!4L=G*}|R$6m{AiKlbR;3vRWu0v>uF(Wvw~-(ZI%@|TBQJPcbS1r2t}c^lwZ
zJ2J4fvV?;U9}s<*D=ty1o-l&L<<1Cr@v1VIuP*Z4XESS{#^M*e>1Yr?Z&)Q9q~Uda
zbLBz6lDFln6DS#Zi)nstQTB9F+!;|h;ls#uMmcVRWKzE9TmkILVv9ZBz@h_5d~}*P
zznTG2KJrW_luj*gWohuDcrA})bXk$<S|QUQ6gqLfx|54vysRgsA!SM7<%Z@)17Gr=
zMP7C`USE)R;%c2GFh?4Ar`tRqvz2BwP?%z1M}9-13wrV$r}Iv6QBS7s@rENb(JN)b
za@MwoBUO0wj&FS}E-YDZE(=Z{6c8r}aaT*sN@tYr5vA(JkaA=X0}LH7PMf;;3r~8x
zu%?TQi6=Po!=cN`s`N>-Ey6UM@BYK8>+sd~qmQjg%`Y-`VR3gRR@rO4PF%M54mz0d
zNcQFVWEylH#s+i7U4xh$fUj|(ZG4+-#_4AqPSeU={etL7P1p1U(&WZs_H<o1F6eCO
zFXKl1nkOJ!*DhCl{&B`f_n6JIXAz#Av)V`}9-j2V3ZB0W`Qrz|l(afsP|;MvWQ5&r
z)1bmDSYJs~kgYAub8X;{XZ5p{oHa@|)o1%xg|hpZXjJ7wE4S%nM;^=FrH4y+f&Fne
zAVK!#=6g&mH7>i8qzdtex_bAQeh$VD-mziu7mZzpHYiTG_z~D5VHeg~@4|3z<xqr8
zKEcm>;j3CBTa}?pXO(yK(j9rL*pvac7q#pFrfa-p_C9epKVZaN?Y-|P+!YQZc5q~H
zqpSRIt8wVV)V#1ADsH;#vaBGQ9jLU89{=VZ;)iyi>Be}uZ_()@j?O)>PrBq=O`xX_
z4CyG3?ds`r**S1m|HixFgRJ)ON<{z}zAY>`tSy`umcdpU(=~tl#ogL@%US7xp*g%J
z;VN2S4kNfm6FYg{($R2n+i6j%zoGH!FJo<Q<`6e<cjYT_^>lrFJx>1B^kXazV#hml
zxa-w$x80Sta}Q{owo2n?PsKx9>9IBXMhh-u2!3|<{`{)FLFeZSOkdC16&Q67{f+aV
zu=X7F8_)-a{_%62;_C1<T%~R0Gj-F}TeLur!xgS0bl~GTg0~-@_99%!#{Gw7p_$__
z3^U1p1kAR{IS7k~-HOk3dEnPw;LJrZ?$@sBEg1J}N_YA4!@1{&$r#7|Y!Q0;)Grq@
zK2u-*^4u?Sa^yLuKb)PNl*@}Vf5V8~D`#iN<@A)}y;!um_T)a_^x?Z#-Vrl1r!g;M
zoco38S$2alLCleR=jV(YGM4cpsvF0pE&>_<v1pUAHDY@*ma&KMmx`B#v7<v4$d0ps
z<^?ClqwaZmL5IZ$Uh<onp7o-GpNqQauWOXzxZv%Dc3Hf*rSTGJVE0FRVU=U#nNaV(
zrG`nEW*R3wn;|+KJco64dE&A_eDd=+PvgJ&`LejUP!=TXq8Gk2c9c%|t{01OexXl6
z8Gm_fr3?Ld;SOG0x-k*W&X0Z)ot<Ee30W+>C@0EC7AROy=;g`yi^YwzGxD6H*0P`|
zwD}Vk;g61vyy&^R%i`x=xzIS19YHH=t7UCt!@E+JHMV6ZQ+8&}mW|b=vbMbFon9Q}
zH>>n5FU)z-Y-Mq-EYHt+=OBwt96!iH(##Amhw^eKW4qwF=4a&A#RNMGa0~KhVT%_v
z`Igz)iN@h72VM|-eJ$DEmPNvdCkrKhk;{ofckJl95uY4Un2Gzh8pEEIomVfE#$D;@
zwZ^Nn($TaRFgV3$cW1Bc@9mZ2!~Mc?GT)Y>bz?!61x9up@r^VVC^pttoc<N1b(u@F
zYr?Z5irrnS%kyOmnvK=6#BL>(FUmLZqHos;xjFie#XxpSv53X-hGfo^%L9v3>{4UQ
zpM@AFk@BE2#4(HhHkK|Hl-Z3$?eJoaXwe5_XpYfa)J0j#+!=E$EL2{5@rk@nUCf+M
z7JRr8w-@7(No9(KTaIdEcN&Y99FfR(z$uWK6M5^g`uyCqk7o3of^<*Z%vJb^Mdme>
zcK6X$_c&V7{>mu6sm9ToGcF6M%bp*gO_3WKmwn_DuFV~Qqpq}egsY{KIs#o=ybgV8
zS4@SY9uC3+=$LQ-b5ih<$N@!$6&cVUy9?J>mEWo_%<b^9FEX-7OPz#Ii>t-!t72Hp
zV!o8J<sEvpJzyq}bf*oXtfME|I%Jgo25A7$0d+qw0nmmK4m)ihra}nk6OVM!nz_Aj
zVERvH;AWe}LdpTVD(6&(s7K~q<M8uMrh_)qEbzlsd5_Ta?gk_N{B0^M4)ee_yeRu4
z^c6;NiNm&P0zG{%L!<Y<^Pg<zjva0HzSVz#2z`$j@@nudTpu3Uo2-*B`!#<F27TP`
zhDq@^c$Z&;ucqnCWCKE2ckjV}1Vo;6Rrg;R&|9_sj7+`o^|yZOZ~u<ND^)3&xww%?
z>`AbTZ}4thAj0V;(1D?8it(LWb>qN~kl;)U0UZt8*o}7}(@_zLHBkMr2~|T>z%L2#
z<PT2$%-3Ns&C!10lhBGjel|^!pB-v=pv5Zo<|VtSklS=MjQ!#!oDDf>@a&J4YFIAX
zyU=2W#xT@}Aacm%tLhrOuX?Y{hSi$R9#%K9Se-x3Uv}pq54UvKc++ToZM{6$eo)p{
zSKT=U{6Ng}!{wtj`r#vYeJJ<`Y4x|M7P#5d;Tz(J7Y6*puh(H@6rt0GuW5zHg-qEP
zW|w)g7?U?0=v-n~2SN`mIt7^iUXePEJm~1)^EL;aW75e33EnkHHU2!9n$9c(Hr>F-
zEt~XgndLSHCGj^4@>!HtoCyp{x&6r^WG6uM!K0FZAx~;!$A(`zKcVw+j=CUVfr~T`
z7r8jLkOw~aEL`E{5;yc?doKL^W}VAJ;w5ZkB40!2BNWM#nlIdveo>JgaJqx@pc%hr
z+7Ad0w^f5bVepYXJOb9AAP;KU`3u*|6VU2Q;{}=y%enIVQXN3zM$O>q*iqIANyotX
zXEuso<-o@rikOZbi(EOzBK4SZ!hne51z8|-XG9$k29hw8M3!F++}ZW_#3VOIKhWWE
zER<E}wK`^spMeee;1r+$0fq;@s-nn@>{y&$@ur;HPI}=-5m3Ic)9+dXs2imn@S9m2
z^&q_S5<UXr^6>%c5lwqQla8a`R#d%c;4*`3iUa+-V=0qzg8rR{2x)<J$z(B(1$6W`
zJ3Uuc7M8q-XuViJTyRf*q5G~K6{527-E9_?1O$dJC;-AFT%6Uei}GL#DgqA==xZ3_
zb)3cnlBxbs5!lYlGQ&sKdfAPYw1h&A*tz;!a21hQWVPAcLz_H12^@r5?4gtC!ipUe
z2;SI^j=BlI_?p7J0r{Au>R%4Z1zJr#3{^M=2imm3I1CQU%VpZ|=OS(bbw1T26ftAh
z-KgNB1)Q<r*YUR9p8Dest@FBi#sI5J!EoDd2oE3w51{g)t{GR|l@2=&8~zy^8ruc3
z5EdT()iCXe&gIn9G6-Oeu1%VoE{wEg;3SVdu?ugMb15hXiFOR4cUa@(sZ4Fbv4HD%
z>+WBYETC--7K`<P3%iLTSL1Jt<-;Ttc&ujPD{lB!zZwU89UMP?;AC&<uy2ji{?5mQ
zF~)Y%cCY;LKmITOYgzbekhG!OnkP+O+v^35Omx>EX{de!djo8}Ym^RXZeKxkamCsD
z_kcTQv&SDh`fs!~Y&CJW=PclK=rTsumUeLM>Zm8-fB_z8<+8~;G&_3`=F@10SA(HD
z#?T*oXnJ>l8PmD#v^m+o;&eq9v9~MbyW(Yw4rA}Ybk$wMg}!(9{?xbBIp<aNZ|zoV
z7aCM=UqipfugTx)cb+Q^-`87VI(I}L_~W#@zXId9A@4@h-0<QuCci;^ec2m?#mixH
znZkZ@*LK)$lMl+Can0Rm+GV7yjc=0)BzU>i{&uL2E?02TWhalXxJ_bI4bwQ)2BYVN
zGk!fT*Er5@es<eA0*%I0HB7s}ak>1^&yveWyH|WJ02&O!HJ|>+KM)_=x_IpH6(jx)
z$L02;I^)-0j=>(iS+&{WKzDq{ApWD*C>~y118Dqo<*56}=c^j2%E&Gr&BJDY4ckm$
z#4QryFs|}yGV!8sM=AI4YI2#DOrbyS%wYh)xQ~f$CehEz)w!;V3r}FP>uX_wqZ65^
zo>t#KWg9<FWXr)sF)z)9EekrzcbP1`G0ySMD#kmt4lv$m<4Pv^1IkUthq|t=8E2oB
z(~}b~+^{>0Z-Qk3hcFsHpkE@C9!N8ba7;2geXegzxa*RPjJMe7!~&kbW#sW!7Vw%H
z-(r`1*x6y?Nn^%2vNMijOyWs$@~6%pFqTCZe%Vv*S{+>1y<jG0ep!^Uko3di2IDru
zcp+deixS98yb9^Do_A$2(GNq2Ne^~G=mO}0MX!_7vvP8L>SN&i9cOl<kf$P+B<BBG
z?rV0VDE{nZ$rzTS^H}hD`D(Yk+}SU$bnPD;m(%l0(QrH-N7T=Ffrn%2Ir@-q#c?#@
zVqM%?(pZq=-}si>gZ1U|aAUcwFVB1T)Y_89aZ9shZDqbZdaz!eKH4k~AFP%~n@i=<
z#!}f{o%gk|JgYEsWpj1Wi&iXZDdlLp_`tZEON}Kj72iGaTC{#i2V+KFJmouFyRTlA
zqr-hK@SdL>mlw~!C_nkhr{#+;pOt4Xcgi6<rOs~3&f!UUvAb6eW$x_md$;1Ny+f6;
zE9cAB)`rVH-{V?QI{Zy5$+fsJQ&tw`FZ=e!s;`GztL5>74a+_=^}5U`)7hOhrSa>z
z^nH49Q!Z}qm5Xc2XkECST_~rQH)VHUx>vmHh@-yUqb}%TrxKqU@Hep-hqL=jc}m`q
z=j5Y?)*j2VFh_Y~=MeJ?EW9vYM|*rHYeTYgJSBC{FC9y#<UQ?&=hnC?(q)vnotrGS
z`M6LOXv*T8@|b)fF6MyRi?Qr5ql{8VyfCV=(Y|ZuIK-SffxpX-^oI^O?r~v$(Vt3S
zx70k}AC&*hY!-Nk#|u!3i@c_uaCsqEa*=M&B}sG&N?XLHzO&c^v_&K@fJtFazudl%
zO`&-i-19Tkal$GNA%dgMBfMw{e@*^Ssr5%b+8n-T2<_I|k{3SNv4>175PH!~;XLQ2
z`U{=Md%Bp%$$SwDzU;JIQn_AIo60;AGNM-6%Y<?JDHq>7pPrJOYM;8c4k{M7gtMM%
z9K^?IL45z7dSY6|Wqnl|x79MqE1wP{yKdGqaxKlPooAuUap6^ID&RIh(Z(<S&EC-<
zWpxK|W5~onE<y4$;r)!PvV{urQPXLxc99uygBholZNx4|H5x839?o~!{miv>;2*0Y
z_#V;dh5*j#k=d?>PuN`kjp2sZEqdwUeW1lR3Et(`;N3Je13EL%#{ORipy#57%SICa
zD~GQu6Oe_k-~JD?@YRItK!p*bOK|*br6EBv_<^*Ja965IZ!&Nug%56JF7?u6gYdRs
z5P*w66#{2ufTViGzq`1FNhTvhZ$gC7Kb_bxu%XMhL181l%FyA_K*z_o76&|V$Io^<
z@rxB!c4G%;kLqtW_7Omcf<RTB`(EQ}@>gscZEPt}Ac@7P#_l+(-tdf%3obkW{L*O3
zS4hqSURhl)4<0=5`MA7gf@db_TPV7~iPHtytY!11MqZ69rrn;5=fRW*lpg{LBeXgN
z32z$x22ynfm@McdeikC22HYudrw@$oJe+ud@<7xBinpzp@b?=rdZ6>x3mpKqKhv4y
z_y)G`v%86hb6db@<)rB>^2Bsn?Z&P+DZ>pDPb|8L9=*kuv{D}J=&&--pXhb*eIxsz
zJDD9nxa*(6D7s8oFJ4KXe#i)g?>+guQ@{h7PP+U#4$qx=?C2kTLt{JiEQVB_k#{mX
zATFYV*0NSymEUfcpXkN|3?7>Nj7xcg%pC>l7;f0258+)-%nU~@kETsj<|7AGj!r-5
zktbon>nBf4EEbV;N@pC|z%hsCbm}A<ev}bp;rItgIevo91UPiUj;tIrpiZ6!SG@hh
z0LI4~N<K@hcv#4xBk45r)|1K)+2n_z<Dye}r9l#Jx3PdW>o^=<@!Zf6zP!@KnfJ*D
z!dTa$;f)^TvP<-&pBpkGCuJUo%L;lEo;-i8_~4zYlX^jik=>11#DtMRGP)zG0Wclb
zIdxoVhIpg2l`Z6<)BIK)J6Ww~w;^&67Xe5UolWly<qa`-s-rn4ogo|F_F7t6c4wY9
zz>*H`$nC$b;9I%C-}QwXv_YC56x>ZC(~C0xnF7MtYA}}}{mia;8{%+0<^u*?jiJ$o
zz+tgj518Qas6d-QJqFCnbXag#32YL!*%MblHe2P_MIHzqAOh&P!efND=%7#*gf2!x
zRUW|#V`Qw2u7S}NNw^w=ArnQm*l6V<dU43`tdF!nIHatO$t+*Tg%>o&0J7N%zlIAK
zZ{DHDoh!VlH6XfBmf4gT@v~=e#^aLW>>cu4g4WVkr$FW2{z#l>Vmka>7R)jyT#b9+
z1(}^y4$O*%GGyQ+2=0`7P_c(<0N@i%wacM@qwUH=XSOSuB-?;SIN2Q%8vpF*@wy^&
z%Dm%2fNCa0EdbI(>yPnI-V2~5NCXWIG&Z@dyAZlSqhGgj$Y<8ViS8f$92dTR?Kl4e
zFMK(@W8gY1z>73Qzdha1-LT^{Ku0ry9+(tHh~PWtsQ;+D(ssCUcj!B}47p-84n*(n
zQ2bgLLE~{3ZW2B!Z2rL;u<J+Jp((I`rE7iK1huE_gbbhXh?lW;W5$1+uHl=TK90)o
z9sUi5_~AA5ldampwQH2-XK{nl9qcI1*KPdJ?S7dd?wAP|cW)ne>o($!Ym}ZVIP}<X
zMh)CUx=~zTuNsZ6YQ!hd`^O&30N?Bm+~bGWZM-ouk-f<;lp-FbYhi5++(W-n`idTL
zkHj(Jmojb9@ial|LhqJX)!l_N7F68b#0A#?Z=AsOW`N6u&%m@DIeS|BRJ)zk&z6op
zbP!;*k5Q>x_P4w730<X(KQ`lgJq6!r#_^M3)Gxe^(yf)LPzcXmx6o89Zu|6l02Ao#
zA@56R<I(s49j?0N@KufYVHt-?Pz?d`@SLSTl5w72s=MIG#M5$d#<zP|0MU4czW87)
zLL1<*iCO4|9_99B{%Rv+abI8KJSMz3x{KgUuzNfLE01Ah>gKVyOvd7jOWA$k9SDre
z>ts9=-`MRIp0SF@tt3yWpu8qtCViPSMkbB9G}dZ)?TKUc4KgNCiY$9%g)aWkHhNgX
zSM(YTDM;oJ^dp=nmLZY^&>@eHc|7M^M<?E4$GDDx6l`Xt&t;8oIKI((v3z;)kMDOa
zEiUIMc@}{*P9ja#BNN$mA(6Nkqp=`zuJIrXg~vxnUI05)z99pX-sBq#cj!2}l5Q1_
z@u7F8NKSua>FP%IE6IE6<Lo%LlJ9u2(DLSOOCPdl5#zo(ac{iKDw;sPEHY+XH_dLZ
zy0Eddv{1IUHg&Cgfoy$czN{?FNS)W-5yp3{)^xG+hxD+^jqz<eHH5OjbKKv+mX7jm
zEb0FKOu4wYE?+!*RetyvpO!D4@09(+Gk<f5>kD1acMi*$RR3mbUcxVzx6=#d?frR;
zIp@pkdvoRL?R2?*GhM{E%q_0yTJpE@=H~ep7Q3eAvTIE$Us8Us$h9H)x7JtyTPmCD
z8mCG%XoqiSb0p%8l#RYv;32<QAUQe9x4HSu0TuW7_}IG}ug@>MxW|qv;-O5RoO47d
z3tyRkp{z5%;CU(8z1Zxs!~!%6P79W`eKU?i#vBCS+~3~b@Hd-MzFE}GE=O6qvgG+J
z@uAP+MH$)w<_ZLD(TcIY>lF1!m+2%cX@Q1slC3JA`6d_(SX6X&Mp1sU03;PF|0w$$
zUAnxySQZy~K^Q&Vx1N`j2GZkW{UjfGPx*4aRNTyIwfc^9)FszX=^+bh%)3Mf%zvhh
z@uY_3g&%Uct<i-{2FeyXU`|gaZ2)Z+i~Hz?MJ`_Mb-fWe^5~{?u_*SY%u9Y2uFw<o
z5(uW^uXvPKd<(EGo)eaxxl*<KXya~#yH{2eex>gGLPq8}Md)x2D;nFUX^Rz)^vH|T
zEP5pmBWq}-e{=&2^1Zch=t`#Nu7vYXIH^i9Qi3BldYw~&U!KnnOx*U7#qD5)w22jO
z3S&p;VuyDREiPn|BbQ`Jn3P}tWTJ25O@{xtnG3tGiVww*#*WS(T(G)#-qj*#_RbvI
zF=Oo@em%0$P6GT!{l~&`eMR^_%s8*R;j4MR8@`%`+!}@Vm92s4YmlM+zgl>IoeiNb
zeEpr@A)Ts9un`&|vDd`owBSS%lqWDiVRvmyrwdHOW46U6_$DkiL9IX7d)_hOMP;s@
z@vFqHXf}<pX<z@F%+=3{YJSx;Z~U8Ex3m@x4Kh4svyAxREr(lI9t1%*(;-Q}!^Ok)
zdfDxJOfIw)WQ&!n;s|42YHT5~ZUvIl>}nF}q_iNR@FT=F2%|&Xgws_UCtjo=);BdQ
z4J}8!Z9jORhwFAd;xHXy8f_YVM=E}n2qz$XI-D$~WI@GtJwSMi#}6xok3Y0=n}Nc4
zQB1Ucs2I&df(`<7xg6H<sB<71-csQsw|@9lo;<Kvv}v@s*=dsnFOCT~Q%8^w>E0AR
z`g$%m9S^jSZQ2Y3)S<LKkPVA8>Y|2*qQB^}g3;jP><Y9%QdJv(z6eYCkW+T+Q;x_l
zT*b$c%T~bjn7Zk5#}T_9czN&<-num%ofs~36ej2qqI_TJtsnmBL=Y~~C-3S43*|)M
z(D*AJ$`Z6JPGkT=nJ0evNq3k|QvygE@%KZyUX)RxLMLWUhsnba;9>5Ez3oY(J4cp1
z^5zz~S|<?o!OSetK@D?yrGbPx{VamH91%{xyiG(|s&Y8&+j4co+kA9jvMY>^9UT}J
zwkQz1-9x9T$w?l1*IeaH$AwNZ3qKjupw4vM=;%@|uX!7bfy%|X@q}|ZL1x8G*=`-4
z)^TlVC7NNmp~K5@Az8Q$p1Msrpp%{YqH=R38Eah&h@a~C0%w&|-_<!1rQ@N)Xo<YQ
zob*|S8Y%L%oo_6dBdd=|lnm|+iV7W34rf)jmNkf=lk7ny97qc`{~JysD^IW++~(;=
zdP%g?9vpX*z{fbd90t>F*<J18@P7)ufq=I?2D;GQezxOJ#Np?%WoxDLy@~`b&d}+K
zGi4kTw|Ng_HUId?kh?|#7Cg1Zyc?#*<6uE_{JOjStBX2i;^?<egD%~yn~oP@5|h1c
z!q0cpHQrb(m%|b&X6(ohA@<;GH!)_TuK6H-)cO{?!&Vo+Ca~|~5q$*PWVY3OTiE6X
zsy`+;4e%T1hDxlYDRhpj(aBM%V>*6*c<8S3<hXQE$2&jW`I|d<oJNaYg~oaXiM^5s
z#7AyW!Rc0YPA(EI7hb-ZEO&>BA*b6^+_tyy!4WikkYDu{KNAuleh{%x{YID5rVAZm
z$FG%zW>5GUx$oe3Hpx^#^lXyc_($gs5AX1^+YGARZvCVymlo@f@G<|<AO9Q|zJB9B
z@Hf7IW_Dnis&(KuPUrNz1NTn9dI1eTLXW$%j{4oA6^!GYd*CoZG=K8Ca&BQj`^G<B
z)lT>tu0kKC#{uJb_g4kM*^=^RbjqLW4|W%q$`4allX9FdCeQ2mHvhyGh+pvWH@3PP
z#GPx@ts2zXTlk)axJPl^NI*~y$Il+sU9We!@3J>N3gas{tUaBd)71^Mu)W>34qv?n
z?g4G>y8)elwU6Mco<?zY8hTuR89yRNuCA<hSEaONh#&Y_3AA}*{`#4sx~;#U1~?(c
zaK^ihg9B-UJCgM@Blf{%D(sf2uX~+eS7rni-}nS8v(n&g4rBukBl<WJW|)TRXNM7T
zPoe{YZzg`#*+K9%Z;cP1xG@L$j?A_R9IkT-ppPDdZWhUFqj&R1=Gfa6XP34SeC+WP
zG<Wl_fsqqbZlbnd4=qiNHU!{gke9+4!_BDpjnL{cdVov3L!L-F^w)iaW)y~gT&)dw
zdvjgRP7llOtLJ5Ze@|m##=R63rQN%Sa^#o-d9szhJd?Y!m@ZO4>k{I2gNgOL_&1Gb
zD2P*I0o@$P<5BwCcq>fW8(q>-D=QgC!ZrLo7J`p>Fus6}a2_|b^kWyxJd%1ncmi1C
z5BwPu@f{<MMNA!mk3-4Loqk!uA)6O>lpc25%+4_3ov|D7G5*R6vvsH1<)wqc*K-<@
z*+;`M(eJS$^9hWNG_G1+_DjU;8yo)4F_Zb3%&szO{6w74lN=TUy(psi*e&G65l#cS
z%A%nASy6=ZjJ)JHyY>Pr<44b%AfOWHai8$sX?J>APR|$%UwGl>ki|7!<iApB{cB{e
zak6-e)d}DCx>DI-ngBiOx?d;iy^{?w6oMn-`24}?(Wzw3t{F*ME^pqJlT*bdK4+J_
zbgVd7s9~aiZmGPPnknbk8lPX>6b((w(aB|b_G-U;`RbtTpDON~d*$f-bvZh{_Ttvm
z?2^pIGPAf=rWe-A&D4B(JF`&k%`BChy6D9>>)3&Hb|t>z%}%Wg#=TO9zp*5}O-pW$
zZ(QUkKjq=v{kL8eSYDV>nXcp4={fa#XUk2QR-R3l)s;mb7s?TbuU_qyFTZ?IUcTC~
zZkNR8EqSGK0Y7%ktu8Nm(TCka=NDIAgk*usC66(_!uiu1@+XhTtIQ{GvGCItsQ5M;
za}LFeU)!6jWov_PwxyiSP*#-}El-$tU@?wkE&0}z`yGmtdPSU+S^W8SoXfWc8`CO-
zEJ7+;$st*!BNmBPlm-^EIP#EVK4}v;Qqp;P#xbk;TnJ^4U0y86vDinyhTVi&43nJF
zAG)BrNepsAai`9a_BQ|Gs4c=9*W-L5A>Rf^rj*b8vcQc#2@$b{c6}tR(B?Un`7!i~
z{?MoE^o`0Kb0g+SQ1M64w1dJ?Fa6y)$w|B5cFxat(w8=gwuL&}>N2u0=fqqIdS6*u
zENiRFUIg@REv1(^Aus4LZz3JTn|6YQaTdz6*vE0Ix+JUH3&)L=y1NV^LfSy*FALc!
zKT4Kx`Z*(UL^r^!(y=l(rTl4hcw_Hr(ymmzDU9uCNT&(7R~&ki2Z&s_3198jmtgpe
zAI8(UJ2Q5`F5#NlpwTrv8tM248GD6c*l}on<~AT}aWr@rkLYCxV_~_zQicKjJ7|aS
z)jZz~?;Fn$ex@*_?du2o8bCSw`Vn~H>$iXV@BEG{5e#(YCunX=l`ycWda?8G7QJXh
zOdi_SFo|tPYa3?8$(K@VTP1M#xT#k9$xXq|q!*PeG+0{WMq^|Qa^MW3${ROG-CG*L
zfNH*BD|)6@Y)hlUk8u7aJr39S+5rx~#zbz{1wud<TIgIsefh78Fx727w!{w|)^PmN
zMB~<FMI$Z_0TXrdt})enfKZJu9&De@z%JC}9}hx0XAd82mzA{@H@4nIrH(P@2}0Q#
z7e83))h%+d$f5@z55XH225i^E7Q1M?SkNBwa{J6&*$L~1GW_(w!^|RCCfehtl>5qJ
zMvrG=U9@ztg>$^(dwIhy8s4bMxymelvCE6Mj_4$^0Cu5{A8)VlFy|pdN4MoE5VhhC
z#`^1g(W!i_KWmp~rNB>h*=s!*=10rS%A)z>Hp+RE#SzjEKlAaT1$+}89B-*nZs=?y
zGaS&5J2rCpEjIChC3$VW3T&RTBRdY8a`SM)j!fxXBt2Z@7in<gZu#&Ru7*J}Wbq=6
zXj-^VUrwkvcrXRuoG0%Gc$LkIChnB-fEVaNE??7oAR{{B1w<u(JozVHIuvb7c;b+M
zjn`Y)ud{=T@*%Z2u<7a#mdU%|tF|Sh%9hF#43!P`qq9lJg0_Lhro{yoi1|Jie>(i+
z7jOPhHZvHZ<3-2lt?H=tN0jJOWi|sT$tD@!GBA+5=A@1qTjiZ!q+NeL!0W{~`L{(&
z$}c)#Aj(3X(~bu^t~6X%>il170OIe1K_Yp%as0)?QjWbKeUv8;M&<iPBOdCp%bd~y
zbL^NNT<QReKy<%~$pBM%h7Kr?$gH$^N17BjC;iY7=WO{lfDMQrWk5jO$WvXWu}dcM
z#$R^h`GJ#<SMJ+Y2@*p0P-3!}U@%$Z&>;-wty@?%e?dpW-6h60izG)iLZh2aK|~?`
z0M42W0r51431y%U`vhfMAD{0P?`NSr8og!a)<tG6O^9kFUkQPp6HXZZCa~2w<H<`$
zp2ex^8(RqNZ0Tq?H%Wl(l|&#8@-QDG<)=7fO?)m2n5JoDxGTNU)y_2tZ**>Q*VH*o
zrBU1=Ycw{-+;+t+NzDfj<IuTE5_-av+g_ObOcwo;Hz>Td&BlVmDVO^!6&DoY8xG!h
z$pv0N!nb~d(@LJG9z|W?sDr`_2MzS5;}Sm40EivHez}ka<iL-g$l+}Q3~s}+W@Igc
zxcpBX=F2Zc+>FEC{JJ;=_N%nbN<7^l%YQ{qSY7Sdoc1ymw}14<|K#Vo@Kvpbq%Gjo
zS**w!@#1gyfEKt5H#CRP=-=g69Wg76(@)|t42x|@)7|;oKXz`Rqd>Mgk;^ck?{I=4
z-8epcad64s<sjgu2%LV+-tZI$!ZiPGTmZVW$K8!{Dx}|lEoSVY>D{-}7dv5wro9Bs
z-RRpD0F|ryyZV5rxDndWnC=dDGkwt(ROoP2>L7X?x<z?o5P#cx+RCzbgNRPrxB%$<
zTfQ}SyDH6)_s~AP0^(oWf>D2CI=8G7X!@)3gA7;mb9)kO{Plyr5_WXx?-s7oR&bMK
z=cwP17St$S<;1QY513AcA~nkxUC{%Vai(whKrs2)?Kot|<mWK<=-o!e-qQx+tbUX$
z%Zu6ks~_%mVN`q@X9ihGnsclkrr1_qhQ>ST1K;>IMDqvJ<9oUu-(SW%J5AWIvQ}C3
z_cI(k{=V$8-s2;{V!swY7n~Z@a7M(^(R3cM+gaVg>F%p<M*>5d(aU}F3XmI59aXYe
z;JVQ`c>iFpyxQRyxr4$^E%>w3;rJvA6pYK^&GCVr%oWRw&FP0Se$m)deDTY7j2C>&
zUnX#wsAW<;6Sp}fBYYSmF)8nFdCBcbdlr4>X39KA?}-nS(oFPnL>)&mGU?4xYl}<s
z8aps)or&u<>5m;+@9sbz!b_G+9?y8}2+&uc&CfHz$&*eP!|9S-$e82WB!}c@SCbb7
zM2bFmLD4TC-moJ|<LatscrnIe?8Y2}UmC4^A-Q4vr?DI9WapQ4CD}Yywh-*DnJEiS
z|Dw~t_>bKv!~-4q(_SWQi-p9?ya`9tafIF_$5|?Vhm)Khvx~-I#G{o8j+$c*#5>C*
z8*?J;RN{MCdwU1v`0T1&Xw1iTsPgc9XRmzue7C&V*)Mzhd^3(6N;m$d)yerqIXpfq
z`-dlGk7EW8bRC|R-J{F0b97c-?w^&t<BM{9{-#`N=y>&-#k5?PH*dTU#y7F{5808X
zaC^t)`HTJXYVW9=Uny<nez|@#RZcmAPvc_!&Ex!%6I^g>oGyJ#&&>NMLUyrT@v^V<
z!s#o#K*=|=c6JZStKB{EKPU%^=TI_n^rJ|#*mj(6(V;M2p1rzZvFvR*ljp^i#y!$K
zi#b#3<6NAbmzOVg$`{XGlou~vN@x2n8%v8SpBg{&03fe9CU0|nrLg0T*#z%u%mjk?
zFkiu92|lV(9-^mxOSyULMNY~cb0(A*Oukjl$uP6i_sh!4N?GK{LzNfm>9o@Cg)7R5
zbi$4;b{s8pw5eqC!bvRzEC6NEg@s-gl=3|<;$fj#B}Z|;w$A<H>YE$Y!<n+K^ssnE
zIX^l+b>1Bu9!aMv7iWC8?b-{?)Vs71Z){H;O}(Cx-ED#5Lk1QU5v;we4rH_|v1&a+
zHt5kAWk0;3qiv%u<Z^orUVJi-nE4kj+@7yd`rPh_7j!IqxsDP>d|{9_G~YAh`9xUa
zBH!|SP#&pHdl6IdLFf6GYJ#L5vZ%?%fo0{>+UkmTNwP4=F{3O7F3huNnKm#BU&4u&
z7m&$Y+C;uj$I+H;*O;P9dMyv-mFG^)*dc&NRs;HB;p<*mp5r@w_*rz*z^{_ojw}7S
zikInl@EhP;xag$$1t!zQBp>jaL_Z7%bf*IeM)}nGQgIH`{NX=<VLU*`XJ7|xug>6m
zKVY0s1MuU1k-Ep(bw_w$Jid3)4&keLz8l^*o*{hQFsA?d(R@v~od@p&EPUPjPygsY
zcbnr(^FUUn3_A0+H0jka5DD!=U+HQXBZH%}5n0l3gtrVxL02A>Q@1sPI84{^1{eUO
zH~~H71dyj?4$}c7RCvdoMsAdAkBJ0?=#(l%QX@;xr~+6J+!aPvds`>^ODOw09C4e}
zG_beR-q?EZWZN5I34<TB5Dv(aj)u}bZFs;YX=n{a(u2(nzAhRzjyU|_gU99Jg9q-2
zkxuUr(Ql+sxzXoArAwU*@2bHqyU(JQ-Me4%MAtuQC9c=%R3M`tCgMXrIb9IwFEJz@
zjk!FmKpI^x9)4AJ{onx!+w-xz$W^(RsOO<`&8`|f)HsrWH>TWi)y0FGBLp}~01@Zr
zc{7IHC~P0s!`A#!V1wkFV|5+^Xb0IlF%5KxTlwI*2mwaZcb`wBSr!q6J9dY%?&xd~
zFW*#Zx=9!;Y&mC3dL{wsSUV5gq0JF>egh1eng`&x8o&oPrX_aVCUib|p#`_{B4tK$
z%A}kmkMV~OKKLOHowk~<@^ikCIuYUT0Wk7XW|T)jD_4Z|VvEeSh>0Bdv#5pH-o$YS
z2iav3ns9!rPw7AxDQ7J1kTN>j*X2+HgWcmpb$Cxb$Wu2+k@8lNc(~B1vnhR{1?*SK
z0}T_z4>{j@%lh)#N?BcBD;w+Ug~echw@c|+(0~P*DPsh8UdSKWNHd+SEcQ$}9PzL~
zjGj0ro}GsrLE-b_YZ%f)r;q^{W!5_rl`b!CDNY71bf#Iv;y6clzTEK=nd0}iywq`J
zw-&nK<GRNhV4NQ3XwRBV!~D`*S>0SNODpXt4ze8`^Y$TinZa8IaXvDWJmEs`8YoQb
zB9B<8b4QV%f>UnpX>dXrUR_u&A3S<eHq^%O))na#Be#j!aVlYCr!Lu#JT_Dz@x}nn
zWppgvVnmOsrnt(F1usf|@)k5Q8PP%Y@!j<qPb>~8z-oHkln*VeX|Z?yF(8R0nC2ek
z9lDXE$__3iGVV!nH0b<*Mr7X7iwv?c!K$9|u6`|Et3qz_tpW6~Gsh-;^RG?^hl%3_
zQ9VqSGT?<NTuo5_TTa+1fVR>pF5BXPmuZ<;cYmTHP2VvwMZ(lL$a8Z_0766`jYKz0
zkTxI^m|RM0EC&pt2NPEuF`!^<D&w}*u(Ug*(&XR;v|(o3-$18?$r@J;;dny>scnQe
zy7IL=YGW&&7$M9{5)GiVkG$&LyaUY!|LWc0IuLzi;8)2i5AF$;`)#-j9|)I9(MWZX
z9zS=m$5;NuLOZJTD(w|(_>cbHA2e3~*5QBlzx};o)E%UrRItuQ$?QMCjnG$UxR%Bt
zOri(CkNY%Q-?0zEPU3S5CwIdPW#MkyEq+`ncg@S$*TOZN?V~uL(Kf$|h+Po8a4G{`
z-=qHcyM$LB0OeIqyGP`|!`&b+*TfGFAxl5W4-FYDY#*l#^EA5p%eS%}#?{O|${I+}
z`pf2t#m?Rz>o9JozPd*M=EB>ri)V=Ikj9~_X17t}w4Ao}aXHOF*-d{h(KY`H#szPC
zLqJU5t&$l`N7+ee?upON5&jLrCVeabpP_%EYi>Dk(}|HKc<kX5-u4Qvx{;v=6V|Ym
zuz`82ziGb!{a%987X=6bzc?G6v_g57vxTJ`4g2X1HQ?LZ`Y?@qPunhNYzSo0pZs%f
z+Nr*g;<LRcjBbN8t;>g#qhF0G<mp=(@E0=q7=RC8R<>z;;QlFn>-dKcW-CY3H4uK3
zRzDEw5e;?I^AD3|x{Ox60<FCmrPl>bWoJL3oJJ6)flut=MVM*|QP~NwwQ?os=qp%(
zgo(4GtBA&@4-*S+CXa_O!gnI{^ob|#q)#%-N$q>BjAYU3gQp*tZ-4uH<>8}`y<3Jw
zfs0FxLo}A)csNc$VeG*-iq6l@J;q_Nh4R4!_B<1{Ow4MWg4ixM(l=v+oC2bJQ`|n<
zO_v>#!Se&oOX@(5oX-n{S*YSAIwpIWi1$l_>>?3w7Im0#W=!OT0VRyFPhM)RV_(KQ
zEC?a1cNwwMMq@qEx_)Uq^IBs<$79|(e%=$;)t|P4bQ1@r=MJj8Tp9n|*VrfN>t1lI
zI`OxoC9hv-yt($`-1^!^S!1V;{J^u|gkDKMF9oixt(IBIa(;1E_74ucXu~3oBOpG-
zMOfx$_%@LBEq}&m>_oe~gg)(1q_oQ1N9ETR(a<qt*2SglU_#}ur6U&3IJv_wC@#!v
z-2A4ToSc^<$$5T$i4^`$mX8XYpHUe>SL6xfZql#F&<F8IK6H^q$a%z<Fj)X1O?=DD
zFC$A29D~S$92fF3<|Q0+I2Y0dpQL!B{A%BCVgZ7M%EjgRva-7D#Yz%Oxn!(6UDL>k
zDu>b|b3LT*{`9@Fw2+0M6{TTqnFWXir)zfhUU~Xxt31}lBGKu&^5;zQF}_xLyW+H%
zo2&9x<Jq0PgYx<FowB>XU#>K+URhcw-~IN7<;kOmUU<7vxqGm&S{^;zR5^KD4i1mY
z%l+f>V)w`kgeT(bJXJ+t(UR|d-Q1`iC~fRCoROU5BaP*p<Xc-^b6q$+<y(DcWq#&<
zdGh3;%FCJ;FHcyEyn0>E6z|Ci_4G`7eB(uyMaeen1-at-dn$kC4_J6<#}ML4{!rIA
zoy7~Hyfm)*I;{qNb9KIa`x}qTcR%^ai*8>&+bRF%FQ1j4eECYcI4q~4rE2h<!0nBV
zvbD9Ig_?V>z3|6)pB-tWnW+cTd#QTqhCuq^B%VX*=;-LwI}tsvq4ZN%DYFZzJ1l^+
zJBr1%GrpImyv{rsb5SaPPDRQO?$*ZA9-(Q<3oqR+D*scy7v!h%eVk)NeJYM<X?IwB
zX1Csl((wn6x1^_OiOEi`;+=yWTggK0k@Dm6QnEUx)G;14%q~$M<SHvFV^6lXRTj3q
z`)$^XcT<j!9hfX0pI*o=eURlGnRs92n|Z0sqosXho`>)KozM;|e%cSx=x_U551I3!
zZaL-B9phzQDkfd%WMTSEdAv4LRu`vv5<Ab~9daBo(U-?wT!xIffS95gxJO@H9-4hP
zz&r(Q%sljq9lp|^5`f$7b-X9gx`$;u>80)@$46;_+hGGOyzQ91m)()WF5`wU60SnR
zq)$cK2|t7(ZD1JIciVxcqpyP7JbQSTUxRnkkbtp72wFbAFF1~`7vyJvzD#}{_`CnR
z|M|W9zx~_)gWpl;Dx_{T=U|v$^Qv}t=wxeQ=oL2{_ZY^T2X08XnKy+S)yBhYd{LP&
z=~O_CfACmCqww~#NAOtfh9msA@b<0RtC$Yvy9ychR*`Ch(1pj}F8cu;-%5j{$2Gjn
zh-g1Nnq7u0t+8lV<K18MAv${e%tU`i54a^4QhVapbfL5jx^@{Ch}Kg<=_JzlGht5W
zb#rsGJlKBV4iOI={L=7N9tupO-i6i01KNvTcEcT>>6mt%F!EY{Z|D@L1H<=@FnMTY
zS5`V*97jk;lZPAM7~<iO4sw0C5ufaIGNGgM>fL2(@Cn0W54(EUK||-!-|)KNJ6^hY
zD@t+s<FO|v>JS~eV@$`M4!4hNkRF^@))9?%bk)jQ^^jGMI_o0-nRt&k*v7#C9q9+m
zkG!jaZ9s1i+w+i0<4Gsad0gWr{lRB&kZ^Q9qZ>t}zw!rN*ZgSuZ!)-?C|HXJJa#z}
zPBK_RP{idZh@T7k%0n)f4_)ZPVQZd)#~*qirIo^Li-3g_rGfPEvrNQAhdFlSahkxp
z{a}GhCSjl#Z-<e)=_<)cIH$SB4@jzVEj;CkMI{!HU`ktH{pyb%oksmy*+qvpDr-r%
zI;QG8EiJF8lQ^w(l7#HY^1x2znzDi{lqJ92$XitE?68<cd3Bm9&uf)Ozm23=6@Ahw
zI^M$an~e%jSTAmhf_x&)$bwwfKm3eWc;U<o2NpvX45=4%Hm|QmLpn(+K>1QfdUigG
zY+N)=lmkm8-d==sJOs)Bj5OMxzhx3V7xAf%l0OW5Xj`ZoGt$@M{GvOtK1xLth)jrD
z3esLyoYT-`sF#0O8#sUvl#lB6E)c8^iwX2~40?jO&NiGOZZDuoCw>l89~0^C<k{6M
z!2<-VR?{`RZT>qp5?h1SUv(Q(WwL*6cAIwsRX?oxIsZVwq2rB7Ju%i7GIOUs;~`tm
zx56;a{PEL0XuJKhe&W|b3uoQLFZ_D9|J8|~vCYdWZDC}!+qX&!5{D6SRTP#+gXta6
zm>B+58j38Y{iP3?sr!`*vt9)DulhB-d8r+;9$g*cACKx^>4v<Ad;-1M^1+uFCqLs4
zt}P03Q_T}@NS{pS*EG#782yy!P7V!XscUh|R`p@KjHWl!b|!w|-P{0!=kTt0@oxHb
zSnDtTIGenttyh!FxWSKqbP#WY1ZZ*N5BS-mA8+FR!=K}huV4So-<tT1uc5o8aR_$5
z2hBLH!CihseDfW;Z6Aai(l@x}cKbd7KV|~qXPYm|Mj-AH9E6m!4%cLff1(x6t{(Ed
ziBFTscK!JK8UQ4qTDS&xxp{s$9kQo=>*>r3Kb14rG1*)BY%rvUXOnRV;n82MeH#ML
z{}xtMRhFJtA2!h3?HZ=fIQ`(+AUNU~^5Di*>1+?bcm@}DbyR412sC<R4n$AJ+l4*U
ztsmC_d%GH7?aLDGgz0fK3c|@v{x#jT_-#UhT{rW#S7@SxTn?w-h+ahF_=Y&><ZM^(
zhY2l4Xbh%_nKq1Tm<}#*%^eu#qjgwip>xqf*^M4e1wePVJF>@>_=a(e>JAVJLbwD0
z$n8=ql;t&1ML-G4yyNZxTpM4dt_}S|59fsK;~dh#KUa@$e&G>6V1Tb_3slpF;({NJ
zfj6WM(;>mhWf|@A<vZh;3N3zeD1NvxV<!v_=San433c4T1)6{K5@<AcT>&gD?r`rp
zAZPT!f($QzGFeWO%&rkHcGMj(Ow6)~<g%xq$CUDSUtd7K1RRST?CfHP7CRPZ+1;UW
z5|hh(*LQVwrL3&2W|xX5otHe$@Cgy(<GBLH^^6@EYj~U?86_vX#+a~%ciZj7xT?uO
zT3E31&Iyg_NEc(HYvvjxdtR2TFGo^Vl}_dsE-%hC9=h~`2rnOkyAqB`b$`R_?8M*m
z;^jumSJRCytUtxw@;CZaG>q%&xUp%R5%rD6cwEeJu&BWJhYQ_#@kjB~mmv+StIN{U
zs`a$AB)w^@$pR(vXI`Z)+R^usEbLriQGi_}jE|Ws@psb{Pma}NyvAuD(%pi_UyI9S
zd0BBNPgtNsxBae?S<B9a?)*I}v1oQo=1@*fPs_1){IF9i-w!)D&~+%ionI@BjImj~
z$}S(Yz>#<-CugENEc*xh<>>Ip3u<2cJ3TMw%mXoTFIv=heUpX3lcOUqevt=Kw91<D
zS^8L5A}tGExM3mvwdm;wa>O9t7voz~=zw^=kSRT|*n6V{9V@;A$-KLNP<9TE%Gssz
zM}90iP{!D~#p2VP^f@=n81|kQofrbMbB7&>bpAaa*N^=7X*C+}G6!(X(UH;tX=f)G
zvaYK9udc3I&vOebNGX5WjU`^Ri!A!Vi!lDSm&$^Bw<@#DRWK(o&8{<H&o9vBX*ppb
zL~?OmQjS%hFRw0CZdsgWKJCi$U{}`{ot%u_iDN}^uc{od6N^PuS|#Rhn8%>bu;_TM
z`anCt{267Nw6MsBJhT%R7gt`WqD)+hKHtq!*}Enec>$PHX4sJ^IV2PG9E-?{$IOpV
zUy<d$>#gEf9Nw*`%G92#t)Al=dDhB~kISp#p=V}V{z7=}E5znUJv4u%jRNL4lxKcn
zS9aQ*{~vRI)@|96<@rI&o74Bln|mWNGPAN$8YR()rlbc%=?kFBroMrAf@gSP=mDbO
zNemb;h6YX50Hkk%vZ)e$02l)hmC7bEZp6Lre*E;EyYv73|2EgMz4kuG-J>canS0wZ
zvuV?&O`Fy%Yu2)b;nOT^A|VfJ7rQDu%yaksZB@kmeEKdYsS9lxx0p#k&s6twDiH04
z%`y5?+6C(q-gw^RD9Y4dvV+e22^MtO)@uClvY68LDJR-Ee0|SS9LkfjbH6um(ofnB
zkiIE-K-U0udUsRh&pKLEXga`l^Mi>$VFZMqLxO2xMOS{H=&(n+!t*a&vw^s~IM~Jn
zR?$xNL9i4a=-oj1xh~Olr1%rwAq5k@#aRU{UeWe(3YNo`!@mR`qx%vbUz$!2UmM5t
zU!TlZ!m^IP1aj2j@BP6~6<rL3kztucZOhP5zK-l0VSYGUtD{SB7y}L2V}uMm8u4U{
z-a^xfa1pL~14Re_5h(zV5st3SPnCl*eQ=IuPi`#O^kEjugy6A4n~NyPH#81!ytz4I
zk9c#Gr}6Nm?2?5zv5msrbZSK7XRsRr@q<X|p$~Hc&1IA;rm?g`<Z%O|pG)*!ogyCJ
z2-GP941`|4{$}|0+iyHbrDLYU^(Kn#MDz^>ziv(+KB{9@NA4$m-05e)myL%EVtIqb
zn^?XX(wVFCSEtTRl>wZ;zonb__03fVg&v4P&jLNK?B~fV($&T3aMS?2ywN3IyeY%r
z5jXsJg6h0INd=!v4GtKzTxzhxfQ^9{PXTc==5^xzgTrBKV>{n7lAcmt(IRD@oa4uZ
z-Z}`Giw#|yqPYPNh{Tf{Y^y0_95|y^dn*EQv6NpwU1q)aGGNJ81*X&Yri)PAj(s!K
z&znx9C!WOM=36&J<N(Rwf&n@MmGBacYz77{D;YU0gEt->UxNmfYX(LdD1ujbOa}av
zEp*nO{?4PqNn1He=9DDC48Q>BaU(N%@J(8>xOFKH46<rNvW{%T^4m?mfu~H+urrJY
zr+E5&c+v5F8a_^IBYDhAe0jUcZzsx(a-w$SM<iy)RYM2c<dcBsYdKb6LlX`r56Dj#
z10v4KW@7e4ZHOn?7))kEkcr4g*9j)euG8dFiMm}WzMmRYn_?ojqX`N}kAfrrYzk6t
z45}iF<W(GuLBeLhmb@X0+Y6h*N}mDw?G0^`yebaStVfO}VB?G9HaLr0X(6M2E<cq)
zCZ=qVv!R!A0jPgW_CtrEe^A*f?;afDW??{Cr7ZHql?HFmIZK=}b*b5Kb-lsof3Twq
zW(r4a3}D*@)#_Z6Yl;s=)ujYuJJf0649!U2@WgRNLwLs#IEc6vZfvnAZ&;K8j0kV^
zWLlY$$Lt{yGL-t+A9PY9kGO+r84+N^Gv!c`DtP4M$U%lGbdJg|%;f@MjQtRpERk$_
z-o&G<$3U5iJMx6aJc%<6Siyyr>Os1LF}6wE-i;Q>0gsk;V-dc1986phQ1`mAMvEdy
zdWh@$tNe~@(ngHSb=O4lg9-;|U2pM#s$lAIqM8On@JOPI-4HdJAap5z3y-Vug9zS<
ztiRfl@w&7;a&6HKxScm_E_GD^)XIR}{M6&RJe-D1N*aGcBUke`99pX~?CXkuQLNp>
z%UB5;KQZl1c*T=l^#KycbXaU-_;<Mc>~UGDcNzZq|Ly<&>pSZ3zw_`nzARsbb^f9%
zoGw8K`ZzuOs(1`lrU$6UIL)5(DE}M}fE&Y`UzagoU?-Z%WI<bo_S8C~;DFwq<0pTm
zy97c{GPPgDk?n=f2GDU|E13I7H~bdjwJ_?ktMinGd?rZ!6`nA=7opOsaNh$KAo0m>
zAH8G}HnziyAnDpKGrAj8T74c1H>cA@iN*Fk#BTB&I!oKa3kQTx{N5;NruvcMO?>(6
zC&aA45iVUA-}F;}d+Zkkrjd85BXRd}!Zc2x!W>t>st>pWUAWAotN0UVqAmR!M>uvX
z$}w)EC$4Eac`L4YH#$>FuEYn|hb4~7B0ignU#yL7xImi^G{0Py!*Wx+P1hI?M}#Jc
zMsA3%D*rJfKiPK6q-=9R9>ox^DKD6jzTx}&Kg!k721M8J;9Txww4w=&;qf(l%BP_V
zc3m~S%$Qyjv8ql*K08Mn9#!UJ*(XG%35vHO31Bm_b{O|l(raPu=YOSHVY)2H^#wp#
zz@K#7{o+>sES-3fSLPcRmu!BWYo0<|<PDIVDB%Tk&9zCFjT$dj`wb2r=vxyn#`)+W
z^Rq0_=V3R1`G+3}6c086F!!3zFqdEfx*lQ$2b~{Iy}t3B{^F8NsLb7YgN3>Kg$#1C
zVRG{Ry>a+4J}#bx{wp4Oz4imH%<-A~)VP>=59!@09p1`et%3CfuQ%MXfPX3b%9~&u
zUC5hM)K%tY%z-KA%*E*A1g-C4;hB89coaaGH$td?(lL4l2M^LG4h#P48*B;<?QJdO
zWs~jBX^wq@Xp=AL<l=0&;M5Y?UQ1!!LwVqvRIJ+^9v%8PMdG=vZBjq^i~}2J%vpK6
zj&ClRzvefrBJgk}Z{{#Z^D&c}dv1x>{@%XdEIQcVAC3+WhQous_#2MRo86sVr^|Pv
zm_zRG?GCSBy%>J<?JIwaYSRx>rj4L$>(^^LC`<ZLUh{^S4D;J3IsX6V^1=^?GB@`d
zh{&e-_U_)c(_rIZL%QdiTBNbFy*?Zr><mA8edO<PF)v3q95J^iS@-r2yy1lIsjJ?w
zmCU=s@9gb))0R_D_I5Uh!`-c6XN$Mtp7{+(Hu=yg#~1RAt?ez9pTeX>&uuwck++JN
z^Ixj&v9ZRRL@86{Rpm%KLB~F_Pw|)svjM^bqH20>6O=pO*5ZL#_^ojaDCNe6oaW}_
zlleVwv++hCZ*b8rs2@D!%|;WOCtgeV$eUZNdz}xL-psvnduE*sKsU(aBS~BR_4=Lb
zAL|RO17*YW_BL&SZ+)Rp-jZZJf{iTl>$MJw6XW(m?PKjgbzX*t%+Jm*Z7wfw6i<2)
zUy(Wg=$4II(js2!rUdf8vjxb3KD}8*y2wnLtm8Bp2*alDHZSm^6Onl!AGnkq%wwc}
zl||OGV6Y**xUOxp;ivLeyJp>pvY{O7C|7vYn@_a&dT@F8^vrL%o~a)@Ic5Fv)W>7e
zx3L^oYe(0tWnSJ5=kDvyy|GOHc6F(~3Z00SI>#nlHow$nCGrOAP)f%#h{$~hKj$lX
z^Sw&M9C0{qd`Zq^8CSELKVeERK?32xo3`xGnrI9w9Cmb3{_!4)t{YSsP`pN*>|g>D
z--#|14Yv?JLGNF1W9DJa8~ha3!$bTE3TNq7gcm%P2Jl$i9zLJO<M5^F^zbXiA+KK_
z?JMAMeLn;~>hOQ}PyR3J3dDd;K|HYP7V3@zbYo7ZGG@e{fICdbQAg0wUI4+5$CF<=
zR$s$UgdK(aG`e(9U=e6M=E&qtCV`*u_g`^A77OupvOo(j+N1|R{B-aka~M4$ZC2J2
zZ|K}`WlClx;&92#4I{EbK}1vFGWZ7W4k2;r@a%@Z#RFoa1JWT;IH1>XBDRk{tPCoG
zZw8V#nD~jF)Ez@{bc3)@zzskIZ2r7{{d)NJ+i!=%!$S{<NZ&UR$Vi`fGHe=n!Cf~z
zKUt&>nVYL`mhyX{9piy416>A?8PwV^z*HDF77^=)C?4&`g}mr5PcU%wfJ@B!Vo=H1
z!EBBoJx_%Z&x0t*0Hhk_cmiGtCOvTXN^g4Vgu%hJZ#Lxh#nG{!fb#c=bdx|^eiKCE
zA)@tePr3OctLV^61vMZK!H7`64Hp>z%DTc+R-tz;&5BF#=mFc8v5etC+9@~x6i)uU
zQKc{pe#Oq9LW3jHM;9NZD?AKVC<itU%}09XC^{b>p|}i4>!y#)lqZS9#{(Mqxry<M
zkPOyR@0i$-8K0x90Tq$ZfxjoGqGLkz{Fw#}v?1a+K6+xX>%Rv<K%Q*-pt8@TDx^>(
z4iG%L0w;;AD|9Z~H*N+MlIzp`aHTxI)8OOX={rvvo@x-l_qmuP@und%!iIWF+HCN!
zDV8?PU`TCO+_yE*Kl=6?PXf02@T&$xH@BCrSI9>DVGs!*KW*ca+lh1|S<nM|^`u3H
z0U=KvTDP<-#o_Hk2I>sZdFqOMQZ@`0nH(Szb>C&|4d>_rSvd~CPm3uJJoWYAM(yzC
zUIQ-;kWV$4xR73zM|T9u5QDe%?dQYpfhH6CyFT`iKF;mh@_cgrVe*PxOr*y;U>#Ep
zBmjUMlF#RMr>Sa?rM9de_OaUBH!oig`#Zcbm-a^8bE(3@<tHD1In~Tsfc!0gc>uu}
z;{!q{-1y=lzHoNGYFZxSNIgwd@Q}-s;JWx)v~HKU@D!cw(TfAO(^aOMjM*G)5wIL1
z;S)|5ydB`Jti)5%aUgT!4YG)cU|y|I@d7P=g@s)4MPHj2)pqDt#&*ONo{f(usN~Q8
zgd2*4wu2?Pu%R9Ek6SL|1ygLth|?fHY4&Kjr_#$681o;G(aKG$b5P<C<h(<f#C2>g
z3J#Bnr_s6uO?>In@_=ah(VzYdJ=G)&B=-|QtwpTOK-Ew26-U0XfRese9wcGWhju|y
zNSxz7`N^mFE@|;cmO4JylO)m<Z_Cl>;a|t(3Ngx>@+7m7gX}nlXq^W95#2IT{t4%A
z7DOjlhWI0dM3A(7q#p<iVh4c)#5PWxMF%G0p&w);Tv(GjAad)UNfLT+paAKRwfUSb
zah?qS%m4NN)|`GF@ZbL*{l^limD4C&le>Z2PlFy0IBs(r<e$^@VEUF;xeHdsE8G~;
z;^D8jzOTs?C&D~F#nWhd-sP_N3t_&HhP@(cHn>;@(4%a!D;qs8`EqTjo{kpQlX4#d
zN;d^|3+JlM{iRDM{DN5Wmm$7PP1da|LGA?(%BAC*hz-#}pLRhncOnKt5<O-<5H|qn
z$O1fm*vL*;$rd8FV?lT*k0orgkfHKj(8~ynaXUKW^n2RUt|%FoHaC}uV}QQnFxsoq
zgJCv2;_qp|7B2k&iA;Wjtje45GuaD@PelVJoKtcTUT_^7<I2xi4oMuZ0XTPEfY96v
z?A0Y?p*3H-y0&eM+~lozh~LD$2jVnm@l@W^zcvusvH6F!%@}Tku3i0^^4jE*i+dye
zc`r%sMY=dW;y0ZOHRT0#;&4=1n#fc+kQXwuY2?H$XzUe*JD`3ojB-sKh~ILKVH2*m
zt1e7<TaA7ZSM`|~;9MTED^H}Ge1>nuGm%XodX1lS=Q0>tl*)^(mWSwdf`xc1pbt+<
zrETF`Sdp-J%8w#obiBE>KO7vsmU*eU(T?U#Pc^&aZJ{-9ny>)R`~yEMuCCd{(p*A$
zVr;^O3Y$61KS=vta~8fy<Ofrk|KHs!6_sOC2@hLl?I6Zt>EijR<{spLcEs^}Y*cu|
zP4hwO9p%9~0!Onk7IS@&@zL(``+-}{ft*)g@&KO~{%>-09%VC@rraS_y0kaNA&kXr
z=BIgcf=xZ<Hq0wUFMhm@!NYva`MmDIyi9mx@dK|eZ%J#pK5CwsO)oZ%j*nlMFUPI%
zb`$f`PnxS*$1JQXUG6=+B}RS-XG6|&Ny){!1@mb(B3_8sPk!?^hohq-A5nLDDxSQx
z$L7oJUFP1kK#n}hC*P0ejU%RO-VD^dkZ=r}SpLosZzpXlf1+jH&e468HPbidhnJf7
zCM`C3q*KDBi=EwFpB%*f*GJp29%Pe;mRXNsGle?B+e3Gv^Tvn5$tMpgvpGrqV8ek;
zDK__bbLwDkd)VDt_fdV+4d&Nu@KAodE%nXomp;nTZzNqxzKd(W<#lm+BYmhoDt*d?
zwFTZ9+mo!sTVH!RY;8Op_BPjseQ$a_QM{Ys%?agmNi`pS^3Cz^@@UUx@b=y5@ZB%p
z4d1^xQ6BI7rsKxuy65Q^(g!2-&2^Rc+PcDcs8sb=?mJa6+75M%jWx=Px8>fyKlR!J
z^K7r5NY4O%zQ4V-PQ75Q<dgDtqk41UI2;v7TVP!8N=yA!`*3^EkB7q9;H4ROqgwba
zwbMiC^5u&cu8Uh+lAVnm<*U}}c=PD;iW8D}YwS+*_Ko4Uf8*8gJHPcC!_M}Ozwh?v
zKl^_8v!8u8eE;^u8(O64{7U|9<(ahrZ#-G%J8$H8ok(@b8?8zg-P5kvL}R0yx`BRK
z-*7vVaeMMQk?M)phO*X3T|zeMbG3IH*9MYNJ3(iFQ2xFeZ9#n$GP2Q2+kl$7NIb81
zDDRwrvA?r9{Pu7Dc=$WN{S)bFU!l^d+BbFe{PI@1;lX|NbJRiAXV#NgJ7M#Z{*d%o
zU*chWjtxy-J`H@=@YV6b@apBU+V8G-+zoHuUkvX~p}TNfq8|Z}m2VurINl##A0PYp
zM9P%6Ca+Y-nIp0$iQZ`QehW{P%<HvO2q}cRxV!meINIdM!{~?l?D-{iM)AR-XK*gF
z4&qNA3zQAclSr(Eu-=W{o4*Z==6tYSKj(P6(~tJ|Pmk_N7nwoK_yWBfSmfD-$vKAE
z?!p?(c>!}=??=8>aS9$vr=Z^;d<;IH#%g#h44ncGk$jnCNu$OiPVY-2{z@>_&98wR
zb@+RK@Y665rty%^KW`-I$DgbuA5=R1>`9pZ%`SRXfO6YU?e^dCu!)<6)C9!B-x{TT
z!8_a+Dn9XAop&T06-UtU{M<bJpJdK}H{ny-G$t7_^@sIjhfVl~nmja_?bEe|`KmbN
z+T3;VLTLMWQpew<TpYu;1Gu*Mm6sN!m4WQ8lX4&lDa0yq@8OgaZfGg!!f=utasy;w
z^77?NZ+vm{qqFwnGB+C&=`U6ur-G*=#+Knm;{mG%baiGn8!+66u=Dg*2EA8Tc_M`;
zRO)RkZq|OHOXgN|eyT|~^J`uSHjg_$$>lSPuX9`>^xW*v{6xg5OrG}Qi5H$)<JgAx
zr^17i;mI(+65jB{ryLKJSMt5Bo3{tG=v6nU1$`N}=uRjO`Tdh}TKxdHIi&swH`W^s
zK2jVm8~u&TCM5wtY-$OHO~2{BV~o>x8Q3q3(xF~)bM)q+jLSu3!l1wdc*WtSPNI~T
zpJZ_vW#9n#2m`+rlnpe_$!3Cb=Z~whu_(XX`LJr9-19lm3{Zv3ncZv_A;Y}~+4YKc
z<d7K3j0YT&8GYB&U+4<?{3e=plzbsKd2~F=2O8@WJrEfBiff(?Pki3G%I0a}>B5eh
zH?J-^0#kJoU0FA(-8>Cbo1bPQEE5`+tq_u2Z6dn&YwVf;@{v}=U~|_4HSx}-El=LS
zK|Cyyc*;#X61n*>>2kSg@PbbLM30DQ^Zt=1At+1YGLhq(V@#}26YW!h)-gYG7Crj%
z+g2n@`Kyi*pS)6Lq{Cp535y%GfIK|k;4F5Yf@3hq(`czz0CJEXbpwE33Kfric6xxS
zAL@4bLW5|8N)Gzc4VB3rZ?b7X>~fYa!a+d1au`|~x?mNhHxA?k3;QNQz#ierH#9BG
zZo4P}0fk{E;oSm_^ICk1m~@(=fT9!Nbbo9b5EYpAS+wR+dH^T7>c_{0dqIm8uhE?!
zFrI+)1<=5jIukh?4+04wQFIJm1ccj-OMl9>!r^O*5%0e*rjJ7w{AIOhK;r4zgm8EX
zBpKp{D!<$d6LYj22_DESH)$0Wez~#8s~rw75f1;bn~<gnS^5kAEHii+-`oUPjU7%F
zVjP`ZBez`EM+5Mthnbv^xO;h?mg{s?_op&2%26EE{TUuGA>cj{7X-%WLO5xQHxbMv
zX%uT5Rslf`kF3t0+$K@3Xm_SDG8wkhbRlGMM?TX9U(TZCANquW3?6CBr)^?6JVY|i
z>;%TevMjir?YJ@YRd$kjKzuXyqE95CWSVSb_-Ft5ukVeo|JJ|z-}A```ICNtvO<2K
zbZfvp^Y`@%zjbS8hi_m^(agdh!xg^w7iEuI<p<{B8Q-knIm6lAj1a*W604&t<d@+v
z(JxTpMbrE916Osc{GkQxPlLp@U~zxyX$gpbB<TDp&(zc4%LIBGysiuUgD>MTOXm(9
z29SRYx6k}W_egI<EGHPd!c5Dr!vS-DmH7g$limH3!1J1xcKW9>O*eKXJaJ=i7B67z
zKO2;P0oVK<`v_g~*qCplG5=<l+q#>0lqo3vE#jy{@;5w9yz*^Y-3Z)k;npv80k3FK
z27Z$lPi%0|$!!|)5&vi`>kQWnu+q(6cJXrj`ZazFIMV>ZSACoCE>`(5O%ILM*O4Ad
z5aL0Y>3crRS5QI=10%ih0}(z`oQj7Hzk*7y$~@y=0v7W;!IY;tEaLlk6(??B;w`Yv
zVxpXVJ!trfQviAjpFxkpJ^JIAu4P(&AAm{+Q{z+P`E_A^xc#DG-gc$A_0{E><|P*z
z(`RFX%@*Dg!T9@L%=J7U;w=r$Puc=C^FQRnojH6Se8mv2;`7iS3(IVLIG%XXp1k0%
zIY{0b)7ZuHLZu@U8a7|Rw>hKIPr8|(uwd;uj^tqB+;d{{XHJY9q<KAj^gIgbRnCo<
zH?^3niNAlUU-+rFPFLYReMB;{NbgfSq_@bO4Je<q5{3FrI?K(xn|ZM1&|H~0wKsQo
z6DX&9_+dt+#Jtd-X}IFl8rC<?hRcgH;dqdj??+47V(*P7r4PX03-#hhKG`tAuwljK
z1e=x2hm#k{Be|euZ3dY*7VvO?Pkq?O;Z$kzW)mA2Y<S#SR_3br3eVgw8((sNQduiM
zY?z(C=M<gy!}X=+)+%GxYFL~3;fFWF&wu{Q;TOOBemG%~{*oj2IJPaavHrp_fShc?
zDJg!~R61e30h^5{zKg^+qrQ3dVmLn7A9gp_hQpop;afHVjt=BsA3okmhIj9*2g;H<
z#G6rj+gpB1h9co;%gggq&Ht~gKi(Q+(}AN4Az+Q^z4GzH4{wK`fA{_HFaG?mhX3Nv
zem49UfBv)K=f8aCwJGZHmgdJ>nlFF+l#Nr$lQ*e+B;fU(*8r%uv?uBbb1#m6e0Op&
zygfM|PNBV!ey^GL=Tsxs4_F)Hp>65}ZHF}-<3*o0xl+E&N9l|FnD<LC0095=Nkl<Z
zhL(OQ`?NFaydU&&ZA#m~i4VTK;g~i;Tf9|S<$Hd_Cm(#HYfJjw+2XCgya~n|MXY)F
z7(TTdHhA)2hqyYEPbKR|$;pwD$fuv{XSH7>PHQ9H-L3Ma`VT+aY92~Yetoni?L;~u
zALJ7Q&_-~QNBShc8L2qVGjsInC$Jr(0X&cgLtA${2v4_%il=lvc6Sq#2IWRil*`Gf
z+M(ZS`{>WKux3Ll{QR9RwHemEyjdqcSwG`pW{$gjKJ2TGb4;Ymi*l9xY?AR77#p+J
zoBN)~mOhvMOS%w-O*ZNT<;276+p60~Y71Mc+q6l))u{Hqze}4^TVgX0Udp@sBgy&f
z$w$R|=I?wxmAl3?buSQ!fc`6WH~kp6xP9+$?j?H(w@3@Dj5pfww|5cG-~ROV5b-O2
z(N1*Bc0upQwJI(U_v9~_^8=RSO@KeLu7=0baoEFA7hs-l55H<yAngyTxp@#2Sh9Pp
z+5Cn3n}wFnSAyU8`ak@W|1{kOa?%jsqmGN+;tcE|RlpdAzla-%gUQ_8Aq=Qs(yK(7
zX3?U0($paS=D{5<3P@N1G<5LcT>(iCpbs-tMl2VDz2*<|;2ghPjX?=wTYi&F%<+!A
z@kJop#l=)61aO4DWHX)J?z&~Q06GrmOVH9PUUbOG2c11G_z~Kuf{v(mLr)3Pv4<Zw
z0iNc0@$$v+^5rYPvd+^Fj-qh7+f-PQb0cEFrcRg+*iS~u{?G=+-Y8)}tD6M=G7Nt5
zHjO%Lg|paB=S^o%e88Iqx}or76N59-$PNCJ1~1QjvnFr(&QZJZiH4);fEv_tQ$%=f
z>}+aOF>&MKri_fp#B1)aUcDM#zIf$Vm+3&!3Hl*Yy8P5!t5Z<CoPd$&UOGj$jmS26
zf~4IzTuv>rZ_e`b1O|HWKv#Z0EoYXZlmFu<`K@CO5~OD~qsT9P2$xSEC|~%>D}4}^
z0gDH{l#3AL(+-Eqz?Z?U2F@H^;lUO1h!4CeH)LT@;6b+rTJZGXMaK1#o4NQBCxaQK
z^DzSt&M0P(L*_FHOui{E4?;Cj@CSJ{*aeWCy5@<i=yQ`yeIq^dk((^jd3wOYAfIGi
z-l>C>dnP0d7RY8M81QG{OBfqn3`DRctv@9C-J4$y@7}%f;LDRN2F?omsDTKndf)|b
z6|mnXL^t}~sIJ~>u(hUv#o?=$!~WsHz=vB&5AilNz+vN<C$f;mZ*++so-RMvS(P~h
zRLV$jqroq99&}0eTa^(TR2(D1;GBUFY52(|#evR~I_d%3#M5!nhY;3>$^#illk)nc
z^sY6L@j25M8l;_G4mXz?r1999;voCBCINegJl(ghiOAN#X(#YyGf)oK3!cPsTckW(
zZW(-|0Lp{`AyWB?&-G2wkZA=Ym&$nK*@h-q2g6Uk{ju$5pX#oXvL-(T<xe|nJ}a6w
z0tG0=rhZ^6253LXptj>+{N*x>(73iWmVwAH(vW8332*8KKa<EZdtH*2+{lF^B<;<e
z7Q`h@LMB+_rzoa1$Pqsd(a9Ra8VD^Q02y&Y(HP;9*U2^1s99Or{7N5z-)*hMBae-2
z1gEb{G$G*9=%WuIngwTmp@e869{DWDoH*tkG66K$a0yLuvd}m}(2#@^zj!N_>VspO
zR%Eg*CNv-tDLCki!x*8zu+&@rPOI?}(7#d_9c~bHi&~2?w3ReGai)(o-2`bjArpbK
zL)-8h&;rY)izj=?L%4;Btfp$ThN4lpRYW|{P1oeHDx^6`OjFdxA90Ltj9MDzBZj!4
zGz*6V<|>@@0nTVB9a9##xGjU-6O85Mh_o+L+8ugxF&3O{{!TwK;Y6;HvmF1?uk($s
z|M7o3CT);!%u@?NQtP%aFG2ZxVdpfZ`&^-g?~F#U6i=`U2lRNmS{gl!G|gU!BX`xW
z9_BQ2SO_n>;joJ1NY^1(@l(wcKDA~Q%;T)|cQ|VSzuQo3z}2>+bCY#kIabIAs+Xw3
zI+78Pe~Q}^ghS7_cNZ+;a8LRe(_P>j7{g?z%>|4nsG%152bOU7m8`o8yj|R-w1x9S
z4J-1O@-zpTK78^6;X6U<K+z=}aGL{XKk>^BBy6GwdUrw6Ta7OuyxM%|LDEiQH*W)l
zkDd{kaPrbbgl6o=6KHpVvpwQ^o<$#s2$3ne$HBa!xr-ErCW;r&#Ab8|3g6>^9>1DC
z>9g@g*$AUF8H?oB!2LkAKr}7e9Y5&wlhb(_@#UAh<c6k&1txdm0ao4!y}lNq@m|TJ
zIT`}Xb#8&zqWj^nz&yP7qPu7<Fqe5T&xAL}2q1!-#|?mM@~a>|d2FCGHLrTUra2o&
z<ZW!r+R`|OW7;-_UzcBF{IxBW8}l#4RpIdsA~rR6fX)l;Ucjb)rI0+Dmfv$vrI)!2
zVXR58D4k>8cz6#qB4xuWr&~}zc#sGF8k{>m3+G-WX1qlf$d2XF)%=-9M;3Tw&P_c?
zJ0o4@OFn^u1dz*fUB!ipi^gbtvx+0mm?zi#+#5ooT7SlgG<o0<JtB+WvQfIE1I~GN
z-ZN)ooVYICazr3+(~)=L@Le&#*`+x+^K%1oA+Kmi-*ao`dZeid9J*u-$nk1?f10^=
z9%R(qR`RnD&nY;}<M5LwWZK@^7=HB4tKoNk`!|OD-R<G~@4p{@@yj2E)3YnDo1kY5
zvUxr#89e;8uA^_(1^mD=8$)t?lTSa^G7#TJy7C&2<k{Wc^S9o1_wq(q)_GEnUNe%M
zJS5GU()#+-;pO4JkJo$q_6O^bw}e<v_~!L%g>O1-9&WxMeevO;%9k&WhgZjkGW&j0
ziie<oc=N`exwzM0_SwdkA6}(poht4xly8ooyS}{ee4Gcd*_`xTTY2O&7|47ing1*Q
z+HcuNY;Q<iT-^+R`STx!zxb;khVzSC<!i@prlBkJL;0X1-kRb#J5^wn8+u*$rrh?%
zn#<?pEN{(mI*T`TR7X6oroIYTXW3I3v4OI)#b)7#>ltf0?@!O14>psjtF#g71aUnb
zQGDu_kC<c~LBD$Qi8tfe<T}v2|J945;l-iKhNBE=V~WUIeyrPl|K^?30S@<fMaQQY
zB!l{kcW0Nw&%XO%cz1Fp8S=Jc%Bda#rmgXo<R`B~usNExfqv9`%D7&u9RkSchjR%t
zCUn%RoF=7wNJn|wliZ|j`2ldH(OOLd`X$vp+Nfj#-28>7Ue_8GHf@5kAT8E6S*zgz
zZ#K|rTWmJ*&^&L;@unIt3sC3Yym{}p9=$23zJa{4W{Q5<h~llfm(nd|$mZ%VzJH^-
z!drCd3pT}%x9@iL54<Kt{XaWZoj$+zCK*bwoHDF4(I$TLH@;Dv9k(0a@S@Dw@PZgQ
z-tjh|%H>psy7%G!de~L=cc1eP@16S(>IL&zmjND?rQN89xuV>FaVo#^1AUkuob3F%
zcJA25;ty(x9xLr-!p8~k%vTcvL>{|h87eN&yMd8!=QhYWhS~1I8q9eCJ#J;V<zB`Y
zJet<S(U#!zX{?6F!rbs>YSZ~w0!4j1jepfZo%s6nc;oBI|Nfu;C*;_Ar*Fs;S==66
zBMF)+XB?(vX}AJ8>M)9fhDL$A`3s0D3qB&Tn@$anS;8Bv;*z=tzS!m!D3X%HGIDJs
z3o<9R-3F5eYcg<saEnSu7-c%b$UFWHOpI-g;fu{(GcM=TN=;B>A-tHjsQ6mi$xR{w
zvK=5@k2t_F+>yg4i4!w=v@>2C7R*_Y=0mZsUb6x8a@f}mof`r-$P7jm!JS^XbLYSJ
z#urccFrXp5yb;1p!B1~7sA9mN8>DYa>Z}>W7^fR9H)Nhj;faw9zLE}c*&IVWI(&rW
z$s9g%%1zKa&g$9f?He-WffYA9^Wf&m0{oiYqz#_3_F#>Vf$9biZ$99<ySML;#j*jm
zrjFk?Nzuc@^+V2Dz_MNml1X{sNSnY-tyg5@C_DVvQ`zY%nKQ`O4Nh{iaY$k1$&Tbj
zLikBBBC!e858!$7r8b->5H_UP{Ho3#Ji4YV@&pGPVGQ!&(b}OHp?jWOVNgkU0eNTR
zk#rdt@RnYk1C8#vIg+j?4@!snLmTvFlj<S^0v2wqL*)ZnA4|zZLS^Dt-gQ&=29xNx
z$%5lUo!n$O&zqwP7_j82xqR>!{(f^wKYz!IjB%w7ptpM&=}F0nj!jX?pnSmxVB-kc
z{WP5TG4Q^ZxlmpE?)zU1Z{EJ~lV5}ChX)JF6S{u-i6^v@FUpBLF=%3fa;xyO>&xM(
zc)$Mfx5LiCo+RN3rW@%XudXvV$`gF#)v8xh(jXJ*B`bXWmY3rA+jbhX)<&q?jBu3k
zh6eQHQ<?B2hI)fuGB9HxraGZuE`^62`lsl9Ju)Ew-HqDhH5=3#d|&WnoCfz-;!zu4
z(!rhvv3m!a+-}lVR3;i+(RO$<hDji@a&|r{5mU7h4Zx%~%Gv`?$;{-Ga5mm@?ln3P
zFVWHlH>AVEy~E)r-~7lM?Yu{4y;$%7>{Hf=mUfudRyeYZ&~jiv*aj!Z#7cg7{1*>`
zL}T8~4KisbUUSU=ECwiv7JU^89$fg!9j9$V765M|*e?2_!k>T%X{*VMyMjV1FcV+w
zHB{h8+hR04Z5K@G)C*KfBQl4q#jrinGz=P`44qMM3bF|H^?XEYWhAD_1#Db)S+v^(
ziU^<h6Aw><6L~7n$Puh;A~ku8<3dD0ww*+4+yWhM)3><F94^8#5!0WEt6Mcl#iMT)
z-MIM#+CjcTrEB<kEaCicPe|B!Z;2<isMDP{dcqgV7CbycQzV7Qi5ezQ8Y0T36AJ)m
zK$yRdq^b6qppLgBrJshLU)Y5R2meNjo=xQ-TmV-PBe{Vn_(~6UAvF3M!zZv`;f@dQ
zmX<faT4tkbOJKVH=zsOUnX~yDgZajn^U^`ezCm*?;DEWmLDBVB=d|#%cuRZ*;9Rdh
zLyPZGn&vi6CJ(kO#{nMY)1YvD$cU3wS{&nc$ey3aHo}keaI9eLE>7otz)#(3{<4J!
z$1%LAZ^60EH!!E%uneY`f$lLdAMq`^S=DKH%#>~q%YL=z#$z5<Q26A25eUPiWj2~F
zU%+yjO+O7yv&1>41A-HM(Ki2-KP3jPrH2z4y5ze!K<ah!5>s&!S}}`$5zH%pJdO7v
zn7_=bc5V7av|$V@Sa8Z?fASSF=yA`=4>P7~Lc_>=LN#1W;>tNj?Fb8mTd2dNZMrE#
z3wh`OJU@Ee9wz!8-|}W$tq!mln{+gKcZ!(sU8-_|#0`zY%rEo?Tb_j(xDQ7@L1VO>
zm-tu9Jg1-BK`(|Cj&!Fww+IuzN$cPqy<fC+ihnk~CRn5u=b7K_X#ogdq*(;=QYm25
zzTg3#MP<e!%m+9UY*RLdZx8JZPd7AYSj#txNWtF=Ue7n8n16AUJ9Bx@6Pd@6Z<$&|
zB`;oYQ0y!gPY*MCoW&fDIgI8OdGko~GWo+s@#OFvUU<~N{Ll|c!dpM+NaMcEk$u{Q
zB}Jz!?6L5j`J-jm+=Vo)WB5YrKMAs)z=M8>pb*VJnO`CgiTfBp#<QfUAM<_Y-SF~H
zbi_;Clo4+7%G+6+o7+x{Z;Y{_!kh{}GzIed9&G(7tu>WNzU`CE0(9!{tSK*SBtYY`
zl?=@BJ$E!GHq4k?YEHJlJN(_h`<uhR{yTri8y;-Fym|Y6;Ap_xd*;rpAAImy&gI3W
z=B#|D{l;@r=GDkYQDC!1ln1_87f`y$%EOa#^RO#(?d{E-yot8GCERn#z?)v`Ayl+p
z-(aoj<DK&LWH{K}8qQDN4uAe<|6(}5JRc5@UMS9%>XdTvbj@$g+)Iu%*F*f#84ood
zAM6gthug!!p3-@y5|n&ws;+Ns4>yW_a&|HN?B~B2e)iq>!}mYD9WKsz%kkRBD@K>m
zw=dEoZ>8;SZw|lxqnE?K`|tcWhu{41E0r;{AN@wy|Lgzz*FN@;`7awC92>|6#s}%2
zHbH%04Tj_VRFSNYedX_9f7|88_n+Rdxpa2!HHD9=JJzrC?|LF#ZcA4@u+AFR?(UBA
zlZ_%yJfh6rpP!o-ZNT+j<w1RAL{HgcShJ%YaMPyl@uuB|%7J?D`uJej+0xvd_Mm*A
zuLI$+`QFqo(bLt9*E)EMi8TN=_AamQh9BOaN`_OF6US42a^2zxL&`RFgZ525{pfP?
z!`G6r`T-lsYEXRYtn(rny+K7?68gg@)kU=r>a*W8m5wMP(pa_e<@H0w^~Tr6y5z~H
zTznCK(#QkL!LyO%euDN5Pu^NpeLz0eg0{DIhAriLeSK{>JlywM24zIqy*)XTT&*2Q
zF486r-@$wR^2PAGzw;ZaL)$(!^Zh9&Qk@THmzQ4SLk`w4_e9HEkSOEw{95(lVz`zJ
z^i^!Qd2@SfV|a0}GrTz3b^GyAq3R#eQNFXdHrx~c`qpn6zI*p>IDLPjHgG+x%RGO=
zI$2Y;`z-1V`fYu*fIh3-*wtqPRc?U$BlVdD(DmT}`UkRp7k<4hm)un(Q&^m&D`RXR
zG)<lwSPJ^MJ@kHDqhHoSnBDQ?p68>7<#>w_nODaJ`m}m5{^4j#@Q{4KuNp?_W4?Oh
zOXTFauRDKpQ1zqWS4+=F9sc{j_lG~l<GLiS8C>9`GKaM@Hl|SyXz6Yerv_3M3^$Gz
zzoL{IekABdClpDyfW&QHTe2euRG}Fa>~w+-#Nh=Rg?R(6<CjL>+!jt(DtaSu5IDSz
zcfx|N$dSL{^&2_Zsx)Lbk)0Hd7Vb&{S{IWz0=WFzf4pG1T!L^!c-Yw*>b#Qz+zizL
zF(`eZ8}sW|uYF!JH@*yzOlD}D!<8EghP{B64mSigXgI2X%@Q9)$U?ps^DjL>;>{u6
zXkpXKZ(M!A<`rL#K7>x;ip!0L8wYZ-5yk+^XZ13$VxZgPVv{IOd*rRI+{iUJ5Dp^C
zB-*v-Y^-oDH6LbeA7wp0R-RuD`*Qo3HFebKN}E2%PLz~nxt^4+jdCYYI`U$t(o3M+
z`WKy}Hh_b@Eq`E~AN#qvmRkdY54?WP0ACXY-8|6&gI`{6X7iAnW8R+Qm;^p-8$HzK
zV!}&5DGPMOdDeb{L<1K-m<cT#kkGMF$zY7*|9rz|;2@zm&d~8C0|SZZn*nJyFY5!m
z<RNKGZpGoLBGV;fE?2UQ?(T^vUU|wUPpSFs2HzAl2-6_W#|;XXv(zaQ*II>(LVms-
z<^hlmahYr|DdXsg=oZ@>c8ce+;KpuWp3*~aH@7@dm$$AEkwxunb_t<!rFpH32WPh*
z?5hs#<&8*&^vJivYv!t7B#ACuFI7%#8ln$q*eqip6S+vF8T6v`>evtDB2Va^$sPLk
z8zkgSdcrs%#r4~3)DL78AL?!Dm5g*HM%FExhCCg_cio&%DT4FA{kr^R28_fbKF73e
zZfiir1Vr&E2hw+ar%ibvH*G2kqYFnqCVxzd4H7{jM-)#P(ZO@I#Vz5}-WVwF?eU~o
zUi+WAF)qf$>)3Kt*hnV}T@(~<a*T$15I@q*zftL8oZV&nZ__0;ff0(|0zxBw3jvN7
zW>CLrVlCJ)#<Xk0>4I&3&10(q!28nDX-M}^{(}b>zlm(c-+VS90~suJ!aw9jexbcN
zB5#ou?^v>->+<aj?)Ibp8HKwAM=ilPEyD5RY3TyCOcV$851Xu}42e_0jbFR?=kOS)
zCc69#ScpG*8m0g7I}XkWZnypT?2v3!bm*J`p(sEcjV{F#nz9lMKRk(ISMxW55tsPQ
z50Iwyr4Z}6)olQ)8S`!R!@SY6X-n4dHXnIIQ?`uy1#gZi22T;v61h>C9>Y&|R6rR1
zxEE;jBVYk91Fp)`?hcQIytc`M#S*^BoB+W>{^0R+_|t#>pa0s9I{a_{cmMsLj`<Pv
z?t;;8Ir;lCLdEG&hQhl_IW!dJ*XZqTQ1m^Ne-*yR*)_-Yer)^=k307ps)=ew(Fqpf
z3CjN%&es_Zq<o7w`yrihcDs!>cOR!kE5C7H;HMrLU@;4`?He4PS-681X&1sfc!*}g
zPhs)T!<XrGZSE<o!ht!y!sK$)C4BGq|KV`sWZdt}@jVVk_D{GVMx>f`(ZnB`7T^6s
z$|u27xc|*R;#L{@4fiXJjxJHwV%$Z4Z~8F1yB~k!I{J|gH$U0wy85t|uMr=&9HViI
zvorCEpV6w5aZTYtG&-5wOF<FVXqtD#6~7i<reF~t|1y3~-{@>8yPi(A>2m4g213(i
zK6Lz=+h{^@qqcTm##FTOkGLsPGf5(-@a_#T_XF{lod<{-m&w?wbQ_qSunyn+6R!OB
zjMdgvcMNOBsE7TNb5Z^tCOm}xJI)AxaFco8t>yzP%3o=|bD_D&waitF@O+<&<LP+F
zls6a{=X0N69_#rj4<|AYG%t9v8O0ovxli8g%DgDY<#FU6M-zHpq&UoH%}aAE9~ZAU
z-tg4;#NV@)^2iJI{i)_a-u!Xdcobu>9+E?cWh@suuM24I$s+i=WcM6La~6M_srlI$
z8+Dn#$1|kCJkSqh3Rg!bW}YQ^n48g79hb%U%w+-YJAOD+GBf95PRV==o+364OyN}<
zA=aG?`taK+o`>>)F>^@fe4>GuWP%rR`QclP+`O^HJP>-H1S4JMh(z9o+Vlny4@mE9
zZ+nx)3-Vj}Mim>M<c|e!>sM(-4(5I2hX*M6-qqHyq4+yHd&4nnPOo1LuU@{C`^Y@m
z4E5MjdE(*Goy~Q>arEuW{b7Im*>L*iyWy|?{Lh9<&8@%r@ox=V`!9x*i<{y7`Soye
z&WS?G=QF;CwPE>BPtOK_0^{`5M=QQNc`y6SZ{wU_UJn;nx87j-`7eGL{^BoxHvHAk
zem=Z=doo<GR;4sjuh@J==UIoK9x(U!H|B+=UJ}X!<NLe2vbTmkHYx<@mJQo&j+s<m
z(Cz8@rRVh7+}yIhT!u0{^t-*i;kSrrGpNIDf%c&|xS4TR|DAO=9z52(dV`HK%4OU0
zc24y1_riI&ox0rS@YFXppIApAUzAVkWPl?sNtaEyS1<O5E!D$o>4P`Ou5NBbC)t&j
z*FW5bS(ic%@lm?id7$B`%fRb-v;m=bn3;H_pMKzj`v7lNAw!d|jMs#yx6&hPGSq+a
z!o$S0N#{Xy)pwu^XGAyy@v-7FBq;m|sdmYlk^2dZ__CJa7XaGa3MN>jIc@AvcF0Lt
zvl+}=fu9^VZ@GT-8(QR(brY|p-StOS-ns2?v?!aXl-=tW2gA|9uItQm*Sn9Z6I;X4
zA?ud=!=CCF8`9J<-pb@?Qt-Q~H+=JLn+--8>Hz7p3C4ys@~{p{yaUBse=fcGnAaO_
zmCmD$`y|yB6v>}VH#;W!z%H1=khj^LALYf3H#&}Ul|AFr0C&6-4up9*K;hO`U>>)J
z@;g=^Cg&JtdkR~`FG7!76%HiNtHXeKdIiQm9Bl;XD(z}Zb6ABdcr2{oOK28A<dQ1R
zuZ{cFQ28(Os{uzHdgIG1bYZc9x8B_#+?9aT-(j&G<kd?c4TP>IVG>L{>@J)D#dzbU
z-CX>a*Nbi=EmTt?zUq>04Ifek@AUY|lLe9%pFn|SBCJuXwU|a-Dxtwb-@+ywm*Oqc
zl1?5QG=KC#IzotrIaFnop%K6-5y#jQSiq!3c#;AqX6D@)<B7zj<0U5z!0G8m!{Qi&
zLLZ;kIC5;-!NH*hAm8{kU0zkTjE#-s3-S~Nn-{s+PER-K=EZ=CL8xwUY=~U(y{?Ok
zycHtDH@o<T7h&17$x&`}$}GG?;}48#kU@If=zL?-0LD*o=!QZGkq2&U{J3#u;0P|8
zE!@C(`hlB{kJ8gXZI^+_{=x9##Vc>@9ltmp4h|V)?fRzn{Ao5ws`j+(N;Q;wX^C>Q
z8b=%hhU&;H5pv@z8tg>2#~snghfBgz9=rk6o>W7x`Y{0YW|Z}Hn`8CRFFIv|khij=
zyL-`~SI!k?0D`VKhno+~`rK|N43vrJuh{&$$rDC|UtL_v7B4mo;jeht9t^fYpZNKv
zud?7~PFZ;aOY-~3KY_nxBN}glD31&nJlLab7_>_V<j8eYxg<{vX1FPP5`Y@y9*?px
z(;Qvlr_{^`lfel<9|X<QC&Ez=7>LLgf$1|aVRA#7p~a8v@P~#VmqSui-Wc@fDH}hf
zCt+yAyq(0E);r$t+8<sWzp!ERYh43*>qGP#3TIGJgC{sMDPzFM=B)?vO4li>zPkQv
zP)PfMg$Hb+NqbBDX)wnkGZP^uV0ks)n_mLTKbwwuipFQkJGOv)C|=&UqMq}lUN*_7
z@ucg?C+BG^|D>%b9*C`P(S~GiYM`pg4YD(N@&>Q|H8DlT)P2%=qI49GiICf;2BqL(
zie4r2`UYpLZ(Fy-+mH@<T8$4UKc!%;Aau~mr)32g^F`$5_O<ns9*`BT%1|DM@pHLU
zJOYBVuet3L`=pIOX^qF0h1@zIyc&A=G+3a)q>qqL_SVI`C39E_X+7Cvf5@=y8n4l3
zOOt=2vDZQ2>$oOqbiRPI%l^2?oBZfZ26+OO%=$1m9_cojNk+ITwuGl&<Og<42i~@)
zwl|fRZ2Ms--h`W0#wh#bc0{}Ri<A9!Ss%uP*8F4?7f3x+owRQ(wlOXExIeeSYVrBu
zM+V~(4;n5G)6Xhz{0Jt1qzS`K7|x^tD)<6=CezAxY2dV#cJ7I{gs08We-h015kHc3
zr=`5fcG~ce+i3vi6Mty2(X+K8H^QTtM$>+yKGgrDN~N4g-?@~VbVD|!9bU3PRL0E!
zFKL9*hypvIN!n>I6%Gaq1}FiiD9C%Gl_U-(y^8g>x$!lJ!Y|OsvfqxNTndEubOkfq
z=qmc=EimtnzF-x;$8ojm5gLbe^tN00NYh{y{!4Ja&Tt^*Tf_w!*9v+EgX1^GnP>}E
z`O97m!(HJxmTBY~`<i(e(Boo)OW0&|HGVOy{8PLsyy)k&fhl}41?3iY{EdLK8OO83
zA@-+FNgpuwLk(9ZsF|N%;q*%vFBkU-5-^#OQ=Fn7{S$8P&piP|k1JXuokaQl>u8N@
zevjW~xZVJB+$wzUZyu*>8`JJtp;4OFh5EG?K83aT#C81m40vr4W_Lp_#C1U2h3~_;
zF5x==gs-qUJ$8>rhpkK;Ryr-3n085+amBBNbtZMKH{$O!;_u^*Zqvy<;z}+M6&zaA
zcd*J2Dt5=m_o<1;WD<;x-|h?+^q=SodcW^<(Aji!SY?3Tzo4N0I^mYxy2`Q{-MErJ
z+-~!p+cEcJgMbYx9>CO`=lWu}m;3Ja(u=E4o_x^!h_@HmJkeNR^8+ttGbfXLY`l1r
zs58t{m@BdF;JGoIKboU52Vsos6C@-NYZ0Ez0E{h=Q&z?W(Ghb*AJM0ICTk0$-UPUd
zCi6EQE^Kqp*5-MOg)j_nPr%EYUz)#o{s}v!oy|bzSIEGFiJ8CQuFVsO?J{rbYINyE
z{Aaj%D<t1GW0Qaa_XfZyYvNGH!Ff(bZlRGLyoOK)d8A9y_j(PKjhCzz2R+F~WXDr^
zXnxNcz!mEcn&+{8;B|uA9QQ|;w5j<aZ&GnIBM%j`=Hg8-_;{m+@?)KVIdwJ?HZ}L#
zKR6hUjt++xGQ2&+@qTM6OBIdZidt9Q*j#_+jiS9BzEiY0oW1+y@Gt(~|Igv(@_aZv
zJQ{xTcmB2E`S#)P?)}y9{`5*UkhkP^#rw#o_>kts#hJ|cz^OL*ZZ?~7tYdsses28+
z+xI`541e|4zZiZd!&_dgqo7X803PVua%+CNv$yB8k$q=-S2AvT^NaNkj{IXIgE!WA
zqwIKp$K{vu=WVAAZ&FZyc<YAs7iz(V>JZw`x6#?aWsU4cX|jQ_>Gdq;^~mJ6msnd-
z4A1o?;RoR}2WK;E@S8t}2YX8U`M?IT-^dd`MRR_rgJb)ZyUIW9D{DBkCC1m}<451T
z91iz(Ef*VK$WDEr0&#R9>o%m1%!GaTC?Qx=lU~>yqWw@;p~0|`OZ#W7gmif*+M7aD
zIwW-cMz+@O8qRAb%0$}6^I^TU58Q-9Xqr#b1s;^q@+MK8EfIznS(G6fZ6_3XbX7dJ
zL#1DBFzKv041>ScU(#mtFf&|uklzoCD^KK~@+KkVNHXLpa*$=-hU1}s*7#_vY-(+*
zpX80Z*DsINU+%iya>~!a;jYSd+gW@2=6lucQ@`Q2x3@jKI6fG5)mHe99Y-J1jyP(S
z%~Fm%<a09AcfOrQ``p-2|E7NRT5ZJrvvU(*&N<GM6=3}~sJcTy;aA#}yWM5`YPZYQ
zu34PXUgX^y!bT&%TPb*lQ`dxO5c<G_ZcO=QEz>Rf`|yJ0ILqL;Ni7IN_?1}<o4{eK
zgBBn>CYbwyRk(u3!V12GW&uQ=eE$k@zZy!9|8n4suYdUa|LCXUhSb5zc1HZ?CT(O3
zPnccMnI^#)HvuA)yYWoaxDzQAI+2NDg8b#Ci*YApf~4a?1zue&lj-bDn(?*xqj!-v
zF6@yWIn-O3cjB2of8&hDmcTUUG2f#NZQNY2z_kB(NjQe>H{~!914^HcD4mYb#y5v>
z?zmY{a>pv8PL?1BGKAa0=)wRh3uk!+evifcm#>E7<73~9$RB-ldg;c{v(fnIt;HcX
z3=y*M&jL6%KAr~Q=@-sVK39i&b0z;ZXCCvm7awEgrf}=c8{R<TRp~Pg7|zup)4BU-
zI~i_-&@pK6;<-9X^5`2gw$deio|y4dAEIHC>+tZfnWN$G=x8`TI=10_<=vg#VQX6h
z6%EXMvlLI?x=;%J<)`SWm7>tjt?<F-s&wVR?Werxz@amaI2gIPOfhsp=Gy#`n;SHD
zJq0NnwgKOyMMsd|?$SUAJ@AoKHmT4hgDbwRRBzN>j+=@a*eOnJB=R9y^y}jc(YFUw
z(!I)vr<o`tZ<gJC@a7`62Pcw&Bm5Z5G5F_65eAKZ6~5IChA7^Ek=~G>NsK2ba(l4D
z&E7XY)m!o4O&f4b7T9QEkVTwlPdUmG4M{UhgkYxnif)awJ~QaUoyiNET`6-WF~sv^
zg@K4@EemN=P_$baXcYhxmw|(-jl)+KlLYcenQp9Y`c-!ZcE^Xj8OZtIFVs$6$bX~(
z`IdO*ttRu*poq7lq_DAWx=gWnBThO$S9x5jJV5ib6dOLgt;Xbyd~?>f+mXu2b)OAi
z>E2JwiKpu;d6S!ZkvhhpmsjY?COVW(jF0XLXZ^@c`AQ(yX^u5yAWgY@QgEvQqejul
zhi*1B_~P~YO-*Px;?N;t?Xr=R2`F+SR~_XD4U6$<vrX^FL!Txcp)>mK9CfSyNB%8M
z9%!3P07lt$O<7?>Ye)Q@tNfKlL3%gAM9|u8X!1}4(ISV*?2La5m$$_TjqrYJ%z+?c
zn_l?PHhv3`IGD(hIB`t=Brl)sxXYP%;gdKx1kF!)d1(hA<0h<8HyVQ(1(9}U?(uVJ
z47oaOjDDkTz>^GaJlJ*%>A<3u#{n_p=9NvuHM;181p@HuTgj*DQOI<qq2{j(@mpAn
zZ~Z|JAL6!f(>Aei+mF+~W)t4R%!40+aT#xZLB?$wjQq$39$IW!-V|zh7i95ufkE5A
zvNpFIZU@9|ZtDPBx+BQ+;AC9NG)O_>;Z{6SF&ZB9skF<EH}nLXx5D7>G>IXD-!41}
zC;o_&U6WJ$3T=tf0#k)rj4_&X@4RF)IkLp)m#j2763Mn?z-m=P7sceBT?tDzaQKXq
z`O|;)>)rTjPNzT0^#Cl;ATB>}!6YAcEc2?kWk$f?oIMoH0<B;vzM%KpH46tS-bjnP
zY;es#(qI}C&iDpjg7bBT11TS9IaHU%u!NU?!h@_h6K+n^`+GBn=|_9uSk@pXyuET$
zt1BL+pvU#b3nPz&2g<+bPWa@W@D<kMCsQ!TErjXke@Bx=K?geAcyysGR9xGQzQ@h|
zy`CB}yzGR>6c3NDqknQwxTzk*zf4njIqY_ySbi+}xpXoM>DAi|08tirHRy5W_OI}b
zd@-!^;~~6{W4+SXIcUTet@~5CU4uY5tm2fsaa-fE>3Q9euHadGd%x>DxEa16e(aVe
zW&z*mTa@yTc;j-hY%=&qx$QC860G6{EaK4<iX`BHIN-5tw|p4Wa5MVeU$}x*G$X*v
zcH~iLt8<M8tMFYsyWwS@OWRQV%gk|$ezWNO%k)77AHXerG+e}vZ;bDYt^~69%)@ve
zHLtk28g8ymhs(2f!-dSv&AGzwy<za|Nj@#WLzT=bnG-XQ1Zpna=Mxu6WURs*G@DSQ
zjs83yVfO^tD7y?KWvGwv%v_4aYD-TXF=D>rH@P$i;TutW%ZP_AncI+R(se2=9nE)`
z<7m#qm<Ybi$;M;jcqo*47AY%S1Z|#1{_Jh}NWJFgv6;tH*+fG|j^}4HiF{^GMVh?j
zl1+uQi_E>S9bd-$o8%sE145!r)=*d{V7}`)JBIZfWYe6V%>{}e^ZAqon;WDFU)E;w
zz+>{q2>{QYWi4q#W{ZV)=IG=LwG*y+9_7S(4GZ}%j`xQbhcf$H!}f;qq58nvfzO}b
z4+lHz!^^{+VQ2lRk0ZRgI2*qE#m~LY!7+wE`R(5wj$ZwE`1ovNxVYvmxR2f(+TYt9
z4v!Cq9nB@t3lCqPou7JL${Qxq3;L#S@f%8a97&kB4mkmbwGr~fpp~&Q`p-A1HoaNL
z!_u!_e={5%9s8{{%Aa|BK07dYV~`{0cp!O4bLV5_k4>I)$-&0P$B)@Sd(K9t%8CsM
z>ju3_{~Q;|X-KT^oS&V^edV@6-J_l&7xQ<w4<RXwtmUzpmxs>T;AGR|V1G}t^QjMx
zQp_g_I7*Q3$+Iz3<9OEycz~D5+j+qe$}x(3@9pStSLuEj&aZBVi%Zs*a>@|(!f&Oi
z{8_&^yWl~0wRz;$*qkz_{_^G$`l8;^{*2q+b~t|8zt@)p=)@b)GL)GvE|0X67A{^s
zqEc;^&t@P;`p;~j!_)LK)*rd0{OyDbSY$+hq@VoLHoP`UJlO?Lp3vU1o<SW-zrzu(
zY>Lr7RaYd7*PcXAIn{|r%xTlE$xCF<dew8U7xC~RblzBI4M=fs)V_F&j8fo~tdo-y
z@w@g$Fl9o1cw>=12i;QMJ3Cv-5_ROG-xAy1;doi}8{WL#`2PILM=JWkcHyos&(yx&
ztKQs7F6x+bYgy3&ZvX8LlwRj<pvSxMmpiUp<*)Hp=z5xRFQ!p2xg%pz2oScK_F*@G
zZwa?R@2lOwoECVDKL9RRu6=mPJjXABX&w$9VV?FwvOENjg%x}W%>syg%>CEK{c2dO
z1HUS?jjuoasRue@2+vx4f&G*Qh-5EeYB>Dl$>K(6OcBrIc0B#C$8=&$P<~paZDZ~B
z0%{8nlwVHlMxFQ#M;v?**^hBxsO>>!!h=otgriX#g|m-OusygZ|M-*JsTaL|k(oYm
zif3D`n(qm?!!#v$IwZW%1#5=6fT)W)ycQm>Dd^1W=^!?Mcx#0t4v&tG{83SEHqf`*
zZN{*qMU2J@Ex?0;oNu`5;CaHqZ*tw_Eu-3$;prf5xP0W);p(W_@bYWq!ud!&4QLp=
zSYGEV20!;gZZNrdYLLWHg_4=`ae2zfZ*J}6qp0NJhz+lU10P?wySp#?U2lkOYCy}t
ztaRxdk|6^bOO3@{8(zNenIt&l9h7)u=S_QXfHh+rz>{!}i_Yb8hz2PPX80~v;e5OR
zW#WNHd!?O?KXfG>W)tyz)G2S$U0yK=ls-LBx*V>!VX7=B2R`ch{!9}MHXGUS;t3sS
zDGLVXHw;|RJx3YbA(iFvtLeHyW1)9$D{LxpBW55&F1cZ&7s|_TXd$+8VNIY11{Vxi
zcoR!Ec=(}n1~`6$tG%tnz%OOdo>&qi20sh{T>gs3_qosuI-)UPlOCyg0|i+%n1sG2
z1$HZac)Q-pMs7>k!fFr_TR0iyV^G6WM4Z>WzqjvW3yu#ChnL4Mha>rSh2PYm6{+v<
z@`j-|Csk%*GOTMr!iP%X>w%BTg0e=B$juW|v=2YkMH$QXlX#MiNeFqT#(49INr?2B
zi6E0GwF4$Qk`+2{V3DPk2j^3M;#gngT478qv5~JPXDVn?Nx8BSt@?0#!$g2I6Pu?F
z*EckY*~tfyeN2Y<*MyO>!El5F13t<%<v>zRukzrwo)wSsuL%^BUItn`k>n@o)UFxG
zBCC%GBr_>BL0x6Hw&}VSv><ge3Xr$Rbje1FLg5)y9Do_f5ay~#pny6xGAU@ZV_I=J
zu0xVWZ3G2d*9Lf`->UZu6qdc^S8mH~TKIyM$v^cPx^^vaEhk>V=xltOE3suHu-(N^
zwlQQ|BDy9K?29QZi_^f{;#tN<6RHX>oPM~;2Y#p9sGMi;{6iPDHwrAs93OcIhdA|}
zU<)g7d^;V%?fMCo34CM3&;MAVWx!7e^ewPq8Xi>mIk@EmuoZ&Pw)Cw#$CXhZYA+c4
zlY&9mZm}aXd7~jYk7-AQ1uwMAAYnv=9jYPEOwItFST+wYT-0WCixY098+kY#u=KO7
zmIxQ}*iO^o@*_u(YOhM%gtBA^SFoXxB*5|cC16V12rdisP$owF#t^Ul;wSI1TkVxR
zcnJUWYu)(z`~Tn{jvHSE(IK(~=6=Bbh-*uT(NL@f(Di=6^w<g;`Qo1A1m&LNfmO8S
z_qF-wbie}boCdpa6;8cuutYET5}dDdoUFJDkH6|XP&n)f?-<oKs?Tl4;Vax2%EH)L
zILNvOy1ZsQ=cd54_?wy2%>5+fHrx59aHUCK5T?QrKQb-SgU9amlv-A_g_q0u3XrFX
zCNTgZTYri@#uq`k?T4b=%~bxu;jZwJuE)XOGECu^i6*uv<S=r<^l61|@>d^hXznsX
z*p$9vOlkMrE;IN0_duBY-_GxFe%32)ndZhrSV7{H9enKg%Q(NpYh`5LBjts8uuKZy
z$2YEpt;Fj$wmDN7ABa0~#yd%O6^!9A)!qvFunCX7y3PRdb$IbhrH~tbRjns^f;VzB
zg$rJ^z86g-3hCnYZK$tXg~K1T;~@a$0?IU)qU+5AFp`*Ck62;Qr>q4?thizT7JEZH
zg)+uL+^2l%Li1Zq7WC)slhfhk<cHz>^klfXWMQ9hWr~e5PLpsg1EL#ec*Nn-kN&jb
z7}KzsKtG6E7tb|Zex4IXuRbEq8w6~W<kSlIqf-{`pFLxvi#ckJZ)2WATAr6NSCjFg
zI7T>2CE1yad0x#MZkpTqNJh=Ep+iQ>#^Wx@487-bUITy+T$HxQYSACG7=q9?Z<hEi
zDmF_zs>&Sv`8wag${d?{E1z<3I_SZ3Ye<_8u=R6#=vv^1Gnr!|6Xheo&E|tQTUe9O
z9H04V9*SiVpYjpg6pYi$oRsp%uen@%tBFl0j@;uLQM}=|x3e|u%l+c$VEFdc;qar^
z$HO--4>Y&l^EbO#yyvYijtL|VZ<5`zej&r!(GPFm4O_dSJ$^N;ZS4+MJgh1C*pxat
zJn%!?2M3%QvSHovU@#9}^S~)(J#AE=r)Qdr^I3yzh_PwNMkn7IqYg|DMzb!EWBIpu
zqfLgq^YAJgQ#{DcgW9Y^5tofej_hMS=9y;lyn(Z`voTy=T@3F}c{uq-GF=Xrd{#la
zWX*%k59%*(cTqO}EQRV5hIJbrW@eoyn?kfHDVk!W*-&1RKlPLPl(hn%PP4OZ`N3mY
zgX0q<uDhhGGRs_^qWk2B*NM;Zn8@MwrM64{kM{S5ql2AcZG$(}?kxY+&As%WG<leO
zb8AC#d=&mp`BXa~AJPSEcqo>26Uqk8Ub~Qsc82b#$CgXH+>S+09P&yXqcXGc1<rtc
zlAq0{bxx#NSNn&ae#qg~I|y1`jU7X4plt)AEepL^2k4CQk(+jG-73uMGxSp`Tl9{8
zd03b9*~HtG&N&8@Z=sPN;<3I)nNlZx0**{Q814swWmtRiy^wg4%^Qc}&xRQ5q-PiB
zl8razIJJk3tYNrTd8!GBr}EF*i}KHjMQm=du18<WhB9xCNldq0ufwViuvyH;8Yf9z
zou3SM*B90Z+6d5p#~3H8+B`6YExQXm#vj9$X+|62uov(XJ+SEZRd)I`!bW3W<%e#9
zMIMWRj@9GjT&AD<Cs_8UpDyF^<md-_`pEk^FwN)5J}VpWSbCqArR4n*(?HJTU-oDD
zKL*8jwSo7)9RCT%jW5pH1?afih|mj^PO~gug2Y)HUd^Y!48qA*1439aNb+JHBo61-
zaKtB`G4dCy<}huAhh=y>WOO<Xwndj6-tjnHcnMd0jW>Pd035@?gmri<UBbI)Teu_O
zFK2FEDWpmX&{YUGXX6{c)|kOF?wJ~qFmC)n;<Sq*_wo}?P;Md~tgQ3#&`rMu_4?JT
z;pq5<H<El~X>lA)al(nb;6h8f8Sqob?ql4%*~O+xj?wdXx~??{WKo}2e$~OU!G&Rf
z$hpy>$;U!7pknZm8$bDGAiz^Q3^cYh=<tU-_Yc}D=EoXP9D9R{O)tK)#3mP;VS9d~
zYuj(F`3M5t6!NBv6({!S2aum}gutC8x=|Rx=}rmYiZnQlv09^#l!JW?9tngY14iQr
zDmUT5v5A={bcB(C1p{*5Xwj$i!Oaz&_^lr{2)!}LndZ``1~}LZye>7MW7C2Sl)N6V
zG9X-Kai#&^<@x1sKMrPO<jw$AY1ECJn=7J|f8UT}&=>wpXo9OF2R$KBTn24uf;#98
zC;j|Jm-Wq!y=`=OV9WqM=VCKZgfBOI;+1@{C6jF04Vw2LP%;z7;xQ?D0waS=mV+?T
zcHK>V<+XEUg@yrBBqb{u9FcCd1<H{;@l89P9@=V~Fg%rZsCsp<f2jJmHLNR7-Z)|)
zsxslN##=Ua(Z38E9%~wed4fV+lU&qq22hG7kj&IE%31MlG#KVBNZN$umZ9$Xi6WCz
zM>1gH^>^|De{uqE{A3FKvx$T*phq_uSc)I@UNouq(lLWJ!aUJYY9E;pa!jV`n+E07
zJp+8j|M|vKZ+fu_#-x!XU1q8`A0-!Z`<rEa$Xey?a!@#Nk*g+;mfeF>2CLKqbS*zJ
zG#<P`=kxPLOL-wf+GJ~wM#*13;oYbV2%P+bcUNVyxQP}0N((9*ZyaDmQ}-~F+ZoR%
z1$cvF3=ude;UtEn?&A~WI^a0`ZIO8MG^qSUu7Hyfgtv@%C$98ErW~Gl&{}3=a*eGp
z-13ZhCIouh%`Jmn_02Ma0F0Z`0}?l0yU`y3jUPVyo6TSyhx^Q?YX@MFFYryTT&)j>
z=|Ta4Ych?x97VDOrt`fIJtz%FGZRAPH&K#-4!na4C)=^W%UwEwCh`Pt-VRHA$Eq-0
z7I>SM{sbI<@G`<%RwBtDt93w_`IwTdx$V=)Y6HscNCq1C#;-8aG2Shx@xZ0>#ETo|
zx~Mxap1zOF_~*ccW6UFnF+BBy`VEQNa;4vJmPa%rFF-g7#TdexPAXk>BgR8O9&tZw
z1rlD^#aC|8Bz^o1Cl)=<ykn`}AYT_pw14tzedCLbFDm6SjPf>E^7F$lv%sO@3SaU1
zxC_weu$RmX9^}zM(QsWb!q0Qs%f<a@2?!&e`!_kM!wpu_3s%Y4r(-@HZlWJUY^HG1
zoa2Ry(&BeziaZm)xG|%ogsli)6^C{$hYdY*C&Gva7I>7s0v70IenbcDKBCW~0s4uy
z{BwNisrSbeEih_Aczi8ih3H`-HB5Xby3yZobDBQf?3%kw(l0;yj?U8`_D11d-gJRm
z3&Sq|MAMsDRwhuG`@7Z`GiG(1gxWoa!WHk(cW#*}+>Z2}Mh5S(F5#^&G+q|^a9zd&
z$hMA}y9^lORa!vO3qiS8>D*v4{BbYVZ{m#}`ze<m=b!M4WrMrr%R238WX3YBaG+@9
z;?hCzG0-)lhEsfr?0cF+btMZ5o960`Zh=?HGUHWT3g>bGS8+|=`Ik)#o&2CIW8o5B
zW-*>HaxVH8X?$;Wfi`Gip4Xa&z$s49y^SO8s#Dd66B|<iX>2)|lWOk3dcoy6^VCz#
zV=pyqxRuHLgSuf}-ZW#rmpQb5;N^#pHSILp8(h*orID+-m_Ble2Wcw}=0`p*(cxJL
zVDrl{)s~pEu_<!RH?D6qkG{JRE)SOaJ6p)_bqnU+Y*J{ht~uR38x-Q39Hjqc{ef{E
z+`*Av-bg`5!cbo{fAc)KJ<N&1kT^PSbEfr8<In{$&D(L1Id8|Xc~jrn%4Y$hje2{@
zRTC!imtHXDX*Zl9k^FL4SCW-C*m&6X?2K<^$()`J?@!+ke7?eZP<*MP%{xC0eDmwz
zVAtOxVsq<9uaAZwe{(qe#y3a9Ph@`ddOvS;ZL{d#HUW@d^L8G-J-fIWzJGf<eE05b
zIK8+XZa-=4%HlqA&CMLS&*>_=Y_Rb?tGykco<c40Cmh-*GgzbYn>eE7O|#v-JwM$2
ze2sMn1lByd^vj%deZw1ReiKJ{O0hNwsXu%}{q4IShBt427|zd5z0Q?95B_Y%*7lai
zo*Xa8M#bTg^7-;mZHjMzaa82(@b2_tc>Dg$Z-)7Vp4)rPg`fI+OB_+?F}HYTorrHi
zWxa+6q1l9Cb%65W0aogtH)EO%6rk5LR3BIeU|o&+!CDx2@<F-s{cYBBd`#n7-ZmpH
zb;0e1I^y)H2hxvdc6YaYJRuK2bM)mE8!vZk3QH#G&w5kcyZ^v3lC-h5DXhAKtgIQh
zuaI$_s`3$U-kNhgKsJRVbLuM<w&Ael2ZtM8Y$PE+n}+p3vqV-7EPl$a(K7#WUu`%1
z($1*y<LNx|r?23_?(`GmX0Y7WxANmRIzLRCi16j%T{d=KzdG`_xhN&_f3I+5-D_<h
z+`lD#U91bzq`o;{-T+s7ChuwspICE37?VGf?w<^NkM6D7#V>z&uX=booSt!f;icNi
z<!~Z5M@@R8O?g5u``eqt&W75S(j!gEk9tLwroa7occFH0*ZNUNl$i5GqD|*=PX_;l
zb1dhl#ced$AnnH0=mFnj1g39LVT*Q|@+aN09g5$MqQSP+%BFb)tK6<zg_q^J(a{0%
zPq^}z9*`&D6?fSz;tQ73S_P*0JlPVjRlFZd@AI+{o-)UT{xQ>myr1X)IZ%8$*o8Et
z!&XP<_{n@dWaH}({_v-&LhcYtcz2Gu$f85TJ%s}}D_$sri?0EGdHZ;74_!LxAcAuH
zbdUpZkb<ue7p7!uh-O)u@FPvIExa5_HsNs<P2w8g^6K;}SWxmMjR`k7963a$nuyBh
zc)E=7gmu^9)lL<#ejKkt^$Q7uAi114eqZDjjqr3_+~l~i^E3`0C1nHY#fz82i<d9$
zW?`S3{}>O>{3QqF3dgzua5G@w%D|DQX}J04`%zbUwOj)uo_NXY{`YwsNu4qqIJ~l3
z-=?a!nizC(ljEr-Zl1i+b-)um$as7-93Q{1;i;ZI4P4l;BOm^_s0MmGap5PTbOTH$
zoq--&OHQH>-J{$Qrj9htp;+b79kGm6Or8^JT<No*;(|6m8E%{m9`MI7P-<>&nz%*d
zZ$UM;apLvCH&P8^Wb&jBn_fKab8UTrt0#KyrBA<>uYpJ1c>PqC>>HH_Z_lv_Cic0h
z@Mlo0vPCCs8jVlkFhJIg(d8|}#)KC~MZ;jkgJ=2umYjaBM~XuZ%1=BtMkqJxkI$m^
z!26lXqn(QkAI|O;o+o_R41)u2`oJ6QT8-j`jAASdVx+Ed6GuPX$l>eFI05l^(~1o^
z<Nz4xz${M<QU2jg$LJ@Bg!d*8X~~Zq8ev#J^|96Bh3p;x!_qoeT~|GJ{bkcp16G#_
zg9t?8ta-&pF3OoQw$7#RtIPA@UU3-gQC=JyXuYd$01WsuKtQkUylmG62947(4Q#3*
zH+f-mn7ZZsN>17uZ3#Uwa3&WRPqHv1Bn;V$KA3E1Af!Pm$1Cv07e^nOP&6)Yl@F5y
zKM_g)pveBXY{{(wbjgI=@S}>8F7c^D1s=>{d$1q}<>Fs-!P|6vLye;!Ya)i67RQ$s
z$y^fG;PS-6AGhD~DZJF0bfAqfDe9iswJtqzl2i*2RGmmzxus{hiC2`^1e>idEyTQm
zCPzgmu)q3qk`7%(50FrYkkeEVyt(D5w9G?4aP$SB^b2o(fyhakaT|{t+Yndz9kYdx
z;xrx!ZTTP_@U%I+x>UY=k#QSn#^~}(n1-@H_d9+>b40tjv|{$fgI3(SmdozuG?usx
z5zAQr;%?B=p#KQSUxBjWk^4s6mN{M1s1G#34?44HebZP5l`koWx=sQah->m=J8sJh
zEq>UvaD@n`Xtcc+pM;zSl(Zp-s}HvWAtt~p(#V6ycI%rOg>7BN!KL%2{(#InG6bQ7
z)&_cAk_Wk6o{=fxx~gH~b~}VlKgtp$Y(q(!jlR)HkdZH7t9eXjn&6T7kN@<a|5`V`
zkhM3Xg1!Q#xOgUGT*q#tnZZK*9#-I2`5(fq!XpmlAy1GE`Z#5JupfMLr&_J1UxfpD
z{1mo~bDfUMW(>0~XSjyKS-6Qduz+9T2YOtO2YNiCz!9J;W4RjG-)xyae$cD^a?QfZ
zp3>~Y%e0V{xSq#|ZqU=mq|fk3cOGV54dh=iMoqMz^FM^cq=*||jM81kVMSI<{Fuq#
zL*b{go6^%ofAMPlW!~;>aax{DC#W$r5ZXoeq8~eL4UWMPKR2Bzj=iy)sks0!(Zv|Y
zZ}d;@5hpXH+2MVEY5sQmkt;ll#{4GT2Di-1ln;bWxXSP3cDPB3cDYQ-q`Vqk@o7JC
zYO$J`#SJX_$MhS%plHf8Z1WS2f4NQcFvrQ4--wHQk!~5c=%4Ch*}W;v0r+#DE#5N#
zO_;+cz7Nq?7&>Eo@`M>M^ObwerLHfAyBp@NSDwefuig~%+(nZL=7S!LX@27rJup*=
z!lQ*TZ6FQkjc0x+dzz1%K<*YF76ci^n<y;SXI<m=nz=Wpb7=l09K1+7>GJjz{98Jj
zZ{-0;8HGba`)97p`ULZ}mPG3pEUuE59+=Oq@$eu<b7SNu#B@_#&v%}(p;7bx5fG`-
zWz&l}o8q@|A@ehjHO*T#{Xq}Ok{!@^VLS-?brJ59AFAaPne+4Wwh7igrQnB}S%0|k
z8c=<|eeL-Z&FR*LgWb*H_;A}BUA%>|waI2hj`u&iz8~J5oDXl`y&v9rLH}g<>tDVZ
ze(~o0aB_AvT;EB!r)xfLe|vY=ZwT>r5wc#KUkqpOPbJ62@Lu+tcPGQi`!mUrZ)dRq
z2R)loJVfoci=+>}JI~yj9I|eK{?Qu`N|RUCTl_Gw$~_MoCtqi0Cj;LLV`IyoA-LnA
zZjSa-y;Yg8c|*F?1>*76)#3iO%k2~IfN9Kob$dTtN{@bHhX;G*rY-DjZ}^QNHo>fS
z$wm257XZ(R_0I+or@kN~>u{R4b3%-M6szl+!u`#B>4AF2TS(-cPZuB`b8t>*VI7V7
z$^71J#+5SR@SqNYqwM)!6nt0*U{fe>qLELv8S0|a_QTB5BND?ZAj#A&&^K=!(H@c6
zI+cvxJiO0aYF>j8f9h;D-Ncmi$$JOmd_}yW$$Ag@W=#iO@$fI}VoF<dv<sSkJ@{U-
zDVLVJM0IEwn{R>jVsF>xG17y}@Qyy}iR6FoO|Kulel@%};QQ}u<leNbz(5~;x+`(I
z*g(ldAyV#Xqv->9Ae&&`<~lu9J-oOY&d#s=W*sLtU8vo~u;HsRG1JJ;(TMvyTdrI5
zA$$*vQ;F88_fM6Vn@h=cJ1W8DPTuwJ`~OKF!Qpij<TZYxi5)YK@3}&Qc^I&Wi>ZDP
z0B>VoqvPM`V+TIx?q!*HEJAP-e)3mZQy8#{2e2ByDDOjHn$M7Z2_N7wx-aEX@Ky98
z{aXIZK>TGBzK@w|Twe~^`1-v+$djxjU`rR(MOfz2Iv6}P&{z0JV>((lW)m6^*gyC6
z!Do;zT-Y?oZI})(Kw#*5!vaej<&Jl>;hSLFA#FHa(hx|iaEmmxaA&`9u?e*DB^GvV
z0yd=xxJ2gB%ca>S#|saf4k~vBD=0B;V~xoTfb`&Qj5|`|Gx*_%KaMgye!*K(FZ^Z~
zuY?~R9}W8l90j<mo8fwHmSfq#pfQn6hwOnTKL&WZiTUXV-Nd}v#fFw&Nx#rQM1u}q
zFTcy1U3og_Y<TznM1z}m9-t71g?#vPLtxY7Sn|Al`D%FeN;eF@&2=~&N<NO7<9I{g
zpnZ-m8!~w!!bTOxAA;4562;^RkI0Na`QRVh9Xa7jmnhM<1`NV_@ligMi5<7%5tVpO
z1E%2h=f)?$2Z^$)?gTHRfe<Ex31pja_Z}F@ZGC3YT$_k>OadERry5xD?JfoeJYD7E
zAvNHD7aN2OmU%Ob%^T8yFLl9BfO+tnkFJsylKZCHD#-KoY*6{8qBwpkO?AL^7hcK-
zbm*4zuSt^)MxM?hIpXq)H*eVGtsvc?HHgCPBMoKXZQaSvMq376$l@@i8Cg}X@RfT_
z@zFJT@!$YiGy$qjE;jDG5hgAke8|A>&N{nG9%Z~4seBPe9<cEjz?TwdNagx1`Y}Ct
zF_6Ew^pkCG-@F-q`Q0yuAKt$8My@x$*o-7Ceotk4<4cN>4$)N&p4q@;!hl>%4wzV>
z?+lzsk0*kZp3)?M4K0eq!diYLm(vp&Hg(yA!p;U6e9)=umIlDFGf*Thx`3Alhty%=
z*z{ro;CiM(8}i*;Gsu?=h#{aluW699sR8+>+F*1@onUaCC*tr>XS~VC6O*btqz@Pe
ziQ>^VJg}C{rl2Rc$ffwA8rH>ce{VnOkN|b2{lvbNTl8STpmFX<6Qsguzp2UzlRq2T
zG7;6DCU)>4o<MP0DgoL8G&0Q^qY>o9o2(Dn<<><eaeQwAXM78>U-C&n%Z<M(EEa+b
zkNCyP0GSa)!cQ9}RcX?8e#ESeeJd3@LZ}mXomLESeXq5*IOqosWZ7lJk2t0+dhsoK
zmnrq6#V{Rl2ygV}fn3xxa1Img$Q#kY%MBgjjw!qR%0uK%-u!QTEk%>bar|#G@)uu+
zIj&JHUT`H_Fb*`Hb_I2?O_z4tVZOS^)IFloQF*}MyrEJ5n+Ohbo<!?1l4~WH7OUC(
zf{5(c+$YOTn4qc^;&S&RSLh8l;F<(2be1zR=OLJ;9sM)TY0R+5M*YYG<zY7%0_B%1
z)&oXZ*iHFW7{WnU@hprPi=AvaEN3fgm$zwQ)zW}Q{P6$DpZ>F7^BZ4_FzMYHS%QSE
zhAD0Ux5pD5%xQX9g$FZ_vx>(m95Bb1o9h_Xh4)b5BVCu5xgS`fpW)`P3U6H5%~UEq
zu8>qQyqF$FT63GgaXT349>MovbNqM*f~4>*%>`OxCR{=9FL(J%y*+*kpLD;75A?X;
zL%)nSR&LsAU{2#aigb?WH>dxc{~_EQKPjxEeK0-sAPfG*@PZy+rfJ33z@*nc?Ve=B
zb?#OEhjG2%6bpFz&-CF_Sn&3ZZhU_DtfLv}I=^ug;S6W|j8+$Mt9#yRUZ5SO3fkif
z=C}oaBg@<_{+>1jaf9m1hdMg);@)9g^N;m6@h5!aAa_BzCz_aQ%M%W_QO*BQ8Zi!;
z`Jr=OXSij$<+zi-%lAZXzY#{ZIL2c_JH?y&(S<rL=xNIIQ27p;`6+KHu$dskM$_%p
z@ae<VaL=1FY$oZ4+RdxwwSbf}uVxO%h7j&H4`N=YxlrbWdE>-wL1UtPYuFhQE@cI6
z9qZ@0q<HG$_|lk2X-8RmsJSy^H^O<?F!yE)>UPA!H)|I@{><~?l3n@XtsCaV&v@H~
zqwZMv&zzjRc=Lfa>+w?BuOG&A`)u>aCY$HFPCHQPK<lxeapaP6H-*C{e?aUAGI8>n
z{N^FpJUoS&HWrwRYo04N^Gs;T8~Es*qw;t-_POVp-+lMp@YjF+b1%~0e0VavJG~ix
z`6h3#yg$A0!ak>u@J(^Xdgy`;kzL;EVy$X@UA#XHZ{NNf{>wlA%i+KLv%ehv>Sx~#
zzx@84c%6CSpY;LqL0nD`k+5D%U~Pr<sFQam!|D6?{>%Wn`Q!~&k$s@tH4na#3|GwY
zF9+V_^MlE9v&Kc3-{?~U?go@+&+~6Z1FhU%6Ozs63V4VUnR$!t=2m6&ROP$2K{Ulp
z`|$}d(joPaH>x()r7HzUeWs&G{Kk>!X!q15)>)|M%-g-;3L}*pb&fhhy<v@lwF>f$
zOgz}knvvIVC`;v?2U|H_l5a*)r(u)2p>mY|lS#?x_VjUBQ(dJ_i$FT%Qx80(t@82)
z9PJKr38H9Ne7l^Z2se4CT0yK``GhFR%$kxnZQ5r${05k4sB38l)K7HT*2}De)N|@R
znWlZF-cyf}H*1RMlKM$pw-4g9xXzL2kSDprMV8<pU$b)`3=L5FaXj&X2FOMp{Z2Lj
zDeJxM?cv*(N5j$n_OQ9Znigwc+0b*};Wf1OaJl?8;-{#iuIp_%>Ll5vTv_i^sM41_
zv`^`Xc(gSP>sGvNcS9SxWpgoYlg&uVc$cF@x7Yj)z5U&d;o0!PZ{1R_$z1xH)Jtq@
zr{%Nk!c73eKs~?SXmAQIH#F!l_+>tEFUJeSow6t>H#RQ&TX4b_`SsApEm)wHb5-2n
z=5b;kreB8D_=n{M2=ltb!z6~sc+BB*>3t0>(^qRL==>u6TK>yG{4x`Kp6^_)&qF)v
z@b`b}s;2n0spF*j1m(m_qYThF=-T>E%Flt&(ZcQEFFLNc6`}RDF@W7rG(=fu9(N#~
zk<NU9@Mh7^gH|=ho)>J2&Q~lWMHP?!iYF$fXc6#{qP;ECUW75yUWYmH2Jo{NF-(i2
zS=fz+4udWl9OTGc3p}MK0DAD@rJJM}(Yew|?(Xgl$1k#>ggHDs@Z$2e1`#|l1C4Jk
zZQ}@z+%n7v3zzuO(PjY2U<Q)~eKxQ-f{!O)u4H`t-qnRWQI6m{J3AeIc=L8RJ>_*=
z-U4F~%TqwQ(dq`gyTcjKyTj3o<KeYseCcm<?d|P)v&&DS=*G;4Q&9vW;iN;3ege+=
zByt=WRX7jq<fMF*P1NWxUg2lZ!&0Mxuvp<!%)nFbCX)=pBMEJQn-ld39c6eTkeg>a
z(379~VNk*#&^NQ(9DS40V4i^&x?+IsBOs;EJZ;7xiW^>zG~|0|42&2c@R3voSqx0p
zH`yS1;*H1T-$DwPn?=<TZlD~+2Q4>RZ#F56O~mPDFHLb1ZGJx`XC-jWYxUGKZ*WP8
zq@Opqrj4$;p=W~&-fX0aAL*GsH+dnD(z5W8S8nDxqJgKt@`RiSrK3EJ#Re{M*f3!;
zjtxQdL8g%tBcU@ea#?_{@}cvg+~lY33Qw(|9<tHojVlJ89>7WG!ZDyj_ms`q`?KNu
z@4p{@{@u@q?|$*!@aFqB!`bDzH?tVT`H3*qr4Q2IC!U&;4ygeQhB@|=kJUoU@fdvA
z7MZAf6cwW4_kf9kufnAS>(8&g%bS5TS&&DDx`$0U@x}$=u79ed<S83h*$7L?rB<+M
zmO&PAQ`U_4Jy;bz`t)0v^~M)Q5!biY+#VThOLF7=G$CcG^m9{IS@IN`WO9m((uA%u
zf4(-+Q6H3Pmjz{mY;5Fht!-xGOM`M`hb{xJ26hVHaIUk)6cC0U5U}}M9-7axjr8UN
zLBRO&5<2(OGLaHmIDl%r2y1DWRzNlj%4u5{OUz#pi{3KYAtU!lqaW!yRs0TXbiEC<
zfGIq2y@3E&sDztW^Q9#O-@8Vib!=tY?+h2N<vSrtjWB_Ag<EMx|Bh+e*z%X@dCO_N
z#L*d&3NC3hJa)O8F`v;N-sMhe@M$<>CLIuF-Hd#QoI2<}ELep%I(+1jgGZcJ(*Z=E
zG(oa$Y-l03ukr$K(SU`&`$bG}<_#S>1muXUxJ_X`q+z-ifNeB>yTOH@L4McbB51hi
z3p=RL;cYzCmPyPDEZ`bHnp+5B^FwE)*U^Rsl9?ZkBwouiejtTmn>?7lxy?mZdbvsi
zpllLuqdY<XCx7~<f2$i`1yeaKL&6>i6fA^eRlVroAs#(`PBW*UhXZqe!7ADUunv15
z9H_;Eo=5pdyerrQJuR?A-(ZgG!+;9wxfd>^37K#)L3ViObdj!y9{*Ta4-0txEeRy{
z$M7>6e}gejgWe4!e4@kj>h19{%lJhI&0^RhZMl0Ku!OI$IUZD(j{UQ6V2<nUIezrF
z=zTaq*VfZj*edtj^kJ*(R!=+k8`Jrx_A%j>dChTv`}igPF+JTFUcW{+;_dI~M%=P&
zM$c&NANl#NC3~4BW4%S5K<_tB*;??2s@=Z0PsESgW^}i3TP?2d54lG;?LMqQ`M(6;
zy9;_6f>JJvw2L$qAA1>(?fa|sz<Arp1)9Zp4JBBl0g{GYE9eOlhCmzZE+HdKJS%?r
zlg}O&!{fI`G+(e&V;m1FnDg9ov>@vUdE@H-o;MOSw~=`+BIuY(SfIwwi{awu4KL;x
zvZE)3^Ox;B$ga9b+>$2?&TZZ<8@$JV*vv;X2I6E69^TA1pklK??)jzC6ceX0E@sOx
zr}0}#8=H>rxy@7NwfVM>=X>y0I<+PPFXoV>?+qNyqp%6j={DI!OP&Lg)Hr1EoUjAu
zBT7h~O`ANW4Wu1-9ENQ_^Smt1`?M6{(F5TgvsN7Pqj3D_W@C%Dh&H`8z^30bHhv@~
z$FY6?=45zxdO4h5-VN_BZikceo8j$yj%Q=zNAac4_06qe)1Rx5T|Amj-o1M_{N-Q%
z)$nJ3_UE$ydU*Se;|3`&HgWh48V^Tjta@{OHJqHh8{Yiz{qXk952AfLynFlB$28`<
zU(BsplW7}Xnx|h~U-~l!)Hyc2{KlUQ^#0s{=$U7`#!$DoSl80I%OAhy;a>O2S{(BO
zHY47io)2f2d{->%K=e47J8(20n`^i7XFlu`f37&`hkDFLN9OjV>+jV|C-+Lj8(*wr
zAd>=Jk31LG97H%arud{tj-<Txrq#CMv({q1inAqN`+Iy+B%c|u%#^pA2ze8cnBU!|
zT@j9q9GS<)3jE0ra<EQBnm*Y^@)37?dviEo4Pj%=GGFi@@SVy`aoDKg4JFFzQg}{|
zA`RY1;uI?C7wv#}Swlk~vK7y6%A*wlYeUc*@GIV(%J##3beOdv>Nu8k?>tgI)OX6F
z>UC%qY{Eoen_>6?$2X3)7#a9pI(3H^_xMiT;lb|k>h;TEk9Ak|0epsKQ)R%Bn|#7$
zYlDr<^?_pv(JO$w=#RBN)~4V|c@lR+^lW0{W}Os0v9b7^H68t^EBTh!tv41=FRq67
z@6Wwfi(WQ1(Cwx-St%#l)aB)s!z!<pKiV{Ti#s;{(14rpbGzs!J229<ye`MX9X-pn
z6poF{e&e8toQph`;{p%4VY=WqEc0wGn`Qb{JOZENJqL$99J~OJ@#*1nbYIG+U^PAM
zYz2t)Yxyq&@ykr`dA<+G^azaK_`>5S;f!uqSt>r;6X4g4KouLbgIaiK)i~j!x(eDH
z`Kx1c27)w?1;%!oFgyqjVGAY<hc2v0(yPmD8baY0f<(cX1~+!&1<q1MP9VGqw|9)k
z&@;SRDzYm*8!X!;bND4v(x^P!GTy2ACb+neL)@vO#+Su|#z%ez2)xd^cW^Mge))2E
z#fI0>p%<<hJb4k1K0=Y(eLyC0ff_XNgD4wb7;e_O0ck+U&5Z^9^Gn^x&(DSnzwUnS
zH@<iU{7yGAHe=qMoOrOoF=`C-*zDTd-}f6`uU@_wUcY`dy#D5!;l<GlKheX{g}#YM
z9v0@UEAA$e&h<r3@?A8e`2NM??~urb&I7mBfs&6Js3r60Hx-b)`j*jnGGM`B9*&J3
za<esc(~-kQew^pqjzfU9`F+D~x~stfH#O4qTU=L{!?kWYe$z`gD@F5J;XXou4HFFz
z82m60;wF`W4o`4!j`W@N?C*k!4!!_9xT(l;S)n+;;UmKbMzaBxH?TYaf5zZC{@g$q
zurbJ1ebJ8rMCw(JY#{Fp!n5I}_`dmv-ukDGJqex-EvG~J4BkC3<7pQS<|PXQ0nRR0
zsk%Hki`yURMIKEO*4iLn5uG<Njf_#g@+@5|+u(>#+LVl6aYq620~usd79e>_>x1=2
zdGL|Xytbds#(Z0iO`s1NI9&Vbu+!5sKk>yARHvsWehZJc?>JiWUIS>3C3wbDZQ{*$
z*4PxQr?G6Rd?dFwjr_zFx@3UHz?M3X&N6`r9vPVcG6^Fq!iz__?GT-E?(+5~m}p%y
z2~rqoLS~)Ifw5bG<lQ_~W!<JVxcZ>Mt@S2==l0&#u)DwOr?x!Ff;9}KUze>+>JIfQ
z+sM=pXkACt?=YAnPtaHx<WRb<H}Wdi-t6S<PEQV9j;-w>3sGdz@8QbZQVN0(zs8e&
zL!8Jj3&j|G45h%ERC1f8F7DgSL(szPE<W(4NmU#I8XdN2kW+5*;mC5E)_yL{u6qhj
zaKy(_Wcw9hYF#WF44Zol5V#-IZxJmFgt(i9K-Q1b1&CqFCam0`$TLEXhW6!PxtrAz
zbN$h;?upJ*hv#pDRs2k<i)n)Q8$XVryvQBA1u;uGB6o{k56KYMb-=vxOPwH;Iz!fI
z4dj7E6!8;-)H875V+|^twDH4_Op-0rh~8<MA8E*;Fyht3#CkHHW?G)i3rh1cgBC{m
zGWbOid4z?Z<0B)y;H|$iu4aZMr0I#Lw&pyG9suh&2Hfyh7)TraTZ4AqpfNIGzBF0v
ztT5%l@-+E;FBZ;qig56zwl>?(EKG;PIAz4erwj2?tcjEME6{~9vVKBpBy^-*9a7@U
z?|9fIYoa92XvHK(Cf={}jjw<EAN=0ZQHPPDpzzH*gNpMQm`3?a*-UXh7q`OZ6szLR
zapj-mfk(qIv+$9oY{3G)gFcSKRTmcgGgyT$e}(noJ+3b&NJ2L&y+%7iqw8_yU!ZI7
zP*{)e!{nQAT@)966KyF@=P!3}7V(QT@;N*(;T;5K^cUTvpS0mAjIoTa^0)p)(8up-
z<(lDS#M*f3dd0`|@wzxYuIxGfA)FaEKfO?e2k<Y)FJp?%pT6=Dddw``{+90OZ+ctQ
z-->vh8^BrKhiF&%{q{qUi>9OVJ++YlzTdWcQQ?ZVcjGlK|K^8+_8dC_xT3%E2iKq?
z%9O0(3*mC#(EB}Svpe|kl)vcgYS>0oreN}yM2@{chgJT<6(p>Sy@GBP4k%pR9}Unn
zQ~HP@jh4y-G<`al0B-6VBt;)n{!lh#Gf8Rk_k7Dtxt;*L>rff}kelx#@o=b*IsB-3
z@%`=a@$PcCzd6@@<yr%yThBZE5Fc|&&7DXBe%>S?u9CCAOA9F=gf`CR5O~u|b12Ou
zT+adOHuG1{U1j^b;hLi`cg@^3-&W2>2lKdA$0Qdrd9Gz1S;%KD%Ull~`$0tJdzvS)
z8G_ttPt0p2i*WbiO9q%rA&ci!N}G9d-rS+AqF)TUW4(q*<{`sc1nGGWuJrs+CGpUm
z_<3_n@*xX;_u1f;-^}l_@r6FUjsitXi!|Yx2e<f?Lgun!<TRnay1w(<SU;Rx4DZix
zhx6<E;ZpO=3*qlHALrpy(q<Epc`@;si<e;z&RWuU-~D3vi@*BW@ZHaU;l+C9*BnvE
zM$q2Dz8CVTHy0PY)%8X+KMW^tei+XA7TLvxH&^&R6B}Mnln+0EN?lQzv+>27T+w?@
zWnn{%x4o=G85BVIWQ~b>Mm?SS6x=}i(aeEC<)N`B!}%p^V(+~PanAR>u5VqwVkb}x
z%3Av4$7T|5?PTLCb9D@KybZ;sAi#QsqKp79>WCjimIv8cv-4*fc(7GGh)+FaJ%o)W
z^1rpYJ{;`r`0X#2T58U2p5#?&u%^N$61ikO3Ay}muM8VmtTlM!$7?8TCbE8^G7_)X
zua1Y~!`<Od?U>^*@5F~S8Qw;!jV;2w&Lx?sW3(4Gi+FGrxyd_tj2|2a$NWNm7l1;l
z3|uy{{b01{;GeoF9<(FTd*ch6_DOwrJ*NG|n3s&(kGadVu#|apl(w$+LHerCq{X@r
zZ#5E$_17PL`*L`3lsd{Afd1@^`Vs1`j~{fIs11rI{eT}dr|cv<^$}UnS7eb4&)E#m
z8r+Wbbab#c92{_zXFfUNwkPqub|_xw7uP;E^bK!Ss{N_?x!t{gcj5=TsUH9vd*rJi
z?)WEs<=svKy0D3EiaUi@zY2uMqCa-(-4w6z@;5tv=;Y?Vf)C(04}|ry%y~@kP<-sg
zu!7ZmR^eC2FL2ny!6Q7zV-BCA{Tg1tSJ8{~Yxyq&@ykr`dA{ivz6QMU^?Sen2R~Jl
za00q7*{0&q2o1v7MOdRV$bvp`<HtCh0AiKD#R(tWiPO+{!C=K}l3*IS-Rg(T+i@WT
zVY-An54anyQH_4b^M!xtBPO|pg3cj+@U?(rI=P%J!m*(u-0slW_y-~;zR_ddnhzMG
z84!lTh&Z_92CH6<4sTx_;L8`UWRBh8vZ2Hgf$+)`P3iodOk+dZ@b=B(Q*J2S)aZcI
z;kLnm1_Jl$)Oi!;;!-#7^9yf!@s<}0^nPMSojslYIp6eBIGZ^;+dT2`YWT*RTt~yv
z;nA?a&qqdi$|Qq9-#}U(xp1Sko!ZbA0wb*r8FCy+&k<1$qjbU}w+BeU`<7&U0Y1}N
zAF^c_+<2ovywNE)SL=(R2s)67jxwmkSf|3H8#Xt*!KFBP)jfk*`!$G#o*P+hkioCs
z_R0VTJ<Eumup8;}T6(>fKG7{@;_M?4g9Z&O7~q3saEk3G)!IfxHV-*>IB#Zgb4E`d
z$nwM*gIWw5UJTBBBZ;=WF{QA4ziHhYP=3W;<>ebc0|lOTQ#w4w<{L4C2Jy;UVBCO-
zn+ZeGWCLuym5FHK(Vl>VkL!WuplmZhuR%IDWCq1+qGKY$n_-y@l-z94sjgag3>+v2
z`4v)rjzHkeH}uF8SPVS*E*b+uUbVkq^AY{C;iZ89Z$S~4s`gZ}WdrYtX?!fD3~k|7
zX>RUn5TiQHx6E9|q(i+@J$7!_*yzjKOleQ|9%wy#sytIuMMPQg{VdX@KA;9BIWVFg
z5ts7wz)X2YCmE;^H}%Z&I$@s30*3}@7eMMVZNZyl3V$j+-!mw5{p91ls<ZdRrwE=5
z+q*nb#D=suz{KrBfuv0x;VC8buLkN#sPd1FopIMc(p0d7#JIhfx$>m+7+|ldj5+FX
zcV|x%I-c?pdq~I^Y0~)d5GPesZVdS5rU9F8yf{HQT9Gs!Xu~w_KXBkCyg3ZA{V&>P
zMl1MoS3;A2vbd0Tm4btuFafxOGsxp0VffJ&k=ec$2RTC{+c!H%pa8F{a83<Eibj^}
zm<j!t!2wS!^dPr+C9Uuv99-xf4jukRubSvE(+KCmq~jEhYxtXI3#<6{>0&bQxQuYl
zWhw)ibVdeH&Y^$)!%O*PQ;kUteFlnBP-v;6#Dz{)aW4xzN2^#q{=}zkDTtb7dEC~*
zdz1rC#y2?~gMjp7!KaThZ$aJDY$3DpiC^xN0(?k|2^o2i?fx#jt!t${%58=fM+SEp
zaQsBa&waRlPRG2GK5QH=hsQM-{Q?Q^Cw`=f<vQeS*<*|9m)Hs@(e(RdnZqmbNW1im
z@x}4YBZRR99(PiX{0c8a|3g}Mp^;1Or^6rr$)Eh%HopG;KUjU#q4Tx?C42MEV9eKN
zU?q+Fl!xPkT^0TiZWXMeEvT@E!hwfqdi*>LFwT8_;34`st`7tHu!6#Y&gk$V{$GQ8
zh{wZtU=?1z1$>3|_8i~)3+DJf9Oo3)$D@5Q2BUrT^w5Iq@x`C}|L4kT5!Z(;`l;0F
zt;*lk@ewzR^B8`PTkyAhgAQ&U-@zk*o5Lb~Obf|xj-UCt&S`poaGgFD!tI;Ff*AeA
z&HNZjYL26=*)D%a<NI;LSzxRd7X0M4*YeJ3<Zt)n3Nx;S$B&(~Cfr0ry=aKh7~`mu
zKeWvb!tRlea9B3J9X7facY$`qO>zadSSCREK}~eQg*~{Ui3!Yc81c6)pY_&gWsP*s
z8~zUG%uRm(jGtf(Yi7drd|<OkV@&hMkXln*2YvV)Kt17`p9H!*tFGW=&cfR+%%?T4
z^4nm1cZ-b!7KoVxYc8WWPZgJWF9V0HUpVItTEb=7&Rm{(67<A(y83xO)8?R_bFOdN
z)SQ{QXFU+>?;blLU8obx?L0rq+J=uj<e@&{AY%^nLGx99HD5%Wn%7ab_nMb6Cquqk
zxAEMOIlg|5UG0;UJ@2zqIP%3hMB+h%yxG8uVGfOap66>W&-~kST+?OlU*EopE|Xqw
zfPS9tv8u+W)Kc=r1_;Lwo}Kenm*&lyj(+k++nVNjnlocIHZ;e~TTIMf`4&8<1#wc)
z$;rv^=FJbo4?nya-oAV12T<AYVnE4y#Qwp-u)DX1%7^odv*G=_H^a%>ABMBjli^DK
zJI&G28QMYzwXh$RQC7MtAFgtgy8_T7`L+I>?-=4xmL7jXhYnyPD4X)N9%Tq8zHx2+
z1=%mJZoN+CH>Ff}+1TNG=ZfGoNr*W%C(ocqQc3xz49ElJ1Xu<Vr>>GWWbm6q%oWHS
z%LmB|^$BA=Aj=&OPTzWM3Bvu|&Ee=^TlFG?b-w#We$h2m%ctllv82J~$`h|sP{yvS
zDm&9P8?n4LrZ&2}Bi=`Qmdi(3as(f14$z|Kwvh!rn~?A$zuP-o{;oe8y~soRaQhY{
z?^OmV6ATcZEF$7o`{XB~LK}dR_(X9(F8z6ogMMT?K;h(>c+g`O$3O~$rmgeXEkoX@
z$9~i9_EtLO$jhu*F@N)S`lUBNaIHGZL)AR|?zO<S9_aNP(E+44>Q)Z)L;10ONIuz2
zJ2==MUcSs{Y*_cr5tck$PJfk$j&JWj4cEM3$NJahmEVRueSd1bQ(wp@`7cO5<Dc?e
zy0(*mE^MN!xT!;dF^sy2y+9k7$H$B~jNJA&zH!?*(gD4@;K6V?m+4pMD-d^3i}9BE
z1U|=q4i0-b*zj2VIeZE2*YE?rh+ZTMW0ij?fJd1LKF_;v`CkUv`1(WM_@V<)=krO4
z@Fy)YB+FMJrYbHlL2!BdE5FE)Fg$5!+<aO%Fb2orE<OJiAj()WxTm<W8xA7#4c?;5
zv?gOSHgsFT8P)vZ(cDJTS<2>a*~wgSino78HhKbOC+<WOXyHvCd9ojJU7JCK(35WK
z#AO5Q{2%Tg_((&(m&LccSTtu7gbv>WCf!Krq3M{BvE*<{t;7D*HU?Of_RT{bI5#dj
z*c;uD*o5J`N4ze6c7Ez7X?W|1<J>N<`RHj3oiqag7PDWxcxk!V@Y>ng^%D*mP;rxC
za?A~#TvR?~gs2Uz)<?SY=Eh<9BDxIvl<j&FAg+}o12>%39`VSdj0dY85a3T<ts~u7
zL>N6OjGLx6QLL8?9I|Nut!~uvXX8=y(8#b!Xd3hlH+aKG^mR-h2EF=C8|kiY_T2c8
zfty~PhwS6@6u)ld=wy5gNcv!)x~>5vn@{M@PlE8~lsB!Wff$31L4z6})1ZMEYT{<k
zagy+0V@c7icjE9=3Vf(ryphA;gX0x>6AONRN=AbbHn05V7v%#F23=;9fduk$^P~Kb
z*JUOeWXl3An<yz;-y|t;8JAm&Pdsnti9deV1>EFGfWD}|9%!NOR?d_;dMBKxUl>%N
zM)b+6|Ing$$^{w}03Y=0%?hQ@(|zcI;~O@%G*I8(8IF%$3|l)pj*qNcJT0}qV;vy3
zq1~8Wr^FMTAv<*ixx8>7{<Qc0i8<c7%0{#56ZK0reGPGSk)F`2bftvIBQLQf@2x(|
zWFn^k=ON|fg<<)r)3Ea3Si1LSEp<-j_EL3?jW30BHv85dAI05r9dTbmomMb|YtE?8
zU@2)MKgB^FNSdebJn10aHdrOV6EqoQuxwBG(CUWIaNlBbNaB!U*_kvvIf*xD2;7lH
zUIR>R(-I)`6_MMe!sAW9LHy8#SQq~0D~IC&rqC5UL0udHdA7Tqe`pC>T+k+SkX%jX
z@hxWnenR_=DC6J&T36-Mn+4EGzGNN0F=mPW&BICo<OVm&fM0|y&!gXl_F{~~Tqkh~
zLPI+&(!_H*lr3@eulj|jSux}XoN0w3t{kC>qvR+a38J2dmQ5d!gg8t;Z-P3k@(qEm
z#<`S?gHHxtErQ9o1n>Alj`G_G6+#)xNV3QuJ{W09W%rRV+7a?OQfOkMGuiyiC*?+6
zCY-x)c;y%u!jk9Up?5fb=y5Bid5bQ^+T>T5YYXy?e4#EIlJJRKM#L{X8_+3pjH#`U
z=5bze<$jYm29q?tqg#?;8vzbFg;%}=RuO^bC8sm*zw_hvj^CF~-ZG7G!kMU=0H<+C
zo3a{)KmOyt-Hk6GGOdD^hX=rUeIy)MqUjQv$DPwI#TD4^&MeT=tfX!BJPauO-1Jbm
zRr=}UU>0!pEkaLMuoSNgUxgEP#6>2;VwUM=e(Eu}hOcm0BNY7@r$LYJ!ydzz($Z(7
z9|OILc5ZrjD0~rT)aa)lm==MS=?i+?WB7-{;-^38Xk;zWNDsY##6vps19RMx-+IJM
zVS%Dw#Fu++R)zJr-d%n<7IAHiuz)jJ%6uyK@;g*^XF9*34{NyIKjFfYU;Gu$eJW#b
zK$tE}aWM;NV)yZj|Kx|~GOpqxNBNg&Ox4`b`)|L=$FMTvy%xV?cl7jqtKt<>fWJ9B
zUMqJG3vs)!-rt8~s=w)RfjMrOU-0F}w%4#o)Aec{K4AqtE^#V!fodsU091vQX<gM4
z%=zXdhKdg+cIMOJ;yDZ70^>NuTfTvPJv{LSn8sw()kL%($#XkH2He6Sk8E!$F@JA!
z^+xD9R@NW#0A@Caa+Dr(Rr3<57p8HDAAMnMlJU?db7;2_f+Z8*OMWhcWLa-u@jB}X
zo+ml!v~?cvgdftw9ywX)W?ssA0P6)=AW!@{vL5ibhkhAu&SU0dk&jIgHsW|=EXS^~
zPE>0v<c&FY9qH)#br^Y$2_F!+(M=v`Wj!Hl1gtrr1EzQ^>~q==Zv=7V+U=d>P;8$_
zq&elwqy6E<;cnhE+E{ZsoFa5`^4{x9=Q5Y)IY}sMFg(P`+<#4T?_F<tF)zQqx%3-e
zrzdQ7oeuXmSKbs=#+{z;Q_34vf(KX01nW1ksV(SRWn&tZt4Yv2$uTac-r&wX58dK+
z{j8tE=)a<4zjY(Gb*~>Adk31|K9kOP1CCQ>&>?Fz2m9N@uEI$~DC>@*B)sY#^^<uu
z=}=BCOQ*-f&Ps#%v}8k0&#y(tS^$V_e34ff2$Vm|#@k7fgLOSx&ED?%@Zw-wb%Sq*
zlU_E-^p8Hw3kIJ&|EJF2bDgAavJfH8l^%fH)KyMS*<mwgd&lcAIk~8<X{q98{j5Fw
ztu%Ps&L{E+&rzX#5+!d-sgAX_0g#Tw6G&IWp&QGfc)aO`EY7gRi9{`)(y$@jCMy)=
zGkK;~=qH}F77A6?Hbe(?@xV`>W$U80i0){ox5LE+Cj^}j?@rE!i_2TJ0UoNp^^uq0
zUDwafhf_|@QeCI5aN-c{`CffXz41pLQhrX^HnPzE(bM^r+L-v#UN>o9TN}gcW4<%C
zt1_jHQOCBt(Y&$AciPb56MssD&(ydbD6i;}{1=S67!Cf;5QD$mMPCqi=nCd^*o)yQ
zA0YUoY5$08wh<E!SO}lN9Jdfpx|m>+X@O?up1diOGCf@ni)cKHdGEpa)j<pJQ2aUk
ztEMaXB6^X0E#HR$KGGX?RXodl9w?iygZ9Q(jyg=DT<LJ~ix0*PD?zwfxDpUWkj*mr
zMP{4@#x^=g5Y3a!gr*iC`Pq&&N<h-~%Oo(CfsR6+gm*GGoF%s#0?<YZ8l~Ow6$DoS
z3Zbnkans<pNxHh)UmfZQoy?9Wz;6;qDny)q8$>QTs{#t2oAs6k5ywYI!^>AMbQ3-B
zlQVu|K|rVJMQ}PzI#73<#BF>)*`_n)M#RkklaAMC5A!w`Z@6&0;rWFJB<E*m!<oPB
zm5+Y<`%?<zNfzYX-`yWx9={llB-eokD1KW=H*f|@$zwWimp6Z$R;09Hfucj}&R(8!
z(ZGb8yKk^W!{$bmn^D~4Q+9ir%HVwI2B<J>A4{Mci}m6g3-lR43TYlWx`22ebTi6v
zbI|#9d2Yzuv_0U|0EPhqX~@lB)ZuMlV?DNk?ftFvDP9Z&;K_!UZ*+3A;pIUHaXpYo
zy5L-&bVK9qIY~FHYvAXne!PL00V{9HWYEdrny0lG*lO_Io-)Be>u+f3#>b!>`Pq<w
zU%es5$4GsCu*%D?WGl=YV3eI`{Kgeehl$4ZSq8o}Ai<CBJqTA>S#}ixHszOr_Y?7r
zEZIPT4@BzBBm*=6?0qv<-aV+1L7$KsPg?qUzyZ#@m@KHBhzD)O^#KE&%8!q#0^siC
zw{-Mp8j$kV5d%%PF=SD?HNdk@H8}TEW`~EvkAL!$;o#^HL43p^ZIH9gpGbeCuc4s_
zxs(^jD|}G-!Y_HXgrgj8$4bK+licW=KCUk>hxhN^$;k(T8I&s*YMaS>Hms;y)ORsb
z{iAMYpu`3s3iQN8I!zwPUv$P%1?wtDIZewzm!lP~H8}HekP`3N^C!dhPR^&_*b1$Z
zah{pbAtv%uN8JW#xRQ&-1h+Hs9dA(RM79^T6-K+`$w|VX-(26+#DJp?r9=5iYlNu7
z8y&4SHDn`7CmMqYc6T^|DP7J&;E`>o$<RrK86ha%GWem#0}X%U{LjU*$tXTil<S7@
z#k|cCFM!9qTiS{T&B(X;jc7>kr{QwdDSCyq`Vy)buGTlMh8*pdt^<*qJV#=O=Q8n+
z0dK79r*Zg0nl=+mkT?Z!wT${F*%nvz0w{m#jp~NXgo}wgw8kMbaocs;uu)w2LZie?
zF9*1hCW*#B;=#3eGJaz$NTZ;!e6Guw<i!_shdu}cr)!^26Jww;zh+mN5tbNg<w=Wv
zc%-x4K_#4j!p$nQfGs0i*oeo&AHQQHyvh?Dm@b#X={lx`4y>ydS60(c{B{#2o{s4-
z;jKH|EpKKK8WZ$a!l9MhQF5_uK8<(b2!_@U{H5#m`=kFq*!Th}AFCkwSRDrR;SbT7
zMnC%G9v0%tq|Pt$=xKn_-|?#ORkQ^aHm8}xDw-ZwW-%v=Y00xd+o#s!%ilvER#3PJ
zKlzR8@C#x1C8%9R+aU2`k73RfG!t$N>vFt^p8&thFh2V!CdelG@G1X<pP>AYh0k#X
zjeAVdl|8F7y;+4TJg^#9W;s4%nK=%W`&aB9Kck!D3g+R$<O<GxCGL(69O*Th9_D_a
z;`Q#4uJaeJ_ZuhuF8G&dj2ivDJO0TXzfnuL-XGlD9~w#kV_cc$pVL0%Z>Skg$b~ZG
zrm>iJOpgnXies-li+oE~Ou^(|)~y+YiCqP$3-D`USoO0XYElanF8IkGz@Ea5Z$S$<
zBUkAFSjJ&3(scYx-;fQ)ux5e_o}`Tl!e(sh9nScM?xE+opqH|7$XK843M{O7G$L@k
zoyFYV^YnY>%FNT5oAbcno#yg)8ry!<xS6!_)|il*FD4FBGFS2(Omj5^fPdyT%pWtK
z%tjFNLgrx1?^u}EJWKPNY3%_X%uzFsW8O$S!W=FG&Kn|1g9Ul!N7-ON{w(@4H)5U#
zEk_xWZc|6%uns}mlG|$vn$P+HPTBRaBlBcG$fP+ik|lo1YSN!*#5)GL$Q$_=Uh^`~
zgFS~<eiaWn4WfbGM?T81PO!bR>kW;K&269d!aNuHEXwoQfLkB8#@Yrr@<DpM>BW7F
zPdo5Af$w>%>n(3}@!hU`Cmv}zJ%)KVrvq`sUw_Qw#rf&*_RaUh#o4Ln*^Htyij}A?
zI}GKiylL)>?pWtS&*&1o^TYav?C7T;YYWi}`a_#mqW#QQ{LmR#hhZyy!jvv$?Rl`u
zh0TZ;hezIQM3uY|1Z-(ueSEk-9LY!}SqHjLo$<Pq#6d37rCuv#`KgQKEAwoxe;_9e
zmA-M*4~NBVUZO!4@S}XG^XS9JY%1#E-sbT7<(}7O*vJA;UY$qyDLzR#E!VGX0Q;1j
zd)6S*?nwu^X(!1?>M17=QCFy+@FtEQN``5xyOLHMoP@)u2h<y6W`mYGKwV&+z-3Aq
zsGX&*lDD{tZ+^0U-)Lp#yt{mOTT`jghMR7@ria^dPa8q_^G|-@UHX(WdKaK)`<$-$
zsvV2W8&hZJSE|<+{?6UyHQ!vj8O~1G47>6J?zG`-9?~a77qp?Q?NElad+?;IK0>nP
z;dI_YU2~eOD_&lldjsu*+8y8WIy%@NUcEY2{W=`Bw>G@-mG6Y{AorTu9xpPSn|A^E
z4n${|au*by97~~ug6(MGu^11%7Zk4e#9irDc#LUAT(jde1U>O7=>7AsUKU_6yruaF
z%{;7!hk_<Nj+j+^j9W%8!DBMa;Y(@1h9~ev^dk9Mz7GR@R_pC?a83Ox@W$8g|K1<|
zRH-@Huu!EZL&gNMSeM4%#1k!7@Foe^-67yh-0DWegHXC1r_s`_%$mwLW8Dd30}V6!
z&AYkP-Ql&1##-`&TILPpYMVcJ*~t^~RpU-rBg11HP_Bd%zoloH0||=Xyyb?s^^A~>
zF`at7lFp_EA2!_Qn^`Yj4#zKE47<9i5awv^gdNS9s?7oOuC&}iw+>wNCIP2TmW?03
z3e8b&H@u2_JzTg`RcC#{_l?-_x-yOpFTbX%4jlPjy_7u1l4oz%0~KUqAQ_q3&BHFB
z=|dN9?jeNSbiVHMbps)s&OQUZ$V(joa$~^Gje(nK28Z0l+D*;*$<0Gd6qZ3EdGy8s
zeg=*Vn296)`KV+D(hSV7iRX<i>4m4R?&Z!o(DEzq9CyggE8ptk2F6B~27Xt3B$d~i
zud~@D<G19bQ-4cJX>%jxM#o@}vLdgXcZ{4U2bmZMFqqof-geouA?3HX*wE6Tz;4~x
zYlDh`2Afw51lVNS<cS%@A)LVi<-!2MACc8y5<T;MEpK#)7f0B!QNmG~E)&VYhd7_F
zGvGn(8f1%x!3cxV=WXEOC*DNE5q2Jgx3|nP=+J=B^%%M~Nf3>XNt8S3Feqo>*f!)$
zL%T@baUE7T)qo56V`GY^*qk?oXG6{7HPh3g7*bImpeG>)Tnrwwu_hd{BD2y3mvIaq
z98R0d5so<)!j(?4GLYTi<D#73zPUbdj8|<w5#4~GCKpclOk9x0ES!c7c~iP<EK+V9
z;c<O+>9X)3TJ^+7Y7mEZtGZEx1&(L*n}O70>b!8#BLhLF;v_O4C%+1_PI>Z61M%B@
z=gga{icj-m6K`{eH{58$lzIAPZ)}pj+_W<`dQ%Q5H*|>{gmFYB{W$c<jE)V^D{ab?
zzHVD}=wNSu*jnFIVC1z1DZ6-$O+Tp+SwZ8lMnDG~fVSm;MMD|Q)!nv2>2KiQ!s6<A
z3vgRx9HKD4=5B$|8ZVqT|0r*e;Dtl~{Kh_*Ku5CIa#$upM4vw2UdvA`&`t!9+uqJS
z+Pb>FH<P^ZOPGBwcOw$!dr{g|QR6fZ{Zvqv(cvM<#de%r5>!aJTO!W3oP?Pt{DLqv
zpTj6K(>Sfk?Ryq(Ys18@J7(c<)uI%X+Zei{Zdd&9b1un_{qQuJi=d_MvLJ44=1uZK
z4^`@JHqIb%g^FZK*TrYXRQJf>Ah{DJ<G9p({Lo6C2xbCb8;|DEKXr{u<hP*m>Nm=$
zJQ{~nKX}-VA0bN~7fXvDU$a_xAaNmaUK(|B8`nhL+;wrHEk2OSI6@zDa)!oX;3|)J
zeQ8udjA60EF>xpSAN}zk|JpXb{_THn`o`BPh)7LVXl8Ve#RV$tK{{Bv9!)f(sX7v}
z5#Phy?=qKBz(imDSce<KX7CWchsVMzewne(HnRW^;LBfOefTOIunLcB#C7=6bq_`7
zt^pVu{+#cGPZ<3LkqQ?W`4O+ht#Ld+Ii<e}EYMfm7WDXpEy6r}Zi+^i(Z)E1!nZmj
zmuV7jb)6}g<11|PN0FdchsV%2uL>ji-f!HTrulW5Opr2jJ0|fK^VYjp`B&kaNB^`n
zOs`W+=#sxN9{0e`@94x6j5N(%(7S=6D^ogL#I<<Mv@XC98p=F;=W@qC;pX8O^EP@)
zbI~7n@;^pv{>SJS{gsc=$nTmO0@L@h#y=UQ)|?fG_9|Nd?{vNK+vV^w;GKKBD_*&K
z93Wr(6H5Nj$6cMv1MwI9D`>EAFBnQ+GCFLciPQqQ9e=cO_GMhg0iz+DIukyNvXs#!
zCe)_vVSzpvyPHct=2n`&+-bgkdpUfP%^P8wplHr=qe;?@#)iBJ#9Tge7MG~>=m!IN
z%Sm$==D>dAiMg!iJ)Fit-DdpG+$eLTEJ6n~786Z26jaWw9n?IYb%1QNseYI)Yden4
ze2}>k^EQm+@@9Yxa&a7>A2O}BC7u|dbisOu*Dgp$L7q>OALg;jgAMm^f!7%%i#JBv
zJQevdIEjx|kiGajJ;(&`Wd6%K!uIa&u%|iOt_&L*e(+ZMvv`i9IU+_sO<Ti4Nuej^
zx#t&G!<#p6hoAlB&xY^5`{i(Q!UNS8-Uy&HnOpDd?hl(=+ny6s7UYX{0M?sMPu>lu
z@81n~H!5o;u}aHhYVoF31GvY0DDUVfif<r!6y)^-=EVS;7Wo_nn~<Fj3@Hb%6ZoF1
zbQI>fh=Zsj=$E+R&6*HvW82Edi=$)5WBupq_Qv;YzR&dS>*HZ-^I6u}F8SUP%v27l
z2UXARRX=!($unlxwbVD{MyIJ8Wbe~(M?8Lt%iP<WqT*3+kXdf=aXrv{eqB5ccDIIC
z$12OM4X<tGv>{?imNtGNP4r=AfHO^Lp(FI2bs6#>U8ww7*Fq->piC*lyy-~YNL!%p
zIfl|9e)={(*HCL=dHV@`Ho(}pQ9EVRi}HSwjXc(A+z!c~WFyUz%XOplCO<L*(NlC4
zh#P#(5l!lF$p?v$jza*l+6X6kjibFfAIT5jG`p5g+4S<8kTPtDQKnAUtfNPM<cYdb
zV>;!8;xesI0#Z6`Q1iJMzK#C!aBn!2%>(Kp_Hl+BEqZl}5Uj5~k**>Zp9EoTk+s~j
z_a|y^*Xs6_SGSeirzWFFO#+v@z*PZ8_-#QG+KS)%qeB_RnfOj-g7i6o;sgJQX2dz+
z@;BTHI-qw0{kXYck!CK}GLN~xhlj!@Ja)_~TH}_{BT)YyO~2U%kI@2zm-$-wqBtV?
zRd}xk`1~rey48Lac;o9I{=pyoG>Nt=i*TID+#W-hD5_Cu^mGj_oaPS1$IredxG7B1
zDm{FKdn(>J9<fm2lxV_3+GvBK3k9FBM8q^64wD0)BP%_%V236kLHtG(7d%SGxJJ)a
z7O6YHfmqlHC_Y?bH=Oe+ce|1B^S=>cqeJVX`RQaBh;e)&-|jj-J{}Ga5B(<9q0HXy
zo_F%o>8FFIlc~-gZWh*#fH#D6gZGU|eQI^6ES}%q(K+7^x9X(Xr19pK%+=+k&xO6x
zpo8;?u`ktm@;W=|?r^N(@k<%q#J9G5Q%aJ$x{y;ZT7g?O-UPI)<eQtu7{<+0{x*of
z9a$L&$gP{8H@C0}*Wi!0r1-8@4FW^YMp&MT@`p<0_8?PqY*b~i7+t`dL8lk<i&kzQ
zE1(;-H@wV;4F)#6GFXL{Fa`@e;dIB7XNt$B){O=-4AQ)T=<pcy=c5bA7n^9L+tNn`
z<PjbyOeDSmvN@Hrj&*~Qj<6|nA89CEv)RL-09@Yg$smP#z;T7t5pR-d@CqF_NpI-b
zCjN8N5k`5SKPoF5XdWoYWKco+{8CmmAWa3METPQ+KO0*bl!*^FLF2@On|N;Sxv`SA
z@ru8`&J$=1qVt3f13Na$d=x>8=hxlEi*(l{GjE)+d1i%4jBNNQ-6mruBr@oObX+B&
zWk8U@5X?E!51x*za9SbZh4a%zl8d$i6E?dzHgZsOjy>b3LqAO=8pl@|dO{;ZoknW_
zWeX0PY#1R!(qIB+`FN91v^-Tu-fObuJ^_6yE_x%2uA2-F)eg+Fl@%tFCXDEn7G+&+
z$^#{}K{j+RFH}d>mcf(6GnK>6-nKVFNe8}ZZ%mqK8$8vQ4PE5(M(lcyxgZ=K*(6Nz
zlv^?u6q7v0bJWM3jjdsyH@;N&kkl6#G;o$>dwfW#kpC_q<!p>0E@%>7A(OvquiOsA
zPie!OK2xEMckUC-+f0Som<T&EP%il4W`MTgB6sly4-N4Gs-2GYiL)UR#fD2ZpPENz
ztOs3~%i4ckj8kg(J6iIY+jhg*8N1z5_4g%i^HD4!8--ssaWD#}8G|Fw#+lFv7}#&U
zvI%H`&@*=0B<-3@?pKa2O&F<b^izC&RX+K7a?p7A2|M&mErBGFS`27Kr*O+CtOvk0
zMWJ7FM<DSMz5Pzdqz)V73i9nOxgi(M_chXTgqS$RNsdeiuw!)LFB80<WKK%9L#Os^
zd)ep`xAmGhi5Iuirhvfdnzn-6XgG8ZF}<^4euVLNp72d10%9aCKp7Q}<O3Yx4QDrD
zF~!@Ko$Y7+V8AC3`LX@OZ+i+SI5Lx`&=}cd{-ZzoHEw+UgWnt9_)0z|^EL3LaUP<1
zEWUNr+`yc!kK5ZLo!i*gLB(HX`tTm+egIsbdc`YKFvr<m5#B@L9>Z4{*U?|N<}aFn
z{QMTel7k)>`~>r`J{YI_AA4Hd3U7YnnxZBiU0A2Z)!~5${bd>yehhE8hh%t22MG(z
zaomFz{ZsfuIE;Jz9@muB<9dG&Jq<xI6OLSzJ1~XCB-|nP(<k-*7O&hZ!pkHdQy<r-
z%@3WX|0JM=>BqPYsIfNqvdfge?B10Abg?nF&R@9D<0v<t$vxt*`IR63GrHw?U>+#P
z+#u{WVKb4lO#G!g%)DMeu;6#JxgV4Jid?}HUT!Bglhb&h`P<OKn1%2_;R-L;63)qB
z&GCXcZtgcnIb(W!U=bfd7jVWm==>F)c*b<54-aO#SGTe%=-Ro%gKS0;wQ>SPVRVu4
zj2~}y@tETiz6a(xzUDN16AWAC(|!D(*c4-qrFqp;jkljZ{Wv@qBaS5GVKq%@WSH-z
zO(Nqxb2-n|ReRdl(DlJIG7)10uR-weuHyR*6U9M(9v-9`Fq3xm6?fJx_&%~fEz#Bn
zm|OFPMjkvAm2mJRzqNsZ9N8>m&cMT{l3(*R&;M8tkl`ekIu4rjBbRJlyitQp3a&Yx
z*C8wud1cn6^q7NduJ(L=Ef3%B?t0CqzR%^kxaR7}x}kaChHTIGg=e102_tNn@t`b{
zoS)^WH$I)g1H#CUVL{)YRuC<D1jihcb(Y)P>*3<;Y&buEKU`g2C@e=B_TYX_3hMjX
z98s6`APfSyUz819=E-aXWxdF2FC9Q@de?%K8R=xKZxZwO1D^J<!JzurbjW5Yb%%$H
z(bMHM-@!k3zS+R}(d*;kXn)fWk-tB^RGiG!z3!#_BNv-lY?k;9E0r0*nvWk|-e3)g
z`j>T^MfoXX<W62)PSPv7$$D4TCiZqVhU23H&BM2i<48s}Y^WIgyiTDxeAX&5$3eEJ
zF>5u^xf4>VGLj{Ap4yZnJfUGDko7N32u+)Mh=HGMiwh$qFOQG)^BSM@$hw>Af-9`z
zf0RvIse7N-l_U#tp>y+uA><;!SpH7d7_3Q8Y4^C3|I97ZW=t&boY*w1bE%D*r}#T=
z(h7O{59+etdX(+@gD#=NEx9XQp!itdjc@pbXPp8@yJGX};P7DB+dIhHmm6Cu+fT#0
z_vgb8Z{G|j@81vC9I+`K-rRl|E-tQyv-fX@>+>@)$~q%?taJ+|ck<Glc84B($s7EH
z3w_0#cs+z`4M|XZ%57gCC(b3h-Y)p8KOJxbWM2#iKF6bnhr%X&%1wx7jOtACy9%g7
zhKJH<rr<F-3%-P=pdUYdHlBiOKL4){;PtEU95wgFkd3e3|3^P1t%|OsmG~%z@Gi>c
zc4)&9Np1!+i5TNc!|7_D80I@8Mm0Qq`Y5`z&Z!}$K=2L^x&_2CPU$s;wNwO+JMuL-
z1C9`<!)3ucK7|A~Ee49#3L*sGl(#bB4UJ5t6QahulMjoIKsMb0n-K?xN5kt^uZP3q
zqnvrG8x(^G9~tHy{dDF}HOO(NN(ZX=+<f6}kbJm&qv3`s!&$y8n0r&^CT}9Ksl`Vx
zFL|r${M<(yUTUCm$(vr6&}gutaL%!Pt`2{H_rM!+`}=#kvGTSGZ!fhc8q@)M043vr
z8v_t-n7Vm;lS;JNtO|_>fR39tSn4gS42&``WB`P`Y(BBk1Pz;YAK9SNVAq>f8YDPe
zGO{V<4Jx^vmiQuDJpp7r5Z7sHu*~4XbPSvcXY(TY;b=o2k*IW_A#eA>vC+i<&2Ni|
zmvzbJm;Kq8Fuev7$jHr$4I+-;V-PT&=bN#H%1?uC4QxDk)F5(qZ^z}GVq{>X!KDVp
zY>fCAg=cvh3tC=R_mgTWZPRIR<V_9%gGmNksw_%_0bkk~10ruOQTf_Ui2B4pmH{s}
z>EJ|-Os1IzM1X5qt7_cL;mM7gGR`0v{?uvWP+!GMhC#J?$c>!j9UzY1BvYM6wi?vH
z2e<PodY+1-#i2iLpn0n)gCE)`dZvxIRVq9kG8=MuGidST#RhG-0J7LD8#jyW(0Wki
z3<^y;lJSEkP9Nw)$e(B+$XiYf=BW3YtSG)S4>U#5MtLGH<;GKIdD4*b<(Ll6VrJu!
zxEa9l<lrPD?BMPC)!InvRmqQznH*_yM>Z_GexierNfO`wV?fRK`}~bB>rgzTuO0O<
zyj_Sc-9}VbsqYL#8Gv4^j^C&)uCH-+^rm$Y0aSL>U!|8mExL7nEsy#H#aH6Pn(FP&
z*6y&sv!j8r`VL1=-T>UFtGdrz=uZ=tKpAWx!3~Wkkl@J=RJ3gF1Os1t?1<ZZj1M~L
zf(THr##;wI<*`vYKlJ^|`4q)WR1WyfKQ<2Q>!WZ+!p+=+k4w4?opIRu@%JSe7a)U!
zn;)xr^sjuFPl7|&@C)A>BSA%QaMfdbaF(AO28Z9&wyBD=3;LVJQwHkhtVsGR9qNeP
z6`rs$9efUxX*_*bt9}zX%!jb_F@Eb29MLHccqkk?OCq<!;Aw}L#cO==Q!YXSUSK}d
zd4-RYUeasz+-xCsobZn9ktIP>cx*FKUQ+JHnQg+HAIlRuq=2Hz6Ir}S%+IuqqVaM2
z5;8o3Pt4$YnbFReE(x}<O26nZPy?Y&9QzxO^!34nKK${UUyEml3_nAo^&d9^%7gBm
zGxNoiCR*%9|DXLWZG6q4<m|z+&Tu_Uagy6ru!?WND!ehXIIDQzr_B14P19({I8>+x
z3;16F=J9*~9H)!@J@j-1MO$I^H<=f39Xx~w=6GC7xaRM11&d*;b(gRLryv~dnl{$L
zj86HOzL}Au+~CvhgrAbKztZU41#?{aAEKLwmAlgJjr*S^yvYqcz5r)Ta6QdKxH-Pm
zhCR|W<miWP=3g}F<BAzL`W*XhVY_1#zA1df$C!R}hoJZy52pD?ym4igX$ZhdUq6MJ
zZtUY%$1NPN3jaCW<hOGLeidzxi{00i!cBhSEyfMb$k9CyTktoYfpVwxfCc(l9qq%3
z;BqTGcde0z`%y<*RRDi(tTwjBlqdQzUW12lW&6I_U^VT8kH2W}HQ@vkzQVZRCT|Jv
zK_N@H9v1m6(}jD$?iu}pfno_Coiom2E>RFyXyWF|CIIvNdkwiY@BVN-JYnPOBX5IU
z$uDzvF?_nc)O`O+<O&lV59Vc!qSh<g29xGK@O4XWnV{}sjA%H(a!5XJx->uYIMx?b
zRlHHerWA2JCt@wZYYeOfWJ8Mi-zIO$XkI8H@geQZ*O~V+ZwBY}3TIlnVce^^7n@$&
zoJg|4=2zyf)}3gaC_*5w<nr7X`R%RE2<267=A^6}FyH1uQ=bwenK%W-pHvWj)1Ly^
zP<osqlk_lt$hK`{5zfXUYb0!3p<gy`2(!GN%db04biwIDXQwB__2s$OeE8JFwc_2$
zjea6}WRAxwvOe%ckF3=&FP7fi1{EGfIWBd`n>O;3XUdnfQ*NOH%1v^_gWk6L`|yk`
zDIwM(@^)SL<qfKPhu+@v*3XUC(h!(8cD{MB?{%QJ?=NICU(fo97&{*If^Z*gskt+2
zWu(Vu(#|$V*r^<p59T4RqvDH{HRob_!kP=~CG{Zfz1jim2IP&pxWBjUV;5O>;G0o5
zw^O}+uDE&IkMd;W%vl}k$Vlo%ZKx?x1v0&-{ve8SEt&G*VRY<GH~AqTj@Q}R!}!Q<
zZ3~ag=z+cta`JxVZ^zvYSGW0I8v6A*iwKh6r0F)KFv=n2n*3yb)_6-7wULJ(yt6>L
zT7=XQBV{bI@vWUT<2Odd%LYE+TO8{|;Ce@VTVdoiE73w%ywHmuplv`K_Zrqy*3(!k
z;{b_3cE9^->04N@VMB|z$-F*$b>+2AKVW}zGn}1W3_rX(9p0Xt4X5YV!=?1bTIGl9
zv*8JE!UCk%WGow)+*2MXPu%?6*Wh;8+zyocA)MO}xmlo#8F3ig{2LA^+8zgt@n(3J
z#|#!}mU#gW@$UiLDttkOQEu{#_7ZeB;!L%m!Kye72<bx~lc(TIXbQfDHre|M`Y!_b
zJ+9}~u%yM$0vlgX{*V8s|Ad;UB#QwF$89W`bVCi^E+NOo8)526Z^w}JiO0@a;Defs
zfK#Yx-E-mYc{d0<$1xw$5Ru=wAk74th^u*nEwnCbuBNKEjiLV7K;zwfj%QCWTzxYs
zpB-i{P+FEPp#{mQa7~_;F^3rK_TT)(hgx1b;H}*~A8GjV<%?`|z@ym6O1uCrI#_@F
zU-n0ye8LZn{&cDt&~f9d1@mk+(NS{lt2b*1^8n>?xHxCiOGX2c%d1PrV^N(O($?1Y
zu)CX^CH(KXsi`COjm}xALBL1r-QQwDP-IgC-xOtBnW6o_d9^$Z5}JG6JiUO=&7^gb
z@It@bVBGoZhGrcozTZyZCaRwoT)`<8Ptp?yJP2;IIJi+}!;KJR5ta)#O6Nf(fgI8i
z`2)B?-)Ufs-vbTw=hv|rEJqG>0N+m@m_8LR-snP3(jl+p<*5c3y!~}MzWqjg27x@e
zF}`Ud4!rF}e(o6<3&#^&93{t~5<hxVvgA%cjx$il#h@A9Y{<cTT{r1F1`3pc;xaG?
zsB7rdPa9bUAMB3a$!qjt{^(gw_)<0mk$5<zT|M*1e9?t;7m@gTaNvAVCluF<*M#Xp
z9!!T@c+wV$qVoilG~tAhMJdV2poRQVpD`)(<jeKG(K@c&lqZ8Y1{+R;ABoZyEk9J~
z1UxAL3^Hqg%(>P#;+H|JQXTG}sQjO(y{ODRIcoJ$DL7r)3!Ar;Gi}egftP43lVoNS
zDo1SaB$evY9TT%mzIakGZy4UWjM&_!4-j|Pb@a=`fTw&akA9m^lL+(^U1}n-hWmpD
zd$g&04Y2t(7;kN!zSqF$?WMo<g=(LzJsDoSIvfs<4?VbcK9mO!l%*H?2sX$`k#G6!
z?(cab#$-p?Q##agIaGEWQA!zblp$^EMuRj7D}MLFZ45tt_2c2&m#>E1jqSt%T;*V-
zzw$$F<0gL!vLguh<I4L->Irz|m$&esALC382Lp1N#7W8F_|08ld5D9h`xR-De(@NY
zV97=w^mA++ny332k*O{NE}=&6T#>l^WTj+=kJF9A<mtzfnBRt0-FDw33=r;()((}m
z?y;5`2(me3C}emx9uDRrkJEufzod$eA!u|!qYaPdE@TnONj&R=b}qN|=FEm~BFc6g
z^$QJU%Qzg(ZVq3$h#@LN!%bDzGyTbvPYQWT9dxB||L8u0K0$;`CM-`<vZ2qj6TBE&
z=a}|$MKKXH(uot*{X-17Cm@LPsWt}_#t6lIye3NYA~$~A&_V6`VY=wvWnexkPxM9v
z`NBk@_6LYDy)hPhEJLFDl3??en3A^9Bq#uZe5twt$7v4IwOe2$7kD>`?G#V>oi_&y
zX_@(tvF$H?SaeLt!dhVI^tZmu%t#}gqIkO=gL41?lTX7>|L6aUV)Hi+|AYU@KdHnY
zg}%&s0RI>R_o?c)>wdio9^_4b!4GY8G^^-+pBL0bvHZY8{NednzIs?im-2gncLxuq
zfmsm-?jhccQCGoZay*pA9KfyO;_=}F@KCv~<`3;-yxeMR9s=&^J*-Nnpbsk-ZPxfz
z(Pq3J;=K}2AHBdd?m6wMxM?HH<-O$R*Wv75W@lKHP7zN*<?m^i<0UxqE#|Ssokb~r
zDM!L*1cK1?rnx#=v!>2WC__&(0lI0<5f=3YO#TP*(!r{9(iT!r3zq3f_u+bqYq5V?
z(S``upL7y82KT&uEVj?(JTK$wzQb6f!sc-TC8z0D^aB&$DQr3Y;8xY~UJu9v%xU89
zO<+|zgirJxfa_4pX*}MqFz-PJ1*_8z5cY8BGI)%qLiJcecJ=X;GddGP;}|EJv+t(_
zYTl*M<kP2X!_zh9O*~M%t~hMO<V_VOq7)nVMaG!Su{?j(e2No0o;`<_aVKwoFfz^Q
zAUrV0n9!dP5IqU<{cMhGV~))FL5{iOJLy;6=s?!zYa7GC0VnAkYhJ}SdHC!EbF^Ep
zz3?^|8RI+QJfzH;1LH)$Wwo&}?CkMY*4D7Sv8E(H4%b(FUx-s^cnDZ>u?C=@zqgJ2
z$_ImD>Jj$_<VH2)g1+W<o>wX#q)i%Zs<H9qe2JPh27mfO^TRyMOPL^M))iPI;bCgN
zVU33Nc6a<BE@_f)WV>pggWv=m4cw&*RLK}#GHd?n2Z%AsQ`Mc=aRZcB-_FRt=fd(=
z9Q@>qjW=u_m_~=pQAI?ol)C_ZupR<1cObuhU{~?HmL_>zSH<6TP~om;qW7kaY);)d
zIN0~5q36MrqvHR)-}&+I@BBCZ-f*XA|BwIAUku;<@-3-&{?5GOQe{T@UP>>NHOxpG
zxepF@{ifEz-kxRQt+5~8z8}snE?ma0=aP%^ARe1mlp$d`p(uG`)9ZJB^GCzK^>6-;
zk34+y_H_7b$#JRj;B7Oq&cni9HxWH$e5v}xL$j=R@Xa^ZQR$Fl{W#9=ij$3;?!EN&
z#B=?fon6cAH9_hJJjD-PTwI-dO~G|jX_AiLAf%qH=i6QzZF3F%c|A>Fy|*%D{Q*7T
zqfDretXZ%wNf|;$d>*E*vP~TckJ_yCyq$_CT-J)<_rdzAwAhGB`)5P3wQT{R<Jd~P
zk;5B(=9%_H8JJ!&aFplRk2o#r(Il@UQJ<wluNAp}QHx?N)%2_b@%Ae9jb2P)>Q}O!
zM7^f0RSrrHeeSJ48+M=H4(r3M*sxjpX&md<ItAeVGOp3hn9p(`lNf@pu-KCyc#Lo2
zMrT09k7?mOKUs0-{@2qLl)J|PCC40kx!^aC_c?#TL->M9-*-N{m(v0YS9aC4g2&Pc
z*TCoGDfm)4;2|AOfUHekk$w&DF9Y#Y1*Fb@9kjOi74VP$!~fvPlYjbu{!eL&Nfc}w
zRrCvz4sff8pFD#G2y#57@5L8<Ezor09wo^r+zTf9g*(jn6Amq8x>Q=mH910J-XlGh
z$rc++035bshpy;lkCet$0I=u|`4qWa1ye{IP3(XPb(vLg0KL!|9v&VKFJ8Qoz3(St
zSlmz9D^)tjE=ApT)xip*xaq*z_@c9Ruy4wAz<x98p0~VsgGx6j-FSFoi8na7x!Umd
zmj)<odil+-i?iXX-B5YzhgWR(c6iHcZM=~)fOB~&E&4Qf%<abF!ANy<B@lxIQRjxr
z;6?)v!fcW&s~&I^I_V6eunj6iH%b}b1SAU);Rzj!`~+t}9^T}^9lmJL&xd;kEa-|2
zI?*FblMVW0A$tAbCsSmEWAUEjA;<cmKb<M(7~g6T#WFZ)aO1ntpa4DPhM7$*1`gux
zH(wAEIyMjv4)_@AhSOp&mbbKYL-z<oN$KL7B}R25x>bD2TzoBx?C6^tXq5@WP0EBi
zz`!<7zF233h!=w~6T?RivyhX)inxPuT7)O<Y#uV;M|0FjYz2rf1Dp(G@cRiHbgsDQ
z+A={0d6F|=S?6+4wn-y(J!$(dUY5rE%@2Jj+<uiG@&ThxS{(3}S*(0Kgm95bDTohr
z)B$9E@&x_ikIvBnM-kppU-&(h|C!23>3M@wgCpw2{fArY;n}mu<H5Yz6*@_|DU`NM
z*)a`J8c0j_(BeQEs#|<#?oMq}5mWDxmHOvFdInW_qY=3%!zVufnfgOrr0ijCT|Jep
z^f2@jq<M3BGn~9TmAOz__f(hR`Noss;Al^irybW1=y_w3dX082FRqlIyFqiq;l<0N
z;b8wDPkpKGQcGxSOwO2`_+!43fp5g2-|Oqk;p44jR^PL|u|53ew?7_!{Oa|vy}1)v
zWze-Zg>YFPWFZbWHCK)X_(3Y3!$4Iq=vb3&htd#nW9cv;IBHn{5nSH@A&ZA$bO;Y0
z!fkR)33Z*P^s1gG0@?{**DYfb5?}nT=gsXf(b^4vA@DmKHy6hUSllRYj1M@5{?V~~
zB7xQ^flI`ZE|FV0afYtKn@#NnJMj=smXontJTM>yxxY8)3rYfoeaZ%uLyeRL4>^e)
z-n2(}w!ABxL4VT0lm#Q=sI5|O<fYA%e<G%Rn5%x2j@%&=P3|8KYjUJMV4#khaA!qf
zNj}V{x7ZUGfAS}H0ut_E(UV7<0FH*IA14j!ebKm12<3-8oWIZ+9Cu7a7L3Bs7||hh
z)hl1*w!;Xy{ZBlnJp$=N*Wz}(Nwe}eLho<#ATwav;td%W*x1iC^drrP3Vw7ZLi>k^
z?ElaH@jnjXuL1st|KUHi6dCV44po;{fy=VFAA^i%mf<1ZD8edT!T&LNdhmU;!Gm$&
z|3KK6rqjdM@&f-@Dy!nv_-77}$vY2Q$@g=zJWy92DqFzmwBvAbyFa3Eo4>&29p?zE
zWGvu5HHJ5ytLVW!#AQ`}9!m2;SwN3rvFEf8$rbu9r45jqmHd~Z1v;h9^Sj8WccWYE
zc3)nR-dw8-ycND`;bD<a7Ukx=0Wg$rpTsIUU}gMA>LX11aAMDSJnZLw6u0lqGYH*;
zUoBg3*tvgC01mBT7RxGn>NPs&J@yIXy=w){=yzc326LS`w9;&3bAEvb%5WuZU>@cN
zab_^lpeNj2dv8$s2!1iFaE#5Mt9!vBk8-by8&%@2GM~dDuVu)bd>LpP3?*1QFn`-^
z?Ad6rlXH#QoD4G=9HEF*q0%T$SB;T0PShYXeXJ574aQis;+p$<4Ep5x@N8{c^P#Qb
z*~aehWNlNl93{_~Q*$BBjoI+<yc54Su6R2tn=XEkQ1e5+pU!DIOwGV=?`#jRU%edm
zc6PlH1r4Y2030*O=7=|wF0&EA9FRFBZ|3lw^}_?c|Fbo0ZHi8FuJ>mb!|6G1P;ttP
z=8lquy!v6twqXYFU8XwW!w*_M7f*lx?3v#h;+x@o8iD?d%_!bB;RrQ0*LVn%O{^Rl
z$>MuXGePfsQ;0XY?yhgd@7b^;K5U{f7iMmGaenEIFXm0C)njSJ^L(E53eSmiijFt#
zT6x#K%jG1<bzYZ(EnuijewZgyPwL3U@Qhs#Pfux}ue@=`1GK~kvSA_`udVT*uc#zE
zZ3W)674pdD*XBAKp3<Fg$n^4Hd-(hR^?!5tji3Bz_=}%?H~iV3|7^Iq`!uZk&AF%2
z$K7yBxt*Q5{JnvqsCn>wV>miE9A3UW)*O7-S$ubLHhllXTg|iij#$=PYO~PgEW-vB
zywD?ZvmtVPcrg4Mzx(6izxuELj&!m$eD}kf;g@ethHKRuW!?3DeJyYDQNGl-E6Q8t
z#>q9*InrVskj*+azAi|EH}ptbbzpmQTWPUDtWu>vrY$KBs<;-NAHMfmMjj$ZR=>Hp
zxi##lzVB>N=bwA6lD6X0l6Zi(l|Q_Q<90@kldk+te(DcpC_iNdUD~ebr6V?end=E<
zUeq_z=STY>fUbx3lC>r3T69Yo8+)vgus#F0E;ioim^Id_BhXWKh>JI$$H+!1<&X`}
z)a%@{TEC=tl!4c{#Fr4-sq3wDMj??p@c{C)%?5Sj$668XXLtSSaJ>C#*nGx==WKk@
zP9t{#x*6w^f@)Xk{|LiB(X663MH;Z7c$5t+#=+gwzyrQTFE<Z6+S}6w%5B}tJrA3F
zfnGNF_4EaE+$tL2A$&oN-6+Su>5Z??$<o8u&{iHE24rpWiu7xEe;J6MDj@a$>tOYm
z;V~FDzVNuxSH7I5WDFE8*dr311%FpHMesn$Bc9N%0`n7Ho=S!u+Qe<i7>I8fOdr3K
z!7I;P*A5p<=tkUl7kkTWbkfF4!4VGQ$dm8{;*kZ97B}(m<KfcMXyJVvGBGatG`z{0
z+x1<B&XHpdU%Y(fr$u<BmIljWI`T0P<Tyd%p>KEYZL=f;7n#=KI$f0#i{{)6@}v+?
z5Z!n%#jE7pbnMsQ!^a-#=EBoCyq0}=#m7f^JzImWtu23dY>TH?xY;W`Zg>o+)HPd1
z40SYF;;I{*I%i}ApmVK3eC5X-y)f20G&1n;%>sJiy?C#mZ(K6j2zc&61)Aoj1WN{M
z@F9KEl9mPG2FZq%Zus2E(4q9<r%yb%N}2}g3;jhd20psB8!Q7L^1+RlH{&?+iuC+b
ze4Y%sR@es(V7Lhr=ax+h@!+X4HWPAl%gx76j6f$I3=np9cRWC#O>rY;;|spr2t^XP
z6`vb916GbA08c)A(^eSzK_+KK=~xGp0SGb90F9@Ikgw6m&ke|N9nVRbA8yigJ{vCQ
zHs=kuiYE&FqA%pr)dLT(9G5`*g*O-Tvzx1#+On*2^F!vRy5V_a2Yo4>EO>_;Iou9#
z!-c2}P=sT`p$HDEIK+o$(i3<YEcwk(@g2!>0dOWS$vf$}oe9VF(=+K(b=Y{-p_(+%
z?|I-Ww~s1N8}X5vs&lZ6o*DGy2m&QMJSAQ6@e^olL^BbRTcoZd=<mb*wfL~PnhDGY
z)wdii@Kkk0156DPc?v6PM0Uze14#{9(G71;_()0FSLfHmySML#leZirL5)+LSbI9`
z>~Cv=qd_Y5-u)HgrL9~j&fWEg;rYhX;l=BN;rPX|-$q3TUNBGs)LCdbyPj_$Sr6jn
ziHiozAMZX6d)vFiPk!{{;oBE%d~G{#R<#nNyTVy*-r(|3nD_*O_x<Ig(#8%+py(Ug
zcyQ3+CK2h?^}wW|52UQQu+I3AK^N!>#)n?w3SPA2-^YWfZP3rw25ZuTMn)GERr3hh
z;(Ed!K?oGcT}~X(*fi8ae4Jcl9Nxkjv9nyfO|P0~FRe8{adtG@!01*eaiEFdVXI3*
z3vkgsantTn|C1K}MuISAUMZQGgIY*rt@0>(tXu?&OR#M6LCuX<c%n47UHT&nKR^<S
zb>qp^VE|lOm~gbssjq46pE&%HP5*Hq3z6dGo^RA7s$oj6blqdhWp?mr^5La^yUNRu
zUdm0T!PLf09h`~dGA`)I-PgSalRS-H5UoUpN#@B0BBm&8H*BH>LS{c=7J4z9fR=Lk
zC(31Lx(V4ths(MU4)_1`r@#J<uY#$MdKgk}t0DFEaiC5<1asal@(#5s?H)&6n$vK<
zUInXo8H9f*YynSsQI0d(dE5un5v&RWw<^X%eCWG9hT$FtRNZRy52Z1Ohxh}%4qad4
zG_0;OD|ARZ1E;oZ5v@A>6J0|;DhrT!&4y3mn_WP7w3)V5nth1>m=}nF=6#U9!&ejl
zeDYELs`E{XC4TsWTb0%v*ZXmIK41I34w#qGsx%6wu(Fri)T;QNn+V3Vm8S2b9W3)9
z4mifIbuS1|=ZYSz>)<?H(U0}MgH`e^0=%bqtMXU+BR`e52?(3yD_BZb-Vv9gH<-WO
zmH&pbtC7k-X`}R`u#U$ZdVR(}(b$c*n3~KNd$H#{)9xoZT#qbV{J?5CmT?91u=vXa
zqVL5vYn(IR=oGJbm3#$#8s;H?=?$2d=cKn~SQWqLKZnIIWMqyS39<clP6MwMG`9#L
z*F7)o&1nkF&=nnIW30`=T82r9FaVT`M}fI7B|ISwf-u1S$<vMD$r|5$-WZ-f*L+>Z
zb5&(_ZEb7V-aQyLHuprQIk4sjoZ`ZqiH#)Ez1Mt>wHG!dZf-L-XJhQe(eZF}cp&d5
z&mY-5B0p>h>}>B0d>5O|FCIF+yyD^7yv4P%vppQ_?|9?uNH*&boak|OeleU~kOrsx
zX#S|VyWdz-IB%kmE_%&oiS(^G9cgT8KEr1e*uY~R&j}-^=a<9Vx9^8{?@s-IG;2W2
zM^PCI`FsO=lg$e8WzJg<QL|abH`SiQTVc#gS;ye}R;1}6DspK4bjt(TGX7?pbmS9M
za4Vnb$Mna>u>kUq4P4CnvJ+4j$g_{`(~q}e$TxM&{J6JD2-c7AdriP;<cLA$>y$5H
z#GxG!pEZxMe`#|ifO^EHkw0zF=7;d!k}m)5Z~eybZ~fQ)4N-j<{_HRRdU$(sDFtXQ
zzrniBbC=nhcW;Mt>5daqvQCu^%e~#*fp7l5INBff_jg@ye|UQ`oV<H)UHSo7$;~G@
ztO14d_84Ws+Eu=h_0$_Aua9<zzxTVpHT>@1`EAePe*Vih!@JWf=ZDDaYfoLb&^vE-
zU0$j_vi6{STb3QwGd6ovM(3(yr>CdhXeB+;-C|={`KP63eS|u|8<y?iaLQK3YajBX
zJ%9I2aSrzOy*an5`pi+J?@r$PD8G91jI{|8A${^g0no0Hox10`MLD&#7s`$F*_@(G
zXit<0;I$#w1rbVG%2Aw&HtK?)@w5-sdHh+AOxoD+h8BL_ti8*6A~@68@yz-|X4Xt7
zhv<d+MOf%b%YBi6{>G<k!IzCM=}ZE9Lk-%DQOH}$)o;S7F7x30-X<Gg>RX<{H=A$&
zDmT9Dwo7hLRD&y6boXiC?&;wJ-yTO@wU5gzY!P}Kt~pJ~7MRlMVS&er@W4FoGF>G_
zsf_CdkEB<2>R;gD^aJ7gdAc6HhIT&wTn6VZPLX~M?=J)KbDuobnJ<II@yV9}8(+`<
z;UD~?pHgs2lH3SQ;sn#N0J*jhyG^G*4V0Tt#cgN^d>wrXpi$-0u*zWgO*FX2g=)!Q
ze(p%Xo5*2QF8{C_uH61L8$8g58CSeuZHgWmhHyI*3)`CDW*e25CxKYW8y*62b8T@Q
z6kjkBAia$Bz>@$BO2{f5GB+M>w7$vG`LWQ>#){m$S;Pl4c|F>%jcd@rtI7U$mu^<I
z;dOR)Hk`cIjp;<@^u65ghO^T%f76Rsrnxb4qvfbWo+{vGglr5{xbbrv;Km6*Uc*PO
z=juS|u<1x`R95t<#DktAx}f75MlA57F~7wmKfn=gscQ;zy-87MeB_%FX`=-55YA6)
zKo24MgQp_!v{8<IvceLk29S=V>7O`+nIGxNFI_`Z9gbq?MwbCieGknW4edro{?R#x
z?{{%yWpnFZ<-zfXe8Y>u2W8IY7=up?jhA}NAcCi?d~5(8jn%+`!7fKFd{RbuT8bxW
zp6lktz%V8sfb~-<k{Le(S7%8kPo88TNqJyXf2fNLFxX_t(^#Qn(+u6#Q!jLImQ%Ww
zCPrEPBM<nY5w33Xq)iEimYZpGM!xHb7H?ocFJk2Ab-@kVub;b?z$Y^J4K?!UcmO<*
zohPkSqNJe+PFL|Kc|<2YSZ>RjFnFR9%Mq~!;gNb`wg$IdL+TV>>Vruep32FfinqUT
zXR!29gSBTq%7QivALYcGNz`99Hb;P|>Y67OssolyC^lw!GwSwI1AYx^z$qvCsXp>k
zJ>-qeeDpQ-51Dx5il@pb?@!Mi$vX04c^mv7k7c5Km1W*e<A@0+MIZ0QMSpb1gn%;h
zx9Su|{bta2d#f~6x2-Pm;T-uL4ZhcvU-AM_meem#Oh`*^Zw9IjLBsd>Q2Ul9en<O9
z-azM!a!<gTjLuM#)qZR;q(3#q{#Gt3A*Y+G036{;qj>{`ci-ClScqvv<aYpVwou5_
z1(6;>&6`W3MR$~v>%S>erjB8UT#_3AN8YBYXF}Vmt#TOS{s=?7Y{WEbWU`Zstfo5;
zu564pd~7s8oT8=9+Cv;$ZvXI6Le6@4WKp5~q*46<bdb0VoPXjAM;jB)`amxE+<wT*
z7@s_nsU|ak&6BoC$au7YA3h>>njVN@NJl&@3u!exKg*;)K!1u%{Bi<h;(#k1l-#68
zT=E+pk->S9+xG{jD?CIFpMW_2fp^~FVfA2*^p;R=>kwNQem)8skTu~Qjf^-Tm<Hx2
ze-Ytt2ISNaCd`89vdD!nQ#b$Qt}FKBMo$9D4o)tcMnl@oXBnCtu3NpChua@rjQon$
zbdbueka$cXH`CH!MG#Bo|N5W((Xa0tU;p0XH@+<IXW%iKG0(Vto$P|1W*(X_bQ!mG
z-_b0_4fOsBe+<Xft{HxWIUc*Y3y&H5lFvXNXAZ0Ib9%0Jt)K-)x)xrhArrsH3FdyF
za9<KueS)b9A72ZIKlDMn?`pVZfBFtb#Z-HeCw9dNR2_lXVJ&=4TmCto@n*&C@jb2a
zEu;4HD}00UW0Hfw5?$-d_`%fuwf&7Y;u!k87tC>e8?he8@X64>739GwBFH23C=d9M
zjL`D{YF+^puBTarbKff7V;gBaO3o;@aESx0e9bQxV}}-I)6tgE#b(5r;|jVNZt};T
z{L8d9qg)sWph_MKA3ttz%Y1E!=s#hFuY40PMw}d`i%tAyhAW!u!ZE>3rf5o@co$`B
z9vn_16%iADCHK6n5*T9~Kk$WXe&c3P_zH^~+g?LYQ}x*4#zj8l9LEOG5!T{{e&S)j
zLjxnv!s-IqJPuH<5d$5+j<z=@8U4Lm26>fq`2n{T;!xjZdyH6M^2qA3bn($IeHi(6
za0{a@1<bSknB&m~X=1Rq(T~ON;RB1?7sJQf^Wnqw$?);!{qW)TY`D9=@I#EGU*F?q
z11WE?aD3pJ-+bEgTU?qqYQ9hS_SVjDu)pVTFYWNXF3qv{_VV7|-mts7<BbkBJdXCa
zG{^j?IM0WpgT3M9@&54o<>B!1Xm>c?+Z>K$_O{nFzvCh1Q;B*#tZR;Tut)qunS<d_
z?(MA&&Bb_AMCnV0ovn3mp1nBOF`pL)yThUUYnspTVD$U<XT!<Kx!0I#!z-V`U|oSX
z+P1u=kq2|7J56-dYNSJzgLLT4N6JiLtW$V1koAwJ&j(Hzxsu-at`~32`1B(5Yr8$P
zN`gSMmSDu=SR^7Ub)LSGTH(0~asiUl{HbFtUi$9Pd*e%U6maBe0_siZYaWFm&(tQf
zu+93N9J+jveW%OL_Qvqdt3%g?s~a{!??mv_^<rad)BIRhVxz?=%1<p%J&kOvi)=8L
z77ywv`C|i=y0Wn@nOmhLZRYZb?)nTb8NQjtMi%k5x5elAhSvvJLm;*1<VD#dP3HD&
zCK26k;)%zu>gnO(-mtHF>QA63$z+@UG7tFrK|V$F`lWR!o-h_|wZ-&J$(rkjWao{u
zy`7C=e~%Mwc;isIYw~$5NW|7He!2ZXEDZEdI`J0`3G*WvNV7h~+E<$s3WppR)@`WA
z=yqLoJ30`FH`S0K8-OGMZ}f_e(WTeeSldv&QAEdeTEer=MtbgliA=cU30;gq7VB!o
zKO56ApVS^)8=lpsEbD#B1Nl=NWJAtp%H$T`<69G*)QiH!E@^kLkglMYA^K<@K|j_2
z9M`V6dmEVCtJ0Y0{Zm>6b}G2yEXSQ-mA}WNT^B6M0hqqwdUpVK6UK2TH^%LLX8Q06
zw+w~v?Z@bQfMVf(_XY4Jd<(vY7Gb`M{)+&9U#5#ipoV`P@Qtr$zyF7Ss0qCiOq<lD
zV`L|}Y&QNTSnHM(kFcU?*LdSf2ZcL!ZF&HWq+RID;1(P5B&o{R(lrfM^T$2$1A^m^
z%Z@OzkMvE^Ans*clVOTDLS$6PC_|&SPq4r%;f~V`BQSU4>a@J6r5l7RbxX<}w{RSB
zMmPfj1|B>u#D>=UlT+JlT(PNz8^i0;=jR%%ygM1*zI!vgJ2@Fn-@lJJJ(d66@c#5n
zgQI-Y>)ab*=ib~xVixAzuq6mj-`w3Y5aDSBg{cGNBbjufEPUsQ1@gwm1|0$&CQmZt
zsS)~1=)H*s&YOEId@C+IgyU83+=!ZPS}|7|+%%x0^Cir5ve6sza-nXJmTyMfHJTFh
zM16D*-fzZ%cesqxiOlc<YaZ=p&CSG5T--AN2+kX1!qrVSH!77on-0ozHgd#MdGg0%
zHOS>0ZZ;0qbTeIlzCJwH4INs{8iN(-hnj=_tRMNwvjL%;N3<M05bgmNLGtIOmiWHO
zDsAKCN4MVeLO$a~=RCtdd4Lms(x7arV<&Cj=yh|SXk-+X4MKE4+R&FCfzq#SDV(_c
z8AMptkU0+NhrfUz(v~yX79O5T4?9!}A&_urWt?`?fj7x<K{nFKC_MCFB)5?cRs7_`
zX$ickD4q@%BbB)H$EMSBm$x@M6^L$4t3U>wE;Gl`Uxuf9cD8pl!N?nc9)hS&krM_p
zH)_kb*O%Tfa>^38-X^Plf%=zv$5VSxJZWGv>A5B^)*sS|m!x+&)6PiCZNlZneC>me
zJLH(dY|^rE^=#PQ6(7|%o-)h;SLR;fq<5!s5ssI}cr$Fz-}z%8syv7(>ZU(JA5Zw+
zl=Ypu!loA+dz3AkbT1B%JONzuV2v_TWXh(MtLqPf8pw$)0Yt%1!R2Tn&FlLMVfhOx
zZps{Ag&2%zX;~0Z$Yyl#jl5Zb1&^SUCm=4$pMOmrU$Et;?NobpR0qWk-p965S4;yW
zEVUmuS0FOS-+mGvJ(?0C!c5F1@#_gSF4PmZKN+POUTz1}8`-TLz|(mV56858p(%wW
zz9{4cQ*r8^khh-|1`)QeqLItI`1N`T+u+R$KM1)M7T$6@F3`f74dh#yxS_>r?bDcM
zS{&QXQ*f5sKV(jfqKByHjJK!!0f#iY&|rg0m~puLM+=41bu4=zaUI`i{FNW@iEW(2
zlt%-B32)9Wo&%Z*zQyo0aqyw&;NUvp?TH0JLkCqJ2ah(^SdY3eoQXGqd5%Pc$4w0e
zqAjRC{jF|%0VP`}?h-&V(yf3#e6*V>SmaTm!4=wdNN!;!Kd=xN@1h@g4EHPJv70eY
zox@^`#W-WU1`ma;!U3!BQ=MC;?_rD^Pv{%IV5I4A1%22&e2x#_T*ff4GhATihs=Fy
z`X@O?GikSQ^mkL3`?2Pq_~V_W(O?1ZfBSW~9(sCy4##T#3R|L;N!?w<&FLC~zv>e9
z)L-k`{5}0dgWK0J9C!S|H=VBJ8Piyd;r?Mald#DR^x=h{<G9xV^SWB5{Mfpf*7C&h
z|Ksk@mNjXz^gJwb*gN-><3CkZ52ywjL=%E29Y{K)=>`CF13DFtrAr{u1PiwzQUF0S
z9Yg0pL*OOINSV=$!d%lo<ea(Z%+&k5?^+Qa-~RT_{Hv;~A+GQ9@a1dr<wJOQgva2z
zGqvRe98-anoguBwDm~K$?MJ7BCtlLvIqr;M<R?HuF7>y_jdhK$(3iSas^rZH&MyWX
zcd&NhB`Q4m0i!IxO&Wrkr^Fx93!YJQEnjKNjrNGqj;%ms!dv7;p40sJo7}V^ghw61
zkIZJ>kA7;yzv3ZSATk8kvSJfA0lDJOu>(v-lcjv#&ir*+n|KC!2#iHFB^>D5cygCy
z6={ZuR6b>N7OA|FCj^&HaU|GsKALv5CDU!oc<Cb3cti}$$<oin|4<j9Kj^&cJvRvI
z9PK@C!#VZ&+~xVPXy=not^BGwUiZiMx;g*n=^VbB&f*2<alE;M#jQsV52vF87RL6b
z`}Ys0&!0T<0@vP-=Gq@Frq|EEoBrkh_kWtc`{tKAM?0M!9Ne3J_|+FWzuKAZ>il(|
z>j;|K|EvG`f13V>|Ka~Vef{&Fc|qy%qx;j3e)!e&_`w6^$!iI@#&Bo4r*N=$*LEHr
z>`ptp!9(YdEW+q?(NB84c*T=E7Zaz1l%FR{xOtP;AhJmJfY&TOdgLdPeDiKyr}0_o
zJDyNsv49&PWzh93R5(vwop8hT+gH=^+n264FH-1S7JhCnMOK?M4eAb8V5XJe0B)ms
zzYfnfi4{T_b&yv(s|;MT;@m!q3i3x>#zh?Ka}GawguLv-&$$%W4|E=0vH)_NC#zs;
zk4I|nN4}oN>jHW5kA8W;D<Rb<s3oqGShJ6DT+@-goS$&Lp>9s+`jWy!oyW6Kbg-{>
zyz5@-fE`~?;2fQw+7OrWU95AR+ud?M<YrXf1jUWs)>U4}q01r}*8qJTOulksJtFwa
z_v!8{Z{Gwf-?FHf>k!ln`#0AvC^sC={V1zRE*r5JqTz9b-n)PDCYqx|UYmH=*Pt$a
zErhy4##`qBc!AF}N(tb<Xt~r8N|nj=;QkPL`&=g569Z-TbuQ)0^%wf5pA?f}t_^a1
zHfap2A7}^klUMTJPyESu7QLwt7SsB=oPCis^wRICr?yF++=R@H_O%Esx+KeYowv&l
zHo5-9jp_`B(0%@|<lZh%8PZ>&%eZ+<xq!&ewG-Uw_!nN{5eG~&7JQ6>QNHoE3(S1W
zn38{$!wI$}4}f=QSm6Yp+|q{dmi+)2Wr3@9Hp2o}^4YL8g3DBX0G6+i;|CH${Cz^{
zd>^>U*M`ylCxb-4m;8-@ooX-2_#3mk3D`}@!q?x<!k1a<e1QlsSh)`ghXvmgmJf7f
z2+*`K%CHeTi714-A&Di%rheFwmE>VwI=%;?(g*cHr1CHo+_5$cpsiO{%-}CxOhwVT
zM9KUj2b^X@AnG{LNC#_+ixF%GgJP6TA>oyPfHE?Gav_>KF8n<%#Y_q~xbSVRJQAFk
zg{hY>UrygWeX7aNchlEj|6+Rf^t<WlvuD%O@4lN}`mx~GUfg>5;)S2?dhz_Z7rdaq
zJ>dzje9w!8F&4{^d-d{#p8#VajC}3KX`b-n)ol#G+>PVAMqDxuU0I~d;tUgo-~?ZI
z*MXUVc~=uH27E67^+CZh;Pv>s@T7%H2fH5BSu9YDEqV!KiJuffzw42-M*8%Rkh1?o
z!UsS7Xmbv|m!uWkg3wf^+!dpCFixC_0wS~xX<w{Wa6c`aMJwMO!huZ_LMFEN_V%Y;
z?&{J3``|zmkpn(Hd*Hjgb~M4@K<LM<rQ;9Iy38qeCOFh(>I8`YMHfGHOJJce<#t>X
zBNx78@zWQr;hB%ylsx`hy!xQ)P{DPRo!#_#e*!@8L2>|-0`LB^ZCkb-pJ!0rakysd
z2?i{DBXW7V4RTNy6)orMvB7?lpL|nh$Rs~wHW!(cKmL=Z{PBg@PZ`QgY>7Y`yz=Oi
z-tb%~{<2SB(1eT0`0c#6ow}zl`^hfh*tew0haLBO<P_-pEPhd9#CWocNj|R=Wq&*M
z_w<-Zrmb94)JIMJ!Bd~9ck;fsx9hv+viM1xNrp2Mq|`U?340VH7I{beyXizHWM{l4
zL$Z|l-o3li{Rj8Gn8k#afRnCAKcI;p)CG%eyXwRHDjTokq;Iki$Ko9Ir64&bY~DB6
ze>6$uHHY}Wzjru&@#Hg~;O*YYU4g*d7v*#6p6tT|vaf;QgZ}=e6@5Se!}>uZP_cr4
zL1#YYa=L#5KgWsm9cmlC5g>%xL@&pyhF~og+Y0P(;OvhGSVy!WZQ2lfvCI1_x&qg4
zm$~;d0`jr~n<O@&@he+1?p^>cImQmjQdTtSQ>JGV96Gop!e`$j4&4&)Wye2AP(fUe
zp;?|+zPK!+m!Ng^L&!R^jYyB<ma0w5&H07`Nr8O=UfRQ@gx-AF!yh8@0Ft+272h?W
z`tW#WJ^iXLI=}pps-Mv)y6j5Urh83fi&nhUM+Ll`bu2k=p{G9h%U;P?&MuZN7;*EF
zul7OIU>;HUlt1|!t+b-nH>vBWjyVxi8oY%P&Gq9r{w)5vdIlbCU$(t9IRPE}gC8PL
z>_y5n|2!s#6K*iJf6)tHfc-3AZ-QBF2qScQlRrw0JlHiv!!f@J+J-a)XeD(mjQNgW
z8oWw3+$28+z&-(_$1uPs->{lfY`~busAqf&J$-jgTS4(0Hq#@`(V92Hb-Wdg7@FzX
ztoSBH4k4tI4Br=ztZm8A3MY8q>$G)h?&SSHdTp!2Hd=>G-k44np&2SY2NSNN!+=+~
zg8d%Sgw|0arV4l3BH%#NT0Xp#dz)P9W8pP@49LST8UGV^upGY<UUZjm6MY8KV`y3s
z4r{A+HYR>y!ApJ{-f1E3BO^V>L-DoGt@k^0TCR^3;2mGhW1;o2LA3rs2z!tVeKXy7
zL(?6GbeBbxnYNa9%CXahL(90(`~-t>BXgK-(GX4}1`arpP)YDrdznugl5TSM`3_3W
zIDa-vd$f-oLQ632pXz;1DYqIg_(5CRSCm$S$6sl<vDrERvNe9pI;NHke0ZQsdZbbA
zlpWeaN5*+7EUeORdd0D~(3=O$IwsPe6Jm#-fV$|kWn-c7%>Z<DEwrM0s}$#A?@p$7
z=WnKyx6h}yub=C@>O>rsw$59*knfupdCG@t0i3gG!gP<fh3K5juR_!L<vpD{?%usM
zou0g&o<HUK&U2k(@y3obFRXq3<l*$yS6|o#FIcFEocl5NdidbM^x){e7bd>?;<M@g
z;ejvQb0cM+=j`v^neOwn*WvDThu6wId**L8o$1`~+h=d4r>}U*i*v#|zRtm-7uPWk
z509pYk9hU{qv^@xhrX#-CyP3#J)K_hMxmE4rqkosI@i>>?S*{3;MM<pC!K3rd|u<k
z3vw4a2PVyH+<g8o8}x7br2D1ow`(@wI(L;f9l1ic0=8m`^B#~suk|>mK#xnFIfup|
zJkH@cpGFp)d0&wYUoVgh*Htoaww2^nbZ%hmxdyf=BiHx%R@|cp57j<<qJ5a&9P>)T
zW8ZAeCk$wR>~Xn-7+=WSKGVKRgHsngv2~>Se|U5-J-oj^-9O@KzPv`yul<t|pTo-z
zG=46(Xj}(4zu=smg}U34eQ!T_BJ3TX6yQAGeI|bM+DsPA@~yf&{m1n^p8PsI+*5zr
zaUJ`5nvB|4)uGEzowKN;pTAL$U%s4JbxL`Ru%D(bp|O~{zt4V>@8w^-V}Uv4@&1eo
z<#M1MWCvCBn^t^c6<)FU%Di&UH7d1_7uRGz-yh2&AUE&&W_bs-ogjTh2$@sX4|xiV
zYlUtnk^Pn`?ZebH{lN>NKCh3@QKf#PgFkr9BATD@Dj)m=r1JH`7kSNIqmweyN4cTc
z*92vmb<h2Kx78lE)CW_}<+vDVJaKrwra_&qc#DktO{K#r9{-XB7QM&=SM6dW`GU`M
zH&4<)>i{DTY^KLD(muYvXIlz0ZwX_1lQ+V)d~Sj+H)3tAeRC*U!$|vR!01O{&Zoks
zz-IsWB#`*`*0~nQKYip^w~5}(q3sj?#h?9?e{k!+{=fg@z9C!Cdo>U|Ls7>8BtlDK
zn3Dzf2*eH*8i0O;OAirZK^DlAW<nyWcSO>y17>4ANM>6s#Tb*0`<-HvMkAlWs)Rca
z^OMhP1CAm@%~zrugFz~gG?B|xeyw~))M6!Dx;_x`*TBVxS~(zQ(Ip3CUp(L8;<f@4
zqa9vX_HH`WM1gnuzI^p+dd*W_FJJnbT@2O?&htff4X_NJELIWxsy-dCE;ae^YYQ*(
zT`#{&SO+i$TSjZ<1>C8?0pE|Y%Mb4C;X|GH#|3@t=Yl?WGURc64w`<~v*=vBPkB`q
z>H97R(Q-F~24NkH$;g9;{%#r70TTmHCKR~7oqE7d>I*(KOWCaA0}eD?&%wiORP7Jl
z3yMq>C5x^OCq8gPCrvZ3py&fBZ9yC8AjYfZxRZmj<PmEbV*)hiNghlFHR)lIfi~2L
z&D|c<iNif!WvGL`%E|<Y`!#q}8UOE61{U*FP_h=nM!ZaRC<_NoKXt|dj|FBIJ3nJI
zR_PJBs<iQ4BGd)_k*27AX%t<Of=nLL_wwy7T%}Hsu@kdQ8Mj|hvhbn0Q{C9M($syo
zv0bAP;NpwYVp+F&ogb5eJjux2G)x#plMh~a6!a4tG9&x+DIv~JX-&x7cLf~$@qzrX
zOaDQ~caQ1dpM#(Y_QQZp_novU;4&~?ScrvE^@k4^)A@VtC%3eZ?Qt?vciJc)FO13u
z%J5G6ma@%0tF~v-$%)1}@7m_cF7`tv{@VA>NekzW9?HwUv!imkud5E<DPO<qT>J5*
z^1xS$aF^F5?(I+ezROB|ME;XMi{$v=F+e9(Y6!n_ls@RYwvNxHSFc`A&%S#(oxI66
z`VQ~!O`m`Hxh74!(~FnSrk5|@xE{#k9SdQ~*X<(z4iE25pMCX2Cp>)TYghH6N$|UK
z<(E6VXm|3Lm+t4>P4-SFP4vgl9)B_Y(O>y%)8~&Ld*LhXLIs1+ywc$(Kf86<x<%M@
zV(y#zqfe3vI<JfqFJjq7sUh~Eh6jv%0ApoW5R=f1#H#s-u}P1yc=eYa`vI3p?FAhw
zV%`_<!+lXy>{&WngNPpYlqY@B<0733+tSBE*le(~*dWi&D?Ulae{DofQhW<&jFBJg
zHy$t4el{#1UrQ<}3H3X0du~0qk^ayk7Y{<lek9TJU4ZUwsf>J>Ji3mN^gnsn4Ec9{
z67o2qpYuhevI9+-Mbd_SCw9yhSVCdYQ%t^{=<y}TWa&Gtzu75Em$w1bW2Ium4NlmE
zb*K~SKYhr5%XAqHmPauFYBn|$HEw;Yd1Knjjvy&{DmuP%x|4P&j%XQs<2ShTE1iG{
zR3tGp1O=gxh&61(cO97=9$TQTHPU57yCq|jgs?#|CwB!x+BNIVN^8b{`Ct8{5`X@%
z`HiowAc?VEKL(#F+j2Kcr`@-Kt4};P*EO)Idz~$)pDOz)a7Fhcw!cqaqbxZ4PSdto
zX_>~|*9IGaQ_vJf`wiP@o3K^Sn}xn9aYd(+Yx;=0)q@tSWi!tLuJid*=~vQ~?}~oa
zovZeh+?MA^bL-6MZYz$LMg@!P$7}`S*mZOuJjSk+Y$Yo(QzH_*q0nu9G+f79(>YEt
zA4{t9BhFYn%7VA=lGw0sj%Yd32Re=^@iug3D7tNWKZ#G9c>p7A#B(gT>KAZ@t}J|!
z&vXb254nEBL&G|sp3eZSFQCFm_@5ZrwLgp#oVf`$^03xOqVGkr0H=|GmasO_=_a=9
z0xNsP%e-PvTF2*tN^`E*Kw0d+<0UVbv8eD<H)GzVQSxQlgheLSO9*Y_%tqUGbc`Bi
zYW_3KzRW}rY{7}Ynb%dWgMRx{-k5I6>}Q};Ef@Y;CeXBPyug(<8q%2^DP0FxPTPF0
zdS`#Sw|_L<fB4z-;Ng?$=)t4u;P6N^zR8tE4lY{nALO-xcl_P3oWs)Ix#*oI%$Tp<
zz5U*=EBySkM}A_A1qseAIj`X9uN_}o$RftuW1VAaPN!2zziN+#2(C}?1`-y&cGGCn
ztJlZV&%ga{dinNL=exJ3(+izH-r@#NZtCSNHe9Pne<B}l;AOFng*ZNw@#4kH>AP>g
zonF7@wU6h@KGy>9!%yk3e^52@pK}6Uqllf$_wk+URX`t`eI8uF4z3_Axt+HOk)<xZ
zD2UG*OGf)k$T)!yTrc3sH=g?99M*3(p+RLXa|6o3H2}(i&G-Hmx34WJ55E~n=iZjJ
ze1LX-^x*#V*Z<ldPJi@=f8gsJKmElw(=UGc^>lJ_;p=Vtdk52r+Vu5XzHN0<3y0aa
zSxjQVbpIaLnRfi7;FCuWrq3T8c_ED_eemJ=i#OBP-@cf><r62b`6LRTF}YB4)OewK
zW1*328IPp<gD;;<2kL)ZBS8Lx+TPzqlkdFk$lr=nJDr}gkV$*;jk!DOgAb=4|Niex
zUw-kKWIy;C!Z*tI`Rg~+x%v=$?wNdNp_A{H@hxz^8`tM4S%9i`#3l=@JG%;Z{R&Q=
zR{ZSALqDlVAN}^JY^zMCJcal!3r-{W8eY+9W8btc)7=D=$JcJuK3*tN`>}udW_Q&a
zve+f2j8>G4LG?u$MHMe~h#+UTE|2;&Ht7?@eT|5`(obS1`$eudl80p!`kWH_CD)4Z
zgMA<S`i<ZJ@(rJ|%Z;Ueqn7H;Z`hK(o!ghw{e!#H!+n+Y7UM?eAKi|H6&@Z%T6hcY
z>kqSR=q+U$|1El5ULbhWiLtS^>H3sdTi<4LgcYAK$_S_2;UwC$@iF@in>=5ixdFy<
z8zlaH?B4{RBKv*pHGHg03%K_OmCCP@*LI*U4DI#j1f#z{`Op8)w{HF4{^5VC($HD7
zgNe9YG!vPBl0vg{vb7B{sv7)_8w*=jbPpUM&9M=-(aJ>^;WICE$mCT?5h8=*m>n3R
zWiYanP?UE5Wt*)B4I97|JXn;0${y%SEw)WA3xI2;0<cJ99W1yoAu~k?d<J4)SoeS?
z4&$F?Q|BC5{kZWZ7vdRk7&wK9dwO>41uuV>>#Zg<Z{JLBbYSGEk@NFQO-^*6)`W(q
zDtMHh4@16@+_B^?&p7B^PA6wN04jZ|gCO7Yy3}lli9H8!J}k)-EZjljS7FIHV(#+U
z-Q$Us437S;mJaB<3S9K&Nhco9=G#BmfXBtuI~tG~_&ES@fu0eKyeOj=)ig+R(D#8@
zWn>_Cqo_StP*QHrPkW?v({^e1?4%qnnZUqh(T|Cfp9Yh@#W2SKP-XN1O=;?%g)Sx$
z9Jr`;?t1Zr$9c2h@^0eett=)`1~2}I&cXzCxLXCcyvV1rO0n)>nSAi!shC~W2MdBY
z!9ffkows~sFy=twJ3v%7J~+!C`UMM!z&tsTPtYkn^^PuS?^iBl`kN0-qPae(8^zcK
zT!$r#A3i`kl}-0YbaF7|4juA<Uy}i$cE|EcV3WF5f(ZjjHeMu=JPUy2>l0F91?nM2
z-I8uY`$l`I{aD<=vtGdRBjl2&MsnaqJ>_NN`jJiQ9vBx~=>x73(siLAlfB3?7C@Bu
z*$3H3$34P%RGSH-%~<d{|DZl`>q2GA9eXSqSeAN|EJaae>7T_wKh3E2y*NFeUhrh)
zvlp7!zMbCt2}<pE)Rij}ubGUcU!*-$F7;CuseSj6(qzCl!uZzKo&t;FeD91ZrTvhl
zA9>+gea{OSwyXWj-|OPT$Y(k^c{aUyaXQ`7<n!Rc{`BP!zVw9Y`3uoCX?>?juFFCj
z$QOUpPJ#Y+_-KFn;)~Cx!}~{S6qTI?Mb!u66t%`eEq5yMZ8G*B^eD@hpM5#~$A9&Y
zrmsHxV%pOfq2_iyVILoJl0@B;PMx||?1usX%WWtge~VO~Rwy!BVRF0xml666ZA7#8
zO%+xI5VM>Rlg&-KpjbALrw&EKIfC<2<R*d!GZ9<=;FF|`!rhxiw7B%&)Aw8;0`-}`
z#RQppWpbVcB!02VHpjOD3Ma!&N<<z-kUjL!XS`(pqi^uOaQY4V0J7Py@I;bI=Vk=x
zqRD^sy<z`HgtRucO+xv&C=#L^=;vu4F-7k(Ae-YYny{6Zwcs|5K*q#2YK~7F6uUd%
zTQaO~dYL30DuZ@UnKMVQ0?e{0T_q`BmyI;$kVm#mN2?NS?JGoa$RX0<Szp0rj7S(D
z{7qDBTH9H~#u#uN*(fP(Y(P|z%EWAQmNmZkMJDuJ*GH<Rcx8*N2}#4Dyymx~0D3JG
zg_f?EA0sGg=n64MP<nGwEEz4Q3Js04_##?M8}k)>3|ov$fBN_T|9jzU6F#Op_QR_{
z{ch*D9>p7*a}%!9-9|S>^y|!BXX{hIx`RIZX!iRBzK{LuWY_bgMP^g_I{o(ZQ+&D6
z#x=0X+mxL717p<j*tW8<#<QmuTw`N|EB=gm={NR<k#`+^8_fKUw#kb|e;%8y<r2{*
zP70bn6Sl&2{Hyeug9|)93Mik;IEvag{Mcf*NS~f(VB7~fzm6Fb@di!w-KUMO_nXMz
zSJKM`-#P*pqxX#l>tWOLk^%-3>2HW)bhZO*E6+9>^{LsQm`S^A#4CI9!OnCU@L!1k
zEfcVN5<+}ko}~>9*ecXO+JvH(th41e2&YQ8FkLcA!;k$WR(Z|(O(R}xCJ%)Toj}e_
zMp<XU{N;-LxbM$a7hRFkpO*fC{Ak7Yt<y>*gO6jUkH_ea%7zfwHZPtOeNf2wk+$?m
zB3|+rrSf%qmachD2ijKbFEHk@hBX~I&nY+PBW&6l)7jss^VFw#qH#I^xIjn0W80b!
zdr?baXJ>!9bC0*c9Ozt*YX*w<kEZ>@2f}&%{5?%FIj_>Wt<I-epx8lA=ZSlJJJaKb
zhkn9m{~j0ba}9y(M)bqId%M%o0Z;u*)49&)j{PmEx`6(ko1agow=bTnE-$Aa|K9IU
zx8=((zI!=+r}Mv)vvbSwgcsi!CvZ;A-1?N)-0>D7&Np=)%V#p)zRG7d{PvR5V-+;#
zwSM)Z%8U<Oi^}|(`pq}6`PMVFbt!+&buP%0HGy1H@VSrlQ@=|c16|L=<6`eN@bw^%
zS08-M#n%a}pT0k~pRdc!dEJ^0_V#=(jU45rj-5Au#atiin|`swclW6q`mq-%)n>GL
z_{z)sXg(#vqU9g`M}KAd@sEE``Cd*x{l$0F*WWz%0u$HG*vnaH;yRtL56K6gGs_0o
z_wMcHdth8^d3679`rRLWHvP^Io=l%Vx-Z^czuD!Rr!S_Tee-O3_WbSi`uKD@InBZx
zzOjG#T7&AJ@4bEX<&)`)C-;3F;hU$gr?+Zn_gDGCUDw<M%L1iu;H3|!KBzAiq&|Q0
zWctg0@O#q_e(;&kFW#P<D$iHbi&yFsye850_fCBzPquMmxvwjUPTTsNlO|D`HsU&>
z-~J@}p33q>WqSPhNM+rb-kh9G&!6+QsyEZA{J;*nZa4j@p9>Zsl?T^`oR2EK+lc0O
zn<(V`G}os5J$3w%%q9DgV(-V6O&j+)IA?#jI|GdhNCt*)-d3HV<MMH~D4f1Re(rOW
z`CZz8eS);Ffl8XTzT~>)MQ(h?dgUOXKVeRGh_!J!6uADl$AaOW`o!VRbbn7{hxSMI
zf%p;#p8Y3jj{7$1D6}B$6Je33|CqF*H3VPefjJN3(O=VMH2jtBf;X)Agi%JQZTp~6
zbkhbI!N>Aw*yPcMH$loZfW*I#{hQ!ZWWSHShL4qLfuT1_<yXmTJJ9A!fB5r)(cgLD
zi{dnym|Zp?j{bDl)i;xj0nyBJAkyK_z+rW5T2Q)~xCpfa1O_f-z!!@F(X6n<OpL5M
zi^h+|8!E6+mq?XKy6r%-K*f$2<fwE8$cW6KbgRx;Sc!-M5FGt*im{4waOL1DTmv(M
zF$Zi0I7CdBxL5i$;PHfp7r`!d;5$8=UccenTrWJ>zkc&t1HW)hTuyXwJ2`vfgVk;B
z&d~vbyFs{9;~h_Ah|a<4?35=<a@UFvo*r=DdH7~gi+B9;kyGw);oZ^<cIa~8_6HPe
zV5LNC7|1=?`eHqI5b?oM<$H(Mpk)B&yH>diK@&Fl<EK>i_q<5=PUZ3fSsN6>A{+KN
z=#wvba6t58jLJqW<vqn!v#K}OpAy6b`Xme=oQRQeoP#V2DI6481jL3LoAyv!&_W!5
zE|`dE^1(r!11)_0C7XT(3gSCE$VMCd#NRtAuV2lmI-oo{q^Z6y#AEU+JF?Br%>vgg
z)f+emY9HJ<0J@IxEb{bok;J1QBGX0TRe333E-VXG+6Q<kV3YoX&*<YPWuUnDNqv)_
z?bx;gmZ@t=IE5QRSirDdyOqb`&BQnIRpywW&M5~vOj=A4o_!Urep2<~2D;WsowI<#
zWBKeG?iVra{e-MMC`-;Z+h<Z0!?)*@)w(c6XU-;`<?I7w{3i`5G&0F}L|aoYS%`#7
z%%ax`udkR`V7@aQu&}8zQWs24Jb96g9p%kalGK?OopmzshS!Yx#6bG0AKKmv#Q3H1
z(f7!kkLcn%b#VDk`>akt{7MMz|1{6;KHmzv=MUKODorN5xBV{fj2lecQngI(RGuoc
zdO+%3CqOKGef#xy@`=}$Y688#Gku{6FYlXw{fch_zML*JIi_J<kCGv8*SG$6cW(Jx
zUk{%=oDMZk@QpDqoHM~!8Q8Zjd2NNtPycftRR8$u^B+uq<B$IO^u^=PTz2{+CkQM+
zW1EvCx3%cXEfM`nKkpAH7!!rgULD$U1~@6Zx{_2QGcR(%!=CMskc?18wGo&Vz4Y0T
ztWtv18FkZi#Gpfg1|jAYTY&bFLxxBp(K5E9MADTX=@=%ZKsmq@TVG~rS;$y)3O(b1
z6K=!R;sbR+eJglhS6{rNen)>djeIGeTt>y&%75sLRo>4OLKk35k~Tk6Q%-m}Zs0>L
zjJiGok(MTH6a%IeS_nMpQH2un*hIZ9S!o>62AR2iQLM6{Usj`nw<NtaWbBB~mQh~S
z16`+)MZ#T9v7K2tyVy=Xu1iR4Ts~~aMj$le<Z3|}1>#pS#gml%7-gxC_zFuR2K#_A
z`HB}$dK+!FCBBkru%|Y$&;U!7=$NO$j_N=78V?$jcB1?<P|nD=AKZmnM{>a;*U;Vu
zg8k{=|F0I(pEG>&!q@Ek7WlXY1O4bG@V;F50-~*mOTv}3RVNUB?A`>;N4_Pu`?4KA
zX1^g9Mg;C3Yd!EDCpYQW!F4?7%)J^~ZjQYTY_nn7K;I^_4X%^BQD3}~+xSXn6c`)H
zF!T@cZrx(+Ptul~H|X5t=Vlv>vOZQceX!jnSn;>$HEkXZ2HA|;!J9t9nzWYL<as*~
z$1-i&hL?-2lzp-FQT>qS^h&;BNNYQZn>6w^aenhLLusqZK;)91HALfBw!qp}gYiKz
zW3-@UY(M86GgzK-lc;>1j!f6tsMGDe+3tXI<gdbc{sojziyXQt)4EUwbPzY*^5o+-
zTwu&|^x4O5Sl%S#Q5IWFP<)ZpX@$~l*t9+R8T*xo(~(a<gSgZS(D`7e^8z3{>G=t9
z=fN>`Nw@9Bv6s0>r-kE#@@@-;s2~&>amx&(awQt<nU*NIz#u%vGdbx+C-z6Xvra?n
z%y~#&^AOUrj$kZr1K6S$zP+YCoZ1Go!}a|J{*qVb%|dLx%awU5GCEK93{G=n=Hj>c
z4$y8EzYY$)Aa>_op8g{6l-51b_V@Ou!^6Yr(W3{`<3~r+{e#_UPv@kX$}2y$zRs!k
z_jdhlFrGU4#kb#0-+cFU`tI3_>E+8;a^ao7H}>xA)HeUd|LT7|y}z|HJ$rpRopIA6
z*I;x$#~W+7xl_(gm)ciOE_edzWP1Dd%r{&=fBt-W{`9Fok#T;?x$T9{ZCUKW7tT3T
zFS00{CzkS*5Iiqzu%Mzewpsl0H|$ksoHIriHhp=l-vv_5)L*xC>U|rG^9WzJQM}#<
zl83LU2$C<0E?gVCXIU1$xOvndRcPEe=!GwzGkx$AUwKN0vojVfmp8*;)91uhFL&?W
zQu{ra{^}q8!Stga{$RrNGe4d6M(6x$3-a^AN6yWF+{CT(ZP}#lI45Hf>F%8m(^sE8
zp8nGBd^!Ez55Jhc`26v7cyD()K0Tkle)??s`8Us|7cY;elT+T}B75>9*HJhRznJ!P
z9{<IY2h)=W`?C3A`sVqY>AM$i)DF2BJ5T!Mx)awWSis`7mR#>)-o)Z43weL(_kT2f
z`S}xHf8-h(Pn|t~_1cSW@85Ig^uagXpPup5SS@~eKXKEz1sOZIc1Ax6?SuLYPo*7B
z4<0_8-idg8d^Wv(!&7W|63*wDT<g(~I_6c3d=}<h^>yjbEvU@;F%UWr*{}|=0@qKd
z;@l9=^%vyR4}6VITFCl32kk8!=d6!eWN9DmTisSrgENk0EcZFI@&&xOS$zR>j>WU>
z@`1MYwG{D?_4N!3W6NS_6%F*(w5xsa$%DP=i-!l({d?M%UF+D6Ov~a+LvYdIIf@Jr
zf0sOn!JOWZZLrK>ALu-0I|I(n606Z2<AyQ4!U?0S(4@x@_*9#l(7cwoD!Wc!;Vbzm
z_CEzaMfUsH1Cai7o%CsgRK9Os*F*Q2-;}-WF#4N?uYdH9J-l(C%l@RnyKqE(C;;)^
z23-mEddW}`Y~klZloF!lBBfEwW>NxPA`MQ^^%<U&AP?(#BV0VP;%D4eN+xI-$#&kf
zolXPLgjYF$Io}2{YsJ=eGBnFuT${m8oGe-@%bXlIGNE!A-sOPL0+j0ru5|i>J~0bK
ze0%E`zxes|i(meHdiw0CFHUop4d3v3^XjE-eDGpJ4%oMJAm&;)W*E%vf;PM|t>Kmq
zxE!$W@Z`#!yZ#mxlPT_O;Q)4cbl?G<gGV12tj7gR#SG^7%+pGDZt*HV4dNOc{ph<U
zXxQG{$6h|<$kQa;;lpI@!NUh$fIHPe6Tg@Mu_)%n5837^8Gkp4LWo9w9(eJi@`m2!
zWujyd-3JzE985%aUD|ckJrf$GxdY^q$piIa9`Y=@Xfj)6<$&W!nCO(n?Sc&UnZ(E+
zn#>C~;-TXU$jm%d8I%J7PeK)Q&_%`vF82ZI9~-%wtMBH)dmo$F*fw{9^rWd42FO?C
z_kmggV6gD_pr{MlMF(c`LWa77k69L2tglUMGU?F-lloGIk(HlLk{4R|lDbS;$qO77
z<;82+vRHg_eaMy<DJ=(+IBi5(t-}FXbo)nE!r`M!AU`HCtVdI3SBqrum)gQneo-c(
z{*t%*g@kFVzDp-}b0JH;qL8*HV}Bn^Kee)H-gl6c9Zj94vv-=Tz2jAfmpUQQKBIlb
z@Bfw*ednDfseYBH;@k3zN#Qe{81Sz1x7ue}c*Y<TR^H*wLX#(yYJ1v%`T$%qjR{P$
zdDr@hCU@t@=hLm7+tcBrJ?)!&(}DIQPD=VUo2r47LlevN0m`W|(g#%sRTb0wiw`<!
zIhme*`*J$gNe=aRS9TwN{%E>?cr?8|emgyVsxigyboXE73ozz-e-hl5&I3(CKl|c|
z#*Tf_-+3RV{&_s#i-t@%Yae3K<I69;oc`lK{A<%^ye3imxLV0|OWz;s+GC6S>NcfD
z+_{KFqu1CA(L5HIp2#s93xG<t+1i+GIa3xnpll;$S=gI8<Y|e5)~Ax_M#!{|5&9LW
z^g3u<yy>=X15TH2NL<vDQG9Zcd7mJtJeG-UdmKf?&5WK2>^Cexhn{Wf#J)k1toM77
zN+3nikGn`SW+jcRpQx?ziI4>r{bSU=vPl{=^7H|KT;o1=E$rXA!nK(jj=v-^haGGq
z!3L0w`?(0%^SDq9O6B@0edEm*Fjk-+Kd5sj+|TVlWaOwHmnrFxRW~Apj(wQP&v})c
zRg4JpcnOfBh^AF}Y;@KLm;utVMJt%AYZKZDL0MP5I7dMl2ss=3KEM^D5H@549$S{q
zuWYMafNct&D}Df*n;me}#RIUK?4Z=xpiLHVsq08~F%kU_ZsF@A;QsYJtn~hLEr`z;
zy-hB(h5pfsV*D-GW-sGR0RBpbbfFDbEo_Hvc>$)6ydnRlU#AOf<3a0v7Z}R~T&J^1
zZMjW(tX*eg8(c>RZsGyWAEic}v783)Kb7`4!LcRD(TcH=Uidayp!4Xuoc(G#wW2@i
zQMU2!n4{jN@3USES*7Iw&o_jV9Moo5Y&&v#v=2nL<rdz$TnW)P*$|P1*7~DAv(1<@
zG;Nc$Ne9|L!dyS_n;tZDJJj{WuFY6oi4|BmI&<zVA3Kcn%qtlmS_0{qDtj?gc2JAW
zP8*yqT+rp|vW&9un1j&XIt!Vzk<^A<r3btC5?!RRU1bLl9rT&fAPd+Z;rwjG`v<x(
z*7=t6#RtNhWvnioxRSQrY^yK(pZNq4!s|Gp!89aG*KJb*@ms+%pt2z(^60Mjsgy72
zF<<nuKob~cfVs{E6`7Wd+lxvAMI%Ox`Bv)#GcIv?q)6$a!NZE@XadiNjb*DTT7mL#
zHwhi-A&WT=Z>2M{3c;`W+R_Z`bZiWLX$8s=w%uIxTrGlSDrC>pHN;S8$yA?uy9)IF
zmd-El>`r%fb*^}qPfF}hcez1%|8P3G|8RQz`0@1Q$z#6??!L~$c*5)U`-|z9zxbKo
z>T^%$h8JgN)4%*T|9blF*|X_V{qE@g{pk;W|M#aK|M<t#kAC!{>E69V=fe|WmpajA
zK9|$_Tn9PjDWQ|I=|VoDe|r9Idi&<o-<W^;-M7>0SFa|%SN-lCuUX6+f_OTO>o}Zy
za8tDVH#ZaOoS5;1#k71Io12?iEYNu%-_Yi|3+K>0t;4yx>sfW<_~Vd^rdJ_1DKymi
zj)I?hQlKTh2vo-opG&(%xK5Mrtnn1mCFit$at%B5Ih9p>KgGkjy!s^9<NDeWCb;R3
zo4@(A1?BLy0?Fi46nXOM(IXa&et&xS;Gtgw$hWjO_vbZ--1P37vSrKrt=hngA);}U
zJHg+)Ql0UAtuLQEoPPL&&!^w{>htL<o!9g2wKLg#`uwHBo9XzNH^^MNp5rgqD$b^R
zyLT+l!WXZu{QCLp>FZ~&d}BR*f@@C%iQ5Nm#zl`$rr;NC$pYbLPaaLb^TRLwYC_(M
z!y8{-zdiNq317cG_Kr&3`%N&eX6l>vp`EESXW-b!c>U=YezI)uo)>$0+YNQVYa!{I
zyf%>qDV}olx6!0Sy}dm?o=))bjHl*kH!@5+v&Z$dBHG14its3>->?&}D#n*A2y=}y
z*ZOFa^rKwAMDNnqGsue=TerDTSho9G7qQRrJBRYYeOUaEDsROco3N+9AoCmR)__7A
zaGuS0a$EKH+5O$=N1r{I?%(4Yt-VLc;}5d2vFI`w5<ACPW-U9?OeM{~=p#gabAJr2
z^-Zyjww-!vdz*2?X1b3F4ZXk@Tllc>H+4t&SU$jJUJX~vx&c!AK6&2+pCbEx?R~m5
zKs%*xeym*I53aV?p9hTodg1H8{zv~^_CoI{VzT2&h<z>hYco-S&p$gIHmV9Bdje*D
z$@PwDCSa4x4$H+7?Fx%s8%thk4iu1wj`bXzkT6RkmX&~aZ^?Rb4OG(Duqp{KUqMOI
z^k~mp(w^+OY-KZ<7hm;Zp9SP&S^x_HYy`XI$pnyzT1#Xg(*aGf>q2>O(BzUe^}$^n
zJR!n@*Z=2#{2!-(|FfS?Prv=vPkC{`A`e@0-6YK<gNxw+7vY&?kRJyc@30CSs5zMP
zlouaYWwDC`GYek3I&i=v&0RnKev%Km*k)meMVMT$CVTEU;b|@oW-NU9Nhciup&uO{
z`5?goiWprkj`K7ccn|9L_70{4-f1lw2Ms1g`DR#MD3#kBSSTT7p&V}is0)O^ke@!u
zp`2p&MS8OoyeI=5ITjtMX)ikF!aH?{E<auuzyt&nIiS&2*ds6HO?{G55-M5t&g6-S
zi1>`2We_r+;9Ta5xVJqK;qF5g3Go<ezdo?-(iVQVzta3tAJ|G+ebPoga?E+F;y6fl
zKDJT)1~KQ2o@kEsgXWS+yidl|y{H7+&w(4B>IAz=`!$0iro5=)v`Ml}{}%C1?3yN{
zp43VYo9LiB)`!FtQU}QM1eF(fWGp(863q)o!ryb}5arhdgXJp#y6xDmVzCjyh9@Fw
zL+7V<hVHU!qW(@3bbOO}(e8-%PLn%6_NxijElr;8XtKJayk(aswD4U8(xhG!tyesW
z`E+{w`ptBveV4nMc*4skQj()@b5am}?GM^l+`n8d?Q5P`Gx2|?KFHHv7t`sRyuR@6
zy<5}$M~Bk^@9@@syQBSxi6D#3`0o!EQ#Z0hnfzKzim7}VgSb=e>{Rx2BJ=w7+v)h_
zsV2YgrGIOB;1AOt`YyAVn(UtI<ffWY)uejq6BM_Z?2ER)zcW30BE9>E9s_s+?Uw45
z1wfAxKGDgfT_vBMJbXO;!SDVh$GjHxZk-URnY@p?p8=H?QycjX+Du2L_|G^@OcSKF
z%Z^Z9!HJ*rm3#vXaxn+nIUGpFr1TxE#U~_`at4wqK79ol4I%c8yiF2#j!U0YgdVZC
z&cAylO!Q*QGU|^I+!sWkZ?z~gJ?0h(A{)sPQ$}scLmSILmv8hr@2l9Z-+VI*BK#CH
z_E0T(^=wXH*>~9gy8lvs*XTxrQBST%i-A~bPN$5n<<e9CXcSrLD``9+Nz;Y0UntFf
z;wQ%<nVN*Jw8syTQ@YCAEE5rlE$R@y(zmIvIA~eeSpXEDKCZNtl&AM0{m=%IgMf`}
z;D5&T0A}CQRTMYPh-9p%M0+XoNjZ;jp|?-HO?@N}>drS83E*yGKvL^%d${10lToB*
z${NJ#MXgaPDybFw?~JgmdaSZqgTMPGl4XNzJ5SmlCS_#54cpKuyI4fqc7CNV`k(&k
z|FRdpHo^KE&~u(mi0r3j=6Y5etSYnX*;|@v;bME6FDpQ%(1x2duCvE}X`t=tE6A1X
zM|}{CGPAybcq~)HR-2+Vb(6a(f1UmaTV({-(Sc9V8717NK?6!8&9*s3A?m6a?G2(-
zRahD)2br$JrgglHKGMdv0A{(_r-j#fdn_sXs87r|4lq|~==?l~HKEkCDJ$Lr*ckH!
z;I924Ecups%&{HlVQku5<U^ZToAjj!gHGzb%iXf1Ge=+O;A6UFhvFJmycwj&<Z@L*
za;W5FR{1qg3H=v4ovuD2JaW?E1gQWUu#sba;W=T!x_pT{pRsMl$#yZDjBHmMItz9z
z7gqZX*d}uM3-H+tL+T(G&*MwLS)kK25#;g8wk<Z>*$jRfsdVK<BK9D3TSgBWHh|8v
z>x+Eny0aevjtMQxG1l`$;n2Eoa8A6)z>D4Bp#w8+<73-I4`?4oy#{Dem9FZ@Q>51S
zrfl-d4o(4Pp5v1Dc?d|+oxch3iE@#)z0v_p7x-AIJZ4TKqG$A2U{DuOJ_rkOWgl;t
zkU0Bl4?%JrwE(58dzI@tQ0!B7&O>z$#rY`bDE_9;E(>7!j`$(3z2D_+AiL9hoj3mc
z-~YSm#f#@YUh+D9o<ic<%;P7YO~3oQzc>Bf@BGembo5~Q{L3Fq2ZxVzUVC@Ccz4@R
zdvV=~n>Jb0IN_-poyYooQ0Lk_b%rl2%Dj5{YI^fFPkOOX^8N$oxw$^Uxr5IcIY0J!
zgvuj7{JKNxS2CRY`g)QINlU)J#SNf3@8vTe+(arrQm^awy9xUI>?&;PtCe(&$aw>M
zjwDo9J%7sB0-lAhUHnuzp!fAK+S%W8QW@6^Uo3Xi;s<iF@5K_zD?E!KI>*P3W#0S7
z@XtT{Y<m3ok;{;Uxx3RTPgd!iJm1sGSb$G$0~fhgMp;<+!rr0!!@~!M(`Sz#PG5fZ
z&`)7KeDuKQcW>36Ua;cZ-RFGQp1#4GY2LWJUp#p<eew86IlP~qzI;3V;+q%KbH&Kh
z5B>C1T@$<%etyp9Q_ko1;oEV%(dh@De>VN-tIz$(6kg3qpW)SqEWRD9JnuO_;94bh
zqPpYNojZ4-sjcY)DtBzZbDNxU?Lsta{qXQ$die0r*AnTMZ@4j7a?}CW7UAOq*Jrq{
z#^T|t`ZNqS{q&UTiSmx;+>`<PK9>`p1wkcEpNDg`4Q+7y4sWqayOZv|gdN`0<JTn8
z#`FW_=Vl`#%H1|XG}U97bbWmedt%w{(4VP?UU-b1@=q##jnay6#rxuDclzOH_ow^!
zv>$4}#2%`g(*vVzc(EB7?1t7cz9clwRMPy5K0@Tz`(x6JovwqmjdfnYhyyDgVdM*4
z((S__zhP^HoAPOxecpi0vR(nHeV@E<f=`kCzV<#{8lacBr%3tx<#jV$ZLdEM82zn<
zuYbH;W28U~0wu#P!N8CLcSSV~6j}!7*|0>tD=Ep?2str)>CVPQEZiJB5#gpnK_?Oe
z#R{dTfbB@dM-K>K_$7PRx{@Co637r+Y6W=G7hEE2isQx0GSk~K9G^tf-vgF-o{(z5
z>w%22a?uewgY)s*6Tdh47hnH!`d9z@UrtY-KAm3b!14UqQ|srB84bKTXmGK)U#I5_
z=Ncdx;NUa4;FV|`d^w;Rzr*B&$rB4@+`*uMe5dYWQ2yxpBZ(T={VgK>?nus0XDQzd
z_<5aSzcZQlLH7bzZ0Gw%UNDm^^4Rr-bo`Kwd65g7)K~pHajJs~wmE2fQbG`ofxUMS
zY8XxGj03XsbQ>v1F7>QHy`nE1IbO-;1D@?69~I<$$A=dVbI`NC`5a)9AAX1vm(0tA
z1OmPx=MxIiC^L3^@tR{Bnw2NASyaoQN>0f6fMz|V%Px5VMDt`rGWM?cW!1jmcOK17
zyJq61zJY}E5?M0N&-SpRI+Xwi^X6F=d(J?7wXcPgX2Au&@q==BA&4OG0$1rEOAg`W
zm3a8$eo-9SnYM?X$$;qcgGE(Z+KZNDgnq$Ceev()gvodIoh}#okrzcbluy`&=E<dK
z>EE)ayfF+P0^N)LL493Zo=wLWr<0~%)2*FJCmg(IUGkdXobXZKXWvb)cpCG?%jwPQ
zH?pBgsq9l?^5JeTKT)Rqc6NCJGOxH~vg~qc^2tO=C2+sv#NzZsCjh*bfRhgSv$KD9
zdieNYI(l$4?e1w`)BeS)MOg%++$@qanZ;*{s$E>bzQkRFcc$|boychNd49@!*-xj}
zFC~A*D+)E4yw8FfcY3{>UOs&_ohe^6r?j*Gu~;W>T^{aK%LMJiw12cS9cdEEs~7pU
zp!=EV)Ms8@NV#;%ru>igkEYKbJ((Uqe54aJ$?xR#5uSW1FWhySF_QvdrDF_&D{uM)
zbp|O6L5ZyaJMo20saqGiVb+0GJO@F96XcUc@ee}#WoVojmLw1{wJbU<U*Oqn)e&Ve
zQ*_nCa(qg?8+B4AHfq;7A8bg*lk_AF@e+*yra9?g-)t)VW0oP`U}-x6#r9&Y`zLvX
zXjAxT1^R;aPf=Z!&=4$M_nofiv@7iNq4|4X)`Rb;aEM7$AM`g;Ozxe3INDmnK<_M5
z_R5K(xP<z5ISkH27A(rf3h@(4*b0o5`l&!3cgmjis%+URh6d99n*diD!c!moLyIpE
zuvmbu$!r8AlDO>5!V9#>0&PdZ`GDK=MXvZUmu-2u$ZP}aMIPI>D5u8ynBB*|*;lP@
zcoPq#P!>rji0Y{P67JZ-)CYK^|K6YeeN%rM@E^j$*Cu#w;|jW34bJNmGef(Nfmc9w
zQuW)bntYPi$Kg7uQGc5)<{?0joA?BHdrjV(c1X|T%)o0Jur=>(G{P1-u$8_+`(C;Q
zn>^C^KGH{g=2h{`sSOePVi6=aqvrO$jL2rb7KC1H($M}S!9RMfKl+Lo=$OVRY17IK
z&@xTNx9AHgUKt2Lf)CpK0qplmJ~ELV=mM9aZ79m|81=}Fz5#7^<?~#bi(K<J-Pr1{
zNqEA3nED~CeDFVXfaCfo*Yt)?J6;!g(bJDH0XF%r%SKr8xBTFbWd^$>qZjh5W6O!n
zn0ZP%&G|s;JnFFlfcb{FAs>X2w{t;Y!|w$+&p|$X`7!nj>?m(+kVlh&*}qDIw5?8~
z!+Z!Zc~r1|%O-Xis~|r2TC5vdKC!sciyzpf{tAkpb0P?JJ_R+AApL<bAHNml93o|F
z9bl*uw&4}A<q{`P@FJI3(xu`5lBzU1knE4_C?eVRBh3w3OIhK$Z&%uM<nt%d$O0w}
z#KoA<wQ%Ph08c+`7}>KK1<yqVO(S0A39OQ|hqe;rhFtuhE|Pz4Jx73ff6n{qU)e8H
z2bfTwaKFj?oD8TN=JWA@Gr3#-hS=lJzM777PPV(px0Cj#chlX8CxQ4h1`AJ|&wCk=
z1*|)HqKETN?3`b4Bju^&xG^&~cyi8%tK2|}kA7vM#vgy*UAB76<UE)k=QwU=m4$C;
zb4?%%9$7GOxm{j_u7OlmP}%!bK$dG2zV@R)J^K2M!1Ev1E<R>qCN~B$24p|#I?WAv
zIs=i67ryXYKVN@Qd6DJ(Ki~UiEY5Wi%9@)(V~fR_2M-?knjMYKr$@Lh#_IxE2*geN
zVLV5Vn_;2jAJ^u5zJ19z?r(d+>2QDFi(rR`d()%GN7IwXkEZ>-17D-V7P`L4S3aH`
z^GSviFC2aT`2KXTx9cYW`QF&qPhU<iUL8+o7rc&=w@|5#{2epeNWs@G>Ux^&QqD*B
zA51^|>I*;pb<a1`-}Ryx*CpQ`pH6SMw!w7)*X^BY{~n(~x$A}7UDXlSKwiIkGri$;
zipr1IYl5>VOJBNw|6qD>ct|7G;$ZqK?LdA<M+cI*@0<F0YZb4KeD&(}^!Du=UjxB+
z>b%>Yrp|&8L2aaR`rZt6PUN`U;kFme`=I1dyK^URxiLW2KKt%{o>p}Gc;84~D9~RL
zk~cBqbNV`qnc0`2<IijxkRILA^Zd$xff@_M6GwGAO<&%>H~sMQqv?>h0?`j+c(#uY
z(0IlN6+y|)w|RYS0DxvHyohbc2U>TbEqMUY&CzBp6MUfsEWeFbIAFeIsExKko{QAl
zM(VO*oBd62pV<Oq*|vhzw%KdAiMI)#BKv*qjneB{*#?sEz#aHLdEE@O@w&bKJYe)U
z3t#_j7QWbhy7Sl>Xd+ph9nWm-Yzz#JL&u&kK?(7}!JQ6y<WMG+N^V^!GtHPaY>U{q
z<XQ*tLDd)#?e@XWNshz2xlK1Y14L|R2%BHa2+d-ZATOsi;7P$*YeR0_EzZ*hNE$$#
z5fOqRahr3>nY<#QZOL=6VyMGc)VZ+F0L~pc{W`-}uU}1X-@Nrfio0;$ym{jb<rrZ+
zq7ic!1AzmR23z4A06AFC7c?E>e<npt$TO&Ukx2tBlc^k_bwJX=l*KRFk-VUJu}yR@
zS}C1_Xbt2_L*rnU0ut5ahCy7vSkkuSnGh@kQ|P@DGBMe!#W)VQnN-Ds^Yej$yhV~O
z3kMuL{CGW2yL9{W(-DJ79J1hfbUIJO$<TZN#J(qc3QTZ#+DE9A;5j09Jax!MAYC0E
zt&0#6+?9`VP9k!E!#;L6sfbisV!j=od_<E2rI(Y|EchsmZI#=;>L7_OwgKz{_D(bo
zq8v~?>7bq^#>5P}@cAXHn(#TF*t1D(JONPNcCqsE-6is&7j=DN=ZsTEJ7T$8ernF{
zD0n(3Sv98!JnGIwIX2T9Su|r2w!WKnDW3%NPbNFQ2o7B_V+?fea>GA90PA;F>qoxI
zF8L}i`pE+E7l+(0B}*OQGk2(ok@`>_p+=p%zkJZiz`OJ5^pbCUU3l?|?|L1-emy<=
z<{O2tr#CNNPG`qw+BaoelZ{<X<Zdgz?bnL#dSAjHP6Sx^Vqc*R$is`a^hf3855Ve#
z;`Fsn0M6<2@qB0hj$bc$bpL4D)5Ml1Ht%Y}$m>M$kq^dRYLW=tqW_ELa`L(i*SBno
z&&kT0w`bGa*C*5Y$%Xe7URQW{bm%sH{_Qi<V@Q*`s()=&nT!L%@LhUzE1q84XQHi%
zDD8o}?kmb4-171P9^XYf)Ct$&;n8%&s~%N9?w{1V+Jd~?cd$YKqT?vSj{B=BBV&M2
z@hq97py_f`Xr4%fs)THgW3c1yCxAuZC^$kl^2QaVHwsrG3t0GE;<BA%MEKktXx{-v
zx8(?ONMqIIb(%7uko^UFFl<xx2G923(QjdwOvVg!Lbt49ChTn;f=|qlQ+>Ezs0XF-
z*~SH+-cOLDKhk%k>~Rr*<qiHMA9Rd3FKCpk`DGgxLg4b#Pnrq0M#aEbmj)2(pI8}D
zMD*r;*|x<m19P6*+H)z8>ikga@yh*1LjA-UdMv~d_ld)N0!;?aT>r{5R*-@Q8yQk&
z$2H;9wfjuTMBh?r4>0Trx3&nDX|Ns$Ws!_+K&H<Nr}>KywevQGG8&;AjQNTxAFBw0
zC*>l}eqfvYoX{_l;}&hpny;-k%?Y$jDDmGUmw{Koc74Rxm87;XI`0CZ@hP|!OiSWF
z|I`0+5dQOp|I6QF;cM}FfyIR$r*4EBL_SG>1FqBYK4199ASIA|DR>?-lnZc;-;mcP
ze4n!9#6G~jk+CR18pn|E__1jh-VJ%dvyH+>7<nUHXYYzW;TqWyuE>y9xJmO<?T&KW
z%R*g7hAp{f?4@2!#uY<r+q0q#$eb5*8s?|Qhki2iX{XjBn$bd}gJ`=OJ<1YVnK77G
za*mYLs67akLdi6OUi$;g*h)=_mKo}9(IYN8sOMLe9PE$vzK!1LipH!~yzA&bs%Ri=
z+g<0D0aeuU4VII^ZS<yX_LrGgHkqepMgo12K`l1QpCO-4H=bCV$_iT3M{t(Kv#4z(
zolv9#XqjyWNJpp$g7p26a-QR)#gNYtf5_9YoG&;f9vJ-ATl`r3TeV3kjCg}-CMGOt
z#Ht=SR+<usjQB&n5|V5ipwUZvn4ma)HvW~Y{!UqFZELJ2Aa;nWE{liGoUe(bLtkXP
zO)MJq-FX5g2Q|NW^f!QWoQCAUrg+54sIqc;?9b*$^a>=O1;#P5VJ(ACYw&f1g78#M
zZv$)jwe2fOrQ>^B=%Dj%V*9d`FRnpXBof}i>4nS45_L@?MB6e~hNit*=U{ht_H>?i
zZ$8g3aPFXU8E)F-x&zmSc(aF}GP;w`Q@DRgl<QW0J)H8<nH-~(dE5Z)#T@)szLuvg
z%eS0k-16rHSSZNTUOZvc=OI-N1(yu>tsy&Nv`s@r9|OdFZH$AL^qKCH2XhbdY`c-o
zc`4T$FmI}O$uo36xg>jBi(t_%`x*Y_8cf^p^)l6G7RV^0cqxqQi1T8u4Pb+M-_toV
zH<Ys<a*QPp-pWMG_xI0O#9?6t-_$-HtK=cC{PU+vbl%Rj1ztgUaIil;c<{ham3{u%
z<9aP7*Ojuc_WJc3KRI=zbNNU2kNhOzt2f8f&%b^){o>mf)9bft8!zVYDn@izY`UDU
zDM6=9l$#skAKiaAef8xR)5AwBK;=9Cl!paWo*X-t?-yrW^LXcLB71vxriTv>rY8!d
zedGU|lj)^=c=qb8Uzd1##$p{$$f>Rl4y3c^{O|`~e3P(%K6i9@sF3=jKHt24HNANC
zdV2NxtrzdS&#E0cr)1IcuCFz6-7`1v_6_4{*DUx^M&63U!k-tZly~<bUxWSdZsKh_
zeCLaHN1g?+8ESt?KTe<UajW_pvgudIB2u{F4!hQ}e~M?U@<%o|1KgVScWzHVczAF6
z^6~z($F;NUjf1@bM}L*pzi{HdzT040#YP$*Vjv3`O-mU0T&D$Qw(*6(X_NY)Jw}+X
zPZapPok}Y3wbInm3}YTjt>Vo*N0@ouUWG32wIIdO?zTLxgH2h|Jw5^7$JPj~mKMcc
z^(nznjavYzNFql5znTY|kp32{za97o|HXfH>(+nskN!JVr*@tyeP4(nqMk{?YPT$_
zU&l?49EAWW#iCxdA|TlyGWq+0p8jdn;J!HHIHGMP68Jr1p_3H&<j$D3rfK7j6Ynjk
z;Q7mK78tq{BMdRnrG<>PFC<z9P`;S)hGguUuGl|0@9MCMEHO|jukN_KtJen=4YH@F
zCtm1cK;`v?EI85GdF1yUcjS1WrOY;EXBZSYa5FagLKzbW*=CUE&WT>c$ekW2wlaR@
z4h#+^9GrYGr~^78&O&Kdm43+b`Z<6Sa?mx+gOBX$kALLj9iz(Vv}AlimVsY-sfT3J
zlPNOd1FQ_pyfTe?pdS0;Co<?0C%fD|lZ8wI(qsrVe(;3ZrOJ7!0}^61cytR?U;ZwY
zCNw!9rOo7(>(mDwcMaKeJu9uec&r-x?kSW9?Nq%?$cA`Wah+R7hcONO<iTVDg59xH
zFtYMRRr1oW+s+G&p5)*z^%i^h#;@y@{=s4g6R6Z#+6d_OpbV&c@dbyS9kSe&L>do8
ztUBzr@dObevX;0Hj!Iy@Cv)vXpA_Jq^Q`hD7i{xS44*#ciJm7Gp>-ehVwfvds@Rea
z_T4T@(68JUou6<f7amKL_R}NCNdnt!>>nnDqt7IVNeiz-goeB)G786M+=cW~6R_jy
z^^0fIv#-CMUOs&`ogJUr27p<aneOb~op$zga&R|KMiQ9Zs&!PJy#4^&J<(<%(33~?
z2i{wK`sQr9JYS67-Mus2e|#_<Xfk(jxIf+FH3FL8@x&JO-1|sRRBnCHB-E<;F>&IR
zki7evg*84<%w&$fzNbkpi)r}r?Ca;#>G8RBfjjb_I<yVdCwGNCeDuiQ0hHJE8Vwe|
zG7)u^RhiK9iI0f%9iBeZ2r}*OY7%~Q=sU9RPzLISeye~1b}{+{?TF_<#&G@AmNFe0
zOyfd~)3(uI|1L~qQPG1UdFugowp6`M9-dIavMhKcFzN-a)+KWO=31LRH@AT$(W5U#
zzc2KA@xyF*sUK|g{+u?%X!Fd4Ad>9cRw!0*q?Cv@TO_xf_x})}7ps)1AN=M7S1Iz(
z--4KQ9Z)W4+FC{uDFaH8x0sb8S1fJ3g<{{`tHtE^!K;&CL@2<u0qM#wY(rMuPm@}f
zfZrAk4S;T3ye!}-$OyF#t;CZ?+6vM4Tyxw%0095=Nkl<Zcn&8<ZQCS`2_DB4+J3Z2
zAp@^zfmwG3@`#4zMq$C6EM)XKAiRYt_I^S}faIvT>^tZ~ES!x3c@{Zf=7Jhp26-p7
zrQAfqLn{nfuOK3;pM6bT4%&eb9d(h6Ym+usX_bhbS=@-zCRn4kosRC``!9Z*3t#^Y
z9(V5E1{N1m7rzFO!>?fDI-M0&GGltwaS2q}U6Q3;1JK={3O4imK0a@gg#x`#YDbAr
z&P#N2Hu<D`$0!(Wc>li6reGUyq-}C<vehulY{-6JKdytdEF(nruK)-m2G~$95?JL>
ziIpwzxif^m=2yI~?uA~E8xl5jXL}3Yu!w=Ucg|UO{8nJ@X>JS8I-zy4^=w5|v(YuS
z$^y_mW|e)X2Z5#wQ&<=aeGT7_E}xk5&YmrDbN?-tW$JLcY3oz-T%~P$lg~E&3fN}%
zy1a6{y$Z+^HtT$h8)kZD6^2b25mFaMx9gh?PH=rSF?Kn{AZ?$7(*jrhTYLa~UP3IM
z%?=8tDb5ix<>fbvL^XCf*<(NL%-KRW!oP4Q3YRSPyNu)`aNfcBvVLOQmeu0FQQIk7
zf|%`e;@O9lT=BeJr`wZ74_NRf*DSzs$}WIe2GCErMmwPwBH0L!dChw5h@jhO0p)D{
zwcY|!UZt&3n84_I{{+}K$#%%vu(rJo@TX7%iI3@wFv^m9Jj0u|JNqz*#?Q0_z&K6q
zdBMnw0-W2jSilXPT$JZH!u20`oJ-u9c646J0tAZ+I)kl!C^wsWp`hDVb;C6&zMpl;
zVu5&^tLKIa%8#!oQ}-74ISp;s3rpVze4Rk$=<)JS4uRfZSXk-vSAb&)fUM7VoD$u6
z<~kVH$$UO9;F_IvaaH3q*F&)Dd?d$p09lf~4=V3_-^7kpFLbe}D83g>Sp1TYd@G*?
z$G%R$4XS(#jO<U|yq?~?c{`m+*4NY|Po8`~o*PjQWRGhK{hB{~eDL6C`q2-+oc_x1
ze>Hv1IsAvq=|c6+Q&lft@yf$n(}Vj*eyR%jmv7FepZ(&gpZt1z%z{rAb@+?}Pg8N-
zf>xx?e9g(%q^4<CZGK;N9zA-fJo%Kuffu5<esFJZ*L8As!Yd5V)aDn{UA5=Y;oa%;
zM~Bk`u0_b+8|C-x6$@X-)9d52>G<T_uWF>;I49;)5B@yME_Hg#*G}+_{5j>l*Kes(
zId}!(3zhGsZ1IE`i+$XfO)FAwTwi5Dmv680WFt?}alM9XBh){6Qr?#@U;2TBJQc|e
z!Zi}fUv1p4)0gw=P6GXuIDJXMok(py-`t7?hPv7xwgCGVYV-lilZGJN`8QATq3YvD
zPxhwI@2h@x!ndE5PvWGbA)E@Ed25*^o!A`GAzhIHIL}XPX@^BF<qocxYXB?$Iv!!9
zfvseOvu}k%L|p}KxN0+g0ypK`=Ofoam-kwb`pWl@*lYNh+^5>v2B<D08g`PZDftax
zDPz`>p<msL5$FR$`~4Om3t#{6AF~dVlB--}oq~3S<f5B8l?hdGQ7K;Jy=S%&<Jd^o
zWQUvQ79eI*@pRTQmQ4UNV}4^k&<0jhO)A?hkcxFK`ppV~%GY+yDt_tAOhp#8ZR`pP
zPYFyY8fau%94RyAauR7^J>#y049X0wK3Fo5xQiSImiqQo4Ukz>P>$5C1}+BD9ED=E
zx8&=S47WiJ!r<`~+r%=Vf!W{oDn4~(n*5C4Jpi!q!lDZG!hl^iEyJKjt^y&ZLxU&+
z>7dj?)s|?{H>dLENtqqnxYWTFNw>o&2Y|{sFH~^wmt5*M;E63s@>9rUgh@oH@*|;?
zPbg=)wrz&Oi-IYeqYfdt9YCsF)xfb=GRTJ<xtbWDrrhSI_uPJni)n3Vgl|?<th^wp
z3~m$2^WCpY9b_rn@&p(Q3|WwLyBHG)??ou@UkWTPg@9~u^7nUs%yWhj5Q?>wGY4IK
z!d5PtSJ^60*E8QxvksO8wk4j+#y>p(IwG)X!G7hZth(O3psLstMe;;eKQGQnGksJv
z`lAeW|IOVzcRU$#`9vch<OyzX+3-c}I`I3h9SL!f-HT@WUB1@@O%unLuU<^wJ^gz6
z_Lu*DdhzYI{$3V~N#5tA&Kn?kVvTWu1urIeOs4RI`sr7N`qiJbALUg0@y&xgp~o0-
zF&)1?(WLO)Pmj5<0$z3a;PKJ4e~|C;@x<4?{d=~<M9ULwPogO^<&#p?^X(6+115#%
zd7_lNwYYoAcW6oAu1;cjn)5<7o_+ITI_rfm$|XCv;C?5+dA;K2Uwtus_4yam&K?Vr
z(w8n}p}yeLn$$NsEJX6CJ^b76C978DM1fZlFfr%k$WNIfN4rRy{L*iN7s2&vM0Sa?
zDW(O98fm3OAr~!sn*K_cxH7L;WU=ra+*#sR@%^1W>3K0(GTt8)stbrzB@n+C>78XZ
z0&%TcJa9HfRAvkbQ1v*1J+p=D;OuO*Klmg}e0G8GyRL|X_o9gRu98oA#3xPva2841
zMOTf99#1xPcoNPoO5AC10d2)AXhRUabinBg>}%8^#~12PR%3%akVj2G$_#w4FW{!k
z&4mPrt0bsHHcbIVvG9Q&k~ZQ)1vQBbC^hMUftivhAIg^L6@~3YQ0j|h3C+}5O|W1M
z*k*)_^r%~gi`3{$GJ@Wa<BBQh`fv=P=!2zfdqWCu)DRY3Y^P45=lW3Iu0AnRZg0en
z%#vq8kv)f$jO~?*%@GT_Jg6?-RorFK|L^@4pZNN}df}^KKxhAOYoz~b;QjMggq!k*
zx8^P>uw`#-2;XCY;_Mp>jCuXK<!;}$H74{;`#0uSyq58pw?!VB$G-utvwKs1o4jq3
zpKNzX8@nN2;3mHoU_V1{u+5m^qOa>_#zboz_E<|f62P4$6c!2Wwe1n+d>AuYbptlb
zvzZp9!oq7$m(g?4S!5P(RyGF|FmDttn01;zz??_L-KCObLs$&1^51~#c%gSln%|W&
zE8>ixZzgb4&Hj|k25h6RXy7%cY#9An@Xa#9qdu-LJL%p(*PN0q`B86e1H64$!$|9M
zAmEBE{oBKN?DerpJzW9!Gd)`S7&{#TCKWB`>l<Y!xIRiI3I?-Gb1DSq3<_SfG8ucu
z?C(sQA-X{8FZ|wz>o`uz_9!E<NiAYz(d{_Uwi`&>Mms84Xvq<1d#%Y?7b&<tk0{QZ
zZSfkQmqZ1kqA2cuXB=eght<yQ17P-$!TLB{<!{m;7l6-vw*-pcW!)4<_ERA?JN#zB
zeyN;Sic=5{`zJj{8(iQdQ<dc*!RILoeJwz;JG=L$2M-=j_a8i&?(RPDQ)1ZQ=2E_c
z#A1NY4MfcoUU}<DuFv#2d-CI2h0p7;Ap6ckc6_6)sSAv4xea>Zi*qQSm&!nVPJgpc
zLtYId-q*oc9PtgEJGn`m>vnkxD*e0{K}fT>#WgJ!ly-LR$bYVFUEr{7`3Bit)sd3t
zeCO)&a^iJ|e1D&Fddj=Ezc)R3{A7Cg;QsXP;>-&%uiqT!x}e5-dJ_v@M+Yo49olC;
z1;GLtucgG!<A;aS-}sMyZ~ALL{+(%W_pZO8#S>M}{M|Om$iJh*z3Jh@2h;xU-t_9t
z>Gbb^@!j;@vscr_Ic@Y|I+lMd1hJq-`KZ&}G@P4v0m{Rt7!LOL{rbftK5?+GKJn<`
zbaZs&C%gEj7e3xqSK7Jre!72fXL@|ZH^6UAXBQu)m-zLH?{u9{Cuf(_@v&ka4B$pz
zRUivgckWWBsV9H`Tr@Sp^j`IMTQbx&b;5VXp1*wKC&~DJ`k9{;i|_crTWt9LIBn2l
z6pLoIy|?e{94yqme*M}{eqB&@^rh?mpuBkEtDokB$2Zh2<p*JN@9OJxW4~%}?n6J$
zxqS9NDcEP*)89A7s_q#}kUHAinSS@PgXzf;ubH%nRF-2v<K&ACrx7aJ5@$bK(uvI>
z9nuvUV9W!-G5@}ng8p{ht2AI+dZZ)|D6P+CsY77X21lEQ>+%J%zh4hy`L=@8N4wkd
z*ajby`&2vI0M(3oPC1a$F!B+wlrd|`(68>r2;EP9GuF1jW_|Ih!#jWTkN?h3DvgAT
zmV#wNA94}hnTSm;V^#QNV<wmnBF23QrFP;B1G$4fa@hijQ=tpjw8q67x@^S&czfO;
zEvqav4tCz@NstdAZW`c(r$MnA*;MqxFrRcUcvu%Z9ZKHhsK{qyNw8JKg+~s8ImGwy
zsPg$v5*?t$XOO2JBhLj;Z0BhVh+N3`lQEL@g;5Ud8tggPGC|?MyvqT?2VU-Q$m7R(
zr6CvSlg5U>+a+0l=L;Lk#}_T}346+~j@91a6Pqr#apjo{s-;`O-8u>iDXXN?AR)~$
zbWW~40g^8D-V+M{s1u8_1qq!A5-Obdc@hIpx_*)f8PlAB^P%e9N6QS*?U%+*4k?-@
zFfI#Jct*mbbt9yI;qdkEJg`IX37GzU0p0tBf(KZWv@wnRM8~@7wCf=I5eH8mp|5uh
z*MTqJRKo}9K#*bg9p_M6%3HQU@sD)$5YVP<uw$SV?PU{7X*%1e2_6Yzml1XK8KkXd
z75Lt|O!O7!6?4{izVH>hJy{rHk}Dh<`t!-C9IzehL>~sP*@N!8%kZVzvL`y0!!PPf
zE~JDCrl)QJ-e1jou)q1{m($Pw?Z2I#{^FbIOq0zI@8T!>h+i)#KY03yMK7Mx<ZdDM
zFD7_Qb_u;0;&*auQjA;pz~UDZC7$SH;VDnIAqYQDeC;3XYofa2*AQ^06GgBMl}`n<
z|5*3_T4kr!nH1&wU#dIp+f_XCs7vmcq7T96o!Rr_`JKN)`kipuX5o$RtNFD6^gYp8
z5aIOz2bv%r>>sG?`@ZWduQ`x?`haxZH!EL!^qqHA&-UHh1ly%;X&eA2+!ujiOGOMi
zQh;i3lMBqHn}L-Nv|azqbbU#u^qoQ$Ir2~ixX;;^jp+}s>KYlVEOp;~6)kX=r(`YH
z_KOD3{mK4R*=5smjOpUjFH<M*$Ukz{#g6o5n@A}(kbrIkKPx}v20Pk{N4rF_9_m3u
z^@$wvNuxu;?pP{()65Gw^`$^%m{ts<jYr37HfY4@<lAY5=FKvSlOhS~l!{=mq-pTk
z#t*xSfI5o#W(Dbdk}n#;9Yd(#`Up1bd4M*+A(Nt4ybh5vl&aHa6rWAmkL~8AErglm
z1Sxp@s$2%s;(dthES6f~8HbUfj)-B4-ZstSZ_Ju)yBGoIQ4-acA`%(QXa6M8){0lb
z?T&~c*qomcf$HO*|JncRA5`kk9sc+K>EF^kxxFIp@Ee9M->(TvKT9a$I$dS(Z>BXA
zPZ9i080|E^8NLrVw=H^3ZPH3^o4jSA8(nZ<l)a)SXnP$Rw$U0!Uh~<GL5wRhz(||v
z#7!GyOy7X!Z3_#K9_&DIyx=pv1L>kSubD||JlW(IgvN2d2nDlz8QRose3LeYapzdm
zx5+}>PTxke?BLhPo72L5JRf-08fZ-qzUU&i(AS~iQ|L*D7JQ_a9C?+@72Xg&MlWOw
zFQKy8;1h?&VV}9L@kQ7n<9TgN8f`3T<D2Qiom3e4n=&JQ9Z#F`C%8i;J(t_`v6%2T
zZ3W-tDIWNX%!r5Iz&tF7xijB$7sWoetGZpw0INV$zY&-p6kQP7BD+pEEy4I~r)8SI
zrU9EYqTr&{&;OPq!FHsPC;;=^Rfy}|#6_-uW{vcjlys%${@3XQBGDZce^X}epa~I)
zZ0rqSTKG=ZuVF*F0P4p>>~<i1b7;L!4ORgM*$6AU^fyIxOPYrp-fmCaoP449T=Oz+
z+B{J>oi0w_PG={_zL}ew8+Uyk%=tz3C92jzHc98)X=$sRyC@@oOe@aYAEFoj7~Ejj
zLfa+ge2%LhIqwnRhiL(qx1i~!DKrN<`g!9~_D6#RY@PIIQ$McHu`rZt0y(e4ZbGi}
z<n@8MnV54z+mjvg;MIjkNBh3U!HvDh@og;l@dqDE9(&5y3o<N_y?-}7Ji0gi=qtYa
zb<fuwc=OQN#e08$>-gkidV9(nVJ>wJuXB2x>)*c14aD!&rm8FX&h-M$_fsmBhga8m
ztf6h|{8asd>u>nSTWnsxe(UdTz4Z6CPI8?={!wo%INd+u+g<z9!QLJ9lPoG`%uRW?
zj*|0IZqDcRio6-?e0uXX-zWpWR6pYy-N8QJH#?a274NAG_k7)CCpTVFH|nWe2js?4
z+K0tfo?c^Ni$$%sZ{Irbbk|FTw>&i`+W>7wy>}1rwH^7&^$xDBQU|`y<I15w<a(pB
z5f7cTbNp!921ppcFhM}V`wK=$S3ToH_Zi3HAMEj}$$QiOJ^7;axDT!JHwed9u)&W_
zIbw(A@h@8QAuhBfUhqJZ7uqI&!Q=Y^D>=fLb~!qbcC>;~J}~GP@2bpp+6W)xlg%je
zQF+0yo#&0fxS)C;>T8a_hgoae&aVR2Uc#_{{MwM+WCh*NNB9?i_D}x7oxk<RfAW(h
z+oekI#rs<o&=)J}f(m4K-3U%g%Db8pA9xZ61^$kL0~HydC#|jXVk4-0jE0c}xETuh
z8*eVqv=&(M07B=GfbaOs;E~K>bvlv@R+_aTvM#@{RUeIy9=k>=!6vvZP*y4Ic%e&!
zDhEOqpS08a;;>?WE60mq$iQcygPsEcjY{n6`{e3;7}T@yLq0x;h(10A=XQgpG}qus
zGT3<WO96#6pEd*pP%nLNd(d?pGBulVs0k)IP+zubT8K8sh9r^b^Q#n9hW0ZAa0x>b
zZW96@UdlJ>fbw863$5ad;Jb-TO`fzB3OUHcmg(kGYNbklE;Ds4J<BNxHxki{g{H=j
zWmA$glz?`ImW3H<Q}?PzXJu<;rw+8Ss91<O$SXJ>+Pdo_8%1qPAUW%+?x;^s_PT7H
zm;P!wt7q?!O-;Oj)<H4{;OH78v(&BZxjqHx*uUtb;k?A`K5Hw;TPMD?<z9qvcIG)L
zh@~XSLztMEWEyq8sP%#zIQxLhY<rPo1HC6%mQ-^f$14|@483{%diwd#e>Q#n(|<R;
zQ`)*J7T>tyuA|$E*(dSi&hG6jPVF)gVi7qLBkIn@!50ND%BtRYvg`cx(sz0B`2YEt
zCRphn0m>rVd^e28`*~f!-Fv(~LK9II1T~o@rmbBDx$iNg>fKKR(uTCRewpa0PNXNp
zS(uWK^ku#i#$8=DLLJqxWq&Xs8;nDK8dFRr!hX%7<ahk!SiWb*xJ3SYEU*)9<(=>b
z+mVmTllFH8<R>Co8RaKdw-x0A5gK`tMCd@IFLl~d2xyC~Fp6f&E*Is6Zk8JcJU-fv
zNUaOk8#%{#(u>RNE0i;J9UXtSOzDh=@g?@s2N1F@7X_6st>S)JI<Z|gohR)noAfXA
zu?M-FavdsoAE<KRBa-B0sF-}S*W_d^c`DrjE^*l*bRF70P(B+hq7CCt$-69W18heq
zbs^OCmpmQNag``G%@!Od9a`h=-<ByJ`e8$K9IAd!l&_X0VoMFwugYe!=<%B3zzwDK
zwQ)>3Itdst*LXrTyEl+1w=l~%Z9ukT91-fmMb4Zg4XIatf|6-oMKgrXrpsC~wvj|=
zY`4hF{G~n`f~AE^VWwHEWwArF;xBm8M4>DYp849kEn24Zm$H+-HT~H?|L4E8g|GkJ
zfBH8yPbPnRXWG~nK+@NTPlA~KmEfklMqcw*v=C<A0t0^wT$Nv7wmpPxu*ut$bLu1d
z9nUhv18%Hiqv2m?6WGS<JUVQ`CckCab`Wu<kFcqOTnE#)$y@_h<*dD;FTAyU)1$#=
zq;m{0aSo72y}&gcxsGKGquw@r4CoQ9X-SX3JZr@tf6xXS<`+KFkZ~A_vEmCxou(&R
z(-+!Wr(vXhj83#pFY$`jaxTQe@34*5^oEh{<6(%A^F#v>PFi6;UR0X-39>%YnAG;7
zocTf0wy8g|&Z=Z1S9(blTXFdSrDdElUc91*x1^IuI{0;RiYpC_{t@@=G_(ptqrRd}
z$J)|ABNM(h!8`4^1L?6dF2NBS14#C<*6YB25xO-Lh;mF%s%2>p6U-1Lbas5ZKw4Xe
z@D|z{U-661X6EU)$S?JoKC^^SOQ=G4W(d~2kcex%0o~+Ps9FiVfb^I@l)yH8Nyp@d
zu7dkb2S=fQJYW|9`FM!hI{z`O<pJo9wgw`@XrqbDf#)RibVdoCYwA4n{QPt}J3XFG
zPu}P}^-Sk7+#t&NOuhxixvAxyxmulbgn8UfH=FZwzYU}h2WGzUCA}@+U*gE+JSeye
zu$bKntqaJKx3Q7f{Z@Zp(^0*2c}y;xbAQegS@7v|<zXOIx~}U9=iHhY85WTM-k8Eu
zUk7{l?2q5*A|CaPZRah(7hk@S3{kEFOw+@oz3F#;`1$nc!J)5@oL}(muJ_Z4&i&t>
zo=>O5Z`E#R7xs@uByNoNlT7wOWvw=}ubkWZnnTJ$anh#XY3FmU6`Ws8$H!;W+c!LU
zbv(U2KJ}AW_{$>P;r_kp?k)Lp{ML(De&R~?L>Ya}je2zDzVpqzyteQRe@|ue_;`B#
z`qlL6<qO$AcAaqDg_~~qmU*tn@LELqqdMZ7V8s0<A#77NUNgy)d&K_coa4)RVUcSq
z)IZluC_mTV$cM$XebLjNOCL$w()J0t=CMG|{nIDMKuLw9%%(-peA#f!Z_p*Lqy63K
z(c!&mkH_fjMtIR(Y|QkTcQu}m$D|cpmLxai6Jo#f>!9r*8*HH=Gx9s$#5)ZbhgNAI
z{0?ScfzbwXUA{4FgH8HXotycQSzt504L&CGDK@qNYW8Was}AmQ5%@%~rfvNyAT6#N
zb^j$@{qt9W@;-L9gwz*dfm-<b<DX=as+pd0ENM_8Cu80T6G|46V=fpGWe3g$z}ktj
zA44r!1^+GSf1r1=4E#yd0dWx$U@#ZtTNZw6MTb}$Beip)SD48hq>?h-W0zfuEpN@5
zJuOM4od|^}YAP>mwjx5e?IcSsbOuYFj^Ifw4isMKQs6ONUpT&dFBg<K5ID_~T0F7U
z7nySrmAb)i^6>#rIWVyDO)1yCn7%v1vZ@CUpm~Zb{0!vSgq{-^)h)3@e0B%TD4y~;
z0LoV(7bw#v*!5zYV&_qnhmWAHJJ}*OP7<yTR%F3)N+|@e5k!C5tmMhZLReNz-=n;y
z@#-!>(qz&$NScJc(q*H#uzhf&ZaDyvwvM7a>1V{m)**k=p_E71cU^N~n><3}8%;T&
z$GaRPrQu1F+Ewzg606vtU*HqedGQ3_C<v0|<xuS=IuU&YKBc1%^uf6gyg=-z1?>kD
zrA!FytMYX{;3`?MsVp<mikveOuqF68G;CnA@&YGc=-AKe4e&*IsNLHxs!V+NpeeeL
z=npQR`ILrSsBh7skt&2duno|!?E}6*&_+`p_mA}VSaQbl<DDYza=N@YpWeKAJw5&I
z+v$xaRq$E-+1cei=9-kL4QV%;<SyU#V*lLXX)pFo_GjflIpqhq`xDu#-Kk%HIQ5)7
z?`ZOOG~GXZAiYd7X%B$Dwxfv*uL{VASmmqy1$g2MIdu9!PFXVv2jn01r<hkOP-efI
zTlN_j7*i-e;8!f1N`}EB%BamaedaDth^l<l<K@LWv&hR&x5*aejP1~N)bG4#qu@nU
z%A|kBtr{aBs6N~c<$os4)gHvA8@0T*YS!$x(I<A=I<=9Vj$4-9KLQH=mK!~QF%miI
zF8N2#!sU+wJUAumcBennXDNeZdZE+%)Cc&|>&d(I@gw%U?-{56(Kl^eI+7=212W0a
zSs=IJBdEu5(%`~ePu2^W4YV>Nos<WH>jXb+QZiPs1p{Tku=ic7u)%L<=NMs=3+Xl_
zujB|hw3Ho6#g-S{{)k3C{47f-U5ggazC@3<CAPXk$KHJ^ZT}&Hn+BfXsPpU!6w!L{
zbh5<B6Ik-3KOxdG*d~Mq*p_go!YDtWFXfIn{SZz9ka~rfas#d(>w6oz65Uef#voYa
z6i_o+PUE+~g=u{l;h|@85tMC7;7RzdGjOMkQmet)=&xN$vAqwed7&U;hoV<>#T`OO
zeHgVy@%c%YACP0%t?AGH?6<e@^`HD7{*OP=P$qvUXw|$n>ES?p`gOsj9O~s`;X25?
z;Cg6pN3Umm!>;M%SMes?<i{qo-<x3*HtEhgsA1GKeUvSHGd2wJrdQf_nl`|7^oC9T
zHrY{T#otCN80j;QxWYWIB5aW-G<+&=giX5hD8!_J-!d>k6Gr+PJPNF27CW0dBRxFE
zh(&JA3vANH^os7nEc`xpnK#Of_>dmyz&1Kj@J)H}>vYz%O?vS7TFIg(y_SQp&}RDJ
z&x$_8S9qm^Jx0LzkYFyXfkwzEWX2p!8`5*QVgo@A8$ptHd@yRel5>Arj#nMV^tHJE
z;8*#Du$g0`**9>DM9(_@gKw)7q>Jy}*H~z+Pg!&@b^WgB;G5~-NfUdkIC(j3-U5s4
z=6<yBfE+(69p=U{CA^SIu0q69<~5DD>ZN~X6Vq2@tkNAiN9$UC3B^IIXk&g|=L@g-
zu^G9@nxH~2w5tKFCB4|~&}ld9N|w<4;O*P?u;`Gklx<%P;f~O>?P0AiiZ`&FScTZ9
z<g`n0A)#Ajk@L9>A1bJ{4^O;~|MGk~J$^GCzkTa-a?VN3`|wWkI)BBwUva2&Q(ym3
z7V1>=HTSKd&n=h$mps7@9^|I5M9d#^9-u$|`WLuq-V|A46`;hR1raG1--?wov|P%Q
zG>a%ZKA+DG&_L|w802dQgu(%IkvTlrpB_ED?>Ea}iyL6G_{H@Y5!kOK$2AO+SSG?v
z7R-L|#b?u(PaaB`R|o36|J?__n(s{K^elp%o%3qJQ@;w4YmF?@aehzRu!z;R+&48=
z(Vi?U<$7gq_`P$RYnGe~^K=;bD@c)R1lT`0Ih|gsKk$9A)6)~bnsEQ#&U7yP>C^9~
z3;D$>8~N^5=K73PE{t#X<*6^8OylXe;Ahj>>FM<1<*Vt%tC!R9n>W+h*{K(<con0M
z>++R)ie45Bd2<fX*P*)Yk*D4mq}6Jv<M+JYkiL_)u|*NMMs-iV^L8!sq>~#~X<Mw|
zzqW4M!~|Z$m}@uH=J+T)I)U_oZa@54aQ8v{5KezO+}oWV?BAVseMu*J{-HO=1=0hW
zww76Bi93iq)0Q~Fz6>$YazG!yfz})OYnuyf%0zPwky$t!CJbx4+w<NAo4lob9o8_9
zN88yq<^}w^{F?yPXkTOFd)P)2Up`%91Zzvb>hjkp(+FO}D%<Ozt@Fsf4P@c#PyX&t
zvKt`aZ3O+#FAVKW7Gwx<R0fG_r`5kNLPl1bkh76<R#6ythLW(24$&KB1m+E~z<rI(
z6jOGSXMLPt=qgeN(feW#QMNf?kP4D?wy}d8<tu^ODlTMB6N=~wrZMQkM91e0bUCo4
zyy9vw^xY18K-4i8sCm+YMKKn?czkzv+T~y*+>2x!Xt<bcd+{`jSPZ^(0hR$$nWmj`
zG7?$QOi!R}T>SR^C3O&BLX!bIHr%k_*m8L!iZNNsLCT~A^kl~QiD!$JE!+BmVTnB4
zcglt>;G30&gTRMv8A>XERd&DwnPgxuKEp_C5ykWhQr#EQ253eu{c~IY!Q0RfWn*2c
zrQ}KbU`Cl;j|t9AwA81SZ4?@I!L5rPY&+I<tw331iu`F)Y<sZ+E*IV5m~WEv;lgd2
zQn=Et7pGXP6lP}m<NQJ+ZJ#h?P?~;)uk<NV5btujzVuJMi;uq5lBfNKFA_d82_jAZ
z_X1Qzk{!5on8?kGaP>=HOW82MfM?$oypU)kfs{XsJ1I}>isK0j3pBj4!F^t;1pG3{
z*2xKX6TO&Dj$eC`gYR42+h>v^n-~+|JKx>n$xZ5l{gQ=iFB-@e--nWCl2<;Os4>Z6
z^2JknPac0h{r>O&<>`<9=&w)z$^YhmJN@qO{GKL~x2LynUwg4?XXmzUGXeCSMzTqH
zv~}#)k}6LYo4g3_I*?ENs8f8L>n{^O@P17u0sV`!3r!r)#FugvUk=ka#+bzMG#lUY
z!UrnZ3!=ijV5GeIRf!&NsAA=VFZ4D~GF5i+^Xn4u4L!WV2Xx$Ls3%#)H*6@za~3q|
z`uIejGlmSZ#w(3Swi}}RitQ4GAKCP4?@!bzY4~E&mjU-{Pfn@Z@{t&O_7@Y1)i3<C
zRx-5>bQ*V?u?Sdtlowt5OP=~sOO_?CZa2|=cO55`<y#K!0@C(@SaL{XA6xpn4A`@6
zbfKj@$WXPmfj%PGMV@roC7p=4@^Tqa0AF;7n@`@x$-}KD8S=GU+W{>(%Z7$sLepa#
zQpu>GA+d=}v5lv_gHzw6DO#$-7A$NrZ}gV{KfZICwy~7`NOgBQ(7XoBIQC|#4->Fe
z0!?z=8<sIG{vrULz5~Duz(XE%p>-K%K%Xcn=jEoYAOGH$nr6OtJEerD&>^nOm+i#p
z#h-+wt~}b-z6%i14oMQ%@ki;BP}-gvpQ#o|9hrv>%K(x2Z7zJ>zyIhb@89vE?@QMo
z^D^}q1^mRH__V~oBDASb2?KbHPbC}Dp4W$)z}!KOZPQE#-+)c~t^28Poz5n0Ok1bo
zhNcOb)7#<|KpO%QwnKV=ZL|?aUdvu5-*mPeL|pVp4=~d!Y|@Ib3Q$-40j}r@w$Xq!
zJ^id@*cvMu<5>36a1dMMn^pt4A-v_-BB;JR(=A(8kzry1KFdUJ6YqF4Tu0B`a1AqG
ztWmbZN@kz|+vtgs_VGsL0Z-Q8t+!Eb)7nPxC7p4IWWrUsnWmrrO3%kT=`#ky>o%D6
zoLbw8t4n&5zXIpg$9;H9J}aFq^@}o5@HKxkJ<|reNozCMHm&e$Y0h^|pYhrE7*5&3
z(>Azc_{*Ha3_;NFBQxTP5z@ibF&oKJV~r=h@Rl^VY$&$92*4>Z=$#gB(Mls1A*V~t
zY0K1{z}u!3bb1UUY@?5~ksdd?E<%94RUW$e3bY3V$lGL#2Q+O{4yyKuIUNYqn4*>Z
zs{dM~gi4Oz%%24Y{CZgHxOD}juZGA9E1h*%%F|)72avuVoP$N8H`|!il#Z0f$K3Ru
zR5*-ji9hI^_jEcvemkAMeK}p6zMkHlA17S$`a8}&FLi#*EB-iV);Z&4VqYtey&jwA
zF&f)(Ze64sp#2}qF~-=&)F{)T0o><Oju}JQk>mn?T0gOAl}0CS!wa2rV4g<eT+wn0
zUbrft?dEfM0Si7n4aJ*W4h{~c2lpTN2_$U#lL;R<&(HM_ztLz0;j)1pUK{xE!O`@i
zFCI-t2Y074o@%<3uGIXb6L~0bQ|_C$C)4rix%&AXf2W-5k=RL@0^Cg7KI1FjHfI4W
z*FfrqexK7Sf61_a$y<N>>P&gOefws5{`8r@S%!Z{D$^m~0aISc@@+H9$c^H8ql)I-
zDjSPgEY@)?h3iAUzMzorn4e9@Z;$;d#IrM%sh-;Eo5e2`vslKf9(mF${^Os1YER0T
zbMD%A)V!2Ac_c4?Bi^;3c9a~~k#apmKet=1-S$n;z7_<?4o{+yroJ4&72KZ);If0P
z5>ir~vlS2DNb@({X^t_|ba3y^^nkYmaZ|nNwo>>;yp1O5ZL;Q&=3jImc*<+$Ok2wV
zE4tAoFJPPAh&NnM>p<Gk3f6kt^WFxVyqjc!zi9qVfNHd_vGF}@BZ)7ct}%kOf331?
z1ICNkF5fypGdbfF;DxWh{kQ+lPs~$*<$b??ChFg@X$_<`2(f$lB8ZgDtpfoYB2)uR
z-;_uO<y4g7BqHI>7d(vf2LU2Cq_k-QslsjB#Mm{6QgZ$yW3&Z`8wbqPf0QVUDC|I`
z8#>PncV!{}FiQX*e1Vl(RK55jG6N&;LuSEA17A<Dd|=UF%7sx547nhi!7~S94WQ5!
zxFGI};?U=S9qf{4{O|=*yu%;ysCBoK-GiP3S_rgNXtJhQi2fWDA{%}+X>EY$kSw<F
z)XXeo0m%vVi2|7HwoP{_u1prV42tP4v6ud0S2*~SLeD`-Qv^0wgd}8JWeL{@lD1_Y
zag~o!qZ&yd31nuUVIq$pbt}~Mqrd@nmQ8`O&s3X4H>w89B>-Ysgoz*|+JH#}lN=Qx
z2UjsD0}m<mfg3$#uGpBjnY?iDF1yZxrv+KOfWm>f{K9`crw`$)Mdw=bqAdLj8QWE$
zjAhi;*l@>C<d`stU=ivgeiun-^6;crlMJY{jQ;Mj0Lcv+u7Z*$ahE@R)Px8L=+}Ny
zNl8zBDn^edt0@1)B_DF-J<m)4Cog*O^c52tIw0~A^9NR454A|m-8`BdP4B4B_mqJt
zVm(sNlU^6^G@;v{{@}-dIQ^Bs{8y(R|M(At|AFnk(j@TZ>le8TN_p`bL7ep6aBAxh
ze0)h>QcHKKE6OCC*gw(yeX-lTmJm0TGpa7O^k`Rlk7Rak?|KY)`}&PeNY4EF#TCf3
z+g;__+hY>S6R0)jc%d(ClXj)<n8e{@CW_p}m-mos!Y-Qgl&&Y3_EmL_-=b5$K*|7L
zxXNpc9fio7qDg-2XD+{aNrzrE<;VV#1sExcg`adB(xA}?$3k@an^!4gjkDH~eBrJO
zoRN<EnV9ac&YLXA54`jU`T;$*yFQ#pLd$1CEwb#_E-$*J*S^MRX#Q=(rGss7x0`VV
z{Q!~dtIJnu`{X#h(QfJi8nVcRXCJ}KM`FrhBbFgjtU43w5P*?1{Gx>4dR5QOz>5D0
z9*>+H$`on@uwgu$dCR8C+;omBDYrJ(1IFzDMTk`Ln2%c`&2%NO3pgrXNWqZrl;aY#
z9V^2bY)4EGWNr_q7rmr_u|Aq6=yV2&q$`fL>I1Ck_z+9Y9q`(O4vaF`Pug4C$5Q$?
z;H~K`DSq=Di>iq5+g$ki<`+Nv$-nzo|8n~Jm%qr}q~~YT#o39+_srYyE52<)yxxSW
z>npIuz$ZZ0(XR@d`t5Vdv|+nwAIq_IH^W%+O&jnxVUyN&MrayT;%&9tIxXAy6@XS@
zlx=9*7;b{=WVg}TI&9&Mz&N|A$8ECU6-NH1wJUmpQGNuHj5EDojxsH|q8ATtXht0`
z?4=|NKk`1L`jA-moNuSauxjk~_-Bgw!8f68by$It<FPIyopac#v!W@)o#0~_=~v`o
zmHi$^#-VtYcN#qYb-Lr`DWw?NQjX2`8bkCKc@q2<ddVr^)+|F@=~-qYJ%%vX^*i(U
zce?WKIR!BD=JYn&k~WQVtmf6+$M`CLlZJE=1go^7O__E5HqKlq{nGRfmS6OXG^d3r
zPNO>oz2<19xnG37q+1vI=nDk?`fWKbL9>85S?kQY`3tHa`6mqfWeYIP>Et_iWP#1h
zp&8K~LqpQURaVP3eF=f)k#A&*Y}|oF!Au)N@q-n<Ap?1nW^&<c*yNE&)M4h)FNU_;
z7RI^+LQmZL#4H<$Yr=*h!Kj-6Z5us7n^|u*>nOiqfsO-=6`|m>->mWBVtRjZI-Q@q
zp3cwSYW{RQogTlM&QIS?mpa#GF7^H$ul(n=`&`E%1Z;`^N@Q~13<<jQv5soG+!RB1
z`>|}ns}GC0lx7W_+aZ&4G|mgT@hSS8SHkCdz<lE@W@fh{$FanJiEu^0*9vx6P*Pe%
zFZ#_Y+$5SOf&f32q`>R=DDQ)#qv_$p`#LAzaeu<L7lDw&uksJziJ-2K{oUQ^%P04z
zM-TQ~ZF%|$U9L|^7aEIS7njqUw`YE>AlCiLz#YDql^ce4!R4FZCL?*W%9Ct6)6Q<)
z+$uiTQQ)~=Fit(FtiCpIaWS2ooJ_A?y_ufBdMV!R>A`~s)4jcYm&MnTxTeCj%sY2&
zpE_X??0t<DGP;yoWgs&a#{6wE^{0HJEDK}MBjT2@Z0q0`<1KmQ`bPY>RM?jDP12X#
zxcuJzn(x<9MlXsfA7uIg1)j>=<r5XE6Z#G9ojOUG^6)|yzc$aWT~~^UQf;&XArQIf
z@OHyP^@IES%GZmG`cXglU9t_v`B(idJamGOG1#ILIuN{dLT~;kM;$ffxMS>!Eb*u}
z;tdOpxI)!U(gIZ$>A9W=D_gy<2iCkbZ1TqRXushmTfncw|MhJv1oL(iKBgZ)hCY$D
z;@7l<o1kP|anb*D=-R#>7GJdO!umRlNd>Xvg|9#UlfUzm98R<wvAEztfdi~M2yztA
z7f&j4+syC11Y7|Os*1gEU>$G#r?1BkIviv+0_li?$O6UhI~Hu)G`eec<R<t+AVT&F
z+L9+|%d=pDN-DZbf#2;9fAUEnjniR4E1BY14}FoHRyI(I>2%{xlcQrD9I*>tgBpJh
zn%q6Sx)511x!B2pmW2onn()l`Uv%uDr@>SMFoW<ecN=gg35#d?MK^;paz1bicl)+K
z3NACa^69}Ub~%7jC=Ohv%CD+L7mZx_#iwq2+q4}H<<(M3`*llxFB0k2caLiW;VIZh
zQ=$*meNkm-n+UqcIxiiRYT)xp159i4bQ4b)#kOPHQLqy&2jB-jnPl(5t;(itu)&7u
zBm_1S0$UE;*_)IyS(x(VLNKC$lBGk0J=z9*X$Q-CwpVHFD!68maecr=pFGmOG*Bi&
zvQX(*$V6)d=zUiVa~1jwevuCz;1`(hz%q?&(=V~a$5SD@#IU971*HOXQa;tQL+ZJo
z96++A-7bphFP;=Q9VvS)x;C~02H<;a%TM_F+nzSW-skY4>k3^eI?hT4zeCGYmh2x)
zIJqlGxynz!eo>RUJc*;cB#T_${hWQ6i=brx{sSMBb<HSHuermB_eayVpFRG3`oll`
zYt!eSe<>YJ<RtpdH(yV`{N*n+Aw2g-S(!vqZ>J}xqGjU7YXB~J7xg*c1v~d-%kQJ+
z8)Rg4nF-iC$y>*hKJ|Xt$)t|g4}6$*?(BGSM*Y2h`Npp{@PaAzLIKQCeMyf=E_a#j
z>~bd)cXK89qs5B3ql-H8B$v8WeG&MuuJ5)|{_OI+V$tnFQ{b!+_jhTf6NyItSizTy
zz43mUy7jnVfyC(t^rx7xqSD<zyw%@hA9_l(jUGQ8yX}av1)Y;CCab}<(O-djEFFXN
z+OCr(M;0Av$9-b6u#tTfpHhY~MraMzLl)PN!B^5EgZHrjf_*`T_N%-sBN{yCsbJi)
zR*EefSa{awXBwi=I-gFL9s3}fZRzK933gfm&oCiI=_pu291w6DA|P!_{-~fgr_t|5
zj~(=E2VMBZ3llVWm!tbu^oE#xDou{YsRvNSfXfo6VqP0YG;nYEN7U)GUD8#7u?=p%
zV*3n`HexwK>MF8LcMX{&+;V0(1TC8L9`rjR2E{Sb>8v>{3Q3(<gjBssxJa#2K09dB
zRcf9BewA6<BQ^s(iFJLLxX?(EPP>HmXaDTaerpR~mlvl$IXitdy?Xv^`u3MUoqqc7
z{^#lG*FT%ij$dg!KAS#Vp1X|fTT~mzAbagPw8z0`WLs#k(24?tI)DhS%Prw!uuXn5
zJ;KPF`HgQ0_Fd?<u;{O%8OD5Cc9YiOI@l&V2K*zwPIeP+;sGDyZ^Aa&(Z;9n#`LQ9
z4D@XQ$T+cuM|^_~aU7O3u$CEl!qzmC*FD?h3Y~75rU}bd1dpMNI}%xJLtF5*{A@(?
zgSHDz^)|S&=r20N)Rof}Ja-w=4jcK6p%~kPS&{2}5*9ikQ==uVRN*$U;;s4D(*m7N
z{h2bzKwrw=yy%&JMSe{is&}Ssg^?^AIoq!T$*6nP&YEunrll9nWpf>l{iNl_bmmOJ
zNO#!CXW^y4#Wu%7#RK1UfQ|#W&o&CL0C>Ks2t5!;q>a~HZpBL)p_%7j={b<QJ(y)S
z@ex$(qIY>JFxNr0g7vzb;I3EXT4c~wT9gf1f;j^YjQoX8Jks%xGA%Y%pyU-=!MOi-
zh^^2Y+&(K<(ru$5c2Z}A(e}3RF_`6?P2@$X?OK1ddo^q;$EQQ*v}t!4r{Fc8dMSm1
zrU&T1EibP=@N^eX&v1UolT2qiXFZwTUmPo(DDMmPM4dNEFXu$)Cj(5z2Bj0td6){~
zdFJd2KKLh)ylp6fO(Pa8nZ?-|PZ@%DyQ9l_H2WdvT`{3p=jU82;95j4g29WP0@)#_
zpTD=QU)zrk=gp#Wy@6QroU_Bn1U$~EPtQ+n@BY!j^!bwq(}SZ!-%yGTFJ4g==AmVq
zYkKt0_<egi+~1uJ_E<pTTn59)-I0B+*<cxecna*@baHy?S0Y09qSoD=X@7TTIy~4@
zz6aCpE{jl-SH8S@5TIUsElK6zNj|Psyt_EJ@9!?v)@u9aQ)lPW<MoQX0#S}ApItwV
z_UiT9>D8Ol>5Nw_Qhnla?UHLPTyx<j+|<4+*0%6I5PHTg+FF5YoAM3&zEN4h^(z^I
z>xXmuyIGKe??p|_vk~C8K@K6$8+s_O%W-kxr^)hK#9Mw#&i?*^!hRONWRG&-E;u(e
z`%^c{&%V-D3e*9?xP|Fw^1q7bdLS)dVc%24YYrdmt9>L3;BK&}VFQA*FUFi-<=r&L
zg*zS6h75HIpLpQ`AIG=kS22AYe5B9w57M-i@@^RiXpQTareXsvekMe21siEgibkFd
z*ckO(pB4m}qpP+DxQW+L&0F@@@+0p${!KZMKfkN6(Ai!Wl3#-BgP$w90#Q5@II#hu
z&GLdO2!6HJc4+SPQnonP2S(W@j`a?KfAMGk&wp^|Z~pP0`~(#W(mP8!76UI9a<IrC
zYQ4mTqZE|kl4O{@xMQAZDRawWEArsB?UVv^qT)FbQw3q>__Y)!JE8!20|LTD?p)hP
z=hm5&Z8+F#ewZC;BYBsBl;ZeY01^=xMN0nZ@Sy>-&mysN`MXOTfV_jI?3gIGJg9nb
zw%zD@46|*u6MX6Wky&4S&LUR^^yo6!Gf?`e1@r2lP5!1QfASvv%|(35seJWk@yH&y
zZCeT-*~AZRmC!+<1|y_0aNsuw74cOT4{$MKpm<~6-d56LitUSi#dm0yiY)mxPJYC-
zMW0dt_~L_x`=6#(B2%}?LtAXhH&0gZku)Uzu#kb*hO~(sw#yYkbW*>Fr5(DS-A*hj
zNp3!f3h*Fyg4sXU8#w>ULmQ?{%uHMtw2@tP2?VJhQl^mwdM=Ot5@z}6ef1~%DimMI
zmskJ!Nf+0XL<E_TVS-5=meZE5N3!t33No_oGKtP#GJq#?$|U_M^_L6vl_oC)aisiY
z6Ti||{1ZKL`YVyTsQ^<kX!>&ceimb3iRb#t{$Ox<Sg^x)COFR6{>j9!n8Te>yn~!O
z(PT$y+VJijCO5mD*kH?dt=R@ONt-8Me7jUWOe9&5x_|U=`b$6lE7R}&-tVdIrs?ND
z{n_-hpZ#?D`s-g#=Vxbr((9J$naB3e6i!cB;F5jmo}QfeBd)Jsy_sIUd^5e(<d7$}
zpdB9{PjAe7J-vSUdOA5ik<Nwlclq^Wf=2!Lb*W6aa91ww+`Z)wwcgpe>krQQH43-0
zAK%-fjL7G1FJ5`blW+TbSv=#_j@aT!IogZ5>vGe_Gs#s()VCti{h9(!5a_M2afYr!
z|D;a@WQPiOy~?)Rw)P7{^#l4-dPl^oyh1#-^a+%iGL<oeiDTs4ANkF;h+L4+M9IXO
zv^^zf8?&DKmUOg1Ub#6h`yqSj;VGy4J9$_R$>h&uIDim#oLl_Cj%^{0?}@Pu-CDJ8
zl8*iO0(+%zo(03hZ@j3q+pq(EJC0965Vkxxe01VtaHOr{Aew%uE0cQ@3Zf<<Td{4X
zUj0{AAm*EpHjk{*<}PJ-{q^`3YIOQXQa?0ipW7b&ECPz)2|=31RqU<G3b+l7+haw^
z58Er7rb5rA=SJ?+&Q5rCfB}k6#3q6>E%c~$?aq0Lh!|-6r{BRzpS3LZj9W;!li9c&
zlE9t~Id{v&9GR9bc@fXH(dWO!(ai>Z)2!dN%`^s#^}#k=6qPD{IKRy&z63u>#29`u
zou9A>_H_E@mm0f&`fsNf&%V`(`x}k9#~O3becYgHaFXwDtt!zc&sCs;rG$w)T!rn_
z-wxcAH@vjg0B&en_zEXb{DvFt4|1E}K40Pcz^J#)!<%f5y0cvMS70lDfH8fY3~&=~
z6GmA)=`sAevT2iegrarW25a70exz^m0*XCeEdDLLHH^BmocEt4EIc$CHh48wn>x!!
zW+kU&=nE{ogs#hxh8fds&PE&-TIi#W^~zXm4D^CdFJVnnDuRmF^uT6%o7`slstmEV
zU55kR3}ZU>nO=MoBNy4Ce{WiLDdOv7lqvz{B3<WAOBLffvf!g`j^yf#0}TkCcu8N6
z6MW<w(k0-KG89kFEbQQx&)<0aTS7A)`A%Qwx9~B)4PGS&UT8~NxP8oAr{mybXnOYL
zHIL9VG0RTRr8*Qxcx0X4;79jrSjxZXY|2`DNYAtlx@n3WsQenoXIcd|;RsfYz-V(@
zm|+XAgBx#vH7{|d5UKZ%!{BqlYHmM-D)lFY&9>|Sdrc!P6|6_KWGYY&&FOs2fe&o>
zNv{h{^LRtZ#q{pox#BF4O`LmkUN|3TE5_z*w9^5s@>_wE$i_#fsywz)Fqgryveo$_
zt8;-Ynh@s(Vd@nTu7mK}!9KTzhh}LDG9xliLF`af2b{m5H|NW?a9^{~AO6KTuk<_B
zRP=IsaQ|?6^7zqoba>!7C#QhE`CH?g+ZrFsUq5A}m@hT(l+cIw=hKdOetN5}`DF2n
z@}!(x?>Ro@Is`ZD@=Yt9^WWQ<4)*qx_tA9!!BMVPNdEHja^ibij#&gc=f>A_U(;a0
z%QwREgcxO{%<?nWl-|qd_tWX=sTb~e8_v$|J->4CC5vCL-cE0iPp324^jm;b7hH?j
z-My3R54CXQHtI`i@-<K9QFfj{yX&XccpW3<<@&~+^6*W|JGr6Pb;ClG_*~b?8*X^&
zE_LXZNHf?zF?}G{K3I7C;2WC>XKEv!&P$sr;G2LOk#jvO-?3-$i+UB&hI8Gv`dH_w
zI30-mMOP3}_#LWZ+Piyeda$?Sr@rpiLj|^30A)-E$5uk)3lEB_nS7I+>A<4HaUig%
z->_Y8leUgG`AbSbHJQJJQ9k3NM;6iXZ<HlkWRn+2g+`$#X&CJrluRwdi%#$jxdCqC
zjeuSD+nd10DMZGwL|b9AOq(#8+6;yPu`N|}EEdC2`t=|@fM_mP2l`jWMxf*Ea6NbF
zfr7=ff$WwX7RbUEuR2893KG&k7dxt=T*##8OfyvBA;Iy2SIOoAa?-O(*#Kc{Onl#Q
zpj2;jfks>|L~AFGz2r4th;i?>!7`MXYA0zB!qY`#gU5DuW?Vy!ge7DbIhz70J}5Bb
z+Ta2>8jfu@%zlCe8!lZCI4*-nDra70wmgGa^dY7k4BQGH;OHdsn>2J^j8Dy(><j!1
zf*CB?aU3f^PZ)5~d!X-h2GpcIkk5G&%Nqs}GNkM-LgaIS6>qS^!IOd@fUI~@gN~4G
zNuGlM<1=NHV&o+&mdh^7^Uj@D9Kx`I{@S2qC(un2eU|0Y-sB~k0#zO@EYk+$0mL2<
zUa&UVCjg{5c)_PW<)zDz_EV0vu;hc2FWScjl5^jq+{Pt?>yphuAFGn6`bPr4@!7Ig
zEsN;mlT@fbbdj~55X~K2itZ;&5I@|aM#z#(srO_zPY#?PcB@RJ!DTZgRv8tsND)9r
z`H1Gch35${@y(JvX<Vo>xi8udz9l~_V*n2Oizfz9^S7*JR1q?f%!v5mr*#$Kt76NN
zhx1B(@bi;M?t9TS-+}r_KJgzLs!!Rb-@}8F{U>)AvCmO|=wtzpT#2LOyO>lD<YR3c
ztqCyY<2~FLnzSkJ&!2ob{iWak@pQ0%G(CIzZ2C9<>fcP?eEZFGeEe3LyaItIz}|b|
z>WwCSOd>DOneg&t7ayg)n9fgme|3GVRrvXt!s+>RcEZPm6$_6oP3G>hIHd`e%7g8D
z_xAD_zG!<2_jdREMAyOLp+72ncyJ_qU;gb)hewCgeNC<pp=)xzf3PoI>Fr6Mr_vZd
zs2}P?L0PyCR4oR-27%h8O;sjNVwmJoJyp@LU61YyBm~|^6#J=S<Tw^6-F?V3$lwZo
zl9V<vIW#|E3ju(?($&^!IxBxRv?apCqL|OPTJrM9dAVQH&k&6b`bzit^h4++OA1@A
z9BAcJJel~mV{Aso{Hh=OfRy64<NjM|ie(>cDflvW$4qc+N6Il(rI<Wy37o%eD`qEl
z8l%LLY8m!R=*WWjPo5@{j16?4k)P$N9$?a*#W#<-<A)e+Zl6<z`8YlX+ew_Kh(mYQ
zM93%~mvOcMoquGlivo1hI<`b&h)qwG2;!HO&5lfB<f3QY8A?We&?e!e4!kc|6>8}R
zi@Y046w5Ww+C|nz2sOw!M(p|!*1E{2jwFvbKgYIGGR7cQyO0248%UlCGV7a0O2~1`
z7w&$dq~~ptkBzbmzYy1mQ2Qks#a!{xcirnZ)W>i0i7!DdeBp<Cjew_Vb%J~Pc6#yr
zyXotn{rl+`KmXU$tLI-&$FE=d+mM{_`<N`CTplMVqZ-}i@sFB8raDyFr}$2#dIB@@
z+l9?+-6j<Px7ij5|2l5?lzz0=VT5h;n`AnC3RvzM*-y38a$~52oMA=325yoWd1KfH
zBYl>;LG~s)+XCfUr3d&l`Z|ofF{Jjcms`5kRUYg>aMQ}pR$DV&y>SSE=4rdqF?CG~
zc}1If1B`U~ziAtELc=DFXp;tP(uhXdLccz}NfUNG9j}G^7({~%ax_oNE;`?n#&1rq
z<vfNK9{$23_sk<oZ$rNsvyLqkU(FsF{^87W(1-Mh6MP;qq-|qPTYN~D47q^or$h7=
znVD}K-q=5(FMU-h1RT737*D*U8{bT~t&}6ACGX@PI^gsinil;{-=ZtvAdm3uK%nin
zT<~=~=WN<G`Rmi<>tFd3Ag!(2yC{R+tnXv>hD|~gk}jBe3QJ8126~4@FL7*lyo4}2
zg4T6>HMCTRtNiVuTKl6RKcu&X?Xc;?%^^7%IG=uZIeoZfkw@V}&S}~2-}?sZq2aE8
zY}r$VHcI3?(g)gWr9x+ZTyr&Iu7z3F%952ZazNyDKFu3X6i1Kq+}I3MriS2i-V*FZ
z9~N_bPAA}6fuZuToPM(vr{8&z=*`==z81hD&7+6+eeTY63SR3+seB`4`zy2e<SXay
zy->4z_tvy~=av_Pe0@OYy7Qd=HrG6Oo68;9o9bz%_=`8t$(N_jo;*65_IB?&U!Lwd
zJ2{=s<yRKD_`Vi54X2zg?}t1Q<#R**X-&5IEP~oxd9#qWx3_Eg*Kd!f7q3qhj;FUL
zXTBcAILV^dQC*MVx`xkl#o#rFl!fnW-M-DDR?5eM7T0q4cG%(J{<ME^&(|{!4)<IZ
zU(-;Y_Y`xZZ@#Ub#X?^vC8(XroA#y*T;HG_vtY$_xO^jx_BlR2@e^|d76ko<9-ex0
zJ#d`J!kHVFzDJ#KEdX5f<QKd23d%S4HOuGP#Qx6h>EV%l-X)#><Nj$KY=ISR2xOf8
z7(QTL#ZHsrzaeAJ71~-3nCa!iHoXoFBW{4-C}nr*2Kj<*^v!%WMM#>Sq~VI4A!+%b
zSLFt{iMI(Klf90AlNLIk?!gwAmD5%vSN4?1NL7g|P#X9?qafpW`TZdl0`{wd^MU5=
z8ocoJw|Ui}xEXe%pk2Xn^mFhC9ue(+^ZW(yiG0w~MRuuz)(brcB0Fd!VKvVqGzpIf
za5lH%2UU$JZJlsqH~xetVu2;Nf>6rVGw!Se!b`sw-?Fa;_h3&NpUZIDsAE?2iP*MG
zQXq!*z?`Ci_@bg{eJ2R%Tu3GFT!eSakcSl&_<JA>jb9I%q<KHCFSIy-b5WE5obUTA
z19@~+2;m$+80b51+Y>LZ_L9DIGZ+)&fVj|$hm4CvGYR!w7a`>!&KijsrlhXNc5Go8
z04HS<?&+7>;3v9y=!ofhQ`&wPle8i>v5!~|vJ!+NQrk*i1|ige2U*%e2N)VZc9TzS
zOmqH%#Vd<98Ws2JnPf?{(2Z<y#PIvDH_E{V58eGBIHvC5X=6X7gU^=LBnn}02~j<k
zv8!n_*Ok)VUqGA>Y$5*Z=lW88Xae9xB-y}f>K&OZV$eri4VVJK7jb!PoV1@tK~dnv
zv2F+b6EAjqQrP1}_r>n#_$AD^HrB%oIYOGD66Pa-54LMEb7<?i4Qj$+VRC?>1pCK5
zO_=z`6|W4S!pWCO9Zz_%m{UzF@%Wj$dsql#v5SeHYDs~K-O<5=>8r1PIDP*3i6)8f
zrf<IcZu;)qZ#7xu=`1dyo=vAb_J5iuyWVPY%R8m#C!}14WCys2N_V-*pAV=qX}o`U
z<jF8kclk91JNGn6+?x*e_gy9yzF6Gi8)e8G+}roZWtkZ6tGq0FJ$!Jem_@OJ=}2kr
z?&6h)Orkv@RQ-4i(a-Io{BL<OD7RGC{HRBFI|Fh8>11q49|rGs>2~!PK-zVVrc+Ks
zfX2or4UR)Mgz3NO(z&bsIOBt4X>Y~OOWyTj5qh1!D5yD2AF&_Em5sUI5SN^J^%Eg^
zpodK9$319@sRxgn1nQwqR#bt`%W|<rh@JvJNbdF6QM$_H{Zw|{FDgIlB8gIUnAw(h
z*TN$;=WCy>Ry<<-GOjJU{8}F#dZM9dJK&;_<|iz&1j%CoT13!P5IL;H4sq8Lx=7}i
z@j%qj5ujd?*Uz@4=QMtj6*6G{BhQ~i*;R~H_f_<a1b6)*J8VrwE+%G}%ZWtMn-)^x
zVerRLM4brVKjgCszsgc6ktrkISvvCdOC@v~L@@$Nd5zOHt>E&fgpuhy2}FS*U9`wU
zj*k7<ux(o<o4<Z0E%ccODI1atu9)!8{(=_1+>aD8-q9b0>Exb|w4a^4o?blrX8QVP
zKb?O0i=TQy?DebX8p~fz?=H^deZRg(G402g<kDpu*oGBcA^djW#%!xTu7hr`>^EzG
zdYxwLZQ34bo4ieE*)ePbs55_y%*XV%!7QK2#0;aYS!M%n;*Idha)d$F@`KKbKBTvS
z=|h=Xc2j;EJ!;U{Z#MbFOBnft&EB)(TW5rU*CrWPXIetTr_qa8u;zs+><X=6q>ntX
z9#gLHMw&vZZ!K%yrfeu1wAA35woPwMTj=ZbipK#GT)}NzypXQy+(aAcK2ERfEqou-
zz&l-d^EnnStU*tp)IuwI(nbYqzJJ9GX-o%q81&}(@Ej~Yq+I~VgnS2SZEr14`qOy)
zoc1v$DZscZ#@tyxx+WV;<6nHI+c%Es*eJOr^gg{VSMx)Q&Zh0~ErQgVM_$Z_2bwmK
zFPkIGG6kbv(^qNtFKAon8?rO+3O(r&*8Fv7SksG$Gedwr>k&gQ{7P8y+QdRjk!Hpm
z)_x>xj!7K?RXM+Lh<!X|u5^n<62ECBizwQ-FQ`{pk_-KQVoc|nmbaetXFp@Hz3nmq
zj^}=rej1?cTno$jVy;KT2A3^JW3(40IJeFM2KtPD$m2)sCGYTz4f%ul8iJ7c5&h=3
zHne>6oNGATgv;XB`Sjw|Yt1d+NcMvla}ITG?{mmGzg(-}^?SV9@bG|F@9jB1z9Y`t
zS~!;{U>62;L9XP_8(?nT;aljp{5`Of)3fRA@tf)R<iu~IxzwDNH@|Sbz@IXZBHssN
z@h9USF3Wdr46bV!0449&1Ohyz28-{FT`C{0CHk6!{N<9z+mmxIdc8f#D-rp81NF{p
z9`7r#Xhm5#?{gVc2YzLt7rt2N<2sa`am<1xi(xEs@dg{71|#$Wmlq9F^^RQ^^0oWV
zUE##MF^DJdc*=`92K?Gb(rT~fIS9|U#(0X6dXQoN;}f;wYelLHvewz0`yae+8)W1}
z%7}!>hr73@2L~#P-?$?Yq%2#p#fNk-;Zk<owD3tehF+|eNu02z1xDP*?t<3aCQIC*
zq0_GKq5+ZV5E%7F*rabuk0c)}{aeEo`=d?;;5t8mn|LFPb+!$9f7u2%X}YY1|4J|`
zQ)jWYvZq9PZ6&TiYYe<ktdpByw2(R`v>$7JLl(aN`0x5vhg@`Yqn9&8T~#5Opg%if
zB^`o5vbaL+PMtK&cF|`?ha+0+GjE1lN%PTwZ>;Y4$9iN!_ri$<&Gen!PPCr!b{GLR
zuo-M(%PfL?nEh0=wjG}>XTEgtxdki(C@<mU*matQ2nLX%7W|jKZQ`R4Iu93&>EQE0
zT7d!4iv>zTw{59-A&ZMAmJ?4SVPCAw1zCnqXyE8jX1oJp-#ZgJjG_h5ql2z^KBnX#
z4_zA8<L9)3KSHUPgETtvBdwi-9xv(RZVhWu!Ls5?=Wm<#0C!x#gwzFjV2imAvk$TG
zkAoC)s42Fbf-R%&A2e_v3wQQe0=Uq1=elr^3<tjVw{$>$=K~`0Jhfv(vYyFFm$Azd
z5Gi8LQG5!8U&%xXCUi>aK;<&IZK*SKkP}j(@iFCv=3FQkN!wE0O4k@=5)E7NCVkdE
zlGKL!;Q*fg#y>WYbw0*Kfu{|Dc4y&-$%7B>*l}J7t&87Gt|$XOpv%O=eWDg@Xn(4T
zm>CTodhvkzCO;eyfN>&-t}`^GRF2?*D}T2$M$r<{7oZWN9u(*=?&G4ffCQIcS<A#o
zeD_hcH>S|H|MJVPHDsaa9(NAivz{k0mzVB)ckk>-nD;k3bMaH}2y&OyyGs_m==VF*
zmtX#9`r!|M<i)19Z+MLVwI)U9p49TH!sjnuD7>6rym&PopPpzEcscvI4Oq*keJR7e
z-MiDl!Jc0~@c7}Q>A}Oteo|{!K2wi->c<EB2Y$Ri<?4lQp4`d=5x>~Sd3QI9T)RvV
zYf=b3Pros#ti>V+#veIDp!BYHwKMHoZNW}Ki}je0zLEaw0NoAQd`eyPnYN_qW9%tF
z`XRbb3L?{E2DtOWZ(@b`Oo*=c9c0UHXl)yW9E)yV?DtsWK2!SHCrAfxAIv3|AH86P
zU1adfHx8J9vGRm6Ss$NC3s>wh51G=(o^MzHN5=frQuJeE^hHS8!}d1HILIYkzUt=-
zG|<L?ynY#r;g|0O!2G5~mY?P5;WN!>#o|C8Vt~l9rWL_93Jo`MgkrdR>6g~A#!^@2
zfO{8CY$<q46j&u@K*!~m0V4Z{ED8A5g9nZAk+vzi|0SW#{Ip`z6}zTLD`DRi7-{^s
zP1@1KKo?ZtJ_66;*dtMTfOJr^PXh8PVS~k-3@!PIrZEqE<TnejTnw{GWWkmY6fbl|
zF1tXs+h=4<^<O{P{EJ%nYA7!a;O<Mp{q!$Sf}OmcUOoM0`sU~VcKYV)fA1&3UcUHt
zI)3wNdVk5c4j3n;$CJt0f6I_<(87!LRN;#75lEX{A7b!&xFY`v@G-gY)9+^-I?P`K
zKGnZZ1^a}}YwW_?3^&0_jswLOSjlX_HeS!KMi}MBa2>$e#vf_d*%|3GzkIqbwB9z@
zl<hFWHvU|j61~Fjp2b+pPqa-K<wqFl3-4<AM%YFlX&sm!Ug5Q`1uLEc8rIsNH;lBV
z0L(x$zmI9pPlhyMvpkzRO#?P**QEo+V;nbKq^<d1rCp&{&1~|!x@cLzbn<YDLkO~9
z#Edr?Ut^<>Ssg1_u4q^(x^-qA>Bh~P>ESKu5qJ6{@|ib9Y#V=jdONQxBIi!f9U-l)
zdr{_+99#!p>$i<{8athqUC(h+ucnElZCMwHHsb=utIpei%~=NfNjnU3*0XNOIJCWn
z(i?)gh;Gt=ftKV_E|)FTfnK5VzRH6(`m=!f)HOo#xi4zBr4b}K*j!7j#(PCZw*|pc
z&TT=PEn%szL_o)Dzz<Z`zWydd=23X0Ae=cbv2*7d#k)%l>F3k!_h-}e{#4<7y7eB~
zJL$dC+*LlJY7;pZ#g@DKn8$2Wf`oy_`5515@^v8FP<Wqf7n|U7Vx0$a{>=GQ%+b#L
z0&VvSSgzs<xdxGIXuM5@YXf;oi0fyO=X?zrt{uF2^JaSf^5t}PexY;reVxPaVNNzN
z%w?AP?zb0|?)s+K-5vEw{c>I4&UCQHxqV*SNJX_zA{giL4GTIfXmQ=*_3PKuv*#}~
zpMB|1GRTsBqOV`7p7^vw>bWeq{0tK0ntEnm!ynhN`Ly*d>vQdZ>xoB4dCLs*V-}B2
zPS2;~lLQvOxR!CqQ(Gz%uPN-$F8KUjnfGykn?lKfR}j8^du;tPUd8CA*gm*!C>slV
z$Kvz09TpbPls>;WbKZP5f}4}2;I^^H+FV{<S`Hi3wHYci^+mhEXYrA1UR*!pI+NQ`
z`QjhPFshn<+KXb?NguK?w-fb6Tclk8p4huR-RJvWdv~WDzu~CHngZkei?;C621`81
z6z+6L8!`!NT3{RQ<4nO;{t&N%scVRAhd^j+_!OG+7)VYNCpp6vTcb_|;JQ4w(Hl1T
zH^~Aw+Nu};?=j9H<X;_RppW&HT-#J4BU2@=Kx-_#kCpXPV6+f>g!W_2Z}7s`-};k3
z`ALQZK|7+DN`vf(BadTv7H}}?2~cHa0<n+^cE*7}0~^UD9r;9D><Pon;B!<o@faLq
zh)6kG9&4e2yG;I9Tqi&MC}`x(GTj3VG>bzmCaF~{P~NTz%bGL%k_Ol&w~-PoP<A*V
z@Rz|}1fh;H7%o;k-YGP%>j1zKx#WGx#5xf9;x2L0IE-$1#jk^dLk;FMw1-@H+(qEQ
z)Y&9o@_>en=&=`9e8F2k1bguI6EG#`9Bj*mtPf4Oh)<C!FZ@xUG=OPXwyE;f_MDp!
z%y1C2@8F{0KM)_#L5$p56I%3UJ~Y*>{kJ6%(I*J9l|Cx)qN)v6!Qq*+*dr=U(#Eu*
zJGzNYiJbei1&HEh^Qlk3?dZ1QVt3i|w)n)e4{bx>ZXl}u<I_Va2ivxri~25$@+1#W
z+A+(|eWm<q8`UPW56%W1(yB`i_^MMDsi23SeBD0i@N<823Ff##(x#Dzb>T{v{f4aF
z&qZUZ8*cGPS&u08ZFwg1!CCpHMD_>CxI75Zk9Ze06Gj$_?oRuf6yH;t_ebBltqCLT
zQ2U$AXu^YA%7MkOiA6C@$Rz)}Km5`3#TQ>$hDEN|nzSNJf8kY!uU@|PomB7m;4h1e
zAA!`7a=3eIx_@xsg|36W!|CXNICnC!Aj0BS?rP#AtL$s|C4DKoh)R1uce`i8?Y^bJ
zol`7~b@@_1oDfjo(w96vpK83typR4XS|AfaNaW-8(9it@Ju&9vi|xjrj46^)g8sv?
zAwILvnLLq?|Ai|Lg8Ppry1gL9f|ts|B)k0Y^3pTNLmT1APXgwH+ZF-3Xy8(mh{|tz
z{BaK}9RqUCBXx#NvNGLz@mYz)(D1>u*m7PiN80w<7Nx_F*mIfq!Bd=n<>UO+u3qSa
zO&b>vn-CSF@3F{2HOU6I9|}lR-uOU$VFR7p>gULJSuBGNu`QRJVuSf+^=1v$a}Mrj
z7WKwVHbo$%{$YFgDCoL2Ex}!^;-u5py8TC9r}1{|Ut$!Fh~m;E%xQqg#==e<#LbhM
zTF1q=pZ*C7pXH=nB`k4qTX4n$ZD;P^bquh*mJ{9kqJAl-^sHDi#Irn7sUz$F>GR|j
z8wT^W5yP>9O-gB|6BUoY_gkf~Q@wEJIb>sqo-E1EKmBL_>@Vn5hrR6pxCF-VF{P(=
zIg!1XE>7P}ubzK9ee?Ctr|-V|+TR3w@#3k!3&zu8A2i-0pPN!*fgjeWFQdOnh^xPG
zu<r`jLGwHhV9Y7`Rj|Yww#q3q+{D|2ZL)+-fN~vg6Gl5X>8xamwh5o2-}t9M>s%+l
zjSrsjae-}eBh36tZ^EcM%Wi~C_?XO0Bd+ueCd03^U_K^pgdS7Yv;`LWI**0E*4d=r
z#B<(T=>w1+#)cXpua)1VjdF~)o3zdHH;lBVUzZ;71kzZ(TDOF&@@tysB}01VQ*s|j
z7y6h-&oMep09?!94$Mt5KXG*wz2D5?`4pjaW;;z2Bw2K)2f2@>LmJYGmGw%7oMwEJ
zx1`+{W}eDfc3P%mLdr_oWP>=K*ZH?0pY+wfF{TqtUefxF=@DPkw$t0u>-gK~9l~Fy
zV{08uUD+{dNH1aWXM>*Fze4Zeea*oJw#saVhBbX0x5}MuFw2Z#v2`7c@=Ar~?Nzwm
z%5M&QUWhLp1oBQl_F|n=YEH#Bs{CCrhYJUu4*PJan1vVNw+TF5FFj66?y4R$t}NwE
znMFaMfW?@6<7vm&NxWF0O3yWrYD@k2`#K48bDvYi1~yx+Vb-hsfMuKs`1+aTI6vcN
zTNu6}w?bP)hCHEMyn8pDoL%_sFx;HY6HMH2oY(NhcRq{Ypt|AO8BaR-Nj04l?%m6c
zuD07jlTK<HoW&RxnYfO^0?@f^coB;4jVm0#Jqh2kIimm=qA3UU!Yc<;*No2asjn<j
z`89=HLxCe1&S|MC>~PJ6r_^|ojcXFnFW>VN8{a%<K}qdPdHC#t7rb<CdpCLZ`7<8L
zdHbZQur7;SZ{EChoxMFio?gFxlkoDD7xP|=$BoF8@${6Z+fM!ZLKd*x!qs41zXTS!
zC>uZM^ILj8_+pU#p?xWt%gVO<qyB9RuKSr{AZ-zT2LKB@EPU<VnGWx%zo;I@z63Z;
zLnJq+$9TbobzHd9{EH3*hes`hKI;u~ee4FV>VKSOP<n$>G93a}^nfk2b)HxCNJH7w
zykdVyj`G|5*hXs@dB$&$1=5b!!^j1{DevzMGPUxz!VF3*J$A(@XpM#UoA7V65PO97
zW6f_^J@J)FAI0bd+TjMWDZ-s>h9`=|rYfnl(c%kM?IpbL>_^vTtg@tY@dYBx=#)@F
zBN|MEGL(c;K05_ZIt&Kp0pX<#h3jWoa~oUIgh(egM|RsrkrzFa*o4>K1Or;KB&3i7
zLoOVegAqj@;3$J6J<gaGcC0~&HR{xKZTQ)yp8zm12Lho!(E31K6DbYjEOd~D$3aBI
zK{j`>@Jmf|!J-E?QC%0r_*Oo#E#J}0#b05W6O=7<;#WUKlE;5VLr^?loU{VQQ(|-g
zF5ZK2Fm^Qzb>%wIKe~}~Ul0Wx;PB$MQT|{7d#fl17pn+-uyWY|WMUuwtVupYM+a{?
zLaUc+!R>g;oN^J$8w5+qb`(gZO!`aSc~m>17l7%1#X6XhO_Gv@(V&F#XieIbg@GK@
zP)g3o^`aMeyV_4#@lTs6*(8&dyY;I66cAB&u;1=qBH&;1Y<ml&mVm)0A=Xh~%)o`1
zk}Y_Y1yND+i+1-N%H)151~@b(67<Oi<na|IIFnpHj(V_vsL3Ofu}mO$c3EWNsknQR
zLSJQnPoK^tj)~}lhmWQo{NP8^(czIMTPM@E-#zscUwj0WNfr|??l{VV7v64%lxYE;
z`r;`ZCW%b!!TnBaO=kPmrS@?r`?ALZyzFAqZ=qLxv9+A^v`t%e9k@@l2goTu*Qe@_
zf@-I5>E#17kL6XjgzTrONBf~X$sQY=Y|tFZFHfYwblc+_^5t_Dd=PHGh08Vz=b3EZ
zVWdjF_^A5#VvlqvlLSn6XG9)<&C?Itq!ssA<UHsbiW!%BK}5<5^ev_B5Ax1b{3@gA
zqS-p~WzRIFD_`hTm+hiQUQp4OT>IjmaMG~Shnt5Lfee|BjOD=*iSANH5&hH-Dzq4Y
z2I47a_{a(^P6vL(?YGN^py`oM-bpVU0SJnPE)cu??2nkZ+fp%w-Qe-Lpz?9PB_F+{
zVl$YbMTWHb6~}|{atwiz`e8#nhKAIKEXZ;DN53ahu!|NRH5w|<MaYy<e6}Hit(IS{
zlC0WuCyXV8N`RMb)0SukrBfg2<7IjSAe0Rkoc~gLMM%iTzS{1<LmqSuP!^womimiW
z_*(o+3FBKz8CWD5=>P16)F6vs=O=Hbm(RbRzWeqU)3fh>IlcJq7t?dar#dG1?p}@m
z+;o%iPQl}_epgNZ>M*+e3E(m)WE@zBZ7|X|_1v1pXM4p0a)o{jM!OvvZsKi&n`Ayk
ze-m2gTD|61auaUSZF<XW%8vL5*U2}`yp8l`cbVlBuE4-sg_)i7Hn@q`&@!VOm}wou
zoB7&PP0Mi%aGK*TX%#N~N>^z|+DHek@&emv4I^*l&oq9fU(G)^g$BUJeX3!UZy0G!
z-=t+qHp+B;t;447m>%(kw!Iz+KAMa)muj1h?RjY9G_h$_cZ~x+6j5)Y&G{A2<7(rr
z=VZ<aNUwQGBLv!Wx{<dH-<IA+3&H<u`P*=ZAx(`3x0qp;n*)=nO*_Gv`<VwCVx@-w
z6yEwuC}#hnUSng~->&v2V3U?0=@cSqe%?=Q#&A_9_Q*5&uIXdCX`|e1STM?k7uclp
zThVJQAq0q<CT!r3q34Ao%)A+5D*=GN$c^D+uuX@s0&F{vZ2=qC+qo`$Ur0}001MC(
zilB|{bA7>2d|hgO#ZzEs!WCbh>fBg(Er{KE&l#b6qq@0fWIo5TA%m|Q7}5Eu=3M(a
zzu(=>8(!GZRJoo(U{MKj{6w!8d6w5#1I(N5SPJ>gsvv0}HW=ie&qRn`%Dx8Ug{6%D
zc*}JxO2su4zN>Y@(^p>1;%PPZL-crxinq^PPG@K5UR+^Ojc<M1yaMOTGA=(pOz%Xv
z;O$1T>vLqqr1&<s7jxPVrTty5ONDowrE?vH#jh;*;5*m$d<}v{DHgbVE}44b8o*th
zm-jcg_!c?x_{6y_eyb(CpmnrQTX3VXY@o0Dp&iJdC#?W4qRFnW2?DAIjQRSd@}exX
z1y7RUUp`my&NmuUF9e=m<Eb#N$%Id+?{X>l8m0Wh$6i<zDyT&((#})mRKC<z>Z|If
zV&te)#>(+n(!BYJFM@4Trj*ONhFgBi&%=Ye)BdjNlYA9oMSwlq?eShb=dl?Fz)dZh
zWBwa55IbII+vEZZA3p-y`3nsoy`~7LGs|7oU%?hzBapl<zcux$+$cbuTW?dQ;5s^R
z6K@-QOn&5oe_#7If=rF}Y(Ho~3Hs@rxB{)Q@O~5ijTU0BDZ#hp;DxV0{^LLSi8M6c
zwbSgls89?Poe(7E8jA<YrYU$>oC{Gq85ezQOFOnGTmW^qw_M3P$^c#}ru(TZGC=@9
z*7Aw{#0V%k5No5z>HOolpB{21F=Jh&lMOgB&V8|6x&|^a9zV)zaa%@o=U@3KqMW`k
zY5^{IgrD7BR5QZ&VC+ee<f;ukP-@_FeMpPC_Qkd;H;TRpiVbkZUc92b*g}Um8Air=
zK;zB;4$K1CW1!5!6gu(8ui_HT2bj`p$be3cfw1B`xbwh|!utRRaFC{+a{!_Y;5x9?
zQz#t3qF+HeJ}7|cpFsxO)OGuyt*fH-l7vjXBNck;i+`~Nv<;9pl~?wwn!&aRY?w^s
zVmIl?gVSw*S<Xd-Up>i<7(LT)-8cw&2_x+gS<|(#O+%EtU$`8RrF3&Yvm@wKURE|j
zCXo^NLbia&Z3%9ILis`tIf!$Rmc3+Yh;0#sq|Z=i?GvR^UX=&N+<)}zr>SU9<ZVZ~
z+D0E=tgp!elLPvuWT}z#E$J$O{p`VLrVQ>2*b}+;o!KYS*og|Ah`((xnV{d{XIHHk
zU-a|SY2<*v3aq~HssZx!s}GrEX`;rf0r&S0rh@~%jdfUeiS3zprzUi?)7`uGrpJ#z
zn+^|;JlT5w{N?n`*WXSro;~;7MK7N}n@-=ptqD>)yB+YMfl`uo-AR{^(4K3eMj6=e
zxJ$~DPfc`*dx5F<`Rt1dwWz~B31E-9rJiD68@Z)g3nHNEvw$UAR4~3%SJI%LMT&M;
zK6ET>D)Orm6_^m)9~Rf4;fsR(7Tx0vYS79*%GC={`a8DIGVFY;M_P1bp^5Lsb?w70
zmxB8f`PhGGZ8P@5kFAVTsW&lPAFhM4t-t-U9mzRW{<!~_obA9h%}UmFzGj4<m_A!|
zj~|v-pi-qDM(p%cvSx$J1@c6+^8su%Or>Ny4HUlcmUj^>3j@DPMH~BGym=fpzl4yB
zC8LCS|HV*IT_?garU`Sv7P5#?u4d7fk{{{DN%+`1=)}ByMyhbjhM~|XR*1DNmL!dk
zoAoMX1gy-29&jErWN*n2sBH~Ct=Hbzpox@KItp`rSZ9=O+hM12L@^uP-cuvZ#&SY(
zj1ZSJn<!Pla`=Iwn?;<u43wOcEk}&)irsgBr01B(@=c1*s8#idUUJ>=N;p5^@z2aT
z{)>3kAvpuoSz~Kx`T}?#h-^}7uXh*I+3_oX6YS}C-%QWHdpdom`001QoX$_*=wzE$
z;9ku8Ujp@3i4mx#LW3cb)K=JZ`y+6Zsq0|&bpx8-ExFZp$v6}VXxgs|pQ1bGyG3^7
zbr`FXF@9Usw(M3Yo<osFn0X2xgHi5N^gcxf8083K88>+yt^o8xfo(d3n|Oo~ZsLPo
zrB~X=v7$7;3f1YEUW^<%i@s<(tmOkEZIcIF;a!tIWb^qe&OGA!e74cH(Hlk@WATzU
z*Z3&cFs2*dPFwTWyaXS;Hfa1>wvaijXpR=Wq!oJJB=C4G+;zzW$Mo<bJnD8Fe57~i
zHrd9r{i?TZ{HtkV&c#E&!e6z?nvWksmLR~+=-+0T=VKee(6|i;e2ye&nNcUStM(Qk
z-E-1~Cu~lS^p-a^^PF@%%JVDEny#>oew7w`@qcYQFv@Q5R<w{jZ$W3!+XT;nhq|>K
zV^PWbCy~~j>8oIt0Aw9D<q{TJV38;IH^_`qS7G6QA6VJ{gpk}B6#`fIjl=8q%Mwrm
z*UzSRI{oC^V8nS9q9#<*c&~Fg^mDv6B5B)I%sCwA%LfMsI=|obd9}aI^o|7`eBvA!
zS#A3`BHj3=^F$VTgxmIPS3qId_Y`&TfUD5wz761<7jllwc)~u*H5B|nlXGsaLvY?q
zzspl;Q!PYY`bjkw?)tTVWX3`bJigUMwp^>g20A<qMxNXpOMbnW#gkZk4=dL@IA4};
z_K@o{x9tzbVxh*1Gb$_fgs)`EBGX;Jf^c`@rfU2kZzD=mM$W6baoC?!a9+8lz~v{N
zAY;LaPiSyW<*sCSogs^1xh9~zl^133Vw~F7Po3e4?B*Cix$-K<T>GF+#4la=y_iQi
zC>wdwZ~P>i`x5fhJD;|wYedLV7SY|t!fB^m+sL&BY><cT!ndA4cI+cRd{l5szQ89>
zUmL`)$_qXLy}*ZQUv-g%FLZGwGC-eikK<`W%Wcw3Et+HgV>(#jSM`DP`#^jMT(xr*
zCwP8C80l`a1W_skLfe21+NKQ2k=8idwd7jPFeJyaZiB7972L$zgs~pC!B$(}!~RD>
zrp7!*)9%qqXJiZIY5+z&b(hrKp?@Wam8EYiBWz%u$E|wR;UE8>bV5If)+Rr8m=r44
zB(x(hLIF^McG9pE3rQ+lu~Jc0P;X7p?yRVGdD+$$LIYxKf>t^R9+!#4N_Rzi5HUUJ
z6uC7EGjE~w#i2?;P81PC%P-IxMOUIb$0)g}6}iPc0~rHcSc+@l?!k^uPuWO&{1n+`
z>%26uQRY5yX1tSp4eEZ<K{Q{m=OVK+wWxJ`ky$ik`vR~J#N^?hYumDtbYHL!=$k8C
z1P`zx=3*o~{m9H0oSj`{WgzXx!R)Tfd#M95%JYR{nW8w@!Ve$drNse?wChL|d|d&4
z*$40so;YBWbR_WyXA)RU3rRBAvOZ;!huA>O5gwcERFR#g4k)I@D~<yK5nNxyPPPEC
zg3{=B+E~?}>k|MM+PS(fV9&{v&8TFkb5m^~vGpvUGFY~-?l+c^V`$?R8M0`?WW^H#
z^UYN#ZQB&4zuUx&_~VW$yeAp1e;QOcZI}s0XiOBm0AUmE<r1UDGciHQ<`p6$2_bFI
z7nUW^lNa`-@bJYN)^~qa8b0>ta=Sl@kDUB)-Zsg^GyN`;DkdY)W0!>#_9>q9BJA(+
z;n<_Z3ipp5OkaNa)pU4t-?muTef{Fq^z`Yo>G|{L)3@JzH@$lKVmdwHTVCh(Yk~Ov
zso)Z*9K816@*S@LI9Ho#g0KCHg=L-wV^YgP9e1&Lzoc%fJk&pRM7z=2(24OVbaF`@
zrGJEIpN00t{UnHO!1ZJxa-iPmbJSfXhaTTNqb4XVe&nQpU$hfCPKWM(Te|Sl4>iG6
zpY`}JARj*gMm`F@Yua)Z-2T~P>^-F7*)OLh<2)EEl27uAZ2N%!!m#4}+h1&3$9cgQ
z4gavAK4p5930~W`CF_9~4SY<ujQcP93%F>ibH{8PHx81B@q;i62$#!P2wpPe1H%!)
zwvmdA@eZ;?Il9SHsgY@zv8za7Y-201P}#bS-POQ7rWIFovT_PPiqpp<63w?yII2G_
z6DI%YSl-YSjEAxv!rUgZ)H4ZTWxOo|>9K3=wGW9qP&&YM-Sr`b<^(fQwnJ$;$o00V
zwU}%UFw3kn9O9-eFzb-ntQTt1fh+Z)O;cvr;`2ZK3;xE}ZxVRj;pqvlL;PlX{`BY5
z%V)owzWe&8(~GBHPZy_e{SonVjp>Y=K5n^A(NxU-S@lD`<S3p*w+i$T7X4ozu+{d7
zuY{%HR$#qr!U{%uh3jN9p4Q&9g%6Cn*U2>8#B12J)A+XP9c2ku;3l46)E&deU{iLR
z&L)4GOhfZ;>Ow8EvNO<D^Z_>M+tOp$gsc2`+4BU4k{kV>c|$<JxDVM2O{qCO(+!;_
zuLNivMtXQa@v$+4WYhLQ6GK;2>4Xus(SS|)F}+C>HmAXx?qtzBd@Q|L&NZ(%D_ZEQ
z`Eb!2#<~f*=oVqnPhEQ6sSv8<{4f0uD|_oU9PKxr^nyFB3>F@E+ks|M;iHUV$#l6o
zTus}!LiGdT9Xu5`kNKqX8-3QlYAxxCXZ~iIfIW|;MeD4YgCeJ&ErJgrpn1fKKGY*K
zu1YWTvR(R)6<l6vfXB`v6LQllK!mkHzm8@gf6^7YOq;><@@1wex?0vD4~tnypg6`i
z0_51pft(PQBW+aWj!RrP^|mFya4R%)dXq057hoOO2Cni5&c}6jHOzJN>jvimTKJ03
zcGJ&}6WsqD@R$2tRlgtlm5Arl^iF|=FxR`rh1<MpPi2rU*BsD4IN0~M%`wTMkr%zR
zlO=wy;&#3Q(YyY+X2A6-=4pM72=pAyHxv(J8{D4LC00F(4ekXaox}H>hJM0gjLYRb
z5=dj$i%z|;wY%%ZJKj*!i#0sR#gkutE#HS+`{VrR<m`MpIlq{8@9j+o`}geMG2ghl
z;AZrkS0C>0O-Bdt^IAa`*T~mb3HT>+A?1er&_=o}obf6}7PY!;sx6nv@~Ef`EKc#w
zt^<FkY}ajopZLq`36WLRt1MhMyff|JyF2ag<vZkzDJ-^GRzkjBHP<KC8d!YWak^NJ
zv%rL9t{LeZU-d+Q$F&eInwFRFsYhQc5l&lhT?pA+&rn|aaqX0AEu>vo%dr@kKEnCn
zJLoDO<tFy^59&tJ_?EuQlY;R#ZJf8*SU~z*nYypGczAek+S{R@V;=+hyT9NKbcm=W
z27+JZnF`K-lUFg!;EO(RRc@r=Pv9mU;)PrEfslGhz8xCobkR0t!Xlk`gjuEn(0N^f
zfz<o<$6%YCt+ua$Yh~48=!dIH14yIe_}2^crEBY9CU-qwAH3CA)0EU5BLP6qzYbdC
zYKS#qbKZfr7RuW8f1fA5{_aoU(l5g<WBj|~pb$G~$5!NvImNd|jq1ToUSxbgi5!3O
zf!fXzbNr2V*|DzwF+r%vRb?3{W!v&qiNeXl$E6S)C0)eQj}L>5AQr(|XxQTbX0p?T
zIEn~EFZ7hIX>HqhhiLgnF7CLf4SLW(&G4YXVu1#J4$5{!H0wxCX?G%%yOWBp*rKHA
za3*6X`H}Y177~I50sWpdk2^JKX9vWPbs_;S%!!sXC4gN89|Hhq&%qyh?o#nYf?xTP
ztjj_Y*5@$WWpugttIo><)0_>ml%)<R2HHkOkaT~$-CEcyzSB!YA9lNLK<9%B^+iS7
zeAwX`p+1vJUV!34In=d#p_iP|LCyn)JaL2ahrzaTbP|bx>xUS>z}r^Q9h;OiR^0BA
zr@Z*&J8D#i)KZr#wv1IQE;L&SA7Nr1J10ihE#zv_L`@==6J)Z%geY?KZ_BIBdI8WI
z|E&n2^c5t$OuUqfJDJL$yo*hgve^zYiiD3ck;86k5Idd#iNmX><e2Tyarkhn@34|g
zdMy)O7MQZA#NrYY=KDwYr_Y|`DKF~vLKDm5;}gGE`pxUt)2ml6rgs--ve|u_g|;0}
zKAG&2zbAv-<+PhSmb|Y@){`)$dDQ*>gQMxm<A>8dK3c0ZlhRB!mDWVk@5ZKtDuAD?
z+R+3`<zX_PswHm<CvdvjLUrY*ER~nrPdwMJ5Kro?lR9@7mz-(f$oG~%32l0uqK{=#
zT6A#AWx9g=?ESp&hI3yoJ?xN98$qQnWs!@v7Aij9UvmZ?2b?u@_XTp&&x$ah$EJU_
zo&3Fj!?S+s-uV{|e*C~p`CB@speI|<slKt?<dd?XV-vQ)_@v1DGknIO<OkQPg_jb3
zBd<+r9<Wd_A`8qU#cd1S=o=iv)32N|E)?6eb&l<Da+#fkX<!~);V23z0C`9ocffWr
zI$dZH3O6gbWk|TQnb_ryj5kRnB~&p;;Vuhe3c-;Rwg^>BwZSpEfc1*zjje#V{bCoW
ze4%9%Fbzs*&X=@9tQk=mwJm=1TE=+olO$&_FO)>F%du#&OQ5`<bc~Q<DE4wP)eMsz
z%pJ)>f+rvAB&jUYL%^d}vMom&c10>p9|k&K+f#bk2O|6z^^LDTFG!E%Wcy+|V-f7d
zvwREe={M8Yzx?U+;`!6*^6bRCcNgb6Rxs93AGtvz<3#V^3e_*bQ-kY}x+Ew`mHryw
zG0cXxz?{!^{x<5hwzfdqZXGT%j4<-9gH0PF+{C*HM!oB*ca+(Ln|PbBDSMON$7DuY
z%iJW}q08v~cSClR8{t#wW4dV#o4k+lM`#+pcOM?xYSu9k-NjA<^uS0TVWe;Jx6?QC
zK`xvXy>L^TRw1;j^o6%h1Dmul9om}Su$CFqndgMwyv(V_5c*Y`5LLS>9vd3x(k>(x
zosLJnB~4x&doqXH<d;jv-S(bzjrt`w(1LfkDvytiGyaY|NRc08oL*#<x;CA<K-uJg
z)9ZYhtAy3Mq!wR72;FH?#hYcId7diV6)(DvHH6Z$JByBKNM3^jogTx=PE8dW7CG>w
z7y8IgSa?esxS1Bbr~^G=p>NnJUi#LWXFlM%v-TCfLmjTcHvyX4HAw^bHra;R?kX(w
zQHDNGU+MH1$Y~p$a1(5k|5d_VaVfEA@^tQhqBVPl$Sv~30kAxAL(sXrgU)@me{#c-
zPJ|`TwXo<hZQ`7n1sKlbX<t9_b*Z@-I;}tC(d~qv%;MTVS<KI}B1*q+`h}kJM3a4v
z>hoS?ZIE+N&hdE-;9ZlPe`)u5u@_?$u)*BNPitLXD*eH)JUlwgn`k(n_2SrjzSVW<
zZ)Dlh?K`HuIX?B<V)pOt`f0KAi+7@(OvfjuzRtn*oP+(n>G7in)58ac(;hd}?%s7>
z;g6rr;`)uoQGCfO4^<n|+mV0xLRnJ0Dzp+T>JU?Rl#|7;Jp~rG0O}sU2!6Gp^svEV
z+%8X))%6bQ0=wkN;#Y2Xj9)Co`I?P=q8&-gR|V|je|eStLbj-AU)#Vp`UX7Fxqjkn
zF>X6kODAQaEp~Y-O+KZVsR!RYEdu_e6;tLV#9w@Ce;i{QKJBdhX)k3p?W%fM_&VCV
zJMA!T*14KXfi3;)V*$J|1~zGIjq~5+t>T1D9x&5zvS2I!8rnh)m@*I}yD8Q51+J#O
zEquEy+vaS_%rKB?_fLUswyw3m0b@O00kLK~L>-7e(f-HaqjtoP%*~+}y?hO`Y-Z~Y
ze*-Kwt_>gcv4S>P`1(74CktOp`X~TDqqS4BGfCX9P|}aBMLnnxhb}Bz5kHt_lBx5>
z6Vuok%`nc-7c_!}Z@CY?*iYKR6=Mg~Hs^5x9-WXa1bj9FYIJ0@s3zk>Qur1%06a<k
z3TKlJ9)i(ATwdX}<$v;nHrjSH#@(5furEq87`o=bHTp%umc`+qO`1Alpsm4EOg}<Q
zN;;UuK2^scsb1!R7j+J}9)LBNr}J^(M4khNtvN4WSdtNrOL=V{WpWN~_vDX0xc-*2
zrUDQ=B}zW-Cn9?x3i<Nmo$tDVfn0PG!^egbE(2wQiC?+63^%^w1F<hI8_j__eJecI
zZPk@Y(M~FbKy1csm!x7R`9VNKfAG*E>lzfAd}#adL=9M%_Ev;JCy6EJ3|w0>5Id$@
z1YUy0MNR#aW9qf(mNlOUPfAA{r>gCQA{80gu{|#ISG^+NPf9?uLV?RvE+#+PDs9`g
z>0eCRz~NB|9E81?Abakc)<r&n{+T;ypi@P81wfbI(uzF^P=eC;MPD*$z9%W@>6d!K
ze*hd?I3%dZHp&is$j(Q0JXvDa1YPyvKKo8Ni-e9Bkp+GwfXbT%EADRMZmUBF7ND5K
z@z_5fmE{gBFF<Jmv!e;-(Srxm<Ht{?hYuf4j~+dqKL6s2>EWYCnp{4do{0Y7;RAo*
z_1^wHPsW*i9vtpXM@RdTxj%jN#plx(Uwtn6f#mk52clu?=>B~#!co<XAFdB@{Ptoq
zZN!~mHKA5uV&}<{%1+Vzx&Wm;9(CQgyj8E%lj{{aeB{UvDan<*_burF?t?hu*G{G{
zFk$3QE@Jm1Xo^{wXEKbg7vWk4+&&Ow(|&haIQ4ypj{_%9KT+t{ThMnDq>H@$kcabl
zh<@9Z_i3k;6HA_S;)7&Ehz;xT!!Bv(gJbFU4#lUA$j5h%QOA<cYaqmTJ{2n<=VNd4
zOp7eG6XPiWp-JHrIp?DBgW;dhEGD#S*%pv5G?M`k+OTnjl3dAU-|2KyTQ6~JgOM&C
z^<aLbUB-wd%~m|uUrG57om?YV+Sx8F+os+}+_u@Lg+M9Q89c{<n7=j=a73`rCQ&ui
zvTa$u^Q{0Zw)B%g;^w1mp#m74K#uw#rmYtqHad3wneLQDh}o*zh_a!tuJo}@eI!NQ
z8*>9IhUf@AXu=&QVZ739m*RULfT81Al@4lT+`OR?2k$MLjrx#I+cr^U_^1E$e`yO}
zo4`1IF`XR0nO?tmI=y`M&Gh`+pH1KV{NE}4-E?;PR=i`bFz94F9|w0m@h)Hbw{3_-
z=X9$2XW@kKfa<WnBD9sIjRqKK+xS~F49IQ^AB8$5Z^EbA`c&Rca+}bygzIYhqxOpD
zxlPM%DeGvX<wn{je2PrNO>!fTt?@0g*V$<LX1K15Gk*)*#NWo-#5;7(Bh6vy@f*CN
zu4oA(T*uo^PrRZlcz(8#e?vqfca^uMjp^%XoAkgoz9XlZ2X)hpa%1o~rD#oCr+ZF9
zHOzbx#Dm<nG;kf?p~|35EE+8y^es9FC9Y%UOyAHQaDL|JFutah%!qGG`)6HpzCLd{
z?N#0|W;h+0B-ZjBnhCn*r9I|!3>1w+bEjM0tispfD!tC%x8$3yRJLb&WExtwq0=K>
z*cBQu(iVEBM}BicgOI2g1>q&$7OJ2#T;W&VvyDodW%;E|JA5BV4*#J-4IBk7+u*E9
z>wdOb{tlb4Va)_yi+AfoJ^iJ5)-7I*_~Bi?9j<e;y?gu9q0Xr}uYa#|eilgB@5XO%
zk9ir#w9b_;FYCKd@9XB$YAepM`9-0E2g$kyvk}PKH@NQc`2?NM3RqCGtj1U0<m&U|
z>I?8$9P;89uX%j$MXy7yL$Ek>E7t&B9`WGkb%s0suGj0gZ@s|B(_CB|;L{1Oj(HmF
zaymUf_xQzAU!Ofbnm&JYB)j)~b1moOEa<ps3R$329$a{oSCoS%#tsj;9-_G_*CM&z
zqxomw<f^~UiEp_~EPU<l?K;gX7xDSxlIw+8u=36F*j2uF?$&iB*V%3H)shlKk0-^>
z&iRfP-|#x~x4X{IFRG3%Qx^pMaX$DYkA1=v@U;O#wF&3#)Qzv5NQ~ngIM;aaJ=YrY
zEqLf$D@q=!k!o}D@bS@?meOtp3<~Hcj+-Mo^1k*#o6GLr&h6>G>VN;9eDQ^Z923YO
zHWN2AejUx4%~NFBHL?~M>|ZYzxQPaUJ5;*CG6ONk9tfL~B|AWDF7!l=FS0AV0<dXs
zhJoy3-6y~{Ti4p(fKQd%gkM!}6MkcMI(%<?t$97%Y_oZ>^WXp3Klul@{^39Ve`F66
z&1_OeAcUpn^`08r8Mbr6tesiJ<@7*<k8n7-%v!(>DG_swglDZ1vJKJuNe^O=5*33o
zCGwyooS2h%Pm^Z60AESjwwMJ_oloZ}oh+`1W1BE2A2fa%1kH%Nc}nIEX6!iQS!2;n
z+=^fgEotr|B0z^7iMt=Ln>3Oc@PhlmDTW7L4aVT`-SGPN;+H=HDS8er)f#bLvdEJK
z`dN5lQ1;;OOl3o9?g#PRvv!9?iST0+!8*YEF=S}yBAZhB9l3QN;fTp(f&<9qg$_PC
z@JPV+d|<)9I2GIE5vtv$Eb#o{cOBFyqisqGIcF^$5HGGlQ`#oUL<k3s95nd|s^n}V
z^@%<wSXdIvGysus*_J=}ND8tD?nfZd6|rE;hd}${X6nkCwvC+W`ayHF0%-6HZ%{s1
zF{=8YT>jY~NxJc<gV;`vy&)<Z{@^GnDyVq;X|wnX4plnJi?tux->T$C1xlax%w!=G
zH|+WS$d$$te$i*rhUk|a{e1TevT7S-y~w08n6AJ@yzCBef%uCycmB|Z8T3*&mXe+^
z<%bVUmSoJM6AJW6%UEIY+Y?IZ*qqo*DD9(sa+$@atdxyBlW)v;f2;`!_Lzv>Q<_QM
zyZ3xl_SBQD3-$d=O}<V}j;GhJ-)LXB@I;9TQ@rL8cCgxaHF>?bIGavSPd#zF)C8Tf
zr2gM4KOU>+)dNiU(pI~>J6;4kJ3sTI{~4`RPWiUOxZp7+KC`cQg04v=G?&O5?c3Ci
z>WlFr`w%%Q4^SEHZ)QnMtkCt>##_s~kD8u3my{Qz?=Z1exqY{m0)9{jJqZqvR~qod
z8vQ2qjX%_{dmq{0hh%A6w@I}h?RaPBp4*LfrAS`D5=Q$$@G-3XPrKuAcquOv_DpWw
z{w66wo!BDfvADrhWP|w!-#pV5YvXK%<Fik)B2oDy9m__tt3q`{M(+Q|-k)t-l5FXD
zm`B9!BO)`hIZf38NKh2XLFr+D>Uc-J0|^isxI2R{LWz`y$VWioMh_tAMnk#-psG4`
zvd_q9cN-D<zyF`p-InjQ_TG^h2P_nF`ZBlK+qA~rZMiM?1=wF<E_3khUo%><{gjb2
z@#w(nQ-gCN4z%xB)3>UN7!JWaprk`g@O&Nh_*AZ-h>eP+4VE$jw2?KO+IN>fGemUz
z%rb$+{o@JtlxJq>=*EBV*I;ftZwS*!*rRpoh{?hmpcyeYO&vECX&=Exxp4?--$IsB
z$DFJra%113?8-vziQS)7dlb)O#k^Pq(l_z9f<bj|5AiZ*Qmzba9(S{gMylWE{sG`n
z8u~R}42#VSLf7r_&;R-VaZvb!!yNS=3Vgk+#+h3njl3#cU#gEDPDdvnO^3&ir(gc^
zH`C|8`s3;0M^C1YKlyBW^!Q_+H$C6t^(g8b+q-Y<pjt2q$jzJ|51bd);BzJjb=U%3
z*X`1~L;KkUp2OY(H~F~De*=2%uflHwH11mNUHsmqp4;F~vEHI{Cm*B_P!@1oU2pQS
zO$P6#Hg37+4%<NNU3M03fo=K?yJTy=BNl7n5BYFlZELT!uWfZhXK-WK78|@~(~RKx
z{<WOb))0QDz*;wcH|2F!*#OVi23Y$6w)x(G!3TAQiq67cY&PQpZjvo+OCEfyeerJg
zz1f#zzd3HvT;I|4evpHW_G@2lzgS$?S;q!+`{zotS^us28@%8fHuk5Kid^eoWR$hd
zH?Zb!jbZk=MRqN|(A&#X``j4EBHL*%Sj%!g>N?hOSSnF9=LQ(;3iwXh(EMv&WVSs{
z!5qX6`3-Qr4{reX`)>Uj7I%g|w+FyOdkiUmcW-aH_h5f|_uxQZ{12!7gNM_TPk%A}
z;`3i?E_F1$eD!*I`SR6tc5&+O@_dh{jeP7W;ln@R<U6%;nmMn@yXWuc{r!@~mJ8YX
zK9<hOZfAXZVa4BD(d9es1HLQg|FIrD;y=WXyqLoqV_3kURDAQ}P`+PZ$v5{1&d<)K
zH<wrPH~B}hzxc&3rvLaK{zucNpMLDcuxHO-P0wGvp3Y9s9Rq$JKR&Uqv(wkpeSM$$
z^)El0e);(m*X#RducjBTdDZsSw6E{!pFDXq{exe9I{oUikEx{OFQ%`)eK!5=m)}p{
zKYcO1Jk9+*MjmzIKEuh;@$}h~kEYK)eKH;F-=AK*d||tnuP;<5N})gULucM%!;g9o
z5B62hbG7^H>9zFUX!7ber?9}Jw&97hCyyUZA3Z+S+?j>0H<Lco<<(oo`gS^38(dU6
zKfiSQ^t~zH$Eo`|923~~!WTf??l1ps1oy<~FXGf(SM}k?%>H4wS-CI46JFdeU_7Y%
ztb-5w<khRw>FoSsI#Vo^#XU$S<M`zMh5y$~w!Y`Ym=HKNMyN>qXeS-+>q<`(?&G~T
z)29y)r^m;l?-8T46^A@ZmAebMR|43+pz$ms1>nJ*^^`Fdjb5`&+c4{IS(gJ4!wBng
z#vB-Vo0cUTnDGT;ty#JPT90IZX3>TkjfEp^U0?&Jts5WW9)a&r=^-8SgY8W)meH{0
zwa*dOZMgx&N`KdQS;Mw6e+=wuhxdhU+s!(07R&i_H`prZ4%l8-yskapoBsQM@qf7Y
z;Gg{IpZ&IP(8FX+heL)ab7N5K8Z@-8H6bMTViSIqqXsyKKNFQ#iq9@Dd3<xgq7we>
zliVzLX0UTk+^FKxuY!a2!6=>@*B(m_g(Vus?R0b*+EV)oPM4scC!g4W_zl+FbYA4P
zFdj+UDYtJ?`UHc=P3XMtEwA5`jG)oz3>gGQoDQLtAem4&PQ<d~#LmIs39J323{XFL
z;0r7KT)6Q|zvItOML^?U8dg)no|`D-KBYptmyH<n$xn%kmmfVAnZ*-ILJd5NUG}3L
zI;pqtiYc+jHDH-G78CsIKZMvIcb-^Xiv484BP_KCb(08}nQU-T*5#pWDaamdl3E2%
zW}UeVqu%)jU-Lm_35#;(F;1hSQ2e>=YK*1a=(rw6@f{oL>tixT(ZP^@D;h;lvf!!y
zmap}->-r`n@APp*88f};g`V0F6x!E68^s<w*Uq{8N`Zw61lE@z6F?=K#cuiKX4A^C
zF|T*@Ok!oi1TWkw3mf{_K5&(?`Q?w@+Yfx{C@yl>Nog*}&SsCEOySI9!j;W9S%5Z^
zw7xW?9;L6GN8i2ZMS0X)WpE6zkb^31!K8@S2RwRs;#Utm)`X3xyFUHolj)OBo=hKo
z^ik5|$CmS?7_SE4k6ii17QcXELF?#<U-*bVR?d@ON`9ZT3gwqq{HlxBA~K1keG&`5
zPU5)awE#-;>Hg(d^zlHg(jIOq=R(|u9P_EabvT|t0k?Jfr^U6?zWN7goW`Hx@Pe7z
zBwrNh&-5YfM4x-{P=R}qPB^ykdZF&VzNENsaVeQ&gjTX1xEkKqC*{OIeWHUAy&fHp
zRdOcO&Z#e<<~-JZp`C=sCUlRn(CAj~R5nW5ZP3w0hxjx8Vw1Aj!c$gqNh1@_;%K3u
z@=+=M>@1thikfzOOqq^ucyI&Dxh2Uer5geo2eva2o)^flcRMV)#;qOj_$VBUXw@(=
zR@%f2ASh@#06vFhL6C))x-7iLmjunY)_$gYvq$?dV!62<@Md0*Iqc4w>wMO2q3Z=7
zok){|O2OrTM>02rQE8GpxYQFY%ti0IZ!vAtFyaHa{IbCJY@;4tnJlppQ%WTE@JA@x
zz8ERRDLPX2L#)Nm^|o}h`}_99*W4DBHx4X#us2t{4)N9W`sGvqAJ{iv{mt~v*MB>`
z{NelQCI1`w`jyWYoLgvRzhc<OB-WG^=<_4c=L7fiqH-r#;#&Yec9TM{L64{H(hmoz
z&&?H@>%9eTl8rFxj5fP$;m&$nbXqprujvs+{W1NNx;Je2+XAC}mmZwkbbz(aHoXnm
zmileTSFlaC#<!)lOgr-oCO36~k+;Ya?{)qHyJTyAjxP&_1f4m<bs09Yn)&FZ2}i%t
z+vLr%f!{F7Gbgfj<~GU~88#%H$3=Hr0G@NMAUd0Q5;&Zc1JCpFa%p|*Sjn*1Carm`
z?=xb_g+xx=_8T3P&94;I#zz@#p?t-2)WQ=pZ3Nwk9UpCTxr=>t<7dG)W!RG_M~o$z
zTy(8eeH2)Hkvx~P>8AfN%{*-`4_~fVk54<;!3#fQvS^FTcP2QYZQgvVEjE?5mOs<s
z&3xtd$JW=z@|*HX1|NY<p@#$$Kneezt1DSKqRes%te3Ru3s&%`l`$nWzjBL;zA0%h
z=s8uOea$|{*cxCX8&cYf-Q8iWzApQww2teV{Jj%Sk|2R(IS*yZoc{oO$A5q+=eEEx
z|2L)Y+q7a{4|u8Xzx@0&cHZ!bKO>2cdj;rOSIN4iNX^`$b-puy0Do64JYy;MBlvF0
z6I1+kg1-w(#*<lm*XJ)5?y&$S+jlG^k*i#u*g8Ew_m4mMKP>JW^mzl|yL0Xh@ZC8-
zf_-&*KD|6WmCZZfW8j_w_wg>@u(<Yadd<St+4*#M@L>A*;n8&O&6&!)^t}z9a^m|c
zHK0v7|NES;#C?au!@cRFM<;nzqT1DqScm(5n+^2I3ICOT=%>maD?K_<`|ss-iu@3k
zdoA<%Q+w_k>?;oHK;8KfKL2yh(`lDidGhU&J`;~0f1u-dglF-Qdx2?lHuLDSaNRfg
ze$EezW6z}xzJIVwU5n=aOrG-M`*`1r;C@Ekui!pK`N4PUkhmqVk1-{J06M<!LN0`b
zzLSG{)1$-t)4?9s5r-(dzTTvd!N(lf#e*`Bj~x2D!JwOQlh~3k<7n0+kQIzR8rHbm
zK;XUTmRfI7ExS$ucx!*!6~4%mN<P3ygGFzt+cg->x$iT=QqE=?VH<B3?9@|3W!w}k
zXut2{|0d|M^q~MxPVeorEqBA+eddWTFMRcD3;}nyoN(g8SHGf;TfC@b=%W&S6P)2h
z+lM~>HFU`=@~-m?@Mw?<#P`8#MflyRrCH_8Ww^Z1-YgDo1@<sYZ|0G!tkw%xU<8>m
z;w(sg)gEFo$lVUKP71UvAgszOCz0i2-rN>Y#Lze=8BVSSPKfa9mj`x$Z<Ch74IYf#
z=MC*VScL-iB!Y{ZbwzM8%ehHA>r*zRX!A|!ibG<$%CXEorGq0p+m80B*tARiZRC?b
zC+b}MBV+4DxxIV3DEQY6xT`qq2w(6(Q~)l$<AS9NhefHEeb!=}_3b0$H9LX*<AW9r
z@s7tV@M`xq?9T$Q1=eIo9X1#O>G%?FN8-8ettZc}oS=*7*hJ8-#^F1Q58vFA-Q42P
zO6oOHB$4VbKLrCW-)Pd#$onar*zuuEG*X^rP$r^Fm|TpR1i+@9eIp;=)=R$$@mPVz
z2RqnawD6!3A%@G);4n*dM&=7Ya%AFr%rSr<%5T@1cFUrO;-dh3qVKk@1qtioE52P8
zp!R2cc)VZ=3m@dBuwlsu1C`@RN*fc7aH;v%J!Yq*I4Zgdqs>vr2MgDkj8;sPuedy+
z<kvi!u<@R2CaS!bn#nYi%Htyzu2`_j!dBkrU4Oz#pDNww$tR9Y>P=lGVj}2C7QPh=
z?Z@%t1swGUWxCo*?i<B}t^d<cItcuvmnO3Y#wNLpy1jUs*JV;k*OT)I#}9GIpVZKF
zzZi)gIcO1;6i4~<_)^Y9TO1|r<$G^Gu_JrH*rfkuM>ANiMDFtd#~d}MEYWU%e5<WI
z!PMsSj>nC?Xir~LpVY~Gm6OFZ?1+tca9ibI<+#yJ9UQaug*Q($IWdLCCcdOtdWn}5
z8eiJ%8y+#TMPJD5D}6wY!}zrsJ423R>x2=o9Ur){qn^&aC#Nl&E?x=(n{zd1G`g>y
zWdu0QJnxlKAESmYA+iXNmx8i664-@1i4m;xhW4}YI;Zs5X#=%n!7IAr$qxi$1_eX(
z1Dy-ZPNWcE_LPtDrh&?oSFC)Dwq8e~y)$-X%Dd6c$c`*~oCtU5qB|s@e8*Mg*}-sD
zqQV<Vik(oG51|;rt;c@tdnw$qAWlUIA(x+IrDKKCh-J5v?GU6OWy**4`1`i-1yt2;
z05n>QcH&pA+=zHRy?Xx4{|WZ(SHGXW`TC3L`Sb6m=g*!_XRluB+*s!eS@=gH%DW$O
zUF;4ifHi^*wj?q+T8^0&+$0mU?RcKMDQP!fpWX+y`idZV!8Q2+YuPrPHH^BW?Hb^Y
z^4mP$=5K^qH^lW+F!~$uHMDG(ykXSarF)yaVbojeu4S_ny`o3?2qSN;GvdHC`I@)l
zLvP4SIP=)d&v>5aC8aaMY*VRW8&BNgfw%8Pr}0ic{*CJ-Ar)IFpLIf%CxP&$M?amS
z0&a9eM&j@#*EA$@7dzT5sd=4`d|S<JmPB31GtZM=mDTgW*0XitYZ}O<ZjqbMXQay?
zg^c`B7JRfTd_xWse$u9X7|T3%EgSJghAOj)G%ks@fo_?V7k$I#c{ue#4LY6GvW%73
zjpH^&2nVrD7_qN;8_8Vd_)$u9hV8o7`BkBfe%`8Aun{x17PVf*fzSg%64znVUj&=H
z4O!JAy!P4UwqDnj_KjU+MhAc~B<pPYia%h`k^AcTt;ss1Teewm!$J7#)XhoyvCw60
z*6YK7V=I@w-y)8>SNpV6|L+%zIegdUKeu^pJ_|`Kd|lN7i0y|R{o(I?@`+vF>q>d5
z2V=-5Klt1~kX89<nCt66$mB~pEL>e(UCBpY!FX_VI4`L1J@rsA;vc}Jy=upshp@Xi
zzwkYtgMI#g{Ln(a`Lh_se`WEMl^2i1yLj_<Iz7Gc6H1hwAD!k^+V}78J8qt0I@O$$
z^TpN0`SjO+_5YhLF3zT3eEzwA)XUROJe`y`y6|J$J^w(LvQCZ;rY9fQ|AYx=G3@wo
ze|q%j#0y{W9z7%-Pai)z_Ujanq;o*xN4;Kf$~^$CA$;%Whra}$dycdh3q)DmqCsd!
zHrlKIFP-*sTa}&L-FV#-5l{yAJ;)KV;h&annGn;&!F?U-+5MgRvv8NbYDf{X@`rON
zZgOv{^-_vMO72+i`Q8*g_2}rHlIo*!-1X%He|%Rni#lD)lSZB~?KXIY1Hl&=c*_|B
z4WkW^b_-nhm%NXGk=L5$EoP<DDbRK^47Bgt+ybLLSjyjSSC<1r>yB_OmMw4>9iWfr
ztsv2lDf@lkgZr!G0j+d*|Fh*zcpqQ2@b%}v_4+x6z8ix~2OkCT5x<9l_wx}(#sLSe
z1F`5n-~g{cP`~`E-@>gf*Y+;omdlS##r45~t(67#7oKh5YnPbNi(e+U3`F=|h%r~E
zCd*SNYucv$D1~dBa;yczPg{XoP&u(8qAYG^n`Jv2MK#Rq!j{<H{3w+$>RZCZ2Tnq(
zAs5O9PJoFB9!`}ZZv!eXUmV<FrOv*gEL++<7e+}r(^9U>7b-vYY8O<{Oqb`$Z@d{O
z9~&?}ER^ch?*Gwo8_AdbT9^wyH`S?&`$(3)A<qR=7c3?aedC^|jJQBiB0hNEu*)YN
zEWyjSe??Oh3gWe1>AL@mCP?<!;={l6(8k5G*(^-huPQeFLft}rtY6A3pjel#gk>of
zID(-zmQGKiu#P`)u5jR;QC^9Jd(7$skJn2YRG)~cQ~EJH0^#WwCIG{0+}wTAlfG}@
zqAQ0!Hr-B3T|C~&)fV`n9Q@F>j>iiQ(W78tw##HdwEZE6k9*NlB}9BNd9tm1`bNHJ
zCMEPgzA!@{1ikpG$qS1bIFT)57T&#kj9bZL;xb5pqMB*yfZU*1QzxQ|PWqOd5B!vo
zNd}WZ(G_<lHmMOCAClWZ9Zg(7-}I+uetn-8P>7pY;+t0x-q%EEk9Ox6s4_hnqQ2%;
z(lO!VRsL9n^y@NRhxGG17OI+0$utU&>oZBGJlBC=9v%9f;MB$ASUxj0QW5*W4%cwZ
z2gDGHkGFIL#~M6=TzJAsJiNM*_Dt=Q0&s|)O*wU?-c`0*M)hUETa#DUo&Jta+Jc2V
z)t@$?EaG5BDnDM#qRn)EpyHh0q@eO>PZs)mK`-Tru<xj7M;!Q7%D9TX#{dnf<Hhxe
zfgfQ<myzIcfeo9@kkM9MH^h_?ALbD=0wwoo^YGU);l2)Xth#K%Ko%d77f!N0Tzqud
z;?x*1jh_bMqio6EN#A8Cw=cxx_7S{e9up3Y<0T}P@}s>6K*dN5)}NK!c9d;v_`+N6
z^AL98(#o#43y3bh=eEEHhz-M!AoBPveUJj`YYNqh)#oJpMvlO2L&^Z5Jdjk<<vDck
zP_XZg&pR>lpQM;+$FWzsty3YmT=N7r(av?Q_8`$;;#iN;7;f}kVon_HTakz;Kl(OE
zG8;M!Vx|nop^dTUQtUgpKjnkt$GXNTONhZr_SZYEG4Rl^9{y}B1;&*;xzBAIoo+vP
zT_z^x*hkz8yVf2Q_xJffzS0gigWG}~?O|hH<#Ik<oWGi0y?i?T@ZDF_*I)e2^!+zq
zPA{H6ou2>jbUHgdb-&WC<h(M5YqkgC2x9<M&el{>dl+ogV1ag(>v6}t0MvHN>oDuC
z<EZpLuwk>s_9lNVYaeU6VZ<9+XPXQdW$eea4Yv6k@filaF}1AIXt`B)fSY7%7<G5q
zz`0Yxc9)-lKg-A~&9X=gvtGeSxBdvt8)fXb@;bp=$q_W|t@(u~mF)nFY~kO`ThqJk
zobz!mzKPIIO|vzx(g<r=%NII0zQWHn#@ozW+l+MSmoBhfPk8H)L&A4CYrS0a$j3DD
zHf+ja+m~pjM|_u@nY5_#RM$ylvu@FYtV9bv>$czUnkN3thhbW^WSqJ2t7}5Zy`kqE
z|0FCL4^7mW7vRh43z%2_Dw)0Fv2ihPoIDVGNyQd4J<9E-X%j2DVCGd?+H|vSk+nWN
z^cFr4KJraI&~^wtUn!mF4aph#lB`mFvYE!%wx?0Q=T0MUod985WplEZ(kq<!cEMI3
z@0r%HP3Na1?NpQMA84vWE+q3_{>I~Sy1F>koa)MR7fYq*@8Z56Qjn8_{gC)>W<@|b
zUtuxsr|$6-e=4Q*wm^&U3Cb(?_^&EnmB)8AFS<xTtSl_~JFug~RszSveT#Q~H6#CR
zMJ)WdmIWjWMIRhI@KaVS2=N_*7c}x5)SG&`i+g$bA!ruA_zs@`1mksrm$EyUo<99y
z`v3fY|KHQ;>lf4KpFWvB`|OkHgeSlx<6Z=BreRTtdw%@zmxZj8W5uyYTU<$Je|q>R
zPkSBgJ(%|Q`Kt&P`S!d3_V6K#UkB4s{V<q?NHu}$!2M4D*O%JE{{de26@33;ukNjg
zN1~mm{|Qf$$sdbUv@JXC@6eWMcd)d*%iHJvMD8DU#~||$ZWU+XoppIdWA5+p<6hd0
zULgkVQS>5W-MeG+_oMh56S{ql>39J8mU;WfCu$r2zZd!vC|Q>;<n~B+#=aA<oVO~O
zH_%GaOAZ8IWFtQFl#Mo<`ipEE4*+)x1w%Q<3AmYp_Bk^)d>97$nrsDYIaFe}Zs+kp
zUtxoe^)}(Q@_@VO;P_5@1H7;PZScYURcg4q|Fz|HcpqQ$6JP0O40Cd~3w5xEoO=7M
zZ)%kDKKg`PgPS(CK%w+%>+jNigP1KRtD?*f!#7T3z|9a2TWk1+kT4wCH@X%HICyhV
z^eF*$mt=5^(P-bjMO#>-<_a;Rz0}E%{h$M{<;y;tNbtbPt_asih|8zVLUt(tH)T0d
z0$vD0mlKpviiKMmo3v?D#2If{U;rZbUb|=@QSKAB@O)erI$R*Rpx1eQhFG!*s@y&)
zv-JhTO-EfRPjxU<JsR&D&AxH1f^q{s_1)V)@c6Jdd{iJ@!~*!jM{WQ*BK%7Y4b+hL
z1+X;4t9Et$6eoq)r*yE7mU)c6=7c6s9L<BmW86l&Z^XkQ{xkw|*yUdrB%@Hsg!i%G
zlrZszz`C&J<A?-bxnFOo<V8TY6YUZ$A2;_UNAYd^Wb6|oI`S)Jo-!hC+QjP%!uJ?&
zjI}B!<4%}p`p#`DH;hs2JZ33RDqdI=;tRf?ypaucmWme!1dbGCLB42Za*)^c5g+Ab
z;zF_cV%0H}AKNL9lky`P=ZYz<ftsxhS$JJyHPD9Ie7wnya<J{Vz<Uu50$n`9?_-o-
zFfsAUxK^SZCSl?qs12x>Cwrc-rH*oKaDA)O<2&`gr!sjopRqx8uP(2qOa8CR?*`{r
zUx|^rB6ry;lX#fu(T_aI#Y7pk{4z{_DO+WvUTG)VxBJEnp-|=!i)^rPf3vwQ(hhE4
zj%9pMHig-Sc+zi-K}^tP!siKkO;SBM#E<cstZGBXQ*eIjOFlfA?J-ep6gq9kFU$Bv
zRi9JLL&tXX2Z=TiT|ACO$3WXqZd|zy+#dK>EXb$>C3Dcam(W*&N#=NpVB@1wjGN91
zK$1V~LI<Z%ijIx_ATEt8ptLr!G0BeHeA$b3Uu*LRnS7fzLp9(SDB5y+FC7SllOF>V
z&q**;!-Q;`l^8(akXtP9&ahclfVPENsm?e}gKu{aJo`mI^`H>SF^&ob=0&F6fD>EV
zlg&`Lb!Oj!_D9Zm$8R{5df$+8X*KxW9(G%%i&Yrxt<}dLfI+weh|(we>`D%gV#88y
z%*?_fsEnAb)x7pd<Eoyatr0$$4MW+hXg7t$>?*JJGVV?j?lFxjJRuMUIL*^esYS4S
zeQt0W)*e8VK7U^qzUF57k>IiFYY6`*c{;s*`F#4GMX)dbR{59H4@%E}_;z}8!RrwD
zuK{kPaBk#W$Y5ATLQ<WVfK_m<k^Us@j^D?Y`*Rbnb2ed{e3wm^F{ayK)(x?wQC9R#
zFzSppH`%t{Hl0z{soQp!$JQJ1rgy=pH|vs*so<L40HbV7H^EwOq;E^s_kO!w!!{Xd
zLk6y6!1-FXg>GKMHvY&PW$c}xTV_sR)VEH!1Fvig7JkwuZ_#c1y3TeUAqj&y)8HP1
z$~_NaYq{CG>FYd`0&wOot&?P4PDfBB6JC0YKG)|>Uigdd=wqb)DW2)Urqh~7{ddV`
znr+Q9zVJ)rb_B@Q?Qoqgoa^<54BM6mA9=Z?ww#2m(#+HOUnx3k-Gx@B<XvPY2iZWI
zS2W;xsqwYEbn4_*>{%yz>f_<~gcW|E+afV6cFi+B@B`pW8zz|@>55Oj*t8w|wf-WD
zzW3(gd(ks@$!kN&-#ZFVie1wiK87?xlg`<#68?i~cN3J(W*NeFY^uFC8*WX&Hr@gq
zSLfGZyWL&Emhzh+mvS1qJmyjSNc7_J+)vbBUYv>MDX{!klw;H8|1RS@rRe^?8HYnE
znZKta7xee({(cbM04%d%ajfguL7&Bvvvb9BaZxcyL5h4ox4QhhZ1Fq}RTe!$87!Fj
zwSSr$?j7Vms(2y_AKVK%I?P`w@an{TAJ^=V|H!&HpRTUnPFJeu`NgGd^i5T^htlOf
zAon#s{q)o6>G$7F|F8e&|6}^@yKkq*j~-5c^v8eX+)ruAH}^q~4~~4#gtldx`j)>P
zQ9OB)YhN|U>lS%J?ds|(KcE*ki*4K&$f6SQ<-SYqS@}MU_{dp!f<8LPlVtog#PJbt
z)#1su$d8ZU!^2nI15iP^Z%LyoRljo2g7)E_L@$2%$I-Is|J*{IKH?q||Lfk1bllr=
zAIlegLtD|toD1PogiE8J=r`rGd+xLKoE+yUWt={k-cBcn_opW(`_svO+91~uoLNx`
z6#WGcT<5Vj&S#J{j$q_lv3SNdWFVEP9u2chu+}Hvz}t88ElXN_MjvR~83uYSyOk!v
zb85zz*aaF}ID(RI@K$gw-Ze;ko6a`*t$yDNs_j1zv?u56lotP;Z-cw&jJ`%#^Pp?t
z>rejV&wtCk2M&5wRp;0qY`!eQv2rew9PDK;qn@w<(ZUX#JoeK)ZYtM_-s;ih6N=Mv
zF!RWWVdSwBj}t%m90OUL8qBVpxf_bgH|i8#Y0eN0*Pvj~z(phTzBEh-Qy%XyqS06=
zSI`)o*kYFnLu^YYAMH4?kwcJYYGG8VUoE+`TWa8&)b^E=srywV{Bmzy-z1U1`lw?D
zVi{U#72hx>htCHmwu4N0u7`r<B7ig}blFwe25g1*LL1P(gs9@IpRHO&20f|B1U;dl
zPVBy!>v#(6J0r$rJBF&acJZ1RT~;b>4!nc_W&=dgEufD4(gHnSe7TqlrzN?VQNFEh
zYexKP=cp)C1QGzf=u-~yRhzZUC_-$wV5;3Ph97q($n9y)sdMPk2qJ(_+Yu{He8E*c
zyodl#s_AlmRa5-*zJc5qUD`r9bib<3i7_Xc>mn&(aL!0Sru}-Mh}@b|wG6~o#FiZU
z*$;m4Z)wTg-=z;DLP6`2vzIs8o<6cVeM5Uk*QS&&+3ZiWV*qC-b|!t9#Ni0qzwl9A
zcmj;4njSo0(TjzdjE@;`D;xEsj{{<NE9rdcaZCSt5|$^d_%qZ@QpMs4FK$*+r>+m>
zQw8w=w}gDS9UM7jpqF~H&27{?;c8ax5*|eAW|d)*CO(P!js5Ka6qE9#_@x}i4wKR>
zJkbxC*b%Q6u4+Qdgpa4aI2Yji9*c*X5YJ=Ju~c~?Jif@4=wsK@^!0drUP4xW@^m8;
z+45mNIDJ~WL&8`4ppk@mFGh6OD)&gS9?hN{Zuqn=xx2(R<3foAC5K^_(c;~wO_kf&
zBFii`D}36fgrP(D<Atp-1S!K&fkY*=`p^2p=e?BLf;|e<%{dH@86jQIvIE5;x`Fmn
zM9Ik$-yk)l%Hou)0Y03z$-Xth<Os8O%x6V7MqoOdU)2nl7i!^FUV$a1Zh=7O2FsIY
zhh~gjy7}d7mNzV|HP)?TNq7r4h;k)w#WH|%JO(ktdr0d8mAmhZI7KE95=r%aJ1Bi?
z#^HwVbPy?-`349yedLN_i?0BRRhD6q{(XGa;fIB8HC}0ZK3!eBnoeInoxcC}i|O01
zemDK_{g=}Z!g)R7#bs_VFh26b75*bK=R=<Y++jH;ea;hfPh@m(-6;;F6^+38(a_eT
zmz?VsPM5MR0j<9+ZNOSyJSU?o`6}H6chTR)U&E-!ex0^K+iXi~*r`K4q*=bQS;HvX
zr2}l!+m%Lr<Eyj_w%L!V1dEU2S?>9YXby<xt<svmhEdk?Sqo&H$k1ip!Dc&iXMFSp
zG`{fKW`VXD(>DIdBOYN3KFB-o)btuQdDrb)?wpV5K}J5%Gk=5EcF8-9w0#dYP5T3n
z4gO#=(%L13ag9c&=Z4yqIP&-+XB@TP<YU_8nObBszi70ZW~}6EfaprvLZibtX8s(d
zrb)%nr&F^+&v=i=Synl<HeCq~gw{i5v~sQ|xu$?Sfh%Eafszk5*9jW=BAb(qGH(yQ
zB$vx(JPDn#j4{E_Auc+W{!WlaMV@?D%J&c;zMP03G2D@7{p2%$ZYS{4H$Jv4a@uae
z0q{=j$US=+1Wd*9out~k^IKA8Xu+2_JDw%=_pwfkey24Z-X=8Tnm_=r6EM;v|1O}i
zX`#6;ZjHi2v-ov&dFJ`k#raF&EIM6j+R9uhV+CD*x0bE%QTe;43s6?yG3!1DV~IAx
zF8|LTAQC{|+RY0eyfwn(N3EAvc@3i%T?8x;@qbq6=l+v0f)}0d<~(II<;ky07O%=y
zI{YO7Ie({cbg-ZQ7(YG|e&`?UGM;(8<GIQ>zq;}(5Kqs}d_#cm?7RW$L^dpbJ%9Of
z`q%&ZUr(=Jy_o*<|MY(|{rXp*E4JLfc>6{*Nn@!^-}yhdu2dJ=;(+hr3jg%=#q{#!
z>*?jG((AM7)v3xlz4Vi1=+i`e2j@P^fy(B-8~0iGp>=*horNbBw)mg0hsQkKbvQkI
zc<hC!!+m5dc=1G8{vV4c(1?K!aoedrv>$h0d<#a9x9FVssWtBNa7?lJ9)-?%zJDbj
z#KseAEPS!Bwy#*|1^UMej*8iB%PSFuzp3^h<sJq1B)DHBJH<iYp_ls(lr>HJ_uft)
z9q&y~RIhzC8)u_lkRxz(9GzkUG7gN*_?$P+XVwco>iRqJEDug{yNxjWxn349;=ra1
zt1;&5Y2j-xeA9lDZrH{H)_mxkD@~cV4Mw|Bi=or`t}?B&L)PHy+<O3J?JDob0%Pw(
zfg@v>n@atbc*lHO0y@7Ax7U625j;<P{mGx}3q6&T17$jVONEV7uBdr(QsA2yWbR_i
zYy&Nsn~Z>5S$Qd2GT$`A3{`ecK$NnO6<(Po*UpO@=H#?f%*MPYw>2U*5ows%5HoUL
zoQ*)M*7D8zT>J=uwhzI8lLNMukMio=u}HxQ+$?s3y?vWynnrhSAH{PH77E?SY7<Rg
z<u!_-Bh`u81lrnD4x0uVC(fRb`Qn2NTe--M_t?jmfPe*ELv%R_+omLZ;)}Yb6>#XX
zxbSlTE5;sEZfn$iby>>dmk^#5ly8?+e!y9nLlT|TIST_WT96z3(um9p0~Ex<jzBn}
zaFZK*%D}9aePqG}pvDBnSr%|{wR!BO<6WKta{2w_2xaCmcrjgY?5t#a?S-040r;lf
zt6W$);Xu>wv}=zK?H!Af-7S6LoVG#@5CY{G+qWwsoAN<Sep1izTFnblEYL8Sa2){q
z%2Pvrk1~bRr>Z;k#%BLP{?L?$GvAW|CIyLs<kvWg<|&?8N4fi%!X+oROd2WIz3K80
zqvPVGv<EIF--w`%7#|Ya3{I8njuc>rlRjR2v@&+kby)bebK9|jQ+C=xaaUa}ViV4U
zgkK@$msi{@q>Yw^Ja9jTul59I%;uL_8IQ!lFmBqu#mJwovV_Vj37H(`mru;XsH30!
zx_awLHx*#a`55usC%pVW7WG3Oe@tE|&N0wNY7alnEAT>~^837c04;o2M>%mR{v02S
zI~Md*zk5u`(x#rU3D3C(U1nrNkmG|zFBY?W{L#_)Qf*k+;?;+=U4G5gi(~kr{orZx
z4K2GIzXX(i%c~EIHlMzXy~`I)KU<;7%yAr9&d-SAcYab-H2O4n#s)qu@rJDVv#s&S
zz}0lNH;7F<0WmYI44Y$4?N1g#sWCoESrC~LPm=lkp&T2M%JBFW?xKaju?<xsw;bi*
z_9KAIc<;3A6NxaJiAK*cExOjRJbdRY(uFOSa9?kuGyAf1@ZlV2e+5ROn_9&;W7j}d
z|KB(R{Sd@4W4XwyWGe`>FYVlh=toa3EVp*mIhMj;A;pJf0^{L>yRB^jGs0H2lFLD6
z!g^={k+dx-e8b?^@`9NrZuw4&#}q!td~kN!yL<v2Np>W%wntdB<fcjZvwqAFhbsBt
zdnuso+qZ=vy+cM`?NMSN(tFdt`d9zz=Y_BL1g=5+V)^Rwbb9^j`SjhlUrt|t{pIxS
z_g_!ne*4w*{MmQY+beGDTuNDA)Vc2I`sMSFjt%n6f#{cvqwbOP<N#p@8h_tZSxUT~
z63_1Rli)-7*#W!!-c+V#BfceBHl*8NZMzLdopmaOIl<YcH>QTh*Yf7?f>CF**+?bR
zZss-2Is*^b;BCR1ZhTB5zhRV*{D$WDJZjW4PA&Kp&y|8Z>o#+tTQ>3l&*zX2_9HZJ
zl<lGeI9ueSjLh_wvetOdn_#!hw69+ft%pqc;FCF-6*hR8dxcI8w0xwQE0#GPB8xb@
zwkhUJqjw9<);#TId?8!cqQXhE$H?;pr0_}ut({ZoBCFvR{nF8-!1Ua1rC%v@*%TwP
z(0DbYZP_d<s`Re$z^Iq==7}2WHtVjEcF9X`L)JFVt+64WQ^}haS$M_dV<ULOY}+>W
zG2)BPtZzR9-N~Hmyq-5L*RhPpC6?$R&vz$;$g6*Qo9$bURVDLCkn8|-<*(v(vO;7|
z@CNw;k<WHpl6Ffn#3l(C>E;1zdK*wi+9E}4mQ7mxbh-&7w@u-zlu@t5+)3ZV`623?
ztIO%0az8C!rs+G%*LPvzKDQJQlkZ>Y+{E``Y%pU1i0>S7ZSZ%~Ve|2pFDStQA^r|8
zUBAT!J$xwf2Bm!0k3($mlY41=6Ucvl@$?t>4JaRd-($FfCw;}qe}SExocM_?7Q*;$
zjz``i!=l^i`Pua1#jEM{+4*#Kb}^lwUrq-HN7FBV^(+5R*Olad^vlnt-~8r}rt^!7
z>D8;V>DBAA>E-KF=PYEwKfAb`E|vIh#yt)ey^!PgRDSWzD;jz2q!-*yPhCgur}%%-
zc^d10c&kmdV<C)%rxQQD_Q3aOc-o97y?C;WH}S9lc7z>sTQwo=Lo>gb6YaV55B<s8
zdyWo|rh~)$zZ%B|eeC|M_Elz?-}2)p#rAy<h=oI*w)41<{$c$c1%DTV5oIto{DU)(
zkv;qIeG%VR%HkL2fy2FNdVKU?dZ@&`Fpqb9xX0=5%BdJRa_5`l5pK&M7QJnGl~cLX
zEL(tfaK0+5;hO%2OetQAze?iWY`>OW=QV8O!wY}A4Y14KHMl{~<rIDmB)D@K1@EiB
z9q_TF592dnPlNl>cgE}>+?sa7-QwKz<Atw({HK5M+wgPH5(uOhwX9p6qyn4^d~+j8
z*5PytSflK?pqnciR*d7LeM8|V{-W>W#k%C;8E1!5?1J;j7jG^WP~6t-hDe$5Z#tQC
zM-jw0XsV(eaf1`LYgRtyO2i^%rhH(@k=aTcH(8_O_Gp_{1|l?qe(2IG6h64lZf39n
zbIgDhtK72B0uyZ}o-$6ZI+f2hoELL}jy`SAO?$u*u|t<nOe{$MQC2$3)>U>H8FJxI
zD_RdH8f3mmSQnly6rAXl@H0O`Tqo_kMlm#USApV1;1?;2DQ5V?1%!o%@K{KIpQ!MZ
zSQ@CmP1zx(Q7I%$^P*SvbDa}J+ki8PVB(NA7O&d@yQ=R4w{=h0q+?b53Nq+sJM*-+
zN|lX>wVOy{p6g6vNj&IF2~;U8m8mxILR4?GCw{R*o%W8o>cA8S?-WRm9B29{07AY0
zAQKG#CmV|w)`HIjg#JQZ8~(+|3pi^v-5!=Hp@0s7(9iCBVpHCWCZ1^6FOvgwx}4~E
zu?ahT4}S86B7N*aR5hyP-?yut#Nl7-NY{%V(sdt_<g1S%Kgm^b`y)L+Qm@!(FXG8$
zh@7_J*F-E3Wh}Btm1C=X+Xu0+;F!>O-RqTn`PW?1fp?|JHUG<Uc6O@CAAdGm<4iTc
zo=K~p{zJQLJcKK5=lT#Eg`wMRZVl^I-z%}W$1%vS+}xJ5i+FBZ)fvE?dhyCawXps4
zq^$eFuO}oB>aDWfNA#Q9F*;rx6zv5>P57{7Y*_BaHTpG+Yu)Y?ZrXzod-_<Z%Iv-m
zj!ohUN`3IO$XvHLcJ=VP$Y_}ygDoewta6fDz&xbk5uVF4VirT%NRH>?u@m_GiLSt`
zpoL?D=j44&v67v(dg$N_%;(>b@s*GZFIo8-VP;v`A9*W8!X*-TN~Ire<>LFAVMtjm
z54Omfj(){ed%zk+I-7Tu@`=lZ6==g24mypeox{p5&@^FnA2i<>NVi8mfkrIXr^gqX
z&U5U6#%0NjW$zN4u>;1u{WLec<TX}=j=tEN+0eMnMt+g=_+UrI@ng1GX^)mI`w|9+
zm~mHBV1-JE$%pm`S81)E>|g!m&rf{)6yRFYzh>q?!FUbg)9=2TzWL^h>B}#FKRti`
zRPsyzx|wSWzohjI0_KqJl4YEMr<Z!{76_qVe{|C3v9&E3(AW+5aK0l+VY9r_U&G8N
zzd`4wv<-f&>?UZt*`6GJrJG>3c@G%Nnr+E<C5w{-BR<0D^Cq2LfGpR3=?r>coq+Z~
z`WoqLGViqSRItrv<N@1wz{t0(=5p-Ji&Eh*BQuY!WPycl9$6>kP0V`MX?-g#GV^xR
z%)^A3M;^IS;Ux)f*@7oob%vSWc(80OQwo%9=8rt^oR!vgGaeG{t#T)w^THhL<aX4t
zxbS(*6%QI?@4=f^Zj^JTaT6Q)qO;H=9^Zp*E}uRNhsU*=iEHDFUdqKb3COj*ZHrg9
z^DqWJGh}G+K`)Ahm;OM9hjK}Z?s&&!4cm0Sq_zz{%CRln+otWJYt(&Q=py&97H-;v
zQ;E~|g|}Zw9lCJlH2WLs#CH_*=CNU4B&!<AtYk|bsPdHN@zy%X2KmAVn$Pzpv?6n|
zxaA9vJsZ^$t7P;GDH=)Jq2nXbs7t&{>`rWEvrE1`WlUIUN0Qpg271kVzZBnqzHjS$
z`Mr0S%Jbxvzps~nzQ_7Uo}3qfE?cx^d%1Vf-Z%~!3--Y{L{`ad9wmY3d2HpJ;rkBz
zd78^leQB<SAJq76@9*&iQuq7~K!?SzD;B|cn(FduIzQ)WuB++w>vJzWvAD-;1Rp&*
znI1oWsPE~#icr4&L)yu233&DDbb9{c<@B7Vv|gPlmi_5hzxbunXVXUyPo^hNKJq=1
zi;FkY>+{R$h4h}kdOf{Rq8t{#SjeJc#CNH9&Mz;N-Uw$wF84NQ$~SM`c~R`_l)rs=
z<F;Z^$xnG5=7+TzlL5Y$GgqaK_{|e>j7RQ=-1BP?Su~@L+>4-DiRDVMu=qt=C#4<9
zxsT%eQEGGFXK)`R$DVs^x!;o?JhSk&uQKR+?yInH$h|`JdCSfI(P52;zAwTe7k!$A
zx?Zef0gU@z+$%a#H?Zi%0$zTMu4M4w2yJ%773Pe0&asDHPHmeq06g?4TgImCXFkY+
z&pg2{9N3g8xvVO8txH`8xXBmLV|^Q}%S{e)H=AwH@|*OK-Q;Cm$2Hh~<^QJg?g~9t
zw)y|5F#6r*yGHGgf*%uWJ@NHtzis2}>jXaE)!<XOc$)=T;(dQP1@9Yp(B{ad?i1-4
z<795d=yPI0jwFPb5wd58DW@~DCHDdnXC#Ag-`vT8Xe;E#vF!s?wE3}V)Ww;TazHu^
zh~^VtI7xx>DR4Mdra9mhm%*}ku0MD(c-cZn6?pfq+M{J*TZcL!!k$yRb7(eBypA}4
z9s2CK(cwo{MdhgqW929IzOf-^B=}r#5OROq+Stg4e<|V{?gCioi%(lS7L>Ikcf7>K
z!G+8?3M{tX)5RhGf7JCbNOAnN1vkO_hCIA1u%u7iS7nE9E>6TrSy?ti=pa2Az`hOo
zruhP;TaZ2=uN|MT5(jSEHyCxBOQdZ<Qh%G555PQ#W;rf_>Dfawlv*FIw&DTa8>hW2
za2{!RTzpH@s_qZRMVzHyZHfZD%dy6yk8SNXb)Ju@`_%^Ogxq9ha*#H4TU5e#^SsH%
zJf#Jf-x`;_K*-`+CK@bcQZBmqK~8&%XZ@7XY4+hN6Bk()b!d<Hk#3d8L`4)8kyJX2
zZTrdOBFA#|kHC-FYx1+Vr}jaINsfFl*|Dr9zWSKwC!p-N+7E%sb2~buH_;b;^@hj#
z`E?MJ==;;@>8t7d^tC5t#OTE?CQk$++pcdfSW-nVXo>GSqUW45gL3&KW@HgjH~1#q
zM$1B8+8_&7cH^>G#AT5WBkBX-*Z(o5x^hf1nPWjKzgWt|&SME%CHmBI+VOz@n@eo|
za&SKrS(2Y#b9_uF8H*%jDoePcwVf9*sjYO=_QkLAOyftfF5|_-KvPM7jHe#0kG*KA
zxT)HPxI=2^j}k}-qOt|L*25ZGaCqjLCYI=EufiP;3_-4w3>Ir^7o5FS=yLO++|%B`
zfN%pIcLwyMZ3@22(PrFaq_eEG-3kWij%@MWP&BrTHEq6z&veV2IrR!Zs7VO5hq|x}
za)vK!%=T#^<0W?+o6ZKbZXrH5%x#^W^~gL1LJ6r__Q1dv+T5i*5+=Io{f@Ri1+tYt
z>(2bvYpt|bq`tyvi)Xio*og?h-$EmcEqS<ZA7U2ic0n#8)tR-@hbCe_=gug{2F1W1
z%Gk@+`;t|A0hE!LL4f4*UVOKQ@_2v;8$O7*#?z*M`IkRG@%7<A{ltxs^V8||v!~Ou
z@4lYCSNi75-%a0s`M1;Smp@EzFVCkpN?Z^9KLzTCT=!)Tjt@}N?8uY-XpnvP|2j3?
zR>U^jPRr<AV3d!r!EagPP5(%kuT5M0tYwXF)7=JWjW%s(ex-)YjTilGF#6f1vzD)u
zcG2@ba7||m+{9m}h7JA}p6N=!$OpFKo>K(3evqW%(PxdfUE|w$Yk*@JeJ_0WM*^*L
zAagM(3Xh-JpB0mk<vdKD^KX|q>*pYLvd>W_fERqzcGmTNfpM<ILFV=<rIrDm1M#=W
zfXundSEcZmyyXkD9@mR%qgCA7(uS<@x!u7dD~9omgT>bE#D;C@wA}5^)gFH7h-n#p
z8r?wbOQ+}_1HddJAL+*1zxn2dUvz2Cbz(z6eBGzrer^Mw7SOdIHVS%)DKhxja~fI7
z!-%}E-_0`{$p(*XOySYT&5LfAZ5rJCqRp=auGHTtsWT8im0P2EW?C<*NS85)s`V{H
zHL-*jojF;l>us@98g$Sp!EzW{$#OIN2_Dth&h20N$Q?`3b8b5&Rj7HfLE_H!SiF9$
zDUnxNWc|Hr<XumI^GeA|wvnuwyyQ!woH3ztbJ^vnQrh~aWRp7ry0jTH9(=#$$uFLe
ze|w?R>{WiqTJu;o4Dx*^$CiKOJ<H_5_oTQ}wFRU%j`Q5ZlTBTx?T_!h_#qZ#y@|;E
zG8S5-Y!j90^5s*s{|j96J{FQ#WCA}wJD*;^emy;V{%m^j;-wd$Sco~+oR24|9FMER
z+ivooW3OLn-gwGWT|ANXZu;oSlj+GPPo|UO6DJmKFk<n^i)?2X(r4l8Qu<fEXT%~8
z_e%U)z^l~7KO|P!{U2ado2ScOsVu*;Q4H#mAIk1cygi0{H-76=J;ipa`edQ)%ukI`
zN9xW(5>I^{9^zYVQg0*Szsz#4k(7HaUR0DXaJN0}2~<C@r=2A>)w#m=y&>wL-kL71
z_|G&J8ZX`69Pix2K!;AEkNN8d#s;rDJmmF^hiZ3T&&Wd9zGB|j{CkgkV;(`a16dyv
zzBl6bfnLT?1Gvj4M~~w_9%sGe_#%)02IJ-5v~`SK*DDxtU{hvGuxvH%8*cIgT=TU>
z2Uzo=yM07ez6ra0HEfo>Ti0b3z6~tBL7{!yPo)n5AIi@*_`Bvi3txZw=f9<~HKcv;
z8t_HZQQ?jN+CP&Q_!P)wT6sj_aq57BFHZjE+j03vnKb*x4gsf$aN?J|=-!ti8MwEB
zhg!01!CY!^+Zc7SABkJT&^n+>9dn9lNn#!M_Jsg@<2jJ9w?}<4`ImOJpJip38)^P1
zv5%#x)^#tzgDA{)j<TTa_;4cQfJRYlii(^I1qv|gf{UEoC|D?b_#(;)j?H$u&U10$
zXAvsfzKFvY0c+WZb%{YTPm;=y@j6Me&%%l>ioROl13hV1OsG;1UvTg&znqB4sh$_}
zC`vqH$%R9FUAPpc0XeQ+7ooh}%T0rTgbYE{9BcR(MlW)e@7T)*+A8Sxg&{KZ<~UIS
zqPz*z6`hz#H~PklFHh$62neq@v<-TqW^2>om{b<u1xC_d=#uwXiHPyb({z%tn9-9R
zHG=CUoP{pi;aIxmO6^~hF>%q$uVC<x+schcw`1T!*H8Z46Q5t=s9-PB_`(Z`Uet|e
z@ni4_kMba7klwCVHLDl0P)463L|cpZO$$LymGY|}t1j@FNc0%vG4=<Vd@zpoJt@ng
zj2Cgd&_loANA)OK9H-4l{8Eg`9Xlpe0Qu#`g_9?Me&Xv)6Qv7DGa2&&mp=5l+d2h>
zFCACGeUf_8=1k<wQXP<I@|!UYo;IjH7T<N|xa8QCENu`xUwBAq58g@b1xrsnSuB$s
zi(U`*{Vr=3vZQ2tw}WyKUeKI>J;lV8IA}XwZzx|mFW@ibW$`Wk{gk2oRvjH%$1h!Q
z>Q9}KyKeBvD8K4p39|Ns+#CZj!?Pn~-VM62i=QtymA6UJ=2*PQtZSTO$3o-e>Bsg3
zrp?OA<6;J6;g14%g&Bp2j@ipD(5P}WIPHan4eeCmK_n;!O&JReZ0Yi%P)hKNv<7sY
zv%`cNh<kW7j<s)T`^c{BZ98k!UO7T1F{<5+<2Qy%#WE%+zG)h}1aRy^iKsj#5IERG
zY=%;0k!$Zh3tBt;DxU>`k|Z`{b=@(57D>tTCSKIEBc_#({b`FF74azET7s%$ikmen
zH17nZ&_-Q|<o2hH_Gp+<a+3Qw@gU^5>ZkdjInQQ{7q1}E(AwnaHZLUU=Y_8i3Du|u
z^~JlZ>CMHf>Gg|mrysumdiwtBFQ%`)_}%pFH(yStuU_iQ``L7Hey($}u7xau>HO~-
z4w%Xfeag7#dE@N*1H%~gj{!#AP2jQX@wo{uw`8*pmA?Vj`dp)om%L%EH^Mf0o$dnH
z?26yeey-^i|EB;R><5311B35f<u$!a-mpuSdBFl}KkSW(cK;MU%fXEeP<qy>G~<Bs
z%p*eE$106{<25#nuk(t4Ju&tirR~Sm;CULfFBp{ue3l!EZxa5%tF);f=<z&4DDC0v
zGhQniu-Rt2?yEIyv&+1c*mNvzlt;eA?dzcDiIK-gWbk~}Dw#oOe3kx2{mX8oYtLM{
zZz@Wu#p08te)dzcpz+uK@zXfaGV1Q{#K_>8wuWP|9tHI^*!AgziFk^EcCe1^EHZQ*
z5p$wUd1(ju2(X#`E2TXNV&PdA&Omotx<5+R{XOcD;tPIcf$)+mnP<w|24JVH`#!AT
z{oNwXh|CS>x{v)Sp|d#&eE8T8-%DtVv=M#YiE{GIYtxu}7EQaDN^_fc*_A`53~i;;
zUEeU)Zv^+3>sh{BSmiypvn>2wkSlrX3A?5)2Fx*0PkowB45V)91w0RIy<4{1U~p9_
z_1~4s>P~4H|IpNeIq=nVpM@`;0DF@}2cdzU$AH>8<0`TASn0)-K9<M0T#j8N52Q>s
zX}Apcz5K;b{9)qxHGaziP|2~A0wUi|;9dmx8(54Hllu%jjrHR7Yd_7!q7miz6@CW?
z{?RJ;1pFOco9|or$+0)n>G|b!c5&tC{9{+e#rOO}UQH-l-r{oghX44=lUOW-vGDcq
z<k(M_(I!{&<3%m`XHn=#wueUtzMsH-2FkgR9eTt{&g%sIv=nXZnx=B-<u!|EUIYWb
zR9$%*>k(z19QpsacwHiWL>u&?7xxhRm5}0;?9I%7SM<Dq$dhF(l9{FrG&WR!?u*d!
zyw;J$zH_w??Mrewv?1-DH940DaBktAlIyN?cqshA+iCCKo9Vvl=278}&ben9H^cEh
z<WAs9lmHb@>518TGY^g8jWS@tT9-Vn4{YeJ@R0@}#Rbc@wjI>hw!ntI#=tJvX4`GI
z3vQADEx*mj&0Yd`E$c(VAHrwDhxYAutn_1`p7{EU-^N}$E`siwn)nqP-~gr|PL#7c
zyBwdwW`s|2U);;U{6g5{%M=3P&Z>-P$UYjxm7616#GS>9ZQ4(86I=v*!dEWbb-}k4
zDKSnuFSe;g);pcJgrH+a#9}YFZH<S~zOac}mszn!&BD37GIET>Z)SAU42psafzBz`
zwcIrW3F$rM+Q1i`=qmZ6%?|VtzfbTmd}Cb6lRe6(q4jau@+WlmOHioUpWTH<>Ze8A
z_S}TG#FkvlkmG~LjBQYI7SZ~G5yxsL;sV76nv=ZpK0y(ell&4YuTkTsI;UJ+M_jMo
zx*TwFPdJq3TKU403%-!zCVu+9IRGk|_|(Z=4EoI4@Xbz1SuDUq+mmA=seK4nB9g=c
zFNsR|_@Q!bM-~I?qh?9Dj=1bg2+sz92rrTd$&KeoiFwAA_@6kpjtJGj?p;YTa`C-L
z(f!Crak8Lfz8gV$G!|pjI<z|uj5<djNPkp3$}>?9O{sa^<vUMm`~(oj*r|Qt`AHwz
z+2zX~zG&K+$%*5keDN838AF7Sn7U5v=rhI=W%ncNp5STnl~?j*@r6l}{m9A-I9{-!
z?uo~7RJ(c+2e}xYl-8t%@$Nr;W&CjrXhL-Q`gA%wKh^Od{Pc8s`QpWN&L7FDF0DiE
zIukba$FHMDjkY7^ee*J2((ur<1yK`1#vwe~jkuMlw->2csFRTX!=KM3v<>xOL5e<8
za^3i)Rwl&z{2v!>CBA>9#YC6MvHW&ivEzxanh3kB$-jys(Z{q|7Rl)Uj3F;35~rBd
z)jo--{Q0$;@Q|wg0WW?-Vyhe*QSyT=7<R&2-dpPyk1}&?g|=__Vld33m|k!usr4g?
z7@n!r$AB?-W!V9gwfIPB!B9!kj@Zq`LR$0}HYA55ZXfGbAS)c67>VDgOYyM}U244t
zY$7`T@T9zApggb?3#T0iqLn;;0Q6e|2rHrY?Z<ie3xrP01=N4gODw?+5HNj6P<Lqy
zch6Waa>uW{RSg@ytrR_TQ*Y~TsIarDAiD9o32g#qE$xVlzROr<tclUVxXo22dh!=a
zn2Vz>$PjkSBp@=!gK<fFoMG=`kwi(Uaj1-s_`??&C8Cprtjp{;ENdFc{jx31O^AL3
z=rezarfY*xPJUka`j9X;cmcgcBGVW8%h!JD>xZY`PG5ihx6|+b=HE=;e)YTQ{Pd;L
ztLZ{t;17LsV?TZDu_cddw;vD~4?Qj$px5a}*rIP6(g4j{>y4@P?v!-B5i%#N<Ka5A
zO~dGC8;tUs(gtjice)L>+1!?FSGJsVX4%T0^#*zyZ%waLL*r|Cr%|Te5?}E~*{Iuk
zmVp;-y!n+{*7Ant-NYv?-yL@A8Xps5%}|`de(}-TfX?w;Uo2bFqQ9BY4ZaEJ`V3-m
zZ0n372)&fE3C!03WS#u|Opv+Ub-PVIxw5{-iCE%nI{Li`G1j^B@X;e4Xv*1?_ji>p
z%X8?VUgVwfW5JY*xh>F@tyyK~#0JlFzRw`{92bFow|t~G+h>ip{hS;~a3M++;~b0d
z;&0fAPltAWNM=ZPqYV7OZ5%ytZ=459wj*h$&r-(B$9ilpr}nk!({rZ*)^^%e9y{U>
ztjkCo^N&77OI)g>UYO3yXLw1YPwk2g-Qz4g4|%cDAGB4;=0shHsqNU;HkL6%?B?;w
zn7U~jeMq;D-*_32nX!QFu^oIYDI>*`P1~lAS`N*<icK35q8N+1VJo*z^83NIv^PlR
z*ChHUC@U8%gz&?mw^!%FFZ3Ohr@?qNV&3Y)k2$Hk-<l#LS4RxLLCGC{8*sO&_1@Wz
z@@Y!$C9vp19=-SnaNR?wc|d55>AGiz6~MhU&W#rrSNX54^9z3$$G}fv9p+Vh$9(VK
zSDj?fVik)@XP1}L%U7q<%a^Cqt5@fK)gi5y@BD{yk#7;UQLM!9@Z@Ovi2oEjIYy$g
z`IqywIzKe#$t)JW4i71hdj;H|k)Z;<y2`x~>dhN>cn#qG9(H-fAWvGIo}N#y`Jb@*
zPcrI8EKfdqI6Xc&m>wSI2h2e3ho%L&SK|AL=5enkKm4bxEBW>73)SZSp)w0z@PuFT
z*A$|0KwSai<Vi0U<47!mot@D>{Eu1ua4(c{+Q@&~pbp+U*z;S3?!RM!RcY>D&PQ}2
zBLL2@hab#%(|vw3&1cqIaF6q{3oQ7oGt#BD=-&-*hzs~*nE}(q+HM<O^cLL_kd3yX
zIR+Pc4Qrb<+$L{8<2FAHxB0tEdA9*R#w1Pz9{{fV{;=><eX{U%@Bj2a{?A<b3i?Rv
z!xVvVogm0njcABOg~OR3I5D6RHynf<kOj6~V2)WiClLD-?nN%k5Lly7l`-mtEj)V0
z(be9EDeQ`u<P$8u0ST!EdVMO2Cf>tyy+ow0NCId(cAKFPEAz{yB(^fSINJor)BIT*
zG@t<A@hAN*BPZX~jf+J%?1Nh^zU0^KBH%*k6QN|Ve51UqD1MnN$&1@1PPk$3uCN11
zeZtSl9&=9YM(q@aE&vju|5hjUH~bocI5=@rhBh93#sL=(epTWy*uj&YANj2lJ(D}|
zHw-G5#9ngwpuUvL`IvaP@FW_FstXo8m0@kwFOPa#A|sVav5>?l-NcHNU3^hB?B&&X
z1stA~IoTuaapir`Mdm(Iplxs6QZgV&*#KEe2{he3<GMG$##QT}K~LtWqs?Q}Fdthn
zj?_Fct4%TzQVM<?znxHrv{RRb0jAKN2>D~>`9c_WWGH&_3t9YbJNlDfwIB$e2qGtW
z%&4DmLKh>IxF$=90|qLLCCZ>LW6vali70*VkNLsl6LXJsc#?Sm#Cnp;`NqPU_*Mic
zTN|c%5{bniO=jVi4|K`>sJkD*=N-@dk?ULY_>~e1Tt22WDdwHpOx|9fz1C#twI*kp
z5Gh?<@(P8-&E!ZmBXs-Y3k-><U}f<be%|}-<J7@m!$dGoS#hJ1i7eoSBgNxmPjnW<
zvtY=iiv=ha_1xZZW?%eD%THuchWNAt|NE5!X$Qtb7N+pYxxvQ-AL)6Ki?Uq@=~12%
zGWh<L6m?deEVq6h-*;cJRsFCdM)$9HJq`g-NvC(*Y-fB~UjpF@1q7{}JF$nF?eXW&
z3hcn;aN59%u>is$<>NvSH|ZaQlu!Q=J3Ld;@glcAmZ5{46w2ITndO5L`-Wg{kA-QC
zoF|GA>!1R+q4M_W@~z%_an|u@SG>hXpm?Rz8rJGI0Nt>I6JD_F7r>uBL;Uo|95WR)
z63blgH^J>?b~#v-jeB1^_eIzn;IR}V)4*b9QgrN#usm%hq<u%!TFz4w60mTh)0zP!
zjMH4(+xMzH3anwb<y%={1x>}5J<RPFnVZ+_93O56;hck(W2EiDgcFqfz!_<)p@X8K
z&b;;m5`kTM2gJy}S}&9(V4Pw9=|B7L%KN_)=-~dI0!=wKXH{4IL#w-I#oB0udwbLI
z@uTVB_~YrxXTP5Q=pX!}>B%RbPoI7E%ju&}KAZLr4w*&xTEaOpeb+m|U3zb$z)jd@
zzOKPqC&%>N;F_jvZ?b+L-#ejvc)!EnUCMc{dfuYDmVS_W3hv}*(79{5+w^{l?KWsF
z=PMh0_<ALLV?3^X5A?2bxAB|jGKBZ}zxg8%fVOPKYiK-Y*T5pbuD1sGo-Y@4G2vHR
zlr4M2eLa}@l@B4-ynj1fw~JxZ0Qe?7=+C=NXq%3?)EyV!MLee}2wj8#8i3_%C09DL
zv@3tz0Ay*irEQ8^XvJRsXIpS3*xSk<X<+0Tmvz2l<0TIJ$6*Vu?+`Q8HMirDd;v>p
zkFCE&n;yXyE>HE9FlEE5Z0KiNZ4Jb<*lo7U0^zs4Z2;+cjt@DPBj(D!L+UYSPpfho
zO&p6J{_pvIjadd%mRM46$X&4~Auk<3dTE=b51o~Z9zWKbT`acI3E4i|Pe^v1wgGLB
zdN#BsZpzh8nCr0>u9rLN4~LmyDJz1NcnjRDaa-)SfxXKKx!V+kwxSL&kkNE}55KSP
z{+cV@yMH*{KX^Fp9etGh0VwI?r-QCo2+86{kDG14eS$Uc#*~ADqv=3Ck-n|_ZC)(l
z-h>iIF~|Qn4svdggfRHB&a9g6;;s<3`?dw=dpti<J$ZOE9iJRdC&!1=!-vPy$<dL&
z^YcX5_donFegE{?^z!u73q$xhQfv>U^YNqO>En+cO^+WwoE|=Uq;t%D>pgw;a(emd
z)x;ZZo;*ID{`gnFn11=$CoYrM6aMh>wHK<m*Ko|D*U^E$X!kjTr^%?t`RRrK5`xe_
zd-Bor>63^4FD%-N|1JCe`RnQ1A6`w*U!BSCr5Bo5H2d}EPp1F;AN_jzkN)_x>7z%7
zeih=k-@llC|Mk=9i|?LK-#$B?o;|;qp1*h{*_nT+O!shaf;RCvuiDz(uZ?pEIk6n=
z@2ic}RwoaqlaphgHCS|etva1v@Ly&5Pd1df$HCcz`y>GOaQuVf!-MHa?R|25G##ot
z9$dYcri&LkMx^Jz#Bm4Oo(SkY+An<?@ck?FjE`4((Y|pJexSz$G&uRDeUGEQ&$8(w
zIM6o0F1!JjBO}(FwF|@>$S$*C?bnfQsu<Q~hTfuBur0<BZj+DTl5WuN_H*#JxP6~;
zcfoZZKiy};hluYc_^JN?`+xCY-+S<n|1|&cg{w~82Hfz<25V-ja|#E9T1Wg|bTQ9r
zoG?w!$^Gg<%+PMmIIiIp$cd?$rE=tut?#L1n8OJqwDw(4yvj;4(hJA-ZjU~JRD|d?
zQkmOFxi!sBeIq0%3K?~~2-4P^6l2@EL92ZHx>Rz6Fq2!)d*OA$%w#;aB?)35B-Dy0
zk_kCFoIvavoLEKpqHD9*L8K+5!$n4CcTRdp=||^o2MG9OCtrOdyE?^QfFi9!`6MnP
zXtam4&l1n_3#rJ6h1^j5E<cMF+|U;v^9l(iqv+W!n{q2f?h_DkdlTE^jsC(A&;+8&
zGHd2nD|sMOF0G8;OpF6IRDwe;4zZMT_hb39rgqjcFY%aPIFO|ZN9epgLm`XrBv3Fz
z0#zREgfL6ZrWCiaadsrQW~2?>4tW|UiykQ-IehEeIb2y5KRzPM7=aJV*2amQ9^!hA
zf9tZb;NpcCe$|twgOGUvg-M=t{X~%_4*Y`0izcGIh$Wpo!NsqF_@h$UE8z#6$KRPO
z@(Uv-9eK)2seQ;FkLu$Ky*vR$EE&UI^dc6;&ro8=2IN%&EX4BbAQpnSAqf6vy3{0y
zcU?2-VWEpfuPc7}Bl^;lEwwYgamvSd5>KJ+?eY47JZ;5OSiBFL_gzy5ULWWu&J-_C
zU>%*D%rTj78M@Kq^#b<6ud_7S^rU+KF#auLG8TTG=u+9$Ry^^=gxOE#(H=_o{eN9q
z6nmg{^}-fyC!KyhBIRchjCQ7<k^n}S>m}M_Kss)4`xXsPc=O!;B<m*zi@f8boq8yJ
zG!MKw6h<AiLWiSm#+w~pXhthUjtR;$ZCPx*$Yg`cJxJoCY6@A$pGB~<lTx$;_RY9}
z52|FJf@sFHSc2l&-{NNhtl;|SoPEQhtZ*X_6It<KnMYPK11;Eg0J*&jlSE`<P6_SL
z=Guj)R6$3T9TskZ=om^*O0DMRwPet#GP>LZ5w<xY4V?o%*I@(Gf|TD{S(Gd)tl8Q9
z%3X5lS;w)aV8_`0?Hi=qrEKLV+`yuam&7+;M-aL_O2;w_LL0!a+~k7BP4uH7%p}~R
zy`dOC(T@z?Y-g>-hZj?uAZuJH{`h#8B=E5)S@VB>;_E{L{nM$yO`Q&a8?522u2pB#
z`RR-4+0*Z*ufO=4>5Je0)%5K9ucy~9o=&GPpG}jlT>#-g9>2Us#=i*0gky-U`W@i7
z;yezF{0?KYOdx`?7^5}bsqIGl4ga?Qnq#^Ra_lYXM}h1-XXCgn=$j1h6??@nrn|s4
z8}?(8;F{i;?h0%Dtv<jj-PQ9pKWluO4zT8pdSkMzk{IC?zrUM=8ld#r-uji`Y{&|l
zx0c<N2aG)SEj7NjDFt!71;Rundd+WOt`#4PJUV?`H#8MrBVKLROldZq?-=N`Y=OD0
zx5)4@#)V2!@SG(C##0V}&xY7j#%ODU8NW`OFwePCU+$U?8vUMEkV+t)@QPgwUoZ20
zec;_d@8X$lzV@n@saZxGBb^kT*58!PH$KS*et)O6U{26VJ~>Ora(x&hqOpU~K00Z~
zCsQdQ_Dzm`IrYBeB)*IJ`&iML8@1AG=9F-vXE8haWEfkW8m{r3n7b#g8nLsk`@i}$
zx|NVSu?5C*yU(M$;K^ZCZZ!N+mU-5490Y`0VHtOGnx(<T9Aujtn%h%ACL3uJU~Hz8
zBdnC$j+8cC(uQzn+OF62{ITHp$`85wbq-8j?0XZ@Q<kz=q~JA(ENZ;F(p*;G>8JWp
zE@El%V@a>q%3q(0au;dY7?%OYS;kt<Pa1oSEyt3@7Z$+q<72_!+oi|Yazt}viL-q4
z4R{ug?W1ysrNljpD;DW4F1=94dFtrs*o!*HhnnNP<7uXTdMQsZv2gYB<*Vu0izF7b
z_(8G)a$J;2dFQgjFXcRZ^l*Ck@Zt36;j#aR>xieFG>7yPUc{n$K|lHE@$}gzA5Tv{
zMsIK0d%$AT{ppj(52sI`oJ<D~cvDb*(9iKn`|?A3VtW1hbb9sr_4NAeYU2LM<44ES
zM-PvtlcRmN%$54e$Fb_jeHos_^1>v2K#+Vv5)KW^Mn%(RMaq^Z;nZ&OL2tVa)i$&v
z{~KmoIrI;XrI$8W!Vkm6G3bL&jiUz|NB5PxO4@vmum@`<WdK&+hMSyWlH9WD@0LX$
ztj1LFh%vd!Q0_8iYdb#V*4@0O?82coa^pRol{>lG9@nuBAa7{>#ziQZ+<?($gs^wX
zw(}cqlQ*DslkP0DV$hr5uI0WD{8XRVcltxaPc373;>$xukfFzba}HuIhg2{)Z%ZF)
z*0btj!|>0b#8x~JNUTZ&iRN@_k9<UP<H$BJ(04hM0JOwXL9F2!lyRe<3w=jxW<F(2
z&K@GVa2D;74ZLtgSv9_0Hy_}MAD(j;BAg3|c)*5lpFLs~9{8dW?NJ73Uu{ak;|(-J
zoruU&9hHO%xB&!S;_4E(0CJ-lPlUxyeI&jRmoa<mqu{fwQlFHyQEr~H5npm}<UH<-
zC9#6_Nt}x!I2XAr49QHzzC}-G_$&tKOfSFjt&0Nl1$+@r@o?(85DxLW4CRzz3dqeA
zFD}S00$;xUVV8Q_w-A0L1Nfq#PcCNkl!yZNG<u9v3dfH%@JOAADLlk(Hv>8C3rN$l
zqLjzNln~ikma;8FnKlAPL|+zdh7qf1SA?xc5WjrS{PJs53CoU+{_=5Z39aPTEC11U
zeAde^U|6uY;$p6gI=sXjIZrS;Vi9;k91?E)6eDe;L`K<6SknLaN&B@v3o?x5D;AE-
zvzql%h$b2=Ho=qJkDgNo9LBVq`2y}uK(J&=@g@-Bac=!E@;_Gesu+rIe!>S!RUluy
zvS0@*V~JlUX;P>O9uq(0`gmdM{Or_EOup5mn5Vq>!&sgU<9N6@JM(KFI7T^s`qxw}
z91#l(N-S>m*aKK}o{vN6y6*B#+{cGU@{@^I_|ZW|rD$Jv96x@op(pZLM53Pb1q*<A
zij?}nOCNa>EAzbWfqL8JETcV`c)nrcCf~dckp)U@ypTwH>BH*}Ii%h0Dwo$!(xxt*
zSf$U-b;CEL{5zK(It~C1dfGu5PX$ZgT*2bUl-RZwf`o~fn|B5Xj%fv{+zVQ;4V>>p
z6~kuQIV9snAeePzFB|93+WEX{wNjMb?8Cg?BSnB;`jgrMrQ@=Kr^b{;<W-j2Bgcgc
zF2G`@D&~=AoS5Z|p0PJR;?2jn_Bqo_8v{Us91kGSqt^oaL9Cti#Ur%<Aj3@cE#Y-2
zThy+D1a9NB-f~`oN6C`clh<Y|S#Ah5@`6JNQ=!$mh!#Ejrmc%P%Y7udrE<PETTqJ5
zg0oLpU;uLAea>?~hM$am=~6G*CGv<`-4R#a2Mv9&4P|A(B)i2gdIM+Xp||V<IbKAU
zA7>4iX=4<-Pf(j<jwiBz_D}!$;_wd-8zKG-ft+XQpKczVryuZC@zHd0^6~VGU;SqK
z<*$D;ee~())2E;R(e&u!PkfVs-e<VligDt8RU(7*xQTE2y+>n!qXf7v86nn7z73YR
zx5FsCO?M6Nt3QHm?`Z#1$^kxfz1GrE-+hq-dS}0_VWYeqW#1apE%JuD>WwhkfQ{pA
z!<+QBdB0ZX=x;|n@cY<nxJh>xjIvQ{<`2wGo_0g=5e>=))fa(}Z<DhROgE!nWShmX
zuXZauuvpJ=4X~RRKHo*VH@w#=B{bI=&1R5`>_xCm+vE+q>BYwL*ybR&PU_J5Yd>hi
z%2`gE%@U<`hX(ZbwZa?L{z7BF=x*vzUUV9uQ(|D$pV18;ZDYjw5*KYc8~M8omGV45
zz~0~6wt)LXx=8IqpZ?ATq>VP3gyPv3zVtUR)&n-*UIK}6LtbEhFz*4%^=6hR=M1Kf
zL=464{#ncn_Qq7O1k06dsYT*Q8NeW^c-khA@`2J<hRxhoEf>?}mH!1socipNZHH*k
z7sO{`fSc`?vbO1YW*qd+u=v@Gg}%sm`7z-800Xq|a7C=n%OlNqvv>M#%a29x?;TFl
z{;|GaA5L$k2l`&D@Ag-@)8fS@jalDQGqjJcNS~L4>3cIjSmkXl`~VfZt4n@Fn{y6t
zRypJcs(xb%Z&6V>`VMaYev?vsa(^NB2-P-<Id7stT`}-oS#j_0A5QxRDUTnz^4h^q
zo;;a;^@~rYk4_H!A6j31{r&XycTcBRuTOoCAx^N*Jq*&};o<b?@nb*9l>3m=#D9cw
z-{43-Kl|+C=~JE%d-PECLe86`vS4(sy7B5o;^S$kCm$XAX{`rplQ;Y~*yZ{3_;`Q%
z?9-F!{@csx_3QKL<=NHr?TfSN>+fGpUq5|0ee?ax>Dh~ye!`3U55M`<C)0oVpZ<FK
zFaDE1ntt)|$@KE{Z2J8-&!;cGeJ&g6%jWwZUU=c|@`~e|HsbogF`~Ar?_T{zn!e`*
zaDOUq+S!|qjufZbje7FxMP4IGy|CqeO`0kTbm{Z_NPAy>d~|SM_1c?G_TEm14>V%*
zJ(A;W`L<6CT)jjq_x+rb^<05GnD83M<2)ibeU&zkXpsZx<bKg&-^Xt3!fV)udk^k*
zT;!>?d~EU4u*)|wtwNn)yC3!`XARr*N4QNs0-87JTXuJM?^X6)q5JK}!XF~0UGP)O
z;)$<&|HJ?IKfjBk9DIFoCbN6(7}LIxRqq%FWhaBa`Q<}Oy2i=j*u+ebn4pGFAz2(i
zHuyB4r~}&u@g0$Nb&!==KsX0{>*q#g3!^s6(4)MF;wGYigBIaW;Rck@iw7We4d2CS
z{CL5qVvY_wRM5jIibP;Dm*L!ExK=l>5pxw6bu_E_AS^0ac>sK&ORU)GVnY}ELMj<I
z%YEX)T)Uh!@Z(mL%mQ6Jv6o1?){zW-{5rv5C*VTsi=--|OR9ZagK{P37YkRq5PA~k
zddjE5@{L#7d6CM#rH?ObAkRb~{wRZsv@YNhww@44q(iK3@U#;PYPMIxe_s?x?{slu
z5h0Upl55TxltT3(YPce?p<q(AL*$k&d6P3>BPM(RR4k%)S;ey&9FLE%0nnwd)n3%a
zl9)vSq|CIV&D{n*m(gx0;2T;T(e6dj^D7!iwgrOpg@pE_f4a5?1qGh#J^hI8J@xM!
zaU~P&s^9~AbI~=k>=mG$EbW0lLzRWHj4#F}iypT0f=Pb)B0_`5PZg@HO!9C>Iq{XU
zSSXSYp)A<v5&773T()9IyAW^3uFYe$CN-t*c0|wTn6RSP3bOY?mEt4re%)pk1~0v^
z#nWB(LAx^XQCa*_>f+*TI@jcp#jwO0zm((H1e68l39phDgye@mcLnmalr+#IM#sQ$
zr}3|tc*2eP^%qYjUhq9>qdfOZ%8D<wm&*{FI#u$yf^t+pe{n1uA8*<u_2xPV(5F3^
zyw0&`Qnx3*q&`PwF^!nTQGd{9rd1CAJLWjFQ73$b#yUIeg333;jDC+Z_#)iLC>L9K
z&fyu0l$kcLL^{aO78f7J7l<yt?A>z8rL7Ec69MQ@euIy%QtV?KPPp3RFrPv@!q{|X
z00){Dqy^EFuzb$}l9f#s7X*Wz%c}C%P#!i!gS&tjzDD_zNDR$W%CQL3I{`X{D;BPa
zP}XgcO+uu0E8jj8#IJG!v!OH6h8Sp~TtQQP7cQ9FgCRQ`25|(SqirjV{4UgaF^gw*
zR}x|XbLaNxI-6;Gx5MlkR&-!EX7H8oWWC6gmfQeH7ouJF17&rYU55Pv!rMg^H3DR7
zMjk&gE2%RWRY!co%Wj5$_RoG^`1-p6?#LAea6ad0;G^l}(MQuSe*K&2vtRzv^yHIY
zO~3s0A5V`SeXNV(zPdT*@GM{ps_)~CJ!5<EzmyPeQ`;c=cY~aV7r4pyUHq>7?SgH3
z=5f3!<$UmdaM!YT!A<t>;~ULUpPF=i;J<H_JImg}U+Y4z+j5t`_t6Dzl8w+>H<h~q
zD|$n&_=wP)E&lFY#tM*+8ngb&%vu|~O$GbWJkQfRGb$6`;~?bXe*id{Q}c;@{5}ta
zT=O>RhURa|Ghbf2>1!pb%NT%88sRp&?Z}JT&@o<Q&+A9<k3_|s`O+E84ftr*X88-g
zm9K2EY@JSl{tn#WJ`p67wZwD(3R0iZe!yw(YqEuQtNXk7bs)}pD<7k8fXl|*TPqp$
zz_wwOW8?XZAbjawv-jTk0?@=$676WMzE(A)4cEM)E8|EW?UwS_dCRID<_brn>*zW`
zmu|6VnDypzwn3F~2k0=mpt;-uw(+*Zb)8MyE^Vm%Mt&RUn3BIG=J$kK3~#z#BcUpu
z^Z50L2h%;xktYB6^MMz@_~y)gH;y&ymL+*dJ(l3_)%u>!;tFqK0a$ps;$9o~7WDnf
zGQYv)XyOeyv?p@(vdE_I0Mhe!daf(Nv+yOox)*UN+W)`JTXA@s42wCUS<Lz6XHTYI
z|Kii><BuQv9uSLPUw-r5^!oK{-)G>Qnmo0jOfTr{^SbOK=lITjpm&lzn2wJQrpJ#@
zrjH&!lFp%Db$E2JKV8T_Piy_aYZ8_78pMwtA5TvnJ)S;#crqRCP1Did+v($n52nL=
zY6JOu^Y+2?!|TiGtM6Y;zx(?6^yRlNrthA;m|ng-o8DYrPKW#Vra$_{N7MiAfBPG!
z$B#~?7q8BzZ@zmzef8bT>FcL2rf+_DIen-6#cS^0<h~+bUHw-b0`6(}zsc2pS8lib
z_!sc~l?N;ws-M(e#C*wrr=92NIBc)n=h1(o_VQLISNfY6KRVf;J~_ER9c#!u;QN>O
ze0Kvlg}Lv_CO;I0j@&Y1!68}(1Yh*+d5{B5llvZ%ah0R^+djO-ZX51BxY}uyyZXf6
z^&VuvO?1bzhP6##tmh`|;y2tT9|6sq^ewx)JHz{we^=-+^JC!;5z{XCsb%?9hyUbH
z|LnJnHVtDr)QO3M3BTIqmnb3_hiuh0X(WuSXuV7v>Zl#;IpJ9XjUQ~p^B67nbMlcJ
z9JN$1_F?K{8=V@%v>rQaId7*qDKuE(yP%O~k4<E@Hz7SvXr?)NEHdhoC$7LPRvwu>
zM`Xk{`pi|7PsglXWrk>KGgI2i@#*A-hgY=4;NjUf{;;)u<@iwM2Is_Ll#>$*Br}Yg
zgN|*uF>Q7^2*j>VEES7+;3z_~wVxg+R>4u?NLk`p7(Wee@$!#e{DLr6q6gm~I-=-%
zVL=<^a<b1W?Q%1P3pN)%w}KtSdmsxa*x94T2YN_T=hhozqMkiE?63f-Hpk?T6Z`IK
zS_QPrU{@mRRIAO*=xE=4Z19n4p}myTSuCuY(LCxGs6JC3U2-t-q#Q|-W#aWXvt9%h
zBy@PNy9oR*B@DtsXPY*_S2d2yQ=hpH$tANyy6%4^FTK$gu7`krt4703ju%u-)VQf=
z>^F-<EEwgFb7@J+%@|{01s8Ms!o#_8^k<&qv{%cV3tS$)M1x01u_@f1T=uIDQ%5G3
z;_(Y23X-~Yq@UOQQBK~i&7ZLPKe03+<Q0UxVvwh$4iAp<w4HQ4v67yD-6cJ2Ss24E
zuR37TiVt3C*snH}Ozpz|e(h^w$>ft?b@5umOmfkYe<keubqS6Qyy{8WP51Ff@)Ki<
zqvI2E;<b@}{UWg-r_IVQ_KaT|WBvu2+Qa;!g{F_Bm<{!`J9@^+1>Fz$w5_?}Mc1F;
zh8(Z6+yfA9Ha-&g9&J$Vi62t{J7viGIJd7E7atSv#DggbN6zAdj|EN40GhF0P=`VT
z<e)A>p45OY7t$%mY1*)uh9F({=!+L4`ANJ^DB`#D1IjC&ZfX0luaYUt@edt;;~W9L
z{gqDK4g9&4Di+HsHyYXth@f&14uFo5+hg?2Iei&38|_=T%T_X9J8b(r=D7_U0~IJg
zF$~i7#s?96i-fcHICab*WhdNYNvMm6$hmjSYmA4SiOTv$p7+{i8Wf?+r~O6Rwepg9
z%lHTfaz^K4Nnjt2Fmh;1F-Yb_7KM<-14)~Q;h2a<x%<Q@JC#|6l=iQFe%0aMA(%VU
zkbdU4;^xZPbawh;divcr)7M}8Zu<JGFQ%v8e>;8u?bp-m(^tL;;8zmU?=tgotHdZs
zPv?9Eq?A%9Xlh$}A4tp#beX$gUDi6?X5ZlPUJ39#t<ELptm3*0j6TQIu#GpuC|~QW
zW$QG8sU33T%`bY>`Z}!bXPSJKkXTM?d<7dm2DnLgo6VZHQ?HT)+vRIM`H<j_@>#!d
z?aT*e*}wxv{!Mb@Va&W)SD9(HQ5FNIS=IBLCM;QK!D!#91eG<f<>t<K%cFCh26oAp
z#I+@pp@kRS+z`I*t9iH6oi^nmigvk>k<CdVt$89hY20kVmt?l_o)43XIndD^<&DpY
z!X3b-%#1qYb$>^<^n6l});u1{ByB(NuxuaA54+^p1Zg?-a@{=FfxpP2yWrbs^Bhad
zciFG`O=n&@%ACy9HEiI;*PzR6*&UtnXOmGrCySiV<$0!EvRSVb@nt><JXh1;WVTFR
zDL@|Fs<W)++o3F+;;L;*qHOfHPOX=GGa*j8Zl}*Gzn(S$p0jId7k^CK{Ez82$lR7=
zh5H$A^tlqsYY;Cb<A=lIy=Bpa|HO8W$#3$Skk>vbM;%Zae7|OVSdgcW{NMVLGj8}k
zk4~hu)qI1@b>LHV@jbM+Spee)yS(y`MKQmjsTR*##yu4Ni|VKr;QWJiMa%zH@%^9s
z4cu2C4rKht75{m4ad~0<LjEo;-%Mv0SN<QY=PysE?|yhbegE{i{|oH={K8MOot~dh
z&z`@Q?8Wrew@;_9zIi%*^ZhgNUbrp~9?)J_)1JQUJUTh{lW9b9e(`2{b#^(uI=z_A
z&iPx3ED-XeVb%z<cy)g|JjlHXA6>MAe6kS9+iLt8$6my9v!z`*XYpnl|MdoaK;phg
z`w-C9O}h=`Bkiv?P8SGI`*S`(w-?SC3x}%L!@~#D!CuCI0bkCri(eezF6E@`S*M7Z
zzsdpd&<hV}o$;EOT=BWw>wY$2@k^fO3q2=aD;EDFta&4S7ySk9=6?;h$=87PO*WR@
z;u4)P{TNvLSf^WqxCP?l-wpiKvhu3K|LOmxv8MrBCsrTKLfW^QMLdRe#v=zQcPTg#
z+X6v$$vB~7?2Lge#A2L+C)Zwh%FQZ29Od{HJ2z_tZDzdm%8er^-#%6>HN$wYlm(t#
zIuI0XoyZNXG@?cZLFL!%fThv&+3tH*W#_hL(lcufHi>^OdNvGW+*@2O+LP63%hG33
zi0{6@`9$q<q3`=fIVZFJ#T9*!N#s3W3@BGx&N*3gDvr5qMCasO<&ty36G3rac#>Z(
zbos>?g6L(@ic%8an=4&_l&2n5wrq&i_V@*4<ng5oTfcXAxzS#}bRqVyDOhZ<QM~4Y
zMLpz8x^5fM#E72dQguv)Q(#Z3;|m{^tFh8{P%yK1)0+}NP#iwvMj#$InNn{2hG*KA
z)xS!u%pM3Y5PgtEyA7!V#JhTelADxy&?}x0w@;vH^R%-)ujdlb9u#4ht_WEAC(11F
zAS@m6ESkM4&j*PR)e;59W9iZ#=1V8PdTJXQxO||e@p$h+YT^Dv&t9+%-=hz<fEe8;
z2!x~6<CG%tnA|L!7Op&~eVD`%2MOh3qdX<IFXgt6;Zl#ZUrgR<63MF%dF7x=*BIi-
zEFSqs##3x>-dwspSe){>R{K(p>mna41|>X=e|S0uIJT+Rl_qqQjxFG0R62f7cEvz7
zn7FcNl}GT=&$%Jz0%<8O`&3N$r7V|TF#s&EL93#Qp}174bIO$u+5x%a#yTRU;Hifu
zk4nUW2Npb2Un>Tu(H$55ah2m3pJeuDeX9j8zz<sn17HU;8h7Q#u`6b)s!bL**Pul`
zaka^?$Hz5$w-L<f8JD7q71h!Wu%nU{ZpSpK2)mS3jYbH;g44#OBq%v7qoG^V%u-n(
zwquNIfH#|!n<z4NW>iX}2jT7XCTR5r>Br0B2i$!OPd;<hAaCb4;i43<V`d!SqWIlz
zK4%F>r6ijkOklf2Wh7>D+YznWgTY#5G(_K+a9I_Ovz+r9@aU{uMCRibA3hL<v0OgF
z4IW~%ULlH8^)|rrg=EE_AvSEO>%uVsRpvgQLMXE<Z9c{eZ`nXi|MZ{!cVz#rz@`5T
zKNiNW;>QmBC*{%f_{k^JC!hav`uNFb)31O152nw4`RnQA;YYeY9P0ePuj3)tGIxKE
z{cGdw4+AQ7hGonKa=prNvI}mDa`V_-!2I{9<1Rn%1MlPGeal<x({?xc*e3a@Wr|_H
z>;Aeij`t|@r|6E<ZGP@l&Y*+-uJ*pk*IG8x*Q_LKp7omWIla>YKHp1k+p#m2!|?|_
z@;t>63@pQ<(>Zk>kKj$;+hCM$(}g#(qjS4WVLSEbe2_KN72W5+Zm+SN5eK$CF@=t~
zrQ72nRPCcMr?uaP`TO?(8@{kppo@7)oViCq47Q8<0;2OhrN~(wByn|V)-42eo5*yQ
z4{+O9Fx@;H&2snyY2P76A3+zo0e<31`P=o<AC6xlQ9%YsStBoPv4-X?ZBX>Ay4Wzb
z^o?_5ZS~o(Zo|NK{WReyOZ_@7_ScIt*O=R412&a!2y+K)#=fnM8?KlAe&D&m4RBBG
z=;?Ho#1E-ulxXg0&g1XNk-hcPM0^kCZ7lbvH}?)?3l5-Dc3|?+KVs$SF52Jsz*zXw
zam&3ie$aY!bmYYM>?;<1ylAyNodm>JzVl;?4hvKK^#U~a1N=i`>G0Y@{^RTM!;|Ub
zj~`CI`usCLrFC|8K0ST*a(eap%uh4*|5@=g7Yk{W$$y!X`mH#=FZm#T87nTIzeP}0
z5B48S4^_@1mBVZNc-`U4S7+1f(=)$ChkH~<()owK`Nj0V|Ihx>^v9o_OrJj7n|}G(
zN7Lc{;q**#{QjF~)8BmcZ2Idje();{pTETZ+~ph`?oFS4bUgjTUq6}t<3Ij<`s|a(
z{=x2r>itr-yzY=E#lHOZ`Sk3?>*<`=6sj#aE*!UX`B88G&^YH=7SvcEWU-9D0O5zh
z_uWS<0J4xqyK*0hC)sEN&h&JJeY)=%BfQ@5-Sp|B{pqtu>cc&bC#Brn5UyOlGxm~G
zCwy3LejxWr#4B5%<%?eKzqGH0S+91XN!!a}EN{Z%&sN%P`1g+eX1NU;x+^*N&(^Tb
z$6b8yg1hK6pm~$NWp{A*K4tEL`1t95*6<->yA3{c8GhpHAOG23{8mF;&c#vEkTA|+
zwG<a#ZUXwF0n1^>Az3uJc;F5+POyvuNjuxvHeiueY7;&$0L(DZmX%Gh(9f7JS?_(|
zr>yXln+_iyd+KV9I!22!xE;jbXb(H&GE}~8pq=w*CNyC}h%d+g;=_5{co#e0a4Xp?
z635rUnUREV`!!#ggNdgE)ln2?ME=PIoY=fz1CArxkTD+QpGGZRl;)dLA$<bH+6zd~
zx){0d)$YjIdmb+Nyx^t8IHFv1vGs{qzWJ3)-&W^DZ~uzLC&%*VxJq9LMHNIhH{eyS
zE+~G@p?yVzjEfFcg-2QQ6MV=~AkF1<9f7n@iH#fNj9;>lsEaQ)`1OJeHmHJ^N9b6F
zXNSV*$8ZWzgj#rDzL9Ew)|k^WJ^*Vv4^K8KCv}CL<DEz?2D5{zotbk@u{M@GJjV=b
zvTzf^t9`>Mi$QKLGQc0KT->9V`cW4BM%=R_+kmA1D;J=$!E<hjg`gzL%NQPc;X@Ds
z_SWgLtsj5(DcT$pU`VZ@eN9<lYv6b2Fxqz?TY&#J%VUS9tPVUGTAu3C<h7pkLbk^r
zxB4k9UL6&^`Kc{zOy}t=KW;C-d;EXapywyu_;o%0Oz&~f>cbOV(NRz}p7X1vIf?J{
zMfq@E6JHjwIDfc)!l|7nO(b~AS@e>Uab*6FM70m+nD|M51353%LYnwFH_>1+;Z$Qk
zL+h(gEEeCSaV2odx1InT!g}!CTXIC~$~5qnz9~C)3`pXG#;OGFRm&iiI}ume3d^;u
zfG0|=VkPU^20q-fFQt^@JcN&Xcw*YG8zB|-ITX9LwOS_4;JI3*4#bRpwVDzc(rCi-
zAwVuxMQ0K2*=5p93gm|>(!)StDL$d|5k#KYOD4d#H^|}vo8=Y`Sg#ZDcB-uS*0wE(
zSj}Kxa>@Xr(#Fb_b$rHy!F2o-yHOX94+ersw=qgS8rqoLc-N)N1C-Yoo9k&1WwK;t
z<`_)3ZV5I(JhPNDOhCIa1hFj}`)m8wAy+b4xOL1(&Ja|V>qi5O(9srNat|9gl((ow
zha85D)N&~MqfI>P!iNy=FaPo{e}3ZY-wR0J2IxC(wq0tDb~c^9dNw`%?#t<$uYW&%
z`^^{1znH%N{+sFQ@_gD;|KI00<v0?aAGv6h);+8qDmTGc^812)d(2T*%Feir*TzTQ
z2xYo6$!NRX+9&x9Nqc3B&4%qAVVgZ^8*J0NDXncrdK<u9Cn@jL-v;e}wrPBokXlc3
z4$GHGTlhQD8b<!8*SrzdGWMOs+>otdv|a0L@W`PHpJhAhw}!QBq-S0y=nZ~yvgmi-
zI?-r4SK>+$<Pw0~jKY=jop4E!c^^6F3FX{aUG%svkMxp87<r@4HeMij=B~(Fr^eE%
zOk0#r@R2{@+Il|S0Om1jY`d*#J_BuD=~$Ls<&k^tXq}`5Uwm|mY&V^@#<=;^W0)5W
zx+u+b<6QHNk9H+f3LP0Ku`lxIY|;yF@da$dlc4c0yky#!EGaf^mpYGeZP3<jJ@7@3
ze$H6WjxBQ>^iz%)mN##dE%+wgsgEsG!duGr_>56f`A9y}BP_Dnmt2iyj&sZf%BHp|
z*}NoR$=Md!^^}mwJ1uQ?J+<y)v$!&IZYL63<J;5CVE;FP#fuK}n~s3_4kO~hbfv^!
z2#|uK!{OmPrWhBpz;cdMn~3H+4(BJoYVa*{FU<12yW68v19&fTTc(`*Wg_61pT9u~
zp7|M1aZ$#b>5%`kI^6f74^N0uJMI~<=*9g7UT^s5(aCgld_3{$MC>S!#W4T>SUp+A
zUqf86ICsI{KwK$ZPN(M=UaUG}!RzeIuME5t|NQboGTtt5<|od6`T3{Q$<cx4k-QdA
zt)}pang8gzR5?6Z#(%9-P0p?R5B44Rp5~ce;FAGu$y;yu-!T>wS=a)+aEC*E`kF(^
zy$ks|V1beQH@uyP`!Pz`s=2HyT$vzoaqlBPj>ctpZtl7df~uchJKCR4_>VQd1K}$>
z`&1;ALofcD2e|CYjWj;w2&`v%u{Y101xDGZI|C<&vg%`Wb^^N6Z$AsZz)z(Y*b=8;
zw|?LzIg)GTlWwv>wp%>IZE@WN;^U|LS;L2j?Kb$(W$?t;gFpF`zxeIqmJ=B{16()<
zAB`2nIC5^faA2av29rSz+7LIzKwp`TTicGlqf$P^pMO9MuON8omTbPZD$k)6hE*g)
zH@-XP^gnwtuQ`s#{LqBv__UjSu&bD1hTnd$f@B}OF~`~UDQD(~W1e-pLMF7cjUt2w
zw_a_lxX_6jve=cl+8HES{6+4Qv<!OTs(Be`%ZWHXwUaUDZm67?DKDwPi$A%N*oFnH
z+|b66KL&EgqU6b|^!Rm3>;j20wl1H_K$`E?2`+oe@T5RV*QAOAsTf?CFhd(1+(_R%
zyG%X(GzvO(Aq%*|;!_s(IFLM~!4nM%m%Z7jct?sN{w;{40K0Gjw})UZYut<FV<+JV
z;ESm&9;0X8h{sh}#3M#D+E#V;&GQlq@vp$juvi5_*rS^|0oX+5whU{SoGH1`(z0!}
zb?w7eZpWoZgP@0<Q?x*`h_}cR3UujM4{l3RUXi9N(s|qvuyqXOODfK1LB1%OM24R7
zJ}$iB2OY-p9Aoi3787sBPVxi`C)jAR<3EAj%Y>;HwEU_;$@rsFzpjvp<o^D&zsIkK
zn8ZHt<dR8M7L(9Zx`!Y9W<iO`C%#mq%X1l;+;Qx~^h8cL@#h>-$zs`d%sx)ymwK_d
z1+UK`__t5{NFDu(N!x|?7yPmT*xTMV@gu}-1~_EWyvw8(G-Tm}el8H=Fks2&V_Lvi
z194vlF>O6^l6%Vu+4}lonRM*ln{WwxaU#0_^-15m$mkXwMNz6RxHtPj10A#n8JMb4
znB@R_^s1z6aS<KsMVHdENMpj#V)|#(6S8U_L-_`!4~>YHf-SYfcgF>^EKQ0p>W5Y7
zcMu&H0Dth|m_Y_!EajW7m5a!;OI9`oje8{Ery8hZVV_vD`BSNODQECo-0b~HZY%;p
zwNvreFzO|_q4aU>g1gUw?B{GYCLd*EdkubO--TKNVYG4~)N^4}jtJT4yY21Fs+Ixs
zF+4JXCR%QE=FWa+|4V!P<-h#-Rfm7KAbqr?5t^ni`}$D4cURNftBdLM<;&^mw_i=)
zee<`|x8MAJ`ueNCo}NAZMqSO1pw#bg-}vT1|0u#E4;TD!pwATpEG7Ops6%1_@J^ek
zq3g6!w;Q1Ct-C}2PX5|{l}1~qQg7K^;F^y++T9iI>KlpEpj&5|TVd^Yg9aby_U()L
z4KVAiU>9#pBdq!C=QPM>9nt_ZzhKl|%YgQ~hF$u=HrdD<Wqr(uBOa%Hj`gHmDOz8&
zkJ+Y^Y|Oq<C4aZ_hihl+1xDLVzRG4EDWs--O3Qf%I(nP-q&5F8v^Xj+bPpXwT_!Cv
z>PL3nU*iqUCvCRvitjcm-e3nbZrW-KZ5}%1GYaT*zX`8;r66Q0+Z*UPZP80PgHFo^
zK7?_{09kaW-h_L;8k0?(;xBc|{6B!aWsR3j#4swAj+xDBIyQ@KbO7rWIqF2u%!+;C
zXO6qjj?ZOPC)odz7QH0E9xT?b#s_H}pJ9*nT6`oS4<4IEwm|GUMGhZQNa=6bk}Pgs
z<H<q#Sl^aX1$=b64YuhnywUef<tEAE8gEIr!M3rs*pj-B>B0P%mRBHx`(MEBUFj<T
z3u3u%r(Tqqe4;5+{E+9HWqGsP#O6hoH_P}!51#MI5XU^GD?VuET;ehJ;+&2cS$}^Q
zAz9sLV7?~a(eaVyYxkzJ(^Eh7#ljU&pYenf^Y{I|z3G_O5c1ay5Ay%w{9v60FvqCC
zdD_ddsdSzSV<9a+qE|h7q3rS^3suNj(7L$r6JSq1el&gd=|_sO?hR_k!k4NppIHRs
z4O1*)ao>a|%2=#BKIA{p-pU6HV=7y9<Q@zEFXjcXgWM~jJYuBY^ZgRl9~td%bab4j
z>y8hnL;2n3hsgY>S@lkvDsex>!F8Vj>2Y}A_Nyj8cyMoeC_g6$JXvOq=vcLN&7J85
z=UA8An0S;)UNWF<`nu_HHS?2Oqilgiwy8Tq@d4Wn;<aw*@UHn0jJ(_Az&0Im?_|4%
z+vMH0=|4)j*~YS)UCwY@*>{2X`00Mu@F8Nm4L)=kS@`<XKmV<7(AjsL3_-<!?=PGN
zasAOwCs2m7_1x8wH6DLxine3<l}&k0UVQR!!{;PpQE|gVPul)vzB#3XIS9OBkV`*Q
zWP!FVC_AIt`evqMl`b`bG3yJsZgeL^pSaMZo!BP<;j*cv@MB--N2B*0zkSd0<kle<
zf<H8dbB+L`(lbC;-T(?)4(ONgQZ6T4Uu<;7HLa9u3VALzHfQJliyq8eD0##;zWj>~
zR+^N!R4?NuHue!R<+F*$1rLOanAgjA3TCb(WyFttq>;q!3qE?pNuO2xKG9;+<u4HJ
zBtKbXJbEc6c*kapA}O<wVjHZdjnc6ts?t~iw()J04PV0h%O6m`{Tk8$BH^U0fc^AX
z1#Di5?A%`HNC2rFTVTmC0g&65W9=`hX`yz@n9n$Z5pUF``v4$kk^61Q`+QL1YHMUE
zPe<k~&vAtWGG%2fEf-c~R1Ld%OuKwwq3s+HYi=jbFJ`gjoRAB>{Lk%G?V#KVozew2
z4~5Rei*I6ut<Ad8u}vj5^JL&QQ=Una+6#^vhjGkI#~0IH%u;-&RX`@2Os=wk<@lM9
z+9ngzUN|zZ61FZ2JL=nhuw^m^j|g2Z_Jy|{n{lcVf31^p6SrlRY$aaDBBsiSZ*+aL
zKpTALfb|O(Pb!h8vF$fyHjNg$*m!Th5S=eN(C#mfHF^Q|(ubp+hs0cN>tw7KTVSkk
z_!t>=vL}T2h-zDCzZ^Hg&@(A@msUZ5u&c6(fq2RvJ!h{8LJ+T^IF5?b@*IQ$am6Y(
z@mY9`qMyBB<sTK0F3sW!3SWSA1Wg-ee5rrr(59;RIhq1!MwN2uHwCoyMme)AV>=&q
z>?uRmII$Gn_O|GyOy&GQfFTT3O0fwB?2K6g;=<0wwF<g1Dh3((=p%Xce39DO78KkH
zCS6ll+Cf7gwp(FKd}dfLIzA&=FVd20b03<Nt@FsrY2_}hApE&^dl$AKIAo%MDBm0H
z0k3k?1tS0Q=l}TnI|f7lY=y`<e}m7bAop$OSJRuz^Xc`=r_<AKzns4N_Uq}JufCYR
z_?y3)o;~}XxMTc;0095=Nkl<ZyC2reTLZX$^mx|@(mB7+haJj*&d@5rh`$eDGhZJo
z{S?6ZHa}xo>vSg=^#^=RcY$qoH>F)R+hCg>X&dyY>3-0zIDw7t4nx|-8)4=zgVggF
z^NR*rZ%ntrF5O-H5s=I}gP!Mi<Xh}pf1R|u#v4=Hj<8Lqq3Q9s5O+?^Be_+PS4uM%
ztpgge=XsH2u30@=s+&G7DSYX*O}#L(8~H5)Hu;-=??S^GY3er8qy8>@ooJEmd>d`>
zXs57O{)={AytU4p+&<#hm|xd{B+iNsjJ)QPg*G|Mwxv<-zM^dA^OkMumOJC~xLfNK
z&!3_hALg2_<9bL-dn3#BYLVfmZIt8KyhUfx0|&d_e~WyR2dw#s%3|gP-_#8h4zJpc
zly5qNzV$Ziu3h1qem7-f61&Qqlgm__%|2YH^-P!tv*lw-dwUF*-PR-+&oe(Ixts!V
zYM=aiYS<=ATDR#sZL>+*2F#$jqn334o|181n};lxxw|3$yEkHpuLIUUI>&=>m(66!
z<-_#d2H&&euRyZcgiraZdkKC!lI`+z7;lB?@A0OQ+a+U5hRP52_ni0<JLT~jMD7i;
z@by~h^z3{(zc`=H&Q7PZ^D_k^8~#u0!JhH?DKo}Uh07vR>hX4ZqeQ>&Q9k!E^j!g7
zKcOWnPZ7^m4*!w%@kfuRlcVElUp~AjCqHlaUv=x|xaLEh{NG-R{qe)2X<s?7XFNT-
zn9k4hmk}&lvFODj7CE%<yU;GO&s&~|J7btP?Hr#RPAA8C738sU+p_?svJUnRZ0EWX
zg!`lFKxDA&FFzPQcrZ;5kM2*$`_PIAd6bmhoC7mY{+x?mQrcqCufCt@(yzQ}7_kj(
z>R0<3U-$;wjl89-Yr29p58C&y=@@|*4_C0w&sYv%KT2@6X#?B+EZ}ohL4^~4H`rNr
zvs>%jR`y*Xem;z^5k6FWH^EOSqia2;TKM{}ek;%Fyy`$5Q0_b*yeM+ApaX=5;nM|q
z)7XGys{>LeBp<?slly*#xH+^TV&^d{1u|=ZBXdC=2fVO)am)u+oC=TJ7LAw}KB{p0
zCOmi+Qp`<eEQ)AY?jlUVE*%tN;~Vv^M#+_%)Xwg8WZw7ce6>eZl!P0>pFP^ne((t=
zpV+~p@~DtF;)y|cFP_-v%ugGYwPOJ~R@#V@tgaa}m`}inxbWoS^G+vM3Y4t`PWX~t
ziesUN)=JratsskEE=_*$O(-5m(ShL;J4w8GQtfM<yh+6uXM7;XP6>{EK@)|Dj8)lU
zD(H)_3UC?rfiF3~s*9q3Rf0&n*tH+kU3$n;1`=>x%*f48JMw`?*`_^sX-Bt?HtJ?p
z;_=QJmRbg{$}x$uz^w_NzE(E2fi-C$Orao+(H_yOq73Qh=td74jQV(khD*FHFOS6T
zO~{~+Q}ZoP+x5azd?1LOL$xTL5EEWi0e0l&JX`!&PYI*=s8NP*(-4(Q0Ubw<oRGyv
zm&E~kkbbmx@jRxXH8woyDBPtE&T~wGiBAERDZ$6LrzUXV!K$_^Zu&cr^G88~i*5#-
z<1oXs9oY#_{4k{pUO5Q`<FZ4?V_I@gzT6&Fr>bl8o5uoG%JZU=CT{*Um;J_uR~ASp
z5_xR<c=H6Kd_faOw<CObjs`w7Jy3eCqx14Hr>X-vvNWOr5<jj#4)U>`TUZK?kA5aL
z1N{!fA9&k}VSS?#*sscquasikbW9wCFiLb^dJMEy%9!;)SoJlCFFM`q3P#6#k~+L?
z3PA18Jne|r?v%5az@EtQThXFt9Zae`*TzC@L2R)2v$D}h-DuIoUyh&B0YL4$V~C(E
z+SMF?bV@I($+YXP%DGDjHaP%dWUTCpE{^xfv290fM?Wcd{E<@z<)uBM!v;W`r!j$&
z`7^I|i4$~a4~xN-JiyQbI4_cStjU=JR8K`Uod+c}HfCE6QIv7?Qj7`ly6wG}J#}?#
z=pt_;fxiSZ=IzcU#4YumO}r1S)NKbBhI7Ybt(2?1%Y>F+^j)p4j3v>J<FZ1Bb`wNq
zG+FuJ#|~YN{WgW4f}a<@{(S~|w$m<n=S8qrFMgPQc>4A9-FIJ2Uw-k|)9-)(*VEbS
z=ek~AO;?v^J}w^Y^TUn&Dw*@U{d&Cl;M5k#F}RGOhUWH2i%s+v*^+L9n-*PegSE}J
zv~D$c9|G3pbn-#@9?<o_3GDZdFvl^#ng`arG2I1r*{x~Az;h~g!zjBAcIkOeC>VW>
zX$12sZy0q+*Lq_m-vH<>@-+>tc_ZDtZP4<LYqVKX+Q1+>ZL+B|!pNUx<YO9m=y+Z$
zpN_-Vt<7XUXq9VH`8tjK?YOR##am>JFZCk#ynKV#K9Vx~DJb3u;*@SPgxKez>wLi1
zvfVtsE5lppG1g9-Hlsc%Z4ut0*Rbd%@uNB+`tSol>3QNVy6S|Cc}RZ`F>mm_&8Nhh
zG&(~Yg-5>WQ;2cbnPi=jUv#uL=|LV|<B_d(5G*~t=q+-qT<6VxMf3m;cVo-<GDJf;
zFp5XMTrZ)QG>pJmJg0&&_RcUTX9$gPuT#ne%3w%rK=NJb?qH7_<<qzFu}Y~=x7+T7
z&dup2*l5`SVjKTpv``kn9!wWbd-pD<z5CGoUzZa1^5l-xUVZP))sr9L-1o{oKc1{Y
zU*EyR=R0>lg_UdhJ=5GP$RgODlYe+Cz9UjW+=ugDCa8mX!VB0xJe2tUbatL6vHV2V
z<)#0l>*Dgl!C`}6!2dTU;CN;Miob54*7^U!REDy7^&pFJl+EIorPN6|i*5YJ82QmY
ze-|Nr+w98^y4-J6UlO~Uly+g^j@K|A_$f3Nr+6*m>G{=kR!^&;i=V@TJ-^EE@Gy&H
zv<-`K_NUxq=s^|&k5uME`FwPI=vNgU^SZ*L!|715^Yj<Vi*lUj@+I7LbRWr&;}m+V
zsy;rtFMpitkU4PcbqbyHQt;%t&WXmZV9_T}TH~z)m?m$%60LO#6RtWL3%_y4GS~#l
zC)hRFCJ(&lm$U;YX{~b|QU;Pv4b5M}c0ZdyB%K=U{{~yjZg+jB^6vui^I?3A@S)-x
z0iR>~vCx)X^G^SFFMR#kpZ~TGiU!E#_+UUVs)W|y;-Xb^pXzLo=rGtb2<wtL@06iz
z9ql;zdJAuMST{UcTBiY#L}6_MF4|vogZfZ2uekcyWAO~K^Tn&q52?Qcn*EwZ(us%+
zXn*9cgJQnkIhIN^IB`K;uVvbrn?It;kM%JNx6EZXpUBWt23S`qZ0pSZ#I6d-O<H5l
z?6<Js+Ec!8I*z{M7azH?Emkdp`42g*FQ%1+iD7O!JaT_Ik)N`&-W)WL1vh_*VXUwy
z^(MclNIyi^?JAv_rw@-k1Gy0{kRQa+6{32e!-dnMMFA!Za7S+ZEp1B3@(-VrD1NxW
zSVcy}S1x5Iza^VN_Ku;zttmeNjFQqSNP8?(Cb{^ir~sE?eSZp9qF>AljVv*NAyY1=
z9?O=(Eih6%W$1E0p6jY<he}isC^gT@W;sV$lx?Vo#fy$FK#*k%FW4qn*Lg|o*GP`p
zpaL%R#1+QkvF+1$W-HkZL7{0?;)xY)OMBUig{5OTNO{!c$YVehEV%S#YxK?;pko0)
zb9<nP&t&u1VSJO7gC%aj^6ESDwButM&!y9RVeT!-`U@UPF2Q<%p7;^d>>FLf$fJxJ
z48lCQqAwLV$5-NTiBZ8svW~$#c0p=0YOJ63vk#AB;!^IuvcUeb6-(P(w%d@};ZRAT
zkx3)n?Mg~&F*e8$FmW%LGsRr`#qp<Uw=^sqlz=_EI?MR5wpd^v`oM#k30OR5ECB4=
zcgwt=CCJ?VE~0(d!Zyvf^~S(iu&5z-t0;@ardXDb1U8J-aBZIy9x!7dT_GGvCkrU;
zw?*i58DdmpS*KaryH6pM7=u=~ykyeyk0xxF0^r6dP%aBf*I-A<;MH)Sb&Z5_3s<bJ
zi@D_+q64qYpgb;%F=yQ#_`yIte7C{~W~IDn7Kp(p@Ahb3>J|jtA}cf@%&4}alMOTo
zew3M2c}5ikzN4^=&&0xxV-~drMw#mg+xqcgS)*OJeOY7njnC4KZ%Z2rSJv$TKF8TY
z{5s~J7ry?(0s8yx^ycz(I(_+UdiKM2(|6x|J^k(P{%ZQY(&?)g(>wq9`nkpquh8LV
z5E&z$XKFBU{CaGA3<U#v6efiSj3aJWx(kfwRv_oWq@MyyqzjC`$JDS3zYA>mtK;t`
z81-$l18$OSgWLR#fR<C){MfSEZl{mdD{_RjZKpM?`J?X0Z)kqzb2sY(qx`n9a}5Sr
zOSkEdyymZY&2P9So8{!qXMWd5baddGwp!VZ$+E$3TMs-^`>@WyZ`ehTY2@$1A(L9x
z{F<zouCvrEkG`eW-NYB3&Q>cPTkC3&B-bbP1&<u&Xctz=l{WN;v|%Hq4SF_N^A>uk
z>&<%6Au$GsXTh6q9nTsf_qjW!TjPhf*eAhHR`jY)=E-63lpoh+JNRt=nAW^C{eC>x
zYKE~rL@8fvB42FRDX``j+x)Sucj<2?zS%_0fuL?a_?X|6VourDQ-81CW|!n5R=jO#
z(HX&Ic*X#@B&W97cFw(f+B{a*kUEl&Hl1=mEI4uS<Mg*mZ-iete_%d2<)|n8B_T(J
zf#w^01kkI+FBXORULT7wRd3wbix9ug&x>F>>XiG*E848)1^4+Y1<kws|6Z4u(&Im}
zj=d-b&8rSsMB)eO=v-Z0d2#FT;Bb1dUyEpzBmez>=l<`00sO?o6I%OTpyL0uxL@F>
z*81L^^xs^)nNCm7r>jdA^->QO*p7}=4vTD}c@?AQh-wQwvr5PlS;x}h)rOSA|ADb+
zcYeYDjj=dLd+}p?Vq@_wPjMa8>kRY%=v<rWuXm1}g{~7Xc##gKhxq5Uio#j=%HK~U
zCgP-@!Tmi*k?kozIk-1HI^@5|kh?hy-JeAj9=iFvav*peTee&Lt@SL0x3I72Bu{^h
zFzb;GHZZpNXjsz?@N38!T5p%WB$m0BvO|lMV+#3o*zIdK*zXN?mfg|qk12l}#LtKE
zHNuCAZ-ft4=Spj+g|Gk0|M8X6i9m(AT#h&ppENBoG-lq+o-cL9T6`Js<#S7X&lLmL
zhOIg98Ia1OVvEQbspO?>u^?fF;gfIq#*R<a!{Z>aYUwxM``USv+$YMM4@IjDMC1le
zX+>Z>2OqqOQ5$=wd>m-kJ60`F<`Ygxvk;Xp2)BoyBCu6rLk?eAPLlCKjZGrc);IU;
z&Iq!!1%z>0$)EW1nL!GDQn3dq<|eN|a$?}}XcnpDE&{xNogo}vOyBwf1*dh}CAx3Z
znfmZ0M;AVF$+@7RJKqSmUJTLmFVIjRvCt(SR+Djl!4eba?Cd3vJ{Qv$DsT3KRCtL2
z^Sd1H8iyU*_C*8HU+#lAG+(vWu8i7PnXOr1br&|DNP}nWQGb&a2EI`Pqbw^%!P3&{
ze3l7>y4{4jo}$V5G;}P^v_qS5MofLP6zFjVji2JCJY<&3yGig_!@S`5*F?$@k;HR*
zq?|;mJ#|I74DCa_cbfcx^KpFWS6TMWSS!2cp-eTn-EQ}mQ(LS@o)SAg=0%VYybJEw
zP=XlBsU6$gZ${KA&KzyoqiO>XZ*FrbkY~XwJc(pSy|APAQgd(itGd`#`HKkxYsQ)^
z`bL#f2eXwCbA%ZV=t76*u~)SB$n6wDDGrseW)i5q7u92f%ylu2EdFdliAIpSrp{fS
z{6gW7+&m>>PVu41k@<fkV-b;aIMGSVIoS9%8ytIcy-PU|)z&45zS+eC;;$pO9s#-`
z138|`n|<4B^i9Unw(Q}BVY@CP=wjDgiZj;&07?;2jup@;J3+Sr<IQE2P7t8%Dy@CH
z^EdqUj!fke5O_Js|MZzvB`-U4(#o?+@Rmc3c!QOKIxoku16_z>85-|0N8v^sbD#p+
zyaaQbw$M!x%W#x;EZ!g(4AMctZPTd$0cr-}d`R{azbO+@%j3PlE^HKe?W`Bqp~+Bb
zq_``98@uiuZ|Ez2`zyoQchMgE<dv3Pyg4U;%G|c_`NUZXj2QpRfBBa`FMRz60rY$K
zy#~X(H`CR{>2&t$`Sjx1chl4FzM6jjyMHr%_4~h`F3w&kolfWHuj__FUeDvP<Le}T
z7*LfsMtz(@cNswB(SLvNF)@$V?a(@R(Om=jM0gh%{cppYzAKD&w)K4Q?l9XGjJ|fI
zHQYt#Ho%5!aNFNhi_w4Uk37RRo?sX6L-4mj>+Z4{W!vgJ%A1~HB?Fq?&=YLq0i*s}
z4z(d0<ZJzr_8fSY#cre{Uy^mK$GIVC@xeJE_U*&C{j@yS?pZeK-GT#MR=($tGV;(%
zqO8%!A`5@1S0_T7b@<G>qb<3aPF0&(9-U1a>%iYg;gJEeKFw-=Br{)WlWv@Pu4(XY
zE2RxPE0#_0{%&Q`H?)Ooo;9}%9!hsk_L=$k;uqdP5A8YgY`5SQT;a+q-PC3zwx}#|
z#;@6hM>ovQ3k_yy-NoiQZ%im`(g64vN3|RKL3d0eY|1ui7(LEzN;m2Be(n2ulHv{+
z>leCh)?;v;to{>#ImH5N|Ha*rw0C^hAyR%y$$c7Lhsb|`<p<yoG+%naeK1Q^j*bg2
z_^3R;Daob)^FR#0d_SYr-}%3<{9|AxmnSwqCikM%6^lPtNpFJlnnB;&izC;gpWeE<
z$`8?bD(dj?z~xb(OHxS4_@Ax&+IZ#R$;puyk@(J{2<(f+Dqfk$0uv$f6qoM>$p?NP
zK0Nm81nt{3&fA^1zeoA6&o8H!uP>$-r<W5Cp%Di*tg#*JP1BJOUL|-b-AkT8lMh~b
z$O6?-FRHzHYx%|Ho9X=W%1?q_`IU)kN7)}9@&8{ee&xwB7VL6QF;C9%s>K8GSaf4S
z?O3|xJUK>v{U2bm+ph&M78m(T3V>Se%h$uh`_suGPnJ1E*B@Aqo#3IHZyot)82JJ0
zI=`+5*hk}Fe1`Ej>Z-6p0-LfGjki@f4V!uacx%3&fGgN$GnTQIn!n9Nu-(_qfW=LA
zH<h=BA5;D|h@TJRYlIIK-v}S9&Xv}Xg|9#TbHD14Q_%xXenRIP4S=o7TzNVT1?9?2
zdTTio2#GZA;es!6ePlDchl!1G_MGS-a%B6dDR|Hlm}?t<yvxlqbjnYG*~KAu#KXZc
zmy1{GQM{A~DjS@zfKi$=lx?`evEzfE@{!94`>HKA4njMBu#3#>&hg_EgsfcvRTC$+
zHZZ4jIT__Og#ecU0cxg!*qDKSu^?dk0?57pRLK6KD;!igI$2n12U(yJGbbSjLnhtg
zW9$p9ME=!;f1QE<_~X|f-Ijp;>mujHEAhFIaTCK&jG<6=`al$Kk{<2BqS@R0Stbg6
zqG+A`nuR!Eup@3_tNPLw+T+LJXqPcX9+`a#_xcQ)jvf2Z-ouhRPbBbZ-$?OMz6I}a
z;sW5q#ZSty9a#HEuYo=)eS@+kj^yyDyE3J;KRDopNMfQNV6;2s;VcY~SLtR^Qy?2^
zgTK@SWT};Nv&*+~hi<kxV!@M*T=F`uxpg@1%oF9}7JcXF1Su$FZjV5;psO|(#dnUs
zY7gy7BXVzpOu9s89VS@jOI0ziiWI~LdxY3{>?xUBK9NMvnsHHn$mkpT(#MZ-CX+5R
z?xQ0HJTD3hAvDJXzvT1chVzuwHB9id2Tno~kJ+=1c!Mqe!`4Pw5Vg0HvZXw#-NzM;
z&m<NbUQN&*I~I>7^b8B0-=8H0cOY!eF#x&vsi(vhqT3dIcs2-b$RH?Iq>cf^B4W|r
zjR7C0@N8hb8VIGZ93!}LVkvywzg3m2(>~bas^skP+j4LCZ#!-AUAi8RNV0P-S!3BR
zkpaDruZpay(s5>AWmMg*gt6Pd%Wr@earhANV6To@64#CKQT7pA7qoE%-izsR2hur~
zR<62RC;l=PtzrH+rU7I?N?7thIM_CHB&{39VC_u1d5$nx3vnDxSbvsAPTY#w9-)b~
zWC1G`9=3<BE8TH)R=91DQ6jt{rZQQgYDx4qq39LcJ)`o|FJdq@Z60sA@zlMuw2Nc1
zsA#|rlGC>t=H8d@viG$~`A}u*|ML@Hf8T*|k@1j@&$tfsK)9MNFJ4clubxdWo_{}m
z^VRRCzxk_wJAL!@SJRuTi|O^t=aQdI``id&9MxF!aB3F?9%C!m#{0VhV{80W=ysXg
z=RII;yUlu(H>~NK;4ZqYGp03c$cxu7%I?~>tv|v}9+R`a*cP<zR{a$Z*u}pI+U6#I
zo!0ekn^C7><TaoDHQqpv{Dzz4qrZlkuH<%Ao=al!t0t{G+Z62L1CS%XaeuE0C;Q?@
zk8AmLn-O1V=JCOSZF*y(UL)PQonMnFas4&z`LWs8^#(e;(gj!AXwOb&ZTF*{bI#bl
z>1&=<W*zgO*#=(tYo0yLwCx5t&_9MXtu)V%z)=|JBtDxy%mXc*RD02}o<*B>GmX$b
zi5LD2x?AYG^UNw)XvQS@rrwgGw>xRaC$#Ia>8oMnjS2tV&f(pp8^vw1b&5kWJKRXt
z(csiRde>6G^WT*%iEM!`FSz&Cxt=}*$W!@CIcs1HrB8248uQ!u8h*K-#<+N^#1moQ
z@09TVR*9#F#Fs9w@|%B5E~EaREbf`PPsQ)QJ>mWZ|J!xtC$=uU0LJSIui#x>`G3N=
z-{vx8>?gu(U;ZRx0qE%Pa606PEA9jMW<|hnX;KV4HFb1+>;)q3Gv!C`@22CUBfnOV
zZ<RbP);_80{{DmfhZc)tswlBiuf6@e{`%_b-Sq10a{A%rtLf?Um(%OhGq=;>;okK4
z<Y;=NL|fqZ;_^zdy!9f~@!^3N)biw)g4D_7N-`F`dI9Y0;?nxml|{HC*&H133d94)
zl?6tba{n)l>azR_L_h7tf?wM`nD$hbUz@n6+)t&ceyYpS-o5GMP%$f~Pl<<+fbRdI
z3Qyi+$&ytLjJo6teqFCA^NU#P%)nqQ@O^hgoycb0r95E6hu}84;U*t}&2l#45udEO
zX9ISXw+rHKw~b}*;qraU-v#mWVSJ77q2e3igVnjx8ocoJr+@kv`HwGD>gI%Z`f%a!
zWUod9q8#F(m}JzYUM`~Ko0FpLZ`Oq07<Q4{lUaDNLU`t9LBSU<IL(%>%T#Vj#Nvfs
zE{B{hMd70qIsR}&!$uBBr}<>&6<7c-y0$7F0Os<9j6NKjX<1pg87uL?j!(dzTbz8^
zIRmR23F(+g<-5RjLPZL`5OoMGD+T_DvEhMEzCCI44f=|?Pns+|;mbF$rH0Czcl?^f
z1(i>M7k32W<(C~wx^TgB9|$pp@6yAWFRIeZFF8<4Yv9d`W9VUyP>_W+KGLx){Ei#4
zF4wY2`IMJ&K)q0vt;+NGfGzNZ!h-}GZb%Y|q_$IOjn*CjE3rP>7Lgril0q{OQOb5S
z^dSndc6^OHZ&Z-uMJ#xZ0XtAi1l$tyKI3EY;f!_xG2fAjMLg+egX2|7L!~_L${#>~
zxgl~4r#>Qa4&S)cAlXcqEr8Qe<D+@v3TJ|b0cE9*RtP=w<Gb6#T8_7Dt<n3!_%vhS
zg(%zsr^jOrESP;O`G&8xD6#fdif;7VE*vK5j3|%asupcdY4Fg-NBjVk6YuO8IR<RA
zb1ZGg&RHbg4dw^cUc#0eQA!_?G}iH9XOS%)Emc<K&c5|2|Bgk@miCBUTQ^uwOqwAt
zErlmSAGB@e92+rva0!sP40608Du2)*_%*<UKqu1R=*vMx?ILpiuqa@&=8?w_JlEG`
zc-mLJm7@c%bb<6N{ERD(7wG1*vn@OsM;AeO?2(gq%+@uAm&WS+VF&u)s~rJSSo~@e
zZA&r9%^>T&XzOQZ?u-|Sp<_;k3xo{57ZbLUBSj5)v)Ea=@B+eWwzKjL)AgHW;7DSy
z;Tv}8HOwSt<g-rfnl!68NXEXMHS~#o_eK)XInZ`}hA~veSPr)s6a;P(uVKw{!6rtM
zo*S#8(8uR>wM&=S_H*BqZ%a^(xS%b%UX7zzx}<htf($M4!kr+qnIEit1h>G5N)T^p
zrR>SuPspF2`1%h6YCO*gAZK3HbOs@R&dz^+@zmGLXHTbZzWUqgi{Jg5>4)#WoK9c-
zFui{D+&5|RDjIG+<HNI<_~Ke@pYrDeZXRtrA>;ctt-C?{+LpG#HocpYcCjnln_z7_
zrdwfcQ|w(Z>fQ!^Es@}^`lIi4x(U|0jIjPbG@tSYXuS~}Yx7IidRx;_w}zIjVYay;
zZG*eYwJXkb*yUq`Uz0K40KK7Jv;GRU$wuB<2AHnSxvm$%b$(!zUrfQM$9_{U$#cSi
zw(dZ0<1cu(P3Yj8Hr13fEb=;kRRa3E+-&b;*32{i7TU9hfhN&}*LZ{7y6hqcx)K5L
z8F<@#n?Kl(_(*r|G~>!GSY##(U-Otile?{<hveqp;KA7Bk(tG|(I-n;QH|ol+vcCd
zXVEcB6Gd_hZ31&M*_Y*&Zjpz*lo6Antux9uQ(UaZ30vb-KLr{#^tPlCw#hp&Ulz7i
z)(Iu&akfhSY{0C$1zOg=?V|i|O`cC|PcGPBY{>0&PS^t3?}AOKea_AV3}D*4+8v+V
zN9F0S2luX~z5AES-zbDDTkr;!`+Iq%pUnn47Ie68<$DyOIVboDt+!X=zj5;YH6<3s
zkcDRlNfjF#KV0X@Ef$83j*q+u#$pjahRwAy>V6%e7moI(w|>Hmy9K=3aBq6BPdlmr
z72>MCWzmcUuzS-Zp8PsF^!T7MyoQj(^~bL(yttTNy?Q-;|LoQD^u?>`jdUJAJefZI
z=y*EZa}mA|i7xd%+~4=Ym!Hl$QmkIA%bTu{^90z{+jr7ey5e5zJ^#-av9qYf6JUpX
z_a}wySXop%=fA>aM@-x!^^;wSmvd(p%9f|`=o5Nn@BZ6q|K8i_SdMr?jrzOStuu_9
zna1bBbM8h7tYyLqU+jQUZ`OoYM8;m=y4{*yV4;`T8?O7=gl#<E4<6vU&h>Htcx%48
zX#=*|-2`agWMkPKTyBf){Q&#)XZ-vWUpK=~i}N~sD1W<Z?S-#@@@IdcFZ77%a^)0i
zQVmu-=kQuQr~<wb&Y>uVZ!B83Z02JF85#bZ$*ys1IJpWh-vDTX^#)|Vxn*)V0A(dl
za~yns#Z%0}vzE`!py-l!Y0(U2i|9Bt4_4I|acy8Tw5}ZaL@l65>)BCcIwQ?;+YE_}
zK&-AjwSXpmay!tcaxrsrT|Q|NE2TYLp_rS;f=8H~IC2vJTR5_InIv=4hz8>41_KwH
zoNyxJ2`ZN)65sY!i3Keo@g?A-ZoaNsdGbpMJA6sz3AA;%sDeV{B7TUM3y+`tlI&hR
z;YGB_T`oSkXuIs#bMhx-+9SSD*G}z$2N-5^H5DQ)3U~7&g%&E7%3^%6DQ7GOEH!49
zlAB2EJaty(bbsaJ1jj}Y+QoNxj>8bm_5tPCgoG-Y6Z_EaKKlvRda6S%u8D>4&~Zul
zmRj4}MBrZsQFm}I+EJS2HnP5~Txfg)>BI6{F=rQ@Y{(D_F^DIeC}_v@9bER|8;_wg
zpz~Bpw4B(jb;{awG?$zAky;KtB#VK#w5>&qb?4!ML{~YuggIBG?i`qDM|Ax$SY}<u
zAc#KBv;&-lpW{=0)F5_gKgP1nzEjuY;z0Dm0UTqxwAx;ATE!8Y_V`A`F@`<~nR)mr
zuhovpt3T{Vn|wQNDS9CfgIF~YTa!f#QiJqc$}O4e&L*gMY-?!0@t5P(q;t2AeWHZ|
zg7`9Oh(DLv?!c=(!b5><C{J~FYgm>XUZmKf#BO%#Es%OPmXGwXmkk!QO-ewBcH1a-
zyMtRT_Rwa<$DErdu)jqH(9h<9quRO!IS0h1>pr)KL?zWmzbA)-lH+!L#Db3q4QNy#
zwRb+WYpJEYwl~u|1Sw1Mw(t|%Ay$w$W*1fii*okR)C9WoD$B4~^td+=pW8(q*nx{3
z%PN)5f$0Pdckxa0B4QtR$fZHuz}v3yCJPhZwG-8DLrPW%46BzvgBhh$hAtnZm5Sk*
z3qWQ9G>jEnL?~yCueE%v!maW*#PDf{u>R%G3t#_%z&HpRZyM-99$@^wdv`TmUA&&o
zPG3#Wo_;rd_2uuTFO{Bt|Lyer`P1q3s~0lkza{QD`JAU~ialuzLVJF53-q~g9mC&c
zaINl8w_T8P(2_nRysy8X5<}bG#eUS8ZMN8L)0_400Bf60YuR1kHvOCAa7MXxw&{)X
z0oJ^cUdw>yjj3T9uVIvN1UBwtR}3tiir<MD0I=rw4Z_jJ^q|vpL-Tvu+Vq$RX2G}l
zb2jj6&LKXc=2L{{{j*;2scYe}EB=Px?K&eqC$@q1jA@`r1apnI@^u?c+UA?I_CMPC
zN4P8ANT<D5WzmuYJ@$!dr%ju64CYlC8_8VE7oO=+$Gk1=xK*!s#(=`iV-#%aciP48
zv?<3Sz8XfElD3`xGCVWBir2aEga~Wq-I!d~9d${&;;N*b%P*%Jwe+`}`u-Teh6*vJ
zZi#{O!|3c9;IsX10~ybg^3Afeb4<iC@~@}c<eip#JS6`6>%P8y=V>q=VMzJ@!Ld#t
zEa-3^>3ed3&uP->`(*PQm@La#)VUQy$0zf6boo&`3s>*Fpmo5)7Yjm1JQa25#U&QP
zSfJv6u57C|<_Y~?2;yEJi$?hNaqpBj;P4|RV)E-0_wG#}JwBO^kB{As*z*5hyh4%6
z@(RaGrPK4b(~DQHrx!0@xebmF9!!@Pr(R?`J~>f2{iNG8@gHIjdFqQ-8T!?S`-xM&
zamuR>@yq>C7Q(163w$h+@ft&3m&pHkv1rGl+UePOr87T0Mzgt1D?#TW0Q!OZm+B|A
z-@f{8e~%H$A|DZ^oq^3feNVpduIIp}KCs|xy;%c(5xaB?%bw$91*6V(+fBS-8*jbn
z1zgwLET>@`Zw+_R8R4ey4EA+Lckf&7b=Y3#KLmW}az2zlylnF{Kk=0t*rmcr&>1|~
z@WmrvW>J<kSjTc~gSQH&2M$y4xYUUaN(#AUZ12%P3A$137}xkWkKCHth*LalT?V=^
z0Ln*#O>OYP$3}UvN}c`m)^a>@0@p>Oe6Y(+;n;_zJ*p56PF}5o@AyYZoT$ODwfU8y
z1i$>i>vCMQ#fb=+b?gr$JWkK#B*L&=cKQ^<*z8brI6BW3UOScTXr;rLa5+l!L2iDF
zBO-iBSTHjcKDxvU_+sY;B%~CSIJg)&uBiFM&IOlx=P5K6&6J^{Jk0@81|{eVE#=xD
z{;dzLGPxMNdm|PXUyCw6sFz*9vA5*n_l4iel!<QCTt>zRxE0VQGoX&XaV3}0_#8Vb
z@e%0Awa3<l!GW*1<<zZ(z3WMNt%C_1C7YDrhzxKtKE$m01h7$Gh@?m^)^3lXEy*b<
zkTHO{CsRlv#Kdf_yNJ|Vn*}0lU3GLy=9p1|C_NzKCuIkVjQ!k<Lf);~BYJRJiSWd2
z-^yYU8q8z5!E$_+zd8Qak2K{d*GXq~-Fl`e)68nq$a9Wmv8pDcG-w=R;aHr<g?)4z
z#U^-U>O1SnLOwfI>vYl%ABO6CwwA{g<#P6Hy!h6UKY}+{niw6Fa+I9PlZ^my_SHa2
zwl$~9Fb~z%1+~EA36C%%i+!J80v>^;Vip|V*1$*c>_fZMt7Ao}Ja`;Ig;yG;$XvE^
z_>>A^sbwW-+z9$+L{^fZ;M%BU@#pq)Ry@!d6STi#xeE-@%2}esaztAPPmXV5a&L-(
zy9FQ;vx~6-Wk9?3@nO~Qf~!3`W-~yAyAw!>vqa$q$%noshi%bfn=*zrswkq0j+xFA
z_rhzf;4?e9cFt*!>UW@Xu$0PPIr_72c9D+wED~(Ahw!wA(2}PX2t<?)CJ^?t!7ttJ
zG!*BUqu7-FyWHZ1Z4<|D26C$z$guN~H2C)7S_M4YI*UCx>cuoR=c99_rS7A=@yK9N
za+|hqN0U$DGbRL{4B`FDzx;XO>pvWz|E~kbNF*L9PK;Y$Zw=Dmhgw(X)A{KO|0mct
zU;dTT58r(;Js17@<xA<v?t!kVI^MY6dY&p5K2`6493LE@DKiN^QVctWpAL)vCA}|T
z(YBt04D<-Id`M$6ARFnm{WTc;tkX8wrq^i~ypQfS_%XI<S$Czs%g6ipZt!caw+zDe
zHKv>3Hoe>A0Fs;ZcJVg!hq{ciSw6rv*~r@^>myIpwQ+1cQq#`Oo`HGdrhf359`ZGx
zL=}5+s;e8>i0Ay$wDzVWYdU%aY@U3XSH7bcXu5fxKiIZ?wN3&WbsN8l=WU=p%N%Gn
zU~F$7^-KzVyKmCkzsOD6zBY4UkuSO}AL(Lk@Io)~I%dm?t~zemv|a5sWqgWOpQR2f
z|Cp`pY+C~3&PN?&*KJ4}Wo}7v+qBiM9V4BzbJ`|L>{Mv<HynF|-e!U)g`WVke`7mf
zC0gRC<caCXqdw%VJJ`^74KsgTuTEvN;d@9KQZu~<alN^qYw0?yU9^2R7_Gn&1&Hsm
zVX?Y$rSB~lYL7Q8sVH%0LH+2PgUk7=oGE_pMLdu%F>o^bCog>^+JbQf!J^cq{}W8*
ziO0ebPeZXFbg-XS`;oBm(@rXb`+JNg{^y-VsPl_@wIcrq=J{FdSYYC*seS_OVDDgh
z_~@Y*toC?~pvpQvX7S69N=<}-XP(k}eRe**Iz7|e@Xd5~dOm&k&DUy&che^yKk-5w
zPlQn|UWrJ&Cr2z~@ruH|X^;6KZQ+Lo)Na1ts5U044lHz`L-}6ZQ!KpVkVUOmD);%T
zQ<Z-?ou2a)nCf@=X1chzn$FJ8r`P8fUg-0W<puZEH}_SyJ@t+kEV(|hRoido&EF=w
zo}<_Lw12~>H*3NxqHPvDu;ABaz$o7>yUA`7?y3je)yD`o$+p3*zV7JmeapQL*W33)
z1OERIWqc@~c-iKOg|B=6=l{$9!f8qe6f91TIdgI23*crBC&W03RFP@-u4HlUn;6>p
zdIC7t&L^;tZ?$9I72V>oH*L=@*D~$kxN-5zNg4w)>!Pca$OOuCDv}<ZVWp@lm1`92
zI|A+G+b8(aCq*CHVnNCRr?mT$u>ca?Y7dvGnh>XDsWY<Duy%Z;By9At50lDlW1clE
zp<jFgbQw{4M?)fo7T;e)?Hsz>%^G0lEi`~$8YAV}!a5E>g?ba(gV1e4ALvA%^3wPG
zG@(z>{I3`|k3F-9#Eo-b2xR9+s#AbtH72vGtGpwZs444>t~UAqpIkUC;me;P_Wf%N
zmns&1-s<82(kE|xzL9Ur=VIf2rpl62x@|F3S(Fh+^6PfS9~T%9vkYY5s3F3v{V+K)
zPG~B}qsH>`UCP$cK4V2xIc?u4^g@Q)1a<9j+s9YTX6eiW9G8^e>fp3q*&xOSaqrcV
zs;5!cwf~Tm2?W}01^8lHELtqw^lv#(S*fj|bjck1!U_VINPlg%N3zseNrY7e9#7@K
z=O3Y=8Y0Kv#4nqKJH*o4#H%H$0ESMFtp>{x@)RY(5GVZmj1Q3tV2RHqSDQ$u_<q7m
zz8Ty2%jA*CQhp&c*PWQLpkh@rV{^|FO~yUv-1xxxd~8Cp0~ZS}iK?VcmHcV_Y()r)
zFTI43oHA?!PMw6#vC^{6k;MbYEP9E-yO<KU6$KqP4D6ITW|twt)iRH=ps|-<oLRAa
zAcL|2wvju-4_wY==ym+s7!!6@vY;`xntfwjahE!IMA0)X#PR#QLaU<82Zh)%dGEz<
z?-%3fv$vT%bL_glKs=IX?D-=Qw+FuYZuw}+x8sZ3<nT$NF+<nKeR(6Zx|NhGhOuu6
z&Y5?`vXS=67_QisSUE9NM6{2vhEzQUuCtQ+#{4P=MS*=QPgRjLq?AEZoBS24k3W+s
zaG|yLd1YZ3=Qz9PYhsT7oIgOSWbZhj;Pn?Xi%YIO*8pYktp;C^c$y24SG%yIST}yx
zqx9N0U}tJo$(Vn*%`L4`RCzZmG<NhMaxpz8ik8~D^%h7~wWE5je|#jS0GNO2X+)V~
zcImdEB>A;X{JH3;DlbL>m`BGZ8xOMA9^bQ2`OqFSi1*+8xBqP9{Gnlp^k?{A2d;N_
z9i1LKBV<GZj5Gjh=|Z2c@AK-Mqv_<)r_-mO|7Lpf`7fqVKK*R^{1?BRKKbaAY47k@
z<_FS!;BnmZ(3GUH%0Qi={ih7iALp|F?!YoP;as(}$NNIh`5S(UuQe?5{oCQkmis<-
zW19iHY~hZ&XhYw<qy^Sxyic2~VOP8jx5*oB(zC4U46MgT!~522TbaA!So0l=SX+H`
zTFdJCzrl}U@Ganb69e2B-z-{=zhpzX3Vb|w(3WEvH4JC_HNfXOywL}~Fl&}Gvx02%
zvF6>T14ew^-sHP%GW%L!)T?S4*7kh&*#@nbe9>9kZ_=CX+0ak2G^~B#ew$xkgJ|30
z0nxjzJ3<y0j8%@#^|-`z=5|&4U5jhOel6RCrOX1Cx68hJB>AR{zJyLZEh){}at-k@
z`e~p~w(+k6Ls`l@%doF1W8J5~W~~-*`IxzY1^T@I(*eJ<*Amm{I~a3dBiF$SH^(;4
z*?m4%>SM0^iG?F%nSe*<E&qje?|6E1?~taQ-n^NnOMUTjE6)o&%$cyu@kra^&*vHX
zgR+$SNvr*Rr#*g5Bp*3D+?$S%52q&|Kb}7Q<m2h1j~@HKyO^W#1k~%-r}FdGHY|YA
zR&U>2c|qvqYo3z2^uksC2aCIl{9l&laMFMC=5o5aIQMG?Kl}8_baH&;R|@(`uzk(n
z-n^TBc=dXE_Tts_^3}N)(TIOvU!;EZ<k9qx{^--`?b#30%jZw0U;gnQPrv;2Z>IgD
z6aRM^i&u{y9xC?zx^MGtdiDBpdiLsKdUdMU&M&9e=Wo<r+!NyexgJbM2TJ>Ue$C-N
zvC8gB&GhW$>*<@P&!=ydc<mvJh%AixwTOp%Zlx>vp~sk?zEg#-PG3&%E?!H<|DaI@
zt)71KrW{yuS|&K<_QGw;f#8enf^YhprSOVqy#)`9ICn5M7K{rPow4jd)4TX<SU>Hx
zOD}M}JYbg%eA=#$nO%U+oj__g#fNM>|J|f}H`rd^4+(!L{~yXPUbcDq?Z5c1@7?>K
z{;&T_x>kjuWPCn(!Y418cGphIF_J8H2^uBkbR?wh!Ql;JL<Ma|`6*4fvO30NQh4GM
zQFu-SzMZL51J!&Z+S(&fwv7U~GOk?@bgRWTb?ODRvW$zt6%M#*)1l2WCGMy7ZBYYC
z$cbAYE>^jjTx^}V@+$^e_=`a~b4Cslz>1q-b!qe7Qu*k~DBM_RKle${4QcUHU*%jM
zi5!}(PV83HhZ9f!X!4Ct09MWg@$HpVVuyd<#B%*;qg*JFA?3+17N!ym?T|$)76rJV
z5gvb_3FL&1f8r>*^=0pqkUsqJB|1K7DR+!a_$0HA;zoyyq6@$e<vRvtQh{y2B_cBt
zbBoY#-R6^s5>8~%W=|^`SC%$aZXfZB7QFD;6#)njAKfY!5OY>*M7d)woy0%T?979^
zK*neT*vf&xu~wZT7hV~OGF%&15}k;wg3Ljr<dFv*MkKW1{KbtvlLnb@_*<o9u@s;K
zR2yo?b#V!uxSbi+b%)*JNVG4NoLNGeRgT;b$V59A+c$0rv>aJd{p<0Fju*bFex(DT
zlU?Pot0ENX<Ho{(%P~)R3FeDuM1quQn2kZ;%hINq8ZT<X+)sgYS(WWFK?0D5;`R6-
zH`H-wzy}o<!!a9>2e)J<wQCi8!nCU>L}gY`N;^wxy*BT*FBN6+we~}~ju}SBEZPf=
zL;n)KvH_51guu6~33l?)`ji2Xp#r1*jxbwbYcv+{(--jUVQ6c4iki@(UACCbCu;}x
zjVwazDet<@$2h(zBh-ekt<Z{t&KyHQP?utC1_U|9CyhQtLz>+&TCxHuWN~MyQvkg7
zUAGb1`2a3qEX3}#vu~j#GfuwEZ$a81F=wuXu2T0ArI8s2tT|^ub;Sav%cnN-S>MVq
z;vHytinbC6+yNyX<_)=u1=_<3+QkSnuL>e-Rzdo1@oB6{6lKRSoyV%7@MbCfH~;NF
zh5B~|9mvn{?-leA+17o?lr*PhUi)Bwe|q@n(e(M}znGqU^2_OyU;M-A^I!cZ)5+tH
zrh~%+&B-6+rjaJ}!@Vm(4ah2k<2yGPbHihVo6Zq`X!t{xJ<rwd0Jp_)=S3d1t*>!|
z{*SHWkFi_Z5DT!&7yNA!_-(s}8{@*?P4X>poBmz=y-%69jf)Lm!f*0@R~fj+#U1po
zYg~`Pv_9QK)QRJIUAM)Ke3vY7rN8-(8ZOW}UEXfsrt@_L`z@S&DKq#;qu?9xD?b}L
z1#8~zI=kD@I|o@hgIdRa-T(H!w0ZH9cUt(scH8kA&vu==FND#bsn0c)Eq=#^59D*b
zea&cyUt@y~^QO_K%dERU+hAMyBT(jC{>t~#BY{m>>uxI>{-TSQ+u{Ij(y?ri^Fzqp
zKy26J>iRMJ%<;Fb@7=(y^1f!X5pnG{-fZ!0u&J@WY(VCFgi4XH>uFdk55C+gEVNAg
z$d-Qu+ndhc+?%fM?N4t!56}HP&y#8qjNzKvyzw}i+fq5NHRM%>=yGq6ZxlQcg{S@f
z`_m)Mvz|Qpc>46yPp6MRdMy6C>Gbuh>HPeBI*{$j(V_fg(TfG57q8E!=PypD)ALK;
zKY8##-%#{r{8)M)KYrxb3SM4ZOlN1ZKiHSgM_vTuJk7!ri)j~EdBx$==dY$0r)Sc=
zn654_E&tW0kEdULdNQ59{9*d~tKUzL9v)5q<$v*?O@H*8-^li0+H+Z|(_R+kc)j7P
zvrENwH9b>$$&*``idSXOPWwD<wl96@=FLiKS=sXcVb7kQPG5cd!}Q&=7k=ss88IIp
z98UYn{g8m{@x%QG7QNnG@LEO|zSPF<w%l__pDf?i(%-Uk9{8yXU#444u+}pSIQ_%n
zw_f;K^jEe8BkwkO!yoEZhsbC{`ld$pR#?yPcY@SzEAPj{_WJ#y;X{@Ahl(Aq+dQ-I
z_25tc?9YFj;hh6&R#q?$$af%`2ius?;(Js;>7-_vaGxkkhh(Kt&SAAE+04LCa1QMS
zY*T(0Z)OHFv+TQkGWclwgdv6x6!fA)*7dYm(Jo6Ls~29?Fklk2Jh3C-Dv!fuLoQh`
z5^WVN1jH8I0OfN)>zIo_bR8oFlqUWZ7gUFpn8hR6SWuaL6xuhXy=Uk09H1^xF|cQR
zMAc}b1bM4da6~LnA>(4eNunp@j)8J2Mw^vyzNj*(WJ@_zCsN|DnKt8e0JicAPMG#W
z%)a4Fx%i<?bF%7%v@FIEbtN2EoJr^<)_{1R@U#^o5QTshMg!i2+daO~v?7e;vu(6Z
zL@{hglKV{TWR@r9qKDW3o=ia(jp&T!^uBl-P7BpeVHd9U3YKD&TGRb!RGcsk9gd}i
zgP?*8--dGN5Q2IK-jVww!_f3ka8tBFmm5_llyWS)2k1^iCVA+TvcQuDR1pZgcYtX<
z(^gMG#VLL?7*F8}K<IOCVc@!NKBt&^ooAKVeoAPdysjF#Z6Vmz*mO*kr7@g~s}eHu
ze5}`gO&1^2_PZbhn5ORq)nR0AZ+y)4N&Ui4JxqJ&PYltAC-4_QYhnZ_M@FmyYuUEs
zFsy=20rK{zH1`2~He@p+tqJLbZrEn@`1bhFmO4l1(2cQ^lm;}X>!7VzwlHq(p{*^k
z6%6xTS9pvc`V&gDZN}CK#y*r%wV!2S7Zk5>+d<5Mq*}8ELKU3$P?DbYv-788un9Uv
z<EPup?n?!Oxwup@Jm}_+Acpm&U|AZ%0{f0iV1p2V7jG>BsFV5ZfCaG(+VysaN9KNk
zhrD3M23|9*7n3=LoM+-2=F%Q1!%)iM7p@!|KU4-j$jIFuWdj~zgS7(#+IZWvyBQ$l
zwCg+S0SkA5lpU-`VzZEeF6}{mis!xf)X#M7lp<Mpt?QhR>rWnJ_8qZtE6otW&;RlD
zGyHU5Trm&4n%=&-n9fgMPS2ixGkx>*7t_~YeKmdc^_SDP-+VD$p1qtN@M~M?b3NMI
zJMc{%j*(5s7|u0R+i+6?;8^qzB^cQMD?rNMOl!DJz75u`@Ka!2{+Kq-v%rd;_1+UU
z%D5w}^&8*ir}?BYYG}`KR(zoKw?fI*=_a^M?>0Gblg^q4J=Uq~xz&CpTYDd6YZ!T>
zPUE8-*x+x-R{HaCywsy*t+%EZ+^AujY@|DH+?7}IH8>Nyfv2o=XB#n{ulZYa!^dD#
zW?i;5Z=l0l)69o*P7A&A8y3t;g@Y6xyK3X2lQ@5lwDh9nO$-QUNu^`nl8-68PRqGC
zc<a>svFTZ7?Wg6?;LXc<Us%zRnK#pI6Iu83C}#^Z{B`>vAe;49De~PZ?(G208})&O
zHaXA>-?_g;SR>n74zhwp=Xx6D{!TZ@B6nd0o3fS*8)#smZ%cih$Fi;88*GT5_OdPA
z3|Vx#9v5)UL+Q+t5m&Mg=I`X<JkT7QC9i&}2=DjrzpHsq=05nSQ7GA-?YU3q`<2@8
z)#rHbzwu)s7Qy@)Y0=!9;|J$wXJ^wXuk^d%NAmZkM-NY?hbPC=f&A>LyyJsCzY1~x
zfov5Vm-(x=Z~O{C?0K5YKW>+dC#D`fiaoD4<S8kRzy19^FQV~^K^CES3hd#@(R9Eo
z77y~@V*E!KHRZL2_xJXvPd@vF>+{!t^;grEzyDkRC)r0Ie=;2(pC~5TJ-8<uu4A-K
z>c~P5{o3(eDqRt$?=R+FB=;JLn*}tUR)cqT!K)50y|BlE*rnpWk~%F-rD;cA)BR5K
z(tB?&)Hk%hOzA9;((7|1iM9kf2iCmnvbI~~rbhjs*Ysv+9Wvn?vVw(oUH-mw!`<|5
zi|0BZyGh<~7oBdCU2rEKH^SXw{m}5C%KSscj@NCT{U2X{@@IekTMnc?bQHyWGlQKm
z6OHR2F1n#EDMd*|KK$ya@)Id1%ieSFYA;97z<r$RglDO7xa`1XZypD}aALG1*3h^!
z&IVC*^mBQbSxPS73@pMVhq71GoXATTmh<S_<2xow%EQR6bdna*Cnpo-le~52edIuX
zL811La_T1;W!YYMw~;PPc|}vaSyvlkv~31Y4FA%c6LpngD1DwVGgKG29I55k8k~)Z
zH$D?Pz+=A{2$Q}0h%1!(9SzDx*9%eMIHz8&U-2l})sca{IXw9w$Al#P0DRfer()&5
z#KNaMba|R9Z6`g;C@C_R>Tu%EZE2*mx?Qu=o?LvI7+KN8>~<9CS5FD=eqqCV#&nr<
zW|inz&wWz5Abc>!m;KR3F=XRI=)q%~GL=;dRq~=KGciZz;-Va`<bE<wzy%j!`wFGK
zKB-gUUSjU{5Wn=0I48F9-Ml5PBrf#5b)D>+^9NbUT-*$$U;O3#AeQiqDasXVJf9f{
z(VmPgNaRl9*<W-h--vzKVwMSsi~^yA=Y7=()fy_#x|T&LwFicA1Eu&W)=ww8luRh?
zLzKDHaW!_c?;IbpLCt2Rvw<fbc4ljv&v9(vE1C^n`cuwlIcRW*`3V&37LyM&de%iJ
zd@2Qtano+m;~1ZB29aAid)hkv<Bz$P=;QL`h_)_m@q?9#`LrhCS<>>E6UB__a2ryG
z?(Wied17|IpkP@1Wg!=D$a1^@q0yoo@L0Q4*QGe}?IKDF*AnFN&GCZFI7~b^2c$2Y
z&3?_c*Ft4!mx<G|D#E#t^g)*{+=ht+=GY!K!WS{a1{X&ABPVlvz>BJqJqfI2=>D?r
zVp|V`Y7cg9xhlh=_KjGnkKdGEQ8GTP3$7zAOyhuM_7^YaqD!uQsW1IlcJL9vPg}cm
z^U@x1;@3AKPyo$K6Lguu`P5DsIr5@~r?2HkCUu%f47Mp>LMg*Bx@}4xByLTR%9^a?
zAdxg9G`nCRE!{9I80@KU#WYJ4K-!9Y@a+zRPE2Mxh5mWr>u2~WupCDVz~VUN6+7pr
zucqfuznY$Y`@8A;FaLV_>WjaczWe%%>EiOj(^0Is(c&9Bv#**9LD%&@nhlL}9FU^B
zEtU@ryUOVDcg1m&oFBUN(fRKLIMLk!cE!A=x4||$V3%$4?y3*8Y(BG<Uh_v8(7e0A
z=wr6q(rz2_EilT*G{QF7O?2}bh-#;N7hf>)Teb~v8W?$_jQuvgkPTYcOfUSj>p~aK
zyxnx`TB1Djj)o1J0dD~J&DE0kT&MXCzVyOl?y>1>jfY<F(7+}y{58GlC%ncj-jK~{
zTY1VrH~2++)@EMLq1$wVFLdTYBxu?DM9gQfwXErsVU5B&ZR%~I!`dRp&**os(NNi-
z%f`N(FT4%C#x32Hujy-P6~j8ke(^yeqkK%34gS`)3*SfeLeFW@?Rjg{x23hN+Y*PB
zmUiFl$60@64{MW7+6CKm-!Hw7kKMW)|Km2E8*WW(bzX8BaV1as$*AwdYXCLBnXab&
zdvClL#exlUcjiYv>tQ44`yK7mGJ$i^{sFsRNw_b04ocqAbb0y4Pk8a#L0o<O_~G>V
z$B(B+#|P8Hqy6dO@u6%ac&D=Qp+w32R~QRl<j^c^otzv_AAfWrn<GEng+JPcC$D&V
ziq{Vw@cKdiTTFg=8tmwRR}3EbNi61bXP1+oZhP|Nlj*b1KKH`hzx}s=GkyKdx6;W|
zX#6Lc<ElBIzsvaA%A2L$PG=Wa)9cgA>D3uel$}p+uC8<@Q=L(J`^NWkScKbC!k(wj
z&Q#77fAR8``<XlxK*Pzmw>Q(hw`w2WqJ@)m&LVdVW*8d3iMQlAxAwUfnDY@Q4_dr|
z->^fcVMAB4O?r`&?~<*7wj5!*-gSQsyZ8XIo8%35(V;1BE9<Ah-D3TzKpA8DQ1FL}
z9k1Iw&kJ8ExDp?02Y1GSXi|qB!aDxKW};OL)8b2~WkBCFTR7U=Y4n|Do9w`K`YMJp
zAEOZefA;?D+mbCi@55H+IoWqNGjCfRPyk4<vt-#GYKE*~k72*qe(|v6H|tLj09r5B
zlPUix7DW<-{4PGqwuUHaNR~hVBtbL<RH3Qv-nxBuC)3~W`{s;TJI+2`=DoKHM1E&S
ztl4~XuDMn;J9fm`=Y++BwRHpZ?Go0_3EG|*+JNrO_K8b&wH@Ws%*9?g&5tSkC`;X~
zhhW_`qT82-8uHONK6h`<(z|&+mpXp^Z3<pqIv<_1PC%W=V24sf0b0glA?tR>nH<FO
zr(e+PO^XLb{@Oopvhk!V1jlj`Y6D1|Z`d<l9(*CnldtnD2eoM74b9J?{P-g!SeL*<
zt$wiP8x6|W9CE-xA&-Cif?Oi@pD2Kh#G8rx^?*ALk)k;1!^V>padvsta|w{pl|#O4
z6SK|D1;NU(7_}Q4h_qzNL=u=`A7daU0b9mtZHcp{WzB4LX-bjbp;~3v6+Z^+MM+*{
z4AQB^v6zJRTX&!Z+18rp+`Me}HbRZflr-igCN_`Jx!a~~hzi`A?VLWZ6sIWxS}Z%q
zyhY-bfh6U_;)t4z-2Atwmshp0%6R#9f>SKJy;rU)$9Rqzsor^cAj8Auj4%0Jw-PuH
zX0^<@j~)6@9j3DCg>E3z@Ppq9f}K{=v<<*^W0Yow*Qn%O{4Fxiw5&3A^HfW6XkGOH
z<|A6IRhReaI(D&Vzp9vdEUoj3(n6qI);z3K-kfZWGCPOzdA?E~VgloCq;~r5@qpUK
zi_Ug|C689NO6YlUeyom(ZCN)*>-13TrQey|HKb1GlgilACZah`A=zxnRxL!rfn>ht
zZjH{`LP_g7V_Y0+6{}oiU;*oHU$r_Ywe<ebEzEE^9&M;BvOBS~j`dIA)F?#(*)D1G
zkhA2avNfJKYkJMOy)VRU3u+shM~93N&E-{080g{KZ_6(1!Z*e3&PVm6#AtSDndie}
zvoAU!<~h`!X<obz*HYV7peF`L_lq$jbx^_Y-I1o2jw7PQ6T8=1`J7zVUemOmp=sq*
zZADG))Yyo}ZPg6Ahkx{s{?Y3PU$60nfPIoB?-LDQJ^GtqpMU=G`0#@tkN1E4qw)Pe
z|I_g&-~G<`;QjB97e|kE&G}U4iWk0_!T+`JHM8<?e)6s=_?lDb<MUq@^t_#S(dwn*
zdeQ2bZCR4o`L}3JV~#rw`*{1}_VF6$@$EzV=3TOF^Ku>>i`m`LkNZ!-Nt}K5t*!NC
zDcfmdY@B2Vv#qb68||ApjjgZ!TV2Act60B3($01n7P-T?+O==xu5w1bwRgK+w)Ulu
z$>tw?lYN%=*=BiO*hRLsT`t!R>zL!G-HqFgX}y=#?%61ZZG-<0w9UW9Xl3247Cv|7
zN2;%1vh2pUObB|~c{y&gX~(wJ`SHv8@p)l^aH8)0X^rvcR%CZ%ahe}q;Tgr<yL9m0
zg)Ye|p_ZkdT7rE(OdCDB^JLpBjPVp+Sx;ENs(ZbOpO-E6arDBov$Xv<_lsKSvnJp~
z?AA0d+Xr1+Y%RMS$&<Q`U-Kg8vzouzr0XD_Kk-+cXwms*m@bTz4!`zW`Lf?rAKiH6
z{fnG@^NR;xhlf}Epvg{wKY#f6(PKX(f_Lxkt?~B5H^zgzH^!~&{Ap?Rpvy32Q6C3h
zj-K)0iw8uV`?CM<*PWX;#{Iju{qXDR;bqqf^mgE1P`V;R9)KN=8`rM*!Pk|;LqCWj
z?~~^*#-qni$J3`z$DKR(#=Gx*V_ds-V|@3!e>{HYcYb^P(I0=;55o@llZ02U`au}q
zDx*Mt(f+xv_ntm|J{~`L>W5j6wY)gW{UD}>`C~i>(s>AX?b=oE3eTP%ji=9^`eC5|
zy$yUutok)Hebzi0mtVZxgRh(ikJ~}yj$i%OW`4wDyv8D{Svh<B3-;Fxkb9qL9FJ$4
z9=Gk|tq=Ehv5yO|x9=p*j59fMpT;>)A7^8CKEEt{S$%wId289i55E3S|K9&uALw(C
z`8S9PoSLo4PH^93HaBK-40Bjz!RHg8B@X=f;U@>Jjm2T9ot=1_?2IhIec4n$wiC0~
z><|%CD=d7RNbpflaq(3#9*s}#CYgtxXga|LS&Mu~Ipz{)yV+UyO-gjKyzn=&$boTX
z*-v2$@K>BuhC@A^IQ^TIT6{rdqM4GAO|+dgEjfGx-?`u*XlyjZnU7iA$`v16y;11h
zf)5bDUny35z#NLp7^#O7cxCmzgB>JyWO7ln{q&_D^0N4UctP=4s}6F~pC=)9vQdA=
zE-Nmm*5oHYU%OVz{{!P2UrYbAelwA%4M(l^(5LuWnjM){n)G1Za&=tn>@#+!^<1(q
zp<@J$VV2@gzVN#aClQw8#U^~>_}hh2WrT~*agMSQnO5bc!sz&tjNHpsvZ=r65Q~0V
zrt$VBv~m+;uYiM&Oz4c=Z;h*RXD<?p+q~ZIJRfDV;M^F>Un*-?PY!XsFBQl0VMai1
zUvMx6V(gSot<I03YEf{HEnMQHU^O2SG1!8*V<65kZgPy+<J$1+M|t`<#(ALqHNNCp
zbLd9{b_Tt;pkf~RR$+J<J3o#QkYba?W5Fmnf^9U-D~B4!-+D>y(vmP%HHWqW^3ay<
z6Jq<b<eVwlYW9Ju%#vRu@k^^+y340u;Ipnf3?Vh4FD5OXnbON=i+;$3>^j5Gm=q<2
zW5AnAIeVP$SvGHffwj4A*ZD{;y-IM8mg7fHhRWS_3rJKw+foC1yfoIGF1+f+>t!3t
z=c>#P+-j8vxn)H}uDHZvVx?-je^kiQ-A^G~B%Iy-qwICCdVeIRUha}C?!(UTEh@H@
zY#xj>a7?g@F}s6WYy>4xs5xL^tST%ITB)(eX_{-Uf*fwI?LErqD7|fLXH|~OQIP;N
z8m&yT5s2rQwzX$t^Q$xuvNf3Gz%GffWgm7>Wa_D<c2|t7+^!K+tvn80{qAMr#kMp-
zDXzzoa9miop%>U%Pl*zp8g{VJTV1MR{AxWs2fujzjjz}E%JKZ@>3ID4hvSnEe>mR%
z(VvVz`R?zHKmN||jgLS4U>rSrI-WhtpYFMOCBI4RQ<e`(7VXX6oqO-@gEg3s<L$B!
zXK~K*zbsxQ*B4jC9;hyVW;t!ldu+|!x15Ile6;1B7T4U<mMz-GdSB*kd@rZr7>{;}
znf*Jw30vRXw%pj-^7(E>`^_mHYfEk4V2)}3!kc}wU0CeXGGlAo>is#?ZTmEw<+@#p
z<+6j$VH*@9&c$AGx3e8=^?45Ei6Gk&tNpXti>;Q~Hp{2k=&t45cs$O-wtNPPd-1KD
zw!sTGw$^LeoBwp(wr-bNvY>0L+2_TliYW(Un#R~m{JZwpPS*MHOLE^Enf*Kc^uDsx
zyPM137(upw&5Nz&UbdKhEbAUaDE!nn&r>g+AXq#0-CQir!+2^wv%sBoIAMpFdui5t
zDf?NzB|n}izxya4X4&=kQmj*aUk;cx`pCVBeM`^t%PiGNfqZj6QWn|z2IQc^b;(u$
zXa_t18%O>I*p<uAmF|VV3C6dQ%(<NVdE6Hyz~?8uI+OoDn68Nq4zG+ue*^4LKD?L4
z<0nVs)6YKlx9RWRxi;Q?aBJMVb<Gc|_-5GCXZ&TSOXHCHCOWV2%h1bT#_=!e=g%4P
zZLItEZjC#)uKU4OT@I){@I$I={E5RuwUT)_=KsY7Pd4U?hfn+oLjGjo!NJvW|G|Ut
z)|(IfOZeaW^Y4$}`ltVV{Ez?ScgBx@^ken-+z-*N^G&d8*FArkF&*t1`*Z$oArHPZ
z@7JzfQ9c#J{T$8Pl`9A1=8dc4x|S=xPE;(nExm1$gJ0XfB%|k;Z+yw_?N+Ui+k~(#
zTiW?tX790<Z5_mxIO~h8+K3Bpm4iiAvvT(MjkO=J>x+K6pKZBo16v-L$GF8={eo=E
z)3LAL9mste=gjyMkmEfyzif{4;>+&i%gRmGeVP5>>)-u%dGJ+~V$tN3QmeN6t`P?p
zH)3Pv(Cfn4%{W|pcZ;=%ZhR!#TFw9)3+}A++TCeA`RacTE23{W!CrhVVD||iZwlFw
zsZfe_q*x(XCYfGdG9brd=y7IkORd#f%!vr4vI;!7q7B>0%#~gjspuXbeq0Jj&a`VT
zVdr3$w{D0fWw@*zZ=0J1(TS`1P}@~IRx7u>nW(G&wiLPRbDsSD=L?AwTRxv?qj|-{
zGWy6x6+SH*Z!Q3N@RbVvh)Mz%>8=AB3H0L4$`|ht<>)c_`U!&27mc*M+3<IC^MA+W
z$8M67Rs^U#T3q(xh|qPQ$E`eT@QM_MH*Ff%<91rpz(Z?)P~#SsEp1ZjSh%znX_ri~
zX-Unz;Z7XykCJ8UJ%bl38d5U0Y?B!LVJ;-wd9e*K^`YV1W=OO#OK<1+N~iDgH=^rB
z?ii@RKF@Vg-B?L3Bu3E$@Anedew9Zu9P;*Y<Z-C2mu+FG?Ogcm#Ab_rv{sJ!;UI{+
z{95{<6qLov3lXf{d%vVle^p+x_ft^*UK<@JzPFK$RUdYmWtCCd?wz`qG?9ALwbHi3
z!g}%8*|vhsr7evjY+rq>bBbtPIhVBELvq`1dq!5fSnNo7m7E9`%NEKXBdg8vLT?3Z
zjw4TTE3RCYd%Z}!ywrSHF0YmN7u_SYd0J<DnU@*l!_%?0Xtnl*WwobW=F6VUXk@|r
z76Pkf;#b`!@k?DIO^eLBm#meO4PEdYR*ilD#NuS)<dxj+1UZ@4F%}#rP}UU7!a1l)
zHNky&|G*F(WNphL+nSdg9;9BgxBRQg?`yFaGn9q2ZaefEdliVK=VKp=QA=Z$Bh9|R
z8OiT{Y}lme$Q@%et&rN6mIMXVieG?kTiyfce@1e2MzHH&VDrkO?ehfaJf*Ct@?wHD
zJAri7)qe2$!Pjeib?E)Ek0bsZ;?wcu(Wm3HkA5<K_=7(mfBeUP=;f1--yhGOd^Vmv
zeLM~i@;@kY4W<3p$Jlf%z5@C@coMdLpHKEIXJNTUDeTL!UCsiG^Df)|XYA{Ds~cMz
z?d*RE%>68R3;S#{PV?{cwas=~Y@20^IcAC7gJb?Pc>A!=-u!)WTi@(Q+S+HERy)4N
zY-@SpOnT><vG{7)+P3<Ry?k0lw^)9+>nA*4rDCpgK<B_E=Atk0lRVqN*6u4q^~D3(
z#U5f>i?O%d?RMGPxBAv5Nv$nzSNl{;LMFXA$8BxTG$A*y^HmJS`kaR`^u*iO^0HLi
zl4q42u=ul(aOD)wX;)5cvy7jmW9+qW;n+Xv^usu;E5=zS&LR`qdATwzel5Ok!+zp2
z@!bQa>SNz>ia&d>u-Mi*q}I~C9M7ZGs(T#kBIk~Nd?9@8uR18bIjetnNlhznUEFV`
z>6mX>cJ^s4=1=-5OFm6(udY|@i#h*7^<DcaFF0nHK|HmYRLd^-%$z-pZ+yM*w~?^<
z+hAI-^I**12(!*lz%wfEFRZlijj1a<06VM)V0@G7=xBWQ=(F+o@uR_i-nxGE^0;=G
zZ)@=<2%pJM{oE6LrX0@syk=9^2+y89lYXS!SN#9J?%cku<;J*vU42}l=HZ89{xgTy
zu8eCMK)wZ5e+p4?JeXrHdC<nUzqt0ed++Y}`q#fczV)qdIrhgt{QmfZ-}~M1-~D$S
zoo7@N4Y!5u=!=4g(jgWQktQG|lte@XML<P*2~wp72vS2(QIIYoy+mv%y(RPnq(r(>
zLMK4zJ%o^iw99vY&YxK`Yt5XQwfD2n=1kL?!Av4)C94xgnW%fI7ctWS^mV`%HKe76
zXWvYWvC|GD4yd^`{xt1^6W8<C8Y9RDbrD7{AA*OQ9>-&aTCO^P{UkGVd6teI9hv+z
zWi61i<ZeJ1dh_QJc4efj=3Q;K7k2=m>#2Mr+p&}{S2But;4eDno=Q0p?q>m6?)c>&
zVE-*Te%9PcF;20uTu~NmO?6#r9sQ1FP!aYqK|!=B&9wGu0g!jyWqC^L@luJ0J@*o>
ze&6$eN66kwR$bSYjBQufZV!BKu~!xityxeXicRsw-kj%Ho(XNNz&#DyW}ol{{SuAw
zx4B*jHI$MF-0wfq0MOF8#Um65xvyD(lWK%H{jBeL|31F!YLyLPHmQrfkeJOIaIcs*
z8K~NscaP=%6c5T(4YK`yt+VM*K7a5xG<H=|L%aCzVha}-R;*t<T)Oxa<F)|}?=leq
z-+Q*TtQb<akRPb*X4m~{fJAR7!}9%57r>0@C&@9#h18z<mMylvb$T#US(L)%V_*|-
zJSoOn8}1DFPdj3!&@e4~2fUbQYQjk_m)sLQU0^VPtp6eAvig^R@i5%yh?{XR5BR>~
zoIr>Gk@1_w780E&OdrCn8A6qr?By|4OZ|(yjS%G>E!C#Fv6XZG5mLgGl6o(ZpV=i>
z?*hduW1~Jq*p(Oc)i<K+hLv$)#F0Z5{V1=A)e`Xa+ll?l=PtSP*q^*_mOfMY-z7;&
zFFv0S2`iwUw=1%Fes<L<@*h_I+zR7ip2rKkR=Qbh_Do`+k7statkOcBmDr&mgL_Op
zC8;Z)W-sl0Oxek_=)<QBosrMH^GbrFs$&Y0rwbj%6`Iebcs!DMHf$sNa0umQx1LZh
z6A4dwUp~c?i@Ryq8`9yQ7v};w50?WD*8J&JR~+>zfM+&dcVT6A|7Wxh_P%oY?_b5H
zVe>{Ai9#yip^_@={HQ^Y+^T%}N~NP?BK}%>Nd~81!|lqpqPYs{XX%x)+}qKiHqu2e
zCk%ztZ%cWATQK|M0i!2xrSlq}pg#8<xlvCgE)Ad>N;f}>Sne&w@dd~Tcul2`(j7@Z
zj*mP@uCDz0pwjnv+}nWb!Cf}#na7+<_{tZuUSOLhXlyASF<P(87ZD?~7u!R+r?U)x
zUH>!<do^_G!CGtXOAx+=M?Yx>6ch3*^O(f;#G3RZ8d%(P7oB|^;(iH9$<}&OW>9zm
zeFZAIrN^9}s!SPRB+KPi54u<V?hiv<7S0!44BBwIw1Px$fFPfK;afd@TE5{Sj7iR1
z8L_9s=D>?7)fN;v#id~r__024vAZs8bLsH4pQF){99d#sFc#=b2)(rhhKJi>Lk+ac
zJje<r4M>!@>~gcBK>0`5vS!4a0dq}kZglm5PIXzvN_Y;eM@MGA#>RxK+CT5zT)4aF
zrI@DBtD}@xhzrcOIoc#2tYylmK-C7NAOyW(n_d1T;ML*!fh`Mg=%}8Xc3+ye%Jye5
zmM~`~G<EDCK(EqUtLeA*<kGqgW8!%<&`eH47OYX`wYuT^b11t-|CAVN{oTTw)B_dR
zE=d@<zI|iy8dTV=%>8qjjGH&Z=H)znDel_2FD$;0R$oA3D(n<pEJCx8qjTKw<SzN>
zQ%6|Y-n2;nT6nw<IDg34N5!N&W;(JvGR}U@ZL`h+56nvmwf48zJXf`76WCF<CLi2g
z=Q3KiJmIBwwLwS5Vh1$ke4)H-ec)taKc*Sb2&?LVBeZ!sOc>e4xiseY3)_$0P20H{
zcw5=SDC!B?ClUsQ9)QeKa~NfEF~s*Ap8fedJ>UHocm=u8<em*?eVeRhIb-1LRT$@%
zIhd+A48aC?z}HK|iRB0~n>G>EM(Pv|#(iBL?L1Vxx@fa)IHELCM;J^$JXom45mmZO
zk-IJH?{@8XYk$(vqdCwi2}n4jHDr5Jo!M80GX*S~@)k`nk!9;dO>CNEPwLlYqS>M|
zeSuU>A{8d%xBwB?S-++fbCbc%A<NP8W!iuNi3m*uhnM5XT!ONONYmiM)9o&d)0ReO
zBqG<#IcX$y_8NquSJhUm-ia&PWF>>dk>F6xq<ryE3=7<I`@V!P>BGiMo+z<}q665D
zq`hDo9@3DI1Ot(wYi!Q=GBWY&09Ya;2{d}vEB)eDa(uQ5;tu*NI%)#5+!1~XX9jhi
zNYC-=@)3Xu*rNEPXV$pn3xA-8?&sHE=CzC>V$2B#Ef{FnYvNRB3C{oH_Hd?Wt%iK`
z=Zs%O?aXCTPiOIPs07@a-qoN;CeV}FETar|Dwea3z<L~0T_`Xz@nE^RBuX3Yjwkl{
zkF}lV`j>Jreni?`gUtja^vkP2EIP8)`B00+Zw3pEr54^<{Fnwc@-Vuc%NCF5ppLH<
zT!L0goje*bgtOMbF{IJIE)*>3UN<TOg!}rt$Dpn(l72W|Z*jy|mN?=pWN}Jho2uh4
z6<P&Nen~SgmyBQh1?#))D#2c8kNU)AYxg$b3&lkv0XVN$r1;|EAFFWYSW-nbOIDEg
zk+x}4fyu{(=b~7a?sSvx2cjYLglRHOt9IgW`egC8omkfwjO#kCC=oz<B~Limh4+~D
zHDl?obG(^)4}|k@pnCzA2J-k~?5>Bk^a^zA8FfuW4@5{5iGH$qmbB20zZkM8FMC|2
zp7!u_sr=2c%yps2MMFQm?YSqhfUBof6NbNyhA9`g!sVYU^T{w~m1pe6G@kvy%zXf_
z-c4Fr2Q~M{l-RUCfsK>r)?EJM^9g;N_9|fWUU8}L&&M?$=$(7{yAw9f-SD?u+G;sB
z)gHuqr8=`Jj^@uYAFTaY){MO~GS+aRKDXp%(e7!?NqY(f6qT0fW$Z7jUc86TIyRtM
zGj0X6Y_#~eI<R<8w;~+pCcLqS85TiByaL|sp*36N*LVAkc8^D7u!?Cy_9|j~Yvsw}
zNRb6{A_C{y+}>DF_6w@ogNW^_?8}&n_zN8u>Hn7(HUFEa7AZETRYR_ikth~Wk5zr4
z`3dhmOq0L1782^8g*bszla;^TqZLYeT{W)QcScK(d!*bl4+SjIUFOR^B3bxt-Wc@A
z2%ViQepl~$2`EdR6;EU#>)WmGxPfkOFMT${#;yE$*?%&iQB8hm>_-2tNkHrUQsrO8
zGG8h@AN1S|oPGYES;*|IMEC3HCf)l-Zf5cMcAl!)>aV+R#g-_vQOyc}jD8~p+$^qa
zto`C5ARF>aRm^Le%j-pN^CgR2d6+@G&;5=Gag!>wCjouZpS8X0ahgBx++=TFud%P^
z?3Xtqb|@l{lVR$OD++!Z3gQj`%Q8CvRia4;!JdqC)rPN6`)|*P68marL`Qw6mK5g}
zCN63+P{cb%%hMfHG0EljfM&*}xi74|c7n~a@C!n5k`I&#3_$#2VLU;8gU8!D>krN^
z3#}hRHE-0J5S0|r5h~d|;hZe`B1EV|r}+Nf=F5)WxoV}+ij*OR5BU<-bq<DkLB>G0
zal|j5oCw?8B{F=cHSo6*#@xKFa=S-s@<Wh@fURjf;bNKiMB&o6shJ<Cy#851^P1&?
zqW?+thWZ;J<?bI&lEOaYg$1ikFJ?peX6H$F9j2zlSBDz1AIf56+_SNc45}{2cH(P?
zx9-a_Ec;{OM|6sx^0{&BdRz$Hb?zJSg-5qy_X0Espr>ma_A)PKv`@Wi>glBv2|$i}
zk=XDHNeq)h0KwGFdqAskOuPMCd(pH7u8f3(`RqNAGDQv@p<e)8LjaCvtj!4$o-_ph
zBqKq(@f%vs^Ft0FS%a-3N<3vW)4LE0EuNuWYzvKy2==uT$L>)Yi5XY~OuH`F*$(UP
z4nLrE(8yG<Fo82mCb6YWS+CZH3{07Qa_sF<*NFSO>ffFX7Nk$Amn8%f5VvN+{DQUX
zB7>(IVMZW~@cwY5!_Lx(>@#h|9+7H3<Vx}Pg0uIs0zjew#C{-ifjBV2#hXhbIIDLy
z=U8@5dvOtGN{fk|ed|P0wL7sk@^BDy_Oa}89FnsNNM5((#j(^S5}<6RVgmM{C(^lz
zih{nd8fD(v&Z^T6`MSN#TIMac6IO>hkZYT1-&q7k)OI3)K?MAytrO8@e01cYLvui)
z?S19#Bo`tZ@!8&^Ia=4VU-lIX-bTl=UU+d|jS;PBnkYd%67`=e4XE%_Qv(PT`Nya3
zz8VT#NSqw&d=<e`Ljwom5uZa8E>dckf~#7uAd?V7rar7mOu=EjW5;!IyJT|X@bM1j
z%z%frjz}}A_D?j<JziIT&y(?8-}lcAJg-vjXYtfwg~rOVriJv}jK_$OTux|Ag^ZH8
zutM)VcfT3kBHyLIR{5>eW9>+lG*Gsc-zhrdhvyr~@z*7(b?^m>X-AMqx1H~bg2k-k
z$d%>*vs4z}S%Vj-nH?YGZ7FxT^()hp_e(t*-d}%z&<McBx`F<<mf20|#p67@#9ZPN
z`MGK(Lvi!!x0!{<W0&9ljtNW=KsyXCaKkAEAg3$>92Td!d&m1eXa6+d0y9wol~f-v
z%Z({C%?I*ECQ}Dn7fH{cz5|D&{$LebYWisTdZ#nKB^{bUzY5_y`&g0uxxlkS^h2g(
zfP^=V9wbzSv&CS9zC=3Uy{n}(cTTmd(wh>`JT=ld>Edh+aiPul(*9Ncuut(>`E<#-
zg<2z-p5Iq(glY1&>ef|sQEcs+^zfzfd(~Ypv=8+RpH`)cO_)gc)SY(?))qMNV8%^Z
zvL)Z%dnrl?R<mMmt4<93S@$R5NSuXsaWF#tx)fmLRJF_6nlsL@YJ&jVNFqGBVKXXb
z<$Ou>OhxMe{b-A^1r!kQRCY@0C@~pauN3Nv^(d)vR4~n0)Id~dO5ZB?t~{Mob;%{)
z@JlBX{lxC<i%sd;>ea<cA^UKqlGoy!Hd5vVw~sMPZ`XeH^EMNP|9kT_)>usT^(`##
z0}G<FHP~fUEW)>}mxo;WLVS4t=9qrZgOeTQLZO)k`+^g#gE6IP=AoLI2qQze2Nz|D
zx4%`r>~o5VZX<pBdNu(VDLMA3TbzAyvOe#UzSqI!7=$MJ_#?&o8mG!=ce9CaFrzm0
zFgxIO>47zm*6$mqK5{kNl5I+ib(epn^ui?poeta=(x*~+$8k!JolU2>f5%Tt8O|h9
z-aB9P$m4xpR6mt=`EGT%y@~vhES8K7?n67us?)dp7VbL<<8I=;s;KMch1x)<rD~ko
z);wL}{e>``d(oy|0Da|Kk~~<p*XG|h$cd09f%x<C6n(ETH9_aQR+WCa=|84sUVcEC
z3^+44Aq;HU87%zU&}c{^EJ~-&|MR*PJEMJsw9^-itBss1PYmas5vCI0;Z2l%E#l5j
z_P)P;GM=J3TE@OM_7anuoJV{CsckMs)HiM%IB)`JUx*o0yQ2g8WMoEU0yU0!#GMlH
zud;V#Y$zaCy%8n{Rd8SS>3=XL7~jWWeZtmmIt^8AY>I-$NWofk>RtA`f2@sIYtP48
z+pu$>@G`TS&VeioMmD#L7~18Ceh%3YMu?CO3+ijq89OK~?<neNilnMX*}d6d?DngD
z)8B3o#Hyrw#u)cO=7%>Ha$TY{UA5c<C(_pww|b&eo#8`QPy1x(1+$E>yI~{=<$v#W
zOMNAuLHM~OgA;^1eT}Qd0`t97?ZFKD*t=V>jW5YDJ~J4Jc>8$rMJ~j*lhQgjDA`^h
z5&1FdNTuMOUj77b|G6zSAs@SjP;9tf+MJutTf#fHMz*|vLyMKtC+rhn^3JE!ndSX0
z+F8S`J7{yJ#b!UBCV#QLv{I}!cE#Vl|CZs(0>6+mvZo=huw*^E*su4K9QFi64n>#z
z!~WqFI`1>&?uI!NMDNykTFiYFd?MW6&&LiQKv`4ea1=51l>Vq3V#Z7}aLYbiX^YSZ
zoC#ZQCjDC;z)=x25{(Rq1R?h3+=*?bAZlB=yMZ=s-c<Nz;eX4MPY-pULm$;nly_@J
z3{0g2*E9jQ2=#MdFpIe|UKijOsvUk1$eaT(*&!SpScEzgatpJMf+mOtZztjg91?gB
zyKS2|BzGdw6p*NQZk9bQN=%#_#r4T%bW*%xh>78~ofO#&Try&(kXVZ#tK-Zz{{tWp
zx7j%yHZs8!eCW=SCJ7ST&5R_@r#Vq7AOVUXiV~Olaxj~V<YI>yE@rGgJ9(T)GDR?g
zaU2GK!|4p>><!>H8-sCLP%Kbls!^-VUS}`!Ktg6`ioy0Q<6GuZVPXE-Rdn+4?^reu
zrb9nIY4Wj6%d-Lm`_=d*gm6<Ce|csHg63C5Z8sX(=7fo(=<%lHp+h~N7q;OM>Xu8N
z<yPm4nIJD{L$QgytS$O(iGRMJVCiCfd&V>sv%w)n?rlRfX<&9ZXNK@RD(mwI({>-D
z8!cZ(8wqKU3eI^Sx6UsM{jc(cG1u6n!Q@QlieR8gTheKauoTaA4p7@$pJbyb(PE7}
zu;DfBQZ>1rgLi(d=R0W~2TnU{2T&bqLG>x%v>uvpG}I#MQscUYhaY%WcD?ESFaCPV
zGT-Fz8~te6V@g7$&Zd<a>ycSm1OMIJT55Zp$o&Q4rg75w4#n~XiQK<zmE-(nMC)v1
zsvu&<N)^HLaQQ4!Y3FZ(dK(*Qn?SnLVEU=I+8bl_)SH8sh$O3X)d!lM@WmYKBh+n*
z9wp6fLIeEya%^^r^e?tC?p>N;s1*_5b{QUFQsu;l(3%Zx0hMIQ-!8UY3-Lv&is~um
z&6~~G@fV7^euLO%{VsOOJ7Fhd0y>RH$bEnprU&V&|F)~9i+BQiv*Bue$+CDaMG6}l
zEgkXXdeSiZSN-n!)><zS9mu)us1qN)CIiB)a0qXZ3+z6r@LAB4c=DAPbN{ZKXYCK#
z(mQ>&1s{kYRay$nBS8)y@jx-%V}aRm)2PZv^*vnaLU$k9O|n-Nq%805E26L6t23_h
z?kGv<YV8`D)Ow%nGfTE#R2$g3#orJApoCszS}A{Ifj(4>U#JiKhVC#Vv>US~&~>S}
z`lklFhUsr!pqTNwfu|(GDNckQ_>0vefrz`mtOA#gKdE9A@w`KU5^Im8=&luK{jz%c
zM(SO<_D}z4p~<a@_boWyni@ID%KK*vDn;fVo2^biA2j!k^<?EMkOxz}Uh^mcoiZvn
zyh)TM0a5Tc^mlSpQ^i^c6?6sqXg()Pv0@^#{*=i(yLgpG{CT@~e`ctg!$oWFJ4uA4
zqEgL$%`cM(s7LtyF_M1C)RRRR^6txztY69cSf{UU2dGD>YTLCwQRQi`rZ2?WC_w8z
z7N@RkDvF~%V4Mo<GCbYpttu27Z1Mj#Z-3dN->&Q?cAf635+59K;W#obe;fGmOr%W>
z1*BTW2rc}vXt4ubv>Utn&@o{pqcf6?<a{0F5Z6@^(XIJf@{0M1f(E9Yg#eBVvnv>@
z!rq}q1~+lqwv$c4bjKzRn>SGRV3?Ehn?uyqs^{QJZ8%kZzg}?Atlhuc)bckVNM23u
zyK0(M81Rfco%(#ToVXh4fN}I@z=geGz45FZS3qxJqNg{k&xmh|7E$Z-&e^WOb;74N
zo!gC#>A;*1@=5jfX5yzGi4{<md|4_?b`n+~(=9_>C4Jg1EcmU&kCgSaeSV4hmu9h9
z8|*CLZ-+cJPr6YV<krd9F7yg5-8SRH_o`Avar{P4*3wQ?9;xF=ll^$Syl77PIR0A)
zBwFA77h>b)WLRnS?p-S%Omu2#txOT#W4$`~-6t@2AzQq;!pnBix-jj0BZBGJwr?M3
zr<mxxA0Urj<yvMsOMt~ikGE=7%64Pq9!uvj?sE@qcqmlA=#ZW10x62iyKet(K9ALU
zpue3ttzBoKgBTLjL5Du);0QSPik|V>7GS^8P7O_HmU9<nstLwBPcH5FSO4jSb_j%F
zT~_0v8b*&8y*e{ha@e=oe<+1aa@C!ASA!kk-T0;cNS~8AHf-?N73}RIuP;Uo*Bich
zA$x2bv?FilgTudNE)(z219<%t8f(kwx_=YbEpY|j+G1v3{H<f*+eST!UX~H8cFx2+
z$NirQM{r}cT9CLpG!?a=B0@cjl+XbT>~45Wo#(KqU?D7{5vz`2NJnad4i<?1tOE*Z
zhkf{EXbfh~u6AfV-5kJCikX~>b>hUOzOMMM{+m04b&=aibR_(MUPqnn^Iz{5CGv7;
zrkzk$H9;IpZsa)qPg?4yO%ph6Bsx)=_%F3^jqyVJNzbzQz%NhDrU2Ke=3r=3fj<8&
z#4bBI(Eg|9Fcnp7se|d<f&-9@bkkZ_?2#J7c+<p13!mgcu_Q&LL9m9bTf}HBKI~~P
zo~NR3DI&Mi19XcqFt66%`>TFO4qwlf%dI7Jb{F;q{1cO&i|>X>{2gs#-rj-ifH7=-
z45{PjPmy#{9>i$ooru)(ek6BO>$MZ}mxDxXG<q~x1+l?9ZYO|eKFRBPrvB=vZ^FaQ
z$Czw|-rd~yJcrJ4HGV8su#x*YVm$_|yC|~pYECH6Q$FklI+a`4(DIS5a`)Ktcz|!>
zFMQ|qL|E-)rv{8z;RMXGHJ5E=$*8MYq}yS~%<sUjbCc+;=8-x)MmmP`ljgwH-RKm;
zQG*^%xaRJ0Nu1HF>G^|#IG9#Fj8>2i>nen^FAuaSEW%9w$;%R2TQ600m4?mA@}fwa
zav@FT^=Wxk`DAxWQod7~eLUxpI;qGwEr9PzA4L7s!dD57jDDOY&j5F&-2B4t<Kogj
zovl;pIR9qkD{q64HR<P-6!^pbrq5Rzt)GOAe!v_yx7N%zR-WlfNYGtT7^8~01s9Rw
zpMTU<9rKquxL)^0+iWE*)hBA@#d60M-#C(%(1n98@^D{!ZjODn{NW$r!j($y4hg>O
zucca-eW}yvyLQK1L{7Iy6Twn(Q><I}are}W5IHvu2!9P?=Ez@OrFHze8z}eBAyxDN
z>-=2hIJfD#Ol{VD=nXXG*!$uik6gRPR9>0+98n<{{*KG*V5K~R-y3}fvM{=taFStv
z?QNHN)8dz2%#PE)39<1NQy(SgFYlae=z5gUW3A>NFFfd1twvn!-aIj(H6Gmd$0lF0
z)>YVc(RDgwum1@;%pAM$`I4BlBQ(j|0_7yfqogS~slhUTD;Xdl%wKs>Ab_)c)W&Fr
zuj)OeS1CR(7qIO@JlOYN<tcj8F9nR=OAENEu6%+%B>e+u@CL?yo2$JaEcWUBqHIUK
z<ir)Zhd>cmBf^=!hX*##o~Bnb+*a$zOD0LDaZAUe&lXEKXv$gdz`YOF{PFlJsq4XI
z8aGWl8WQVZ`Q}$|Bl{nn{)lzzsU0i1JoMrF#pffkIdAOjeFi$5%{S2XX%g8Mu~6T`
zA}%T1@Lk}Ycv+bb8{OLPhhJFYijSa=q9F_vcG?>AU*NlPl87P~;P!aA<&JIb#6bC!
z(00V;a%VV!hOZ0el#(dr9A?PQj?X(kZo_Tw+(`Iyi_xKliLT>rqAaHOMinp}B8~y=
z)G6lkvxAnj1Yd9xEPqX=2)-vEs-5ENEpCklS>%=9PBPw5<41UEnVw6Sy=T!e5c;Rb
zJ$dLalZ>CmuWxWqzV}Iz;z$l`1~tB7caCn<CO9^x3t5{ru>??;<--<_7W)L3z|K|Y
z-D|d7)=^Q;1>$C!RyD$wsXA}odnI4uwZzkNoF+j}rGfObfbdbwPQ&qn+6V~mZTC*R
z+WJC1>{G>(jq%!}1Cz3Z?>8hw*YntQNyXu(OI{2TvL_?Hzy0i-eR*Tyb_4Qd{t>=M
zf;tpMep8PGi~4$qSz{Cs(ryh|EW0AJVla8fIY(II3`vJqW|Ag%6g8`EP(?LZ?6I%8
zwc4Xo@?1C0s$xG>S$SXP-@gX=a|$lfdx(!p6}Q9lj+-|bZZ0;{0s}ERk3W9T=q7TR
z&j;Hu`^Iy;UY<I!HMw&Vkb2rI`-hnxAo*2{$wE%@zj1KS@EjC=hP=Gj*=@|B7aLbT
z*>%N4d}HUw;yA~&>fk<8b^hVnAv;SnI<yJMUQ)!UBeuMFaWMdv3v&erNQx4LCKM9a
zXtSv}A`TF-#VOQ}(3Wgi<I>N-xhG2;a#-v@?XW|OoVv9sN-T|m(jtTG=jyW#mbjgm
z*6l{SoY@$#7>60m>7cQx*lh?XjG9;589|9$PGmF)gqu;OE^fb=YCzoDnqG>0YcCR$
zf%K}^p5b)G5V6d1b`FU_pnvsW?93y@MDDhcxCktjksYg-9*-s<bo^MeR&Zw{6uTPZ
zUJISsX=}()FH4T$-AaZ=kk^ULRE%0aNJ}L0r*<UGfFN#~4<gt8{+SNk2u<h}3)t48
z0!7J+=X-sY4JQp8nCN*TY9}hm+SNy80keH=Bl(O)7COV8=S74U*O(jGzrP6iDHvcv
z?#o;b251XrUlgxj?hc(nh?4gYzWD#MaX1Nag-FkFT9KN}NOm)1+4^LTY`98sSzlO?
z1{Ss2J}-_D*wU-8ndh%C6m~(j1M4LBT1Ix3aJ996o)kLc-sm5qCF0KIyJMb6Pe*Kp
zqThW;on(x4L3e-VA3ityu#9ym^G7ezdkyZH8zVP16-+mNz{(e&{GoJB?_(?Z-<7+(
z+&gZE3bd?#^w&(mss4AFKaI30yT@<(@X_}DUe{dz;h+EW?xO#rjo6ymv6(C6qmRB8
zbL;Y~^l-)WceQ=R%=+!!es!&j8GQ;y1PHz^%D)+`N*{{0tGgoW&bi-J{ned2LtP0A
zd`QuM?O^vFa~aQGaP|j{H=Ua${f)Qr(OY(Vy(W+qmh=s)p&8Hz3T`y546W<TWyrR8
zKlkDC|I;p3Y24(KXAWPQ`dsMQy>M{8Jvls!v>4*wjJ_PR<RDcefh<w3h-`n&b@Wcz
zRLKI&qCP`JB*|XH`PZBJE3X8dpXG21Qw}kfA9sQYu~hi!_1h)lG@-bWq4oIdU1N8*
z{}n^_l~Tt5Rgjhf>~b#`%a&LA%qP=`TZ*r)N33vV>xHG;2ANSbk6ke?J~5PO0J!mw
z6dIJ~81d=@>u|`r7(sV|*XzX~`Bfpv0D#q|)jBzKQwHyMmwWE?r|w}#6IE|MSNv>s
zdt216kdeXUAzTK+@Z*0f&|_|o5i3C`$cb>n>oUBVKz!!8{k(^>{)=KuQzNejxYJKO
z$tp(Q`VZaQh@N`BI&^~8{`|$4@YZO(^7+~QOcKd#Z`ca>+VI)>m8_T_8mpjf0rhM1
zQT|IE5wl;DOLBYY=@ai291X6yTHLc$<`S&;G2#{&57qNNwJgJBu_blat=c8HFT~so
zExOk)Z884wW<B}0yOQBD|K&xNl#ZF+L1H=Mm8|0_RrNz?Fi9&~CaJf_e}t!4L)da3
zY}bgA^Pcpfd=7$TTB3oc0tsa=D<23%E(t4VngX8aU_S13+qK4iS2@P-O#KgVjC<M9
zVZo%wOMh4LE)Vx`Q@?H4BO55S>aV4nsa@?I1aOHpMo#_x*PHmhB^@s0QV3yO8){Hz
zTt1^bz;M)P7AUH8J;O~Lb>RGJ8>Ml-r`M!L)C<mxb!yZFxx-naV6X&+-hVimCW)ei
zgZ;YTZOoKqjTqkl?njV=`v$?7g=h>qFzd)L$9aU3=oua+bGZ@xWTTSivYh_-_+ftm
z>KA8Ew#SpnT=&h${Y?^;;(|?nlTljx+*ob&S(RvcR8rQuL!C`Uo}C#p_B7Yp$HJfD
zec_U78e7Q&g<8y-M&KIRzbn!D$8jvdi7X3(wABp^>eXOWX7-0F{~quNC?K_l*SLbH
zTFX}CD&tYiQWH#``AVM4PrgFOA$WJAM`wJ>v{h$Z?Wh9<`l5gLG1N^;!LV}*W<&hg
zcyS{kaZ?MnMwb8Fxj}a6(N_J46gd2QxcVR2%6{#l=q+~W*-hzD1xwcj7*;ojt$E>S
z^ETIO#7h0idbAye)uyNtj0xr33Tl7Ahenm@pCnN)sKc86dcPNX4QX*Wd$&gnvQhY_
z>Lpg^@|APN-I&0^7EAv(>X!aNXv<h5$RCHOCHl{R$+0}quEVaG^L^R_f0_*TF<SMk
zMDj=K^ZpJUF~Cvj6B~Zb-SynUfYRFY>No}e#A}@Ivc%an6lsZdXf(qaqQs0y((<LC
z+Zbx|cA~pM1}?%LFx^1)(?<Bua(XCxt7ZT#ZCXRX;#$ZaEoEPcjm5*jM-DSuU3H=4
z&bkOjv<4SD5`t0j-y24W`611ghXL%_7-BGjQHx@xlD-01A=+@pCJ6?nZjv^koZqzX
zP|g%>LW>u-DH=n>quTr_kA?FAiG>r9h@G!gTqKD@=Ol4<0i0T5lxco~!N9NeUjY&!
zNukNqeSRb~UM5hxDcBvpIUU^DC(6W@iyngLjHQt6bqK)n_7>nKfdPQe9%O+A%=(;N
zmi)-G#UTR8s4SSFydow&H0|8IB+%sdno!=zj<Tewp0e^qANX;`#>6hZQXKaH*y@21
z?qTM?qF<S{)_@{aOhgIA^R8>5*6_55yc1yB8Wc`3+2D6SY8LtY_v3%LFDjQ^y87Q^
zw1gPy7o0-+How&@oEs%yHRJpUL*(eFo7AS9(?+aZeO5P@`l#w>sVcpOw_YH2Rq^gv
zuHhqfZ|7rm7mVK@odn#WI3c%l)K(T?V(pT5Y;F~2;l9XUP*%U-)vH{pUbGpw{==mP
zZ)W?%qxq-`GU=u*HSsy!=bVDc!j9B`#Pcyx{m#RKyzR#{<NKS?FY60R)cM%D*HE>l
zmIlOC6@=DtDZx-KcEj$Xc1uwMbJAiVWnQ$`kK#6mc=lfQV)^||?Q@u!Uj4MqZsNN9
z!|yyDjDP>-Rb+PDWZ77pTkkV}E#^5mRFnQss_I1t=U5d^@oLREeH*A)k$Ky>m1*xI
z4}NApZZ3a#d2~wADv;dnsC5L=QaS|;Hg<Ojbu7sR27jPicE20X8ZDWQFK+j?c`KiR
z3QtoW?S+f&*vdam4xx=(kcYKmzgNTrNqNoS2ZMak1#OwU!Vwa!1_=zc4!>(s^$*Jv
zPVgByu%CFrD!<l&Ib&Df$$fuUwRU%ObKa!~lXrGO*;ru-Q7VlBR?YWzwB6}F{r686
z@pZ1>dhRK)nQ=|Ee(~=4<So0~ao^wGJJm1WVP3;0`MK)P;)L!fKvTKo8Op+$2LzQ9
zedjdZo}a4ZX4CZFR~o1|<wn`DR=d4HM?2i9lGyDAtEt0nS=^a%%ZNta^Vxrwzo|VA
z4zarF^K?w~wrH$#JIHC``FwOD@)UyCp{h82Afiv{z4*ft*Y}<u_#JJNq?-jGlOB2}
zX9N&O1PmA%4O;a%OYQZK{U*)Xp4H~I1BU5ByhjS;J07fb+n%a$Oa4Jwd#Jf7hS9z9
zeITlE<ZMB5#|tGhl$S!r=7O%(W9Z{I<>3uGCmTa`#}b|hEW(EPH)*P;LcljX0Ny(I
z^4Z%q^-JB*jR9x!qtpv-OJTvFhrV+2ngIO-s}}?1nEyG(k#qlW-wBCpA&y<)IEpAQ
zt`Wk(hY;r`iG9IXB@8<enyeMPz21-wv0f64S`)h?fJA=0Vm;dC3{1}BPID*F9aE{r
z?b^{1#Bs>>_(~n{z!mm7Ber=Zs$AY=o!lc94**q&JGk{SoNn{_OfS3VKsQ$y`<p|a
zNN2a1t6g|y`QCSyn-73z*JHg+VHc`*lVhCzvOlXq>(?_!#T;yBP^Fp9=$8|C((v!%
zkiK!zYTQ!N$bJI-;9UkCZM7u?;he5M_snw~KbBrRd@?psEMt$ov0k^Z71jz8ovr?e
zv)7#2)DxSpxoJE*;fj9;2X&ymAfY7SNIR)HE<5#0$xoQ3;W=g3_i5^&3dW9`<$#?V
zGciAA3OSaTIH3kI-AA9@S=K$1t6>thgWA)Q+7WO5I`x<LJf6K2cKzReJ<lzMu0+MA
zNSc322tA|M#|Y903>Om`-JB|=rfm(oN)L2dMdZVNZJGGTyRU*H6zQ$G6YqR?<^O6l
zrAH>o_$8Mnhu#tN(mvXJ!f%<`7$H7-F6-bm6C^2#XqOM}TIGUdyL(~+Z*9#)IK#<@
zGY`AN4OaKH!>WO3ub>FlAu38w^$%SA8~J{xEg2*Xh}?~(Ww7LbNt*(aZPlSkSh_lU
z>3_H<5EQxBPB=z%7(yM2M15YCOliS!evK1I?T9sCeOO3`6!Tv&MulA*$xOx_;y{Ng
z5Pf0DpSnLyGQ!eeL<rqB(uZlvS%Lrv?4@vGrz+;+j*~V}Q)D180ZyC45s}P&uqcNT
z$;lxi+3{d8K*a9Y=!=9b#q&Exe`2P#pLM{8HlWNjh!`h_L&2(S&D4VshZ(;lQ&70_
zVU}wtCeReAGo;jUyMzi2!(<j}pKY!69YGu!BG*l#A5As*Z0atVqwnMTf?T$xMA_^y
z;T7xs4)Nvk7WKy3sRR#je|IA<Zr3vWLO)}6{Sa1b^B{(sH+dT?MwM5Ur_ORgBB|S~
zgM*;&*-Y2tJ*O}J>IfKC$l<@|t5qg=K@1Ub6L*XtZRtI~<7~R9^2qxMv|{J)^wke3
z`FrV?xzs6BJs)jL>wuq78iG$n3ad*0q_^YT-FwHn<@IuJzm+GZ-%76iowIdzY)6jl
z_>?w%)VN7@R&`|aw~Z9uzFPPB<>Lu6RrPVDuOyD6RDw)?Hr4NcywH(Rp?&YE0>b%N
zAs`{Pnu)#;nwKrK+hrqY(DcM#ccJ`Y{6cE2Cgu38MKbb@AEIO2BimAOn3q-U_pz}2
zt2>?ROPP7;fYOBztGrrpj80HHU~6&yHRh*zxcc{o{j)9p?3N|L>^r{ly+Lv%tEUA6
zNCNFK(~KxH*%(}^T-Wt*=+xbPWSn|I?!x7%g&X^sQI}KvAMjiwC*Fm4<+-+7*TeKs
zIu#zHFWdKWbC#7Vx2A7qy;wATai#Y%efJ#d@o!__3!y>Uk5So*`<IYorS61ll92+b
zGlt)82<zk)DP|4E^jSTM^9>N&50%eZw_b3$1wk+VAUk1i^He;bEc2fTE(=*cDn0<@
zGv>jCeA~}@C{#6FWs5r&5bAQ>KPm4@zPC(=wOHdynKZGBOTzo@GYcpAN)!9Mwg#`b
zlOgig^yK6C8XbnAd?%$*x1YQ;bV1D5${<`yj4Qlr2X^V_PsI?MW;V2bxD=AC#=%!r
zmODCU-%5nvTYPxANup8iesWeFS{``wO~!Wxo-zR{^!Twyw<*Nzn?WhRm6U;}ZMv5&
zCr{dyto6lFLCa~sWR8ev8ijo^?*9%MP<`$Cl4E%LkN>1WqJ+DK|9@L`aNpB6AgU&(
z-&e~SIDGN+v8CS(j+fE9<ai7BM7d0CmNhl_X&~B4FTdlZ#9(^Vh>35YNa<d;k`0*;
zKa5_v(^hZRv@!Sj{wGJvqW8ZCf!y&EnCQUKjyKvlUkgoQ%L+O##gBSzS$+MKlA(V_
zXQs=?cJZ9D;;PXtC`DL2zT(fDwU_E@Z+idCL*7Z1?a$v3Qq%!p_lr5trD->BU80DT
z`}27R56z@GAx9l`z>Z#9b9Mg9eWZQc?ZigjnLYL2n0*Q~alLkHMRT=VW3NwS<|w)6
z_GI2mo#edZpu&+V;~My&CO93|o{?a=>7GQ$ST)~8UJ74|QuJvI#VbDT$%=sZ*-4XX
z@fl@Llr;tRtvmAuyimN?Mln<VTHE^4v)U9nF%-tZBkMkQde+cOD~DS>tWX=GUquH>
zzhulLBj)TlzzwPN(sI61``jL?sXz4&7Kt_n*C)*BSSX*^Fnfg5GO2pLg?YRYW23h8
zdHQthTr18=^^eMOz?6cN7)#FYtia)tivxbg!lf16fulWZ?|lSufVY*sY7r*GVWN%2
z2K$wuk(BVuB%gJ)CyZmEAB#cDpFjf;DhlnJkc5OSyUFjXy5&)`Wj9_nt6f7R6n};F
zt{~K-;<~j@i=!V)=H?o08JA_{M09_obnfQ=Oou$)r~gK#t|q;T-NtS`FX|ibk9hXB
z-t4ZA*5l@l82dd%UDf~)uk@`vkF<S0Bds!=>FEAsEAN=-UcF|JlWM2WCfAUj%-Xm_
zF`jI5WioyAJ&+pEyz|(|{$S>F<oiVAz=bk%DL+$MNBnV)aGN>+Bt$sC4<yDwF**`p
zgf@cq6_B{iT16%jIQ@X+?ZYbVL{o710-M}Jo#-!OOo9((J}ig80(2^Jb|Iv;^5l-W
zftwE4!w4rfsDpP<LHsWll{dMwT6zG!P1<d3rL`vSF7sj;(}z*MgRfwb$b-U2%`yj|
z7<P{v3}&Zc>7$*Tc?ipvI7v7-(0HuP!Nx1?6pBP1UP<ycVmq9OWRVG?oHi&MPh%&8
zMGQoGN3r`ev|4szWJD8dwx71q21w2s<-u__APmdH3@}C~UmGX_fbWgd1hLdZDM^Hv
z{`2TwE`}LZ{?hjTb&!9buC>dRfW6Y`@NNfoT<O-UqTu`MlR=<U6@AZfuo9w|H}DWS
z3yzdRa+L$%dA$FVdY*^v=qu%pO`Z1M3p@SpW>!Rn(bDWv!XlaVJBuaz=!!sLy$e2h
zZ%QcH#BT=tu%hYB&f?u?*}Tp3G{grRldQ~il!$w|M!9uIsjt;@yGO5q-$Qe|2XE3Z
ze`w^wx56&i*|bs;tr1bWhB84#9?DS7EQ7)pg@%`~1@q6<u}2CsQLk37%3YaLJ!iO|
zWJzv4vAA&87F3`ZrjdyRhnJ;7o$}X>=ibM^diba<HuEPXG38z5QfnG9@b!@T@(11f
zvZ*4IRvMuzM+}xH?U~b%-Q68yukq>1T~CFd;`(Y$EO`~qwgS9pDAZG~zc@FqE%cXl
zWyyo+zGFuVp1Zi_7%!roME-?-T{M;6g1@44Jw%+V3iE$7K(3PiYQecN_eEZylqTn5
zUqyWN)^AXc4YjaS>HYN?@%XQw?`|FEvVjnI2leqU`cmFCe1gDj6<eU%g-=VzoKH?G
zlB;?X)2r?u73%G)msw8EmUSw<))Rf%6hpd{|B@+SY<1prUNnMVOKJrE=U$?YrYF@e
zH$*+&*!LRnH+<FPM2|uj%3oLWQ}}AfW<r0aWq>O<9x6z4Ufg*}?Rb<gKj<ad$}c+Y
zH7|Q3Z0Smvdw$hD8D39Haa}+?W#3XX+u~weU;W5KpR!Rtft$R2Ky!-PD)l(B;NbU(
zwhNc51?sjuADn)4aS$3_87-R?BGWTId7qYWn^a?zD>WY;-#=L_pBQJigjgK%pT}=_
zxzL6UBYsUk@&Bb%yOKW<N7cS~sTKdgRQ!+4&3;x!{l~OJ>=gH`qYi?51%z~ccY_4X
z-d&Gb2?fl}XOX?RKK-oQFuQc3=%{i0iEj_um!N3ptIxn)POkoYr=4uY?z&v9OZW83
zYE%-+sReU=(H>JV2oXi!mm3=tefdVuXYi(I_<z%=ak&w*7k$<}KkqMh&z#-t_>fYr
zHldN@Y%_A@6Yp3}*+MbA2YKO@T>QO^ZZ&W6F&X7BTB%VOtB?mmR6K=l^w@q_3A^`7
z%r!;C%OxpicDIh)8BT0%3+~KgK$EDL{ZbODJo1oE3=I~Oa1f2LEQ{->l+nD3taA^`
zS7pX=Qa$MWLB`a<#}%)4miiD`wf=5KH``MbkNsS^xBR6uD4LPcauy|+)kY3b{K%gh
zjG0~3+0K3WLZR2Zw_mL+q<kP{y@uWJBc1vj9gM88WoJ3dLa4F$ufN^2x<9#3?%)38
z8P0)DA1bJ0O=qMfkRvjt-qk$H0=C}k#aa8?yGhW|4rqsUNl%JJMc*j6>-YisSj}`c
zUgMPvD_z%6?zP`|@Ncg>eeNv#4_pobeCJQ$oJL17ai_(!MxuMAJR;pTj={ip0=|=h
zp>f9}<WI}&p7mVdzdw8X#_-q!Zu@@$yuVc?q5X=DaJC398zVjV=X~xNU5%&xszDOe
z?km3Nuge-U{jQf4Mss%g(-`Wq2L+#=l=Ty)y@xutBX^UcJzg_NUqV0R*t_y$9Nppl
z^*QT8KknHR9mYD!Nk*RqM5EeEUk*6inr&@S6C)zFa>#kW>-P_wQYyA$JPxjKXC{;T
zoA;EKUnYuj-c)&YMd)GZZ?7N=SSha{*`t(`sNY1gM010kQ3fvUkil<e@Zz+#o75#h
z@D+wHf%ESW>j%^Q;A~nY^>FC|G)YvwED_8|-A>$Qwqk_{Yyn;@yY?{fK@bN-?t8;`
zSH?qnkyoB|u8)={c^wACiY?1%<);6*4+}mG2T0vI9lYR!7tHsZqx*^o=%IDX-lw%v
zIq<1xQxHxEQj~Lm-3JSECa?#!owiUNG1L5JA=692dUP7V0ZiY~12Yex>==SDhY016
z{V6++ho(SYq!$y)&fqN7`BMf%_V-A2uw{H*5OFDlxj?7QC4+=Gdq@_WmI~&iV0D*{
zz02>3*QRAJV*$DC>EDS0`PtK%I~JinSzqbT2gF7-1oR${OpZwpq3eU{#V0{b{g9oP
z7CMD<l&Eq=9?GG;Td-XiV!1z)uwwqp`~~FkFY6nIRn({}bp<jO547@}nI@^;>!gj#
z1(~Rfj31YPp9rAhQv$QOXDWv*77rKw<&9YiM80(z?~TpK4eXIG{@W7`h)Gq#tX=Y@
zq=|;72y>48{w8f;5nHCg3Iu6473>*%CA9lpZe0;w9D@%He3A^dxYv97ws|!mldk@*
zg3n8_tmSq6tyR0O*Y@k0{dEqzC($AQi|)M(BWc{d(*3K$`7&N{x9JbEoIiiJkZUtI
zdn)SseF5+-%`mn6c;dEro094J7S|Kh(tM&4`A&(#sl|nJYqPtr2^VRACy2+_wZz~O
z>F8ok<i#Mwr#K%6na5$SEhNVh$!PypO330waLRgva0b_}8?^KnXFh%~|9hep_RVg_
z;oc{moV@tsrQaPp7ao@uRoQ8{=n*$}3{3{>QI2=-W@Sp+U+v@b&N=(^eO~A(?X|<|
z$QCj`+6#R@Cp0iuyh;Ov@NSL-O0O@Wzb_rXyxAM$y|Q2YRSdRJ5oHcPg=#!g=$-l5
zi6~lF*&cuC_NZK)a;?11c|pmm`AwE+tV(>BQQXVjONnAQZz0)EtEKKZNmffa%}T3m
zBeiYlX_Y=OBViFL{5o1p3u*SDS~+8rOXCn(Ncz5c*}A8fy<X&chpX%LGkTNZ^;6#m
z>yU#UO?b(nb7i>(vbE02=zdQgrD(~Jbkx2qtV@b9rKB{T-Xe#x)+&5!rMIRDW51tb
z1R9jDUi&uKZSy?H>zz*V-AOBzzgvaJY~#X9y{0)s6JPJ*m6bG4n;O;$9Cl8Y$K9K+
z33<Fmh!Dsg<u|_$U=$Ap;2b@KzaKV(>in|^@N`&Cvv|W8F-pAVmJ>O_gR|sX9k`z(
z|HB87s|7k)%BXRfc_m_OcTZ-<w^IH=!c`N59B35vbRP)!|M_6l-gNnlSK@^ov*3<0
zy5U`I+_C()bMmqsZpi!c&6ySyejVMVIOk`%I%*!SX0YKk%KtLH%*A7k;hUUOn<#(A
zOqBmNJBP^5*{0_>La_{#KXGZ(KV-3?thId21{8T;BnV)dS1@2CpVLu)@-5shQeSTQ
zUHN%w3UG3C{S$e#Hkq!Uu-z+`N*EMoTK-zqt}OK5xmC>&3MaaFd6j|pVS_6=3vV~^
z2m0uBP`~#i$*KHcV#eU~!PVd<8wtnzc0Rr~Za|OT-w*b?{d%z$0wr5ZjeKr1H@ETQ
z*KRV(x-uDHi+FPdkXyJ^OTAtD2^=w)a({y)Pd(}OrvjTkj_zJwr3%FWB2L?&L(AjQ
zJ0ABY5|g#T7q9VpmVLC)lCO+-n4jX+&0_T9(ZltXU(xc^jX}a-^K7JErHhr)@<aKY
za}BLbqu*u+j<%Lc%3I$LJPT@bb8)zLGu*;Ku(4|X>VPO~o<9|v2GJML=O+~`dI|K(
zecf>FNF+hw&1UF`bLA(Bnlv)p!tAYlOixtiJ|il(oc=v%!fl~I@B(ieUW348p4b0<
zFyu%JWM<A#OXG04IRp#Znr>A0B}C{Q6l@-5FobJ;r^RZ0XD)E)b<Fug3wxIyvfkQ!
z_}OjN%1>+~wc<=W5APFnNZXJ}D>591D4S2@?zy^>cO1ST<5IwjrazF}FJ-1+4jfKR
z#;UXFU`8P^8B6A3<ME~ImYopRcZkp-*H4GiY+d2k5)@C175#6l7q3K-zVd=3u$;qL
zjAfZ3!r`0&c3=}@|Nj><;BufCpoBL3*49+Go^w4Au{&o#q(*XhaSSm1kAEn3rM}Py
z2%GB&NMIfL@$!5qI`!JN9=?KP<Ai57Icya7Iw*rTvViUS$vzamG(@X!Z&{M|p#men
zuOp<Xwd8_jDAEtpS@V*)<^vvM>KO98`2p(-srs=I8#hYeyabz)CS^r)BKQMeYp(je
z@Bx!UaJ8Sr+Yu+J$Hi<@g~}He`rC4}&DDp{D9GAM*~jCePqjCzA1nFn)!Gk7W^8ep
z{Q^x@1xu*iMLyeHN^C(_<lf+cy<d$93SH@l#kt5WQNo)`{Zd#-EyZ<0#$Q0=YwBbC
zQRNIhfzN{t7U7iC%K4scqm`DQ2hUpy#c)qP)$c{=u$CIGVMk_XcK($M$k}DmWZ4T(
zdd7fErmppWV7UI6&ckh}CFw{vM~FY$+kgpG$asb1Fw(%U|HSimRv+##QfbZ^0Xj5^
zEifOnCyY01W)Mo|>SzfuT<)*vmpeg=6g^d!j+P&tY!VdsJXo_-3^{d~_rvcFL%8qR
zN1tKiRh;MtkCr~tQ{QD!`AYoetM$&YXr>`g-JrKNDyAWw0|M8(;V3lm)CIAnY|{C!
z7GYs_bLZDk;?SjZ!KTHJXs<FQcjNx^5xHrqk1XcQ?uYA{K3Y>%Nl~tIYYr>%9vVJ7
zAm?M?<L+a+uOG5$bnas=?%8X&%j4A8xU+;5eU~?5fX_x9%VBlLi0e`_67xoHyYFNZ
zFOL1O5GQl6Kd0PYWvRD)QU9}e(fWb(Rf<}*S7LqwUTpH^ihknK+u8c)8&c$F!s7gX
zj{@;1dGHyR%D&>kV!Ud6ZrbI=>SGh2y4ePK6-hzq`h8gPq+&vytJ(SecS-RU<dtWB
z8>3gFIA=9@_=%7sF#yyL4ZhXg5C2WRUM1nudGh)q&R@f>-_&y94osR+y%N2w+}j15
z1_|5AUI6-7=I}o@Jel10wC=_~Js+3wPg49T**$77_smN5=gNmwEd^nJv(LqqbyEIy
z&a0j>`_?Iu9CH#IdN=x;jS(;Ztd*A)v)xi+65m2&MxDJv?zN3Bd~*^bFcTF9*`mLx
z@D+XDy&1wV^D%JLyLuh#{Wtyc^0lRFNi)%>uBvT|p6)|wB$~}gC7JK5j~?T#{$PF2
zcGB>p)|GOBkx(R&jDb{!S$|2ldTfPzK31fC$X6w<bUgSkHaEaWpGrRy)x}L{4%Qu1
z0hEocJU9K+l2l|ZO0$flK#8oPzvfMyUd)TF3Shv+p4I}hc8WiBCWp8`%|&D%{>%Rl
z-U&&KJQB%B^k;SU`TK?FZ)b_=<wgYXIL=LyLEqBJt)GTgoU5<hfgyrEc>N(0uQIv)
zi|y~o!L;YN%PAK*)|HSD`FG?UbV90=N$<Pn<)AiTX(SxO){Ls|aIi&vX2E>%&)#o&
zy*`(oYvyZ|%Ck?Iwu)c?jRZF>quye~{wXgY?$4w~>EFWJ6w5?tW&C;JwY3#GNszQ;
z<`z~CoC?|;ZtjWKOqNy)S^x23nH1uuz3k7Exyd?T+-QaUXH6a+{(i~{&vo^&#vzQV
zAmBvL`knA?&$BQof}v3<H%8u6%aILAr1}VM@=k`VH(vi5>LIi(G58h;UGnO9`f(#|
z#H&n88j9LUmo@!An$E(n2{+v0Vj&_-ML}tlP$>b45t~IQr6^rWN)DJb10_Uaj8G65
zOhiPayGvr!Hpww!#9#wPj~;RNyZ3Ydi1&P+_j%6w9%<-mm&8iS1GX_M%oEM_un`~=
z@b;DHhD|KX@3T!xX$OEhB8(DweW*pPl#OEYnK-Wh?1}!(GnGH}$9s5A(gck!QXT(l
zG*w<7)x(yRtnxx7*ALyKe2Ti+jgISZx1f$9Q}4~rEMG_OZhWDxum4#S-z-^NUl(k?
z5aMy4ALM6yIExFh4O+>FB<*1G8K!uiu#-=ldy66=G^!?e|2@NbPa!eFUuX&f6(@$v
zgrhYzdpwmGlm!L%DF$u9!dP$qhv>cOYx^6T;F|(Rb4LvR-G26l;}jk1cfbHcVB6@V
zsbj?x3JG$&ZOt#uUIf=v&3d&mfm2mg68361X^kY|P7aw#k>JDry#*5DD5RTTEqFPQ
z7e72TIqv)5Q=2k>O>?TJoILHY-AH>pn&TvT?v9@Dcyi6R=R~gNC2`aT&gx@-!ZvPn
ztlHmU*J<KS{u*g{_5nlL{RQGWGKpV388dfQ^-S-y;Q`F>@Btmf(i<kN_Y<iNJ0ST%
zOC|0MvH;j8;<wt;BN~)BJiMGi2}iAe20dSLNDD?{L0%Fkc2^Bh%j<Yc00eT9`f#tS
zVPxVIa3qvZ+0=`oo^>o226BpO<*7ZmrN$!yyh`16H^>U)m=IRje)I3|UcSaCKY-Wb
zUFJlz`IA0Lq2Is{fA4wDCA(&0O2x(=dVw)AVB|c8P2tx!H{4!M$&Sb4mQQm@#hf+d
ztnn~uE~i6<#g7UG$FXmu`1i(JRQGInppBYVA#NtgXSy4r*5!$mpfH_)6PCx#2?wIp
zocsa(i7yiN_C!;f`{xx^mR_D4CjS)^*+HoYEBY)6JMtrc`dEK{?l2$8H0q{><e~KL
zosnR&q~I>@5MRohXztwE8kGNCX!dNX)+ErKRa^gsv99)9vc(TzIag|cElcj_P{@GM
z+1L{9(j1&tk-tsfXEDML$V-CmztSw*Qo5}Z(6ie<#;Q@kK#>=$^|i-%%<WWPB@P%0
z&F7q$H59GBsWQ;RBB;@=^l?PjntYk(=snz(v;g;x`{HyRZ}}SRQKmnK)|0o?m$~{|
zL8dg!<ehnb<<-|0BkOR9SCDvxFMYNLuPev|P+_#YVAx2jO**+v!N^rQ?pYT8Te;q3
z!6E;&8|%ILts$gl3BmZ<VpLc;dBM#EuH?F)wjfgdMr!T3e@Yo^+S%POYh|CfD`T06
z8j|p(!h{tj)Wc1g9*6#&8wC9>xML<_cE98II9XEok(HkAocD{}3(>fkCO6+NE5G!s
z;>s#%qfgygw&Vy`W9}58MG~HINpKaPedevBtzT$%=aM*EyjAK>5#!8u5LA~5nNMPL
zPThT9Dp?zK`{@tOXE2XaTMhM9Ki7ZDteTpi7Ch|2!ZTjQK3tu+eEMsfLJTXi_}s%-
z>YKaS^d~sTmGjy^$4BBVpy(>J`twgeMMn(c7oPC$o&_}Z*V?DV&%90jS1WF<&FUH_
zopW73y6tmGn{<{>8MbUFtklLSu&kF;xKk&p_LJ*L&W20#&i@YN_GjvY)^=zz<-lv~
z?s4PQ?g(>VfmmN6{ZliG&q=E}sk2apG`4L{8ko7x_sZK=ojIkLDQ;yZLoSXtP+|cb
z7}<8%<jHUGrMV-&9Rk_P<dPnw9Acu&JrWMBJA)KS606HN?{UvyL|~)4lI970(?YN)
z753%9*xGZw-(1V1Dir@3u{Q#_ZZVn%J<d*LCU%u$A82m|zBjOK-JEWyTqXHq4}r5a
z5P_2dq*w>9V($yF=9N+}qh?9SZfl_`?*6WHnPz7-<fOu~u<VC_R7hK@&A6?V_#R*N
z-zy5<*wb&zi-6sgbIvhJ2nY45<{T#iV$OoWJ>%RVjEwpA?}m&KN}b|RCq5KpT{Uu4
zGUtk!pwHwdxt%W(1+_Jw?9*3ycoym$mc0;&d8Mw>;<|P(1e=m{p5hQp{=sIA@B*Y5
zyYPN)G$q%~^2|5?@f(yG^sdDYZST@wX-(+TXgJ+!?mx487$aU(-E>Uct4pr;U5%qE
zy2tiL;u*c8ZH)UT0vsp{I#^Da(<CL~P_ef~Ztsj>c#g{>@Wqxi*l#ghTxSwv22aY>
zXP%M{r0ul&(HX+FvN^cX%?#!>e%}y8GoRv3VNk;Ui`l9IXPZd&C(T8<`iRa~-9fA;
zrL*8;#^~OUM!^|-6OV=yl0X~|wC`up?4{&B)dZBH7h7-FxbQoBIAsSq@f4TI-_nHm
zS9z|LlR{w|trFTF&1RF!rBq~iTw#pj4!Y9rUK3s^50|t|pFY+mM*5$qNn^)Pf{7h;
z7Yn-y>j%MUFjTne-a%mVYWcnNn=>6EP24@qxAbhxHaa7xBrc~ztUP9~dGeY2aK3`>
z7A?SNFJtT|VR+A<{fIy_-Y%p=-jC1D$lN1|iSWYNktEqVyMsZ1%PoH;bpZz_v@Rn0
zk79C><D&irltrnd!fcc}SbWm8IV!-tD}%Qvg5AP>cu3)fo=v{_=S3cQ*go07|4~rD
zv-N8=a()?_*~xkNacr{k{Lh{X#PtJ$*M6@ere1%a7rKI+tp&X9yaNJ=JV^1_oQ?r|
z{(cl-v2Is-rz~KaIm1uti=W&_62zf<6+m(j7bxA5ys0up-aPZ*SmKtyEB?-cz}QRv
zBEF4W_72*MzMo*kQ%T!mX6se`>)qrg!Gy}N&!5&{n!GrdKXU8coo(iQ?SaS6>=|gS
z=Ju7$3uc`CtMUS*Q>m+4=_?;4J)LweBL(H#mu~R;IF*<c3dZq5@78(}O21}x_Sb$y
zL*#<fFKxC8Rzu#{KB<<hEiFm534S3TLkg~Rla2J5*;Z$(+^ZhES`qew^fpdb*FJ}n
zYBB;8u8uVJyslP=shS8Bo?0?H9kJIPb+g`zCHYRh9R%kUc^lD|bVklxC3;h`c3jik
z)7c^<^GzqEede*p$+e%xe>g0wR?E?L!`%OrLmw^|a9w@jbB#qW`eW(U?BgDCrZ}J~
z6R^T7>nmHU5`k>G5VrV8tLVIvRhjM0Tr5wPlZg2%Qym%9y}i4QupKn{C~U+>saw&p
zVXnCPjkye>e5EP@=Raz!!rY?Ab3E{57$J~fE&Zc1`YW7<eti9HR5fR}l-3w{y0FR5
zxR%gQkQ&(z9YmKg@Mk;*=AQ2`!7kaS00jp@SZz*6->8Ck&rxcv+%wO-2mQmGBJl=m
z^gW<sxCyZ3>ZR(_`HO_|aY<!8Y~9?yF?o*kJqN=LtWA7i>BnvgXGRY3OuP3Ql5&sw
z7n&K`QMRpWbe8n=x#d42J+k2VIqS6~*6+jkAqVJv9%XB8&J={Wv)A}X0}U<jh6-GH
z-O$GjhuIPbtXnju&!=7&iC7G8IDQpbypeKwd9%L05y2sZKYgcyx*Fi&7dTPp?K{x`
z4QZfPLgIWYCLT6>Yce9SOWQ7W4&WEU%>|v6f#;-P9W6J1=KxRiSz#-T3ZBCO4#Uw-
zIx^%VlI}s=D>vcd8+WSd3M+;w=}SthK*T!u%A<Yz`I9}1nYGNAx&wFadauzk<gq)D
z-JIkAfNHY(M4?;M)Rw;i8>O(Yw7AJIjhefnPF2vicoH{w@NStTfoDc9c7XZIJ6y*p
z%wiomfmKY$cM%YYg3-=u5M(V$i7j}e>QV4XnKUwW#J`a|cBPR;4L5{2;sCwX1J0Kl
zuk0Y14{!c)*2=tIA8^oh`ZshbcT5n4Q@_Kf1t!8UPLg}Wyv1SiuocnKB&-Lv>Hf8U
zcx1Ydp3R^#N=`Lq9#E$n3afNQUwI=Qh_POc<kj3c$D*=o*zH(kDUm+<U?(E!Md|g2
zg~>Z?KT9AdP7(ln31jR)EARNG4b>qjB%e-1#5WmDN$H~^=TL6Rw9TfLCjX<QVb5JG
zW&^7Nf%{IuTpIS*vx+{tBYK&Q(fv7$5D+7Pl#Sy@2%O9fZj_pViY<XPs_qslyoyCr
zf{T00G)+b-mBBS|@;_MU-y30&gc%#`;NZhqMS(afvp$-H19gWLQX&}!cQC&|2=+{|
z48NKhH%c|Y06^ep(CKmL^4o4yZRVnt#cUipTOZ8Lz})l{I)z{tJ-7V=g8pBMmHFG_
zLMm54dt>l(^|hi?dKUj==@3Y<ThmQ*Yih-K32DOn(sV^IH)ZC5#glHjM?bSQ7d6n=
zwKzIv>xfa5L)d+0(!9|jnyJ@*tvAq?sJY%>Qud*=g*_S`Dg44O|3aiLW$<3G0880x
zf!ZtV;x|)F2y~&xs+@7py7Y2{le--Ob;A2vNp}eUwBhH7kKRKw`ao^_vR92p=~<eT
zt0ILdu?n!BB_6~uL?mYeqs@1JuBwCf_^iLJwj@TnTYA$zGh_6HczTF9b0Ws9mFr5r
z+P}Fh2f3}|$VtKE%U6xvc>VZM%g8U7*YW;X&+d>mBfY3?lZQ=Bcg%c!j{Ahf_9(}1
zVDr;2AdLekHK=Y%4lSK6*fzlPfz^7;^SY_!FYAQ&LLKQ31nm6Qhq>-<wJBm3;L1nf
zhC@;G$+bbFgf_ZtW$&>LtX^S%dXR0Hl4aaUxuUhB6E7eD-n4(c1#b_{NTu{gEQ=s|
z4}qFeT2!X=188-eWx}P8^q9+}xFk1|;MbCtuc$6Neku)=9y;;w$Bw;Qa?HO<dD`xG
z-@Y4MOc?ZqzM0pbpBv}?UFrq7&Qn14&5V0mtuEf2@Vg`NC^BtSI+KFA$qnrW^?)4R
zvvb<BR6^a97fW2ZlGFd)HS@yXOh-Jq!+jxh`;)8Lu5>!Ax6VsG(i_9{^?oD4nxHv2
zel<J%)y}4n7*VZ;JPJ>P%2jx2XRMZ&J+lK_f5_+YGSym5bD&=q50f`gmRG<b3OD-A
z!rx##mdgx~hpC90qBoY!K`CQjR5RsNpOyZjfw!JJ)yKTf4rc#-rQGR;9Cv=4q4>D+
ziM_ZQEusB5A@$PGlP~?c-gcGrzn__`<+E&euQ<A61-ebIEB4lz>nj-iI}(%9ftJ7Q
z5wD4}%j0J|1Qv@Va+w{-beOm$yPEg@0~MU}3n^*QHn7UPXzrYu^7glP+jEcG9hF*-
z^fA0AdZA)WIcUK<iqG#Ydz((Y{X2mTHpqipmXuwVRM`&SG`j`R;&+nA_1^Gs*`U{b
zA)ozP6Wzic<9DG<agZGyk-+a-;Q$hUE%~f=)6ov}Ya-^_>Q%?DBiVU(rRjQ-cAuN(
zI*S8Jq#dn4OYGA>YiJ}~Qb=ZHbIL2@rr5>lGRzFcM@#?9bx&bi-jrxmkz&@8r2j+q
zlGc}WPo$&7%@`ZV!2h@-a7F$m;3}~m>^>FT7JMxHqqtP*41L;~{&>?maN$eHBag#6
z(&3&{yL+}(-ua?*{Z@6Tb@+)onYi)U*NBWAUaJp;LR`hRLq?Ran`e}$@-5ME_AA-C
z2TiWp0+Ty6RyOoZyWT+8AdHsufkb`y7i)|%zpDka#Luq6=&Ee}g&VToRp&tyw2Jxi
z!HuWr1LE`p^Ug`y4W1&`sOTqCbE9xX&{bwv*nUuNwc&`{g<L7mOV-Z}Ey#gKhc$o~
zS^NBGRnLey{UG^>$ra57WwEGq8J8pc$KRa%DE3VmmmaiM7gyQ~P$OERHh&_c$-#7V
z<jOl8k$zb)!6(9r_fqVELwq?YER7}-5uT~NixQuj{FWBUl2_?A$qdu%;Z6B9BCmLK
zV{V@H9J<7)-~V#ecoU^?<)+-z`~m$CYZOgH>fo8sx=QEDQzTU0!5!aIdL>Z0wU2H3
z>zy~djYekkJ!BlnbH1mw#`B1QbqsU)&P-}ehrk1O*AR@mA;PhTra47F-djZ-d#I{#
zlX_Yig0oB9IXO=%Hh~=y{ue{QBM@@nr+TnV^V~b!*c(gt?VQ@{*zK$)2HH+EbXvD?
zlm1JWZVbwRLJpty%1NW%?N29ZlF;CN*TXr67CRy`;s0UT9dVg)6n*}DtRrc?BcZbZ
zrVsQIP*8;Upj~qPh!;x<;Z<(FZqoxI$HO|VRnEf`mc(u^5JzTvmIyw8)J;!2r0+kX
z@|1fb>&fTMPe91X!4HAl^jm-?sj=dbmra5eKTo8&!I&ofP=TPM$>tLPRAFa78IH*l
zDug3n_;LJGZZTpa&kHUvS#91F^%CHHkpzufP5<0Jvhp5CRt(a?x`|MqJ}|If{dlig
ze=aYCRbn+e+DuUkZurOLHIH^pa&+C?&ND)L1b~Gq`{BpGfp3pNbhlC@AX#0aenwF#
zwBVgn2!BsSp^G+nC`Iv<uvM~7*{q3s4_Q~bD?PSCcKu@tPw$-4>@ozbC?t<hsT>|+
z3i0^|nFcX+&ukbAND0E$?=LNuI|i?NcQHHfqxcq|N((@ow@9zOWm=}@MUt5v=DNcQ
zR>3FYmk8T*rooRV6+%ZZ1a;mk5sEKI*_%qArT@4uFUuZl)B8yaIdxy{8sQq%q9uGz
z$alFSZc~w1WL5ji6O4HRGG~Gg308OJ`%U-9%nix<s}Rl@v9xr%-bK$o2h4~jP>T=B
z;Eul|*$JmaA%8M$j|vhJktr=b;{AcAF0ycKa-ZuWm47CUYx93-w7dFsdCatcWhmQi
zZ0M#(*Nn{gt*f=XVX=4li!04A+_v*uNV77u&>{fg&?o>(zeeea%sJOz`$}JA-DGdP
z|2HoOdnv%dtv?z*%o%9j3w&~GGY%l=QYbpE0{SNSt1SOn5w?HXS(~r}`CU3x8k2u^
z_IY*3xs9c?FnL9<PR>u|cOEL2@*zx!ho>L4{+j(#v#kI3)>n*6uQIbFY1a+3ozO7H
z^sdT}c(vZByY{|9yuu}v?M!@eVyzgBqftX4vG!Zl>r_^ssw}*um;(`P6h4^uj;pxV
zd8`FlUCqjsyc^H0|N3xPnzmq*JF%`FKJ$@rvW)-zmDqR3db7izvPOU)Joe<$#@9%g
zAI3n#?7*k*2lrcL+FvVX5fi)X=cHtz=F5AcSDvm+To)5Wi}O@wXZr}ONU_6n?`ed&
z3MPvP8X>Iz{Tw9$RhcfYd@1RgF_)Y)xpYC|VO_OIa-M)r)mNKoygs2ZM5`1pOHnP7
zPQQt~k4bb{m7q;iw^7PL@Mc#N#HTm;sIz|cNbQvWPW|o(m+IBnNJyg`(SJ9c1ZP-4
z7>pR+>a_-$BFn}ss*7bgm0JbB-#r5cTGj@wZg=lF3e<|4l@<n4!E9ol6r>#GR&A?k
zWmbkFMg0Io73ztNiCg?}*!j$}ywcGnkhCuxp9BOY2ab$`329T~j&W;@;kf|CLJu<0
z;~CV`=^g5&n1Ql;s}$FOZgOK&5=&!+EdIt*WQ+G{+~3?dRz~}Df#jgiv<E7fxs<B@
z>f<*kHx2#+SK8J_LOzOvmb_n~Y(`v-Y|=07J8!JE+1-pew?k-o=JNsJQm_&Psdn-z
zpoF>UJ&;OuLs=~JaqP~ML9TdQkR_{Y;lv~s-QRSGPEdwL)VO9>p4i$#Rot{Y-V^7!
z+zN|p9a_3z@s`UY$#gxrO(3{o=RIdNSpdg#M}x;5K5@<#*JP7p%(IMOTQ59TwHRH>
zZHrB7_UADp%M%<bGLlhE#`vi{uk{C+sRDCNmjiJ6$Li<D6)5wS!ERjGq%!N;0NvhU
zPg>`;=i2I|=MfpR(Fmr`>|uRRf>eVKTrKLo?g#IuG^rhrNuMwVXriTv8ZRPKO#mSP
z-iB%L2lJ@~KOi=RKo2JriQIrS;tf1IK4kZA?YHr0{f_w=ye3MGVSo^5I3etHqY<5G
zL}o0Sk*2|ckNjdC^$&Berf5qdocJ8)+<T2A|GMheyBAwpFB5V8Ujg=C9qFF_`x2Hr
z4E~+=ZZr0zj{tu(ap1^{utcBNi*ZAyK)I>601e!f#cIz?+vOiMS|`QP3`cc}I)Bhn
z*wjD&TVfZVlPh$3Lo3IVAGx#lB$kYs{_SJS?zw*sI`|u^xV-c9Zk<;|{Hv65pSX+5
zyAl7CPT`r~xr-y(bjp6pKk^#6lFs&8;h;^bbXVn8qM}S~)`G<9py_8Jt4AXrFA8RF
zo8^eI2M^t+nQwm8%-{ZGqG?|KCALdc1Ijk%bDqljxFx;XRBVZt`}(F1KD}@AeBr0a
zm(0wOMfClCIeNgQ=hJ;7>(GRcg0GsNbzmV<jLL7B%%yN_jsI<>7psRS(tdTL)2+$S
z&p#Ads#pP9C9gtGti|*)k9WQtdK`jV?wwZw41QzPBWoH$PpU2)2j9-IK7AC@a;I+7
zLAyO&_aS%@KVs%TB@L95$N`=fE9zxV1M_#Ok<uL-@i>3(x$Nep^n-WP@6?lEmtV6p
zmfh)N7U-{rN(J^|k-AJ&-HGq{jFjfs)4~$?<=u#n-Zlp_!!|VG4Kv4hpM(uUyuS;0
z>ig3Tli2Q?y7{t~#aYLHxUiR!pX5*YA>?f}f6Rw+3W0FN;YC%lw>wwoHoP(Sr)7MS
z#)@+mmjmA&0_x{XC}&Os3`gxOViW$NhsxMemLuYxf3h<RgpXv-#)q|i4vj!sBgZ5?
zkc;^dK+L!M#+VMF>eP|J9R<7j`<zDFdJpXhPztlmEAHx+{ihLbodyV-n@@(x&xOei
z8NjDY*)d992bYA;4AagfQLq_Jo_<p$I@({Kz6~d7TF>J~^8?3K1^Tpo<8+0|F(%*}
zo*2L11=B#Shy0m8jh(j@Z&OeGqWo-$iQ*Qnuga6<)_2}2;cL`GN%AoXcjMPA7jIP>
z`6d-;%|%C*&OdnnR5$3h_Ags`Ie|FRyhXg=)AgV)NYT2xm6B@7UfU@V%hmk`S1L|z
zQ;rC=2Iq#upMI9OSfwsrkQp*oM!*C4sz#O)o?#Dtj<p&j+6J9d0Y;p*Q1v1;Z4*E7
zq(rd_HSv-fjhBSnBr#9-lKyooW7*0y(L$>H6cuyn+3zfkU-Hw#V~>{<aPRU8<P%Jw
zmX8Fw``49(W~@|=e8|%k`+LK9K$8_oTFnQMK@0sY{}??wFHTw-X5ce0Dx^Z+W)MS)
z8F&rFJ<(C~;0r$UV{9QDvCRXgl#2B0TaU}*f~CK0+#@ZF0x4VxYZ5KWG4y-hykTCJ
z!c+b=j-CN}=_$Kkym4R_O#KDDKOH--HZvTN@WGnBDpH2$z&E`W>n7OoE8B1y`i=R@
zh>suRx}UdptkZs}1?a}J*89;Z#G3A!YdimQ5>9Z8!RogiN6FHvOW~I6c&~mnv1B9R
zlMPkJc~Rq{{7}iM?DfVN!DBYfXFDqh*C#gn1;k3!YyC4J2X^UHM=1Y%?u^CwdwJlZ
zm^OoXvLi?|tX9~5bmAZ>6ICM2Q<+MFRR*$<gG1#Nrh}i(Zw91i9h3-8Vc|X?h$rd|
zmTRlAdMNm69J;W@H_f_M-o0v{GZpYLz>S6MHagX$%(9m(^G_cGeGjwRK~{q1EBhSR
zhVitdK*{_`a8A9+!R5$iiH@d#F6(sVX|6>EfL1>)$U5G7wD~_iIh0aV6iK|{yE@5Z
zTufD79Ah|TIwjGJC_KUMaG}2jO6~Nc9%pz`512CQC&GFUQ`=n_`IZnH9s$_l{@#W@
zh+hC4@Sjg(ac_Z^%fNpIJqX@R0(6$65r?Kj%)h7;g{bhMbdqxftih+If*Pc7OHA{U
zTRO$TTIXLGSS+(8K{TVuaXh&k4qiohsw&vJAjL*p7)ssmp@%J*Wk~nnhT+b3*>6bD
z(Rs&HJa>S(7VN4ht{_1h+uFVrq2>9x7dvdO+UFLj7$GAYmBwWSaz#5v=t&jw>Oii(
z#yaoqpQk7$j|Z-4mU>Ak2z!mXL(UDw@!5WOmyuTA7H8}Ee%mc?jvFO?>2ciZ__(u7
zZ+v|FLauLgQC_~o3NTw<I!lA)m2$CM%wvOddf&gIpTH|ddmOv0PQq?;8Rn&ikl+1S
z_FvD5N?AI|>j)Eqr0A)$H%34|hBZ5*&r8P;A1o>I@3}EsKQ+`nCZ}{YMA1&_ygsU9
zM;F>iEs#_$`FFlA1c)5h(WnREy?>vGTDxwvun3eLe0NydI&<gCz>(wi>lB%Og!cFB
zK2~;!bku%8ZIt^)VEZ={Ny8VY&gGw5Phj`2oypkhgZqsZNQWFpDF9<$<ddtbT%Ks%
zIIVGa>!sUhkx`+rr}JDID*P<w67MV8j26Pu0KjOz_3?IIZ18_E4CVZFvnPR3GvGZq
z&r^g~SjkpjnMftw_lU5pofM5Kyy&i$9LFR+X!&k_gpa%x2C)=3lEN{crbku@fi#!5
z)O8l)S_O+{_uzti4qid2jN(p}Fz7{>;gv_x)ue`6QI|i_4Y6fp$y(RE3(RPPTC=$L
zPmx!T)<3^S(voSH{+Rg+0byyc-Cn%DO@Fgwj!yB(GrL!IRCB);%Bk?MJ>$}3lwVU<
z@7yRQHS$g&f(d(9yC7JVxvTw76V;sDdDQmk(eDNS&9mC`7Zp7p$Y@r_!mo`4>biJt
zR5l#QxwizH=5)(pG+7!e$Js}AAAX6Imne)XWs|=$BEmslaifBi-nK7EQV<4TgYrs#
zSH+J{_@8|5sQWQ^s@hCfj`IyPPTgCuw2&7EXJc*M%p;SPuqAB<I7IOk-m9EMZ&vxf
z*YD<~VcTKyxA@<i3J<6^(l6w7$wKc6^kK~nYm3iy2|wO46O=AwyB>Doxe3{@v_Z{b
zT>dIY%h^}+1rGLXU+OxGZU(+q`Plhods~qqkn+E5$*QCKmww1cdZEzVDS#rlxUqX|
zdi5Hjnz(=9w?0zO5J@8UF25?{2{*|UZC+lUeu1YR2Kw52?e+5VkGJjGdj)xq0uh;K
zxaLuxXk>;_xno^@XpuK?v%z&0ShPzR^vcLZ?Tf-dDM-HRZFBr=WYfa&w4bt7<?5i4
zquRtckNEfqQ5o7OO%@)nbNd#ltC!3Hb*l;V<F;U7QfqZ_L8&>w($R-%d9T*GM~!hs
zaQYQT&p#)BJoeSme+e4|J^3~(h^I}Q+E(f~6yKhq8g7oau~{1JzN=CNE3X0VJ>Me$
zzPF*Adg1<Wa}}7>lHrY%YUiR~&ZC7D(kdxOA7Usse``1AK{UK}z4xNs3R}N8Cb<N<
zXRjc~Cj@U#;WbBH)TyNeErBStVB{J(gCbCY7MM~1y@}B(7YpIL=VcRzmah>wT~$ke
z^e`o00)bV#0B<;0@Z4`r@%bB<ioVu#Zmo^1B(d9ZkV3O%sG1F*)jU`Dp7q;mv^ju3
z&X)*-PSp9ksBWMnf)w1XD39>hcQI<ixwyDP%GIeFOj~i=Mwq~%N6=beh=P=5ql^^a
z#0`f303mS5u$(b!jVYvV5;|JDQ9g7_`X|vFS$dwlo>t?8z&XtYtarF<h(JgZ{*!<o
z)v&7mfDqzF;WzL0!Nc$9a)9}2EF&G}Ph8u`;PvOVCNmVVbN8MS<V2cva~By*fWfYe
z?!Ss0%HTM2rsaEtV%VntZkO>zBlaI6{F&i_`?aEv-2QgpJ9UB%J;(ZT+NMJSL==IZ
zi<;gZH1T@jUYgEtXm93R2O1y_VR=`{_JH5#XU~}_&v{>yE4_Q6N$jGPMFiDC9g<zE
z?$pABsEDOk+IJ(b##}>uPd0MBJ<I(YHB9-zL0?VX{oq6u+vnC=34CZ?C(vdyI#GJ6
zOz@0_)blR5y4(uzYb0`UVt+8jb#$dmfx6-OR$nwl^bF~5dd$tBb1eG;=81Vx$L}{O
zJolB;>2HVR8fEsfl5F;;toHn%C0HH-@A!-9!CtpstXcev*Qf)Om2J2^I7xlnK{vO&
z>Q8!j9{IDbIrfsNr25}1jZH!_B>vIX=t5$r_^a#EO|z=fm%FvumD{>%<k_2N?tKD#
zYfUxeQAz=oqS8}OH-as)GL?3-)4Ihu%8rJfh<_z@KuNIcrVHGddnUz4Z5yuozOU<q
z^^?kgB|*t|*wZDxst9e0N7{2jOZwkMw6S^Hx2Ldgf=5S>M+m>>%0uHH@H2+d;cn8e
zFUu)I$s1!uUwM8EOJAdNomPIC6x$`01G11iv=Nd--{Y**%kG$Dy?%@MI>wvVtTtGF
z>ybq&Z_*@2<^xo+SvR~1^!s`%u+CYi+uk_l4ucSiW1dPYgWG89e|C-gC}8Kh{`6n)
zoUoOl`?*AO`x<5)K7E&HF9S$b`jgq)Mzypz;TLMGRDAgpyB~tmUR+%ZGFW)S^!36}
zl*U!poTNxKcyqq){nZ?!qpfq^ovtq*b{9<_r3dd{>28QWE5L$nGqJ4B7|N_edYQ^N
zyhN6j-M8`@=Mm_UDE@TAEjRhSRX+AbB1srJ_FdYFyLvu43okM;_;9uFWv7(_tw;03
zL}a{2NySRCG+2;5=6t#B4;3+9Uh;J}OGEccGf7D)u}B%H0YDfcK0f0QZz~;Io4&z1
zg!(I9=#Mo|j@#b(5yF$C5<)aGs154dlD>C2^VLpWVQ;4Lm7Ld@l~|?WOH-Xl!4m15
zlC4j-G-O4MikK%$Wvc8dX1ON^Z9+|NSHw*?-MCHp5xk}=d^8=_^5X*HQQ^B~-*2Yr
z_@|04Q#0YlTo*INmwg0hyb>3*S2M4PE5uqKRa(nY#Y4q|;P}1Xur`K+%$wJ$9<(|U
zY^xfyx5I#XsF2tycLkc{il!RUA)%JFRfj)#Bi~_}x~FMR#m`~oLFx-{c0A!v|23bD
z{D!lnw5!)lGZq5`#mrrjJ^hR|8KrZ!SlRa-rQ4Gjj8!2Tuh4|2OvnA=NPeoxx!ys6
zou^N>>XgY!%)L#cqk6r0Q1l#GM+O14#_sDt4N&H^ovcpKGEUthAU0&a;IJy!TF-af
zZFq8xdd3wRHdu@H?34O>kRMRYgN7n2R}l0dHiEy?Q2#T)aZtn^0^OdFCtg|DaMM5D
zdxGc@Yt_3aQE-?%GU24+uA!7vs$hcmy{DylSa#dDX!4!X;hSyMA&ZQBgzW+-F}vS(
zvPmE;ZJ&eQ7I@g%P=w3KC;b86-CyRoF0Alr|4<H}NXNChf{AN$#xACd69-RT_UcCc
z+xP!HT<O&ux0Dj}cy3tZl>ujCO^84BfX$MlQHBW9#6Mv`Q(cUUW5pv=hRIWP>(Uk=
z*w#DMfWxN(B&%4ead8|VWHV`lIb@xgG^j7II7ac9XxME5pg{)<w{hKOn!C;8Qu2$D
zff>$5;{y&&X3UawkW0uY#|Sg<)#_cn=PdTp^2j^19aGzSHXpBx<*OAuA+{*k)bS~7
z&rIbUaracy(aU<jRAwu9(o$OMx3Z%`!p<Kh{buX4O@D%$?l5XPLpFCqvf0Cs4at^|
z=|=8l<#u5u%_Y72xHHc_r9TooqeHxU`0r`qk&@^e64O=YDemT-#(CTKQ#lF{@pFkr
z<#T?AAzvh%4MFEu&Io>SPrkvSzI0GRBia5Uwii%y0etl)svw6R$fEZ1sh@vJ>E}z}
zUATExuw5YM-lud}gKYg*LK&s(;8Yb$ttXp}sXN0_>whnv=LJxxh69~t6O$3atnXYv
zaMC;dugwt~6e=(6BDIsr$G0yE24qNWOj9@C@+U7o4vjQxNcG51I1$jW(o>itJ>mrB
zyAb@opA(zfd)}<N*AvW0AgKZ7W*^_>()KTzV`c3TeOM>(oj`qWs<{qYN$+v){;FJt
z^YYgOQ{jG+Lx1KYI-h^hA6h873BydA%&iuY`-5t#oJz|DVX5%$NQ~m;&z1HVI2&~U
zaTc$z-&;P!d@CopsFC#D{GN!KM-X}^_5>oD>ZYyJw)NQX4f_VsYfP_b(0?Kd)-4$F
z*k?(BP5H~;zq@00j@_PJ;<ux~3x3p?Y}+t$=!*enlQ>TyXr)n;sH;NIBL9V;V&5A@
z<Eu+K=f_eXVikwAfn!ttEr0Q|4T-AEz3r8f!9CGP!>NCc_Q^Z`Rc3A1_v{lT$-_ay
z>4(pkd5GJ?rOS|+h<<8pR1o%B`J+o<;IHiRsre@&>aW>{wI=i&YTX8Bd8h`5A0h?t
z8kfSijae?S+fs4>q1rwu!P_{$zt8LrP0^I5c52u5M1u|7a)q^d%unvgd8moMk$d^X
zLsK^=*KBplYYh!!QyNo{yU_V2VMtq%X;GDVs&JU9LE@i+_7*7rV*0+v-(Qbo=Du~!
zTTLzWu%=0Ni{~@RR`CU18xg$bG}au~FOxswAnjAAs$n;BA+wyM<u07$jESz1Ki}Py
z2AGQ@0REU!TD&kY(Gx$U%N4wghk`tvS6=8g%9Y(3>PY`A<4sv!s4R3PV&VHNJ=_Hb
zcJUL<JL1L78;o4^#@B~<$4JJd`oFN%N+19o7P%Hzzq`*lnoc4fMUt=?fehak^*^F^
z(NXb8)}LYW??)@t3n-BwV*E0+_0u62?>LRjxJaDw(k<Pr-<EF&LO-nN?#9gCg1K$j
zgD$DWgk%QYZ~CkXH-_EVxmP(2gcRnI^JJDueN|lnQ<=mB{hpYkIsOd+l;S|<;k59F
zL59o)qA<;`RkG-rkfPOm$8?twUszW=x(eRAVS?WI#aU8JhE;X?@B81+$y_l%vbG3m
zELgxf!$pkP0-s{-eUP^1MbJ8R1J`O(DE^xp5q5NHoQrF7<7me2ot-wZf(jbB-Of1`
z1gbWk?NOHgG8K~Zj;o%e#9l-&mB2BVeJt`_w%uvjS%k8VdO>yg5GVn^kH8}PG5@$*
zeHPB7W8r}D+3Fm>5$b3a23uK_ioYOj>zG9x`YXL`&Fu@B;)j43X#++$f*(|)$ls9V
zlI3Fjb;m>1pOHyB-uu6+|02oJj)zhEO!$vk%@1ZI!_Z<VMt2MV;O*AYR#Q9WuUhDM
z7X#acp*uhbH)kHM;B}F}$+Lt~P3!<H)GY*F2vfu#*Cas;kL2}cLmAU#?e6UlYja)K
zsj8N|xx-Z*O{I+zQUh!G)A>sj`x3S456!d?dh?KClfnhx$y2b@*>sD5m5KGuZYdQ&
zuC6&%6mwdoO}s<;>C|-ML(uV0-kaYKG`_ENvG`gp-!Mo{+iHsf!FIz^+Xrkg7?u1_
z>Q{iLlHpb3uZ4r%$9h*!j}`euea#rdd?k8w*z%KR&pUgX;PEnJ0{Z7ZyuZ#9e>&3p
z756Ka4a4|!5hfdPWQ{RbTePv7U+ny{UD?~X;=aoBy>JJ@!ADJ<K4x-IK2-~{!+<75
z+&H13&tO)r01EuFn^f7^X}V>qLtZ`Y@!H_N+xSZ!L625f7oJwl*1Fi?GkH0$e+%h>
z|FPJ1yc*@`o&lZV(5J|}9%;B-sQ-^)$ZXSABR093zlOtOKzLW{0Jdu)36XtU&zX1L
zeGLr%YRd{)A4(1W<7;!R1QKYEw9K9U@cF^bbi2EpE4ioDwXAjkyv{*=fAi(F8lUw%
z(S4;|opL&&R(knDH<`$bH_LW#NO8Qw8H@ru^Fe-GeJNhU$(1j1S#m-7cfV;eP}lCq
zz0-lQWf~>9#ICyHxOM&jVR)MKykdMZ0MyqZ>)1`+=pRwu@gPsUZB3Ed?X$qUBzJPh
zD9XrBoFg9Z*%$n=p!`h$j}-iV>#cW(f5Gon<<eUfgdOdker+AaL!67k@3qhS!dtm-
zN4e?WMt00h_+-K(BgO8U{UdhEhTOQHDy`ag%d~WBqk3FzlY}*8UZu=q)r#MSTmO{z
z^-oe;h<w_g#3VT(??Nvn7f<#$q`2mcKXj`~cmEm8(%vroll}4^)fl<!N?eJS;!|G3
z6WfDKm)_{Z1!8dTbLz=9Ha*fZ-y`bN#~oV7WJ<5EaXI&|tK$$y8zy>ygsOK6L7GD~
z5`Pv0WCfBF`yZ%xGocma?mhhQjJQ`88naCO?D~B#_G|i4(~_*vL~?w+LjfOtXfN5-
zgw^3^+-H>wDz*N*y$U*DmTz9(0VTJ!1+itxWhYC;O$@fCOGBqID|_t|%X;gZ#zmF_
z)oT-czH7}uV$F$%Y7ot`5`r*i{IVH8?YMkn_l1TQ>GE?hchZX<%BPhXx@_b(8)bN#
zzKV#>gd{c=CO?-pH|$AN<yw>;pPgU7M=}EeEKf<R9FNesHzZnsYh76Lq|cPqKmolN
zwPCbD@Xq?YG9j6yB#&nqcZGXG0-?w+y9GxG+-gz%yQXXL6Z;Lx6N<G}N66~reW7p7
zx<bpK8!CYo^aHFnVPn*EN|Bk)AO6uTvlB#Lv?y9&1TV~w)eoHCuS&Fng25MDP`lb^
z;VP)KUGzs7?15{>taI^t+6XvsGqd-T6`k~Kv>9BiaNwKSoosv6sq*70I;G_na_unU
zL;}l6;PTmt<Xhjj;oFvwj&~}y{H)d#?D15e2476tL`~w&%=vjM`Y3U!#tNRC=`2x%
z%?T@Z`~$SZqT4^vabfQ^70)oA?skqtKsR>RCU5$#5&7{D3m)k&q{So(shr_A)ePRt
zL6*Tb7jsE~CWfym40YHJ_}Mi~`{<fEUOBZx_yZksWj>k~RtSO#7n&Tb1KoL=dp>SG
z?{E4#xL>_P8a;5}^R!+L9^csbK-*fWWaMRAeZA(QQ()`fW|ne)18YHgfBosfhCzl$
z7S~?OqpZ14e8GDM%qQ`_OsLKG-}2_?{tSNix!-KffsXsZc~pxCZIEfjV6IKPVxHYU
z{&e%hM7x2O#N`1zH^E|SLbSFas5}<S^@}As?g{Ys?S!V%*}ehag_nKmS`Rt74_>xx
zzWBX2;=o>bT{6cn;MzE8@$y6mr?Zu05wW}RGm8;O`O9On&w-4ERX5l3$MM;#Ul6ws
z(;jiW3(Q}4$T;OSI&UPNT6?No@)AqJw9**I!x-)1KijCPBsiGT2}bliiSGkUEzCla
z_?GPS-9uP)d}=19w<69+o3Xg(=}X@e0yUTZL@lL2Ts=Qt8U_sXRrGpa?QK(>vJlqd
zJ$SDBzzk@-TdW?pmGSiHjRn(lak&E9c|c5F>TxA_AM=d#1qYxkocB^TAA#8dx%Po|
z3z!WDrqJPkxWur^nwl?2h@!Iiz7B?0$kBYGC&-rSk7@YM0l4c~ve=#9A)WOe+T&2S
z?IZn2yK$XMDtPH%qQ4YK0o`t=0N6+N<mV9Qe*xL<mI94lB2^jsBP?FUz9}1t;HXWS
z_AY%}sy95NQRQrXp=YJ^O-T<EE5RerXvWSu0<}SYl+_26z(hAo|Lo@gjHJmZ=LTyz
z4h(FWs#Si>0nJtQ8aW7ErS>t$zfeV#7PvM4B>63JTb|yY4-kT01z##ZD{+&e474xr
zAJ&yWOEi8uJ^F@UO=F|n@;V6+d*i!daBJ<t&x7a2R)+ct>GMGzKY2PVdxZZ6)bLTV
zFR92nZ-l-6x1gj`LACDBfA|3ba!tV0mJ;m!0-aHd{y(0<`jet-Ze6$;c>x<~dAP)U
zIR@z;RrzaD;ongN0nU-0a8xKe`XP;GcA*>I^=B$w1#{8Vz@zWCgGe=N+swnaiI|0`
zcVLfOh2$G)Y1X}<8(*`6Ml<Rn{@R>Lg0cK|$RofduFSH&1&>)B7O)Zo#gSTf-0dX&
zwPj@a3zHj$Lnr5k_0G~8eI;K%|BZn@UgfSeR=6P78u#q#)N^Qr`dW=P$vAl4n1R~4
zf1Sn{<fOqP)Ch0<Zea6p)uKsbkUis13uNRS{{v|$t%Cg=8)!sEo}Mw{!PBc7EgAzX
zDT3CcO3e5NNEhg&BQ4za(<?!D)Hz<CvAb_)-8~(aQ^Ne+abYo=t-bBb&HjNP-YIKr
zs&<343kW-#AqGIwjC_mgx_nQs#r)Y-{ou<@Yv(=!2A(qwvM-rI-3ZZu(Lh}Zgp@N!
zz=rLnA)%?A3q(BH-%ZYA%?r!#NJVPUfc71Ilb!NaQ}IhOt8MDb`Gpkm1DT?&K@!3i
zAK3tAZdnzmv=BP5=&yThkq$(cl63^+{*iHgF7=FF-RF|^qh24xhl2G3ukLin2jQY@
zMytdIecsgVsU<@@_NSE`nIk4rl!3oaqLY|(&uz;;F6mqyJL7ujS~`zC*ih}~s!vNE
zEW+P-2My|FtlF4%FE#AW|B5M1La0g%z&Kz#EUqqgAQjp=b$xoECV0E&zr#t*Zs*Fz
zdMnJ5Ym!gVeZl~I7?H0{+8~KJTQiD6=&nH&#IYYElEl22?G@TJfqz3-8MH7PcK0tW
zw$-nXE0+OZ=xRffj--_7VRqPJbG>JYi?o_zN{&qL8@z8_NM}O^x0ol-%C6aFP|&Q3
znsS$g<t4}A;R|cC`#6U9h5VoRSq@j#H7A{HY@?i{l@jkMo_S-P(wUODnakc^Yx<S%
ztXjyQcw@e6<dS6FCX4tl?5uH6&+ozt?lp6Q0ga^?#H+n&7Wun0t(IMR+h=JG9*OI%
zW>)P!hAt}6eyi1d=XR%My4=(9ecXae-m+g&P^uU(EZQ^mlzlIi=K`4L3d+8cIX7x_
z_uG#@Nqw_Ri+c(+;fafx-8RT5!x#)){uRfXq<XedVkjYb!9mG%t5W})(?<afCm!Af
zVb6Ayc!eQd@0{VMIvIFWe-|$7ekJ1*ayWjaN&epRCJ}X1`=CK}p8vo3NMm5CuT}7Z
zioD3F1Q9^5aNwzEMUm~7tl+Im3y(BSjgLmGEc9C=+r4&aoNMgwAM*Y1h$Dv+Am2QW
z-ey5B@;RFN{Czq2=N56oV_>b_ey-@7#hhWV-CEd<t8<<H6QbYzWfAym<8&5tY;BZa
zd+XEQPZPgP2gG5t9z?Y^#b@J%kk7TCLV5kZl)w+DSZa^|OIhZ)N1cs=3TCgE{;U-A
zi)eq&nJ`=7B`v2nba#)aY6?n~(v-VY2ebo4QJXJ0uD880Onl_N_kNPzQue&ns-f8%
z$())7yr8PyWgof7D{B<~uSrV`qluWToZh|rQlYyy*w$)3p_OjcJsOmWj;pMxnERaB
zb^R5F|B~*Q!7b+usNB<vLr4X(y1&EbR_P-d{A=n^YPMp!KPIV?YOj0ScID!;*3vt*
zwd9$%{494+bQv-ENCLOP*u{j=P@}|pz@8N9YxMet>5diUXvFXa(F<$L%PTKX#8FOA
zGpceK{&536mbTsJ`T(cq=XJr#_QT*Cae@4HiKngMv2XL#;bD1Y&gIN@leJfVN1!Sn
zt2nR3=Id+V&^Fp3UA^Agc<r{p@ZbtZ!fT-b;A!oRo}BBzWN79je}@@u;9<p^g4L)y
zf`#K8>M`nPx$iCrnT?KcJ>O{D)pS_p9i*t*$N<yQ?5V+gNfQlE#<>S2ZfI+b@5~Wo
zBu!|`&mRmN8hHCO-DuwY|39qZZDvc<+k=vpKRD~UPtC!kks4?u1UAV7QT20aIKuvi
z5rOQRzmYI5TF<=<jPdyOl3(xaS5e1g?0(10ULX#4UUes>eG?Rf<t-lK>J@Lt)x6`U
zS$oDx4^x01MCNb`rW#T}iBgM^qvOhP$OlD9v&oTfnm!NwrG^5z#SUuqu6VH-Te6Hq
zrOximD?!v2ns)B;GOULEV_sygy8%H*d8D7Cmb|3(@<@CMgap#J(&$c2m|jr;=QVG6
zPx#7)@CBZcdZmhXC1tO5JJDPn-GA+`_>AdKpeoh&!`-G%jmuCoBdx|a=E<r^*K(lm
zw*f<)wPr)|dCvplmmEu1xADk-y`FwT9=Koaw(ro-zsJlgnI4$ak%+8zbu=+EwY=B5
zbj+#tSlUvz*CpTOvUWb(a*r#j?5(Inm2TnL-gmvtkhkfhah&7x+RSFZXn%r?i+JNy
zs{hM#{pXupjbn_?u?+9z9M>Vrnjai+a)vQ^ZP9Ri+Iq8t`UvQ5Yf4ITtw8Bsox*Bs
zT9#^L6=uCr>n$u^jxy@@^^yq6Wn8qNVJOwUve?$7l&3j=tql}{U=-wexvji&+#bKb
z7Bs}+No;ypBRw^vesJ((V{n-+v2_e;3t3tR7lOSPUL`95Q{O&%mY%M3Kk+O$*A%)v
zc=8ZLt$Sy(C|(j_hr)gRy^a)LXJCMTa}k~-xrZ9TyzYMAUX+6P{AmsAXPvC|ub<Vu
zV{t6OJ!M4_A}T8o?d~NRP3i{T8I|3ah%cXR6;_u2o&#pCjha60YSqVR;wKB1p1F9n
zJH*hRsh$sf&i<=kB9vRQsW$g-eekYOjvMh;oHqLdnuksK74wF}o2S6Si5C%&2QL99
z&Aoy;RgRl{pC5t`e}x79O$px@{bJ@b5Z3kt7tt*93)OesYK1LYSI2>}aPCA-2c~=@
zEk@qabk9>~VDD|3_fo09>h8&G;;hFr&v7zD=)O!SbohxfZJRvPcuub_{u`^-5OH|%
zbCN>#$8Q3@LPc#6!y4n(FpZtC$Gigu=F70b)Z^qc7NX7jLR8ERs8sZ<W3ywZnpC9Y
z_?a?UO4H~qtjyM3ePsnP3Dx(>yqvaQiup6L!Z@{N42Kc-+WXRCy>&S8YG#f_Z6*3Z
znAH!qO6Rkd)^0V}!NPai&)V+`^dkX}9Y4GaZZPBWH5CC_a2NiuGDygTWH$O?l|H{@
zEJ7Ea+%D7AMLhYA+7c#&sb4_fB2}Fg(~%%&8)C!*R5z+Kzj=sPI)9OjNJD0MH{7au
z6Dm;sk{?=^siC`e8d^E~%?l;<+OsD;t@hom7}u{o*~7gg$48wJCTDV_1Y&;KB=J{9
z#fh!f#xVuw-QN0>YOFKs`HKBkNTOkIY}QkauZj0M3YCnixR}o%w_a`xpT>tJHmjIm
zZ59omjKlMSOpH|Q3aJ$?-?>_KAQu)tlNFg>Ub0K|np3#fqT4MtTGq@#B}lYSq3=4Q
zd^sU^j8xB6@I^NAdCS7ZyjX#1o%azU!5@!vw)Wdpoeha<gYE>$DcQ;fQRpuNwXtii
z#kew-2LR@&6O3t>>o{IUlpQciKA)+g$6;T5ACWM|%aL@Y$EVO}t}pi?)#m!-Q6Bj<
zIxuQyQ4Z>8)VIiM(%oGuDM?FEZu_;O6lt=Q**c~Pb>@qt2V{{Vb{IYiMkQV`KkVPO
zC$}pA@x*eA@$)U7;{MO)=uWI-IL)_L{pgd$@cLc;SjGcpJmxjVwYhlluq72(8&tH+
zx+;A+Sl>kN1Ko3(Jahig{|84_l7)3lP^yx?09PK#&{=dRhI=Ez1tfN&>c8tbmb_a`
z+dfwX0O++n^LGPdeD#nift+g<GAo_890}x#CKodgqmvbDoC=fj=B)Jq^54n%T+g7U
z%h;Z<I_ceu&ZanVGzo!%>h)AtC#Qw4B#bs&G;}C1&omv34KcY`^N9_g-^l>$W3zk#
z-XMDQ#Nfd|<igcZ+ifVHSy`?YEkUwQfU;O0;%TObGbPO_m#(eG6qQ|U#e#a9d>X1q
z<940N&E`)I8gF1Z%6EDk&zA5*{}}c-K}u_xIUfsHze#Ql!n&aZhW4I3Ocv3={C#OY
zGAzB^W4Oy}4fq__@y?0IWwqNs9pgjkZB&*hrLf<S+@0U+H8L)?yooQi=3(Gy`FvL=
zZZZ-)j6}}C{@x&%UcRxn`09kp_3(JR)p%1h#48t{2!q^8<~MFT#wa$CBTxCSS9XTU
z<DF?n9#Z){!ApH*2O}~Cqww+JYNhUKzS}ig{1t<Ha(yZS-i#0=;KJ^8We%x<DK0q5
zVajo)o!J_Ff=>C2Y?*kz-TmpzhjgW!9m~(NbIDi5oHPlI9_OYSN^{=~{Pp>zaex_r
zexmp0MxxTq0KslbY(~1AhnTEbLPc#t9Cop*S`%vH75X4U`K60Q@v7UBJn_ee?eNdZ
zO6AiZluj$lWsO};Pl(@JzQ-KWiTxB&9KG;LmKTSxB=GB^(+<kvW22z6WU>Cx(|lA8
zX+_TYK+m}XW!q1G5mKPT`#qclFJ><SK@}Z9!|t^H818>*n1%c{llWwaLeg4HLNy#Z
zoK>6G=jy@2wFrgz?8o{kC(BuZ-!(dsD-U&pFKe>w8Nda8don>=lory1#p5D+ZY1=c
zoAWv;-NW_Q0u0~pw-`B!8Ws1+`QLvQTHWV1FCunYW4@|DN>#we1vO6{y_axuW`3<Q
zDqXF*oS8qLtsCQ4ZzdfOMhjCS^tUFq@?i;QUw>90<zD~tL1r*Uv#s|$i?iJ0kqRH>
zzVYe;t7qG6j_<cKRzxMYM5g|_fv-Ar&&$S}?tjYv9{}4xB)>%XE3jm{r@~oE8>~`g
zTv$jeI^K=j&XX31qe8up_=q7!{+?^bKJBNzc){lFu;w5-BKwEr;>Q%O|EHN_liI#C
zjt*;rjXf&G?MP`s_njQEt=<`lcZ5VoxyH65;wYO7wb8QsWe#jf%c?EQxS?~n`Uprq
zyWo+(`{1?oK@&k<xs^|Xw}^ACVQZ_5IxlbvA~(nV;e5UmhdA;Q6Be~#4X3<L=wlPh
z(m`?Xxp6h-C@4Gxp{2Er)LzG+x#n2i2QTO0-IEV$v8CYDGsCWk7IL1<o*NG_YF%?P
z=RB(onf`WtRWo%)l67CStlBWWoto(}Me>~CNEtP5$vHnLu4}ZtbW<zPxl>_$c-Iqe
z@+RE*D({ZZG4hpU|2fdCwa>OCUs|LdSKBIb5hvRQShL-lQF%Nd#+>&M#$~_GE#@>0
zVNdS1;0+^d>}w1yklg`Vm234jw~>8?J$1Ft9Qma>c4s~?z8JsqAH06>^%}47(?s8l
z$hmS04$oRHAC9ZnZ;YF_?u^@a-xzPc_4V=AJ6{`j?mhH_u^av-*ma-#`0+VP3!ewh
zdLVTk_?(&(`-wQKpHm-vfvWk-#LMKrXw0jOf!H~EzVYxyJQUi8i|ilg$@?mt9^=c5
zdtXl5&JllB&9k3l;UfGk_Khd}*}m#0<+6Rh%h<;CjCTL6H;e1TZOrkyg8jVHv5&Jq
zkIrCK9C~y1bKAE1wmh&u_GI`Mk~i2s45P$l4$VKihLd{T%Rk|`T)DJs+?-=yz8&n0
z<@k9w556W_`TZGUgJZEL^c{OIEBXP|zT99ZpGdxAwQEYtvU6;nkPz(IzPmUj_CyeK
zR$TU&-f!}L4LUaQJ(d-#Jlju_XXGJ`z=r`o!8A`z%d8i9(>CS@Ckuz%{sr-syZR=+
z7N670m;V81YnGTD8=tlCwLP(SIR6Xm%HdV-1D`*Byln6J!5Xo-|B6hT-1b;0y%rk^
z(P{H7G{8A?eoxvy`05tSvMZC`rd=LmM_==-^>4YeysK}q&llTC_WjuUoZ%Gy4D-%M
zJ%;o4F(WZA*`e?<eST5+Y4&{@sA}Hl8~Gn!fB!eMZ*#Lt2NwsRkKvFRP#=`O0U;6J
za|!t+^@Zt!+j{eqC!yRX;gM5f@%3-)+Kte5>*OFGO5Cb~m5`*%8ybD7iAIh{tX*%_
zmj@3Vka!qVqAH7PP5_Eta!|F`oFL?L21NCddFi-{DIx3POE$?|z2P}ep@e1Jjkar9
z`Y1%CvT}56lRMbfY)>Sw-I4sf)MD%Tz|AdEr!YVCtfXt>S=DyqE4{$q(jvbc`4$np
z%nc{-XGhQSCk&ONuTk760G;20xO|ugN9i~A=RC}k#!Ov4$awq!3xPKR^VF}%*!h!(
zo)7c#|H8a|u!-m2JdmCTbmBkbvLk;UjX5&UzT|KnJzlKz6^DkR{1mIj`__T#$4?6V
zFt3hbN?FNPi2|mBep!gC{02ER(bVNPz<QG=KgXL>&^3~Djz8blE+1*UcR3#l+By$5
zJBitD*}N`OPM1`^l@TCYODFLsaOIJpm<!kB)z;go5k{<@$D%J4sPVUd@(;|ksO@b<
z5P(|EvHF#hGn>cf<CvE{wq=gm#>GXr&fvQ7MqD){p|%g`p_4Lt*vX_8U$&N`@uzz%
z$qJ*`X{m!Uuq=uFs(h>x?T2PiP$)+7)Vqvv+v;F>E;JvZu%a&KS=K@C`&ie_rPvH2
zpkLPqGe^1I_OJ2gNX|SEM!n4;z@qzvSD6&Y`vW{W=fx^n^S|#*bbpowzmM@FqidGU
zX$AMj7_`qJ=Uftg65wZ@^N7m^%n|1*CrdtZCz4DB$7;oVT;}`%BV=0tq64}STNuPh
zhuZ$QYHU$fpSbQHIkxmgSR9VASPtg9by!NkcCkrok6B5U{h?|`r{D0BADJ^ZLNvim
z#dO!v)KFt@RT|cvb$3mZ<D;xHnQn{nkqX|-_uwmEjVbRe`Ed<MotIb5htnY<J2@O6
zx$<Jmm{_&H5r53(v#IQ*8a3{k3a?zBWP2Q*I73agQRzZUDV13HWi+SyOv36rM8A$>
z>T3}ve#P$OZa)pHWAg#*j1SQvUg>Au?#=`7>A-#c;OjMB<7WZuVc6%7oIBe-A37Yn
zSH`vLx5xFH_r{$!?vFR${@S?z=G*@Bhz}mTId0v$I}Z4>=i>4UV7}g8QoP(@=(^T2
z_u|-%Vc#73V&M6BxzbNX=J{-#m$whE5_=Z*Jw(}ui`4K%JP3Q4F`qXE)@j@mYub$K
ztcSwquIC)QlQ{dgAGX{d*y60dw#VQODz+KN&MuqW?rYan{cnuITo-cAV=K1f+>Qfw
z+d3A#SzG?@c3XX09=_wYeY_R>eEYUjUni>=R*?myllwf$ejIyqEZbM@Qh*?F=AYEp
zf{Dv)=dY*~lsDRHWBk2Ece>qIqipMz3szp|S^e$n{HGq1_0w8ouIVh$55I_hQS2LQ
zBX4!57r`9vyOE4#HSgwVK7G%})V_yM4fJJB@%h!CefW%Dwu)oz7%Tbo7O`cu#RdJo
z9-9__R-bLbNe{l9cd4Hido>lm%#(A&d7=4x?qBA2yg%RQ`sP6AaF3O1oycn~i8bGP
zje}RqWw)`g+H<dVyUph7njXh)t_kR6ntf}ZS=zT-e_F6B@9JCZ^ToDnKZ&=q?@Vxp
zJu&-W-ubA<aNa%|;lE^Og^Tw2D)=h<dKFORzVp@c;LA7nuxJI%r}n%UrR}vUw4asi
zo0}ZzD+enpd!V7?=cXB=jD5m)pT!_S=e}HjBFsFkeDdpG*<Y}}l|b&2lO-PK6JmE3
zuQ7LDc_1XM&uL3iArz&r<VM94=K=$nQZql8shhgWY37upJicL^8?l9@CW-Vb-8XQ{
zp0Ja0CyQgfDrIf5our_9At%5_AN*)CUW_AOu;BOo@{V&Y$xl4MMXfxP@}iUddeN2t
z{l%A=mJ_aie4=7e91lf(-5~uk-}sVp<R@>{DgBAZ_{rVh;GzZ|a`+(`ez7fA!Tz6M
zVvyVYIUR3CjI%%){)vpkCvtKZp9J=M@?3o)FtWHmGt!R*_K$Q)vP1>_+d|KelE80_
z&Vh}w(3K-qgXUj3Q}LQRi3R2M@xtNZDHE3(^l}Qs4y_o*W?S3m+(zaSn}?%B9pkY}
zPJNyQ2$s!3(pYkH+-AF%KAT%E^vtI$it%<-yu+=#$x8Che3UK6VrTZomM%saU8Q7x
zTnIcanhCHHUaN_Smhnoi(~ol$eEgG+-0a#(?dTwEw2oRlHqQMdj)*1R99iV-ZMv{3
zu}fd*4zONBTH^AH2Qg+JrKdIcwE5_~V3`lSIB%5^lIO#$?tDSTd%IA3YQ%02I{qr$
z^U5KDsrYcsv9t@^8B(~NyVs77(G6MrV0o!ib<m&oJO6ps*6oJSyrpqY=houP67nD=
zu-$!>)<kNN(5lDFz9q%|Vw2=Jt<2;lJz5mp`O}T@8^ytpT=6+q>v+vmE49y8eZ2HO
zopB4*G|vVHFPbyA!Tr{}`fOjW#8-}G-h-twx&2-%L{DZ4^%$};--kIhFNqwylXu0&
zM;fq4@85zo9ecG+Sz>DcFjwR%+elRvgdMTAG4tU45!oF;kzPDNllUavDc#Qavi;CI
zyn8OAoe>Sf%t202F$?9zNUR4{IiZ`(c;O;z=>%#WEuu~H`Ot=m==UqQp?Lhj?y7zn
z9H=<q^`AO?jo0{#gPa$2y!%+@q;N1UA0Cb?*RPMO*KUnFcOQ&*-}|QjAJ`lBAC3nP
z-Wk_!+}7uSL!Cp!<t7o=-wI#ftK1lfKD7uwbEw1qoDp!M{#k*mjd%i1yMekzznAp9
zoCg=-ZgJk&ZB6@bz@G)5|Ao%;N_C$xr{~Fi8mE01*%t6lW1f{~9@9w=!{!|4t?eYv
zKKt%=Mc(Bk=6D=9tCmwAh6(%Pw>I`5+cwnx{M}gN91&4yiFPM<$=*jNxjC%2`}}Tg
zzREIj*S=z(eYKrbe`EL9nZd&%RVqnKXckZE@3=Jxf7g3bj$I^fYq#dw_9Oh~i=TGR
zVeg`^_8E?q>sU?FeTS#tNCa~xTKYX+cawJQD{o4<!p7oE{_WV0<($dq1Ua{PSEP=M
zqR-928qamSZ#t`Pe!>fPj%k_k?X!dVK}~;FX+LXZdY)Tq$&_&opjzr7))H4O{;dyi
z9_`%0?zKe6%<%-Q{RjTbP0l2_{{@V|XCubt`>P&%J>aU?-F~$owVzkZWw)`g+K)f@
zn&)yJ&u*>>=w+IHYoFAy=4s$RmP=T_4Y!MZ+{W71F80Nonup~SKd}FF%BeUj|2`z<
z6*?7Orr(#vSJ~HT=sI@Kjl0<Q;46RX@E`oT;&NaCZZ@$S`)x-jW$fg&U#pgrZ+C~c
zjhh|Bu^@}x-(E;E2U^Y>flrIOx)S9_25OVT9Q+|6L@cv`!Aguw%=%rk_FHm3vf|+X
zaJYvhSLOg8I$^xc)@e0{5)%+1bJD0jg@uBoS7(Z%o_bO-s@+LbHdf4oxn;6sANJk;
zdMvv)u@N(uj@OEmarI1D;kv`6IgjK#_!6t$5c+}9&Q9mxnmety{`Q{+B$glYXq7jB
z`gRsSvhn*X-br<0%x~{7pV=-Q-`^lRxP0gnXn(_tZ;8q6nz?Lo8Cs_{`pp9_#@9CU
z9iT<!Fe)CDp=LW?Qca3Lhg*4)JCIqxLXGY*YQbF<Ynf}>BBd0s>PJ4kS@C~`T(ieZ
zfaX-e$p_?@?wj=;R|`mfjfVh!Gf46Ag}*SZN|(QLGX_Q(#>BN9ZPgQ(K<qOkQgSsI
zio3HG#+I2I<=C(yYq2Un{UK?)^}~8%9IwRo+gQ19O&@IRN<h#FOa!IOHfr%aBWtM`
zj~8m$a?V3ejZFpY*L{-i3vyUU%Y4iM4cR;&-g7~6WROL-dz~AsPDP)aeC~0wAy!#|
z4b>W@%(fF`M9Xcf&b`*kM|^e0y3Zeeuqmmh**4aWM^n`jktmiQ?<+C*81}UkzFIhr
zne*P2h~dGt<rpZbOSEpV?PTP`Ffsfcb8>szIWLvccM>=UKOOJb=-7ADOG{XZki3o^
zn^rpydnWs05Xh12TH{i`+5R2t8M2n$$E4`R%G~VM*vGOt<@5iia}xOEv1Hw8VO7mU
zA|?$sXRu_gsPVI^YGlB_=D}1p))v+=Ih2+8)EATAcLck=&KUA;lQaw^=MVX0Oqa=R
zUK4@#13Vx-X<hs>*LYGF>}3e{7hkd)b~x+~;^mB*SA3zz9x;%_ex)7Q`|b?87~{r1
zqmnhp?^A1X`#3oIeA=L%{-%8Pi;NLFEv5LZR5f`nEdeu&9RGrHk&SOjFP!SjIoeiS
zGT3h5lTF57{gq!`&0jt!<27F6H7<ypYjgO_@-Fxb5ZA}G>$k??)$8NV8xO~OU;D<m
zd-u(8mxo{v-x}9%+!_amSMqHzpA>WsnR7|ZrxVUuoSitY^}B+s{5IIn#9>U0^Qr6P
z)7V8Ge$CiF$NO;+?rFG4o}KNs2ic0J<5lWD@4TM3k5}PgJMTzN>xVHd-%=pwE7Y{b
zNu2$5Y_VI~!-?&<y>VJiT)&-$Z7lk}BW%vIeE{Zr4_yB#)|VsOa}AK%eXVDiUErPL
z_SNRkhmv#OW5)h%xNAdR_2c~g=iAjb#qR4*>c05K>A82YYe$a0<x>YAE>1cthM%Q5
z*cMsK?o{q5hA%m{{Z?Lo2ApDM9)0vB?KYlRuQ&Tr^Oo_<Vf_0J{F_wnC*rYwQk`r|
z!Y%gao7>|cU=wr!aC~erTaahYR^FN89iNUYsphazONP-ruO)ld@x@#~$G~U#)v?5y
zzshdw9E!I3utf%K=b_hfjgf%J0(KsJF*f(p=Vl?bgxm4vlzr{D+LEpFn*2-u!fH>w
z^O(WD2VWfnlCN~OiSN%G=aeKl{cN$5XTlft;A<cMv>d=$&f?%dopLJ9%D+YUuh3!P
zqW!)KzRJE%L)WoAKhD@!>wmobkFWOe4zJzZs*j0s_@)F&O7-#3LG0j(aX3r~eq&iF
z(#?>K8%*TmM?CnBws*CZ8QnJosfQB?c|tee3<|Q1RRzN1H;K`RFRuq7U*6mTCdd7D
z8h?F(yx}E({J<&n;}{H86E#=*<dKdHQg;$8{Xo!(18XvMpc<*%NGpn3KbmoV(AA@+
zJTg$L&BiK9;w4U1ygL>bnGyNMtKYZ@;5P=;l2ja@z&R0eK`06*{mM3Y_SBuz%RzmG
zH{Z0%|JFKs{>&$T-(r-{J*vAN13DLf@Uz1z6Xo!uJlxO*`QKm4uY6ooF8!z|`feV<
z#Nj7rS7%MOT#&lA%s1nbjT{)L1&-&IoX9n1a%7(-MvbFvMG54{w@ISn!Jjd+HIu0u
z-eeptLhA9PcA(|#M$RHr`tuD@IcS_)i+$p;oxFc*2|FNN?~|)ls#%e$Q}b*kubSn(
zOR}c`np?4`t@;D~AfT~O^^^IN+_Cl|hvPuK$ceCyK)Ow4*`9UTt)Vmc(jA6hLfMkd
zj?!C;l&q`QbKC}R)Nnwhw7+dGTPyx?>9XIjKgJuP`0h(`=kN{kIMhweidSl+FLS%>
zdsak_eaRDJbcxG1y})8@cKE4<KCr~sc>rs)tf$=iL8kP6$Yd13=NP^EPY(KhfH<}3
zt9xIvnrEq!RdsP_C-fPyCcqyaNhCG>OVEkaxjmk6QPZEy@Muwx2@!|02OKWyn6E25
zsFxuR>Q}x6*%2G$%(t&dmhIYH4+B|($ATVd+6>9pv*B8tq&*1nS|o{rGbYcDkdDxn
zNODk*^)TD#wUmBOtd!=gg-NdQtXY8ksJ>nh%C%v5!M0|T=D}H6x=EekF8>>`^b(_B
z3stKaGQ?tL00{c&nUm~}x?|A(9(%<_fgp4%*p3j=q8`I+tv<S7##R5HkF!CllW(e-
z)^x9Y0Z^>0W$mbh0FO;jgXG*zDmXH)iYtN*u|(OxwIwPMk`ZbuQ<1>fRCA{?H(_FA
zo{xc;nvdj#tX8^3+hUsM<MlVbUgI_X;$XcdJn2xvUyH8suU@-8u3o)0?mc*GJbe4T
zaqo>c$NdNIjQbDY9ye~?7Uj@?6@vF^|6%|sl)Ljglg|}(@N$#!XW}q+rG*;rS(wHJ
zX1V8ei<9U1WV}qwX*h4b_1Y|BA*x43j`L)>aP8#1V9a^>7s-1P7yD^h_Qf%O$L!>3
zoBRuBPyCF1`N3KCY{KcY$pkp3^}2m%Fyo|t4Y~IDmzrt;{h49L_957E-l%%G{F$TX
zzPK!!_tp7$MDG!8>ucTj(l$PD{6QGl?z70zxB0<78#s@hEp;kyziV&IaT|JUTOay5
zgS$w-70blalzrI-lCq6gOO^?H`bcW+x4zB|pS*J<F-z9H+b20@tm8|&#HYj38u9u{
zSiWtIAQfr7&nfsCnYQ{{Ap5c2rhS#1SxyJ(cfw}K?wrk0>v2fwkh#y6o73py56lm1
z0NOSORokZ5?g(<e>wnP#I(BXs`SRLt-{X#trS2ybe1Bk$RomNSjZxoI&d%TWD>9eK
z17;Z@J4^9Vw*Ix<%>&l^^ucKlzP7E46Mvc8!YWe}Id<dNmh+frT=c<L?8klk_$TFP
zIMZ1i{HIe+#aa1hg!>AeO}I$EuZpj-uhY<V_&j<7_SKpPU;p4Ye60FCWacO!ha4wh
zS}q36v<)ApQKa6N<B1e+Y#_PoK+}>)NhY<nMquHaSg}nR?N;k2VM4$8k*eJE<(nw{
zhM(H}$k~9arY%cVN`j4CiI>tz0iyfU2QBATq~9dNUt}$L$fO*ON5VmD+jx0p8@}bn
zj~2(4k3O;ThPK?;b6~^Xhy}F%FyYzl+@`HutUNC28}hzEZ<$!emy;(Y`%el=_BW&S
zroaBSL&uXpwYYvRA}Z(8r_U7kOg0jR(O_Kgm?v5^UwjVUtm;oC>Ls`9aAbw|!zuUQ
zPTFet#+m!l%l{9IdR3#pB*ud!k@!=J>}e*+XWp_c0AtDAV{v-BsMGF1vZJ!$Y~OJ+
zH*8N^;vsKIUYpB6O<5~2zw-*_RTMQ-N4A6wu}|^oRk~A1_E_e8RRJ+FS0zSgp%1qI
z>RI8rhY5!hf|FGq$xCl-3)?<&CSK|iw+Q9Sd8X$JxMp`a>2AYGyRyBeagJ1BFDoXm
zJIAX+J4?L^WpyGDe&&>>Vx3F68eao6AFBJWoLa5(!76c!vxvT!E|+uhsyu+4yYtam
z!M5f?=PntqWq=)9=Toz8yPiFu+>G6kENw3kk8Pz(rk*u}qifzUcAaKbZhw@JJ!8(k
z;QhwTAp4yx%)j#S{y|?MVe?vXH6Qu`9wK6li;!~y+fPgpIM<k_RX>z*oR-c-EG+XX
zPs5VyS?i8XCL20(&ElHno5zY2Ib6WdTgTUG_CvJFC!Lztp1^L+X<5lua|Nn&B>lu5
zseDLF(#~J;<zKBZY_#MA#+uvx+tz*AwAdr#k>3Q_)uu-E6~THWGg&QMzIEiTy&)Sm
zxaFsMD{HEoZLTLzacwgsn?H>!fVZ%WHM7H1&+G-AH@qU@I0eLjR12RFXQ}+&52)Ro
zMBv()ieoTq;j_-Mo4l3T<6h4Tg6BiLJ-OH6CaZb+G18q^@dn7h`uf4wYrMw4<j|Sm
z;M%x+<@&gG<JNfi<~!s5gSW@M2XBry-+E`f@y5e({l@KaRsKtt56$;^NZ0U=uA4hP
zcz<Rdd^I?IUEma^(sQoN7}oRn={cI$@U_Iob`e}8_GS8U#56W??8`F07n~Ott!Lq%
z5cfsQ#d&k|=XtO%f;{uSaNfK-yc1$h%XP9Z#~|3{P-ArLTMua4`Ve6q*Xj-S^>165
zJJP-G;pQPp!E)-`G0PMcd#+EgLGl?!&d%N20KHQzeI)h51(*`Ci*2o_TN~#zYT@3@
z(wUKt!k@FZE%J<GxrLK#>&SL->|v0)-rCmlPnK;T==<ga%yA86b02HWrQ!b^v4f3a
zr_bFQkhp&B2aii$#xm8`?=f5K8_x>1*KQu*KC1=ucxE2(eI#m*T39vnH~GP)EMYbz
z=PuR^?9#;d>7~aspKo>`bGCD-72E73u6fLLo2HzRxzE^rYHwfM8(9aU8D;bwTjqX2
zt^<ze=JS1gV16(K;A99KW%5-xJnj!TN85^<yr~%{-0lm^52O0e5w@S}l`VJmspU26
zOAC+MSDX7m%>YzKe4ww3ItSS11rFu@hqW!g`&(-Sv%JL@@!)G8_q0A6&U6OP{<Bk^
zi1X9|!n=4!=gt2`;-}H)d9kmy`Fy+2?+0Ih?;reUqSc|T#&6~#6M*jXkp#@FVfto{
zi6~gxYeP7!bztNKPY-<}0Lh&bxHQ?hN#Z)H-{hsy8X>N5;wfHiq0r`l=2j^urn#!z
zh!I2~)YoM)q-P6qP^UD>Rfg&$&JndvOdfz$-{gkuG+5LLdM3~@Ng9q@n3gTBE^?Av
zKi+V%XBv3ZEHZ2H=(jPaC&-sX?x%rM8Oqd=q~L#Vd7In6K1AyIB-UYCEf2m94s;^R
ziTnA{kz531KKfx5E2f!;Nz4T|8vVe`c^OUY{CIom#K=Y5dNC>nbJ+6>M?TLB9m(W+
z>cY=|T2YR;xmnG(rW`h&@&CnY{5<R;9dSp_uXxwBjvtv9iBpbhp$<AilhJn9MS~~f
z2mbb4q`XR9Zpb1t{)8t|FV~1K$?ZW*?)1@iy6~}6^PUF<xWutfY~D;eg3ykWW$B~F
zkSvir=Sa@2KG3;&kkndXqfLF4Jn41?H{MsPA5VsTPT<8OV>>KX6BcBwk*m=@!}ug<
z<(T5o!b(16vQxSn=F2KBC+jhqfT{G1Z=nU&W9b;z0IC4g{O>tQylj`vG4S;2I%KmI
zhhlsGcK`g4pBNXT<AhvXbg-YAh0DyuP<t1|n5zboxDK1GI40*1SFj-V3Zq-k2hf_@
z)GU5t;Z#m?n@!)e(tT`)yjveYR2&bJ_=Y6cM%mY(7q{r-_7<fd5g9kRV>83wB=fhv
ztRXNKOP1YO5nMAZO4pdase>7cPvHE3iZ9s&VzNz^RMOoIi)qbO9V8Z!YTBJy3J2Mc
znmHmVEe0HwyFg@9dOkY09o8gSEFTG``B?W4XDhUn9Hu1clDqfmMR!aQ%Z{bTv|F3A
z9-EKYlL8@gL!0gD+Tp56_xjQK+f(F#@+Ziv&h7V0v`~D?_cx8K#D>uOp5;VH6u~w`
z2};s=I@XPRwDz}dBK{>o_#U?nT1uj^d+FI;-Yp5UAE?d#p>u#qU2}L5<Hy(6F#I^-
z1hcS7;CxCVkC9k4FY20T?Ho{DYV;`TyiGXj4C=uaX0B^IdDLYxgl<@|RFcWP#p`c;
zy~b<&MModPryX9+x$Hl6$p8GhcJt17@Foww-WvDbcsSmE=NsdV2XBqTYd6NVtJnQ)
zFwTedvBU5$EB;I!#uqt#!7aW{sJwSH`kHRGT?8*1H=}cOEse8sZ~YgI2WR$C?H1?B
z-B`F^Ks_@~;+{FC$v-oD_O{V8pE*9Y-+ex|%iQN44z}Ix2WNL*4d|b?t{L{AKmYF=
zPk%;Sahx^x=}Qf2E4{IA3^Ur#dRZ!u=TNN-$*Yr0*kE_eTsEzmWKKv^$<7yT2iy8i
zu9yBwEWq8z-?g=z_PqU_WcTOY%}AJ)Ex-4_JUEhXnv0p+cb*!vFZ;*3bc*hu)a(71
z4d{Ey1!g^A`{oZDjaDq;<&$Tyd_!;PR)TmtdHh*G^=hX`H++U)0T+EqF)NFl9J_J5
z$HWqEM9aON88a(6vs@h1y*+mW+jw<a>1Gu-P4<U1T$~+wQFhRF*50SAgwK4$pIr31
z@Qc%O39-*MfE*)NAMRsc{MzQe=4|hWX~6exg!r>;bF|G18g^b~Gh?T1`z)uge%R&t
zbrQ7WdB8rK=Tte`vdhm{S@z|banb{?#=e-?PO|S}YwLMF4;*`b`VRKhavq#CwzKfE
z{k;l)8hxG@FH>6{eEt33c=4Q_vQ9iXLGeJRH;WJT<#fmmkmY8MjqLPpDz>>xgcf8S
zBqk*5%G%XIM<DsJiVW6W6#ciYnQKHNH+FNv@iEr}l07%u@W7$PtZ_h}(q&GI03>?Y
z$Hq|{MBs@md6`J6<;^N55y}g!xk>I=LAoh_@p5B|H&dw&vg+22R(2IEd)pF|0WLSO
zlG0+{c(}oZiHA)N#9<$Kcwm)pR@toiIv58E;H~@U*^%?Dr^qD6=Lb_=GjRvpHpVYq
z@mvT-%nZS2NiACHf`U4L?GY{WP>mBE7amWaJ+;%nk--L~CUL3b84u^=a~|U9Pj8U@
zAaAL^_Jba`r;DY8otw4hzzQ+~`4>2YTZgw6lNZKKeK82qd7E{og7cT0XsK%WvvHmP
zbY#7JL!P;&CB!Pcq^Qy(7$}X{mJ6h!bS|q5LQ+S;?fFQr6$CsESC2BSLDh&?gGo6v
zx4oT1TDrCscs`eLpvwS_$zcwBAr8B8+0HBRnFou)A_%i1v7Yc{?6P@mYm7ZmENiXg
zeQ|DH)!AbS53|<f#x9O;;#>Cjm0Y9U+DE(pob1JNejvW&Wa-Rix&_Vjsa`EUhnp_J
z^97@UpOZl6t6XZ9hdz!;9h7FPYa}AQdo0C|tvl(Q1cqs!=UH;)CSN?2u_(T%b2Mv?
z9*!9JNy)M5vy5|Ns|RFK56B&>xv5!PHQBBf?4!x?Kuz`tRZE<D#fQm$_z4M5I(2j|
zZ|5YEw}?U{t?gG{%u=&2#GXL;t6wy*URPoZr^1(Y^jf<~eDi7=*4(9cFH5YotlQH%
z*x9EYO}^&JUvpsD*6v}|6CJMSLqCqNan0=_c~P@|mYOAay0Ag8*VJkg*V;7LsMfTW
z-t%FXSJ-jHKj*0`FbLQHy5y>H`K1sh#YhC~R4b|LSv-4k+bq52qh3(~y4~g1vZiek
zBI7G1PBmKC2c-EpLPEjrv9n07Ye$@GZY@h5{ODc<%Hgr#){o!&a1*k@Iv-_Nh_C<D
z;cL9cUwrf-e9^;8XQjid<KXhuasB3P{~5%4{4cOK-yLti``)-O{if{vzpzV}ujq7j
zDZdoAxj}e3_MKjS8m9s3m}*f8bp}maC;VIk*16h-H0Q;7J^95T*Dz~suYy;J-xqsY
zw$u7LwT_pKE1bkRE4KZt`-Yz_lYIt^**51t)>r9#p5F1D7PA9<+9P*3RM1uI#-5bx
zM0>RX{gh~wBI9~}6P#Z2t{-XrMclxj1!h~zYIbw0CJmJ%)AA;d5-;0qi=7&_aTE6W
zV%xQ2TFwnd%d*xzmKi7K-H(g$`FzZ|c=Ai^`f_c!i#gXSueKmM?#%ff$hkySo;f$0
zBaGDsmN}aGo_jbxR<h{lm_1`-jPdtak8@p1*;vCPM#b3H*o~@uPRu(({GA8g{#plS
z)DrBgyV2+M#>`!O-QtpHt!(o?zde@b+}nN{$2eMSp+#3dciZ-o!Ol0luyu@ZtS(yk
zwr6SX+}dMZ*Cn$bU+ZmK+AOOrb)g@N(d$^<OWXWeUdOWY@JqJt2khjDxY%j}kK1?p
z!+t#97GKnZuakMZIO&|a4~{=SeFyvcJ`4NCbzZzof3J$4MxW=!%hcuvU;pVp_>DT{
zXlM40&WeL~IbUrcEAQkXo7X_>n0}+Trp7li`@oi=Q^4y>bt3HT15{ccQt%hH6_Uo$
zH%Z+>gXa#2R|a>~y+~DVE@al8V@pC18#+1Rfl3)Kv*t-c!=biKj!4uJgRta<SA7#B
zJvRa;-5j+Q=lP%(_lw;O{g4HYRlfPIRcu)O_GoV4Z17}iYiy~QKiLGG{Jha~Q_2sa
z#K}Kultl?r58HxkH`(fam>TJzZ>G_gbEOwQoJtPJ@Gt4Z>w!;HJeQ--r_YXZQ3MH2
zGCL3U)I<NFkli=6_4D-UQ`zF<!Pg=G4@`NVA0_sQa_50a@{u4HV9_1NMZO<~k((Jv
z4YZ*9Ty2v?K1*0L5Bd6u`b3OXx@=BFtz<)GeQ9}<uu97hA(lKGg%NMJ=h!!TCG$-|
zN6;BM8U)DQsLdRa4<nGnl+^6jDI}{N%N1M1+U6DS7_-fxdxx)X`{w<kTx)Kv+Mb$)
zBD+?%7*n_T#L1T4i{1j;PBq4lPE7d?$g-F_`KnJe_tK6`i1HBwjk#7IQSoG{BC_8r
zM~b%3H5VIrT<jmkOE>H#b-P(Mw>}o(Ckrjt;mMS}G!H_yHOIf!oEBa(Np_tDo8x2h
zT&qcO`ho&icTje>EkmJh*CRl!#=3!&anOhNTKMXls`@-aEw(sNGI_Bq7hB<x2Y3!^
zr^u0DBWnYlhe%j_SoSzN-dYsa=_OI5L0W3Vs&hrn(gcWRYH{@vqqB|;t#!~_ZUSjw
z)cE3+#q-kHVsoX46TjiiRebl-!}fwrenf9G)!MnUPMoD`u*FJWG7>jC5$kx-PLrM0
zxpN+bLu?R3hxMGN(y}FdNxAMHYIsk<MvTVcs$CDp$R57L%uD!QmCeUi+q{zIc_>JB
zJ@%s)ODmUAA=Mb!1<(*;w~YaMKClWE&|_=8BYPGyA5DTUdyz69l0DF{d)>zY3+P${
z&zjm8Id)tgsclY=&JWa77ZnjK@@!n79<Uc$Ov!AUvM?UF(lcghm9JAq4evkr4}P`H
zUj@4C*LaQB_^Dw&6EyVR%x8#$E91tkd*j|4545~B-hSun<L!69G49@bIIi8i?f)5e
z=|G<obZzfnc01#Uo^R+ZI3p<M(i!{ey(nJpG?&j4fcR~Wsh(3x>m0s*D6Dy39w+6P
z`vE8MUbO~CSB}#%U8JAA+?TBXWyXzfJAc3tQ=G5$CVkGh&ENW&+hwZ@D~C2RrsSGy
z*tRQ<^|OJ!`}$@ndwLV3pP6rupA_=lQvd)TYDq*vRC`6U70$7FuDG3eiH>Bxt*<`Y
z(e*med1mfmUkxYKeYTx8^4`3)-tkF1DKFTY=bW|`^x1bx(BLmu;lAFVvCTEd?jpSK
z*KM?YaJfMCdF}&rwsADz@cNuV?sm<4!ETK$Z|3$G<X7I#v*l7lY~dhr{hNDM$FbHW
zW-hP|QtTV)imqW|5VYOT_n~XvaD9wu!P?_8qihi);h&GGljIqTU2V&>&7O1a_-EU!
zw{2>fmE8wlyBsgkB<Ffz$3H!++CEg<#_Y%umbpzXI9b}J^`awZiEsPhE139dS2-~j
ztKHouaMMPf?JcjdTU_+PSM=jC`||Izo$J9@Q8a>mU7ZUjjrSy+G_LdFqWygl_-XWc
zUhJ!_&oyA5KM%fs{WrwoCacctTJkNDDWG#mwF{rrHML-B96Wf*{TNQN>{4NyFGp^2
zK#+|C%YK%ux#@wF8<On6)st_)NRAl>#we>%rf_nTJRHA?VsjHA=@*N1T0RUBhYzC`
z4p_By=~8RyB4vI#Ie2ZmxgAeV>dw=s1luVAo;j?7%!Wx`o73}E$Zb~aa+?H9R9~6%
z2C=+9GN+8o=N{^DhVo}35>H>W^rAo>{wx;^)#8Ptv#Gny2S^F-yBMT&Uzqbhw9vUp
zOg!HX<4+JC$;Szi2WtJ>4D^T3DQ(HDkdwZSj)<iMVh@~`KeO1|<Jq&P=Am=)>)%x4
zTVb?*@J3qK=K{H&QQPxMZ}bIEHgwapa*VwRjlt^%IatwVrpcE+od>_eBvn4&=)+gG
zRItW5hE{(2I?*IH*0<TrhY1C;E_R~aFa6`GO5-#)c92?<3)x4CWG$h4`!E(Cb9(t=
zPObNRLa~_(r?x@G25lq09BLfrgl&kM+In^?anLmuKj#x$GB(i-*E;p-+*;g!R62Kv
zV{AYOwQVV|=|=bEvdAfWpU(_J@F~jmt6j$kDu(`<KsGA-CpF_<`Xp-Y$F+tLp9{hs
z8_4xXb4VpW=LlP~2Clzt)RE*Khx^Fk?-f2jkhh8qQHtxP&(1kl7mBG4MbN8jMoCIZ
z5~u;CZ2K;xQdx4tUmA~pNtT6_G^_;6*d*8_ArF(QU%{s+1$dwGSoFjESxPR)0td;b
zJedZZ#F1Kj<z60w8yyLORvrqXAL8nw7-Wpt<tCieZjz%j(w}X`@pjTr`-#Wd`$tDF
z<3a737fV{78*KL~9_A*}T3cf}zaped^;CT?ao9Q`cE*svrG5-Z&6$L0J|d8vLkKms
zRvbC3!ss|>MMz&(9H=H4xgIfBz})9IhG7Og1%~ZtL3;ORU3NRl;5ld+dqnRaJ3-{`
z8be*>a8pOyYUKh^I&VqjDm&`ZRnyzqoC0%Z6RSD5f_X0`hh4qK;EB=;dUY}nZGnS}
z2&!4^i($Ne@bwz6@s|#L$nQHG@oBkqI1a8{8#iv<9rx}&@Si|@``xdPH{W`9+<D{v
zxOwZ&xO(k|xCj1R;Q#6Cb2dK#**TxB?+8Jk({`bASxJ7ED~0&x`n`MyvkqsQa|`&}
z9F2Xk&F$P<&YCN=oz;Kg*f=^q==amc@p11*^IkRnW#Gsu{-jJV)5j8XM(jzs&+2pb
zd*-J3J8!+sYxYiD<zszXzP-N1wq7$GhckOQo?~B-j=(AJNoWphdN`^6*nIUw1u+(u
zKV2u}ngt;->vdx|`|`Ku>m_blL`y%#C5F#Oo6y3_*<zn<#;!m2EB4{o10)TE*c0>1
ze3I{Cc5JKN+QeI~OM&dO8a+SJGygNTeIC=hwSDqsL$cb<>H9%pI!4aa=m3PBKJuB*
z>|=cgi8H6%>AI9T#)ZIJZ6`g*Gu18rT@JbTM}ZszyD=xTcTjn^=Wejr#Nok?wGGY3
zM~!EjV_#0N*wxM)<T+yN>(+9$&G2!kb~lE5jItNE(K)ch&M(Hszwjnb)RtPtpU?H4
zn~6VjI|27R!tYp{i`z@-yo%5*W8)!v%l#m2(I-9<n`E_RkB%N2ZTn|9eq-Lpo3Zup
z_7j}c*A|XEqd!Pza?XK^<_Q1%F<(TD=fPLi&lioCsXKq_@Ynxe?DO97fe(G<>f3d(
z&HeyA_TI_sq+wg}>Mn<4EvqT6fli~<Vumb(Z}znbU5-9AJ4bOlHU}vJhRH!x$qUEU
zZtGyi2T8H?1FCYR>e)zMyrhgQ>;J#K^V@UdHnK3Kv1B>6EdP<>*v@WVY#+hd>)hsP
z_O};#z|D@NJ>NN10J_;6&WvKOFXjvh6sk@Y3g~XI*-dg<uAJSOpYzgY`?%KAkl?=E
z|I({TVcFM)j3e_9ZpQ_!3Ab%xSDfZsUxnwja$Z^F#5v#i4Q9$z1~-$Q>ji<EP(E=|
z9M6sX2xpmC&$v-VKRWS}EerRv(=9)X$c-qm$d8{syx|5FT<6am5LXXhNsFIvmid2)
zS*7O<%*1nBt@pf<Ee`Q{s*iDl4o6S}aw(a({LeC83>?pPqSrP|Am@iRW;M_eL$do*
zn;WhCuOjr8vBx@UkGt9~Rpf#?b;g@c;BmW?Jm86+jvT673^Ir*`vIOQ(?|OdfWGRb
zEF}avxhmeee>ggJZOe|AT-vS}b@S5hDw3x|g-5)ogv5r6CbdA~RC;CxLHlJa0k+6U
zJ)e|)Pb7>)X<K_BC(#+P^=`{F@-ITsQHj%~Be^t9h1=!tdBD~|XvvnnKR!s+tOizm
zkmHx>qiB`-_L$3sy^QTK2&yaD!I8kJC%HKw3J<F6F~HDY*}w=8cr3##u4+EBh;r^9
zm4EqCC7+%T6hg<XeBv7vT{V5Ek%xU)OE^wU0Jhm0R2z*M6xLCfPor#ARliyD?IQ!-
zj~jR}rRaIsO=Tg<8m(nN9Lxk7@(Hg(#|sK=)m6b59DuGZz{fH>{t|CmLCs;oQqflq
ztuC_<NLE6f4;I_hfmJ4pDT2*3tLWI)^BPdCQDSi7@tkX3Vh>2QnU9uDhu>q1GTXyN
zY!$CHL0Xi+bb&BiPDePv5^#!?YVj!B@OALA`7bDsOa%&Hg)B8UGm6?g7`%k`%{k6G
z*7g%5W}8JeUvm5)#J6pwk*q%{kQ+pY)BfX5R~!b+L8v@eo6#!wYo>rEBS5lgos_i?
z#r^9q|Jjbe1-j=aJmCqyYp^WTQ|rXb7q3o#`0~x^)gNA;-hB7&^rxTx{q**SpZra*
zFTZ+odinCD&t3Hvz(;vrQv}X2s8#f9Qx=^4Qo!1pwKD77b&NiQoE!G@2*jT6#m=$!
zh1c9mv5z?Sfbw2*zb~%&oX<yqzT5j4M=|%lkHICsyP{p|>!SU!euX1Db8wg6UAg4F
z>fN}O(|~vz=GC0ocFpPiZ1ut3Cj`^>ps~)~!|Axk=A;o>X#FOF8k<p-IJm0UeGs^{
z#A+`3++33nAL4B+u_jt}k~`Zurk#D^@>c!$kwp(ya%OBE-x|J#L(Ic|OLhO8K!VO|
z0K#gQndVua==q?6#C31wTye!dm$SBu)=@{cBz+BeiQ9Q-KFJ|U*~wy%%x~Y>C-vks
z_Bm-XP@e1Mi)sA7`b%&(=lx8W5K-9o(>FI7XZD@{L*Ngg*Kx6~PwTO=KW;9;yd{?F
zH0E|&V)e}>Uo%j*Ho+}^*JVg$gu~u2uF0VtHH^ODwX^Hl?Y?dRHXQpTH|k+?TV<M4
zuR1c^?Hi*|Vz<b9M*&9rWAB@LuBD#WG{SSOXY|u=#t5%%=X&a`{s?#3*ZAO}vDW#%
z2vM%ZJ;K9%CO@wC_kT=&55cb;|F^(n#!%n*`iG8b?e6UHIuf|hz{Pq3Y6;H4felE`
z1xPNkil|kD1tH(cA^{FMyc{Yo8mK?yAful;EJg4mb|emD#iQwqIM(qX9+zv8=@+3F
zi)xNQ7F62NFG{0LX$R>%w#wB!(>xZ>McFXgj?J+`&KTel)y83S?QuMJJj<a=E{<QX
z)wdR>j5Y1SksQ~=vw!(|4v1|FK6={6qZ!|B14C4SC%NQCPQC@T48VW3SI(HD!wyyP
zk+togeXLJGNRdyTIC^e;;KIq2_7A=_#-}M!CN8XeYm0dN#Gx;Tvgl)xpLuG1cppDQ
zI$pH+W>@0D$ap`ZgX?&TbEAH_`St$Rxxw>SeHH3G@hhZldOCY-;`0WWZ=U_nKmX~I
zjNHJ=m<Yi)*BCQDq3AS~%LfVQXo75}<d!*e4>3fmDshq>W6?auoVemrJG=5T5UH$&
zfJ3!o(U#R*Xzpp-vyASkTmUv4t7*G)jX_WRtTl)6TkCC=@%r@BC+?=IZ^0xzx+CD@
zRc+ZaVhE`jx4F<guZtfXx8*f*3$;PK{&u3F?dJ@O&+(FMRVrR?)FlgaUeC0l`#5xL
zTsseM!xCL)yE$H!^yLwX>+NTt&wiOu6w<k_SQQ&>X)NW9O6Dvt;@v0W<Z>)RO~k<L
zWzN>)&Z5a?RT*SPm=9Izdag_5qo(t$oa4XvxLsP)$k}^@{m6rF6DO>S3QDfT@!C8<
zOtj@^9b9|Hm$lOt<_cMAaWmHi$7V3qc6N?VQSs=DIAq)T<@kojz62+Ij6jZ=7A1_%
zBV2qP!?E*5ZN6^VTJr%3ZTz@l?rhYYuU9BKCus6kqVWK5$P|TS_6aqTjXf~viC@_^
zv19L?Aw({JwS7Fm<5}w|1>*=xswC5IXP{DjwcK3RPQoU3^(9@-H{w4$Y>Kau0lMYm
zrP_p|$o_*p{y`Ja$Lg-JI5sGS?AbS2KWhKLWSTwvw7(7L3pQ80bR#6r7!hlh2PIu^
zhZW=x&HnMRuc?)r2@XE&3DkYrpoiwiar97KQy`dOKUDS>f!KKSVT<Ts1rFnVEy`YJ
z+t=`<l&EJP8MeZ;%^rHj>JfS~N)};k4k0<CkdLd#Pv7`@!V{iw9jrb4V|tw%zTkg?
zed(V><fjmS<R;kLA5O2|e0zHR`Wt@}>{jm|IIaEQg+~XzWLCUh^Ujw7tzNYfJ${iO
zz@_Dkb^luAoLeM=OL?xKuePu96=<NBnFp4gYCn#C-QDZ+8lX9^*MlQmpOcA)a4$kJ
zI?7$q9@^L7U-7#Zx9#&><{WS?#j*B|KJM!MT5jMl{;oaaFb2*I(vEt+KTx;4Vsi36
z?$SP1g3N8(aB$n}rSnSAZB;&ZF?s9Xax4M5vCreo5Wzy4eo|I^m#XHr&32?&bYu8-
zA8NM_tok4cbbI*IcXhS5wH}}wdSNpU4l&y+5A?R2`fxL*)`wGRN6&otJ@Q1}GadH<
z_~7#j$1K^<eiDBg>DLd$iW|Q2QD;%3SKKv_3_kspZ8=N!)9?GDu>A-_UwNn0>yh%8
zw_BG$zkN;X7(|;l&-OXYeJeP#ySzJxOCWU52K}V3t|yM>1rB<1i{F0jP1|zm8`k@}
z()6Q?=QnDrj(X<NH=;n>{DIQysd=z0^Nle3TRrSt3q@iyeIsV}(Y2%ax!SDZ`bO!1
z^=?w*x8^y00sV14(la^j=RRCl1?IM{WcR5*<}3U>i&nZ|l{@AotlT};92<A>!QTG+
zx$%`bd@Mf2_Zoi9*uNznGY;SQ`mg`~hxY3%>Xec7r~S_t8)~uc@RQkG!8>aoX0iGr
zB38c&z>OsbrMw-K*zV12K)in14}Z6dCEa>l%|6P<Wyg+RvQ$}<<Q93LBYf0_c%X})
zL^W$FssPx^uBWwdEEA*Ta&CQFdt7#ux?|4^c2VX6r)uYpw)|)U^I=3)R-q(brMaSn
zsHf9$YXUzQj^e$cVBsLPCqw;KZn;peHgjsy?2i{LbX?Ho-K2j$5iWX#;bxWh3H><-
z(fo}qac^$O9i_Ipc<6mY>wV_KH^ao^MjDRX^x~!zecVb%UY<ANi$~AyRDE$1%O?|j
z%S?apZf@!)7Zr<}Y3nz~DiJcbNQ<9p^o=<BRxZyyFX~4%BLeZD2h*%cjUP3UD4m6*
z<YGS)rB7H>c|+bun>lt#FM)$rD_L~uSPrI#RK9px^Wh8ds;xIcO(U_%$vMgn6|H|3
z)r^YoygLR3e(^CJBmJZui80lPOP9W^b%$n$TrzziJGP4ap-PNIu+h&MA{jKWcDeLg
z8p#sI+<1J^0as%V;%%y==u*ZTh?B`N;)>2K%RXLc+E0dxmxus+BKAol<+=i=i*OP=
zu{N<QV6QigwgVqK78h;pKeVXOjeW>6ejICwo6<_vVL88w-ET}|t@%wOGEXZQuXxRu
zj=%IenASJ|bK^L0p{Q^hkK|4$-*9%@f>2U5(;YcW!3z&Lc%_o;M{@2Mi5+LVBFKN8
z=Zw$J;fuJLj|3DdUL4d)4bVrE9_Ptk+O)zEWb2zNuJok8WPYAA<qS9DhTF2>xCm3b
zMs3}wtHBt(4k5eTdSqXgMT%^1D}>6`$;F2Ou{G>LOkzu)xB>&FWopJ{d(Q_}71ot+
zwOr>}JWGwXx2<v;`qdt5<nFIq!Q`y|EfJYoy6J3mkJD$ZBe%xTxU=modOjeehXQSL
zT046g=q?6m^eBZiE8S`g>y+C<nN_3Z)`Qtv^}jrQ<Le1ec)|tD^M<tN%8^y%=EWDM
z&tH9Y`ttR+r?>BZJpKIVzntED`%cgAPj9~Y{&e%%OP`zG>Kxyfv1TBFb+p6<k6$DR
zb`bCrkn_XN=<%7IgIDMyto;SM>(dv}=dz8wR_j&A+b}rdJQjOj`gusam3z&L5x}_?
z-_||nz~`<w*!tfFtk~px!5=-AOMBraM%4iQx}CtKKK3wc<!<_5=czX5&%C$a#6{-#
zUUlS={U+-Dh5nu!<ZPMfE5>d=XgTMs7LCDC2%jk(`8a;stG(w6O|q5Db%aLS@$qBE
z+xG)?QB`|UY=D2AEe{U$h5Smlx2^3V$7SQgez>+*^jKE?+V84Amox4S@L{oDu+GIA
z18my_n|&%X|D_?Dw@r5_z4k*sINfu(BJ`Zca_*bR?2qB^(in&z#v=mPl+@QH=LOyG
zA#9k&yXS<Qh&B=XR=2@w+Jw^Ur)<n?gV_E+I!+7M`<H5es}D}KVY;@i8cUnhkV|Ww
zeM|43EW~1Y*4VS3qC58GIvLxr{hc;v1d@5iCw9<!p3K?g*t*oM&zi)0c$?Be*@MKh
zRQ$QdH!qwS!yoQ1Tl^+3^^|F7xJF+~a4fC6zJ)tzB?w2deWv;htIr-sH@(1JK45R3
z*Fp}kn6AY=!oz&#oIeJi;(HChX6)Y*j~U1KsY7;nyHjZKbub<537G>Tau$HTz+{_w
zS~yx&9Cl&9F({n_n3T4wDqZK}zK+pj3CjW6%~CFgxd&QOshx*m9c>>+;*<@2CW;(Y
z+BIVUt}w%iO3etqFey2FG(=1PUFPCt=af+9D#VZh?b5d{Ig(!~RrkeGX)5pcZl41m
z7N?9WoNR~jVy-%4z&?3!%vE3MHm93#no%JF8j2G?+UX;q!|ir2WPF0sv9YVoSoro9
zZzx~f<P#J4GTK|89P|l=_A8Y)TEw}@#m%VO+gqPlz^4;DY5Sx>4SvIxlLXp{6JO}J
zeB%o}-YT&Bn`Pw7O*Rbh{SUA;58OP9(imGqeK6N_$xkc7@r6QyHfIGcQafvMVlq(j
zsO@b{ho`1D$sg*bWnO@6_N+t>$IgX_nM<>XEu~YlS8=3Fl*F(_2KVAe`Jos5#-B$5
z2{3ZpqsfKjG&~>rBL~`H%RIhW#ZwGb&VsuSs6Y>h@s>T#dPadqa;|~JWQE(_7WPXY
zz#RN`(mB$ZOB09_EptHZ*+1qX*rBXAy7ckqKtuqu-`>x2oW$;Mh$6)+nG(WrQQ}4d
z?90|yZgDG==TNFGndU@d$EG?D$<o(20i4b$K8cgM$_nhX%!_^GT`?0#RdEo|V}gm9
z{^FMyAjg*k737*v74ej^i1^L5Jf4YLJnYowB<JiaK^XK6EB$55_|)(8fpw@R!X>^5
zGC1hV+wwMcNvt2RJV?(@jV~-p;5?Wg%Vt?0kHn)dBrkD{Igdc`+=o48T;Tqx%Lk6f
zF#*;GDfpa0{HPgS+A7Ieac4G?X^7J6gJWANe<|2)=bpalmIx{&u&9olRmx}{SmwhU
zM{svM<Dyt;cziXO@F`+rG`*gp4=Xv^(>j919--^nBTQOyl;F6~@Ftk&!wP$*N8V)B
zYd{-qS?19}+A*=aJ$s;#jyQWCH(H+N&NCQJk|gu6DtW|SRc((|wy0;yk$sw40HE>8
z7{73ex`4*AN1WN`d}Qtf&l|LlPd|0|geN@VBJ!H41J{elCBV5kJ^$?U)90_gJbm%S
z*QXzT{Q2~!pZ@Li?e~8?y?y(~)9bIlJ3W8#LhnCx&hOW!cxF}Vl`m`J<3Z40YdPCz
zrd%JLERJ{OxeYuoM&EC2cW<kQ;l8o@e5Ucx9QORZ_tAcM@7!Fdyud$-vyP>4@4jWb
z!#c2urzbt$<#+G#4jy$s(bEU#3P+rY{`Cf*pV~x`_gMj%Vnbi&DLNk;=NH0$Y{B-J
z$+>wogw-c~FP@UUMyR&z_bwPxC8KQ&;I^sZyX{Bw?B#m(U*l}|1=g>A&l!DwK5ymY
zANeQ_x>XO>@eZ5lJ?3t&*vHBF_%?<|rAOzy^4D*ba6X+Ihf{~N@1x;sH)CB2!fStI
z_96Tck)s8iM<?vi=qUmm8TfM~si}ES<iDox&~u&~CW(2S+e`i>UxGbJW9?_{)Je@4
zD-9goI6TMRt8k=V{P}_vTeyLg_!N`Kn7ywl@4j)<`<G=qnz_}V$;5(}n7q$O-F>uv
z)?llP%v%n&sC}xoB%yh$Uw_*xXnoeMl{f2()omVS2b0@lnEf^;Cvo;SzSf&cty_;F
zDaY2q4m&<@I>zK5$*Cx2zblS(6O1j=-Ms7ei#NVBoSK~n<657`=<W14e2VW8{?+k*
z4|vR&#!nskLOTBS;zKaLz2b{V(r4j@U=qHCtU~Z};dvhLv}qBsKmA|q#N(K9sAz*`
z1gX<EFN~F)`ire|wqD9C_${m90J8Q1fC^niA@KlZx#uDxPrXKjU$H}#q5F`+c9awy
zE&dez848g6!b;9~X(Akt<8tQHunWHYc^q(Ja2Lpe<x;flUYG|!jlnvV6~`(88I3QF
zSs~eh+;Bzy=q&fm3UuZULv#}PBn6A-?T1@)xZvCWT+u%}_&%TN;G0wac`KQ?x#dqj
z+}xaQ|M$HY@6SHtn^>~xh57w0I=;<?LuQN8IrxQJz9@B)z)dvb`~QbYg|ELICOi8k
zWAP0xaq+?Z85><zZa5LMPmnSzlHCPFMQ}5h_ZW6gDQd3F2WrDplWiUWUG!lk7cQz-
zw8a~`xXg*$6$A7JaZ|G(*?puHKIW)~#e#P7p@*Y!Caep`DYGLnKekwj6n^)apDimL
znu_PGJ@Q|onkQf@eHcVmjomi2EmzH-^zP4I(M3=jp8nW}y!~pc6w5K-Qe{WM5j|_3
z#q{I+%cnSvXaSa%fjf6?Uk(KlXJ7HoWs;Q0<mZNEF#3|grtS$l50t7Ms<)SkDr$Qc
zWmk8ZKV-)edTv|B><T?oN5cvXGiR^)HW86{YB}%ZT=cZmK{z;2WX-dw$L`omerh`o
z?WoW~b<=Bvm0~26axMmON_N|^sfx?79gE5eUP!Crs32XroA3Pt*?J6F%WS2OmL-gR
z2U@f^*=C?DonJSdbEzF^#zQ8DN<@`sMRN6lEF0OxQc+YnXJYsM5tDUz+n17jZI8yU
zZGv;Miu}}A<ma`2gtN78Vy~shVx?`r=sjPU^~Xpy^C6l1s6vVX8+`VgpkhOh!hB~l
z-A26l2gS02IhfWkdPEs(Y{|JIp!e~PtilJW>WbIEY$eg;8%y;@zeRQIv1SlQTy0?6
z?cR~fK}Zx3ooPb{l0OFGBwod~9s_M2j5u(d@@r+z6Yx`we#mZ^*Ygpx%8KoxM8Y`-
zxAkW`&FEUOiEZ-7(=UI0y7BddCp_VyI9fLP@YZ{co0q2-FJ7I#`ue-mPk;W)>D{}Z
zPhWrc{psBgKc8N|`9?3QdOyKyAL}zKm4C8be_E5gWD(}Uz;m0|KFe*~^nDqteV@@K
zzx!m)8qMI3K-^vI!R>RUV7_KwK&8L(`dqPbHO2uR<Fm)L7_0R@{oO04;+R|Wu$$?D
zzF7VkJY;^3xU0tUJ>#9nqTSDg$r<j#^+`<0LPQ;pnx7eudU7&rnK?M)eaFZA`-MjP
zyFSo^_*X-SY5K)Rk@FhsXqS&;=`lhiz^1>}xkTREXsgEV5>-W9a|NB#R=?d>u(f&H
zRg{Ub7=-x5zMsRRIR%z?@mD)z=<zgsodzc!6}SDK`I~*)z7lWhTC(0Yp_JP-4Q2eT
zSCf9*pX*S$DbUHK9(^je(U%MH{aPWk0iaSF7y3Dx*OL6ayzXw%USP%$EYJ3?yEANW
z4u13p(6`YSNX_ftfT-&Y2qd=X7Cnz!_}Wd$=Gb#?0mAl&_|`c?CWTd>cI3X%gRbSa
zQLl3Q@yxZR^hX{Tk;B(})wsFMv%hhRoH(<+t+NlcU6G{JW0xC1PP^+lZz~UY{!Cr{
zw>f}y4wR;!ZHJw4Foy0kXnC)r$v#7U#J`07K5(6Xw|(932Bf;?cL|U6qtnOxeoDXR
n_*cgI|Hh||<?ny{k7xe}r4NniG^T6e00000NkvXXu0mjfPzEu=

diff --git a/docs/_build/html/_images/joint_transforms.png b/docs/_build/html/_images/joint_transforms.png
deleted file mode 100644
index d39d15336b282345b653aee0ef67115a967c076e..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 479613
zcmZsDby!s2-tP=UcSv`M(xG%oBPrdDfTV;BJt!g44I(8VAl)@I(%lR#(%mI+H@|bv
zd)|AW`^PZP-po9#wO4#UwFpyJmB+!PzyyInIEo4~8XyqzAK;~fjtbmK>vXgL-oS1e
z@=~DEVXAH50@>!3$}144A{y(jISO!%;i91L1_F^VKfb_<8cavPO=~+XJ$FqfI~Nse
zcXKOqcXQw#2;{?H>0sxg?ri7e?#3W&=)(YI5awoQ@B)7NvHx?$!~2r&|9^vrhvz>Z
z`qx+fxnb=09IZ<a1bTR(DDz4SYO<G(1~amrY8^?MOHH3kTJm8*mg4%T1*ZJ`><gU=
zzD)Z^oCp5Dk`?rNk}_XDN3yEOBrqr(X6IfGy%p$tJ71kXxA@>SoK6{HJpq#>96r3g
zNE5kix#ZaNJ3Lxs*@Wlk$}@8&53IX382!)NuHzfK)rtS>7Vv>)99Ozc*M?e*;9<Bg
ziO9tVPfyY)6OOh2_05@Sc07HL1|#RD#>U3n>gq`|1L5=}1tx})l9IUBc~*aVV=Jxt
z2<`0bVt@W*)o*k!{B!E{{Q2|Pi;EBBFKlDZc4yQmC@69(D>a^~WjI(DxNVJQ>j3Ri
zG&W98NKRIIL6KNl$@|H}H=hHMM-}itEq-j?;=h{r`N|F);ql?aYd5#5_L`d;Z~X?>
zJaFY^W`IdtUv6NHc99Ob(Hez1u0oyd7{jM`pLZ+t2JZ`^Zv)@`)z;ND6ah7xcZTST
zRSWoDxsZ!_SD$}!TkDRD|N52UlZ&LpeLY^Afd5};pA|1E#@g|fuwHNe|9*PTHB`L~
zMx_FI_(rSEs0t8QvD>{W)Oi}{3b<)$a*}BvnM)ugM8FUo4zq=pypISqtN_`hot|nL
ziJ$rBn5-DjqHe<6vh%5qaJU6rRwT(qq2<cy!c!pHfQcU}d~_*8Z}%xSw<Iow3*7!Y
zWHV{85TWFO8hxaCa9$FJ;nO1Bvf9@COn(DmVc|bx^0YO(TrF4M+u1;5G|sm64~5xf
z#KgqYYLdne_jf}%xG(|h{wJp`Yi={fo;AEdJxDtpE(UFx|GQ&p3}~*L$@ZU+iork8
zDgD+HbqDA7TW$}k)F*BaTbJRz^{7yYvg3i;Ybc(HO||saGneoqtrZ%Vh(k~-7=o&o
z`YW82l8zBOsd+;xTHwHXMDU<f8n2AM+Ts^xVM$4iCHw^coVzx%5USiJL`SvxVct<-
zzSXHC2*Y(H5bY}qwa@$*7yP*1#{Q%G;;i0nq;Qi@?)PdLgq!jB_u4N%_miLURo=oV
znhyRuBTG5h9$UTplEM9nI!e9w^?LIDRqebJAHL^d^CbmcCtIoK=j|J0I8xlPW~Ug4
z4c-samE@J2?g}HqX?yP<O8tSmEd-6CEwUD5A12dL3h4qlwQ*kfTbc@+QEvu*+>@3!
zu0|`yYNwo;F4nKYoqodwik9f^7rK!W>UZOsDG2qsl3rDNy3a?mL(5|k<abcN7DaL)
z{BH7@&{o!`V)o7J;gkZ1%axqfP}1zf*DJ(SiQfbksK!ZbB4a?wBiHD}yG1*?Juq@k
zRdKT^ulBo9hP_j=vTj|X;cRS@t$rbVc07B=!-O@*b$D{DGEEwo2LD;Ph~NJ#32e#L
zsq;xqHGQcdN@w_3+mmIm_BX?2HQB9KsK^iva0N)K0%|0;&C;{bFq7d!p_TT+E)HBf
zeevOw)BA&PncSa0<h)&lit=-djT%pd`gZ-sPChR!GcFonI3^~4{)1`!eC8B7y?(~9
zDdK@0JI>27&oNi(R5*Gs^IC~1@o;?CAd8I_B+QY#y%kekD^MN~VCUcv7cy0pSNiky
z>zd)@jlo1K`KE;JdXrQ>d6KHb^rb3zb~_x|a;zZN<6Gp7*jdPcy)k{RCARrTjEnp<
zalgh$kFw^=Bp%Cdh2#tCKc8bY(mpF`vNGmb=b?^O-st{V{KJ<zFfiI*@WOM$%8{Ro
zhhyT~km3)+9tP&|nC;0*X4Kyl6X|au{U5gcV70&cBC5BK(Mv34S&M41=uC8O1uBF&
zJ%|bHwg=O6sn$KWbvZiM-jI_=qjUrN+kf_*8IpKl+1ZN_;khfiJX^dW{Xx5hWV_7*
zV!?aC$|cZNZ868|b<)S;NGlk}!>(7}CfWe9_bX%~<Z($VD2<(k1&+6ts=n)1iGkXM
z4K8@-_I3ZRTnjp`T&phRb+(QVEVPvkwt78~R|!{kMSP&7(LYYHI$f-+<zuM<+BT(j
zgDYtmdysHXyuu$#B1SuN`;Kv)#3FV7An8jS;qwXCl5{W1C-f-<Vx-SG5)-EJe3eF%
zl7{S=jc*jGC54++5Y;rIedTBwR?uj09WNr2_oZG*RE5g?kOzv%Tb0yrS{s#ds`*yO
z2pvWD&Ahx8lZ|cDpI(vG%_rZhiyf1E`Pr_2pl;Lq))gyzp9o<Qp&+vFKcOnBIHqqs
z#E$X3ef^Lv%`D~$cgLaR_%tAU5=Nmv#>^{7QPBV?=XnZ8pz~0Z6~)|G$wez!>sbdQ
zIJwA6e3Tj9O5(yuXTWCB^xn-H(G{JuGI{O=L$|U6)>?24igA7N0!Op_&q;-2yp#G-
z@Be2wfDQ21y5#UW0<%kcmK3vZcMao(qNfLC!Zup&xSPGOn+EaoWPX@3c$!_!df!b?
zl$EQFDNkuYJ1J37TG%qfx~)JvwRp1BF5xA$qSv;m=fKJOz@L_Rr$qmn8j?6fGDG5f
zEaFIPTIS<M;n@BlrVq1<KX373YM+Y?jAneprsY>Q=x5DUmM1LvOeRSABvPTspWj;8
zI^SP=qvX5;fqUB9A&rxQEQ%y(cggf4Gr?RmhbS<4x@u^@((U2hC)se3anxZl%g||&
z*8@Nkv1L?>N-8Uc9F_Vw4QS|oW!O%3F4;-ettP(FFZ81ybaT+oEAIQy&&$DA_|%*`
z3SX6u<U)0$30tFPm*rF=r}l?w(OCw;amvuli4`JBsi`TqIY)ZuwRf)HIqSAjh4PI}
z9PdLOX`aU%Vd3QX04Z0cI)dAmUkk%$S$pR7FW4)>!)4R%gW7zpsq=X1A1d*ehYvLR
zqtymKJ)C}E7w~aOY~XY8Q-9Qw?=Lo;&H93Gx-YVtf0lWA?oET<WN?c;yGt6Tr@z58
zF~wXpAIB?bU~YLo7(~EZ@zm*q%l}$ONy)%E+NGXW;-J<bjntc^k3nauWKWDei8C`T
zx9%(U@&E@NJ?*q)n|&x61W{*w@$1@QSSR_Ym$QjDC*rEkuK1j=;dY`mVL)De#H?Rs
zmhnR3B+ET7ZhUT%E{^5d3w1JY15HD=#h{l5g3{fpDbK=UbF7n%KP2Mr&%qQmb8F$s
z2CS?@MI1lxELem<_@mgPA;~0}Z8iP|PoI>dEkU3;>=_QA9{P$e9kVBhWlTyeNrxDk
zy0>l{?I<q{G6qxX<_e%%6q7oVg-|d29r6kBF;7X&2ixE|m&he7c+P!ZVX-lq#fNu>
zVg@k38Cx`1@Fv+skF<l9=v!{QEdBi2TxXD+X#W<UU)yvM);XWaM)<L|y~L5sM0c*t
zeuiRRZb^;yM)O%5d!Z?(n%H@@=uK;AUUjjNPN_qTmSI|gt%NrwjeFkuIwDkK9?yBQ
z{yne%o}SRv5hD|&wofz-OWvoD-u|eBGuwtd#9NJ9+1pcuCMwOg+o!0+yOr{__##8=
zXmE|j8yyC=3H?-~c<%WLH8X~buNrAFVyM#wwnhI_SX>*HXoO<i8kSyh*s2XUZ*s<y
z8VIv%5!ItAljBFG*pQE+kFbcQgS60(czJ1R)CY$TILS%E$;)?531Ve}xQA$-@Amht
z$VlCPm@yu2ep|C}k}RQAo2Q>Nl13S%@uB%^OyJ$Z>}f>v@WrC1DY?+Q1S;RrS4r;X
z-4W!KwawDsy_<P8oanPa&*?}cs8NrrakI$o@|}IrJ_nPFg`>!<oi+MkWhP)BN>36^
zFO~>^`(Y}+`HpNk0YPiy`A5U08BSuX6=GvYPo+N;TNi}h2Ertt?{uEVG|&C)9~fMc
ze0ky0xfFo`3Kc`^W*Eo*Gbs_(0N?<})dsmM_SI~*@wN%V!slE9Ad-b;4r;HPaYmIP
z4ql&Mub(>_vYlpny{4d~WNw89T9J;-3@pakJNpbRm=)WKaxRX&E+i9B{2789KYP_j
zn8`2QJ)<+wOTg%0+S1-Aw}oZioKs|hx^wij%dJ!c?FEBLA^%pKf?{EC^V4JKWFrPW
zXKmB}e;`s>W~lCvIU|dHzA!@;wV1f`Rjb1;jX=EY2S>><v*KKU{Lu}W{=^rjiuOXS
zUe0hO_s|9?XZZdhXgF-SWv7?+@evt98NVfjp!H=n3a8fIW{d|&23)V8NifZwe1_X3
zr*4$gHL`*9p~>LXYLN{bRNwDnW+Y42kQ`!!Q;af93;`2`CcLLsd-2JUxt+!tYPqVd
zTWV~kR$G4mtFPLFEq|R{#KZx5vr7GWX8Nac{*>a~WH<{oh)^BX9U{6a)O&1OgJYf=
zl&52R7nxezrC6<Ak9R0~5ycom1SZfeG-f*&rSW6ZG&>XqEq#xQz!n(ItDU*AG{9SO
zZj*v-zgEf}HAW8@QmmR*ltLagG#P<h7TClf3R;s)XET!8g-s6fG(hh>^AR8RDGG<S
z;2iqT%0&N~+UeK4emmsd&f5`!!6>z=l3Y+2E900RK_jGrD|Tf@F6xl7*Swig*_j)r
zL0A|AWq;>L8qVsS{$HfYIV186fSNNROe+&PMFxKwgK<!eHLjgbT{<npRW1W%_RxC7
zZOWfW5n*UrnF_+5;G@@TW+)DHz!iR1+~)bJ#Ebf_W!CtH6+7t>yvSvZ?9R<^Qo0#g
zkY@N4sxaq&7{98Gffeq+UJ`?_4X(k_`WYrO_~}cGycwa1d61QAk%~g$v`CqlU3LLT
zT9Rz(ElfIVE=F%MYw_<z!agNhOj=FtVjL!DGAZw)6%vdcK5=wrKj?07wTWvW0m5=2
ze#+9uhJ#PW#J|OR6)Xmd;k(f`oPKlR#fX7)m&Y|rIBy{jqt9@F&f-M(V4uZWKOqYt
zrGiE9gKb}WN2#lVn6mKolBcAg&yZI8Uw<6cmt<N$egVPcr42jL_~e1noLO$z9puS)
z+9^nmm0bXP)!|UVxbMo9;WnxH{ch)=f1p1E)$4Z{i`CRmZPyX8OT9Q8(QN|*gO7de
z&tU1y8k(Ann)#|V+%uQWhklg<oK3x3V>#AWN2^KaQ~v?Hz;GC5KJu@_7?>PaoqO+)
zQW|zS*bO}mS-gUUQeR$ZgMvbxfM=`I4?_$iqG3(qQqgJ7BGuz?G0vb;L3pOLqS7c_
zh9L>xEP3%VCeX5{&k2o+>ADfD!Q=gHGT~inWm^??J*ta^ns5i@99N{eoM;~fCU;RI
z+$hf27&Y~3xC~Wa+_0J*gj=38y=!0PwcEt+Y~mp)xI#%|_{kCk2c3;QRmUoO!TZ&9
zP;AP3A$t=YoUTrtEV~z-KQlXn!zFry(e<0^E(fQLqz!ANv1X1=lMREpH}`cq^fwA!
z@!)R@@Vf=9Eyg*>=1;bM=90}omlR4af$lFy@Emr2lyBRsyP~6CWN&v{8@Cw-_P-Fb
zVRMV0jY^+ftwj%f6LR}gn4eL}M7iz0qECi-C71F(dAQTWXP^1dYs3e8lG^vE{X>b)
z%*X$=#nj^dYp-7wL)oLpK)Q`vU9C}N22XZOV6}ZhGvLomcCZf@J=@B0eF&h96LD^s
z>v)qkQ_KSbn26P8x@WE5YLwW89yz~5itkk?ZPJC-Yu2Kl9M@nZ^)ZtD5J#j8kon%2
z+~?;$g=3r&`1G*fEl2fas^I=F1Hr5zn!heJTFS&WzG(X%{!l1-0jg^b%e3XKlFNB>
zn&h*QYGbto_d2O-y;A|GuGRK<9_rQ?6c}VMl8i^{hgzmnW)GEH1U9v>)(@35U<X~;
zLE&xuFJj!?e?%VkY@{(bvlq(`_qsc~yX&5<x9R&>mmOGKWa_Jk-dINK{)FeQTiKys
zJ)iVwYilPJx7XZZ9M-Q7Br6#jrv2Sk3y2HGBJF347PT-hU0+`>gSfYe!Z-D}dZ<0W
zy%O#)1iOD*qxml}b8Yr}WM+=Pya5_*wcek&fpflBHPY`kB4(4JCMhUI5iR%|J(b)1
zqwfY^CQliQ<ToGjW*?dHPzQed0H&i95zo>Eug&yIK*2cu=BvVsN4$%4*zn~G=3KzT
zo%P9|KG(m;Jp(cH(pIBcuaYYMuLWX)`^TuXL^+71Qg=2dTAX#p#HYkdY$#J=*$gv!
zH^b+^E<*+O1Vlyu6L-}JeMHcV3jcE2&^b}R>TTcNT-8uoDHnLNJCf3d`#>}ECIQ??
z&i7WYpz425m~W43&@c>;a@e1o14C1?PyFF>(KC&ODkj)Py3*#LZrT6jI67dVC0=%7
z-5~QQs`JyqVZcLEu)s3Vfs^0iTwg<(H^Wf|t(RYkZi-)J<Mk!jP5kMky@Ey1N|>;+
zu~A?AcNAv~+5g#kpIsn9q88jIR5pfWY~Lv<g*#41_r?i%pV<llBFqq*6VL`vXTcdc
zSzdfXrx+;X@)Yy6uj>8jmSv^Gu|<RN41%a}I<81pQTM54{hD?~E$jC;b3SFwR^z#f
zuHED|DQMKUPkz4vRIvZdpi%*@fBrNb-GUNVGZTQz5*9pWjFZaE_YpN4I6dMfMh^>{
z;w*eN!_|8!P|oY_+rX=K%=cENE^!W}%r_FC=amf~ws8nr3XZU_PG>cul*?YF>Q)V(
zi0OdTKpY3_zWRzO`?Uj=zvKp9bDVPc6|h^C;Z#k><t2UCG7w(Q^GGnbu3l0)WIHE&
zFQM;Fp#Op3`rS6EqdM(()zWd=$`ghI*^&mhaZz}}0Q@3hY9el+uirgC(q5L8aVXuZ
zxz6xyox8#hyS!qPc?O)zWk!e&XW|bbE~V3-+HZT79)!$>!0WL0Ht8Noeh1YPs<kx6
z|DnIh1Lxh3e9N0swtfZfb;{1T!JKSlq0gJlu5U2|Vrbg0O;b+eS5XyQMV|9HY|i`f
zqr246@wYr<n^YMlJ~9l63H-6rc#)6`_OqHMdHdzT?%ocM;M90(mHj76;_;*bN<9j)
zl9&yL;Wqrs(tdY@*xh#mI|Uo&vP~A<9&0DV{vEWp;KI0T_?c}}pHXsGzfzv5tV|N~
zgocR$>!-7&^};qH?3ToXgE$$lhTIe5xG5q?jbE0&y^TqK4KBj5{{yDjh@B6$cclM%
zP4W4g-5(DLJ?I`cevucc?=?-8=Zi;osSQX{&?43-Ksb~xTz*_L_7P>ROQxD{oJ<@i
z`9x`I{MAElcg_sZ56J9`5FGi|o3+g6qAulHB%ZkrcauetTe=>(%80Wn>3!Z7R;Le?
zy4ue>S#xwYognVUd72t|jw3fCVkZk#BYim!H~GazEy|4`U$HVgJ8Et-SOmq({)?~k
z=AqPo+8}uD>U;m4dd}*P3%Tx9&<~_#|A*UU!yj5wS{M(I(NzL@y`9xiTF%VBe^_w5
z;d6JZ&;syVs1LnpsrnW@YF5SJO=>?{d1&iPSiv}m%!4G0VCHy_f*|_q4Dk^*9v8e6
zB(=_*SIM8um!(DdJlVQ0B9H|E!Uen42UrWUNwemR#|<gI-t8X4I+{${ty~MHwj+=K
ztlzq;sV{;iX;6B;N-y`Iv)z2h);|Ps44xJ+D#QCM_h+tVpO;C;yl3Ft1g4cboiOOx
zm!;R$4YHy%)wYv-=WqD>#$LZ);H_|;-0hIwd=1x`d14X@wrW`kZ$XXBEw|3$kgb>z
z<<EVK%n!}{Dg&S%L`?~ScZP5h(*(dX`0(mBhQDdOdoY}I0!FkDRL-L+x_gB&yag>c
z_Ef9;;~sbP5po9JBQ$uLy=LAW0>L|9LQM}fPbBs>tz&<uwoAq~C|*?KJR5R+N?zy1
zQRRhe&gh|9sM4Ed8;lUmTic3z<NYuYEy2o0!WpCeP_KQm&oN$8OHL}2)VC|c8w<e|
zl)RX;p%uE>%E1||nE5uaj<}FXdJC_fDb{BMHn`io+S*!|U;oA{HO&906&>CH>|s~h
zktH=|ij0gLpPaPXoRbOY`b^*o{KDYJWM{Ns2M9fj6WIpcnJ);9qDPV#OLKA}Fkff5
zE!u?tNL$;)Y38{0=XhlTI1X%0!TXAgvCT&I=Y(jzV5`Om#6f3QY*O}$k;Wkb%^e&A
zQkA65ihK3_L<6D?;*=z9!$D}3%)nGD#(9eZN~ZaQ3HV&ljwlOz#>O=i?M%_+!OEZp
z!qX>^rILOt>x3lw$;qrlgme*)nQW7NwKxK~pk`OF0gs21j;Pjt2*ombw|PK=Tye`r
zI~U;UOZKz4c>D;vaURGr=>4qAEBSr|gCz0j=fw(^bG%4dYk!~qmI==&3>1~^<##p2
ze(NIY+y-(h4rUVtxFE{k?{%@*C)#DJX`yKE;th1W;Met@8XgNPbe2@a`D$XQxVMnC
z>W5ia!}l`Km<GoH^e7jiS4pGCs4X0Vx{oGtY;Ev1m}YHXKx?;eXJ`-w>4-<(k|=Tf
ztE!UsHiLfA?pa~$-5Z^$nSqYEGQEZtWYG-4_ur{}_9|SD+R+A5cr2MHB1uXcftqRj
z*K%y8_1@$9yez<R`TB6I%>TToWLjkYU*l=-6ZeQEu+W|Bp`x4;ciU<lqUV4@9<Rl!
zUC&zi5Qx@)Mg1Y6%<tStF*!l;ArOLNGR0DbS_L=(VPMOnVx%bT#6<EfFsp}Bw3i~-
z24q`h7{3vVf^0CGQG1FG`eZ}t@#Zi^Chrydr#IS(gk$k-JXyF_wvy}?OBRj@noLP*
z_$o+#Xbv9u{Y_Yb*B7t{s=XJ)9LC&CtEX%H<}H1Wk}&T^;V&LPWC!rCqn4b2g`U}B
zJOrvb`;Seq9E@3XbW6)*cijdPgd&JN?=X1`*#kgkUYO;bRLawt?*@1o(Ah7wM?3Zt
zmOEWlxFaGr$OXtItARZx_SJZj19n1Y)P5cFIBG(m?}38*wAne5{M~YkzJxbl9B=Ic
z3W%~onmNAd-1hNYj?u1=yM^ad)0ZTGb1l%|Se<u!uoPt+OU%n(hM5IiXF=HgXc@89
zi+TI5&qYIz%#io!&FE{p)1uaDmE|f+I4T{sd9T-KOJ9}^_t5w_r0d!^gOf1qn6G;y
z$oZ@ueGKe&z^~XPF?kQrT=)Wjt%CU&>J~qwQF;A|4R-!_5i8|jek|Z`zsRGim0Ex$
zLEKy0`I$*JnSMG>^{c2}0P17X?6uM-EPB$nY1MU{KaiVl^V)ZV2DKB$2Q^q7l0|38
z(22JvL>)4*Q|dQ?*yH?JISwHu#bc*20_AkO>%3!Q*x*3*{i;DN;9TOwA(2zsFS>x<
z@a^Y$MJu^aF?eM9!$yfQ-B@wbhffwn_P^W39lg0=7-T5z?Ys7Y7ty0?H%gybOx{9s
zhMe|C?-QPMD2nV0<KwGw?I3S~+sd);(B!~Tpzt>43rj;(=%fJ%27M2%g1xC8sb{Uu
zAg!T>F^<OZ?lpwy*tpgTP*@s$9qi@2vv=Sm@j#}}F^6)R__y)GCRX|}vGRWLMRRLy
z#8<*3F2ZtDx*;hYxSvNy%nd?>VPDoJoz5uLq&L3H_jHIwq~1!ceC9X~nxI8^*vsMJ
zL^)zH^r7QNz4Uu@R`gmP?r-;vY;0D{NxvN9PzjFSgWN<8>T~v*4qG&S+JC=1{SA1D
zxaTN;S^z>fmnl*A?*TE?c`R!~_~qNQ=S4W9ObqL&86d^63{JZPh5m(y1;!a;w@7xw
zN*MsDV|j*YseWUPCCITJ4HbW?y~d)LTRG1E(6fe})n$*Dp|ckzXZtKesf|vw66Vsz
zCrQ?ZhQ}c#dtA&qLmB846cm>WD&rBYKJvsQ!U3husuk;utPzALP>tqZTA$Hmk*^E=
z5N)$BFhMIFvR9(W`tB$C2V^~Ekby71Lw8mwo*vsb1(5P@^IsDxw0Vl+YeD9dkv(%R
z`-~X|)dMMDx7m1J_$BxrX57Zxp|V@O$h}9h70lisrMBWIpgbR{iqwa`olET?)yr;1
zbjAOA0Q69hDO^cHQ{-kd%7uO?JB#Z=0;(}H72Wuu_gB5;Vz5cx>JFv2`RaH02|!PJ
zYi3N?xLW^K78^BGsznPCip6C8k*2+gp>pk|ul=X|+)R?jh^$5M{UL3=3O>Zz{uh3z
z3I3<@VBb~w*DTDJ%Vjt?#cP!g!25M7xE8`k2EO*tCUyH9RvwM`a<b$R6KiM~JiWua
zh9w+U=J(b==e0dDgNOv_gsg@I`6GEBxiE@8c@`o->4M?;(`c0d#Z12-Hwi-uU&+dr
z8OEv*1#<uG4<hl22A>w#ha!RkIm#7V6&l)E+lVq;p<%130xiOvO7V6gs9Uk_;1OOk
zs-c;8=@H>~-RRHr<b%eLm@U1<&*C8JgoGHymJw&9_~?=l@6|auzj99a;+BZts!6G)
zv!M7+h1kKFQGF)OxjA1caj0$Ml1cMPKU<~mwJTPi3E(ylrt;;xtZP?@Uw0}wZ49QK
zT#p2pgy2waP<A&N8OA=cmVahpv)e!PEaV8^r=`L8cPy`+l8x1UvHnTV@`GxIpLtSo
z#+^oXLT~-?n)UB9>h({wVsfs0mvJ)ePuH#6?~On_Fro~XAZw(=BMVEe7Z-w^no*&I
zS2Ev`rD!|V)jRa&iU#?A&-mQRGNPUyu|izIt!MlqYSfZNv59X^O(0O5q6qyIW*Dv@
zKR-d0cx;=QyY`^yLjh$ppXjk*?H4Vv8wqGSxN8(<8+=Y(##b5qJ5Y``&?YFI;UXMT
zPVfIt6#3mP!D=xykkbUN=YN{EoU&<TYKFVYQwOZe0f6K^H@!v7<W^NHhnyF&83CTw
zVZ#5@{bq1QK1m-%%W3-IRvn?8qhrM>;29@5o(%GhB!tC3?Mo<}zzzXVb>N&GlQ2@n
z$(OyHu|C37D+djS1qUGwf~=v@s~qxrPqQnSNazL$)n3F2gBH+Pj7MXWE{5TO6#!UE
zN6YXycb_-0+4iktVz91cO|Nm$6<tQRG)_`ZW1(5CSm<1Gfr*>ZnK%Xgs4S!M`2qAm
zeV(3@r7g-HBzj?L=^Nx5m!9a)t0m>CSSsmvh9%RZXGi%AU682|Oyei1AU-z{8KcvU
zebi}T-T?{ngTemVCT=~nM<v-+=T7eY@vSH+`9|%ShFG{KO<taDHnPD9F$vi{73|7@
z8l3cftI#2UB@dQUr$$)0-ybyG(w=OJ9K^94+(uWI7@1pp7yWERNS^<}qhdA{KDn#h
zjGAzi5C(|RZ~+(HWks}~@*^~F68N`vK0}i92od-u$Hy6TjXe@Y0ifu5IwU|k{Kbo~
zJnRm@`A?Uxx&<6sOJo?{)Za`Ok=26wk6uB3thA%OXwQjKSBqm-k8%T%${+=`eJbe^
zJrb0t5O(ruA=sdJD}~ba9*YjOR~}V+XN|~=%)4ck9F3dFk~Id;OxhSk*<sO_(pWBL
zjlpCl<c9bdQNBmr^YTVF$8lU!%IUgA?~p}=J-Zn8y7GHvnX$pHRapZ7jS>DaiZzw+
zNf3Ri>0DEk*dF?WthB`&di+}BTV+>tC{Y_eRKA`)6V@t6&4z<Fae}TU!P|xOw1VOe
zHTSFIz!87f@GC)1YcjUJm!Q0ocSV@{l%9xdg)Z^H!o72O65X3nGzRn5Su!(4UOWf2
zS?*f2DBKrEsRDxGj^!NLZAP8bgE&4L<~}#wOTJ3N<ESqh>NLyNv=ygy93+<m)2tSM
z)yBX3os^yYfE_W~!lBz#6-6epk|fS>L5%E*8zF1u?}O-8ti;{Pr9mj~>Xr8qT`n;1
zyV)r91ROlso_smUb`AMS^+F~Dmz<%saVwP9yaPE^*yG*mmw|O}ktx!cG=9ggUcIYh
zuTPR1+&23GoBvN<vPs<a=_cuG6bf<Q`aOUap9JFhO8!tz)ezkMH$_7{d6dogUjf;d
z$Iy$=<${}a`$rT&v-#dZ9816`-=7~PqPldubcuL+(TtK?eplUbctd&+ro-M3mIgVX
zgOIGiA4^V4*}RQP;M151uRy#oDpMiUe!71Ar<6#~&G5Og4r}p^N+7D<gLlyM)k672
zj&!e2PJq7N7cxWB#yZG^$qjiOsoZ6y{bM3dZKbA%cC{H=!Gn8151)#YK64+)fV{KT
zQ3`G6=B-7FHqo1Rx-*A!?bMj*HndTO--B|<7(cYxWBGtX6*}-c<ag`%d#}K++X`;9
zx>v|y+#q`>s2lq!_}7+17MF?1yK)UK`v*pnj3%9)!La+Abm74FFrNN3W(C@E($R=o
z9MyLc@+{vsz!(e2WTje7K|aVx<)DZ&Pwl5iFRGhQtEnX^=?4c1!dO6|vp8yUaNNM0
z+5A?7gG%kh4xdZmCXqKDA0n6EdT%x(mjqIM*w$Zbq*{jRb80Ssfa1#Q-|;PV-n*H!
z9BnIS<cy%noTUreSgn@O6D27Mr>TG(MK|s=3g?~AvE<4=7(Dotx&N3LPt+A5!LSvB
z9+_cT-%aL=Gow+tz<?l9nG2wrHB%eZyh{`dc(}i8#j;sS!4*4GZ#Zb&b;wlsx0*Ne
zqc8B=Kn;P|=^)=p7m->o5Uvulb8@O}u2PrPuE9<h)#j{Ud9Z35UFCJ%(}Tz^-pB;g
z`}dX7bv+r$zhO#!ql6t%{Mr-ssHACL@`;II8z&48;|YwFV-6l8`|`)KG<s)-6O4n9
z%>}oN8(s=~278`0{JMlPWpZ(XdXG5Y!;MCb4TRZNs4@b<2-E|;yvW;~j#?uQ&VO?6
zW}c?#NQ9poiux7w0;+yI@TvfCF=0>Pr3GobOu=-%9*Vq<k`Bt+gNPnCCtvNnB}ywp
zjtTn($P=WL)e8CMwV!O(g-OiVo(UJBU8F(FNb%)tcK||Po9*yAC!-TIIkct$S<QZ}
z-GZ1{+n)(rV{96y?^j>bxv_m%r77h>OOn;y#NjG0NCH=bmwAoFHe&0Z##bFOO_{Yg
zmf|>|%o$e*WMNW5_`n|`NJk{Gt?0B4At_+QS*ID}-Ku{UjWx)4G7=O9J<}%X@Wfb0
z=~`Xwske>y=XEl(TR|x}EjeEeM9YN~zmKfx=24FokC+tdI*lz~qenaqlY!;V&VtgW
zm8fyHBj!+M`;CG>{OT@Nfzj<YymsumF{!mGbZ0kaN`5%eNENY(Vuo$qWm3VFjytaV
zbz2w!#j1|VWAY8^Zh1Pe71&RWjhW*|vG=KoU+0P*wLwl+KH-s+w!GhmpAJ6?&_m6C
z{w)jgOFv>*oUQ19zc~w@%YGMgQ^e5t?5p9Xff(|p-V`En#;|dX=WJZb*VJ~YdfIj%
z95t>Eb})I6DQ>tt2b;Z*xN~+|PzdMriO6LU|4L<+n6*zWIG;iTEZ9zjaV01C{0MTG
z^SlP*wDGrpWi|tRA0`Y-87Jh4LHgT~gCU5bzUnVJYl{^v1fpm|GegM&w}3z2$bQPj
zcf9dSxZpec&XD^=jg#<|dIo+sec~%YMjB&~6dfLe@{9RJ>kO-eJhf`=5`N6)yK*o-
z0f7Z(LFX=}&)Czu-$d%&_aFW$fETu~mJFeA@G$I+B)91A-Yd*23sd0(lou9i@QGPx
zu7%F1#;YkvF*p|H0_^~wZgB;`2Jtz+4pkc?y1UyNBWh<7`VcLiXJDv;1W4OaSYn|^
zjV&W=HmxxQ4}>t=rzLlsgWz#BCmP_nMWUIg9)eYW`=KU4f~L25F60LBFhsXBjhSE`
zgBX31^~o>uX;HE4!tYZzzxhicZ4;SA-h+4RX=e4Yi{EER1f_RaZh$4<u8p@I`Bs?d
zW;~1$QH`XE&37zu!jGgL6N2o9wZ-Zu+7%DKKNB$iNzhcSI0Crhe95C&(@aoPS2ut{
zYD5S4Qs1g;YWggr#dPXVrr!;w@SNn(-YJM5hpCY%d;S|fI(~hGQq9QFHQA6LoIPTB
zq)_##sVUd~&5@*Rk)edXL$9)Z%5HyC$T<-hdh`zICmp<LmM<cv6I_7_4hAWK=tZHD
zG;thEZcI<{KRJg3RnGujNO_|P=d7HC$|r8s5pAUn3*#`?pgH-j^C$!9H@E~Q0{sL2
zU5_{Nd<9n>m0KnN!~$)I-2*GgHG?QXn1J0FVne*738LG{5`&SNXa(_;K*;7rR9iug
zAW?^l5qt8x?@>z@z7f>mPi;byW#@K&<}JHK6845gq1qLT{IbGN(nz}&oO@Lbs~SCw
zSLi(fbE6lqOgw!`(NPYSXs>(aA2rH^k%0Tti*)u9fP7C(&CWgpsC4ZDbS*NN&p8sE
zV{@)6EzwUC4V!FbPVe7J@>160^OEgwvsvv3K3V4qm=l*aldNqR%aNa#n~DFS96QD+
zPXR?@4K{nO#q$Cz`o8P(y!}Z=+tJzEPKvwvZzq87%3&tpa?m4jUm*%)-CO|<N>U2f
zfo}!;U87A_^FN8N91*`wK}#Q~p|2=K{{=svHXNdiU_#!z3;Qr<O^TI|{PbABZajA%
zkrsZppenqb+fgU*)S-LHW!7=o@3ID&{hsRIjzhD@BRe2IJL<%nHxX|AyUvwYSXk)7
zKT8?v;0EMOTd>J_yQEzKZB}yI9bwitrEALZaMdG$opF};30Jt+tPL=jj3JyAdd0x(
z^49LN((ng-F~stZ&&Z;Ge?ZcgA_O>mzHpkhEKDD=mBB4Gdo_gpZrsU*T>pq{EM5mZ
zTn1bW!U@1^Wv%z#bHH4zHBhtsS3N4-e+6tkoKu<iXDki32ThKeJKT1V4iPlEwicaO
z1^lEPAiF^Uo&_xkEVz7H0L&KDO>}aur80umZ0|?C3vdTnefFE{uqgz~U4GTjo10$X
zHSJXnu!lcYw7cpF?!*;Q7XQ17m$o@^afFB6R{wDMaGz=9y`2}#m-e)l4rJhW;FcK=
zpjW{ryYDLSQO?%CuBk8WbyoN6Us`r!3+)zH4QlR}VDEktaLxb(YKaPbm$P;}e|1%f
zyR(_<_w$Z*efbi1cBcyaN)dhlNkhfX47m3)q|6q5_32+G)66_0nro?NjYeU@-@ls@
z$pLq}rW^F3N9dn>)@nl+e3oxbu#$mvtaJ01OH?DHR;dH{Imz`djzDitv?#-3a~2iC
zF4@>>Fd}`u(bR+i!%vPnuFN!?l&f}dR5It>?&0n-nalq;e1n4|QVz>zx&Fx87tn#Q
z@4;TZ8R6(|e*^%F$FoM=3So%_tCRW*!E9Y!{<|T7Wy5cJh8E2Wn(9XW!!$Nh*K*D=
zpwamJ`+Yc9%dF_gbUt9;nMJb~Brvusnp&7@f*3A1NKO^T6Y$wBNNe@XjOY_ATfm-&
zKo17z(;lY5t=G=fisVv{K=Pl`;_dTCa`?QIcnJmYRWn5IuhuHf+L1&;S#ZhM{DG%1
zFj=T+4Oo63meyTSEVyJ${$N<KiOc=%3GyTAyLnBBPZ%6yo~&HXbVYaS*g&GX<1t+I
zz(5`{f1Yx9iikTo>%{AGyHPG}-WHxWt6(9OOjlp8NZG5qTyh<gCGDH7u*6}nHz9sd
zw85(>^Id1#y-=G^L05tQiBsC6rCn0^%n-AaWat#@9B@0&Ycu?<W9}U5ko^K5nY~&6
z4p2=KNBl3UcF&RLwT_$A>US%8XquL|HDQcLc!$+Pj)w&~QE_ag_1{O%`@dCacO<u*
zG2g119wGcke3mSir3nbb>-Zmrb~qev=9lkvq5wU~!>tz%QlO-u80>UET512#QfIpx
zM5Ox6muTKH+O0DbkL%m}|Ln>bz+-IQl_?-tLV<V1c4e283=OM&6MS=lu;=m_>k(Ov
zgr_tb8OC3|w;_h~UFG5NEg|*yd-!c)ZeEW_%4aw<S|@?)pHJ9bz88x>thAvo;<}`6
z0F{<Z?^4a|@aTWiwQ>C@G5<wN?j6Ir&hl#xX_0+{QnBQCz<YXOug`mr8z!4)GT`c*
zTdg&1CBk9JgCAWvW$YCO|62@qz1)xU-xjhnZJOR;C&^BR(tw}M77Y*&GG?e)<#XV^
zT6h9>2Wi-3cx}Ah_X)8f&Hhem<kJ53)z5e}@wpWpAjkbU{h(l^H~_0Y5Dz0;um^=z
zuDAW^i#H`o8o1Ab%9wggY5RqIePpD`Wx%2-!{V`ac%+-`GUt()YA@g<17|{PBw5ii
zQlKVviK}+JV&8(A8|-iIyU8gBzxD^T-UHj1(<o6H*Ey<dD6M~edS}|$f6R>af2Jm3
zswRd213WZ2%jKG=zq94_CLlH#HqJy3XtV1Wnx>GNh18LOu*&$_o+uHald!SXsEH1$
z{|w)2VVGK9ul&JhV^5Y+!<wELk<@257m)-m_Hg&;1ENA&fHV#a&XE4?`F2F@bHe8w
zu`g4<O%3=JoeAHyS>m|J&HI*4pWf9(CQOI2P~JK24^mDRjq*}y>PpS$F2ibp9OOBX
ztpK+HJpqSEBkR+<>Wps}ZVn7@FA64ha|jbFnNpMLFWl6NJ8m?npC>a{nZ{mowK6S>
z9Hzy+emY|<h8xU6K)JOx_C>&6MMY(`ze#hCqr1i2&r99$=M@@;KbinaKe%wVX<xVX
zMtgNI<d-iFwXnPK19}*;daTN2B$EW_XZi<(U9fZ6m!pA4?|`y6J|N|K@zCFf?!D!0
z{u>g79uwc@Un#@lw+^SuowtsUS(w}Xx`9XQXePKCY4H9exwW_c{Til?iS{8A_MRA8
zIa65QM$ozjZhf85L)O`2h5=V_1=b<(&q<6&Im^2V7@3+yds$(&K8JjE=(g<j<$geN
z=EEPBU^WKpbpoLiOWOFmWzUpvZvBdJgPxT5?E`2w-<5Tfy>O1-Sw??p`QxzlGSo*&
zU`@97{NgihL*&qF^k2`g@R7b`yJS|WilHeWY2-B_NunzE@4Rnn;nv+AkmyP1+{JR{
zo*rz|3uae(Tb$88_8H78#lMZ0#a4O_9);;8HcmHIqd~uAuNOv4$)-U|@=lFWf9h;|
z|8Sx+^kZcMY|^;Fb!&5ZIk~PJ9RZqmH3lRjbxcbR2>v@z+#A!^Gu|!ow+0$N{dRdI
zduG8luva^6Ii-d@sh<<{B@oq-W5+3Ts0xA_^Iz;TZ;Pl=W!RviR#F~Dzt^q#r;oh{
z>tUgH-`C?N<UsDT6z0wu<WI!)n`twRIBm=zhxC<nm$;o)v_hbcnU;%}r(YQ(&KH4*
z0i*phGaKm+Mx4VJ_p4x+$$6#XveBLvWGK?YYV4FDl=Gx+ElT9;CXY|=8!i~ObM4*5
ze3?40ZL~1@_5u|ca*VvUDDIP6VQL6ONq}h*oJ_siI0;Ds3g03-c9MS%PCK_lkgmC@
zy^|k}2XwO{)X=$NhStO{`zXW`PY_0hk2ms5<ycbm_h&{2c8CDuvErdnd*{KAG^aRV
zdKJ2K|7S%qKaRUGmZa;{0M`1CTbh`VhXszR!}>fI+1&57s;K^3mb`w;cUB!vK4dWk
zoLun${QN{M84e!%9EcA82;+t#<4n4N&_GJ9UtCYG=>J6BvS{gkMblJ$K(!ZxiGi>^
zj^0sD2@^cjRaeWfp8ME%;*%&mqk|y{LXO(j(6>4v{XD-@7-1OJ;UmYS9g+B$yis+$
zXVdSve)2gq)peYL*g^Pr+4kSKuz%fPUTp8Qud%i9%t9Au8%hPp*vGwL(v!&=hN}eK
zenBcaK*%`bZTctI=H%NG1GT*2v3FwmbtyAb2!~<O))b;)OT&#S71R1u>b+L>)M*oB
zw*@_rS$3VCgc#av2OEwWx1@IbwDj^ZuVr&NXVN?B;z)obmKT5Yx6ecfpN^aEwvmjy
ztm~^ruu{5dhVbTT<lP}bE1aZO<f%a%(@#dRX5||SlF+Ws!)5kpKn!cv>6II6>p_7V
zLvM>qNH{v{XWr3n;muSvnxV-Il#EatyfykhzSn*;pIzF&5m_Vl%z&P}B*4}>7)UJF
znjaSoNhc!wf%tERC$#_g?-Mp7H0R*ABN-fA@S~5x$5S&iX)_uXA7;(bK4HRyL}R=|
zQJfA9s%>Mz-*LQpR<naQNwToPSs)c%6R$i!yx^G`+WVu>8K?@ey&rd!*a*x}vgG{k
z;0!lfMxNHM+N<gOQPg<5hBWBEqnQ~ePve#AA~$GAm~tT}Nw|WuA`4~9NHTu_+tS|7
z_L!h`D(ti^hg4Sa>WGqrKP(sTmbIEhN!(v`)ae467zBCgeGj!?2wf)hcNi*C;D=v+
z5SKU}_~pGDrG_(IsDf*?kVpN0=&&h(N#|>H%1k%{S~g7@Rw~)7rWo-^hEcf=Z!_Q5
zo@I)v@Ipcjp3%@jRvsf2_@Z>Ek{?fj$T$fSdatd5CxtpCc{Dl5Zmd!=GPh+EEcuR+
zK=GYlLXR|t9vDt$L=F>&^=!%H1v0@CKuocf8bw5A0+RE{PN2X_1Z^Aok~|C_AI%Z0
zy2WSq7SK8VoIU>`N*8`B^n(f?6^4xe$4T6ttsk4Fvi%-NB&-lxey4?stlFWdSK=!=
zQm)86fQ=^=(<Hm`XJ6!2p{A0N9u>>=M5F2Ide(JSv&Tf#7KFRbMzoA_ebPfb&H6lJ
zDw+>d0u_(i7pwNPMNj@z9tPPN^f<1@AxtHS{jl0uJk~h6g<!-NxhCCGhFBuw8?Q(>
zIO5lzuNb25o@_Lrpm(yxxkbctzn0=EMBx&2koB8Qt$=j)E4{WZM^tD>>q~%M4o~30
zdtn~Z=PCEYV;)_85wR6W_4Ro50(t!OT*Z!c&9g@u_LD;3k7F#;_)E}16psnKECv&=
zGap%2dwUBF6{AfWGyS(qAI?0AOQYenJzp&rW3>WXcxKL=3oep9-W1xW-H=JRecUf+
zoROlVdfmi{SeWPR7{nOp!iw;t6XGQCF^O#^Dq|84#i7)xue`%@uEX(4E#Z38b|cyl
zC3>O&@n|LjbE5@06FxO!?@$et-W|ZD_~~i;w%|PQ=lM2(|Ee5OznD+nT-f^L7yvXp
zxN8Z(*1>#%({&?871K5|=U%<%wOISVk1@FSan?F6j&Xmb5pe4{pbdl1__ZIo&OgLw
zi5viTj&&WMegg%t&dWQ-j`PXf#-s7kTjD50E@2hZo&tEWPba%@3fK=6l)l5Yc32sB
zTg4n#Fu<wk%@B)f_YJFeDjo2A5?gXJlI$Q5aTc+<!y7;!L&g(?*PYzt9@nSVrDSZ9
zwORB8AGWczyg~i^ahWVDNciYk)m=@aH+DBT1;*XR2g6uF=3-ybS{Y*DWhLM%uK^A+
zU}z`uKV2cqP?2nKDFYO~eIzSk*Y*ta1*Hd!>pbd%U#Kt&I%Y@E0LZkO=NSc$EgHVm
zY|v>$nA+b}+ai^LwJ{RSPE<ei?(UT5!GiK>vl*5*;L+by>-aS6(8%D?&;F@m)tp&|
z=$I_pJ=U?q&yTQ1<;94Z^QLKUu;_9EA}Bx0(mtC}yTvT%Dk_oNM?se!be6YCPRs@n
zSo*Q{dl8ek!QRjP1ye)KlKKZ#FSE}<!zruBsKY0y459X3H_G23JGR=lE}W(N(k(FO
zi0BdeMDQMD7JQB5T+jQdY1XxRn0)sKW0_k{Z8nt!uAb>EIvCPGh;TKW_<Mbd)4LKG
zJVRw(@(Zax)3yKkqg>t=wTxr2{Gx5}Ca7dsXp_pv$-AhcVhl746ha+AB6R1*5g`5;
z>>PO0{|*x!8L2{IA*4&pFF5u4Ik@e7d&%eUC7o{ZP_Q>8;fT`vjR|$DT<c{~DJg(=
zTnpm~@urCbnk<R9FF=)mtp)+;aO;3JQPn<XF-Ed}J)h!zc-$i)yc&E7+?{A)tPF)*
z@e?Zx{}0!)`KNsD09rq98LG)>hgcjmus!_KXqHfk;<c`3(cc{Z#CGg9WOdXLI?;0B
zbEvX*UONfhj11^?ixRqu^(Su{ihhVF_KG{&YH$KtM&yJ~+PSzS)n=doYJRn-`OgJB
zDzFLI5@xLLnCx@(1!)p&2(t(0vht`EBsl;MF)~>Wm;W9Qf1$9s2fhya6H8jStzFnF
z0eJCMZb|Jhr#<B}x$Ac&(J{&^*yb@*@z;cgE3u(H<mEVN<uhczt!d_0h|#H(*H2L?
z+C`z>Z%AWGwR>LHaXc59*494^v@}<+L0^f9d~<8=!b%xIt<Y^^10)AVYyosmXC>DY
zl24zVGbvQUZ+_0<pw>6LQy*BB=glQqq`BMPVoCU>imeR?kEd;|QfnH-yx4|?ccq}f
zc8fyT9kzI6z~r9D^9JuWZ>%mTslwmRuNdVBH9HsGwtFG6tdt!}kfgVG7CSoj3c=Ok
z3>>jQm8jslu)yKHS4b_zTS=Y%B}!&CX}R)trG{)zO&;8AWr4kV+@#0^U5UooGxa@|
zO%?HUT76BI)dMizbwDpD+z<O@Vl-a_XadYX!CB{dILt)ozq;?Om>q3xpM4{Mi2Yqd
zF^~CdAN2YJk2<^H_N#qqh09oE*s`_%CsVKW#E#UNcWy8P)9f;5=9n}-n<(GkM48-b
zgXdv)S0HyFwa_$IP*pV{X?`pLBDQy{EWTjgxB@(QUmRqsL&+VK-jI%_sKL5pDBR_3
z=RfBzAGfCS;oiV+_DvIyLEE|o*Ko(zLm`Xszvb-$G;|X8$I)1MGD$3aBs8Cos*#@M
zoi1(=pzI-KX-f*9ja<B#aD&pn$DZtn#~AGvH)&&#q*@6=!;C@dUBxS7GkcPu0!ru3
zC8+Oudo3Ec01<>xfP-hl?LB@>fXFgb6J~ZwX!>38=syRM5|V_2l-j<IvSDRO&*rsr
zQKR;N%14zj%jL;;tl&W&y!<P%uV@{^cvkbe$;$E8kc5MmY(UZ-GTQ?D=DzNw@?~Qr
zSz#AD@%oT(AUjpb8~>6EIQVhVKEa9Xj!Jhv+hI_{1r6kIV5xoqOKOw6>DRLj_nUxy
z4{{!-?eIJfStCo46Pv}^>S;d9Y+gA!WBiuX0D6a@aWWv~PVymVyz8mL{*2)ZQRZ!J
z6M|8F$ay$RzI9X+7K;YCSIZhFfCn3DBa*mGoo*e#-)7aYB?-P*A?;NKUy`)hTGANJ
z|D2tOyqo;;O22~tmnCc3kO3K!iFvj+i{Ogqd5UFJg(Y$jA^{oyGjKrV4bv;VOkCkD
zy3>q#y;(K}(F-=y>({G>S#_Ab#}6cLBr<HB6Zv1%MuN4Gdy!|)Yer0*1--6TL+6F0
z>IG!~TXE>YfX2Tsdn1K#yV=ga+zLkz%4ea7yuhXvhi?K2X3m5y!^OqL;2f2SjU-RU
z#xEQL$KA@tqNL(%hpI0H1TNW)8>=^a8KbK=(`{O>oDtP_V!B4}Wk-PR63f%U^sAW9
zhmm2C!@rI1LjxYkI3}||u>>H!l{J(B$LlvAPj%a0`?hQX_JhyRxwGG{p;&{_aH+F0
zp2T055g@o`;uc^UkZk1I_w=rX_s7zuNx<ECaFMern@;hT5O5^GRPz!@{!h6Ly_`9t
z0)S?jQ0w!61rht{VyXSInVqR3A6$vQN8*R;(YL1l*XvT-GrNhpCfljNsoTbrN!j!?
z){P(aV)y*NlBlVv`SE7|^s|*!zXWnU{}g3}J4z`p=gbE>so*2bHYJOpuP--<Lp`zS
z?SmkW3Y4b@TO~srhX{9K!=$yy)J7<$ug+3>hrR7&!JArBrVObd0+Mk2zDMP(gUcth
z)(OvowgM+u$LZ%)Gdqb-OE67nkmZFcCyYg>K6x&ri^k(GZdKW4x6BnAxyQ1i*!7?I
zvQvIhAo#uhCti(TO*!smkU5-^h`vQ&3A)xeSnC(&xA{d(BwW^l5RNJEgizTv?Dsvl
zGtiY_XG+rB)kj+=*g1_8&I0@=hVPj0+ayk7gZlsJ-J`qfc)GK_%|q7nb`CyRL5s&l
z=CGmqR8k7)|I7t2M#9&KnU=Q!p)}qF2}vzCnSvL{kO+JVg=e=npx)YN_h1WY@14T@
zTtGcOa=c?*k|?c1FY0poWl-y&ziNBO@Ga?jE1mnJ489+;Ac)-rP;W@2ZdE^fyy}Rt
zp!Pc(OQJ)raHa|)o~=7V118AsO|+orxFSm*1W(lLK?RA#SplZHH<_$gcIbELgSC7^
zvg;3=mkD;Z$evF_l8^?UdWUv6?OmIRJ(=G=?x8iUGyTsIKl_Qt+NTnpTnlK(BZM!F
zeD-TTIQ7QR*NzGLP#$X>Aoz)M716uI9lG8GV@5lU2ov9SRcsnU#d}{HUVY!0cHO(2
zi9M|3O<>kXw!Rw9@M-t|U|cIt9kAEeL+i6sv1x4iW1_!Yf{I*A(*W%jI}RS`P{?#f
z$H;LzUBR^&XNA}?&AW{AOFU&7TA|Q)?AsXN@i{G-yqUd?^4B0;e>QUOtzsZ?$9U>k
zWU^-D$7+AOGIqIgUmR%n<ZbhMG>;`?zdG&zA?v;4sf_>s|LdH?vG?9YQ8p#pF+yY{
z$=-WJL<ix>%nljJ$Oze+GLA#YUL71VGcu1^5q_8N=l%YCf4AT7ul_4e*L9uO^?2SN
z_vgLo7fJg(xB~N%yvV88XS^4pzr#(f{=P6k|KUxL#<Z@Y1paY`+jg`NalO=6JgG@!
z(Vt=N!x<UA->(<_5q!3j)+w3!fN{9h5@1diDc9TXosxsy%6aJF;Q@v?OgT@HLMOwY
zNUNdaZ`7cBRrB+X9cMrG1`85G%ztTTK2SHRG&lKkf+ZcP^*>x=XeY#hXIqK@2DhV)
zv9t#^;IIQOH{-K2<CNRb0+=5x6oCBj;vIctkm(n{itiI<+j3FoQfm>>QRbc-zmF#L
zm=dH!cK5P{{p=8l4U=r6Cyj6Z)4Fi$Iw+*9^e}3r#*ZsA<0$UVS+65)y0cI}DOP$D
zn|vyteC-;x`V&KJ&a;m|J{aZSh~-VD1#-2HLNrfM;D{BGA)v%DH~wZs+uVv-_^-0l
zJS|AOua_nkQuC>B+rZF*Z}-xI8CFm8O|Kn)i*t}H@<UUf*Z!nArOguHmvjg6G;4vU
z1s7mZ@%pnx_PfweeGjw_h-p2Nj^F+y1DoX6E%Q2n@}>O}`!50HPrn|^=aGu_;zO+w
z{IqhTRGVB6StR9vRZZ8>$C{_gfB){>tPpi1xuZ#mk&*afa90XrIW6*F4M5usP7^mX
zP58kL)u4!yo#Yy<#Mq|)DQBbV`803wIVXE8bK#&%8{ZG^5BGG>Sig^~Jt%e<`nn5(
zBWc&~4jO+^-$9F?kVv!|8B35Y+DuCw-nS)9eht8H+L5Ro$p#VE^(Rl;dh5&q*X}Ey
zZZY5T&8cQN4CTnYyNSu+8!i##o63{V^r<0sE0CqFsj-DQzUJdB|BAK^RXG!1TjpbF
z8>rAj(;(<`wAvtFugWy{Atzv`WeUmU5o=p_DLEov2j1(a*rs-ekhUc<sNwSW$K<xe
z>3lZuZyd&CLDyxuAHuwZ%Ts<*po_qk+aSVxPwa^`do{grguvxt?xlJA-yu)jZ`<tP
z(_d@xkxxu(u<38<6Yp@(WYHw0byI_wiQZD&w9;-8y|Xrnbxf~y4Js9NKX3J8bV#c0
zg-W(B*2Hm2=bOk}y24$*OFBJyo~zLv*8g$xw+uBcxt;oyp#L3YnqR%R7b3@cvcl3E
z#OPz<ypcWN_sR9O>xX_ZQ6-`X06~(Zzvw_<h(IOdrH>S$u7D|u9!=ICPE95%j3N~i
zbSKi|FBuv&j!<UKBh9;6<`8P!vMVAyVV9{J>VZTamqkzSzZ-tvN%;=`C#fOT576RW
z5kEYIrj9agS`;k&y1u@_iznO@L6NP0lW@xoXycu7$g~Ny&~6w6N$Ad-XcINLk#j`l
zRmlGF8}=5w3Cz}y1B>HIKh+06yfIipN{wMW#2Jh$nLe#9e>m{s`I>{KLqvSaC1m3H
z@pJ9BX+@q0e$+VHSnXZ?u5JZ0^F6L_2HL7eSr-`AUrPFvixclgvMO3~ThLGKlL?z2
z-h~1Zq7AN3xH&V;iF)&~QIy$&5^WMMCc}hl?o{8Z^jQ4cL%2z`+0RxlnN;LEu7NAs
zEu4=M&SYDw5Kp<o9GT&{mVK6tQ``aHu(;M<=QM6M|0%69$l~{{6nOM8zDY6{N3`7_
z|F)q)Ew|5op!Gpzf_B|W>@mV(Iwqixvh`L<os5kTa-*gZNZ4_`i#^y-E3dBn-0(n*
zvxiV~^48#5T!g<9SR|;4$aKr(tzCk$XOiJ;C_eB6#QRsbj0YPswMGf_HVaKuU~_M#
zjx;p(bzB~cwrt|uR>z2Tn=Ya-hrg~U(oUhpiN;$JCAH~K1WrbI3taa`wbj%Otp|>N
zP3#1M{i=UEo*|Nn_4D|+^ARx4wz{vQ6CbpGH{CsV``Z&Clu0uY3r}&|771b8m($I%
zp~F}is;^3wNlOgTQ)c0@P-v{G>J-sWMUa%ART_2Qym%Vq@->D5{u9;w0rl(%qE{>Y
zVcx{xq35jUgTZ5vjcXh>I{-(_^#OMYrG7i;U6_&T?`(AxA&+e6RM4t59h_!rM}&d@
z?I#5dfJg5vI-YxEj6z~n<R=NLt@T%2|JpZntg`^O7I7W5=rKUk?LYq`R6w2pMJV5b
z;zrRRR3vimMnmjJmubh+HRU5pUbyEQFg`Ejt75|EEuYgy6W9aR0R88qoR1H%z-#4l
z`yL!LVDaX?4uC<Xf(oHA66(MXs|M(;BhXtnZ5N&tzG{1mQXo5)x4xOOCnXT+WN6vj
zY&hG=8y76B^|45O*3ldx(VA<-gqke`-?5sR%#v$+QLJ?&^V+F|&jAf}K3lA!FT}*O
zJ_M-}UmEjEOsoFVpXH+mM3b2x;Z}sD@`B8b-mvsQF!U%GpXtn^rFwwY8vii~Wb`v7
z9|EtRMM1Alo@FLLzI=D#nCpG-MRh2q-}~TigAB^*b=<iB2l?LyOm<5tHP80MV8|ci
zq}=0#fF=ezX;pB35B{UNEDh$o+(jf!s?25cVUXjnn;n+#9Hj{97veM8!DN&^tKgKp
z<R(51eZo@xIxxSS!y~(a=;aE3i_SW#?kKdOQ(jCekIsz|m?=H^Rb|&37}vPLv`P|H
zBuwzsUT8Y126l;S5aOoH65~{H%A74RBA>Gxn&3QVH)Uz~TQocxR$;P3J#i4nQl;n~
z%7K4n=V7aH!FtRJB)2dAYBn8!*s&0ucdIk1mpQ14^_uoNPG|e}%vl%Ir*#dZrCO=z
zaEiW?@Ydbwolh3C8HRLs7d5zQ?zq8xi;^V@TFoj?FPa?oCG}XtN$-H{-&IqM;wN6t
zfIOy=0i1K8X)iR@@Ew=uRbGCFYST(*xeWYt+~se7GNbE;7>;f1T6a!qsi{O;z)^jz
z*V<#q86GS9XEDqp7)$h1fXMv4goD=7;98j|m5e(A8YaAYehhL!UDme55}{r7O*L&R
zF%iYA)S?g6A*okb!fKK#Wz8O&fz?SN6|!{)eDeh3u<5PcGceh5Tl}?F|DSjtBnPgZ
z&Z$Tqd!LgDT{#`jpreO!PiW%%Ik_%V{~24+r?iNSy$FKA!*4=$Fk;HwFitsy&@e0s
zpMQTtTn~Z<ssRCU;(T9NBk>B%yj;tP!Wh_kM=JX7CD8^10<elUKJRP(q!q?nR<F28
zxB*gXS{C(lDl`zMm9k}QZp+TiOt93ztjWwrEr_>;QN!m|P|E$V>v$Y1?e%Ukl)7C*
zML9OAiDhTyTrCSJD-K6dEQYm1Cs1t2&M`_af<T&D;>pe%03~k4$tM3~0K<B0i7>P7
zpe>D8AIu`F5|V$5Z|$l?gFmzf@a!lo5G4xVs((u0KITeJ3o${t+Gx6$hzZfzkWo_|
zSL)H5Tt1#M<G~V)5R4GpKBW<h_hoqS2><Q<GXgI=4~g7zx?K*MM4Zpp&*k5JiCcTf
zs?5v2!syYxl3!0fCXQb-U0ZH(8D8_%adiIy;5VSC$m4rTwLdHZJge96k@gjYZ9C$B
z<lFNf<XnDFbeO5r(>`2&@_QR9unqQE)09)F+e$7zx@&Kefw1V;9Dl1n7$*0srXd>Z
zt%Cv4bdUlo1}~2{;%`p_`CX=yP&!I~zv<b7hO-FCAU8WO5&k`3bur#}1DF@@v29~O
zB%J>u9@yEb=xrerZS{&Rr^pe;a^}24HfQ?`%Q#jIxs<)&aQ0C_)?DYA`-(|G$twQc
z+XdkpuCuKj2uo#?@T&60e{lb{{;NkJZI(wJ&l&ia=}YrUEKa8sUOp>kp@-dGatTjo
znpbDZV-sC%eY>UQq$IdVMGbiJEIw&AQOc1*5dxf|zu$KWQcQ=hQ(0Qqhw@W;kx(_0
z3Qu(aCAa#eDdi{eA{aV~X*D^t$YJC78H(B_LjVQMLtj|8>;m4@mLy4lI&i5{CD0E)
zCu>c?oa3O;j7aKdoMRzsN-EQY^h@%`ORwouUsi1FMWVsn#4omC=$!_noJ_@WW{c8O
zDMxgePHF)6)$Z*&gY=K&?}w)^>DpsoT%Ru#6)j88A-cS|lc>Nq;H2Pm94ULy$Ln!<
zAH&;py?K<ak74hUP#)5M#6XENL)}-Qx~b$mnunR0iVa!=?jY}ZXi7Q`CLIpT=_#I$
zB|eofhxRYcJYlC~Ar+O&n*e96<m8YIl|SM0pQ^vLZ}vzbt?3TFTyGWl8}b#eX{R(v
z2J)R0%J)=4GSf!~s6Ex&J!7y;39aHDE_=%Vg>p=;?$o$G>))9pFV!7~-Z`*^7CgFG
z_-#Kr*82@*9MKjBTbX`DKK&H~H(II)YnlNIhi%PW#l$<^szj)J(d<hv&4~<Ya!93l
zCli>0ThW6n8e9zWt0+g5$Rd}`TRs|Y^28UnRiq)-2*~V!7NMFda3|q3;H{uyN4tbb
zx}g5PBvfsUEs<?RbfL=)8Rji>)c@0D)x7uJ_whgDWNQ&>euK6gK7-<G9|-=<7|E{&
zz*^>W9{xBzQ<ribRszF<h`K7$Kd?BN)znAiejcP=*Se0up|sU(aojl;V^0k-Lb2ET
zt;rzn%{Am7f?iYiXL+{Ipy_i=q7^STG?Z%9pht<9@Fe|LLKVc;@KIZ?hP#3iy->pY
zsro_);heDTs>bVib7mika{{z<6^;#IF-VQ4#N~`+dp<ct$-(ZQGU&LklgKt21^#EO
z<S_)MY`7aopx3}GnpVHEq2c@R62yD%TXFwlaKi~khnR1VK*$+j$wfvy%dwN6QlY9j
ze%;$?B?QJcR{s(J%jpM~E$caTEBeSKTF2MDbB%fNkQV)a?Mvc!15ezwyAd1wqcxvC
zebTqxqUE%@>e;AWqnQEEA%*$Zrs>4!7b&?3n$MP*Is~pDig}4>(rH@`dZ+Ar(-dd5
z9+<VpudKX(F;`ZA<3kQl#UB0l^~>4g%U@%6{HLee!_ayXpJ{vK{I;htCjWDxh)k_r
z-Kb?uV!Pwoq76fzX7u0bOIRwBaNt-iZNXHbQBbPOX|J3=_vLu<*Kpa*l7gRfxqrgX
zhuwE@hTy01eE*X5cs`}V*na(b3*p(eM;}>ITp=KR()eYo@{48DB`ZezP+`wsZ)sMp
zI6j^wQ3J&>pNKa9??0af`pl+`DnQCa{R;Ky&%ONT<l^cou^0Y-s6LfyaI(#)b(L?h
zG36-R(_ep(+*Yqbb3vSB=UH0;UgeE(#IaxL*!H%MUx(12qKOo%ce$djT#lph`evB0
z{pi_@Mn>)^V$Y8YSN#Wwnu=wpSMKR#@gQ4eeSOsrjsG6rJk`Kk#;js(INR;gRL@Ak
z((4bCze+SxW*+3-NQ<r#7NHB+*N~4GeV_Iy<<+GIypnBr;S0YyEA!VoGttv~Z`ckj
zel%9`$_-vW^w*|Wmcm+Ff$mB3#s|BtXHWlrFG}xE_V?Mb3@4iW210MS3On_f%nCj5
zcX*HT=DYvrq^aXa@r&84?_eKLgvsLhn{&IA|GaK*{QygYl@$it6wi;pKbhsc<IXwU
zG0uft3}rNJ{5>DYq}9?$xKCFnb9TT0OV=4&OKr-P+tD@-w)-8al;BzaucsCZCn2%7
z$UJuFp!n*`5&m{vcb7_Xv$D$5hRgYrnRdkL*(Mu_c1f?qso!toK!%8)d^SCnC?-tI
z_89Trr<m0HpUTWWAIaU<mpHvYPVZf?R+40T95NA*qrgF=&RYw<JzOLkN61c?wm%ZI
zQ_`#@o3MOz!)+$+%Tcq!g2Qj<gzS;DW3Rp8ZZgQl+;*KY)E#gRigWXAoH$iiUv=HF
zsxN=QI$;v#U9Zh(=dNS*tzO=EHnYRNRLFlx#w6f-72kJ#y<TDVXMZ-ty)RyL@bc$A
zPtKmj{X{$+h*_YJ2uul058(?CJW`h!{jaH^QS>S^dwb1?%SIwvVWn^GcdXSv6~6^L
zs=!Q3u0?V5hpFc_W-Nf}&5gXh<r56$t3|?`Yl!Gt44K>}MM_Cx2eFOZDTzB^ppW4!
zRPmQ6gDP(5IX!9W?(g50H8I&A{Gf$WZ3#BWvFm4B9Q8)SHvz-$fri!`m{B+)T>RGu
zW%4|z@5hGS_w$cb{VQVKlgA_~(=_23A9^SCF3%pT3GoV_v(S@CX<31wk)w$k)+CSI
z*OnI`ohl7MI!tcDBa>VhtCdYo)mZ@TkctkT374QkVQw<+qL2R`8>@yGm};~nFaQ&U
zs|^7xj8)a`Z<eGgS8NskSbp^N{pFkNbChV*_sPAOL$5qxAjS}j+o(F<p42i35XDEf
zKTb<GZM4{Q$|;__VgJ-^e3#>soy*SPY%@L$IuD?vEi-p-&nA||e0z^gopC_Q-x^3^
zDI+3oxRvsVE!D(qQjc+-#?(>StiRv7tjIJ+e3Xcr1|3t#I~^kncgy^`b2IDm;p+^c
zP0g)mrOJHn?2h;yfw-dFgf^bHV<y6H#7)~vft+fp+;W+<j5~-%YXyg6v8E%NiiC8Z
zRFZ+(i|QA|9Y$^on3FE@qB#wJtf|)2J<}?RgnJa;4|$~k_cu!!-k@Z!F##q==#w>S
zxA{?M=V6nB3|3G6yYUzDG`H!nxL%qDLVfD9g9p`&p&=Zmk$(#kr?c<$iR)eS^Tf2{
zo~$Jw4>H5cb|hpKv20uPeRtbGjk}{|tKEk<JZgV_HSu1#+rhUZEX;_}FbD-()?i>B
zu_GAO%pab$SI^$preG^v%{BLx$x5hcKEQ2h`{T_)`@lY)s6OzO)d7uya9;eU0eLI}
zieIMWZXpnL?b0KjdMwWuzn;wm?xr1HE;QM%))tKyiC*k~=vp1$s!ZRP^lMCik%x(S
zjs>ViPbyArDWxI1_*Y6sR@GA*19H_d=`SpoZ>pN~<4K3KW%Ws%LthNl+V?E7i8<V3
zMUL8u{?$%)DGNI|E-jyYtVZq)5aH_~0K^C-2P9MOh>nuz6{hK0>vZD^L{SW<7(wPy
z;tkXk<@&D|?ZsyKML;<1{mmEskh>dlUPMT^7w`%)vk^jUV&oc9LIgA!&FVO})f1|*
zn3~vFs!jW+CWy{{FIWgmYLhw_0)3=}vP5W}%__a+wn<)?9Rv0TccM2c;}!?k0^4&+
zW~H5RD-`wlJ;Yu2_#YcY1kS?e-=%(u;tBJ>LG)zen?F38qZc<OcL9z?%0{QB>6`L~
zZePM0lqfRGhpzh)A<E}AF)OZx35jlC>p$!jf+>&PEh5JzidC2jX;FIa)Sh{FEu@6=
zQKUH3bVDjnq|%q~5Teii@x*2b>vqmdCm!j#QYepzel{Vx$g-GvVG!Js##ONjl63DS
zQi+7dtWY;colyx-VzWF#!9r~?w88ZnjoFJqh9FlXGjTth<8}ev!>p_pkNYh7=iPpX
z&vj*<l>M%=1Tv8|4bGRwOw+5GYkn+QCAZCNe6Whuw^0rbpA!zH;MRP0UmJM^?HYS1
zOkHQRvU#yLeFk3ISq+P;Tboy$(9Kqk-ysfn)TKz<Y>PgKqselZ#MX~Yn+&kSPBxU>
z!uKb*dMM`#L+Z7&5pGxvbweQWAjYlSKgTx*)_**8wsK|JqY<6*bqx|o7k7>ANhp**
z=(zl*y72eYb4P>m?C<N!TNo*&zbjVuKXcCd(3`ii)_)LTH+&e%BVK5Mh;g<<8~N(h
z1J!pF1nH^d@u;gB`@z4%5i>YmKRkCA98uJQ3H<<sh~3)q$9p@9|FEtt<{v<$e*Pf@
zcxr(Ym=bmu;}&;c8aNm|p3>TE;lG|cq&ZX~E`j&4O4<JcW~~-Kz&bm^2WJ2;xB-e)
zlr>DPJ{?N5e>T8w4EWdMF_iLXx!Cl58ISv6+K8B*=fxIe$_xm;BLXs%c9W2{g99;I
z#VSvLEUD9u*i*KDKVMG<#fkzJp)X-jh#R0)PZYhvtPdy1^F&p=WvQ^Lf(ZdZC`cY`
zlmZ3f21QiJT>y)cLuD~?2%?LwXQhbjP#puZ+?ud_1~?e~A-EQ4siYB!!D2=)7eQ$$
zXHXsqf}Bgf2mT(D73-6i>@8rXw|KKR<CmSA2*7%jA_GG&eq(&q<crYWD~`uANAP66
zELLJaO#XCPCQO7%Qygb%UBW2BT>q#|7F*Q2A$U)f+t3xvCq=ECD>~{y)68bTfPfX-
z&8Ab?1_+k+7??-3AuS-Nl9YyT$WVhpB@bKC9yKdzN6O85N(3*S(NmS1&MT(D$!`3h
zoNT+x)HYk~xd2*dE%h^AitW*3mmJg_%pH1rKjOlc>QvW%v(qi^IM#{VEmk(GE;NeM
z?&|;rns9PiF$c2Ui7OTg;$yEjN^B_U6vWm9Pk+{SHfcrds}u_9Q%N&oBdRrwcAyz-
ztg-#KDz~Ij!SCHh+r=s2V;|ij{THdK40?zO>{?MCSV6b@o^=htpssV)61xxAihY7i
zuMhb~7h^e*icRRUeB#=EN$(*47JZf4b|e$$&ZzX^AVfdGAMUI>x<IwL%thoL+|bsb
zr*tdPD)=>Dy6D>n65pzeu06bdKCEDMGW&vy1zVN-^7)bVOo*n^){0z_ZK5qWI{I%!
zfjb3z=En(Jn}3hjb{;8(R1SVwacnvJ>{j@t?$IZWgtSK^fAdp!KUa}6OQ?PPe#73#
zClJs>zR`Wbihv>Z?>@9DIXu;+qV75*l74lUN*_+$GoX|si!KO|Pg)i)73h)_@(?{k
zFkrS(GWC#DK!5)o2S%Zv#(+3Ak17_ASpI6Y2Vq~#Zc>^3xf-J4JJ)}ZLz!Wm`OU@R
zsOY%`QlOe}1CqK~+<QbLt#ZlM@Gs98@CAb781%u2&j{c^IiXOgl<&`&xe||TJzlgn
zQeiNG34X*`U@j*~{2gXX`G`$SY7u@sA(E7=ho(e?1w-w$j?$r3V*R#zW3RQ?_*clR
zO%e??u3<&ppRQp2K#2REd^hT*e+5xrL@}q|61y){jD;*Gte!$&dHTIco~3*~AdO6-
z_<jJ;!ydn7So_s|j%Lccz9TtxL<kVQ1ZCaMm(DG>onXiT5H~`Sk8|}{m9E(6qU4Wa
z{i-2{!?M=#;(LhldK~AM2`V&;WeBM3lSMF|175b5z@0SaZE>K;_Ng;nH2n5NBJkPe
zN9QJ4D~1{Np<K*hGb+c141L{0%x_Kc2m?`k#!(|BP1-4x+s1$~Q1kECcWqm3_@Q!a
z#A(Cde;B`F%P`gT-YffzMbM5Y+<?NUzQ(zw8s0N3v+Fd()BlLdC&t4h<S6W{B<fCh
zwJ?8%=o(~^nB4yP<URl0BBgRWELYI--yXBQ&cKVCPcC;!Tflny^|z!d>1PyNJW5}d
zwK5g$_-Aao_0OVSy)Xce)>RJS6uzK}%Q#wT>EeIbFL2uIww3xn-{wQTmuIcrpS=x2
zdUX!wJcFOE3!YM?UhK)}-(Wwy`HNKoV_67!`m%j;oF|aHe3b)tvVP6JGNfDuyU1f`
z`BZcQDFFI2>viTymS%ks@onWDbIG^$WUVzmmi1Z=s#;6(G2YRuep_1ZF9?xjY%C_9
z#^)#8r3cCyQp)3OKX}xSz`~no#o`Z9usmX$SLS^33^kq%HUS9nO)4JFyy*6&A8-Kr
zk9@ox5OXFY?&cDMCJL+hFB&E`PCb}*te<P0W4^msrn93}HjPr;tnlNc?rZzN#L0<5
zsyO)00c053P=DfPVkGoWxTOPHrfq1Opahyh4b|j8i6uaEP|W}yNP9K(8&X(5#AyrM
z+$2@N;HiHWM}N*_Ly*|5$$f}KGi*XQp~cmrD+QOSA1rsf98o32V(N*%<PMF=KQ$-r
zO-sr*y4`aCjeAvFKT5#>a#rfFmo7f2o0{h}=*-Q$a(cOLZDK2m--B|`rP)$Yz%qz3
z4-EVgbP38r1+F?*MOGcbCqMrC9PswI65a*$9zXEOdU>|2y=ujj?HSnfSgk<9XQb`t
z7p3!-)Oc36525bElzp~z?yc|ej!U8hw^Ceb;fa6}67`Jp58+X#6TIp+g2g5l#kM#F
zQopTDcxTK2T#t2_PO)#m*Ml%HyQbIlbwosMI@7Uf$;bBR=`Ps;-*Cl==Jlt?d^^kc
z;6lsHnwm0s22b=h#^{i`OAyDP30ay2i1te|)I?d%J%l)h;V-E^Olp6VDrs})O?b=*
z(^Q?J3r$(zqpun4MlFqZ_Sp{yiE^?-dvijq6?!Gnl3kL+hj6iVF3P8u>k8fr3!ZPb
zDqC>7%2!6kpFcf=%^Zj4Of~~w+1~<Wat^c49<v}r<Ze3+vujIrS~^CP=Ui2>t{ngE
zU;aZgg03O(J=A=!882%yQb0(?$egs|aBu~9zfR$2b2sJEyrsw`vP|*egDq0czgJUa
z3VZ#+ZSw3P(XGR|12bwAp8EkC_t51cp-Z4>+MDx+Sry$-XZgi}6qCcsZpquk9u%~p
z`uy`X+Z<6EMMf?plGICcd=V_n71igop$Gr4LYsY>yI-9gmO6G*djg*})jE&v1`b3g
zW^?;ayht-IREd1jQ@&gEfQ@&;wT%48!==KcP93C!k=kO^o2Oe<O+NqMTk{iG7cmxT
z`AORXMhJx<NM${@nr<Z$xZ6^!cJ0()2b94y$JiX78j{zXz50m^_2%3Yo6XWm+-Ikt
zlUf|Q-zUyM(=#MK|8a6aF$Ni&{3`sx7kwF(EQp7F&UFZU72?8Yk1mK;qeMwznq9K=
zJ>_lgk0Uhi^v0&}nUP3Hru<R6408(Xs#&+KcCtwHA5SQYnL!dy-S{GG$9?-<_0D6Q
z6c3j7NPab4;j(<tp=zMx-%aLb*Ql_0e*X7<@7R62mn1~LyL?!-Yk1jAwVY;jBjpWv
z7@JaYR=vt&4^ndfspXG3FU9>QBLLN3@4#=u<Al<pqOaC-f7f$87c}Boujig`JFS8?
z{<lkAAX1dRwPWNtGSDven3;6Ycebu_k&uuX6DTOZCLk2{f{`Z)&4<aLvFU$LWc-lp
zGNpgL;(WQ<DUbG;x>4tVWD}D2#yL#kq)$YPaq)ZRu<!4+`XFw`WAB6@D9!PbiMn@8
z<n!XjX}dro@3{)5nXIIrH^O}+gfMsROGEWx$utHOd8XrUr1oxOD)$vv<WX^>o(OtQ
zCzN=P?l^g}7j-0sx$1P>0bgmJB{yevTo54(-I;PSbJ&Fh|MYvi{*;VR_s`0&Vz~zm
zFQ-n~hM9wYMz;~y5J}~+^X#Cd(Uuu%VeKg-=?cs>Ogx)_BCea#{;5t$P|f-Fn$fQg
z*NA_=ae@v*=a==3qxGtZNze9_wCvX`B|hGA<n*R1TZt$$v*dgB2P|V7{P}pEDW&tY
zbzePGF(}~w|0(|h*`ri&x;`Owzc<znY;3;0YUG_<`-M9_V%7gT63I!l@M;gpx78;s
zgy=q_TE96Ujv)0?VWBkx3K91J7H-ntFyH1Ums1iJ&<7=n3a0=+M~g_9vGST>>-&SX
zn}qKn(60mEl4cwN36ee1M%7v?Fy*atR3%Fp<hf|Z1Qy1#;5WqK1Z-U^;-z|y<a}d9
z7!HV4^r)sa1&h&EL$EtYD)EkL5ovY+W?h1U<lK;Vn%l}!#;sqN`GNR-Sb8!}=~RVd
zZs2zYR`q~pJIc|mL-WR~-;CA++;8S%Ed*fyNPi2GxAq)>WqPfFH*Yg!m>CoMX*uW`
zZ>$I>d7^Nsf$4=35)C(=`h*SPb-Q%=WNXOV19F5qy1pk2(vYZrED4gij5+ZL&vANu
zyHGOiNVV*?nC3a4j8TdzQ4pw;&%~~jDAzD5=q*#L3EUbbmxdG9{S*oiUE%of+g9+f
z`O#TKAlgP1&Fw~yd@T&b+0oh?h%~fN{EnK`PAhS6(J<Beo&XM?uep%TKP%^v*!Sjg
zgE#m6?s<+)E?)FcpWOE?f0~>wav3ofvR<}SRzRt!aw31={CE1HM*o`vxJlYVlQ!5+
zyZm_$Hl99rQ*Nv)UADaZJ7Cr5^ndr!cvxJ0muNh{Dn?y3Y5ji{rSs*E41ha3hO+Fy
z#%KiXY!h+=^W>-P=;Ly{ymbT{ULSfJkiq)f@86@^uU}?Qit=jp23t!uN<y2EYrrQJ
zX?6d~M|W}%=u8}`3akz*ilVlF1^~XzMA9&GlOhM&Rb}jBZq{?xY99x91xnzEr36Z_
z%qbsBtj)(Og0B#NdnZHv%tQ7PVC6P?B<17*?P6_^$Cp}cxnUJ4#}Bq?BAw{;;hL&e
zrEHq1JDYFDUYA4&K|A_x4Ip2su_|lw&Rf;jK{JZZ!3kAA=A=F}9n9tYecw<pdUc6&
z>c-Jcch6go)zS{QY6sazUZ1N8=`4$lYo62Qf_TI!;eXy5JN^9qhVnS&))#DI^_O?(
zHP|fA<{Egddt@n0NJmIV1qs0VB#<W6M2qo%Kc;~_sqRyR$y)dWG#?$05FIxmu(F=3
z7)7FjCoj0!L#4kW>-WL)=*_8f%HKSN1QWrdqS1Q;MsdQm0XYx4j#y@=Hr{<84+(3?
zaoo8JZssTpdw>>lz!Bxt1RK|tTdchnAJsT<wA*<z1n!IxB=c}t!2elM#OlU#!j4e&
z)35G?{y@2<@{@@M6<gwLc*Hf1(w)LD(%z=J`b$8FSYkunkc_PFzX#&RS$_2wu7Xe3
z!hc&<m*aJz2@1Q9yuqa!-_EGy%DvWh#&8X8tCG2Y<K_akDk3D6Z%lWb?|4dp&Rz`5
ze{yYbF+>B%;cj^_QSX8RzlJNEt?~C5INSO8icBO4Oy~{3n4h$$vDNqwz+w1uP+Jz9
zgyHFuqwdm*o^DyrKSpkV{So@*dt4h~QIsr(Gvqn?@(5jKOX3GWCEu6N6!v%nDm6@9
zl<ZL8eK~{z=q~RhLqUII)uAP|w<bz|A2YwqvjUnOO5^<FBB(>X>t~3$V@IX3#)=)>
z^T#)#4@MQmD8*pZ(0VaWYcLrs_BV@<(eM~V(V|Tvu!%3|cp@e(D$eWL+Ul#0h(@j{
z?rUB9N0SG`iy?|v(5jzq89dD+rT8v^jG;-6ZmZ_4Cdh&{0qXy5N1#adgf@wGfl2&!
zzK*p;ELKmNFDZH-mi>pIn|#ubX;b<aSs<!HKX8eox8W@={SCM-g2B=j`DjdvJHa1#
z=ezk*Vg!&gB1S}SO(zG_HImGIP(V?lDXZ<z3QEsiVp}@P66_S}|3VFH$y)c1^l4Ey
zIF$TgLgGDlrp-nm|LUWU6I-C8kcnzSOO&c$fn$@bd-)+LvWUVx(E?%`M}%ZUeYW07
zvS6;hw*Yx-zMj$0DN8mb>lQsa3Rpi|J7x)zvH15deCij|%0g1i-Vgd35#}>2cWdCj
zrX1LT`+4Q-GZ*1sJ$0uJ2m$%W8OsdM^y&f8kiXi0$J5&lrRg@V9K=^G*8kQ1fv*g1
z;L4<_*;}<GwflcSIoiFz(Cvkej+Z}qx!zorV}k;oRZBxu-(bj)`vZ#W%J3@|yAX)k
zC$i3?Pa>_h&g$9X7jwNQ>vp_&e!EH7At}r+xuvy}a&dC=!LWQd3ZcV-p}Cj8T2!k(
zvT~oGB;OLpL&y6e9e8_><L3R;w<g)hK#KT#<kSisjTUgXARrB_K`PTCh!6ySzA^(E
zft2NM8;AeNQp5Wo%u7Y#{PZZoW_UPe!s+GXIqs2HC-o0XD(jy?h(m9MD@izVlk1a9
zV=^|}lq#}nY&tsMz9&=RgUBKop2e-xFXc}M=3{k;dGqDDqZQ-8w~sP!Q>LW&eQPEZ
zj=e$HN4`PI=J&W2z`T+-d{Ap(GMOWRih41BIJ5&-3N<7<l9y6LSB%oRw%{E2R;KJ0
zububi*dXmUXo$@DT2tEfsQ_|M{+@)U*}>rNemVM&wqbd;<qZK=b(IV%Jdj{FU-@#P
z))jnfzgr;k8I9t=OX!yD^MAgwRh{qsW_?dp`gjr4W(Hh&!czq&lYtweKCh>^yk5&V
zRqkKdwxUc=yICFMZS1)DfVZDra%Dp;Ob5H??5H;FD#v=PGqx(mu7+w{^lL*|PM3Z9
zPo6m4X8_yb(pMU%50hz3=h92zbV4uhOo9zLBANYeI<kfTp0vTEGq~8Qvgz^FA-qj&
z`94&dS<$cj-(S&?w@Y%{&M1f|znEHixtdX7EgT1MEPdQVo-fW2KJWYzlcdaT28==X
z$|!}IfLe*niFda>dGg?Arb|mOde|ILMBLQ1@_O_hVluXZW9_w?|8U(p(}i4kg-+#b
z?Xt#R%SYR85l)7xw|Kj5+2u!J`Q>LptBTjQ68u5e7qU@8vCi1nUIy0e2qH^pcfMX6
z*CMnW$_>Ff)9Di#brz5`!;(laJg_R^Q{p`6+9)G>3^2K#S2ZBtgbJ)5X{*=kx$i#E
zhFpI}k*Ds=i;mm33t@#vvi=*_I~|0dgGTa`3tKJ`>owOM7|KOQ$`j1CMbS5W;0V*a
z6~Jh$mkH~ap@Xj_>mup&4DZsd)L}LZDlJkyVb*t>vYfO-WcwevScKI?q(Kmn%W9d!
zJ4N6Y8q&Sn+a#%l3O!w?fCRLEdM4HvLLC1#y&qYTb-o0+qV-PXc*!SK?lfdy1X#tN
zG+C~*O_Tg1d~ZXyS)hYcU6Q?5F3RnRukhT54gUMw=yA0`&3n71S1Tx}t#HESY9-rC
zu-BDJ6||w6R=u_}9!90;JnyfaCiJEtK~}4>d6|?&!d`Rf)Bin)n_eBn8(%Yhy=q9F
zWDYrAm#|KHaEbfv&~&~a82iZN@Q9>Kf#e6Q5U>DVAa!e+=(7%1j_U{Ii)`Vw?s@_W
zMCSQI;JLUcM<pnH7NRcQ!$es^@Y1a8IKqoCPt5{FzxXYWISWfCVLB~WAxBJi-JYg|
z5CQ%li3%dfkAWRPaxMF#1F{~FN-eU21n|U(60fdUa?_uE+IKl~ci{rO<6X8p6{;f1
zYtz}NiILGyn3vv$Cns}rhpWCPIvu0846Tp0fV_K?Y+#Z(%xp*?3oWSz1$W!Ygh;5!
zkf%0_SNY_#MSrHTGmWhm$!uN;rzPq7rUAr-@61!G6~`$py^t>;U8tPY-ixCw|4v1O
zA&N6EP8`M>>TAp>01q#E>JIq{!_3sAbq-Hz{P-78Q}tvq`{VI1e$qDl2H0a3fKLlD
z`l#j79Ma%sR|5(@gGtRQC_YPxjD&}IUm+A%l=;=Fwtf@@27Z^8|M}s+6v1S_JJ;ep
zZ4-4Lgr={QYX5Bq{GXSjM(3-BJGhl?02D(=&ia(<TIv42_*uM)(qk7L3k?%`PlbY-
zL*9p~6JO^e<MoO;?;R8*87ZPYu@DsBlQ{gh$tb5c%^k}!BHD@ugxuY!o<o`mj)}x7
zkYg}+YW=9eH*-`#p?NitG0@dK9E7Z=X(VMQuzW_@3>PC{T;_ly@|A#l{jl(980ABy
zMN)he{}a_)Zu7>d8>a>pD?Mb15_V;_D#?-8ztxk+6LNaS#nu4iU3mRgY+s{{(6FMN
z#oAxOJOcAK_mXL#A~RlGBg&_-3ZaJK8VO~U^$*@5&ut!IBq2uU_2iM9EQvUUIHJYg
zwck3K+$#dqbOQG+;wS`D#%Rr!OQZSOd&u+b0df^jiA6TMr9>G_8atD=-mLj-%|+YE
zr-26FU*PJxlq*t&QN-N&5|5w%-9D8!ceTCYwOYP@|9?MQ@3z^i*tKy3Y>%XZtl3Iq
z2d#DF`Lf8_kGHiW?Psk2uBp+DzA9$RMHL&nP{)u`!&D$)gOZ=}=Tt=YO<|l3FZ@{E
zL;r=Zmqt;mtFVJ&Nr&3eu%KC4ifc5tNC>VeL%=muLe|v3$B1z68yYQcV7qtDlI?p=
zwdCx7b^F9po1tbW--Y}8*ZgJ+Y=taJ>#-bqP3*C+d{>WN$M>uB6*Y!VF(uk`Xm<C5
z9m7GGtheZt3m3O$#`VCoeU4GMK(|{Wc@kP!aUTJTT(V)FkR{&wY8NkF_k^8Ayrwo(
zT%p(TVj3bqd;pz?Of|Pt=FePGSR3#AgXB;Eh?9mSg94o5ty)bkuqJT;QmSU4cY$B~
zbn!n|ANa==USO^#zaI{2Osw93ir_g*UBQ)SfA>ABQaLDP9%v0{Ir!63HFMMP1=iB^
z_q97AGi#(n9l~X45Be0XkO7ckyc$dZlBoF(kqpOhIW;^tW)`SUkU-qpv|ahfu|>P~
z>z+gu$JVF8N3o;2t^0ktq|K}hi7crw>zcocAJU^GP$gpl3e#B?11BKp{DGhOeawkE
zzbzcU!9VYSl5FiI4sm#iQo$93uuo7~bL0exSAcTD=<5|xe+2OtKFn+09y_O?Jc2*n
zGN82e-@$;vr;*>8PB|_t`usZY3<Z|{_m`djKd4N;E2yOt3b{fVZr{0cU!F7lk*>@=
zzkfgPdwXakF?b7@$h}T9q_NbRzJVaatNH0-q#&LMS&Z8L8uUHQr`JrZz2OfHm;#eC
zU!-%^@v~&glU%aX$7T>Q4{%V!r>}{30>WLvcB%Xip*dIXPZFnbDvZkh<O<zV5fNPh
z2ijzhTD$wv_g<IuEnnLmk6{D<x>bL;)X$~9!-(}8%a)9XAJf(|3UrS1=2A~%8;U%R
zjkdssU40=bxHYqBK~cp7+^=8=?h!Hm5dOapl7OZwWF~eKgpqO`IL&KpBif!*aVeSy
zpM%}s;V5>;OkI*<cE&`Uq9gi_)GE>EioxeNPABB&{o2GTc1wF?(-I*OHyPTz6e2Sm
zMZ%&_+~1r|X!&w}b3!knG-U)YEi}Pa+eudXIxy5<;K*>ROx^(q^ioh!`2e-~d+gux
zFz(=`#5}c-x00ROOR7D-`c%6fj6$l(f-x#OybG!%-J6}tKlpZq8W9=98!pXGhf1xs
z0&e(ybT(G{?K8;#Ww(aR+_J2ga+5qoN30|ItTgT2FT(k#m)H8Kr3UqAt=}eDo_76x
zuM2!=|6$7MsnJUMtK)j(s2!Gq#j2EIQR~Z5Zlge#Y5oEqd1)zep70svcp|zT#De8H
zYJwj7x6Cvs=+!1h#}j(0Zx)U)I85IruJ|Z|iTow!S7~FVP*HV%cG<K!is#tLmp$(F
zs2#32cbi4imLi9!e?^#e+k(0L(b~1SMC!9QgRf*sx_YG2l#OP=6*XkPsS2O=n`zcJ
z4o#DUt^NJrS=eE1d6R2*q|ZDhJN*o=^>sL|Sb4VgPF@I(qmBswPKJw_3?a4iMkwN(
zb4Krtgh>NY(6Xhx1q=Uim8QdUnpEc7PJ3#tIxIvEcxnfo>)+uIs2%Xe4)IHET3lTO
z0^v?o!^uVI3$Lv!mcMC*FiQrqB#Sf6KIF`*?X_*5;<Z`)_te^R&B6WJN1u5Ilh_uf
zs;N^$Y`+-irv2cc+(K!}f*9uKFMgkhAIxN#`7M6CGEM_O?k@0wj?OgeP<?D-@KJ%m
zlPARu|2s#md{G8YQX4ZY{Z|>Twun&l*Sui=4EmQ>4+AQ1W~JwL)OK59LpcHJ0nT@W
z#<mcEK(v8R1#qa=f{;a1Kk82$PhW4CvEIAOZJFOXiH2Ce8r8Rq>_lvT>xq>VgF35b
zb=`tMEVfSER@kW_i@>5u*|{B+R7F2xiCUOSXo;Cv<s;VNG=JE40IqFz8lkX&%mZ#8
zu|xymixHdvkddd?f_X$lOnib8yhDxt70N&ZV`)8trK8ktcNgHw&xvS(Yo;K;n=gu>
z$b;>S@}mjxp%QV40WHJKC{e^~`<%9uSNoCQ1O-&Hj#3o%0qa57H9FyA0z~n|CY24h
z9U30@!i(oGfqri*X@lijbZ|soq->Zy<8=wyhc20>HB2@=vB@qhc~W`8$&Vj=N-#6C
z#}YPL3-)|9E@4mVmg(ur3AQV{MUhnPCu9=cs0^PGdj?M;k&bh+FP37x(d4Y@w8LyY
zWeDiklQ}Qvl&3kwp_tnEz|Mw*$`rSkhmA)6w&NMI>la_7nd^D@eZ7*`W*ne1fzfD4
zI@A-*jir_1xy@>=t~67ZWY@npSA|>4ZGU*z#Lh$`jw}mO*|UXYtJi7Y!*$@Wnki}P
z{x(mmsaDyJ*?n{(e}Kycm5Z}aADeiCD<?;&zsVcmIHeY6mRqXvZaZWnq=DBD$je7R
zmqeYGMyc_{k?HFiqy@AM;|TJ`tWKrW&%VvsnMrWlJ=*OZj$@*FeBwj!u|m^n(ul;&
zN1kk&&AP|bahGxPv-L~f`U?<IFmsdK`~>Z*`aMCE!!$f)m;`Cj4lLNx-T|vkmNg{j
z6ia46dKdxkbU{asUBzKmVe#PbAcIfmMf{)@paOYoZ%Ub?aPo%WUGh$5JC|&Uk%-b6
zx&P6{nUH<RB0BT{%+4&Y%B`=z+G0ijH~l!?z2f>9AA*&QtMxdHwGGI9+na)dDRv1m
zJt~ck!|!h<Ry}<iQfVwKTE3H13VJ1RQ2Rw20_iR`(v3!;OsNb>P|>}U?VmZ7le%<D
zOS!SM%I0~^vJ#??_#`INMR|uL40?R(Ew5uWxxX9msb$ETbkxcB27jI;Ze#OTW4hak
zKWtj3;`OZdIT&u82tR#IomMR2$F*}ntN7uwfl(nIKQQpTf-4$-^v-C;&AP&Wt926h
z;XDr1{_*WYspy@*t=b#-s+?t{gY^hF4In%+dRJJL4FJ&8P{(Q_jyx*qJ8g~q+`|)Y
zID5m#MQMVPV#KLkK4i^+09ck`+0APtq?IpSeX;$-;cON>Fq<}u=GeG31S4Q*f)UOh
zsC<CwS=|}RbPFa1$A@c1i{JD3f;fXlXO#4!?rcly%-V9|tqT&-@=)f`0)$~6)46?4
zs2YM!*!B=*@t0>=5v!8n44uxLA_K!8f1`ajLPJ>3C}R-!eZoaDtBwx=Ie$ATQG8gs
zQNlk^FPrvI5d_0pq)-izs*Ga_OOo|l1Cmsg0Il+y+b*2?rbK=Y?9Z7K1%_d%^@0-l
z5{}gpgvc@3mW1C=7TZ{WLDL;FyG7<a0&Z{vA8+njU5W1m*V#nOFQ$_ZO6B}Ln1CgD
ze9Cd`5HlS(VwQ5qVcPm<<Cz>U#(20m-nh|ay~Df8Lg$MYQY`jK?0NE)BEdV`9~fOa
zk345|<5G7$J6iG=rV_AkgPmds1mhC-yT7lwJ&zmwPHqvV_sdl6jPg@TR2CM;d}H>V
z)~sn>E}**>uKzs8Uw%*8a_F#P)r23Vx@o^#L+QrRtf@ru(5XUnv}R%fc<Fj`ht3+i
zU9kSgVvz+lxu5M%z*mb+`}OXle0BV~{Pd-`amaK))aURaLb{Pdt&Tq&)$PCdxtW=%
z41YAVYU?Vj@zw_Hx`%mz+u+`UP^<3MlCQ@9yG(RX@d`4ex!Pw1%z>=C^#bUc<W>UF
z?w??VvT2Waj|TG%=}+_tLWUe8Y+U?BKpczr6ACBt%X$IoMjfsf!HES_|6T+P4oHc1
zf(#1zI?Ym4xmW8Gz_t~iIL;J|B460=R}P)B$ZS+=c6{-{10xbvNCWOYylW|mRD_%(
zGl~3`9KL2x>Trb;As8zYN@HBivf)5qzE&0T15q$gxp5;1An@`44J>IlCZy6n)RcU0
zlMQRq?UpYBd{T+nT@OQ>9^G=o6N{-WB%3DGSi?5EZ)_Jl`FJiN+HH6PUtSi++FUzW
zhVz6en>ta8WZdn8I-`f<91!%0_Yz>nY73zp^$_lSp+n-<Jo3?)QB4HTV(jT5?yh^c
zG~T(pOsK?){2k;`HCb;QiJXt9byP4!PC(U|*Ya683$A+&EvmNQEP2F8@JGyY*Jiks
zX@nn5)1*)3J1XMC8m=1qW5n2451BZ|cnFmSiM4w@D(B@i!!$7@(JU-Z^L0i!Ed`XD
z&}2wVs9balKx@FilJC(^AT>F^by>@azh)2iLDZ#hoxm^hfZdu9+ssZWEL<$PW;uve
z#Rf-UJ>;f0V`}pf`DMvJQsbIiT0k|I@2Ka|9L{g1R0QZm?u;pI-bR-E9%mu|R_33W
zrhuf<I0~2IX1YUX%v8_T;KV|8dy&iT<3GhPSC7ijxq9t|YK}vdOXIUpjxQ8&-Lm@|
zcizl^&}CdhuVma1veB5#p4qYT$>!l*-LtND+?iw}52!a;5&o?Quc-IYVYFX#Cl4xI
zJMJO8IOGnY%?>(;vvu{nb{&sUx?#SqO73UgZc%Ti939@^Vtl&}wQ^tR-5ppMmVStP
zH)XuBc_pN5S$pTU-Sl5s9_W-kyUHZeV$YY`Pa3Me#owJ50L%Ns+2BAlZTf?`Rq%pq
zO*o0&tRX|X)4u2S@Unzwp*(I>K|a*x_uZlkeWMpkmS|#pJ}A*bG3LR}BA>zXpwZ#r
zk)4YeNr+euv2}nf7IJcv%p1_deIfRMSSrzjMSuIi78dSJ@eSo@e&0uuXu>}c0x&dG
zgyuq%1qIKe9ECD3<&3Vx?$ui{z3!_u#Puit3}b^RgRI|oB4YwA2!i4rWbcUmypxRF
z3nOm%FkI>pTsoYIeq|jQ-r3ZOO{qH{15GC8V;b(6gG6Wj5NA{)`eP4tk&xBK%)Lt*
zCkxXg@AhxG?!p}>5sw3FBqgdDw=`q_0=tWMqAaXje<gjp^K6Ul{MRi|U=g9P4OC92
zj_$k<uG9zn$miXT^;GUqYFklqE#?EU3fO2Y|Bd-K<wU!ke$!lm;(FmFa4`naan3_c
z-pQ<Egg^K4y9c_<vWoJ9eYe;#T5DwbjKjP(-P0)-&{&fvFMHR@=}@=a1ddV*IZvU?
z_VLzX+;OSx)qA<vf^(?BHBjgSU$&v*zUiaF1w;;N7;}C(swzopIGycg#bi;MLEHB&
zf66S$t#m)#ChWz*EqONy*bPJ%rH|fq!`badMlZ^lFz%45^0LPxiZ`R!y`#TN+IACP
zC6KG&;GtTm{C~&f|AUk_8E(Hh0=3etsKATA>pOFS^M{;&K{+Zn5hpUQ4S7I2&jx3R
zO%Fk_RF$ApsExv>$VKR1r#tCB%lUj9LXpaC`KimU@|jvz>nMy6S-KXS2P}W(;11Ow
zKYo~#{t^YOL&%9WI}5tnJpp1|Z)_Bk*D3QWauq|GrC8K8vh~RefK5Qdzjg1=&($0;
z^HYs{GWoy_wyez+WequULTR->oQT&+k3c&Y%vw3>X4G9KOKX6*pZcU<@fq@RI@lTa
z1W4xG`~%hF$1v|#ugEPD>qnl0Ter9&Z=F$iRrznsT$Yam%64Sbo%!tYD#4|&arX^z
z()+A-dcrt|u$S@~)ZhVHK@swB<<l*SQ|fDS&9#`;^=YyS^{v6*&)MX8j?zEZJZy$2
zN%5xAMMq>0p6T?eUrsBL?}?%5z2cv(w5XfK#rqd`4BlO;ET>+}_=_%xJBHxNfTz@$
znCzC}c9-`Q>+b^a8L(DsxwQzXP?u>kv8cP(wxaFwXnd#c4G1g4sADeb?GKHpPHYoQ
zBEr4uC~J5}2;ET6Y8zF()T~Of+M$usRR8QK#P*P*4@HnSpqlsnyFd4SY?L~h_vMT<
z&GFT`jLn^;et(dCD%~h}s-UJFfE(rw{CG_FWlSWu?cUjvNubnG%1P`<m7eMQk9SSJ
z`aU%{KYe1(>9LVB72nn;aur8k8EC3kf`0#aqIfnl(tn>n+kWA{{Rok%FB)Kk;aNKV
ztG>;B9$W?qFTZu`*3a^~;bg0S%S?XRK{Lb*1ci&jUD~%q%UdT4@LxIf7|1jx?OKDu
zeQ77jWPR46G5{gQjZiw`H`&<`Bv9SWUu6PG(({E-wDeQ01dIYjVofJTKn&6BR3*`e
zmvhHa>~SwrI{ebp@u!Zt9|CZ*!z%IbK#=(k>gF+ckcMCg_jB^2{)rp4zGCJPdwt@<
zgjcf&=dE>O%(;!AYW37405tt=Jictp<Y7<6&3f(G7Tjd=Xm`Ao5V2|iqGZP7{*3AX
z%%0enuM@fgJ>7q;M6*;fP0v=XhaFT|MqMWq<=p&P^@SgG8`-sV4@9kaQXbU{?>byC
zH~9@Abj89T)C7!65!eFXq-ua#3ZcY8ch3F%jbdhRD(J+Dk<C=XQZ!zNKu_}jjcQ?@
zeIeI}!dbZ1g<DP5(FkLCS2UG0Npyb4Bw;tSyoNJ_2$MzP1J^Q{E*-upMX@9lI4;g;
z&2k7KCOiF_^X!MZy>sGmUjV=y<ufD?d9Si~PvNM}_&wQ(SDdBUa)T}T(T|4n4RBd)
z%2j>F@8!=o8vR#-WepDB)c-BngR57r7Ob{8Tua#m^(r^77Kb^^)H?X>R(DaR42X7^
z_at(8F6<e%YhBcrbaFF{l8y)?CJ)_ah3Uzu1WrnAL&T-Rs9H;Ik;g1%c1P~n+hiB5
z?t4SuEyfePhi(F{fEelyH(j8~(pxoJsLBi=xNQO_hQ{rGYZ3dxw{xgA2rjNl#f3vU
zi%Pj~!V*byl*0uKlx{~T6ag=`zke0S$BfF!*CMMIBhfr3{VMiugxnAxlo7y=BbrB<
z18<J57cYb=qprykRD)vtV)X$SR+Am^imVb+2Waz%ktah5IDH$J<X0vn&|0S&Tzx|K
zRboqUNr5$lmJg9F5ZQfASol6kIW%Cov?s@jJV$EZRx%5h@hY4!^&Ub{zCuM*XtOv3
z`IwL=NO>Fzj`=KY>ZldOgJ-|%>FL{T*wiaBENppu3~3b^9Axjk?5));xU%Br&0U=B
zdH%m{nG<kXXJwO`P{GHKHx5sm|E^{2m|Iw6zX$mp#&*+3kL=b*Rdw8kVoCp~BCYe~
z@w(8#>JA2;U^Fv=kvmd<<p1I7E#sn!+ji|4hLjjWNhwLGA*54K0YO5LX6Tj{7<!PF
zmX-!lP`Z(!yStH)E*V<-T|Do8?tB0CeB&##)|&siuJb$&1w6RU^Avhw94Ip}flM^k
zK&A<Z6BUQngIuf8m5MM#wHC^HqE^t+OxMJ`VVg7~pU^UGf`@0hFc3?d-r)|0(8>x{
z0R|FWroi)D+9u+NTuhIW4CCAA5}KOBi7n7Cl73VkyY+m8D8b4gGRx`tP;jP}Nf^VD
za3}uHe~WL_5wKZjpk@T75_d1kkOzJgj8}SqvxK_a_Q53JW?k1`x=dOCxFv6Ob#1O3
zGM(D+_yMN5?9>0e306x2rTATLva1e5#7gkk?5xew`Mjg_1@PY3q^pY=@tjmC{s8f|
zShca=+j=Q&`sw%dCx9T<er4Y7QJbiQ*F9R+<5Bkf!f7XLx9bx{>5{3Fo==kwXoeyE
zUp0e}gb~-yD#2#Nxh_HJX-E~|EIHr}dOQ;e8qT3$iWTt@vZ1J>L%jTKQ`6BE70T8f
z<}Ih5=_S%_!|<&8nY^cj8kO&i)sJP7nXEX;qrQQL&u?1}h%Ir}8@L0o+J46WdKzfO
zYSn)94D9pWPe#%#=WazRPUgc8JfvUQblz#|<!Ss>IhTm!30J<~*<Bu2&A6$1wBHz!
zy_)o?<*%76)$WAXy-Npma(*tOF&0z!!~JC@?Et)*hvc&qV&mQa{+sNs?#m(ixd6FG
zq3Q1CpdWY!TP?{)(sAE;++G~D=1FtJ_0?ItUbC2&m_6xyRMx7*3p++JlkN2OKQl>0
zo&3q~a{Ftt1nX!-G3SjaDNCQ!&BH_F4l2qB&O_7~9%cjMUE|77D*ct?KSOdY5mK-<
zTomzRU}FUDIE{CIDknanD&5Y{Shk!}InmlLvMwsmI?&gq(^k%O302F6{yx#|cg;A+
zvHHE+AL^aZ_~6CKl-mt7UL)IS0D(HJgBktO+<+M{-!H{9<wL(#>fVQB^1pMkL3bq_
z03TO(79!W@+<R%<@i{-Z1(J0Vj@G`EP2tjICt+;bWr#k*VgER78bgG1GDDGzS4%Kf
zuQUdAR(KdIOH)09C(O{nUgOhi0mOwXG8d>xwTTP%)T49$;XP0JErB!RgXbHu-cM{*
zlOEV1?Se4PmUtIZ>T%D1nl+BmpH;2WZDdHT`YQgxJS)#DRzEBNDoJ*}EJ8eIKOvZW
zaFmGRlhJ&^aetI}o*vcJJQk8&K_nomjAge#WjEVDGNQh-=4VC5F7n|p&Thhy`4bCA
zsVbqLH5{<S=eDY6reC@C%e5I-m&G%)f(*v#=n5m#Mx123sOWF-$a2t;rg1BYj{unP
zS2?$$NYw@hyT{_4Rc}z@cu%__iDHUy@EVguhWTR&rgdAgji3%T$}E273FF8;h=S)|
zw!={Oc3`ENu!4Q2!CchX)<u%AS&z4@4?3S2j^%a9oEH?EGMSUI_`)ARSdb-EDV)!u
zpx)p@6!*~FJa^1bX^TctYfCy+gNlAPtdVGTp7_cPXNqSx-YgTRZ9!MQDrVoai^{{>
zEbsb2b-K^rp<aFLr*^N3_&)P@__dwqU5R-AnqR5cqyGj)_wR$Ej(Wc@|9H1zQS_V>
z^kGuh9`jU;3;QuN?AHZ)#-dFL*jIxe5?fy7Z1eGc^qoY++3)W0hp83AdQ7ZsadY-w
z>h+Uiza9(zIz<waf$;=Hb#p>JE7gnJ*DpluRNkZvIVt%0#%RjZW+PataD9~6w|v9G
za4Bj(SR50I=t5_ld6q-rUzzlxBhw|%#)$sOTpM+Lefb_LIHbWAP9~x%q;V<Zi=>&8
zjU-WMm9hij^|=!EeDtMM(x46_lab^9a&)wme=6|mtRwGDwqoo|?!b<XqfU<h>Cplc
zR4J6Mw8V(@I19mL{Q<ww@*l3~pCVXq-8sDlI!^L@<|nj)?|?am@~auvD{vu~Xz*N6
z5REP1wm|YB@rAmvJ7t2(F+z|}jaiY{HKsf%u(?*1^mgnP>N!Fa+NE+~yXXsy_h4nD
z8|Oq9^9=(NN4QWyVB-Z+TjmLpHPtkLw@RQ^kR{&k4|XEW7!|~<3TwhN6Qp^9FM?BL
zH(K2S508v;R%erLn8DM580Ya!366x9*<30ufmD&oQ$%LnLNfupc0&TJe@*tw{+5Yu
z7A6aZy(xpaP1*YE_+yWJ;6kco;zjG87riqkzgZJMvzfefI>(rD&SH5vy7`EAusk?&
zR|G*9pfQtlVH`Q}d<m55)!WtP3YsvB{!w-7asv@MGT`^NuTP&P6CbjFGgpTU=X{$n
z2766~TKZL_(j`yB`~<x9xyQoAZJ{foxQwisVt{J@=<%2^0F!TqUkMC+&cqDY)}b+B
zHOX`p44nNa_AhEh6k^M-T|B2ML(dDP8OQcrrEE~AcWoweu{A1xWX~tenEkzUvY(<)
z+W*k3cfeE8B{foc_5gITYhT+R?KXMV@SkmDeU0k<`Nrlqkk2{_hWR`Pyx@E5KIoC0
ziecg5WW1|AHbp~283+;|c8HySvi>jy`@=Gsi~v8003}i$kM|RWexLJzFKXuasQ(;*
zP6P@f#KBmIS=wyaA?hKM1OiRm7Wf^A3S&}3Mn*4_AA)EZxX^v!TliA5+SNqxno~cN
zAwvDTKdhE@P$PT|TZZ@YSrNabJ)hQPP$-d1B*i(9$nieJ{>$qi;f}<?{#d1oN*|YI
zz<pAV1=~k;KzIVg_TJ%$EO*jgIYHOK6#dl_zF>vJHfUH2IcKJ)?3BcL6$&x9u6023
z&S%R>vDiTUA9;b-3@*ecEjMWKE?~*P?0m$r3yI`H4zgB1k_oco1e@4Ox8MoLZ)m(9
zkeGb+#L{k~hH{bL2+7>1!B&IQgmIp<Edh~8fh}{}5$=5%y<mZGmVLTtA&c@W?scdB
z2hJ}V^XhunO?W#>?mB!4LV{I6ZvvWiL7_I%n0!V0H@q`FmyCy4c1h*nNRY(@#CZI!
z`byy$DGa(y&711o6Q-0)goOHjSJ_AswalpODkoeFsDdn;4=%@b)!E?VW=#G9E6DH_
z_hnQ2GI>7mOWvF9q1-mZ<xsTde=*IqSs(zi%0IKYnUvaMfoK^xYRyMQb`S7Gg2nk&
zHZaxM?7rM$G}k3>LilQ&>^1yfh&OY{eU4)|<~}AcH}f4|;tPLWsA}<^s|v5f5a+bo
z^s>cQB8Hi+mAne7!z9kj5MAAIfHh2fjK{Suib-@tP$`r$9{$ex*vITQIFU+t#3feB
zb)sG3Xnr)?VrwMm$7|i9W}~XmX<D-<t@t%o85K`b$fih-`x<AW%idjfEtu7^9W-eS
z-L}nExSw*+blX^EKGlZbZe&{g@6SH4VZT2ey(t7JwES52sQoIgpQF~AM4n3~M~(Qf
z1WAuCCcyhHM%e>dLf5m8!}t=s5T{pL<gnrNRhCo=1U%DolyxBD7}Zg`({J}q^c*Xb
z6Px?Q|MP`|uPC$yEK@3LRN*<g$&yZC3m*%>o`?{^VTK%0jf`9#I<<NCwuC3;FUe%K
zh^ufhal&CIILKF23}8BEMi6f)H#ZFmeSEZJzo?ERhF>2NKUw4dcJpocgjQ9&u@i(h
zEmq~DW))sEVC~sRgaA>hg_Mam{s#2{6|AFB>Feiyr+N7H25CQUAvk@Z(icw)i%XKu
z{a2L8Yr=nf(Cg^4Gai-0eBYhTy23=yHRxgKB%;@3-My0jS4u1MyMj(@?5Un>@)jFb
zL;MPS&5{;Qcr9dJcVKQji;kZU4r(d%#oiAf^1$q66lhHAE(4NiN-uHJ2jPuiXQXg8
z)Hpm=uIWH6t;~$!VOivbWwB-gwreEYTp3~jwOC-5JvIDG**a~&KU%qBAazkL?8!=O
z0&o~Oe=OjEeC$)0VzoiDjE$fwG{bJ<-2(E<guzang~!-0+J+KZ=Oe5FFOech)@-8b
zZTwwK@3N+TH$V2!3>|#Xi#hz=Y()lgyNv1{DT{vzZnnBpl#CTKSS{U;(bwv=&owsW
zucnL5rdjRZK!iN$bcZFFBwQJhW~Zh&9sjn-Hp{EYYtL-R9904>yo%m^LHZ?3-c_6Z
zeK)TA`*~F9qkR`ik#FA^UU>NCDkB=}exKTC^r7C(sp;Qu+0#81J>4bLvfhwN{ZEGZ
z;6KwIyQ}+!OxjrE`66fcW7o*4+FFXCFu)|_A9wF63U+iX)fvK;o61`bmI8sSX=B+*
zH;?oPaKXl)wctg`y;3ZO7cETDAb(KE?dXF*pdQPc&WpQ?kMcvxB|!E@UZQ+AYV-H7
zGSHgF^kta1a<o%THH4Av$6+;ue%6x0w|}f$me!N?q=nwSYk$=fmfFy#P;A!i3yq!n
zeV&vbR~L+l@o@(F-pepi!g;+k-*5Kq?6#bIf41RSEI4KXR(pFbyFD6Bba#oxZ-3xx
z-rDx^v(KTfO~ZkMjn7%HEPm<aiG~zgLxy`&D}5I>Gb1JmG>4QHvS<{vPeNzAEPpjA
zt!09?wcxOW2h0$Ni0RlXQL>{6=Us?WK6O7n9;9<d3I=N1D#<*o!L*H1=@{6>bIcEl
zo<!O(R>E9ypS@_|Kn;SdkTPdE_-b0=XLu*%AUPTCMe70BDIZTZrA$P<c{E8`ctb!y
z74W=T&1oKht$$h}S;n&=X^3fFMB19|&R6gM$jg(vmu)B&y#2A8bhCCZJMEDNmIbnS
z|9vP(0*1d`!k<k|ixofM9>3i_E3@Vo6%D}Y`~hoIUr;z1p-MI}56A#X48mogpi@Tt
zVL)OzeRL8Bk+d-3qr(!EUi24j575luqC#-xcllyOkys4a!|cn`%R0+XJNaZH8Y`_M
z#11F&XeBB@CMS5%sw9J<xpt-qH9=H-U5uLaG7+89gsMK6mkCpNCY#Q<%kqj!Qa)|M
z(d2s&zU>8to|riRwSS(8wf~ex926!wiLSw@P+y|tZqNIt>6*4T*W7g}^<B`$j0r`Q
z5;#5eW4(WEJEDnma9Fyy+R>o8UA||4DJHJa`+zCWJ%uqJdZ+!GlQaMCvRpCzk}PQD
z6}~I6`ILDk+T}IOkfZYUykOGpUmQL0_Vf}e6J#}$C^{%4)bkP$q}5(;$7q%4<aI^H
zbsME>pnD!HKg1{qUWr0*{9y_1rZOgyd0k!xV@JE-qaO<442*j05KOC;*VQ`x?5}<k
zwl^^>aUAIbn2c(fMGp}MMB;nA7hS?li~y6%BPZcMp^rKE{q#JfqvHOxW^F4kCJC@{
z&WlQT@HC#h6<v3=nhn|O$G`Fh8zAiwFUn`dmLt&5+IBJMbG6Ukw*SOs#|{H8F28wI
zXX?tf^1Zc5CurN*@y@yoEZ$3m6bYim@Wc%;9LcP?`BFf3z`kQaFM7t#q&yazd4|{X
za%x(8?lYDX`JP!w9R=~jzpGePG(X)yx$l?6_VIl?G$MY1?It9VH>)R)iKA}vK=RpN
zt#!-5QZlh>TBvrAAHr@$@Kpf^aqn5iSd&%5j9kQCPD?3I{np33c%!?m(5YI&@PY4H
z*C*je+wm$1;bz??PXPTcpt-sN+t}D85BqOma)#E@`U-ekH>ZS;G~id;&i0r5T8#Gi
zBK1znY3BJZgPqZn2u{W-^LcLaoBK}XK2V=1Hu|Yunv~}7OVRK4IBWi3Czere4PfIY
zX{bsYvZ|)2vHxF3Hk|t-8zYc)$fN;AS6M8{`24zl_t|K`&m_B{7`!k&!*yQzUTBzK
zND9Zg{DA7J5bOr}fY9Yu`bL-=oR%Hm>r|A40}!5OUV1K5qa!Yn?~qaw4P7?6vQz=D
zYgT-k`p*Wa=_06|Ox-Fe$P>%(q>PK+^)E*Ws_n8p1u;ZM-IBn)L6VnBR}yqNwX`yX
zDb^#e?<jD4d=YM7rl*jiyn~=>(m%K^1jy;1nFgKch~c*F>4JX}-QP1^J_T0*98Mni
zf+}Egn+|}r%w?P}uqh>%#;!{HB}CGRqMR(+FUO;5XC<ozCtb9_f9yY9Vz`>zeG)kQ
zWpJQ%sC+z+fyVi?dZTD9)(t~lf4bhFpy<!2UF0aN(MlDu`1IIyYW65KhMp-tgpiQq
zzx&5|6p$AY(*J!)8h+jdP<3HIW+@>fqSN$f(9ut$38*G0X)6|@X2y8l%MRnS<-`SY
z<|Q#cj1}Ui3wnP<YTFCd`FI&xtH+-#Jp4>zrfn}a$Qxq=X|Uq54sX$-#Kv20$0t(2
z-gt06oBM%p=8qrTtCl6B0!&%^G@BFCE)B8@N7&RXoR(hgQ&Qq2%s(DoEG@0myZtce
zXz}vqCzamzf#9?O)BntlBKxcDmHgcYDj&@{rhek}=<o{*>N{?bBasLQI6<s@ZcJQi
z7u|n)T*5Mu`GPGlkHqt&703K=vN>*|&2wx#F`2L7fMqk|4kCIH3*mB({_p*H1_ThR
zXP=~Tq{SHdo=#$Kj<n*?-!nOWYHDkj2AXtMP>k6V1lben(MLkwgIAEwG-EUY&Yd|w
zt_~_g0pZwRcT{1n<Q&U@+vS?G%gA$_)M<nAE$@R7t>oPd^4P2!0<c|BLi`$}e46Qs
z9Agmam~Ql$SzQ72+p?B*pB@MIr4>{WWqKhW`XLjO?{|el{~xRYX@1ng7^C=kX?e{I
zG;WQBnTatWgPEmDvV?<Yc;Q<6;`F?ksaLiw9iXC)%IfyE0JgmU6BMuUy<aQW%s5@!
z8C`w3H?N&F*3@VXiL4Tc9-xlDmXr())i0u^a8^OOX{47IXfxGbxfrSXSvBq;1V}<=
zafAJ&h|__La+)Fm!g0CiQbHJhCt{DgdOQq9R`MS6T}gb##QvxH%HV~9D4kuP>7%jO
zK0Vnu#zb}e7&c5g5RL?-XmrYj8{(4(kG}8-0>UKtd@;$%{SuO_YqU-^uXicwV!njV
z1M!MIsqC&DGijx4YBW)g(2MyY>m}T7wyF$Qpi-Aa$x8CHvdL}jNDrp?{-c`v-*t~O
zE9tA$C`jCkO<;bgwUYMVuQBQW{~Dj20)+i-v4c8Vj|Fvy`gQWuXEvXGJsLeF0Szj4
z`KOmDY%7^*7-g~mvntq@v15M0cNP6(6^>lMfr|H_RzFoUvTu6#H}XC5&N$+Pi|lEf
zF&Sxw9QZzh4zD)8{0-jlve#i<S<%W$`U-8~R3gq@QtWCIlxV6`@lI&<j&u#O#wb$w
zWa=$$Yhz4xvnh3Ehv^=NiyhnL7nkZ&>zUEnuek8J)*DZ^)|^LA8`N?%Wz{xK!6x9I
zDLJcU<w}mxtR+3Y<6AYJ(yNzCzJn;wmjjZPXBlg-lW=0sIUtxj3;)b)m8wG~%pgx=
z{^FY-KUsW2a3fdfF5gXyE73n}gNPmH;#;?^f3m7#@4xi9ZCAf#WFfq=b%`G6oQd`2
zuxT*tCnzq8tyfC!wUYJ@k+{k7*R8+3xtf>CxIf`70u9&yy_78OFQuegpySpLw2|CA
z|JHBry=?gQ?UfOwwB05)A+;v^x|aZkUKEXR37VUjm&mc=jXI^vFHyB9`hc2TwaF$~
z137j%-If<Aac+kdUKjmuYUiDCS3$f6*e{k9F*`ihn(usR%?VC_ANZ;^on{<-`A(Mk
zhIzOp9yg9X@E9TO9O5;rLCoIf$w0Ij@bo0=s3nHC5T_C7EGi0uWD~wKD9X&N5;JJ#
ztUv8;M8-JSD;?BG5xe8UsfnW~xfsK7Pb;lVv{lQVN_T^(E?>Bi?me``wA^!hL=vbh
zA@In;0y_nel*r`GeN2An9qH_7&~J=Wv0HJa1HZy6$fe(t)XHB6%iEo#ltVhpVk~&4
zuK0m2!2U4%MscUkm7Vv!@w9kLdL*<0;gPsqKL_N5F^1}ro64uDvi44UbdWl<p3BLU
zu^0|~?mKM9d~R6#I6_R@iU@iOS#RP?-EW(9Ly|%MDkb6iSpdZ<O=-@egpzH#d=!ls
zcpsEZvuMfethjw$rpj_0wMmmw`ZQi(BxF!T6@#Ly=X!NC#J-R|_Lcuk^XJn@LV1>g
zycb%w(x8u!6meRCMY|c+bh}HLFb+4~GeKMNoooWv!C%IAbt9Ra;~9F=XQNhCPPE<v
zcaRC8BAdN8Qx;uO8JsQ|{O94_Ust{?DeUH60OR?m_swU|!TqMIjo{D!9p@D*0pq+h
zu8P9j%jxE%`+ji?#*@_c{n{m-dn@GL)fsEHEHe0#+?WGBfJdqlREB++JFMi9q8OKq
zkimnUfVsBgzBcvfyX%0kPq!Lz=n=6J*Go9XL?_{@{(84wzunG0M+>WDBz4>29%Q7{
z%zxWFRK_sig8$}%E_lW;i;Jy|K%0$hv7ti4xga7C!RunE&v!c6e;_pt+ts5@8i)Am
zs5kt^Tg0i(TdmJu=@!!?WOTVLk>G++)#lQH53kdbvi$ELQ^PCXdpvGtT{;)Q8_K?Z
z_=Me!3`ti{$Z74(k!k(uSeCx$gnx&EO^K01_m$s?&cy`O-x4lAgJqrju4hMCWtZW~
zK0JVkm`3J7LCA0P^p7ba5KRJrEm6uQLgHLOON)4~z@J1^Z*UbLk8@)t>LsXxI6pTg
z6SSE3v9!bB+A?JC1sk}Z#{YoC6aqZ)<FH%;Gc00^P$bO!AvU()7GD`2Z)LWmJp&$P
z5Do<>#;s*8TM7idp(Vwa8T}LeVL3eCjeGKT`T5H*COV_Yz)QTebhSsdN3g07J<-*8
zX|<ht0JMiX@;A0k#stZzATQ8bUetxzT+miY*d^6)^!HaU=`{}t(3uWUsX2KWRu?H)
z40MtG;3PCK)!z$A32fE#_Z4HBoa>Gg>+-6m^2K;&>VZ8d!DkhTrysY4CDZk50N_(c
z!4D=b@`lAve@=<^%!tvGX{T+ABCT=HURmEEBZK0JBC!~=RZAQM&|)(4U%7+?oxZJ7
z#fSBB?E6RDI%hbORD3EfV;sf=C7ES<Z(z$<rY2MmKaMX<Jr!)0v@tXOTab2zC<%xF
z{hAc3MUFoeJ7`Tmn-N>7m2I%1^BkaEPOO%b#@?k&(YO?NZ7<u%Tv+I!%dLI4F@5yU
zURcT1wUU5FNcp~Pbzin)&U+QbfA!z%+5Ub!oyNugBNM=mpY>LRzj$(V7~^li^J^>j
zjTfJJ{^48u?n`YE12%8nNsMJKxC2Qx-r=AAe!<v~ns(y4l%|m*jG=G4t=M}ATr4;B
z(8iM;Pxz3hE~^hALuWq^to@`L-iPmY{70e;lxFR!_6%1VqL#9mFg5&YPliTD-J-sU
z*G@8Ob~!k+W}81#yh@&ai((V`OpwH0qyK|7J(5?3P$G$vyKFUI+#xO$c3l%i_2y=I
z)TPCPr%DQPb6@)+LmGk}b_uC93;6X7=%6b!t807v<G(3Y{CcHVefzpGCi#{^Z;p_(
z-19m&nqk#mISNN+-qSul^1Tjg@6iKKziz{Q>J=Xu9Du<E47>tO!4fTY1wZoMxa<D=
zbn1GT5`J{>IYs&B&kAp_w4h!{IoJ$M-cD1H?^n;&d?E|HQ7bk$_9<HR9Qprd><xKz
zCdv1q?(x4(aY7#I<q^K9ACcPzANu!d5|pPebrlebQDO4-DoBkM$}_N8H3&sLrVD`s
z+Lqm&aJy!98C+B8eN}y)8R?>|v-q119Zuv6hPSvXNG3ev2*2rhO=Odcp{i$b<$MeC
zDcE=C#1ljjlIdFn{%K@C43sWX39kz&vN6$rq(D@Q9+7x6*ye4Ecp02?4%1$BM9biZ
zV9y!vBE{*)W|JjUTt6qTbh?Y&K#$J(-1{oEra!p%vD?R&bHA~)K%iu&`F#(PY^w1f
zQ4$*R1oZ6CqW40)A~hTizE&9+T)I}E@m5>5DV*IqWb4g(Q=;9TXX|+6nI}2IeS4ZR
z1Jd%FiyNw}v~_Hn>4haug`T`ZuzFgCY@;xUsbFqJV@>jI8E?Ds?Zm58`5^K5^BOf%
zWmL;V-Y)d6(l5IMGYx)ojNAQW{&ur=kw;V+H|P6tE_xGh8OL~?tU#=q)x@>KD1Gyt
zbbaU&QJ~8up4xZ`m52q3i!hUOBi$`1<FGGsm(tUroDUw-&f%F<A;i-)a6KZtOn<g<
z9_iM_*UVQ$KShD01DyjlnAgc?_cG(7C`TVTsgoS2;=4pCvj5J5{s+9t`V?rf8+!a^
z0T`%1JZ*y7%>KGsWmv$LQ92MKL;#lY%4A1B3qzMg(TP8Z5aGKVC+7(l0<$5p;Vzyz
z8~Kly+xs2220J~M@g18G&kkj2ydIPz>qC<lm)C3i;Z^cG5t49t75zAcb00b*d0}Hn
z1wsW%slniCUHxX*-ODNB6d(n@Lg->A9r8od+W_QEL3Iqu$2Eh(4nnGjT?$k^Lu_^7
zZOI?L6853r4o3*VPG+$5oe2!|3oOt@^`=?dbGF$#CQNJN{>|F2*f_W&q3U#w%*QRc
zfFl$EMz{^7ajr*@_bhLVE3}uXzb%BF@5u%4Q{q2QmG}!WCBZL;aJ@B>NKV{5sO#U!
za_?#V{Ighq9gEnQ=gN)>^=a$i-%i~mpZ6iZ?%{}>`*4BeS*!_J_q-<qz6?j|fe?%X
zl?9R+z>$GCHd<{C#H$`OrX|t~*;ZYzXPx)3#fTVa@#x#0?l&#n9yK;Eyg$bv-K@<Y
zh@PRc{^)z?4)NQYDD7K*aup7Srji*m)bL<gZ{eapwpS9Dwr#uN0qN(OgLvCN*uJ3W
z$ZAVT`QM<t^`}Vn)W|m-RX%-P(V>r4;t*=!B&7EHiI{1M*w2E4mkg2dE=AH1CB$bN
z!y@jzUa?%24RM0*4*jBUC<8s0ORfS#beriLuz^je6oxRk11bOh0VQMr?28`wT1=wH
zwvUbk0T)^K#PD&eNZC^}5HlYuh-4FIR}Wn03>~IE9hDN&1znGcz1kIMo3pf7#_REw
za1FBxGsnUn(i1Ck_`5#)JSg3i;=MdBkNulM#gy||JzqMzCnS|P!w+})F~3M82gOv-
zmBWPAJOZ7KaWc90tJWqgLCSMheu2keO+n()(~6B$jmt0N-iTL}u#xm7dOuD|HnjxN
zeD1x+vWQz8b|~%(>>(YyO9}}yeU;|R$<j3-RPoe@OcX#d-$&hIx5vNl>Mecd{_jr%
zHzN?ncr(@(jK2;9bCUs#U8xi6%=F$pD_fn6t@qJ!Qh(*(J~>8cEg|fO3=_&0^qf;b
zmk~u;W(*s>{G2w$@3Xm{oNT~A0&7I#8WX{Ofke-~Jjj--UF5`OEeLc$IfrK2pd=ta
z+puU$-otl92^h&R2AqjRdVX@bfr$4L^&`|iQF!*@+~Bs*#Ij==fP~@KU}w~fD;+x|
z$aoJKN{j#nqo?#{;`qHm0zJ!Nv`B@B6Q7-1|E4bSzvJ3)*mkc#mBZQf*wJ>;elXTT
z2NUixMo^nZ$&bc&ubPiOGXJ!*xF8bOZw`2qfJMinWH~5~qX1?wM<G%KOY9TzTrCxu
z(-O~isC5N>zY#UP*m&5+4>mDREaHF(w^Z?_CNXY*jU25R={t<K*w1E6G9gvv>XI(y
zw%8}m?Y;ndyF!>>z&T~(WZJi{lHmUNlYATFS&3BiQhJ$9TMc7q<z-Rf(=`E$lf&=5
zYW8Gx4Xoo&#V6iTF*7DSjD;9uWK$x0J*aTMHb+#wks-aoI>Ue*reMe%c6UH_xe~ND
zpOjg2oc+a^!#Hs-D6V)ZGs9Ar_qrdW5x0;TGtrs!to|!gp6^Zq)y<E=MTFFx-ehp!
zoYey2{mM$01mvr3a^MFZ<l7C;Z*IB9gm3*puz;cgh%3PHOFN;$&;y4PX4|AXDI>ZM
z4UFk`vhB~B+KKdv!CYsW?E*37<E+Cnbwj#Lk>RIdjW&=2SPN}%8KH4(nRXvWKRP5&
zqL<j3p11e-adg^Um0HM|XN@kIgxpD-Rgl*JM7f{OeToZihP_j-y}RlGqB_Z|v=v8E
zZSK!bx+1DgHH~>g(|3zcGnSGjrlwepF8Z|Be*7%qNGG}2(QfA0%<^yQSUvo|0TRXc
zBDcIJFFx^30@DF?08Fee!g<sQ^Haa48Xar>@>_WTQAJ5X1d%*QFuX|a@a<rP2<$kB
zSZhm3BpXl7EGJ+cp@o&{fJ$L<=q8J%@LlXWB>tglML0lMNYJL}*~(@>h^ru4!9Z#m
zs*TV1aaSUVln-JJ3UYqu?A^QxZ^1st*h5-RwBk#5R|R?yF%UTus#1fV<b=pzR<NE~
zTY&I(TA2#`SK!%944-R|FC$~jK(wn?cfv9_ZgBrgyo$YTwj$aOtwq7uo*=vy()VCB
za0M~POv$vh^xW*zcQKM&Z*`ycHl9~Xfm{ock6K7&aI*uw>Ov2}$8+?)FJ9|Uy$)a0
z|8)l@p9|Zh^~JR0V{O41ArTiWegDvtW6SKKL-8}601uMql6Iay>#@Lr7W_(~d-i>n
z<emsrL|3EPfal_*yvBb4Ny4<AE`2>H5yvC)m$E^~dW~IlJvYK)A=5ienABCSQHE~{
z50vn0f>}#?1S(?(H`J3x;^;#@vt#`;*(>7ab}%za>!0<hfKi$*A`yslbyy22@|7Xk
zkJX9-NsG<Hfn0~3XY&Oa0opBQZ93pUV2Tj6mx|Qbz)T~fi@;%X|8rqIQQc`kG<lWG
z94Psxj()kzZEr~AkkW#?d`d@t&#r4O#}T1ZF{Q)igCj*kO<3@HBhVZ(>=d~rFidL3
zhgp0M9%-x5;WTJh0rb65Ug~@^rYgkKA{^;{QJ;p7v#5_<+<*aU=w>@qnrrFytad$4
zp03>WJ&n_PqW1rr<*Z=>o`q2lh9d!jk;eV2-<*=*8~_?$gZlpq+_%tlSQJ1lLqH(B
z=dYAoVpwQ<u&A&O5i2PbGFS{)aIu_jnr!487y!IwW3>diw%{=+EDv@?_{EyU3*o)|
z>M#c>15*^xvul@FU|fU3aE(EyW&V$mAx=_%ArJr9WASDW1G`2H0)qCVEi#KOw*67z
zY)rc6L6nx&@IHzdT3qa9P%cR>&Pg`+GVVIY7OD9eCV#LPpjcZUTQan{0Tql=>^+iI
zxiQ-HZ=WNtKTQ(f&&-p)WK`6Ahc%*MAC^JI_9Dm|q@NC*_jH-+KRWQo>18^bwey(u
z9X7O&oZ;Y7ZJ41N)UPmb__a<Xp4rs?P;@O8QpAH!6tXy0pl>IBwM;Sc=@=o*W4Op5
z{M)C&=g3UHFN~ypZFt&{?R&b*r$h@>^4;zE+vX32%8#WNhoY#&waB>**uNT;U5x+w
zM-WGJmiKo#CAQ7DL|0>`xwcHc4m%~(m>bJ3#_x>W?7>eyh5`$?(ayU^=TS$D9WM2Y
z9D;&^v(g{Y^9`n*53crq0~uTUT_muwYxuva*7o=h7Sjz&PH{oPj*AkyWyY;yl%cO;
z=<LHEBN$&GiDU>E-}_(`f!^ZEbeyA8(2_C_HgUTlTHq0oGKdishI9_weV*~Y$;bh2
zm(U5pg)z=$E8`j$r1AT(ydZ+Y7|89xDgT&WB-IdSDxctu@??4%tkRXJ(;Wz2K1#yt
zmTgEdK_M(g;Rofeu+>U4Zq>W8N1Jt7X4&Mrni0X>_&2$dwz<AQ^l<EZ+1>fg)pHwr
zyJRS>Um3x!cC7=~L@%E^-T~e)(_cK3g{`>u^%mdrGC#<-PyM014%<hMc8wQQ#=rN)
z0j)S-*9vXb!CfWxB^nkk(Xru;_U?|NHd!Z<K+9ao-P36YFM<GF!cY!N1V#TnmEjb!
zs)qK51iLZq;Oqaz@5&+WV%#AJ&^6*EdZaBk+oaU^<IgV9s)4?2AP%(Vg_zGD+U~_n
zpzNCr$OCkDc6ads(EI;c#7);^YcFLQmPiF@kGv87%?vDxlvg4FC7s;bJ%wOyZmx}L
zI_=SNp-uz^jOTH&5mfptA)w-Z-2c3Ka5Uo_^^1am*HI2*@n*YgelNvc3NiEie#jiN
z(0nK5joMf&H|s8exI|L0hkj~5@S5*jpCf;CtAU+mQUH4p)8=II@zWrU)*W$X7Gmp2
z9)kw;zFT=oT}9+diM;I-H4V+W^!abLbIF?`iHcmsO!zn}RIns5SznxMZ{a_C75j{P
zXF>M80fLuEHBv47MVs+!E^^Lv=f=r7$}kKU$&;;k?0O-JuHW|qe7}wxU_0TyknGhi
zvb?mx9^`D@MWnG;spH!!%s{(~Z~q|WL!<a9$J7G}&#Sk%&Hq4%;2mFO(X$$60%~v@
zuDyIqnI^^bM@bI6hD~bS?Abfg;t8{}Z|3xa`Mr@q?5-c6$Vk-u>f~pS-`fc2vb<RO
zL7rN=d7xX~b?cxv#_q3=dh1}+^twcUhM_sF!QgL^5lix*AuB~=or2$2|4j)|{igG4
zA5s4)*^33T(jMW=R8G~H6fTGqBC<Lvb8=}5pXaLhx|z)=Zk*0r(NEAS_1gA~AL3ma
zJ@Mav%9|s}HG|zpN~`D1ca1*4w#jHS>Ur1>;AY^w7b|=PP?GLLryHGc+y9B8>Mu*y
zSdZqm)lPE+88~9bHloMgLtJo)J*KQw%2rRq-K`G>+v*6b-R`#r9tMXg?SprCMveCc
zoSAByKRCXw7?u&|Bde)UdwPx_?F6vFO2f+@7)KKY3>H00Pg3$yM-UmoZ_*=7WfB<s
zhcxhI5DBDKM6O4;3?Ev~tFBOoWA<uyP2BCXPoQg1glqcf)zg@BN?)dgj1Cy&RP3fj
zDe|hvMJZ&3c+5Fdl3e`AGteROsTQ2vJa=+;!mx0uSFJ&Iw^GfOdw4M+TTJ$wwhdKV
z0~LiY^ekwv+?;{=GTVW)Pt8MRI6f$3jb+a*+z?mM3bj|O;Ja6?kd=z#)tho!weR2K
zEF#Q|AIEmf$z5vztABaC|G0S&$8Ux6D8JNi@bY=cW=`Ux-L{GUb_S;Iy#=Qa4!!~q
ziD6kSIcWXPj%_{AX681ZJe}GGz|zS8c<l;sF8^)5yLDsC$dT=-e#%>1i*|3=`>~6^
z)bY1;^nweu+N|Q+T1QNJqJ_v*y-lLH6n%YQ_^li9XCK(ycwqslKI}foQ=aQq=-PGL
z%qGmh`l{y<1TZ8udHtqyK;mVqruu=cN3{4J|7i$^ND}(2s2D#uKa$=P4y`5oLw&A}
z)AYiG6tKvl+1cqAl%y4&U@S9I$-lyZT!X%W%x7?7F1MIwwZUAEU8M@AsbJn$aW?Mn
z#O&;yqX?=e2I)saZm_`bkktHGt^^B=c3k12Wzdv-!SBAn)vswVo*M74&!%fgW#?Km
zZcMOh$`zTHaWtoGOdcDa<<F0Dfzg>J<VdVrL+ZK^NPR?*i7NlYVSA-7H(Pf<lvU;#
zbL2;eYH~8)A!4>A2qtcq1-sJ0-;s7Le3##Aadl0VPKG7MPQ-;)p(E8#UXtsb$xR$u
zxGERkfE!bu&*^`2iz)e0<qe$k&k0LSyt^ueXl5SHga2aNP$l7-WbY~VUNYd973117
z;4q9kTMKAKUrgRjO1ay4av7PrhDesN$Use?Lvgb;`qT22HA+I(8@M0^|1y?S&<4oA
zNVI}jU_;Q>j9)B1`>Xn<m2&RbT)tn|%U5<g#`FU|@@dzbp-XYP=XdhI#KrS94**Eg
zkLx`v>3HDoyP5Ln`d`B-K;%>b78S+X_Y2H>ESBe><q9CmEV~ARQ3=5N9K{gIyEBli
z>42O*6vNz!sHcsOfkXKPvE?-;!Qsc7B^)}<`6hSAnj3)JJuGbu3LMCJL&&-ADE!-j
zcz-QEMgZB`$ovVgN=#q&w&^(-hn=Ax+HUXh8;{J*vDV()@OalBSRcf0Hw-0Ycx2h!
zKwg6(TV!aAu;VQ8ze$svG%WCUEZvz-MO*-GkyLHE@59s_mKqorTT4?XE$Ec+XJj9>
z{cfDTL;MOE8|8?N-TTjyQb#5bVe_wor3D7b&&jg;jr<NZF4~9v-kUZ3J+xXI4W;|s
z-+cM<%5P59kC9&LFM@_s&D6T6Z7#<!8%m5fKTZ#d^wXhF8_N)?i3^_*qr^n$PQ>wf
zk9vT};L0S8*n6$weW>y(+P!Wzr{LoZ00G|||L>#M;BH~U1y?2C7!HUKW6fBp1xg&|
z*|Pfpja~FW;FO<rqdDvB=bM;%|GK$S!<f{A*Y?>E-fwP6HnfCdtaEGuxaT#S$Nao{
zCcZ$v%dbz@4%xzAOiyF%h#E(qwY_3~wE{<yXAtqejZGzQ7?Vaa@`hc|eXaQe-+TGw
zvpX%BeHr|OoU`BOPjA=UhInd3fNNFMIrS<wSJJKwGI?PbQ7(AadKMXA_BRhYQ!C3>
z{CZk|_a~q$gw@>!A0@1ayAJt^`qrgXitWSB^oq~0G5OCd_m7PwV+OZ{d83_qtxl3R
zzsb!53{kE^cUt{N)4z`0q2X}7_dfYaFH@VoRZA>d%^>;sq$fmmh8d^WZiIAxW<{o+
z{Y%p%OZGC1mzbVOF5H^9lhfS*dt-1>vF#>VeL8J#EHdfpeY{$HzEC;bcLnv%Dl%+H
zvdjl|Whx>om>2i%op?6Yn*RqktDR+QZ!d^GyJxbBiHYq!W>g0NFDL^*^PRNfs80eL
zK@l^_V8Z%V_~dgoA2Qg|GLKT2qHI7<K*2?H5PT#N<3h`^MR=1A*%<aiWBRA?g2Hjx
z{VFaeY~Hqh3D`6emrfL3CK_7%_SG)D=|kc59{OCb`P1n(XKC^Z0My@q^M?Ouonp^m
zHClj0L&M5A#7nj`eZ8<KQl*1{8JyF#!$+qZY14I`%rcgmIP|ID#p&^4xVWAX6&vgy
zpo(a$9eXce-uX85*RSV0dG<$JBhN`y2Y3(Y<|rQ$F4G`Q&C_cz8l33uJ6&}zdj4pe
zzvN=9yo4sUYjeeV%PS`Yuk#faGDO%FdLkL$x8K+c8*u)-bE(cqbD11`>$MJBOYEkr
zImq+#4oo1RHZ{ZiXza0!(*lXbMy^p#fKlMOAlHGqf6iupWui6MWuU`?K(mbpP?nu_
zF<1Jwx0nfSng3pZJzQGI_66YjVRpDOFC^2UjR!~o6odgsawApSXsLvGgPU9k;&QTX
zCTkEj>QJXJ)%REkQwif{muh-LGs$uKWm@~lk@{>d3v2vQHrh{M`e>VoGaI|GfI){q
z1-VuAN+_LvhGFX^);AOBOJm=0x3&VbWuARBI1R+NL9b@HD^qIbxPgy&F|<_Vx0$K_
zj!(DH?%|b_Bkq#yifBk8!aaq4nUgUmva2gMYoS{JFQn5o)+|#aGg!p#`9Z#YmkjLy
z*n$#5XIb<cElM=aH7vOpT@=83op2lC`UxvnFXVLrUTjZgyAFv=($RvdxCGl{iiqPC
zs^|;p#!o25V!i6uJb?IW8ZoVsE;8y@3za-kJGOTJ-$k#1{&ye?a;XV`u^+olsFMp(
z*#dHFaw4|@zHS97Sb@IUN-n{Yu3Q0m(>2B$^e?a!`n4+JF`WcT8S)fi)76ey09r^{
z%H(%yRXXeY@ckVDWy?IiNL%>3)S||YRdqStj&J)i%xx36S*tJ8C~Rae5rsJJg4VPk
zqRWY|>p1e9MFem=84*c)G?w)UHo%x%>3r!4!Y~67RNZbov1vYVuoKpNf2llsWw(^v
z{zY3s-}40>oyc2q`Bn0_IM+G8U7y~7aIn))^p+Eu+)=9YwzCcPiCp7XuPiiNexnt~
z49L})C}cTL1tw0+Bj(oi6q7I+!*hPU-u+tqM{PS?ML+>u_DAS1CV@%u(fe0YO}638
z{hVv`H=+H7oEyE@JGU;SbgTQCxxqG^=+`^y-am`c1y}u(W9ZE#eev&bMEQctVenXm
zWt8Ud%)72T^fY!TajY|tbsN$Q2i#FtWwbZouL-tUDZ(<u`)YS0+&~aIJ-T!Q&Y)Mu
ziZqIB*6N2)wwgjd`~|(v_-*)>0?(L5h?X)!$9E>RaR}Xl@wOa}YeK8>s8X<1thjjO
zu<-d?L(jk$h@NtufY1-fTjfH<p0?<OS;8_Z%XZAMw>S*p?uWb(UT{39>DnOXwdUL}
zbHWL0dP=+EIb4Mfz#f9y)$zZCIA)q|2o_EUct5{bCM!ibJx0`+K&QO4?9JgiUPYaI
z^{wP90`l~Pnzo+1m2=UqpaJ6Y^1ooil@uM$pw=-qskn^Fyk#BO>@SOSlT^a+qcpYm
z4O2k;f{_A_$hnW71Phq*Mi*nFQ|2Fv|F2-lprqCa*teCE;XOw5UbRs)F%S|G23I~Y
z1LAl9#$BKFO@^)GFc4_(LF8krfKw1QwaE^W{e1qYI!GSCM34YEp{HziJC>M{vO(uy
z62!_TWdT?X;xp1#q4{dJW4BOW>zf$M4+}793okkPk9n622>p%kx6S?-uqmdL5djcd
zpOO-1r1U{4!~`pR?S85gs)iGR!;i8eJG(l{Z}}{DAVR#X_-vDAZHGv~F!@hr$OqYB
zjm^aS;1TPuLez0pncd<zsx-(q7hhOY-FD=Q-<PY1?<#$kT6Rj+{ymH&5o&)WBY6o<
zFXVv<^n9$<QMM1%rgyS67h++8`@xf<Q`z&92a*Mux_ZWPm<Zr=QrddjxghO+ef_n}
z&ZQ)}9$k74^@ztpr(9HTDT02Dod?xmfdo~{+31sfIJn!1yt9}$hZ4_Nl+3?F{YUoF
zw8Tj+LuH}N8#q^Rz`tCCuPgbdUA|xXpl&kSA2I8=&XgtADnWYKVoGbQ3VP&L=74_7
zjF~{J(?aHoMcF>j!4-J^BvY$Mz3mR9e?m`F7D7A{R;`TSxH0twi3O+z)S4A~YJ$uE
znUj|LZmm3Odn&_%j3ot*5}5dsh1O4l-c7Xgxi3rDg<MU4|KlzEa$xz=&eh&gY%|?;
zK%+#-e^WLNv6&me!{Oul-v>|H{e$PBH!(>CH3wZx|9$DyJM-dY;m=~h0~%yE(z@O=
zo#d9hO&Iye$>h!1WSO!M9}d}aupBN8olMY|$?p{#+(9^0R|()h0Q0gDNrJ$JOCZ=s
zSs6aLBRym&d8t<-Ymzm=ADEv!pL&^%xo8I>UZ(T@q%Ivn7l!qIrKyaYfxyIhp<8k#
z^Gy}h8ocK`aML21O)-L9M>a0@{@N>%ShfpE$vG0uKNIrHhHO{@b%^w=QZ#D19?e^;
zm;dNAs3!L+GNF58KybPDU%=ZIB^e3~a+AQ8T~5b!(A$XQ>HhuNMV5xW@}JM9gj1$h
z*`#k2E7?}_jSvduwT2#IEc8tm9n+p&r&BFk#`POWPQtE90;02KQ@=Zx@%^!>DTiMA
zFZ%OqrI#72CU;ql$H>y$zT1CzHZVyUTw(?WLlX7Y>e3l+-RTMr@}i=~2KUYqovE7a
zlI#Mz1K_Wh&oN(mvEgy=v%nJXE;7X-P)yS6`Kg7hG+%+^zY9|Bn?6gT=j0!rL0K7+
zjz7E5@OSL?0MiQT7%A%XVp8#visPLxxsxX3huv_v|A)h2x2;3+A>Wenj+T;!zZ^#X
zH`TtOH483E_!htY?{4-Nq&eaL`82~k88%Oue2)i(UE1*2Qk*HICI8Y%zM3Oi)F`=d
z|8~;%TDS~SL|t1Yfi!TCWJMCAv7E5E^mHComPt$%{1IF60ExRtGVc+GO$vSHDvpou
zM5UbQQ98^i#snx+eqMXytD{3(ho#&`7OB=D`e<?^4wI{0Dbhu!h3+4YIg=(-l>=2J
zl1X1I<j8JJefkpdilpmxrVNurF;WY`2ZqW)*$GUKtQt9&X`CW{xj5SXd{OCjI<fPb
z-dQ@Al1S9m={1+Cl;MN8^Cf>H_eIasLD~^uWW?T0Nc)T5P~(AzrbXyPiJ@3;XqKN%
z!;cE_-7in}^Gcf;&R^0$C@C^ZlW?o~s=3h@PI3OZNq?<plT11;JpMZ+lIgB&!O3ds
zG`zz5bnj^dMKY##ItWi2?rGt)v0m*^|J`F!pri7Ulxq%hNW#Vdr^;1htG=#pceHUh
zwiv508Br6kyg=b4cO(a@{iCj=dwnPe@<c~JdHBVHq=i04r<RCz_?I^<Hpy;V(SiFp
zO4zd$w0$v(JjWlQl^GTf7hBoxhvLiptN1164FlJ<k6hNj>^wRfSMB*-yTCR2wzgG{
znKZ6)SlXyY^aOW%p|CPoZ=7eb5Ot3F@5fL7f4<%*I>|ZIimV9(5%n{Lu1G4Ti{IW~
zv?-R0imAiQEV*KP^VN0L9>vq8Zj-BonS_T@(lCPl@R}b!+l|j~=wh?9zpBy{{IENS
z#*775^7@f<adh#QaXgC^ji%6U=REAb+x;TxT|+=CHql~$LO*ObE2{T8DWo89->==|
zaYr4{iT7LHTw*Vk@HF{*U2yVq)_Y#m-dVD3|EmrRnYGtfT+BXLt1~+n!lg*UzqWh*
z)%zmK!=lFo2TNP4w>inrcXXBLNyB#=dt-b==c7#dXs`rh!GkQ5Oe2)MV(2Evh#0{%
zHaoaa4yv8~FpY=dFWB#Jz+?0rd-qlyr|om{<izexQSX+JW_HPheSMQEuF*kyCdYXH
z&(j~rwpf4SW9*M|w&h}@LwREHV>FMxM+b4Nw!WrzUtGU>FBCR%uAMkFcrA7c-wDBu
z-|WT~B1xqw7OJzLLwylGwa6pgvGUFF)L>m%6a5l$k`QhMTWpmP_mjzniEWEJdm1(l
zOncZgrxWZl_5JbfgVNMgmQDo{BYaHl1M#`-R4l4b28Uv?R+|&T?|vEn_@+AA>YB~d
zU>5KY?viJIy2bswa_qEm?)BeW4I_xxf9_m3yt!B+_E_&tXJ@C&ePt79sy_uL&8_d^
z;%{#~>No#34Dxm$JImxKvnewuBM`{C--^F;?{cOU*;(gyv{6&v+&p`MlYld}QrcVG
zPPCd`pZ<QL@RGnm4Vz0AVac=N6PApx=Spl?R!`>trty3?s0#KunH4$i$m8lW$daWz
zzjht&8^7}usoQx$W3x;D^6cZ}9X>q{ydFd+6Bm52!<VU2^t0zFU0Tc2OkPx>&KQ&(
za{$8+jIurTvdq=@k3~BRMf0kv4kEAmcG8pC?Ka9Cm=?bVJtONP+yeKMoiWsc05>ky
z>(e2X_@@@{uq$0obuG2ErW7YMd`oDG7C{B|iIW@oeF8%vBx~(m95U=cDYFKFAQ2-B
zgAAy92(2#D25e|g>%f5rlX(uEhGoE%hwaNm&x$d)km3rp-QgSV($8%JVF3l<{hW{2
z{w$_^8?7O6X+nL2ds%#^!Mv{es#!Ax^to8rcuFVE*yf?2wcZVeB0<G@FRT1{@x6wz
z*RvwR<71%uU$)dBhJAMB7|O?HKIr9q>H0;+&$iRKbHdj>Ev%2j{Jy`Lk}7sAr-PtH
za11ckkiHY~yOC5*QSKp(7jyYFD)~hA#27|aH+lu96-RDdy=l%??`M=Z*3~`wC8vp!
zUvWdqFI1B4JrdOKJY{#Pxo@g`TSlD-{(G~d{KSE~z4}X66pdBGPacCh-n!gZCR$0?
z>upNw!~T*6Z#CX>4J7>L<g4QQ%lkpJi!H=8$eAz!zj1Mjm6fiG>j7hMGMS8nJ=|r_
ztD(kiQnu5(3*0@mSxk;m1#UQvQ|kInY+#?+RyL`&xj6e0?s<@C*<M$cX?EvZ>Lcy>
z79G{I+#0AoKkM-ZoGM$&y=Y*kLbXAA{6G!!GsuY-N<wG2u!<%>7bAW(=ah$gv?}HS
z66=9s>O5Zc#KOb~0xN)o%vBW_*PWo52B3t`I>nyS`L@(;)rVoZe~Y+rQ=ZpU7A|(X
z$`zPVnns<IH)LS!qrdbKjipf}!OhM!`noj|Rd<vck!DJQ^b$4JG<I|tqq992Wx<y3
zH1!ujVnRxQDlQ%L4iU04t5<T$TJ@OWS$WmlUshjrz1%rj0n`sE__!)Ubt;TYL5ZZT
zvDpy*)YpZ-2p}y_B%U3zF2^6Ck9NKt5X(r>Ss<!MlUqbFoy_fD_L*xy?Oj0I2B{|z
ztjQl0emocKB0?u+_C(k7JfnTq<sDp}O&2MeO%RLK(q<NQ@j@EhL&>>AfaZ*z+;Zl`
zKQL#rT-02lj$KR5cU#Vo?|GUX=WMPrA|<=J?RsQ7#4AlnUvt@V6@8Ve5h-r=#(#vT
z<uPTnXh+?XY-ahdn`AS1xR_4vqrbrVg~G(8vN7KWzi@A6`F&BoKKS7wpC#qHc0ZB@
z_RV~Cf7bu=IrWakae;HbD}8kYfC<>NjeM&U5{mT3fBFu?D+W*R60CQ91Xu&GDLX0u
zhTB$|wwP*~C9HaW91So3q)vnh(gaDMms=`{AapIQdQM8FL&i9`#yX}g6?v?zTp%pR
zM-Qp1Eqi{DM35MKU((1TzBK6xLFbF|v#%2Qdpwb-_CEjHI?GYquM}N_qDC};J2!@m
z{xLaL1EB_!vGb!j{e%AKc?TB9sYW)KivM_tu(qU;v<g2(F-G8QaKCvnBbCbtykUlt
zuS<qJda&|S=G0wc;=3L*Gx}O@TRT&i$<EC==K`xGgi_6ngLeiF4<_jWcbB)iTQQg-
znsax@ZJ`tO1^M*<qv|WeqKw+DpJ7PpP=*F6K{};}kWfkxq@;#!$r&UDDMd;^x>KZE
zbU>uLTe_Q}qy)b4JJ<D|_w3*EXJ+=k_rBL!_gbY>9h{6hAvd5e(^hOIZG8%VPi3=d
zEJpWU67ncE7OQROg-+3U#0a%fuk=T7a}6U1?t$F}hn%|R?i1)2{^B19*HA-3b6(v)
z5x^E1PxAGCpWofzn=Fq}*Ie>UHku`q8`50$%;E5m-^<KB2jv37fZMSmlx^qPBe(|Z
zZv1jjqLtKcHP^kVGIb3Yqw(yyv<JXFi<8{i_VJj^F>_?adzQ17=}G$u;Sz}FNJfi#
z@e|XK^pM<;^whQ^GnI2eMbNwa0(Emn`xL@8l?*ic+1Q)q1tkqIZ2`CRP`~;ppi>Nn
z{rc8O^Ml(d;}*+%JN$Ep^>N<Q<-f{);omynzLR(=qb5%?*JjYTiRG%~6ZKtNtB+rG
z=x=UjzxB>gom-Un)%#d~szKw<IMGKShri}8a!-=x-=_>E^x%rF$=_9hG=dk}uH$dB
zGoa87x^rOu7~-Q5JjHO5>ZkZ7>YQ!iz3}ry)t-~PPmTOCtX>6u@2D&V4lt&?38qcH
zIIsS@h&lc%VmcEO6Yi%YDt#E~a)Zzsfoo1#!-CNP?MY=V-<DOZwZ3&-XiwD`XC`g@
zv&a1&_^vUH4#`H3^df=<&3H>H&tlPg$;IcNaQW{~s;wApT?UQZs{(xm&c0=N<p~IZ
zC@^>8E|do40o-%K2BF)~9kJ|8k%s2b`a+LgKG9^(d)nD$v~}>-{?adE&}k{D`E2lc
zJR=*DJm#E20Gt0|q%)Ca|1o0<xwBc+M5C^6L_rsfAdmh9=P`~tb4y>Y!{jdt`}lib
z$c#_!@jpb%e{;F-g-yHW_!d8JHK<+K?qcFqAMcbY3=z6xGOf3e{$ysh@8Kb#YK2#M
zik#L0;u?H8X8Pl~PN|VB;*OQ(+vY*sVpS8u6a=*=%+z^qGZ>eW70801b7nKV`Oy<~
zC&Sv^bm<`M?7{Wd?yPR{JJC?m{YDaVf#?vJ5H8P4*F{J10{^QVhCDnof3?D&H%4w8
zqjwiS`?4vuYKE0MncT#h9y)m8?G5w8BSy?lyy(RS2P&I0UK-i&Qo5U%n(@=jYVK>M
z3HQgW?9hVg(l%^nrC*!sIxa;hVVrAvAi_{0kN8WCltw69lWD~|)>Sf{2_D->QK@s7
zouIR*medWO*%;Q&7XG9{O{}H{UMW_mfXazyX^KzFeeVZ0`xvKkFA)9Yt{+bA8>TGW
zn$sBA1_R4SX}N?hKWqw6V(c#m-f)Q_!B^Xa$1!*&ED0-b@k){J`23SOsQewLPVdl?
z*iMbRCCvO7D`bzBK!AOT>O|I0`OD)W7=HIwgF6FdIk|GrlTRGFi@Ye6N3m`unescz
zZ<8j4vMBl$BdZ=VeN|-Oin$JYb;^s5woS2>{Cd;tb(OW5=^5g8tS(f<qtsA7c$IS1
z(XIV-z}8OUZ#FcL;`=kkwWtPIK)by8F?QwXAqoa$of2J+P$$U$tXi1g>rGyl_eDV(
zvm}%=Pw#ttI1f$>my|n|uyw)YhHfY0{V^RwAzlshVT`A!G%hvyWLL<z%<B7R$s?@s
zNn}okm-_quvvwO2|5a;Ii&kuj{n|0iz@02X`TP9*!d`Yv0GdjccbL*CBq78iayd%=
zBl-I`@9_x2_!oqV9jswi!c_udN4btB%guNSwYsbjKpvmT!W@z~)|+R8wD}@<PsGy5
z(hf@)Kc3Z$$p4`2Elv#GEAMI$6F^W&te7bJ#b7NR#hIj^TD2e}JfHpCt`tTDe(B`Q
zt+A*_3?n+28A+y{hoA=Y@XUi>QJjZm<0a!ewYhOflbWkuFoa1i7`HHPGOSL{=4C?8
zT9ifsD#1x=hwGjsv{^QiaI7uF2c2i3R)u<VrCrQxYBi<c9JpN<S3Q<#vz9l<{#3|(
znLcH-@kh|d{bw6J>E$x(C*$fR9(cSZj{qMT?!=zX6I(TeO}+Y7b5+0baVnc$86)j)
zdh+sT(+((oPp!CiT_NJgVNPf<M(T!Qa?E>Sf79_5p*Pux@v)u0{gtAcRo3ts#snW<
zqOw3nlC#7kG5co;J6lQW8%Z}$g|k5A)4NgX5GgFrWS@CY>FtkXwZKOp4*BrdiU)nK
zt>%N^y*3#T2d*Ct;f+;fN*~-q=sZu7lMET|hlap(DcU8@EHhUKRAc~)wH`lKUe@PQ
zz<fUphzT8ZpquZO;ui$yJnnx}oc}_~n7=&ha7tk63*~4AGYW4a<&fOcSINe;<oIf`
zEtcj?{=hd2z7^}<=pH2Rl4i3k<>{O0BI#FtqX8R_cXGCW@|V=Yt|IDIw@7<8Vr-Lb
zo7wbX{?JC345>{$)dgp-bZ5<-^V&4WNC}VA>*_D3lbcB@7DkBAYxZ=&#u$Fm1PV?K
zSb>ad5t*_zO&+6lL&iO?`g)x;o}f-HKu90_mU7KSxfa9OV4pRcvYOJvWY&Jp9q3OT
zk{hjzA6EG@MhK)9Ab7dNGNo4met2YMMW1Z^S6luXzIY_#KUAaM{n)waHHsDO*}gky
zo|A~@*7iGDawL45`l3+HY2nzk<NDJzZ&RWo;#An5>2|v1I$1(l>cV3E;K!j@%s1^)
z_J$sgwC<WA*~c<({}7e?BL1^(5|5#Sjs?n5wAPNqdir0h%5c-ayh%x>{q!s8HS1;Y
zd2Q^sD6p%zLOH<6=Q$B-2ZB`Qf;{`V6^aGk3FZb7<m1Ivbw22VWQX`$s9R=(f0JEG
z9dc98D#MvNxs(N!#nvTUizIpTCL@g%wvetCPiV~X^6rz0aSMOZs4we#nt`MOaJ{AT
zorwN&S-&`WGNvn>O+39VVah*#-xc*#PZXDxIMYsYM_$UX?cU-2C5Gjg5MGK@+_mm!
zmR&P132m{_`F3sJVu*~ko1pj)!n!K?2(C&#F31HKHl7(@HkkE&S&gm?AeeZ&>%4v{
z7(@5XsV3jR;b+F9s5B>sFyG_zcy`3t09)O2YOp*sB-3zvqeoW+R#>1h&>buGW8dZ`
za9&E|Fzh>DSz|Q))h=Q_e{;#iq74AJ$Z|MnKR0BW)Y&WD1sQkiC7C`l^RMkSJ}5Yf
z&o`?$zzszMxp&3Ft0d`w0U!-{w2(k?1cEU}O+`Th&T=qPb<nsQ(@Ac1^9Yn{e_;3x
z*W{_6N&x)$O@=!Sp~6eh2jqPy*)V=>Ckh8^xvGaym}tT(vzjE8P|V6!L&1U0NmF1j
zU6`|&wJW52#k$8ofi%C60b~B#dI9!FlL#A>-2m+mW%jcd`Uyw*CmJVt<IhHR>Y7}F
z0?kP<q_Md1CcQ`UX^Y5Jcw>(TQG1Rn$OquMir)kaX!t@QXFK_u9%qss9~%2lket**
zys-Ycl@#C$<><3aq-@IrEz3~ZA2KyUM5$wZvnun**$mM`ktS+%cN2;l8F>2mhH31N
zcaPh0-Jd->hL#*@M^CXsswKHqm8`LV^08M8#1T?t1W&-(pK!?tFy$u6cZ{>2x---_
zUe0_9nVQ%I`4r|u0Yp{-d$ZS?y90zs(X*8%+!VY6J3aV*MVa`bpKcf_6CYFH-#2bg
z^D8U0!-(lPY}4O_S=rd&zH0j42U5Ue%=^fio1Z^vx8kR%awLSso`%K-yNmmA{#}cF
zb)TO2vtP>d6R{kTrGroqM`7qd@uvuuh!;_nYJz@(=7R5z-nD6W>JrA#pu7ayk9d}!
zh`@W|PXS>l5DP$&pNRbOWO<-5f<ywqc-e3@z*^d%m;8xCIB#Mjs9G($<nAT2yzf_<
z8@Q&02#n4{&LXcPRoso(^*fF)5gals`8*0<7M>VN&DiOSZ-$hHtN;MI@AQ~^J1yAa
zM=OAX`!=IKs5yI}z?aE15bkKvJv1VHZByi!`JrBx294!rps(wCmWvle0ziAQ@16Fj
zC5dawnI@`3RL(KTom6l69pdRMAJ?bm6Wd;ti%bCz*Hf{K+YBn7LtaAThIpFKdZVnu
z^Q{pe;rYHKwL^q(fuw}qhRgW#h%q6n?P~FY`jrgkDdJb*1i=Mu7r%a`=)zpyji)ED
zd}Q^StS`NioPn=A;T3;6y5mqF^w`+cK%XU^Us6x^$E%@i<HPv9f%;mBAaMvLGOgy*
zWLAI;B^HY2h=hCeGQ@|@BkXJ}nX{kV{E=fL=qTA^zUI_aLs3r1Gz2B|6Hut}%mj4A
zCaYGzDQsG3H6;lJNx+|h#Dm1wez{~LpMam9MPb5|jPXeyBL(oo-qPqwOM)!=gNEB8
z-qL6xnJ>rOTLfC(ZvvGXzcEG2Ia>_)*x0gWp1JaiU<>@+YeXVc6>MUuN%^erL`f$3
zu0?g0VWP21aES17gmhANAf|fDb+B02@sONKO}1Y}fqPrAx1&Zc-jFsEff^Qh`#8-O
z7nB9^N_d|2hG81-Z?5T;T+ev@2z86H^eqqpXmln?w8+3Y;#d!<m9pN*orhSbQ0Hd;
zpmGuWB$s>O_+uf?6O*Ucr${QpR`+0Kw{|h$`kx`O-GV#v1^<6m#iAx-HzI{YGu>{N
zP$NT9&)wafjW%kQgH(l><{_z^v_7$S!?a|WWT)T~J}4D<0mW<Oi$A4=&4&JN9EL6d
zHi95Ms5F!~%p1bjJ?mm~DlYPkfy(Pybvzp`YbLa`$J$(J3>Lfk8GPFErR(DdEh5C;
zb12{z<Gb<(k9#{M!^JIHlKlI<+F!zti0;~+lL%lX0JvDY1}`(J@z3)!p~^@cG?ptb
zluRH*ax`E^Q@B_T2|~WJr4nFk2d%|Wv+P?i)VMhdQTOh8AGQaRWMjjzaYt`2JvO@a
zemJ`#LL^sr=<;Fu(X|Vfg!bysld8336Kh))dglEISwtuR1q6^Amm*x?*NXk4;>`Q^
zbJ*_W<mS1bZ*)Re;`CY8LoqLNk8sh6Z6y87MTK+0n;9WP^y1w!HvXY<%KFs0nL_v3
zKgKf>vsP*QYpp7oLM<Jg_&fnx-yVB}zbVKfpeZ0UG>)xnNcM%R36cQX=G3cr$(H>{
z6&f`)+};Qpm8KxSWZ3)WoEUszt8@{<T}~UWG}+`Xo3F-0{hcxch%3PdCJho5u(niN
zq)7|anZ;bP58U}vrtLcvI$>4T2IOUau-Xx(LT_1GtdwdBHXZRE6)r-y^s{C*kXZ`y
z!xI6=wV@g{##;7Cpx@oG;|1RAU|<VRNEh1>1vF~2(X@~C^gfQ7KBfP=i|7C>_PB<r
zL*vM!NE%_Qil#r+L)xecmv76sjZ5zc^FI$N{%7y-|Ig-83#Pqg3b@E2q7hO`<S~4B
zd3jmi$GuB}nR)Wsu>X{q$(Vn;NJ;6ZNQT8shz>$B+WC`Vp~=HaK)O#hBK^L-9f+;F
z_`50-2#C))#<KHijb_3B9tqj>L!?yZ-~LH#Sn!{U&VpY(-CnGy*lWBt57v<$%{lLz
zS`6TKMeR?%m-<_Qr@A@FK#V_7@+j^mOLn&WF0i(~asowC?A|Un+mptG;(jejJ<V6a
z6eu?6?I*LbmA4jr{@VY$_b!L9?SPOL$L#yo0D}8+-bf=qVYTY=)1!#PO?XVPmwnx7
zs{ioK?n^(pw}_jCTJv96|FVvqWEu_n_p$o?_J222>=!n*kw{I8{`U`~e9P&M@!S1c
z-O-%H_nMoFiwl=_+8Xl_EetJUtM=Su8F5!Kt8GO=ia^3N5ZIwhMC7BinV*JcCu1Up
zn1+hb^h2D{5+8@+9sh^rcxkN5OGwQl1Zgn**<8KIC`A{k$ijBDRrBVt#oRw&W}Z-r
zCQ`AYAKoW>_tR*y53$hUsP-aCTI~2S_|5fU<Hui5dYNtxuV1x35wSVl`pnNf3t8mP
zS*aLKvWW94h<Ww*d;l?0mKbnjm?j;0eYP&J$Bkq1N9A6i?aJM1d?~L4vzR)mHt86!
zI)$RXa4`+@?e~YXhYe>Qw!*Bo&(e371x>$Oq+Yx|I$gUS7EUN>@fBOfA6v?y3FJRV
zcU`?V!P%m`r7!GpS(L|huKQeTKh`-&zP~0TcTiGcwLEG)iTi(k!3D*^zR4J=f3rq)
zQ>O8JufI>HDt>&x8|U0Gug{3_OEq@w-ff^}bXN=~As$rHls<E6{JTO(<X2YwYQay7
zLo?H|3;5xGK0OmX(%Ne|vss+goT7-Dyk8%oEG?m$>9@|+(=!;Sx$jPN5axS6@AFFP
z&Wx&JurQmRe@Xx++sfaKnqT%9o5<Y^m1Om|iC&j4JN!BF8I%MIpA{Nqom(xrWqqs|
z@*B$&DH!s-It2%z<KpF?=db14>J`*ZqxRuJn8t-8{?i$^pzC5vFDqRlh0q>;PKgmQ
zV~`kSwdS-legL2o0-=oVi>?dN9Fs3?=-$AuWg=UbeSN{_cvQve@bb{&wcKns?&FYS
zouX%DbUBA1vAQHupQ{tukP#U=N$n0i2^sE&yPxWHY{gtzP?-WSv8?QfkTLLFuj6aD
z@a`<Rjk*-U$0C=~s4huFLdRe+VV<bB`WI=b!O_6I3kSMxrMcPQ5MhsuIuNzv;xJu;
zN*aEK4Nub4#LV}1Yd8t4ehXI*A*x9FRZags5#Do3u>k|ho&G1FGdpoud-97vc}cIC
zg-rtnkMZl)y?^&D(H$lziXX59I;1r%Ah!PlZ?2#9b%Q1N|AognZ8jlJHhT1po!JSg
z`D)#YFo6!*7|D%LPjBx%9p9V8fBe<n#im@J9|Nc7mPL3^yq2^@*HJ|Hn-)ohr>mmw
ztp+p_x64a=cNKM^Ww{!+bId7p1D`c~Jxz(5U%y%Z?7qGJYMRy&|2>&6g?m3aKRfeN
z{1^<<hlf4BEXmx1niUaryaqXZ8>Dqjom;|-{djhXgtQP%MuTOS^wuWbpI4t@_t3<)
zb&|BuiD4CsejCYNBvUAruz%6m)1#xK>nIRXcQE}oxEf$uQ)Ku;3B`x!7u6qutwBy`
zbF6Q}%Cn$aEZ~KT82Pf)B$X8<)d7ytr^Le^x}|mCQwgCzG=qG1sp=0)<%a;GsU_GZ
zCZdy09;ognn-(q+vO!16TFS`w+R{%qiOBCNA=b1q0L8jet)w++WojkI{MSYiYq44)
zg*SZ%y+#3%qGuFkH{&^qMqh(ij%`&8?(t7b3jK+^Z(YSjn3d3x8iADYM(o?BHHyq?
zHbly-W6gir?2m9T23O}rgAgZOg^Ye$h*L`c`~d@kBBZOcuC7!1lfzFbHn5gL=2FV;
zCvI}Fv9i`>@?yowxn4QZbf85MUPQGli%0aEO~y$&D-^5Imb6COxeF}%zL9=L8EKFI
zT8~`goMPZe*B<iub#h?;>)BQt+Wew`we?8~&$QrSZ)WygBB6!StnMc@%lQ6OYM#5q
zHXQ6~g(PclD^NTqj2Oemp_DQS$pR}Y)ry}CoIbCmOm63z&m0zV{+o1*mX38?w!Ha<
z@m)|ElsPxSOo;lP#kn?Lb}n^bipFEi$X7t@aRa4p!d0o3MXH-ncyY?S<H9}j7Pocx
zzk(~b7jI`$C>{Y{m>Jyg52+*lseJh?s9@u9OkpxDL2N9Z%Rh#}5r-h@4MXg$VYvsF
zcza8N@hQo6i+Yn&#*KU9&m_!Z==jg)yc(o$nDEB#{qX)g^fQZl+u;jrdiCgJ<6}in
zZs^Co>u4MX>zJGKMD77fEk-E5lojLE=??PXuB4TeG)`<xoODqC)s*5!FS;h0Xi(|O
zxhBKpq`fANT1~Km1)JUyOoAjJbE*P;5DB(HMo;#rP4Z7p&6a$$6vgeIj$|g<1{(=*
zoPvf~=d6MQ(sd1;*D>^sObaQXw#EDOPN)+q{0;}iw%7SAwfMfCn72&zsI|L&s61Kc
z@BOuV!-g!2d=-`Rh#4+V<)?MO$h~!6Q<Mmk3J)yIj_dty|9hDu&23!UQ%@jPRXci&
z(HL0*d`PaHHjLnz2_M_ZQr5p{`+Cg^l8*7f6*h0D2cLo}ZS4HHf={9BOAOjPFQ2HS
zyn+W~vh`4w0YLIJPeIm(S&=>Zdn@?kC3u+a#?M)N^{E*aviqatSPZ<41+`}<zL_vI
zK|g?p-+}uGL+)_Gg9W>SG-`v7iB0W|W_xSm3cALFt2%_Ejtc$9<$roJ+Ni<nsz%fw
z_h`Nujz=iSX;i#V^<fZ*Yb1T&<~We~d_S1S8P5y*A=Y{FW9c8YA8Msvjep*xGgtnd
zY>e?cf9xN*$&yX_MOE@DY?F2_z1+Vg-D;I|W9%UVOaXJ59zpXL*4m_kB(lrI9wn9S
z+eCv2iwPI2>@~v{RDK|mhVlCE|9+&rtNatvKe79lck;69hsD9^_Hj}$Nu1$W;m+$4
z&lXIK=F=ooj^ECrkNyw!<Jb#sd{Dl!V%fO!wPX$BZ=RGX^Y+<P`RnzY^Ti~T_x{@8
z^+8gT&*E!KNRmUcr6(-bat|a4&_b#K&Pe-aTOk@HomDazdE1j*WfnK=;QL@|EedAv
zqJ@Cvg=GsX>H|OxX@VpHAe8HaUBUIGf?)X=H0=uh5^{<n#|TmZAPZ7C)7X)-xelk4
zBwgogzXfG1J9)!=Gif&=4dWmK`AZ83U8g{pc$f%_LoM&;`t72!&HStutgCmv!+5n=
zhxg4RgC2@z2WePb1GE_v=uHwpy*z+)DaTb6gv{|z_^I?$1&&=1#l+cTT~J+n_uzQE
z!?thkP1bK8<8-&LrxFKLzn#Ud(<<0%a_7obYTVb-80jkJq7uYj49<-n7FZtLByMbu
zUbX6T_9L{LQ?9cais}y^T}vOO{^Gh~APeRs{ZS+&cJkFy><=!hms1c&?R=2BdE73{
zoCYRXz{BldJ#WLh*YaDfOB7o(5_A$gMVG0%W>$WZt;ja9M6#NZm1ueL;qDvtp<uSr
zTMe)sC?S?9(|50u@R3FOj9SvsGCjY*4*pAar+6}zUMof&QgeM<Of)g*V!b6NxD+mQ
zWTMT4yyyQ5<&f8PWjQJF$`v5%23HEYikDG)?c{pcWlknt0bk({dZt(yyj>=2HRT$U
zyn&y|L*pMBEg9{N%&7Th>^w-aJ8I%G(w&BAhnW0H&<~qzT!(T?YEG2%2iLS@wp=?&
z4n;aYoC8zaX3Si<{b<8%t&0Wdm&uR59$F~C7$r>B^1<&3tISPhVe74D7s=?|lL_t@
zHW&VXkFKZ+9bJ5P<O6dG50>9<XeG<uF4XlDm`!Zdoju1`<~y&rU2PZrZoF~J+1An3
zwceYqa_?pd|35EVvsaAWb#5A2vad0axOU2uH@T-<BZ_h37q`pdO!;0zvenzfK5BP-
z8U0}FRr(leg>9rRq?07!i$xKxSeV!no<ZwnZ3H(M2O5&{dBU5?ygdK=r1T5mA(9*Q
z25krG^c;GY^o91!CM3MGIUtTH#}HD;AR?!5&(=N@$s392#lZ&f4SwI74#uOT{E@U_
zn7=p_rQc`Mn4<3zHA%yN-~1kwyyc1UPdwDJ8cr!V({qryJvz1qw~sE5y6t{UnZ!u)
zeE23Y0+UA}EL>ScAWzj2JIUpu!F^TQ^oRKDwZ3u53<sne6z0b7|NIAti@a=X#ULsu
zUcD*sSKRt&BfMF}n<sWO(y52%^~0(C18-HQvEmwf|MQ;(^T*bsUpY7koM$7H9_mb$
z{RpC1!p}mg&R$>k3m!X<gnJ(j6?<V`EVj+kFRM@D*CT!$e7^Bs9LT0bhc(+W2r!cc
zcL<(DNZ>Hx2c>7u&YFZ!U^yjI=i|r%ooKt(>j-I8(sWB%Az5?W7Ok*B$4;;T`8&4G
zxJ&D<)M&5nM|2ncoa<p|d`Jl7r)2Z+9jN5Xm*-WIL>4?Lcz)FHc0}a~roOdu+FxYC
zO4?}c6+>e`g!tJBKDJOtDa5v&)1UaE)gCKs-7mB-)z&)pTCgL){bdKf8?k9`)Ybmh
zi>zfF(~6{iweO#OQ!r!1wDhY3K2rWgHY>!-on&~;w3k)tcr)e?+V41BSEZBMRRGF&
zDQFW}@h4<m?fv7$6M7IyP;&oAOMlxXweD&DvCA>;ASPUlrbOvj{afV%i9HuC*?_Rd
ziG?3x^9BC~msbHPGRLHETpCf>m~>97%XZqohM4?45I=w{+ZdBy)a}0%@-Nb+&h%P(
zw4I{(d6siV-uJ4~$2+X)5Kdh`aZYX-CG(bZQgqtjSY4N?xAn)Tb8@;PU<t9v60nzC
zG=(VAxVT_#ybE)*{u5Hc{&bPF&L~SiintmRBGHEl>EL0I^sIdQzi=NaKnB4)_3r7n
z9iK&i&VTVe2?+_^7-64-K=jY_vif*J*T&b0*6YurXd_FjLXM)yKLka-ApC3hGMKUw
z7ye9dl}b?vBR!HqmnlT)_1m_7N4|!FBE2JhZUw@Q9<8MQ0v}BH5qwf)ufj|kV;KM_
zzy-1fghUa&@8HQR_f`MKf($yvr5=ZCjPeJukALz7IoaWzwSVc@jd2s1#jXW~ydVL{
zNo7!%0CtPG!mPmiAQ_UYC&R(yF-45Hq82_-CTY+%-Ryf#3ODc);0b<c-XBlR;sMtO
zkt+=Gkm1|+gx|If7sYH*FU@XqW;}v1`My8+<h_5q-b?55x#@h`P||D1?k2wno1YU2
z?B}}9^jq8&#nit|WnUnA*-T?(Htq0Y2Hv<89UzqKwFaNJyBVIXE-ZtGAE_O&NJoUf
zZXt_KevG?b>s(9z9!pw29sFX8KPi4!D;g=@LUqNF4uf-GBI`-7=moXrIB?W(vALL9
z6e$&Als70(h(Vc14rCjSp5QeJGqGx}Og|0O6@vRAoZ0c8YX;8x2@xpBj=IdMv>3cO
z9&Sq#E5UE%Y?@VzqxI}6IH36T@&>_4k_16EXoIZ$g9;-eEYoS@tngKL*6oUfQ{D{-
zpOyGV<#G)Z>?2PH=Rm7kHB)2~MCJypJ!Mc+T&z9<PBH~cN7O|y_A^8HZCFa|46x2&
z9S+2-aM6$;4c3zF?n#s_%X0fT2(7_b^B6YzSZcnLfz%LlOKLw5_{yei_G8%KATLu!
zROS$8@)N5}llto~QVUh5ye&ke8kzGL(VxC+fIy3GZ1Z{-0ii@$(fpud!?*AHy>(4=
zn^rp~dy}<!hg>B_gv?D*RnekTfsZ~KnVk;w2qZrkt1{P6*il7<bR;?WLtL-oF&eUi
zFNYSWP(+dUXn;8pjXspKhFdPq!C#=H`q{nFhm*=Ekz%cxdGYSkDboaOXO(;}<6>U^
z&*d4CJ@%WKoGPRY+>edeXZ_BO|Ff>V)`Cn3^_yrD{`o$dyy>`)+wf-^MtZqZQb9CG
zjLGQLl-T#|>F+n#RtpmV1jmE$;Qq`={4h~v(Zx@)*j$i7df2d3#e*b9*3`2Z;B`(B
z$wOj<+dVem9Xfp3QQCso!tYq7wNe>UyGvo-zWe~sv(<fyosslJbryILMcf#Co9i>B
zN8+p$CU6!Axg^{i_<TU;Y&A6^*u}_s$9R)nN1WX%!vX>M0m{lZ{b>^IB?Gks1?ars
zcQs6o-Cty@^@dDROpNTkgth8taZBjRaQ{NykIMp?%nwHJl!p1Cv?vHjxSz|r1KSCB
z$n7w1<UIE=iQK=%2ETvEEnGADC}idB2E<yTLd&ge`nyX=pF5EsI5mt>3o0YsoNVhi
zWP%F;?bvnsU|6SFELre}X(wRmo(V5es~#tBMZ&;uD?^4N2g67_=X-4*Qo!&Ctj}Do
z?uG?K4r$Kl!?y!jqLV-<K)0mX7~r6f8w|Xyd0b*WHm9EP_OX>OqHf}&r;ESHKDmZE
z6QyjQD6`Mqq~R=oRBB*>Q(XLZc6#P&8BTg$`6&&Q7Ofcw_5l<^(4%s-QFVjJK<Ukm
z4BE0PlmRS843zubz(B~c6x0{*bR*J<JYqY2An?o|ln5wfdR>8HjCY<J6IVMn_XjEk
z#)dWL4dQ2~<lEcht+u1GNl>~5Q=|eZ<y}N<ClTIg8DxFN9~#%%hh<ttkuOUezsb_s
zpFTKQv1SyaZk;y)9ZNg*JMKhpQSXF{^PX*ngD~zkl0wSk@wUQ>V=o<zu-#bApd4B0
z!m@w<;x1+FmOk(DE4<r!dt5{cH@$fCu8MaYw$sUUBU4bT;9#p6U!p0k`Sl~UnvPAP
z$jhhJp&N}iEaA^LMgo&w`+9!P`#}vkJ^9a`J-%9B$86dMESmp3m?CLKpWVB6&+1>Y
z>eX(=?^y-pcMN?mp?dlI{fC}YqMVsozzDxY;HBWR7x-X_O^5+9rN04!hFn9oAa!fu
z%S2ERMI0jp<ko&69wR2|cQp4V=R3TMy*TK=yS`Q$+!9=+aSNZ;W-=%II1TH88N`c2
z!J)fbBOC`$Ei&f->6~a5`d+Q)pD_w-dH>f8S>IRn5trq^-@R3;zwN{1C+LYA`9^xT
zJUGjK4gEgFk^Z&Vgz1LjUupH4($|p{(fhhH7o7lb%*t38HkQqV`H)$3L-EUX&-u2=
z!FH!oX;i?9jsD9NTQY#{l*8RIed0F<KHP4kD}U2eI!&xs-S#2P1@n8+Q)C_wGhQss
zR%>a>{j$UZ8vD_R=9mdpaTyy{&{dRV^`OaC3NAhjgGQ%$KW0oWC05B4v#mXRT+5e$
zi-PegZ`_BqegmU>9qCFB)&}N&rA1w?w$c!6dNL<!0gPVk?F~`&pZ&Hu6)3pk5n%Hb
z+TrkFmDY}wES%XSxIZW#Fl%8pzkkM}WlpYLYOu*Bj9npg73a5Igl+^K07(224z`S8
zx3a<kTh#HB4a4O|E0R(EqutSEwFfS#=4cT<nSE=gFQls_W`?x0IBu(m^^}b+(2nXy
z!!f=P-AA{0;7sz)Fu&R>??`xI+bUV@vD2v5#*62yD^!!jM|?Sp5<Z8Fony^L#TVyP
zH(ds~7k-lk=!{TRErN6~kv-k&M`|H~lwCM!YD3;1$|2Iq6p8+HmL%82SCE9ln4c-q
zhR>Lqv=aSqtJrFL;_D?jAK*z<FnW$LxSuZGa^T&sZ}{^cIvjzxIX;G#<2Y8++I2Db
zq$4_<f8PA5j~>Q|s%#y5=4IUT_r&0ph)+PFfV*NIJxHa}{s_McEtjDT#Y0J`hFcJa
zxc3~5x_2SjTj!r})Vik79<2@Wc;c1_`vx(<Yqj}uV1Ctw6c6OiEz{kZ7F6MipQk1b
zRe9Epx3UY1=%qiAk>M-S@Taj0J~-48IHV#B*M%{$HW?7VUc%i*^ZXhf(9leLs6kqp
zO<ejk^c4k~D7%v+Wju}Sse<-cowNczh@^B;Ne;*|5WW|~#Y~`Wl~VXiitver<J001
z)R*s*qpSC^Rpk>xtDKo1B1s_qb5gKLrxdT>Zidnp>9187>ZhOKnt0Z&`5ahw$}yTk
z)=nP$<g1A=Bha}KvV`G@!3~mbMN0ELnbP-{qUs4M3EY9iVU<$}{|}qwCm*Y@?NU_=
zOl{X_-fl45^ud9bTl^c$Xx!K~Bey;09h-Dx+vqAbt<IF@PQ{OZd%|A=&s*u$6n#5E
zE$#km1Wq{B7||>L0NHHV6!MES{~Re~M6rahVw$YI9ZbE`N^U)jciNHK61H1QJ%TqL
zGzQ<I<nPTT@h%ON)2woFw)3snPz!J0z*<!yJ&7h<M4%oc9?)-<etzhWCx8d{5#&tK
zMT&L$np&$bdXa2t8JSv;C7sV;khb7&Q0wfK!m=8t+oRj|-T1%0wne`$nvaHq-PbQS
zjTqBT7u79Co*CBKeWLe6|8j357FWfP276Nf1&5oNV%jeACd`EYF~(h1@9WpE`M+D_
z)Lx7{EKZium2{HTR+c@gP2ea<{!G#@#LwJ>M~1@&SONU#0FWwKCl2c2qt6*i1QV^g
zA8|SbkB|1u^YDmNrHzofAuXS}`g^$vaoWc0b?ige2bh>i;H{j>><B=FgRdfi*T^<C
zoQRGaaBFoN>~B}%(gP=*LK#~K-AV#o3O{#9z8p`I1d$2GHIke%GeMcpz#+@?aV6`p
zxi}_9ZD~FL3MJ3Rb>*2%CzNsiZmmSrzHIT-_n3JuaaekX#3Yvie(#O?`xT-xCO16N
zN5|l4v@Ke7tQudIQ4soYr$f?vuJ1C^ux4$`|2z@lv0?Wy{k!)*M<9ccQnbvUC)@M>
zth>88{0#h0N8*?}`11-|^-m=8d!PFJN%FpGEKD|zBv`>D<{8ns-Mdeki9!WD2Tdf$
zNXMg|108@al5Yxr3mhYjT0IcqShE*`NZVweL3;m~BU4LfD2mBYiuAB;Zk+9U`JE4g
z8}Hm&82V)<%|Z^!mMhEWk@d-(6?;&A2B$zPD+SpTl|{;bo;9emN#oIa%pAq|IlGiI
zDMSF88F-7};^C?FN&4S4#m`T=X7t6fDp~sGbj%&ocR#Pyi>saSV&b~ndR@%L6(=Tz
z`I@qumc_m1QN_43@5y9EH}C^3=ef+umy>~ffzv(^r(kybLD67Z>EL9k0^AiSB6<cY
z(TV4+jmD^zhaZJ^%wD!9v!|0PCb>1ITQ&UP2y^BfASHc*if}p!m8AZT^W|Ye^fm8a
zt-dZocH`ampXT(5jr~62d$bF5#NxPDCHvBq(Pdh_uggHEx1NwJH(}Y;bc~4dcJevn
zJ_~a$$^`1r$*j%bKW`3GBhFn`AeE=9F9)VDp>@`q89wzxq7!At;u3r9tZPy4_=6w+
zEEQw&*K@1&W)P{>zC%$*Pr35nE)!-+Wo|B!iy;SxNa0^Vvj}Q*`O@WPQ!pOS`LSP(
z4fDg&*#*w4gJjcXZ>J6ab4A4G6I7`W-OtK$#>`4*-8iQV@dYW&$b7kV8(C{fad;>h
z<#7q(q+9`I3@1JEzc4jSDq8#gQeU4V$<((vl0iyeNJ!|P{|JU3&zg=A`Zs*}FF{@Y
z*?%86Bg}!**3;MTHMGrOCzZX(URz)9!N72whp6LfrN`WBlfPZ!%WmDjc=itpef;e8
z6R?Bwp@fCTwcAf3KTs+N5^17(k~K~6RF{`SK_385Kb7eMMEdsmV5CVM?%2+<-{2|{
zjPhl@SXLAqmLL@DWR#D5gBdoCHSwB=2{O0_|9Fmfb1hnnvGpc1b-(*o^V8oKaYM1v
zMA_<YE^TV(cf0`la-;~_C8+)YHg!NP!kMaMSr);|AHO;{Ysg3k&QM_sbh&(K5xnh4
zuSCsFQm-*|#8Ao*yq;<xlM_4b=^6GZIH|VYN4@m_<(!zt<wO+piWh+{ZzF_|ilHJA
z{3{I~!;fZ}Z-=-5dMSsKJoUqcCicgVpEuuLInH1@;Pq6;@=pWeX4ft~{{_m+RbW8*
zoUdhNbzZ^K?I3XEPfVkwq@$y=mihMJ6N_TlMQqERe5m;W2HR}-nSc_Rx!Bkbb`3;$
zB?}lxMl#W-m%Vm8i6dO*&O*f^DNw0seXO`_vhP^JKyANP3omo92%P5=N67b(?a&1M
zpQERibS7?Q-)`x|E2EiKq!(U&LUm{T1xBFO%{XYlt~>r~M{OKjyo!>`aR&5Sbtu&v
zND^prrp?`in^V}R`p>Y7=^kp>cW~UV^L53*%5)!(?Wv{0+y@F=1r8t2Vl7$RC{Did
zq+?rZ+CC`oKV{!G%!>R@;Y-4dck9OQRkROJ|9(69W>{#9Vurb+j-f@#cXGHUVB&6M
zaU3Ux_t-ob&zySO`L$A$>g`;YAYGwA;s=6q9+YL{BFb7+Ye(Zw+h^26CLIrVuJ~1~
zY-!tyLB^THWoX)4VAh}^u9-7|fwbY?O?2Ii_J@UFPP%MfqJkpEkQhp=XJ$2w#ua;>
zzi(1}DRl;=b9tfSvpp9S&xWS|^bN#5<BcDDJX{+oxGvnBv&plceX|^c_p>e{BEpG!
zKM)hgG5X!haeKTtUOiQ$XW=HK>;Pkf^?wh6&c8#jO#bWYO(~aWJGn}-h4*38`D_>o
z#No$&y>aYlN#lzAJ&l8lRe0z^fFCIl{xW%&qi-j`tN>=JU>X!WePOqEqUM5R!~38=
zt^5EkFqkZ3e0z4(bUYLB!F#BJ<^QWI3O%P~n%I6QMmTr=N;O2X-l7??Zh@Zw6F?26
z5daSrkhz62YGDLqQEt}RL>A+vQ=#Drq1wAQA<y~8+*P#<L>xUv977DlEyg0jyYRZH
zW00xSC@8XZmvR`4A@0s`mMnRk!gX&LTU0U#gxOx>$A4j69ay=l#k1=Sd*DZ{Ahy#l
zaSl`gM)1hE#A_)fbfIOLa)av6UzO&lpyzNyPz-)jzUI8p7?!FYM!&3tw5Z-tV>lHv
zE+K!U`on2nP(v%rYor-;Q1y}13kFV{`K+gp>t)pjniK<@fZ7dMKQ$YDwHCz92V1gq
z#+?m9#AjRo+Dw^j2(EV1#=~cI@H(klLK|qEt$2|3;~;ZY14Odiu6*2-pN8Xt)hzQ;
zY7oY=+bwFLsLbn11>ZOLk6DITc2t0tfAXvO-~dY$c}tlht;N|v3jH+4@5x-YEUYHY
z%Y?p*HyM&AJm3E`5?d#mrPutzieX<aajtz0$Q)a!81(*1gT$cFm$tHkT{8no0ofVw
z`J?=JYD+%w&72hUq(K+&GI_T)^c?%_A@Ruv*pH;%t#7R50%op1ZoFv%CiHa7%=qyb
zJY${rgkhaTGmaHaRv0*VLv;VYdEbBJ##fAr;pxBN7vb(`i4BNw=fc}VnHifj=e2?)
z(<=>Pp-43%jrKJnr=AHaES_dO$#%LdfOu1;pv$4uG8xul359)|+-_=ti~@uLR*rN6
zs~O}UImxZDjFF5;@@(>{2VD;R%(&bD5mJostK+Gx*r)f?gC+U4^Tq3;UiuHVH}yL5
z{jnkO%pIwCj>^_0EF^2$O^^x5a(=-!Jy;G9+<WxVmXfR^GC;!-c(xlHq(uj40m;X+
zzF-d%Tc6H~qV%%jfY+iJ^?;3Qjl%ICk00__Kd1M=0%0nYEqGztW7EtGti|<@dF{mp
zdn$JqlT->+Y*Oh+_wkq+3D<=?WzHnitOQ8Wp1D~NEn|q;bqZ=gSo5iw@SjsrL{%}K
z0C;+}CH}+-oWE4u1p%o8zQ|F{U#a0CLVqT{S204gqeO{lC^<M0)i922DMw#`DlI`f
za0AzaDi&KCs}$rmUCWy5PwryZ`WJ^=sG{|>KQte^;SK&}p7-pleyt5cNAcsx;eZ`5
zwx#Jvz{dNNmDa~vn_o6kiaaQJKTObIKlC?l>b$YIJY97vODn)YWr7}=l2~l^kC!O5
zdul#>Nz-SA=;Co-$_*l1jP2>@4GOaukgE%W{ni8Vwh3>nGPQAAJ_qFS1FuR2Q?f6Q
zngLu2{h_(dOSZt#VcDII6Cwa68@iM@U}(qbub;o#Ti!!|f=8c!fD(pKXLEYx4b4o4
zHZSmq_stH`Tlyh1sU2if`!@*D{D!|(>C+&10C0DzrX>F|OQ1n)1c3zy-&U)EFOM1g
zp+CBrwz{T2@t+77UWYUTmO`8}pJ}xF3q?J=4jK{hxv;M{U(T>56CE>7bmEC9#z+DA
z&Zy0I0erETn1$x+?$N{MI))TN40-!uwK-t`k0{pIwMno3w~N%z#^C?F*{1k;Fi*B@
z{8#KEALjpzj*YGF)eO}OboIn@_L2NJWSLr&()sSSXQwP|VFmC(^`5){Uf}0J6ypv7
zY)D5xN2FSN(TT~M-Jz$V#JEDroTg78UCCRJe#mEdCi4=}E;oooNJOSkW(j%P!jEzn
zP$zD+quC#atV=>eV&x|Nd-W9|;3PM(NEZM*jNeogi4UdA@n-im%MD1%JGVHsyjZm;
zd`ghx0%?F!Utz8KJ~I?S=ROmGP`v|1{;3QUI6Z2~k>mKRZ}fzI*=z5+s&Ay5^t8Z*
zL{T)IwK9>If62DEV}^ZY=<P$Z7NQ(DJb2`liESRSwA%)rH94Ol!znBDq1vvQp^Rd4
z>Khg95;Ls<0^_GDQ9G=S`v4Eg_mw7ns*)-R-DF0=_W+q?8NYO)Cq9ufNRmcM?ACxR
zYQ_L$fE0K{RE!;sTtLe78#42>l<NNkdqUtNZ~K3!jcr+Hyj9b;QR>2dkB4tXPXj(E
z;6a83nWbG5;A(&>`^h(7C)#;f7R%a7)AbwgP%6MSC_Ml{Qe9*A*=a1l<RT}|*12_v
zpi1%yezuA+1h&D+XdWqXhKZRrJxaxJzwfrRnu}zGxl25wIaG&AMN;%=2!U5|@(i=d
zWnSjZSUz@rYwMr&U3X%rX;0E@Lu4JZ5+8bqkcz!oZUmI#qy3u4q>&oHCe6BDT#>(q
zMC8+N%Y<&I$3yQ;VPg-jGR}#1t6z{X^O;@azwbys(OhCc`Y>Hm#63gQVzR*eD&G(;
zAj#zRvA~{wT4vgEojB7S$#MLM(T09;B<mq%=UClX=nWQ4(W{q+^Cu!Q_Cc;KGK4b|
zyQL~Bh}sXw)1s+<h=Zi<bOIR?n!p9e@lS!?G_VbeZx$h9T?-nu;mu!0JQJQS3Q)r<
zmO%K~`L|AdBiZY^7Ne`L8_9sLNx+-haKCNa<CJn(m6}T-s2M>>-Fq5-5MOYJ;x3zA
zQ?s}mu<A^!h)WBtF%7s}_ul*OLf(TZ<X?`D-5d9+`_?e5M;u0z)h)4^&WaiL|M^$I
z>uWxfXFd_FJBrO(@{nCxSwKWkKv47d(>UUL7NwR}iC>p#ea!*P$S35eqe@aZ|B~ba
zo2tbP2g@7{%a7Btz-5UT?CxlLUL(q^kOD}c3I$Y;`hPv@aO=F_BS3jPqGpClK>Fag
zpugyTkdZk$jtPwoRg}Wd5GJLSv8lNyZHQK8agFpwf|Yr^a1H>i?lkWg)FBOIm%}RL
zbzz36d1O#MQywjpAOu_4oyqH7UsrP16-U#rr!pLYs9-x*f1K%-QCuZ2!I%<UutEYU
zb_+9JGx}mwQzu^B8U)=sS!P;%=sMsVzH|H4!DX7;v4_FUWHY-Au8g%YN9#I3&R@P!
z;5j>dD))4`PlSiFSv{JDzcc!>h3Nz@Mx~R!1#G@SnoKpwx}jV?$e&5PA-YQ9WU()l
z&By5P_DDhKhy)udL=7)Y4g}OUXotfPN|y=0ggb=0{-)?NiW+pxEz9=0XTM$qUGa=N
z-gaE|S6hJO6SB1?U$HWoKf*Zs4*22jRhiPWWDmlg_rHxqCI682i1fN3$tmxDJNiOA
zGdxT82sTVVQmE%KI?#1CuVuW&uXJP2dVa$5aFB+HnKPn-6gD`mOMfVT6%>$D0WYco
zi^>^2$<o{`!&6m4O?8K#Qy%KN<NO6LcYiz=1q86iQ8mbT)FwXZExKcKtriwjg{qfX
zVz#fi-y!8&QdS@L&!YyHgx+rPzo9Upe^A)h7sO2%BVChB|38M?Gc`r@zFRp$4&#pu
zO^vv1O+A{_X=-bDUadwkyd*eledko6OrQA_;tQFBz%c}80wf+1W6pw-BEyVMiv>dd
zKt4kXAm5+>Go+Ja*)xP5tNRh`+PiUsz4J;pty(z86pyb;>_MkX(?)pChgSq{WH71<
z@?%z}B^+F#*k~a|en<;$3+NN>6ObSHo@M*2<a6$Wwv$p*Uh{HYTj@hP@P4~MCsOQ7
z?*2M$*OgoexH0zrW!Fc?tGW*&c2R64A4GPqo4I&Pc@q6IX(`fgt{gO{xiJu3c9wJ3
z#D%P}@tuSL)}B?fkd!{63{>^RTboGkS;HV*ELCi;30^fMWYYHCB`Jn_1-(dynfNdo
ze)y*}TZ%xdw(!}bT~;M5G$pe)P|rY*q^ZYUO%uGvO`eE$9Z0w%Ls5;&Mw4&b9HUSO
zmq$!+4LbZ)Z1oerBWbL@?s%~c6&+_6tTRch{@SK~TX55rT61dCOFKp+-4bM&?3-p*
zET=zo-k!-iuHx+RiJ2>|NG`mxuKoyp<xCZgL)Vgy*ho3X&^yo059Tpn1=9;dt7~eU
z85(~D<Na19{jb*VXC$We3rZDpPAy{+|2G<k5pNb?8tn(U;}}4saeMN&RfQsnC>$vt
z^^4e#Ig(Nf(kdH-4S!d~%EAm(&_aU<!R$OG@(=VTlB;55qPvLCK{yxDq76tG{1^E%
zW!h#i>II1@1Hg0QJywkd9*#!{yi{lb033TB@59KzNgYcKv2rj8{(V-WgcM>UkWMT~
zC01k(51#&<uQhy@{1!($PklBhJc4(+;tuoz;2tN1QesJHliYwSm(u;nH}hkt5tXJ7
z1VU;<lH;gX{QOPjaRX1U1-dRycd;1GqXP{N(gegX@`7p|q5NDijOUBWF1Hge_g^%#
zy|DkulGFT79iraE*EW!_u|4e`Qt|C}#UvMCF*!{ctk}Z7L0Va_Uoo*INQRGLB;$73
zC0%LYot73}Wxrm2btcmvJN^!vJCe;~I-<{015A6ZHLJ^+rkpf){Ktvy!tfq>FhIrX
zP`w*89*z!(@lhS0UJ!-wroG~Mzi043h+6e?PHQ*KC#8hK8Uop!9?&NoT_J4oOqbd*
znvzq92LxJcZ#Swl9bTJEvn^ML#2!0Y+1-GASj^e-KZ~812=H@Jez6(tPf<B)qYd0U
z>w55<dWqp?DVX$`ynN8V+*9}8&J8oZxqr=1E>>gkxc++(h5$p1W1VfJxzfh6dU{P)
zTHwCLH{74CD?to6f8Q?jcJ_}Dp9{iy6vDOa<A<e7h6FA;Ji&=aGUzBWX=dC?QZjv=
z*N#ge*B2zDY%N$H2lyi<RK0cL5CFLg-b?V4CHD3!^NS7gQJnsk8d7A~CaI~5ZZ9|Q
z-v_Z>z8>S0!}7@)4{V_^e;<A1t|4LAHD3<3b)2!!^dm(oL7~7sMn9-nbA#+y+EK&`
zu!S$cJ$Z~)e@MLltA=D6YYl`&(fa&1h_^Oaw5OZ08~g7s>c?ig@>1K>(wRV5ZH#A&
z++btY6}cGv>T$S{Oj@4i%)FKS<sRn$?p5#k)H_?f<aG1zM(BO=@cFE3mW8PuATSvR
zYvSo7#wW<e%C^^aHZeGfLaR?owe@`x-9SHJrH=_P#~WVwC24HN#}&`}+lH?6I`i!A
zAy@t6gBF@@6}w0No!?C&h5hHAc4ZvFOyh75)mkM}o`>ALC_?)h-&o!1if=qv_h3Rs
zdSbM#H1-#4p7D)+(-g=3xa5&6Fo)gIZHQ22=|JZb_d!2P4JcGV+r=53C~>`7Vrr^+
zbQ0{tbaKyM&?SBE=VE}4jtS3?z2+ZlbuPR^GUw?;(-`Ne?L=}Rx8M6t2Pc{TOQs!P
zA^x#xlo<aW&Eer3MPB+3#0;i^H(r;RO<_HpEpHRK%zNbczL@%*90$uLRAvchiEE7n
z5{qyC>m1*$%?q^vJP^brhfHjXnv)a*)FC9wyH494VDlEp@g`$0D-%-Pmx59MgV4{f
zod6l(IoJ*<ndl13<2@L(p$OTtT)@IDev)Nj$75=Or0%s{zWnp)r6a30pI%>!aKzb;
z9pMkdQQT~bP52A1fk=p7E{|B=qkeT^QhDhkfm{NwTCdG<s13CqK*<<Dne@sa)rE#U
zrRGAknEUPT*|~(cIHTv5&wPrDueiTeyRgUU)9D*GNPnvi5EXZ-(~GgPf5N6=)SxBU
zCU=?uRu>Fwd5#)`H2kflqN8rk%Q}k#fu&o%k%+9)Kf;9;xE2dnI#mtgNHcE36rjZ4
z3JSN#zPmY09$*z<J?#P8;n5fVq<PgTBJ*K*K#ws35I<1tm$6aCFI3zjuegvnoTzR$
zSr4r$Ke7!FCQUJBt6~y7MP9bgJp?29XF`@TeLg$;=1wK;{0L&AN-p{HcMFAa#nrOi
zkcOt7iKmo@T%MTto$a$B&a=e(Le3XH1q9CObG!fUo)b5%zuI$OC$3&xR_P|6mz;eY
zCi&m@vmfSu4&}lO0@;u_?F<*|a{hjC7=*_d8zXzw-J0-k&@DsCheYS@2z!h0;=?bo
zNWDmdNh(K!WhmL*a>@hRHpp;(8x*$|*oE665rDk8k$uMROG^Giw+5}_4xmUKN~&D*
z+bw57fo%psF~rNvd`W9Qig^{MjFZCB{!Y&YRU_2_nO>GauxBdi86GXxrl#*YwxLd4
zGPg)E_==uSH%T**>p2I`tI0CdskabwpfJML`$R0l4s!}6kKZvKZ~7mX`6Gp9N$|nY
zVLpJqP5CFf7t?&>VN=1>Y50-2{>>ajtL||2+v(jv-6L~d(z9ujSGrFI{o*+M4iJfm
z@-~=@%dkW{`v%#N?Bb8jsUX4j)OSdu%n7m!tpCT>Uq)5ky<gnu-ZYX5(%ndRZCdFP
zq`SMjH`3DGB`uv2n?^udx}+tg8;R%3GsgM<?laEeC1Wsn!Dg*%&1=rjJVYH-o@s=P
z?EDxO7If5hO>#1+-q8MeG45~fj(MwwIl+4p^W$$=#r)|5!dLK3Iaa}hn9Vq9MX6F_
z7O4lg$1WGg=jXXW{x^HoBr^bl;jLZI#e9+(z5Tud=jOH47#5oA*V&zY-G_zX^L>9$
z@s;9}U-Uy?WDQ4VP`Cal2d>Hj@MawW+jbUvdfWf#)_E`fJh2>)C2JQzsgrYYsbBHC
z7z5b+&j2VelO|2iB~s7t>_=io24T_VJuNOf)Z%BZk|Hr(TsW%j^ye=oi#Pi_jytEr
z1#v?u<f~L$Upi5%D|Lef(c>S1?l;W<2o><Z;GW%8+!%NW1x`H^)9}ER#b?Xq^^Lf%
zMz>j2J4IN#*wzA=8|DuyaDvXGO_RRyw%wNNSa5Jz(x!!IX_(@?H=!H9q3U9sy^?RS
z&DiBnsI!SDq|XQXX6F27o=>q6$zHDlD|B~oXBpcmlMwCCaj@b{s*?eP?6vvU+hqXd
zL#zh+u3K6n6Fiqt>)A=KEQSxH9ywtXE{WOrH0Cx93MAp?a2>7VK(K;%+Q5Ss@#`-(
zFT+iHR3qxk%-=a+*P?<h8A*MxbI7J^OpJ||FV2LLhdUAkK7U_mUD)`+tE0Y<n|M)6
zt812aaqhIp0mOGOp#_@1oE(7G+6EYb&RxlEX{Rhhq^9#O!AHuIoKd9#Y^6DV^$+ak
zU+g-5U#qI?)2{hlrCc7a8It+`<=CHYD2+`acKz?)N*VZDvA^sXWa)Eglp$v!cHe({
zxXgT%CJgCwJ^sQ}T{}h!d-D9hz=Pf6yCFNa8nXL{?AxWgF(enPzrwv9j^_CDJ7ulL
z6?@%VZ9|5!Y~1W%K=?lOfFlEWv|v_>gR-2|(2YWLJtu=uc`ZP1HU#i{Db`=zEJ?lk
zyj!@u2aaRtMi4_LK<E~OrcA=e;nPT0@`cmCHLm-UnakPRw4FLXtF7(DRu@ZoHCT}_
zCUhE2l^uUw>p(eJWQ=b*M`5;w3M&FuiBw5sMhH5#6H7svcv{IB+q|~<M@5C><&j0o
zylnE-S3?R7PL>ozw%nUJA9C(uc^BK<eC%v_Y{+Xz&Q`2O?=u46U4D^^k*#C3S0{!q
zJB5-)Xj+MK6X=<^q7Tp&zkU4XUTN=SyHRCiscl4LD>^4fQ=W3QX2&uZDx&I%|2cXP
z-&EcoEGy6B-+BG|nYqpxY4QsPoEHaK^c&Mrj*ZNyCejy#kZ2c6RXn(=0mA~EFTpOO
z{kI#PBx=7St7*J3UUBKS=1DP`826I2FDUXi=aiX2%JLZ=!GK-?OGiCdWhY}MpZ_pP
zUsQ_VCI=dG5&1sbfs(-}VsCYZ?@UsRX&JS7f~Yrp>B<iO?A-H_yF;yP_^OR*SQ+^v
zX5)yG67Rayp0K~Tfey<*gr|>mE)(%WmSpV}6}GD|Zx?_eo1Vt~O!pVio?ITU=H;33
z8kC>>PycuO-rv1kzSA@1YJV0?<5DI>pY<>W_b9kJUrMsetE+iyT7ebI!jelD@(<$p
zLrywL<4>l0NnwN?5?B&>vA4c`ZeqAi`eQ18(h&$|BV3=TNxhu~(5>MqTCjP=-hU>X
zL)juX>T7dM4k{p#RbWR!z|JQ*FI0fSnb#5}d&BShdtrf9pn`TPVyl(b$S4$|a^==m
z>?Z_MCdu%kx0w702<e6&X$HIeKF^j42)eIl)T{6Ixb*0jax?Y2+bRl~bz7e=@ZAiR
zUaVnNR~rh`Kd5sK-e8Sq&UsL`$=Dbq+K&S5cRYV|GQ!^&g?9cpL^Y(or?@Qnx<-f$
zOW)I6CpP`rp+`^Z-%kKy-V3}>ac?;lg9}MOs{L^vOt}L8?$;omj@B{-X`=`T?@7zm
z$s41+U2vk7FlAQ5a3Ib}d>GtY8$byPzgMw9k@{lFgQR1H?_LPWkFN5xveroV)C?;D
zA0!0MM@M*TuvC7la{+OOh{DYs@Te^fL-I*>NGKu+G0X}bL3|<n(`+aNGBULJC=+bF
zUqR8K>#?0~xh|~C>4ITUd1d>~DZB$&Q8)#jNMZVGg>q=gOjN6KI?r1;hAFIV(l5?M
zY^Z3ohWBZfvEEqu2+68^OUSh*^cK9<8R{QdwC{x<U;A`jjVgS%jPP+>q#K#j>1T?L
z5y5`!qhoC(&X*JEh7346{hMnhqvoN0$Wq=RT2BPp@{)KAd4r9aLH<x#$myQx)U?}u
z=9A1Ex7#Zy7h^%nETh50#fb1s&IC{AlX`#KR{$hiM{WZczU}++7}ZLKTAa6D+Si-S
zF&0h-|BrYwa=4p<?82e|5)fOt+@CfFZkJ|Kb?I0V#z-5h`hvcl;O4}B8st~lr$MH>
z+97pE(H8lk;5R%2A6sGxrI^S#QoJVi{8xg5;(B8EP@E7YQ%&>V+=bV|nGyvjPd3e!
zkGJR(xfW<7hZ{TQs@|hI^%){eWD_;boGJ9!hLV0_jKaWSKvTC0fP&GzJ-B|{+UJZa
z6=UqkN@jtujZt(61jBaGCO-?-GlC3cAgqmXSD6cEZ;iU!6soQQde2T1dm48LCWFCB
zIUGxi?&?wX`n(+1`<fu^M;Y>U)DNUUMfWc<)2<SO!-!JVOz`XZz;0w@qq3UN-_|)v
z-`3I+iygJHvQ#Vj_4bz^UoFdgwfi+j6;s^9cc`H=AXYvGks<olIj7{$+R_$Kob14w
z1@TmInrpPM>yW^++cl7V5Dn}my~l`<%xXnC1-<8R+{20m3r8d8%6F3t5H85BaRzcn
zj(yY(mv;r9PEL%z-DzZ+2*TN++OpUpz2x(@=7cvMMeQtnE+=-)gta1NZ~C;nMijb&
zTk--9fOMCbB^vAEa<2GDK@QJ`D+4GlvE1pyDe0XmUEWNj;)fRUIF=o>39`p-2naht
z6cppLp76o&7d>7HA%wUHnP{s{$_JScrz)_|l3ZqR$#iAH74y@pJXc}vkK|%>=4{j6
zy7?DGaiXw^^e(9Rc|o8R^`FUINX@a&*W!&u`1o^XRXlrSyz`aGt^#^e(_srlu6N(;
zSh&Z}atoywa?+esR@G-B|CP~iNofl=Ok=Dlvc~8T3s~xTS$IEz+b_Z0F@9j@XkmZE
zbVR4A-Amnq=A_PoE>O+(NW)KGqzP%Cc!03R^2?H6aWJ!a%Dxgu`Y|2rQo5y&?Xmav
ztaFQWi2uwX-*4&b^Ryq1ou2H7UZG(tX8i1?UG)X&MX>98a;N9}{sEEQT;J-^bbw-9
zs}1w2gFHW6B<AG-LE^J8yH4+mxzCsXJ8Nw^0Xak51~6}9e+Pa{&e^#x+iNotY4ho-
zCQudte#%P;Ga%~P+al>(A+VNx<ABHcWNgd3|D%#7`n{1(Fk!g5&CC+b)a6>rhhk5S
zw%<acL|)(xMTFZa8YtX3awb@mj+uItc?3Fq`Nf8XX))Ran4|43JX8yl5vN;k!q3a>
zGPA4)j*F(`_CFg}?xV6bV2iOJyK8KbP{5rY6kmx?&qNHao0rbM=@{oYqPeV6_6Q5T
z--rn*P(mSwM&E;&H_6|ugK9x7;o|z~2^QE;_U>`Qa-{e&@+%}(qgrC8UxG`l`4-#3
zv3jB)A{qEkh2tb!iU?u_ff!mgq>BY9f(g%1UTsdbLwTaPyx#o4jYH$=&tod#u#TdJ
zC*zhtfj^b&2lsD(5-?a-VR5K@RCCjIYO=oa-;s+<=DHu7#_D@h%sU^1fLB(k47G%^
zfbjesRt=ZfvbPoyTfhb!_yp|{0dOKMao)pP(~hGo>$u%f58Z^(q_tRV%<DMbP8ji$
zc@6xclO-_$SSS{w3O+fl{aWhQIy*e;c>Ev8n`ZBT0$oQgd;^&dz^MmP#hXM_hE#){
z%h)EP{z0t;1o5Z1No$eIP;S9w=j<G%Ws)dX>9&<W08H|fD$<qndKib3SmGt~{hHL_
z&;#8kZru}FeUwS4DpDE_)OZ{%;Q{QjEgPtYH-Qbp9r}^I>w7c0ezn_;V{dG5Xs%s0
zK{g@&btlS`v>nS<wE|}gT1LFpqs65G@PG&=u&*dT@sa4>i=zK4GDt=iqZXb(Dd>Yr
z514761{keQirkMI6~<zzL^)4x&jmnUd*^{ZUL}skYUpZ5cv9y2&ZJ$dDG!%mbo1Vu
zmvl}5i%@M4m9ihXKR!p=$^$iX)A*{usQK}R!P4+ev|%ZQS$a<D*|zDiz<Row>lo-;
z$(Arm^!rO<75uZ+B9_s8Y=sM1fl-bFd4cHPtN)JJ96Rp6yySPr1#a{Rokv(ybpYix
z*Sn;O?aNK}|1G}K<|pWZ%~w3Z0ezmIs*>>4Cq{sHlI76$6GDZe&ceVDU*H*=%;+78
zG|fK8&dG5PLYyA1C~|ZNBPv1VY_A`l)9Q`rKgH~jXIJMy;)qR~os%>6$s~jl{O;yF
z?>p@SAxwg)*)`5oDM|0x31x#fA}QsFqDMk)?A{m**tyqI3ovkr$Nc)F?I6`ufkpws
z!a+BH?OUOVA)VkItOzq{S;|ktA46ZeXR?K{-w{Vl4y>QSLzWr16ibq&$4*yaip`^w
zo%yiwO!@(yQ|ML4PDRhE#)z5o6<ca@mh+nWKkIL+@}y<Yqr)~_v9ArpFVS|B<eA1O
zO5@lQ!HmpG<Q0yM?y%F$LParTqo`oo5Rth|=T$8sd5ejz`E&c&HqWjfb9+>{HD<@6
zpDPYEJ&up-od+N7DhHI?<-XabA+A*#`u?0%<!-yH$r!`3r~gMfzqn3)5!DJ}2Tt|H
z;<eLnCM>LTw0%(#-f)*-6vD>()-XQxui|m&IciX}b+Rmw)2P-KKC&5jAfCPn;T%-8
zXS#!x*O>7nRpflp@6Xu`TG;vWFDtLjJkk?)!m6y~{>u|qQc|kus^)fsL>I~lhUnJC
zmVl%f&j5A{qS^L4AU|nIa4JnSXV=iyy8Ksr?3eeTFG{zZ4xgz05a(`@v9Q1n;R*dZ
z*+z2(Z;?#Aw>l5+F6Sj_Z~F+0+WqOwG(@~PACl|73I<tyQzEqv4vu6(=iFo&;_CLr
zK1B`LdHf*k6@CpY!#&*~r&~`0-`+9$Y<=+F$@?y6p_6IXHge3loAf;V84+zef;zS~
zkQz{x^*{VKZG-ocw&_>d!T_|#-3;j!bxnO*AlzH>h(W@yVi356Q7#)vU`GXegAMKm
z-5ND+IDalch4za57=fFhfkaBABQaAaqix}sCW-C_g`6!8AAj(&r=Y+W8;&;VKnB5+
zP%00@)z9oVF>`64V!0c;aa5~?l`ziQdU3oy)~p=d7QZDTWAz%DbbuvcU-b+jo;#zj
zxL`qb7f4&wO1#x9DKL$Y<0vKmg2Ib_pbYSrtX@v*0R}sUUX?SmUMPJfCc7+Av|Y3V
z-U{+%t&`9l7+qU0SNXV9SWL*~6{J5N2{dWxB@2G1t%lAZowLU^#U+(DM#Q2fFVdXK
zT@3$^j8*zlsCTJwHo1W9f#=rZQLyd#qf4)aWaWcUr;p5lf*%d=O`_%>9Gpk@G3M4E
z6`hK$W4v7A*PLe$u^3mSDcul?f_4RVbNYlB5kjsFrj#G20wB<JAl!e?im;$tRvi$&
zOPLE5<I7bvB0u3gubP7hcx=p#M{%+x+JLb0!=ZkEn%q@|H_!;(kN2mvfMU#l<aEI<
zQ`<Sx<h<B}SOUI(>%xm~{&#EMd%q+^-UyYi0v56JxWI=#$mY1fE{XT;M!5OI?HQ-n
zsgd`?*<i7LyN9?9{RM+fiV^4)b-a@xMXEO|Ep4VJAr8uINF9n5T+;@vgck)8MWB~`
z@;uJ)Zf_OToFK`I7CgpIa5y7E!(8bqh(v8b-?TFZjxbStM?tpXKNGIPj4lfi_Ax6a
zByb~hmu>DAETP+HjN=4pUT|YkkQ=-sDEdnf2}VQdQ$xQ+ZgDjRuxT2$s}2M|A+(J#
zrs%qmLljH+aqyo#_}&YxH8d@$we_D81WZ@m#E7^*1wFmG);T1FI8qr!^$f?4EcPMh
zd*r}XvLob#B_^{LPNsbeVyS*tm!q7;<e_nt{%dvE^!v=tDw-B?^W86x(dx>-<4YuU
zLWg&xtZBp(nYen^IzoSQ*ys&bnD~Lv!hLIh`&@61`U6F8I|kPDiN!>UEo<kH$NKSV
zNBdu(7VmtNYhs}u36Ki1NW8Zb60@`IZvLy?J-@sVRkgJ<j>^8Iks#sc0p6QpK=Z+`
zA7O1L$hKYU^p9Tx>Rifps@BGM0gt1uB{cfbPN5dq2zzV$X=*V3eko?vHVDDMSwNub
znNxww1W)h+)&Q{>Y>BxYlMRpFN~m#V5K;tHphOcpE@~QLCHS;gJ2P*9NY6psmx4GS
z%N*fQWkWo*#{A9PkY^?7hgE3RTu7_U>I_#%sMmGze#c)o)J-Zg45zSC60*-ROxczc
z20kF6V`}}|$@qcU?keFQe8(b6zzgh2-ex<bj??9#uV?XIJ$`W&s#k^qcIoqM5yRYT
z8kP_H*U}~MbyV}@U0R$06KXX#%L(o7pR2>j-x_yyJ1?F?7z}cePsU*qqL9Tq#tR2G
z1K$1bDMR*h%6w8&qkggFhmv3IPfI&=-dP9$qv6;(5wBf3p^5Fm@}HFRfHO{AOD5q(
zmUQKaPkYAV4CXC!suPuN#`gTZRp(TUNXK8@EB~avtL#E>f_F=qZ@4i1Ty+>WPef+s
zwF)4pV!b{)v*n@iOM82TZU`Xj!Mg#4HiJZPP;f8ye2@o=I5FIe3M`HI{u$JY0S-#`
zl?m_L9^mRqp&dilKWhIF`>>Ulg=7gV+`5VuE_Z=C#k$|$(*Wd7`*B-EQE~B4rWzQP
z(BInNl>ne-@Vol|Nli6*9}8ULG|{lhT!4BT&b`?tTg?nU$K25zp?UPWm6a8W*@IrR
z<LJ&%7rO~K0hleafbUi{r&(dFi;S_az*k*lfm(J$%)e7g*M}=gKIWw(%||ZggAa4f
zS?d-)Y`ek@Q+D`jkX5MnZ%!(tEo@@E^?fORW}&otR?<_J5a)<3_V-Dr>g(i|lmN9R
zT9O5Epa8$Li;#qe1-7+mOroQujZ4B)^|?}gHOWdPeC%^!36H}szMUf^Gu!HtXtrgF
zI6*eHnmpDgdkPx&ZYcQLQNrH7s;IP7UQ#mH5AgAF+Rjx{6#sW(p?sNG4h{}z@S+x-
z<>bD2i<+64RX+5Abk(}q`jyhW>YWefcwG(q3v0Zv@vw$q0RkQTM=|SV)})iR6YG1s
zuWkFlcbU|4*?q((`30fGxMBAXd#z*FWf%8#5;vA2Tlxj_fAC22RzvlL?5*x!ERdRa
zA_{%sXg3~TJePJNL_4vj+kKBFaXnk!2?vV%KrH6da!+J%3@l@Yc&0-((;HcpZ!m~g
zoydmmQ`RM(`YuB@l5$;nJB_#S-)s9R?}_f8v08Y8rWVG_ePY~c?=VBNxN5-&)-7>4
zx!P}_mm#y^gJqxMWWy~v2EB*@WXrMzXYxo#)Ito8Ox{{YxE-gFYp`oxz<q`rGebO?
zu#b$5&9~e@8_@%=|D6|)kaNlk78x`==J?0!<%!-zO#XGy|I0KjD~#n-8PsvH%=7n%
zpE?W_Cnl=CA1*L0R<j6SAGMa`we&JNd5yHO>HZ`9JNL3Y8@ji@!zIMq_;O|4W?FE)
zh1KQsEVL&B2Na}<zcC{AYSjc}icqvpkPcSaX-_XPcEy?k7JEPJ8o?{;LNJzAyk3HH
zGfg81CzznyVvdt})>I71AS<Qmim9jJC?qQ#C_@%u$m0gZW6!A|xqm$tBk>q13b3>3
z_59o1o<H0(I7;)KQugc4;E`QKuTR8wQ)Sqg{F@^`3-hL<lAk`&-2}XlUsQJf?_tX2
z|6CL&kf^;1>IKl;X)vG>POfv#mD8qk21UTC0HOt!ynK7eRex>zB~=`&Jx2ZgQc=-R
z{gTdMLu@mqS#VG(v%P2Nj`*@27yc2}v3bWpaZ`ndxWSu>hEC6Ro_XulXspLy>mIms
zep-QV>bBU4d)Qg+)sm`4t!l|=YEct%ucx|-EY<MlqWj^~Iw@&#MT~~{Elksi#=`5E
z6fBeQ?Ayt9i%YETP?4$#Jf-4%bXq!g;+AZmajpFqvJofN5NCZVQTKIuJL+Pz8_6!=
zFQU&d_{<9Y_>f5=3Kh!*J#Q0&eG=oAgBHgSd!H@UHMIXN^cOQ|m=aH0;WeLO9GM!h
zJ6yUJT8Q~)95PdiJuX{krQ<|n4UWpc7v-X_cN6(F=dGjkdp>Uj*>i=#U0$wekkwc2
zcQm=#&J4L3fJ^$WNLF88|JUIEP8qhE=e(B(^9#P`0eB%d>8pohlD^BbtB=%k`AOTH
zeglsMF`9=3to7z&QBHOZh%y(_MK^7ggwQWv1o|%k*f*5;kqfo8WLptJ<N3@fE@V-f
zgzaV|Tqridx#3%X>NJ?AY`n?i62GHbt5x_$v-2RSWLL_$39@dcs#H78;{Lfr<aR5D
z=(j^+`@!Y5FsTb9xJeKX6kROzPMxQ@4#odRX1EdhoanZCxHNYD=Ozqm6vwU)vm9ga
zDAU7p+z1SRtaAUR%9H<ETS?1Ir{G}mjWj?`UuXNl+W)r<t+IS_7v&V#7FX78Iog{B
z&lMO5TKTLm`Hkg`eOMp~0ou+yqkkqY2`*cw{o9*&$grO4Oe8|qLnk>9=tf=uxcv-o
z2g|lzRm5F<{>G&hjQrLb3K63b!~GK?7JKRhN`Ut-M?%?`yvI`M_iqgSod^~uTng-F
z0g)Q-UPfMdPKu7wuM-n@P4eCc3#!nI(OCn%1;Ly{OdQ}n6qo|H)sLI#j3!OdAlVFn
zEk_t68~qlPaL*-miQ(<86t{4j{ApeW|L8l>SQhVOzyMp~%<vK%$P6CeLkX#~j$;hZ
zzMjQ#e9<{WPz_VeB`iR>qs?5(EH+m6)AFjY7Cp}1zyLSK(v29;eMpSiR1=GxY9D{Y
zM+MC)to-N`sF%ASc^R>5uzooq&fEk-CD=BDK3d+nl1|IlH3b=fsH{y%k8<pA?E-=8
zD|+(xhSMjbEOSpq&sOCAVnn?C0FVOx2xG%F(Ib^tUo{PH<Th#|`5^Sd?85$gT23<&
z!_BI1joabe*j&OgFj!WR{QrLLAl~)aPQ9q-vzmrKfnI0&#l^+uBm)bJ68k+Tz`TyY
z*fB>(_7)04azU|CLHQ{$^yB7U6rP=+hhW=_J;S*wjN5?Kok?$qkfN-9)th!BBe+Jn
zj_5>;8r=+*44^zhjAJ$?&K|x{wQe7E>E4a~TJ={-6m=(KZYZ3!DBJQ6rR;P9%yqv;
zk%K=SYEHzkrmf60BZ*?W=NP_CYD;{rs@b(t%~Sm4Z$pu*hVRcOvE`?Nkx~Jj7X8l(
zB%dsE)i2)~=ol|>Q^eR7@DZjZ;pk-WX@qkle*;-UX$Mdk`sugApdX`|^2)2k6ha8X
zg2I)c<f^V%f94KRFM<xEnc&<L<9?UV6R@qzr-v<ZrWTd_Vf2Y6Kz=V|l9}s!=0>GE
zO!fAv=kQA#Fq&233zYsgqKE7-L^W>1r^A18xq*;Uiz&C2><S8@!nf@OPqt&qz?&7z
zO*4Ia@iJdqGBHNZ5@eeMz@vauK+LcMkR=oM<7cbUZX9#2nl=k5^yC9k)}7y>^W+qx
z9x8M9Q^;Zp_uv|V`$!KJ&HRO_%%zk;_w9lSL{k}ZxIwv7ziAP8qjTZfFOcabbQo@x
zzR$Q;g6OVT<Acn%D;5LWZiWLm?5&KiIS|)wZpfJxn}KQf8ata*dCVLt|Cdsli%5}l
zb#2n*yuI$K9m!=-uy*EZ?ixXUeTthx+)(3r*;tYv3%&|cL_<AXd;~JF&_`EtgV^l}
zlKs>Kx~L{c&=15+tRti{j%%puJ$F*wjBkciBbpeIt<xj+IM<GVG^Gl{&Y6-`g?-Oq
zO(jfk6==_Cy(-DFtNJDSxUYN;*j?mW4akK&DqpN?&D%l$89<(2mU|Tbz!C>&R(<%v
z>&g#wd%dKNTm}d&WdZ^^SNJzI1ORkl4fm3S@;mf?Y`+k(SMX`4oUBP?TW_~s2NO!N
zE0BMFhb{&)zMMq&u1*Mp-LWx8{o@G0Fo_S(ZLgI~Tb&*A9-=r9ZzcF+0Y;v`#bXHS
zQ7izVp!*;*qY%1HBJoQt5QlN+(pVsvDjeQ;9mmt&fJRTLdb!?Rg1#z+Oz($>4tx|n
zr@%M@IjpNo`j;Vx`Rx5hQ0dk}k^t$nV4a4E#~P#C^-hjYb9-G)1~!JzO(Qn~*j9bn
zN}o77zVZPL0o4~Uu5CvTpVy8NZ2B;A_C(%dERaQC?cI1AGt@hTqm3PM@LTXYk!Ds2
zMYYP%JOQ4D$6wy|rWuP)rf4O`-1TxFRS<u<ku#3MC<ZgeTOG`^f8ldPd{0<I=TaQ~
zcl&*oBpz?sW1J-pWnv@z!sk*vGTS0#n|1N`f$){-hm|#mGw}7ygN3p+;I?#P&}%W|
ztsZ|pYJXsdLz*?85-{OcCr}g0?r76geMxT4X$WNvA)HD(;&vP$@cNPcRoGAJ4cI~K
z6SAF`15`td4Duah*yvc0s!RrP1aj-d)m$20<{X0rqEOsFTO9M0r~RAWh{uy%kR)J>
z>~;JzuPvUQBED_1r15PGXW*iX?6k;H>b!<rqwuKZ5^~Fy_td6+Wf@waUa4?nRBLq_
zT%6zy<nXzYb;CC(zyj85+PT|aI}UC1`W3j*L6Su8j<LpO`_50f${WVY(QN$1qIjOT
zmu6)ol#t6}LEx}~V`8rOwm!3Pc6oEv^>DgvDF8^@4ZqdYGbR?)bLv4npOR-=#1eSn
zldJ$GV~J^LmizyG@U*?$()WO`%k25_Ce4cbU7EfJ@L2S`n$?uIScbGdbp}RV?z7!l
zm+veF{fd2^{WXlsfQMi~;VqDwe3KqoY~u%?ijt#C(#;40N~lbU(-Gfq`r>iW8+M}%
za5}-i9+&bNwG<<yvqB+<YC~Ul_ECpU=<z(D?n)5EL76~cm0s_oTFg|kLw7Nh0;8h7
z1%%Xfad@|G90YlMQxQG<FO8sFAj=*sP(%-lF}(*trQeq0-^G(HAg_Gw>_!A(AMKDG
zfS|xH&$T5Oveuy+gua(Jxy^k-l7OQdA~#*?E_{l+r{D8#+L3;C=5XIn(b(N#pfts2
zlwa+E7_|^X->*&5yQiGU1wZjfBkDW4u&a)+Rk)tMv1+ke+*O6$o>9!+t55n6FT$4V
zc2bgVm12@RGzk5s8{0_YnL4l>3jGiAoy>&|R9Q(9&nyi%$q8@{td0GAVq6Ep0xbN}
zqZ)HMk>NPV)003N=SHKa5@t_aaTs)mLu#0-A4k0hvMx#By`k#S9Fj;dNlFM@HCP-2
z-9|g&>^sM#Eu|gY%4W(yiz}P!*jNxkmhTKL*}E##tjWyo$dL`&+z*<n7h=jN>+>a<
znMAtEcDJV+=YWrPp$X_((>iWDWMyPvsCBI&kNN|aiY2j@N2HJtkAs}cEWh8lDJoYh
z&7@y>;nxVAz#2YjYK{PqJl180_9_U#@!%aegpd4>qf>DN@O&3xi9V>ririaiH``^b
z0PQsAi<WsA-z9w<Yzf_dlTLl}B@f0YxM}wC-+HAR_oeo|#IL0fdMz~bsP1r7#&huO
zP>gPDD9OCQ01jh_Vo{GDJhXN}XdQII4T%O!c%%mWMFr@?Ci>LtSr-E~v|iubUr989
zr~pbGz}IpxZ?a3wvW_>FRjev@o+ffEp}Bzj3XxblbuzW5$16Y{<}d4M(=ajq>~%xH
zJnaK#3Dso&3N#~(O{oP+0QEyqp8F`o-(Yhuiq}MztBRRelMH%dzd`W^r5em6aT3bZ
z-(F=0b(ql$u+a_El3-eer>PEd&%JiN8P`8AhY0K&rb@eB&QP*x$xux)q6hV*(&3&x
zp~ekGi9%i%>2<A&9`IZ^d<m}<U>t`(s@;uE*|dt3UkEXPm{podpWFSN5Vgx@{teat
z<QQNQZ?;Rnti_=TRGTv1<FkjBk#9|;Pk+y^A`g1^h*sk?`dpRJZ%exQ=zRAljacE!
zJQNJS9Z8=XFXDfX0JMbRo~#I8FG#+7_wH{t#v70uGybVBoL(iLTn^APMIZQdk!Sv5
zxqgx64}<_Gbr<Y^SJM0h1@I(}AR;FI^V@Tc#}PeR9HNWIqC@liuq#?zdonO})EH&9
zKBqgfV$}p@RFH(spF-i{ft#LzxhPhX5E|T+tJR;4xTZ}V`QZeEibs&&#Mr#P*CQAq
zJ$^zeV?PzsF+V%m3=M3C!8F$bXHnNJWS%Ih5_Q6VoPFGTP$n&|(@%l6Dp%QoKmZs0
zk(bgJL<Cu8_Dkl*Tit_=14E85h`^m5<EsKBH(S%d1mq5Na|m)<ij@hL#)*%2|9NV|
z<tU~osecv=rxI%1)GY(~$*jk-r#Y1*diDMie;&Iw0b6njJ-W{^Y|BgkX@lilmCKWt
z^hF)DyC@89m=(t*5-_feCV6&pYIdk&?5_wH0gT#OqUPV#FMrw6o_AwR|7{Tvw6ZFA
zaz`D91=zpE$Jb5&r=~3YR!^H2YddoeXq;vS6c$ORoqj=kzzkPr2S6NpQu63l8DVT*
z^Fa^ZfmEwmFv;7qAWu5>EYH^`k`O6x>DO`-|0a3+Cl6s+VTOS@D*;#PiJ6)2f^Qa;
zJMS({8Ch6T05wNjNpPPg@;y~**MH^+<s@M1E;AwWukxt=OLfP!j;{Q1=fC45O#gqf
zCl@uz0a@2v3rKDzXAjWV+6xSD@9PRFbL|y~#fU+uEr@=7*!_i=GU^|79XI@-bfCB0
z(SK1+XjsVHP~rq3+qI;U452bOSl>YA1Su7&jSBx~U*}<=AY!-Iw_%mBeKiCKBZ{q?
zg}R>`t`4EOzC*J@2DL8CXNkJ?U;`KGToM{@Qq|xFbrYp911Z?-=RP3ACo3Y`HC4fS
z%#TYEY^!I^y7{5C`I4ml0U+<VU(9+Nz=_Db{;YmZE9|1V;uBWoRt+!Ff*;zQdYLS~
z+#(T>eUY6Iwi`8Qc6C|JI+Dto^70`)9~U|qiWT}xF-YoLY4+W^{oDYd2<+r4m36vQ
zWC24P!uNt#F;G>dD$R2agLL2#gabA>dcupzzNu<8`iqF-j>b?3_ygulduQoWva)Q*
zkt+FAdw|EG8r9S<vZZnN`$%6ew)l|d@WBX}K=aeC-U;fU_n)hEg3x_naZw~b6Yk?8
zr}M;idmz2(n-%;}YCS~H%x5RG(Ge*0;B@D)`M#YV7kYS*<(R;`sC|80lp7*;QC<A3
zorbM!9uN=8uEV4>UV<epHd5=of|Aqh6<%92rgoMXhlVa+z{_}iS>6u9_J?`>Od}UB
zyWwQv;`({iFVqCFIeDnRS&R2{Y!^=?CN_%-#!9*~^rMO@QEbxJ07uv%0onBb&)3!T
zi~KwFLN6O6ag8R7sWc8rc$(97{8^%lnLH1~Flm@{IcFC7g5iR3segOxwID8@)fzRY
zgGmT7BT#Y4i1|^_RcrDL$F8g-6~(zP)u7qnH;)LsAk1`f5V06#cddZWpUJu@UY0c&
z$TCok@_R0qz-@{36+t|$wBFh@4408Se}oRJ**e(1c7qqoC`NRL@s>$$$o<_R+8lBA
zUVN^N82kwZl+zXl_5n#&#sC{5_f9N6PNUVd{)5X!r`*)PUWA`bVRDNPL5lodN+=iu
zw#nBYE=4>#5o*_)`gbU|6=mE$y2o6Z8%t#Zf;z92r1sx1`3N5K@fsImp4x!T2dmM1
zUb6fmWW2fp1mmIYLD!M$5>Q5n)h8(fzfOgK3-v?3Pm`9Wgg!@#iHn|U!2*|rS2S1>
z)nH!8L<rY(AY=lXJ@sc~yH(dAUODeRZo~@XYT20n?vPefRy;1(1M5+y3$&WS=9PZ_
z^FFrnB7&arZW^m{QN}kzxyV#l?mpTXL5u^>PR68QO$RTx)_6gXE1{7x*^EUr07WdM
z=DU(?os*-7Qn7wcqrMo!K*X6*n$`}ye-EVO_!4pJr%vQdoKI3N2eQdW%=R<Z>4$x;
zIYhHvHHf>fS?LH+eU_!Wt^cDmRbC%ectTMw9{<U^0w!|qP+<x~LKl+feOh)-EZkDk
z5;g$9OLO%zm6ZXbhCA4gXnydOlsjyu%G#AVG99S?rFl!XSxu-q;YeA!m9sxdRvEa)
zc3!!r-}U-m@=Uz7x9j>U{^b9VpvvL;&F$~BrVA&h`9Vkx#$S=sZuqGVUr(mx|5$`d
zUm{{ucsS43%Z~rALkMwES>O-uG7lERcVPLD9s~d(zEd>KJ5m~fR7}RUR4(3qF?%l6
z49(*$F37@2Fuf<Yw!M4Jiat6d)2sGNoU%7`BH$l6r}x?S<FKnTMIE75uhRChQ)Dp}
zzPTNE)zeqT4RQZOg*1mv<Ub3|)%@n`(3ny?N2hTgQAjF|OP0I5&^ooBwrOG4#gnqB
z!k7s^HI_&MQ&)N}(9{23dhpjSw);i54A|NPCgHU3ec0;g$nv_mp%W`s3P6mPovqLt
zS{o58<Tt3)?nMfA)KYT-2)RL-tJ?cWD<GB1szLgEB$QyjH&-Kaa*3NZyNE-_#Zc<=
zejHRI6kO-^K`<pd+UYmr4~1a5?pGVLVti%t5)}xM`;Q>$e2jSlH8K1@;5WMpBy5I;
zZf!G~AF5VVGDu&U{Tr!Hdh5^EpkfcxwC7WQWYx1w=q#FG;47|{qKkn^b_QM#5_{5=
zTFRW5Dy9DIz)bArZAFqfYbgYpt)O{h@3nS+RSE2Y-l(9j4~(~rDoboIMoqd?>?j;9
zJN2OtFbHt;fCz==;i5nq0BvnD@eoQgka~`~h#+`|@QW38DpAALPm~eTKK;C{Omxtb
zu0p^2Gb?|0Tqxdo=cjeWMbAz&+3sPsaj8MK-F!56Rada%rLeu?RBuC5hif5^pylQ2
zx&X6w+b5vTLMs*a_=Q_TQ`=vqqS0;NZz(jf?U5C4M`;;qd8tm^@!en-&Q(6k_^u9&
za*=$`86M;`ljm&OKmsLm|6_1&Dzce?-d|J=Z7a^?n{Ye%i(bJu!Ofn!Ytz$<Sno6{
zta{$FoVe6b(|9=fai`-swTe1CoP-$Re-gL%qZ_c|k?3ddy4{R&`CB)*#9D!`@0I{=
z1Arj1*Nc|q%iBHht+O4E*vjEcB|UXSeLlVPNbG4rD8^RgQbL}l6s`z+pdlj)c#&I#
zZ~Qgiq=JzM{-J#{=^d{ZLx*Ub4)tnT2#caSb;CLBbT#Cr#9h`1$SFEdBcr#7f0^kn
ziR~tXB7E)KN=uc`%SO)-dNwGp%ypQbFKz3up6VE})|g;4PeDonA$h%NilZUMW@Pi#
zJ$Sc@daTuTzI;r3XMbIbZ3&&J^j4qpQBWXC{pG62ROtB)y`?Oe6doJI%@gBgSga*F
z{i&OfEZ48zmxb7wy<2}ut@!;Qm_f>lm*VriE1SWGXslJMD8TX9(q>Wov+~t$$ACNJ
zoL}maF+a4fl#L<E*({q-`W{uLm%u59dj^#~l%8N!2~me!EFZIJ+uBgWM*nLn8N)!j
z02Wv8D*$&yy@PH-jSJAdf-&AhwUZI!?@puG{@hAprfz<vk|(Ztf=B-Oh!vY^;D1km
z4&T2KiMRrf#D#U1dsu2Ox|?~p{bU=C>{+dP5~Ta{1K$db{{2;2`I4xZ;=n~xqH3_@
zs6Fi8RHBad^OR;ZyuV;W{qQ~Xx8uID1bfhXHzoreug0_JUwPdcxdPQ&A7%8*smV-c
z0KDUSyQ)X|KqBnf^PlKOQqNQLjY;voQ}KM=aU5T(yJ2rg9U(&m^ts2Mssd34hJx?8
zG<$hit0tPBgG^?EcfVXQ@?-?@8);R=l0TwYnRoDtm@bW@+mFCb=toHB4X6z}FH#Rx
z)W664``LsQoeFdKJ=Pc3#FfD{ig6$%^!mzNWGfOVdf~eu<CndoO#UM{nZUc55MfE`
zisEI<uJ*jgHh}U=x#-75Y$w^xf~3oqd775!<!@)pjK?edDY=UmfY6_ot=Ob~kkfzq
zwCr16p6xyWDaag%SM5-GKHcq(EIW4DM@63itn=p1mFFp{-vrwzeMH7yCPra{eu^z5
zQk-IFTN$UFlTirNxK2_;Fk^Nf)J2ovSJ9uKpjF-xFzR;Upco=lQC#QSXnEzFuYS5-
zU8pRbOv+A=6n--}WT#H%VqMT12XF9<iswr??cvEsB=??fK1uzN9%dv~_mOBVAs|vN
z_dS}M>!B-4Ku?1^S%)}F)%a<3s`$5v#cfujiUj=GBp7?WX(s<q$L-OWs_I#=d;{5)
z-L&%f>h(K#oDjFEe<A)Rm+;L{RUCP`Ve=Ei0|zS!-NrSA<CRA`c#u~T7-!YjtN_Mt
zVG#e-I>*bIZPEC|?@;JI`KhUHe#^<fQN@X6CrHQu@zXq=FKxB6sO2`X!Jl+9FP23j
z;8Ib+bSju-Uzf9~?@1clNe|dOLl6UNY-j^j=)NO<6_<e_G%l9DIXGzU)lLAwFoHR1
zwhCRd=;>0xmC(urTjd|Q*ay-w>VhVwZ-t!K2P50lySZ~rq8c?ho@v6Yq^xL(RJ^6&
z1g~L^L9p%U+UCt#GMgje6E$1A$MRo>;PDofFD~bm!d_WmAf!&YzQWBny>sbV{BaSc
zm>NU@dxc7gSp6$d43aCfG%d{Z``DbsJIrff`ie{X0%;gLSw`)>=`GzZE1yy6`|~31
zJMG<v!$XWt4n7&lz`@IOKkwDkRj$)-yI!6L^GMZRb7vw5B<FVLm=uPqIMG~#&{m#Z
z{4ZVO1feH|?vdPe-`)X2QuK=Zynb^h2{Giw1NlO{MK0H16q%^i&#tPdyghXH`MXoz
zKE8sU7p`*r4d){5lG)m6C&P01;qJmd{CM8LR}qg@k0J1G#Blp;kem-o@jrrm9^K1R
zbHV%ykR_5OF$Ma2H94x;n@g13qufLRnW1Ze>bLPNZ5}}tK`#RWquv(I*6Y4(H?ML~
z+$r6LK{o+X(W;5DhPG^EDrOBC1<K$DFu|6H$WJWxo+a53PcS+|7B<VVX%1UA2?cx&
zsisnncm3$MXl>G4cvX-H@7Ne8*4ty_8+9@gSPJ`UD^?8&Bo*pnud0SFrm{Ba8}_#5
z)TDQB*Wes!5~K}s1L4DrKqR1dP{Lf^^q(<TBY*lel_s}#_7{W6Q$6<YEc~`25%$EA
z-&R*h8UB5o!PcIMc3)eFFusk&*XoeOfjsfggjG`GcWoeC>FIo;=F8(LUnGV!9>u@I
zUt2WfT%ujR4Sz)PL)m@n3MF-F#l$US(+ZEd0~f)UaYz*zgC8FvbWPDGq~`4VBxBCZ
z9^%o^W(7(vUK=G;xFSyQ+~c26<86ocf(K4HsS%QzH#nsA)2S~injR2x;!+(yZ`08J
zbYY3RLnVNZ*bqQ^Z=zI7%waxOS;?E<56k)Z(cB!A1L}!c!t7-T>Pr4epj;6lSDGo8
z;WdcCdD3@N`DANzI6a=yGnhjoo9UP^htDwJip0R2%b`ftz72=wkCBp^z$nQI(?TM5
zZfqP;;4}IV!9bQHMshI1CEk8pN-MdH<c8P{?!6`3_@nVr)AA-U4D5Qdmu@A}EtM;s
zPcWZRvC$GX$@B?gXHMp{_MB;BYim9bO$c~bn1y{VZ+}Cpi{;aX0ka09Vc=tv9qer2
z1)=uub$BtBL)W9n{%nP2{M7#jO<T$T_mKmXt~pS|#?MyNbT09FcegA73qgz__rD9j
zri^C>g1JDrQXB?Uqj<vHk{PBj4G^9f7&LvJ=@YC$@FQd!{5{>M8}%H~@9p3U+@TmR
z`lSNT6ZlQ{clK-y$G7MY<$u(XP@zPk$<oD&a*0K>bg~IMJxem`K3MF<mL6eMzdi(X
za0?qqg<RD&<6I${f$UC;z{4tzFb5E#7ybrMm?k!vEJ_VA1K1soY`q5|Usl$NcH}@S
zf9;5$nDyMz!1-N1)>X-K8M!|)2!UM%$$x!4`?~P=h7s-~*lJP|-m}0`zAP{%P)kaY
zfH=F#a03Z_Cg_1c2sf7c73E%lGAT>cS4H?nf1kgsX^w%p-Xb5B?~#Yz>+UO~$<93F
zJv}*l-!4lBx6jr1`;gPMyNNC=<$ZO9YW}i}!ylQkOhaVM)X4I$FFNhkTgHWQP$D>I
zTzfW;lMnbByN}*eXbxhcj39xUt)V49MqbsFTMO3Qo0aN5V_TqTO-lxQ74W^iXmwNy
z8;u$;ldaV7+4n144T<8s?xw;gS{R0s8>Q}oYmtmh9|hJ80qG}~^!ShO4*3-)R+G`F
zc2~?fI1tlh<5#)+FUVPQlvVqb8|EU(^hr?5b`Sjo@&v_QeJ*LAeydw#w*ON3`Mo@w
z_S2b(a4koiDCNK7rqU~%$Ik=G4;o=Peu;i(ce@_3_m}(B-`4`d{}x+=r5Y)-OGj<1
zTN^}UZ_hRXVKn|@q12*JmKDIQYG8i`tPaY8p2rXEL6%)1cx(pD^ZK4SEY={utu>T;
zc=j8h&~+H7PY+K1j}}V&?!{7)>$QoBrmhbtq0s~!5@bGmly{p^Y-WSMZ{iQ7;!;K%
zBYF_f6)v{RzYrF6_k~_XpT0MZp}`x(>O%);9<+GI$W~wwKhhMojTApU2yFEOGg+Sn
zSB+WERqT+t+}@6Tq8Vt4)Q~<#HUP2cl-B3s<Z_R@GGBf<threv(vOL0ZV|bn5Qf6b
zfUK-ZWt-VWB^1v@`u?FngKjWy60i<~>Ksx+32hd=cEJd($Pr-~q^#p{yRZLD^UVxG
zJT3*XuyPeDgF~0Bv^UzEFDbh}|0Y1KK?57M&Bd_`9G~Q7u)2N?>Eb2oF05(IduWx?
zudq!Uwyw#syW^7%T)sS+5)SkgI?jBZ0F-K0eZOEuFq)Erq3j<@i-`W5;wk<NIKw5E
zJJ9i1raSF^jJ<=P25EZ<DX~^1$I7CauYUGDYA@+;bJG{hu{qQS8w4pcx8?czrqHAz
zMrNuWBFGY^0+sjOT2K<>D3a+35T_XBVbn)cDAQoK;122G0EZOC$DMRA-4VS92jZKp
zsvjiVBFA-C-RWpcSSx7xrglPyE0*GZ@wgYMofhL<-%oSlOX}tbT%!-YLR+s6NJ5IW
z2C^+tyA^UrKkhJl??o-`keQ6PQhP@8@2YpQ@_*9Pvm<VWo%W*jq^K|X99EgX0Ik5G
z^eo1az?LO9R3(gRO!#Jn8Xg%Fz51c<gJCSWs6jRCT<+h7i9XzL4p2o?KdmIR&-I$>
zS~NBts}kSy$6HOUymgkFec-Wu22k%Db6v{I%f)T}b75j6yr_I<WG?jCPM=}GWz6Y+
z?`GZJ$2xWMxuS^ompIZKJG_~~yBH$u-f1KikQTdDz8cag;w=h3crX~{galS9gBC(O
zDL73*BJ{ywPH=<*ySAKNcnp$u%5Hk*Byk3{e&l)N8+nKR1+=&GUVHav6+I7=j=d4a
zC6o_))n-M0fgBrQf((RrB{xZ^PX@vrFXvk$=C2qWaSY9E4_()0d$spD9N)O4`%SMJ
z3$FSJ_@Pdh3W~@eO;!nyJY%xq2Z{=nw9%>FXvGiXbIEa1H=-E4#0MdJrI5=uPs$u9
zk#8AClv3ndT6{>EB|YgRA7(11yw9CU=5uTa3F+6fo<5^Ob7lVJ;ijP6#M1<F4$UOr
z>T7c&-nAg8z2E%T@L6cS#(?xDX_fT4$wP~wbS&z9u(g=(eY%S5()C^oS`n<oyS>VQ
z<JN&NUeoF^41kSpH-+O3Nk8A6Ng*6gkx6?d>W8^;=MV6=Uj1CCYy7o?Vv=noWco@E
zzA|(R_seC0CIdw@_0Cn>d%{CFqb_L{ty(c9p`*qo2?0nBCgfD?4mcOK%S&@#RSSZ*
z7DBzJ75OlL4?(Er%1{FJpG0}q1c0sw<RK8#+yA8M7CEkU_N0i0rEdZ8l7V6O?H5q`
zMLZJ=OOx%3p85ahbTRP$^HyLpf_ikQzhdj9{N)UXZU{*F(ml5v8puxA+sC7(_qEw@
z4Of_L&v+xvpC0!VM#U>okmOgW`Sioup?F@nvon0duy5&?Lvnm4Oh$MOOnUHNuwj6G
z8)*%s{ue4Euo89Z0<Qr64pM~*tP|{`B5(W7oto^5cas?Sc8k3#!0g5!Mqb~r|B%j{
z9}%x9fd-Z?0<XN5LsQTusb~oOgNly;dIKkO2c?(5NBu}UgDo~FKw0B82WDO*#5rM{
zYKVY<V(vGZ`^WR3lE*)%+NkfNd4G{ZDIraAwg)UE%}FJt!b5mz^No$*c(<!%>BXy^
zz3$HNk*5x<r0XTRgO4`3Cf!3CNVxl+OBNs=9@%DJ0BAIRbr=Gc9wF2vGrbb`qMVBB
zR@gz{JmH)*2Xe}6%7N3dSyQK^tnrUN$DMEd&Nw&s9{L7&d|3pzH-7>oy&m5K?&6cK
zXM-@nl$19Z`n6h_<@G=v?-I|v-n1P7{uc~FhxQ}h<<s>*)EQn{2rXbbw`6?xu6~RT
zuo0S_Z~t`JPLS4Nuyt^7`2U_oUi{8efzSU81<v|$#-=v_pKSy7+)3xZ@#rcbO))T+
z^okyShlb3)8ON+b1E~|Bnc#t&P$uAqjXdYLE8`n!@UJ6F2{$}I1&b-OJ76R+re68V
z&R-PRM(F$qR@9By`G_c4hU%tHESEIYr1N(O2u-uTs2IhpA;F3CMp0Ma8PFyeB*^p$
zzAevyy&}@IQc`22N!*dIQ`YZFu4viT@xQJjrE_C}USr3_t3xXEP>PmsgYG8{Bq-02
zXp-Mfun+%`$DBp<-VFUfUY7y)XR>A{%9MiMePAnoyJ+hlw$lDQQgij~9-fe6EFHT4
zh=_M()@L}o&TcyNv0t?;u?fam<fOoa0;I(ga14b+0YD0pI{^g{Pp0D6E#8%$nC@!r
z(VqUl&B`1-WcxJP{uO^yBfENkW~MZ$d-7-o7kx{;@h0gPPf}4oJBDOxh}+I+#tWoU
zxA5qOYA&H2XrQ!<3<kQ9bcpE{(kT!T5qHcxfc|N~+4H~sHt+k2K|Alqf)pUug@J(q
z<Qh#2tWn~y(V3#fLr@8;W02UX-5PBzGFdj27xo7PGVp8!YZaSWTqbTFUAXM15g}VL
z)5<jsui>n3P3>F9uI_n{T#A5Fk&=mIar*`gn0?8h|D}$BTs`{a77MrVs?|1^%CudP
zHGp*<Nwu#Zg^%%VmGbAEH62E*b<QzeC%>X#x!O*jl!J}&?}$*46eQ$aVXaH2vj%$2
zZJZ6<Kt&EPZnx5%v?Zt*(b%}Ui3hv+-!m|+!7toy2|cx7MoU6FCZnQFK@^sHXkZhR
zt#|{@mRUXUn8f1>nIIye1K-RJ$`Iah?ozVV+LDJkVzqt5An~b#B;-J+{(8xlP2g_(
zszx^A`#jM9y71MAq~O5|o)>XzVdb-v#@tp>-X`Mqwot&s0`|x2KWT;z#Pqd?SDJ?Y
zZ~ZQ&w*xSp;I3<w{Ydjd$Wb<DtM!Hc@xi^m&MrHdPR8{ow=kD=cW=`FkNtniiPoU_
zRcy!8`il<`a=%w-!EIs$-jS1&V<w5who?@6fN8>3U}GZMdKOvWb;4^|*>b<q!4Ru}
z=Auy#A@kpSb5vQMG^~yD*7loDF0<pu2=7U65(7NU6-qQUw95`!%>SgOs?4E|Fxl~`
zgK+O2RSe7zJL>{Sv*6cR!tqyOR!v7=1e(4M&lUE4%KmDN|7kv5nvn9FYa`6czpx5}
zS9e5DL)UiI->YZ#o3rfRL4CJt+|>#S?T3(t)IEWM+7Rb@Fyg-amB<nX4bZ=Qj3rhv
zeKHh2srJ-oi*+J)Y#!6xAEMDyof(5Pf7@po&+DQf2-y<L^O~U)8FN)y?tOcVIPE+E
z1B_w^h2*|*Rq~n+ft_Z>W;zBD4|Dx4%>%gsaDRtw3(prkC)(IDC&xpV<@1__KR1+I
z_0<zz&y!qPPn%YG<6n=)#~HApj%RB4SR`KlrlRM+&`C>7Oe_K_FPm){h+&nAd@D}D
zBtT1(sL;Y=Ra3Ji6|RBdUj`j*Z9=`fIxWi+;Ax)rh$Sd<y<&<bpdwnTm_;{z#r6qE
z2ZSOyb{u(3iFD=o4qU}`nyKOEI<J0c4w4rp>y14BF>g5Hy_?HacVH82pjM)k($oI;
zT~oRBb--pgc(tmy$w3X9cVOf9hgBH8_xJVI5uvo3Hz&laF<$)AS3<m78#%-6$)5Bl
zM!k!r2kqarrcdGf`0iJ8Do$*9vyGG9K9N#H2lgha%Ig1OIDtKP((M1E=HSvzWo1ne
z^q)9)l$QV8{J+^-wA63N25qWRyMa0Rz?SC;37x>GKqjPy%&8m9>G4;qTwXQ>e<x*#
znhT$>#kh(KooDWqsVNkovIrS!jY++7Kvw4EH~KE)fr2Qtp&w*PYFo9z7-mU$is@K`
znegeN)ATq0)c?cQTZKg#_FdmIbR!_$Afa?iBS@#9ba$t~(1WDXA>G~GE#2LXba&Tx
z`Mlrz-1o+N?r^X%8*|JR=lNf2{T6KA8AX?R&&PKuv4=4-3`YcGye{rEoV8}ANfk@R
z(xJ{9%bMnp97^s_9m!$NOY|XO)Mx#hIDVY`iRcvABTBieo`2s<(la|}jwWwafyGM+
zTS;rv5Ke#C>!9+_Jnms03~;%O<b>6bMhK$mqZxJD4J5z76dBN0p93<gi6-+goglH|
zIP>DC)_6+|fu9;l+(2BLXff<!y9M9)?-PD$nEZRXXu92#Gew4&rYZ|~eRjctV+GVJ
zSalaE#ah1ed6v=?S~V~G!l4Xgou*B$D%p5n1jIezgD}9!24&YEk9;wZ(;z-;T{s;)
z)>}TYwH1B5o)=@Dzacr~PetJ-8;+k={aF{cJo$IWqiy9Zor$KM!f^AG&=#V)c+1}M
z!~shA<p72M8<~Hsl-J!F)BcO3UD<f%+HpC&BfaDoF2@`+{_VH?d3t-O6r5Rhn_cBt
zo8iq|O2N)J$j`LzlFj1WutY7>TzAMF#}J9j-dwDhLx0_mp_zu~(o{W`&U*yZY>hn`
zg~_^nq8)?rq60@tjd^Ynxq-)lLWttJC}UK$UH%r+g^fg4qC*pydkdJ;qQ{dam6P@Z
z$Ks{MY1a^6$7%Wcp<XX{Ef3$vmhw0}+yU75S25a_R(CyHr}?O=rTci9r4qTqi6+qH
zd3SDvx>NKm$|=t({%={lA(i+Px{$Uf&IT>x_~et=FvL^aiuAS|y)G`U^R`rNqFPtP
zV^fv455I0{y=2@~J#sMjRAzIZ%1V?kR8)=xM<a+Bc{^G`jh-VhLfU1{ThHBh=h@Qj
zDRP$tFA}dK2|gq6%GI8Q@9IzK(Q&sY)p2XNw!(#9Akwn3)qEcRku13M{<#Ogefze+
z>;@S43;!8kJHMVM837Ge)DlvM+&0tn#qo)_Gj+(IjdX~Ve!rm<%;<;dH@zZVd+@;q
zXAYQs1bz^6keQU9D}ls2Rut+KkHx}I!MB=KL`w^n?_gTbZa@j(aaL8fJhlX<=olLv
zZf^IENHhFtc*_nL)aK%e;5pJjmolOlIir>Z)^}{*3sEaV9}I@8M@ek{Wb}a$LYayX
zG8)n8h9PqTpm5DtFJ(mvgyx{XUtl~tgAqxh@UJO2<K+qCrf6HE>OhBamYRcEWd1Fq
zMJ3!6zHI0s_Q_)J5;xEmC|DkX&laVW84`8sq<-Nq6NG9{4fZeSOjfyai9n+ib<F9m
zvrhakVTjM4x{%DM11I-5e=@fT_ZvKaIQm}1N$bi1Y<@y{-o!fJfJl<JjMarVz53_t
zXWT=RU2>;qnOq4uGgBXR=q5k(9rV%?LoPv+%BEa{(^|==G;_8w-xv0)MrYHn1f38g
z-<J9HlWfK`%8|2`gM~ye!wJ)Qf0w-?rAlnL`)qwqbGN*Q;Zs+8Gp2B1ma^c5#~z)z
z1`ZQx6o3tO>oxc~1Pj`@?$*{nxga%Mg~d*koCZf>>i=%+n}=zf_xE=PS6bHCLgWMV
zhZYO&IWR3ypZhOsYw~oRraWqf)zaNGCA96!J~Pa<g)iTDoG{U?+K0UTzNdc~ZH9d3
zhyUJd2#_JQz&>YaJyRiOfoAyT^2rcxh%Od!!Wo+XA=3rpxp!vkMpMl4*m4%#o@55M
zQ^Gko1qVfORe(Q~a*S25y9G9I>UG#>SGqT4tl}2+l(9FuZFJ^5`vo!*?HVX#HM}4p
z-~Rn-g?;m$uB6SX=$mGfwey1Wq)CvDu|q`+yW7QT!FZkIFT8$y9$n_#HAR$@@pIX+
z$Q9cbl`O9Z+4#`lHlF3ckHdZT80>3=JJ_XTH50dkX4#VPt|4eb&jN&y<pDmHm}648
zwW>Ut;|Z<f=TDFO)i-?HSBTetCEq-xx=J4cjSFBJm53>43?-q<0##wKi$MNW5Nt;2
zQKRYYt}*G<TPXcDs|(Qa-m0ek!LIcx>-&4g*o!Lm%Nn7(bw88of-iL^ssH7o3H&Ea
zzZ2L}5C*hBz~{&`9lUZe|LGK9sMKC(AsHF*b$d*5($Fd<@VzwfqO@ARkucb^sb=oU
zXCo&T>?7!_KyS_X3t!O@EG_~Xq$2G#XL#O95nwa4aBbxVA&K&z(ZVJv>6(8I;{OeU
zh$X>B;p@R>4raUpiNJwBoqf!dyU0vimTnWl%Rqkvqex7FS^@eflAQMRB78C4Ey<Lg
zH_hCd8Y-w?<Hn$@mVx~y0G*Y}@+*#NNIaPdEH}Yg7(3NP?{fK88%tobT+%r2k}eb1
zPs?DVCrRbf^fNg)7|IMdn}R7vFjYsBqNxD=WxSW^8c%KscaV(t$t*&4fnMXNyZ|Mb
zS7U5W#<=BhYRCM5jKTPYHk6QmmX?TMn}B}kZsL{~L3PD76n&dr0Thv2$(M`|(X%i?
z1A%!#R>P<nBt}wtAhw3|Cqh{ej)`?_l<6Gaw;r@O5FuE0E5sFNW30od_a~bXozc#R
zP0Rv>wAej|qS+|2GS0J}(N1T+aOfv)ev6uW_^(;3JxOVtp&ZT`BnJu_)8r7HynP%Y
z55>GSw`|U@Y8Hde46oCuG|ur}Ii=w|>pz>cge?dc7JWvihHR>^MXh!M{5HZCZ$~$=
zq=M&g4YsF!-6C97zb^Un(DUO>A%<d7ggdO!%T8JO{5}u>9Yskd-IV(!3n-3m*2%T+
zE_8L_D#5eDl~935Tfv;hW?Ljo@q93=3jhxscPNiVoF-&@b}DFKm%PPnZf^SIk(?s_
z_0m>)aM(pyp#+4`<V<0KV}CM}pB8H%7J+_Z&v*TVx;0oFwA5s^Xv0lsj(-{LhVcZG
za_%Vv1jE>m_P!;Tkhne}J3g1zOM5x~`c#9PL^iIdC535iMn-jOOXa^B&;mOJ8N(_Y
zyjv&aDIa{M7t!p+R-9F*-gp=-P87=Xf%=DXtU_|?-SV$V98NSYwl5WBfjk!wf*Xpa
zk+l{S6@^q+SC^9gXV;`V68~?i1WcfS{B6Hcgdi;dZR3vMosHz%%kUd6=o}ZNJVNOj
zW9?vDPoSq<M;VhR*)s{XYwMx!V}19&mz*{9GeHFiwdE&0EH$~gh<iAg%E0{vR<4qR
zA&|?SNy1$}I1iSwV(NX`XG(Qc!2l37gcZyPh209NC!QNXjqDWM!c^**f_npE1*7P%
z8!a2%VQWehM|0{4e1U!h!68auAL%a2q#3;IfD23v?Ty}U=Ki=P+do73t2@{cOy~2E
z=)zJ+*)kjOH|=Iox~j#!GH#{D<<u?MS_cw2E5Z2A*_+eU2wd<72iY$LFr|w>68LR%
zU02!Pn_b*YBWlyiBVpEE#d4cARM)+Ao7@!<<{s{N3})Mp`fa)@)A1u|PmSHi;4;I|
zaWQU^`PB|Xd7u2*x1NULeYNtV+xM`XIK-dHN+t$BJZoiBIk#vN`eNPS4?a%N)wGt9
z!G;#F=;{S~8#2Qrk$7Um;tW#~8?_Qb{PJ<u<XIwQ#*WZP?eN0~<~epegMj?AeQQiO
z;$~-(HUx`3;vC6n3ah07<cQ#06Asz<g7?Pe*Up#(O;V~GKieEM<_goW3`QSiiW(1P
zV4l5xk4IpFfcyn2KCE$OEO%P?`UIQkams|}I3J6>R2$E2rcJf*Tt8jcNE!^hbbh+C
z+nqCZD;Ar%G=Vi&@l&U$5M0fPy+qwR<=pZ?!JrusRm<F6nVD&muT*bpV?+?sv!oXE
zr-rmUdPg^?h=4-`51Q;=p-MD{tsk=c%M~;pDnvuRBXEu239D2H3>!}2^+`>bQ@K<m
z(w&N8<sSbGs;dO)I=3D<ZcS7%=AB}V_<5ea*=hxi8`%sGPHtZ;p2=<(^XQ?TVDGnX
zjyiVB*Gt>lTBRB}?>t+s2DlFkw+fWo-X5q{Sq~Z$Uf*ugG<6H!!fU-qW3=C63`{ua
zlf=}@?O>lW@8CHlx=^-6?knz4f}GsNr}dfBqCt*39m>x+axJlm;%kiM73fY4_;yA+
zrQSB1M4hs|@nN2WP&d67wN+AwR?hAs^p5jWWMb{7$F<kE<NCGLgN^^|p4;A@@%$GD
zhJzZd6vXR!FNlVoEp`;Z^c#HL-q{&!F_#n4*0wxDEhj@$Lo+~A3KbEcV;yCCIUBUd
zMqtx@pAc}QiSpS%#sKYKA;BG74;Ys<lnoAyMKA^;?dA$~r66(J`~=eg3XWtON9-aq
z!s&XOoqqLX@Yyh5<y4f390{BP3jHD+Y9BgTwTOVq1b=3aK9;3x8CvBlR!dX|WZ!?A
z({VE&cbw&GgHTMsX>ZOD)I}Ej$H9(LewlK_$Ywo}E1%IdLTxe>2v86{uD7V8JAzP-
zQ*|^<o_qKdw$#2g%Nf$tt80vWdG`vlpO9!r5u{^OTm!Ui9KDkjp5y_cJRk%ECfmgi
zOpl^?zXiy)s=RGxH&F`Z`zHjZjHALCtk!YOpM5pIEt2&PLSr!y$hbfV9sM3LEHUQC
zlN9mSR{g5f?}S@Z>-^sI!ucUx@n0IB58<*u2J8g1X9=Sp(g*mAR<Y3vU?~O0W2MPM
z(Aay=4`G`U8$|m^I<`B*H0(kze5;3PCPDh0TrUYmp$zMYXM<1cgkTu(7>HPaMMQNH
z6iw)}gX#4<(+R^)P_#`e|6@4=1<qySldm7hJ}sNiOA2YqR4Cw(fRA%`GU&PcBP<43
z17vd1yNUMkaX*NB@NDXpndWkLA(g(}37>bTGsMRC<*}?uu4rz4Tt6GE)z4eMMguSC
zjzYOw<9o00%DT2_=?3LX3;L!s5%4K+qv+T<O(~zMHQns(flAhCM6BvhgdULn6|S3O
zcs~6YIkYS!96Yyhmuu$|7?$j}fzje5LLL>^21}rVx%1KMlQ2vWax-THo4n-Rjdm&V
z=Cn=(p<lVs<JUGpyA}R1Tc`W$Nr(7WCh`ux&5r8P=HN)19m%}WpmWR@S=z7{HT!re
z-8Jv3<KH~eUd`bMCj`atGi<Ou!{m=xpQgY|AMltWBg$Wb-pZ5M-R`*Te|TK^o*%pA
zTIGstlXqOQGC-m)3gZdhf~XS_%pHpG?KTebb80-sk_!_Zuf%0r6!2#j4S!laZo#=f
zhc+IhVeJa`H7C3ICv<JD`lyWxm<9(F!ED{#`iDXP)mtq9W&)=F+Hk2|{u5}i5vG5u
z>AnXBsrf+oTB?=kaKij+-HMAAfGv>B@rq|=ciEwG^O#`g*kU_kT?83-kX%Z95lcKI
za3H8BXh#_R3MD{u5CKza5S(-%cDu_1e~QQ~;W&;#`Vgl#X;Ntf(9vc(>BW3rdx-^t
zDuUTUWQ;h2ZWtJ>ND7u{Ymp4mu(P+dLdc`m>VKMC3r#bW)2&NU(;?&LF~c3*B6MPY
z_79|rFsm#zsPhtO{t)c&zS|+r^0Ixb%fL+`C_aA+4A!D0|1hfJDdO<C5*U#la{|BQ
z5uQGq9K=n~LHQY0YTV~)J7!Io*-p-5@We{SZU1<ycSXYTH99>ZQ-OC@s1=Hj&WbS6
z*AX(#Z+Da<nM{hue=^T=EY$1k&dx#~!40AjJr$uqln^+kXt_MK)27Ffp+#sc(WAa?
z5=W>xw6n)%i>F5%SJqT!;7PkLWqfbBihv0iuD8j7jx+9ufsRR+UVq^^Vt`BrCpBOB
z(wM#T{$+K|uZB-f9tS8j%p3r+9&{u=EuMg3$R!+TPy6t%&g!V;a&&ThoPsBUfHGd!
z)8Xv+1F8MGuj6Td)MTcY2S4VeG4oj{1qH5?Z!L17>##$Hb+%B@l_1R3Sq_^bq_<(v
zBf#jyAT2u3jMDsvpkgTUDL9Q|2X_q3q&n{%lu$nc@!dWouMn|stF(T;HRV#^LSf@0
zT;=dLf$o>f`E1ADbFu5eMJ~%>F~|P3@QPhjb8!zs*Oq6f&u~)s!)Vj)8s^|WUD~>>
zvdwaPyTH}_X4RK+T~!uDv%?z9b7NWV4hdb=`t4lDn?$qz>Q^_$2Cc2a-SqfB9>+G%
z!>y3!FHWm;E!l*q#Ng*GjSb!{!kmbG*M~SxjF(2U_}9J6@ukF<oz*jfNV0H^mid62
zX%)p<-fm33Cl6r6M)i-BngWP<bspvaIY<1T%X(^rXN~cnfsoWl1iX#J(`ki)$(w=Q
zfh_fTkCKd~1ToNzD3Ds0AH)`_F3Ki*mh{C4(_DNWM}qDwOaSyA@+P3yy)4hnzzxl=
zMgL5@(;Q*d!2IuwuLH6a1qm)Sem0IBcBmdIhYCMd*V}|36Zt!Om^y<R!_eq-yyID}
zsYlUMA!`K=(d&o@6}kh2&+7#1-E0<8c2+$WLJU{ba*Dj3iZ75RgLRmkJ-Sj`{WHFO
z+~y!Fl9)74b6EjKWbvc`T=A;YDamHkk@0_u39S3ELEmz_d-&m8KeCa@)?JWa^|e*!
zc>RxT^r^sDHtP^0<N2LK+sDi7-Tk%Y^%9H!X>e<TGh<Tg>FIGCe4pQs%F^;i+o1_i
zJ+``W6l?B-XMbaW4I!?ei`U>LKyF9>P%3+cYLT=ke-%h<%PNBVYo#RppO$`XEh91k
z*wvRWkevVV#o1xo5y9*8QI9}MrFI*=faNeU4S0h5!(kGq{8y=~WgE%JnI%>ybZ~4e
z9zcBXBlc#FH;6(kD?t3?#Q<*2Iwi({jB@2llU$HC6laW8G#80~+r2>oC-dtv>YGiX
z4U1*jJ%)|2oYSmy5cgy?8UFSLyjWXK&lEg?R09=7*mXEaP426}-r3P#l8m4}vNz$v
zqIXha3rKN+Jm_ifVteS^{u08zb+W)l>j>zpS%4;{;q0J5x)|t$TET++F01^WW4J>1
zqh5#I4oE9^J_m%I5nKi&e+RS7EBP$~QP`Kt{T=rDcwcUarVXkA@<f={l{`?72cA3g
zv?_zTj$9W!W}{%vZ*&E{B2}e+pOpTmK$WFL6nH(}Ao$g-x*x6&epvB1ZFL0l_~o8v
zPZM7oB??ZG4C%fIgv`x(Ke!)E7yelY455KL(+r&hu#Gf$PMPiPwNSg~hifh5)jOv0
z-)&PnW9cj*_&j#4vj3$>e**~6@W`7U@`~tXOqw<Q+^)fMr&uR7aA6FVx)3_PjJdQw
z6D$lJYl<@&7_doRuMJst#s=FDzFQa`cIrAp?xpWOTDrlIfhV9A=ZlcdFq``ATDRtN
z(p3dgeiyW|l1bo|L8mBmpW~Y#9DT=@)zrlIQLEnk6;M*=P`3AQGIlw}1aPz+vLZCh
zN*gt+*LH^!r%IZhi3SG-qE1$tW##3=K8ZeLR#tv79ZuM2-TaRP<o)BHRkAY~EcQX@
z2t^FL>jUqCjr#iaQV~V=eTRkDM(POGMrC^-HynvOOJlmCKORif?Azpq_5B*==le0A
zxav_phL2h_c$ah}mq0f#a@iBxS1ql_Z1Q(5%MOyLJYj0{-l9@;3MuqM@htW2(lN9H
z0r<M~U4iJ;oaaOWkc#gI)9*i8Y376lIf{#mSAE}a;t>aj^=?jZMIa}d>3kI3!m>K5
zR7fX1bmoRNFv8<*G|&->BhptPnN2gwx12?*<>L7Er1eEzonEJ@gbV>AR;&8;KVPsV
zDefPO)Bkg+g;K4}lIq~#;17=SHD!2>5m?8P`kVGFbQCQnP2otf67+gX%b&WAP^y;3
zS-8*U$u@PBXZ_AUYiZ6i0A+jX2Sc%c-z-!8vf<&#b2;+a&4eU*Dca16K@jOb$@5Qr
zu~j|5dU!L#CHQM!u?54|*=Dh(vi;@xaVQvz+IxD=I`G`}Sn>g9zDJLqTOAba5a!GX
zHnRR0HB+`?X@U*IZ?i$A@m0s@(0a6_piW15SEAVD?*&SthK7xmgLiF>HtpU$9Zl{<
zZ)fhsTJ-kC(=xW3=IbhpF7w#IuHO`XzfpC_*3Ic||MA*>y0<Q$L6f$yMcneuvo-JK
z5cyX0`P4S~1J72r-S%m^7hS#&uej;AUbKTu;4587lKJ@0`~_%0|GR1pE|U3bSSx2S
z$=HFp>i*z3N9Qz|*VUH=!w{yU<-_`KD#Vm;U>*)KBjC)0b?Dk$ysJS32KoP^WIkFf
z?iNzY2WDsO#5unhY;yNs+H}YGaralB_(%)%S}{xC{ma>qd6a1CvA!kKirA}OEzF<*
zF!DD2w3r#R-YXknKX_CdW~s<DOIit7+RVD}XZ5!T=DV0PkVJYkLpnapS{O@CB<LFL
zD!ROZ*RJHSRIRu%GSO+xc~@0kJS$Qz$uO7;aZ4Od=g^V~!5hdS&l~hV-!$21X}lQm
zM|M$cL9o51Yc%e*xj5=F+M4hYexr@cdt1e@b>ySXNbY^{6HWNm`rcr&{twTHN;~Nb
zquH{3bq@a9MTLqI18M$++~2olHcHv(#wy9;-?(P+S<8P_DZ4#34VgUf+rjbfC#*p7
zOXdx=mU@2IZI1NraBoTaR50&k6(!gRbv>k5TZ^S%O1@U{qEmO|an&j0rOk2g5x{R}
zKLziU-k+0N`9vdnEz$^5=>S~iW*1)7cH}i57EEpL$`;@i!SMR5-!~IP?yU~QM}+zR
zOH{N>18!&os;8eis^maL?r(%|5oCf}*K>ODfEc{kIA>RW{f8ZjyMFLmtcTODa}GrR
z{$BC;RXZ~*Tl`K${tVj7cyr(nH5<HRZ#*xNb}8ju1pLFiFXV#gBw!C^EgFJ;m^cv4
zyfLgI%@UhhyM3<1$RBhHtmduRL&2!euDYtbF6Y~;XJi=&`;IE41RD|ZGExzUZ<B*7
zsLR2Ot(0h#*2&S<6$oR6H&Jhw1`?D$b>M9HiQq>9hz}JT6%|I|4%H>8sz1>*T(xx$
zC>h%$^=na65dzv-R@SQatJge(!lQzcK<^2WeWoq02-P#jdae3pzno08K&ub~QB1&8
z{-C^BDl<;6*FJmj)cbgy_mAmUuF_LYQ$a(P2S=%s@2L$l!>^=G9i!i>HLt|(7XycC
zT9zwkHyxy+`2fO#deJ2%x}S?N?-!*WX!G=PEIhD{W|ZHS#tr@zSN@_(YeLhboY^OS
zxhuQ<qY;}yhq*_k4b`&FVDNZJagxs^PW-g=hrVEhusq7EONv}^K{UpMa-!e?hQ`ym
z$DjH`RZ-KpE%NXF#wV@?)9;$mEfVlUDQ!%I#^*>bHA;7f(M1S47&FTxipZMOKCF30
z9mys3rF{Tca<<O{EnwXX5`Wfkd9$J)@@^!<)-_M>lNC>lPNw(ih?nfc9#|DY4zptq
zlIa{6(*dOA`BqB{IYEuAjhvR-b#^oKr>ovEj6TP_445XSri1YK+WP$6y>{4t|M#xn
zuJr~E^_|@?JD^7XlLn*h(V~?m=NKR$FXFP|tj2%)aNHwyKUu?rZU_V*HIdiL7B@U$
zKYL=%|0$3lSlJjYC82&H?-JykYz(Et=mZI0hoNpA{>|fwn$LPRbv(4?0otAkMVK;B
z0NqKM0eIY5A@K9BG5?&=&|H7ZNhN-om~Yc?;3sP$J@epd)Z%%zGAWqf(RpbuxqJ0>
zDB`+g-WeZ)^P_@A5^vxM(}yIwk@>ozuJt>kRJA7<dm5D>F6S9QbjX2op=8tP+>=00
zg|VLwl0A3*2UD9|A6diNIx$wGPedR|KW{K3vwR>%$Y>Z3z`fL6t7XN`%@RWm=ky1E
zTHHe;Ep6xkN~K1g1;J{{ma!F&Tde|o8!y3TZ(`&6Buhz=L-SjewO=xcF0ZM|7n3Tg
zb~O`NeX=){VmZ#Laob={2C0a>Cn?qlra{C#HZ=$bLB_Oac1-Wk(NdtS&cJ~cMxBgo
ze45#dm~nC{&8JiN8&+<M#h`DfI|Ee?=~Al_D>RKpW`zpj^8%E@i&{%cEDSQga?6@z
z_?tf&I++gsS|pNA;E_E5o^y4yw9eIeveav@D1)np4DZR?QTT4!?=%ON+!QU!jkL;~
z*oZ=;ugzY6laqNV4Vo&TDwis+(k4sC1Frm4j}wN4<3(csql{Y^$gZlYY<<ah;yIL5
ze4W~{eY+jJ%Ut!{H_y?zMrta0Ea}4b<(>iu%<+|$eK>&wwq_~Cmz-aezbmi_Yutua
znxcF?5WFuO^}z6{7Z|~*7)Y_kmyXL#s?+VODIHBsP*CXQ^`xsv=Ij%&+bEm4bSA6Z
z<QgN0j3likR9m;>8980t=-Td*?33a|nt<=DtXu8D*K3&#6aB9@`YbcxE=^FE-2WHu
zEiP-a+&0`eN|WqA0RFeyWs>js$t74Z`}VBskN?acJTiF^5~2IFS18+--ydyDedGw*
zUA6$<kXz(?kOn=fBK@V%7qKL+Ob7*a65Tkwl33EVr9P|?J)#lW-j9#Hd-=T(`Y22#
zMA#<39Z3lBz41E>(psu*%N7-%3oqxtG|JzaPvAGjwzr?87i;*JTWYJKW6++UkDza$
z_o40}5r$I4>#VXA*r@__Y3I`1QbrHiL76qj#4_=H-ezf|MSWO|4vIy2;Yr$-KT7#m
z!AyK(*9u&P%6A2IjZb}>wcNj5u<m!>%$;6=ueW1b&a`A?Kne}#MvT7X(|+e51<>i4
z0xai6Ycd^%z=hy`0`Bl{<JS5w<x_Z{-OHSU@G@b3y_;{14O1h!Hau#;AOJ({F^}`<
zq4?{O_Fc1u(mUe#Q%x}9@6(3|B0PMV2flqtd|V|~*29VE96^*8txv<AiC%(tLS!YV
z<KkG@{!(-qa8Ro#8&xoyE3^BBUm)k@GSbuQ)W$`xrHb!uzl)jPyX2Fmd|$!`3O<j(
zmhaa+RqIQ6tt>v6i{FQQmFHW7S{|-q{@_OOlanw7y&NNXxDbdvaDGE<mw2?2cBmXq
z;Z=F~&JZgl5jRakBm2p=ZX~y=iElu`WjyvVg@;I=3&MTnB|;~D#kWcN>y(h=!R|;5
zb7-Is(^;<|oc9IN_&CZ>!^Jd`w!!HYH;`V#96INw#2yX-+M?D4Jli89G20|tJj^5b
zY4Lf?I!DN~i@xnA22tM8ja-fiG4zpmT@9}2X+)ewieW?n<(AHV+!g<66|md~Hv0iT
zH(r365UVpLq8<VKfuW%ovzdOw%G))|`fs!b=167o31vH2c@k7ER2guvcz!srG0ncH
z1mO^6a3gEpxyT=4lnCrlBQ&MU@X7BTh(~e)V!sS)5jf7&HM9Ckw8S+)bWkJ&)=1R7
z$in;e8PVKxO?LHsLI>?{YxKu8!|YrwI$oL@nkSzwiwt*Y#HBVdO_<R_ol`;Tqma2K
zOQ8<VtDb1~g)E!8Kct<wY#sXygYN{%cw%ME3GFcmX1S;h=hy~G2|5^h-8$c@&CNf<
z?-wd_z9NhiOb=?LD^J$PN*#)o*a9W3s+fa)A$gihsP9nps5&Gc;@ZqS_rw;v*fI<U
zs(9WNn*MmlS*C02yyzR9C@5u^r1y-1Pn47zn{!;iW54xK_h*r}xN5O%!tPkFEdIwS
ziMDz8&Pfc#1lX2wkL3Y;#wvMfT*{asYog=lD_O>4u2WS*Rukbi@ogyI+bNrEKk;{F
zwQax%@(zg&Ldk^p<eTUPUO_{z9#8lJ&EP_BcC&q+htDxfnieUr4uo{;boLkplS~{d
z0trY=a48mewkKK)-JPPy42(r(T>70v`cU%0EK+T&?0aS53*&IQ!x0lr;pf4Yf-HAA
z!F{z}RvwxOD7jV5cJ2Yk+tS!emKLerP$I(ECYb=8=D~zYBNM<AA3vayKUgRh5d$<z
z2;AM=vIsQaIs731ASwuSxc#dnfj4U}g9x7j{Lw#&{XkH|LO7w~e)RNHdCSEw5`Y0)
zeEc#v+Vaon8o}-PpG)gPvmV&G`953>vDj4JoNx9m@ft`lj*p5S%OiLArMO1Iie=Kl
ztE1Wb<KuTwux0)R<BG!i;`z%*ktu;#JFAPT+9m{~Kwr7s(T+Cw(m@onl2qR@P3Y(b
zftu_;XoKnBG)&H5l*FW{O<=^JTF{JNd@VyMl1NISEsNsHCI;VzXSFS3q7656b0}j&
zBD-!L1b0=2vXucCdI#DYSYueFuL+2usJPS-X9S>0HbFxWwPg8(Ey&PCrfNMkP>oQD
zQEW-X8(@*3r-9Q@CRNZ*uqLJ42!H-W(63U7W<*yCxEZ(W;94AIgTmVpit_n>fGB{S
zp^;ZC(5O&C@&|U<j@;`$*>VrlB*p}%sB5d3!Fln;+J!C<-a7IrOE30Eq_-(wG;9Ed
zn(5cC$*#DdXzzUMNe8s~jTjwD{&!5V2Si+ZF`Ba8s&=xlXodpk7vOW;oZO_KbW?-p
z;blsJSeL1KXhBXD?#1d^BevGx9VwN;QN=i&a{AaZ9rhB;eM$kGbK`F6aCWfPu+zia
z4wP;f$QNekZvN==yyH-=aXbe!z~Sa`AO5n71NPSt4!<AJ-u`J&Vu53l7!;j{DHsLn
zV_Y6nk3k3+)1YS#iig}JMbIQ8tXtuLUOQ1)Vu1++PeYYV5Se>HC=r(o5bU7_c=@*W
z*Y#NZ>9(~qs_f?D6a#~URaU^3>>s|iA0W!;?`F=ABitc`X1`>5%Tn-|48G+(s_5EU
zvReUU!Mz7yiu7dOH<w=I9OiY$jOp{V?)597kA#B`JdV0E77wrlm1uiyqW_!}i*&Vl
zJtC<6*UL018aN&+;o{<&TDM(I%KpKvw9jr&ca96_0oQ=&5{y8|&`&EUvlW)CdoZ72
zbffSR5RN*#5T*VY!;oPKNs|#cO(5WR4m@`dxS+zxtoI;%?(jo2!hDQyE`^d2YGRNH
zfaqsyVZ6bQQix~NVScvP2v!aF<J}L^kVhrV4vfs`oW9Lqb}++o7gw8@0p#6rF7uX7
z{#*kHlvM5vlV?zIL2C2Io+*5WU^-h7WbU5>?mqG^rMu4b>JWJNUU;(roz+0tLW+m$
zjHK&jw&}Pm7JD}XolO}`_(Lp^*B^30h3B$pz8)lGz{Jrh+Sp!`w+bWx6!&OELb+k0
zq+m0$ZLS#XV{6B4A*be6J}QgZyOfS{uh}I&aaL;+@ze3}55Cq3cUNRq{ubX#ea>KD
zVc2&<ku4Hrb25#<KU>_PDah_XIH?_~Y#<c+-jEUO5`4ak;?>7SnZpvEM#mDGXtu^9
ziHOcL-Ol_caVPIOL-9!c^@m~WER~C{;ASWXV3SuZl@EeSfD(c0Ldv#rJ8s?b(%;P1
z#ih>XgNnMk`CQZYeJc1r3KdWK9enYnKX@aPk}QdltqY(I|56a103+H3qlB2(G3S^M
z{qs|<Z_2X@af{h<V``ZI0_J5XvtHHbK=Xwo<6i~44CU=*&o!l~ml*3h9o6y;AI7N0
zU{8&K@C=6lr<+++psN{>P))=st#hTYdIGPUX>a#ocm6@n+;)Ryo*F7i<tc3$OfyOQ
zX+MRue&xzxI?#-tMthm(X*Z{2==e+E1qhD!>w5e@E;17n6H}<6$lUa_l)HNq;R0g1
zm#;vpej$-98)~UB&u9HN>j4#ljM4&iNMmpf8{(fD=mJy!6n#d_g1XKp$}!$an?nuZ
z9C0Cit`OLBatWF@sA6yE{Q!(UP*(@T7=>x_1w(>bA06|3sB{OBNYHi78Euzr*i$2h
zd9scM&9hT=3CRb)a}+3;(VzxfcaV4WOHIJxbkK#bbS$zqoa4sK=X-*ZBLwL*#?|l*
znKltN5iDnYw(kO`cg+5BkaAw7oieN>nb)c^5;D)&05cMFic^;)o(TG9`PD0cE4w(s
z{SaEAZkuts&m&4)T#E)RJhZT>K@UdN5hr-@>pe2)lmD5js7S`z{XkDLXD>IG3+cbM
z`sckLh^CsYv(o(kwTgPha`d^sJt5nG@DNjJ)Q6z<8K{M&j}U%o0FLyCj(>mBA*iV#
zpO!Jz&5Qo)-riqpZ~s#d(23LGX1Tu(&%MbSOXg9Oh#;emk}5iCe|?#3R@D31Q*`VF
z(IbhZhP$1o6gx`T{hf#w3ImD4@1vA*O=kB$?BP1_$o|u|>jo$1gvJOve@V8HDy{<&
zy2J7_OIopc?{_Pk*j9ehe1Uzx{<E*d^ZUiKQSM5H&Cv>|V~2EWHQyE%tA!tbk3ZSd
z_ZWhE30S;_a^SyTQiGrv>GG@bDnrGLv4rIYEqdqJ+o|FOT&HTQ0sbeh3Qok1a_>WA
zB1$+dw42n08a@1n-4D6cz~4-pJsUUlvbV0<$=+(8|A(AzDFc}PlvaFRpB(|D@zOtP
zF^NMr7W>nvhw)ZufMv(bK)mb_8fzk5B2g-&Xr!nhE^UD@IxjF)D_ag-nd=3l2Bk&g
zi$w49KQXX&`QrDc3r8d%9zj=SD>?OVsM_#0cj0@}P)AGz=Drn<sa$n*nb5ToWBuPt
z$T+O#y|p_NOdN(*SqA=9FZhAX2rLz7rwwNb>}<+2d7ND_SCBTW0}GJc5S-G6^AGht
zw7b4+Ux%cu*}6-F@6L$+3r*G!(n)+?Kh8|WVt?+ES<KRC9?Lu&*Hl!g-?$BegYw`a
ziR;`hTlT#f1clwxUPjP?(x)U{M_poCyeqolOJL|z@7#69X|z&ZRP>?oyqB0s;5&d<
ziaD-Z)2&-@x3XAkYZbD>=%fZ=MIgQ2fo_mt5%3(s``*KhUWS9M9RhBVP6;yDJA&WX
zrL75g)w#`q+8`&4UXeVj1yB!#d3#?oa{aJ;vj*wjzrSSMHUc*;4yZ9T8qN2D`AuS4
zZ|jYqc=qW}GBt98pi$V#0a!)7bzEWnikEe&)^IgOCZ<Uv(H^5P$WRkG7>|&`DmKiC
zeb^}Bc3c9=%;Y<6I-BrhJ4~O*>oSw%0?X1SQ75!lzNmnoB>!lE*g;0n|M>gXyoj4b
ze~j%Jse4zjCj#zD(2DjV8^w#@rQ?&}!a%&MpfI!j$f~n@`ZH_hoKwv+0K83JR$?(r
zcg_=Fcdy|*a4i6^`E_5EoZ<xkYm?>pZ~uNUK$jj1%*>}m$63F#<&-tJH*9U~BvtZR
zxCQufg(`;1iK2n!3Ctls-rW3rKP!$d(8<PDaeg9_(}mPqP|Y>LDJVfDp`W*-QW*Gd
zRg<7v9YO$wq=9ZaS7Z;bh*WSuy|)CDRw%G+aC+x`y7E?H(B=Q59&wE&pVUtWLIyIi
z6fqvsucD_UN1#(lD03N`8tVk*@8RC<@QPBFP+_LA&01|~AV^g&z*0RCl{1Opj8Ttq
zjImfWA!J5C6mawSPmAs2^`S)fPH3@zt92}Y*+31}C5M>GOL;K0E{Uc*z#>o{k+=Fl
z7rK7HPUQ(zns$pYNUa4EWB~Cw7kst<f-fTM{8&~q$-QD8lf(!Ve4<z>z~l~cz1sV|
z(Vqt59uKxf|BSPfEId=|^LL1^G*0A=={w0Ttq-Qsua_q~@(A5jHAyR8JOiVw_Pv;}
zSLBiCr)<!bzv^MFw@}+FSR3l~81nabUwnlSK`@5|<cs2Kvk1q3C1DPENQ>*!y+suS
zHy~<+GXe2>@746&0(NAPik1@m$m~x|<1vjoJz~9cd?4hLr?KMg>@0iLTx+@7A?;N=
z_esBc4B0*J>>Dpf`PFREx+C}iN?$#R>#{@pLa)ewY+bKb|D>5;lzLUhKaYUvzy%;j
z5=2saqO!z?lk#(0M6wypR_wRD^j%59IT9bxxUM?{IF6s|sE%IOj(9*3U<~bt3D01`
zzDHWmTZAwWkRfOcB9|0|Kgad^57slr9Sp>jRAe2Ewb(}H7cN=`3MG-%X=Nd7nmm+w
z7TisZN`N!bg6x10_=Tk<8>$`KUQ@RU#fU_*`*0APKXqhYTLH_sssM4)xXamrU$)fF
zbhGq^Bu>;N{+>bDw16ecf-=scpdIFatGdSV6$O$w;=gS|C%zmaz0H31x196_VXgnF
zu<syDs%R1Awo5f}!%~$_gcMHx%f;?|lhAV>Fb>j;y4Q3qo052etzyIT&xfr7{rXPP
z{4y!@P{fCXo3ESH3Ks7iE5ux_?qKa@^tYJP^9WmZTd7qQz8dq1h@%t>uX1PY_5W4H
zovW2SaXcRSkxqC~jyK0-cy~0FK~eBIH1zQTb!p;ZT?{iOIJ}zMAocI+*<t;vyDSjm
zo8jogq%FC;_n7Z_Ti)<*U=3u-h_oFJAHe%SUXWAka_m1@^<s;E@EI<r9XVd$qc*-y
zc^zfVCm;Tun+R2PeLTY$9wZbzkDq`pO?QB&&R)<3T|eqGd|78AbYBID&R(g)Wt`!=
z3k3EdnZs>$#GpE`=!bZM=pZUlWOpzrFy<S`czKAEpn4?W@U{1oc#FpMCE*2o$iXBU
zG4fZ-d&G?HrXNt21CTL{=C>KHd3Ws&&I1?%7}W`5(Z}IrxkE+o7Hu)5zE&t51twrl
zs#Bl(JMsCE*Oc^gXNj+=NQtLlZ!zN$$2#7#yb!c{_5s-(MXVC}OO62dO@<Jp&}Af;
zU_D!|lkz++OR(b|P?3H+s8D~i?l8kV@~7eNf##B6d>aKRnOP{V5~^KUZf)%-4r6rD
zwIyeNem-ECF@V@smJzOp3oi|*p2GhCgGLxsrzXo3_OTw*fCus+mE*AHPya|2vxkuD
z$vu@Fn9Xu&>lUStubuaA$vB1x^bYXe>Bw*OJvadJn-l)mj$A-Y7Win@MnLTH8D5S_
zt8TcgGKL{p|LSmVY&4mAb?AQpezbMWYr?Phidv_aw-k&0P1z-M!O-OVbnhm&k^lJc
z-MIc?e3gF3#}5H&0}Ic}_n1^W;6Sr;Jzj68$8Y!E@CLEYqq@6m@~b0jl$d;RC(;~E
zh8Tge#3~~aW(iU|EIWt{ghB!<dA2~A6M2{dYxagr0`hHR!0#3d`b$aV1fI6?gj~e$
zYBPjxu+UCh|CTHF(6v$bp=M=TA1lMp6C?~VtOQFvnc}QtKg2qkb9;eHK1f9noc&2g
zK4m=^%){;A6hX70BO#*b4i5*VW*ulA!Y+1p-WO_em-bzvCWsMGj0c1NJffDD)*z!t
z-+iy0Y39{@+5LRKp@@z!3zeHjKBenK@df5H4-suUAEOg8=2m{ZP*UWSLaB-V2-Xln
z;PyI;hguDBN=~nQ7<{QV2U<JmGDueKbGDm&-}@=oeEhU@SQc=II~f$)<CQu!l~aR}
z(4(m@J<HNjk4B2KBZDNwRR|XAAcBUcpmG3>nl+2Sl%OJwa8vq~m=o<GF_piN9~vF0
z`dK%W=@xskY-Ubq@;)Kj*v3F%_n@I-h`n=&`5QxvdS56TyfF-BBJA&LTX`NA0t7r5
zmiHlNilXWkpS<7Aq+z<_oJ$i@dTw%PjGYr0hi`5NsZqipMTfrZNQ4toy9b;@UiRrs
z%$LiGAUP}ShW<4&(&&?=?l%_HPCx9wc@SNr_Qur7-EScBUfB6?jBP?TR$Kt{ta)Qu
znfy;u1K`+t_WS;*aW}0FS#hb(=J=`pEt%t*_hae53>J{tzFIHpf1jUX$bVGDQ1V~C
zk*R-=>=5w=wznFH!Adi1V*(6*2pBbyA49D!k0yjc;74Z|nKyx)QwS9LeeUSlcy|@p
z+<RR5@b)lLP%Pq^OKq-`0pe_8B@se`+E}38)l*+!QwOfbfCJ$Iw53mgLl%e?Zh^KA
z4z1XSJ6w%YOGs1t*nu}!fMX9k-|0$lP(mY4dO>AbaOgwRK{9!G%k1^zE2#P^rZz)e
z_!qaUbOZstVQXG!jw*uL83ZOe)?$UV9yt}2Cv;s`1ZzA@iN7V|?19Q>w%@XNdx{){
zUFLwfeTghEjwwCi_$FDkMnR$OGS^*Bn5De1z6uOiBHRN3ZSTqUhChJ7dps=xNQl<S
zNmSQd9NUk)t{QFfg<e4_@38kNRTzqfV}8Dv;_F`I>AeqHAv&LgNvTv9tT1%c$UVN0
zYF?oC=M27XxTsyZB3%!#|NP+6S%pG4PCK-N<JFaQ8T&`Tkd$R{EafmY)!-Ja)nt~Z
zq2+6ylaHyj@|6-fcWY8$TYFYQd5KR^U=aq+kc`zflqbt(rG{}ZUxBfxU1}HX8F1ee
zIxc&RN<sQ8N51ilJ8ITyd=2jc9{k05`t{iuooBI++$x>zW>{mf>T4ffhJ>zvoU+T(
z*N1l-uk+2ApyndW*G>L`tHdh_Q3|lAic-bn=OdIG;SWMz-{Qi%{mRJDFGmVC^vcL8
z$S9yUbjzAYb%D>B16oV{ujJCDFV@nXPkh1B?4OikpEtv9mz{1Fo>N+$Tdhh{<92WS
zOf(0!tLkQxqsx;OQ9T|m1gw7DJ6oN|5>p(-J}%&tkoUO2)(@jQjR+2!wBnsFI`ToW
zJ)1L|g^vX?smuH_cT#W6n;VCJ*we?C`jF_gHPH!LANWHM?SGSwS{zp1hE^WM%#pd3
zd|F-%`1ZtgQhxCj0SW0{$ZZ|fZ9rp=9sc!IKB=QduN1!x2}f}$*PzMozKbz}eew)D
zXYV?f!D}A18O-JB7p{Avgo04nm%E^=Fh4zP=NwY9%Z95g9HF=*S_PCya|LtOV)1=J
zJ5IVWOvT+vffpU#^6hxt6y3TpmCZ#d!Gm*;KIR;b`ZOFHj=4|<B^!azYB%L}SbRMn
z8~1xKi?q+97W@|3^?Y9*OR9ru62OIt>P`5j`r|i^1v<K<CgN&pPHxjnn=RG7#PzQ5
zD=&qC4g7rUZW6h(_0Gmxbk={?NHM4pvBLYhSg+@#sD9Z|HmPN#ZeN<tlRYf4)n@Sn
z3pV&2c*=F}pg<&A+(@^<K@nDm@dFb|Pkz;St+{DtR}unig?;VEx&{PUMr<%jM?cKC
zs5Tj1u09+EJonanLOhj=nue0)``$cUKbY3tE4NB?Xekj<ctAb;%ip#aV|-P8lx|pD
zQ0RuI_3ey_TP4~$RJtS9J>kmB?GNUOukT#xF*?LF3RToxScIxIz^4^SmHZY4hcS$C
zua*U$6i=6NvWCG29hLf6Cw@qUJY%$T!Oi3V$4vw3fdcE<#O6~a;Z)B31L^m-uiwZM
z5T?J7pZpR*gQ)fN6UE?N|2my{n-K_)12G0;aQj9*c%1)Q%TJC=`Hi~N^76Prm*;Fm
zm#cBVYAL5~9X9e@MLA+?hpJSJu8kh1X!FSw4VjgVd^lpik@;dhTEn;Kg9MQjy((dF
zm&p_fEar!hl8xopzbRp1-bH^C2r=Uv+@2hHp|NFCv%=2~;(XV84B1B$<}YG&Iby;s
zP9RF7($f*G%g@BaK9};zq{SZUM%&z~c#(^VhcOW&Zbgi1y>Q<V;zz7*r=1qjCIV5I
z)#99MYt_)ro@bhkER)Y?y`L#s8`D{OJfrc8D=fXY8e?BmjMyX|2nguL(k>}t+^k=^
zoqkX8BkWs$SL(Wf#5rl~P~m?a`_Y&e#@6Q~v=bMd_3B8r3VQ+`3niUXW7S;PsrtV9
zc97G@Gkpb8zjyod{^dm*;|&Y>uE^Dco3h;{JXB53gk|Wu1YA(JW^0hxCqg<s)9@n^
z%AH^;AfPv7ryc2iS(ZI{VO2KByX~qVY34qCo-W5P{P)Ym?Oe$|`W-_=@$A)Sk6sUM
z4Rcelb+6lOlF9zuskf}A(>1PxcTV1MlUCA2QP)%s;&ED?t@UREp?%}aVuZmGA+r+g
zB=3uGp4?nDwM)p3LUYb;`q@uTn^hOEOc>6P*ljgQXyxd!v9EaysgaJK`&f(XROd-s
zg=GzYEoYkA@p$>-cG7*0&;{4c!Cq?rhuLK~>XFKJzy->;h0CG{;NUf;BhsF>=<=OL
zdFkX+K2|#975wiqSNR}v0xl7LIgi*EXL0|cS!4bsiPcTUeL1QxPf~FA-ojaTPjk82
zv50BvLX#119fWU18FFG#%DAS?IQfSGwVwF0QV~^e{}UQWlJ&8IeW0kc23k#g@nc2w
z?L18NGEYlp6@j4xwfC%iD=b#QN>-G(TzC1PbA|e{iKtnV{q`b=A8$Us>;OY4d(?UT
ze$gH2<N14^5Lwi7lpo)&vqAD<M!TQryud9`cf9(OTIeiU1qDwwjMIIJBCDK^#t)4%
z%`F)#2S##zSkDQBV`W+qY}Id#-h;RldTab?DUWzZ6XkeE>p=mPxYdq~-8!&LTeuP5
znPZl6h2WJ!BdR{ZdO5vZ3(J!~tVm|1CQse!LL`z_xHt|(FN!In1g^i>9II=shEDfi
zy5F3tEI;xci_PV%rPSA)+gv`lt~gx3Fkaf%Feh<3+9s&Y&|=IyIh3aP)OOT8Xdjau
zHJb(h)fl*kCTUx4e-T`|cN-f6gAQ%Sj;ppACb1CPKm0YHbzeiK%X<~Ns(`3f2E#wA
zKZ0WydkiDEHV!fe4`%p_UG-VLHo7;b@P%D&$5cJc)Z7P$^nTw6-i_Ccgj$?q6w$EQ
z{H^m?nNiDUY2WvmtU{p3c~*RsNqQa+LVSY|_DN*S)7n4k@;n>a{NxGO%6;;wEW9ET
zP8_U#igaoWYDQ>VjTmfs*Dh8+wx7Ry1!{05DHc7Yx7}JL*DpX9X$VqWWla%!DvHld
zW_fbTc6l#&;t(MVY_Hw_LAr>|08YyPUEH}aTVx$XBqVa>UdL8b@oiVRcQ5Vf&(`h(
z3>Y%0ZW|`5(-}Q5^>tfV>B#di&gy|0!@lwzktWqWny$J>w&o(q6?sD;{CHgPWk(nl
zJo&9$Oh?BN;`IT?*DI}9Z_=EXe$=*tqeVFJn0Z5LSXt<O%j#MR@T%pu)uHx&)`UE?
zctd}c!H)XBQ_0A=Yz!HDnRLOzc;t9R+B5321yi%6rN6~>+GyL&WnZrPWGsqXSXCvR
zLEGAM<Uh8mvtT<xd?70-j3qyE;^VBcUd{`BNSf6I8M7Lv4!%5tsrt3NuB#;{#oW|D
z8_WG1uFq{WeAihVlZE)+jpsH~O?$tEZAqRV`+G6^$h`*&cp&p8V?Xy3B2H)TRIfy0
zTsPSINMaO(?zegqW78g{^Ch%L!W#9%;lk6z*L-(tt>*GJBZP&#E@IL{H=6<U&}qu*
z%}ss&v)*0j^t7hN!OSVF+slh<yX@GaSvvp9>!#kz2SMl4zCQku$I80sYoT*mC7e^L
zwLf*9@4c_iPR@lcPaIaB&$C6m$=%ZV=mdw1`#|qx#cXfATR@zxx2;{DB+Fjq8RGL5
zTh1#uzxS+V*>VC~-|v9GJRiW*Xa7GQ7MYou-={14@>DK6D;pbmTmJ$~ZvNHlQBKY>
z-jeDampaL#_aVaLzlIV665)+-%dNFV-Z85(Dl+cmm=t!RD4@|-wyS%8oz?pNjp(jN
zX7ks#6^nXl9dKQl&Qy%4Yh8fVZ}x`xU$Q2?s42d-CHiogLB-w9Fl}2y?qV(U+YDo|
z81%F6ZdA94_<45eY~xL=)Kj)O@p9f_?lYM)X*arqjTnz>h76r^mbJRrDNdN<Uzx7b
zF}C)Wr|`Gi1H$Z*=l=AaB-|{zN^ELPz{+#WLl%Z~fgea(t_E<fm7m*OU0^({oO;KH
zfBYMKl`s5U=x}pd`*_kUnK<TsQw{tq$8Rg=0n%8?>hp<f{+QQ^Q`pzv-j~U1{<w>A
z^lH=;<Y6_OSNqMftGlZflkRJQOP;F7GN(F7o!7hgS+^WIjrLz2_B>XdA0v1<eO{jS
zx+Db;zJzfA_R{~ml7VX&Q-5W0B{}?Y_IA=Dh&&o$wnV|BUH_HyWJWU2hW2|8Ei#lq
z_M@$=i^5i#+wo#r@X}jV>#oDjB0q+=3@DYIj2AFF&b57=ii|~WVGwD?vF^KQ9(sJX
z?h7BoR9~A}-{p>OAZPOv7ZGfv-Fft}5X0D)-iOuBNNKpcEqO{@0~3y-K3COITQ%ao
z`=rPD8e9aAiuYM=MEGd~GmPWC;BDg=IxDwPuOMtK_wfCmq%cu)gzd1VEho%L+j*GX
zvYpGHljB2K#MZyKSRA%G-uP2T3)NG7LN}F2eo9~ech$JzQrv>#L{?rOSIBd2NgN;t
zdf&wsPb|h{%H8{#Rpzb;4KjNLe<n}Orj}NMVa{lTl8~uCwy!@JVpl2|X>25XBMRZH
z2+bvAKvlIU85ZHK9&w&33uoX~9-Xt*7r#wiy*|XuKh^c{-dyUXaaqRH{cgIs)Sg~9
zEpu{TzRdk<GxhKGob2cjaerJ|BjZ<$c*`@KFL-PbW~aq`S_AmuSxZ*iUY}2SMr{*y
z?>h1GfQ!H;gh#I6)uDW?pe;f`mB;pFb#D2Nx>x(9##ixFsTTEgk+Q7#{fM;{^3WdF
z{nNwC>eWVl8;#c^v7VxsAMTm0l@+Ra17NZb%+>qyziT4z8*~{~y<{PdQ7`Icx3*k*
z?d7I^%4Nwe-5av%X7LivAIR__qjyc3FQg{xl|)bpN<v!unjZ37d!pd#-&T43f7mb|
zV<uSupl1zGwZ+r6;so}FE+f;WR7ms8Hcy!e-!@)F`{@$m4V1(+6XW%i#PsySO-gw-
zd{Ed9ocHYP{r|{%>%S)7zi)gSr6Qmrod%+GcZqLG8U#j{NHfOh9P&m~Qjw5W=@<>8
zQMv|<bd<yx8zl{s_|EI`y{^x7-}m{;{(v3ldAwfFWbjw3=dU<v+H7x3sFnAte@E_)
zWeX<Wzbpv+^5e$r)9E)N{;sdGx0V1<)r%IpfEno7pCjj-RK~2mpO4cze;Dh@^0fV;
z5Y`kF6pXL^??X-cWiuw7Yj3}RfFe?kpUmHcHCI%YEWY;d74d^UC#hG`>RIfi_>l_8
zr{_7srEbOHD!wvIkb~68wCoT(YadfUudg6l@8{c{c&9pQGu=)4l}IcDw5`jH!_r9S
z`E+--uH!^a<QIUm!wZm=gS`|4oEyGBWLp5q4N@mgohj>-nIXz=EWJIR#hK%SAR&+p
zS=cLp7^7!6|Luud*|ef9ezTMOY+m1or23WmHrG%vAMWJOZ{lPs{Tz26gwOwFJ=>;W
z7@6zLyX@O;omXUIvItD{bfNT0w;&sHoJ_XBS5KDja(Z)hYAzt}oj+H9-W^6KW~EQl
zn9BDgjD-21wRj62Br1rtGm$v$|Nrr{py*fZ>~ZTtz1|JQ$#*H@Id%gzf@fM5r;rwh
zxusE~@_ZL()3IzzQV0!?qcLAhJFjH>#UD5A^0ke8h{9L!zy;ZU;*G%CC-L*_(ei(g
zp)r+vbVipuVRvx{MS-cE2Jot^*CyZ@Wc7zPe(Bk{A#Y+}??r+ArgStE3UkeE^_aLv
zMfTSzQ{O?KD5~_PEb0w#r)Io$Zz^sPwkDY?3urFFTrpjoDUSmD$dMm$NKwZw^e}Bc
z!Hvo#b87Y!5Nf=>W%Vdl{Ho@;W7-5}aI2Dkx=^u@G+_zZE+iL9$7@AVBebM1FQrR%
z-)o~w`GvK++-LHy&!M#G1D3?=6_;7?)rmm@7zS=n3rm>Wpx=fECsT_?B`b5nh#w)|
z?U5ukmx1SeRhU@_KqYg*V`7|fnG8It+wZx>Wy4h!&ojTD1qnecO=iV3e{9uKOp)&7
z6LQ{aX}3!Ivy!H)ARInyUhig-Ul0(!aq2s}4AFUZhJ5|n@yGkh{~3piH+D_nR)(aD
zSt(`luuxjf+w19w^67WF!X_9_3c+~6ONgXg+u`SLA|9FdDIT^dah9+eocU6_sB({3
zJ3ayNqhzTy&&t0~X}-nUoK;xnPT+3#0LV&1r{&4WwANCy6m}ZzZ&XpF!i@yJBA*_w
zhnT;GYQLZabhU{}y&NE^gtAq#<z77mO0)+^CdQr<_e03nj8bXrJ-Qv}8P7`RK3;ye
zTA!7n1C)$TPN(`lXdDQoL-u&qh})*cWM^9Xgkd|XT9k<>b>aO<6wigqY_mi~g(Q5I
z50TN-PAFhn8@R8-2Is$}`xz#O`O1!#^2(w2$vA2<tF)a<^Zz`{a<-H{lr@RD>(^Qz
zz&+woIWJn@M>{5!Uu`&&6<XeMBoiwGd)k3=gH#ajH3q4|U~v@lJd{P@bBJ0+n)NzG
zS-O^ghRCe4TqZI@7f<2u{NdAi6+ZG1(!0_A`O|17#J{lNdO~(Zy*bM|=$f=>+Q5qL
z((DTj>%(}_4O9@wpYT=qP_az@_vkPg8`Bz8blss3m?3!9taNb}(4H-H+BrAj-KFYf
zWf_85ZW7j{lRKqxC7vRf=kyib(jFVj$Nu*$`u_b+IcXaG1zHX#mFW~k(aJqu2t62a
zU2F})^D3WdVMx9R)+d)|V)q(fo=J&axa@FTf4W2acHSi5dG1v=6q7DcIQ3Oe`A*(^
zwpEOc3T5MKewFJB)|U6fQ2HE`Jpt4kL<?)vluCu(wbj<QobpnP<AlB+&UGp~_TpS0
z+kX%9^|w*9QONgy3so~e<X<{t9Rpmm;RjH0+uyb=pbDI(+Hf5@vM<Rm@n52lhO^tL
zw2?i;UQL_;wEB=xU|gv4K&_8K-Wg5xySE|rhPhGH+p>_e@@v<`gy4W{zUPrlMJd{R
zfg-FNdp!&2x(dvFh!ZoNXSqL>o~sk8qr%lAV_`fMVL3q`w#yn@-g}F(JN@T0JAmB4
zKIP>Lvzs^Hq)ZdDPCA?2hDjlu4O;I5ZC1i{T{xMxKQ8k2f2lm$Tv?2Kdz(_CVld3y
z@<Svqu-Ru(v1zy=G&_3wWBYf2&c2nDNY)$xR6(8Ye6xpgBEs&YQfobgyZ7dX`Nu?{
z#8{zImVZM&wn@*9<=0CqMx46}SShYa?2D6?#iaNxc`^QjhRZJiU4Ogj#P-@#A1h$_
z-5{@&1F~;kgl#&wblEq_UL(%(jW~H~A$b8Z4FF5$<F*kOx7p&{;lky4x?h<}dsgYn
z!~w8Q|AB@wmCO*<rPf(3ERH4?IX!RVNZIaM3C&%c&KNRV`jVb0FtSF2>X{tr%a)~E
z?Je+q<wbw(nv34>c9qs^DeSsQ_1wNBOyeq;Hm;<}?$TxG>l>aNS#H`ae_b!n=hD7r
zsewcTPGUDnZT-igWqy?4zLS|bDC$rAjS0<#eoi5U7uAOuN+XKlIsWOq1R62)NDu;V
zIUTtE!I_Z3b;W(L(|uBKx#M;MW|Q3_{vbm~LL&5$l{+Hs4-_2;|8D&G532RG!wl%7
zh9|`jYL5o?ii7%fKa26Sk*)=u%HR=qV|N&Z5@!GP1PEy{TL6Zeeg8hw7gQA~@ULTJ
z22(Pl*U`%KYY(j}6@lY?=X%4gjQSwan>mxzidAoU-k%pZ0zq%!q5?W9?9|WU_CW}?
zB)fYRUW{3G|H8@+RHs#0w*bDK<gpXrUpO_F{88f$O5k9vOZ{O3cYS(d&Ax0bzM$#S
zjs^Tx2|c9E@w9S1^nnb84?qF~G6Humeqx%K-<tfa_Fh3SaEeNx8rdc)oP=$>I6d|~
z$`}muGIESWC|g@*ZRYI?PD4Bd#so{TbUye9>Tv21Pdl|*Pa)f-YAf`~zTpn;>U6!X
zFv2`;uCVSPuCJhNy@O8$UY(tB&_MR!@t4Fy$w(?<p6rP@H4n%MEXE?;M|87j?+}j2
zkner|>?o6Dt2dP|e+#(A4SF%Agg&A6QN{}kSMbA>x8=UP-}c+av`c)ekOid3SfVJn
z2b>~();E4hHeK4Xw#^YL64{a;Db@1ZtrV0sPn=t&D!sDrad=|I1j{C4KB6m-?CNsu
z+v6}yO}C-TnPk`&dqdQp8y<QR#2yy|Et$CT#9|JMUFjXXT>nhYGajI`$ST~59lU@-
z#FsW^mT|a>vy7lG`oqGwks?%KVbzO+tsp75(1raXnHRC|Jo97UKk$fyt}L>KEb8=(
zO_$4B74cM2CRj^+$6*fSz}1G6_h{b(wc~RAm-A&=i_`|M>!(UOtYfQrc~CC|g{?)+
zH6*)Z-xLeS%7i&;IPcP8eb?l4M?BFQ>xpM{BoW7=F0c`%3^$hiO{P>JYJ<sNn$r4n
z6c2VC91mE(b&CQ$K-F;baq7X0aPNYQx2=i(1(JJ)LV(>O*N63>if3?*l6pGw)1t{y
z=(yDHp@VL|-(QmZG?s##;T{bClrQQoep!qcpX1|~7b+(9w`l!oN_LBJwrREuK9IjZ
zs<Hj>!jU}#Cim+Ol5KFw8G9hg*cARtM#a@8IQ)<SjGtAg$c3#UZo+&;MvwTf_Hl=#
zCK_QulD_qEY(nTnCYnn5cjV+GhH)49zh0Yo5*qx~ce{S3@Iivi{fnim3)|48Rk&F`
z2d%8we3ihwfQs1~MS&VZ?JI>-Sb|?Ng(NkHcRa&&0jH<#(8Ui--WNf)!Iu8uylF;h
z#z+SRa~I27Lqk6*rG+mA*Gu7Rjt?$=w0hSkG38!|+o^%(&DkybEd^dv3H&1<P>G7Q
z_j6jV&TCQ58=ddHFT-^EtQNfTRZ3L&v=ThU5o<>6B%)?Xq0-n%6I*FDV{&F=&;>gB
z!Qc)mMB?_R7RH56UQ6b^k;9W=Es?gFo8l3nA!`&!ASb{WiN-3Six{y7a&s_S+Gw>K
zd0hm2X{c~jZ!-lufHRNEld4F?cv@NhIw6O@Ij=>UczuIh3~U3u&bR}(!QBuz4t2d~
zmUmG0)S>TF7Sqjak3h^3>vX+hQDlk8gUGBI=|LkNzgf8xAUCC3G2P?)Inz--!ADL&
z3b-cZ{Ao6Ga;Si}f+m^Vq^S!!P>n@++MD~ebgJyD(0V2Til^690Gx-iRiYRE=_VBA
zS6$=M1<MQ!DxC40!ROPiE5qxPVud$?xg1mY?GT?M0OiLu8z4=fG(bH><^F1ywZYP^
zST+AT^eC$idkM~+UKg#rCtNJ2bW(!*Cm0aRkhNPLUuSLmxv!IU!{O4pwvab*6InNx
zkMf0bvSGcF`~GE^9{m12D}rQ9<}SaS2{`G2)j#9R9Xe>G%3S-l6v>JBO5mX>9vP(3
znOB-w{+OjC(C=e|CR6xjyes0~CkmxO?(6+#2k4<(6P3m(N3(ZTN<vh4@$pUpsJ>eE
zPFY)blqdOB9B<X=KJ~oRN$T1kh1lLMm>X?B4$_pc{zgi05KE>_z;oCkf;olEr;<5k
z{Ztij(qwKyoUG4+f_oKX7q#h72aGyoL;n4vyU2M)uiDP0V;8nqY5WM?S%~oKL3=?L
z)SCW5W4H~_b2Mg;Y%uN}5%d$Y+D4YNzBYwiBZA@qP86DDXKjxXOJF;a!f|OzP8I|k
zCj#59?$U@fj?kn<Fyjb9?O&}4AD!q9hzI=za{j__u4nsLC_=XJCvAQ1pLv{-v$Jy+
zKdF51|9>#Z8%g}nqk_i6f4kcxxBIMrw$~l6-uB0jL$2)H7VL|_!_YkkY?qOj8-H2+
zpoMc;EBNJ4hYw7MU|E`2Cc9|1yVMCWShd~U)QVo{kN175ld3bS)7@n6=OMS<MX6s@
ztAm#Foo_oudO_**G$$0ibd(TQQ=Qk(EZ^?Pe0d<)kF}Vz*r<K|A+e{%=P6BdxF;mz
z!*0D@LlPCA8<U1XJlhgF@_VaYZu%}q0<+N|$WX|l`Vmf9NI(-VviDhqTxA3iU5<v!
zq`q9HP2H@lApAq<R^^d?0*De5EKK2-n)teS2`(lJVC?07?ReBE)5uM=*Qr7s_WLQ@
zQ4`egxsw)Bs26k!(DcuoW%z}KlufHQ#e`j{8Rm}Q-_3P!b0Gg{gOo^E;uX=~d!bo^
zC+)g=q#s?Tn9@ub_pfy5gtAIPwKyFd2vfhPcTH0;&HL6cI^a3;%2dC%Tl7S$uU9W(
zCu?qqVNyW?MX2IVT&=UBsseVWaJGuYI5$RpGtm(XD;P!6AKQqhqk{U5;-@`rYB}dW
z=j<aqzb;~G?$4;r!EN(sR47}AlLyhG#U`1wGGesz79%j2F6VkVvlD<F?)%KEQQ^CW
zD_kE&{%ev6^DVSAvRsxY-g%&H&?J{Hj3RVcwrc477<VSYr7-%-;DY^5CykHw#D;Nd
zTpoCP&;>CxbAbW)*O<1Du=%;~hbckfaMc8?XX{Z%NZpb`yMv!T=1EaNsEHnAu{3J0
zEK)&pU18(p4yZ*omKk=D6PxJVb56P0yGC!9CF8beNGL$4r1_8($@;suMl)${&(k4T
zzV(YI{zN1<uY`TGt+!R|T?+-Y4Q6<F0M?k(Lg|u9N89@0qPqaVI2f=$_()_-UQc#^
zoOX0Yv!v4Bq%Q#3N@7__O#VMx<7|*t(e|iNhoqcfLW%47=^hmEpJmQvl}`on*ze>p
zzp?!!3?Ex`4k2vs=v7R-%HK{5iok>?#uOEh1=(lMY*~FPzb)wT{f*_mYS2k5oZ^~*
z4-j$9D~D<47Y#R{a28MZ4-n~}J;t)mlvY8wt}U19kr&rf6e{@i=?!X|*uY;PyW703
zqXYD4%<^)U8J2J66TR8AK}vMCqfqqV%jM%FE9uym@+@zhtcB>3svf5{M#~$#toAT$
zGe)p!eg|n-#JW@a;Oi;U0oMT!fB*i{q{MyG{ejR+yVt?5RaF3X{oA)Rp>#9a7Xd+D
zEsnQ(+U94$>s9{H;o$gsj$(i&AV=WUAx#;}1-b7jZQO>k3T^)+ty8u7YAOjr@19^h
zF@+aC5i~zFeb!;_rx)r*$of6WY}=_HF_25a(AdB2vp|5RbtsYBBUwr_SFL@PO4&A_
znnv_$DjqSMihR0K;ZD`i4MJ?ow2Ko4YIc^5ACbSc%s!xcQTME6A!H$AQ9^c9a+PAb
zraif%d9r$9?2h9$PmywG@xT(xb3rE47Oe#Js6X`u6!LUbBC$xTpz=71q~f+5Do4A<
z=c%R`RJOl^P;UK{=821=y$ylvY^2{c9;e%M3E=N{+=7pI%FjIr10d%qJFjn0ycy1h
zBZI;TY`J;9G#vz4>QB-Xo$DDt+3|7Nt9-kn;|T>#ihJzq&9eICt=)!X1dtNe|8b`F
z!(U#l<}t1v&RWWuT<cbxU*~Xzy4T=JVYZ|0*qYO|M>l*oCoI2>c#2+)Fm%%N8x@Z5
z+*h(hq&1=Ln{l~OU29V4C<5<Md>RQd6uOUwua^(EYJ<x`P~wB~6(8GZ3Maiv_FeI9
zG2xL1zb$fu`&gv?d@9>P*kS$%g6XJs*l4gpNIfa@qXkPnM&y;|w$s~fzBIJ~qp`k=
z5yaZ?oKKa4IYff-d=p=L_AYlv=qE^yA<cPcS%P{|d_m=CS&^#!{9il}-y<9jZN?X7
zKD^#`juqDL;|Q?&+i>uIk}-v#wfCydBq9dSL&bipZ$wZK|3o-UP3~Y)zsP;&kn(rN
z+)qwZiP%3XIc}nm$n}Yp$=72J41v{faxSoxc_&a8z?3Ett8->9Q}YREEwEwUZmtj&
zz_`cwiLqNGJGzjfyp<6B<y`RNww}v9p-0(j(nCzK$Ct}|$0>OPs}u0c>Qh50cUGZ9
z>7y61J|{AxU)2@~m5PJp%mR~PxaV$$TVOd;ML<k&<iFLkC>-$m?;n1@?earD`$wt4
zTEKlKkta{Jg`gO=oh(_H2pM~|NLnS>1m#YxLZ;eJ|1Y$BX|m^sZBtV}Qkm~l0m?45
zVPlT&Y6UwLJ>pWieXLEFdoT~UxzFoK#Q|Q4=Ma~&k_-z7H_9P%_mQtNZ{y2r66#$#
zNqB^KXYwm#dkhD-+xW^aSw87u+wOX7D<SA*>!Ak6sf|)Oo*XyRD8TWZ;ou`b)`MI0
zM3)A#@g_5P0-_vfY|&Ep-mg)%`eVrY?YY8ZR}-rW(I(@kx!yCVBl|Ea)W4f{Uh;tQ
zg=d@mWx}-uJ_Cq(>hJ1d6sfY6p1nU0zQn_CmkqH52{9`<mL;ZmqCz;=9CeP!;1mXt
z+5%PJ_<?5%5~J=a73T{uH-SSInSxL|jf?sRQbzrwZ_n(LU-O)`*7bIY-?(2naD;89
zLD#SEQQgCCn$J*<GRRP-v&nB8^TMcZ`JrU0|5?qc2t&E+dx(9Q$Rxw0Bx7Zha&hMa
zH3<}@4Ym0akhKEk!!lX>a5JHAU3*dAIerI-=#K14jFLnZ^0#Oe4npNpy>QpF&lEdv
z(7$z$Z_d={y>-*XOawj4mYfX>JEU>f@b>l5YZOT<yYkrR#3TYpwOi?Hsig|LN@qel
zWh)PY&0LEj64G$z^01{~rVaWtNaoV*ob4~6*+)5h5^vUHYX+E~7;{ydmfg#%TXdZ-
zne$n7t~X;#Nrx<&5F?!mgTD6m7F|9C<w@)iRfBGi3a2?HSQAMQC0)mHUY!qi@Nw>6
zYOS!6qzQ2)70LV`yvkqvS?%5)oK|)%$+wF}x;r41jyG)Wfj!<pkbl?ak$%|1eUw9z
z1N8G18I4pSYv_j%!L0?Q{d3J~#%h<i)vPzL^q%@dg8e-OqX(O-HfaUWasfLYuEm^Q
zkV*JmZ^%5diX3=%vQH5mMa$T(@*YOc8hNX|Hr0?T*4q!JZ@p_FnO~exfUwiC)-TDf
zn3EvIeo?b;7)Od6%`iQ$m`EIsSX1S&17kuggHh79j5AV2QCz`xaxW9p)Y^#}=lnq>
zmL^6f5q^~k{EfPPY(H}!NQ$))HOrRN#dr2+{FkIs!`FiRyf>{Uazlt;&l^P&G`^W}
zwKx(h!wUOC0rHb<=iiuX&s`a5*FoR!KfU^Hd@E__VII(<oYQIs(8Q4cz!dyf)Nl(_
z3Vd;d1StbIedJQTiPJj;<qYQPk@20>r`Hr%+*7jz_}ef?lZ55Lut{c+rGWl-Lj#?~
z+ldts+*XjOcnapzf<>kZa3-gcpDKK*Y`+bV5X3{O_<^giDS0eS-5!l%_9u-`_ts<l
ztI8*?2)t6+xaSdgCQr~zx1;!_ie#DCw0AjMOGc>=oZsNOL^5sppgDOK#a(QXVy{Xr
zpQW-KWccY5AL|96)ve^sKj`^HVqVP-n-p>{+d*$#q0{ATfXRjGE3AQro3hnZney#f
zA4xlnpl<;+R3Fxsl?U853kvF=kpT|cJ^syLpoFuAP<9#l%Z&2e8c(D+I}U|7ZEG!x
zuG_CxzlHj5v=v8yIr*r{C#cpx1by-3&I~Yai=<$-u;j_4wi7uy;5a}Q^8)Znu1A$p
zAyhO%NV$dMJ-Ym5wW$u4moRQbIR}SHApHN{Rd4afI4d1|#-DUtg@rce)eLme=bm<-
z+8pM7vusYwcg|6i<34F`OVAr#*u%#|u7V>OY`}nI{XEcN8Nl;~!5y!PGr1FI0C4zT
z=g@`JAQGQSn)VxQvQ33p;@mTh`|1i0&*{8!QyZby6m(@Yx9o0pzM<v^_lPAIIGQvL
z4qg*jcDQjf(x?0b(l!b_FY!d+qfZ+jT=J1ME1UBRuNgn~9&^tM`^do`#=*aJ$`z{6
z>i48DV@!i7G5(vp29%m0G4tvgb@HxmnNO+IwU<U_(<PmH@Zt08?4ZQ?=il#xuh}`^
zDaKrCn*q|?YzE)5^B(qpQxf00B`3qYT_(L*EWloNnt-kWe6lg$NmCv(DX0{{!DZn$
zWlF!KIx^b);^Ov{Mw1MOGW%(X`*qf=c=INxa6rzK1&!GCQ1q)Xw+D=f$@!!B%7f0_
zYTt!*_AE#9pRSD(glnlRwqe2*dqx`7kEcG=iw@a=8V{po9pWl&Cg>4^2~6uDz~`&i
z#-i0++BGhFpUInFd;hao3C^imT0tB@RpxQ<E7_ChOE(a{ZkJ!r7ICg03IsLnOCyLR
zbR<+!!#`e9*=O1w+jLwQjqg3-ND1C|u82|Gvd3@QUIk|;M})8H{135bINtk$#4TvD
z=b3d7r|=|N+0`PA>Fq?JuJ&%wXvm6{r6u-`$zkr75VHQMv6xns(r<$|M$QI}5Fpyl
z8;=34aFD=St6x?!MVh@l5p=zNYK)lm$SM)W${1t^oKf~BlYm-1dzksW>XQRS_nlZL
zh7F2NLr?Rb?OXF81lZB<s!H1L!pGOEtd4kHw`5mq1gMYcHQk#=GQuErWaV<Q(oY;x
zO;X(eH)ZZ*-d!WY1N}TixixYDT=RyWAMZDCGL!uSbU?%un4x(6x2(;uhAn<AW*&+w
z&DqJ6+~TyqIJHZ%1kj{UZ$beXH2+8+;7vbsP@NrwEg)sU6$Ni7=h3zT^A4Zf8vIJL
z_5oG+2r*=KHDqQR6@$6Eci|HZy{b!@JK|Ucn}@$nNJ+7%U)^sG)iv*(@$>vh`PPUo
zsI(PYNrZR+q@jFRniu$5j%GQpM)nsonD=)NU*7J;-mUT#W?FOg9BGV_!?GF+jR|m>
zttot$yQq1Ha}BT$!go}mhH@=#B}pw6?Xu$>$P8ZF5Mx>@&g|NKSf&$uGt0b?-nbgL
z_;et#KYe}fYs82oOnY9|$%lbT;;1%Ugvt$GXoC^)sl9$QXv|Hh5MC?U?R<4%+|@dG
zoTT>4Aj?^eO6d}>AMw00Vmy4$QBN_3b3Z92e>eQ3#g!C@79a8-a^k%d;_rBNadCma
z*o}xFUqc+dW@_aQ!}8l|Ra8`vZmkpEARocsyL{A8YUjuI7`IFp%q8Kc7b~&oFUksX
zC{I0J<)hQOqi$=*Wv_dXdqR-Cksle4aIV!{qPzQkzrMY8Nof(XdwB6Wr`_B96P?Nm
zr^<6HWt<AF9a*utI8ta~a*nBGC0G`k^78gSdF0%HMF7=7U%(?a?kdv^<uBq0I^03B
zCD_YkC(w=pP7f##vg*7g6gw|+<v%>)Y8Ql+lVjx|o>3K8<KpSypzzsAK$U?$V^AKY
zwrCYuF~}sJEa5O}SN&NB{=o^Jx75nKXN`@`D@x_PTqS#Hc@L<0?eeDP$K8}SQ;gao
zjzA?AcPH)2i?TP<(FJ#_$gNCTJ|L)vM2~>cot$Iz-G*k|TAaihN(;hv?tAx0gnFwu
zd^kwxW8<L-qIUmdWm6t|Oa2Z?C)*gm>4*(4P3#UO{|4@1iO6)Ceb_C@7`C3|u2LxX
zCe~GNR$8DilfyXEp_UFiTf(?jA)RX2_@rV=4@qeEQ=(&x^dR498l}7=y&mLh6Z_Of
z4;X1>Pc?0V$hAxJXLdeN+mh%sT=XF`YU=z*Bfa)TC|fo@l2Qa{`pl!ef7Spn9{T+!
z;kk+Q$FE8_^LG;Q?k7yv!T5qbNcWVP?E;CKPtHAG^Ch(v7d<g9ltA76ZyuO|!BdB#
zpJ4$(7q;<iuhD-SbK2g=8IqqEe)tWqr5|h*D$Lr}Sq1(Fr`*}&+&0_L*<esvZd!_j
z0?0K1@M}qsJ+JB=yf0Te(7|KkCo8lOz*t3@RU**EM}6NrlC8k8)2{3|)%OUq2;j)O
zrdd~m%kW)v;#IenDl)I7)fMO6VvodrrwV#)2||@EktA0nWzH&nqbhU^IUFfO=y%h+
zh!j_$ncc8y6!M(;$*FQfq=x!Q^pU|Oxj6WGr2PD0!@!Bwkf*n)hPDdp33QXJXbqBP
z(5XEKOu>vKP$cW#IJE5^LQ~0dq$6J;n@{?j=d&(%wF9>9G|6sXpv_&=E(-tU$(3<-
zXkvM0pqUvjzgZKlNCt$y@wvyH&>ZRMP|p(LRf~u;%S+L{3zx}@bLRy(-R$9?;3b-u
z%eoR~AsF6QEWaKJSH0IRm$eUARxFC6lPR#_zK5(Uk9dqO<1G`cUg`c~DJhcWp8*CM
z*y-qq*UqF*w)}7-+vw&n=J5Q-&Zc@Ewh)?wnP%N7l4`X3iM6`y=#QFQN?-i=2WXeU
zdG^yI4|9*<+95mK3vln+fKOa(K6J9EJ|(BM-jWqIJ`4Ic=iZsj8)cZED%!nq6H{Vy
zCPKSI>%S1*wz0c*<q2-~b^TWYxj)Fy=4^~{uWUn`J7=emDv~B-jQ>X$^P*0af!plv
zN+B_=<D7>SZXAwZ&x9nvbai#{+O|X^49Q)(db;!SgS)6@rKN5Wsb9J4Wm{YMc|ZWq
z6iyHw1}y6jh=s-mI_Hz!Ww-nln@lcae(Da|02D^rTHT>=r(^oQUO4$_VrhQX(2$e_
zqjY<SZ%MqZji<bK?K|H)NxDtZJ15t>fsUQuf|JQpZ3G?rqBp;tcTzEGpmnG|>{qwe
zKeO}yMHPqiH3<MP!CwVM2stA$TBbp8CWG%I$qaY+b7LWACamC#isGU5!{LR+j}19h
z+A8r@9^fkIJ9=Af0U@}02(q`NHhkVIR0A%8MUD0mLk#tow*5X(`i?6T^{vxLM5X(x
z|7sbB^HUi_y8j;eog*_2I-OufH==`^5+fH)Ir-N*<I7(1?h=Qj0=WL!<ThjY>(!_%
zz*2W|$V4-P!N|ts!8)q0{2QybqjNYj6%6}k-b<dSezPj{X8C7oDY~%$f{RtR{e;b!
zR%RsMUKf~xKz|~yu`k&`M6pxz*M=sL5`sRbp{gk*t~!lYDSBBTK?AOOdL0sJMUbuq
z@g-hm?Biz^7d&(-vw@MX&y2wNPmF^)zmk$;(BT^8o0AkIHlxX6e`faxs^GX+v9V;*
z@n>OoWr=c`*2N=->0o_CTlvq=QS^KhY}M+2&kQ6z!uU35&4!3hdqe%vRTNS3MDf)w
zZ+>GqVRBN=)d$%MG<M+pZho;mZ9?Jge6MSt@*#MPWa|Lr!(~;TxI?wQZw;uj%K)Kp
zpqC>e@_Fa`1zO9#cz3Y6$&kMj%;W_lwV?A;hOwS%3Z+02m0)H^^_b*yvPd%9Hhm(<
zFOh1Bj*?(V)dp(AIrrLAk7x>?c#bR|vzWTC^36Qs*0?AQetR-?*HkHwZm8hxhtvZQ
zXPxzDJMzuPb`?97sn8~6A0;%N-!tImX;bb{UfzL&G>~V9;hVPdFy4daF~tkK`S%0h
zXMnD3Sbg$3vdThS$WEK$IAcPs@tp6%{yURoKyC1H4sU+SX0yZm<Wt)c_E;ql46d_1
zPFY=v{Nlz`Tq~DIrqX+3_*9Pad+J)Kf7TewNl99hLErsk@gZ?XM@02+P@?|;zH48&
zu|8VcL$9c#o(3S1Hvb^0y}4=-eMNL}fWZI-T4!19*unlm)FI(4%+|tx-ED*=h4b)r
z34Z)aQM@<A2*Jf%Xd}?&k;Ls~<&Lm*bh-bwW0Tty@I(Eoq@_8~bnPCp<7{VE;ic>V
z-*~NyO_jH}w7Wr{;)Ipds3QM5n5NBbphIcZ1|R!=hvs5GtwZOpMr-0nbbtrp_opHS
zp78E!`GEiR_mjk6q`$w}dwB%^_Ccafha7H;q#Tv=5BJVEYW`+4Iz6YR0`(^Q1iAsH
z0vdMyAS%SF+!Kl9p&xT#(%OtjRjtKXcs0HP4Y*OW_Mt&SN3@Y?<pe$X#ND@GJ^y2}
zJY;UQ7^phmR#b+aQJI@Dpq&F2Z1CETfm%@rBG9Pvgwa_b5Ns8BADHN~D0xmc^-2EP
z+ch76%x}<!{ZpznxFCi(NrbBwbmMVKWl!4~Gu)`*kNb@5M;k??(0g%8_UebF6F>`<
zcwako!s^@up>#*Yhl(=@chL-8*_$_Hzp5G>KKlmkl*Q<AawX}POQya6aR1^dZfmD<
z_0Je%`F-a+qhYzG?y#X!Z9%HYvoXjrL)_l<Qze%h{95^Ex;*Y{g%Bp3BhG5VOvTR2
z2r$^pLFf!KxgxZn6m0WNoJCvWXtd?M;vHc4h?!3%ON7n!&$_o5kB&8#K#{*P4<1jB
zQ38Kkf?h>4SRuZh-IyjE460UxtO$Wy5m%Ce9rc(!UkSOrO*(z?SCW_yxI_i@=S?RC
z1ull;9jZsI$M~4cu;ZV+3*KEc@nugGS6N0ib1$}>hYD%`BYOL0@E3Dnd`5Li`06{y
zr1V+n^7yA~mmLLfZ$n+aR;7%9=f(U5naK5YIyn5H|AV{A?|V=yBz=WTc@Kcu)!Zaf
z9aLL;Y`;KJ3V2`;&2f=W_7$FG0vs#NBkYt#GR8h(1HrlAH}gQP-Yh((FYNtN=0qyX
z6!X~3`hr&FjYh}3{UVI(slNmRG*^<f$nFG_cT1C{eENPz?+<vdiE`&?jd&_F&2svq
zA`wf^FLm-3_sA<%E=n0Ob3tx5eE|{xCJw=D_E*F4;nQI&IZd{40wS>~oYewkU3IsS
zWV0&0RgZ^{9FAVHcZ&AO$3f_3wAza12UR~%KXj?Tcd;btTvTa&Jk=ipBQ8foC^r_#
z=JE3jhjqQCn-$;bWMp>L39Vc(z3x&z!K|Nqls=6{5VJx%N92o;wS4D(yPJd5W<{$@
zfv&4_>-973;%Jh5clOluh3}FZ>|jwI71CY%u!1={X9}4*8X|lXJG$H(wEJ|FGxo)3
z_Gwi3*6&<rusM3@`@vS&>d!+ZM;1ccNes1+rUu7(X`RdX*T;SG>mL$BUlLIP?a2)o
zw=x;C|8)mLNq4Ys3n|#bkGS5nq?;>G_+95VRVpLxZ)J?x?fFyW+_tw-H*`ElkyX0-
z!hFO0k)B5hsKRnB(y_b*P{@~>qD##;=1^)rc6(D|DH6#iWpSf8S-%h}E9fo`7~%!W
zp=Wk-Uhb99mZLN|;15>z`-q4J6<DPs78&zrss9wZ|3U*hbrz+?SYh$=6TQXZ#{=%n
z^G;FxRFCS+iYblTW-+T&q{Dc-0QJIogdwAf8P>L-O!CePyxDXN(TLd6%wyEH@#t)P
zsi@SKpyr%!EPU>>I#=j7&L1%^LmZuP<!XL(vEEwnR~fCf6!=d@!&F!cSvB!|Jw|(?
zrsu*HQ{GTFqa=<N3ejsF9TF^+*q0rSxs=ZF#3{z)$d48Uymh7@22B@0w-&nVXDiTf
z0{ZO3kB}(I>VRK^k<0zYth;)<Pdc-s%c#r+djBHN*J>@d2+LXetxTmdpTfz#bF^iD
zZ?n2vNS+|D`@$WVws|`0@T9i^j(V$MrfH+r^YUJ*;gzvk_#=X~RmW9++bdoDj4-Sk
z{&jMq_A_lpX+-`ZNwg<vW$HYdv+)1SP`<p^`!8zJ+5`k`Li_H!|JiQTSH8gOPc*=A
zi?m_tgXE>78M*<<`5WQ+ptT*FAukS1l73$+G(o*{R+MY0VsCpfMGo+<z{w7u$WFTc
z)+K}ryK|V+9}#epQ9MdEXZ1ePWsKURvO@Gb;%JkfTDWTXPhkMEX84fmx3Nlo7Ug56
zT!Rd}C&MSumU%aE4>0CXL7g>_Oi>Y-O!%&rX(FQ8>u+q~q~g(dMAt)iGmrTs5XXIj
zF`Dk>w<>D!;Us|GNA3@nBt^VuPE`5aAjrjxl9ed;goR@$7Ariv?a7aTZCZj-1ZFl@
za9c2MA06<F-oqH8taEys>@N9cX@g1FWBn!fs>)IGCx4dzjn3i{I|p;eb#5!Wt;r$N
z)UrdL_D!W(rKg4+DwYJC)%3^Y9F;p4B!+D{x?Y~2{l5@u06Ig`-N7-2lrr~+Bn8(^
z7rEDVQ~}Y2gv`ycI!g(~DpJ3qHl`msnG@96xApUlvEq$=)@R?A$`rX4(&kd`^nEMt
zAYoyIqM_Y1&BXm>aI9%f|HGxtRSU)f#s+k4s9`RqX5@Z@KN;Ik#v@<Todm48C2H~{
z!Gju^G}F)+^#)^vP%8Z3X+Kb;$9RsKUp&rQyf2H6)3kMW^o*FPRL*J-**@C~GtQWj
z6wf~SD*mk^00+_-_xBshC6u&56}}1vc!iko2Cc6dMx3}>Dl$Hpm_ajQHNV)p9(q<L
zY##1ZCbnI=xvOJy!q4k1U-Fy~90PQ&>PHWeW*1@8eG)Z-2^=M13uXlTA8VY*$<LN3
zXuaEGX{h?pw&w@sGus%{%+6r%583I@4{*#TSL6~>G5vUxkI|pb-8O|&&GSuC>Ioe3
z^ku=@H5JzTWs&sWrjP~)y|?#D7E~D1U;xd<o!aXfZH_`|p17Htv{a$Zh6-wXK$tnO
z%!kFQq{&c&YrPft@W}GR5R7ZAfJW-&?XDyzCTZ^qoYF3gq|Ahi%^(w_f-vQ|aZycK
z9go<MlP&zPs2Q`%38wg+xb_JtC$~h2_3{Ivrc(P14B;bGF_;vyiE@eBtdB-GGckXP
z;fRUnBJEOs61{^pC9%ZXgv{~9kDK4c0<6J&Y>H?20M|#2Qa$RtDo!C2vkKwsx0z1X
z18s)QU0-Nk6namM23?cQi~m7APsA-1nYC{fltlPB>FSa)#scL2Y}HQ0v52$&ANHZY
zjMYy5UmwiXS?-mVa~4#6Rj1$?t?a`#lKNhMqDVh+cLqzb|0Acp)nF$vo?PJAYo|8(
zmA1gYOlzBrUL<u)SK9z6v7ndcET<%BoC-XTmQH%LPLUIxpDi)h8?Pnr>z7~V6YM@e
zAj_300T!uAU0&o7Y&7>W0gN|rc#^U)uCEB|SqTIm1W=odb*|sssV8q3j~t=82B?~?
z52OUT82|yZzE3$)X6V_J!E(5({YZ*3wTjk@>efAG=uG`@iYJ(ZHpU6YTEZRZ=lzjR
zz((VvqUr~po=5k+a+Ju01rXRF$yv}(&{DMwl%q*W8rxo9@+O`RVLr#^1tt1E-LcGk
z3J`^pUvs$m<Kv?U3!NqOTyD7Ulq()>8BtM^Iho(!ivQIYD$%RAP)A;Fpn>S}Y+}z0
z-e3}WW(nT#T@RlsB5d1)*-;t(ewI<+q>Ni@{`R;L+svZBP405o1jUs)7rl6PpN2`<
z51{VhRc#WYy|w<_ZVY`HiJSod?@W;pgEyfWY3uMN!zjLOsI}!99*JxC;=8=pxHdXC
z=mvAkoc@$^xziL>W9l{92;uyY2)yZ6eWoEf^GQ0ZuT60D!*Z1z+RG85#~R>?+}N5!
z|Ja=e%1RYdaVaQ{g#G9SJg@$#?v*IT>U<%r{>3>t<+Cg|#8Y}nfxfq|nnsn<C(PV;
z#HBXM=luck`Bqi>pR>TAoj;-4NN|S6SZa7eHU!2B`+|AW$SQ66%7HoKYF!5w{$}Q%
z7XASL(Si5sR$6((S(2vYZwG2wm48F56rS}C2lj|(c9{Kce{Gs$a54iG8ZjAl%5)Yg
zqBsRSaRv96F!hs8uP|{**3}^g*AOd;e`r_aww~|aTU)CCT-+yu86qzaqwquh=^}N_
zjE$~WiUQD|T!+`3qv+9`c<i?MM~%*!EE`KoSmwHaU7xeU16jK54~?QbkZ!nmB34o`
z@msjjqoiV;=`ex0R_eG!C>nHN#ClO8Y*i_(AZkm%rI*?uE>WaPdS6mGz@n>h3<)L1
z<Jp!@)iY$E4*w6BJf1BfB9hd8HUcR$Ya^JrdRJkpC+`ARvYdhT_kxdq@Ze*lg=5vS
zqB-C3FnRM`rOgD!_sHkf3rUI2BFYp6UJ3jXT(dk_`A`t8mK&WDo!cW@Y^Q<d;PvhO
zdO(;`be%&Tw|160JOo0w9s3G)S~u!L%Jv2KaqH*0>#=4C-IleGCNp5vFEVhU)maok
zu_-5ycEei@6%pQ^))ViTdXK@E`P#LYqBrv3*943Z_#NFjED@|?`*8?FvF_VI(LKgh
z9`XjJ22k>|o}5T~)!0b!lAdZXY#zIHfAGgERfO&KN9Scao5@V$7i`XTYD&a$zq6`l
zAObG;3vzji+K@hq=+8$mOLx{SCPk)<ZOz+23sQwHbu6R<!{UAxPEJb5dP?GE0rcFF
z3V#*uebs`$FiSpj@dM>n%BvKQ;O>NP37D5=CYj)kXFEZFawrfB2jDyJ7~D_+xo)MG
z>2!`zGNfJ&?FWogeZR94Ik%<|DF^V!y~_qR3b}8|lvN_E7C5~BQH)eNa(+rd+_1a|
zAl2ce^8&iDWu?%xtJz%ILe~z2XRA~tPe{D3nR4m-2iJT7ESN9K!}1qZ{w|Ljt1%MY
z0Z9XtY?tCnWv~m^=Q}axO!;-&yDOWI{0<PC6CZIBJqN{Hul(vLHGzU7mfK>yoy-8W
zuvhD{b~Zn&Nu8bg3b1NwA5Nx&QRZ|Api~Kip?uhst5~#-Jfl(q4&`Luu*$@#7}O6{
zD4Ft#(1OKn)?2ENvqIFH_gjHwhtGBrHiWB&{6-fVf)ccro1~mS@h=QFu%Z)Vn@l73
zZqx+nYn3fx9><G{?hsk?jrtnSciut=&s-Z7KG7wGaJIoL_QR_2I(pZt2+DnQgJJ&G
z;gl*J__HwXw&P8RC*@0EO{PySL4UW^-Lc4J+Lht(V`6h=?K+#?<D5()=#{-++7La?
zC^H4J7>MYU!PQk}K-`W`O0UdMh8NA`a9J?)ml)0rQmK1G?`OU@^R7kNsq3ihx&LR<
zI}9LBH@2tzb?*62=t>tL5C{ngiMzQPTw%V!3!64fLgl|wqvbSTLa!shIyKXv2muxI
ztE$(v56SJp)X*je>lI|fq}}JPjIIlSar<5Pn}QgM2xpt(Zy}8Rsk<LE^dL`JQy-lv
zpY(lIEUHc7d5RK52-*mK>~W}Zp-rpI;(T1yS&>>U0LcIWz{R!djvV9$Rh^)qP8CKM
z#y*v<?s;0f(fbi|t-oJAB`>dfbC27_&8Ds_*{7Z<4yybZXSI!TemY8^t20n$4B7Gn
zFuJ>SelNOwHL~)B;``)K+c-*{W$NofL`oSSvLDmiM%+^D2-Rp_Yta6hkcmc2_-k{8
z5AGS#$<I!AlD`0u9WmsQrTPG^6b;NRWvNQzMz@W2T|&H)LyjoQo^6w<lj@>p{Jnp0
zy<KZQ%L`kN^iPM(QQo!4C9TL<SwOf175BJMrxw>|>c;i`ls{EZRwttQXVfoVHI`X2
zG-|V&8vS~^*XHF3nWXb*b@PK@ynkd!_)l%v#7EH^H6Qi_dGwm$<G);{<_0})`YBrE
zUdEf6(3#RcMU~=&M#5#RZ8x+>q>&*Lg!vLlPgwWl#%_hrT&F4a^U1Gx*<EHUGRl#7
zr;9Ul)PX4n{N+WSUMH%RGE}_r=$$OXsCIjy(-HWwz9LyWR%|Le`ccwaWo=0Ysp}fn
z^xD+%8*L7RF-m;S!ci&XnH|UctIG^=cUTAfMIIN^bhRuUOjuxJAKKZN)hPGaNu*KV
zS~Cabme!o?i_U!TGUvsKsn_NrSG)U6zi)lrPlbD<Lj8WYyL37~TbJ{qXx&1~el1Pp
z9M}1-e|BE<(z5=)cE-bimDevFNqZ)e+W~lVa-!kU<gBpDV$VZ`TF4&GZz*?@F8QMr
zRv(CP`U)9PB_d9iGFd;LxD2wGpBNTE<^f_Vs^~uGrNU|JH*_wfO$P1xJOmZ0NgP(<
zRT~!YPyEF_B-II;1WGW73o()Z+s{=6mV5!^HjRcCiZ#tHV%+=sMeQYvj#bJw{j&If
zGb&i*il+pCW#T#LO`~h?Kbn41l>pMOi{Y7#YnX;h&GR(<KCZN7GL5zZx(67FXkMs1
zHGMf#xlL2X2lsOFVoMs5yYUtTTuzNV_+xZ1s@~W{rW<kmZyhOrxkt3|DM=PkZOj8M
z)FT4&5dbjFW-oaXMP1j3Xwex_iyEDa>UdyqS?G%qzTzjAsp_Q=nBxYWBPkp3@?oad
zO(CB6&<^3;#IJuklt^I&MFDM_)_ag=6-;d#YWVV6)brJS@&9cZbtP@(hXEEfFSUY~
zscq9Qf5#~2mlzinl8S~DFIJd3G?O3cJ|{(wss@wzw4a*hb)rDydU+Q4?3UNJ7<D@7
z?0!6tnKJjpb`jo6l6F^9pa3e!pwZ3!y1S%7jP@hyrv|Y{9~ImJ3`GalfHxN~%Ace>
zidQ_L`se$(3D3*^NYjPN&Ebk-JOTod%kTDzI2=~1*<Vo3`nm7gWz0-5Znp(G1z@16
z8sII#jh&IBRn^aEw3Sk$iJwzX{m^Lw+*aCKYuk+u{Le}ro|-#>e_t>x${ydn&0>{&
zW9+MU*>IRapR;2RMKjbgmFidTU~IkhPVZaE{({-0DxXPx0r{reH6AZN_ZfL=*)j~p
zGc3u|ow{38zZ4l{i`OIeK5(d|fZtS=yrgY&Se1!z)p@5oSa`=D%n?rzi+z9jdH=x(
z$-AHS*NQ$lH<y{w6T{A|`DeEyg1F_ZZ0XSdA4V!(mBdJ0WFw2i3C6k4A;A-w4quM?
zHI&2z!ByM~B+VIK5E7&vOmaVmcKB<FeSs!^5cyYKwFa4b)tw27@8m=Hl{fuxlHej!
zn?D9HPU~1X&8)TCk1$MITtwLmKH?%VR#=4giu1ytqVQ6eMVv``-fopqL^rLr7Q13~
zp^CUr#`2H!&HFhFffDtSxbYe9Wlyi?ikoj9(3xsUA01O=Io7{&q^j;i#=gH-bm`DH
zz&p<bTS%p&Ey#Z4OF7$o{Dy7~HG5tNwA2sM!T6yf?du;&r9F=VTiKb>+@va=pzQm_
zr7b+|;vcxCS5`T7UB)q*a;zja9K>}b7%P$L2psbHyubc@@CGW#E6$3sK5D3X|FCKA
zgL2?20XOMq7Zw}~)sA?@siq>~!3b)SNwX5y;WKyMsJl9+IFkx!a#U)!#5j$Dl;2;T
z)f^nZ<IPFO%=HB6xLSwnZ`F<};fHK5eMlJL|2+SN7zUL6RWRX;za=~;H52Iz2MuVW
z>t}2-{~FEngAUdUaxV{yTuE34B)>Hzp8Ns)`ZaMQyX#{wx`aV~I(?U+_i__vb+@|L
zdQ@+20e!lC$Q+O-YQtqmCguF{lJ6yZZp2QdgT3tT6f+D4fu8NZQ}D^|6v<TeW+^Xk
z;wuw4i~I*sMtw?o<GKr1<nlpK<_&;4Xo>pcXtdF-08tdu|3)wEex%M$o0}&DQ}|7i
z;P)&bVvQfgWY8}c2M!(C_DF;Yj3ea-sjM8G7)9Im<1wym@VALGd<Iu<Botpyn|)t;
zs$yzOj41%Sdu2)a6wO~*j>gtV6@@iMxCNisXe@d86?M3|Tu*;8v;qIzTj6heS6Rmv
zt$D3kG`{uZvhdsS70kkabHmzfmnOEJU)U-;A!79}j5YlqgTLXpvVIJQKUfa?$rydQ
zx4L>(|Lnxw_7CQ8KtEh8y#L96E8zL~Fm)sie6wW*xcdHIYkRzmPs>iTdoam7n5z_k
zafNRGez5e(=Fc1H{7X@(yUb?Wgjs+T;MQ9TH)tZcM;6sj6;ihEi;PJzq4ELnGbr0#
zW$k$pmxO}!b9im023)*PptMs;v<qI5v%LJc5B0f!;m6`-)hn|2PU+4v%wvaV82eBS
z6$!6*<vlu~PqC5wRkTPW$WG2YchkS6$>-oGbKe^_bRs@YS<@Lz*|PNhkUVBHEv(nV
zc!nwMO<XHqnJAq0lTpA4pN9N*>0dX!EkDfN1)jk&K0IBsmlMh&>aYPM;Wyt-Xc$d(
zx<5SiRoKU9ZbOL^tQwiqUqadyIGXz+;~{i9p6}{e7c5PEsJ6qNuDVQsR#lq5evK$h
z6|`jhLQSHgxzu#VtIw>r`!c21BAAr+KP2Ae{tsj-apNyy+St6^F!#fWl!zkkjMAZM
zUE{$|6lHy!Mp&fy&!058j?QqUT*^sE{WZS9)y*pK*Ej2*R0!Ud$Su3i^sF=9i7EXE
zhb^0UP^5QrsERFQO9jK8^q#*Meh%P>Mar(k)7<<o2iE0jLNVRg!^g@@x?agvGLb4?
z@hXh`^3yL1#nX9u(>LRguj=Z%{p2OyY~p@5KMqjW8@-><^Eo69=ZJl)Oj#4@=;tBk
zN!HCy7VXdwX=C?A%$-_T-?$#AdUUa~(sxltvg4=XZS8raF=9<+K9+)fMCThxA!kFj
zIpL>WT)iE1KiJ>$oS@U%CBjzdbCC=2_5YUw-Q^o`9;d`4q<L7_$eElPaFQ<^Moe}^
zuRXz!_qFo2okk14O@I0LQ@T^(oqnEIVDRgDsTeggY_NHg$!lYD&0BX?c`47OsT{8z
zj%%|Nsb>=R%!{i999gL1s^61gz_n8+sy67?4iaLom>adm7o>1PXDSnd3)*P$(QtDT
z&`aWw3Iw6|3PQ*}KGAQ=r2kC?U}U(cB}IWLD_)%Al=Wfj!&(aq3wr^it+Vi?{iLl&
z^*=Itm%Gg!EW39X66;nI<O``tZs?MN0v>DFnUibTyVZ$u^IB5%q+ReV!E!>pIRpLu
zpiz$s+rc!F{rNbz(FEh(n}2xt(|Dp5r=y?vSxR(U)+sc2(9{?m#p_?7_$X1IZXjaL
zzoCHSxbLOS^!}NZFfG&9M&Ioy>s+MA_5cp;c9O8tt_4>CA?0Ik0F4KUJtF)fJ6sc6
zsL%1=s2Ttbg6}3M^F80GEiAa!G5QalX$<i-L+T4U0(%ngoF$$^R)6d5M^T;GBKG}Q
zcXl2m%O5{e%Nn&M1ZMiNLU}f8&h3m^x?m>(i+=W$Zd16*j!R<IS58h&Jrd8p2l?09
zbi2DvY036~jdwu|)J<RG-Mqa;wp(`F@y^PB%rT@&?*HQOX%&8V{#OixA}J(8zvHw=
z_<p#-ct>JGx9Uz<2uF5K%jyK%66Y+aEkXJGqhVKN#N=}?ZHJi!HlZ4}<Ot3k@4m&u
zb+#krhkiGfq@%m7)0}Bgk;N@h8-7ZX|FW|1?)j4{1+B1YSG>6(oH@bA80SEH8>sst
z4Pf++;`-$fRbc6=txtX0l{acs@&sVMZ#UqMTXKr0u{+nXn_$4_k5le(U`1Q)!&sN*
zv!=0sJfyw}?Cknmn`!`V9_~B-8<^4_8`D%=BurC2cWdWt|BzQvI-ITB`?J}rc1hi#
z6U1;3Y$%|emkSs9UJiD_m0OtzaeFR(a@<G9o@|*pfwpST8M9vTe!p6xY>IQN7UhV7
zU@W#Jtb6j;YAVCXW^A+dl+LRrBMyGl9SF7LV~0|E3TgY;VLHKcbBn3H@nzu(|NE=_
z*O`PVHm$HoQ93HoFCd*lAySN$sJ}G@Ug_`)U9?HJ{L&f9+mMfVGEPR(?@7zKU_z5n
zYY4ovBpYT(L@_Z*qC9M<IK-6||39|QIxMQb@Afkcp`sw7ptPbWEz&&(EurWzz|hj&
z4FjksNT)PPNq5K4NHars&d?wXAT^xL{a)vJ?(02=|M-uKz4tHHcdgI*^PQj7c2v&r
z&f6O)EuX|wuc1l`{h2_NPaG%oV3|>KPVNBONtZCnh-OXYlMKQ2H~4t3GiuSI<e8T*
zf_3aHa~`DpsQ%im?^nx~?VK{Q7Ub6qdYRN5hc$Zod80Z+Yo4((ce<n9s~l785%4Na
zO|?2AOhLP%B8$HE^apQtmv6kxd*CSB=fYjd9z;+JoV+zk(E*Jl57>~geO;1RB(bg&
z<@Z!g;kyzl<4pAaSRY0Ad|aRAypX%UPIq*6B|EKKKq@k8bQ(tAP+d4s4fLFDlRI|A
z+eL&uYp(Lz8cj<1R485dEWlPjPAI_bO((*1C6B%uTR-zf2NMfI9#h^_m%C+2{g((*
z2Q5LUTJxD4u<>;$FpHY9(r-n~!okkYvy-;<-x75)Vv3{;9S?6nzAm-IG>HnESrd5J
zaLW=*EHgh+2m<$U!u(-#fMWPWV{Wxd%m=H_WBFEPZbU%n{A5uH(*n35Pheb1`!18B
zGqv+4@wPhv&SG)yBOvhp=SQdr_)Ul%W)HLqm}!-nC?jhCmx~k)BAAv}<#N;Z=oFu&
zvdNdZU^>|^pv?v_7`XstNz8gK3f*cV^^m2jV|j%tR`Hm+7sLP!I-#(og}vyvV`b`|
z^9v2F*N@}gMaN<&mrvpRHL9CX^URXOLi~K_SsW#ZHD)BVK88dEbcM^)W{vU7$+*s#
zw?6UH`=+tb`4C3CL`Sf8RaF*=<T>~jk^`B^JBN8@VpWJcLD=pZzDC`<vjOL4^Ij{^
z$WEE!K$eL*tJo@fGVzN$E87;LUiZsMjwBz@*geqOCA|w?@fdrb_<7n(!BuPTcT7l<
zP{ALvZ>sc6L$|z5^I>HAbQq*91VG$jNe3StCHv?N&d??nEZc<X1Z`b^&6S2sZA4Hy
z_l<;l^t8%8>GzXdC%-+BL^8gYOhSVsT`XprYRqFr&Ahktz-&j=5z%q~Udr}X7M+vm
zH2{;k95?EidVV_^0kB^{O0sjibw0(N@26mxs?Auh%aQ=Yz6tItBY4?nb4`izyMR>x
zgqlg-G!5(n*Kjq;m9ik%Cne^zm8YjM*!-#+_%kZ1w4BN=&mc`F)=zxPv_h!ZN2iy>
z{#oC>r?;jzD9Fp~?;v`;A-O~v*mR$ylnvd1M{DJfq>B4}kqF;#W%mxBIt0ah_OpWy
z5$umX;{-=(;MifQd_3;R4`3>wxuwHzd_{yPu-_f5K0<QtvATRj=1N$lK)!L1ckxR_
zIrtuUV)ts+tVbvYJ4(67rm&)ub5cT?(q~xVH57zGzBK@{wh-p_s&*5tq^_r@CkC`v
zpaHnd=HKRjQ|~Z9Fw88f_GCTJ$S6?n;$&<g%i<?l-KsZmm;929uP&3q@`aga@X)aX
zoq(oBC0qB)o{$RUTE@iO4!%nxqEAhJI}S`vZhIkGxFn%q`!d+Bxo7TqzDLABx7;j}
znjA2tiw#0H7ZzRmf_gC8J|3?Tb4jy!y&YxB?<JngP1UX}-DG#YN^@&Syt+~`TAX>D
z2eiqGbw4+~rlosB3UgUG8B*7;o~_;szi#(?<hs)Rr_brjsit#I487@`q?iMqBjr4$
zDC{Cg*vw3~TC#VXYJD2m$LK|fo(v1q#;^$plfCj=!{gK%H9X;z;HT<Z3F@Be{nE^w
zP5ih4#p~mVe-dw)kCKQkJgJ!HqesVtF;M(9VTi%S3sP^eS+DfxSe@6}K$I#iC?wX5
zw2Iwl_R<c_o286eIBdKdHGM(hRX47GmZ1x8T~~B3o79F3fB7vE-YGj}igTQgCJ4oD
z)%a=%a}^Im(!NiPFwH-Gtg1d`)m;kIa9_!YZ45};<LlTq63h<MG%-)Kx6eP?w8&S%
z8sAW@pa(r;{&tWtnpN&GC@~QZ5J|q<tZs*$X?oWTl(8lY#RrChCA~Q=5<O3dJ}OX_
zmgpO8wFf;#2=DXt@UK`;4bj-Wndzk+b-;%6Y0#jsleuR}U~`rfPsu(kP58#W5tk*a
z(7HJ1q#8OFk-45*R1b|xur<4>H5+1Uas%b3AyH=snUzFH*BGgo6bW&I1=_i7UG(ry
zn(r$Dqg#*m?5&&If5LpdYIj@6F|`u2dgtpJ7^%}u>Vqf$>mo6k0V~o?rVv|A!#C`B
z^8%+~o06NSz)deQRGRR&+^{_e>1^wJI)z#O9YyatEu0Ic2|>I|s6Rjt99t~twFl+W
zmiB7{WTpy%{5REBB}`z4Wh(omy&vB*ub3Iy^FrNY8cT&r{*st(?sb=ba#rb=B7A9O
z9KXG-Q_vT|dn7>aYz_@tpVsFZ*JS+_{7J;FxZd{(mm$|Qt<%x0jrq^lEI})joP6yy
z9scsO>tr-JQHRBc4m$U$4k4yX4ssz1!n_wcS~X|%xwV?g@q`3{Lg!Nf$+3;n5y@Zt
zIKiI<f++tJB5GL_{1x+x1DNvMX(S<~*n-Fv&u*5kH%#wj@V<TTNq&@r$NBxy7<<xt
z<zH748;QDXy1iGW%>5D@7Z!?dMW&2U=)bGuSm%v(?7mj~y(uqse{Yw4X|&`{7vvPq
z0_Nu$%e80H9uf3a955VUAW}Fyvr^o$9I@J`ONd8vlT~K0^J_m~c30{4Ht5yND;YLw
zCnK7O5TRwX9<)`Ob=0QQebcw1Y~B$dhEB1a;W%OX2I=qh?xf4j!iF}+4~;UY<(b}1
z$^|U4<NTH2fX-LGyaPsAg#aFh=<WLB16G`PLPrNf`0Vkd;!w3;Wycl#gW@-0E4j3y
zLTn3r2~Ci%9ab5#Qv2N*mwuIuUqxXOJhiTZ0s_%0qASr(_#wGK=kgHWe(8%d5kjx2
z6t%tFzYsP5vjzSl5D5R+fye5xf(xG?$_@tW;(ypF)t>G3VANAjhVj4ZMW*Zn84dkx
zo_&gum)idA3k}$(S)*B^SxmbdZ$#C^!`oK1W<{>VHu{q057+CA5~wK&3D9@3Fguc@
zF@Y!o^jr=7N92S0_aTbEU*@wD-OBHNerLDSBh@oS>=!wCYI;)Ol3*c@c&{GtR!sVM
z$<gUn-JcoOs9ErHCLcZgWK1QGn!I-QmXxLZrAkqL59}brT1*S_aV81#dc`7c3iMEc
zQl&&HAr-R`w=sIy!AH&n(?oqq+eB<tf1OD<B_THaS}qaSu0_5%7}oE{wT?=@eiKa9
zOnII?v)@@2nD;;=J*jE>tjB4KTkk7jB8jPC-xHH?u*53uZH$1cJ?F?T8Xme?916*Y
zv-+i<Sjucud;hk+#gc0xyZr;v>4%V8%GW(1j@FK!y80I^2mMwnCsCf{vJqM-A@vgc
z2>f9G1#^Pw0U$<HJUkK(CwcN*Pw6BOT`y53e-V(gPEoli0lIH?RgvA&#?%wOxyCsk
z5p3w20#Vm?JUL6nr>)KP>&EgNV@tbS1w77_OfoNya;*)m4ib+M2j(M`lj;E+4Iiuy
z7VHD{A`2VMl#}M7vus8Zr1nDOrU??$BSaPt#~rM$Zzui8$%(x8fo_HU-m~oW<K>2>
z8m$TwWwDj0)Ym)SJD!k^t44OS?zqf?%1Mq(LfYDWN_Y9*l7A(KlrJe4ed~2|=QbWo
z^mFkxKvkO)P`uNLrI0@Ut%|-l8@R|At*mtF^_)vhIP!iD?*p-ORafT%mVaiLIE|Ci
z30b5Lq-x8Gyz1q=)dQs@zjXLn30hN7t>(8MS6rbMy;_?q>V|xH7=C&inM&_Sp!`03
zXZP1<@qGW-!yBZ_5H(P;b=@Q+;h38C_uDsI?9K}bq@~5<Z*5Tun(M5i(V$J#1e1D4
zwk1${K=U@LGu!z5<r=n;faK)L8@5dBKXg(mr2bnnFp5DB(C|{~v(en?BJEKA-HEv0
z9z&#4xU(G$e}f!ywmda#=B%z8*F$i&>p{w0lE?4CK6L-OA9|QmRc+3~s#a4h*IQ5`
zx?}&wck2o=enZgrg{fA}-dVZJNLZfV%s1?SryL!>&0D76?Q&f7)gfChqAGJJACLr+
zPOG-I@!i?AgD3eJo=}DJ+jM|m3os+!hX)(S2<+|Nid>4G<|R6HT;oqc=<RPAD3gXy
zlsLe|d)f$H8Vmw<(U~z`AlP?sHnM2kd>-ionI$d^Q>JuCL5MaGP)`q@!VqL5B+HJB
zY<XN=xyrXJ2JMCkzDx(9c}}V>FDiO+XJB$aPaU#Hew=3aPSBN)I!3RX{`MjrCL&;Q
zVfBs@w6o?|u|eH{%Bel^KF7Xm8VKouE`1+UPdyD-KID&=WNrhy_sN#60?uoR-W-?g
zO3_n&IdIqlxIuq*T84kA|IGh?UYj6Gi>=B_btWtI;taP2xX*8Bj5O)G8BDqqJ<gmU
zYZMBZGSNkG^Q5x9H*F;ei=kg7W&uB<2}WQm*Q`PWP>elnY32ven9ua|V@fPRpYA=W
z*#FpHZsH~~GmK$QE7bZJdc#tnIG9P?`|Zh=b)MmU4j-B^@m-21x3BBY&fifI{gcln
z%dN(3WjXeK)Fi(-_QsXhb4<|ikz1gLY|$h<UyRbPrTD$bUoW5F`LvD}D-8Oc$qRjv
zwpP?L_OYGr=I`9zpPi*WEt89IkNKd?*u7yT@eL8p^%&_FNY>)cbxQ|Jb*`iebswpj
zR{uVpgpOBraCU-2v*?<}SgN{l={;}fCdVb+bHfr!8}mdT&?KBuYkXtCF2C2g%ud4<
zW;=i92}eUoOzflN#?fzdEM`@wT@C$XX$#u|x4xa{Ic={6@6EMzRP}BAUwWjFn4f*x
zwAnl4bdxM=dc5g=A;uwN<cHC2!ks66<0)i+d+U#M4UT_Xz^_Y41HI6s_?oW)rh5id
z+D*L8f9WiEx9PumEiZVlO&n}%i3{A`5#6EL5!~t9@H|1oj+2R30o1!VVsZt%${BW$
zB@#@TFn(u1T1i)J*6}1hG|Or9O%b3q0AGGo3O-|adYXa7CQS11i!#z>(mdPjf8iw&
zLng+7Bh)G)_q|UrsiEy-C7n!(v%7$<d1<uq2C@Z;*6LZA7an7Q=)W_Euf0Ro+Qfo0
z6J=PB-VSyf-fUa5Ad2U3YmMwL3`dY|m0p=IZkcp9H08Qx50v<lD~DfAGh}?ECw|M{
zFiiuVzT(*|`aZnTLC{)dEPqR(Onj2atR#AmeWo_4(B*l>u5pC1j5VKlW(7M-irBRX
z+lAh}UYk^!oISELxSghcuv~V^`ePb8&ZowIDH{EZp1dP8!T)MiYloLTEMp_99j5(8
zG{-ejGiXufQ-0TU>Dhhex-hsGSbg^~G{={F{dWoLL6yFS@F;J#u9K|TK?|>3Lf_gC
zjnMwcZNJ>pNyY8|+Do&XcT2>x8V-|({vM|C?oVLewdy!B<mdLkK*Wr|y?)<{qpClW
zaC6Db(btM09>f;*KA>SatRER*w}h|A^QyXw;OOMH$jEk>28xb?j{<O2o;no@Urnbe
zK`GWprw0C3#LW)ye3=~?)a#Ud{sv|sCc8C!MzLmQp{P+v)oG7_A9Kj*=1%~v%-KH$
zPhOqqusOZb%-cJ4^zs=~=~@ZESG2R|slr<GU258=reD->A)28u6f({AYa?jx;9eKc
zFlw%_`a#2nvRu~Xd|+*)$jA8-6o9_W5Fs5+Q`}7^oa_E0QJO_xUWMM3{=LN;%GIJG
zwwB?v5~Gm?G%`Nru3q>oU{v%tyUOQ;-(05kLC=p0f7_*&pdJCEd!$UFmE#p%D{LIH
z@qllH^Y1dXd|6Jpe5F=WA_k;Q(>~3_E!)g9Z_dK&Rb$y5mG^!t;R#RD*=sxZYlf7d
zYdO4OUK|DYN;O}Qv{VO9E;qjzUzGBFCF)}OA;R#UZS#6<ZJo+UOXCUl?!$fM8s~%v
zG{HCK697d!N~t7{ABxrb?dP2~XC`o~D<sL)y|%+^Ccf8vf5^*~xq53N#4^PF2DXbT
z`Y@@R;Yp||tKO2w^~ElzM{N8a1ZmVc=hYi@>($<wD;9SHwa4YovjQPh)5eod`SH~4
zc|>M_BbRYfy9^JB8d2OdIVsM7k8Y~rjuq~K^-pT?b0TY~cyXI^%AI?hFlXKC@wgNF
zD46R?F;`5}QQTR557f8Q@O@<ED#cDHU2VW$7VUE*5GQO<`h+`}Bu~duv;M2P=15?S
zISDqV?^e8Lj#Rm(!ks26wLHYZ3i$A}ndplVx935^>=V7Y-*S@pJlBOJFP(Gz)aX{h
z`_nx>S5Kn?FYdP~wSG<qp8cY!CjhbWf|mA7s9}3vBYUH;j;8^i$*T-pE#}ZbH+7Qk
z*&=Ork(aU4ANGG8JonzR^LFIMym&LjR<7Ud@6^lJnPGdmivT!QgL}fBhntrkB-pm1
zrxaJilH`uUz=Ufm+zacBfLm*>zF%0+V^pb;*L*5g`Luj~Djo$k589S&*cfA0nQJ&a
z>!=E`bQ$Z#2hb)@{B^k)mcP0FsPfnMMX@fmfyc!}^7q(rNqmnjIdjFiD0as0mn&*$
zzGmE?^Hu1nLMd+R=uPnVC1aSjyV9t3i7D)HD=%)8&c9kPloUylav6MNz7pBoF+)Gb
zq?(w(<U78Jl%9MN`olDY^Fi0LJBF;Gb-HRM0t3W`Mxtqmgj^0|RPv>^)a9&6Dl~0N
zJd!7LHGh<%hnZc7QpTr~9LH~NFQnMqR7%8CUf~V#7{|AUHgPEhI{Nw}gq|%N>U4tN
zDIIf0t=zlY+48ypCHQnew_%nq;XY;fmmcBxY;JwvW$5k|yxH@2C4aFY!y1(=aB#+l
zMIT@5F6!wtLE6Ks(i6x}2BhOyV1V<#_Ls0_wMhS1^rI#qS3N|Tb<>R9CX4wRe0rU;
zaLvFvZtAMbM4y`i6PdM!ZQ<D>BKOFDyRm%fV?W6LBr21x+C@i-61!$OPvJJ)IP&sk
zNlggz{n*mo9_OQ%SCyV0PdH`rA9)K_Zv2!IhkQK#5%K}rG$HD{^dcYsc~5YE;j{Dw
zpZCtHaE9~CfHh-YPtB8w0iW2y8f?NBUJszFFuT9MzdHLr3q+3DKVr$S*vab8R_ls}
z21%*Zqk7z@;K)A4&wE7SBMcg&v9Yloj5735orRT=`@}*NI}a(dFjARwx4nqSlA_r3
zWe7yq<Sk2d!GSl{ylIo78-=g22<KUvoNSX1d~ulRyhUXeLECeG@Mowyi9AR^3VfA?
z%n1|-XZ07I_NW>2qh$GbB0?!qO=}0J39*-BkTkxs20>>^3bIvD5S-0V63rP^xosBW
z=C4nKoR@HwHjn-p3L-5YavXt9x~Q0f(ksoCheAhX8}sY%iH=<NYt#bhv@5DINM?lA
za=2RBZhuhz{Xm_0JbtON=M?SC%sA6QJ}l5ATEX;8${AZ*&~@x*(y<Cy;z}%eEYXj=
zA$+m(%3?msbY+S`%9i@G<HP6Om3Degyg8YA(7P-S&sFs{OUZxC3?fCiftZxiV8`9N
zxbM~=1LNlQN7MC-4=8BtkB5a%6s5thJ4UMSI>IbZc5An11*%?Z=gxfb<1nHS>fg>;
zjAB<sC^$kH1a{^0kczu;p3yhTS1e@23KLkc6_%cuermBFGgxx76-}aXlVv^#tAaNG
z<bh+l$LgbsYOYn@)24Ci{#uiWOA;!!fsoB;Yw!439J6BMaV=*Nq_;mLXU#ottri}`
zNJR67ul+J^whX1CcV_)FZ?aqoiC%C?*(Ytl$32sD1@G-O^iO@1ZCt<7SaS2-<J}ui
zA+XGt4vD?%W^t#>zrLj;rJj7J=lZj5G};1Xq=gwDtCLB8a|^xgr0o+cbav9YHpNWP
z-wXn#BEBMkt(=+MQgnylTEE2jD=&|8685k0%B8nMz~cO%$XEH&k-*gmv`IGM3v9hP
zr(tt|(wEH)*ged*r)$;#;a_Fa`)G#V)vD&TEV(wFyTn_XpEN;*;+0)&<XaJ)X+Z*9
z2(DM$Pr0RbEDqjj54`d@9$;;DAa=UXa&5YJ-kCS}7VLF5g+{Fx!giR&blU>;Wkdh1
z@r2ROF9#)efpfzr51eYT9LUwV%FLwT87pr%H~!R-;1WafVl(;)ESbdN{dPQ}vpeYP
z>Afe5_3eZ{_l5kUXim0NlB&bxgRv!aWQB(LDBH-;GeODI^v=xoh8WHi&*xOm@3P!H
z7C)Gmkw}lzm^(;|4ussijbUb?-X*yOx=C}$RaSkoRIzkH#+a}2{`{L2ArYYyws~E`
z@?OpkihuO^G`aD%&rOo>9^NuL>2h@&ZOph`T+-f#v-yd!d#?sANWf{zre=p(nkVGD
z6|Sq<PUd#vMx|f#lYo^-KFCY>+wU?i?#}rH{d9Y+X>CpCEjL2^u>BiB4cVWB^GWn%
z;5TKLM};$zA=I2~jxBDuO-vsEy`FRp#0;j&n)(T!F7X>OIb^;onr!*C-Xt!xyxO&t
zR@DA}SjDh{E%c`l=T!XxgM24P$nen7m#XpETb=W?X%D%_i8H?AJghxb%AOsW_dmlv
zyh1ieI^KQkr@3z4+sv)6#)K_%i^S4Eg$2X1OAR;a2Yv`SExUuNf4?NS+#<zlEr0J3
z;N1{fr5*Tr{42H3+}oPq#>9Hz)i)j4gq!v%*kh)5-j_&N8uC4+1_=y2FVFJo=Ldy<
z_e?SG#D4S2;4|o>H;&k!B(t#ys8OkEvUO;(s~hnhsrz8ntGZ&DeH671N@_gN1<8l_
z>U?&1E{d+|<_n!9k@x8MOlRIeE>}6TZ=P*XvV@N*-ILG0aNmEkrk(98_0LFZngAGM
zB4c@U=<Gt<o&J^MTLsV$LpjdvcQi7_8n&I+{K<K~Fdo3$B!Sua>&wgp8YgvBBVavQ
zjvsYx<ZoC;LeRBYvG8#SP}0HW#Vz0`5d1v2DI)kyEM5NRHiyMZ<LBlJ6t6^MnbeBs
ztDaUt%eyL3Bnr!x>D(dh#R4>g^Pd6j&d?8L85r2A^=f96LK_dOV<PSCCP)0)gul@o
zEVBWVEI&?mDLx?*<lxM`*7nu;5#?LgZJqfMG-6wbu>dT(%*!l1V&j5_g95Ie0H>0e
z5BDi4;7=X-%2>!iGa50P&mrbnSkGpjVAliZ(2O#9(+Q&2T>*DgGzQ`@@#*vr`m)N?
z-F4TQebt85btg#>YexEm9SkQ7>BV13MCO%=WP?tt4E06&HY|ec%V(;s!y^Qa-M;?m
z_ojh49qXn&XF2u~J<-zaFFWJw0B38p_urDsFEuM|{?xsuad7*H+$&s{V0q!{tmQc}
z)0%&1uDq7s*{fJ)iW9@W&?GGxClXxFocmTKYDqgabrSf4dBs-BUq1BUs5WK;{k;<r
z1q*4Zh<2Gadg1Zs%~Uq_mrG@S%2l;DK9iJQ@v4&-6;Xw%z9&Pa>m7Zo0j82%%Bhd;
z`jWt$s%o^-)5N)+N!otsccaMorSE)@bnF!7L~zg9*?%1$IP!;)<~x0q*ZRg}#h>UH
z*{G~lJx=<w;^@OhUtr}+JOu9CPC!R;M+L9XpWwASMQ$<-|E9s~0*))|h|Nll*d&<~
z+Q!{aPJ_?MKE+H^1D8AC1<}S}QZ6pgNv(6SS<cQ`3w%b9VI*XpsE3N2(4*lj$Kn9}
z{OOj4Ai0-0!ql%lBqTU_&Ao4ORsc{&Kt#yvcJ@q9!F=#dq!dva<Y9hmB`}XhJ@-I?
z^}&DOQ+8l=rqv~NIxwc0l*cZ=!~*mr<fKgC`^XU8yu8$+u;1N$R9hzz3_kZn&K)|r
zjKvmTi8MJy2p^QJwSB9^PNk<ifLL-3uJqitaoeL6w}_i|Nqdpn!))dlCVepU`3q|d
z2I)98{*X1*F>D{f?d{JxSNr#+x#rPG&{Rfc;#nF#L9UU%%K0gR-7*h)PMqSet_?6j
zV>YgdtUulY=eEf}binU%nHsjScKX@&ADsD<<|L5e?)qonGCgCL@zA>)EOwDG6iQ5y
zDebp7YbQe}dj&|*-(Tv_9I5C<H=ef3*bijkbB(as4XL90Do?xa{<UH7KFO0^rC53B
zE<x9Myfe+q^?L4}jC+M_9p}1jzjlJk>7d}Hu?R?mac`C4+B&L!{P5Hho^>-4!w<eR
z(ipiQ>|9V-u2TB!2AHo8|N4gj<1ze`DW(H^1pn7cfW1ug<{0Om{q2;d5YEgitmXJ8
zHJY8|pjd61;)a$(-@7>dH7sCYn9wk=kM7lRRl>N;3ii!<s90E-f1TgbzxRYBBi;!`
zJ|+wp?}=<tjB2JXQhM6m_Hx9B436ByFF<$QyP`IjQ6ygs9?plALlZ=lDppjHsn6>;
zLzScU+y_+~Ch3p*yH|<{kv@Tc7^ys8-^;IqJ{<hQ`H1$tn%b@_%oKX>2xT1bvPNnH
zoqt#flI8jn{p&4=nw)j42B&O=6@u-gjNjV!H?;KgGuVpld82CF-O7AZ9(5<@3TlnA
zt~D1DCweeOsBN6&W`>!fC3X_YvH+mbE2%`ZBUW0n;o^@R&`WXu+@k)820v&bE8{{a
zYFR2`Mq%qMKo#-o4K*(^wP6aiR_qod>?+CXV9N`%Fnp=wl2XE%+!Mc@%fpLuR#Uod
zatFMy{21=4>09Ceh%fcH*A#trypO4BI^Vi9tO&iO`LC7HTIHV)3-B?;rtBx-GXAN(
z0?s3DZ4qS~>o`^5&pJ)sc^z!qAn5NgwO}`gla}^A57xUrKFlwW#8&>!Z$PuW#D{1O
zE>(ihMKEy+L+wJxHH+7C|593JUVBG1&L$Z*W=G%KX?;8{n#N(s`n&1*H+c2$VB;}0
z71lV8sqT>&+1l~nlqd)5<>LsttBFFfCPv-A>9<aB>)%bAtlmMlYxYx9W?hA$+jBDi
z<G*t9AaGXo`A!Km&hRJygY+r}x~J%@zyGu*zf0DA&c7_W@6~2cn3{e!8;OV7ZW(UW
z>}h5*A5VA5HL8fL3^>y7T)$c>1cT`?kF6czL-u7{i!P-<gFD^;lSXD!MBFO%vWR1&
zpi1R37D*odUoh+>QHZB1Y^gfOtjxl~LUCiXAh><3<r6suzlLEA9jvc$#Ar0`#gudL
z4wI6RzO{5zpWJv8rl{)O9i?t+sYI+q)_FE!zl0J*ezQ<2JQ+{DYf)I%wTDa5>aUa9
z8-E}CXVO29@UVHrQXlKZHF@6tU6GY`k(Ks<UOUZpb9Ts#JZcU&8Z)b!Ggxbht4`dP
zG~ZsLn<JbPhgyFvtnVtMG(wkaTL1c)cg?>kbCIXJy}ZiPiC*G~W0AM%tnO^b{E*px
zQkjvd$EvP+%boa^Pa@2oQmCD4i0pEnxaILw=tX3JeuYe~k<F;#o_!dWsNY!gTk9vu
zk&df{yqk<yl$2vr5Cb)Cb^U#AGX1Uwjkfkul;Oc;)lYoM&>9riI}3{t?Mq$*P%W{i
zck8mnD#>%a3j60WsrYwoImbG%57+-z#mt!7UqEXeoK<VNF2Dqh<pCuP=j@p7z9ZdI
z#gp7K&QW|uqJvIR_i;@s{jqUg9p=WdSXYz2wO%cco;H6ykI<Wi2LpXt^l>*RK4-4o
zsIe;HFrq~q^dFDxhG~m#z^skZ*9RJ0H0}sToL25^8~<AODID#Xbjt~2zHimF$dISG
zBy&<r@N4l@`d&D^^Q_Y~ng1Z{YQG#$$WBPB=h?Pq>c%BeHRCe7{$~g2W)C2$iqJQN
zr^8crKR#|dDs6T7&l)zg`kre!s1iO0lpkU42Qoz3G#rdNp}=Grt^m_wZ9nzG%W6+w
ztfHzaTTc39Xzgr&K((*0FQhJLn{sjqpDh?FRPgf1%U#SV3_=e)$?``$d8{P4CC}>e
zIN7D1xx6M3u{u-J*7awGUd>a&4V<Ete+KDjY}_NWNp|rZDo6G$o_(Dl>i`LHDLh~2
zr>x;9Xee;dn2oM=u~825JWT@0t*h_Uljnn<hg#DZkd17N9m={-C)OlchHP|~D+G#h
zb(u3gbN50%da7&%)t5_>`a|3-mNrwEglMku9<Z_uB~-gu>si5CMmzw9KFU?C-gmsP
zb|kb2ilNZiGGO$4>8e-#wRL?y>2AP#>3r2Gr%E#XA`@Z$s}2}zW%r<CIl;i$^v2Y9
zjKCajCs$CCYRta#YBG4$rrVdY&-%VKpABM&Ge@u8Wmr$hc_Eh~ZA<7EKg&c5qH;V+
zh-FA+3o=wa%c(T8{|ldn5XRZAZmT+qX|YPXgYA7j4yv9^^j9JJ+D!!d=TjS>3As?F
zME6krEEUIP>R20fuDGLWt0w;p7Zy*Ch4|SEg$k(~C|_ypmOBM$96KD-#+7n3<w8nc
z7y2e4kD(fMzY~}{;8W}FDlcEYT+`X{(AdnX?Gu`63(dV05HW&dh^T+W;ma^^RY23f
z<_VY8n?JZ?KTQBr?LB-2(9@%wA|aX#pm^7M`3GLloz5Q7vA^{IMCs^Rz*hL_05Nnf
zcd-U7B>jN%GFYJTD~XugzQ2UQP|UF@&*VPR5B=+lioYSfEDhZ5L8|ch8yYvs9~!V%
z3Q5xk&y{iA;AT#aVc_;QveXjyd6^Kq4>AMhL`wU>914Xf%@yDJ6J`HvX>n-9wN+dT
z4Wt@C{L^jLuYTwvndRFe`^bG^o=RT+c7WY5P)S&XjKR-8O><+ML-(hXp^zA<PkEeR
z{CFVpCh8%9wVb59ukW~^J9T3}L_Booiju&&2`g>!U<WHp#52u-i4xksL|>p>cCU5y
zZ*4p~ZVPSQ7$bZ5mhG<qkZXJt^GTCH=L||24pH**EzEWc**4-fV~Dx$k|DGOz^?O`
zpbZKk%oMQY$}&;%qrYB79v|-*@PMWv&%m$5NjMWL5(TKT`Qq{zj!uWAQTK{P8VW=Y
zDBj09cF5}3^cb5b?o?sLz!5-}^;PQd)~L|_^)E)vV@W*%W4%<~R-3~if@zEiBLNl6
zo1~k5X{d#+J`K1aPyfoL`SP1<!<$esfUr4%H17PfwKn0HC70@%70auaGJAQs=I<9_
z-D)_REm;zJQU%a7@g2gAe6vR@Wbyxf@<abTc$lrNt=KspW8^9sTuufUL(QB0Nz4z{
zN6gU~3MJ5=VZLWeYrCzKVM{-J8`?stU)l)N3i~Db1rfakfiCb0ub#it=-uLRJT4N;
z4LzS|gG?8%lBvTti?Luqc(VP%QstGW<8NY=mOpx3zwRfzXJq}pRb-HCxO(+tyZOY5
z_pw=zpve8q%px-dYSZ#pKB&)(`oM~uHNj~8hDJjL^_!;|#Yp~)?`9xDAil$=tSSP7
zt-$!Gqkmp_njjw}dLTLS`?8kntK)Qcn}RWd(t0EE4d(xpso*uO4Gc(F)|37ykFc?%
zk!#rLlREt>ShBMC@!C7@0)g>Oi^H=RZ1HUCg06lvj4+x07RTq5A{$!bxw2L#(MK{T
zoaJ$y-SpnEt)i<+)3Hb>RTPhJLYRG$v^Vqj0eMm8wSB!A8e`;#A<AR8bTP~^-@B^r
zDoy}PYHuF%`Czi!Fa2&X2R74kq=Q??pRaSdb-kbeX?7X6PQ9Tuv<q;{l72M0-^F2L
zKL?dyc8P1Wa8_;PB}qG7Djne~<T~qBvZ+IDhKQ=NO#B6)8VHNzDsYHw>7dcES>vWy
z2`6KD>!$qaW<hplj4*G|-*lIeS=**_pvUh|XgN?xCVM5|vS$kc`<G%Yd+FBOjv$<j
zHGMwN2Fz6nd`7ZRkyyttw>DOEb$!kmCY2D4?MtNP3uHPO3>4btIMJ|BLQZ=!yE=Ky
zs;Js})xAtpykS%?0)-3`rGaii$?ZB`q<S*%&P$9*rWPEZTA#lJzknOs2QF)M6+A<#
zHTFwVT_KvNCNIr5eD18Sst3Ie)9L<j3^AnINLvBoBP9a1UY;P*!NRR1zhV4XU!CV{
zvRIa2Ca%%<RKVOs3Dvw-5JQeFMaXItmW8~;gTF*l{?vmHl%PE>dUzB$<BBWkl2!hm
zv|_$#T1V)8vfBm$;g{sYbj!tQtW$5ND|ywF7OTN($Gf<ftm!bRZ!RIxe;zwW3sLrL
zv22Idln`91?-X<BkvSvZKdU5u_KEHeJQXv#OUWWY*aKg&6xm4O6K!xOq3EWD6Tb*)
zsI?<A{zMCUdhzL#X5%bBqGqJ@u#A5{ve7T;Vq64Eln2CUTez-hQWL4?x0nA?`6=zk
zcVH+fhtQ#2IX2fV)rW-manFssk+v5<c)j{=)KULlLC;uwAtuY?vGKeLZ;3ACKxlcx
za7Dop7$AV0UEWg!w@$0NxE~@``Oixp!~r%t(GE2$sey18&&qGtkMs`?B+RMrnjzrE
zvy)@Yj>&ge;YVN0sWArP{R#oXh(y4LLfan3dzt%b5!1S&1+%}D;Qp~ZIrIM4@<f+>
znVpcMyL<Nv?WIMx&_3&}wzjrMw|YvX4={=7u*qBO`N}^wG&Cgt2495k1MKN#vR9X9
z566YG18vVgf^H#^pQP>6wMV6$>SPAcUYmDX-rn#kuEwe0s8_3Pto1TTcu-YawUb0F
zh>9%XEm3zbC@BjVg#0U(n6PK>Sp7U`O0c4W9hcY=LzMw;HYu-~rXZM9RCVwn`6@C+
zk>p2E7P7PX+P{fKvk5-<mBgQfqgO>zE}i&lZ&Tu~oTnH@w2ARA2=skxmeNHnN3v6!
zvAqPHB=#N#lWsf5j#4+~gKHWT40t<W9!ajT%7yY2org{Gtao1{zha~StQM_@5=_v{
zdLSQ^L4c&A-YjJhMnXGB`rGtxUHLHffv1=y#6yAdi=rtt=`nYou}bf`vXGG7L#BK7
z&BLn47Y|^X(Wx88ULdYmZ4fP|yxKLd;6^k2FR_Hj+@g?uf9D5Z4al_YtB8Nr1_ok!
zD@TiOLS$>!ID#wE_S!p_V#NdpN>(5-b$?uzrXZU^&W|Y=o>~=hVCvt}9+ZMV?M&hD
z(=RlmfM8B=Wu!wdI6T<02nU@O9SuS#Hk+MydcSMP)Bs@oz0U8LU-uuNi}L48u<nKZ
z9R{-ft=S?TQ4XgwmWZnDVX4^6jTXj{@vgj}W({i!Fv0H|+Rxv;3k?7iz*}vr>Kd`%
zxz><M?7g(kzejq`Q#X=}f~nWNXEB(lM!*rDb!oX=TaHpyDKi~V0nFR3i=F@4U4Hzt
zy?AojFA*P`@V&7FLcdD?MALv6MI0<29`R2DBaoY_v#(f9`I1SQ$MQ5lq8K!Y+N}i<
zvqY0r$yBH!U)($wlf4uCeWUgi9g<fFV40Rqb6*Sn6xD9x5v;OeuG82S`^66(|HL`^
z`B4JN1@H|P*Ou6kU3K~1l^pthu=YwNWyX~bTRvWBeYGL@_vg<T3gG`L=66dU6^T$C
z{~XMqB&HUS`CZ8WKH_;ZgvS@ky!+syU{VXyot_SyB7iD@?&O$Y6YFGRHX;RS(fwGb
z_5oPd%8~w0GLJ(lNA)qI$)7f#!R17Ln{Te2mB5JN;1?0|f16J5+~)1Jo{Ohu0xLS|
z>Nvc`5-8TVOk4}{+C%1uQK!bwRt4fIl$lOjE~6I!<NCGSp84mJ$PX$NU#mQp7=f9c
zD5io{(T>5>++l%6;*r||>8vNSDtTOksAoi#)kZ}6`wl5W#y5{_;sEon8m;qDC!B-K
zPLLA*nVs6sfhnHf2b>DZMvyrV4RAG~xM0!HB+<SHR1(BF(YyujL?T3rJq~ovcZ?`<
zh?*s6J8X1$E=CC!6i8#HtZQPS%I>+cv)#gRUms1hOw*=_+VKlDbQuo=zKWvKa%K_Y
zxy{vUh927>KF9m|S}INNZTS5@rLa0)K@Ky;yb0Y$lilud=4<k^bH>P>%<kU+s;pA@
zmX@WRb5a$Yb95XOW?94fIPF}Pk3-Bwq_qJ+rq##$@z0DE#TGtNqc0Vh;%}aln3PyR
zIpZac|L<j%gRbFnmFfigiLVQ1US5Gi_+toj7pK0Lkejui%;#qC-vA7pS>ZByN3X#w
zvS&83?Vs{2kR^(rfFurfpC`)MW6^?_4VKL73JOHSD2YHzMDp*!IQ7lNaiG%#zKW#W
zGaO<~lUfDoI|1z3oJZauS!09YDOVl3suO&dk)y||OJFu*mXyfeo5%VejE5~>pA?x#
zp#^6ez<1izAiULhWoyf;c`W*$4PSmlxPtdbs9Wr%^Te!4{oq7kOVroHk1}tinm)_C
z--e)MeT#$CDSCKWZnmjqc|k?ab68RGd_x(u)lB)*jd_!qj(bI;*|EV6jd|M>oJFTs
z`_>lau4g58x{`X;`B>|LA}1KR-)xtUUOPiN*Qsn?|2sxFo*k~Qjc7IzFbT}M9BX}n
zkk%rlp*Olp*Uvf8yFq=nn@@dkl>!%im{yK!T?P%tXPT)I?pZ0=>vtj8oU$9YM{!X~
z4yiR7l=lL=U%aj_F9FNW*h&p2J}dTGDl|XaOv|0?BS!{1C;V}?HI9HaE85-2dCNL#
zk1aTvIJk`@jzkIAq)!%xX*z)4XXwErL<dIhy<I0Vu)Gq3S<da`8vz9)1w!_%kZu)q
zTk;_isgz|xjEO3dA$sQn;fxj5tGj^fK~VC~SJiJi@7u%c)^qa~$?noP78JP7ey{#t
zqs>*O9xO%va;PXaeoEMZg`B^+v8Z}6$Z5uTjO3f=hV~rYsN>PApPIWwZz?JRje-v%
za92SN*WbZywKvCNdEO`Gc5e+DCcUT4L2)QaaP)YItNLHt0~4dJ4z{oozZJkqL=!WB
zWl$ZL#nnHcl;T{ahMGgtw+Cez!Q_B%_I6oNu*uO=PGOjY5qiBDp)2*a!}{2#&@2I>
z%l_IO<gdHewD;kDfV|xW!=zgCl{GOTuHG1W#&FAI54&}%x%<{5vvgag2aw_}ty4U$
z8f>dWw30~oxY~hUCvNK5%~3T}@Wnf-+es0an|7Vn6yroQ7K1c2GBqJ+y2r5yRHndu
zG@oYb6tNSLvKA@D+GBJ0ioN#mV}yyLZ6-TN-RRq>?nr8($X4ZWU4AKY@~z~0J3l5M
z<rzTM5FZnC6;Nvnx)#oF3Sx6q(!~D0@U}9jUG40lceT|eW2kuP(Nbp8Rw8OdDCs(x
zUx>@F<k6zW31=oB#)r=yex0#R9mA%JZrx3LM$0KcqGHcih)N<nD_X~8!5RhoqYkHM
z94mmZXVvPQwxK_ft?;znOR-%P#CC4bV4~ExWGZ);gqu6&;NK3Xe`YPx8UMInfSkUf
z@SK+;l=7C`_Pqf&+3HflmaVT+nMt1_F4fm3p$gOgHRSpj?=I1o+!?NBMwkjjl+w}-
zC|IQV0>T$lP<sXi(*{2Y&o-W}c2d&wh2A{+VojO_HUO+lH;^o9^&-ZGnkCh5UfyhR
zy-AV;VncGW1b<S|XzXyv@tmo(EBNrewV+TimFxix9#7&%q%%zZbb-+E0KnD`RsoK=
z{Lpd(^ZqFMY?oYG00;N+lzBrq{A<e%ZX6{Ps7yK~exC}En4wqxw(8uxGNhuW#=zd0
zcAK;28cmBD`(mkiqF1?T`?_@29PeMl=n7HBL%qk@3f7nwZ-=SE!(zp^(lEQp>-f90
z+d+l)k3W>e1}mv+B|d<Cw>P)!UW(s<*AW$liWO2;y)BmYZkDt*VJCZNe_`2k;7s{6
zTNVEX-fVoV2UDHoZ4gR${GQjpbc7-I^!@3I0Rn$os!fJgkVDX=9=!r8v=8{iv)0Nu
zg3~-^$40|hGruiSHhma2R-VcgWZeHcv*egzwXW|NICNQ8ySTH>%Tb&CtO>XvGLwDL
z29iIQ_5#ePwk;#QIMOaV*+t8vnOt{J{$Ef90WDxGjHfg|KRGCQO08{{Q8{6fh_A;<
z@_GXwsp6$*E4er^pi1Vo+g9#z9{=TlB*spd)g~mr6?Mx#(`IdpGNXm)nj2Uh$wUv{
ze3s4%?O6e&wxrNXq$p7-1ZF5-G7CianP{~3s9xp~`NYW3y@jl@>&>=tQ!@|;BH3UX
zg?|2sDGplnH4l3X<OhoJ4;t?Gn;hY}y8H(mg24GamKJ9e3j5_^^kIwS-PKCo(P&Qz
z@98rrw|SBVZP-d%5}75h{U}nN#$ghK_X|cb$dihcFa-n*-^oE!Y=7(rf96{6*V~Q4
zH<GKMm;2T|c_mn_`F-*_s>DBhvvI}s_}<x@tgc7o4j=Javp=cJgVqFu&Dg)2-}rC0
z`MzF_HDN02PXRi1vs?=~?b|4Hgm4$Cq4QK#*yx0)NzJ++e{>CJmZ|?)FJ7-L(KE~F
zld~&D)Y@R`vSy}`DzWev|5erN!KHq8X$V>ils+>MBm7;$w0)(<TuOYGozk3@|5`@L
zFQ=Q*51ShY#CPh)rx$(%^8`6^eR|CayA4e4y{pA5Y=xyI&)yd)P3;`C6`PdmfA~#f
z$5q6o+1Kny6G5Gr(2^Ts3^-dxa#M<-W$^lRAV?4Ti4^kN0i-YTC#J6bGRNu?oNB@E
z6J>Hr+bW34P?Bl3FZxwMgkJ!RR`z$dA{Q558Q1HoZfE+$9pgW2fC*%{=D7BgRdnKj
z=1p7=c%4h%j~5!ck)f*A>n{wQ*AdvO7qcEc=e@8NYsWhOf4GRFZ%!|cnO@d}XqL?+
zS7s#D^|%6;*~<YC(qO1bmo;^(&hs<a20BAYE(P}8VOiwZCIiP7%^LWKz`xJMq3!SB
zuziwABb$+Dg+|=#oy2K+Ops$ba-E%fJFkwgi-y9QD;_wW^{p@?X&$@gD)l@#GA@5R
zlv#SRBUEXsB2-x;6+i3LU)6X=x;*JpRJ&{y064&n(J5!kk2^n8Oxe^~57+$bLEr&1
z-zxLyVtX(qPxl~qeA>Xw?1^%M06QWPReP11jW|iW&c#{`%eW)Y>ov21fkm->j!8^r
z(6oa;{d{UhDCd7DLYe@y>j`tUedc;l6N#0{YR_^rcBC(9hd0^}E%$fRS-*8u*P0f-
zEAH+PQ4X3(&$|{mKynMk8a&j`wrc%)GOJ9GCOx1grC(b>TYilG`fy<+OM)2Tr~am1
z<|SNFdy&FJBZQ{={3pyoL$_Lt@uEWsKqhd<dU|eYx2Joua7q_~dnJ3EG@fE74!8gK
z7(U0=GLz@tjT+(WT5;K-EWYmK%0>E+6E2=ibh`H@RO0zrrV!unPQX~YNCu(F{gcw|
zUW|B8CVk6JQ=@V`dX~RC-!;8rsn|os-1_U?C}@*wn)1_O_Qwhuh1(&F)|&gerL~WJ
z5c3?3PjCxz*GKd79k%?f0w}llLt}WQrMaGrg)ySbf~Z^HFs%=G6rO(GKzD|DD*+0*
z&mH<hx3OS)j|t<rObCSt3lcFgnEsy!V%Bpx>}&XO2}xY58Su8@vdLCSPRLSDj4f<D
zS^ukRiNO7JTE;cT+svAD7QZNSThS6LD-fiq9xL2Zu2WqPf`4$qPp9Xx(tJWqe(O+4
zDihN(`eSBbA*{Gv6k7G0e9lAex5ZKbpZ%<V0?8Fi!MDkiD|>zbMNg=Zo5aL*#EfM;
zR%>Zy<N%-6z-tDKP}b9>jw&U3Qm}RY1rbCASJfOgDV?$>XYS$6A1ZBtiq5Y3KwgD{
z9}E=XX7+h2jaEF^O_tOoE>duGix7PTr)E+2m?!d6Ua~jX6K29?+#Zi#b+mL%tu4Mr
z{rD}H(oJ%!fawMeqW%@jZ3t$WTUDAZC+{fE^*X1(@i_itdy&V)d!gx<E3rNAY>VfP
zIW=6#dW}Lll#|;Fj8L|0aR<{jxauj0HJ7*Mf=3nj17h*HkT-TFM_SxLZ0ntC;h*&m
z4;6&$p304$I<H1<U)Xlu9f|12NADf_MChfT`QW_v6W79DQYhqU6sZI<`?!Hb^=*Df
zJ}NO72n5o?cE;TQ)hs5T0XvFQ(G{;U%<Sb5t2j1Z*jD$_Ec;efRh3mnu}^0S6FM-#
zE4g2B5yhsONW+swlOV(c8nfvM)*<f#v<P&8JJlqh=6m4Fzf$q?6fw4Xj}S3_pw<W>
zr_tagw}f%igwfP5=G?&H4GTmEB5e$H)#gf?pI6l7exDNq?h848qj0E``oaDUvr)%1
z<S91+z)87suS!0xu>9_o^O>T%c)TGmnp{9$(xF*9mARnYR~`C6U2LU=H+ul#B)c@#
zAy@}E>6(N4t-L(vuSB2rxhrj#kZ)kqrmGwYa}z|XM@a#d3cPM3m3&2g1Fj6aV{b<K
zRPq9o%PmoKg<taJ@<i49E=Eh1TJ{8(vkJtu0HaR~xzUZ{X8SRQ#M8ywM7Wd+^8#*z
z8e^tB^-rt0j0QC@4+2MUkwqEx{I&dNbDlqVVvO%`^1^<VwSC*9zka3t)@aA0QsOb2
z^)<?krvr<9AtDj$%M&){%h|ehuUEf0fV{<L*Liu8SQ1<~Q>R3$T*fMUcBl?wfX7#4
zDk02W6_aRMVgJpwv%=D2>AbC=LA=RQ!b)(SXfXOJUG7g)_^go8knmQ`Ko+hN{Hd_y
zRayF1Lb<`!P2pc8ck>M2bq}1%=wx06ecxZIRFxfY-=R(Xk(0OKp+}I~aBc2#ZIk8@
zH}~CkOeGH%8B{hSO2jjGtfv3^^@}~(p__f}6R(#6tMh*bUT=nCbAUmIr}E<zH^UQv
zG{g{l+mM_aQT<fMzWq)}UJTQLURdzmSDu@P^Tj~G%3*yZ7`<bw&A6>S|NRrs#?3lM
ze+Sg~WO}-t=$+FL5=91Kbih?~nMw)V__+}*_$66F2}Q{u=@p=0t-AP&fiaPq9Vc}i
z#TWfghW7OjdGU!3wMowX5Gx37v_02#<FK-8;6*~>m{83o5GyzGyDtBcL|J7jb>A^x
zYFoXDUGzL3F$)kfy)0f=&z9O<hI$$%;#(kB{fj$FJlWN%MaDWGcxGQ^95T&w^hlJ-
zDUy-sq1pPXqkW*kqb)Tb09QwQt^H8#1b_^f-u`=Qjss%c)-hAR^p-QF=2zxxZ~PMV
zb!djX@S~~}1Z$z<r+Fuhyx@1i%;<YR&#0`~yfs6V4^ASoh<NmkjSXZB?#D=b^It>L
z^V>aPQRS(UDim%7$$rRdWQjQbsh6ZJ8IOCzL)-S$r#r@TVApOj%e9}DCaWo>iK{?M
zEPlYYhUs|myZM}k6{BN$W$*Hju4}*H-ULC7jFhR*^uL6Q0#ijE1NrOja|xmI?dnAJ
zfssTqho!7<%6*+GK?EV2^%p5r;lHcRHWnKR!z1rB;oo<--iQ1)&Ko8Dvd4g^seO!1
z^?4SE60q|ypVRN<cHE=vfpvangW?wvG-;XFw~i!=<sqvsiNkl6+J9WXB0#(RsDZEN
zA`a`~E_7l<7@u!H12<WF8FqevnXE6~1CaqeC1Z%&8&qCK@3F_m6{JDG$h)w~`b=N+
z*4)XsD@4(_(~1d4i+uTpdSQ5~yDVd3DnhkSG_gT`NmMwhPuhDC{{3WOPBWN>Tak>}
z%kBhV)kL}gn~c+T4KC}_@vz3qd@OHaV-vb>{4(&rgBj0%0IGrq2M3wi5-aR$z;R6{
zf!|a`>~P!&gOk2E#*Z}-Dk6TMwdYPo?Hq{|SL2jm%QAnVNb<on*PLsH{U@)kk#A&q
zp<3Giw8h$&m^EhFvT&(#AvLGrBmw*f{&%5vple7_zPZ)3{&U~+$ytP~ZU#<9GML_u
zR08aoLB_y!(w<bN*7U*Hm229M7}xwl4N|<`8|;L$+GJ@sSK-xc=PqEMY(<j|myVN>
zfx-GWhaLM{2<w893$dKOvEsyb_Sq1M;buv6nMW=A5s4)O(*F|`Lz>h)=nMD<G6pcz
z4w7^W6i5o+z!M(r%|>e_WM^#Q8YNGt>;~h&ybdY(K#OLf6LG8`Jv6v+wc8v@-@KSt
zBF~<xQUEn-q}7MK#g`F?1VR>W)P$kP)85gOj)LuLxLadlV5k%c9~AkM9~UF4dh3)c
zAZctg%TYI9);f<Wjrtw^f__LHC+S}NiszP<<+h*qdhvGc!tn8ypMXnDT@^y<h%S-z
z&`L$P;`*x#<?f{quX|V6<80*ln>1?3LS<)EyAGy35pW;p)C!>rT@kVPs#Y?Stf|RB
zfZ*Tp#%G(p7-&OO^u7%0p2X#4-~#HZj<5T>Q{a*ArZ#@>(D88{sqJw~2lO0WHzYhw
zLT2znxc&I$#*~De1Bpbw6AfwlPPhKO>bul16G4`Oj+jSt<MA)=JYlBY@0Zx;$|-aI
zaozpIG~XvRUK<_0alFTt8Iq{5haBk>o=f_N_^iZJ^AKpy*8N8e6Wgv{LSH&*@ZSKK
zX)=%{Dsfl+_p*372J}=GvY9WvWT*3UYgckf9+reD+q$*Vqo-yMD=#bdUeN_(0i-E^
z{u-(-q7+3M#?`@{(absex*tMJvt9@h%vS}`<c7a7j#y;i6pejDv&glHdS;&kZ$;|U
zY|_l{RTP`Xw7(#2Bbu?I17An#ae00o)i4I2id;yZx`zG@E6Oydg+0MvW%u;W6oilP
zf$b6XV-&DLGo(Kr=rbagsP=lV3Vrn1brH5kZFdbyEDODT0G=wRbsMG#4;G|K0|W6k
zl<dIniI*yg1(NE>(g+$dE%o%{%4B&XzxsYHJ?l(XxzA=>dB)N~ge%^)Jlj7RaaDB4
z{O6e_yr%KtScGw8=$)#b(C~ue|3lV!N5k1h{aOi<h>}Q<=p-b%=$(j2qDPNTMDKMl
zdX(s*mm#9}&gi`}I$@OPb;KBL7|!E8=lkAwt+SRt{h>T_Kl|SMx39|ws~-#xm*yv}
zlr(pfw;O6cI*^u`S#OMnT{yJk8s}|F;J(ZUu>Ym=x{NPdv(Ak2U;EQL%jiaiG*ag0
z@mwsWOIoNmy~0&3NkIkv{uKDL$d`3bE=KNhN?EskX$Lc6Szm7d-qsrNv`DAP_M<Nt
z>F_K$W+DR4)t(MDCq8uAZJerVjCFWy1zPByk|=#ebYsbVl3rWq=r%cj98@~Fo&u6M
zt@2(w{awL4a)7%%T<cA#nU}*QMmKNovqK1ryWRtYerclbsUmoW6ftIGPmJt7EE`2s
znqx`v6GW&v(~@BJps^$rCuQmQ3ixMY_PzN1{4g=+Yc)BTm(9l*SfD3F-jsuGwzx*4
zWx(5F2o`#;$8*x7_=g0`@+-s$D3!Mzw1LuTakkP<=2|93p+XuLH6~_P#*W#(rC&Mk
zifwfDTbmZ$?37BczAx~&<wor1pcCtxU%Vml{u@ZAaNFXYWHUJWUgNSe;nMw3F6W1`
z1T?nse+%yd3;BO4?;;?VkhD|LMJ%6W=mXLo`29_}<R=j6oUA-%{S7)vc5h>8S#QUk
zoZqYyBG7ZxsHU#nCeZME6uvO;)En^Bs9=rHv8Tkez7m3CA;k*V2+O3Xm!+`?owGs{
z6c0uR;s%w<$q@j?Fy9;0q`XAFK3!IJfK72-XJTu^jvW)`tcA;f^OM@JrIF=;$LtOJ
z%uOATQs8T48r038Xwb(dD#B}7*>7iHV?cOI&F2bZ5I_31MATK@(4hQJod<Afv3f!J
zG^%om-V1KE-(F-i5(M@48BGacD(&=lVaqzNG3cV`oP`+!nhpDp)hvj(Q{U}Mjg$+L
z2CUbePlPBsUG0_RbQ^B`BPVrJO+!8(5InH$fB2W<Y^SK=jV_&OZ!IaGYY7q7N81;8
zl$Bq2@!`1G53|&-xc{0fc9}ix=M!`w#Bu?(QsiLUtXm2s6R<C%!h{S~n7`oQ_=_5V
z*e~LX^|5uFMTaSlrJ7KHFgkdGPQt@^S69H3Vlvihq4I^I=T0}Dxm)d@(7M`<(;9Z|
zOuOrI<h!B~Yxvb>4RV8{@FMu65JL!#tWKLjZVBZ74B?kuz<MVY-^4rg1~xd8Vauas
zqpcFihM2(G19&s=<-G|RX7`O&s_*7sF2$C`<Oq+erO+@TY$vyO_;GH3lrn~(bqk}J
z`l@y2eI9=H>re3$BkRovh4fp1v7tyM@8;|(TyDazYO@h)?;V*UY(rf>bO9SWIUIJ4
zbeB<DGk`(7M##?iwgy(bwc(rCrOQty<joDL9;ZXLVxdw3fs{YWLG^xJn`v8*LN&Ek
z=ERp??AF_Ad94rKCy%)!X}cl?6&cT82NBTOO04M{Dec1UI}rDH<CgbrI$25SNSkGe
zY@BsFXmeJ5!dt6lV1MfT`S2y?)hJD9YIT2+(Jz7I1;|Dt2l9smnxD>#C2o!9!gtP|
z)fXOovNHm}7SY`UEPbQDZ!X(z(msFwY@$YR=#uDk;`sj^&HA_d<da@Y#}Fa%tIJ!*
zMK>+FgTyChW<!?4wrwwPLi12q*tYFtUwcf~Lzl8j#dyTzp}pW6`gWYTY?|guDZF(2
z*>D=ogJGd!z%SECFxynK*;DAv*z(QYz|Toh17C1r0gc}^F&ZFlPSHnp+&R%@M5Hg>
z(NQt>gCI$};zNZSp%KD|e3g8pB!b-`*BtMTk_7?QbHnfjOO6o@J!rZCJKGq^+yt}A
z`xrMnXo(m5{`64iQtkYU$IZAe(yKjYB?w&|?x<tOqkh$WF3^%HuHvp)_3esYio15%
z>lwdEn_#N*w76vgtqNrs0jWZ}!M^GGrrG!5+4J8m)R1UszNWNpYMT$0<;7c9{X-@G
zc=7AVq&zk*IL?o{hgI^s4_PFI6jR5kLH=>*WBESAB%jN!+(`Ab*`~HC^WnWUV))hh
zIN@@${`^6q8jf~Krhpqa1<F6!3Tsmn6lGXhB)Vz!B+&6C#{u2<3p?j;j;5^bBSh2H
z_&Em4ixkI3*WIHP-@nihsCyZV$nHdjAxhx^Y%j~$$$t%5C&v%DKN$PiW|B33r_-aL
z+8nM&5ED^Nptk&hji{hNGQUQkr&+6<U_;~stKnLi(sjFWVgZ8e7jbv134m!ZMV^ZZ
ze(Lr`dHd-jwIs)8A5`uHRso?X05roV(LVE}*sC*@w0TOPB5IGT`Gt6c`C2$lpL@6H
ziap!07jR>ol3kdhVm}t!bD#OX`y7&Ej!x~}e7xK7OjWoyW`W<`;Uq3v#A0#+eobT2
zl@=eJYk7axJ2AI+#QpGR1^KY0)X9et4T)V=&CIiy$|-;LmmODZ<&CVE?Vz(<c^}LW
zDu_jb2mBY8?)~k|*OV#b7VwneJSw~+P2w4|^PjaJ7n0j9Vhii!PEm7=*JRpK{3y^(
zocjGcr4=w`1rHTTt*~zuaAf+z!wplH&X-8K-#l@ss;GCU-z(~Y%6cgqc;G0J*V(kE
zn_hN2yI)Y{_m!uUwHJpxxoB}z??<nzU0k~XkqH(DAy|4vCw5Ac`^1rb_}E=jGfHR5
z&w@Xqic||?JvfwP>L~W<dy%rOa6#kRl9xP8<g(qx`K&b3o@&;43y26G-X@oyjQ>At
z$~bOKEY3#MMK|I|XUMJ)(1r{YxVtQx!NPWuqYo3K=5JuX8vE?Z-M}ry!`Ggldw#^Q
z%;))S_ii~{XL@ggC_z3sVoWP!{?c*h#2>xt;0#nSN>V)fK=bvBQ+7{C5B{ABu}cX5
zP6JGYJWYi;i#(J=+%Heb<h97*;wnZJSI11G?@eMtD&3tc1@5+DCaKh_u2s%iNs5O%
z(jU&|-{W=1Nm|$IUJ~F$febAPwnqnTUK$cGVb^0tKu3^3*3jIT5Q!VN@)B#80FN||
zh=Wk}kVEpp`(KBsK*Uj(A05Tiv=>P{%CtQWZb$BQA7`^U{G`6lFC`(@;yX=ghsZ{t
z{kv`Cw2Ah2^OUxA^<A%#L$3EpzAMRmQ<E76-tE{sD44IrKl(Z#52|YO_pk5GneXlX
zmfe4tIgJFjy-Nd%S#{AuCpG<M4@Xa6IFQ+TT3aqF<zS~=y9eNB(Niw_3(JsyGN}Pe
zj1<%biR(=5(vY9zIgaT!yT`J+!2Rc%=YLbuLwGURX0>&El4tUg&*-&qo3p*Nn3iRY
z^aJUl?pH<Mi8spj!V)=b5nZxQ$mNc_6AxE<@_=20gdEM#)#B<2LaX{y@laU4@K1(Y
z>pTv0XD4}a)@L<}!OG-=>-#HJI4RoOqcNbHBx3NQaba}}$Eu$M!ZBK&%j@^hL;xP4
zi7Dg^fd72|Pt&O}q)3xMR`x8rJJf5`83MGGAk>=A2q`!f%rjgyJ#Q|T=|oSa|GDsS
zw*Kv$%B1kUGnf){5>PqQ?~3TZ{S|>2Jzvy!L>S~|&=$E|=U(Nq@Tjs1_m7povu+#(
zrJ$pg{H@+cS>GmHpfvU{`&sY!rV;ftca%0fz;V5$4?Sv5G+Pkt*zHOi!L;QskU2Uq
zo9TRl=!4IIqt`aAx}GL_tnxXz)*9oVX&oK=iOe|a=`OB$tTz5cja0*w#I0v@W@ctk
zhBqgMmAV3ZTOA!8CQttbfVyyPM?C=+i(3lx-52>=FbDS08R)W>(|Odw^0k0I_+<JY
z;H~Fe7wydZgo)d}oXZIREMM!BMLIsOE6vDX*-iK37p#|4&L{`{t@z`AgfzcnOLUaM
zrOubjA8+g;e-b%`KG{?A_4WN+W1(s@$8EpV%xBsgQ4wJcT!|YTR|Ck`(Er|wegBba
znwt=v0ogT)C_LPo9xk8Q`T6+?7Kx*<IsXK#rhjDN8-oc)Ed|@xoGMY5+miP;uy{eX
zA0*i9i`{T(B29RMf~jqZGfU`V+mHHD?tKRDvODGvs%92*lkyTh>~!ONzFbQjdLL%k
zpDHfYv$q^ipN-SQn(DyvHD-mybyLqqTHHVSlI#k&?a$pZ!}G$qT7{NiFyQt4$=%%@
zJ*d_CC=EHBDJgl-dWqP9*!6z;cPUc6up3A&DC2d50qCd_AdNGe*#BKUS>Jk6sy+7G
z@x-;<OMmy^Ad=FlWCwEI%@U0FZoGvHQDoA;_!MZ!5-%Qa?2L&M+jtk>(bo3W4e&Cv
z={K}WJtX>6KEjj%(<A1lA5Bok%5(c&EZA3IYVd(8R4hrn%-h6PHN#-IMC4}I^~F*a
zU&0XN{PU>O02$K|Kc}A?d97-wwaFLP^_qz}|JkUz>0TUD#{O0j)~?ZPnZ4N!v4$u2
zLnS%gXh%LU3ANU&x)=T~4ZobESaQ+;P}Qgyn;bWf?Vy_Fw5{A_NIjpv!_21XrEMVu
zybiB9{=qDJGe_COd{5wh0lHy+Me3$JiYBbD+YZvQH`9~HtVd_7ZbYDF7r!#(z+{&~
zY1(TA>>uoxSW(;N(ff)ve3TIoa!|^I+OhAm7nMVeKE9aIksp;`rV>pnb=h*(W8Gkq
zJgM<I+Vw6Ux!0f(mChyLm^?H8Xv-de&`8%GS}6nAqn@a2H=!9~vpLtPsIm#S%~Ot%
zogSrX!d*pi%s3bC!oHLU^Pr}0)9U;-gFtZqY={%ami!mQEq#Nwh;!aVY5KKq!jSCL
zS0cfM9e%^ih2Ich5E<Q^x9)9oy>eaWt+*UDk_nfyUCmOS8o5Xi&Jea>Kftlpg9{1(
z&_FGcLi>%xM%Gy^$%V(Ad5<u{nWO2w@L9K$eI9aXu0VbR$5!csG%?iKX-++z<6HFZ
zQ$r=*`gvBG<MHVHr(`TSriSkCZz<`w^Hru`JuNM*U%CH_nFfxbx4=fsVm=O>SR6Hm
z8nO4`b<+b6$Q#5(iJ#6*&CVu`0&Fwwx$)a+(m#j*8uzQq42DBR4IzLLw#cN~7i5Al
z8TZnf-s-q}6B#7uy3ay7sESqSO&f&8LcOpU+a7)Yd8zW=F<=W$#E&naz)T)@E(Xx)
zrNqGGNKey2lgan1w2i{=nX$;G!sp{>9+VNbn_#yZovflHw^{^t8--h(qG$72hQUsL
zEVfwmHOlNox?Tj?Q9vGQl4ub#E@6xPhIs=y4eMMwsRbP^CmmyTqQdLsN$mQSB~P~Y
z&?-it6}<MCw$nKVvaHIr9K^JQsNTf>>C7Arnc2fopAQm}l~7^y{Adz-$fp@i@*vy$
z)vPIas@zNuZ!O|vP)|vY)9AHD+=84`FJQzKhUH;}%A@foo6G7fy7&R=Vi}nz6U|G(
z9Tc1dNZDAsEV@CJGbiJ}f<fGoEg>Kq1vXkqP9n)%{^&pGDamGu5exgU450%dJsH|2
zTRX{Crd|_AyC&#sX1cQtiP?~FPp1WB)(FgUxReSQ_W_*Tp5(<DEM30@(1TAL{m$7h
z-0Q_CHrHMVkJt&f=zomN5XQ*{6n+Xr&Zy0$4$AOSWJ|a+U7(!kdgI3aq++DL-1G>@
znP&aCnUdm!N78)5ZF+}g6AFmV^=*P;`aIr*L-_?b?6K;0Y)7ye{40R&ZejdSMXa@T
zJvlSI%0?D9?nQ?B4RkG7l>lQzFxjf-d*P|{0baZ3Y&b7I4WEVi%{P~wdaUAHZFpj5
z97*)EB-Cn@iN3!(R!AVU`4mfXL=J1@N<;DhcVTm#514I5x~ojFoCQ{W-E5W`=D(Or
zel>f%Mm9XDIolKxvWEG&zXVr4xJX9<A+x^~TjjnAZN7)&mSAQ3XKk}TCcB7zt*ykj
z<aa2RuQgM}EKSg-^=%onmveoadjH8QUq8GZ?_}<w_sXL)H0_}qiKuNkNfGc0P>kLh
zsM72Ac3A*i*88n1V|yVnMF_V<R%Fj}A>07%Zi3;#$?5F)Tp|a&3GL+13LkM>P#Z4n
z@7M*Ta}SzCCIXtZY5T)LJ0Z@`2-PjIo)NkeIuN=fN*VP46ALOc{*c^=ssvmmY^iJ9
z&VI(v*X@x#yee4Sg#>=wwYe|Ij}b&d@tH!)Y<oFBv9RgW@I+KPK7N~yu9YAxl?5?K
zFl@2-ZrG%i2WrhfsJJ&wj8##>K#1G{Rkm=N|4rJkp6JR~dQUY|`caF+!xunHGAl1i
z+KO-R>_lHIl`iO-{`8&Rh42;s;y;tVzn4y+xF3P^_m@*nqr|LvXzE(;4!gP$ao<^z
z;)%hj!EUA%D`d;+{w930$Q&JQ`j^NL>%KREbbD3@+}BxT;rQgIADLeiaeO%scix)w
zGB(4BkmAtAwl%oHm6qxPn+L58VNadfnn+m07T>t{GT;rCiG5iY@DnP`x}I%b&E$fh
z-Z5_>gLqZL*bZIP;|E>&Zu$ILfWx=pn-%0Wr{YpIF9*i5_VRB_?oqU*@}!}_s{=se
zy_mX{b3s&;ss$nZ`^{`BC@-(ZROxv3iFMZ}yL|ZBrj9#eDA`?D+-ME-1%#*RiC;bp
z>(AND^HWMH-Hqb=k#<Z;Ow#9E43lo6arQy)exF`69ylEv(BILL*TVbqTB3JAFJoqg
zS}NK#tr26^lj;|soGSbRe3Q3;!!G_DmQjop<4y8=r82|&x){4q<XFr@0TL~#Tg3<|
zmRpA&>fo<!Zt}4qS;B>!zZgRZH@o3iVi?<v`F=W!7B@vhAGg__=7N%m-rNqc(ap`e
zjMB<`of<$Qf8N<a+AiEZ%su>KBl!uH3O+TiG17f`&0OiDWTP3V+1)K=ETwF~Re+rA
zxFri9GuvX9pM-7AU#{lC-}yB@RUgLe=;Mvzo-E-(2oD)e8R=j}g8o|{Xll$pjt+Ts
zRIh{tLB!teHaKMkxG%97TP7cCqXX^p7h2DEYK(>I!EH&oqxlk|CqAc4mLrBYq=!c^
zpVR7FewaYd|H65{f4|25=L2LtlFGjmZg^!PvRm*AV126qNXx*9K8e#xKF8H=)QP<}
z6L>|YFJGRkk$a*Q^(#|pLF~v9JCA)|XA$dOduRrZG*7q{q0-+63WOD7l8-fcoR3X>
zRHN}u(ApKc1&wl_D}ue!zikH~nmu5I3kQuN6Tqp8<K_q~+0~8_6Q*66{~>0^BGz@U
z>BXTcso1U?!-rtq57+-Yg7f>$?@sFH+cwXrET^<E96{0&u9YhI5bb&_ie0LML0==L
zp{i$MI(&P9DYm7?8efl^_07$|Oyce$2wON#$15qa$Jj<SH}g-D^6B(EO_ZRZ7LRDt
zNeD09BMk8yF{k+;EP|5nZjiYYeB>OKcI`%#QwPP22Yr86UW9T|$v&~i@z@7_OIKQ^
z7%X7WmSymYt{%|Kwc+xyyS6_(z~z?`J;2eX*;CV07Toli^_h7OoPPJ%m~3&$Iy)<l
zm?tTAnUy!T<!0m)h8VO4|D+SJ<#k|rHmwU-DlfJr=x)tbKqg_%P8WW$rxgx3Z-<jK
zpuWT6u0}nAN1}m+7)CcG-7=ydy*Y5jLJ+XEBSM=jD4>tdCc>tJ{z~m~ZknIaghnQ<
zo30sM<Ca97FwNb4o3gC6b@a!eu#HO(DxgP;&9`GZy5UpMH=w7Rd`X!;d6u+s1c~hv
z(BpD$lWj~7cIFA^p7{YWtx&qrPZ^pAy)kdke&(_@d6b?^_r&Fy5=jdD$qG7KFEd0j
zh4kdFnA1iCbW)NO_Hu<t00G&`T$AK7BwLv~cKN-F;N#Cfx#HXhUT#YE`%q?FCx{t_
zspaaX&dOFSZDoxJq*Tdwr*4?exzBl(2H9`<!-Uo#ZV$iC-(r(KYa~vo@q$dGlb?JA
ztUhX}tMjVLh3Ri93YWX@(U_<V9ZtgLc`npWbbT%A%BI?CirMJtlz`^o5duf&r216C
zX}8<2Z<LB|&33q%1mn<w(V*jOvQ$NHh(>>uFK=WVS+cuC1(ATo=nNs(=t7eGC-qCI
zQKE<jkLaW^oPUd^n_D0`I)7WR?LtxHV*SUY@K!q6vNC|qLTw7R31cl{bNPbMr-E%M
zGu89Se6J3UI+Szz#L(l@t73-cIwr>Vjj-9tclMB-^-<t2;}5d2wfM7)!J87AOyo%N
zbn;#HZ`#^n(fkc|G6wOutO;I)ud$eSIGmt$*x=o9W^x@RTzc$p<a4?kSK7PhxGeGc
zQP1gK5|ya1gj;Xu@&%e<C5lwL#B#kapqc;5K;ok~x&=_cO;lm%(4j26+Hb-`&q&Jo
z*zvJtum%|E(qrhx2>h;$k7RdlfPD<wx~_+oC&8t-y?!=BPIMM@vvS}(`>s_kEgW{;
z#65!a1F=e``qEtzaGv*xTGF02*M^w(Z-2v*F;phGMRlt)T}68%sgx=7&T5Grs@^+-
zY(U<V6q~+DVvD5rS-orv2HPnCRzyTOv4sX%NgPTT;balA4PZSo#X-MPzQL)IVrs{p
zfzJz48p%VjiW~h|TJ1BUv9(L}y$tLdPZ!vbM;I9bAJf@nNwm8Dv`!n9B)E`fB}Vv;
z1XXPh7!XAjus<A`EuCM$xu4SYP?RG&!@~@#y0wu9_5l_ptlu%r`Yy$Ms#(B<U0U3E
zvl+Vjl-~_0Sl6p2LY{e|$1TzO7UP60Afq%_VudC9S0Rofu)l7`?d{?SJ-zwqtt2iD
zBCi?5o|-b0OeQQ5$>621lGl3ngY2@*``ALq2Z9Q$^SB-++wd4=<AW?qUd&ChzsvC9
zjn+0BY=iLnl}o6!C3{^!JUOc-_{bsrnwgs=5+)j3%@VvyKtYSF=A*pmZQ_ouJ}pjj
zzK;~^|M4OV376lj^_eC=d~Z9IXnpOb>iJvVFtx00GZnSYA}O44UZ|XZn0Z`^g19#`
zxqUVpX}-C>l%xIkft9^U{^yC^iC0q!!&}tQ(ogt0lA53-elmpuY(Jv<ppAr`+i>jX
z1_$Ha?R|x457e(o7G;dhyiXITGPI`Qj(%rn*Bfy<GWxgVU0*9<+G2v&>=PXV;rgH{
zvPGPESEtcRJg%|l`rx+Q;$uYpm<EJWL*{~+5JZZ-T^W-A(M@+Y!s32M{(HL99n2oA
z(^WbAhZWzdK=7RYT#T&**wHW<Kahy&bJcWXLxQosHKT9~aFi<8=fM-_-gaO=;ux7s
z9gFJsrje5(c6zBIr+F@zY-PC&f&7p|hqtNXj$0O7+wHEz!`p)!U6F$si{s=oQrFD<
zf&|+Zu#+zs0t+WVs4=vtlm(K?YU6P;lJg7n-WQw8jf9iYG!fH3Wx>EA<2DkT!_20;
z^YW$=C;j}hlxedIaL0^^{kT^kt$1%E7*tG3P|-)LtJWo`gk=5AWXDhSSV9+)&Ud-7
ziNQ(Hd{HBUeZB8fljVii4k4{C_5XI|T3(b+Q=e<?29-;SjigV{oNK-k6hw#DMDL>0
zZldH^=QU{FdD^=Bg3BF|uu$bjXLJ$!a%3)FD_g#j<fjSBeDvH;7b<zfXl1YhkHN)U
zeN){=nQi$(RV8Y4+B`>#U|BF#BC|v}CFIFWld}ZXWWPWZGp2>ZQT%qTCRm;oq$Rm;
zOd>Bw3;MDj(*l#Y%(+$0TMov*a%}l`UbaoA2Ka|=)M=dh8UESbHSLR1R0cM88SUW1
zDZHi+`)P{W+9XF(cv)J2r3!^ZujR{In+}!xbql)@NAw<EMVQU_FykE+Y?Jn1uSyh=
zJjpLkDcC4xwX-<3cfv2VTkp+~a6{=FaHAW(J1X7*YZp5lGV=UvsKj=FPq&Ma>%r$b
zSvdi)PSx7^>C4O2^y~3S<_(eZXiqGwl1fh}A7>HPAz@qNjov|G7t>f=3oMU2k1M|#
zk93`=sa(`muHiiQ*IAXX41I{J{Cm)Be7)6h@x0n)&`{#xg^kiOsN<>S*Vza^4Rmx*
zWc{5aJy-nP8kq3k9DIUS=MQ(9n>F{lch?vF<3FRXGl^)gpQ3>r)g&NvQW(1KQ9<0c
za@~GAo2==M<H|Bgn_(ib_uH$?lVx-0u$rkb1uG#@OVh;@sv#|$0=9nx@14&qEkO+a
z22kWAuxaJpQk{2#$dp1yK>sbTZq4E-Wep%}<_VZ%lCJ?3OFq<j3;6ca7}TIv3o8aq
z;`^S?=jL+ac8>YJ@AaAn2>X0B(BO%~_ivBg`%4$t1UbUrYpSoLCV!ZKPl%FMe7D;z
zTs3i2ttE{VG6C4R^-4Uyx|w*7pw0ottLnzKt1w^znOBNxkJs+I-XHBZVWM?*UFf-%
z$-H~UFMjY=79>#6vtTthjt-l>Wv}n}$@j{TpwNoq!-?@gt>WkcA`77rtPFCR{PFHG
zG$E;{)(fh9|07ph>@z49ou#GPUej!P<`O22MDA(YB#D8`*VIM+>Z8(a$nGs<b5g=(
zr0ww_jrTr4_W+yUHvE=rgo}j^`4a<<aAka?9nDMiom=YLeR&RTL^>jqE*&HyMrg`_
zhVs@y#Zb}XhOJNgefVaMgytQyY2{wC2i1QkTxO0Al?6FIKmL2b(MO%udj1U-k_!i*
z>T4m7n~I}`4r~8y1ZMl{0sRh)TCLq;BMP7=a~`zagio&a)aS^Fp?P)~%0b|HCSRZP
zmctMA>6a%pb`AG;TPMjNac-X1*OnV)UfoJb^yw=XfR9X4$F7o=Ec3x*Q`fJKgG6zA
zd0``~5tocjbvg$e*q!YHh-mgdaPu`>yF|YfYY>JiTzBsa;g^f5_;7873$e<a>h{{K
zuOHsAPjnaI_9BSC$BJMuzb_5`2U057YWta@1=$BQLw(ha;71{bNUlI9BkGrDQZc@R
z4Vr)%usD<VrgJ(v`?%g$0HQiJwDHU$Z~<Zn5?))x{_bxm)sXBg!A3E|>gyMAl5!Se
zO5Olp&?));hoAs^Z}3x;wi8y`8A<r>`9~U`?GBLmFgU>|?im0zaF6z!F8gbO2us^e
zuiB$y^AnKNe%aN6eX1ce)Kuh_KRpvhHIrLL=fkV-fi5hxy$V7BR4bmAGop~(U&Lbc
zm|eiX_s@*}w|QoU8?Z>ztN-7(QNcfNBkMnao^ui=#7lDqIm_wj@aH=ixXfXp2H99q
zRWs<L{7om#+HUlj%s!I%bYSlYbBS|{TL6-s^wXINiRykc=+~nGvp6c2+0psIVa6v8
z)u;PI;h)0i{Y)=z8mx8pKPQT?w8*U>R(|`&-H*AP-7&nkfw;f^?oDyz055~A7}dog
zVl}H2X7~jcMXHS_;28L2;6qk4YV{af`g)JHOh5s<42!W0wn{XQNo=Y{Wr@5mgTT9U
zHze$wAI^4gkG&ck=mBCxOaAP3)7X>FBBb)4`&LD94A?bW*mBHO_h_;aR_u|!M78~*
zcf7aIrL@S1^T-QtUptCWymo&x0V7qwZW}CL?^P3BTPzf%w<KNl1UIlUt(cCWrlX~%
z3VVFtvym@<{Hy4i6_vbD*%p>A&DJL{hCZvz?airaz$vyDdKjMtZTK0JqhI2FTu6xv
z>Z4r0&aO?;Pv;cA2(;`unavA9U~>4@#su^G*qeC3e}dK~iZqHRE|+hv?+T8Ms~+I%
z{@Y|mMc)Q`e8YUu#fJGXi04Oiua0@o*^Th&!A=3wQu3(e*+S)c6A+MH6${fvE#F|4
zEoahgO^y^wG_6!rYCuou-46%lt5d_=xsKJcdtHbej$@g)2nA<+^Xj~|&r!Y65-&Lp
z>Ap9H@>^7ss(5~qSKFu5t;e0iwoTUtYQ7jxfY$B;cDJv|1;OVQu%Fh;wG$NJtTsdj
zN={%|4dUdt(>oF`nv+GIT+r@>An6Es(wydGr=m&(USLD?aF@jQBy@5dZ}1oFhEA%g
z%eA9c$Y&i+h=;1-_?GL0O4vG48<yr@93Us`@>-x>+;%0<e8+uq>Y7Vd6&7YDLDz#K
z!5@Vyi^+4`Z|N!;PS*T{e(biN=(j0aSbg`2cY7zXC5L_Vq9;6p4##Q!9>)kYeUL^R
zzIrTtQ0mrHs`tTm_;vav?cZZloFNgrQ+-8frt`&S0#7Nu{*&cc*QInNQK@p0a1rph
zr5>`Sw~}eOcoB>7Zev}!B<3hY@_o6_0mG|ODy2)h^Y1Pb_xkcpqa|3uKu%SDGEY(7
z5CFMORs%7uWS0MYoX(<dKTb&ljE8G<Hy(}K+2Nn8t*veLD=raD`KD3A%x))C)YK6!
zw?a(A^G=EhH3%}E&$u&4NUO%|8AlA6sL}(dSCL1`y4HQdpS{KsQC$oCkMcC_d=s9K
zL6;qgv~3<3=WG3K2l>6ode5Xr|K_v9-5TTHck1qF58_z2KbYOX_38g5Nf&<6_i?~o
zrH6-pRcsD>5Q`e=+^nH14}>h}cboOKy!<%EemSv~z#mVKv=|f{#cF>;<M34;BMg;V
z_!qDCYE|of6E$1Qd3Q^$p89E(?FAsNv2@wtiKiqvfEW&H|EWO%2OzY6S$i~!e1Mg<
zxbz6qUnKtwcOoOYIKx&ji0wu3YsC5>VG<%edeQYHymj{Qov4(y1Ubq0`3~1#C0{Du
zGaQxPZtcJj_EF}IO4TS&s{C{@sU@Jn5=(K{Vt-oZx7%IB0O+OHXS=2g>`x<f%MVXt
zX283IRh4|zxJUqQmh7B>$LXLM>0E0(s4}I`{xQ=psWHHRH{ae$UO3SrE3IyN#&a{U
zaXJl_+_jraCr*(tZDiJ>JA&MW8iBPBrHWhU`97g5d=M7vn6wGkS^XrSv#ynUs}Z@K
zh^vU~l~hjvG^;pb(fx{BFfXccwuKYCEkpN_j3seI`rrBip?F(9a82JPIZ+Aa-Y+<G
zs-C-@;x4h{<2@<&r#=8;Pdh-FW-xD-q>rfCzS7z)BV1{k?3Z;kP$rPdYSLxb{qo~m
zT}><QQ*{}3a-Q~k{!E;>1mtJ0T@>ia!k@3a@j1L437I@AUufP~xXl5lV_K;})mS*M
z;4s%RVDjMYYNBx`2J4Yqff<wu^JPXHieN9_w-{pW+iLLUI+M(jhPZkrN=NC?*YZ2c
zQ#UU@A(<kj8^TCMRUI2+QB6o<{Vu6XwBRD`<##a3@YKjsZt;Hr3a~}ih;q)#f~q`m
zl50NHFHMP(TIxK;R+SnUK25B!D`Ln(2XII|@?g@`ZcryyZ-V2zHhDCt*0TObZ!y2=
z1}bdp9HMm!E?aK#l-WfTq2Kxv+ZXt0N{zywOWmc!$q-46a5u{^v%wZ?x^6|-w?&UA
zNL2Tk;)<l0)o+{zXy|c3LzD*Aif%>-!P@?VM3z=c)eKvOnJ2jheEGdwRRZfb)UaI(
zA~p%kT*Bp~_g|Zl{^JjA;cEql!-U6j@OOzGC&8@qCe)0rx`;EH#(7hACU6;<mfEw-
z!)c|he!E{|rD)b4-!g?oKJ`K+C=xB)q1T7(JmOr<e>+ieNzKo1v*iMC<0oeT167;`
zj2f@}wF_87DINaR@)c|^Y223VpZx6zXg!QX4x5iB(9IU)V(HV=d(hj6aHM(y?O8+L
zKYPxFTwFWW$wnQp2SZM}HZNyX7XNf=`e)DEj=b@<HE7qp(yeg*Nwo~81|EZeIYoV8
zB2or$i*TT6xC2QTQY7t;0bQsy{DR!=Om=bK>h5WOnk^xH*9p;Gm5|Y|q{a$4g&Sg|
zSd?K5c#r*YCf<r7EA!Rh!0t0HwveCy{37vPYzl2Ie@R4XMc|^zC2rm}OPPBTrw9{=
zUMuMS)Qs4R1MWi_<8h|L&#p$EzUVuCv62*T9Scx$W|-%Yxw(qZCb>Dj8Roe;rRecw
ze`@wZJp$>Pa(oVC2t1-gg;pJdptUXQ(_fV0af0|g+9|45?;uMuy;>cnp?_TAGD2Ax
z7#^jw!N2u(`n|x!*3J04zINK;m*u07cjE0IPYO7a<M*hoITOED1I=@rVbL+CQ@mG-
zt<RG#mH!%L{FNk|)FeSPFXlP}W)LY}=U!?vp3OP9z-(>wNz!R0t98@vASx>wl3WHp
z)>XSHYwFd;@j@^pTJ>rhLP>;v5NNaDWts20W*;%feDcLfnD}VEB&1U2NPDky*KxXy
zD$s<fug=2d^X+cdYADlc2x(&f?G9wBq_0o&53aT=Bx)*Of<MEt_o*cswG%1&?<w#d
z!ST;AKoQ=sCAKBJvl`B&q?WIwcst<m1a{!b!@F*gZjtJ*B5m!IlrMa_6c*?%qV6UV
z2VhfU5ewm$q>M_n1Aytl&uq)?8@rI2u5(sO;fn6KVB|B&j=PIp?(H<7%G9UJ>N@9G
z%4PmuZm(U4rEn+XahWTw<^=)@wElieqZ7KT0G!p4+5}+<FQT4ceQLTH2oa@izcYtL
zD)nRu%d>>G2xkOWYIJ}MuA)oN`SL9}Wsco0uFHE|JwFEI&hE9-GeK{e1P`)j)PjPg
zY3G9tT8lVS>s6P|Eq`8V+MpFXvhG<Fl*fbQm!-Pf9<ANK^W_5`bHJ;KZMvOY<Da<z
zk?)m^j)|3J(lri9yMCuU^l9SZ-q4ngfV2qw!HLd71CJ^#FqOmkCN(!Aj<M_u%ZSZ?
z%psiN?f0faS&zrI;Aa6kg?FAH4_MM6j-U9V=@f88us*b0JHX=JGb)y!9R*dIU?Zd4
zEV>^*|Fi*8-k!vBvq)ku?|U^LHw9jxi8*Z@lkOW<zNv*pB=@27X+_seu8SqD`JA`&
zsO%L8T964lAC$dKJkOezGiEq+QfKH+8~ILqxRgxKmL=zP8EUhA-9cJTTHupn!wk-b
zm4KQph*z*HwY>0_&F0@*2-9ZAKR51wd4^9Gb#Fs6J`CJZP4TkyXHzU<f9@4FHHl(?
zLD$D(Ey3Z?-iQyTED9fgK%j<Ji-+x4ciZ(gqL|CH@qW)$SDF>I1u~sHCy45ADGSuj
ziax}8Fslp-5H;<qxJyNrC<>?0Id&0nY$;*??w2ty4l~UmpTHvj(>XhsQU0o=joP^n
z7lI{vK~hq@@=%_>EY*mVtx-c}kC8|PM+942ikx(va!k!IM(GQ`t?77=XTkk*g9O}$
zyDv#rZ*!}83F5OBXl8!%Rn8q*FKR08Fv9HDCuDxO0{3D_R^jtn9em*BAD8x?y3Pnf
zaN>jjjcr1)#i+eg*)It^;f|Uj!@i)NY#N4I$EEYAn)Y679l+R!S1`#uqmDDb=j=zS
zFwutg+K`$iV6VEfCP<7{Fzig|RcVq|Gwu0C9;8O#$^c5CLkf_fDqDZ?LouV%q-@L2
zE+XQO`TOyX+tqDxsjTIpv9Beqk~MTezt{iNkmx&SBHBG?Gc*NrZx_alGd><~j>vOm
ziVC(k&+4m9zRK^0X5X;?mf8YRKlTk}-^tF9No`^+K!{24A2E;Hf^8DO*-O9FNcQ{p
zT6*&w_zb(W#aO<42S0|KHJ@LH%84-oh8DXWKo70d($eyZpFg)Q?1$P7a^X`S{9tnU
z#$C$uuk$HkDBc|a93l;DS^<Q(h0fQJwP|M0De=3f?Y{kqn;n0qsh*|Lpt!mBLsv&Y
zqe&7+#OU_~zBjJ@o&3!4rDtcjth!rbQcqXp_SyZS%#xeoYm(%j{9;k;Eq-NyrFhOy
zDg%9taT6b=8z7D@DbW6{bK~UvWt`x3r4!*)@+1_Ww{GjQ=AJjxM&L(&qd4=CZ^W2C
z<2&W@*P7tpL9_s$esb$f{1v(|IOz#6qyZI%SD(~=HeftSi9MTHu~d5oI<)vwcEy5?
z!;Lua;)24m;F=4z?=PF3gs`ZlGu`v?wlXszC_rF9rB!N`KCn~2B$DB?nSwnWe(mgX
zyLAb8R|S@u-r#uyU|o6LtxAa%KipM<M~AD*-m7Gm0N>IUuKMo{$Hl$~xZzwz&Xnl<
z09co(P|LS)Z4OdppcQ)F>S{n)S6`R6S8f<Xdv@M35pHmXc4<F`P&xMby__-2qNWf2
zoBW%7So?$jN4TI)3?BRUy(jmJO4Hkg&ArPz)ci8N^<;8m{la?4Dm~{Z)jz#ux@dd+
zT>N-HDfc$B)oQB;%eX<ITKni0&oxo73Gls*#!HHPd(4h(q<zA>PAJ-bVc^H4!1C4C
zdBEQ?kz#ARBTjeiki#%b0u0?84B64M#XigDD5WB7AxcC#-4pFo&n<@$d{j$)5oIEk
ziA;nc_+S_nEZEFIw6!M@4Gb?FRBH#K^vm0>Hn{QuQSEO_y(>zU|NcR4uY8vY-CQOK
zM?gUMc`L5VEg0dSIuvN4bbU76naKS!iY_r1XQKhF$L^>`d#dMf%mYJ^%QMom>E-nL
znxlgyGJ1mHNKI!J#gCSYjV^2RwE$TKh#j9je(6r4A;b5~Qg)ZJz}%mW$XZYIUim_W
z1UtB=p~jip8XhY98!$WD0-IOb$N+1B*w{smmg!(#j<w$IUQWGkXsUC^*Lx589a)wg
zPbN^|(V5BQSJwd*QOB(Z+uQHuXk69NU3Awmr$Dd4<?5&}B*E!FF=+6wTV|QoYnofD
z@7#$!AjNrqTN~l_f{{NmBe2#5ys_5H8h4APrl%7Oka0;lG9==B2Muf6CSSx=YWsGS
zt%0T45Ws@#vQ0}M(E0?!=3*!%EM+&ZS-KNUCD2ycx7qzkDMC)u`6kS+gsId>E};4|
zyJe43iGZq3X0U*|W`o{uz|zY^kx#0!R!1E0VW{PH)ybHD2xnHQyTsA<9TR?bqyqUF
z(GucOTSdJj5^wreezmt-n`6InI%3k03$w8gBa)kPS~@=s4f`!BMol#mYJPVmEMu<4
zPCs>NK5pKrGc)t%7b>H^O_=Bum4i85-~t2Ma+0w@+uHwGYFr5<DSM{(pZ|R3?NZ}q
z#*D6oz%`H}?@TxWs!j0iBF~T84P{`xLMx}-7qhqXovrLHse3;|NKR5KD%WyzI3<$$
zJE90kTDJVJ3dvVbz_kL3JFPOhb=zM?h11;#VO9eOSK*1f*Q<l~;~cdJa;5lI28E4=
zzEbJUK+BbFEbqH3rPT--EY6VzEi{?>4MjogPl?`2xM{VzDJpB{6Kx!th|DW|%T1k&
z6}TS@07A@ed&!Xs-SN%duF8Hw-21)WOUy4<WvWybz$RMHhIh)fr!!~;T;Bj!vtgY|
z%a64`!xlZK3&nwfkiv`HQsmZ+`JotA@yi*!tNct<kQC0(Ggety_Ltf!l}2p^T0Lc<
z2AK!*l>p#?ORHx|l=VAN&Y8iZ3Hp|L!DU}J21X4PhO11$!+pS!F*=M8ZpB=g2PcGE
z8eR&m>vB60zY6g-r5*S(TjuRfdu)a5JFzf~5bd?<B;11ih~0t|pS8*-ak8#`4-el+
zwJwVVHu4LyOwX-m%k+0XcX2=X=j)nLq5tUu-tiVJ2{0RI0bEAT%Y*vNclLVTM?G-~
z2`_N*AB=&v@M*-<8Go7bbv|EeQLldeQl^&ijjUlw*ylgF2}$}Z-_k&d+F$&I6FpM6
zoVjM$%i84yxptfHT(a4ti>;6Y;X=4$MSB6onXm0LBnfE7*}R=Ne9e?j19dx~TXUOu
z5RO5UG;zsiAz;am=K3)dSJXs9h7`=}b&}Wi-SSo5nl;^cpFLjrLW1Ae*;>7eo{^<C
zQ2H>=P-xv%;iv*S8xh=#sFE3G7%YDZUfg~16=R-k!F|pDtLydRS@_}o<%gRt>xCOU
za_mQ>)noUUxxPGIrpxY&Q7D-jT74O>@H7FT9dH?C9?!$B#&_Sw@dH#e$v$a2M#=rd
zqPfPGR+vrs%W`$;KJ!msg39Ka3K|rKY^W$o8KVw)mUVNeiV2Q0^wtLv5Xxd_m9GtA
z?*+RlT+RF`cl+F9zUej)AAFVlGT*n#72k1qPt}QiJK*uU_o(!I<vHQF>$<$t!K(XR
zP%t&BZFc(|x^n5om;mE?z*f^0LeJ+cu4Z_Jwf_&IzVyMsHgfE1w;uGeSdUw7RMIwx
z-LOsppP0Jf4Pqn7FmW6qe3#oj%L|{g?aHYo&zLg!pD$k;=Jkk!;QA#``Ct+{&RT@5
z&w78X$RD9`>V$-Z*XJugCoY2?R%I(vghhpFb1cN@H*X_9DS!D8<(0yWyrUzhP8h&3
z(=vuM83}h(X^5pQ`oslm8D=6$F23jeQ?BNLDmPF38ZxB8{-G*svqcgV&2aAVXSDcB
z!i<W$LB<eW9v7_S`=Gp=^%-)so-3(<Dyd#kL23)UN;tsZzzKN#-ba#;tFnqFv)oS^
z8*JHTB6t}DPE!1aZ_6xsu$t?w<P)riyb%6;6X|Gm_+z~R(YJ^5)c-Gc0`pYt6Z5c{
zpLQS07Ev^B{7%nt0%$*7tV)CpOOuW%<q*(|Z?75A2tRsE2rphC6&roPS>!+bbJ3?w
z@NvA)oyV5%?<eW!;$DUlWWAi2mHYeA`uW!`q#$0BzU)M>cGIX&+ALn5@JFBBVDIbG
zH2WO2d4`Gn5ldb64L8w(!gS=a68gcT-x_P2oxGZty;P5m4g^2B9at37h-xAYK{^oI
zWdl2}eV)T%`d?)YAIlm}Kc`68`=|5llMd*i`?gy)OBeb_?3<awb~sKXxXYiy@L#hp
zi08w}5F93vUPH$fH8pVYKJ3MJ!ABV;*VRuDPote+ng(VG#MrW9j_J_p#^>GrgGE*Y
z?#mplLr+SosGl-FWcbPYH?{E)C%`)FlGBm4i*Ltp^Q)T9_OAciwGU<<EHHtASDbV(
zw^p{aL$YFCHG7C*RBVV>Kd(a5xkQ;u8;RJK_iU_r&C@@<rJJIM@}8~7M}MSW^&iRr
zWz6bXsK(DiGGwGkToOZ>_u+Z;-QS3MY@LBa)hKJKo5s!hkmY;mCyov96OF%;cRieY
zzBTzcoUD2!Zq7>?RxZzAmP+~uVk1HhV>Z@W6Q^qUK**i^7tE|5f!wD}M%$0&DLFc@
zI>jW<|2>-?td;i!rB^PtQ#+AJUR8`bKc=ApnpIBW&!0apf$Y89XzzM)J!YEA*b?b=
zK|D*by|TW2QjGVI3uy1#zHR!pBhSv&(CqT#^Qu9*KZVi6LqeY*5-!x+O$Vi#g(#Q4
z81PG`CtL{M^|O$2o*Oc*nSOjmQz6Y2gQ%wx<bMoy^~n>=rwBcESSIioR^+BhcLMZQ
zdeUA}OT5aoV50HaLT+iyyf3B<!j6^H5GjdFa+;6wD=Z&Kf)9(Lo>4K0pLke8nlE<~
zp2j}9EA#xV3g^Wm_PNt3sUYG)8OAY|opIH9+Amm`Z*x*PcVLe>qLqU$dKz9&n&QlF
zrFT_`MUkDny(DGpS3(y~oiJ@U(#2B0&(*BNPIx5$aZrQPH8gzYl$XEtwDMBMuVLR4
z_A*oCwRg<>Y`4nss`iYb7TL1D>0{>C81;vHgZ+gCQ=Yka+3H9KSPa$shY+Hu4R!UV
z%RLB{J25iwW&TTd4s?`PZ-wl9C|1A<3L{-9CyBhW>`r1kPW>Y{R6{rWfI<8^V&R45
zw9zXVp+uT*u#}h9O?X;^$dsixsaBJ@dH=U^aV9MqP}~&1EpxpjIJ23@D5kl-&qybs
zQjgr{3c~KW^P&7s?Dpo)#N}vTjbbrvtN#5wB~(-i0~E$NvAG$7O0g%S7{ep5NRwzo
zymQF(co*{?MRi7|BWvjF3%Wj2QcN^MczkQx%2+k6iGFCxu~;3DWxa-=+8S{`<D3SY
zd%oxyjOhG?Y0Z(7+pX>a4u7hxhZnuzbS=HSM`_;dp?*>_K@ZJ)c}Ku2sZ-(R!EQ*J
zugR}qysHs>2x=H7k_bC<wG|ic4B>&uk8WIc9`@O2m1o4Q)*Ohr=k!6B^K8x(j=(IR
zq7U(x-67(e-}h1|i!putpx_<R`8SV^@`jW%LR){eqHr<2YCXK~E>IMpzyPtA*Twpv
zaZcmjXvsl+EkjP>aPq;dj=+t?1-l>iZP%WzU)F)XDf12H(D&qQ`v1L~BihKxoO<1Q
z02(D1P$QH_&`BD&ErHBw>O@R?A3%?KsZP%Qc;5}aGEaZyXgPAyqi3kONEUTh>NSrt
z8P`@=tTEb2chew1qxUEdVGyU$4JGq<y19`^;TGgiZ0Tx0>hoBOd+?=@IoTs)V@Igh
z5y`f^=&=HW_~*9-d>`>&b%*nC?Z2I%mwi3N`kX`LW6!u#y`s+B-M<0RK>?ob9E@9`
zSV+952Mm(4e|nm`r!)6h8ockmFz754b>kKgQ(rL9j%E4KLt>Re^^5iN{VU?-59J%2
zKdTf%NE_W4adyaZ@Hz*h$r)KgLPd5px9ehK{kel#UMr~9kPN^aci(rD9ISfpMmem|
z?J8no^KzJuV`B6uett0AfVNKm=|?n9MK*Onr_12lN$-YS3w{nM*F5Q>e3*Jy2tr^I
z)6j>izOSfNA24WTNkLKCmq!u*el|#3`9YP*i--QwJnyLrDbBuMw8vNz$`9s%^&)6Z
zwV4jy#L>UR{{87sSx1Gr=J<Nb9x2@cBD3l6tuPdIXIAMB5vFoIl`zL;*WQyeN6)vd
z25u%8#{OvMuT|+=3f0w3H406x%*Y~*Fa^U}4vx6PWnY6Oi)DSzrw)D*7iimqp+A>&
zzs@r53fvtTiiPiS#?Qmq7o*?XGWd58e7I^2=`Y&jivPYZSq5+1Vv*bvmwA0OaQx;i
zC1!|s{5h-MH_GLv&GFcIm>AvSBNO#hyk9b1e4fOrA0IQHePj65{dpoJloh;GRHV_W
zYB(X<*JmMlovcOQl>6N3h6zDw^4)K(6}UTmS`(IDMK&vGYdmifR8*H1X(~|v+{btn
z0DJz>?FzqQaeUFbOlNmcnjOi0$IL#SmgQ5Vd7y;=LU@^KS1OZ?L2mhzT+nk&NA-uk
zq=rMmg_^+AyatI?cO2CZG)d4yF|*^0Ez!4B7sT{eA8_YomSdG@pVQuWZq*Yd|2WQ3
zQ;NN5L4HR7lIr7$N>e#uAmW)*QldaHd;tKwUr0z8L{t6W@+J6a*5u9T9HIe$D!Lf3
z1(!j5&^wK*YXImAfc;=<Xng^(&HDLTbQ0y6$$<PT`W}x7P{&4YdI%)RH2z*E7Isg#
zusw{r*qN*MJx&mrGt7|~8ARb=;0@Z5ox$Q7N8{`ppu*CPEm#->EU1HY?93bzda}mu
zl>tX)=yS$8Qu(*t1?!Hk4EIAk8NXG*!$rA{_tnVbh{A#jJ|5knedzjNk4cgXJ@umZ
zhodY)``x$=2tAt?n6P+P4&eet$?uhJlsO%&d%DQCgFm@9s%_*9$=A@?GQWQCkYYn5
z=x5V3BD(cnnZN+GCPT-c+T16s9He)jAb$L5A5i@>|7TrZ#gZ)I_1g(m0z}xmMyQ00
z0wZ<*+7L<i9g~pjwWR_**S;OL-H+liKSL!3?_}=^jdGrS6QI{1{q|lAO>z{hVw`3n
zKPp?r=uGPSpectKg?IK0`o>p{ctg!TyDNqv^i}pwfXU?aH28HU1?+z0PwKMc&J>n`
zPD9@b5$4?bxL@{-YHmx`S??&bACY~RG3yYAeyV1rRE0vH8!o)<Ouh5#4b1Dr@&hM(
zj2vh4CpW6g6lx&^TRr$2duY^b9j{)Tx0vojHfOWmki-6S7CPj|N52o_dCl6oqT{{f
zG8l@Ts#>P#XO$r3XMX01KKZ^b77csKi~}(PaF#)W8(OAvQZw&E(vY8=bXyaGq4wRJ
zXIKx(o4KTUnu@AEK3e4MX<2U?IEz4D1Uue9&DZ7Hk^A(nY51gI29fpnLn7BKc8E@E
zAY3pA^z#ydw0VfvhsjN!-hhfYDl%?ckhxA>JN03<;=Mz^`1lHuoJ%;zn15dr>#~*s
zZ-<oPwI!5jf&s2Xf3|G+dMHy=jOwp<1Lm1vW}u4`Usf_GmJS4&A<tjD0GU7ipJ!I<
zpJ&!+T4Ph>nV?-d3bCjgQWo5pwAe@It0!^3f*bbz`-C#N?_EN|QNud%EW@{Fu5a=G
zmcr*QJ-V)%+i|&$Y$^}iww3!+8O?p13hnG8@`IN+FllGwwk{u381Cf$;^K(8+s_q5
zIr}v;^I_S6;biplw<7dRex^B$nxO%>9Ft_^#thN}g%7`Rl2irq#zwzY7&keEa&9o;
zKqVHY2<|<K(9TwH$!|(b%0$L6w;#=;SFYUGx3pCfW3J^?-}neT{5DU*naSlj5aMU~
zQsvpE{hV78bu*W*g;$IB&4FN5o0q{)+XXk|$y|A^*+QMC_+YME$EwWw?}g0)<;;z$
zYMpdf8Nr=;_qd>D;dh;V5wWRzWm>}2;I+t4UQ3`)wJ`c5wwIOL#nWel`pI5L-PN!`
zsCoKNvyN9~KC^CzoEegsDfx%L&By33R`<$l_51m9s8%D~#<)<n6;ByFqqphMVFHXs
zP&&XXcBMzLZQfQcMA9!~8UNv8W!W?{_$=18tjnP5a+>txks6hq{o&)skGECS*L7`d
zO5i|iv@r|U*=aOQP#G9<{9@Bu7%cjq|KHi4HhZdGH{9c1ae)3Amzc;53^}Iz8zACd
zgKak$``^G-*)=v~Qb|;iNxF{3<3f&m=af0wFZe=Z{(P%Z((sK9ClSo*e9doLsu$ag
z>|||-DCqvDqCfnx&#$l<zez^@Na}wjaly0*3%rZ04|B-&MplgQjLMbyf6u^Q=og}w
zTv*@mmh_cTmGSvP9YeR%Wju1Yq1(i<j(b3$C5gm8CMX}OQ|`K2C#vJnK5bl}Wi(Is
zX|(G)hDox}%-7FKyi*$gKRlg<Lsa3`^$nDiQc7}$Zb48=a%e%iLmEV+V@T=lMnbxW
z5RgWs2L#Cx7<%ZEA&2~q?|1M06Xwi0&$IX1zqK~qD1I6lc++G-_VK}&<@H53q1WEi
zN~zBIJ>kk<p;_=|;`0UEy{;}fFTQwM1ZoZvUAM1X=W0Ytp_;+qU}5fd*T<}EIoM#w
z@3E)15MIpl%k*!c_pUhyta>UBL(oV&mnL|#c1}CcOqVeFY7>p9t?^YjKBPFea-hO~
za>zJ;`|sf=gq81(*7>1<C1gWi96QVB6@)C2z?Kh2z=WEjQ*AH=Or=>J=9{M1{`2(B
zuHRrSxe!>02B!~l*0gJY!C*HN;exftM08w$sk=)6ZqXM-#<7lvkKY0K$M~ZIB#NA5
ziUtJFn_{}~H_!&{UUi@HfB?fw8pjyT-GRXsJ_&E1FLHNMDn(~2;o3h!TMi>Xm6j+q
zzrI`TFB=<5)cL;1VS4Z_zTtF($}l@If9d5;^%abEk9wXE&_xio=KlEejrOcS6zHS?
z`OwPwH{BYOi56EIT&xF+|NZ98u4yw8UN$Lb3EH#x?A3+K`TS7k?I=nVA+TEK6%^D4
zjJI5X(%DJ(I>v!);}i`A6vU3|Q%#_A)R3?UqZ#)Pq;a;Dlzz3ME5^cOG?t7sB6D&-
zXp^Qg`*Lw3**_gaQ$D|cd7kcAqI#1PBExYBJ^~)+xSz%Ra@%3&Bxa3LrxRh~Gla{U
z&G1efQv`67^6{ol-H$ld|I<zNT=?*TY(9S91C)M|cQnWJ)kZBErXrb~k!-@`-tQIG
z5T*guGT)vjnd(=&s`8aSbJaPT%^2cZ$i`_Aa2&qiTvoKdXn6Bo$1R)4HT(WpxrD^?
zDb+lXdrcQ?9t50<B<Y9X@a#X<g)-T|Yxs3@y3JKdx9aznP*`=E`+f;N^F3X&FS0Bb
zmWk2zb0X{`Mj&mXM-*;lVrN3^^nN*V2F4qmx)0qKTNs{n$s>#df=7&N)#^Db-0~j{
zdggIicL&i`Ne@?iulhYK&BT0QVDz%EN}59UKDx>clNlWsN~tFn*e9T1lWGAk_MP~Q
z7D|_;ys#K2A=hWfsh=IICV+i@tD@^Y|3jL*)hoJ^`V^Jb?W&jAqO<wUqF8$ecf?Ab
z4%YK(vreTW=y=Kb=6MbU(Z$JkspIC`0;48jxqxFIti9xewbF-UkNtl6hWqDfs`lKT
ze-FT{Q$PWWtsBp;PQNSoxI8`g`GRNO+iW5A&7?K}k+J;o`F<ASe^ot&;Z*mZ1JKX4
zPzsZT=4K(kZc85E1wUxr<-y!u9lGOia^l6&2x}Z<)G6rVTeY(5SbwCVE#}NYKbU70
zm(vwlm(9Uc84O^ZJo=JutYeO45_5y;Zla*eyG{}l-<EZB?0GA`kMC$dA5PD6rz(K_
zgma$S(sSYV-T5@z0;w7B#Rr~kjzhx4Q<+!)>qC=b;tGcWG0#Y`jsfj6{t?|0&W+ZS
znY#P>Iy(9I3`LK@+Qf<RyDmn$1@4$cb)&l2VTrD?x4XXG>fysROYxKZylsad;rAD~
zB8Sg8{mAX9z;=Ob`g$5A>rYy?4RWv#uGA@p>OE+U7d{l1w{674>kI`UbjROw4~OqI
z-g%}NVK5Hpo8^b=U;GFYCQZGi&5w`QSMH4Rj~g(R_a>vGl{nnA^>UB)&>VKUI$S*x
z{+e?<CT#C=-O{*h7o?OH1h*M^1RckvsRG9JHBvJ35z4mX8c{k}-`i+CQudsuRa$>#
zyG3w^=;NQ$7LJivHpGfci}QV_;oPpbeatzp)4I#BWH2DB+}MV#4-@AuU_8Bod!|6|
z>8oT+d)r|<7^_R=vfc2iROw{zb$(ZbNwl!i*jQ;DFf$;oF^@^&$v$<3+KZkicG^68
z@6VVJIHHT_eS;sZ#EHDYo%G4lG-3Z`v~+A3m|b|y!Xx3Z;#GZxqAJW2(m}yDg(`pW
zc&rz-Gzqw$${<0fdXbc~<n4zbkxjZW&=`yRwCln{DT&$#PRi$3S_Z)%t0rA+@qZ>}
zqV=}FS5zqZdmnN`_rk_oi3!JhuOBbEO{$<Vn|aFNgtVw`Z2A;nz@T;YU$Hl)i{f8H
z`Oew7j)00E;(v>=nT2lVtmlW*B?0DnZ4-ZtmGQONT!URbfNl>)9Lp&l2VTL~T+<a@
zEQ?f$yV|Ry)7sK4Ab6b6R!>u7oO{AQSb66ao?99h_<(@YE{Nizi8oZ;CQssAwYmVj
z*rA}?`rXM>Aqfmd;tQ{aDeeX*PcnMBnFo0KE#p%n@je}=qMnxjXr!jYj!W(=zfP5o
zYds3M-nnp>X~sgxo-$iItzh*&Oe*iEeV=!ILZ<N;6Zl>-uBf?FE~~nt!g(|4HQPwR
z&mOIW-vJS2*-=Vl7RWE)bP9b4{U#v9;_C?f;;5X$Yl?m{m|dXZ`RjcO2|_C4LS370
z@rD&0<z|_%sEY`*I-trIJ5${BwH!-Iaxr!bVr$BVbDSGFckOChgN}v5*?3SSV}?zY
zj=o;ac;sLhIeiDwLWx`ZshQ%=w`jQ_klWz91-9UZ=#o2d`rA8w)i+y07;sBVn@%&F
z?lR`qEOjo<bVgaD)7~?45cy`OsLC)x?0lAbqM||X?9GlFmEbxnrpr!ZXt75%*{bg7
zd9kb^)T8|M{2)V-M%x5w%he`tk(FYiQPV&me;t%F_W4@N`fQcQW^b#<^rqDi+MXvN
zE4%8wj2KwMP+@QI?!y;i7bJBN>A0y;`t6E1kBi^nIbMFs=Be-!Hx`{=3=&19<O(HD
zMIAn)uSDprUuNHy#<L~!6My*V%1fLkhVNirc&d+@m2z{I@Y+alN8-av34tS*6V6#|
z;oERFM0$Sp`xVRA%?S_FuXXWiGLOLq^8>Q>dED&f2{U!--n>dkzx)8=&nFaJyii!D
zq=YRhZ(hDCwWXc;Kse5G|7&cDEcGV3z4iJ!{J~t`arC{%m*KLqw%G3W`*Rt)xixs{
zHQy@79G|idq77@33XmVS=S$<;0nNh&sJC~(LQcRDI}9M9cMiIo>i_8$<t%>w(+f3s
zP2La0)7Qvof2*%Y{jrEo<+JGDsx$o5B1`9YqCPY>mTuM)HZ)r}n(qJ5LzgPVmtoGb
z*hewmjkh@Tpf4-kQ<83khfCY;r!p^;QapXtWH3@PGuxDl8;S3GoWbT7xY^h4>Cztc
zykIBcK5AW-UNKxudV~RU@e_3E#V*gl5I7hevdo;*H3Kt}*a@R>8Xe}UoVtdnk(7px
zFvt}9B*w?$B#WGbRYeANcXqY3m?2@qv}=c{@f4qo8U*82Qu*Y&$}|{;8*$k~D3>XM
zM4r|`ibxWJ8y<mvG;R&QO{Z5;qz`Rt&p$wdzpJbf%SYgKab1#Qf(4`soZp}f&D_->
zpghS!IP%p4Y?7o+I3s-l+|L#YsRJPs7!}xe0xk1Z8|eSA>_l|R7&$o<vF8kQNH344
zB00Z)T)xykn9F{5O*V8=6Q$8Kbf^lUtXH(87<-H%u|m@4VlVqMGJ=-*f1mxVkw`3%
z|AY{*bw8FN-I~^ACUh`sW<IB7pLW2ILN6x~>ob0YiRU7xTA8wbE!bl6C-YJDcz&Y^
znn>2my^OrBR6lFgMC>>T;{x64q{<~<lLK1+))D&!)xz=wDL}_*Y&sB@c{QG)N>=Q$
z=NkW^qd)tVh7L1J<pl^#kf}5uK`Em;QIjgH?{fR&HmeI198b%`c-U%cMUfNs-tu(D
z79rPZ*bwU?Jm}i_b3ysCQO38y|24@6-*mSvnDVA5zOU||f<;LT1PPjIh=aWiv7eL-
zDRHJ#houZbaT<z?$qZP=q9ssM@v0~~r?Ch+6qIo%j!j`ywv2x-t&M8^H0SEKPmmf!
zTqBPFQadbr1HJV+(=hV$$~`qoC$bBtfk=~%%XewmjIJ-r|3x-mXXowYI$y@Sea-#-
zb3q2ob7QZb*EgItBtIt3T+{hC=t_KMID&+ciG@YZ1iDV9zY$04I>Oh=Mu;!=pOH@3
z(b#+GQQK#KlZJj0zD<b#+IEc|*6#1tx5+6!9+s7t-;LhRw8>>(Xmy{o5RZ7;i8F{B
zau3aPF*l6Vn)W7e4u<?SErYVN0&9?&d1wWHi1YWRzrzs+J9$NxWYe7gj7yLt9^vGf
zTyaz>EU?tpPOV~-Pk`9DkuBa2=xKY+yA9dn<}Cl6?hjEWbFZpS-#B2V&6Ld1`WqU)
zL!&p9svy9m?{Q$Sp?1QlGRwf1$UaeiBlvu_L|q2<074IvEu+?W5GYHp#^*1RKpBC*
zGtx*x)j*3Ojz&7=op6I=1Bg{RWAcsbV7{hdPTX(^+h3}ubg-moU*uEpq1J8l8Y-W@
z85;-xGF|)u8ayc!U-o+Av5ZMk+a=@nAL}jivf&=*9r+w8JjI*Dw@7*>_Wk9|ud!3*
z`V2}nH5lXP?kw=g_7mZOxqjJ7tW^xiWo#@Dg}yhCU>a|77Qlb-YuVT7@5Ke~M4hpy
zUz||52s*pbK+JRw32j5|*QGzp^Mg%!)u)PpTq3eppSD2mNEJ&QG6XfV7CoX<OhgUf
z$zK#Eqwn5ToKrEaZ^ljeml8qc64WB-c;tL4N>tJZscR?{RZ=g3>(wuw&YOD&(&hBz
zfTH7<IFNDv@|}dSiJm<l_C^-4C3Vws`Kwe`p060<ph4n?-?Gr}PkQoA(+L9-i%vd1
zAvX4|1ty43q8C@!w<?+Zvf&3~xo-?b+;@tOZGlvmHuR#;p`w977NOy~=kNd4p8<qV
zUE~S_r?Dc#ENsauYZrZwe{rl;R90GT{mw^E>zZ!2D~<@u3mAKQw|ZKutulPoYH#){
z=#DVD55X$9CmbxmAX54xbx2Jai^}D}Y}i*RBfht!nA+wdqds{h%aFIcT;uk{WG|PB
zpLX|PB8Wp2)Ky|FUvN!0xYB9F)>!|@8S~X2WQGwzJV9#_{)IH)G`5GheZ)!y0YNTo
zLPgb0e;vx%zLaf#ir5-t;jg_?;TRKW9ZStIR<E8QW|!?HKb1j~Oh0QEH?LJKOd>Xh
z;E)u{7)LCEy<;#xY(?Es$>>aZ{J;%)2Jkj<zzGGhUB9mhgaup{A%v%zK~Iww@%yMS
z@UuYJDamYjk^Op;U%5h#opKgl%ljfy_jvbBTS!K*0`~@Q;bLcs>_uveNMhbfLn04t
z%RtBa?0dU=JI2NXIuUZ0kJhQj)Yal`4Hw)$Z{j&e2(7WHJ3PU7@9E9@<VhJ|oBqD2
z{1R4FqI6*q-{J<>XN1j!2hica4r#IJus+#Z#;b(IJn4p4qSn2uemtM}R_bZFRx2cX
zb++HpAy}*TNMrQEN3Y&8G8;fKFMS3ak-(xv&b9v7WWZksiD%p-cKKh={0-<npOf)+
zUKEHO^*%-K)^sV=m~^0k9T%y=MB;k<FJD?M`2k&k>n5D1Z4X??)AH7U@Q)_g++4eW
zPgsn2KV(nR7SvB{Hp^4pnzr(Og~Q1$x4u@*xH^@W2ygqlwrstoM!P64#F`>iNg5Ke
zVS~fG8o>q*Ns>T3b4@x`h|$FS{4E?7__#1W=y()_apEE9vn1q(g);?PdeUr%(-eJ#
zMgIA)4qq05DD!<JkqQ%Q^SO5CXZNYdZ3A7FoTXibBlFdbAx-M!K{2~i&@R&FE$eX)
zNt0UYi9;fI+(139L&}!a=tL2Q>U&ahT^gp4EP4mB>}&BqcPCGvpY6w0C8XmxY8Qwn
zNNl?u9EG+Y8>J3+UJ9){FzlMbTT#=0>MX4OE_W}Qs8vjxc;@(rEqSe357mV|fStv@
zb_<=3O<max{?$RpaGcBhz<u>_{cwk^rKT&bcBSWan~}w&S*jg4G|_2xcJ`7Utf8s7
zUj2R1C0+(b<B4>30b*=-9M?7=sQ)Sgb*DhLEg<r#{YcK(U67ae9ih<jJ-_?=v)@Wb
zUwTfd%WwO-(4No9KajsQByBxSNy9zTFw!hcG=VsSdL}Fi1S+Bw8k_3IMQD<7SCYYu
zS{|T|jN_=5HE>S`E4NUl#EU6n#Sn&`3kI0bC{%VDvKU(S+1%@^M7fwb-ilnWQEZ!d
z&znFOVm*lQ5pD#>VljzsZ}^S)D2GDNO29u^@$Fz%!92A&V#5!$o$u-6({DNi=tsN&
z3!L?x0tEz_r0>GoM(6RB1etWVI<G!7%$3v|{2+Of{-tvLqgMZ$>)NSU+RFI=oqp$>
zV!zA5Ojx#KC#GMN)bix3l#9dp0jWpNljnXm*Me1gb<%MXw_dU2)^ze4kzd69S_!pf
z@7<ku-KDyYy+I5+{O4}#mx#m_E*w|a*DsgRmiVWXMn@+DajW`raWw2lM@MKNrEhS%
zG!KS(OaraTz6ziiS`<B>#u3m#Sgv|*c*(?o+9^Lp%uHZ@Az8<b2=ry6>fVJ=Q-6GV
zH2uOHU%#I`fQG2#_9i1?{_H(W2Fpk$BJK+KrGhR)>CU%m(<*%w77BuPGkcyrI*9N9
z?mwhdo6$Y)jSgz_P3T{%eT%tu>=#{BkXZhZFH00))kw*8N^tJoi{<F^f&wsY+?@#j
zD>0+ZaibQOVwk{_Lb=!pvNCS|(f^#iO;2XXj4UYU1p#Qx1>kIUeVvEXdDG)1ZuP|P
zu3h~|7J#*wT9iM=4m;UjN_gx3)Y;fd7&lRL7CDm8h|4eLs{J8m4uNfB-5wB8Vc&*a
z668bRNA(1qF%N0QsgTzLmNb@WmQX`#mNz`V=J{_wP}cACB#ed;uWaDg|B8&6gszG4
zg~eDjbGqywra1rj$!`D65<6WIM}QvkhNNmxFwFa5<<pmUAkJ4~99+{~hpu?2j&)(Q
z`wR!v&a4C}Y9EOdwy3+_M9jMMOg%!6T*piwD*>^JTaL)7a?RqB5;;IjWqfHfno^qQ
zZxH{K%6BD{VyEf9G|K#AfJTwSP0o}!ozhH7NchT0NGiy0_7h7*0wB>*?Z=~&zMc#2
z>7z9%$;pLRVMm_JMB9*lj5MKQ_%eFanX%r%6^p7z^`?*V&*Rt>L!u{EG@BovP+>RN
zsKwJQnh?)5oxPapzlZO3QTrMDYGiXz&9lHNJw;~_xyqjtj~+luQ+;VvTGY?tem6Fj
z@S5Ca_~u(Wd_8$%$US*;m`Y+affIRls!2Y0v{nMIo}_?VxTX*Pnk0vpW<t@j8%mj)
zdKGtp_Y&o&J{7-ivCz}YXvb`P;fJCvX)E9buSP)f4NgR!yo6$ZB>^(AQFh=zr$PU#
zkSF?r&$+d=wM8JQ!l|DJbAaq?2#|5j`hSwo1vzK^&tcv!uAZ<blV;VsLFXomRyT;*
zFxLi$-4nl05#JllM)y6Vvggk&9mg&>&BS?`$T$-Sp9|QY`nu#Dc6=xJH1qLWc*QHD
z=6v$}^e=7Qn7QZh=GZ58#uU5B&XSk-^a0o5eYBKG!#~NyaX4ea?n{gi^6}!^*HHo7
znC;sFIvm~&b*3E9+Q(yB%Tak=N$pJo%RyoMq}mU>uHy9*JgyM_c*fZX4huqg*h)==
z@+``cl&suUQK!XX1fyVJMuyS<TP)Xgt^)3{GI4IH)XaeewGTSz&P*$Ql5OO<kS5rV
zisE30RXDHLepW7Sh=_A3KSlegcJ9pxC^CN1)zMKdy`Ze%b|x`P+JODS04=C{p9#4Z
z=tFvHbY#EP^a|d)UOt3X50sfI7yerd3z-0t`Bx_|2Hg2MFy0-lm|0fua{gEBXrcwk
zw%DwyDz4ZR>NnmE2Up2iPub4_RtkP|zEqPJ85t>ZgrW1<10^OVW+7~>tWvo+E2q+-
zCb!9jU^fB-Mfcdw1lHzsyQ2HNgYC79Bw5Yr0bwQDZ4~Y%U)O&2G}45%)^9z84$?m(
zjpzg}b6;xxcPW{>A3ENJ^&~A>nch5Dbb_^pT@?Jj(@0($!x1)(prp2Oy@Fe%s~w7i
zQF?>jrp~7+O%ChZ#vX~a)Q%~?bKjhbqc_9(Ym|Xua~_X7jKI?93IC`5nCYH(e)*wa
z@thxRkAxFg<p{s@I`}FDE*#Ze>G`JOFKN#AjQXgL#@hKHGuGv5c3qoP!ZEIQ(XCKx
zuOt3hjw|enZffUcc1D*{L7vy!pT}s+ezP9S?`Gk4`r;2g<`n6ki>m5q&uo2*<1*CN
zub$u(BrfM9t#osBAF3{!n1*6_h@X%<%^T4RKN_t6m>v$LtK8jX8lH{@ZA_)(G+L`C
zw(8;C^qoU=MLm`bXRdy^TQQ(0<5W!!=m+$EWn2yZKC#ukMNi;4=*G1XlE#a#W;Md%
zQhu3pb9DwWS>zoGVOmT!FDR=;WtQv=>p0sSK1wnhYJOj06*j?Jz+As-a5?31J>7tT
z=RVJWjgR28moeL$o}NaZ%{vYQdGCx*SyiX*{`@$ksC@hIj&BtxW2Zg-cNIIB0S(hh
zEqs|VUJMMJw#^4ifMvx6m>usJxzhVz<jbYK$ZiBu2+#o6yNz^xV){N{h;u66a!}xI
zSo)c2nvXDRR>^mt7Hip=Q0DasewMG}yp>-1p1kuTD)MfGOZD}0U;kMUoHchRi0wNG
z6r%{z1<vb7;nhai-NfrK%{`Md5?qA`SPxGnC&;v>xA4Ut+=H&x5K}*+QJ)YK(I6*J
zAHvW(7@SN-EZ(IBi_n&?JMXXVJf|h4SZM_#r-NUC`4fn6VLTrh1)5)xNR9+*uW^hf
zh%EaCjN^}SE3{P7JDBH+qwo+)9{X~wTFq#So*r$t%kg(c^c{mpzfLhl=NtJvEdacT
zYR2aIk`^i^{hmQ&u&3`!>ow-UpTUPMy!AyudX%y!-6|89Nvkkf=zC@xc1x-@M{zT&
zEG)0wH{_~tOmWQ!eUbRK(nw^pUWN8LKLkhrfESVPn@f!%;>U{<-L`C54GdrZ=`T8T
zuAHtkpFxQ^`{JB5pPzj8a$GOMxR-4I;em6;JnDJl>@5FD+Bko})(UBEI_JLsjbbMi
zG~cq}u*h!={zHMYNxux0VMCWX6HQ?&ta)|7VM^oIF6VJHqQP?mSrB5aEYLUxC45`x
zD5kZ;-yBvsY-d{YIZF*cqc5qC8mu;<QbtAbjVY^2Cm9LuEhX)XvzmCvp;U6eT1i$m
z3n_Y^zTD0xuH=uh2|bk3SIP152YB0;4aqjayOY)jQbt8>>9B*#_7?w*MdtvhwNSYC
zQOJ?;MBBD_P(1B9f?}%s<V7xLI=-37#zCy|j7?6x5bJ`U=}>&tnlM-Krvsr02Px+k
z|3feF=GQ0Oi4J0ud*Lm+&Zb_EforeBCbWypD)sV4e(K2fZ<VL}jbxiJefJ-hS6|22
z5Ug+dZ*-gpmr6aH`Q|?5-kTeBg$V0&@(YT@;h0M9q3J>J;o)^)+u#IQ$<_JcUHIML
z`~MYC@&5`a%}o>kfa@c9GDje%RDJdg&RGKx)imJZ;^M5Z+ApsibL9r>Caf5U|Be!(
zci~_^VL4R>ybNDjsny$3-pfG(h7XU+0|lgs8ORl6!Wfnlup#pdABb%sE5v>XW{_Oy
z4`czqBIKf3*Ku^sD!YuH#ZMm0z?Z;i{M3S}t6r`IMKSzI<*8tWje%KtlqlANP1|Sc
zc5zx8pXT7@ibL2+MnCT#7-&!MiOUUVq?v->B(G{EjqvhJng!qp?1V8$J!ha#H2L!Y
z{eplf|J8KPJP(h-L?L5iPQr8A6#A1VZ^CGf7@N_?T^>%G6GS)!mek|2!erf@a@N*u
z4n@oo1}}|$+E{xlnDq6GU^i(S@XbYrF=7gw!e^*-PS>X4B|$zNtH<-@FcWTLe+Y+_
zlZ4P~cCd7@LD>Ak2AUDP*>*M2-x%uc6o4suf@IwpzLBqDaJfekh0IqyMz$b5FL{93
zQ&!7v=j-OYja{!n5|<nCt05!>UHTX6UlR?vo?<MD9iC}=@6MQ7y!qM_**HQ7T|O9?
zMed~~F(BG81p=Hr33l5KR)g8)F6$($LqV_hPoZnVj@9O5nOZEo0Y@cn#&kw+&2=h0
zTz2aiPfX;(!tB1-VoPUk2{dgJ9gv-e_)MCQKM>wsXSLGl0*=<-h0M&vG0#-8By>l;
zwiWd!nu6Qi`=JN=ajCPi#psK2*1-u}iQQ8M;f_>*NuZbj@yO-f$-<Ue<WHDy4v46=
z?RGCYYwDFM7Sh9Tv<l`DqONzN+E2LWN2dEddnu|T%zC+l5;w8zhR#242JJRP#%e?k
zc({6I+YoOLxsJS;wl>j^&_(x#VE5^&_~rNqK*=d*l)9ej=-O-1sE1M=KxO-#Z{B{d
zMo&)%cFH3`6I3%%&FlL4G;@kmT=icLpKmh@5=6_gg>Ct5U>tV+!fo-#BlgW7&e4T^
z^S}DN;BvU&sa)Mg=i1B6%)D(*2q;|1z+z^@QK2=0$k^EF+t>dYbz02-6M|)aaD<GQ
z_Wg#{bY3cHXlS_ch2MYpcyOXvdA&H@WP%JglE;ikIIISSZ&Fp~6v33LbU?&6O<eQ<
zjpOCKi7z)rkQ5R@-*iLawk%7N`MG^1H0P-sHE0FRDkrxU?X3z;?Tlc2>6N*eOyP<3
z)SW>(mOt@<f&{o(iQ^cR=E(DV_ah)<jU|b1UzO2RdH}!;Ts{KCPHY2tpl%vq4n^HR
zl2;&n-<;1&lgOd1M}QFWj+l}2@eBRy)VE#HGz{l(Yi3xzk6YK=AT81>Z=VP7-)gly
z-Gh)hRnbUP0X>O;-FPBTK!ltL22;*!9_v!}*d^4}?x9PJ>+pg{N}Dc3RW{t&G<dUF
z=J}AEU$Y}bWP+DN!Te#j&j$RR^LgYhBT;0CdFKlV0ck{ivJn+hCIr8Sg2W#3Hg~?b
zu`+dV*&=*+cz=GVDzASy!Ff$!zc-pU*@CAX(x+{roeF<3S{&Unz#jOOcn1T=PZDs9
zLwW3RZA;v-J<>Jh{ASoVJ@_l9y=4FoNk8`Xi>PBV%kkz14C>BoO155EDd??%c99h*
zjaYvw$Aic9{PUI6323!yKPl7&jUPW8lH-x~ok&8pe&Ji+H{afCDwh*%3(Jq^;^^K4
zMy<W^JLK**(E@k$qQ7j-w(B26*LKII=;t5OdZ5B>#)38P*O7=rvL5Ng79?X4my8Vl
z0u*Tj<|6rxqcJz&8ir-0rBO15ebaL+4`IJT%H*0Zi!+=hI2hvRHmY#P1}sc;MR<?O
ze4N~Dz%3uBd9<42!1N(RRIZ5b>O?khKG#sMV$5<fZvaj-Cc%!sF*+?v1MtvQ9KQtR
zVxKeroBu)}SdZ%d*ZI%^oeynTP|yPmK&xrP0n?1*M5DbfW~(sXO<O^S)fWH)v`5G&
zD24$Ym-*iW-_zO&?8WqeM7W)j(KA6if~r%heTgx|i?5~vU-fx>h9sCF$=5FY<9ljc
z>1RZy;sA#%o$}5frB)@{U;7HY8u4n49=3Z23ie90X{JZWsWC_qTj20Z+@X^`Vi2N?
zCv(&mC%NhH->NbV;*bZ@fJ%oY-E(3fLNjGwYKlD(F(>uAPUVi+IV1t{3abxa5U+`r
zD)2pnggFak_p=>zOQ*}Wf&jcE-j?EL=<!o6HJwT|sT9zQ1dAbcST4*7i#92kjh~5h
zU6Z09*|%jNDt(Z)De}E}96uH*0%gJ{K?yDeu*<57ilhP#@<;EzJ~-(y=ZRq>UBkG+
zt6PVoFkg6y-{3R2+iahp@*^5Q$p8igk{kz%!u8K5&suc}#V?4V5-bc|B?8GEx!%%~
zNrp{H)LqYcGp)MIq;sFZrNqXuPsVP)mBQ=jJIe=Fx}T0F<V_mbSu!O0L2kzbl$AJ8
zokXUo;+e{VODJ*&R9NA=@lHay(nsrTBngm6lG*-j>*u&T6HqzNv7rQIlK>qn<Q?G~
zxn0RTy0b;<2}d{tbD8nf(z@$*PL+v$>|OFt-$b8J$%S|46()$yO1qcLjoS7d@!JO?
zNX}f`gLJ}ou{e`g(m5F~3Crb7I-ep-!+sO_VpG2q)AaT9&5!+o!nWBt*+&1!@f;L&
zY=`xzCprV+8`<EVaH4d8?oj#u<Uj4r9}hrX+EcT%wA_MM5}4*l1auP62&pk{{`~n9
z1<;YHHn~u_TR>mgRzCPg76Yq{g9#O{KJ+PJUWuB963Ebqvi6oSyl?Cy4^*&IEz-4X
zQQ{~fGJm4`BYike{)v`{tQ-KX2W#+rbQd>iAR&No@L{FsI>V;pv?&;W3I?%>gWobP
zr=@GR?ncXbgS-o*gIz+0b~7d7d?WIa1ugv0@rT~~v31I&cgYRSJkR(0vi1EJNX#K=
zBp3>E(h7tWq1v6EZh$52fF1U}3%!QEVU3J$qGfjc7MgKVsH;pw$R=Gbsm0aK9{R+!
zdm4Ka?rQa7B(<d>!HVD^7Sm`!5?(qv3#5ZYR$+<Jfs#krk}?QeUrs`k;(*FLGL@G^
zWh5Ycxgam{L1B)v@sb9`+~^Bwtmp0}Tm=^i0g#GQr}b|tnF>rO8V1!raxviV%@Rex
zKmJV(&^%9>NE2=^(Y-xIVw{?f0i=g0&#B%>WN(T82DZz*_Km2ay_-mkasP@)%}n-7
zVlK`*ax7aDBZfk{X5-^-WyaA|G;G00!X+RhnOAzg_hu;osXMXuEndJBn*0p2Gxw!&
z@fvhs`_ivK-5&NX(FG|Vwy!hQ;v-;DQnctFJw@=Gsd7|IK94hZe2^N32nBtpX*2oh
zeL8U!myI%ICE+~&hcd0Vv1_6}pD!WouYR*i2|9el<-&4Vi+%0wV=r{E9DMq^(zMH;
zoYv+)rQy8LzauCPkbrk+T)5B)$;pg>pD{%>Y7|Ta83C>TAD9>)pKj>bp*>0`tnp~<
z<?%~xC6ZsEG-eli=3W_cbs(7tOBqs-(S;7500}7c`u!hz(+0qj5c|SBj&H-d(lt^T
z_Hlzuh9ePsXr0$UvX>;$1*M1&DSt2RhCzThaEAhGK|vg^#3+U3ax!E%3!R!qj4xmO
zT4WRTyq3U?{#fHW)fgE&^9rnACb5b?+e2ePP+aLrX1CT9DLq_Wrd1c3V7>*C#+eO$
z0g>jT>a5I)f<P$M9Zw!Y6SGdJ6CDN>3)cDdhbEI{6qq3-h?z1D-N%WBy-18JbX=F<
zwbfR+l7KEx67G<^8~wO;wlYMoXm|kpID+l6axnE>gJI~dnm=CV%Qzp|JMhDnfmW}O
zwoN2;@6#Rw^)fN+lOsHY|7D~iCecwr0DO5-bbyi<;U}Qt8${aL94?ynkL7YZWW7qQ
z!%zRvyWV`*nYtoFG~<1}i%MS4w8@(kFY;61+8J{<zjY1tv2}rD6)Rml+EjKwKbW)5
z#a^W-=k)_wbfvY*(>dGm0Myt|I-j@paVJ>4&kc*Jqc5pm7FnOz;x2fa;)>o~zt3vp
zsaJ7v6^u<aLWH>r-QN6Nsjm@5gp*n2`5XY$^ToH?4<|tWA~UBhY|5su^j)5~h$E7{
zG#l*f>}*o;zpIJk-+Pf1`INPS#_JZd>28K?*$v<z%tJ#%=$*=%ow~!$E7f!Tzbk_l
zanoC`3)T}vBgk2-S?LH!q*2tQSS`J3DxF#i^x*Qr;Ahey3|l`a1orr<lGX-akIKTF
zcrfk?rETQv3}wya(nQguG<1<K@L<2AGAzD8k#2<C>32OC)4^YBuN~CG<%v3#JyiFr
z6F)+{L7x+RAPt~fr4Kt=Ut20FYNH6{#Ym~bvhk*y9_gx*MG~Ep`#Y^OOd7pxaJ<(Y
zF<2xKIRegP6KdMV4&z8#qMgYvA5@)t5FtNNPUPB97gr&1l75y<Z?|rWZzdn0<`@DZ
z2{;jUcDN;0n3K*4?A+Aac?<pFA@iItm1%5l;58O=(pA{;So-c8{h7b4UR^h5lMyOm
z>Q{MBq&z@yt=ek&h(O*+$>T+*VE<1XyQ$6y<I?%v*0Z_#1H*GW=})`=NSt9{{KA&!
ze_76hTJFitI5`9yOa*sl-T$4&Ipo#u(k?cc?Y$@}EmeN|HbzTVcQ}Pz(~YnNjdqFU
zH>2b{9$a5v2e!&8TV7ur)w_0ePL#`F$0kTefy^>C--Z{8L^PrM2zqEq6!riq$dU{?
zI7Ux8gb<40rET45!a!cBzXC^EL@-X=*z*-lQfQ?F*RD|_mNWdlF`JvbDRs&N_KUH6
z6-zjZguT0hH^PTy(KFrq$BKrw*oa~il*PPb+Y)(5YF@y8Q*v0$!>;z}c^+OXI}_e`
zQ;M(e*oYe9k@Sewv9UwMBY;eQ3Ag-We`U9vq}$VS%O_5~aeTB(xGT~Lx<)SDURDCb
z#@DTHy#6jvuRakT2m+6I>gnmR0Ryd~l85z`x9O%@A)!mIW8ROb&8s#hl-+$Ao9&lO
z&R;(=?N2?SWtf(yVGK;^eV@hovUl3m4*}~~Bm70-k6Aa6pF=n~r{MFpt6iS$@9tB5
z!SC5O^*xf!*>u^zi}RRJWE}!vD5jN_?iTw0-J)mu$0hcZqLdK{YjL8gcgri`^fO(d
z_!K(@gyD4FWo`#Z{qV3GV3no1vkj$TuvGA4s9{&QO_GG9EoOpHN@u_*d^3-CWj8~#
z3iP~qqYkFf6jsIULHyx(tW8cl_bE%B_)Uke7|It~x#^VYb}6W_P+l3bSi3Rz<lQOb
zxwK=mRBzkjF>m|VJn`UmkNZH`@to95L4dXP*9$v(32736uB(q3dA8aGAis!DQ}5D8
zS)h>#CzFv0mXvkz0o>w4!v&;}jHJ9)yW7*LSo<r$%r<tqE0Yyq{CBMAAZ0&&4AXtG
zv}sB~P}v0pwE}a+i}4QoCFTFAqz)ekE9!cVH@-N+foI3?PYgSFX?FP!BL!A@j|-zB
z&{FXI`e7cC-bR%#ud*Cn|Axz=M}Mj5f;&E^`%0`<g4&m#j)&qb8-=|`t5m~Glk<3C
z^v-?JZ^i*HsVk4VjnVG7A}@SW(**6}kb%o6j&O<4df5B8(&NE4Bp9=vEH4{U8Bbr0
zsO3g&zLk(j^chUx_&(-j^!^XexciUB!9#H!gw%1HYv#K+e}=9A9vP_n*NV!goaqxL
zXx%m@a$m`~S~Q1@85P)Y7d^R&S`j(nmJSiO>=X{`(qNGZXd!g}y;6hnWF_{K7j6g_
zMpMox`%t}lD@;?!jXG!CRS37#WvR?^S<);3yW<?}J&Ly6gyo!Gq!!8#Pb1=HE;YU{
z`Cph2{&AFe^J;ZDc%8jq)5WF8|L^;9QOA@tWi5L5Qey}oKY|YN(Pj4+NB`G@OX{Lp
zSBj2z4S}skF~sUaYBt`vvS2;2;H+0UZ<I^&+`m~;z)0Rb!Bvh)J~@t0-YPs#7SYY+
zO=w;4DeWE0Vf>~lq76S=l&|6Q>3$$=^<&hVChU6GtL3a)p1CpzbH8Vy#eI&o+MVtv
zu>RMw_7Aq-NQb$&c&^U@w@fLNB||<DiZdkRKPT12Kb5&M!G7Sc#+l>u`xO5?)_dmF
z#iCPQ=9~ByjN8#OU7g(Fv2Ub@xmKf3T%pi7$|%(?4f5zn0arxPr{Q6a6B8Fl-4yK)
zOxC-9*s6)f9`y_h#~%-k32sr${k%Dzm!&{g9T^>uxZKV6$qs)mi8Roq-)L;V+^KXQ
z9S*yLTgy)W8+9-eIgnugw|MP*0kWYa3=0CyXMD#`Z~=xM$-TW0<Ho<h2xMxd8k@%H
zvuZ_Wr{W~EJGZ?;A+4Fh19ERQFx_>rK^ZwL%>Q6K`1i^q+Q{I0(oc@~<Cr9g)d_zV
z=rc$&lSMN*(P;$9(h9{v*~+cVBAs%SV}<ASgzn#$^Jn@qro5QyNZu(s{R++;Y@((3
zLmGP!t4O}3QlkEA#VfWK2K|t2ys9Z~VGe0Z+sX;>)Ve2vS>kvAlw)gQ;8q0ViVAsu
zs($&of>9|7=^R3jfOLKh)G+qBA(Perm>=P$OB6Qpm|4Lo)Ii5rHmSDu;eJUcjRl<u
z3V)q3AgfhCM53)-gLJ{8s!gYkbBQef`~iu)UxFYPG9R<hVpN&Q`Q=I`6Gse7MPEft
zM?L(bt~eruxfT2|1<Oddg(nAdn(;KjL;ZW%3*Ii}hdLoQPF-JXmG%*)VQ((m;cv*t
zzE+#&IJb{{;nR)}0CEMAkTw9BjQVyuo1v-4fy?7K|K~HTT^UnW0+R6M9=LuMQ+9Po
z!m6kfu6mG(n`j05q@a1wmo(vWPVh*Utg;H6VGBr~X-Mb0<oXw$3;Zxp%l0o+tXXhz
zlOqLmx$sW0i)<J*<nJ>y+_uNexzMpJZ0vjbNP~a>^&vr`b$lmF^?Z}|QU6Wx5^(~K
z?9SM6!3}E1a4YWE`1fI8CGB~%4Hen&w|Ak?MPlcMQ2xY)Uk_k+EnTp&!iA32S?QDd
zojRqrp5kLM61?WoTc5>k^kMXKeSOX5yUeA)4@RsUTy{U=Z61Tnj<bZ&x{rE5x0IRq
z&a?X7G%xw)VLHDV)QYO)g+DsD{hg+I#Fyo%V-|U0m-H2?pPH%dr+wCX$ydq*jk8N;
z0b&Rdk7SmCElIz%^ChnKUIP2M#|0?)EhBhVoq9>+Gh$x;e{m^hpZKXu*vk7`p#l`K
ztHkX^Aw*t&7{F**2WVVlRFmRTr~%K<sHo@3xcqMK&hCR10iIGxo+}4fXF}-!q+CFX
z!!06ShpJooO#ybf+{xP?9UWr;ZAAu26r!>s(^!kEgR_#DGaoS1l!xDwB!+~A4Uy`y
zSUr8y)X*Sp`Zv-841X*Y24Sp+HEb!f58fdw7s3=BeZoz`ZF$MF6kJcGSkwY6u_o>Z
z2GvWw(x8alaYmf&Dl`!J{kV2U<(nvkeHKk;1EUnHIrExa#U;cCaUgio{qYKG_wrI`
z-&%CGeq|~$#^Hw($`w7@bdiBPxVo-LEa7uH2vCrSDn2A(z)72!GkAlo7~Bks)pjdq
zwM#jaHvk>lVr#Inb4V*0Q?D&_NN=h0)F)nZ!YhjuEt}7DSxA5TO;8}8ni7o552lUr
z3K_ewzDvm1*mfEDl2^{AgbNv*&FkRg4xr@)j6W$G3cJlh!d6J8)T*gGA8K;x5Ob$S
zV@vJm)dy6kKlZckLN!MdN6u)itgO_mNl6C$z)>;&`JIv?<JKy8_$aviVaj>Qlwvh2
z&Fj#jw4-`mr4~!KJqAxA36(=`E9<w{+Tw5S1bEXG=7YnB#1Y7<{gRs#-lXWm2PD=C
zb$z+NQFK*R6%&_WS`wp(Mh#tFOPc6dh86PElooM9xPGp-cnfyQFVEQHW=AnnoL9Mz
z3MMWfN$BeaH^kba8j;Rm@j(^N2wsa{^4h8BM!oXph2A;Nbfw;>!_J;m*d$36vPWOm
zYVI+3u4IE^_uPGx1()M9$wX&uv5A;yr`K9rLqn0Y<73X}Ppt+#BWyx?bq?B|j8J=y
zy*VEQ+()HXKRom0<LUTaPzqFnA5zjk{8#@!3&R!e`TJgP2WaDpIi+q-Cf~LG+)<aO
zW&X!?XlZK?0o#|k0O#d50Y_k)b_^w`P##;i&gcvKK_5+b?_WTx=iMHBRxlh9W~gRV
zBE#sz)IT%^4b4iYaMpRZJ{;jC^NC?A!Qyy9Sz^G4Wnf5`QwLx*HJk{#B>_9y`!ebv
z+qmJezb$3IUqBsp3hO^k$9G3Le3pbf-JW)CkYfe$B+x@}w-egYN@UbHI80A9aOx!A
z*b<FfWmXm<f_m_S`I&@H8h+*&>1Y;$N9bF#4E3403m<Zb)Y6F%%otNh0aZR;<hz2u
zI<06Ua&CHcMd87xJCb*#R*+sAKS)R_y}B%f`l50P97t)cfJF->7K)%^BEE^BD;jHl
z4Y{Wg0E?$75W6d_w|XlXKS?;JNY7`K<i9M^D^dap-4!GbA{a&qtuv@GN(ZUO9k!$E
zja&|94K_+m{h0mEk;(-?H?+D4HH41}>jKH0EG-Lf;+<C6O9zw3#46nj;<t5kNj~)Y
zm9?K4;N}Qx?#Ra`T}lNvI&LYWC*{x}&BU<mXa>keUzGv7d%PHzEdI;&%0t&vgD=Vh
z$8pI#g>k8ugE&L_O<Jw)J6p6Q##O87`mQJXv99O$DQ6-jWe7Cnd$~_n=qA5Isz_25
zmX5QJ#qPV8E<t&SOC!XtK}+1r7%t~+{R91ba9*c@Yo8cjls+LL86b#F3`G>JVHjd<
z-&|5sDKCgLACkQ_(%$Q|xDPP7%SP$;N!YW4H=CLGJk~+PHOWO0IDp;~YfloW_@xB|
z`CKo;x7pu%&290J;~U8*v6rwp^zV*yV4Q7Sf;b0>oC8y)luB4DT=~BN4i-MWA=zdE
zpM|p`657_&3{il^C>4nNfq!b-XF(Xb##&li@1VAU*=-8UxNYYrQ5DTBu_Y28|E0Fa
zSE+yVXs2!#^%W7f1YT-z1hg$8OlAV5rDNA%7&~u*njmpTWtylJ@LmuETgGt$6#}jq
zqQ9f|<y&`nh52^kvq~uNy;k;?xl%|yE_z&GX*W};VccZBwR{;7V+(F`Z;@gSNu#IM
z<9<d$F!hbS$HsUuDoZ5N{Zv{G^sYIM{z%>r0|~+^JPDF9wtEUDVh?%CNGlvh$?j?u
z9v-2D6P5`!OJF$xJ;1LEVw3WfZS7>z;ARoiV(#aNeUDEO#sHp9c=d!avFvq0=qo1z
zj8qY&75MD4z)sBDLX+#z@LJN3YCHG#&%T$xyh=g5@RRxbGTPQ|aJ4VGxa0ff_HuWy
zE>Zg9jX*=E?P7jQw55J_YOx&+>(bm*=TX>H$89UmGI8r+r^!x9w>d%57fv!=6IS{?
z+LYjcu#A(EiJiTk^*_|SGai_#n$a&mOHR1Yqpp@6T?x-qeUwsY_|s}7bD8nhFpQ#Y
z1a*%kfYDCQW|mLiIYEM5MAz~+>X=#xkV!%YigJ=wVm!Gvo57?`Ujm|^mw4i5U){2S
zk)Il=ZG-DH>A5SgN%p{%dMS2=g~%T${dnrT)ON3{UjZ4f<5o_t2R^ImcK%Oo;AKG(
zSa3(Ad=a!yBgFlOg7F;uF4ts*0@HY{i8ts$Mt=7lt?k8*wto>sF8x$CUN%v_wPL!3
zHYMv6_(#f+YXAbBdJ=vAE1}oyhT#CdSfWq_tR23Lb*8{&sj>U#^C)YWtfckP#5*bx
zY-7rCPr*x#njT?bY1P^Tkvp<_DaPiCiqq`yfL5uJb=ApMhlqwhTF`^PHlbULkjE4E
z%9MlV<vc)K5g(;HK`mu%@FxiJiNH%+EaG_>O>df~ZH~Y*4pA_`*Ve#_l$d4#XK=e#
z;`8&pI+2}blMVr*1nlm<QleGDa&i&h=8&6RQ{W$38E*bNKsV{KD7?y1ZmAu^{a$BX
z3HoVTK;Oie3-~NL;D!B<-=HMwqyni^S|HH+t*E<^;{_1!*lY6);X_K|k@Fx@j)n6~
z1<6j)B8K~z0}#&A?S-cP-x*H+vuIxb*N!ycgCKmKgp`yN2^^FEHF&%L2GcLX#>10_
zIo=#jl&33CbOwl*i6#d_D1e}y9yTIOl4rl!C=bNZy8E{T2qu6A`LW1NxlzxSUYNR&
z+z%dERUU}`E`gnO=dv&FfN^cW%rm)u^m`Cp#tmyZJIl;7;&@tLvUL{rrx8jI_1;=o
z%suk#mIf0<J`Pj*=frG{Wsn)XKStq#|2#_6v_Fyex}RFmLG#FpM%~>VMg!j4c(tTh
z;@^B^3y%#9%gRw>M8x^XfRk4{v@B!<u1&@a8}#UWE<i1Zi$3o37k=*}IAqpEJqcl>
z_W!%~Vq>9vG490l^t^soXVb##B5qp@W@cvmHls9D;=au^z^VzC<v{G6TJ5k+zU)&?
z#|Qol@EC>HvDWxcM$h}2c;|_o3Ct|wZC`_b32o40oA%$IzpXo!O~yPReEx}g{?(Ik
zSMcKXo%rgokmSz`jMpNuSEZHFhn>76=WP1RngGT9-I~gn!WcRtBhG5VxrO1=O*2dk
zoh%8uTYPg6Ng<weX|UQwkQaMHQR}%xb1X_VknNmq<PyolV(~vf@Hvqg;wqma%<P)5
zKx1NZQhS8Yt+A|3A&pBPtXcAQT!E!vOG#PT3P{OA13PX3dZpU%``PH?D=%Mkw1$!P
z&VF^?N%IDO)p6ir?l>`+@U?(g!WIhyk>TF!7$5|YImcl_FZ)5u^Slf;pfpb<TJ(8`
zhlgjG=+hY|f*zd4fI#f}j}JQHegEEGW!a!nLmIC=st7*JyhR4HnJtRBp*9IlD0<k~
z^R|>=o6*RfaC&mU{YblNaE^2}$OHPf^vm2g?yB`O;rra$q0IFk&_=(yn5T4fhN9}8
zT5V&mG1VzkN80CX(9Vi5@V<64ow5FQ+$JI6+2cTeIe%zF3){KP5KV+`xaDJ?4>`H#
z1dOg*bP9#}y?nQ}_ftb(!+@u&nj?+<Q6DTu&3NH7Ql5TVK`0=Dqh}r9db|KXa^HWI
zwf<lLxr<Tx+|rgIdH^(*DL;QoQ@wcqna1t6RK3%RGzLJyEHUWy=Zfcm`tbpUZ2AeZ
zP)!<1`snC*L$K@i71?}#sOx%Ns-H(Z9!e!0WFDRWqsmtCxlCZBs^x(g`YGUX?E6q~
zN=i><beaOa&=UiHuQ$+bcnjE3cPo2yqTiaO8dyF*KIWrf2#osmP`%ab5of3d&w9D7
z>m5*CrZMS0##i9tDR*+mO)LpiNagL$(+OikZ2f#%RTWicR;<EuQ{#p(Ckc#UR$kwD
z8lQ+<A1+cH)^X|Mv?$V^_o%P0S1M%vA}9O7_*)uxD(BfkFi%1gAD;`NmXYcFQnX<a
zVB$1@wBY%026OH46v+?)*G*my_GEEWe~u3EJ3SXsk2Gui4aK;t>+S1kw{eo4utEEk
zIXI0?sB5~?f-$e+=@itvqgra>_NI8BtW@FSu$Hl1sxX-$M1E*>!lmQ;Nb2A!p4DOj
zsEF_UIpLFh*%Y>8-g;@?8`5?}<D3rM{E^P6S)*nVMPfOmU<0sZw#cK(J**~dR(MRy
zeOP{HASvLM{=M%ppb3^&1{>sQJ2$WyAF5|nV)k5INgF$csJU)G^F<-_3nN;Rbaj(r
zB5wQ)E*^G-O+^JvXLn*`;m$WbURKYWrGMMC-sVFRlL{s4JF%0mZ?gmxKNz=VuztF`
zIZp?;X4A7pAIFaW-9xe1Ko?gR`2r*Ub02WaBt}KyZT~~FIR})g{vq}NH@H0sIL?6J
zousU+ot)D#v?q=*ZNiqnY=Z6}o@6eA$Dj3Y6+F$Y9IJ(2D-4(u;PFu$#;4EJrA>6S
ze>vNcHJ(n0^B5FHBv{eNe7s*dY@M#V{o*}rXs6MVt!hJ|mhKTu%n~=+;)S#LW_<;B
z0fAz$-Vl&8ml#|IP%9<lnhwVPSa``NjOrZ5e)_+ry9n{E)dtERF>@s7g}$1fPUkOU
z)L`jLRn&UMx$O}ZKkl+Yma3%0ppn&Vp-+}7k}_+$@D%fXdF|YCg0}k+u?Ve85!PtO
z+xfg@G_R}QXD_6qRa!&WA&ao<D&!Z%JeSIu&&ea)qF3p&vQ`sA>qPEkvSKsS-<stz
zQd-gaIp5Dya9uF3>35-jo$}tGuB5Qb*s=k2X7e=xkEBN7gP5!7h8Ly>-YE6VS}^1T
z9Kpp^%RJTU^9u9zjpY!@l~U-gJQF1XQswUC*T+%T)pB{F6g4^Rv;8`^EDq*S^<$U@
z9|Vg{)SBr2HgG&Kv49URaBN|Xge@hwpjJCSgnVs#ost?}<~9(i@I_mVLlMYFk#{ww
z2kR!fr$ln0W@Y-DqGQt}y~?h$blKjJt}*>I4wltjHM(<tqC<L+`K)rY&fr6QuO1l`
zXLbWd0~-+x8k6f#j-G1zd)*ilnN-v7&v@ANw<?C}Wxw_}lcc1iB4B{_Z=weO5%|zp
z^}ko@bLxL3%*zZR=K|l;X<ZlKAlZ%f|7QJ*Nmy7vJQ)a3mj`GQ@YUY8_)+$OcGo)r
zzdY(y2ZxN8SJ_-P=Wp~20m~|-_fSQJAV*vTxwHs~Sg+_oewdKd=hUgGGN(bdYBl!2
zEifL=9k3xtJ&XYtuX+Tn#9atqm-@*ts8u4IK$fD6dPZWI(&wHTPoO`bRHG7c!oi)c
zr;A;Xtfx3*+L8wP`s)FS!4lblNR@~(HPE%eK)J-9Axan-yh#f-MoKC91~T_jIvSZO
z68C=hgrE^zg<MjEb-{PA(L>A%>QX^rNA;$;Un4!wt%wOxFjiJL%)*TnOpJuIgMR4Q
zxt8x1&SjIitV1{9@76jCMNxxK@3_B<pM?mxuU9$w(Tls5xOHw4mT%3d9C&Ziw%E;Z
z{X!JDQ)AqA2v!ssvj{-)9Xi@N$^#d+sflEor9(VKi-j0Y%`A8jNvX1GxGJfWpKJh^
z=Z)-*hpM8QROr+PQv)s^qkvS)7GHe_QH4RnC3rTo73mJFPO!{Ai)-D6p1JFEzal8b
z-!#I@ZJm)L^1xo2>HjhHo?%TT-q-f1C`b>zLqado1f&T<LYFF?1Vno8D&0Z}ozOvw
zK<E$z5eNt<oq*CoDIz6QLFrO`Pv-w~p7+CioolY)<ea_NUh7_kt>KEo6A5cVfM2V`
zV(`ne`@>>}fl(u{AA6zpdg?pH3o`la-PS3sPW!QghhJ6mMW2M~ofz(%E1G$Rn<;hH
zP9SO%gkP0+pY%BFTxztmFSZ)gprY>)IYu`-UxfXt_SZ#@uERD5Zuf*f8L(t5!aqWU
zUZlrh?<nf9MSD;04rYjD2Jf_6afJL*(Iq3gDs;vPU;#H4Iyy@K`cp@L6ZFtm1Gscy
z{Bn}AsbF8OHE{A7dmreR?pOWZa@+b_lH4j&;Kkvz361!j>C<~RPWFRq;vOD+&5;{m
zSzU`81USl!f6Cvy0lMijiFcpD6xPF|UB4xazU#3aE5<Jxt>JK0)kZf}jHw<GWu1|3
zjPl87M_WnezLndHEucE?$U=(KtCC0%-5e7XQZycebHIp3)eAB8Fyce)coQJPV$}oN
zNH&b3+B8tg)H^VUZf43hG%qrGSuTyyPcxTX8k(fAFnwTCwgFGtqUO!`zWJQ4@Y?~I
zQIc(pCKKjmxoSO1Pl|Rg0(qx9J)BXqj8apj=*O7dRiP24`dD+p+6upI%=p>8P#5LV
z@ct&rC%3xcg~2zPIW9iFF<qgC3TCF8l>aDX+w7FamLW^%2QzTDozAUDZ)dpOy^(O9
z<j?MdIrwZ2aq8GhJFg%AoKlbl_>e;Rqd(aQ2#1}S?7w3TieL*E&l5`BYuGWw49TY_
zw3&&d>)3?r?SJx|)KljStv@0QRc~WVOTz|Jk6ey|n=l;PgF5&${ckA;V-T!UM;AWV
zEI;{WrlaO%ZlJHWuh`dHrI8s;(*5qUs=axSBVhjS6{B$tg-%AfN^XjGG;hzSg!9AR
zu!nHOgbeY7d(`;rASc_0e1<)$)}O&z*Amye1&-sr!vkU5U8)UiJ1Q-$O|Suw%5)bZ
z4scL2w-4RLpRCF)osdUFX{BtjLVi<^+?{_Lw=(!-jX!W0@i?$VsUzfPawLfwn*F#4
z0sk<6_WkY0w}Z!7k3sW+-a!vsekLw>UmUXBBtEw$V2B8j()9mQV^Y?H-Jl?1(P3Em
z&->3}(P?opF{t`_$)Dxxu33a22Bor<S;HK#=-ng(5v-JAOC-;K#A+wJR)K*)5lmb`
z&RQ;!xT{eTpM^yqob>1@iCw@Rg7C%rud#a!!GUCeXX9(VRa=!u7Vg7C!NfVM>#PBf
ztVK7wa}o3yIM21N`yh*&N7sbjGeP=CI^{6z5{rdG7{FDsp6h_*>sIr6!senz#0-gF
zZb)68p#7>&#JZQf=W=I>Vk7eVJ!2Fj>SF-4gl10mbzT+=`rMcCl3o5;ZYlxqD0z8P
zD85bz3tDg8Z=k)vVM*wsq#zc1!fxmXs_}HK@}EV8(&vq6O1ewd(#){$Ri3{XDXt;8
z7v<ZCOf-HD^|vWld73WfH;*OP$M@02Xh$&Y*ZXiXl{Vwvae~G_&nugM-|V8+y+;RR
z*(!`vU-1^?hv(2aI<D6hCn=g*lhvdq9na|V9e%UGg#GXg4FA%xGMFY6c$k}t?jLOG
z`?@wL>Xc}trOnRFH>coN!FKkeIOzQ1_HZWityxXd<K6&)&a^dBVMaDpW<XfK`)oNs
z*tp0PlQ#WF!7?M)8$EY9>y@)Jepg8Cv);Vgm}=4cv|HWt1tVQIv?l<s6;HQzZm<&d
zs|nh|flyPeA2ih2iu#Jh{W3QM$EO=!V%MHpC)<bE{391P2mK&j+_saw^cCO!D_Q`b
z&0m=q_&qrg7C_Sz!?mhVVi>r?VU)R)!xGx~)bWaz=KOCZh+Lo}2w<YeN!eejLKLy0
zs6d!6WJAw+8eZCV9zO>hd08yT2pk+jVjt*$5}yBS%`Dyp;8@~~|3uYF8Xp1~)qSHd
zD%yIR$6p;<PK8uIMrWQXIEjpbv)-4B35ZVja}||1f|ysWFF~|OPV38y#iWhVD+yti
zrj`|u7;q99K*IC|e?qRPN>Sf5Mx=_FV4nn=&4~{P{z*Y*<(cqF5HtF{0c}Is9qBVR
zsqYqjNTxutS-!B}V&Z%o^hK~frkop=D~A8AOn-mu=VlbzpbkHy(#oWb)U@>t#(|ZW
zA{I;4Oy0FDK=_R6k&S)<s^vAz>a1)z$cI&x6gUNY3PT;_eL4z5n|{f?;7gc%>z4Qv
zc;wa{Owu{Dg)3nFb_BU98Bpq@g?s4i`6}y~XKiIsC`ib+nQv|M=a7W@0V|ohF?Mep
zev+n(hI@mfzrO&HJE_rtU#{yEtsaE69)Y(df#JR#Ax|v)!+H8WmMwhhtMAL!pM`;e
zM&}=1S~jH*W@?pDvcHEzrsLc7JX^rc&X2LdqFn$>RW;}+t<o+j)!<l*bIgp~Z4!8?
zU0&X<7){uD`e}XCsNj`wT#};E=k^tv6USciyWf%@Z_faDvv5Jlu*UnXYxIcp05|Gv
zYGDOLNRhu1-k`!SU`6i6Y4%#d-F8kzV-jo>(Sw)X`Z4BKsS8#&Cxy#wTR(cWc6<V#
z5`XR6`0*fC`F^$&ao~{5g572(TL`%Y{u0s?7=?tbD0E${Dg1bR*q+L$75pbvANMCe
z$cWMBwH`E@GCk8K%;NfGRFUi7OT9B25O9sZWe=eA2^oz6HAVz*%3m~EzNPftZ=)m=
z2h8v%t&?VfyBZAlKZ{K{^R4oek=>khex+3vYhnrROTLE`>81raV)5yxs>8{;)|c7b
z38C%X{a_`OX@6cyY`Sl+hLblo!S`f`aj<>+XgshXSv!j)>KYAm1dpl>wQct=oiIs%
zynd!04}1MJ!RH66^7lR<Y}TY7NPyxwR69?g*Ph$KrZz_O%$}f`MjtgWKJQy1N{jh*
zHCEky#3qczm&oL;r3|)ZUC|Yj*gmqW7T@uSp|4!H%l9Hg3842~*#Vik3V2UNxt;V?
zET8&$%EvK^We~93%Aw{K>E(keeH_9siK(WfAW~Y_U|%buIfTsaxIhkXJC?_j9RqA^
z(tg>2Pw7^G1@KgZW7jC0YMK`^QC9idymaocUgYWFgNkFXONNw@y7YsNLq)Cmw#>}2
z=JLwQF*!bhWhfA^tx9|R{<Y`}>`Ys@N)QN?<M=;?MSZy!kaFFxySlmk>z>Zsl=E9%
z7pkVf0%*p635<D<X9jdE?f`tRT=nKjaMX)E5{ZbXkHSNx(o_I|_P~^h>77VcVr#xG
zdCv81=Gm+L&5YJhRWKb&Oovm{>0V{|LN9MnqyP!X#{nGZc|ng^+`3)TygPHBP$#rE
z0eDd)g|}^C$S_2$GQfI@t2I-CU};~(>AF)`vNyE;2$|2Onf0ADuY2nDVZ?Z68gPaS
z1OKwd(ma3}=?e_%TjG=bno*#?XvsQm>Q(EN%ioPoOuddDsow(ZCT*bjOxn^F$S{6G
zaS^saKHZ%4*y>xB)!{}V5jmW7J!PB~J!MT;YZea2;LnxU_b$(eU*2S0T)3N*0q8D3
zJoMIw5jDt|`VVXw&;4A11w1}b1I9?J@*4c(_kpd%&Amu~uD){ydc0Eh{@vp5&_9<=
zxw{>I2B2gNPb)fzK^n%noas%XR|R9nY=2WHcv{H?VvIAZvvqcdjT#<ehk`d~E2=6p
z(!vnDJc{lwQL*>@iwlUcZH04eUQF}lqsiZ7{$#t_Ei^@44SDdLgMf@R(+~aGb}Y4>
z>-QN&xWq@64aojI#Miwck?Z<`I|r^r@=XLR+kPv!dUSNucejUekN~n$`1$^A=|wsa
z75<k**wkd3&k@x0W9EHQlass7L_C||Sar4t&}}hk=%?8n)n##qcxAXWS4WCXGVz@C
zYTfnAYA1&hddWX^P4-8rK&h~UcOmi?FQg{shkp;rO}U59D4**kl44f2s9YRJ<Fnp~
z=iNyToOw2BXu+JmxH$BuSMiz1vlr`A^v3fterI)Dq>mCJZ!tybhFzBQi#zkMcjaX>
zqORZe7kGCkqE#1uuO9az;GqvF%9?f_H?sF`;PdAx?2`Ek9!SJbHGsI)`w8}HfS7hB
z&lvjmOi}D(2Yiz*uM$Z4z;!GwnlsGYaX7y%=+I;pxx||*SBxw_6AldjwkCOc*80AU
z!cNt5HUj&dNss07iyckM=~>|`P3dknDA(7cKaE6D@AJj|%i+5n)BYc?GSoS9uvc7-
z>(NLyuI`O3e`HGmV}e!%yqit&O)yx_`ur#Jtr3KSncHx=EU3gwu1dZ!<xUbiWOsgc
z1Hwl(FAaKLV^<t-2!~ljRB;PW)prV=-6qxc|42tP2MA&amqy^cig`LCMJ>2~hliu;
zt)z(JczT!2K8Iw1U5=ZwZ*94dz!d8FppGz)5(Ivo6M{)^Wnf5}e7rE<RFMDkt^EMa
z$!;Gf9<Yx5TM>X(J{W=G<8OaX?xIqj8+7=quqF}?i@(VT_H{gv$VLGy%a8yxd@$n{
zzPvcmNMgG6iI?t~@E11l@ve`UP7pk&$>|*NYk5%rjk0n5;tLvB{$v81V8GoV&F#Qq
z|4~YYp7H>GwTLK9mw<$cW)IUDURSc`Qco6DEi6p9o2<R_@1HLQ-*SSY(WP2R#nC!%
zm0J5L-j=SvU^%8#M);ajOdyi>Vnn!4-PHn1)L+;AqvdE(qMjO%m^78{Q*j&RG?6PH
zxiURgPwK)*n|y%PK`#Rj;CbxaceN%uv%X{~v3B=Y4IM6CmFUdI7>aT>`+#qbqbQ7M
z5z5O@)GMt@Auf<wOGApaGr3=uh_w@yZt#3E(850A(8A(YmfC%Dsgfd$^-s_Vb%LgB
z(Yhr3%-lm-P6&P$bzKW!vukJmX4=SLs&bGl-~ss%_{~EF$~Ms@0Ok$j`|&#CdTz>J
zdc%)q!pE(H$<&rNWdi_A)@sW!ggm@@4B#HD)RQ=F0Xdg_Ur)sS{)>QL#mp0qZud^P
zCa9Iu0|Ht$g$<2i2H-JGu_Ls|RpG(7b?wK2lDH}<qd0Jq@LZ#ILA8k7(GlxPo-ii$
zWOUcM+p<PP=+^XNDpC&ej{0GmnYl@g@D797v}Sgz<6%>QcCS~5_CV}?!R(_~){Hs)
zd;`z5Vp6cP)??daZqk3~y_Cds<4Ey|`=aChx4W&<s>F<9oo^Fyft*CpI9y`|yr5rJ
zOfnvvm#cwhD-!xn!RgRk_IYb{Pc-m^*5Z6*%gDrdbJBhgiA+Lq{CV=CK_Rf52*Qn!
zhX2{Ey?v?C*P>6vwih3`w9Ze&s+zbhWX5m$qvU6^nXTw;Bb1Ul4lMgtcrMiE1e?T*
zNoY`ph`>xujol{LMxXbO@I1-Yl;;uZvU?7$?Eeik)!p;;_4PLwlDsT^<(cw8Pfl(B
ze~)ui-7;Ygy}VIbS-ApOG?V%BpQ;joXo5mQV?FG^ILa7myQX$mXfw%A-{WSIewY?V
z!9fx5{^Kh+iflSO6S#HZ<Nc>^7IF8clF_xmk3OphkA6ceTsPrC&f-R-vU=7y9;(ip
zG0$ZoRP{tB-H)XmTJ>??*&q!<^D_88?D@paIyHGoUUPjM$*Pk8*ie7+FhE`eHBw>X
zSx^q-Qmr49fgU1x&@`Qqa=l+XnQ2FCvo&k^3_oJ~koBrX62_@P;6mhS(PtOTd3>z)
zr(ub13InrMnPOd}=w&7RE(2Y<%PXuBO+sM-hGplycFf~xu9i5r(k;0CyOf`s3E#DZ
zMbqNH5YZ#j(&?BBL9R$NALbH1OD}fuig)vk&S_)tqiRyVqQ;rt8^%WTnEvFPA0l7*
z`89&VckAfjp*0~3QiwCJUv^ZQL%CliFNy}yjlonNVv-4Gky0>g(DPpJ`=`a}v(|7p
zW(`uAfmL!@J-x=P#%_dYvxB$Od%hx({=Cb5CbSm^%MY>IR6flo);)9Asc^{_s#0g?
zf2Z(dyq&7~OqU3*?otv9n^CoAJ&tlJ|8OY1m+bNB=Z+fmWUBj6VtbVX<z%RZ_18m=
zD&?8M`8?*{s>(yHxDh9H7^-Gsrzf6*P)oFPUfAzNV#0)j2X0^}z+0<pSH20rC;0jE
z8eD`q|4s#?%|Nh%DV!UirR0SG%APz{0p(jAGVo?HPP9cc=wKa;Xz}0m#jviBE__L*
zMvZM+rr#jBZck%lUzCIV1rFQFt`Hj>9Je0bm+8&2`!+7XAECFOr72imJt~P}hFt@H
zO=cxGfsueSWD~PA$*EF9nsj-XUKCZAnr?MU6qOiVGtA*&yTQ!%xnEo}ggo$hJ@Zv9
zmQLNl+DgL>MpjU15&zsf3Q*6;J$t$UgbCn{cy!fh<FmtWa8aaU|J_xF<s6B{MVg<u
ze!<$YFsP>l4=M~TdL6TaV)Onp=)Db%1*7_xQyQq8-PAk`u8!#+++HGQp2Q&vG{qw*
zmCJ@~qso`<$Fz#YsNr+0W9w8iaV_x_RHVW@+@V`!BCtmW?UPmFxc$@D4*qE17+=No
zgPODED#ZX2Su~vd80MvGZo#v?bQKdeD%tkhe0HX_MV04{ndHoq38YvAUc*a?H;F<{
zHJ^${tF~J<-5trlbIq?1e6;@aYG$SzLC1bPBp{OBKCmkUeT#Y^8eXK3xqk07BlsE7
zzQg?cBFT;Q%{Vwh&JM+PI#ZU?BJgsy!^FJZ?~aA<skoq5j>zDAsLM{i!f8Z%e84dK
zFoV7$*>pI$eu-zl+G*j?{H*=7)5I&k=AVi-6X{N=9yqXe@|lYANS`8=F5>Lz>SLD_
zt~^BeZP;PLoo~qi#zNj)LIVbi*eQ6r6eG0y7p5vskWLnc>Kz6Nu57}1IS5o8qu;zS
zm}&A7yBmBok#3>zWT(35b981Hq{rGatD^I!V3hd12(H*Ke57tJfNpZ58`MP}N7|lT
zonGL|%*tJK@=3^sS^<^?PTE5--wzt5R_tcr;z7!&erSlNaGo$Qkcljp0-W5C=yFKZ
zZVxO%%SAV@Q6*YTyG<Mb*S=(FTiv6PQ-9S7GT&)I((o`r@5GQo4NjnBCf?X^WD!{g
zBVD*<dll3G-TdB@Ooa&c6fEhY;VwIBBpoD*f@^X=Iz+}UM~m}Hs`B#h-IR!?<F1d>
zd<!vKbDjRy=rf9Grx)V7NS8xNldxl6GQm{A?1ByN6f%QfM(FHyvw(tOy{H2DapDHh
z6-Mt3m@}&Py@NY9eFgBE{)}IAdKFaCvIo5Xj!99yp|3-BRM1;aXxn->T949DOPi)o
z@0^|NowKIK=?gm(NqQgJn|ntZ(z|2smSETt-|9<o-u)y!her?1+W_SZC70xk_!AxP
zfF!>|^>viqs{!Z%q|2>Bzk9j^*D!07y!pe#(`)dikH>b%s<=zw=TRCE?a0}NnBvEM
z^VOL}nZxWQxuTkhmup!^C_S59cu%X$ZV1<RXW>xTpQ~dMV}~fWlq23qW3#kDfZQuu
z+jf*gDpx?*2*f9WrepQ7F?aLkzqSSy!sDQPFE2QT-*N}w0;QPQgqrlQK}`Xh10S%>
zVh?^U*xVpz`i^HH5J;?Ta(Vq?2c<g6ZhheFl0vZ}-~9gNJpf3w7OJpGvAF!qq4R=+
za|#rbv(TTJ6H0mq;GSV`sZxi_&yT5-7op1T1_UyS2=Xy^6|HPi;V7TWEUR)!ip3ZG
zs`21{B+c?uKm;Sb)6HB@elzm7`4Ud8-*EamFJ7o9*$J)@ZPz0ICT|t<@|QUtSICj8
zv%+@qj7}$><UwF0)3WUBiW7d2v^!?7fP>b6IpV37PhdIc)s**|x>14<djQ!*+5h$p
zHIXiol-Zs91=UHXm}9vPRV6becv7`t{@J<81lNV}l&sV$PWj_UK97HFaCQ0}+?~|Q
z6QkQLb|2jWd$xT3z#A(#P6h^LpdPHMIpL!3cJa>Qe{$+WG?%TfHyuKUo&0Kw(xWzV
z2W5skPxtP?dLF5z4~6`u$Ua<YHCo#_p5l_Pi8*N7Gj@40@lXt3%d4PZrF3!dAl)bA
zR9qx<qTJAT&Zo~q>h!-l4R$i%<hM)J8R83IidKKF=BVNV*5}mpSOePq=K}5sCdEvX
za7R7)EFgedl`tHex%ea^U?LH5A0kiT^5ULKQ2;Qq=sVsO-1iM1k-P(gJ+1#Sb$G+V
z+O)!mJiGpVHGJ%<p-_h7N~UHoyDO8ZU##{!QJ5V_S_5Wn7~oy^-1BCgkbRgH(77P3
zOXE*-2p%JmfZ8(}+TVY{>7CAN6suYM*k`n=IR1+|VC<SqS$w{eDsRA8G=j{9c7tyg
zcR_i9jC%1015=dkN9;Be1DSk;OuWhMaIw00kJPbvwv+ACGW(M?<zPW8J?#b2`dN=+
zo(hd%F^8?&Uf&ciC~rH>0rjp+C9$G&zCY^vZ~$iz?X}SU|Dr(Z1Z(Rq)jkj)8~^|r
z{CB=REzX3{fS3iPCnv{_euKvR%kp>|^uxhNNRwO4n3#SDjrQ6*$O{o37;wAQZt6Bq
z2!m`2^Kad(C^!F}-W>Lb`@zb{x2BCZP#lkuXyQpfwUg~f+gM~BpCPR5dUOm;jOoti
zhmi-6PrpH6$!cJiq1=alF{I*waKziDPS7_~Dl3r3UAJd5bidcYG-}_kX4l8FR<^M>
zNKd7+TXbK{VUWa!>Ou4}WwMf4d~!M8P3z66D}|SxWMcimPmtGd?en3L8bGw=CYZ|M
z01umQ`R^;-oS$&VV||tx2yth^RCZmwuXQjtPD3}jiACoEf82l%)El#HuP~eaBNHn~
z5_aE#rE4PEA7r46*~FqEQpT5Olk}3ep5$b_1wZ!qrokcqi?WG6myx`%_b^7FSX9O2
z=uF@?WZ{dHm*YA8#PkG?U;=SDZWEg9aj@RZ0D(BU+e8(uTk;<_M>PM$W$XdzlTB3$
zV3|&(v^ccwSs_S2KwY2^!3_1E!!v3(xD)f_=`7$80ahclGY>^-yFP%L77`W?$j6?}
zSPEKZ^V4>mSQZIum+*y`{Alx2c423|^{J@X|F;0hRiyocw;p7i=HT{yCF@_2_8HZn
z>s#D+F<7>7{P#)PDen0!XnaRh=cfWQ@~I4;^{%M1s5;(=h$!vxyjJbLWvaBwP8z{1
ztoZL_k1Ip^nTPW{7JDCMwSLT5F3m_CC%P*QEAGcD0^odrbacWQ%H*l!Ji>)NoU#h=
zVfcGOjXwm4EaHtI2sd{CIQsYRr$E*W6(3LjQ#09RymSDtatP>Th@fU>W+uhX3%_{}
zzYF}&;KB)6Ir3=DY+_`)F_`=%P|YF5&{R-F>P>Zn>RN@JfV+D3ET(#@unLI4H|jV!
zyJei&tLd6D-@;Uk%B_CkQ^re7JPy;sHbGn>L>|wkbz$Kz{RaAcqB`rdRQ0aAR1dKT
z30JxLx69jeciCB3&`I>)GspE#8z?tb>?QA;tE9={nHY2e+x_hql8)@ML<}gc0KJ22
zql)|L{K6`U2JUO_d2h?jq{NgA=;KsV;vLbG+YH))Q`$fK3kX^qLz}eeUYS@jpyBXH
zN_<wRcM)$muj}JTwRr1TlYb0aT3T8G*d8hWc47Jb&&@*K^@Uo-ep0+qg#c+L7`O0w
zPrmIMg$?T|Z1c44v{b<Z@#!hKxgtPF=P<E)F<FA&!zg0zjUF+M;+ojxV#S9}ilkx0
zj?$pLmKOFvlk~Gr(})LmlV#rxdkc%6O%ubnIN<=i(Q3h<Kuucda}oqXI!oC&M@_Br
zbyayqs&`Y;$|K~hF*g&F1Yt4lBDe)v>>!Vzq$<E6=P9e#UMt`}6%&zy4Nnfw$fpNZ
zZ{i?yEBXP`!T}pCFI<fyL`6O(v8>fUBZ5A)>x7n6R1xoqcOqCn_LA1ZXd%buh;%g`
z^NPkV9^826YqLWEnSg5Io1!LVj(#CE!-_xUZi1LoVWP$DdA~|!7PLUFylRn`(BSBb
zce&FQ19~}e?Itn={SA*vIM9L<`JF_TXW28=*II=RlobKS^BccPz-y9!@|$J_EY5fC
zNS;isl_&=58qZ`<vw%bq!-}dw7f#JU*7I}OBCbik7976Qo)@^9Ex%xAZEY=HOuUc>
z9A>RMnzm?H{%-9|6ogG-R-hncFXlb)c*`q#h5~O@4j?Pl+E37@f&3H-jpq4@)LdAL
zdTPbG`u&KwzkFUa?LE#FQ70{`c>kMCzOe9onETUm9NNTWa!iC;@Z6H8c^OuaJ~gra
zf@;V`!{5r-ygWbdKsBo;ir6fQSDR8W^57~%;X__?fY9+s+DPoWi8RY~jW8aPKW~Cu
zpDO^F`Ubj|W!Sr(h)AtSMm5Im%^sgI=9V~fD@)TtVJzh8pzCIXA`Wk!qQmc+^cb=r
z;1;A3qdsS7h6Z0v*pInEM8snltl2Cy{X4S9VN6EGa%}!2PDe=UDiug0f)ghqG%=}f
z7D|7!*^fG?nsiW=fkcC$XPu67Yi}AGa!8;VnWdY6*7ip98}dSdo~Lko#8XX!f5sSI
z>%VIDXP7Sc4p1K{X%049*?jp7mTcs-nRs9AK13f?j#S*^=HB4k;}VwB3-|2^Es2;T
ze{xn`q!aE&SEyrmXt_hXmTI#m>c6ow*7E7@ZnW&%N1~nSh#jpoH<GoH>C;-yC$D=4
zhxv~gK57NR6n(v3uH7g|h#e!f__fYtae1y`^fQ03M#Xs?cJuBrQw68wJ=|RAqSC1H
zptdmYyLR=U$Z1@lQvpy1n~=J>sccPac-&2^|6lfl+%EP1{^lXlYeU)GgsZ6{V0@F0
z5AZ+^*&^Klawop}?dO!|*NC+XltK@_l#*g?5z#t&U-IkA(W(xlMP=oc>S~4t)G;`0
z;;w{Hbg~nGd-P#4ioK_ukZ64K29Tm%2R1;VR%)*yuju=N)s?0np%!U_Z%i=JT+1wI
zu=n{n2?vo)*>C3rStxmLWV8+$WC|xNO;yi7N2@yV%s1JQ1X?r$T-H<8H|sti;y7C~
zMh#o2)eZguqh2UKpKt>}XQVQaMPN=@$Lbb!I_Lp}9x(`Bz-DXp^8nzRh|j^w4EY8o
z%htfHde<0Svk|n9?16QKRFjCy&=;N{puECGljN)r!FTaR25=D&ueB=i$zx!Vc5O$u
z4sUrMsvG%GhT>D;2>97N|MptouV^BGs05ZwIlBP^(X=tHZN{&l+38^O$>DTaR@NoQ
zse6K{m#VV93#jx@h_m%OI8_?R+CSjIHdh)@;YH?leuAu-=tV3V{mSK9N=guoRx|F9
z_qaen>h$GW!dbjG+I>d*Vk2&B^Yv>rZ`0<-lMbu>AMUg#Gjx2K+VY|>ZSWLTj+ed9
zP<m%POkbcswWVj@Vbm{4d#5O%)|@~nB)zQ+faF|T|4s(tGS><2Fp93})E;(@@N*x6
z>UlUE?aT&Q!r^e}q9RqmoXIOAV+zP~HB}1v5{IN>Yf*A-!&+l)J+>@`k-u<4wU+EP
z4b|~-i!Vk^kdj0cATyYW#2>47I?Kcn06BGmNj>(SRwAjHPo(!o26a_m*$QJ7lPUu^
zAz20H=~Jx}5|3j$5+rtHNYnheB1<pDwyX&GjeNnaUCVL2D4IyLZ02t5lTGkY1+5ty
zS4r`>l7EdnxU!0M>luMP=`0!}bzzOGCd=$8!L!u#iBG=QhNUyqyhs^L!^)=_PglzF
zT&4IJ#iU9E5gV}CzDP-llI6BxHB&m&Fo1v&DMvbKgI(0{ZiWjTj#D<QW;j<_3L?*i
z(E((ro`!g?klDVb^c&?#B1UAF;zm5Hsl3BD8rm;*8ZB8tg>p*Y35vwr`gHZ<-pdza
zu;H7Hauca(;GM9@*5IBtOo|L<q1?+>H-V?s?MlEZd`C_-u^k=SP6Yd~idBxf^?;ZN
zJj^Uf8K^dah^eC56Q0UT`fW!dwUpGDoK)YO`D6S>ezW|s77@fq+;&D$&UYv?CAkg1
zKPWgC4Lwc?ye8kj^feWa)!@j$6{*(zocYu8eld!5WzKJSQq7_h9{yMDv2_Q+??Bg#
z?T;^Y^6`g&20j`JW8d!!*cek=9WD?9c#T49$gC}_hL@n6sL{Shyx(cQ2=`fD(TnJU
zYfE{X5hQQED_REZFkV&Z^45Z}JVN?Mg{35RKk^}-pbw-L(-eY~AwyITxQvqJN94e!
z4D^{5A>1uPnPxc)gBslo)8G4I+{C}U%Q#Y3FaKrh?C5HY&PXDPr=+65v77Os;N;g!
z?nB;Xg;Ul+jhT$<8f0XLXuz6unaccZPzWkv;$=ufWDEB~f80ah(dj}`+OLa1tx%rn
z?>=fObaNisy#lwhtp(k>YB(b%jIaVFHqd6P@o)u(WJi*h+-m3;%Ij0~n}`V>pUaDW
zSPYPBRuhSfaW9Uu%O&&lYg<h5AioQI^IqlFOK1=xg`+qWK$*pZtgR4tT#UhF-drD;
zQsSPW?{6ZB&|=h4<m33%v&0|hTd;W)XPsuNNpp)T$cd!Me42x3`F)X5s2<=*$ZIs9
z++X+c%yBcz*U@l=*nsNh>~@^$gbSRm7shtm{}!Hh@JSNLEL%Agv)_T%>Ua#~L^3o+
z<h-F~PF2=X63r4blo)MU*ETdnh5aN%wliq;WK}v|GBE!|UG%R0Yck-bzB(j$46Fj{
zl~w)EDcR-MD-)|*a6;8&3?j+Mv}gWUk0(E?Dg2f{4BDHoU2ALW;0W^$34PxDhzi`G
zL@lW)c(2+v<2;gCc0t)B+6KtkKcW=BX9;fi{|pq^<XuZxm%?mP?fQR}rl7u?1`Q)(
zU-R|YeD)B6ju$?~rxIuP%4rJho%)qaq)luKg<YWZE)~5mbDeXO=!!#pu{!0+S8;`+
z)O1cC_bL=TehnU~B%Z7pIW;!l2&EUnACEj1Yt7I)=@8pJ{#8q6qvUfzfk(Xh^l7-Q
zM9IyomTJ>9-9{wUE7+PL@kHng%IdjdMVi<$=}Mg1xWCC{L<CXb-T*%yXyTr-{abBi
z%M^Zt@MdG*BLuz)vIe2Xj5H-Hi^*jgla~ufTzkOTPs_nLVInsdD2);iZz7x~a$aP|
zt*re~;;AbAbZpd^tA+o%7XUAuJ2@zpeE8uXb`H<Ky85ou&DQ*{<rH0CFzchm(f&Wa
z7&7{0Nx7sLo0@7nI+ofyIJiuFUhLxlk8@Amy?HTn-nurL!ZAHJdpmx+;NDyN4gb#P
zGz*Q9)wxm{uG+%yxx@1rKtA<m=9LYaE_^jdKZFZG9|I5mdsSbO|N9Fr0jB5rwftc1
zGQB+GPKk*0baRlO4lYCIx1NvnXNucgg-QM{-*%sLffBY^*>PJ`$Nu4N3no!-X;?mq
z8$?~X!^SZ5#xkl+(kqu!`IEP4)DV#8hQ7#l@;rQF>D9bFW59<$>vFpQm%0CuIJSA#
zYHf6UG`4{*#L#Zq<I@6-ekwJQOz)U<d_qNfg$O?9i%FSl%d_tLJHOlc)q$&`MoG*Y
zoKg4l6)gx{XnJ1<-O7^X@F%;8o>@m7yo2BMf(HPoDcWrs^Xk1Nk8rDGG<%%Zo0+OQ
zhhNa4;oY3T@fZ888?ope&TfYN&uY=p8;brSPd}Xd4yE>KmscO)T&s|886>U{48%qW
zJwW38x!jIGbl@KK<J}e{tRz;J0sfG6w6};(J}F-~S#IN5kCXu8Kn5K%CJy#S;MxmJ
z%xvxBjg`Z-_imfG<@*fP&MZiHo4_@{!@6?fxCbM849*Xgls0?eLG5&NX>!XlnJFa}
zlDns6Jo{BPZE3g42K4gKe>~PU%Kj>2y{kZ5fjA3@5IH;t>K*sK+!@yTVfadhJm53E
zaxC2>!CYHn)#T|r<5IQ(0eyOPXW?Nt8=2-Dkg4@uL1|PD>FDbNdnH?#e<Q~8>tirq
zRWGdx$t}c(kB7p(pB8LPOxl9zXXj#F>3x8>>49Q5T~)zUvtIFV2G*n7#4#`u!}W@T
z)nK<oxV@JM-Jiaw(FXoP?Og3EU)yYK=IX(LOaY1>SibaVmtLyUR%;GwS5z2+Xjnm_
zZ%K8~2=|6r%2Wr@aFu|Bj3g20hnmXjsiW&=*)^#aY9&HTKv`--dv$KkcZ_a!yEl|2
zsx|E131ok!M@<L*G)svJaI~cWoLtlsk$XTbNWj`e%gb$qQa6O_7d0|B0!`k$zrl$6
zr*WG2YTy*H9ez!i6SZzMwKGrJoai3@BW&y#_kcI~QQUv+4|OhtW1)I4pf<0R+_YQ4
z#@}g6qj>P=hjzA7Sgg8}n!37c8XPb(;FbTJB#F|Emg#eft+DXk&*I`%k53NVUmXDY
zeuN0KD@F+`2(?~g%+zw_GTIZWGI@_y*Jv50PB&^PDhC#{8=y#$C0b^vF+G$UB&*2+
z>3TFT^ScTOi4`C#jFg3Za6?q3r0C`AMz&kDdXR`kx%D>7(@jrQKHDq)Ol{CF-ff(l
z+1Pn8+0wwQR+6$ZCg2fd8$$$t#f=DlJ6P3V(@)kfG;@E#Vz!y7!hf4r@)od94)RIf
zl2>e2b=zNm=JuqGr`b->5M;FSxmax6uE;poFnm+173FxKfmNWg(C9`qw(^BCv#-u9
zeJ)9I5erz6-3H$s#XCjLvi<yKSU^+5eF2p!sDi?>GCs-q0EQuRL`@26*$|@_r-M6`
zeUu>3`W`2Avw3T*z>ieNgwm)szFugWk$&d{Ikr!ocx&#l&NSbQIs({o(OO)pKZX0s
zsXR)4PVoV1BVZA@1vpp=tIhMl*IZNa{mptb>@5MSFXE618TpKqnHVSXRD)^H|3nW8
zg6P2l{{I~lMRny6tAAuP06mg50aH~|lm6?McV2NKc0=#(p>aa4!jCsy2Vb5ngz^Lb
zG-+c;Mhxiv2%#Z#Nl73trXapv8|$?CDEonNB@KndxP+)>$t>EW2jFU2<`6@8J1W(C
z?%!_uM2-7JZ3v3Y67~z=3V%NMGDP(oS5ALsY_3nhl2Z*3gT;eh#0k^_33(McJVxg$
zc^!<Vr?qiUiy7uZZ_3@?i3V`I8Q4?EJhkFJJP*sa1o%GHt`!>B1OmP~?FQkdArCPT
z!SThcv=~`;JWe%*x7qDo0vQSP{_>pk82k*Gv6T;;!VQ{40S)wcopXh@GvP6$$<)+D
zC%*N@w4jfJcr~2`zOwdEReuU7P*n6J!MxjrTSZ-N^-M-?dEbQ`cS7Nzh^rT>6y?H8
z>8GiQ4dj&1Yf|CUO_7-4fma(l*P*_Eb=U3HhW?5u9<ubM%nNyy?Q6+9a_1I4%Wa@T
zd3x=LP=_!!W<>w_JkNi^Cq;=@qo~x$^IbdAt3XsCN$J7Z@$BS(;nl9++`yNm7*sE(
zx|1(srkN><@Q<7#gy{ZU9xEfb0PPAI{Tn1WiPoZ(1LQ7V!2RDd4;+NhEiIO($3MlW
zlzzYd(J-!O?ZSEk+*sZK4;vC3alRA9OOqFQ-<E{iFAwaIY!8b9w<5{8waK<&L}sRH
zC6vbp#mm&5noKJkW7Prf4c!_MTGndsV%c!u7i1OcY4l5!g(e$gz6cATWlbxQ9489n
z9JSSiMCn0}%WoG&hh<q9K!Av@(Kbd6#<?fZ$qp^G_1~ORIsv+VSWuab-BHonMz9Dv
zKw+RBe)@!Bb3!V1{hrS@#_$B`k5)!*+WqD$AV0AemEDB;vxeSNNRDrKwcJowq0of8
z$bGAewyx08us;QPHbnL-HpVdW1W|s}WF3Rcp}DWdtS>8e7g4?qz>kVehz%n$A46z1
zymnXv)I)rKGNq87TSY2~1_WxVW`WYpHfOTGpB9$3>lJSZEaX)}y{=&Z2*G9Q!a`}=
zifzH=$YR@$V$r4}Lq_>)kL8^cvA{@{P{36Hg@b&DynYYe9_}YpIt*n=Sl-E>Bovqt
zLVo}+&vzQ}MC+^{CZ`<wtf4$^OwsCPL))q5A1n3B*Z-q9kawB00)x;f$yYY8ncpy}
z%WPu}-0zb#5c((@^*K2?+4}+XksfS}2Vi7ZSOLO=j;?OPrNr<Y`;)msl;5kj<<~#}
z$^h2L6?FbvFoHr^{Tu@NctaCHL>ouS72<#0;0a|XSseqL<p7(<p^-77s*69S$Q4A(
zyB~{_a;?!ajt~(M%`7STSeOFzPn&ofQhnw=jSu2VPS(4=bFOz&B8BClvpv#3F$u)L
z^J`6cS(Z%bjlmWA)u4XqaKo#Y$-xu!%zYzYb8(fEDH3nr8|7f#QVNhNGownPuLHo^
z@u8<zY}JPeX$PO|RZQO-(^Db|jo@O%Yo)Uo%)x^w;R1c0hZ%$VHlSOYc_-L1<>aSu
zstOI|bqJ+Ut1HiF392A)wb|5ZnsKEhk3=eJz;{?6)3%xXt(pkrCK=+-G1ts6UgMj@
z{R|~M)AZ{L5}9rm$t1}vzeFXq=FzgbKaY+>P)WX-&1X@@=nA&tg>}w7V)zwBodf9?
zQ@j-s^S5Q-@)1)51V9-OZA4|?c5+pwV3oNKfG{~NfD%v92*N1pBJUF2y7tr{#LD9G
zK*I?2s&>OlyS(Z_=6Ljg9%CNYB1ZqOM^ronIKruO+xtb#rHyw?3?FUJ>HvK>c;GLY
z!Erap*_h3*8xg$yPJX4fy88ZpPhX!7(1b_nr1qU1$m|drJ->$kxhQcM%%CB0$3DIH
z8SI@bs>x<gCP`rt`5l5wT<$TEu7BNp|N6+m3mW=6$ady$bM9$f5|C0~A)x@&)uNBo
zDs42&kt#*xFSmiCIG$EeBZ8@y*su=^A5L0<SjJOJe1>X+>-ry9ank8dynA1&R|1OU
zYCl&amhGK;(l2lnL&EbYUJ+UD`a}dc25$8}B8fElIB6gl9rjC&4&_B&6QCvkgdGn8
zMSG(uC*GB}SPpss(*-DMf(3g6wg$>;jwmb;?NIh&A4R(R-EBm9!)Z4L4rA`WG{yW$
z4kPDPYxW6^+NqrIJ8@wJ#fLzcK(9#Tcm#YYuk_Ca!9Mo1J{A?L841I;dv_pb;vRc^
zhZp^n%}lpNMX>Q}qkvB5BdZ;pt+{YlzFhBJxQs)J|0f%;jpBauioN6J8`sOvcaPx-
z>mjEbWzwCgPON(J8$6*=#8e#Bpx4rvp^f7XKEM1>o6X9m&z)4yOSQ(<_8y?Se9l_?
zqVxaS@$5AE2-E;o^`usvZv}s_O-*5s+^(hg>`T$Ri`)Rw)AIts-hU5RKpRLk${U6Q
z*+xvr;?o}Rh>Gg9UmRH_6j`;0U^8C!KTJ1;BxLC)+P-4Frcwfh+`akox2>!Y6lP&)
ztY+om%~3a0PkMK)?7eC_P)qYkP=rj~3%k!ff7rs`Q}IDnNFpuOwlp-Sn>aeRrQQNT
z=h7tudU|=-q;a=Mc^302Y~JOLVHgECDV+11Gw|q(dZPqUhT<jA<QoqrWzFc53GJaa
zJQZX5$`pUYzS-G8=|*id4;sG1dJ*3bO_3mCu!yGPc_MC!vA$kuqx2|EA7<nIMP#4o
zM)CDf!F~g?OvKjn1~hpQ;{(WxuFTr@EspHV{iMB5YXw>XYLN||^Ol6Jr(xb%Mw^hj
zG{8PVCveyX6jA{H>ldp4_pW7>{=k^X2|iYX=PX*#8kaF{U#2(qr*Z#a{BKH@{eQ1P
zWAHRY#ss6sDR|p)D49?&M6gatYj?vP97@oorLHVclb_f%j&x6Q6SeOWT9?f}%l12S
zSV^BHSVDX-p@zMmJIQ#lP6j3qwr3)ZL90}Xqti~tnXI*0P}aA}VolPdlWwY;2I!N=
z%9u|6pk^$F-j4C-Zi!v5o@)7}(xxpFWe0hgto%mQ&_PNxYwHc*rNFfA$RpD1fx_RM
z#0UkTBLw~W^*x8%!*?27#?}8|+D~2O0Mezj2h=_>kE>(;R|B0G?!c&#`dG11Iml)M
z=q0Dow0Qd~;_ot|!0_)U-FKBg6BkjPpmBumD;enPYZKw+<*oPMFrs%SIEG2(VZ<QQ
zb6A7hv>tz#(-)g6(;EM$RmxK7?=`W&s+<?+i6B!HxA_9FPX?fV8QO2l`=z@$fzBS5
z#7UEwj94l%zgY>EfJtsa(@M9OeuL66_W5^43o}GA?isUzjC&Jr4)YB2k;VTSY6tQW
zr)=u#>YD~zgsx6Kr4##v9ed!TBf*shn3ZhgE`Kivm>-oInLXh4%MY}1RQ%OqoS6|^
zVWs}m{a%o@IJ_k<AR+<-efU??`acI#Lb5KsZ>!8tZhfk@*;!h_Tel{Mfq~>+FHnr=
z*9V*?w8o#wEoDJ>CLZWd1};;x!A5s)cgAMGrfQOQvDf8CmC3ldmVPg~u(EFJ;%ur$
zeFS7|8cxP~#937R`Gz+DP?EsU_xp*2p5bVHTC5y!<a3#-&c+l4jqBdgvbN4wAKn%D
z*Lpo9(bxa%$~FI;E@fyBPBlAZ%ddXDrcUuS;Rw2bi9BEpEp3?V=*WEZpV#ym;5BuQ
zIJW?=byZdlRrFBbGcB@jcJ=V6xxBKJpvKO$*T;s4X7fb?(8MZoi)ru4cM@M6&za2+
zkE?3iQ-CYC5y1;kQB^fw-b_G?oO`oDzxd?p=W7t{5n3ZRIw|^>->t=--LB4&u?cJ2
zGrX9yL~l=OoGwTjEzOn%3;sRxNICxdRapTJYFdt3jljz&DnQ=AbM8#i%WIz6p<CS4
zo;bX^5!d}&Nn~%N&Q9Y37Uci)8zSg<r)JcjM3Eupi|JRmM1L*5)RF44-ZSXL{*D^f
z6Zztd;JO7c(9=#yi<55vx^IQIox^ux&JcrQxfrC~m6Z7>3y3IkW(eS8eEkEvmY)+|
zwdDD$Q^IB-O~E=X_We~cpFt7|s#MW3cG0%WpF}apxFe&#A4BK(f7itRDLP*zg51Tc
zlhF@Yb_rJPRBaUSjoXkw)6x(LMpHA(`<G+JW;(*dzXA_GLEAQaG@rJPt@(t*xmp-n
zSCA<-AY)RBHZ2iwh@2$cjJ0OO<Q^2=T&n&EabbQX&4Zj7>SlpYSJb@*GTK(saohAy
zY+Z8Qi{6+6*a5in#~Y)DDG0+7;TNUXAtghIXty4<2(3s3s^#7n^phA-deOYoH>4sC
zjqx!P^m-L;?&iR<HWyrNX}Pu88KAjr>E48-T=y}RpOia-Te(=F75c>=nsO#QE_ins
zfYIIg*6b&|x#>Sj#?bLh^+`KMUMFu(!VJ32zo!v>y;kR98Iez7W{RCHO%y3zTsw?r
zYUSHf@<3CmJelVUW?PoK$?;XT`q$uG)X=rQA}|1puf)xN`C@)o=216^1yJN9s?>Sz
zLudjIq_B8%<>nKBk-BwT!+6d{OV=Xue(SMG_Sqb<(b}n#+lNO~p%^AGd*ZC}m*@XI
zK=S`SK<J--(DyYIZ{<`ee`6oEi)JYX*6U>`1ICP#^ZnsNCcj|tLb4w@zde&M)KUQ$
zWdEh$H7~q%(og4u$xF=rs*Qhbw~GWR-#1fXpyfz2UZ-G)eJI*$o*DK15en(}T=TAz
zj~3SRt*da~v-2xv`mist_1g5lqn;h5N&@RkaJlnjU-mz=H#G(&2dj8Kdfe8q20Csq
ze6tXFZ8gRzs*jOO(8jv9ZiF(gkZO*)TIB(>{@uaq+;WI%Lc-%|C|`X-nocbF^yx~o
zn_-_!-4#KfLm>=is*$JJjhg}WxZ!2XWb+x6W!rsg&unr^N8KW_!FP551p*KR6wUHi
zLe*cTy1W$2Q^aPlAV(xCSkWzwCfIt5ke5C3fMIoiKA_wtwcOj|&a9>qk0bnP^VX~H
z0Z*fnA9tMZ#fjLo!vSiL3<5u$gycS@8BjN7gCwf4pVL1Fw4+;n9KqG<pCrt3ZKf8k
zSG~CmI<!l<=<1D*^%ByWpB?mUN=O0j?eFRS?&2(D|F_#!sP4>#mGxso^^cBZsEX2n
zpmZmpNwNOzTRh-@Hx8_Sy89@>T(w!konB7Eo_7BUfrVLR<}q)Vbk{iNjj!AO>ZgDm
zdE8tM1RqB)EbUrjrCKcl?7~2Ko6l&9jma%J(KeI&_k<7^l7m7eFLi7T;C$C7r3w7|
z5ERykw9?i1p{|$!YAMGO@Q7ykPV@QG8QQ>#K_K-5=9Cx`*`hhXSM;%8HuDaIXCltO
zYFg3ZU4YP7vZqvlix!qgLv0-kTb|Tv%NUjcdF4j0K=XY(HY4nz-vOb@$Rc{Bh_b3{
zY>t0pB4Pgo_PqqAUS7@vp)`WCQ-{E*BnA$tU4U$W2?>pg1|6FN^{9rjk5YizxgjHD
zjrws2<ohE;E0zQ04_g7#yF7T&+C*~q^u;=q2hV$VM!O8qZE!Xfqv>L82Px;kV=RQv
zBBsQW4H5TsQUHPYG0kS67A+Km?6!aUFNxU2#iixyFJ-M>vkn}ZwIE}uN0^XXJhW@&
zCe~W0(o0>RuooNsJDY5?|GT1{C`8P)2nYxmWUC*r(f;Dbo726Jf^20aQU3l4u8%>=
z73C!h!DK;K-T|%E>#;R#(9Yz{{ZNU0CzX5PnBN}(=j>=<(SYHiq>D+Ndd96=RU_-o
zS#J!o*T<0N++$?f)53y!CB~T9ooODGNE$<XLDZs=kyX_r3X5K96CV1GN@TU@O$o`4
zW_WGFZ{c!M%5Yupwe;x;my=J$+z9OC<_Obj4jyB;u?MIab!J8g;ZJiiYhY@WBl2E1
z<B3^|B*=EA52K4z+Dk4t<?1OXLxrxCV1$LM>^%*$(JlL^M}aJYj71*m=hz1ki<JjD
zT#Cj^U0u>a7@)6Km^RR&aMW*}_7QJidQZSQ=e>4``!c<GUGFM1GJwk)32ZMVwOpQO
z0RU}`@rFmz{W++)pC@k;cD8+|?T?Zse1?IR_H9s)F+@bG_4^AlG{B9MhXuGPa$Ci<
z<gjF4Oax0V7fzRyHgxLDOLortP5MpvjBV!@^GqDp{2iQT?+|`3sE>0#L*OPscMtH|
zvW`?#2?i+kE6z5rqgUx$03fMVqkco6UT^5$UUV+B6QtWT>)Gl8qbu?R0!UMu3S(N_
zv5D*D37@p0zwF^B;$P6Pi=MhT=Bvlv0(K5f*;NZGQ&VtT>N4PSgL4nJ@b*@wxiYht
zob^{6WC>K%kWs4fQW!FD;0!*C)7$qeeJ7S~K0MpjeePTSs)OS4vgNf$JIQ_t+aqeh
zs2_IUWbL_5kk7^gm*!;Ze8?AgqZ0RHsA1ipl4yU}Ie5hR36P)0X>hSnRv?qnkB1N%
zXt1Hz#8s87#l73->2jw?BG@IE(S@@94FcY<vcl+0wBW_r#^0frBXVLBUKGGT;~UE5
zO-n=Lzka*w>}YFRzgVQnLCBN=hkjL%-t#l0czd3u`<;iVf=zH~16%#8|2N<qAqfMH
zy?bJEtIv0be{ntEY{;N+FrX<hh{DVLBI@qJJUua<w};XTxT9+Jv>IqrqWa6z(&%Cm
z6Q@Yzqj)6Wr(fsJk@2VuWU9}~$}0D*y|2nsTdBTEmr<E@DjJ-wp882ZWYji~P|4Go
z_tt8hC+|T`fcw2@eVW}KZ@mcm&KPhu#dQn25?j`baOO7}@@MVN2ev!%r^M@KpNndf
z^Uy$%2r9FsMh+1;_iB^n#Pf&w0KENve#qtZfkE8ZcEI8=w7-qdaxa2bFBzD!a6WzP
zWks25w!yZO4JUj14pdqntJQ$DBel_tb0d6TGv;UVPRZX(QQh5t<xu9?%4926$Tws%
z6eHf9PHzPYfDMBT`+k{TCQDup)9odA6qQhW!zSut`MMZls<&p^T19mAQndoFr_j^$
zOAqKe)S{WRVq??5HRJx$hEa&#r-mM|RII8A4|6yLiCGa($K|Qp$>Fq(!(6jZz3XJ<
zp?VicPK%C5_=w~27h9kYhEZ<)^-JRq;MGQ}x&21(Au~z*5u=o^ryDg52OmP6O-KQ8
z?iGWa4I?V~`-33-97ZL;Zn3#110~<SMG?4H^=|u0-@Yl6DXSr6FU>2`>ESUsb|{jl
zR>)2kw7R<5{oQLflB#!-U&HM~-Wx|o>C}26cq`O&HWuE!CI99}l`x_0Hg(@DkT)e!
zlIbd$JNH@qE#l{~5zyCdzdFh4okt;ofY1D0l7&P7g8E8KF1w6-7G@~;e7A4wq%HV(
zYQHfZyV!VRk{2&|artLpZcQn<g~h99<`$0-ui@$WvHgZ@2$IzC?>De9^xwh6E|}ie
z?d2D6EM*K7%_i(gb2s|69a2~r_^(Db>|LL2xi<l9M=lma8q)x#EunI2wyBnfpT8>2
zPJpM&MleaLo8tuKpP+TagW^1_bRovk4k^)<Y`Z#?UGIz6M)0OzXJHUAe-i|z`o+zb
zO-9+w$gc1>jUd;pk=x7q&Ng~T>)SD+d*USV(^Z_QKR8RD0{&Ity7sSswZw)|>_4QF
z2R-3~(C;jw>ghpkf4-X<nKp>Sjg`w4tx+rg;Eq7PUZyf^kZGsBx>{=YDiUn@&b7qQ
zvU>e%jn$omu}3?S3>4pGyX?lt#tvlt7QsKLLlX>HPA=lzxh+vjMS0|crTQ^Xp(*uu
zflA8sEWg#E(f1$P3cZri2aG4{I*kNB|NcMq+M7GAKQS(_kA_Oeb{ag(aW*SWuktfe
zcr6GU_7PHL>}Sq6ue|iv|D<F)t1g0~A^EM~7#n|iDeC3b6%sNxn8J~{y6q9p)dz%-
zjDLv#nFAQ_#$y*i7*q1e4m_`w{CT&x<{mg&<{(uA2vh<Pi3iaU%Pd1CeQ?K1)vNBe
zo!G}>^&@Cg5}g%}P*1~{31CV2lQqA1`Pu|v2MXLTCo3S43|F&Irb<ro*wmCi1dDm#
zE!i~UM@2XPnMehf<Z4Jg1g^42RDCGViEMqyJiyQ8KLpCRyeD4mKi~Qt@clIG6}1<f
z6FWd{8?gp-2NyZeg4PD*<CMq&7xuyr{%<Jeq<B=$7^zsmAW70q%_PG<t>d>zX$D^D
z@c*7PLKmYtx19!!@7C3?Y7-2;@%^`eznSTHnf|o0J<U?0fY=xcs#!wuTy}4ttKr!`
zveK?#t1jWGsvdX|Sr1M25$!t%*TG)YIbT`3R!_zp^<fFV!u0>q^_F2(ZQc8@ibzU#
zgVdG|=?>}cZUO0TP`W#$8|m&22|+rg8#XB|{jSYXpWpw(dtH7wNB3TPt}(|PbL4$r
zgZ#J@Uv)>p3hb>Syqhq$@b+%s-cj#y)JwVUkpM>>99UvHYvXH3=mc9n;SFM%v;YZ&
z{k7Ssn@PF-7e#$_X2@uf*&Pt5g7=+F<X?<nB?q&lje6pDHitH~G5{;?y`OIhkOZVy
z)M}UbIbSLMEj-U*K|os~!uzQ@$fM-k#}^hD<K%9;7d|eM#0OuJ4TfX#p*kI)U-VW=
zZDO8Hpqhy3eoJP^dq8}>71wbX#{CU9HdHQWELsDW2CJ0B7xB9&A22{+cE<!mWIn(w
zLag{2bqBd+3&$#VM4XZTZB)Ij&S~7TJgs8I@cs2y*)+dJqg_xY1UMgF7|S?IHdw6O
z=*~@5Hif1``#Koohvp`*?xL2f*oTG;DmX$dw|^1sUEEV>7zo!S1H}~3xsGn|4v0c9
zjcHXG%_qnKj-J{#DUk!rn^@x(wDfD4$3ta(uM|Aa;g}hK^jZq5c@C<rcn1_Rk9<%w
z62LjLMQR^&*QeTdU*guauQw3Strt6(5TLuBhzc>DvYaB$&A&xLf|1J^E&s;%{C-c@
zAHc%P9850UPMR5<_GS+OXZ`yaFk9av_EjdeKqzXL6^mJ5=32p4oY!ZBebAR5J?2tP
z=pJB?A10<>s5X(3F@m*xe7IMse~rf~e#Syffa+NrBhI;ToZXzKU9fZ;`onAOO;JVB
zg<?ZE&=@{(TOF4_2?ukk?_TsbHoHszub5f@&$%6W`@1ozjn_y?aE^2fMa_~$48KP|
z*(ZXu(#_57?Ot%P@G7c7V3rXFEvnN1gWl<LVa5CGym&Cf<|;CiNv72=3cTOGQ9PFD
z<AK`3iFUjrM!RpF3=-5QirB;6Rt~#CnsjG^5@b1(JPElR$C@G8r3z8Lhj5?7b<yOf
zhV~J;NuY*KRpjCE^RQkacA_Hee=fsHz?00viCmIe`5zqsjJhw;w-y4xv`4hYkL&l!
zxC;PUz|1aZ&WO6%ix9Ee<c?+#fA1aX@4M?YfVgj)Neo{xyj7g#(G-8}J*PM?R{XO_
z(n%D1+Nn;r_#PpIsW68*%ABa|uT#(0f}&aKy%K#MYNVUNxFC&`EkKp5f)3x@m-7ne
zD8;q?3!@%BJ{{~H|HXP3{+a>CUjc+UhzdJm^;L&ry{OsSU*D`nfABUYG$EY`9(aS?
zjm!}xiWP0M2`xfs7|EY|BCn=GSCiqX_JNQ2{l~$#ty6SwXt^ZaBe_(YGaYLE->0O#
z`|rGkci=slevY#^DHEX?jHlNJI!01Y(>#+M&uXeS+iW1Bt9ME{Z`pP<_xeSo8~XRI
z_JL+&iZ(dN%Ikl3RS$Gh6Bz`9Gmv%3VA?$zMqnD=QR^hqtz<f+dcM1R3(~Ayl=}0d
zh(;EW!Ej&&^+xLP2D?;OQjZJo!SKaeP*_pFe|I7nuRF;}I@uNdYtHYR0f^@SL(eS5
zW>xJHSc(@oE)EM9BTGd@9^`Z7@5p&*3kSDy>ISuM@zLbyjCc^nBa`k-??Gg5-!c&4
zkGY)F1e6l^(eU>|BA%rO_V*h013SnPdq-`6d~%sP)V<qon65Xo5XfD+X&L^Fx6pO{
z`RVGmve$7$HV%NvHKNEiqWSrFp=EMRRWTKSSN{rvoRHM7P0yd^lN4VzaP}`@^qb}H
zuS%BqT}Jn`dn>!^-Ppi|MV<oZC!aVx4*7}@G)e3#^O2^cRr#-JXM^DA0fHV6hKmS~
z&cH@Y>gn)MoJU$%0D_5+Lo_DyZ54xJw;5DJe7MK?cz+IFRngPA`&NLY_72gg{|(nW
zjZ4p6X3o_zZU{gmaFtLG55ydgw&}Fvc_(1dsUQ9Ct4x5(^&v^RmzGt3#Nj3zc0QyR
z=JybXB<^xzH`)Ldg3O1UM4y>-*U5?hKf7}SG1<TpxN%6cNg>TU_P-mc@`r+jDab8{
z9#=l!%yT7Qt&5#@Et?LtCh^7GG`k-M|J9-=UBE+4rqEcP>63_RfFLJ~hXh9@=vecJ
zAziN<JmG0Hn<#H%5BXaV#mONx3)oZrIQ=JKGl<g<XwzJm#O|ErrO&-8?(g52Xba5K
zfT-Yi=o1*CnzX$+j*&oGF+2a>_c`EcLcS%hiM*tQM0i_x;CK>(hDErQZ;dO(e<R38
z;8`x+@#v$drLpz*OMQ_M5Mv=?jAe84wghZ9+(rke(>s4rQOOE~sHC1gK1wSj{9WJ)
zb(kv5hwYEwgZ#w))keb!A#pe4weLQE!2;ZCu?HHMR;M6xXY&o<D<;WJT>c5!rHJN#
z&nDrMTNXS(yOajYsWsKb7#=ZcIeWx>UK3<In<^%$Y~PyHY*IPbLk06FhV26Y68y_k
z8udf87P2S!<m9X(T$wz|NW@op5U3eAxP{ZltL*lA+3mLT5KqQkZOGm<+x)pixZ(c@
zW{#X}-N<A%btEP`ldmwwK?{m(lT!QEA40nlk@gq0pYS=vegJD6Ijv_4e+^MPDScVa
z?Zk!s@gWMzwq#(1fS#2*40-t+9V*t7jsCR);e7xn{gb*sU|1n7Ng>^CGl4IBsKnJt
zCV-7g^kUZKBXBH-Bm(jK{@}2)Y640M5#h;KxUkUpbb2m9;A?i~UIV&7(py1Gg937p
z61~sc{u%fP$aETHfj?_iR3|r5JhXuCYXw0zVAUvTUw+0vvJ7(m3tA@@nEx2XKFFQ$
zWS{V8FKq{~k@we;V<b-gH_-Joq5wRnPnCk4R+%f)zL4PIaAV?nGy9($`|u7Ph1tU2
zi+l)y9H_o}C{+8FWgf*q@0+amBYz?4P$59x5Xt+1LUiVL7wtWGPxV41Z1(=Aa$!#@
zXL*`~bW&gF2Yiwr(6Pdjs>|kIt~%*|&6Kn9U&%Cl4qgDSqi>B>MM>^wd;GhUtBfFV
zA;kW5q|YT5G5<QDF@Bu2Nr1fnU$ZmE^yGZ`%Mg|W60f3*Y4%{?+zbmcNx~mV9QelZ
z&nn_U@)kxg<h(Pg8o-t>kOE?8PILc4@UVnF0bYTT5;qHW`3Ctl;13>TAPn+fbCki-
zp-B+O31V4N4(FHtrwl=m*CaPiG<MU%r{Y~E1HLimEO9XL9?5kdNn*i>>R-5mhyCwy
zAr_O1|L?8=kHA{UL4xSWe}AfYA$@Yr0uvjwdw&-MEan09M_zwF*thbOI{;Q-eG@%>
z&H*`Z|1JOj|I`2b0AkN?e8>nxg#V9qS{sX=Ia`&dSATlKZ&S!W{clOY9zM0e#6)XU
zgTuHW|Dq+U;}oG>{x-HaM=|EcoxzE_ce=yXAn?KJ%Id%R?Hm$hYq)X#SRKTTQOG5a
zu<~1Rd0_+5tz|{xbVinHVD^!S&V=T)m0$t;;$0o{ev!-82I<cJ%8S|h8?LxN5OqEj
ze<aAr7BhHf2q|Q4?hFydhc0?DsN@-+z!!n&t}g2tPAP|Did<T$o!-^mu&))1nT)6E
zY2^K<uzh-e<z+elB=5e=TplXDzQ-MFjq{mVUCE6sz{H5AAuq&Pd#l=eXRQ6rP-sC)
zqg8$eCBQXamD_n|U6bLe)z0_uG5G|5v4~MyCH(0bZ3jS+bQpWFf0w%EMTKKR!wNnH
z7W6^}oub~PxB)w_xKd=h<1}O{Ue)!F1Kv0D_bf$37W*3F`mFIn=8-6F`^}jzj?^~K
z43Z>|7*dHZdKsrB*m>J@ag!z*!Va!i^U)Uet`sdiNZWOqYq`umH8ccL97ic-e<yJG
zZ_oN>dLb?V8k<wX^ttPkA?}k~Yc+{w@rEpg^^I6N_Zimr2ZPY0YP(>RRL4n{w_j%{
zS$U+Ln>SagS01esztLN@P8~*}%tmHBa_Vy3?LV_p5{7tZSZ%yB+j0L{)49R_Tm>Kl
zf{I5(Z=evfyo`qvdcFTHN@8s>Uw80kd|E8i)g#@Zk976Y=gb-G)YU=GL%Eu7-62BJ
zKlF*4AsTAAT`A>oW>V_)2o?MN!)%^em;K+9Gl%!@`2Ze9e%|YYjl=tl^?+pNlV+-X
zLzC>JYHO^HtFrgOWv{zGuA-|+PI~Kz3JJANz5iq(C7C(G{5xPjZ{^)rsVt3ZO0}eT
zpiq-<{ii;v6%_3r)L@0v=M3MYf{g1B3t)O!6SyWL?t5l60NOw93L5$mN6gTyX78PH
zlz*3@i%rDP_Q0YYWLf?i)p28_No6Y?xrft3&$-%>_udw8dwwYb9?SGQ#a3#i1;Q1+
zHoNa)Plqbp2Pf>wlS9l^K#N;ehPXISPQs~2u(wuKWB?xg2f7Ep@Hd*_g^fo$1a-m#
z9Iz3vv)N52gqkUT2xtC}u+h>}T6IDFcV-<+=jVIZn|u3vCfBC?-{;~gZqviVj-%;s
zd7SEMdGy)QmHudi>XTFWf?2A%wJaizD0ABiu;>RkBwLq_$Lmg#oZl9ru$N25?N`DZ
zb^Fr2N|ZOZJ;eN&tygo^#3eiE|M}3CoPSQD**T-96L8t8bk(z6vtSQG!oAYEz=|7Y
z`XUeRzdOiSj|6-^7d2WV?GW;dIQxFrWA1wH)%1#$Owo28iykLP3>!FoTCI%2Z`CmW
z#-@EUoqz%#=>Kul8J&4FU;mWhLL^>4EK3E<I*>ygAoTw7QN}#r&Us_<bgli(R+!NP
z4Gq=sMdTmDr4{<;PISgmC#{?8Ku#2p%t<;GgADlP9sXQcSg-?qX%ok4y-BuanO&|>
z<9}<C?DqFMw8Z_h?q;wsa0^>10Gw6MmWG<F33D^}zv&QPS4}GuR+UYVS6SfO$dUW>
z32AV;MMCrjVxKMH{zz%guCO90oWUpN4vf(1&tH8vN$k@wu@?!i;4kBHw0#KuJ+o=<
z?}bqR60kFnuOHw{%??58yAWtIz?SK*6Mw&U3#7f@g&FR6gARn#ATKTt55Z7uQ%Ruo
z`1e=uG=MAO{Tq@Rj-iElMmFtf#&E*K8+VQKOZ8;i;SKzFJHE*yFO}+kxr<9@?e&F*
z_~1D+8@W7J!9Zo~1b2@=s*(vY9+3BlAljU_VdtT%OHk33w0$?dbkr)~+~$_&9~k6T
zBv=eDwK)s8x7|%rZSjP0g#{sKOF2aiREimGia_rB|9u=Qpm}Uw5Fv&WzRtt5fBk*i
zFMQ!FTPlB}11VaTih!E;J>d?a1{6)0Y~!UJX!NJuPF$FzY_vZTcsZjmoJOO%d*P1@
z_dp8>*9e9T7*%cnP^qenK1mA~lg{E*PxewthI!q9KNkN;m;s#nSil)L-o*osFTIkr
zkKjj1mmhAE-B2;W5&W@$pjfM__a_WVjryZ1cogES#(0Wlq(G2MM;OU1V3H>1__6xH
zeX<E{g#_=1>A05OP$|`N;*HD~kMgR|rQsI0w~t?xS{XN7gnj-ixrYxwbu9re{G>GD
zfh42RW&^c>B46LdSnm+ftyMi|FnJ9)PXqL`nfEQCh@KF$iXAYAs-_!Zp0REPZ7}fB
z$}RujuHyc=_{VqEPnO6N%q%9|cQaxa%hDUJ#8Qr@#hNqR{2LlxseVWY;@{Y_GeicH
ze$oQbniYw`(?!oBZI#W#Ki=EZ%*Qxvr6?hA&CkfI$kWzjx~oyl&*vfS7WFXB!t_I;
z?OOg1{J;un;h7HG)f$FGiS~*u$anh9PY=!62OsB=;-Z4m7@%vgJjB?grNb~nEEc+f
zzLKGVx&+n#Bv21Q52%8QA+mw^3gf@O-&#c)M`Z$>1f)%3Fy_q;>_F7cm{vE6`KG=C
zc>I2TaiOv#i#_D#tNs<iP?w=c>{Hju|BKeGZF2lHVPWLX-S>m(*3aY77O%KGZFOmZ
z97VL&HvL+C(WA(g(^7*e58>JUy}k1q(z?2?M-+J#%rsNGL%ESP&kl$~W7JRa?S829
z!Ac)zf%6Xrx5I5Z+0G^JPl#5TrIr5PX1_iPPklfkorQFw^$!?%H+<;{6wr*|-kISq
z9fOBV=mVzX_p98)didb_b>?*R3P-m{Wy}O|jSm^tXKDZ{-}{mT`df#X30lLd(aBkH
zp?+1}LE<7%MZ6xpk9uP3Rap|*Q}Su&y5G$saJ;dws3->L^9BK&90LR5AR3yu!9huY
zz&-)s=9NrkvwEV*+1xbNXm%{Rm4JIFh=yd|PESwC>F8vl>TFhYff^At7z~b0O{FQ*
ztedEwt0*r|luhR{_nzS7=i<_=AS5QPFnKQ`vJ08R_zz!tffG@U*Z7A7P^^Hsd3}-3
zUYK{+MGfUJ>A8W(bY!`;Z?v6Hxv(APP=4}D#~h0-XZNV!qu~)z-rL|)06h(l$iB!D
zi6^}b&EhX}Xvp~bxY7I{XG@ID6Q=+cL>-lg%~5UgI?RMvey?mG2nIWk-)-#^*x#yE
z_IqJVwB%l0T+VzhInL9{O!_I;zCqFWz;nU)^Ln1;!JX6`$A>hys<4n+18Cd}Wqmn+
zcYUT3ECb*g@H^Ps8w2d8?sotdPi#Ga3R$XA<Ji&Ks!f4sf`N%CAtdB$Y-%b6)Zr;;
zXrvUClrB8K>6n`r`3D4)VObXKW4Awh_INHKBB-1;>ITa{0m)$cMmogU!!|=YgE)a|
z#8k7b)!q*(RF|~=V>mZJb;1vkA-bFnbZZ>-eNOsbIRf)!*dmeuH&o+pM7;=95?|Zq
zE;T5d{t(YY#%$I1#>f~*b-05pf~CXbbZm$~R>;JSnu7pqc6Qc&Xz>CTsgU)rvso$a
zh1YWCm8(9P-x<_+)(=B!IX!J~ZFiv2h-%z?CmA0kP^_7lnBGoBTZ+_s0XUp*o1*Q3
zmOS|)Hjhh7OnpR#tthLUZ6RUdSx#vuiykKwrzvFcdt;(gq$)8-T|>hh2x(~wfI0)o
zH*OySGg$Rd$^DfYdywp_<pLz{a!BxW;rP(~!2XKrV&ip~U-}%>a22Np_H+>G^K#i9
zUWf?yvo<9Jc~#(4DQHOKt<F;E?jMHaogb3<W#H2($jtnp17_dhAEGA#mIRA)oF3`^
z2>h?(mH<g^Dy>$%G{p?SVnl%gD=yl*Jqi~>yYX)D(K49fs9Q+%<NnNn4d|iHQ!tok
zv&QG~;q7{~y?6_s*Nqtl;lf=EarREq(`AbPf7i;3A8S8+1*Y=gbu9Yfajlf>J}afW
z$1kj|oJa0tLnY$^rbM;t%C4rptUT{Ieyl_hcGN`2z3aY6@iAh|TxNe$K6|}S0{G;J
z-X~lVyHlE1uZ^_y_oj-t936pD)hR#*vWFrcm_@dUq8bp8+RnvD7A4N&kJTMXS|C!K
z5FfwWYCFL54cop#t6|lUX@FNm&(1F9n`ofqOzm+VSoU@>%?;t4V6Tf&bvxcDc(2v%
z>^#n_4VeEYoRJhESRYHWP7P(aFQlq)Ih?PXtWM62wAynWAPq80-h3XG16ZT~qOS!`
zfJ<ip;MCbWwEFTjXqgV^;oEpJ<f+1Lvh8<{-|u4q$PESxSo@EeJLNIg3|?esi!0l!
zqfGs#mL<wE8!VJ<Mb<(6@1?X;w@Y$0h2Fd&C!^PFU=Qkc8`=yyJCd(@g`Watqf=@7
zt>No3NlC%Ttk=lOah+D4#v>8n2R*(VVA`Aj1|}F*JkKHkhWUX!i8y6PtB;%bJ4#4r
zwM?6F23g{?d2>KdpK>MXxclV+C^<Q~8fd{OR0qswf?E?>W^^Ii>+yhRMx?N(3|<>}
z<|^l@wEJho5)0;)04IX{<lUWTf*)e$7SATZsSVKH`F0wMPBTx)YB4EO-156&a<UF-
zRncLs)f9A5V-u>2c91hE_U*;~S6^$ODH52(Xg1^OaV_#276m-7ldGp`Plu=b^3IE-
zz`JeTK;z+hR5#Qk+h#%g)g!;h&U1Cq1n(0TP;~T5*w4?h1CbQzK2fW@RG6x(aB^}w
zEN{MCUeKJqbeXQBvCzPG&I@w{w~-6DAM-XndbD#R6RizPp&$(0m;KxX2Gjsl86M{2
zP0tm#4I<;aN+E5>ix4yPFxC4ODGdwDN9$mT1Cso5elhh0Y)Pe>%IkC+k*R&pJ6O1M
zvv3?SAkkg<YkQ~CYfsw+F3<R#!Q$G{o3O10BY=#COn)4X-DD60wPE+e(9n>>?5_fV
z&xkBV;Y$vzn4mR)YDBiGz?ds8nnc(!X&@FeNcHK2o1crbPDQ`~UfGNzE%C-nIkQ=_
z5R5ErQYtc#{P;ZO75(9*Rh-ckc2DDDCEaAD*L+FmC#9-4r6r}@K;cItiT7x?Ah#y6
zo7C?O-W$FfNnN+nMH`^9BI=pvaWp23oUo}j5@w@~71nEb?NsHP)@%Ri?7HCC_6j+w
zCCc|~8uyrlw7Kkdz_3c4mOp4d^sS}qp+l($*a)@Zi{kvzt{@so)oGE7Xzd<{0nU6R
zPVh|7g{ZNoWo|&lJ(gFCI={KoZmwGP?N9V2F|45qm`Q2y+gpHn{1-rRB$-SNg|n+{
z0Sj+&*Erc=qi5G@$9*?~p#(4I7&g9yRqI(pi*1Ir!#Ad^Cg9~wh%$-9vG<Wfp>Mo7
zZvJB-Xh)C&QE(t82`Ywei3(3b1$PSh(RZGMkks@JKX||0AsiYSdfjzsWCZKE-Q~eg
z@r9W1*DLLTGDS@TNKz=XgaIvql8*Pc->V2*eyw)x&XlE|=Na*Hjo?B{1LCN<<5CwK
zG34xoNCkL(Z-hOnvdflrR>bBD2FRu+1?7KjYHFImKLX%dP{j!Nn_oI~k8g=Sx7i$s
z!D0Cj9v?Gt->a-hZ~F5?Zc&-9=1nB*#7Gv)C)inVVh7>gLd&Na-p?8R-C*V)v@QzC
z`=4;5p6NvTv#i6{R5GeC=plT)7a%G4>`_(CQ&?dQlOocEThVbF_W7cyM07gvSb_82
zE}qDX=B?%>ilFsccL3SE-H1A6%g5=r02gG4JT_ve7sgxAk8w_R$6a}l%~q>oUX3~}
z)(}Gw<|b_-R9aGEM?_L$+ZG>uB=8qlVOy*DtP}_fe5nx!z58R}->Bj|PL_V);)lFp
zb>E-ZOUHaK(B$xg-Vik@`jXQo>cs*BQ&Mi*iF8_(`du$T3K!k8G)HqrQ3+_$T+Ki;
zGJ`K#Kz<wAs^=z!plIzgp`On+N36b!RW4AD-9?%tUhnq!nx$W|oO5B3is>|!(5g&>
zchs_vhLuq5XDkF7#^2ASKB6i2Y5Iq^K$|6)&0GkWSUqcR#l&&NOC$C9F!}x}`5DH+
zC(V1%RDn4x=b8dZduFgTEvJbMPeI9q%s};wMrLckH1}zAY$9onj*}bQV2R&V4ul9{
z+36r#tbcK?%4mxlx_PE-pA(7>I0L{a^3#_8#22)_7M)zo1f<eyKbB>45zChm(%aA$
zJU{qWqm~`6gs&Q_wC}ve+J*Rex*S1>!DH(ygJH5o3ap_~zqWaGzVD|)5STdkym9FM
zWn0}!j08jcB^zKWyZt4ok|057oKGPUx%4-KvUp>#)$eu6)oA`MT+dgXChsui9Wp*^
zUhuz(cj81Nje7)BrGN^1ZG`ANkVj6Cd%&(7@*0Tm4*9ZLpN+t|FMTLmbb`vA@qT{y
z+o!W|f9)LanEthZ6~$CaQKx5{^$@+StT%ixtUQsHPT(GNMXvpsHin!h8o(8y&Uvim
z`M72wPFLF1{KFn(lW!)5I*Aa-U%8_TOj>`V>z`5$)P_)rDjMEP=|ki*!C0^L6>=B=
zEh7GsLkCj7nDXH6<xtE>4w^c~4pE;~C(~?*7SSNi=Vs2&%;-IJFn|LMflNozLLd;W
zX;#EOcU2tgt=~8jvzP6M=o;6EXG2`C*25SiB`BcqE;Yg@k6~04J~eB|J$HI1Fyju!
z-Vy1=Wr{3OPIz(sUBH-DUJ3f^ty$Zc+_&QC3ZqfZ{?F*MI#@-qmLLbw1>x`#dYmmc
zdG5sE;B7~ob^7xx#VC_rF?n!F_Wb2E0uSH(^fVc*6<m&{i^y$a%<fDbf-KG)mYhuR
zx3j1vKfvMH03bxWWZ0MdzsFI9&EaA<m?oDp2p9saY_1SwCLRE4o6{Ab0oo#9mPlOx
zC~Q&%v7Jm1y@7}&YztXhRF)LD8kWRb=1wl6o+q&<?#0>zef=#_w26H^{$o>9%68v$
z?7ve0EJI}Nzt<7mBE3fLz+>5dFB#J8`fz^(&6X#NMptVo<A-)&x33Z*Z?AjPG!|u#
zusF(El24~%X*YtyjyKA))QV_lY>LS^3{V9uu_wjorz$MDSr^Yk_gSZOvh<WvUaNe8
zuLbAxd%csV)YW)F{`hP<2MR;9i=^<*$~_#Hc_tc*(B-1BM9%J&vemn+2J=NXcmvLT
zR4nu;r29w|(zEWfk-qd<oxmhrcQ~8Y6M#MLN>|%;Enq^d+-jk|CBO3~%j^a~Pg)21
z-TR@jwLbwAp_2Z*iTFF4o9YG-iQsj@+hYwqmrZH^{t>a>1!74F(ipeiH4kNUK%LH7
z_Wp1<7MkgPcXM{V?7~<?S0R#~xj$lP4tvSE(CWtWVK^lK-@5TRfYy#CrcI(ADk-=d
zN>kqH^LVT0`Kt@&i)T07?n)aJn*H|kpQ*jq7hyurOQ{Lpk|W9{z4?5I@R`kP^|!Xw
zjax$QRVrXb@aTlHlsfL>%bP=AWV!j;jhibAzD;9mTdh6H=J$dI!+bT!<P7~J>v_!E
zphJAwOfzRt|GhF;V61J&f3R70)D9^0AYnvhJGr3lij2`p3Qrx}j%S-U#b@h!T}->~
zW(WNOK%V7ZvS(BPfaKzGoRrl!JxJmh>ipN6=n&wv_3Dj%J&(h8ka78S|5}|`KxIti
zN=Zzgl&AL;hO*g#+Y!gb|JKYUG;WgVERFkQ^1dRcIBQO4o&!*;Z(gQVtS_OHBk<V`
z`@(S>0jf&TN;e>wSTcQWke$8WjR#OvH`ZpjuEQ-lFIY8hL~=7)evx2ul1HdJ@yL>>
zSz7w=K#_YZWc6-phSPTb68*E}gluZVJ8Mp%HH%v^I8XT-w_YaJ5Vp~4eF+P@M7!LW
z2?CM}Dv+0G$t!m@)g&m+)KA^49E3>I2zV$}^$gKWxiU6c=HV@sM3SRFQ$|<6IkOq%
z|9+3H+lR-YBlav?X4w6vQALJAPoK!b^Lz2ZxXVtu%lZm{j~`$ButEIT`JE~=d~RC-
zHk%q;E;n9QQ1Gg^x0h^F;oF7<L-Q_~9Xs>8B$JAN=Zv#cg66xUPc`JYnaF*~<5~@L
z7{9HcA|n?l{-`5E$JWY*sTbLSr^Ww!xrRa8#G7f5#Q|_Xq<Shd=OPKYD}io1FpKkc
zvZ>)<Jm{z<aDbB33={pc-5aD{(@ZbQ`yY@=H;FT2hMJ{9YTx<a>2{d`Vt7>`5A><X
z-XnN=$B}7laP0Pr0MHQ>GQ857azdy1Pq;C?w+y(XynToE>zJlO*Yyf2uc+D?P`2SU
zMUIuK)U?geIrs!e;64wkL(6X(`84TIsiwsKRMpAW+%-C92q=@A;Q|v<Eg+5)gB!h{
zVch;uz}RFGmn`+2uNUxkR^Kk}4~n&HD@(B5MG617(Qfau8*UWB7|;l~^Tcv|*W|ju
zSm+{SId7l~qL|RalHLybc5GbSR_g?l%kaC|V2en&3hBNrF363>Xavm}wnaip+MIqH
z2eIc!Y~8u3Yhx`8U2|7NBGR*$M-xj$10((2sTsZ4MhyLXrUhC0_iTlDZ$ubh8cu5I
zg}Or2FpkT?!J!xR0%&WK((eIq_#+1;xhfHv#udcF!#aO=>74ADTb&H|>}>cOi48eO
zpCbAP>qvG<;N&sv!DbZ`sNb+n{pb&8i5N3cO?$Oz{VobLpI;`h^mSHH1tZx+7h~5V
zgQCdFUAXM#rL1-i>Z3MQMsd+IX4RgMLwG7KBH0r92^MQ?cT4vb#gBXF_2(P}<!`K;
zU%em<WzLG-#ua2J_<YOOyof_To`;*}9cQBZLQyL}$^s^uV*=LGaj!}{;cWH@MI*p4
zZJtYQg?7*FHRX$&@iHauyCS)lL|oPI#F;BZ$Sbx?56<5y)<0NI7QF86aT!MFZ2;PJ
zJ%J0M@!@9wWVYi06KIj<Zhw66aD!MqB-MW>+MaQhBbeA3HuCmHZ`Y;Q{P~i79xWYA
zNk)cm1(mMfqccCWQgfQWT##<CRpvhOfj6s^Swuk@Ps#vHcEOKVf_E_n`oiw+&GG5X
zrJsEX;*`B=Eq|gx!61FJ%Z=>ZFumyg$}acqyHU`KDxE9Z$Mc}0vR30y3D4Uo_n>c#
zq^&5Fy_wV6`=Jcpi!1ix7J57yk~pNhy-qjK`ml$<hK1ByL{fM#Y@4$nBwEi}&K;_a
zO1Efwp+)VKyeyX&tbM=|R4&#u+E}@nkb^fQoVzi+F+r=b%?MqWJ)s%;dD=uEsOtm2
zpghJa(4Z(Q_Geoc16yp_2e!g$9&`agHPStiEJrTnxOnv`m6qGGPd%UD;hiEni|B)D
zyhw#E%!7})39z-j=^g+SF2!PiYV{t0D$Vu<WHf5MxPwDXZ8440;TJIFD+$`~#nuaV
zS$2Z~CidpvjP(Fwe%SrbU6U|sR((^MjRdg}q<|a7xSNQ;_^{!VGc3Omea*>wyo?%i
zxV|VBVcMU>Z#W%yoHO%E=q_77hf)gTr;#N9@;u=XUQSI+r&-(iYpti3&SiATW!-Ow
zxD(_>7A~(63rAI++EM_6s66+p41C<zZjRBwJR^=<JbA)+U)1MM!9fc$nJ6lLH&i^7
z4L=z^hg}+^<piy)8w;X~FeA+`S#{RQmdjum3Yuhnf72hbF}yQ)2AGFBBRC<PWfJO7
zCi#aj&+L{iR((Cs4c#pY_=rM{9k)bX&;8%H^CL#dyX#O@Z<Q0dJeQ}?u%-T{5WTU1
zaaltX)Pn+BP#x$m<r;v#^nNso52mYcH)^^3rKIis7iV|fl(4x$Wo^p!b^yZA6=*N+
z#dVvIpdw{s!x|>j9#;l6=`u4ji*y=+XnMGD7OR{OHSLCfqQF87gg#&(o*cbIzt{bO
zi&+^sS4a0aCkGK3CZ%1w90nXItTtIM8J^LaYTqB+{x<xvHOJ@VGMm2DX#yl(I4<Xm
zMAbDqyxUr)&WZ3@J|I4vCUvYrFdFnfDamp7oU2$Y6TLM7rOnaHE4&VF`I#E7FKQ4%
zhLNs|-~M@$6%RJIWQZw8M}x1E&z3hT`}7TilO*#hfxY6d*m2*(?GjO6)Qz`cYv-KH
zP6_6NXP?SyvZ)0Wb}g1=V^v+=n4+hRF&0uXt+Of}4A+}rr4_24W*~$vf7QIIvumxx
z&zeO((8oNje*t!XpypCEKHnv#GAT>87<T=M9Yyxx^3_=y&De=sGyIl17qfUyOyykz
zty&a|NEAS;v7i&#URlY=c`Q9>G74l6&YfF*&Wsm5Pt5J<L_`kk*Q&ysva+55Im;ek
z{yhL-rzL<eMEHODwF2+00dbr$X$CZCVK~^Px&R%f<xh&HT-&ujrA}kK7Y+Bi@^(yP
zM9h}KKO&vBo<ectITF7bAN9+xIxF`kSjluofM>UMNzz1O^}ah11Igu`SpkvXkz$ra
z93{uiZm!*u)9hLwzRfpa!ijc0Jf4Bsn8X8$fNKCJevI`M(vqgGyt+|G?I0RQdL-jI
zmR>*F%$LhPh4(_;t9cK`-B@!{kuQ|@m<y|4WR<k*f?pxL?7>l=`Tk&Kbt}vy@vb!)
z*3;eeL%MxiL>iqfH361sC)l+SHwpQkrB@)K4plZaL;IQ7OPYHl+H}*tE^L(c#Era#
z3Ju!$)@7*BNKV9RbweU4ypnz~uEZJpHr1t7`S{!#dQ+CPnWf%o%CW7)Hq@+=sIBUV
z&*j3D{fWk@iu#v1XN<K0L4jFL>d=uP%b^}eJ~*Qf`?x&7crXXhQQa%E!!Ux0iD>|g
zEV}{&d?yVZ_pgz8&ruTTb@tMKXvS_m=mT<wp;V2oQAa{U!~2M23H*Pe9~Na*J-0jK
z`*wR6S-mqog4+|LU6(CzF#e*&R{`vCN-RiZXSD2tW2D=GA9GSw?Wy-&w;D;0_{xR0
zed7@*wj9vc9UNOOw`!zGA3(7g^&z@|>X^kQ6aeHcWr-PMbRT$U1%;NDN##B+He&Ry
zw|Qr%%Z9i`CB8Y6VL~dm5AI8hcIBMV+=~x(@23<3iVA5Y*|&-u$BVe8tbWTb>Y<gc
zLhINhTdX?b&?2$J%OH95*B$JZXR)iKK#8hp%TS~_-^+y1^{@Gnk)kVdBaItb^g%!C
z&%sC|GCIfcJJ6ec7T~S}Z$`%rX8KLhFcb6~T?1blL>D0LJ?*mB6nqx35qfAms0321
zW+&<uVJB4GPN45>B--}KwT;sRan&D`Hk9qm4?Q0^vA!F5X}s%3#Aq8GU)X+kI`kXB
zO|<K}-Z32b<Q*IapJWkKiA);63GV{<>fOKu9i^d-Y9dW1*dc2U+jebey9(m^XZ?$l
zdaL82e=*kbl~$Z5r~dkh&f%~j#^JHB!(sA0m1FNZ6@|aG1J<>Jytp1Ib>$jQNlYg&
z!p0jEMR!XD^z2Stg&7ju_Ey!7CB^@2cCsL_`-Zx1eY3;=7*aFBCFLG89Yhuls~dZx
zfkaMJ=7Y3OQc}bg*^DIoX+X7F+$tTa*Ip*ny}WU=P?JMrU)qPgd1F^TH?_=MdW_FP
z(OV<>#dKcsvMw>WC~!ph;)}LHc-uNoPG?o%yU*}pWBW7cU5ycghSZ5E)uw6hDej*!
zk_#;w5&2>kl>``qKtuc*-t)A<b2<vt>skIXzZ!oc(4*Y^CVZudVd*mDvY!>Z^IRr%
z-^&r>B%-K>wnS8gDeYw&p-;aa<7w<#B|9NBD}0=_riB2<jAC1qxR>;S(${Q{*x6o#
zAi}_Y;#T~MhYNWgrz3rkn>8?(%_E0%+PM#*I_{!K()NY;gTN>-dgw$$j&qlVJao}|
zd@Z2wQ2*H-#(M2(h~4oa2g<#b=SMypB}NrkJ8345pVtR~;zHiZ126y{3{oo4-=8Tv
zCGZFY&@{rHU1dJr0quU=PJ85|hk`N`(H(?^)X#7t1?<H6Vqc<`GS7a|e-((t2p!)x
zUO&t(%j}a)=R%|G$ju?mnQsw^#CEB*&#u9bwEi%$Uu8T6W>ub$Qc+jpIA9n194XJJ
z#e!*u_4O*|<yRR7j&d3sD#z7)TC60(bZbS~@0%05t{i2H-d3NpVBb?M&2Zib5~dNo
z=Rts-H38EO$zMAx2ute=lc>)(JHI91TA}u|UzT!^(fGK)R~@R<C>(2)9w`POqOzV%
zqhMT9!*5%U3y&F@&FWol!HFs)VkxcWscV$<XOE)8y?!lv*655wkREpS;EtQ&c4yEE
z%;zYMNOnc~5?^&G&P<o64LP|0z@qDsB_sW%6c!3n3aovBi5h@_v7Zb(++ps)>!-^T
z2{E)jPu!-7xZD-d1k8D)ICL!P=EeyC(fk77DnsV_&{Q5j+*(e_Wk<Gk0I)*G^cSDt
zioi5opP+kJU&8CX4t^W!t(o#VC<lpH=^$sI2jzvHJd`M@b*g#3<8Vw~lcFNH_Q-tE
z__>0>y#0o6n3O;I{L9GY$UUYRujPXkxY-Jq^HGf2Mm@co;;4hy@|YCwYjdByU$fk&
zB{7KhRzT^eYLNIG(XCQqEzh^3PqKI3SpSf<BSh@u>Q7`rJ(lBlYs|}mY~n`g@0&vb
z-9#AnX{6`EA~l(5%L<(oZCEq8da@4%4L+g$1(K8kn8#SdCi=k@<7jiSvNqA7kIw@^
zH*lVLfu{o)-peF~gBo*LH6UwjI~}BP1*n$+#CB}`)y~h{gbo-x)IWInoj~P2HWa$*
zgdr4RSUJs_nJ$1BIiLMWz(S(E4&#2cGx7+h=}9GdpedHV?rx^_X0HSXh<VBLsymX9
z3(Ov_Iv#rg3<*Z}b2ArnsQ4WzJkb34D|n@>`O8U!1h<i}cnaI*D7#*Eu|Y=BANx+F
zb_J-3#+1Ve-xPYi&1s2Ez@P}1@skhf0@NmGdP4}`h6_}EKHfAc;Kc3h8~-S-#t40m
zrcgHQfYE#>?|7K7{iX9INN|aIs~AmnjiM4YG?-S+L|@c39{hb9T6FBkbl-iy)g^CN
zu+C@w%y)oi#94`6>?4b6ni@Z;_-S%OtzFlVcCyTcM>iO!*^3Ed$=UHm{cEyL`y<IY
zj|xBVquA_n0a#u-I3hU4$c4I;8Fy9=#rWqQFp(GZHN#z!z`;@X@@NS^CMJga_D9)B
zaaUUuZ&-(pb5m(tMn-#$<2)oH!GJ_01f6*a#*O^%q=lrTQx%?c^it^~O>yN-G6?YD
zFy|h&cwCCM8uS4!MvlYUN!ua11`;~DVI6n?Lszor9tbl<VwJ8!*Ab~hMKN+xTm)cy
zvu^I^?hND_)9B<CZut7BN)=qg%wc*-U^AnYHC`lC^}&9iH8J`47W%6*RK%Mf2<`2R
z!{h;bv!H@#X7a9`!jj1iV@}Ydh;Vl}&g==}SsK5_{@OQZrnU*?GP^~)4xPXj8dv!q
z@~PgNcakfm;Hp;$nfhOZ+_G`;dnMp?+@+s?UxF8l>N*db+~2XrLX{bGg9{kX!Fc0e
zH{Cy?xcYuoIU)IUEm|HdLv?@M`hsw7B_hAqy%g!T-%fYedGqc)xF1NmrZuSQM6iq@
zH?ISeWfT%^j=GA7A6B|Qd&#`8f_3`YEd)%d6GZWx2EPoh{7Sl%#{FF?@MaTla%dEg
z?xpa*c+c?XF-U5tua$;4PI0C?VqqyiQO^%Z`L&*naIeefQ175~8X6kf_Ticab96gS
zkEC<IA<MB7iWCUKAXKeWA^XI6kp5mwgAsl*>)f5}OU#`yiB$lsk?h&!2i)viDVr7Z
z;S6(w8>b?9jIH_0*Yx%(7{S6LV75?u>@<}7lfr!kG7h+{fIjnC)>j|T5GPq8Mi_l?
zE<4c_E**@_DSbY`Qx9?ag+!2h1%#{IOe^GtF6w-lhg<q^y&&=%YRLA&JyJ~`TF1vd
z16MJd8ch{^vSm3@5dv-4Vh`=LD5}V|C$evKAWF=Z(13|Fsl~~khEVX+z3Z%OG~p%x
z?%HHP;W^Vl(^sje9Ex=RGWE9yUXdy6HuOl+1vTv_bHlEi@kBVYR7iAMDPAmI2_a?d
zOOELDZRd)pp)AbI@?wIK3p=hrs@a798<05s6M{g!3gq?lT3o5x97&SbX)W=bI30>k
zO$Dy0-$Jo;K(6{ND2WzbT3oynC})x%#`13bcC9cV>Ab1<l$f4}sG`7wFw-JEEmYN-
zfJ8(26|K44o(+l+$<GJLKc}#lc}6y9Uyj3|;8+{TqVkVWl|u{USNmuuzXTv&U;5#E
zksN&Xf^YP&x;8eFwY=x$)D1<UN`ApzTt7M{jOZH3G;FzpZ$uwlMO%~KIY6#xo!JJy
zLmnh9ZvlHKRZA$ub!VzwKtXqfzbkl|SX6f4(O~siLO7p0!j~{AH79L3vWL$l;xj#8
z@|!l^k>^V;UEgG1eY7Rago+9&zi}l@oZoB8(|-%9FsJ!({qfU)EsP%;neAaTm$cWF
z$)e}RMeGzFHa68LHWt=8;Phjx9R>PmM>%P^WL3%@&(eg<0Q%`|`sef?icWb5IdWf7
zJxSJ^-;+1%5?VzoNXEt|U^J>7BX#)(zPS7|k#XYtD)UK7ZJ+y#M!?sr0`PNOK;*B_
zfB?PZd@qI3;6-;2R50T-=0puGwWzs#uZH5fI1ILqE*XlOV35kTdscsA`c`qQ?;w%Q
z;SKn6lSw`~`}bge$2=yEPreup44`tN-9xFzYTrX!>iN?97f3o*j;+@NOD}?6#N&^w
zj`4~$q!5Q5@r3##s_SR6hgxDh-?+fhF<>CUp#CWru`pZVq~5*{y<|*_xz#>LT+v1p
zHCwL%FR7-NmYO!0;Q;_p|9Za`bvr#I(wR2~_GuNA-*W34SJfrSZ{il65U@pbTllo@
z$PwZ`0xK1fTGOA&$B-r*TIMEEa7DZmwgDgN*L<;(ZC>*7ewt`j`xY5&<GNqgkYhTU
zVXvdAsv6pJwjsyYtH;oUA@_qu)ZGb+UKEh&b$aE0)j~)mfIWBP+nSj}W2$dclrkoU
z(RUq&wf&IUV26>@GXV1-+IR@4i9jEBJhrz?F#=dio&X7a<Kx5acypat2zfw!H=%r(
z9yUw@m7{u<XJt?J1otr5(H}LubzmMgeyh_w!A$ZEh(JNW=qJUJ8Q&ao1R0JfQS)(e
z=o<NB76V)JmM^TKXXI9~PL!`9_^saYLJ~5(mb*`^^?Y%v1XdaGr%wLp3DA2%+3DRX
z%UK6SsGu%P-AML`^~PlTVmpo_eFwo4CVmMKKB@vjp<Pd-5KEHK$zcb?;8iwGqKqWj
zpk&S*l<$ff&G(v4;|oY=Tx&t+Q47pNJo>mFEA6%6j~8}Q9e!k95ERAdyRSIea>%6#
zkI)VIbH9e3wDQZK5o#c-fO%z7dla&}YlcX~gTuaP>z7x(3{XsS0C-R$=cb)KIfR(5
zzth3%dwC9#UC4kzsOC3-<Sw64mO0uG*hd^ALws4FvMd^hHIDPx(Qi9-#-6R=8`N2o
zRQZ*_W2mZMTf;rFWoiON5@fm3u9LDhX<lp4^GaDI`N1rm%L!m(ChH%|7WU)DS0S7V
zw2|rYW{3%v&oby>e#JG0Ipg$_M@RQy@;!{opyEd08^<{-J^eIe43j%~ZY-b{)2u0R
zq8w8uo%MW3x{Z49Q@p}Nm8c!afi8_6$QX-f$W3k(qO#)%2}xg@X0=-_2KE`q)kRb4
zqXbO4eL=nS9bTB>l*c%uPO&TPTJ$ZF$}Xmn8Ep$C+n7Um*{cAbM3HRkPMC{LT^O;F
zY(G5)%gIJrUQ8&JudF8N=@&6h+`KX8Q*IuPiVqrei)k3<*rc{)#!NJ6Ym)fjg8}_>
zB`T7?nZ%9WujvqW$WrQp=j@&P%ERm%A0j@_#^650PpeMcl|KCbK`Bm<<mDpZmO&~8
zBJoCwj1^uGg>L(w_qE0T^(^EKliQ5c?&DPr%6zRIQhK01{64l3eUkFb{GfZp6`hB?
zmKM$0YYb+f=fgiG1y4hkYEgnJ-5)3n!9$oA3U0dORVd^auyL6R&Es2Yi7<?-8^oE5
zl~PR5BtXTkp@K*l=&B(5fEG~E0e_#a6^7GYw4w(5N7|rG-vgQVV;1}Z8OX5ExCL#1
z7mvNc2sZD}a40U~Q_WN=W5xd1!%p_Vx#;3)bFcmEInx4p06YQ9m0V1buKlic<;w(<
za>RQ5sOF4~b5f7sIE=DbpO@;)l1a(&`|jt~N_B-=X91{GAX>#=G(#+BGGYYZbxjNS
zPR!b9M@nZ=-1!ptY^cSj&s3LfS2(#yht6CA?8F&5KiKX@#-A0Dk0;Jb=ep0?b(+&$
zxZJZW#Sk<Qlai92ik|$A2_pVB<f0Qw6T~jIuS#da8DQGX58(8Aww_N&QNMj_uV)UZ
zaR)IZMFE%&AXL7imk-9b?P>bLb4%fv!rHCVY%`^*YyddEzqaORsy2q)jufT5uo!g1
z5$GGA0q|cW_V&(A@GVaJfe^mSwD+_ymWZ-40d`snmT)KNq3U?w8M@nze&Vl+^gi$X
zDevg_4q-jI^ew93kC1(&S}-VVPf>06dE^VjIqX@?)4Z2)yyMQDV0ZGgkGzSC!yjn7
z&;)RCVyZ==+7?U2K8Jgae;M4%VO9%skQsMUQ$gxg{CSh35yFSXD~pQ^&08*|5J}J1
zhAc<z%fyt>wlsecWGGRbG@x3vlfxl=+*|}38M>8jE!W1o%A<ybXG#J^Dd})NcTwep
zUz9mzlqX+wRkDj@3^p^SIcpBgLa@i5{BYqGX5JA*qDmZU73+&XE3B)f7M;_-IK#0f
zqCL5zsK+kri2vrpjfgzM?#<9xMYgL%_Pc?_zB<duY|?$(1ryV?UMk`~0BtJ_V0z0c
zEky)+?@+!H_z~}>)pP@u$!lJSuG^#N52-vg1Mp9ldRY+MPn2<N-v^DidPGy+c?g<+
z#PV)DWdfPd&Bw0O*Ns6p0R0x=rgsk8((y^Dfe!$FAr6z_OCpa0rZo4xqUzG|N-+%+
z)HG|c!?&o@QPXnA0%f-oJRf4sviXqV{642e*sk9SX@AtM?7)M<BuD9tndNFq6&*lp
zfLQDAXVT{6I`#*deOWKz4ZB2$;N7GVorKo>1BI<J;DWj{<uHt>pD&qUfYo!78DWub
z%cfn<6>!!cz!|x73H3r1+%J)Iw7V`xoAmMC&RXDE&0P7ellRcMNBSwL`JoiVEBYh8
z44~n9rGfOrUMn;b<DTpF^ak8?|Lix8Vb{Bfjq(_I`P{I|`iXdo$?I*W2M**SsVwS(
z36MCG!e|VvjU#wGxo|WlcQj$$vu2jCA03ehy&q*kLx`7W$=`n;E1#5)9_y=~EhxkD
zRN`)5@3H&0!ke)iQ3N<|eYqx!iZ)e%uZ(@iAxIN2;x-{qjVWJP6QQ0?!6^)>x2q2U
z|DhYfL#K(^1alv73K$R8fP_WgM9X8u-nya-xNB%IfrcXC6etgfyzeW&4F*vBAT??T
z2-jV!nsGV$nxrJE*H%=ohWrr!3+9<%f!w*ri2GUUAEN!KD1MaIppRbEbC&mA94(DR
zw?UW%k{YFxC-N>7Nue*4ZY0RGgsNDp-F!j&h^nCyf+e;=aD-X5ecpDC-q#|mm*!Gh
z6h<nJ=?c$sY?X;GhjYzlOEJH&4pmX}dfR2X2{z!~R@u?xD3Z4A_`*sf8tWTcDm23e
zuScGt+$vvEa`3P<!-8kD?s4Pqw5>Aa^sKtczfdkD=2t|}V(;)M9ZVXzG$YRL(0`P(
zn-J5j=F<ZU#1ol{gu|wO>^IEXFxXBDHfz+ek&KG#Hg9>}HHv9B%Hn~-MN5P9Dq!l)
zN+6ZbR+&zd4WS3RXJO;#zHASG&Jn<7HNTt9t={5X7VLx=@P*AP<<<~l><|Cvy_j@#
z!Mzq7MUd?bsnHV}J)`=*;UzjfW3Z8`XSg>YRdt<Sk-g{3QS*>h$$AXH+4ZhGT(X0v
z&rE?nwsjmsmb>6!SQqm*pdr6ODPQz*z1f_&7WvVpyYb2qSo-NL+g<0}FH`Lu@e1yN
zXt3~aFR0*<<tMJBGZ~Kj$mH9$AJH5Io(JX^(DDy!!xZF8e*efgXfIPK7(^;y&XzV;
zUXYXwD~$0>=@TKs+BYZY`VqFs%A0alcwIib5Hd`kg19JtG1sJ_;!vSU^vEtORJv99
z-LpOR{s&L<p$P79va}3S(GwOF0zC~RWmguZy&<EqeeSZ0G_Z*xpY$4&(QAT0-Ip}r
zGGtM;k=H}c8q9+ZYoB4gaZ83h-$Do7PWJDPX?%?mjLd9%mfVQfZ@(G15;b8HdAoCs
zj5U?5z={8Pm#xo>8eC=&tbWFKBeJ<3>mb*YYSS7yG{Q}2M`swFkPrZnM{3+q@ZBEN
zkWExI<Up)G)>B36lLXB;d)9yV!1HY3qhBx;oq0AHqNzPD7Z>x`Z_mQ2en&p`%%WGI
z?ksk_kUICi0aLXC=%DKDp4TJfDibAY3~4~E42Rh`xH0f0*CgJ4_gC>+BUF^*9r2Qq
z&pjg2Y$wM~rL>9`d_g%h7Khlj^XKEEZd9S6ERkfoL0z`Hnz_*;{pdrAAo^>iSPJWU
zl*Fn)|I}JFctP4!#P@s5B$PyvL_xvAj{D$y+;g6>gg%kEK{V9zZ{z(&7eVFqQ<%0~
zMAqNBTgj;lx~tFMiNjIeE^=}w3Mi--gMvP;tawo}zBQu5xShnbOxjfM#~zU<PVB5U
zPW07L#z*BNyqwwRlRQ?AN&we`!g@bb8)EiO^kwZ-(}Q8&cq+E$ZP0(om5G`K^DJ^;
zUVP!O2I<Hc6La4y!aF+>1_taq%-0Ei9*$?QnVuB|D!)`Iu^J4nmQCN^Gj4Le>D3^^
z;Z`UcZ?(dfPUCnUl{SygaDOOH!2f~6(8wqzKA!v@0{{Z=*f<;IPePk%Zhkdr28l_D
z|GOWeQHHnspEifUwRU$JcjNI2C(``{3#i;~&@ZtA1)8K~dq}y)1l;oC^+Ta_GBfno
zH2cu_t?MwvAVO!H4DTCryWzTzbtO$5k{LGc9aW39r114Xlbk>L!cW#aDyMgaS{EU9
z-qWhW3es?4C@~^~{2*^P&eC0L8U|LRlJ>X=Sk*+BP0Hk0F|N!zgDeqOCDbp<m#@Ij
zcg94wFrx+3g|xoF2fa|beNKiY$^R}hxc?ZVBj>!G+FyYLW4tsYu7n3U)NAZFi<Ihx
z_JB)8Cp30548>IUQv3PjSq?R!tV$J}xV{_B&L;=lXA&(Y=4RULUU%-x^>okXxFfsO
z#S39IDSy_!D%jD<za4zFio-F=V{@ZjvMo4hH*^-!61`JpMeOjP<jpKmPkNJU>J)$~
z(1a8qmq-GGeg&Hzynvmu{F#ScaK?1KTpK1$t$ehmhv+k>bxQfNYA}(}<SDH*5&H|d
zw)0mIHVqL`QBr<>VgOR6pRQri7wCDBiwy#HWNY^SWC`vwMyLn^zSle;-&b(sXQM_O
zcuYTA_&NeKFjREznXZWGB!H_f#C(flX+=#Wbj`u#4{nII9D9YiwhLAb2IsruM~c{q
zI@>mWnvZ364A6dF-zBTNenC5${Pq>xI%YKgit#n&eTZ!@ngYn6w1_z*6T|d{f?8Pf
zbY^#Up{PQHEK8dg*J!7=&4WC-q*lSK0<&sA)k`NXZATnA>Rsc~L?)U7uZcDaiOlj~
zF;w$?ZM*2A!>~0*62TYMdl%9hOdnXXa-s;-2yY9<p}#PG(_+oZQvL$3j1hV(E2+ZU
zLw9~Up-oNMq}Xco8)|2&M`0TSd|%Un*Q<~S{gs|Q8&0}@>umZ4k6L;%N(=w$iw~JK
zO5Wqjc5+t9IGAaKW3zSjI{&tkMVP!>I+k{{jeKkGI+wB0s-5!#ys@Js8_7UO@IdU+
zwtELzwOi+@dg5~`u)}O~OR`0qT%8FPL3^Uf^kPzb-}E-vY~Gy2{c7*C6hoV`5Ha9R
zk*EG_b*ni~G{W5nW;sIRF+ziY6LLC)ey7I|1Z{e8suPX&XmQn=$CMM0h%FH+e5^o*
zarLGK5>Jqg-J?w<szFNgp^Ob1sn<sn{I48udCtd#zwrC;6_>ccMcUcfH3EraJAg_g
zgvE@V`ip#?A^QCx!cqVcwTVA7yp@b|0xH+RmTLSuWnQ3wl@>YgFhM8~Y}n#Zt|r84
z=)4MgK@<>&MR>utm$#BKN%kBmpvJv$6Z-w9oUpe}$*7~duxIcDB>x{%Z`~GE_x^z*
z4blwVAV_zolys-kB_h%wF?2I@gLDia4JsXyL$`#|h%j^vAbmE^_jjFh-oX4bYp-?R
zpS=EE#2U%x6H=_<D@_fNKXTNc_v-{pNoTxn)SrKeH!n|82DhmZ)NG#O70t0H_ok(J
zZ?1h}dPPE@VpS8PLVF?he8122IZF#avyfa}Um;e40D&r<Ou>SO`n3AAhEd_<Gd(A^
zz!>c&>LGI;*fzjVS<DTA<u4OUT!&hVzfP0<y9{1HK@)|1t7(J7aOxIOVPy0@2FBLv
z_KXpL`Eza`s|CX}ae;9%Q<{`SU14rVi!VL{I_~B?!lgKyqR#YT<i4JED}mYFDW>I1
zT_2#&-3$c?D4ipKbWq%x$YT!3vRhrS2$FK{x_;-2KW$bin)&o+J%8G)|0DeY#h&8V
z%Nl5SI<%V8;hvdQW4ISTQz{g0zmibQh0(+m`3%pc^CClZ-LU5EHa8kzS>Xe6wfup&
zGSTxfffeWl>$p%T<*52X;0B<<EFMZ@(k;chA_nZ!;l-?SWQM>%+`;vC=gyWXGy+^&
zMRX}<<L;2?ru+E|J4C}pIKx5r&uP>qUKrEvKX#rZT@yEL9rbes6^HV^fcivAMc5^Z
zna&vDGv&)nMZ(I~I653<O!^bCgT%o=-_y=GV)bJqzi~7Iya465oTGMqy&JA~Ac8mM
zl$aUbGet@v$qWRXP8USCA1=-=EZ{c9#2WHORz7;7IF^OHBt+IDZ{Un2VM(F(ywfun
zH1;4}#U#&5_0|g|7fn-$WPZkt<h~;oMO7iIf6Q;)*7NqC{`T7F-56ocisDCE{J*P(
za?ALRh@VaE^jR&{2gc6^#F?+}_SFc%t7-6rdYE#%8$Y@o#whP%f85yEP*zt@lXl0K
zJVCm-5J0wecW+Gd*UJI}Ja&WI*6TOUPirCSf6TQdCDmRM!E9|q5X$$b@pwBp#0o#P
zD*!J_#<ey4)g6&7mvMX@i)k42sw;GF)66a(gUE5dx(ASDajykPZOO^W5zx``;1xbq
z7%qLFFYQmhZ(?<x!eX`yqSdbc!Yz~_Vu8gDj-IwsdPyu#xXU}bhNm!CTw6_`u6f~U
z>`?kh1|Dj_Nh>1Yj2&OW-bo_95Z=df`Yz<kzt960#vWRI!7UaQTCm07noQZ*c^F${
z{A<{KI0YxXKDRMjf(6?EwD1d`v}d~LJigd4K&gN>bF<;phX`H8H6*UnEZE#b8u;=K
zLK*3<lukJr=f@1_A-Wfj@N=@2LD8?fa-}UKnAqwDH>ULpS)uf1fu6AIP~5DKBPoXB
zSZr@+elC#va5qg5GK@~|=;_=c<SaO5WHc@C@xW*4wz4&k=$aBzL4Mt!A!?Sv8H1s%
zB#)mzU^)TnA6AU~Zu&Mn>x|6J3X*u|pmiasqP+)w)3-LCYbSAs|Cp53zH<tl-SH7P
zzEBtjjAfgELYK^*jiz)whsw?CbaS8uXs1%hGHJApLF51nd`7^){|~$-ec2Dp(FrgW
znU3o2f<l>=<J-fU$}a<h3+2{(6{Gdnoi0Utcyatf8B3k_m%sm=7d*xRjaRK|xc9eL
zFknSuf%;BQ(=w-o!;3*_y-bCSCXk^Gv_4Aewq(CJG&v0m0mWxgc)Y-XCLF>(!4FM4
zkjUsV%~j23#8E6C^U?-sv`7z>(8LbVyezipY}COHV$Wd9OV$RkCGw+JgzQ=8v3P8k
zc{Z)@?FjWJ3|TWoZU*%;?87;Z>Rw=^&{DAmhuW-k8K=C3Qm5JF9lmK(yDES^7~p~y
zY}hy>B(Xowm~bI^e&T~??1XimBPKQLVs89>(r+?y=pg_K7R8{bi0#4I>}6)jKPVp1
z#hT%Ei`%oX{W6DNyVGLeOb7inYUq9jNwg2?W^Kug5wcE=x{L_+XVAi;Osz5<*?=N;
z3a;26Y(2!6=>>Y4OmkWK&{dhM%+6`V?KxH^h^R_+naEHI{ZgXTXL!dq0PMQ#22;V>
z&wfdlXxPAoVBRhS^f(ql_d-vh@=v9nL*qYxKfOT#fH?V9Ed6`>5c{aC^ke1r#e(qv
zF2Ibi+><QUr>#>o@Ynff?k#i07)S5sZ}T6(%q@G=c;BucFSTh{YZ?QT_mBa0&CTfZ
z9%;8NSRxE4`C4YC)i#~I$_M@GH=An;N$IV=&`xf6eS8{Yi4$(Q#F2BOXKa)G0i)AG
z{E!t-2lG@J3)-%^>^fbeQTg5scdjYo=8v!ta-NK*)U@p2LO(OIr};&;ZFsHat`)z~
z&!QoN3=AdwRfiQd9`Bth%XY#L#A(HdwOhH0R@+Vea@x@=TH-vbZ8VlvL1Ti~1{C#f
zpJ`o0n5e(gp(^FtJad(UE@aPI)w?}59Gg6sdP&=PrLbEfz+3s9P*Eh{I3s*7wa><i
zRbDX!zOCdoY9rC!<MB2PU15f~tS)L5<~zx=0F_TVR~Dfh&_h=fbi#`ouJ_Mppg!O!
zmzi~)V8$t7h@u`)Ob9uWtTP9PFVOcfEe1^){Zf+^aSSURRjre*ErY#?sp2%klT%RU
zBAV}t{GI3&b$i@N*6DY+=hf`GA*W*1X~a{&XA>D-{siw%_5lAB8;MZ5kMFev`TeRx
z*2%*Xxx}_-?%C<UqM@x`AZO%YMil@eJc-u;tVRjW<&Ah;39Z08J(>w9aozgle^pL(
zx+SDlDCzjfX>bdC!pvM}KpQ&RKxq~^Q?8)H461#u8~Rb(KKD1L(Xva~sk$mPz9T+y
zNNM%#C<6?kh3&1rTI%zc$$pl5IzA>p8^}73BE@=*R1!A&Ts(VmCyos|BU3w+<}+i^
ztn{Unh;YO~h5+JF>}fVlL86{i(xS{4X>%Y%3n?ZC7f%l;vLcZl@Y*$jVp;Qjqf<b>
z`h)Dj+gZ#@u0FXqMVJ=^R5AHHZDYH|sjD7b5>#X+^t&@#|H#W;DUs!QbGG83j{30)
z>ePQ&A;w#A!wS=ok-)b4N#g1*+*{;BQ8H;z->4V#fn7VZEU}?dA63{5hK6N8V9yYt
z0?Cuyao|`-NYP4w(HqNiIWCKfcrnNzDVt-CGCvT(=W;HoAw&>oN$i9U=RAUdg!#{{
zB^=4WlkfK*MX!1+gMNe_?oz&Ccc@B~x+G5!`%5n9_%0y=y6)0-T_$-2*ezApUR(IK
z0HcH;DaX)Zl<mGnn~_F4E)O?YFI<dHY}=Q0(*lJnqw@trWeAn)f1ligrKk7<pn*GA
z#Q6sk!;BZ)PP-jxn%r5BozAox_n<4Rt^8$nb9)BJ>+wro$M@ojEC&O9*iT|c`6#^f
zKgyA!^%#agGw&a_F#XINh0!xPHe!(MOY+}jy+zMQjb5uftC3toxRr3Q3#zMXLYF<q
z_4SIC`36x~+;z1}<P(^`g0_q^D0^3}F7ov4eE-H`l?su#l7oJnL|6iJ01?M+vyP9A
zFFW|lHMQ$9Mj!QV)_IJ(=U+5;H~4~tE_YQOS3+tiJsr4*Xj-mwvcvXk0K|sV=)!e~
zjg|^cEaPp4ZI4x9M5d3@DXUYf+kV4b7PM9pgOh5$VuW$du2$r`u;*z*l&MXjjr%?)
zQ-#j@8J54fD-PfY+ic@}`IlOEs#H(6@4?;O)IYzNVuzbt-3FSI3<c;NY=2sRZA{x&
z7&N}pBG3owZT(%}GG(=7<DW4@EO@dL2E44Rr*Ds>4HKTqZef67Bcc0lo9C>nc-PF-
zH11-5VGAfLAcHNpdENYax)?j3E=CHr0eB1IcTPiEklX)@N`I8g5te;)e2g0wc*F<v
zad2SaeKI|s#dx0#v~VP!9U^<fU9{f4V{vZXCUx6=@;NLvIprXdfa1eENf_+=_wN-I
z6|bW3XtedVk|*H!Xeqt=%FhKe5NK{MzTCYatWx+^avQCmS>*f^R`y6ail&zt`wz-!
ztq|?DJ9g$?|5INjIL!JDlOh4sqWiCh9{&ic5k&Mso7d{?hXtmhnTr5{0nxVZP}y>$
z9^oxr$6v^clCzi)>5(%)n!E8$8L5vNgqD4Y<4^(2b~j>*Mb!%pEU!Z9tQsrGIAvG(
zF_g*Q-}FptDD{k0y)>P8#~P{W>I@ewRhv@#QA{wBVg~udX{7O{c!6$ABv^7>&?suV
z6!f^OUIMP+v#sU3{X@=VGkz(R#EX&fujznwA^b&~;qLX;HIs9q1SE5_G4&)Copn6)
zmmms6&Y8Tff!?|{Uk{o%>G0cV!~`v)=3O<qIIaM!sE%wPn6OPK2fl@cJYZ!2g{eQ&
zvjBrOjL!mKG~Us_(|TB?g<V$T00{{8f~f~?*UcWspY83|N#MRzbFHRtPOR_#&GbCK
zPvx<l;A$L<4ZqcXt;~1y`S@6YDCi9uz1*X&6X{r`lURi^$FOVAhb+{!_?Z93e+-84
zU(JI$lWQ-a)m(l;+LiOls@}HNM^Jp5n>WoR@6TP<bf1b_&Nd1*sfb0{$OWB#>6?ms
zj$B&~vuEC3rJaZ0cLMj}rjhM)rd-s}O^Ow@k4aiPDfkd+kTmP3DMG_EmSR$x_v-Cw
zXHz%5pA$d3GyTx<FDj|*z?`}o#6#;a1dF%rdd@!9!InnSvOiIwmHT!|dt6N48d5j{
z8YG0q-P_qwC;Qf#RZ=w>sO_ivv=5O8%5Jrf2&G9on!KPxJU%%BaIVfroXsRXkT`{;
z>@;t{cPvmLFF7XmyN=1&p9}0}RvN|YQybgj<Cm{oOccNIhFV_wv9_Ju!HOr-t$*Uq
zI>9j!M9=Ow78mi8tjqI7;H_i_l-52<d+xy@ghC|z=utZ}JJNHv`c&NuFvF6%Y^DpL
zVi?ibX_-#zjVQuvgIBcg<pEl#oQ*jVwfG-a(HmqAcafY|1R0iJKYmxYh{Xw9zrOmc
z*4qXCqV{K*gAmy}IVmZa%B6!aQTm1zs2S<}8lFCuFGVG?=Koa`PL3oiy9{j4p;l<E
z0m00rnE~MEs}@&cO=Qr`)Od@nxJmB5IH&VvAxivFeV2msigNM=&Brr;_hqTwRsLv_
z)&IEC*nv%~K3bN5Tj~U#U2MVl3)c^a%x|6>ZdJuBte4-K%X+D_wj&n80xE&cF%l>(
zA*eE{`;>{SL>C|k7{X&tPRtl7NB5p&GR)O?t+K4feBmLXxGUP`$Z@M_s&BA)TSBZc
zbbwws8)f9%ew`P?OE~4fl?hK+qm|DpGs*jD-+P7%#S@X`6<4HmqwI>AcbDb!vQCs6
zXcKa2tns6B)+-1k`TUJS#Fe!4qd)#68Qg%SsQT(Y`%Om9(AYGC$S&FEPU46!A~}c}
z?D_!05hIIH8nw|v*rR2rK_p^(=?0ok%&aGzLHjnUH#*^>**p*%kN`?gmp3O0^lLqg
zPup)|TF}p^)lHFM?-M7XQ9iW`Bih_+)8}}%bv4fx9piV>&ebwa=|!b4E5gjR=*aRu
zNXqPbbV#F%M2GCe;3)>sG@||#sQHPZsL}@Ct+vtXv2;xU#?@9@CPgJ$x;jG0=u*y@
z7tAi(F`fYCL8^{ZI$PI)si_c0h;qMv8_mFGGZ@!1t0faptR*=e@cLM{TV$d7O&HKQ
z`K)>66IGc;&b+97C7^zL-w?ViT`vxnYHA=0r(?b$8?X6F*Slb1-=;=c)>5fp{xv1&
zOO;aegVLjxRAQe}kRMtYdHd7a2#iA3*Gr-dcz(GeST836C_1@wab9|^xkMTK&m#qB
zY9U;DT|%1^b*M6kpwj^|W1yR@9hDs5hSTcI{Z9w{7TPbz3*Nll+4ucP0ND;W7#KJ~
zTJ`o5`$N8))I@}bg0U;fsIH@t@1n{K7q{u=g{X2#^r^f-C0WH`A5y|+GpMP0jX7$d
zgAo;K@;-|&Y45qyLH6sO@Lp9NO6sy=VSQwkkVsAOO9C5X#Xa5GzVVl=^M)k}&yWss
zUKhn}S-o+1l?G3ocOch4OVmE-REjc<j=I&_(q8^h3>#gwJCMnAtte+LWk;MknOIW4
zqbirxc~v?DtcmXXU6JzkZ@sb~cvGBIt-Q;k*+nHt;Vw7e+YrBd@k-Sr581+CB_)8J
zX;7l=<LD?P&&k9YSoQSmXPzX^EX6FSJ}R<Qvf>gMuAX!gM6Ht-Uu}`fkFvCOR4g)=
zvFKIItz{bP(ABo3;6c-dR8pguK{!1wp_6_^i0>&)y@=y2Np0n7B?Sdk038c&KklLe
ztYuAUkXuWaR*_X76u=$)z2)z&HL%KOyfT(`ro)`=_LN6fh%|y*{7K~Yq2*{q@T{U&
z%y%n1ww??GOKrV{=M>q)WTL?d4;aPQWc;xMiGmpr!Ibb_bXC#+{%x@#%*2>pt1C0e
zR=HH-zm51^`=6NX9?u(g=gCekO81l6KbEv3GPj@7R`h{na&}*88k%@LEh!Qu8ZXf1
z=H?W1wDwFB;_9Do_X)RnZ(X01;<V@uhbZV*M|jR2zv!KHH@DG(HZzW;u><AaR|`t8
z*H1$Tm8!v>SI)rU-b+I|HWITB%W9nV`q&>vKXy{wXF1edSoX4^Nnqg&EnyBMq4G3m
z<GUR=A@3F!M>B_@+#{sTFmgrDnqqTQ8}u_a@e+BGayY>ygeW=9v*(37v2yqWQas7!
zWxC-KSc;Z1%|4xG84G@5d+yCY9D<@(qw*!BoTIGpJde;qe^+s>QT+RHuS^iq?v#(A
z=*CfvMV;XR39BZY8i$#CE(f+hySYPyw+=o8gT&#iwruPXX5yG^?+JlH?X+fgfwRDc
zbt7%Z!Xk<@GyN3?q@(A_{wI#{zI^iKN)~0tE_*_V9N&OSLdbSyObOyE=%7PU<&qFo
zl-+v_!AE%CcbybkbO-vZhhUTjnfj2+P<=jZ^xgul(UZiWJNaDCh=_>Brwjb<{}WaG
z4{rm|H!5r3QZEeJU)`n1*_Hm4*E_<vEOmudnh78Q$mf6e@K3fgjsa8gF8M6&B;74R
z!jULQv#@*Y4W)*xB14x}e(d!ZH62xs?*IEmweuqyx3OFNt=*ut$clvatJs@XP_<gG
zcTgMQd0ha}_8S0Y4Tyre00l2ZPoYE5ssnuQr-CmB>O5D#-n+up5yD($>T-W`ag?&o
zngN<&P<lx<%50{p*I0ki2RFVS16`Z9y~}<86<zgpIdsuFa`shoM-Z<mEi3E3Ry8up
zS=BzUnzxD$jE`bchd*ZdjNJ442j9;(3cqUTNZr7=a3wDu41DJq0$)(YVgxs$g<S<x
zIQI9z=+WI)zt3Eq)|RM{*!Qqa&cXzB{o=nY)@navx9T@DGC&`Yw&BCle%fLr2%|r=
zdzgTK<hyc4*lm9F4GDBg#-hVSg$?)XPLP04Pro0O>|8X&?{HX5O39N6!zV$U3$|>0
zuv(bvOQyV_gY5D}vLt8@u8z@jgCs_7p*CuREZXOzS}3wc^gA}AoQ3yDU#0cz!CG@$
z(R};*yIUG2=mse^s>UYXJI}{GLe?eQwrACi(}`@f(%`tt_S7<&tpP{W87+T5zQ){1
zm?Zfq;w&#N8WxfF(=cR#Eb$o$8d6=~hX4)qBJkDGzuUFPgeaF*B6;zHRS(zgk#rxR
z+a3BXeY;9{7jOvuvw1of&<DG#q21q8!<n@DiB5`aKBwg(rJO<poDKoti%0DpdzbR@
z>oMt<Dw8dN-_}enW73!QUH4vQt`BBkZ9$Hm2dX$$JqpB5R5(_0QVen$KRi7UYP{k!
z-Y-}t?__W~PEEsKqrd<2D4_h{DcnUKKr!ES!mw9axdN4;D=XKHV>);)0@T4%2i=`U
z$Hggp_`nXNummvz#kl$QjqHF2qYBZEoGK!9q=I(&&ssEe&}x!9Bl!bsZ$RK2v$4-s
zcl{om?d<;ie64rSH~BT<w5Q)7-kL{*<-?IN>dy?eQaH<fDii{;LDosbj*`U&4Ur>j
zLt%0{Q4U!b7V<p_)`Yh`-|3JBY>hHng;B%sS?j+k^RksDH<TTJPCR@bhh&g}zs>uv
zh${D0FK!=w)=6bSDxZf5$k%^U4~g8cNO%%SoEE(vC$%@-<ZUa4U+XVHvk`|kk!;W|
z9MTiaKJ5c=`Hs5V>f%H=4E|ibq9*d~V|?@{&Tg+ot@uzgo5G=il-Vg3SHJqY0<tH<
z8h$>w|C;YG?HL~DB(Ar~QN8~+QKjSG&o*y=Bu{F};$U=KvMMxkidT}nlAU^f*+06%
z8JhE=;|5h}LSWW>2|syz;NU#7i%8R7zMzh}0u&$#3S(P#e-MjAk50gZ{oHC>FY<`A
zvaFjp%SZHBVpwRk0wH3>)nk;Xz-|8oh$t_R=A8r)(_VnKcLQ&&&jqihxij8X_pfz#
zo8)>*GxmPooFH7-Xkx0%D0qyX+p|akv0p=j0@~GrXhdNP>g~k79oOzETNe8Fg~s9r
z1qFK2t>jT%d^%k$L7vZOE&ee!k$$}W348zFG-3D1P^uz|*ZP+;S9HhDQ_P{X-E*yi
zt?CCPs4JcaW)U&x5W5ToYrTGrz8drx0NA1^4sb<29T-0;^hFWcsa;|LH=ZhHP;FTM
z&(%(FtVwM}1t#;TjQmQo_cIllp|)pVYWQQII4sd0P+0umjax7?n$jBvS;S0gwX6k{
zs+cguKQa@9Y?^wZ<-U78ROs^;fy20$whtB(FK)vj>rrlqdK=X<=(Mwc`d6Ng7>!=#
z+*HEgn5Va0Bo7~;SF{b@;FKSqGLD6NQhitJC>k=%#fn;Et9H7>E8OQM@C@GmnD#<1
z5=%KqxD6`|RXlhDd6{mYpHhnF<~*!J8Y#uyq_%mqb#xVJU<wPO8d6BFOej<=mfLi%
zo!L&paIY1)<w0VY9$h~S*oP|H39~6th+VSh-=<Z<zBK3Y8i%(APw^?ACP@AJg&*E$
z?wo6Mnd(6k2I--a<Q9LqbX8Zhy6B1Vt~S*?(p>}fcId59%~iL`F*u2ELcw+J>R3|5
zulsfG+?I-2EueN%<2&58Gmi!*7X>h4@3kI!woquv@7g%Ot*0nSpbZI<zTkXnX!R|)
zS-(v>o_ty=CMD0~Itre&P1x?;NSr(uM8LTLzZ<TSK%r$-4yX@gul@Y#*4f(h`Tgio
zOoR8<ocRsQ2EK9e+ai}&tETr1gBoHqFCUK?pHR1dlOt^1pVtSjA7F8}pK@$A)hf3G
zgVfE6VW&pRLRSUK?2EON26_HaQ2Ex%_k!rJNBxNpp*+p%ndk5RJ5Sr|^!(DKO!@6t
z#L@xCPkSczlCo-(o#g~GzV)D(cabRDIFNZ@?5AqK49rg)R(L=>6O64qH}vu_zNH>u
zN1inGLn9+{fKLgUeD{<;@o;~;?`1>Kds{TO{d)1VYHTt7Ope|$57R^QD!FkLA+xq2
zO_4P-_mQ*Z)B*{GXt%fcYXWz5qp4o+Vr{{LGz^v5N32)y<0{WY+mPWVM3=&`kTvw5
z#G-#1$?ou2@|nE$D_aLfw0_H}hUQZ+Ye4rnAYETs++BN#qw}YPeo=$Yo3@V@Vcelm
zGD5V~G*g$kjxFV)=^yOYR*}L<7)LF>EGmM7tvvM`-_vE-Lw=`#HnFc!Gh<r!;5k33
z3emQG(k70momg4iBCb#2P{r=&lF|5HP@X?-G3-b98+3Z@@T#9O;OHpi=Db!?&E{!8
z?z5sSA5{_CYKs_00IdknP%d^-3yA-h;?%i*z3|&s?Az=qrdUw^u9)jt>iSgj&MoJl
zBa=&&o2>i9o6u2y%Ly?$jGiL;+`38Guk*V6T(TrRMCNEx_BT>L_!ow_<#|uTQ(b|U
zFV&mMW2rrxrAE7!<F4D%IIo|T1{Jld-aFO6n&JwqDLkl(tviD<<bb7<)w6Hx_H;e*
zu@_$OfQvsFWQfX62inS~^iD*?-j<vlZ@X;42YB{dSu<duF7rFM#@JGk3nb8LQQ&AA
zjpu-2(92<!(59W5xwZv{I5*O2H|uY<Jh_(mD+x(`^ky)$g;Ry05SY~d_l}KDGC3>r
zsNY<QQJ)SOy#K{+Ojk9SD&~ULMHxgA=(jFY!OiBKi5{g%mC=@-naRY$f^KByMN->w
zruw8T;NigsoCJ(O)(A873o(oOlL!1Xv-=M1ssIei54f{Je0YOVzR-ZAMt5&2(Jm{J
z7MM^gNb(D0B9p%*O_*DXUvBfWTRQlc=s3b%@1`4JP_~?W-`KKJG2d=#a#0klm&FoO
z+2|s^)HF8pdymei+SD)&xqLslOokHN<?Hkt1Dti*l~h;5%6r6v&Be`_HBF<WWatE?
zb@b3Pq{Ngmk8*BXg~fzvgmY&o_^t2WYDRsj<}@NXtqv>AK<A{f!)A++y_B`l<w8@s
zTs^PMEgDkXs5D@lONh`0?J9K4!iLXmA-aqT@YLRG1d+}C^u979XLFxNZ<j~U%#E1&
z1V!^&DTV$Bf8oB4t3M;{q>e*Vwn!Vzy&o}pIG|}7Q6pWj2-`Fp(Iuo+wDh`#>G<fq
z@ta^P)uApWfVGys-)$4_EF@RAN_Ku7wSaB*8Fw=)-{bpFzgnNQ!)W=79t^sEXr?&?
zNXB#n$X?IV3Qh0*CnbmP`QKr80J3)lhF=SUj5=x2vv-7xd#NQZ+|s{JB19YRPGj7n
zzEkaTUvKFcnkA$!fgx`|+h@MSm(O)eT%_~O&#>pWo!v75{mE5g2Vo79$wm)dK@Y~h
z4B3^nO%d-^vx{@TNtFzd_sYX7%G<a9w-y_?$x$Oj?HM(5&s`90y9%OKvdiM}rL|?G
zROa_M7IAGLlhpzi#&EJbq;9|#4|ML4q^GCvF1HG3if*7@y|L_z#AO_8cysq71O@P6
zc^8_se|o>diTd)wB}TT@f}#2a4bj~Cv~@%UV}OOMx8eJime3L1FpSjeuQwO1%BpZQ
zAl2wu*^B3zo{n2*?4uZ)w|gjGycTLPXlztGO6@7#N|(P=YsTV*M6d!WUQX;6fySY~
zTpAw~h)Yf9l>WXA=Y9WUC-}TIpVbr&W@l?JRaTl-M`o3jX0fXYoAF>P!XO!Vm6Km)
z@s^$yb@~u0gTac9S6YKg?ztgE=1Ew_O{T;!>O+Ga;3?!sU9j0u?`^8@D7Dkd_~D3i
z_>W@~j}UBu(=Z5r=o8cMP#<Mxo<dkJN_&}j?imf=7Y=~JlgsP4(p$~LMJc)`f5K>8
z4s6O{MdGx7GsrL~FknL-WA4j~+_D3jODr;KyA9%p_rTcW(wgr9aqxUT18o7)AIM7y
zl`$`R&K1j96RQN3Y61HJLwv~9w#H`)D{Xl~DLy{JOtViEM?f?6^-@7-DDsm(f@tLI
zmt$Q_JpgdvNmtWfJr`hoATSmGCy8bM9uuN~AK$i(&oxMhZT$VtjyhQIL2;kV_U|R#
z>BT(75fjSffLD|S<IcN`+q>0hH|fd`Ri|SqnWrOX-rHIKZO{$WlHo9gje787<W@Hr
z@3ikC>d_(?vMk5hb}i;F1sDJG@NL?xO+rje>|giet;`E%W(FX?RaFP5LFb%=LZLB8
zN<9hG9%Fos34OO97Dea0u&EQ49|;3UU~jZAyK;~xt7>Zq9=PwbuSQZU*Idi=F25?d
zwO^D@>HE54FDX;=N_vutnrEMz$wc&N@-)=3aQ|onFj%8<Bu{Bw@UJ(Oly_YwP|lwM
z#r6?$8g^{dm&{uan3+)Su)bGVon{6pxb>*t1CPaZ--L=^l2ozJ?t6!%N@~%SOOsI>
z(f{+MxA{)ULZMj%;|@6OK;_QB(X+I+RM2Rp&X9NGC0Kr`rtfQOru53LIz`$`A}<<E
zLR%?{IxmPzq6a=Vq)L7H2T(?6HVyr~sY<`{xmb>gO=P4-6HO>^C<D#-PsRNpGJZLS
z03keB(8IaA4=(s&Ak8<mhggbAEStLFrL6GVJA==-1JWT*w|jOsfLzXz+Txpe)Itjx
zj_~l@j#{E@6PL$&+Y#lfP_E6}qf;zk1qH1*w``yz4wm%7#e<A{{Q$Mr20$`pT=`9H
z0q{wXC~yuqT;JAh@kws*Ug=nV4v*ZL-!OVC8r2U-Gt6WNGrU0JUEatH<!-Xid{7D(
z2-4h@rcDJ?$8J^<b#(3BOXb($qX-RpN)su0iONl+%7D(y&8}&?mc&QR&Te&N3VUP2
zE$z^s05!nt2l=}qv6*#6H#n5(qNHDLQO2dv)AqvOdru=PXhn~4=32hN35Xj^nOpqs
zXk-eYXz!mq+9E(Y0FZ9`TDm`xoDUfYK%#y7?L7ly_{a)&xQI}|1YkpeWs`AEb9mE;
zgSbC`m%FR)B&Jb2->*MuGSY5ytAWGjE32e=Ba!=kGOPC7z8+JAQP9`c%ZYZ_+fuwQ
z)Rk=mGD(61y0Qx4yIYR+Jd^-$ZdJ=#s`^#Eo3DVkZ#^tmV{m(Lp561cuh4@xwJ4mU
zUv_BKhn<BM@9KDaD%CYyL;Jw4?3ir{Hl|W|JJDx1obTd^*|eW1-Zt{~ldxvPEA6D;
zyc(*G*@FApC$?<WbwSj>x3rB*EEjxvq+RO0sdu7#snHDDMmj7ZRjX060i2*6k9|{$
zuw1R)868NjE<<S??{9(wMdPdws+Jc*FJRjd=7~p=-==m(4~e<z<1&Z8dM-2s(c>cB
z73&F*`%QrQ=zPqk;qW$Ls_@$?&gA{x7ehyyK*GSE&S4_Mf~=#k3B$`kf8dTNZC<{)
z*k7>+$l{)0l0SiHth=Uy>zsnPh)khO0Y9xlx0S(`#D_&!+P62(czdU67a}E#pC<7W
zZ7^oj?nKqsK+BQ`s^R-o&;ouVE2c+Xa@0$tyJMBdDEov~Iv#Z&^5gt}$l`PD(<Oyh
zyP-SY=wAPOOT_-|in@uiO!^R_xJcq#jhjBRw$R2NE*R^bOAhM$MJ6NX<n_IjpTE#*
zrOk`6EuLD8NF(R-ty-JIe~iuFzmVrqQrkC|UT9HmZ$%FV4AY(|t)Yl{a!y0H72gvR
zGAeYy`PILr{$s+R4{TAQ6T&Ay<GqI%K^1g!-5f^MC%G?Au@yALK(lXhgfq?UAV<yV
zbQD%>N&9idlq=qx6x*j4<aK1|-ms!JU250OvXVt3bEJQ<>Y(eE8eu8=7^suq=WA@#
zf^v?yP-5`RHuEv3&N50s7YHrC#$)3}0<R{28wi?cVJVx?d6j!~bN&pb{COWdg;e7j
z9!kTqA!Ir!fP%K$AA8ObTU!4*B>AW*<HpCIxwL0J>m1Rf>i)+#own{ps-G-F5;6;o
zJ4l*MbmrNPS_G7LEYOoP0LrSi!Ab7P@t0Y4dU|+&{?TWz1g0uGwc+M1F4wR7HFvub
zZDvPHPF8f`&Hg%q+&)w;w8v(H?g-94w-9VkCmX~np0V9Gz0iPwp9go+0b7pxGSHAq
zE)&>+NI#EqUI1bjY4K(5cmcNz=QbQrP9-J+Cxp4}+itY+uR|%!BB%Bus%RO^@e3Zf
zCLj3WdqsSgm_+N20+75C6|j|Wj3COGsXWFQ_<ozvg~fMJQ<VMHrl@)!O)Z*^CZHXG
zjF=V>Cgh>MJ)yOt!HM(nJLTAX_#b8J?n*TCf;!67gN0@CL8L8L^4F7*C(Y1uz{X``
z|J2ClZ|t!pkQK9;5Sy3?tI#QVF%&{KFY|95#UDt+CHnIN|DV6}OHZ^0^+c7zJDV2~
zXa&kdnBrePR4It+ecS}(l6?goq&hTP+dG50pA8_2Yl0%T6UNW6yiZ)izwBYNRi&rt
zV)J@->r539l+Yc^#+w&>uu^Pz(Wa^`$J-OFpB`;8<*;hL9NHLQVZl!rjPT$zIFMWp
z*Qx&|Q0N1_8l6IxGGe%AQJWOAUx&Y{h>}RycQL^%gQ#M0!+zZRVEG)_z}(-ooo6YG
zXrn|D_>9x2g{kqYcQ>U$S&({+-d0{QP4GoZpPWY?=p~jjkhAwlkAa#I_X49n0~+L*
z0b1h(SlD5jF!^yG);2FG_nj;@Pw^!xbqemh6weAB<?~Qv+LiQ!?LiMJS_OsD6Bb!g
zMV6HX?*mWSJ&%$$;COcSf><vjW*4yPo@upcwPBy9A{yWYY$L)=LCXmP+BM+M3JXBk
z+SU{I4S?OuJscN`VrNVC1u8N7i82qn+`u{-Wc&1p{Gf`^5(x~eZv0{QIlOa>(!|}T
zP=13i{GMvh&8Izt(ey@7*MOb5sNCP34=>1r=z(v`+^rkU=Iz<C)Xkudsm%}cSr@D@
z=9K5&_;V*w0e=^7kCW!XgY)1BFWw64n<lztrEV#Q8<uLm#WaKeo|8Z`#w{DZ%_S<{
zjoFt|jV=U6ljhpFxe4<YwF5%Q_tV^*syBT!K)i`1eq~%7jvK6j@NQa(j*IR@!s{1v
zu5t|RiVmmZwl^CI4r91>T%l#ZjP=dRv38t|@6l??H+8FvuXH_DL-HT(wWi00S2&Ba
zqNPDT?!vKZ6Q$qhsdxqiE?Y-`V?%dd=>(rv{5Xj^VkUgc>V7K9W_dm04R`(}ip$Q3
zju#>Jz#|@*Y1vHWSD<n}c8&wY^d!h4n+FA#{}|q4a=KBj2*xb&j72NvB-0!rQhsdv
zb3F)!1)#M=k@6c38;hz{3X;8FaWuN-p1f_k<}aW5szD8j=SAiv5y_J}_S-uzOia8#
z?Esug%S}k^O;v<_S)Wv8C+c?c`rHk~QCe}^V7s~qu_a{LOqzxA)OafjTI<qR<&5%N
z#=@iKE?xT(u!%;ae6oCda}iJ4Pl(-t9r?rt=X3Q=DPWcT8Wrsw??y5MRd&mO(H5E-
z6edaD6S{E;$G#7JhJvo_%YG}Xt9@18(qgPScMM2?I!*_vSlO^1#DL^u>P2JOBeF<q
z>YRt}Z+A6?J<lW{vnLO2leYI^>C<@$X;gG{IcKRS<ysGroy;z7hi{%=#}|e1k8AGd
z%xtubg-6bG9=lvY3loRl*yuMGDXi7(uT_)uXEga}cpIv&ESe}V-4a&IWF)(1Q+Lq4
zR)CZ!;B1KbL{rT?Pwfgsm#6E**4XOcV7`oeAgnp{x#c6bposeSw;`qNW+RHudLmcM
zl48F<h8U{!Qxn1OKt0(gvr4B36Uq9h!Vp)SZM`M}c^{~qMs?Z${itN{>)trpqIfrw
z=0XXg<i}NeWyeq&)+f<9Orb9_Jt~sFFs&~lsM;Qkab<U?Z?|;n-K*#&ixy!r!A79n
zi;D&*r?6eLrxBre^X>sQ)<WzDbzg_yNS*tbic&1ojXb8lMWtZzGbOD=U#Mk&eYA54
zVqZ6&_s{X4v#4In3DSGB0trZ5wj<(tYG+REpadmB&c>EQ6h~masg<svD#_Yv;&rw>
zt&6sO7F|m1TZ0d(;ES#>0KSu`b&Xn$kZv9)E}Zvw_WO2?*2)(%|KcjL$BKe~G%HNj
zEq=jimQ}#0XFXR3(-9n<Env=bcc7@_a*G@sg2$<(`KEX>>7ifIO@Lgo;Lb)Q9%<RI
zIr$wJx5uN$v1N9HR<1S@xwWf_f_qa!NeN$u`YBcMiM?xb8mF6llc0(5(bH4#?_>dw
z6~0UNeN$Sw31@~35G!?yR@8D?kgj(!bJk3izA0w6Qhuu_`r}yEGnLh|Vu9CtA^~k^
z16M&xdiCE=+RhD3dFrFwTP-%aizo&>GTgg!XKRa$2St^uJhBpTTqVa#Q7q&>@ixJ=
z`@et(n;+YU=}Ka1S34^5GmXiAN8@^XitcZFffV7(315m%9%(F1V}eXACM+*Gf>v9q
zE^-gCF7Q9*j*GuJ6}7d41rLaq4y4D=eF-YmN_X)e@3K|W`EL5UjIv)?iOP|zc%l($
zuZJHl;n)49fLZ?7hP5YBPJWIUPyQax&4)kD9*?}N!~Ysvx009S{Q^EFFFJ8cl4IA0
zDyTY<6>d6czQ}n?&@CZGKE5}tWnwu+HRHu!C&eL#moYixa9+&!n+%oO%cWcKj3hMf
z%aKNPV)1F_Glv6zG{jNGU|1*@#O<h6d9D<F)a37aZiTUf4ubv~B#i3L9mc+>mTxG}
zUv27hYOfbGnh}S8Ds`+10h<=pqT_-8H7uu+;m<1Ops{)2Ia(vj-{imdTWgR^b5^z}
zM6tSlIZ1W<;cJj-eYjq4iWNGTcdTe1^Ik=02pkpk-hPIg#uulX08!3oe(M*+sDE~G
zL(O_`rt2A2v}VH&PHk->iX>lnSJ>zfw7X|p7$&XBMuCU8PTVmJdPlgC4e#>$sXbaw
z@BiEMTQp{m)YUT6JzZaKueO0f#LSHDQ~ztYSfrFBkjVBF3j4xt3GjBE)&m$B=D_#N
z%<eGpZi{c~@DBfz>wo({WMVx=3&O;x>BS|novveNpv7fG&o4CI(7RuZ%iaslTQcB@
z;I_-jQQr;Ei*Wbcbn4K5jBa_?ushJj9tbkZps&ik7yIvEG#Gy(q@{_GKkh<(>HmoQ
zi&kg|m^bwQfao97U0gNI9Rns)qo1^25-UOy8v(4N(s1=k&d9#}_nz8lMb5X@^o;25
zmt@yE&|O^}ZVG;q#Y7$DwHSXkibeCj!+^5AIU<tz7d3E4Q#LKFM)6gnt+Wq2D!4Z?
z1ql<PZK+a)D06(3#4#=eg{HmafT&x)RAizv*@<@j!b;wlEMxYDww4!5yO3O&Ouy$7
zD>kor0&DLdL-AhLXs2v0XbIk^DoT2=d;1cVBBws5!*&*hK7n%9oHC~3s<0@K0_v1+
zm0}*SlZ8x9^9h9F>lu&5LV&nqnG7Rdk_%^v+8{Q!qV{}r%C<>y^c~kmSU*ZYY`P|8
zJwL0!zdcNBS;9XSp6`X#jcUz%j26b_zed7Em*+rSXA)T#>}$J(6*EpImfp=Uz<dvV
zS_sL*3ePV5XZVJtgEUH*DVk+{Btfrz{X4k#1LlblS)j?FSX8&OFZ=;C%;RBiZ<+#U
zdJ1L3Tmg(?>ui)q>FUTYK~Paq@k!rbp8*t+tONO6$&HADz#Elg{vqnOfCUDRW9VKE
zl|8SN-1|!B@rRxK1dK%ffpMbdW@<$-lWfHtL*Tgjo-V_7H>Vn@{)MOfD(O{nm;D#+
zzWek1OB}cNwCg-MCfH4_GWY(wEovV(yot-FyRCJ8|In!Yj|I*M%gpLWjRPM)i}T;c
zQ}DjFGu=1qwRW{L!|(U+L?>wlk^{-an-}flH;L5@wr+QGX6S=_N|SWQ5L4o!F*uQ!
zC#Qu?s;HUrk$;&UqNQ#R(Zj5l|BF4p?!gX2kd#?aSCR8CP>9f};i6zsfykF~3%I|(
zt68sy8OxvGf4jy}nt~0uk8v~A3$O+=S9lT_rmNLSq&F=i)X(4lIi53A&8{5#LQ7|6
z%Giq<TC6$2!bIwpryr1ez(<OS&B+P2<(1!IQpbEYYQV<tE85T6B%VB8BqtX)G^DFz
z;meUI=8*qTTTn8pHj+;DEJ7e-Yi~gU^yenuvWbL1>YZDvK~E8DcJ<Z5sWx%#Ix)3a
z^M04$Chh1u-@A~BD*XFg(K*&JKFb=1BzN0@BmE;%!nGF!{ZX`aX%91Iu=@sla~(>#
z-7*|~_Jgq7xdI-Lo(C3<dpu#6kJQOc@sJuJAAui@8<D%Y^{@TIc?fuBtWK<3-O}#!
znb3kMQ{vaJUz2ll<D6I8inI!(F>!GRz{j0YIWAokRKg#Vyntnu4h1*_$Uf<s`72ji
zqD8lSjDMyYyq4C_uax&+c=}Es{?zvAp(BB0yVN~cO!2RcU#r};=S90+J)FEKW1e&?
z;8TcHT(2cuIL2|S&@y~;vikg7FH?}6Pw2btBY`GRkw);}ocX{C<Vvb9yN5YT(^=_+
z<Y|FXpz+l%q?@tHnbs?6bS1TehYo5wQ2L*k=A48G)nP;rQ76PdRc-_pqCk&LIxNFj
z`DjJ{@b4$MpyP8&1e(ahXn?Kd*%$X-^`)0Nt1j8Nur;e~d~nxK_?8V;X_rM>plv*%
zddiM$gowbANST}N$r+7>*B8m-{wqGLB2Tc6?Q>KbE0805Y|$c*(=DT@XT5JvLj&2{
z@*-E&uZYX%NaB0QERBt0x^+}bM37?21>8h_8(h@Ke|5(w(B+Jc=K~b3H?-aREsgWC
zrvi$$989LV)s8rzHP*MfEt{gOn||q8{Pm$fIh+<<CPPMku{w9I^pD*+-RSR~58B}v
z70PpyAH*m{<<dsY+GUv&D0FNI3oOXN9l_13uAY>n_PsG0b1(vq&0M{rNeVS@Uw>dS
zhep3Kc&DHacw|u}d-uO_8mZc5Pf?w;F)!oQ<*mXftibmRe%W0tzUFtAvP#Q}WM(g{
z@H@d5M48KgsB22?af~4PY{q&li%&lP5y-vQ9!0_f0Pr7AEPsEMXG#ciCXTpLzbG9j
z78&YA)YkdXdZMU*{%(uDS8c3%vRUP;+qQS}Mh8#rl6;&E+leB|L@(iC=csJxMIzIv
zS(Huv^-ilM`|^ggxT|t>rP0X$G=JOPPvq^5nHM_C<c&yMKL2Y|*;`dF2eAy9GjxC^
z5t@0&8L}y*wQ6{7+y2b8V+7AkzXz`(g&J@trbwaaq%nl15l<88m`P;+`PoSgo{5@m
z07Aq5K8yIAqAt~B=(<T%)HuKI1p8rCCjlY3&?(E+4Hp6y&l%qfqjOm@5;Rh5<Ab$<
z2e)x!!`;BUqTdO2d8R;6e&X;cv~K?83y|htO1Ba*ILC$LbPq|tSJoZMqKHPW9tga6
z9?xh^uM@phVyUNt6opm4-#zUyGv2{XP)X`VRZdkDzljGMJ5#02tPHb$6f2u{Z(mUw
zU8EYgj7XW_Sa|-ZY*wc+>b&b~@MSb4-{u0hag|P<O7zp(vhI)4t%M>x9N(nIrS*ka
z%zT`Ef+AH56xl(pdvp$n#-x1rsw-r>jT(QFZ9uCHSL_NmCIm!I@v(REndiJvqiMSJ
z^2GjpUqodJ;`<MN*91Km3m}$3Nj3XQkZ_9a>WYJB;JeX1u9J<;3k_euClp;X+>;e}
zh~OqE)69cB1=|A#tp==W;2J;zgz_jL7uMpRJb9%i`+M>{a_Yf<UY2<^!rmAMTMSt^
zU(?pr-PMWqJnWv3FHs#GwSpnpA(j?hfb54k;Jb7PFdkS9JEA@AEg|JTz9j1eX{T-Z
zlvm$__B}*e)o?teUkiH5|BuQa*i^Fh=S=YbOP6&ry!C#mIpp1We5oiK?pK~RIsu+Q
zB4(46Qg|0C9qM7~C|L8}Ko8H8ygCTSAHVWn)NG@v=;7@?#`kE|_Yp;p<Xh6eeVPJA
z&AR$KVMGR(bS2u!1vmq+Lkd3vOPrr1WuRneA!qD?<XiwFCMdm$EBFIx)ytP=s7i!&
z{;nnmvacH|IALj-jV^ELD^-hjyRe@c&T(*&%#1@U9UR}#>Z@d6c@C}Nm74{ndhQcf
z-%Vd7<Dk3V<g+4Yh^~EZ7TCfK9$CVU+6l9<f^Zzm2Xr2^1O##yomZ*|q$2;;PC_n9
zQj&#ekJ(tiufH`Z+n7-#BaIQSowV5!JEgHuu7Kwl8rk^~`=wvrtapm!<z??{!0qcj
zjUcV+A$Bo*#J&LqUs`gD&GTvT(L}iyU@tt@l77c8kWG@UJ-^8nJEYuTpWDd}H~`)5
z_=<6`RsP+(Vhm#^H<fEhJ7$`wQ^N9yopBPC_`r8`yN}9dq@kwP7a9-Wo&O>gNgG)4
zqJBn9NyiyTqGz1>zYxho6pN4fER31~kxg`Q^1t_}Nby-%5_(FVZ2<k3<?!Dsj<vq4
zVUHE&AnLa7%lACB**9eHr2>nX0)B5YfvXI@yu|+TzS?D`{+@!7U+kbWo_B%$ii1al
ze|}s)THf!pZ2mTBU=3(}(e*cN$I4+))%1Ra#>e5>1f2F$*Ya>gnyj18A<oX_1=bx$
z<NpSU`-B#TPG0!FJujkQfu3>V@L5x$Ew{$t!tqt_2uXkr@rqDOy&OY0TONpW7aMZs
zI@^jT+`o6or_AsIw1Pfg{HTmimWjb>236yB>|uY3GeQ`(eOZ)o`oX4aO)=$bWW?3^
z;~({kGA_%z4<H)Xh%YHk!>hIj2F-dMmXl31u@~aLjL7U0vWYgqEMUWie}(VM#Hm>v
zDm0suPxge$D>;gqxWZ?B$w50txklgf69g1tG-yrppk%r6%l26*_$gmJa}B3ndQ!M}
z>N1>4&7oGs8;S|zw-i65L^)3VDgtqZl578v2AdC8FW2T9DH5jcxNHgDcba3$e+f)5
zU^dux*AS}g7Y&G@^=6f`rR_N0x8tTbDynlN6bpdW?o6;LM?#`Wcc7>?Q(PmC18ThI
zJz{nvTsvR$9&n7>cj$Jmg1KccJN7+Msd=Vzid`<zMM30j^)tq=teTH)rH+0at0`He
z@(q}{E{)@FwmCi~^n{@obySiTaBC<Ur()*1kFd??Io}}Qx?MK_0m3LS+tdi~VEO=b
zB~WC1K3Iv5d?`hh;rjj+MlL=?36z|L`IatHPO=>#gJC-)Z|4L}OU*xpL&@)M8puVq
z<MKjHTmp!7W$c6ra#OD~{Rn|NFXy>YFFJBxk5@0P<*8Sl363Psen~AmNfzK_wjgqH
zcpvV+@;-FiZ*11N{OFh)0OvYhqI9_JZp#k5`ftwQ%qZNwEjhBMiPf=?45_qQ(N@#3
zr2A6kQSa;~;P2LxqkQu|vYA|05BIxMh})SqP));3zwOT{6{oVe&FeAB+nxZ2^v_;S
zu)Vzh;|!*N3OSW_(}0J?I1Oxd-v>8wBS#2Xq@$uyC46~C4dLv#z==)M)qiQlo8`_n
z)suc#!f6C#3p5&L6h?v(ms?%(Ia$1D*sna#d|PUBXFiHX_;nR;DU$0RXbi4nLm(H^
zI}!VhI|Z|i0&@w_&E1p-7xPeRvp0S#ZA6`P5Z#s6RL?X^j4RcZ482<6@Il0{!B@##
zc3;@YB53sz>37nQ_YwwN#i_B=fry;@v$hi%@9k?{Y_1+6N!kggdf=T<Kcjq4gf<X3
zWe(HaeKlrmGeMC!?d`PhsGDLAmn?ZT<S6J+D_tfXQs*e+lkkEsLoUC2Xa{X4n61rM
z#odfS+|@WXJXqTjN)Ow&JJctZ>*3S&#|_51BFEl!d}?0k3u`uua;{-Y{`?sN)awue
zree{XA*<b0XbERKppZEQ>bBbInt%jsN&u?JzDsJu{Jl8#jke}O=8uy1$HmuLgfTIN
z7`j6!qr&GGJfs=%fti9-JUl@d&0~I;FYk}#;{2~bIABM9=g-$j!VMd>ELv-rW-|BY
zZdn&G`!P*AJyb~ejxwl|)E+W7aIZ)4lgSj_{edDUjP+q)PUYmAfC1#o=7vu5;YOMx
z&LIMK!lvZ-1Sk35cNLqsj{jXSGB|EnPSaO|9w+KUEY)AeyfTldNW>zHUHb(7UG`HY
zHIHzqAfk#%MiIuf!)Md4f^17?SsgnG30aM`k*u(GE)}E2JYS%jf>|~(4$S;MI=m))
z*F}L`zO~_cmQuQw)#quzlU)2l?WSk=0Y36!H?at0-zbvB-#<jE4ZR`1yNxD+_E}{+
zf3$b-E#Ac==3_?!d9ad0Yl4fDj7&(r@zAM}4^Ti8HB60e0u4ivk<$|@DVLf>iP_N0
zc*2UwN#G3Xd%n8cXk}OCI+a*LmaUaq2b-s7w0zZUAhe;^*zW8iD!Ji|{xb{<%XQzH
zX>~shKe#40xcYRFXY>%Dnk#Wb>`HPmXrG%!^!PQN`e8d-=LXz(Bx4%(^tXS;_Bb$0
z6g>ZHk~#bhUNw+C<f{nePh;0}KH|D}Do-ljXRE2ETu?{<fHPa(uDiM9mMv5XFwZS}
z)3d72=%|Kh2XzrE(a7BT=8@$~M4FtO$D8xi+C5jtwg~<G=r5Q+A*E}0y<3u~ZKeZ>
zMUP-Cb`S-u{Si<3vRzHiUSOrCeH9&8Rn+AWJXx<2>KgKtP(e$`4c+0A|H6e?lDU;l
zAa%_X1PzC^t&jZ!7LO>f0M`UeXL<92;08TS7+Zm9T#mg|G+Ab6j7#b}tM&)zE;B8E
z^#+~5k-Vlacf!$b-fKm^qZl{uW6<KDH%*Khdj3}!_4z5e6ikc8gXgW^J;`^%@KzE3
ztv2e2g~T9Av(Nh$W%0o~0lLj@|8}+N3N7~O@3}6Sh$J6A{I5^P?y(3clTpjyl+|^^
z%jYtQNSVb`oJDonuy-jWSJFvPDgSQ;4!|F3)P$f?!uIjYK^BX306_$zM|5WWYEbY?
z<J(Aysp{iCjIH<!_MUwE78Hr*-R<=GTCe4_t**F5(JJ;QlV{rnc{v$6;YZA4y3`#t
z3(<jNkniWVp#E*XdRrmK5(K^8LF;ST;f1Ig>1EP9OD>{*ubmU~bmy*D+MLBdNNF7|
zRWc2M$Q2>XKmN83wtvM%^CIOUlw?Wl`y76+l}@6E-ovt1NBBtWqPHfk5+&)Dy<;yK
z$;Cib{xE8s380%>3RYQ5$zDkUlE*-nRVt_q(aXRnzh#w*+7Hg((n#T%EJ_S@V*CYh
zRfN6aeP%Ej`|F?7D-ntvA!8(zog{`Ap6t4>E{^*K`3)~|OaGCPVi{yH)Jm{b;fD6@
zi+N>>315D)aS+}SAqO=A`7Q)X3}OT~%cI|axLf`|Or7;#RO|P*l`hGlOX-r5?(PN&
zX%G-;7`nStx>LGEx*MdsrMri&@8+EId0x*SFhAXU_Fnf|?{!^3ApdUY&UlSC`dIc<
zI4G&1LZEkgp$nAGVfvoL^AQnfhM@w;_T+jO?zk&-1>yzLU?Tev26q4!aSoV07PH|2
zoMShVz2FbyCZb5uB~&me{`}a{X|OJXx48wwk+9R@St)c=r(Ymn+2Ddt@UCF6BB!VS
zJ6CNnxyu(Brjf2?Vq|fhO&+wzTNLZ8OB#h2hz(g4@#WelyFEgz{9n`5zvNwBEOnlb
z%Nwk&$;p1pukgtSbVf`bdD<T;$7?75M!|&$LK8zRcMvT(#v@ohTXT-Cufb)7KS-`J
zJgwdm_14Zsc@rp}=HW^5JRsWg8)Kb64>GNnh%$+(x%?$W5BpgNNeviN@Fb3ocy_Je
z`R_|8R8@~JsMdbHXaOMpL;!EexQC{`$9yBoDXx_erI)YZ_<<+$KIESp_;8lc%rrr4
zIFxJ>IlnBkIOy*ap?mhkcCWqi&p4c)ouP*xMC8<TG5c1Q$2`S1rSHK4>g_CO3enY6
z9-YXEss~Si(P2DG55n&t>L+a%M63U`oh6Y>hP#ZuL75A9lH=rv8B1yWx&?o_Ulm-k
zKknl84pvCIBI|;8SpN8Im}sg_Tctq%;_rd1UgDPNfOcd3-VujYsD`f-psI31zgG}W
z>^^wV<arazEH+xom()%?Fgd*3HmE|ksEJq9$!$<v>Xos%{cW>xtt5Jo^M!%9AQ^w1
zI$}b8!2WX-IctTmaKwiXm_0(V(H|&$(*pz`L?s_<#meHumNPu(A%&Dp@bx2e4wYl%
zupKSsw)Zd6`g0^3x#3&{`%zfuF(2@R-5Bq)V^ZrBqI2>x>D!HOf{j<$M8&>}Pkb^7
zZ9{EJwl|A5j&E~c&Wt|}>M|^;64?ECAK4ir41M<J)7al+<EU6iLA^}>p}nXEfVw!5
zCyBPTv$HY6xz+>pZ~aHopc@Q7;tYEM5IQd^kTuPpbnk(EuJRj17wboV7mp*x;JH;>
z%emN%rV=#wM3$}f<C&Kw`FLZ*y@a(D#<wu99huSlBYTXg$95`DTgHmJX9bM8hS3)e
zkI>5Ba{Z>Bs^Y$ph=soQYfl79e~j>L3X@^@1a=rHejhO?08iXZux>!`6(1aC7tN6T
zi2%ve%j4Uf%m+juHOU0mPf9lIZ9Xr1V&gyBOVpY*rr=KeLgY{w0$$5thrJwqK&c(O
zBR%0hAdPeqd|u2?;As2G?#qe$UuZ>pZ!(3~nGp~|P5d%fC30T2tFatK6EtvJpgP)P
zB_)nJ8NgoAc=RZd9j<=NnLFRUrH^g2g;vu0O2qAZUs$b@*CLcIsgRT-@t9L3wu{!w
z!c&}Fqm>+1M6$vH>@6(ga%AQQ_72FXtdgM*-|k6lRI6&1&b@FlyXl4z5~C($Ls6d>
z>5@r_W2oKRlP{iyw|2NN=sC#d#PM$}LRg(Q8PcxorzT)~ec+_AXPnFA6u0l+_Hd;v
z)#zYsi+y8|^Tk@6K6cn;!5L)-G}VLi;i;{IE;{!uRUT@~UI3P84kxyfC4MLQxEd0d
zizh44I&NGcs}Cza5i%<AGvgxtmJk_oUixk=)ovVrfNeH=&v;~oVRyzn=d@PTzhR%|
zOHDwk5$>>%bMEe<GYL7wvBzP|q0gY!z+0@cG*A?moH$vI@@IpTZ`Ft2Oz6(V-&a?U
z$ls7<&Z>8rFHEGTU9IgMqs<<=r{jk8^Wl1so;?j5vN%M>f8#xdKbRs4ol@X^=e%rB
zet&Zq9yh7t@yQDAmEQT<jK|}$eUF5MbOsc2n?Ou5jzS$kCcXT#_aR@)wag4*U1s0l
zi8x|AGUL24djr$`wnc;_^JTv-W3Uw5R}B+iQ41Z|z4T7ZnbRb}Br_}hg7gu*?o0-*
zW^nXI5=WS6|Etx;^TCeIt^OeD*0h{o*WusC0zk_Y+;_7Ne?~?|?j?4fR+f~mq-ok|
z>J2hnHblg8-9=31x|04yFz&Ni6bqs*Fy1-VHy*8DQTj3oPBPd=6VDs@l)ZP5(;*U@
z%>{krK)C^nCAXj&Q4nC|Ms6oY4<R*M^kYk$gbEC-EK8~i=2s@0E=vD+uPV!x9w02y
z@a;|Au_Q)?CKL^7V^nbC*xv5^OT~NaZ%c}98MI9?n)ds!y^_~fc~xSZ#KQ#8osC9J
ztcj}EDHOReB?Fw4k?GC%N+>e3oAFy;o;*mfMAD%7`h=K4e!yGAH`k@NMomy7D}Ysr
zd1ez#Q!yDigRawFGvDyuHNwh`fMudI+ClsSQES_j7rL7CPsq-XfX{*y`Tn~h##=2M
zY^c}j#?c~rqTnmoPbg=fM0ebNFdwe<M~4;`G}=PgkA6*E>c)NaODWr3m-s}3SY#kl
z#&w4Yaa`@##k^)!lXh6&^9+Eb?6GJdp{@knEv-WYQ0Z2%Wn76T*YOefU+hsANMnrX
z#$+dne*8V)VFji~wk|gufb+?TYa>SN&Uv%Eel;$BM(}B0@zo~V#z)iwfX^BLnNVzI
zCSxU#FL_{6k6BX8u8s1A)5|52m2w!mJUj2J^mb;6@|<22?=`mE&9gO%Hr@{3Zo0yf
zmHw&;Gs?d;on2h}-|@qLqNSxx=60a^ZedaIXbDJ)Z-Fnel7_DD6r7l`R8k@tqf$g3
z4d>B$J9%!MWbv|IT;?=LU+*5{wQUIO{Vx|V2AY#Kgboio5>)S|85Ps72+K9eX`_DD
zLOP4CG-;GN(huzU;ondOrIe~10`Q|qgyJ21@<}6H#BZ=v*|omlp#%wn%sQ3tf#EZL
zb!2Wj8XQkZ_Ypb`PhLaon~$G8XpqPQi^&MM%bVO1ufrN@O@7-6vxonP(~=wu$67;s
ztE<mY{S}{qSPL{e>rH9$D-*G%j$J9uEcFZ)DGIE7Cu4R;z|%7?S}yVOr_DOD@n-fR
zkr`IQ^7H3ypz(knX4#vfbmZDE1bc0H$busWy@EJ5;ZRCLbkWe#BQBU9oc(OM37%Ur
zqLg`mKG`27u5?czTC1sQdPGhEL+4q*by2h%+A(nD(hB46kWdkwRnexta^>U~>I{bz
zC!?<)#(%%34u;Z=c4Op}=XB_1;vu9}aw7gP1`9Nr)l{5Z^%l~iMwPBRe^`v}UZ{S{
zrDG{JrhPu2`^0R0gy*^U0k9mQ_0T$RCurmX+hPNwqkZwTs`)*RfRXgX0{C7%hhT4A
zlhX{bxF%3$Tn)~)t2ylvhD1yXdGn(v42R&vA<0yaQD^INZ!xLZ4>Kh$d3o+Mt7HyV
za_@RYY~#vbKHl$l*_;Eppw?v)Jl9$eWjT+#hW?46%3zpb6(#~(oS<JT;UqYT60v=R
z84~}&ag7$*i_6M-9`7#Cfi317z&@1>K;hLZ^&)|_(uk84f?d=o39(X`DebVmhf-^^
zsQ#^m7l|xcMPaG=RqjcEo2MJ@$mAKpB%vRa|1vV3^`M)i;+jb-n?KNF0yf9a$6F8H
zz>q!HOdPD?6G-|^q#UZt{Fii>-0L>e`ztD3FISj-GXUqK4V`&7!lM92BuG@WvN;gX
zx?F?a8ikjd!Itv-MjDLY>A7qeM?}#zq8Lr{Ns;^=B5C85yB}Ap(eQ`48<`S`cWi=<
z22Fp_>aX~p{Zw*eEEvOoXg{B)VRgN_nA)`U5Jr4dFiUO2`SSyb&a4(g{tn*$r;A(T
z@}uEc^GNN8fGjJehmcsE(3wtQrBC<dN-l~tT7rP%b3rvq;5?sfB_Y)%MJHxwHZO}=
z-adD~!(Z>r1JUy&pev@<9KE2s^_59rA!wC4I?w@oPIuHn>7XQ{p)7jDd`^)xJhL87
ziGkh(9cg!N_^JQ0o5TsU>7sXawFePUA9JYH-7OPXsF02EqZzCAS{s~Co^8I=#22d8
z(Fw&;XhuilyG7>#bUg}-#j1s2!`Jbu3Y~zbwUP;ul7E}T<?q}0goMy}8z9jt2+$(3
zI=5zUsrmc_EaMJ=ZF2Yqb4@Hr*bJxO+ic!)NlT6Ng<T4a1x=g-tXb2HCyX}kOa4c*
zNmX9hU$XF7)Sid8z4lAj%IB{S){Z=ZT`iW6uhsp=jI`q4hMsD!>V6nR0|Vyn%8AR7
z{{|A#Zv?r<Qa-`Y%*^af79a!05%LWkuo|e3bdmU&+#;ewwG#6K9}lMJNdrQ^_v<Ws
zrV6W=CEje?xKl7QR&3U<PFeyc>>Fw@6q-0rnEgKe`mYm{9-M0;RY#y9L#OcGj#j*0
zT&@2%-t$&n$iei0fGHzrI~k@Ki>JZ9o}e=%@`Mmm6G2?8X)3j^BPV9|Q1b|^%VO)Z
zl$R{Bt265sFfNaiJtX?~oR!bF_2Qre!+t<|RU|1BnWtN14~c#l!&mOaM=S0NX_y`g
zwZV+6CGKWdUiPe2yw~JrTgubUExeZJN3Lb647HjH5^`cg{L*zLhOb)Y#zUXkRVGVs
zMy!-NK)xfgpv@N3G8Su5Br)(-T{-SnCxGa32U&8P$%$+exnU3><ufLU##p%~IADea
zFX0|oeCUdwx#xYu&9qECMmzj@p*@nZ32xZr>X2qfx$Jx87$IzR(0IRma6z}jHuhde
zW-!A_fy*G|%f4{YE}7<?Wr|m1rZ%-<6<_>~UYbB3q7*JkWlBD2wiCl-|C)@VFnk0M
zQ3hNm2viE`G%rzA9<QrDm`CHy&Nhf40A<$Z6_}}2nBH5H1?+7+p@$j2q(CS}_daP&
zYS8{>39YlyL3?!kn?-3AUyv>7$OpxV){LpD=#6zbmK7HL3_-FMydc0>*sFPH>{@Ia
zFE_#E<JocIW4`#Q|Iz0kDc);FvkS=ciU*K`dq+~)k%7Q{vdgj^mxYU&n1sXu+75pN
z4Ytq!hTBc<3`SpE;(1RfT}wGO4>v2jbIo>g$tb)upo$eixxiIF;HmvL(<EnQaPlu1
zDz!;-5CO0gea^Mhj+Qoe3@hiS3oJ8VW>fTYWu>EYfzx2IoITWVx)J9#$2ZkRu>P`H
zeW=7e_%8RtbR8|5Sty7YC}OAA8i0w!#^AYyD;}Fj?2PO87{e-EF=f)Qs@EV8)2M<&
zKU%?`JzkH9>Yg-V*f+Tt6-+3ZpfPCLN4L3UYzvcNaO{aq1D5~zsUVlZxBT=AvY+VN
z5dKphB2zC^I)nYEswDk10iR6X;NoT!!!yxkW6;(fmoXGW?<XcnJX-qk%~}1CrId@w
z_lI;;<0`+lxYygB3>HeeGws3eCUqs|OgAR%+x$Pw**kKl->&v)0k^uuGaf0EP^aou
z4JJ<y-v_x`9j)q?B*bQh(3!dNcyf4hEAH_tsUoToh{cubbpQJ*_z&xKG8jp|_|^AM
zdwR9hSstt;9%sTp7$5qskz3_{)P4pKVoNUn2LEMky#s=Y36r;}sec`SKe?Co>uaU!
zy)SQ63r;xWEfVtR3sk?ephVBd#5z{1c&KIMet`&w`wX)E0`=RApDE9<!*8MlG0j^R
z%wZ5GR{53xjkG+roprOW*wJfL#P@?6fh!xCfB;Eh1c{%5@{AI=@0W3rNKqM9%;I+n
zaAmJuVR>hEt^0dEIw7Nj)!hCXM<;?dk6&*N*B;)6k#5n3OyG-aXz3EG4Jiug<l3EK
z7lbuzfcd`{RtYOx&tuIEjk$Nvq7{-`JkzxbdLG?p_D?*hDxY`u5_fe35zESr2lVCm
z7pIV<p*7{WptLYeuPqQpF;X#s#2JyuEe9vr2<YaYwtffxj7UPQjbQK7lImA~FG;+j
z2O`)<+GHsxXBI~BD%W=sjTW7>-rpKt{We7Z)Juhda+My;!6HH&*xFa+7^TNK)bK`&
z4qq(hhn@|Q@sfa$VN-~{!dLnmE&m_DaP2}nD)23wUcabLwQ4mkF_zv!Au|HQT!GIH
zgs<--w9`7D5>38T;prV^zHqw+MctezANC$(c!`dYyLZkl(DAnOCU=oz4RXGz<7$;f
zM%`<0t;Hb2Vn|W#{+nVEaW|rsfZV8edyjdG+z~%LDmb;qT;zkg8_9vEz57X98oNbM
zr$7AWgu8Wba+{Tsva&DW<6euK_W-Ds##;liWL=&>czGjUyLlG-9>_R&`2t*bZx<SM
z){w4$-~J}vehOSj_K)cDzpa+27`fqiBls?C?OdhtbAe5%_{#}?we+YMb>fNB$sa7P
zC-hQ$p6P9fi#LMv>KfBV3!*lEhtD-WZ06O=pRyy7)g>r_>>HxzD{N^*lnvtSP8MzI
zKSg!=9F94WvZLzne;*8kl?XA)syFZpijHC4#4*0JbZx_6+gl_rQFzfHCySXn{$@wP
zoe>0gEW68m=41W)g3m5sun_^qpV7X4uaRXPR|RRvfy~hAt&ZmrOJcQ*7`?b*EyYrN
z_T&oo8>)~Rv=pN%(OUJ5NB>P~xO2@7_e1Oi0nU@FNpx)Sjc=+bNdxk%PlqiPELnTK
zn??&vup&URfnf%|m+l&O3$M$=q(E5gYvQ1F8dGH+4IOQvUX|d7ADXl>p=ZBK!CE@5
zL};>Vc}qevrk7Xe%!cfEps)2z6a8|W`rZ2_P{SFGl{BrS(dTq}l_`f};O*q;sI0$`
z)a-iFttZ)aPVwkKYc8ZSe2W~Hk5lKn4mKC}1((qd*ap^$#1n)sTurc1M=Pk?+uQ%B
z=VvtZ*MS|Y-q$dC>{W*IYHT+Ab!y%%P$_tMm&q7%(-_$Bc>|l_4*x2+0m_W00b<G+
z@ot`Wz4gY|eZ0?R%e+T{eUtqLY8paTsc(aGMJ{2)dg?mPs4IVnsZ-THWBzlQ^t@!=
zJ(h_Fu>%AiRjHdXbRYg^nm(m*#BqOybBHmz_xx(4f9`Swp`frit*+M7azSUTo-P2P
zS+!H3QR^r7Vb(yH*!)b`_Q@n4#N9xmtEN|+TlZPw$goJ%Bs7F_Xt+SC9^1X4+Kd1e
zJIst2(xSA5q}ChPh)P$3oxbC?whZ4(7{)}>dPJ1C#`;GnQ{zqju|Iz=*A4q;Tg_+b
z8y(}n0{gaKc^HWuJ2AgDTF-vj<k`g2+>wS)7i@u>gw)B|g&x8lWbC3r6W7e9kT~&C
zcMi#3eJDK+7w5w^bSku!s{SREHD@C){b`K5w`vs&S}*vys%6fwk`w8-eI|iW^qAcP
z7F{Bvdj|b;R)~dX_N=#s7<b|p90oPzalKZ$otlbE-j5#zbCc9c-4=A7fq$yw$?HkK
zB|hXpbfQQmk)v0EU7kgD_4MBipSRnYeo!m+6&GgG<*)!jGuo5(7q72h@o<Y@&j?_S
z03=Kk{zEUI-9F?}>*Ny$K$-P(IZXcXK_)&WRklBW(XkgG>hHvmB=R*BjUFd`cG&K2
zjBkR)qGyW}lkxQ-{re>ILer-lpT~te(N_s}AeCAo3Sv@hA*3g9g=WW|0n~blp0e4{
z#D<P;f36Jn56*ZMh1D+q_A6&iF4O&a{Plwt%@PJB8Qob;Dp}ggWT=EbB-8<jRe)_K
zM#Sqs&tY8X2+>%}*cY$D$<k?4cAX<6D53Bg=<{4_lrki;%I%xuNCiyXWYKBZB#{vg
z3nKqMJeEkkkU*oD7pcsXsRcGx*gSXl<nmoqRVht!!%V7<x@YG(KEf_#yD<IXHhSUu
z+3FT50q!keV6XVZgT9Yis*f`a_-!X1m%V%hp5Jem0KWDiZSpT2>yL2sgQf;>;`B&C
zAV%knj&+@bn~?k^R`z;^HAp@37pT&&+F<}AC3^H%5RoxoB?#P|inl(Kr%39KNT<GY
zmaLgbL{QPcc@1zs8#pd|9I10@AnbBEak)#Q79#uA2RB@E1!sHo##ew18uT+q`jP|B
z^_#wq_s07O(%fC4?reBC9=@w-)ja^|5I^`SHo*`b8HeGY;Mi#}@KS|`eI0^?dFAh=
zWp~j5=8FG5nb)sUS}3B6%gYr?Asrne1;HmUTQmyvq{cqlrQy$5Z|J8wrPPn42L+@A
zuoT`zl0zDbudA@FM`{59oVje3Zvpg2^>z8e?ZrMFh2q*bOc7d|s`x0+KB#d9He%zk
zN@2AnJZI@mnvY1mUDRex(F6udP^t<8ci$JZWOO~p-)*ZUrLLMDX{)Q!=4McX-6mQk
z;5DR*X#3y6WZi`NfW&B#Sk8Z}Bl*e%(LMqNE~~V(uZa!E^%vuXukUHvK!u0rdr#w^
zCG3KYa<c2|jH@v&yHW|>*05|)2J-PsNY11x!2{e-)U6$Iw-*LGHADGqn&&gF!-P}1
z>)#H<aj3hQC+!Bzh)I6f;vCmv$Z)d8u93>=Q4lOBTBQWseI`$9USew*+spd2JA@6X
zADhc;x>?OCTy~*-DL8vNcb%+Xd*^INdC+DaDjc)U{m0JJRgO#qJG+Px_MQFg)?I}Q
zsl*jXwB_8f;Ho__(ox+MJhn^f?z#4#zxJZ-&eku(S8@uBdDBDE6%I(kf}gKg7k`<*
zngBg+*MnVNo^D=O%*B`Ow_0yjy8#NAZ(JPaG%WfAaHY5L|L16}u12(et?9GZ4%-6&
zu-1pYj=YH+RJ~S({|>nV*giEJPW*U6{-o8c^$BD1HLLzp;T^!rQZbd5R1@VgAO0&Q
zCR+VCDH?)cf#cg*uOEW-^9dO-m@`+b5+fcL`ttMzW~w67&kw9oga9K;68;H&r42jE
zb)C%F@O8$pTCw#Rjl#!Qp(YnRC8;pP^Q5Ebs$Bq|)=eS%R4W=A43&$EB6UWoJ9)P=
zqaR5&zLO#mf&A{rq;9O!-q%Ii^t+r>VQlw--aQBO6`b!QyAoeZFN&74p{<84RsE!$
z6TOp#8U;)fr&azGGx$!q`(YFl_(6@%y1KbR9>@3lFjXtRs>4|*fv5F1g7V!<y+Ucd
zD+ZiZWv_le$$oFmPE*MGBrUpq4ujDH>ZZQsgQfP)QEXSJR%__z!YIe1`|w~>rIRR9
z)TTa9Sar$FU+F9U)mw~j>#_a-uD1~hB9M2<YEDALLplucsNyuxK|8ewzSP%6N@$eO
z7i@)W^6dB0TuKKMcN2#QP}zM}!882^pOG!l@UH12K3$iE?E~2_DWZUGG+s3w+wNdA
zNi8k>OwVhr{KCRlqGrF(Tj;O&8TQVAl-v3R6-f820X$SKRHQ}8-n>Blsa2k*VGFMV
zOoEOT%KtTF-=SCo&j+cKLqEIKA_$9GK}<|6fYvcjs&iEx6nVgF^f*C!e@J%~>~=kO
zz=<y``@%M%yJ$RZa?E_)^1TO)N+xv-+jv-ew$hG`v%(3R;3|HiB1m_|dM$4;^O;mZ
z6(wUCVWNnH_lE|*EuCzBG>$mSuklMN$~-R-1X)JoMd?uLU5)CaIH+iNxtIjopfjm}
z5ck2&==9zcdsH!@oU^!E{F37jB-`a6gQ2TuRUgJ~d3>rdCUK7F%&(Z}LN^5Be5=DM
zG3UyvjH7~|(@lDWxOd)=8F=nBE%8Bzov`tY(d4`7a-zg(K2Q=Rs@8n1eI<;wJl=Fa
zduZ<xIGgG?|F9-R$)vwFDy0b$J&*DO!U^iK_20BmMcg7*bR&$4o$OLmlJ+>17SbSN
z%1WI9>LSjBx|S$^L(FY-+&cfBSGW8w6WWo}DfOip36#mC-o)S}CJyghE{}pv82yXh
z&$XsjE|9dpLyOcTkZh@=hfp|$Az=wYE8XPaLIcBIeEI7*nTsVtmbM8X^AnVUPga`4
z@i<L}K#1<y2$DHbFQ~166?oDjs<;fesn+)+*(dt@;XMsEs{T2dy>l&=jZyjPwg11D
zF-{ePiJo3|XdpRA3h)d?d>ZYjGVI=}rW^TW1Whc&0Q)ZJS?GkhoRvd40961LX>Nk_
zWa-)_607Z=8g9)Fqbu)elGU4v?ZEQx8^_P6CeSORprr5P{*)pOmD^8-z^R{wT&_*;
zqC0(LquIweG4(I#+Ts16qPy*t!+0@1Mvia#qvlQ$aZOmQ10D9;wSA5uPkUz?ISOhk
z@r#`+DFtas`)dFE7=^w%(3|~3oF(z=!V709#F7v@(B8o&Pmvf~#Dp>JFm_?$)E&a<
zrFuzpz1Ui12~PkdqotH0Wz&$rY!rv<Bj(?D++;n=NTD8PpiJ`Xu-D?D@8%QJheHJA
z4`fwQYWj7dsX9HaKY>@Bq-q#4jE@F!K&|YF?=BTLm-S*S5-GFEg2a18t2t}hZHsOS
zAw=1qoyosk==9vg-XVl#IwO9_arp7{*X10}Uu5Cu>2`B>o4?)$r6?)QYM{7}c_o<+
zT&cRD8^eRhoO)^-7!*@G+r7^{%7sfiy?^<~6@DcZB!0+y3E?2hByT#@q{^tQ6ij?G
zHxt?*2<Qv=&N`r9<N6>oI35r#ZaCo$5QvblvMNV#JYwm89m-rR8NqT73ky$;|HPY|
zJ!J|@m3xwP-*KwHJ9;Z-{I535sQSoTDJBgVwoOh>I$c{IOyt2cHXpJ6ve@-<8PhRt
zyM*SDrT34mmA-f522Zh1TSNr}5o#y^spnKS@ike+&Mod2iTY3dhhs1ALR=%Wn;K9A
z`Oio_xS#31K38>FezIt_Lsl*_(*M23A;ss;x)J=)=tJ<fw*ly`p9m}x7uF20O_?#Z
zGIskG>8`_tg%*Rl7SViUp5H3eYVv39NUg>O0fPiITwO3mUleg^@Ii=e?0h+IlY9$k
zB}-29rax}tcs{)I`P)VM^G1`$h&XO~IDjnd=Bj}-+G`p$c;>8*356h+3V?j@ymIE8
za~G2*7a7ST(#F=Vwt{nc%W}nt>@juzI;`3~5z9g>v<W(cPG~Cu95oG{a>Li;fK2Kj
zCaxpr9pBuJ9x)ZYa4?V4tb=x?VorRC`c}B+7MM@R8|`G$cuW}oQ%3MRNNS#JZ09>v
zar&l#$3a=v4Kqt>TJHj0d7ZB$4*UMwBS+!=DjJPAe!4fHLl;z5h<lJiQ)d4W5BwGF
z<fV{d#M4o-(DDt}lNa)a^}C7F-al7Bl?qE{1bO&c-ays6FO{P<JK-P=jQFY;tSY=M
z-3T2ST#s4h^c4m1S)$enty3)J<WNQh9xQcSw-fN#zM+5r{{0_?H`fi&ozDeedGl?a
z%?$2WGaA*7hw82wK)r2+@LZYDdjW4Gw@W@DvgUDjxqe^l{mqyD!%#D*`=4v@unn3j
zYSKsT-)V#<Js}u-<2n8aB47&I#IjlS&d@M_QH(xho{D6WFWQJ;-X`G>e&fYdk>|9l
zRnv-C7;{kFBrnYrHNSZ%q{y~_WO{rM{_bjEqqjRk!QfKw{^oe7IH!JL4$%1|0d|Dd
zy=?+Qwt_LXPi#pW$5Rq&dgA?55c1xy>n|p^N-rB@x+y6~JP7Cn+2O8y*jNOL68c;M
zXbu9AWrzFl!wYtsmM4<u``i9KuPgY5)79pjN%gQz*{DxS(JYztJXaplDUH&$U1@Gf
zmaJ0Qd}|283Pp&=XTq`7HA*mq!MPY%>S=6BZqg8^1CB#ulj5By@f{umO&?tIVR>&6
zJ5Q5NeA^OBW6t^L$WAi(T=58zs<_Wcel4R8^&u|Qq1Bx<a_mq9!_{f`h=y*w3>XFD
z^rIbRb!6Z$E&@52Jera*tYO6hx%lm_&TRGm%f3~xhu{VO4&`6UF>JH6FE@2|Wqnr=
zM)6siruAFB!j6SsfdW8W!>mkP%7c_<?P}N)u-8S08-06f99uyu)EwiNtxo9~dJ}Sw
z*pMptayQnnov7WzUT@Tk)N=Odpwj2oi>(o&3iL*=uIBWU0I1e~RogqcFZ}s+YmpJA
zl6C|*1xya4|N9Iy0j;YrQ2^VBgys2goq#qBE;;)8KsNO)({t1>)RIB_>%KnNowgT+
zP7G*fT@t{(g`uIhWY31DdfvrENz?Qhm=7Xc;S7pD#n~7xZSO3>UQy0y>qVdfUj|CP
zFEQqex;yQ@@U$8DQYAB_;Vcqs-mempHkc%L=-)*pt4@y;Q?{75lx5xJsc+R^2J8>g
z<25-Qhp`y<g$OyHT5<l(R7LINH%{BkbdcOw<+VxLZIK!s=x)6c3~67zD3adXaz!35
z)MO#@wXW>Rllwv&$N(qd^4{b(xbt&D$G%I^ek8Txp4Fv}ispCDH(``j3}z^aJNy0x
z@ooY{{4Xj<GIQAaqzH+=DtlnarL>^Xd5bd9C2qc^leAg#(GtrEBbEh5@_`Jw=J~xT
z;&(9NP3W8Hy{Aw7zW{&Co8SjTdxq_N-uPUCO4{p*VNK{lM@45?G@5vo7G&hP)3?!E
z%PZskJy(<o_2Kn-(Iy=U_vshQTQ>e!Gv7Irc1Udb6E{sO)~@vaXqp7fBxXQB&8O~(
zn1zf<6koW<u6l&7BUC}^DR_Taoq4vPxIkE#n6^W-g@1o9S&O!NJt=)rXUsukNNNZA
z-xD<^vUdIdXzYQGSL}};n{AV5C*f}iQ51ny@^%D3*FwZ57Qo)i{iGoL!%bQeBeKzT
zAzI3~GE^cCJzuN(-^k+{PSd98SY~7by++u3TmphHECnuMB75G!)%W}rnmD^)Je08j
zi4EECty6a@1`ZfqPTs^0^QAFzuu!E+6|pW_bPu9UX0(?^dRUAtTD2)tQK_=#gFItf
z%K^RRi}?*m0b_gBd){-}B~Bo{?UU>0$<MS-6CWg;(*r=a0y{L8x{!}2S*pH0!sIdO
z1#=ELw-S=*!yzedynxISD125Lb`LIXZwx6L3=FU^npRzQS0uRiR+D*flRWgGB*lKd
z$p!szpv7as>$F;B;{OBHR1|L)yiIvXiVd2Q^1N}3G8Tb$c~0^Al4#==!wwgw-O@L&
z7`>T&I4V1k&=Df%@Nv?0!0X$U8$rn<ai61fJ%Rf$`*eSlLTy={oBNC{4@XR2$r}0h
zO}6(fDYSEtMpRl;79qn!oqgGXJL5$<w+MfvEjfsP)MQ3KXbypsW*4}Gl4qVby+dPA
zlg;X(NEGc)aS-85BdMppC*lF<5x<VUL48WjFBuCL6^`mE3=arbI<i>pj4zlA_C*8v
zXnpx31fPiG9ua%O#k0Z4d+5Rz9Tx8$JTGY7A4S{%d^INkv);bVpvo-ZI)D=wzk>Rr
zk|bPyD<ozOR0qQbuhl`XGI4YVCG?nZ$irMg_u{)4lmFHB9G2Mv00u@)O>MxqNkx;f
zFW|HAOo}a3y$-hFyZHAtEp^Nz9@%Ku>ud16=p4TaOG7J*(R4LH>RqYtTTZ!9nyy5M
zUhZ1=#(6_F&6#M%N;@7EyIGEpf|aT`26exSH5w}F9148rYuoxfnv@@VP%VqbB8cHI
zn?#!TfNitc_QD8GdUK=W&=z6~SWwUyTh2tBi_ze3_|MQ!va}*jUWPh7$66z|wyaJZ
z-d*<Hlrmdb`;hc9@H;*{v55J4E>Kke$box1;h<2RP>>>@htXA|zD64B+5&!3_9>?E
z$<^G(9#SHunmDi!XAfnqR3yr|$yug8?0mLLpL2SGxO`e?-kx6bEYRPrf??aH{xtgg
zseoEn2~yW@Y}WYA-@FKXp}f4|R{nUutQ?c?z*ttCA~;I962Kr^v1V2Q2nX9mHNJ=;
z7a8~>3(N~LYUr1aZ2aL`W<BK}SokI*uGKV*xp|Uzr2~sGb=5*k)mlT*>gsow5+xpH
z%{x4O^~g#iGL-&s80=*-A2273|0DIPPQ)&upo;~Z=GLyi1l-T3B*FdV#$R$Bb?D3V
zTED<FkVkPr?WQ}RzLL-X{E?&8_XJRQTb^X&vB?mByY(3dpv+Iq0(c?}@Yqq$YrX!X
zouYZx-N%y8#A8w3=qy(RQzxhS=ap{%J_M*-^HQDw2xcQHPut{TYcS3auZ#JONHv);
zJLp09+~1Xal59U0N*uAWM%0@Ioq4AACV*dX84F2kP%zA2!OhCtC{@F7%T=EFdI1^z
z$9#tif%Hk($n8$RVb$L3u5yZ*4%f;k7aVkjpthGa18~br8lCO`xf_!US~0Ic?-H;v
zRuqu*qS9UJH|&e-wjDd>Y?gHjep<(yw>_05U2ofWJ(+CX1wwK5-WQU>CN^w2@CA2N
z6h79UD#cVr1z}TDk?eGfe+M@xesej|bYc+3G?a+{K9C~uy);|KJ{u1d*E$%_Swo!?
z+>r~Z%iSBwlV$o{D_wDg`%DvIqPlFP*u`P{#SvpSgBo;9bB3J{#9TW<jm;}@haxmi
zwPj2W1Jk%nKqWN=q&a`!cS}0K8m$M3T5iK`PXV3HNgVJ23N1vfR-TWb1->{}alBGS
z&I(l|V+JPE4jJb4l9rB!rN#Xlg1K!kdya1DGc=y{#{<Vj2-|Lh*BwtZuV>?3z26J!
z_|=^7DKDN|`Z1Y>-Qj+>ZD)#i+-d?!vMX@;o+#7$^lHLZ>1%R1H(GYwJ=Xm7Eq32o
z`aH4q$i<7C5QD_HT9Y8~^~{NL?voaP#T+}+EG&$f<aR334hk{O4JPk@kK4;}FHE}$
zT<j66kijE;+Ats`h)Xo3GNmrfO$vdxVA_mB3lgk~ACtx_J3S!$Fkn*VM@R3|EphF;
z%)3D7TVtt8u+OoB`!=rv?l|bG7cpzqzmvL7$fc*vJ3z@1N&BZq@N0tNU8N`LXYJsg
zsc7NHY=skoNKp}zlMH%m#uR9(^pd8G{bGTW#&bGL&uya!{P78{jK{xCZ7`o-CZ{XP
zjhLI{2_H=$-=?$`yeE|VD_OeXjKO34X}dqPYR!h*p6V$J3PEFw#ANF*Dq#yv4+`rJ
zoGu&trgE8u-m@=t`zDh@o>lP)zHb;*3slRNr0kg41L%f>fx&Kk*1>VO&z2+8AB)&g
zhG#iLS_aIa_FzK8==v_UrvxJzLqbgdbS0YW9#P_SJ*;c=1*iUx!`qtNidS8v$sjqq
z9~*T>S<R{TSSRQxqwIl{O)|6<untv-ZYU`hBaXcef44=-7~XPtnT-l-MInZt*-v6*
z>hln$Gz`ZXB3Yn}86cKtSwsUJ;Jf~Dv>BN(Vg4TT0oRa2xVjU1)e3p|pSYm2q!--p
z`%7d}diEocues3e;?wpQlFHU=f*@4luyCd}ybK;EF>5_voy*JJG0sZ~mW18pbsFH_
zj|(t(|9iCq0(v=N;kU0uC$-O?frAzXHg~*Vo=<Aw66)WDk?!G-kLTm3zcCNxHLRf-
z#-W8`uB#P}hoBlGz8^k?>SApwi+m$a9tLCiS6}aIoL=7MdNy3EyG`#_lA^@Tlq5X+
z8}`JKT$M-gTIub<e5rD33p9{P_b*ERezVZ0oT?p(zXl1Oy2PW^YT2vA4UhS2+$U<b
zv8r9ko~zr1z9}_%8^Bc|>zp1v;kjgV-^gEO_ow>y`C|CIuS`p2b%^OYeRD3tTQ@YN
zK2mtH1lHxLZq9c7_U#^l%_wd?<w;AmvB|@rcD;}M3ne@7!sJd9|7vH%d#bM2jHO_?
z^F?)akqr(t!&1)rN;w@-QlAv_dxwR%j_*6<8q`XBB)O5LZs7X>%*MILd<_(9KW7Pb
z;j+$!hBQmA2_wmx1Kqd2)KwP@VKEvq>sl)V8!jvnjQl@|bF{Kfw7(p=ClMS3AdX|K
zpPQ~|B)E@0>;(wHA}haV7&Ju_Z@MG9VqXzz=VpG(u93w%iV;0==-DmcfNuf$UZ*5C
znm`itu7w+N5902<lUR>L71Tp!5WE(~&CnaYWTHH}vV}Ni-?mlIv5z?@_~$oGelAXI
zBhMiTISy#^blV)Fuj~Pg;q=>=uM69H_CS-bJLQLjOe3B9Aj_WNGbd@Ph3_1}ir@d9
zhvBypU(7dU)v^2%-u(xg95HC5ODn>nD$#SC`5Y3xwsN%@zboUKKAE*G&VVQ(g*Ux}
zFTq!;KZJw86N;qrJ~aMnSwcXQ@sOFw57pf|tDx3=*tdc_pGX0i*6U(S)S(5fI+IZ8
zEP(8Ep*dsLw8MVq8{9YE7%sew^i`=o_9RKfWnlORW^v=WcbSs~-6h-2+2Qe9#yr!4
z!YF;+^iG>!I(<ZlVC(Zu9GHRd2{5o5A{DJ{`_gvWc^#-!-jrjJ#Mp4QRqb+YBWZA`
zJ$Fn6gK6#k=kJGxm)QRGUwiLO;*OjplG3{EWzpsdm1M+;VS^*@Jy9D<SZOk`t_kvQ
zyI4iMs~Pb?-I}XXcHwLOusq{|h3&Ncz5tYAf26*7ub=r{LalgT&$J361wvR`6+nZ_
zftF)<2iL0_mE<5%J*dyOnal-$3>P-g*A@<z;Q8p^OppCQe$miS9yF;*%Yk#`V~-Wx
ztGtU*l^X_W%(al4>D+fo#+cDWI)as*x+6%`s{Ny$1>L*q#ou?FYHVG~h8$+}P5UVm
zg(AI1E@AlUK?~ENnZyL3?}P>3H|3S8i?gk1F%l6HhHKljVpC_jkiA~>?{>3{?|+%^
zMV2?%7GYGh&qbel-)7Ma%o;09KU4wbKHjUH@!yubrk9E4bpnvieA;-8&J32UuQz*l
zBzY-=K7Mih5%eN#=xS~iiSdD8Q!~f5GwxH9#{q8!LLdo`HN-7ysd?1S-3;F>9@8v|
zPA2KRI>lpHoih1s>zwOiu${c-m}yFb&cqShv>i9JN6-~=h$J$(EZ6gq(;NcOM|=N7
z>virGeUumn8a+3P!y+3Emv*O5Xh&2%%3^thA#2S6+qsO5pVUmaL~L4d6!<693}nnB
zGYh%+(H=L=0P9@eX#vqyX7ecye1_msSh(G3+SJ5xlVWUoro|+C^WkP5^rWm@f{OH$
zj-^JniR4o){d7Q(bV;#M3YmE8iLrl@<bC#Lq(TkZH9>fyCp}X?ZZBZA@*GhMn<w`i
z+-oa2?V=yysNnUq;zE(O7{E?!T#O&8`Lno&h=Y*$T_uL%N>+J6FnUe^2>jU7YO%LY
z`JJv6mrgn)`o(zt7$JfYogs)VqmYFWE?n{9kIt%UNsp4ecKLZ3ZT)%9cOOxFJU!CZ
zjVKQu8wp4iV?|9gC9MWO*gz3`gLtpn<ZpX2`z3s}R6f(PIrC#i#mBk*dHb-uO^M(C
z0oE1C_oPm;kWp4+jQ5n!^>JH6f1|tc(C(6Zlf`lmqo}N;>(^&|7bC-WvlVj=Hr0<Z
zubTt?q~7tT@W}uF@r%dB#pP#d>gi<$a7PewDh#PR`!iBBnU)N1`+GCbsqROils(^Y
zy1YE2%e9_JdU!$Bn4ijcDgJgO@qHl`ug4I_N1|uoEkzS7GOGxDb@WY~L-P&8(XvT>
za)(W5`ooq{NF!EHtm|xbpMAK$$>Mt5K5<$yg)vD{75>eps9ag<9*o+hedJkLqO6=T
zUTK~lVwBe`-f6yqT<f*yV@<6W{UD$fz7-GFiuTJ;49WYXkd07lh^1ytB+?}Qek*20
zHVl?0F8LsZZ4OfEKr}t_g6c>qDtF5Kl+3?#z`HGFZ~CDxDl5cYKN$*pvZom)PJxQ2
zhbD@{25npayra!pAtZ!J!79+sm_5EvmgX#<0oDlXsxF2BmAzkfz&IHy!w}TEw-f&D
zEdVuk<I+0JafxJ3&z)NTIlUNufP^*P!4xFQro=dODBH8%)=wyzfyVz*qlnZ~?%uIb
ztr^vAzRKv*NHz9xSnQ7EuLm4GJT>1KgGPaYgz~h7C(<|C;Vu4fdr#~6{*RipjMCaa
z$2w6{G%(<}qs8}aX<7gnx?&b}upgJZ*eouumeY_olS>A3Xu@usbc^1&VM?1wy1hYF
z2vhakc2zfOw<sEazz5Nw?3P@lsoCw#<1KYA^9#ge=5@{K*_m}ghq&`uN&VdX0J1XR
z_7|6x6-{5!0QD+Ud9~tAHJ`l)L~4LcSAxL%Ehp6fXw}ccfF4b~Y@9TRo#7oJ`}gTg
zKO)%hatE;@+xh%8=pa1cu_(~Ix<}iwGmX_tj4XpKje)F#8CuTulb?5anG1%cZy$?f
zBcyIR*}lFs5EZ0xHg89}CkE6UMZbFy|C=)v+|?gs^R}sjW(Q=_QB1>DU=0YL2Lq<~
z49$c@G{35%G~Gd^Ge5zCxn`R|4mL%M%MWPJR$Zy<nW9nM7@;e%=`HV!vor(jaF-+U
zA$-p7Hbe5Hnp8s^l=q8^JZ;{h>qf+2!TGKuyZ^=d%|pUS79+C~w8;4rnsVl+$NYB7
zvQC5ovx2C1Tz%o+5qt}~mtDe0Da<#topQRSV%aBo&WL7mv>!*<4c{=_p%nft8Vi=B
zEfC!JNjJ~f34`kWQ*=lcl&2kzm@S_t9ssNAd>EY&CBrkInX08qDP1MJeyC+CV;7y9
zKuCaz9yTSHQ%(#2<z{tdL_#A;z{_raVp}8C4C)gBUDD_KcB}hkl|PnKLXC7e=+eOi
z`U4kwl>nf`dL4BidY2jEhgb^6bIZ;C4yhC2WB|A@CwpKyJ~wCKNRzT9$V2SY<#%Wu
zYOslQrb&Q%V(G=LKd^K+EH4EUC>r#88p{}n?T<G)n7P^z@$P(^o~3=<ZH=xb{|KFw
z?nrj;WM2A0Jlqvb>KOspe{f7ptX*KniUHGdOhQ6nB;i$iIKgw9$K?p8vR=i@yIX(b
z|0iG$-`4or=?6>0z#s-N$D;Tyr0)J|llI!PHf?sLrJx;<P&!VHHLW}q8`7+XN;ywC
zBzrwQ+-@9;$vrI<CVO4;yEpA4sYVQkke=OjX1;|1?avnpHq~(z>@BJs-%HPn>RwCG
z@WkQC{gnjoI^kU%WZ>PLmSRj%35BlN2w5p$Jag{zfO)St{^(^LGs2$S9iBJyZX75-
zbRl2x|H|`{A3K+@ED45q8*h3gp8wtpSWMFM#Uc(G7bHduLWxOjK9xv1_-?m95vq~3
zwT#qqH$t|2^OyVtvsL%WaQaaRzR(~Iu(4@&qa%eWU8MUhvQFuc_wM&~N(ondfE-f%
z@!yS(yi2K(jSbz9dq3YFvq20igv1H5Jt!ZZR2-))uBY#WS=M77vd^jE9FC4-TY4jB
z?3laQVV+Ze9Lxn@%PCZMgn+#;1mTZ=w(it5FDVfl6ON7eynl%UxXBx}BYgTVg@aRZ
zK}|G6)!_QM>K#u)IlMe|NsMqq5Cf+fGqS9$dsMe|4jwQ>w>1d^`!2gOXhoan-IU%a
z8YVapSfbKh-D=j9^;@_-&7^E!ajoA<H>~G(?j@o7A01hGTrZd<U#-?QIDxexN3-H?
zh8ItttQOg}EOp<OWguB*9Jeh3qoR-hsGXT7=T-k33A}wftY-En**7<r3<$a!o&9}7
zdZI-lsiBneJV`pVKuC`c8mFl_MlML{eY*ly9rJ=k{uyLELOnQtKOq-ysfe2t0|=){
zviyctEc^N-?%5M;<90ej=SYvZukZJ+mPi)R`m-AGMW3fJ*X5zSZ(rCEnacYz&dIlS
zD>8TthVzW=DudA_zYE^J`Tk@YfI14@^y}@g2sENP>O)DwHdLr&-cF}v&TNIKR2Gzv
z=Z#2e_VwLv&TLsMb4U*&7W3wC^s)v)6IHVP`c^y&VwS-qy}qhLFLnjZZnB;qD9{K?
zezIb0aOV)Zzs(3dQFny0jdkd^zGr1>#cWN7WSL_qteadCU|zJ8BZhuxFqG^p+F48&
zxX7d7g#%BlD(jz}9QLK8UGrW>m<5jkyTE0j%Fk^hMYSZlnM->TOi7zq!4wt&g7yE)
zoE;F=R^-Cfx$G)7R?G!0eVA#x2pKm<)xnVyRwA+lSQOzEnU|JmkQSCQ5fwdJADc8L
z2U?T<wr4wVp~{)U(;%1H)q^=F-gpPh*ER*x@CGcx1MQ@`S&vgTPzksAO~=W=c&a_h
z@gxdb`XU;z)r!E%%6ir_oq(Xsp6qd?iGs&0)J2-<@|bn|II^o;zs?IR)1ST84@C33
z4ux2kXc0gUAyJS9QBLaL67r&1Sz9|C&?_hq*@I1qy>8d{W-P|^bbg}PBzOr^rsEKg
z!0){I`f-x<*_%?Pb~v_$45n1?RwtmTn0h&w=^6G<g9Qi9@*~vmwj0=*+AoKSXl`Xk
z(0y<hAQzgT>0s0kNCY33a^{i~lRcyvo3gmXNKMgq$18=}L%8@j*X1lT?CG0+hhKp{
z(qzxHz6hjO+^<Sou|+X)2HJDd#1&L>SohT=O1Ly{!Eh<V+J14uP_DGzkqcPy7EBk7
zKxU}6l5KICcyp&!RNlq!O+qDj0lj-T#^<j!^)qFTStz$+s9|&O&-WE8=UseVg}WvB
zWSC0)YbBsGekbItbD=|?PBG#ic51eqtoPMs{NaSwj{y@dW!0s5J=q7n>9WB)F&d6u
zYr6&LQ^c9UvB5Z8?K@&e|A>;eZf-ja?<8kPG97>d@;+LJi$Y}8LECj<lm)=vJU^zt
zvhD|yd`ux>+;1(?%PY`fz6>wr4AZCK%&Kt9=%4R(Y#HahnOpGHTt9#U*<VOJ`RbP|
z+Te)wYOjc$LEq{>24F4qA~=BLQ=*Q<<pVLaQkI8T{XuxG<NZW!>$2(geOY$BYm9N;
zwYN5N;o-x|1tPY<%FfVt!9NceFE<l6$!WPt+UoCLZw&6Q-NOGG2Z{Q=LqjruQa(xQ
z*dI-R7tITEOorb8@~BIbGT&|idzRznUp;GCDo@<JhlS-*_iL66_f;sLGRHdN+i=$}
zr?R@cqKB%_t|%-4tsuWB1s+ldGXfI9z2T2lQ5MMhu<dfv*n4Z+H^Md5)1Q#ogBq^3
zHJ`4H$jjK498UISj6-RAtH!5N^6|fm6$y=gdP<kML!>c}O5I%(9ndoCG*?E%o*{5d
z3s=7Bghv1@EEz^bhVAvp(I<ZH&t5rvrd8c(wH)hpg;{a^OGrqzgOde2>|hb)Z0ToZ
z9m_oy1@D_fG|3Oa3bog69x3;3`+REcrej_{?j5XSLn@nFNYO%4IvOr;@6TGQb$+a+
ztPVC7osJu|v%!fn)8v;BDHL`uQ}Y>&b-kPVGnRpdCSc0L_+vLkWyI-;a($W<LzEWX
zUXN8BpQSVhQ|-FPn8$XMiVM^Q)yVT?_(tpi$>H6vncdz}Nm75@y{64aX59qOC%@de
zyzJ!PlgD>IajQx7>Xn<DulEv7n&@AM5zv2jGQ;=&TCPIi(l*@>Z)wI4r;chvy>gF#
zf_4?~=K@ha28w(GTckVfWZ7-5RSLO;_+uabIGZrjA8Oaf)%Rd|4S9VgtY2TPoQh$6
zQGJRQx9-0DUVT&;pP0B2L}=ZS6HF2Y0IBu9$UUlS3Xd9bcSzBj>lFF_Hxt6;C^BvQ
zuB%IoiG|fm++Q=pPs4<M69c(1(x*kE)&K{)+!U_i!bww9T8HJ2;TWyUkI5|1LsK|W
zwR*8ZYY8&UkRD;Zxjr^3vABf!b(9o>n8ojlF}CV6;3Q1+S?^KKUw6_E>MQmQg65Zg
zdu2?wyIxGU_mshM$X1v_zlw?jh6aU?H%x8suf+Gvfl9)_F|^3@%8FLi@&19bR)qqH
zX-zIQpH0%Un9ieXrKiUR=O^iRuH;Fk^%oa6?B#|-&bg@&<AD^ji1N_dhr}9?!Z4~^
z#^;}CM04jf{LAl}Ud^zThD<_XDuCTsT5!*}WYz(_pnp{-@mS(p!@hf;mpNZO$#y!g
z;d^{>s#bnP?DGUuI6uVS41op7OLrkF;v_tidr`*o62Zb$MPza|^h)2W1eYZikyM|h
zW9eh&4**8t4!+x{6W#=?MEfpOLJpltkFYf-1&YFra%REK^T3@6jSR9U$>r}96+NEo
zM;&*7%~4rD>o{@DU0+hNj1U1@ggnTs>UZw^;zLsi%=B)$O|ChD4ODruiS2$4;P20~
zYQE+&7DHRMyEojM`2ef{lXyauQAh=dj3r6Snf$4e>&OttTAvUTnRG!OdVuOWV0)7L
zer;*ZPf_p*q{x2-dX=kMR5ZfP={jeGfZJiOb&abZg{x^x|9LY!dYv&HwG6KPGRGq7
z$B>jhUG@F(9$OYPr6KC(K7qZx&L++Ggm+W;gm(RVX(8=o2bX1YL9NAG1NNstvTva3
zg)?&Op8dr)xy&m})W&<105(K2ibs+79eTUVgr6tR>k9XM<y!tR)yC@(xd0hG8nPq`
zz1d(WxzDuU+t;89P1w3lnl*@w=vq<g*7AorZnvXhge+4iMdasBTF>u2y+!9VRW}~g
z-FW&|-I@9#@X4d)1smPhtJHpU8OXV-avb=JeCZ>gNi<gnbsLg4@AV=63j2Uit=s#d
zuSRB{a+fmmYXv(LQE$<X8%l<}>oH?|NDN)DdC0vPPxE6xJ_1riquVNtCHwSYgtY+A
zq9&ZI@T7M-%E&yAV!A`z#SnOstMxQUMoDXSaMeNC$}edclM2Ob-#G79^1NaS!_AOC
z`29MO$-ZmBVa&}kxCIu|?rFVo)16;OANu^Buj(<a6OZ_bC<xR`Gg)m$K4kz6!_MU@
zZM!#QoWeYpbT}hgzT>v8tkgK@_N4h^=+K0{pn)E-D|o+r|9JnLRsJrly9;pLYlQMf
z`Sk+-=P1-z;P+Czt_$%ACN8^Er}_2iz4qCivzQi;)3@F`BWb}OwUcWe4_|A1fOiqZ
z^H0yH=H*9LR{Iq3T)s3`I(Sk#BrME6s~kRU{t#{VsB&8ge1<NY%hQ+lqA%GPjfx@@
zp7wtKLB2UqYFz&RsCvt=DBGxQbSUX=>F!kN4(V<fN<ctr1_Y#sF6ovOM7q06x`kno
zMj8a9V~9OI@B8ij?e$}h`}fS6>%P{x&bWczR%oYZMFtD!CJ047+y^OePqNR$yT=2<
z%mcQJ%YCn)Z^NwVDl?mPL)y{H@tVBXFt@`QFJtC*sG(IB;Y%iVpmtCG<x>1{4v&^E
z-Hz6T`>oDzPX2gBDYJ1K*cG6p7zZGHu(LDo9)%HUHjStF&jqNgjxk(S17C;w=A8-&
zOJ9Z@<dY7*wQ?fDWaFI`0ig6*xP1&byDvJ5RcZKWa3>!e`I~I;t)!){&=Jeu*3#`e
z?Hsch*(6*MCX@7vmg8F6jEi0TP`uk*YkZPptmeK0sU>PN{(h#f=gEX};5cMcei6;@
zEADV#X`(>Fk#OatB)cMs;g6kf7)1?-vf4DKj|6<-JK9tw*|6qIY0Uu7e-Yfa&%O62
z3RnZ&<nhj!gY|al^XJr1%W3osH%oxZz>a_YO&%hKoz12Xk39y$O}g;RS{Aon<O!4i
zyGNXm_5|kc%13CTzfW4mbG?9mwQ<(_X1BYs#n(NE&DRMj=al3_jP$6XOuD+4J>r3X
z@f`;T@7E{S5e|MP+(Ftz#zR&J761Yd_O?D*L1g`?Cg(4=xFy)ci?Y_zE8XCk3p&?p
z?cwCgT9k@iN*4*=)8*~(itvBUj<+N^NSD5RSr5yU7P{|O@E#NI^UUjtfZLkz42oRG
zF-b9K=+q+Qm@JC3l4yBLH|vp0>ESa-l%RHe$bHcD8j4Ipg&17Ap`!cH^vZ%~v};u3
z1mfcRD$(DPJ~a^;NsdU4{s)Cz>2qQtA3?Vajz4o4Q7xtw6)3~61=`X0Q+4S(^h2yy
zQ^!qE;9hsb7lakwO<EGyO>mRfO|JLu6c5wtIy}eY<;&vJ{GN&}DBpn8ZTwFij;3Q`
zVXyX9laiyWr?)Eu@SWTs_gM(n$&~o%WjKwhcKm5iz<ALT;Q-4Q&j{I0W5RGUP$g9T
zn&08+PRw#lq+Z**V<aLU{-KKb3Vn=!x?Tan%6gd@+bdeW|DDEEC#?oqjgIh11g@Tx
z<a;6k4xqM-Y+5($T<SNB-S%2L;`+tw7ygv!9;2HRBEVQ<y2?bP<B}>h0!m`B6TQn+
z-`3*F$WL+tw7-qJVCW8ud|y^$v~vCCMQ5R6#rPVzdk5W&&jPEyFIu6FA~Z&5V<lFx
z*p9}@3aD_k^je|_f;v;#Vzw?m8re!Nii#=uosrK5LY&|q=u@Bghlfb#N5~mlg4H@<
zuJ<e3QOCVed#MY`Hbmy39YP9K9|)(to4V(0yRI}fk?<(|v({x+-Eku3G1Js?l!=b^
z7=ip`Xk<i?{7i-Q7x7?pf6!(GaF_jmi%lDTeVlR{aO#JW-_B=JufVD`g$b>oWpEvL
zX-4w$nBsY_Es}AVpPEvW#a;8*2jA}G3(zLB@GDjqm)04)P^`=(#i5TvjTO;JjIUiP
zS-Swk*W?Vgka2L20=i$98+?bPwH+XENL|l9g~xVXpkMyw(`df(Yo{FjnuWJSa#;(r
znn^=EzfJK}*INQvT<3(M$r@Fl2YVwuK^IiTp+DS*k@nvyd&Bg^1JaD>v8UKj*&SD$
zp)SQ*;$^<y(BDo(x-(X$KO@zEWOhWqfx2}-R<Os1f@5jM*+hiOpxQ8B7}wqy#(<w1
zg9>8BabHmf%XA_e<&!4Ulf3x}XT=dhYomtcw(C;^{wi0rhITZ>(lVnNGwPEMo9M7C
z8gC~!xd|#+Do}8GFv>5G(UUWDw9b#^QnfDn=`)_MZIYn+B6HbSN{}6X*eqrmJPuSq
zjkI03J@i@GwY=~Cb<?QO6#EE2|5kG)>;`N5w#TU722d7T$r%{Fjj0kF#xU9O?p-g}
z4c|`i-)3^3d;?Ps>Pc~{`Na~TB)0QgjW<m(mdqqmY9?MzNLF^3tNiXadxC+P85bg@
zhV$|APqc(5LbuQJc)7v_Xe%fnPm*i>p#86Auq_cgrP>g@C=e834l?OJVmG!C;X0ud
zpW!gZ7#$;Y4S)>AHFMO6KKh${JWg|e%8+6=3VeQbF5VHyS<`hx<-QXXz>0JzC^H+Y
zZ<085i33=@zYqKca+nwJ2HMBA#*;mM*Fr=^<B09akS^F^7k^MqxrA|Ntl<azh!N8_
zU*ww&iB(nlt`K!Fggz)Kq_PsB;k$xUW?IgcCt^g#ZOWbhu+4VN4+M+s@OI2cGM+jg
zq>JAqkg4RgoR&@-MNq}OLum(l9)PCQ53X)oifHL{w1YeC4mzFd3P;4(N3@2?69p=a
zkx8Y3-Jt_4{gWoF_5wbaBi8n4GytIDp%}aC@@jWI=jIocO<+;G)1-da2{K^}(}aY1
zp+t+N3f^DOQVYiO<4tz1WZI+l<*TO=8yJ(^lZ1?6L}B&&(|anZrIr#Rs5Gyf16RDr
zu~IqLczuS6rSCy2@*bry@T18rZ<t}i7ka{iq*{XlM`xbN#G`|Guaj4|_VFE!NBm@M
z`;eTlR>A38kD6xTDLZ3#hJU%|=W8w9y1e2xb=1(7mDV<+>o>HiEnUYh*Z0p4)6bb(
zuX0nk*h_Q1WM}Ujl<1Lm<fG_Rh7myb)R0Ed(7F`fRIrkGVjvt{u&{iqJ-0^)eg5{L
zSIJQ$k93`<)ig>+MokJN;`R#U5cGzC4CcR5-O9rshQx~Q8AW%$n}ZGkr5(SQgIH?X
zT;n_MZ<@z_g3jtQQnOuU9A%8YdNt!VAQQ76N^=s{+FqNm*e?n6%90uaKSJFkm4Id~
zm(Q27k3Plv&wWl3`ba~zVXf`xt|{I5La|QLAEb@UMb;xJG)<hPd$wqX`+?GT!T2M~
zmRE;AwHWj72YGW}Mg#3h#2i+B_hFFgT}_v&XMV@Dr_m07ht+xvQX<_E8~D0VbBZ*S
zn9V~IGv<%5AdMHEQqmNneuowlzrS*Fl;m(Cx&}2Uz+j`&oQ=L<eu=!jq<Ag03(rn8
zr*_`97<WHoAwl|8%HOe1<l-cdI6%60NW0(_-*o+wdb-Tw1D<H_uWd#7P3!b_{8({g
z`?OR{aW5*jv?%J}^!)1A(U^QYoiw;oY~~RclRC~Ji3Z1NMjg%$(#5tO*@?HsMlmR;
zyKS7KuZ(Y^BE$8Iiq{Wm1(!KtFB9`lS4;+W?k^I?`j>R-JI-oX0;dsq)VCm$F5fGe
z__pc5>xT4fF82L!Ug~2L-|Ahr|KB+*6ShL|bRJ6g(A!H5H883&wSyY`s>MrkjbH69
zxrz%gymJS+-{8=!aGd;=viwMoXU|QLM&#t%qB8k8x_)0H1J83V;PliMda9WePOg#U
zSTSOBwOkaBg1!?h{j4<3WpAB(ZwF!4`P+Lw@kVH@R_mduwM)F%?bsnnK3DLMUVgfA
z0{S=8!Uqx&@X><mskDE?Wy(lxLZEV?^|o8~&Z<&58a5eD>$Vq-6nDoG-dJ&iKXg7f
zF7iqPbI-^xw3YU8mBMBDuYeRn_Fc-}PXn#jD^&3qJoZrt>-?y=1TXDlAAhA}u<+W!
zwmaFz^^blR-7z-x$D(-BJc`0RFT{>-&E*9SmC;{oMECB$m5t;ciGGz(UfzhyMpHlt
zF@0O@SWzv%dQL?41*r}_n!xz{&#a-G!9~9{CBB?!0_~a4H76yuJIUcvl_><J!4e2Z
z0#|D9DscD197ES^6xNa7&Mn=9q1S)r=}qLHmjLH}cj;bllrB<1`>P9IKMv5%mlcx+
z+Y*VXONkFLVs00Hp!5NyR7%-ZKMbZd>^y{T?d;DMoe$7GX-^?Spx}%juX>)2dL}q-
z{C9V;usGQ`D_vJw>n`W4-n`yl$DIyFPQ?E2U;6Ju42kg>_k`dncv||hqP*M|VTN;e
zw*_K84jOX{REj4zLJ+&2IH-jh#vy6RD^il)ZW$x3j`ABJZ3pS`so`AWJDt)J2GtLn
zd)rOkK^hkDtQ1@5GV54B%#wL-ihFC(FR8-Dt!Tv`iry(tNUspsq{}Ft1Ky5rs5RXC
zb{a&RQtI7^125tGb@bxz2wFsi0FtJEU(}JR^uhZhPb0O)kHM|SH=n$L{dYnmEH~9R
zQPe!ak}Gat2Znfoa-Z{t?)A;>Y|qNv@DEGQ$FKR(Q#m<r6^%I(G3n9hn?sN}lJMg>
zD_4(ymtr}S8e)^+5=Um^+0)XOD~pT>J&#4E6Y*wGtuC0_|DD#zVLIvc#fJ@La9BmW
zO1NmION?TqP%M-#^ZI;L=aN72(Y@7fnQzpS9L&w{TzxG|xtC8mzVYcCP>)v;K{-=p
zdiR0m2=$X)C5Fn2=9FQc<3uO7c_-l|AYj{!W$ja*ZL~lMBbTNWmKaGZCoXETOwYU&
zb_sM{QU6L~jjnoJ)0}Uq*-49v?-ag~uc9$Yf&Hcv0gBopV)b@Yx9*gn(ER6zz3w>0
zV;tWFH0I)&nd{^BAxq}p^05Loob130s?w?|4TN{)bKb;^2h`}lV>wEH{bazBbGEWj
ziy<`RztU_4#YmiwFkQ2~eDfhS$9v6(Y$%PxU8t^&MkQ+HS2xl7JjmNH?D#zPS^d?@
zhF`oyyyU%#N3_fOY_i(iFi&OMn^4BNvhqc?H(Lo~ndJfX`#a;?4}9I0)wjQ$wyWsM
zQ(#Xwk=tXZ0jh>w-o@M^-i_+wE$Y0r<Vo?(-eU>tmqCMZd7L4k4!7)<KQ`mVZ5K<V
z>^950@^7i4#qPh|uabXaN8Kp*GhRt$p|waQF>;Iho?dRDNy7?t3rsPnPW{3mwP%XJ
zJB}9_-14Q@yL+6xbN?JlI<K2$+WJ!$ILuWMAzgsNH6&FE^^IMsK_WV(NNffqcl^p5
z>ZTAnIg(&^nM#%nwfAS?hns6>Lrq(ZxM=k;x!Wc-Q)F!o1U?U6yraSor^GbYa{wZ(
zEdaY5`yo}AJwuPEG;G-GOqHY;y(Yh|ejW#Zvmp02eQh$#GOjax6nXgL7}iw$ybzy$
z`;K}r<|B{mjHr;#aG~MKNu)DL&b9UVOa0KerRxJ<$+_u@*96}ZUobRPx=?>qWGKZL
zQ4cpb-%lI$u>U#;QLC<ky^#@q*XVtS3z2Yn)An$9nI-V9Qawx5$Rc18dR~_h07Q7}
zzt=uNU~9=0)$OW7aa8tguQ^;VwoU`oY~kj`9;lLZrX5~YK=)CW^Nt>PhY;^{x~Zw@
z_7>f9t9q8;pw#F4vxK0^gpAMjnh}uJ=_XL$;9y1`Uz(2~+W%Jug$SkvW#RybQY&Zv
z-xO5IS3rXuQ!f)wEr{~#qsy;Ue~q0T7BOm`!$-o}q#mk2Q)h#zg`*P1&ZCMIU8@eb
z0t$oZV_k1opPGh&g6In>ub<`0OaklG#XY}si<KEn1s#2AJ4X+4W2o_G;p$qL;AB@m
zLpJm_OIUfbgiMj5XjdIpx$*S1YVGr0B<kl@WH+OcFP-Tex#nZzL;^X0V8D<@Hmjh?
zv;oy<ZA;$Hs@^ul<^dGQGSvEvE}httU%h4Vb1GEnAzv)v9igi-RBLFM9utB`pcf1r
zVCB+9+P2Q&!Fe-fluqKp`0>kwZCF7vInVLJ$H(7ZlKZEjjHqJyIQ+Yqpm6cH?%zN=
z_eoJ|kPO4Io`)@qtR57^<s~yqoT(jVli01k*Qyp9h^~svJl-qrt+3p_MN7qp&WxgK
zud;0C+OjG3RQap&NgcZYgQ8@bbuuIB!}sl8H96IO%MLt~V8Zdubix--u1W?9&-hdf
zvpKJ>GSF&KAY(7>k5im#HjIsKq7+_fkw<3NwU{xg10d2dIXKFvO;0M8(P`I5`NVrw
zsR@CXMduO1oqhD`0Rdero-p%kSp^M^IEnJ<fob54@Xx@x;cxjq)R&zx58q*yL0zJ|
znvLxp9XqGbPp8M&{4I`itaqW(w<veNrUMPd3XP|gKE2{#&GS$8Pn1Wq4df+(0JUdg
zm9N5@QU9k$x;B2i-;QR><B!CM2S8TB$t~E4*>*z_>Ent6ouTM#s+!0?kZ7kBZq56E
z_!6v($jwiOr&51nk(g<eFkll&@<d2h@9*XC)5)FgM`;@g>fa`N%d9$d<+=9q0KvPC
zYZ}7#Ujh)F(S6Mp#9Q+;An!hCaAXTL8~XQ}xqTo>DxF9aP}Wcg_e%Jy9mk00<`WyB
zh1YhdW*@boYWo>O`bU)_2B)_28^G3OfZQANHIFfi=t|X+_|vsv7Sm64w+|m_H#T~U
zLXIBtXyKw|>(7sfphH6XKQpv`LG!cw6}{;mer2B$f|&{^Tm`9hMM>qeupd85*5ZTG
z)5LKz<#`29x?0T2$t(+dxVq9=mRg@38TQP7<WTCl<G#oIq;6JLIJTlW&@dFYguf(o
zpC9w1GCQ0+4SjClZOojg*<hVrD`ny|XU1X1j(biz%C;(+5Kg1JdQ!z3!z<<E*d|7`
znwc|)lt6mQ&0&>&N9{`v{61D(yilRKK=8<il8)+7F*-xIQ%QmvR_6H2{CaIh_thQh
zBX;h&+3kT%Os9o6>pf?NF2nv*M@VN~*CkG?#U84|!TiRgU+?}6*^>d`0Cy&(_*8@_
zV{AmHXuFICeLF>KLDhWK)lYK&4vH`(!p+HZK$<V5$R)8ld6j=R@ydN?GE-mxxI6v5
z^v{=)I-BA2Za**5f6bw$-!POuT0uziUw!PKpPOT9Sr3HEEaDm%VE!6Kb9XNP#h*QA
z+c?da&L%Uu`PlBbFgcKO9{opW9?Gp#5=dB$HKHt^)9s3#`-+WV26s<yTywdX%AuaW
zFS$Xr#tn~EGE%EGzf0x!?U(WZK~IYWZr_0qqR(_v5~Wc;=!IAR*w1Phz38?bQ>=&E
z<O`R+Ej<K$_PaTa(TAxhjBwmdirfWX5g`d6NvTjoewwDU_#PmINFIt28N!3er$LiJ
z!(Y)51jvjk+8j$<`yGJzFV7z3swvzR&HEmQUujQ3MM3(?Ee)gyovn}62c3Oa5e=Mb
z5Pk?nSX>AaB(pJ-$@eoUNtTG#%6!TMOL|2-D|TO_4;}#*5kpdATJ{SO!SxhmUgAIv
zotW>(V)_G}DK+f2=^*H2&w|+@%+GZ9t8pH8V?&?#jL4c?|5ae5GF%B(vFufXPgM=#
z8Jl|J>*VY9Ml9&wH{gwldjpHL#f9N`iaoDP_s<jxV+?{liGkt$ho2s#tHd`*GRqgG
z3LaClH+yq)$_Wj?ps?3-H<5=|hxdV4n%eu!=@wM5Vw{>Y-nUZJ?5hJN_PH!{Pk$qK
z5Ekmqe&!AD<EW+lji*E&KOP((Wdz;z)$A<z(a!E5>2pU**~g<wmM}Kn{sI*xoyKYJ
zG=g>>uPmRNs6{~jL&=ASA9`H#6a|znNT~n8JM#MLg#SGtp7|7#BHtrK0lV??&djB?
zU$UjqG3s_;fheC-%GrbPHe*^VKcYDQt)>_4z|4H>smSS!tnsr}Yk3>*Fyix!G=PfA
z4={8R$&Bp#SCW^$TG5zx6D+bVLhjnZ4bGPl-x%5-C87j~Ele@Ewm<tPblywGxer#f
zjFq=czXW(rmn<Eby%=b{>9H*IX?piHHQfvw^ve}AkEHOW2#4avHT27iYKR&@vmZ&#
z=RT42vx7tJ-vw)!C0|CaH@7qCWWFU95A^2xG@!pXCcNS$bGcP$d1T@du6M3BZj;*b
zbnAN!#ig*LXpqLEl&OETh?S9QYiVTxsuc#g5oN}AriZqM!N@CQHtbz)>z^mRVsxgM
z`|8^UuaIP_KJ^=C&rK{x6{0Kt?q+v@H1J}kxzG0;^HP@9s?qLBEVkSTNS0LRpEXEG
zaIxK2G(ZtQY&U;x)t(TW=?Yu1PjJyJRK00tf#DrqWr<7(-tVts(I5fAZl9R)kDYn1
zK`LgkD`Bk|!OmXK+jyvALZ^Qf#gnJ`=(!9)NX=o3x~MD+)EfU*G&u;HD>5QSC>Oll
z?tc5s(DQzI{i*rkvPAj{-D#OaeRIKou!AWyPfwBx4g396jw3_4iT26&Y}fj>*`@O7
zLh{5I&N8WGG?sU0^yDO$L{RRhGvfa_7A`AebHK~ZE)6-N7!?1E9>);2iaa?^QHwKr
zBoGZAJmL*pOCMyJ#8~VWz}JMGF!7k}(+A8b74aeweAj|*i_@Htvqbgo3TWg}oe%oR
z0XCUj>G8)7+S<vkaM*iH0(8J;9)=&YLbY=4q%QGsT3JK8?b&j5(nZhTZbVDyJ^%5>
znSXQbmA-cQ=fgGnx+uj^70^|5CC4?3TOl!!`A+(7!8Ymg@P!A#=Pk+cj2N>UJ^VuU
z)R|6=Lh#&R9HLVzt1r-)5^FY=*e-_fBb+nL4LL~w4@^MMb7XAUH}*iC;(24hD>=jB
z>E28O3bZHNHVuHV{wYhN;U4`|iLAjJwHW2>exKPelRyg+ObT@>h=d)Mm~kmPLxT;g
z2~w0S<2D(EvL&GRrSM@68o&LDnj9Vh7AljQ1<umBf}KhAbbymtWlu6ar(w@ETL5&;
zfuW$9NTgh;EkMh+g!Aqf-szm>6@Il?#p9vS<4>7PZFNTM%LCuIw}*=aAMmW0*r<=h
zwSXuD*nrocvJ`C$R%)&d^oyMWco|BQ+IBKM$PV*9d9~j9S9(a=zkUzuIj<Cdez?FL
z1Nt5mUw2$-NwHoG`GS;57u@`}sLm0>@k38f-sAT5&{4s=#}Lg<A-6T>>Jws)r!{!)
z`W=nh>fgYC=Hk8;SkIHE&#m>NUzwaFIbPGXS3tGxKEen%V`(X*Cq=8^{~l5Aal@#>
zI#s|@FJnYDh(=enYfV@`pN+__+8(iXrN(j?0Z{sXz0cQ)m;Z#sgdw~mrbqU^BLE;7
zW_>P}4FY<0ng7NxVrhHc44~yC{PL-B4n;IaAdL4rTIaqfHv)-Ukn)&G1)sHvt)_bL
z*J90}6L4~P&E=kdT^CSVML%+DsOb?0<v%!G5Z8MHv(?R->OWpa>s*x*3SReh;AB!7
zqcm;;3;P`>ZBMEDp$$P@tq~Rn@TVeqYijAOYovDo<EEOz>QK4hutg2%&Ex5EJ{ZkN
zvW%`Gd4x`Bgi}zMw5{uX4r_f-u0j|-dAQ?{p!XtgH0z-VU=w7%BP-B5!qBDZ9T{mI
zntYpq%8$~YxW)a`j}f(=di%rDyKo{#pgs$_ao$RI(*kaK2S4t2%#I&4VsQ0n5qt=t
z5t}7iOy+=o92$MYP|W<VZEQvQE@5=%i7$I%N~^(fOG6w|uy>mKUb@v61$0WlprWDJ
z<xjEA=BKZIu~mx8d0klq1UUwkK!5$VrguT9Cr9!}t(=1=Xts|nv_17dOY!Vp<|z$Q
zQq1wm!EPi9)lBng!iI*BTTu$dUk=XV<4_C=3VuYu$0LzamkcN$;rKL3&D9oJZ(L%+
zPf?tnuHBvuO}*UCNkQQFbRR?1$6e%j9&_J`RrQe%h`qo;2{Bv2N}*!xHkajH%s)?{
z^8>9QQ%$kaAQ5>gLFt0AS&y^<9yJnXpXLkLAeIa-`~RtZVFDiqfIIlpAp#w7A(nf0
zp<efEy=RI(tWBO<*ra>JWrw|9W{OpgAeKbj4?@^p7SB2%dfj0x;{I_?L!_KZKW<W}
zS3`Iag3qbf-XwjtiC_sN9pNd<XqHuW8pzO5zx{Jp=07TWp7W7&9I6QS^^{k}>#EGL
zV;NNG2ai^rx~6xCA8zjzQ6VPb=T!mWRdQ+iia@ur3D~X%&CNu!a0Oqk6aNVX`<xf8
zl$GT!4yx3CK>ILlVt3?Ql}3`pGSz+&`nP^XS2b;w?WR?}#eGF*uL2m*yWRr(ps1Om
z?lVC(?Uz53#ql=(z7!<`eu#g4R8k2VlKojX1Z}yn0H=_vD!B5?Smb45f0lG<87rpT
zODLD*7X%c$Sa*}jR=Yx~ua_|--#_`#GVVknqZwk7L`^&vGmzUr#I@o0Ao)z2*BLnu
zlrx+R*$oqlSbVl3Pszv_Vne}<*dnOlSMQ@|G<#R~qs5+#9C}joown^`+m{u|%*!t}
z6Ca%ea<Gcx0GAN6UgNh<VlnLm_pxd`jBK`|^uu?T?Y*x&=60ePNkyQ}U5DySmUL~L
zCERtY{V2aJIDR5IFL(m>$2)?`tCA7k@6u=BECHV!XD3lk7aAq*Z=^;Ge1T`!ch^Qe
z%pLf&SJRnWkk_))>JR3rIDOyR9h8)m5RJpvOHIf9FuDE^7tVX(ZDVGZEjnz8gG?Jy
zi1*)3WZJTtls&BC#eYMLr1w(0=TX6m2x5OxpO^t898-cSVsRyEm8;+8?|97N#!Y9Z
zg*Z7P4IkzcC^=A3fpbg_;$B$=&*;jLfEAILyrh2%-veIu;|2%WAV{h&s58jvA+?hd
zo=3aYj@d#rs9qzhX(dc9Qf>xU7|~>w@29>>9Iv!BT&@ZnJI_0%6%1w)GG4%`Y}X=k
z-A^hv-R~^BbuAuequmGTw9+r6ZV<*--#zM`RpQiL)?DsZUdUu|B9m<3m|eR+mZ`1;
zzc_%`L^6LVhkdt9=AY{6+q*{r#%FC`l)iDE5isU<JvJSu8LorEERGU{KN<3QOLkUY
zHt_9R9Xj5eh~+r+62-hExq5~$f|qtnKV;TKsfGg!obw>oty6CoZfAxn6OC#7u$xe4
zHohR&lVT#it!#g~G$bV{pl3zjd#==^5PL&{P#;0hvewC-dXQ=4GX6{|TR-qdYIof!
zoE&Qbhu;(yiv;_wu@n)~re;aA#5Eg&Evau)MH=llcjjHy5%47Pm}a1C9vByBWSY``
z%^gWAlig(Vb*=0gc@m_9EQZO~bY-wBS^yi;F?w?f%+6?Iq|O8*znq)acZUWtW!R=p
zh!0<UYe0>!h&cMR59FXgSy2o+`-(rprIp7_{PtSSPbpHiK;-k;(etU2`_cQ{t&1sG
z{@A6>j2kx=GY*xPb0^3HWyC@`2Agrvw^`mWIPAb_s89|hwtf9e_y2~)r?p#gnw@_M
zva4K=TRxpX9F{8!_w9Zwr!AxUc`*S_;aWviI!5pGc)#m7NRLeDtNJ7&2ei*SqyecM
zu!iLT>+g)#BvDa&gXgQL0%D{Iq!ysgnpqB`JNGYExjm$0kYrf7Qnl<hqMJ3ouo1d`
z$H|z6K-9v9z|;(NJN<NO?H{Ap>5o^M%zC^`x?9$ypKQS1pAgO_oA1iH?-?n@zngzI
z_sQb+Ju*O1WPT^Ss&eWZr!MN$rk*2VswwRm(z$aru-zCfxkl+d#^IHRbZl%>{aU8f
z$uX__=*bYS49ABo`h)mC>*@TE`63j%vs00Eh@`pAHFup>UVffuL(e8S3~QnL!1<Un
zLI;~;dDMDVfZLa~YP7@{^l~F}sWkn$yu%C*F^pm=3E{KB$8lPz9F8N*@^d{14+BU)
zPLwHmPAnl@Ak(u#dku$gw`NfO>_&SKM?}XGU?{S`fB6>46>|=alXj&lpfVs04Humw
zIZAYB(ZmcJg73oeEw#IMR@m{YC5j_bD`>UO3-C>mRMQ5@!4dt!?uP@{Z=$%DZ1ECt
zXJWs#IOAxcS8tyihJJ40&~vv@R2gH@b&~hVoP-~JIjR+DgwZU(4^W#e|14JoIQyUu
zYp8v0n7XSr49x+Upxj6_i~&Xkl&;79s>S51&5vompySgvf&%-=glYDZPo;nH3P1kY
zyEfHs_Vx3t;yfhp{)^@_K($71WbT=USe()&@de$Dr^hBd6sJYuM<<KUirDA8+YB!>
zzgO2f{KZ=?^EK%a-Z05qVN}zsdN3XS@8aW2SZ47oOZ&gq<bEA{`;u_0!Oe1FR`tZe
zSaP$$Hs^y~GB(P#ApmZe&~6Bsq`gtBXefxFjaxgBi~$^m&Dc*}Mv+C5y&s`r*z}W{
zA4r8xLQ0DiIPf-VxqHZ{<Imnb`Xmi868*W{`iL7^<%UQ<B@&)be^Qt@fz{j`%R8!P
z|4+gj6nZ?FARNF(PDT5JJLPp#jmni}w?;YNY0{w+FClDgyTXrXAybU6?O>|?1b;h5
zPib^~I-8j-SBL4*uiSRgGQf(!Se^upJ~f5fDopCP>#LiBZZet$r^kbQJf8&8HHWp$
zVB8znFYcu7k{h!Y8th1ocB#vzqn?MWpZMDe@W-FtzK`iNFWj8FH!kac)9F3N+M$Z{
z9oVG_?&U}}YRhrwVv}KwW;$dyR74zJVfwMas8t?<Wfiw#mx=g)b*~#Cqh0Pq9}#y~
zZiS&tQG2R<iWqQSPg!eITIBb^C$GHrA;;ccIt(Zk@LA?2X^;uCN7nSrH8!r@(Vrjy
z#cY*|IMa&{D`t>DRhJv-Fx2RT^;V|fxc+$Vc{+U=XDqLQ33aursi4Z=l{Ve`K{!8O
zPH-n06rl|v^To#7RqwZo_u?tCV0iD|_-peQC^HTr886m&gx~4o#e6X1AOo+JBJP=5
z{)Pz7hy)5)#b`l3`)H@Z^7i=;B|_bL-=Ys`rF*^|cHayqA9F0pWevJA3SvRH5*0n1
zXY`~ZUV^dboA^PC=c&3fO!%B=lCNg>?DNf8YmPAW|D9KSvXd{(PErO6{-=n)cKO|y
zhC>(2WE`}uTRD5bI>z>)k1wMFDL{qaNTZz=zXj==rA^hR@etBuZlF{21?;xKhUe!t
z82}71`hof;cr{rizV=jBRYjxX*&n_~p~!7+WChks+o#|AGMB?}{hHR$sqPxN_&EA0
zIrIq(vgU?}?#S6Ox04y|>124za|>OnGfM4S6dE*%e_hmi6Pi%`(MHGDe6wNP_%7?x
z%^~$L1E%_MSKFSVt(zM9dao*ZZTcy+Yq*IzxwK2Gd_8LfDI>8hu6^1646S!>ZMbP%
zleoq4XrmKc_Yn{Hd@7s~NFS6_qwa{uCW+5bvyk72V*G`6oExb%hZtRk<smU~QX&yM
z3T<V7rAZDGP}xY~(a#a=+$A=})&)G}p(}XVx*V90=k%G6DA9Au7$URcs1L8oVwHF2
zXg(33e+iwxC-@U8J}A(vh*)1`+VCtFm(x$1aF449+a(!L2yLiqHsEB<;kMA)sEs(B
ze9=&ba}*K`g<yN~(j=CiMosAi2;ry!L?co?oTQC5<upP$0$J5bKaB?P`DCHnOw>rs
z)`am~_#=JVqls0^SV`UHXq@q)fyIf@Wno`>bC+{psBb=DYUCsI6a`%p#2?7V3l*`q
z_44miD~;>=-A4we$KvBPE80h$kT?;-0kv5FzoB%g!d(X%b`{A_@Sf)mget`?-Sghx
zSAdH~%J;CW#FV$+o^hAlRh=3#(XvJAbFw?ZaXYh#`X6)Se_N-(o}IIj^<o%d+ob+}
z@{7Sehx{(>2maR0xA7N$*|kK%lo{8dziAXGAskh~wPG9QhMtN*8$a5M&6>_15kTSP
zu+lHY(KpCQV5I1js5r^$g0s0nYGKOyLShL0N#w&jpLk&_m1QZg+38<xGzya!f~>(S
z)z>xz%)HlDFtJ+Al}>Dl{USyl0Ox?l3;m<#ryJbutMz~yeV$oA^Y44p0Cl_(?{6CS
z&3NQ3vT%9QI4TB%J(%)Gch1H)WSjY6=E&(7;VWR7`_uW09RGU#o*kaz(D;LhpgY|I
z@tR-|R4SKh`GhE|)0hz$xou$ft%fw1UEx{qN&GZLS-6$3AP%%a%^n^)<(`4W=P1nE
z*a)g3<ivs?7a)I#wIaC~J)ex5hX>|HHku&6CN%8W5qh-P6Wypo{qRD;_|?}hfZ7Gg
z>k5Z+e@T1UR(1fPFgY)8%xjV>N|!+mVdvBxACItW<c6O=1LNrJajBS1K}l04l5Q+Z
z;DktrEqj%DPi1PKjS&<F&E8}$>HS!Cz|@FjW`zBNdc0`vvH`%9-mFcvL*xRq1?8Qy
z!9KCv3-7yj<hr75THi;Z;iKAW#GZ{&b)W(dA@Pj2lHl&2anHVa?SJ^rXOPQ+;nZi^
zZ@aSLAreprVSB#RdS>=s4``81UH3Xen(kP;;=MaG>M_07@L5F!d7qzn_bhWRN}0XO
z$<rXW5%VVzW+7<To$ZF7_N?E#9?q8u`kdQh&lYD6B2>hDmhdGQC+9oaHIe?1JFDZi
z|1MH|jL;UN>jNG3LxTTabC-W&YYP#vdt>F5jLGl$fUW#6XmN+da;aD-LHlTa%ZL3J
z5*)qgyu<}GmUhCwSJ%L(1Pu7|K*=C?;^K8Zguk=zaWc`O)xdKj3%KP5C5?R-Viw=x
zB?N!adF7791PY{Q;uXz#V;J|^J~r4X;;)hi+F=B`lCIb{A^D6d*0YY0sdU%2Kfhxi
zZoqZua(_Ssn~hY`%hXzil4K%jiB1l+IUc<N<++AkXYZol-zbw#6)@9BLmN$2`mBGu
z&Rv%mmNn@%`Ialh?!Av7MWzFe^F~uJ#fbc=N_j%}N5oga*X@Jw(o?hoPUuwE;UH+R
zPpDaib*y)UfW@UjakkzN{s?9#<@X!Aa2dh*buL1g^3*ME)3dQZV5hk!GzFxE@=O2E
zZb_gfov^ULH?3)Nt|yUB-YrMenFq42Shu>Vy?ax0aJRXO4|MQ;R<z&UNZqaNtnWYT
z>C$NYZeIJYF$<@CGiy{g84?FMJHCOGj7M$`A(z+l0!+k&%RW#a_zSnbL~)8~2jL6~
zL%vpyHbRp1ZL4Xm$CIez0<Cvr`x)7hN3!)^E9sxvRRfUQ+^WN|<M-amwZd^MkVTt<
zn7AZ#-d3XIBKpke0Q12V0SK+rp9btLH}#-TV_HoJcc0<5T{K&!2VN|5xF;c}y0#yD
z|LMjF3&odmu)6rfS#t7gi@`fj1ZFP{H}T^|Dg^+8DyC;{PdmtC06t+SL?3(yaD_7m
z_&M%(wVt-MwsTMYOAs(K_oJSn4!WGdo%loBe{Vg+boXCP%K$`kOr6;Y^Q%+1%1Xtu
z!@+f5Uxd_(4iBcV0qsS$K{XRnp)?8y;na0!gC<be5fZwpYv5fTGG5pv(twENxK~#a
zZKyIN<SEoM9fS5$ujgua)BRHN!-{E612`p;tkruYBaID0XPpWm&r#p@8AlPq?0Pg{
z!GJ9?c&~ZXJO}P+o+gWzQ)yEN=+`-~aL?{z5@~!nf(w1^4!R?U6K4RSO>XJsEH;mi
zDaz7l7!W+y$F~O^)6aRGZqS~fs%D=e@8se^pUc2P^PAWSW}u0%@A^F-cy+*Tk<PN>
zln#rg=j!yCSN!G=*CX?Jymx2OWo|jHw?)G=^sO;NZG^)8jzYWuJqG_DhmMg#wMdz2
zgrS09KPTqE1Se~OJ9?vA0`*V$T-%^e>~6&v7PGbr$&4QBG=<e7&C2E}AB|z`02YQq
zCO1QVyX%C8Sfx7cpvd0WslC#i6kOr2ZA_jy35}`>7~a0T^c%EJN&X>_1NS2WFpf)J
zkgVgHa4d=g3r(F8UhHz+P!Rqn>incrybMab_LBi21RrRgtdeq<;h0&ec&B~bk!zZy
zzpyC*SFlhr!7NY$<p`6X5==eK0AZh6a@W+DZ-hnyx5|r*BR*XJ5k3EIr}=bQ|4Yz%
z#VT^OCg{=WZjD@un=0>H@333Gkjs)Se>^qVr2Y1HEKlQ&M?O49l52OG4l5fxmL$cX
zMxv}7r2S*v(r2j^T-13M=a!lf@W%#akPKj+|8!ZhQ$+XZVDc9Rh1zX4!Yur%5WZlN
z#p{N=G1~gNJE_{{a$aqQqK|(grBe}G^`3@ihNsGkxjh3pk;7YG-v4%BvC3ad<^PxQ
zsPvD_dUA3SEb+!>IKJs`=885IJqh45u2A$UIndabL8-L|53v@9tK(;$Pni)}yX+du
zGe^~k8XCCC@Q8cr_5A1v286cHdiPTGUtyFvE$&ya{|1UL9ndIjBNrW;^lvI$Dau)t
z<X6I<k&Th~zr8%#AARA)xMBz?V7#J98#h!>LraNhU+jLYqUvPAYx&xobW$h9&+erz
zvbD0*blnx6(0%Ov!=+;mh1Ka~BhV9mt-787D6xQ^`sI&=F?Bzh7|6H%eJbfnap`vs
z>N1bheN$svh;3mg=1|-Hn0uPqeSy~-3^E|q6bG;|L{>k6qu=M6bmwc^H(a8yf$cZn
zGx3?x%Akz7bJO%nQ05X=#|+>?u)kHkWS#ZTtzl_R<``O|&u|qWKYFxnn@mQ&9;$Bm
z*jjnb0Szf?v-dMb(3VHgpoZ9C*>lu|fSp1{iq9-<cYu+wVV?FMz~X~8tDoQ|gfsd_
zZm?duNhIq6gLQLdu{G16OV9FNf?A9~0futVlqKX-M5QvOrreVF6*Jhxb?O`BaC;DC
z_qXo2lapo5<>|r`8o#?s!Sb4tC&rbSO6DXr{g1ArM0D_H*(`pKY;*}Vz^B&JcQlfW
zDUmlL!__&Q6;qN*A3nU0zkqvrjlrI-jFVE<$u^e?9R|L<VG+rmQ6x8K#+UdD3Ap_m
zt>xbz4_|2Mw(Id*US2a48^exVbvv}!x|?h&-T#7)j#)A|YkXPxjKDZQST{=F&n9fQ
zA5}MKuQ2K?+VoykmZN@lS*-1=MvAJ`Z&~xLxFTkpcypZShwvo*w|3QT7)uYu{~vDG
zy+*j&#z9Fw?<2Dy!hMZWx|?1BM@IIR@(aHbFzc^$#p38tcXFpryJ*bE23OI(MFrxg
z^auxiot2BkJhy6ZW~2ZNx9rB5j&Jc^F@HMOBSjB~a#Yg@=`k>pGg`2RsoHuSwklsS
zk6|Zm`M1}V1Q=dj6u5TR%U%Bb>1VoeJ+|JRRQMv)bI;s6nt&FQokCJ&r9J;XaVTj^
zRms69HAaNqiaJdblj+A5F;Xx$Dya7UX~+19wc>O8yDU0knoxFZjM)T{U%>p!$DWQK
zRqOOR8fp6GRW75DLg&)c&Z83qcN1n_<Dz;9v8cFajt0CGQ(}~h8gEa|!4HyXRHEMM
zaI1=U9KA~=qjD@iSdg}Uj76CY9ZhxauD*h?+IogPnQXM8gi%TqK97eRZ#^*Tr-+IL
z49>s@9EXAL<M!_W?+m{&;NuLJ`z>0fAibCKx==VvkxYm}5kvXfs!7)Se(i&_;x_8H
zbC6-HdpbK7GAP@lxrdw{rkv{*Ed`c5enC~03+bzT>X$8j=h9$ND2S(6p`dn#ZC(R;
z021c4!)5Vd5HHCkd>Wp>N9NTSF(c}5j!bL`n~|?ZngHtsVLCBpn3_Rabyu9=@z3=P
z{D)CG)ANUK*Fw!>-rYCy6{e5s$!9$j;cCb0Psi(m1NuRyH-TD#=ZLfv{ZfsX9;c_<
zgy&DnZ%&!##Vn8(UE6lZu=9r*jK2LtB@hg-JJ+Fa@RNA)laa8vp6HiUR6GBeLE%N;
zJ%6gOWi{XnGgz+R_sFX_sI*5{-blm#fA~5jJ0<OplDmuDNv;vPl6QIuTNnVDk#bV=
z5yIiuE(??HX$s6uY(FZMMPf#B2Vc<b!M=sx^F8WTmm7<C<vd*k{RGy9;BPl9du1qq
zX<H&Rq*e&Mad}Iw--f!&Et=m0SqRT}Kic58HfXm155DI==&v5{zKhw-Wvd=>AypIp
z>wLl|ZZk_yPPx?Pk+E_Iz^kA8flKA6a={ZVMK4224wgaT*R=cl`N3BIs|HfunwGyD
z{70mKLC`HevuRsD%G3b7Y5npEV&h0y?I%+M?Nmn)%C0NQNnPHJo>C4Yj<j62jBesK
zXK0zPWkyEe&w(tu*4WYzao@tzXO>}$Y0^Yo3N#?n$4&LIaSj~Yr3a@_exbE(obeFZ
zFezl$MxI_Z?4Ai?82W^+HB>GGJv5}TvKL8{H@?_X4<?)mUA7|_bLkv>#-lLUZXci-
z_Gg~{^;@lp&pjSZ@Sx81SVQc_UO7xJu`RJ_$QZrV^s3#Ua^Q3A8UrJ$e?mOkloZ6x
zt;LcKo0vZ~5qQo)goO=2$tzxU28iEBCmJniS<nuDYZV4ohlRd!yZGVH`hyZQy0mbl
zV9v8<{w1skptb_WC{<?JDinxPq!1^NM%rIp8)kmz+~{J|wHtSi?&cTAohP|CsGLOI
z&xpDC;6iY7m;Zd1Uz{I*;929B>wgw&*@%qh!?;d~hLK#(Jn_o#C(Vn`C^<-XzX_m>
z;E9{7UorBXM5N)vv^5S~0WT`mV}anXk-*(o`2vXE+D<b=u+zd<&e1lu2O3WUZ#v|u
z!&kJ5Z$1|O56MV@0r=*>ea?P3)A{t@1h_5xloMj0V9<<^U3zZxe#C)IidKJ1LUR?f
z!r*`3?+gyRBn&ZT7P@r~2_||o7cbjv3}JX|30ppQiU!>Sv@!U$ZvaW9W7LLIS3dT*
zHNq}uu5Ft(58ox#R<y3zlPI}72P|IHBGEw<AZY?AJTNm8#0XW}jaE{VPSt%E$IyGR
zEZwCSwqn@|_w8Ud{d6W{27j7<u9F@`68>|4bnHiLVA94`V2g};W#OIMdUbRR-@gRw
zHXRZjk&iPh)a#X}TV#pig^jkC<dfZm#b=e@=hn<fLA}gA9a)XMJ82a^<$PpzU(N3_
zn)bl~Y5#({JQZK>aWGQ3Ei@l3W_W}%Y{a&aiZ_k=O39@Mn^U?26tcONEc)3SdigJY
zM|I<EGjV4T%yj-$`r!n2*!gxRMJ=l3tFBOo;S!3_&&>YhQstx8n{WJeK10dRoI&k#
z5OYpkPF>xYi+{_nlM1tV4EV9K;I-uUvC7!B`SXX4Ky#O~G-IA1Gqmz*GH<J`I)0$J
zhuosKyF_M@U;d1=XW*o`<sJh8H`IbgspY_s=DFxM2A4)e^809TmETCmhl6KE;an+f
zMk_9BMEFcPTnnb4&;+9p6HL#+W@D)_xpm2+4vg~J&x!t?|1E&O?ovFUdSir%{Hqj7
z#L=<k>k3OL-#-2_%2$gtDk8frJ&(t-CtDeX0JSh}BHt{M-l%t2b0vF#t7MG3fX5?C
zX~3q`Z67`!m}cl7y|wKySv5qsd?k5+7k8`F@IDyd-wNYaUWBs6b=CB~%(%n*ls6!w
z>NUoesMcBs_a^eg+D*67{b;Pz?U+$F<MUtq=R`~aj#dQ=vF_VEei(uz_srYns~&bl
zCi<L*`#-dX`N}W9(Leon+tZ$LV?F<oWD|Ml;qZ^Rn>aq{S45@jor!uJw=2`fkbcx`
z&9t(%QBkL9K~E?O8yh{>Nu&0IR5zO8Zg^RR?raO%e`#JnuiEVW3J9}ELA`i^1#Sk&
zerZ;n_nvp3vcpm$acy{_?b<E;BVk(yf@?dvFE|o49qvw?JCXwOk}EM@oK!lpjfN=3
zP7Wr`o15WN`A0!wmGIQ`%f}pjR$eyt7(h>hdRA*NTYpFr7mZTPz&Np8cy_W<Mgw#M
zq!NE7R9#_PmxZ0a7agT-GrFg}EMOGN`T+-Woj<0B$o-=Xy8M_G=gs8E=z0&LW%m$<
z@?OABK9HjKavS9UW`*Ry<QA1%OYDS-GVdsolMP5EOKkum43jFOIwA^7Hw(J-I5uh_
zPZBx;sVgp$VlQMG3xW1Px)$aDQO~ltWqOm5qGq=TYW%9;?yx|jLm)H6t)&{JpE-7e
z#wOZ~gkLF>(>mM7)nI9idfrE|FEK_zAQUgx4`3o!=DB93`SxCX;j1q;Us|_RG>HB^
ziH1O_(mKCoW5Dc$o@&juN`lj`gPLgy#0k?8w95vy8+}#|B951cl;dJKn*a<UA%pRe
zNn&N3xch^B7C_aXp1f%!3<JpbOO9OM-AMVs#tp}^`GBmihVU(nT5AyUiTqJSs<v8>
z&p{c2evJqpditql0V&20*q5(=HyHENEhan#0g|cT4F^38Gx*%SJG(J>@HQ~*ZIyl)
zkfui70({hxy8C9}hL-4F^$if+JI7s|>#P4UE|$&?vt+3+^a&--z5VfeIWp`EQDIC+
zO>zhhFp~O{&)s!v$(~>f@c-US4Rc_LB@VMrd4{1_Pvv}mcvrW$B19Bo>f$pp@ilEv
z3`Kv_>ZMS>k$4qgL<B%<o#{N*(0>YeC(@SOR{XM2@|6~X(E?zG2yTSQmW&MW=a@c<
zX1}{$5^)ExG-#$VqzR@J-Xiaz3W6=uwtpj&%-C1M#IN*4*f3;cni|^F);u43q7ig6
z>Ali3vhtUEaShZUwtVMxCi_|$%Ob2j<=$`cYK)Df2KgGTj+oyGc^6k*N3SNU(urnj
zZYrc}Pm4HGY!}&TV15C*jQ#3*c*-l_QEIK{r<PO7oc1`UB(o~tOa;A+mmB5CFAeMM
zmOXorES2(;ls<#0{I6{oVcAy|-L6|lU$^P6u9csklD9N!=y-zc`uq(J+C~VKS@bzQ
z--nDFPYj??Id?0xzS6r`MyeS3rSys4#wuJ)bFPl2bCixbCwE(U{^p+GG{;XV>;tk1
zCL4=B@r{XL9iD<Tp{uHRY>Wc)kK!smvg&K658EQqrBksUvv@O7EF`ZjYJ?0qz@K+A
z#T_193p&|RNw02Hq>nN^Wz0N=#y3CzARLy+{yrUJJuv8Un>Df>zEF*HCh3Aoeb`QV
zu5QRrzUmsl%5OOFe47QKs<94L;IgoLbi7a;PLuL|KD9uMQBoAm$RKtc1g?AAbfv5B
z<Ww;aI|6E9cmwmskAKaWd#|#%1;Llm!!c0iARrRAsAtWAs+*Fa`x1GX<pF$Y3GJrW
zWgL1hLd~o!7r?_@^e<gBvFz`~*Y7kvy3>d^lI4i?fs@2H^V;|Pdu`hPEfR>JEvI||
zMivb}tvUY}USZR3#Vw5|Kl&FIf=P@;{W>n2xxX($?FB$KuUVo^icao(wsD1wN%RQS
z!m71(k)buAy_RlR@v$Y>?88NuN5lhY(ri*xw)zz$Zzum%<}6>jx`i+wnc>(AF}3X<
zPRnkJFDGzQ*yK*|PK7Z@2+yM%;t9Il#maAro}Z$3^NwiVvLt<jr{q7_9<U|3Jr3XO
z+y`~2tVF%+(yVFQD(CH4$pNcS*x!W4jsv}kA2Z5N;p^o6raI6gzQm}%63MV(exwY7
zyB_7Sd;<RJ)AoO&lg-n+bw-3nv{WD*R~L<U@%2#qBa?4#nOc`T7n$fIgz}XpnYTlY
z0sxm@nGi$1mG3~W7gGk0ExYX({dY>n##DdnizZwjS0M;5KV~8_s5qBp*y=mz{V|8Z
z!MB%=c?AZ)e;oVv)}Zy5=xJ7Cw9wcn80hpXikmV24vgt<PB`>yUaP(cYV;l<Z9Di#
zBX;6)>wxbLCN}NRWy~?y<34)3TgCZUk9^#!#OS^cRTpTbKx!tl40P5wFCop}pY=M-
z8N=e)r5M8eG$Nr&ZZ`5WNuKeN>w;w(S}xeaX5BdF>5_PgxA(Hz1Yp<CgW1Us8ec@=
zM#?z&P31`0yC7OLjTQ;7V|F)akG}XJAS4kxGT!p+7@cm(W3NcH6V&xYdVW4`m=S;7
zpL<oz7ci=evXR-ilEBBXNx_t1qU*nrxaTfDTbTKSYW;Rdc1i@Yj9^o*oAEX`F8@Ig
zyjsd@RFKZ=b6j%l4}RKg_4r)8FRQp{6XhZ_FlsoT5+F}*Gn+MAJ5R6Q?5g|cE-}t$
z#Ju9Y+;a5bAOG)C_<w2!{NeHbs@YMdNGYM2)3uG%Fqp(S1RRajyNht>CF#Y@c@4m{
z!gS48G)ZCWHvl6@;<+GWq6U0qoO{VMB21{QO`Kn}*PqU56p>!Q*W?sLzM5JD41gpV
zIhdrJcWoa%DO}((Mg-m|Ak>Zpnt+vUOj?qfvos#B-$>2+W4#651g~cV1+CXH;=_WT
zt?%Ziy;as>-L4>93iMPe<}N-Z<mV@;=+ZxBk64{4Z&&xVazFg7{_qnC|CMT*X;rD5
zQH5;#M_F|DdEBtUZm&9p3f`s=WFayZCCkcjVwbi*Ice?(ILG2E33U{FRt#3LPYe^_
zc@Zu{9YaV}dt_I(Zcj^04km@5^7a6e1fjM<rY&u&qg!Is!=O!xS?Tw2?1RsPXA|gh
z5m^as?!Qa7)vCpJsRq%1-b9V7oesK>csAYEI_3Zs#g-VVpEzjV<wnYMGZL=hByJni
z=qXhT@A{iQXu>pJUoqn^auNpg-eKk!yCK=O!ii3VJREgbGiyP+l<|hUe!5@`u^{(n
zk^^?OJW8FkodGIT-#JDS8YAw_8U5*<mrHv>WGZK!y_Dz_3-J0kdYbptAhG9BJ;$%J
zQSfgC0~qe}2XB1xoVzvh4c<Iu85^*W13%XJUlPN{|85Rk1q5<~^a;uC4zvR0@T$8s
zR=(6Mo?4Fm4$ZrL%6CB;0sKE)y=7Zl?G~<`paF_I1b4R<w<5*0xVyU)FQGtj_u@`*
zCpbk4#ih7wpcE-kTy~zdp8Xzczw<+WKr)Y+d5`P7&S9jW(SDWie3-zyZu9t*j#Y|0
zvQ4Chk(n_Ll=GhRs_mlPRv)Wu&i<S!+529NqiI8F2eJ8|xg9wI_1*gF#s_w90H<hy
zQI5gZvd;hWXG$a9ZaeUHgtebkAo%%l=h3~7%M3HytUzy6EG?C8TU4lGY5F@d(P&h3
z92gtt?H?N`J=iTtJBVKE=N(v$Zuow*qEBi~Fl>LOc7xJ=^{f@A)o$lE4U7zjCeEKz
z5{I+hs72sNw>a<Jl;w5=pw**5L7?;sQB@eAJ$oV>e0ay1yWpa&XleZ7Dlj_M^Bf#B
z31tV$LedA`k!`tWw;4bkyzcg|ZU4>_-vwF=!sH|C7NdNzPo<?$@SW?y-lg?=q?;_-
zwmSv0i#K(IavQy2#R)RofkPc9bFQt^W=hFaY@frgW)4|4P6R)Ee*3j?)^8soAgQ|I
zKda3T^N{5Mp2it@3)Qlf4-2Bl9Oe{ezTLj990`Io*MamozN``%D3XhO8FHVz`kNI)
zfOjZH>jLk=@<h&6$MTK!8-1;>yj(SVA2R-j|Lr2R-*Pi1uC`6n0?FMfM1jDw-urzB
zL?ujSO2qWcGykm)hEE)ZXhp?q2W^`yt}s>d-^$~zgWg%#kCeKOaXxMb=)|mk+kg<k
z(+hym$?pC72_j&!-w~F>paY_RE`U5AemnZEcjSpxp*L9=;r16#Skz<{LXO3ghwAey
zC^6py8jQ6QIB(d5Kp$sl)xYDphi)7JC~Wuic2cCBd69U&)ZNc}xlIO>mBRKyKV?M9
zdnSGuNa~G^E?_d~o3H{Fu^R`Rv|PZ?2Lb|Dp77N**Wcc2_Cc#JHDA#h8v%zcOWT(P
zYbDrH#Tg1r8T@u3PWksg%lF_kH0#s%n7elj3%b4c^QD`<HTq58rK0mO_8(8k2nepE
z-izf=>s099V?+N<TW#e6Jw9#*-CzcO!cG*~jph6Op#0bS;atL_AuF`z7C#7r%vlyG
zb4~w~3zoUH2F<z>#5f_}@t@snqG6Q56a-WsQE%#X)@?*uM!RtvNkW+U$DwiZTQ2id
zd@{ZaG|5iCm1Jo=kd3zWB<Ty?o0A@$Wq%zv+{gfW^~*6&5erlXQ1k^Y8@qOETuH)s
zPaT|6Wjopx4;;wV^Lt!}yHUpOJLMG5HcgncU4yEiBB$pr9pO!ZE=|jVul|0_q3pdh
zp>TZ?-&R-bhuT@hY2|OS^#+=iJ>H{sik=s59@-L~$an0zyxb*lRHwOn5NJh)H~re;
zj2hA9_b&?995teN@i>+LId?=OpBB%@?X(kk0s}J-t@THtzM_-i!Tf~dOcR#HV|ltH
zI*w-ePI9bUjif7<c_9h8%|aXVCJ3yp6c$szCj(B5Pjc#&8{T?limcU}b3LC}`OO{}
z2j!B~!t*K{T0^-2F8!jM5f|58Ei&Axe)A3_;36hvz($H%@?=w(w(Av`)g9Kg!>f|B
zu|=aQpVmIv813rSYA3-ne8mp<?NL~i)QVjZpBdnXV;p?;Vf~Coz!3-4=#uEg0D)`;
zyj1!m1A0BhexRfyl47yux7pKTp`D?09@{eN*kcweKIO&PUgnIK`)E*Q5}W;+^R;wd
zqi)v@A+P(+mgCd$7sQJ}PG?7rhD72l2S-Qi<;HI;Xb}B${T2Y=<w`;3t(MN!w<h3z
zw9Kh6`lNbzvpjrdL;QG-&=SY}&-ZG}n^6iH&Z5RLpLUrgHKK0o*-kdb@TReEb?GNB
zVZqO^bbKxr510lB0pJJcOAM$6@Bh{XJ9FN`r2c7+{@;rt0+daYppvs|17B88)cS>u
ze|Vy!0hYfJ7DlAQ>LtD`LT=3}OSC^Bb-?UE6v2n$7ADP!Ioz>21f*g&(iRv52(1as
zeFb$4N=N73_(}GEN3ZBR1L2Ra0;o<d0BBHoycf}%SeH?X?bNU#?F0h3Z=p*=y7XdH
zoE-XIN7Le4&mas6F^!KeH_sVHA4azV19^}vLxtqS_BBu3_XLk#ww)30_CX>w{$qw$
zIcg7|h~GD>L0tiqPe`v=!U5Nt?h!uoX&#e7UN0hh#xH*q?xxP718-h_C;6BJX9YgW
zouy6~KH0p^5{U2e$*9X=#Eqw^oM0>@HP0rv(DO|a#2}pC{haWp7mkGJt~z54C+!Yx
zpO1NZ&bWfrI3+yrd0q4oRF4FgMJe8F3}hnsq>AX3Uur)P`sD0)u|afI3<urwjY99z
zdn5{0l61cIzbShP&8~5{n;(M72=}5T#lm}IyWZ?tE|DV>k<!Tz=C!+fqn`!<+uaFK
zi0B49q3E)wVjFx9PPwk+Ufo>DBhl~aw~+4UM%ac81}ENm&!xUv`Lo9DhCagi;Yp#Y
zHW<JfuP9}93HA`w!ahxnJpPsQoyAr%uRZa%<pH|+LFuIo2IurkSst7HQGnZP;eNA_
zhiaP1rY9E+j1Q&aZr<1mG7^8-f5H&@7Cp2pA#8h*fpDhHOk%jh+PLZQcT#S8%<od#
zHTNoA?8)p1P)6Wo_9pv3>!R2^qO@tBCyUhIrGr)<`%uNH$1<?7nx^mof2_I4bLj9Q
zGf|uc;VgItL9nTgzeHGdYuoV}d$IRDJ12VhIy{SeoIHDs-yc&#zrCJG_~=DFfkZf4
z2i-w`pxmzL{JUV7pW;yUpI)=P)~35CRy%J#?SBDYWHHNz1~bHdi3PNbEU|#rx9!Re
z3hS59eQjRQw}VdR5^P<x^Y{rTBeOzv)v!P<)(az9$*yLR&vC+&bds1@f<|)L!AI*}
z9TeQdc-xLnvn_EtTOnM1T6pyIvXtz2a|u~a5H{159yn6x2E5UGtt}8l3b2U07UR&k
zDb(~m+p<7kV4|;Y%h6HoxjW^v`gZkg7_bj8Q(#6t>`$8OwT*^iX0tcL?+U_T5{X&e
z)Zg|Qk&d)Eak&#HM<YLwBk5eh)Yy?X5gb(MWdqV@)~0jJcfTL<l68%rLqgQ&$2dEc
zYaedqYV=Uge^%BbSXpdG(uv{#P>{?i*=tHxisDTN-bvI7GI&q!ea=8|RX9)sf*?jF
z&w?wU*FQhTldLY=rkRZ3SkBBGP8oJWeO>7KcsQkU9f>$|5rX2?7o-D*CPXJXNkcw0
z)3{R7UXb9`Mjd0pDO_fywcyJSRN}QkR#?hxUwD_+EiGyg!o0<jr0Zz^elW}?6{*@9
z4IGX_sfZowpy|p4fADkBHxrmvU9?B4XGcW>#ypxX>c4jREJM&b;vr6p=;2Mo>Ntfa
zWgFrl>WIfFk9grer2Rd0G3Gmf$JCAWm)MFYNP_?YM^#URQ|{mmR`+D+PKSsW2b?NB
z#al#4?=oGq5*EEUSUsi0^SHGjCX|{x$4$zP;~rh?KK1s1tA(*yp53~NzBrl`BJ^u*
zUX=+PrA}q>p#jf0yWs(qnhBuY%7i~mgfmO})K`|-DP21Ll2O`0D4rIRz)`;A>agyg
zghVH72fnS=`l!vMB;)oGCLzI0MF}hIaOTBk*Z$b%Qi^xcR0Y`kZ>j#D&v<DS6_v(-
z8<z)4<{LPm7brh632{3CuW)E=;(G4LNTwInvq|D(vEcp4@Fb_<ZxvZM6i@{ZBM2qq
zsP+1EpfM?PY&;s?GLr}K?{^EXD{kjDDVn!&bSUJ#-bCyaEyslQIloa^)!WMBm7m!t
zSX{t-ik#UHvsCzvA(QQ{c@i`$U2L`K;abx<v~6QpO+HX8XZ2Aff9<@zmz{!JPqIB6
zgLpKaLWI3H+bfp%W{`&T!xIx`J#b;1B_ahINB9NdOi{C)lL7<MFDjKqTlL+cck$-w
zZxZ&G8E&;^-wTc2neQwfa0=s)>mks4#(hr7rO+3#S`Jac9ZyxC_USm7b4IJ%wNF{d
z{ne#)>PAL<uMp*>6hF{iWz|bL-Z#VIWt#Q#hxg#&AifX!$f(H(zcNK!bEeXC_Eek1
z7mg7rzaF$7CeLD^y*cuhi@zbX`MQI#r9jj>f@hxmkl)ykPjsJ&SPXWq3GUZDIdMv|
z>j%5<sf-EdC2AZ97Bo2llPw$9VhMH+?XGDWBrE7kN`hImr0g!NjDruJm5r&Gv0lpn
z20S=_Xxg>3WL*;hwYr(gpJcaD5uNrbt$y3tzOvG@3~O#O@%l}7+G`g#T(nLQLA_~Y
zr1c5M-l*&ngt>5uc)xNT6RyOE9rAULR~fX`=&`SOmz@x7#28PddtZ}@|LZ%sYpTd8
zwD>BUxEIv^_}f-Oq{QF8I{)dn?FD?OJlii**M30u&pc2Lt{$<T+hq}!6mMo{B0ax^
zmvK#H{gy0pgP-q0&4`f6XIT(N1TwJT)TdcqO3wK@DIZF?k^dg9u>{MEspwP+0u)y<
z0B&6#)7CY%4yMYlp1Q&gz8#P*pyWP<M<rk2-T!G@?6lbwH-hO%r9Zy+cUjE;7@7X>
zwTlz=eb<c)uH+D*ed|*;Qa2ABa-E*Ql@@Q#3I5cCYxBN)pPx!ohtrjy{||!TuVapI
z!}J`k1`9KU*qDKhih2ckcC$}${Au*W#=lob&sMMRf8`;!pUNw`uHFdQDyG7GcR!AV
zbUwUg+uORX`8x^kqa@jR4xmK21u(n?&Mq1~rJy*0Y3e6dudJCey6-tqHX8?R{zQ*)
zr5yJ@Z_c*aN5P>g`DTyskT)K~vh-BLNiq_^w@T#ob~lJpgp+n_uaCN)=MW1k0bOMS
z)@xFgzbbJ(PAaL$W1G%7?ms0eM5E!<3KWc-DiE2T&Ja%{kQhulzY{sy6-$UxHaHyq
z2~d4DYM!=%6t~lOo&uo*M$_)_f?UoI&ovt7Dnk1L3c$K3B!)zn4XJ1s?T{{3hy@%q
zKEd;T5LYoez+qCzdJ*Wi!cPAsx_(VJp4Oui$R(%Q{UNpe9EETpa;Q5|<fISftHGa2
z3rHT05GhNJRP8&~b+RqE=k(5k$F#Ghrq8o>;)}|Uiz?tjJ16oRy*XLhSBXYG9ktbD
z9Ct{bVbTQf?>5zX5?;qjIWfy|L;Ze77xHMI;gEjo21&HXkiGUy#L+>2M*`T>NAz8Q
z6|(n?xB2Znj-$7Vs=eG8ZqmE<JbGpEoG#_i%KXWHKFU{==zblBe=hO7E3p^n8fbYW
zAT0~p=V)7_4ZgMwW+XOp3kknQXfV0_vN-Y&)9$&$FCaiAAle@y2R-;>_am7B1uX@^
zq}2QVsQkF;ZmFrW;uHI%2(^<Yw!~w7$Z&~Uz59#4|Kp!w%*mt|j*}(ntIL0+G^6vn
zrn$>llTZKHC>GN7p!hduo2K_i%Njeb$M9P+<J14=32R~kyFXe=Ob$z{=6KcSG~WXm
z1$~y_IlcKnHp4nSG)OG~5aL(@3QOICh#Gg;f3rv2Q;;U&e+etay+(Z{6E<}rCh2&#
zOG#uf$+XVBYq1iT?F0LJUqqONKn)C8T7zw<rTH|f%shWfD!C#4bC&Ai_vcE_e$8QP
zv~*noLYt!$w2r4TuOjZ(Q*j{ZQEdcYXf0h}B+4)1bCQ1ia{N}`|CqyLs=b^9hlVPr
zefRJNkzDq(eWwLrNm|f2_u4tTr!1&(5*|bQ<UJ3Uo{c(=kJvZ<eO{&<1?Ycim4AIB
z8XT8A{td)Ts<9ix>jqZsW{2r$=2)H;qiE5J^@fvb0;rXqf_nV9CIvjc?6+4i_W_$N
zInP@z-uUMtWR%4>Ot^w`bBL!?=}-?DjJyX)sA|8U#INUymw4}hyZ1oulKfkfncq~i
z-&%;Pqv<?w|05z?F#<l0WS}%VE}xnkEs$AmxZJjz=)$PXQa@HmCVBR-vg0JV=XDH5
z41PZj<*^9*3%K<D&8#B@vQ?B6BS#eb9xZMqM4DpZ0`zakI_9bCzGCvkgnpK8tTb3#
zSXJ_)SK9sxQjbW%_5FIe`JK8a0^tvPv>O}3gYV}f-;|i?xW&I{iMyFLCq5Kvh~<4k
zImyBsS+lLr+m&b@N1rSqchiN^2L`lEBz`j{to`kO;gp};^ETCUp;OW)J!r4nv<5?J
zP8L9T&2n$+;Xt5Qh*>5LX4VLO)t@H!vBm38xJSRrO>0}X^YmtbmO*K<qo&|w^4$Tz
z!VgOb?0ppI4|!8yr1?mJ5T;CIRv|-xe)mJ|V=~6B1#Y|&$d!F1k@#0Yk=93R`1z-s
zT5C$e?BD<QJlzJr!h169J)A<G@OF=pW2D#A%C8QQ&wt?iZ7k_gog_z&#-~9k-1xHS
z-U;kOzP`qZ`mBGb|N12mn+t?X=l)!ei>IQ`tYO>a99XP?@$^0KT-!r^{;-TYYnPqn
zSw*UXo;-m6*pC)fd*GdWZp}dvo!9)(-7>nB-@-^kcqof801<|(4ThOApH7RP{iI70
z!;$<8NB%59sdT}H)y5RL9z%v3Z25sO5o=vrTNWej2jpAHOc(INSTJZgJFx8Pc^!7*
z1Kq8x>ovh@OLxC5Z9X_ax%+P07D>3!=LafbX(5&s%*Hx`?rI{8$B~_dXeX_{vhSR5
zMoLB2gv;?C%}IBR(ueC6i%Pj}ApDM&=#MnS6!7~)%Q+_A-9@B1qXMaNv6TlPv;Vbs
zinwX1I7o7dQ8H5^WtPWh^^9LGGhDn?;7-UmweuB7VS~TVK*jm(b5}83vr^&trvSU0
zzWMbO5IxZ%j(hmhOyhI)NlOf7^$NHwc*LG~%<QJ_k*wt!4^NyxH5)NiNMS<?{J|Q~
zTEliB^edp6mh)Ip2>|sL&?m?2La=pj)NjbJq@FX<>H5B6Sv&jL!@Oyny#GD+$=`bH
zB#~qo)fR3W)stc&2%L7mR)4;c;>Bt6)@X&h@!cd@yWGb8Xw3dBM*IYahi@|CE108^
z9;}o)T67CSN40_~ar-U$n~`Rq-qnd~u;IpK-px`wb3A0ka}<>bA`upF5@Y&kv5cK^
zA*20A;=Wh{p6MS<##F@*(<1#fKJ(95M_ufOqlrp&RxyhyG{sPaAEcQmK}s2%X~B0H
z!Cyh6+aDg5hQ%5GLBLSLmk-dEZsFpEj03Pt7&@$DQ!Vj-3(}S9P+nmXk%_N7aIM=z
zvshaJV&9-gq1b^yb(GtPGB6urD|qCForcKkH0a(2yjc2)K$4t@EQbXMQ-8<5E@%tj
zT|HZ9z5RqsbSn3eDS3)+O$6`_nRxj)i<X!SmRWd{@|jh^y^7u|KcE(sT?<PEQlSi_
z^~KKiHZ$$OobabSB}Ap6Nqfk<C`>MWunEob8f~y=$~m6>`hK5cETR3R_GO`MZy#HX
z{_MW0H#v8;&4988q-*d>!c7xSCqBPf>1ELFJqtqF*U8Z~PksqoSevC6c$R!1Tr|g1
z+%a11-B|M#OskO8*MPFpY9--}{Am(<%qqA{tYhCUD!E0ZDU0>ai;4b(u61jb19Fe|
z46$atkW3VkpAG)bsU|EA;2FE&_s){x^QtYQPH6A{K1Cph8=ju~aRLw)?Ny=z+Yx*2
zy2Ri=9GBgqIRTicnlNKgr!(XjqVgYg=<Ua_YSC^{aOXbS!EeW&FZjI6OaKLRG5|>e
z_zG+4>Ji1@$lqJ;N-f7K4e*wg_!#pYlnzZ|0>LQqSF0pHJVtKCF&TRV?l2I{NI+jF
z)p~o5rQ>4>-)Fcv_2+2!2^~vpvDLlTPrDOy4iDcq#A@des#VP6A$MsNg5K}frQJL&
z%PDgCYV4CSOaDQ{pQ4f6RFV`O7$KCK6p+$4eYs!6MlY_`B4D8IHLO<RY##)h{=@Hw
ze>e#Ls6%`4xINf;z2bOc+phwQR`}yXrz*U*7uKA-TO+?0@h~DM&I>q&_zEsGLP%iu
z@H$Z@2Imx9BzFh5Ja9NJ*0vO356=b9m8-RVe1X>h9AE$w@$7#{utV7*fAMzA|MG_Z
zCPVMSaENj|0vpWbb%Mxg`M<#pBrehdn-DtwaX$;`)mAky_vc3!>pc)3B6Pde{5__m
zLPWX-M+-jcT3}G<B~1f>D!KBwY&B`qd5wMS>rg<=A)uCiB~e^g;lTfE)-Pit**=7F
z!umvqny@k4)^uGrYNG7RAhoEbHET#6ff<q%oX1N!FwmoR3&9cWcTG`@sN0xdV)mgB
zWbO@~*GH^P%;n&?sBsNa;_Tf~^jmxhQtt^`^$#sVQHcLwxA{7%>QdobM9{DIc)1?Y
zLBV^6d=CQ4D%kl90;iA0-F1T~c-#nDB{+bf@YOgeM7mK|YKrMFH!H|+VYg<>;}<7r
z$kmNb9N{9JCk>hle&Zil*Le03Qvj2362wjz)d+l%!z9J$o~tikJgF;1n6l$-mOhG}
zf7pdGKIF!vMuZS!=zSp-740zr3<b<dTAdm{tfR+Sjl6MAtEgIz$YdI>ln2mw;7l4?
zoBt$YRW*;d$_L2OgpSu`+1`r!u^gVlhi)3~8P`KJr*Lguc59U5`Y{_ROnVafaYvAz
zYqKB2SzIrG*yMLUs=EmMoa4xc?M{WD>Ul6qy)9W1Y|k|$hBE{vsukC9C!AK27?L#}
zPy3S6`$~YOR)V5!WuI8D=vm@vPJ%)!5Mbv4_k%1)&=L9fKEs|7BQqlZk&y{afLcV|
zwje@Hy>O^{B>(Y8K8YJEQ~baWA`HoQ&jJu)*s}^U^#1v#*OhSCZ{y%eOg{5h7+Oa<
zTc#XAEP!NZc5fiHt!tm7{j69XX)C#hCNRT7=~2NTYM3%W`T4;0`2c#kJA}ab{D<nL
zq3dD&jPv>73F$vanv6`FI0~ae#{U`s4hPYEz_dQ#7ds{{tUFiU(riah18*w+Y8mZe
z+V*%NY(0>!b+NIk;19R8ZF(}E1z43LdcnILuZwVZ=V%tdBfMj_!CWgl#>jQvJ_a%!
z;(5=<KqYh7QdFEvi2@{<!^#H=jl~?eAp^(V9=FC>Gr)P5v!-)8k^9n0F5jV3L261R
z&Ylq3%8*&Lj-x5OY?;a`lMhioM!!Xu*fu-<Hqa0PeNLCTHa~7bx-L8FmJeLF`(-a`
zZ+c8O?rKynjxfph_;!cK2;s2KzU|K67%nj)gn+J&qaX7gb1G3BdZuk>#Xn-`$@nzD
zfDG<oO?j_EP)tcLnKz{!Ri_4iOD_sy9xw@C`C-&3!mpMO@Jd_79vgx}jao2{e+j`E
z#4Rp$h@Wg&YcH_)y_2|n#5PpAO?o<Q3<tq97l6(T&3LHVc@SR$+(U;nP#<w<ezP6%
zW!@!z?P}wgbT3%WwE2lMD5l3&xP5L^8VSE}OVT#S2jX$@b!JLIj@=Q%4%sf7d$q19
zmG+_*m$Py)F1@>mWQ(qnNZ|HYR=O)ES6AAf5?>6*aBL97^Wo8#exTf0Ok+EZ`?k`&
zCoYY^3lO<`ihP~r`F=ROeXtv^Y>H&-d5$M0{qTN!c}Nd{*c6uXK40#Nrf`m65?H^b
z6MdPObpF!&^3c1lEATR&=Xv_)Y(M`@mh&_#Q3-%MDG9+J3cqy&aYq?)eD)F?18AGx
z&t`mvE~qy9QtpVyycZOuLdevZ@g7ZOxrSXXr`r}0>pK2xO@zM&=E(tfDz1j#g(c26
zSYq&x)7uV*e|m7b&VuU`6a=a@Qd|Z@H2NB8+i(#W-9BRQoLmXP*g3gqOg0~&GWf8l
z-9V@4>^8;MXMvvS`aMh++DW?#vb6$>t)}4k)>Kbh$*|-T|7p{yi-3J6CjSoh$Mkwl
z&jXD$pqV<7vt6)ljL!+M6u_ylO+6qhXj9}oyIyA4waXG@Bn91ec{yO(-`o_IBU1j#
zPCPwn+_`5cAi=!3+%*OPZ`Xu%f2Xc{0#RSWSIE|3jm8Q5)bCAHmOvnY)AN9`Lp@nR
z#4*AifZnzwJ=2#-Hfr@$82_xL`x`Q-{)e=0tI9`u9rxwqEi!tJul;)7W*Hr=p>XTV
z)76%D!S5aloK`1C^IsQ*g#~?;*FhcczLQ%aT*=08&L0yy>4t<8q!v&CRs7ClA67n;
z=f{jwtp)oHVGC!AsSh|4iL|pOr+~JQoakeHzJ-g6zMf$$clHL%J>5U{8IQivw0O#*
zv%8`{GvK%Zn@s{1M<_tCqdy)Zn^MI;jI#m2oPy|v^^7SN97lWu-;IZ9XkU@&8Va3s
zO`*qtWc5BL5Dl-bHGJ<Bz+I#ZrnKw}99&mC<ags%qF6Dx(4W<T^q}i5Mo2uG>$f<h
zJ;lN^4!X^UZwXegFH~rRHgxyD+hhW91|yK5hZTFTN@8-Y!#oZM)j|H|EdTLZNBvr5
zWBY|uS^cj@BB1CCI3CakR;W@)&t;JsgiuRzt((cExD^rt+Mi91TsO=&$+nnn$G-*1
zowDb=qyO1vfg8HHSAjQbvr-#}SS!1>Ai_X9MMCa#(VX<awn6Y-ub|wxQhqdVH<q?H
zb3dV}2!L#eR#~uPLl+n~A%ym8tOr&x(%Km)UMrYjVi1O~vxXq)O+jG0ogm9eYg_o~
zOmCU1E%*`rZck3)F%~~-*nqq4Q<n*{!^$R|ql-8-dA?m!WeygJNMMP~-p`_7B4bVA
zMoHmv@ZHGBpM_RkTBAh2z6>GAB^DI;dkpQQnf?V}a$;GfV^A5J#d6>iPz*G!IyF=c
z8{0V;AVm>W4i%qVY&Nb080?v5zcwA#fqCO*I&E;)`c<LI5jE!WVY>xw-&=BWmv<Nh
zk|&?G=ylYxJNZ1aM@6rNC~PyrC1-*Q?eTBJEZt9qdU}<LZb3uutIz&+p{&m>qBdq9
zYrvqbb8E4(vwoIo-oP4&Nc?baI>Tv<Xgq^YNzhbRX4%vidhq-opDSVuwQmMvp_dJU
z*okp>((WYpSqO}Rc9x5@Hz<kZA`{?3+!L#h`(a&%J#P=Tc9ast^0oHf)5Rva$;!}#
zL{LFZd;G?B#jS1_ekEe~AF=-L;0)br9#tD#>?-j|K~UW#TzidnPm)&IYfsM|0HGzB
z_nuDvY8Zm7#Bcg}#QXJ^-S5;$@nqDL<lB0C<Fr<>+$~E$xyI{CR?{ObfkPC~68CC=
z_QWcK1-?rfYf~N7kq}+~9^`}UDZ`{n=FSx%Uc)z39d-F4`5?LOs@mhksr)p8r>Ftc
zo1|;_a}94(cfz<7#<?<k81&PT>hLcyyFSfg(1kyS<$es<Gb1<n!x1zPcn(6_y`@-f
z`*9r*OxyCNrc^gS>&vx|;r@JF!jT`=Vm4oG@cK(6;lbeqRw=loR%cAW<_Te~Pa_kx
zM1Di<R5CvI3!)!L)~1VDxnOaMy&{Pq@t_X{QF&uVA5aI-+53Vx0>o^RiVe{t;HH+*
zOn4)w+juMc86bxXLtY~c*GAwX=a=`oIL1tr3tL2p))!#e-iz@%*=nt<s7IN7ZtHTT
zb9m*I6g*Km$++$u7K+%YZ`kJv_CGp4{^K&og=r!?e(aUk-Ru^tDWn5~Pu*z2L4N12
zmRt%twRi~SYo|(?_5YIAqX;5x{A!vsQ3{=7pXn9E{;?}Sw)NFB1?LEv_<nRYGSLJa
zDk{L-BWW@wx&=!7ZP!{(ifo=t8v(xCt0)6kxoiUOuDiW2^d=adDMLS*$o%`Zs~96Z
zzQc9<@8}etvp^MY*o5b;a~BKjM%<f59$CMHehaiYeL<5px^Db5QR5=^m8gA}(p^yV
zRb*O{i(4@d{*^FA2d)V_?%hB#4yL@UDw9V;8dw9@rMssXZ$f0sbrkryNo_IiR--|E
zxt0i986xF}kPe8g=9Zj%y1Kb9Ij0{s^=gO!@5Ss0%OdzDmjq8Pi+k~~oR&W;Aadc>
zB^5_M^jp8*+M;1!>}PP_+(uy!p`%{)hO2ekpQ?g6S0&sl*Z%C1@4;Ii^>wbbJ(4|l
z(Vka6o4g>5E;{<y94`za&WjU4Ai`+Sr39nG<?lAThh{Y(Eu<8@IL%`(;Xc*(1KP+L
zPJL0JCzrzY5kO6vu}fVP?8ZHVa{pyFnwoe!3t(D+52+cEEe$%nu4!=v8Xo|ZD%%Y2
zzy7R*x#1sV&apO%`Xj?+9`cFgWX%eAk?}5MW?mi8!5o(=7Be%}hu2SZZu{3P*-UHV
zI6R;9R)|qhmCP4#xp4(`5N;VM*aTcUWcJ<-lx7lQVZ8o**4j!s&35DJzkoj}eK1+#
zCw+OrP&!51-$l0;3(=5dd_tj)k$vH7+pSlNpYSwhl;xDmwZuJP5bd##kpcLv`Cjo+
zVKtxrDg4W3vb$P6ydTfy>3$<kxx}#M-jL2|%=(WcbHD3yvdNa$!R|PHD){b2-gTG%
zL5TTqsOWzqAXY6`*E((Q-6=vg9Rt}mWndMEVNQjQrU(*h1Y~#X#3Zj7o}>iWwa;Q!
zIW9BQ8VcM9TFQ%0p@_5auOBF8?MW|d#jj~gvMjoq`}Hxd7T^=<i5ab}WXs5xOreO#
zrIX})dUQ7rFcIBIKUa{l1uk*kV1+TStnGexSay*dyvT@C`v!7OdCOo%!0mPuDt^_P
z6Z5rOqv#qDPbJ4u(gG(EDEZrz4MVk4lC$$svHf@B*k1DKhr4cjzWtjzW5x{Q%g?hR
zXh3}krr*uS90!N0q^Xz1(%GLC5?l0!ZXsteQnEX(Yh^dQQfD<YaH;+n%TLKk2{VHB
z4G;~xe9>lt$Ru^t2Dy2Y>`5Q;5D7qyL%Fd)f=mFe-DKWPh+#Dr(tZ%NcwXd80oh@K
z^ZdSOUrZ2Ef64LMn;!U)hDw(e6QO7pcuO)iW@DbM@5zH=<ZeA91+!MT!aWt(5~$_h
z;u<2c3lNnfp*lrF1#cRfv=lxK`IlL+PGRZ3O|uq3O>=A?ZkS_G3>(2pj!)AT9eNu@
zp8ogcXRu>1*UZP6>%o2n1&~|$q^nfM<yxve#mSXhChQeACSdewhgasl^BGfO@uTUl
zI`@)^gtzgCv%seW2h;QM5$3><Tk8L~lL-SP#e!5nUNxEih?BUB@&a4&z^8?R|H~(;
zhto!^RcI8mn}7IQZ{Du`8^D+j-pJwPyA&9URDYXuE@%t^O6Jo7src6&$OEZ$!bF%c
z=_^0StoE9~msj{sOF~&8SU2l~LU_3RA1YMAs)T7G^>4xP%tm}a)H_SumMBgd*GzC#
zdqlUJ>3WF)l2i8ToD`qFq#tK|?EmX#*PyS))WxKn?xw?9r^$m4$K+TMJaVeM+j<2g
z=aS(qkW??<d>Lp9sEX|yu+2S2Yr-3kLj*>mPD)Qf4v`kJ^9MUPP_X2`DA&EFn0F^F
ze=@~F0-i!N1e>4{U+J5s>=Uj-R(+fDsR*`5JnhNsuBZnj(Sx!Vv8V}<<=sAM6cR^f
z(*aUNDl8@6=&V@^-p>{!<ak5h{2*V1J6>UL-SNFd;b@iFai4#Ko7;g7nf&mtBb;Gh
z+~=t5c*QzDtZC^EM&E8c9|QIiI$d}z?ZiM_pkG^G!o7mA|Ec5ehhs~Vm_x&8?vnX}
z%T#v~iCw)4+dQXNtt3{9R9E*(7!NT<;Rra0eM!f0Mj!iLtm4J{OhS9!*|E5#_T%rb
zeQDm^Qq7K#vJ<_mKGit;!v0n*LL~mf1s@LhR&j}Ys=L{%=Uowu-YICP>XB>`2if~S
zamb7RY71}Gb!PGH*o$AxPYOHECpg2^|DW|WTnO4^wQ!{Cp9{VN1!+fc$H36g`-~hO
zm?%|42XMAQbhh}~2+Qi6?SahPO3|32*;ea=?*yfyC6OFfSXBiUBpJ~WNct3oU2ZAR
zUhGi|(BR?ouuJ#*%2)5SqUJnDcE$qjawuJ~$`;n~G$?qv1i*y}hAx*v0{~jTot6dR
zvk`>P0FqH)eBKA=P?wS)n}oVH%J2~aVt7{5_bsw)|1F~=!T661NQvV8-uN-%2zK*g
zqReXA^I0*ns+1hKg@O8<tkK+FKJq82vF<OJKfT<-0=Ey^qJQo|O2fvjW#JSH`U@Qw
z$!Rb9HqLLm^hov1rG`|7N)tSk>h9m46Ox#@s3;KDwjvR};Na!9peQ!#E0B~-zSv!7
zZOEpv{@kkmQr3#CySKq)AL8RcbTmzg{h=~KVZJEl$=3!A&~l?;nghA{$(hD%`*98n
zJzVeYXHn){BFQ#fUHz(W3T(VdAKG~8+ON+wc2=&VEjJK0l4slPwpRp4AP4&pFf|S;
zCpT6-e=9$6UeeTrlt{=)7xp=*z_x}dpFFH#CwAAbZ-oWjI03nN-c!STUfip1f7Ian
z5LqzjZ3U~}A5dC3+41esj-3(_7^R^0;cGs|-d!;Tvj5rHi>v;dhb5kyb)WplLpceu
zibyaV-uv$l&ofb}KyUl+EKJy(jK?BO^v@uEx?T<!#@#r?Hoea2#s4W@rQW4FgdBOE
zc4B{4$CZ$bty1p?s<%^~dn2Gi@TK6xp$*jyUeugI>XakY7vkv@wo2Xt54t4zl5ch{
z{4|U1U=OW~0!L^`F%vnCM6`MUb^xGw_+jrZtlV9tCtBtOFa<4dKE_;{U1h+X1~K87
z2LZ^XR-#wln(&%opNE12(rxSLb1YXE_3}FfU*j=t4;E|A*F9nE$;YhuqL?-3GELJE
zR~=20+{L_LTCorbjT)_fq+y~A7hdbqrhK$fh`^)X1tecN=Ut20UqUYRhGqp5^wKjo
z_fV}U^>$tm>^?ojA-zBg9m;Oozbu&NR!M$88mn<|d~FdcvwxpagtW2&ieReGZnYg@
z#XMf)ovz6>2k@{3lg|g-h%sQv%Tyek^CA<XAm#kL>oY3<#p4GnD_P)nF2!R%cDw)a
zAy|5)>gEi%3(Ak-occqIapN2_>1<TjryA<WGNP+FNTNQDKs|aLd{&INZDx5gIE5*d
z%|V<aPVRQU{d2;jzI141P-v_jG_lxIF#S0OV~83(`uulR*=|)`(`Q<6qP_)XqARw5
zEwbzz)!d(SizfB*I$R<sCB2uqXCu1_5_wPN3h*UH%KwYVX;bI8|0Cy$kH-r2WYcBj
z&qX8U{uc6SOj;m4e3IS_m;4~K!+tpHeaPAd(|DZ4JT)4X3Zso2bQ@wJs40KFxwt!k
z60q`EmWj!it<66KA@+zp3D5+qvvW14$MO0dbr>jOwf>&y-EV`JWkV=X`}%p1M90u)
zlp+JIAn!Z_R2J$=0Ei-IP28h)KLw-Q)(-g@4H7kK_=uhqW%Vmfb+nb5&uJa!ARufM
zo4+@=qW%>(6%o#lqGby&oHao`4$ddz_}F{ubI+7jJqFAN=usCe={(+FWn$g22s}kj
z(<;z$T0|FRVZw&vA$D-@m?ogm9J!#WO$nSBp7`B^lhvcG#TS+v{?;a>=YCk$Cg0wZ
zw}0l{p(SqOuSFpgr2cJC?AQ-k8%yQe*NOJ`ic~}*h;pVVcC+ngA0&QzVwQy%B<aCo
zk)LA5v{tErDHaXaZulR3e9tkgiZ(Sp!V>WpFcJ-Y8)~<CSG}a5z&-fW(>YvFL>#qb
zQOuxnLsWM0UOe<?ythe8=58O<*~cNk8`F*)9N+1+<Qu=%<Gpim*i_YNymlQB2zAe?
z!pS&|?|hm9xQqJ&C=VvMzW<W13OCOP-w2yJ_B@RyrMPH)|4z~S>lobfM=JqO2vI7{
zQIWe7|EqS8{Ch&rQ>)x`Tf+PAtTD`SJd>vpUS=3GeQi4l77Dzr_vtR#tEeA+*X`W9
zsYlcvH<HmDoO{|^6_lwfguC)tDF#QK8Xdk63R^+L%WTuY9Um6c#2Vp`iFe#I1K)l5
z+>o9-EV(hG>VASrxni}JfJV~x<Pv3*Z|os!9hO}fa;>4!IJc$o`hoqd;0J!a>tfJB
zGf##|2O2cOZ2~~^U74tD+m@G}<_Sm0RIf;BRqR706ka#z1{DU(u+-hCiXf3Nbm0k@
z>Ofg1Y1X3pVr{?C5IasxVBV=?fkH$ueczcPc6s1!8s1VB_ME6tPQu9?ErK>~U20ik
zH*X>!O&`%Bis(MjDC`)0B<#I=$u$qP0H%eTUxVu5*{>wZ0^VT{-@2%P4R|YB9=>s_
ztBOOZuT-V?sxJ-4^n2`^^nIQEXZLL2#Av$^YKf=vLuLc_4a?ml^C8X@(_l+8Hgg$X
z!$Dci_=nAHyV5D|bwNqJ4Y9O65Elvn!D3iBmAQ_WZwjFY%6Nz7?W~Ez*pHl<gT2Rq
z)aYUS6di4pp^!9uT5%0LPKJRA9m+XGbP^GLN=y>kcbJLZW@i{>b@u>nYrHq!w%yL|
z2t{7Cqgdtg+ZG!6{bKinF%0+TxOV^5x!dOoW-&HK9`gG1$prAf_4f*dmX;QM9#qz9
zxS=%ly(&w0G7kCbdV((mnfV*3W-18<bh={<7K#|T^_w-(ah0jjGY*J5{GEz9c~VUd
z&wI@?1*SCk4(zkGG5plfj7&_TaHvFXl3Xh%Xf@K5Hm?JBhbJV8uP(u0p`OMOT>%cy
zvMc6Vkl)1UBCkrF(6)<??Eb)7`O5%ht$h7&nk1)|y{C1H=@-cggZ=tjWF!D|ZupGm
zOwPjf3(|^fzLKCY5mcx`rfY{9?#)8ifC%Yj-#!|i`ssJ_Uy2e8`-3MlcW?X(wNK*?
z@_nkG?7iEL#ayo9B_LMuF_ZmjcPZj!XP02x$a)h1xe(shCBxTkS)(GSJTMAE_nJ{>
z^hQ84#BC<_J;Urnkzk1S6o<#*Hng}AR^d+6sPeH(dHZodp?tAKs?M`peDDp&S`HKo
z=pgudETlkBOxV>&+^klh{CEl1KOUbCn7{s9{?zgnPg%T|1s+8^<_O7c<5R;%-2Iia
z9{nJ^r}@$Kf!K$?r~F*>UVa7ASdl_xPvsX#so`8U#Z<}$-V&1lO`nKs&!7dx(O){_
zGw3@yG1PWc=UgnV8uK751Z)nY4^zA1?CHlge|au5)%}e!L{~@SLig9^U``J}eeR%7
zfjci7^zffpQ(Kz>6BF~|*!0EVpRQ*bmTFWF3Kj~o*i_}Y{l6~re_r=iIr#m<q?1JA
zr{_C5UYX%ZgQR|d7+(9nKUX`Qw1HQl8H)N4Bx_$C7TzMb@MAF14W>-^dT^-<Qei>M
zoS)xJF1G3h!^KYuvmuHdQKg@X2!$mhj)&s1s!uq=vr0>0^`4D)c@49t9BQ~Ul`E~0
zm_le5AvV?kM8fw!)NG^n=C&(QrWOqC$l}Ep=WVi&B5~SjM<V#p%)SZLpxgG#q7*Fm
zetQx6(D7ssz!oqzO@KEGPqrmB`(o3bd$=4k6i;Gpr)Y?ezmtE!4)=im0EL$1_SL7<
zP9mpZkwm3LubGSwu>(Upyi`f|eq9%4NbVTf=WA4i-dQ9znZ1I%B@pSmL&{Vo7LnXT
zM&dO17W(wjMQsFPqo5I<cGpKX(#bWr#J8ySF^=$66yvezh=<J#YM8R2>i1H@2nU11
z<2*U9lfwtH1xH<(QirzTdwPWEQ&TqF!ph>mD}BGO?#-XyxLk$f9+)!ob^$zSYdbCD
zpc{10Ss|AkC}#F8_jh~%=_3JgX8BswEwS@2cQw2bR^D0s)Oj2MR86+1W<)ry$Z~G&
z=~l=7GYlJkBY{78q8RsMxg3unwMzFAq^j90C(@@`cSU48_UYqU!Y2Q0tjyw{kboyo
zzj^N9)-yI075{ZTD8J7J)9ECagXv)tEAye>e=0*<2=ghcrBV}Q*)Lzc|1E;^5tEBV
z3l*3TeEu<>Mam68)m4iKpA$-+LlpGb4@?4#HMNerW{yQERzQ#;oBc{%4^B9PRkJGR
z!tU>hV%UVda~N8w7<nIFg*s^7M84!Dih~*Vp#5(`cUhDG7+_KVcn~Cwxb}xl&ZCQD
zGlQ;gIrzsmY73V%oH)NY-JVv0eL?*4=2r_46lia%Z>?u!D*Gf`VU*INwvYe91VFqA
z9dNut^0dloHoaeb`WXt4GH+UmMMQ6-7zW7ZA!2^^5H|fB$1UhE#LP@BX$q5U01Z>&
zXv3m_v+O*mpc4uc%jxe2Y~~Ywv}cs(C``uNYdYIRrFhKF^a-@JwIyy=JTI96vE7q~
zQVf4Nl~Aq5`W^pv3o6#wDWpRnYs6#{t0Md9-RzeeIkoy2Jc~DawQrB-E)bjy&51qe
z5jON(S#HP4A`gCv0oBpMOZy%CH+J-#H)!I$k=e<O%L+oTtEz%W4;m1;cYckfUKhWu
z-`)Y`mb)RJepkTg!sjLuV!AmtD1GXaDiE~PR26P>d&(MtFDZvnE@dplf73oydC1|`
zxzHk$pJM$T0;s9xxo7_8l~QhJCHwJl{uzl@=cps_<zbTxuGURb{_^^VN02yNLH-W_
zo4Qf5bJupww3GK(7{6)$;;BA;S@FMvpwY-U0quu{nmNqkXi<;7L~?rm3ps;sKF77!
zRQ}qa;;T1XB-ww_pql9YS2(J?>-7eljH)rm%()`yayZnV!7gFBH2nQ5M5o{IR<WtH
zKWTMnEaJCw^j?$O02I@grqBG<U#jusr0%;^s(X6jV_T3c*k}OdEd3iU0h8}ka=5n3
zpKPDx?s5vXuqMjNHgWi{pgDNQ$=#ZjTeG4M><Y;A0@1rAOCjtvMz72J2g-1=$4{{y
znw=~i=c?i#$e+ZbgJqcH;R(T3A$al)HBx<84}%sj*CwfYW}?xW8i4{}!y1hm#cAH}
z*9e0h)(8SMay<~83)v-(Eoal^_++7OT*Gs&)L1u2cr#Ns<v`(Q@g;=2|3{4I`8Ylx
zTis4f+``3LXhT_e@ds7JrOTP6ux-oPVO#7QqBpKL)5JJdgfVPOw1~o4O?`X&x`q|j
z^DK39H%yX5E(_@tJ9@=%)}m`?P#J@XhC149ik;Zp;7aJJ@&+LhtR+MHwWs)i-s+hg
za%_ipvxn^zWN_$S9dBdm`^lDgx|un{O0$|SXhGj)bfDv%w<_5drM^M-p2$y_A(9h=
zZb9p?#~AP@pH*6(_iq-h<m20zp{D3W?I1>jp=H^X&zg}ZotsZCLMN|tFIkPo{~<7}
zSNYGLn{4^02(}PPC`xaTki|^@J2o!3kKc743IC6E9A2GfG9YkWf;Sz6xYXfWnPRE(
z;mBiq>xMbG+A23fH+vX|9HeH_1{<xyNnbTA94N%df#U=H*6BXm-g;$pFbZ^;i4W}z
zLr*9o-;S)T3uD~Dit(%j(y}i<eNxsluM9RSqC=Iv+9ZJnc?dK>tpRTbhV_!~B9Wes
z(Ch8nKjx1F?+JQBU;R)(0ka5|vme1vMk^WwI-suA-UZkgf&qK?Y(5i_I#>URz)7Yd
zs-D8d9fR(xllNM0l-EMy@Hl$kO$n_<7>LWk^CM?<mtDG*UQ{1Nq`qynXzRUwx4G+U
zSLm{CS4oF8rP+2*NKwIZR=L99A3w?HkL%vO@lBIrQVf^qXw(2kQ!Q*{!6~7pD=8^q
zQ$U-}e5zl7%&Nwu_oC-g#5nKs_=~6j6@>6MJSF*8`<`pedLoO4H;FHK++wgati5g4
z0#n`HXGK9cm})TqIZ^feBpTUj@7Zl*TgURFZO$ezs?u9LxijG-N3u8{*gO}~YF-@3
z?6$F_*=Ih$B^t}7ZPJQfC#W(9NrwR@MeilpxUAEe4<!i=x6&tRE)0b?raUp-Vnnba
zV`Ffal)Pcc#HkxOc<uWmmcx9n*``blz-oe=msHFXU@o!VS~$_&{;RUeYP&x?{rIvk
z1o0x_<u-vIyT)!XZ#Xb1&dAnWecH49|M!fbsD+-awyjR}LG$3Nr#?zT<FHPfCzxow
z4-zfu@|gJl(({<!d#o#`yBjrpV-N~Rb6JasON?~HwfG_`Hd%J!bFadht+KwCa=S@Q
z$w`>w(ig$N#pgqaa0)mnI${$Fd2&ev@Smul0AG>N=e}XdqQipC&azFA?s^qdGI(lR
zqrq_E*(BfP%yF{x`z>#&WmdJ72jkUd5?*Hz+@Lj6CxAKM-gAoQfdfhqfXaB^b@4b^
z$9{ZvWW-R`H0Ph)Rjeh$@C?L4PJQ@Jd~CJ*&N(zbl#sU~Yo{v32F<MeiBM(!VQLIh
zUXSUNhjsYcD|%pA3&|oALsC_eoMz$5F;CED5pm&6XPXdaNw$Y@OfvVy0K2iODuxDa
zk;&+~&;NnPAW-31wD0XZyR!@3=zZI4pV~bl;J4h*HhG0FEYR{2kP50Bkj48II7kr0
z=wSF`v;1<kl4AF{pC&`Rx0X0NLGT6e%{nK4Ap-8JnL9M7t4wAMy%3s;Z|zPWnDOX8
zYQWvDl1sR)Uz5pfqI;EdZf;DTBpxxYp&!C-)Z*NWbAQshW@9!lm1j?PvfGscdw4T)
zyldkdVV^EcCeymUKM?E*_EOLX@lNrwitkp^(F+&q_FSZ0MC*3-_qe*cX1IR5%!cc>
z?v;z8c`pHn8M?3NnGc2jA6ccfma#Dhd}Z<*cnGyS!XPk66G0xPvU5Juab$c_NMdd*
z)Hqt6i~3aN(d@IakIqC&C=RsiWAogMn+@{rr;zUVFfxfC4{Y?|Ux!9DQ?#;3G|Lqo
zIOD09ZV%k}Vi0l?8&|>xNYfA)BQL4+Y{~sQZSwdq)!UB6gL}j9c8U;`JvKz$2E-b0
zhZ`bd7`t)L79X~|n@OPxZ`5-F*L>dgD%z$*V23%z&wNz*Jf0j7&Aok8E-zE6gCtK6
zCV}Ed+%X~#?wT;CY9Iu*T&yvLxjG!Cxo&*;`Z_^S8oZU_K9>B~v5C+jlE^e~Q}^-X
z)Gi)R>wWiMcd{=G$W|=cx^av=rg)xQ41hGp^^=5eq@P1s7(yL|^WE68-eW;WmJSrA
z2zgKyAqxi<lilqnH-}KQztfR#`23^uGwk6DXyd)#K8vmy`f<)QAI2bN530=E)$f{^
z*7&GHTUyf+%`eurQw`mcxSkVHBEt-KgpstiC&xoSHvoLD*r*c&Yy<J#oD(zW!lEZd
zOcNM>KYoll$SLJI{PSlCcT`DLxPQ%lZ2bLW`<TL6zPf2=KG}(Fv_(C?MF9p5a&%Lf
zch0?c^;i11&NoWq$k~W<I)|1O2tFoXJvd~Q*azSIH4=Whn%mc|^G~ZbXc_$|?g1~=
zUBAyy%zGE~89^548w-CrTvq?Yr(uRiO4T_3PLZ#EqcKOUA^;CQ3+{OpD=($*_akt|
zvb&kplhTi4$QiTZ29`mj{;Fk$4obaRK1?}dV}X2e>7%<~!6lPhV!vCQz9OVt075|u
z1KI<OIs8q)NyJZDA=dBJmP}1%P=7Ak(m0{?dU^gP6Ht<g=kF)G?8Br9@1)`vDAWcb
zlQGIpe&uM3D(;4m_n~arvEb0kE!0KSBo{pwya=)qHbz<tjWny->8hapoXPhW^_v8>
zP!xR>W5U5UL)4=kC)qj|t@vu09QgeCeTm<#{oN!p^^eH_rNThyh6gqN^wqhzJGA(n
z%JiR(@jjKZW>cW9>hPFUP>2(9G(<cnzi0R^6M6FbaOv0)k`{Tc2#zVnSa?MfAvo#P
zRV|O~#;*jG12%Tdmww?cj6+$SZ48+G+wsJiA5({iqm^94U}Wq6aqI|$qbOYK-x=N&
zsd~05SvR^C&9Z#?iA#*JXj+kMsdPwh2M5UoIFfZHgfl-c;S89p#U31(M-x<zr<jW5
zXgd}VHn|eBG(?iMobiDw17E6<53nb-53p6=QByf4);mnrT#jF09T0>?`F=R$8%sLl
zL2bkW^(jWy#5F{pKg8bT1YIv*T(|YE_n`iMxg~BJtC8r`h*viHwcfLTFARSWsAq=5
zUG4A_MDy!^=ShC^r|X>)q;ldyJ>TP1_ChW2qebFbKRaE4fE*s=+sst~BY*^8yvOWB
zxi)QXhfz^Sgjg({oROyUUvVL^t?E?6u{X}w+z5J$T@(GgDE*yo(}}K={o0;LMo0ko
zMlHQGL?7@XN+3Y;N(|TBmx$jGAOx@tdlQ{fnNmyE@%2=M_UVd+`kla|%hObM6Ec%p
z$ePQEE|MEvR3Puv(u``9TPH`4GK$BLeno?cZ8fD~ntL!nP-+j}lN|=sA7;Q5R8H{s
zE(LlFzN~nxX+emklt{Uxlsxy`<TQ-&0haTMbWg!pqR1rMsU=*tc?Xw?`%&kM9K8#7
z6`l|`Db|y+iwodQHPu%G5t~MBlv;!4jw^?Fw}MmX;>zKOwfsX9QvLE}=1BFN*T<pS
ziCF9!{c_ZWdvByy?1-eONb-y?6i>?H(gk#vEP1w)uDO;j5hmnvWc5)JO_Ib<eSVC%
zQeSt|_BB;TlG@@f9bXJHeN-ds5OrFPpWni&!;^O2t(EV}$ur|{5jjS2pU*<&Jf}z>
z{qyYjrsxZ(cl(;+SzE7I%4rG6B!#RC!%uhq4_9Xy)@HYL>jZZzQrw;55}YEXK#{g+
z2^4n;1S#53oZ{|o#hu_3in~(??ogx!io?mf_x|=d-+6vK`JL-Z=9+8HG47!sHOQU@
z9RC%F?MFUT>T**xQ3x5Y>;D#CvCyow9(g{VE;7=iyLJ{Lmgvv?XXj3`P^sTXi=OlX
z<zIlE|1fo4TF(Cb$;tdGGPbIU_gfhf*#h?T^gPg+n?jo7JuwiJlIr>5)_al(L>s!@
z{M{H;v4TwO>A~}tZehKCp#|)X*sO&jX6BLh=tLNaj1G91ba@bM#UDDPU4(-{xKRuB
z-IUbOo@qE=SsY1720MUE1yPdCqMh6j^cQsX70iKoB=u`r1_HXM6Fp+t`siKsJo4Af
z<L;^}SQM){rNEA3k5*s_HAEkt?8babaC&9xR12ekk3p_2$lMr$CKy(WR)6bU#m+|i
z`pfb-{<Ayw)4!8c3nhF`6K$2lS(W>tD41rma8}%0E46bhVO$S4t@yMKM;I@jd&lzV
zDsLL2^<8d^%hEkJVRG}@4s-wr`9tGfYTBW{1MXG_XELa<+?^gziwF+f_Vf$m4#w?O
z84YBM$hpttx1^KWH$MslhVKfZgNQK*I25{H;DiD&#G5LOOGDB*`(f6Y8x3A_@63S&
zE;GPui1(66{RU;nF)XFq!xjC|ohB(|;OIH~=`-v60QEg1bNJZsWgt*Fi!i8L&_Baz
zPCV}@vUY0~SR&<fVelN-iCln8D1LmyBtl5!bus@i8xf(s{jN~f=fZ2oYjV#m^Rxd!
za9+BQb@YAT$9p9tqNuIOb4c?_GiA5m!~-cH`<G^wBH&Pk)G)v*YHTp-KY5uK)b<zq
zQ*}-&65PK4Vbn;5-OCmr?Qx<%jet%sUY$6C+aO~io+lGOWvJpkrX;-X2SzH|A88{q
z#;{QL_fF`{Bus=<CzOC6+VLaG<f|0GT9_w%1AqpyfKg&C;B=#TY9zavp8&W4xUQ14
zb60!eXt&x2EHR1M3zu$(7+ruf01&I7w?#qLvf;bPf%H~uuPHiFuis#Chg^_gJbO$N
z^|lJRbZHs6vMHYE?_RYN!P}DHPRA6R$CgJ=m4gn+`2B>hqw-r9G*KBp$mG6T=L?t%
znOIPY9G}K@efqkpsVd{py0ZH^aYkeJO99D?Xw~itGAvGD5l@_Tv|a9~eP_gv>t`!K
z9<vim9io1BmARS2I0?-oKOatj7K@zu5ihM|hksT1Oxp6W(wt0t#+e1e&4BS|MOSMm
z-s|`Gxkylpg!SlyXw&?BQGvI>DZT#b5=IZmpyYx~oE4i4U7HPkf90y-DIuTxU}2^w
z4Y^##oMygkC*61*GQ_scjYoLZ!a|7&c;O_<HPmm0k1_#F1!+;1tL`?gmXN)ql%Cp6
z4{k8DoPYNOoa4|9Jkq%faQ_9tELNB{-<UM7?rr4}!$9Cpu<pWTo>gF>O*GQ{MzF;o
zFlEk41nuC;^iNlxPZCyqE*_9&AMg)C@P8idS1(??5VRa7RU^<S<+|CzGNg@lB6>ke
zqC8rNPGb2oOkZ2b;+b0Q7j%l`wV*m+)#qW!S;!WTPq3#bYjuoJbh~3vGKwbA&4FF<
zZLEmsgAd(rl@%8+ciM)CKwSf&w~lVHFdYE?iisqR6io)}I%*#?1k<%LM4e=;tGg<z
zyv%g<d%<x8Ie5D~q_|If;0?(@e-89rW`1U1iD_UcKVXP>D>^_-Q6KEt=CqkI4C9_U
zc=D{6@~A$OEm5MU*XMJZ{S=P0E0wFBdLB*-Iu&neTl-XJy1Tr5ZWCQ`duZL9=+A{c
zsRHM!D$MMR08qo()08hb$G^2_x?w@e9;~55>u;kk+6g?F_I9agCemNye8Q_al4c@S
zLW6aJW``Zrv8$WpX<-HJXYaa=v&Nw`F5UL?KiYK^?gB&rZwd}~uhTZM#4ELpLRotJ
zuotbyQPW{Ys1YnnUAp=2*Eoo;R5{982%jC1(Xv!$j1LKT^Nn`Vz(pB))DHBjl;M1H
z?^<FS_OE!261pw=lg~cnayz~z&o;p)sl$8o>wrVtk~<(?gDIOnlp4>MYO$IBnnVk&
zNpQCt+EMcQbMzo<im!2Bev}mMoewKY=Ja2tgX^}-r>#9AC}7{1ZI4#z)8ipB7btBj
z1pCd&5PMGiKLIaw&=O;@e`bll-}C*u1^DMZ=sK_w+e@}EHO2UhT_bVTSiVEsA;M)W
zHWZIy<N^;XPWy8goc3yQcY%_Bc(IgCNuQcp7c3wrE4qEKedX{gKnb3|7_{Y4Hyj#d
z^-884?{#l6SZuRh<GR?fFu>jRx4&yZ+?Kl7kPJ(;N&P5WCO&AKtOwntjdZ7<*^``1
zvc(-hc*epB<~zo34XyrT+M%6qfhuDMK7i3o>A}&kdo5@%<h3sSA-Bgqs`)Iwd0H=*
z{nu72m}p!i@jlpg*~`wbPgS^@W(-s-yine`JRNo6CbvQEm<DobHkv^{8zwM<kBTkC
z`vIOHtb?tay(D8wu9B_jtucZApku+$f6%U%iq4}StvaWi6?$OqC&X3w9OkQaR<LHT
zU?rlH*BOS&KY+cZ*%_>WpZyuft-3eRd!COg!YjLFT-;J->E7{*5@K8=j+W!=GSZh(
zy9`m}#_zmqS$Y(Jz1`!<$L~Z=5x;zTS=H!Nj_IM^NlAY>kLdS7E>#~XVAl<Bilrhy
z6ToHI-``Db_qT6a84TFrDwKiQyx|<3ai|bXwH)@$9Idw<n$RN@b}zn(B-A5c8I>M4
zY)@-P1J+qQcK(NOwFE+R$ZPVk7jD}%*)yRU%TQao8-WoiS)KZH`dGqthrH3}dv`~&
zPv^3hNTS&zW%PSw^31sqvM>DaWI3@BC77%H-#x-a^dO^{`=K^c6PfUPYd{T|Yxhzj
z!71WS?7qi&AAl}bT>(aNK~n$$$Q3-%1QYNMz|K-){k}TmQ+XnYV}ZQp+W`AJO!>Y%
z8@L{%V|G6zwMGD}!OBi3egF&G6&+qe+ZrxVHM;|bL*`<AcqgO-pAj}cb0S+=Y_0T+
z&=RP74WNS!4*nBlAlZQLaN?r4k$xakxOaTv$cK}{qD&(feI^!d=oTKW@GE=jF3;>}
z@!KZA2lXmH!waDRY7h)4=p>|=`||bTj_ctOTefq1S>pRrQ<b`Iw$R(;9D{z&B*t6>
zQlb7+8lH9K!_JcA47F&%3eZ#jz?$x!jtLcyNY#WdK7W=OujJJUu|jSI*3l^Ryy*Zz
zfCV^x0Hjy>=FAEybAd?zq03@Vw#QpcVU&yAAv32v^xcyg6G<mOLVq5NED^$J1`CPH
zm${XO%^T$5uZ|j1aj`w~$T&<or$4S7Z$btw#|&M+AKdZe<jLUT^g-nYU?L70qVsCm
z1<~$uZ*gI%P2XuJe@}{#zl~)9R^yGuImev?PR#T_zS8Xuu0@VR*@jZYxiF$zLm-dD
zx>By~@Z^<Nw&p`=D>x*;Qwq}qQQ|lwji>S573FnZXIfy;3n9Y&J*Kx(uNKZEDVm4>
z6n!nS8nyY^>DPw%v=vq7ar}Xr`)Xb6;%Kcx9fR0w9Jz+V{9jwiNPvr^$FG{x<>h5d
zBz|@vLc&c{O)-@&L+Otw&^ZE3-8*oNLUhi@)0Or^9gPSj!ktcdU#yS|8Poj5bJboW
zF$}EX(Brb^&|xH2$$jzhzPImyP5gs@qf@>bW9EwE{0SI>k=c`^cp{9jci62*<EV*9
zsNA$mxvi0pA1I-9=A>kqyi(lW5TsZ$0UklDlwde3_iZ7Ubb?9Coe;lu%<+oDw*$8n
zm~TcrXa@^qs&!R%X)26(h{b0GVFk<nDW2Dn8O5@o5lsqOgk>JAQ3yv~>KT=I>z@}3
zx5^)Oq(66DP`+%Tc3ylT=I@uFXwDY##vq<B_|9kSyX81X*4N~{%2vP+z(-hz1jPu7
zMSq!Pim?y8<)A`e(KWiB7h=_74Al})X6|U>#&O1=jY}bj#RT;hS<eqAf^!r}B>@C{
zpNqt+X>$4SKLqgVCa%8}DB(oM=7Jb7&L`0z@?Mlpi&Q(-Z4wuqSzDaBR)QQck;YQK
zK%)auetPUn3<NPRUIRCY2D~GNx*IAHgFoD7FV5;itA|iDLcQo{n%PA`Cx>6f=d&4^
zUobnpyKO98a(&m32KIQ6aah|`1WN_AOxGAN^@lZ<4-j{rx-M}^E!)taw6~kbn;iq+
zUACB?boF>`S?btc`fVXqC!S#UWFCV<Zo)NWZs|W6ysCR~%(|&PCeR}%m#as57s54@
z0pvJ(x$NFK6scX_QzxTvU+0VeuZ&sLT|{Jc=ZivBUfxEkLRIy`-plDRN8vs2%6*;q
z*w1xT1<@6=j1SNfVh`Oy;}xE!xs)C)_j*~=6Qlv1XFPm4a=5sgZ5?GJxY5CN;I?U_
zj#-mzoJ{*END)_ePx8$*?M)FRe?LSLMb&cX^2UCF_cAxOEJ5386gfy7d{Sic0_+x|
zcLH9$^Ol9zF-50Alh<HdlfA)-Wcd`h&G@l_SMhCJ1{^3?`AsurwkUZ-T0v*>xn|^e
z48NXV(QJ6`*11kkbWpv1VWX^io?;brxPw(10&`75=_+(;|EB8Xdxm$K_(mZ&2yvTp
z=aQR>&W{8llvO;(+EOK3%ICKrgT7#*N%ji#;lUMdJWdJ|NOg&TKO9xsLA3Q36_@*I
z*8Sba6nkGGmzIFAZRUc_y+oy@Xxh!P5Zu1kc%^K8x1~O_2&gtn(_y}WJ%d7PkKgwo
zTXut5l-+xV>(O?b#oWEw%i(*=s>GfznBS_EZ<|*R_GpjTj{M_$bktIthvzZQ1qe;v
zh9}EY2E|XaOP=z*RCZLrqr1F~p4974`puN`(HY&eE%JMSdu#t~U#-k1zMBoRN#!-b
z9s6jVyQRI}US#k>ao;(f7%fVb4nIaTAvWO4UvLkn{-ZQ;v<yQ{_EY8OYhv-MOehaM
z$HbYT(_^sI<JRZ-7J1@W*1vBcqxh~qx>Ws0MH_W#9lPfLU^bCB>ssW_%%^!xiVhh9
zkBt5jiB*Um>k=Pwd%rhWq)qBSUviq43e5YGjMWh-Vd!pNP3@A4tpVx=FTE|!a8O&)
z_{3g=a;?}|IrWSsJukx#EDhr^p`|wPpO1RMU=iEK{w+C+uHoR-a+%O>P`L!MShzpE
z#|b^(a6z)l8&z(W@wi)SPF19hmxY^j^Di(X9JZTRO>2HM-8d8n2$GAY7x*9R^#caf
zFJI^UTIpK;`uLsrHQXHUbL)pb5gQ-t%wKmeUaX_G)GcPB8Z@rupAblflHr$X>CDB=
zYK#-otA9OvZ-P+nay=CK;(9f2#38nV<vT=)j9EUQUdP*#y>oE1ARwBS36r4WAzch@
zA_vhQ@X@@v5b935$(q7opn)#Jv@}B%lND;Zi5F*17M?>s>{V3O@@(KZ&bhEPP!H%G
zIYDI!kV>d^4=6%OuH02iijyY3fZgZHaN*^erwTttZciR2f>QczOcu-ASBq7_y7dHZ
z#!B2(L#I#raPUwYg`_yPGA(ISHW3lhZ8l%x5woNL9cRm9_Zy-z_uc1{fAzKJm$^WH
zMbV+D1fN?+)*tWFTua`|nd&0QHf4$G?D9??JGzq@Q7Y+i$#vg$aOV>AyHIb?5n&3$
zx&4pDcl%CW{7!(Rnp;-A!5ZwDov{UDZtKjAf$e!ir#NLgZ#e>Yo6q}FkU|U-2ZK{7
z?!YhEmPu)s|1-5fer`u*@PAi8&2){5iqh?vkQa5y0k&X;j$(Da8Gd^YBog*-(iHv3
zva-JUju|sSd<d(&X9&etb<njpjd~BI>i}brb~1`L!&Sa*PzM<{4b^dx;&<Bfqf4#%
zJ{}vq8@J~)3vw)XN<J4YVbP<{%UB49KJ4bSH-n=8#Gf#)Od8f=#sB2<5zl((omYr-
zX!6$`U`jbWCRO(vXmg<cXxJl`E@^@)Mma;7;$&LNu7Ap@*lsxo<HMMQ+@aYzrVR>r
zfc-Dj->k}Zz);#KnZ2ws53AvCsqDvFU&p7;wM#|B==pnz`A@24#FXvW-I9&Jy-{}*
zsoR}r`+0qxH|_k6sru(*<LV3S=Oe9G`b&D-*McT0Cq4OKdr%6D9z^dly`iBtK@G!)
zxo>|82lRjr2QO359f)Xwxs^WpT?@TdS>M!$L-cmD4`&5XTb;U=LmB2RK(}ym@!wc8
z=-NfJC0;1oz=u?^?cABpniis7yKMAh2@k_lB1|mI_WIrFb3;qUlNH+4tc8Ljk<&->
z+7K^e$SDnO>PYNP^^7=9L{qHBvGbop6-OK`dT|GSZ@KM&5ifE)62Dw~A*K{MdR{=E
zS?cfs_Cnw9(4)qw&>tVLKF_BT8*7+SfLp;`nb&e(CT+74Kt{CYn5BFqlN8GIuvF2g
z9M*r*PLZG?Rzc@gDRIOUXIJYUqgT6icBsz^{M1cqL~0=Li$>zVvP}cZry+CyIeM%$
zhiF{8#a#ytoR~7%;6~T)#TGC+a1gu>KMHhRO{4x~{AQ{>u$*~AjFP*Ss*<MkJU`^J
zm2$o*^}W)QmnM951Eb59US^4liHe^)*x*$V-qLR3**sn-c1^%==ypES!TmJ%y;nV2
zCdgMVpVd3k$(z>eWY{rL1$px`h90NZ^Uh&(V>lvdv+zLo*53<%*H5@z&j5@E%m|l<
zj(19|bI<vSPMVK^ghcUv{-1LG9$4fI@Q&CjNJi6;l=GK6hC%iqALYWf;<BwTN0+%V
z?A7nuc2+ao@3r5v(<gPcUX=fR_+4uIIY!5L?29WBvwJ-9c6CSzi{s|^pP*iqfgT(f
zK<R5x0zOgkIsbHv<;_;SS@dBlQQcU}FMY)83%}L$a3+^F`oHUmIBldHL)badsr9>=
zAGXV3o8S?Y-bx=gql4N}%v_x!v#3+A*)=_%0#7v(a=Ncg@7@3*#Ya6fZb9lo><K8P
z8#0#0Tc2J$!*Ecd!g^Wa%59(Fp7E@V(&bn3a{GDA4I#_ky)HjvEgR>)smaFyi$6?8
zUvI%fhFdBE^oMlhB@5yD&f&h1<brEc96F6b)zz_&mtAKyyto&?B~MCB>p{4cbDbQ_
z+>1bnMdv=Wx9569Wx}4Z6lT#gr`CFVSiY10wSDXFf}zBA0BUdHZvZ;mi716x;uaPG
zj>OczR^_aP1|oUGcK<qiChPMaAN|t=dP8-isl_x1Xh6Alb}Q?3acDdje?ML6Br~&f
zB7{~#Sy7&owM0!$@#frxsq^vD<)X#n*ARa-AM90>gz#_JFgcLMD)vBK%iL*&udM;G
zQT%(e3%(~ccslqqPzNG}dNf#Hx$Ru+$cKLB&Un{!;X5sftz>?Cto-asm80ao*k2B2
zi+LQa^$_M+d<fuzxkA8eO`))QV;n^gFuGq29|Lqf9O2JC$9oB~rfR-kfRGoXFD};h
zpL{hV%gouNB8HWa-s;(cY-c)|1&i!bcs>>`RA^D>TrfmU_H!oOZR0xMra$;fq5J~w
zTtnp~uJCGzr{4@RnwBa(;8>s-Z%p!Fasq^5Z;fyXteF{vILFax*eGc8sz=qbV$yW9
z%x#yE>A3z1(23uli<RIk(BvC52ezk$_VatG*SJI@H*N^WDssR@A;;{xhiu`IOL7<a
z5qk@-eqhC06H0q(*nkk@N|(gd$_eHJm3}U?x-=J~hjH+EWJX@f3{>dp4_EvLo9bw>
z$#@x1gEAr$>UPZjhx3ri1#~jHBBAkdrZ#_qi)dVR9yI6XQVKN2ES#WfcCG};y9He6
z!WBMhf^;x{fY)1ebdDD+TLjiy`W|Zf-PJ#&3PrcL_Sk+khalXZe0go>%5+@&$n*w>
zlemlSY;q8P2d%uWPd=*5&g|e0ZSU;i;z)cSzRCD^9?ZXIB)Z6CWLBUuXd8(hV3U@X
z)>Z6|Kp{#WC~n4j<t};@a><#u9}p0mPow0p8K8#5S~hIzBu~8f@Y|051Em6MA@v7T
zrPh}s6j&1Nnl$4yor~8WfT6e#H&~c`X@cJEga;$<loMFx#X_H-v}bN%tq3X8wA2;z
zp0}j$qsmYInT6^jia*}^=zApLk?OzZV9F#plS^f}l2$l%9F}ie0B=O2)I~l&n#e*U
zXX#~^B4^hCTL7@?HYi$EK03gJUkp6Kc#)83e^#1=&U%Ci`UX7Bbr){%Fr0TLpt&7q
zzqHIt>^S4~UeYzE{Z2<S(YO-ajpPcfADC+qlmfMTZm?$8jU!CL{=O`{QZE{nhxCJi
zt|SrIiW%PVnA$492p9|?YQ(<F#;jete>?h7o|^J*@oid>Xx2HV=LdA7KU8%X`3Y|R
zO5dPYbiZ}Jx<mtyqM3!KdfvF$ZuS)9bK6U&@x{yJ=Jq08C)tWdpLM6iT8yVQP&*k=
z<Zxd*DVgA)C9-+0T&Kx^=OZEt$8D#@zXnEXor>@26JrG~YsafssoYWz(4MECnT_=6
z@?)S*TgNx`xvCvXMI?j$9i(6(-MYXUl;7@^zh>Oqw}*f2fjvW;98+>)0868HDq+?O
z{U5L0M-isH<kEyh=P|MR@{f&1w%4D3EA5TU?RnYw-qCsvKg7tq@G;pvA8%emN0+Af
zS1m!W@^-D&E;Sg4GXeek*HY^G*1&yJvFjUaD5A=!&Eiwz<9n(UmG;^~keu!(U7mH6
zq2J2U^*_#9s)T;jVpLtEjt7#&Wel5W;dpEaZfdSmXPArT1?k2&+g@|H82URpfiU&W
zXVmo%qJ8{auPcbOmey9R#CI2AhIFl&W3IE+8Go|u@BQ9$z%rphod$z;A(=DkciyMt
zK>66u_MdPDH|S8R6z4i`xggPCa~JnTnieGSmDwV?EU!VlQ|!-K9AUL-0bKDd)_UHK
z#>`kpuxK%<J8JKe@3_n@_>K~$-FW>@mTe%f_*I89ZOYExSp3q^QZ|ld%eLq$Px6fF
z-ooK|%%IO>1lVMt<fT?s<*Hsj`VZnNwU0aDIIW|n^U_GPn-+^Eyh?O+a|8E;r{4o>
zd&+<DF32+!g$k$-n6*fnZm{9g`Qnm`q741R`{5Df+B7DifM}9@aH$BsK=ctKSzZg;
zn3dXywx&KDmWoeFe^s&u+jSo4RO{sNjVeiqQI={{@;O_Lg(8CCyvI*jOS&+V>6T>n
zT*{5^Ag?V4^cqWfsg7@cnqCW38X6G{fGx_lARbxfVEX5%*@kjM5xWi9^m5}=%`^cc
z4@-*MLo@b2C$_9U=$Dh(b&sWNR0w}sEA<n8x_^hh_se<|2CD!a0=!y{XRZ?2^{&F4
zXQor^bMX8<cnv|CylE?Z<-6V`M~+?B;ZhBnT|*$T8)jrPi%+|@#`b@^*{)zaJ3B#T
zn*7$o?V*I7*D+33F734`EX`z2X<wo$?U$_b3ZwB$b$~C3#M7}VnG2m2qk^kG*ckk|
z?*^3flw~-r!y+a{O9u4ec8yL;L+<XMo@!%7GA%Mu7Yvho(QHBjabg9mA~f4`D8?~n
zQ4)do7s|>zv*LUEVgn3)CgPtp!|<h%`7Mu|QXt=oFx0%H1oGP`zig}o2j}@Q?>1S2
z1syybZKnb0)*<5!+P8nd`qye1%k#Vdi=d-^%5qtH$kiTWJWXtJSwKy$A11alKv8cL
zs>^*HUBP9Q1a2_%IiT2a-}3ycw&j0}+MU<L{^3(xMW<C+`(qZFTxOyfhc$sH*bY<)
zXjd{V1ueNb0p)nEj03(710w@?f=p2=A<{(_eaf)fNNEWE-by|aWwncyiX+uC1v<hT
z!av__7kCZKj&H-`Pb9D2IIQ?~c|(Ugf5dRm(N^h6DF;3JT;^8-rMNOQ!&f3TnG!#8
zkU04KNqXdkYH33ib)agvS2|~Isf-w8uD_}op#=rdoapURbvHh8o<4uGQ_B1%wZ@~$
z@rKxoTzi_3c5L?^ig|PqIw;9w4#WdClm3}JPlrBV#7uni)l13gO=ir+OV;R_0@_Xx
z`iZRT_oO2+MIZs#TY^m&#A1}Bd-o@D$=IvUF30fXaYcE1)iKoKPW1RxUn-D;>wGpw
zwM(nk<CFh-IzG-VzRt$8f7g?L-uun(NuBi@Y~ps}6t<)7ZEfS;KD08T3f6oda1{0=
z!(0|n2zp=c)do*CA|WP8@&s3O(j5#~DwJ{pg4<ONVqdPDZ&@(Q#skCRzbbFyl*!?l
zU?ja!lN-rY?56!wZ`l~&QD<P?`_1)yqFy!>OPMqp1b!VPv?)O)I=rW^MqwLPH1NUI
z8==8s$-ZZ#3g7zl8I7NjPyd`Cd{8AR#7d?1_rm$D58bz`4_r<G095=8RuqXJ8)faU
zT>ayltzT;Uuin9B*MibISH=wcF5O-We`8>2k@Ki3SozZ`z81d}-a1sRmErM>r1Xja
zl=q<1yM1|@c{iGeC~<EGw7{nUVoGuZVywu~1SjO%Yq_^H!K1xJdbd}Fw@Li_ST<<0
zORjC61FQ-%Gbl8~PH)0kj#(64KfQ9|>!vfEFJSi?AcLBkzOvL@CqEzAJ;eH0^z8aW
zxW1l|`tCa@KqS*yg=2R4ILjA=R(_G{=y`2deIOF@^_lXEqsW_&+XRXgMYQ%(VF>?b
zOydi~5RUClz*)EF32y4Q)33#qJhU5@!Df(HK^ZymT`O@HdeEhZ?B?hBy?C{Tl{o7<
z;|08~ft}WKf5tUdCDRbBupvy;dt1gfbZ8ArU`My_EyZGn`)UK=0wmR<eCs^v>vl_1
zo!nC3vwg2-%vP@FbM-50d1)sgl&<NtBlDuVV`}egtC!IK{Z$zJ>HedCCnowsi0Z>J
z(j@Q?<3Buwd$yuhyyd?@!CC<1XrVgAwD%sPtE_Bz9Xvt2lKWwmC28m7G`^L_rxLe@
z{g7MUt~fL&BKZ%6O@1Ve4l`Yd2_A9HWPA(&R%nQmb|nedP23_R-py~x)x}{cAlQLv
z$em1b6rj%1j|nM+g@<c_NP5qyHZr|#)?pGs1L_8)3)!XwoTz}HTFU?#e`Qy9IlT4Q
zzs0oPe#)#UGAExuSUlb&KYKGsi@81IdfsQZ^eamfWYf?wb%6#0$W8Mibb3_@D8Xi!
z^9IkWsgpLwfgHll=PiS5qxAwJDbu6xv0Hg)gA}SzIJDlK>#v+OpEMoR`g*;3+O`>+
zPmmTy>t}`i5(;*uZjOp`mbvY6zjx=|f(&O=zrLI8DJnTkO;&Ryv^loBgarQ{hDZY|
z5N84;QCUs{lDS62v1SCKN8pAV>Gw}GmFEoTKPTfv57^s%4&UL1(?TwamhomW5-&Y|
z`qp^GH)7Nkp3`!ep9tqteP$~!nU0ad1?+Fv^=@EGsa{pfFe9>+rSyN~>7gkSJMqYa
z-H#<ROEP&0%}63}VCgP)-=w!8ZHO3m9)vB7WqgDp@2FzHb$Z6&1jLXM?JRsQ9Jq_R
z<etNq&W|Kp!*8M+ZUqViojxdY@YGFA*aFO3cDm~W38lT=5N-te#+^)ZALCz28TAr#
z`b}#{XGUhYkMG}0o=@8&o7)_#F8}**_ML2ztKkrQ>R;AA1Y5U1=4kdtkV>B`Lv%+h
z9Ig_a9J&AJh{`%dE`Yqwcaa_>#eFV0r$y5hz}7WmB+kA_6QC|GCCc|z>)WPHl${?|
zRDkcrotk1(3GNDM<6?cgR0~*)=`LS97>2ilWoW7$t2O_O-@$@|GJsX3Xp*-62P~XG
z)e0Q@N=3QAc7$Wrm~QYa196_o-FyV=)Sk3Ktn%l*!yLgy^zE-!6*M@k;%pwp8|}>2
zq}@!_b!<Jf%0&jb)K|{~hvM<3r}i$0s|)_VA`~y-o%v`9phGC(Q#VFMCe-_HMMFdh
zz7P!*OeOTuPCuR=C4Dt6a;(tGpdwe<iVh)>qC!-cSEv1r%tds#Qw*wS81K=hc*)_c
zN2tYbVvV?ip<n|xgfYAdW*JH?-Jm)FgpoU`-;(l|os&SopNwH|zrkR^ID5ff&%E)%
z%p7clh}e$Vlt!PK)IZwV*SvdPB52Ldt43peTfY;hrVE+7!<^RzGP{i6Q*V6td<SEl
z7x7{lQ0JzvDtCpWT~#O8Y}kEvzE4|lTK7FB8EI7Z!KN0mwjsIQut0d1&~taw)kK_*
zgg4C$C=-0Z@>7iN)?W$rV<hh;dl(mb<uf!7ruyJQ$JKD+H+##@lqo8$H@$f+9kV1v
zdhN_mkMMy-ZXXoDu*{P}>fz4KYXwKHUq(N`-{_9fvg+~7)V^QwyIOHQ?YK40MwR*q
z3x(;DnXI~^|M{*uxI*iO@(+Aq5*K_)JFpdt`&j_(U-=DLHadEG8Rll&uTR6ib1ooe
zOP}Ja{xXy{59}$J*Qb><b<nvI73&flx;6WV#W`UW)HZI4u5zIis!AF_5`YxshyB_3
zW}!~ttIBgHexymEiru)x5zBh9>Q1_1q`pYlIia*(<>P!4FWltJN`?~w@_y1~eb?m)
zvWA&@U6DKQ2bkW+NNHUt%V+lX9Nb>Wu#X9?Q@o8L2c@vX`kLDC<=4j}IGo~;V~mG%
zW`Pz)%o-pe3=EiPScC}`2hl{QZT8aOUy}yuB@;9D5nOE|pcZ|UTQMU>o-_USg6ADB
z(Q|t0aXV^lYF*#JV$$y6i0j@jSGHa0N0FOY+eAEHMB)4aq@$Y0%&3k6_mN6E?>k=|
zw!mYm0!d!@8OZ2!R<f3>MwaTQLwRy%T&LBghXU!=Kj|6^GQ6SDRrPPFsGT73`T^?&
zy)*Q9PJ;%*hY7$KYa*LSjMZPF?utP<IQ_oThN3=WAkg9~c{p-6vNaS>@d)499ihU-
z5!U(yfQ_f;X+kPd2ID7KU76#Zt7lU>OO?m#88|!i3^Tg*=Jf>;;02eb0Q^m>9LFoP
zPtCQ%IKl26y4w9Ty}I|zdw~IkFf)KSxK7&<-8-2$n7bl^douB;@Z99Eo(%=BzQ6Hw
zyuvm%B?}slN3ZMXv@c>mKln^jidTM$iY6VV09bkJ=cIMhTKZg7YtrR^!M{;9LNY|`
zlXs#Nl%r|l!nX!gkRc}ehiohIjf=Uv;(DWL1sMu}57ka5=Vh_U+lv=<-*TW$6Gp<O
z$pI}FuMm!1hB0&Mn-Ibb8jZYFUPr<V%N3$$w?iTA4U~=d?#%tkU}Jv;nb$!saeD;T
zB$H^dnEN9{v0NzKJn+|uIMLHsRmtmkvp@~9ey(F?u#`DG=ldjE8QsZOnb@bVXR^6z
zWF;H`c34<62IB<%#y1Rkv+l_cd$e!i<s@_BFA~+4Rq5I){A=tgVb4sMCxe<IsZL*N
z$`7E<r8mtR1Z%X>(>`svG_ADRCkbD%(Wqe0jZ-n7`rn|sPHT6RY|TbJox%{G6pRi^
z5(kB_f-EA3cgBvSY|!~DVzHum9%)!;2~ASz6jFK4=hxYwO!y}l+fs1gr<mtky`=B^
zu-_o%r&;i%!7VIMMy_8HK`=VMP;{g;fx5gZ6tG6*5>@;H!_uL{z$czp*^&E7c1t@@
zj{Fd>fvV1fDIBm0^BA#B1>z_4jLp}r><S#mxE;gW8)0hh@39f6r*6`Jz^=Q`r9E;$
z7Pi&ji@teRMCqVEOs;agMZKZF{Hd5~Tld*LN+}2xl^Hm(9kcZ_4xPY7b<^AJ_7@;T
zhl%z*>OLIi<x!+?C1f%1elozC9tPrE%AMeobDA5fp1NFReIIj@03DaoscPSJl|*B>
zn)I`;sO0VBop?^uK3dI#X{9wO^uN@J=uFlKq%hSI0qP>AK_U)*t=8&<h&cY~wz_x7
z+2htkvWc!_Th%J!ZFb7XhIBZ3@*y!BXpEjDCqg3ziA2=nRie_Cmjd0GXJ++dR`jKc
zbnFPm27HD9^N>u)0lpp23Os{3d`xtKuJTm<Wj&DCf25ygP_9by@yAd?iJ#`(aI}M+
zvW@I?e*?_B8;!;vfSRedJ+3BxfDtZAk1)@0>ya436<cRPeNMN?FqK(<5jx!z>I;V7
zL1qo^-lKfcvh)aY>cQ`}mGqEhkSFbl&9qQfTT4t~O=1tK(#B4N%xnR@?dgVwb{2jz
zH8TGl_Tg<&vMbTszgFEx`Xa3r#Jf2ei-QiQ{-PLgl#O#E|M*1cxpakt@_8<N=?TOo
zY2&EeWwGr2RR!M}%7H)oQf|dFlgBA=G8S)5vIN-bHKau`&6MUgsYLg^VyGI5K!J_r
zB<_v3jp!H4Pl<$mZ+h4ex-hwUD!;x~A(H~AG$&WwA<`tg9M9blg<hrbqaCxt;QB^J
z(lb#<TsQbG{&2Y42iK^?rspDoRrQO|ZHP#JE6OP9BxMYag~qy*-|jwD^QG|myBQZ2
zQsdJhoa}~^Ke3r#7!CHKS3>0esLIRSK=SK{J}-7JvCv2yT#v-Pl%+N1`hJ@U#s2-q
z@Ci!j;<AQve3|Aup5z?MZTuJK+qZ9Vj>}C$s8=i8no@sYo(IwMcTq7#Yp^t=yGia}
z=`7tk6D0{`96{y(!Qx5w_ysGRoSClJlp)nCe$yav1YFtExAh1WK^kl<;@lW3T`oWT
zs%r3epB3$M`OZR1zGko%arn=f6G_n{xi2?+KI{Z0L`ZY9H{2NZ%qM3S+WuyuWWii$
z%qla=b^O($yimaB%$RenpF#@B=2e<v^Dr6doE>c{@mC0j1ugk*3L0w|e=bWbdvBJa
zvDhtDn2mT_>8h&WU!D?m6{iVvrsKP@8|zrX+MA}%jJm;2j6Z(Bd(o$1q&5{BD0*P^
zd4S3Vrj{AHMBnse9T|Y7Vu?`5P8W6hO(Y0>q4)9qP1nl&y;+xz3N1@6YjlYb;p^ie
z?wx|CwfL<szP0lRsaeiF4_SwY_kG}DoTo^17U0>WGN=m+RpW|V*0a%@sPDeo=QSd7
zrr{2!DIXw|i!+G>QI-@$xSe!M$l?KN=!|keihEfle1*CAUW1T~Is&$%OaP5;*<TF8
zGkt0pM`nV`>mhrP<X+{UXi=``n%iEn>Gaz~+v2%U;hvqOkPzD`D}>9I^-=oF1K=El
zyDLNp%R9#O<zPL{t3mMzEX;`9BRGm^dY{%rPGOj>Aiy{ta~@Ok7WuVKnS?f?3(JHw
zgb|-A_>KcdBW5x~%Ua|<BT;WN2j<g<rz!rf9Rh#P`OYSOx*kQhgiyzq)i-b-m!b}P
z(vx8FZySW!yS%B)_AbiwIcag-Nzz*~^?W>}sPjmnQ&R=>lYmCOfSm_Ki@CYh|I0>+
zqbO5|ib*ioU@K!m5Q_MF88fn2(E3tgCS|XimTY|#!5ZrL?6CI)M>G-idj4tD*~vXY
zy-{AxEc*)+xoYqhx=j#nNR;EbZn$py^b`21)Jc%tKu0ami!`bGM(N#yh5MlX(qouM
zo$10D-$`7%?kiHa#e4fx{(_G7n6j@1jKN`@j0cj8ZA-;AzbzHS8oC>b$4KxOn%*9`
zug#FW%@ZhBHRln%#-|?j+a*}|Ix4Q_+=zAx!4h@&v(YS<A5YU#1&DU|Q7^Yfe9WGk
z*BR<-9W5&6bfI$peqcdk+NWBpp_dYk@pI0?Z}@gP!Np+z0AZ3pSGfO*M@m*orcL8H
z+sHr5X%_`?IeIOp^?Uf>MP|^Smi#YNVOq0M&5Jo0zOXvm!GiOK&kM+&M*juD^PYPq
zG|7Wi)giEFl;OR!F8)Ne;pk1EJH}GJLCM!;NWEgWq`xfM(*DG#<KV_@LtwR>w_l(c
zz<FH^K!Rz_GcQJ<>bD}laOA#P=aGO;Pvponu*@r++#aL=4x%(u*CI~R0y9X%Uf@R8
zb-Ekln!;fge}1@B3t_g|;bUx`q43!GWbOo%127@}T$Edr^>%&XH+@vIjdr<jkTGd%
ze1caFgf%lqY7^$cq_CTAwf+VgaD5M*1mUeW&+KZ1L}F5c&1pgB8o7}m>EOmBS`Uys
zhUPEhFIfK&?oK9jJno+`BsR@#9!V-mi?MMHgXRqJaz7+*;q$dglBv5Ca@z>w(hvB5
zZc80KjUq@ehpnSyW6l_(SfehLq<4pX!Qa9>A||9=Z#;9ASw_*$HH~H0q5DcEFJq7j
zN#I8UhQ3B>2LA<4T*Q=?_c?>6#zi(E$WZO9C_Yo;F-g@V;er5BEU%eGFhLxh!fRFr
zV2%9#UMh_>X-%P3nIb$Gpe}iObyM)?7Esz#oOQGt{2fK~Z}l@VkK$8QnwJR__C7e7
z9CG9H60|w~;TIq%Es!NPc+c_7>DzNT6heTz!$|;z)sbr00uaznXrqs`U?wj=i)Ej4
zhF5ZtiNiU}VzIVRGhfIG5j=2Li!=)-c5-7+D#d-#$ike_)t7~6<RX6Ka9l-AIQ`Zl
z;J^j}hpmFi*)3X#8#S44^#^(*2nx3cf#iB8Q6aV;_x*U_rT@4rv^OPCSAjhtB|X8!
z{K|WgQDi5qa-_K_lT(Q_eZ)N5uJe;b{Y(OwqW6gE(5U9r-XyN?sAiQwC3g+Uo{@k8
zI0L4<gqj)tyEg&9bDF00_<2bfh(a0cEl+-8a8z<}Lxo=;T%t03L1$qun-@3VeQQM7
zNe_Glv<8sI_dEVOS2?_mo*&Y<9PAdt*h7?R9eTLtmSb-n)AD%?)?+VKT{5x>PDwRT
z+jv}f@<!mdkS8*oohWp-A5ln#={T*?lP$4~CiM{Z>x>&nK&<ugd9C?Cy4uN~kq+Fm
z-KgU_%QZ%EZ|yszMy7va@<yn<&L0;!di&48rP-N+BW&-$n)Cc$Hz;&;+G@xk5(F`H
zaWHdiJ;K?1;F2b2nyFa*P`AdhcQb^Q^WjN$)j;Ufmy@D+_u|tErMHRVD$BaQM0-{9
zvceY^A-_d3URH$g3rTn8Dr&1BbE^$IRP;g7MsNBl2PA!8giM)q60KCp(D6KH1LZq*
zkz)hGZF?F;Fp|od@gq4^24lH{SLt{!+6gP}pY_b?1#ELv*ye7yRYim7-q6OldBPiH
z))4I`RswhUnI7)5est2dq(NZgxp5Ba=u)hN7=%^hH6cF`BbNRvM7)PIVLHn9I4)F8
zR+~W%jW?e(NZ-#rx}0C3yY9Pb`O;&c#BPu<>V#@H5yw>#eI^t!$`Ilj<+F4VTFGS2
zwyR)Jxb>sGN8S45+s;BFOy#gg;%`;uI0faRcZRb)iCqzTx8tfE!^Aov!9c?n9_=b0
z!6#fqTHWMCT9Mx2&iif(PwKqFpm*_mk&_du$f8h-;b9OEo6gx<Ok$3;j{aIO#|>H0
zsf*5~nU=!aR4wJ$kt(T8x~8?Qg}LI7-W+e>c2qd!fB?}omlfa&hkp0XK}t~OOUYAJ
z;Q`abOYnv-vvuJkrL8QX+{^idaarVE^gG}jCV)C_9qxjTh4L54??hox&Y-Q)SuodD
zA;!u!kn_%dYcN46$3CH&ozbYq%CxWww_WwpvzsW>Dc|12)g6lb_P)^RW8nQ^Hm2pk
zF8-r<$T?rt2U^m9U%US+mI?qUZVklHm)5Dy=`a{QgrT=OkB4a0@}v-t6ZM=!DoA&>
z%zEv-X5D&yr*(E?YiVqd`GzZ65;P1tI&_i0LvGpeyA4F%g1L<_-DjlTP2Ss8Q8MHJ
zTk+oYvQbkIYy>3#GOA8PC-iA8>^j+q$F<4CB8pM>;*xTb^VS62(3o0#8;uAg<9R87
zsXgiE1otZAv1;UPN_GZb<f^_{?qJ5TCV9Z&;o!7eE-yJgXNY=y+w%9-NVzrd_X-zg
zDY()d>?(~rTIMGgG2<s&u`JOM5eMD}En}pYOIlmyI}?rTz%#SfT`O8cB-vcM0}VpN
zl!g+LdA>p~O-^U{4d$BjbmTYv=~KIBanS%5rxB`dAs}^mq0(Gn9yUj0@EtYlu7J%u
zlpNZ`K}yO0A)Rkmcm$V^js9m9ru2ScO(1RgxnP!c6xX+H*+8vLE!xjC1T6*1EKFDV
z;=TRp&xBRU%>giAu%>@jXaRKm!k+(yk)AGpu(&#MmCRrE00YT`)7Ib)0o$WFJR59t
z)nV;n2+4I~F}5=)@{>gkqfn3>5G>dGydo)X=h=^7%*eX?m?H6tho6Kz4f424uK=Kh
zO#eFI5+l<d>qb{lJX_Bov`+#}7-iDVZJhKaLVr{#JDBtikXdv)>T2NBH3G9x%)`hY
zi!^20qvgaNm%e0aF{$lZ)%3qN|5<O{eBrPl9#Vh)NySr3E8rjb%l|xFL}&OMNEP?&
zw6kI>;&ocX^)Cki@kpl3vKw?|e}kM+PiY#z)bQzjoC;K0%N(kvG}|x}7^{{kK4QEf
z)(QC5i{BZ~ewJUo;M_iR6;yyRfYlNY5c5Lf{ifqYYx4<dqi$4Ae@>G4s~SsoWJ}iQ
z=CkVAEJH=FU<I^S-T(}K)a*j2gv;pe$Af<VZ%7wWJs=eJRcgZ<MX~1lsguxv4k@Z4
zO<-%B7z4Ikn9&!<%O=y3KW4kd=?@@Ywsj^lX_p^0Mm=c>I4^}MNEqtg5l>_D55Z4<
zlX~WIm9=a?UuadvXS0ykAQL}aJztmr42wwI^o%*j6^Fye{(L89oPPI|hzSoG6J(!`
z#iSrJM{8gtq~hh;btB%NX`xP&)UH;y5k6J_dIT{pG~aT2q}CeZ;`tI7=@PZ4y$T+i
z?ksUwS-f$9@lyWkW*%JzTNiDT6NC+dO12wvAPFb|Jne0nfB<a`%SRIL&a-@nut>mS
zQ&j_P+sV~Lpv#P}kNKs|9n!l!0bE3ZSi+A(RPQdrvSWOcJwArtsO6kV0Wi)@QP18i
zm{>JhW(P+Z;zt-FBZ;mDHUrTa*d|`Ce_$PdmF4VoN9E{|^Y>@p65{*Oi)nv%ANnf-
z>FlHBlO_c0Z=SMmg-j?P$T2)q8k#zq2~zz_4`O(FNa~8a9XFMcHPIy)|EJ?s)iu@n
z2Ox6m@i1V8j-#XdZ{~cq^hU&2TGxHpo48^xy`LDudVZ~OEPT`%y-R(oac@~f=PTWl
z3CC#ITLYR@gFs60D5M8gBKbI6klbc9uwh52W~*rv;=OVkOT>z4cgqT)MIinUfy)@x
zZEwF+6XOlcq~PO?(F~AgOy#qMJU^6*;gJ#d?HV$@8d1etnR5M95`yP6pbqKe^YM+s
z?_-#wL(iyv!Ee1|qN9WvKh-AGT$Ki26hGU0=K9LjH&l{iM_VDeojy$9N-B_g_gdVQ
zL-L~XJQqp%O<oEgBAWIUxdMVPdjs^VKkp&7T!##@#b`9Q^-<sKM3p0(l-32?kF+J}
z5+#Fsvh%s><4*PAkgoX^auk~zH2X$c+d+jBflPDqX{phPks>~Nysz<QyeK+YaX*aG
zD3(eGC@p$l;`sIeOCC(TW~Y(sYv+Kyig1$L7u?_sIWU86FNR-rG^6Qs?R97Q+#it-
z`i{M?awd4~S3Oc(&o}O84SCoNQN5){?q20gHy1%>UiK)JoX<}6r<L5=Rct4tO|mLc
z<1-w6(j`3C(c-qn8hG|7W0$|vnP7^&M@^5j{*Hw-N1iwc{X76LQPI#!s3tqX`Xdy@
zz@F>HNtXYl_1ig>>cMPG&NS3#CFR+&JyR+E(Zpub{!K3(prf1e{YH}epyQOv<`E^_
zS0lD1KA&!LJgZVr$110VWcC6h{{M^7u+4O#3Nk3THuf9#v*B(L#S<PAKx_aXf(#k<
z?j3V2<+Sa*e4ZMrPN_tMF!mdhf7K`r<CGR<%JXZ-Ueg4kue?q`#iPI?isEl=SLbJ}
zalGrhlspw=&g4%b@|W|~8tm9FrrtAAd@N<!TRV%}Tx{4HW(<AK`;BqBsm|b|ZyMom
z^i>0m5*nMpeZ9<3hiUBZhQ0^2Ti8EQcZt%|N9T$%+{GM&6%TN#v1*F^oON`*0dTZ!
zS9zmlMyQ&lWA?B2LvkLg6`A`#>BkjpxeOV&guXr{yP8ZL+JsM_bUgmT+;JcI{Py;n
ztBA<okv$oB^0&0glXt<Yx6c?P3TY)u0<^YwGQ8^Gt)XM;NW34|Gr1E}nH9>zI~_+>
zHb}FvlpLdeHxD*DzrRx1fJ@~<aZD<qeSg|qPE=>Bp>lU1D$#dsK8bu7ac-)D${w=Y
z=YpcnGmUMU6%aM@!nKlz(mJpuv(WCxcjWQLr$$v5K#hj<Oo81GpiccjxI~6~CF-)E
zya?y*%cl=XTDY7qV3oD{)|EcTh^hR!*8b=y>R>@Y?(XEQfcudX$;i*e3?NGzfeIS9
z-dY)bqpZ|D<)t{TOV#(^dKiNO)%sm>Dh$*cBPvL)FF!Q&JQ^Gbzl06Bto3o}n09O<
z-fu?=jVDqHS-rN`%ehxRcNI@vuSE`A{uS2T(XHe1HdIx{_CE)V_bUmv{n(33+U17I
zk6p&zJ|o?!$wo4^pbdkO>c3iSdadFZh3ba9+6+HBUh410g4MFdWLiM6U&gvR$3z^6
z#OyK^9hQEc9lnz35tsFjSev9<vgfVl;G-00%+^*0dk4c;dBb<J3-*Svk5C`S#0_b}
ze_qL);L>fkY+-$8eG|<U<$%Cu97FNDAZGmhY|mS<yO>H@%a~g065<jxrJifIOGC)M
z#SgF2B{bPhI4wh?g}4GGUF<9y7v#}${`w8xG|*!`mG`_&8V?QT{#MY(KCanDah`~t
zMDoG^RmdGDQuakf-#QYA4mO1k{z1!}^tx^hUR$W9PxLovfzOM{D9beCZSeM}tHcra
zGae7uiFfGk9SGC%b#nsTKly`%QCx{{!|Dl2TQ`%u(-RBS_dw^f$6m7(`v5ASiOmR+
z%`-2C!PArt8(Mz{jl>y7M^s*^U56Q6db>2V;R>ZHe9|53epH4ewp36PU}2QZH8u%1
z&2sg0&Rz7F#pk1NU4NTVz2fT^8^<t}GQpWOTkF&+=aE3=*T}pSNc1l7D)E=NqLT_u
zd4*S(`~%eTl$p*uDtmn;Le5}b10%Mur((vxyD|9yr2f@Q?(>q=r2Wpy=Xa3;b8a`#
zUC!G#Zz{i)sv?RFeY@USpo-C+(Z0?|71O<VnL`GfFaLOsVbDGpI_$qC>(6tSzFe_s
zr!^V=4&eHK@5Gi3uf;6q<_5X}RpBdbo+<vr<Qe88oWYV&T8$_JDXfs@vWNutgmUWn
z5d2Ov3vcLz6`eKdhIv?x_)5bEVl*(6g%k7Fh7}MHU;kCL)M4?Q%L;`WD{c;-%ZWi5
zi)EdOwa36k3fRTe!H#?IL%i3$9;^(~#o{hfm(C(FNcwr<l|@L~t5cBS#VwOJU;sBT
zwqbP2JoZ!z+6pFW?Y(<=7+5+F1O4fwteJ&QipoogJ*k%Q$bvD5rqn-o{p}JzSM|EV
zB6G`R#@`0`OJ3ZXmy9I3NH~#|9+<qesgKws7}1=-Y95_Z7d!ksm=K!biWMZmrp&<@
z%NDVnkL$gC=}fEVUF7=q^=3b{>dlf_qZk$$zoWy!XXGV`ml>@A*uR-;E`=-%P`G6f
zE+N+bOx>8|6c({10bv@ZM0Vy#=6eakmaa}G!qIniL=N(CF5Ca8Pm;<ijcFj*o<9@&
zHf`S5RR_HCB`#Klb7_bN#s>!EAVZlLB_Jy>oku7jE@zU968aQWKvI?-qQ+_f=civi
zy7#PXJ08|m5ygIL?5YG{?d3kl+Z%J|VEZ94)s%`GZ!M3qX~5z9wdyki%@(f1P9ThR
ziNcnG=Y}o-JW9Eggu^<Ke11gixv8S(Rcfdq6$2qeW!TGPetU~bB$K)>F2~zymt_t5
z%Qi}jLljVPn`!Q7<k6Spx?3myPSQWo&C_$F>tVSI=}UH#{|9*b--mK;Zf<C?b^tb%
z{KC*;IPuFD0$7C6)j?O2`*|gGF9%^r5^3Yf@1*xX8@7rW1$-nF3F4mn|8mPAjE>@;
zL&MI<`24;52;*DnCvk^FLLKHjDeRy1@&|)6ieWS@M>Yj?&+s)JEr_*ni5eex*ZO5T
zc&SZrlmSG@-_3UFrZiR!dZJ2iYC)@k7{w&$V`9E#D5&S+dN`@6=uxF(B5u`4IWFWm
zIGHAhJeg{5VXi3+6|&??|0Nxr88Qk8WaAU$I8ZR7W|6pQTw)Ilpg?D?IAo>vH&I|!
zRp`L0?5uOuZ`yr~n|5<+3YcPIpdhZoQ=IcDHv*~acA18M@7f~kW5G}mKR!jJ=J7O8
z46E1SRlnqher>@HE1F2vlgp}*a;}Bh19A&!G0a1PXqn{buh0QzkH=oxh(y*$X_XfL
zz$TB1@ilO$O|p&TiM)x)@w+4I-3b^HS=YV>AVde2;vC^I6b~3we^nqD$XIO)Zb&B9
z34cNJsfQ=S8bk8`Q1#YtQHAgN_6!3<cQ+y-4bthLbV^8rNOuig4xrNAjWp6-0z-~;
zGk}1U)R0mNBJb?IKl^yUd;Wm=X&q}l&vjq-c|u`kE!s%nd2}dN){dPzWqL!B&w$pB
zB>FCou`2<VkkTl8SB1etx~MT)O<nIOr_o?f&>(cp3kY}p-N{I;gWDC}x;c?wi3A^M
z&)?h2W|D8sLY!FW6S{UUi44o8sK?1A{nCMo2w}SOLZa3U&qO)Xs~(xlK&hYc)P=nB
z>WM@3q1V+-e)TuD893Ja*@*k_aPI%+7ztSZzv%NLG!Zqqt*vcNn{{U-zPVx1Y*F@j
zu-2^p2QT$r;{4#ZFvY1#+KMR92$0^Ri%sz(Wc{^?&veSuiIXnF4<wL0@i&Vbj>^xd
z0*Z@OSj8qGz)k%n4u+1#YhUP12mksE*_|?tp$R}7@?;ZaNJd1qE@-GD^xPf#<A*?k
z!Wrc0OL-u#BWXh0EpAV*TM;l~(o_~b{AwZR$KxG*QszZ54zG#;IPQLKJ|8Cg3&NRk
zm$27u7*Kfj#9HwpOwYpp<oL%qKCw10vjLglfnBazEr5&#xjXjKAonKIZ?Q(M{f!es
zBr}w|J_yDC%WXsk081q?By;8XYkc48yF^p{^-kk6gPUaPjpv70=LxnK)Gob$GQkU7
z-*08N=2}QO854S0fp1_Pu-+B>z(^AlDY?3LsR@fH!c~u8k$``%qsG6k-Im^<t>FHY
zAIH$W?(4h46`JwAw}BqQGD7b=nTr~xq#@d+mvu3hpaE>IrZy8HOo*@iq6s+*Ns*bb
z@&1t`Shhq`Qj#w0B$7Tv2ZO<8cWkFK=|~uw(li3F2kd{CuYC=-Tf!P<|JIxV=mV(W
zU*`p;yjB>%i#rEmILorjIYjYrnhHgKi4)Qs_39?&AMf$pC=O0xHxB!hpagwR6kZXv
zPn{rPK7V8O)42Q4a23IYaMA9b+8?}&PoRgrLC%yt6(0>4FCQPe*trr({(Jf2vH67<
z7vSgo3#xY2eZ1+Gq9VNiW-6;IG&_yG-=_K>5W>|&i|co>Y}tFG)5fRD30bfBvkl!K
zZB=OYBPJ^9tIpH0j>046+dn%VRfl_vgv=o|DzA9I6fMaX=YE^-vBpH;)9)<FcFpqt
zB&~iw=dNU3q_`m>mZxtsi6JgAxSGqhO#{bp%c~(2w!~W=F3*7;?kzRG!)~#aCr|%=
z=NvsDzsKvI#sh4T$6z5l!V$C~PZ<1|)P#{4J@_#^6>pzXFe*ttPvomoA*fJ=ff<B3
zg&rjYR5!^pZ@M8P{&g8hUKVxcOjJdd@N@MPp#6Id=e@mnj9!4S=U(RM#T2YWbi`jD
zL~O=2>1usBt!b+IyFSF_l(#57CZ-P3xl?sahyM)OQa``RjwLnGuehiWS55imSeR>y
z%ToPt1nWi1#CW`<wz_Q0sygtOKLeBvPJ7V4IfX^tm(1Bg6c-I&K643u>OMvOS>Hrr
zzA@lQb6zj=D{&5&#B2fMv9IUn$NPrneM$be!8qd!s+=2OjGT`xCh2JNu%L`?G<$@a
z@}Rq=|75RRN9cr?)uK26x!0q2SQ^5K<j-G=Iu`hSChIB)l+~-D9kE8x_3@|N5(#2`
zJGzADby2)>T;2QZViL#%<-&j;s5A6Y)?mH78I`n87*l&`X8fp`>m$c>A^B_0J8S0Q
zLHkc|tmqvBMoJ3L{A#KK>z!N(m8siHxWLxG9_3xe`fgpPk$}wY#P{e#<!eB5i2L?R
zr(t2_9l<QKS7D^Qyz&2Cf2QI$bCBHeP+|FnbF$p{>Pyp?Ph@RAmJxm-joBO;yqvoe
zm?Ef!Nlt#LHTn<vTe~WeH#(XW@GvO8A&t`->ufmw@GL(h-8H!SLCN^%QQ>(V8$^Fe
z$tck$|BmUUMQgml0t=7X0-A8*5ZW2317Q-|mau8UBw&prHu?q*J?-qO&<3)se@jU!
zka;^hX80$`T<khyhr6VfT@e6IaNIbO$388h=}oT2e<ptc#-zcBa_D;lq@Q<%Q0;wr
z7*om#_U}<9w9=n-KMZ~wC~$g6Kdfnda38_l!TAz4euwz!ga4@RglIvZa%;-l8k;<=
zLthvh{alhJSTn&X_?R9D+>yIW*DLegyO8fJe^_t25PD$;{s<JG$n;hb+fgnwQK_90
z8gX%P_5Jugg<+o+lQuIuq2i6GX14&XHXa4T7jeVN5<G4<mAzLo*lGiZyQ|)=jW;j@
zgph%bU$?}XqlEVL_?PApS0G%F_(sk#00oQsdfK%4d5SC2n!Le|sPp-E_W_&MGgio~
z4&#d@Eh31pGM^hi&o<m=DdwOV4zS|QGY!o9=s?yr7BVyabLjV0i{cd{$_gR{?zm7}
zR3!{Y3HN&@RtYr_litWUy&gY}Xv;|bIEV~@ytqvJ+_~8_B#VKS{nm_=@a^a@0#y_C
zPcHf@+n&-L=&q_mO7~mBuIU9*$2CIEPvFnn_fIm-eY4|egkA;)2JWJ<opt$$QfsI5
z($;mcx8dOlul_sthrRhx_nynJ(wK6$6z4xiQ@i4^BU<zp<ma%Mc-q(e&2wMerf?XC
zg~$%~BzRW-vH(A{t?gYlwu;jnHzkcr5C~IACN3{eNliWypm90Hxh~KVg~4El<rK7~
zhl0O{DTMJ`2C16ellYFF<tt*Orhr@0A6nO+1&-7Wm@Eizwi7y3PKr`1k!)%ZW8e?o
zN>FDWSMo)f@4ajmG>4MKFA)03FTVD27InLcJ_9DY<m?2Mz#zG9aX`tA-y|M`19aD|
zc@6KB{aMeGWQ>I<(5Bp-`FCv)LhhLD(s#VvVOZP{gU=4H)~%P6IW<>|?rZ~UJ6o%+
z)?crf0oE{LVDjYHgBTHBkKRX~(m!*2+wcCU)$s$Kz2TW3aQyZ+H?&D>T!(CH$VOuh
zCj5lKuuk*w1V3u)N2~F^twk!2s|Z}-zM%LCn&?J2sZOoq4?B1JuMUP<U#$kr1**(c
zM~!`{5pKW7R}_>{rvnjpjtH`GE!{mTyh@6<A_LSsXZ?%RImr~&_>5zg@7Qiwvm`|<
ziI~3HOm72}q=Zb|wWD8&Bh%;aHn1A*v-0X$)Pb}G>X2_KH4v=nXWU~=S}>h1J7J!F
z3`N`rEQW0>UYbWE{a3yp=nNCDHf0|e7+76-rUiSnBc0RtT?xk3211;$xB)3rkqnm{
z=A+pG3+ryv{SN(D;AI<Bepxt`-Y#1ERCslJt8Q&wvKDaf+H^4+<ifuy>0M{~Oz`~q
zvxEPu{GpAf6;?7aF#$7pMiMiNRVWY~qznjIjb4q0DOgk$v)youoshaot$B82`^pyO
zhHTDkP8I{H)@PLPLu@r8LKz9~Hgp|$(v4RXFvws~vqCf9QM}TK-FNA&iMuCV*?yxl
zEk<2o2y?4+vX#F(flq5lW>R->8e<ZBlz9(B^<`gv4}N!;Zrya#e%9_`Jg`j=J|U@m
zR%<y=q_HW-DR9ouCw2@0kH}B&qj^>CMdSf_9vj);V?w`$ldjgo^SQ-PgUYld4U-c4
zS}{U1@h>PHUFYi`gltc7y}kiob+>3=cU5aI{;PdW&-@~d@V=$?s#e4u3B6IH9K8~@
zR5?zQW)eLa>92rRm?9DdzwwgHuh=}`JmY9txxQ9yIhey>DcM2%9nc&Orq|2Oe=)V`
zh^tRH-#8#D6;jp_CuxnsBEeZ3_X^p=J+IwwRU62c+JLp#bcc0Qr~yBrdUi_5@BCCw
z1??p#E0?(Sd`_>uGeQgN9e&IQ9(#4y0a+m?R2s+tt8}f&hjta#U(CLncWRXVmLWg_
z^~;SdoN<3U!gQv@W@5o_m$vW~gskX_5h<OMD_k|}mY1cSA)!Y{N*jl|b!N}XII3C>
zO~6JBJK>c<lV@C3m2iZn&ERIbA4K*3hVLtlYX#llJXzyH@Nxl5eH+gVVQkj;{tNfr
zo8kDo@$SJTjD~&IseiY5-oKk>J#)h{D>x0a{NMT=b;HpgY;@xLU%#2ly4e!t@xOmA
zC!d5VB`|QIjmx4HYb=e~YnM^e^`C<9{j}t1Qr}3pE<3RB=7=b-onlXh+0qZ#_|pT>
za6Uq^WpL)Do_We;^DNeeP6uqV>Eqquqi0||1L~DjvqA#j8b2EIee%FmRCc*EY-L@W
zJBQ9L-#L9~p49vpG>p-(r8fAa)l__mO>0`kPWzOvx>E9<w^6OZ<(KE6hc^0UWH9hf
zGJ#My`YcQcfDPT%Ovm9|hD+wX>F|b5Uwb^|D~i0Kf8(k;<Nly_z$|$(w@7_>iw0G@
zalO15(afw;*PE$>DDq223P*l5SZ8edJE5v8ElX@hxCr_Du{28nj$vjqsBBt%FCPFW
zLtE!AP*p*EGpd)wt45N)sD+WbQ_{Jy5SILL0|o2-9)nW&Ao`9J^9&H~93AlZ*nG<b
z2+Uv1zwLY)FLLB6lA_e3&fgT@%}(#4AvF03Qw4+VT`KP%m5eK_mx0fo{Iphuebj8)
z`p)(N5bc^nc&g;p&R%WQlTfZw;+SJ`<`=r(>9;o}eA_EL1x`BIFT~Ld_+eV$ReKrq
zxX<uF{-$oUmf8LsU!0;y8!3+|W9k_X9k;bf;411!s4jcI-+6*JR3IW41tR+f&YHxK
ztV<^vT%(8QUMVk)(LAG3T{{miD^!9o{ct@TUGv=c327$wjQ({!hD!Mnh~c#zJorc4
z@k@8XqA3EVjB3)zv|QrXGO2q)tQOQHf^3`TcP)_ADnm3<c*w~Z1%3<xkT8cB72;EO
z_Uw-4)z=2PK2t4WHeAC;R!tZC=bbKa%*6%}TU(tk-UB|^TMf!ZS3N~+xXP4Lmrp-~
zev5~z=HA}zUpH3oC<a0iwAiOqT_zuEeh`Noyt-K}@zMM>Q4ryE@Kx?HH{syYv~cl`
zp$$xrW|sX==BS}BJI^D<@Sh`lA;+3W$FA<|Lf9D!-ApfJW3?=6TmCCb25MaDTMF-F
z&8GjbeOvnR%>=YH9)=rw9z)K)1fw>xFM0_YO-(dypUwj4gEXp>KbXc@rbszRq=~4+
zdHU~0VI063q89XY@0Zqd==F1-G4=h{1;N169!J31UziIR;B7XL&1-UknQdI^e4ygo
zu|MCfLshvVQGpSnT=T5DKhI@WUzP(X-7lS#MEtxnbkj5t2jil^*N|QS``pNh{hS02
z>&D4sJdBFVMglM(@?E?$5nV>4hec+$?>|ZKLrDCk-j}%qrS5xkKXGo^9F)uRETzW~
zpp8?cM!g;0F_CQ74&6Gc=lDU5@89ma;*0I)9b(EkCR7C{QV`nkZWBe?|B0KsZbd^x
z0MGFL=w&fil)mqjYA1b-^=zDBhuh|~w)xQEo5Gh4hL8{S-sUJKURZLc6{BtmQBi2G
z#jRZmYr(g*LaB_|=RM#hdN&<i%7+Uq8q`L@z4rq!0kJmYa9A0=aL`q}^rWFDxT_kP
z@wB=GTM!elH2_jFy`<^c)9;Y+>;w*T@2O$NjX}TveC8y{MLLroFm(#*N;>VFmv_g!
z`bqoh?s?D(wWg?pL>QE>Hpp0Q;j8t@^YYRTYFX>1Vx=d@eqmf{7YjYQ0;mL|r^e9*
zAA8OVV+qw99;jXGKW{N3P)%p6_)2{hbJx2;Z~(G<!}iA+%7R7vMg252<B!TcM{#%b
z<D>ik%9G+tI-&bxrK+i)^bByPe9t!cVyP2}vIv(qnFiP~Kd?Y1x}23%K^}RG6@^sB
z<^9{cRdhy76rI0a1H)$l{krT`dYB?l&xa?0AJnCF$nx~YRU=f7)Qlm{{lC66PCn7~
zbXyaUP_M*Ngyj$(=(QHi!S&DVZ^-<PGj^zs5v8AgM&G|E13H;l$Dni5LmPhX(=Zbc
ziN7j~pl6#RV^W}IItGt0ajfq^`H1YhV9W<mZAt$WsXZEI3>UM=Sh2%fNyrCh*^7>{
zCdHrgScY-=g%nnGvOIH7JMG|%aO_A;+-0eEJ|bPLe=_!hm)4a0zP|qx`{hLG$jLzH
za(?0~p<2s7gLZJ!`Mb2M{fffm+Ipvqw|`%I;_)LdV;@};%Di>Ei49ZaqlQGKc;Q0v
zFYOGnOH#BdY}EzCJ#;I%=4~r4qQag+m2?uv&s4^O=dShMGaojx>MGkGNFHPyMti@h
zrjPngL}Ju|4P7wc@f_yONAV*`hIQeoqN8*en>W4(4}5|ZmHn@nPb+y_yDy3qc-%cu
zqQpkFSPsWX=#<&6nS&WLY-emv={vm{!WKYPpZ^8AQ?Go3UngnFx%0XM{M@R3+XHjA
zQM|c$1>;s|9B8c`yPuQEaZTMk5D+szqZ@Sn@4@<~DngRufx>XB{5zW!lbYJ}O3^ft
z)Y%bTB7a~tD1dhOW(%HF*8Iz?b*gFtw{A9$m0?6;+n?xySMDl=bjfcoKwn2kh+x;3
zS&B>gmN|7N@x5N*U@n`OAevjVt$+Tu8%>tLm-oph*EX&#_{OR7pS1h&$@R^RNyVaw
zZTeN&XeuXt8a-7)+LY2hf40(C6}V4g36C;*F>$an{8jY@tg~}sYfpMi<snapOUR89
z4(76rj}!rf#NsEZuA@-i=V7<oSFEoQ0h!qG9*wO?6#AKhIyVr|j5`NHk~F#V!4<Y0
zNWfbOjSO?+yaJ!zsT@rjf?PT68vJr%Yl#<JF@bS7=K~#EB<kHIdyWo}ZP2Y$B;8@4
zw~Gd&6X1iJ?AxDG*vBMcB8L?VjiB@0ejJAAr7-VCi;~`g3uHi?Sl`NekkV{+NSy_p
zHNoG79RvQ0B6HmQ>NC~eUzFWfRuBvT5AJdMy4c10@2-7Fb<<&5tWsMI;FGoYVFWK1
z<eX79-*c%(--e!WuXqpR)Cd}`xxV?8`50K{Vp8i2yrn{be{C%@2)Boj6=Eb<6qLG5
z`Rv~^M3#I2HP?>xvJuW5!Twmqqht^hol_Sx1=82sOH!Mulnt3EP@Pt`1wq24$b!38
zPt<i)ali-p0ey+9W#MKj)woxwL?w{hV`ma1r0bdEcNq{zQlR4`K^DGK?qSf=0^9(Y
znWCSuI&#l>^XYx0l#emc-}Byca(<7_doNqK&YKxPXU$+QL;n1p<<u!9Dg<1qAUZAK
zfO~a$OMOq7xl%gXMa55#-*KscCmUp7>)5a(MSq<)f$bdOGbL&R$dRN0)=2C)bbUtg
zHr-`PG53NDrWqS=X`q(Ha@MYW_Itj26WL96TiyZVVm-`@CBz$?hy;elo+RHkBD_GM
z|2J!JV3^Dj8kDNlmFB0^Wa8!#E&|z)jw0AOAe<)&X+5-WMg^b>PTX2FYjW!R867)$
z&{|-6)6vNctZ6)EO*%6YJb`^md1SCf(ZoV=OVv78nO+8L>D2H@0#-BuowzX;ccj_C
z;O*jUF+`c?pE+yOo(`-)P0aO<1gR;NmhG;;fS}5^XRqET2HD@}%5oI(AZR}f)*MFl
z3$7`7_JFrA7fM;Ea7kVpl%G9GN+4__kdB&S^yTU{v&RDiq9vyMH7#ym&6#;tT(_vU
z{N!9qyxFMLX-@zVRq|TG6>%V?PkFTzoesCsf`qD)6rD7>Mld+{Yw7Czz?iv}N}_%&
zgtu!?gfiLUF-H#n(irF0o3yuiYXLo&{~WXb`l;N$$_!W+2-Qcn5Jh&zt5Dk=dny4X
zWjq|=8TxM^%cZJyOe<d+HDx~y-_yF^fQS?r>;n?S%pc?Qybr&#gR=mbVde!-$Bud#
z176n~B4aEHYx|>}JRqNlf`72k5#keg*&+oJ)MUjpRU9$^MbEMGiRo9M$2{iZ-lle)
z85GDGs9n(O{qa-xb?0ACIEr#FFxsBmFUZdOxF7y7Zmicjn{%Gs$G>OATHH}BWZzpD
zmc{=teHx*st@_m_7Wy;Ai-?AeQD)y5Vp7{l<5>(uaukQTG42pW19~Q0H=4&&J8`qA
zZ))YBO40Vu`4#Ef4$f-q>>uRB-Bx_Y{k<4MIF=*nN&`{1r1|e;YqP`^Eo3iCHRj$0
z1R^a0PO~}78r2FTW%ST~>C*qV8bqfou_NZT_Uz$-PJ+|P!67LEThv2yalIzYyLQpA
z$!f2<jjAG+j9}ZUWQ42%W1qIc8n3lYNqUj@`gHl}ln_S40O4Gjys%)@<}6^1U;VPA
zJP)f87C!pZ+IhB&^}wxq9*Mf%5@bEU&81<1vx<KAVCU-`f*Y?@%k$l)V{3NQTfgw&
ziuC@&_J%NNA0L}sUhq_ug@7pILU$FAA(0~1+kcd)h;kFkiYBlGU>(!oLJsCVi!*2^
zImeW#AH;=1DM}05=OlLki~!)=Zm_cPI4<>scK(UeU(ZQJ>w7B&Iy=r&Ia!X=XLX`L
zo+@G2;^K_FXy?J~A^!>E?bG@Xm7~<C)p!X%E*3n6;Rn?oK?XDWjUv<=q&(Gf0~JtI
zV&}ji;P;d0FWz~<4jGG<g#?Fpq9mXXK(#H6aXf!D&5i<teFp0X0H!gh05#5nYE)*-
zDy?=18|O9b(2^9h{$>K+wJA#SIsRz@`+7`&L}jx6O1s8Yw<pZ<u$Ql_mw*2s8|Bp%
z)a}Ldwt&;6P;_yuhu|+t6p;T#?C5LUCmpahbzS?FmGjqr3$z5}UgeeSMHb3tyixCh
zfrT)$2Xk%v6=3N@2u@M};{cyVGhKnXh{<SB+ivuN&swxuH8Jv6sWSk<;W(cmb=)l)
zpv4pl@7&1}*EJRI8(rKRuQPpE-DC|?OngVAEl0BzupKt_XPwjb^3!#bPb_HNk13|?
zjn~}Ah;WXg|MH3OTP|KO_u@V){m;6^gNNct$LXnC)fYM-TeCfS8o1AuQ&`!(=o7L4
zU(R#JTvUW47$pTl>5e<~IUnT5ardX-e^QbcV=-uqR|~=EbeN_mUs~&ZXR!PT&tr=u
zJ}O`q<hnuSLkbSQssJAr#k&9u=SkR8Y#ANz>hY&=B~%HAeUzxb$pHPP7)1f^4xLIT
z>_9*C;VQaA20u3cEy0m<gKE#pJ)|r@p9Vl1hXFW6bR=PJwd3_(hhe8^9V+K_qJ=u?
z!Qj@=gP?L8-_954>oz5XDE$YJ9Y6r#r?T>k$9!|jpYW!qBW>|3f3*aj=P^s3f=I_V
zhxwl+D6@33bpt-WHdNja_Ss^g!e}5J_9qc$D1hMGE-1dH%Y+Ra^^0x^<D9xe-(W&?
zinWJW>ei%3{^r%J97H14K^MlOiCCq}`)x(!;p(Z;Q}}zC2%FCHD1e`v;k`FT8yF19
zrv~l<QupHKQNvXZFhTRaR}o-TO$3U)y?cZbYS=M_F!rnk6s(%>O*QKudty?=*VwI@
zK`E3v#4cg=Ou3R#yk$Ngp(80bTB)%$cTz4Prgsu`rg@>{-X6o8zUO^zf19ik&6>zi
z1*A=Zq=bDHkK@dosdyYI^vixrS8C1Rg+k6fZAfC+VfDy_b6t(&$An=sbL)kWD@Jr1
zBTIa(MkW_lDzhIf5&epiWL)^;b%<ReXfX7P(3B)2vU+JEGuGX|N3y(7!dt1}%VBz0
z*k%%b;Kl#>reT;g*+hGLI~3&}3`LVam8T)fv`$o$S#~Z0WNPU@*G#OdR3?iY<-jl6
zxg_MDoCe{50V-TrG11^Vm<)ptW;$k3YI%&0F(w=E1}Dw);PtL7ag$?`i(84Ro>`^=
zU~JCtP(a25Y*YlMLX7YYbV`U_7saqd79iK0zqSR&Mz}~$CI9|Sl&Liy0wDAc;bKuv
zO8Ab&0#d<TD^*;nO2<~X*o>rfpT_Ui1LAkIqbtSY5;h_@$i_Gz>B=XJy@J^vwO^<9
z!ifs_e@}a^_ZTJX;>m!(2dyc4xs*hm)bnfa?t9{KNfs|OkOoz0nt5sV#GQYN@okd<
z@)P}|2!cAAv<Z-s#CRnrLn4+{YHe;ybj(w)eBgSi8Zj~p-Yxl*Bammg;9I?V;BQ!6
zU|qgpsD+&AP`#y0GCmtUOOh&J-J~K(cmS17mfs%*(yrcWR^O`ROM~q7d<cJ|s9=Vd
zEk=RjGajt#;(Us~q7>-eTK|+7tMez#Su0DBc&<&>W&B<sCj%)yc<XO2H8OXE#LyEw
zGsSBwA6NmaAoyZHiTzqO(~Pq9k`#X=;mJMe?S0n|BAbE@J%^g7XBP^SFE#~H>7D{u
zo}|-4MI}?lH9?X3CU2L#F)X!@xTX`9=&IjKs{LgW7dm%4asH6wEi}o4UbMK`fR!WO
zv%EnA|4?;ZW@0toZhMap@Bgjj{KLfRG5Rk`xi_#n;Zg8EwRm%q;-cGuk95l%#dT{N
zJwc*ki<=bt6a8nX*e^CX%ZflB7Ug7s&O%(V7KLRPJcUA-#X=0{*fsOuN|@h)P35OM
z0~B-5pwwKkh=6Ti7Fc@ZMu}QIL>8+L3js7Jz9L*B=yCS*3&ZiEr0V-S!H3`&mBF9R
zYy72rdr_-;LM*@z1))DO4xr;rCKzw&Z9(ANqSKgqn5N><T2;lGED#@%2^Z_hG(tR+
zyK4Z?(4Nb$(ieI9wyMjHrzYg~6O52X&8@SNCZ#9=X^jb6+U5<;84Y#g?xU0_15d{k
z+3NyWx9KErXJFmpr{nXnPNyE)5qZ*ekD&|;_x4OM;+2*`nOQ+h4@5A9s({d!_Z$6W
zr-cjaGDvb*VrSVPKUePrS3^4vxRI*cyQg5o5ejj#HY)%H0XKZQ6_(a{brO3EeAKyJ
zm|CPB@<b`$$?Q;rG=o-cD!Vn2Lo$s-RlRz|G_mM=xcN6GlahO6q~;p=<0tHYMo-fi
z>`Mja^m)Y9)<P^;i{vaNiA#0IliX_1e1kCi8+x`&`X-Nuy_(@erB%tpyr(X;MqLKm
z!a<4#W~mE>8Z5FX^4G%!pf)c;xZzDg;YPd%LWt<^R9+8z!N;nrP8j}@6Bh;uK@Lqc
z&FG_c^G2e$q#kM%s@rL^*d`6VLqn*)tY2<)v(E{|dz+i)zbp`vcx-~=r-A<gXZQKj
zzi<4ds>)m7*YkEfTWMDAy<b1$5Fo(I>F_GBL|@1RMBiKz|M?FBkogN38~a52%gv2%
zwp>V;x36!Ej|k&IGh`KSb-t0@UabH}-&D&apB41PThwKFA0y-imUbT#>l^Y0MGuWt
z8qrjoxLOb^L=9vduF0Z8CXzB|Y&5}iix*Gr(Vp~tC3LArT$!}WtiuI>59>|UmiU&+
zxhBflmOJ=uZud*QN7!GQ;fy=GKF*&6mNrBMWc+ky@^{pmGj-Hxv)wxlfFiy)?F6k3
zn^2PV;<XruHRkdrUDzkMVzuIRT*4vQg|yxFr{@Qq+jNZk`KQ+%1-~p`<NhqGA^VH)
zj$W`%uk9TDG=gM2t4$jV2qYw*>HH)od2ZwC^|njS2vnA*15rICHZ<aLcqt{2pvCAO
zvrp`oI~fyDTEfmRyym;0oY(rTvlFijMaj9?BIzb`owfJ+@=g2vL-3Ky=M2lCuEIWF
z-l1Vv+$gIlAK5s=7XsBC%`5$?H#GMi8JSM2^LtAZF(~gdxsxTa21kN#<?4<p;p%uc
zTr+XrZ;c1ZU~F8oQk=76&X>1Dw?4o7x9&VMEV?lfU-mgll-MkpU-$izY6t6f7_h-9
zrit?HLE*+plUXwp9}4-5PNW8ydfjO>63O8CMD=T#r>`K$4@1k_aBKIiGzM*$*W6ry
z+pegQqm&b&(eWR}Bd(;^S0=*zCgeU(rRmtgdA|s@vN1Pzd`fOQ5@aD%$`$`Yx8Sq^
z-<M)-PbUud(t7VtyHk-^Sej-6r+uvde{Ze+wd+cgA~&}Zl3?(OxZ<NCV}AW&(P^B*
zLPoLiz=(JPA&DMo15&1>-~J;w1R_kBR;Bale!?$}lz}{J=Ljx|Ndxo;%mWxP4+w)q
zHfy}=^5q#sDkf>|KCWFVz~h?1hIEG?W5uEu)5eq0arRZ`s=axbe_x$0x0$erAw0PR
zX{#k{afW-VMesGd5hC!TUlp~Ktz1G<|D>bBnHz@Xq0pdV!sSLuI9B4t80*4`$f@xD
zeO&;vhf}_VJkvGt9QI&jc&rIMtJw6+UzjYeJ6ron#{9jmzVNIL*E8>tp7gKq;{s?T
z;Dg>Q5>oGCVv-pw-0OR+4N#HEigOP0HXoMwboG0-D|elNI$G%6`k_e_tf~f~1JJi+
zQ9H9g(%5HR==|JkqCh>xdYv5e<yJ|LIOKiH1>xN*cSb;gktwY03t?BN79DR>V0f>}
z!w_GU^Pm&&9>!L1vQs?12*%aXYU8SdaaL?ey0djb*OketckR6!k3tUHeNnw$aXJ!3
zJOX^z^{J*aHo;Xxjkiv7txggqRSvDLMU{#1@9kQhHaSx@3<-x(?6mN|({eg9I_#Ov
zr}{n8Q-QD{@P0oObmo<mAml1z>4Pe!D77m?+H*|?KWp4J@q4ioO-!l6Y%kkj@f-e5
zCV5X&{GDv50$tA|!OjXK|HbJ?p?41ePTKD^;Oo%JAJNj;`!Kn2YxUe~CvC67j;u^`
z+x2E(#f$b9rE<;c24Z_6W;x#P?e5wZ3=|?l>HbIPo#cm(H_Bmq|9J=Dzr4P_ZVLYM
zYgd|+PBu?6)^rybondjhrUfH`>u5n-F}4Bz9yOcbe)(6fZDod%fHXV=9zO|#u)z!T
z4VkAQ?3T1emorYJ|5ufP%8xnzFkm1bRos5`50J<K#~1kyN}1`ZW;?9E(z$E{TpzwG
z_a$DZ;4=zf{*oB=&?3B1A?WzFK|FUBMyC^$uRh({(w9r-VF=*&mZ?QOsiQ?#aS}U2
zNl2ogb*o1hc9??j0Hyt@uW81ZHI)Ir9pDV0IDmQh40?p1&4|Vx%E8h<5(=Y)#njnV
z<!!hKFlz791*X*JO)%O6n#awdOUY?(5VR{@Fb+cU`LIuvUUl}#3e-KW$G`1|&j6Wu
znT^-|YDyXWF}Yqu>!Q<=-?pSnJU-3g`ki_q7s3LeuoT@XUQ4dlZn7emoxG3gQKAV?
zT`$VeOYB{uO7SvFnVh=ilZ4=fcLgywmI)60;d;|fKWlqdI)BVL+Y7WUsAwQw&&70p
zqD0bpFqDXeFg8%LeCvtZ*VxGx8j`DNdvHd523RTO$JdSG@xRFtI6nq#+nFa#j7u_N
z7m0c#*(VboOB~4x2u+ivl@C6>0i3Fg<T_*K9QrI@+S0ai3O5ns+&GdnoAyQLp~R;z
z4%GL#>IOr)^jQnA|IwWfn+&~+6Q(`r{%0V&en1R?g$rHd82^Z)&$o6A-I7)c>j_p#
z_N|pIp7wG!6AC;tGM713KSOls70R+emn_)(e(Sk6^RN9EF9X3eFNzj8vHmxCWgPN(
z{6FMU;<uS^LpxjA+7h@oG`o&S6lYWypT;Kwl414wBqj!KTEzz`(nDRPM?qF=Ot40m
z94-X#!)6oEh{OT33&4i+f>`MQ^JNt0I&$k}DaqxwZjE`*^13YGn3g23r^ACFMf+?6
zTVNdNzc8k8tO{Q<j8%-=7)i65(>b<i>KZ@C*wFK8q>VX47$cZGmz~vN<U)_lOuM3B
z0TvL?lR-?g*%SD<<Ps*d_0`eIMxWyq1b7rrN%;wz8@Qo@;*Di1H)SVMWKmn9{1}A$
zt>x+s+1Be)dX8JgbK*+CT|BlKp3NDnW=6fCJGAxI3lj&u!<eZ*wcwdm<h$@=dkIpt
z@PcJGIdaV`&FeMorliv(xQ3thHg^15pT%D39xZ+0?}<>PO~i#jdXvdzT0a@o;wZpN
zI#xH45W(ndb0-s>D0c)A+aHf9CxNjprPXy;ejCl^Jo-FKn$ZeAw-rFiF=o^WLCew2
z+NQ2)71OlCf_wZZ5a#@YZBc3?ANvSZJEI`9af%u}DCxwv3^c31Pagr*sL$ydS3`Xs
zZH9#_Yd1Q3YLtpkb8%~C{Zr&({lx&4osZtkG@DEI49u7{D|Se_LVq)x(Xv3s4d^Rd
zW;MXR>O%pJnI0{itt?Ge)fIkGnmJ<R!MAh}(K(IRP=1TnVK@<KLQIB?W;7+F>$!{k
zCromx(8AiGz;#mOX}_6e<zjuy9N_M8o$oHsl=JNIBoTkTUAs`uwRP9`WX}Gj0pzxA
zUd&!zzW2ZKpe`6?j@189e9R@BTwKze>wA1V5*UVm7lz(B9!mSY1a-+FHObnsnA6g6
zhE%m<ag01|H1a<`vLfH7h#$}21|5F<D_!(1Q9;jMl>lcZE9bOwF9GBQP#}X)J6bv*
zfFwZi8hC&f2@#<3J3DFrRdOT)ChPCpd<D$&(w^kHu|^mN9W0}^ga1Ws3Z(U2y^m;I
zC-u?(pCn>qseJGp9fdMzov3n4m*=Mq1{0a3b^-s=mNy3=`eMcz+vvtGzAm%EnhdA`
zptNWV(mN`8`UZMw>aAfFefu6ENzmC#VF6(4sMpiZa>mAf^Ywr=I@=T!e|N9N_<rPS
zIL1142^$A92Ys89EDcWn_uXrNA*B87$>zF(_wCmApJrz%yWJRDbQmnajKAJ}H61Tl
z5*N{zbGtQ|CRc+h1xVlf04CYNV$Figp-=s>7X@uciF&uI5gb?ct}V)r$3Q*)or+=e
z0YjwzE&W`opHzApYyEw_3odWGOEO7*o;}r9!z1ga&}9XdFv*vNJP%p|#GXPe`56<M
z=WZMRmAyS%&xwd*DiNA%l^E)1d<#&5hZH;wAR(t23*5=;*Z_Q}p?7~_awm!eq}jM9
z8%O~am6^SQB1P;s?YoG*r)TAJylp`q9Y6g;f#rWuyRzP$W&HH{pJUAm3u=G_EELO3
zn()QSOWQgFS?06Gg_W8>Y45Y6je-7(>3SR_>*<z@iS%{^$zDpr*V1+<q0Cvv+w`w2
zq|-QKcuCJ)SagxAD-pV2g~f~ip2G=zZ?ukFSXjui<_`%Sx_yWq5(o}D4kX0WjhC?v
z4-dC!`ZD$J>Hqo%Ba0=Wdn)ek^%}W*F+X;+jGHReh~z>Ou^u}XU}PCoA20c|;YIV>
z-CCg%+c>KZtAe(lhC0Nb$9SSzoEj>D&p!)k@#!M067tJiNHw;ZFvv(T=gb>xG$Ql~
zGcHT_eH<%ed6Ptnu#L_Y`J_UTK^DayQRc$f;T{aySgjXUbIiD*=uoyuV7h9({MH=~
zgD@^D=JWMQ!Z_y+09msmNXbm5bpbsbFa%&@%jhfXN~ZT?oX&0+^Y*U!^Hbu^7ORod
z;0UI-2s%MAylpl@Y*t3?z1lxnpzzYx9B+do-|_G<&m#3=9Zg0LW8f-4RwCu^%gezS
z+bHKODfPDXMgwl@#4|Cq)`!u&i)kGD*9yr&EP7llG51kHi0om1_b$zvXb7GQ$)!BC
zmJ|7uDf>loB;)eMno5bt%+psQ&X&Y8t&}r+X!o6WaDp7K(Ej#q7uEKLSTeimRb~ky
zAB?@%vb}GHQjiMbh(X=1+;Q*;S-s!tnAl(@#C~>Ff1A;D;IbBc7}z9Yvq_}M8FQpk
zzAYgTjDTmCeE@M)>16G5Tc3$Ac_u|S+MR5=T=nmCz9LJl!_TUpbDmOFW-j^1pNSFE
z9e?G&ksFhf1`y7x$xCds{DHxc-CfhcEJr~zV~-D;br;iN$tj!JTHZYGm{O9p4~md;
zFKD}_)T`hJX}j{xVJaat^=nZIiOxiwg!SXV{NxkQASSOwP9~^%*uz6s_=qC4DILFg
zSd~t633O}Qsq<olev;kq?{Ck5U(tx4YdVY|Co+s4&nL}9{&Lt@0sl8|;^yH=y+$F`
zwWn1A7ka60#?ghm=N$*ShW<>3IIwB}*ogB**9YCmNtV^39)XdItqNC@NwJ6hz?6jO
zrvgVZJhmP0an^uJ9T2qOPTgbd^=|Nd?yq$k%xRbyghsRyMjd<Ib!c`NyP3?AktTfP
zvt@O3mb|Q;sJ`eu5tfeU<Z*ZnQ21quSx1E~T{DrC-!kZ>R;OvnvI{vnF1XJjbGsAa
zbS$2}eG0i75-Cctx!7nTrE%7s-6iHSkOn{lzUl>JSGRN2t0>Stsbd>4nofK=x>x)H
z=-e>^kmyx+VjZAWj1hS4eUq~7IV`Kl_hRquzW;Yu@<$S1#H;mDLoBI_tO7&CM$S{4
z%~XLU3J<Ey;VzbZ0RhUk!zdu$TAU|}3p+3tE;8U)VQpz{1ne8HQwdF;g-vnpBsfRg
zkyaF@J1mMlFtMDN3-2t<bIjJT6Jh?8uVtP6nYN&_#LLybBAC603)eFXc7A(O7wPj&
z)OPp|yJhLc8A~%UHPptE=^LTTv2$h=bSvqy3?a6xa<t=en->KuadaJ3FnFATbp4yW
z!=%GVMI@iNCJ;vLMEoOADQT`*Y&46CFsvVB=)C*8?PKgvRlHQRxILJU-7yvM(Ccrp
zaC8s2_EuyvOjggft)Z&%Jni6py_Y@O`+1ZQrQeqT{8;1jUaJya#5&)?*~bn|6?2_N
z#?ctcF>FZ=3Ipr~cD`JSp=Ef+Z7$iJ6|;tqhmWbM(+?#>(+{zy`A<*jCOnO;60v%$
z{<}hs^(E;))?Dv*xXNp2j=1eo4d%Z}yEhssCy}PMzsv{^RA!}rur%#9?Y2wKA%`@T
z*-b8;s2Ru>E@bPZ06I$ObZ&wU_b;j-Gm9k^eHSh-2Qq2V5+>}Z{+ycpI1chtUfN2*
zizHk($F&5$ql0bp(}So;eA00^aP@&49WqO6rS+4Z^+T^-?vG~$y(54{n0?Q>^SslA
zGXh9Fl+3whLscX5$_@{iEtTCP72@9F;U_%886<@uPqB*fnr|d~9uETx|3KL0Q1J&@
zxM>S4l1oVCg~XgzZ;>Hj%S9L88O2e!(4)c33suq<5p~F9TC{Yd#v1;idnv?eQ`@Al
zZ0pX-EUY3=8T9CXhq2;ony$}ZKm(BRD}PIL?4sU#IrR7Wqf|!5sTtc|i!y>HO7-*m
zrLZ!qKvqTc{z4I0`SVn4SXrK^Any_$pgOoq{Y{I{PF44cIVYn^_I)aPtX{7ZFRr_#
zfpTp1EUsIQB|9vH@vqr+j5Xy;jF0i=FXlNeM1r1yi~;YB8#A=z4o6-hzUdYrbV_fP
zQVG@~*Gwk@Vs+`WLFUL~L){v}(m4<I5}b};jaw|o_G9y#8-IS;^nTec%FBM(k8Hz2
z&=Oasr~TCOTbS@!Kv}AzGkdXW<QSaGIYg$hH!Lbv4Al%<4cRHJ`SlrW-48`m&fMTN
zm7WM8TJ^c273aGeT4831Ls*!%j)h+{l=)JlP3UryR!b&tV*EN={!XT9um)*EP}h6s
zWZdmrJCQGKn6S7oj{NuFW{&LA?i9EIFtgg22g}Px4}8JF3#6?KRN7-E{7)M3f+9UL
zGgq3Z!t6`IKl%7R(^BY7ASrq#tXHXDTQo9Amw9;R`EHYmwQ`6Zlr-LTq%!X!dl()%
z*8#p4nF$>V2%m$cd#e-5lvD}lbtgb5{qWE@Zh&M_egWdBV(XOk_Y_oU%j3w=2Pj}k
z@(g2lJi-h@&maQh#`{D5^}aRd_1nyDIT+ThVhMO|QZ0^piVFXOYJV9yUjnlw{p}pm
z)m*g1<%q=^M*m6V?UC2ds&|D$dCVkHJJs!aRz)*1#00vL_#Uku`S>%mXTx+x<lmsv
z6zNEoZZB1$zZ0{acSFXw9l%{64=iAT;cA|U6d3%#6cS)FB@Dn9bepG?tmm7#37Y13
z8(`&jDAnoEyrKU7*B}EBt13VxPyp3p)mHs^AD{Y(C2q&JMm<g)(hrt@a}w2SpiXuC
zOTA~@!?JkK1_Ih|%1b|gNT9v%$CjN?eYm`VLY)6vnZ-iHWVI*mWU8Gf@+@gZ3q|~0
zW;}8>UgVn{4jk2+o<@r(&dC{blFuI64?8tBwpFTwNygOQa`?0#c0?dW%vku+3-3=2
zZm3BN5n)$I3aLcS>O`C&D*@o9-j?r>qjHYKe%T=q;uJG=mbL$>Ao5Qs)7Do?vSzy`
z1b5P6{?MOqD9%mpmV3#Ppj6VVzA;@DW!z<LS@`1jEBX*VeiO4^x!?A#7&|Alm3;{I
zs_8ff4ZNCy3WNP<Nj1jSh5k}toSZnxqie_3CwiNIfB!sJMIO93S+Jk^<?LO1(5YI_
z|C53+P_cRwm-c@JpcE9ls4vxZDPcguujJQ+BVKh!f!WWEK`#g$YM+JZ$O^8pUr$Rb
z6R==zaU7m_)YaDre<)=9+l#*k({^0vHk0Xijzb2dTJTkWqBA*9W=j6)<EXq0P(&oF
zrJO{B9IlgQh^<E1z9@N{23|@AR|p-NC?47bx|%u5KpRE#;fE>s0XU^7KXrm-7LSTX
z(5>zmDTu&QW1PoK?faT75dm2oIa__NknqNxaENB`kt3Js<7)7XvJ+E2ud3fXs=>No
zW^+Byoylmv=!M%0XV`4sM~J@D4<V+~79G79mKW2tq6Hqm1^)qJ<K4gY8L$|h2$T7>
zM(LZQIcse_I%kHFUaYI3(<`&BD6(PrNL4c%$4P^}6X!XF(Z^KHeXfGOk+7;7qr!|O
ze<M}~RZeE;lp1qGKogzi{JwhWEmL!yr<9%I>eG0E4nK&;rDyIvPDM_;9l<fMM1m7B
z0S%$5+KF2f+>t9DBjo=m!ULc_4;(EcW1hVx^anBgL1mn&rVptqF?L*VsP6kkSwkmF
z?>PANzjO9hj(&=r&cV4dj&Y<Foo#|*e*p1&$(#<u7BCf7?dDFFmYNs$)iJXw>=<rE
zeA=pcRJo;*TQ9`DHHC@6!*6yLKQS0If8V(=Q(LQGSH$BE%4!9$O$-RXF>6N@fE-Pl
zMd|xV3D5&<ge}VPmUk!&uc)oA;h%+u8VC@{gH#W`lkCw7F-Ypl2PT{a?@QZ`#L+%y
zvz=R5;e6Y7nM*YFOX|HQ78d*eIbel5p_$~d0!wfq($lhiBBbhBS^(9uQtppqj2oFP
z-X{m@^O*qJ6MJ4hp)Xf+RX8rVclH8J5N2`ugc_V@pT36|^;{%H0KO)iMshVggOM)T
zk~qkZ@f;dsM#CV?06;Fz0B%y4n71#M;BTNScaH&D+5B3fyg!Bk{(&6KqQXsUtuV%|
z)y2T#fWZy8!C4X|Nf@i>_O@ifL<|TGO@Erlzm|i^`@W?U+%G%Qd6no1GSfY)uTYHm
zMYq3f&8v4I<>aK9&qwNWktII&k$l!u?4t;Kn1NCW9DTZ@@pD2XzdV-5BA-&2!>W3L
z1rQ?P+DkH(=x=$0nMTjP^u+Sxl|r?=f#5$JBc%{obKwj(jn%?46;^g)1j$aq_u6)K
z8>MtpY()$$P&truLRb_l#L(0*b*O_REUdZO7Kdcu&*dJa$2ZZ2f1G&5x{XAt&Ngqi
z=6>mlO+LXBECtj0>~);xdF!k9?@0vy5UW!}Nu^sry^pcGKJhnRV_h|Y*(Fh2awWG*
z-#Ho+oC$W{{X6i_J4uL*O%Q^Q4)-(&Ja=wWxquW(zqstU!fO{_cVFw+PTDnw)44A1
z@5R`M-L5s%bpiBSbPM%X9`xDd95%BtF$e!hYy^=Vs8HLQ;#-?#1x~68W8KdTI)TL0
zHKb}D%}PU`P8$ORJhF&+VfX)<V|PHgotz_SBNA{)_x!8t>(QVe_ZFyz)uLAy%Tv^)
z`wAGp{=nIuJyVv{{`Wr>I!3}#!=38D-^|7T`>oTcwdgUSqf`1vFDpVZo)?hrb(H)*
zLn3E)Q&s*o?u?)^)nOSN5026QuQ=9he?cGQE#cbISMpQUQv1ttODCDV8LbdWom*yv
zI_6eKuS$s;e3AyZ<sqJ~joj1&sm<Y4;qT#BFt@tk?saO-iG@i7JZ`Y~?_Xn9HD`9*
znR|0FJng}r6UObWLMjA4@woy7$dPC$b{&1b=SxaJN2{k&`+1To5r<KhL6dFC0J?pW
zDPgJdZJTj&mI&xhlyyOhhz5wr*q$KlJs$H_hk;)LfEn=8gOCK%inmi=%q=_nYgNl9
z=RV*a&OabVpjLHm`cu@G?*<k1K$V2v!dmyd-+d3KcaeG}?~3Xup?`sIYYsyDe}n82
z<!p9T*ZV4a`pY19aEBu8vx-q&BI_kU!eCh@XCE&Ke;riWwV*gVcFp7}al8v5i{<^f
zo8JQmNqH*$7WNE|8-y>8j3tF9(}jkd5Ea<m3b@s$CaQ|OJ&S*U#E6`ZGe|H3eO1@%
zdM7f1wz{+y(OF0<IMDHGt~BWB*#?QGnnxfIriv8tKm<<aMFbS>J-MfFFlxaL-g;Sf
zL+?=#A|ykc3_WQqv_Jf+J55|cv$^P}ilLbeGV`T`<~57g^A`)C6RO^>`YPKj?DA%e
zyDd!cQp%p}o^i>}Yy%=MC&TLdy4`vA1fU9g_+ViS8h`zg^Bs))aPS}Mx8R3;s)X(s
z<GLzkqNqX<$CNeihs7o<)%2Fu{LGSjzwzb}9v?ZXM46KCaGC#_BmeJDM1FQA3}eGs
zwlAn%-|)%-BE4Sr{3rUrVFqBE4Wl4E@M$+$Enn1DD+4~^@}9jj9iGz*!=Q^&`YMk4
z_vE*hUX2J%x$%dY6~XCj5Kie0<xE2OZE(Dv8$sOc?XT5Wx6&j7jjLMf3MT9`Lxw9h
zndNO)c+el&sr>D}x!~LTGtz9iY1nHKiA^)5X7$omoq7EIu{B$=R9$;Q5QpS>`40qu
zZjQDYquA9yBvexjQ-UoC>~P>evLH0SV~kv>Iap8=oX+!NT<0UpRsYu6%sZScWjs%<
z^7HS4j=srvH`Cn_^36cCGiAh-)R3LLrNIOz?B#yH?q}hyoriJGwKX82xZmdA(!;I5
zVqeC(N(|#1?wL4|w?(|?l>WP~$j#y^3O%_dy%X}=$KrGSCQC~9#o;AThu4@Sd-#^3
zovdN7z<lgpp5$6p;;aS>&-g=`Z2TBmVV$i4sqVM+C3^J%<xm+rRKaZJc5uXya5dY&
z1L6*-^k*iwvHwKv?&)ocp}hdl++<i&oRC&^Gd7z^Xgaj*C$8cCQLgt7z<NT*h3&tl
zZ8?WubBo(KX}@Q9wIT25`dg6S<r*kvuKt7t#p~Pj*7a(zj<kmsE-KMK-6#o~-eZ7I
zdXC7L${Ys4Og&BGfp3u~ehjp~=Pw(;z`{6X1HdaZd7tOXq{hzfCmpip9l>!d+gwn4
zPQ*UZcj!t1`k|tsXBdQ!0U1mBEexCIBPvp_d@KC&Qfk%Lf6F$WeOM6pQ=JZ<ya{>z
zy5R2{^y$evtdfYw_dhE&SY;a<77<=6BSfh-Doh~<!2P;s;Xnwst{U!doMRkI&P-Dw
zsBw9i1f1vs3L%&h&*O)Og}KoNZhfz7kOFdHkZoat8u->6d{QP<KLxQ*XP3MXZXZwK
z%K+1WL;)_|u?*}XjCHbkdJ?~`obi>ybd2|3v-14bg2~u^>gS6f4S=CL<!`Hz>s1w5
zyD%!CKQ48$JWCxZ(tJ9A@YnUVqow8;=YG{U$V&#8#3qgPnq2QE@bPW(l)1tKpKL<e
z&Yhz>oRS4g{w~IO&Hr^pqKJkCHp@>7+^OrHTydgrC#FoA8+ihov?P->SdJm+^b5!t
zh9{Yz)a`egEY53G0_~o8U@duk4aHspx-pNW^H;JD^y4^`B1zx4H{NDT3)T;EqpH3R
z)S=VMvHK51xr$ayciyt+743uaf{`IgL(Wr-K2@v2)Q77f{gSQAZ{-9VS?{F`YC>+v
zJnUJYpO@YP5aehD(M~6Xu%WSCC4D4Bbr8=owQs~&U1I3d9bZo%fyh<a{qmfrWX9A-
zNh{S%8_nO_c`y#kOmrjVeo|`r&Ke8>EeRPO)9QXk|5tKCEnLFP23(pVi|ewV9fn#O
zF}e-Vj{P3l>k-$aT?vg25+rU&&^czQr~?Lt+!64ypC(_hh-QntKg(9A9nVtY3qF9G
zAQ~&3gX2t-u0PHvK({TVOKw$#BZaIp7JfR*9I!B3GF224pKHGI8eMZUmh=0<n<c6_
z<wG>~ln=kGQ?XwBStVpxdlUT$%#8bDPvuSeA7tB|_M10v(s_*<;9W;$M@KH1j-H7G
zTT5eM2IU)<WXu67tPC!tC{o^nz*Pnzu-)A41+<842D^gPO>&Ov5AA6o#m0G8s<xyG
zwhZgAe#E^H)-Aa)qab8H&z4NHfR&U{t7>I87MXPViJllq59}-u$DsPnqColoQS}#o
zO@ROUKfZv0q%@2a36T;386%YjNs*FLN;<}*VTjTx(j%lh1O!G1ii~a;EhRMs>DKSw
z=X}oZch3Dc+>iTy#dSR|Aes8w8tdN6oE!zuq&QkE@$Pf0mpa^t4ZXAq^~=jh`iGrU
zJFMAP2?4w__Aciw!MVQ+okGRqS$eqMlbI6cQ(2R);0GRZ0L}$(QwLhEotXj=;W$(_
zL9Xt)-AkJ81>s+EUi3W1&x;0_JhCukGrL(;OV+7vKJ-XBkSYuElk5PfEMWVvlylIA
zffi2Zy$6q166+hD%T>qm<}r~PG+65U#1yIZD2iAQU|uX-eSRf`N*;nm8ucYsdphqV
z)f<KXxL)yHUxIM}A{0!Mp7bI?ku#i_V4G<QXX}r&D9Yt*!)j-{ilqr6CG*qK?d<73
zTt`*<FQ;ZZ;?bAvTjoCsVQ+GM0%n-7oqnbJO{y$KdscRF@OXKb+9@uXfONlM1j8w^
zWkXm?)MuG+J?FDJY$W9adLK_~U{XT|i%yO2L(OW3)9DOUCyuhT&ZL_Z)SXGpRv5eW
zHRfu&1`RJ8_rO{~D%&Kt)^0^cWx1v<QFrEUz7nW=6S@a5%+ZZG=BWQxHd{JEh-9u4
z%5}pjM_4*XCowpeotg%H8S<g(JFOQXe)H7Hkw?FtBWyEuxZ?3-bjt;$;PGs?bB(Et
zvHeiCyTav8yNvfB-wTVDygXU$KaLFl^9$0`E3fR~YrZa_y~>~?XZ=YT#+tKs3Z~4F
z(1!F5#aB*(zz~m7fUMc^CgaBhqGp<!?0Um6H_H2-J_L+|nPBxBl)4=yiLU}Ov)EbB
z&%Z=@qhSEjA9ZX_v#cKGzLWQ(v6Z67G~$_+Vg+WA_E~4!*Lj9S5)r*od9`V-(Bn2j
zkG12-KsLGAXPp5#GkQ_kzO-X*>M+6W$==!d%YUS>%BULoSmriKT7Ovy&)m6x`xl~n
zE1eZTg=w0o=?Ccm@tzIwNeJ(!-c$mAhuiJhwF?zL1P|Z=PY4a?9ktsD0H-9#&LHpO
zuYYsk-kcYPjiFL#XV~hOMI*bAj_a*y8Y$M=1PMLV+O;Ir()ukXc0En(q9{l1eYp{3
z6=7+VmiIM@k<Ww7=g=*n?<*&)sgvEVfqB`>bOwB?Tm6x7lqA2S`>3;J`I^BhCcOfo
zc6_ZxVW=ipw*RZ8z$1j?=gY)|n^*CgZ056kkm6p&q)-*Ma>a64E^{6>Ge!RC^J1yE
zH%6yu9}l)s5<tbzweiXu@7FzwINH2ON#Fxk-6Tu#MLQts3x`Pr3{QZZ4ZZKT#Il!P
z4RxG_-bN_z@Q*|vl^|=(C9W&)@jAc$24Bko%+3a^`AkyUQqeNoOU?&6;71d?vEJI0
zJHg`)p;fr3LQ3$FXXt687W#$HI30pWcJeD_+r!~&A%>&)<RO$J&!k4II&x#CnZN#3
z(8}QV$<FJQwgJ|E9Xb_SEO`I;4WY5kVLHt8ar(3hj*e2{IO8PV>cYmf);fnupGUZ=
z>hSE}6{o6ndDz>}WP*3^ZGoO!Y<530s<?sD{!d9yubuHwg<y-f9fDF0(h>X&N>pbe
ziKV>^Ps|*TJaNn@K47Hc;b3r-k)w6SZ|`VCv^}s;4Y(KnVc$W-`+a|Mrd6Sb+3l!#
z{2#E$6q(7gPp7gH7^*J6y`az*?iSZs)YkBDL3PtP`)<5K;R^8bQ2CE;Vtg3Hy7tlU
zj&!<L%h5J$&HPY_#B%Hbxh>@bb)}Cc;V4RtI8pM3TYp38I%^vs(B=G>xg0(j%FmAQ
z=cZMkB^F~>s)MHE>KQ(dz-&CkL3dNFe#hVH20dBk04+@Yrbh@J?u#1yeJuhs$<lWO
z?A^_6SN?nGY|MgYebbX6@rR`FL;N7_c@|GUf;j2)wLqO*HKXx#`{s6hF7BtUp7#s6
zvulwk5`SJ${zgln0gHsp?uIL)K*1GHlRjj${5g)CirXN`G+BbB4k)50id^$I$64(q
zAh(nx#+5>uy;&g5jZ^9@!>xTwzNR?yb{zw|S@7hkzZmFv1#ZWoPCmTH*~97E*I^Tn
zWbXlZ{8WjU*xR~`VFbwfxARV))V_ZAEjVv4r<y-2p{+ee{bbrft6N^C^=He^z|L;V
z?y(*OQ7=~VV=49PRThtqu28X4A*}X~v&ptXOEzat-y&?&v0aFpv{Yj4uaC;`nz0S3
z?G`PV=(D;SVA5q^)Xd1*Ej=njlHkuhfyTUFv1Psc>QxY@qsojOprJB2w)ifs#WiCX
zdMy0KBHQ^ZH|OJyvL+2WcDab4#KdIU5fT<}GLCdvi>0e-pRKzZXU#i;Ru|C?U&0%v
z@&bQd;~xJ1^U+Gxv!vWP7_5=r-X;<4si~>!{5hAcwy*c4cp-n6`wdJV`~1CaXpc8?
zyiHMAo4<0c1IGI9TZou+FQ!B`DB-E#0a1uSz<4x(9JqUcHRm@*(VjF<xAzb)cBp4N
zOaDeIb_y!q!VV$Y&^12VG$_I5SQL?8sN~sT5D;+hkuH-CXrb-6&Cr~`aTxH(;n<(x
zSD{poE9fJ_4i;~|k;;-OrC_qKw2mVk4TTowi1BTXz&ewO@4*^9Zh(padZP22Y$_uU
zZGZt6L`TdZ@+BSo?^yA?s!}^)rzHIdry~ci{;4G+v26kZK3({i4iC9IcRI&nAg#9P
z@^&uRy`bpR_^ecG&Y^9biaO4gW1}jumxL?QbCbFnwqOWP-AL?z@gh=;=)p>P)<=Mg
zfl&P`D(?kK6N|$)7B>8p@473$6M0c+eT_?ytvt)$7mFz62+1G6`iyZfrwbS(iLZ2J
z3sL;?Spzc?ztz$5!@|m_5^t-ZHZ}R3)oOY&Y_I<Y!}yxx8{1}3CzZ8L(1EP=)3ORI
z^z_tb!FN<L*P88h=0h4ROWxn*_BRSi@>!p`=1-5d9V{GH4!3E+p9V^{k$}i{YL4U0
zR7jQcEL5YGI(kEa0gz2^&F1BJwtCPV!GKoV9B}fCe;bUs?e86j4*b@_;>VZtB=-Bq
zpq&_<0Zh&s;?jK8QUiCgLy`&jvKuljfAwB6_~7>=zJhK<S)=M|hfaoJqkeTxpvknM
zkxBVZXD90V>g&IN!*6+cWB+@ewWac!Fh~7tN3KJ|Z;;Fpa_yNRji_|~W?NkkkMh9r
zgrbX5S55&SiZFq<lj7VF`?~0RmvVnz?L4K76&+WE-2&Zu8v0VgfXjsM#(g`5k$IC$
z@4+N~TtixZyI`XYir=a~ZV_OhUHg2evxqO5Vew1_5NMg7S;B_u5>V5Ia~g43L%QU2
z1X1+yJOE_@66#^uw{GTI<{`fzBoyiw-BrZQ8W#7_6I|2c)dV_bIY;Q{Az&OcV`bf(
z4?Qx0ap&9t3sm>qrSXXasDiu<o@vYyKLl8LOji0eLbg+A*L2W^BAyw;F9kXyQ5vxQ
z(lac&x<%|l#pp6~k1<T)k-T@TaLU^<XLo$LL$Z4h?v&lm%WWrRIO7r7Yu|CnCsG<(
z`*u|yEa6t0cLk?kuF<58H|gOz%dpup13q<j4AdHd$NHaMI4jR_+2CcqLrBc&=uRQk
zJO&RoYrGPf6;l}8zMEKuDbY9>-tv`JEv)#m6*wyA{R3HGVFtKs@r(v<eV{_i*^Jx3
zjkn!g0Qj;O-Sq}~9A)eNLA(`qF7nGMjlNfsZuR)beOdBEg5E(eY=#`fZQetpf#0A{
zx@lbl3p$K&+rA*x)$Vo6k}dQ>tOos<>BkcaqF?UWYmtpw_bWwD-)x$O9nb{y_}t!z
z@>|aUZmqWHJH_!o^eq_?MeLhFee8v6n@WYT%%JH(%B#1e0fP}~g`alXUceuoNB~d#
z_H=vRx2%Ez=4gHnai)eG$6fO}9mmo7Lx=!=KQUvXxL?xQJg66gf&1;XY!<e(%yv+d
z_1^~&iZgOUWFWT`@V)o{iz(+MWL*M&86;(P=Vr_J@G10ZhTsYxivsU8_EX$@t&Z~=
zx1QgEJ6+1W-JgOgXR>O1<wqublmqDPaCZU2p4ptf!rcaSh6xt-vZ(r-S8$YElGVLB
zghs{-SSFC3mnlgW^s}b{K3YplWea_1c1!D?wBXiGS_3K{f}F?O2o@sBy$8d(2Yzq2
zWDm@=ly;L%5}1HX6g^vY8aKRhj3#P-5|~4F;)KV!vYm&3K@i2ExPr~8H8HA<-F$hY
z)6A1idf%ozE76f1y)oZAK(2#~8<8Mt2Uf)xLo>EVsP=<G1Sj1!e#W%D7Q~b^buNxg
z?Fc_wM`aqq-hL5jGrnIU$)_Mtg{Wfistqn1W}uWK>*2(jgrrpjm;`0q$U)4xdssaB
z&88$-CN9##!H{q=B5Tk)9x51E`|nLs52>~0r!B&h5+~L%Le`8c$EOr?CkuK%#qoVj
zkDN(UQ<=jr1nKEgpQk&i<H%7ZUfJMKf9=#_w6PXVN%j<rR-VxqM(0QS#P!>oC-o3E
zd&NOEMUIu=sT-R=8FhkL++!FVrZFvAP6cBdk`P=lLUEd5g%?$8Ru;B+o?thuU@Z>P
ztpDfDo~iW_abRwSNM6d^RJ5h?7++Ndd+Z=0<z%Xx`Q;M<PwV-yh)Liq9%+@xUFF;*
znXzJ;*}MGbj!zY)MEDmvs3-G4zBCeMyax2K%v%-;ir;8igUJuQP`3~jlu~3pYuNlI
z<9z!&2|e0*+;DchMX5ZsDyaowujS*@gob%Oq2iEDZ-d`^t^B|J@9m_7x#}FIQYS~2
zmVPam%M}?*>UHbd)6TT%>E8?XsG<y~T4wOrI3Neb${adS@~~`Q)rr;WnXVwEnPg&o
zg6zEm^skNAF{Ey{zLB$;okEwX9=pstl29F@*sxu9q#UBcgi<kO*!7?e^<RY?&|z_q
zaUKIzC)<x?iF9JO%4Q?*l^ktVP#ihB9*8tTK|7bfoowVs`&k%ZhYd(sM9I!ez8ySq
z6{egUx=1Pc0%t{pJ+e#Yfm`jHKabUdESvUoR(HlM361+rw{)X~+Eh0m=CAPr=o8<G
zi@sO+MSSx0vy=3Sb<VKeQ>X<apy*XQdaBc6zmM%j5!4Exb_``wRkK}t8%1(8Khhx|
z+C;oYT?y~7*zEY08T#N&DkzHlfgsA!IJroBKC{u^S%O>O&Uo>*S<AWs0XREm)dc@{
zZ~+&1fegcco2{@3eKzi1+!?%lhH_$is-g(jt-OO;BXrp-qicU3;$Aw=Ox|xm7H@XF
z$NZhH!{vPz=@sVZHEY+#g)UkB=b9a%*0%gKED~UEIG(OULs1S?>EtO5K8V@ReB#le
z-wKfv#LO`CB5RTNCCd$GyU~US^4pzh_z;ac?^F<(%jY>1yeH{r%X1tms2O$Fea`#^
zf6LIT!0ndG>R*~cd>=v0WzC{F8A@OgVg|0<ls%Pz;e)Ps@vhw*zx=)&V%CXSfUT`3
z!NtQL)4f-uFNFrdRhMvlhbB~{lM3bzm<5+$sP9g5*)*I^`Lkaxn0wbP{3+K857?DG
z9#tBjo_<u$qNjIT^KS(QVJP;gEdF2Oe`^SdEj)I1cRyyiVfviFMR2e6NP-ybHAS|v
zbrh+s=Xd1+Rg&vTs{+0>;$}wxb!FEN6{)^HXf5sS4>}|;Zr%qBdSu2Ckm9k630kmr
zX?7|<+XI-%Wvj(+0XCJwsf)&eW;Z#iYN^db7OzoC2g{rQ-b`DO%t(~6m0DGqIWGG?
z{Ia$}EF4S?P;}yCr;VBX^!U%g@6=*)UkM+!Gt%n<aHZwQJ%>9~f&i**$E9=XiuNTo
zb+Im~A7>+bIOYzkeM#}^t@8pZ%i;1jU9h!O_&*OY`^3+Y7ux{(S|VeLGJn6(hj$T-
z$X2NF2H9t-I_WjC(g%9^XLZM9a;|kU!@x4IziHTK*Q2rZh&B9jnWnVQGy-l%91Hrh
zbqNJMBFG`tO-H+7;Fntuw6I#!Po`98fqbD=D*3@&am*eL2Ur7YxjGWm^GU%8xtAUf
zM*MotlQG(hFevpo6~^FUXMxT1*gdg;mt(wrqG;}MX~*iWoR?qqr4o@_xnYv;WXO8I
zLr3-wWLM-T@bUeEY<VOd;}YD|TRXngV$@oqY0btsma-PW)4#U#b*YZ<$C(F}D*i(p
zmHl)h#F;;8v7zRG7DL#gzqjtuW7M#&vV{hh<kj@ELG!~vAV{Y=5k~#+YTS><f=%ka
z<4-yb@%SqlVY{hRuC-q=lLi~f4*WKO=RlHsM@|#K)$<3B&-^AW`Dr-m*Db)cEWQ*C
zQ#HQ{=>liY;|IXLzX2}nv-@)+GtTb+M)~b#A4*kQSXdk`?gtyYw=ZANA4fm`4@y|n
z^Z7)z%zu~rTWWlm^8B408i1tFp80HhKocaE<#ni#ioVL&pUEiV$k(H}q8-Ys655nA
zO4eE)y3nHgASj>^<w~>`<$+B6qT%xd-wAh)LR_p1+((28A!`B57MnbL)@Rwa{)>fy
zDgg1fw(V}#uNv&Lz?P=KG=LBNRf5E_+-0Jc>j3$qP#M&ZM<|o`tO>GGnyO*~7>N)7
z(Uw?Mo?KZ>EX~<l8NRIWa4j+iR(5B<?3%~Mh=em`XSeaIvcx;+h{+zofd$rh=(q<k
zr2gcmF?3l?O9bKeXoEq|vtydqA4E@bH7M+$=fyT#UiuT?$-Sq9S=&HWq-O@wCz0F+
zyh324smMM2_G+mHP%h@UyD86k71Q~h_w;z2TJ@dvHYW|@PT-VJ?SA-<B9n)UCpI)(
zh7~M+O5pF&Qi^#nzv180v_^lm_gG9SjvI*bD+azFu8`htYiCHl?Ay;k#6KsqwErO;
zdXGQ1!CiQ9>ZgJtY@9>G*j9*QxQb==d)W;6<o$^t7l5<u@;5l|P}&U^x*-ebYTK*b
zRhZ^ZbX`&1{KL~Y`9ewz|K*}E)ak_wX%TrVNY*@((xbgoT@iVIy0Ka*Y^s?dv>Nao
z?Hubd4!j#`EFKzKk^J-HhrMl%G`M@&L~Nh$s%aQ&0yQ_PCI>b1h4BIf?j^m3#IIH!
z+p>GI2<F`SNcW&~bBOgQx`OfcfH^SeL>_^c2l0AdJ>P<Nv-|EiBFMgEyQfKgy%;*B
z{oDHImuN^!^MUTDpwa*N7S&y6YP$>)4J0Kc9bH%xsG2?A(Ws}UXfzyVCZ+3wZ_-1;
zlUF5o-$a>fu1&ItEKl0}Xk+RHr1zl;NIfE66t&J8;Dt`a`2Q#yk;HmeOAqiqJexuo
zKqC8{I}}k1Kg+t-?PziGqDlMtTExu2%oXSzU#zFvgIypUUE}=Lbd=KQUciT<`;25*
z8m@O&RUO}o7*_os6LrV(Vl}aGp82J+%UmUpv?#HKkOVEn+Ibv2I2-WpKJ_Zaw(oY3
zaK|L_b+>0hXlv*lL-`@l@n@~Q+JHF8KaUuS`s|76UF7XUl8v&mrX!wdY^L(Dd%tWv
zZZ^oDkqz;#1lZ3nJvjI{)Y=hfHY~YbGGdC;=wbKa`4efV9paL4CN2u|C0myJTiOk~
z6;zw_+^UH*kL31`1Cyql`2my|UIiC9Tv^?xDrH2D;}=!`Sjeaqlam|dJ@x{rBWq&B
zV^O^E#CA_ZjF>OEI9vbekc5GK=%%*XvAYEpwjQWIqE{1o?8F+;CYMN~w_yGOpzOr5
zs%Y*h`PEajmkj0DNF&v%dUx`@jqat0tN~A&#itUH<aCJ%&xK4r)a#gG&mS(ElUnSr
zQ1=$EzT-PwJ|KoZf4vE3sB|2c&~SAKL^+%4S@Kjr!%7jhP0Eut1Hg%bcLga>zrr4`
z{WgAb`7(>r0LTN!qr~PISM77`hMz5aFtQ4u<T;m#AIp-e)<1hCQ8M>8G2V6bk*F}d
zYO&-hg_^_V-*Dx}@||BsOV)j~76bn66D7K7Avd@X<o-)^*jvN@+%Ai}a{TB&^{0xT
z%9Q>azm{M*kS3aj#H&UJq@|?9{ch0g)Qn#0XO-BRK;hmZDS0$bCi2h{BQN!t$Zo6P
zG_12AF8c0STF%7G1jlO9o$(<Lkui^zg_p+rLZYSYP;bdjfDp9yg%S|tqNOkx$~5Ws
zA~>!>L!~a@?FfrsSI|1y&Dz4<lX&YOX}vpY1jF*tT5Xqhai|k|6QU_uX1;otdcLk;
z&Y^1FMR|uCD5SVg;1I-)qf)JJ{%eId_$j%fxBz;a8de*X6%EKV3xe4^4E0>_nv0w5
zcV0y1`1d1=i7P>p+ok;rD<BZjj<1TJo`9}D2l=GrXp_H#-#+m0AT!v4++47%UWHpY
z;UKTC`=cX-bMO3pia|>=KyzcObC@PiRufP_TtRuNFDjV7(lv?xD321Wu6@M*?+D&>
z<ej5g=+ZOzJh_@RUM)@*mAfBL@b8x4Jg4NW_u9hcsx5Ov%op9KUjsV(g1bIRQBklm
zvWt1Htf>A7{X+1eEe-y4#9e)=!|poCd_kTO>BNO<W71~V3S(Cbg?`yTRVFL6sGs5q
z(6r5)C#dz|k~E!BEyP?vo)ft**onChQGjkpf<{~kU3u5(txgW}BDS=ERWpLaDEr}t
z9Gn_vdU8dIWfAy3(I(lspzlCUB5)Pmu19oG|1Be&E(MbeQ^C>&#NmH^k3OkEc8ba1
zni(%{qVedXgV7sWV3S|=KlARo`SLXh*N&nTvuIbcg?do;;qEF+=~@`wgh5tN;xMyG
z|L5v7LZnTpF4jpeSCo7jFG(>Ma+gtL5>a|yyOH6LlT@#}_^Ue20Wm6Q`F|g`le4pd
z@uIudQNtCJsWRSMdWUqB+@4G;t6emYb{~u@e!%Fb;8j7sZ{0yVcd0S{c9HfF6b+Z!
zBa1{p@}<_K@q9oUo*SkDR#8Zy0w(Fm?kz)ZlS+|qS3^MJRCFmt-m6|2@;{3rLHn-@
z6A=sPliJllhoSB4YA*S+M1HjI%_3wP#b!{3Tg6%T-zzqnQcNR0g=7I}h`zy$JS2hq
zl=m<XAS84a8h5pfha!N5kIeXW&`5K>JolC**7y>a&Uggm{E<UOH!{HdPuz<n+}8D~
zWef_8B8tDoZ0FWPFq+URQUGq{VSRaV3Ywkww7}Cfp1;;xm?MfrSACn;HmXNgaI6#f
z1@HIqLqv@?mf6%H1p2i;b#=|W0hcht100Jg_yYox{2KF;i~Ggq12r1O_&UH_pe~D*
zdyuKmQ-d#pxU4IY9$*@724wBkF&s3+)}9_t1zVd5YyLiywfkuhnXdF_GMNhf5DO5f
zei?ft<h7lEQ5{ehw6lo4tHtWDr}5s^_i^PKf7yafAwIT_D~QEf_4gCB7wpnt8~w~!
zq<rFp+IrK{qkp5SDwbk>^|Kv?>3oc+c5uP!)e&WCx$6~$F*Lwyap+Qp8hY9QW2%wn
z5TsDoCWWl-{{HPTf@z~iSp7bs1-2Xd#lu@6DRQ{iw?ijPHYcGQgrC%UZ^X`c9xhst
zTJSH{CSk2vu1?9SB;a|C<lbnb8xH?iBd1O@aG**dN2QPf&&IrqL%zXRHTCb#VurmF
z{v+hso+z<9yZiIwd8jVtfZG2@#jzVD8QEo>fCTctIkp%NWFK#uiSd+6$rOPJvc28c
z%;A9sc2;@>J(2f#1hn+<FmZ+5NV_Ev&=m%CBA@8h_Q&&(xe)*65EQITBAw55!l22c
z&I1j?vOE@TjH|4;UxH0}>6rKqz!DMFPxd=EaKY*M*dJ8ql47h?x22t9ijtCloG`7U
zqLaWE=^F98cdeuVaAoI(GJij<$bG@j1RI%&L1|D@JPqKz*mpTqKyH~2L()TOKTQxH
z`OcOnB@w>)MuICLE57w$U~=vusZx~?Ayrv|J9-_lxS=!Ih2QrY^%~<tjF|qR<T3H7
zkWmx>egV8~=4#>0T4UqSf*hqzm*a>#g);tP7r-t6kT|mX!^dG)_v26cbp$b&zhi|E
z9U!)fZkil9;iXv8#k~=cX@RP55kx<$B86lmm-e7j*b<5%Vkg6&xli;G80wa`yXOO+
z<dX#nk0u9wh^|4(En7>VFkX0*QUlFO_ploH8#~J&SuUnUS!9SwGHrZKO%rlJ&h~vk
z)luc}`)Yv=BsX_E6sCPCqBv%~`fWXprOw{FT`fjrOmM75CX3=yZ(pRMIk*Ps#me64
zGmXsR-%F@l{SLaBA1^6MXzrk3Ud34W#O%#VpI+BqY_|PfI1B}eXeufwM&HO&Q4r14
z{o{t`Z%o}j{;_|uw06|!-=56-g%6g|Vjt|$`Ov8HtH>hLt`_{;MyS_m?UGO6q1f4A
zK4TNfxi5!WVdL6g;%`*P8bpD7NE!T~l5864sPNjWV=jAmQPanXfphY@Tlf0iQh4hK
z`+NwO1C;`Qb_`iw;HaR^O9lcTqWk%flEVKq0_6nxMa#uTmKxHG5G+!IoTy(WKMR9|
zGkSws3;?QsNYc5r<8U;~j(U!b&7^KX#p~EhI}|8y#sg4Z!w;Cs8;6TNYUBER=#RKu
zkK+1(ua$VT)9$#;rkzoYJuzrFhF4bbJ(eQAlr(wd@dSkzlvc=oaqG$27ibg{Ak4f|
zT2m?Pq2W<1F~0DG<sEmVC`07xW%-=TkAIYr`U>&!SpYF7<cf<{*sGm%n%f9>K(Z`w
zSzXtbjj{60^6x=;DN$61a$&RS?}t0qK-r5-33IKmsiPBQ^W=}BW=M&8;PFeu8X&a?
z77>_qit`@_6d-L+@SpAgIJ4zE-46WOLO3MY8S$#vai@T?*Lz7z<OK3a@)!TJG~}Wk
zQEYVjT>~VLvheQJ`%9pHr7@d(rEVT&^^<9%q~n}$k7Qnpd!7?4-M!HghM4E8%HJ|n
zT^|}iL*JU-&bPu&(%U4g6yN<_OTT_ngum}PX{~+lGZ#~*2K-b=u6XH1;GY19D>vHM
zCNI~L?x^U<F*^}Uce}o1v05tMk@QbE{lOrdjqxnRzD5pQI-!08!$aUZP;d5HEYSWk
z^Yu}dSG8y5t5cpUW-dyYW@c}?N7127&pTBJGZF0T{s-HyY#7}TT+P|%N#H?LEHL2f
zL2Wm=lsZN1;IMNJjM}=n>3e6H+`#93;0Ms{3%$3K2IwDc!P_a27iza^Aw<6}r0Wgx
z%lwD@PJ}|kB@oM|4O_1&elpb?vAp<~-sfKjkAX9Z9rcKrCg0a&F%Q1@Wu!$#NkP?j
z69Bc1?2(qvli5ls`49Bye~x8gu}=1Vaz-KMlrNf2<Ii;EiVpH2v5IscB~YV~FIRiN
zQ@J*)22w)33CnPYKP~~q9H|VQal%{{0bj50JYT8Z!Fv;IHCj2&_9nk6X;&&u+kU*&
zz(=azNOB%8yL@6~2zjPkc7Ph?eSrH^1e7dmhV)1@$89*~IfnY`SW^NDZ5+IGjpzm!
z-V%Qs!%B|a(GwM)VKCLq;$StVTqEKkEOz)6yFvpfm<fd#IO&kAnY;Z<AO<oN2l++O
zW|#e_vm;uXc48E04gS-`p3EjCj~Jbj|DKJ#9;JVK)fbl9!7Ogg@qnASx&x)}{gRhB
zrA`;9m;-=B?BNbv>m_YZddH6)rr7sI7k@me_<e?EcihWcV_AEYuTO*spNRD1c=lC$
z^V0@PU1pF5RASBGp<R9JkD_yssi^}%fzxgIsZDk*<OJ<GRIl+JtlL&x)dSq0)k2Fq
z<&U2kT#Z1}0WIkXE>Lb<gcgnZ`{ab6NY*6D?HJq`GvHp#JW>|fgxR1Vm(p19oe5=S
zbCTXO(AwvKHb6fB_$$QvGKZWLaSA6(GlG?Aj+e516(?y^dS8{<65!EoxtoLpon3wm
zDZV5A;`U405#B|7YW2<#!#=Sl{$RBlQGK=LYYjz?Sv<!+#c3*UZXYP`Y;|4jx^Eht
zoKj_{104pbuNt4@L;z_gTdjJ)--UB{lOLPjd#C_E=hUF+Fppm+)s9pnddm(vI}+G>
z@0UBj#eyd%3zl;B^t(gIO9i6~;8O~6NTcCYytwQ=_5ZdKmHz*%L@O2L<(lJ$Jd-ms
zuR~s(O+C|P_U%fMBy#~0sax^9e)?<~O5O5N$ee>{T9K!Mr6`o$sss?RJc_o73W!~!
zC1pghE=RZc*o%IiQ=L@lN7ez-Q4bUa;q(SPOFR0%CjAR~m_v#OJtTW1N*jA1e&hH@
zn~Oi(C%#gvw_U0R=zj7?E}+7jlaARACZ)MIcqW%m*Vp{NLaiOJ24Mj^=@~~NC6@Hu
zRTAIENE(2etMM-fg-(Q7jL}B0n@nWe^uiVwMPqB}9SFd>EsILVm_d<-GzpUxoo8yK
zSiPh}qKzt>ygnkH{G9ok=<g*|AzNH<gWP0#cGt25mo)WdzTEPo7S%n6S5@0)u^!zg
zXrFlbio08(jO^6Tqo|3Br<=@Dcbxg_nR-b18FdJndRl(RF4G9^eLIo0H`#(4Xm|+x
z=k(}zlJ9oOcBK4FiKI<;*5LZvIrVezV()aa{z<LYs0O~a-|H20wV1t}_0D=>)8Q>_
z3O&N$;LPIjl`lygYky4)39A(V@XR|}xK_!@(x1X&W#7Gz8XLE6>GjsX+l1gUe1xnP
z6qXVF-e4oh^9es|VQ_q!i|4SeyxwxhjA+KO8=a0s&f`jBl<-y5SuB6ff0$wom*Hba
zRy@!yiY){=mLq`r9R1w59?=f+Y?+2HrUA|t4xA7VEWb9WGBjHpZKNPcJ#tNpY$}da
zju;XA{<s)Cqm-DU_b_dgr}x7*vNGbe|Ndj$AUJ1D(w<A$?C@OlH*qx7guVTbo!dXE
z!aw|svzQ<)`_JKwa%*0JWGI9b>d<U!$oTAiG@X!4EE$H5^0<yC-Y=tJ1oIfAgAGBK
zg#c0}_G$_?mRY!Z;Y`x~^2ZWX)CRL&<a>8_0z4_bKNfZvAHK|ONw_{Xz?82JgSr9X
z%{grin#%1$<Hz>Gb;qL7vrHfwc>c}_R$F82q#wz+zCo4_UH~kQ)vmsGZwV+^yJrK;
zOMhgQufOct+o`CWHpu;8+Q|82Rj3y7W$VD^t;d2`86ZLtgm#QXc*=P<tT&@%GL5&)
zRR<G*ts+q<Xp#9#QYqFoza0Uv{^#SvbtV=dn?@_~VMc&UOisOWI;0oyWk=TRM*gqw
z+?)CCe?iSRIpV{-B+DyMW}iL%1a}<ztyXvQ^7J=kZ?hD;lpZZ`NVBu=e5T;b&aV0b
zYpt~4@L26g6^+G<>23P75Bp@#R+b;832c66k))|j*2JNTyuja0TW2pl{V65XYegyx
zP1ktp+w<^TYl;#{uPUzw(B-lJd<E)vr}pgG`7Oko!H(Uv=i`A6CGphQjWdaRTo&`E
zRX+hh>uEgKGYXRdQ<8!ram99prIHa&=iv{22~#dSlJu`-1Io1KyX*QRNIK3vQ2wbf
zUq4klY@Hq^M@QCvc{uM}d4n<%qG+3NuC0{*U8miCQ6@>!VRUV!eElLRcqWM<UTuRO
zMse)$_h^H%wq(RHfa>6S*+JH)Z4*7U2mU5(BIubRZ)1|C*pH5%Xj<?Z@CX`<?=Z>n
ze^Wvj+0UFku5tO_BI;w#(Tk%^%KoB!)LGfHlHG(UfxE<#c%iP8uvV8mHSjKvod)<t
z1+gmANcYN<e(S)$4de^Al%F})KD`kJYZj&)`kO!TyPsw2m{#rCorw5FGnsXCA8cYc
zt&S4OnAm_v&;pO5@jV=bL}3dl&oslq&udJK^|*kBIoHXdZ;NjQqqy4C0W1J)?E7KA
zXJWYV;ddBY@FMNR2}7D@X3jt(I?$k|b$-!!33l5AOb>vxF_DanshMYTf7N<FaqDb)
z!LC|_WBc7yT8yN-V)iE6b)|X5<G_^c0ZESL*_)1!-n)0L`4I~4DI9mHy?dy7JhcJE
z7ydc4C@e-3q~gc9*UjeZAkxB1hv);Yz5c`X+kw4z^s0xF)Ft*>LT+Qbn@(gZK@!pl
zoYNnuRfpYne1*ct`&$7-@{Elt1r|xf;+t8>&217+ns&_27Ca-b9NJKGgTH(q2r&di
zef(vUfbknUVJ_JoiK_JbmN*RV9L>wE7BqM50F+F6OiDA0tNQKir+%6k(HMlj>M%*`
z+-@_RaWiI0Dkd|+CpPv48B7Q>xXCM`_r$^i36iLd+Z|51h!Ao2%j?L9(637odop2p
z-dyavgtDqR{+o35-1(N$C4JP1Z>zbP>#%c=Ixe;4xnD;eHL7*(p8P?xWlVK0F@8aQ
zIC>LWHV+#&0n4^F7ZW>UJB(6>TOa=w5Rb@hDk#J;Cxc>G-*iFKPW#2fHzW8pAu|Z%
zc7+SIN&Db)3^hzmwYjv(ZSlVIG^|z{2D=xxnQ4m?F$O=}N~!uUuB76W{)?Ay%TKH3
z|3@)KEBxZiQ$kE^m8zbRa76An*JYvBT0h22B9f$-c}8P}9tGx~y-uNZwsr+Q)z1>n
z(un{(qMs`{;>kmC$?^Y@&_{=i+met&{1?PLzpvBNmJFkbPl>ijzB*9JY7w5+JdGVe
zs^nn=go2W2J<4R>T;4wpiUMQEtYVIrgj7E8f#}<sj>oc&nc(t;mo$G~V5}gc03U$C
zq|87P1CpN<dqL7!vB)d7HU#_AYB&kRJ0i9-0Lth%qI1ZrV^x&Obzl+X7p#ElXU)DS
z&N&kR0J@Bh(d8MxUlSo=1686vFA6`okwx4~(IYZk%{=b7yR4xz$1YHjC2}@Sq4oNz
z-9Z)e&+?M)7Jwj|UkI7FBlWL7l{vlc(h<<ltxvMMPDfjJny!FWosc!=MGnmE2;Gb;
z(zyxW_%oZAfLvCsiBdiPNxKyqe7^Ktpre##bWQEm9VlMuFP4%vGCZP4?{8@doR!S+
z%;TG<zoW}WF&c42hZS08m0~zKI7FLcPG9BR!f>puDb#vfhpCyWvEb8>UzulDzBd?o
zYgnm*VjhOcF>crDlN~gx?o@a#%wlv&G_ZxHZ}%7$Y)w7o_eh$AuEK9r+FK3^*En(~
zoR_?_!hF#cA5kA9sK<_G*v{RoEousJSp$g<<Ufxp?<~$R;gb!Wt+`G}SgvFfWAL=g
z1MuvJN~`TYvk}xSBGx-PpCQ1RKE(HhY>xG!cylZF8gozS6PfH%N&meCX-9Wl9fiM&
z;D6K=|8xH3ZROz>7qbKT54-7fdq<<Nwxq_yj=83J;jHl=1>P=qtGony$=<UotXWC#
zD)O@b3R%YUo-UhGM1G3=JNV-XBBJ#2@EVn41I{b0EAEmdUabmd%Nx=Scs463Ji7|d
z0&SA+-f*#bics2DMam}<pIPZALkM7r(w2xNs=B?ZNz8rP__DQm-eUZ(?M*cP-Puk$
zB@GuCUn-m$oqs)qJxK?EsmpnyCC;UMQDKl0hQ02iM#!trdu&btrV+F6@h%;!1X{4^
zNl_n+O0yDY2ZW=B(Mnh5xF3ggKh6>F!8A-PYpFO?Sc~HrD(8UMZBSTDXQOoCH`z^^
z_M2?U`XPabbEo0g7=J%_?NOVA{9n}L2Tp0G?K0?y;NK6}rL-PTH>!cD!elx>(27WE
z#dQncx*or%rMD5KXyk*mg`0W9c#v+cIkHkEx0)%?rxTV~Iz5QTw-XO@fb*Mwqqou}
zw7JMfW9iKRpQw3FnFd)-rqd^A&)lN9U)Y@fLObC;rP<5`h$jvaF$ZWS)C43+mW+T9
z!v06rd(ZMXs+DqXZTk!JRwp+O=S;P8j1h4YF919Zln4^CoYiK4nRUfi2;{i2airyj
zq5*W7wh0KdDje8Z$RMZfmCP|Aft@mk>?O#3WZ<L#?3(CSYOy)(j>+Z?9T9KE9m62p
zh@AJQE5vzI3$-y8POzFIg<k8wGAAwl>D3(?1oS)ja*zbQnse;kSCrPjdw3^93~4JY
zcfK5U8mcoY_~L(0RBuR>?~hLzTM28W3TXs$KP<>!011PkIR#+=shzP-%0J;1x1?wj
zwZn#)VJI`_V@X$|Ca0%CpW_imE^_iGg@|qd>~QA|Fpf%LnN9Hs*Q`_0NeuE-cwGWv
zO;iqZ)8ciJUcuboAU=~Y%j>+3b{>4p033-%>l2N*BDo72K4|tdE(Z%H7GaT+S}dW+
zBeAsmBo88yvFgdh^7iJi>6=6X-^F-Ck&+)Oz-;)*jdhKv*EAr8RFR8aip`k*ZHdvW
z1iLO~(Ll66Q~q@+k#ECsuZo*a2J84uvnZ$G`M4NSwHd)(JhEW!SIo6wRLHx^3|h6f
zUSSr4pzd0;N%jDq5$hCarwt|c`?dG^tT|(=RlbO<{k07HK!X}*LXMh<FOdIvf-`HF
zsYu=u0QK&`>yC|<7aLk^aKdgiDp?fYR2K+=m9>W|;<va4*9H75?QHI73#^#u2WJjf
zR3R1J{MjpI*sPlGZa*kT(0WoKlKeW;{CJ<!s(*J5GD+rlf{kl5?Qo~<kSX&D&*Ks~
zSQE81E2>+VnHZfzCVhX<0;4rp9I9$-W<*WTbNSgkU^HbBY8G&V)n~v;v*TItF{Bs#
z!gKaR^sXGaODw=&7V8o0=F6^t{3^uDR*ym=CgZ?j&!;29sf9d<z*aXwq4*EpzL>^C
zV0RQd)xFKG#})xpFdYQpqamK^hL5iGe;A<9TqTWDS31FERNs7R_rZ%l-?YlrH{w7w
zQg$8#tI0|+C!bHo7faUATYIhhZ+=m}wfrx|;TFh{9ZAktx|2ayI(z%Swfe`oA_hrt
zINbL49F525(yTDMYt{G{+luaIY9-Pt%{qjJkc&9~^=+5!1ZZzkxRXE?(WerL(c_qZ
zXJ_h@)m;+R4CYp_7i8#{h2m!X8gJDXq?_Xytya<Q!K2loy96k{wZ`SkU;dk@fm`Wo
z6C;`u0d<c-Vb;x$uZP1Tm(Txzq4TS|8Nb+ko!tgQT)71ssL|c7N;`zSFEP*#&BtfD
zBINf%-#`o1j*J;e&L-c2I6$W`c|{aJo_u@@rD+2)7~<3Q6j5H#5CCvVe#k$7+27xQ
zOBvt=qSQ--0WZJ5s@#TFbOkgV!Z(-tJnK!BZn3?uXKc;ups!r^Z)O@EGGXveQ-N_b
zxlPvP=IsSvJ|CAovq$u@yzURPi1Q}uQ7J}@d3|e15^8L+7BT@Bl0FPi7UY4TF7q-Z
z7yXR(1Ine>)rBANucQY|Tx_`^982!w-r2Ajn;7|9Z3rpt4!MNU<k6jJyNa>B@^4Rl
zWY+%UMx>&f-A?3D=Q)vWKty%hLP$XSyp3brw@QF}^NfjT*iE0b$QzsFq{oDfV+~34
zuW7!)dqH0SMWOu*0dF_~?=^@w+TeqQ{fB5TZMlad6g{AH9PXvHNIozMc;I;o8X;Q5
zZhYfK2FR|E*;c=P_#Httn!4y-s8dgM3AdPo*U@nb{c*8NxLVDKDo8YR?x`=`U|;L^
z^Bq25vwOPbjh7>kd{bKPoIrI<DsUi0;kj1-e@%ljf`XfIvS5e7^gBN<r<*?PzqKs;
zKU!9Wdu(3bBjDjrU7t+i-;AURXo<dwFRz5NT5!b{Kpuq6lkk&uaok4bCgl@T9IT{p
zglxq^CSZ68UlH-!dbf$zs<$0IYf+cB(3zgg5J-#c56UYG#VB#I*NV&0VSP;Tyzvv|
z!27rXX^$j^drni<+0L%AE?hkr5C6xmMn?6qP&e35(#x{OA6Js7;yy64A^cev=|{(p
zfb@QZ{v9rAT;9Ie`B<{IGm*J1n^1E;MK^#E1-NXKCh}C-CZIGyq^fvNV3`3E)#H$O
zI{S}+`-&CA_9P64PJv=e(1jP`z-i)YrJuJx-VeQ#nn^O#CH?%#<TPe2KcL7fy!!pE
zolcE*RfE%Jqpy$VP}RVVY+@y|D$5Azcbd|~$fprha@J?|k~j}We-8~v*#N<Z`=ox@
z!}~*iI<Y;U#eT`j67UsU#ECiRFrIdpU-2R8J(AS7AH=ayzFk_q%?s&Qf7bKTNpCSi
zZ174tNLU*lSYnk^%Bpc6&gIolQO73&K_jNAJ5!MGin%GnNzeMZm|5qd7O<j6zl6n4
zC0dK)FGSICcjYP4D{zCj!a|3QG-}}ad%1cD#(QyRX)HIv(IixSZ{6mSyxn~x>1abv
zG*HZbay0Afcn-;0w8Py$w@F0K_V*Fvrt{uEYjaOBHAWmp)SRsfq>|hC7_OjcBG$P<
z&(=pL>bi>rWl20APRlxuP-$gzHbRa~`~AD*uooWehiy|1=UXD9ccZ;`Wk#C%k%DUD
zwj()mg8xAfUi!CJ=NVT<|Epl8r$4rNMsRqr|4+oXg5_f1(GYY$K{=F~{rAUIN5SNM
z+Lha%^nyZ~5MQ|{?QO>OL|Yb#n^LJ}_Xt4&BFGvI;%1huFE|r2cGSI*TW{xBd3sEr
zESH5o5U5THktJHb^M_ZZ+AFMMzp`7gPIi>&b{=9k)0p6|6)bZhrjpY-a*_MZ=I8k#
z|9O8G*aU30{AKlT>58P%&x*%2l3jp7p|3rU0o^8a5}S+9?)PNr6gHP`RYw;Pmq3S`
zgp`4_9-@<{tA9s0K9HV0h1TZIh<JO(0i!#mo!V()!FpJ_Bdu()Ucw&qLE>G!%Pkd?
zVw`4sxlmvqS*dqmM6j5xPTq&I5Ohp+kubj<{s$ewJotCro5Vu6eY%zEW_tX><FKw{
z3$0lU5lB?oaY`-yIuUB5EEe?BPfb%aF%Dx1m`TQ1;uiCd)*lyl$=BJFPEy<c25R!%
z?;LLVK8{-w(tfBmY-}Ni2@|AJ$-B2~eD*4E1LeuC6>~zLJ@wO*vam4u(oA!ZWv`xq
zO=oeO2bgtgCne(ptM-)Qz$X7<iO>27@Ff_akcD%DlgrGBDp6WS7yq|?%skJ#mR$$v
z;D!pRbni@yk7-erYJWWP>(t%VwQ6^%{YALHjNR6sFn^09Z#U#mvpN62nBnMl^$99G
z*RGU9FWm{pSqQFYq*`+RrP2Fm>1N@D^vQ2or?xPK0+i0*A__0tk+uYWxksVnQ%gFs
zJk|iu3D?V+aWTg!ydCwqb^0bP65tswvcMzeyxF}r?1vSfVp<a)6=c0U?c04n^$-!B
z<%99!l=GJWv=}%4zll~RXy|_$!$(P<;qFW7(4~Bp7~8YQl&|YSUP;j<Ky=Hr`ke``
zVS@`4Crlf~R`|52Q$!HdD^4Qp`({g&|Bu%Df*5ZD%4?ft7Qijo1{eSu?EzN{-ifC(
z_(|YZb1r~qNO$NWEF1P3p$m+Fqz;iNR?AVF?I6G-it2Tu&N6j7a08FZKbya<VPLY&
zN%EhXotAuCn~f;#<;S5i*zxVsuPQd6D4yayW}x?Uk*c<=QxnbTv-<sO#~ai7Zh;2?
zGLN6|U<moGN}#XeKvbmn^Xg=0K>33=IUc`FF>Vrlp;+$DaEk8Tx`Twx9WkxYIQl1B
z<}R&u-DU^g*6z6yp3*Vi*_XH{$+r!t9G-8ddkD8?lZ#V7p%C<<gc9hwEI;O;#815?
z!CA;4Wv!;I9N$4FO4VPm+j6zvPL{VOk<N<&{HqO`lk-yoRu@g~eZQ<+lFAhpL=jS!
zq6!zu3Ia$ZCbc<unyZnetOD|`>sC5R7Qh8&aRF8oF8`DsZL<x3&>Jrke^Cr!{nNvQ
z1HLPu6Yy}{x4?*;1PKg2wN;E0BWC;{KV-?NW`s?oDk%Ic9&Z%#a5mY_pRmuJnqNp?
zO<nUV#8E{4Jq!8gGmAl1n7^5|J9G#+71$Qe;-kSs&nyM*p*8mh%u~Mjh;YgjOM|+x
z=3&y3#<4+fR=g)2xC44LffD$?_LRSh@<Gsx(K}^79pk^z?>YP+VR3DQ+hLgejP!6`
zQI?{d0NWyLOz3;rfOB>c>xFFDCRr)F`{q@4$^Gk@TGN_VA%}O~?-2OU^*R+`@LdIR
ze>;Ky+dCFN&3N^|t?MmK)6CzStD^k;ClC_xQ2ux*!OzE{Whb0`-WztA1__5*pzI}e
z%@(=)Jbxu}JJj%MM_bpGLTKG6F&}8rPI(+F9(PU#obvsaiI(5d+3&EC@Symc37m|E
zS3VYn*W3HA$VS%Sf@IxQ6<ZRm4xr0OJ~9<Y*IVx*0Sw*^3y2c1E3CNmdGThw#J9=k
zDgX)!AUdV|=TxxjE6jrrmi<8=UXlLkFa8ZuEK9N54uf+paBc!KX-&l6NW51EvhDF1
zSeKdpD!F_0?Nh`=*@j~j@~QCeQ6|K2bFwC^Y+*|IX!D+6_CHb*1{*+O-XqJT^qM#9
z3fZmw$N{?v2HtT^xy8m!Nq?7`VOGWW&R|B@;20A+2wQZlG4P$DVs(1mk6Unphj1<u
zljF$s5}`;Yp;y@SY?BMCe`~q0nwg8VGXqrQIk~VP^z|a^tHAyZK4RT^qBo~In~35u
z1n#M`qQ15?Ls{`9zOvP@xe|bn#JpWA8U+0;Q^Q_jwM+KnG$MpNUbDbn6yl-kG6mzO
z-RVK9uFB*7e#L-*S<a$DXR&WXs^6KRih#+@DxKi6RTgV1m3E(-_Qy&V3cE}_TgkcC
zg=ks7am&j(-5?|RtL*lNwd<N`Z)Ipo?s3W|f(C=vXC;@~l8nl6y*Ei8WcZL6tK0&N
zyLadAck%@nvAx9|!3F3E@yvJO0rcm;CN(xWp7(pAZhnX%DOL}+4OHVaLnue{lycqe
z3R|d8XGw^}@F;4d!;fv@n|-bdSNPAieZBRqm6cCBE~jex*Zca?|I8VTpz_kP|Lc{z
zCD^?7-@C@(YyTU(l|ui@=={8(%$wa>YiARpsgse`JhljR_bg@b;u9uA;{~}##u>zM
zL@TB=c_0wL`C2=oliTO^wjruRCF?{FN;>h2PAa}dO11Tg`dFb`YIg2k1_Q!e`+*g>
z^P9u*?`J$4%utHW4k{Ww+ChX10Ne7NR`?HSr;_3c+e0<9^gSyBT>#+Uv4-+v!$0;S
zA0zQE+tjUigFbw0Bn;uax8*X#!vhBC_Jnsrfx>0=5M|jva#DPXwaWDNB`=B|%UEx=
z+^ueX1w~E3-hL%W+d4Ts^d~=w3Ffg-J?x2*#%ip`ICs(XNfvG=6j>7|q7DE{2bXNc
ziDabq>?$X|645iKs$V%?)V*PAHy=JZ?=;5#=~bdK5xWn&e}Wf72@9evCQtMCn01tw
zlQ!c&4;OoT{|F_mln>C$KaonX_zBN}z@&$Foi;O%*}bLVh>h6mLlSu|tMgV-euX|7
zc5BL4_-2ug?v~zTIOW#mHyT_m&WRh!%&FV_z&Qu0^Kh9BdHkA4k&ea+)>4XmWQsGr
z@CF-v&a}Umu}Qi{vJ6&F+87S}_<HwcbcA_Wba@O9pGvz^a<_mU<ZmsBC((EtfQ09Y
zhV+qZM^Nc<Ym>a+$gPW>Y#+?zh1l{VwHyDR-XAuCOB1AXe8cs95`)@Sp#dUsw8OLD
zvu3%@lOu|W6WSrGmCVKS)nGt+U}PuyS9|zojeLcRxu4q>Vj5ZnSd?XHc0iX(_K8)!
z#6EWmKI)d-sh2VHt7`Ch(-$YT^zr&hND)5spAkTQ%`Cg%=BMWo$8og&e<amRc{{sw
zL;r6|miCs^%)RyZmvx!AvHk(O{iSxVzip7vve%KQXJoZXit0N8Wjah>r--MChG529
zOVWI!2@-ioj~@ojq|=BnO$MC@JBcfqcWx711B1Pm9}@|6zW=s=uROOHf5B~WJSZPk
z@-!{du#N$T*^BYxZdT}6%>KHa4i1G)M86BDx@fK#>v6v2jq8}>QVu-yOd_tZ&_AI!
zr1W;K4bJU8GL}{}DkOwOoe&Hv7%|Y}LD!E!%fdhU)UJI+PlklR-i2DL_TX!=M-m@R
z0?RCn+xB0;BU1d)w@~SiDTjf9&D@@cz~@2w;)z^n#iem~Y$z*=*kFI7#`B~+nk~9~
ztJ`MS?wz>7rq8_VbAuKW+$=#RCg1)gR_J--NnwS`BfH{mHBJG_wc9p&SFuYt^C%5>
zdu`w0Z=_PZ)|nr57VmDFnf*eX<t8@$8w^SMdRg)HLr;(D*^?e!?8Juc-orFAxp<W4
zzu#YPq%JO{I+Ct!8#acs1t*HPF%vgT*|Te>90COZj_jCw`(f@*X?6CdbSl*&hKi<y
zqVGGf7v^yz0qT80z1u!#VN{X?5u9F`8b^x<(2_>Y$ILlaa&9ykKBIg`M?WfIv9FQK
zUikk}b=Cn*_5a)77$qr6j+E{mj0U9<kdQ7xNoghxL!_h|X=$WO8fG!+W=MC3#OQc-
zfA8<_d7j_-^ZdDU_BrpkuGjUORGd-p0v(77q(&&8JV3~~8Eq169wxCBF2C|0hTn{z
zJsuM*+NJm_l{?)(QPY09e_$-03%GM`qe4$_fPW=sQn)`_%7gs!#24qh&KC8ylx_!9
zU)@;Ku{eDws2S927<#;_87RHvIhs4MnFMJE#EETrChjhON=QzMyFdUVIGutAA+8C3
z8tnQ9K2`Gq<l{^{w55+e$OuFn+<zF1s>;bX{P54Y_dnNF_d|yYj6{}^r2gvaig$$h
zrHj!?TDzZVmP^A@V!;w6Hs~clqi7{3s0U}4!9GfD_ea&|U}iJ5U01cB)u{YNOORvs
zn<8zXww4T^D7m0`w7^(H$*?!h7|^=K`Z(JgJGaH#EpAJ~zU9Rh$=Fgx_@jwn=E2D&
z11=`RlA4M@tpdobVw}v1<hP`|u!fK9)YyOjq(BhwE#iyoY@?6u!umfdW2q&)t)}Xs
zicI>rx@rZoNMd-v^x2=Q<_UF4nbZL!+tl;ajLS*{YJ`Pdg7uu;g}*(HoYPjxggF>M
z3N5Qg?+2j0O~E-7H^|nel+%v)`8jB;x6YEe6WCsVy=I*y;N|kOhpx$<#4K@)8uSw<
zDizgKZN)PeDVc_;uuAF0C8kCT@5Z|(8$cYzmEA30r&*?U7Zc;n%CqCo-8USqyVdP$
z&B52r>#!RC9O=7@RTtu1sR60CGFf_cAEy1g@2|TBEbp&S$$2r9)l&jNN741xzX(vR
zsdd;c%Z*s!%OyW=Qz&c@xy*08&<3}Gz!WVUE07Bs_Q`ZyvbP=;((AjcaIP>ASVuX#
zTTG<#c{3>5I8>B#R7ixgDJ7h=cq#Udgh{F4{U1N#j&xYscNQzb19=Pa&N0|E&I{di
zEhHAxd7O90k4~||!>$o-#{1@+u|K6Jy6Q1B+X!hV`z=9>FXB6ICt&U~3c+XEXTBMg
zzM%!Qi&k1<Pv|LqHl(5xt7(_&w8ATFR5M`<cW>y*n#*mp^ggxdY!VA@2LF8(>+7{Y
zT$YsKS>N+1n*8n>LpU9xAG?-wMkGv3@Km>aH=8Iw+n=?Ir8ha-4u0&$g++EEOZ~vl
z_+L-v3)sywj{oY**gj)KkNqpZqA>)#M8NHG1mo8{Qm37_T<y!DQG`mNss!z;ZT0W>
z>NG@ftbsqJ?SU-~)e#uW8m<vV>#FLJWWhHMBrnSIywO+>PlF!E{`D`X@=@hyAjhd?
zbz}F*>3rSpO=<%hr#~<G&7-S}j8R$!*aJo7N#B*K#Fs#*B#t52`|8i!2gIch@`>>k
zJ{k{>53!|CR5Ok3s;=_@$cnmpCW0EgUeYODv(4PF6E;TFWO-RgvaN`M{sK2Q_|qfn
zn5jzu()60+U<&{yLk>?e@WpDhP{sJ_s45nco~v3`nq><&(|y?QFtOGD_z;P`RH5OR
zPSyybAb5X9ts3$yPBKVdKfiD%?pz?RR)d7v+iEjXgQUy>hTW>w^B&^pF!~Echv8Z`
zcD4|q1EH-QP0rhSX;^0Lc-}Hxi?gV;pxmv?&AjxyJIrmBgX=gAL!7M_&6SAQ50X++
z#67qBi+9e)$z+3>2K>QyrNQ>G1Y*u(oj<z}WOjeOqJ~sD2R|uTb+<k^@!OIWFM*b-
zn?2+9{OR6-<HV|NUX4j|JbRj%)PLt9&3uGaV;tJW&K!uIBg^s2@z)&~{PwbziVQ_m
zy8?!a%e(7;Y0jmZ)=s9RtDO+}8Rp=h%|`hjKwSlkc<|8H(ZzSeF1EaDOno|^AiwF4
zkL1lKLqJ1BSB=hjP!5c;7lQ)uXuR&j1M~BW;?~u0ob1girQAB{27rrKQZ=U=BolN}
zJMaA*W6<b1`O$4L7UpXp@W(n~2JEJ9$xb#b5$w#8D<W^VqccXNRC6?C>Y5rxNX0JT
z=6BQ8UR+x7%o<Q+DI&1-rYig8W6YRER|itH5&GYnaL=#bbU~O~@Ye(P``dGlMK-J+
zGV)ccf-cm@T~_9;W)o9Q%A1WTThJGs20}NGJ~x&bcIl>eyS|j822504F+b|MFG4)I
z#7MKY%%h5k-e3O%*4a(_cT#Tjb>x*EHUgq;`s!>>E4-Yu+T!#1Pe@XU(v^b+AHSk}
zv}bHwnJA~dtWdx`rD7>67ig8%)8_<#7xN}}YvZ>skg;ewX{R^i_4qLOEW?u*%P--p
z-4GSMO&2QvlU7>MmWi{t)hpB3k@Cr6^)vueq!5T(u*Pd(9eQ^`O58Wn<=X1Io{fv8
zf%tglry=y!iQT-KoM^`pBA%l_JNKHVGh>5f_Hb|lWAAjC=N+U#RU&cWwLLf2>cyUR
zf*IE!Q4z{6sgJFAGdr2M1nGG-iJ!)Mswuz(0T-lq`WDmevvEp%)deT@cyi~h!@X2p
zf5X-|SZ3O+KT7|N^6qyHzWBlNm(HrdhGKYB=#9Wv@huoiwNd#zv}WOiBu0am%%u&}
ztUF|43o_6Xy991ZJb?{T&wUoDi))yyImz*=)Q}VbMfG{n!J`JAac#FCo21Ogf%#HR
zR1Gh_G)?Jpvp5!SJ5_2iAs$<0M>a0R-wnC;iR9hCx%}f?_-fSd{;FoZ4#}X=)lcR|
zNJSG1JqE)ru(H{jgl%Lw5p3y+RV-)Xa(HGY-nR6^I9J)f{TXVt+%}KI<!b>Y$fn_T
zTV^^cZwv3N>JVCVVBz`e=SVYi<S6ev9@hnX@z>u~n9z!gou+zQqb6r(XAksbA+7w}
zEyBjbWz(BuLP&gTz4-qZI^GE=5sc9;Ui7R<C01ASS3FU<+My<6I(w&zT?qsO4YvyX
zM)&!D5A>Naf_s+DfSkq&TR{+IK$ZdViz*2OYx<XRU@^|DtUXIvv3#dRMD`>+fW1}_
z4<Ey2)#taUP;ee_I+hG)Mi|^$B<7Ob`Zp_Lcc-}+xP8MOndq3kVGoSQMI`fx;#)3P
z|A0@&i~_m0b1=b;27NgiTq;5kgd3yB5ROy!Gr~<Wt~cA%fGaoV?M@mikk`baJ6GyN
zX7ZCLA5yMns@?1q%|whAj3lf4>bk?dk6_7>L{&!C?b)K7GD7#jP9BAKz$X7Zv1MV1
z-SZZx&fJE$ibKW}Jdt7NFD*8dw9{7>C}{<eNNAYiByF`ICGXR&o`?9E`T3%obcLFW
z?6DF;qKceET+69Fd$xa)oVOp?8rKK}&O<KqcYE#f(igAKxE!*}&xdlkjPx)+sQB9B
zybHzS`N)M1hqjyB_VnIK*BdU>z6co{xx@yWg!8sAM<Xx<YwQ@-sy`Y$!xY3X-MqAF
zyqwY`ZKuA8<I%w&yED%3HmBTQRsEUi<(`7ct7vz^Jb5fm4FWC55tsTCn}uQp#riu9
zAHumlfg`6>5rD2Cwm8kX_9A=#^7_ism+OPazbsp{y6!^wJkYq_F>!Hm%UVsX)|I|f
zOE-NQ5$8Bo8f}VMrfCwroYv;!LQ`#fJ9V==FbyKr#%1(XKV9RuPVa2k>QDdUDPj`i
zJ|JKlIPG=Y#Ljy)!+XJ|lp)ifX*+lvLk;$b6Z2e*OD*<ZO)8S*u=#?VZE-Jz+<cOR
zVSgFUeAH`RYTI^1;HUJTnG4|y*gq)HVv4p(Ol}h9GN&$h;f;Zz>Fc-TKc@PLksMVn
zhYDbll{>y0D<D)}RDN@7yiwmAqhTX1NS4%}noRnBf}_QqEmMaZyKyBv4f`$8QLpf(
zAme0QQ#W+Mz83kdMA|*OrzlbyrD~@jXQZu^(=?^`NY9S=oK~w&ciRwS(3UMlPAz%k
zM}SdJ&(*Sx0K#KRYNMJ;5vti?7>vd;opF^2oqNB<n-Nwlubtt6PbqPs3y;k?h$FH1
zL+n~wZC~!5gCFi*^x$<lp^F9}cx&Sl(Gp+r@I7#9SFbL!b~gjNS}3zcuA0B`F_}F0
z#gXTRd=l$&q_jcaPtmmD!&eJH<)oiQK08o?c(-g;T4YtHVlutVtD0Skj(la#e7LgY
zAATHPeP~TrHlJEFl-Uxzles)K<{RfP_1XZbbGgQN-!&I>J)1-oJ7s<?xgIPx`vjzS
zvW&z`uhtv#ZudClMV+R2FZXIXjJxh0kH5YppeM~#d3{c*^*ob7(4feDbp~{?u3*>3
zswYLdX<*S=kI5}78?<*owc%F2o85ckBDdDkNCa*ezQ{!XW;mXAmDX_E0$u>*5eCBw
zX&*_kzPc{*2cJd+v$|eRbRJK22N(q%x^}Z!PD`-jkur`qRSr|kMVnFOKR~esC%b4F
z+APxGKCGVSqUaY$^fr?=XSlxFT)Rko_Ac)2vn(OWfbIi3dc3o5J}$DUld`P8UZM<W
z5Q&@G21QzW>27GT^W|Tkujv}`z?(9g)6EFi`CTvCp|EXf5WH0WkF2@WK5Xb$BuJEJ
zY}#Gg!1=4KU}b8?Ws_E=#CAQxlP(=HytRw|gGH_nT)_w>!*v*^8eyb;QCFV2mL)a)
z{ubDsZ~sy1{?9dXUkvc5FsQ1>CB?`4>D_fg|4Fo$5TLgBW%f<Z3%&$aH@V9X67AV1
z$aGx43B4b;;ZC6`Rm@E-l18|(Ze#Yf^q`B47~&5sAJfL!lEE6-xf4fy?zhx*Oq@OQ
zPfHc&6nwSFu|-+)Q79wvDOhvMIc4OGI@u=2i>8SN_4E=v&S}rmpiX7bK2aIg(zfO3
z`+%~Ltpu9wI--*83a<CATwjgdfB`lPaP0AocBjds=G8~zrk#x>-ZJA}Z!=iNuC^|W
zzJE&VYtzRPFXK~kv?0NcRD*%^6Z&v5m?}(eGbHZH0H9_L5-U}*?f`nkkfYB%n10b8
zS|$cSy8fZkj`=8VhQq`ItJ{Oy!sGPhNZj7UVpEG$r^7y(i^q`iD<POisKw{A9gW-k
zLIbun?loMxZ_n1k3en}kmm$&@e+zrPSqdtSt{3lvP*uWYT17Rt<8r@i76dIzwD`v_
zCr@6F+eYBB?npaveH^VfRGULNTc%gFm;`UmXDu{OV%f~QX|G24{~{|rq0r2A+(qtS
z3>u>kN%*GZn#kOLjm}}iw`(A}Txw3mWQ6&DLaWN!+fDB_?{D(z-}F=d6C1J~e}8;(
zJm0B#oTHMProG^)8-#uw%wuTjpL4Ttc$?eUyenHcWxexg68596>ea_?TX(YhQXkv_
zbN1wPORkEk>gJQbmHo*QOj2%@xnrH4VX@~e);SG^c7{^5hoK79-V0yWK7GEXgoj`A
zAO7h*iG=htB-_!Q<dx>0T+!_$L@E;#imPK*I&t@J3Q^*g4F)9hHAv%KgxtK#pH<vU
z<bg;KCFrjgFW;batFqPBp!g#1$!2^Jzh}v~-juptIz#i;CHS;nKVu)mk$LfdO{N;!
zuKu&LMcv(tT?}Awz2r-&cdVL(x<GHot(CG6iNDwvDYz@XLn3n~gB*OS(Qqaua{~9&
z<@)7wu)aw^*=F5&f=ea^`V-)VJlT+DzWU$YXi448I3xAEZ-sxN*|c_-)*s+Mb7XKR
zNG_F#Fdap&<QCYxHf->p|903)QmQ8DP$}17`*^#Rx_O8_%Iz1Rmf&?9`B)eJ*fw=<
zcX><qb3qqGjC_u$tfG2#TqjKP^%B6c=nWP(oLFAA2QS?`gvG&Vj!K?%lW02%4zeU~
zLD`thOZH@tuXgyz>dY-V2_}`vWOnyV((^t#C!L!-$z<VFnRTh~)8TIg3AkxDJu>%b
z2eCU@?9(Yq)VdnM@y*QSt*d{z4gud<o2jv2pJ5dNI8O|FG25dx4a+NC{uHOfokY2d
zA9BLtbk=#muHa*iVBOq+W3_9VwzC<h^}PGby!BYa-RUIep|+nnX;pb?@|>nTk~LmG
z8fRWl)_`~KSd_>k=xRVlu^b!J4@Dc)t*4y!hgRA=H%9!v!8He<fPyOaOuS!1T}$#0
zPbzjLzQ4*MYdTvv$vyv#CwHMB;WN^@fBYddft|Sf@7=kYb99^pA~z66e!c4!>^(rp
zwT2Lqr*WFmt{T?)K$w2L(TYnhz!QE_vOp50YWOQk%%NVxbMW^RBo-s&md)OB6!NiU
z<Xx*$I^%aS7NdeE*Y%HE)wE6%38M}?c8hfLaE!NVkflLP!)%_a$1x*9brOiCOpaU}
zC_gO6*yQUgasG+{b+DPo13B8!!1(>7g?fjtC7~7)|A-2<*eZ8)l~VtwD=GS(bKHjN
z<iFHf!{j(`g#X4h!fl4@L*L-vxidia9)EVSx4@Gp=+L{c0NvQ54Ty;d?=Gu>);Hgq
zS+NeN|BjwK<kh9^^X?)hZV~<BAtNWBaQdSItXf2iWd+wMLOaMiHI*HUj%-R|0yunu
zlu(C{bX@p$6Q5*Alp&1X2eeq`sjy(2fvUj4-?vzFke22~@r7g#I*ExQ79+eMLH43t
zaLPbStch0Zl^1}g=Y3onxK>4vF6C#Kjg9sF(3m_Oy!SHtATULA3s)I9*b<lR-ej{}
z4Scxtwf+=o-9nLw``l{S?+HaS{s$nRo_J2R5YI=;eCKA(j=kPE*IhF#UKp{_jBp&&
zL6vF?gF6?NQJD?Ya+4C8Nbpd(jdAsZ=f~C25u0)6A1Bx;*EgGTd0AZqJ|M~4`1>Qd
zwbG!S(#p^3l=8sUyMx0!t?G_|=!#{fTcstDxZiuzb!WiudZoK&0`ApHRxS*0$IfYQ
zfU5raoQ4Q!r(DJ2?^bsYV4HNLt@S)syh8>sd#zI*Mqm00*z9GQfJ^wWP$G1{K@p0*
zczfX8pX0so>E?An;H~LmI%&WWSvf@~`<eK5Ma#QDXw545VY?}FV49J$>X{~){V?Z&
z%hx&%SZ6K2rGcc@GZjXm1j<qLWJR$3_H<DL*lfJ+vlE}cUARRs!=4RYxu)8cP=W~N
zsAFI?PnB4Kwb)%IV3oQU1eA>$6xs+HGE0se2NXOLWn})lAPp3E@F+ytyL1mQbl)8J
z7nhdWrG#cy&)uK7-WzN|iTr%I2mXs##(Z~2g5Q@t5qDjX<`VEmcY?v-HOxFgoYQU$
zF*wE&LIHqc^D?DAuvZk<vLt`wT9I0I<+1sk%j471#rLT1jM<PD6SrC3(kx5%T0`R<
z^9l>RO{v?rNx7-3eSp2u0RR$TVdwY}z48NQ_lJYV>|V?%t!nvZtREYa<0M_fS{JN|
zuLk*We4tl<Rh$2i=Wk5|afq<rj2kt8!4vRXydO&g;OASApA*FzdW-l)<;%#Hgezo=
zyK7M+(6me`?6FEWd*EQt((fM!k|$Ol1$uI3+$O+0U-%|52p5bSEPKw=<%@@MX|9(W
z&>o+Y_kI1%%~s(%QpAY8U)-wjsDXJ<YJ*hrq53YO?vqFUyQm7&@DtKV1KqYXtwXPK
zi0?-b)Yvd@sC@V7?3+~=CaMJ0W&eDef?|#IzEasc4v&4MBF6ZsAEa`;-AA<DPyU_$
zL5l^4=2T*kYV_?PK_Hur7yXip(;RL01%s}ms4eN$_ZqZxK3Iw~=z8tI7oy5_Kk4n*
zFd-XSJh}!|B@8*iBJ&!W(zHfArMZGCCB@2mKVs{KzCV+k6zf>ZHjvkvA|3`!uExvp
zA7b$T$`aTdUk(qP67SJXhrMtMr#X`f<YX&)J_JexESC4O5?D2A#BC_h<Kh?2EWU7&
zvy?Z<4QmOIkdq%cG*h41n=mF(-Wf^2dd_VhWT)HHgi^3HE1pbemt*7^n<JhBLir%(
zNQC$29q=QoC;pjEldbul)!#PW<1)5@2Kg%e#g27T-L<>Vnri4@+K|iR8Oa8k<P9p+
zP-|LoVET+(PKvRY|DI<SnI^SXA(K_=A#~_92A`!6G9>85_R{%>IRB%fUfj?4)FDrG
zazj5@<yh~u%!|tNJ@*AECT=Jld9>9&>Oe`aN~)OHJtHkRY71ew4t>V+RV!a2pVW>T
z1s&O^V=(aKL#jn=v8#c3L&Sd#a9NWq^*#QpZRzAfC3FAK!Fku!y*g{C;SstqasYQL
z7QnW9N-*79Qs;-|p(L7?AI|U>!1N&7Bn_yV#_gvZUmPNeuFrVPlw6!rt~w}nRyLW_
zAo}v4^HqGP`{Ql+Z(p@`n?2&WHNXI-$ypv+X;N!@{n<}yfP$?vXJ|XwFqD^KNDA!x
zR~&t<4(|$!F74~V!Am;qLB(BVnah0p7G#)QlLF@k7@G5l@<bY#%YpgJD?x(i#7kHS
zy4{1?srt?q-O94ZQ-i`zjzdsCte(IF&pcW+oYygz-jD<Ko*9t%Q4|suy>0T^3>L6D
zW6rPn=Ma}bbrbxl8C^4O6WdI3YT|C}VUg{P*X@dIMDI>2Hn=}APd!!LAkG-P@Ov>Z
zR;fmda!T~LrRPFi1^Aqfs#LvOs5Wm;GK`zbYUWLezwc{bUp<^=Jq)--7w)7LUwum+
z?)}6vEHz?}UY$GX#}2UOcTXBJdwnn-$zLUHbLx%3yy%*mW_3VW5bu6i1lE9hg<;Z>
z<pr;BoN}&_=2>Ares}m*ebc*s^`sP9vL(a#^SiCPW=INnxavJl-zEkb1UcCI18#N4
zy}Ew!=wYwO)Y-0==N}U)+qpLB(m>$sFv{dHc?I5<gO!#1KY341l1T&{Fb$39ScefF
zm4C#<ogZGxJi7;y{j2u$QiuegIf7O=!rL)39XWptCveYSKI7ZR)_ykbJboSNEd-4@
z^*oQVMkvfF0ItzLwoh=kT^+*nBI^4orRuwEPr;7}GL=*g9;V?sqFjfi3_OtgPk^Rs
zYSBUt?s^)5NfDRGz&vQWUi8;5=$&-BngQD1?1};Lp@#UWeq^I<==5uS{2n!t;=flj
zS%k<ZI}YP;lbX-!HB|O0I)QCtkB$Mdfb^mt4CC_ih_pBXNFilqD(7~tx<~{*iY_7P
z$c~qWqnd+|`}pS(jDhH4d8HRr2K|OV;Xoqj(6@Cx@Hmd!OqM%qQ5WkYmTO4KhSo{`
zQ{}q2%dhBal3B&Zz|E%;-d7!Q3ZE*P((#@=GFH)AdSI1CQa?~dTCn3a=N?r*<fc4W
zHX+o1>0PxjPCKQUYT33hsGVp~ZN`;%cUroR@x!SH6<FV?I$;kjarP6`s`^3Vl$rW5
zJ8j?x_ncT7@u{eTSDo=0u$O=jDp3%@R6kF@w_{8|wY0BWQIH=DAZ!f!n2b=qZ;A{A
z7G6NS)}1pJm+ZbKrAxTKTfFyo>1=84-mmgl3cgJWIQh}sd)Rf}blhH3krsS=d@n3c
zKgdw;=-@nB6<5*V&HEPzkUQm%5~|B|la8M1n1Mgy`bA|KQ1^De?Iq^%GWbss6P{0x
zM!>eb4uxv=x^=Cq;TBK|Sax|TR9$9Y*lSDnHiseF4g=A1bP_i3bt3bo8#bMKa-9H6
zD?Ol$+R01L7%Fe^y8QbyKX`+t5Xpffh5P%MSFw`*?4a?z$w$ERTqXHeaht7qm$l(u
zlMpB6MUJ6ij9s2~G2bx!_}+=~|53qD3&`0*XdqR(`?Jl%?*#}AhEb-O$xf7ly>rhf
zYaI~O6j12aogI*|Y5XJnVY#U$cmNCGMuYZl6RqsaHgN*a(LzD;%S6V$Io}$ER<7{j
zA+IrU65D3Qx?&i=oz=T~3#U)>g)rwq!j#bj+%q_fnID3M3A>gQ`VF=l_!By<8(D~i
zpuqfMamEklOw|`LdE+pT0D9L@+VYc+=fG{sZFjf;6nju%a&t}(^VmXdm<Tqp*p~(h
zhp4Sc1r61&uJLPyr}K&kHp{wIRst8l!sm5Y2j-K5<$mXb{*p95amr7D<F=ff$_}>b
zu`+4QI{blpH|z8B$~M>RlcLr1i=9%B_%cB40tAO3Z?Ddti6^J@&x<O~vYS_5^bar5
zarI4A3A4!SdsjSnmg72XfA_i4(v;|f+|4p+zot0QQiYsx@l~7(C}l5<_u=?1T>;2-
z($+h$5+|S7p}5QIhA&kg^{Tfm^2Zp&q}%kvgkF~|{R;L#MSD-?_b1IaV36lVfzAFK
z6nIs>Q~1mqYfP$t<F}pc$)`Ed$;I3G#nj-N70bTxa=u3|cQS+)8Ys6dviNAl-X@R+
zEP;A?a&i|t4KqokT^v$vY#}%v0G^%C%LBW+NqVJPy>giL+Bg(HQ|)HT^)_wqPXmW<
zvPHdY^4s|x&mS;p29*{N4W2wJ0u_;adqwy2R&5a46Mp(~&aWud4#`ci`xp^HnAT|y
zIC37|l8l|UY_Yindpl4v&3ZnoAvRLumOWSEmXoh8OJn%^Xk?(_VPyIVZ^zx$ZgHz|
zeDO**8COvu$<kWyV|bT>0O3E5La&hW>Sy;LF@={WTsr?wRz`4G#taNAjfAqls!O6o
zhNW#n_{!rcrK?-Cf**f6A&#V4SAZ){RDUdARz?mn;}?9ZoD9YURl4W3?u6q)S}etS
z_HA6U(~_^eJBwRah<f#6X1}B>lY~;jRlKZEzo~P-*A9%G{DXC-nNxnWm$I_kWJ?yo
zY7&@S%x{8l;VOKFyMtfPCnG`uk7#q5iV~9}PHU9J_%vR^**C`2fhyc{4hW5%3ZN=A
zFL7g)t>LBDx&M26t3-(QNUJ0|CGHmYHnX;TDd^E3>V%+5Pc!T%Zhwr-CF3GID>;d2
zmREKT6R02RSR{Id(9=$r*0<vPt!^<`DaXkN0N)Jnp-|>5vldw)`m6TH1GX{ITX1LE
z3Q3mP{K(zmbXznn=`PfNNNY&b6Mv_tuB5!F1leHwq^nSOkt+FJ*?Hq>QdPaJvqhC4
z$Jlp0CKs%a!Y-ne`Efa|L&E%A8`M4z^haTWjpr^=k2$54li7E{m1M?Eusz5|c~E>c
zPL7R5KdhF+t>r4=L1LTc7jY5=r!Tb@){CE!LIATO8*hllA0^jRj&}!%H9_c#UP5KX
z2158;j1h|2p-toKRsyLyk)r0=1Yh$UbF+m^{b4$<38!9!A8W}*VKdT3^2sVQ%PpmS
zZ`)Lp>Uiv6mP(ug%5>}`$9neOz2V?ZuK8Y-J2^^DfK2N<&B>sUG<#sH5>olRP(je_
zecP~gPodTQ&nrKt6a~JvH-_8(o@e}?gCd14=@p}w-<#3CpNA!^JPhWmzNYb|ruIld
zDg33(+A8xFveonQ$Tltg!dhum_Cpj!?1mD=yYHILAiK?NZ6@i~Z(%CWW_JGlApd9J
z{7)^&@qnq>aq`RZj64MeMH;&;f6Wil789iwcKnCs>&kX&AU=sP`4LA7U<;>@B~!&d
z&eo6Rr;?_+PF%PJzE~5+b?!T#HK*W#XDe&sg%rSN;4>*huppVYYHl(K8MR)-AeclE
z?638%sqCYoIuAM>%6?FWv&&n{7Wp+CAO0nVR-gokvu%y5EnFTzlc+^C3QQtb>!B<K
zzYrwsk<3B|47Z;u1);ot0u|`o18Qtk3F=;W#vQRXk;8ielQYHQcsE?9IfO8Ill1<J
z6on`k*tIIn#)X*^SK?T}U)kHIq#ZlBQ5(w-`+<bq`ElQ#fl0TTboXQp{WJ}__a~Bd
zOICxC!NyNgMP`#M)23SEB<i~jHdE`kA3;_D!T@|p)g-^I8rnG<XsPX12G(b<Ae&A(
zOMuq4q@CODbwoR6kCmqyBx3MS!?}(*YI?o7v&}1M!EpU*alFh}U+#PYhp0+Yv$7(d
zm3MuiLzqNsM)}uTIqf>u!9Z&*IVUFZjXPCKT%vfju$RQAk`IwVZ(QsIV8qJxSk_p6
zXsDOueiZjw*5Yk$>v(<J?b+-)>D?0P>mKq>1IKCL5xf^B#Wr(++Zl|9O*NRq?%Fk0
z6>r>N+vj?7e7#;D7OMNzcGifgx6>^U!HIa{Z%~s`f{ijq;t{g^gDQov16oW$ls`{g
zq(Q*z52ScQ+Y6`@l?l25_4BV1a>QbKtFB*u%8Sl*D#7U00-a^H&B|o0G>T8OSE<@d
z;~KuY42EpBFW5UOcoJm4TOZ#CkvbLFFfy-23jK7@paB9b6q;8<x)}aIOD(zk`Ul8R
zf6C8G3nspWfUa&Zn-I?br9M^c+;?`RC8@?Zfp}GA0@lhDQ8(`=Ey1|tUW^^bftL>f
zN?>M3D<`GX5Kmv`3>ixfvh5&pt@v{DM$6Y%b8`ow?$NkO2w{Z5BxV@Lwaz&OUjT~&
zp7=?Ls&;+L;nBYG2h%JSh)g$ZYvN9SB+m=NV7Pxx>ipnS0pV2;!Yn>uU0}D-{8g^1
z$q+R_C}U9sB<`r^aI}4v#LM+;{JA}i1@R$Xjda2V3cJ4mzG-Rpe2{O`(Rx?*oSDv3
zK5PI@{^h7vfr6GNTS0kC7`&JM`57q`fGtWZfRoC6><qc;l1+b~S>?3e3o!a&f!A;F
zdZ+AiCDTo2wD#Gdi0oz2TQe`_%k_(yqrJJciU1n7qKs|}SJ5|Vk(PN0Ddtl=HUrc;
zb}?F?dTSc2Z-3d5^ckLe_1~WuS~7z2>#sJ%yScm<1DmyM>XnU?D}ft>)lVAU0SBB*
z*{~Rdn*p>NA9|Q$Yz(fXDj4zPEjl5WuO0|>i2_F<bUNZh`O&5XDkTC4X`5edI9T*k
zO+cUXvrjNA+o~OtlGff&pIe7uS!9sYpiS=}WVRD-+IDW`X3W+I;nEMGPerfNwhx2%
z+h~MO|62aOyjne;ZRSt)XbUs>VvlQ5&;!ENC@;20kNaqPWcTy8ztz`!|15%}RbmHa
z#|V4!XL0Wh<ME~Sfa-%^y^;N>!|IRY+M9?au){SpcH1D<CT;Q3x*Kry$x!!mG<ejj
z`2p9iG>dg0?oTt&!TKI}L_*k`ZoAFP=uN;-8~JM47Kk8s79Lz38<?esd_0)qQci83
zY~1!n&Ct9npcV4YXC;cGsJ(sGu@|TF=3g*LiR4c2^B0pMCvuY{|4Ov3jvTH|w>^*;
zKICGQ^zG^RnTQ$pXo>hvH{#^;7<Hn7QMH21w>Cra0HdNbvAhJM&;$D;C2$N$=Hua=
z{6yO{3Z*k&nUUwAqxh&7xj#OfAA!vbp;G{VI0K_#@%PsE8ZNa+tHNS?ieku$*_gZx
zJ8`^LYy_hhto%iivI8;YSU$B67AuJ1Dr`XyD9E{_1&>LNnD}$26>3Blm#!#;_;~vJ
zC&X@$#n!-B&_xm7?g`i1qN5eB6~7H$@1c%@YqhZG2$vKDtwv!#4NqL~@q^G;Q!g9J
zm;$*E=+mj(MkJRv5E1%Ce}S|PM1oR_i6Xq&dIRIuoi!ya_0G3lYgDVdh=m{t)C(el
z6%jnvk{{JW#xV`aeu2iLT#efqKh@G-$Tv(D^}Fk1h25hQWsW}|HmdpCc^{8k2d+HM
zOHg5O&w;2li9#!P#jN-0;Xd}iC7x9ESo0qBB4KUpt*5q(su=i_{t^wLb9^k1#c=of
z_Cq2!nWK!K6B7Mdb}}3eIbXkKJf=&YXgxWnfsq2ETUsbxl00)lJ#H?J*3z|V{O5Pq
zKJp~Fd~azmLXPtg`>8-@>;F={=ncZn-)-QOMK}Ho&~|iDo{QKgyOqRBJ0V)}O|kDy
zV)NK8ZVP{;qfy;uf$XT6jRpG0qqpMAIhO37dmHDPdw!X}8~&Xs)3_xP2l?KRU`M@c
ztMjep3+BIcZcv}bJj}SGg*D<Xx>B@uan<$esL6sB0~dU<n@%1iOUL<z1OMULoth?+
zD~piT_?pk7f6hc=QRH>g8b$@_bl{jTtSS7IzSS`Ne+dG$$sN9Ai8+Hw%Cz!iovS0L
z*v2Tz%Rl_U<<NQ39K|Mk8W@Go?*3s(mb+#I?mk+!^amRN{W=d6)ujbK`LiVQvKte-
z;^#0g{6Lpb*Ob@{2*8(DlfemJhn|ENe`{?$oSL_;<udUs#N9ukFZ$l}v$eiF8^ATk
zF5mV3${iQPRUdakjHm4%`t$nGy{W7$B!ZeYTiJ}3Uou-8yoJ=@_da+UHDGKXwqMmX
zmk?LX5RuY(l>L@HhiOg~fi16S;a3o+Bbt9sqnQ?2LJG)d{#{;wWiHQ_QyM90q=e<M
zQvY2E@}@-5^GmWNIsOG82FJEBJ4ti*y4XYK!StrXwall9kxC<;K^5Y(44#3xkH$Ww
zn>IOuhLSx;^36^wi6{xf&Hcm(hJ^T_Rk)}(F{_-HR&DSNGC1WQp&#S=&{M$Y_<%K*
z`rt#CCzF$suXdyrE1I|4`swoGzE$d3J-Gg0;#qk(q4_Csl7Qv5;VF{hw^%M<ej)}0
zO$vATt9)*3@rB1`yxQi0eZeK7@)v)gaA5{)w!{h0@*^m>N!5>i&})gO@Sj;@ql3H+
zy^VziU(b45ywNkviX3+J9NGm)O2g{<6&rfcDXcMmHpT0rV#Vg7oX@%o8K(xp73UlD
zBS;z3Ftu3=`PDnCvbaDlrZmo#KjDO$o16tSL~epCx_?(2@aU-xX*f%JShz0l+7uA(
z0GaE&4MC;vun3VyBe;8Ok~el#tD$-?^J>n_^6!@21H;pHd<HxpH|NOKR2pj@FrYRu
zL`nngqbbcH1-#eh%`Q14&pt_`C-lZf1drNQ6a>uwr!UdF|17NVa<eId^uMQGo#R4>
z?*_v%DKovp`N8}d`U8%%6uJ@TN7|KvKF<TY2VBMo96H8%HK4~1&*c!gV?mR63)Vh%
z)k|n=<WrS?CbMxJLwiOJf?s=hgkwX1w<?T@_A_+1Dn#g|Q<>syMeK3x))rN$I3x`o
zg?Io@n~c&J!V%m#{CX<&?TvoQm2ysP+om}p8xE<M#l}_}+Z{lA?{HE5CEa#!?H*!B
znY+(OJ6XRnm*_8T9m+%CxAl7ryiaM+VYV+_$cu|8{<@BjaEuct9;p(`^NlC?*@==}
z)n!Yr?qFM{U;2P@>nui)Jxscus;j%k7UOGw>T>_5j--is9!*+GVCGTx&A>Dspt+<x
zg{f3t^D6;Z%)=b^0D_<*KP#X6uE66&?LaN9uW{k-@RswAO_}2cx$B1Gy7)le%fZMk
z$*a-D+l=f#mw{Q=k$xc?k{VOMRTH6ptIDddDLqMn(rhXI*)T}zBqeQ<eenTs%7@E+
zxTSMucYfj3@Vi`Cj;SZAy830d--&@*z={ywQ2=$Aljb>5c@rIqNpDD)+QIQNvU1+R
z2HemL_}E|Vu}cSZC7oN9PaKMcz=0O^$XmPTCJ|&UM9`|PfY)>{F>hK5dRL{C^>Du8
z#PlTPAmZdvVp85o01exjU)6!@(l(JfmBFs-sY$J*D<(jO=m6@-w(@--O8^_&S$gS=
zT5Cdj+l;2MAsj3zHi0k;NK`cD;B{idannFA7%xu9t>vok)m=<HKC5gJeU<C%JCb1I
z1?k(XxA>&f^nD}#UJGDynOP$5wMF80`M)_8uRHJ1HR&gt!?UO&79EWt@1_)T7onX9
z2P+MQV8W#UvH{)F2fvdg%Ty1U*OSO)7H`GmxgB%5sw@kl7rCpGx2Lyc)yL|si6^R9
zxLy=;)~;V&&9E_SiHOIyMtWo`v*%+61DN2=SC|lnjwy2vu5f4tfF}7-DG59TVRFb(
zK$1a9t(Oz1VBJZgky@mqWLia73B)f~rTB#t=0x}xumxgZtETlNa2N0k@*KbzB9?v9
zxp!i1a{Ee6u3k0n^sJ!9=>SXx$oPFp4>PawjL<V>af_G-{KZ&6<0{+K>WD~im=Y34
z5Mb|^D{S-%eN^%s1$)6K-UVyb=jV5q$@MVz*rcUHSl<A;W8^lUD`g7vEA`+RAuq7P
zI>KpMDrK)m<NK1Ywb|s{u@IwA$p3!0KHUty+pJ0qscG*<J7139pTND&h)s6B!n8Gu
zZO9L=(-MXpny}rIb7V5${H_8;^U}Z!o|@6cct%NDx7UY>1c6Hop~++<E!j=9plT=^
z3P6WNaD`=+j5|A9j+}&a?x~k&$U0G<mHO?~A^%x&{y1@Bs=AFS)&W^YC}c7qb=}Yw
z*77Hx0;?|x-NgKZCz@HzCv~-8T^m(N_#>IE@oLHJG^~SPHZERmJImOtK2N3z)h+&D
z8h>QBKV&JfvZfK1m0t2lx<Zu?+0++-J<-9l;}hLZ<;~z}atb~`W1`*N9FK!r4Am6q
zo4jHBx;1eUE3LS0n)%lmGm+&!#KTW>6*SPT#$G`rCf$aVd={(~l!LVI^pdfp1qGRg
zWv?p(7Z+>%*)h7(|0VNF#zalBxd!f<CLby?_-#G$OnI)K)~B*V8~ea5Yz;2*{%!sv
zE?<LY;@^s2^6#7<4n8h_O%@`z6q3^0VZHe*6r7z4Gsh|HeFP%JpNbZBs8@8LXCK?;
z?6BBUAGKi&oOV<+{>Wnm5d2lsVn}SuhZYNi!XryoAFUejOeT_hgGp%36k{F?@mP?O
z%ZonBCrQUW`4P(xSMWO!B^E0A;dJtV%uEkKn60@!&T3LtBY+iB4kaZitu`?HDN|Fy
z3i)0pgeN$<Y8rB-Yt0+-77p?hPJ1Z$x#Aa{@&>`FnZb`S-{yl9IKdKYIlM@b19J0q
zUhHTfdjC9Ta4Rx|BaNi$q)8vS=ve;-fLtpd+HYKjRrBzk$W_dQ<SL3j!}uXDAGco5
zE@}p?oN9ERY8;Emorw4mvRuP)ar3SwO4nwbU_;V@%Uz+zm!)m5-Ii^m;H>3t;aKmL
zg2P4XFI#4LDy4$w`n>ksRe<O|puuV-NQ1uG+dmODX~IJ0CQ7QlEgNklU~pqOEae;*
zI%EuY0MH7Spj1&WI?wI<o@+eryjY1&rjiK6EYgxK@6nd&Glc>fyL{|}nRdWo;sZVk
zBuKW#KbC$x_kmeq-i#W7@2UdcKEObf69L<G7|a}nP1x^^xGSE2e?2etxOZ)t!Bl%^
zQIfF|XUxB0Y%?{)y_0b&s(9uY8PQp;s$`efgwlA+KY?|mfXi=uMp^q<P&&A3)-}pH
zHw^GPwRiz5rJSRAt-dbDO;4x`>i1_8dNm#Ii#4~yfB_R3SzIqzHV=KlBV}l8T#cm_
z->Ld_yc9~BsmH&y{-1ZxqJUug%fC)us%ke0GvD>p4Q7RjM1Ra?XMguW`FaY^Z<4<S
zgbulI697eigap(&3V;uwa%6dfvp2>xn;#VpAp6U{8>xm9P%w(M?F(JAFS3G65qM>q
zc<P}|Aiq&IDjJrljGQb(^9H+2j3Fl}4hBEx$DR$-yT{4g6^-Bko^o@~^F0VBB+rv!
z!bCkoE^;`?EcF=Udk7r9a3Q7qr@_h@*KwXA@9@uQ3WDh1Bo}2iO5j2q)UL&Dbzu>+
zVK$?R<{J-bOVN7haqV?xn=d@_5%WG2H+v+5IV3I2u`02pVHBa&n`thOg|NBw+FX*n
zDr+Sjzn*<(I{Creqo^{@qgO~y9#_6Vfr6e;U{YN6@`EBmLq4aPEUS6lkXY(!S|sm!
zZgY*$^5T79p5sYL>~RIgR3vQEV<gz&dt!UJ#vv!sBNQ%s8R~L5vx;OL68pgf0aN2w
zE@x(WEx4H(w>5AuaW`FhX~xWwQEdm+<6>h5|L3Jn_)EWikKtbrM8+5kA>wKuvL2pO
z`SXG3h~pRoe};Hk&Ba`{$x+hobz|nY$yBkC{0)b9t@qr+9c~kj<ULL8thRD*&Rk<=
zi4fgB(V99RJSPw@ZHMcFVb9B0dT(l7%*9RSOLI4d^W+@&XdRu3%5_j-3-ORwoLn-b
zapt-&jHxEx^dCT`@VScNB2y5n)S0)0;wFLpIyR{Ukpymh>pL%nNLn0;GlUrQ>|!q;
zvW(rj9>|Q@)2PTZRx8L3a7u~|Q-amqs$z_<KFb>n@W|J`%1aSwH@j++s<PqpTMnbD
z6k}exY~9RF?opmcDF`_Ix3Y`7zrg!<%-2~2u^XZzG~gPG#jnlv-JrSl3vkF+7kc;J
zZ%5<OjgNw)v`vOziNfat4|Y18;9yNn(YHpNF6>F{KAZ$3mXhuZE>%)!AyW`p0_}6j
z{6}leIq$gMOc3>B)*D4!kG!)JF?3sc`7d4HfpXQIF3lA$(@Wa1aKL9fNA9Cl;Bc|7
zN)qw}ZDik$HQ7>gb_z6WrFYn?vd^t}`RyyS^8Th}n4jjlSlm-8wXqW)%|Ea4DpX&A
z1Y=9P_!CQ3-|dq%?H*&h#h_(Fq!2i|uc8$LWe{Yw_#*_;d~CWee1|h>ktH{8en~PD
z`>mr%@L>=&B1Fc%>>5swa~m_w{evw8e|A>6SC?Zhj33JMrS;&{u6#uyYO7<$=l9^_
z?M}GVl(ld(^!--K&Tic7qF4`E+N7rVJZtxvQ+HZ`c){7ilgsJFn~Gx365(@>?WOd&
z<sl6}4yK+wyD3I3IWrdr;_f316<M((;`S-MNIrkla6SJw{Tu&oa}=9kdJlc}55wS6
z)Aqf2t@(7(Zq7W%gnP^ZXIU-ch7=j@ugaTek-M+!PTRlMQok29mopYO984aHH53|e
zb>p#Z=WkqS`%IEEFE<R@NK|H%80zoqQ#bDPV`aM-IRZ;~8=PNstZ!`a@O65R_269*
ztIjO+THp<nnF|OFT2DuR)rJudS2K*?SQ`Zh0<0Q3jR5r#TVnG&W7-vBsvY<3LlV0(
z&3k_eHU&gp7};vcjNRItz1Bh63{?&DYM?joJYRLil^!H!$jExPzMb<ZmUu?Sg}{W?
z;Fh}g(o5lPk@`veufmS;y-UZDGGQa!9ZFxuW~ir^eDn2#>iAEGLIo>*vEE5|Ut((u
zuaU(HmSX<XYh%2J0q3-kWz6PTfHBE_-$}G=AVT^>5pIR8fkbo0WGpwntb54B;X4Os
zz!OZ5*~b2z?A;X&hz>(Q?RwVdmf)sLeD^?WKT{ak8bUx^TO}{M*+)jqt*X;RK7vVN
zY|IeW*fRRmH#Q^+Xg0Byq?!VfjR1(Sgim26FZico6p)Xpp2y>!N#ahudPnyQt~RG7
zAKl^p#sYc|k0Wzq&w84{5qSF$C1QgiS3jVHr7Oux-GZh+zoR)x<LSwh3@pt{EprSF
z)!E!FdV7)ud#;@S?W;Ithnxy@Ej&v~&~RQ87s`k8Xe2X?XM%vmIit~fPHA6_ua*_o
zTwJWwby_T-=I`Ggm^TSSxzfMO+=#j|g>K1Sd?RI-qZ#cThaFP|A5j^klPFDK+nBG$
zf!46}j4ivOiJ{LVxd0Tk$?sQ~mrHgWL=3?Q2=0qri`sLnz`&!cub2?4-nQWqDfwN6
zI@z3hM(GCAg_C&Xo3WZZd6%2@?xes7GXl0j^1GYi`?)dNz@~Ge-}kN8KwmUtfa#au
zJJ)q{xmyd`S#xdY^tc|A97Y*Z$SI)SOeC(R5`d()L%Ys<ni;ATDr)eL9>QFGR-Q#P
zIjNX^-p`r$_<~WTyw!D_XO<7hlKlZcXc23GFp{}Ndzmw+@6o{LMkDA%haaPm5#*wU
zOSvD-nh+2(uRRI?-!Old7UFO1j`ucf+vJTUi&+!&?BkihD$Y+S{}(I0daV3AKfheh
z>i)Wh9d1GSEWf^;eiD&<MeOPm8rW%rByl<wG@5w9vGwNjKb0T9;3O=js`tOLgVV9`
zFE1<BzRtRcl!9iu0^9ZV$xC%UiIz23-EI8<*TUKlf)M1ESCVE7xI+L_IQvfTU}ehh
zim=x!i&StbNdf*=pO$x-0w9f+Qv3vnd{fGI=FyN^P0&pr%#UwV$t_!lkXsuXqbl>G
zRV7gDJ#6fy`ykNMt#9d6BQ&VM%|d$|=i}k^d8iK`*PFT7w=z@txEDpaE4gY`@Zv2G
z?W-o(2_}S^$XmQ!R6!A``QF<<zU$F%q&p^n{-JeDixv~!P=hoR9xD?%!3j7PlinBM
zAA$Q|pbnP9N3EAC!OV3tY_>je0j7M2g<`XEn30asPGT=R?`e8XX9iv4rV=)Jte5!}
ztmhe%GTBG`bg)rqq~N@x!m#I7kB|tDxjd=&VuhjeWJqd%x6;CMzGCf5g^X1S%b4-~
z66u}%flqR^f_9OIZhCdYdZohKjhXCH0K*jD0jFyW5a!h%Q!Z!ph4N3W*8p6ar#j6i
z>A+ob2kY8nnRvS`;zL}!b0>2t=|R#yQgZE|RNI^Fyu0#_+kMw`!_J+T!mz0T%!@bA
zV;Q*jq0<RDbbCzL{g<#hxtzd`>F^(FMqpm)#?;9%+BM)=&}DkCadpd?@~7J~?f#v`
zi?ojG%^<bA4c9y6T>DI}O+0BUBLkORxwb97AthYS20ysX=dFoyy+OA0m-OOFnVUY=
z^E!t*_i<FalIQRu8CpQuL@Bl-*1bD}?!BNk++9+Z^b`us#d}J2-Cm(&O;;z4H;iIy
zViDP8h}7EQTFUjZl}&-3yjR?H;eSW8Ya&Kf&mi~xR>Q<Ox6PDwB@}d^ac*s>eJGe<
z!}xcR@NV|o;`yAXwi-44>6qbv^1XCk4)Y#-{clwH>8hctYfZ=HpsQBbW1gAYwPbDf
zMGE}C6FwvFbqFcTeK+vFly#eii)O^SyO9Mvl27`g_Hai^=3!Vr`Wyj|6m64sC(724
z|Jnmu1#w~ZyI{7|6`;oDwq7Z5EcHZY6+uEM%ZQ~{6yb4*JMALVsFxv6KQ^ZXx(IiN
zYS;gC+IK8{?3HMynCYQIP9S~q7_N`T*y)}<+TVQCFQ4Xo+-m^qEW;i2(|&{=!&gKf
z0^W7xx>P`SJ|1vTFp);k9|(Aots_L@EKeGNpxrb{3r%Pbo+?7d?a)_Hn8Ky|Gtd)z
zOsDb1Gryc7!ttC(e|-%Wkr$W0w$U)t8q{uf{(p4#MvFXT#6lZY@3#BlB;D*;)hK*G
z54DI&PORU-<(0|i9WM`0!;Ep)fzLI4D|-1@F*!o}ZpMFYCe_C2t4{iXr{zV#Y2JwM
zh;RKH+gKR`N%3^?9x~59aCXD<9#Qabo#%ZbP1`Mbf1-<>7k`XNJE!i=beb{*UAzL>
z{YO+|7_@FU0EDfZ23fG*_lyf4Zg%nV$Ofvr_|5vS%N+W~EBtCdXw!IiF!zbhx6vo)
zW<}txfKE1Gx(ja(DEuF7L?zAmzTN8f>t_uC`-=xjme(tR_2~01Hcu+PWX&wGKTYmA
zc4yLGPJ9QZQbp<A&n=f8{`>}U%^mVqZkiu2a%Ij~-U~9{M4we+a(9hKe#!pQiI$4L
z3LIrkcNu7KTV_A^x16wLsUpWSNz|iS*8jytfT$jK?`i7Aqxf^Gn_CEgj%{5g05j(2
zcQ8vke*He6muL&A?)0XY=h?55kWF$%tm(iS(ex*b6TD_EFDQA9n3|fJaF+od6Ddwk
z{Uzaeopa>>zX60)0o!;Y=|^u;P+pZ?Do?g@kqpofe?gZjT>yZrQ=v0sF)AMC91@Vc
zUN15(F@H|nLtgw!7fxqq<Qt|Df{jo4O>pcP&^_Dx&r94BbumW7XV{g5=j_s2Lk})?
zC7dHuB|$8s>Qb?f--SBU=yLl*IUo|iHO69>5cE#D2t#hEFSnL|^imP^lFkjg9~-4n
z;|9rfn|zo(zMRCwV5i^zfXYm--9=i+7RdmsDbBJbk!}bl1sLSxy7U~4S-Io5&jS=R
zcKIH11tLr>G-hnbIORcpc}VZbB}J^5oSJm|A&;n=#_!OyS%EG~ZaazzzJbwirNh4;
zmN(Gb(fza@)D7mP7)kPjL)Y8_gq8h>z3{#s>CL<SAU^h#JNG<;eOU}VbXunh-UcP@
zSDNSs=F~gbrU&e7co<8#R!Z$W4nEsFMwe=-No$->tvNh8?3@KQepUdPL3=P@L|KWT
zn-<X=SyoNtxw@w-t0w(zn#bMgxHhfXDkSf6z^?MsPwo3tZCWdTw9_J|=fWiomp^i|
z`0d8iPv;tz$zgK0#-7Fh0LgG?Wsei%si|-7&JUAwBG?n!ESU1?aPIYh9<u1(_<qv6
zb&=yPrwVqWje1P{Np<Y*%@4V23@#&PsPuqq>|<4sY2vJg(tJZD*q=k~@br<}H2JQd
z#HQ|val6e_>m5V7ef!-v*n`)FVo&1C={$U$u<XG8A$D_(H=@Dhph6DKsm<(M!%f)&
znwVBQl{K))7UYHHL#Eyn!Y#8evm(P>*{;z0o48h*qYfeHa9ykSq?5%7SOZmEQzPC?
zN8frWAK~W#gI&cC`ThCVoc<gZP%jJ&<FEK{3*9C5cA6=gN<a+x3{zH4_AxIBkaOv_
zl;f*HRGY4JC77)m-AE%v%gqY9NHo+My;S91YJ@s!gzifmj)_JyF><B07`TO$Rkl0i
zt`Y+O4_9yb7G>Lae-AT+(v6f1ARrCWJ)YD6(%mR2BQOXf(u0Udcf){$bc1w=QUXJF
zcb9ZOb6(ed|F-RZ_BY2raBO>g*IJ*2r769^;LP!%UU-=0WAGzAmCdn9BXj9bc4p7U
zin2F!XZ9&wHSQYn2Kk0psj+`dyut$4eQq}hT)y5kKYEoF0eD%8nddqssBP9j=zEZ9
zBm%fv&(<lnv`_%akuz~SZATLW!1IHbgD|Z@ngFp2;^*!2pwAs=M$bm#b}ycg)*SQW
z<P^x`Le>`yEi7kxaXu&SF-Qn@kHJE*3gcSN6sse;0;}HUfl?@j&u?Ny`ffo}P^<w-
z0z_+yS{DV!8%$kFAb4V4$dwJVwW!~3McqSNFR$7#JdP#P^$<&QPXE)NnX$Ads*o;*
z;c%<jKihwQ)`gcYjxP*zGTkKEucPl;UbwevWETo!aF5<3Pixs9GhGh_ubq|qx0D!H
zD{+ar#1)aqF%vxOPUcp51tCORHd0}PWyh(Dw*!muQNDZDf5*%H^jjb94*N)+`u}mP
zyL+x6gjdE^44oWEq*eb*@^F_?b+PQ9H283v@OK7=DR{bBIsL6W_^0r`0e6E{vJ)8T
zaIXy>;jVVv;!8sDP3C540p2i#nyYUZL`<}LedJ9BSP27+-QqxScX8YDIAD65WS`J4
z$^6(V;&=RB)Rk=}nizk^;&1o(jp;ZUc3Ay-y=k=%6cQ@LhFzj$ef%^`Qx;Q;fOP$X
z3)t{s<Fjds^kmKOHhyqlS3)k<L^4EZ{Hw}2H)JK{{~tY$xe()<`SEW0O(dQ|g(BIt
zPv=I5fEYbf^YT;9Yd|EOgD)ew_6?K@6vUIE1~Qm4%Ar*I>_*s4kL658%5sGLzIJlg
zRdgql2-Ak%+gM5d43_nTgt$IcT@gUwK@ji)tIUF4gVW6)D}F1xpmGB)Bc>p*P%)Sn
zMrK6wlCvvBBM7?}0Ad8~#L;fOyp|38qEBzmRgOD@%jfDA^Sz+sn>C}QYyEO676kCf
zwSNugzEv1%kKTf(5)h{lFbL)*0L76=9BW6Q!h7H52_3tz3>Co)^8h-pOfvWY+g4-(
z?gK8&ShJs7OXslmU6+s4;B&zwnbNAW@0G;WEQ}vpb<r8awV=<e5wys-7RtHz8I1-|
zP}h~`yC#&!7mM>PS1{rbbofAf6+yq*Hy(DzrlwjCa<i#@bVkBfCwnR6((CefXpXh*
zEI%`*`)Q-*W}`-Ad})H^ZTo|eAJ-MmGvrbz6^cHAa_JXIvXo0_&C{#(QvLT2hWA}&
zhGT3uxAwBUa{edm(?+E?)(m%JXWX`o5p<2FY^bIVk?rRurd(F*d1+_HF$I+{XsVRE
zWg_Xks?)=Jv-78sV<qq=(Y-{Y3qJJ76OMwVClfbl*bC5-B)t5oQQHQs*3ru^`!Yx=
z#4-E~*kcrD{G_K6YM9a5jwO!Lm9EB()f<zJSx}pK!6%7yf64k3HGJ-q|C(UlI*vpd
zdbi$2R{wIlWM*;Eg=sB*`)KaUWik5eDqahZ{z@AzX19_fR<9|PM-2Ou{!g6_-}9D}
zjKdn|Vh4<<Rt1~nyOA`+T`oGc+>TQO=%De$3~C2cgm`ev-fGbx>WuJns2DGPMkC_(
z+|M?UY@Qudz^@(4!?W8Auh>4p)mX7pwVQoIaZ|<(#V_MI%9KGnPYlg4wAoYdVp}Bv
z)g`rPq9HZg3eAUT4P(7SjVLk?6h-K&Dd1gX+B`ZSaX}6eV1gLkkgzI{i$u(cX9h?q
zEG^En>m&$w0eY|sOP-n@-{psD${vhZ6~Yn>ccYZ>kZN<{;?q?c?b2B6A4%YGdUNEV
zCK^Ub##yCQ1s54Hg@7>RV5ccV0f1p^YAtWknCz~w+U5kt7yC!T%aUSjUaDNI-LCLw
zu6jNO+zoI@U*NQr6ce=%^oklw?h9o1ohKLK0O*<qQc!v$x3e|Al?BoE@x{nJRBNuw
zH>8?e?egEZ2y0rhE#BYD%YAqv=D%R(rm$RpgHhmwuBtVCb^I<2{WV1S7zogq_zcgX
z-$he5<+Xi3%LY?0&uXxm`@LhJw0ODf<$Y#MoBty2wEx9JblD7z&7x(i`Tb^jtG1tN
zRKf?xmeUcnb<6wZu)n5trAfUf2u)zjbfAH0rO}v-X^j>3sZ86$*K{evnb)*uJE)u{
zuE}tz1oy*|t?*aDwiD!o9TRjRAK7^t#XSS^TZw&JP;M2A4W?lVb=GJpmxO}@W<!q{
z>zNf5Y{>Uf5pNMsXN8Oiawo4iZ}#;Nga)GuIMc#S``XXv@8cxCNYmuWXIgX|A|jN9
zd0$y0%T)~bcMsxoEN@u1-@Szo+Btg17Ck+!`VZpjzo#VT=Tp!=y){4dO7c|~gYIap
z<8!2lFbsjUK`>0?;Pw~%F@PDsGaILv*NQQ_2@JrSiY4NU_@|TnIDpByDVrT9Hw)N(
zDdUf{S40xUY&^Ddlv@8ZRFv*Mpf@2>qmkXf{cOVO#doOG05)7*TRt+rhR25)kr-dK
z?mo$bd^5*dmLei1vB^X%a2L~{p)({eBTMv#NN1&}GJ7|IeD%8;Olez?M}!g*On{T5
zf8}nt_)a>QK_k^I7=ilUF!tIK^6Q6W2-dsn8O&Vqgjvr(U6~E{Gl1$z5=)GUnj{}*
zT?r`5CI)Hc#i)HoU5A3UrB(Nrxx^!BbE|-!90`5<lZk+C)CY#9f6!-o+fzH@gZ(AO
z>^P5p`VN+f24^>Xg;N4esycduDGFN_@+<~NYh|>L7k-g6dIMI_rkp}{13%aw16(bG
zHAJB1lQxAe{>SJX3~C5%=GZO^a>%@-xXOR<3X;Cei~1w);b)AxT{T_r7n9@L+<_6j
zTGnA*AG+T{5X&5?(B3ZA^$jhysC;R?lNxdVVR6uMVR3cYE$2KyD|<WJ#>wP=!``qv
zS)Nq7s|+BG-nj5PyjaT)>ypG{-H`yYA?ArQy#9#%GlY=lX%`+0*y+p5_KAIma@UbP
zvh*{!$D}o*7{uZ2Wlk3-QtSE7t*_X}wRg2A7k_%syCG|K$yZW-I4*28Nc&#9|J9Ft
zvtc~Rht_6r{6YP~Tm3L_Op_t1EU@OluvqG9CwA~z;7N6n>Y<>c3-Ek{XM9ym9K1Ci
z<;Hnm>vf?1qYl1V8O#c<+T6j^oDG-2lI)=gtx;{hmF0ztJ;8k9CU2j`6y)T{|2zGo
zWnhSBl={$9Y<cI18CBN0pA$p-N)RIk#e+eCi*c7yOzI0{r|<uNe{DiiQkq(j-uT3X
z!CPnNL{{a<XKm`(4Bv+Cj`zty*$zC>c<Oq?R^`9Dnbbe6`>7d5AT_bG$CT*vdUhO-
z>-5TE3JQ+o`1n5Ad>Vi2_G|`TOki*_wzw+!J=S7&oF)5bTR<|)@)n?Zz+}sezdjR+
z1A>U|h?6?xAL5m4odFNB8@wKHLIXUnIe_?R)7O*o04DnvkKSZ4V~JzEI0~+OC!KY1
zj@u%(+4${%5@%^(ID3+msa_f6KIcsV5NXeIQ+|eOfU-c4{8sS@AzrsE3PZw(XQ8>1
zj@#8vk_KaI-1;zLu*a5P3z|8^z=D6`a}LBHe4*XG*1|0b){KGv$}4OvG|HhZSX@2Q
z-eo-S7Mckp-+u6M@^}7*-p})LN~r{GU@GwFWTzy*$u8WEmvtqPtRY|IK3PpS^9piK
zUvZxo1pb&fcn|RkC_7nO*lNDM8uYhVxQLwHyU~lolgxY5t2A99%Ve{n;e7$KWb?m!
zxDWcb;$BBrSBP-pssF9()xt%qeYXhmoMN&^)3WZBbL%~q=RmP&uRej7&2nR(Z2*e?
z!YH-OdB<4<5}Fo$dgO4ot2E)qHBvN7Z}Y3;{neivu|G`yM;U*&HT~w&eS~9rVDDd}
zi>K9yCX^dLSW8mdrw0G#_%QWR_ZNwwV@Y}KSu~EB8AXP=XC%P;tajMU*L>IwJS@B&
zlP9Dkm;Vt>rD_>!J#vtCrc>rE-b9VGu!k>Ntm{m;_F+|Rl)ujsRW_iA)brL+R5?Ra
zWNAj(Mm+mCaLfd)IjtJDZih{A3-=8Bc0OevpJk|Rf9F5&eWhVKckqR;I(9VL*L<%m
z?(XMFl)WL##G^80Rfc|)jA<)oG}WhX81sei@6rheLdEjC+hzT@FLpC-i_gFB;~|{f
zknhF+C3oPoxGUg~#ImW`jDA-?zeCA*^Heag5#16RI6zbaKG1H}(J5iTO*;X@Y%fxC
zeiq{Pecr5}H8U6HH|lm$0z`)GXV+2T#UrI0o5_rTk@c-CVvn$vr$OkHiACF^Uu7bi
z^{*d&64|Sebj=OtYw8G4NH?LJsDp?hdbT_RY~t9}r0pA;h%K+Hzj|p1lLO}9GLZbj
z>iNpKFok+1fccDT<-KQ!x8$kUnSD$osA6kv^-TmTDWB9%zjhBJQ~(8V5VwcUX2gS{
z;2_}S29Mj>_;aYGD&z-i)RX!uz4>O!DLwHov3f=$TS^DB6dD=mYT_Q=xg{i_$c|{U
zbN{W<U84b<cUcwq&#R29Y{%aHf(DJ$H#-Zq6D4T<IiC>>s~87^um<#eO0NVMVCg5H
zG~?3d{gym*bl;`#j>}_^8V)*DV`^5EJ*hVA{j;uSyPCTGm*%h;I^oo);CF4m04o#H
zdFhX_o+=---Z&TDY;f2tGA)8F?<^nW?`K_Ze<0e;Fn4u_ibTifg-y^nstn6O3Bbz@
zNz)`AeV3+yH%Orh&WSU*u8`i#;kTamlE$qyCVmg{ZQzHy9Hwem|Gv0d9H1-FSKM&Y
zBnLgeQ5U|4K;wKKKT$*)nTs)W(^cB!d5>&Zp>^s|Td8V>3+cSc`niY3GUnZ2*2bVn
z&#pX7sL8NhS4SCZgO>K0%cMx3EY1%P(S|M_chdn1SURHE+77{?(l$t8txQ1)W+Ijb
zGnKtw^5H_UwfEr^Gk#wEg@>v;4^aA<x+W7&#hEARAqeI_Tu^bsJ9YK-NnQyMxxX2j
zm|&NYk->d0HDh9H(y3RNEU0$yFq2fWV4f1GA1^%r*vs=juIi+MBKx&mCG7udfacLN
zKb+EVevUozI-j;*52CSn!dL+eMl*iEx#N6KY$njNg7X2}E!Tv|ogPrZL?)mU_&9n)
zo&o5p(}u(rM~99GU<qJ{U*A8<TF!U1RmA7wgy-s!x0{lhk@~EDlsZLPEE3Wd`7Z3V
zbhb;S;sN*zd<|0yi<RIJtaj7-2VN%tY@p*-a!Oj*wQEE$!xOci@sY#uNn-|Xx?3Rb
zz7K$3rldh1O8<-Zv)7;+p~LTQM%l${72+113?UAY(MjyxPnI>tZo&a1xhzTKg+17N
zkcrM!VPQaetOY@I^%r8CH@7>ONA)j8+HqMJnv<S}3B3}H{i!OlAWv+HUu>aLiHjx)
zzGLfN=7!oUXY`Fq%$^&6cDzH+hN^6LJV*)Si~0+f0Z=fc(q@S=FGphk*>T;!oIabH
zenw?(zb5a}sGz|x$BlPz%k4R^UhITYsk<GOyVISG!WeS=`CPnz4OD%EY!ZlL{M|;+
z(DK$1RFic5390s{EqA!FU5^0rFa0OLH+m00S~fT*tf8h6j8WcPDUsoD(Widxsc(OK
zA5NT=ok)9zqJ68rAj&&yo6LOMm|MQYX*PCo$o{rv%J9?3CMx&13GvRhGZ}vR>bDSV
zeQj7|6%!FB<XKmZqB+L;M4XWq5b4FfeJiLz6e5pqoj5Coy_gIN1;F-jB0;~o-+%iX
zj$`q+<oA^_YdV`GakvWC1h;vNC*&yIiUIK;E|B7BZXk)47nf?9Eolf9;qIlc=JPdp
z@!-V+J6MqkQ_;>#WLPyYHRU$6a7{TqJy&syJIIV=rxkb8%Dh?~%OdpZ{Xa-)rYiW<
zoY!tC?(%o_G<J^#&-|`;DOnL<;^1@<+o5(%&`|vw;x>4jreY%nr(qc1iymu4$e%xE
z!>)_RpWxq?=7yG}_}&`;ehQQRIPe#D+JYEiPm!VpAeJuq4F-fyNb5I3DD~3~>$3r5
z*L5weKB-vMRbeiZU!SdZj7-wkRQ|K8_$}0e`vO}y=SUK)bT~RfIVvn;Wqx;32A0ua
z>!PfGeZWaH6Tm((AiQd3eS6j{NSrF|p{q3UEXTP{fo`c=ty@&N;>4IvAR3Bu`+TQ4
zITY5hE-z_bcx*VF-Z9aYZu{n}0~2o%ENlRXFR`@;DUps!E3*nvLuMwKF^2I_kwhcc
zQp`0rR^70InBzO6RI>!A0_$`B00Kb0r8MJ?x;f<G>@p|*Oqs_>{k%(XYW6jR9rwB2
z1$7o72kTO-@LME}))byj!Cm;cSQzzABRbcIn)Be@2hP<oNUjnK>?<zQbj`e8BYVlL
z0HZsixsT5;t%uSamXmv;VJ_blT=R}0OM-8@gb*y)_W68IF#aDqDs{r_&=m9;Zcn-A
zpK+~f-$P6X-ZTG;l&HQzzipMP-Ex`d8MS!sJZH)0gmlVYEhk(o#+D+$yh3I}>^1mG
zsua=DA2VBBC09JAMduUy?agr3n;JbXE>fa;ballT-s4=PE=Qa_k9q#&A{W_eJdSmH
z^Tw*qk8M#8-Z8y3d3?v6^-`M8H_0t%3@S!#VK@0Eah2G8rP<m-=Is3Z&z}VDefc!J
zgEq^F0orKWlYBL`T~CSAqqaY5K`=x{#0lYJuinC~wtrh0<Z#Ck!p|%t${ouOl>aAS
z7gYT18?$G()WU1hG7JxAXJ^M|Q^T=|jit2OrxApegb|M|?o83lD+HwOTdialMht0E
zUs%qkQxysej=#haD!*Ej%GP-crf0fDJtb++!!&g7hJpbvrC-|C;bs8{%xv>?3hEVz
z3QE7dghe!Q0OuuJJb-T*UH~fekl>LIo4ibn1@@hHOEx(|JExv}PrBIZc}9u;?)imm
zmb<F+Ls>x+!i*jP0x#>(c##c{td!|2qZaUj=3MjD+k?nmF<kPZfooK{>v&}i-3I{P
z5=mK;4&1t8aR-^whc8C_wpvkAGv|>Z(;%j;^3D^uqmEjc(=K)vn6fL3M$X}rU~iw9
zSVwo#G^GZXVAv|*s?+CPKzp4?;=Okm;A+bOeb)0VqfM!z8SuuA^?{+;s$XV8tLXP-
zTmc&=jr=8@j=aa0*s?okN+p5)HvwSxdJ%ZY?jGfKlw`N(i4u_<fcgdogrIfuRVsgH
zJQhziTh<ZSj@=Dwk?&qbefZV@s>M?-^;L)?9<emG^zxlxtdBR17Cdl@dz8<9HUu>^
zcP%xg={+nh?4AxJFJEwk&5Z~bi|(xc`i4|y^40r+>CbCB9oFoVlhfvCd&p~A50Oz0
z;%jro?0kxnC6UwdRm>@}PS2d^>xBiPL2R_35j9$&jHu*>KOMR#MP3fX^_Mw3;<3?R
zBB8sLS;m2RTd`l%OoS_nq4_BI;2;)boIxX}X7#4BfgoaiN>hB(ov_q^z3^}>frorh
z6{WmqKH_*aOS2n87aZOoA>Q4%4^=rhJQM>P*1`k!eC>kucGxsV>ZCo3OyAH=P>utL
zyn6racl>V;8|KTD`g}^=YtE^zVdXDnh#W!{G@zHT%Z8C9uQyB4m6hk?S9~n#uW#q@
zt6)Rw?V(|TCVH~iLrW8l#XtqP4<&*Ln{;SCcFFG=IB8x1wT&}>`GII`b;BBW3TYH1
zd0nwfUHr0EA`n0WsJw0|69jb*=wjKz@vxjro?5i_XLO4n(a9u^Cv|8Wq2`!gUNq#8
zYd<{3%itP;MR=Q->k>3QZCMn$Ne?960z3yFefKk2sJ{Wuxu+*QDF7`U1?h&w8(83(
z8F@9_S76()K;-Vy10ETBI{p0xbbW_Z>f{NTSj@f<qW+zFY9#De5pUw?y6o^OZ2ab}
zZ#c(fsyKWFH{GJlj)u4)<a=p0rau{B`MZ)Z<Kx`Vm0l|H%#ZzMCS+)COWd~}(^uA3
zQwkg4gaaB=5S+1{Ril9NFCZ6ww>{V0Xwk3!B`ppS11hl2Ds7e1X!1@ZUnPJ{oH=aw
zgfp!|8wF3cIB~Op&T!LhRLI|*HSIW!$=5Zf_%NLn$qyR<l%ChYH)}b8Xy@8JT(5b$
zr?|e7#NB5X;*Ju;2PTZ=4xNY({_X*HWBIad7%Sf^3;vOXan?nj^Y~MJ=l4quPzbx@
zO3RT`VaZq9<CV@{eg!&u`clt<C(A&mNqh21m3?_hdHcNXnwze?UK0RV$0J%M1`p3Q
z-<7C27||L0*b|5xNl87<)Xy7iqCod1BRZ-KG-S5h1q%22nVDpl`up4!7;|(rjfIE+
zIlEpNC(<>MgI@fCB3x&5ZgfAEo;PDU--&EmE9{e+om8p`35iu*+FF$_el3Tpy<@^@
zY5qTgT@(aEl@_D<9nYz|)*b-DUBhr^;}#v?)o-w?u?bO$>#jt)VdmBj)8o`yQcFXe
zWFp^gYlf?O=}KH^c78MfTEO*P>+&a9pL8R2|HQ~IO`~CFbLZE**5<N>P$wMnBfK6Z
zE&L<eE!Uvg!u`Y`AxP_Wivb0uk#duJBx@!s{%7+PgU;}2g8|1Nc682nii_BG(Z(-I
zxvcCF{5+j8yDON&D~vlI3&a=i`jQa5(%ib4t3dA})Gnx$UU!@oXq&r0%fpK$2Eb=L
zBj1)W9t~l^;z4iS=~7yDFG4Jb&)B1f4;Z@{(UN71@NFruYf21XUvkU~X^-`S(|Qc4
zNLpyM%wK%n?7exTf%hx}pfe9z^qe!u6JWSW2!#<>VrOd@H72qwly3?s8=o&gS*8D0
zG=%Y#)|gH*10oR72)n|P2}XdD|5K;jo3{Zu!c@rJem_(1`uQ)=wBN>^Mw<hV+o=qn
zAO7@BTK$Bn<8Cw{6j=6BTQdJPWhhQpXq>-FN^;gD2n`A>RlJ}p;~&w}lwNk+W4j5{
z)29=-Y<z%5?MZ!?yXbO=KlLb!bB!`8Z3xdf!sJIshKFN(lC=4mz*{pB!Rfz5nmhnj
z0D1Fj+j^mfLtV`s&(nv(o5VW*Q{xv3_BIeIWc70d>c(M;Rj}{~;0t-m*SjZOX#kI2
z@mlvO;beH98)V-t%$;5;9fg=*LWiGN3602*iZw5_ZqUYoHo3XDY|!Zv6MzDLD`@MD
zPhj)L`#PEr8MLu%rWQ_?crDGsXn!Zn2-`2dXq6)8agmo@h=0ldq=(JnMI|M|nEY@;
zv77s(xd&~!x3;77^WzaiiAma@&#b}#6q_lB{aGo1G69Tp%REQH{pU0F!B5sp<`#Fw
zde~Ws9kpFqaT0+T$^zDYNVnO^W3!J`ZsxxHbAt7w&ygr~k6JGnT>)Q1O?`n}*)s+n
z^=@-s$tF|_3;m)1q1Kyw861I}U7H<jsk?9qqc7&4#qlsj&#s^qO!&&RJRWHOwWd2x
zvwm_@M0trWN@k}FboA280iPgDAwy=)w4WlY3W^L8@@fmGupYdm3ScdZ5~<knx>H&V
z_6{@g_T<d5Cz0H0N5mTRdcJUjci$wnJx`ho$S5UN1tvocaRmga6S-2WH)yLdt*Uv!
zo9|eFKY$}8-{jX)v29)}TY9fFkGgV|3p}@5n$@9K9(2of34=qIzXB&BIoe;8aZFG}
z$~tCZNu2=b5I0Es5+R*nSDtG=2u7p8OL}Sv66|)xU(pdM=%0)EXJEGh75>J2muxA0
z<B_n<PsPOc>Y^g9VwY{0|Gei_isha8RhL-n>+SOfAQZSI!4Bc6CJw-49oGQP;j|~O
z0ld}UPbGjg+02TsKzK{89%1)J-PHx63<nZfDP@S+KF<ST?A*ya2QcLm=n46vdXl*2
z{53eDWnllcX*Eu_ZRs+*t<!FGP63v(&*e)Qzc|o>GzkGJr5}%J<S0PF&AaZEMWdn}
zI+>4`asKbo2j30smc9SCFAY5mE@UDv5Ed7AOP((Qa~F(r9-%WdYE=q-COoEpxn@IV
zYh3B1)P>@{543GdXUGoLsE=ygk9vO3K#$7ad0dZGX0U6Fh^wssPPCbxDfInD9eM+d
z#(o3@Q<LAEqp#90ERI%I76=?>rdaO$Ouu0J$;~Z#FRq(x1|CXqaB=N$<OV<W>M>It
zdo1a?^0^sfGw<aV(ptnJA~lr<r`LZaR9MpbRhmff7aAkPF;xUR4I|9$+1-O-K}esi
z)?ls0f91=CfHmpBKY5+g$d?(8?=tAD1_hPVXWf&;kw)FxptmA%jt*CF{_wgY0jZZ#
zO97I`*VP97=Kd1^3vyv8;O<NVvj|0nLJ@eV%kg~>IK9SH*+c;R`ki?F>M7j@G*u(F
z1LY?(c597(pDmj_UJKF-Ck~s<#`2?D)5CgFn({lb#mf>yO+`x9e{{bupf@mvMe>z5
z;yQPOumsXS@*(Ndp~jV(sTPFo?&mC~VroGH4H2?m)tnb8_E(~$3*BU;N(41{@2nZ$
zVbZ~X^jyDlkF6OO!Dvah?ba@$LYM;#Z?ET*sFeu1;jGStuDWLTo~@;+;9xn3Vg(C=
z)Y!RKXC0(l`ZRa#>N=MhnfFDIGJ^@tk)c=^h6~9Vwd)FSSTPfeSt*8v?!KBVa`0Dw
z<bc<#tL;5P_hG5^Pp*hp>XRe0DR}k1Vw$~MPmoN7O)vRX$EK|H13FB2;$QM$lI7jF
zWi30AZtjK4YG>`^uYZ14bSZ(~DupLiduo?=He~I>44N0*l$uA~{s^fJR!2Ns^WzRl
zC^nv_SF1TK@0kikkvIQb_@$3u@|*L|)4pFqJ6bF0*8W>wDQYp6TyXY22-6_7A7`v|
zoJ`xh9)c%*H<D?;cRkOHAEk`$*DX2|H!OT}rN!vYC1zFaNC$PNjC5FrXX$YPw=!9e
zqN4BDv5||H4DXrN<P5U^yJQI+9UP(u{qIkxZM55O#tkjiEm8uWQ!iYutj7zPG06;I
z$x*&6sdo8tkcj;a<C?O`T8qUEznl0r<aA}^simeBLLUT(r!4J^lM>mm6)CQth(+Kk
z_7CH%(H@8;r6s#@4^I_{r}FfhvZE!H62D5CLns`7Zu?3sF)VQ-+_W|EFH--M_SBEe
zL<ONk>OYz+l5j?vfdm5k4z}W$*J}&F-WK5j9S`jZzzTVFnb0lM4ACwqowad>m_oP8
z0VAnW_k?tXDh_mam((9q-u2HSXztIi+VJMGHoT`3Tj_DJMZ23@z?9P2VC1t#cHC@u
zjj?ZQVPR~O3w~KqN>8d(`tWWfk;wp$W3zFRF2vB!AFCK(TI=@d7w>9Ks(~=C#ErpJ
zTJZ->=YNpik#@Lq7ii6$&mZs7SHFEQ@>_fng`{9`q(4ZWenNK<g6CI>$M@q&KiQnV
z;t_85cdVJ0cQ@Jf>K^R;Ho36105zW<2ZC+;7UqHYhAbCe2QqGuYrhuJO(J{aRLM|^
zaJN6V6a2zYlqo1ivUMGxUI+YlBZlj8_jkdz)1P_)qT^z1SL*+vqAot%`18~7@Ay8;
z_Abs4krA3F@K9Awiu|UEF#uBHuB8z0s?cb5-T|gO`2;rN2I;qiwanf9j4aQI-2NDg
zJ~0Jn;4P(&m#l>|dGUykS9wDw(dl)NmuRYrC^y{n)fD{??5-+0+~neAgMl4hsn(5r
z6I8~XCs`e(bK~P|m?JY|Y->~IP9yHNga1CZzqmYMV&X*dBFW~XQ8x`jpA*<n@c&k3
zbeFLBO{e9pVv3CS4}=L@ElkHPm@DQ$Q6Q`-Fe?L@_U4@X$6K?so5AZCEVumS%R<L!
ztC!Lu1pxATok!5V-`=ytZy|+PGB4Y3>mschpsaT*nr|WM9=ch;vk5p!#9LC|y+S}+
zl^pD7tpaz#%f%Lv{3zD-?=V}ObD<>a7)9+GHet)4t|9%CzjJ(4OL?;=ojbhAE)lI#
z_YtBrQUiEoyDSPR!cWgzLBUUkAz9B`bPmVk*n7=@vs^$p#f9H>Er4=^_c6@0U~K{J
zivjn0n8l0|U~wYVxa=8OT0SyB(k^B$*tbWnqR$DwX0aRX&n@{8R=;3Ql{&Kje!XB#
z@Y|_XDx3!Lxa2uq`GNML>k9~pq^%z*mOVnoiC>$8&)YXc+5eU{5NyyEJ=S4s`nz;w
zuxC(EiyBaEHb{Px^}JM+J6%t>xBjC>dKvu@MbdYK^1v^82zV))FKsWRO43VMW*a63
zR%iT4H{!@Wr#BH(Unhd(Mc%M6_8n*jSBFn+N()WtBIR`yjSW-Z*$8|4K=1ECvo?zl
zyEN<7-eHW{!VZ_Ei`Tpj%U|f%GViMW*~?qbQwi&W$7a;36Z<?+63<;)^9EzMVir_&
z$<Trt13r$T7pr#L)vryTa8&f^D8;#@(wuD(--GL?2(#BWWT=Ls?z-tf8wI01DS+#|
zH|82!7T>xUV1uYm{rthQk<`e3A$xoeCnw*n1Jh?#Jc9&3a3?h<^C9uVPZ+jFyY(Xa
zG4IFLn@mgtU3Jj}!}!hle%#CqEpnxRTlL>v?SI$ZXN9_P&y$T_)+p&WmTedzJh4Nv
z^wA0&>)RLlj>-D>0A9#fGa5*7H)5qhYqa3`rKXKYMZsIeIxQd~{i!YR2%-D%Hs#I~
z8pPu^iWOzDP`yOs@zuH?f@StkI6V@og$j+=>8jG{PSafP`?F&+@R*N+Xo^v!0bp%F
z@#+w-E);hm@*uw66#nx_CLNhu{9<yf9n_dRjSKR!1pJB+M|ac}t)*Gu^9k)Bw3B!^
z+<1X2MmZvMNYj_s_?V~qF6mHeIV^$dKn6^&^~2c$zM%iOZaA@J#dvh14&?AA!mXlr
z&gVMSx*5VaVj5tCGjlUkv}QR9Tl~g2BDiaE+|Hlwfm|-(<|HHS)7FhfWJ(Q~tt%8&
z=4(79<;j$tcQcqf>RhVFez#2hi`D=eipL+C&&p$O`$e5i8f<!jvM=?PB8)Y`8{ixf
z!hiR}U8%(o4fFF0sH*QqhNUy=jP@~De)T2%>gXwkT)$|u2lolxpT3I4f0fE}_7tn;
zHq9+2sK?CS$7u1Qb+2{zH74V^zxvnP|3TLHN6nK{Q4x{eLkv3g52ME%vs;;`mfhW|
zhfS}X-1D{*%X=QrZrkDCcK`Tj+}%bemR;=h=Ott5&dI@gpLp6GpKTkSUDXPeSO)lr
z?G_L6Sv|4mIlRT7UPPiv@aTth>6|qz9#x&{Sv}`G+mn}>ohZDT8Jsodjz?FEaJ|0U
zUJ9eA-SX_==lG~msj=5GV*h>Nd@6a)byxj`q$K;MsYYK5w|A~RWb>avcCCXg?fGBY
zh6#T><n{E%G|hN#qYeovNA+|5b7+>Cm16v5&aGE=AUBuh<1hd6vwFFy=ti?rP6f=~
za`tp!DR8k95g|E^-&X7trYK~YY`htT>9|CoV{@lt&~^cCA<_|eU%Y;kd<ZjjE3Nlx
zaW;mlAPEF8<CE3rzieZXWLDp(t~Hklt-E{i#90&g{rL+|Z^lf7z=|(F&Y1xSKRGN-
zT_zjhnY1t|H!@kT+C%ZiFK)0`j2k)mnz2zvjE&>oK!K8iOna8nQqMMV#}2AI#+90F
z@iMH-w@rp>qdabv%pevo-i^<x9{S7E1r}+iuNZSoTiJwrp2r;!K(y8Gd|YeY2FM(i
zo?|6ZbhsAm+CZ!~1Ai(i1xA;6KhJ3)DZhpcFrK$GEjW63T2W=?wI8o1^INf2d>#8m
z!m{-IhS0-orh=6UgEskLLsYLhl1uNLI#%-LpIb3;K1qgGqry6nrv>u93JeKet+$0V
zC46SgL$-3exAf7hBGPF^88N$(Y^~0bA0aIBZYO-LB@FF1=Eu7(m5u(m>4-MX_inCj
zB@ZL&8!W=@zZvWY#jRmpA`nUQG>Jf`zkPBK#-1k;lWR$S+exuvzVLa(q<Pg_-L4Qk
z2I*U)WTEtoTy}eZ?RIf^0^j@-hI-+EZ)Gg7n8oVCj~}qA&iv*Y-`Hn{oa9Yzq*x6`
znUe%HsAq)V`m~_PbWF;xYGjoAjUQ8DcO9EOC%?fN>Rw=>8aHc@aJ7`S^!#xgj+1RD
zZQXs^P^Hsov)wGYz0>;4g~>Ty>IY91)3-<_k%#OuK1DDM23z&tbw)9UVO;YVt&>DH
z9QtfZ(ZLh&-m-scMaMBGm!`S=taI_<tbc=K@D%vMs_c)sfhW8aN*CwItZSQZOzT`l
zbAfcQYd7Oy<jo+tjU@5uNB>}9kCgyB%p8eu<^o>VBR^))*;2Is+!a5^={C`S{Q*qo
z2D!^OuI^lQ1cn31V<LUiSx$1xjY&%{l&uyKCO*G-SGGQc%99u=<!h%~BimciOZ7S6
zdv{>{@&~N&4PA=i63IZzL>*OLL4a;RnDalYtgjgne2hdye;n@$0VVfyQrAMO8(wMg
zs1&Ng&Pz1XhccIm@e$`Q@jL*bG0{bc+>u;=g2GIm1g{FmkNEL7`*$`$fr9xomZ@y}
z9P&HwWc5dWw^p?vppHN0N*AB2Nb!*4+)2<N`RJG{_^Ub|_j>;Hobq1K(cv{8-{8nc
z39Y!x8FN|tMdGhrzFOkfqF$<ks}~F;!al!#BX?q<Lu1cRj%j$4UZZ`Ji4sKG9d7a%
zh}13lZ2Q3_4bCGa1`8gN(M<M)n7Qx6!IJc-%yz)av@gL{e(PfL0?CB}{^bZ!mFMn<
ztBJZ+Hch#^8lN5AOw|QxH<_b$P+U`;4(kX?#n;Pa*S){tJ8e5JBsZ^Qm#IrultpCr
zkibpn*+{donv8X8_T%8m7Tdz#lHFyRlRr^HDKdM?@;lsx9ech@0th_MiM+S_iYuPK
z`EhKdrwtpU@}Zed5cdQLxNAPOj2QX&_{S8(;s?{LUBmJux%pk2dla7S2j7--NrY-7
zqf}zE(2k?_^n;}G{~n8qigzO0m@1RcB8aT-b16dhYZw6)YQafX8}Eh2F{Ihf!~T)m
z8*v(hi)S3urMRRb@O3r6e;EtB^?mgl?#dYE=X%FZ#6lwXUC^O4&+=+~a=T8ILAp%o
zw!uUG<nf3qO$p^9Jm=%x{pBzdOuC2}I>9K<%ZDV3Y@wps{*>GJK?FDX3k>&C$$sqm
z_i(P;XZk)Jp*SRldcQL+Sn?<{ME;B9N2F>7lcRS?Uf2H6WIo20#3;((4`^l)qv4}Z
zcPQ|Cxy5C|hU`cY>l`#?_Dj`M$-g*&_YO?#0zcu;*|CEO7&QVY#lPcvxgQGC*Avz*
zzi2F2RCc2Yv*t4r%Tfq>dbOi!TnQ<c!1S3%^e)$Ve-bKwPDD*JQkmZY>)&>-ovi}0
z$J*g5eE$xv5FUrO<yF0Cb6)2=&g}>0K$ZI^P4$*@Pt1aSwQ%!vv9!}H$OacIYP#r1
zhT4CRUe!U}m2tl=pO1L+NZXJ3{PygXUY@Y_rv{=FSOb^$-j~%6V!NxiTW*{US^q@z
zcv@TXnK9}pRs(&ooL#@Jh=btlBD2t#=EK&THB7fqoY<m|?)^`>OT0L97q-TmgQj&C
zOdX+9+q>jw**K`6h+y|s3;HSBfM1{aayk$tv1^N%c8~@_HcF2B!mCa#&@@9C_CY8-
zdQjW5o==wP@%9Uk2HC)p-ofaW_Imsw`!Zz|4nTH6?3mz7_+=iJebg6bWLK3?*r{xn
z_w4$r$c?Q*L|kO6wsFN7af+Brt8=gMMkEZTw)A04Kj+a}cDAg7<93E|()bq~wob+W
zo{HIJk4lXy|7pEl=bvqxot<4Tt&8RJv2AqMkd6e?@sqv+nxmN>SyDjl6d|5lu0-U9
zMi?2AH4s7#mXC&(R%hel6v8i;FzzJXd9y;SC-D+(PdFH5UGLK8nZVd^$D`PnVxRoQ
z_*V~#GF`IzV)&G}973QCCm<U62q81e`gwRf8)tLwbxeo1XQc2W+!tG!S}p^|`6?iN
zgM5v)4SQVg->bPr@UmpUvw3w%n(IYauE)ktq7%?%)-Wz{S{^-ctuWfR)gnrGjkD}q
z34o{G70J<Pfy?)&EMoTSn|q7H%=p(6)&`~GNC@@E%KI@O_GjlysDlDlJEz;TydB3_
zbC})KSB2U`B6I#Z{}N@3_EnUSsI2!O9%<o(fjnV;4%T>oOXVO`cg?R}^d?E9PI$F<
z_g&EaTTuaz6!9H04Gu7g7EN~^di?U({&zXp)}W1#JR`(4c%|MX+xwGSN}R+TVA3GW
zIS)lPU#^J6s~kK{U1<ONG*GslZ6G2QzdcAO!Twh|+lpo@DcG6%Gb`w(G4sNQ4>Fq|
z^*#3<xZ&KePJKo4uq@YeyxP5f2RiJ1xMA|qmaT*RcxOTlpvXG-H{bTNiRo@jxe8`j
zCV}T29a*r4g%X*Q=Q<aihZCLd!X)i9Icy}m5_2_fz74%jt#zM09IsnH;N~t<tR@9N
z9c+{)#4f0c5}mUf6h-0v;84M_iwF~zr0i1Lm!HV+<{~x!I6(Vo3)3MiWjnG-V(vP4
zCeiR=^+n;Z)Ah<RU&9Rr$t=H%M!NWG|ENd<<V*p##{a$ND{7tnn_hY0fA6zfrCQOM
z`C|D<aFcu}9GrHoB~PcYg8VaXR2APFfChAIZ5gra>_3{>-BdS~Bkr-D?6k$MDCEl0
z7*fd)ZKpY)gb-L``)Ch34cgnr((j*gU;8(I%u=N#1b!iVc2OH<`5mjam`J1&@EE&^
zD<7*4^oK{i(Z~_0cA*3Qns)_!+c>yYpNMlofRPUv?DbFL@$qQ4edlQ4#Q^HdYL_E_
z`6W4*ZM};(8e>P*f&%5ezVSs@yvGpvEaJ=zyvh)$Kkk=CbCO+}Rk}5%$H*@=%O%N7
z=DMJOeyOwi#CMdIde%<=c)8@rS9vfNqJ>6dA{j?y8&wX=0cA^<yOO-lcl#@_7v66q
zD-*3*zY*k@)t0&%xHGZgjDa&3I0~oRx&8P!%rf8a;J6mMKP=m%YSi~!nw`_aql9;C
z>Cf)*0dz~aV7&O(Brf&pb<PK}t_35n0`(jjED>(M_46b^CH1HS)ANA*-HMFfev46r
z*yh|1@BTYmXCs|-x1~RsKIro>1eM6wl3l&!vE05*RO#-z7}M0VsQZYZ_kSP0O-{N^
zO-((2xNCdx@Vs0Owmq2pMCSugF!ODk>@Q87ZQ3BlRt421g8*FfaTHUsij3ELJ8fAf
zxo*A5eLMZjfHf1ZMd|qxl*M@x)>qGw3-KKjVAC|KWV!QA_t_>FL)v~L?zN<|>F%jn
z?y6?Uq<y6XT*7b^;FWJZEm5c3Xd4qbsdCqR^%vcA*gH{bYWwpOj)9x-2a}dY$^bLH
z|6OC6M%emp8~v34YMGn8>XGT`w+|>5;8=ijpsK;%eC{&a0cNKXt!Jz6c9K7K*Ae43
zvlQ>H|6$oEfhA3K=;SE|5iipWrH+gvUp_y!wqC*dD(umYz<wrqT+)yt%?!oIMN(#{
zufQE`2}U+?_Qd^r+L_VM7BAmi%38l5&7M3QZckkj=Pg6^r*RZm<W909>TKJ_$6h!T
z1k@oEaK0quHQVCEOE+2G?ge$8&B7m|?Shl@Gcw0+N{CUuv%>&J2Wc&w*OHMqSplqH
zh38*5Y;^zAt%Dps_a+hu7i)cn8nh~_e&c6y)&h8<W9+J0I#)Pb$fe-Ekvm&^T~m@=
zeq;Z|Yx@Y@dm~Kcjq@b;rg4Nhna0icv=`py?!q=NQhZ=H|9+JPX>do_eFJ~F(MXnt
zW*mCRvCbJ*PEZpjD=ZRkoIAhiUv`KlF|;Flx9p8{8PQjvwVl88Fep)5?fdp+74Gr^
z{g=+~+;dE0%)6S>B|YNmI%+@o40T5P(tv2dXbY^0?>UizTS7N7Xl=RecixR6j%;G#
zur>&<Jo#Lj-|)AntLTkyx2p%R73rFLZnq>CDI5$+c7v+QwpSV-zTNGP^H{?c#(6(6
zNT5!Ge68qB;L2+SzCho-N2e5r15ziBf)G!}?SZ6B%9bW+RnXfPzv=q7OpKg<f3oKo
zM83Xz1u2##7&sP-5I3!zJ&c*XcqgbeGdIJ0F?Zt14g0s}8-YN)U;Y%yyHlh+J}Ua}
zguaB@FHH>8$ycoXpJmtzJ(+3ou9ILQW)KILH~$SKBy<2I8LB&NX+iC^br!db9zj(=
zD<$xbq2rjW%=T&?qYcS1dHREDuyn3_;@73PQ-R$%uYcy~h-Ub^N8=Q;2E2vq5o7x%
z-htv>FX5_?l5$P1__9T&?)~b|2cf#USM1mo*8SM~ltUcg&0Xvj&pE9Dm!MO@EWA~~
zSS?dRs!Zmv|KxV_UNvB0j!Qfgx{pmBTV$XYF2`?`=J!44N}jBqIqPu9^=-8CJfDg>
z4}4htYgVSn(w)?SW~(tc-Kw9humY)&r80fHnk;#e*Z^SUD_r>CS|T_E#%tzTCpHrB
zw}s+$G1qQK3K+6gxyyEG3#lDg+?f5H-#*)7s!>%gJ_u+=@6@IfoakeHKKa|o<kN&^
zq^MjC=)GacQn=`oAn0^zmOrf2zP<XDw2NyOBu}!*79E27MA3?^yo)BVFDzcEJ^w(c
zjtj3#H_w?L@&r=<EsOs7MaNb4M(V+15@_jryGkbLOLLE6EEiZ+th&g?6=Yr#4Ibdz
zVrjlQz(`5fHf^yct=l;($)F5*IjnlMT)AJ>-D0~USV%}Q#xJU%vA7xAg<Ex;^SX4T
zcEhtBC38_WHZ-5)*4MJkXI_Oy{-rc`BxnPJaf}Y94=P{P%e(r$CytV)O|O)cqgZM9
zXI~P+8>xJt()|V?MZkZOO@^JJd({a%9FQPUPuFLtE@2yN|D?BG;$t6okR+>CXlX3>
z?5tKKk)=Pg#d|u9MLFte?ZKS3VH8VrSm^(*Yfgvdj-dAgB9lE|h?wUOW*6^&G;I8>
z`gDYicuOP@VSHCe6ZFrmEd3*(6<<33wr4x0Q<HRw`Y}gO%4F&ACj_K4-r|>467D^M
zjxe^$cw?iZR@q$m)AezT)?tL6<eaTL#XNj>P4L7I<C=V9`_>-2YwQ!IL;wKL0vIw1
zj5z*$+pHd~PNR-_>%Ro~Z3Ns@UkH#uq#gU&wj=Iu$z0ph8m_&lFldKN+|B~BZXZ`m
z&$ml1RzVu6xr{4+isLEX`tk!r&ED7oEoz<)wVhcYN!pXhhwuwx)duLs_(eE!@?>=~
zqgF1*9{N-lcj2AO0-#sG@FEClhqlUfF}j=P;G?O5p3oN+$!1G`KGxmX*+r+fq4%JP
zs9yv2eC!j%0*Q=uWe=A({%pj^fdw2dzD2u?w@zpdmBCjhQYw;{J}nv&;lfx6-cs&!
z<}CH2=oUt_u`CgJ<)%{b>9yAf@u(8F;oHA56Ttk*Y}~@v>H52nRid&135ElaB-AW^
z#qi;`80KPAJ7n4gVs{t9x{yg#I3m90K3~{tS33re)en*Bu^97QH8WNHKqlOWZ_jY{
z6t>%Gtee_j9q6X2ONDa!VbLh-6ij&Q#Sn(UO709AwTR7SjLmV!<ZtAlT)X+X*i&1~
zl)x5;i*({}*?O@hPFO&7Rez-fnqR@4SX^rCYmTVpL`L#gJ^ie@5FyjYms_DFlTsIv
zJSJ6U)cqN})v#tBi)JGXhnSttrlzK5Dzl(iImyHN-$9nSkn@gK!U)-arw)kDt<lW;
znmlGQ`10pVj30D#EkOa>wVe)t%_mAU*Pka`ge9xERn7SkJ3hBO237Y?4Z6uiuvZ9E
zuqvKB8ol!f!X5L=lFFTZ3t*fxQ#s0{Y2WA}0TBW|58aVZ1*rwOf<meULdt?ySD7mY
z^Z(kesc>U|8kjtO!8D)WS>7a4ANVl?>^Fc&^n(#Inc343nu49rGTaxCuCVe+&&L`p
zKr79CJ~a?uL#7@QM5hNJegYeA23mYMecCXnKdG_H5R2Bu9r~`W{rfR=haUo}(_WxT
zA*+9U5k332Qb1`g768==KUw;$r^;a3sM*gLEEVcq91Vg3MQL5`e5P-#9xLz;wQ|hF
z`m?c|kzP<2CXEj0TbSFs@0%cC+a~lq4Lq=D2-ou5`gIVBsqt{fcShR2(|BjNMZH@u
zi+IRJa@i{zwDnyou()Es;NwQN!>!MG$|RweO6h^q1oI&fhcELZdp{1`XwZ0Oy++GR
zo#Zml`zRG=b9-6+nrL7=xxH)If;Jz-869^5{e{27XE&Ky>8f#xz>F<9+L<54;{cLW
zHpg{-xl+F$Op{G*)?(QpIQoMc-i%|dO&ZitY<zd(2TDhsnq=UC6GyAxA(#kk&Q4;-
zEN822m(m7*N)Xhq@&E=2n{5xuPi&HDP^<{I)8WF4;F~APuTzjN21%Newae@+KklUm
z7t$HZtWYku{~*zgIEM{9^I`Qc{T$X`?+pi<Mx^#i%eFAaM%(QXfDb1Ee*6ZfIjV8x
z|6ZX~FsU02aambDYisMj|I+??xkDbUKqZ<_9&!UVuSp4k^bQpj#~)JGw!X15aGKDq
zjm*1cCz@Db7kQ}@0Cq(Ra2%F$Ib-sxIR+@<jik?MBqPYRbqTshvm%3t3vyK_CPJ86
zbNMA*!zHK=m~{vqPnF)-hQQmjt3$_LFbFbsckT{OELkxpchxFnRHT};+$(nsZuzV@
zRF_52k<$Mby8NI9Uo9Xs!=ik2%N=Oz%Ri}^9vXMROy#C=%`zbBuHKL6xl0NW0el4l
z>*jUuuEeV@-x1B<7QuceNjxS)hMh6j8m>#Z6T=7A{c$TM5GHb%nAJF6v6Q{yK6lph
zDw%d~_wwcMGdG?43AgOr4DaJc45ZQsk;%Ub@uAd@FALqa>gFgg;!CGRPT`|k0@vrY
zOhf__lQLtt3%*KomvE>(qPCY{y#p&GjgWLLS$Bsnu^|)6DB~{<`O!H2T61ZkKQwcd
z2h{O|z`3*N$|h4Q$|(nxG)x8qF4^tg4@jO)1)C{c0KoGjCU1>msO%4itXBf`f3;i-
zl`<zZJGvZ)1G0=)$?veU<iM4xC1eMEFU{j@O+nHF2WNuwB*JG__l1v@i+EQjHv<}A
zCmq^qc&{C|1W~vd<*|=Xa?gRw^v+e^>2%Gq8>yXrUIph2{&5;B9^DmgeZ!}xR?V+A
zzcIJtV|?8xyNiz{0Vc0d<~q8iy6!Ld*vh;>B5vMl?udcUuQW%z-=Dx$)Aaq91J=ju
zb$2#N^v9SrvY_~%@rOQ&|DBCeQJqb)1B=_P#2&u$Shkv7Uth0b%tn7#nI6PFMSAux
zKlhy9u!F-yoB&oaUlM_oK;bzy?D%V!x{d4J02hO6-$pv%W1BvipJ)0?`tbdwTp+_3
z|07bCQvC#)mu!e>sIO^*X?GJJ_7(Q9dO5CG7~=aY>~!#e`&HtSJ69ZTS0bNkj7Jtd
zvZ75_`Uv<|%Z0PKz2g+6g6WeY{VZf*rg+O>h5J=`X|NtX7^5T>Vdi3bd1Xrp=ST7*
zbw;M(ub01mLwYOnimDa=?#x}qdKX_tbwZ4Unp=+ba|N<PsCDe{_`gVc&M|?rmvN6h
zJs8~luc-=O`xUS%&Dfg<<&ZBw>e$AP&hkt0>k8GcUONabGDjH8T$he6Ga(t;X}1%{
zF4Oq*g}K4>X!Z>H?qrH~$#2yz+v&~}FuitJ@rTxr#Y|%Q?i*+IEgP#t{R7$->D)3L
z^?WM2c?Yywb4WtxY8EWxdP6{z#Ps_^`Yr0H8w4scLB{4L|Cv5HmXAs?Ht??yz{#;U
zFH2=GD?bE8JSCG^yFsv_+Lw<2b{MGvgmyhEMaJG0tIkh&UwBy%T`seaZboHn;}7po
z(g{~AikDx_gr!>_1WSHE+DJ|SdyPi3Wxly7jN-V_#YPVJu3voM(y6_X`YlM~9&7hB
zmePY_yA`jE303U#^aO4(8E<cK?(6Bw9*`}4IAKP=kVylMJt?iSC-p8-9V5nm`j>%W
zFJYeEP-VCCbup)E#PIj;?A)#uj^kX@ZT@ZhzY2=w|6_rG3Xi&v|Bo<Ttc-0Sz6v#0
z%-6_>ty}QwV{oI$ero)KtRr-PMf8ve;P{Wl=XM(K7a@XfZR2a^EZj*w%iu9mErTFz
z2r5SopBNQDumXL_OA&-7k)a+T34mfZV(<L|t#B3)L;61)<>Ed$>v>dAjE~aA!}p+f
zPQJccNwW|&Ov2eiLPtgb6iD8YZaoXROgmg$G30Y3Q>(@%VLpw>egTy}uI`H}VTqQu
z4ihsj3o$FPDc-UVW#Va*_{%#|$IYMV%u-VSKdR0%9?rJy_EQ+W6TOG%U35l7^xnIu
zQDY3z8Ig*pi5^4=q7%Vrqjw|PkRV#L(M2!WbMNPV-goc$a($n<esi4XI@bCx>yWLR
z5O~{G=$H+!bW^)_pBExI)VBPY%{@!F8!S)onXVpDu~&I?DIG+l(t}{HTL4JnrvTht
z7omD;7w+Y9riZWpnR+*@RTwt7{3w)PmQTm{R$JYgGeH$9$O^9q0mx2XUAe0)IoiM@
zGB!XL^-o_w80hrNiGZzTI;DT5P+BHL96U9+GQ}oRH>`oBb#Uhi8HknIDl8QcQTdqo
z;ES#j4gvybm{r77O}0H+Xr$E!RwO9Vj4|n10-EDaSd6(v1s9@F5{<t_;KhulS-v;G
zt{%anXsptl{o3M}p3J4h1hnz1WBwMzGq<Uk{_Ius*rU^s^jNghLNlBDLD$<QJ|QNs
zA;sn!9X)h{yT}o?N12muaVa}jQn|53E)yLqS>}8I*m_y6R{x&MPxFVoi2Tw=>0tg+
zaam4Ql`BLoV1KMl`1skp@IL{M#jNzczDLa`Gd8>Q_ootV>fPSO$J0VlrGm!)ygqi(
zo*8*pdXE1Q4W{VIPq3I_TxQM%M`fS*@k8Jrnx=|tfpT`s9!Y9T%1S(EO}{I+G=DtY
z%F!T6Gm9mGcH7#lDOE^4+0ib>%tFRNk@eBAt!Wz*U+RR$B`@V5i?>Kq@VStVgC5B4
z`ICv?gjLlywhTdU!>@>vqvM!HyvFtO6~={gLp|{ho1T@Ly{_+{AU`1h5QN&5$3Yk-
z(MRt@;~eVvi#iYIo;h49=T8K18TRjj4(dKrGg`(H-fETO|9H2R112f)ppH%*m18B)
zFOhgn!PHJWAC02PpC2=L50Y~S=x<nW^p4u+pZ%&nE;`oJtJ6{`JO%;)conl;6W4A4
z=ACSTWC6u__I1^>_~I8xrn##F0jt{2D{TItAx0tH$0MXHu~ru|XT4jB%#?$Fg~;n@
z(0CXo{o1idUNuIR><NE7jm2OrX|kKVZ{aA$ZXYulgH5Ct+>w7Czd)HajDVggjEh3u
zY$~Q69&JK_7?m9ch`Z-pbpy2)+D@GzQEK(E4<qP%m;#r^rrPvY?|40Uo-Ig+0Zsu=
zXRvTkaGUm-=3<+9(KqgO#jlU<0$K$$iid!=Fe(2>M;~(neZ8N;^)sVBI*gTl%)_fN
zpvuoXJ0|^k3zc<O_Z3G)9sgVjc|`iX$?5drmr#L3LBG>^F_kLw`QqahtzP>mn+vNY
zdRqFx$TF+JHc8*U6naGFMs)kHnfDuNnZe4n`tUok`vsf{d+BrvyxV<78da5*qey>r
zuF_8y|0&jYNcsQ&<YY8O^rOKm-FuN`LM{)-x!l^>k=mG#msVyXE&*etYJ!?1hCpG)
zW0&#|kc`oRFXw9@^$R6x*2Rz|5xs0-%1)szcYUyqr<j~)s|U(^L9EQAVWVFDGoGoD
z1NMOC&WQXx%+(n`YdMK|Z;~U>rqm$lFlWm4tk+N2R;);L%${=ItBhppkJn0?FDcYB
z-{i01r2gNJMYC4R=lD$dCKykqi&O4Sm_{f+?%tM42geyzZf;uv7{cpv<3-A{c-bcv
zYR-dg;IUT(FRHSh*)4$pWx8tg7jF$$$pLE17gWYNaOo-pUeV3>&BtRBcJ&DW-eIIR
zS3cN_aOcyrnLdTCK<eEHz7eeM*ZsreiSH97n2EN!V^wgy<Q*}}&GOY6{q|;Ya4>6?
z^l4d@($*JqcA-3TLcHcFUjIKDD`YdH96}Jk3E_Z~z;Ais*v=la^=_;c&=;WrWpo^`
zdi`ggrFF%=^y%v`lB7OGQ=kY3jaZ6)OK278LoiQ$*t=acJb$?_sgI1bhcRuzIVXGO
z?771G0PCz*e(k^*vkA#n-~}{b%*?E<`C<DO&lM^Be0>4-P2(hI`#Nm<Y%8QD^!3j1
zbr@CI)=0Xt1HdieVu>wQ>a%;w203UJIM?JW5NV}q@}X1bQ+swfl>t&HYPDVP20yy$
zq`8r-&uSN%Td%&SU!INOt~>1$Lz^LT-C5pA_+hrwTRWEf!d$->s6!Ude-W_2q~IvI
z(GJ+Kx*cmQR77A~tsu?*Cb5q33+EYg?!UuL(B>cP@d&4dgM(;^L0(!(UZnNq)m5rR
zNpeYaG-SDw;DI$mXQ+{CKq}-ts%CF;rSTWLEgM|{xEM$?{}7e<C81)<n+sK9!&UXM
zl6{iog_Qa82zYWaiu9gUP%6%}@XU-=AV*m?+T4MLc$_`bHpZ<DvfxT4vDra1ArmDf
z8$=l`=0{A~)c}-Sm=d@O2c%6UOgPRXC08hO9zlXda!5U5r>uv)G9Zjie*%OM<vwSS
zSwkNAu&P-)4tpc$Q~wpeYM_6>eB|;EC&Oq2UC{<G1n~-hZ=6BGk_06*@Dy|;7}o-m
z+de3EH0)()R(6a=+5vI`+RYNL{bMRD{Tav2-oBsiuKs5EJwk&_W!>>J2sAOSis^(h
zR}IFq;9OutexAO96aI2E&2lvj+p?ZlQxoLJNe!B6@{PzbX=JsM#{`1&lrmwsl{(lh
zuO+-Zc-?2{dK>+{nLdeh6HIQ=Osgf_HY>z-#(lCH@YGVk^Ax$`wGOfDkwZnhp~ShN
zA95L)LIFG1wI$#uxo&8X*l6aXOCC(ln9<wpM3KVK$O<Z=vV5!M9DIh=)50(bG~;~0
z)UXf}iQxz7!g~I(u2->tE^3#@zvA<uTMt<Vp6jf~TgP}Hs{<uwl*-V|aO{V-5Gln(
zK<J<-;b+3vV(1sScOu%iSbm>w(jJC;Dpa6rTF6#Hm2A`+NGg-T2fH6E?32B!nje{0
z+Y76ysX^8e0Y(<D|9`Q`Yj3{!=(lK;)#~{{-o~%%bs4g;QC+6ldkt-h(L#IqbzKLS
z$l}zLI}O2VAu)g#<+?GR6!Dzn{lAUVLVM3VasV}55i!SJGdoKU3YtNAsg*PYLL5Hi
zuk9f_x>N@U$B>I>@Ea?t$mfrtgq|S^QWAh2%4?7Y!UAs;6F!5Aqw|${Lr^po#`%u&
zHG@Z6h@lAwi+?NHD5|OyE`Z4wQ?hP1W1*~=dR)j=`L$yh`m^W<LN1w&*rmSUOg3cZ
z7d40Wr$1LE4n!gkI0VThCTxMF%@=`o$h7s?A1naCc<l2ty1I43fZxTBI63Gzz&Tpd
z<J>J6DIN#i>Ij3gt(47-61eN{9$VZJXG@ulXsci@#4TQS#q$qR7URb`;nT1w2?#5a
z#-c61-|V!#AIJ;sWbuG{+t46kpqH@-0bISkQmqB>MbV4(Vt24b+4LKhi8C<6FraxC
z=hX7i=<(?=U`g_LyGSphfuQax)6!Vf+303lBb`6^JzQM^nziw1!v;jLifvf>y)*+@
z8GqQV_&9wyMz+!7%fyG5L!iCQWF@mI)U+m7Ty3*0gmF?L4;LLNo6ZuFRZ(ngQdx|X
z{3fuF6X0Z=Hfaqn%5pER)@i)$TmzVbpRkAVA_l+I58Fx*ARSna9loCK@MNuv<thI9
zENLv&QH(7^TjlD|C+LhKvt6D=eC%v|^$JqXc-BK>FmQcwjElxw`L7KmXa23@-Bzo9
z&Pt2BrA-==Y;pfb{x(vnzYL4o-ya<M&?)@v_B!PM`uqRpn#a(O9y;;AjMSa4tNhew
zmMlg2tO<wHnezNX9oSdu^N<pGX$>=zE)@9&8OH>viYj@FP&_x2Hx4zH)PmumikLRP
z{*f#nG>7yOdc(4E!tGW5GBHy~m(2fQ<|!IZepdST=3#^k$OCm3K;BQWM4G<T1Q{!`
zN(t4;M`LLXEaY<R%_u=Y1cDeK6shdsPJ$ZTpjrC4;T0@RR(b$l;L;PtW>iyf6g*si
z<W%WLz7-DI-)WD1L8DcTg~f==km6I;zU;b^n|0ZODXsW1Nw<58aLKg%*mXJC0;H}M
zp28uFbz%?BCKs~RECw?-4-T6r#OKAhUptu>sbGnj)+8F8p4|XyOMVe91ZkBOJ^FEa
zb+1I=Gt*QvS8*yYPmzd0BZVJZig|1yC1t>l22m?NNJ3tDjCAXxKqD@r#QqE$rl=<Y
z7>f_=%o-0aq@~1OC90H%IZ5q^=D0-cu9Wg;#0IR_Kb=)>V2ZZ82c{fqmWRQS&BH>K
z(t)Yg7V84jBXtmlZ;hMZMwEqEy|^4j&ot!a;2-c~koPJt@V2G`lxN_Mew2J4n5*BO
zIVj${y<P@n!e&9%^+}?Fr@<JHn-;5IXi+|1H4l|pFTY9KFvT=KMTvgfmVuL30f|3z
zKUonZG4t6GI-krvtF_v2PaDu~Q(1D(Y^rY84;z!s#V?{i)~MED1m6bb<B%(RO{?VC
zCT&J=&Q!XH5m;i|)3B2qr+Q!3lt-H;{r9G(9TO9CG~97LJZQo7Piv01>g=D1{I54}
znkv~TP%j>wG2zk2qS75dKp9Y`oLF7LH=kxNYaj^Op`nMYcTt(sgZ`bL)N1A~+lzgy
zpt&zYSnmlR6p7gIubEiY!V$z#fW)R{zb|v>CwE6rf+DnY@DDGgRedo0hXK!fq)U#)
z7hzmQGM{}>?f1zLXo7+Md9ybVx&RR#KqGlZNF&uAyJss9g7_#KeY8JVegb_W1s!@a
z9R|#tdI+uoy||N(5?l`{5^X9$xL!6$we!_I+|SD!jDRe~INh47#D#P<6Xus0|9yE^
z*rHowG5hMN(b9I6v|gRl?~(4|2)!VE6$>k1Xr_WA&DP|_k6^@PvV9rw2|$mpmBMeh
z#JtXYN)=qNlK(}!TmR)1DQ5E3Pqh*Q1SJPW#(i}pn?~dlFQ<1FO_6Lx+b3(YCXbcn
z%4OTlcQoyVDEIr-r*XQ|(XPap)hde^ec9|)#csGA`&eaFb;Pro!(O?zeC$A==0*=c
z!B;$T^Ec}v<d&aT8?pkY_$#RQ8o#bP>Mcx>mmpA*)r{WX5}HEC=Ak81NqZlw8*ux{
zV`O{iOeH6=-;aDh$WaL>$EbrMsZvD)oWXBrQ~7Ff<{+NqV}@j*bLHGKgY%$I!Hb*M
z5){{6(J;xXRzI3&_g})hE(D){E3>p<jgyOOrjLA^G<AZ*Wn|h$MF5Kb0hj(pVyh-`
zTW4H&;X`lSur~EWJCCdj8cBiw`D|C~P2nT|-d?$Ozao=RR216X-91{^kCX{%c!(E3
z`?>$6Dk9dIl+Xy0<yWzsROY}}RiRL%6_tCbzb-L}o()Kc?EX$*z@{~MrUyz&Z4ip^
zMDFcmufS4WqA3D1evsoMps}u*rHn6!@%b8@!Ze?jm(gh;1)vFERri9n)&Y%_c_GX1
zr_<|J_VM<Mo`_5XvH+~hYd;U6D?R*#oAq6Ek0!zLIUuNiovA(tC%Jsz3+_r{3E(;I
zJClKyNu5)`i76YTRC5$vrCShpHx?x~J<5szrW2g6Lwi?2BeAs-Pmj}LZIGO@<fE=;
zWW1GZ3}L`v)y9!<@SEdLA)WN?#=BpKx5Gqt(-QU=r+jtPl{!sC=WJWZ8kuu4?&iA&
z)t{STGE!r2T%kf8Y)yAz8Ye*zLtDJSBamY<sM)c&SQnW)xxy5XNMFg$>Zo<-_H@L7
zC*d+YWJ~#Es%cf%2>W!Jt<C&ps`gnL5c4sw4(_oh3*JZeo4~$h(o9&w#;vUD_O2>s
zndcP_60b^@2eog4h|A-OoC9TvwHkQa7~QQuU`c?*2`VLLYn8k|-@A!5*#$<`Ij~~2
zL&Yg!XWi*&5A&%YZoPEu5_GkFXso+*k$Y=8zW%J$Qte#@IHlHVdzYo$!QBgXUb_Ui
zPwzv-1>~BTeO^C&40-+79)+3#%Qs$*bTPgmY2JP!o(EOli_3zCZOXFt4Ot((p4$8I
z!&)xns7N}*vgKJNi+m9pjT0mPBO6Ukl>AkdQ<FYm{%@|KTpQ00msU;3`LodNzBuMn
zNp9#+Y;bz)`@zo`q4Q%@`CCLCgG6u+b6~w%_shB|p%=`%Q_5KD!3H5eVR)U@Zec3K
zhtPcC)z6XNPJ!S4gqdT6vLzmkYnw!aP`K{7dK$++`IZZ^9IqjVH0s+#E9f@KvYIyh
zU4Y^Q9yU;)@5rG;lVv>mY?H<6op6qtdRaWK=BHa-`pcPyz!M-&@{SjN%FI%ekMy+<
zfhKBpw#_>Ep@Hf{izHI{2qo#hrl8xO6e!EP`3qYXp!fJJpwFc`>aY8+GhkS$HtO};
zwd0jL6@-p;-ab^=xK%=qp$6EptGYGUN8b&h<U}Pg3v;2;*=o7nFz@xhH-fBCshf&W
zT&l#mDpi>OgYWPHP&ajPKw0xl+kCyiTWaI;bXGYs2rs~Se=`$+H`E-|`?QIB#Afgk
zcBj$<->|eaRg{BtX|x%oMBw=4OK~aOpwU_9aQaZzm<9NNk#7rzzvA+<4wF+BIZC!N
zOJTJY=?~!8CO8e+viBylvn)sb*{nHNGk9?<7S<j-_O5e>xb_~*Ljogv;J=!$Q#B35
zOPfYZ->cZzslkx+J+N@uu9IVV8xJ10zW?eI8%(}&%B7x{Op!*?Yv<}D^VRVkjxkZ5
z;F)Cm5hjZr{FzDH(93Ior~$&=J|GDHtG`>R-V=08Y%_krHcRpk>`6qLefIT=Gwz08
z7qDgQweycrKxQ$x@TR4#Ov}7t{3;PCV0zeb=iejJKMxfxyRhAcJ8!-lRaaM|ZK=>o
zXSPS4(IXxc&I$Ayqg=#g>C?ZNnD3X6kl}wekefsup&UHF=67<cDlT!+TpLgWH?&x}
z^T=Ac$fifB^P{61Pa3x{>D}bUk9WPGfCXcnhd(k4X;yz=E8EkGt3j58S<wd2bSkp*
zbRX@=2kMPh&mS98X@QHl@7bBrlq3Nkv3Q0neUu~)e#CSa(1Vn^gV))rui1N(G((Kw
z|1iehP@Hh9$+A^s)+MYSd-nirh!Cu#T|C(iv>DxXwS%9q2~RUnCI?Z$^p@Tv1m=J)
z5GsyXsXAJ%c^I@Nr=HqyIXmE^Rs+tr6!bJgA;dLwH>^PT0z)bVb$;vY#GiW+(jK7P
zwBn~U>hKABY7a11_Z8&8b&X-I<68XX-;r9EIHDuK8-BV+7p{8v2)!o5lRW#g&z)<n
zvN6yd*PkHSo-&K&B(P(x`8w^i*OhwcvnM^Bi<;_%)U5Pg1~;lKNf!aOiId3z6Y7^Y
zx%HuV0q;MAnd8rcAPL)w<MDX7`qguQ-Nem+t+@M6W4Px&QH1@7z4wB<Pql<dq-v6Z
zkNOBnhxd|wSzswTdYlNCDJc5Y9Zljr)6nR~crg{8uY`EoJLdP<Pa@^gD^O&3e|V`?
z^=aF^{HIi{%t(gZ+NPhg4z6~|RA@`4rvUIAw_@WREcz}E@lp3z7)n6SQgaDwD%H@?
zz@3Z)j0a7W{%45cTqeTEibeln;krOCzv-P{9S(~TXLeg#-%)i?bE&qg_<<7#a*?ce
zL7r=tpe#aagI&pi)z5+h3CzS#UL2DIH(dTW;%GfKeKK#TfPuy|Ye<QGqgAlwgC^6m
zEct%w1Z{aX{rU1Rlt)!Sj5G<Nxbh3HS>B+1w`Sl-_;L7~OG_x!@gYbM;F&O10;z9p
z@-@DIJb2e40A!E;SmMx_zpy(%i#B*!kMGVP-ORvlSS-=1&7>`s2pAL69gMr^dH3iY
zQZ9~~(D*3QYr;>wXiEZvU_J#Pgd9}j-z&&#gIxT+e>W{eUC3x5xVMwHa#o)O6mh%v
z-cAN`zD%k56jXitslrl@Y)5zMyZqhPXY~8K{B_n5$Zu0q!<Y!>9U_kcqyZUHA_(;n
z*X1*+p{dH_la548z6}ABudadiw%a9xRx4IQNE+x{NbOVVHvNj$@oStZ{3D~Eucrwq
zJEU=uJp2q7Jew?ax9pc7x}Rc9%;)lJJMu(|4EVwwf0k<@Rjzei()<ivPf?yz_axjD
zU0$WaL3_~Q_aWj@8A3-%o&b7(Ous!w^lTo(=&7?J&}#;1>JPE{+TVO{-D_?ga4}oI
z&MP$QthoB5t^h4NU&rR|!8TB95b*cFwx)T-N`q<p5mF>wb`4P%O1_iX9mzq~yeT)_
zw(8BwoXj-F4ye{0`IM_lQ8TrXvk>+9KKFtdtY<V+vVVK<`%?7Ym+Pl^3+9DY<o~1#
zPB@K7{w_;{d;h!eD7T?dKF{}_Z!*H6(;Y>cF%D5}^HkB2dSHEx2yR<G+3q3ZUw+RI
zl=ICqTL?Usd=oONJ_CX^n};EP<*ZT(5tJw$a0EOs8A(jbS?R3<Y4s4p^1B&)v2Q20
zGA}{`IZNpwHK;B1g_7N~mmKMTEI}U1f-k-B_Vh&Ga<{$zOw`nO9yR5#8i9wkD<NaJ
z+(P6GtyY2A08!`H;WgwePeNoy6fh+Sh4BL2qSo(*AWhY6XDiN#=b`f<O`QN>(*z)u
zN233IRid+Uro%*VG~a~?Qr<o>hVsd<EpQINHCfEQf+{3uhkjjNALMWfuw*GqK@9qw
zF#_zG;c+|d?cp73(%k9ufQCO=qH1NfqmsC+q;2J*FuD8Pz}6$^P|?}wJm_}@`=Y`-
z(yl{Xar|^gr3fScxVm<JNno&D%>$f1zAXdil)-o^uoypzR@<2-hVFTKy0)I=6DG)<
zjvqi}X{?g&SA`j27PhCCpW8NB+qR3ap!S!XL-+n%uO>{^T5F3Y%t!za$2!<gofg7l
zQe=M$^afKXv7PSlE_gLhIsKd*eF`bX_nT>R-s;}@Rd5V;bp5*pADe%b(^6pSQgpl;
ztm09om?-ORt3c-r;sD=HybsB5%LRLpnp_T(-spPX9FmS7uMQy`Ck2rf0f|-fWgF)E
zB|GU1Z)(NNBO)zX+m!OAuR0XXSzs~?0ja9ZCMEiF9UJI8Q6C>4$a=wlUZDT3U{5$F
z={wHV;Scif<py_cLyR<@q!6x;wF&BjT~V>p<~62OAfTh)gJtV(qJmEQGEHuU3tU)s
zUCLTUx^#?q0DQOm5O`fM%uTsk1NmKK&3pE^fSIobZ@>(AL2CZpuro(9aE0QU0?X#n
zP<YM)I$m!qay!;gtBHpodnmk;bPRR;$aKAM_94EaV^^qWrvMD_RR9)|r8fTZeSlTY
z!;s{xM@q6;tQ_0vW^bqTw{^Y;bnmqZou92FsyH7frhcDr9A3{m;Z&kbN(MlF{n6G`
z8wIIUy@g|?wj#rNwlL%jPFp=ZtD5*rk1LJ?6>00yRoRhYIc;_8p>QoG2KyT8)5uTx
zzSeI~l@{Ifo3(47YFa@gmOY1CkN`(iZ}V^$3)mKbHYozHkJKd2+8~m)9)^0Hd6dnC
zChi8p!+&SH7EL&>17%RvA6k4COGZTZ^tYRY7xJ-UcTgVFR^t&Pty!zunD=9)X}J7#
zFr)<7-BgKqg-~E|v@)8c3gUyy$a{xX)GR+VS0vHjyaW~aVPFP;9s%G{bW`ZB1UC`D
zQT)hb%^&yl_-6~7CN<+qy0>Y!u9r=HLZ@?Jw>VuN5$*~w$%CbBkFmKgUi7N@Rxh#r
znGdqwpvLg+mw{3hZR8Vv!p9~saF3s#lcSf$T7$>-M2n`)38B7KXmOlndBRchii)0v
z|I1VyGZRzFKZ<f+e}84S09n+(Z_Gb`IVZgTX3Gmq9WI6}TyS~K)2GjC_oM#e3RSxB
zuR<T(FmZsF`BoB4w$dC(4GT2!BKH4&z>iMj(%`(1SkC?QL=k_U-^qSPkD5w1FG3AG
zoeTioNoM<u<4uVq@RImTLRYxTGHy7{YKR{gZV)v=;<-vox>skPoHs(YZl~ftm8$5=
z=$uTVKB74z{$!!4N4LZ5DLY)83BLiko6<`$4{utDHByKOpLEvQ)0JkBZ0p@Dd{T`m
z-rCFA>B;0`l}u1z%?qs3oAf4mh%9_M9a3G^0HKA@MklEk)LehRN5fDOQ~VfzjM>ri
z4in&>pVHCU4*m+W!-Vjs<JtgE?Tx!YHNOcU6M!5lv@Sshjmbh3Re;nq>20IRciKa0
zD=a1T1iT+Xn+2@8spnmgPb*<Dl|ft+YI9x)XX7GR>GTXiiHrS?(J3!~Mv?N^L!6Ge
z^$uz+VSOIN_N*9`Yd4|Pd5%cO_ccS6f*rB9C6lGz%_ZR~ARo&ZNkCaPS>Jcp*_?8y
z#tUZSUmNE;CyjVZ4cTr)%{Q8{4a%Z8ldh26^qBtl_OaMN_VN0GTS@l~pJ_CCPjIw_
zY{z}0bYb`{to;ai3qyFw7qCT+jFka<TLFWB_1~fPrRYar6vA4B8s5XZGuJZn!@A~c
ztqv7euK*$IESd@BGoI(Jr>6GEyd9)0L~BsgT6E@5MJ~)1meW795gmJ0cRSKMYY5nv
z1OB31Ftez#^*C8Q{}ug`gq|rU4tYd}Gi6WFeSW3u6a8QH#{}BsaPh%GtDX9yTjg3!
zE@&^OSX7TNo$Rk%hj$!gMJDcQv-+iDt!80&=T=c?o}BoXB4GKxZ{2gOENn#2G%ZwJ
zfN}f{lsSkzf%f?m1XHv^Rj$woc3<iVXZGHNs6=$Ad28Mu>upCM81^%%Dprv2j9FEx
zAqOEybB66<kNN?4Q`|vzvK5x_CcA#zm>?Q-JCY3%XZfu{x?;9@-Fc8H8~o>HU0^5{
z#d6+%v&2yMzA-ky?0EzSJ|ZSG9QBd<JLkm29s~Chd9b1J#=C-p+z`ATkQ72>ihBNz
z{E0p_l-z=a<Hc8Vz)S$S_($iZk6Qgcb_(@~mt%!cC=*}OE5i5reHotn-IAj9N1KiM
zI#<oa=`7CfW4se3_x4$%4hSBLCXo{b{5XYWPR+@D7^~dONxt-E)gOks3$R4Vk^s?7
z$NGwb=7Kd(ZY-ToLFD%V<i|N51@UB^|AKaBHvnFzUc{KAl`H^%1iRy$qhS7rE1ljD
zg;VFB_il}|_io~x$VmLXut4ag)p}nSA*wCCIP&v~Ex0zM&ku4ww>3u2!90K`hF0V4
zi&~L`<p58q8}>^#EYhlo{raYd+BO>$-4g<qcO-e};Odx(1e#Bg3tvF@coK|Hg_0s{
zG0(BfGr#GHT*p*qvVLSMHvX}Ck|c^%)_a(4Q(in+PmQpIkR{ts9>*_VhzIir?Ec<y
zSv!gJ?YMbX&WTxA5dOzDD*-90oY=V+XsBT~7hBHo-vNd@3el*}-kkk<8W>|7ah&~Y
z>Rg%V04O?@*d48=SGo1v1(|o+@G{G2ve~J0OKPuoTC@ho<-<vbLZAa=^UAkHn%Ldd
z)_?XCCQI&;<K1~7!YJmYM~tMLHR};NGuRXJ*G?x*H9`?)F}6#zs+VH~_o^N*JYbXo
z${aW~S@>NL1$GFD&0H}8AlnT>`4}`G4We+@@SIe6*ccy|qDUOZC4Nqh1C<^@z>cNQ
zZRGXglE%7hnnn;uN@##?b+c1B5D)8Tb9ae+tf$i#mi=H`<RCsolHmh8I#cE$ldm|!
z^JFIDbzNvo89+vBRJd`O8;Rk?EY^o*DXGZTBWeI0F0F~-KHUWVgRJ~sdyau~`pj<{
z0y=FyKb0h~drY7lZFQ#<oEMI-NZu5B4Mj?)e`ppi`F?JxDZ1d7y|*CKGwG3z7PVH%
z-%q|5b4A`yA9b{kv$QSUE&k#8daE_9d6yC!CrOFf?lN1UP-L=2>>Qgy`6gO`&+JM_
z&lv;IT~H{G{)2^|&v*l)vcrV2QbI{j!;KW3*IGxcXj%Nty3~niizfLIaT}`U2sc1p
zaryQwo+d2v{x4PzH~WgYi!-&XFVi=Zwe`O0X!6sjD<fvXpgmxCb8~O;_P2RDsEk*q
z<a6{dq)doLt#{&9o80y+-E`urWc0fx$rEX~`}$*V^9rh@Or}@IX<B-!l}foM!otFJ
z9+SNAfs=qBn_XF^>+!Y6!Io(CDZ1*EzjX096<VvjILJ+p_1}>edH*4DBE94K_gXn%
zErDHyt)1flJEa1xuDf)y0zN)xd_MT)Y<0eARv<Sk0KwoXY$mHDR7^*}x+esRfD-K8
z-|iRY*&M~YmL0REc(V~|#_`7SgLu-bftzLCF1m-r9EjCxRoGM)h?2)6lbl}1y)d}_
zwdB5&-)?d){k-cK@r8n5PA)t|FF7GIti%X^to{VQ>|Gm9_vh3|8VgT>*gzPlwq9&d
z;L(r7`dm`sljoT>)#NO?UmY5*mJ^!hA5tA^M6oQ<7~Wx!n6FNsK@Uh*I_|>eEwv@R
z0{7bdc6Tmm2aMgcN}S1I+pd#z7dd?qGlpGR#7X*t;qKez$w+yV-K)o1KLpOyxhdC4
zGg1PnmrY$9fE5UJ*}s0`;xme*&J+QL`-W$+fwvXyK`wgqmKH3gDCU`3QbRDg=*Rk-
zrwf|!rSTV&3Gz?s>0kX>+%o1lsj%cu@#=7N?|PrxXbM1mvj1r@C3ak&f{el8i1B7X
zrGD3kk;va;gVK=$v9&4ozQ@mW(ZAOH?hCpYp8hG9Ka&+dnIA8c6SQ^Nad#A}p-Gah
zaIW`l#NQ~<b8x?fJzBZ${$wFl{ie-v$$lexqx&>0u$A#1ca>be=MCt}d9S0;AkL=2
zWTP3xvfA@IBp<62LwEJ^fjvrzWuP@((j4!F<WaYj&-XS$hMW%DaH98ZIIdlBVPQ9u
zVtfLrj@~;Qlfv2y$Nj@iFtfuBCY-v%{^)=5AS%q(I7P8oXu&TAGWMkzzA+@zen@uJ
zm{)HZj%I{q@7FB_$%k-IR56h$+tE2<Cr~C}9e34s2<ieM&;2WklQxxrH07HnEEG!5
zDg)~ux8%r%K!Wqnm7#^inp^1c@J_q4Hwhuo>t^u!W?eoDP{MD<1;w@WX75!3G{dpw
z@MTh#yJ4SIY<USe>t0s)*<dSSYqO|fPYT&(6iLyU)1DZLYfNcg)+~vF61pp}#8y)z
zWVxZdX-&ew(?)%d>db`W7lyz6w-3B@?gViUw<lxSjLck1m<-uigqf}+^>BVWyx(@(
zS2osB;pe`r3zk?yj0E|(l5L(h`&CvZ0u5&F*R!7~(@t2p7j$2R5oaqTh3XR56_bNQ
zgj}Vm#td1!10-1|<*?%sG1yW<zj!P%<v!^+=dK@d&9FzfZqA;1|4nKn@%tZ#P+wn3
z5N!&1xqq5xIHRt<=Eg|kAEvwCy8fE?R|ZEUPsbF-?IN@P#EIYh?d}v#TkwjY-)Tj%
z=83%#Y~mq%;E>kMN?_WxtZvOiCOgDra<?vFWa@E`k<NmeZO~-1SF_6YBy1*n)oMoe
zzLR4HPAxBk2&_bMx7`uTP*G9j&iX~X!StkfqidP`Zd-$<;xDBB%X%;VDyjQ}6?A(1
znikWNjvl)Ko-^*Hojdd4dMUYTSH0k~hd#0>FMCLd12H8{qc_h1CVfr^WiydYkCpsf
z6c2TGFWFgW!EE>1ZEmN1|1K_QQ}g6|!K}xLfDxYm|B8m}#y$&3v%MD4DKHj`O}&>F
zZoC-f5MxQU<k@~%!quW|MrJk+4f3lZ#jWGIh$)x*AZo(*bKLQkQJ<;rB^UQ451GmQ
zoM4ae0cjKFMB@d(NAW%*=u{lQ29BC{zO1Pq=xlR3yRrbfJ}C4WNqTr|?yG}6KzrWu
zEDbw;y+VZKN-?xe&3@$^?LoD2cz-5|N+0BpYyi!OSzb!Ti<QugBmNlXM7x3>CVe5;
z!J{@Urcd0y#^ylIi;sQxDMWv}s$ZYnc$ApNq(DtMI8mH}@&=%<f3ngN6ux*_NV`cD
zz~4<$QGN|>3$}H)`mS~M$E%Z-a5JlZAm_u}&53HZf}xIY?B$P#$ZSy4&@XzAD7$x%
zD4+0(pAc4q<=pAFcv&NM)W~rCdpRNdN=y^+%@#@;>jJth+w0Rm@2~_SY(3x8LEU2>
zFN`|bdBEx`JLE!&AIeuR+GYs$ta|pa$~f#xk_42gL-L>YM$0kfgj`%rO0I67i~wB%
z$2LcH=y}}>b#o-Qsg5$C;s}SASQ5DzaIkOL95Zi>2*>eN6eEDakOadkO<Bfcfj1~e
z5+FZ=maz~%U_@9E9tXB%M$P0Mf7u@lGQQ6chyA9y!ID3Aj4+MVkr7o%y-w*$$gc_`
z%Wv4c(E2!>to+jnVpWE=<VC2UouQJ5555&V(4;IFcQ;tV{nR_gA0_MA-|+35c0|O5
zsChvZ%YTHQ@477&_t(6{AO42|VAoJEDG<1Jx-(ZOUL@1slXo+h_mY0tGwwIOY+<*1
z8)P}H%8Cj!PWX72j3%(AV%?Td<6ayht&jNQz2`ArzpID{KB!#1DozL`kx0$s(R_;W
zN=$r#Xn2sjR~(;-T99W>cr_&gP3>El@t}UKM?CcB2RQ+QhdbaQ)T$n`&+JPPtyt#W
z^o)U|szv}v=66@zA^!1jzm(Gu&>+WaY7kVH40$a`ai*I|xP)dM&$sD+K3wf|TM4<N
z$EV4k)2JwQj}I8!eq6Lm1lRJ<Q&W<s@9xz55V(1uv<};PbSR&gCgARoPUrjx^<hn^
z2$_Wy<o3#~X;OV%a`xJWX3488KlNk)^EvZ4e)F4rFVE{^{r&!BebAyyG1hn7ZOaWd
zbq&^+*}S2je%);~7$_?v38-JcT+M1yjLvBDLEabC!Ju=j05YeJR;!fr;mE(yY5`1r
zc9vg~fmF5`D02ke64?g(6?byr+O8r`5kEM;BG<XSp3gDv5y6x;nD5(m;Eb)H@1m5x
z(HPK?VuLZy#_Ws=C<aUce|&tk96lWT)#;Q5t~n&)_^y7$UTB-)%W7}t`jhpp)@@4`
z5A0l<5@*{{NCX>9?<37m!I0Nn?r};ZR?AEPf2E(0o}rodKi<<P5M=brt+e59KsH$Y
zsM6gwXRYzbXkgV++vD)=ONE@CIoh@OZ7=u1&nH-yC+A0*OG4v59~{OX+Fg8+{og+r
z+*-nC;n`0g|C<xefq(W|8+o`Aj<9{Dt5Dv$OM@v8_zBf`+FRr!+BSlRR7*6o^!P)X
zt5zCGq)%3dE7?h+MfelKX9^%N9f`mS1PPQ3gkwnx2v)sTOQ01iXSATVmCsgTYJUxe
zSiYnY3<MhK>`swfh7{#`IF?7th$TtM=;(3F^5WlNFiT;&uZ2?FbG^6^|DXq}rYvgQ
z!90<JEde1qEF!$~cCzOPEy%F;ysT~Gl`cX^HUvOCVv}hdD1yjF;m?plwPSy8+ivJ0
z)j)8BdDAu^4G`xsK~XuDQS7*V(zrl_auVR+TXhL<VZZb1ix-4J2#^}aZqFlVF{-$O
zH&k3#a3%YfG+}RAE>F=-hq<a!#?Uz|ofci7G`9E&!h8%ceN6s2x!XrG-C*fxiKpdN
zNt$i)Y-7b-FZ`1}yGL<IEx8@C-Im6$_g2{7rua)H9h2cwiVK@28NrfkWFL^DtW0y^
zD~`?f#y_MnbC1Fx-Ec2^%1ng`W%@zQI1is(mSV&?uf8MB*4wI<0>;6WCAIdV+iyh(
zjItz|&;a_+O`V$NSgpTtAg-ae>i(c`@?a2ab;tzoxv90NICWajl2TqyCmn8toN;-E
z=)4H^!zZ6Lg;1imt=x0x2-fv;LL*t<3CMO#y<WR8#~1Msi78Ko9Vz~Zj5_KveV0f&
zn07cRQL=aBZfuaCGL`c0<@|pV)2}W6!NAd-E!OeKzUSM#nRQN_>nOn`7n;KP!aKIJ
z40AekWuFt40$G^eAb<Y^2MAE_^W;NDD5XjgxVNRK23DyDV*@0h^qoOwHD6Yc0ba%6
zVwLmsPBRF@@v3>_*m9N!Vc?0b6k;4Tjc{I2WNn6gpQ`(j0i~MgM!YU+(@*z+E}_Fc
zYIa7x-=Cv%+g1|IT{)DZD9vM+&q<55hL=2ukj}9w+6IZlJ{<ifS^YYUfG;KVO<}F3
z>|$vYPSJR)8KZFP4Jxd|V9+HqEFLLjGS**mn#&OBlVvAL*>Ge31<=PvZ_lrb-RH^d
z)&k;X?@u$;tBM{5Ef@jsz<->bUl66|fcuus?=IYFlK1$#-L@QV+dS+kkkyjf)}=f^
z(xk3kl7Kd%hh-@}@jycB_axWHB{o(P2p(LzB?FS(f7PI=5ENdS^v?OqKx}VomQDGJ
z;x^l}OUWFb+xxN}8{0>h)~j@<G*@^_>Nl;|KR`!++&z?6<mhomZ&6Rvln=Ss!<b+6
z=o{m+cMX-H&+?C72aJ`nmth~?G+kRcvzFy)?#g>sikG0S3M60sB)8z#V@LRH{JHjS
zGHcD<?>g>oT_x|S)bW($FQoA}MBa}{cey^GS+=gw>#tp+{e8nD&7_y+S#{+YK0j7y
zV+#0llBv4-ZfSiw@VAxaddBao%aHZfHRWJl(^C6aF8Lh>ZW!C!UP6vNTucY@yPzuk
zKQn1uchbxV++gwFtJG&)aSz4d<q79FWT5SGyZmUcg$aS1+6qt-76<qp2*TtpNd377
z*=#gkT7otlQGiO>Urs3K-k5zu7tv^H2_TaGjUc=Tu&&Y*4+0$tSWr@eg2?w2+p564
zOKw~Pyt7I-S?AKNX$S!h)=t@R1o@ZMZZ4eC*zC)4rX!}0n!ncfD_Jfoqu!$fh2g#6
z^V<tH#sEGdpn$>%{kMLdVt>I(5U!&I(R***$PvnR+}JldU)Tb7Vx<DKt`FUX+Gf`Q
zmfdYDO@koF46#feu9EPpCUajWIHPm|cK{2|T;<bZ9E+_zVw+drVN@<8tl;=q-e&VF
zh3@#oT9Q&M0YVYKkn_B#4;1um&DP>3w5>BYNuT6JHxY5Xz>ihH<0ZN(et0z<^kape
z>hjc7#0Yv>NNU_-{WDh^P18f-yM{H7&^Go>7l0-NgFI&94QSKmu5k%Wr|y@`BHILK
zw9Gv76)TL*FUek(2M0?algEs|lNMbJs^tmbf#iaG$~|SdPTA3<cHn+A{LVR2yNM^T
zqsjUs*!H>A_u7oHeuT*wz%bsO^n`$<zW_dkD?6P*|IUG}2RNjIj3!=|(P)JVAB-65
z&>RwiMM^fhr(@Y?YFEDbjJp$^S@7Di$IL~JEJ>;;sOK1M|11+~OB27Ns)o0z^e5&Z
zC%Z)`J{4}o+m>f@CqQ#wOor+qa(jEae0RT8(BU7?#eY8$1A(jWWQOd57I*~BUm*VA
zxZp?gaP(Y<;7jh$Tdwa`QZ9|AOaznhZ)9IfK^s7Oi24GQafkaJwSBjA{T^>EXb*MI
z7y<*hDsWoceIr@EQ7;RpnaHQvU_JRTv*2yVc2f*r0f@)HtZrfI^{gls;phw7tUXZb
zd$9EDnWo}NE$Pa|l6F?3<=@Q=;|!y_ZBM-vJEi#$6io$J&gOX6??fSnFH@6Vk(*)g
z+<>teCUAW-fNG?w&OJkpRLEcR>I%~5Us=fl+W-&iw|l-nrsHegy&YVAtk)z0r?}h6
zgO|jwqpsg8Th{zaQ+#AfSdz@{E3TJhuTb{mCao2u91`mCQh}$1DJ%8qKi<@@FU41P
zTy`z?2;rH;Z$NkoN8Nw~D~j7B3CryAi`NKgR(W>-jUxr7ih7JsKCRAoo<1opO!)nt
z%|cij)0pD2)35_uDQT07FoVLLMr74US^qggv8ht?rHhZ3r+>bO?saDeb)$Fc`Akfj
zIW|Pe;yCHprGW$ZxoB#|(=HR6kxjP&AYkTi0SwN?b+PDfc=vLaUiC2{4F8G?t{;zh
zt&~>37OW|eYw9^i4ulXe7~s4U9vV?a&NE?r;gMU4+w&PFCjC#uRI<&7Zj!1h+y{;J
zsXnSu+J;}RaA2=TVmD@LO9I&f_BZRh;l8o!0LtMTBc1Q^^cDymd7#Wb1=LOjv!Ou0
z5qRyfh4oBj#@M#6<2qm0)&^50q9XAJ;*d_J5nukhhQK|QX)j+ooa1LI$TVr;C|SHX
zY+v7>hBdws1IIWoaFDjujRZ1%g@JRa$b2eGsdLbnmqn`RkB8W9)kXZREocN_j+NRR
zg{(Dg;-=Yivrd8N<G1c0yz#?{(oQHFD?iz{lQ)H8igr@4NY3pC$q$AwBp6FQNqZPp
zdiGW9k_#5&*xaGuVQ9n&^5dZOh#L7&oXV4zs=d*l7b*_rx;>C<Zq%jbZ^i9>c}KJu
zTb&b-XwRgovu5vd{oY^%ojS-G1n?7G7BjZuCaX$ZjfE9EEdtz-kERal5^APC3VxaU
zkc^3N_6SBP!-4ClyJru#IRVexps|ixOcm4QULh8z_JvojuevxfusUgk*d9R;la~Z1
zy)n~<5VeMnGRt%5U9acK%~3tbZ>&UOLmY18B6o<ANNjmW&(`289@fUzSBCxmJN)oB
znc#`VbZ-A0pS_nDfLc>mo%EgK6_)Rs&@+;Ek`+XND{+#5=$bUw!aQ|%X8=sOmywca
z77VJhZyI~1nD1bLczn82%g(qXEyE7adrF&3PyPUrk*hIZn}8l7tl0~^-Fh{wlxdSJ
zIgC+HB5~}!<+jVXe1VX3bpL)F*8oDo4R(S~UA_(TqX8eo!ftE+hUmk1Uxk&kmX9>-
ziT5LMY^}v?g60uHGFHD$M%moH5bb+UP6aUb#bL4J)vhaZpU&E=)}Mbpb%-2D`toCZ
zoI^<o`OV0p^&bcPC1q)`#orj`SDqPK;w%5%Kj=5DDIaH|HSO&D0+PO6u4liEE7OGD
zFZU@@ZGp_0Ola0~J;4nd$Va9T5Er3mxh>8<4EuD)D#hPb_0))L4Wn0O;L1u2>jsz(
z&GQlFJC!AU`~(D7<>Qz42wQ`XYOuk^e>9c!s?`K+`iWr7Do)Yul&y)0u|G3quTyv_
z_AW(VVagv4u_h(M4ExDs9Papj%5{SUM5q^yFJ}q7^zH`YHMBZMx75pr@%%~Xquswz
zM~KS4^_%?AX_5t6ny5$G!=D#c?>`?kVp~~GI0|@Ce1{|N)8A(;%F;KsOUlj$OA;;y
zE_S*A)tm70pM7B>_v1De>-Gl;H`u~l&pwjXajpKz)W3?HVx9k36e6)M)C6o7do%&Z
zEOM;Aj^c~lGyW*i78&`=zq%rNjAoRAv_O{|-jW;}UrwFWBR#Z4hV4|6V6(LVRJ029
zo?Gd3pO!dKiBp>B`z!fE=kr6;MjEtY-Ptx*a=44rBgTJmvJtpE{!*IDACXv6M!bRl
z<UK)<6|f!M&ZCc6oG5_JBwyNdmYWP%Zl^rDx078Fe19`=hqUHd{M*lh*rz&+NmPJ_
zt?8P@9wk6U=bk^`z{$Bchcto>QGv-wiu{#&NuJ~$-r6S_b8II~0hY3mD~DHO2oC6!
zR7}R@zj)O`^B4iYxKJ4b(94%S;ySD&b&?5wlUv8$Ti0{`$7zFz`r=~sG|}oOLsWsi
zeE%=&Sny%TEw)41NMUmznH$#{T|M|ojp<4HKwh}&-?(4+hK9h$6<a7FGBQ6qpmZI1
zYO7xr*%O4KJeNIK+dUr|WgZg@7y8{2lt7MmM(JGooU@@`_}!>yXY!)^C@362h1Yv8
zmn1s!^it|RK-bWVm>%=iz4!ZD>eo~S-x!l9G>z%PxlzPpqI=P2?hBf4YAK$*p0Wpp
zf&L<$2wl0vRF1PoTk!vyueCGyTN8`13Tr^r=uPKMcmmt~n5-7W=GA3q{iK#D3%=Rk
zfrv#yiT6I5g$0PVG8nV%D$&)@4fH8)vP{%C$JSFGyjb=7{$oiKFTlzg2K08-hw)AP
z_8+#NyE4oi_Wr_lGlsvl`8~u<ly~e45io76opJrcNy(hX4GVt-dq2y;@KZ-hRvkga
zTcY#c+}(ZZ#@jpDPaAI{Z<<r&CPThKD0klWHc^nwyz|;hRJi@sO1CmnY}wjAu<YHd
z`uVK?@`s$|%PRr!Aws<kc?Wg;O>Ld!L=Le7bN~qFvzC3h4&&Uaqnq%WVCgbB9^cj%
zsb(pE3)f7gUu?vtrgkcbD4oi<)2FlNzm<ZQ52m%=6FimFH-jpUr#ny{><r928i~ci
zc(EH+0#)2?atajjUZwZ0j}ZGoEc8Lk!2s_OSSBJ&me|T+q9AU*E&k+uGBh#0;^I|a
z6-}k`<hA2dL~;>X`ZT5Kp{|=S#EV{g!GEEV(MYYTA#nXwLMcV{=MW~`8RzEiRM%hV
zqkm63dW9bqf`%~drwwa3BnWQ7PR>4)U#m%*eD477?lLVHL`&ocUw)|B()IfVO)1w0
z5)7fQe2l7WI>4pLTGOU*kFhh#eDn%T$OS+HV=794uD!?UzT$jtcH*N(g{qSH*ZeQq
z9bW?9nKt<=zGKl$cD9{rD~S7CY`Kj#YHh17{K1+LYiha;1=9JwOkPZ*<JkCvT@HlO
z=;upl=1Xiajd36-U)~DNM@!Y!bX^6IPM@`yDg4Rq#WgQSxs4K+6<xAL)yl>9Gs(MP
zwYP3?>-*~B=#Is7tOqp(Ex%xVUMwwY6}=@LXrJy)tltzrLN=W>W$HWz;6#lT9B$`?
zY_|frPktqaN2FF*ngGS&@)LC;QUL=cv!DuuQQC8)=|+*!*@~LU8GV?0zJ_9gC{$16
zL!BXa%%^7z`Ekv>4OASaDZaZG`@yGd8}ZVE59&>e0F%ql7W7s$E_>l^#NU${X0;ma
z6zIL4>Yg6_mOBj~cWcwqa-cL8sDx>l+acFct`niKj1?~19QgO&y^`Sc&26vI@4vq-
zxO-XfBbb*m-AU0&J69p?ESIXY)hPQIbSSSSNv@8ZRSY#to}y@mNv^Xj-S8O26@a5(
z{&;TA`qkxent;Moc^LldFOlUNx|=guGlzmkn&rRetk@?$gxp_Si+(mj*O7!89Tue+
zD!Ljt-Q{7WxT+h#{Tv}H>1+|Z6Mo%Fbx)SV%URR6aS@z4Z!4|QzJC$ec8@*tcf<b;
zC8LW7hx`VMyLcL({!*61#qu__CJ+>cP_bCS{)s~E+e!8fqvM~PmG3yBK0)&EQ0m&?
zwcL7Of>9zdp_shT9&Q1Ji5vsX5*by$B`<-tC<+b4o#7~$Kc^gLfuV#e(L{uXHCL05
zTi?){$oqL3J}q-<YS6UOYe`+P{}eg1qrRfEovrsqzan6|$c<I?R3LU^17Py^<pJB&
zPZ9D1%jFuFgm2)0cPk>6xRi>;H&kga@9cB$#_<Gn!S=;SZJ_jYL$BT&9bt+GnFiRM
zW|VDWyV?eE2|B<j_q8Y!A|D5797|h0X5`ja#^H|4?{@%9^2y7xh$t5$j$Kd|#St6q
zlp3-AyRB;qs=4B)TTEgEC8zj~uNBBSQt4227s76XODpf+k8uH?mX^qd<>JcTH)T9~
zEOzbW2~#DsZsg&bdibdjyl83J4nLS{cJ3N}4*m~6(Li`0X*EH*<FmgtXIfkL&vNY9
z2R>BG5dsRnv554aG;S)jReP0rB?Z;2uQ5B+J)+Xzn9dSgN+}2V%P#p106zz?+E(4q
z#cY5$W?4QWEflv(6x01=);agiEoadI;6?vLrEW!z#rbF+uC@$b%IoKd4d}B<1{utc
z$m&In6sOS~u{-mju4lbWfD44bErw}@HYc!eCQ!Rt!INAr!sI+M8L;5aC!XH|8(uuH
z6}$>xT6ycfd!i$)^BdWs?9s4lENq7~{qf^RN_9b%i&;*^zXzOu9u%pbUS6)lX~IWd
z9j9J{K@Z!{+V-{J@cMXz?Hr;vv+U3PQFjumC4fGf7u-pGA8Idr$sXt!mjV2n6cJ)5
z%JTGL)y+&y56$L^{NSw+BChKg(LgMd34|DbZ4wz}ENxcbZGOev?VdZeM4JJc()>g+
zN}pwaC-GQ4#!ctY`kNVx<16u{5zZ6p4NFEx-Qwu#l5*Zo#k^={{`AiqTJP>cF=y2I
z%%uYG-6e(|SNziDdMyJ%l6DwgvN?g2!ts~jd0C<-nOy57H=2k96uSX(k*d7@-W+)j
z;cGfSk8;T0aLD-3`WF`;liE9Ft8dB6N1O%nHx77g{DXGsN#mz&kF9RFQuVWl0iBV^
zSuVEM7Du@hDc?fZx_mtx{L-sJ+;V~3h<0SidR^VuQ7XG^E<V}&LRIBPisLT4V${oZ
z0hfQve)R>?SCEpIaV`Jh;t4H9a}-DCww5x_VA!exw_2qMYXXI6!d}z*vzT5BskoPY
z%aoguK12-swjq<Z8SM<e<jy<1Jp+}3zrsz5q%Csr;i=2tTzr8;P$7j?I;U2R+^J#g
zNstQ5sg=MfxjP~PFy%x#`gE&#2A1I^(fAaB+4!><Z6hnECH;S>dh58R-?#mLW0W*f
z(%s#o2Z(@33rM2~6Ho-%Kze{mC@F$8$|&iQk%CA!jFK81BP2$H-@QNg=Y9X~-}V3N
zzw2?m&ht3W<9V!Kpcvk-4`5ZjUbgA0RqXv9HYRwbs=Whe%N)Qrery@|%--Qox1j9y
z&icJCQN7Av8M7VTo#Qq+KG#2JR&w`2Nb}A$Q+f3Sds&L6zsKB$sSY3KinpFPY~hpS
zNF$%sj!sN4c=Xuw|BnYbH>EYLuvCG5qATiupk)_nqv1s4tNS<eKU7A&alfMJ4$oHN
z`)6vPyB>5Bk};@{jrv^r_I0=Y8<PnbNEy9Rccg>>Vx*~$$g?DI>7u?i=_%34YG%4e
zbsLQ=pg$l9`h2+p9su!u;4uOym_uEz9{y_1qbxuYS!K=s{VvmR^i+;)J?)ghA7X(=
zM}!PIGbC`1$$8{{vC(*~lXd8%@?8S&LoWjzdr9K}r`_iVbD^|B`kg2*tZC$nju)Te
zgodvZU+LQmw<y%j6wr4X!}sPBY$!f-Y!UJizWcKnY3)V)9B>bSwcM#BW^P{GerCw)
zaToZ(`NfQ)HBGqa-iY%gcaq%QtWN+MpFd;2()ki5_B4ts-fHhfAj?AbLpR|9?LzDz
zpUSN~pw^Ny=!|iK6g|wL5juVHi8m`i85jy^2Yp%38V}vz9_I7@UPRXuRKa5p|E4<i
z3*OaemJ`=?dq!!HeDggb7R(8JgCd;qDLx)h9uo4Qh$5L@n+u(!2oJvNH@3G<IA=eM
zNb-BY&?aO!l~1ds_g=!Ek?-qlS<TQd<-bQW<jjiBX4FIJIe=B~?$JktQ?4gsh0EtM
zrYf(pUbENBM+kwwei*bXQqDmHQMl9jVBrCoV6*i!4VdQCNB<Qi^R)mpG}>Si2rX11
zI^yBKzWQQP7_am_uAd?o5}@E;{`q4f+G<<IvP`>s^iS!icig08imhjZj)Vhx*nHG9
z{V(O_=4NF@UOB#Sg#Rlk;~`~{_Bb4_I^;Uw{9@`yc98CB{LN-OMNjMpz9A9!pJ^Wr
zBZe7N${>MXyv`EJ$xOHs1P0}aY|N4E0&yi>!7S2~Z(tUgya1A1*1E;REq55${=4<c
z`t3-4k(is6f&1G7%=GE<DujcmHf7_OMXx3Y-mn0GDu90IGgh6G;BSGI)^7nS<#2nm
zMH{(0g7RfNXA?|=`npapPTS=L!Y}%(r%+Vw0Ud)M+qE57O3h392w-;i00<c%0##0Y
zW0@I(ReL;7c6Xd4!miE4&J1~c-~ib~9-shV@7+JjM}+9lR6R>i2^N5|-qUMYpVXio
zch)8Arv|O9ms{m@A0K6WS6Hv@-%PBq04w12<(*1c?KXZ1isWG<$;Ywhr?-w>)b+{+
zdi`=3bm_>pV+Mb5cQ<|v7-R<DJeYooRx+?p^QYWW9<KyqJ{#;BuECnAtXK>s0SoJp
ztVX;X&kK~^xwT%33qHVVJaL<uD~WEAQ7*y;>OJl6hG~3{DxJFi42K?Ym<5F0&)6d?
zo<vfByhVq;w9C))W+xjC5ji#Xnf5ZHhjZ1l=bdk6q>Fqa`Y93pY)fanM5)PtzN#7d
zS?{$hbpeoie(P7s69{z8rv5(B7{Sw;oZ<6lLHeo~$+F#dw#GU9kVCVrw0xy^*GU9L
z&|zeF6@lJ)eq`&E`_)1A@p&?n)ZU+!B_{s&Mej*BB~St^dz1;(epT69)*imLjoSr=
zk3)gX{zJNva}!_wfl~i_%B4(4E>lC#Qw7NabCi7UwO-Ee-!~xFM7~QzX4`}FHMhet
zt0F!TG(l$0<adx^bxk+@KnQ*G>Foy~MuZgMDIvDw8u-q%P#By1Q&WpwmJ*`fXcK*f
zQVJ`5tws}`%TUk36?<Vn9~~H*Ztix$X0r5$PI)hO{Z3W#gKee6d;Mq3qOpPi3vi9d
z<J=I3qYN!IiSD3-qnV$7qR@-w<4Y|pe71RSN}Yuz+1~@{GNmnN4Z9x|#N{pd0BJoc
zsQ=^-eEP-qqDis1JAZ%FTL}R2lFOL9^=wyNoGO)J3#J=N-!))oaSPSQf^AgZ#^vKK
zX*5(uoN4;DDQ#-v)XX<TEdCN*kH3)OLmt^rE<{=dbEVF;c7F`8ZnR>PkhOcbLp?8O
zsXe0X;ku_`VP)=hi)A27bMt-wZLV7hqu2|BU0@IWtxX~nfp6I~?RM&#&R}&b`Er^p
z!xsASt@~}UKFh|cFE6KY@Hc3klb4JqTft`qR=cv|!Tysw#m=h^_&z;kesFpFSbe%_
zr@cB*^On=6-wn%;B;D~6ws^z?z0+KcQ_pEd1I5T>Tq$6lT@X~X(E_kv53j<q8Pgid
zJ?fSZpxS1N3OyYj<c1DSHIJRwD~BrIvwLmrMK(wH)wb{DXAPglDyD7GHa@TTp<>wY
zHGfc#Sb!vzeRhq=UW<%)wg+A__sFoVBwR>8Tl;wch1)};XU`E6j&`-r673q1!yF95
zDt&Yb-&x>m9>m7^_#)_sXWbu-8(EVFR<Zx{57R=e{)0(yyxz9@xF2=3#Ch{2tl`U-
zFBrm^*M;D4>|2kBh(f(x7Cy7&EC^#8sJ1=Pj1NKzl$8*<$>5{hGyd`I;2ZHMVTEu)
zgt7?<{z89gwfIC@dJ7Up!xCh+-ybva@+wLxMhy2VSJ#g5O6py!!U2rhzE%NcjUq8S
zB+b_Uf_10Ym2x2BHX(bBNCMPm<%~$j@=FS^paYdo9(p|nf(N|KsYI@NH=ch;;c+jo
zO*;J@Y&X&{9@x<fLJF};Rw7r5h_~T%l($#sfCr+82<5-lQH5B{aH}}pq&N8{tWG+0
z6^2{UQ*+%17$3s_B6LvQ$I6pJh;x@x7x^y0eLS#o?-A2v=83xM(ckiU{Lt$c|3;@`
z3!`1C4-?5|uS1@jrbDTygegS4vQ&%`ELz^_kgLGzR*<nrlY{(iMC-JJHh4TZO|7#(
zjs-UBw<vfGG><;|@&NLDehOhSoc1C}8mv&WtOn^=Y8{b#bGjDRG}uzku(g499G6}F
zzR2lhp62z#X8e}@c%o*Oqy1DkdRM;@CLt#ekOS_&GSUjgKQSXOw&nwfy~Jd!ik-zp
z*<KK9L$?B6&<@u`$5qOu3D4G-?MDWVwX|&*2i%idgh?zMSx0Q9P?7II)i;0H31q7w
z*QjxCFNI$r4h;y^QxCS)$)zkcDjr<he@S6CdhBFpK^BZx=Q6dso@1(w@FZ!$5=(od
z*(_P85$wNJ(7$G_Z8CJayvOQ4Fm*XOLMt-)x2;E6`#t!@uhPvQv%c~9hUtM(=;~&<
z+}hf^bFyGtJ;DBoFX>3@8f_+X0#~zE1fKXkNi>?N3dnqja`)P&nYJRzi9+8qt2Gns
zKkr!HkRYdtoI)TVKcBM;vN;i!qYHQkV=?SfPqL^iSO$_#woii833nK(3jlNExE=rk
z<^PIGelFsKBOJax0KDlV#5|QqW9lMA1E^LzPrYbbkL+9Ockz@0n7bJ55VN~U8diTl
zKDqV%;6=ED;MChO1SXS6IVS@I{^Pv-ebGua?t`A1oFXQZ|I4iSit$t7BDsA4;uZTf
z9gBN%+ieKfKAx(#VYwk+NWGi{>SAB0IErDdDYV|#p8{^s6J4af7@HOSy?8ORW7VZA
zvLbgHN8ts$7iJ0=d0PVdEOL8K>;&v{Ppv83A7h6>*VgtIq{0{Hau**sh~4(k30AdB
zK)yJdIdA;%ojQ&+$~Pia>%uwYLuU70%@STEq<!Fd)l^#*LlxumZuWWm{R1=!AtB?S
z$Gu6?w>Cb?P|qWK;@g7JRJFEyRKrZ$-Ae>nh0s7l%@#F`Tr52xLsj<p3Uj}Xs!)dj
z11yK#c#q%8GB5`Yz{K^_QYQim0FNPob2&R1Jc=$NlUZ_)vSha(DUW?QTK2zhPRW*#
z2`sl>N14yazH4v_E@`aVR+4Ap-rvY(Tp98i!`nj1nIvlK6RPh=PSh6Ip+A-VCsKjX
z{mbQ}{}w`M5)*s-tWx_X%bTvnGgf>zc7vh5ff!}03P%`xw`CF0Iak|D`+9tT&Iw-!
z$I*sAo>Mu;Xz<4T{8*qfva5h`$Y2eBLb<FhmF-PRO9LX>-iWDZ@noPUAf3$Dyv`ne
zCN<~c)Hawc$>B#`?7(t1N%PBfP7zIc)FvI1H#JCZ$VGZMo_l9%;VjC*`32D%lhL-Y
z3mDoKlnzh;SX<Kufr>9wv;b6ctAQ?GyME8q@8qO}8;E4kCT1Q75+j62HG=drhtha)
z;u|tD%ue9jnX&*01}nK~kXKJQT^H%uGZlBPU!3QrdKpA_7-D%D5{~9n+J%YadI?Mz
z>ZS*Zqv&oyX)58`VX~CKeZ7}N)k%GJfK`OI-0RQ9Cu*Yw5ufvYylm#=a~R@RZV92B
zKA}`ewGa%D?dmchsl~j?^397Z+=k`m`QzBB`ESrdzTnSlv7ncu>3<3_!EqkNr3>89
z#Fm8pcj*kc1DK1xwLGb6iRXRgZxauk9rUO6ur#{RaSE``Nlk!;XQu?{Gx6r(7V;y#
z`+>g`o3p%-!r&uX`8j}IW>YQiE1poix&Q7j9NFeZE5)(5t5TDHEai61>L8%Lh%>tB
z)ogc1npn6!SVu`ClQ-LTrhxq-ghWiI_nfx{T_vgXgQ0q;jp_JGidsjC>fn9Uc*Xnt
zN-U22)X@^^P}krcus=!k+vsK3JDh*vu!ILO$jhFwtI=m;DEcCI$3Od0-a`1u)at{s
z$UPl<P4`E+f1Nr1yZQd<*-BUJVcKxeJ%yJ7{11?LmvOG0M+tEsVG7zo)W@FkqzCND
zX0}Nz`HscQ1)e}E;GJQTN|8z}Y(v-!3;dN&MBh*Rt1Y(XveR9YR?MnBVG=~_rS~@Z
z!rwAJ4l+$!>zl=B{HMQeJXsq;8mDaM2vOl8-f{tF05k~4+F0xaQbeq{32=B;;%Tt{
z{#)!|Z$s?WGp{EH4TKlSSqs>+<E+APpYE~o<D>pLFtT&`6-Wf=RP^y}MwQ+G{F3$n
zuFqsg`0MEdB-%P6;%*SNAIsYtZ|o8Uf%}7x)oxtx<T*Gg_fSDAz}&bvpkJ=WLFpqA
zPwTR#QQ!f|-?7OzLsc#&9{bw-^Rnb`w99u}PGKQ%>g}_g1%3Uw^WJbQ)&pQ2v$s+y
zGCO50{TEH;JZQ5p;f7XoTwT52DT#&~nT?q7orHy1=B9f>OClZ#1m}=+d09RZRSqX;
zcvW2Kt9GsJUIHSlKJoYCFuK17GWnIht(>gjs|m3)6Tt<`?`R@MCgZormE$dzf6v{?
zat`yO!+fABys#2jn@l7dP?y7`%-V{0pxYO^g2}+x30aC~VkM?MHSdGf?=_Z|@6S=p
zb?QW^ygs=a$&HX+HW?rG9S1awNZV|rN*@};5`5lKvO+if5`T~$YI@!FZ~)uz=fhDy
z+;cpxLI+<m<ABzSv3?JN6mZMsEefjLKyN(lvf8l+zxDM`-~R8}CY!)dP)jctQd^Jn
zS5<iMfiu^NN<n9p$5|m(b6vCYa_buB(Q!UJMZfg5N2q8@GB{b}>D@Ao3Jr<(#Bz-y
zY`Cc0d}jA<ISN+GS#C>>vR4f}etMW9u;neV^z^6y+mDf4EcjJNz#jy86c>YOP<FB!
z+Y4msp*#lk=Vu^58pbn$#mE#teAxu=<@<cfB=6<m_uj}|Vq|wbuzl;-BLD*un*!{j
zclQ?GV>K~7ama1&Wx9JD5GRd!h`fB7h)9C9b+UKQ-AUMzlA=k!XP44Wp#EcNeI%^z
zhZ_13$jO5AG>f5_L}8gCk6HJgobI2-J(x8=<)nM&r?%Tz&-?cUFN+Bfn1oBNL3W2L
z)B2%|q0VL8a=e;2op3%Kmr&wRyvBv|Z2e?zZYh<WrSeM{GXv)3ct7`xYM#2$ZN_a*
zyo_{wWN-quS7$1f8F6(g`l-Jb8Sp~#_w|QuHlGl$RM#+?kw3?xz(<K%Lkk<Kt~%c~
zI$aNf7_mCepsb4fOiabe$~m5RrP{dV4uD6Z?(4=qkrVQS5@Nj;?#$T_rAj?Xlt4sc
z%q<A(=A=)b!-Iw7n9F6{`^J|FL44={-ib-Pc6jLv*&t(by`HOdbVDb^$j&t{Di8ct
z;w!u3HLP9~4pm*8-cTF#-JlY(goXHXBrKO)4umYVH%KNbfviC-AKII~oo?SLE2i9t
zx&G;3GHTIOQ)7<T2Ft#5;6v1y+~3}_KlrcW!}vRF7*bV?56Sz_ahR=tnM%U(8sfV<
z=47nXMx5+nTWRYU#3jer^B%%+hfvPC)<{*NKDJ&^lCfvImVGqJG{PA`qscFUGM9`J
zgp1`YL^XBC6Ox)*I#U4Cy?Q1jNf9I)tcL7GBt*wb$FGa%A}{&hO7LD2Q;Ip|*KE#A
zMI1%7aq7k{u-29Bw1)Xw@Bn_G-OCkySZPT&%<qvjbuNML@3TXaNFc?VT>4Xn0Nz14
z#0QssA_#!Wt<7syKda<LFS*h|18<`da&0*PL?flR%hJ4p>CKljQyDm|GZ;*Qr`Wi2
z5l^QYcntJ!cboc9_CR>4*n*1n5}%gCJfG16!DTZ?O4pS<H2EtN!9*B7B0m<Yv#@lP
zp#3VMSDL<W8sE$t0=?FhQR#Yx6H1lGvuu@O76GAK>Q6!kMS=cyIY)q=>U%;&mE=(L
z9|Ra8Q5<Vm-^Q%KJzu}hGkY}(s6hN^Yu~vUtx*eE7eAOBS#|=mS|gqZ>YV;v;v^On
zZA;k)ornyDYcy#f?5_P!;=~>?m{}U+=V<>HHsm^cVCQH(^6LuY0D70EV)ED1-*(Pb
ztq^e{oX{UVb^T&%_o7*9O8hdxn_hF2@DDJ$OmxRG*x{)en9@}_%VT}cY%Ah&#EHwp
zQly1xd$a0BctBM72lB_OANcP!xc5croS$-T%MN8n>IiQ|c^_ncDJD71b}Q0s+@@5V
z<I2m+TmRsJUnFG^EhrvjhU*lqhW=})`KM<ur{A5f>M*l1zwXVsif;Dv{&MVj_8QOV
zl<TQI22(Z`Q|b<W>#J@3uFV3-A&X;jB2A!VqK+l?5?*V3woF<D2%R6i&>>0$Zr8PP
zn$bWdlfB*bmhN+NbpeTU+khX-K)RITZNy=f#EmO37XD{Ob9^ve^dE;z%_uS}8mhYS
zinox@^o9UE)BD_~1Sm9Smcacc`Btj=N<Vj$W2i?Fvn7vlUmhoAoKTCd=#V3QsY**m
z%IR?u%dyWepndf50SO1CJ%0@Xcm^u^saf|V^1}+iuG%~O;&C)Z?q>5g%gS~-;Vj@e
zNvKmC|KF3G&nlxe0&lr~zFj_tFy%V&Pt*FZda$Y|80`JQ*tZiAKYtpFmDy@6M*P~<
zg}3+)>@3WX!=4k-RQ!H%S-+0gakC>zj~BXzTZvlJmiaPigkuD=Dcgui?s$$?W4YPv
zc3AgFz&#6{(rIcM*G#28e+Y-F>9G{goQHsLwC2+V@2Y=#FkEzpoMcODo;r{4EW+0A
zTaYJh)O{k9b1d_)*}S8hVebd8K<gL!0J~G>10++$T>eH8fG<&?Zv!-<H?_7iwaIdL
z!y_FJV-qk;@n1XtI3&ISjCAT(x$qCLOcNVY+8{;@a!8sRzi(LC8Av2>giTtr2D^G|
zG*eqWE`Hr=I!xQvlg7=C+VSr}m74yF%7@IJ_eDLNI6vB8&77@)8b9Cm`t+aCI(66+
z^T_{>D*iG9SNEFs)&B`W#&_gFm$QrChf3bEj6)J3zW7CMgnT$We1eA-vwX(0OE$&a
zMga`GM@+|bU%;LsD^RjcJlvQ|GY#R}MrN?aQWfwDXfh~^*DcTVcnu>S*$D^bV6Gb_
z--Mb6k}1;U<T>-l4&_WC8op(G(Rb`<AwYbETvm%eiRGyt8?KQXlFN^4w6lo&f&9o*
zE@#FsxZT3g)A-CRK8__2m4skQHSKp<N@+mz-2A*%vgj&BD5Mk^qGvS_P6N>zFKOS^
z-DWl993!~4c#>TVjifW)&~Svl7XW_$fXpT6@JhnttaA)@71sS`TOJP<jFNcwww$#t
z$YM1ORg~=*fsAACfTHl45}><?ohzUqpB40)Zz2twqA8pS9hf>Sdt6Tbl#=AQ88ym{
zuU<X-%1QKGB(ZbLR5<a=jzuNKG?HHgUG?)&l}^Yb%N?t%&=YClIXSy3X|L=)^$fM~
z(wzG1g+D&vvFQy9OkXb3#*@2O_-HYY8Y!bwL9fKUK@Xh&-E9x6yop@C32>kWzNF?o
zpg~LwRW9%vQua;8Np!&y^F!G7tPues*$j{#myFKg$*D2f2R15oH+_gZizHGj&mX>h
z_$U0R=x|W(MKNtX4qwNmG!G-#0u!;@7b+3e{QE>vAIbXbids-tuS8x22e*jXet*!_
z)wQ#Ttho2D@>k9!|0g5;zc-|az|?9pQI`Yn7WO|L)m_6)Ii?z<O7JiC>P0;D>qyOC
zr~5l##=!2$6~4oB?tM_`u*``E8~dGc><i8x{T?<XTM++5;S}DK|J9>J$+rmjmN)Xa
z@{5Bv;0qGc3XD}E5APR%$GJIqY`vC~;quym>Jn?Q3qqLNCFY%&%IKiE-}ugSSdg>B
zg_JDEtrWzrL{d-L=0rW~WfbYhywBo3YORp_D3_J%eR{fs7S1E)*n-1+R-rL$(2HP$
z_E}s)JN4e^sQt%_&;i2}D6BcVcIF&!O<{2lHh!aZH~{~2XLwW)7u?u|KtG?Cfe^uO
zyoeJ+dNNHbsKHiz?(;zTq9u}2p~gbIUeo+7*?bj#cE;wWF(JT#0#*t)s#kt~M?Slk
zF}krLBG+SW4lSPK*sx&sstulffm_#7hHJ8`0a(2#Pqeosy^6?I3B=>C!OheN**2DP
z`c%J1*}zH(Mx}GR&FoBLpLHKrTq(hze-<3G^KinjBdbfjy;oPzzHu@I(sc{ifN<Ec
zwkKg7e*@#jkH#9FQdvJzk9fz2X#FaC>(11v4b*Y|<;|*k@cuigDSv*5cu*_dk3Gh$
z<$;=hrx#6XyaHqKIEp50>z$M5GxzdN_$VDdYiFJQd>Hy>Jl>yQCq=0v-$tN~+h}!k
z3Gnn|6;m`cF&|THtYAc3hFTWV2Ygsrlsm%DtBf6D8<YNVmj8Ru7F^0@cvnNq;b?q_
z+<9Ej{`JDXzOAk8p~XGIEyBNr4wabSY_2ml53MPWKVt_7+}I?+IRH|Uzan6By)=V)
zMHYw(fn_d)fe>=x5c&Y*lqm%c6M9zOUIYY(KU4N+8R-SXy%G`}NS^0?m+*>GHgaQO
zo43RT`nI^KFWM<>EKWhrU904#2bEuL1tY&=mwGn{a%~FQufO%;ehWkcvLvzy(WrrH
zd+5>EK(V7+J(%n;C_g28(5c%O{RrSYq9ygRb42SR&|YAS3vM5E3RG^=5RZ#P_1C}W
zHIz_^B$ClJb1Vds*XwVDof{CfwKeqE9MQ_j0^WSuVS;=XG_3fpj%*!*AWQceeTgw*
zqo+5|rFJxP4UMuNj{59JO|*bWt->lFt^_!5P&2o2+fTeWe>Cl^Ly(~GBj2M+Ri9b>
z{`(Vd_H+-xdzd8Zo6QJ7Ju?zU<Wf=Bs#xdavHfJVWyjvE+j{@Q^F8a5x61qI$?+D<
z_k|MYA}`_Xjpb%pLH5|&2t?bEi1%I*XQ6G!w%YobC}k8LyrM*0S;fiXIeIQKO^xaq
zYUXfagIb?{H2}V5q$ni#nHAi>9wztFz0A=4QTz2Du`k+XL!n++`E7^K1qFO7TsqlM
z(!kk}w1f<&Rsr@#?1vj4Rd8Tovmmb-)SKzZfC3XZldXmRSc`t0rI;C>_S+vF+vK$r
zE@o5PV`kOV)P#?zi}BF0m;O(sTUh$JoaKL7YTS?eD52KFPZfNQ(W>XG9}m&y#ten!
zw}o<EF914{5z6>bk#6}v^Y~qxu*Q2!5FkNluHKOTQ=k-X@^CM^nL_HB$)LR4qZ7RM
z97wE60LD}T36`|<<&~ecw-lBvV|1_-+j5)r0)%=XpAw>bBG!}g=6tzi66$2Nl8;{C
zMhj5I01J^wbI-k`5yR@c>l8?70^zR%)=3V^ItV?k*+jt|*=u3@hzaBvL;Tdz@mJzc
z@=+rPK>m+kPl2S!SIhBNCSp1?kxZU~W4iraiT53S8mZQapOzOdb+QyLvmgO0o-C}4
zy6!&K--LK@adLw=PJO!=*#ypyEl+obOfDB@g_yQqQkX~xc3+KUYIScEMEr~nn&tKe
zn3~xEW$Dq9;hItBT-Ql^KhKj#A3_Y~BSRL<4xo<4EC$IH?Qdpbmx4bVzV6v4YR5m+
zKlxT(kuN!GW6!t|=CfwGdn?xXj036w4_kjLn{=dAKh2MIxx$=kw=@K8`x;HkCrisR
zjy<laqn`1QQ}`k)GyExAmS-MqV*9kbS9p8OM>{|iVS4%=6n@mKKR$BwWO2m=>-Ig_
z!2o2~>CvWj;`4hT0;*r3{udG9kI80$V*y7OX(u!Iq|MXwfvGQ?(~Ik^dFu{0YQ?Yj
z`PG_GtBAc)tLV~&0;Mh3e&RxANw0ohw#t6<crKe;33jsVJyk==-Ty=bXu5xPJjkf~
z?_Vg#zYMS5sh?EE_X=v9_u@P1e;~^FRaMgGBD@5CWWOY*AaJuj>w7TSUpV;@eh$BW
zcgjgxO;Mux`bFc~JFStVt4+i1m5n-}-|W4L281JDBr=Hi*3Z5)ag|q}i*{n`)Fhg4
zoN0u;rGX;3M3IwcQ6GLBLEC}%xZWG7ztn7LSn1;;<Z&X&2Hu7Zq&!LAU+-!8Xr@vg
zO2_xH;MS%{Vte~2pI(8p_=*q#SrDTUTc$rDc8)ri=kIw`Zg#<K<+xj@U&HmA!xb&u
z;0!a<Vd^4Tg=d%%#TJ3;wQ>N-`>w>8DZPGn_L1```JuwK?O<!l@<x9=rLbiDt4C&u
zv?4c`=qX{c8|q5{W%Za+Yv4+M^^>^02<Gih=c(Dx1*EA=G(NGdaiVU@qkxD2pOkk+
z9+1W-7HM~Ksc~0cuusU{^!EwrQ3DgzWy=8Q&Z$gdh;TpwV};>R*o~rw+l;j-)MR4{
z3vV8(nhl5#k*zOBN*^hErK!bX<6OC2o;+wQ@EpvRa)l^+mdn(Oja~X>3Bp>3NS-}g
zL5VG*B)u=i$8b|P?oX`Kzd|H$JF!P)0OWjxwgc6+R&;#2ZtBrS2<-Pts9-XfE^Gst
zL_j$+8<1jv9nF8TNPw|k)X!Dn9lGQ50UMjVbH(D)3FQnKI6=$?`K@Ip5qRLh*2ddf
z>vMIeWqg(?w{5-WCncZtJD>l#G5^UT`PblJuk*Xw<^9#N<js@APdz0Sts?OnH>gbY
zVC8e8ghJlUw<|nT>bEhbiKIpuiwaf=4Pgk1A{^pOnew~7RsT&XaUD9O#xq=}u1@A_
zg!`sLquq)L6Pp@9!%Sl~j!=-%vqo->Mv<P1@tSpm&C7_f44?ktC_hb9x>U=3ncEZn
zaRTC}r_+hki4k`S8}1|fOz>g#W#(hTUKe{A-N;`#^PGFon?3;9tG)bNOYgQksD(eQ
zzZvCH&B^{uc4x->L8&d@Y4O*Fh%(T;d5w>8nT#2Z^u{{UH>-ol<CtX}`FxBJGf)D2
zkoY}gk-Tke&($0Il2+;HUA|e$a!X>!K_mycPjAa0GK6)kg>(}YIDgcuSTNP9$f&RD
zLGZ-9$}!Bc5n5L6L<GsAp)?w5r3Ks(&W0$YD<?<)j%ZQ-0lOGfpz9&Cx>F%JPwzC>
z6~5PkhhT_bf|?K2cRa?#z2Ccb-gIr@!vqyMR|W<v_Hz$NhGeeQ;d5YZM6nFerkFBn
z++euGnL4rD_S3s9_ovP0?3c4_1)1>=5t10N@HaO#@Ck+67QdnMj(^qG(8!km)3%|C
zjb?979UgI}&qH-eAFTy>hvjnR0zW64JIQ+LJK5b*HC%JM-7ZV4HyreI{JXiguQx{e
z$NB;#wAuuGcF_cjJ)ZgHAO^~7gnCce>}t%D%Ll5f5Xr41*tP~Wu&MYyASNMk8BCWS
zn=OF~Y{yvrV*<BIV2LTk{_E9b9*y1qKI{2WGW5LvpAo^T@y%T-PW8;7Pn)YXUruhb
z-|pXjj)x{93=q$##KR5PoY8(AMDo4uUmY|0`N9Y+G}Xdhh8BXk5{Q@M8ciW$;m_Fo
z(0|RWBf=gMV$r&`mG@Pt1l)ULm(;nZMy+Wi@j0E!vk@u<T9&eue$1N62=X@RVsm6n
zvhkR<_v#~-JmLwLQ$jJNtJj~q7Zj?NaW3`L0F{cBIj@Wnz|WX&+p0?fU`K$>FAymJ
z&rQk&mGk5ZIu<_30w{?3ZrAI+pc-T#d|FK7DXO8DRl;D_0bO<fYON+QcIW&7gmr45
zMetL;j7`A4S;?-v^H%dgjU(z@$GYn4lf|d#Q3Zj>Q!Laa|42Gyrx+vk)pc5qrt5A;
zjBktXysTx$K(1GGMzk9B>S$o;m;Qa^bq?(yUan)$jSsdb^t)|cLt(}`GqoeT&64yp
za~{%|y?({QPT$P|0<U3@thWnqfVZoRcDyHxDJWQ*(QuNv(K|rMi$JzHn;1KI`7`BF
z7n-@1(!og=17N1Z@YtYI4MKP>RaN8jk5YgKK1RISJKy^+KFpNe$owql;~<l8l@H_Z
zjR)c7xQomuK@jmt;=%<y6)Dmp0UxP{+Viq}wNSid$rZX4G`&(vIgF~_5a3yBZR%xl
zcu-98`BO;tt_oY|#nn;i3k9M__zfex*)HXqjjgR{cRXh*5&8W|`Ry39f9{NiOWpBy
zsOrD3&9x-@0*77maJ;m^zc%d=Bx@uZXJL5QZM!6oNx|#$O)tO;7ak|UHIF@>0|Xg?
zd-z=I5JrTzMVz4>{B{8<NmgM&h+7xRWBmqKuzxZ#ye+hZ;st9QU(?V-$C^t6g(UD^
z&$gE-WA0%pVf8-5r=_n|w*vg@Z7mT|H*q7(NdmW!LP_Ci<l$j_mLCl?`ga3_kQ=n6
zOg)+J>Xa?Zr31rW&q*ONngNN*IdXQnubT>M3MR@$@op)~D7PSK6dlk5p|Blg%0z+r
z#2gR*;8Ke2`^iAKx}xSXBE;ZCFVr!81TTLSI_837DJ>3rkgiUzzwzn9k=O%JG9K(H
z0(d!Yu&%l+>$|9IxZC~#KTS%x5*;<E0G|LbG08J&QNbkWc3iEKi`T0)`|#pTOmBMC
z_x^hNMPBSk#nh^5Wom;1Gsawr?Pa6GWEI2bH_qs}hiTUxtzQyJrgoH^+h(DUOLrNr
zFThe*8v`B1dX{<1{#?GxtyOa16K{<+Xs*R4?0auOnV#kZc5>%x6{#i!{`om{Y}Olb
zXamhd&F4_9^yJ*y1HQiXXyrGP?Xm5JVk~8h6O*&2%2%J#jY40~O{y$-8Ao8BS-@ac
ze<r9VUV{$fmk)CQX4`oj&3}<_z<XsjfWt%S*6m9+hpKq-m0Wd735WXI_e(x(0TSE-
zHF!JJi@hMgcqzrl#j)h_zL4MUvpWBWy#DWL<f-#o(rVai#I9l;{{|egC+czrKRnrA
zd<kT43)_p@YY#QuT1f0l)h7ELM*?~}j!3@;Bp&xFm~b&G(UavRisNFnGI^`IM~lIm
z>j~gvb(0H4sZ)QDm$1zzJsvZ;nNC@dG^d>K3v?w2nQzE>50_3j3P$zeZ{3b*M=Xm8
z{HW7B!{hS0_kF6%Ag{#tnEY1+dxyerte0RbEy$9ju{4-=77rzIw{bv8J`VtxgK}2_
zm0Rz0MbAc^60jS$a#1}?Dw`(}x}Pq52mE}2bF!pVTGJ9;)R7hOUTE>{?Oeps%RxBs
zdFh;VL!OHz`ncvE0b0w}ic+VBU<|f(6?@FQGka+=bZ)`@pCCUV%071a8A6)QFEef?
zM+4uLM(k<aEF3@%pM)Lb4lILz>`7`s-3|wPc??GtTe&!(ELfn|U_{XeJ<Tq+_4Uoy
zgR#ni_sMG7w>8H&;naW~m|AIvr^paL)NCrD<#DWqMCNVl$Z^q)`v&qdPX~1dK6M{K
zhVwGYJ%lQ(TT~d_THa4LF>`gPKP?3)u)&bu>p7`BJ9*kBL--InY~LDxT>n9YY3lY#
zEDrZ)QzJiYb98X`SsQIw=4imC;Y=AFlFVaDOKcBnm?&I9EV!V?^L{P%`ewx7fScsj
zdZTV}wAJ@g$fU15D2v=|UNj@)wPf6IMHYKzCoxEn-08k^wnd5M%#X{9izk%-Eqt$&
zC$u!t%-_P4x&Hkz$&TQSoNHF0I4hThW*W>=d2Y&F#>-YWN84CI7KUV^5?;{c^LA?3
z{lV|?#t>eRo55scy0NvD^={@!*{k7=<Ccj%A;swz@;hR&vF~V6&q((gh{L?E=P+LP
zBMuLwy_-1|r|KLl0l(dg^F5PHDZ4Tj8RKV>rAIE`T8$3mA|BNlbQ$l`SvRN<&RUx?
zy@#3RlGL@mHnvc!9j}OXgc}0F+ajxv!5CfRnuS@z<ut!*yT}!>w`A7R#O_-_aOp~(
z5eaV3s~4DAT^s-|^b(ZG0pB@iUXV51E}vJF0?-gZKqpT-J)+6+?V@+!r+f(2X*AXy
z9mpm%ZQBc%seW?8uG=}z;J3u$(^K~7qmQWNv&9wWR5n1=HDRr|7mH!yc%9nOyPhkH
z<;82+j34?#HO;U^7E*`6U?tc@VB*@vj53w3YytzIA*qrzit-P?d{BRPY5CXLgJ=w6
zKPgjIH#Z4*+=DrHuy2glm^D{MP&80;xTs!k=e9d?3J7k(9;<F;AK)4}2M4XZeP@0k
zpD&lz9j{Zs_&aOY*i{rKO_Ku)Sfz?r-PGz!VcEfj4hA`*#8by`-FfY7QN+ntC4-@O
zr8T{k>{M&e(0loK)I@sa^y@Ps;Be8mWYhT~U(aH--XD_-cUrSj_l!ZS_<km6FwIJ)
zen3#TvIfqEqCeW|EC%;Ok7wG|fHVG{_|_1=t(cR|p*HmmKfUf`d1V#-e`9AF$#fLe
zSyZonS6tiRUmHAZRLddc+yyV+P<wtQ^PiJ|I#5$dv{3;MF*?jIgC)@+)dA+IF7~G9
z%;8%y&mu^F>BsZFfL*o_n@~{D%F#9Zu}wKSM!fLJ<y*QxDCXmCU1?-O^_hUiCbwdr
zd+j5A)!ULjt<`>>mWL4zAW33FMnkxdDLzl!sO#3l_O*usY)0s7o|f9RUCf?va{aNf
zvEU=K7jRFY4>3gy2DAt*RpCTSCN&Y<CJ<9xEI!Qy3bt%?0$nVhD^}_>&hh113s|v^
z>3FX_pZ7RY<(`Zl9{)iJIM}yfz)dt}bowc5r32hyM$j=6&9$y)Q3okE1{>XDU%S^5
zE4D6%P2xkaFt1)|2PeUD<Y0g#AbpFFX8Xm2+elS)#kwT9;hjE0a0hq&wYr)H(Vlkj
zys5{_6RoBKk_Urn>j691HQUa_^nrk?Qy*t35~eA_`=FnQR%<1}6*ND(<w=|6FiGII
zT5_QrZFYrRT1pbH^+e*V{mKj`LzUlDZ|nHkmC291QCT4tszw^E`*5`|Kk_NYDHC3u
z8f{0%f?@J(W|YQo3+rCXfN9rof=%CV&*s5BZ!N?#0Gm`!vK2sbYw1en%eHm5eEf}2
zbB=PzKxo+Wmx*Mhr_ms^yGBY+<L!?M4l4v+l5~OJb@tQbR~!`P%NH{*2s;=^-jq~K
zspTIvzV;sKWr25o!}Q`e8gL&U8Y~X5|1S%P6CYiFiF5o5H-Kv))vnLFHBy)y`RXcd
zsHTYR4R6Jb(@T)AQb%`Kr^1{0XxhlKQ#BC}QK51GUxfPIn5CZ+-1Z(YlPf&1r@S4S
zPoQNwkk$ASg(~8%TjE_dmkQ&hKv}LCX+vZQ{G2*ZE+b}L^N8D(bZ-}@6Z87#D0z!y
zx`Hz$>R2Ip8AwTkEi_~{M}y1K?_~<{Bq0X%1VO8#96B-<Q5IrHMo~OmhflVFA3cAr
zxq!(m1goDCNC8u1;-XKJNfnuLA1|1B!=Qs&>LHA0-;VS}xche%=1l|eQglXM+{4c5
zK#`<9X5S=ys4>6}cY$Y=;KJ-);fKBQybR1$p#{Jp0WKO0VN_I}JMSOllOa4lk8kd&
zwcj%l%-7cLqrVqTa0hr<Nh9T{BdH<Tu??80?6FqBq}kHfye$AFD<XCh#)Ew_j@#Hk
zcpzgyc@mik=^pB8aL2+iB?Vo6kK%pcnKdsxz%LCPk1}f-ev{L2PC{zjIpacE^f;>j
z{W)QF*&OppfMuKVUUZiX!Vkmm)6TJc<*MmuiF^X;EQ&`xngTZwBjhs!Gd;$f%e?z@
z!E<GqsWPS;)IEK0?+%H_RIniHtrpw)``IeZDPT*3PN=M7sf$aaJ3&hW^}J|`yFHh3
z$=ColytCzWAy65|a4EMPV`|q~R3xfLNl*RXea)EkFQm%|@4tUIZ#wXr3m&MSsd~9n
zGfq!r%oR`E6BqY3B|CqS|4z63J7+OgIl42@o1z>A`*c2d4sqbGm9DQhv%MRg!w&yY
zLXBQzfx{1<U(1d`_0#w!%p>fL^P-jBYh}rW2mW*kcPCqcYC1Ro8t8_jXU}Mxh@I;^
zbNx{Suerb^+nBv#z=KwtS{6rEiITyul_8+i^ygPUuXPoT$9#aCFpvv}L2h&yPdJXS
z%UVRz2XKWnau8wj-v4W%cL`FhRQE}J2U~T!J17py!h=#|N!$2^-=G<=V@f24z|?9k
z1<^TIySw{~gGgU?0xcj1&Oyz`G<1gf`{`2JUq8t~c2&UUPR4~|T&r8AL-BD9`{`Q)
z^V1y7rE2ukEJ2V==?iN-Kf-2eYMry5&q!}f)Zi&4MbTr_?m{J}TTg<9280!Fx$vEi
zw`7%Sa)hfXXJ~21c4BcY0om@`zvvW`+qts*)Tba~BY%3RaoMn}egFr-GI08$>3L?#
z?aW60G&)2|_FEPfFu<kQuqY?K3GL3|njaW-Y<J=bIFeTgnJU0)sx|1hNz-DyP{X<T
zdCU}s@@BDA(Juo!$D#!h>kWQo>PSgFU>4udbUx#yAa*Uq^sh{PcY~iDOk4Th(EW@4
zH$3>)8tsptOj|Fb#`Vd6&;SBh6UI+d5IuO8Nq-oK1X(UIpg3>2)rjBm75%?9-I7Og
zug}UBFT`i6c+u#eF2Ll#!(PE%tQN41O5<@@mF8O{qM4T=&M9P&1CK1`Wf59|fe7*q
z8T$w~Jydw)RXFK2Mw})1GhNo%K!9|QT_f^v^Mer92;$i`Nu`^I1S)bxL^N5!vV>cR
zBpK57zy>OKIY0vp1Axw5Dc8SZ$tvg&AWnpn4iB8FvT}{Xhpw@Xe$<LVQ<Mwj7M`(B
zCZ9*zS{!CUi0IPme?j&3lCGPxu0z#%qxF35p`7+Je-*S-2ZP_TKYL^=6bAO!emk^<
z!<$odV`iYgY%ceZC45d))4Iw}!b*vmD}95+034IA1b(i8>?UxP#O!I|C+kVJ764bW
zH`;qW*xm0&_+8+ObJ=GU($l;vZT1)`L;gef`Qc%sBm&8*&LctXMdzQMSUHTSK=PXR
zON7n5<$2+6nFVK20E3<lPO)P1<MOiUAhMT_NrVS3n3>P_1bp5rv%P9$wDhZ_CoG0t
zr6dIpLVs>Z%(6kunLpg!Q8JqVzJwov07ji5B#)M^_|_h7v;njAhlwj0Zs}r+_R8eZ
z8RHdhF+W-UhzSwA`VdRZ2GM78hg?0pU~yZ?4(?{#y0X<~Z2KbziE~t*URb-BY(Qal
zWOC4(tb29IcJ(r#Uq{QT>=&Ad=jidL%{nY3wdzo*(G&5N)gAh%-`=X-B*d`RhlYlR
zmb`NK_UHd{_N=OkH14W3gkBi`-MSkQY2_F^(O9mapzwRj{O+IN^*X@ou{I7b?~6Lj
zw_c)|R`Uo8F2<q6Z}j2}gG^CvYV7*@J_gFAzfXr+xm%Z9DZk4X8HcesH%vdl)#-8K
z2KVorRpSA{&^c*TX;H){zvZW$f!@5dQsY+$ZxeEnoF_3W6uP(DRdsF}3V9&8U!A`-
zTnITBLyVPq_lkIyGPkj{$=Zd^xg`bdtvv1Kd#&sMz~5v2yP6Mey{=xMuB)}nS04S@
zHPNO}d@N#T2nv{w^_zaJnh+u1-&&}r-C}uaNG$PA8Gimx#uruU+_s07Nc99VgCg8a
z34lI;5{y(2WfPg?>FqGj-Gx}rxiJG*d|zS>TGJgaIbKp?Qd%B5wvZSOup)Ffru3@?
z5lZ}r_Run;oEAYt?xFXdO*Jr1%mQIga=Q>=l-qc4$<*PFH-sfuI4~sX+!sZy-%Z$M
z^X-Xi2(U-403)wJ1)m1^Fqoh@FTDZZ;YV8C&LN&1W%*c-u_I?lk`LbNp!FDVXR84D
z=wo#sgLg>PlmS<HJI0>+t@m<B$6!~snD=Lv<9r3yoOp5x>`%5$#;GY<jJ%UU=)zF{
zWMFC4z5Jb-uh?N6Sdpl5BGX?@99VA3TTOQMowE*YjcC$bm*FZLRxj}<IU{-c@yURa
z5X}%OPdC0#>?Fl~rrPDz`T+B#-+y;?L;Bp8G&rES|KC*TZw2C&<KK%Zm;Hqn;<bCa
zx@%?i3rQL`D;h=%iF6p?1{JPffGd@_0VYi-^VD8rg%@8*kleS^wBU&nR7;4<G*Emm
z_?C_o_-pwR*^!u(3O}VmnQ@W>NQ&}OCZleefusylq|<daKgt0jQA4Gko!<c`jTv`0
zopPw#0*44wIqLHDiu0sG69LMw_b&pJJ}^n8vimuJhfv*bJ+xaC_aXzuZx4xQ>Y*sp
zSOFZuALkTTdJEPF`tK2!<!VrAXV@Zo<p`W?XoKfWA-!w~HS~|il+wuUeM;;t0iqAk
zQXR!4PqWHfEiM0{??rZ&HRS{5KJ?GqhSkfCx2A3*TP(Zvs;-%l12(!2ncfiL#u>Aq
zA879N&WXvlE#G$+Rs)r)1O{KSh`kC>qjoC&9l!H7J0+p|mXLrz&%Qj@t9}*WA1Q_H
z4wtOQA}ZS8Qq;b6qdJNuOf&6xQJ<%di8K;Rj(Je{m<zmB803&vq@+0FHVpFl>bh?@
zT-V^>7i?B~Bw})B<ZfC08j)vp?K=BL+I1`l*Yol5;^BU{H+abT{T_)rr=J5OZ0kpv
z_U>dLOzT)l=R#6y#x83OGL+uGX?+a-`?{cNG4`D?<FF$3N&gRT>6D3=ye)U299wJa
z^kw+w=SX*)aT3-`d=O+$Fd7*dImoT3P%zBKutuzB8~Z=>^8cMN4@ZR@DTbP}uxOF3
z9;|e)8UH;tcKLgAeFC=xYauJN=irW?y>=d9;#J<-m|IK9s!OaCI#3|Wki=teyols~
zFu4Dk7$%p`ad&m&YEgvTvhydroI%w8=-<{(!xp}SO?B|7aExH&+vqC7e*BmglgGYf
za9n!_+%feEo#6NqQCgBETE28qtXa8G&$(00wWustVAR61R1G4e!lly{)|T)~v`&sh
zPYgOTwh2J0ga9OZCR3CqIv<JOc;HPnXlrK<8D|;4<O~KWg1Y_B8W0GBohfrd`;zYN
zw;*YMb<W(M^2ar7M&D(oN#9^vuF~r5t|**7_r?=R-Qs$pfvo#+`+SskrfwTq`bf?v
zb<RN0lS+QcmFI*tPQ4A<5o@{s{Pwe<n00FWYShTD#CJA)Ty##l-ISYEL$ww^y}v>G
zOivSn?{?e<SXl0`ro+69?g%1O$v+@|$~h<{BHtZmS8;o1Xt-~bm}F|I8FQSb@5;}U
zK-$06<as=$0^4N-IFv0oGLnJvfNKUxwa!^2CA5*Y)jizrWWQrS>_O*CP)icuw|(zW
zd3~ed3n*cC%X&C7E63H_lk-|Y?hr#gOjka9Sy-I6>2>JMxXXN8g&`ekq^3J34|uur
za*rs%0cK2E$OS(EY@!@X+y_VH*3CSg-OclX6wb>Wkkii%Ts<K=F!=OQuz2PUJpVrx
zNKUm)04{XyU$<Ha!NqZ*hIr1^s-WODHSOwk;osw&okeSWSxH<eQ?9AFm5zj&j$|jR
z+6h_f=``{IO1`(US@IpyXOj>tk*3u=>cWeEcG8e6aux{a4#mLJ@F!!M-p1l~>~^-W
zH-z_>h_xcrQil4lE_E#l!0=Y`FC!N{h%-&fO2vi8?Ov9XTemB<Mks^C3^a*Cjx!l_
z$%GTMKptb%20GWM&VWT^s~cvoh{1BR_5i<^N`rCOX36uNC{BfmDW8Zv1KkGPM}z_L
zPdb(r<}B6MOA$gSJdcUf>7mGa#Bkm_%r}ppL0*GF?k`0kWG337A7DBqkqUINaZ#B8
zqw4+TIUsJ&vm~(ne#<Fi6F%y8IW7LauB{iXMT~SkCv|a5|MxV5!E~$S=xePO$#pI)
zg+XT~zo&l0SPj!=>%t)U5{_ro-uLEE>U4($A?PI_dU6Lh_>3O%ouW;Y+*>K~MWeb$
z#du4_1_PAe(_*`N?F9D*Hhe}NupclC^Q{X8{94=j6z!4tr0=?PbeKTF<%?8YWut$)
zT9-1xJAFs-um`@3@L`r|XAO|c3>53BzaVX$Jq*oEnKa;~4-OSsk9xS~srDXW?_QG?
z#I33e2(nm~{fGxLmg7gROg}QY9God#CCSQ{s$JO8fqR}58rs3F)f=?KqGS1i>18Jc
zVI5=%^NWj;ctM2nDEFsuz3&V7i{>KH|BMLjx!t+)dH%bjy&Y;6y0L%7K2s;!7P+_Q
zdA{fOLsa8pG+rQFE!@GFHnf#Eh@0&Z()IQSf=x4_UxC}}k;{Jr)j1tX5<tWwGaqYB
z&5i+%#Vg2-2SN{W-rfekMi|M3piYjsKf!WEAyVprFg-ThiWdye#6ADQvlZygmo?9s
zc{ehLMmM!5C}3{d1VR2G5-ZsUOe*~>Xtr6DCjWJX*TUOWpM`mdgNp{<i0}$>O$#JZ
zWPH2sajaG>8hc610W^}4kSI~}J)taVa<Lbe+?o#aWpEmM8dqouN<)$wf6MEhMid}8
z_ue0C_=u;Jiu>fyH`h00xl%uPwrS>-B)t$`UR3&Nss-1xLG02w_GPtcVtdn`FVdKw
zrsjS?{LQKlapKh|=exKs);*a_p33%$*q`YJ%<|GuFiWvx_YGmxt3iVLPtkkT;PHEr
zWwbHCZ}ZfKSxsLY42a|vw_<j=#pOkv=pmU$l>zZ*T}+_qymt=1j?g#b{fRMyH_nbr
zdY}_vejjqB^5LVWP|20JmAdtMSSU7|AbhKV`K7-Vn^8mO#X>XKdE*$T;{jr`vThXK
z;5fAnyZrD@pzYqJ9Uf&YC^7gz*(hojR=5yxBUU`FYWrH@VSWO3rd}S?Tw{bT7flK&
zC?v<fR|s$AcS#MOhUnUHJ*r+IF>IBj8_rfAU$;m?+$m4ZfukfOBx`s(eXmu(dtTnL
zv9Y8e1~~u!uf56Po)5X)J^a%_-9MD2vg^@UezUiKQ+|x&FAgK@MSQaT5x~5IA;hY2
zDx11uc@lkJ|EbVI*i5oZGf=U$b$L~wc28sWUoa`cq|pMr{D(h_F-5(N?#BT)J#=&<
z@3|%+EJkeVbWj%0I|h`Zw~`DmSQ4Cup>lvpE@rSRuH==cL^@l-tFG55PK<V<nH4_5
z*@p?AN%J(VpLCPtIntlV?Hls5S;c^z8<8{dhY^#d-?{<>0bJZt<00xGBcYRh?0VXG
zpV=#K>1yA$Un2amqN;H`jY+PL>D$`5I#Fl-2A#6qNXHa_sB$uk81SAzagXe6i9&M0
zoeOib;Nr(oveiGdr$S10Y(nJxALDc8f{14X*}JtPh_f6EJzU{)eJ2tgWF{@h%beOS
zo;%KQ{qNgd-p*2XFP6rj?I+24@&da!uUbR2D|#IU;(>(>p$LaI)K$ouGUV||b~Zn@
zr!Otv**(z=F*v;Rsw3En@Zxa!H>%flLZ5O&mOLGt2VFq>G#IiiPI=fyI<Vy$kPlHC
z-u*1<PWh_qdD-xck}#}@yN#Jj?WAwL2{z8ZXvGXE4B`Kw+D2<-4y|lhe<zvYQU9>C
zL`enr<l=#Bku{%_tXnI76xd4Evk`baM*zYt4E9<H>@klRs%Zf#)Tv9*Gc&>lDRvHl
z>Bl3}-2ZK8{xkp64kV8G-xPxzt1Sb^i=;3mIdyf3e<YHi4VA^j@^XpG?54f<%jJKU
z@22yN;PuHE+dy{5nbQa0+GipFxkr^U=dNkt<Al%;CYhPF3Ity6!@%2sKz!J;B~N@r
z<U?sP<VTSxRtlj{;aY6I&Maapm3`qqn@N+xtn5I7RPfFJN7Y&QMb*Are}+^fm5@dd
z5b16hLP8ppE~R@07*KK$>6Vi2l9C>VmhNVxo1u}Ge&;#A^PF?u>ra@^z4xAbt#y4@
zv0UQ1uC6#y!_%L(=XJVA<(d^N`L4S**8y~xPi7(+F$Y;TjEe(89C`pM{@_jk2Oib}
zAe1uy?V{|Ee`nO%_jS@pAU2Q#q>gcO8dA2E(HK2IBN4XU^LE@@*a*c0l_^|X0&`Re
ziJ9V-WzRdLxH}uLy2+0RxE6}t*NV`55|_5y>Curdm!YwtRkS)vDuEIXy8k_hm>3s}
z%5yH(@Yw?%gXCul${@HCcOrf(-atB_Rzy4P&)VZRcfD{RTcFmJ7uevu&cIWrB+DXu
zHa%1T)b|!CtTryBmoIu)Ju&=bET05VjSvR)B6k4K#o3#_Jly;~(v(Rd?k3hy<}z!F
zhHFip%F2<7hs@J|m%!bBKf-jrGBHxI|7KSr8b$OyI>MR?bT$>&b2WW)25Wtq{9O`P
z;mmiies^Klv>&nkNs1>Kq<j>p63gGg59~Eg0S~#(e18x#3F%rCR$dTC)&~<_Q;lZ6
zuA_e(L%I+$EAS`%lgIORCtqp6JCH{hm}_!sb(_T3nt4q$L#h-FVEBj6`tO|cvn)V}
zB+p_OKBF@J2^FmS-^NT#O<BU6!=K>_D=sz@*Q&owp&>kM9QHKO34FUM*102B{t`sb
z_A?3H)sY=B4a4-ZtcZySTbb6PLFb6DhOp+`U0<Gc=P3SQ7^e6kuO`qmGQ5U%Y@}Tb
z03>X`Rpf{HN`r*;%z6_QlP4KT^4C)_7}kCPs}rZ&CcG~eZMuNA*4$5*|5krh&u?6z
zhxpSfx>rD>C(L>-pje+C51HTVMD4ZEUkCV(p3-GA?Sw>nQ;9O*n19{d-1~iAkNZj4
zQ+2@yRR2Zb*JGe1@Wo(57<*Zm9>WY2Ej-2QEB@}6w)tuMX$*v2Q|mAyWl&X3#b;f2
z1flWU#OM(<5l6ZyNXuO^@D!=$Pp?H=R9a4!=s=Dr<T!bxJsb7o3EbF{K)f^0V~1N@
zuD%!Ki`i3deIV5&iycf~ld3_48T2MkW=*?w8t*n9_hwby?W1;Dtsh9kyaM6t<UcQ=
zS$n3nBe_-<f;*p#&?(5QQ2xZrLe1x&2xB*|V3SWXyksP3?I1g3QGw3or%CGhrE5tw
z3Y=C9DWiGg11ZQa@!WfbT~WwpGsR-*<f%M2ca1H|j}^17DK&&B`U-Kxkn6!MYR^|}
z=9w@ps_GF9D__y8k|hvm$>>62Bjxi>dbG}^shShG;lz=XB#H(xnw4l3!G+{9ZvJ#E
z-#cB*Zp82N5e+E2gR%)h|7E+_ZQtJ<CF9cjXxcjj&Tpc-b&{p8^Q2SXmG)stjAbLK
zXf`qgro{4IeeUF>T1f`M-ch}r<efP9D-Tl@83fRPUQ*ElB#Nt&;)?!=!aj_99lJZ{
zr?HoKun=TYo8i?JgTJeOA)^EyY!im1;Ukft<N&NJTs@IA@ERa+qPA1^YHT*KqR3my
zA$#oLI>GOzx`vy*_m-sCrA=djxkba+nfy0JJ1)q;CJ)VNwtOMYg^n4Q<Y-DG-xgec
zz`jLs@&3lO#2Y3lq#zVYef*b9TyDTv7xa2JX!V+^`T1^6B2W}>?bNR)h{?8)l6FY2
zq{BzZjd$l-DnH(Jao96F-COVBzF6X<;o0Rh{u#QFDY<E*_^vmzAeh#~GypRDdHBgn
zBKZ>cQ({GAr0@Q3izX?`5oNsclZnflLcMb5tWLWgDh%6_o20m-z1kL~7SZ{6qYIWw
zOL@biwcwR>%k)6WO}i|ZAYisY7lM@h$W~lGO1?Xq6h+3Xu|N+!0w0%u-!DGv&E!*3
z->q%D2B-;?^@($%*VF=dR?-@s8Z~^$x35DK&?PWBBg%U>t5?+@6>Ux9!W$)yjp6=V
zeKT*Rv(6^%5DVHd6B8<sx;g%W6^Y&Zi~NBJ01hlng(N$bVmXh?VqMB@*s#2j4>$q6
zBM@>Ugr#4Z!8Y8hYSg^A&RNXtZhucO3i1anYth?xzmE(@$WHMLyL|rlej^pSJ>j|R
zdubrq_zuw6u;ir@9L&dq$x2p|g!NmL0R;m=PX&vSM|zFLnx|=7)twpEpa_uXV$_RJ
z=~Dze$}>1y3|gYuu=6n@xZiZ4S`EkgOZ=<HrX_5ObF14RnDP4w*b)dE@QaGqkt`Wk
zOC-ac=Gl(JtOBmVYLe)!{-=yNy9D|Vs#|kuuh|Er`CbwBQnJ;HoEGNOvz+9up-C0a
z({vdjSn9QVqsU?Mm_SzX-gdqfMmO11&XJ<<sfNcueE`W_hIZg75AJi@vDK1;5I4tc
zo!e`pvovhN*M5Wm*IQAen$(PZK5}3p22iSmI+*>N_i*^ASm_?fi+9x?WoNQ1Sbce#
zX0<ujAKmjh5wQWkzmm&oV!HwIh?g=D2*wFA0Q9SzXRR?xlo~DRBlo1N3ET}|r#%Kx
z&l9FHMc2ORP2~&r@hb3Jo{%TWCbN033qx$(traV7yWaxT?I#isjCOJ8J_!i?F+Ymu
zGY~gg@M`fOv3%PN7eU=V3+=hr2x^K#VS?jqFp)slZ-k~B(U@EfbfgWZGA;UisP<3V
zdF4_cftl^yIJHwYfr2VZ{)9li&&GcoQ;SeB@7!oUiIAqm;~6|l(8|yOF_sL9HgHrU
zx$E}YjK8y%7U%;G|C$bpe;nomB;^w)msEDrBwq}zq>tBmQ9%W0zZukUWtAtg>-8i|
z|0gZ`KWU`M-5Dd7#H^dftoSz(vc$^^3mpCzHvTN!ZMW6F<u6Jadv0@%BE~*Eor?{B
zH&mejS;xwmk9SFWWP{IE3X?Pzn&L646VJEy$g1<Qo*4o*7{6Y@DRCtKio|9m(0mD9
zBIZC#stHZuUTRxCdlDq(WbKuwmXG#N+W{2R=GeqalC{;j`v5Xyk!F{U*BO|!{*>l&
zs3me?n-2gEtb+(-Sl(Zs0EtCLAay?E+ZtZ`^9fOE+DGk-m_5bZI6Pp7N(!O{#&g1-
zI@o%|hi&z8Il&Zv7>Kl$=dGfZ<rcu+#ZI~##u&0cr4Z`6(lM7-ACybneVU2-s_&%S
z=&Uv&A6NS<)dK1ez4ic_x|+?y5bN!BokSte_Jf{QVD*lqzHf6in?;ne!G<qjo$<J4
z>f^eY2#v76`+%g{(~43$^OxR5Y_RDsEnCIOO8l5&Wz}^1i)H$(hXDv@5Z^x}bq(y{
zlHx?`bCUCx^qRtCv|BS&#^nEYBSxCm&trTJa8tdWEi+i4Tk^sx{d8*HT3y%SUV9^{
z>man0C{M<V_Avg`Lhur=r=wBSpJ(4G{VUtqCI^5_S-QUPk-E<f<ah|TqBoq%&%;y%
zvt6pE<L1do)Sf7{m=dNV<WaKd<5|}0{Xav`DkiHhk126@&#z<<m95A)@+|2<N`i}8
zmE~=Y9qKXEhuoeX1p|b;-79Wl=$rpYCAl9F_Wx1SrZjY5@OkGZHNRB~zZd9zjd%?y
zw&k0p%Z*<_Pp%Vj`tCB;oNwQ*Jx`<IS>LLjs=*713xkoN1}27OlXVcKRL1k}$$dc<
zR9y|Y3}}ds*RTujZ{Z2nio2hEJu#{vdO!xN)zf0l4P!n#Kx9PeG1*<2xLkM<&dJQM
zAg;*1p%07ma~i4lIp1)moKZ!x_N+fSv+h&-Fsa0&BXD7{x_Ir3l3YZ3#DH(4CKFyd
zAq1fC+66)3uX6?|pyj&lg+Gn?nI4&V?T>o(u#IJ!k?4k@UE@t-UrgoH*8tK_FqQHl
z$eUeSw_uf+D{gNy)3V;xUO7)4?IgFldYbdU?ljI<<dF7vTOvT+|LUli85&|%9I($(
zQAhrc>rtiUP8I!`HLw+Vj>!l73^ph0(+={6rtj1CJc<1A6cxq-+;g6;llpkGgbIkv
zq);)iYQq=<C>Yo)4G8`W^+Fr&j~O0~g&S4i<IFYh8pd6}#sAf!h2nhDak{p!w!1Ip
zWTHf{V%2T>aTKr@P!EvbGOvsF97xB3j|vimGNpeMW9fbVeH}yVbUamWYB(+ZTe(d~
z)<g{uU_p<$4J(HUm5%FknU|U9J@n*QS9E-lT!9?LRfwrv8oxJ`IMv9IWMFV!tTrRw
zZjC=Z!Y%Hb{peQ&shine`m{bv-&dA8xQ<RxXQWGd3l)}|Vu!OoCxZqRf4Cp=XZpA1
zPf*8cxz)sPUD!`<rqayC%&hs7R_5m|v=imo=3r(N+jdDGJ~%y=8!L(3%em;_tSmh7
zPL8U09VfS5=YWP%`g(mOykk?LrKB-}8+WJVRYHso!6n&_71mM8mqXo$u&he6Vlz?^
zc<iOFr@F2q0o(VxjE+ZgdK1k`bxh_g**!TuHWZI-o`M4zFYYM1RxjA>?g8#6C_;Ti
z+H_T?iJq=>bU@tO$37}1ps87vCrdW|I7mjf+{;5CiAWaymm!Ac&P7FrtjgJL)cPN>
z>->rxDo7axbZ@{O^P8-4>dUM)%lHEG5hanp5qeT0k=L#-qR)&ncpGje`3P+-)7y>6
zc$_bXCk2T5xBtwG`5OR5pi;hS<pqg`Zc(-TwsA0s%`(2=06<n}qg)2ZdkA5*_LRwY
zUdW(|gW&y9Z+d|R)0SV2WC9s+#rfhN;YDb`#BANijQN<a{072$G)Sy^gTTnAKp-vx
zYDzCekyenaCzu7An!yyPq8IgUbaQr#7c$5RrW#SmeEcR_Ykt`a2NnTmBVj9a7jqhu
zG!wD8ZfUItgi*V5yZSxV;@04zA-D;<n#>aJQ}Eo_@sj~ji5yB@WJ677y!%wl8fr#T
z#x9?CaZ@ryi;oI^riU^Qj=acD+@c1#n-JOTtRQ;Gb{(a-o{)1Tt&R_7`QG1to<%@Q
zBcsWY<&g+i(*L{eHTrXRZ_ny<YxpGQ;aHjwruHT<#KmWw+V%D?1oNGa_AcykULbj>
z`ur;CIn+>34agB&t?)Aq-Y9`Rk6aX@{Iml21(gRhugj=x)|z`B2mcXsu71LL*6Sdf
zZOSr1uN=b+ngK0-S?wkeOB%$UP>YQEcqVu4TVb$jc+8q9SoswDmVj?fo4-%!edd||
zq1$kskXe4ALAX-@IBoU1^hK1r7A$Y(xA}@JfR!y@uAOJ%Z!r*gok?h5*6F0XwW8cT
zY9dClMWv@R(l6H{?r;Gzq&+X6SuR`VgoU69mzuZ?zwTc_*|y|&`JAga=et+O2Ku#M
zI@l9XH3l5=v{AO#T;}vc@k_tTIcvM&0(MqeF}`;;_Q4~bt1!9=xU!&wT<yc8JNJ@B
zr=yfg7s4gKFZkbJyn<->yduWqQO1NxBv7`#oT#uFT}VgJpKu9DMSRA3xg&9{b0b^y
zKBIx}Dt;8Zeb~!jJr?SfSsbPtjWrbHCDYqM$D?Q0(;qTO2=AxEkj<T}G#G$>=-up~
zbJz&D;fG>fUs$XKln*({5D4EGZv`N~hYWeAk3<!<ns-KdEh8QvbLVwIXLjxWDX_?`
zV5b`I(T7|#Y|{y9>O|lkRp#x~kA|Rqzz`MHnF?l4(;2~s#h?ri(qreu)jW5P<6ek@
zZ-?@A`nWZ0_hW@N!Vs?i)iA4}#Yqax%%OUFFUS2q0;}#5a*o*L+u`MerbC3v)=-MY
zNQ6aibbs4j!-K9xFUH3lnR4l~@1nl9?r0zY=+Bd`c@K@^W%;?e)IC{Rf$&L8lOgP+
zM?b-5%5DXK;Icz%HA?Le8t3OLsn|=*e=t|?I}+vA12b#Zwq-?K8t?l>uayececSP}
z%?$WCCyv2qK;pBQj#e}$6}5rw^CCm*X#6mcbTT9BQAq}NKnwL0C__Nr{yj4cM=v|?
z1=dJWq};&di})T=cr*BJTy)a@RxQuqJP9L$eaK$V3i}gahQQ>hp15#+{8}`7pvt1n
zPyy>{(yqu^OHli3RvI%gU51$H@xkc{e%nF%bPMCtt{TRv5et*2tFdWzbFT*q%d3|;
zTBvxG%cVeO*b+5U_V@LKMc7A}`?0zMb1R>IY3aX5MzjzTV-eE(zO`&T^v0Z|{2dPH
zyh9)X;^|R7Vm#nvFUNhdi56v-jFwkZ^c)i)z2Y~aol#IE@4Cd<q=4Luy5w^{iq(O|
zRpY|7Mv&JzBCV63Q8}${iwWCA&KH>i*mWP_eDiL3Pu~HQS`bPRN(IT!*}V&P^SCmQ
zzb=h(S>#+r0K}vMYv+diQW)<j9hQDA@R>0$>X+AWtz%w5O!*EMlz$bd*t0|FAYr?+
zY-Dco!$altL>9*J4z}9aONE|ROX;)m@Z$oN<c%I`mq@{l3XK0Ni;ywH*YEqKuUM-;
z$bpk%5!Yr9d;W?40J?E-xQfD40GjzF!$v}6&YOZ#EbCMFm=ck8#Nq3~D>LmdoEy?)
zT!iKHv?Unl693Pj_No$zTAp?`p)+f`C^NxR)%^1)f}klLV+E1SvRGdhe2hZ^6MS{a
z{H*msV|J|ZW|v9D?)bGI$m0dYIHq7xdp{$IXpt{B_lNm2h8=Z{&b?ZJhJ?bpcv1*3
zV38E2gc0p%`d5k}IG7Mc7-|a}^-*Ol^~8Dia#-ZZt1N}gv=`F-3H-903l~S12;cls
zEq}5WIb}lVsI8YEEQ0e=6*6rQev~5o)fqv0(jcI}5H5;BE``|KN9-!LL9eS-2$9Ut
zdFx@JSWELdRk6`^+%|6A{H%BH+Eq<qyxL?o)4^^40WKA&Zrs*EfYLs-B(AA3*xd@7
z=T`>~yoW^LGzvfTquJGI2$cS8kLvKzPE1BozV&5iw}m{Yds6Hu?eI}HQS1*lhdoUe
znrx9SB{Jmb&#6`XW-PqH3C=A%H~0oRXY>GfaHqLS{RGg=VD-+}dwJq@!5LI%IRY5o
zhwfE0$NI)LO4BAoORSbsbQq}FijCU_*h2SW=8_Q}nS|AV(fwP-kqKC)9-3S5e9Vx=
zgvm|2x=870)Qr6}esu5jZC2V4CQt00c0rB`pM(HIkr8y5&#?JZRb#RRK_j~v1Uv8F
zD{gMCYSzOoYI8M&J}EeOg{mXq*yVn>t*XK=9!msqc&dF`@L+NM?hOHq(<J^d27r?)
z&auqJ=UJs6;Fhh?%EYMtl;E#Pry5ml+fy1@tOpng#s-%8^!|OVDuki{qqeS@cB$!;
zFThzv$lkn7HxQ+2&l(F<!lVQ8tOfUrf74$5`72QovAPY5nw<Fj<O_It-dT;-9?UyE
z@>r7uuJe*MPx>NZcu)RP4~ALS0V9qSkE5n1Q9P)CHelH-YgWR^<}v=HD@<|Lyfeeh
zZsav{BtiI5y~>Wmy3{WiiT@CNWsJ$dtX@;5w3io#d)_2O=FVBb_2N&kX!{j9Zi;5q
zp{s=h%k<w@A7YC&OP4Tv-h;qzfSN!VWhaCzuv`qfyngZ8RkeC+>bME~_Cb{EKC<F5
zf%@Bk$V2WeGo*^p;s-(E@)CJE)2b{{Cik&vUa=7ca``GV<+%Lu{$Du9pv>#fYbiC=
z30B%FqBZSmyWi{fi7Eyss&KTJ2t_t+TR&DbSVjc|fL<-+f|vp8tE^}H+dh2364_gO
zZb!Q|hvZq&knDL-uFSaV>)o%gk04aN+#1Bex}(YWKIe}rkY>-#o+frss_aWysJq>S
z0#p^}m7Vi9U=+(%(|HSzb0+rtHg}QJR8cFd5-AEmh?SZk&beR7vWNsRO~ugr*&6ml
zzxxzS?!krYKT->WwvLXf^-hyF(TKHwzc#l$cER@gn%~4o{#$E3Z`1sTE4j%;f2D+8
z#fkC1wYZ9C`};%`y8G@VLmpmc{nEynuVLL!r5|fRb;=g=EdT}6r8Ho|YNz<~b<X*q
zSN4N<zRVBl<MaXtsy-%LuBXs7aMdaXV>S-Xc_tCy1v@*T{{$#ZBV8wm>X3qpf9(B4
zPh?u;7P?(HBRZJK-8Mf5Z6eqvXnv!!VGDwP2dfkiiWy6{lf$ff?yRvlAOt%!UU$YC
z1dYw}<?=YplL9Rg%{ggonuNHtmpZGNxjPttM#4`pElce^&>tJDwdWE;9Sk6LSY;c(
zNbYa`DLwW8)h5qvn^Dc1Oj~{)$D0_c;QVA=_`~xwXe={9m7{dAk4-%MvkFJKzK8E?
zSpSz6VtKV{if-nebU5+ih=w}0(iA0=+B@m_cqxKU!*7d*wNvaO0UT6^HSZh)7v5mB
z#|7hmop6|?8{4~PBu*sTLF|~ehovqO?GK~%5nwocUFl%o!Di-H<50xtmgHc)!L?9z
zA^vTLjG=q*6<xNEDk|(wvj4m=-h33T*;}9A3<YH7jD2Dpaw!oJ&mYqnCE+o_>Mllw
zo+gO-=3lvHz4`N0ku^$B?0i;fu~)))xG#oq&kcm4Q_PfVD4Q`4N_!gR-q5GyQX~M~
zq1hHgQ?+__dsZXGJr9jfDvH?+{%^O{8@26!7WhVe58Ye`Ig@4D>F>tr`&)f)zEw(P
zkF+|MKr&&bnBdYT1=zge*i9x*H`RA**s~od9XJ>G)2I;WG5?`-742J*bV%;_Q6(7W
zoIm{>-mv}>Ll*2y8#2}*Km@Yi#kU6n7G4wF4uVK6aAQfjFrmc<!EQhLBr!kbTD&I4
zQEI=TD0`|QJi)jc%vij9+Birksx4G*es52^Ze?p8MX2O3RV6lYrbLTVp~|{N09U?~
z(%yqn8H@)#vz7JQM}r|<glDN#6PL)2EVz1l3#?Irz|0}}eD-x`47ydX0X&<9ar*F5
zGjj(NkR$&Wnq87{)6(F<Vy4PX<4TJP>CReo0Zdal@HANEay=KE+ogD`gNj|aEEyf}
zy<B8%I~8wZ8kM?n{51MI_EFReYqU&P`qbPQD^HvE;UI+b{rq@iLSx<MK^b1$_L(}H
zhjsX1TXYlgb5%Uy8;lTg@9LqFA1LMMz3(L*J(4;#YM1(Ok1$vsz&@L_q*Q>cV=)=;
zl0}oa_%UcpaAar3JA&38PRlucLwLv7NT2=cCetBSES2Kt<%FoO$M*SWQ_Bv@c2!h{
zWS8HE<+<~QW$yfF-5=<7Qf@c8iO-FXa(Qh4HoiS8q)>F=+XS!M5M$D=M|&0SzV&Jv
zP0IO2#WQm|xAl%S&-Jf+$WSG+yw38ti`cfS#_Jq^#AhnR^q(VgM>;)mDMRCWoJ0NJ
z<bn8cR;lC1DrRcXKPxg<gKf7epYN|w@kc}G@{6?_#MzM_WKbwV8eQQ;HSy@zMDy8A
z=Eq|L3R3I^WksV#o!jluPujOL#=@&VI+Yc6lO5o*?XoHGV4yg`-x9Up`5w-3hYb_9
zc9C=us(mj&e>G?SlXxpyD<IYrJ)=7=|BI-TOAJhn7V}HTMg1Mpl1n);$DD1Ls3N8_
z;0YK5MW99p*3y#W3bI)jqt8DXnZB%e2!)n(<p?r`Pzly3QRfF|1_bu=wClLI^s_*w
zaORu!ESEMB(D;I7%;!KITfBD3_Z{W$Km*Pu4Xqv!q>mv|0$q%>^S1};|H;^F>hM>F
z6<5WV*y))5iYi5`Zl6UX{;9lWkr#YpY|a<AKb@Huos~zOB~CK`qWci_#En|-CJX|>
zw^zvzCuUUN*GEg9vEpwa5}SVA=U7;ieaCuxo?g>vV`j8ihtZo~&fOmGYdAGU@Vpnb
zeXTE-pyHHM!?yB6$z2Xx7c$x|rQL3)&D#%2dWMuf&mFFrp3BoUxwsQCf{cxm%3YRa
zSdFk9>a4uiO_!Xg16GvqlgCXA-*GZPYs|BD_yfGYc&KUXOkM)&z)kh^<Y-jf{%O#Z
z`^AS;ax)f?bV)rxVxlkE>~Kk3S-_NmHr-*}ay^plaR|L2mT1<V%ac>xmjJKcu$BC&
z8xMebPyac)z=Fquy@Cgk7bP84lB9RVwB4iAF=UbVOZ2XA_cNw{!G5n_@7idk2|0a3
zdo|#v379>1$Mg?Xq#h=m&)Oa3F?$`P^2}@*M1A_xC8jD*L{tPmWhtSLSk?R6#!ESd
zg?SdjBD|f?C9gbU=3l8Lvp;2s6?o^Dxwm}-nR@)^CH?KQckxR<kUtuAk0`$Dz6@VQ
z%PAL>&O+&S%&_6{fCP78P)g(GQiHe9_t<zryD{l0ic<jLSMf*SFDk?KHU$<QYeYWW
zQyrZ-(@63KEp)YJx&(m$AR|)r2=8NApqS#dM(I(HPjbOBkV5IIlef?Nhw%pkAd}%T
z1(5(SVM_-a7>iKf%EUSx3#!5TWQZ=?yn52mn&e~Qbuh2T<RU!NSF+{t`QldMv&XW^
z$$^W%(Y?#Vt=WL|b=$3uTbu8JgIPSMQGH+1o)qK!efM7VxmgSQXv?j&YtSIZ$l^no
zMqk@?gePY=lA*u)>7~W%iGlzN^n%5N!lELy5!z{2=#4dIPoM*?v&x7B7>wY8+@}TC
z0CVl)6NZBTYPOO6JLbJW@x(?ohnB>!(=TZ;UJkI1GG3VPz<sY($a)kCcx1tyZ?|c1
z?8ops{f+XE>H*M<)*?p7rtV97F2e)mMYH(bM+hp)qRx{Or|WTx)iN}Ji+Qx5RfEK$
zAP2p<B%B4c8CF?dKQEi-#t`vzkK%OwQ=(K^2n@OV1%T^&Sc4bFivv-sf5G4z@{XA@
z_Xp&8{XoB4hW22d!}M!A2Z#8$IFh~ambY!u|9^t<-4rRhcM<wul^}c$x&T(JKzHN2
zmgO6C8E~O%!$UZ{<)M9n;r{pJOyan9vzNL9ajA@z6i;`z?7+(rN^J}W`)<xr7^OLE
z0Ix4uRNEFl_O}{N&_HQh6J0C<Emls#l*u1XoqrNS0?vL_jPtvWP9WA2T2w4okahIx
zRaSDr&AI;M<Gr3a+1>S9@WGd3F&cj{-L}J|u`1G4!n`XCX}PziucaF5)oH+++NFY|
z#j{=-qL2Fe;)I@<1C~_3{J1^};4(?0$3>EJuZF&67ts)7Jp?h3Ub;9pQ7WOj;y;S}
z8-Nuy--PdWCuuswp7lCihK&l}fB?^Sb9ba>Z`v^os=ThfHO|i;Ac2lI4K|ziZ4XC7
zn>etW<%gp-riU#5%Ljo=Y8cmh3skIt$60?3cXI-|x4>xmI;n+7Fw-e<`D$?e<oqGG
za@2^+IW;xBbv=LS=B>)5dl`i{qrX3#jYzU^uf&_~ljBdZpem(Hq9`^P#EZZ<3Sik#
zm`_IsrX+rZ&y+Dm<teaRS0eg-5mU4(1X@{DNn_$)BK{q(z0kXRLVtF&9gVX4zCOsw
z^&K-kW#7o$&5L#>hOlD4hH;h>VToF9_x31*(|lI=^LfV6LP8)b`<Xpuf@;Cd0Ol3C
z2sL5jDxM;1{<i*pk(w$!p>Z2lBw(^iWLG?-lcK~gee4KWgrmzYZheBltEM7u@;Em)
zR;opTa}acAObv<`0B@i$|2N~v4&pTt{LdR`m$m#&1UcIMTmBohtkn8!;_Cjq?H=ih
z7VGxTgqcWgF<=0@{PMUhrMK3DO=}*K`7-`JCHz8+bQB~2AgkeN+KXZS@=1leDmnsN
zpyGS93U3T}Wlo;ud0m%JZAr)n%vBnnf?OGDW2vOt#^rkbsgwsdu9{*#T4`o#;p1k&
zt~;JtrUGD(h)GejYrkY5B^0%X6(IM-I5Q_$aBlfPI4DABOT8b*RK<_9y!DyJSD5hz
zKz*=DoMAHdCxpO~m@ZdsFqwNB=dPpJiy7SHwZCv^0cTl`9>a3^9<SId=gtFN9Ug8f
zy1v}Hq&b_tvIn8<R0@XJcM2;iixRUpJ3CE-g>$Xf9}IYG>{GW@&*$6j<}WV!M<tky
zX1;p=y`!MA+Iu)>v)iZnS@|R=A_`$6ScCDBuK88<!8&^N`OPWzyivOaVqy1A?Wy^v
zKiamD+5Wfv6!;@a7zQh|0c`?$nJvAkdE0HA^?hGj)Ah|pVhbF1TZhHOXB8u}1O>81
z=q_IVfj<ge6=0$jb8E}znxExR@wAqXw|?TzPhoFemR~jb91BqB9bo9O1zInA-|-<+
z(%q>}feb-7#mbS2NfQp=7J2S<Up9j-xDK&d77*G0%PeQ$`Ch)0)OT^+wu=^_E-n6K
z?g>_L7W9yI*57YP${sDYUJPxHE;PEKFU<e3{m_4EzAQ@pE4Ru0AY<3ne{Ciw7bq$a
z91^lQl){g*nVJSM9@eDNG$gJF5bEBzD?tT{M<=MmOkQVq#L<TtA(->Eo(`=?s{o=K
zjnpM<MpAdQZ4TZu0P3>0J>HsUIS{XS8*(HL^}O*ir;P3Vfa};k`$pzw2GfH+2=K+4
zY>Q!1Gh#0;72N{YJ@(Zb&|183;eFsD-y#M}o*o^TN;whnoY}r5=>d}8!lz5pbU`>1
zFO`8AeAag~+Jwp!%xY7J7;;$nyLQ=$0A(vIs)CBv*IE(ytYlf3#6uOLYl6F9cq!b-
zcYIsP(nH@+?KD`^PR-w3P@9}go3+V#;X+%jv$5S~22#7HHcDq){C5l~ZyJz;TLkg6
zW<?rVLLm#t`w{MbVZS@N*=9ySO;3G<>oO}{wLe9;`(mfYZ^Njz(;`F#HD1okIzp!d
zWhdiitxHwHlUT;Eyl2UI1q5dJpjxc(>u3uv$;EZIq~{{JS^>EA7$xlT_NCDAY2Is}
zx1z|7O>9!@80}}}ey#9xvBEDnG_5<E;F(6k%T6k{sEFY;51(HbTcph9-wfktelvN`
zK>k8letY|HAu&EW1Ae{_<p=b2`@-Qew_m>1jJSoXH>B%<tS87k!h^3dRB^*GG0{7I
zc6D+Ta)#HzF`-P%J_z^s0%L4P4TMVNblr0ZuZ&%YTVKxoy<g)eEu_?c6HfkJA=e7}
zQBo4u9YHZD{eXHf{>RW+#{7DRdbrKnEGf$Tc5&fxvs7E1Z4js@7QME;EAqk4@mq*8
z#o0~X-3u56{erf*v>>pC1{~1tc{nj<Ql*p+)3yjV_nj7`B8vHLp{A?d$@%2CS!Cj6
z+A>}&fK`)3pwmFEUHUf@8V=u)5jc-sq(Tyl`5c--qA>~3-T`n-{lMVb2y)x~JDHOr
zPmY5j2hg4hM>AG6hF>41jl^vInm+~=uM1^}MTebvgyt<7(TS}s<*`tnw>)PWX>$z!
zXbO0et{`BgUv9yn;xugU`0h4`9*csE=asOvQC?C*c>t_;82+*8@6(27T&+DiQLYaU
z5Y^WEnERj(_)O+qf9a&d{BB6RzuNmw1oVXa;rPM8&)cPCy3+f4T)I|Z+W*|;zQyP6
z=)PscgfXgUw6F3ugucJmlQY@d6X$2JcLM2e+3^ltRK%GBW*4jk1C=;!T_ag>gqd#D
z_-ZAbRJ|0pL8fh{MHMxYk;P$xX`UD`7DZ2joVK-V-LmJHizor7@tR&7uR(JjE0GIj
z_!*Rq&>yPxU}Z}_w5rpYe0Bpz-7Wg5O8gYWHfpc!q39M{)J`jZWC>ggXhqQD_gIj~
zf$ByIHHC@V3EL%3gW-)Tzm_V#O&9v;WxeV2S~~Nx5_|x^aD9LYj9`ogH-GnD_$-2n
z;`o+Y=f!>Htr4VhoPFp^n>9Ubnf|7v#Xzg*_$k-2TXh@s-9IZg|AZQi5$^q&{SpzM
zxfD&%?6^Dhd6H(UGrx-YZ>q(0Y=oif)WR>`5U|7T{n0~f!QrqHm3z}7QP{$nFvi~<
z5f<l3lk2;Fh7P=v2HH9=)%(8qoWNIXPFM$6>-n`80wG)(l&Ie?S98TL5=0QD7=Cjx
z;k}f>lZ2@8uL9XDCUwr4IDG^;5G-g*VV_aJXNN7&4|>AJfXs<nb@r1Ap!gUDz+#as
zB78BeIa0@huvc|g0JCq*iu4aoyXppz9S!u4&AvJSRFPM3a;auLBOc816D;XX#Kc8I
zg!hMj2&2sG?8}u=Us4T1M>5^WlPcwO;f;UMw=tPT$)YTc06!HXOFyU+M>ibQc3OE~
z=Xf{V7HdOf65PQ$BAt2b;8kwmQ&VPCqd4R8n*qH`Tj_I_=h=^$bsg!mgL)_w$dY7#
zSaUfd=)Q(J+a3P&y@YBTNgMf{`rjzxY~9Z+UmJO3^hEnz-BFD86*3D}f3|>4Pi1#%
zi7CUV7$9aLQ!B9(m5UcQ;i!Wl2b?N!R!~Jre5f&4)KzkXm;nzw(AqD)^E5@rt<wqd
z;69>^iQ}4rSuHD|jJ_z{W5lZ4yUNa$AL>a9Q8CnHs@&jTkO|HfwhPkw_mV)Uk5+O>
zT7zmPFVZpz9Q3~>6SXG`=#8jds8Vq_@`a=v5cSNRxq=@}aOBIjHV96+TDUm1C}we@
zMF%BhJIyDzWNcO=5_eX$%b$mhXfF~2xu6L?l9@4}2qkD2Vkq*q5(_P_Y~rsR9KU#B
zz}MS<vN<>$haV{?vXNUs_>Ut1#V*bn(x&}a!+S>jrptdP*-u&O{&*0Yo#}lj`d_*H
zqt+YelMFl{zV1;~DkOg+rL!BLHxd@K7a@BD<KJyUuVM#lV?8Kl3V6mAiCZWG9M4n1
z4aED;^w+;k0{rnn>o*k95m5H&V*@<<xHGSr<MEz?EET8HVb47q%S8`L12NRCrZiz}
zgX8H`ZJWG+bv8>zRMa)Z_}TWSHS7{QK&Ht(e#qnq!-D;>9K~?TyWi8*Ai|;-(25}^
z0ZN>|DhD4+^4r5Vjh^VRvC6k_EyxLY0AQ9)%etX+29ZdB1|9gPt{_xbQA3Hk{#_65
zuenY4hr7$cquZtjmy>C${}@EqtZnydqHfo<xA*iuic*)SZLiF363r5&A10*<)zxPx
zY-YV^OAo6&m!&zVk-~jMnLY~j&Wnu8SDg<mZFlVM1ccXjNG_#yyG6G!CA$o)&EE~6
zydZ~isoVo!z<WPBC286*=}4e{f~QXv8O#8YY>e)0Yv<kVPdorFd0&%Bvcr$bjefAx
zMiH#NSH8?=vP#HWpEFn~wO~D8n9vdRj|><o-(UQaiWE7)M|<+MhTzv|Z*xr`=sOmT
zj_4kvL;L61`ahy!_<fVlgKMxIakXn2zL&m6W`#QsZ-tg;QTJdn462}p%xv-hc*`G=
zeKK$#0e*YVF2aA2NwYE1C$>uHzH5AlZ<B6D_-;B)PAs}(#JchKBxiDP3qAos|33um
z^78VF!B~?f4>f`LFhuqL<&eIcs{N&Q*3IoVj;Jj3+9^ohXuC~*_~s+XFFL^rY24h?
zUyaUlCS@z9s$J{iW2GS!%I~P%b0VN+8VqkJ;u2e#hQ?sQptmWpACoJMdXViDH2hQH
zuibxbGef@G{c&dV$WVI$B}3tJ6*L2kQ0Pj;f;Wn445T}CZ;oi`pWzC0JeD^Rw6p^s
z*{!tl1SBR0U|((&&nr-I)oqXqZ%Y_w%JFgI)o_Dtcs;$^lKj4tG32Zli*AGGQs!Cb
z$A#FJN{>!iFLx-!xU@;11Pr`uO!AaXWw2`ng_S*D4p5YP&Dk!088H7h4)B%lSB}9(
zU;b}NV#lM0n3hw(PxSA6(NL7H<~5ps8LRnJmQ^W^kyr-1QcqI`Tc*X!=~Fot>3rS2
z>3ME(@hj7&1WnRMoFFBP$FZ?e{i=M-{3%QG*T~cI8dmR%eK+ebAlYiuqegzW5viN9
zmW#R0eb%EJUy2JX*`2cm>G{ndXwXTM^dfUF*LPOE)>G|E42E!od0C}jV8PF_xAyjX
z`C!i~S<%IcmiNxhXRl`O&zVgGH`ac!Aq91JDYO;A#kBNlWU!jYJhNla__|I9SU-Gw
z!KhxQ8l1X?4?Z8$M*<38R@K`N3e*o9x(#dNG$Cwvx<rKzJccPmiZT&97np2N|9Ziu
z^KdsZ;n&gX<}MO*{jG4quzrQy9lSF*S6F9tW<1v4Zpx4vZazJkbt%+<WyXnjPPx{X
z>z2by`KRVjy3y=<(X1zM%x&hzDZ@Q7oG;2>^0?}e_=gJ{g3cGtu9d5QRo!RAQYI)C
zj{ttlkc?ro)Z1ASO7W$If#K?kS`*j(RgH8Jtq;alRt5Oac|d3$80+0luV2>ZmKU1=
z>B&f$cf2z@QhlBV=O_O)Vx;apls0|SdS)5d{r8+Y#vBiq{tv9^Sf}SQqN2V5PVMR(
zz8)uUUzgaz0jyg9cuZxixnxfbbJ4ycBi-QMeT<X^3JP)Ru_h_pr~WYJf4pD}-(|$G
zwH@ZV=q8YYHm*(Z&QiOZFaFItE1j*G5q11FHTy$wY)UY25C<l5o`1-dr)e--QAk!%
z#X}Whg#YK4UU`4wmlyCKV1n88N5N{E$%f{rLXk%sGf>aIU-P5!5%*`A6}?d!K1PR2
z*F*+72-DsVwqJ7u`#*w#zqMM`nslfT6Bxe$v|)H%0G<>J@dg7>>DYOKau;40&bsW&
zACwflT>}{(Uqze2dkDv)*>ve+=Nx0Cu5EXV?l8K@NJ(ukCp8v)f+JJLJe7`b%%sbC
z4%bd~EHD^=UsKpuHX`{qmT>9ElSf3(yY9va6=z0Rt>z+UPyY074WguPO)q`qr{Ztg
zaiv!Lig6>!=GSXZ=lW9V_GrM6RT}I^)#PBU?)N9T&+RLK>-t9Rv=@miKl%~K18_}M
zJ)os#q#wVc*t34s**inXZYbfr=kw815Xt3hO?=4RZwJxutzAGwF<2{Uj@PYhicr1`
zwPhdWiYs3^sf48bBO;~&kX{?|D93ujHNK}<KQ9QJT3>)f+R3-P1*BFvsO=3$y<LBu
z-%773(mTQ*=NU!7`UquKa@S(+kHUV&FYtKv5|_jWAZb$2JFJ6riP__+@D;bzFh~@b
zmvcUNlwVk=P^gjRTr~Q-0S9l}wtD@%CmGk(^nb^9B7se3WBLAI*>Y#1z{?^O?Buh_
zZWc!?UW}PUd=x&gl$~l)YCOfeM2RK@8x1XStpC>5|0>5qsx=i&u1Ca98@nDX`QDaR
z6I3qTa+Tp>Zy%^B;9mT$IA>fK^iKQO1At8^zVaMYX`9%RL!Yvgqu2UA!<32%>lY7s
z)E4uQ%7B`p8gIW!_an%ku_>4cHv?ZY#AeAu;_zb~2>WJf?f7(~=e=~2QPxA$JPWh2
ztEbCN$im-^hRxI)X(o!r8;HS-W!96(GPdP}x!25Mzxaq?3}3xpFN$J4U8}|*ToXx6
zcRfEZ<e5>*Xw8S`OX9kZ67<kp#0(#%6+^nT9z}LS$dLyPi$55+eoQV;T>5v7RHU`i
z_q-5CT3@;y()>QXl<hEu_Ersl)48B8oq2}!TM*fI_Bd}NcW{}B!t5k^s`Lgsp>vuQ
z$~+Pv;@s-?o~=C$NZC#hUEh*MCV<Uk0eKG+;vL<U<+a#CC=dwHKg!Ef#&|jJ5N2pM
zrwTjmL16C|JgYM38J~M;y)W@NH`dUs#(^B(bKYJDB*y&844tb({t()3rzA%(hJ_?M
zgGb=*yw2jdz`YlT9N6tu(&oaN-~V#-b{SiTJ7ZSqH@zu6AiCuek-WOZMZVxN@w0&`
z!_Mp&vhF(pD09Rj#TehBXtpzB)u)>#xk~;OE5tHw&VCiPi0fXsmCl=xdv(d9=2Oed
z!^JrRvXF(ZkDnoS`#-l-ZI0&t^HMTP6Wi!*`tbR`hxX+r3N=X4_V*1X@2@4EBG~v`
zox|ooep<4bpMk$Rt6Xzg-N;hpe4Eq9x&i#UemI0xR}pfv5U>~k7!ga?$(d(S7Kw#F
zrsE$CNNk~ABoGzeU$p<GRWuNt_`A4eN3TJBVptgV)nenWG0^x<JR2iuc6|kp_4(M(
ziZ1mD2-mgHr`qS(F}#4l$d*|ajEH%{c39}&_hh0CbX>nQ<>@GKdrkFzL}yuTQm129
zP>>1h^-JDe_g7153<?r05JSc!U;S$Qef@Kum>7rW$qRGiAtLh-ix<wJdqqUr@VAzU
zr?f|po=gFqS6-$=5D{ct@OQP(D#<F~FP~oyj+(vDYr7nn(zp?H`G9wOCezeVtNf-!
zGC61eYVvP^``ZtP_y+ge_LM^Ip@?lnS1Ly2xM;%X=X^4m`fVmsmsg_>@`-;P-$g!b
zx!i*=ZW<&U<fsi-ns;Z6v1Wb8<MbA-U0flkV*REm7!#E}N)u?KF8#O%=>1IINj6Gl
z-pvxprR4(6IGNpnvHvPp%3tsc7E%lK_`*DLHL<*C!$%bL{=!RL^!_8V8qd5Nu{HOY
zDdRvyNiTv@Y&>)QtC&4r_iimPtY;|i2Y16>XEecgYw1Y<+bqTLtbTK28lc>Gk4AVz
z3hII0O!x4<oA}#h`fP4fbRUnJg0R8$1}#>3gSx-6Fq>dFyAFewYTaBf5wT%wslU%)
z`#&7x=4L>7CGgdr82U=|PAlW7@!uxoUKA(MM)Uv9HAi?@mg@l9NnDJNJd4H>(U9L|
z({jJ`!<B}ev$MM7v-p--Lip@USD{M`&r{kTlS?beL{Z`4uj|H64KJQuh#SBgQ7qr7
zZ<Hi|zy-dfiYu)LQ?+zOVG@USw)MP4!Pniw;(PW|gL3c5(7k!bgtB9_<RJv>k;}yM
z=suAKyCTu=3Y>i6_Y5XT?Oh1Tsl_tepASDg!b{i>Nel#r>j<|av$(C1{Mr%{W;5>&
zlv=#Z=Z>;(qB(AF>Z;wl^bM?C{vZkIm}Ruz@sZY@0VFYll>nq~*TK6lK<ta}i!$e)
z^;vByX<mGjRIHg6rGzztH40#4hj4h3iNyOQ|8DW7&fJ78PO8F~*4?9q`6_4_SK(g4
zE6>6OgKgOFi6<lDl_nT(s|&%6b%Jll$KG5E5Q^Dz+@lE0VrSi&<f_yBF7z6&Du$l|
zBp@kObQX+;Z_9auWcJfW!|-sKv<zSz51MV~A9%YO<L%>sneckNEC`{PH5rl;x&v5E
z0jYPBdzyd=x)}>_G(+}C+lQ`D`r*h%#8*gf%AL=oc_lAwN=v0FIz-Ilmhe%q#sbk!
zZneb|z^sTZr1(v@RHJpRX@+DgmA7vd_lyjwX)iX&W)UM8Q=G8e_j~q?0<U<M$BlL7
zB+v$(#m}r+uUdYJd4zM_6hSyRx0>SrYUPsBEp>w>@<)NG*r3qt<Zrg~a>AwIG+~W;
zr+JoxgM&FpVhD2e#s7C~2=$Ss-sR~&C;cYf-6gD-$hQ0U_xG1=d!v$FQ5<nKqNt{{
zMcX0v!KD?KtB@uVms;WD#z6T{c|xT{()D0AeKw4Rw+=j2+$&`{)Vu9*yN|xjY3t##
z<G5hh2a(VZ4UF4no~g9v2zT$vZ>5(;^8`I<$NQL}%z&l%6p9cD_6sy|3!c4b(fac|
zobd*}YM7SKUJ(v98;6kMn!5WH8!4j37bIvq=g#86?XO1$zIk60#X6*247~Bb%NE-%
zR-y1*CQ0c^DQdU5ehRfcf6W*q5-3ub$8{*`A|LtLBxYJ__aa<*`Z8CCt!lKmFp8E@
zxV3wPK$9gZshRl&d76U8QX8^yLqzdSX@RX-;KJ;~;x*Nw2$iV2j#}LFk7->}2iIP?
zAHO??*b<LdK7g*K4;mbtnu`7IEOR8hj<Ti*9_L4?EUJT;nhv9S-ZFuruUHI+yi(np
zv5e{6x1tne@~Ut6;0OCu`1%n7iRO#dBR4j%L_lRBRa5{D6N{&dp=>%eK+#FmTzHR2
z#Hp7Uu)WsXb+vE{>*2?&a!WHD1dctSVB~6lSwzRBvjr5ULqg=+F94yqSc9|-0U4O~
zhuHl4t$bmlFT#hOE2yhcQ%=t%36%8Jw>?u(S9wxIK&TJ%%alA+1yIAnnZCO@_}kc`
z=JL0=c!6SpTdzO2*6Pgd*Ah_)C<WtFGNK7Wb54ns=nvDy&!=0XdjzbWR-Zn9ZaLg_
zFrFU#&+KWEo^LGO9eNI?*kq=7DOjHGPM<6u`RR#cuq^S=-3dKbe+b1i)=G4FD#mRe
z=ar4`{36{p2mAv#ZOa&2$j`(pRkHA*p-On6%d~_yD3J9*|K_7T6PC|vaH<n?toeng
zSKzW7NPNw@T{2x}KP$~~?W>mgw{Io0>-It82_KP@66?lN%IXXkI@jM?cCvmDk)2{s
zb@r2T>V4^q{PG9Bqo7Jw<C3Pjul4L`S_{%?dPkjza5wZn)HoWiylpP*qkG2lsGY~L
zI<T;XjooOd@Me0p2=w)`Cm(tNLqA;4`0c7@B$#$7Lkzbm=#4q}yy_*}b~+}$J?S0T
zK9<bU69aB)LLmVQY%Yl(3`SNd%c+sfBP3iyGhQW>RUo`e!5<NX&lOulTcr=C;_K>Z
zhgk~MJU}j}C913TQQdyMd+QiMJ4s^Y!W002osim~(wNAD3$ucHw`sPHdPsD1#0KNJ
zvy0Yc0nmPxrTab3#z(b+J^S$eIl(Cd%-m`xf$*?+_uh5Jcn#8ng!!f26*3aN^ccvh
z0w$JghJ1Pg?3Z~LX%p%G>0>1EL`9Dq_x4giA{F@Jxhg+ITU;o&Qoj7$Q<l$s?^{b@
zqyq9l>U-MXqfE^?fOUgkX|(=YJ!IAwFoRD(ixW-!%tiHC;+?OxkaRJ!>|W~UsWMa2
z5~)Cfirf_)TiHRvb&|(FzN>OPi?XtESkHUALt;*fmGhpYfBq%A05(zQEuiw`{|s`k
zz-?4guL8anUh#<RFe-{#c!0BEA$OS0E7X=!xZwELiCSI7)?C+{k7rwgPxm_2tX*ye
z(PS?r*Yp<)l5_GOGJQ3QJT6{V{1E=5g{u*}Y5x}bo7y?ouJNK!ssx9OP>hOuWdM^-
zGrcWG_KSPm?2n9GT0S5pXPu!9M|mKQKK%T}rPqLD{-QY~Vihm4M@G~mn`Zb$eWfOk
z@CWzQ=li7`D>Uk&G;YYlzeOCh(mF{5)a;kFGkvKo)Y?W?#NeP*BY5$eiAcHaS`MhY
z)tTyY*3Am#GWH1C+YOyz>m5v4*S>ihTB-mY+VO=2u}OaRPa$=zOxmRWVddH+8DcQY
zXSMeCux_0=y3Ax0e)<8qadLfySbU~YdH4mQrMOsIqO%bv-pVw{<z97|ag(}HXvUG8
z)^_s#?sHpYrZ;0|M}iJwMlM;k$(_6{TTAidkHQt@pkH$kK{}v<G5t-wGS4YThDQ0S
z4YHz#oj<JSsX!U1t0J!*UrC*y>76lIkuj>ryF0SxCNhQsSOfSC7wdG0(C1Q@&Ryoc
zOpDjzDFJ_T;J6n5BG`pm=fNzV9Ys3O!m_YLm>5}9TEw@wMYf#!D`fN4==T?@=)x{o
zfO3qNM)x}(f38QnLN6vB6XVs=h%AVbLt3mu8-jDIlv?g24c{$t9e9YKxy!YYG@{xM
zae-8>0?@u;HIXAz7p2b(sLlVG^KkhD`~GjtRAvOU)KO7UC+qYN`iG4uIyUw_BgXcG
z=IjjBbL*CoR#umUy@4=z*y2g7nA(&<JQhH+q~$}y%_E0;sOTpa&Y7&HuPVW5XVJQx
zUX{`Ha9g;h(Jq2(!W~$ITq++CjdzzLg8dXeGkh34tZ^F^i+kLg@8Sb;j6ry`D+j$+
z=ykx$1ZOGAAQDLFwlpUlqQyShM#f2dWTx}n!>7nAv%(Mb!y1<02LW8T84kq9F;nDP
zloW&5wQ8cX_cWL#00yxjd*AZo3mfl4wlSAcUcyGW-_>R*b6-DHaF7p1u%Zt^S$q@k
zJ8UCjtiDDqOEL)#XYkco6k+7EFGI9;r$dW5b?ZAaBmA_d1T@3N6n1QwJ46yqigd(1
z2cwQ8RgzkchD%uAWcV+_EZj02=zdY<w$jA1y-rmV*q?sxW`1dBNX?v0sVr9fzD{K*
z^W2$JyOY?HR!#OB6}NoQ6EaWZ3KNH}4|zXhzOz(}BzTWb%^stlY*75yuh&lQxdl64
z#C|afPJUT*cnMfzmW4KLLCWpRC5#<OaG#q5H%Kz}L=S@RFi}SKCcb^9u=KH4U#Sv?
zy0=9UtzM7*n3~1?m_|CHJThtXG6vzH3m7Q}byT+eVFKnut;uzjr~Or9iODP^vyZcA
z<2sn_P<4d}L(|$$8J!~a#2mR0dna`?DDwll-9Wt3G$cqO_Vh$d>2=|}7JcGq8@boN
zm!~g`vgbkEJ~HH7#=qZd5Z>+ENKdj5wcw%fk2OKCmd^!X@0lr`{U;^VAMG0i6B^-*
z*#g}|Q2OpesD1M)X#P{BY?ZU{%7W#(YsZeI%U?CT`g6@#&O*mPEVoe#rxgIewni8!
z;=b0sRsJefLqOfZeYJVMp|jg%eUXpkXQgu`ty>+|Dfh(aL=`|CC(50cPEM;S;^Q@p
zs9oX88k;JhrQ`om_0@k-cHQ1XNJ)cqgM>&)cXyZM0MamYN`pv9cX#K|FffwRjl|H6
zbculUJNI**b3X6Af4TmEy|1;`{?>XKCFNHLq^5JJE4|^5{M-8e`_p%lg4c8NB?*er
z8$+9~dORL1@Hz8!g()qwTs+L_BdRxUsWQdr<Pm3C5J2?kRV5zhq45VWuH=<JAXA+S
zsW=LHqur5bn~|Tzr-CXKS12+cVs)1fz1tNBDnn~5Ng`ad)4FT+j{=5vhckyxO!?Oe
zD2@|-%qbap{Q>Wq@|p#|tf8P*f{z(UEMa(D$7M=fGRg=F9+ZQ*B&Yfm#ECkqpdMBU
zdrSq=)jZ>T0^-9})KfMu(zdGC>M&cKU|9)}mpIqVBFM2}tyDBfVBjaur|B}R9@YhA
zo}>{jOo~EElNJ=}%_#Q}Bo1wl<Q+P56$+|^Nl{(IX;>MHhgU1iyneG5`}f;ZV%&)2
zRIbeu<u+`eI?>zkDWnT7YM#c!SKSvl*k!h{C!mSMo5_E`z4BOh+g}o~YBK4d95<7Q
zh<}d6|1IBvz80>}7iRwq)U&!dc>?GD<7znjex>GfR9UVyRg&E4Q@hUAn?lf{(Y2uT
z<6Cm>7dp0U^LHnzylZ*{PG!2Q_0FcWUskaAjilS5BV`Nvgy2PTt0tRV)_S!zjZF#s
z-C{FQ8kbf7eIl*Te;|kmudumXf7D7~B}^|=Jm9$edOLrD*n{VX^l=ko=T~k3;uEBI
ze^Hu5>d(9wb&RMI<_<EuhP2}*T9~VN_XX7Lspz)d5eGpQVNEvZyA=ESWx;9+!7!Q7
z(BQxwEuEsVCyX1MDe2`@B<!?AEY6fme)dNCw8I0`V}B?9QDN)MTyQ6aU^jsHT15SD
zPCt(cWWPU}VpgYOKzLje<>E1a@{xEqnNnU^wOqK)?h^&wJ3}MJ!MqJSY2iIR{y>`R
zF9aHD$GbBOU#=DWg%CDTSTrnyd9fZCdL&UIRf#(-gQN$`bMn19kshp<ER;~=6M)XS
zE7Y@&Uk+^`ywr#+3q_*pM;GSQGTyKC1Z_*>21%xWLA*s+5QaW1=4q#{$r~jj2TSDy
zxJE`~Scu63KPLugJ;cThT$kdH3(Ur`Dvzr@2){yQgC!xJ@>IQZJCk0`0Q^q8AxV<a
z$#D<f!D_QkYD#ro444s29d2Im#eA8eVZV`$Rfr4ob%%H%RBz>!X?rGmwDKsqI!2`M
zdMr;Moz@#_YtwsrdM<B~4dukwP>KDPnEY?!3jZ9N;`7Hj`>m|(=AQf8@H4mB&%e8y
z^zJIgj&B}4<S^I|RCG2`-I~3T3kvq|)NA>Pc;oegFxgh}>ouB&tal%t_8%1B>)i;A
z81g3#f1DZ%Kc~1I*N?~)#mr)OIy7J3c;;Gea&jG#38gG%pJ&>zGH~9=8B+#f5aG)_
z*0qn$w~}Xk2?-X&Sf8yy2(I8iyRq&Ltji#Nt^#)l+m@%wz0vk;4|aob_}8u8PHv3q
zrLEGWj|;sNKz~WSS@B`#qjd9jFEzfJ;y882#q)Ae9YAYoP1EJH?|3|lzx{Ji<9j`?
zH#XNx=_MqzdZ()0pFfDf;jqYHljB?TUeK;^oU&6_4#x$H{Jtov&-M4&H@_qp39ife
znAU92caf!oBSaIQnZv_RId42)FfTCCboAb_W|ximZM+oDRjYMBU)Sx;Fn#REa+YcQ
zfkIt)Ahc{YC7i%IUr8`z%z5~C4h3DyX8+ZZ6u0ncm(kH$PHD6r9bGnL$0Y?JzUcB}
z98wRmL|N-6`G}0#ct`+Z$2OvgHo42NfFDblnE<<l9~MF?lIQJP2_AbsH%g6iY#I^_
z61zH<O$hiwhpht;Va5o{=OW9VmwR*L`+=uK0ly%2r>R?<@N6cN5B>f(nJ0Vf(-Y^8
z3fEJGMFvX!QajIsF_z%0IZSXUg}sX^%AyUX^3gLy#DII@x#}q8(~odFgJ9KcK09sH
zqH2M}J9*r?AJWe%zCM%;Z?vgP<436-Ud;Zt$O!dcJHOmGyDeFzmj>cHp49u_xwT&}
z7{RkLd0dd0k!*Wsv^$5|gHTF?0cR9PWts`1(&#u%bv?3*@)KVC8-8uJ>2kGPZ#vu6
zLf$-i?RxAdL{~RSkDu`#RdC5@xL$4cRIl%mvG#{0gUzF*!U4BrVTZ22-tus1Xqx3Y
z{w7|&Epa;d)_d7j=w*dPPP7PPjuAI+6msJsdd)+_M?!#mmbA1Fig_d8y|LGwm@Pk<
z+{QOw-YW3-$o`+h5^ylqIL<RBii9$@yu3ztl|E{rE|R(I+YR>Jcwn$~c1tufPC~~;
zCW+w2mMB}(45@l(9Xq^>FK;~nw(p(dbHWm{Y=DR-DbA^TSVIT#dgJy<BeOqwPS8@L
z*ihfVQz0&tOY4gy#jXz3M=zQ1hrAHQ)*n;r1oa2+<6%A)Y}z<+zF!RtuvPZ3@d+G#
z@e>ExTTpgL*3{PARCN!@6H7|2*AMzf{er&~R;L`}1V8?E(_6tU07B&vE%u{Y^5%?-
z8?<sO4UaiIB3qGi^3wqd!(uxh<{aLCEH6QXRtQk}iB<zVWh8Dzm=IG`5N!&Bm**ZW
zI&r-RLeNN}qsaoorfkP^H@JegWR{LsFQB1hfj2@Ee;y)-Ij`*g2FR@H#aHfkoH9N%
zvV2w=HAGPekPoVKwojA{MHyi!&!L+V15H)GxzsgwqZ+0Z`)%{{|Enyzc$ZaU9Norg
z?7%GF!S(Le)=}^72Vg9?wYj+@FZLTbsHhU+=TaMX-B8gfgY=blXQ1R*KZ-m~f!v9Y
zu2eu(X+g03wb^_45L8~j91GM4L#dOgi%i`xuORNV37?N~UmDDw-?FQb6LTW$V+O}A
zt6lqD2ORxU(z+u;eb0PDMP*|Zg80n|Xh2HONhhrL<xsibA!s~Skl?EVi}-Mwau)gS
zuNukUkFqayW0KySWRfM?uAtmi3!J^d^k*8_%~Wc1{$FPU-al5PKE7R;d2Q4xxt=#m
zWR81}C>GXAbG>rXVYq`><U%rF|NYV%1B@eLP|fe3cHPM{#X_rtofTcKk?x;dyLLt=
z6PeoP6*Ln#^!^ypY7fTiKvF(Jcnv_3G-1~_oh^56gNm<twI0^1q4Ijvz;0=(m5=+Y
zS$O<RLpn`u8pg>;sLA7TL-U$wImfj27j|vi>ABe{j(@3KQU`Sx-x}x2l$~DVPL~Sk
zptUivCrKPvEd?}E=4w>#5t~Nym{zPis;eZDc;=UFlDp4X*Rg^tS4$W%LprTZf3jCu
zn_*_TQfkQKktheqWqq`~^=?>J!J>3In$bZne_t;WYw_kbpT&P+;^~5A@>MpCL=kP^
zC>IhvH`-^X$Y(xS#g!&hE@I7kHj*7$_Bq#58CtWU)+tI4AM#KHPk2ypbgXS|X;CyW
z$rIc6{L8$hbo2YSuc(fS>$cdD#y2WBnDU=>T?QWyJl1G6)(kd2*zH00csOmdar#~F
z3>_4hNO*jjesw!qhRX1|Csn_0L=2|GX8mIgd_#r^nNNX`A45VTI;;dHzU1g2x9($g
zddcld_WJt$aIWCc)!tnd6s3uIdTdms^T@RfqGUV!CHC?eb$=UCe=fcK-VpDvHHrJA
z1bQU$_K9ThK%Wf*Nl=V}MOa4?&-xO%aqkT3>dNI;G~BVkj|T-0cVQce59_YbRlY->
zc5s;rQQPvJsEp344$Bh}cIpUq4hOe5R6p+ie0Hx^{JYfI@&!L0MRaiT)&qSmpgR9?
zyY2ps=G4T<q^VVi5l6AjSQ0n43@?3_aA3OEp=y+NSN%PUPovkD_oR*Tj5evtoTtrm
ztadWFEyOSEyh!E4xevd_QfTG!K6~W~{S~`R1YH|O9{qabdG@^wKs<zO>+d>??f3qM
z>)LWgzfzR3U_Ox8N~WjDeAVMdsHanQl~j)wJgI>EJwo3Gz>yUblflS%^t1RZv$Lyt
z()``Js;W>{VaBs`@|IOEf#KCOOH10f%xyxXo436(V&$&+8LQ9Bs`|M;g|>eY&`=$g
zVAokc5Gr&1xFZ9mH@{5LJCS!d?QGxwy2TtqJW(MIEnTP01xBczkg~;L4Apqm)Ov%^
z_t#uH2CYAu?NcwdciL4nH@ARtA)3W<b`uad0P?P&m_j1psc^^l`X2TVY5ji-R$T@U
zejnW)h~8_;f2xtIkrI8p`0=~@<)M3IM%Pao?8NDJMXh3-;rlbn(BVdI`OTxcONPq^
zb`5PFnHN(jpTS|YmpD&|0s;bZ#d6zPFKSfd6m(ejRWV%_#3naRHrMx~VSHQz9#gEd
zv$&~+!(4h3|L<^S;JU*8g0lUA1k+gGJoedM>Cw;@pla1l+&K}6rO=j6;)%o97b(12
zm<hBV<|ToC-ByG+p5la?EG+S*kYLS5HbvYaDMzDC9b8?RkiU<a`Y`!qsQ1(^Jt7Hn
z{v*Lc@m()X^nt)Yr(2nQaxB?uY7N`~!_^Nfa$0@%b-F3p6euZpnMQBiO*;JG>;LXc
zrlc~_i5uW;8Bo^G@@_QUGGlJ!P!$L2o()d>#0(vb0iup6%a7F;NMal#Qs%o~e9Q4%
zFt|r+scxJ<gLyJw0fXQ|3D)xX#r+(+VEgUSd0G5%F@Dz^0I#+kj31DG>gt2&1Ni(2
znO6}4i<IMbrZ2NifhwOQ3waR4wGN4BtX1?_05bUmzD3*e<KmVB2`Oep-(<v$&>RX4
zdPcm4^DbxZeaC6{P8gdfFxS^7*GuoL$Hd<Pz@;F2JdEjZBOFfDo*z~jG>+?9rboY0
zGHL^@RRF@RpMrRZNX_?oasfHAH{SUBAshS<c2v)9lw_IogSbI5J8ChY*IBK${OXri
zuX(`BGc$$3PP@a2M@d&g{=5G^EAO&6jf0-9F({dp;>gIzcv|l#oCCG#_19#$e6~MB
z{G?E0JJHKq-^3BHY2_o+GG{qCTVA8&k(>ElseV@lIpFDG(H9PPnSv&^r1~6xJa1!D
zn}dX@+<DsoNROB#pjD<`TU3<}#hu!l!40Q6JROW3_t`vl@aWB3j&?L7=GebWsO$_v
zlHo<1A}zn<raWc;Ab@;T(64ZH{i=}=##jE{oskro=g)`TkPGWcGbubn?ts_(bCSe%
zxUPKy_NJ6%?^;2-XJ6!7o>!sNkM$yACoDd8J!AMPoM0#8^b^A#QmgK5B63l?$oK7>
z?#sJ>F5`#47oR+s7rYYi+Rr~<IOiaOiXk+P2bKN7IBqS}I;$~Y2&<JY8QdL7X^@7R
zQDpHsX1#}Aj``*mz>zgen=Ps*5#yj9X!MkruFB<W^T_u|ZC;{#tFARm$TnFl8^6p*
zHluF=L-xuY@*2tgymBy|X@Hu@<{L!ItyRl_CUqHt2%vg&Ov^Eb0ueEch<6`gv~?^a
z|IJCOoV}9t)$wd#c!$gmSl9f7W1|ix%V`VgnBbwBZ29?`u(Tmc$f!Mzc(GmGea*I0
zVUV*w5Cvw6c)>X6<h8tGy>3_4*OJHZtENAXJ1(^&o+riZOv|4u`3sPBV3AAgk5(n6
zRGbQ2{SuBt%|0<P(dct!{mNsAI_SFxqo8x=di_7|o@1KIjK0c%g!t<J!X=$}yw>9+
zKJdIDCSst6<`HS)(UJXKFHT@VELD-3Ad*a$Zxxa9c@)eQNp<yjTz!e$<w)u6#ynZn
z{tm@h(~8TLvvXdHwT{Tg#hGoVX;<i`+D_9p-7h`u_bBq3@1%yGL(D~sidOx1DrV0r
zZkmF5$;O1=wQ~Xb(D^-bLj45700<I7H1gjE-)MIBS}<%vZXBB|(EDXVfI?drENcZc
z5*E-(63cWKQH0%-E5py?d<(Xrk@?++?v4HzJniih!$uQ@?#GJ>^UQB0D;-EK8Xrw(
zH#&qM>z&usje}y%tjyT5J&I>Ef)9;bA3j|=kqVSB$#i&}x!r9(zPMgd3pLe9352J`
zR3wbtbRyJSKO4>VQKkl*j_mU*YHk)xq#$6q?WO7+-D;|4=Wy7tt>XB26P>5}QTtm4
zv3+lq(~DrQ(UXhGTDdK`x5_<rGO~xw_#@!4QHqw^6h}hrZz4a%0{kmAN&3VVj0U{+
z#VEPeVNfB#cI;Wm8KK0|F++w&-!lv$CHrNoqfHBkL$crO+Hl`rb`k+rRrqjZ=6Bkl
zcKW2<^F7A2V4t`6ASI?6H-c964StSKIGOvk*Sg`!>UP(U(qT_Z?P(o0gl<$4a_n^m
zl5Z7E==#ktUK(BcGmoXyfqDDCowFJ&sV%X;rR=Kb=I8BpsR0qv*xsvfcfxi1^OK*q
zpKqt@zsm@y=&f?vaD><&+DtJ~Eslxxi{}G*@g!8L&otLDou;q65auWUem){ChqE;J
z?>`%DPc*M3yZdcrY@vvbIfaH=|L8Q_8SPqeTqu=L=~e2q!)E*`K<AZ=NwE_SEmQbX
z%+ZpcSsW}K^6HGjOjgHB8%d$*{PkbdxnDwl+CcED3)huDxHCF*L33k8&#CKwS8`Ln
zZyssM?o^%D{xs;)Re6@;5(q!&6M7~gHa$@lK)#FEkff({iAF>TMxbO3K?x=pOvEO-
z=t>jHG}}L^WlYF6GPGJ7cyYJoi&XG*>AW0q9^<`fXztQTi_Lm}s}NDc;wLEy+bv$%
z&F(g3VEW!EXPq<04D)v?^MBwz-up=!<O1Y22u<`H^qa1y4c+|{n9bYj(CBvJqtrh4
zv%+Oe`M2eg#8Y7761^VPifON-@$B!R(d@GG2-cUY<otKYCvHCxHucpZ(yp~z(|xbV
zUVSVS(CaU)3>p-&uS!!hn+@(2g*ocf$gZZecy$8$i(GX9d~2a0J!I-F`I7q@X8f)f
z5CJ&=RsB1RWBq6WGgRz1g=fabkh3Z;Mb5c6?lTckO+7WJPI<dZVEe#BDcwBm6kc|t
z8kk9%EJy{$1(gZDg4F5HZx-S5y0qX^{9oze#ATIQg{6Z%vj_&01T?@P>ZO*xbFv5F
z_w+q27y>v@m=bUSmxZNnTCJ)_V7oFbENqzsRTHD<_#)|F>ZiJI|G;wmX>-ZNt;Wnn
z)!p~=b-Oe2)T*Ku&Pv4+RmwNNGRSu!W2|pGob8Ea$t3G3Th^De-}WFxNWXoT<B;o!
zze!TGB>i0y37p1;Be+jNY1pods6<%n)l$2ykq{RTNwO{A^iAuS{Br;MY)!iuao90n
zap1=7q>F?jI3_#5X_0<LN}4uEXbAD!XVV{IdPebofzI!IX5xy+1xo?gnmrP$SODpg
zL_=FUz}KiPuI?OSXslOkEPv$ZedlXeI0#2+7|bp#e>O6aNY4MLnCsrQu)!WS`+6_?
z_F1Me{NrfshM;;{4!49Hs#hLDWzI@u@rUL1>!>DT0kaaWgThJ0G&#c;$2)@ir}`7?
z_M#UH^GLlxSkW7oo2ulS&=t=qaVBjh^%J+f8H#>@vcW?Q$UfPXJ*mLKJ=SB}1W8Pj
zHTGK8;<cGcpN%&QtFn;Yb7C%A?_=p&Z)(GfUh`o4TCjsevFPz?OX@CrsTX}Pg}nZv
z2cpLCed*2GAO3cWLt1;Pu3d0Di2!#$_*CTZdg#lpAqBUQN?a%Q&qp;0fsN+Y;wS&p
zGz!!Jmdq$eb3i`i6p)TYkY&1g#A>p?-EGWz$HfHF;fY;p?UVgk*%{H4aJ3NKJQMTu
z)KqL&e^Tk7MdLlEDL+ym8EOgl$=$aPQF`Zi4^f6Ibw_o6|8Amv$>9^A`MinwQuV(~
zj%Q@FkiuLt^zrwY)#Kl%AXtq_r=e5FA!B7_Wy9^n&ES*7D~D}4YJNlgCM<bX8z^$J
z`W*QxJ6<-Z%8S2C;D>c?t}iWJQE}Lm5-w3m8-H)Oud_irzDqHcw=9Ct$@7=u4R8Tt
z0XC6EQKVl|#kp?#So>EgGkF#4gWk_li{U$9cvjEBn<@mu`^|B!7QPNLUkQ{|Ewby_
ziprm|E}1wv>N1A2=xsRc#o=zZ?GOwLF4;|r_+#QO7(&O`dBTb{kQv6yKZ-7$L_6H8
zk;^lT1pPSt#D+66)fmV|{CK(cyEno)iCWLWjT|7X-E%!+v*2;ODpX^hE6OB?IjLx$
zgRuCM*?~urJ_2W|KqM9jo1L0&nN#(8{wAq5v7YYRmlq`&x*LX-y}NyNY6u9N(QO$3
zk0j0-|7i>*bBNY5lJ9T%;X}JD_%H!yk(@%Lj8S8FXR|)3B9foBg<C{R9f?k{p#A*J
zi#r`RqOp(X6643iep+Hb#OVbFj#ZNI3}N*n;Wttd<d~OjqI;hr+?mG`f9ca*HXybs
z>exBCc4~Zp8amrqF#xdf^9N`npzP@Dp3SA^U0y~2>#9mQI6Cc0{Q-rn7qC5x)>D2Q
zxglEXiM@VvFixQ@UyaOS+pcM?BrRRMySr=aJrD0bjN$f_Kh%6>Dw%ZB5jQ`F``{9S
ze@RYE+yfqWsMD_|g*mR16#@b}*Ifc%9yZczjoQlcL;(mbqkNI$g5AeudLiT&5+F@k
zCr<#v00$)7-bC4@Z&NBc&5)SMih5OtKP9?yd5_M|)mEYHBh`uXG_xm<-3dFeCMBS4
zP9VY<gUJ2~$|&^dmp1}!7|Fo#+hNm+f~k`gYkp^9QsJS_JWa0{3MMn$o-a+6%5o~8
zo#)8)ep!cHlX2%eqSG?cdQtcACeF~`sIe&oLPERulke~D+D&yOyqUdR*gzEtEq?sN
zQ>C&*>cSt>_|Aw%PU~pSzLo18^DVb~=7?Hn5iVXas@{Ycy^r0&Oo)(n(ho{JzM>-1
zHlNxm8NfeSjzJKFQrY2a%sH2hXY5TLpRLJa?8E^R+?0sw+|O-*DU0q=O$2%tIG#~U
ze}bc#choF$e@-JLkRchnk_J;Jv~HM<AV!8=#Cd81ZHNBGAdLeQSDhmR7zHF+s8$MQ
zCJ%|gRsgO<H^0A(m0&u}kMKmzbM)6Ia*5aW`R5LHDKzEz$UlBlp0Edum=M<#tFXer
zAH@jxnKge;qm!j~q6mA7?}Jlm2^b1yl#?LO-(tA=dxjZ^ffZG`LQb(!vzAB&c^wP0
zQ5jETx1(KwcUq8gm8z=hYicey6<5(seTVwmekeEa{t)BnJ2uBZ9Pc$s!jrn$kub;q
ztUCTOHGwT;pL4&B6|B*%80cy1%I<m43zUbKw>0lGdmLPTqUsHWZhb^mZ&J9t8ewOQ
zDs#7pB?v>jBz%m=z^#WvtH+Q`lwX(i{5Q+>A=#IZvC1k9OiigHCj_d<)-3_$D(VPk
zV{wlnR*I?TlLk~>;r_YwGP*{VmO0R+|BqM^&YBBSbX&ribR+&OrOh}b<bDtk!M?&r
z{%f4V+E)v1SV_~6nyLu8+sW#i4xxa~Zvp52N}vtbm^<mg+k;JC_m>$rMYZs$fLyDo
zFNAHpBSogBb9*kuW0hV`dy8Wx3kuD0{t<h(F6c$LyIAB&OvL+o)uAf;X(S9>Fi6(B
zQfueJ)uo0P!-}d^Ch04R=AWJe#@jCAq%_tK#gm_iQ2o#fEMc8_-aYKcg!_z>-!U%@
zb?>Zi#h^|T2MYV9!JjID4_E8v@dH#G56Be!ddi%F%QJj$g@X_{tJVT@?Zo-tb|hDg
zC6NUvg0O5+9Qs?Tz;hL$lVH7$?y?Rh)ZoKa$w|s|mX3+<4Nk6^hV39GNrz%y8es;K
zPy1;buU%cSTIZIIXP(c@&ioK^$cXcOqd;0*UU}6c#C;<k<|bop3ZR$7-v>=z>@~s8
zbFnL(er?Uo&BG;%$-An~zR7Y1o&U4@o(VeR)TRLm$BkJR)TaM;1~Syd?kw*f_r>d4
z5}K)!bzNN>9=C9U5`@nn(m`j*%*frhR({3C1=lM&ywHKUG(l#f2u4GJz;dDQb3WG2
zC(v&s%Y%Lte+r)8aK_O+6DE{9>8*+)Ev~e!Iaxa=v>%p1S!cc%vp1isx-<YzT(UI@
z^T3tERNa;9r#b39O&XG|d?skt-AjStqGCd-*ew>ryW^6lY=-<}8t0>vlJQyo;|N^C
zjD9MDNFBr%@vr3xTkhVS=J3J<D)1sQTr<@?17@tW*5o|4W9+1qLvyB=SYTKbOl_}1
zcax>dN0yS+Q`_lh`(K&LUrucVk$T-evzLnu3DQ?BeVVFE?OV@ib~47;WfOg{%>2HC
z)6o?Sour+=qFw+Uk?pOEPPXv+_dn}jPUd=bor<8;l)v*kE6+Qk@C8(EB_~M|?&scH
z2U`{u@n{x+X{UJ!Aaq2~9WK<ba!W0^gDQB~Q0oVgT$>Qxiob&B)_Hqo-Xe(hL40#X
z<lHcDcKBTKjh%g!YwXWYEv9|?kug8f5Mz(U^vG(hClHR@?w9NuuKJGqXSgeiRmZ`)
z@Rz&Xf+Lw;xrrY@GME%*s>N>@_nMdzHhJ);<v{20xe&^+pLa)Vy-MEG)Rzf5xm#b)
z<cMW+BvYT5no3Wkmmh_<i>E0hg7*i-qA|iM|HiwH>FjB818$an!(BuTM*H)`rv94b
z;ttErij=egi3{G@pT8qlBK{7~e)v5BJ~QMe^>#Py=nbi`D__|+e_70Qo5PKr1t(QB
z0TJz`@U)|m&{$&!R0VZuvzKln5xGNOGao4_qYov=A!y8tzgn;*k(3`~VLFfS5O8z3
z<H}2%J#ARZeptpI$?}ZztL4Y@e7h}AStpP{|3SY&7LiOyMSzjpY^Q0|WMLT?&cz8N
z1yHplMjtC#sV)hfI{1Q&VpluPu-h-cIfJxcQ4v2KiZ&jc9q(?pDz!tLB8)*~dk-zb
zR<bj*jqM*^Sz$K{eYtr8s=2Bz5LOy#N|t#EoVH^}R?)5V3qrLJB^Y95T?P@p;xGK@
zeEF;rwY(-sEoj~y%8u+NupOC*0%^^vdRxG^J>HZIBf=8T0=>;}ho{m&NWT>!UVLb8
zvSpO|N(x(o+=2aUl_Yq6T=!iEgJSdB*8%>3Mo6$GfEoJh4@4Tl-3~DXiHGGE24)H0
z#%h}Y4M~TFYZqeMNft&~&Osbpm~)ol`%QJWnE{|=ZKX)0sb?IjNN4`~#4@gIWdj|3
zNBeW9P3?owoeZpiUJ*|BMuklY4S_qct@>Q)Lh)%#bs%>)PD!n;*p*a1H7CnJI=B|q
zZJcXpI21>W84C;RKAsppGU)nx2V47}JHj(zZodp$deATEzfdUbQ)#^aF1_1U81ZQR
zr^&Z-3k!@bEiH`>OCQ4AJEKU9ubt2tsm>`}(fA2Cu>lEhTq{3y6gE}7xt?xbH^y`z
zG1f?hGN2VkboMS_$H@HQdAA5WGDc#&Fpv;f)o{cD6WAp&Uzq&<LTSoyGxA)G0gP>!
z&J6{r-Ja(7SnqV=ZvsjRr4ccY7^9A0L?C2sxhb!+{ECWMUF{P9a-B_7HKGjsCDuSG
z?YbjoAbcs%?l*)H=Ab!^VrlgG!RP+EsN480%*jt94U3XReC;V__5RU$_ssDnfw%Pd
zdIr`VfK#Ubu6yNK?-mKHSiqVnH#v2Y#}3EyE&zk7nF4CLRx-t*2tWE`L`U;h@-7a%
zgC3!TX|SW)A+-ahWXAp;7Eq#LWX)@67iNRf5fq`iLslsFORoAdmsb=*+U1*R<bI@4
z>m<=SgfosF`dA!+eKcGwv}4<*W5GA2gS03a$ZAN59V3&5)xZZ{*(6>&*CWAdc=YTp
zKtCf$<!B^w(dJNjSy6Y()x!y)Z*F>r1Q0xDuygfolX1T2^?;}HsqprpOaK#yS`c-M
zUkmae6?cB^1c&crEzieqTUibJ+?#ptbCpwZXnn_ct)`@dm3i0Bzf@GnL;PYru&Fq3
zR=xAZWVum=w1^1_73?aNb>LVe<$y^<`oHxZKwkSf=31s(uy-fxHRkP4xo7yK<WYFP
z;J5VM$Ngx@a;v6yf=ESRd2J3qwwi<ZhbnIZ`@0<5{%rtY^$(M$Z8QJCasmRe&cVAX
zmY?TXJvgW;%tDr}^Uy4p{(S&{5Bqi)OPdLf?wJmW4YO=fPu;96iZXxH7B`eoL8DXo
zxCJ4kLbkMJ_k%2s^bA{)tHvs|jrPagkMWnYrH$s*geyDmsRVTwPS$9=ouA}CSVAo-
z-!EU4yyv}cYz~SuHu?)$h5YuzSSl(wEs$s(uKdi#V<YoUOj%W;ha{F)0F~z<GP1*U
z?4_ZPHz;E_$5OHqNI-7`xkf)I3^$f>Fol%Mbc-{BpvL~AB1`RK;HDI>)5d)&nBkKN
zYW!i}*c4<E?%0703spy*xD{b9&k>Rso*Nq;E|@C~!W{=|86=FJI8Yj|f-?c6NVNgo
zN~Uto7!(}QEW^dq+ZqC-?-nsLPlO^Y;SC2p=7^j-%phmi>$@WRO5&ympF|4i*_{Yo
zbN}aWE!J*NuWs&XvnE@t_Ekax;x7%i3T8Xvz48IDzx(AfkA~So(_y?A<U5F3ABj7T
zn_tYhf`vt?J2#1*3wyfing#VtmfQ7?2e-avbKAr{|3Jn)Yo1k`C3EX4{ISLI@8^Oy
zo5^>5C&(5`rg%_b+r2{2!^*}kJZfL{DdoMr&#V2wCBaS|Irwo`<lG~C>dFqr?mz#l
z^aOFrm9LHb;TSZ&CKAD=Mni=+yOj()1y=~alGp4_#*p9vN3q6EMg_M#qc?mg@+K@f
zxApYw(n1u#b)M5VXYMhdfkXE{xF^#L(aSjSvXCvcT!9UQ=Y}kx?CRaT?|5#Y3HEl+
zHD{=-95Gm*aVS2Eh8Ty6j=1i=P|SNLc|(kcW+jAq(#yn3!uyOzcmvIwkjwT64SOP;
zMhgd%M1D=+Xu|FaRR_qBYejCH%Pq82Meg>pDaB8jwOXgfe$@6cBzI7!xQ}wgH_JT)
z$nDqx2C{X`9j{q3A)zVk;2kmIOV_gDN$Y&j?&LP3-_KNFIjgg!$h0VHN9Rl=VWlxD
zo_u0Zh;nG`ivFkwfRD3aKOSYMuX|-opg8%VfHXhsf%})#JTog6ZE%u7`{cFF5QU1#
zDkaZQ7&KworA+%v$dA<=2K3D#-YGA6wg~_nvtH#AUn_#RE`eIAFTVjf-p`m7JppH5
zC(!=_jkYP-bysS&wi&y*yT9p~3F)Y>pQf@;8_yFp3dw=}&#q#BSw~9?_}|VCrX?U|
zJ(<J8zv17s?h#H`bscT`C>03b-_fpsn{ZLZ)xSd&3cW0_4)X0qRZdlG%mgGD$xQd>
zQyvQ59{!H}aEXw=>+7oEVoMRah^8{Pbdmzf1QZ)gcE%hJMgC0z4-7p;&XKcdA??Vz
zM^S~f8m_BV_eeDAwyRJ%#mlZ!t)B}tAbFKEsJF<%kCBBvYx)oO`_;Qe#DPq(xJ<gS
z``ctA)4u_?!?u~n>!0<Hbdfz$qBJwrpw{nlli!thhc;Gv&eneW@oN7-Pv#_3p}ylV
zxT$7p?mwIOA+W@lYtz*;o7B9LEAz|H+xp|a$hCO6geGH4K6G|Xj@XV!fnF3$_v1SA
zYRd`1sWWDjyy7EUf~yaBioympm+X4kPfJ}>xOI6A{ijG~s#v1twTfISCeC2@wNUq4
zQzKt8fFJkz9Kn|>d+nA^zSDic&Nh}r<ha{JocIDzTtNR@ojtNZU>|)^<}0vVzh|n>
z`Q+#{u(fGo;ESNTVm+2TOQMw@mOX`lBbKb+`?+b>MvHOrNm)IFiWx@IL!IA69x;MS
zCd^JrjREt^)1PLFWRGymmQ>~CBi<WBrujVtV*X>Yr3F8VJX@Uo7xwY|`%PhCFWr<~
z&%~#6JR2?|*urh!OEcOnjG#)I{LI$OfIx4ZSDLE`MaK<Z8C=(|)@IsR-yD)*(v@a#
zIvOdO;gR`6V1xM(J0y#M+~UWe?^20GGZ)gNj%sluJy3y}P}|ifyLj?^7hwQc%KGW)
zcDuObcB~wvGwn4Ike>!xcvMMi)f+3?chV+a8Cc47o0t_(AlC~lk4Qd^)k;70)I%{z
zRt$OC2uLp%dfNWH{D`x^O&2HN8`o!OwyGX@X|z_tT9u_G@i^TTFJy5KnA8t*66(Lv
z%lQOToNz)f%d1+ErF7F~it|2W%Q5;<u3!dpO0dE~vnA4#uSQb#XRjv>9?A$&!Px&P
zO;Ot**H79LM$d_?bz{iALX~PZt-#e|y}%y_jlGQm^!o<$Rtr4{ZDMUwn26in8WOKr
zZit$hZ+0{ya*)DPLr~F5wJ!jE@BP(1c=**VH=pC0MUiC&XceAXC%w4(N1E`rYXzQ<
za)dSN_L?aJV?}+Y79H;u7c`0G3#^Q9^eaq9+IU(44BHwi!L@5&g(s=>Y_wxnum4{C
zoLjBo<#inqz34&G0s`5z%Ty8x)J0~H+JER%4gXiB@Xv=*r}JU$JPpTx6R9%)q_&>x
z2*3TGmpVt#o$}F=Ny;t_A~UYu-tmC}K7*k3&1ph#_U+c}i5IvlyMceiddl5=aNLTs
z5~~)#D?1fsSCL}2P3OmNbqn>YC5+Ev_gl^)OoGSN0Uu6YMt>U9N1Hs>!u7HKo-IJQ
z%x$E_k+;#|+_&BH>71~crKEa=exS2oyVE3KE5`5n=|$nmQkDRof)5H6eWAox5mtq}
z-}nAwSBA>!;(@};ke1pF%hh}9fXbTul;<<pV6P|J#|G29$D=sm>lO{0T%YjaJkQBX
zIUNE>e^zY26;$R_YSpZGUjfe6Rr8}3lq&Ys=fx^?^!qy1G_Ra$yOgWSQmDDr+i!Tq
ze$sv&zuzLvi7iQ%;X*lJdE=ycK=g<F08A~@*;e)r9gqdDwRz<q>1Zeth)-RjRtSgX
zYX=(K2b$Y#*<=+FCY#M29OShzws#?SNkZsI2$`@OzOY?{385sJweMwysQj5ho4}+D
zb@Rr#I&xukU{K$AEW>9~yxjCq%J`A!)OIZ`&DxiXVC~N*iw%v|6J)W~dEV<jAAEjW
zLXY(nGCQ$-uD=(-eS#+LZ(z_k!2r#!<uy27`=5nviU;gGs;b`20)c0Weii2V9<khQ
z6L&MihuSQEyAGZ=W-IfG|H3W(p8d=X0qvyJ>`5PeaNKRm)xxwvSBa4=DXCTNl@&0-
z@#)-h>8d*o)Fix{weOT3*4?Q?H1CiY4Qd@77!yx3POmM~h)Na`@Vy43y$)dWxR11J
zJga)R72c787sC5bNb&;vu!IJ;b8%|zHgc_cB3<oR*5`N>Og)8<vx^IRWk`boyP@Z<
zXDmX$QI?Jh{2m*I{&;#U7}$(7<FOn_d(Zb!iLo~(&YpCr>Wj1STrQrNOpzZiINp6k
zhoi{GO>BfN<$vCW#Y(P|cmuzVQZf%x@t;ivPxVux?H-HkBi7auBT1ix3Pn(Gw_r=-
z-J$r-?9v+Qu@h=wx}eJpsV<!UjmfBGWhfA{^aE9Cep|Di9=1H7ibAX63$mTLLJdBr
z-868l%vb}R-csUYPI$W$6X(B9ad_{ofkTt2HqOMQdnAUaQ0XTh8m%(HRP&|7{IG^K
zS@iUWmCX#5XK&{on{(nndS%K!>2S<M0yP22!Pu+YFA+`u>Vbt0`bZJ|{jRb%-Xozv
z?v>$ot;T={uYV-Zdtt;Tl9W70MycORg6UXa^jVyi_vwl?nT)oQ8*5tJf`F~!CbZEe
z&&kFPWP3$ST{D!8OQ+}H6?#w;qJefh?=`JSU8V_=#dN}Ch@0UUp^unf#^c`_az80U
z<oZgM1%VHoN8UIhE3bAUs}L7EMw+9ST+)#=DJSTP%e>uZ*;MDVX5}cY$Ymn=_pwvt
z=_Q<&n=6m7X9ZYP$^QB?N~{^6yF_$2@#thE9AO#*56FaYdr$HKy;7G{o{>wgYGQ>(
zO^Dm`hET`8hDp-F%jur4*38UjbB#l@?ORT$-G*dZ>V`9ykgDN<=9V)f6^W4}QH&)-
zG1hG<-Wvf#x-1C+00mG#Gh*a{yodoKf@Xr}%X_Y!Un>Fds^t3U=*=!tbWnV1y!Cio
zsCL-zIJ6%pT4fUpjhrJzH>2QL&WoTxCvir*<yOd+h{<AMZPQgWGJ^i&R)I^*cb9DP
z=OOQ~1<5u(Q?CCHwO*0Y@#_C_0M>XMmwE2sT<N=XpqIfBsUK){iRC|RSe)=#g5ydn
z9MA;gHeRCb6akt+RP6@f3H#(y`k}GD7IF6poXD8)1jl2f-?IpzyJw@0H=T2yMjQiJ
zdzGEOoq1%UE?+H~Wav!U_Hd|}h<{OXoP&XM;-l~Ir9~S{Na(O|Dp(cLr*e$j>$YrJ
zR(F1~%>^G5>8g(4MpKE%A-rjsI`c0oc)f|R(`V>@)bS%=KB(gHXle{PA8&NC#}U2M
zuC*1<=URZo0#qu`&s(OW%VL{5!TWJMC&n}AN%t14nEVy<dgo7ogR)Ml!2A4EEsftr
zUIL|l^-+*_GulF}j3K<6wC-qgbd6+Di|<v{ZT}<yL@EW#v71@EPGWeG1In}!dkhD#
z(QQ?F;WqJvwJA*40B@+u0N?4668Ec9GGULPAL&k7``NiGInQ2aYBN(YV*u&HzNi)U
z7k#c71xKClu(PQXf5WM4xG?kUS=CcSDW3wp@Q{bz8n9J;;lwE^JRHUGU>ZJki=*>T
z(n2C!;=Nur`Kj)Q(LZ+?<@P9b&M$Gs|J|;L!nF4WMY($Wxy=&488K6PujGb@XRr1`
zj+l05gE}<3{w%iTB(u$b(t*a5(U8Y;vSXweX*Ve^{Mx-AK9I!6jTu7s+1aa=6mJJ_
zaJIAiCHU{9#E(sJJjRKEja@RGv}bhQQlh);n!b7H3Y>Y^y(NCE!(8K5jU&}~C+P)7
z7MMDSAt%0`h{|g*&E6aMv_Jg%O$cHOn9Ri$q1RSqx?cT@Wby9p>AvD8yRfeQ8(=JT
zu<cw1(q}e^%=$|&D9rHN!Oe<ux*frgPz0G1|EgVBH5?7KRl!9~6_{-m+qO?dh;F`F
z;q&Z#1?l>LzNvzlf`X~7rYLG>wNc%&6PfXWU!sdF$3fOSGH-Ay_GyVnPq>^au4{VI
zc3(P#x?z>n2C`pd9(e22cwXze1}Z+|jwA5iUi86W`<8<$ArAOd2C1S10Z3Rks`6QP
zP9A3GT~@BXKLvTMA6AUnFi4o)l)Bruq03Ds(6D%sM}=d<*@>@^^Kh+XcIkyIceSes
z)!UKQa7nFt3n(JQeNs?fNlB@su~946%^Cj2YZ(cJZT;KhZbH0ed5lNef|Eu4HQ{tu
zIDOPA-jw61E}y2U%PVR8@x(myvG?x7=6GTw$iICRx<k8fS=^cmlT4Iqy&M`r--$~+
zX3T#Fo7Wrhq_-ZMntI5RYJFJ0Gcve+4myH8q&+JHTAMy7L5Dp(b={s)X7`c=i7V(6
z-JhXYK|ZRJR%qKKKdxVIu->wFn}%Dp4IlbIDgXXtL_psxX)#e*B4V*8XS{96Wf+>w
zs&Hy6tz5LjvD$-_A;}mW$EEiPp_Z|Wy_`I8Vz;z?^SKj)uHgzkh~}$_9vT<wnFh3E
z#POlqGDyncm+MuHIdi%x*^fmY2pQ_V3-MxB^`t7z1jLf~9-i5a@+krc5qV^{t1#+M
zT#=L22uTH0=2|{BsE%KZfENrYj4!Xiy2~Px*`;6WxT#mD-xOl=U9hpwZe{7#i+XMs
zH4Zr*x!kwd6{Fer9e&3Jsls#+@MKH#sJM&))OaflHdL2dWQLnsB|f7SU&9H|P)(E9
zu!+~c7Hg|BMw0pQdWptnFd|_0ts>BM4(MjWw6^Q%$ZuK|?UqbPd2AE>52UG+=BxEU
zwd0lA<)iTN&H>)@^b-I14UymQ|9OhXe$Ke~uWd(2XP(@02hX$7GdiYh)nWOuN4hoi
z-Se+;Y5^@GZ;w%<2$iJctnA$m#qwde7h)Q#F|0t9(#g5Y09;p>c${^#cmIuP6g^o@
zO9OU#4x7YApG&obDcNMNF6%SC4svhEyM;;l^I2d^QXp2mNjpo9-qS>n<cSmih#8B6
zUx0O1TJ$T#&1QZ6gOx<+_j<>83aFWhH-zX7*IX2vo6Wva$uPw7^~%*j*@Ev<`=y}I
z)$PJ)v3lK1d9W*FCqd{t8SC4OphVnMoeJK-E(Q73Ln{ij<u>4#N7<Vpwb)2||J2E%
zP=YMNuSFy1lqxG^Zwm}0`*Y=_5%bb}i-8sxM>VIc_3oY?zr|dKmUf8f8bj$k7f)Pm
zi&5=+WVh>qMIP#omC*y(-{!4%xRv>p3MRWLxaKl<{*p)^^eE$Bn%g%vEc})ePXA;K
zPd0u2lG-s<<Kva*PS8<rNV@t>V7DSRy;)AGg}0<*TFv9Euf`Z%Zt0c*a@JuvH0*=2
zF4_tGk++Ub%55F<cwHAXKlxJvV$ftgk|XG@<$NKO^Dl&6kN^5rV_4$MIt@8aU1;&+
z_lBhWHFY9>{^7Uu)}!!WqZr)ak>GgedY-x>)U^9+^^;q`g2dfLPV^q!m~i{mfraHO
z4I|nj2^>~jI{T0c58AOvU0vqTOT#|l5NGpyH6eIkkoZBu33{>OOA#YjwefVwp7}xK
zA7yZKuj<=eaG<_%E~OT*L;K7@NHyH^qjKy>G7G^_^$J_<a+c*?7YmTnY4mU6w-sWC
zHHLlvy+9w2tFAe2y6(q4juz;<J}cryA0En6!wc~P{>AyEW4!t19S#Eygjm`^4i_$c
z8cOVjLz@Gx5?!MvWML}r9^0~~RSKAhtVJ83;%E)u_Hiz*+`c@z25~2HdQi6w*3~tn
z_W0*)8siqYG)70@LNiE1)-YLSf2w5P*ysm&!MxNHYTJIKXY6<|K#x4#zJGl@{<gB=
zX9!o(p+s{DdsB{gM}gZ3Qh4+y(G|fnTH}x!5MA4ka-q!(28mCq4ai;rsB?YA;cGH}
z<BGT+-mM17PK^@yHCjqbR_3x7V>wpD8<?1|7t6&a;RIfU0%IM&V>|!<c3RJ`75vj)
zGzk;XCG5Dl5o>CRMrFqLX~)Y#yB*6I?LKoelP%WNz=YFN@pQkYh_R^^p}k+PUKn8<
z9ZEopZwV6~3IMa-0AAaarv}+vaSZ*W{wp5w3f9EngHMZSG;|BYwHz<NH0<%zd@q>}
zm#(NP|3-b2=MyD^%Mic44YdF$AK=>&To^H2Ul?jN^7o@&HYV{hLx<jt2smrmu>I=N
z@k4vDMmUPscY`Ijz3xuYYx~AL{p+0Xw#F*1`fAy0&y@E8I&(5P4F{M@1~V1nvznd*
zQE%LCmRMY1#>XAi+uLM(nurvpGVeTdxbSS&?ovo-GQ6ZtapA1F!n*#x(wGJZo~@NF
zgwq9r%#u;wxWDQUppQtGGpEa)i$7fmv5g%TfC7^$vSBj)3zK%c+^Pc~p-yEGLwgy4
z;>IBx+q7xt1Y{4@*!hds%VgIz({EFsvhM?YV=x$JzwI!Y8#bE6(+k+!O-;#qh?iy}
zxkn4*x}-DwB9>>m5}uFaqTIx6>`&=BplcatnH2xUUU?^*elnx$klYFcS*mI#J1l>f
z%@)k@`IUJTC;g9rGsoknVp4K)Z|1(|;xLjRM_p)v`7_c2>f0yflK-m1^>hNl9K=5O
z9=sDDPJFBhbnCk~bW>cAXm>i$@oWlHYpuwRtrH0he1`4)f|pLuQls6DD)V;B1wpsI
zzZP$rD^<!#qQ-Y}$i!s7j0{2dXIX~RCPm?9Q151iB^2n>x1u$=S;1r%;Tej8V%_oJ
zLj0JFEZ^Q*#RY26X}MpXs_4T})gLe=#&%R-yQU$laqWkWw62$@^Yp{na#rN(%GzY6
zJQKE|@eG>eYQ1*$oLVd})K)4v*pK?NLOcjEr=-cLnYQhrjLJ}J^4YFfq<!sd);^;$
z6(@h8)17q(p<~TsR<+Vbrc8A*{S7PCsoP*>qZ>nyE}s`Td#cZ-U!#rf&;&qCs;SjF
zLCIWJyTog6TiFn*Kv1oVpuu|Tbv=1a^eq$LJM{>7qjAc&oXayU<d+nbgKXE4YOkmL
zPo2HqiCpG`2}F?YL|2!7q)p31Dlj6Hug`nedNG0@lWB3Q&Dwi5UkS0lcQ45AgiTU%
zYG<2SuB<W6WrIMCSN6B^?pPh&AG*;$%xxp}&77uL6KUR>>Im%GpNdktOyw~(JqYgW
z@7^^x%IkfqQ-r>F%C`K>i>m%P{%L10R_LWX&$ntJ<x1$o|1ZIW(U$#FU2&GM1NKmt
zlEiJB-J2$a)WPK48eq_QLLG(sK;W>~i2B|-f`!%H<JjJQd61UVyA2lVynqnnpc5!G
z$3b+}fhF?`Q~OZ5+3p{mkY_ZJ*NqI!YcvQhA)hV0!r4<l1|ynBGYZQWM$$$tI-4EM
zBO?gdYE3UO8n{pC72v;et?7nIy@);G`}l3)vp_-e#F0?nL;QBTrPzB>#IFdwza}&3
zy<hfjf0}dwOE%F_Ke;Gu9YhP}N46}?@sCVT+WT-l*tVN*J7s>bYRnz7q`B)0S8aUE
zkFR9K2CfFhB0~mh{`v>gm|eiNg7f>+nOAbky=BBLUdH^IFsyoH=Fw3+?dFeLT1hJt
z(+33&QkEjI*!N3YOP9K5NoH}PsQ+m7Pub|;F<p9O)G8_=Z7IQOvAX!<ruN-ax-Mmu
z(%+)q;YL?hDuD(vb3)d#sXUq$>L;kqA+PXv8S8QZOr&))e?rvi3L}r4f^^0?h4~iF
zx*V~U@L5%>(1}OE-D>s|2#gOno=y`GK&-$Sz&ke=JZY>2PZYN-G@>_FSJYh7$f-*r
zX2@`>A`x0$_Q8P2%_4l);FZ<Vm|+IpEP^KDK$|@wzyHjtA1gEE`#0tUJ>T8uz!LwJ
zz+-|6zfloB+>?=B3km$}aCnLs*}maxx7=1FLpaq;Kh5<Xy?T-x`+c}S5|I9+l>ah%
zn4faogfen1T{isoibHppkJdRRV;NsBA+X<rtCai%m4em!PtG_l;-%QC^G(_MA94mp
zDmLUEfGT{D)ozlxKkx&-rIp?7STGmwN07&d89TzL$hVU<A@r89%_F=xfoHr(u00CB
zNFp}(C+{=f8@1MCAD`&mim}9t-9X?dQ}N0bJZ|y#_$bqEjyI5lySrOY>>~7Gw?T{M
zC)Zlf$nT(<=|P;HDmrd8u<d=V3aPNjsh|dl&4&A6QM=Y`9;@<lep1|#dHdZ^w>H{P
zaW<>d5uwj+Q~5}gb7wizCK9Grx8u2-=&cmf^!uY~)wcY6&KIwBq=kq+r~^!p@IbA$
zY4!V@9J;+7Ztt~RzuTzdO=qPAeU%D9=t9R6*GH<FS4SsYt&m8F(_J}o4FA2acFRJ@
zY&ahJjB1m2GFpSC1l<L`22qqn&)W$1sV=9J&!KTQ;4CjYt06iltm8Z+Leaf!+D#s#
z84GXgkqJT2@!h97$-#(Fj^!{u_PDT1R(%RHak??VU0+T4E&FyTjo#=etJTG3uOTbA
zs7R`~dOycKPVQr89L0YaZQMCnP~{|bjsxE9|FpeA9)wb@0=9l{6Kd;%=s$OMApX4&
zp!Kdd{@~JV9E5qjpp|gSwvm>ffA8u8ZU-&^AEn#R8%;-lqce><O4|(UsXf;_!HU_*
zX=>U(S~vq|hJOih$%rsf&@)~e15^M>oBarci?Gl#d*W9os)j$R4VeRR1kkqu{ef~e
zOLB!E>w{IL>PhVVR-|xFMhS9wtshxD1$m9Buh?7Jt(&!5pPtsCGr*hX1n!dzGW&nr
z=fZ(cu9<wJA#ypkfY@qqhY|(`24SSEMd6(D!+f#PUXHT}`)s_hYH>-wQTN}y?A>$k
zwc3?Ac*w^aRqx^5Njg+Vb9Nya)sBtljeM({W<|GI{%$4^)nlmtU3pM)rm3%z)BOU@
z93?RZ6UENrWq<zO_^Z3gd7`=qx#rKu^ve{qBq9s#8_QWTjFcM$>v(!=S2@*2e536r
zh&iVH4DyTlvyeL-fyu}u?1o$7o(1src8*CB_cVg&p!VOIz;GdSf`XF5#0WXwA+I3H
zbVKD_<XN>-L88t1vli^diO<O$OGPUOTq&I=G&C+_V0q~i-ehHhhCw^z-b>7UOW4Km
z-0QVcN~88xH&h`S()cvD)i*{1gtcyUve_p61J_zcbSSt2pJNs_l3nH7U2PJU?oxD4
zSQ7SI)0I6Pa~HmW?Ckz=TO4k5cjtI5+j%K{$dRu5q4xh{?E`b!*=%1OXl-9_84))7
z412kxIy+}wp~fi;$WEOj+L#)APF-DCU=bHT6Lo`oe3wkqPgm>4rv4vSZy8lp8@7wC
zMR$kN-Q5iW64D@ow19MXi3^l&k?!v9?w0QEUbKMpnY?>{=j?rc@@Eakc;++beP4+X
z9+xseMn9k7aN;}2XXnSl=DFyHW*2X1D$&Lck|<?pi*-D$I-)}EA8QcL*OQ2YzmaKI
z`XVQRDF3si(4r1bf_uvkueZR2x{@ju>ie2_z7vAH6&WhK%Tyo2Gao#`$W^rc^<yU?
z%J%f2-%~^~?K=cn(&L9M%@=BMQW3q-8iOa@VMU5^g~cb+;ziadpEIc=38^om^@cC^
z^GDe^j*?wrC8mpc&bxnd=#Tg;_RFc@J$NAh?pgUq#=l>}-rcYo#%K6!UpdV+WW+i9
zE;Eq-8)n~qn)i{*BE`J%hB}7r{0dvoUT(i4O&ILKWJ~A}`<EzE-)7!}B1X>92put;
zVQ^5@KOp()I5<Cdpl9y`PlsPthilX@k{78BP2CJ$zOIGq%!qV@R|{Oo{l=)%!!CdH
zyL9aYv;^1kqB`m(3L9<N3o<fktHuD{->$rjb!yW2(~)&uV#!{714a=VIp~RRd$;OG
zXECU1_CCKzbOpn#q=O{NivqvtyiAlGqp>M<CauPMl*#zgvwdib&fMBRdEqAd!3#PU
z5ZN|g2^4?^E^hC6DMDhDZxa2E^+rfAK`Gnvg$fI_dwE4VIyVG^VYEta$q_DYZo}pJ
zEi}Tyx|<iT#^cWwUDrFV2g4b*uhZbxeLJt~|6ZZJu4i61=m49f6&tc@Qjwxq+eb$(
z`*X#&4fcryxBFF39rMF$FRN59YcA~~iSDf|<XlpxEj!=!_5Owv<3aU0&OdlsEBs&?
zBc%#+tU*tTUMg8DE@75MZIWH(9(Bc;fzHVlL>o}dnb~*#xEOF*FIT{`TW9L}Y0{_4
zj2et6(UVqdAB1$KFjNb)L%4WLiF?&2_J14ml*X&lN(lef<`D72|0zfy5eFX4O)tMq
z-v1XNZK99=)8owy>fzUeN}&r6LRV6OKV31w+E;q)dgKk|9U7s2VHMSv!b`h*f(rP*
z^WhyCu`A&36;s5dK8f6GYf1Bzk{3jfLgy`iEbsQy3RZ6spT(bKB(aQ$U7k73C7B=9
zt7&SU#;)%2R@XEG2IE#0U)TK%yoCTv&{`OK>GTN~awU$?-;e#}rDajU;6$+Nw(I7S
z?#WW6a};%nT?LHcA%AaqT{<sp2m>OlY_kBZb=bFMXBMci(r7(fRef-GB*1MVQ9s-f
zsiax%!)5mV3s&8BQh*tAYjZpGQ2%-SbMFl$k10#Qc%cc4Ba2%ZvyJp!Q9pGSl=rJe
zGzgA_S1NaH3>KoE!t}Q^r@`<>r-xD&QB$b*a*YODkOV*(qHYne1AWaEj82mVdM+k7
z$?890jFxSSqzYE;l6uzk3f+8{Zc7s$>fJxcD~0*<uQ&oWFfNw^RZk%Lk(FA;^nmWz
zReSh#fsoQB^8DY=%ZHM7iTa7rIpCq$bam)m7ooH`cB)yFZD<Cvtj!ZrBKDup(i(;_
z*YA0oj23V#b<`-YQ;Ks<U9%p?t1E^?F~u3Lr3$<7og1-;S(j;MeoYa-6f*q7SO{%0
zM;k?jXcB!nJ_~K%$8Z8sBa3XnWebR$CRI+;R7wgblI<Q@A3CSusH&jOPqd=eEqs#@
zVK*sIq(4bt(zfeBozcYKVTbaW!8-_NOh(2ok$@Uh=ztg#3v599)bx^{x&VJ6i8WBy
zbA749AZ{PNlFI)%eng_lrE9r`T;UhQPvBm_9GB3NhsUKi038?a&#F&N>kF>oue8yd
ze<w{T?HeASZx*A5D!zLc_&j-w#}K#w)x7I1cu^>Czqe|-*H0KustFuSn!X>Zy;wDg
z0><<NY--t|+@V-+K2y2FQrZHh;IPPdvmZ}Dtd1K$lTGuJb$sxTbYRDl^_c+|lGMjc
zyK6g67bf&HqgKplxM&63S?S&|*YbsqHafHk7TUSIA>!lyf%|^PPe9tX_L;o05>*j|
zdBd28Zd99yNx$Z&9OAqx5u$DRWFX4@TA@SKiNUr}oqi1hKV_Qto#3=iC-{RDJk^X_
z#=C5((SwEHuwrZbHQkl{_g^bi(o#;M7Cq}Lz<spH^|E*{Hs#CjsG(=L5?=RJL@gLr
zZuT)9_J_Zh0_j`_z&rnhfF{RRCFU&SzN=|b!n(`<z~cr66&;d;{nS5gL=2EkO$>oQ
zy?)q;WJj~Y**>h=R5jZmC-YjzuK-n#I>IdM<4wMy%vL)Ex1#hNpEYRHL_=4WQ=r47
zzv?z{9=`2-LK;BjT|Z!xC*dd#Wxcs3K&E3SAfiq@VgB2{{BtbRuA2Ww$&GyGclRo8
z^C096`mS*yIUg6ib)i(<p)V^N`G!Rf!NP>kgCPA4N-oTi)}}8k8QnY7NJQ1iH7c`G
zQ8I`WU7T&LtjRt?qlBsbgaxDB%5Cxwfw8jmhp={b`_+rwH%3bN%2`<g>FWDvs0_s!
zIHx=i?F8Y`BQ5u`X#SUT+PlsdZ&kN=O$Cl!Q_rnK01s4ef4tS|pjm$z<KMl>c#`(L
zOKKRQJ|m@y@^MIodR(#_AKY7tn`=#yd9K%yYg#14rmdFA03K5+S+cuR#IJqg<Z?e<
zA#aXgj+#RULe;(n2yMa4A*i}cFt{sT?7`WN8oQN$+$X5pn@JnG@22E433(8dzj;%;
z37KOeVfrkq-})VB0>f}8v;MAFY?kRS><!fuo@6UzR0|v3_)8*~^U$2*ZGCZt11j1|
z%xasRRVb}GgExZ>$+5;kl2de@aL_WbpXYXC4R_?6QM9tI-Da%b9(0V+7plHd-LT?c
zYQnsPu1tNIi_A)K$R+5Y08_MOkdj~jqdHx*+tOk6{)|@jHw1OnXY!99QFM~&K&9)n
z3}bm|>`mm;eyABn>GE6dsm6y^4D0La?pF^G>h*O$h4^!Y?EiX>?2RtgJFsd=<@D<T
z;_E#6WTTipn$@M|8e!4nv7}q=m88D=V=hL0<~i<ND<|aD-l%E!c&hRh$dgppGPkeR
z;L0N=mCn4;oLT*0lJC*RoGZ?;#RFpoU4`0hgkbZ1YCesU>OZ1Z+fA7Xz?+z$Ed3t9
zY`KoHUyN!UJb6)7G~5-_I%#asKJEu;^rn9YND)vj(fXi{G8dm$@q&w(ei;oj;$Vr7
z;`#N3y#L1Ot(X6TwWv`i<*TD{p6o9$r`X7+tr9pE-{e_trfq+Hg;h6gHA0e@En^^^
zrCV+^fFXE{HlO8e?%DQu<j3nYb~|q^+8AOII23o*`q5AWJ8Y$C%GYH4cENLr!pJ<U
zgj*I}z^eb-i}(7;S|#Q8aFNSr&g!Q?MyENP2SH1{U~rxlR1kenFH@o+1;BvO;?9Eh
zmYt2};RaM1CK{nSYX1=a7`H;QMlp}Zb-r_uwpQ<_3po=dA#BKW8F|+v>W|Aw{QRAd
z(Rf}+ucdi#G(%fdE)6q>26U382{gBD)oRRJJPItQ^x9-P20EUjILz595pn9Fb{-s2
z>Cey9Oj|Rgy@Ab)7-V}$bB}TU*Z00m*s8TkMVK!FW1vrO?zDbDbz{%%NZ#&6-Z;I-
zd9KsdWc+Pp=d+vLbWz*;WM;Ra`-qgh%xd0i8V#($5c+fmEj`Eg^WqF1q@Vwf9HwLZ
zV+l%Kq$MBydMO1;V0w|{BddU`&U3D5ix;}viQTm}d`K2jl8GW?B~43H|E$4o_<Wi0
zT1)Llyh*WCGq>t6@G@CbBx6zpHrc;c1qrN3Kmxl=`Zj!-bxw%bX#|^~`!IAfj(9@T
z!PA?Gutci=M1N8u$?@N^9+CK~!1Zc+H{Lg+I%Hu%g2^JqHk4*;e+ECQ9gzK-T)wN^
z=G;yBI-a||o1Yz5?Xm;c<a!KSL};iVKoQA1U8gLxOq=Petq_5vHVQCv4yo8Fxw>Y_
zRR5;)RTA(!9*Z9@i|?HFs>YfErB1av+NFqjR824h+^<IYM@9`F+3#jyt#_HD=0-*_
zH5uM)xz^|S>M<CRpSJkhrhFZ+7h^P69#Km$=0FgLLlYg1q{o|k3tv`c3#oBAq7g&q
zfI*2&EUn~a)2Lc4Bb~!HTui4!=h>3m-;;#UT1{`N!y4_Kx(6lQx8W?*5n+9qJ*{wK
z3!Ar2_=pQXr?!bkEZEa-{uhvIRX*62;aYXP&2-k|k)%wvt>|KDF6@s~@v~9>?mA}a
zKJOogV?q6Ep*eEcpg_kt<``cNnIX(THwzNk`(-g9O;~iU(Su#ht)0~BJ3oV>$^xOW
zPvd3cz%x)$DVE-hFfD)=lUC5Cn;yGve-OPl_mmm+T{W}jXr>4`dg3@Z`v0(zuD(vN
zo0Q90>=0R`rF)%(2w9{>4-&;eH;V4J{D9~;zzMJjJ*ba|K3-+K{3d@okiA?miIVrW
zZL;{{o$Uq*DJ*>Wr2EUN5Uc5Zw#;mK!HFqbt^4?s{_T(-HO+>Z$WR0<i4n0?Wsx<b
z!$d%2Oc^@P1RfP_GR_7ZRlX(2U$doGQK+kp{s&CvKs`PV7IW>HW+#?6CaJd4DQMs6
zLupB!VP78mVD)fWb*)#;%Bt1=rY6CvT}6glerd}+`RY5p!cr1p!eicHVus2;6Pmc3
z8K9VYv;D&131r+suh}YKLcOUNW?vUo8WUPfANcLi0$v8$@P9i-y!0~^q35qCj7@ow
z<{vvff^l?F(yv#{k07R0Z(5Da(?BB?j7^54NDbm)$&3Ll_i1WTAsF$k`03cHcD8Rh
zYGxNfI93df^lJ8rK_B%QKG%$o_(X!`XxSU<kOjhL+;cRFy8$}4nYY&?II6lowjULg
z|B0mT+*k=ys6m{q>))LRx5l{ijGZ_lMUV#GIiaXn<o9tMZzXiR9|}=k-Z|zUqWQUc
zn(DG+V6I6ltGhiyf(QDl)Z8n4QbfhP^|vJzow%l?{e`^$;Lp2BqPMawRvp&~Uw`!0
zCk$2;8=4&r^M6cKl*$u%n8Ns6)1oFMbUTh2^WfgPP0D2?L2kTc5R~^SbMZp&Q<9rP
zS2V0Mp8xPooS)}^eMeqG?_m}t`E~81M_=3zHxB=MbzlJ9f5e~Z(_DV}9BE|ydIcot
z6dlU508sTqnp2A5nvwlu#ypXZD3$9Ig2H*eia^jOxIs1VVby6=e$`L$aHM4@l7C8_
z)pcDG_`ZX1Dl;k}-;tsjddUga23P||Mt$SwrkSedXBO^IXq&-ZaF&1T@|3spP`I~v
zw-_$%#Z|1`7Qf}=&Q}>6-+tFQwJUjr#2MEy4x9I6n)uBd^}cV_Vf>>RB8$pWCWnOF
zn>aQYV!q{c>4As*MCZc!DsF$PgB+>Dq^`lDW$02}`fvZ6KB=msM!p8ph9nZ9zd#&m
z#+T=Y^2&90IS!{{ixqD983AU!-DEq{0TL||_Wfq71h|^DQYFm4O&+#<aJjjRNRbj7
zQE3W#5i<qc5{eY+LX#7a@<M?lZRX8(zplnx$hRK)Lr2CPs``*HRyKnM>p7do5y5Wl
z`7FuI_bc(*HXlo+^dPQZ?{%=|icteneh+ML+C445D8H@W)#hd_=tx_hy&<N4r`}6g
zwKoiPUDrNLmRG%EUHZaDYa0(n`<hOlabU%rADiZW(v5<&?+$}8>M_4i2o-tjjghW%
zt=O>S*b|68o%jT-v|A~HBwfRg%WFU=9|R9ZfNT?ew8IIurpc+LUA$F(D?cQK;f(LU
zr3Gb{*kZeb9AG@2JJ=Cr+ure!0qS~TRV5((0=&XIuV->L0BdPw^$W;gl5F=JQMy#^
zu^~hbZ>c;87J5R{QecLY?LD{H0|h5OnIUrUt>mqKhK(S<`KVf_ar^;Zz){j2GwtJT
zJ4;})?w3FKAhpV=wV9|{FJ~0M7b@@S4t*a7ZyXpN3rW`HUZdf%=zZqQfs(imZX!$}
zpmSb4P(JFP>*!sTb{*5FQQgMd#U3XUK~vePO@=o%y3>Lm!aP_j8N67K)sf(Pw;Q?3
z+(BWY<bi?(Ul5j;9DnleSv+S}+D?a6Z<}k4+`N#`Mm-MbrM39`lQu^*Z$fWGdmpFH
zL7hUU7BQ`+vT=O4our7ZAyVAlr99UY&7L#H)(rF4M2Z4|gsXg6H+DE08ih*Y&0Faa
zvnZ9dE#`EK*V*X8(0dJsS$NJ9^9@BBx=|t;ownn*Ps!RXqmiQ@ZJbXz_r$i4xEp7V
zMot8Cg>6(aQABpU)PoFc$ZEqJFHmpV280R>C#@p>nCWqU5}Lb_A$)6wfot3@XT7i1
z6L|$GVP(tYOS2Qyf^epOKTP#>(*+^}O7vv+{OO^04aa=3P#*E81M!`U(T7~#OKTGV
z87B(45vu-AK-|Ib9F6=^xrx2D(Ymzd+-SK)SA1Lp-TyDMgw%J-T`X#iLf24Y+1jz^
zWVYph-Ubmt^M00lf08aIqkKQMp9iorWJ<Y<Kew(t-X`Q#+d_aCu0!Wlyow}vi1(o#
zrw;j_wL)BRn(k0<x}pn^@*9*X{*NwE<C`Qz0_xQ=%lqnMB2oi?M8}1Z=0E0P(7`je
z^(wfc@%EJmi$lN;I6n~0&ItA+dtwOPT<TC>40dmOtS^A<h*bUU(%s`4#6=Ub&B!3C
zdrKpL5IU{CBH96$O$m%rsS|1rq2l}zYOxlJHFa=SBad)Y&B27=2;phK)GMV3e9GUe
z8>v-$+Wq4^ru1$SO?9HS#d*neUNU-jL+ux+xp5Fw4Vi9)RCkg}uU2olx=9o)HdP-p
zBhc2(=^i2sshjH*Ow=Vr<_mSmPfi8TJZV}38HbV1>v_4O#KEPJY5QZCCS@yGnWHqk
z@!PKbis(H<jq_9+=n3Sj@&&Aip2{HOPPSB=a9#|oV}SF;iLhI??kYEgFGJ>NQV7p0
zT54iWVot85K^~wy|DEMj_I`k5*kIYLLjMQo6Op_F^sroNL*&A6FXsN&%-%N92naLf
z$a*)6K0Rjt7D&ExJZ=2Xzx8+0GWDB*Z1{g~`hYFWr1^Wy6ve6WhK!O7`+-gS##c=$
zhV6xSvaHZxe~(nXlwq4+J*=M%bl|vddZEf>7x?(M{--W^4I8)BZ!7iL`+WCjzk|kl
zzFf&R?%nn7O<;>OOm;yNtiEDZ+h^1Mw<yM5NYQHb^8SujeLvYKTwhuXB?iM&_S2=5
zX3J&f{+ZMkw)1?Fi!R6boDdG{M11CWp8cibO>?!(X2fBwID+vFz8uF1-tBGnlW5Ob
z+&oTi)iZwYA}@x>Q7>zNb*PugVY4fhp2mzZtc{XbmspkfRD!bEGIe=R8BA;9F$*52
zBE&qp?ESscX3dx%VrIzD44!p_c&h$qlcKPAa6jcr94-4|g<FdncN#3ae6Qvq3ku|e
zlis4qybO9{y_Lex@+vP8Se%WohX}&YIen@RSms2^tga1e-gX@ttZE>}+*M7D<!a?*
zs1DQehF*Lr<sO0n-xZ6i-3``RoimZ59?iS&#u-JPN6=^xBFO;KUx4gf^Bo&;dvSYG
z7MLS*_dbGW_D=WTv+~ApQvmD|goFLOd_)!rpz&;&aB#Zp=!dR)I8J*2A?|KG_$6Vo
zErx8-YdlbX9`E`~nP;Mq$wif{;EVN-uvBwxu88Zp>Z5?s)yTa>lgOBoigukq3?<Nm
zf$Z<sV;7PL_`dLy;?ziCz+}L2^eJO5MaXXEKP|EW`Gs(zEZt!LcXr@@Kc1caC}@}D
zBY*qN1eqf)A)#PWGu25#@NdfcqYSaA&C*UYQ$9W$ja^PbMmhf$7=zA)L_ooxLRxN_
z=x6q0ua2ghu$9JJyCpDA;p%CLahOm7jF9>qo;-(H`kPeKB-8cCk}<S5N2yAkbgr`R
zEkK!O$35yI4uy~QL(Ub4>m_%vvTpk1r7yrJ1MgA?SHHvV1b;RiBAB0qpe(MZ-p1=f
zv7{i%wI&58kme8*SL%+fwDL7J(!FYuCc4=^#aVY31B=_niAyA`*yYC?3aWi>`;Sgn
zu*^iRa^H%61R;|uSZ-yFVnB5`1=`&z-0Bdw9WC~P1CgYar7Td-;f&RB*`^<t#yRS|
zp`44JfZU&y*K}RsiMPxa5!4IT(K4H#7ZNrfSh$MjPzm=Fa+hHBK5hA_QEHs>#<l4>
z{FG>G=?{JdY@hL1!`O*Rq439sO})S4(|cHN=V6qazX?Ofgijs|Ea8j|xg#E+oTBE2
z<1t2S?{BQ3&G3?2Kl9?ruCE%EV>68z_Ru}#rU{NDoy2AWk`;HBbahQnxK)Om6K-D0
zH*aySQ5}Z>j{soYoRxL#=|i8(g4x<K_F0)Zj7TS?7x#WO5VUaxM6+Tn0Z`IaM-&hN
z3P4FBrovpmGcwr8pU;DKO~@Y-0Y{0d5D>Gl=6x2I0?50BZRv_@l<ar><bk|<JZ;b%
z42fl<#nx6o8An%ys4DzEhSdW7Ox6gI`|lFxfNH3c=rWknDpcRpd1GRSEk@0?o=Hz#
zEz$RpQ%9jZWhv#W(w)rb<XI;1x{e?NR8~8?Xm3Q7xQDe1)HbnZ<U&d6@O~zKne4<r
zpMQbqNM803XWyEEWZFUk@#BR16a_x{gQe5~`P-6XujZ9ZlXAU8d9F4!Bub8PRSF1*
zpwR!zi8m0WQ0&#<-(oa&$99yM3At7^+{`q>L86iQ@Wi!Z!9!Mgm0O4CXQGDvqyGdE
z2^Ly4=Rkl1n+J5i(+>xKX)kn_fpeasY6(<7<q;|Z(z~&$QKzu|>W)Lk+T<Y$<BUoQ
zGuQ+Hoi{p+o{YPAVG(J;{HIdS!V5a=IELFwyf_Q{A!#_D=d8aR(0He|buzV#!Dd&I
z5K6Ex9{ouTI+d`JhC$uluHkxLvRV&O%07@kmFV@#o72}%g3WOWoE=x*o~kV_FN<Ba
zdnx<K)96{?*#k5bGra)qvZt$#228KUyn70}{>IHMgyRSb#67Vx{;2<EbY(~tR_%2j
zn=aCNcIGG`;R8TszD^xO>Kp^v_=8rIpE(a!`ub^)>n~3kuc+aF>l=M!abY<2)s+5N
z)B8tkeq2nDMn}3GJt%;ZjNhoxS-J>ya3QEIWI*lx<}Fs>%!%ec#`U^`ZBm-bVjrm=
z?V;&Fp{m1_F|xxVI?*ObS=xHoFy6YPaQ^z}`2I8DVB2oXZhf^amzqhW=w@IyW%Hk{
zqv~z+!jSlKf;U)j1DazXWQUf`%%k^2?6jW*d(#tKGE+U_{(?%O`@q|*$c;pzl&pH+
zV%zB(rrmqkg^=!iIL6<W#ZDmhv>X%)-Ksva>@;esBFA}B3|SI`CQqFjMeANLhj-&M
z^j|RJvdP5w*8UjAJQ~b7YNmq!HzW@v)9so-WtU1$5@v_eZeKAB!G5{ccx}c^1(Oz0
z{0$^n8?OEX&iQzYM;$-s)?-$z5{rl@ItBkG!X=Y%ePL;Nu$zYRr~Ra&rJ`dZj>J#l
zUFF15rkXIwd-@jpoLPao3T3PNWTJ#iUwDyg!MH`_K^ig5;aBW#4=OlUdYYL77zaF$
zW1XZ)a0ePX9lmX?m6j#7buH`c0&eR-2=;KRXUxCmdF=mGw_L%*<>zM52)h2UQR?km
z{z%}QgG&BA^;47n58L$YzT2gNw78Xd5+}iTW_I<Vz;U~O=slr6%`4}(`*<}@m~Ha?
zemY}EBjf(}+;3~kMMD?;O=A8#K;MR_2;fz}(5a;1Vm;7fq)sZE?Z1!mbPnAN{4D?7
zCD)RHWt^*LfY-0(j-l#5Rv`~5QIHX?n-auBj9t8FoWnb&d+_8JyGWbCtVnMpIP)H;
z*tuV!{I#yPpG<u&=R^x;XGVoq!QpJQhw@+_P-fzH<^t7|*cv3QM@kaJR;NlL<-&NO
zD5}+0@|@HKVU$as^=UpMkFtMkJP{x-+CxII;|jSqBgx<_&ys0Asd2`4ur!~iF*1u<
zh59TtjY3NL*n`18uY-LM$~BI2`))iE6&{?OBT7UOJ^uWCh`h^Ndus?D!nRz&Jr6dC
zAoBP_YXpeptp(rZH>;vqzjeS574n$T*%7hYMrl!|=JF9!=F490-1cxu*zU9xFl~iP
z{joaUWX*yx*fKN*Ru&=^2Tl7bHM?ok&)l0lH+}k*-Ns|Gp@TR#_1k~R0T{vRi`GmG
z$0iX7s<%0jFQe)ZlSJnB_RKv0Z3g94TDTv*outDsPxs6h+HGz94}7ug9evAh=ktv^
z?83VN@Y>H4%YXB;$x12XAH0!yA$^wlWN!xmw*i@~Kkd~f_9#>4OUhDHNn8tkj&;|k
zPD>;j_tU+VQ^X%{8am?g^XVEWudA8dKX7~t`GLHv`<{;7?KG375_Y0+fvKyM`y?n=
z0)HaUA|i}drb}rk>?j-+<PH+iGc?&}){?8h`$?fAALFy)DiA`Yqfu`Vno^=6{$#zC
zyzRAnzoYZ#OuSdHP)N5AjfnLfjzv$*dqwAElmH@x{fP}U2<kq`^#({p`eh1Sui7{j
z`f1S~y1#*;Ku;B&S`qn)P=m%wUa3PspsOt)wxU_pzf_Nq^+c($+`&c71U)gne+d4(
zJE=ek<`)q%&+nFR9AcTDF1&`a1C4ybNorOTWD1H(%<@Klno-w`9X5UAe~WkpZBljD
z@4%DW7TI*`*(99H&6!Kp-#^M$@DDF#wkiP6ZaH@VAsnrsk;(UHwYk2e5u(thIS=z%
zZ$^`|g#Bh5Sd|k&)Ib#{=j1WX5WH~wrK^>LexkFb;;lFC!ndUJ%%e0oa|xTtyX4yr
z&6{f@aHimX=`ir2i+`U!(jP@o`Am;)lT+T~-T3$TJu}4n0_C(~jMeSPt5chbr&pwv
z`Q6GtM`!oOU}D`i`D;}m0{@tT3E*=i2MM0rMO6p)R&D;*CapFYKQ#J!x?MO`oT*H&
zFM=IFvoya?_WjZEu9sen(tXu60gsTd#F_OC_x$Rg?OIBv59TvGj$-gcJVc;R6LI+9
zEzsM$0cem5z-gdlxan&<LZPcPK!Vv=Qf$&-$P9i<d2(a8gDbW(e0lCiy(Z@CRc5W{
zjVdoC=M$$kw;~{o%L%OY=bcBm8oXcPFw8?AS<oS16O`2I6;Fd!5yNW;VkG8<83e*y
zZmN?cuXEqO0g8Z)>NeC{p1hFh7VZ$``fkU^XD)Ma*_2Bvv1Nt8(U1aRt))AA{I|j?
zip|b}BWSmT&RFhxi%+Wa%HvP|4MGjc`ltvCBP#)0D@(klD9|_HuZ@_VLZx9cN{uuL
z*~9IGZ{D8>g64Qy#_$tAPaZ%JX#Y?`MqMm45Efe&!Bsp{Nc>m5&ywKZA<JC*7pIlI
z@BA8BT~X+G<4N$ZDxcJl5hA9xk%&h0SH+u8WaeBk7kLXw%}(^-w;_c?%TtOlx3v2P
zex@*Tm{3={Bj`G=58E-#(Q{bXb?9=mb)0jTXPS@kQQv8{GOYxgPR`|Pm=23}r9|~F
zA~O%ts>_nD5oS6&P}7As5D_z>nj~fx%989=n=1M}L~PBmyYJGEd%eui^PHsO%4k_j
zmxK@l7qDJeXeqQcEdq~$e?3-vKy*Y<kPV{N|6Kh3o)O@~ZWp?KUG6?KApY#=TIv9+
zZ)nJF_;jST+hX`o%`-PE*r51E9Nr)F9>OIL)6ctntGdkMvh~^<^Q{a%VTt%j<d(#!
zB>&^F?Izk(`7eo>Tmk)$5Xcr`t%j-5I<D#;@=QFiz+GpvR-q5_PP3Zx$@NpqDFp`D
z2t}(z>ccydfj1TS1fiZ3ktLlDS#st?l`sxnyC1(q!6(Fb|1yBkl$XbQ&9n0Aud+!j
zlo!lo<)+z%<Zd;h_*qB?F4@Eh>zN?l$i90~gb{%86zGshzmRQdwns*Rx*gHwu#ru!
z77Exy7(n<}($0u0mbegw7{SW*sq5MjM{opzKs!<WI6kh>8qU{sJ-hi&x;6e%q_Rl4
z!K+LT>Dyycz)uOglI|@wDXAaIltswZ)}~COarAv}DZ+o#!Do?Qbxrw%c0xhop0P=>
z+OkMSKJbe^O*kXQXm|)C^nI79qS=1zjZTT<sT9N>N7}^Tv!M^X5q;NOB!M^j>9N!K
zQ1A9h(}$Y!q}!WM$e`5TJomC+paGWoul|sGOtF8NbNM^&=oF`IXWuHq$ul<M{|aTl
zVklbDD>P0T|E!Qegxw7l2G53<hhbvSjpY!+5y2~!`GIUpjMj@d#PJIu%=%qXh|tZI
z6>?BW&})%xl1pYZYi!Ev)K>6nP@^P{(t=rRO&L9Qm=)YlrRHCCiB(MANqtAmZ}&Un
z@lyG2uc3q5I>iC=tt(MdzE#X*k(aLHrZJ2~?Tir8ZyerkMI<$KHbx;$1c}U$Fq%w)
zH>Xro?!B8_$b-@YA6C312;fOpk0W-sYlS3s-D^i&@8}0Fi=CQ$$NKC_9T)d!^5Fl0
z?!mCQD>B8u2<uGp(Q2F#SweaWRXbTpOeGXUPw4xEa!|T3Gp&K}R@isblxdA?AQn(o
zvhh&%h_!zX=0FbRoPbf{aW3k?2+f8961?+$n0)vLow}J2#`V(E4x|30B6@17K&x9R
z>5npsYWs0Q3);+Tj%*I)i{t1JtI|F*VY3xmX>IOE!i^?{FB%iy9f~P`;>PG6t-rk)
zy6bwMPx|vj`xjNCz`@Zl6xi0s&z_^$%HvZ+F~+{wC?$S_$C*&|zwY$iR%LOCj!a?(
zvk4tdIHqS;%{4;3{p$%LE|Mt}{N)D<41z~|=_;enKpFDKOYMfFJ~uj4IHjt5wjbxl
z%s08q_YC4h-Ie0pzAfK<c*BftQmEAAb!`RzBz1`}gt+$ND|lR51O1qb7-1lr|GHv-
zIGpu&1U0fH#{1h>d_=j<CoOwG&RQ!0?~D8f7Vyjep;{`|R%|FuQQrX*JHYh!Nb#{(
zJP~o32DhHr+Q+i{Y@$}14&i)@G=_@n&olBrU5YWT>QdT9)ZLU1tvyCXyg!xPe3Si&
z5JC<2zOnphgf`Q}q;`#nvpZ{>_6?V^UqEwgKQxFoq+4G4W3^2VUoeNjgo7g3LY>Q2
zi7TsGwcMoi8`k{G=#PTIZ?Vd;n3ySS2$1~K0yqwBts^m9kGQxLCns7YcfBVAY=}IN
zRxeWk{)Ci3K*|gz{Z_O(XSZ)O4SO`Dj`Rxy35Lay7KZtUg#VVSaX<S3pO$`h``c*b
z0iEQ7FZ_$+GlewWqzM>n3(m!vC<B`6BG~kwXW4|K1{R;%3fGx~rxu)~K@t*gff#JO
zMA%*?uTQ2xLX2X8!~oIx-5-08f1P^XWR~5TF~$ZKrB2G6nfkw$Sq9@Xfqtl(@w#LC
zpDFCU+QA)8VmKS76YA9Y!($bsfQo>o;oHYjd?K+4IsfMGN^v*ZQA&a4*l&dfs@bV}
zEcLPwJo;N<mKMHCA{vUy&V<5cE^T9I(;v>PQRUAWAo5Ub5?L0z9Se7H>H^+;Op$W{
z&J?^|X<g*rQ8G_uYic?)%9xK~d0oj$4ba~p2E!(THw?lW8DS>C`~OJ9&J1H-Z@0bA
zt7FHmevM+e<6d~;IPQOlSKz$0RCv4^U(1pfD!(lK%WBP6JN5o2e>bk;Bl%qaNIrFu
zPb^Kxrff@!8`Cfc%tIU^<w>dnNgy=JM)1^ZTw{6q%=a7#5!aHbK6WXvF<^))3_jZQ
zfEj$h!UujRU81!gC1si)Z>G4mQMnNIUnd_Ke`?mFD!-7VkM8$HcM$BGT`bq<f!&aK
zX|==XY=rom-4eCqf;%U=xYCG*Nq2IZ;LllXS!;hg1Zvvg!akD3fsJ6+944`G5h#{C
z=ZL-{kI_8&tbDn-`r+H=BL)MfW*=S!lZlXL36l6&o$6&%<n^7ih&jp;rqc<bQJd)+
zW!oXhwXYo#Kv}x2hn7SER+5<N{i@kWt4rMXx5hf4Dlq1NMHo}!JyrhGiP&4&oF@{i
zyHSB2JEQNBCH2}uhb*Wzk>#2#d#(}On4ArCT)nADH-0zH+Pag4%v4?}*7LvhwOmZl
zp{>ceCsf@<#rx<C0xX6H9!pXZ=bu^#{B}}HEDLfCf(Pw#DoG|*Cr`Kf<j;AEi__fn
z^|v7GbyTv7?%Tn^!AxnIhQ4=SpFA%UDV?@upY;hcEGiJ$E8qa2#kKd?Fu?H(4L%kk
zhqq|&{<-H1Q{ojc;T=zWw9h<|ph$7AAtiYgS$GWY?YbUV0D<vNYo7a-IYASQvZU1=
zPbc=5St<5$Epu8H`i1FS#vok4>g2$>d&;Q<Na)1|p1B0<I)sCCY?jG!LkCibjO}M}
zHqOBars`4wVf8RgAbZc8zTV@sDy6S}JAl{wnDD#A_l)xv`bX04vy5|h^l|Kz8Odii
z@pqbj#InDZk?5EETTb*`gbIIfGyfD`H6`Jpp^}HgyTzwqk#{r9pL=d7vOtf8?y+`R
z&M0*jAqXRa#?a3HKs{!<hJAkktp4zr+e1{y74BTrbm3m2XOY%aFem@4kVJfhLNpzo
zC?z>pW4q8v)7*H(y4~_N?X!G=p78)Jd$6MDB3oy=L(n{=0h~!#IMURxzwq${hnt%c
zg*cQ~j>c_%`sSOUarS`W*8!E=`K)byy;(c!dJ&uX4LD?;$TSB<mu%W!qBw+lJJ*23
z+B<#zE<azNf3ZXKO+kBtBP_Y#d{!gD(aD|VN&$vCiT6?)f5bNH$B88daAk}$$U-qR
znp?*9J~~frl>#|+OL6rk7_RAwlxOr~0@maN_)RjujL>Z<z7%QS&&H#<GTq%E@y9Ro
zXQ4IwD$#vmDzyrICWIBHcX6Ap(qSpfS}5BPV`+_-IJu*aVunQTJdOz|K-@Pk`@g+F
zfV?ImWfeeP+9-k25uqIyAjjOw$Mpuf_$w`~tg2>dN#}FDP15WB_UqfjO<uhs@?hnI
zU|SZy)GfaYQlK(k(gOkVo?*6l2?Rrl7o`}6%3U~PO2?@U*@0j~6~>DP>WU5z5yU>^
zbIrrCP~_76KJa?#ZM|n1Y>ngQ8nS2!6Peh?g-!j>TQQO|ddz0ys2Dz=3c3QvdyrqT
zn^33Jqtk+^F!E66FgVr~V;E!wqD+Fo1Cpdxm|jSNyt0_I9)572R_B)krf@fSB}Su+
zWwKd75~E+}9IuwTkK;Yw*`8EVxhShQ(_FI1+SM6@V&DyJi1V_&(RPRcvy`0FKVnJ{
z#F5s(Ow(DF=4DU0i3x|z9r(hE(f^nsH5(6rDY7s(>MRot1c8v@c0E98;js%&zd+U`
z;vkkIgKhhmm3OD9y;kr{t-07J8#ecE%bD$5MBx^vJ~#FYeppVK@srnY3D@_0YrpF@
zuT5Rh0q4wK`JfVfGF1xeFud<r{_6-`-7xxqtz2`1#hOBMpHt0;ZvUs3@0Nt2I@(Ru
z09kBzx8&17z#Z()5pC?(m0H0ft+rM#K&M$j{mDzi#@@~_q}$`j8ed6N6}&CxF&9^{
zz)HBb13;;yiT@ir{R=Rn!dQcIu06-QyVrp;oqn)&lRf<H7wCc)(pp+*1tg<nzBe@)
zLqNMKm5Rv4PXcRq_N5AGSmyec?uV>KYDAI%j?PmleY2AY9IPy6pE<!%9J`~Iq8v^C
z<W*xGBD}=Jgfqgo-oJm~W<i*~68G}V%X4fi9eItg2=ftXE>tDme1bP#66Hn?t>aaI
zjD$Bb)eSm0(Y{9sso&D8HUu*-3G${XwpcuIWc!nRS6{6ztak1L>WF+zqCB^%Zmi*;
z((tOK3t6YU#+gS<xk(Z@wOto|t&7f?cJw3`NpPLu=+BJd@`P1i#rWB@20p8sYr)3{
zl(^W!H6Az|Dv4#8KMC|JuJVnUGu<ngz>_%X4A>e4y23(ovtSLuIoKx2<r*90uCq{2
z9%(Vlc8*)tn^_tjqj^?HjI$6G%m}nut8qn#y-SU7F@u`i|9~u97-N9=LV1sxNiUCx
zkjzrN%lC9Hi1fTiDd8Q8SyO#5t1~)H`68XxbW9!mxv~{Y_v6-BPaMBlBwT5iiz)%u
zfR|DkwZ?g_GejmbX&pO--~N-VE#Q<1f6J*T=H-;$YS}LOoO`{$8}}NaE?e5jD5gq^
zQr_r-vg#Cm7T357Lh`PR^!6m(JVG@^U9|ydZp>!P^m~*=AJdtU5oWejM<7}v$>%aQ
z<M*_(uu`>pU7p2_1JD~Zg3iA}k-}&Dnt|ZWA7^cUV6HIt@N&8$c{me7Z5;sT@<`(S
z`;DGRqfa&#dj!yQfwM6PGYFEYZ85uD7La;$&U|`^;*?I4lZTm6<qq$M6MT|3DT0Lk
zdw|${CGML5<66Lf1S9mp+t84OZbva>ips<mwZEOP8h1Xu3jTY8zybh25?UkgsMBhg
zBWn4GBd`t<LzhE112*WGd=Zw;F~2+^e84JuNKm%D(VJg-0&6dt3k3%&U~L3jvmn>=
zooetWONy>o-K;co_CpBHqY4}r^GN*O$%f5Qpg}AJaHfmOCV5#vTsQ;ho0R))=avti
z+`VXW5gq5BPeQF{cH==<jt6n}HOrWUS-O@Cd+t}TRK1p9>&ZIUH?%}y#Fbs?{F!<r
zTh(I0W6EM`s~JaZU-ivla%O~RFjlO_NkM`|>6DPxzwbHmbyw;_On5LH{IXlL!Wh?!
zxA~y#>`I~%aiZAX4GUH28zd2W%;DTle|VfW+<S7O#ELuEm4)r<Sp>a`+tyAVO+&lA
zNG|{k<~Vq!?sh%;zWfZvX;O7l)v+%Rce{txJyR6}bTPeuCik?n>-vxJr;X#p&4>T{
ztG#<|b=Zx{{@dR7Duv5@?0N9Bl2SyePTdDFAR2ds-S-03(#FQ+{Vn5g23Yd#MsNND
z=>o6;Et+)*ZhL>6QHGVt+lqQ;tv{Y@A&ugAe(uex59MHj0BSNC>lHQQ4+ga+4Fd%e
zObBccVu16^3`C-kphHfF_Y0fJes}Y)f7CsKXv*@8PI%1;IoLL`J_8-^55=^07Pg}6
z1#MHBq~-}8pjLF$G{WS)h1fb0Y9h&kvtyMjoe7-n``|mjL>}6GKyx`_bW`Po#N)X3
z<NamW2>)K5rjq>Gfd#;@8<NMgM9M=S(B$g8mG}!NA7&BoeSDieD+ek}`aI3YbvbSy
zpRTMk_<3omJWXgi-y>(5tZw)#65{VJ@Y1z(`mU#YO`!;&74DMsfB96SpChm%;F}7O
z=xR*-zBTm}WWkL^)u`jlnf8Facx``US8rIOCp**YjCJsM<bF3G1R)W}6xa;*&=8L_
zF*1Cc$*~$;r%Y32y9e3UoM<p4=CheF!(e4)b?UxA-4!;x&;9fEBmHMUtX_fk-@)SY
zZ&`ns&+|P>S!3W}!`uEwtBSD!$~?dH$|>aM*F{>CoFM3KV6eFQzeKOE-Cu1wuUl5C
zWB=PrAmvtVO#J0>rbd}rtNdfZN|f(k>cv{mutde{f>QrLCuAo?IS#i-x8FQB9fByj
zTunWVcNNU|Qmt{pB)pTs1UO&NRt0T(+7ChU+&^*?O%((EZ#{GH3>d?~C^&@SLm#17
zKXOLe?I7|%x)e|ju~U32U=alRg7Oasj)RI#SDywZ74E>)pW`)m>??`bP#e8`FQ&U_
znX?UuzSq2HJK>{vX_opQ1Me|NF8BDa@G%)U2Nv(oQTqDopnMnQazHELPVx|*p$kz*
zwi~#+5rD2AG`2zK$<V11g^&JBwo;22Y$T;2;+i%o*Brw6bh9pM=k3n~>n9Y|S$d46
zdjs`5hBGQT&zpsZOSb%KrGoXOR{Lbp_`~JG5$7;GIs2G$bL!N4eH~3E#?Rm)HMJLv
zpsjY8hs7G{;Pb8f9{smzL5(OG*si=wo|Ai+^7|34{=!km{>PMA;1J95stntCBl7|`
zw>V@T19|61H*xUf-me6_UZQ@EO$&FiQXt7_TkN)Jd}yie=5Zmz5`(*Wv0R*Luhj47
z9;Sgo7O+tZW%r`<^KT8k<j0lr6NRH*hi0wJA&MrR3l$!rS6>86)+^fab;M()1CSb`
zEudoc<WB*>^sbq_1cPGbCPt0?K?ry(X~xoiPghrW7g7A5-{buErv#Cg$E$lEomnWh
zYrv>iO*gl3Hk}6e;rE}h0zKHEd@D{scp|39gR(ful$P@H#s1N`PB*XrnG1Pd7bv+f
zv0j1U>}0d*!M+Y3cKHC>u`0*sHg2AaLS6X2V*3-q0HJG-OgcX5rnMzR(GFCILSk}Y
ziglpdZ0sJA{t=$5O$&o!^*9}goj~NM$$F5nA+cls;~vK~A0CvVsERiUg5f6yMV(B~
z+HaF^j)A3U<c|E8^jOt3cc{91z1+nUoR$}zaWr+qWT*u*P(EX;1_PekH_m}brYI9Q
zRp`*~!bMSYw6;!;#bE_7*8&Z@DLi~-yfeSxTC1<}_K_C|La(`v8pDia{cx>8s)!sl
zSH<YGg(ie?t19M0Va6U2RdY+$KZv~KvDZ~oVXfviqhUcfo67#}5Xr5H&Nd&S{J(HM
z_kYq|-E4~bNg6*)gIz%k55$6f<$h&$=nl&2)`7eoKs~4DVP3-D2MuMx_iGn(s9)t?
zSiGM;HY-;e7`z2s*Ynu4b_bZ!_8;F69>qxr2E-!@s2CWu75%?5bq`>X+A~9a)hbK=
zBY?LhtA4%0lRq_1U*!$2IW<e?l*PY(;WPnU2;T6j;{d6;re>U1UQ1&UzrdD|j(#qd
zJ7y<CtOEzOK9qm(9ovy{-?16*tek55XRtJfXO#sDS1btgD=}^2w{kh9s^;}zj2j0Z
z?BkVoN$Wg#iMK{*Y7=mqZrn3cHOm&IzrCCfXVfb&1aAm8D2{AfuXm}0$)=mSN#hC9
z!6Jwl^BB33WxNu&YP~d@nN&`K4!_@I_qPGcy&qiC@&i@lJTJAFjoEi_oGZDuODCyQ
zzm2*jWAs&;l#8UT4kd{2ui-6oIfl~5p8*pvl-aZEIcTAsU&Dr<nSS<P=NP_nIhM|(
zJf2o|+F9i>(>_b(F-K=!e1G3phesvxp0Qbu1{L6Hp$~}+WN}HLIVkE2N>C3(<Nq)|
z8O5ocSsJgm#w6TmPeXzCh72_xIH8`K=t^ke{9U~chig1LP*nKV^7Fd;@lAOLK}-2b
z^Sl}D;m_^z`d;7CHaiCc7%cF#+Y^pT+gba=4u8j8gv??P1}T%4zH1zb^9mKBfoS;E
z(T~>~Yy9_v-Ol<T2ir#^d`xLFFOO@@|4(G@$-msCW*;3Pi&y>Csm2pnk#}7inds?d
z**YGb?k*ELN>{b6E^mY@Kx$E<H;Ix>%XTeF!^6Wj?qA+_M&Ixe)=<l6H>({(EcMua
z6qbF3MttKqCP5S|^dzvCkjRZ2WFqCb>R)JS;Ze@`jTV9Y1{40QXmeJ5zQ(!)e>sU2
z^H2<(LfXWOlaLgqbn{=gWzcPKvwl)F!%T9?$}T86EevvChG5z1erD$C06&=zqlQgh
zx_APKO)CM>PyaKS>gT8<E)|V8S}JnuXq!2&p~`==A&`@WfIUhxpN6lmiq0lr)eW{c
zcET#rOmRvCM99i;B#VMjpkA)LnWs9~?Xp(fT2}Qa;j7c%#cv1#6pj^>U--hIQ*y9`
z)R7`<Djyfeesp=7?^CKtjO1JGQ}t>Rh7twOboG$(3w#of<gIqjd+INTmoQn=e(>a6
zKwP0J7v7Z)r~yAvAlS^C>eS6GS{p?@^`8<Rq`+;10^Ep?&Wbc*n<7v<`;SSyzXiP6
z<Bn~VaePz~TqNaIV)&;^k7PO!NqX5s{;Wz}VsX0C<a9Yiub^xlAt?**pVc!t!W{A6
zx&CC{m#DQknzZ+LLm0X(;Qjn-M&;YI-&n%@X3qZ)&-_n<@IOAc*J#qO_Nmp?rWad%
zN55M({881%g@6$8BHdHr^}9%)=c|l)FLiwwv{6wN(+lrDfW7iNbhsE-oRq51+LSa2
z5c01@Am#tpL!ZxA*&QMM18(E9d4aR;bvDT2-HO8{vT^9AmRlP9c0?=J+J&m6Cmj0T
z>o_F;;LvjS=qQ2RaPDasW><;2nNVoWL;V?QetNzkH;PC18;8I=N5@SXsL$cq*xjsl
z5k@aai&D9`5uz{6?<d}w=yEZwnu@rW;APGu^=3;boesqa&$MMM{o<~6Kq6OdAD7`%
zbJy5x3Ur0*Rt*u5MpzYtwZUIn5kzw0$s>hQBMy7LI*YV#ej#P5QKx+BKh>C*uX1+J
zg4%HEJ|Bu9W+u@jTSC3)rv6gXuGr7;!u^A<@^bqFwgyK7O`^@PvVPRn$5CEx;Al(=
zIiLE5x@FUboX;}K;tk`ubr$;{R{bQ5ZxZl&b_KL6!bL~YnKShb?s<(u=LH-4JsakK
zetCx8uQk|bJU-ucoHdUDr_4NW;9bU|W4fQ({9Qnpz6yg2)`^`Rz%e7=Z1x_OIK;I*
zf%LxyHAiLu{|2}ptI&g#Oi1;Ak(A4?dIOK2(Gdg8el}~Pz}4M#-54GnjpH5{(GYuB
z93O63yJmaKQ|SHW?()*9>FH+q+BLg$c6Rn&B;YST)}|b@&nN$jwi#uh=u&4Lb@1!U
zyu5~A7OwO`_1n;9os?58EL;-T7?_Z7t~884SZ{j;HLGR?XYQx9DXBU!Col@7X{w4o
z_=gCr<Mf*2cEQ(!QKx6Aq#%?;7#nlhfkdjbvCN@`EahB!-(75#v`o+%e{e_W7zx}b
z0=S=W!>3<vB5X^FetWoWMbl$%L6{9xG21Sj(iC`U`hF?$j>9(uoXcc&8OM7AAOyjN
zM!KAAWgXX;u;;*NCZHFJUz~x<E8aq>t(FUeZ?oTYS&l|%;|l77i~EToAO&*+Wk-p>
zVy0?@^q~KZC|M@G+WzE?K#hq+xmSX7A-K;h!+gNbFsedf4%Is0>H|$YZ+uZY)pl9+
zL+-XiB<XnVG<5$)kN@b|v<>S}2^W$7MkFcln;lfbj;b!rGLwU%%Bet7d9L(}fvjQ-
z!BE`#^G(C$Y=y5T6BE?>lfBn;%8is)_O8i(VGft=6()PdtL9IL$|Vz+RXuJl=|_o`
z5sCESn>`X%vdaD^<mCVOyaFbm|GuxDZS7R`_jMKDlz=MbCD>u`$!joaMCF>-ahZ;F
zanE@|-r~A6Wm*^uo-#ejI|1jy^?)nH1Z_N{S!fKX0mmq;**3{Yo1tudju8P%c3z29
zmHY?7jL`B$BkU&^`@TbN2Blg|qxW#XBjfPS#__<t=<iGMRQt{Qk2l}Lqgj>_)?AS7
zhmXLGCtLSUAbnZF;qSO{<|$8tlPb(WOoUnZ+ogsiF>Cz%j^gu7uEKzw2Z;EPAQ`UF
zfsR1@48k;43RFmBt4Z?I|KT+M+N9$sZ30#ki8wB;d-unvg@Eh5kEKB8WA)6;WTMcs
za&~HqJ~f<eHg+&&vn8<j9NRoN!{8W2s0v5$zUSe^V-tZ8fn3ttDuM+&Eqwh@IW}ti
zVjU;?MGoGam}8I}5;;~X2_>X6MUPQ=8>BFi&g?CPz^ash7Dn&M1gB>);Sl=qKq8JP
zOT4hPJbL_s@?$E@dxgd+DX_`X(7fmMs%qaSue<4tXPU*u(*;0B@ETy56i%FQPmC6{
z^Qh2!K08uE_=-DAA4se?N(*Bq6k6#?G{J_Hdj+i=SNxBg_-g;`KWhdZe@c+et0*ie
zJ-A{lRZ!!<`TBa^a2<Zv!}oG$NK(Y{YH#!sU@%o)McaeLnkBupJdI;{5+!RW?HY;c
zjVq;BdO<48R#C1c*z<a}hRgYN%`hvvZODL6JIjsuJg7WPqDr8#c!lzvd5b@0fUU_s
zeGe<gN2Qp=?%A7Vz{sJ3G?fSSHMWrWG63d4=CTbdDc+%V4KJEzKnchO<{viZRK1zq
z9?gM77ct|%;mx)J>A_&k44N-JKf_@>=Hi3V$e06fbN%?iV?57}vCen_Hdzjz7q2Qt
z84D4$YzN-hGkl5qoJDs0V}1d1Up~q*3N{y+qF4bUzegy6>abJN-tp;AKBAstvfvjx
zppfHV;_^W<3OrBWVgIojA^AR95dHPz67-_7LbVa}ncZYBB#ROmBq(MOIeDtpQtw3O
zumBF=E`!%g<k84<BbDG9ddlx~Zi8zS49QX!<GhxXc8efdu}+<Smg?ofN~s%2P%gDJ
zVwp$6vfncHbUP7h+6*K%-5rk8p?3meNiKncPgP+~*e$8aulz=U@XIR(fS`ifAa4<=
zveSZUFn0cdq!5qT{(kuMAA>;raN+gz8qBvz3G6Zn^D%kbk;2ODKnnD5oR$G&ES=J?
zu`y*!dwZ?<KWeh1dDwI0FaHKEhg5vk(|vt&KDuDUY|I=(g-t3M%F~!>Gy6$E*-bhy
zLOPMcOW^q5Xhykp?_jdCEvdX#XBt)htR?W~7ho`2d0>AIm&9%Pev<s5t#subij?V0
zZMxBWe1=-&+-cV%I4-)b!r!%g$!>|H6kHyyVOjYa=0VocpoO<Wl;2>(>j^k#qPDez
z-{BDFRrRP+sKXIZKIFJ5d{y@2`_l8n1?B^`7VVybG`Ie!umol%{y}`X3B77W@J{O0
z|HIc?#YGji{lY^_3P?97(%lV`(%s$N-JyVVNH+-5-Q6I<(A^9$bclfTUOdl!_xt-!
zzIDVACu`PyU)Mj~gAg&fti6whzVp)XK1LRYE8yVBcb9Tah9)AbEGhG@&^NcqxCV6$
zm*kyo{2ma-U%EvAN$z7&XxtatEnvoqN=r(lF?Y=EMF(1*l6455@C_E@)%1+CRjpAX
zP7#0VYbBzm-r{u$s82nGTHAA?-0T#F6N9CO#e@VF%#)%t?HbSuW*w6b4mxsL>!0j1
zEvHg=4GcaF@)a>K88ZoJ5djL8^-23gd_!-_rjXGXLXO>&hRM_rxY-&%mKu3kH%S0E
z6mIuB4l6tN#g8R%P9z2I{Nn|8k9AbeRMaF)NK3sn_#Y~e`0Z;1f6aS2-kqtR!q;?n
zlP`}-ga9mYveBhSEI0)+eSW-SH|_=b+$>w|i9AC@c9G)GQjP*4C&bAlR&+4pk>tAD
zH`67^YKnLRR$sRY%T&!5)vDz_GUshj)xNR9Q@8D9MUEVVr3Epo(18kxp=H>h-Fi1_
zf|g+XwPTfH)7WrbGn~~)b^Q3rev3KP^EG7&l-Uqmi7$0qbWj;$VXgp@v|$)~@ifIP
zgL&qvi3Ky^U=fm+Tn`M{iJrmO_sRA3215L%8NAhz#vdo(VoAohTlwIeKc0Fm&N65I
z%2v5@d}PC@m7I5+eCUQnKJX=$Qge~TtMbHJlXuaGi@zHF=bSga%!x3=E4`9WiCZu?
z{FZkaC%;68N5hk3{GDT0sQ@<FHwLvVE(s!`3s`4@gqVhzY7GP(tl-MkMjpGONYJhR
zbOUo}`Be&Q4IQB$SYEA(7!@qNV4>eTE9Ix~#Y_=<M4?Iq9P%VzHH(xjhdmMwalR#D
z^K{VY{#*Nakqg}$5xG~<GLWGvC%+q&-ff%(H380hc_0vm^oAdE$^{(Cp}I-B!Jhy@
z3z<zU{C`&_;tIz9g&ES)w45?|oN_<40oYN|QA|Zm%_6^R>%Q<s$lKB|M5~F#o+qKZ
znF*t`2r^$8UA^RE%WXprNxLCs>B=%)SmT9@&tQbqq%x*=YCTGG4v6q$yWHYAr$_qO
zpxGs#na^Kb#u<^w3d>m*!AX7JlXD&%?vtL@g(0y2wkROSUmz!oUk=P@g(DnbXl{5p
zO|vinP_W!Hiq2mu2yMTM`m|Z9qx{UBRPOk8g>U(|Fk@U!-Se@eM0;IFG@l}3aaNq-
z%c|^g&h%q&+sJjw2O>gc)L0*o`jDhKPVd+wDB4q~*uc@xU+}DMMTqEWzJ&M<l+<G~
zGS)J*8jySxohpD0kn+(af8cu=9i@jMb|_JhD%N|eg@4!9)9r}6B1-f;XbQqdm@7Tr
z*ZnbYC|4a$rlOQ7iz#{9?0CWrQZem{No$&Ns)b8Agip6v-=T*uRM0oSCD&^Jaf5}h
z&H+Fh!aV)UdQ+Oj58UP`bjT+YCVoQ1iNLT9RG-)$Z&<cj&Eu7oRVg@7{MqN}(P3`e
zO=6}!-@BIG)rwy<VG5Qt^7zmS`9xYOcNSayKX;`-($;Q4$$$I}*RP3C`~*A0B={sA
zd%)j6^7BQXFXMw`%*=}P1Ml`+hY79TH0=dG9UJ}D&@w2l(a_Ma8Rseoc}?3a$}ODm
zL{o<A)XqqLdavrN*8TatmjtJuXjE&km>-XTG`1oaMSt)%wwFwXkX;jRcY&zz_5S^8
zPwX=DXHkxPbmQ<hv{<X|g&owfbj~x^KvWpUcp>bLG6y;Vhkq2osD>%J0K=MQ@Nbj5
zYGTW4N52@hMI%~%4_E!xLSMM_;Xr!hZ;=cN1d)ZXQpsPAYXWO3GE(6&%A86d`adT4
zU`e7{8*ma%@r|)=wLtkO+vL^EBsT5mXuZjI@XfGmAct-33Y4L_6uQ5K;<Pg*Xr6{}
znYuq<a){h=e9Pm0qUyQ`J5#PexIitXjRpM}Z1&M+n9?TM-<G+`6=Y5}cXc80jxD^@
zK#7&bq!4C;HJu&WPh*n8OSiscU$A*5DMy^_2L8KPS?x*7nsst0x0>2psvU{WEZf}-
zQZCo096fV|5Vv7Us>8WhVj=!m;>q6?@ljQlEHwjiNU%zHr29>f9e&KTwW>*SZPM;V
zz^fh|*8CrJ)ah0!&wq6@uSY3Q)JMf^+E+Yik9hN6c((6R+5H;eS;t>)cF6MI=g6Ce
zT7y*d^(zguo!a*B5WLpsf=c16dJ&A_v3j+kisA!s`B)c-JR1Bzl}b~IkO(3T;&Pc_
znueW;CEE~shXQkL%F>MI<k^N;%Ci)=CEtqaomlO3Q?Zu^e}u!SWOjcT!Qui$!-uwM
z#pb1w3(DI^2Qa4CKUke7-Hx(nbgjp6xrUS4vtb#d7nbOb-5uNAotbcyF$qS}ZEAb<
zi$g*ObrsA!#;_Ky1ZhEV-}lheThd^D%-~Sc7mjqC)`6O-CdhEeR75Z$(=l>&G8`CA
z{8I#O30F&Oiw;T7R`>T2u%6aB_&oH>pgXp;$6tD)Pe~Ap-q-2wG^}bJxCw!c5~?P+
ztmjOC?%&Y{$icSic>PGja!B`|2~Vb#fLa57D5b5POdHOs`}&7r4MP0YrE5~aJ6tu=
zx`dc!s#gThnOyt8o^dpC$#+9!ljjq+c0EeJY`8&tniI!5GrBEKexJ||DU2lnawxvN
z?j4mL6ryL&r<IwCt-JR}Zy2jt6asEnD*pe`c>vr?{337$^dThH1U((={w0lAJDRW9
zd(Cif>D*&l*)o|cRjGG@NZkd!oWI!MFG!R5NVFVdUm(2+c~`wDiePTg-v&gxP=@v0
zaWjhVvZ!lT+j)b~=zLVhOP4X$B<z+gZr_QM)|+immiSk=xlfxUO-BB9k1Aj$0@>`k
zcWkO9wTP2enU=f!EzVX9u8_yzlYnaG>pZ)l_#{b4YZ%8aye|qQP!6|*nQlX|FyiP<
z$MUR>gjj~MMY7GVr|Q}fP0EAB=#y=KBtW+`*2><@6>5^9(_3a!=77pR(?os~r(JVu
zh2jhs)6#*;iWd|sgaQi(4&n9RhwC=U*)7?u5x|gt^KEW&Z-cmZ&izmuG0G$g%<)`b
zc1ZVYev81!X>p@nhQ2saWz|B-Jf=$#Ctub5A&m+eS+`?8kny=8>Cq4ki!*rXq5e~|
zaev*7Tk=9^4om*~MKtip2j18QC#c$)?xFSKy(fy_?3-3?O<gLZhtv0eAU#5K*GvyE
zXD7=IO^I+#lwF<A8n1iD5<R>?z;q(cW{j<bVdk4p@{|;Q#VC`KfJ^W_$})S~^1ZJj
z&<FDWJVuV({~2!vbmn3N_P(n%4$vzS@z?Bpha1cB|MTl8^%eg4Vbl}&XR+s5L-^@X
z2?!v2cC)duStd6@kilX>i1|{@Vynlc8NGbh_$0h{BkKx3WFED^o2LBIhY`i8@mcip
zr4&WWMnB~&{V0KmbA|6B34X5kdcme=Cl0Ru$GhuQ5O-@!WQ%GV>Q8+l7;*WObuY(U
zJ2Vil{7m;LHV6lXtM$!q9=V0T9KN9+hc&gR>h~zMjH>ib8#s4_M9G4Q);~2Bh=*EF
z1#_Ujk-=Z%MT6C6ueotUho;=+xQ8Bl6X80FfN8M6tcs}4>)ti1EXMRX+BBTcwd`}5
z#HLXBpxQixIt{S`?sidk{3hu8UedH`6)B4lz4zk*bDzGgVz45PM=#`>DWrBE=m$E)
zng&OFo1i+mY{2+;l7ufYyn2;-c;0u>8XSA`h}gRgpH^6RJT-SZPf$DT@ou588VW=U
zreSofOBtK-Ec=+PS}g$4Lc(<_M_DKO+H6h|8f~6!;a#1E=>K_vwhF&OS|RsS^v|qw
z%U4QD0gD|A-_M*bWB<p>dukLCVEB^h>(&6gMR*5F|L9Qex6A)A26p;J@e;?i0%bG5
z|2VzeIK2j3Eq4cWykd4*2cE}OEn_pqa$K;|0}(ZW*A@3GV<`;9MMVQJae#ofB)fid
z8iBlzO@fJ;0*iu{v=#;7pnxy<sY5{wGl%sm{`3qR&h*sqIP)#Y_Qt_$>jP;c%*__r
z?=hE$@j2&U2#8UY13jO8{rL&`uMD=?(1d#T=TNZ$$&%tjV`TW14xRBYW^+$~VAZBy
zboNVYGzx{x-*?tLrtLX3!XG3E+^AL59y}dBB!HZF5p72td?|336+mL5O$HT%tns?H
zP8G-LM5IOh2evr58B^=wOJsY?SJ}U&Bd+Al>O3ottmW5tW>vno`<lfAPNXO}ea--i
zyO|<6TUr30X00%W4gWgF24m>b-L&M>hfX;3#-B`qN_7O6wWpN=LcdvpF6S(rDKz<+
zHT~WWYb%W+J99?&{HPxW*AJv5AZ&ZSjgjqD0M|7bYlh;S1&dWS$?G0U&@A^Z#lhA-
zp=Yh<ldzXkfjw=#TUJtCp5%IKG;{kXkL!i+{H|8gqDci@h~Vr9nU2f6a)8HZ!KTB%
zV;FJ32_(S#zs%VGyv1j(i98V_lU0vJgP%Ot6`ED51T6qR6WhGu4T%LPZj$9!lyH|u
zBr!hd5mIkDghJ=QN(Bx4)#mdFb9!TcD>@XN2sD_j`MLg_>!JBGrspRFDj0>rgosa1
zxOyM*yGb6HTUO)q&V!|sY~==VgpRjHVmNN&S(rE8{>x_zA+13pp@W$$UM2Hx_Nxk&
z;czb;okA|dZ~qO77~_`h6?Sn|!LRg9gD)z%CgDiHVx8ak08G3yq3NY)*Ei*;)al-x
zZk5#B>Y%?4p6K+A4R%UUgDG9B_?GpYpTn}THAKy^V{wc20ZVN_a4Fg3vRfet9h8b;
zOtoh3PK9s}gUBbmr3Yk#t&({f$MOf?gfz^==xvEQ?(o%)s5gtjW_#+=Fe9|oyJn+0
zjtf2nR0r_Gc11$4J#rhDa`4;JhKDY<rr^m`<>D`Gey($}S;0$~1xdt4EVq4|TR2v=
z%7W*m{<gfcCKozeg>LF8NLlhL&MjZ4Bo)yk1c<5F&bh7mK7s2QLuX-SRl&dve+;Y|
zsjnBMy$1#Bfy`kw5&p>f;);())in?!DDnT@6@d|FQ3=FjJHn*bEWZ!VVE`I}bj{9k
z{m!OgFE=0edARyz4a#5pT3epZU!I-HY?@l^RBj0sc~mK$B3jcwOqKI&-;}RbTsO#T
zHmC3NV{Pf<=r_&TGI3UsBMHI<gCeolq-0tU%9VR!Bh58c9atD<>B38C6&a2?nz)t%
ztO(r9ijRk*?9fIuumu|+C2-}R7)@JZ>P{?>_q!!OOm@=<Tdr9lD7BhlglK$(X&vL~
zAfRK#m`eUYfq(rN`AHGDT+HRi1q>9eavW~kB;ObJO)ZId9<8lj0=Q3K@0>^~N4SBK
zT~0%A5Qzd3>#65a??Ba3Q0*~qG$_m6fSRoGd~)I7=kjyuLyY?aMBCYT#~F1bv~*Id
zZhgKRA}A$U+1T`J0$(i;UWJcFA~+)`*hwx&%19^PW`2V<ya~)cRQ1a@`uExI6P=$B
zb_0(;JN0=J1gC<^?^TYmrQWw-Rq3@T3cIggyS}_<P@Cjc$P+f`{#(pqqlR}Lcje2q
zr<DwRa^2iiysF0|KBPn_fKVUM>w8nbX1o9QsKhXcB@@zp1;UTa7Ake>z0M6^G4Q=`
zV#ldPkaOO@L7K*A5_EE4S0pdf!?0dlTdPbMJ6UD6Tx_5^y|L1Fbm2|<P#7ItwEd2{
zm-+|7pZLQwMcBCuFL*`TqbijP;W9p0*$~?HY(xnOq@PF0)X24(ZQ=#mQwc|*zaD14
zVdQX0qD*iGenwhu4OSRB{!7ppXrNFRCKfDgyDN#{zUv)1UO(x*-uVHh2IZ85<K}bp
z>4WbAliK#qYL^tT1wq~U=V1L{c8`0HB{%7HVGwzaWKn<Js82@g)Q~UMRJmz|kd@kp
zIy-w>klnJ>gIG>=8$z7eaNp9m&Ye)S5Gv+R0;&jYhjC@y^738lac)F$2r!zzc~&QR
z!Dj+H3uX%3>QQBkGF;{PS5BdJ$p_3i)-G7wiz^NX7}-`z7KkJv0&x)IoCa_bQOK5m
zBb0}LjVT}*k$<(m`lP84{Sz+|@~Ug>I&g~BA1%+%teHYZ3LC)ua=?Jd`xeq^=y8V&
za1vkaR<ITfZmyO8*FgsmkpBhK061-%I;tOTmAT`D-D5BZ$a<fdnxYT^<`5wM+QXx%
zeGlXYu*uckP+%-+0Ge;f6JNhDRZT-m(%#m!=V}Z*R(gZj7UkBca?iHvG6){;qWVig
zwSZk|i7c=n@XCnCVBW4)a^Qlgqi$2XcMV58@P>(!;W3Nu;Ea)fl8agz6ZOLT)2$y2
zFF7SdAAO7sc|#wI<UF-UPE*pXjizij$obWA9lmqmQd<9eyH9ZShD3mHVTxy^m?|K~
zmY;Au{VN;5fe_}xwb=fKCQ(k`Y<&5|t$lzs7;kloBvdzqGQUs-1~X-H609tOkuc)P
zdzAi`th8t%clMgOFe|Dfb49g<%rZ9zZ<J5~W6im5sYrE_7i$7Ne8~*tE7eB;85$aH
zqr#Ra@e+k4uq#W;ckK|mZM`~q)<(+Mvi?ab0`5=3_dvJ4O7gZ(%YzAUKQB&{SZbwW
z-Vc9@$zTZgW^%)R=lrO*d7tY4^01jUHumHJDTTXRXzA8?4^5`IEy$`%TZDO)Tn=ak
z#A;%Q1oPV(N>^k2FDLv;E$!L@%6ODL$L>2zK*O7sMl&gR74skQ2EY5p!^W+x*TNN8
zKuIc&gwF5gBvH??O;Y$}d2+No&677`e&<$pp)RdB!m<GNXk_^ojf6ftQ@?2ldyA?F
zVpF=$uVv;qoMGxIRCFN`t|<G~0SMYHi0mPa0A#PO%wPLm5XLh+Gt=8uks{egx8&`d
zR7R4MvVK79%3W8n%;5jA;qeFmanmPdsun|Fbw2S&g6H`5MAz%{+2rT6OG1Xc^4&)*
zVyAI?fjM|un>>`Lykz4O!PrwoPhmLQY)KQ^5iLuR=#Qwq^O&P$6)y2YC2gKroGm9H
zI#7|Pc9Yp4GSZv`NbaaYg@rjL_e`JA1zH?ZI=2nSywiy`VhR<v+ic?ED<#BLSs>~t
z2~O`7>(${rib_|@6syGm`4XQcG78Y3XwD+zlQqGMN08fblPHyfG1#)m6M}WUPFRRe
z<3>+>Wxw$I#)uVd)DZ=0Dv3NbKAt)SCEwrk`#H6e2hZs5&B(E+AXTI-+P#LR$9cyR
z%pq00m{dLoSf>QFmjCYp{(3)sdUdz@Lj^Vyy?P$j?>Us?+Fe`LJ+puqm@++Rw9#W>
z3^9+Q9ry_-{@ZhVIrsW=yQNQ;;~g?usv|@xsoz@+D6akt%d+~GjGjT~s_-O2<d+el
zjBu$~Zi96Qji4Wc7X0&vG=K=4IRkM~{p0B>rP!G$RFzn8!9z@eW=&@h=&s+3#ZD&3
z;4-zg?Ry4_qFEE}?fmm_PIoa&k$CLu;em*VNek?e<|{&^hLfA#MhWksHOv(B`dRJ;
ziWxgzJP@uZ>Y_ns0o+yWd$t&h67w6}R$pgM@<`au(%>q|3R2YFf5Da6(!)~8U)$N)
z<`TrO_rPV+a53&p3%W-V#9|kLh3y5_QMH$b0{Cu8Cd3w@>3(rsWqx#UgMH~1Bde+`
zmc5v4_IYJ)j5!)10i-vA&iZiJXgHwes8jy8eP0M@(T1`it@d>zAoF4yDRbeh%*<V-
zp!?+i#B5z3FUmPOiu{^z!SpsajPu_f&>CH3(V^!ng93ZE!|SlGJ;p<J&{1EH{XFUo
zSHY5iUb_G!iVXbz?{nOQ>C%dTiZ_bieSv`Br1n!$gNH;pZvX8-#M`vj5U+LDG^4;a
zPcMI4zu3pcmLO)MhnrgqBaPU`mY*alhk!l5fBLSzT?TC{LfcF`3thR<;cK+|eyoO-
zhVz_cHMf4!F{;)Kkec^By-vDKvGVYH)H`gNc_~;RwYmBOa&7YL<dDXlK1gl`x;k-C
zj!d9RgTJl}V>X9S#H^$5mpdy%d!bCM@%LjU{P&(pQq86!T8J#RBIi*_oszNfTSDY|
zS87l=SWqqt*?RVyHoo%gG-bV)-0eqV&a>Y6$Raj66k?PJojXl74bRH#p39L|=r=#u
z2Inwz4>m>qn9S4opH#y(-*Xrcr}%9W;o1r!H?UDl&UA4=s}AGUKMj&G*KD%W>X37k
zvCi6>Ed_GFp*%Uumv}rE*mvNS7K~(XVatam#T48A9BrJ<>0tdFgGDQ78-DDGnh;~J
zr>aT?{PxdZF~k!23@RE@r$qx>eVR!EIxWL%7(%|`m!ghEI~w3)!jQXeDu5%(^)mO=
zM&}Ale|kGhEc)NPo&OT+uLof18%#hP7ogbvC#8lqzoM!trQ^I?PxyX1e(dRXlLeUf
z>K-p5MCR^Y&4%Mr48BUD9F{WO0v-mjqvK;WKUw8VUp5Ki5IWwpU^8h0Ya^LH^1@hA
z^lZJkB)D>r9LVc4EH+F9ZPHu+U+T%|k)eLEx+YE3*z$sgjfbL1Cj=AwNCwiXTkBhO
z*dN4F<M2p}LPZfwLSzuQ2Eon=jPy6ijr6OuA4%uFp{C(uh!s}xZYE{Q=0{bb_|EmE
zYIiVBG1rFkQlGJs*z(RIx{H(qx4XJG40RDm1rfFkGsd4{4Ev7eYJu@5_z0vAFn0+6
zr6|<pb6NwVaa?e^xGHn|hZLuI#Y!T(*x?jH5wMKM#Lxu{L^_+tFE@;{nrAJvSF*}Z
z`U(Xh`A`AI5<2TSvGl+v%5cnR(i1*u`NU?}ESCWD%trpcWcmD*z%qN;!~S;~OWJGl
z9nrg=9c-5u_}q?(f6Dp!9A}L0mfASXSG23c{zCr#xz_@0or@-AkMireo$Y1q1>D!`
z0NMH)iSr-G7yym4=X-*Jsu*s8GgI$)tf{%#Ry<-702@)zBC)V59?raaDpa)ozJeSD
z>#tp{%Z)b4!}=YcL)63C2TNP_kE$nd`2#)05c!og?bD2h39I#hYFIZ+i&BN36tW!^
zaTBdM{9Vr3=mgae8qMV@H!JmEVG8V~S&;naOVyjc%M)*X^FQA_I~dszl-l3Jj1G>3
zhL!IkJhZ(d?Uhq|SCd6VKWBib?FyntU@oe^DyJy@ORe>erR)*a<KF?Z9KO21CXBJe
z2duR>Yc@{Jc}>4#?c9GnlVWN5RT6zwK$=uuH^e_~d_V&#N3X)2sa0puZ^?{t)6#QV
zg4-sIcAGzACu#&Ufl|!mKWE7)i^NI2Rf!<Uk3MUA7yoX+AInYA6?H}<#S+ESVft{c
zHs1R$SGl5q(YRa|DTguM7NM71Nq$islro*<+v2drh}FSS1v*X1F|;|0ak<tcj9c~<
zynXGj-sTpquB%fSeoIbEF6`B;?>tUG$U59T^81+U`>T|{6J_iUh_eE6mHyM*^50p!
zv$Nf9Dqjo*V!d^dd)q1f=jzG=3eItzRp9w`+kGn(@D#;RJ>{72ny$k9jz;l&W+u_y
zFi&>@Iwh_Arb2Lhee4PW1xR-8%*6=Ch7RU1*yWz&jJ8+K^pu+>EDK|l%Qzjr0epsG
zT?l(H71FpXi7X|b8tiL?^H&LyKa)-Y762_B@L_|_cTiy21jt%CKKfd}NDW>hidN9e
zbA?1AFXB121M!%>Vv9&WllkGl-6#q%aV_lL5)fhhB|qVh|HyX^GUe?Ic0NgF8_7cE
z)cjUpxE@(nJ&jlom4*~^7Da~Ck(D#KOwV;rR~I0NCegCRHP4`9%gab=hlH(amWo$Z
zpxT8-&X=pjH(?5WjKj@G{)+}occ!_K;*i0jXgbJgjF<^<7Ulq2q@&@}f?o*6EGFPj
z%WX4PqPv6_S$UWZ^Y<>R*$kLdsUMD}mg|s26|yrpaqHY_ZRhC`mI0)Iyy+OwbVXZC
zWODr(eIG9)?U&UF8T?%Tcg_c!<hJ@RX!-ZC#;f51h?pEb>?kKXC6^0O|2T@MeQft?
ztN8E3pmPPC)S|tuiw}?w>3cO|1-;w`?a3M4Ckg(Y5_cVE8yN*8hYEKqi`^eNEXNbZ
z#$;n#4TGMpS3a0N+?+Z}J1Br%xUIhsw3qCN4p8Dt3AhMP7rguT6}9ZI<X3wX#hEhy
z=rFl=_<Y)l>QM9yMK(#hL6Fx$`ZCb;IBi2Dscln)xRn&3Bj935!+XxZ^1-^)nOQgv
z4kY=(^LggUWug?#_-8K+DM+=wF8s(kCBIu9cFlb7!$9~Om+iO;hB_m?bzW?Jcf?!W
zA|cD23|^EBi1%ZPTM?^Exlk%j{?3}SVkr_5fdZ1?YNmmDS_#6SU5JUVzbB(8?Wz2c
z=}35W<4H2V+jd)gv?c1C&wJT1&w^k_cfVpCxOinx{gp)ggvbW$G;*%TdTjkjQD6`8
zBvW99ZT`LcSOg=!Z0VsvKWI)mV$K_Qw9fWov%mnBXgGusBFi*tj7w)ys`8iJzHyrr
z`kABlEy_*EVa+<e@2FnVY&ZJlGVqtty$<qzcA9Yg$=)^num6taUtQ<XZ$yNpYXNlv
z2jGPf`=8q4XAc0vunW1ftLf5wUk+qG7D^okUPc>@20e}hjRNrBJd@X}^Np$PwGxfp
za-1Nx3P1*pQeKfI6cSr{Sp%!jB(AZzrXX$JAN?!#!Bi+4O1!#RESMo^h_SBdq)A^M
z%9<q%Y}ifzJU@LX>(U^Gr`0aH><=z51xefFYuCCJ1A{sK3w-|vzU0WDoqq!Ns?de<
z6^l)^eNp8~Zem3a1ijYM(s*g6mW{b(N=v93vtcxZF()|If)10HRe_1Ku-CH_1_aoq
zFd8B`u*w>oN`=VhqEI2t6CNn-jfCR-1Q7({0E?v<TK4m#=2tjzb@Nhy<^c(u`^ue`
zOqyk0cq{$&diukG4?_5M_J9M@9NPRwr82uQiG-z|ZL&dB<lUOxWR{T95&U3T8cdum
zM+DGEZH)s{DX)uKiOXMtO{my_0hdq5nz>Y+^RqwT`p;9|lT^_L5YU-ySKTH>L`1~-
z5o0cw!=_Q+XNQ1L;Fy<3tL_T0o9P0bWUqTu)qAaV?+e?~uEgCOhgZeV>r;~_6-z|D
zXyBF8z5y1x)vw}+(8GZ0zrpe*8=3qm@?`TGINx+bGTrg=I0<C=zT7IzC~JIV_XKj;
z<d{6LK|!kQds&hPzWk0Muy}#oyqh!M0;UjBd`(uS8-8+B%JmYHn9ku0_@oL<@%1H1
zaMso7a++qYD~0;te-tQfO#BogrR!oheaAszDc~RVH%1iE)LJ_+=QNHFC3&W-K)*2r
zhLusxzGash+4>QL^Fv2UUzCv~2<2wKTtM!4TIH4z$3=EOtK=Qz^f6*4oRA3{geAVH
zpLC)<KaKbWE+0d_ZLky{=?lC%3|vU*INT;=K@RDTi2iT-j*ErC`+=YB%$?J%Y!3xx
z{VMo&G&<t93))D_%rJhZd;>OC!S<d)CAdxSS*KluSExrtn#TS)b-4%kQOB<PI%PCw
zrqz6Do+6~P;O`@g_@fa)b+TuA-TI*v30~b-IsajpYj875$1}O)x^Mcf_6pxw$$}H=
z9$?CV9qH8!9_Ot=x{c8m>U!ky>Nfe$Luw1imy=#CV=yFB4x<T_x(Q@Z00q*l)Gi1P
zX+65_c?s}&Slz5v23kIYHUrbvS_{QAaT;ZTZX+ly3M3LlpD73|@{2RzpVS20zl8s+
zAAery_<THtwRciwSJo#qugnY>>G&uHp&$5uRUW|Fz3M*^1kLS{S;tQ5R;T7dWEc_<
zq4<}oBw8-pvJEv1G4BI2e>s$}A5p+tB8dN?mRWQMjcTgVmUPi|MPtxxf>TzC=k4zz
z)*YL}yr#|S-K?5L=wVz|XderR>8_`(Z4)s@*+f9x+2;Q9D-OLa7+SNj|7a&KaU*S^
zOzG~Qyz=qXl#Pz`V^%~dxwL{=+7Gd(H9#FkCI1lEm~QkhgcL=DS?D%?`G6|9G_Bn=
zAfuu!?xx>Pd{??yf(qQ+hqnnpUXZSv&dEkKlX!v8*<poTL{-)F+ALm`Yt%|zt>RL#
zWK~>*Fw1j>{*RYHCB;#j6;ooeWqdMYb`YU)W!hrJYmOF%-ZfwSingDGcv&|;e87kK
z|9jyaZ&`2mfBlbQ>J?t#dpK#poqO7T8>HN_miPRZCbsR5pt9pKAW;YYx!h}Xn*ANo
z-4fI>>80ZZxh%<KCSxz#iqnmCj(p%70zx0M<fE#IR%o#3aI0b^1W!g3jWyY`3h34y
zYl)(CO<xxY7{ryvKjWn!j>UGDXS0#0R}ak&e8_sgwLr|LARqs<@}$h;@L*;d+3?Xr
zEEZLIKD^_aH|!V9j~a4moK#)^Fe!msBLG<k%<9UMU>8tuR0mv!lPLZL4eXfwAz~u3
zI}>$D#>a(UosI$8jd3K797J-Us?v;>D>|f|Pf)|Vk8E6$H{J)WnZtbx%O6X|9FtRW
zju^ABJ@!ibxVNoiJ3YTO^}4tyI&;Eaq+3z_LOxpiQ%g`K5|QPkeyi*5f|X;zggG?S
zmGRxXpH$0$b{Oo``<!HcYjWe56eQSSxbgheXtXAmF1e&x5y6Ef+c|U%xNiOcDP+I5
z`oi{J`QgVN@)z}w(D<~i(&*Ey(@TKuReZH`)oKX8^g~iqCeQupt_{xrBx<}8h`>`5
zMsqZ*a(g5#a-Z0~Z~d><=gQLK5p-nedq_<v;8_{k@oEkL^Fq)Fzw>AtZgeVid#^tI
zDaKW(cb(urgvZ1xs|3=o{uN0a9DR1+-6=neI?YmK>(Jz1#EEN^zix7_n?Fa0rrScL
zeU6u7`g{_fX|-aap^wA;>-q~vY^E~@7)NEQX);hMW0(tf&0THF*r2Dc`oWq2I{pLa
zVpg$1LZ;|)rgTLwJIiShhbQg>LP$S_X%SJDhL^F#xW%=j#{fd|N%qg9GOWJhI5G+r
z8K6K7>+pw4sF2g#!>m~&`&_(Tb^R5aW~f+M9bikz$Mon9bS#{cyzQ*W!t<!%(rj-J
z6s;*8NWiF<tT=hxZ!;UZ!;4*)Lijs06uBDjCV2Ow8ls>cEay`3(8-28t+@cRZuhOC
z9<LA*Ra#Q5cotxp{P^#Z-N#0+Acj;(HmmTJxJ&CeYnKZIqTyr#w+y3hqs*)MLg0sJ
zs2NML@ulC%*Hpwmz}rFm&)dOW7kwzlB|ar5kj_ULJ0&*@m_wcme@%w=4J#`ylDy&8
zl+kcQ5BV|I@OlF|$!{O<9hDaC20cTAIP3XccV(9A%p#<;dtTM1$x6a>r>CbYeAq2(
zfNeH^{X_Ja-R^?X24&v!3I1zo*?R#7q7SXR4^&~pk#lLb3Ju2hgc#rt-{J~IKaj`=
zXSq8g!RSW;c4(J6c-+;&m<&kN9SvQK&OU}SI<<2Wb?dn`ze%tq?>}$SQjK;M^Mz$F
z79Oss(<V2x^Zm^d5@B5YNBp_?_bXY<%wycxGI$=hGsV)L!o@DKRm%}-jw^@j1}36E
zlCH6L$&ZhEinE<jsyqeHwlqubikbC?Jb9#y^-*LHZ$w<gEJEN>nVYUMqT*9KX%R<}
zMHTSb_DNJk;vsRZ=(xi=2zPbup`xGnH_m2%9D%2oC{=EFLrY{I=FhB7RI2st9{nXH
z#&M=1vy`%;mprBwsys|la;LoQS#5n+%yeyP!kpH6<j!>TJ}@ykxZeC489V%$L;pH-
z?*TELxI|w1{+C048uFgx6-A!hJ+jO7*wi9S04VOY?3uKg%5-@#eJ@d0`b@O2UAX=T
z7#*Zk8@f#XEw^uuWq28WguZOAyWX6$q_w<Uht*dsY$(4bK6mwboYaksS&2O8R53}b
zH@!0TTVZd@1>(RJ*VossYc3-;ZV}{<TS$Li@m%0%V1ByN^ruWIK-Teb2u}3_q?pSt
zINc$hK5{zr!Dd}f3i1=Ms%D$mWPg^4zP&;xWe$;QU7q3V&a5y%^BnbNn?o-mwpq(r
zm@d3NhRuY-%=<1nLb8h;M!NED%zSStz!5vc&C%ByPYa;$=6)1eF7pb%7n2aEK&8$<
zKG>CM<S1{JV&i;gUEMI^5Fd;9_L7$73Wo86BMK(^S2)a3E_c)YyW(w_>QBukhH4jJ
zdp{YmY?$F~cl)wgGn7aw$bsB@gq)$9bk0y>m=KK{VHoL5Ps!MrpT%OZR85e5Ingg(
zgN?iVftvi9g-xRBqI9f0Idg3~TiH>|ZKl|nB2$k8_MOrK%^$huXzA$wqg?rOX7?<v
z>9E2C_h9jxhB=+CCFt6$-sTFg%TQKM&H`v*`187yb`S!xQThMCcE_UpRb)0EyO9+)
zO@V76t>~}gHb(Q<=>%B%<<95s*Hw=uD}Fg3^|bx=_jR@)MeLuys`l!V3;Z`gK4y_-
z@Ug9}O&9;T;-#nSAvmt^u><5!Q}9l=v_J;`hm1yfaN!$G#(xFRDp<I6W_{t>y|A8j
zX5t7Prec!qGe2%O;3(l?;bCrVI6)v4lg|=j>e#S0R}%qNc-&d*n3%91Zuw|N#w{B%
za*X>811&C$)d)vqDKvlAQO6gMgtbQDGGtDf&hY@_Sl}(>iN{FnWWVVnhz;&myjF*B
zm|F``RG$|Ug`xkS+d`wQr*jzRtS&(ze>79!XMy!LyGn1<6jfsnU8!&%lGQIE7V=pP
z;!F(|$tl($1aImUX4+~}6fjee)09bMf;DbG1&bN6$`iKfeDBjJ<<>@YMr#>caD@{F
zJD{Q=5DwSveP1;h==zG_uV{a0u3=DmV@TfI%>LI5fpGFGC4Ea1|Kfms#kIjZ=v}&w
z$KG)tJ$7MND(|~^quKGotMWyHL5Ccu=JB0Y?4=+ib-;6@l0Qejf?>&%*j)vVCEWqf
zI33Rr(f;vKnc&PNy86m5D+o^>)gk|%N3FMG<qva$W2@N`g?~pCHRJX}T@Oi9KRhi2
zgS6i2Z&q@4Kl-6>1l$^MXw`{@A(j+2T21Dboa*hGt5VSTtmI7i`ka$4hsMfv-u{Uw
zvelY@<i+Asuy-+3UuSa!?GNqO<Q-BaiPRlcnGA$`pS3~98QbufbkmtCfV`UG44#qT
zP#FFBa*fMgxVP!t3k)@#*a#>%z2E87oL|h8%r2~gOjHh;WtQ;o$r&qK_Z&OQk)sJ3
zzgX7f*zc#i)P-g!jYPFm%*&%0Yb#haN7eJk0}z3(q(d`nJv&o<xL)6|Gm9gFa!W(m
z8&vc1dIFHfuaCTcN1#Is;JK%gCwB8yN82Rgo4P%A3CVQgs8iQmIb<He9naZdPXhs)
zpI6z~Yvt@ll7%@TF|8W4wvAtF5V%TlpniixZ%e8geq0fF<V<=vPRfK$@Hx-x^|3qT
zf@1}>!W}H}^dip*U#6obvzUHo(;N(>8W~D!Hs9s-;D)1t;Ta7k{PP@`J+}5Yj2%k-
zPnfRcC3EgtZN5$TGd;Q>=1$c7cC*kXB-mi9h5SP{G;*B0u2_NO`GCKgt==S2N2}CH
zz^r-eAn6<3QgENGIl&$_*)3*@-j`%KegaI6os&3xZ8)Q!b5mzg_)Nd*2bLw7S*(RB
zNY_n?T|p#v;kW%;@U75Ca7ETat$y}|JCMu-#1IUx^O#U@dk^%%;Ucrdn1)<gOGx|)
z)@`!gE{GuO2e#*MIp#Kai?Q#0aijv{!}HSQ0XP7*!Fsi?t-@o!#M+>fBt{m%X2)>m
zyU*nR=gwM9C9N-tGsuRn_FZ$v*0B1hi<vF$M|B;fEHTA8nWYY%Hn;82rheDJwIgIl
z%#m`Ma7Hn*RrWqZSmmy!2N%uF)6G*v-4HSIRjPM|CKmXa2m_Mnk~vov4Cb^pNE2=i
zMEd*L5n9W5k$repohj95IJ7%pG1N8KV?o67mP_sx{>878LX?P+xe$#v;okN;x1nwe
zBzWXur=^|Eel6{|E|1UYM?z3;&5_GvH)g|9LZ`<+YvW@*#}7>lD|EklZ}wQ_xF4a*
zY;QQ{efB2&1+M~2?e+y(vK(5`bbJitMTvF<BU6&rYR5eF)L`&2H;R9@uUas0g=%zf
z6r45!Y=vYznUMUF|HYP=yR|lM(F5n81AgvT5l!mtUwJ&nhxY{E=`~Pp*S+$}aF~pQ
zv!>KO;L{}}3&y%ly6Uc5d@A=>rRK3&7TNIQE4p)`a;2*M+IZe7eK`?jRJ-iMb~$?_
zG6jKlrw=LwN`G@GS9*NHUWj-Dq32^EGwj;3nPWB>=#Z57aK4o~dU4^*k)I13eZ1lB
zm2qaM<Z!J#oBr<@G%H!J8d_7XV;2pF?iB2(m<IYQ`K~4$Dhr840tHpoo^&w>`i-58
zEMd8$9U{;YKAHTcUKOcmW`&6!!UHSoG4L_ORAqI(g$`?YK7+{!r1;1iimV2NTk$m|
zYeMbLxtAVxKpb{V6k8_VgDCZ?8YU^#N9*ZA4JN6lHHlnn%jSKof@|%VGD>b?jm8@8
zm8u+a<n~KwL;oI1E&XoIgpu{KjkGIFaOw9oLlUn@>^IKzwPv}qCmwk<yG&jbd3H^g
zhEuf>jn?e_dfRT+N@Nd#-<wx&6s^_%^VdO;eOk+S{|n!#_MxMtg=FF8sU?`)3ORRq
zNS#NwsqLCu{oyPSTE+|gMd(4z(7{w8Mo)nbbzwqG@rT?d1mVH<v^*~DFP}}|f;oev
z06;bx;2q}}Ip|ah;T^ln@DWAU@Szf4O)PXZ{DocEQXS;@aikgoRvDc!`6LgNigtaZ
zIbLFLUHNh<JoyDOk6Hz?)_1#Ts9gP-Jf8b|=qOhNT&ek4=*hC~1)sG=mOJ)FZ;C%;
zPU?)OmV->dn`<45r%~q}@uJBn8>qNgiGopm`dz-)Q!P$Lvn5`?nOgEceMQl3utX0>
zBMb$m<ll^*#<bIykEpDVhtJ+S!&eiU3tfJ<JHQ;<2bikjRf?oc06W#kzqKtj&SAu0
zGc#jcg75#fEfpVn4ZRu;o(*bIal(g`X=Hx?n7D{7z*sfr;~4+)2_^)cVY-=_-6Uf|
z5>~&mSE10e-A&KqTrqHoz1@$)UKmf$B7P9eT7ZQwx1^EDXCxxwL6qnE*S=*YdRvuM
zGrer9KKgq|lciuKJ6HDAEovKqIXmqtD>`e6RYc9VAs@HOf9%8x;`&BVeFchPLig<D
z;ou~(k8mu_iR{@B5@FNpmGDDQ&s&7hFA>6(g1csQ?2{2G3HpM4Bl|x0W~lT$J9xCa
z&4>+2Ofsvi8F=c<JI>Y7Lq1M@E4l%l=OK5`ZaNZ|_27q$bEhkx29yLjo*7kEmmP^U
zzw<5vKGzS}*w|BbX3}fZw~CFGx;UFmN|Lv$<bB^Q&PtZf);qEi+7)y7_g}CQ)xN^=
z{kn=PSx+)qo7Rlaas0?6(ms|fkhSlRkrz2SO}o{3fsD$2ATZ?PaxqicQb+5U;U#BX
z29Az1dFdv0x70UJobf*Hr}S=r3X9zGr@~6~{{d7|l<3A)RiaTPB_**2+`IohUVO{a
z^%r}g#wZ>!Z6huu@6n=J$+cSZR}yud)g;+FLua$&Il5%|e?vOvAI;5ZQ*#tJOk9vb
z18M77CMo476B4M{q4w_2ibN{l&SQ+Dx2k>Fm~dH46^TvRBTT3bTOobua6fv-1j3c|
z1?xP7`)%{`sHc#lQk;GVvY8Z>F{9eET$@<v{~@~-F8-uGT}KrQ?KNg*ns;d9H2E=7
zZ#=>nKIEwc9=Jjc*G3q{)643!DD92J^=25(w6l}s@(fc%)Xtu74$LIB#=BTiEP`h0
zcs%MZ8z8AokKm~^2jBOPet)XX8G(Zv>ksS^AVhQ>sJ!a55BEkpk>yB0y|fC|@T<Im
zj0<W$e=Yy<9@=f0OCb^wKY3#j<R7=1y?F)B>(S_7Li4s|eZ@oo=7reecPiI>&St<Z
zY34#jT#wbn5dKT$Q1v=;sd_W#+1cVEBcBN?lWVyOFuahhGQ~H`GS`5UAXk~4Yj+4)
z;w#nkR5YqZuGh*fF}F2$FxVJpL-|35ea%03&r`Rgn4j>>78NeyFe-GZP@IhZ3a)4g
zE&33a2BE*O+>prheu;fC-e2&MG2QzTnT^-#(d2$OQ$p|f?_#_4to>Nt19SC}U~V07
zr4sNR*)`<6Vhtm!PCc<OaER%5x_Zs_eeojCkq>9*@dZj7@zj^#NtI^p`w7C|JH2E}
z-vmtDz{FOgr}}1=nah38V!OSDJe_hvTMl=X=ZJ#OOYNgyQ7?~8a^@${^R(NBQYWx+
z*QjBh3C!78P3Q5fDlVU;lj<j{K=KNnt1jZ76+$OOOTl*qHDdQ-OLcPVLuj7`){5Uf
z<9gZuHa8x%gV^<Wq(^oC<<thB4<s-(3`%19HK)5Z5lW?s%ytCe+dpTTjgEj~w@a?N
zSfRZ~;)Wyn*$*zm<kGYMwA$*3l9(FN)ZXG>@+}6mJalbwFNA%wT&>Wcli1^9c+2M$
z=THK^owpyDID6<r;b-14Ce!KeCbqx7xcl+Jba4mypT%cE#2}8}jV2=_BM<<Z8Ja2(
zpYHH-a$Br>_tMuzuGshe8Y!D4D?AR$*-zTC8FznuoR#XBs1^%+o%nGjr%4*FB4EO-
z!&@aowma(z$lUA8RX(IDzAzmbQDKsJm@MhiyEuECl*iR)t;~NgJ$yTQVfSDXedKXn
z!^%{SKIX%sxtq?kQ3tE_;!aGK`$%&CbWI?v%enF(V~Lc5UkKrOj+OrScDQ)+mfh)z
zM%b(zKkY>O@KJFhx<^A&@huh>R_ni==!q#m&F(tvYUk~K%jp3~yhxnDdFMO1G-k;r
z+j-|nq1!Eh9OucRB^-CzUyC+U8hDYCH0v)HvQ&}q^%ZsAX9FU2K51KJA2&0&a%W$;
zZJ07x<l{gg$gA?i<rDPuntlj~Uf{{+7~CLU@(TEl{3|v@NBg_AZtPOG5I2%3LtoZI
zXFao8+YV#xU&F>_rs7_Hwr_6232O-kap_3wsbZiOk_({|BW`h3^-b(Zu4NC2SRH8G
z;_OD$xmORrch<U^l7pk9(<jqX(X!kQmte<ag>q2`xSMla5e<vZz$5(=9y^@^(c30v
z<@?Wl=?=vYrt+fJDVM0Qw+>-8d0a5Xn|HiWn`)=MlCxAj;YDTK<Vx(_LNv|II@6?r
zyp}JNw8BA%-e05oEvLrLTWEydlj`j&4YuA9Ux5F5$2@x^A;(mxmzkFDR=4dF;V@{9
zH|n^V{2ERtEhW&wj?XrNAs2E+CgO426-pW{lMcz=MU~d09#KB~$!FG8)EPnvnO1+k
z^tH(di_Es>lGPaSGcZIU9ILxOZE1pZ6%FhUF%Jj4L(Gun&!Ykx-5@mwfZ|))MmqEG
ziv?J^3Dp0bN<`>)4&TO+M;cT%UJSz9_gM_B{eJ0=jw`BRGa(Ugt9me<SlykNW%IpL
zYV|n@_?W#4x$VMETpO|FLv$nw4H`h_p`fQ9Yy7F0$@%MxtG(HJtIPXWN#DCMq07N{
zCH1reKEGP2JVnU-Z#>Y5I7h#D^=Vf|nw4sFGBRp*3^BU$6d;yOx#H8Exbu8<WuRL%
zfjjqN#nOD#o$a@7PU#Qraw*ARv8*HIlp9L9QcFaAN2H(2wT3$6<IiA`ColkRF{i{_
zrB3$hfYJOqkm8WR>hB&XoLN;5R6N2lsmWDcUHLjy>f{@<O_?f7q7k;Sh8bn`QOAq3
zIt=K4>BCcZRNFJFkM+2#`Pj(G)Pc6iJK_{^)kV9n&*4~$PQ8A&)}*$!#mY3)!8DZf
zu&*zo#&&?ezudvoPV7dQ%Pzr)yGMpJU^fgfen)~@6=w+GSjs>v7iMGC&MR2Pl{PuV
zN)9mF%k}4tye~Z6Wbv!r>!HYYImaihN97e4zgc?HMOq%+go2*^)e!i-Hehk+jSG%_
zn5K)KGH4B(U@8UgZ()8B14~c0OXIwuzJG$2uGXDkoMj?h8NZ^0E9$Bo;aE;ztPhO#
z)6MW&JjO(_8O8PKn32ovKN|M0F2yKmgACBO{qt6aNi=%uZ+Tf9f%6^%-`|6&`mung
z`o8Y1u;C-FQK0WFZ?|_BdG=rX5<7XT2c}`|20u8*e|wHGy3fgcJwS12cD)ax@5=@`
zF<h#5q|BK~2ffvx%fk^WOQ*4}Iw1VBDc=KH$jtUf>)0Lt9!y7m`^JV!DeV7*u(nXn
zh=NZa3EW~PgOOO4?~bkV0xZkFV_ec_oIYLFg00anUOd7j?7yOE77S?)n_3u;RDIE?
zW@c>8N8E_VRr@SpJZ}wY{`Lnt-tAC(7Y`qT>zwU2VuAe4#nBkY7Ed44VleNsW()Oh
zf$#INnRb)+wys;}BnW&uXglji3U@`rBpkbA!!<+8_9owS^T<uk!8H3MZQg>O0)uN+
zS@*TYdw1JYqxt#cxu^KgH&9=-9Soi~6}y6lrFV`P(es{IxW@QQa5xtTU%)=6&fPRB
zu@jxm;{QCduUc}Bf5$IZCYVJM6^7>YRZavim|W^loJV5zWbx&(EcoQ~8!xog4BZ89
zT;Kiq%`#m$#OM9grIe7J8G@%UISd?`W_YR2Sif#gS10oHp856#+dd)mJ1gBb|1+K<
zwZ_xH)STZ3Hwo0ecvJr|g1<YYP2#Crsjw){X7Q%bOvyRZ)AD)FlYbf$7$eje`6lCN
zMti&PH7rbXj19-v`i~m{J48WWVBfQo^{Pr926Lcwk=c7?Jo<Nqu=YO$!@R2kH>PZS
zvp6*9&~nF>^v?N#&i+5|o$=N*%dTYVm<q#sdT`CgLg(aB+v>CTlKRo!93pgxJZ$dT
zkK3mua`cyKwLYJ|&t8f*u<a@EE>`LBsM!qOxhe_Y{V4lk23l-YH9_mFCo}9g*VXqu
zEH2@49{Wxgp=;1qm$ItkzznPXoJCyn<?DbY#rM>cW*$UDs!C2X%+w_K3T2h_Hmg9U
zsECslkRvWvRazyj*)>7Mpsg4Csvbv!u}dIvOMPm^@wzAcFR;&v?zwPk882$wynjhy
zjdC3$<tpW8u$hYw40-fztZQz|%M4YxuU|8-QlHP=L=-1EcyFzHZk_R{N0bt?WrQ6Q
za>C90)p(Twjv1T`X`WpnT{{|#abLPTHg-YOFSgg^I&}{gf!s=Tg#5%$%~1F$K$t3J
zS;Zb>V4`br7r?#wwn3dL-(hLwj!f_I`HILnH?PbYE#!2<ALG%jzr6dRe}9rPE%9P<
zzhojyrLXNTPh;d@u-|>Y-z|eehpoGNBXaz5dAl1_k<(#@phj>8$?NtcDpUpz+&*^~
z$chQ}dy%2>i2#NGtz3a~tn;pHll1G$?<zkXXh*IBuWxTVi?Y=`vLmz6?2SlWH$d-d
z5mpW#_nsd#dNPzp@>pW*ROpP~>yg!O`t&t<M_f_SfyCQ@PHR7sVv8T!*KO}>sWz(Y
zFSb1b&pIPi+Po#@S2#&;mB!bYv(@Y^0<4~`_K$(rt<`9)^7n9dq_Gr$zPorKF&y&5
z$Nw%N6Q}X$Bwy~blBXcySr>QEJV^7ZBS*u&cOFoNld%<7Rb7$B{AJ=RT#0IhU@xDO
zidBb3rx+}j-<7=@-SAh$NL$y|Pru{4ww<&k-<!CZ!H|dD?qNsPdt|S2u(;PI1G49N
zf`jkK)4oKJOj%%`+N4SQaP;Bk^;s)S&59yFJ!Qg=TlKP1ZEmXs?RP9eZ&-H24X)e&
zX`J&dhMSP{<gg3=#%8}axn3&C*ym4oXz5PXg0fA(!Xp+u`iB?jC-s#NP%Qq%{M}}B
z!p?F2R#Hx2#E`K!IDX!Ck8cuO^VQoa23cFduObH{Iwjgaw~tlZjVwOii7bp*Uk3pn
zB}}JzhkufZSqL`hNeGqP{{p7_=|-<ouLWaX7Qa^mJ!&WS1<2#`-3T5_si5HEnqRX5
zx*xS0trPD!d>}85pW-`^3RhC%-&bC0)bb0PuVhIoTA=KHFfy56bUP*IE4M6}<@zqL
zz<O5XSW^m|{EC?$afx=~N?tgA6UKd^oW3F@o0v5rxkPaA4c|n4FT%DK?nR=oAqW>#
znYT;Y_OHu3f6rmx?+SYV%X{Y>Kk=Y57$rO<E^Y42e#16Zc;IfWGH71E(aa}R$DPXc
zPwi2x*xJBb5qn)#AZN@`y;`5JLaSyN$qc&P{ubZtpcJ{p_8a9oP*Da{PcApd>NHqx
z&ubYC1;h)5v3z76-cMpOb|bbr3e+~)uBD>)HriwL8=-GWNkS#(^^2Xi2}#jJ)|2nG
z!e>FMP{)1u$5}XQVyuQDWv+6VT~C>>rAjoCsTAg)w`DisDv>w>&mq!8KaA!^#?_m?
zI30V-kql6jLyms?Ry<S9BN7fcN&sh6=d8225r=0KmG;B<m(y~8gaJHb#9V*0=3i&A
zY<l)@bBGyclX8wH@0^x3Q21{f-g)t+&GT&*2AY_kxaND(c$rn4`|YfaSYGG3VR+8t
zfhIEwjB71O2RAqM@LOwIv<XK<;Ou{)hsSqWaoh}54;&eA50A$V=UX8`fcR5(rs;Wx
zw2@wB74Hr^w#I6xevxTnI==^fb@#n(!w>V9=l@gQTmD7ah3&!u0!oTVmx3bFAzdmU
zC?FvWog<(y3^AmXv`UFemx!Wt58YjZfHVv-z#vEtHBx(V-`;!gXa523mv?<4W`66s
zR-Ad9$9eL6wn1J2zexL3#~|k|L&uKHn^cl>&5lr+5mc?13w%<ky<ylYh1_L#G%0%B
zwe8WLl<>*j1}k(j!E??0N=!w=ZLP|jKh7=>U_(@!JY`ugnB?CJ93kfmyg9f+3-tX3
z3;b>f<{@H}_zr-B9*l~B%C4>S&Ndf~MYoZ&v5Pq=n(c`wH12ACr2PJD9(!c$q_X%^
zr_Rg8ZDX2ADV$1NtmfscEBWe3!P0TPDBWiLS>OvDHfh%{M!Q?Ut4pZ_p@PhNfrN(+
z3(w?ETY0B<es?BTI1^umJW+ip4S8sD>fQ2{Ap7eFr}o~I;*JoVz<K!q6wHJoqHKn7
zYYWHtiB6D!&yaY!CT9bx=ag2K%M)h#NH#FdA+>1!nB7X>K=@9#twfq5VcQiPMs&!s
zT`<s$r=_n=$p*VQbV`1k@#Xr|hJRn9S)3Q!+m9EjjyZoSDoc;Gpevd~KIU@=Ph5wm
zs9cf=7CKdu;X6KbAE&B=m$G!(JbeTE#`daxz{a_2Uw@KgQ?PojVp~zm=|1UW{lNAf
zIS6lj!5VUlr1c1QJRa5~VT=?Ae;1z~c(`<Cxmu$vci3sZ>tiQ(npWEs(ndrg7i2?@
z{cNC)J^E#u9I_xT;=SNSJKCQrJv<%W`{)o`>}cYd8r1dcAu!q1{ZeM!e(GB`Xv2qg
zexU`vaSJ+2#S2G|L6s{L>bYoQKZLXB`beHCUZbL(X&BBz+pHMpeenEPxh(@Mqz7AG
z`gb?gAGL`Mn?Jle_5Qg0`%gw9<C#z*R79l#)|b)ksr@&x$+_O%@T@=_r-*ar!>pY>
z7qb(-IxdF1HDvfDMzF<AWK{_k>|qDtjqu#qo_GNV-gph|vD(TP3@~(EHB(IhXv&`Q
zwr1uHsj%ZRY}#kmk5SZyIzHdiKm8)oO7w8D<^qqwl(%I6iBxOWg?taOUUuH}C3LtI
zZ6z{3_qX7UEl<lBqwh`YpCL;f8-6iucpaOfB+G`doVlHfS0oF|KR-S*%x|L_9JRlF
zJJDWDL2_83=hfg@L;eQT;ZCa%^5>rjqZN4ao|6-07&)KUdT!d9eyg;i<J0|=9@#Sd
z$ib28ktp`hM-Nn{?r5%(BK=fEH^19fyri%irZh^-*rtI!|JX#581gIEc7q-JCvY>o
z#uk>c#43Z87Ljw8TqNn_?ITr8f6J#Z8?886lVXm?t^?t9|LkZ}>F}+ErgoxnE_a7O
zbTd?oN;F^@SYzP;{s`!PHtvDMgT`7#*Dtm6&CGV+<H2^*C{=g&eoHIOWF1(0&LDNT
z^>VSjKZMiuWv+QHZ%>GQxJ!u9Jbu{@(%nMy9m44l0l{0_xY*1@)Mr6Ez2QZD{u~Hs
zf>&i#l_h3kKjxEi=B44D8n;feO|Ec}VisTY1w@9nV2(QhZ|T*Jr%m3mT|e}dpP1#p
zjNmt+N)v1#EUG`PnSgnp?2U18akavp*$8(oQE%2Cup3uC54z!L7QBHKS2pyLS{5)5
zlsq$iQ;TMcq2cg<t)B-;O0syUn<`rwl{qwd2cGzhH}}}t++_MOty_{e>|^pGiPA0J
z7P9eA_6T}b6PaE;PZO@-@*cZpkAC^VV()-3_2%NV_XH*n?&c@2r*=}F%62YThdD;{
zTa%Q1$$Tu9{_#KuYj&zcl6UkO71rKgrclyri2!vMc}?#IV_v8)rAlLc`{E5OLaz@*
zjD>e{moGjb#ro$9E>0u084YEyFL0(=$MlMAMPiP1km^Y^r64@8V#3(8^esYgs(;_4
zOj`UJWZQ*Sa3Hv`lic>in}dPD&R7F*6~7khh?K0Xa583S4U-E7j06|iWIdF>UY&31
zQ7rbtJeihpy#x31`abc+I9KnE901@Qu2;_XohU!C1DbZ$+F`jwP{@*}lehfo*U*u%
z&pkL9df2>vUheBuUMVrhzEwq<N<}xik&pdiJN>MpBr(mlB^CK@$jgE!C4YWrwP(4E
zX!-qceI_X<^^580NqpKi%UQRmWHb(6Wf?3sKW_bJYI3{l((_+C)W1ugRy?~Vqvv3Q
z#>&;GJ9s|K1N?m@uT%d>SwyoyEPS_Bo-&PPsZy84#x=8IJNG!Yq`3QLUu{UvM9IR}
zf~nSxvz|){BSv9z;?Li`;Ctp(>T>{RjWoJkY$Kn4z0YH0A=W6j6^6y^BUOrX@45%`
zlBhkZwJfXO?TeU}Pz&(+JOj}T^Nl4C$jlTLe*D^{fh2%F^stkS#8f>%0<i@<%_1C-
zzEsxNf=_zUUKDCplPvoI70|khs@Ra;Ireo-_kAKq3Lz)Sxw30#J5WLx<5)~_RCRK9
z@0ILuKfcI)Idp?<Sa${w*Eea|(XS_N#a2M$aCJScJJcKF3TwVXu&^*W#Tfjb?piCC
z4}6jnu7S$;c#<y<#H{CL|A`&(2_Fa*n+>C62?IdNB<+ON1oBQcfr$B2zJca5ctg8h
zLmZC|VF?QaleV4bG9~op(ETe;E^Rg#7Q37wjS{`qO@6LgJFc%9DoLNwkA9%*+g{sQ
zJ4<kWcYnG>kNoZH)JqhYhy1Le2Mqq{w=V23)C-Z!<{wT>D>kkg7HDSkHjCG$F|;Yp
zysgCg*&_!_ccNHyRE0P&GhBSLrc!6sE)@N6m@^%!aE|I0wqhB2grR6WGTs#loQd6H
z&43{k>G#=V@Ri4*c2kOxK@;V=Z@O~s@bZ#@G)GG(M*s|2>hWu+y*!tjn|r=<_V(O1
zHq2_TXK4jpM^^ZJmNE8L1auQ!Z&>A=nG;rXc6ubTe=r*x*>cHYhzL^RTD<Dsf9;ep
zb)CMEfZnd8gtwn8xZXvkuZ09@OL9s?OjJ}aatZS}Fm(Yd731jAVDx>$8Dy*=RCZrl
z7v2Y34n4Iji5kWy_1a68os67~y~iQOtqr^G@%01{#4f?51q^PlEJSE9mWo`1I(G|4
zHnL1?)Gl~>?z8gEuu4ohq#-V2mVd%{`y5F;O_pm9tcb6>Jefg>-%DaHuG${iWS&Z~
zt3k}=6Bbuf1PPR6J4#5E`3CE|9iv6IaD+e4vrQN6r+aE0#su|vET|YN?O#oy?gX9|
zL@Xrp399IQeXRE6Nyl#EZA_HpY}4%zW7en0yVu_GK$ls7aAp`YMNr!RbVhxM5pqp_
zdi^$@*~H+?fu}RKe41+cIJgR0@RlpgWjdFa)0Nk<YJYp7XzR2^sFV`p+h8n05>S}2
zz_uhr6Z$sm@jb?U&t3lCi{y<L6bnz;QVv{|ENo2|gts!R)Gj5Gciy6^AK38b`HiWo
z^dm6QUsD~upTfnDz7kvaO*`{}ExvUr!2%v64L3b$JUdcUe)!OOZL}a!c!@*qpxY2O
z9T)U4iX}@Wf<C+58DNS*+*IfX_SeDIVwG2<MHQ25%D(%Q1`(k|o*|Q==F6tS@gKf@
z%lUV)Sh~K4qkYYZX^f;k!y00w{kFZ&J>sksJV<`8OxU7Q`inToB=a{S-0gCS1;1^T
z@IT8xBjqt>rVo~dzp@EUjKS6Rp+yH^1{Kw8Be&OI)hY+xfF@nB;t=7Pk{X_wGyZZj
zUvy-8%BX+&*c2(nhsx7keox}uDnO)g_Vp?O<Mm{uF5Gv|k;@%5V=j(P%fWKIbB(>2
z;1I&#Y~tWq=$?1&#MrfLuEIK^@I|{o6i2`*9iTK7AfaZDyrmwWxNjp!^2?=P-|NpL
z_sVPo1uEZ!g-L~rhL*u;Qg_xn;3E3ZsuusW$CY_2`odrS#9`DN(dqI&$*UT%bq|21
z^*Iy&$aQA9(J2k>w+^M9_ab)rA`ILfaf&eo)5BY$fL*2d>KZD$-g#|8UUMm45#nA_
z6QOU8*Hms^_`V{^lj8XGtfTKBS~Nlnoo2(rH)WWCvV9XAK+M9CGU9P6MO*0gozjQ%
zJ(li{?nR&%#d|pn1v>i^Qcv6g-5XSEh@j7~oD8unBwE%o?%yRj`N8c6{%A*Lw>pa>
zEb5RRw1K;;@0N5;fM_k%#8Th>In_k&>neTQ%s|95xNI999}*=teMP8kD3pI(fw^(8
zI>5-2K#9jx<BIvXe2jKXf562@kqER4Dgx_jo6sjvkq^UARJG6)AGa!oSx=i9<6)i~
zW<0*X7v+nYTt$9E)dd3>N_C_3Jt@e&{X})G%1Ym1uNVy<EzZHPsZ|6c1tGXXyJ=TL
zV)E*^R}Jq_R$||$DlYR_aP0TY&PYHZqxi~2lE==4MF_ZJUrZ~VA8E3;6Q(n!Gg~u7
zyUN+B++hr**ImEY6Y<<^hEnM1Z|Am&ul*V=YlXE`7Jhql?>|s~a*w*?>!;=5yR~5=
zRR%XhSj5&T2d2Kpc!jvOHNl#*zuJ}%Dv5<5$Rz~jmtUb6nzq+-HpAHyu18<%0OBt!
zbMnWl(MMq}!ISc@8~R^nT9l<|E<vtCJWF~aM$GXO3B3VL(}vh~vnv$CkWJ%FHiAbv
zN88i8V<JK+x2RNGk~~h=OF7>M(D6Yi%xXQ_pK8g64k#Q^oEOb9AvtJh0>vFhuXA&g
zym=FDXT;6DuDTm9;01bvXp;Y4D{6!z9bK(Ln#sh%Y!pBQ*D=fV-jyqnMzs=&_5vyT
zwx<~%(tbW0pGrttmFrcM!(opFOUJZa7$lw<G<R6OCBLP6)e6V)$neqf5h*3)B?FJI
z*Zxdqz<BYD%4MKXiODAPLQ?1l3vik%wj(kFX}$633us$<IVo(%y=dI}0;DB(<}xxM
z)Li&7YEvDVTDtdh>ZAQ;L@--zOPUV$EK439cUe#XRSgw)A9`Y#pHJSxzfmkEY`bi1
zy1Okgjc=9ENUr#yr|Rm#YDd$s)5CX&8&K4I*Na$gA%k%GpE|M1c&tU=Pzp=DQu*>1
zVO5ryid=xVzUhyps|WZT75$y-#IBs_7`h((VM{d!HoE`XkL@S|mS7nR!_KomL6Qyl
zjZN~oXw^&DtbguW!)UiWLBWTyjFB1s3zl20ZwneUedZ(I8(~L_&o@lf*!-P=NZ+yR
zFX{;-zHp|#x%{G7q>t@n9%JLS(<WEpXX3Y(mQ$obs1VFUJ@~`DN*=g(_B$=*9Z3|<
zdhaP<(A)x;#$>W@0ECt4IEBo94<bc@`Y2K=)~50o!&`k~19N~~qxp7*uk~_2zHGug
z`PwGRj@-~13YheCYdl%<nH|L{-`ch=Ye<mjrbf`OBF2{oj--pIOH&fU5<`m_%S&9N
ztgNIP*R$jfMgo$4^zyFxPM=|){}%P+W^FC_9hA+Mr*WZ^S0*BmV<16)(S^@4spuEn
zuA)*X+~~OJ@%_ftU;MfstZuber@-<mx8ZJEzN>CQj;KLd4#<>CA*&1RdbD`4ea0q<
zxBt1{Fg|X03?jWNzVj+oHDhdvrQFT<wBN_4JjN8@aNdnzW$c^rm~z~>U9to3HH}QG
z2hsQ%a=%97aPgGNEMNRi)t6r4s-qqH6`%W#ZuVC~hvFP+?|)Ha|8nn1+rT!}(lMS;
zra{CKs*NRIJG8La>W~a`t2^87z<xd7v|E|Rx2v`1kld1q!l-bo!8she2+KX9l1Yd0
z#!b#!pDIL)$mwiJm@BP#H~y@-9B)?ol!dtX#!lC6pUqG6VB23@8BvF*7*X*a(5}^e
z@_D{fHr_RQlZHkSh;`+!kK_6-oE9~nPQq;c?C0*?{oduwafy0hazn~{=h^ViWRZp9
z@M(?f(&MN4HZ7T@6<0h?FI3tLb6`ISaFBoe{(RDusIr2j{jmf}5<`Ks7npJ8yL9vk
zHxBz?bERFP-_lSep_o#l@~bN<+39j<4DP5a+Keiq;LEh>YU`_he43rh(-#x4Np+@l
z5h)B-77I-%z@#Qp%KF#iSW`3VzveZ=9Np7{+o!4t;^^QLm&sK-WU!w1s*RV88DSBc
z)9~flq`4or1tusL$L)J(3+il5dFP|}#;XyxK_N(BM!+f4d8f}LTFDnfH?O%i)a$F$
z-D{m35q6&Y!2N7}wy_b<nlXXEW^O)pn%i)av>aU~;?d!338}1qc|uK%o@S#Ke|)d(
z532fRd+S1V*rCFk)QxLhZ!QAI!dr!+@X+792`ln>s(0D@4e%>;(;DZ*l*wUf&niac
z`+Z0q(_p%pH`rf{2&uQzN0~_yWC)y2>IE&i`$nfyO0Hz&TiFVy_DUkw<jVZSwdbye
z-vc<(AKy*oH&=xHXG{VDt_&mlP{2G=j(*Y{M6Bt~T3Lb~J2Lh?Q1<lQ`XOGmH)_(K
zFC74AzQ)V!sBQa`LxC;I3VItenEOxQ@$_tzdC#uF;M$baZat7u$A~l`y=lrZS55)C
z(RlCTT;wZ#;jRuy2EbTK^F^ssk%e6P!+6Ox?1k}L77|gnbKIx($VOuI)tL2!?5Mp5
zSiee}yFA$K^`%B>Y*$ELwJa_=(84u{<zt-UhenF&Ij0jn&(oBEErz~^b(GNTN%0}I
z37Q*{ViAt}eQ0`8q`E*g&SG~Vj=9+*Jv<3&P{S*s;?&L&a2yCAKl6d4?81(nsWb+k
z?h*VfGGSq3)^80qTeF;IaQYN*M%sV%N`^OXIf?ACe?S1UB%RBfobM)?JcU)?iP&A?
z0l&UxuK3zs{?uZGs&_QhPiBYr(sc*7WLQNNzka48{rM_LeNf#ZQw>S&MaWo2tV$X*
z#PFhUG&It2$ECKc3dV9W>K}&-AFEFf9m<G(<4;WfI&e?J8Wtb_NCEYktyib)C)0N`
znUsF-Z<!!pV)w_aZFHCe64fq&WIrKILp=}vHmP9w;D`31pQZako93|dI}_5Ao8Q=)
zGFZ5wtII|_`c)x_#>e+tb!ob1$--@}4^4)1I|?a}vHVG*%m?^VF&tLPCo(WK>He~%
z4QqmYUz#k#@HZ_<v39AL=r~W9-(HEQQ%F5rRJ(+rWz~$e=H%xrs-l>|mpFpJV1_)H
zRKDBiRa`*R+jOG5HN-s;<a65-QgT^uhTqa$Gpi+^zgnjGI4pY%h6rqo*?ZIG?}WDs
z`e58F`;~`99rR^8r|Bu|1}4ZGNvM9iZ%&ma;E2E(s(C5L9|olq3~yh(2RRX7f=Z+q
zJN{u$CTF;&)aK1)CBC{r?2N)KbWG?6G~4^?XOl_8J;mb!Dr5sq;A6+K>{$uf3*oc3
z=RuXdSX?A9r^OB&C(G~1xUIZOPq27R;4Ybo7-oL`+>?*WSfvlv^i@4!q5F2m-4qB@
z5BMO>68YCpFDkCCAgq)q7<W07)5eJBQJWq<^O75iS8SvPN*<a3^v4~s6s?=ox<C2f
zQ(9C^zj<!ePie117yEPsbyB-E@!lyICGVhL?TyT&!gjmR@<;}eEa<+jT=w%d7WB3p
zZ}%8thi=>())HzVw%BVrUWe~l)^`~jEejF1I*9&x>*)fS;Bv~s_>pEvURPCTWsc7a
z?=+c(E_V|z95|~SGs}s33s3sG5EuT*p*BxDA+T-lYej$bL)R<cw21irsu<#T#UEmD
zO;Gn@nvA<>obx%b*y!TmH!TiRw?P%|FVL-rrWZw^9Tak)XleyrUu#4t0{Fvde|-J+
ztraO?NF34fmS<D00T)kX`rZ=xX%x1}8hV(NT*tCslLME%Go`LHJuulgRUVfduO;}}
zvT4cbWOl<>n#84{;EPAik#s6$Cs`%Tko0P5Wj%^-U)WFwwJCdgY;%-r^vtR?w~JRa
z949+C*+>h)<jr8?(+bF6aelR8h>~NSc9Ag6>4Ar)X-yK`Z11!j2)vu${XHmr7EdhS
zI2}}}aCVvNVWe<e<j-tP7O}G!fgBKfR+&wCc|F&d@^_SdFlEVQZtbOMDH0DZ)#$tA
z3Ls?|j58MRlxNWE>`K=k{N_jZF@M#Z97#JR5;21`B*_X@mu}ouItz~N`h_ZGLmzRz
zk4kJ?_X=pn9@?y1fT*h=q+Q+<&E=iF1X12e7jZbf+e=Is$4CE0tEQgb&0Fbu82se)
zCI9nY?$&Hc8O4OG;yaba1&A2jf_)5*)1iNZ()M(+Mm|L&v0|!PU|g3B)$6ktbpN~Y
zXVLSdIKXYGWR}0R4>L$T+L0C;_u1J@5Oplr8>_R1){WwsK2nLs2+4{eQSD8Q;noax
z+)Gy~mv)ZvwK<M)^x5lGkz>5KI9n33@;^_Ozo1>fV738*DoHby0)4og_X!$jy0<zq
zSZt+4yJd?0%JueG`htF;FQHY1>}PEnPo$CB^b?s)a?94ch6eR)rJqseVg&mG)GfJ#
zB%@2H&779|{xQyXnwJBt&b%1jT_MUlsV_zaT-?%UL3~(i<Y+d>leWC{T{#v$px52P
zo%+ud$gN4VnXO)-V1Q@@knSrwCEw%yU21i&m9PJy!BO5B!Jbg4JK7Ka0}D7<s_qvH
zve_HEC_VOMN6c3iWBsFx9TbJj@Qigf$9w$z*n^n%BArj)@$C6#6@(&+^5G)G@pCw%
z5}p;`q;j`Yc7*%dC?#Og5COeEBhJG;l)2Yx=Y;w=@HdhAjaM;t0?m@M{tobOFZWjG
zfl1hgXB$vDK;&lq@x`tIh>v~&asH^EW^IwY!^L9DyV?g2#>pSv$RG1WFOPAZ1%Lb}
zcH5<6E!?RcJwKfjM(1MsIP*HNmE4b||MZm`l<IZueGgzte~!qQDtu2gE&Ose<i6Ua
z(w}oXG7MJQwWOLVGjA5jCw0YiOZ*48X$wVeMxq&@#txMm5i0L-3%?C3n?yhQeCx#>
zdiP8DR_;h{j`kh5N-ao78A_3D3LVQ0i`&MTh1)xBX&i+|%S7U=cn(|g=Mx)jJbsXM
zrdobAYjw4g!OpOrr{j#RRg@2hKF&`e9jA&xlWTtSx^dX%;!&-~x+n+Koi#6?Y>M>9
z^<1gX%;haA2?zH;;k^z>%YFo1BKV`{M#nT5yH2*RX=n51BjA+GT`qZi%f2K&$Ce=_
zW)<D;Hza*+BYj8=kJE7Jlxv^w7o#@jkKgU_&Qo!7Li<iO9=!GN<e1d3-;5#cpeE4$
zVCf#P=KPA`HDacrmS}(Sk*#;|<-IC8kCWLhXH>r{`*3-h%<sn83y@PP({1Mb$5DNb
z_ocDVwAfb#AFoJ^OMg>q!k>KHZ$G<{-_Q0sLWYBu`-ud(IK5zEjN%z<_%4NK(u3;P
zK|F@#7biVeVZ3O@d=S}dF?Axzo-UVpx%<5D7-Q1rXdF6`a=;}pb23dr$@~c<m7k9l
zvekI3+dq)=L4CVm5@a;eiGP(^pWty}KgP4wR;_!~7%PyE!-ThL#L9u#eV2$#To90M
z#^cvHztN)Z)dsYD(m)JK4e);x_;`AbQHI%5=954V-Ry=_oPCB_$}al$dK9({IGjER
zc6HbEv9oN4)ZjkD)(a`s+|2K?xtX^KOq2}ushiY0uw6}KxNNFca@3R#rTP**S}y=s
zNl0dM!B(kiHeokEGqY{E9fe=f5k|OQQmQUl<6hX4=16uxJkg__{`s9WS?~EfOq19(
zh5g@&p<d99OZlJ?o1$5GHc0WN2?~AkYQjgn{_^F^94O(;R&fP}UA8M62u1EH-2BcZ
zR6KOg6c9gX8oGf>l;j_UAezD5hB?Y_;Y&=O^9{q{Via+~C(tMyVkZ5V?6-bEOU>>r
zPn4-}Du?6-lV^uXiCW0aNeph*HJq;+DR93dq#1?v^vG-U-B)fsU8nhkZZAgtTAyJI
z!QuCXPGU=-6*(`-Yxn&QwiC{RB8|8#&9-^@1r;ovCAW#o=2})2rc>3`j*zXVd-t!G
za_5t0w4hH&X~fp~zG(O`SUB=VWH3}Ung9r}Q=#-oo(YRM7kQ>msI#fx!SHkU!4cHc
zat%7RR^8JCa<=j49`9F=-n>zBFO+m~QSFTv=s)-j?Tr3z(b!O(1`7*<HAdz8OMP5i
zxI_{W%4O-4wCS5BY?x5D;;(Tt22}IxaNvWgtEcFfR(D7t&1O0W*2JLVU~;lVkv&7`
zg#Bpv{a}tUOL^hYLSpX3km!)GR|iXW5I?|_ZNBj5dnsPCtbIvb$U}r3K%;)-);y)*
zTU7WhL$Sx{0L7z5ymm4&$3I?)0*>zIgNd{76c#hjmb?o}R>GbA1*XCx{ussAQ=4^7
zGLJ#95mPTHl<9(R6&h7mKHk2?M^pjITdb$6ToUiTxEjxI6a=Ub*~wyj4cR)6N>u>R
zGyg}fUd~nHN_Z3?*Xz~uP<i+;`FfEWjU<J2*P<_tScU)wqi&@#N?6F8{-XOn_mSrf
zOkT>X{<6S7W;7mnKHDehv9VNb1@}5)4yu}yilTl#39o<2`MWXw{<w`&U^2G4lC;d9
zSG3_5)V9UQ@u=C}y>@D-`nTP6c{<36`n0se3)0wad_{Jec>cuRjVb3lP$&NU?E!0p
z9FoLhuB#4XPHpd{hqrq<0&Ym(nlY7}b%*PeQOJl#wiu;3DMxILd{3@oEJ+JXimFsR
zJ<Lj4>dciQ`3k&y;TQ`59HR9X4Z-v^GK<@PItQU2UAdM=bCO{1++hB=6hVawVDP2X
zbE3?1rQHnX^!`979gBHApDL6ni!NLCC1AZ1LQB&fx&Db-gEhpvWaH|^0E(sb&8@_o
zo1iLaCO|$={@Wu))wI0I&L2;-(&?wvtJ|zfa1TC~Ez@@hhzZn@rZPlMbQg?T({!^`
zg~nO^_^cAq_W;1uQZvNY#<--VC(6r6T1Ov;!BgEKCDtlG3(cnLwoe4jYAg>nW^6N+
zjdy)NnjErI(F;a!8c-qlegzdv_6MD;aosyMHeBBfbZ>^D_#BGN#BCIR|8gQiIS`5^
z^h?r?pq#NS1&5TQaZXiq@EhY|iWh9_H+>(=w8$u2EU%k5dW9QG!%&kYILAn8D0L`0
zwb&1mLt6Q!r{sK-2C#JiC7*9pMlye7E$p#NdidhW-62>nOl_&+ra5WYhdKqnmu4}g
zQ`w&_m2cX6u7z97zPy%)d0!&%_0lP;;>1Od#i_TwkEKw17_4idDbO;SYO5xw8=RSU
z-qA`29uvtlY@+FXR!%H^akIRtdZL>Kf;x{7^UBJuuQE-~-oS$UnR`a~{YW0V`7vz@
z98T*Y(lo3fjJ((Q?2K)=4Cm+Ost4XCW!$RcHU4x7y|E)_tS2{ByR_T7zgST#Y1z{8
zB9TdKj2|<Hdls?UUARRXJIr9wQ7RJ==L$%nIxA=WBi_*7PL3%V*;6y5fgoW<*Lbv>
zLR=`mtd4yTX_;ck^1~$apid?Aa?E?ntsknWBo)VuxLf7BnTXLa%^n2DC0|;bq_z^T
zJ9Z|A>B{{UIcO2oW3@79aH+Bv5NBe-q)V8XwiYt8zw|c+-sRJ<X7fK~Ij_p#e@I4|
z3hJr<xG2~P9nm>&M5JdBvZ~GzJEt}?(9?$jW7`s4=gr-;*fIR&%T0tV-^6ziuCaZZ
z>$J~qSErdS$28b`s`A)ghWn)Ul1l2gI5%O+J1^j{CZr|uQW@bnpSp;?GsUjq8y-$o
z0O##`w7~HL-y%svgLh)wpO&@y^df7w6WKM-pb!&7AJeicfT%M;WV^-c@?15)r2lx!
z&ixj`q~dCvPQf!9od0>Co7RYS-h7Rcf&L@^qP`GoXuB+AF7R?Nm7{P!oFx7AYt|0?
zg7v9NC}<{^DB;L#+m~p?yW&e{8OUfV@)}CVfc+$5X|2{i6%jV2Rza35VW1#rmm#va
zpi3y0w-IU2Q?hKQr)CHmxQ*y+FMuwxU-Iy#VV<L>X=)t994gG_6n8g62}2`lEbi`I
z7*cgpG3-rwl`U`OeRlnWiZ!dK!lspN?u}fL7C~?JOG85$@paTzPtBE&mq*5D7U0(e
zEqeKz3@hyEtfA2UI(bOhwm@+~!l5bSXOS-$ckA)4P^vn?$V<K>VH9(R(}p@;^PvpM
zqk87FB;FLtOv`V&N<hT}Cnu+O4=82W%{aUZ=_uNR6FjDW+jL|WZgm`%bLiGBi7p5i
zQO0BsEYl4<EPG3-qgYDSL(hlMXr2scrXQeYh<?z#6;zdtzSU_%5d2!rzxm3WI43}a
zB-zZ$aDVHhs!EDh7td#2CuKj9Yr~7_L;m=9?dGj@rNEDViH-JRkK=f4T6h)MTV)g^
zHWOlFqhE7|77K}7uILT9QJ^6}YwW^p7RRgquEg=q=PRT?>#yjzqa-p0R?C7!5IU|?
z6rWCSeV_ir`?S5YM;0Dq^Oc#X9gz1x*M3`QyAL|I{nk|33&UEQM1*UL<dS*(9=8e-
zNhepT-1@1kakQrnBmq5X@6Yzb%e4LY?NmR+Olj+5X^W#f+cI0Jj6>TFThi{UKfCp4
zN^iyWGVO27IHX6sX5}DTR905LIbfA?Q}UJTAd!hOX~lXByh2ErMK>dTpUf7X7QP(E
zoMje=a5k@ia7RAfy2bh4v$!TI&O+96qb=kzt%g)*ZZ+URPIT^-B#MgQkP0Y#^nsiE
zOT;Cub|!b08@B{z^n$k$RTqQKmelRSbSXMK)?bIx(MJd8{8~;tbb+LcjPugaSV=_h
zh!|x;a+O73GxXtQtD`N^Gd?<%aI<0Ko>8r^P+-ioL$K4idb}|ql;ku<QbeR=arc`a
z?`e)OW2Z|#$4qH5Ep9i?J2P;7$>yLM4cZQM@?}lu%aB!)Ac-^eXhvAN1M<Z6e!RhV
z?dHBPbu~3%7JHS&Htm0o&iUzKHhC(o9wj!?5zXd6e?Q-_q~)SM#X6%g9&PxdWA1GP
z86)HtH}_}YePky`r5GFVA2YCt*<w0p+_kHoN#hJD=Eo;{2B23~Ed1i19-B7EVY`AT
zu#FASfVwDb*IRju!%pF-OpA+3O+WPzrxcrB4f!zWrHNd9KH9=<1>cN}&e5l+0L&3=
zkS=!m>%AvnYrxo{3^Sk@WHc-piVU0vmBSiSJbZ(UO(pgy1C`DmQ4&kvUwy@SZruxN
zTIdK}y4?Au=4%MqDdmau)x1xb?UE2z1z(DWZIN+>-cHBS{G`jX;94<8zL|JCS+SNF
zgQwTp!sItV54)tD8dhb)ynN-5seg=pR2CP3)dZFW3_LyG6r(Q)sdR_9z?{;=9pZCX
zF(d|I&J;+-=Z-TqBYH{O{&LH`FTzYHs(m|~g|A@7EUP-192-_8kmWlb{(^RsC22?y
zg5#!G)Q}DX9H8^etrD{ulE@FGWT=1=LXJ)Y`A)<_^_B~ODS-g@zpvglPhL_5`ku4p
z{Q@~T-txpm#;4vLU&$!yAw}kngyP5$R}^cEQB=o-GmLu4kEu>X4eH#1G$45y#Ud%u
z%+e}>jv=QrM#5h3?lU5{*Tl4;w^)M}^()>iYW{Qb&!=&HrSC{FA$2>NfQd*9QZM2Q
zb!S#Ykx7NYD0cfw!wO6^4?E_f#ePW*M1xx9J19x?n86G;CmvvMK?Te|Bi8iGCG&qD
z8QdmN;!6A9f4Wuv_dH;x*#GgB|K9TdPhW|Z6I4)OKQF63qx-vdAG88f)!2~$w{l(+
z2)HYhKpCnCVp(I8>)+GI=~~M3j<&mXVPWCi55}DW`%D<nnB2U5I|cf8p{1okZ2k4$
zvB(M1TllVyD3+3dGyhGszdQ4U1pN8GQ4L(d*Qr+?J$mG|k{SM^&Kpq&sKM3L)V37<
ze&l^4*vD>Fppr@e8hdl4jakf=!rW(FoZqnc!U5Vxh}~<BfOhEbtc#vtqZg~xwX`}*
zZF=jt#sE6H1U@74-|RQf&TmsE8-prrqO@JswE$RdKpNqq|32J~(D^f~1ricpVQ1f9
zOOv8Nf6mGRxWIqM_P2XaZc>*yvM@0*p=q>%%?u<#hpdc?)HgT?jG|-zzWO1+K)erc
z@9czvI!d(8#(8Uq7drnfbv{_6E|*HjDT&ScJ}HyGx9nyk_!4I)on(HYY4wn755J%w
z%17(p=lbWd-^YM1FGTHc>2T3cB0;rn80>>tqksJYuq)1o+9{A#H>sf&#&9Uzbi9B`
z-G5Jg8QkQX)Dh9q`{-Wz>EfE_gPZ`aN=C=)wfWz&S%78Ri3PedQ!SPgJ@}@%zc3R_
z{r`Rdly@7(FR`UDp063Y_c7gnJ9)n9fN)593R$1k_hk3M<-ezDI1soAzBeU#e_o}f
ztv?WT82WM<s8cN4|D5=PbZddv_QGvn|FnN!rt8-|a4;6%UcL4Bc~0=Ze|xP2fCvuV
z-Q9^|c2po$61OfIH2p3%HV_1ytwBG*s@)-VkcFTudll%^pCZ|LUWiPsaCiZ<br3!1
zf3v&mwkbdM25z+UyQ-<^*YKzNM6^BsJq9>7H)kgs{AAB-JX3u(s-F*MQ<{TDSMRKb
z-o$_W81gOsB^i)r9iQ$8ZyjwmLIITpS~U=G2G4;+c1M~g`VFWI#PskLwH@PeN`hKj
z+W{0ZgX(lttWvK)3KFb0NJeG>h{}Iw<e%+&!bFh(qKhT|Cx2KCUa{o?kMpzBY{TJ3
z?Pi_H^!leeckZBLBPvBMt<BU{US2mWfRsLa<hJsoSuel-&uU(xypKoxfGjfwCf#@E
zQQcMxIa)w<AA~A10V%-p8V;xB4zjoa?IQ5sS%^K>3`|XRqnPE^N{@uk)FcU{YNkkt
z-+yXsoa`BciHLJKm_|_hpW^K{=?wrhN*hG7aHoKYLHJqMl~ngpBW^Dc-yYB~;VY@w
z1KBx|`!fKR9tkNhz5%dQ)N)%FK-$iwEms^A&pI`pNT7Xjnw#;d7;ibigSj$McAH+%
zbY<qHq2&=z-X?&)uCcMP`Nw!{q8$l`p_e&;F%t<oZ@yIhd-?GxgwoQ#7&}Uc5x_5B
zPP|Wx{T{_6*1^i5MbUKLd6hq=Uwgf@SFp7?m{bHrr&2x9-r1pUV})jD`O?lK5j_kB
zNB~lR4%t<BU%%OScKV$c6J%9k3S34J`=8%6ZrEBLJvTH=%-QO_6Xq;thiWURM8Dlx
z>g5s@9lDquO3v^QT<mR7Ua|CDCI*2KY&f(I2@m(F{Cg%BoyZm!zd(EK<{l8Umocw-
zK`%)79y)J!>DRw>uLCD?*#QWvimL4nmi~nhXmjV`?y%Nig@b|b-iW~j1TRt!b)QDp
z`_!A5QWP6DOFB;89Tu%cU{Oj==k6DaQ=L`r%&`k(`t0;TQSR&*XRFg5(0+Xr?6PD*
zwy#29^CEbui1V8K@}1I!AGgPKNs+mr&BCzk)xScE*wu9?l&k>6FFo|uGgM1$`<ZEN
ziNgY5vP;?esAtd$BM-(97c>iz95O<}VYCm_ol@*dWo@)GSWqy%(@d430XeLLXetb*
z@$}Cc2=t}pXh2hPHlqZ}4`u@zb4<bejcJS|H)i<%xtt>Dv>z?X_cU808ks6+g6zBa
zA;Wj)4X8{{ZSEX3pG7PNWMyUbA1UO2$E5sS!>Vgt-H+NYgp<<;S;f&$3C55jA)=8`
zn}kMiFOrHBEt_8P-%nG74Eh1+h%JPL==(V-2!L<3)_LQKty)RWb1p_LA`0x1*`+EH
z!fuP5*-z3|+1NUA??O81fV*0Is|wJuW(Ay`F0YAw!udJ+rx%I4ff|4$V>`svNEZ@z
z6L>IjXlC(9qSft^{=H|8xqpSVv>DI44<B;D#Wf1z&j2zrG3LyGqII1454fW*E32d`
zqnnu%=$?H8)^_$uxoQl%Emevl>RFynR@i0qxqbAYrU{E1hsGo!wa)Ank*^9(Dp9ho
zC4Zu<SX+^)3uCrFx{x{w+<6=JV}<K}yDQGI?M;D1Wg+oxp$uAb+aWm=i&FxX)7Ef+
zplc9D0AAVqiWYy>6%|k!x9Re#SI3Gr!RtD|b^Oo?rMZCZa6u+33R+mo-l@!@2ZgM)
z;(6xvFXsmo&y-Un9P3OTK-KR|8BXl?Tp5@>uX!Vs<v?NtjY`{OUAOTPu<;+VGBWv*
zP<8tp-LKL1F2U5T4+P6BdmXN$-eMp|=1L@qzR;#Bw_m!ngga4VAM{^?beK*2V<o|!
zx&N`L4E%0uKQf6xB4LfKnsGIfLlLcKS0Wl&5SPE&LuDdNgzdn`O9=xGB6A>6>tmjE
z?Cbc$0ji<S#5HVw<{X##fHwmOFst9=CE>Y?Uf{gb@u8xz+(S**Bggpl4!cGn5Q0Aj
z+b(F<DM#tN=(690_y_kIw*8RRr(3PmXjSSIQ5!PQ9A@o5mGc^ilv^&fc7vZ)>bVk;
zx+KnKHEzvav0VKS9YMCkm8_TtD1Iml7L-u4yTMI?rmN4I3!DQ*R8EuHN)^NGj{viU
z3xMd`<w7u=B(o3o;ac*8jE8JFY-4sqUmh71(k7^H881D%LDWdq$PjoVYXEqKcmGrK
zZCUr8Jb4s|_)%4{3rOX+qp6Mn2(cJIGyrA*sigqq?;)5A$OBkoU-wW*uV5{Oey-kj
zbac8CGF4G9U`u~@>EZQ+rxb&P3>u~`?Q|<MwFfjsv?aOGj?L3m%w(10>5m7D{yic8
zFerUvMOp9}6sCGEk|+S1?~<`!d;HI)rAre6E|w!EPvW`t@(s5WZA<V_?Uh-=9^8)p
z8{l9ipDlC4%_yh$A2-u$m;y+GZO+3d&chCKkq!6q8%%)Z;d%34F@0{LeR#^=Duogy
zw*l;sXdn1Wd;F-jIxA!OQW(&FW!pbF>=xM8<!Uf*=<Lsx+Mv~W)R8V!L2dxphd!zk
zP?N29H<UEUOQ`JzDB0y0<$d;d;VyYwj6HbLkN;_dIIwkJMYg>0O9^7c3W0|ui;Ig&
zR@1@M3W+(%(F8P82GwmhB%NakE8n@k308!x#PJ&eOUf@F)!Wsz75xCS5T4o>g%S4$
zkixQZV#NM5G)CTb(*FeLKuS<EyZwKgfv!Dd;-rE<RtB?N0M$bh@x_y)1I2Qw=5T^R
zVg}PWHk)>Vyg(~9Y-UN%`g50u)8NvA5^F6`Mj%zYk(*a2>R6nwCL7;*TI2<XX~|!O
z>&Xv^k}IH;sO3X|-8xX)?haVDlQxO$23$nK&$1^j({fwlcb5LAfgqi~VEw#ZVx&ph
zxw%+3f<+<N5#QH+HC)Dhb?$S7;7+tD1pyNYiGn!~1%VMA(0<r?&<uGS?oNy<gihHO
zM`hi9*}-fAL5S8y3pAznGmzO~%m70)qlk5gXGFw23m?W;|CL>;%fsyJAIDsQmYqcM
z{=44z^mFA)SQ&dR81xE=AxUfn=TaoeE?s)zDk;*=HIK){1c2@HfYdwuvsR?(o)l1f
zJ6Hg&Y2(hG*J3}K|Hj0regqc02lTNl1)I!W)v`69nm%AIY#G9)kOI#`O})IlvYXx?
zU?>{!IILYVzT<8Cm_@*h+&7?WdTU3{wJ1*RgpU$ml_u+nax<&G;|Y9||FcHGw0$BI
zrY~Sz4)zz?dpobp_3F<O-l#I{xPKBYtf&sOoUY4iH5_aQ{2i<|O>OZ{8$_M~8-pHR
zxBU<>1v1<Z79<IR`4t!xrF*LH(+inL0{?b%r4f@o*>vkF(8POa*~qcifu9vg;K%7(
z>)^IZ(1lR4NZiHY2{ZUwrNR=}WZbsk(!Wy?D3Mvob8(e4@zQwX`rl<tI$uVRIS2wa
z`AgO265Xgw>ed3#m*v(ltn?lQ2JBf{*5r3%U}}j4M2}dU(G9o&OF#Ro7%y*6?Vd1+
zk7=s&XOX+GAD;p}7n6bU3+$zJw(4=~NszDymL;%46P=@Rvq1u*a|pk)lY?{nf>w`Y
z0V-C$TlD@2NHPt&|GWf$;3a?+vzn=K?;R3dZ52Le4hbhj*0`@F@2f#<iLIx2f@Q;e
zRU_ty0FLw7q<FCWeyp=UV6!{N?%3CAr#vR7;|)<4s=wQd1Cb1-Tu8wx+ll9~42WzV
zfb%#C?kx132e_nmEUqNOU)32vV`%PYjM(84Kz!rU$%c;gp}PIj{O=QW<?tUfQB7Qx
z?gqO8{QSz~YMWDd<B3Ndu$hdEwWK`@fi-KZLomf*6xxoro!<Io!t2a0<#eui@4IcH
z*`k=mA1Et_OuS7_P5nL7+kPSD{wNS8eS~EGs0q6a$Dm@a`IcaCeKF@2YbOO3prTW4
zrq%-5k@_eYG*9U3moJHHtrPH51m?UPi|{_OAF$lE*NRJ}ubo$2rX!vNY5AvdYbGZ$
zd8d|i=H5r*AcdwAwc-C-FlwjP6U@suzg-f1GAA(v?B3B31=tN68#=y~9ZnKo*uk?;
zo?u%rgQ^^qM*0U9fGL@rh=f9To`dfU)Ku8!QU&W*sHm!9K+Ptl3`0!d4EVt@-cVRu
z&e)V}8>oYnrk&*!CzJ=S;O~s)*&uFFs-~_k#$r)zB*D*50ko#*|Muq3o&fjh?@$xe
z?T+}re)%t)#L4+S;{SEm|4LfUfBf?Q^ZoxXzEVhV%5jKHHaJy{+r0qBBb6u0g%2!)
F{uctL%I*LF

diff --git a/docs/_build/html/_sources/basics.rst.txt b/docs/_build/html/_sources/basics.rst.txt
deleted file mode 100644
index 7402b68db..000000000
--- a/docs/_build/html/_sources/basics.rst.txt
+++ /dev/null
@@ -1,214 +0,0 @@
-Basics
-======
-
-Initialization
---------------
-
-Before use Warp should be explicitly initialized with the ``wp.init()`` method as follows::
-
-    import warp as wp
-
-    wp.init()
-
-Warp will print some startup information about the compute devices available, driver versions, and the location
-for any generated kernel code, e.g.:
-
-.. code:: bat
-
-    Warp 1.0.0 initialized:
-    CUDA Toolkit: 11.8, Driver: 12.1
-    Devices:
-        "cpu"    | AMD64 Family 25 Model 33 Stepping 0, AuthenticAMD
-        "cuda:0" | NVIDIA GeForce RTX 4080 (sm_89)
-    Kernel cache: C:\Users\mmacklin\AppData\Local\NVIDIA Corporation\warp\Cache\1.0.0
-
-
-Kernels
--------
-
-In Warp, compute kernels are defined as Python functions and annotated with the ``@wp.kernel`` decorator, as follows::
-
-    @wp.kernel
-    def simple_kernel(a: wp.array(dtype=wp.vec3),
-                      b: wp.array(dtype=wp.vec3),
-                      c: wp.array(dtype=float)):
-
-        # get thread index
-        tid = wp.tid()
-
-        # load two vec3s
-        x = a[tid]
-        y = b[tid]
-
-        # compute the dot product between vectors
-        r = wp.dot(x, y)
-
-        # write result back to memory
-        c[tid] = r
-
-Because Warp kernels are compiled to native C++/CUDA code, all the function input arguments must be statically typed. This allows 
-Warp to generate fast code that executes at essentially native speeds. Because kernels may be run on either the CPU
-or GPU, they cannot access arbitrary global state from the Python environment. Instead they must read and write data
-through their input parameters such as arrays.
-
-Warp kernels functions have a 1:1 correspondence with CUDA kernels, to launch a kernel with 1024 threads, we use
-:func:`wp.launch() <warp.launch>` as follows::
-
-    wp.launch(kernel=simple_kernel, # kernel to launch
-              dim=1024,             # number of threads
-              inputs=[a, b, c],     # parameters
-              device="cuda")        # execution device
-
-Inside the kernel we can retrieve the *thread index* of the each thread using the ``wp.tid()`` builtin function::
-
-    # get thread index
-    i = wp.tid()
-
-Kernels can be launched with 1D, 2D, 3D, or 4D grids of threads, e.g.: to launch a 2D grid of threads to process a 1024x1024 image we could write::
-
-    wp.launch(kernel=compute_image, 
-              dim=(1024, 1024),       
-              inputs=[img],     
-              device="cuda")
-
-Then, inside the kernel we can retrieve a 2D thread index as follows::
-
-    # get thread index
-    i, j = wp.tid()
-
-    # write out a color value for each pixel
-    color[i, j] = wp.vec3(r, g, b)
-
-.. _example-cache-management:
-
-Example: Changing the kernel cache directory
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-The following example illustrates how the location for generated and compiled
-kernel code can be changed before and after calling ``wp.init()``.
-
-.. code:: python
-
-    import os
-
-    import warp as wp
-
-    example_dir = os.path.dirname(os.path.realpath(__file__))
-
-    # set default cache directory before wp.init()
-    wp.config.kernel_cache_dir = os.path.join(example_dir, "tmp", "warpcache1")
-
-    wp.init()
-
-    print("+++ Current cache directory: ", wp.config.kernel_cache_dir)
-
-    # change cache directory after wp.init()
-    wp.build.init_kernel_cache(os.path.join(example_dir, "tmp", "warpcache2"))
-
-    print("+++ Current cache directory: ", wp.config.kernel_cache_dir)
-
-    # clear kernel cache (forces fresh kernel builds every time)
-    wp.build.clear_kernel_cache()
-
-
-    @wp.kernel
-    def basic(x: wp.array(dtype=float)):
-        tid = wp.tid()
-        x[tid] = float(tid)
-
-
-    device = "cpu"
-    n = 10
-    x = wp.zeros(n, dtype=float, device=device)
-
-    wp.launch(kernel=basic, dim=n, inputs=[x], device=device)
-    print(x.numpy())
-
-Arrays
-------
-
-Memory allocations are exposed via the ``wp.array`` type. Arrays wrap an underlying memory allocation that may live in
-either host (CPU), or device (GPU) memory. Arrays are strongly typed and store a linear sequence of built-in values
-(``float,``, ``int``, ``vec3``, ``matrix33``, etc).
-
-Arrays can be allocated similar to PyTorch::
-
-    # allocate an uninitialized array of vec3s
-    v = wp.empty(shape=n, dtype=wp.vec3, device="cuda")
-
-    # allocate a zero-initialized array of quaternions    
-    q = wp.zeros(shape=n, dtype=wp.quat, device="cuda")
-
-    # allocate and initialize an array from a NumPy array
-    # will be automatically transferred to the specified device
-    a = np.ones((10, 3), dtype=np.float32)
-    v = wp.from_numpy(a, dtype=wp.vec3, device="cuda")
-
-By default, Warp arrays that are initialized from external data (e.g.: NumPy, Lists, Tuples) will create a copy the data to new memory for the
-device specified. However, it is possible for arrays to alias external memory using the ``copy=False`` parameter to the
-array constructor provided the input is contiguous and on the same device. See the :doc:`/modules/interoperability`
-section for more details on sharing memory with external frameworks.
-
-To read GPU array data back to CPU memory we can use the ``array.numpy()`` method::
-
-    # bring data from device back to host
-    view = device_array.numpy()
-
-This will automatically synchronize with the GPU to ensure that any outstanding work has finished, and will
-copy the array back to CPU memory where it is passed to NumPy. Calling ``array.numpy()`` on a CPU array will return
-a zero-copy NumPy view onto the Warp data.
-
-User Functions
---------------
-
-Users can write their own functions using the ``@wp.func`` decorator, for example::
-
-    @wp.func
-    def square(x: float):
-        return x*x
-
-User functions can be called freely from within kernels inside the same module and accept arrays as inputs. 
-
-Compilation Model
------------------
-
-Warp uses a Python->C++/CUDA compilation model that generates kernel code from Python function definitions. All kernels belonging to a Python module are runtime compiled into dynamic libraries and PTX, the result is then cached between application restarts for fast startup times.
-
-Note that compilation is triggered on the first kernel launch for that module. Any kernels registered in the module with ``@wp.kernel`` will be included in the shared library.
-
-.. image:: ./img/compiler_pipeline.png
-
-
-Language Details
-----------------
-
-To support GPU computation and differentiability, there are some differences from the CPython runtime.
-
-Built-in Types
-^^^^^^^^^^^^^^
-
-Warp supports a number of built-in math types similar to high-level shading languages, for example ``vec2, vec3, vec4, mat22, mat33, mat44, quat, array``. All built-in types have value semantics so that expressions such as ``a = b`` generate a copy of the variable b rather than a reference.
-
-Strong Typing
-^^^^^^^^^^^^^
-
-Unlike Python, in Warp all variables must be typed. Types are inferred from source expressions and function signatures using the Python typing extensions. All kernel parameters must be annotated with the appropriate type, for example: ::
-
-    @wp.kernel
-    def simple_kernel(a: wp.array(dtype=vec3),
-                      b: wp.array(dtype=vec3),
-                      c: float):
-
-Tuple initialization is not supported, instead variables should be explicitly typed: ::
-
-    # invalid
-    a = (1.0, 2.0, 3.0)        
-
-    # valid
-    a = wp.vec3(1.0, 2.0, 3.0) 
-
-
-Limitations and Unsupported Features
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-See :doc:`limitations` for a list of Warp limitations and unsupported features.
diff --git a/docs/_build/html/_sources/configuration.rst.txt b/docs/_build/html/_sources/configuration.rst.txt
deleted file mode 100644
index a19f228aa..000000000
--- a/docs/_build/html/_sources/configuration.rst.txt
+++ /dev/null
@@ -1,138 +0,0 @@
-Runtime Settings
-================
-
-Warp has settings at the global, module, and kernel level that can be used to fine-tune the compilation and verbosity
-of Warp programs. In cases in which a setting can be changed at multiple levels (e.g ``enable_backward``),
-the setting at the more-specific scope takes precedence.
-
-Global Settings
----------------
-
-To change a setting, prepend ``wp.config.`` to the name of the variable and assign a value to it.
-Some settings may be changed on the fly, while others need to be set prior to calling ``wp.init()`` to take effect.
-
-For example, the location of the user kernel cache can be changed with:
-
-.. code-block:: python
-
-    import os
-
-    import warp as wp
-
-    example_dir = os.path.dirname(os.path.realpath(__file__))
-
-    # set default cache directory before wp.init()
-    wp.config.kernel_cache_dir = os.path.join(example_dir, "tmp", "warpcache1")
-
-    wp.init()
-
-
-Basic Global Settings
-^^^^^^^^^^^^^^^^^^^^^
-
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-| Field              | Type    |Default Value| Description                                                              |
-+====================+=========+=============+==========================================================================+
-|``verify_fp``       | Boolean | ``False``   | If ``True``, Warp will check that inputs and outputs are finite before   |
-|                    |         |             | and/or after various operations. **Has performance implications.**       |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``verify_cuda``     | Boolean | ``False``   | If ``True``, Warp will check for CUDA errors after every launch and      |
-|                    |         |             | memory operation. CUDA error verification cannot be used during graph    |
-|                    |         |             | capture. **Has performance implications.**                               |              
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``print_launches``  | Boolean | ``False``   | If ``True``, Warp will print details of every kernel launch to standard  |
-|                    |         |             | out (e.g. launch dimensions, inputs, outputs, device, etc.).             |
-|                    |         |             | **Has performance implications.**                                        |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``mode``            | String  |``"release"``| Controls whether to compile Warp kernels in debug or release mode.       |
-|                    |         |             | Valid choices are ``"release"`` or ``"debug"``.                          |
-|                    |         |             | **Has performance implications.**                                        |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``verbose``         | Boolean | ``False``   | If ``True``, additional information will be printed to standard out      |
-|                    |         |             | during code generation, compilation, etc.                                |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``quiet``           | Boolean | ``False``   | If ``True``, Warp module initialization messages will be disabled.       |
-|                    |         |             | This setting does not affect error messages and warnings.                |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``kernel_cache_dir``| String  | ``None``    | The path to the directory used for the user kernel cache. Subdirectories |
-|                    |         |             | named ``gen`` and ``bin`` will be created in this directory. If ``None``,|
-|                    |         |             | a directory will be automatically determined using                       |
-|                    |         |             | `appdirs.user_cache_directory <https://github.com/ActiveState/appdirs>`_ |
-|                    |         |             |                                                                          |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``enable_backward`` | Boolean | ``True``    | If ``True``, backward passes of kernels will be compiled by default.     |
-|                    |         |             | Disabling this setting can reduce kernel compilation times.              |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-
-Advanced Global Settings
-^^^^^^^^^^^^^^^^^^^^^^^^
-
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-| Field              | Type    |Default Value| Description                                                              |
-+====================+=========+=============+==========================================================================+
-|``cache_kernels``   | Boolean | ``True``    | If ``True``, kernels that have already been compiled from previous       |
-|                    |         |             | application launches will not be recompiled.                             |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``cuda_output``     | String  | ``None``    | The preferred CUDA output format for kernels. Valid choices are ``None``,|
-|                    |         |             | ``"ptx"``, and ``"cubin"``. If ``None``, a format will be determined     |
-|                    |         |             | automatically.                                                           |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``ptx_target_arch`` | Integer | 70          | The target architecture for PTX generation.                              |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``llvm_cuda``       | Boolean | ``False``   | If ``True``, Clang/LLVM will be used to compile CUDA code instead of     |
-|                    |         |             | NVTRC.                                                                   |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-
-Module Settings
----------------
-
-Module-level settings to control runtime compilation and code generation may be changed by passing a dictionary of
-option pairs to ``wp.set_module_options()``.
-
-For example, compilation of backward passes for the kernel in an entire module can be disabled with:
-
-.. code:: python
-
-    wp.set_module_options({"enable_backward": False})
-
-The options for a module can also be queried using ``wp.get_module_options()``.
-
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-| Field              | Type    |Default Value| Description                                                              |
-+====================+=========+=============+==========================================================================+
-|``mode``            | String  | Global      | Controls whether to compile the module's kernels in debug or release     |
-|                    |         | setting     | mode by default. Valid choices are ``"release"`` or ``"debug"``.         |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``max_unroll``      | Integer | 16          | The maximum fixed-size loop to unroll. Note that ``max_unroll`` does not |
-|                    |         |             | consider the total number of iterations in nested loops. This can result |
-|                    |         |             | in a large amount of automatically generated code if each nested loop is |
-|                    |         |             | below the ``max_unroll`` threshold.                                      |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``enable_backward`` | Boolean | Global      | If ``True``, backward passes of kernels will be compiled by default.     |
-|                    |         | setting     | Valid choices are ``"release"`` or ``"debug"``.                          |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``fast_math``       | Boolean | ``False``   | If ``True``, CUDA kernels will be compiled with the ``--use_fast_math``  |
-|                    |         |             | compiler option, which enables some fast math operations that are faster |
-|                    |         |             | but less accurate.                                                       |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-|``cuda_output``     | String  | ``None``    | The preferred CUDA output format for kernels. Valid choices are ``None``,|
-|                    |         |             | ``"ptx"``, and ``"cubin"``. If ``None``, a format will be determined     |
-|                    |         |             | automatically. The module-level setting takes precedence over the global |
-|                    |         |             | setting.                                                                 |
-+--------------------+---------+-------------+--------------------------------------------------------------------------+
-
-Kernel Settings
----------------
-
-``enable_backward`` is currently the only setting that can also be configured on a per-kernel level.
-Backward-pass compilation can be disabled by passing an argument into the ``@wp.kernel`` decorator
-as in the following example:
-
-.. code-block:: python
-
-    @wp.kernel(enable_backward=False)
-    def scale_2(
-        x: wp.array(dtype=float),
-        y: wp.array(dtype=float),
-    ):
-        y[0] = x[0] ** 2.0
diff --git a/docs/_build/html/_sources/debugging.rst.txt b/docs/_build/html/_sources/debugging.rst.txt
deleted file mode 100644
index c006bd226..000000000
--- a/docs/_build/html/_sources/debugging.rst.txt
+++ /dev/null
@@ -1,86 +0,0 @@
-Debugging
-=========
-
-Printing Values
----------------
-
-Often one of the best debugging methods is to simply print values from kernels. Warp supports printing all built-in
-types using the ``print()`` function, e.g.::
-
-    v = wp.vec3(1.0, 2.0, 3.0)
-
-    print(v)   
-
-In addition, formatted C-style printing is available through the ``wp.printf()`` function, e.g.::
-
-    x = 1.0
-    i = 2
-
-    wp.printf("A float value %f, an int value: %d", x, i)
-
-.. note:: Formatted printing is only available for scalar types (e.g.: ``int`` and ``float``) not vector types.
-
-Printing Launches
------------------
-
-For complex applications it can be difficult to understand the order-of-operations that lead to a bug. To help diagnose
-these issues Warp supports a simple option to print out all launches and arguments to the console::
-
-    wp.config.print_launches = True
-
-
-Step-Through Debugging
-----------------------
-
-It is possible to attach IDE debuggers such as Visual Studio to Warp processes to step through generated kernel code.
-Users should first compile the kernels in debug mode by setting::
-   
-    wp.config.mode = "debug"
-
-This setting ensures that line numbers, and debug symbols are generated correctly. After launching the Python process,
-the debugger should be attached, and a breakpoint inserted into the generated code.
-
-.. note:: Generated kernel code is not a 1:1 correspondence with the original Python code, but individual operations can still be replayed and variables inspected.
-
-Also see :github:`warp/tests/test_debug.py` for an example of how to debug Warp kernel code running on the CPU.
-
-Generated Code
---------------
-
-The generated code for kernels is stored in a central cache location in the user's home directory, the cache location
-is printed at startup when ``wp.init()`` is called, for example:
-
-.. code-block:: console
-
-    Warp 0.8.1 initialized:
-        CUDA Toolkit: 11.8, Driver: 11.8
-        Devices:
-        "cpu"    | AMD64 Family 25 Model 33 Stepping 0, AuthenticAMD
-        "cuda:0" | NVIDIA GeForce RTX 3090 (sm_86)
-        "cuda:1" | NVIDIA GeForce RTX 2080 Ti (sm_75)
-        Kernel cache: C:\Users\LukasW\AppData\Local\NVIDIA Corporation\warp\Cache\0.8.1
-
-The kernel cache has ``gen`` and ``bin`` folders that contain the generated C++/CUDA code and the compiled binaries
-respectively. Occasionally it can be useful to inspect the generated code for debugging / profiling.
-
-Bounds Checking
----------------
-
-Warp will perform bounds checking in debug build configurations to ensure that all array accesses lie within the defined
-shape.
-
-CUDA Verification
------------------
-
-It is possible to generate out-of-bounds memory access violations through poorly formed kernel code or inputs. In this
-case the CUDA runtime will detect the violation and put the CUDA context into an error state. Subsequent kernel launches
-may silently fail which can lead to hard to diagnose issues.
-
-If a CUDA error is suspected a simple verification method is to enable::
-
-    wp.config.verify_cuda = True
-
-This setting will check the CUDA context after every operation to ensure that it is still valid. If an error is
-encountered it will raise an exception that often helps to narrow down the problematic kernel.
-
-.. note:: Verifying CUDA state at each launch requires synchronizing CPU and GPU which has a significant overhead. Users should ensure this setting is only used during debugging.
diff --git a/docs/_build/html/_sources/faq.rst.txt b/docs/_build/html/_sources/faq.rst.txt
deleted file mode 100644
index bcb924c51..000000000
--- a/docs/_build/html/_sources/faq.rst.txt
+++ /dev/null
@@ -1,112 +0,0 @@
-FAQ
-===
-
-How does Warp relate to other Python projects for GPU programming, e.g.: Numba, Taichi, cuPy, PyTorch, etc.?
-------------------------------------------------------------------------------------------------------------
-
-Warp is inspired by many of these projects, and is closely related to
-Numba and Taichi which both expose kernel programming to Python. These
-frameworks map to traditional GPU programming models, so many of the
-high-level concepts are similar, however there are some functionality
-and implementation differences.
-
-Compared to Numba, Warp supports a smaller subset of Python, but
-offering auto-differentiation of kernel programs, which is useful for
-machine learning. Compared to Taichi Warp uses C++/CUDA as an
-intermediate representation, which makes it convenient to implement and
-expose low-level routines. In addition, we are building in
-data structures to support geometry processing (meshes, sparse volumes,
-point clouds, USD data) as first-class citizens that are not exposed in
-other runtimes.
-
-Warp does not offer a full tensor-based programming model like PyTorch
-and JAX, but is designed to work well with these frameworks through data
-sharing mechanisms like ``__cuda_array_interface__``. For computations
-that map well to tensors (e.g.: neural-network inference) it makes sense
-to use these existing tools. For problems with a lot of e.g.: sparsity,
-conditional logic, heterogenous workloads (like the ones we often find in
-simulation and graphics), then the kernel-based programming model like
-the one in Warp are often more convenient since users have control over
-individual threads.
-
-Does Warp support all of the Python language?
----------------------------------------------
-
-No, Warp supports a subset of Python that maps well to the GPU. Our goal
-is to not have any performance cliffs so that users can expect
-consistently good behavior from kernels that is close to native code.
-Examples of unsupported concepts that don’t map well to the GPU are
-dynamic types, list comprehensions, exceptions, garbage collection, etc.
-
-When should I call ``wp.synchronize()``?
-----------------------------------------
-
-One of the common sources of confusion for new users is when calls to
-``wp.synchronize()`` are necessary. The answer is “almost never”!
-Synchronization is quite expensive, and should generally be avoided
-unless necessary. Warp naturally takes care of synchronization between
-operations (e.g.: kernel launches, device memory copies).
-
-For example, the following requires no manual synchronization, as the
-conversion to NumPy will automatically synchronize:
-
-.. code:: python
-
-    # run some kernels
-    wp.launch(kernel_1, dim, [array_x, array_y], device="cuda")
-    wp.launch(kernel_2, dim, [array_y, array_z], device="cuda")
-
-    # bring data back to host (and implicitly synchronize)
-    x = array_z.numpy()
-
-The *only* case where manual synchronization is needed is when copies
-are being performed back to CPU asynchronously, e.g.:
-
-.. code:: python
-
-    # copy data back to cpu from gpu, all copies will happen asynchronously to Python
-    wp.copy(cpu_array_1, gpu_array_1)
-    wp.copy(cpu_array_2, gpu_array_2)
-    wp.copy(cpu_array_3, gpu_array_3)
-
-    # ensure that the copies have finished
-    wp.synchronize()
-
-    # return a numpy wrapper around the cpu arrays, note there is no implicit synchronization here
-    a1 = cpu_array_1.numpy()
-    a2 = cpu_array_2.numpy()
-    a3 = cpu_array_3.numpy()
-
-What happens when you differentiate a function like ``wp.abs(x)``?
-------------------------------------------------------------------
-
-Non-smooth functions such as ``y=|x|`` do not have a single unique
-gradient at ``x=0``, rather they have what is known as a
-``subgradient``, which is formally the convex hull of directional
-derivatives at that point. The way that Warp (and most
-auto-differentiation frameworks) handles these points is to pick an
-arbitrary gradient from this set, e.g.: for ``wp.abs()``, it will
-arbitrarily choose the gradient to be 1.0 at the origin. You can find
-the implementation for these functions in ``warp/native/builtin.h``.
-
-Most optimizers (particularly ones that exploit stochasticity), are not
-sensitive to the choice of which gradient to use from the subgradient,
-although there are exceptions.
-
-Does Warp support multi-GPU programming?
-----------------------------------------
-
-Yes! Since version ``0.4.0`` we support allocating, launching, and
-copying between multiple GPUs in a single process. We follow the naming
-conventions of PyTorch and use aliases such as ``cuda:0``, ``cuda:1``,
-``cpu`` to identify individual devices.
-
-Should I switch to Warp over IsaacGym / PhysX?
-----------------------------------------------
-
-Warp is not a replacement for IsaacGym, IsaacSim, or PhysX - while Warp
-does offer some physical simulation capabilities this is primarily aimed
-at developers who need differentiable physics, rather than a fully
-featured physics engine. Warp is also integrated with IsaacGym and is
-great for performing auxiliary tasks such as reward and observation
-computations for reinforcement learning.
diff --git a/docs/_build/html/_sources/index.rst.txt b/docs/_build/html/_sources/index.rst.txt
deleted file mode 100644
index 1c7c8f690..000000000
--- a/docs/_build/html/_sources/index.rst.txt
+++ /dev/null
@@ -1,192 +0,0 @@
-NVIDIA Warp Documentation
-=========================
-
-Warp is a Python framework for writing high-performance simulation and graphics code. Warp takes
-regular Python functions and JIT compiles them to efficient kernel code that can run on the CPU or GPU.
-
-Warp is designed for spatial computing and comes with a rich set of primitives that make it easy to write 
-programs for physics simulation, perception, robotics, and geometry processing. In addition, Warp kernels 
-are differentiable and can be used as part of machine-learning pipelines with frameworks such as PyTorch and JAX.
-
-Below are some examples of simulations implemented using Warp:
-
-.. image:: ./img/header.png
-
-Quickstart
-==========
-
-The easiest way is to install Warp is from PyPi:
-
-.. code-block:: sh
-
-    $ pip install warp-lang
-
-Pre-built binary packages for Windows, Linux and macOS are also available on the `Releases <https://github.com/NVIDIA/warp/releases>`__ page. To install in your local Python environment extract the archive and run the following command from the root directory:
-
-.. code-block:: sh
-
-    $ pip install .
-
-Basic example
--------------
-
-An example first program that computes the lengths of random 3D vectors is given below::
-
-    import warp as wp
-    import numpy as np
-
-    wp.init()
-
-    num_points = 1024
-
-    @wp.kernel
-    def length(points: wp.array(dtype=wp.vec3),
-               lengths: wp.array(dtype=float)):
-
-        # thread index
-        tid = wp.tid()
-        
-        # compute distance of each point from origin
-        lengths[tid] = wp.length(points[tid])
-
-
-    # allocate an array of 3d points
-    points = wp.array(np.random.rand(num_points, 3), dtype=wp.vec3)
-    lengths = wp.zeros(num_points, dtype=float)
-
-    # launch kernel
-    wp.launch(kernel=length,
-              dim=len(points),
-              inputs=[points, lengths])
-
-    print(lengths)
-
-Additional examples
--------------------
-The `examples <https://github.com/NVIDIA/warp/tree/main/examples>`__ directory in
-the Github repository contains a number of scripts that show how to
-implement different simulation methods using the Warp API. Most examples
-will generate USD files containing time-sampled animations in the
-``examples/outputs`` directory. Before running examples users should
-ensure that the ``usd-core`` package is installed using:
-
-::
-
-    pip install usd-core
-
-USD files can be viewed or rendered inside NVIDIA
-`Omniverse <https://developer.nvidia.com/omniverse>`__,
-Pixar's UsdView, and Blender. Note that Preview in macOS is not
-recommended as it has limited support for time-sampled animations.
-
-Built-in unit tests can be run from the command-line as follows:
-
-::
-
-    python -m warp.tests
-
-Omniverse
----------
-
-A Warp Omniverse extension is available in the extension registry inside
-Omniverse Kit or USD Composer.
-
-Enabling the extension will automatically install and initialize the
-Warp Python module inside the Kit Python environment. Please see the
-`Omniverse Warp Documentation <https://docs.omniverse.nvidia.com/extensions/latest/ext_warp.html>`__
-for more details on how to use Warp in Omniverse.
-
-
-Learn More
-----------
-
-Please see the following resources for additional background on Warp:
-
--  `GTC 2022
-   Presentation <https://www.nvidia.com/en-us/on-demand/session/gtcspring22-s41599>`__
--  `GTC 2021
-   Presentation <https://www.nvidia.com/en-us/on-demand/session/gtcspring21-s31838>`__
--  `SIGGRAPH Asia 2021 Differentiable Simulation
-   Course <https://dl.acm.org/doi/abs/10.1145/3476117.3483433>`__
-
-The underlying technology in Warp has been used in a number of research
-projects at NVIDIA including the following publications:
-
--  Accelerated Policy Learning with Parallel Differentiable Simulation -
-   Xu, J., Makoviychuk, V., Narang, Y., Ramos, F., Matusik, W., Garg,
-   A., & Macklin, M.
-   `(2022) <https://short-horizon-actor-critic.github.io>`__
--  DiSECt: Differentiable Simulator for Robotic Cutting - Heiden, E.,
-   Macklin, M., Narang, Y., Fox, D., Garg, A., & Ramos, F
-   `(2021) <https://github.com/NVlabs/DiSECt>`__
--  gradSim: Differentiable Simulation for System Identification and
-   Visuomotor Control - Murthy, J. Krishna, Miles Macklin, Florian
-   Golemo, Vikram Voleti, Linda Petrini, Martin Weiss, Breandan
-   Considine et
-   al. `(2021) <https://gradsim.github.io>`__
-
-Citing
-------
-
-If you use Warp in your research please use the following citation:
-
-.. code:: bibtex
-
-    @misc{warp2022,
-        title= {Warp: A High-performance Python Framework for GPU Simulation and Graphics},
-        author = {Miles Macklin},
-        month = {March},
-        year = {2022},
-        note= {NVIDIA GPU Technology Conference (GTC)},
-        howpublished = {\url{https://github.com/nvidia/warp}}
-    }
-
-License
--------
-
-Warp is provided under the NVIDIA Source Code License (NVSCL), please see
-`LICENSE.md <https://github.com/NVIDIA/warp/blob/main/LICENSE.md>`_ for the full license text.
-
-Please contact `omniverse-license-questions@nvidia.com <mailto:omniverse-license-questions@nvidia.com>`_ for
-commercial licensing inquires.
-
-Full Table of Contents
-----------------------
-
-.. toctree::
-    :maxdepth: 2
-    :caption: User's Guide
-
-    installation
-    basics
-    modules/devices
-    modules/interoperability
-    configuration
-    debugging
-    limitations
-    faq
-
-.. toctree::
-    :maxdepth: 2
-    :caption: Core Reference
-   
-    modules/runtime
-    modules/functions
-   
-.. toctree::
-    :maxdepth: 2
-    :caption: Simulation Reference
-   
-    modules/sim
-    modules/sparse
-    modules/fem
-
-.. toctree::
-    :hidden:
-    :caption: Project Links
-
-    GitHub <https://github.com/NVIDIA/warp>
-    PyPI <https://pypi.org/project/warp-lang>
-    Discord <https://discord.com/channels/827959428476174346/953756751977648148>
-
-:ref:`Full Index <genindex>`
diff --git a/docs/_build/html/_sources/installation.rst.txt b/docs/_build/html/_sources/installation.rst.txt
deleted file mode 100644
index 2d166887f..000000000
--- a/docs/_build/html/_sources/installation.rst.txt
+++ /dev/null
@@ -1,93 +0,0 @@
-Installation
-============
-
-The easiest way is to install Warp is from `PyPi <https://pypi.org/project/warp-lang>`_:
-
-.. code-block:: sh
-
-    $ pip install warp-lang
-
-Pre-built binary packages for Windows, Linux and macOS are also available on the `Releases <https://github.com/NVIDIA/warp/releases>`__ page. To install in your local Python environment extract the archive and run the following command from the root directory:
-
-.. code-block:: sh
-
-    $ pip install .
-
-Dependencies
-------------
-
-Warp supports Python versions 3.7 or later and requires `NumPy <https://numpy.org>`_ to be installed.
-
-The following optional dependencies are required to support certain features:
-
-* `usd-core <https://pypi.org/project/usd-core>`_: Required for some Warp examples, ``warp.sim.parse_usd()``, and ``warp.render.UsdRenderer``.
-* `JAX <https://jax.readthedocs.io/en/latest/installation.html>`_: Required for JAX interoperability (see :ref:`jax-interop`).
-* `PyTorch <https://pytorch.org/get-started/locally/>`_: Required for PyTorch interoperability (see :ref:`pytorch-interop`).
-* `NVTX for Python <https://github.com/NVIDIA/NVTX#python>`_: Required to use :class:`wp.ScopedTimer(use_nvtx=True) <warp.ScopedTimer>`.
-
-Building the Warp documentation requires:
-
-* `Sphinx <https://www.sphinx-doc.org>`_
-* `Furo <https://github.com/pradyunsg/furo>`_
-* `Sphinx-copybutton <https://sphinx-copybutton.readthedocs.io/en/latest/index.html>`_
-
-Building from source
---------------------
-
-For developers who want to build the library themselves the following tools are required:
-
-* Microsoft Visual Studio (Windows), minimum version 2019
-* GCC (Linux), minimum version 7.2
-* `CUDA Toolkit <https://developer.nvidia.com/cuda-toolkit>`_, minimum version 11.5
-* `Git Large File Storage <https://git-lfs.com>`_
-
-If you are cloning from Windows, please first ensure that you have
-enabled “Developer Mode” in Windows settings and symlinks in Git:
-
-.. code-block:: console
-
-    $ git config --global core.symlinks true
-
-This will ensure symlinks inside ``exts/omni.warp.core`` work upon cloning.
-
-After cloning the repository, users should run:
-
-.. code-block:: console
-
-    $ python build_lib.py
-
-This will generate the ``warp.dll`` / ``warp.so`` core library
-respectively. When building manually, users should ensure that their
-``CUDA_PATH`` environment variable is set, otherwise Warp will be built
-without CUDA support. Alternatively, the path to the CUDA Toolkit can be
-passed to the build command as ``--cuda_path="..."``. After building, the
-Warp package should be installed using:
-
-.. code-block:: console
-
-    $ pip install -e .
-
-Which ensures that subsequent modifications to the library will be
-reflected in the Python package.
-
-Conda environments
-------------------
-
-Some modules, such as ``usd-core``, don't support the latest Python version.
-To manage running Warp and other projects on different Python versions one can
-make use of an environment management system such as
-`Conda <https://docs.conda.io/>`_.
-
-**WARNING:** When building and running Warp in a different environment, make sure
-the build environment has the same C++ runtime library version, or an older
-one, than the execution environment. Otherwise Warp's shared libraries may end
-up looking for a newer runtime library version than the one available in the
-execution environment. For example on Linux this error could occur:
-
-``OSError: <...>/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by <...>/warp/warp/bin/warp.so)``
-
-This can be solved by installing a newer C++ runtime version in the runtime
-conda environment using ``conda install -c conda-forge libstdcxx-ng=12.1`` or
-newer. Or, the build environment's C++ toolchain can be downgraded using
-``conda install -c conda-forge libstdcxx-ng=8.5``. Or, one can ``activate`` or
-``deactivate`` conda environments as needed for building vs. running Warp.
diff --git a/docs/_build/html/_sources/limitations.rst.txt b/docs/_build/html/_sources/limitations.rst.txt
deleted file mode 100644
index 5640103fc..000000000
--- a/docs/_build/html/_sources/limitations.rst.txt
+++ /dev/null
@@ -1,56 +0,0 @@
-Limitations
-===========
-
-.. currentmodule:: warp
-
-This section summarizes various limitations and currently unsupported features in Warp.
-Requests for new features can be made at `GitHub Discussions <https://github.com/NVIDIA/warp/discussions>`_,
-and issues can be opened at `GitHub Issues <https://github.com/NVIDIA/warp/issues>`_.
-
-Unsupported Features
---------------------
-
-To achieve good performance on GPUs some dynamic language features are not supported:
-
-* Lambda functions
-* List comprehensions
-* Exceptions
-* Recursion
-* Runtime evaluation of expressions, e.g.: eval()
-* Dynamic structures such as lists, sets, dictionaries, etc.
-
-Kernels
--------
-
-* Strings cannot be passed into kernels.
-* Short-circuit evaluation is not supported.
-* :func:`wp.atomic_add() <atomic_add>` does not support ``wp.int64``.
-* :func:`wp.tid() <tid>` cannot be called from user functions.
-* CUDA thread blocks use a fixed size 256 threads per block.
-
-Arrays
-------
-
-* Arrays can have a maximum of four dimensions.
-* Each dimension of a Warp array cannot be greater than the maximum value representable by a 32-bit signed integer,
-  :math:`2^{31}-1`.
-* There are currently no data types that support complex numbers.
-
-Structs
--------
-
-* Structs cannot have generic members, i.e. of type ``typing.Any``.
-
-Volumes
--------
-
-* The sparse-volume *topology* cannot be changed after the tiles for the :class:`Volume` have been allocated.
-
-Multiple Processes
-------------------
-
-* A CUDA context created in the parent process cannot be used in a *forked* child process.
-  Use the spawn start method instead, or avoid creating CUDA contexts in the parent process.
-* There can be issues with using same user kernel cache directory when running with multiple processes.
-  A workaround is to use a separate cache directory for every process.
-  See :ref:`example-cache-management` for how the cache directory may be changed.
diff --git a/docs/_build/html/_sources/modules/devices.rst.txt b/docs/_build/html/_sources/modules/devices.rst.txt
deleted file mode 100644
index eda532d6f..000000000
--- a/docs/_build/html/_sources/modules/devices.rst.txt
+++ /dev/null
@@ -1,200 +0,0 @@
-.. _devices:
-
-Devices
-=======
-
-Warp assigns unique string aliases to all supported compute devices in the system.  There is currently a single CPU device exposed as ``"cpu"``.  Each CUDA-capable GPU gets an alias of the form ``"cuda:i"``, where ``i`` is the CUDA device ordinal.  This convention should be familiar to users of other popular frameworks like PyTorch.
-
-It is possible to explicitly target a specific device with each Warp API call using the ``device`` argument::
-
-    a = wp.zeros(n, device="cpu")
-    wp.launch(kernel, dim=a.size, inputs=[a], device="cpu")
-
-    b = wp.zeros(n, device="cuda:0")
-    wp.launch(kernel, dim=b.size, inputs=[b], device="cuda:0")
-
-    c = wp.zeros(n, device="cuda:1")
-    wp.launch(kernel, dim=c.size, inputs=[c], device="cuda:1")
-
-.. note::
-
-    A Warp CUDA device (``"cuda:i"``) corresponds to the primary CUDA context of device ``i``.  This is compatible with frameworks like PyTorch and other software that uses the CUDA Runtime API.  It makes interoperability easy, because GPU resources like memory can be shared with Warp.
-
-Default Device
---------------
-
-To simplify writing code, Warp has the concept of **default device**.  When the ``device`` argument is omitted from a Warp API call, the default device will be used.
-
-During Warp initialization, the default device is set to be ``"cuda:0"`` if CUDA is available.  Otherwise, the default device is ``"cpu"``.
-
-The function ``wp.set_device()`` can be used to change the default device::
-
-    wp.set_device("cpu")
-    a = wp.zeros(n)
-    wp.launch(kernel, dim=a.size, inputs=[a])
-   
-    wp.set_device("cuda:0")
-    b = wp.zeros(n)
-    wp.launch(kernel, dim=b.size, inputs=[b])
-   
-    wp.set_device("cuda:1")
-    c = wp.zeros(n)
-    wp.launch(kernel, dim=c.size, inputs=[c])
-
-.. note::
-
-    For CUDA devices, ``wp.set_device()`` does two things: it sets the Warp default device and it makes the device's CUDA context current.  This helps to minimize the number of CUDA context switches in blocks of code targeting a single device.
-
-For PyTorch users, this function is similar to ``torch.cuda.set_device()``.  It is still possible to specify a different device in individual API calls, like in this snippet::
-
-    # set default device
-    wp.set_device("cuda:0")
-   
-    # use default device
-    a = wp.zeros(n)
-   
-    # use explicit devices
-    b = wp.empty(n, device="cpu")
-    c = wp.empty(n, device="cuda:1")
-   
-    # use default device
-    wp.launch(kernel, dim=a.size, inputs=[a])
-   
-    wp.copy(b, a)
-    wp.copy(c, a)
-
-Scoped Devices
---------------
-
-Another way to manage the default device is using ``wp.ScopedDevice`` objects.  They can be arbitrarily nested and restore the previous default device on exit::
-
-    with wp.ScopedDevice("cpu"):
-        # alloc and launch on "cpu"
-        a = wp.zeros(n)
-        wp.launch(kernel, dim=a.size, inputs=[a])
- 
-    with wp.ScopedDevice("cuda:0"):
-        # alloc on "cuda:0"
-        b = wp.zeros(n)
-   
-        with wp.ScopedDevice("cuda:1"):
-            # alloc and launch on "cuda:1"
-            c = wp.zeros(n)
-            wp.launch(kernel, dim=c.size, inputs=[c])
-   
-        # launch on "cuda:0"
-        wp.launch(kernel, dim=b.size, inputs=[b])
-
-.. note::
-
-    For CUDA devices, ``wp.ScopedDevice`` makes the device's CUDA context current and restores the previous CUDA context on exit.  This is handy when running Warp scripts as part of a bigger pipeline, because it avoids any side effects of changing the CUDA context in the enclosed code.
-
-Example: Using ``wp.ScopedDevice`` with multiple GPUs
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-The following example shows how to allocate arrays and launch kernels on all available CUDA devices.
-
-.. code:: python
-
-    import warp as wp
-
-    wp.init()
-
-
-    @wp.kernel
-    def inc(a: wp.array(dtype=float)):
-        tid = wp.tid()
-        a[tid] = a[tid] + 1.0
-
-
-    # get all CUDA devices
-    devices = wp.get_cuda_devices()
-    device_count = len(devices)
-
-    # number of launches
-    iters = 1000
-
-    # list of arrays, one per device
-    arrs = []
-
-    # loop over all devices
-    for device in devices:
-        # use a ScopedDevice to set the target device
-        with wp.ScopedDevice(device):
-            # allocate array
-            a = wp.zeros(250 * 1024 * 1024, dtype=float)
-            arrs.append(a)
-
-            # launch kernels
-            for _ in range(iters):
-                wp.launch(inc, dim=a.size, inputs=[a])
-
-    # synchronize all devices
-    wp.synchronize()
-
-    # print results
-    for i in range(device_count):
-        print(f"{arrs[i].device} -> {arrs[i].numpy()}")
-
-
-Current CUDA Device
--------------------
-
-Warp uses the device alias ``"cuda"`` to target the current CUDA device.  This allows external code to manage the CUDA device on which to execute Warp scripts.  It is analogous to the PyTorch ``"cuda"`` device, which should be familiar to Torch users and simplify interoperation.
-
-In this snippet, we use PyTorch to manage the current CUDA device and invoke a Warp kernel on that device::
-
-    def example_function():
-        # create a Torch tensor on the current CUDA device
-        t = torch.arange(10, dtype=torch.float32, device="cuda")
-
-        a = wp.from_torch(t)
-
-        # launch a Warp kernel on the current CUDA device
-        wp.launch(kernel, dim=a.size, inputs=[a], device="cuda")
-
-    # use Torch to set the current CUDA device and run example_function() on that device
-    torch.cuda.set_device(0)
-    example_function()
-
-    # use Torch to change the current CUDA device and re-run example_function() on that device
-    torch.cuda.set_device(1)
-    example_function()
-
-.. note::
-
-    Using the device alias ``"cuda"`` can be problematic if the code runs in an environment where another part of the code can unpredictably change the CUDA context.  Using an explicit CUDA device like ``"cuda:i"`` is recommended to avoid such issues.
-
-Device Synchronization
-----------------------
-
-CUDA kernel launches and memory operations can execute asynchronously.  This allows for overlapping compute and memory operations on different devices.  Warp allows synchronizing the host with outstanding asynchronous operations on a specific device::
-
-    wp.synchronize_device("cuda:1")
-
-The ``wp.synchronize_device()`` function offers more fine-grained synchronization than ``wp.synchronize()``, as the latter waits for *all* devices to complete their work.
-
-Custom CUDA Contexts
---------------------
-
-Warp is designed to work with arbitrary CUDA contexts so it can easily integrate into different workflows.
-
-Applications built on the CUDA Runtime API target the *primary context* of each device.  The Runtime API hides CUDA context management under the hood.  In Warp, device ``"cuda:i"`` represents the primary context of device ``i``, which aligns with the CUDA Runtime API.
-
-Applications built on the CUDA Driver API work with CUDA contexts directly and can create custom CUDA contexts on any device.  Custom CUDA contexts can be created with specific affinity or interop features that benefit the application.  Warp can work with these CUDA contexts as well.
-
-The special device alias ``"cuda"`` can be used to target the current CUDA context, whether this is a primary or custom context.
-
-In addition, Warp allows registering new device aliases for custom CUDA contexts, so that they can be explicitly targeted by name.  If the ``CUcontext`` pointer is available, it can be used to create a new device alias like this::
-
-    wp.map_cuda_device("foo", ctypes.c_void_p(context_ptr))
-
-Alternatively, if the custom CUDA context was made current by the application, the pointer can be omitted::
-
-    wp.map_cuda_device("foo")
-
-In either case, mapping the custom CUDA context allows us to target the context directly using the assigned alias::
-
-    with wp.ScopedDevice("foo"):
-        a = wp.zeros(n)
-        wp.launch(kernel, dim=a.size, inputs=[a])
diff --git a/docs/_build/html/_sources/modules/fem.rst.txt b/docs/_build/html/_sources/modules/fem.rst.txt
deleted file mode 100644
index 359e42df9..000000000
--- a/docs/_build/html/_sources/modules/fem.rst.txt
+++ /dev/null
@@ -1,386 +0,0 @@
-warp.fem
-=====================
-
-.. currentmodule:: warp.fem
-
-The ``warp.fem`` module is designed to facilitate solving physical systems described as differential 
-equations. For example, it can solve PDEs for diffusion, convection, fluid flow, and elasticity problems 
-using finite-element-based (FEM) Galerkin methods, and allows users to quickly experiment with various FEM
-formulations and discretization schemes.
-
-Integrands
-----------
-
-The core functionality of the FEM toolkit is the ability to integrate constant, linear, and bilinear forms 
-over various domains and using arbitrary interpolation basis.
-
-The main mechanism is the :py:func:`.integrand` decorator, for instance: ::
-
-    @integrand
-    def linear_form(
-        s: Sample,
-        v: Field,
-    ):
-        return v(s)
-
-
-    @integrand
-    def diffusion_form(s: Sample, u: Field, v: Field, nu: float):
-        return nu * wp.dot(
-            grad(u, s),
-            grad(v, s),
-        )
-
-Integrands are normal warp kernels, meaning any usual warp function can be used. 
-However, they accept a few special parameters:
-
-  - :class:`.Sample` contains information about the current integration sample point, such as the element index and coordinates in element.
-  - :class:`.Field` designates an abstract field, which will be replaced at call time by the actual field type: for instance a :class:`.DiscreteField`, :class:`.field.TestField` or :class:`.field.TrialField` defined over an arbitrary :class:`.FunctionSpace`.
-    A field `u` can be evaluated at a given sample `s` using the :func:`.inner` operator, i.e, ``inner(u, s)``, or as a shortcut using the usual call operator, ``u(s)``.
-    Several other operators are available, such as :func:`.grad`; see the :ref:`Operators` section.
-  - :class:`.Domain` designates an abstract integration domain. Several operators are also provided for domains, for example in the example below evaluating the normal at the current sample position: ::
-    
-            @integrand
-            def boundary_form(
-                s: Sample,
-                domain: Domain,
-                u: Field,
-            ):
-                nor = normal(domain, s)
-                return wp.dot(u(s), nor)
-
-Integrands cannot be used directly with :func:`warp.launch`, but must be called through :func:`.integrate` or :func:`.interpolate` instead.
-The root integrand (`integrand` argument passed to :func:`integrate` or :func:`interpolate` call) will automatically get passed :class:`.Sample` and :class:`.Domain` parameters.
-:class:`.Field` parameters must be passed as a dictionary in the `fields` argument of the launcher function, and all other standard Warp types arguments must be
-passed as a dictionary in the `values` argument of the launcher function, for instance: ::
-    
-    integrate(diffusion_form, fields={"u": trial, "v": test}, values={"nu": viscosity})
-
-
-Basic workflow
---------------
-
-The typical steps for solving a linear PDE are as follow:
-
- - Define a :class:`.Geometry` (grid, mesh, etc). At the moment, 2D and 3D regular grids, triangle, quadrilateral, tetrahedron and hexahedron meshes are supported.
- - Define one or more :class:`.FunctionSpace`, by equipping the geometry elements with shape functions. See :func:`.make_polynomial_space`. At the moment, continuous/discontinuous Lagrange (:math:`P_{k[d]}, Q_{k[d]}`) and Serendipity (:math:`S_k`) shape functions of order :math:`k \leq 3` are supported.
- - Define an integration :class:`.GeometryDomain`, for instance the geometry's cells (:class:`.Cells`) or boundary sides (:class:`.BoundarySides`).
- - Integrate linear forms to build the system's right-hand-side. Define a test function over the function space using :func:`.make_test`,
-   a :class:`.Quadrature` formula (or let the module choose one based on the function space degree), and call :func:`.integrate` with the linear form integrand.
-   The result is a :class:`warp.array` containing the integration result for each of the function space degrees of freedom.
- - Integrate bilinear forms to build the system's left-hand-side. Define a trial function over the function space using :func:`.make_trial`,
-   then call :func:`.integrate` with the bilinear form integrand.
-   The result is a :class:`warp.sparse.BsrMatrix` containing the integration result for each pair of test and trial function space degrees of freedom.
-   Note that the trial and test functions do not have to be defined over the same function space, so that Mixed FEM is supported.
- - Solve the resulting linear system using the solver of your choice
-
-
-The following excerpt from the introductory example ``examples/fem/example_diffusion.py`` outlines this procedure: ::
-
-    # Grid geometry
-    geo = Grid2D(n=50, cell_size=2)
-
-    # Domain and function spaces
-    domain = Cells(geometry=geo)
-    scalar_space = make_polynomial_space(geo, degree=3)
-
-    # Right-hand-side (forcing term)
-    test = make_test(space=scalar_space, domain=domain)
-    rhs = integrate(linear_form, fields={"v": test})
-
-    # Weakly-imposed boundary conditions on Y sides
-    boundary = BoundarySides(geo)
-    bd_test = make_test(space=scalar_space, domain=boundary)
-    bd_trial = make_trial(space=scalar_space, domain=boundary)
-    bd_matrix = integrate(y_mass_form, fields={"u": bd_trial, "v": bd_test})
-
-    # Diffusion form
-    trial = make_trial(space=scalar_space, domain=domain)
-    matrix = integrate(diffusion_form, fields={"u": trial, "v": test}, values={"nu": viscosity})
-
-    # Assemble linear system (add diffusion and boundary condition matrices)
-    bsr_axpy(x=bd_matrix, y=matrix, alpha=boundary_strength, beta=1)
-
-    # Solve linear system using Conjugate Gradient
-    x = wp.zeros_like(rhs)
-    bsr_cg(matrix, b=rhs, x=x)
-
-
-.. note::
-    The :func:`.integrate` function does not check that the passed integrands are actually linear or bilinear forms; it is up to the user to ensure that they are.
-    To solve non-linear PDEs, one can use an iterative procedure and pass the current value of the studied function :class:`.DiscreteField` argument to the integrand, on which
-    arbitrary operations are permitted. However, the result of the form must remain linear in the test and trial fields.
-
-Introductory examples
----------------------
-
-``warp.fem`` ships with a list of examples in the ``examples/fem`` directory illustrating common model problems.
-
- - ``example_diffusion.py``: 2D diffusion with homogenous Neumann and Dirichlet boundary conditions
-     * ``example_diffusion_3d.py``: 3D variant of the diffusion problem
- - ``example_convection_diffusion.py``: 2D convection-diffusion using semi-Lagrangian advection
-     * ``example_diffusion_dg0.py``: 2D convection-diffusion using finite-volume and upwind transport
-     * ``example_diffusion_dg.py``: 2D convection-diffusion using Discontinuous Galerkin with upwind transport and Symmetric Interior Penalty
- - ``example_stokes.py``: 2D incompressible Stokes flow using mixed :math:`P_k/P_{k-1}` or :math:`Q_k/P_{(k-1)d}` elements
- - ``example_navier_stokes.py``: 2D Navier-Stokes flow using mixed :math:`P_k/P_{k-1}` elements
- - ``example_mixed_elasticity.py``: 2D linear elasticity using mixed continuous/discontinuous :math:`S_k/P_{(k-1)d}` elements
-
-
-Advanced usages
----------------
-
-High-order (curved) geometries
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-It is possible to convert any :class:`.Geometry` (grids and explicit meshes) into a curved, high-order variant by deforming them 
-with an arbitrary-order displacement field using the :meth:`~.DiscreteField.make_deformed_geometry` method. 
-The process looks as follow: ::
-
-   # Define a base geometry
-   base_geo = fem.Grid3D(res=resolution)
-
-   # Define a displacement field on the base geometry
-   deformation_space = fem.make_polynomial_space(base_geo, degree=deformation_degree, dtype=wp.vec3)
-   deformation_field = deformation_space.make_field()
-
-   # Populate the field value by interpolating an expression
-   fem.interpolate(deformation_field_expr, dest=deformation_field)
-
-   # Construct the deformed geometry from the displacement field
-   deform_geo = deformation_field.make_deformed_geometry()
-
-   # Define new function spaces on the deformed geometry
-   scalar_space = fem.make_polynomial_space(deformed_geo, degree=scalar_space_degree)
-
-See also ``example_deformed_geometry.py`` for a complete example.
-
-Particle-based quadrature
-^^^^^^^^^^^^^^^^^^^^^^^^^
-
-The :class:`.PicQuadrature` provides a way to define Particle-In-Cell quadratures from a set or arbitrary particles,
-which can be helpful to develop MPM-like methods.
-The particles are automatically bucketed to the geometry cells when the quadrature is initialized.
-This is illustrated by the ``example_stokes_transfer.py`` and ``example_apic_fluid.py`` examples.
-
-Partitioning
-^^^^^^^^^^^^
-
-The FEM toolkit makes it possible to perform integration on a subset of the domain elements, 
-possibly re-indexing degrees of freedom so that the linear system contains the local ones only.
-This is useful for distributed computation (see ``examples/fem/example_diffusion_mgpu.py``), or simply to limit the simulation domain to a subset of active cells (see ``examples/fem/example_stokes_transfer.py``).
-
-A partition of the simulation geometry can be defined using subclasses of :class:`.GeometryPartition`
-such as :class:`.LinearGeometryPartition`  or :class:`.ExplicitGeometryPartition`.
-
-Function spaces can then be partitioned according to the geometry partition using :func:`.make_space_partition`. 
-The resulting :class:`.SpacePartition` object allows translating between space-wide and partition-wide node indices, 
-and differentiating interior, frontier and exterior nodes.
-
-Memory management
-^^^^^^^^^^^^^^^^^
-
-Several ``warp.fem`` functions require allocating temporary buffers to perform their computations. 
-If such functions are called many times in a tight loop, those many allocations and de-allocations may degrade performance.
-To overcome this issue, a :class:`.cache.TemporaryStore` object may be created to persist and re-use temporary allocations across calls,
-either globally using :func:`set_default_temporary_store` or at a per-function granularity using the corresponding argument.
-
-.. _Operators:
-
-Operators
----------
-.. autofunction:: position(domain: Domain, s: Sample)
-.. autofunction:: normal(domain: Domain, s: Sample)
-.. autofunction:: lookup(domain: Domain, x)
-.. autofunction:: measure(domain: Domain, s: Sample)
-.. autofunction:: measure_ratio(domain: Domain, s: Sample)
-.. autofunction:: deformation_gradient(domain: Domain, s: Sample)
-
-.. autofunction:: degree(f: Field)
-.. autofunction:: inner(f: Field, s: Sample)
-.. autofunction:: outer(f: Field, s: Sample)
-.. autofunction:: grad(f: Field, s: Sample)
-.. autofunction:: grad_outer(f: Field, s: Sample)
-.. autofunction:: div(f: Field, s: Sample)
-.. autofunction:: div_outer(f: Field, s: Sample)
-.. autofunction:: at_node(f: Field, s: Sample)
-
-.. autofunction:: D(f: Field, s: Sample)
-.. autofunction:: curl(f: Field, s: Sample)
-.. autofunction:: jump(f: Field, s: Sample)
-.. autofunction:: average(f: Field, s: Sample)
-.. autofunction:: grad_jump(f: Field, s: Sample)
-.. autofunction:: grad_average(f: Field, s: Sample)
-
-.. autofunction:: warp.fem.operator.operator
-
-Integration
------------
-
-.. autofunction:: integrate
-.. autofunction:: interpolate
-
-.. autofunction:: integrand
-
-.. class:: Sample
-
-   Per-sample point context for evaluating fields and related operators in integrands.
-
-.. autoclass:: Field 
-
-.. autoclass:: Domain 
-
-Geometry
---------
-
-.. autoclass:: Grid2D
-   :show-inheritance:
-
-.. autoclass:: Trimesh2D
-   :show-inheritance:
-
-.. autoclass:: Quadmesh2D
-   :show-inheritance:
-
-.. autoclass:: Grid3D
-   :show-inheritance:
-
-.. autoclass:: Tetmesh
-   :show-inheritance:
-
-.. autoclass:: Hexmesh
-   :show-inheritance:
-
-.. autoclass:: LinearGeometryPartition
-
-.. autoclass:: ExplicitGeometryPartition
-
-.. autoclass:: Cells
-   :show-inheritance:
-
-.. autoclass:: Sides
-   :show-inheritance:
-
-.. autoclass:: BoundarySides
-   :show-inheritance:
-
-.. autoclass:: FrontierSides
-   :show-inheritance:
-
-.. autoclass:: Polynomial
-   :members:
-
-.. autoclass:: RegularQuadrature
-   :show-inheritance:
-
-.. autoclass:: NodalQuadrature
-   :show-inheritance:
-
-.. autoclass:: PicQuadrature
-   :show-inheritance:
-
-Function Spaces
----------------
-
-.. autofunction:: make_polynomial_space
-
-.. autofunction:: make_polynomial_basis_space
-
-.. autofunction:: make_collocated_function_space
-
-.. autofunction:: make_space_partition
-
-.. autofunction:: make_space_restriction
-
-.. autoclass:: ElementBasis
-   :members:
-
-.. autoclass:: SymmetricTensorMapper
-   :show-inheritance:
-
-.. autoclass:: SkewSymmetricTensorMapper
-   :show-inheritance:
-
-Fields
-------
-
-.. autofunction:: make_test
-
-.. autofunction:: make_trial
-
-.. autofunction:: make_restriction
-
-Boundary Conditions
--------------------
-
-.. autofunction:: normalize_dirichlet_projector
-
-.. autofunction:: project_linear_system
-
-Memory management
------------------
-
-.. autofunction:: set_default_temporary_store
-
-.. autofunction:: borrow_temporary
-
-.. autofunction:: borrow_temporary_like
-
-
-Interfaces
-----------
-
-Interface classes are not meant to be constructed directly, but can be derived from extend the built-in functionality.
-
-.. autoclass:: Geometry
-   :members: cell_count, side_count, boundary_side_count
-
-.. autoclass:: GeometryPartition
-   :members: cell_count, side_count, boundary_side_count, frontier_side_count
-
-.. autoclass:: GeometryDomain
-   :members: ElementKind, element_kind, dimension, element_count
-
-.. autoclass:: Quadrature
-   :members: domain, total_point_count
-
-.. autoclass:: FunctionSpace
-   :members: dtype, topology, geometry, dimension, degree, trace, make_field
-
-.. autoclass:: SpaceTopology
-   :members: dimension, geometry, node_count, element_node_indices, trace
-
-.. autoclass:: BasisSpace
-   :members: topology, geometry, shape, node_positions
-
-.. autoclass:: warp.fem.space.shape.ShapeFunction
-
-.. autoclass:: SpacePartition
-   :members: node_count, owned_node_count, interior_node_count, space_node_indices
-
-.. autoclass:: SpaceRestriction
-   :members: node_count
-
-.. autoclass:: DofMapper
-
-.. autoclass:: FieldLike
-
-.. autoclass:: DiscreteField
-   :show-inheritance:
-   :members: dof_values, trace, make_deformed_geometry
-
-.. autoclass:: warp.fem.field.FieldRestriction
-
-.. autoclass:: warp.fem.field.SpaceField
-   :show-inheritance:
-
-.. autoclass:: warp.fem.field.TestField
-   :show-inheritance:
-
-.. autoclass:: warp.fem.field.TrialField
-   :show-inheritance:
-
-.. autoclass:: TemporaryStore
-   :members: clear
-
-.. autoclass:: warp.fem.cache.Temporary
-   :members: array, detach, release
\ No newline at end of file
diff --git a/docs/_build/html/_sources/modules/functions.rst.txt b/docs/_build/html/_sources/modules/functions.rst.txt
deleted file mode 100644
index 5fe56ada3..000000000
--- a/docs/_build/html/_sources/modules/functions.rst.txt
+++ /dev/null
@@ -1,1992 +0,0 @@
-..
-   Autogenerated File - Do not edit. Run build_docs.py to generate.
-
-.. functions:
-.. currentmodule:: warp
-
-Kernel Reference
-================
-
-Scalar Types
-------------
-.. class:: int8
-.. class:: uint8
-.. class:: int16
-.. class:: uint16
-.. class:: int32
-.. class:: uint32
-.. class:: int64
-.. class:: uint64
-.. class:: float16
-.. class:: float32
-.. class:: float64
-.. class:: bool
-
-
-Vector Types
-------------
-.. class:: vec2b
-.. class:: vec2ub
-.. class:: vec2s
-.. class:: vec2us
-.. class:: vec2i
-.. class:: vec2ui
-.. class:: vec2l
-.. class:: vec2ul
-.. class:: vec2h
-.. class:: vec2f
-.. class:: vec2d
-.. class:: vec3b
-.. class:: vec3ub
-.. class:: vec3s
-.. class:: vec3us
-.. class:: vec3i
-.. class:: vec3ui
-.. class:: vec3l
-.. class:: vec3ul
-.. class:: vec3h
-.. class:: vec3f
-.. class:: vec3d
-.. class:: vec4b
-.. class:: vec4ub
-.. class:: vec4s
-.. class:: vec4us
-.. class:: vec4i
-.. class:: vec4ui
-.. class:: vec4l
-.. class:: vec4ul
-.. class:: vec4h
-.. class:: vec4f
-.. class:: vec4d
-.. class:: mat22h
-.. class:: mat22f
-.. class:: mat22d
-.. class:: mat33h
-.. class:: mat33f
-.. class:: mat33d
-.. class:: mat44h
-.. class:: mat44f
-.. class:: mat44d
-.. class:: quath
-.. class:: quatf
-.. class:: quatd
-.. class:: transformh
-.. class:: transformf
-.. class:: transformd
-.. class:: spatial_vectorh
-.. class:: spatial_vectorf
-.. class:: spatial_vectord
-.. class:: spatial_matrixh
-.. class:: spatial_matrixf
-.. class:: spatial_matrixd
-
-Generic Types
--------------
-.. class:: Int
-.. class:: Float
-.. class:: Scalar
-.. class:: Vector
-.. class:: Matrix
-.. class:: Quaternion
-.. class:: Transformation
-.. class:: Array
-
-
-Scalar Math
----------------
-.. function:: min(x: Scalar, y: Scalar) -> Scalar
-
-   Return the minimum of two scalars.
-
-
-.. function:: min(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Return the element-wise minimum of two vectors.
-
-
-.. function:: min(v: Vector[Any,Scalar]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Return the minimum element of a vector ``v``.
-
-
-.. function:: max(x: Scalar, y: Scalar) -> Scalar
-
-   Return the maximum of two scalars.
-
-
-.. function:: max(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Return the element-wise maximum of two vectors.
-
-
-.. function:: max(v: Vector[Any,Scalar]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Return the maximum element of a vector ``v``.
-
-
-.. function:: clamp(x: Scalar, a: Scalar, b: Scalar) -> Scalar
-
-   Clamp the value of ``x`` to the range [a, b].
-
-
-.. function:: abs(x: Scalar) -> Scalar
-
-   Return the absolute value of ``x``.
-
-
-.. function:: sign(x: Scalar) -> Scalar
-
-   Return -1 if ``x`` < 0, return 1 otherwise.
-
-
-.. function:: step(x: Scalar) -> Scalar
-
-   Return 1.0 if ``x`` < 0.0, return 0.0 otherwise.
-
-
-.. function:: nonzero(x: Scalar) -> Scalar
-
-   Return 1.0 if ``x`` is not equal to zero, return 0.0 otherwise.
-
-
-.. function:: sin(x: Float) -> Float
-
-   Return the sine of ``x`` in radians.
-
-
-.. function:: cos(x: Float) -> Float
-
-   Return the cosine of ``x`` in radians.
-
-
-.. function:: acos(x: Float) -> Float
-
-   Return arccos of ``x`` in radians. Inputs are automatically clamped to [-1.0, 1.0].
-
-
-.. function:: asin(x: Float) -> Float
-
-   Return arcsin of ``x`` in radians. Inputs are automatically clamped to [-1.0, 1.0].
-
-
-.. function:: sqrt(x: Float) -> Float
-
-   Return the square root of ``x``, where ``x`` is positive.
-
-
-.. function:: cbrt(x: Float) -> Float
-
-   Return the cube root of ``x``.
-
-
-.. function:: tan(x: Float) -> Float
-
-   Return the tangent of ``x`` in radians.
-
-
-.. function:: atan(x: Float) -> Float
-
-   Return the arctangent of ``x`` in radians.
-
-
-.. function:: atan2(y: Float, x: Float) -> Float
-
-   Return the 2-argument arctangent, atan2, of the point ``(x, y)`` in radians.
-
-
-.. function:: sinh(x: Float) -> Float
-
-   Return the sinh of ``x``.
-
-
-.. function:: cosh(x: Float) -> Float
-
-   Return the cosh of ``x``.
-
-
-.. function:: tanh(x: Float) -> Float
-
-   Return the tanh of ``x``.
-
-
-.. function:: degrees(x: Float) -> Float
-
-   Convert ``x`` from radians into degrees.
-
-
-.. function:: radians(x: Float) -> Float
-
-   Convert ``x`` from degrees into radians.
-
-
-.. function:: log(x: Float) -> Float
-
-   Return the natural logarithm (base-e) of ``x``, where ``x`` is positive.
-
-
-.. function:: log2(x: Float) -> Float
-
-   Return the binary logarithm (base-2) of ``x``, where ``x`` is positive.
-
-
-.. function:: log10(x: Float) -> Float
-
-   Return the common logarithm (base-10) of ``x``, where ``x`` is positive.
-
-
-.. function:: exp(x: Float) -> Float
-
-   Return the value of the exponential function :math:`e^x`.
-
-
-.. function:: pow(x: Float, y: Float) -> Float
-
-   Return the result of ``x`` raised to power of ``y``.
-
-
-.. function:: round(x: Float) -> Float
-
-   Return the nearest integer value to ``x``, rounding halfway cases away from zero.
-   This is the most intuitive form of rounding in the colloquial sense, but can be slower than other options like :func:`warp.rint()`.
-   Differs from :func:`numpy.round()`, which behaves the same way as :func:`numpy.rint()`.
-
-
-.. function:: rint(x: Float) -> Float
-
-   Return the nearest integer value to ``x``, rounding halfway cases to nearest even integer.
-   It is generally faster than :func:`warp.round()`. Equivalent to :func:`numpy.rint()`.
-
-
-.. function:: trunc(x: Float) -> Float
-
-   Return the nearest integer that is closer to zero than ``x``.
-   In other words, it discards the fractional part of ``x``.
-   It is similar to casting ``float(int(x))``, but preserves the negative sign when x is in the range [-0.0, -1.0).
-   Equivalent to :func:`numpy.trunc()` and :func:`numpy.fix()`.
-
-
-.. function:: floor(x: Float) -> Float
-
-   Return the largest integer that is less than or equal to ``x``.
-
-
-.. function:: ceil(x: Float) -> Float
-
-   Return the smallest integer that is greater than or equal to ``x``.
-
-
-.. function:: frac(x: Float) -> Float
-
-   Retrieve the fractional part of x.
-    In other words, it discards the integer part of x and is equivalent to ``x - trunc(x)``.
-
-
-
-
-Vector Math
----------------
-.. function:: dot(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Scalar
-
-   Compute the dot product between two vectors.
-
-
-.. function:: dot(x: Quaternion[Float], y: Quaternion[Float]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Compute the dot product between two quaternions.
-
-
-.. function:: ddot(x: Matrix[Any,Any,Scalar], y: Matrix[Any,Any,Scalar]) -> Scalar
-
-   Compute the double dot product between two matrices.
-
-
-.. function:: argmin(v: Vector[Any,Scalar]) -> uint32
-
-   Return the index of the minimum element of a vector ``v``. [1]_
-
-
-.. function:: argmax(v: Vector[Any,Scalar]) -> uint32
-
-   Return the index of the maximum element of a vector ``v``. [1]_
-
-
-.. function:: outer(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Matrix[Any,Any,Scalar]
-
-   Compute the outer product ``x*y^T`` for two vectors.
-
-
-.. function:: cross(x: Vector[3,Scalar], y: Vector[3,Scalar]) -> Vector[3,Scalar]
-
-   Compute the cross product of two 3D vectors.
-
-
-.. function:: skew(x: Vector[3,Scalar])
-
-   Compute the skew-symmetric 3x3 matrix for a 3D vector ``x``.
-
-
-.. function:: length(x: Vector[Any,Float]) -> Scalar
-
-   Compute the length of a vector ``x``.
-
-
-.. function:: length(x: Quaternion[Float]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Compute the length of a quaternion ``x``.
-
-
-.. function:: length_sq(x: Vector[Any,Scalar]) -> Scalar
-
-   Compute the squared length of a 2D vector ``x``.
-
-
-.. function:: length_sq(x: Quaternion[Scalar]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Compute the squared length of a quaternion ``x``.
-
-
-.. function:: normalize(x: Vector[Any,Float]) -> Vector[Any,Scalar]
-
-   Compute the normalized value of ``x``. If ``length(x)`` is 0 then the zero vector is returned.
-
-
-.. function:: normalize(x: Quaternion[Float]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Compute the normalized value of ``x``. If ``length(x)`` is 0, then the zero quaternion is returned.
-
-
-.. function:: transpose(m: Matrix[Any,Any,Scalar])
-
-   Return the transpose of the matrix ``m``.
-
-
-.. function:: inverse(m: Matrix[2,2,Float]) -> Matrix[Any,Any,Float]
-
-   Return the inverse of a 2x2 matrix ``m``.
-
-
-.. function:: inverse(m: Matrix[3,3,Float]) -> Matrix[Any,Any,Float]
-   :noindex:
-   :nocontentsentry:
-
-   Return the inverse of a 3x3 matrix ``m``.
-
-
-.. function:: inverse(m: Matrix[4,4,Float]) -> Matrix[Any,Any,Float]
-   :noindex:
-   :nocontentsentry:
-
-   Return the inverse of a 4x4 matrix ``m``.
-
-
-.. function:: determinant(m: Matrix[2,2,Float]) -> Scalar
-
-   Return the determinant of a 2x2 matrix ``m``.
-
-
-.. function:: determinant(m: Matrix[3,3,Float]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Return the determinant of a 3x3 matrix ``m``.
-
-
-.. function:: determinant(m: Matrix[4,4,Float]) -> Scalar
-   :noindex:
-   :nocontentsentry:
-
-   Return the determinant of a 4x4 matrix ``m``.
-
-
-.. function:: trace(m: Matrix[Any,Any,Scalar]) -> Scalar
-
-   Return the trace of the matrix ``m``.
-
-
-.. function:: diag(d: Vector[Any,Scalar]) -> Matrix[Any,Any,Scalar]
-
-   Returns a matrix with the components of the vector ``d`` on the diagonal.
-
-
-.. function:: get_diag(m: Matrix[Any,Any,Scalar]) -> Vector[Any,Scalar]
-
-   Returns a vector containing the diagonal elements of the square matrix ``m``.
-
-
-.. function:: cw_mul(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-
-   Component-wise multiplication of two 2D vectors.
-
-
-.. function:: cw_mul(x: Matrix[Any,Any,Scalar], y: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Component-wise multiplication of two 2D vectors.
-
-
-.. function:: cw_div(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-
-   Component-wise division of two 2D vectors.
-
-
-.. function:: cw_div(x: Matrix[Any,Any,Scalar], y: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Component-wise division of two 2D vectors.
-
-
-.. function:: vector(w: Vector[3,Float], v: Vector[3,Float])
-
-   Construct a 6D screw vector from two 3D vectors.
-
-
-.. function:: vector(*arg_types: Scalar, length: int32, dtype: Scalar) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Construct a vector of with given length and dtype.
-
-
-.. function:: matrix(pos: Vector[3,Float], rot: Quaternion[Float], scale: Vector[3,Float]) -> Matrix[Any,Any,Float]
-
-   Construct a 4x4 transformation matrix that applies the transformations as
-   Translation(pos)*Rotation(rot)*Scale(scale) when applied to column vectors, i.e.: y = (TRS)*x
-
-
-.. function:: matrix(*arg_types: Scalar, shape: Tuple[int, int], dtype: Scalar) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Construct a matrix. If the positional ``arg_types`` are not given, then matrix will be zero-initialized.
-
-
-.. function:: identity(n: int32, dtype: Scalar) -> Matrix[Any,Any,Scalar]
-
-   Create an identity matrix with shape=(n,n) with the type given by ``dtype``.
-
-
-.. function:: svd3(A: Matrix[3,3,Float], U: Matrix[3,3,Float], sigma: Vector[3,Float], V: Matrix[3,3,Scalar]) -> None
-
-   Compute the SVD of a 3x3 matrix ``A``. The singular values are returned in ``sigma``,
-   while the left and right basis vectors are returned in ``U`` and ``V``.
-
-
-.. function:: qr3(A: Matrix[3,3,Float], Q: Matrix[3,3,Float], R: Matrix[3,3,Float]) -> None
-
-   Compute the QR decomposition of a 3x3 matrix ``A``. The orthogonal matrix is returned in ``Q``,
-   while the upper triangular matrix is returned in ``R``.
-
-
-.. function:: eig3(A: Matrix[3,3,Float], Q: Matrix[3,3,Float], d: Vector[3,Float]) -> None
-
-   Compute the eigendecomposition of a 3x3 matrix ``A``. The eigenvectors are returned as the columns of ``Q``,
-   while the corresponding eigenvalues are returned in ``d``.
-
-
-
-
-Other
----------------
-.. function:: lower_bound(arr: Array[Scalar], value: Scalar) -> int
-
-   Search a sorted array ``arr`` for the closest element greater than or equal to ``value``.
-
-
-.. function:: lower_bound(arr: Array[Scalar], arr_begin: int32, arr_end: int32, value: Scalar) -> int
-   :noindex:
-   :nocontentsentry:
-
-   Search a sorted array ``arr`` in the range [arr_begin, arr_end) for the closest element greater than or equal to ``value``.
-
-
-
-
-Quaternion Math
----------------
-.. function:: quaternion() -> Quaternion[Float]
-
-   Construct a zero-initialized quaternion. Quaternions are laid out as
-   [ix, iy, iz, r], where ix, iy, iz are the imaginary part, and r the real part.
-
-
-.. function:: quaternion(x: Float, y: Float, z: Float, w: Float) -> Quaternion[Float]
-   :noindex:
-   :nocontentsentry:
-
-   Create a quaternion using the supplied components (type inferred from component type).
-
-
-.. function:: quaternion(i: Vector[3,Float], r: Float) -> Quaternion[Float]
-   :noindex:
-   :nocontentsentry:
-
-   Create a quaternion using the supplied vector/scalar (type inferred from scalar type).
-
-
-.. function:: quaternion(q: Quaternion[Float])
-   :noindex:
-   :nocontentsentry:
-
-   Construct a quaternion of type dtype from another quaternion of a different dtype.
-
-
-.. function:: quat_identity() -> quatf
-
-   Construct an identity quaternion with zero imaginary part and real part of 1.0
-
-
-.. function:: quat_from_axis_angle(axis: Vector[3,Float], angle: Float) -> Quaternion[Scalar]
-
-   Construct a quaternion representing a rotation of angle radians around the given axis.
-
-
-.. function:: quat_to_axis_angle(q: Quaternion[Float], axis: Vector[3,Float], angle: Float) -> None
-
-   Extract the rotation axis and angle radians a quaternion represents.
-
-
-.. function:: quat_from_matrix(m: Matrix[3,3,Float]) -> Quaternion[Scalar]
-
-   Construct a quaternion from a 3x3 matrix.
-
-
-.. function:: quat_rpy(roll: Float, pitch: Float, yaw: Float) -> Quaternion[Scalar]
-
-   Construct a quaternion representing a combined roll (z), pitch (x), yaw rotations (y) in radians.
-
-
-.. function:: quat_inverse(q: Quaternion[Float]) -> Quaternion[Scalar]
-
-   Compute quaternion conjugate.
-
-
-.. function:: quat_rotate(q: Quaternion[Float], p: Vector[3,Float]) -> Vector[3,Scalar]
-
-   Rotate a vector by a quaternion.
-
-
-.. function:: quat_rotate_inv(q: Quaternion[Float], p: Vector[3,Float]) -> Vector[3,Scalar]
-
-   Rotate a vector by the inverse of a quaternion.
-
-
-.. function:: quat_slerp(q0: Quaternion[Float], q1: Quaternion[Float], t: Float) -> Quaternion[Scalar]
-
-   Linearly interpolate between two quaternions.
-
-
-.. function:: quat_to_matrix(q: Quaternion[Float]) -> Matrix[3,3,Scalar]
-
-   Convert a quaternion to a 3x3 rotation matrix.
-
-
-
-
-Transformations
----------------
-.. function:: transformation(p: Vector[3,Float], q: Quaternion[Float]) -> Transformation[Scalar]
-
-   Construct a rigid-body transformation with translation part ``p`` and rotation ``q``.
-
-
-.. function:: transform_identity() -> transformf
-
-   Construct an identity transform with zero translation and identity rotation.
-
-
-.. function:: transform_get_translation(t: Transformation[Float]) -> Vector[3,Scalar]
-
-   Return the translational part of a transform ``t``.
-
-
-.. function:: transform_get_rotation(t: Transformation[Float]) -> Quaternion[Scalar]
-
-   Return the rotational part of a transform ``t``.
-
-
-.. function:: transform_multiply(a: Transformation[Float], b: Transformation[Float]) -> Transformation[Scalar]
-
-   Multiply two rigid body transformations together.
-
-
-.. function:: transform_point(t: Transformation[Scalar], p: Vector[3,Scalar]) -> Vector[3,Scalar]
-
-   Apply the transform to a point ``p`` treating the homogenous coordinate as w=1 (translation and rotation).
-
-
-.. function:: transform_point(m: Matrix[4,4,Scalar], p: Vector[3,Scalar]) -> Vector[3,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Apply the transform to a point ``p`` treating the homogenous coordinate as w=1.
-   The transformation is applied treating ``p`` as a column vector, e.g.: ``y = M*p``.
-   Note this is in contrast to some libraries, notably USD, which applies transforms to row vectors, ``y^T = p^T*M^T``.
-   If the transform is coming from a library that uses row-vectors, then users should transpose the transformation
-   matrix before calling this method.
-
-
-.. function:: transform_vector(t: Transformation[Scalar], v: Vector[3,Scalar]) -> Vector[3,Scalar]
-
-   Apply the transform to a vector ``v`` treating the homogenous coordinate as w=0 (rotation only).
-
-
-.. function:: transform_vector(m: Matrix[4,4,Scalar], v: Vector[3,Scalar]) -> Vector[3,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-   Apply the transform to a vector ``v`` treating the homogenous coordinate as w=0.
-   The transformation is applied treating ``v`` as a column vector, e.g.: ``y = M*v``
-   note this is in contrast to some libraries, notably USD, which applies transforms to row vectors, ``y^T = v^T*M^T``.
-   If the transform is coming from a library that uses row-vectors, then users should transpose the transformation
-   matrix before calling this method.
-
-
-.. function:: transform_inverse(t: Transformation[Float]) -> Transformation[Float]
-
-   Compute the inverse of the transformation ``t``.
-
-
-
-
-Spatial Math
----------------
-.. function:: spatial_adjoint(r: Matrix[3,3,Float], s: Matrix[3,3,Float]) -> Matrix[6,6,Scalar]
-
-   Construct a 6x6 spatial inertial matrix from two 3x3 diagonal blocks.
-
-
-.. function:: spatial_dot(a: Vector[6,Float], b: Vector[6,Float]) -> Scalar
-
-   Compute the dot product of two 6D screw vectors.
-
-
-.. function:: spatial_cross(a: Vector[6,Float], b: Vector[6,Float]) -> Vector[6,Float]
-
-   Compute the cross product of two 6D screw vectors.
-
-
-.. function:: spatial_cross_dual(a: Vector[6,Float], b: Vector[6,Float]) -> Vector[6,Float]
-
-   Compute the dual cross product of two 6D screw vectors.
-
-
-.. function:: spatial_top(a: Vector[6,Float])
-
-   Return the top (first) part of a 6D screw vector.
-
-
-.. function:: spatial_bottom(a: Vector[6,Float])
-
-   Return the bottom (second) part of a 6D screw vector.
-
-
-.. function:: spatial_jacobian(S: Array[Vector[6,Float]], joint_parents: Array[int32], joint_qd_start: Array[int32], joint_start: int32, joint_count: int32, J_start: int32, J_out: Array[Float]) -> None
-
-
-.. function:: spatial_mass(I_s: Array[Matrix[6,6,Float]], joint_start: int32, joint_count: int32, M_start: int32, M: Array[Float]) -> None
-
-
-
-
-Utility
----------------
-.. function:: mlp(weights: Array[float32], bias: Array[float32], activation: Callable, index: int32, x: Array[float32], out: Array[float32]) -> None
-
-   Evaluate a multi-layer perceptron (MLP) layer in the form: ``out = act(weights*x + bias)``.
-
-   :param weights: A layer's network weights with dimensions ``(m, n)``.
-   :param bias: An array with dimensions ``(n)``.
-   :param activation: A ``wp.func`` function that takes a single scalar float as input and returns a scalar float as output
-   :param index: The batch item to process, typically each thread will process one item in the batch, in which case
-                 index should be ``wp.tid()``
-   :param x: The feature matrix with dimensions ``(n, b)``
-   :param out: The network output with dimensions ``(m, b)``
-
-   :note: Feature and output matrices are transposed compared to some other frameworks such as PyTorch.
-          All matrices are assumed to be stored in flattened row-major memory layout (NumPy default).
-
-
-.. function:: printf() -> None
-
-   Allows printing formatted strings using C-style format specifiers.
-
-
-.. function:: print(value: Any) -> None
-
-   Print variable to stdout
-
-
-.. function:: breakpoint() -> None
-
-   Debugger breakpoint
-
-
-.. function:: tid() -> int
-
-   Return the current thread index for a 1D kernel launch. Note that this is the *global* index of the thread in the range [0, dim)
-   where dim is the parameter passed to kernel launch. This function may not be called from user-defined Warp functions.
-
-
-.. function:: tid() -> Tuple[int, int]
-   :noindex:
-   :nocontentsentry:
-
-   Return the current thread indices for a 2D kernel launch. Use ``i,j = wp.tid()`` syntax to retrieve the
-   coordinates inside the kernel thread grid. This function may not be called from user-defined Warp functions.
-
-
-.. function:: tid() -> Tuple[int, int, int]
-   :noindex:
-   :nocontentsentry:
-
-   Return the current thread indices for a 3D kernel launch. Use ``i,j,k = wp.tid()`` syntax to retrieve the
-   coordinates inside the kernel thread grid. This function may not be called from user-defined Warp functions.
-
-
-.. function:: tid() -> Tuple[int, int, int, int]
-   :noindex:
-   :nocontentsentry:
-
-   Return the current thread indices for a 4D kernel launch. Use ``i,j,k,l = wp.tid()`` syntax to retrieve the
-   coordinates inside the kernel thread grid. This function may not be called from user-defined Warp functions.
-
-
-.. function:: select(cond: bool, arg1: Any, arg2: Any)
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: bool, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: int8, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: uint8, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: int16, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: uint16, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: int32, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: uint32, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: int64, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(cond: uint64, arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``cond`` is ``False`` then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: select(arr: Array[Any], arg1: Any, arg2: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Select between two arguments, if ``arr`` is null then return ``arg1``, otherwise return ``arg2``
-
-
-.. function:: atomic_add(a: Array[Any], i: int32, value: Any)
-
-   Atomically add ``value`` onto ``a[i]``.
-
-
-.. function:: atomic_add(a: Array[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j]``.
-
-
-.. function:: atomic_add(a: Array[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j,k]``.
-
-
-.. function:: atomic_add(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j,k,l]``.
-
-
-.. function:: atomic_add(a: FabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i]``.
-
-
-.. function:: atomic_add(a: FabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j]``.
-
-
-.. function:: atomic_add(a: FabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j,k]``.
-
-
-.. function:: atomic_add(a: FabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j,k,l]``.
-
-
-.. function:: atomic_add(a: IndexedFabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i]``.
-
-
-.. function:: atomic_add(a: IndexedFabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j]``.
-
-
-.. function:: atomic_add(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j,k]``.
-
-
-.. function:: atomic_add(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically add ``value`` onto ``a[i,j,k,l]``.
-
-
-.. function:: atomic_sub(a: Array[Any], i: int32, value: Any)
-
-   Atomically subtract ``value`` onto ``a[i]``.
-
-
-.. function:: atomic_sub(a: Array[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j]``.
-
-
-.. function:: atomic_sub(a: Array[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j,k]``.
-
-
-.. function:: atomic_sub(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j,k,l]``.
-
-
-.. function:: atomic_sub(a: FabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i]``.
-
-
-.. function:: atomic_sub(a: FabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j]``.
-
-
-.. function:: atomic_sub(a: FabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j,k]``.
-
-
-.. function:: atomic_sub(a: FabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j,k,l]``.
-
-
-.. function:: atomic_sub(a: IndexedFabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i]``.
-
-
-.. function:: atomic_sub(a: IndexedFabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j]``.
-
-
-.. function:: atomic_sub(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j,k]``.
-
-
-.. function:: atomic_sub(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Atomically subtract ``value`` onto ``a[i,j,k,l]``.
-
-
-.. function:: atomic_min(a: Array[Any], i: int32, value: Any)
-
-   Compute the minimum of ``value`` and ``a[i]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: Array[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: Array[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j,k]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: FabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: FabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: FabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j,k]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: FabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: IndexedFabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: IndexedFabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j,k]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_min(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the minimum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: Array[Any], i: int32, value: Any)
-
-   Compute the maximum of ``value`` and ``a[i]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: Array[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: Array[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j,k]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: FabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: FabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: FabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j,k]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: FabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: IndexedFabricArray[Any], i: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: IndexedFabricArray[Any], i: int32, j: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j,k]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: atomic_max(a: IndexedFabricArray[Any], i: int32, j: int32, k: int32, l: int32, value: Any)
-   :noindex:
-   :nocontentsentry:
-
-   Compute the maximum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
-
-Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-
-.. function:: lerp(a: Float, b: Float, t: Float) -> Float
-
-   Linearly interpolate two values ``a`` and ``b`` using factor ``t``, computed as ``a*(1-t) + b*t``
-
-
-.. function:: lerp(a: Vector[Any,Float], b: Vector[Any,Float], t: Float) -> Vector[Any,Float]
-   :noindex:
-   :nocontentsentry:
-
-   Linearly interpolate two values ``a`` and ``b`` using factor ``t``, computed as ``a*(1-t) + b*t``
-
-
-.. function:: lerp(a: Matrix[Any,Any,Float], b: Matrix[Any,Any,Float], t: Float) -> Matrix[Any,Any,Float]
-   :noindex:
-   :nocontentsentry:
-
-   Linearly interpolate two values ``a`` and ``b`` using factor ``t``, computed as ``a*(1-t) + b*t``
-
-
-.. function:: lerp(a: Quaternion[Float], b: Quaternion[Float], t: Float) -> Quaternion[Float]
-   :noindex:
-   :nocontentsentry:
-
-   Linearly interpolate two values ``a`` and ``b`` using factor ``t``, computed as ``a*(1-t) + b*t``
-
-
-.. function:: lerp(a: Transformation[Float], b: Transformation[Float], t: Float) -> Transformation[Float]
-   :noindex:
-   :nocontentsentry:
-
-   Linearly interpolate two values ``a`` and ``b`` using factor ``t``, computed as ``a*(1-t) + b*t``
-
-
-.. function:: smoothstep(edge0: Float, edge1: Float, x: Float) -> Float
-
-   Smoothly interpolate between two values ``edge0`` and ``edge1`` using a factor ``x``,
-   and return a result between 0 and 1 using a cubic Hermite interpolation after clamping.
-
-
-.. function:: expect_near(arg1: Float, arg2: Float, tolerance: Float) -> None
-
-   Prints an error to stdout if ``arg1`` and ``arg2`` are not closer than tolerance in magnitude
-
-
-.. function:: expect_near(arg1: vec3f, arg2: vec3f, tolerance: float32) -> None
-   :noindex:
-   :nocontentsentry:
-
-   Prints an error to stdout if any element of ``arg1`` and ``arg2`` are not closer than tolerance in magnitude
-
-
-
-
-Geometry
----------------
-.. function:: bvh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> bvh_query_t
-
-   Construct an axis-aligned bounding box query against a BVH object. This query can be used to iterate over all bounds
-   inside a BVH. Returns an object that is used to track state during BVH traversal.
-
-   :param id: The BVH identifier
-   :param lower: The lower bound of the bounding box in BVH space
-   :param upper: The upper bound of the bounding box in BVH space
-
-
-.. function:: bvh_query_ray(id: uint64, start: vec3f, dir: vec3f) -> bvh_query_t
-
-   Construct a ray query against a BVH object. This query can be used to iterate over all bounds
-   that intersect the ray. Returns an object that is used to track state during BVH traversal.
-
-   :param id: The BVH identifier
-   :param start: The start of the ray in BVH space
-   :param dir: The direction of the ray in BVH space
-
-
-.. function:: bvh_query_next(query: bvh_query_t, index: int32) -> bool
-
-   Move to the next bound returned by the query.
-   The index of the current bound is stored in ``index``, returns ``False`` if there are no more overlapping bound.
-
-
-.. function:: mesh_query_point(id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32) -> bool
-
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
-
-   Identifies the sign of the distance using additional ray-casts to determine if the point is inside or outside.
-   This method is relatively robust, but does increase computational cost.
-   See below for additional sign determination methods.
-
-   :param id: The mesh identifier
-   :param point: The point in space to query
-   :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                  Note that mesh must be watertight for this to be robust
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
-
-
-.. function:: mesh_query_point_no_sign(id: uint64, point: vec3f, max_dist: float32, face: int32, bary_u: float32, bary_v: float32) -> bool
-
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
-
-   This method does not compute the sign of the point (inside/outside) which makes it faster than other point query methods.
-
-   :param id: The mesh identifier
-   :param point: The point in space to query
-   :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
-
-
-.. function:: mesh_query_furthest_point_no_sign(id: uint64, point: vec3f, min_dist: float32, face: int32, bary_u: float32, bary_v: float32) -> bool
-
-   Computes the furthest point on the mesh with identifier `id` to the given point in space. Returns ``True`` if a point > ``min_dist`` is found.
-
-   This method does not compute the sign of the point (inside/outside).
-
-   :param id: The mesh identifier
-   :param point: The point in space to query
-   :param min_dist: Mesh faces below this distance will not be considered by the query
-   :param face: Returns the index of the furthest face
-   :param bary_u: Returns the barycentric u coordinate of the furthest point
-   :param bary_v: Returns the barycentric v coordinate of the furthest point
-
-
-.. function:: mesh_query_point_sign_normal(id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32, epsilon: float32) -> bool
-
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
-
-   Identifies the sign of the distance (inside/outside) using the angle-weighted pseudo normal.
-   This approach to sign determination is robust for well conditioned meshes that are watertight and non-self intersecting.
-   It is also comparatively fast to compute.
-
-   :param id: The mesh identifier
-   :param point: The point in space to query
-   :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                  Note that mesh must be watertight for this to be robust
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
-   :param epsilon: Epsilon treating distance values as equal, when locating the minimum distance vertex/face/edge, as a
-                   fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3
-
-
-.. function:: mesh_query_point_sign_winding_number(id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32, accuracy: float32, threshold: float32) -> bool
-
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given point in space. Returns ``True`` if a point < ``max_dist`` is found.
-
-   Identifies the sign using the winding number of the mesh relative to the query point. This method of sign determination is robust for poorly conditioned meshes
-   and provides a smooth approximation to sign even when the mesh is not watertight. This method is the most robust and accurate of the sign determination meshes
-   but also the most expensive.
-
-   .. note:: The :class:`Mesh` object must be constructed with ``support_winding_number=True`` for this method to return correct results.
-
-   :param id: The mesh identifier
-   :param point: The point in space to query
-   :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                  Note that mesh must be watertight for this to be robust
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
-   :param accuracy: Accuracy for computing the winding number with fast winding number method utilizing second-order dipole approximation, default 2.0
-   :param threshold: The threshold of the winding number to be considered inside, default 0.5
-
-
-.. function:: mesh_query_ray(id: uint64, start: vec3f, dir: vec3f, max_t: float32, t: float32, bary_u: float32, bary_v: float32, sign: float32, normal: vec3f, face: int32) -> bool
-
-   Computes the closest ray hit on the :class:`Mesh` with identifier ``id``, returns ``True`` if a point < ``max_t`` is found.
-
-   :param id: The mesh identifier
-   :param start: The start point of the ray
-   :param dir: The ray direction (should be normalized)
-   :param max_t: The maximum distance along the ray to check for intersections
-   :param t: Returns the distance of the closest hit along the ray
-   :param bary_u: Returns the barycentric u coordinate of the closest hit
-   :param bary_v: Returns the barycentric v coordinate of the closest hit
-   :param sign: Returns a value > 0 if the hit ray hit front of the face, returns < 0 otherwise
-   :param normal: Returns the face normal
-   :param face: Returns the index of the hit face
-
-
-.. function:: mesh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> mesh_query_aabb_t
-
-   Construct an axis-aligned bounding box query against a :class:`Mesh`.
-   This query can be used to iterate over all triangles inside a volume.
-   Returns an object that is used to track state during mesh traversal.
-
-   :param id: The mesh identifier
-   :param lower: The lower bound of the bounding box in mesh space
-   :param upper: The upper bound of the bounding box in mesh space
-
-
-.. function:: mesh_query_aabb_next(query: mesh_query_aabb_t, index: int32) -> bool
-
-   Move to the next triangle overlapping the query bounding box.
-   The index of the current face is stored in ``index``, returns ``False`` if there are no more overlapping triangles.
-
-
-.. function:: mesh_eval_position(id: uint64, face: int32, bary_u: float32, bary_v: float32) -> vec3f
-
-   Evaluates the position on the :class:`Mesh` given a face index and barycentric coordinates.
-
-
-.. function:: mesh_eval_velocity(id: uint64, face: int32, bary_u: float32, bary_v: float32) -> vec3f
-
-   Evaluates the velocity on the :class:`Mesh` given a face index and barycentric coordinates.
-
-
-.. function:: hash_grid_query(id: uint64, point: vec3f, max_dist: float32) -> hash_grid_query_t
-
-   Construct a point query against a :class:`HashGrid`. This query can be used to iterate over all neighboring points within a fixed radius from the query point. Returns an object that is used to track state during neighbor traversal.
-
-
-.. function:: hash_grid_query_next(query: hash_grid_query_t, index: int32) -> bool
-
-   Move to the next point in the hash grid query. The index of the current neighbor is stored in ``index``, returns ``False``
-   if there are no more neighbors.
-
-
-.. function:: hash_grid_point_id(id: uint64, index: int32) -> int
-
-   Return the index of a point in the :class:`HashGrid`. This can be used to reorder threads such that grid
-   traversal occurs in a spatially coherent order.
-
-   Returns -1 if the :class:`HashGrid` has not been reserved.
-
-
-.. function:: intersect_tri_tri(v0: vec3f, v1: vec3f, v2: vec3f, u0: vec3f, u1: vec3f, u2: vec3f) -> int
-
-   Tests for intersection between two triangles (v0, v1, v2) and (u0, u1, u2) using Moller's method. Returns > 0 if triangles intersect.
-
-
-.. function:: mesh_get(id: uint64) -> Mesh
-
-   Retrieves the mesh given its index. [1]_
-
-
-.. function:: mesh_eval_face_normal(id: uint64, face: int32) -> vec3f
-
-   Evaluates the face normal the mesh given a face index.
-
-
-.. function:: mesh_get_point(id: uint64, index: int32) -> vec3f
-
-   Returns the point of the mesh given a index.
-
-
-.. function:: mesh_get_velocity(id: uint64, index: int32) -> vec3f
-
-   Returns the velocity of the mesh given a index.
-
-
-.. function:: mesh_get_index(id: uint64, index: int32) -> int
-
-   Returns the point-index of the mesh given a face-vertex index.
-
-
-.. function:: closest_point_edge_edge(p1: vec3f, q1: vec3f, p2: vec3f, q2: vec3f, epsilon: float32) -> vec3f
-
-   Finds the closest points between two edges. Returns barycentric weights to the points on each edge, as well as the closest distance between the edges.
-
-   :param p1: First point of first edge
-   :param q1: Second point of first edge
-   :param p2: First point of second edge
-   :param q2: Second point of second edge
-   :param epsilon: Zero tolerance for determining if points in an edge are degenerate.
-   :param out: vec3 output containing (s,t,d), where `s` in [0,1] is the barycentric weight for the first edge, `t` is the barycentric weight for the second edge, and `d` is the distance between the two edges at these two closest points.
-
-
-
-
-Volumes
----------------
-.. function:: volume_sample_f(id: uint64, uvw: vec3f, sampling_mode: int32) -> float
-
-   Sample the volume given by ``id`` at the volume local-space point ``uvw``.
-   Interpolation should be :attr:`warp.Volume.CLOSEST` or :attr:`wp.Volume.LINEAR.`
-
-
-.. function:: volume_sample_grad_f(id: uint64, uvw: vec3f, sampling_mode: int32, grad: vec3f) -> float
-
-   Sample the volume and its gradient given by ``id`` at the volume local-space point ``uvw``. 
-   Interpolation should be :attr:`warp.Volume.CLOSEST` or :attr:`wp.Volume.LINEAR.`
-
-
-.. function:: volume_lookup_f(id: uint64, i: int32, j: int32, k: int32) -> float
-
-   Returns the value of voxel with coordinates ``i``, ``j``, ``k``.
-   If the voxel at this index does not exist, this function returns the background value
-
-
-.. function:: volume_store_f(id: uint64, i: int32, j: int32, k: int32, value: float32) -> None
-
-   Store ``value`` at the voxel with coordinates ``i``, ``j``, ``k``.
-
-
-.. function:: volume_sample_v(id: uint64, uvw: vec3f, sampling_mode: int32) -> vec3f
-
-   Sample the vector volume given by ``id`` at the volume local-space point ``uvw``.
-   Interpolation should be :attr:`warp.Volume.CLOSEST` or :attr:`wp.Volume.LINEAR.`
-
-
-.. function:: volume_lookup_v(id: uint64, i: int32, j: int32, k: int32) -> vec3f
-
-   Returns the vector value of voxel with coordinates ``i``, ``j``, ``k``.
-   If the voxel at this index does not exist, this function returns the background value.
-
-
-.. function:: volume_store_v(id: uint64, i: int32, j: int32, k: int32, value: vec3f) -> None
-
-   Store ``value`` at the voxel with coordinates ``i``, ``j``, ``k``.
-
-
-.. function:: volume_sample_i(id: uint64, uvw: vec3f) -> int
-
-   Sample the :class:`int32` volume given by ``id`` at the volume local-space point ``uvw``. 
-
-
-.. function:: volume_lookup_i(id: uint64, i: int32, j: int32, k: int32) -> int
-
-   Returns the :class:`int32` value of voxel with coordinates ``i``, ``j``, ``k``.
-   If the voxel at this index does not exist, this function returns the background value.
-
-
-.. function:: volume_store_i(id: uint64, i: int32, j: int32, k: int32, value: int32) -> None
-
-   Store ``value`` at the voxel with coordinates ``i``, ``j``, ``k``.
-
-
-.. function:: volume_index_to_world(id: uint64, uvw: vec3f) -> vec3f
-
-   Transform a point ``uvw`` defined in volume index space to world space given the volume's intrinsic affine transformation.
-
-
-.. function:: volume_world_to_index(id: uint64, xyz: vec3f) -> vec3f
-
-   Transform a point ``xyz`` defined in volume world space to the volume's index space given the volume's intrinsic affine transformation.
-
-
-.. function:: volume_index_to_world_dir(id: uint64, uvw: vec3f) -> vec3f
-
-   Transform a direction ``uvw`` defined in volume index space to world space given the volume's intrinsic affine transformation.
-
-
-.. function:: volume_world_to_index_dir(id: uint64, xyz: vec3f) -> vec3f
-
-   Transform a direction ``xyz`` defined in volume world space to the volume's index space given the volume's intrinsic affine transformation.
-
-
-
-
-Random
----------------
-.. function:: rand_init(seed: int32) -> uint32
-
-   Initialize a new random number generator given a user-defined seed. Returns a 32-bit integer representing the RNG state.
-
-
-.. function:: rand_init(seed: int32, offset: int32) -> uint32
-   :noindex:
-   :nocontentsentry:
-
-   Initialize a new random number generator given a user-defined seed and an offset.
-   This alternative constructor can be useful in parallel programs, where a kernel as a whole should share a seed,
-   but each thread should generate uncorrelated values. In this case usage should be ``r = rand_init(seed, tid)``
-
-
-.. function:: randi(state: uint32) -> int
-
-   Return a random integer in the range [0, 2^32).
-
-
-.. function:: randi(state: uint32, min: int32, max: int32) -> int
-   :noindex:
-   :nocontentsentry:
-
-   Return a random integer between [min, max).
-
-
-.. function:: randf(state: uint32) -> float
-
-   Return a random float between [0.0, 1.0).
-
-
-.. function:: randf(state: uint32, min: float32, max: float32) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Return a random float between [min, max).
-
-
-.. function:: randn(state: uint32) -> float
-
-   Sample a normal distribution.
-
-
-.. function:: sample_cdf(state: uint32, cdf: Array[float32]) -> int
-
-   Inverse-transform sample a cumulative distribution function.
-
-
-.. function:: sample_triangle(state: uint32) -> vec2f
-
-   Uniformly sample a triangle. Returns sample barycentric coordinates.
-
-
-.. function:: sample_unit_ring(state: uint32) -> vec2f
-
-   Uniformly sample a ring in the xy plane.
-
-
-.. function:: sample_unit_disk(state: uint32) -> vec2f
-
-   Uniformly sample a disk in the xy plane.
-
-
-.. function:: sample_unit_sphere_surface(state: uint32) -> vec3f
-
-   Uniformly sample a unit sphere surface.
-
-
-.. function:: sample_unit_sphere(state: uint32) -> vec3f
-
-   Uniformly sample a unit sphere.
-
-
-.. function:: sample_unit_hemisphere_surface(state: uint32) -> vec3f
-
-   Uniformly sample a unit hemisphere surface.
-
-
-.. function:: sample_unit_hemisphere(state: uint32) -> vec3f
-
-   Uniformly sample a unit hemisphere.
-
-
-.. function:: sample_unit_square(state: uint32) -> vec2f
-
-   Uniformly sample a unit square.
-
-
-.. function:: sample_unit_cube(state: uint32) -> vec3f
-
-   Uniformly sample a unit cube.
-
-
-.. function:: poisson(state: uint32, lam: float32) -> uint32
-
-   Generate a random sample from a Poisson distribution.
-
-   :param state: RNG state
-   :param lam: The expected value of the distribution
-
-
-.. function:: noise(state: uint32, x: float32) -> float
-
-   Non-periodic Perlin-style noise in 1D.
-
-
-.. function:: noise(state: uint32, xy: vec2f) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Non-periodic Perlin-style noise in 2D.
-
-
-.. function:: noise(state: uint32, xyz: vec3f) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Non-periodic Perlin-style noise in 3D.
-
-
-.. function:: noise(state: uint32, xyzt: vec4f) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Non-periodic Perlin-style noise in 4D.
-
-
-.. function:: pnoise(state: uint32, x: float32, px: int32) -> float
-
-   Periodic Perlin-style noise in 1D.
-
-
-.. function:: pnoise(state: uint32, xy: vec2f, px: int32, py: int32) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Periodic Perlin-style noise in 2D.
-
-
-.. function:: pnoise(state: uint32, xyz: vec3f, px: int32, py: int32, pz: int32) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Periodic Perlin-style noise in 3D.
-
-
-.. function:: pnoise(state: uint32, xyzt: vec4f, px: int32, py: int32, pz: int32, pt: int32) -> float
-   :noindex:
-   :nocontentsentry:
-
-   Periodic Perlin-style noise in 4D.
-
-
-.. function:: curlnoise(state: uint32, xy: vec2f, octaves: uint32, lacunarity: float32, gain: float32) -> vec2f
-
-   Divergence-free vector field based on the gradient of a Perlin noise function. [1]_
-
-
-.. function:: curlnoise(state: uint32, xyz: vec3f, octaves: uint32, lacunarity: float32, gain: float32) -> vec3f
-   :noindex:
-   :nocontentsentry:
-
-   Divergence-free vector field based on the curl of three Perlin noise functions. [1]_
-
-
-.. function:: curlnoise(state: uint32, xyzt: vec4f, octaves: uint32, lacunarity: float32, gain: float32) -> vec3f
-   :noindex:
-   :nocontentsentry:
-
-   Divergence-free vector field based on the curl of three Perlin noise functions. [1]_
-
-
-
-
-Operators
----------------
-.. function:: add(x: Scalar, y: Scalar) -> Scalar
-
-
-.. function:: add(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: add(x: Quaternion[Scalar], y: Quaternion[Scalar]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: add(x: Matrix[Any,Any,Scalar], y: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: add(x: Transformation[Scalar], y: Transformation[Scalar]) -> Transformation[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: sub(x: Scalar, y: Scalar) -> Scalar
-
-
-.. function:: sub(x: Vector[Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: sub(x: Matrix[Any,Any,Scalar], y: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: sub(x: Quaternion[Scalar], y: Quaternion[Scalar]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: sub(x: Transformation[Scalar], y: Transformation[Scalar]) -> Transformation[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: bit_and(x: Int, y: Int) -> Int
-
-
-.. function:: bit_or(x: Int, y: Int) -> Int
-
-
-.. function:: bit_xor(x: Int, y: Int) -> Int
-
-
-.. function:: lshift(x: Int, y: Int) -> Int
-
-
-.. function:: rshift(x: Int, y: Int) -> Int
-
-
-.. function:: invert(x: Int) -> Int
-
-
-.. function:: mul(x: Scalar, y: Scalar) -> Scalar
-
-
-.. function:: mul(x: Vector[Any,Scalar], y: Scalar) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Scalar, y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Quaternion[Scalar], y: Scalar) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Scalar, y: Quaternion[Scalar]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Quaternion[Scalar], y: Quaternion[Scalar]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Scalar, y: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Matrix[Any,Any,Scalar], y: Scalar) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Matrix[Any,Any,Scalar], y: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Matrix[Any,Any,Scalar], y: Matrix[Any,Any,Scalar])
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Transformation[Scalar], y: Transformation[Scalar]) -> Transformation[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Scalar, y: Transformation[Scalar]) -> Transformation[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mul(x: Transformation[Scalar], y: Scalar) -> Transformation[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: mod(x: Scalar, y: Scalar) -> Scalar
-
-
-.. function:: div(x: Scalar, y: Scalar) -> Scalar
-
-
-.. function:: div(x: Vector[Any,Scalar], y: Scalar) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: div(x: Matrix[Any,Any,Scalar], y: Scalar) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: div(x: Quaternion[Scalar], y: Scalar) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: floordiv(x: Scalar, y: Scalar) -> Scalar
-
-
-.. function:: pos(x: Scalar) -> Scalar
-
-
-.. function:: pos(x: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: pos(x: Quaternion[Scalar]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: pos(x: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: neg(x: Scalar) -> Scalar
-
-
-.. function:: neg(x: Vector[Any,Scalar]) -> Vector[Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: neg(x: Quaternion[Scalar]) -> Quaternion[Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: neg(x: Matrix[Any,Any,Scalar]) -> Matrix[Any,Any,Scalar]
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: bool) -> bool
-
-
-.. function:: unot(b: int8) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: uint8) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: int16) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: uint16) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: int32) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: uint32) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: int64) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(b: uint64) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. function:: unot(a: Array[Any]) -> bool
-   :noindex:
-   :nocontentsentry:
-
-
-.. rubric:: Footnotes
-.. [1] Note: function gradients not implemented for backpropagation.
diff --git a/docs/_build/html/_sources/modules/interoperability.rst.txt b/docs/_build/html/_sources/modules/interoperability.rst.txt
deleted file mode 100644
index 5e4472ccb..000000000
--- a/docs/_build/html/_sources/modules/interoperability.rst.txt
+++ /dev/null
@@ -1,146 +0,0 @@
-Interoperability
-================
-
-Warp can interop with other Python-based frameworks such as NumPy through standard interface protocols.
-
-NumPy
------
-
-Warp arrays may be converted to a NumPy array through the ``warp.array.numpy()`` method. When the Warp array lives on
-the ``cpu`` device this will return a zero-copy view onto the underlying Warp allocation. If the array lives on a
-``cuda`` device then it will first be copied back to a temporary buffer and copied to NumPy.
-
-Warp CPU arrays also implement  the ``__array_interface__`` protocol and so can be used to construct NumPy arrays
-directly::
-
-    w = wp.array([1.0, 2.0, 3.0], dtype=float, device="cpu")
-    a = np.array(w)
-    print(a)   
-    > [1. 2. 3.]
-
-.. _pytorch-interop:
-
-PyTorch
--------
-
-Warp provides helper functions to convert arrays to/from PyTorch. Please see the ``warp.torch`` module for more details. Example usage is shown below::
-
-    import warp.torch
-
-    w = wp.array([1.0, 2.0, 3.0], dtype=float, device="cpu")
-
-    # convert to Torch tensor
-    t = warp.to_torch(w)
-
-    # convert from Torch tensor
-    w = warp.from_torch(t)
-
-These helper functions allow the conversion of Warp arrays to/from PyTorch tensors without copying the underlying data.
-At the same time, if available, gradient arrays and tensors are converted to/from PyTorch autograd tensors, allowing the use of Warp arrays
-in PyTorch autograd computations.
-
-Example: Optimization using ``warp.from_torch()``
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-An example usage of minimizing a loss function over an array of 2D points written in Warp via PyTorch's Adam optimizer using ``warp.from_torch`` is as follows::
-
-    import warp as wp
-    import torch
-
-    wp.init()
-
-    @wp.kernel()
-    def loss(xs: wp.array(dtype=float, ndim=2), l: wp.array(dtype=float)):
-        tid = wp.tid()
-        wp.atomic_add(l, 0, xs[tid, 0] ** 2.0 + xs[tid, 1] ** 2.0)
-
-    # indicate requires_grad so that Warp can accumulate gradients in the grad buffers
-    xs = torch.randn(100, 2, requires_grad=True)
-    l = torch.zeros(1, requires_grad=True)
-    opt = torch.optim.Adam([xs], lr=0.1)
-
-    wp_xs = wp.from_torch(xs)
-    wp_l = wp.from_torch(l)
-
-    tape = wp.Tape()
-    with tape:
-        # record the loss function kernel launch on the tape
-        wp.launch(loss, dim=len(xs), inputs=[wp_xs], outputs=[wp_l], device=wp_xs.device)
-
-    for i in range(500):
-        tape.zero()
-        tape.backward(loss=wp_l)  # compute gradients
-        # now xs.grad will be populated with the gradients computed by Warp
-        opt.step()  # update xs (and thereby wp_xs)
-
-        # these lines are only needed for evaluating the loss
-        # (the optimization just needs the gradient, not the loss value)
-        wp_l.zero_()
-        wp.launch(loss, dim=len(xs), inputs=[wp_xs], outputs=[wp_l], device=wp_xs.device)
-        print(f"{i}\tloss: {l.item()}")
-
-Example: Optimization using ``warp.to_torch``
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-Less code is needed when we declare the optimization variables directly in Warp and use ``warp.to_torch`` to convert them to PyTorch tensors.
-Here, we revisit the same example from above where now only a single conversion to a torch tensor is needed to supply Adam with the optimization variables::
-
-    import warp as wp
-    import numpy as np
-    import torch
-
-    wp.init()
-
-    @wp.kernel()
-    def loss(xs: wp.array(dtype=float, ndim=2), l: wp.array(dtype=float)):
-        tid = wp.tid()
-        wp.atomic_add(l, 0, xs[tid, 0] ** 2.0 + xs[tid, 1] ** 2.0)
-
-    # initialize the optimization variables in Warp
-    xs = wp.array(np.random.randn(100, 2), dtype=wp.float32, requires_grad=True)
-    l = wp.zeros(1, dtype=wp.float32, requires_grad=True)
-    # just a single wp.to_torch call is needed, Adam optimizes using the Warp array gradients
-    opt = torch.optim.Adam([wp.to_torch(xs)], lr=0.1)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(loss, dim=len(xs), inputs=[xs], outputs=[l], device=xs.device)
-
-    for i in range(500):
-        tape.zero()
-        tape.backward(loss=l)
-        opt.step()
-        
-        l.zero_()
-        wp.launch(loss, dim=len(xs), inputs=[xs], outputs=[l], device=xs.device)
-        print(f"{i}\tloss: {l.numpy()[0]}")
-
-.. automodule:: warp.torch
-    :members:
-    :undoc-members:
-
-CuPy/Numba
-----------
-
-Warp GPU arrays support the ``__cuda_array_interface__`` protocol for sharing data with other Python GPU frameworks.
-Currently this is one-directional, so that Warp arrays can be used as input to any framework that also supports the
-``__cuda_array_interface__`` protocol, but not the other way around.
-
-.. _jax-interop:
-
-JAX
----
-
-Interoperability with JAX arrays is supported through the following methods.
-Internally these use the DLPack protocol to exchange data in a zero-copy way with JAX.
-
-.. automodule:: warp.jax
-    :members:
-    :undoc-members:
-
-DLPack
-------
-
-.. automodule:: warp.dlpack
-    :members:
-    :undoc-members:
diff --git a/docs/_build/html/_sources/modules/profiling.rst.txt b/docs/_build/html/_sources/modules/profiling.rst.txt
deleted file mode 100644
index 315f72b23..000000000
--- a/docs/_build/html/_sources/modules/profiling.rst.txt
+++ /dev/null
@@ -1,44 +0,0 @@
-Profiling
-=========
-
-``wp.ScopedTimer`` objects can be used to gain some basic insight into the performance of Warp applications::
-
-   with wp.ScopedTimer("grid build"):
-      self.grid.build(self.x, self.point_radius)
-
-This results in a printout at runtime to the standard output stream like::
-
-   grid build took 0.06 ms
-
-The ``wp.ScopedTimer`` object does not synchronize (e.g. by calling ``wp.synchronize()``)
-upon exiting the ``with`` statement, so this can lead to misleading numbers if the body
-of the ``with`` statement launches device kernels.
-
-When a ``wp.ScopedTimer`` object is passed ``use_nvtx=True`` as an argument, the timing functionality is replaced by calls
-to ``nvtx.start_range()`` and ``nvtx.end_range()``::
-
-   with wp.ScopedTimer("grid build", use_nvtx=True, color="cyan"):
-      self.grid.build(self.x, self.point_radius)
-
-These range annotations can then be collected by a tool like `NVIDIA Nsight Systems <https://developer.nvidia.com/nsight-systems>`_
-for visualization on a timeline, e.g.::
-
-   nsys profile python warp_application.py
-
-This code snippet also demonstrates the use of the ``color`` argument to specify a color
-for the range, which may be a number representing the ARGB value or a recognized string
-(refer to `colors.py <https://github.com/NVIDIA/NVTX/blob/release-v3/python/nvtx/colors.py>`_ for
-additional color examples).
-The `nvtx module <https://github.com/NVIDIA/NVTX>`_ must be
-installed in the Python environment for this capability to work.
-An equivalent way to create an NVTX range without using ``wp.ScopedTimer`` is::
-
-   import nvtx
-
-   with nvtx.annotate("grid build", color="cyan"):
-      self.grid.build(self.x, self.point_radius)
-
-This form may be more convenient if the user does not need to frequently switch
-between timer and NVTX capabilities of ``wp.ScopedTimer``. 
-
-.. autoclass:: warp.ScopedTimer
diff --git a/docs/_build/html/_sources/modules/runtime.rst.txt b/docs/_build/html/_sources/modules/runtime.rst.txt
deleted file mode 100644
index 766adf1c3..000000000
--- a/docs/_build/html/_sources/modules/runtime.rst.txt
+++ /dev/null
@@ -1,1491 +0,0 @@
-Runtime Reference
-=================
-
-.. currentmodule:: warp
-
-This section describes the Warp Python runtime API, how to manage memory, launch kernels, and high-level functionality
-for dealing with objects such as meshes and volumes. The APIs described in this section are intended to be used at
-the *Python Scope* and run inside the CPython interpreter. For a comprehensive list of functions available at
-the *Kernel Scope*, please see the :doc:`functions` section.
-
-Kernels
--------
-
-Kernels are launched with the :func:`wp.launch() <launch>` function on a specific device (CPU/GPU)::
-
-    wp.launch(simple_kernel, dim=1024, inputs=[a, b, c], device="cuda")
-
-Kernels may be launched with multi-dimensional grid bounds. In this case threads are not assigned a single index,
-but a coordinate in an n-dimensional grid, e.g.::
-
-    wp.launch(complex_kernel, dim=(128, 128, 3), ...)
-
-Launches a 3D grid of threads with dimension 128 x 128 x 3. To retrieve the 3D index for each thread use the following syntax::
-
-    i,j,k = wp.tid()
-
-.. note::
-    Currently kernels launched on CPU devices will be executed in serial.
-    Kernels launched on CUDA devices will be launched in parallel with a fixed block-size.
-
-.. note::
-    Note that all the kernel inputs must live on the target device, or a runtime exception will be raised.
-
-.. autofunction:: launch
-
-Large Kernel Launches
-#####################
-
-A limitation of Warp is that each dimension of the grid used to launch a kernel must be representable as a 32-bit
-signed integer. Therefore, no single dimension of a grid should exceed :math:`2^{31}-1`.
-
-Warp also currently uses a fixed block size of 256 (CUDA) threads per block.
-By default, Warp will try to process one element from the Warp grid in one CUDA thread.
-This is not always possible for kernels launched with multi-dimensional grid bounds, as there are
-`hardware limitations <https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#features-and-technical-specifications-technical-specifications-per-compute-capability>`_
-on CUDA block dimensions.
-
-Warp will automatically resort to using
-`grid-stride loops <https://developer.nvidia.com/blog/cuda-pro-tip-write-flexible-kernels-grid-stride-loops/>`_ when
-it is not possible for a CUDA thread to process only one element from the Warp grid
-When this happens, some CUDA threads may process more than one element from the Warp grid.
-Users can also set the ``max_blocks`` parameter to fine-tune the grid-striding behavior of kernels, even for kernels that are otherwise
-able to process one Warp-grid element per CUDA thread. 
-
-Runtime Kernel Specialization
-#############################
-
-It is often desirable to specialize kernels for different types, constants, or functions.
-We can achieve this through the use of runtime kernel specialization using Python closures.
-
-For example, we might require a variety of kernels that execute particular functions for each item in an array.
-We might also want this function call to be valid for a variety of data types. Making use of closure and generics, we can generate
-these kernels using a single kernel definition: ::
-
-    def make_kernel(func, dtype):
-        def closure_kernel_fn(data: wp.array(dtype=dtype), out: wp.array(dtype=dtype)):
-            tid = wp.tid()
-            out[tid] = func(data[tid])
-
-        return wp.Kernel(closure_kernel_fn)
-
-In practice, we might use our kernel generator, ``make_kernel()`` as follows: ::
-
-    @wp.func
-    def sqr(x: Any) -> Any:
-        return x * x
-
-    @wp.func
-    def cube(x: Any) -> Any:
-        return sqr(x) * x
-
-    sqr_float = make_kernel(sqr, wp.float32)
-    cube_double = make_kernel(cube, wp.float64)
-
-    arr = [1.0, 2.0, 3.0]
-    N = len(arr)
-
-    data_float = wp.array(arr, dtype=wp.float32, device=device)
-    data_double = wp.array(arr, dtype=wp.float64, device=device)
-
-    out_float = wp.zeros(N, dtype=wp.float32, device=device)
-    out_double = wp.zeros(N, dtype=wp.float64, device=device)
-
-    wp.launch(sqr_float, dim=N, inputs=[data_float], outputs=[out_float], device=device)
-    wp.launch(cube_double, dim=N, inputs=[data_double], outputs=[out_double], device=device)
-
-We can specialize kernel definitions over warp constants similarly. The following generates kernels that add a specified constant
-to a generic-typed array value: ::
-
-    def make_add_kernel(key, constant):
-        def closure_kernel_fn(data: wp.array(dtype=Any), out: wp.array(dtype=Any)):
-            tid = wp.tid()
-            out[tid] = data[tid] + constant
-
-        return wp.Kernel(closure_kernel_fn, key=key)
-
-    add_ones_int = make_add_kernel("add_one", wp.constant(1))
-    add_ones_vec3 = make_add_kernel("add_ones_vec3", wp.constant(wp.vec3(1.0, 1.0, 1.0)))
-
-    a = wp.zeros(2, dtype=int)
-    b = wp.zeros(2, dtype=wp.vec3)
-
-    a_out = wp.zeros_like(a)
-    b_out = wp.zeros_like(b)
-
-    wp.launch(add_ones_int, dim=a.size, inputs=[a], outputs=[a_out], device=device)
-    wp.launch(add_ones_vec3, dim=b.size, inputs=[b], outputs=[b_out], device=device)
-
-
-Arrays
-------
-
-Arrays are the fundamental memory abstraction in Warp; they are created through the following global constructors: ::
-
-    wp.empty(shape=1024, dtype=wp.vec3, device="cpu")
-    wp.zeros(shape=1024, dtype=float, device="cuda")
-    wp.full(shape=1024, value=10, dtype=int, device="cuda")
-
-
-Arrays can also be constructed directly from ``numpy`` ndarrays as follows: ::
-
-    r = np.random.rand(1024)
-
-    # copy to Warp owned array
-    a = wp.array(r, dtype=float, device="cpu")
-
-    # return a Warp array wrapper around the NumPy data (zero-copy)
-    a = wp.array(r, dtype=float, copy=False, device="cpu")
-
-    # return a Warp copy of the array data on the GPU
-    a = wp.array(r, dtype=float, device="cuda")
-
-Note that for multi-dimensional data the ``dtype`` parameter must be specified explicitly, e.g.: ::
-
-    r = np.random.rand((1024, 3))
-
-    # initialize as an array of vec3 objects
-    a = wp.array(r, dtype=wp.vec3, device="cuda")
-
-If the shapes are incompatible, an error will be raised.
-
-
-Arrays can be moved between devices using the ``array.to()`` method: ::
-
-    host_array = wp.array(a, dtype=float, device="cpu")
-
-    # allocate and copy to GPU
-    device_array = host_array.to("cuda")
-
-Additionally, arrays can be copied directly between memory spaces: ::
-
-    src_array = wp.array(a, dtype=float, device="cpu")
-    dest_array = wp.empty_like(host_array)
-
-    # copy from source CPU buffer to GPU
-    wp.copy(dest_array, src_array)
-
-.. autoclass:: array
-    :members:
-    :undoc-members:
-
-Multi-dimensional Arrays
-########################
-
-Multi-dimensional arrays can be constructed by passing a tuple of sizes for each dimension, e.g.: the following constructs a 2d array of size 1024x16::
-
-    wp.zeros(shape=(1024, 16), dtype=float, device="cuda")
-
-When passing multi-dimensional arrays to kernels users must specify the expected array dimension inside the kernel signature,
-e.g. to pass a 2d array to a kernel the number of dims is specified using the ``ndim=2`` parameter::
-
-    @wp.kernel
-    def test(input: wp.array(dtype=float, ndim=2)):
-
-Type-hint helpers are provided for common array sizes, e.g.: ``array2d()``, ``array3d()``, which are equivalent to calling ``array(..., ndim=2)```, etc. To index a multi-dimensional array use a the following kernel syntax::
-
-    # returns a float from the 2d array
-    value = input[i,j]
-
-To create an array slice use the following syntax, where the number of indices is less than the array dimensions::
-
-    # returns an 1d array slice representing a row of the 2d array
-    row = input[i]
-
-Slice operators can be concatenated, e.g.: ``s = array[i][j][k]``. Slices can be passed to ``wp.func`` user functions provided
-the function also declares the expected array dimension. Currently only single-index slicing is supported.
-
-.. note::
-    Currently Warp limits arrays to 4 dimensions maximum. This is in addition to the contained datatype, which may be 1-2 dimensional for vector and matrix types such as ``vec3``, and ``mat33``.
-
-
-The following construction methods are provided for allocating zero-initialized and empty (non-initialized) arrays:
-
-.. autofunction:: zeros
-.. autofunction:: zeros_like
-.. autofunction:: full
-.. autofunction:: full_like
-.. autofunction:: empty
-.. autofunction:: empty_like
-.. autofunction:: copy
-.. autofunction:: clone
-
-Matrix Multiplication
-#####################
-
-Warp 2D array multiplication is built on NVIDIA's CUTLASS library, which enables fast matrix multiplication of large arrays on the GPU.
-
-If no GPU is detected, matrix multiplication falls back to Numpy's implementation on the CPU.
-
-Matrix multiplication is fully differentiable, and can be recorded on the tape like so::
-
-    tape = wp.Tape()
-    with tape:
-        wp.matmul(A, B, C, D, device=device)
-        wp.launch(loss_kernel, dim=(m, n), inputs=[D, loss], device=device)
-
-    tape.backward(loss=loss)
-    A_grad = A.grad.numpy()
-
-Using the ``@`` operator (``D = A @ B``) will default to the same CUTLASS algorithm used in ``wp.matmul``.
-
-.. autofunction:: matmul
-
-Data Types
-----------
-
-Scalar Types
-############
-
-The following scalar storage types are supported for array structures:
-
-+---------+------------------------+
-| bool    | boolean                |
-+---------+------------------------+
-| int8    | signed byte            |
-+---------+------------------------+
-| uint8   | unsigned byte          |
-+---------+------------------------+
-| int16   | signed short           |
-+---------+------------------------+
-| uint16  | unsigned short         |
-+---------+------------------------+
-| int32   | signed integer         |
-+---------+------------------------+
-| uint32  | unsigned integer       |
-+---------+------------------------+
-| int64   | signed long integer    |
-+---------+------------------------+
-| uint64  | unsigned long integer  |
-+---------+------------------------+
-| float32 | single precision float |
-+---------+------------------------+
-| float64 | double precision float |
-+---------+------------------------+
-
-Warp supports ``float`` and ``int`` as aliases for ``wp.float32`` and ``wp.int32`` respectively.
-
-.. _vec:
-
-Vectors
-#######
-
-Warp provides built-in math and geometry types for common simulation and graphics problems.
-A full reference for operators and functions for these types is available in the :doc:`/modules/functions`.
-
-Warp supports vectors of numbers with an arbitrary length/numeric type. The built in concrete types are as follows:
-
-+-----------------------+------------------------------------------------+
-| vec2 vec3 vec4        | 2d, 3d, 4d vector of default precision floats  |
-+-----------------------+------------------------------------------------+
-| vec2f vec3f vec4f     | 2d, 3d, 4d vector of single precision floats   |
-+-----------------------+------------------------------------------------+
-| vec2d vec3d vec4d     | 2d, 3d, 4d vector of double precision floats   |
-+-----------------------+------------------------------------------------+
-| vec2h vec3h vec4h     | 2d, 3d, 4d vector of half precision floats     |
-+-----------------------+------------------------------------------------+
-| vec2ub vec3ub vec4ub  | 2d, 3d, 4d vector of half precision floats     |
-+-----------------------+------------------------------------------------+
-| spatial_vector        | 6d vector of single precision floats           |
-+-----------------------+------------------------------------------------+
-| spatial_vectorf       | 6d vector of single precision floats           |
-+-----------------------+------------------------------------------------+
-| spatial_vectord       | 6d vector of double precision floats           |
-+-----------------------+------------------------------------------------+
-| spatial_vectorh       | 6d vector of half precision floats             |
-+-----------------------+------------------------------------------------+
-
-Vectors support most standard linear algebra operations, e.g.: ::
-
-    @wp.kernel
-    def compute( ... ):
-
-        # basis vectors
-        a = wp.vec3(1.0, 0.0, 0.0)
-        b = wp.vec3(0.0, 1.0, 0.0)
-
-        # take the cross product
-        c = wp.cross(a, b)
-
-        # compute
-        r = wp.dot(c, c)
-
-        ...
-
-
-It's possible to declare additional vector types with different lengths and data types. This is done in outside of kernels in *Python scope* using ``warp.types.vector()``, for example: ::
-
-    # declare a new vector type for holding 5 double precision floats:
-    vec5d = wp.types.vector(length=5, dtype=wp.float64)
-
-Once declared, the new type can be used when allocating arrays or inside kernels: ::
-
-    # create an array of vec5d
-    arr = wp.zeros(10, dtype=vec5d)
-
-    # use inside a kernel
-    @wp.kernel
-    def compute( ... ):
-
-        # zero initialize a custom named vector type
-        v = vec5d()
-        ...
-
-        # component-wise initialize a named vector type
-        v = vec5d(wp.float64(1.0),
-                  wp.float64(2.0),
-                  wp.float64(3.0),
-                  wp.float64(4.0),
-                  wp.float64(5.0))
-      ...
-
-In addition, it's possible to directly create *anonymously* typed instances of these vectors without declaring their type in advance. In this case the type will be inferred by the constructor arguments. For example: ::
-
-    @wp.kernel
-    def compute( ... ):
-
-        # zero initialize vector of 5 doubles:
-        v = wp.vector(dtype=wp.float64, length=5)
-
-        # scalar initialize a vector of 5 doubles to the same value:
-        v = wp.vector(wp.float64(1.0), length=5)
-
-        # component-wise initialize a vector of 5 doubles
-        v = wp.vector(wp.float64(1.0),
-                      wp.float64(2.0),
-                      wp.float64(3.0),
-                      wp.float64(4.0),
-                      wp.float64(5.0))
-
-
-These can be used with all the standard vector arithmetic operators, e.g.: ``+``, ``-``, scalar multiplication, and can also be transformed using matrices with compatible dimensions, potentially returning vectors with a different length.
-
-.. _mat:
-
-Matrices
-########
-
-Matrices with arbitrary shapes/numeric types are also supported. The built in concrete matrix types are as follows:
-
-+--------------------------+-----------------------------------------------------+
-| mat22 mat33 mat44        | 2,3 and 4d square matrix of default precision       |
-+--------------------------+-----------------------------------------------------+
-| mat22f mat33f mat44f     | 2,3 and 4d square matrix of single precision floats |
-+--------------------------+-----------------------------------------------------+
-| mat22d mat33d mat44d     | 2,3 and 4d square matrix of double precision floats |
-+--------------------------+-----------------------------------------------------+
-| mat22h mat33h mat44h     | 2,3 and 4d square matrix of half precision floats   |
-+--------------------------+-----------------------------------------------------+
-| spatial_matrix           | 6x6 matrix of single precision floats               |
-+--------------------------+-----------------------------------------------------+
-| spatial_matrixf          | 6x6 matrix of single precision floats               |
-+--------------------------+-----------------------------------------------------+
-| spatial_matrixd          | 6x6 matrix of double precision floats               |
-+--------------------------+-----------------------------------------------------+
-| spatial_matrixh          | 6x6 matrix of half precision floats                 |
-+--------------------------+-----------------------------------------------------+
-
-Matrices are stored in row-major format and support most standard linear algebra operations: ::
-
-    @wp.kernel
-    def compute( ... ):
-
-        # initialize matrix
-        m = wp.mat22(1.0, 2.0,
-                     3.0, 4.0)
-
-        # compute inverse
-        minv = wp.inverse(m)
-
-        # transform vector
-        v = minv * wp.vec2(0.5, 0.3)
-
-        ...
-
-
-In a similar manner to vectors, it's possible to declare new matrix types with arbitrary shapes and data types using ``wp.types.matrix()``, for example: ::
-
-    # declare a new 3x2 half precision float matrix type:
-    mat32h = wp.types.matrix(shape=(3,2), dtype=wp.float64)
-
-    # create an array of this type
-    a = wp.zeros(10, dtype=mat32h)
-
-These can be used inside a kernel: ::
-
-    @wp.kernel
-    def compute( ... ):
-        ...
-
-        # initialize a mat32h matrix
-        m = mat32h(wp.float16(1.0), wp.float16(2.0),
-                   wp.float16(3.0), wp.float16(4.0),
-                   wp.float16(5.0), wp.float16(6.0))
-
-        # declare a 2 component half precision vector
-        v2 = wp.vec2h(wp.float16(1.0), wp.float16(1.0))
-
-        # multiply by the matrix, returning a 3 component vector:
-        v3 = m * v2
-        ...
-
-It's also possible to directly create anonymously typed instances inside kernels where the type is inferred from constructor arguments as follows: ::
-
-    @wp.kernel
-    def compute( ... ):
-        ...
-
-        # create a 3x2 half precision matrix from components (row major ordering):
-        m = wp.matrix(
-            wp.float16(1.0), wp.float16(2.0),
-            wp.float16(1.0), wp.float16(2.0),
-            wp.float16(1.0), wp.float16(2.0),
-            shape=(3,2))
-
-        # zero initialize a 3x2 half precision matrix:
-        m = wp.matrix(wp.float16(0.0),shape=(3,2))
-
-        # create a 5x5 double precision identity matrix:
-        m = wp.identity(n=5, dtype=wp.float64)
-
-As with vectors, you can do standard matrix arithmetic with these variables, along with multiplying matrices with compatible shapes and potentially returning a matrix with a new shape.
-
-.. _quat:
-
-Quaternions
-###########
-
-Warp supports quaternions with the layout ``i, j, k, w`` where ``w`` is the real part. Here are the built in concrete quaternion types:
-
-+-----------------+-----------------------------------------------+
-| quat            | Default precision floating point quaternion   |
-+-----------------+-----------------------------------------------+
-| quatf           | Single precision floating point quaternion    |
-+-----------------+-----------------------------------------------+
-| quatd           | Double precision floating point quaternion    |
-+-----------------+-----------------------------------------------+
-| quath           | Half precision floating point quaternion      |
-+-----------------+-----------------------------------------------+
-
-Quaternions can be used to transform vectors as follows: ::
-
-    @wp.kernel
-    def compute( ... ):
-        ...
-
-        # construct a 30 degree rotation around the x-axis
-        q = wp.quat_from_axis_angle(wp.vec3(1.0, 0.0, 0.0), wp.degrees(30.0))
-
-        # rotate an axis by this quaternion
-        v = wp.quat_rotate(q, wp.vec3(0.0, 1.0, 0.0))
-
-
-As with vectors and matrices, you can declare quaternion types with an arbitrary numeric type like so: ::
-
-    quatd = wp.types.quaternion(dtype=wp.float64)
-
-You can also create identity quaternion and anonymously typed instances inside a kernel like so: ::
-
-    @wp.kernel
-    def compute( ... ):
-        ...
-
-        # create a double precision identity quaternion:
-        qd = wp.quat_identity(dtype=wp.float64)
-
-        # precision defaults to wp.float32 so this creates a single precision identity quaternion:
-        qf = wp.quat_identity()
-
-        # create a half precision quaternion from components, or a vector/scalar:
-        qh = wp.quaternion(wp.float16(0.0),
-                           wp.float16(0.0),
-                           wp.float16(0.0),
-                           wp.float16(1.0))
-
-
-        qh = wp.quaternion(
-            wp.vector(wp.float16(0.0),wp.float16(0.0),wp.float16(0.0)),
-            wp.float16(1.0))
-
-.. _transform:
-
-Transforms
-##########
-
-Transforms are 7d vectors of floats representing a spatial rigid body transformation in format (p, q) where p is a 3d vector, and q is a quaternion.
-
-+-----------------+--------------------------------------------+
-| transform       | Default precision floating point transform |
-+-----------------+--------------------------------------------+
-| transformf      | Single precision floating point transform  |
-+-----------------+--------------------------------------------+
-| transformd      | Double precision floating point transform  |
-+-----------------+--------------------------------------------+
-| transformh      | Half precision floating point transform    |
-+-----------------+--------------------------------------------+
-
-Transforms can be constructed inside kernels from translation and rotation parts: ::
-
-    @wp.kernel
-    def compute( ... ):
-        ...
-
-        # create a transform from a vector/quaternion:
-        t = wp.transform(
-                wp.vec3(1.0, 2.0, 3.0),
-                wp.quat_from_axis_angle(0.0, 1.0, 0.0, wp.degrees(30.0)))
-
-        # transform a point
-        p = wp.transform_point(t, wp.vec3(10.0, 0.5, 1.0))
-
-        # transform a vector (ignore translation)
-        p = wp.transform_vector(t, wp.vec3(10.0, 0.5, 1.0))
-
-
-As with vectors and matrices, you can declare transform types with an arbitrary numeric type using ``wp.types.transformation()``, for example: ::
-
-    transformd = wp.types.transformation(dtype=wp.float64)
-
-You can also create identity transforms and anonymously typed instances inside a kernel like so: ::
-
-    @wp.kernel
-    def compute( ... ):
-
-        # create double precision identity transform:
-        qd = wp.transform_identity(dtype=wp.float64)
-
-Structs
-#######
-
-Users can define custom structure types using the ``@wp.struct`` decorator as follows::
-
-    @wp.struct
-    class MyStruct:
-
-        param1: int
-        param2: float
-        param3: wp.array(dtype=wp.vec3)
-
-Struct attributes must be annotated with their respective type. They can be constructed in Python scope and then passed to kernels as arguments::
-
-    @wp.kernel
-    def compute(args: MyStruct):
-
-        tid = wp.tid()
-
-        print(args.param1)
-        print(args.param2)
-        print(args.param3[tid])
-
-    # construct an instance of the struct in Python
-    s = MyStruct()
-    s.param1 = 10
-    s.param2 = 2.5
-    s.param3 = wp.zeros(shape=10, dtype=wp.vec3)
-
-    # pass to our compute kernel
-    wp.launch(compute, dim=10, inputs=[s])
-
-An array of structs can be zero-initialized as follows::
-
-    a = wp.zeros(shape=10, dtype=MyStruct)
-
-An array of structs can also be initialized from a list of struct objects::
-
-    a = wp.array([MyStruct(), MyStruct(), MyStruct()], dtype=MyStruct)
-
-Example: Using a custom struct in gradient computation
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-.. code:: python
-
-    import numpy as np
-
-    import warp as wp
-
-    wp.init()
-
-
-    @wp.struct
-    class TestStruct:
-        x: wp.vec3
-        a: wp.array(dtype=wp.vec3)
-        b: wp.array(dtype=wp.vec3)
-
-
-    @wp.kernel
-    def test_kernel(s: TestStruct):
-        tid = wp.tid()
-
-        s.b[tid] = s.a[tid] + s.x
-
-
-    @wp.kernel
-    def loss_kernel(s: TestStruct, loss: wp.array(dtype=float)):
-        tid = wp.tid()
-
-        v = s.b[tid]
-        wp.atomic_add(loss, 0, float(tid + 1) * (v[0] + 2.0 * v[1] + 3.0 * v[2]))
-
-
-    # create struct
-    ts = TestStruct()
-
-    # set members
-    ts.x = wp.vec3(1.0, 2.0, 3.0)
-    ts.a = wp.array(np.array([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]]), dtype=wp.vec3, requires_grad=True)
-    ts.b = wp.zeros(2, dtype=wp.vec3, requires_grad=True)
-
-    loss = wp.zeros(1, dtype=float, requires_grad=True)
-
-    tape = wp.Tape()
-    with tape:
-        wp.launch(test_kernel, dim=2, inputs=[ts])
-        wp.launch(loss_kernel, dim=2, inputs=[ts, loss])
-
-    tape.backward(loss)
-
-    print(loss)
-    print(tape.gradients[ts].a)
-
-
-Type Conversions
-################
-
-Warp is particularly strict regarding type conversions and does not perform *any* implicit conversion between numeric types.
-The user is responsible for ensuring types for most arithmetic operators match, e.g.: ``x = float(0.0) + int(4)`` will result in an error.
-This can be surprising for users that are accustomed to C-style conversions but avoids a class of common bugs that result from implicit conversions.
-
-.. note::
-    Warp does not currently perform implicit type conversions between numeric types.
-    Users should explicitly cast variables to compatible types using constructors like
-    ``int()``, ``float()``, ``wp.float16()``, ``wp.uint8()``, etc.
-
-Constants
----------
-
-In general, Warp kernels cannot access variables in the global Python interpreter state. One exception to this is for compile-time constants, which may be declared globally (or as class attributes) and folded into the kernel definition.
-
-Constants are defined using the ``wp.constant()`` function. An example is shown below::
-
-    TYPE_SPHERE = wp.constant(0)
-    TYPE_CUBE = wp.constant(1)
-    TYPE_CAPSULE = wp.constant(2)
-
-    @wp.kernel
-    def collide(geometry: wp.array(dtype=int)):
-
-        t = geometry[wp.tid()]
-
-        if (t == TYPE_SPHERE):
-            print("sphere")
-        if (t == TYPE_CUBE):
-            print("cube")
-        if (t == TYPE_CAPSULE):
-            print("capsule")
-
-
-.. autoclass:: constant
-
-
-Operators
-----------
-
-Boolean Operators
-#################
-
-+--------------+--------------------------------------+
-|   a and b    | True if a and b are True             |
-+--------------+--------------------------------------+
-|   a or b     | True if a or b is True               |
-+--------------+--------------------------------------+
-|   not a      | True if a is False, otherwise False  |
-+--------------+--------------------------------------+
-
-.. note::
-    Expressions such as ``if (a and b):`` currently do not perform short-circuit evaluation.
-    In this case ``b`` will also be evaluated even when ``a`` is ``False``.
-    Users should take care to ensure that secondary conditions are safe to evaluate (e.g.: do not index out of bounds) in all cases.
-
-
-Comparison Operators
-####################
-
-+----------+---------------------------------------+
-| a > b    | True if a strictly greater than b     |
-+----------+---------------------------------------+
-| a < b    | True if a strictly less than b        |
-+----------+---------------------------------------+
-| a >= b   | True if a greater than or equal to b  |
-+----------+---------------------------------------+
-| a <= b   | True if a less than or equal to b     |
-+----------+---------------------------------------+
-| a == b   | True if a equals b                    |
-+----------+---------------------------------------+
-| a != b   | True if a not equal to b              |
-+----------+---------------------------------------+
-
-Arithmetic Operators
-####################
-
-+-----------+--------------------------+
-|  a + b    | Addition                 |
-+-----------+--------------------------+
-|  a - b    | Subtraction              |
-+-----------+--------------------------+
-|  a * b    | Multiplication           |
-+-----------+--------------------------+
-|  a / b    | Floating point division  |
-+-----------+--------------------------+
-|  a // b   | Floored division         |
-+-----------+--------------------------+
-|  a ** b   | Exponentiation           |
-+-----------+--------------------------+
-|  a % b    | Modulus                  |
-+-----------+--------------------------+
-
-.. note::
-    Since implicit conversions are not performed arguments types to operators should match.
-    Users should use type constructors, e.g.: ``float()``, ``int()``, ``wp.int64()``, etc. to cast variables
-    to the correct type. Also note that the multiplication expression ``a * b`` is used to represent scalar
-    multiplication and matrix multiplication. The ``@`` operator is not currently supported.
-
-Graphs
------------
-
-Launching kernels from Python introduces significant additional overhead compared to C++ or native programs.
-To address this, Warp exposes the concept of `CUDA graphs <https://developer.nvidia.com/blog/cuda-graphs/>`_
-to allow recording large batches of kernels and replaying them with very little CPU overhead.
-
-To record a series of kernel launches use the :func:`wp.capture_begin() <capture_begin>` and
-:func:`wp.capture_end() <capture_end>` API as follows:
-
-.. code:: python
-
-    # begin capture
-    wp.capture_begin()
-
-    try:
-        # record launches
-        for i in range(100):
-            wp.launch(kernel=compute1, inputs=[a, b], device="cuda")
-    finally:
-        # end capture and return a graph object
-        graph = wp.capture_end()
-
-We strongly recommend the use of the the try-finally pattern when capturing graphs because
-:func:`wp.capture_begin <capture_begin>` disables Python garbage collection so that Warp objects are not
-garbage-collected during graph capture (CUDA does not allow memory to be deallocated during graph capture).
-:func:`wp.capture_end <capture_end>` reenables garbage collection.
-
-Once a graph has been constructed it can be executed: ::
-
-    wp.capture_launch(graph)
-
-Note that only launch calls are recorded in the graph, any Python executed outside of the kernel code will not be recorded.
-Typically it is only beneficial to use CUDA graphs when the graph will be reused or launched multiple times.
-
-.. autofunction:: capture_begin
-.. autofunction:: capture_end
-.. autofunction:: capture_launch
-
-Bounding Value Hierarchies (BVH)
---------------------------------
-
-The :class:`wp.Bvh <Bvh>` class can be used to create a BVH for a group of bounding volumes. This object can then be traversed
-to determine which parts are intersected by a ray using :func:`bvh_query_ray` and which parts are fully contained
-within a certain bounding volume using :func:`bvh_query_aabb`.
-
-The following snippet demonstrates how to create a :class:`wp.Bvh <Bvh>` object from 100 random bounding volumes:
-
-.. code:: python
-
-    rng = np.random.default_rng(123)
-
-    num_bounds = 100
-    lowers = rng.random(size=(num_bounds, 3)) * 5.0
-    uppers = lowers + rng.random(size=(num_bounds, 3)) * 5.0
-
-    device_lowers = wp.array(lowers, dtype=wp.vec3, device="cuda:0")
-    device_uppers = wp.array(uppers, dtype=wp.vec3, device="cuda:0")
-
-    bvh = wp.Bvh(device_lowers, device_uppers)
-
-.. autoclass:: Bvh
-    :members:
-
-Example: BVH Ray Traversal
-##########################
-
-An example of performing a ray traversal on the data structure is as follows:
-
-.. code:: python
-
-    @wp.kernel
-    def bvh_query_ray(
-        bvh_id: wp.uint64,
-        start: wp.vec3,
-        dir: wp.vec3,
-        bounds_intersected: wp.array(dtype=wp.bool),
-    ):
-        query = wp.bvh_query_ray(bvh_id, start, dir)
-        bounds_nr = wp.int32(0)
-
-        while wp.bvh_query_next(query, bounds_nr):
-            # The ray intersects the volume with index bounds_nr
-            bounds_intersected[bounds_nr] = True
-
-
-    bounds_intersected = wp.zeros(shape=(num_bounds), dtype=wp.bool, device="cuda:0")
-    query_start = wp.vec3(0.0, 0.0, 0.0)
-    query_dir = wp.normalize(wp.vec3(1.0, 1.0, 1.0))
-
-    wp.launch(
-        kernel=bvh_query_ray,
-        dim=1,
-        inputs=[bvh.id, query_start, query_dir, bounds_intersected],
-        device="cuda:0",
-    )
-
-The Warp kernel ``bvh_query_ray`` is launched with a single thread, provided the unique :class:`uint64`
-identifier of the :class:`wp.Bvh <Bvh>` object, parameters describing the ray, and an array to store the results.
-In ``bvh_query_ray``, :func:`wp.bvh_query_ray() <bvh_query_ray>` is called once to obtain an object that is stored in the
-variable ``query``. An integer is also allocated as ``bounds_nr`` to store the volume index of the traversal.
-A while statement is used for the actual traversal using :func:`wp.bvh_query_next() <bvh_query_next>`,
-which returns ``True`` as long as there are intersecting bounds.
-
-Example: BVH Volume Traversal
-#############################
-
-Similar to the ray-traversal example, we can perform volume traversal to find the volumes that are fully contained
-within a specified bounding box.
-
-.. code:: python
-
-    @wp.kernel
-    def bvh_query_aabb(
-        bvh_id: wp.uint64,
-        lower: wp.vec3,
-        upper: wp.vec3,
-        bounds_intersected: wp.array(dtype=wp.bool),
-    ):
-        query = wp.bvh_query_aabb(bvh_id, lower, upper)
-        bounds_nr = wp.int32(0)
-
-        while wp.bvh_query_next(query, bounds_nr):
-            # The volume with index bounds_nr is fully contained
-            # in the (lower,upper) bounding box
-            bounds_intersected[bounds_nr] = True
-
-
-    bounds_intersected = wp.zeros(shape=(num_bounds), dtype=wp.bool, device="cuda:0")
-    query_lower = wp.vec3(4.0, 4.0, 4.0)
-    query_upper = wp.vec3(6.0, 6.0, 6.0)
-
-    wp.launch(
-        kernel=bvh_query_aabb,
-        dim=1,
-        inputs=[bvh.id, query_lower, query_upper, bounds_intersected],
-        device="cuda:0",
-    )
-
-The kernel is nearly identical to the ray-traversal example, except we obtain ``query`` using
-:func:`wp.bvh_query_aabb() <bvh_query_aabb>`.
-
-Meshes
-------
-
-Warp provides a ``wp.Mesh`` class to manage triangle mesh data. To create a mesh users provide a points, indices and optionally a velocity array::
-
-    mesh = wp.Mesh(points, indices, velocities)
-
-.. note::
-    Mesh objects maintain references to their input geometry buffers. All buffers should live on the same device.
-
-Meshes can be passed to kernels using their ``id`` attribute which uniquely identifies the mesh by a unique ``uint64`` value.
-Once inside a kernel you can perform geometric queries against the mesh such as ray-casts or closest point lookups::
-
-    @wp.kernel
-    def raycast(mesh: wp.uint64,
-                ray_origin: wp.array(dtype=wp.vec3),
-                ray_dir: wp.array(dtype=wp.vec3),
-                ray_hit: wp.array(dtype=wp.vec3)):
-
-        tid = wp.tid()
-
-        t = float(0.0)      # hit distance along ray
-        u = float(0.0)      # hit face barycentric u
-        v = float(0.0)      # hit face barycentric v
-        sign = float(0.0)   # hit face sign
-        n = wp.vec3()       # hit face normal
-        f = int(0)          # hit face index
-
-        color = wp.vec3()
-
-        # ray cast against the mesh
-        if wp.mesh_query_ray(mesh, ray_origin[tid], ray_dir[tid], 1.e+6, t, u, v, sign, n, f):
-
-            # if we got a hit then set color to the face normal
-            color = n*0.5 + wp.vec3(0.5, 0.5, 0.5)
-
-        ray_hit[tid] = color
-
-
-Users may update mesh vertex positions at runtime simply by modifying the points buffer.
-After modifying point locations users should call ``Mesh.refit()`` to rebuild the bounding volume hierarchy (BVH) structure and ensure that queries work correctly.
-
-.. note::
-    Updating Mesh topology (indices) at runtime is not currently supported. Users should instead recreate a new Mesh object.
-
-.. autoclass:: Mesh
-    :members:
-
-Hash Grids
-----------
-
-Many particle-based simulation methods such as the Discrete Element Method (DEM), or Smoothed Particle Hydrodynamics (SPH), involve iterating over spatial neighbors to compute force interactions. Hash grids are a well-established data structure to accelerate these nearest neighbor queries, and particularly well-suited to the GPU.
-
-To support spatial neighbor queries Warp provides a ``HashGrid`` object that may be created as follows::
-
-    grid = wp.HashGrid(dim_x=128, dim_y=128, dim_z=128, device="cuda")
-
-    grid.build(points=p, radius=r)
-
-``p`` is an array of ``wp.vec3`` point positions, and ``r`` is the radius to use when building the grid.
-Neighbors can then be iterated over inside the kernel code using :func:`wp.hash_grid_query() <hash_grid_query>`
-and :func:`wp.hash_grid_query_next() <hash_grid_query_next>` as follows:
-
-.. code:: python
-
-    @wp.kernel
-    def sum(grid : wp.uint64,
-            points: wp.array(dtype=wp.vec3),
-            output: wp.array(dtype=wp.vec3),
-            radius: float):
-
-        tid = wp.tid()
-
-        # query point
-        p = points[tid]
-
-        # create grid query around point
-        query = wp.hash_grid_query(grid, p, radius)
-        index = int(0)
-
-        sum = wp.vec3()
-
-        while(wp.hash_grid_query_next(query, index)):
-
-            neighbor = points[index]
-
-            # compute distance to neighbor point
-            dist = wp.length(p-neighbor)
-            if (dist <= radius):
-                sum += neighbor
-
-        output[tid] = sum
-
-.. note::
-    The ``HashGrid`` query will give back all points in *cells* that fall inside the query radius.
-    When there are hash conflicts it means that some points outside of query radius will be returned, and users should
-    check the distance themselves inside their kernels. The reason the query doesn't do the check itself for each
-    returned point is because it's common for kernels to compute the distance themselves, so it would redundant to
-    check/compute the distance twice.
-
-
-.. autoclass:: HashGrid
-    :members:
-
-Volumes
--------
-
-Sparse volumes are incredibly useful for representing grid data over large domains, such as signed distance fields
-(SDFs) for complex objects, or velocities for large-scale fluid flow. Warp supports reading sparse volumetric grids
-stored using the `NanoVDB <https://developer.nvidia.com/nanovdb>`_ standard. Users can access voxels directly
-or use built-in closest-point or trilinear interpolation to sample grid data from world or local space.
-
-Volume objects can be created directly from Warp arrays containing a NanoVDB grid, from the contents of a
-standard ``.nvdb`` file using :func:`load_from_nvdb() <warp.Volume.load_from_nvdb>`,
-or from a dense 3D NumPy array using :func:`load_from_numpy() <warp.Volume.load_from_numpy>`.
-
-Volumes can also be created using :func:`allocate() <warp.Volume.allocate>` or
-:func:`allocate_by_tiles() <warp.Volume.allocate_by_tiles>`. The values for a Volume object can be modified in a Warp
-kernel using :func:`wp.volume_store_f() <warp.volume_store_f>`, :func:`wp.volume_store_v() <warp.volume_store_v>`, and
-:func:`wp.volume_store_i() <warp.volume_store_i>`.
-
-.. note::
-    Warp does not currently support modifying the topology of sparse volumes at runtime.
-
-Below we give an example of creating a Volume object from an existing NanoVDB file::
-
-    # open NanoVDB file on disk
-    file = open("mygrid.nvdb", "rb")
-
-    # create Volume object
-    volume = wp.Volume.load_from_nvdb(file, device="cpu")
-
-.. note::
-    Files written by the NanoVDB library, commonly marked by the ``.nvdb`` extension, can contain multiple grids with
-    various compression methods, but a :class:`Volume` object represents a single NanoVDB grid therefore only files with
-    a single grid are supported. NanoVDB's uncompressed and zip-compressed file formats are supported.
-
-To sample the volume inside a kernel we pass a reference to it by ID, and use the built-in sampling modes::
-
-    @wp.kernel
-    def sample_grid(volume: wp.uint64,
-                    points: wp.array(dtype=wp.vec3),
-                    samples: wp.array(dtype=float)):
-
-        tid = wp.tid()
-
-        # load sample point in world-space
-        p = points[tid]
-
-        # transform position to the volume's local-space
-        q = wp.volume_world_to_index(volume, p)
-
-        # sample volume with trilinear interpolation
-        f = wp.volume_sample_f(volume, q, wp.Volume.LINEAR)
-
-        # write result
-        samples[tid] = f
-
-.. autoclass:: Volume
-    :members:
-    :undoc-members:
-
-.. seealso:: `Reference <functions.html#volumes>`__ for the volume functions available in kernels.
-
-Differentiability
------------------
-
-By default, Warp generates a forward and backward (adjoint) version of each kernel definition.
-Buffers that participate in the chain of computation should be created with ``requires_grad=True``, for example::
-
-    a = wp.zeros(1024, dtype=wp.vec3, device="cuda", requires_grad=True)
-
-The ``wp.Tape`` class can then be used to record kernel launches, and replay them to compute the gradient of a scalar loss function with respect to the kernel inputs::
-
-    tape = wp.Tape()
-
-    # forward pass
-    with tape:
-        wp.launch(kernel=compute1, inputs=[a, b], device="cuda")
-        wp.launch(kernel=compute2, inputs=[c, d], device="cuda")
-        wp.launch(kernel=loss, inputs=[d, l], device="cuda")
-
-    # reverse pass
-    tape.backward(l)
-
-After the backward pass has completed, the gradients with respect to the inputs are available via a mapping in the :class:`wp.Tape <Tape>` object::
-
-    # gradient of loss with respect to input a
-    print(tape.gradients[a])
-
-Note that gradients are accumulated on the participating buffers, so if you wish to reuse the same buffers for multiple backward passes you should first zero the gradients::
-
-    tape.zero()
-
-.. note::
-
-    Warp uses a source-code transformation approach to auto-differentiation.
-    In this approach, the backwards pass must keep a record of intermediate values computed during the forward pass.
-    This imposes some restrictions on what kernels can do and still be differentiable:
-
-    * Dynamic loops should not mutate any previously declared local variable. This means the loop must be side-effect free.
-      A simple way to ensure this is to move the loop body into a separate function.
-      Static loops that are unrolled at compile time do not have this restriction and can perform any computation.
-
-    * Kernels should not overwrite any previously used array values except to perform simple linear add/subtract operations (e.g. via :func:`wp.atomic_add() <atomic_add>`)
-
-.. autoclass:: Tape
-    :members:
-
-Jacobians
-#########
-
-To compute the Jacobian matrix :math:`J\in\mathbb{R}^{m\times n}` of a multi-valued function :math:`f: \mathbb{R}^n \to \mathbb{R}^m`, we can evaluate an entire row of the Jacobian in parallel by finding the Jacobian-vector product :math:`J^\top \mathbf{e}`. The vector :math:`\mathbf{e}\in\mathbb{R}^m` selects the indices in the output buffer to differentiate with respect to.
-In Warp, instead of passing a scalar loss buffer to the ``tape.backward()`` method, we pass a dictionary ``grads`` mapping from the function output array to the selection vector :math:`\mathbf{e}` having the same type::
-
-    # compute the Jacobian for a function of single output
-    jacobian = np.empty((output_dim, input_dim), dtype=np.float32)
-    tape = wp.Tape()
-    with tape:
-        output_buffer = launch_kernels_to_be_differentiated(input_buffer)
-    for output_index in range(output_dim):
-        # select which row of the Jacobian we want to compute
-        select_index = np.zeros(output_dim)
-        select_index[output_index] = 1.0
-        e = wp.array(select_index, dtype=wp.float32)
-        # pass input gradients to the output buffer to apply selection
-        tape.backward(grads={output_buffer: e})
-        q_grad_i = tape.gradients[input_buffer]
-        jacobian[output_index, :] = q_grad_i.numpy()
-        tape.zero()
-
-When we run simulations independently in parallel, the Jacobian corresponding to the entire system dynamics is a block-diagonal matrix. In this case, we can compute the Jacobian in parallel for all environments by choosing a selection vector that has the output indices active for all environment copies. For example, to get the first rows of the Jacobians of all environments, :math:`\mathbf{e}=[\begin{smallmatrix}1 & 0 & 0 & \dots & 1 & 0 & 0 & \dots\end{smallmatrix}]^\top`, to compute the second rows, :math:`\mathbf{e}=[\begin{smallmatrix}0 & 1 & 0 & \dots & 0 & 1 & 0 & \dots\end{smallmatrix}]^\top`, etc.::
-
-    # compute the Jacobian for a function over multiple environments in parallel
-    jacobians = np.empty((num_envs, output_dim, input_dim), dtype=np.float32)
-    tape = wp.Tape()
-    with tape:
-        output_buffer = launch_kernels_to_be_differentiated(input_buffer)
-    for output_index in range(output_dim):
-        # select which row of the Jacobian we want to compute
-        select_index = np.zeros(output_dim)
-        select_index[output_index] = 1.0
-        # assemble selection vector for all environments (can be precomputed)
-        e = wp.array(np.tile(select_index, num_envs), dtype=wp.float32)
-        tape.backward(grads={output_buffer: e})
-        q_grad_i = tape.gradients[input_buffer]
-        jacobians[:, output_index, :] = q_grad_i.numpy().reshape(num_envs, input_dim)
-        tape.zero()
-
-
-Custom Gradient Functions
-#########################
-
-Warp supports custom gradient function definitions for user-defined Warp functions.
-This allows users to define code that should replace the automatically generated derivatives.
-
-To differentiate a function :math:`h(x) = f(g(x))` that has a nested call to function :math:`g(x)`, the chain rule is evaluated in the automatic differentiation of :math:`h(x)`:
-
-.. math::
-
-    h^\prime(x) = f^\prime({\color{green}{\underset{\textrm{replay}}{g(x)}}}) {\color{blue}{\underset{\textrm{grad}}{g^\prime(x)}}}
-
-This implies that a function to be compatible with the autodiff engine needs to provide an implementation of its forward version
-:math:`\color{green}{g(x)}`, which we refer to as "replay" function (that matches the original function definition by default),
-and its derivative :math:`\color{blue}{g^\prime(x)}`, referred to as "grad".
-
-Both the replay and the grad implementations can be customized by the user. They are defined as follows:
-
-.. list-table:: Customizing the replay and grad versions of function ``myfunc``
-    :widths: 100
-    :header-rows: 0
-
-    * - Forward Function
-    * - .. code-block:: python
-
-            @wp.func
-            def myfunc(in1: InType1, ..., inN: InTypeN) -> OutType1, ..., OutTypeM:
-                return out1, ..., outM
-
-    * - Custom Replay Function
-    * - .. code-block:: python
-
-            @wp.func_replay(myfunc)
-            def replay_myfunc(in1: InType1, ..., inN: InTypeN) -> OutType1, ..., OutTypeM:
-                # Custom forward computations to be executed in the backward pass of a
-                # function calling `myfunc` go here
-                # Ensure the output variables match the original forward definition
-                return out1, ..., outM
-
-    * - Custom Grad Function
-    * - .. code-block:: python
-
-            @wp.func_grad(myfunc)
-            def adj_myfunc(in1: InType1, ..., inN: InTypeN, adj_out1: OutType1, ..., adj_outM: OutTypeM):
-                # Custom adjoint code goes here
-                # Update the partial derivatives for the inputs as follows:
-                wp.adjoint[in1] += ...
-                ...
-                wp.adjoint[inN] += ...
-
-.. note:: It is currently not possible to define custom replay or grad functions for functions that
-    have generic arguments, e.g. ``Any`` or ``wp.array(dtype=Any)``. Replay or grad functions that
-    themselves use generic arguments are also not yet supported.
-
-Example 1: Custom Grad Function
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-In the following, we define a Warp function ``safe_sqrt`` that computes the square root of a number::
-
-    @wp.func
-    def safe_sqrt(x: float):
-        return wp.sqrt(x)
-
-To evaluate this function, we define a kernel that applies ``safe_sqrt`` to an array of input values::
-
-    @wp.kernel
-    def run_safe_sqrt(xs: wp.array(dtype=float), output: wp.array(dtype=float)):
-        i = wp.tid()
-        output[i] = safe_sqrt(xs[i])
-
-Calling the kernel for an array of values ``[1.0, 2.0, 0.0]`` yields the expected outputs, the gradients are finite except for the zero input::
-
-    xs = wp.array([1.0, 2.0, 0.0], dtype=wp.float32, requires_grad=True)
-    ys = wp.zeros_like(xs)
-    tape = wp.Tape()
-    with tape:
-        wp.launch(run_safe_sqrt, dim=len(xs), inputs=[xs], outputs=[ys])
-    tape.backward(grads={ys: wp.array(np.ones(len(xs)), dtype=wp.float32)})
-    print("ys     ", ys)
-    print("xs.grad", xs.grad)
-
-    # ys      [1.   1.4142135   0. ]
-    # xs.grad [0.5  0.35355338  inf]
-
-It is often desired to catch nonfinite gradients in the computation graph as they may cause the entire gradient computation to be nonfinite.
-To do so, we can define a custom gradient function that replaces the adjoint function for ``safe_sqrt`` which is automatically generated by
-decorating the custom gradient code via ``@wp.func_grad(safe_sqrt)``::
-
-    @wp.func_grad(safe_sqrt)
-    def adj_safe_sqrt(x: float, adj_ret: float):
-        if x > 0.0:
-            wp.adjoint[x] += 1.0 / (2.0 * wp.sqrt(x)) * adj_ret
-
-.. note:: The function signature of the custom grad code consists of the input arguments of the forward function plus the adjoint variables of the
-    forward function outputs. To access and modify the partial derivatives of the input arguments, we use the ``wp.adjoint`` dictionary.
-    The keys of this dictionary are the input arguments of the forward function, and the values are the partial derivatives of the forward function
-    output with respect to the input argument.
-
-
-Example 2: Custom Replay Function
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
-In the following, we increment an array index in each thread via :func:`wp.atomic_add() <atomic_add>` and compute
-the square root of an input array at the incremented index::
-
-    @wp.kernel
-    def test_add(counter: wp.array(dtype=int), input: wp.array(dtype=float), output: wp.array(dtype=float)):
-        idx = wp.atomic_add(counter, 0, 1)
-        output[idx] = wp.sqrt(input[idx])
-
-    def main():
-        dim = 16
-        use_reversible_increment = False
-        input = wp.array(np.arange(1, dim + 1), dtype=wp.float32, requires_grad=True)
-        counter = wp.zeros(1, dtype=wp.int32)
-        thread_ids = wp.zeros(dim, dtype=wp.int32)
-        output = wp.zeros(dim, dtype=wp.float32, requires_grad=True)
-        tape = wp.Tape()
-        with tape:
-            if use_reversible_increment:
-                wp.launch(test_add_diff, dim, inputs=[counter, thread_ids, input], outputs=[output])
-            else:
-                wp.launch(test_add, dim, inputs=[counter, input], outputs=[output])
-
-        print("counter:    ", counter.numpy())
-        print("thread_ids: ", thread_ids.numpy())
-        print("input:      ", input.numpy())
-        print("output:     ", output.numpy())
-
-        tape.backward(grads={
-            output: wp.array(np.ones(dim), dtype=wp.float32)
-        })
-        print("input.grad: ", input.grad.numpy())
-
-    if __name__ == "__main__":
-        main()
-
-The output of the above code is:
-
-.. code-block:: js
-
-    counter:     [8]
-    thread_ids:  [0 0 0 0 0 0 0 0]
-    input:       [1. 2. 3. 4. 5. 6. 7. 8.]
-    output:      [1.  1.4142135  1.7320508  2.  2.236068  2.4494898  2.6457512  2.828427]
-    input.grad:  [4. 0. 0. 0. 0. 0. 0. 0.]
-
-The gradient of the input is incorrect because the backward pass involving the atomic operation ``wp.atomic_add()`` does not know which thread ID corresponds
-to which input value.
-The index returned by the adjoint of ``wp.atomic_add()`` is always zero so that the gradient the first entry of the input array,
-i.e. :math:`\frac{1}{2\sqrt{1}} = 0.5`, is accumulated ``dim`` times (hence ``input.grad[0] == 4.0`` and all other entries zero).
-
-To fix this, we define a new Warp function ``reversible_increment()`` with a custom *replay* definition that stores the thread ID in a separate array::
-
-    @wp.func
-    def reversible_increment(
-        buf: wp.array(dtype=int),
-        buf_index: int,
-        value: int,
-        thread_values: wp.array(dtype=int),
-        tid: int
-    ):
-        next_index = wp.atomic_add(buf, buf_index, value)
-        # store which thread ID corresponds to which index for the backward pass
-        thread_values[tid] = next_index
-        return next_index
-
-
-    @wp.func_replay(reversible_increment)
-    def replay_reversible_increment(
-        buf: wp.array(dtype=int),
-        buf_index: int,
-        value: int,
-        thread_values: wp.array(dtype=int),
-        tid: int
-    ):
-        return thread_values[tid]
-
-
-Instead of running ``reversible_increment()``, the custom replay code in ``replay_reversible_increment()`` is now executed
-during forward phase in the backward pass of the function calling ``reversible_increment()``.
-We first stored the array index to each thread ID in the forward pass, and now we retrieve the array index for each thread ID in the backward pass.
-That way, the backward pass can reproduce the same addition operation as in the forward pass with exactly the same operands per thread.
-
-.. warning:: The function signature of the custom replay code must match the forward function signature.
-
-To use our function we write the following kernel::
-
-    @wp.kernel
-    def test_add_diff(
-        counter: wp.array(dtype=int),
-        thread_ids: wp.array(dtype=int),
-        input: wp.array(dtype=float),
-        output: wp.array(dtype=float)
-    ):
-        tid = wp.tid()
-        idx = reversible_increment(counter, 0, 1, thread_ids, tid)
-        output[idx] = wp.sqrt(input[idx])
-
-Running the ``test_add_diff`` kernel via the previous ``main`` function with ``use_reversible_increment = True``, we now compute correct gradients
-for the input array:
-
-.. code-block:: js
-
-    counter:     [8]
-    thread_ids:  [0 1 2 3 4 5 6 7]
-    input:       [1. 2. 3. 4. 5. 6. 7. 8.]
-    output:      [1.   1.4142135   1.7320508   2.    2.236068   2.4494898   2.6457512   2.828427  ]
-    input.grad:  [0.5  0.35355338  0.28867513  0.25  0.2236068  0.20412414  0.18898225  0.17677669]
-
-Custom Native Functions
------------------------
-
-Users may insert native C++/CUDA code in Warp kernels using ``@func_native`` decorated functions.
-These accept native code as strings that get compiled after code generation, and are called within ``@wp.kernel`` functions.
-For example::
-
-    snippet = """
-        __shared__ int s[128];
-
-        s[tid] = d[tid];
-        __syncthreads();
-        d[tid] = s[N - tid - 1];
-        """
-
-    @wp.func_native(snippet)
-    def reverse(d: wp.array(dtype=int), N: int, tid: int):
-        ...
-
-    @wp.kernel
-    def reverse_kernel(d: wp.array(dtype=int), N: int):
-        tid = wp.tid()
-        reverse(d, N, tid)
-
-    N = 128
-    x = wp.array(np.arange(N, dtype=int), dtype=int, device=device)
-
-    wp.launch(kernel=reverse_kernel, dim=N, inputs=[x, N], device=device)
-
-Notice the use of shared memory here: the Warp library does not expose shared memory as a feature, but the CUDA compiler will
-readily accept the above snippet. This means CUDA features not exposed in Warp are still accessible in Warp scripts.
-Warp kernels meant for the CPU won't be able to leverage CUDA features of course, but this same mechanism supports pure C++ snippets as well.
-
-Please bear in mind the following: the thread index in your snippet should be computed in a ``@wp.kernel`` and passed to your snippet,
-as in the above example. This means your ``@wp.func_native`` function signature should include the variables used in your snippet, 
-as well as a thread index of type ``int``. The function body itself should be stubbed with ``...`` (the snippet will be inserted during compilation).
-
-Should you wish to record your native function on the tape and then subsequently rewind the tape, you must include an adjoint snippet
-alongside your snippet as an additional input to the decorator, as in the following example::
-
-    snippet = """
-    out[tid] = a * x[tid] + y[tid];
-    """
-    adj_snippet = """
-    adj_a = x[tid] * adj_out[tid];
-    adj_x[tid] = a * adj_out[tid];
-    adj_y[tid] = adj_out[tid];
-    """
-
-    @wp.func_native(snippet, adj_snippet)
-    def saxpy(
-        a: wp.float32,
-        x: wp.array(dtype=wp.float32),
-        y: wp.array(dtype=wp.float32),
-        out: wp.array(dtype=wp.float32),
-        tid: int,
-    ):
-        ...
-
-    @wp.kernel
-    def saxpy_kernel(
-        a: wp.float32,
-        x: wp.array(dtype=wp.float32),
-        y: wp.array(dtype=wp.float32),
-        out: wp.array(dtype=wp.float32)
-    ):
-        tid = wp.tid()
-        saxpy(a, x, y, out, tid)
-
-    N = 128
-    a = 2.0
-    x = wp.array(np.arange(N, dtype=np.float32), dtype=wp.float32, device=device, requires_grad=True)
-    y = wp.zeros_like(x1)
-    out = wp.array(np.arange(N, dtype=np.float32), dtype=wp.float32, device=device)
-    adj_out = wp.array(np.ones(N, dtype=np.float32), dtype=wp.float32, device=device)
-
-    tape = wp.Tape()
-
-    with tape:
-        wp.launch(kernel=saxpy_kernel, dim=N, inputs=[a, x, y], outputs=[out], device=device)
-
-    tape.backward(grads={out: adj_out})
-
-Profiling
----------
-
-``wp.ScopedTimer`` objects can be used to gain some basic insight into the performance of Warp applications:
-
-.. code:: python
-
-    with wp.ScopedTimer("grid build"):
-        self.grid.build(self.x, self.point_radius)
-
-This results in a printout at runtime to the standard output stream like:
-
-.. code:: console
-
-    grid build took 0.06 ms
-
-The ``wp.ScopedTimer`` object does not synchronize (e.g. by calling ``wp.synchronize()``)
-upon exiting the ``with`` statement, so this can lead to misleading numbers if the body
-of the ``with`` statement launches device kernels.
-
-When a ``wp.ScopedTimer`` object is passed ``use_nvtx=True`` as an argument, the timing functionality is replaced by calls
-to ``nvtx.start_range()`` and ``nvtx.end_range()``:
-
-.. code:: python
-
-    with wp.ScopedTimer("grid build", use_nvtx=True, color="cyan"):
-        self.grid.build(self.x, self.point_radius)
-
-These range annotations can then be collected by a tool like `NVIDIA Nsight Systems <https://developer.nvidia.com/nsight-systems>`_
-for visualization on a timeline, e.g.:
-
-.. code:: console
-   
-    $ nsys profile python warp_application.py
-
-This code snippet also demonstrates the use of the ``color`` argument to specify a color
-for the range, which may be a number representing the ARGB value or a recognized string
-(refer to `colors.py <https://github.com/NVIDIA/NVTX/blob/release-v3/python/nvtx/colors.py>`_ for
-additional color examples).
-The `nvtx module <https://github.com/NVIDIA/NVTX>`_ must be
-installed in the Python environment for this capability to work.
-An equivalent way to create an NVTX range without using ``wp.ScopedTimer`` is:
-
-.. code:: python
-
-    import nvtx
-
-    with nvtx.annotate("grid build", color="cyan"):
-        self.grid.build(self.x, self.point_radius)
-
-This form may be more convenient if the user does not need to frequently switch
-between timer and NVTX capabilities of ``wp.ScopedTimer``. 
-
-.. autoclass:: warp.ScopedTimer
diff --git a/docs/_build/html/_sources/modules/sim.rst.txt b/docs/_build/html/_sources/modules/sim.rst.txt
deleted file mode 100644
index bec52af0b..000000000
--- a/docs/_build/html/_sources/modules/sim.rst.txt
+++ /dev/null
@@ -1,157 +0,0 @@
-warp.sim
-====================
-
-.. currentmodule:: warp.sim
-
-..
-    .. toctree::
-    :maxdepth: 2
-
-Warp includes a simulation module ``warp.sim`` that includes many common physical simulation models, and integrators for explicit and implicit time-stepping.
-
-.. note:: The simulation model is under construction and should be expected to change rapidly, please treat this section as work in progress.
-
-
-Model
------
-
-.. autoclass:: ModelBuilder
-    :members:
-
-.. autoclass:: Model
-    :members:
-
-.. autoclass:: JointAxis
-    :members:
-
-.. autoclass:: Mesh
-   :members:
-
-.. autoclass:: SDF
-   :members:
-
-.. _Joint types:
-
-Joint types
-^^^^^^^^^^^^^^
-
-.. data:: JOINT_PRISMATIC
-
-    Prismatic (slider) joint
-
-.. data:: JOINT_REVOLUTE
-
-    Revolute (hinge) joint
-
-.. data:: JOINT_BALL
-
-    Ball (spherical) joint with quaternion state representation
-
-.. data:: JOINT_FIXED
-
-    Fixed (static) joint
-
-.. data:: JOINT_FREE
-
-    Free (floating) joint
-
-.. data:: JOINT_COMPOUND
-
-    Compound joint with 3 rotational degrees of freedom
-
-.. data:: JOINT_UNIVERSAL
-
-    Universal joint with 2 rotational degrees of freedom
-
-.. data:: JOINT_DISTANCE
-
-    Distance joint that keeps two bodies at a distance within its joint limits (only supported in :class:`XPBDIntegrator` at the moment)
-
-.. data:: JOINT_D6
-
-    Generic D6 joint with up to 3 translational and 3 rotational degrees of freedom
-
-.. _Joint modes:
-
-Joint modes
-^^^^^^^^^^^^^^
-
-Joint modes control the behavior of joint axes and can be used to implement joint position or velocity drives.
-
-.. data:: JOINT_MODE_LIMIT
-
-    No target or velocity control is applied, the joint is limited to its joint limits
-
-.. data:: JOINT_MODE_TARGET_POSITION
-
-    The joint is driven to a target position
-
-.. data:: JOINT_MODE_TARGET_VELOCITY
-   
-    The joint is driven to a target velocity
-
-State
---------------
-
-.. autoclass:: State
-    :members:
-
-.. _FK-IK:
-
-Forward / Inverse Kinematics
-----------------------------
-
-Articulated rigid-body mechanisms are kinematically described by the joints that connect the bodies as well as the relative relative transform from the parent and child body to the respective anchor frames of the joint in the parent and child body:
-
-.. image:: /img/joint_transforms.png
-   :width: 400
-   :align: center
-
-.. list-table:: Variable names in the kernels from articulation.py
-   :widths: 10 90
-   :header-rows: 1
-
-   * - Symbol
-     - Description
-   * - x_wp
-     - World transform of the parent body (stored at :attr:`State.body_q`)
-   * - x_wc
-     - World transform of the child body (stored at :attr:`State.body_q`)
-   * - x_pj
-     - Transform from the parent body to the joint parent anchor frame (defined by :attr:`Model.joint_X_p`)
-   * - x_cj
-     - Transform from the child body to the joint child anchor frame (defined by :attr:`Model.joint_X_c`)
-   * - x_j
-     - Joint transform from the joint parent anchor frame to the joint child anchor frame
-
-In the forward kinematics, the joint transform is determined by the joint coordinates (generalized joint positions :attr:`State.body_q` and velocities :attr:`State.body_qd`).
-Given the parent body's world transform :math:`x_{wp}` and the joint transform :math:`x_{j}`, the child body's world transform :math:`x_{wc}` is computed as:
-
-.. math::
-   x_{wc} = x_{wp} \cdot x_{pj} \cdot x_{j} \cdot x_{cj}^{-1}.
-
-.. autofunction:: eval_fk
-
-.. autofunction:: eval_ik
-
-Integrators
---------------
-
-.. autoclass:: SemiImplicitIntegrator
-    :members:
-
-.. autoclass:: XPBDIntegrator
-    :members:
-
-Importers
---------------
-
-Warp sim supports the loading of simulation models from URDF, MuJoCo (MJCF), and USD Physics files.
-
-.. autofunction:: parse_urdf
-
-.. autofunction:: parse_mjcf
-
-.. autofunction:: parse_usd
-
-.. autofunction:: resolve_usd_from_url
\ No newline at end of file
diff --git a/docs/_build/html/_sources/modules/sparse.rst.txt b/docs/_build/html/_sources/modules/sparse.rst.txt
deleted file mode 100644
index 42fb05ba1..000000000
--- a/docs/_build/html/_sources/modules/sparse.rst.txt
+++ /dev/null
@@ -1,19 +0,0 @@
-warp.sparse
-===============================
-
-.. currentmodule:: warp.sparse
-
-..
-    .. toctree::
-    :maxdepth: 2
-
-Warp includes a sparse linear algebra module ``warp.sparse`` that implements some common operations for manipulating sparse matrices that arise in simulation.
-
-Sparse Matrices
--------------------------
-
-Currently `warp.sparse` supports Block Sparse Row (BSR) matrices, the BSR format can also be used to represent Compressed Sparse Row (CSR) matrices as a special case with a 1x1 block size.
-
-.. automodule:: warp.sparse
-    :members:
-
diff --git a/docs/_build/html/_static/basic.css b/docs/_build/html/_static/basic.css
deleted file mode 100644
index 30fee9d0f..000000000
--- a/docs/_build/html/_static/basic.css
+++ /dev/null
@@ -1,925 +0,0 @@
-/*
- * basic.css
- * ~~~~~~~~~
- *
- * Sphinx stylesheet -- basic theme.
- *
- * :copyright: Copyright 2007-2023 by the Sphinx team, see AUTHORS.
- * :license: BSD, see LICENSE for details.
- *
- */
-
-/* -- main layout ----------------------------------------------------------- */
-
-div.clearer {
-    clear: both;
-}
-
-div.section::after {
-    display: block;
-    content: '';
-    clear: left;
-}
-
-/* -- relbar ---------------------------------------------------------------- */
-
-div.related {
-    width: 100%;
-    font-size: 90%;
-}
-
-div.related h3 {
-    display: none;
-}
-
-div.related ul {
-    margin: 0;
-    padding: 0 0 0 10px;
-    list-style: none;
-}
-
-div.related li {
-    display: inline;
-}
-
-div.related li.right {
-    float: right;
-    margin-right: 5px;
-}
-
-/* -- sidebar --------------------------------------------------------------- */
-
-div.sphinxsidebarwrapper {
-    padding: 10px 5px 0 10px;
-}
-
-div.sphinxsidebar {
-    float: left;
-    width: 230px;
-    margin-left: -100%;
-    font-size: 90%;
-    word-wrap: break-word;
-    overflow-wrap : break-word;
-}
-
-div.sphinxsidebar ul {
-    list-style: none;
-}
-
-div.sphinxsidebar ul ul,
-div.sphinxsidebar ul.want-points {
-    margin-left: 20px;
-    list-style: square;
-}
-
-div.sphinxsidebar ul ul {
-    margin-top: 0;
-    margin-bottom: 0;
-}
-
-div.sphinxsidebar form {
-    margin-top: 10px;
-}
-
-div.sphinxsidebar input {
-    border: 1px solid #98dbcc;
-    font-family: sans-serif;
-    font-size: 1em;
-}
-
-div.sphinxsidebar #searchbox form.search {
-    overflow: hidden;
-}
-
-div.sphinxsidebar #searchbox input[type="text"] {
-    float: left;
-    width: 80%;
-    padding: 0.25em;
-    box-sizing: border-box;
-}
-
-div.sphinxsidebar #searchbox input[type="submit"] {
-    float: left;
-    width: 20%;
-    border-left: none;
-    padding: 0.25em;
-    box-sizing: border-box;
-}
-
-
-img {
-    border: 0;
-    max-width: 100%;
-}
-
-/* -- search page ----------------------------------------------------------- */
-
-ul.search {
-    margin: 10px 0 0 20px;
-    padding: 0;
-}
-
-ul.search li {
-    padding: 5px 0 5px 20px;
-    background-image: url(file.png);
-    background-repeat: no-repeat;
-    background-position: 0 7px;
-}
-
-ul.search li a {
-    font-weight: bold;
-}
-
-ul.search li p.context {
-    color: #888;
-    margin: 2px 0 0 30px;
-    text-align: left;
-}
-
-ul.keywordmatches li.goodmatch a {
-    font-weight: bold;
-}
-
-/* -- index page ------------------------------------------------------------ */
-
-table.contentstable {
-    width: 90%;
-    margin-left: auto;
-    margin-right: auto;
-}
-
-table.contentstable p.biglink {
-    line-height: 150%;
-}
-
-a.biglink {
-    font-size: 1.3em;
-}
-
-span.linkdescr {
-    font-style: italic;
-    padding-top: 5px;
-    font-size: 90%;
-}
-
-/* -- general index --------------------------------------------------------- */
-
-table.indextable {
-    width: 100%;
-}
-
-table.indextable td {
-    text-align: left;
-    vertical-align: top;
-}
-
-table.indextable ul {
-    margin-top: 0;
-    margin-bottom: 0;
-    list-style-type: none;
-}
-
-table.indextable > tbody > tr > td > ul {
-    padding-left: 0em;
-}
-
-table.indextable tr.pcap {
-    height: 10px;
-}
-
-table.indextable tr.cap {
-    margin-top: 10px;
-    background-color: #f2f2f2;
-}
-
-img.toggler {
-    margin-right: 3px;
-    margin-top: 3px;
-    cursor: pointer;
-}
-
-div.modindex-jumpbox {
-    border-top: 1px solid #ddd;
-    border-bottom: 1px solid #ddd;
-    margin: 1em 0 1em 0;
-    padding: 0.4em;
-}
-
-div.genindex-jumpbox {
-    border-top: 1px solid #ddd;
-    border-bottom: 1px solid #ddd;
-    margin: 1em 0 1em 0;
-    padding: 0.4em;
-}
-
-/* -- domain module index --------------------------------------------------- */
-
-table.modindextable td {
-    padding: 2px;
-    border-collapse: collapse;
-}
-
-/* -- general body styles --------------------------------------------------- */
-
-div.body {
-    min-width: 360px;
-    max-width: 800px;
-}
-
-div.body p, div.body dd, div.body li, div.body blockquote {
-    -moz-hyphens: auto;
-    -ms-hyphens: auto;
-    -webkit-hyphens: auto;
-    hyphens: auto;
-}
-
-a.headerlink {
-    visibility: hidden;
-}
-
-a:visited {
-    color: #551A8B;
-}
-
-h1:hover > a.headerlink,
-h2:hover > a.headerlink,
-h3:hover > a.headerlink,
-h4:hover > a.headerlink,
-h5:hover > a.headerlink,
-h6:hover > a.headerlink,
-dt:hover > a.headerlink,
-caption:hover > a.headerlink,
-p.caption:hover > a.headerlink,
-div.code-block-caption:hover > a.headerlink {
-    visibility: visible;
-}
-
-div.body p.caption {
-    text-align: inherit;
-}
-
-div.body td {
-    text-align: left;
-}
-
-.first {
-    margin-top: 0 !important;
-}
-
-p.rubric {
-    margin-top: 30px;
-    font-weight: bold;
-}
-
-img.align-left, figure.align-left, .figure.align-left, object.align-left {
-    clear: left;
-    float: left;
-    margin-right: 1em;
-}
-
-img.align-right, figure.align-right, .figure.align-right, object.align-right {
-    clear: right;
-    float: right;
-    margin-left: 1em;
-}
-
-img.align-center, figure.align-center, .figure.align-center, object.align-center {
-  display: block;
-  margin-left: auto;
-  margin-right: auto;
-}
-
-img.align-default, figure.align-default, .figure.align-default {
-  display: block;
-  margin-left: auto;
-  margin-right: auto;
-}
-
-.align-left {
-    text-align: left;
-}
-
-.align-center {
-    text-align: center;
-}
-
-.align-default {
-    text-align: center;
-}
-
-.align-right {
-    text-align: right;
-}
-
-/* -- sidebars -------------------------------------------------------------- */
-
-div.sidebar,
-aside.sidebar {
-    margin: 0 0 0.5em 1em;
-    border: 1px solid #ddb;
-    padding: 7px;
-    background-color: #ffe;
-    width: 40%;
-    float: right;
-    clear: right;
-    overflow-x: auto;
-}
-
-p.sidebar-title {
-    font-weight: bold;
-}
-
-nav.contents,
-aside.topic,
-div.admonition, div.topic, blockquote {
-    clear: left;
-}
-
-/* -- topics ---------------------------------------------------------------- */
-
-nav.contents,
-aside.topic,
-div.topic {
-    border: 1px solid #ccc;
-    padding: 7px;
-    margin: 10px 0 10px 0;
-}
-
-p.topic-title {
-    font-size: 1.1em;
-    font-weight: bold;
-    margin-top: 10px;
-}
-
-/* -- admonitions ----------------------------------------------------------- */
-
-div.admonition {
-    margin-top: 10px;
-    margin-bottom: 10px;
-    padding: 7px;
-}
-
-div.admonition dt {
-    font-weight: bold;
-}
-
-p.admonition-title {
-    margin: 0px 10px 5px 0px;
-    font-weight: bold;
-}
-
-div.body p.centered {
-    text-align: center;
-    margin-top: 25px;
-}
-
-/* -- content of sidebars/topics/admonitions -------------------------------- */
-
-div.sidebar > :last-child,
-aside.sidebar > :last-child,
-nav.contents > :last-child,
-aside.topic > :last-child,
-div.topic > :last-child,
-div.admonition > :last-child {
-    margin-bottom: 0;
-}
-
-div.sidebar::after,
-aside.sidebar::after,
-nav.contents::after,
-aside.topic::after,
-div.topic::after,
-div.admonition::after,
-blockquote::after {
-    display: block;
-    content: '';
-    clear: both;
-}
-
-/* -- tables ---------------------------------------------------------------- */
-
-table.docutils {
-    margin-top: 10px;
-    margin-bottom: 10px;
-    border: 0;
-    border-collapse: collapse;
-}
-
-table.align-center {
-    margin-left: auto;
-    margin-right: auto;
-}
-
-table.align-default {
-    margin-left: auto;
-    margin-right: auto;
-}
-
-table caption span.caption-number {
-    font-style: italic;
-}
-
-table caption span.caption-text {
-}
-
-table.docutils td, table.docutils th {
-    padding: 1px 8px 1px 5px;
-    border-top: 0;
-    border-left: 0;
-    border-right: 0;
-    border-bottom: 1px solid #aaa;
-}
-
-th {
-    text-align: left;
-    padding-right: 5px;
-}
-
-table.citation {
-    border-left: solid 1px gray;
-    margin-left: 1px;
-}
-
-table.citation td {
-    border-bottom: none;
-}
-
-th > :first-child,
-td > :first-child {
-    margin-top: 0px;
-}
-
-th > :last-child,
-td > :last-child {
-    margin-bottom: 0px;
-}
-
-/* -- figures --------------------------------------------------------------- */
-
-div.figure, figure {
-    margin: 0.5em;
-    padding: 0.5em;
-}
-
-div.figure p.caption, figcaption {
-    padding: 0.3em;
-}
-
-div.figure p.caption span.caption-number,
-figcaption span.caption-number {
-    font-style: italic;
-}
-
-div.figure p.caption span.caption-text,
-figcaption span.caption-text {
-}
-
-/* -- field list styles ----------------------------------------------------- */
-
-table.field-list td, table.field-list th {
-    border: 0 !important;
-}
-
-.field-list ul {
-    margin: 0;
-    padding-left: 1em;
-}
-
-.field-list p {
-    margin: 0;
-}
-
-.field-name {
-    -moz-hyphens: manual;
-    -ms-hyphens: manual;
-    -webkit-hyphens: manual;
-    hyphens: manual;
-}
-
-/* -- hlist styles ---------------------------------------------------------- */
-
-table.hlist {
-    margin: 1em 0;
-}
-
-table.hlist td {
-    vertical-align: top;
-}
-
-/* -- object description styles --------------------------------------------- */
-
-.sig {
-	font-family: 'Consolas', 'Menlo', 'DejaVu Sans Mono', 'Bitstream Vera Sans Mono', monospace;
-}
-
-.sig-name, code.descname {
-    background-color: transparent;
-    font-weight: bold;
-}
-
-.sig-name {
-	font-size: 1.1em;
-}
-
-code.descname {
-    font-size: 1.2em;
-}
-
-.sig-prename, code.descclassname {
-    background-color: transparent;
-}
-
-.optional {
-    font-size: 1.3em;
-}
-
-.sig-paren {
-    font-size: larger;
-}
-
-.sig-param.n {
-	font-style: italic;
-}
-
-/* C++ specific styling */
-
-.sig-inline.c-texpr,
-.sig-inline.cpp-texpr {
-	font-family: unset;
-}
-
-.sig.c   .k, .sig.c   .kt,
-.sig.cpp .k, .sig.cpp .kt {
-	color: #0033B3;
-}
-
-.sig.c   .m,
-.sig.cpp .m {
-	color: #1750EB;
-}
-
-.sig.c   .s, .sig.c   .sc,
-.sig.cpp .s, .sig.cpp .sc {
-	color: #067D17;
-}
-
-
-/* -- other body styles ----------------------------------------------------- */
-
-ol.arabic {
-    list-style: decimal;
-}
-
-ol.loweralpha {
-    list-style: lower-alpha;
-}
-
-ol.upperalpha {
-    list-style: upper-alpha;
-}
-
-ol.lowerroman {
-    list-style: lower-roman;
-}
-
-ol.upperroman {
-    list-style: upper-roman;
-}
-
-:not(li) > ol > li:first-child > :first-child,
-:not(li) > ul > li:first-child > :first-child {
-    margin-top: 0px;
-}
-
-:not(li) > ol > li:last-child > :last-child,
-:not(li) > ul > li:last-child > :last-child {
-    margin-bottom: 0px;
-}
-
-ol.simple ol p,
-ol.simple ul p,
-ul.simple ol p,
-ul.simple ul p {
-    margin-top: 0;
-}
-
-ol.simple > li:not(:first-child) > p,
-ul.simple > li:not(:first-child) > p {
-    margin-top: 0;
-}
-
-ol.simple p,
-ul.simple p {
-    margin-bottom: 0;
-}
-
-aside.footnote > span,
-div.citation > span {
-    float: left;
-}
-aside.footnote > span:last-of-type,
-div.citation > span:last-of-type {
-  padding-right: 0.5em;
-}
-aside.footnote > p {
-  margin-left: 2em;
-}
-div.citation > p {
-  margin-left: 4em;
-}
-aside.footnote > p:last-of-type,
-div.citation > p:last-of-type {
-    margin-bottom: 0em;
-}
-aside.footnote > p:last-of-type:after,
-div.citation > p:last-of-type:after {
-    content: "";
-    clear: both;
-}
-
-dl.field-list {
-    display: grid;
-    grid-template-columns: fit-content(30%) auto;
-}
-
-dl.field-list > dt {
-    font-weight: bold;
-    word-break: break-word;
-    padding-left: 0.5em;
-    padding-right: 5px;
-}
-
-dl.field-list > dd {
-    padding-left: 0.5em;
-    margin-top: 0em;
-    margin-left: 0em;
-    margin-bottom: 0em;
-}
-
-dl {
-    margin-bottom: 15px;
-}
-
-dd > :first-child {
-    margin-top: 0px;
-}
-
-dd ul, dd table {
-    margin-bottom: 10px;
-}
-
-dd {
-    margin-top: 3px;
-    margin-bottom: 10px;
-    margin-left: 30px;
-}
-
-.sig dd {
-    margin-top: 0px;
-    margin-bottom: 0px;
-}
-
-.sig dl {
-    margin-top: 0px;
-    margin-bottom: 0px;
-}
-
-dl > dd:last-child,
-dl > dd:last-child > :last-child {
-    margin-bottom: 0;
-}
-
-dt:target, span.highlighted {
-    background-color: #fbe54e;
-}
-
-rect.highlighted {
-    fill: #fbe54e;
-}
-
-dl.glossary dt {
-    font-weight: bold;
-    font-size: 1.1em;
-}
-
-.versionmodified {
-    font-style: italic;
-}
-
-.system-message {
-    background-color: #fda;
-    padding: 5px;
-    border: 3px solid red;
-}
-
-.footnote:target  {
-    background-color: #ffa;
-}
-
-.line-block {
-    display: block;
-    margin-top: 1em;
-    margin-bottom: 1em;
-}
-
-.line-block .line-block {
-    margin-top: 0;
-    margin-bottom: 0;
-    margin-left: 1.5em;
-}
-
-.guilabel, .menuselection {
-    font-family: sans-serif;
-}
-
-.accelerator {
-    text-decoration: underline;
-}
-
-.classifier {
-    font-style: oblique;
-}
-
-.classifier:before {
-    font-style: normal;
-    margin: 0 0.5em;
-    content: ":";
-    display: inline-block;
-}
-
-abbr, acronym {
-    border-bottom: dotted 1px;
-    cursor: help;
-}
-
-.translated {
-    background-color: rgba(207, 255, 207, 0.2)
-}
-
-.untranslated {
-    background-color: rgba(255, 207, 207, 0.2)
-}
-
-/* -- code displays --------------------------------------------------------- */
-
-pre {
-    overflow: auto;
-    overflow-y: hidden;  /* fixes display issues on Chrome browsers */
-}
-
-pre, div[class*="highlight-"] {
-    clear: both;
-}
-
-span.pre {
-    -moz-hyphens: none;
-    -ms-hyphens: none;
-    -webkit-hyphens: none;
-    hyphens: none;
-    white-space: nowrap;
-}
-
-div[class*="highlight-"] {
-    margin: 1em 0;
-}
-
-td.linenos pre {
-    border: 0;
-    background-color: transparent;
-    color: #aaa;
-}
-
-table.highlighttable {
-    display: block;
-}
-
-table.highlighttable tbody {
-    display: block;
-}
-
-table.highlighttable tr {
-    display: flex;
-}
-
-table.highlighttable td {
-    margin: 0;
-    padding: 0;
-}
-
-table.highlighttable td.linenos {
-    padding-right: 0.5em;
-}
-
-table.highlighttable td.code {
-    flex: 1;
-    overflow: hidden;
-}
-
-.highlight .hll {
-    display: block;
-}
-
-div.highlight pre,
-table.highlighttable pre {
-    margin: 0;
-}
-
-div.code-block-caption + div {
-    margin-top: 0;
-}
-
-div.code-block-caption {
-    margin-top: 1em;
-    padding: 2px 5px;
-    font-size: small;
-}
-
-div.code-block-caption code {
-    background-color: transparent;
-}
-
-table.highlighttable td.linenos,
-span.linenos,
-div.highlight span.gp {  /* gp: Generic.Prompt */
-  user-select: none;
-  -webkit-user-select: text; /* Safari fallback only */
-  -webkit-user-select: none; /* Chrome/Safari */
-  -moz-user-select: none; /* Firefox */
-  -ms-user-select: none; /* IE10+ */
-}
-
-div.code-block-caption span.caption-number {
-    padding: 0.1em 0.3em;
-    font-style: italic;
-}
-
-div.code-block-caption span.caption-text {
-}
-
-div.literal-block-wrapper {
-    margin: 1em 0;
-}
-
-code.xref, a code {
-    background-color: transparent;
-    font-weight: bold;
-}
-
-h1 code, h2 code, h3 code, h4 code, h5 code, h6 code {
-    background-color: transparent;
-}
-
-.viewcode-link {
-    float: right;
-}
-
-.viewcode-back {
-    float: right;
-    font-family: sans-serif;
-}
-
-div.viewcode-block:target {
-    margin: -1px -10px;
-    padding: 0 10px;
-}
-
-/* -- math display ---------------------------------------------------------- */
-
-img.math {
-    vertical-align: middle;
-}
-
-div.body div.math p {
-    text-align: center;
-}
-
-span.eqno {
-    float: right;
-}
-
-span.eqno a.headerlink {
-    position: absolute;
-    z-index: 1;
-}
-
-div.math:hover a.headerlink {
-    visibility: visible;
-}
-
-/* -- printout stylesheet --------------------------------------------------- */
-
-@media print {
-    div.document,
-    div.documentwrapper,
-    div.bodywrapper {
-        margin: 0 !important;
-        width: 100%;
-    }
-
-    div.sphinxsidebar,
-    div.related,
-    div.footer,
-    #top-link {
-        display: none;
-    }
-}
\ No newline at end of file
diff --git a/docs/_build/html/_static/check-solid.svg b/docs/_build/html/_static/check-solid.svg
deleted file mode 100644
index 92fad4b5c..000000000
--- a/docs/_build/html/_static/check-solid.svg
+++ /dev/null
@@ -1,4 +0,0 @@
-<svg xmlns="http://www.w3.org/2000/svg" class="icon icon-tabler icon-tabler-check" width="44" height="44" viewBox="0 0 24 24" stroke-width="2" stroke="#22863a" fill="none" stroke-linecap="round" stroke-linejoin="round">
-  <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-  <path d="M5 12l5 5l10 -10" />
-</svg>
diff --git a/docs/_build/html/_static/clipboard.min.js b/docs/_build/html/_static/clipboard.min.js
deleted file mode 100644
index 54b3c4638..000000000
--- a/docs/_build/html/_static/clipboard.min.js
+++ /dev/null
@@ -1,7 +0,0 @@
-/*!
- * clipboard.js v2.0.8
- * https://clipboardjs.com/
- *
- * Licensed MIT © Zeno Rocha
- */
-!function(t,e){"object"==typeof exports&&"object"==typeof module?module.exports=e():"function"==typeof define&&define.amd?define([],e):"object"==typeof exports?exports.ClipboardJS=e():t.ClipboardJS=e()}(this,function(){return n={686:function(t,e,n){"use strict";n.d(e,{default:function(){return o}});var e=n(279),i=n.n(e),e=n(370),u=n.n(e),e=n(817),c=n.n(e);function a(t){try{return document.execCommand(t)}catch(t){return}}var f=function(t){t=c()(t);return a("cut"),t};var l=function(t){var e,n,o,r=1<arguments.length&&void 0!==arguments[1]?arguments[1]:{container:document.body},i="";return"string"==typeof t?(e=t,n="rtl"===document.documentElement.getAttribute("dir"),(o=document.createElement("textarea")).style.fontSize="12pt",o.style.border="0",o.style.padding="0",o.style.margin="0",o.style.position="absolute",o.style[n?"right":"left"]="-9999px",n=window.pageYOffset||document.documentElement.scrollTop,o.style.top="".concat(n,"px"),o.setAttribute("readonly",""),o.value=e,o=o,r.container.appendChild(o),i=c()(o),a("copy"),o.remove()):(i=c()(t),a("copy")),i};function r(t){return(r="function"==typeof Symbol&&"symbol"==typeof Symbol.iterator?function(t){return typeof t}:function(t){return t&&"function"==typeof Symbol&&t.constructor===Symbol&&t!==Symbol.prototype?"symbol":typeof t})(t)}var s=function(){var t=0<arguments.length&&void 0!==arguments[0]?arguments[0]:{},e=t.action,n=void 0===e?"copy":e,o=t.container,e=t.target,t=t.text;if("copy"!==n&&"cut"!==n)throw new Error('Invalid "action" value, use either "copy" or "cut"');if(void 0!==e){if(!e||"object"!==r(e)||1!==e.nodeType)throw new Error('Invalid "target" value, use a valid Element');if("copy"===n&&e.hasAttribute("disabled"))throw new Error('Invalid "target" attribute. Please use "readonly" instead of "disabled" attribute');if("cut"===n&&(e.hasAttribute("readonly")||e.hasAttribute("disabled")))throw new Error('Invalid "target" attribute. You can\'t cut text from elements with "readonly" or "disabled" attributes')}return t?l(t,{container:o}):e?"cut"===n?f(e):l(e,{container:o}):void 0};function d(t){return(d="function"==typeof Symbol&&"symbol"==typeof Symbol.iterator?function(t){return typeof t}:function(t){return t&&"function"==typeof Symbol&&t.constructor===Symbol&&t!==Symbol.prototype?"symbol":typeof t})(t)}function p(t,e){for(var n=0;n<e.length;n++){var o=e[n];o.enumerable=o.enumerable||!1,o.configurable=!0,"value"in o&&(o.writable=!0),Object.defineProperty(t,o.key,o)}}function y(t,e){return(y=Object.setPrototypeOf||function(t,e){return t.__proto__=e,t})(t,e)}function h(n){var o=function(){if("undefined"==typeof Reflect||!Reflect.construct)return!1;if(Reflect.construct.sham)return!1;if("function"==typeof Proxy)return!0;try{return Date.prototype.toString.call(Reflect.construct(Date,[],function(){})),!0}catch(t){return!1}}();return function(){var t,e=m(n);return t=o?(t=m(this).constructor,Reflect.construct(e,arguments,t)):e.apply(this,arguments),e=this,!(t=t)||"object"!==d(t)&&"function"!=typeof t?function(t){if(void 0!==t)return t;throw new ReferenceError("this hasn't been initialised - super() hasn't been called")}(e):t}}function m(t){return(m=Object.setPrototypeOf?Object.getPrototypeOf:function(t){return t.__proto__||Object.getPrototypeOf(t)})(t)}function v(t,e){t="data-clipboard-".concat(t);if(e.hasAttribute(t))return e.getAttribute(t)}var o=function(){!function(t,e){if("function"!=typeof e&&null!==e)throw new TypeError("Super expression must either be null or a function");t.prototype=Object.create(e&&e.prototype,{constructor:{value:t,writable:!0,configurable:!0}}),e&&y(t,e)}(r,i());var t,e,n,o=h(r);function r(t,e){var n;return function(t){if(!(t instanceof r))throw new TypeError("Cannot call a class as a function")}(this),(n=o.call(this)).resolveOptions(e),n.listenClick(t),n}return t=r,n=[{key:"copy",value:function(t){var e=1<arguments.length&&void 0!==arguments[1]?arguments[1]:{container:document.body};return l(t,e)}},{key:"cut",value:function(t){return f(t)}},{key:"isSupported",value:function(){var t=0<arguments.length&&void 0!==arguments[0]?arguments[0]:["copy","cut"],t="string"==typeof t?[t]:t,e=!!document.queryCommandSupported;return t.forEach(function(t){e=e&&!!document.queryCommandSupported(t)}),e}}],(e=[{key:"resolveOptions",value:function(){var t=0<arguments.length&&void 0!==arguments[0]?arguments[0]:{};this.action="function"==typeof t.action?t.action:this.defaultAction,this.target="function"==typeof t.target?t.target:this.defaultTarget,this.text="function"==typeof t.text?t.text:this.defaultText,this.container="object"===d(t.container)?t.container:document.body}},{key:"listenClick",value:function(t){var e=this;this.listener=u()(t,"click",function(t){return e.onClick(t)})}},{key:"onClick",value:function(t){var e=t.delegateTarget||t.currentTarget,t=s({action:this.action(e),container:this.container,target:this.target(e),text:this.text(e)});this.emit(t?"success":"error",{action:this.action,text:t,trigger:e,clearSelection:function(){e&&e.focus(),document.activeElement.blur(),window.getSelection().removeAllRanges()}})}},{key:"defaultAction",value:function(t){return v("action",t)}},{key:"defaultTarget",value:function(t){t=v("target",t);if(t)return document.querySelector(t)}},{key:"defaultText",value:function(t){return v("text",t)}},{key:"destroy",value:function(){this.listener.destroy()}}])&&p(t.prototype,e),n&&p(t,n),r}()},828:function(t){var e;"undefined"==typeof Element||Element.prototype.matches||((e=Element.prototype).matches=e.matchesSelector||e.mozMatchesSelector||e.msMatchesSelector||e.oMatchesSelector||e.webkitMatchesSelector),t.exports=function(t,e){for(;t&&9!==t.nodeType;){if("function"==typeof t.matches&&t.matches(e))return t;t=t.parentNode}}},438:function(t,e,n){var u=n(828);function i(t,e,n,o,r){var i=function(e,n,t,o){return function(t){t.delegateTarget=u(t.target,n),t.delegateTarget&&o.call(e,t)}}.apply(this,arguments);return t.addEventListener(n,i,r),{destroy:function(){t.removeEventListener(n,i,r)}}}t.exports=function(t,e,n,o,r){return"function"==typeof t.addEventListener?i.apply(null,arguments):"function"==typeof n?i.bind(null,document).apply(null,arguments):("string"==typeof t&&(t=document.querySelectorAll(t)),Array.prototype.map.call(t,function(t){return i(t,e,n,o,r)}))}},879:function(t,n){n.node=function(t){return void 0!==t&&t instanceof HTMLElement&&1===t.nodeType},n.nodeList=function(t){var e=Object.prototype.toString.call(t);return void 0!==t&&("[object NodeList]"===e||"[object HTMLCollection]"===e)&&"length"in t&&(0===t.length||n.node(t[0]))},n.string=function(t){return"string"==typeof t||t instanceof String},n.fn=function(t){return"[object Function]"===Object.prototype.toString.call(t)}},370:function(t,e,n){var f=n(879),l=n(438);t.exports=function(t,e,n){if(!t&&!e&&!n)throw new Error("Missing required arguments");if(!f.string(e))throw new TypeError("Second argument must be a String");if(!f.fn(n))throw new TypeError("Third argument must be a Function");if(f.node(t))return c=e,a=n,(u=t).addEventListener(c,a),{destroy:function(){u.removeEventListener(c,a)}};if(f.nodeList(t))return o=t,r=e,i=n,Array.prototype.forEach.call(o,function(t){t.addEventListener(r,i)}),{destroy:function(){Array.prototype.forEach.call(o,function(t){t.removeEventListener(r,i)})}};if(f.string(t))return t=t,e=e,n=n,l(document.body,t,e,n);throw new TypeError("First argument must be a String, HTMLElement, HTMLCollection, or NodeList");var o,r,i,u,c,a}},817:function(t){t.exports=function(t){var e,n="SELECT"===t.nodeName?(t.focus(),t.value):"INPUT"===t.nodeName||"TEXTAREA"===t.nodeName?((e=t.hasAttribute("readonly"))||t.setAttribute("readonly",""),t.select(),t.setSelectionRange(0,t.value.length),e||t.removeAttribute("readonly"),t.value):(t.hasAttribute("contenteditable")&&t.focus(),n=window.getSelection(),(e=document.createRange()).selectNodeContents(t),n.removeAllRanges(),n.addRange(e),n.toString());return n}},279:function(t){function e(){}e.prototype={on:function(t,e,n){var o=this.e||(this.e={});return(o[t]||(o[t]=[])).push({fn:e,ctx:n}),this},once:function(t,e,n){var o=this;function r(){o.off(t,r),e.apply(n,arguments)}return r._=e,this.on(t,r,n)},emit:function(t){for(var e=[].slice.call(arguments,1),n=((this.e||(this.e={}))[t]||[]).slice(),o=0,r=n.length;o<r;o++)n[o].fn.apply(n[o].ctx,e);return this},off:function(t,e){var n=this.e||(this.e={}),o=n[t],r=[];if(o&&e)for(var i=0,u=o.length;i<u;i++)o[i].fn!==e&&o[i].fn._!==e&&r.push(o[i]);return r.length?n[t]=r:delete n[t],this}},t.exports=e,t.exports.TinyEmitter=e}},r={},o.n=function(t){var e=t&&t.__esModule?function(){return t.default}:function(){return t};return o.d(e,{a:e}),e},o.d=function(t,e){for(var n in e)o.o(e,n)&&!o.o(t,n)&&Object.defineProperty(t,n,{enumerable:!0,get:e[n]})},o.o=function(t,e){return Object.prototype.hasOwnProperty.call(t,e)},o(686).default;function o(t){if(r[t])return r[t].exports;var e=r[t]={exports:{}};return n[t](e,e.exports,o),e.exports}var n,r});
\ No newline at end of file
diff --git a/docs/_build/html/_static/copy-button.svg b/docs/_build/html/_static/copy-button.svg
deleted file mode 100644
index 9c074dae5..000000000
--- a/docs/_build/html/_static/copy-button.svg
+++ /dev/null
@@ -1,5 +0,0 @@
-<svg xmlns="http://www.w3.org/2000/svg" class="icon icon-tabler icon-tabler-copy" width="44" height="44" viewBox="0 0 24 24" stroke-width="1.5" stroke="#000000" fill="none" stroke-linecap="round" stroke-linejoin="round">
-  <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-  <rect x="8" y="8" width="12" height="12" rx="2" />
-  <path d="M16 8v-2a2 2 0 0 0 -2 -2h-8a2 2 0 0 0 -2 2v8a2 2 0 0 0 2 2h2" />
-</svg>
diff --git a/docs/_build/html/_static/copybutton.css b/docs/_build/html/_static/copybutton.css
deleted file mode 100644
index f1916ec7d..000000000
--- a/docs/_build/html/_static/copybutton.css
+++ /dev/null
@@ -1,94 +0,0 @@
-/* Copy buttons */
-button.copybtn {
-    position: absolute;
-    display: flex;
-    top: .3em;
-    right: .3em;
-    width: 1.7em;
-    height: 1.7em;
-	opacity: 0;
-    transition: opacity 0.3s, border .3s, background-color .3s;
-    user-select: none;
-    padding: 0;
-    border: none;
-    outline: none;
-    border-radius: 0.4em;
-    /* The colors that GitHub uses */
-    border: #1b1f2426 1px solid;
-    background-color: #f6f8fa;
-    color: #57606a;
-}
-
-button.copybtn.success {
-    border-color: #22863a;
-    color: #22863a;
-}
-
-button.copybtn svg {
-    stroke: currentColor;
-    width: 1.5em;
-    height: 1.5em;
-    padding: 0.1em;
-}
-
-div.highlight  {
-    position: relative;
-}
-
-/* Show the copybutton */
-.highlight:hover button.copybtn, button.copybtn.success {
-	opacity: 1;
-}
-
-.highlight button.copybtn:hover {
-    background-color: rgb(235, 235, 235);
-}
-
-.highlight button.copybtn:active {
-    background-color: rgb(187, 187, 187);
-}
-
-/**
- * A minimal CSS-only tooltip copied from:
- *   https://codepen.io/mildrenben/pen/rVBrpK
- *
- * To use, write HTML like the following:
- *
- * <p class="o-tooltip--left" data-tooltip="Hey">Short</p>
- */
- .o-tooltip--left {
-  position: relative;
- }
-
- .o-tooltip--left:after {
-    opacity: 0;
-    visibility: hidden;
-    position: absolute;
-    content: attr(data-tooltip);
-    padding: .2em;
-    font-size: .8em;
-    left: -.2em;
-    background: grey;
-    color: white;
-    white-space: nowrap;
-    z-index: 2;
-    border-radius: 2px;
-    transform: translateX(-102%) translateY(0);
-    transition: opacity 0.2s cubic-bezier(0.64, 0.09, 0.08, 1), transform 0.2s cubic-bezier(0.64, 0.09, 0.08, 1);
-}
-
-.o-tooltip--left:hover:after {
-    display: block;
-    opacity: 1;
-    visibility: visible;
-    transform: translateX(-100%) translateY(0);
-    transition: opacity 0.2s cubic-bezier(0.64, 0.09, 0.08, 1), transform 0.2s cubic-bezier(0.64, 0.09, 0.08, 1);
-    transition-delay: .5s;
-}
-
-/* By default the copy button shouldn't show up when printing a page */
-@media print {
-    button.copybtn {
-        display: none;
-    }
-}
diff --git a/docs/_build/html/_static/copybutton.js b/docs/_build/html/_static/copybutton.js
deleted file mode 100644
index ff4aa329a..000000000
--- a/docs/_build/html/_static/copybutton.js
+++ /dev/null
@@ -1,248 +0,0 @@
-// Localization support
-const messages = {
-  'en': {
-    'copy': 'Copy',
-    'copy_to_clipboard': 'Copy to clipboard',
-    'copy_success': 'Copied!',
-    'copy_failure': 'Failed to copy',
-  },
-  'es' : {
-    'copy': 'Copiar',
-    'copy_to_clipboard': 'Copiar al portapapeles',
-    'copy_success': '¡Copiado!',
-    'copy_failure': 'Error al copiar',
-  },
-  'de' : {
-    'copy': 'Kopieren',
-    'copy_to_clipboard': 'In die Zwischenablage kopieren',
-    'copy_success': 'Kopiert!',
-    'copy_failure': 'Fehler beim Kopieren',
-  },
-  'fr' : {
-    'copy': 'Copier',
-    'copy_to_clipboard': 'Copier dans le presse-papier',
-    'copy_success': 'Copié !',
-    'copy_failure': 'Échec de la copie',
-  },
-  'ru': {
-    'copy': 'Скопировать',
-    'copy_to_clipboard': 'Скопировать в буфер',
-    'copy_success': 'Скопировано!',
-    'copy_failure': 'Не удалось скопировать',
-  },
-  'zh-CN': {
-    'copy': '复制',
-    'copy_to_clipboard': '复制到剪贴板',
-    'copy_success': '复制成功!',
-    'copy_failure': '复制失败',
-  },
-  'it' : {
-    'copy': 'Copiare',
-    'copy_to_clipboard': 'Copiato negli appunti',
-    'copy_success': 'Copiato!',
-    'copy_failure': 'Errore durante la copia',
-  }
-}
-
-let locale = 'en'
-if( document.documentElement.lang !== undefined
-    && messages[document.documentElement.lang] !== undefined ) {
-  locale = document.documentElement.lang
-}
-
-let doc_url_root = DOCUMENTATION_OPTIONS.URL_ROOT;
-if (doc_url_root == '#') {
-    doc_url_root = '';
-}
-
-/**
- * SVG files for our copy buttons
- */
-let iconCheck = `<svg xmlns="http://www.w3.org/2000/svg" class="icon icon-tabler icon-tabler-check" width="44" height="44" viewBox="0 0 24 24" stroke-width="2" stroke="#22863a" fill="none" stroke-linecap="round" stroke-linejoin="round">
-  <title>${messages[locale]['copy_success']}</title>
-  <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-  <path d="M5 12l5 5l10 -10" />
-</svg>`
-
-// If the user specified their own SVG use that, otherwise use the default
-let iconCopy = ``;
-if (!iconCopy) {
-  iconCopy = `<svg xmlns="http://www.w3.org/2000/svg" class="icon icon-tabler icon-tabler-copy" width="44" height="44" viewBox="0 0 24 24" stroke-width="1.5" stroke="#000000" fill="none" stroke-linecap="round" stroke-linejoin="round">
-  <title>${messages[locale]['copy_to_clipboard']}</title>
-  <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-  <rect x="8" y="8" width="12" height="12" rx="2" />
-  <path d="M16 8v-2a2 2 0 0 0 -2 -2h-8a2 2 0 0 0 -2 2v8a2 2 0 0 0 2 2h2" />
-</svg>`
-}
-
-/**
- * Set up copy/paste for code blocks
- */
-
-const runWhenDOMLoaded = cb => {
-  if (document.readyState != 'loading') {
-    cb()
-  } else if (document.addEventListener) {
-    document.addEventListener('DOMContentLoaded', cb)
-  } else {
-    document.attachEvent('onreadystatechange', function() {
-      if (document.readyState == 'complete') cb()
-    })
-  }
-}
-
-const codeCellId = index => `codecell${index}`
-
-// Clears selected text since ClipboardJS will select the text when copying
-const clearSelection = () => {
-  if (window.getSelection) {
-    window.getSelection().removeAllRanges()
-  } else if (document.selection) {
-    document.selection.empty()
-  }
-}
-
-// Changes tooltip text for a moment, then changes it back
-// We want the timeout of our `success` class to be a bit shorter than the
-// tooltip and icon change, so that we can hide the icon before changing back.
-var timeoutIcon = 2000;
-var timeoutSuccessClass = 1500;
-
-const temporarilyChangeTooltip = (el, oldText, newText) => {
-  el.setAttribute('data-tooltip', newText)
-  el.classList.add('success')
-  // Remove success a little bit sooner than we change the tooltip
-  // So that we can use CSS to hide the copybutton first
-  setTimeout(() => el.classList.remove('success'), timeoutSuccessClass)
-  setTimeout(() => el.setAttribute('data-tooltip', oldText), timeoutIcon)
-}
-
-// Changes the copy button icon for two seconds, then changes it back
-const temporarilyChangeIcon = (el) => {
-  el.innerHTML = iconCheck;
-  setTimeout(() => {el.innerHTML = iconCopy}, timeoutIcon)
-}
-
-const addCopyButtonToCodeCells = () => {
-  // If ClipboardJS hasn't loaded, wait a bit and try again. This
-  // happens because we load ClipboardJS asynchronously.
-  if (window.ClipboardJS === undefined) {
-    setTimeout(addCopyButtonToCodeCells, 250)
-    return
-  }
-
-  // Add copybuttons to all of our code cells
-  const COPYBUTTON_SELECTOR = 'div.highlight pre';
-  const codeCells = document.querySelectorAll(COPYBUTTON_SELECTOR)
-  codeCells.forEach((codeCell, index) => {
-    const id = codeCellId(index)
-    codeCell.setAttribute('id', id)
-
-    const clipboardButton = id =>
-    `<button class="copybtn o-tooltip--left" data-tooltip="${messages[locale]['copy']}" data-clipboard-target="#${id}">
-      ${iconCopy}
-    </button>`
-    codeCell.insertAdjacentHTML('afterend', clipboardButton(id))
-  })
-
-function escapeRegExp(string) {
-    return string.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'); // $& means the whole matched string
-}
-
-/**
- * Removes excluded text from a Node.
- *
- * @param {Node} target Node to filter.
- * @param {string} exclude CSS selector of nodes to exclude.
- * @returns {DOMString} Text from `target` with text removed.
- */
-function filterText(target, exclude) {
-    const clone = target.cloneNode(true);  // clone as to not modify the live DOM
-    if (exclude) {
-        // remove excluded nodes
-        clone.querySelectorAll(exclude).forEach(node => node.remove());
-    }
-    return clone.innerText;
-}
-
-// Callback when a copy button is clicked. Will be passed the node that was clicked
-// should then grab the text and replace pieces of text that shouldn't be used in output
-function formatCopyText(textContent, copybuttonPromptText, isRegexp = false, onlyCopyPromptLines = true, removePrompts = true, copyEmptyLines = true, lineContinuationChar = "", hereDocDelim = "") {
-    var regexp;
-    var match;
-
-    // Do we check for line continuation characters and "HERE-documents"?
-    var useLineCont = !!lineContinuationChar
-    var useHereDoc = !!hereDocDelim
-
-    // create regexp to capture prompt and remaining line
-    if (isRegexp) {
-        regexp = new RegExp('^(' + copybuttonPromptText + ')(.*)')
-    } else {
-        regexp = new RegExp('^(' + escapeRegExp(copybuttonPromptText) + ')(.*)')
-    }
-
-    const outputLines = [];
-    var promptFound = false;
-    var gotLineCont = false;
-    var gotHereDoc = false;
-    const lineGotPrompt = [];
-    for (const line of textContent.split('\n')) {
-        match = line.match(regexp)
-        if (match || gotLineCont || gotHereDoc) {
-            promptFound = regexp.test(line)
-            lineGotPrompt.push(promptFound)
-            if (removePrompts && promptFound) {
-                outputLines.push(match[2])
-            } else {
-                outputLines.push(line)
-            }
-            gotLineCont = line.endsWith(lineContinuationChar) & useLineCont
-            if (line.includes(hereDocDelim) & useHereDoc)
-                gotHereDoc = !gotHereDoc
-        } else if (!onlyCopyPromptLines) {
-            outputLines.push(line)
-        } else if (copyEmptyLines && line.trim() === '') {
-            outputLines.push(line)
-        }
-    }
-
-    // If no lines with the prompt were found then just use original lines
-    if (lineGotPrompt.some(v => v === true)) {
-        textContent = outputLines.join('\n');
-    }
-
-    // Remove a trailing newline to avoid auto-running when pasting
-    if (textContent.endsWith("\n")) {
-        textContent = textContent.slice(0, -1)
-    }
-    return textContent
-}
-
-
-var copyTargetText = (trigger) => {
-  var target = document.querySelector(trigger.attributes['data-clipboard-target'].value);
-
-  // get filtered text
-  let exclude = '.linenos';
-
-  let text = filterText(target, exclude);
-  return formatCopyText(text, '>>> |\\.\\.\\. |\\$ ', true, true, true, true, '', '')
-}
-
-  // Initialize with a callback so we can modify the text before copy
-  const clipboard = new ClipboardJS('.copybtn', {text: copyTargetText})
-
-  // Update UI with error/success messages
-  clipboard.on('success', event => {
-    clearSelection()
-    temporarilyChangeTooltip(event.trigger, messages[locale]['copy'], messages[locale]['copy_success'])
-    temporarilyChangeIcon(event.trigger)
-  })
-
-  clipboard.on('error', event => {
-    temporarilyChangeTooltip(event.trigger, messages[locale]['copy'], messages[locale]['copy_failure'])
-  })
-}
-
-runWhenDOMLoaded(addCopyButtonToCodeCells)
\ No newline at end of file
diff --git a/docs/_build/html/_static/copybutton_funcs.js b/docs/_build/html/_static/copybutton_funcs.js
deleted file mode 100644
index dbe1aaad7..000000000
--- a/docs/_build/html/_static/copybutton_funcs.js
+++ /dev/null
@@ -1,73 +0,0 @@
-function escapeRegExp(string) {
-    return string.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'); // $& means the whole matched string
-}
-
-/**
- * Removes excluded text from a Node.
- *
- * @param {Node} target Node to filter.
- * @param {string} exclude CSS selector of nodes to exclude.
- * @returns {DOMString} Text from `target` with text removed.
- */
-export function filterText(target, exclude) {
-    const clone = target.cloneNode(true);  // clone as to not modify the live DOM
-    if (exclude) {
-        // remove excluded nodes
-        clone.querySelectorAll(exclude).forEach(node => node.remove());
-    }
-    return clone.innerText;
-}
-
-// Callback when a copy button is clicked. Will be passed the node that was clicked
-// should then grab the text and replace pieces of text that shouldn't be used in output
-export function formatCopyText(textContent, copybuttonPromptText, isRegexp = false, onlyCopyPromptLines = true, removePrompts = true, copyEmptyLines = true, lineContinuationChar = "", hereDocDelim = "") {
-    var regexp;
-    var match;
-
-    // Do we check for line continuation characters and "HERE-documents"?
-    var useLineCont = !!lineContinuationChar
-    var useHereDoc = !!hereDocDelim
-
-    // create regexp to capture prompt and remaining line
-    if (isRegexp) {
-        regexp = new RegExp('^(' + copybuttonPromptText + ')(.*)')
-    } else {
-        regexp = new RegExp('^(' + escapeRegExp(copybuttonPromptText) + ')(.*)')
-    }
-
-    const outputLines = [];
-    var promptFound = false;
-    var gotLineCont = false;
-    var gotHereDoc = false;
-    const lineGotPrompt = [];
-    for (const line of textContent.split('\n')) {
-        match = line.match(regexp)
-        if (match || gotLineCont || gotHereDoc) {
-            promptFound = regexp.test(line)
-            lineGotPrompt.push(promptFound)
-            if (removePrompts && promptFound) {
-                outputLines.push(match[2])
-            } else {
-                outputLines.push(line)
-            }
-            gotLineCont = line.endsWith(lineContinuationChar) & useLineCont
-            if (line.includes(hereDocDelim) & useHereDoc)
-                gotHereDoc = !gotHereDoc
-        } else if (!onlyCopyPromptLines) {
-            outputLines.push(line)
-        } else if (copyEmptyLines && line.trim() === '') {
-            outputLines.push(line)
-        }
-    }
-
-    // If no lines with the prompt were found then just use original lines
-    if (lineGotPrompt.some(v => v === true)) {
-        textContent = outputLines.join('\n');
-    }
-
-    // Remove a trailing newline to avoid auto-running when pasting
-    if (textContent.endsWith("\n")) {
-        textContent = textContent.slice(0, -1)
-    }
-    return textContent
-}
diff --git a/docs/_build/html/_static/custom.css b/docs/_build/html/_static/custom.css
deleted file mode 100644
index 52a770bbb..000000000
--- a/docs/_build/html/_static/custom.css
+++ /dev/null
@@ -1,37 +0,0 @@
-/* hides the TOC caption from the main page since we already have an H1 heading */
-.section .caption-text {
-    display: none;
-}
-
-/* note title text in white */
-.admonition p.admonition-title {
-    color: white;
-    font-weight: 700;
-}
-
-/* .admonition.note {
-    background: #e5f4d8;
-} */
-
-/* left align tables */
-table.docutils {
-    margin-left: 1em;
-    font-family: monospace;
-}
-
-/* inline code snippets #E74C3C, var(--color-link), #4e9a06 */
-code.literal {
-    color: #4e9a06;
-    font-family: monospace;
-    font-size: var(--font-size-normal);
-}
-
-.sidebar-brand-text {
-    text-align: center;
-    margin: 0 0;
-}
-
-
-
-/* normalize size across Firefox / Chrome */
-html { font-size: 100%; }
diff --git a/docs/_build/html/_static/debug.css b/docs/_build/html/_static/debug.css
deleted file mode 100644
index 74d4aec33..000000000
--- a/docs/_build/html/_static/debug.css
+++ /dev/null
@@ -1,69 +0,0 @@
-/*
-  This CSS file should be overridden by the theme authors. It's
-  meant for debugging and developing the skeleton that this theme provides.
-*/
-body {
-  font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif,
-    "Apple Color Emoji", "Segoe UI Emoji";
-  background: lavender;
-}
-.sb-announcement {
-  background: rgb(131, 131, 131);
-}
-.sb-announcement__inner {
-  background: black;
-  color: white;
-}
-.sb-header {
-  background: lightskyblue;
-}
-.sb-header__inner {
-  background: royalblue;
-  color: white;
-}
-.sb-header-secondary {
-  background: lightcyan;
-}
-.sb-header-secondary__inner {
-  background: cornflowerblue;
-  color: white;
-}
-.sb-sidebar-primary {
-  background: lightgreen;
-}
-.sb-main {
-  background: blanchedalmond;
-}
-.sb-main__inner {
-  background: antiquewhite;
-}
-.sb-header-article {
-  background: lightsteelblue;
-}
-.sb-article-container {
-  background: snow;
-}
-.sb-article-main {
-  background: white;
-}
-.sb-footer-article {
-  background: lightpink;
-}
-.sb-sidebar-secondary {
-  background: lightgoldenrodyellow;
-}
-.sb-footer-content {
-  background: plum;
-}
-.sb-footer-content__inner {
-  background: palevioletred;
-}
-.sb-footer {
-  background: pink;
-}
-.sb-footer__inner {
-  background: salmon;
-}
-.sb-article {
-  background: white;
-}
diff --git a/docs/_build/html/_static/doctools.js b/docs/_build/html/_static/doctools.js
deleted file mode 100644
index d06a71d75..000000000
--- a/docs/_build/html/_static/doctools.js
+++ /dev/null
@@ -1,156 +0,0 @@
-/*
- * doctools.js
- * ~~~~~~~~~~~
- *
- * Base JavaScript utilities for all Sphinx HTML documentation.
- *
- * :copyright: Copyright 2007-2023 by the Sphinx team, see AUTHORS.
- * :license: BSD, see LICENSE for details.
- *
- */
-"use strict";
-
-const BLACKLISTED_KEY_CONTROL_ELEMENTS = new Set([
-  "TEXTAREA",
-  "INPUT",
-  "SELECT",
-  "BUTTON",
-]);
-
-const _ready = (callback) => {
-  if (document.readyState !== "loading") {
-    callback();
-  } else {
-    document.addEventListener("DOMContentLoaded", callback);
-  }
-};
-
-/**
- * Small JavaScript module for the documentation.
- */
-const Documentation = {
-  init: () => {
-    Documentation.initDomainIndexTable();
-    Documentation.initOnKeyListeners();
-  },
-
-  /**
-   * i18n support
-   */
-  TRANSLATIONS: {},
-  PLURAL_EXPR: (n) => (n === 1 ? 0 : 1),
-  LOCALE: "unknown",
-
-  // gettext and ngettext don't access this so that the functions
-  // can safely bound to a different name (_ = Documentation.gettext)
-  gettext: (string) => {
-    const translated = Documentation.TRANSLATIONS[string];
-    switch (typeof translated) {
-      case "undefined":
-        return string; // no translation
-      case "string":
-        return translated; // translation exists
-      default:
-        return translated[0]; // (singular, plural) translation tuple exists
-    }
-  },
-
-  ngettext: (singular, plural, n) => {
-    const translated = Documentation.TRANSLATIONS[singular];
-    if (typeof translated !== "undefined")
-      return translated[Documentation.PLURAL_EXPR(n)];
-    return n === 1 ? singular : plural;
-  },
-
-  addTranslations: (catalog) => {
-    Object.assign(Documentation.TRANSLATIONS, catalog.messages);
-    Documentation.PLURAL_EXPR = new Function(
-      "n",
-      `return (${catalog.plural_expr})`
-    );
-    Documentation.LOCALE = catalog.locale;
-  },
-
-  /**
-   * helper function to focus on search bar
-   */
-  focusSearchBar: () => {
-    document.querySelectorAll("input[name=q]")[0]?.focus();
-  },
-
-  /**
-   * Initialise the domain index toggle buttons
-   */
-  initDomainIndexTable: () => {
-    const toggler = (el) => {
-      const idNumber = el.id.substr(7);
-      const toggledRows = document.querySelectorAll(`tr.cg-${idNumber}`);
-      if (el.src.substr(-9) === "minus.png") {
-        el.src = `${el.src.substr(0, el.src.length - 9)}plus.png`;
-        toggledRows.forEach((el) => (el.style.display = "none"));
-      } else {
-        el.src = `${el.src.substr(0, el.src.length - 8)}minus.png`;
-        toggledRows.forEach((el) => (el.style.display = ""));
-      }
-    };
-
-    const togglerElements = document.querySelectorAll("img.toggler");
-    togglerElements.forEach((el) =>
-      el.addEventListener("click", (event) => toggler(event.currentTarget))
-    );
-    togglerElements.forEach((el) => (el.style.display = ""));
-    if (DOCUMENTATION_OPTIONS.COLLAPSE_INDEX) togglerElements.forEach(toggler);
-  },
-
-  initOnKeyListeners: () => {
-    // only install a listener if it is really needed
-    if (
-      !DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS &&
-      !DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS
-    )
-      return;
-
-    document.addEventListener("keydown", (event) => {
-      // bail for input elements
-      if (BLACKLISTED_KEY_CONTROL_ELEMENTS.has(document.activeElement.tagName)) return;
-      // bail with special keys
-      if (event.altKey || event.ctrlKey || event.metaKey) return;
-
-      if (!event.shiftKey) {
-        switch (event.key) {
-          case "ArrowLeft":
-            if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS) break;
-
-            const prevLink = document.querySelector('link[rel="prev"]');
-            if (prevLink && prevLink.href) {
-              window.location.href = prevLink.href;
-              event.preventDefault();
-            }
-            break;
-          case "ArrowRight":
-            if (!DOCUMENTATION_OPTIONS.NAVIGATION_WITH_KEYS) break;
-
-            const nextLink = document.querySelector('link[rel="next"]');
-            if (nextLink && nextLink.href) {
-              window.location.href = nextLink.href;
-              event.preventDefault();
-            }
-            break;
-        }
-      }
-
-      // some keyboard layouts may need Shift to get /
-      switch (event.key) {
-        case "/":
-          if (!DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS) break;
-          Documentation.focusSearchBar();
-          event.preventDefault();
-      }
-    });
-  },
-};
-
-// quick alias for translations
-const _ = Documentation.gettext;
-
-_ready(Documentation.init);
diff --git a/docs/_build/html/_static/documentation_options.js b/docs/_build/html/_static/documentation_options.js
deleted file mode 100644
index a2012ad39..000000000
--- a/docs/_build/html/_static/documentation_options.js
+++ /dev/null
@@ -1,13 +0,0 @@
-const DOCUMENTATION_OPTIONS = {
-    VERSION: '1.0.0-beta.5',
-    LANGUAGE: 'en',
-    COLLAPSE_INDEX: false,
-    BUILDER: 'html',
-    FILE_SUFFIX: '.html',
-    LINK_SUFFIX: '.html',
-    HAS_SOURCE: true,
-    SOURCELINK_SUFFIX: '.txt',
-    NAVIGATION_WITH_KEYS: false,
-    SHOW_SEARCH_SUMMARY: true,
-    ENABLE_SEARCH_SHORTCUTS: true,
-};
\ No newline at end of file
diff --git a/docs/_build/html/_static/file.png b/docs/_build/html/_static/file.png
deleted file mode 100644
index a858a410e4faa62ce324d814e4b816fff83a6fb3..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 286
zcmV+(0pb3MP)<h;3K|Lk000e1NJLTq000mG000mO1^@s6AM^iV0002xNkl<Zcmb`G
zgHi?o6ovOGdxdP*AltSE*&JruJwUGI3!FN?xxO>s`hMrGg#P~ix$^RISR_I47Y|r1
z_CyJOe}D1){SET-^Amu_i71Lt6eYfZjRyw@I6OQAIXXHD<M{a4P!N^sPbQKi=?mBx
zoos%BSoiGXjr-;%$QixXMOVNSUNp6L0a1Oz&cgu)wqE?07u5I7qrQIu4Fij)Y3c&0
z@0u_#NH6I?Mk(n;dT}d~^J<WkTLqp|RW-hV56tKpXqu)k@V{?amI+5DOlEU@funz+
kySsbM>fiX^GbOlHe=Ae4>0m)d(f|Me07*qoM6N<$f}vM^LjV8(

diff --git a/docs/_build/html/_static/language_data.js b/docs/_build/html/_static/language_data.js
deleted file mode 100644
index 250f5665f..000000000
--- a/docs/_build/html/_static/language_data.js
+++ /dev/null
@@ -1,199 +0,0 @@
-/*
- * language_data.js
- * ~~~~~~~~~~~~~~~~
- *
- * This script contains the language-specific data used by searchtools.js,
- * namely the list of stopwords, stemmer, scorer and splitter.
- *
- * :copyright: Copyright 2007-2023 by the Sphinx team, see AUTHORS.
- * :license: BSD, see LICENSE for details.
- *
- */
-
-var stopwords = ["a", "and", "are", "as", "at", "be", "but", "by", "for", "if", "in", "into", "is", "it", "near", "no", "not", "of", "on", "or", "such", "that", "the", "their", "then", "there", "these", "they", "this", "to", "was", "will", "with"];
-
-
-/* Non-minified version is copied as a separate JS file, is available */
-
-/**
- * Porter Stemmer
- */
-var Stemmer = function() {
-
-  var step2list = {
-    ational: 'ate',
-    tional: 'tion',
-    enci: 'ence',
-    anci: 'ance',
-    izer: 'ize',
-    bli: 'ble',
-    alli: 'al',
-    entli: 'ent',
-    eli: 'e',
-    ousli: 'ous',
-    ization: 'ize',
-    ation: 'ate',
-    ator: 'ate',
-    alism: 'al',
-    iveness: 'ive',
-    fulness: 'ful',
-    ousness: 'ous',
-    aliti: 'al',
-    iviti: 'ive',
-    biliti: 'ble',
-    logi: 'log'
-  };
-
-  var step3list = {
-    icate: 'ic',
-    ative: '',
-    alize: 'al',
-    iciti: 'ic',
-    ical: 'ic',
-    ful: '',
-    ness: ''
-  };
-
-  var c = "[^aeiou]";          // consonant
-  var v = "[aeiouy]";          // vowel
-  var C = c + "[^aeiouy]*";    // consonant sequence
-  var V = v + "[aeiou]*";      // vowel sequence
-
-  var mgr0 = "^(" + C + ")?" + V + C;                      // [C]VC... is m>0
-  var meq1 = "^(" + C + ")?" + V + C + "(" + V + ")?$";    // [C]VC[V] is m=1
-  var mgr1 = "^(" + C + ")?" + V + C + V + C;              // [C]VCVC... is m>1
-  var s_v   = "^(" + C + ")?" + v;                         // vowel in stem
-
-  this.stemWord = function (w) {
-    var stem;
-    var suffix;
-    var firstch;
-    var origword = w;
-
-    if (w.length < 3)
-      return w;
-
-    var re;
-    var re2;
-    var re3;
-    var re4;
-
-    firstch = w.substr(0,1);
-    if (firstch == "y")
-      w = firstch.toUpperCase() + w.substr(1);
-
-    // Step 1a
-    re = /^(.+?)(ss|i)es$/;
-    re2 = /^(.+?)([^s])s$/;
-
-    if (re.test(w))
-      w = w.replace(re,"$1$2");
-    else if (re2.test(w))
-      w = w.replace(re2,"$1$2");
-
-    // Step 1b
-    re = /^(.+?)eed$/;
-    re2 = /^(.+?)(ed|ing)$/;
-    if (re.test(w)) {
-      var fp = re.exec(w);
-      re = new RegExp(mgr0);
-      if (re.test(fp[1])) {
-        re = /.$/;
-        w = w.replace(re,"");
-      }
-    }
-    else if (re2.test(w)) {
-      var fp = re2.exec(w);
-      stem = fp[1];
-      re2 = new RegExp(s_v);
-      if (re2.test(stem)) {
-        w = stem;
-        re2 = /(at|bl|iz)$/;
-        re3 = new RegExp("([^aeiouylsz])\\1$");
-        re4 = new RegExp("^" + C + v + "[^aeiouwxy]$");
-        if (re2.test(w))
-          w = w + "e";
-        else if (re3.test(w)) {
-          re = /.$/;
-          w = w.replace(re,"");
-        }
-        else if (re4.test(w))
-          w = w + "e";
-      }
-    }
-
-    // Step 1c
-    re = /^(.+?)y$/;
-    if (re.test(w)) {
-      var fp = re.exec(w);
-      stem = fp[1];
-      re = new RegExp(s_v);
-      if (re.test(stem))
-        w = stem + "i";
-    }
-
-    // Step 2
-    re = /^(.+?)(ational|tional|enci|anci|izer|bli|alli|entli|eli|ousli|ization|ation|ator|alism|iveness|fulness|ousness|aliti|iviti|biliti|logi)$/;
-    if (re.test(w)) {
-      var fp = re.exec(w);
-      stem = fp[1];
-      suffix = fp[2];
-      re = new RegExp(mgr0);
-      if (re.test(stem))
-        w = stem + step2list[suffix];
-    }
-
-    // Step 3
-    re = /^(.+?)(icate|ative|alize|iciti|ical|ful|ness)$/;
-    if (re.test(w)) {
-      var fp = re.exec(w);
-      stem = fp[1];
-      suffix = fp[2];
-      re = new RegExp(mgr0);
-      if (re.test(stem))
-        w = stem + step3list[suffix];
-    }
-
-    // Step 4
-    re = /^(.+?)(al|ance|ence|er|ic|able|ible|ant|ement|ment|ent|ou|ism|ate|iti|ous|ive|ize)$/;
-    re2 = /^(.+?)(s|t)(ion)$/;
-    if (re.test(w)) {
-      var fp = re.exec(w);
-      stem = fp[1];
-      re = new RegExp(mgr1);
-      if (re.test(stem))
-        w = stem;
-    }
-    else if (re2.test(w)) {
-      var fp = re2.exec(w);
-      stem = fp[1] + fp[2];
-      re2 = new RegExp(mgr1);
-      if (re2.test(stem))
-        w = stem;
-    }
-
-    // Step 5
-    re = /^(.+?)e$/;
-    if (re.test(w)) {
-      var fp = re.exec(w);
-      stem = fp[1];
-      re = new RegExp(mgr1);
-      re2 = new RegExp(meq1);
-      re3 = new RegExp("^" + C + v + "[^aeiouwxy]$");
-      if (re.test(stem) || (re2.test(stem) && !(re3.test(stem))))
-        w = stem;
-    }
-    re = /ll$/;
-    re2 = new RegExp(mgr1);
-    if (re.test(w) && re2.test(w)) {
-      re = /.$/;
-      w = w.replace(re,"");
-    }
-
-    // and turn initial Y back to y
-    if (firstch == "y")
-      w = firstch.toLowerCase() + w.substr(1);
-    return w;
-  }
-}
-
diff --git a/docs/_build/html/_static/logo-dark-mode.png b/docs/_build/html/_static/logo-dark-mode.png
deleted file mode 100644
index ce0016561a8ad6995fdb5729fcb78b9cadb56829..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 8504
zcmeHLc{r3^*dJ0#$x<Tgq*Q1YW(+eKWy{te`w|&rFqRpNT?i2+T7>M%7Foj#lO^g^
z6fKs**h+}8M~vkg>fNsI`{%u`@BQz6=9;;l=bYdDyU+bQ_c`Z&CdN=-i<?WB3jhFc
z>u75j0RU|MtoxkZyI8;I+)M@lz!@88V&QFs^#gi%ksJwbc%XNH2Ofy`CpZEC{=K8g
zmUsG<_D2o+Tt|kV<p=tUg6*F7XT=Xn+?x=lU5txM2}n!OGtoULxVHCJ{YYu(%;crZ
zMqUlq%_xq(3)01voi($}?-L!C)AMPW9Vs)iE8jP|+L*b|i_fQCx<t1vUZZY?3~R}z
z%dQiEnKPW9dpH?r)p>&*bthV#G$&5Yp$&t05<X0ZJ>?uPb*Fbf-FO<wHv_6~32tf}
z^wA&4s7iZg+o2K+xi=2NXQUZ9yiy!X8{GSH405T4d_ie%x`l^XiE&BCiq`qvFU>ty
zf56QuJB-?r(nC~7KYu%5wR-NyYOI;g40y%(7USXi{D#!3f{ej<XiCiwrp$U^Lz~Sd
z+0}#V3zX&g^{M+di}qK02U1s~duy@=?XJVBE}eQK^r7(;q2ryBsF^`Q*>nI;oHT~t
z!(F-}r!CRHSNg70;z9Gmi#c%P5Mp4-qul#EA)}%d#&1BSQ&~z9ANBMlTyHz|`gTm-
ze{(-zkIh}Ekj9S<&*0_Ay7+?{`lr6{1Wf8qaHQfzzqH>c0wtMZCc*4$pMR_&sgAsz
z2S$(HK4;hB8l62aapEP<+xv04%=cNH-^H)+%2ibey#fMU3e@%?a!fRL=j`B}noZCX
z1_BES<u6UV5M^`Td0+3z81&@C^a(D@!EU{EJNa6*V27DBSE!4pQm2CJ1=A0OB^VlB
zHqX(Fk!)XDlUU?ZWmVK;MstFc5N+vo0i;Lx^Tw%Fy-%yUPm_JmO+Teiyk%-1jh9~(
zMtqxncP{g-Q{tl!j_E3WQ?kWRXXEij;7jh`Z(Y&7pw|2#O6x{_c5izR*qXwq4W`h~
z7zQ=3_RVp3ez|ECcnc*q$?4Kx{Jn+OYf6L15s+9S3AE+mX!*3?k)kI(0rZ2?M3#~i
z%C*tSAY^oy-5#CVQBGg;z$1aTn3*0lmzABA#X(6wvGMwt{uyd;WG{-~tP)*k;vXL9
z)Tx!3krw)KWQOrQpL5W?0c}{ABPRkfD|Tr|w7I!$%8=bImD9S|)im#vTCyKa=l?_!
zc?8jnU5~3cYR^gUcJX?8>nuUQ`gRHOtN=%wOc!bn-TTVGBw6W=Pz_H4q-gP`=VM9t
zXSN~-p7-j~?ZZZ@TzDw_cAN)(gX!5?RQRfedR>l5_e+~(J0{ec)KM~)ZJH4-p_uZ+
z;$h2W?r*1`TlEYxZHtm_pKvzF54xWG&7GofZn9^-UPtsA^$})LI<PJ1b=UlOv+4Eg
zcTzGumW{t&rrU0I4pV%!yUL}C4{;D)e)KUFtw{QEcd@M7g3|lG*Aahjr%Zm~t=gfQ
z6QY;h*N@MPUKjU_6UB9b(9=kUu$^pfwQ~^oWCmwJLSh{c^b$b@+a=d3u8<r;lKm`w
z?<28Fgq$4ZOtC&UQz+ZYe<NXu+WgL0=DZ4RuolOUxR@cFW%jDMc*#Z1{xjstLkke;
z1l2)Zr25>EPcq&|70q?SlujM{l&%26=N>{tupP@I=?!SV4%Q}xRPp=lry_f8hCBLE
z4=YxxwT(G-#_=ARpInXhYSlt>b#S(s)gM#r-CjuHxX}kPPq%=Rol+uKh*5RvN;Sm#
z8~LgiGVYh45lM=B(t6Vb{2rO_5yf12-B^_XLiu&PJJ(&GDs}$E;GvbyLoadGQ<k#9
ziAQ3hlJwua{61g!E}HP|LvkKHQ6wh`R(&*+BN%XK8m?Z$N8r4wn*G43!7TnbG&I8e
zXrsYu^zr<ZE;|v=)I`iX?^Z~s`K>$iW*}Oow#h+>qDEzBU?QY1JUe67{ee`uuSs=4
z-*0}n1O;I{2ejh-DTYuk;J6b$r?XCUHGwKBlJ)#!y7c)2w(sU9+YQpgOYVE-z;;UI
zsx#kpy1rb$L-BomYI?OGb82Uvl{sZhj-Qn8Q_MCk#i^i&>_v}mPJUZ^`=j4BG;o%u
z_|9;FaJoZu-g{m1ehtST!%#}nLpkj`!V`SYR80>R28dg$@!fJxJm)B^xZ;=)cNkl`
zd*;sVc`9tr_`Y4mvc7TRz?gwWn{v^ll(yD1ZeYJ}?K64ULb=_Xkn?a8t|uxZIdh*w
z3V&gP+@V67>j)~n8G)klybZ1@u9zQIIb|}?_Dx@m?a2;&7BE%h<*_KST`z8`otP!!
zE(r02=)6GN3rC&<&o-S%u{7D^|6NoN&U2r8vmnx)ku#~3Y^r-ZNPJBf?riRPr@oqY
zUm_2?U|<>t)Q_E!z2<b8Bx|tJ6)irdvbHM;=cyQ|gQ~!aEF8?c=@l_=$ES;wnQA3H
zOKCB+e1q0zQtZlVvNb0I_PcZ?+Gpq?W7HNtiJF|4WLPYRShTEmU&P7mdB3VGs^w9Y
z^2%lUn9SE_SqKH8JOk<XF0ZpHQ(ZLDiqrOZErv4zOKF0%7m+$@pJQ6koMDHL$t6b^
zZq8n1zh0PT(ruY~q4I*$-96I7A7OawXOyZ^a0_3(QKz+GVp#p&*ZnYaSED5<p##Ic
z8V?mQNT?4#EtfAn)AQ!d(zVuS@vE(CmcjGc^^<)vOm*g_GOr!OXCr@@ggD{Nf9?If
z>#T$NDnVV{P)A+;_j8$b8m9%{Mr${p6@78(gRZ8G*eRpuN7Q{PfHETmrYF@R4P`68
z*?FyBJ<hFj{l>!~E00-R`M05w$3+oEz}+ENsI^Mcwc0V8dR!oTWu@sBT2MUA@5<Z9
z@?HC4cN97vo^Vim{nhNnkSzOdEboP@c*k$#I|<=qKWeYVm-=~Tj0R74yvu|n@0*Qz
z{^Y#ksS)mx35?#@Qv*>f9%*n{U)k~P&KQZB66)G-HDAgaVMos$M;nM9IizRv*jr9k
zL}u=w=c!Vc1c-x8J9%K_vDxI~V1tb&Lse_buc$ZJW989O+Pv9Waa@IEy!oa_UudK(
z9T49*(gL}ycf@>WtmxhX)ySzI&o+nrD2|?9O&BhzD;a`)8$xX1w6jInzQLAiI*M(r
zqw!w%Tu(~l>k|_<;|`r>Z|ZA&;iF}|bfrxUxSG=zHKiSrqE^29Y34h{(JNurS)@0N
z;=1?`oIrJ(T;i1ob-PN9Pm|&^(~N1xAdImQ0+2_0-Q+WD7-!W_!UR^GWPv%4!jas;
zSO=0l9_;V#!K$MG0Od3O9$1_!-WzC-cOnqck~6dhNg%-iEorWRfnq$=@y-P8Krg&;
zpuPz%&=rStkUXQprR<Ml5xC>Mu|R)!HzFD3kCxo#MX~O;njw<FZHTukTG9ez2vjF|
z;el{491I0%`V)NRB~`e9%3cnRC?kz?zbIHwXh~;pZx0j%;^*fF_Je^*UQQ4>BoYaM
z%0uMkK`aD_96<EO`h$q%qgxa|IW+KOoEO2vn?NE0w>Yu(Bp+|Iq$H~!`1|<WJusL*
z=!xWCDzNl`_+vdFa$qRL-5v5*3$nMSFN@^Yfc{SlvI(o)gc#w;Bp)vvUegy(^gjAm
z3J2UD_8vZ7ZrkBF;2?N6ygLg@W_gwSn@cSnjNu;^TM{@C+&#9fSYrPT=}mC_Q>?#@
zZL4QHoWBl)W&Q{6-_XCuz71xfFc_2u3FoslJRJ?R<d%Pw0|`fPKyBY*kw_Rq5rzlB
z73Ca2aD+V^q=bb!g5XG~99~fYi?_$&|3alhBzt3tIQ$kBiyTZ~@!*m6IE14j6a;gy
zmj}V+VR9fW9<KnB!#m>aadJ>4MJ)6$6b4=dR#sx&{yM5HDhC!777NA074h;Qdn6VC
zg2Ui2kUbOu2Prx#DmcpH<P;s{<+iCDaHw-6FLx|!ItlJrCp^T1=(OFjB{)jePzNn3
z5B}+OyT#BA>+Q%gKuew{5PkgrgqjfC@y6cREj8s3N^&rm5*#iEha#c!N`HdP@LptA
zCT?-cLBX(}yj#<PVg<tz7Q2<FEP`zdRxBuWFFe+p<Yhu4xuGSu1_0c$+)i(x^3O@p
zCXiW(fUS)GmGj1U&!0~}mw+2#y9)^1PFoZf_tOa(>x*~TcEqy#*@bh)5}oj@_5CZM
zes3rIACo1IN5T<MBp#$7FUOiJq?{56sRV_But<cxgOWYeK~WL+cXl$#(c2H}g;#ZA
zMaqhXl|b9k08eg9DgAe}pEG_-C#XCO1cig-<W1xhQ1Xf>c{vbN5e0=xLVg7d*;>`V
zN30C_Uz#Xy1OBoNu<U-ev9=f1ZUy;cyZWWsmW==6^=mHvMGq|M|D60IegBi|Ke_&q
z0{;m7Uv~W`*FRF=AA$eNuKzc=xc+?M!4p|kkRR)Hh9O?K_1E$)dp#`;z~<H`t1ka8
zt7VUewj~(=*vr55WqVH#Sz|SFdh1{`ImbBm9TE}9iEzFE0C4%~XsDX__cG{b0=s?_
z<{To;=Wv=`v1D7~vM)4_dJM2&drRf6%-{1E=&(~RS9VHZh!%cJz|IUxp|;e7h~iuq
zLvxmkPnN~O-`~%yzukVGgXQO6xXSLrpF!jFM_9!rAtsvfo6oM<h~of2*j@m;Ab^(&
z0D7>U9b(^l-~a<=A!l5|L&VabduptSdNQx_Xq_k*<QwDjG$>-*eUZr1)4xQbr2gRN
zTPA({N{jY?<inQuA|VLR##ELJf$Z+;oLk5^_dK(TlKM!D&4l`7e;wm^9E5b8<zHD+
z{Pa-(yB|R{b*%DCA7*N1X;NK7+4?i`71F>*zuidpnz6XLuJoI1bOS{qnCKe5u&$E3
z=tAt7<{evW-AFt4+`xtc`=SEA{hRP{=SwY{w$ZBtE~=b;m^8os4kq7RRTP5ZsvXeO
zSobO-;k{;YSSooSKDvY5FAb529r}u_zS}5JS=^N45l~O*!#uTAld6BvShsjGcQLoA
zk>PVDQHUB`UDv2NRK54HU?1kPa~b_kW8JrS#_<e@w7;M@RWn_`z2^IAHgX9yQIr~7
zxtYf9cf~jOw$Uo@?tW|bZ<1UgJFJC;(WlOCF8dw`ip)f|VZSPHOzS^%sv{S8?VP4&
z(`Fk@M-i2^NqJlmvwgk`GlQ*hUxx8rz3{T?RBCjzNi%doZ9Re9PM*2!)h?)SN+~N;
zu=8pkYaj78Q>gk3juwt&_#PlMmDLd%v`SAB=#T1XpBt2gooN_>4q{&TN>7cy?L^5L
z@d9-J7sVCv;$zMtJQB+SX6Sv=%N3#I0$+1@PQOxCo~T(y+brp(dTWnpPC|RW%R)`?
zwf=-X{SO)^`1dKkQR{LDa%V5Eertmp8t7uf2vBJ<EBL`o+WSUOfF*IhQ7jv6y>6mE
z1U~A#EICqoWXEs`P((@S3{i2eGU~03>(D?q+i2RR<rn)ejbb^UFS!ose^Mc@Z4`Ks
zC@-C;G>^l_+~V~|X2*}s^RiX<C2?|ZGCOdlGnGr<6XPZbsoa5PzJAw2lcIlASNjs?
zUa#DqH5p#AdUfxr;Y3GtZPdkA{jVZh<N~ju@sWLbAI9gcKi%obw6_SD5FoZKq+e6+
zAQyP(o!Oj!GKRoj<MI>3rsBT%xKdI-l4|2+=mSMLpJlrYd*Y_673qWV7e~jH8$x4e
zY*rM_Yg5if_B~n^m3&!3TAl}us<4+#guJeLUtHW&H%!jddR7s8)TBnDZ=Vj&rw|2m
zDlQgvKH$lR>RLwGjOK@QuKs)181Ho886WE$z@u^Mue+j%j4&E=YD9}`$F6B-^?ccV
z{oPNhZlVwL#O~-RoA)z8=UhZ}TF=7LJ22H<UqN5$1$D;xA4DbHK;2?0%!Ibtov|1W
zu<h)y@@mg*+aGc;gWWGd#x+`|VytyOOwZS-u7QpkYkxeEvw!bCy423sq{x<y%6Ojd
zi1pSTeqd}L!rO0i(>0Un)IJ(mbWu!rcYpa!t2vAU$7u~HSKo!^IghE<n(CSEeRV|7
z3<D=$x&U|lpxGx8vqRLRqR(*dWQc-Um%8#@f$E1rF#Kr$d~TAd!Mh|QmnVkL4Pd8Z
zgN#F{7m`O!sZyq6&Ao-@*wcDWutK{5q}b)LUTUep*D);&^i1yFS9|*|$YwSlrq?vY
zMw{<&>p8~zooVH7lYe|Of3dd|8|*AVtT}<++5YKZn2!&QmQBLa>VXnD#xBUTwbsLI
zVa&R>NYHLdopm+&<LJ!ksOb&xK?yGFtY%ENUB{rcKGoho2rEE*1^d{5B94`Y?H^v?
zZRwMD3!Uq-a35P$D>5q`$3NnWy4--0+Pk;E$z`6<(vZ2x-F~FmS23cy+_K=eo{G;j
zKRuxa*K!_-w3tFt<5FdKW4iS{rc(=}Kb0w0y}OUpL!mpa>N<|*9=Aw;z(++W1{j0F
zPI}1hr_<OcN*J;07b^)rSi7j^jo_d%`t$XY%r_3Em`kP;gH{JPrcb?buS!QuE=@Nr
zKCnCE-P-H|YyR*=MD0hJ=4eI!$|j*<fopjs0(Qbf(0VkEw|3@Qh;82pX~~4vJmDFY
zZ(#N0ZVPC`@ta_KhJdZ>qh=o$iy92=(OKqUmHPUdsol+D<9XR#sA3$rsqE@^kJHmF
zmbw!juRpt$-;B30yKdPUNL3GB5^6D0&ngb)kIm(gZV-F&p7F336kywn-CaYuVOw=N
zUjLD+Q(sSoo1P&~%B)SLw64|WvXdom_zWerFaMS7zRnXn1*Su*Q$I=!CGX<V5+H60
z5PNX+b%bB%pglOZiH@eZ!_t%qk=DkyifdYJUW5s9bVkr@fxU0wUE^KMBGo4%_A0T_
zAMV^QFA4VTYadEtofY{m=3%D;hxGYOx;TkE4>|5~Oi$w<IbGNjChL8p8T-1m<k%Oj
zLYMfXBV*;uSbP{$MlwWpEv-j7*DsgZqb1TMPF0U$><+W0Lm5`;ex{r@h7;m$ik<^L
zsWuULZLs0&jN-WOh{fI=6m_jt(DCHn((Xov>|>6SMgc0F{ow++3!!RU3m=R&WqcWi
zc#QR}bn#M=V~TuplLy_EI;*S~^t4^~(M<>N`^w9hRg8uk(MjBc#A(qoGPhM&AIPI+
z+;D6amXgxsP+)hE%5FZ67~H_d!dDtEMwkt8^$T!JQxe{<$Moh?<L`f}Z7XWal;{zr
z_F;1Jvb|O|zb>VIsNn>is<$WbR_`bbM3kqIZLP;k0>Ps;c47%v-@Vg$df%VY5oK{c
zt2E~HINgqHUK-q;DNsFl%h+J4KFW8a0g|*?u#qxXMY!~dekc&lBfVKie{Ngc_4vr>
zl6tFqe9%<2W#6rNA(FZQ+VqP>&jg0)W4Ue%Zn|xRvQaaA+IwU)RV^ESWtX^%xJ$)#
zo(u2dzo_DA;2yQ_#u`uCk)c*<w7vZzGUe6Jg5FF<g>LxL4*Mq8dOzTD9MjZT1H+H^
z5ON;0oJFNgtZd-2Qdf#vm^t546$Z0MfS~=eF-yUSBd*#tL0a_XS(iaZ#p01sBqo@p
zaX3@Rtlnza_S3-D`Q5%k{Thidq3$qwdSK^Nh{k2#?#7d6M+d&9br|b1lg6-uPo-rf
z9vz?xaL#ZI%BU21pJ^ZSp4!2EBg53U-r(rS;ygahs)SB=vn1{9SWe|@5aHyTt9;{W
zGwq~1foWgqDN;n6t1Sa%zV0sx{umq+eGL;{o8RGD=l1a;(E`9l0*BbVkB*4*Xmokv
z{#G<yV=gd{UGVHsiNY10hx7OCIKXE^L*8=yy@J|(5wKNG{X3Uo@g_Swc73?oATVmn
OgN~-YMv0nT`2PSApK<8`

diff --git a/docs/_build/html/_static/logo-light-mode.png b/docs/_build/html/_static/logo-light-mode.png
deleted file mode 100644
index 3f2fd3164c3c52f91ceeb455a71aa1218d1ec937..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 10324
zcmeHrcQl+^7q?CbQGy_3j7}K63?phnMDHz(nbBu3qxVkKBm`+nqPOV1gb>j?VF(eu
ziwIG_k$d0wy?1@zKkr)K``>*mvz~MI-oL%~@9bx<bIv0jt$UQ@%;b1@c$DgDDtdT$
z_%pcoY%&tuFC-^p3lHzweP06;cRi>Vz!mLc>)?z4xcj&w00?gfTRc4P>G9|0o}BfR
zDi`(Spb*e*`OT+zgL_m>{#Nj3;vCf<?WWHw3JpyypLdfIM3ZHQo^51EorYzIYFsru
zwy&)Ir2X1sxigB!>duCbcDek{U}E}V-j9Qmvh|bPKH~_O&zV=%q)34OMa$Q3X$%>Q
z+Jy1@^5w1RX**ZGA4v}Ry;%_GY3Vpr*oWwJzoT1&y!S;4)W{rX)Rt{EcOKSh-#DT?
zmt3x7xV7oAjlFQ<@#XK+N7p>Pzu267NDy<8qQh{Df9FxGi7So#%7`~8r`1?VAuwii
zWe~19nQ9i)v`WFBk<pUH<r4nEtI9FZO0R{uK7cLgRG#w}(=(nQK`-hP4)abAPQ2eV
z8~qSGKKFZ<aI*Q_vUr-z!_Q?M7CXpOihb}^Kk@yD+2$#^@{bh>AIgg3$K*Y6FDKQ-
z5c3TM+?oQj@WH+85427-k7aM6<?H3F5=Qp>@}vXDt=5*c1Ra>Heb_!bb!RxK4_1k#
z@AgTinbM^$cUW__@3gk>d~to3s&t=C!w3I;1Nx_R<<)xwbM;qMMgm9;)A53uS}vqZ
ziHSPjZgPoU*$kvC=Q?X6TEcH|&2sf0w_ZPH3#JySE=w~|z#Get8PnU8IZP&7#LMg)
z!4&v}DbRNJ05O6&A^o4^C2D{Cuz10>?=kQ9=_+UV@w&BGFF~<c#Oa#HV`YsP0~&`B
z#HZx(`$d{&PR%lWr4M~6Eqynd6Xj2s5o7Fc<K;i5!TRMWviU0e<-1;C1eY!1n(F-A
z+RS8kDmGHeu-kr{^^@!YS`hT~n%yg(6k9>^Ap6l$<vbycyZojS<+ihtn&znnja)On
zOEuV?y8Tk^u`gxk-;XKBMp&v(0ZWWOwa1QX4inj*jy{sM#Y<7v1bks7u#@b%onYYP
zxp=cuI;$1<Tr~!k+f75*=E&EpdFyk;snXb!lMGji=D;?=i`q}bQQJ~aPUvehbVd)#
z9{Ny!=l_xUQR%>~zb7one*9h6t}##G;SrW_DIu;JqAA<P>Rg%L{ZkooGEsxn(rig;
z04>(ts~~=tK2a<Aq}u7F#A_eJ12$>59Sa8f#zA(+h7b6O)d|aDlfuu$EAjZ@bAm=B
z&!-#24YG^*NO*LfMx3%v+db7wF&H0Le!W3{eej@hs&4Cf&JIN~@UWrwx#L-35b7}3
zFc8f6g(oqnxl!{DkSM@w_^@n^EmX(rp3KkEN{<w7F}M2sgx5WSvOgZec|?*Pcs^-Y
zF3MjFJeipyJSIOVbx<o@KrQ9Lj*93m_C8mXzq;LM?@Qi`GBE;4JR$L@_OfzYGpqff
z5w9h0jHM#f%E3RoO$u|z;CqJ`_X#Xnh+2kaX5rzr_*l5_4j?<;;dh01<b<8P(E>Ty
zCo${f@+5fX3jCOQ974r;G9%6;77sHNNUByepotxJ;;N|cgX?jGPSBbre{Yl75;?fO
zOq_(B+~ldU15pMae}5f^!A5IV)hQe*5?DS8xL%<_8)5mg;cZd^IT1bYghSC1U^A&>
zvx8qZrI}aH#kjaqp`hp+y*vFv9y)i^(AoLBRw{76e<i8o^eZiIZbM_2yJrM=A4(&*
zJ}%A6siU*092>?m$d`H%X4!zD-CK|QdYyIbb>D;g>RT!%3L6%=h-($w&P)@0<)&_!
zteLsZ^e^>yr-rS|Bl0bsHHk=0uWU3BOEz;OcF>vZIUl1^>HWnSR9?adjU2^#C%;RZ
zuWz=j1sLl{GxZR?Dp0zMG_kfUP8jVBwWQ=@G9cyh%rw9j30zP3+EO5;m(Hx|M#@jJ
zsQNjnoajxZ)^wzXxtT&RL?9<6|J>(k{Ca3HDgU&A0TB7)vx(phzNC5c6?*A&qx;nr
z*~2?b9c{5q@89cA%5On~>zy&!qQ2Tn3=M^yS-YW<AVo~&Cm*8tD+ezah6B|)qZFho
zUlUZjK6yT|;g@6KdH2N!{-}bOhr99>f|YO86pU)Fj3rZb$ZX5R>1{M|v2n&%+z<Ev
zqV*|VUD<4xs-nm~>j9c6Hi?{nI(lF`dt52P;E|4#xyv4{HEHrlrCv??*$u!TUx4}E
zaque?yvsi)biUj??q@zxz~bFE=EL@Dk2{lQ2-_N(s1_QaaS^$!xA0i-*=r<(8>}NM
zKRnbag-x;hMBWtec5}RbsE|yi9e%6to9G7yIqeEri#@&&Q}i=WiX8ln+Cb(p)R+4n
z-w{&zHirw?nS=G2Eg|-B+HIwntLd^%%)S9`%<AlXAH%j$Z@*usOo*F+G^K@VNSc;X
z=JA<?Bu)3@_vabun{7RQW+1;kcejs)U!8E9PJf3%Hm5nEjf9%P!V`(yS)*czaO1og
zyzO{hN`pw>XiAqu*-o^&pEVB{d7VJjqyl(THGw(cW7cf+iZ!ztSiFnjSA^5sudWPL
zo}}Y(OQDG%ZgM)3484=yPwDQCgOPMFuSD8Bn3^S(bE;`ICsVYIgO1!2&Vi_us3uiL
z<_FfSB#C!xii|<{T?9{moF@9_3q+)HD#16NSy#7F@h^6BJi9)Dc^GHVBw64$%~R2&
zA6waAW+D5zYnPXxuButGse82}I0>7RhI!#;8yEgc@6%Ayeh2D?@G>xh$3QBwGZ^ST
z7(WeIAjWT_fBiGQI*jD?aa}SPtXOE50&cE+TWHEk6eO0PIq>+71f#C&gzstwI-U$@
zGx2EB-^l9lElgG^N|<o#Qv;PB7vo;~f?T-y6?-PHtOk3Nl0`)a^0O_+XDx0#DiMck
zjlC-i_5hDY)-}NCq$!1OaCf^B7yFf)Ytp}_n96Kx6L=lUBV*;o)5cmhj5-;H#}IA4
zWkXt+4YzOz0e;Qu>YQTrDDnG(3EuU+qg!4ihaOTaL37&aCuD)Gz5~j)Wu)KUbw0&l
z58w^gugZGGg*cF_=KKN<)rq_a!G}C&+qATNSqv?34kW5@pV6XfRG)Y%$kPKx-IS=(
z;^N@ly#0-Uf5LSL-}|JU+bzmq^wp<@7QV-8TuS*YKT{?6K7yG8NVVw4NOG@!Cj!fb
zfI`b24|h7@5hw+)Y#7mU9DyreD;pA)G|aA8c^0RtN)g#p31#YE6+lakzpVV|Z?EBr
zDlZ7*fQ~d+TQQo*CQ*6Lg;YwVT8~Un1NDg2xzvsKpKx<CQ(1eVG(`#G$f5#)WR;wU
z?zc^K*M=Ket72njSA4S$6iLHa-}jbbUO5U9eiP|NzM$k3&2pL&n=t&+c&5>R_H4__
zZ&<2rLw^b~s~49BsfoT;5YqX+fSTL#x!tOR$%F6kbd!CL$goGvL7|d{^;nPY(*maK
zE+d=v_75XECAv4S^PZY&m`}>EvGa+tcO*DHIJp`$?fIBm`?Ha7k%FSwo5dRjQ)5iD
z<%;q7rys^w0|$;|pH?rVzp`nx;TqJE3A+W9&9okwdue`~&FBW6mX~?bD^l`vYPu>9
zOF0_}sZP)CZ<TKj1LNTwtuo5qWvn;yuXs^#s`n_~3K^H8eP;$o>9HtRQ-dvpAA~5k
zW8Hiv)q5(}2bvU=ethLPt<sUA(D6zmznK}lpT*NH<ipt*+rvJUH2iWl{ES&>@sZ9(
z4Y>^KE3NQ)ZXY01VU8?**_5sbrd!<7Tgo3mF&b6N0`6kwA;6*vtCrQkB^oUjq`i9d
z(+-Xn#@w;3u)5Kl$knqKwM&Ad4|FAZW#VIUMyOZ43qaX~8^^<r5~7krPLxtMUm^k^
z2R<Y)v$Qltk#*u|x9nYf0{n0D9eb?PWnHyB#(66_3r&Y#f1dC9CXRlK^MfE7k&jzg
z6LSR&HVFyBUVBK0Giz{~9D06rWfF;NVJQk_u#sDR;IJq><n1&fpnpMHt8ecZd24r%
zTGeCZ*)St=NcFa~6sh60zF_fw2@j##+~T0t(6an1hw_OEdbMOHmG%=6Hii~>QUfi#
z_?D)#8DONM{QUiNFYb1+Mdzd23`2y?;+!=kQsNEhy#|KDSJZy^Ba>dM;CsiC(zTK<
z0iV=vJ{!^fqNnRDppAJ&{PEG8>SNn*UdEuI`lcz}btMz&j_Gh(6~`BM0=Zau-)!-}
zF(z_l{K6~>O3rN{2G)NbeQZj}Vv>(O=oBNR@9oJ<z0%}^wC}M6@bDVurVmWfapDKq
ztUa|_35C5otD8tLX5s&F1(ip4#;ABRt~Bc#lDs5X=&@}xFXSiYX^0j=VQLiSZDmQj
zx5Obgv})0nIrnrSrj^$$6c68GjNUv^DU4Ay1cd)R%+6}~3A{K|IpF4X8ebnivx#5f
zkU3`m&SXjMu9?9kmD!Bgm4lw`DlwttKD1rMi&|k;MQ^m4Jy$&0@VEQd*etKw-n|0w
zN)&P(Sz+Aa5%qIA#|tmbD1807RfFYDmUgFso_RLIK%8XUjW%$dP*CVgz1T+6NTJhu
zd_;m^=3e8MBgIj9u@a7n#099l<Bpb>%c4W@i&c4+_%%R@n<q7q8MfzV`jndV^`NhD
zk3F8R5>0k|rJeg3%qH9VJclHDUH$Iy+$5cgKfsb=Ms?oN*H={K>5YOr?B2jA5+6bi
z)buAoFiqC+ntUN0U<(qXT|>1WMX3;#l?&~%f2azGwNv1D0&JnW2MtS~(1gjZ58nMO
z3$cD)h0bRbN0?Cby$kM9e;(;&^Ti2_>N;b;Uh}ERt<v0-Z6vV|d4c)3U(0Y6?}s8M
z2UdE~c+RxXDD*bF)7p0fP|m)bCu-0t<9+$Hpzuhf!U@O63fSAIQ0hjLm}~d0<r9!$
zW9SHMdmCR868l}KWL&t%`zP)wM}HkGJF8EP!SGWyPoo(^4wtn)8?J5+r6$<Xr`_2x
zKTUKI^2<fHc|hq*iQ?yotn7G#eDRbN@HMTod1EzYhU>;26eaeDimFC3s&sv@`9{<q
z0u&)*9$E;MaFI95MZT22?SG;+;3rc52tbfRj+|@bTZ&#rR|hkNE2$6XjVhD+9;Fp9
zMMjhj<GVl?cDH%ARXh_+wOw@l!_FvjPnPSUoKY<Ok4f%qYYN0LEHr*bcl2qwY!-e{
zf!9chkD7tD9b$!M)DFAm#!VxoQW_+8%QD!ANGH<hmEH(bqu%j5wodKy7X8TU*V`N*
zpi^lcsU2b(3H<8rt`DTRCa0wMs7C54lVQ)f!W>5k0>iQrWa+VTD3$Pa=i$+XEcIf6
z*eOe|4#(;TjaE<GF+1ErSy@M2S@~b5YTQ}+RlpO7+6RT3omQjwRry(O>vgazdsG1U
z7xE2xl_GQms@AR1XW`tG>M^lz=gnMqVdd-d3&llYMF6tE;JOBBz6P~N7aHV(HgeM2
zU5$Q;jb6b+#iHX^;|L3FnbzP+eal9j^8y59P-@+9gzY-!X>#bwQA1SXJ1@8NrGV|h
z;S5mP)!j!Ol@FwDFHkP5X=<!AY2Sn*WVHRY<ZOqCAKj`itvmf#zgYHOjN>jhMEfQ?
zgN9|XyNCb_|K4@C+wUBcL2%0f%$J2?qmAMK?elgWMRU_-a6go@Jeo_5Ix91Qys(Tq
z&yb^2<;71rwsZC_&=U=IW8%1*6#0q~n@4RI^Io;KZs>MRa<1TXM(p_w{Ln^)A}DsQ
zt7aCAe>(X7*~H%Di}!szYw-yTvIOne_njX1^nV8TvI35?dm}g10$(VVlQm@wODzS5
zm}k25Z<RDBj*<eDEpt%8VanE3DpMP5JKI~^TXSMt=Ye>l5Oh3^PU|Y}_Ji30cgta-
z`2Y-aK?*_PE;a}uZ=@^k76cDZ?vA%B6y}6*2iPF&98eJ8PGc(&-~fjJjU_aNHC>ev
z_6};kXoSA6mI2Jy2_^#v-jOGl^9JJpkO+4uz#Hj|!hpRYz~8uF-23G;2nhJC;_d_i
znrP|(lwHsWfVhyjkg%Ytw}YoBP@WthhlbmN^;GWu34t3yfcEb0u3!+z%gam1OH9ZG
zZ3hyOk&yuji-JT&1#udJ7$1~7)LRgR;kbnO149LYfuS8--5p#|fJ;oMjf;mn1PH{*
z0so2*>8h#uZ+H~uPZn@|fV`ouAQ2&95E2Rcs|Chg)e{HuXF&f?3ycBovIwMyz_@sz
zVF*=E1j?P`uMlw9zwKQ;(9XZpfx|!uX9N<biotmm`J2mo>Y6(Lwz#Cg&H?HA+X_ea
z-!$DFZ2v>n-(tIz{7&buf#A&ljr%w4f93vL8K<SG3084|d0d94t^xsG`Uk^ZU=DEb
z?^hd988Jx&LP8KGY%3)wE)J6vlopo~7L*be7nK$fg~FvJZ2khJj>5P@Q82_M6b@X-
z0f!?hE@5i}m$VTS78j8b6h}zH1)*?+s31&I1|}^cCMqn9u>A{!HrfGKl~CuuMs*1V
z$3aQJMWJE{X)!@tF)0Z_aY<3AAXG%gMi4G84TB29VKSmn#BV4#41CuGjfCQs(*X&!
zLx5aScE1Iego71z)FD7oA>seb=r}{&ZE*$=-~$JghxdO31`bGszB}}iO%X|4v|`fY
zBBCO=n1%mhZG=E$aFuw8DIzQ+_J`zhS-`ktaKu6{>l6p@8y1%fSQ(9gy1SqaTwI(X
zz{`*TmzKZF8zA?`qNq7wa2h_B75`_=>m%I$82#}HI6M3n0RX?t77T^`;RFNqM8JPL
z!rA>Hg4sh+b_m?}{bxb_Yu@4iwOBGDl5lBTgrJBx9Ew{k8@Qk}94aOVx52S1CJdE<
zLB;=$j&ZSd_kyAkigvh6aoOMs=yx^%-rrR6{axG39&yQ&u&5ZW90Wx~4MZfsqLN?<
zQ9)rTFpg8up9zC5zv_QwEC>3(IFb9U@Rv4#v-@KX*Isbl3iR)G^(SYSH2x2tKiA@a
z=m7`)kCXq3-+$=(hpzvMf&WVRpX&ODuK$XG|4R9v>iYji7x{k<JO~tS7vzOI&IHY<
z?BR|=BsLoNRPZh?|1z8Nl5jIuT-8i5cz6`Fm;d-9vw^3$Nm6%pO;yqrqN@zd;(L4Q
zS~#_FbrnSe@9C{sP0vviW|Dc3M_N*G<_Gd%dg{CBdIA7b{IL%GzOke?Mhc(tnS8@d
z74wXB#>mx-=Gi(5JabshpYA*>&0dk=3FD_dvp8!yF3m0?!@2q&Jl0!kbc-UE9iqU8
z4v(ULbp(>^#&YA~g;3xTFyc|y;Q?In73K*p2Sh@6yFuypp;eM3O-Irh9a9N&vI)@{
zW>*#FD`ZWyITi+UgYIN>XgV{kIW<Hd((E-dkWkmbsHU`O_JmgCmXnUI(`1B$g5R^6
ze;NQTC^PN{Qn#ei>`gG>uku{egs@c=tpSZJBeqg@Ayz>|^EMqX5#z&ylPbLz%fSfV
z90)ee2#Y;&$~r-5xzJX#N*u-XretagKcA!#iS24-aIa{IGH<V_aVvl9c*mS_{<A5)
zdvPhPP#6CPaEWh2Jk~KoFM&(uP1f|4cU|q-%PYTxmXqq7>t>{A_6+7dmw|64Y6Q4k
z^+8?Kgi;7Bg@w35SZ07sMu$hl+QgL^%{i}Tt+6ggulW${q6t5{yFTgr$AnU<C#l!#
z3a|%$DrcC98}UE03L5ao7w(N&&4SEd`wZ~7Efy?ml>iqe?N<T;v&e6DG<(YWr~9vR
zt{BuM5+h$5`?d7(xP5Vz?=_IMnJojAOtY#-VDBzXc*rI+w?p9O$$DSY?{(Ojj>PLi
za+;{#+=t3$bm)$j31%~2^lr~S{|31TeCE&~G=Rm-jxL?Qa%f05@uVKVR%|eC(nDUm
zGrmnGFf&p#qgOcmHr-_hv=LLTO}}0EA>iRWIaa)_3=tfcxUb%aCbF4R^2!cr+&Nn3
zch!D7^*Ed0uxv|LX?__sm`HOn=tJO@rg$QZmqgm04k)@_M@sZ<%U`cKw!C+UzRvu$
zYDX<q!juKic{vFxbt}ma8}#V-Ew%vdbz>1S1LjbX$Y-!)d8@;9Etk1>zbdvA{q*&G
z3P=6wD)98@+37toLvk#pAK;o~i1tv*@6TeJ>^-=bx*=OZxBDSuk`NiIpVt%}?Y(F1
zS8FMg;<;Db|0@`)G@rs;GphQ^bC?y!$q92WuluMOw;a<Tx!F7JLzGp*;+xG0)P+4Z
zQ{*MFR(}~%ZuTTES)a&ad@0|&yw-bfXbCcNeZM6kdgXaSCHKad`)<Umt;tlL42+@u
zNsW5yp{p_JKy7#P+`hGU$V(1qeL2_at?Ep|jL>c5Diy%>nhebf*-Y*Rmuu01nm;dG
z_E5sr@MGetM=CnEti(N|W|>ecNL%*c9xasv-E6PwD}_=sv-6G6iF$ZtNXejE+eyoN
z)<S0JOGrLd{%z1^Mc96B&M&#ivA0*#G9U$o#M@~GSKF}R9J(dUr-9F)Q(9(ZfjkS_
zQ=UsQEV~T><vzsQw~zzH0}6Dk7FTDisAo9Sxz7$;e!!dnE%Y^V9*KLFj2kPI%R#8r
z8#?m?JmcD&qn&NW8QoSX?`VA`cDmW}UVy?1U%m9N*lR|Lf~6ki9ZqFIipixGP^Sae
zjWY&+p8BQuSvR(WM*TY5YXa<E1mo9@5OG|aY4}FJlQxj3H?bCYqz1)qP&7#{PN(xo
zVs==M`G+K3$fxfSqoP`BX{753oG=Zesup+S5Ah<98@A7LUlvhRc6+v#KUbWWX8_wh
z${n2wc)11?n#?+qJorv-t~0x>vw~mGz4GZ$vb&6bCiyF7Sl7FG%RYeR@WxPY(uRCX
zwwQc9;hvYKT-kfg8#Af&O3spxV`EE3)3q1+cnx0jehvEua>n(a5@OwmHUHHUFv{(~
zby_p>9TC{@u+Bx)*V6o(H^mHF{*Q1(n`Lh!{~-Sd{fAOaFkg9}(<>bZ{*ZdJQUk(@
z(?Y&`D|&QfGs>ZB+{5lax1)0ckao>-gO~??@{d@?Q35Zj_6Ak<&kKL~Sw5+*@6-I!
z&3d*Kqxj_mvfN3=Yqez{pjFa1C;t77tQLr?=O?$B#n&@I6f+^~eygQnNK@#SUXuMg
zM;uo}8GWtVN~lg1=g-AV_#?MHVumznnmj_CA+5%gxfz=rSCyn<?I_QehXMvYH@$Y=
z_n@g7HL@5T!{@tOPFf`h5B%bVPen5fE-LPBCw|UD&4uq<(pMx-S|&`=ocIaQmO3$R
zCBxDKOcVL{2bzTf$SZ~)&On%s$#GrEKqND88%r}W`a-pOGBo8ShjmG-k$cVgMKAXT
zT7fv=={mGRrB~V1f8&RiRaDg`*6oP<?s4!J`%>}3NagmZWcl3VvUYcxmEm1|<L8Cb
zFQdJu+eG6E+L59eZwW6@ee|epp<k~yM!G2GbdBn@KUTak6qv>JdqaYSb1$@!52=={
zkD;t(hg-HEafBQrL90WgqHz@2%-0B%iiLp>qtv}!HgHMl_P8F=w+Z=exyVD^ABKsk
ze(r|byuHUa=UxN|pCM*ui(XeV+2)Xx2rYfI?8yKwSZWX8=eYBw`m$Nli(CGJMUXNz
zyqh*j9T6}K!AiSgrL)Lq&PldC#%`Jl^&D@1SC8Ao_P$-$3%o&KsJwK3c;9cuU{Zv%
zVVtO0$SnP^X)EyKr<fMqpmY~jlV^8Yn%HI*sy33~>$>T=vEEM{=c%6d{-Pc$U*1A~
zPN`|C_kC*QHYqC}JzNqegi%1$3jCn;;<MU1exbcKUM#c}){y+N)C-Fgs}Wjyn}E$r
zhOu}41pC6x?$HD#BFg-CFZ>*;EPJp@StvZjw&OWc{*gA5QcZWjErqO+lzyEv2Ct0x
zn;rW71YR0k6XM=dGj})HGJKRvpsXyBn~Bj!se&`+wYsoOnye=tvdKe~L6)mO1DX0;
z$lRG}?s%S_OzHeI4|yr<EhQIEsw`WBY_MN?Mmz0xlf+(U{9L;LVlYX0u{{XcIuQcO
zqr!Le=E=WkF%gqLJ$AD68E#2BvSa){eB5g1Hrct!Ryx1M3GXsG@ZVWOXM0cX@{f4f
zr60hti!)z(<1^&98Y|-+-wn)FP0<NE)jg~v?dsi28QMCClAP)4X7ucqe^{12Cu+v5
z#xY4uq0mq9!9}@8fF2b}tv>sDBfuosgsOzsz-O^2*w6e)p)3B#i7Fjl@O$FRoeG)7
kpF5Sy!T;t_a5-a?x?3B4s`UuB0mQ>oSJhG}Rk9BKKfs2~hyVZp

diff --git a/docs/_build/html/_static/minus.png b/docs/_build/html/_static/minus.png
deleted file mode 100644
index d96755fdaf8bb2214971e0db9c1fd3077d7c419d..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 90
zcmeAS@N?(olHy`uVBq!ia0vp^+#t*WBp7;*Yy1LIik>cxAr*|t7R?Mi>2?kWtu=nj
kDsEF_5m^0CR;1wuP-*O&G^0G}KYk!hp00i_>zopr08q^qX#fBK

diff --git a/docs/_build/html/_static/plus.png b/docs/_build/html/_static/plus.png
deleted file mode 100644
index 7107cec93a979b9a5f64843235a16651d563ce2d..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 90
zcmeAS@N?(olHy`uVBq!ia0vp^+#t*WBp7;*Yy1LIik>cxAr*|t7R?Mi>2?kWtu>-2
m3q%Vub%g%s<8sJhVPMczOq}xhg9DJoz~JfX=d#Wzp$Pyb1r*Kz

diff --git a/docs/_build/html/_static/pygments.css b/docs/_build/html/_static/pygments.css
deleted file mode 100644
index c2e07c71e..000000000
--- a/docs/_build/html/_static/pygments.css
+++ /dev/null
@@ -1,258 +0,0 @@
-.highlight pre { line-height: 125%; }
-.highlight td.linenos .normal { color: inherit; background-color: transparent; padding-left: 5px; padding-right: 5px; }
-.highlight span.linenos { color: inherit; background-color: transparent; padding-left: 5px; padding-right: 5px; }
-.highlight td.linenos .special { color: #000000; background-color: #ffffc0; padding-left: 5px; padding-right: 5px; }
-.highlight span.linenos.special { color: #000000; background-color: #ffffc0; padding-left: 5px; padding-right: 5px; }
-.highlight .hll { background-color: #ffffcc }
-.highlight { background: #f8f8f8; }
-.highlight .c { color: #8f5902; font-style: italic } /* Comment */
-.highlight .err { color: #a40000; border: 1px solid #ef2929 } /* Error */
-.highlight .g { color: #000000 } /* Generic */
-.highlight .k { color: #204a87; font-weight: bold } /* Keyword */
-.highlight .l { color: #000000 } /* Literal */
-.highlight .n { color: #000000 } /* Name */
-.highlight .o { color: #ce5c00; font-weight: bold } /* Operator */
-.highlight .x { color: #000000 } /* Other */
-.highlight .p { color: #000000; font-weight: bold } /* Punctuation */
-.highlight .ch { color: #8f5902; font-style: italic } /* Comment.Hashbang */
-.highlight .cm { color: #8f5902; font-style: italic } /* Comment.Multiline */
-.highlight .cp { color: #8f5902; font-style: italic } /* Comment.Preproc */
-.highlight .cpf { color: #8f5902; font-style: italic } /* Comment.PreprocFile */
-.highlight .c1 { color: #8f5902; font-style: italic } /* Comment.Single */
-.highlight .cs { color: #8f5902; font-style: italic } /* Comment.Special */
-.highlight .gd { color: #a40000 } /* Generic.Deleted */
-.highlight .ge { color: #000000; font-style: italic } /* Generic.Emph */
-.highlight .ges { color: #000000; font-weight: bold; font-style: italic } /* Generic.EmphStrong */
-.highlight .gr { color: #ef2929 } /* Generic.Error */
-.highlight .gh { color: #000080; font-weight: bold } /* Generic.Heading */
-.highlight .gi { color: #00A000 } /* Generic.Inserted */
-.highlight .go { color: #000000; font-style: italic } /* Generic.Output */
-.highlight .gp { color: #8f5902 } /* Generic.Prompt */
-.highlight .gs { color: #000000; font-weight: bold } /* Generic.Strong */
-.highlight .gu { color: #800080; font-weight: bold } /* Generic.Subheading */
-.highlight .gt { color: #a40000; font-weight: bold } /* Generic.Traceback */
-.highlight .kc { color: #204a87; font-weight: bold } /* Keyword.Constant */
-.highlight .kd { color: #204a87; font-weight: bold } /* Keyword.Declaration */
-.highlight .kn { color: #204a87; font-weight: bold } /* Keyword.Namespace */
-.highlight .kp { color: #204a87; font-weight: bold } /* Keyword.Pseudo */
-.highlight .kr { color: #204a87; font-weight: bold } /* Keyword.Reserved */
-.highlight .kt { color: #204a87; font-weight: bold } /* Keyword.Type */
-.highlight .ld { color: #000000 } /* Literal.Date */
-.highlight .m { color: #0000cf; font-weight: bold } /* Literal.Number */
-.highlight .s { color: #4e9a06 } /* Literal.String */
-.highlight .na { color: #c4a000 } /* Name.Attribute */
-.highlight .nb { color: #204a87 } /* Name.Builtin */
-.highlight .nc { color: #000000 } /* Name.Class */
-.highlight .no { color: #000000 } /* Name.Constant */
-.highlight .nd { color: #5c35cc; font-weight: bold } /* Name.Decorator */
-.highlight .ni { color: #ce5c00 } /* Name.Entity */
-.highlight .ne { color: #cc0000; font-weight: bold } /* Name.Exception */
-.highlight .nf { color: #000000 } /* Name.Function */
-.highlight .nl { color: #f57900 } /* Name.Label */
-.highlight .nn { color: #000000 } /* Name.Namespace */
-.highlight .nx { color: #000000 } /* Name.Other */
-.highlight .py { color: #000000 } /* Name.Property */
-.highlight .nt { color: #204a87; font-weight: bold } /* Name.Tag */
-.highlight .nv { color: #000000 } /* Name.Variable */
-.highlight .ow { color: #204a87; font-weight: bold } /* Operator.Word */
-.highlight .pm { color: #000000; font-weight: bold } /* Punctuation.Marker */
-.highlight .w { color: #f8f8f8 } /* Text.Whitespace */
-.highlight .mb { color: #0000cf; font-weight: bold } /* Literal.Number.Bin */
-.highlight .mf { color: #0000cf; font-weight: bold } /* Literal.Number.Float */
-.highlight .mh { color: #0000cf; font-weight: bold } /* Literal.Number.Hex */
-.highlight .mi { color: #0000cf; font-weight: bold } /* Literal.Number.Integer */
-.highlight .mo { color: #0000cf; font-weight: bold } /* Literal.Number.Oct */
-.highlight .sa { color: #4e9a06 } /* Literal.String.Affix */
-.highlight .sb { color: #4e9a06 } /* Literal.String.Backtick */
-.highlight .sc { color: #4e9a06 } /* Literal.String.Char */
-.highlight .dl { color: #4e9a06 } /* Literal.String.Delimiter */
-.highlight .sd { color: #8f5902; font-style: italic } /* Literal.String.Doc */
-.highlight .s2 { color: #4e9a06 } /* Literal.String.Double */
-.highlight .se { color: #4e9a06 } /* Literal.String.Escape */
-.highlight .sh { color: #4e9a06 } /* Literal.String.Heredoc */
-.highlight .si { color: #4e9a06 } /* Literal.String.Interpol */
-.highlight .sx { color: #4e9a06 } /* Literal.String.Other */
-.highlight .sr { color: #4e9a06 } /* Literal.String.Regex */
-.highlight .s1 { color: #4e9a06 } /* Literal.String.Single */
-.highlight .ss { color: #4e9a06 } /* Literal.String.Symbol */
-.highlight .bp { color: #3465a4 } /* Name.Builtin.Pseudo */
-.highlight .fm { color: #000000 } /* Name.Function.Magic */
-.highlight .vc { color: #000000 } /* Name.Variable.Class */
-.highlight .vg { color: #000000 } /* Name.Variable.Global */
-.highlight .vi { color: #000000 } /* Name.Variable.Instance */
-.highlight .vm { color: #000000 } /* Name.Variable.Magic */
-.highlight .il { color: #0000cf; font-weight: bold } /* Literal.Number.Integer.Long */
-@media not print {
-body[data-theme="dark"] .highlight pre { line-height: 125%; }
-body[data-theme="dark"] .highlight td.linenos .normal { color: #aaaaaa; background-color: transparent; padding-left: 5px; padding-right: 5px; }
-body[data-theme="dark"] .highlight span.linenos { color: #aaaaaa; background-color: transparent; padding-left: 5px; padding-right: 5px; }
-body[data-theme="dark"] .highlight td.linenos .special { color: #000000; background-color: #ffffc0; padding-left: 5px; padding-right: 5px; }
-body[data-theme="dark"] .highlight span.linenos.special { color: #000000; background-color: #ffffc0; padding-left: 5px; padding-right: 5px; }
-body[data-theme="dark"] .highlight .hll { background-color: #404040 }
-body[data-theme="dark"] .highlight { background: #202020; color: #d0d0d0 }
-body[data-theme="dark"] .highlight .c { color: #ababab; font-style: italic } /* Comment */
-body[data-theme="dark"] .highlight .err { color: #a61717; background-color: #e3d2d2 } /* Error */
-body[data-theme="dark"] .highlight .esc { color: #d0d0d0 } /* Escape */
-body[data-theme="dark"] .highlight .g { color: #d0d0d0 } /* Generic */
-body[data-theme="dark"] .highlight .k { color: #6ebf26; font-weight: bold } /* Keyword */
-body[data-theme="dark"] .highlight .l { color: #d0d0d0 } /* Literal */
-body[data-theme="dark"] .highlight .n { color: #d0d0d0 } /* Name */
-body[data-theme="dark"] .highlight .o { color: #d0d0d0 } /* Operator */
-body[data-theme="dark"] .highlight .x { color: #d0d0d0 } /* Other */
-body[data-theme="dark"] .highlight .p { color: #d0d0d0 } /* Punctuation */
-body[data-theme="dark"] .highlight .ch { color: #ababab; font-style: italic } /* Comment.Hashbang */
-body[data-theme="dark"] .highlight .cm { color: #ababab; font-style: italic } /* Comment.Multiline */
-body[data-theme="dark"] .highlight .cp { color: #ff3a3a; font-weight: bold } /* Comment.Preproc */
-body[data-theme="dark"] .highlight .cpf { color: #ababab; font-style: italic } /* Comment.PreprocFile */
-body[data-theme="dark"] .highlight .c1 { color: #ababab; font-style: italic } /* Comment.Single */
-body[data-theme="dark"] .highlight .cs { color: #e50808; font-weight: bold; background-color: #520000 } /* Comment.Special */
-body[data-theme="dark"] .highlight .gd { color: #d22323 } /* Generic.Deleted */
-body[data-theme="dark"] .highlight .ge { color: #d0d0d0; font-style: italic } /* Generic.Emph */
-body[data-theme="dark"] .highlight .ges { color: #d0d0d0; font-weight: bold; font-style: italic } /* Generic.EmphStrong */
-body[data-theme="dark"] .highlight .gr { color: #d22323 } /* Generic.Error */
-body[data-theme="dark"] .highlight .gh { color: #ffffff; font-weight: bold } /* Generic.Heading */
-body[data-theme="dark"] .highlight .gi { color: #589819 } /* Generic.Inserted */
-body[data-theme="dark"] .highlight .go { color: #cccccc } /* Generic.Output */
-body[data-theme="dark"] .highlight .gp { color: #aaaaaa } /* Generic.Prompt */
-body[data-theme="dark"] .highlight .gs { color: #d0d0d0; font-weight: bold } /* Generic.Strong */
-body[data-theme="dark"] .highlight .gu { color: #ffffff; text-decoration: underline } /* Generic.Subheading */
-body[data-theme="dark"] .highlight .gt { color: #d22323 } /* Generic.Traceback */
-body[data-theme="dark"] .highlight .kc { color: #6ebf26; font-weight: bold } /* Keyword.Constant */
-body[data-theme="dark"] .highlight .kd { color: #6ebf26; font-weight: bold } /* Keyword.Declaration */
-body[data-theme="dark"] .highlight .kn { color: #6ebf26; font-weight: bold } /* Keyword.Namespace */
-body[data-theme="dark"] .highlight .kp { color: #6ebf26 } /* Keyword.Pseudo */
-body[data-theme="dark"] .highlight .kr { color: #6ebf26; font-weight: bold } /* Keyword.Reserved */
-body[data-theme="dark"] .highlight .kt { color: #6ebf26; font-weight: bold } /* Keyword.Type */
-body[data-theme="dark"] .highlight .ld { color: #d0d0d0 } /* Literal.Date */
-body[data-theme="dark"] .highlight .m { color: #51b2fd } /* Literal.Number */
-body[data-theme="dark"] .highlight .s { color: #ed9d13 } /* Literal.String */
-body[data-theme="dark"] .highlight .na { color: #bbbbbb } /* Name.Attribute */
-body[data-theme="dark"] .highlight .nb { color: #2fbccd } /* Name.Builtin */
-body[data-theme="dark"] .highlight .nc { color: #71adff; text-decoration: underline } /* Name.Class */
-body[data-theme="dark"] .highlight .no { color: #40ffff } /* Name.Constant */
-body[data-theme="dark"] .highlight .nd { color: #ffa500 } /* Name.Decorator */
-body[data-theme="dark"] .highlight .ni { color: #d0d0d0 } /* Name.Entity */
-body[data-theme="dark"] .highlight .ne { color: #bbbbbb } /* Name.Exception */
-body[data-theme="dark"] .highlight .nf { color: #71adff } /* Name.Function */
-body[data-theme="dark"] .highlight .nl { color: #d0d0d0 } /* Name.Label */
-body[data-theme="dark"] .highlight .nn { color: #71adff; text-decoration: underline } /* Name.Namespace */
-body[data-theme="dark"] .highlight .nx { color: #d0d0d0 } /* Name.Other */
-body[data-theme="dark"] .highlight .py { color: #d0d0d0 } /* Name.Property */
-body[data-theme="dark"] .highlight .nt { color: #6ebf26; font-weight: bold } /* Name.Tag */
-body[data-theme="dark"] .highlight .nv { color: #40ffff } /* Name.Variable */
-body[data-theme="dark"] .highlight .ow { color: #6ebf26; font-weight: bold } /* Operator.Word */
-body[data-theme="dark"] .highlight .pm { color: #d0d0d0 } /* Punctuation.Marker */
-body[data-theme="dark"] .highlight .w { color: #666666 } /* Text.Whitespace */
-body[data-theme="dark"] .highlight .mb { color: #51b2fd } /* Literal.Number.Bin */
-body[data-theme="dark"] .highlight .mf { color: #51b2fd } /* Literal.Number.Float */
-body[data-theme="dark"] .highlight .mh { color: #51b2fd } /* Literal.Number.Hex */
-body[data-theme="dark"] .highlight .mi { color: #51b2fd } /* Literal.Number.Integer */
-body[data-theme="dark"] .highlight .mo { color: #51b2fd } /* Literal.Number.Oct */
-body[data-theme="dark"] .highlight .sa { color: #ed9d13 } /* Literal.String.Affix */
-body[data-theme="dark"] .highlight .sb { color: #ed9d13 } /* Literal.String.Backtick */
-body[data-theme="dark"] .highlight .sc { color: #ed9d13 } /* Literal.String.Char */
-body[data-theme="dark"] .highlight .dl { color: #ed9d13 } /* Literal.String.Delimiter */
-body[data-theme="dark"] .highlight .sd { color: #ed9d13 } /* Literal.String.Doc */
-body[data-theme="dark"] .highlight .s2 { color: #ed9d13 } /* Literal.String.Double */
-body[data-theme="dark"] .highlight .se { color: #ed9d13 } /* Literal.String.Escape */
-body[data-theme="dark"] .highlight .sh { color: #ed9d13 } /* Literal.String.Heredoc */
-body[data-theme="dark"] .highlight .si { color: #ed9d13 } /* Literal.String.Interpol */
-body[data-theme="dark"] .highlight .sx { color: #ffa500 } /* Literal.String.Other */
-body[data-theme="dark"] .highlight .sr { color: #ed9d13 } /* Literal.String.Regex */
-body[data-theme="dark"] .highlight .s1 { color: #ed9d13 } /* Literal.String.Single */
-body[data-theme="dark"] .highlight .ss { color: #ed9d13 } /* Literal.String.Symbol */
-body[data-theme="dark"] .highlight .bp { color: #2fbccd } /* Name.Builtin.Pseudo */
-body[data-theme="dark"] .highlight .fm { color: #71adff } /* Name.Function.Magic */
-body[data-theme="dark"] .highlight .vc { color: #40ffff } /* Name.Variable.Class */
-body[data-theme="dark"] .highlight .vg { color: #40ffff } /* Name.Variable.Global */
-body[data-theme="dark"] .highlight .vi { color: #40ffff } /* Name.Variable.Instance */
-body[data-theme="dark"] .highlight .vm { color: #40ffff } /* Name.Variable.Magic */
-body[data-theme="dark"] .highlight .il { color: #51b2fd } /* Literal.Number.Integer.Long */
-@media (prefers-color-scheme: dark) {
-body:not([data-theme="light"]) .highlight pre { line-height: 125%; }
-body:not([data-theme="light"]) .highlight td.linenos .normal { color: #aaaaaa; background-color: transparent; padding-left: 5px; padding-right: 5px; }
-body:not([data-theme="light"]) .highlight span.linenos { color: #aaaaaa; background-color: transparent; padding-left: 5px; padding-right: 5px; }
-body:not([data-theme="light"]) .highlight td.linenos .special { color: #000000; background-color: #ffffc0; padding-left: 5px; padding-right: 5px; }
-body:not([data-theme="light"]) .highlight span.linenos.special { color: #000000; background-color: #ffffc0; padding-left: 5px; padding-right: 5px; }
-body:not([data-theme="light"]) .highlight .hll { background-color: #404040 }
-body:not([data-theme="light"]) .highlight { background: #202020; color: #d0d0d0 }
-body:not([data-theme="light"]) .highlight .c { color: #ababab; font-style: italic } /* Comment */
-body:not([data-theme="light"]) .highlight .err { color: #a61717; background-color: #e3d2d2 } /* Error */
-body:not([data-theme="light"]) .highlight .esc { color: #d0d0d0 } /* Escape */
-body:not([data-theme="light"]) .highlight .g { color: #d0d0d0 } /* Generic */
-body:not([data-theme="light"]) .highlight .k { color: #6ebf26; font-weight: bold } /* Keyword */
-body:not([data-theme="light"]) .highlight .l { color: #d0d0d0 } /* Literal */
-body:not([data-theme="light"]) .highlight .n { color: #d0d0d0 } /* Name */
-body:not([data-theme="light"]) .highlight .o { color: #d0d0d0 } /* Operator */
-body:not([data-theme="light"]) .highlight .x { color: #d0d0d0 } /* Other */
-body:not([data-theme="light"]) .highlight .p { color: #d0d0d0 } /* Punctuation */
-body:not([data-theme="light"]) .highlight .ch { color: #ababab; font-style: italic } /* Comment.Hashbang */
-body:not([data-theme="light"]) .highlight .cm { color: #ababab; font-style: italic } /* Comment.Multiline */
-body:not([data-theme="light"]) .highlight .cp { color: #ff3a3a; font-weight: bold } /* Comment.Preproc */
-body:not([data-theme="light"]) .highlight .cpf { color: #ababab; font-style: italic } /* Comment.PreprocFile */
-body:not([data-theme="light"]) .highlight .c1 { color: #ababab; font-style: italic } /* Comment.Single */
-body:not([data-theme="light"]) .highlight .cs { color: #e50808; font-weight: bold; background-color: #520000 } /* Comment.Special */
-body:not([data-theme="light"]) .highlight .gd { color: #d22323 } /* Generic.Deleted */
-body:not([data-theme="light"]) .highlight .ge { color: #d0d0d0; font-style: italic } /* Generic.Emph */
-body:not([data-theme="light"]) .highlight .ges { color: #d0d0d0; font-weight: bold; font-style: italic } /* Generic.EmphStrong */
-body:not([data-theme="light"]) .highlight .gr { color: #d22323 } /* Generic.Error */
-body:not([data-theme="light"]) .highlight .gh { color: #ffffff; font-weight: bold } /* Generic.Heading */
-body:not([data-theme="light"]) .highlight .gi { color: #589819 } /* Generic.Inserted */
-body:not([data-theme="light"]) .highlight .go { color: #cccccc } /* Generic.Output */
-body:not([data-theme="light"]) .highlight .gp { color: #aaaaaa } /* Generic.Prompt */
-body:not([data-theme="light"]) .highlight .gs { color: #d0d0d0; font-weight: bold } /* Generic.Strong */
-body:not([data-theme="light"]) .highlight .gu { color: #ffffff; text-decoration: underline } /* Generic.Subheading */
-body:not([data-theme="light"]) .highlight .gt { color: #d22323 } /* Generic.Traceback */
-body:not([data-theme="light"]) .highlight .kc { color: #6ebf26; font-weight: bold } /* Keyword.Constant */
-body:not([data-theme="light"]) .highlight .kd { color: #6ebf26; font-weight: bold } /* Keyword.Declaration */
-body:not([data-theme="light"]) .highlight .kn { color: #6ebf26; font-weight: bold } /* Keyword.Namespace */
-body:not([data-theme="light"]) .highlight .kp { color: #6ebf26 } /* Keyword.Pseudo */
-body:not([data-theme="light"]) .highlight .kr { color: #6ebf26; font-weight: bold } /* Keyword.Reserved */
-body:not([data-theme="light"]) .highlight .kt { color: #6ebf26; font-weight: bold } /* Keyword.Type */
-body:not([data-theme="light"]) .highlight .ld { color: #d0d0d0 } /* Literal.Date */
-body:not([data-theme="light"]) .highlight .m { color: #51b2fd } /* Literal.Number */
-body:not([data-theme="light"]) .highlight .s { color: #ed9d13 } /* Literal.String */
-body:not([data-theme="light"]) .highlight .na { color: #bbbbbb } /* Name.Attribute */
-body:not([data-theme="light"]) .highlight .nb { color: #2fbccd } /* Name.Builtin */
-body:not([data-theme="light"]) .highlight .nc { color: #71adff; text-decoration: underline } /* Name.Class */
-body:not([data-theme="light"]) .highlight .no { color: #40ffff } /* Name.Constant */
-body:not([data-theme="light"]) .highlight .nd { color: #ffa500 } /* Name.Decorator */
-body:not([data-theme="light"]) .highlight .ni { color: #d0d0d0 } /* Name.Entity */
-body:not([data-theme="light"]) .highlight .ne { color: #bbbbbb } /* Name.Exception */
-body:not([data-theme="light"]) .highlight .nf { color: #71adff } /* Name.Function */
-body:not([data-theme="light"]) .highlight .nl { color: #d0d0d0 } /* Name.Label */
-body:not([data-theme="light"]) .highlight .nn { color: #71adff; text-decoration: underline } /* Name.Namespace */
-body:not([data-theme="light"]) .highlight .nx { color: #d0d0d0 } /* Name.Other */
-body:not([data-theme="light"]) .highlight .py { color: #d0d0d0 } /* Name.Property */
-body:not([data-theme="light"]) .highlight .nt { color: #6ebf26; font-weight: bold } /* Name.Tag */
-body:not([data-theme="light"]) .highlight .nv { color: #40ffff } /* Name.Variable */
-body:not([data-theme="light"]) .highlight .ow { color: #6ebf26; font-weight: bold } /* Operator.Word */
-body:not([data-theme="light"]) .highlight .pm { color: #d0d0d0 } /* Punctuation.Marker */
-body:not([data-theme="light"]) .highlight .w { color: #666666 } /* Text.Whitespace */
-body:not([data-theme="light"]) .highlight .mb { color: #51b2fd } /* Literal.Number.Bin */
-body:not([data-theme="light"]) .highlight .mf { color: #51b2fd } /* Literal.Number.Float */
-body:not([data-theme="light"]) .highlight .mh { color: #51b2fd } /* Literal.Number.Hex */
-body:not([data-theme="light"]) .highlight .mi { color: #51b2fd } /* Literal.Number.Integer */
-body:not([data-theme="light"]) .highlight .mo { color: #51b2fd } /* Literal.Number.Oct */
-body:not([data-theme="light"]) .highlight .sa { color: #ed9d13 } /* Literal.String.Affix */
-body:not([data-theme="light"]) .highlight .sb { color: #ed9d13 } /* Literal.String.Backtick */
-body:not([data-theme="light"]) .highlight .sc { color: #ed9d13 } /* Literal.String.Char */
-body:not([data-theme="light"]) .highlight .dl { color: #ed9d13 } /* Literal.String.Delimiter */
-body:not([data-theme="light"]) .highlight .sd { color: #ed9d13 } /* Literal.String.Doc */
-body:not([data-theme="light"]) .highlight .s2 { color: #ed9d13 } /* Literal.String.Double */
-body:not([data-theme="light"]) .highlight .se { color: #ed9d13 } /* Literal.String.Escape */
-body:not([data-theme="light"]) .highlight .sh { color: #ed9d13 } /* Literal.String.Heredoc */
-body:not([data-theme="light"]) .highlight .si { color: #ed9d13 } /* Literal.String.Interpol */
-body:not([data-theme="light"]) .highlight .sx { color: #ffa500 } /* Literal.String.Other */
-body:not([data-theme="light"]) .highlight .sr { color: #ed9d13 } /* Literal.String.Regex */
-body:not([data-theme="light"]) .highlight .s1 { color: #ed9d13 } /* Literal.String.Single */
-body:not([data-theme="light"]) .highlight .ss { color: #ed9d13 } /* Literal.String.Symbol */
-body:not([data-theme="light"]) .highlight .bp { color: #2fbccd } /* Name.Builtin.Pseudo */
-body:not([data-theme="light"]) .highlight .fm { color: #71adff } /* Name.Function.Magic */
-body:not([data-theme="light"]) .highlight .vc { color: #40ffff } /* Name.Variable.Class */
-body:not([data-theme="light"]) .highlight .vg { color: #40ffff } /* Name.Variable.Global */
-body:not([data-theme="light"]) .highlight .vi { color: #40ffff } /* Name.Variable.Instance */
-body:not([data-theme="light"]) .highlight .vm { color: #40ffff } /* Name.Variable.Magic */
-body:not([data-theme="light"]) .highlight .il { color: #51b2fd } /* Literal.Number.Integer.Long */
-}
-}
\ No newline at end of file
diff --git a/docs/_build/html/_static/scripts/furo-extensions.js b/docs/_build/html/_static/scripts/furo-extensions.js
deleted file mode 100644
index e69de29bb..000000000
diff --git a/docs/_build/html/_static/scripts/furo.js b/docs/_build/html/_static/scripts/furo.js
deleted file mode 100644
index 32e7c05be..000000000
--- a/docs/_build/html/_static/scripts/furo.js
+++ /dev/null
@@ -1,3 +0,0 @@
-/*! For license information please see furo.js.LICENSE.txt */
-(()=>{var t={212:function(t,e,n){var o,r;r=void 0!==n.g?n.g:"undefined"!=typeof window?window:this,o=function(){return function(t){"use strict";var e={navClass:"active",contentClass:"active",nested:!1,nestedClass:"active",offset:0,reflow:!1,events:!0},n=function(t,e,n){if(n.settings.events){var o=new CustomEvent(t,{bubbles:!0,cancelable:!0,detail:n});e.dispatchEvent(o)}},o=function(t){var e=0;if(t.offsetParent)for(;t;)e+=t.offsetTop,t=t.offsetParent;return e>=0?e:0},r=function(t){t&&t.sort((function(t,e){return o(t.content)<o(e.content)?-1:1}))},c=function(e,n,o){var r=e.getBoundingClientRect(),c=function(t){return"function"==typeof t.offset?parseFloat(t.offset()):parseFloat(t.offset)}(n);return o?parseInt(r.bottom,10)<(t.innerHeight||document.documentElement.clientHeight):parseInt(r.top,10)<=c},s=function(){return Math.ceil(t.innerHeight+t.pageYOffset)>=Math.max(document.body.scrollHeight,document.documentElement.scrollHeight,document.body.offsetHeight,document.documentElement.offsetHeight,document.body.clientHeight,document.documentElement.clientHeight)},l=function(t,e){var n=t[t.length-1];if(function(t,e){return!(!s()||!c(t.content,e,!0))}(n,e))return n;for(var o=t.length-1;o>=0;o--)if(c(t[o].content,e))return t[o]},a=function(t,e){if(e.nested&&t.parentNode){var n=t.parentNode.closest("li");n&&(n.classList.remove(e.nestedClass),a(n,e))}},i=function(t,e){if(t){var o=t.nav.closest("li");o&&(o.classList.remove(e.navClass),t.content.classList.remove(e.contentClass),a(o,e),n("gumshoeDeactivate",o,{link:t.nav,content:t.content,settings:e}))}},u=function(t,e){if(e.nested){var n=t.parentNode.closest("li");n&&(n.classList.add(e.nestedClass),u(n,e))}};return function(o,c){var s,a,d,f,m,v={setup:function(){s=document.querySelectorAll(o),a=[],Array.prototype.forEach.call(s,(function(t){var e=document.getElementById(decodeURIComponent(t.hash.substr(1)));e&&a.push({nav:t,content:e})})),r(a)},detect:function(){var t=l(a,m);t?d&&t.content===d.content||(i(d,m),function(t,e){if(t){var o=t.nav.closest("li");o&&(o.classList.add(e.navClass),t.content.classList.add(e.contentClass),u(o,e),n("gumshoeActivate",o,{link:t.nav,content:t.content,settings:e}))}}(t,m),d=t):d&&(i(d,m),d=null)}},h=function(e){f&&t.cancelAnimationFrame(f),f=t.requestAnimationFrame(v.detect)},g=function(e){f&&t.cancelAnimationFrame(f),f=t.requestAnimationFrame((function(){r(a),v.detect()}))};return v.destroy=function(){d&&i(d,m),t.removeEventListener("scroll",h,!1),m.reflow&&t.removeEventListener("resize",g,!1),a=null,s=null,d=null,f=null,m=null},m=function(){var t={};return Array.prototype.forEach.call(arguments,(function(e){for(var n in e){if(!e.hasOwnProperty(n))return;t[n]=e[n]}})),t}(e,c||{}),v.setup(),v.detect(),t.addEventListener("scroll",h,!1),m.reflow&&t.addEventListener("resize",g,!1),v}}(r)}.apply(e,[]),void 0===o||(t.exports=o)}},e={};function n(o){var r=e[o];if(void 0!==r)return r.exports;var c=e[o]={exports:{}};return t[o].call(c.exports,c,c.exports,n),c.exports}n.n=t=>{var e=t&&t.__esModule?()=>t.default:()=>t;return n.d(e,{a:e}),e},n.d=(t,e)=>{for(var o in e)n.o(e,o)&&!n.o(t,o)&&Object.defineProperty(t,o,{enumerable:!0,get:e[o]})},n.g=function(){if("object"==typeof globalThis)return globalThis;try{return this||new Function("return this")()}catch(t){if("object"==typeof window)return window}}(),n.o=(t,e)=>Object.prototype.hasOwnProperty.call(t,e),(()=>{"use strict";var t=n(212),e=n.n(t),o=null,r=null,c=window.pageYOffset||document.documentElement.scrollTop;const s=64;function l(){const t=localStorage.getItem("theme")||"auto";var e;"light"!==(e=window.matchMedia("(prefers-color-scheme: dark)").matches?"auto"===t?"light":"light"==t?"dark":"auto":"auto"===t?"dark":"dark"==t?"light":"auto")&&"dark"!==e&&"auto"!==e&&(console.error(`Got invalid theme mode: ${e}. Resetting to auto.`),e="auto"),document.body.dataset.theme=e,localStorage.setItem("theme",e),console.log(`Changed to ${e} mode.`)}function a(){!function(){const t=document.getElementsByClassName("theme-toggle");Array.from(t).forEach((t=>{t.addEventListener("click",l)}))}(),function(){let t=0,e=!1;window.addEventListener("scroll",(function(n){t=window.scrollY,e||(window.requestAnimationFrame((function(){var n;n=t,0==Math.floor(r.getBoundingClientRect().top)?r.classList.add("scrolled"):r.classList.remove("scrolled"),function(t){t<s?document.documentElement.classList.remove("show-back-to-top"):t<c?document.documentElement.classList.add("show-back-to-top"):t>c&&document.documentElement.classList.remove("show-back-to-top"),c=t}(n),function(t){null!==o&&(0==t?o.scrollTo(0,0):Math.ceil(t)>=Math.floor(document.documentElement.scrollHeight-window.innerHeight)?o.scrollTo(0,o.scrollHeight):document.querySelector(".scroll-current"))}(n),e=!1})),e=!0)})),window.scroll()}(),null!==o&&new(e())(".toc-tree a",{reflow:!0,recursive:!0,navClass:"scroll-current",offset:()=>{let t=parseFloat(getComputedStyle(document.documentElement).fontSize);return r.getBoundingClientRect().height+.5*t+1}})}document.addEventListener("DOMContentLoaded",(function(){document.body.parentNode.classList.remove("no-js"),r=document.querySelector("header"),o=document.querySelector(".toc-scroll"),a()}))})()})();
-//# sourceMappingURL=furo.js.map
\ No newline at end of file
diff --git a/docs/_build/html/_static/scripts/furo.js.LICENSE.txt b/docs/_build/html/_static/scripts/furo.js.LICENSE.txt
deleted file mode 100644
index 1632189c7..000000000
--- a/docs/_build/html/_static/scripts/furo.js.LICENSE.txt
+++ /dev/null
@@ -1,7 +0,0 @@
-/*!
- * gumshoejs v5.1.2 (patched by @pradyunsg)
- * A simple, framework-agnostic scrollspy script.
- * (c) 2019 Chris Ferdinandi
- * MIT License
- * http://github.com/cferdinandi/gumshoe
- */
diff --git a/docs/_build/html/_static/scripts/furo.js.map b/docs/_build/html/_static/scripts/furo.js.map
deleted file mode 100644
index 7b7ddb113..000000000
--- a/docs/_build/html/_static/scripts/furo.js.map
+++ /dev/null
@@ -1 +0,0 @@
-{"version":3,"file":"scripts/furo.js","mappings":";iCAAA,MAQWA,SAWS,IAAX,EAAAC,EACH,EAAAA,EACkB,oBAAXC,OACPA,OACAC,KAbS,EAAF,WACP,OAaJ,SAAUD,GACR,aAMA,IAAIE,EAAW,CAEbC,SAAU,SACVC,aAAc,SAGdC,QAAQ,EACRC,YAAa,SAGbC,OAAQ,EACRC,QAAQ,EAGRC,QAAQ,GA6BNC,EAAY,SAAUC,EAAMC,EAAMC,GAEpC,GAAKA,EAAOC,SAASL,OAArB,CAGA,IAAIM,EAAQ,IAAIC,YAAYL,EAAM,CAChCM,SAAS,EACTC,YAAY,EACZL,OAAQA,IAIVD,EAAKO,cAAcJ,EAVgB,CAWrC,EAOIK,EAAe,SAAUR,GAC3B,IAAIS,EAAW,EACf,GAAIT,EAAKU,aACP,KAAOV,GACLS,GAAYT,EAAKW,UACjBX,EAAOA,EAAKU,aAGhB,OAAOD,GAAY,EAAIA,EAAW,CACpC,EAMIG,EAAe,SAAUC,GACvBA,GACFA,EAASC,MAAK,SAAUC,EAAOC,GAG7B,OAFcR,EAAaO,EAAME,SACnBT,EAAaQ,EAAMC,UACF,EACxB,CACT,GAEJ,EAwCIC,EAAW,SAAUlB,EAAME,EAAUiB,GACvC,IAAIC,EAASpB,EAAKqB,wBACd1B,EAnCU,SAAUO,GAExB,MAA+B,mBAApBA,EAASP,OACX2B,WAAWpB,EAASP,UAItB2B,WAAWpB,EAASP,OAC7B,CA2Be4B,CAAUrB,GACvB,OAAIiB,EAEAK,SAASJ,EAAOD,OAAQ,KACvB/B,EAAOqC,aAAeC,SAASC,gBAAgBC,cAG7CJ,SAASJ,EAAOS,IAAK,KAAOlC,CACrC,EAMImC,EAAa,WACf,OACEC,KAAKC,KAAK5C,EAAOqC,YAAcrC,EAAO6C,cAnCjCF,KAAKG,IACVR,SAASS,KAAKC,aACdV,SAASC,gBAAgBS,aACzBV,SAASS,KAAKE,aACdX,SAASC,gBAAgBU,aACzBX,SAASS,KAAKP,aACdF,SAASC,gBAAgBC,aAkC7B,EAmBIU,EAAY,SAAUzB,EAAUX,GAClC,IAAIqC,EAAO1B,EAASA,EAAS2B,OAAS,GACtC,GAbgB,SAAUC,EAAMvC,GAChC,SAAI4B,MAAgBZ,EAASuB,EAAKxB,QAASf,GAAU,GAEvD,CAUMwC,CAAYH,EAAMrC,GAAW,OAAOqC,EACxC,IAAK,IAAII,EAAI9B,EAAS2B,OAAS,EAAGG,GAAK,EAAGA,IACxC,GAAIzB,EAASL,EAAS8B,GAAG1B,QAASf,GAAW,OAAOW,EAAS8B,EAEjE,EAOIC,EAAmB,SAAUC,EAAK3C,GAEpC,GAAKA,EAAST,QAAWoD,EAAIC,WAA7B,CAGA,IAAIC,EAAKF,EAAIC,WAAWE,QAAQ,MAC3BD,IAGLA,EAAGE,UAAUC,OAAOhD,EAASR,aAG7BkD,EAAiBG,EAAI7C,GAV0B,CAWjD,EAOIiD,EAAa,SAAUC,EAAOlD,GAEhC,GAAKkD,EAAL,CAGA,IAAIL,EAAKK,EAAMP,IAAIG,QAAQ,MACtBD,IAGLA,EAAGE,UAAUC,OAAOhD,EAASX,UAC7B6D,EAAMnC,QAAQgC,UAAUC,OAAOhD,EAASV,cAGxCoD,EAAiBG,EAAI7C,GAGrBJ,EAAU,oBAAqBiD,EAAI,CACjCM,KAAMD,EAAMP,IACZ5B,QAASmC,EAAMnC,QACff,SAAUA,IAjBM,CAmBpB,EAOIoD,EAAiB,SAAUT,EAAK3C,GAElC,GAAKA,EAAST,OAAd,CAGA,IAAIsD,EAAKF,EAAIC,WAAWE,QAAQ,MAC3BD,IAGLA,EAAGE,UAAUM,IAAIrD,EAASR,aAG1B4D,EAAeP,EAAI7C,GAVS,CAW9B,EA6LA,OA1JkB,SAAUsD,EAAUC,GAKpC,IACIC,EAAU7C,EAAU8C,EAASC,EAAS1D,EADtC2D,EAAa,CAUjBA,MAAmB,WAEjBH,EAAWhC,SAASoC,iBAAiBN,GAGrC3C,EAAW,GAGXkD,MAAMC,UAAUC,QAAQC,KAAKR,GAAU,SAAUjB,GAE/C,IAAIxB,EAAUS,SAASyC,eACrBC,mBAAmB3B,EAAK4B,KAAKC,OAAO,KAEjCrD,GAGLJ,EAAS0D,KAAK,CACZ1B,IAAKJ,EACLxB,QAASA,GAEb,IAGAL,EAAaC,EACf,EAKAgD,OAAoB,WAElB,IAAIW,EAASlC,EAAUzB,EAAUX,GAG5BsE,EASDb,GAAWa,EAAOvD,UAAY0C,EAAQ1C,UAG1CkC,EAAWQ,EAASzD,GAzFT,SAAUkD,EAAOlD,GAE9B,GAAKkD,EAAL,CAGA,IAAIL,EAAKK,EAAMP,IAAIG,QAAQ,MACtBD,IAGLA,EAAGE,UAAUM,IAAIrD,EAASX,UAC1B6D,EAAMnC,QAAQgC,UAAUM,IAAIrD,EAASV,cAGrC8D,EAAeP,EAAI7C,GAGnBJ,EAAU,kBAAmBiD,EAAI,CAC/BM,KAAMD,EAAMP,IACZ5B,QAASmC,EAAMnC,QACff,SAAUA,IAjBM,CAmBpB,CAqEIuE,CAASD,EAAQtE,GAGjByD,EAAUa,GAfJb,IACFR,EAAWQ,EAASzD,GACpByD,EAAU,KAchB,GAMIe,EAAgB,SAAUvE,GAExByD,GACFxE,EAAOuF,qBAAqBf,GAI9BA,EAAUxE,EAAOwF,sBAAsBf,EAAWgB,OACpD,EAMIC,EAAgB,SAAU3E,GAExByD,GACFxE,EAAOuF,qBAAqBf,GAI9BA,EAAUxE,EAAOwF,uBAAsB,WACrChE,EAAaC,GACbgD,EAAWgB,QACb,GACF,EAkDA,OA7CAhB,EAAWkB,QAAU,WAEfpB,GACFR,EAAWQ,EAASzD,GAItBd,EAAO4F,oBAAoB,SAAUN,GAAe,GAChDxE,EAASN,QACXR,EAAO4F,oBAAoB,SAAUF,GAAe,GAItDjE,EAAW,KACX6C,EAAW,KACXC,EAAU,KACVC,EAAU,KACV1D,EAAW,IACb,EAOEA,EA3XS,WACX,IAAI+E,EAAS,CAAC,EAOd,OANAlB,MAAMC,UAAUC,QAAQC,KAAKgB,WAAW,SAAUC,GAChD,IAAK,IAAIC,KAAOD,EAAK,CACnB,IAAKA,EAAIE,eAAeD,GAAM,OAC9BH,EAAOG,GAAOD,EAAIC,EACpB,CACF,IACOH,CACT,CAkXeK,CAAOhG,EAAUmE,GAAW,CAAC,GAGxCI,EAAW0B,QAGX1B,EAAWgB,SAGXzF,EAAOoG,iBAAiB,SAAUd,GAAe,GAC7CxE,EAASN,QACXR,EAAOoG,iBAAiB,SAAUV,GAAe,GAS9CjB,CACT,CAOF,CArcW4B,CAAQvG,EAChB,UAFM,SAEN,uBCXDwG,EAA2B,CAAC,EAGhC,SAASC,EAAoBC,GAE5B,IAAIC,EAAeH,EAAyBE,GAC5C,QAAqBE,IAAjBD,EACH,OAAOA,EAAaE,QAGrB,IAAIC,EAASN,EAAyBE,GAAY,CAGjDG,QAAS,CAAC,GAOX,OAHAE,EAAoBL,GAAU1B,KAAK8B,EAAOD,QAASC,EAAQA,EAAOD,QAASJ,GAGpEK,EAAOD,OACf,CCrBAJ,EAAoBO,EAAKF,IACxB,IAAIG,EAASH,GAAUA,EAAOI,WAC7B,IAAOJ,EAAiB,QACxB,IAAM,EAEP,OADAL,EAAoBU,EAAEF,EAAQ,CAAEG,EAAGH,IAC5BA,CAAM,ECLdR,EAAoBU,EAAI,CAACN,EAASQ,KACjC,IAAI,IAAInB,KAAOmB,EACXZ,EAAoBa,EAAED,EAAYnB,KAASO,EAAoBa,EAAET,EAASX,IAC5EqB,OAAOC,eAAeX,EAASX,EAAK,CAAEuB,YAAY,EAAMC,IAAKL,EAAWnB,IAE1E,ECNDO,EAAoBxG,EAAI,WACvB,GAA0B,iBAAf0H,WAAyB,OAAOA,WAC3C,IACC,OAAOxH,MAAQ,IAAIyH,SAAS,cAAb,EAChB,CAAE,MAAOC,GACR,GAAsB,iBAAX3H,OAAqB,OAAOA,MACxC,CACA,CAPuB,GCAxBuG,EAAoBa,EAAI,CAACrB,EAAK6B,IAAUP,OAAOzC,UAAUqB,eAAenB,KAAKiB,EAAK6B,4CCK9EC,EAAY,KACZC,EAAS,KACTC,EAAgB/H,OAAO6C,aAAeP,SAASC,gBAAgByF,UACnE,MAAMC,EAAmB,GA2EzB,SAASC,IACP,MAAMC,EAAeC,aAAaC,QAAQ,UAAY,OAZxD,IAAkBC,EACH,WADGA,EAaItI,OAAOuI,WAAW,gCAAgCC,QAI/C,SAAjBL,EACO,QACgB,SAAhBA,EACA,OAEA,OAIU,SAAjBA,EACO,OACgB,QAAhBA,EACA,QAEA,SA9BoB,SAATG,GAA4B,SAATA,IACzCG,QAAQC,MAAM,2BAA2BJ,yBACzCA,EAAO,QAGThG,SAASS,KAAK4F,QAAQC,MAAQN,EAC9BF,aAAaS,QAAQ,QAASP,GAC9BG,QAAQK,IAAI,cAAcR,UA0B5B,CAkDA,SAASnC,KART,WAEE,MAAM4C,EAAUzG,SAAS0G,uBAAuB,gBAChDrE,MAAMsE,KAAKF,GAASlE,SAASqE,IAC3BA,EAAI9C,iBAAiB,QAAS8B,EAAe,GAEjD,CAGEiB,GA9CF,WAEE,IAAIC,EAA6B,EAC7BC,GAAU,EAEdrJ,OAAOoG,iBAAiB,UAAU,SAAUuB,GAC1CyB,EAA6BpJ,OAAOsJ,QAE/BD,IACHrJ,OAAOwF,uBAAsB,WAzDnC,IAAuB+D,IA0DDH,EA9GkC,GAAlDzG,KAAK6G,MAAM1B,EAAO7F,wBAAwBQ,KAC5CqF,EAAOjE,UAAUM,IAAI,YAErB2D,EAAOjE,UAAUC,OAAO,YAI5B,SAAmCyF,GAC7BA,EAAYtB,EACd3F,SAASC,gBAAgBsB,UAAUC,OAAO,oBAEtCyF,EAAYxB,EACdzF,SAASC,gBAAgBsB,UAAUM,IAAI,oBAC9BoF,EAAYxB,GACrBzF,SAASC,gBAAgBsB,UAAUC,OAAO,oBAG9CiE,EAAgBwB,CAClB,CAoCEE,CAA0BF,GAlC5B,SAA6BA,GACT,OAAd1B,IAKa,GAAb0B,EACF1B,EAAU6B,SAAS,EAAG,GAGtB/G,KAAKC,KAAK2G,IACV5G,KAAK6G,MAAMlH,SAASC,gBAAgBS,aAAehD,OAAOqC,aAE1DwF,EAAU6B,SAAS,EAAG7B,EAAU7E,cAGhBV,SAASqH,cAAc,mBAc3C,CAKEC,CAAoBL,GAwDdF,GAAU,CACZ,IAEAA,GAAU,EAEd,IACArJ,OAAO6J,QACT,CA6BEC,GA1BkB,OAAdjC,GAKJ,IAAI,IAAJ,CAAY,cAAe,CACzBrH,QAAQ,EACRuJ,WAAW,EACX5J,SAAU,iBACVI,OAAQ,KACN,IAAIyJ,EAAM9H,WAAW+H,iBAAiB3H,SAASC,iBAAiB2H,UAChE,OAAOpC,EAAO7F,wBAAwBkI,OAAS,GAAMH,EAAM,CAAC,GAiBlE,CAcA1H,SAAS8D,iBAAiB,oBAT1B,WACE9D,SAASS,KAAKW,WAAWG,UAAUC,OAAO,SAE1CgE,EAASxF,SAASqH,cAAc,UAChC9B,EAAYvF,SAASqH,cAAc,eAEnCxD,GACF","sources":["webpack:///./src/furo/assets/scripts/gumshoe-patched.js","webpack:///webpack/bootstrap","webpack:///webpack/runtime/compat get default export","webpack:///webpack/runtime/define property getters","webpack:///webpack/runtime/global","webpack:///webpack/runtime/hasOwnProperty shorthand","webpack:///./src/furo/assets/scripts/furo.js"],"sourcesContent":["/*!\n * gumshoejs v5.1.2 (patched by @pradyunsg)\n * A simple, framework-agnostic scrollspy script.\n * (c) 2019 Chris Ferdinandi\n * MIT License\n * http://github.com/cferdinandi/gumshoe\n */\n\n(function (root, factory) {\n  if (typeof define === \"function\" && define.amd) {\n    define([], function () {\n      return factory(root);\n    });\n  } else if (typeof exports === \"object\") {\n    module.exports = factory(root);\n  } else {\n    root.Gumshoe = factory(root);\n  }\n})(\n  typeof global !== \"undefined\"\n    ? global\n    : typeof window !== \"undefined\"\n    ? window\n    : this,\n  function (window) {\n    \"use strict\";\n\n    //\n    // Defaults\n    //\n\n    var defaults = {\n      // Active classes\n      navClass: \"active\",\n      contentClass: \"active\",\n\n      // Nested navigation\n      nested: false,\n      nestedClass: \"active\",\n\n      // Offset & reflow\n      offset: 0,\n      reflow: false,\n\n      // Event support\n      events: true,\n    };\n\n    //\n    // Methods\n    //\n\n    /**\n     * Merge two or more objects together.\n     * @param   {Object}   objects  The objects to merge together\n     * @returns {Object}            Merged values of defaults and options\n     */\n    var extend = function () {\n      var merged = {};\n      Array.prototype.forEach.call(arguments, function (obj) {\n        for (var key in obj) {\n          if (!obj.hasOwnProperty(key)) return;\n          merged[key] = obj[key];\n        }\n      });\n      return merged;\n    };\n\n    /**\n     * Emit a custom event\n     * @param  {String} type   The event type\n     * @param  {Node}   elem   The element to attach the event to\n     * @param  {Object} detail Any details to pass along with the event\n     */\n    var emitEvent = function (type, elem, detail) {\n      // Make sure events are enabled\n      if (!detail.settings.events) return;\n\n      // Create a new event\n      var event = new CustomEvent(type, {\n        bubbles: true,\n        cancelable: true,\n        detail: detail,\n      });\n\n      // Dispatch the event\n      elem.dispatchEvent(event);\n    };\n\n    /**\n     * Get an element's distance from the top of the Document.\n     * @param  {Node} elem The element\n     * @return {Number}    Distance from the top in pixels\n     */\n    var getOffsetTop = function (elem) {\n      var location = 0;\n      if (elem.offsetParent) {\n        while (elem) {\n          location += elem.offsetTop;\n          elem = elem.offsetParent;\n        }\n      }\n      return location >= 0 ? location : 0;\n    };\n\n    /**\n     * Sort content from first to last in the DOM\n     * @param  {Array} contents The content areas\n     */\n    var sortContents = function (contents) {\n      if (contents) {\n        contents.sort(function (item1, item2) {\n          var offset1 = getOffsetTop(item1.content);\n          var offset2 = getOffsetTop(item2.content);\n          if (offset1 < offset2) return -1;\n          return 1;\n        });\n      }\n    };\n\n    /**\n     * Get the offset to use for calculating position\n     * @param  {Object} settings The settings for this instantiation\n     * @return {Float}           The number of pixels to offset the calculations\n     */\n    var getOffset = function (settings) {\n      // if the offset is a function run it\n      if (typeof settings.offset === \"function\") {\n        return parseFloat(settings.offset());\n      }\n\n      // Otherwise, return it as-is\n      return parseFloat(settings.offset);\n    };\n\n    /**\n     * Get the document element's height\n     * @private\n     * @returns {Number}\n     */\n    var getDocumentHeight = function () {\n      return Math.max(\n        document.body.scrollHeight,\n        document.documentElement.scrollHeight,\n        document.body.offsetHeight,\n        document.documentElement.offsetHeight,\n        document.body.clientHeight,\n        document.documentElement.clientHeight,\n      );\n    };\n\n    /**\n     * Determine if an element is in view\n     * @param  {Node}    elem     The element\n     * @param  {Object}  settings The settings for this instantiation\n     * @param  {Boolean} bottom   If true, check if element is above bottom of viewport instead\n     * @return {Boolean}          Returns true if element is in the viewport\n     */\n    var isInView = function (elem, settings, bottom) {\n      var bounds = elem.getBoundingClientRect();\n      var offset = getOffset(settings);\n      if (bottom) {\n        return (\n          parseInt(bounds.bottom, 10) <\n          (window.innerHeight || document.documentElement.clientHeight)\n        );\n      }\n      return parseInt(bounds.top, 10) <= offset;\n    };\n\n    /**\n     * Check if at the bottom of the viewport\n     * @return {Boolean} If true, page is at the bottom of the viewport\n     */\n    var isAtBottom = function () {\n      if (\n        Math.ceil(window.innerHeight + window.pageYOffset) >=\n        getDocumentHeight()\n      )\n        return true;\n      return false;\n    };\n\n    /**\n     * Check if the last item should be used (even if not at the top of the page)\n     * @param  {Object} item     The last item\n     * @param  {Object} settings The settings for this instantiation\n     * @return {Boolean}         If true, use the last item\n     */\n    var useLastItem = function (item, settings) {\n      if (isAtBottom() && isInView(item.content, settings, true)) return true;\n      return false;\n    };\n\n    /**\n     * Get the active content\n     * @param  {Array}  contents The content areas\n     * @param  {Object} settings The settings for this instantiation\n     * @return {Object}          The content area and matching navigation link\n     */\n    var getActive = function (contents, settings) {\n      var last = contents[contents.length - 1];\n      if (useLastItem(last, settings)) return last;\n      for (var i = contents.length - 1; i >= 0; i--) {\n        if (isInView(contents[i].content, settings)) return contents[i];\n      }\n    };\n\n    /**\n     * Deactivate parent navs in a nested navigation\n     * @param  {Node}   nav      The starting navigation element\n     * @param  {Object} settings The settings for this instantiation\n     */\n    var deactivateNested = function (nav, settings) {\n      // If nesting isn't activated, bail\n      if (!settings.nested || !nav.parentNode) return;\n\n      // Get the parent navigation\n      var li = nav.parentNode.closest(\"li\");\n      if (!li) return;\n\n      // Remove the active class\n      li.classList.remove(settings.nestedClass);\n\n      // Apply recursively to any parent navigation elements\n      deactivateNested(li, settings);\n    };\n\n    /**\n     * Deactivate a nav and content area\n     * @param  {Object} items    The nav item and content to deactivate\n     * @param  {Object} settings The settings for this instantiation\n     */\n    var deactivate = function (items, settings) {\n      // Make sure there are items to deactivate\n      if (!items) return;\n\n      // Get the parent list item\n      var li = items.nav.closest(\"li\");\n      if (!li) return;\n\n      // Remove the active class from the nav and content\n      li.classList.remove(settings.navClass);\n      items.content.classList.remove(settings.contentClass);\n\n      // Deactivate any parent navs in a nested navigation\n      deactivateNested(li, settings);\n\n      // Emit a custom event\n      emitEvent(\"gumshoeDeactivate\", li, {\n        link: items.nav,\n        content: items.content,\n        settings: settings,\n      });\n    };\n\n    /**\n     * Activate parent navs in a nested navigation\n     * @param  {Node}   nav      The starting navigation element\n     * @param  {Object} settings The settings for this instantiation\n     */\n    var activateNested = function (nav, settings) {\n      // If nesting isn't activated, bail\n      if (!settings.nested) return;\n\n      // Get the parent navigation\n      var li = nav.parentNode.closest(\"li\");\n      if (!li) return;\n\n      // Add the active class\n      li.classList.add(settings.nestedClass);\n\n      // Apply recursively to any parent navigation elements\n      activateNested(li, settings);\n    };\n\n    /**\n     * Activate a nav and content area\n     * @param  {Object} items    The nav item and content to activate\n     * @param  {Object} settings The settings for this instantiation\n     */\n    var activate = function (items, settings) {\n      // Make sure there are items to activate\n      if (!items) return;\n\n      // Get the parent list item\n      var li = items.nav.closest(\"li\");\n      if (!li) return;\n\n      // Add the active class to the nav and content\n      li.classList.add(settings.navClass);\n      items.content.classList.add(settings.contentClass);\n\n      // Activate any parent navs in a nested navigation\n      activateNested(li, settings);\n\n      // Emit a custom event\n      emitEvent(\"gumshoeActivate\", li, {\n        link: items.nav,\n        content: items.content,\n        settings: settings,\n      });\n    };\n\n    /**\n     * Create the Constructor object\n     * @param {String} selector The selector to use for navigation items\n     * @param {Object} options  User options and settings\n     */\n    var Constructor = function (selector, options) {\n      //\n      // Variables\n      //\n\n      var publicAPIs = {};\n      var navItems, contents, current, timeout, settings;\n\n      //\n      // Methods\n      //\n\n      /**\n       * Set variables from DOM elements\n       */\n      publicAPIs.setup = function () {\n        // Get all nav items\n        navItems = document.querySelectorAll(selector);\n\n        // Create contents array\n        contents = [];\n\n        // Loop through each item, get it's matching content, and push to the array\n        Array.prototype.forEach.call(navItems, function (item) {\n          // Get the content for the nav item\n          var content = document.getElementById(\n            decodeURIComponent(item.hash.substr(1)),\n          );\n          if (!content) return;\n\n          // Push to the contents array\n          contents.push({\n            nav: item,\n            content: content,\n          });\n        });\n\n        // Sort contents by the order they appear in the DOM\n        sortContents(contents);\n      };\n\n      /**\n       * Detect which content is currently active\n       */\n      publicAPIs.detect = function () {\n        // Get the active content\n        var active = getActive(contents, settings);\n\n        // if there's no active content, deactivate and bail\n        if (!active) {\n          if (current) {\n            deactivate(current, settings);\n            current = null;\n          }\n          return;\n        }\n\n        // If the active content is the one currently active, do nothing\n        if (current && active.content === current.content) return;\n\n        // Deactivate the current content and activate the new content\n        deactivate(current, settings);\n        activate(active, settings);\n\n        // Update the currently active content\n        current = active;\n      };\n\n      /**\n       * Detect the active content on scroll\n       * Debounced for performance\n       */\n      var scrollHandler = function (event) {\n        // If there's a timer, cancel it\n        if (timeout) {\n          window.cancelAnimationFrame(timeout);\n        }\n\n        // Setup debounce callback\n        timeout = window.requestAnimationFrame(publicAPIs.detect);\n      };\n\n      /**\n       * Update content sorting on resize\n       * Debounced for performance\n       */\n      var resizeHandler = function (event) {\n        // If there's a timer, cancel it\n        if (timeout) {\n          window.cancelAnimationFrame(timeout);\n        }\n\n        // Setup debounce callback\n        timeout = window.requestAnimationFrame(function () {\n          sortContents(contents);\n          publicAPIs.detect();\n        });\n      };\n\n      /**\n       * Destroy the current instantiation\n       */\n      publicAPIs.destroy = function () {\n        // Undo DOM changes\n        if (current) {\n          deactivate(current, settings);\n        }\n\n        // Remove event listeners\n        window.removeEventListener(\"scroll\", scrollHandler, false);\n        if (settings.reflow) {\n          window.removeEventListener(\"resize\", resizeHandler, false);\n        }\n\n        // Reset variables\n        contents = null;\n        navItems = null;\n        current = null;\n        timeout = null;\n        settings = null;\n      };\n\n      /**\n       * Initialize the current instantiation\n       */\n      var init = function () {\n        // Merge user options into defaults\n        settings = extend(defaults, options || {});\n\n        // Setup variables based on the current DOM\n        publicAPIs.setup();\n\n        // Find the currently active content\n        publicAPIs.detect();\n\n        // Setup event listeners\n        window.addEventListener(\"scroll\", scrollHandler, false);\n        if (settings.reflow) {\n          window.addEventListener(\"resize\", resizeHandler, false);\n        }\n      };\n\n      //\n      // Initialize and return the public APIs\n      //\n\n      init();\n      return publicAPIs;\n    };\n\n    //\n    // Return the Constructor\n    //\n\n    return Constructor;\n  },\n);\n","// The module cache\nvar __webpack_module_cache__ = {};\n\n// The require function\nfunction __webpack_require__(moduleId) {\n\t// Check if module is in cache\n\tvar cachedModule = __webpack_module_cache__[moduleId];\n\tif (cachedModule !== undefined) {\n\t\treturn cachedModule.exports;\n\t}\n\t// Create a new module (and put it into the cache)\n\tvar module = __webpack_module_cache__[moduleId] = {\n\t\t// no module.id needed\n\t\t// no module.loaded needed\n\t\texports: {}\n\t};\n\n\t// Execute the module function\n\t__webpack_modules__[moduleId].call(module.exports, module, module.exports, __webpack_require__);\n\n\t// Return the exports of the module\n\treturn module.exports;\n}\n\n","// getDefaultExport function for compatibility with non-harmony modules\n__webpack_require__.n = (module) => {\n\tvar getter = module && module.__esModule ?\n\t\t() => (module['default']) :\n\t\t() => (module);\n\t__webpack_require__.d(getter, { a: getter });\n\treturn getter;\n};","// define getter functions for harmony exports\n__webpack_require__.d = (exports, definition) => {\n\tfor(var key in definition) {\n\t\tif(__webpack_require__.o(definition, key) && !__webpack_require__.o(exports, key)) {\n\t\t\tObject.defineProperty(exports, key, { enumerable: true, get: definition[key] });\n\t\t}\n\t}\n};","__webpack_require__.g = (function() {\n\tif (typeof globalThis === 'object') return globalThis;\n\ttry {\n\t\treturn this || new Function('return this')();\n\t} catch (e) {\n\t\tif (typeof window === 'object') return window;\n\t}\n})();","__webpack_require__.o = (obj, prop) => (Object.prototype.hasOwnProperty.call(obj, prop))","import Gumshoe from \"./gumshoe-patched.js\";\n\n////////////////////////////////////////////////////////////////////////////////\n// Scroll Handling\n////////////////////////////////////////////////////////////////////////////////\nvar tocScroll = null;\nvar header = null;\nvar lastScrollTop = window.pageYOffset || document.documentElement.scrollTop;\nconst GO_TO_TOP_OFFSET = 64;\n\nfunction scrollHandlerForHeader() {\n  if (Math.floor(header.getBoundingClientRect().top) == 0) {\n    header.classList.add(\"scrolled\");\n  } else {\n    header.classList.remove(\"scrolled\");\n  }\n}\n\nfunction scrollHandlerForBackToTop(positionY) {\n  if (positionY < GO_TO_TOP_OFFSET) {\n    document.documentElement.classList.remove(\"show-back-to-top\");\n  } else {\n    if (positionY < lastScrollTop) {\n      document.documentElement.classList.add(\"show-back-to-top\");\n    } else if (positionY > lastScrollTop) {\n      document.documentElement.classList.remove(\"show-back-to-top\");\n    }\n  }\n  lastScrollTop = positionY;\n}\n\nfunction scrollHandlerForTOC(positionY) {\n  if (tocScroll === null) {\n    return;\n  }\n\n  // top of page.\n  if (positionY == 0) {\n    tocScroll.scrollTo(0, 0);\n  } else if (\n    // bottom of page.\n    Math.ceil(positionY) >=\n    Math.floor(document.documentElement.scrollHeight - window.innerHeight)\n  ) {\n    tocScroll.scrollTo(0, tocScroll.scrollHeight);\n  } else {\n    // somewhere in the middle.\n    const current = document.querySelector(\".scroll-current\");\n    if (current == null) {\n      return;\n    }\n\n    // https://github.com/pypa/pip/issues/9159 This breaks scroll behaviours.\n    // // scroll the currently \"active\" heading in toc, into view.\n    // const rect = current.getBoundingClientRect();\n    // if (0 > rect.top) {\n    //   current.scrollIntoView(true); // the argument is \"alignTop\"\n    // } else if (rect.bottom > window.innerHeight) {\n    //   current.scrollIntoView(false);\n    // }\n  }\n}\n\nfunction scrollHandler(positionY) {\n  scrollHandlerForHeader();\n  scrollHandlerForBackToTop(positionY);\n  scrollHandlerForTOC(positionY);\n}\n\n////////////////////////////////////////////////////////////////////////////////\n// Theme Toggle\n////////////////////////////////////////////////////////////////////////////////\nfunction setTheme(mode) {\n  if (mode !== \"light\" && mode !== \"dark\" && mode !== \"auto\") {\n    console.error(`Got invalid theme mode: ${mode}. Resetting to auto.`);\n    mode = \"auto\";\n  }\n\n  document.body.dataset.theme = mode;\n  localStorage.setItem(\"theme\", mode);\n  console.log(`Changed to ${mode} mode.`);\n}\n\nfunction cycleThemeOnce() {\n  const currentTheme = localStorage.getItem(\"theme\") || \"auto\";\n  const prefersDark = window.matchMedia(\"(prefers-color-scheme: dark)\").matches;\n\n  if (prefersDark) {\n    // Auto (dark) -> Light -> Dark\n    if (currentTheme === \"auto\") {\n      setTheme(\"light\");\n    } else if (currentTheme == \"light\") {\n      setTheme(\"dark\");\n    } else {\n      setTheme(\"auto\");\n    }\n  } else {\n    // Auto (light) -> Dark -> Light\n    if (currentTheme === \"auto\") {\n      setTheme(\"dark\");\n    } else if (currentTheme == \"dark\") {\n      setTheme(\"light\");\n    } else {\n      setTheme(\"auto\");\n    }\n  }\n}\n\n////////////////////////////////////////////////////////////////////////////////\n// Setup\n////////////////////////////////////////////////////////////////////////////////\nfunction setupScrollHandler() {\n  // Taken from https://developer.mozilla.org/en-US/docs/Web/API/Document/scroll_event\n  let last_known_scroll_position = 0;\n  let ticking = false;\n\n  window.addEventListener(\"scroll\", function (e) {\n    last_known_scroll_position = window.scrollY;\n\n    if (!ticking) {\n      window.requestAnimationFrame(function () {\n        scrollHandler(last_known_scroll_position);\n        ticking = false;\n      });\n\n      ticking = true;\n    }\n  });\n  window.scroll();\n}\n\nfunction setupScrollSpy() {\n  if (tocScroll === null) {\n    return;\n  }\n\n  // Scrollspy -- highlight table on contents, based on scroll\n  new Gumshoe(\".toc-tree a\", {\n    reflow: true,\n    recursive: true,\n    navClass: \"scroll-current\",\n    offset: () => {\n      let rem = parseFloat(getComputedStyle(document.documentElement).fontSize);\n      return header.getBoundingClientRect().height + 0.5 * rem + 1;\n    },\n  });\n}\n\nfunction setupTheme() {\n  // Attach event handlers for toggling themes\n  const buttons = document.getElementsByClassName(\"theme-toggle\");\n  Array.from(buttons).forEach((btn) => {\n    btn.addEventListener(\"click\", cycleThemeOnce);\n  });\n}\n\nfunction setup() {\n  setupTheme();\n  setupScrollHandler();\n  setupScrollSpy();\n}\n\n////////////////////////////////////////////////////////////////////////////////\n// Main entrypoint\n////////////////////////////////////////////////////////////////////////////////\nfunction main() {\n  document.body.parentNode.classList.remove(\"no-js\");\n\n  header = document.querySelector(\"header\");\n  tocScroll = document.querySelector(\".toc-scroll\");\n\n  setup();\n}\n\ndocument.addEventListener(\"DOMContentLoaded\", main);\n"],"names":["root","g","window","this","defaults","navClass","contentClass","nested","nestedClass","offset","reflow","events","emitEvent","type","elem","detail","settings","event","CustomEvent","bubbles","cancelable","dispatchEvent","getOffsetTop","location","offsetParent","offsetTop","sortContents","contents","sort","item1","item2","content","isInView","bottom","bounds","getBoundingClientRect","parseFloat","getOffset","parseInt","innerHeight","document","documentElement","clientHeight","top","isAtBottom","Math","ceil","pageYOffset","max","body","scrollHeight","offsetHeight","getActive","last","length","item","useLastItem","i","deactivateNested","nav","parentNode","li","closest","classList","remove","deactivate","items","link","activateNested","add","selector","options","navItems","current","timeout","publicAPIs","querySelectorAll","Array","prototype","forEach","call","getElementById","decodeURIComponent","hash","substr","push","active","activate","scrollHandler","cancelAnimationFrame","requestAnimationFrame","detect","resizeHandler","destroy","removeEventListener","merged","arguments","obj","key","hasOwnProperty","extend","setup","addEventListener","factory","__webpack_module_cache__","__webpack_require__","moduleId","cachedModule","undefined","exports","module","__webpack_modules__","n","getter","__esModule","d","a","definition","o","Object","defineProperty","enumerable","get","globalThis","Function","e","prop","tocScroll","header","lastScrollTop","scrollTop","GO_TO_TOP_OFFSET","cycleThemeOnce","currentTheme","localStorage","getItem","mode","matchMedia","matches","console","error","dataset","theme","setItem","log","buttons","getElementsByClassName","from","btn","setupTheme","last_known_scroll_position","ticking","scrollY","positionY","floor","scrollHandlerForBackToTop","scrollTo","querySelector","scrollHandlerForTOC","scroll","setupScrollHandler","recursive","rem","getComputedStyle","fontSize","height"],"sourceRoot":""}
\ No newline at end of file
diff --git a/docs/_build/html/_static/searchtools.js b/docs/_build/html/_static/searchtools.js
deleted file mode 100644
index 7918c3fab..000000000
--- a/docs/_build/html/_static/searchtools.js
+++ /dev/null
@@ -1,574 +0,0 @@
-/*
- * searchtools.js
- * ~~~~~~~~~~~~~~~~
- *
- * Sphinx JavaScript utilities for the full-text search.
- *
- * :copyright: Copyright 2007-2023 by the Sphinx team, see AUTHORS.
- * :license: BSD, see LICENSE for details.
- *
- */
-"use strict";
-
-/**
- * Simple result scoring code.
- */
-if (typeof Scorer === "undefined") {
-  var Scorer = {
-    // Implement the following function to further tweak the score for each result
-    // The function takes a result array [docname, title, anchor, descr, score, filename]
-    // and returns the new score.
-    /*
-    score: result => {
-      const [docname, title, anchor, descr, score, filename] = result
-      return score
-    },
-    */
-
-    // query matches the full name of an object
-    objNameMatch: 11,
-    // or matches in the last dotted part of the object name
-    objPartialMatch: 6,
-    // Additive scores depending on the priority of the object
-    objPrio: {
-      0: 15, // used to be importantResults
-      1: 5, // used to be objectResults
-      2: -5, // used to be unimportantResults
-    },
-    //  Used when the priority is not in the mapping.
-    objPrioDefault: 0,
-
-    // query found in title
-    title: 15,
-    partialTitle: 7,
-    // query found in terms
-    term: 5,
-    partialTerm: 2,
-  };
-}
-
-const _removeChildren = (element) => {
-  while (element && element.lastChild) element.removeChild(element.lastChild);
-};
-
-/**
- * See https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions#escaping
- */
-const _escapeRegExp = (string) =>
-  string.replace(/[.*+\-?^${}()|[\]\\]/g, "\\$&"); // $& means the whole matched string
-
-const _displayItem = (item, searchTerms, highlightTerms) => {
-  const docBuilder = DOCUMENTATION_OPTIONS.BUILDER;
-  const docFileSuffix = DOCUMENTATION_OPTIONS.FILE_SUFFIX;
-  const docLinkSuffix = DOCUMENTATION_OPTIONS.LINK_SUFFIX;
-  const showSearchSummary = DOCUMENTATION_OPTIONS.SHOW_SEARCH_SUMMARY;
-  const contentRoot = document.documentElement.dataset.content_root;
-
-  const [docName, title, anchor, descr, score, _filename] = item;
-
-  let listItem = document.createElement("li");
-  let requestUrl;
-  let linkUrl;
-  if (docBuilder === "dirhtml") {
-    // dirhtml builder
-    let dirname = docName + "/";
-    if (dirname.match(/\/index\/$/))
-      dirname = dirname.substring(0, dirname.length - 6);
-    else if (dirname === "index/") dirname = "";
-    requestUrl = contentRoot + dirname;
-    linkUrl = requestUrl;
-  } else {
-    // normal html builders
-    requestUrl = contentRoot + docName + docFileSuffix;
-    linkUrl = docName + docLinkSuffix;
-  }
-  let linkEl = listItem.appendChild(document.createElement("a"));
-  linkEl.href = linkUrl + anchor;
-  linkEl.dataset.score = score;
-  linkEl.innerHTML = title;
-  if (descr) {
-    listItem.appendChild(document.createElement("span")).innerHTML =
-      " (" + descr + ")";
-    // highlight search terms in the description
-    if (SPHINX_HIGHLIGHT_ENABLED)  // set in sphinx_highlight.js
-      highlightTerms.forEach((term) => _highlightText(listItem, term, "highlighted"));
-  }
-  else if (showSearchSummary)
-    fetch(requestUrl)
-      .then((responseData) => responseData.text())
-      .then((data) => {
-        if (data)
-          listItem.appendChild(
-            Search.makeSearchSummary(data, searchTerms)
-          );
-        // highlight search terms in the summary
-        if (SPHINX_HIGHLIGHT_ENABLED)  // set in sphinx_highlight.js
-          highlightTerms.forEach((term) => _highlightText(listItem, term, "highlighted"));
-      });
-  Search.output.appendChild(listItem);
-};
-const _finishSearch = (resultCount) => {
-  Search.stopPulse();
-  Search.title.innerText = _("Search Results");
-  if (!resultCount)
-    Search.status.innerText = Documentation.gettext(
-      "Your search did not match any documents. Please make sure that all words are spelled correctly and that you've selected enough categories."
-    );
-  else
-    Search.status.innerText = _(
-      `Search finished, found ${resultCount} page(s) matching the search query.`
-    );
-};
-const _displayNextItem = (
-  results,
-  resultCount,
-  searchTerms,
-  highlightTerms,
-) => {
-  // results left, load the summary and display it
-  // this is intended to be dynamic (don't sub resultsCount)
-  if (results.length) {
-    _displayItem(results.pop(), searchTerms, highlightTerms);
-    setTimeout(
-      () => _displayNextItem(results, resultCount, searchTerms, highlightTerms),
-      5
-    );
-  }
-  // search finished, update title and status message
-  else _finishSearch(resultCount);
-};
-
-/**
- * Default splitQuery function. Can be overridden in ``sphinx.search`` with a
- * custom function per language.
- *
- * The regular expression works by splitting the string on consecutive characters
- * that are not Unicode letters, numbers, underscores, or emoji characters.
- * This is the same as ``\W+`` in Python, preserving the surrogate pair area.
- */
-if (typeof splitQuery === "undefined") {
-  var splitQuery = (query) => query
-      .split(/[^\p{Letter}\p{Number}_\p{Emoji_Presentation}]+/gu)
-      .filter(term => term)  // remove remaining empty strings
-}
-
-/**
- * Search Module
- */
-const Search = {
-  _index: null,
-  _queued_query: null,
-  _pulse_status: -1,
-
-  htmlToText: (htmlString) => {
-    const htmlElement = new DOMParser().parseFromString(htmlString, 'text/html');
-    htmlElement.querySelectorAll(".headerlink").forEach((el) => { el.remove() });
-    const docContent = htmlElement.querySelector('[role="main"]');
-    if (docContent !== undefined) return docContent.textContent;
-    console.warn(
-      "Content block not found. Sphinx search tries to obtain it via '[role=main]'. Could you check your theme or template."
-    );
-    return "";
-  },
-
-  init: () => {
-    const query = new URLSearchParams(window.location.search).get("q");
-    document
-      .querySelectorAll('input[name="q"]')
-      .forEach((el) => (el.value = query));
-    if (query) Search.performSearch(query);
-  },
-
-  loadIndex: (url) =>
-    (document.body.appendChild(document.createElement("script")).src = url),
-
-  setIndex: (index) => {
-    Search._index = index;
-    if (Search._queued_query !== null) {
-      const query = Search._queued_query;
-      Search._queued_query = null;
-      Search.query(query);
-    }
-  },
-
-  hasIndex: () => Search._index !== null,
-
-  deferQuery: (query) => (Search._queued_query = query),
-
-  stopPulse: () => (Search._pulse_status = -1),
-
-  startPulse: () => {
-    if (Search._pulse_status >= 0) return;
-
-    const pulse = () => {
-      Search._pulse_status = (Search._pulse_status + 1) % 4;
-      Search.dots.innerText = ".".repeat(Search._pulse_status);
-      if (Search._pulse_status >= 0) window.setTimeout(pulse, 500);
-    };
-    pulse();
-  },
-
-  /**
-   * perform a search for something (or wait until index is loaded)
-   */
-  performSearch: (query) => {
-    // create the required interface elements
-    const searchText = document.createElement("h2");
-    searchText.textContent = _("Searching");
-    const searchSummary = document.createElement("p");
-    searchSummary.classList.add("search-summary");
-    searchSummary.innerText = "";
-    const searchList = document.createElement("ul");
-    searchList.classList.add("search");
-
-    const out = document.getElementById("search-results");
-    Search.title = out.appendChild(searchText);
-    Search.dots = Search.title.appendChild(document.createElement("span"));
-    Search.status = out.appendChild(searchSummary);
-    Search.output = out.appendChild(searchList);
-
-    const searchProgress = document.getElementById("search-progress");
-    // Some themes don't use the search progress node
-    if (searchProgress) {
-      searchProgress.innerText = _("Preparing search...");
-    }
-    Search.startPulse();
-
-    // index already loaded, the browser was quick!
-    if (Search.hasIndex()) Search.query(query);
-    else Search.deferQuery(query);
-  },
-
-  /**
-   * execute search (requires search index to be loaded)
-   */
-  query: (query) => {
-    const filenames = Search._index.filenames;
-    const docNames = Search._index.docnames;
-    const titles = Search._index.titles;
-    const allTitles = Search._index.alltitles;
-    const indexEntries = Search._index.indexentries;
-
-    // stem the search terms and add them to the correct list
-    const stemmer = new Stemmer();
-    const searchTerms = new Set();
-    const excludedTerms = new Set();
-    const highlightTerms = new Set();
-    const objectTerms = new Set(splitQuery(query.toLowerCase().trim()));
-    splitQuery(query.trim()).forEach((queryTerm) => {
-      const queryTermLower = queryTerm.toLowerCase();
-
-      // maybe skip this "word"
-      // stopwords array is from language_data.js
-      if (
-        stopwords.indexOf(queryTermLower) !== -1 ||
-        queryTerm.match(/^\d+$/)
-      )
-        return;
-
-      // stem the word
-      let word = stemmer.stemWord(queryTermLower);
-      // select the correct list
-      if (word[0] === "-") excludedTerms.add(word.substr(1));
-      else {
-        searchTerms.add(word);
-        highlightTerms.add(queryTermLower);
-      }
-    });
-
-    if (SPHINX_HIGHLIGHT_ENABLED) {  // set in sphinx_highlight.js
-      localStorage.setItem("sphinx_highlight_terms", [...highlightTerms].join(" "))
-    }
-
-    // console.debug("SEARCH: searching for:");
-    // console.info("required: ", [...searchTerms]);
-    // console.info("excluded: ", [...excludedTerms]);
-
-    // array of [docname, title, anchor, descr, score, filename]
-    let results = [];
-    _removeChildren(document.getElementById("search-progress"));
-
-    const queryLower = query.toLowerCase();
-    for (const [title, foundTitles] of Object.entries(allTitles)) {
-      if (title.toLowerCase().includes(queryLower) && (queryLower.length >= title.length/2)) {
-        for (const [file, id] of foundTitles) {
-          let score = Math.round(100 * queryLower.length / title.length)
-          results.push([
-            docNames[file],
-            titles[file] !== title ? `${titles[file]} > ${title}` : title,
-            id !== null ? "#" + id : "",
-            null,
-            score,
-            filenames[file],
-          ]);
-        }
-      }
-    }
-
-    // search for explicit entries in index directives
-    for (const [entry, foundEntries] of Object.entries(indexEntries)) {
-      if (entry.includes(queryLower) && (queryLower.length >= entry.length/2)) {
-        for (const [file, id] of foundEntries) {
-          let score = Math.round(100 * queryLower.length / entry.length)
-          results.push([
-            docNames[file],
-            titles[file],
-            id ? "#" + id : "",
-            null,
-            score,
-            filenames[file],
-          ]);
-        }
-      }
-    }
-
-    // lookup as object
-    objectTerms.forEach((term) =>
-      results.push(...Search.performObjectSearch(term, objectTerms))
-    );
-
-    // lookup as search terms in fulltext
-    results.push(...Search.performTermsSearch(searchTerms, excludedTerms));
-
-    // let the scorer override scores with a custom scoring function
-    if (Scorer.score) results.forEach((item) => (item[4] = Scorer.score(item)));
-
-    // now sort the results by score (in opposite order of appearance, since the
-    // display function below uses pop() to retrieve items) and then
-    // alphabetically
-    results.sort((a, b) => {
-      const leftScore = a[4];
-      const rightScore = b[4];
-      if (leftScore === rightScore) {
-        // same score: sort alphabetically
-        const leftTitle = a[1].toLowerCase();
-        const rightTitle = b[1].toLowerCase();
-        if (leftTitle === rightTitle) return 0;
-        return leftTitle > rightTitle ? -1 : 1; // inverted is intentional
-      }
-      return leftScore > rightScore ? 1 : -1;
-    });
-
-    // remove duplicate search results
-    // note the reversing of results, so that in the case of duplicates, the highest-scoring entry is kept
-    let seen = new Set();
-    results = results.reverse().reduce((acc, result) => {
-      let resultStr = result.slice(0, 4).concat([result[5]]).map(v => String(v)).join(',');
-      if (!seen.has(resultStr)) {
-        acc.push(result);
-        seen.add(resultStr);
-      }
-      return acc;
-    }, []);
-
-    results = results.reverse();
-
-    // for debugging
-    //Search.lastresults = results.slice();  // a copy
-    // console.info("search results:", Search.lastresults);
-
-    // print the results
-    _displayNextItem(results, results.length, searchTerms, highlightTerms);
-  },
-
-  /**
-   * search for object names
-   */
-  performObjectSearch: (object, objectTerms) => {
-    const filenames = Search._index.filenames;
-    const docNames = Search._index.docnames;
-    const objects = Search._index.objects;
-    const objNames = Search._index.objnames;
-    const titles = Search._index.titles;
-
-    const results = [];
-
-    const objectSearchCallback = (prefix, match) => {
-      const name = match[4]
-      const fullname = (prefix ? prefix + "." : "") + name;
-      const fullnameLower = fullname.toLowerCase();
-      if (fullnameLower.indexOf(object) < 0) return;
-
-      let score = 0;
-      const parts = fullnameLower.split(".");
-
-      // check for different match types: exact matches of full name or
-      // "last name" (i.e. last dotted part)
-      if (fullnameLower === object || parts.slice(-1)[0] === object)
-        score += Scorer.objNameMatch;
-      else if (parts.slice(-1)[0].indexOf(object) > -1)
-        score += Scorer.objPartialMatch; // matches in last name
-
-      const objName = objNames[match[1]][2];
-      const title = titles[match[0]];
-
-      // If more than one term searched for, we require other words to be
-      // found in the name/title/description
-      const otherTerms = new Set(objectTerms);
-      otherTerms.delete(object);
-      if (otherTerms.size > 0) {
-        const haystack = `${prefix} ${name} ${objName} ${title}`.toLowerCase();
-        if (
-          [...otherTerms].some((otherTerm) => haystack.indexOf(otherTerm) < 0)
-        )
-          return;
-      }
-
-      let anchor = match[3];
-      if (anchor === "") anchor = fullname;
-      else if (anchor === "-") anchor = objNames[match[1]][1] + "-" + fullname;
-
-      const descr = objName + _(", in ") + title;
-
-      // add custom score for some objects according to scorer
-      if (Scorer.objPrio.hasOwnProperty(match[2]))
-        score += Scorer.objPrio[match[2]];
-      else score += Scorer.objPrioDefault;
-
-      results.push([
-        docNames[match[0]],
-        fullname,
-        "#" + anchor,
-        descr,
-        score,
-        filenames[match[0]],
-      ]);
-    };
-    Object.keys(objects).forEach((prefix) =>
-      objects[prefix].forEach((array) =>
-        objectSearchCallback(prefix, array)
-      )
-    );
-    return results;
-  },
-
-  /**
-   * search for full-text terms in the index
-   */
-  performTermsSearch: (searchTerms, excludedTerms) => {
-    // prepare search
-    const terms = Search._index.terms;
-    const titleTerms = Search._index.titleterms;
-    const filenames = Search._index.filenames;
-    const docNames = Search._index.docnames;
-    const titles = Search._index.titles;
-
-    const scoreMap = new Map();
-    const fileMap = new Map();
-
-    // perform the search on the required terms
-    searchTerms.forEach((word) => {
-      const files = [];
-      const arr = [
-        { files: terms[word], score: Scorer.term },
-        { files: titleTerms[word], score: Scorer.title },
-      ];
-      // add support for partial matches
-      if (word.length > 2) {
-        const escapedWord = _escapeRegExp(word);
-        Object.keys(terms).forEach((term) => {
-          if (term.match(escapedWord) && !terms[word])
-            arr.push({ files: terms[term], score: Scorer.partialTerm });
-        });
-        Object.keys(titleTerms).forEach((term) => {
-          if (term.match(escapedWord) && !titleTerms[word])
-            arr.push({ files: titleTerms[word], score: Scorer.partialTitle });
-        });
-      }
-
-      // no match but word was a required one
-      if (arr.every((record) => record.files === undefined)) return;
-
-      // found search word in contents
-      arr.forEach((record) => {
-        if (record.files === undefined) return;
-
-        let recordFiles = record.files;
-        if (recordFiles.length === undefined) recordFiles = [recordFiles];
-        files.push(...recordFiles);
-
-        // set score for the word in each file
-        recordFiles.forEach((file) => {
-          if (!scoreMap.has(file)) scoreMap.set(file, {});
-          scoreMap.get(file)[word] = record.score;
-        });
-      });
-
-      // create the mapping
-      files.forEach((file) => {
-        if (fileMap.has(file) && fileMap.get(file).indexOf(word) === -1)
-          fileMap.get(file).push(word);
-        else fileMap.set(file, [word]);
-      });
-    });
-
-    // now check if the files don't contain excluded terms
-    const results = [];
-    for (const [file, wordList] of fileMap) {
-      // check if all requirements are matched
-
-      // as search terms with length < 3 are discarded
-      const filteredTermCount = [...searchTerms].filter(
-        (term) => term.length > 2
-      ).length;
-      if (
-        wordList.length !== searchTerms.size &&
-        wordList.length !== filteredTermCount
-      )
-        continue;
-
-      // ensure that none of the excluded terms is in the search result
-      if (
-        [...excludedTerms].some(
-          (term) =>
-            terms[term] === file ||
-            titleTerms[term] === file ||
-            (terms[term] || []).includes(file) ||
-            (titleTerms[term] || []).includes(file)
-        )
-      )
-        break;
-
-      // select one (max) score for the file.
-      const score = Math.max(...wordList.map((w) => scoreMap.get(file)[w]));
-      // add result to the result list
-      results.push([
-        docNames[file],
-        titles[file],
-        "",
-        null,
-        score,
-        filenames[file],
-      ]);
-    }
-    return results;
-  },
-
-  /**
-   * helper function to return a node containing the
-   * search summary for a given text. keywords is a list
-   * of stemmed words.
-   */
-  makeSearchSummary: (htmlText, keywords) => {
-    const text = Search.htmlToText(htmlText);
-    if (text === "") return null;
-
-    const textLower = text.toLowerCase();
-    const actualStartPosition = [...keywords]
-      .map((k) => textLower.indexOf(k.toLowerCase()))
-      .filter((i) => i > -1)
-      .slice(-1)[0];
-    const startWithContext = Math.max(actualStartPosition - 120, 0);
-
-    const top = startWithContext === 0 ? "" : "...";
-    const tail = startWithContext + 240 < text.length ? "..." : "";
-
-    let summary = document.createElement("p");
-    summary.classList.add("context");
-    summary.textContent = top + text.substr(startWithContext, 240).trim() + tail;
-
-    return summary;
-  },
-};
-
-_ready(Search.init);
diff --git a/docs/_build/html/_static/skeleton.css b/docs/_build/html/_static/skeleton.css
deleted file mode 100644
index 467c878c6..000000000
--- a/docs/_build/html/_static/skeleton.css
+++ /dev/null
@@ -1,296 +0,0 @@
-/* Some sane resets. */
-html {
-  height: 100%;
-}
-
-body {
-  margin: 0;
-  min-height: 100%;
-}
-
-/* All the flexbox magic! */
-body,
-.sb-announcement,
-.sb-content,
-.sb-main,
-.sb-container,
-.sb-container__inner,
-.sb-article-container,
-.sb-footer-content,
-.sb-header,
-.sb-header-secondary,
-.sb-footer {
-  display: flex;
-}
-
-/* These order things vertically */
-body,
-.sb-main,
-.sb-article-container {
-  flex-direction: column;
-}
-
-/* Put elements in the center */
-.sb-header,
-.sb-header-secondary,
-.sb-container,
-.sb-content,
-.sb-footer,
-.sb-footer-content {
-  justify-content: center;
-}
-/* Put elements at the ends */
-.sb-article-container {
-  justify-content: space-between;
-}
-
-/* These elements grow. */
-.sb-main,
-.sb-content,
-.sb-container,
-article {
-  flex-grow: 1;
-}
-
-/* Because padding making this wider is not fun */
-article {
-  box-sizing: border-box;
-}
-
-/* The announcements element should never be wider than the page. */
-.sb-announcement {
-  max-width: 100%;
-}
-
-.sb-sidebar-primary,
-.sb-sidebar-secondary {
-  flex-shrink: 0;
-  width: 17rem;
-}
-
-.sb-announcement__inner {
-  justify-content: center;
-
-  box-sizing: border-box;
-  height: 3rem;
-
-  overflow-x: auto;
-  white-space: nowrap;
-}
-
-/* Sidebars, with checkbox-based toggle */
-.sb-sidebar-primary,
-.sb-sidebar-secondary {
-  position: fixed;
-  height: 100%;
-  top: 0;
-}
-
-.sb-sidebar-primary {
-  left: -17rem;
-  transition: left 250ms ease-in-out;
-}
-.sb-sidebar-secondary {
-  right: -17rem;
-  transition: right 250ms ease-in-out;
-}
-
-.sb-sidebar-toggle {
-  display: none;
-}
-.sb-sidebar-overlay {
-  position: fixed;
-  top: 0;
-  width: 0;
-  height: 0;
-
-  transition: width 0ms ease 250ms, height 0ms ease 250ms, opacity 250ms ease;
-
-  opacity: 0;
-  background-color: rgba(0, 0, 0, 0.54);
-}
-
-#sb-sidebar-toggle--primary:checked
-  ~ .sb-sidebar-overlay[for="sb-sidebar-toggle--primary"],
-#sb-sidebar-toggle--secondary:checked
-  ~ .sb-sidebar-overlay[for="sb-sidebar-toggle--secondary"] {
-  width: 100%;
-  height: 100%;
-  opacity: 1;
-  transition: width 0ms ease, height 0ms ease, opacity 250ms ease;
-}
-
-#sb-sidebar-toggle--primary:checked ~ .sb-container .sb-sidebar-primary {
-  left: 0;
-}
-#sb-sidebar-toggle--secondary:checked ~ .sb-container .sb-sidebar-secondary {
-  right: 0;
-}
-
-/* Full-width mode */
-.drop-secondary-sidebar-for-full-width-content
-  .hide-when-secondary-sidebar-shown {
-  display: none !important;
-}
-.drop-secondary-sidebar-for-full-width-content .sb-sidebar-secondary {
-  display: none !important;
-}
-
-/* Mobile views */
-.sb-page-width {
-  width: 100%;
-}
-
-.sb-article-container,
-.sb-footer-content__inner,
-.drop-secondary-sidebar-for-full-width-content .sb-article,
-.drop-secondary-sidebar-for-full-width-content .match-content-width {
-  width: 100vw;
-}
-
-.sb-article,
-.match-content-width {
-  padding: 0 1rem;
-  box-sizing: border-box;
-}
-
-@media (min-width: 32rem) {
-  .sb-article,
-  .match-content-width {
-    padding: 0 2rem;
-  }
-}
-
-/* Tablet views */
-@media (min-width: 42rem) {
-  .sb-article-container {
-    width: auto;
-  }
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 42rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 42rem;
-  }
-}
-@media (min-width: 46rem) {
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 46rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 46rem;
-  }
-}
-@media (min-width: 50rem) {
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 50rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 50rem;
-  }
-}
-
-/* Tablet views */
-@media (min-width: 59rem) {
-  .sb-sidebar-secondary {
-    position: static;
-  }
-  .hide-when-secondary-sidebar-shown {
-    display: none !important;
-  }
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 59rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 42rem;
-  }
-}
-@media (min-width: 63rem) {
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 63rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 46rem;
-  }
-}
-@media (min-width: 67rem) {
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 67rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 50rem;
-  }
-}
-
-/* Desktop views */
-@media (min-width: 76rem) {
-  .sb-sidebar-primary {
-    position: static;
-  }
-  .hide-when-primary-sidebar-shown {
-    display: none !important;
-  }
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 59rem;
-  }
-  .sb-article,
-  .match-content-width {
-    width: 42rem;
-  }
-}
-
-/* Full desktop views */
-@media (min-width: 80rem) {
-  .sb-article,
-  .match-content-width {
-    width: 46rem;
-  }
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 63rem;
-  }
-}
-
-@media (min-width: 84rem) {
-  .sb-article,
-  .match-content-width {
-    width: 50rem;
-  }
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 67rem;
-  }
-}
-
-@media (min-width: 88rem) {
-  .sb-footer-content__inner,
-  .drop-secondary-sidebar-for-full-width-content .sb-article,
-  .drop-secondary-sidebar-for-full-width-content .match-content-width {
-    width: 67rem;
-  }
-  .sb-page-width {
-    width: 88rem;
-  }
-}
diff --git a/docs/_build/html/_static/sphinx_highlight.js b/docs/_build/html/_static/sphinx_highlight.js
deleted file mode 100644
index 8a96c69a1..000000000
--- a/docs/_build/html/_static/sphinx_highlight.js
+++ /dev/null
@@ -1,154 +0,0 @@
-/* Highlighting utilities for Sphinx HTML documentation. */
-"use strict";
-
-const SPHINX_HIGHLIGHT_ENABLED = true
-
-/**
- * highlight a given string on a node by wrapping it in
- * span elements with the given class name.
- */
-const _highlight = (node, addItems, text, className) => {
-  if (node.nodeType === Node.TEXT_NODE) {
-    const val = node.nodeValue;
-    const parent = node.parentNode;
-    const pos = val.toLowerCase().indexOf(text);
-    if (
-      pos >= 0 &&
-      !parent.classList.contains(className) &&
-      !parent.classList.contains("nohighlight")
-    ) {
-      let span;
-
-      const closestNode = parent.closest("body, svg, foreignObject");
-      const isInSVG = closestNode && closestNode.matches("svg");
-      if (isInSVG) {
-        span = document.createElementNS("http://www.w3.org/2000/svg", "tspan");
-      } else {
-        span = document.createElement("span");
-        span.classList.add(className);
-      }
-
-      span.appendChild(document.createTextNode(val.substr(pos, text.length)));
-      const rest = document.createTextNode(val.substr(pos + text.length));
-      parent.insertBefore(
-        span,
-        parent.insertBefore(
-          rest,
-          node.nextSibling
-        )
-      );
-      node.nodeValue = val.substr(0, pos);
-      /* There may be more occurrences of search term in this node. So call this
-       * function recursively on the remaining fragment.
-       */
-      _highlight(rest, addItems, text, className);
-
-      if (isInSVG) {
-        const rect = document.createElementNS(
-          "http://www.w3.org/2000/svg",
-          "rect"
-        );
-        const bbox = parent.getBBox();
-        rect.x.baseVal.value = bbox.x;
-        rect.y.baseVal.value = bbox.y;
-        rect.width.baseVal.value = bbox.width;
-        rect.height.baseVal.value = bbox.height;
-        rect.setAttribute("class", className);
-        addItems.push({ parent: parent, target: rect });
-      }
-    }
-  } else if (node.matches && !node.matches("button, select, textarea")) {
-    node.childNodes.forEach((el) => _highlight(el, addItems, text, className));
-  }
-};
-const _highlightText = (thisNode, text, className) => {
-  let addItems = [];
-  _highlight(thisNode, addItems, text, className);
-  addItems.forEach((obj) =>
-    obj.parent.insertAdjacentElement("beforebegin", obj.target)
-  );
-};
-
-/**
- * Small JavaScript module for the documentation.
- */
-const SphinxHighlight = {
-
-  /**
-   * highlight the search words provided in localstorage in the text
-   */
-  highlightSearchWords: () => {
-    if (!SPHINX_HIGHLIGHT_ENABLED) return;  // bail if no highlight
-
-    // get and clear terms from localstorage
-    const url = new URL(window.location);
-    const highlight =
-        localStorage.getItem("sphinx_highlight_terms")
-        || url.searchParams.get("highlight")
-        || "";
-    localStorage.removeItem("sphinx_highlight_terms")
-    url.searchParams.delete("highlight");
-    window.history.replaceState({}, "", url);
-
-    // get individual terms from highlight string
-    const terms = highlight.toLowerCase().split(/\s+/).filter(x => x);
-    if (terms.length === 0) return; // nothing to do
-
-    // There should never be more than one element matching "div.body"
-    const divBody = document.querySelectorAll("div.body");
-    const body = divBody.length ? divBody[0] : document.querySelector("body");
-    window.setTimeout(() => {
-      terms.forEach((term) => _highlightText(body, term, "highlighted"));
-    }, 10);
-
-    const searchBox = document.getElementById("searchbox");
-    if (searchBox === null) return;
-    searchBox.appendChild(
-      document
-        .createRange()
-        .createContextualFragment(
-          '<p class="highlight-link">' +
-            '<a href="javascript:SphinxHighlight.hideSearchWords()">' +
-            _("Hide Search Matches") +
-            "</a></p>"
-        )
-    );
-  },
-
-  /**
-   * helper function to hide the search marks again
-   */
-  hideSearchWords: () => {
-    document
-      .querySelectorAll("#searchbox .highlight-link")
-      .forEach((el) => el.remove());
-    document
-      .querySelectorAll("span.highlighted")
-      .forEach((el) => el.classList.remove("highlighted"));
-    localStorage.removeItem("sphinx_highlight_terms")
-  },
-
-  initEscapeListener: () => {
-    // only install a listener if it is really needed
-    if (!DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS) return;
-
-    document.addEventListener("keydown", (event) => {
-      // bail for input elements
-      if (BLACKLISTED_KEY_CONTROL_ELEMENTS.has(document.activeElement.tagName)) return;
-      // bail with special keys
-      if (event.shiftKey || event.altKey || event.ctrlKey || event.metaKey) return;
-      if (DOCUMENTATION_OPTIONS.ENABLE_SEARCH_SHORTCUTS && (event.key === "Escape")) {
-        SphinxHighlight.hideSearchWords();
-        event.preventDefault();
-      }
-    });
-  },
-};
-
-_ready(() => {
-  /* Do not call highlightSearchWords() when we are on the search page.
-   * It will highlight words from the *previous* search query.
-   */
-  if (typeof Search === "undefined") SphinxHighlight.highlightSearchWords();
-  SphinxHighlight.initEscapeListener();
-});
diff --git a/docs/_build/html/_static/styles/furo-extensions.css b/docs/_build/html/_static/styles/furo-extensions.css
deleted file mode 100644
index bc447f228..000000000
--- a/docs/_build/html/_static/styles/furo-extensions.css
+++ /dev/null
@@ -1,2 +0,0 @@
-#furo-sidebar-ad-placement{padding:var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal)}#furo-sidebar-ad-placement .ethical-sidebar{background:var(--color-background-secondary);border:none;box-shadow:none}#furo-sidebar-ad-placement .ethical-sidebar:hover{background:var(--color-background-hover)}#furo-sidebar-ad-placement .ethical-sidebar a{color:var(--color-foreground-primary)}#furo-sidebar-ad-placement .ethical-callout a{color:var(--color-foreground-secondary)!important}#furo-readthedocs-versions{background:transparent;display:block;position:static;width:100%}#furo-readthedocs-versions .rst-versions{background:#1a1c1e}#furo-readthedocs-versions .rst-current-version{background:var(--color-sidebar-item-background);cursor:unset}#furo-readthedocs-versions .rst-current-version:hover{background:var(--color-sidebar-item-background)}#furo-readthedocs-versions .rst-current-version .fa-book{color:var(--color-foreground-primary)}#furo-readthedocs-versions>.rst-other-versions{padding:0}#furo-readthedocs-versions>.rst-other-versions small{opacity:1}#furo-readthedocs-versions .injected .rst-versions{position:unset}#furo-readthedocs-versions:focus-within,#furo-readthedocs-versions:hover{box-shadow:0 0 0 1px var(--color-sidebar-background-border)}#furo-readthedocs-versions:focus-within .rst-current-version,#furo-readthedocs-versions:hover .rst-current-version{background:#1a1c1e;font-size:inherit;height:auto;line-height:inherit;padding:12px;text-align:right}#furo-readthedocs-versions:focus-within .rst-current-version .fa-book,#furo-readthedocs-versions:hover .rst-current-version .fa-book{color:#fff;float:left}#furo-readthedocs-versions:focus-within .fa-caret-down,#furo-readthedocs-versions:hover .fa-caret-down{display:none}#furo-readthedocs-versions:focus-within .injected,#furo-readthedocs-versions:focus-within .rst-current-version,#furo-readthedocs-versions:focus-within .rst-other-versions,#furo-readthedocs-versions:hover .injected,#furo-readthedocs-versions:hover .rst-current-version,#furo-readthedocs-versions:hover .rst-other-versions{display:block}#furo-readthedocs-versions:focus-within>.rst-current-version,#furo-readthedocs-versions:hover>.rst-current-version{display:none}.highlight:hover button.copybtn{color:var(--color-code-foreground)}.highlight button.copybtn{align-items:center;background-color:var(--color-code-background);border:none;color:var(--color-background-item);cursor:pointer;height:1.25em;opacity:1;right:.5rem;top:.625rem;transition:color .3s,opacity .3s;width:1.25em}.highlight button.copybtn:hover{background-color:var(--color-code-background);color:var(--color-brand-content)}.highlight button.copybtn:after{background-color:transparent;color:var(--color-code-foreground);display:none}.highlight button.copybtn.success{color:#22863a;transition:color 0ms}.highlight button.copybtn.success:after{display:block}.highlight button.copybtn svg{padding:0}body{--sd-color-primary:var(--color-brand-primary);--sd-color-primary-highlight:var(--color-brand-content);--sd-color-primary-text:var(--color-background-primary);--sd-color-shadow:rgba(0,0,0,.05);--sd-color-card-border:var(--color-card-border);--sd-color-card-border-hover:var(--color-brand-content);--sd-color-card-background:var(--color-card-background);--sd-color-card-text:var(--color-foreground-primary);--sd-color-card-header:var(--color-card-marginals-background);--sd-color-card-footer:var(--color-card-marginals-background);--sd-color-tabs-label-active:var(--color-brand-content);--sd-color-tabs-label-hover:var(--color-foreground-muted);--sd-color-tabs-label-inactive:var(--color-foreground-muted);--sd-color-tabs-underline-active:var(--color-brand-content);--sd-color-tabs-underline-hover:var(--color-foreground-border);--sd-color-tabs-underline-inactive:var(--color-background-border);--sd-color-tabs-overline:var(--color-background-border);--sd-color-tabs-underline:var(--color-background-border)}.sd-tab-content{box-shadow:0 -2px var(--sd-color-tabs-overline),0 1px var(--sd-color-tabs-underline)}.sd-card{box-shadow:0 .1rem .25rem var(--sd-color-shadow),0 0 .0625rem rgba(0,0,0,.1)}.sd-shadow-sm{box-shadow:0 .1rem .25rem var(--sd-color-shadow),0 0 .0625rem rgba(0,0,0,.1)!important}.sd-shadow-md{box-shadow:0 .3rem .75rem var(--sd-color-shadow),0 0 .0625rem rgba(0,0,0,.1)!important}.sd-shadow-lg{box-shadow:0 .6rem 1.5rem var(--sd-color-shadow),0 0 .0625rem rgba(0,0,0,.1)!important}.sd-card-hover:hover{transform:none}.sd-cards-carousel{gap:.25rem;padding:.25rem}body{--tabs--label-text:var(--color-foreground-muted);--tabs--label-text--hover:var(--color-foreground-muted);--tabs--label-text--active:var(--color-brand-content);--tabs--label-text--active--hover:var(--color-brand-content);--tabs--label-background:transparent;--tabs--label-background--hover:transparent;--tabs--label-background--active:transparent;--tabs--label-background--active--hover:transparent;--tabs--padding-x:0.25em;--tabs--margin-x:1em;--tabs--border:var(--color-background-border);--tabs--label-border:transparent;--tabs--label-border--hover:var(--color-foreground-muted);--tabs--label-border--active:var(--color-brand-content);--tabs--label-border--active--hover:var(--color-brand-content)}[role=main] .container{max-width:none;padding-left:0;padding-right:0}.shadow.docutils{border:none;box-shadow:0 .2rem .5rem rgba(0,0,0,.05),0 0 .0625rem rgba(0,0,0,.1)!important}.sphinx-bs .card{background-color:var(--color-background-secondary);color:var(--color-foreground)}
-/*# sourceMappingURL=furo-extensions.css.map*/
\ No newline at end of file
diff --git a/docs/_build/html/_static/styles/furo-extensions.css.map b/docs/_build/html/_static/styles/furo-extensions.css.map
deleted file mode 100644
index 9ba5637f9..000000000
--- a/docs/_build/html/_static/styles/furo-extensions.css.map
+++ /dev/null
@@ -1 +0,0 @@
-{"version":3,"file":"styles/furo-extensions.css","mappings":"AAGA,2BACE,oFACA,4CAKE,6CAHA,YACA,eAEA,CACA,kDACE,yCAEF,8CACE,sCAEJ,8CACE,kDAEJ,2BAGE,uBACA,cAHA,gBACA,UAEA,CAGA,yCACE,mBAEF,gDAEE,gDADA,YACA,CACA,sDACE,gDACF,yDACE,sCAEJ,+CACE,UACA,qDACE,UAGF,mDACE,eAEJ,yEAEE,4DAEA,mHASE,mBAPA,kBAEA,YADA,oBAGA,aADA,gBAIA,CAEA,qIAEE,WADA,UACA,CAEJ,uGACE,aAEF,iUAGE,cAEF,mHACE,aC1EJ,gCACE,mCAEF,0BAKE,mBAUA,8CACA,YAFA,mCAKA,eAZA,cALA,UASA,YADA,YAYA,iCAdA,YAcA,CAEA,gCAEE,8CADA,gCACA,CAEF,gCAGE,6BADA,mCADA,YAEA,CAEF,kCAEE,cADA,oBACA,CACA,wCACE,cAEJ,8BACE,UC5CN,KAEE,6CAA8C,CAC9C,uDAAwD,CACxD,uDAAwD,CAGxD,iCAAsC,CAGtC,+CAAgD,CAChD,uDAAwD,CACxD,uDAAwD,CACxD,oDAAqD,CACrD,6DAA8D,CAC9D,6DAA8D,CAG9D,uDAAwD,CACxD,yDAA0D,CAC1D,4DAA6D,CAC7D,2DAA4D,CAC5D,8DAA+D,CAC/D,iEAAkE,CAClE,uDAAwD,CACxD,wDAAyD,CAG3D,gBACE,qFAGF,SACE,6EAEF,cACE,uFAEF,cACE,uFAEF,cACE,uFAGF,qBACE,eAEF,mBACE,WACA,eChDF,KACE,gDAAiD,CACjD,uDAAwD,CACxD,qDAAsD,CACtD,4DAA6D,CAC7D,oCAAqC,CACrC,2CAA4C,CAC5C,4CAA6C,CAC7C,mDAAoD,CACpD,wBAAyB,CACzB,oBAAqB,CACrB,6CAA8C,CAC9C,gCAAiC,CACjC,yDAA0D,CAC1D,uDAAwD,CACxD,8DAA+D,CCbjE,uBACE,eACA,eACA,gBAGF,iBACE,YACA,+EAGF,iBACE,mDACA","sources":["webpack:///./src/furo/assets/styles/extensions/_readthedocs.sass","webpack:///./src/furo/assets/styles/extensions/_copybutton.sass","webpack:///./src/furo/assets/styles/extensions/_sphinx-design.sass","webpack:///./src/furo/assets/styles/extensions/_sphinx-inline-tabs.sass","webpack:///./src/furo/assets/styles/extensions/_sphinx-panels.sass"],"sourcesContent":["// This file contains the styles used for tweaking how ReadTheDoc's embedded\n// contents would show up inside the theme.\n\n#furo-sidebar-ad-placement\n  padding: var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal)\n  .ethical-sidebar\n    // Remove the border and box-shadow.\n    border: none\n    box-shadow: none\n    // Manage the background colors.\n    background: var(--color-background-secondary)\n    &:hover\n      background: var(--color-background-hover)\n    // Ensure the text is legible.\n    a\n      color: var(--color-foreground-primary)\n\n  .ethical-callout a\n    color: var(--color-foreground-secondary) !important\n\n#furo-readthedocs-versions\n  position: static\n  width: 100%\n  background: transparent\n  display: block\n\n  // Make the background color fit with the theme's aesthetic.\n  .rst-versions\n    background: rgb(26, 28, 30)\n\n  .rst-current-version\n    cursor: unset\n    background: var(--color-sidebar-item-background)\n    &:hover\n      background: var(--color-sidebar-item-background)\n    .fa-book\n      color: var(--color-foreground-primary)\n\n  > .rst-other-versions\n    padding: 0\n    small\n      opacity: 1\n\n  .injected\n    .rst-versions\n      position: unset\n\n  &:hover,\n  &:focus-within\n    box-shadow: 0 0 0 1px var(--color-sidebar-background-border)\n\n    .rst-current-version\n      // Undo the tweaks done in RTD's CSS\n      font-size: inherit\n      line-height: inherit\n      height: auto\n      text-align: right\n      padding: 12px\n\n      // Match the rest of the body\n      background: #1a1c1e\n\n      .fa-book\n        float: left\n        color: white\n\n    .fa-caret-down\n      display: none\n\n    .rst-current-version,\n    .rst-other-versions,\n    .injected\n      display: block\n\n    > .rst-current-version\n      display: none\n",".highlight\n  &:hover button.copybtn\n    color: var(--color-code-foreground)\n\n  button.copybtn\n    // Make it visible\n    opacity: 1\n\n    // Align things correctly\n    align-items: center\n\n    height: 1.25em\n    width: 1.25em\n\n    top: 0.625rem // $code-spacing-vertical\n    right: 0.5rem\n\n    // Make it look better\n    color: var(--color-background-item)\n    background-color: var(--color-code-background)\n    border: none\n\n    // Change to cursor to make it obvious that you can click on it\n    cursor: pointer\n\n    // Transition smoothly, for aesthetics\n    transition: color 300ms, opacity 300ms\n\n    &:hover\n      color: var(--color-brand-content)\n      background-color: var(--color-code-background)\n\n    &::after\n      display: none\n      color: var(--color-code-foreground)\n      background-color: transparent\n\n    &.success\n      transition: color 0ms\n      color: #22863a\n      &::after\n        display: block\n\n    svg\n      padding: 0\n","body\n  // Colors\n  --sd-color-primary: var(--color-brand-primary)\n  --sd-color-primary-highlight: var(--color-brand-content)\n  --sd-color-primary-text: var(--color-background-primary)\n\n  // Shadows\n  --sd-color-shadow: rgba(0, 0, 0, 0.05)\n\n  // Cards\n  --sd-color-card-border: var(--color-card-border)\n  --sd-color-card-border-hover: var(--color-brand-content)\n  --sd-color-card-background: var(--color-card-background)\n  --sd-color-card-text: var(--color-foreground-primary)\n  --sd-color-card-header: var(--color-card-marginals-background)\n  --sd-color-card-footer: var(--color-card-marginals-background)\n\n  // Tabs\n  --sd-color-tabs-label-active: var(--color-brand-content)\n  --sd-color-tabs-label-hover: var(--color-foreground-muted)\n  --sd-color-tabs-label-inactive: var(--color-foreground-muted)\n  --sd-color-tabs-underline-active: var(--color-brand-content)\n  --sd-color-tabs-underline-hover: var(--color-foreground-border)\n  --sd-color-tabs-underline-inactive: var(--color-background-border)\n  --sd-color-tabs-overline: var(--color-background-border)\n  --sd-color-tabs-underline: var(--color-background-border)\n\n// Tabs\n.sd-tab-content\n  box-shadow: 0 -2px var(--sd-color-tabs-overline), 0 1px var(--sd-color-tabs-underline)\n\n// Shadows\n.sd-card  // Have a shadow by default\n  box-shadow: 0 0.1rem 0.25rem var(--sd-color-shadow), 0 0 0.0625rem rgba(0, 0, 0, 0.1)\n\n.sd-shadow-sm\n  box-shadow: 0 0.1rem 0.25rem var(--sd-color-shadow), 0 0 0.0625rem rgba(0, 0, 0, 0.1) !important\n\n.sd-shadow-md\n  box-shadow: 0 0.3rem 0.75rem var(--sd-color-shadow), 0 0 0.0625rem rgba(0, 0, 0, 0.1) !important\n\n.sd-shadow-lg\n  box-shadow: 0 0.6rem 1.5rem var(--sd-color-shadow), 0 0 0.0625rem rgba(0, 0, 0, 0.1) !important\n\n// Cards\n.sd-card-hover:hover  // Don't change scale on hover\n  transform: none\n\n.sd-cards-carousel  // Have a bit of gap in the carousel by default\n  gap: 0.25rem\n  padding: 0.25rem\n","// This file contains styles to tweak sphinx-inline-tabs to work well with Furo.\n\nbody\n  --tabs--label-text: var(--color-foreground-muted)\n  --tabs--label-text--hover: var(--color-foreground-muted)\n  --tabs--label-text--active: var(--color-brand-content)\n  --tabs--label-text--active--hover: var(--color-brand-content)\n  --tabs--label-background: transparent\n  --tabs--label-background--hover: transparent\n  --tabs--label-background--active: transparent\n  --tabs--label-background--active--hover: transparent\n  --tabs--padding-x: 0.25em\n  --tabs--margin-x: 1em\n  --tabs--border: var(--color-background-border)\n  --tabs--label-border: transparent\n  --tabs--label-border--hover: var(--color-foreground-muted)\n  --tabs--label-border--active: var(--color-brand-content)\n  --tabs--label-border--active--hover: var(--color-brand-content)\n","// This file contains styles to tweak sphinx-panels to work well with Furo.\n\n// sphinx-panels includes Bootstrap 4, which uses .container which can conflict\n// with docutils' `.. container::` directive.\n[role=\"main\"] .container\n  max-width: initial\n  padding-left: initial\n  padding-right: initial\n\n// Make the panels look nicer!\n.shadow.docutils\n  border: none\n  box-shadow: 0 0.2rem 0.5rem rgba(0, 0, 0, 0.05), 0 0 0.0625rem rgba(0, 0, 0, 0.1) !important\n\n// Make panel colors respond to dark mode\n.sphinx-bs .card\n  background-color: var(--color-background-secondary)\n  color: var(--color-foreground)\n"],"names":[],"sourceRoot":""}
\ No newline at end of file
diff --git a/docs/_build/html/_static/styles/furo.css b/docs/_build/html/_static/styles/furo.css
deleted file mode 100644
index 3d29a218f..000000000
--- a/docs/_build/html/_static/styles/furo.css
+++ /dev/null
@@ -1,2 +0,0 @@
-/*! normalize.css v8.0.1 | MIT License | github.com/necolas/normalize.css */html{-webkit-text-size-adjust:100%;line-height:1.15}body{margin:0}main{display:block}h1{font-size:2em;margin:.67em 0}hr{box-sizing:content-box;height:0;overflow:visible}pre{font-family:monospace,monospace;font-size:1em}a{background-color:transparent}abbr[title]{border-bottom:none;text-decoration:underline;text-decoration:underline dotted}b,strong{font-weight:bolder}code,kbd,samp{font-family:monospace,monospace;font-size:1em}sub,sup{font-size:75%;line-height:0;position:relative;vertical-align:baseline}sub{bottom:-.25em}sup{top:-.5em}img{border-style:none}button,input,optgroup,select,textarea{font-family:inherit;font-size:100%;line-height:1.15;margin:0}button,input{overflow:visible}button,select{text-transform:none}[type=button],[type=reset],[type=submit],button{-webkit-appearance:button}[type=button]::-moz-focus-inner,[type=reset]::-moz-focus-inner,[type=submit]::-moz-focus-inner,button::-moz-focus-inner{border-style:none;padding:0}[type=button]:-moz-focusring,[type=reset]:-moz-focusring,[type=submit]:-moz-focusring,button:-moz-focusring{outline:1px dotted ButtonText}fieldset{padding:.35em .75em .625em}legend{box-sizing:border-box;color:inherit;display:table;max-width:100%;padding:0;white-space:normal}progress{vertical-align:baseline}textarea{overflow:auto}[type=checkbox],[type=radio]{box-sizing:border-box;padding:0}[type=number]::-webkit-inner-spin-button,[type=number]::-webkit-outer-spin-button{height:auto}[type=search]{-webkit-appearance:textfield;outline-offset:-2px}[type=search]::-webkit-search-decoration{-webkit-appearance:none}::-webkit-file-upload-button{-webkit-appearance:button;font:inherit}details{display:block}summary{display:list-item}[hidden],template{display:none}@media print{.content-icon-container,.headerlink,.mobile-header,.related-pages{display:none!important}.highlight{border:.1pt solid var(--color-foreground-border)}a,blockquote,dl,ol,pre,table,ul{page-break-inside:avoid}caption,figure,h1,h2,h3,h4,h5,h6,img{page-break-after:avoid;page-break-inside:avoid}dl,ol,ul{page-break-before:avoid}}.visually-hidden{clip:rect(0,0,0,0)!important;border:0!important;height:1px!important;margin:-1px!important;overflow:hidden!important;padding:0!important;position:absolute!important;white-space:nowrap!important;width:1px!important}:-moz-focusring{outline:auto}body{--font-stack:-apple-system,BlinkMacSystemFont,Segoe UI,Helvetica,Arial,sans-serif,Apple Color Emoji,Segoe UI Emoji;--font-stack--monospace:"SFMono-Regular",Menlo,Consolas,Monaco,Liberation Mono,Lucida Console,monospace;--font-size--normal:100%;--font-size--small:87.5%;--font-size--small--2:81.25%;--font-size--small--3:75%;--font-size--small--4:62.5%;--sidebar-caption-font-size:var(--font-size--small--2);--sidebar-item-font-size:var(--font-size--small);--sidebar-search-input-font-size:var(--font-size--small);--toc-font-size:var(--font-size--small--3);--toc-font-size--mobile:var(--font-size--normal);--toc-title-font-size:var(--font-size--small--4);--admonition-font-size:0.8125rem;--admonition-title-font-size:0.8125rem;--code-font-size:var(--font-size--small--2);--api-font-size:var(--font-size--small);--header-height:calc(var(--sidebar-item-line-height) + var(--sidebar-item-spacing-vertical)*4);--header-padding:0.5rem;--sidebar-tree-space-above:1.5rem;--sidebar-caption-space-above:1rem;--sidebar-item-line-height:1rem;--sidebar-item-spacing-vertical:0.5rem;--sidebar-item-spacing-horizontal:1rem;--sidebar-item-height:calc(var(--sidebar-item-line-height) + var(--sidebar-item-spacing-vertical)*2);--sidebar-expander-width:var(--sidebar-item-height);--sidebar-search-space-above:0.5rem;--sidebar-search-input-spacing-vertical:0.5rem;--sidebar-search-input-spacing-horizontal:0.5rem;--sidebar-search-input-height:1rem;--sidebar-search-icon-size:var(--sidebar-search-input-height);--toc-title-padding:0.25rem 0;--toc-spacing-vertical:1.5rem;--toc-spacing-horizontal:1.5rem;--toc-item-spacing-vertical:0.4rem;--toc-item-spacing-horizontal:1rem;--icon-search:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" stroke-width="1.5" stroke="currentColor" fill="none" stroke-linecap="round" stroke-linejoin="round"><path d="M0 0h24v24H0z" stroke="none"/><circle cx="10" cy="10" r="7"/><path d="m21 21-6-6"/></svg>');--icon-pencil:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M20.71 7.04c.39-.39.39-1.04 0-1.41l-2.34-2.34c-.37-.39-1.02-.39-1.41 0l-1.84 1.83 3.75 3.75M3 17.25V21h3.75L17.81 9.93l-3.75-3.75L3 17.25z"/></svg>');--icon-abstract:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M4 5h16v2H4V5m0 4h16v2H4V9m0 4h16v2H4v-2m0 4h10v2H4v-2z"/></svg>');--icon-info:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M13 9h-2V7h2m0 10h-2v-6h2m-1-9A10 10 0 0 0 2 12a10 10 0 0 0 10 10 10 10 0 0 0 10-10A10 10 0 0 0 12 2z"/></svg>');--icon-flame:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M17.55 11.2c-.23-.3-.5-.56-.76-.82-.65-.6-1.4-1.03-2.03-1.66C13.3 7.26 13 4.85 13.91 3c-.91.23-1.75.75-2.45 1.32-2.54 2.08-3.54 5.75-2.34 8.9.04.1.08.2.08.33 0 .22-.15.42-.35.5-.22.1-.46.04-.64-.12a.83.83 0 0 1-.15-.17c-1.1-1.43-1.28-3.48-.53-5.12C5.89 10 5 12.3 5.14 14.47c.04.5.1 1 .27 1.5.14.6.4 1.2.72 1.73 1.04 1.73 2.87 2.97 4.84 3.22 2.1.27 4.35-.12 5.96-1.6 1.8-1.66 2.45-4.32 1.5-6.6l-.13-.26c-.2-.46-.47-.87-.8-1.25l.05-.01m-3.1 6.3c-.28.24-.73.5-1.08.6-1.1.4-2.2-.16-2.87-.82 1.19-.28 1.89-1.16 2.09-2.05.17-.8-.14-1.46-.27-2.23-.12-.74-.1-1.37.18-2.06.17.38.37.76.6 1.06.76 1 1.95 1.44 2.2 2.8.04.14.06.28.06.43.03.82-.32 1.72-.92 2.27h.01z"/></svg>');--icon-question:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m15.07 11.25-.9.92C13.45 12.89 13 13.5 13 15h-2v-.5c0-1.11.45-2.11 1.17-2.83l1.24-1.26c.37-.36.59-.86.59-1.41a2 2 0 0 0-2-2 2 2 0 0 0-2 2H8a4 4 0 0 1 4-4 4 4 0 0 1 4 4 3.2 3.2 0 0 1-.93 2.25M13 19h-2v-2h2M12 2A10 10 0 0 0 2 12a10 10 0 0 0 10 10 10 10 0 0 0 10-10c0-5.53-4.5-10-10-10z"/></svg>');--icon-warning:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M13 14h-2v-4h2m0 8h-2v-2h2M1 21h22L12 2 1 21z"/></svg>');--icon-failure:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M12 2c5.53 0 10 4.47 10 10s-4.47 10-10 10S2 17.53 2 12 6.47 2 12 2m3.59 5L12 10.59 8.41 7 7 8.41 10.59 12 7 15.59 8.41 17 12 13.41 15.59 17 17 15.59 13.41 12 17 8.41 15.59 7z"/></svg>');--icon-spark:url('data:image/svg+xml;charset=utf-8,<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="m11.5 20 4.86-9.73H13V4l-5 9.73h3.5V20M12 2c2.75 0 5.1 1 7.05 2.95C21 6.9 22 9.25 22 12s-1 5.1-2.95 7.05C17.1 21 14.75 22 12 22s-5.1-1-7.05-2.95C3 17.1 2 14.75 2 12s1-5.1 2.95-7.05C6.9 3 9.25 2 12 2z"/></svg>');--color-admonition-title--caution:#ff9100;--color-admonition-title-background--caution:rgba(255,145,0,.2);--color-admonition-title--warning:#ff9100;--color-admonition-title-background--warning:rgba(255,145,0,.2);--color-admonition-title--danger:#ff5252;--color-admonition-title-background--danger:rgba(255,82,82,.2);--color-admonition-title--attention:#ff5252;--color-admonition-title-background--attention:rgba(255,82,82,.2);--color-admonition-title--error:#ff5252;--color-admonition-title-background--error:rgba(255,82,82,.2);--color-admonition-title--hint:#00c852;--color-admonition-title-background--hint:rgba(0,200,82,.2);--color-admonition-title--tip:#00c852;--color-admonition-title-background--tip:rgba(0,200,82,.2);--color-admonition-title--important:#00bfa5;--color-admonition-title-background--important:rgba(0,191,165,.2);--color-admonition-title--note:#00b0ff;--color-admonition-title-background--note:rgba(0,176,255,.2);--color-admonition-title--seealso:#448aff;--color-admonition-title-background--seealso:rgba(68,138,255,.2);--color-admonition-title--admonition-todo:grey;--color-admonition-title-background--admonition-todo:hsla(0,0%,50%,.2);--color-admonition-title:#651fff;--color-admonition-title-background:rgba(101,31,255,.2);--icon-admonition-default:var(--icon-abstract);--color-topic-title:#14b8a6;--color-topic-title-background:rgba(20,184,166,.2);--icon-topic-default:var(--icon-pencil);--color-problematic:#b30000;--color-foreground-primary:#000;--color-foreground-secondary:#5a5c63;--color-foreground-muted:#646776;--color-foreground-border:#878787;--color-background-primary:#fff;--color-background-secondary:#f8f9fb;--color-background-hover:#efeff4;--color-background-hover--transparent:#efeff400;--color-background-border:#eeebee;--color-background-item:#ccc;--color-announcement-background:#000000dd;--color-announcement-text:#eeebee;--color-brand-primary:#2962ff;--color-brand-content:#2a5adf;--color-api-background:var(--color-background-hover--transparent);--color-api-background-hover:var(--color-background-hover);--color-api-overall:var(--color-foreground-secondary);--color-api-name:var(--color-problematic);--color-api-pre-name:var(--color-problematic);--color-api-paren:var(--color-foreground-secondary);--color-api-keyword:var(--color-foreground-primary);--color-highlight-on-target:#ffc;--color-inline-code-background:var(--color-background-secondary);--color-highlighted-background:#def;--color-highlighted-text:var(--color-foreground-primary);--color-guilabel-background:#ddeeff80;--color-guilabel-border:#bedaf580;--color-guilabel-text:var(--color-foreground-primary);--color-admonition-background:transparent;--color-table-header-background:var(--color-background-secondary);--color-table-border:var(--color-background-border);--color-card-border:var(--color-background-secondary);--color-card-background:transparent;--color-card-marginals-background:var(--color-background-secondary);--color-header-background:var(--color-background-primary);--color-header-border:var(--color-background-border);--color-header-text:var(--color-foreground-primary);--color-sidebar-background:var(--color-background-secondary);--color-sidebar-background-border:var(--color-background-border);--color-sidebar-brand-text:var(--color-foreground-primary);--color-sidebar-caption-text:var(--color-foreground-muted);--color-sidebar-link-text:var(--color-foreground-secondary);--color-sidebar-link-text--top-level:var(--color-brand-primary);--color-sidebar-item-background:var(--color-sidebar-background);--color-sidebar-item-background--current:var( --color-sidebar-item-background );--color-sidebar-item-background--hover:linear-gradient(90deg,var(--color-background-hover--transparent) 0%,var(--color-background-hover) var(--sidebar-item-spacing-horizontal),var(--color-background-hover) 100%);--color-sidebar-item-expander-background:transparent;--color-sidebar-item-expander-background--hover:var( --color-background-hover );--color-sidebar-search-text:var(--color-foreground-primary);--color-sidebar-search-background:var(--color-background-secondary);--color-sidebar-search-background--focus:var(--color-background-primary);--color-sidebar-search-border:var(--color-background-border);--color-sidebar-search-icon:var(--color-foreground-muted);--color-toc-background:var(--color-background-primary);--color-toc-title-text:var(--color-foreground-muted);--color-toc-item-text:var(--color-foreground-secondary);--color-toc-item-text--hover:var(--color-foreground-primary);--color-toc-item-text--active:var(--color-brand-primary);--color-content-foreground:var(--color-foreground-primary);--color-content-background:transparent;--color-link:var(--color-brand-content);--color-link--hover:var(--color-brand-content);--color-link-underline:var(--color-background-border);--color-link-underline--hover:var(--color-foreground-border)}.only-light{display:block!important}html body .only-dark{display:none!important}@media not print{body[data-theme=dark]{--color-problematic:#ee5151;--color-foreground-primary:#ffffffcc;--color-foreground-secondary:#9ca0a5;--color-foreground-muted:#81868d;--color-foreground-border:#666;--color-background-primary:#131416;--color-background-secondary:#1a1c1e;--color-background-hover:#1e2124;--color-background-hover--transparent:#1e212400;--color-background-border:#303335;--color-background-item:#444;--color-announcement-background:#000000dd;--color-announcement-text:#eeebee;--color-brand-primary:#2b8cee;--color-brand-content:#368ce2;--color-highlighted-background:#083563;--color-guilabel-background:#08356380;--color-guilabel-border:#13395f80;--color-api-keyword:var(--color-foreground-secondary);--color-highlight-on-target:#330;--color-admonition-background:#18181a;--color-card-border:var(--color-background-secondary);--color-card-background:#18181a;--color-card-marginals-background:var(--color-background-hover)}html body[data-theme=dark] .only-light{display:none!important}body[data-theme=dark] .only-dark{display:block!important}@media(prefers-color-scheme:dark){body:not([data-theme=light]){--color-problematic:#ee5151;--color-foreground-primary:#ffffffcc;--color-foreground-secondary:#9ca0a5;--color-foreground-muted:#81868d;--color-foreground-border:#666;--color-background-primary:#131416;--color-background-secondary:#1a1c1e;--color-background-hover:#1e2124;--color-background-hover--transparent:#1e212400;--color-background-border:#303335;--color-background-item:#444;--color-announcement-background:#000000dd;--color-announcement-text:#eeebee;--color-brand-primary:#2b8cee;--color-brand-content:#368ce2;--color-highlighted-background:#083563;--color-guilabel-background:#08356380;--color-guilabel-border:#13395f80;--color-api-keyword:var(--color-foreground-secondary);--color-highlight-on-target:#330;--color-admonition-background:#18181a;--color-card-border:var(--color-background-secondary);--color-card-background:#18181a;--color-card-marginals-background:var(--color-background-hover)}html body:not([data-theme=light]) .only-light{display:none!important}body:not([data-theme=light]) .only-dark{display:block!important}}}body[data-theme=auto] .theme-toggle svg.theme-icon-when-auto,body[data-theme=dark] .theme-toggle svg.theme-icon-when-dark,body[data-theme=light] .theme-toggle svg.theme-icon-when-light{display:block}body{font-family:var(--font-stack)}code,kbd,pre,samp{font-family:var(--font-stack--monospace)}body{-webkit-font-smoothing:antialiased;-moz-osx-font-smoothing:grayscale}article{line-height:1.5}h1,h2,h3,h4,h5,h6{border-radius:.5rem;font-weight:700;line-height:1.25;margin:.5rem -.5rem;padding-left:.5rem;padding-right:.5rem}h1+p,h2+p,h3+p,h4+p,h5+p,h6+p{margin-top:0}h1{font-size:2.5em;margin-bottom:1rem}h1,h2{margin-top:1.75rem}h2{font-size:2em}h3{font-size:1.5em}h4{font-size:1.25em}h5{font-size:1.125em}h6{font-size:1em}small{font-size:80%;opacity:75%}p{margin-bottom:.75rem;margin-top:.5rem}hr.docutils{background-color:var(--color-background-border);border:0;height:1px;margin:2rem 0;padding:0}.centered{text-align:center}a{color:var(--color-link);text-decoration:underline;text-decoration-color:var(--color-link-underline)}a:hover{color:var(--color-link--hover);text-decoration-color:var(--color-link-underline--hover)}a.muted-link{color:inherit}a.muted-link:hover{color:var(--color-link);text-decoration-color:var(--color-link-underline--hover)}html{overflow-x:hidden;overflow-y:scroll;scroll-behavior:smooth}.sidebar-scroll,.toc-scroll,article[role=main] *{scrollbar-color:var(--color-foreground-border) transparent;scrollbar-width:thin}.sidebar-scroll::-webkit-scrollbar,.toc-scroll::-webkit-scrollbar,article[role=main] ::-webkit-scrollbar{height:.25rem;width:.25rem}.sidebar-scroll::-webkit-scrollbar-thumb,.toc-scroll::-webkit-scrollbar-thumb,article[role=main] ::-webkit-scrollbar-thumb{background-color:var(--color-foreground-border);border-radius:.125rem}body,html{background:var(--color-background-primary);color:var(--color-foreground-primary);height:100%}article{background:var(--color-content-background);color:var(--color-content-foreground);overflow-wrap:break-word}.page{display:flex;min-height:100%}.mobile-header{background-color:var(--color-header-background);border-bottom:1px solid var(--color-header-border);color:var(--color-header-text);display:none;height:var(--header-height);width:100%;z-index:10}.mobile-header.scrolled{border-bottom:none;box-shadow:0 0 .2rem rgba(0,0,0,.1),0 .2rem .4rem rgba(0,0,0,.2)}.mobile-header .header-center a{color:var(--color-header-text);text-decoration:none}.main{display:flex;flex:1}.sidebar-drawer{background:var(--color-sidebar-background);border-right:1px solid var(--color-sidebar-background-border);box-sizing:border-box;display:flex;justify-content:flex-end;min-width:15em;width:calc(50% - 26em)}.sidebar-container,.toc-drawer{box-sizing:border-box;width:15em}.toc-drawer{background:var(--color-toc-background);padding-right:1rem}.sidebar-sticky,.toc-sticky{display:flex;flex-direction:column;height:min(100%,100vh);height:100vh;position:sticky;top:0}.sidebar-scroll,.toc-scroll{flex-grow:1;flex-shrink:1;overflow:auto;scroll-behavior:smooth}.content{display:flex;flex-direction:column;justify-content:space-between;padding:0 3em;width:46em}.icon{display:inline-block;height:1rem;width:1rem}.icon svg{height:100%;width:100%}.announcement{align-items:center;background-color:var(--color-announcement-background);color:var(--color-announcement-text);display:flex;height:var(--header-height);overflow-x:auto}.announcement+.page{min-height:calc(100% - var(--header-height))}.announcement-content{box-sizing:border-box;min-width:100%;padding:.5rem;text-align:center;white-space:nowrap}.announcement-content a{color:var(--color-announcement-text);text-decoration-color:var(--color-announcement-text)}.announcement-content a:hover{color:var(--color-announcement-text);text-decoration-color:var(--color-link--hover)}.no-js .theme-toggle-container{display:none}.theme-toggle-container{vertical-align:middle}.theme-toggle{background:transparent;border:none;cursor:pointer;padding:0}.theme-toggle svg{color:var(--color-foreground-primary);display:none;height:1rem;vertical-align:middle;width:1rem}.theme-toggle-header{float:left;padding:1rem .5rem}.nav-overlay-icon,.toc-overlay-icon{cursor:pointer;display:none}.nav-overlay-icon .icon,.toc-overlay-icon .icon{color:var(--color-foreground-secondary);height:1rem;width:1rem}.nav-overlay-icon,.toc-header-icon{align-items:center;justify-content:center}.toc-content-icon{height:1.5rem;width:1.5rem}.content-icon-container{display:flex;float:right;gap:.5rem;margin-bottom:1rem;margin-left:1rem;margin-top:1.5rem}.content-icon-container .edit-this-page svg{color:inherit;height:1rem;width:1rem}.sidebar-toggle{display:none;position:absolute}.sidebar-toggle[name=__toc]{left:20px}.sidebar-toggle:checked{left:40px}.overlay{background-color:rgba(0,0,0,.54);height:0;opacity:0;position:fixed;top:0;transition:width 0ms,height 0ms,opacity .25s ease-out;width:0}.sidebar-overlay{z-index:20}.toc-overlay{z-index:40}.sidebar-drawer{transition:left .25s ease-in-out;z-index:30}.toc-drawer{transition:right .25s ease-in-out;z-index:50}#__navigation:checked~.sidebar-overlay{height:100%;opacity:1;width:100%}#__navigation:checked~.page .sidebar-drawer{left:0;top:0}#__toc:checked~.toc-overlay{height:100%;opacity:1;width:100%}#__toc:checked~.page .toc-drawer{right:0;top:0}.back-to-top{background:var(--color-background-primary);border-radius:1rem;box-shadow:0 .2rem .5rem rgba(0,0,0,.05),0 0 1px 0 hsla(220,9%,46%,.502);display:none;font-size:.8125rem;left:0;margin-left:50%;padding:.5rem .75rem .5rem .5rem;position:fixed;text-decoration:none;top:1rem;transform:translateX(-50%);z-index:10}.back-to-top svg{fill:currentColor;display:inline-block;height:1rem;width:1rem}.back-to-top span{margin-left:.25rem}.show-back-to-top .back-to-top{align-items:center;display:flex}@media(min-width:97em){html{font-size:110%}}@media(max-width:82em){.toc-content-icon{display:flex}.toc-drawer{border-left:1px solid var(--color-background-muted);height:100vh;position:fixed;right:-15em;top:0}.toc-tree{border-left:none;font-size:var(--toc-font-size--mobile)}.sidebar-drawer{width:calc(50% - 18.5em)}}@media(max-width:67em){.nav-overlay-icon{display:flex}.sidebar-drawer{height:100vh;left:-15em;position:fixed;top:0;width:15em}.toc-header-icon{display:flex}.theme-toggle-content,.toc-content-icon{display:none}.theme-toggle-header{display:block}.mobile-header{align-items:center;display:flex;justify-content:space-between;position:sticky;top:0}.mobile-header .header-left,.mobile-header .header-right{display:flex;height:var(--header-height);padding:0 var(--header-padding)}.mobile-header .header-left label,.mobile-header .header-right label{height:100%;-webkit-user-select:none;-moz-user-select:none;user-select:none;width:100%}.nav-overlay-icon .icon,.theme-toggle svg{height:1.25rem;width:1.25rem}:target{scroll-margin-top:var(--header-height)}.back-to-top{top:calc(var(--header-height) + .5rem)}.page{flex-direction:column;justify-content:center}.content{margin-left:auto;margin-right:auto}}@media(max-width:52em){.content{overflow-x:auto;width:100%}}@media(max-width:46em){.content{padding:0 1em}article aside.sidebar{float:none;margin:1rem 0;width:100%}}.admonition,.topic{background:var(--color-admonition-background);border-radius:.2rem;box-shadow:0 .2rem .5rem rgba(0,0,0,.05),0 0 .0625rem rgba(0,0,0,.1);font-size:var(--admonition-font-size);margin:1rem auto;overflow:hidden;padding:0 .5rem .5rem;page-break-inside:avoid}.admonition>:nth-child(2),.topic>:nth-child(2){margin-top:0}.admonition>:last-child,.topic>:last-child{margin-bottom:0}.admonition p.admonition-title,p.topic-title{font-size:var(--admonition-title-font-size);font-weight:500;line-height:1.3;margin:0 -.5rem .5rem;padding:.4rem .5rem .4rem 2rem;position:relative}.admonition p.admonition-title:before,p.topic-title:before{content:"";height:1rem;left:.5rem;position:absolute;width:1rem}p.admonition-title{background-color:var(--color-admonition-title-background)}p.admonition-title:before{background-color:var(--color-admonition-title);-webkit-mask-image:var(--icon-admonition-default);mask-image:var(--icon-admonition-default);-webkit-mask-repeat:no-repeat;mask-repeat:no-repeat}p.topic-title{background-color:var(--color-topic-title-background)}p.topic-title:before{background-color:var(--color-topic-title);-webkit-mask-image:var(--icon-topic-default);mask-image:var(--icon-topic-default);-webkit-mask-repeat:no-repeat;mask-repeat:no-repeat}.admonition{border-left:.2rem solid var(--color-admonition-title)}.admonition.caution{border-left-color:var(--color-admonition-title--caution)}.admonition.caution>.admonition-title{background-color:var(--color-admonition-title-background--caution)}.admonition.caution>.admonition-title:before{background-color:var(--color-admonition-title--caution);-webkit-mask-image:var(--icon-spark);mask-image:var(--icon-spark)}.admonition.warning{border-left-color:var(--color-admonition-title--warning)}.admonition.warning>.admonition-title{background-color:var(--color-admonition-title-background--warning)}.admonition.warning>.admonition-title:before{background-color:var(--color-admonition-title--warning);-webkit-mask-image:var(--icon-warning);mask-image:var(--icon-warning)}.admonition.danger{border-left-color:var(--color-admonition-title--danger)}.admonition.danger>.admonition-title{background-color:var(--color-admonition-title-background--danger)}.admonition.danger>.admonition-title:before{background-color:var(--color-admonition-title--danger);-webkit-mask-image:var(--icon-spark);mask-image:var(--icon-spark)}.admonition.attention{border-left-color:var(--color-admonition-title--attention)}.admonition.attention>.admonition-title{background-color:var(--color-admonition-title-background--attention)}.admonition.attention>.admonition-title:before{background-color:var(--color-admonition-title--attention);-webkit-mask-image:var(--icon-warning);mask-image:var(--icon-warning)}.admonition.error{border-left-color:var(--color-admonition-title--error)}.admonition.error>.admonition-title{background-color:var(--color-admonition-title-background--error)}.admonition.error>.admonition-title:before{background-color:var(--color-admonition-title--error);-webkit-mask-image:var(--icon-failure);mask-image:var(--icon-failure)}.admonition.hint{border-left-color:var(--color-admonition-title--hint)}.admonition.hint>.admonition-title{background-color:var(--color-admonition-title-background--hint)}.admonition.hint>.admonition-title:before{background-color:var(--color-admonition-title--hint);-webkit-mask-image:var(--icon-question);mask-image:var(--icon-question)}.admonition.tip{border-left-color:var(--color-admonition-title--tip)}.admonition.tip>.admonition-title{background-color:var(--color-admonition-title-background--tip)}.admonition.tip>.admonition-title:before{background-color:var(--color-admonition-title--tip);-webkit-mask-image:var(--icon-info);mask-image:var(--icon-info)}.admonition.important{border-left-color:var(--color-admonition-title--important)}.admonition.important>.admonition-title{background-color:var(--color-admonition-title-background--important)}.admonition.important>.admonition-title:before{background-color:var(--color-admonition-title--important);-webkit-mask-image:var(--icon-flame);mask-image:var(--icon-flame)}.admonition.note{border-left-color:var(--color-admonition-title--note)}.admonition.note>.admonition-title{background-color:var(--color-admonition-title-background--note)}.admonition.note>.admonition-title:before{background-color:var(--color-admonition-title--note);-webkit-mask-image:var(--icon-pencil);mask-image:var(--icon-pencil)}.admonition.seealso{border-left-color:var(--color-admonition-title--seealso)}.admonition.seealso>.admonition-title{background-color:var(--color-admonition-title-background--seealso)}.admonition.seealso>.admonition-title:before{background-color:var(--color-admonition-title--seealso);-webkit-mask-image:var(--icon-info);mask-image:var(--icon-info)}.admonition.admonition-todo{border-left-color:var(--color-admonition-title--admonition-todo)}.admonition.admonition-todo>.admonition-title{background-color:var(--color-admonition-title-background--admonition-todo)}.admonition.admonition-todo>.admonition-title:before{background-color:var(--color-admonition-title--admonition-todo);-webkit-mask-image:var(--icon-pencil);mask-image:var(--icon-pencil)}.admonition-todo>.admonition-title{text-transform:uppercase}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) dd{margin-left:2rem}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) dd>:first-child{margin-top:.125rem}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) .field-list,dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) dd>:last-child{margin-bottom:.75rem}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) .field-list>dt{font-size:var(--font-size--small);text-transform:uppercase}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) .field-list dd:empty{margin-bottom:.5rem}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) .field-list dd>ul{margin-left:-1.2rem}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) .field-list dd>ul>li>p:nth-child(2){margin-top:0}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple) .field-list dd>ul>li>p+p:last-child:empty{margin-bottom:0;margin-top:0}dl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple)>dt{color:var(--color-api-overall)}.sig:not(.sig-inline){background:var(--color-api-background);border-radius:.25rem;font-family:var(--font-stack--monospace);font-size:var(--api-font-size);font-weight:700;margin-left:-.25rem;margin-right:-.25rem;padding:.25rem .5rem .25rem 3em;text-indent:-2.5em;transition:background .1s ease-out}.sig:not(.sig-inline):hover{background:var(--color-api-background-hover)}.sig:not(.sig-inline) a.reference .viewcode-link{font-weight:400;width:3.5rem}em.property{font-style:normal}em.property:first-child{color:var(--color-api-keyword)}.sig-name{color:var(--color-api-name)}.sig-prename{color:var(--color-api-pre-name);font-weight:400}.sig-paren{color:var(--color-api-paren)}.sig-param{font-style:normal}.versionmodified{font-style:italic}div.deprecated p,div.versionadded p,div.versionchanged p{margin-bottom:.125rem;margin-top:.125rem}.viewcode-back,.viewcode-link{float:right;text-align:right}.line-block{margin-bottom:.75rem;margin-top:.5rem}.line-block .line-block{margin-bottom:0;margin-top:0;padding-left:1rem}.code-block-caption,article p.caption,table>caption{font-size:var(--font-size--small);text-align:center}.toctree-wrapper.compound .caption,.toctree-wrapper.compound :not(.caption)>.caption-text{font-size:var(--font-size--small);margin-bottom:0;text-align:initial;text-transform:uppercase}.toctree-wrapper.compound>ul{margin-bottom:0;margin-top:0}.sig-inline,code.literal{background:var(--color-inline-code-background);border-radius:.2em;font-size:var(--font-size--small--2);padding:.1em .2em}pre.literal-block .sig-inline,pre.literal-block code.literal{font-size:inherit;padding:0}p .sig-inline,p code.literal{border:1px solid var(--color-background-border)}.sig-inline{font-family:var(--font-stack--monospace)}div[class*=" highlight-"],div[class^=highlight-]{display:flex;margin:1em 0}div[class*=" highlight-"] .table-wrapper,div[class^=highlight-] .table-wrapper,pre{margin:0;padding:0}pre{overflow:auto}article[role=main] .highlight pre{line-height:1.5}.highlight pre,pre.literal-block{font-size:var(--code-font-size);padding:.625rem .875rem}pre.literal-block{background-color:var(--color-code-background);border-radius:.2rem;color:var(--color-code-foreground);margin-bottom:1rem;margin-top:1rem}.highlight{border-radius:.2rem;width:100%}.highlight .gp,.highlight span.linenos{pointer-events:none;-webkit-user-select:none;-moz-user-select:none;user-select:none}.highlight .hll{display:block;margin-left:-.875rem;margin-right:-.875rem;padding-left:.875rem;padding-right:.875rem}.code-block-caption{background-color:var(--color-code-background);border-bottom:1px solid;border-radius:.25rem;border-bottom-left-radius:0;border-bottom-right-radius:0;border-color:var(--color-background-border);color:var(--color-code-foreground);display:flex;font-weight:300;padding:.625rem .875rem}.code-block-caption+div[class]{margin-top:0}.code-block-caption+div[class] pre{border-top-left-radius:0;border-top-right-radius:0}.highlighttable{display:block;width:100%}.highlighttable tbody{display:block}.highlighttable tr{display:flex}.highlighttable td.linenos{background-color:var(--color-code-background);border-bottom-left-radius:.2rem;border-top-left-radius:.2rem;color:var(--color-code-foreground);padding:.625rem 0 .625rem .875rem}.highlighttable .linenodiv{box-shadow:-.0625rem 0 var(--color-foreground-border) inset;font-size:var(--code-font-size);padding-right:.875rem}.highlighttable td.code{display:block;flex:1;overflow:hidden;padding:0}.highlighttable td.code .highlight{border-bottom-left-radius:0;border-top-left-radius:0}.highlight span.linenos{box-shadow:-.0625rem 0 var(--color-foreground-border) inset;display:inline-block;margin-right:.875rem;padding-left:0;padding-right:.875rem}.footnote-reference{font-size:var(--font-size--small--4);vertical-align:super}dl.footnote.brackets{color:var(--color-foreground-secondary);display:grid;font-size:var(--font-size--small);grid-template-columns:max-content auto}dl.footnote.brackets dt{margin:0}dl.footnote.brackets dt>.fn-backref{margin-left:.25rem}dl.footnote.brackets dt:after{content:":"}dl.footnote.brackets dt .brackets:before{content:"["}dl.footnote.brackets dt .brackets:after{content:"]"}dl.footnote.brackets dd{margin:0;padding:0 1rem}aside.footnote{color:var(--color-foreground-secondary);font-size:var(--font-size--small)}aside.footnote>span,div.citation>span{float:left;font-weight:500;padding-right:.25rem}aside.footnote>p,div.citation>p{margin-left:2rem}img{box-sizing:border-box;height:auto;max-width:100%}article .figure,article figure{border-radius:.2rem;margin:0}article .figure :last-child,article figure :last-child{margin-bottom:0}article .align-left{clear:left;float:left;margin:0 1rem 1rem}article .align-right{clear:right;float:right;margin:0 1rem 1rem}article .align-center,article .align-default{display:block;margin-left:auto;margin-right:auto;text-align:center}article table.align-default{display:table;text-align:initial}.domainindex-jumpbox,.genindex-jumpbox{border-bottom:1px solid var(--color-background-border);border-top:1px solid var(--color-background-border);padding:.25rem}.domainindex-section h2,.genindex-section h2{margin-bottom:.5rem;margin-top:.75rem}.domainindex-section ul,.genindex-section ul{margin-bottom:0;margin-top:0}ol,ul{margin-bottom:1rem;margin-top:1rem;padding-left:1.2rem}ol li>p:first-child,ul li>p:first-child{margin-bottom:.25rem;margin-top:.25rem}ol li>p:last-child,ul li>p:last-child{margin-top:.25rem}ol li>ol,ol li>ul,ul li>ol,ul li>ul{margin-bottom:.5rem;margin-top:.5rem}ol.arabic{list-style:decimal}ol.loweralpha{list-style:lower-alpha}ol.upperalpha{list-style:upper-alpha}ol.lowerroman{list-style:lower-roman}ol.upperroman{list-style:upper-roman}.simple li>ol,.simple li>ul,.toctree-wrapper li>ol,.toctree-wrapper li>ul{margin-bottom:0;margin-top:0}.field-list dt,.option-list dt,dl.footnote dt,dl.glossary dt,dl.simple dt,dl:not([class]) dt{font-weight:500;margin-top:.25rem}.field-list dt+dt,.option-list dt+dt,dl.footnote dt+dt,dl.glossary dt+dt,dl.simple dt+dt,dl:not([class]) dt+dt{margin-top:0}.field-list dt .classifier:before,.option-list dt .classifier:before,dl.footnote dt .classifier:before,dl.glossary dt .classifier:before,dl.simple dt .classifier:before,dl:not([class]) dt .classifier:before{content:":";margin-left:.2rem;margin-right:.2rem}.field-list dd ul,.field-list dd>p:first-child,.option-list dd ul,.option-list dd>p:first-child,dl.footnote dd ul,dl.footnote dd>p:first-child,dl.glossary dd ul,dl.glossary dd>p:first-child,dl.simple dd ul,dl.simple dd>p:first-child,dl:not([class]) dd ul,dl:not([class]) dd>p:first-child{margin-top:.125rem}.field-list dd ul,.option-list dd ul,dl.footnote dd ul,dl.glossary dd ul,dl.simple dd ul,dl:not([class]) dd ul{margin-bottom:.125rem}.math-wrapper{overflow-x:auto;width:100%}div.math{position:relative;text-align:center}div.math .headerlink,div.math:focus .headerlink{display:none}div.math:hover .headerlink{display:inline-block}div.math span.eqno{position:absolute;right:.5rem;top:50%;transform:translateY(-50%);z-index:1}abbr[title]{cursor:help}.problematic{color:var(--color-problematic)}kbd:not(.compound){background-color:var(--color-background-secondary);border:1px solid var(--color-foreground-border);border-radius:.2rem;box-shadow:0 .0625rem 0 rgba(0,0,0,.2),inset 0 0 0 .125rem var(--color-background-primary);color:var(--color-foreground-primary);display:inline-block;font-size:var(--font-size--small--3);margin:0 .2rem;padding:0 .2rem;vertical-align:text-bottom}blockquote{background:var(--color-background-secondary);border-left:4px solid var(--color-background-border);margin-left:0;margin-right:0;padding:.5rem 1rem}blockquote .attribution{font-weight:600;text-align:right}blockquote.highlights,blockquote.pull-quote{font-size:1.25em}blockquote.epigraph,blockquote.pull-quote{border-left-width:0;border-radius:.5rem}blockquote.highlights{background:transparent;border-left-width:0}p .reference img{vertical-align:middle}p.rubric{font-size:1.125em;font-weight:700;line-height:1.25}dd p.rubric{font-size:var(--font-size--small);font-weight:inherit;line-height:inherit;text-transform:uppercase}article .sidebar{background-color:var(--color-background-secondary);border:1px solid var(--color-background-border);border-radius:.2rem;clear:right;float:right;margin-left:1rem;margin-right:0;width:30%}article .sidebar>*{padding-left:1rem;padding-right:1rem}article .sidebar>ol,article .sidebar>ul{padding-left:2.2rem}article .sidebar .sidebar-title{border-bottom:1px solid var(--color-background-border);font-weight:500;margin:0;padding:.5rem 1rem}.table-wrapper{margin-bottom:.5rem;margin-top:1rem;overflow-x:auto;padding:.2rem .2rem .75rem;width:100%}table.docutils{border-collapse:collapse;border-radius:.2rem;border-spacing:0;box-shadow:0 .2rem .5rem rgba(0,0,0,.05),0 0 .0625rem rgba(0,0,0,.1)}table.docutils th{background:var(--color-table-header-background)}table.docutils td,table.docutils th{border-bottom:1px solid var(--color-table-border);border-left:1px solid var(--color-table-border);border-right:1px solid var(--color-table-border);padding:0 .25rem}table.docutils td p,table.docutils th p{margin:.25rem}table.docutils td:first-child,table.docutils th:first-child{border-left:none}table.docutils td:last-child,table.docutils th:last-child{border-right:none}table.docutils td.text-left,table.docutils th.text-left{text-align:left}table.docutils td.text-right,table.docutils th.text-right{text-align:right}table.docutils td.text-center,table.docutils th.text-center{text-align:center}:target{scroll-margin-top:.5rem}@media(max-width:67em){:target{scroll-margin-top:calc(.5rem + var(--header-height))}section>span:target{scroll-margin-top:calc(.8rem + var(--header-height))}}.headerlink{font-weight:100;-webkit-user-select:none;-moz-user-select:none;user-select:none}.code-block-caption>.headerlink,dl dt>.headerlink,figcaption p>.headerlink,h1>.headerlink,h2>.headerlink,h3>.headerlink,h4>.headerlink,h5>.headerlink,h6>.headerlink,p.caption>.headerlink,table>caption>.headerlink{margin-left:.5rem;visibility:hidden}.code-block-caption:hover>.headerlink,dl dt:hover>.headerlink,figcaption p:hover>.headerlink,h1:hover>.headerlink,h2:hover>.headerlink,h3:hover>.headerlink,h4:hover>.headerlink,h5:hover>.headerlink,h6:hover>.headerlink,p.caption:hover>.headerlink,table>caption:hover>.headerlink{visibility:visible}.code-block-caption>.toc-backref,dl dt>.toc-backref,figcaption p>.toc-backref,h1>.toc-backref,h2>.toc-backref,h3>.toc-backref,h4>.toc-backref,h5>.toc-backref,h6>.toc-backref,p.caption>.toc-backref,table>caption>.toc-backref{color:inherit;text-decoration-line:none}figure:hover>figcaption>p>.headerlink,table:hover>caption>.headerlink{visibility:visible}:target>h1:first-of-type,:target>h2:first-of-type,:target>h3:first-of-type,:target>h4:first-of-type,:target>h5:first-of-type,:target>h6:first-of-type,span:target~h1:first-of-type,span:target~h2:first-of-type,span:target~h3:first-of-type,span:target~h4:first-of-type,span:target~h5:first-of-type,span:target~h6:first-of-type{background-color:var(--color-highlight-on-target)}:target>h1:first-of-type code.literal,:target>h2:first-of-type code.literal,:target>h3:first-of-type code.literal,:target>h4:first-of-type code.literal,:target>h5:first-of-type code.literal,:target>h6:first-of-type code.literal,span:target~h1:first-of-type code.literal,span:target~h2:first-of-type code.literal,span:target~h3:first-of-type code.literal,span:target~h4:first-of-type code.literal,span:target~h5:first-of-type code.literal,span:target~h6:first-of-type code.literal{background-color:transparent}.literal-block-wrapper:target .code-block-caption,.this-will-duplicate-information-and-it-is-still-useful-here li :target,figure:target,table:target>caption{background-color:var(--color-highlight-on-target)}dt:target{background-color:var(--color-highlight-on-target)!important}.footnote-reference:target,.footnote>dt:target+dd{background-color:var(--color-highlight-on-target)}.guilabel{background-color:var(--color-guilabel-background);border:1px solid var(--color-guilabel-border);border-radius:.5em;color:var(--color-guilabel-text);font-size:.9em;padding:0 .3em}footer{display:flex;flex-direction:column;font-size:var(--font-size--small);margin-top:2rem}.bottom-of-page{align-items:center;border-top:1px solid var(--color-background-border);color:var(--color-foreground-secondary);display:flex;justify-content:space-between;line-height:1.5;margin-top:1rem;padding-bottom:1rem;padding-top:1rem}@media(max-width:46em){.bottom-of-page{flex-direction:column-reverse;gap:.25rem;text-align:center}}.bottom-of-page .left-details{font-size:var(--font-size--small)}.bottom-of-page .right-details{display:flex;flex-direction:column;gap:.25rem;text-align:right}.bottom-of-page .icons{display:flex;font-size:1rem;gap:.25rem;justify-content:flex-end}.bottom-of-page .icons a{text-decoration:none}.bottom-of-page .icons img,.bottom-of-page .icons svg{font-size:1.125rem;height:1em;width:1em}.related-pages a{align-items:center;display:flex;text-decoration:none}.related-pages a:hover .page-info .title{color:var(--color-link);text-decoration:underline;text-decoration-color:var(--color-link-underline)}.related-pages a svg.furo-related-icon,.related-pages a svg.furo-related-icon>use{color:var(--color-foreground-border);flex-shrink:0;height:.75rem;margin:0 .5rem;width:.75rem}.related-pages a.next-page{clear:right;float:right;max-width:50%;text-align:right}.related-pages a.prev-page{clear:left;float:left;max-width:50%}.related-pages a.prev-page svg{transform:rotate(180deg)}.page-info{display:flex;flex-direction:column;overflow-wrap:anywhere}.next-page .page-info{align-items:flex-end}.page-info .context{align-items:center;color:var(--color-foreground-muted);display:flex;font-size:var(--font-size--small);padding-bottom:.1rem;text-decoration:none}ul.search{list-style:none;padding-left:0}ul.search li{border-bottom:1px solid var(--color-background-border);padding:1rem 0}[role=main] .highlighted{background-color:var(--color-highlighted-background);color:var(--color-highlighted-text)}.sidebar-brand{display:flex;flex-direction:column;flex-shrink:0;padding:var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal);text-decoration:none}.sidebar-brand-text{color:var(--color-sidebar-brand-text);font-size:1.5rem;overflow-wrap:break-word}.sidebar-brand-text,.sidebar-logo-container{margin:var(--sidebar-item-spacing-vertical) 0}.sidebar-logo{display:block;margin:0 auto;max-width:100%}.sidebar-search-container{align-items:center;background:var(--color-sidebar-search-background);display:flex;margin-top:var(--sidebar-search-space-above);position:relative}.sidebar-search-container:focus-within,.sidebar-search-container:hover{background:var(--color-sidebar-search-background--focus)}.sidebar-search-container:before{background-color:var(--color-sidebar-search-icon);content:"";height:var(--sidebar-search-icon-size);left:var(--sidebar-item-spacing-horizontal);-webkit-mask-image:var(--icon-search);mask-image:var(--icon-search);position:absolute;width:var(--sidebar-search-icon-size)}.sidebar-search{background:transparent;border:none;border-bottom:1px solid var(--color-sidebar-search-border);border-top:1px solid var(--color-sidebar-search-border);box-sizing:border-box;color:var(--color-sidebar-search-foreground);padding:var(--sidebar-search-input-spacing-vertical) var(--sidebar-search-input-spacing-horizontal) var(--sidebar-search-input-spacing-vertical) calc(var(--sidebar-item-spacing-horizontal) + var(--sidebar-search-input-spacing-horizontal) + var(--sidebar-search-icon-size));width:100%;z-index:10}.sidebar-search:focus{outline:none}.sidebar-search::-moz-placeholder{font-size:var(--sidebar-search-input-font-size)}.sidebar-search::placeholder{font-size:var(--sidebar-search-input-font-size)}#searchbox .highlight-link{margin:0;padding:var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal) 0;text-align:center}#searchbox .highlight-link a{color:var(--color-sidebar-search-icon);font-size:var(--font-size--small--2)}.sidebar-tree{font-size:var(--sidebar-item-font-size);margin-bottom:var(--sidebar-item-spacing-vertical);margin-top:var(--sidebar-tree-space-above)}.sidebar-tree ul{display:flex;flex-direction:column;list-style:none;margin-bottom:0;margin-top:0;padding:0}.sidebar-tree li{margin:0;position:relative}.sidebar-tree li>ul{margin-left:var(--sidebar-item-spacing-horizontal)}.sidebar-tree .icon,.sidebar-tree .reference{color:var(--color-sidebar-link-text)}.sidebar-tree .reference{box-sizing:border-box;display:inline-block;height:100%;line-height:var(--sidebar-item-line-height);overflow-wrap:anywhere;padding:var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal);text-decoration:none;width:100%}.sidebar-tree .reference:hover{background:var(--color-sidebar-item-background--hover)}.sidebar-tree .reference.external:after{color:var(--color-sidebar-link-text);content:url("data:image/svg+xml;charset=utf-8,%3Csvg width='12' height='12' xmlns='http://www.w3.org/2000/svg' viewBox='0 0 24 24' stroke-width='1.5' stroke='%23607D8B' fill='none' stroke-linecap='round' stroke-linejoin='round'%3E%3Cpath d='M0 0h24v24H0z' stroke='none'/%3E%3Cpath d='M11 7H6a2 2 0 0 0-2 2v9a2 2 0 0 0 2 2h9a2 2 0 0 0 2-2v-5M10 14 20 4M15 4h5v5'/%3E%3C/svg%3E");margin:0 .25rem;vertical-align:middle}.sidebar-tree .current-page>.reference{font-weight:700}.sidebar-tree label{align-items:center;cursor:pointer;display:flex;height:var(--sidebar-item-height);justify-content:center;position:absolute;right:0;top:0;-webkit-user-select:none;-moz-user-select:none;user-select:none;width:var(--sidebar-expander-width)}.sidebar-tree .caption,.sidebar-tree :not(.caption)>.caption-text{color:var(--color-sidebar-caption-text);font-size:var(--sidebar-caption-font-size);font-weight:700;margin:var(--sidebar-caption-space-above) 0 0 0;padding:var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal);text-transform:uppercase}.sidebar-tree li.has-children>.reference{padding-right:var(--sidebar-expander-width)}.sidebar-tree .toctree-l1>.reference,.sidebar-tree .toctree-l1>label .icon{color:var(--color-sidebar-link-text--top-level)}.sidebar-tree label{background:var(--color-sidebar-item-expander-background)}.sidebar-tree label:hover{background:var(--color-sidebar-item-expander-background--hover)}.sidebar-tree .current>.reference{background:var(--color-sidebar-item-background--current)}.sidebar-tree .current>.reference:hover{background:var(--color-sidebar-item-background--hover)}.toctree-checkbox{display:none;position:absolute}.toctree-checkbox~ul{display:none}.toctree-checkbox~label .icon svg{transform:rotate(90deg)}.toctree-checkbox:checked~ul{display:block}.toctree-checkbox:checked~label .icon svg{transform:rotate(-90deg)}.toc-title-container{padding:var(--toc-title-padding);padding-top:var(--toc-spacing-vertical)}.toc-title{color:var(--color-toc-title-text);font-size:var(--toc-title-font-size);padding-left:var(--toc-spacing-horizontal);text-transform:uppercase}.no-toc{display:none}.toc-tree-container{padding-bottom:var(--toc-spacing-vertical)}.toc-tree{border-left:1px solid var(--color-background-border);font-size:var(--toc-font-size);line-height:1.3;padding-left:calc(var(--toc-spacing-horizontal) - var(--toc-item-spacing-horizontal))}.toc-tree>ul>li:first-child{padding-top:0}.toc-tree>ul>li:first-child>ul{padding-left:0}.toc-tree>ul>li:first-child>a{display:none}.toc-tree ul{list-style-type:none;margin-bottom:0;margin-top:0;padding-left:var(--toc-item-spacing-horizontal)}.toc-tree li{padding-top:var(--toc-item-spacing-vertical)}.toc-tree li.scroll-current>.reference{color:var(--color-toc-item-text--active);font-weight:700}.toc-tree .reference{color:var(--color-toc-item-text);overflow-wrap:anywhere;text-decoration:none}.toc-scroll{max-height:100vh;overflow-y:scroll}.contents:not(.this-will-duplicate-information-and-it-is-still-useful-here){background:rgba(255,0,0,.25);color:var(--color-problematic)}.contents:not(.this-will-duplicate-information-and-it-is-still-useful-here):before{content:"ERROR: Adding a table of contents in Furo-based documentation is unnecessary, and does not work well with existing styling.Add a 'this-will-duplicate-information-and-it-is-still-useful-here' class, if you want an escape hatch."}.text-align\:left>p{text-align:left}.text-align\:center>p{text-align:center}.text-align\:right>p{text-align:right}
-/*# sourceMappingURL=furo.css.map*/
\ No newline at end of file
diff --git a/docs/_build/html/_static/styles/furo.css.map b/docs/_build/html/_static/styles/furo.css.map
deleted file mode 100644
index d1dfb109d..000000000
--- a/docs/_build/html/_static/styles/furo.css.map
+++ /dev/null
@@ -1 +0,0 @@
-{"version":3,"file":"styles/furo.css","mappings":"AAAA,2EAA2E,CAU3E,KAEE,6BAA8B,CAD9B,gBAEF,CASA,KACE,QACF,CAMA,KACE,aACF,CAOA,GACE,aAAc,CACd,cACF,CAUA,GACE,sBAAuB,CACvB,QAAS,CACT,gBACF,CAOA,IACE,+BAAiC,CACjC,aACF,CASA,EACE,4BACF,CAOA,YACE,kBAAmB,CACnB,yBAA0B,CAC1B,gCACF,CAMA,SAEE,kBACF,CAOA,cAGE,+BAAiC,CACjC,aACF,CAeA,QAEE,aAAc,CACd,aAAc,CACd,iBAAkB,CAClB,uBACF,CAEA,IACE,aACF,CAEA,IACE,SACF,CASA,IACE,iBACF,CAUA,sCAKE,mBAAoB,CACpB,cAAe,CACf,gBAAiB,CACjB,QACF,CAOA,aAEE,gBACF,CAOA,cAEE,mBACF,CAMA,gDAIE,yBACF,CAMA,wHAIE,iBAAkB,CAClB,SACF,CAMA,4GAIE,6BACF,CAMA,SACE,0BACF,CASA,OACE,qBAAsB,CACtB,aAAc,CACd,aAAc,CACd,cAAe,CACf,SAAU,CACV,kBACF,CAMA,SACE,uBACF,CAMA,SACE,aACF,CAOA,6BAEE,qBAAsB,CACtB,SACF,CAMA,kFAEE,WACF,CAOA,cACE,4BAA6B,CAC7B,mBACF,CAMA,yCACE,uBACF,CAOA,6BACE,yBAA0B,CAC1B,YACF,CASA,QACE,aACF,CAMA,QACE,iBACF,CAiBA,kBACE,YACF,CCvVA,aAcE,kEACE,uBAOF,WACE,iDAMF,gCACE,wBAEF,qCAEE,uBADA,uBACA,CAEF,SACE,wBAtBA,CCpBJ,iBAOE,6BAEA,mBANA,qBAEA,sBACA,0BAFA,oBAHA,4BAOA,6BANA,mBAOA,CAEF,gBACE,aCPF,KCGE,mHAEA,wGAGA,wBAAyB,CACzB,wBAAyB,CACzB,4BAA6B,CAC7B,yBAA0B,CAC1B,2BAA4B,CAG5B,sDAAuD,CACvD,gDAAiD,CACjD,wDAAyD,CAGzD,0CAA2C,CAC3C,gDAAiD,CACjD,gDAAiD,CAKjD,gCAAiC,CACjC,sCAAuC,CAGvC,2CAA4C,CAG5C,uCAAwC,CChCxC,+FAGA,uBAAwB,CAGxB,iCAAkC,CAClC,kCAAmC,CAEnC,+BAAgC,CAChC,sCAAuC,CACvC,sCAAuC,CACvC,qGAIA,mDAAoD,CAEpD,mCAAoC,CACpC,8CAA+C,CAC/C,gDAAiD,CACjD,kCAAmC,CACnC,6DAA8D,CAG9D,6BAA8B,CAC9B,6BAA8B,CAC9B,+BAAgC,CAChC,kCAAmC,CACnC,kCAAmC,CCPjC,ukBCYA,srCAZF,kaCVA,mLAOA,oTAWA,2UAaA,0CACA,gEACA,0CAGA,gEAUA,yCACA,+DAGA,4CACA,CACA,iEAGA,sGACA,uCACA,4DAGA,sCACA,2DAEA,4CACA,kEACA,oGACA,CAEA,0GACA,+CAGA,+MAOA,+EACA,wCAIA,4DACA,sEACA,kEACA,sEACA,gDAGA,+DACA,0CACA,gEACA,gGACA,CAGA,2DACA,qDAGA,0CACA,8CACA,oDACA,oDL7GF,iCAEA,iEAME,oCKyGA,yDAIA,sCACA,kCACA,sDAGA,0CACA,kEACA,oDAEA,sDAGA,oCACA,oEAIA,CAGA,yDAGA,qDACA,oDAGA,6DAIA,iEAGA,2DAEA,2DL9IE,4DAEA,gEAIF,gEKgGA,gFAIA,oNAOA,qDAEA,gFAIA,4DAIA,oEAMA,yEAIA,6DACA,0DAGA,uDAGA,qDAEA,wDLpII,6DAEA,yDACE,2DAMN,uCAIA,yCACE,8CAGF,sDMjDA,6DAKA,oCAIA,4CACA,kBAGF,sBAMA,2BAME,qCAGA,qCAEA,iCAEA,+BAEA,mCAEA,qCAIA,CACA,gCACA,gDAKA,kCAIA,6BAEA,0CAQA,kCAIF,8BAGE,8BACA,uCAGF,sCAKE,kCAEA,sDAGA,iCACE,CACA,2FAGA,gCACE,CACA,+DCzEJ,wCAEA,sBAEF,yDAEE,mCACA,wDAGA,2GAGA,wIACE,gDAMJ,kCAGE,6BACA,0CAGA,gEACA,8BACA,uCAKA,sCAIA,kCACA,sDACA,iCACA,sCAOA,sDAKE,gGAIE,+CAGN,sBAEE,yCAMA,0BAMA,yLAMA,aACA,MAEF,6BACE,2DAIF,wCAIE,kCAGA,SACA,kCAKA,mBAGA,CAJA,eACA,CAHF,gBAEE,CAWA,mBACA,mBACA,mDAGA,YACA,CACA,kBACA,CAEE,kBAKJ,OAPE,kBAQA,CADF,GACE,iCACA,wCAEA,wBACA,aACA,CAFA,WAEA,GACA,oBACA,CAFA,gBAEA,aACE,+CAIF,UAJE,kCAIF,WACA,iBACA,GAGA,uBACE,CAJF,yBAGA,CACE,iDACA,uCAEA,yDACE,cACA,wDAKN,yDAIE,uBAEF,kBACE,uBAEA,kDAIA,0DAGA,CAHA,oBAGA,0GAYA,aAEA,CAHA,YAGA,4HAKF,+CAGE,sBAEF,WAKE,0CAEA,CALA,qCAGA,CAJA,WAOA,SAIA,2CAJA,qCAIA,CACE,wBACA,OACA,YAEJ,gBACE,gBAIA,+CAKF,CAGE,kDAGA,CANF,8BAGE,CAGA,YAEA,CAdF,2BACE,CAHA,UAEF,CAYE,UAEA,CACA,0CACF,iEAOE,iCACA,8BAGA,wCAIA,wBAKE,0CAKF,CARE,6DAGA,CALF,qBAEE,CASA,YACA,yBAGA,CAEE,cAKN,CAPI,sBAOJ,gCAGE,qBAEA,WACA,aACA,sCAEA,mBACA,6BAGA,uEADA,qBACA,6BAIA,yBACA,qCAEE,UAEA,YACA,sBAEF,8BAGA,CAPE,aACA,WAMF,4BACE,sBACA,WAMJ,uBACE,cAYE,mBAXA,qDAKA,qCAGA,CAEA,YACA,CAHA,2BAEA,CACA,oCAEA,4CACA,uBAIA,oCAEJ,CAFI,cAIF,iBACE,CAHJ,kBAGI,yBAEA,oCAIA,qDAMF,mEAEA,CACE,8CAKA,gCAEA,qCAGA,oCAGE,sBACA,CAJF,WAEE,CAFF,eAEE,SAEA,mBACA,qCACE,aACA,CAFF,YADA,qBACA,WAEE,sBACA,kEAEN,2BAEE,iDAKA,uCAGF,CACE,0DAKA,kBACF,CAFE,sBAGA,mBACA,0BAEJ,yBAII,aADA,WACA,CAMF,UAFE,kBAEF,CAJF,gBACE,CAHE,iBAMF,6CC9ZF,yBACE,WACA,iBAEA,aAFA,iBAEA,6BAEA,kCACA,mBAKA,gCAGA,CARA,QAEA,CAGA,UALA,qBAEA,qDAGA,CALA,OAQA,4BACE,cAGF,2BACE,gCAEJ,CAHE,UAGF,8CAGE,CAHF,UAGE,wCAGA,qBACA,CAFA,UAEA,6CAGA,yCAIA,sBAHA,UAGA,kCACE,OACA,CAFF,KAEE,cAQF,0CACE,CAFF,kBACA,CACE,wEACA,CARA,YACA,CAKF,mBAFF,OAII,eACA,CAJF,iCAJE,cAGJ,CANI,oBAEA,CAKF,SAIE,2BADA,UACA,kBAGF,sCACA,CAFF,WACE,WACA,qCACE,gCACA,2EACA,sDAKJ,aACE,mDAII,CAJJ,6CAII,kEACA,iBACE,iDACA,+CACE,aACA,WADA,+BACA,uEANN,YACE,mDAEE,mBADF,0CACE,CADF,qBACE,0DACA,YACE,4DACA,sEANN,YACE,8CACA,kBADA,UACA,2CACE,2EACA,cACE,kEACA,mEANN,yBACE,4DACA,sBACE,+EAEE,iEACA,qEANN,sCACE,CAGE,iBAHF,gBAGE,qBACE,CAJJ,uBACA,gDACE,wDACA,6DAHF,2CACA,CADA,gBACA,eACE,CAGE,sBANN,8BACE,CAII,iBAFF,4DACA,WACE,YADF,uCACE,6EACA,2BANN,8CACE,kDACA,0CACE,8BACA,yFACE,sBACA,sFALJ,mEACA,sBACE,kEACA,6EACE,uCACA,kEALJ,qGAEE,kEACA,6EACE,uCACA,kEALJ,8CACA,uDACE,sEACA,2EACE,sCACA,iEALJ,mGACA,qCACE,oDACA,0DACE,6GACA,gDAGR,yDCrEA,sEACE,CACA,6GACE,gEACF,iGAIF,wFACE,qDAGA,mGAEE,2CAEF,4FACE,gCACF,wGACE,8DAEE,6FAIA,iJAKN,6GACE,gDAKF,yDACA,qCAGA,6BACA,kBACA,qDAKA,oCAEA,+DAGA,2CAGE,oDAIA,oEAEE,qBAGJ,wDAEE,uCAEF,kEAGA,8CAEA,uDAKA,oCAEA,yDAEE,gEAKF,+CC5FA,0EAGE,CACA,qDCLJ,+DAIE,sCAIA,kEACE,yBACA,2FAMA,gBACA,yGCbF,mBAOA,2MAIA,4HAYA,0DACE,8GAYF,8HAQE,mBAEA,6HAOF,YAGA,mIAME,eACA,CAFF,YAEE,4FAMJ,8BAEE,uBAYA,sCAEE,CAJF,oBAEA,CARA,wCAEA,CAHA,8BACA,CAFA,eACA,CAGA,wCAEA,CAEA,mDAIE,kCACE,6BACA,4CAKJ,kDAIA,eACE,aAGF,8BACE,uDACA,sCACA,cAEA,+BACA,CAFA,eAEA,wCAEF,YACE,iBACA,mCACA,0DAGF,qBAEE,CAFF,kBAEE,+BAIA,yCAEE,qBADA,gBACA,yBAKF,eACA,CAFF,YACE,CACA,iBACA,qDAEA,mDCvIJ,2FAOE,iCACA,CAEA,eACA,CAHA,kBAEA,CAFA,wBAGA,8BACA,eACE,CAFF,YAEE,0BACA,8CAGA,oBACE,oCAGA,kBACE,8DAEA,iBAEN,UACE,8BAIJ,+CAEE,qDAEF,kDAIE,YAEF,CAFE,YAEF,CCjCE,mFAJA,QACA,UAIE,CADF,iBACE,mCAGA,iDACE,+BAGF,wBAEA,mBAKA,6CAEF,CAHE,mBACA,CAEF,kCAIE,CARA,kBACA,CAFF,eASE,YACA,mBAGF,CAJE,UAIF,wCCjCA,oBDmCE,wBCpCJ,uCACE,8BACA,4CACA,oBAGA,2CCAA,6CAGE,CAPF,uBAIA,CDGA,gDACE,6BCVJ,CAWM,2CAEF,CAJA,kCAEE,CDJF,aCLF,gBDKE,uBCMA,gCAGA,gDAGE,wBAGJ,0BAEA,iBACE,aACF,CADE,UACF,uBACE,aACF,oBACE,YACF,4BACE,6CAMA,CAYF,6DAZE,mCAGE,iCASJ,4BAGE,4DADA,+BACA,CAFA,qBAEA,yBACE,aAEF,wBAHA,SAGA,iHACE,2DAKF,CANA,yCACE,CADF,oCAMA,uSAIA,sGACE,oDChEJ,WAEF,yBACE,QACA,eAEA,gBAEE,uCAGA,CALF,iCAKE,uCAGA,0BACA,CACA,oBACA,iCClBJ,gBACE,KAGF,qBACE,YAGF,CAHE,cAGF,gCAEE,mBACA,iEAEA,oCACA,wCAEA,sBACA,WAEA,CAFA,YAEA,8EAEA,mCAFA,iBAEA,6BAIA,wEAKA,sDAIE,CARF,mDAIA,CAIE,cAEF,8CAIA,oBAFE,iBAEF,8CAGE,eAEF,CAFE,YAEF,OAEE,kBAGJ,CAJI,eACA,CAFF,mBAKF,yCCjDE,oBACA,CAFA,iBAEA,uCAKE,iBACA,qCAGA,mBCZJ,CDWI,gBCXJ,6BAEE,eACA,sBAGA,eAEA,sBACA,oDACA,iGAMA,gBAFE,YAEF,8FAME,iJClBF,YACA,gNAUE,6BAEF,oTAcI,kBACF,gHAIA,qBACE,eACF,qDACE,kBACF,6DACE,4BCxCJ,oBAEF,qCAEI,+CAGF,uBACE,uDAGJ,oBAkBE,mDAhBA,+CAaA,CAbA,oBAaA,0FAEE,CAFF,gGAbA,+BAaA,0BAGA,mQAIA,oNAEE,iBAGJ,CAHI,gBADA,gBAIJ,8CAYI,CAZJ,wCAYI,sVACE,iCAGA,uEAHA,QAGA,qXAKJ,iDAGF,CARM,+CACE,iDAIN,CALI,gBAQN,mHACE,gBAGF,2DACE,0EAOA,0EAKA,6EC/EA,iDACA,gCACA,oDAGA,qBACA,oDCFA,cACA,eAEA,yBAGF,sBAEE,iBACA,sNAWA,iBACE,kBACA,wRAgBA,kBAEA,iOAgBA,uCACE,uEAEA,kBAEF,qUAuBE,iDAIJ,CACA,geCxFF,4BAEE,CAQA,6JACA,iDAIA,sEAGA,mDAOF,iDAGE,4DAIA,8CACA,qDAEE,eAFF,cAEE,oBAEF,uBAFE,kCAGA,eACA,iBACA,mBAIA,mDACA,CAHA,uCAEA,CAJA,0CACA,CAIA,gBAJA,gBACA,oBADA,gBAIA,wBAEJ,gBAGE,6BACA,YAHA,iBAGA,gCACA,iEAEA,6CACA,sDACA,0BADA,wBACA,0BACA,oIAIA,mBAFA,YAEA,qBACA,0CAIE,uBAEF,CAHA,yBACE,CAEF,iDACE,mFAKJ,oCACE,CANE,aAKJ,CACE,qEAIA,YAFA,WAEA,CAHA,aACA,CAEA,gBACE,4BACA,sBADA,aACA,gCAMF,oCACA,yDACA,2CAEA,qBAGE,kBAEA,CACA,mCAIF,CARE,YACA,CAOF,iCAEE,CAPA,oBACA,CAQA,oBACE,uDAEJ,sDAGA,CAHA,cAGA,0BACE,oDAIA,oCACA,4BACA,sBAGA,cAEA,oFAGA,sBAEA,yDACE,CAIA,iBAJA,wBAIA,6CAJA,6CAOA,4BAGJ,CAHI,cAGJ,yCAGA,kBACE,CAIA,iDAEA,CATA,YAEF,CACE,4CAGA,kBAIA,wEAEA,wDAIF,kCAOE,iDACA,CARF,WAIE,sCAGA,CANA,2CACA,CAMA,oEARF,iBACE,CACA,qCAMA,iBAuBE,uBAlBF,YAKA,2DALA,uDAKA,CALA,sBAiBA,4CACE,CALA,gRAIF,YACE,UAEN,uBACE,YACA,mCAOE,+CAGA,8BAGF,+CAGA,4BCjNA,SDiNA,qFCjNA,gDAGA,sCACA,qCACA,sDAIF,CAIE,kDAGA,CAPF,0CAOE,kBAEA,kDAEA,CAHA,eACA,CAFA,YACA,CADA,SAIA,mHAIE,CAGA,6CAFA,oCAeE,CAbF,yBACE,qBAEJ,CAGE,oBACA,CAEA,YAFA,2CACF,CACE,uBAEA,mFAEE,CALJ,oBACE,CAEA,UAEE,gCAGF,sDAEA,yCC7CJ,oCAGA,CD6CE,yXAQE,sCCrDJ,wCAGA,oCACE","sources":["webpack:///./node_modules/normalize.css/normalize.css","webpack:///./src/furo/assets/styles/base/_print.sass","webpack:///./src/furo/assets/styles/base/_screen-readers.sass","webpack:///./src/furo/assets/styles/base/_theme.sass","webpack:///./src/furo/assets/styles/variables/_fonts.scss","webpack:///./src/furo/assets/styles/variables/_spacing.scss","webpack:///./src/furo/assets/styles/variables/_icons.scss","webpack:///./src/furo/assets/styles/variables/_admonitions.scss","webpack:///./src/furo/assets/styles/variables/_colors.scss","webpack:///./src/furo/assets/styles/base/_typography.sass","webpack:///./src/furo/assets/styles/_scaffold.sass","webpack:///./src/furo/assets/styles/content/_admonitions.sass","webpack:///./src/furo/assets/styles/content/_api.sass","webpack:///./src/furo/assets/styles/content/_blocks.sass","webpack:///./src/furo/assets/styles/content/_captions.sass","webpack:///./src/furo/assets/styles/content/_code.sass","webpack:///./src/furo/assets/styles/content/_footnotes.sass","webpack:///./src/furo/assets/styles/content/_images.sass","webpack:///./src/furo/assets/styles/content/_indexes.sass","webpack:///./src/furo/assets/styles/content/_lists.sass","webpack:///./src/furo/assets/styles/content/_math.sass","webpack:///./src/furo/assets/styles/content/_misc.sass","webpack:///./src/furo/assets/styles/content/_rubrics.sass","webpack:///./src/furo/assets/styles/content/_sidebar.sass","webpack:///./src/furo/assets/styles/content/_tables.sass","webpack:///./src/furo/assets/styles/content/_target.sass","webpack:///./src/furo/assets/styles/content/_gui-labels.sass","webpack:///./src/furo/assets/styles/components/_footer.sass","webpack:///./src/furo/assets/styles/components/_sidebar.sass","webpack:///./src/furo/assets/styles/components/_table_of_contents.sass","webpack:///./src/furo/assets/styles/_shame.sass"],"sourcesContent":["/*! normalize.css v8.0.1 | MIT License | github.com/necolas/normalize.css */\n\n/* Document\n   ========================================================================== */\n\n/**\n * 1. Correct the line height in all browsers.\n * 2. Prevent adjustments of font size after orientation changes in iOS.\n */\n\nhtml {\n  line-height: 1.15; /* 1 */\n  -webkit-text-size-adjust: 100%; /* 2 */\n}\n\n/* Sections\n   ========================================================================== */\n\n/**\n * Remove the margin in all browsers.\n */\n\nbody {\n  margin: 0;\n}\n\n/**\n * Render the `main` element consistently in IE.\n */\n\nmain {\n  display: block;\n}\n\n/**\n * Correct the font size and margin on `h1` elements within `section` and\n * `article` contexts in Chrome, Firefox, and Safari.\n */\n\nh1 {\n  font-size: 2em;\n  margin: 0.67em 0;\n}\n\n/* Grouping content\n   ========================================================================== */\n\n/**\n * 1. Add the correct box sizing in Firefox.\n * 2. Show the overflow in Edge and IE.\n */\n\nhr {\n  box-sizing: content-box; /* 1 */\n  height: 0; /* 1 */\n  overflow: visible; /* 2 */\n}\n\n/**\n * 1. Correct the inheritance and scaling of font size in all browsers.\n * 2. Correct the odd `em` font sizing in all browsers.\n */\n\npre {\n  font-family: monospace, monospace; /* 1 */\n  font-size: 1em; /* 2 */\n}\n\n/* Text-level semantics\n   ========================================================================== */\n\n/**\n * Remove the gray background on active links in IE 10.\n */\n\na {\n  background-color: transparent;\n}\n\n/**\n * 1. Remove the bottom border in Chrome 57-\n * 2. Add the correct text decoration in Chrome, Edge, IE, Opera, and Safari.\n */\n\nabbr[title] {\n  border-bottom: none; /* 1 */\n  text-decoration: underline; /* 2 */\n  text-decoration: underline dotted; /* 2 */\n}\n\n/**\n * Add the correct font weight in Chrome, Edge, and Safari.\n */\n\nb,\nstrong {\n  font-weight: bolder;\n}\n\n/**\n * 1. Correct the inheritance and scaling of font size in all browsers.\n * 2. Correct the odd `em` font sizing in all browsers.\n */\n\ncode,\nkbd,\nsamp {\n  font-family: monospace, monospace; /* 1 */\n  font-size: 1em; /* 2 */\n}\n\n/**\n * Add the correct font size in all browsers.\n */\n\nsmall {\n  font-size: 80%;\n}\n\n/**\n * Prevent `sub` and `sup` elements from affecting the line height in\n * all browsers.\n */\n\nsub,\nsup {\n  font-size: 75%;\n  line-height: 0;\n  position: relative;\n  vertical-align: baseline;\n}\n\nsub {\n  bottom: -0.25em;\n}\n\nsup {\n  top: -0.5em;\n}\n\n/* Embedded content\n   ========================================================================== */\n\n/**\n * Remove the border on images inside links in IE 10.\n */\n\nimg {\n  border-style: none;\n}\n\n/* Forms\n   ========================================================================== */\n\n/**\n * 1. Change the font styles in all browsers.\n * 2. Remove the margin in Firefox and Safari.\n */\n\nbutton,\ninput,\noptgroup,\nselect,\ntextarea {\n  font-family: inherit; /* 1 */\n  font-size: 100%; /* 1 */\n  line-height: 1.15; /* 1 */\n  margin: 0; /* 2 */\n}\n\n/**\n * Show the overflow in IE.\n * 1. Show the overflow in Edge.\n */\n\nbutton,\ninput { /* 1 */\n  overflow: visible;\n}\n\n/**\n * Remove the inheritance of text transform in Edge, Firefox, and IE.\n * 1. Remove the inheritance of text transform in Firefox.\n */\n\nbutton,\nselect { /* 1 */\n  text-transform: none;\n}\n\n/**\n * Correct the inability to style clickable types in iOS and Safari.\n */\n\nbutton,\n[type=\"button\"],\n[type=\"reset\"],\n[type=\"submit\"] {\n  -webkit-appearance: button;\n}\n\n/**\n * Remove the inner border and padding in Firefox.\n */\n\nbutton::-moz-focus-inner,\n[type=\"button\"]::-moz-focus-inner,\n[type=\"reset\"]::-moz-focus-inner,\n[type=\"submit\"]::-moz-focus-inner {\n  border-style: none;\n  padding: 0;\n}\n\n/**\n * Restore the focus styles unset by the previous rule.\n */\n\nbutton:-moz-focusring,\n[type=\"button\"]:-moz-focusring,\n[type=\"reset\"]:-moz-focusring,\n[type=\"submit\"]:-moz-focusring {\n  outline: 1px dotted ButtonText;\n}\n\n/**\n * Correct the padding in Firefox.\n */\n\nfieldset {\n  padding: 0.35em 0.75em 0.625em;\n}\n\n/**\n * 1. Correct the text wrapping in Edge and IE.\n * 2. Correct the color inheritance from `fieldset` elements in IE.\n * 3. Remove the padding so developers are not caught out when they zero out\n *    `fieldset` elements in all browsers.\n */\n\nlegend {\n  box-sizing: border-box; /* 1 */\n  color: inherit; /* 2 */\n  display: table; /* 1 */\n  max-width: 100%; /* 1 */\n  padding: 0; /* 3 */\n  white-space: normal; /* 1 */\n}\n\n/**\n * Add the correct vertical alignment in Chrome, Firefox, and Opera.\n */\n\nprogress {\n  vertical-align: baseline;\n}\n\n/**\n * Remove the default vertical scrollbar in IE 10+.\n */\n\ntextarea {\n  overflow: auto;\n}\n\n/**\n * 1. Add the correct box sizing in IE 10.\n * 2. Remove the padding in IE 10.\n */\n\n[type=\"checkbox\"],\n[type=\"radio\"] {\n  box-sizing: border-box; /* 1 */\n  padding: 0; /* 2 */\n}\n\n/**\n * Correct the cursor style of increment and decrement buttons in Chrome.\n */\n\n[type=\"number\"]::-webkit-inner-spin-button,\n[type=\"number\"]::-webkit-outer-spin-button {\n  height: auto;\n}\n\n/**\n * 1. Correct the odd appearance in Chrome and Safari.\n * 2. Correct the outline style in Safari.\n */\n\n[type=\"search\"] {\n  -webkit-appearance: textfield; /* 1 */\n  outline-offset: -2px; /* 2 */\n}\n\n/**\n * Remove the inner padding in Chrome and Safari on macOS.\n */\n\n[type=\"search\"]::-webkit-search-decoration {\n  -webkit-appearance: none;\n}\n\n/**\n * 1. Correct the inability to style clickable types in iOS and Safari.\n * 2. Change font properties to `inherit` in Safari.\n */\n\n::-webkit-file-upload-button {\n  -webkit-appearance: button; /* 1 */\n  font: inherit; /* 2 */\n}\n\n/* Interactive\n   ========================================================================== */\n\n/*\n * Add the correct display in Edge, IE 10+, and Firefox.\n */\n\ndetails {\n  display: block;\n}\n\n/*\n * Add the correct display in all browsers.\n */\n\nsummary {\n  display: list-item;\n}\n\n/* Misc\n   ========================================================================== */\n\n/**\n * Add the correct display in IE 10+.\n */\n\ntemplate {\n  display: none;\n}\n\n/**\n * Add the correct display in IE 10.\n */\n\n[hidden] {\n  display: none;\n}\n","// This file contains styles for managing print media.\n\n////////////////////////////////////////////////////////////////////////////////\n// Hide elements not relevant to print media.\n////////////////////////////////////////////////////////////////////////////////\n@media print\n  // Hide icon container.\n  .content-icon-container\n    display: none !important\n\n  // Hide showing header links if hovering over when printing.\n  .headerlink\n    display: none !important\n\n  // Hide mobile header.\n  .mobile-header\n    display: none !important\n\n  // Hide navigation links.\n  .related-pages\n    display: none !important\n\n////////////////////////////////////////////////////////////////////////////////\n// Tweaks related to decolorization.\n////////////////////////////////////////////////////////////////////////////////\n@media print\n  // Apply a border around code which no longer have a color background.\n  .highlight\n    border: 0.1pt solid var(--color-foreground-border)\n\n////////////////////////////////////////////////////////////////////////////////\n// Avoid page break in some relevant cases.\n////////////////////////////////////////////////////////////////////////////////\n@media print\n  ul, ol, dl, a, table, pre, blockquote\n    page-break-inside: avoid\n\n  h1, h2, h3, h4, h5, h6, img, figure, caption\n    page-break-inside: avoid\n    page-break-after: avoid\n\n  ul, ol, dl\n    page-break-before: avoid\n",".visually-hidden\n  position: absolute !important\n  width: 1px !important\n  height: 1px !important\n  padding: 0 !important\n  margin: -1px !important\n  overflow: hidden !important\n  clip: rect(0,0,0,0) !important\n  white-space: nowrap !important\n  border: 0 !important\n\n:-moz-focusring\n  outline: auto\n","// This file serves as the \"skeleton\" of the theming logic.\n//\n// This contains the bulk of the logic for handling dark mode, color scheme\n// toggling and the handling of color-scheme-specific hiding of elements.\n\nbody\n  @include fonts\n  @include spacing\n  @include icons\n  @include admonitions\n  @include default-admonition(#651fff, \"abstract\")\n  @include default-topic(#14B8A6, \"pencil\")\n\n  @include colors\n\n.only-light\n  display: block !important\nhtml body .only-dark\n  display: none !important\n\n// Ignore dark-mode hints if print media.\n@media not print\n  // Enable dark-mode, if requested.\n  body[data-theme=\"dark\"]\n    @include colors-dark\n\n    html & .only-light\n      display: none !important\n    .only-dark\n      display: block !important\n\n  // Enable dark mode, unless explicitly told to avoid.\n  @media (prefers-color-scheme: dark)\n    body:not([data-theme=\"light\"])\n      @include colors-dark\n\n      html & .only-light\n        display: none !important\n      .only-dark\n        display: block !important\n\n//\n// Theme toggle presentation\n//\nbody[data-theme=\"auto\"]\n  .theme-toggle svg.theme-icon-when-auto\n    display: block\n\nbody[data-theme=\"dark\"]\n  .theme-toggle svg.theme-icon-when-dark\n    display: block\n\nbody[data-theme=\"light\"]\n  .theme-toggle svg.theme-icon-when-light\n    display: block\n","// Fonts used by this theme.\n//\n// There are basically two things here -- using the system font stack and\n// defining sizes for various elements in %ages. We could have also used `em`\n// but %age is easier to reason about for me.\n\n@mixin fonts {\n  // These are adapted from https://systemfontstack.com/\n  --font-stack: -apple-system, BlinkMacSystemFont, Segoe UI, Helvetica, Arial,\n    sans-serif, Apple Color Emoji, Segoe UI Emoji;\n  --font-stack--monospace: \"SFMono-Regular\", Menlo, Consolas, Monaco,\n    Liberation Mono, Lucida Console, monospace;\n\n  --font-size--normal: 100%;\n  --font-size--small: 87.5%;\n  --font-size--small--2: 81.25%;\n  --font-size--small--3: 75%;\n  --font-size--small--4: 62.5%;\n\n  // Sidebar\n  --sidebar-caption-font-size: var(--font-size--small--2);\n  --sidebar-item-font-size: var(--font-size--small);\n  --sidebar-search-input-font-size: var(--font-size--small);\n\n  // Table of Contents\n  --toc-font-size: var(--font-size--small--3);\n  --toc-font-size--mobile: var(--font-size--normal);\n  --toc-title-font-size: var(--font-size--small--4);\n\n  // Admonitions\n  //\n  // These aren't defined in terms of %ages, since nesting these is permitted.\n  --admonition-font-size: 0.8125rem;\n  --admonition-title-font-size: 0.8125rem;\n\n  // Code\n  --code-font-size: var(--font-size--small--2);\n\n  // API\n  --api-font-size: var(--font-size--small);\n}\n","// Spacing for various elements on the page\n//\n// If the user wants to tweak things in a certain way, they are permitted to.\n// They also have to deal with the consequences though!\n\n@mixin spacing {\n  // Header!\n  --header-height: calc(\n    var(--sidebar-item-line-height) + 4 * #{var(--sidebar-item-spacing-vertical)}\n  );\n  --header-padding: 0.5rem;\n\n  // Sidebar\n  --sidebar-tree-space-above: 1.5rem;\n  --sidebar-caption-space-above: 1rem;\n\n  --sidebar-item-line-height: 1rem;\n  --sidebar-item-spacing-vertical: 0.5rem;\n  --sidebar-item-spacing-horizontal: 1rem;\n  --sidebar-item-height: calc(\n    var(--sidebar-item-line-height) + 2 *#{var(--sidebar-item-spacing-vertical)}\n  );\n\n  --sidebar-expander-width: var(--sidebar-item-height); // be square\n\n  --sidebar-search-space-above: 0.5rem;\n  --sidebar-search-input-spacing-vertical: 0.5rem;\n  --sidebar-search-input-spacing-horizontal: 0.5rem;\n  --sidebar-search-input-height: 1rem;\n  --sidebar-search-icon-size: var(--sidebar-search-input-height);\n\n  // Table of Contents\n  --toc-title-padding: 0.25rem 0;\n  --toc-spacing-vertical: 1.5rem;\n  --toc-spacing-horizontal: 1.5rem;\n  --toc-item-spacing-vertical: 0.4rem;\n  --toc-item-spacing-horizontal: 1rem;\n}\n","// Expose theme icons as CSS variables.\n\n$icons: (\n  // Adapted from tabler-icons\n  //    url: https://tablericons.com/\n  \"search\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\" stroke-width=\"1.5\" stroke=\"currentColor\" fill=\"none\" stroke-linecap=\"round\" stroke-linejoin=\"round\"><path stroke=\"none\" d=\"M0 0h24v24H0z\"/><circle cx=\"10\" cy=\"10\" r=\"7\" /><line x1=\"21\" y1=\"21\" x2=\"15\" y2=\"15\" /></svg>'),\n  // Factored out from mkdocs-material on 24-Aug-2020.\n  //    url: https://squidfunk.github.io/mkdocs-material/reference/admonitions/\n  \"pencil\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M20.71 7.04c.39-.39.39-1.04 0-1.41l-2.34-2.34c-.37-.39-1.02-.39-1.41 0l-1.84 1.83 3.75 3.75M3 17.25V21h3.75L17.81 9.93l-3.75-3.75L3 17.25z\"/></svg>'),\n  \"abstract\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M4 5h16v2H4V5m0 4h16v2H4V9m0 4h16v2H4v-2m0 4h10v2H4v-2z\"/></svg>'),\n  \"info\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M13 9h-2V7h2m0 10h-2v-6h2m-1-9A10 10 0 002 12a10 10 0 0010 10 10 10 0 0010-10A10 10 0 0012 2z\"/></svg>'),\n  \"flame\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M17.55 11.2c-.23-.3-.5-.56-.76-.82-.65-.6-1.4-1.03-2.03-1.66C13.3 7.26 13 4.85 13.91 3c-.91.23-1.75.75-2.45 1.32-2.54 2.08-3.54 5.75-2.34 8.9.04.1.08.2.08.33 0 .22-.15.42-.35.5-.22.1-.46.04-.64-.12a.83.83 0 01-.15-.17c-1.1-1.43-1.28-3.48-.53-5.12C5.89 10 5 12.3 5.14 14.47c.04.5.1 1 .27 1.5.14.6.4 1.2.72 1.73 1.04 1.73 2.87 2.97 4.84 3.22 2.1.27 4.35-.12 5.96-1.6 1.8-1.66 2.45-4.32 1.5-6.6l-.13-.26c-.2-.46-.47-.87-.8-1.25l.05-.01m-3.1 6.3c-.28.24-.73.5-1.08.6-1.1.4-2.2-.16-2.87-.82 1.19-.28 1.89-1.16 2.09-2.05.17-.8-.14-1.46-.27-2.23-.12-.74-.1-1.37.18-2.06.17.38.37.76.6 1.06.76 1 1.95 1.44 2.2 2.8.04.14.06.28.06.43.03.82-.32 1.72-.92 2.27h.01z\"/></svg>'),\n  \"question\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M15.07 11.25l-.9.92C13.45 12.89 13 13.5 13 15h-2v-.5c0-1.11.45-2.11 1.17-2.83l1.24-1.26c.37-.36.59-.86.59-1.41a2 2 0 00-2-2 2 2 0 00-2 2H8a4 4 0 014-4 4 4 0 014 4 3.2 3.2 0 01-.93 2.25M13 19h-2v-2h2M12 2A10 10 0 002 12a10 10 0 0010 10 10 10 0 0010-10c0-5.53-4.5-10-10-10z\"/></svg>'),\n  \"warning\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M13 14h-2v-4h2m0 8h-2v-2h2M1 21h22L12 2 1 21z\"/></svg>'),\n  \"failure\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M12 2c5.53 0 10 4.47 10 10s-4.47 10-10 10S2 17.53 2 12 6.47 2 12 2m3.59 5L12 10.59 8.41 7 7 8.41 10.59 12 7 15.59 8.41 17 12 13.41 15.59 17 17 15.59 13.41 12 17 8.41 15.59 7z\"/></svg>'),\n  \"spark\":\n    url('data:image/svg+xml;charset=utf-8,<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\"><path d=\"M11.5 20l4.86-9.73H13V4l-5 9.73h3.5V20M12 2c2.75 0 5.1 1 7.05 2.95C21 6.9 22 9.25 22 12s-1 5.1-2.95 7.05C17.1 21 14.75 22 12 22s-5.1-1-7.05-2.95C3 17.1 2 14.75 2 12s1-5.1 2.95-7.05C6.9 3 9.25 2 12 2z\"/></svg>')\n);\n\n@mixin icons {\n  @each $name, $glyph in $icons {\n    --icon-#{$name}: #{$glyph};\n  }\n}\n","// Admonitions\n\n// Structure of these is:\n//    admonition-class: color \"icon-name\";\n//\n// The colors are translated into CSS variables below. The icons are\n// used directly in the main declarations to set the `mask-image` in\n// the title.\n\n// prettier-ignore\n$admonitions: (\n  // Each of these has an reST directives for it.\n  \"caution\":         #ff9100 \"spark\",\n  \"warning\":         #ff9100 \"warning\",\n  \"danger\":          #ff5252 \"spark\",\n  \"attention\":       #ff5252 \"warning\",\n  \"error\":           #ff5252 \"failure\",\n  \"hint\":            #00c852 \"question\",\n  \"tip\":             #00c852 \"info\",\n  \"important\":       #00bfa5 \"flame\",\n  \"note\":            #00b0ff \"pencil\",\n  \"seealso\":         #448aff \"info\",\n  \"admonition-todo\": #808080 \"pencil\"\n);\n\n@mixin default-admonition($color, $icon-name) {\n  --color-admonition-title: #{$color};\n  --color-admonition-title-background: #{rgba($color, 0.2)};\n\n  --icon-admonition-default: var(--icon-#{$icon-name});\n}\n\n@mixin default-topic($color, $icon-name) {\n  --color-topic-title: #{$color};\n  --color-topic-title-background: #{rgba($color, 0.2)};\n\n  --icon-topic-default: var(--icon-#{$icon-name});\n}\n\n@mixin admonitions {\n  @each $name, $values in $admonitions {\n    --color-admonition-title--#{$name}: #{nth($values, 1)};\n    --color-admonition-title-background--#{$name}: #{rgba(\n        nth($values, 1),\n        0.2\n      )};\n  }\n}\n","// Colors used throughout this theme.\n//\n// The aim is to give the user more control. Thus, instead of hard-coding colors\n// in various parts of the stylesheet, the approach taken is to define all\n// colors as CSS variables and reusing them in all the places.\n//\n// `colors-dark` depends on `colors` being included at a lower specificity.\n\n@mixin colors {\n  --color-problematic: #b30000;\n\n  // Base Colors\n  --color-foreground-primary: black; // for main text and headings\n  --color-foreground-secondary: #5a5c63; // for secondary text\n  --color-foreground-muted: #646776; // for muted text\n  --color-foreground-border: #878787; // for content borders\n\n  --color-background-primary: white; // for content\n  --color-background-secondary: #f8f9fb; // for navigation + ToC\n  --color-background-hover: #efeff4ff; // for navigation-item hover\n  --color-background-hover--transparent: #efeff400;\n  --color-background-border: #eeebee; // for UI borders\n  --color-background-item: #ccc; // for \"background\" items (eg: copybutton)\n\n  // Announcements\n  --color-announcement-background: #000000dd;\n  --color-announcement-text: #eeebee;\n\n  // Brand colors\n  --color-brand-primary: #2962ff;\n  --color-brand-content: #2a5adf;\n\n  // API documentation\n  --color-api-background: var(--color-background-hover--transparent);\n  --color-api-background-hover: var(--color-background-hover);\n  --color-api-overall: var(--color-foreground-secondary);\n  --color-api-name: var(--color-problematic);\n  --color-api-pre-name: var(--color-problematic);\n  --color-api-paren: var(--color-foreground-secondary);\n  --color-api-keyword: var(--color-foreground-primary);\n  --color-highlight-on-target: #ffffcc;\n\n  // Inline code background\n  --color-inline-code-background: var(--color-background-secondary);\n\n  // Highlighted text (search)\n  --color-highlighted-background: #ddeeff;\n  --color-highlighted-text: var(--color-foreground-primary);\n\n  // GUI Labels\n  --color-guilabel-background: #ddeeff80;\n  --color-guilabel-border: #bedaf580;\n  --color-guilabel-text: var(--color-foreground-primary);\n\n  // Admonitions!\n  --color-admonition-background: transparent;\n\n  //////////////////////////////////////////////////////////////////////////////\n  // Everything below this should be one of:\n  // - var(...)\n  // - *-gradient(...)\n  // - special literal values (eg: transparent, none)\n  //////////////////////////////////////////////////////////////////////////////\n\n  // Tables\n  --color-table-header-background: var(--color-background-secondary);\n  --color-table-border: var(--color-background-border);\n\n  // Cards\n  --color-card-border: var(--color-background-secondary);\n  --color-card-background: transparent;\n  --color-card-marginals-background: var(--color-background-secondary);\n\n  // Header\n  --color-header-background: var(--color-background-primary);\n  --color-header-border: var(--color-background-border);\n  --color-header-text: var(--color-foreground-primary);\n\n  // Sidebar (left)\n  --color-sidebar-background: var(--color-background-secondary);\n  --color-sidebar-background-border: var(--color-background-border);\n\n  --color-sidebar-brand-text: var(--color-foreground-primary);\n  --color-sidebar-caption-text: var(--color-foreground-muted);\n  --color-sidebar-link-text: var(--color-foreground-secondary);\n  --color-sidebar-link-text--top-level: var(--color-brand-primary);\n\n  --color-sidebar-item-background: var(--color-sidebar-background);\n  --color-sidebar-item-background--current: var(\n    --color-sidebar-item-background\n  );\n  --color-sidebar-item-background--hover: linear-gradient(\n    90deg,\n    var(--color-background-hover--transparent) 0%,\n    var(--color-background-hover) var(--sidebar-item-spacing-horizontal),\n    var(--color-background-hover) 100%\n  );\n\n  --color-sidebar-item-expander-background: transparent;\n  --color-sidebar-item-expander-background--hover: var(\n    --color-background-hover\n  );\n\n  --color-sidebar-search-text: var(--color-foreground-primary);\n  --color-sidebar-search-background: var(--color-background-secondary);\n  --color-sidebar-search-background--focus: var(--color-background-primary);\n  --color-sidebar-search-border: var(--color-background-border);\n  --color-sidebar-search-icon: var(--color-foreground-muted);\n\n  // Table of Contents (right)\n  --color-toc-background: var(--color-background-primary);\n  --color-toc-title-text: var(--color-foreground-muted);\n  --color-toc-item-text: var(--color-foreground-secondary);\n  --color-toc-item-text--hover: var(--color-foreground-primary);\n  --color-toc-item-text--active: var(--color-brand-primary);\n\n  // Actual page contents\n  --color-content-foreground: var(--color-foreground-primary);\n  --color-content-background: transparent;\n\n  // Links\n  --color-link: var(--color-brand-content);\n  --color-link--hover: var(--color-brand-content);\n  --color-link-underline: var(--color-background-border);\n  --color-link-underline--hover: var(--color-foreground-border);\n}\n\n@mixin colors-dark {\n  --color-problematic: #ee5151;\n\n  // Base Colors\n  --color-foreground-primary: #ffffffcc; // for main text and headings\n  --color-foreground-secondary: #9ca0a5; // for secondary text\n  --color-foreground-muted: #81868d; // for muted text\n  --color-foreground-border: #666666; // for content borders\n\n  --color-background-primary: #131416; // for content\n  --color-background-secondary: #1a1c1e; // for navigation + ToC\n  --color-background-hover: #1e2124ff; // for navigation-item hover\n  --color-background-hover--transparent: #1e212400;\n  --color-background-border: #303335; // for UI borders\n  --color-background-item: #444; // for \"background\" items (eg: copybutton)\n\n  // Announcements\n  --color-announcement-background: #000000dd;\n  --color-announcement-text: #eeebee;\n\n  // Brand colors\n  --color-brand-primary: #2b8cee;\n  --color-brand-content: #368ce2;\n\n  // Highlighted text (search)\n  --color-highlighted-background: #083563;\n\n  // GUI Labels\n  --color-guilabel-background: #08356380;\n  --color-guilabel-border: #13395f80;\n\n  // API documentation\n  --color-api-keyword: var(--color-foreground-secondary);\n  --color-highlight-on-target: #333300;\n\n  // Admonitions\n  --color-admonition-background: #18181a;\n\n  // Cards\n  --color-card-border: var(--color-background-secondary);\n  --color-card-background: #18181a;\n  --color-card-marginals-background: var(--color-background-hover);\n}\n","// This file contains the styling for making the content throughout the page,\n// including fonts, paragraphs, headings and spacing among these elements.\n\nbody\n  font-family: var(--font-stack)\npre,\ncode,\nkbd,\nsamp\n  font-family: var(--font-stack--monospace)\n\n// Make fonts look slightly nicer.\nbody\n  -webkit-font-smoothing: antialiased\n  -moz-osx-font-smoothing: grayscale\n\n// Line height from Bootstrap 4.1\narticle\n  line-height: 1.5\n\n//\n// Headings\n//\nh1,\nh2,\nh3,\nh4,\nh5,\nh6\n  line-height: 1.25\n  font-weight: bold\n\n  border-radius: 0.5rem\n  margin-top: 0.5rem\n  margin-bottom: 0.5rem\n  margin-left: -0.5rem\n  margin-right: -0.5rem\n  padding-left: 0.5rem\n  padding-right: 0.5rem\n\n  + p\n    margin-top: 0\n\nh1\n  font-size: 2.5em\n  margin-top: 1.75rem\n  margin-bottom: 1rem\nh2\n  font-size: 2em\n  margin-top: 1.75rem\nh3\n  font-size: 1.5em\nh4\n  font-size: 1.25em\nh5\n  font-size: 1.125em\nh6\n  font-size: 1em\n\nsmall\n  opacity: 75%\n  font-size: 80%\n\n// Paragraph\np\n  margin-top: 0.5rem\n  margin-bottom: 0.75rem\n\n// Horizontal rules\nhr.docutils\n  height: 1px\n  padding: 0\n  margin: 2rem 0\n  background-color: var(--color-background-border)\n  border: 0\n\n.centered\n  text-align: center\n\n// Links\na\n  text-decoration: underline\n\n  color: var(--color-link)\n  text-decoration-color: var(--color-link-underline)\n\n  &:hover\n    color: var(--color-link--hover)\n    text-decoration-color: var(--color-link-underline--hover)\n  &.muted-link\n    color: inherit\n    &:hover\n      color: var(--color-link)\n      text-decoration-color: var(--color-link-underline--hover)\n","// This file contains the styles for the overall layouting of the documentation\n// skeleton, including the responsive changes as well as sidebar toggles.\n//\n// This is implemented as a mobile-last design, which isn't ideal, but it is\n// reasonably good-enough and I got pretty tired by the time I'd finished this\n// to move the rules around to fix this. Shouldn't take more than 3-4 hours,\n// if you know what you're doing tho.\n\n// HACK: Not all browsers account for the scrollbar width in media queries.\n// This results in horizontal scrollbars in the breakpoint where we go\n// from displaying everything to hiding the ToC. We accomodate for this by\n// adding a bit of padding to the TOC drawer, disabling the horizontal\n// scrollbar and allowing the scrollbars to cover the padding.\n// https://www.456bereastreet.com/archive/201301/media_query_width_and_vertical_scrollbars/\n\n// HACK: Always having the scrollbar visible, prevents certain browsers from\n// causing the content to stutter horizontally between taller-than-viewport and\n// not-taller-than-viewport pages.\n\nhtml\n  overflow-x: hidden\n  overflow-y: scroll\n  scroll-behavior: smooth\n\n.sidebar-scroll, .toc-scroll, article[role=main] *\n  // Override Firefox scrollbar style\n  scrollbar-width: thin\n  scrollbar-color: var(--color-foreground-border) transparent\n\n  // Override Chrome scrollbar styles\n  &::-webkit-scrollbar\n    width: 0.25rem\n    height: 0.25rem\n  &::-webkit-scrollbar-thumb\n    background-color: var(--color-foreground-border)\n    border-radius: 0.125rem\n\n//\n// Overalls\n//\nhtml,\nbody\n  height: 100%\n  color: var(--color-foreground-primary)\n  background: var(--color-background-primary)\n\narticle\n  color: var(--color-content-foreground)\n  background: var(--color-content-background)\n  overflow-wrap: break-word\n\n.page\n  display: flex\n  // fill the viewport for pages with little content.\n  min-height: 100%\n\n.mobile-header\n  width: 100%\n  height: var(--header-height)\n  background-color: var(--color-header-background)\n  color: var(--color-header-text)\n  border-bottom: 1px solid var(--color-header-border)\n\n  // Looks like sub-script/super-script have this, and we need this to\n  // be \"on top\" of those.\n  z-index: 10\n\n  // We don't show the header on large screens.\n  display: none\n\n  // Add shadow when scrolled\n  &.scrolled\n    border-bottom: none\n    box-shadow: 0 0 0.2rem rgba(0, 0, 0, 0.1), 0 0.2rem 0.4rem rgba(0, 0, 0, 0.2)\n\n  .header-center\n    a\n      color: var(--color-header-text)\n      text-decoration: none\n\n.main\n  display: flex\n  flex: 1\n\n// Sidebar (left) also covers the entire left portion of screen.\n.sidebar-drawer\n  box-sizing: border-box\n\n  border-right: 1px solid var(--color-sidebar-background-border)\n  background: var(--color-sidebar-background)\n\n  display: flex\n  justify-content: flex-end\n  // These next two lines took me two days to figure out.\n  width: calc((100% - #{$full-width}) / 2 + #{$sidebar-width})\n  min-width: $sidebar-width\n\n// Scroll-along sidebars\n.sidebar-container,\n.toc-drawer\n  box-sizing: border-box\n  width: $sidebar-width\n\n.toc-drawer\n  background: var(--color-toc-background)\n  // See HACK described on top of this document\n  padding-right: 1rem\n\n.sidebar-sticky,\n.toc-sticky\n  position: sticky\n  top: 0\n  height: min(100%, 100vh)\n  height: 100vh\n\n  display: flex\n  flex-direction: column\n\n.sidebar-scroll,\n.toc-scroll\n  flex-grow: 1\n  flex-shrink: 1\n\n  overflow: auto\n  scroll-behavior: smooth\n\n// Central items.\n.content\n  padding: 0 $content-padding\n  width: $content-width\n\n  display: flex\n  flex-direction: column\n  justify-content: space-between\n\n.icon\n  display: inline-block\n  height: 1rem\n  width: 1rem\n  svg\n    width: 100%\n    height: 100%\n\n//\n// Accommodate announcement banner\n//\n.announcement\n  background-color: var(--color-announcement-background)\n  color: var(--color-announcement-text)\n\n  height: var(--header-height)\n  display: flex\n  align-items: center\n  overflow-x: auto\n  & + .page\n    min-height: calc(100% - var(--header-height))\n\n.announcement-content\n  box-sizing: border-box\n  padding: 0.5rem\n  min-width: 100%\n  white-space: nowrap\n  text-align: center\n\n  a\n    color: var(--color-announcement-text)\n    text-decoration-color: var(--color-announcement-text)\n\n    &:hover\n      color: var(--color-announcement-text)\n      text-decoration-color: var(--color-link--hover)\n\n////////////////////////////////////////////////////////////////////////////////\n// Toggles for theme\n////////////////////////////////////////////////////////////////////////////////\n.no-js .theme-toggle-container  // don't show theme toggle if there's no JS\n  display: none\n\n.theme-toggle-container\n  vertical-align: middle\n\n.theme-toggle\n  cursor: pointer\n  border: none\n  padding: 0\n  background: transparent\n\n.theme-toggle svg\n  vertical-align: middle\n  height: 1rem\n  width: 1rem\n  color: var(--color-foreground-primary)\n  display: none\n\n.theme-toggle-header\n  float: left\n  padding: 1rem 0.5rem\n\n////////////////////////////////////////////////////////////////////////////////\n// Toggles for elements\n////////////////////////////////////////////////////////////////////////////////\n.toc-overlay-icon, .nav-overlay-icon\n  display: none\n  cursor: pointer\n\n  .icon\n    color: var(--color-foreground-secondary)\n    height: 1rem\n    width: 1rem\n\n.toc-header-icon, .nav-overlay-icon\n  // for when we set display: flex\n  justify-content: center\n  align-items: center\n\n.toc-content-icon\n  height: 1.5rem\n  width: 1.5rem\n\n.content-icon-container\n  float: right\n  display: flex\n  margin-top: 1.5rem\n  margin-left: 1rem\n  margin-bottom: 1rem\n  gap: 0.5rem\n\n  .edit-this-page svg\n    color: inherit\n    height: 1rem\n    width: 1rem\n\n.sidebar-toggle\n  position: absolute\n  display: none\n// <debugging things>\n.sidebar-toggle[name=\"__toc\"]\n  left: 20px\n.sidebar-toggle:checked\n  left: 40px\n// </debugging things>\n\n.overlay\n  position: fixed\n  top: 0\n  width: 0\n  height: 0\n\n  transition: width 0ms, height 0ms, opacity 250ms ease-out\n\n  opacity: 0\n  background-color: rgba(0, 0, 0, 0.54)\n.sidebar-overlay\n  z-index: 20\n.toc-overlay\n  z-index: 40\n\n// Keep things on top and smooth.\n.sidebar-drawer\n  z-index: 30\n  transition: left 250ms ease-in-out\n.toc-drawer\n  z-index: 50\n  transition: right 250ms ease-in-out\n\n// Show the Sidebar\n#__navigation:checked\n  & ~ .sidebar-overlay\n    width: 100%\n    height: 100%\n    opacity: 1\n  & ~ .page\n    .sidebar-drawer\n      top: 0\n      left: 0\n      // Show the toc sidebar\n#__toc:checked\n  & ~ .toc-overlay\n    width: 100%\n    height: 100%\n    opacity: 1\n  & ~ .page\n    .toc-drawer\n      top: 0\n      right: 0\n\n////////////////////////////////////////////////////////////////////////////////\n// Back to top\n////////////////////////////////////////////////////////////////////////////////\n.back-to-top\n  text-decoration: none\n\n  display: none\n  position: fixed\n  left: 0\n  top: 1rem\n  padding: 0.5rem\n  padding-right: 0.75rem\n  border-radius: 1rem\n  font-size: 0.8125rem\n\n  background: var(--color-background-primary)\n  box-shadow: 0 0.2rem 0.5rem rgba(0, 0, 0, 0.05), #6b728080 0px 0px 1px 0px\n\n  z-index: 10\n\n  margin-left: 50%\n  transform: translateX(-50%)\n  svg\n    height: 1rem\n    width: 1rem\n    fill: currentColor\n    display: inline-block\n\n  span\n    margin-left: 0.25rem\n\n  .show-back-to-top &\n    display: flex\n    align-items: center\n\n////////////////////////////////////////////////////////////////////////////////\n// Responsive layouting\n////////////////////////////////////////////////////////////////////////////////\n// Make things a bit bigger on bigger screens.\n@media (min-width: $full-width + $sidebar-width)\n  html\n    font-size: 110%\n\n@media (max-width: $full-width)\n  // Collapse \"toc\" into the icon.\n  .toc-content-icon\n    display: flex\n  .toc-drawer\n    position: fixed\n    height: 100vh\n    top: 0\n    right: -$sidebar-width\n    border-left: 1px solid var(--color-background-muted)\n  .toc-tree\n    border-left: none\n    font-size: var(--toc-font-size--mobile)\n\n  // Accomodate for a changed content width.\n  .sidebar-drawer\n    width: calc((100% - #{$full-width - $sidebar-width}) / 2 + #{$sidebar-width})\n\n@media (max-width: $full-width - $sidebar-width)\n  // Collapse \"navigation\".\n  .nav-overlay-icon\n    display: flex\n  .sidebar-drawer\n    position: fixed\n    height: 100vh\n    width: $sidebar-width\n\n    top: 0\n    left: -$sidebar-width\n\n  // Swap which icon is visible.\n  .toc-header-icon\n    display: flex\n  .toc-content-icon, .theme-toggle-content\n    display: none\n  .theme-toggle-header\n    display: block\n\n  // Show the header.\n  .mobile-header\n    position: sticky\n    top: 0\n    display: flex\n    justify-content: space-between\n    align-items: center\n\n    .header-left,\n    .header-right\n      display: flex\n      height: var(--header-height)\n      padding: 0 var(--header-padding)\n      label\n        height: 100%\n        width: 100%\n        user-select: none\n\n  .nav-overlay-icon .icon,\n  .theme-toggle svg\n    height: 1.25rem\n    width: 1.25rem\n\n  // Add a scroll margin for the content\n  :target\n    scroll-margin-top: var(--header-height)\n\n  // Show back-to-top below the header\n  .back-to-top\n    top: calc(var(--header-height) + 0.5rem)\n\n  // Center the page, and accommodate for the header.\n  .page\n    flex-direction: column\n    justify-content: center\n  .content\n    margin-left: auto\n    margin-right: auto\n\n@media (max-width: $content-width + 2* $content-padding)\n  // Content should respect window limits.\n  .content\n    width: 100%\n    overflow-x: auto\n\n@media (max-width: $content-width)\n  .content\n    padding: 0 $content-padding--small\n    // Don't float sidebars to the right.\n  article aside.sidebar\n    float: none\n    width: 100%\n    margin: 1rem 0\n","//\n// The design here is strongly inspired by mkdocs-material.\n.admonition, .topic\n  margin: 1rem auto\n  padding: 0 0.5rem 0.5rem 0.5rem\n\n  background: var(--color-admonition-background)\n\n  border-radius: 0.2rem\n  box-shadow: 0 0.2rem 0.5rem rgba(0, 0, 0, 0.05), 0 0 0.0625rem rgba(0, 0, 0, 0.1)\n\n  font-size: var(--admonition-font-size)\n\n  overflow: hidden\n  page-break-inside: avoid\n\n  // First element should have no margin, since the title has it.\n  > :nth-child(2)\n    margin-top: 0\n\n  // Last item should have no margin, since we'll control that w/ padding\n  > :last-child\n    margin-bottom: 0\n\n.admonition p.admonition-title,\np.topic-title\n  position: relative\n  margin: 0 -0.5rem 0.5rem\n  padding-left: 2rem\n  padding-right: .5rem\n  padding-top: .4rem\n  padding-bottom: .4rem\n\n  font-weight: 500\n  font-size: var(--admonition-title-font-size)\n  line-height: 1.3\n\n    // Our fancy icon\n  &::before\n    content: \"\"\n    position: absolute\n    left: 0.5rem\n    width: 1rem\n    height: 1rem\n\n// Default styles\np.admonition-title\n  background-color: var(--color-admonition-title-background)\n  &::before\n    background-color: var(--color-admonition-title)\n    mask-image: var(--icon-admonition-default)\n    mask-repeat: no-repeat\n\np.topic-title\n  background-color: var(--color-topic-title-background)\n  &::before\n    background-color: var(--color-topic-title)\n    mask-image: var(--icon-topic-default)\n    mask-repeat: no-repeat\n\n//\n// Variants\n//\n.admonition\n  border-left: 0.2rem solid var(--color-admonition-title)\n\n  @each $type, $value in $admonitions\n    &.#{$type}\n      border-left-color: var(--color-admonition-title--#{$type})\n      > .admonition-title\n        background-color: var(--color-admonition-title-background--#{$type})\n        &::before\n          background-color: var(--color-admonition-title--#{$type})\n          mask-image: var(--icon-#{nth($value, 2)})\n\n.admonition-todo > .admonition-title\n  text-transform: uppercase\n","// This file stylizes the API documentation (stuff generated by autodoc). It's\n// deeply nested due to how autodoc structures the HTML without enough classes\n// to select the relevant items.\n\n// API docs!\ndl[class]:not(.option-list):not(.field-list):not(.footnote):not(.glossary):not(.simple)\n  // Tweak the spacing of all the things!\n  dd\n    margin-left: 2rem\n    > :first-child\n      margin-top: 0.125rem\n    > :last-child\n      margin-bottom: 0.75rem\n\n  // This is used for the arguments\n  .field-list\n    margin-bottom: 0.75rem\n\n    // \"Headings\" (like \"Parameters\" and \"Return\")\n    > dt\n      text-transform: uppercase\n      font-size: var(--font-size--small)\n\n    dd:empty\n      margin-bottom: 0.5rem\n    dd > ul\n      margin-left: -1.2rem\n      > li\n        > p:nth-child(2)\n          margin-top: 0\n        // When the last-empty-paragraph follows a paragraph, it doesn't need\n        // to augument the existing spacing.\n        > p + p:last-child:empty\n          margin-top: 0\n          margin-bottom: 0\n\n  // Colorize the elements\n  > dt\n    color: var(--color-api-overall)\n\n.sig:not(.sig-inline)\n  font-weight: bold\n\n  font-size: var(--api-font-size)\n  font-family: var(--font-stack--monospace)\n\n  margin-left: -0.25rem\n  margin-right: -0.25rem\n  padding-top: 0.25rem\n  padding-bottom: 0.25rem\n  padding-right: 0.5rem\n\n  // These are intentionally em, to properly match the font size.\n  padding-left: 3em\n  text-indent: -2.5em\n\n  border-radius: 0.25rem\n\n  background: var(--color-api-background)\n  transition: background 100ms ease-out\n\n  &:hover\n    background: var(--color-api-background-hover)\n\n  // adjust the size of the [source] link on the right.\n  a.reference\n    .viewcode-link\n      font-weight: normal\n      width: 3.5rem\n\nem.property\n  font-style: normal\n  &:first-child\n    color: var(--color-api-keyword)\n.sig-name\n  color: var(--color-api-name)\n.sig-prename\n  font-weight: normal\n  color: var(--color-api-pre-name)\n.sig-paren\n  color: var(--color-api-paren)\n.sig-param\n  font-style: normal\n\n.versionmodified\n  font-style: italic\ndiv.versionadded, div.versionchanged, div.deprecated\n  p\n    margin-top: 0.125rem\n    margin-bottom: 0.125rem\n\n// Align the [docs] and [source] to the right.\n.viewcode-link, .viewcode-back\n  float: right\n  text-align: right\n",".line-block\n  margin-top: 0.5rem\n  margin-bottom: 0.75rem\n  .line-block\n    margin-top: 0rem\n    margin-bottom: 0rem\n    padding-left: 1rem\n","// Captions\narticle p.caption,\ntable > caption,\n.code-block-caption\n  font-size: var(--font-size--small)\n  text-align: center\n\n// Caption above a TOCTree\n.toctree-wrapper.compound\n  .caption, :not(.caption) > .caption-text\n    font-size: var(--font-size--small)\n    text-transform: uppercase\n\n    text-align: initial\n    margin-bottom: 0\n\n  > ul\n    margin-top: 0\n    margin-bottom: 0\n","// Inline code\ncode.literal, .sig-inline\n  background: var(--color-inline-code-background)\n  border-radius: 0.2em\n  // Make the font smaller, and use padding to recover.\n  font-size: var(--font-size--small--2)\n  padding: 0.1em 0.2em\n\n  pre.literal-block &\n    font-size: inherit\n    padding: 0\n\n  p &\n    border: 1px solid var(--color-background-border)\n\n.sig-inline\n  font-family: var(--font-stack--monospace)\n\n// Code and Literal Blocks\n$code-spacing-vertical: 0.625rem\n$code-spacing-horizontal: 0.875rem\n\n// Wraps every literal block + line numbers.\ndiv[class*=\" highlight-\"],\ndiv[class^=\"highlight-\"]\n  margin: 1em 0\n  display: flex\n\n  .table-wrapper\n    margin: 0\n    padding: 0\n\npre\n  margin: 0\n  padding: 0\n  overflow: auto\n\n  // Needed to have more specificity than pygments' \"pre\" selector. :(\n  article[role=\"main\"] .highlight &\n    line-height: 1.5\n\n  &.literal-block,\n  .highlight &\n    font-size: var(--code-font-size)\n    padding: $code-spacing-vertical $code-spacing-horizontal\n\n  // Make it look like all the other blocks.\n  &.literal-block\n    margin-top: 1rem\n    margin-bottom: 1rem\n\n    border-radius: 0.2rem\n    background-color: var(--color-code-background)\n    color: var(--color-code-foreground)\n\n// All code is always contained in this.\n.highlight\n  width: 100%\n  border-radius: 0.2rem\n\n  // Make line numbers and prompts un-selectable.\n  .gp, span.linenos\n    user-select: none\n    pointer-events: none\n\n  // Expand the line-highlighting.\n  .hll\n    display: block\n    margin-left: -$code-spacing-horizontal\n    margin-right: -$code-spacing-horizontal\n    padding-left: $code-spacing-horizontal\n    padding-right: $code-spacing-horizontal\n\n/* Make code block captions be nicely integrated */\n.code-block-caption\n  display: flex\n  padding: $code-spacing-vertical $code-spacing-horizontal\n\n  border-radius: 0.25rem\n  border-bottom-left-radius: 0\n  border-bottom-right-radius: 0\n  font-weight: 300\n  border-bottom: 1px solid\n\n  background-color: var(--color-code-background)\n  color: var(--color-code-foreground)\n  border-color: var(--color-background-border)\n\n  + div[class]\n    margin-top: 0\n    pre\n      border-top-left-radius: 0\n      border-top-right-radius: 0\n\n// When `html_codeblock_linenos_style` is table.\n.highlighttable\n  width: 100%\n  display: block\n  tbody\n    display: block\n\n  tr\n    display: flex\n\n  // Line numbers\n  td.linenos\n    background-color: var(--color-code-background)\n    color: var(--color-code-foreground)\n    padding: $code-spacing-vertical $code-spacing-horizontal\n    padding-right: 0\n    border-top-left-radius: 0.2rem\n    border-bottom-left-radius: 0.2rem\n\n  .linenodiv\n    padding-right: $code-spacing-horizontal\n    font-size: var(--code-font-size)\n    box-shadow: -0.0625rem 0 var(--color-foreground-border) inset\n\n  // Actual code\n  td.code\n    padding: 0\n    display: block\n    flex: 1\n    overflow: hidden\n\n    .highlight\n      border-top-left-radius: 0\n      border-bottom-left-radius: 0\n\n// When `html_codeblock_linenos_style` is inline.\n.highlight\n  span.linenos\n    display: inline-block\n    padding-left: 0\n    padding-right: $code-spacing-horizontal\n    margin-right: $code-spacing-horizontal\n    box-shadow: -0.0625rem 0 var(--color-foreground-border) inset\n","// Inline Footnote Reference\n.footnote-reference\n  font-size: var(--font-size--small--4)\n  vertical-align: super\n\n// Definition list, listing the content of each note.\n// docutils <= 0.17\ndl.footnote.brackets\n  font-size: var(--font-size--small)\n  color: var(--color-foreground-secondary)\n\n  display: grid\n  grid-template-columns: max-content auto\n  dt\n    margin: 0\n    > .fn-backref\n      margin-left: 0.25rem\n\n    &:after\n      content: \":\"\n\n    .brackets\n      &:before\n        content: \"[\"\n      &:after\n        content: \"]\"\n\n  dd\n    margin: 0\n    padding: 0 1rem\n\n// docutils >= 0.18\naside.footnote\n  font-size: var(--font-size--small)\n  color: var(--color-foreground-secondary)\n\naside.footnote > span,\ndiv.citation > span\n  float: left\n  font-weight: 500\n  padding-right: 0.25rem\n\naside.footnote > p,\ndiv.citation > p\n  margin-left: 2rem\n","//\n// Figures\n//\nimg\n  box-sizing: border-box\n  max-width: 100%\n  height: auto\n\narticle\n  figure, .figure\n    border-radius: 0.2rem\n\n    margin: 0\n    :last-child\n      margin-bottom: 0\n\n  .align-left\n    float: left\n    clear: left\n    margin: 0 1rem 1rem\n\n  .align-right\n    float: right\n    clear: right\n    margin: 0 1rem 1rem\n\n  .align-default,\n  .align-center\n    display: block\n    text-align: center\n    margin-left: auto\n    margin-right: auto\n\n  // WELL, table needs to be stylised like a table.\n  table.align-default\n    display: table\n    text-align: initial\n",".genindex-jumpbox, .domainindex-jumpbox\n  border-top: 1px solid var(--color-background-border)\n  border-bottom: 1px solid var(--color-background-border)\n  padding: 0.25rem\n\n.genindex-section, .domainindex-section\n  h2\n    margin-top: 0.75rem\n    margin-bottom: 0.5rem\n  ul\n    margin-top: 0\n    margin-bottom: 0\n","ul,\nol\n  padding-left: 1.2rem\n\n  // Space lists out like paragraphs\n  margin-top: 1rem\n  margin-bottom: 1rem\n  // reduce margins within li.\n  li\n    > p:first-child\n      margin-top: 0.25rem\n      margin-bottom: 0.25rem\n\n    > p:last-child\n      margin-top: 0.25rem\n\n    > ul,\n    > ol\n      margin-top: 0.5rem\n      margin-bottom: 0.5rem\n\nol\n  &.arabic\n    list-style: decimal\n  &.loweralpha\n    list-style: lower-alpha\n  &.upperalpha\n    list-style: upper-alpha\n  &.lowerroman\n    list-style: lower-roman\n  &.upperroman\n    list-style: upper-roman\n\n// Don't space lists out when they're \"simple\" or in a `.. toctree::`\n.simple,\n.toctree-wrapper\n  li\n    > ul,\n    > ol\n      margin-top: 0\n      margin-bottom: 0\n\n// Definition Lists\n.field-list,\n.option-list,\ndl:not([class]),\ndl.simple,\ndl.footnote,\ndl.glossary\n  dt\n    font-weight: 500\n    margin-top: 0.25rem\n    + dt\n      margin-top: 0\n\n    .classifier::before\n      content: \":\"\n      margin-left: 0.2rem\n      margin-right: 0.2rem\n\n  dd\n    > p:first-child,\n    ul\n      margin-top: 0.125rem\n\n    ul\n      margin-bottom: 0.125rem\n",".math-wrapper\n  width: 100%\n  overflow-x: auto\n\ndiv.math\n  position: relative\n  text-align: center\n\n  .headerlink,\n  &:focus .headerlink\n    display: none\n\n  &:hover .headerlink\n    display: inline-block\n\n  span.eqno\n    position: absolute\n    right: 0.5rem\n    top: 50%\n    transform: translate(0, -50%)\n    z-index: 1\n","// Abbreviations\nabbr[title]\n  cursor: help\n\n// \"Problematic\" content, as identified by Sphinx\n.problematic\n  color: var(--color-problematic)\n\n// Keyboard / Mouse \"instructions\"\nkbd:not(.compound)\n  margin: 0 0.2rem\n  padding: 0 0.2rem\n  border-radius: 0.2rem\n  border: 1px solid var(--color-foreground-border)\n  color: var(--color-foreground-primary)\n  vertical-align: text-bottom\n\n  font-size: var(--font-size--small--3)\n  display: inline-block\n\n  box-shadow: 0 0.0625rem 0 rgba(0, 0, 0, 0.2), inset 0 0 0 0.125rem var(--color-background-primary)\n\n  background-color: var(--color-background-secondary)\n\n// Blockquote\nblockquote\n  border-left: 4px solid var(--color-background-border)\n  background: var(--color-background-secondary)\n\n  margin-left: 0\n  margin-right: 0\n  padding: 0.5rem 1rem\n\n  .attribution\n    font-weight: 600\n    text-align: right\n\n  &.pull-quote,\n  &.highlights\n    font-size: 1.25em\n\n  &.epigraph,\n  &.pull-quote\n    border-left-width: 0\n    border-radius: 0.5rem\n\n  &.highlights\n    border-left-width: 0\n    background: transparent\n\n// Center align embedded-in-text images\np .reference img\n  vertical-align: middle\n","p.rubric\n  line-height: 1.25\n  font-weight: bold\n  font-size: 1.125em\n\n  // For Numpy-style documentation that's got rubrics within it.\n  // https://github.com/pradyunsg/furo/discussions/505\n  dd &\n    line-height: inherit\n    font-weight: inherit\n\n    font-size: var(--font-size--small)\n    text-transform: uppercase\n","article .sidebar\n  float: right\n  clear: right\n  width: 30%\n\n  margin-left: 1rem\n  margin-right: 0\n\n  border-radius: 0.2rem\n  background-color: var(--color-background-secondary)\n  border: var(--color-background-border) 1px solid\n\n  > *\n    padding-left: 1rem\n    padding-right: 1rem\n\n  > ul, > ol  // lists need additional padding, because bullets.\n    padding-left: 2.2rem\n\n  .sidebar-title\n    margin: 0\n    padding: 0.5rem 1rem\n    border-bottom: var(--color-background-border) 1px solid\n\n    font-weight: 500\n\n// TODO: subtitle\n// TODO: dedicated variables?\n",".table-wrapper\n  width: 100%\n  overflow-x: auto\n  margin-top: 1rem\n  margin-bottom: 0.5rem\n  padding: 0.2rem 0.2rem 0.75rem\n\ntable.docutils\n  border-radius: 0.2rem\n  border-spacing: 0\n  border-collapse: collapse\n\n  box-shadow: 0 0.2rem 0.5rem rgba(0, 0, 0, 0.05), 0 0 0.0625rem rgba(0, 0, 0, 0.1)\n\n  th\n    background: var(--color-table-header-background)\n\n  td,\n  th\n    // Space things out properly\n    padding: 0 0.25rem\n\n    // Get the borders looking just-right.\n    border-left: 1px solid var(--color-table-border)\n    border-right: 1px solid var(--color-table-border)\n    border-bottom: 1px solid var(--color-table-border)\n\n    p\n      margin: 0.25rem\n\n    &:first-child\n      border-left: none\n    &:last-child\n      border-right: none\n\n    // MyST-parser tables set these classes for control of column alignment\n    &.text-left\n      text-align: left\n    &.text-right\n      text-align: right\n    &.text-center\n      text-align: center\n",":target\n  scroll-margin-top: 0.5rem\n\n@media (max-width: $full-width - $sidebar-width)\n  :target\n    scroll-margin-top: calc(0.5rem + var(--header-height))\n\n  // When a heading is selected\n  section > span:target\n    scroll-margin-top: calc(0.8rem + var(--header-height))\n\n// Permalinks\n.headerlink\n  font-weight: 100\n  user-select: none\n\nh1,\nh2,\nh3,\nh4,\nh5,\nh6,\ndl dt,\np.caption,\nfigcaption p,\ntable > caption,\n.code-block-caption\n  > .headerlink\n    margin-left: 0.5rem\n    visibility: hidden\n  &:hover > .headerlink\n    visibility: visible\n\n  // Don't change to link-like, if someone adds the contents directive.\n  > .toc-backref\n    color: inherit\n    text-decoration-line: none\n\n// Figure and table captions are special.\nfigure:hover > figcaption > p > .headerlink,\ntable:hover > caption > .headerlink\n  visibility: visible\n\n:target >,  // Regular section[id] style anchors\nspan:target ~ // Non-regular span[id] style \"extra\" anchors\n  h1,\n  h2,\n  h3,\n  h4,\n  h5,\n  h6\n    &:nth-of-type(1)\n      background-color: var(--color-highlight-on-target)\n      // .headerlink\n      //   visibility: visible\n      code.literal\n        background-color: transparent\n\ntable:target > caption,\nfigure:target\n  background-color: var(--color-highlight-on-target)\n\n// Inline page contents\n.this-will-duplicate-information-and-it-is-still-useful-here li :target\n  background-color: var(--color-highlight-on-target)\n\n// Code block permalinks\n.literal-block-wrapper:target .code-block-caption\n  background-color: var(--color-highlight-on-target)\n\n// When a definition list item is selected\n//\n//   There isn't really an alternative to !important here, due to the\n//   high-specificity of API documentation's selector.\ndt:target\n  background-color: var(--color-highlight-on-target) !important\n\n// When a footnote reference is selected\n.footnote > dt:target + dd,\n.footnote-reference:target\n  background-color: var(--color-highlight-on-target)\n",".guilabel\n  background-color: var(--color-guilabel-background)\n  border: 1px solid var(--color-guilabel-border)\n  color: var(--color-guilabel-text)\n\n  padding: 0 0.3em\n  border-radius: 0.5em\n  font-size: 0.9em\n","// This file contains the styles used for stylizing the footer that's shown\n// below the content.\n\nfooter\n  font-size: var(--font-size--small)\n  display: flex\n  flex-direction: column\n\n  margin-top: 2rem\n\n// Bottom of page information\n.bottom-of-page\n  display: flex\n  align-items: center\n  justify-content: space-between\n\n  margin-top: 1rem\n  padding-top: 1rem\n  padding-bottom: 1rem\n\n  color: var(--color-foreground-secondary)\n  border-top: 1px solid var(--color-background-border)\n\n  line-height: 1.5\n\n  @media (max-width: $content-width)\n    text-align: center\n    flex-direction: column-reverse\n    gap: 0.25rem\n\n  .left-details\n    font-size: var(--font-size--small)\n\n  .right-details\n    display: flex\n    flex-direction: column\n    gap: 0.25rem\n    text-align: right\n\n  .icons\n    display: flex\n    justify-content: flex-end\n    gap: 0.25rem\n    font-size: 1rem\n\n    a\n      text-decoration: none\n\n    svg,\n    img\n      font-size: 1.125rem\n      height: 1em\n      width: 1em\n\n// Next/Prev page information\n.related-pages\n  a\n    display: flex\n    align-items: center\n\n    text-decoration: none\n    &:hover .page-info .title\n      text-decoration: underline\n      color: var(--color-link)\n      text-decoration-color: var(--color-link-underline)\n\n    svg.furo-related-icon,\n    svg.furo-related-icon > use\n      flex-shrink: 0\n\n      color: var(--color-foreground-border)\n\n      width: 0.75rem\n      height: 0.75rem\n      margin: 0 0.5rem\n\n    &.next-page\n      max-width: 50%\n\n      float: right\n      clear: right\n      text-align: right\n\n    &.prev-page\n      max-width: 50%\n\n      float: left\n      clear: left\n\n      svg\n        transform: rotate(180deg)\n\n.page-info\n  display: flex\n  flex-direction: column\n  overflow-wrap: anywhere\n\n  .next-page &\n    align-items: flex-end\n\n  .context\n    display: flex\n    align-items: center\n\n    padding-bottom: 0.1rem\n\n    color: var(--color-foreground-muted)\n    font-size: var(--font-size--small)\n    text-decoration: none\n","// This file contains the styles for the contents of the left sidebar, which\n// contains the navigation tree, logo, search etc.\n\n////////////////////////////////////////////////////////////////////////////////\n// Brand on top of the scrollable tree.\n////////////////////////////////////////////////////////////////////////////////\n.sidebar-brand\n  display: flex\n  flex-direction: column\n  flex-shrink: 0\n\n  padding: var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal)\n  text-decoration: none\n\n.sidebar-brand-text\n  color: var(--color-sidebar-brand-text)\n  overflow-wrap: break-word\n  margin: var(--sidebar-item-spacing-vertical) 0\n  font-size: 1.5rem\n\n.sidebar-logo-container\n  margin: var(--sidebar-item-spacing-vertical) 0\n\n.sidebar-logo\n  margin: 0 auto\n  display: block\n  max-width: 100%\n\n////////////////////////////////////////////////////////////////////////////////\n// Search\n////////////////////////////////////////////////////////////////////////////////\n.sidebar-search-container\n  display: flex\n  align-items: center\n  margin-top: var(--sidebar-search-space-above)\n\n  position: relative\n\n  background: var(--color-sidebar-search-background)\n  &:hover,\n  &:focus-within\n    background: var(--color-sidebar-search-background--focus)\n\n  &::before\n    content: \"\"\n    position: absolute\n    left: var(--sidebar-item-spacing-horizontal)\n    width: var(--sidebar-search-icon-size)\n    height: var(--sidebar-search-icon-size)\n\n    background-color: var(--color-sidebar-search-icon)\n    mask-image: var(--icon-search)\n\n.sidebar-search\n  box-sizing: border-box\n\n  border: none\n  border-top: 1px solid var(--color-sidebar-search-border)\n  border-bottom: 1px solid var(--color-sidebar-search-border)\n\n  padding-top: var(--sidebar-search-input-spacing-vertical)\n  padding-bottom: var(--sidebar-search-input-spacing-vertical)\n  padding-right: var(--sidebar-search-input-spacing-horizontal)\n  padding-left: calc(var(--sidebar-item-spacing-horizontal) + var(--sidebar-search-input-spacing-horizontal) + var(--sidebar-search-icon-size))\n\n  width: 100%\n\n  color: var(--color-sidebar-search-foreground)\n  background: transparent\n  z-index: 10\n\n  &:focus\n    outline: none\n\n  &::placeholder\n    font-size: var(--sidebar-search-input-font-size)\n\n//\n// Hide Search Matches link\n//\n#searchbox .highlight-link\n  padding: var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal) 0\n  margin: 0\n  text-align: center\n\n  a\n    color: var(--color-sidebar-search-icon)\n    font-size: var(--font-size--small--2)\n\n////////////////////////////////////////////////////////////////////////////////\n// Structure/Skeleton of the navigation tree (left)\n////////////////////////////////////////////////////////////////////////////////\n.sidebar-tree\n  font-size: var(--sidebar-item-font-size)\n  margin-top: var(--sidebar-tree-space-above)\n  margin-bottom: var(--sidebar-item-spacing-vertical)\n\n  ul\n    padding: 0\n    margin-top: 0\n    margin-bottom: 0\n\n    display: flex\n    flex-direction: column\n\n    list-style: none\n\n  li\n    position: relative\n    margin: 0\n\n    > ul\n      margin-left: var(--sidebar-item-spacing-horizontal)\n\n  .icon\n    color: var(--color-sidebar-link-text)\n\n  .reference\n    box-sizing: border-box\n    color: var(--color-sidebar-link-text)\n\n    // Fill the parent.\n    display: inline-block\n    line-height: var(--sidebar-item-line-height)\n    text-decoration: none\n\n    // Don't allow long words to cause wrapping.\n    overflow-wrap: anywhere\n\n    height: 100%\n    width: 100%\n\n    padding: var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal)\n\n    &:hover\n      background: var(--color-sidebar-item-background--hover)\n\n    // Add a nice little \"external-link\" arrow here.\n    &.external::after\n      content: url('data:image/svg+xml,<svg width=\"12\" height=\"12\" xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 24 24\" stroke-width=\"1.5\" stroke=\"%23607D8B\" fill=\"none\" stroke-linecap=\"round\" stroke-linejoin=\"round\"><path stroke=\"none\" d=\"M0 0h24v24H0z\"/><path d=\"M11 7h-5a2 2 0 0 0 -2 2v9a2 2 0 0 0 2 2h9a2 2 0 0 0 2 -2v-5\" /><line x1=\"10\" y1=\"14\" x2=\"20\" y2=\"4\" /><polyline points=\"15 4 20 4 20 9\" /></svg>')\n      margin: 0 0.25rem\n      vertical-align: middle\n      color: var(--color-sidebar-link-text)\n\n  // Make the current page reference bold.\n  .current-page > .reference\n    font-weight: bold\n\n  label\n    position: absolute\n    top: 0\n    right: 0\n    height: var(--sidebar-item-height)\n    width: var(--sidebar-expander-width)\n\n    cursor: pointer\n    user-select: none\n\n    display: flex\n    justify-content: center\n    align-items: center\n\n  .caption, :not(.caption) > .caption-text\n    font-size: var(--sidebar-caption-font-size)\n    color: var(--color-sidebar-caption-text)\n\n    font-weight: bold\n    text-transform: uppercase\n\n    margin: var(--sidebar-caption-space-above) 0 0 0\n    padding: var(--sidebar-item-spacing-vertical) var(--sidebar-item-spacing-horizontal)\n\n  // If it has children, add a bit more padding to wrap the content to avoid\n  // overlapping with the <label>\n  li.has-children\n    > .reference\n      padding-right: var(--sidebar-expander-width)\n\n  // Colorize the top-level list items and icon.\n  .toctree-l1\n    & > .reference,\n    & > label .icon\n      color: var(--color-sidebar-link-text--top-level)\n\n  // Color changes on hover\n  label\n    background: var(--color-sidebar-item-expander-background)\n    &:hover\n      background: var(--color-sidebar-item-expander-background--hover)\n\n  .current > .reference\n    background: var(--color-sidebar-item-background--current)\n    &:hover\n      background: var(--color-sidebar-item-background--hover)\n\n.toctree-checkbox\n  position: absolute\n  display: none\n\n////////////////////////////////////////////////////////////////////////////////\n// Togglable expand/collapse\n////////////////////////////////////////////////////////////////////////////////\n.toctree-checkbox\n  ~ ul\n    display: none\n\n  ~ label .icon svg\n    transform: rotate(90deg)\n\n.toctree-checkbox:checked\n  ~ ul\n    display: block\n\n  ~ label .icon svg\n    transform: rotate(-90deg)\n","// This file contains the styles for the contents of the right sidebar, which\n// contains the table of contents for the current page.\n.toc-title-container\n  padding: var(--toc-title-padding)\n  padding-top: var(--toc-spacing-vertical)\n\n.toc-title\n  color: var(--color-toc-title-text)\n  font-size: var(--toc-title-font-size)\n  padding-left: var(--toc-spacing-horizontal)\n  text-transform: uppercase\n\n// If the ToC is not present, hide these elements coz they're not relevant.\n.no-toc\n  display: none\n\n.toc-tree-container\n  padding-bottom: var(--toc-spacing-vertical)\n\n.toc-tree\n  font-size: var(--toc-font-size)\n  line-height: 1.3\n  border-left: 1px solid var(--color-background-border)\n\n  padding-left: calc(var(--toc-spacing-horizontal) - var(--toc-item-spacing-horizontal))\n\n  // Hide the first \"top level\" bullet.\n  > ul > li:first-child\n    padding-top: 0\n    & > ul\n      padding-left: 0\n    & > a\n      display: none\n\n  ul\n    list-style-type: none\n    margin-top: 0\n    margin-bottom: 0\n    padding-left: var(--toc-item-spacing-horizontal)\n  li\n    padding-top: var(--toc-item-spacing-vertical)\n\n    &.scroll-current >.reference\n      color: var(--color-toc-item-text--active)\n      font-weight: bold\n\n  .reference\n    color: var(--color-toc-item-text)\n    text-decoration: none\n    overflow-wrap: anywhere\n\n.toc-scroll\n  max-height: 100vh\n  overflow-y: scroll\n\n// Be very annoying when someone includes the table of contents\n.contents:not(.this-will-duplicate-information-and-it-is-still-useful-here)\n  color: var(--color-problematic)\n  background: rgba(255, 0, 0, 0.25)\n  &::before\n    content: \"ERROR: Adding a table of contents in Furo-based documentation is unnecessary, and does not work well with existing styling.Add a 'this-will-duplicate-information-and-it-is-still-useful-here' class, if you want an escape hatch.\"\n","// Shameful hacks, to work around bugs.\n\n// MyST parser doesn't correctly generate classes, to align table contents.\n// https://github.com/executablebooks/MyST-Parser/issues/412\n.text-align\\:left > p\n  text-align: left\n\n.text-align\\:center > p\n  text-align: center\n\n.text-align\\:right > p\n  text-align: right\n"],"names":[],"sourceRoot":""}
\ No newline at end of file
diff --git a/docs/_build/html/basics.html b/docs/_build/html/basics.html
deleted file mode 100644
index 6f9607131..000000000
--- a/docs/_build/html/basics.html
+++ /dev/null
@@ -1,525 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="Devices" href="modules/devices.html" /><link rel="prev" title="Installation" href="installation.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Basics - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="basics">
-<h1>Basics<a class="headerlink" href="#basics" title="Link to this heading">#</a></h1>
-<section id="initialization">
-<h2>Initialization<a class="headerlink" href="#initialization" title="Link to this heading">#</a></h2>
-<p>Before use Warp should be explicitly initialized with the <code class="docutils literal notranslate"><span class="pre">wp.init()</span></code> method as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>Warp will print some startup information about the compute devices available, driver versions, and the location
-for any generated kernel code, e.g.:</p>
-<div class="highlight-bat notranslate"><div class="highlight"><pre><span></span>Warp 1.0.0 initialized:
-CUDA Toolkit: 11.8, Driver: 12.1
-Devices:
-    <span class="s2">&quot;cpu&quot;</span>    <span class="p">|</span> AMD64 Family 25 Model 33 Stepping 0, AuthenticAMD
-    <span class="s2">&quot;cuda:0&quot;</span> <span class="p">|</span> NVIDIA GeForce RTX 4080 (sm_89)
-Kernel cache: C:\Users\mmacklin\AppData\Local\NVIDIA Corporation\warp\Cache\1.0.0
-</pre></div>
-</div>
-</section>
-<section id="kernels">
-<h2>Kernels<a class="headerlink" href="#kernels" title="Link to this heading">#</a></h2>
-<p>In Warp, compute kernels are defined as Python functions and annotated with the <code class="docutils literal notranslate"><span class="pre">&#64;wp.kernel</span></code> decorator, as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">simple_kernel</span><span class="p">(</span><span class="n">a</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-                  <span class="n">b</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-                  <span class="n">c</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-
-    <span class="c1"># get thread index</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="c1"># load two vec3s</span>
-    <span class="n">x</span> <span class="o">=</span> <span class="n">a</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span>
-    <span class="n">y</span> <span class="o">=</span> <span class="n">b</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span>
-
-    <span class="c1"># compute the dot product between vectors</span>
-    <span class="n">r</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">dot</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
-
-    <span class="c1"># write result back to memory</span>
-    <span class="n">c</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">r</span>
-</pre></div>
-</div>
-<p>Because Warp kernels are compiled to native C++/CUDA code, all the function input arguments must be statically typed. This allows
-Warp to generate fast code that executes at essentially native speeds. Because kernels may be run on either the CPU
-or GPU, they cannot access arbitrary global state from the Python environment. Instead they must read and write data
-through their input parameters such as arrays.</p>
-<p>Warp kernels functions have a 1:1 correspondence with CUDA kernels, to launch a kernel with 1024 threads, we use
-<a class="reference internal" href="modules/runtime.html#warp.launch" title="warp.launch"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.launch()</span></code></a> as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">simple_kernel</span><span class="p">,</span> <span class="c1"># kernel to launch</span>
-          <span class="n">dim</span><span class="o">=</span><span class="mi">1024</span><span class="p">,</span>             <span class="c1"># number of threads</span>
-          <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">],</span>     <span class="c1"># parameters</span>
-          <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>        <span class="c1"># execution device</span>
-</pre></div>
-</div>
-<p>Inside the kernel we can retrieve the <em>thread index</em> of the each thread using the <code class="docutils literal notranslate"><span class="pre">wp.tid()</span></code> builtin function:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># get thread index</span>
-<span class="n">i</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>Kernels can be launched with 1D, 2D, 3D, or 4D grids of threads, e.g.: to launch a 2D grid of threads to process a 1024x1024 image we could write:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">compute_image</span><span class="p">,</span>
-          <span class="n">dim</span><span class="o">=</span><span class="p">(</span><span class="mi">1024</span><span class="p">,</span> <span class="mi">1024</span><span class="p">),</span>
-          <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">img</span><span class="p">],</span>
-          <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Then, inside the kernel we can retrieve a 2D thread index as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># get thread index</span>
-<span class="n">i</span><span class="p">,</span> <span class="n">j</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-<span class="c1"># write out a color value for each pixel</span>
-<span class="n">color</span><span class="p">[</span><span class="n">i</span><span class="p">,</span> <span class="n">j</span><span class="p">]</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="n">r</span><span class="p">,</span> <span class="n">g</span><span class="p">,</span> <span class="n">b</span><span class="p">)</span>
-</pre></div>
-</div>
-<section id="example-changing-the-kernel-cache-directory">
-<span id="example-cache-management"></span><h3>Example: Changing the kernel cache directory<a class="headerlink" href="#example-changing-the-kernel-cache-directory" title="Link to this heading">#</a></h3>
-<p>The following example illustrates how the location for generated and compiled
-kernel code can be changed before and after calling <code class="docutils literal notranslate"><span class="pre">wp.init()</span></code>.</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">os</span>
-
-<span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-
-<span class="n">example_dir</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">dirname</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">realpath</span><span class="p">(</span><span class="vm">__file__</span><span class="p">))</span>
-
-<span class="c1"># set default cache directory before wp.init()</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">kernel_cache_dir</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">example_dir</span><span class="p">,</span> <span class="s2">&quot;tmp&quot;</span><span class="p">,</span> <span class="s2">&quot;warpcache1&quot;</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-
-<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;+++ Current cache directory: &quot;</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">kernel_cache_dir</span><span class="p">)</span>
-
-<span class="c1"># change cache directory after wp.init()</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">build</span><span class="o">.</span><span class="n">init_kernel_cache</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">example_dir</span><span class="p">,</span> <span class="s2">&quot;tmp&quot;</span><span class="p">,</span> <span class="s2">&quot;warpcache2&quot;</span><span class="p">))</span>
-
-<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;+++ Current cache directory: &quot;</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">kernel_cache_dir</span><span class="p">)</span>
-
-<span class="c1"># clear kernel cache (forces fresh kernel builds every time)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">build</span><span class="o">.</span><span class="n">clear_kernel_cache</span><span class="p">()</span>
-
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">basic</span><span class="p">(</span><span class="n">x</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">x</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="nb">float</span><span class="p">(</span><span class="n">tid</span><span class="p">)</span>
-
-
-<span class="n">device</span> <span class="o">=</span> <span class="s2">&quot;cpu&quot;</span>
-<span class="n">n</span> <span class="o">=</span> <span class="mi">10</span>
-<span class="n">x</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">basic</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">n</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">x</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-<span class="nb">print</span><span class="p">(</span><span class="n">x</span><span class="o">.</span><span class="n">numpy</span><span class="p">())</span>
-</pre></div>
-</div>
-</section>
-</section>
-<section id="arrays">
-<h2>Arrays<a class="headerlink" href="#arrays" title="Link to this heading">#</a></h2>
-<p>Memory allocations are exposed via the <code class="docutils literal notranslate"><span class="pre">wp.array</span></code> type. Arrays wrap an underlying memory allocation that may live in
-either host (CPU), or device (GPU) memory. Arrays are strongly typed and store a linear sequence of built-in values
-(<code class="docutils literal notranslate"><span class="pre">float,</span></code>, <code class="docutils literal notranslate"><span class="pre">int</span></code>, <code class="docutils literal notranslate"><span class="pre">vec3</span></code>, <code class="docutils literal notranslate"><span class="pre">matrix33</span></code>, etc).</p>
-<p>Arrays can be allocated similar to PyTorch:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># allocate an uninitialized array of vec3s</span>
-<span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="n">n</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="c1"># allocate a zero-initialized array of quaternions</span>
-<span class="n">q</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="n">n</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">quat</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="c1"># allocate and initialize an array from a NumPy array</span>
-<span class="c1"># will be automatically transferred to the specified device</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="mi">10</span><span class="p">,</span> <span class="mi">3</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-<span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">from_numpy</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>By default, Warp arrays that are initialized from external data (e.g.: NumPy, Lists, Tuples) will create a copy the data to new memory for the
-device specified. However, it is possible for arrays to alias external memory using the <code class="docutils literal notranslate"><span class="pre">copy=False</span></code> parameter to the
-array constructor provided the input is contiguous and on the same device. See the <a class="reference internal" href="modules/interoperability.html"><span class="doc">Interoperability</span></a>
-section for more details on sharing memory with external frameworks.</p>
-<p>To read GPU array data back to CPU memory we can use the <code class="docutils literal notranslate"><span class="pre">array.numpy()</span></code> method:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># bring data from device back to host</span>
-<span class="n">view</span> <span class="o">=</span> <span class="n">device_array</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>This will automatically synchronize with the GPU to ensure that any outstanding work has finished, and will
-copy the array back to CPU memory where it is passed to NumPy. Calling <code class="docutils literal notranslate"><span class="pre">array.numpy()</span></code> on a CPU array will return
-a zero-copy NumPy view onto the Warp data.</p>
-</section>
-<section id="user-functions">
-<h2>User Functions<a class="headerlink" href="#user-functions" title="Link to this heading">#</a></h2>
-<p>Users can write their own functions using the <code class="docutils literal notranslate"><span class="pre">&#64;wp.func</span></code> decorator, for example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func</span>
-<span class="k">def</span> <span class="nf">square</span><span class="p">(</span><span class="n">x</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
-    <span class="k">return</span> <span class="n">x</span><span class="o">*</span><span class="n">x</span>
-</pre></div>
-</div>
-<p>User functions can be called freely from within kernels inside the same module and accept arrays as inputs.</p>
-</section>
-<section id="compilation-model">
-<h2>Compilation Model<a class="headerlink" href="#compilation-model" title="Link to this heading">#</a></h2>
-<p>Warp uses a Python-&gt;C++/CUDA compilation model that generates kernel code from Python function definitions. All kernels belonging to a Python module are runtime compiled into dynamic libraries and PTX, the result is then cached between application restarts for fast startup times.</p>
-<p>Note that compilation is triggered on the first kernel launch for that module. Any kernels registered in the module with <code class="docutils literal notranslate"><span class="pre">&#64;wp.kernel</span></code> will be included in the shared library.</p>
-<img alt="_images/compiler_pipeline.png" src="_images/compiler_pipeline.png" />
-</section>
-<section id="language-details">
-<h2>Language Details<a class="headerlink" href="#language-details" title="Link to this heading">#</a></h2>
-<p>To support GPU computation and differentiability, there are some differences from the CPython runtime.</p>
-<section id="built-in-types">
-<h3>Built-in Types<a class="headerlink" href="#built-in-types" title="Link to this heading">#</a></h3>
-<p>Warp supports a number of built-in math types similar to high-level shading languages, for example <code class="docutils literal notranslate"><span class="pre">vec2,</span> <span class="pre">vec3,</span> <span class="pre">vec4,</span> <span class="pre">mat22,</span> <span class="pre">mat33,</span> <span class="pre">mat44,</span> <span class="pre">quat,</span> <span class="pre">array</span></code>. All built-in types have value semantics so that expressions such as <code class="docutils literal notranslate"><span class="pre">a</span> <span class="pre">=</span> <span class="pre">b</span></code> generate a copy of the variable b rather than a reference.</p>
-</section>
-<section id="strong-typing">
-<h3>Strong Typing<a class="headerlink" href="#strong-typing" title="Link to this heading">#</a></h3>
-<p>Unlike Python, in Warp all variables must be typed. Types are inferred from source expressions and function signatures using the Python typing extensions. All kernel parameters must be annotated with the appropriate type, for example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">simple_kernel</span><span class="p">(</span><span class="n">a</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">vec3</span><span class="p">),</span>
-                  <span class="n">b</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">vec3</span><span class="p">),</span>
-                  <span class="n">c</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
-</pre></div>
-</div>
-<p>Tuple initialization is not supported, instead variables should be explicitly typed:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># invalid</span>
-<span class="n">a</span> <span class="o">=</span> <span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">)</span>
-
-<span class="c1"># valid</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-<section id="limitations-and-unsupported-features">
-<h3>Limitations and Unsupported Features<a class="headerlink" href="#limitations-and-unsupported-features" title="Link to this heading">#</a></h3>
-<p>See <a class="reference internal" href="limitations.html"><span class="doc">Limitations</span></a> for a list of Warp limitations and unsupported features.</p>
-</section>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="modules/devices.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Devices</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="installation.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Installation</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Basics</a><ul>
-<li><a class="reference internal" href="#initialization">Initialization</a></li>
-<li><a class="reference internal" href="#kernels">Kernels</a><ul>
-<li><a class="reference internal" href="#example-changing-the-kernel-cache-directory">Example: Changing the kernel cache directory</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#arrays">Arrays</a></li>
-<li><a class="reference internal" href="#user-functions">User Functions</a></li>
-<li><a class="reference internal" href="#compilation-model">Compilation Model</a></li>
-<li><a class="reference internal" href="#language-details">Language Details</a><ul>
-<li><a class="reference internal" href="#built-in-types">Built-in Types</a></li>
-<li><a class="reference internal" href="#strong-typing">Strong Typing</a></li>
-<li><a class="reference internal" href="#limitations-and-unsupported-features">Limitations and Unsupported Features</a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/configuration.html b/docs/_build/html/configuration.html
deleted file mode 100644
index f4e4539fa..000000000
--- a/docs/_build/html/configuration.html
+++ /dev/null
@@ -1,530 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="Debugging" href="debugging.html" /><link rel="prev" title="Interoperability" href="modules/interoperability.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Runtime Settings - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="runtime-settings">
-<h1>Runtime Settings<a class="headerlink" href="#runtime-settings" title="Link to this heading">#</a></h1>
-<p>Warp has settings at the global, module, and kernel level that can be used to fine-tune the compilation and verbosity
-of Warp programs. In cases in which a setting can be changed at multiple levels (e.g <code class="docutils literal notranslate"><span class="pre">enable_backward</span></code>),
-the setting at the more-specific scope takes precedence.</p>
-<section id="global-settings">
-<h2>Global Settings<a class="headerlink" href="#global-settings" title="Link to this heading">#</a></h2>
-<p>To change a setting, prepend <code class="docutils literal notranslate"><span class="pre">wp.config.</span></code> to the name of the variable and assign a value to it.
-Some settings may be changed on the fly, while others need to be set prior to calling <code class="docutils literal notranslate"><span class="pre">wp.init()</span></code> to take effect.</p>
-<p>For example, the location of the user kernel cache can be changed with:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">os</span>
-
-<span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-
-<span class="n">example_dir</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">dirname</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">realpath</span><span class="p">(</span><span class="vm">__file__</span><span class="p">))</span>
-
-<span class="c1"># set default cache directory before wp.init()</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">kernel_cache_dir</span> <span class="o">=</span> <span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">example_dir</span><span class="p">,</span> <span class="s2">&quot;tmp&quot;</span><span class="p">,</span> <span class="s2">&quot;warpcache1&quot;</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-</pre></div>
-</div>
-<section id="basic-global-settings">
-<h3>Basic Global Settings<a class="headerlink" href="#basic-global-settings" title="Link to this heading">#</a></h3>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<thead>
-<tr class="row-odd"><th class="head"><p>Field</p></th>
-<th class="head"><p>Type</p></th>
-<th class="head"><p>Default Value</p></th>
-<th class="head"><p>Description</p></th>
-</tr>
-</thead>
-<tbody>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">verify_fp</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, Warp will check that inputs and outputs are finite before
-and/or after various operations. <strong>Has performance implications.</strong></p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">verify_cuda</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, Warp will check for CUDA errors after every launch and
-memory operation. CUDA error verification cannot be used during graph
-capture. <strong>Has performance implications.</strong></p></td>
-</tr>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">print_launches</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, Warp will print details of every kernel launch to standard
-out (e.g. launch dimensions, inputs, outputs, device, etc.).
-<strong>Has performance implications.</strong></p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">mode</span></code></p></td>
-<td><p>String</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">&quot;release&quot;</span></code></p></td>
-<td><p>Controls whether to compile Warp kernels in debug or release mode.
-Valid choices are <code class="docutils literal notranslate"><span class="pre">&quot;release&quot;</span></code> or <code class="docutils literal notranslate"><span class="pre">&quot;debug&quot;</span></code>.
-<strong>Has performance implications.</strong></p></td>
-</tr>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">verbose</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, additional information will be printed to standard out
-during code generation, compilation, etc.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">quiet</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, Warp module initialization messages will be disabled.
-This setting does not affect error messages and warnings.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">kernel_cache_dir</span></code></p></td>
-<td><p>String</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">None</span></code></p></td>
-<td><p>The path to the directory used for the user kernel cache. Subdirectories
-named <code class="docutils literal notranslate"><span class="pre">gen</span></code> and <code class="docutils literal notranslate"><span class="pre">bin</span></code> will be created in this directory. If <code class="docutils literal notranslate"><span class="pre">None</span></code>,
-a directory will be automatically determined using
-<a class="reference external" href="https://github.com/ActiveState/appdirs">appdirs.user_cache_directory</a></p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">enable_backward</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">True</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, backward passes of kernels will be compiled by default.
-Disabling this setting can reduce kernel compilation times.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</section>
-<section id="advanced-global-settings">
-<h3>Advanced Global Settings<a class="headerlink" href="#advanced-global-settings" title="Link to this heading">#</a></h3>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<thead>
-<tr class="row-odd"><th class="head"><p>Field</p></th>
-<th class="head"><p>Type</p></th>
-<th class="head"><p>Default Value</p></th>
-<th class="head"><p>Description</p></th>
-</tr>
-</thead>
-<tbody>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">cache_kernels</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">True</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, kernels that have already been compiled from previous
-application launches will not be recompiled.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">cuda_output</span></code></p></td>
-<td><p>String</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">None</span></code></p></td>
-<td><p>The preferred CUDA output format for kernels. Valid choices are <code class="docutils literal notranslate"><span class="pre">None</span></code>,
-<code class="docutils literal notranslate"><span class="pre">&quot;ptx&quot;</span></code>, and <code class="docutils literal notranslate"><span class="pre">&quot;cubin&quot;</span></code>. If <code class="docutils literal notranslate"><span class="pre">None</span></code>, a format will be determined
-automatically.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">ptx_target_arch</span></code></p></td>
-<td><p>Integer</p></td>
-<td><p>70</p></td>
-<td><p>The target architecture for PTX generation.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">llvm_cuda</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, Clang/LLVM will be used to compile CUDA code instead of
-NVTRC.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</section>
-</section>
-<section id="module-settings">
-<h2>Module Settings<a class="headerlink" href="#module-settings" title="Link to this heading">#</a></h2>
-<p>Module-level settings to control runtime compilation and code generation may be changed by passing a dictionary of
-option pairs to <code class="docutils literal notranslate"><span class="pre">wp.set_module_options()</span></code>.</p>
-<p>For example, compilation of backward passes for the kernel in an entire module can be disabled with:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">set_module_options</span><span class="p">({</span><span class="s2">&quot;enable_backward&quot;</span><span class="p">:</span> <span class="kc">False</span><span class="p">})</span>
-</pre></div>
-</div>
-<p>The options for a module can also be queried using <code class="docutils literal notranslate"><span class="pre">wp.get_module_options()</span></code>.</p>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<thead>
-<tr class="row-odd"><th class="head"><p>Field</p></th>
-<th class="head"><p>Type</p></th>
-<th class="head"><p>Default Value</p></th>
-<th class="head"><p>Description</p></th>
-</tr>
-</thead>
-<tbody>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">mode</span></code></p></td>
-<td><p>String</p></td>
-<td><p>Global
-setting</p></td>
-<td><p>Controls whether to compile the module’s kernels in debug or release
-mode by default. Valid choices are <code class="docutils literal notranslate"><span class="pre">&quot;release&quot;</span></code> or <code class="docutils literal notranslate"><span class="pre">&quot;debug&quot;</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">max_unroll</span></code></p></td>
-<td><p>Integer</p></td>
-<td><p>16</p></td>
-<td><p>The maximum fixed-size loop to unroll. Note that <code class="docutils literal notranslate"><span class="pre">max_unroll</span></code> does not
-consider the total number of iterations in nested loops. This can result
-in a large amount of automatically generated code if each nested loop is
-below the <code class="docutils literal notranslate"><span class="pre">max_unroll</span></code> threshold.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">enable_backward</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p>Global
-setting</p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, backward passes of kernels will be compiled by default.
-Valid choices are <code class="docutils literal notranslate"><span class="pre">&quot;release&quot;</span></code> or <code class="docutils literal notranslate"><span class="pre">&quot;debug&quot;</span></code>.</p></td>
-</tr>
-<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fast_math</span></code></p></td>
-<td><p>Boolean</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">False</span></code></p></td>
-<td><p>If <code class="docutils literal notranslate"><span class="pre">True</span></code>, CUDA kernels will be compiled with the <code class="docutils literal notranslate"><span class="pre">--use_fast_math</span></code>
-compiler option, which enables some fast math operations that are faster
-but less accurate.</p></td>
-</tr>
-<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">cuda_output</span></code></p></td>
-<td><p>String</p></td>
-<td><p><code class="docutils literal notranslate"><span class="pre">None</span></code></p></td>
-<td><p>The preferred CUDA output format for kernels. Valid choices are <code class="docutils literal notranslate"><span class="pre">None</span></code>,
-<code class="docutils literal notranslate"><span class="pre">&quot;ptx&quot;</span></code>, and <code class="docutils literal notranslate"><span class="pre">&quot;cubin&quot;</span></code>. If <code class="docutils literal notranslate"><span class="pre">None</span></code>, a format will be determined
-automatically. The module-level setting takes precedence over the global
-setting.</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</section>
-<section id="kernel-settings">
-<h2>Kernel Settings<a class="headerlink" href="#kernel-settings" title="Link to this heading">#</a></h2>
-<p><code class="docutils literal notranslate"><span class="pre">enable_backward</span></code> is currently the only setting that can also be configured on a per-kernel level.
-Backward-pass compilation can be disabled by passing an argument into the <code class="docutils literal notranslate"><span class="pre">&#64;wp.kernel</span></code> decorator
-as in the following example:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span><span class="p">(</span><span class="n">enable_backward</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">scale_2</span><span class="p">(</span>
-    <span class="n">x</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">),</span>
-    <span class="n">y</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">),</span>
-<span class="p">):</span>
-    <span class="n">y</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">=</span> <span class="n">x</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">**</span> <span class="mf">2.0</span>
-</pre></div>
-</div>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="debugging.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Debugging</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="modules/interoperability.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Interoperability</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Runtime Settings</a><ul>
-<li><a class="reference internal" href="#global-settings">Global Settings</a><ul>
-<li><a class="reference internal" href="#basic-global-settings">Basic Global Settings</a></li>
-<li><a class="reference internal" href="#advanced-global-settings">Advanced Global Settings</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#module-settings">Module Settings</a></li>
-<li><a class="reference internal" href="#kernel-settings">Kernel Settings</a></li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/debugging.html b/docs/_build/html/debugging.html
deleted file mode 100644
index b2ed80f09..000000000
--- a/docs/_build/html/debugging.html
+++ /dev/null
@@ -1,406 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="Limitations" href="limitations.html" /><link rel="prev" title="Runtime Settings" href="configuration.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Debugging - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="debugging">
-<h1>Debugging<a class="headerlink" href="#debugging" title="Link to this heading">#</a></h1>
-<section id="printing-values">
-<h2>Printing Values<a class="headerlink" href="#printing-values" title="Link to this heading">#</a></h2>
-<p>Often one of the best debugging methods is to simply print values from kernels. Warp supports printing all built-in
-types using the <code class="docutils literal notranslate"><span class="pre">print()</span></code> function, e.g.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">)</span>
-
-<span class="nb">print</span><span class="p">(</span><span class="n">v</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>In addition, formatted C-style printing is available through the <code class="docutils literal notranslate"><span class="pre">wp.printf()</span></code> function, e.g.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">x</span> <span class="o">=</span> <span class="mf">1.0</span>
-<span class="n">i</span> <span class="o">=</span> <span class="mi">2</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">printf</span><span class="p">(</span><span class="s2">&quot;A float value </span><span class="si">%f</span><span class="s2">, an int value: </span><span class="si">%d</span><span class="s2">&quot;</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">i</span><span class="p">)</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Formatted printing is only available for scalar types (e.g.: <code class="docutils literal notranslate"><span class="pre">int</span></code> and <code class="docutils literal notranslate"><span class="pre">float</span></code>) not vector types.</p>
-</div>
-</section>
-<section id="printing-launches">
-<h2>Printing Launches<a class="headerlink" href="#printing-launches" title="Link to this heading">#</a></h2>
-<p>For complex applications it can be difficult to understand the order-of-operations that lead to a bug. To help diagnose
-these issues Warp supports a simple option to print out all launches and arguments to the console:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">print_launches</span> <span class="o">=</span> <span class="kc">True</span>
-</pre></div>
-</div>
-</section>
-<section id="step-through-debugging">
-<h2>Step-Through Debugging<a class="headerlink" href="#step-through-debugging" title="Link to this heading">#</a></h2>
-<p>It is possible to attach IDE debuggers such as Visual Studio to Warp processes to step through generated kernel code.
-Users should first compile the kernels in debug mode by setting:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">mode</span> <span class="o">=</span> <span class="s2">&quot;debug&quot;</span>
-</pre></div>
-</div>
-<p>This setting ensures that line numbers, and debug symbols are generated correctly. After launching the Python process,
-the debugger should be attached, and a breakpoint inserted into the generated code.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Generated kernel code is not a 1:1 correspondence with the original Python code, but individual operations can still be replayed and variables inspected.</p>
-</div>
-<p>Also see <a class="reference external" href="https://github.com/NVIDIA/warp/blob/main/warp/tests/test_debug.py">warp/tests/test_debug.py</a> for an example of how to debug Warp kernel code running on the CPU.</p>
-</section>
-<section id="generated-code">
-<h2>Generated Code<a class="headerlink" href="#generated-code" title="Link to this heading">#</a></h2>
-<p>The generated code for kernels is stored in a central cache location in the user’s home directory, the cache location
-is printed at startup when <code class="docutils literal notranslate"><span class="pre">wp.init()</span></code> is called, for example:</p>
-<div class="highlight-console notranslate"><div class="highlight"><pre><span></span><span class="go">Warp 0.8.1 initialized:</span>
-<span class="go">    CUDA Toolkit: 11.8, Driver: 11.8</span>
-<span class="go">    Devices:</span>
-<span class="go">    &quot;cpu&quot;    | AMD64 Family 25 Model 33 Stepping 0, AuthenticAMD</span>
-<span class="go">    &quot;cuda:0&quot; | NVIDIA GeForce RTX 3090 (sm_86)</span>
-<span class="go">    &quot;cuda:1&quot; | NVIDIA GeForce RTX 2080 Ti (sm_75)</span>
-<span class="go">    Kernel cache: C:\Users\LukasW\AppData\Local\NVIDIA Corporation\warp\Cache\0.8.1</span>
-</pre></div>
-</div>
-<p>The kernel cache has <code class="docutils literal notranslate"><span class="pre">gen</span></code> and <code class="docutils literal notranslate"><span class="pre">bin</span></code> folders that contain the generated C++/CUDA code and the compiled binaries
-respectively. Occasionally it can be useful to inspect the generated code for debugging / profiling.</p>
-</section>
-<section id="bounds-checking">
-<h2>Bounds Checking<a class="headerlink" href="#bounds-checking" title="Link to this heading">#</a></h2>
-<p>Warp will perform bounds checking in debug build configurations to ensure that all array accesses lie within the defined
-shape.</p>
-</section>
-<section id="cuda-verification">
-<h2>CUDA Verification<a class="headerlink" href="#cuda-verification" title="Link to this heading">#</a></h2>
-<p>It is possible to generate out-of-bounds memory access violations through poorly formed kernel code or inputs. In this
-case the CUDA runtime will detect the violation and put the CUDA context into an error state. Subsequent kernel launches
-may silently fail which can lead to hard to diagnose issues.</p>
-<p>If a CUDA error is suspected a simple verification method is to enable:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">config</span><span class="o">.</span><span class="n">verify_cuda</span> <span class="o">=</span> <span class="kc">True</span>
-</pre></div>
-</div>
-<p>This setting will check the CUDA context after every operation to ensure that it is still valid. If an error is
-encountered it will raise an exception that often helps to narrow down the problematic kernel.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Verifying CUDA state at each launch requires synchronizing CPU and GPU which has a significant overhead. Users should ensure this setting is only used during debugging.</p>
-</div>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="limitations.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Limitations</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="configuration.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Runtime Settings</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Debugging</a><ul>
-<li><a class="reference internal" href="#printing-values">Printing Values</a></li>
-<li><a class="reference internal" href="#printing-launches">Printing Launches</a></li>
-<li><a class="reference internal" href="#step-through-debugging">Step-Through Debugging</a></li>
-<li><a class="reference internal" href="#generated-code">Generated Code</a></li>
-<li><a class="reference internal" href="#bounds-checking">Bounds Checking</a></li>
-<li><a class="reference internal" href="#cuda-verification">CUDA Verification</a></li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/faq.html b/docs/_build/html/faq.html
deleted file mode 100644
index ed2625772..000000000
--- a/docs/_build/html/faq.html
+++ /dev/null
@@ -1,422 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="Runtime Reference" href="modules/runtime.html" /><link rel="prev" title="Limitations" href="limitations.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>FAQ - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="faq">
-<h1>FAQ<a class="headerlink" href="#faq" title="Link to this heading">#</a></h1>
-<section id="how-does-warp-relate-to-other-python-projects-for-gpu-programming-e-g-numba-taichi-cupy-pytorch-etc">
-<h2>How does Warp relate to other Python projects for GPU programming, e.g.: Numba, Taichi, cuPy, PyTorch, etc.?<a class="headerlink" href="#how-does-warp-relate-to-other-python-projects-for-gpu-programming-e-g-numba-taichi-cupy-pytorch-etc" title="Link to this heading">#</a></h2>
-<p>Warp is inspired by many of these projects, and is closely related to
-Numba and Taichi which both expose kernel programming to Python. These
-frameworks map to traditional GPU programming models, so many of the
-high-level concepts are similar, however there are some functionality
-and implementation differences.</p>
-<p>Compared to Numba, Warp supports a smaller subset of Python, but
-offering auto-differentiation of kernel programs, which is useful for
-machine learning. Compared to Taichi Warp uses C++/CUDA as an
-intermediate representation, which makes it convenient to implement and
-expose low-level routines. In addition, we are building in
-data structures to support geometry processing (meshes, sparse volumes,
-point clouds, USD data) as first-class citizens that are not exposed in
-other runtimes.</p>
-<p>Warp does not offer a full tensor-based programming model like PyTorch
-and JAX, but is designed to work well with these frameworks through data
-sharing mechanisms like <code class="docutils literal notranslate"><span class="pre">__cuda_array_interface__</span></code>. For computations
-that map well to tensors (e.g.: neural-network inference) it makes sense
-to use these existing tools. For problems with a lot of e.g.: sparsity,
-conditional logic, heterogenous workloads (like the ones we often find in
-simulation and graphics), then the kernel-based programming model like
-the one in Warp are often more convenient since users have control over
-individual threads.</p>
-</section>
-<section id="does-warp-support-all-of-the-python-language">
-<h2>Does Warp support all of the Python language?<a class="headerlink" href="#does-warp-support-all-of-the-python-language" title="Link to this heading">#</a></h2>
-<p>No, Warp supports a subset of Python that maps well to the GPU. Our goal
-is to not have any performance cliffs so that users can expect
-consistently good behavior from kernels that is close to native code.
-Examples of unsupported concepts that don’t map well to the GPU are
-dynamic types, list comprehensions, exceptions, garbage collection, etc.</p>
-</section>
-<section id="when-should-i-call-wp-synchronize">
-<h2>When should I call <code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code>?<a class="headerlink" href="#when-should-i-call-wp-synchronize" title="Link to this heading">#</a></h2>
-<p>One of the common sources of confusion for new users is when calls to
-<code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code> are necessary. The answer is “almost never”!
-Synchronization is quite expensive, and should generally be avoided
-unless necessary. Warp naturally takes care of synchronization between
-operations (e.g.: kernel launches, device memory copies).</p>
-<p>For example, the following requires no manual synchronization, as the
-conversion to NumPy will automatically synchronize:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="c1"># run some kernels</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel_1</span><span class="p">,</span> <span class="n">dim</span><span class="p">,</span> <span class="p">[</span><span class="n">array_x</span><span class="p">,</span> <span class="n">array_y</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel_2</span><span class="p">,</span> <span class="n">dim</span><span class="p">,</span> <span class="p">[</span><span class="n">array_y</span><span class="p">,</span> <span class="n">array_z</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="c1"># bring data back to host (and implicitly synchronize)</span>
-<span class="n">x</span> <span class="o">=</span> <span class="n">array_z</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>The <em>only</em> case where manual synchronization is needed is when copies
-are being performed back to CPU asynchronously, e.g.:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="c1"># copy data back to cpu from gpu, all copies will happen asynchronously to Python</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">cpu_array_1</span><span class="p">,</span> <span class="n">gpu_array_1</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">cpu_array_2</span><span class="p">,</span> <span class="n">gpu_array_2</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">cpu_array_3</span><span class="p">,</span> <span class="n">gpu_array_3</span><span class="p">)</span>
-
-<span class="c1"># ensure that the copies have finished</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">synchronize</span><span class="p">()</span>
-
-<span class="c1"># return a numpy wrapper around the cpu arrays, note there is no implicit synchronization here</span>
-<span class="n">a1</span> <span class="o">=</span> <span class="n">cpu_array_1</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-<span class="n">a2</span> <span class="o">=</span> <span class="n">cpu_array_2</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-<span class="n">a3</span> <span class="o">=</span> <span class="n">cpu_array_3</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-</pre></div>
-</div>
-</section>
-<section id="what-happens-when-you-differentiate-a-function-like-wp-abs-x">
-<h2>What happens when you differentiate a function like <code class="docutils literal notranslate"><span class="pre">wp.abs(x)</span></code>?<a class="headerlink" href="#what-happens-when-you-differentiate-a-function-like-wp-abs-x" title="Link to this heading">#</a></h2>
-<p>Non-smooth functions such as <code class="docutils literal notranslate"><span class="pre">y=|x|</span></code> do not have a single unique
-gradient at <code class="docutils literal notranslate"><span class="pre">x=0</span></code>, rather they have what is known as a
-<code class="docutils literal notranslate"><span class="pre">subgradient</span></code>, which is formally the convex hull of directional
-derivatives at that point. The way that Warp (and most
-auto-differentiation frameworks) handles these points is to pick an
-arbitrary gradient from this set, e.g.: for <code class="docutils literal notranslate"><span class="pre">wp.abs()</span></code>, it will
-arbitrarily choose the gradient to be 1.0 at the origin. You can find
-the implementation for these functions in <code class="docutils literal notranslate"><span class="pre">warp/native/builtin.h</span></code>.</p>
-<p>Most optimizers (particularly ones that exploit stochasticity), are not
-sensitive to the choice of which gradient to use from the subgradient,
-although there are exceptions.</p>
-</section>
-<section id="does-warp-support-multi-gpu-programming">
-<h2>Does Warp support multi-GPU programming?<a class="headerlink" href="#does-warp-support-multi-gpu-programming" title="Link to this heading">#</a></h2>
-<p>Yes! Since version <code class="docutils literal notranslate"><span class="pre">0.4.0</span></code> we support allocating, launching, and
-copying between multiple GPUs in a single process. We follow the naming
-conventions of PyTorch and use aliases such as <code class="docutils literal notranslate"><span class="pre">cuda:0</span></code>, <code class="docutils literal notranslate"><span class="pre">cuda:1</span></code>,
-<code class="docutils literal notranslate"><span class="pre">cpu</span></code> to identify individual devices.</p>
-</section>
-<section id="should-i-switch-to-warp-over-isaacgym-physx">
-<h2>Should I switch to Warp over IsaacGym / PhysX?<a class="headerlink" href="#should-i-switch-to-warp-over-isaacgym-physx" title="Link to this heading">#</a></h2>
-<p>Warp is not a replacement for IsaacGym, IsaacSim, or PhysX - while Warp
-does offer some physical simulation capabilities this is primarily aimed
-at developers who need differentiable physics, rather than a fully
-featured physics engine. Warp is also integrated with IsaacGym and is
-great for performing auxiliary tasks such as reward and observation
-computations for reinforcement learning.</p>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="modules/runtime.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Runtime Reference</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="limitations.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Limitations</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">FAQ</a><ul>
-<li><a class="reference internal" href="#how-does-warp-relate-to-other-python-projects-for-gpu-programming-e-g-numba-taichi-cupy-pytorch-etc">How does Warp relate to other Python projects for GPU programming, e.g.: Numba, Taichi, cuPy, PyTorch, etc.?</a></li>
-<li><a class="reference internal" href="#does-warp-support-all-of-the-python-language">Does Warp support all of the Python language?</a></li>
-<li><a class="reference internal" href="#when-should-i-call-wp-synchronize">When should I call <code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code>?</a></li>
-<li><a class="reference internal" href="#what-happens-when-you-differentiate-a-function-like-wp-abs-x">What happens when you differentiate a function like <code class="docutils literal notranslate"><span class="pre">wp.abs(x)</span></code>?</a></li>
-<li><a class="reference internal" href="#does-warp-support-multi-gpu-programming">Does Warp support multi-GPU programming?</a></li>
-<li><a class="reference internal" href="#should-i-switch-to-warp-over-isaacgym-physx">Should I switch to Warp over IsaacGym / PhysX?</a></li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/genindex.html b/docs/_build/html/genindex.html
deleted file mode 100644
index 51f68ea69..000000000
--- a/docs/_build/html/genindex.html
+++ /dev/null
@@ -1,1980 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><link rel="index" title="Index" href="#" /><link rel="search" title="Search" href="search.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 --><title>Index - Warp 1.0.0-beta.5</title>
-<link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon no-toc" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon no-toc" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          
-<section class="genindex-section">
-  <h1 id="index">Index</h1>
-  <div class="genindex-jumpbox"><a href="#A"><strong>A</strong></a> | <a href="#B"><strong>B</strong></a> | <a href="#C"><strong>C</strong></a> | <a href="#D"><strong>D</strong></a> | <a href="#E"><strong>E</strong></a> | <a href="#F"><strong>F</strong></a> | <a href="#G"><strong>G</strong></a> | <a href="#H"><strong>H</strong></a> | <a href="#I"><strong>I</strong></a> | <a href="#J"><strong>J</strong></a> | <a href="#L"><strong>L</strong></a> | <a href="#M"><strong>M</strong></a> | <a href="#N"><strong>N</strong></a> | <a href="#O"><strong>O</strong></a> | <a href="#P"><strong>P</strong></a> | <a href="#Q"><strong>Q</strong></a> | <a href="#R"><strong>R</strong></a> | <a href="#S"><strong>S</strong></a> | <a href="#T"><strong>T</strong></a> | <a href="#U"><strong>U</strong></a> | <a href="#V"><strong>V</strong></a> | <a href="#W"><strong>W</strong></a> | <a href="#X"><strong>X</strong></a> | <a href="#Z"><strong>Z</strong></a></div>
-</section>
-<section id="A" class="genindex-section">
-  <h2>A</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.abs">abs() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.acos">acos() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.add">add() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_body">add_body() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_builder">add_builder() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_cloth_grid">add_cloth_grid() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_cloth_mesh">add_cloth_mesh() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_edge">add_edge() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_edges">add_edges() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint">add_joint() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_ball">add_joint_ball() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_compound">add_joint_compound() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_d6">add_joint_d6() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_distance">add_joint_distance() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_fixed">add_joint_fixed() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_free">add_joint_free() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_prismatic">add_joint_prismatic() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_revolute">add_joint_revolute() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_joint_universal">add_joint_universal() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_muscle">add_muscle() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_particle">add_particle() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_box">add_shape_box() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_capsule">add_shape_capsule() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_cone">add_shape_cone() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_cylinder">add_shape_cylinder() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_mesh">add_shape_mesh() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_plane">add_shape_plane() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_sdf">add_shape_sdf() (warp.sim.ModelBuilder method)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_shape_sphere">add_shape_sphere() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_soft_grid">add_soft_grid() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_soft_mesh">add_soft_mesh() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_spring">add_spring() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_tetrahedron">add_tetrahedron() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_triangle">add_triangle() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.add_triangles">add_triangles() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.allocate">allocate() (warp.Volume class method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.allocate_by_tiles">allocate_by_tiles() (warp.Volume class method)</a>
-</li>
-        <li><a href="modules/functions.html#warp.argmax">argmax() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.argmin">argmin() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Array">Array (class in warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array">array (class in warp)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.cache.Temporary.array">(warp.fem.cache.Temporary property)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/runtime.html#warp.Volume.array">array() (warp.Volume method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.articulation_count">articulation_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.articulation_start">articulation_start (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.asin">asin() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.assign">assign() (warp.array method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.at_node">at_node() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.atan">atan() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.atan2">atan2() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.atomic_add">atomic_add() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.atomic_max">atomic_max() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.atomic_min">atomic_min() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.atomic_sub">atomic_sub() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.average">average() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.axis">axis (warp.sim.JointAxis attribute)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="B" class="genindex-section">
-  <h2>B</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/runtime.html#warp.Tape.backward">backward() (warp.Tape method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.BasisSpace">BasisSpace (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bit_and">bit_and() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bit_or">bit_or() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bit_xor">bit_xor() (in module warp)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.block_shape">block_shape (warp.sparse.BsrMatrix property)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.block_size">block_size (warp.sparse.BsrMatrix property)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_com">body_com (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_count">body_count (warp.sim.Model attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.State.body_count">(warp.sim.State attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.State.body_f">body_f (warp.sim.State attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_inertia">body_inertia (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_inv_inertia">body_inv_inertia (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_inv_mass">body_inv_mass (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_mass">body_mass (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_name">body_name (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_q">body_q (warp.sim.Model attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.State.body_q">(warp.sim.State attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_qd">body_qd (warp.sim.Model attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.State.body_qd">(warp.sim.State attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.Model.body_shapes">body_shapes (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bool">bool (class in warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.borrow_temporary">borrow_temporary() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.borrow_temporary_like">borrow_temporary_like() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Geometry.boundary_side_count">boundary_side_count() (warp.fem.Geometry method)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.GeometryPartition.boundary_side_count">(warp.fem.GeometryPartition method)</a>
-</li>
-        </ul></li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.BoundarySides">BoundarySides (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.breakpoint">breakpoint() (in module warp)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_assign">bsr_assign() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_axpy">bsr_axpy() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_axpy_work_arrays">bsr_axpy_work_arrays (class in warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_copy">bsr_copy() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_diag">bsr_diag() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_get_diag">bsr_get_diag() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_identity">bsr_identity() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_mm">bsr_mm() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_mm_work_arrays">bsr_mm_work_arrays (class in warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_mv">bsr_mv() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_scale">bsr_scale() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_set_diag">bsr_set_diag() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_set_from_triplets">bsr_set_from_triplets() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_set_identity">bsr_set_identity() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_set_transpose">bsr_set_transpose() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_set_zero">bsr_set_zero() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_transposed">bsr_transposed() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.bsr_zeros">bsr_zeros() (in module warp.sparse)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix">BsrMatrix (class in warp.sparse)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.HashGrid.build">build() (warp.HashGrid method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Bvh">Bvh (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bvh_query_aabb">bvh_query_aabb() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bvh_query_next">bvh_query_next() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.bvh_query_ray">bvh_query_ray() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="C" class="genindex-section">
-  <h2>C</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/runtime.html#warp.capture_begin">capture_begin() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.capture_end">capture_end() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.capture_launch">capture_launch() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.cbrt">cbrt() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.ceil">ceil() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Geometry.cell_count">cell_count() (warp.fem.Geometry method)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.GeometryPartition.cell_count">(warp.fem.GeometryPartition method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.Cells">Cells (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.clamp">clamp() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.clone">clone() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.CLOSEST">CLOSEST (warp.Volume attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.closest_point_edge_edge">closest_point_edge_edge() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.collapse_fixed_joints">collapse_fixed_joints() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.columns">columns (warp.sparse.BsrMatrix attribute)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.Mesh.com">com (warp.sim.Mesh attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.SDF.com">(warp.sim.SDF attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/runtime.html#warp.constant">constant (class in warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.contact_count">contact_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.contiguous">contiguous() (warp.array method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.copy">copy() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.cos">cos() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.cosh">cosh() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.count_contact_points">count_contact_points() (warp.sim.Model method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.cptr">cptr() (warp.array method)</a>
-</li>
-        <li><a href="modules/functions.html#warp.cross">cross() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.curl">curl() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.curlnoise">curlnoise() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.cw_div">cw_div() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.cw_mul">cw_mul() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="D" class="genindex-section">
-  <h2>D</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.D">D() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.ddot">ddot() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.deformation_gradient">deformation_gradient() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FunctionSpace.degree">degree (warp.fem.FunctionSpace property)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.degree">degree() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.degrees">degrees() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.cache.Temporary.detach">detach() (warp.fem.cache.Temporary method)</a>
-</li>
-        <li><a href="modules/functions.html#warp.determinant">determinant() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Bvh.device">device (warp.Bvh attribute)</a>
-
-        <ul>
-          <li><a href="modules/runtime.html#warp.HashGrid.device">(warp.HashGrid attribute)</a>
-</li>
-          <li><a href="modules/runtime.html#warp.Mesh.device">(warp.Mesh attribute)</a>
-</li>
-          <li><a href="modules/sim.html#warp.sim.Model.device">(warp.sim.Model attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/interoperability.html#warp.dlpack.device_from_dlpack">device_from_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.jax.device_from_jax">device_from_jax() (in module warp.jax)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.device_from_torch">device_from_torch() (in module warp.torch)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.device_to_dlpack">device_to_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.jax.device_to_jax">device_to_jax() (in module warp.jax)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.device_to_torch">device_to_torch() (in module warp.torch)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.diag">diag() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FunctionSpace.dimension">dimension (warp.fem.FunctionSpace property)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.GeometryDomain.dimension">(warp.fem.GeometryDomain property)</a>
-</li>
-          <li><a href="modules/fem.html#warp.fem.SpaceTopology.dimension">(warp.fem.SpaceTopology attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.DiscreteField">DiscreteField (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.div">div() (in module warp)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.div">(in module warp.fem)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.div_outer">div_outer() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.DiscreteField.dof_values">dof_values (warp.fem.DiscreteField property)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.DofMapper">DofMapper (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Domain">Domain (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Quadrature.domain">domain (warp.fem.Quadrature property)</a>
-</li>
-        <li><a href="modules/functions.html#warp.dot">dot() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FunctionSpace.dtype">dtype (warp.fem.FunctionSpace attribute)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.dtype_from_dlpack">dtype_from_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.dtype_from_torch">dtype_from_torch() (in module warp.torch)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.dtype_is_compatible">dtype_is_compatible() (in module warp.dlpack)</a>
-
-        <ul>
-          <li><a href="modules/interoperability.html#warp.torch.dtype_is_compatible">(in module warp.torch)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/interoperability.html#warp.dlpack.dtype_to_dlpack">dtype_to_dlpack() (in module warp.dlpack)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="E" class="genindex-section">
-  <h2>E</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.Model.edge_bending_properties">edge_bending_properties (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.edge_count">edge_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.edge_indices">edge_indices (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.edge_rest_angle">edge_rest_angle (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.eig3">eig3() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.ScopedTimer.elapsed">elapsed (warp.ScopedTimer attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.GeometryDomain.element_count">element_count() (warp.fem.GeometryDomain method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.GeometryDomain.element_kind">element_kind (warp.fem.GeometryDomain property)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpaceTopology.element_node_indices">element_node_indices() (warp.fem.SpaceTopology method)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.ElementBasis">ElementBasis (class in warp.fem)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.empty">empty() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.empty_like">empty_like() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Polynomial.EQUISPACED_CLOSED">EQUISPACED_CLOSED (warp.fem.Polynomial attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Polynomial.EQUISPACED_OPEN">EQUISPACED_OPEN (warp.fem.Polynomial attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.eval_fk">eval_fk() (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.eval_ik">eval_ik() (in module warp.sim)</a>
-</li>
-        <li><a href="modules/functions.html#warp.exp">exp() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.expect_near">expect_near() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.ExplicitGeometryPartition">ExplicitGeometryPartition (class in warp.fem)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="F" class="genindex-section">
-  <h2>F</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.Field">Field (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FieldLike">FieldLike (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.field.FieldRestriction">FieldRestriction (class in warp.fem.field)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.fill_">fill_() (warp.array method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Mesh.finalize">finalize() (warp.sim.Mesh method)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.ModelBuilder.finalize">(warp.sim.ModelBuilder method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/runtime.html#warp.array.flatten">flatten() (warp.array method)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.Model.flatten">(warp.sim.Model method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/functions.html#warp.Float">Float (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.float16">float16 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.float32">float32 (class in warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.float64">float64 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.floor">floor() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.floordiv">floordiv() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.frac">frac() (in module warp)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.from_dlpack">from_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.jax.from_jax">from_jax() (in module warp.jax)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.from_torch">from_torch() (in module warp.torch)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.GeometryPartition.frontier_side_count">frontier_side_count() (warp.fem.GeometryPartition method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FrontierSides">FrontierSides (class in warp.fem)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.full">full() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.full_like">full_like() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FunctionSpace">FunctionSpace (class in warp.fem)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="G" class="genindex-section">
-  <h2>G</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.Polynomial.GAUSS_LEGENDRE">GAUSS_LEGENDRE (warp.fem.Polynomial attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Geometry">Geometry (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.BasisSpace.geometry">geometry (warp.fem.BasisSpace property)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.FunctionSpace.geometry">(warp.fem.FunctionSpace property)</a>
-</li>
-          <li><a href="modules/fem.html#warp.fem.SpaceTopology.geometry">(warp.fem.SpaceTopology property)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.GeometryDomain">GeometryDomain (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.GeometryDomain.ElementKind">GeometryDomain.ElementKind (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.GeometryPartition">GeometryPartition (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.get_diag">get_diag() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.get_tiles">get_tiles() (warp.Volume method)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/runtime.html#warp.Volume.get_voxel_size">get_voxel_size() (warp.Volume method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.grad">grad (warp.array property)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.grad">grad() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.grad_average">grad_average() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.grad_jump">grad_jump() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.grad_outer">grad_outer() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.gravity">gravity (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Grid2D">Grid2D (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Grid3D">Grid3D (class in warp.fem)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.ground">ground (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.ground_plane">ground_plane (warp.sim.Model attribute)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="H" class="genindex-section">
-  <h2>H</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.hash_grid_point_id">hash_grid_point_id() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.hash_grid_query">hash_grid_query() (in module warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.hash_grid_query_next">hash_grid_query_next() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.HashGrid">HashGrid (class in warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Hexmesh">Hexmesh (class in warp.fem)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="I" class="genindex-section">
-  <h2>I</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.Mesh.I">I (warp.sim.Mesh attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.SDF.I">(warp.sim.SDF attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/runtime.html#warp.Bvh.id">id (warp.Bvh attribute)</a>
-
-        <ul>
-          <li><a href="modules/runtime.html#warp.HashGrid.id">(warp.HashGrid attribute)</a>
-</li>
-          <li><a href="modules/runtime.html#warp.Mesh.id">(warp.Mesh attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/functions.html#warp.identity">identity() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Mesh.indices">indices (warp.sim.Mesh attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.inner">inner() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Int">Int (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.int16">int16 (class in warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.int32">int32 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.int64">int64 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.int8">int8 (class in warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.integrand">integrand() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.integrate">integrate() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpacePartition.interior_node_count">interior_node_count() (warp.fem.SpacePartition method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.interpolate">interpolate() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.intersect_tri_tri">intersect_tri_tri() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.inverse">inverse() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.invert">invert() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="J" class="genindex-section">
-  <h2>J</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_act">joint_act (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_angular_compliance">joint_angular_compliance (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_armature">joint_armature (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_attach_kd">joint_attach_kd (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_attach_ke">joint_attach_ke (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_axis">joint_axis (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_axis_dim">joint_axis_dim (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_axis_mode">joint_axis_mode (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_axis_start">joint_axis_start (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_BALL">JOINT_BALL (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_child">joint_child (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_COMPOUND">JOINT_COMPOUND (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_coord_count">joint_coord_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_count">joint_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_D6">JOINT_D6 (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_DISTANCE">JOINT_DISTANCE (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_dof_count">joint_dof_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_enabled">joint_enabled (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_FIXED">JOINT_FIXED (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_FREE">JOINT_FREE (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_limit_kd">joint_limit_kd (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_limit_ke">joint_limit_ke (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_limit_lower">joint_limit_lower (warp.sim.Model attribute)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_limit_upper">joint_limit_upper (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_linear_compliance">joint_linear_compliance (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_MODE_LIMIT">JOINT_MODE_LIMIT (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_MODE_TARGET_POSITION">JOINT_MODE_TARGET_POSITION (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_MODE_TARGET_VELOCITY">JOINT_MODE_TARGET_VELOCITY (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_name">joint_name (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_parent">joint_parent (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_PRISMATIC">JOINT_PRISMATIC (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_q">joint_q (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_q_start">joint_q_start (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_qd">joint_qd (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_qd_start">joint_qd_start (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_REVOLUTE">JOINT_REVOLUTE (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_target">joint_target (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_target_kd">joint_target_kd (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_target_ke">joint_target_ke (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_twist_lower">joint_twist_lower (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_twist_upper">joint_twist_upper (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_type">joint_type (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JOINT_UNIVERSAL">JOINT_UNIVERSAL (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_X_c">joint_X_c (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.joint_X_p">joint_X_p (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis">JointAxis (class in warp.sim)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.jump">jump() (in module warp.fem)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="L" class="genindex-section">
-  <h2>L</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.ElementBasis.LAGRANGE">LAGRANGE (warp.fem.ElementBasis attribute)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.launch">launch() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.length">length() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.length_sq">length_sq() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.lerp">lerp() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.limit_kd">limit_kd (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.limit_ke">limit_ke (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.limit_lower">limit_lower (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.limit_upper">limit_upper (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.LINEAR">LINEAR (warp.Volume attribute)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.LinearGeometryPartition">LinearGeometryPartition (class in warp.fem)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.list">list() (warp.array method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.load_from_numpy">load_from_numpy() (warp.Volume class method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume.load_from_nvdb">load_from_nvdb() (warp.Volume class method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Polynomial.LOBATTO_GAUSS_LEGENDRE">LOBATTO_GAUSS_LEGENDRE (warp.fem.Polynomial attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.log">log() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.log10">log10() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.log2">log2() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.lookup">lookup() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.lower_bound">lower_bound() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.lshift">lshift() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="M" class="genindex-section">
-  <h2>M</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.make_collocated_function_space">make_collocated_function_space() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.DiscreteField.make_deformed_geometry">make_deformed_geometry() (warp.fem.DiscreteField method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.FunctionSpace.make_field">make_field() (warp.fem.FunctionSpace method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_polynomial_basis_space">make_polynomial_basis_space() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_polynomial_space">make_polynomial_space() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_restriction">make_restriction() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_space_partition">make_space_partition() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_space_restriction">make_space_restriction() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_test">make_test() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.make_trial">make_trial() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Mesh.mass">mass (warp.sim.Mesh attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.SDF.mass">(warp.sim.SDF attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/functions.html#warp.mat22d">mat22d (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat22f">mat22f (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat22h">mat22h (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat33d">mat33d (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat33f">mat33f (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat33h">mat33h (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat44d">mat44d (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat44f">mat44f (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mat44h">mat44h (class in warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.matmul">matmul() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Matrix">Matrix (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.matrix">matrix() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.max">max() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.measure">measure() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.measure_ratio">measure_ratio() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Mesh">Mesh (class in warp)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.Mesh">(class in warp.sim)</a>
-</li>
-        </ul></li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.mesh_eval_face_normal">mesh_eval_face_normal() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_eval_position">mesh_eval_position() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_eval_velocity">mesh_eval_velocity() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_get">mesh_get() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_get_index">mesh_get_index() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_get_point">mesh_get_point() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_get_velocity">mesh_get_velocity() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_aabb">mesh_query_aabb() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_aabb_next">mesh_query_aabb_next() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_furthest_point_no_sign">mesh_query_furthest_point_no_sign() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_point">mesh_query_point() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_point_no_sign">mesh_query_point_no_sign() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_point_sign_normal">mesh_query_point_sign_normal() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_point_sign_winding_number">mesh_query_point_sign_winding_number() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mesh_query_ray">mesh_query_ray() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.min">min() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mlp">mlp() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.mod">mod() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.mode">mode (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model">Model (class in warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder">ModelBuilder (class in warp.sim)</a>
-</li>
-        <li>
-    module
-
-        <ul>
-          <li><a href="modules/interoperability.html#module-warp.dlpack">warp.dlpack</a>
-</li>
-          <li><a href="modules/interoperability.html#module-warp.jax">warp.jax</a>
-</li>
-          <li><a href="modules/sparse.html#module-warp.sparse">warp.sparse</a>
-</li>
-          <li><a href="modules/interoperability.html#module-warp.torch">warp.torch</a>
-</li>
-        </ul></li>
-        <li><a href="modules/functions.html#warp.mul">mul() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.muscle_count">muscle_count (warp.sim.Model attribute)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="N" class="genindex-section">
-  <h2>N</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.ncol">ncol (warp.sparse.BsrMatrix attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.neg">neg() (in module warp)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.nnz">nnz (warp.sparse.BsrMatrix attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.NodalQuadrature">NodalQuadrature (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpacePartition.node_count">node_count() (warp.fem.SpacePartition method)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.SpaceTopology.node_count">(warp.fem.SpaceTopology method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.BasisSpace.node_positions">node_positions() (warp.fem.BasisSpace method)</a>
-</li>
-        <li><a href="modules/functions.html#warp.noise">noise() (in module warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.ElementBasis.NONCONFORMING_POLYNOMIAL">NONCONFORMING_POLYNOMIAL (warp.fem.ElementBasis attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.nonzero">nonzero() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.normal">normal() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.normalize">normalize() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.normalize_dirichlet_projector">normalize_dirichlet_projector() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.nrow">nrow (warp.sparse.BsrMatrix attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.num_envs">num_envs (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.numpy">numpy() (warp.array method)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="O" class="genindex-section">
-  <h2>O</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.offsets">offsets (warp.sparse.BsrMatrix attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.operator.operator">operator() (in module warp.fem.operator)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.outer">outer() (in module warp)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.outer">(in module warp.fem)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.SpacePartition.owned_node_count">owned_node_count() (warp.fem.SpacePartition method)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="P" class="genindex-section">
-  <h2>P</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.parse_mjcf">parse_mjcf() (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.parse_urdf">parse_urdf() (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.parse_usd">parse_usd() (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_adhesion">particle_adhesion (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_cohesion">particle_cohesion (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_count">particle_count (warp.sim.Model attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.State.particle_count">(warp.sim.State attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.State.particle_f">particle_f (warp.sim.State attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_flags">particle_flags (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_grid">particle_grid (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_inv_mass">particle_inv_mass (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_kd">particle_kd (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_ke">particle_ke (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_kf">particle_kf (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_mass">particle_mass (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_max_radius">particle_max_radius (warp.sim.Model attribute)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_max_velocity">particle_max_velocity (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_mu">particle_mu (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_q">particle_q (warp.sim.Model attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.State.particle_q">(warp.sim.State attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_qd">particle_qd (warp.sim.Model attribute)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.State.particle_qd">(warp.sim.State attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.Model.particle_radius">particle_radius (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.PicQuadrature">PicQuadrature (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.pnoise">pnoise() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.poisson">poisson() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Polynomial">Polynomial (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.pos">pos() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.position">position() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.pow">pow() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.print">print() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.printf">printf() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.project_linear_system">project_linear_system() (in module warp.fem)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="Q" class="genindex-section">
-  <h2>Q</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.qr3">qr3() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Quadmesh2D">Quadmesh2D (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Quadrature">Quadrature (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_from_axis_angle">quat_from_axis_angle() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_from_matrix">quat_from_matrix() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_identity">quat_identity() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_inverse">quat_inverse() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_rotate">quat_rotate() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_rotate_inv">quat_rotate_inv() (in module warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.quat_rpy">quat_rpy() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_slerp">quat_slerp() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_to_axis_angle">quat_to_axis_angle() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quat_to_matrix">quat_to_matrix() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quatd">quatd (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Quaternion">Quaternion (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quaternion">quaternion() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quatf">quatf (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.quath">quath (class in warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="R" class="genindex-section">
-  <h2>R</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.radians">radians() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.rand_init">rand_init() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.randf">randf() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.randi">randi() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.randn">randn() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Tape.record_func">record_func() (warp.Tape method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Bvh.refit">refit() (warp.Bvh method)</a>
-
-        <ul>
-          <li><a href="modules/runtime.html#warp.Mesh.refit">(warp.Mesh method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.RegularQuadrature">RegularQuadrature (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.cache.Temporary.release">release() (warp.fem.cache.Temporary method)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/runtime.html#warp.array.requires_grad">requires_grad (warp.array property)</a>
-
-        <ul>
-          <li><a href="modules/sim.html#warp.sim.Model.requires_grad">(warp.sim.Model attribute)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/runtime.html#warp.Tape.reset">reset() (warp.Tape method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.reshape">reshape() (warp.array method)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.resolve_usd_from_url">resolve_usd_from_url() (in module warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.rigid_contact_margin">rigid_contact_margin (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.rigid_contact_rolling_friction">rigid_contact_rolling_friction (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.rigid_contact_torsional_friction">rigid_contact_torsional_friction (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.rint">rint() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.round">round() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.rshift">rshift() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="S" class="genindex-section">
-  <h2>S</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.Sample">Sample (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_cdf">sample_cdf() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_triangle">sample_triangle() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_cube">sample_unit_cube() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_disk">sample_unit_disk() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_hemisphere">sample_unit_hemisphere() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_hemisphere_surface">sample_unit_hemisphere_surface() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_ring">sample_unit_ring() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_sphere">sample_unit_sphere() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_sphere_surface">sample_unit_sphere_surface() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sample_unit_square">sample_unit_square() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Scalar">Scalar (class in warp)</a>
-</li>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.scalar_type">scalar_type (warp.sparse.BsrMatrix property)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.ScopedTimer">ScopedTimer (class in warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.SDF">SDF (class in warp.sim)</a>
-</li>
-        <li><a href="modules/functions.html#warp.select">select() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.SemiImplicitIntegrator">SemiImplicitIntegrator (class in warp.sim)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.ElementBasis.SERENDIPITY">SERENDIPITY (warp.fem.ElementBasis attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.set_default_temporary_store">set_default_temporary_store() (in module warp.fem)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.ModelBuilder.set_ground_plane">set_ground_plane() (warp.sim.ModelBuilder method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.BasisSpace.shape">shape (warp.fem.BasisSpace property)</a>
-
-        <ul>
-          <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.shape">(warp.sparse.BsrMatrix property)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_body">shape_body (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_collision_filter_pairs">shape_collision_filter_pairs (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_collision_group">shape_collision_group (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_collision_group_map">shape_collision_group_map (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_collision_radius">shape_collision_radius (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_contact_pairs">shape_contact_pairs (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_count">shape_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_geo_src">shape_geo_src (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_ground_collision">shape_ground_collision (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_ground_contact_pairs">shape_ground_contact_pairs (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_materials">shape_materials (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_shape_geo">shape_shape_geo (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.shape_to_dlpack">shape_to_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.shape_transform">shape_transform (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.space.shape.ShapeFunction">ShapeFunction (class in warp.fem.space.shape)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Geometry.side_count">side_count() (warp.fem.Geometry method)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.GeometryPartition.side_count">(warp.fem.GeometryPartition method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.Sides">Sides (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sign">sign() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sin">sin() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sinh">sinh() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.skew">skew() (in module warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/fem.html#warp.fem.SkewSymmetricTensorMapper">SkewSymmetricTensorMapper (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.smoothstep">smoothstep() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_kd">soft_contact_kd (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_ke">soft_contact_ke (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_kf">soft_contact_kf (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_margin">soft_contact_margin (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_max">soft_contact_max (warp.sim.Model property)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_mu">soft_contact_mu (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.soft_contact_restitution">soft_contact_restitution (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpacePartition.space_node_indices">space_node_indices() (warp.fem.SpacePartition method)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.field.SpaceField">SpaceField (class in warp.fem.field)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpacePartition">SpacePartition (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpaceRestriction">SpaceRestriction (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SpaceTopology">SpaceTopology (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_adjoint">spatial_adjoint() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_bottom">spatial_bottom() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_cross">spatial_cross() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_cross_dual">spatial_cross_dual() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_dot">spatial_dot() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_jacobian">spatial_jacobian() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_mass">spatial_mass() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_matrixd">spatial_matrixd (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_matrixf">spatial_matrixf (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_matrixh">spatial_matrixh (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_top">spatial_top() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_vectord">spatial_vectord (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_vectorf">spatial_vectorf (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.spatial_vectorh">spatial_vectorh (class in warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.spring_control">spring_control (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.spring_count">spring_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.spring_damping">spring_damping (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.spring_indices">spring_indices (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.spring_rest_length">spring_rest_length (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.spring_stiffness">spring_stiffness (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sqrt">sqrt() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.State">State (class in warp.sim)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.state">state() (warp.sim.Model method)</a>
-</li>
-        <li><a href="modules/functions.html#warp.step">step() (in module warp)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.stream_from_torch">stream_from_torch() (in module warp.torch)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.stream_to_torch">stream_to_torch() (in module warp.torch)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.strides_to_dlpack">strides_to_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/functions.html#warp.sub">sub() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.svd3">svd3() (in module warp)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.SymmetricTensorMapper">SymmetricTensorMapper (class in warp.fem)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="T" class="genindex-section">
-  <h2>T</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.tan">tan() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.tanh">tanh() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Tape">Tape (class in warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.target">target (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.target_kd">target_kd (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.JointAxis.target_ke">target_ke (warp.sim.JointAxis attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.cache.Temporary">Temporary (class in warp.fem.cache)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.TemporaryStore">TemporaryStore (class in warp.fem)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.field.TestField">TestField (class in warp.fem.field)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tet_activations">tet_activations (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tet_count">tet_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tet_indices">tet_indices (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tet_materials">tet_materials (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tet_poses">tet_poses (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Tetmesh">Tetmesh (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.tid">tid() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.to">to() (warp.array method)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.dlpack.to_dlpack">to_dlpack() (in module warp.dlpack)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.jax.to_jax">to_jax() (in module warp.jax)</a>
-</li>
-        <li><a href="modules/interoperability.html#warp.torch.to_torch">to_torch() (in module warp.torch)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.BasisSpace.topology">topology (warp.fem.BasisSpace property)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.FunctionSpace.topology">(warp.fem.FunctionSpace property)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/fem.html#warp.fem.Quadrature.total_point_count">total_point_count() (warp.fem.Quadrature method)</a>
-</li>
-        <li><a href="modules/functions.html#warp.trace">trace() (in module warp)</a>
-
-        <ul>
-          <li><a href="modules/fem.html#warp.fem.DiscreteField.trace">(warp.fem.DiscreteField method)</a>
-</li>
-          <li><a href="modules/fem.html#warp.fem.FunctionSpace.trace">(warp.fem.FunctionSpace method)</a>
-</li>
-          <li><a href="modules/fem.html#warp.fem.SpaceTopology.trace">(warp.fem.SpaceTopology method)</a>
-</li>
-        </ul></li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.transform_get_rotation">transform_get_rotation() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transform_get_translation">transform_get_translation() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transform_identity">transform_identity() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transform_inverse">transform_inverse() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transform_multiply">transform_multiply() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transform_point">transform_point() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transform_vector">transform_vector() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Transformation">Transformation (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transformation">transformation() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transformd">transformd (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transformf">transformf (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transformh">transformh (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.transpose">transpose() (in module warp)</a>
-
-        <ul>
-          <li><a href="modules/runtime.html#warp.array.transpose">(warp.array method)</a>
-</li>
-        </ul></li>
-        <li><a href="modules/sim.html#warp.sim.Model.tri_activations">tri_activations (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tri_count">tri_count (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tri_indices">tri_indices (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tri_materials">tri_materials (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.tri_poses">tri_poses (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.field.TrialField">TrialField (class in warp.fem.field)</a>
-</li>
-        <li><a href="modules/fem.html#warp.fem.Trimesh2D">Trimesh2D (class in warp.fem)</a>
-</li>
-        <li><a href="modules/functions.html#warp.trunc">trunc() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="U" class="genindex-section">
-  <h2>U</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.uint16">uint16 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.uint32">uint32 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.uint64">uint64 (class in warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.uint8">uint8 (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.unot">unot() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.up_axis">up_axis (warp.sim.Model attribute)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Model.up_vector">up_vector (warp.sim.Model attribute)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="V" class="genindex-section">
-  <h2>V</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sparse.html#warp.sparse.BsrMatrix.values">values (warp.sparse.BsrMatrix attribute)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.vars">vars (warp.array property)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2b">vec2b (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2d">vec2d (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2f">vec2f (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2h">vec2h (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2i">vec2i (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2l">vec2l (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2s">vec2s (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2ub">vec2ub (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2ui">vec2ui (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2ul">vec2ul (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec2us">vec2us (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3b">vec3b (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3d">vec3d (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3f">vec3f (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3h">vec3h (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3i">vec3i (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3l">vec3l (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3s">vec3s (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3ub">vec3ub (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3ui">vec3ui (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3ul">vec3ul (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec3us">vec3us (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4b">vec4b (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4d">vec4d (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4f">vec4f (class in warp)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/functions.html#warp.vec4h">vec4h (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4i">vec4i (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4l">vec4l (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4s">vec4s (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4ub">vec4ub (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4ui">vec4ui (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4ul">vec4ul (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vec4us">vec4us (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.Vector">Vector (class in warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.vector">vector() (in module warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.Mesh.vertices">vertices (warp.sim.Mesh attribute)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.view">view() (warp.array method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.Volume">Volume (class in warp)</a>
-</li>
-        <li><a href="modules/sim.html#warp.sim.SDF.volume">volume (warp.sim.SDF attribute)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_index_to_world">volume_index_to_world() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_index_to_world_dir">volume_index_to_world_dir() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_lookup_f">volume_lookup_f() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_lookup_i">volume_lookup_i() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_lookup_v">volume_lookup_v() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_sample_f">volume_sample_f() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_sample_grad_f">volume_sample_grad_f() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_sample_i">volume_sample_i() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_sample_v">volume_sample_v() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_store_f">volume_store_f() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_store_i">volume_store_i() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_store_v">volume_store_v() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_world_to_index">volume_world_to_index() (in module warp)</a>
-</li>
-        <li><a href="modules/functions.html#warp.volume_world_to_index_dir">volume_world_to_index_dir() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="W" class="genindex-section">
-  <h2>W</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li>
-    warp.dlpack
-
-        <ul>
-          <li><a href="modules/interoperability.html#module-warp.dlpack">module</a>
-</li>
-        </ul></li>
-        <li>
-    warp.jax
-
-        <ul>
-          <li><a href="modules/interoperability.html#module-warp.jax">module</a>
-</li>
-        </ul></li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li>
-    warp.sparse
-
-        <ul>
-          <li><a href="modules/sparse.html#module-warp.sparse">module</a>
-</li>
-        </ul></li>
-        <li>
-    warp.torch
-
-        <ul>
-          <li><a href="modules/interoperability.html#module-warp.torch">module</a>
-</li>
-        </ul></li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="X" class="genindex-section">
-  <h2>X</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/sim.html#warp.sim.XPBDIntegrator">XPBDIntegrator (class in warp.sim)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-<section id="Z" class="genindex-section">
-  <h2>Z</h2>
-  <table style="width: 100%" class="indextable genindextable"><tr>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/runtime.html#warp.Tape.zero">zero() (warp.Tape method)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.array.zero_">zero_() (warp.array method)</a>
-</li>
-    </ul></td>
-    <td style="width: 33%; vertical-align: top;"><ul>
-        <li><a href="modules/runtime.html#warp.zeros">zeros() (in module warp)</a>
-</li>
-        <li><a href="modules/runtime.html#warp.zeros_like">zeros_like() (in module warp)</a>
-</li>
-    </ul></td>
-  </tr></table>
-</section>
-
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          
-          
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer no-toc">
-      
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/index.html b/docs/_build/html/index.html
deleted file mode 100644
index 67e67f171..000000000
--- a/docs/_build/html/index.html
+++ /dev/null
@@ -1,566 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="Installation" href="installation.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="#"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon no-toc" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="#">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon no-toc" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="nvidia-warp-documentation">
-<h1>NVIDIA Warp Documentation<a class="headerlink" href="#nvidia-warp-documentation" title="Link to this heading">#</a></h1>
-<p>Warp is a Python framework for writing high-performance simulation and graphics code. Warp takes
-regular Python functions and JIT compiles them to efficient kernel code that can run on the CPU or GPU.</p>
-<p>Warp is designed for spatial computing and comes with a rich set of primitives that make it easy to write
-programs for physics simulation, perception, robotics, and geometry processing. In addition, Warp kernels
-are differentiable and can be used as part of machine-learning pipelines with frameworks such as PyTorch and JAX.</p>
-<p>Below are some examples of simulations implemented using Warp:</p>
-<img alt="_images/header.png" src="_images/header.png" />
-</section>
-<section id="quickstart">
-<h1>Quickstart<a class="headerlink" href="#quickstart" title="Link to this heading">#</a></h1>
-<p>The easiest way is to install Warp is from PyPi:</p>
-<div class="highlight-sh notranslate"><div class="highlight"><pre><span></span>$<span class="w"> </span>pip<span class="w"> </span>install<span class="w"> </span>warp-lang
-</pre></div>
-</div>
-<p>Pre-built binary packages for Windows, Linux and macOS are also available on the <a class="reference external" href="https://github.com/NVIDIA/warp/releases">Releases</a> page. To install in your local Python environment extract the archive and run the following command from the root directory:</p>
-<div class="highlight-sh notranslate"><div class="highlight"><pre><span></span>$<span class="w"> </span>pip<span class="w"> </span>install<span class="w"> </span>.
-</pre></div>
-</div>
-<section id="basic-example">
-<h2>Basic example<a class="headerlink" href="#basic-example" title="Link to this heading">#</a></h2>
-<p>An example first program that computes the lengths of random 3D vectors is given below:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-
-<span class="n">num_points</span> <span class="o">=</span> <span class="mi">1024</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">length</span><span class="p">(</span><span class="n">points</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-           <span class="n">lengths</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-
-    <span class="c1"># thread index</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="c1"># compute distance of each point from origin</span>
-    <span class="n">lengths</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">length</span><span class="p">(</span><span class="n">points</span><span class="p">[</span><span class="n">tid</span><span class="p">])</span>
-
-
-<span class="c1"># allocate an array of 3d points</span>
-<span class="n">points</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">num_points</span><span class="p">,</span> <span class="mi">3</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-<span class="n">lengths</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">num_points</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)</span>
-
-<span class="c1"># launch kernel</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">length</span><span class="p">,</span>
-          <span class="n">dim</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">points</span><span class="p">),</span>
-          <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">points</span><span class="p">,</span> <span class="n">lengths</span><span class="p">])</span>
-
-<span class="nb">print</span><span class="p">(</span><span class="n">lengths</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-<section id="additional-examples">
-<h2>Additional examples<a class="headerlink" href="#additional-examples" title="Link to this heading">#</a></h2>
-<p>The <a class="reference external" href="https://github.com/NVIDIA/warp/tree/main/examples">examples</a> directory in
-the Github repository contains a number of scripts that show how to
-implement different simulation methods using the Warp API. Most examples
-will generate USD files containing time-sampled animations in the
-<code class="docutils literal notranslate"><span class="pre">examples/outputs</span></code> directory. Before running examples users should
-ensure that the <code class="docutils literal notranslate"><span class="pre">usd-core</span></code> package is installed using:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">pip</span> <span class="n">install</span> <span class="n">usd</span><span class="o">-</span><span class="n">core</span>
-</pre></div>
-</div>
-<p>USD files can be viewed or rendered inside NVIDIA
-<a class="reference external" href="https://developer.nvidia.com/omniverse">Omniverse</a>,
-Pixar’s UsdView, and Blender. Note that Preview in macOS is not
-recommended as it has limited support for time-sampled animations.</p>
-<p>Built-in unit tests can be run from the command-line as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">python</span> <span class="o">-</span><span class="n">m</span> <span class="n">warp</span><span class="o">.</span><span class="n">tests</span>
-</pre></div>
-</div>
-</section>
-<section id="omniverse">
-<h2>Omniverse<a class="headerlink" href="#omniverse" title="Link to this heading">#</a></h2>
-<p>A Warp Omniverse extension is available in the extension registry inside
-Omniverse Kit or USD Composer.</p>
-<p>Enabling the extension will automatically install and initialize the
-Warp Python module inside the Kit Python environment. Please see the
-<a class="reference external" href="https://docs.omniverse.nvidia.com/extensions/latest/ext_warp.html">Omniverse Warp Documentation</a>
-for more details on how to use Warp in Omniverse.</p>
-</section>
-<section id="learn-more">
-<h2>Learn More<a class="headerlink" href="#learn-more" title="Link to this heading">#</a></h2>
-<p>Please see the following resources for additional background on Warp:</p>
-<ul class="simple">
-<li><p><a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring22-s41599">GTC 2022
-Presentation</a></p></li>
-<li><p><a class="reference external" href="https://www.nvidia.com/en-us/on-demand/session/gtcspring21-s31838">GTC 2021
-Presentation</a></p></li>
-<li><p><a class="reference external" href="https://dl.acm.org/doi/abs/10.1145/3476117.3483433">SIGGRAPH Asia 2021 Differentiable Simulation
-Course</a></p></li>
-</ul>
-<p>The underlying technology in Warp has been used in a number of research
-projects at NVIDIA including the following publications:</p>
-<ul class="simple">
-<li><p>Accelerated Policy Learning with Parallel Differentiable Simulation -
-Xu, J., Makoviychuk, V., Narang, Y., Ramos, F., Matusik, W., Garg,
-A., &amp; Macklin, M.
-<a class="reference external" href="https://short-horizon-actor-critic.github.io">(2022)</a></p></li>
-<li><p>DiSECt: Differentiable Simulator for Robotic Cutting - Heiden, E.,
-Macklin, M., Narang, Y., Fox, D., Garg, A., &amp; Ramos, F
-<a class="reference external" href="https://github.com/NVlabs/DiSECt">(2021)</a></p></li>
-<li><p>gradSim: Differentiable Simulation for System Identification and
-Visuomotor Control - Murthy, J. Krishna, Miles Macklin, Florian
-Golemo, Vikram Voleti, Linda Petrini, Martin Weiss, Breandan
-Considine et
-al. <a class="reference external" href="https://gradsim.github.io">(2021)</a></p></li>
-</ul>
-</section>
-<section id="citing">
-<h2>Citing<a class="headerlink" href="#citing" title="Link to this heading">#</a></h2>
-<p>If you use Warp in your research please use the following citation:</p>
-<div class="highlight-bibtex notranslate"><div class="highlight"><pre><span></span><span class="nc">@misc</span><span class="p">{</span><span class="nl">warp2022</span><span class="p">,</span>
-<span class="w">    </span><span class="na">title</span><span class="p">=</span><span class="w"> </span><span class="s">{Warp: A High-performance Python Framework for GPU Simulation and Graphics}</span><span class="p">,</span>
-<span class="w">    </span><span class="na">author</span><span class="w"> </span><span class="p">=</span><span class="w"> </span><span class="s">{Miles Macklin}</span><span class="p">,</span>
-<span class="w">    </span><span class="na">month</span><span class="w"> </span><span class="p">=</span><span class="w"> </span><span class="s">{March}</span><span class="p">,</span>
-<span class="w">    </span><span class="na">year</span><span class="w"> </span><span class="p">=</span><span class="w"> </span><span class="s">{2022}</span><span class="p">,</span>
-<span class="w">    </span><span class="na">note</span><span class="p">=</span><span class="w"> </span><span class="s">{NVIDIA GPU Technology Conference (GTC)}</span><span class="p">,</span>
-<span class="w">    </span><span class="na">howpublished</span><span class="w"> </span><span class="p">=</span><span class="w"> </span><span class="s">{\url{https://github.com/nvidia/warp}}</span>
-<span class="p">}</span>
-</pre></div>
-</div>
-</section>
-<section id="license">
-<h2>License<a class="headerlink" href="#license" title="Link to this heading">#</a></h2>
-<p>Warp is provided under the NVIDIA Source Code License (NVSCL), please see
-<a class="reference external" href="https://github.com/NVIDIA/warp/blob/main/LICENSE.md">LICENSE.md</a> for the full license text.</p>
-<p>Please contact <a class="reference external" href="mailto:omniverse-license-questions&#37;&#52;&#48;nvidia&#46;com">omniverse-license-questions<span>&#64;</span>nvidia<span>&#46;</span>com</a> for
-commercial licensing inquires.</p>
-</section>
-<section id="full-table-of-contents">
-<h2>Full Table of Contents<a class="headerlink" href="#full-table-of-contents" title="Link to this heading">#</a></h2>
-<div class="toctree-wrapper compound">
-<p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="installation.html#dependencies">Dependencies</a></li>
-<li class="toctree-l2"><a class="reference internal" href="installation.html#building-from-source">Building from source</a></li>
-<li class="toctree-l2"><a class="reference internal" href="installation.html#conda-environments">Conda environments</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="basics.html#initialization">Initialization</a></li>
-<li class="toctree-l2"><a class="reference internal" href="basics.html#kernels">Kernels</a></li>
-<li class="toctree-l2"><a class="reference internal" href="basics.html#arrays">Arrays</a></li>
-<li class="toctree-l2"><a class="reference internal" href="basics.html#user-functions">User Functions</a></li>
-<li class="toctree-l2"><a class="reference internal" href="basics.html#compilation-model">Compilation Model</a></li>
-<li class="toctree-l2"><a class="reference internal" href="basics.html#language-details">Language Details</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/devices.html#default-device">Default Device</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/devices.html#scoped-devices">Scoped Devices</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/devices.html#current-cuda-device">Current CUDA Device</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/devices.html#device-synchronization">Device Synchronization</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/devices.html#custom-cuda-contexts">Custom CUDA Contexts</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/interoperability.html#numpy">NumPy</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/interoperability.html#pytorch">PyTorch</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/interoperability.html#cupy-numba">CuPy/Numba</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/interoperability.html#jax">JAX</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/interoperability.html#module-warp.dlpack">DLPack</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="configuration.html#global-settings">Global Settings</a></li>
-<li class="toctree-l2"><a class="reference internal" href="configuration.html#module-settings">Module Settings</a></li>
-<li class="toctree-l2"><a class="reference internal" href="configuration.html#kernel-settings">Kernel Settings</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="debugging.html#printing-values">Printing Values</a></li>
-<li class="toctree-l2"><a class="reference internal" href="debugging.html#printing-launches">Printing Launches</a></li>
-<li class="toctree-l2"><a class="reference internal" href="debugging.html#step-through-debugging">Step-Through Debugging</a></li>
-<li class="toctree-l2"><a class="reference internal" href="debugging.html#generated-code">Generated Code</a></li>
-<li class="toctree-l2"><a class="reference internal" href="debugging.html#bounds-checking">Bounds Checking</a></li>
-<li class="toctree-l2"><a class="reference internal" href="debugging.html#cuda-verification">CUDA Verification</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="limitations.html#unsupported-features">Unsupported Features</a></li>
-<li class="toctree-l2"><a class="reference internal" href="limitations.html#kernels">Kernels</a></li>
-<li class="toctree-l2"><a class="reference internal" href="limitations.html#arrays">Arrays</a></li>
-<li class="toctree-l2"><a class="reference internal" href="limitations.html#structs">Structs</a></li>
-<li class="toctree-l2"><a class="reference internal" href="limitations.html#volumes">Volumes</a></li>
-<li class="toctree-l2"><a class="reference internal" href="limitations.html#multiple-processes">Multiple Processes</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="faq.html#how-does-warp-relate-to-other-python-projects-for-gpu-programming-e-g-numba-taichi-cupy-pytorch-etc">How does Warp relate to other Python projects for GPU programming, e.g.: Numba, Taichi, cuPy, PyTorch, etc.?</a></li>
-<li class="toctree-l2"><a class="reference internal" href="faq.html#does-warp-support-all-of-the-python-language">Does Warp support all of the Python language?</a></li>
-<li class="toctree-l2"><a class="reference internal" href="faq.html#when-should-i-call-wp-synchronize">When should I call <code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code>?</a></li>
-<li class="toctree-l2"><a class="reference internal" href="faq.html#what-happens-when-you-differentiate-a-function-like-wp-abs-x">What happens when you differentiate a function like <code class="docutils literal notranslate"><span class="pre">wp.abs(x)</span></code>?</a></li>
-<li class="toctree-l2"><a class="reference internal" href="faq.html#does-warp-support-multi-gpu-programming">Does Warp support multi-GPU programming?</a></li>
-<li class="toctree-l2"><a class="reference internal" href="faq.html#should-i-switch-to-warp-over-isaacgym-physx">Should I switch to Warp over IsaacGym / PhysX?</a></li>
-</ul>
-</li>
-</ul>
-</div>
-<div class="toctree-wrapper compound">
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#kernels">Kernels</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#arrays">Arrays</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#data-types">Data Types</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#constants">Constants</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#operators">Operators</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#graphs">Graphs</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#bounding-value-hierarchies-bvh">Bounding Value Hierarchies (BVH)</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#meshes">Meshes</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#hash-grids">Hash Grids</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#volumes">Volumes</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#differentiability">Differentiability</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#custom-native-functions">Custom Native Functions</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/runtime.html#profiling">Profiling</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#scalar-types">Scalar Types</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#vector-types">Vector Types</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#generic-types">Generic Types</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#scalar-math">Scalar Math</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#vector-math">Vector Math</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#other">Other</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#quaternion-math">Quaternion Math</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#transformations">Transformations</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#spatial-math">Spatial Math</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#utility">Utility</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#geometry">Geometry</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#volumes">Volumes</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#random">Random</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/functions.html#operators">Operators</a></li>
-</ul>
-</li>
-</ul>
-</div>
-<div class="toctree-wrapper compound">
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/sim.html#model">Model</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/sim.html#state">State</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/sim.html#forward-inverse-kinematics">Forward / Inverse Kinematics</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/sim.html#integrators">Integrators</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/sim.html#importers">Importers</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/sparse.html#sparse-matrices">Sparse Matrices</a></li>
-</ul>
-</li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a><ul>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#integrands">Integrands</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#basic-workflow">Basic workflow</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#introductory-examples">Introductory examples</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#advanced-usages">Advanced usages</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#operators">Operators</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#integration">Integration</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#geometry">Geometry</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#function-spaces">Function Spaces</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#fields">Fields</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#boundary-conditions">Boundary Conditions</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#id2">Memory management</a></li>
-<li class="toctree-l2"><a class="reference internal" href="modules/fem.html#interfaces">Interfaces</a></li>
-</ul>
-</li>
-</ul>
-</div>
-<div class="toctree-wrapper compound">
-</div>
-<p><a class="reference internal" href="genindex.html"><span class="std std-ref">Full Index</span></a></p>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="installation.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Installation</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer no-toc">
-      
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/installation.html b/docs/_build/html/installation.html
deleted file mode 100644
index f5cbace02..000000000
--- a/docs/_build/html/installation.html
+++ /dev/null
@@ -1,396 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="Basics" href="basics.html" /><link rel="prev" title="NVIDIA Warp Documentation" href="index.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Installation - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="installation">
-<h1>Installation<a class="headerlink" href="#installation" title="Link to this heading">#</a></h1>
-<p>The easiest way is to install Warp is from <a class="reference external" href="https://pypi.org/project/warp-lang">PyPi</a>:</p>
-<div class="highlight-sh notranslate"><div class="highlight"><pre><span></span>$<span class="w"> </span>pip<span class="w"> </span>install<span class="w"> </span>warp-lang
-</pre></div>
-</div>
-<p>Pre-built binary packages for Windows, Linux and macOS are also available on the <a class="reference external" href="https://github.com/NVIDIA/warp/releases">Releases</a> page. To install in your local Python environment extract the archive and run the following command from the root directory:</p>
-<div class="highlight-sh notranslate"><div class="highlight"><pre><span></span>$<span class="w"> </span>pip<span class="w"> </span>install<span class="w"> </span>.
-</pre></div>
-</div>
-<section id="dependencies">
-<h2>Dependencies<a class="headerlink" href="#dependencies" title="Link to this heading">#</a></h2>
-<p>Warp supports Python versions 3.7 or later and requires <a class="reference external" href="https://numpy.org">NumPy</a> to be installed.</p>
-<p>The following optional dependencies are required to support certain features:</p>
-<ul class="simple">
-<li><p><a class="reference external" href="https://pypi.org/project/usd-core">usd-core</a>: Required for some Warp examples, <code class="docutils literal notranslate"><span class="pre">warp.sim.parse_usd()</span></code>, and <code class="docutils literal notranslate"><span class="pre">warp.render.UsdRenderer</span></code>.</p></li>
-<li><p><a class="reference external" href="https://jax.readthedocs.io/en/latest/installation.html">JAX</a>: Required for JAX interoperability (see <a class="reference internal" href="modules/interoperability.html#jax-interop"><span class="std std-ref">JAX</span></a>).</p></li>
-<li><p><a class="reference external" href="https://pytorch.org/get-started/locally/">PyTorch</a>: Required for PyTorch interoperability (see <a class="reference internal" href="modules/interoperability.html#pytorch-interop"><span class="std std-ref">PyTorch</span></a>).</p></li>
-<li><p><a class="reference external" href="https://github.com/NVIDIA/NVTX#python">NVTX for Python</a>: Required to use <a class="reference internal" href="modules/runtime.html#warp.ScopedTimer" title="warp.ScopedTimer"><code class="xref py py-class docutils literal notranslate"><span class="pre">wp.ScopedTimer(use_nvtx=True)</span></code></a>.</p></li>
-</ul>
-<p>Building the Warp documentation requires:</p>
-<ul class="simple">
-<li><p><a class="reference external" href="https://www.sphinx-doc.org">Sphinx</a></p></li>
-<li><p><a class="reference external" href="https://github.com/pradyunsg/furo">Furo</a></p></li>
-<li><p><a class="reference external" href="https://sphinx-copybutton.readthedocs.io/en/latest/index.html">Sphinx-copybutton</a></p></li>
-</ul>
-</section>
-<section id="building-from-source">
-<h2>Building from source<a class="headerlink" href="#building-from-source" title="Link to this heading">#</a></h2>
-<p>For developers who want to build the library themselves the following tools are required:</p>
-<ul class="simple">
-<li><p>Microsoft Visual Studio (Windows), minimum version 2019</p></li>
-<li><p>GCC (Linux), minimum version 7.2</p></li>
-<li><p><a class="reference external" href="https://developer.nvidia.com/cuda-toolkit">CUDA Toolkit</a>, minimum version 11.5</p></li>
-<li><p><a class="reference external" href="https://git-lfs.com">Git Large File Storage</a></p></li>
-</ul>
-<p>If you are cloning from Windows, please first ensure that you have
-enabled “Developer Mode” in Windows settings and symlinks in Git:</p>
-<div class="highlight-console notranslate"><div class="highlight"><pre><span></span><span class="gp">$ </span>git<span class="w"> </span>config<span class="w"> </span>--global<span class="w"> </span>core.symlinks<span class="w"> </span><span class="nb">true</span>
-</pre></div>
-</div>
-<p>This will ensure symlinks inside <code class="docutils literal notranslate"><span class="pre">exts/omni.warp.core</span></code> work upon cloning.</p>
-<p>After cloning the repository, users should run:</p>
-<div class="highlight-console notranslate"><div class="highlight"><pre><span></span><span class="gp">$ </span>python<span class="w"> </span>build_lib.py
-</pre></div>
-</div>
-<p>This will generate the <code class="docutils literal notranslate"><span class="pre">warp.dll</span></code> / <code class="docutils literal notranslate"><span class="pre">warp.so</span></code> core library
-respectively. When building manually, users should ensure that their
-<code class="docutils literal notranslate"><span class="pre">CUDA_PATH</span></code> environment variable is set, otherwise Warp will be built
-without CUDA support. Alternatively, the path to the CUDA Toolkit can be
-passed to the build command as <code class="docutils literal notranslate"><span class="pre">--cuda_path=&quot;...&quot;</span></code>. After building, the
-Warp package should be installed using:</p>
-<div class="highlight-console notranslate"><div class="highlight"><pre><span></span><span class="gp">$ </span>pip<span class="w"> </span>install<span class="w"> </span>-e<span class="w"> </span>.
-</pre></div>
-</div>
-<p>Which ensures that subsequent modifications to the library will be
-reflected in the Python package.</p>
-</section>
-<section id="conda-environments">
-<h2>Conda environments<a class="headerlink" href="#conda-environments" title="Link to this heading">#</a></h2>
-<p>Some modules, such as <code class="docutils literal notranslate"><span class="pre">usd-core</span></code>, don’t support the latest Python version.
-To manage running Warp and other projects on different Python versions one can
-make use of an environment management system such as
-<a class="reference external" href="https://docs.conda.io/">Conda</a>.</p>
-<p><strong>WARNING:</strong> When building and running Warp in a different environment, make sure
-the build environment has the same C++ runtime library version, or an older
-one, than the execution environment. Otherwise Warp’s shared libraries may end
-up looking for a newer runtime library version than the one available in the
-execution environment. For example on Linux this error could occur:</p>
-<p><code class="docutils literal notranslate"><span class="pre">OSError:</span> <span class="pre">&lt;...&gt;/libstdc++.so.6:</span> <span class="pre">version</span> <span class="pre">`GLIBCXX_3.4.30'</span> <span class="pre">not</span> <span class="pre">found</span> <span class="pre">(required</span> <span class="pre">by</span> <span class="pre">&lt;...&gt;/warp/warp/bin/warp.so)</span></code></p>
-<p>This can be solved by installing a newer C++ runtime version in the runtime
-conda environment using <code class="docutils literal notranslate"><span class="pre">conda</span> <span class="pre">install</span> <span class="pre">-c</span> <span class="pre">conda-forge</span> <span class="pre">libstdcxx-ng=12.1</span></code> or
-newer. Or, the build environment’s C++ toolchain can be downgraded using
-<code class="docutils literal notranslate"><span class="pre">conda</span> <span class="pre">install</span> <span class="pre">-c</span> <span class="pre">conda-forge</span> <span class="pre">libstdcxx-ng=8.5</span></code>. Or, one can <code class="docutils literal notranslate"><span class="pre">activate</span></code> or
-<code class="docutils literal notranslate"><span class="pre">deactivate</span></code> conda environments as needed for building vs. running Warp.</p>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="basics.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Basics</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="index.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Home</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Installation</a><ul>
-<li><a class="reference internal" href="#dependencies">Dependencies</a></li>
-<li><a class="reference internal" href="#building-from-source">Building from source</a></li>
-<li><a class="reference internal" href="#conda-environments">Conda environments</a></li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/limitations.html b/docs/_build/html/limitations.html
deleted file mode 100644
index a0129411c..000000000
--- a/docs/_build/html/limitations.html
+++ /dev/null
@@ -1,382 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" /><link rel="next" title="FAQ" href="faq.html" /><link rel="prev" title="Debugging" href="debugging.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Limitations - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="limitations">
-<h1>Limitations<a class="headerlink" href="#limitations" title="Link to this heading">#</a></h1>
-<p>This section summarizes various limitations and currently unsupported features in Warp.
-Requests for new features can be made at <a class="reference external" href="https://github.com/NVIDIA/warp/discussions">GitHub Discussions</a>,
-and issues can be opened at <a class="reference external" href="https://github.com/NVIDIA/warp/issues">GitHub Issues</a>.</p>
-<section id="unsupported-features">
-<h2>Unsupported Features<a class="headerlink" href="#unsupported-features" title="Link to this heading">#</a></h2>
-<p>To achieve good performance on GPUs some dynamic language features are not supported:</p>
-<ul class="simple">
-<li><p>Lambda functions</p></li>
-<li><p>List comprehensions</p></li>
-<li><p>Exceptions</p></li>
-<li><p>Recursion</p></li>
-<li><p>Runtime evaluation of expressions, e.g.: eval()</p></li>
-<li><p>Dynamic structures such as lists, sets, dictionaries, etc.</p></li>
-</ul>
-</section>
-<section id="kernels">
-<h2>Kernels<a class="headerlink" href="#kernels" title="Link to this heading">#</a></h2>
-<ul class="simple">
-<li><p>Strings cannot be passed into kernels.</p></li>
-<li><p>Short-circuit evaluation is not supported.</p></li>
-<li><p><a class="reference internal" href="modules/functions.html#warp.atomic_add" title="warp.atomic_add"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.atomic_add()</span></code></a> does not support <code class="docutils literal notranslate"><span class="pre">wp.int64</span></code>.</p></li>
-<li><p><a class="reference internal" href="modules/functions.html#warp.tid" title="warp.tid"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.tid()</span></code></a> cannot be called from user functions.</p></li>
-<li><p>CUDA thread blocks use a fixed size 256 threads per block.</p></li>
-</ul>
-</section>
-<section id="arrays">
-<h2>Arrays<a class="headerlink" href="#arrays" title="Link to this heading">#</a></h2>
-<ul class="simple">
-<li><p>Arrays can have a maximum of four dimensions.</p></li>
-<li><p>Each dimension of a Warp array cannot be greater than the maximum value representable by a 32-bit signed integer,
-<span class="math notranslate nohighlight">\(2^{31}-1\)</span>.</p></li>
-<li><p>There are currently no data types that support complex numbers.</p></li>
-</ul>
-</section>
-<section id="structs">
-<h2>Structs<a class="headerlink" href="#structs" title="Link to this heading">#</a></h2>
-<ul class="simple">
-<li><p>Structs cannot have generic members, i.e. of type <code class="docutils literal notranslate"><span class="pre">typing.Any</span></code>.</p></li>
-</ul>
-</section>
-<section id="volumes">
-<h2>Volumes<a class="headerlink" href="#volumes" title="Link to this heading">#</a></h2>
-<ul class="simple">
-<li><p>The sparse-volume <em>topology</em> cannot be changed after the tiles for the <a class="reference internal" href="modules/runtime.html#warp.Volume" title="warp.Volume"><code class="xref py py-class docutils literal notranslate"><span class="pre">Volume</span></code></a> have been allocated.</p></li>
-</ul>
-</section>
-<section id="multiple-processes">
-<h2>Multiple Processes<a class="headerlink" href="#multiple-processes" title="Link to this heading">#</a></h2>
-<ul class="simple">
-<li><p>A CUDA context created in the parent process cannot be used in a <em>forked</em> child process.
-Use the spawn start method instead, or avoid creating CUDA contexts in the parent process.</p></li>
-<li><p>There can be issues with using same user kernel cache directory when running with multiple processes.
-A workaround is to use a separate cache directory for every process.
-See <a class="reference internal" href="basics.html#example-cache-management"><span class="std std-ref">Example: Changing the kernel cache directory</span></a> for how the cache directory may be changed.</p></li>
-</ul>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="faq.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">FAQ</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="debugging.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Debugging</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Limitations</a><ul>
-<li><a class="reference internal" href="#unsupported-features">Unsupported Features</a></li>
-<li><a class="reference internal" href="#kernels">Kernels</a></li>
-<li><a class="reference internal" href="#arrays">Arrays</a></li>
-<li><a class="reference internal" href="#structs">Structs</a></li>
-<li><a class="reference internal" href="#volumes">Volumes</a></li>
-<li><a class="reference internal" href="#multiple-processes">Multiple Processes</a></li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    <script async="async" src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/devices.html b/docs/_build/html/modules/devices.html
deleted file mode 100644
index 31c18bee4..000000000
--- a/docs/_build/html/modules/devices.html
+++ /dev/null
@@ -1,511 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="Interoperability" href="interoperability.html" /><link rel="prev" title="Basics" href="../basics.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Devices - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="devices">
-<span id="id1"></span><h1>Devices<a class="headerlink" href="#devices" title="Link to this heading">#</a></h1>
-<p>Warp assigns unique string aliases to all supported compute devices in the system.  There is currently a single CPU device exposed as <code class="docutils literal notranslate"><span class="pre">&quot;cpu&quot;</span></code>.  Each CUDA-capable GPU gets an alias of the form <code class="docutils literal notranslate"><span class="pre">&quot;cuda:i&quot;</span></code>, where <code class="docutils literal notranslate"><span class="pre">i</span></code> is the CUDA device ordinal.  This convention should be familiar to users of other popular frameworks like PyTorch.</p>
-<p>It is possible to explicitly target a specific device with each Warp API call using the <code class="docutils literal notranslate"><span class="pre">device</span></code> argument:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-
-<span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">b</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">b</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-
-<span class="n">c</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:1&quot;</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">c</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">c</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:1&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>A Warp CUDA device (<code class="docutils literal notranslate"><span class="pre">&quot;cuda:i&quot;</span></code>) corresponds to the primary CUDA context of device <code class="docutils literal notranslate"><span class="pre">i</span></code>.  This is compatible with frameworks like PyTorch and other software that uses the CUDA Runtime API.  It makes interoperability easy, because GPU resources like memory can be shared with Warp.</p>
-</div>
-<section id="default-device">
-<h2>Default Device<a class="headerlink" href="#default-device" title="Link to this heading">#</a></h2>
-<p>To simplify writing code, Warp has the concept of <strong>default device</strong>.  When the <code class="docutils literal notranslate"><span class="pre">device</span></code> argument is omitted from a Warp API call, the default device will be used.</p>
-<p>During Warp initialization, the default device is set to be <code class="docutils literal notranslate"><span class="pre">&quot;cuda:0&quot;</span></code> if CUDA is available.  Otherwise, the default device is <code class="docutils literal notranslate"><span class="pre">&quot;cpu&quot;</span></code>.</p>
-<p>The function <code class="docutils literal notranslate"><span class="pre">wp.set_device()</span></code> can be used to change the default device:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">set_device</span><span class="p">(</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">set_device</span><span class="p">(</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-<span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">b</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">b</span><span class="p">])</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">set_device</span><span class="p">(</span><span class="s2">&quot;cuda:1&quot;</span><span class="p">)</span>
-<span class="n">c</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">c</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">c</span><span class="p">])</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>For CUDA devices, <code class="docutils literal notranslate"><span class="pre">wp.set_device()</span></code> does two things: it sets the Warp default device and it makes the device’s CUDA context current.  This helps to minimize the number of CUDA context switches in blocks of code targeting a single device.</p>
-</div>
-<p>For PyTorch users, this function is similar to <code class="docutils literal notranslate"><span class="pre">torch.cuda.set_device()</span></code>.  It is still possible to specify a different device in individual API calls, like in this snippet:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># set default device</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">set_device</span><span class="p">(</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-
-<span class="c1"># use default device</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-
-<span class="c1"># use explicit devices</span>
-<span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">n</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-<span class="n">c</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">n</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:1&quot;</span><span class="p">)</span>
-
-<span class="c1"># use default device</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">b</span><span class="p">,</span> <span class="n">a</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">c</span><span class="p">,</span> <span class="n">a</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-<section id="scoped-devices">
-<h2>Scoped Devices<a class="headerlink" href="#scoped-devices" title="Link to this heading">#</a></h2>
-<p>Another way to manage the default device is using <code class="docutils literal notranslate"><span class="pre">wp.ScopedDevice</span></code> objects.  They can be arbitrarily nested and restore the previous default device on exit:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedDevice</span><span class="p">(</span><span class="s2">&quot;cpu&quot;</span><span class="p">):</span>
-    <span class="c1"># alloc and launch on &quot;cpu&quot;</span>
-    <span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-
-<span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedDevice</span><span class="p">(</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">):</span>
-    <span class="c1"># alloc on &quot;cuda:0&quot;</span>
-    <span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-
-    <span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedDevice</span><span class="p">(</span><span class="s2">&quot;cuda:1&quot;</span><span class="p">):</span>
-        <span class="c1"># alloc and launch on &quot;cuda:1&quot;</span>
-        <span class="n">c</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">c</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">c</span><span class="p">])</span>
-
-    <span class="c1"># launch on &quot;cuda:0&quot;</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">b</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">b</span><span class="p">])</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>For CUDA devices, <code class="docutils literal notranslate"><span class="pre">wp.ScopedDevice</span></code> makes the device’s CUDA context current and restores the previous CUDA context on exit.  This is handy when running Warp scripts as part of a bigger pipeline, because it avoids any side effects of changing the CUDA context in the enclosed code.</p>
-</div>
-<section id="example-using-wp-scopeddevice-with-multiple-gpus">
-<h3>Example: Using <code class="docutils literal notranslate"><span class="pre">wp.ScopedDevice</span></code> with multiple GPUs<a class="headerlink" href="#example-using-wp-scopeddevice-with-multiple-gpus" title="Link to this heading">#</a></h3>
-<p>The following example shows how to allocate arrays and launch kernels on all available CUDA devices.</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">inc</span><span class="p">(</span><span class="n">a</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">a</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">a</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">+</span> <span class="mf">1.0</span>
-
-
-<span class="c1"># get all CUDA devices</span>
-<span class="n">devices</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">get_cuda_devices</span><span class="p">()</span>
-<span class="n">device_count</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">devices</span><span class="p">)</span>
-
-<span class="c1"># number of launches</span>
-<span class="n">iters</span> <span class="o">=</span> <span class="mi">1000</span>
-
-<span class="c1"># list of arrays, one per device</span>
-<span class="n">arrs</span> <span class="o">=</span> <span class="p">[]</span>
-
-<span class="c1"># loop over all devices</span>
-<span class="k">for</span> <span class="n">device</span> <span class="ow">in</span> <span class="n">devices</span><span class="p">:</span>
-    <span class="c1"># use a ScopedDevice to set the target device</span>
-    <span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedDevice</span><span class="p">(</span><span class="n">device</span><span class="p">):</span>
-        <span class="c1"># allocate array</span>
-        <span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">250</span> <span class="o">*</span> <span class="mi">1024</span> <span class="o">*</span> <span class="mi">1024</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)</span>
-        <span class="n">arrs</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">a</span><span class="p">)</span>
-
-        <span class="c1"># launch kernels</span>
-        <span class="k">for</span> <span class="n">_</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iters</span><span class="p">):</span>
-            <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">inc</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-
-<span class="c1"># synchronize all devices</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">synchronize</span><span class="p">()</span>
-
-<span class="c1"># print results</span>
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">device_count</span><span class="p">):</span>
-    <span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">arrs</span><span class="p">[</span><span class="n">i</span><span class="p">]</span><span class="o">.</span><span class="n">device</span><span class="si">}</span><span class="s2"> -&gt; </span><span class="si">{</span><span class="n">arrs</span><span class="p">[</span><span class="n">i</span><span class="p">]</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-</section>
-<section id="current-cuda-device">
-<h2>Current CUDA Device<a class="headerlink" href="#current-cuda-device" title="Link to this heading">#</a></h2>
-<p>Warp uses the device alias <code class="docutils literal notranslate"><span class="pre">&quot;cuda&quot;</span></code> to target the current CUDA device.  This allows external code to manage the CUDA device on which to execute Warp scripts.  It is analogous to the PyTorch <code class="docutils literal notranslate"><span class="pre">&quot;cuda&quot;</span></code> device, which should be familiar to Torch users and simplify interoperation.</p>
-<p>In this snippet, we use PyTorch to manage the current CUDA device and invoke a Warp kernel on that device:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">example_function</span><span class="p">():</span>
-    <span class="c1"># create a Torch tensor on the current CUDA device</span>
-    <span class="n">t</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-    <span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">from_torch</span><span class="p">(</span><span class="n">t</span><span class="p">)</span>
-
-    <span class="c1"># launch a Warp kernel on the current CUDA device</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="c1"># use Torch to set the current CUDA device and run example_function() on that device</span>
-<span class="n">torch</span><span class="o">.</span><span class="n">cuda</span><span class="o">.</span><span class="n">set_device</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>
-<span class="n">example_function</span><span class="p">()</span>
-
-<span class="c1"># use Torch to change the current CUDA device and re-run example_function() on that device</span>
-<span class="n">torch</span><span class="o">.</span><span class="n">cuda</span><span class="o">.</span><span class="n">set_device</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
-<span class="n">example_function</span><span class="p">()</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Using the device alias <code class="docutils literal notranslate"><span class="pre">&quot;cuda&quot;</span></code> can be problematic if the code runs in an environment where another part of the code can unpredictably change the CUDA context.  Using an explicit CUDA device like <code class="docutils literal notranslate"><span class="pre">&quot;cuda:i&quot;</span></code> is recommended to avoid such issues.</p>
-</div>
-</section>
-<section id="device-synchronization">
-<h2>Device Synchronization<a class="headerlink" href="#device-synchronization" title="Link to this heading">#</a></h2>
-<p>CUDA kernel launches and memory operations can execute asynchronously.  This allows for overlapping compute and memory operations on different devices.  Warp allows synchronizing the host with outstanding asynchronous operations on a specific device:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">synchronize_device</span><span class="p">(</span><span class="s2">&quot;cuda:1&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>The <code class="docutils literal notranslate"><span class="pre">wp.synchronize_device()</span></code> function offers more fine-grained synchronization than <code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code>, as the latter waits for <em>all</em> devices to complete their work.</p>
-</section>
-<section id="custom-cuda-contexts">
-<h2>Custom CUDA Contexts<a class="headerlink" href="#custom-cuda-contexts" title="Link to this heading">#</a></h2>
-<p>Warp is designed to work with arbitrary CUDA contexts so it can easily integrate into different workflows.</p>
-<p>Applications built on the CUDA Runtime API target the <em>primary context</em> of each device.  The Runtime API hides CUDA context management under the hood.  In Warp, device <code class="docutils literal notranslate"><span class="pre">&quot;cuda:i&quot;</span></code> represents the primary context of device <code class="docutils literal notranslate"><span class="pre">i</span></code>, which aligns with the CUDA Runtime API.</p>
-<p>Applications built on the CUDA Driver API work with CUDA contexts directly and can create custom CUDA contexts on any device.  Custom CUDA contexts can be created with specific affinity or interop features that benefit the application.  Warp can work with these CUDA contexts as well.</p>
-<p>The special device alias <code class="docutils literal notranslate"><span class="pre">&quot;cuda&quot;</span></code> can be used to target the current CUDA context, whether this is a primary or custom context.</p>
-<p>In addition, Warp allows registering new device aliases for custom CUDA contexts, so that they can be explicitly targeted by name.  If the <code class="docutils literal notranslate"><span class="pre">CUcontext</span></code> pointer is available, it can be used to create a new device alias like this:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">map_cuda_device</span><span class="p">(</span><span class="s2">&quot;foo&quot;</span><span class="p">,</span> <span class="n">ctypes</span><span class="o">.</span><span class="n">c_void_p</span><span class="p">(</span><span class="n">context_ptr</span><span class="p">))</span>
-</pre></div>
-</div>
-<p>Alternatively, if the custom CUDA context was made current by the application, the pointer can be omitted:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">map_cuda_device</span><span class="p">(</span><span class="s2">&quot;foo&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>In either case, mapping the custom CUDA context allows us to target the context directly using the assigned alias:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedDevice</span><span class="p">(</span><span class="s2">&quot;foo&quot;</span><span class="p">):</span>
-    <span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-</pre></div>
-</div>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="interoperability.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Interoperability</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="../basics.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Basics</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Devices</a><ul>
-<li><a class="reference internal" href="#default-device">Default Device</a></li>
-<li><a class="reference internal" href="#scoped-devices">Scoped Devices</a><ul>
-<li><a class="reference internal" href="#example-using-wp-scopeddevice-with-multiple-gpus">Example: Using <code class="docutils literal notranslate"><span class="pre">wp.ScopedDevice</span></code> with multiple GPUs</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#current-cuda-device">Current CUDA Device</a></li>
-<li><a class="reference internal" href="#device-synchronization">Device Synchronization</a></li>
-<li><a class="reference internal" href="#custom-cuda-contexts">Custom CUDA Contexts</a></li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/fem.html b/docs/_build/html/modules/fem.html
deleted file mode 100644
index 7ff7f8fbc..000000000
--- a/docs/_build/html/modules/fem.html
+++ /dev/null
@@ -1,2063 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="prev" title="warp.sparse" href="sparse.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>warp.fem - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="warp-fem">
-<h1>warp.fem<a class="headerlink" href="#warp-fem" title="Link to this heading">#</a></h1>
-<p>The <code class="docutils literal notranslate"><span class="pre">warp.fem</span></code> module is designed to facilitate solving physical systems described as differential
-equations. For example, it can solve PDEs for diffusion, convection, fluid flow, and elasticity problems
-using finite-element-based (FEM) Galerkin methods, and allows users to quickly experiment with various FEM
-formulations and discretization schemes.</p>
-<section id="integrands">
-<h2>Integrands<a class="headerlink" href="#integrands" title="Link to this heading">#</a></h2>
-<p>The core functionality of the FEM toolkit is the ability to integrate constant, linear, and bilinear forms
-over various domains and using arbitrary interpolation basis.</p>
-<p>The main mechanism is the <a class="reference internal" href="#warp.fem.integrand" title="warp.fem.integrand"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrand()</span></code></a> decorator, for instance:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@integrand</span>
-<span class="k">def</span> <span class="nf">linear_form</span><span class="p">(</span>
-    <span class="n">s</span><span class="p">:</span> <span class="n">Sample</span><span class="p">,</span>
-    <span class="n">v</span><span class="p">:</span> <span class="n">Field</span><span class="p">,</span>
-<span class="p">):</span>
-    <span class="k">return</span> <span class="n">v</span><span class="p">(</span><span class="n">s</span><span class="p">)</span>
-
-
-<span class="nd">@integrand</span>
-<span class="k">def</span> <span class="nf">diffusion_form</span><span class="p">(</span><span class="n">s</span><span class="p">:</span> <span class="n">Sample</span><span class="p">,</span> <span class="n">u</span><span class="p">:</span> <span class="n">Field</span><span class="p">,</span> <span class="n">v</span><span class="p">:</span> <span class="n">Field</span><span class="p">,</span> <span class="n">nu</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
-    <span class="k">return</span> <span class="n">nu</span> <span class="o">*</span> <span class="n">wp</span><span class="o">.</span><span class="n">dot</span><span class="p">(</span>
-        <span class="n">grad</span><span class="p">(</span><span class="n">u</span><span class="p">,</span> <span class="n">s</span><span class="p">),</span>
-        <span class="n">grad</span><span class="p">(</span><span class="n">v</span><span class="p">,</span> <span class="n">s</span><span class="p">),</span>
-    <span class="p">)</span>
-</pre></div>
-</div>
-<p>Integrands are normal warp kernels, meaning any usual warp function can be used.
-However, they accept a few special parameters:</p>
-<blockquote>
-<div><ul>
-<li><p><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><code class="xref py py-class docutils literal notranslate"><span class="pre">Sample</span></code></a> contains information about the current integration sample point, such as the element index and coordinates in element.</p></li>
-<li><p><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><code class="xref py py-class docutils literal notranslate"><span class="pre">Field</span></code></a> designates an abstract field, which will be replaced at call time by the actual field type: for instance a <a class="reference internal" href="#warp.fem.DiscreteField" title="warp.fem.DiscreteField"><code class="xref py py-class docutils literal notranslate"><span class="pre">DiscreteField</span></code></a>, <a class="reference internal" href="#warp.fem.field.TestField" title="warp.fem.field.TestField"><code class="xref py py-class docutils literal notranslate"><span class="pre">field.TestField</span></code></a> or <a class="reference internal" href="#warp.fem.field.TrialField" title="warp.fem.field.TrialField"><code class="xref py py-class docutils literal notranslate"><span class="pre">field.TrialField</span></code></a> defined over an arbitrary <a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.FunctionSpace"><code class="xref py py-class docutils literal notranslate"><span class="pre">FunctionSpace</span></code></a>.
-A field <cite>u</cite> can be evaluated at a given sample <cite>s</cite> using the <a class="reference internal" href="#warp.fem.inner" title="warp.fem.inner"><code class="xref py py-func docutils literal notranslate"><span class="pre">inner()</span></code></a> operator, i.e, <code class="docutils literal notranslate"><span class="pre">inner(u,</span> <span class="pre">s)</span></code>, or as a shortcut using the usual call operator, <code class="docutils literal notranslate"><span class="pre">u(s)</span></code>.
-Several other operators are available, such as <a class="reference internal" href="#warp.fem.grad" title="warp.fem.grad"><code class="xref py py-func docutils literal notranslate"><span class="pre">grad()</span></code></a>; see the <a class="reference internal" href="#operators"><span class="std std-ref">Operators</span></a> section.</p></li>
-<li><p><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><code class="xref py py-class docutils literal notranslate"><span class="pre">Domain</span></code></a> designates an abstract integration domain. Several operators are also provided for domains, for example in the example below evaluating the normal at the current sample position:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@integrand</span>
-<span class="k">def</span> <span class="nf">boundary_form</span><span class="p">(</span>
-    <span class="n">s</span><span class="p">:</span> <span class="n">Sample</span><span class="p">,</span>
-    <span class="n">domain</span><span class="p">:</span> <span class="n">Domain</span><span class="p">,</span>
-    <span class="n">u</span><span class="p">:</span> <span class="n">Field</span><span class="p">,</span>
-<span class="p">):</span>
-    <span class="n">nor</span> <span class="o">=</span> <span class="n">normal</span><span class="p">(</span><span class="n">domain</span><span class="p">,</span> <span class="n">s</span><span class="p">)</span>
-    <span class="k">return</span> <span class="n">wp</span><span class="o">.</span><span class="n">dot</span><span class="p">(</span><span class="n">u</span><span class="p">(</span><span class="n">s</span><span class="p">),</span> <span class="n">nor</span><span class="p">)</span>
-</pre></div>
-</div>
-</li>
-</ul>
-</div></blockquote>
-<p>Integrands cannot be used directly with <a class="reference internal" href="runtime.html#warp.launch" title="warp.launch"><code class="xref py py-func docutils literal notranslate"><span class="pre">warp.launch()</span></code></a>, but must be called through <a class="reference internal" href="#warp.fem.integrate" title="warp.fem.integrate"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrate()</span></code></a> or <a class="reference internal" href="#warp.fem.interpolate" title="warp.fem.interpolate"><code class="xref py py-func docutils literal notranslate"><span class="pre">interpolate()</span></code></a> instead.
-The root integrand (<cite>integrand</cite> argument passed to <a class="reference internal" href="#warp.fem.integrate" title="warp.fem.integrate"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrate()</span></code></a> or <a class="reference internal" href="#warp.fem.interpolate" title="warp.fem.interpolate"><code class="xref py py-func docutils literal notranslate"><span class="pre">interpolate()</span></code></a> call) will automatically get passed <a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><code class="xref py py-class docutils literal notranslate"><span class="pre">Sample</span></code></a> and <a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><code class="xref py py-class docutils literal notranslate"><span class="pre">Domain</span></code></a> parameters.
-<a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><code class="xref py py-class docutils literal notranslate"><span class="pre">Field</span></code></a> parameters must be passed as a dictionary in the <cite>fields</cite> argument of the launcher function, and all other standard Warp types arguments must be
-passed as a dictionary in the <cite>values</cite> argument of the launcher function, for instance:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">integrate</span><span class="p">(</span><span class="n">diffusion_form</span><span class="p">,</span> <span class="n">fields</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;u&quot;</span><span class="p">:</span> <span class="n">trial</span><span class="p">,</span> <span class="s2">&quot;v&quot;</span><span class="p">:</span> <span class="n">test</span><span class="p">},</span> <span class="n">values</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;nu&quot;</span><span class="p">:</span> <span class="n">viscosity</span><span class="p">})</span>
-</pre></div>
-</div>
-</section>
-<section id="basic-workflow">
-<h2>Basic workflow<a class="headerlink" href="#basic-workflow" title="Link to this heading">#</a></h2>
-<p>The typical steps for solving a linear PDE are as follow:</p>
-<blockquote>
-<div><ul class="simple">
-<li><p>Define a <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a> (grid, mesh, etc). At the moment, 2D and 3D regular grids, triangle, quadrilateral, tetrahedron and hexahedron meshes are supported.</p></li>
-<li><p>Define one or more <a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.FunctionSpace"><code class="xref py py-class docutils literal notranslate"><span class="pre">FunctionSpace</span></code></a>, by equipping the geometry elements with shape functions. See <a class="reference internal" href="#warp.fem.make_polynomial_space" title="warp.fem.make_polynomial_space"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_polynomial_space()</span></code></a>. At the moment, continuous/discontinuous Lagrange (<span class="math notranslate nohighlight">\(P_{k[d]}, Q_{k[d]}\)</span>) and Serendipity (<span class="math notranslate nohighlight">\(S_k\)</span>) shape functions of order <span class="math notranslate nohighlight">\(k \leq 3\)</span> are supported.</p></li>
-<li><p>Define an integration <a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.GeometryDomain"><code class="xref py py-class docutils literal notranslate"><span class="pre">GeometryDomain</span></code></a>, for instance the geometry’s cells (<a class="reference internal" href="#warp.fem.Cells" title="warp.fem.Cells"><code class="xref py py-class docutils literal notranslate"><span class="pre">Cells</span></code></a>) or boundary sides (<a class="reference internal" href="#warp.fem.BoundarySides" title="warp.fem.BoundarySides"><code class="xref py py-class docutils literal notranslate"><span class="pre">BoundarySides</span></code></a>).</p></li>
-<li><p>Integrate linear forms to build the system’s right-hand-side. Define a test function over the function space using <a class="reference internal" href="#warp.fem.make_test" title="warp.fem.make_test"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_test()</span></code></a>,
-a <a class="reference internal" href="#warp.fem.Quadrature" title="warp.fem.Quadrature"><code class="xref py py-class docutils literal notranslate"><span class="pre">Quadrature</span></code></a> formula (or let the module choose one based on the function space degree), and call <a class="reference internal" href="#warp.fem.integrate" title="warp.fem.integrate"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrate()</span></code></a> with the linear form integrand.
-The result is a <a class="reference internal" href="runtime.html#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a> containing the integration result for each of the function space degrees of freedom.</p></li>
-<li><p>Integrate bilinear forms to build the system’s left-hand-side. Define a trial function over the function space using <a class="reference internal" href="#warp.fem.make_trial" title="warp.fem.make_trial"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_trial()</span></code></a>,
-then call <a class="reference internal" href="#warp.fem.integrate" title="warp.fem.integrate"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrate()</span></code></a> with the bilinear form integrand.
-The result is a <a class="reference internal" href="sparse.html#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.sparse.BsrMatrix</span></code></a> containing the integration result for each pair of test and trial function space degrees of freedom.
-Note that the trial and test functions do not have to be defined over the same function space, so that Mixed FEM is supported.</p></li>
-<li><p>Solve the resulting linear system using the solver of your choice</p></li>
-</ul>
-</div></blockquote>
-<p>The following excerpt from the introductory example <code class="docutils literal notranslate"><span class="pre">examples/fem/example_diffusion.py</span></code> outlines this procedure:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># Grid geometry</span>
-<span class="n">geo</span> <span class="o">=</span> <span class="n">Grid2D</span><span class="p">(</span><span class="n">n</span><span class="o">=</span><span class="mi">50</span><span class="p">,</span> <span class="n">cell_size</span><span class="o">=</span><span class="mi">2</span><span class="p">)</span>
-
-<span class="c1"># Domain and function spaces</span>
-<span class="n">domain</span> <span class="o">=</span> <span class="n">Cells</span><span class="p">(</span><span class="n">geometry</span><span class="o">=</span><span class="n">geo</span><span class="p">)</span>
-<span class="n">scalar_space</span> <span class="o">=</span> <span class="n">make_polynomial_space</span><span class="p">(</span><span class="n">geo</span><span class="p">,</span> <span class="n">degree</span><span class="o">=</span><span class="mi">3</span><span class="p">)</span>
-
-<span class="c1"># Right-hand-side (forcing term)</span>
-<span class="n">test</span> <span class="o">=</span> <span class="n">make_test</span><span class="p">(</span><span class="n">space</span><span class="o">=</span><span class="n">scalar_space</span><span class="p">,</span> <span class="n">domain</span><span class="o">=</span><span class="n">domain</span><span class="p">)</span>
-<span class="n">rhs</span> <span class="o">=</span> <span class="n">integrate</span><span class="p">(</span><span class="n">linear_form</span><span class="p">,</span> <span class="n">fields</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;v&quot;</span><span class="p">:</span> <span class="n">test</span><span class="p">})</span>
-
-<span class="c1"># Weakly-imposed boundary conditions on Y sides</span>
-<span class="n">boundary</span> <span class="o">=</span> <span class="n">BoundarySides</span><span class="p">(</span><span class="n">geo</span><span class="p">)</span>
-<span class="n">bd_test</span> <span class="o">=</span> <span class="n">make_test</span><span class="p">(</span><span class="n">space</span><span class="o">=</span><span class="n">scalar_space</span><span class="p">,</span> <span class="n">domain</span><span class="o">=</span><span class="n">boundary</span><span class="p">)</span>
-<span class="n">bd_trial</span> <span class="o">=</span> <span class="n">make_trial</span><span class="p">(</span><span class="n">space</span><span class="o">=</span><span class="n">scalar_space</span><span class="p">,</span> <span class="n">domain</span><span class="o">=</span><span class="n">boundary</span><span class="p">)</span>
-<span class="n">bd_matrix</span> <span class="o">=</span> <span class="n">integrate</span><span class="p">(</span><span class="n">y_mass_form</span><span class="p">,</span> <span class="n">fields</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;u&quot;</span><span class="p">:</span> <span class="n">bd_trial</span><span class="p">,</span> <span class="s2">&quot;v&quot;</span><span class="p">:</span> <span class="n">bd_test</span><span class="p">})</span>
-
-<span class="c1"># Diffusion form</span>
-<span class="n">trial</span> <span class="o">=</span> <span class="n">make_trial</span><span class="p">(</span><span class="n">space</span><span class="o">=</span><span class="n">scalar_space</span><span class="p">,</span> <span class="n">domain</span><span class="o">=</span><span class="n">domain</span><span class="p">)</span>
-<span class="n">matrix</span> <span class="o">=</span> <span class="n">integrate</span><span class="p">(</span><span class="n">diffusion_form</span><span class="p">,</span> <span class="n">fields</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;u&quot;</span><span class="p">:</span> <span class="n">trial</span><span class="p">,</span> <span class="s2">&quot;v&quot;</span><span class="p">:</span> <span class="n">test</span><span class="p">},</span> <span class="n">values</span><span class="o">=</span><span class="p">{</span><span class="s2">&quot;nu&quot;</span><span class="p">:</span> <span class="n">viscosity</span><span class="p">})</span>
-
-<span class="c1"># Assemble linear system (add diffusion and boundary condition matrices)</span>
-<span class="n">bsr_axpy</span><span class="p">(</span><span class="n">x</span><span class="o">=</span><span class="n">bd_matrix</span><span class="p">,</span> <span class="n">y</span><span class="o">=</span><span class="n">matrix</span><span class="p">,</span> <span class="n">alpha</span><span class="o">=</span><span class="n">boundary_strength</span><span class="p">,</span> <span class="n">beta</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
-
-<span class="c1"># Solve linear system using Conjugate Gradient</span>
-<span class="n">x</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros_like</span><span class="p">(</span><span class="n">rhs</span><span class="p">)</span>
-<span class="n">bsr_cg</span><span class="p">(</span><span class="n">matrix</span><span class="p">,</span> <span class="n">b</span><span class="o">=</span><span class="n">rhs</span><span class="p">,</span> <span class="n">x</span><span class="o">=</span><span class="n">x</span><span class="p">)</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The <a class="reference internal" href="#warp.fem.integrate" title="warp.fem.integrate"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrate()</span></code></a> function does not check that the passed integrands are actually linear or bilinear forms; it is up to the user to ensure that they are.
-To solve non-linear PDEs, one can use an iterative procedure and pass the current value of the studied function <a class="reference internal" href="#warp.fem.DiscreteField" title="warp.fem.DiscreteField"><code class="xref py py-class docutils literal notranslate"><span class="pre">DiscreteField</span></code></a> argument to the integrand, on which
-arbitrary operations are permitted. However, the result of the form must remain linear in the test and trial fields.</p>
-</div>
-</section>
-<section id="introductory-examples">
-<h2>Introductory examples<a class="headerlink" href="#introductory-examples" title="Link to this heading">#</a></h2>
-<p><code class="docutils literal notranslate"><span class="pre">warp.fem</span></code> ships with a list of examples in the <code class="docutils literal notranslate"><span class="pre">examples/fem</span></code> directory illustrating common model problems.</p>
-<blockquote>
-<div><ul class="simple">
-<li><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">example_diffusion.py</span></code>: 2D diffusion with homogenous Neumann and Dirichlet boundary conditions</dt><dd><ul>
-<li><p><code class="docutils literal notranslate"><span class="pre">example_diffusion_3d.py</span></code>: 3D variant of the diffusion problem</p></li>
-</ul>
-</dd>
-</dl>
-</li>
-<li><dl class="simple">
-<dt><code class="docutils literal notranslate"><span class="pre">example_convection_diffusion.py</span></code>: 2D convection-diffusion using semi-Lagrangian advection</dt><dd><ul>
-<li><p><code class="docutils literal notranslate"><span class="pre">example_diffusion_dg0.py</span></code>: 2D convection-diffusion using finite-volume and upwind transport</p></li>
-<li><p><code class="docutils literal notranslate"><span class="pre">example_diffusion_dg.py</span></code>: 2D convection-diffusion using Discontinuous Galerkin with upwind transport and Symmetric Interior Penalty</p></li>
-</ul>
-</dd>
-</dl>
-</li>
-<li><p><code class="docutils literal notranslate"><span class="pre">example_stokes.py</span></code>: 2D incompressible Stokes flow using mixed <span class="math notranslate nohighlight">\(P_k/P_{k-1}\)</span> or <span class="math notranslate nohighlight">\(Q_k/P_{(k-1)d}\)</span> elements</p></li>
-<li><p><code class="docutils literal notranslate"><span class="pre">example_navier_stokes.py</span></code>: 2D Navier-Stokes flow using mixed <span class="math notranslate nohighlight">\(P_k/P_{k-1}\)</span> elements</p></li>
-<li><p><code class="docutils literal notranslate"><span class="pre">example_mixed_elasticity.py</span></code>: 2D linear elasticity using mixed continuous/discontinuous <span class="math notranslate nohighlight">\(S_k/P_{(k-1)d}\)</span> elements</p></li>
-</ul>
-</div></blockquote>
-</section>
-<section id="advanced-usages">
-<h2>Advanced usages<a class="headerlink" href="#advanced-usages" title="Link to this heading">#</a></h2>
-<section id="high-order-curved-geometries">
-<h3>High-order (curved) geometries<a class="headerlink" href="#high-order-curved-geometries" title="Link to this heading">#</a></h3>
-<p>It is possible to convert any <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a> (grids and explicit meshes) into a curved, high-order variant by deforming them
-with an arbitrary-order displacement field using the <a class="reference internal" href="#warp.fem.DiscreteField.make_deformed_geometry" title="warp.fem.DiscreteField.make_deformed_geometry"><code class="xref py py-meth docutils literal notranslate"><span class="pre">make_deformed_geometry()</span></code></a> method.
-The process looks as follow:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># Define a base geometry</span>
-<span class="n">base_geo</span> <span class="o">=</span> <span class="n">fem</span><span class="o">.</span><span class="n">Grid3D</span><span class="p">(</span><span class="n">res</span><span class="o">=</span><span class="n">resolution</span><span class="p">)</span>
-
-<span class="c1"># Define a displacement field on the base geometry</span>
-<span class="n">deformation_space</span> <span class="o">=</span> <span class="n">fem</span><span class="o">.</span><span class="n">make_polynomial_space</span><span class="p">(</span><span class="n">base_geo</span><span class="p">,</span> <span class="n">degree</span><span class="o">=</span><span class="n">deformation_degree</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-<span class="n">deformation_field</span> <span class="o">=</span> <span class="n">deformation_space</span><span class="o">.</span><span class="n">make_field</span><span class="p">()</span>
-
-<span class="c1"># Populate the field value by interpolating an expression</span>
-<span class="n">fem</span><span class="o">.</span><span class="n">interpolate</span><span class="p">(</span><span class="n">deformation_field_expr</span><span class="p">,</span> <span class="n">dest</span><span class="o">=</span><span class="n">deformation_field</span><span class="p">)</span>
-
-<span class="c1"># Construct the deformed geometry from the displacement field</span>
-<span class="n">deform_geo</span> <span class="o">=</span> <span class="n">deformation_field</span><span class="o">.</span><span class="n">make_deformed_geometry</span><span class="p">()</span>
-
-<span class="c1"># Define new function spaces on the deformed geometry</span>
-<span class="n">scalar_space</span> <span class="o">=</span> <span class="n">fem</span><span class="o">.</span><span class="n">make_polynomial_space</span><span class="p">(</span><span class="n">deformed_geo</span><span class="p">,</span> <span class="n">degree</span><span class="o">=</span><span class="n">scalar_space_degree</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>See also <code class="docutils literal notranslate"><span class="pre">example_deformed_geometry.py</span></code> for a complete example.</p>
-</section>
-<section id="particle-based-quadrature">
-<h3>Particle-based quadrature<a class="headerlink" href="#particle-based-quadrature" title="Link to this heading">#</a></h3>
-<p>The <a class="reference internal" href="#warp.fem.PicQuadrature" title="warp.fem.PicQuadrature"><code class="xref py py-class docutils literal notranslate"><span class="pre">PicQuadrature</span></code></a> provides a way to define Particle-In-Cell quadratures from a set or arbitrary particles,
-which can be helpful to develop MPM-like methods.
-The particles are automatically bucketed to the geometry cells when the quadrature is initialized.
-This is illustrated by the <code class="docutils literal notranslate"><span class="pre">example_stokes_transfer.py</span></code> and <code class="docutils literal notranslate"><span class="pre">example_apic_fluid.py</span></code> examples.</p>
-</section>
-<section id="partitioning">
-<h3>Partitioning<a class="headerlink" href="#partitioning" title="Link to this heading">#</a></h3>
-<p>The FEM toolkit makes it possible to perform integration on a subset of the domain elements,
-possibly re-indexing degrees of freedom so that the linear system contains the local ones only.
-This is useful for distributed computation (see <code class="docutils literal notranslate"><span class="pre">examples/fem/example_diffusion_mgpu.py</span></code>), or simply to limit the simulation domain to a subset of active cells (see <code class="docutils literal notranslate"><span class="pre">examples/fem/example_stokes_transfer.py</span></code>).</p>
-<p>A partition of the simulation geometry can be defined using subclasses of <a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.GeometryPartition"><code class="xref py py-class docutils literal notranslate"><span class="pre">GeometryPartition</span></code></a>
-such as <a class="reference internal" href="#warp.fem.LinearGeometryPartition" title="warp.fem.LinearGeometryPartition"><code class="xref py py-class docutils literal notranslate"><span class="pre">LinearGeometryPartition</span></code></a>  or <a class="reference internal" href="#warp.fem.ExplicitGeometryPartition" title="warp.fem.ExplicitGeometryPartition"><code class="xref py py-class docutils literal notranslate"><span class="pre">ExplicitGeometryPartition</span></code></a>.</p>
-<p>Function spaces can then be partitioned according to the geometry partition using <a class="reference internal" href="#warp.fem.make_space_partition" title="warp.fem.make_space_partition"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_space_partition()</span></code></a>.
-The resulting <a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.SpacePartition"><code class="xref py py-class docutils literal notranslate"><span class="pre">SpacePartition</span></code></a> object allows translating between space-wide and partition-wide node indices,
-and differentiating interior, frontier and exterior nodes.</p>
-</section>
-<section id="memory-management">
-<h3>Memory management<a class="headerlink" href="#memory-management" title="Link to this heading">#</a></h3>
-<p>Several <code class="docutils literal notranslate"><span class="pre">warp.fem</span></code> functions require allocating temporary buffers to perform their computations.
-If such functions are called many times in a tight loop, those many allocations and de-allocations may degrade performance.
-To overcome this issue, a <a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><code class="xref py py-class docutils literal notranslate"><span class="pre">cache.TemporaryStore</span></code></a> object may be created to persist and re-use temporary allocations across calls,
-either globally using <a class="reference internal" href="#warp.fem.set_default_temporary_store" title="warp.fem.set_default_temporary_store"><code class="xref py py-func docutils literal notranslate"><span class="pre">set_default_temporary_store()</span></code></a> or at a per-function granularity using the corresponding argument.</p>
-</section>
-</section>
-<section id="operators">
-<span id="id1"></span><h2>Operators<a class="headerlink" href="#operators" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.position">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">position</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><span class="pre">Domain</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.position" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the world position of the sample point <cite>s</cite></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.normal">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">normal</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><span class="pre">Domain</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.normal" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the element normal at the sample point <cite>s</cite>. Null for interior points.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.lookup">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">lookup</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><span class="pre">Domain</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.lookup" title="Link to this definition">#</a></dt>
-<dd><p>Looks-up the sample point corresponding to a world position <cite>x</cite>, projecting to the closest point on the domain.</p>
-<dl class="simple">
-<dt>Arg:</dt><dd><p>x: world position of the point to look-up in the geometry
-guess: (optional) <a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><code class="xref py py-class docutils literal notranslate"><span class="pre">Sample</span></code></a> initial guess, may help perform the query</p>
-</dd>
-</dl>
-<p class="rubric">Notes</p>
-<p>Currently this operator is only fully supported for <a class="reference internal" href="#warp.fem.Grid2D" title="warp.fem.Grid2D"><code class="xref py py-class docutils literal notranslate"><span class="pre">Grid2D</span></code></a> and <a class="reference internal" href="#warp.fem.Grid3D" title="warp.fem.Grid3D"><code class="xref py py-class docutils literal notranslate"><span class="pre">Grid3D</span></code></a> geometries.
-For <code class="xref py py-class docutils literal notranslate"><span class="pre">TriangleMesh2D</span></code> and <a class="reference internal" href="#warp.fem.Tetmesh" title="warp.fem.Tetmesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Tetmesh</span></code></a> geometries, the operator requires providing <cite>guess</cite>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.measure">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">measure</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><span class="pre">Domain</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.measure" title="Link to this definition">#</a></dt>
-<dd><p>Returns the measure (volume, area, or length) determinant of an element at a sample point <cite>s</cite></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.measure_ratio">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">measure_ratio</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><span class="pre">Domain</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.measure_ratio" title="Link to this definition">#</a></dt>
-<dd><p>Returns the maximum ratio between the measure of this element and that of higher-dimensional neighbours.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.deformation_gradient">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">deformation_gradient</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><span class="pre">Domain</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.deformation_gradient" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the gradient of the domain position with respect to the element reference space at the sample point <cite>s</cite></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.degree">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">degree</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.degree" title="Link to this definition">#</a></dt>
-<dd><p>Polynomial degree of a field</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.inner">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">inner</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.inner" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the field at a sample point <cite>s</cite>. On oriented sides, uses the inner element</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.outer">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">outer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.outer" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the field at a sample point <cite>s</cite>. On oriented sides, uses the outer element. On interior points and on domain boundaries, this is equivalent to <a class="reference internal" href="#warp.fem.inner" title="warp.fem.inner"><code class="xref py py-func docutils literal notranslate"><span class="pre">inner()</span></code></a>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.grad">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">grad</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.grad" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the field gradient at a sample point <cite>s</cite>. On oriented sides, uses the inner element</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.grad_outer">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">grad_outer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.grad_outer" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the field gradient at a sample point <cite>s</cite>. On oriented sides, uses the outer element. On interior points and on domain boundaries, this is equivalent to <a class="reference internal" href="#warp.fem.grad" title="warp.fem.grad"><code class="xref py py-func docutils literal notranslate"><span class="pre">grad()</span></code></a>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.div">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.div" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the field divergence at a sample point <cite>s</cite>. On oriented sides, uses the inner element</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.div_outer">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">div_outer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.div_outer" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the field divergence at a sample point <cite>s</cite>. On oriented sides, uses the outer element. On interior points and on domain boundaries, this is equivalent to <a class="reference internal" href="#warp.fem.div" title="warp.fem.div"><code class="xref py py-func docutils literal notranslate"><span class="pre">div()</span></code></a>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.at_node">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">at_node</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.at_node" title="Link to this definition">#</a></dt>
-<dd><p>For a Test or Trial field, returns a copy of the Sample <cite>s</cite> moved to the coordinates of the node being evaluated</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.D">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">D</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.D" title="Link to this definition">#</a></dt>
-<dd><p>Symmetric part of the (inner) gradient of the field at <cite>s</cite></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.curl">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">curl</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.curl" title="Link to this definition">#</a></dt>
-<dd><p>Skew part of the (inner) gradient of the field at <cite>s</cite>, as a vector such that <code class="docutils literal notranslate"><span class="pre">wp.cross(curl(u),</span> <span class="pre">v)</span> <span class="pre">=</span> <span class="pre">skew(grad(u))</span> <span class="pre">v</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.jump">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">jump</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.jump" title="Link to this definition">#</a></dt>
-<dd><p>Jump between inner and outer element values on an interior side. Zero for interior points or domain boundaries</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.average">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">average</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.average" title="Link to this definition">#</a></dt>
-<dd><p>Average between inner and outer element values</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.grad_jump">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">grad_jump</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.grad_jump" title="Link to this definition">#</a></dt>
-<dd><p>Jump between inner and outer element gradients on an interior side. Zero for interior points or domain boundaries</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.grad_average">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">grad_average</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">f</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Field" title="warp.fem.Field"><span class="pre">Field</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.fem.Sample" title="warp.fem.Sample"><span class="pre">Sample</span></a></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.grad_average" title="Link to this definition">#</a></dt>
-<dd><p>Average between inner and outer element gradients</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.operator.operator">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.operator.</span></span><span class="sig-name descname"><span class="pre">operator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">resolver</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.operator.operator" title="Link to this definition">#</a></dt>
-<dd><p>Decorator for functions operating on Field-like or Domain-like data inside warp.fem integrands</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>resolver</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Callable" title="(in Python v3.12)"><em>Callable</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="integration">
-<h2>Integration<a class="headerlink" href="#integration" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.integrate">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">integrate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">integrand</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">quadrature=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodal=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fields={}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">values={}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">accumulate_dtype=&lt;class</span> <span class="pre">'warp.types.float64'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output_dtype=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">output=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kernel_options={}</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.integrate" title="Link to this definition">#</a></dt>
-<dd><p>Integrates a constant, linear or bilinear form, and returns a scalar, array, or sparse matrix, respectively.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>integrand</strong> (<em>Integrand</em>) – Form to be integrated, must have <a class="reference internal" href="#warp.fem.integrand" title="warp.fem.integrand"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrand()</span></code></a> decorator</p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a><em> | </em><em>None</em>) – Integration domain. If None, deduced from fields</p></li>
-<li><p><strong>quadrature</strong> (<a class="reference internal" href="#warp.fem.Quadrature" title="warp.fem.quadrature.quadrature.Quadrature"><em>Quadrature</em></a><em> | </em><em>None</em>) – Quadrature formula. If None, deduced from domain and fields degree.</p></li>
-<li><p><strong>nodal</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – For linear or bilinear form only, use the test function nodes as the quadrature points. Assumes Lagrange interpolation functions are used, and no differential or DG operator is evaluated on the test or trial functions.</p></li>
-<li><p><strong>fields</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Dict" title="(in Python v3.12)"><em>Dict</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>, </em><a class="reference internal" href="#warp.fem.FieldLike" title="warp.fem.field.field.FieldLike"><em>FieldLike</em></a><em>]</em>) – Discrete, test, and trial fields to be passed to the integrand. Keys in the dictionary must match integrand parameter names.</p></li>
-<li><p><strong>values</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Dict" title="(in Python v3.12)"><em>Dict</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>, </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>]</em>) – Additional variable values to be passed to the integrand, can by of any type accepted by warp kernel launchs. Keys in the dictionary must match integrand parameter names.</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – shared pool from which to allocate temporary arrays</p></li>
-<li><p><strong>accumulate_dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a>) – Scalar type to be used for accumulating integration samples</p></li>
-<li><p><strong>output</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a><em> | </em><a class="reference internal" href="sparse.html#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em> | </em><em>None</em>) – Sparse matrix or warp array into which to store the result of the integration</p></li>
-<li><p><strong>output_dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a><em> | </em><em>None</em>) – Scalar type for returned results in <cite>output</cite> is notr provided. If None, defaults to <cite>accumulate_dtype</cite></p></li>
-<li><p><strong>device</strong> – Device on which to perform the integration</p></li>
-<li><p><strong>kernel_options</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Dict" title="(in Python v3.12)"><em>Dict</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>, </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>]</em>) – Overloaded options to be passed to the kernel builder (e.g, <code class="docutils literal notranslate"><span class="pre">{&quot;enable_backward&quot;:</span> <span class="pre">True}</span></code>)</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.interpolate">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">interpolate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">integrand</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dest</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fields</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">values</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">{}</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.interpolate" title="Link to this definition">#</a></dt>
-<dd><p>Interpolates a function and assigns the result to a discrete field.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>integrand</strong> (<em>Integrand</em>) – Function to be interpolated, must have <a class="reference internal" href="#warp.fem.integrand" title="warp.fem.integrand"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrand()</span></code></a> decorator</p></li>
-<li><p><strong>dest</strong> (<a class="reference internal" href="#warp.fem.DiscreteField" title="warp.fem.field.field.DiscreteField"><em>DiscreteField</em></a><em> | </em><a class="reference internal" href="#warp.fem.field.FieldRestriction" title="warp.fem.field.restriction.FieldRestriction"><em>FieldRestriction</em></a>) – Discrete field, or restriction of a discrete field to a domain, to which the interpolation result will be assigned</p></li>
-<li><p><strong>fields</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Dict" title="(in Python v3.12)"><em>Dict</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>, </em><a class="reference internal" href="#warp.fem.FieldLike" title="warp.fem.field.field.FieldLike"><em>FieldLike</em></a><em>]</em>) – Discrete fields to be passed to the integrand. Keys in the dictionary must match integrand parameters names.</p></li>
-<li><p><strong>values</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Dict" title="(in Python v3.12)"><em>Dict</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>, </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>]</em>) – Additional variable values to be passed to the integrand, can by of any type accepted by warp kernel launchs. Keys in the dictionary must match integrand parameter names.</p></li>
-<li><p><strong>device</strong> – Device on which to perform the interpolation</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.integrand">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">integrand</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">func</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.integrand" title="Link to this definition">#</a></dt>
-<dd><p>Decorator for functions to be integrated (or interpolated) using warp.fem</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>func</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Callable" title="(in Python v3.12)"><em>Callable</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Sample">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Sample</span></span><a class="headerlink" href="#warp.fem.Sample" title="Link to this definition">#</a></dt>
-<dd><p>Per-sample point context for evaluating fields and related operators in integrands.</p>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Field">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Field</span></span><a class="headerlink" href="#warp.fem.Field" title="Link to this definition">#</a></dt>
-<dd><p>Tag for field-like integrand arguments</p>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Domain">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Domain</span></span><a class="headerlink" href="#warp.fem.Domain" title="Link to this definition">#</a></dt>
-<dd><p>Tag for domain-like integrand arguments</p>
-</dd></dl>
-
-</section>
-<section id="geometry">
-<h2>Geometry<a class="headerlink" href="#geometry" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Grid2D">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Grid2D</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">res</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bounds_lo=&lt;warp.types.vec2f</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bounds_hi=&lt;warp.types.vec2f</span> <span class="pre">object&gt;</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Grid2D" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a></p>
-<p>Two-dimensional regular grid geometry</p>
-<p>Constructs a dense 2D grid</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>res</strong> (<em>vec2i</em>) – Resolution of the grid along each dimension</p></li>
-<li><p><strong>bounds_lo</strong> (<em>vec2f</em>) – Position of the lower bound of the axis-aligned grid</p></li>
-<li><p><strong>bounds_up</strong> – Position of the upper bound of the axis-aligned grid</p></li>
-<li><p><strong>bounds_hi</strong> (<em>vec2f</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Trimesh2D">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Trimesh2D</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tri_vertex_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positions</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Trimesh2D" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a></p>
-<p>Two-dimensional triangular mesh geometry</p>
-<p>Constructs a two-dimensional triangular mesh.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>tri_vertex_indices</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_tris, 3) containing vertex indices for each tri</p></li>
-<li><p><strong>positions</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_vertices, 2) containing 2d position for each vertex</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – shared pool from which to allocate temporary arrays</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Quadmesh2D">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Quadmesh2D</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">quad_vertex_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positions</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Quadmesh2D" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a></p>
-<p>Two-dimensional quadrilateral mesh geometry</p>
-<p>Constructs a two-dimensional quadrilateral mesh.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>quad_vertex_indices</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_tris, 4) containing vertex indices for each quad, in counter-clockwise order</p></li>
-<li><p><strong>positions</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_vertices, 2) containing 2d position for each vertex</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – shared pool from which to allocate temporary arrays</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Grid3D">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Grid3D</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">res</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bounds_lo=&lt;warp.types.vec3f</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bounds_hi=&lt;warp.types.vec3f</span> <span class="pre">object&gt;</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Grid3D" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a></p>
-<p>Three-dimensional regular grid geometry</p>
-<p>Constructs a dense 3D grid</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>res</strong> (<em>vec3i</em>) – Resolution of the grid along each dimension</p></li>
-<li><p><strong>bounds_lo</strong> (<em>vec3f</em>) – Position of the lower bound of the axis-aligned grid</p></li>
-<li><p><strong>bounds_up</strong> – Position of the upper bound of the axis-aligned grid</p></li>
-<li><p><strong>bounds_hi</strong> (<em>vec3f</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Tetmesh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Tetmesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tet_vertex_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positions</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Tetmesh" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a></p>
-<p>Tetrahedral mesh geometry</p>
-<p>Constructs a tetrahedral mesh.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>tet_vertex_indices</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_tets, 4) containing vertex indices for each tet</p></li>
-<li><p><strong>positions</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_vertices, 3) containing 3d position for each vertex</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – shared pool from which to allocate temporary arrays</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Hexmesh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Hexmesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">hex_vertex_indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positions</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Hexmesh" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><code class="xref py py-class docutils literal notranslate"><span class="pre">Geometry</span></code></a></p>
-<p>Hexahedral mesh geometry</p>
-<p>Constructs a tetrahedral mesh.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>hex_vertex_indices</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_hexes, 8) containing vertex indices for each hex
-following standard ordering (bottom face vertices in counter-clockwise order, then similarly for upper face)</p></li>
-<li><p><strong>positions</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – warp array of shape (num_vertices, 3) containing 3d position for each vertex</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – shared pool from which to allocate temporary arrays</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.LinearGeometryPartition">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">LinearGeometryPartition</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">partition_rank</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">partition_count</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.LinearGeometryPartition" title="Link to this definition">#</a></dt>
-<dd><p>Creates a geometry partition by uniformly partionning cell indices</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a>) – the geometry to partition</p></li>
-<li><p><strong>partition_rank</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – the index of the partition being created</p></li>
-<li><p><strong>partition_count</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – the number of partitions that will be created over the geometry</p></li>
-<li><p><strong>device</strong> – Warp device on which to perform and store computations</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.ExplicitGeometryPartition">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">ExplicitGeometryPartition</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cell_mask</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.ExplicitGeometryPartition" title="Link to this definition">#</a></dt>
-<dd><p>Creates a geometry partition by uniformly partionning cell indices</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a>) – the geometry to partition</p></li>
-<li><p><strong>cell_mask</strong> (<em>wp.array</em><em>(</em><em>dtype=int</em><em>)</em>) – warp array of length <code class="docutils literal notranslate"><span class="pre">geometry.cell_count()</span></code> indicating which cells are selected. Array values must be either <code class="docutils literal notranslate"><span class="pre">1</span></code> (selected) or <code class="docutils literal notranslate"><span class="pre">0</span></code> (not selected).</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Cells">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Cells</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Cells" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><code class="xref py py-class docutils literal notranslate"><span class="pre">GeometryDomain</span></code></a></p>
-<p>A Domain containing all cells of the geometry or geometry partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a><em> | </em><a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Sides">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Sides</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Sides" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><code class="xref py py-class docutils literal notranslate"><span class="pre">GeometryDomain</span></code></a></p>
-<p>A Domain containing all (interior and boundary) sides of the geometry or geometry partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a><em> | </em><a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.BoundarySides">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">BoundarySides</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.BoundarySides" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Sides" title="warp.fem.domain.Sides"><code class="xref py py-class docutils literal notranslate"><span class="pre">Sides</span></code></a></p>
-<p>A Domain containing boundary sides of the geometry or geometry partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a><em> | </em><a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.FrontierSides">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">FrontierSides</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.FrontierSides" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Sides" title="warp.fem.domain.Sides"><code class="xref py py-class docutils literal notranslate"><span class="pre">Sides</span></code></a></p>
-<p>A Domain containing frontier sides of the geometry partition (sides shared with at least another partition)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a><em> | </em><a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Polynomial">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Polynomial</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Polynomial" title="Link to this definition">#</a></dt>
-<dd><p>Polynomial family defining interpolation nodes over an interval</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.Polynomial.GAUSS_LEGENDRE">
-<span class="sig-name descname"><span class="pre">GAUSS_LEGENDRE</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">0</span></em><a class="headerlink" href="#warp.fem.Polynomial.GAUSS_LEGENDRE" title="Link to this definition">#</a></dt>
-<dd><p>Gauss–Legendre 1D polynomial family (does not include endpoints)</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.Polynomial.LOBATTO_GAUSS_LEGENDRE">
-<span class="sig-name descname"><span class="pre">LOBATTO_GAUSS_LEGENDRE</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">1</span></em><a class="headerlink" href="#warp.fem.Polynomial.LOBATTO_GAUSS_LEGENDRE" title="Link to this definition">#</a></dt>
-<dd><p>Lobatto–Gauss–Legendre 1D polynomial family (includes endpoints)</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.Polynomial.EQUISPACED_CLOSED">
-<span class="sig-name descname"><span class="pre">EQUISPACED_CLOSED</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">2</span></em><a class="headerlink" href="#warp.fem.Polynomial.EQUISPACED_CLOSED" title="Link to this definition">#</a></dt>
-<dd><p>Closed 1D polynomial family with uniformly distributed nodes (includes endpoints)</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.Polynomial.EQUISPACED_OPEN">
-<span class="sig-name descname"><span class="pre">EQUISPACED_OPEN</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">3</span></em><a class="headerlink" href="#warp.fem.Polynomial.EQUISPACED_OPEN" title="Link to this definition">#</a></dt>
-<dd><p>Open 1D polynomial family with uniformly distributed nodes (does not include endpoints)</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.RegularQuadrature">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">RegularQuadrature</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">order</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">family</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.RegularQuadrature" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Quadrature" title="warp.fem.quadrature.quadrature.Quadrature"><code class="xref py py-class docutils literal notranslate"><span class="pre">Quadrature</span></code></a></p>
-<p>Regular quadrature formula, using a constant set of quadrature points per element</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a>) – </p></li>
-<li><p><strong>order</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – </p></li>
-<li><p><strong>family</strong> (<a class="reference internal" href="#warp.fem.Polynomial" title="warp.fem.polynomial.Polynomial"><em>Polynomial</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.NodalQuadrature">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">NodalQuadrature</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.NodalQuadrature" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Quadrature" title="warp.fem.quadrature.quadrature.Quadrature"><code class="xref py py-class docutils literal notranslate"><span class="pre">Quadrature</span></code></a></p>
-<p>Quadrature using space node points as quadrature points</p>
-<p>Note that in constant to the <cite>nodal=True</cite> flag for <a class="reference internal" href="#warp.fem.integrate" title="warp.fem.integrate"><code class="xref py py-func docutils literal notranslate"><span class="pre">integrate()</span></code></a>, this quadrature odes not make any assumption
-about orthogonality of shape functions, and is thus safe to use for arbitrary integrands.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a>) – </p></li>
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.PicQuadrature">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">PicQuadrature</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positions</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">measures</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.PicQuadrature" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.Quadrature" title="warp.fem.quadrature.quadrature.Quadrature"><code class="xref py py-class docutils literal notranslate"><span class="pre">Quadrature</span></code></a></p>
-<p>Particle-based quadrature formula, using a global set of points irregularely spread out over geometry elements.</p>
-<p>Useful for Particle-In-Cell and derived methods.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a>) – </p></li>
-<li><p><strong>positions</strong> (<em>wp.array</em><em>(</em><em>)</em>) – </p></li>
-<li><p><strong>measures</strong> (<em>wp.array</em><em>(</em><em>dtype=float</em><em>)</em>) – </p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="function-spaces">
-<h2>Function Spaces<a class="headerlink" href="#function-spaces" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_polynomial_space">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_polynomial_space</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geo</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype=&lt;class</span> <span class="pre">'float'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dof_mapper=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">degree=1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">element_basis=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">discontinuous=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">family=None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_polynomial_space" title="Link to this definition">#</a></dt>
-<dd><p>Equips a geometry with a collocated, polynomial function space.
-Equivalent to successive calls to <a class="reference internal" href="#warp.fem.make_polynomial_basis_space" title="warp.fem.make_polynomial_basis_space"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_polynomial_basis_space()</span></code></a> and <cite>make_collocated_function_space</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>geo</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a>) – the Geometry on which to build the space</p></li>
-<li><p><strong>dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a>) – value type the function space. If <code class="docutils literal notranslate"><span class="pre">dof_mapper</span></code> is provided, the value type from the DofMapper will be used instead.</p></li>
-<li><p><strong>dof_mapper</strong> (<a class="reference internal" href="#warp.fem.DofMapper" title="warp.fem.space.dof_mapper.DofMapper"><em>DofMapper</em></a><em> | </em><em>None</em>) – mapping from node degrees of freedom to function values, defaults to Identity. Useful for reduced coordinates, e.g. <a class="reference internal" href="#warp.fem.SymmetricTensorMapper" title="warp.fem.SymmetricTensorMapper"><code class="xref py py-class docutils literal notranslate"><span class="pre">SymmetricTensorMapper</span></code></a> maps 2x2 (resp 3x3) symmetric tensors to 3 (resp 6) degrees of freedom.</p></li>
-<li><p><strong>degree</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – polynomial degree of the per-element shape functions</p></li>
-<li><p><strong>discontinuous</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – if True, use Discontinuous Galerkin shape functions. Discontinuous is implied if degree is 0, i.e, piecewise-constant shape functions.</p></li>
-<li><p><strong>element_basis</strong> (<a class="reference internal" href="#warp.fem.ElementBasis" title="warp.fem.space.ElementBasis"><em>ElementBasis</em></a><em> | </em><em>None</em>) – type of basis function for the individual elements</p></li>
-<li><p><strong>family</strong> (<a class="reference internal" href="#warp.fem.Polynomial" title="warp.fem.polynomial.Polynomial"><em>Polynomial</em></a><em> | </em><em>None</em>) – Polynomial family used to generate the shape function basis. If not provided, a reasonable basis is chosen.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the constructed function space</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><em>CollocatedFunctionSpace</em></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_polynomial_basis_space">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_polynomial_basis_space</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geo</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">degree</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">element_basis</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">discontinuous</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">family</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_polynomial_basis_space" title="Link to this definition">#</a></dt>
-<dd><p>Equips a geometry with a polynomial basis.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>geo</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a>) – the Geometry on which to build the space</p></li>
-<li><p><strong>degree</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – polynomial degree of the per-element shape functions</p></li>
-<li><p><strong>discontinuous</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – if True, use Discontinuous Galerkin shape functions. Discontinuous is implied if degree is 0, i.e, piecewise-constant shape functions.</p></li>
-<li><p><strong>element_basis</strong> (<a class="reference internal" href="#warp.fem.ElementBasis" title="warp.fem.space.ElementBasis"><em>ElementBasis</em></a><em> | </em><em>None</em>) – type of basis function for the individual elements</p></li>
-<li><p><strong>family</strong> (<a class="reference internal" href="#warp.fem.Polynomial" title="warp.fem.polynomial.Polynomial"><em>Polynomial</em></a><em> | </em><em>None</em>) – Polynomial family used to generate the shape function basis. If not provided, a reasonable basis is chosen.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the constructed basis space</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.BasisSpace" title="warp.fem.space.basis_space.BasisSpace"><em>BasisSpace</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_collocated_function_space">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_collocated_function_space</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">basis_space</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype=&lt;class</span> <span class="pre">'float'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dof_mapper=None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_collocated_function_space" title="Link to this definition">#</a></dt>
-<dd><p>Constructs a function space from a basis space and a value type, such that all degrees of freedom of the value type are stored at each of the basis nodes.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>geo</strong> – the Geometry on which to build the space</p></li>
-<li><p><strong>dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a>) – value type the function space. If <code class="docutils literal notranslate"><span class="pre">dof_mapper</span></code> is provided, the value type from the DofMapper will be used instead.</p></li>
-<li><p><strong>dof_mapper</strong> (<a class="reference internal" href="#warp.fem.DofMapper" title="warp.fem.space.dof_mapper.DofMapper"><em>DofMapper</em></a><em> | </em><em>None</em>) – mapping from node degrees of freedom to function values, defaults to Identity. Useful for reduced coordinates, e.g. <a class="reference internal" href="#warp.fem.SymmetricTensorMapper" title="warp.fem.SymmetricTensorMapper"><code class="xref py py-class docutils literal notranslate"><span class="pre">SymmetricTensorMapper</span></code></a> maps 2x2 (resp 3x3) symmetric tensors to 3 (resp 6) degrees of freedom.</p></li>
-<li><p><strong>basis_space</strong> (<a class="reference internal" href="#warp.fem.BasisSpace" title="warp.fem.space.basis_space.BasisSpace"><em>BasisSpace</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the constructed function space</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><em>CollocatedFunctionSpace</em></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_space_partition">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_space_partition</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">geometry_partition</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_topology</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">with_halo</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_space_partition" title="Link to this definition">#</a></dt>
-<dd><p>Computes the subset of nodes from a function space topology that touch a geometry partition</p>
-<p>Either <cite>space_topology</cite> or <cite>space</cite> must be provided (and will be considered in that order).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a><em> | </em><em>None</em>) – (deprecated) the function space defining the topology if <cite>space_topology</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>.</p></li>
-<li><p><strong>geometry_partition</strong> (<a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a><em> | </em><em>None</em>) – The subset of the space geometry.  If not provided, use the whole geometry.</p></li>
-<li><p><strong>space_topology</strong> (<a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><em>SpaceTopology</em></a><em> | </em><em>None</em>) – the topology of the function space to consider. If <code class="docutils literal notranslate"><span class="pre">None</span></code>, deduced from <cite>space</cite>.</p></li>
-<li><p><strong>with_halo</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – if True, include the halo nodes (nodes from exterior frontier cells to the partition)</p></li>
-<li><p><strong>device</strong> – Warp device on which to perform and store computations</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the resulting space partition</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_space_restriction">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_space_restriction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_partition</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_topology</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_space_restriction" title="Link to this definition">#</a></dt>
-<dd><p>Restricts a function space partition to a Domain, i.e. a subset of its elements.</p>
-<p>One of <cite>space_partition</cite>, <cite>space_topology</cite>, or <cite>space</cite> must be provided (and will be considered in that order).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a><em> | </em><em>None</em>) – (deprecated) if neither <cite>space_partition</cite> nor <cite>space_topology</cite> are provided, the space defining the topology to restrict</p></li>
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a><em> | </em><em>None</em>) – the subset of nodes from the space topology to consider</p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a><em> | </em><em>None</em>) – the domain to restrict the space to, defaults to all cells of the space geometry or partition.</p></li>
-<li><p><strong>space_topology</strong> (<a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><em>SpaceTopology</em></a><em> | </em><em>None</em>) – the space topology to be restricted, if <cite>space_partition</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>.</p></li>
-<li><p><strong>device</strong> – device on which to perform and store computations</p></li>
-<li><p><strong>temporary_store</strong> (<em>Optional</em><em>[</em><a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>warp.fem.cache.TemporaryStore</em></a><em>]</em>) – shared pool from which to allocate temporary arrays</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.fem.SpaceRestriction" title="warp.fem.space.restriction.SpaceRestriction"><em>SpaceRestriction</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.ElementBasis">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">ElementBasis</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.ElementBasis" title="Link to this definition">#</a></dt>
-<dd><p>Choice of basis function to equip individual elements</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.ElementBasis.LAGRANGE">
-<span class="sig-name descname"><span class="pre">LAGRANGE</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">0</span></em><a class="headerlink" href="#warp.fem.ElementBasis.LAGRANGE" title="Link to this definition">#</a></dt>
-<dd><p>Lagrange basis functions <span class="math notranslate nohighlight">\(P_k\)</span> for simplices, tensor products <span class="math notranslate nohighlight">\(Q_k\)</span> for squares and cubes</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.ElementBasis.SERENDIPITY">
-<span class="sig-name descname"><span class="pre">SERENDIPITY</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">1</span></em><a class="headerlink" href="#warp.fem.ElementBasis.SERENDIPITY" title="Link to this definition">#</a></dt>
-<dd><p>Serendipity elements <span class="math notranslate nohighlight">\(S_k\)</span>, corresponding to Lagrange nodes with interior points removed (for degree &lt;= 3)</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.ElementBasis.NONCONFORMING_POLYNOMIAL">
-<span class="sig-name descname"><span class="pre">NONCONFORMING_POLYNOMIAL</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">2</span></em><a class="headerlink" href="#warp.fem.ElementBasis.NONCONFORMING_POLYNOMIAL" title="Link to this definition">#</a></dt>
-<dd><p>Simplex Lagrange basis functions <span class="math notranslate nohighlight">\(P_{kd}\)</span> embedded into non conforming reference elements (e.g. squares or cubes). Discontinuous only.</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.SymmetricTensorMapper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">SymmetricTensorMapper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dtype</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mapping</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">Mapping.VOIGT</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SymmetricTensorMapper" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.DofMapper" title="warp.fem.space.dof_mapper.DofMapper"><code class="xref py py-class docutils literal notranslate"><span class="pre">DofMapper</span></code></a></p>
-<p>Orthonormal isomorphism from R^{n (n+1)} to nxn symmetric tensors,
-using usual L2 norm for vectors and half Frobenius norm, (tau : tau)/2 for tensors.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a>) – </p></li>
-<li><p><strong>mapping</strong> (<em>Mapping</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.SkewSymmetricTensorMapper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">SkewSymmetricTensorMapper</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SkewSymmetricTensorMapper" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.DofMapper" title="warp.fem.space.dof_mapper.DofMapper"><code class="xref py py-class docutils literal notranslate"><span class="pre">DofMapper</span></code></a></p>
-<p>Orthonormal isomorphism from R^{n (n-1)} to nxn skew-symmetric tensors,
-using usual L2 norm for vectors and half Frobenius norm, (tau : tau)/2 for tensors.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="fields">
-<h2>Fields<a class="headerlink" href="#fields" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_test">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_test</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_restriction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_partition</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_test" title="Link to this definition">#</a></dt>
-<dd><p>Constructs a test field over a function space or its restriction</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – the function space</p></li>
-<li><p><strong>space_restriction</strong> (<a class="reference internal" href="#warp.fem.SpaceRestriction" title="warp.fem.space.restriction.SpaceRestriction"><em>SpaceRestriction</em></a><em> | </em><em>None</em>) – restriction of the space topology to a domain</p></li>
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a><em> | </em><em>None</em>) – if <cite>space_restriction</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>, the optional subset of node indices to consider</p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a><em> | </em><em>None</em>) – if <cite>space_restriction</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>, optional subset of elements to consider</p></li>
-<li><p><strong>device</strong> – Warp device on which to perform and store computations</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the test field</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.field.TestField" title="warp.fem.field.test.TestField"><em>TestField</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_trial">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_trial</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_restriction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_partition</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_trial" title="Link to this definition">#</a></dt>
-<dd><p>Constructs a trial field over a function space or partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – the function space or function space restriction</p></li>
-<li><p><strong>space_restriction</strong> (<a class="reference internal" href="#warp.fem.SpaceRestriction" title="warp.fem.space.restriction.SpaceRestriction"><em>SpaceRestriction</em></a><em> | </em><em>None</em>) – restriction of the space topology to a domain</p></li>
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a><em> | </em><em>None</em>) – if <cite>space_restriction</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>, the optional subset of node indices to consider</p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a><em> | </em><em>None</em>) – if <cite>space_restriction</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>, optional subset of elements to consider</p></li>
-<li><p><strong>device</strong> – Warp device on which to perform and store computations</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the trial field</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.field.TrialField" title="warp.fem.field.trial.TrialField"><em>TrialField</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.make_restriction">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">make_restriction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">field</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_restriction</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.make_restriction" title="Link to this definition">#</a></dt>
-<dd><p>Restricts a discrete field to a subset of elements.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>field</strong> (<a class="reference internal" href="#warp.fem.DiscreteField" title="warp.fem.field.field.DiscreteField"><em>DiscreteField</em></a>) – the discrete field to restrict</p></li>
-<li><p><strong>space_restriction</strong> (<a class="reference internal" href="#warp.fem.SpaceRestriction" title="warp.fem.space.restriction.SpaceRestriction"><em>SpaceRestriction</em></a><em> | </em><em>None</em>) – the function space restriction defining the subset of elements to consider</p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a><em> | </em><em>None</em>) – if <code class="docutils literal notranslate"><span class="pre">space_restriction</span></code> is not provided, the <a class="reference internal" href="#warp.fem.Domain" title="warp.fem.Domain"><code class="xref py py-class docutils literal notranslate"><span class="pre">Domain</span></code></a> defining the subset of elements to consider</p></li>
-<li><p><strong>device</strong> – Warp device on which to perform and store computations</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>the field restriction</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.field.FieldRestriction" title="warp.fem.field.restriction.FieldRestriction"><em>FieldRestriction</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="boundary-conditions">
-<h2>Boundary Conditions<a class="headerlink" href="#boundary-conditions" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.normalize_dirichlet_projector">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">normalize_dirichlet_projector</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">projector_matrix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fixed_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.normalize_dirichlet_projector" title="Link to this definition">#</a></dt>
-<dd><p>Scale projector so that it becomes idempotent, and apply the same scaling to fixed_value if provided</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>projector_matrix</strong> (<a class="reference internal" href="sparse.html#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – </p></li>
-<li><p><strong>fixed_value</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a><em> | </em><em>None</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.project_linear_system">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">project_linear_system</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">system_matrix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">system_rhs</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">projector_matrix</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fixed_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">normalize_projector</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.project_linear_system" title="Link to this definition">#</a></dt>
-<dd><p>Projects both the left-hand-side and right-hand-side of a linear system to enforce Dirichlet boundary conditions</p>
-<p>If normalize_projector is True, first apply scaling so that the projector_matrix is idempotent</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>system_matrix</strong> (<a class="reference internal" href="sparse.html#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – </p></li>
-<li><p><strong>system_rhs</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – </p></li>
-<li><p><strong>projector_matrix</strong> (<a class="reference internal" href="sparse.html#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – </p></li>
-<li><p><strong>fixed_value</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a><em> | </em><em>None</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="id2">
-<h2>Memory management<a class="headerlink" href="#id2" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.set_default_temporary_store">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">set_default_temporary_store</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.set_default_temporary_store" title="Link to this definition">#</a></dt>
-<dd><p>Globally sets the default <a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.TemporaryStore"><code class="xref py py-class docutils literal notranslate"><span class="pre">TemporaryStore</span></code></a> instance to use for temporary allocations in <code class="docutils literal notranslate"><span class="pre">warp.fem</span></code> functions.</p>
-<p>If the default temporary store is set to <code class="docutils literal notranslate"><span class="pre">None</span></code>, temporary allocations are not persisted unless a <a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.TemporaryStore"><code class="xref py py-class docutils literal notranslate"><span class="pre">TemporaryStore</span></code></a> is provided at a per-function granularity.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.borrow_temporary">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">borrow_temporary</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.borrow_temporary" title="Link to this definition">#</a></dt>
-<dd><p>Borrows and returns a temporary array with specified attributes from a shared pool.</p>
-<p>If an array with sufficient capacity and matching desired attributes is already available in the pool, it will be returned.
-Otherwise, a new allocation wil be performed.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – the shared pool to borrow the temporary from. If <cite>temporary_store</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>, the global default temporary store, if set, will be used.</p></li>
-<li><p><strong>shape</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – desired dimensions for the temporary array</p></li>
-<li><p><strong>dtype</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><em>type</em></a>) – desired data type for the temporary array</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – whether a pinned allocation is desired</p></li>
-<li><p><strong>device</strong> – device on which the momory should be allocated; if <code class="docutils literal notranslate"><span class="pre">None</span></code>, the current device will be used.</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.fem.cache.Temporary" title="warp.fem.cache.Temporary"><em>Temporary</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.fem.borrow_temporary_like">
-<span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">borrow_temporary_like</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.borrow_temporary_like" title="Link to this definition">#</a></dt>
-<dd><p>Borrows and returns a temporary array with the same attributes as another array or temporary.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>array</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a><em> | </em><a class="reference internal" href="#warp.fem.cache.Temporary" title="warp.fem.cache.Temporary"><em>Temporary</em></a>) – Warp or temporary array to read the desired attributes from</p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a><em> | </em><em>None</em>) – the shared pool to borrow the temporary from. If <cite>temporary_store</cite> is <code class="docutils literal notranslate"><span class="pre">None</span></code>, the global default temporary store, if set, will be used.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.fem.cache.Temporary" title="warp.fem.cache.Temporary"><em>Temporary</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="interfaces">
-<h2>Interfaces<a class="headerlink" href="#interfaces" title="Link to this heading">#</a></h2>
-<p>Interface classes are not meant to be constructed directly, but can be derived from extend the built-in functionality.</p>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Geometry">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Geometry</span></span><a class="headerlink" href="#warp.fem.Geometry" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for discrete geometries</p>
-<p>A geometry is composed of cells and sides. Sides may be boundary or interior (between cells).</p>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.Geometry.cell_count">
-<span class="sig-name descname"><span class="pre">cell_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Geometry.cell_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of cells in the geometry</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.Geometry.side_count">
-<span class="sig-name descname"><span class="pre">side_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Geometry.side_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of sides in the geometry</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.Geometry.boundary_side_count">
-<span class="sig-name descname"><span class="pre">boundary_side_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Geometry.boundary_side_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of boundary sides (sides with a single neighbour cell) in the geometry</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.GeometryPartition">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">GeometryPartition</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryPartition" title="Link to this definition">#</a></dt>
-<dd><p>Base class for geometry partitions, i.e. subset of cells and sides</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a>) – </p>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.GeometryPartition.cell_count">
-<span class="sig-name descname"><span class="pre">cell_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryPartition.cell_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of cells that are ‘owned’ by this partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.GeometryPartition.side_count">
-<span class="sig-name descname"><span class="pre">side_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryPartition.side_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of sides that are ‘owned’ by this partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.GeometryPartition.boundary_side_count">
-<span class="sig-name descname"><span class="pre">boundary_side_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryPartition.boundary_side_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of geo-boundary sides that are ‘owned’ by this partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.GeometryPartition.frontier_side_count">
-<span class="sig-name descname"><span class="pre">frontier_side_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryPartition.frontier_side_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of sides with neighbors owned by this and another partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.GeometryDomain">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">GeometryDomain</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryDomain" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for domains, i.e. (partial) views of elements in a Geometry</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a><em> | </em><a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a>) – </p>
-</dd>
-</dl>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.GeometryDomain.ElementKind">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">ElementKind</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryDomain.ElementKind" title="Link to this definition">#</a></dt>
-<dd><p>Possible kinds of elements contained in a domain</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.GeometryDomain.element_kind">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">element_kind</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.GeometryDomain.ElementKind" title="warp.fem.domain.GeometryDomain.ElementKind"><span class="pre">ElementKind</span></a></em><a class="headerlink" href="#warp.fem.GeometryDomain.element_kind" title="Link to this definition">#</a></dt>
-<dd><p>Kind of elements that this domain contains (cells or sides)</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.GeometryDomain.dimension">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">dimension</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></em><a class="headerlink" href="#warp.fem.GeometryDomain.dimension" title="Link to this definition">#</a></dt>
-<dd><p>Dimension of the elements of the domain</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.GeometryDomain.element_count">
-<span class="sig-name descname"><span class="pre">element_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.GeometryDomain.element_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of elements in the domain</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.Quadrature">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">Quadrature</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">domain</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Quadrature" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for quadrature rules</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a>) – </p>
-</dd>
-</dl>
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.Quadrature.domain">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">domain</span></span><a class="headerlink" href="#warp.fem.Quadrature.domain" title="Link to this definition">#</a></dt>
-<dd><p>Domain over which this quadrature is defined</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.Quadrature.total_point_count">
-<span class="sig-name descname"><span class="pre">total_point_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.Quadrature.total_point_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of quadrature points over the domain</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">FunctionSpace</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">topology</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.FunctionSpace" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for function spaces, i.e. geometry + interpolation basis</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>topology</strong> (<a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><em>SpaceTopology</em></a>) – </p>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.dtype">
-<span class="sig-name descname"><span class="pre">dtype</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#type" title="(in Python v3.12)"><span class="pre">type</span></a></em><a class="headerlink" href="#warp.fem.FunctionSpace.dtype" title="Link to this definition">#</a></dt>
-<dd><p>Value type of the interpolation functions</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.topology">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">topology</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><span class="pre">SpaceTopology</span></a></em><a class="headerlink" href="#warp.fem.FunctionSpace.topology" title="Link to this definition">#</a></dt>
-<dd><p>Underlying geometry</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.geometry">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">geometry</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><span class="pre">Geometry</span></a></em><a class="headerlink" href="#warp.fem.FunctionSpace.geometry" title="Link to this definition">#</a></dt>
-<dd><p>Underlying geometry</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.dimension">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">dimension</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></em><a class="headerlink" href="#warp.fem.FunctionSpace.dimension" title="Link to this definition">#</a></dt>
-<dd><p>Function space embedding dimension</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.degree">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">degree</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></em><a class="headerlink" href="#warp.fem.FunctionSpace.degree" title="Link to this definition">#</a></dt>
-<dd><p>Maximum polynomial degree of the underlying basis</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.trace">
-<span class="sig-name descname"><span class="pre">trace</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.FunctionSpace.trace" title="Link to this definition">#</a></dt>
-<dd><p>Trace of the function space over lower-dimensional elements of the geometry</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.FunctionSpace.make_field">
-<span class="sig-name descname"><span class="pre">make_field</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space_partition</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.FunctionSpace.make_field" title="Link to this definition">#</a></dt>
-<dd><p>Creates a zero-initialized discrete field over the function space holding values for all degrees of freedom of nodes in a space partition</p>
-<dl class="simple">
-<dt>space_arg:</dt><dd><p>space_partition: If provided, the subset of nodes to consider</p>
-</dd>
-</dl>
-<p>See also: <a class="reference internal" href="#warp.fem.make_space_partition" title="warp.fem.make_space_partition"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_space_partition()</span></code></a></p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.SpaceTopology">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">SpaceTopology</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">geometry</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">nodes_per_element</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpaceTopology" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for defining the topology of a function space.</p>
-<p>The topology only considers the indices of the nodes in each element, and as such,
-the connectivity pattern of the function space.
-It does not specify the actual location of the nodes within the elements, or the valuation function.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>geometry</strong> (<a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a>) – </p></li>
-<li><p><strong>nodes_per_element</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.fem.SpaceTopology.dimension">
-<span class="sig-name descname"><span class="pre">dimension</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></em><a class="headerlink" href="#warp.fem.SpaceTopology.dimension" title="Link to this definition">#</a></dt>
-<dd><p>Embedding dimension of the function space</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.SpaceTopology.geometry">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">geometry</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><span class="pre">Geometry</span></a></em><a class="headerlink" href="#warp.fem.SpaceTopology.geometry" title="Link to this definition">#</a></dt>
-<dd><p>Underlying geometry</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpaceTopology.node_count">
-<span class="sig-name descname"><span class="pre">node_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpaceTopology.node_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of nodes in the interpolation basis</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpaceTopology.element_node_indices">
-<span class="sig-name descname"><span class="pre">element_node_indices</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">out</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpaceTopology.element_node_indices" title="Link to this definition">#</a></dt>
-<dd><p>Returns a temporary array containing the global index for each node of each element</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>out</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a><em> | </em><em>None</em>) – </p>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpaceTopology.trace">
-<span class="sig-name descname"><span class="pre">trace</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpaceTopology.trace" title="Link to this definition">#</a></dt>
-<dd><p>Trace of the function space over lower-dimensional elements of the geometry</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><em>TraceSpaceTopology</em></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.BasisSpace">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">BasisSpace</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">topology</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.BasisSpace" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for defining a scalar-valued basis over a geometry.</p>
-<p>A basis space makes it easy to define multiple function spaces sharing the same basis (and thus nodes) but with different valuation functions;
-however, it is not a required ingredient of a function space.</p>
-<p>See also: <a class="reference internal" href="#warp.fem.make_polynomial_basis_space" title="warp.fem.make_polynomial_basis_space"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_polynomial_basis_space()</span></code></a>, <a class="reference internal" href="#warp.fem.make_collocated_function_space" title="warp.fem.make_collocated_function_space"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_collocated_function_space()</span></code></a></p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>topology</strong> (<a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><em>SpaceTopology</em></a>) – </p></li>
-<li><p><strong>shape</strong> (<a class="reference internal" href="#warp.fem.space.shape.ShapeFunction" title="warp.fem.space.shape.shape_function.ShapeFunction"><em>ShapeFunction</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.BasisSpace.topology">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">topology</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><span class="pre">SpaceTopology</span></a></em><a class="headerlink" href="#warp.fem.BasisSpace.topology" title="Link to this definition">#</a></dt>
-<dd><p>Underlying topology of the basis space</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.BasisSpace.geometry">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">geometry</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><span class="pre">Geometry</span></a></em><a class="headerlink" href="#warp.fem.BasisSpace.geometry" title="Link to this definition">#</a></dt>
-<dd><p>Underlying geometry of the basis space</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.BasisSpace.shape">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">shape</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="#warp.fem.space.shape.ShapeFunction" title="warp.fem.space.shape.shape_function.ShapeFunction"><span class="pre">ShapeFunction</span></a></em><a class="headerlink" href="#warp.fem.BasisSpace.shape" title="Link to this definition">#</a></dt>
-<dd><p>Shape functions used for defining individual element basis</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.BasisSpace.node_positions">
-<span class="sig-name descname"><span class="pre">node_positions</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">out</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.BasisSpace.node_positions" title="Link to this definition">#</a></dt>
-<dd><p>Returns a temporary array containing the world position for each node</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>out</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a><em> | </em><em>None</em>) – </p>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.space.shape.ShapeFunction">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.space.shape.</span></span><span class="sig-name descname"><span class="pre">ShapeFunction</span></span><a class="headerlink" href="#warp.fem.space.shape.ShapeFunction" title="Link to this definition">#</a></dt>
-<dd><p>Interface class for defining scalar-valued shape functions over a single element</p>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.SpacePartition">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">SpacePartition</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space_topology</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">geo_partition</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpacePartition" title="Link to this definition">#</a></dt>
-<dd><dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space_topology</strong> (<a class="reference internal" href="#warp.fem.SpaceTopology" title="warp.fem.space.topology.SpaceTopology"><em>SpaceTopology</em></a>) – </p></li>
-<li><p><strong>geo_partition</strong> (<a class="reference internal" href="#warp.fem.GeometryPartition" title="warp.fem.geometry.partition.GeometryPartition"><em>GeometryPartition</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpacePartition.node_count">
-<span class="sig-name descname"><span class="pre">node_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpacePartition.node_count" title="Link to this definition">#</a></dt>
-<dd><p>Returns number of nodes in this partition</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpacePartition.owned_node_count">
-<span class="sig-name descname"><span class="pre">owned_node_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpacePartition.owned_node_count" title="Link to this definition">#</a></dt>
-<dd><p>Returns number of nodes in this partition, excluding exterior halo</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpacePartition.interior_node_count">
-<span class="sig-name descname"><span class="pre">interior_node_count</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpacePartition.interior_node_count" title="Link to this definition">#</a></dt>
-<dd><p>Returns number of interior nodes in this partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.SpacePartition.space_node_indices">
-<span class="sig-name descname"><span class="pre">space_node_indices</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpacePartition.space_node_indices" title="Link to this definition">#</a></dt>
-<dd><p>Return the global function space indices for nodes in this partition</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.SpaceRestriction">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">SpaceRestriction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space_partition</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">temporary_store</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.SpaceRestriction" title="Link to this definition">#</a></dt>
-<dd><p>Restriction of a space partition to a given GeometryDomain</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a>) – </p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a>) – </p></li>
-<li><p><strong>temporary_store</strong> (<a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><em>TemporaryStore</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.DofMapper">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">DofMapper</span></span><a class="headerlink" href="#warp.fem.DofMapper" title="Link to this definition">#</a></dt>
-<dd><p>Base class from mapping node degrees of freedom to function values</p>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.FieldLike">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">FieldLike</span></span><a class="headerlink" href="#warp.fem.FieldLike" title="Link to this definition">#</a></dt>
-<dd><p>Base class for integrable fields</p>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.DiscreteField">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">DiscreteField</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_partition</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.DiscreteField" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.field.SpaceField" title="warp.fem.field.field.SpaceField"><code class="xref py py-class docutils literal notranslate"><span class="pre">SpaceField</span></code></a></p>
-<p>Explicitly-valued field defined over a partition of a discrete function space</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – </p></li>
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.DiscreteField.dof_values">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">dof_values</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><span class="pre">array</span></a></em><a class="headerlink" href="#warp.fem.DiscreteField.dof_values" title="Link to this definition">#</a></dt>
-<dd><p>Array of degrees of freedom values</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.DiscreteField.trace">
-<span class="sig-name descname"><span class="pre">trace</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.DiscreteField.trace" title="Link to this definition">#</a></dt>
-<dd><p>Trace of this field over a lower-dimensional function space</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.DiscreteField" title="warp.fem.field.field.DiscreteField"><em>DiscreteField</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.DiscreteField.make_deformed_geometry">
-<span class="sig-name descname"><span class="pre">make_deformed_geometry</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.DiscreteField.make_deformed_geometry" title="Link to this definition">#</a></dt>
-<dd><p>Returns a deformed version of the underlying geometry using this field’s values as displacement</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.fem.Geometry" title="warp.fem.geometry.geometry.Geometry"><em>Geometry</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.field.FieldRestriction">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.field.</span></span><span class="sig-name descname"><span class="pre">FieldRestriction</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space_restriction</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">field</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.field.FieldRestriction" title="Link to this definition">#</a></dt>
-<dd><p>Restriction of a discrete field to a given GeometryDomain</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space_restriction</strong> (<a class="reference internal" href="#warp.fem.SpaceRestriction" title="warp.fem.space.restriction.SpaceRestriction"><em>SpaceRestriction</em></a>) – </p></li>
-<li><p><strong>field</strong> (<a class="reference internal" href="#warp.fem.DiscreteField" title="warp.fem.field.field.DiscreteField"><em>DiscreteField</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.field.SpaceField">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.field.</span></span><span class="sig-name descname"><span class="pre">SpaceField</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_partition</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.field.SpaceField" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.FieldLike" title="warp.fem.field.field.FieldLike"><code class="xref py py-class docutils literal notranslate"><span class="pre">FieldLike</span></code></a></p>
-<p>Base class for fields defined over a function space</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – </p></li>
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.field.TestField">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.field.</span></span><span class="sig-name descname"><span class="pre">TestField</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space_restriction</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.field.TestField" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.field.SpaceField" title="warp.fem.field.field.SpaceField"><code class="xref py py-class docutils literal notranslate"><span class="pre">SpaceField</span></code></a></p>
-<p>Field defined over a space restriction that can be used as a test function.</p>
-<p>In order to reuse computations, it is possible to define the test field using a SpaceRestriction
-defined for a different value type than the test function value type, as long as the node topology is similar.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space_restriction</strong> (<a class="reference internal" href="#warp.fem.SpaceRestriction" title="warp.fem.space.restriction.SpaceRestriction"><em>SpaceRestriction</em></a>) – </p></li>
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.field.TrialField">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.field.</span></span><span class="sig-name descname"><span class="pre">TrialField</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">space</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">space_partition</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">domain</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.field.TrialField" title="Link to this definition">#</a></dt>
-<dd><p>Bases: <a class="reference internal" href="#warp.fem.field.SpaceField" title="warp.fem.field.field.SpaceField"><code class="xref py py-class docutils literal notranslate"><span class="pre">SpaceField</span></code></a></p>
-<p>Field defined over a domain that can be used as a trial function</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>space</strong> (<a class="reference internal" href="#warp.fem.FunctionSpace" title="warp.fem.space.function_space.FunctionSpace"><em>FunctionSpace</em></a>) – </p></li>
-<li><p><strong>space_partition</strong> (<a class="reference internal" href="#warp.fem.SpacePartition" title="warp.fem.space.partition.SpacePartition"><em>SpacePartition</em></a>) – </p></li>
-<li><p><strong>domain</strong> (<a class="reference internal" href="#warp.fem.GeometryDomain" title="warp.fem.domain.GeometryDomain"><em>GeometryDomain</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.TemporaryStore">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.</span></span><span class="sig-name descname"><span class="pre">TemporaryStore</span></span><a class="headerlink" href="#warp.fem.TemporaryStore" title="Link to this definition">#</a></dt>
-<dd><p>Shared pool of temporary arrays that will be persisted and reused across invocations of <code class="docutils literal notranslate"><span class="pre">warp.fem</span></code> functions.</p>
-<p>A <a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.TemporaryStore"><code class="xref py py-class docutils literal notranslate"><span class="pre">TemporaryStore</span></code></a> instance may either be passed explicitly to <code class="docutils literal notranslate"><span class="pre">warp.fem</span></code> functions that accept such an argument, for instance <code class="xref py py-func docutils literal notranslate"><span class="pre">integrate.integrate()</span></code>,
-or can be set globally as the default store using <a class="reference internal" href="#warp.fem.set_default_temporary_store" title="warp.fem.set_default_temporary_store"><code class="xref py py-func docutils literal notranslate"><span class="pre">set_default_temporary_store()</span></code></a>.</p>
-<p>By default, there is no default temporary store, so that temporary allocations are not persisted.</p>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.fem.cache.Temporary">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.fem.cache.</span></span><span class="sig-name descname"><span class="pre">Temporary</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pool</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.cache.Temporary" title="Link to this definition">#</a></dt>
-<dd><p>Handle over a temporary array from a <a class="reference internal" href="#warp.fem.TemporaryStore" title="warp.fem.cache.TemporaryStore"><code class="xref py py-class docutils literal notranslate"><span class="pre">TemporaryStore</span></code></a>.</p>
-<p>The array will be automatically returned to the temporary pool for re-use upon destruction of this object, unless
-the temporary is explicitly detached from the pool using <a class="reference internal" href="#warp.fem.cache.Temporary.detach" title="warp.fem.cache.Temporary.detach"><code class="xref py py-meth docutils literal notranslate"><span class="pre">detach()</span></code></a>.
-The temporary may also be explicitly returned to the pool before destruction using <a class="reference internal" href="#warp.fem.cache.Temporary.release" title="warp.fem.cache.Temporary.release"><code class="xref py py-meth docutils literal notranslate"><span class="pre">release()</span></code></a>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>array</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – </p></li>
-<li><p><strong>pool</strong> (<em>TemporaryStore.Pool</em><em> | </em><em>None</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.cache.Temporary.detach">
-<span class="sig-name descname"><span class="pre">detach</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.cache.Temporary.detach" title="Link to this definition">#</a></dt>
-<dd><p>Detaches the temporary so it is never returned to the pool</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.fem.cache.Temporary.release">
-<span class="sig-name descname"><span class="pre">release</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.fem.cache.Temporary.release" title="Link to this definition">#</a></dt>
-<dd><p>Returns the temporary array to the pool</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.fem.cache.Temporary.array">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">array</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><span class="pre">array</span></a></em><a class="headerlink" href="#warp.fem.cache.Temporary.array" title="Link to this definition">#</a></dt>
-<dd><p>View of the array with desired shape and data type.</p>
-</dd></dl>
-
-</dd></dl>
-
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          
-          <a class="prev-page" href="sparse.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">warp.sparse</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">warp.fem</a><ul>
-<li><a class="reference internal" href="#integrands">Integrands</a></li>
-<li><a class="reference internal" href="#basic-workflow">Basic workflow</a></li>
-<li><a class="reference internal" href="#introductory-examples">Introductory examples</a></li>
-<li><a class="reference internal" href="#advanced-usages">Advanced usages</a><ul>
-<li><a class="reference internal" href="#high-order-curved-geometries">High-order (curved) geometries</a></li>
-<li><a class="reference internal" href="#particle-based-quadrature">Particle-based quadrature</a></li>
-<li><a class="reference internal" href="#partitioning">Partitioning</a></li>
-<li><a class="reference internal" href="#memory-management">Memory management</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#operators">Operators</a><ul>
-<li><a class="reference internal" href="#warp.fem.position"><code class="docutils literal notranslate"><span class="pre">position()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.normal"><code class="docutils literal notranslate"><span class="pre">normal()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.lookup"><code class="docutils literal notranslate"><span class="pre">lookup()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.measure"><code class="docutils literal notranslate"><span class="pre">measure()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.measure_ratio"><code class="docutils literal notranslate"><span class="pre">measure_ratio()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.deformation_gradient"><code class="docutils literal notranslate"><span class="pre">deformation_gradient()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.degree"><code class="docutils literal notranslate"><span class="pre">degree()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.inner"><code class="docutils literal notranslate"><span class="pre">inner()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.outer"><code class="docutils literal notranslate"><span class="pre">outer()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.grad"><code class="docutils literal notranslate"><span class="pre">grad()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.grad_outer"><code class="docutils literal notranslate"><span class="pre">grad_outer()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.div"><code class="docutils literal notranslate"><span class="pre">div()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.div_outer"><code class="docutils literal notranslate"><span class="pre">div_outer()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.at_node"><code class="docutils literal notranslate"><span class="pre">at_node()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.D"><code class="docutils literal notranslate"><span class="pre">D()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.curl"><code class="docutils literal notranslate"><span class="pre">curl()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.jump"><code class="docutils literal notranslate"><span class="pre">jump()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.average"><code class="docutils literal notranslate"><span class="pre">average()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.grad_jump"><code class="docutils literal notranslate"><span class="pre">grad_jump()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.grad_average"><code class="docutils literal notranslate"><span class="pre">grad_average()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.operator.operator"><code class="docutils literal notranslate"><span class="pre">operator()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#integration">Integration</a><ul>
-<li><a class="reference internal" href="#warp.fem.integrate"><code class="docutils literal notranslate"><span class="pre">integrate()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.interpolate"><code class="docutils literal notranslate"><span class="pre">interpolate()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.integrand"><code class="docutils literal notranslate"><span class="pre">integrand()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Sample"><code class="docutils literal notranslate"><span class="pre">Sample</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Field"><code class="docutils literal notranslate"><span class="pre">Field</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Domain"><code class="docutils literal notranslate"><span class="pre">Domain</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#geometry">Geometry</a><ul>
-<li><a class="reference internal" href="#warp.fem.Grid2D"><code class="docutils literal notranslate"><span class="pre">Grid2D</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Trimesh2D"><code class="docutils literal notranslate"><span class="pre">Trimesh2D</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Quadmesh2D"><code class="docutils literal notranslate"><span class="pre">Quadmesh2D</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Grid3D"><code class="docutils literal notranslate"><span class="pre">Grid3D</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Tetmesh"><code class="docutils literal notranslate"><span class="pre">Tetmesh</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Hexmesh"><code class="docutils literal notranslate"><span class="pre">Hexmesh</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.LinearGeometryPartition"><code class="docutils literal notranslate"><span class="pre">LinearGeometryPartition</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.ExplicitGeometryPartition"><code class="docutils literal notranslate"><span class="pre">ExplicitGeometryPartition</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Cells"><code class="docutils literal notranslate"><span class="pre">Cells</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Sides"><code class="docutils literal notranslate"><span class="pre">Sides</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.BoundarySides"><code class="docutils literal notranslate"><span class="pre">BoundarySides</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FrontierSides"><code class="docutils literal notranslate"><span class="pre">FrontierSides</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Polynomial"><code class="docutils literal notranslate"><span class="pre">Polynomial</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.Polynomial.GAUSS_LEGENDRE"><code class="docutils literal notranslate"><span class="pre">Polynomial.GAUSS_LEGENDRE</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Polynomial.LOBATTO_GAUSS_LEGENDRE"><code class="docutils literal notranslate"><span class="pre">Polynomial.LOBATTO_GAUSS_LEGENDRE</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Polynomial.EQUISPACED_CLOSED"><code class="docutils literal notranslate"><span class="pre">Polynomial.EQUISPACED_CLOSED</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Polynomial.EQUISPACED_OPEN"><code class="docutils literal notranslate"><span class="pre">Polynomial.EQUISPACED_OPEN</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.RegularQuadrature"><code class="docutils literal notranslate"><span class="pre">RegularQuadrature</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.NodalQuadrature"><code class="docutils literal notranslate"><span class="pre">NodalQuadrature</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.PicQuadrature"><code class="docutils literal notranslate"><span class="pre">PicQuadrature</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#function-spaces">Function Spaces</a><ul>
-<li><a class="reference internal" href="#warp.fem.make_polynomial_space"><code class="docutils literal notranslate"><span class="pre">make_polynomial_space()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.make_polynomial_basis_space"><code class="docutils literal notranslate"><span class="pre">make_polynomial_basis_space()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.make_collocated_function_space"><code class="docutils literal notranslate"><span class="pre">make_collocated_function_space()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.make_space_partition"><code class="docutils literal notranslate"><span class="pre">make_space_partition()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.make_space_restriction"><code class="docutils literal notranslate"><span class="pre">make_space_restriction()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.ElementBasis"><code class="docutils literal notranslate"><span class="pre">ElementBasis</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.ElementBasis.LAGRANGE"><code class="docutils literal notranslate"><span class="pre">ElementBasis.LAGRANGE</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.ElementBasis.SERENDIPITY"><code class="docutils literal notranslate"><span class="pre">ElementBasis.SERENDIPITY</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.ElementBasis.NONCONFORMING_POLYNOMIAL"><code class="docutils literal notranslate"><span class="pre">ElementBasis.NONCONFORMING_POLYNOMIAL</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.SymmetricTensorMapper"><code class="docutils literal notranslate"><span class="pre">SymmetricTensorMapper</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SkewSymmetricTensorMapper"><code class="docutils literal notranslate"><span class="pre">SkewSymmetricTensorMapper</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#fields">Fields</a><ul>
-<li><a class="reference internal" href="#warp.fem.make_test"><code class="docutils literal notranslate"><span class="pre">make_test()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.make_trial"><code class="docutils literal notranslate"><span class="pre">make_trial()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.make_restriction"><code class="docutils literal notranslate"><span class="pre">make_restriction()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#boundary-conditions">Boundary Conditions</a><ul>
-<li><a class="reference internal" href="#warp.fem.normalize_dirichlet_projector"><code class="docutils literal notranslate"><span class="pre">normalize_dirichlet_projector()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.project_linear_system"><code class="docutils literal notranslate"><span class="pre">project_linear_system()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#id2">Memory management</a><ul>
-<li><a class="reference internal" href="#warp.fem.set_default_temporary_store"><code class="docutils literal notranslate"><span class="pre">set_default_temporary_store()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.borrow_temporary"><code class="docutils literal notranslate"><span class="pre">borrow_temporary()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.borrow_temporary_like"><code class="docutils literal notranslate"><span class="pre">borrow_temporary_like()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#interfaces">Interfaces</a><ul>
-<li><a class="reference internal" href="#warp.fem.Geometry"><code class="docutils literal notranslate"><span class="pre">Geometry</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.Geometry.cell_count"><code class="docutils literal notranslate"><span class="pre">Geometry.cell_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Geometry.side_count"><code class="docutils literal notranslate"><span class="pre">Geometry.side_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Geometry.boundary_side_count"><code class="docutils literal notranslate"><span class="pre">Geometry.boundary_side_count()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.GeometryPartition"><code class="docutils literal notranslate"><span class="pre">GeometryPartition</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.GeometryPartition.cell_count"><code class="docutils literal notranslate"><span class="pre">GeometryPartition.cell_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.GeometryPartition.side_count"><code class="docutils literal notranslate"><span class="pre">GeometryPartition.side_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.GeometryPartition.boundary_side_count"><code class="docutils literal notranslate"><span class="pre">GeometryPartition.boundary_side_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.GeometryPartition.frontier_side_count"><code class="docutils literal notranslate"><span class="pre">GeometryPartition.frontier_side_count()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.GeometryDomain"><code class="docutils literal notranslate"><span class="pre">GeometryDomain</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.GeometryDomain.ElementKind"><code class="docutils literal notranslate"><span class="pre">GeometryDomain.ElementKind</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.GeometryDomain.element_kind"><code class="docutils literal notranslate"><span class="pre">GeometryDomain.element_kind</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.GeometryDomain.dimension"><code class="docutils literal notranslate"><span class="pre">GeometryDomain.dimension</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.GeometryDomain.element_count"><code class="docutils literal notranslate"><span class="pre">GeometryDomain.element_count()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.Quadrature"><code class="docutils literal notranslate"><span class="pre">Quadrature</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.Quadrature.domain"><code class="docutils literal notranslate"><span class="pre">Quadrature.domain</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.Quadrature.total_point_count"><code class="docutils literal notranslate"><span class="pre">Quadrature.total_point_count()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace"><code class="docutils literal notranslate"><span class="pre">FunctionSpace</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.dtype"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.dtype</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.topology"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.topology</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.geometry"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.geometry</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.dimension"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.dimension</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.degree"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.degree</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.trace"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.trace()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FunctionSpace.make_field"><code class="docutils literal notranslate"><span class="pre">FunctionSpace.make_field()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.SpaceTopology"><code class="docutils literal notranslate"><span class="pre">SpaceTopology</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.SpaceTopology.dimension"><code class="docutils literal notranslate"><span class="pre">SpaceTopology.dimension</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpaceTopology.geometry"><code class="docutils literal notranslate"><span class="pre">SpaceTopology.geometry</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpaceTopology.node_count"><code class="docutils literal notranslate"><span class="pre">SpaceTopology.node_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpaceTopology.element_node_indices"><code class="docutils literal notranslate"><span class="pre">SpaceTopology.element_node_indices()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpaceTopology.trace"><code class="docutils literal notranslate"><span class="pre">SpaceTopology.trace()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.BasisSpace"><code class="docutils literal notranslate"><span class="pre">BasisSpace</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.BasisSpace.topology"><code class="docutils literal notranslate"><span class="pre">BasisSpace.topology</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.BasisSpace.geometry"><code class="docutils literal notranslate"><span class="pre">BasisSpace.geometry</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.BasisSpace.shape"><code class="docutils literal notranslate"><span class="pre">BasisSpace.shape</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.BasisSpace.node_positions"><code class="docutils literal notranslate"><span class="pre">BasisSpace.node_positions()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.space.shape.ShapeFunction"><code class="docutils literal notranslate"><span class="pre">ShapeFunction</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpacePartition"><code class="docutils literal notranslate"><span class="pre">SpacePartition</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.SpacePartition.node_count"><code class="docutils literal notranslate"><span class="pre">SpacePartition.node_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpacePartition.owned_node_count"><code class="docutils literal notranslate"><span class="pre">SpacePartition.owned_node_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpacePartition.interior_node_count"><code class="docutils literal notranslate"><span class="pre">SpacePartition.interior_node_count()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.SpacePartition.space_node_indices"><code class="docutils literal notranslate"><span class="pre">SpacePartition.space_node_indices()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.SpaceRestriction"><code class="docutils literal notranslate"><span class="pre">SpaceRestriction</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.DofMapper"><code class="docutils literal notranslate"><span class="pre">DofMapper</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.FieldLike"><code class="docutils literal notranslate"><span class="pre">FieldLike</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.DiscreteField"><code class="docutils literal notranslate"><span class="pre">DiscreteField</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.DiscreteField.dof_values"><code class="docutils literal notranslate"><span class="pre">DiscreteField.dof_values</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.DiscreteField.trace"><code class="docutils literal notranslate"><span class="pre">DiscreteField.trace()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.DiscreteField.make_deformed_geometry"><code class="docutils literal notranslate"><span class="pre">DiscreteField.make_deformed_geometry()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.fem.field.FieldRestriction"><code class="docutils literal notranslate"><span class="pre">FieldRestriction</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.field.SpaceField"><code class="docutils literal notranslate"><span class="pre">SpaceField</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.field.TestField"><code class="docutils literal notranslate"><span class="pre">TestField</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.field.TrialField"><code class="docutils literal notranslate"><span class="pre">TrialField</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.TemporaryStore"><code class="docutils literal notranslate"><span class="pre">TemporaryStore</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.cache.Temporary"><code class="docutils literal notranslate"><span class="pre">Temporary</span></code></a><ul>
-<li><a class="reference internal" href="#warp.fem.cache.Temporary.detach"><code class="docutils literal notranslate"><span class="pre">Temporary.detach()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.cache.Temporary.release"><code class="docutils literal notranslate"><span class="pre">Temporary.release()</span></code></a></li>
-<li><a class="reference internal" href="#warp.fem.cache.Temporary.array"><code class="docutils literal notranslate"><span class="pre">Temporary.array</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    <script async="async" src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/functions.html b/docs/_build/html/modules/functions.html
deleted file mode 100644
index f7819a3bd..000000000
--- a/docs/_build/html/modules/functions.html
+++ /dev/null
@@ -1,3001 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="warp.sim" href="sim.html" /><link rel="prev" title="Runtime Reference" href="runtime.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Kernel Reference - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="kernel-reference">
-<h1>Kernel Reference<a class="headerlink" href="#kernel-reference" title="Link to this heading">#</a></h1>
-<section id="scalar-types">
-<h2>Scalar Types<a class="headerlink" href="#scalar-types" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.int8">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">int8</span></span><a class="headerlink" href="#warp.int8" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.uint8">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">uint8</span></span><a class="headerlink" href="#warp.uint8" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.int16">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">int16</span></span><a class="headerlink" href="#warp.int16" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.uint16">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">uint16</span></span><a class="headerlink" href="#warp.uint16" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.int32">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">int32</span></span><a class="headerlink" href="#warp.int32" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.uint32">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">uint32</span></span><a class="headerlink" href="#warp.uint32" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.int64">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">int64</span></span><a class="headerlink" href="#warp.int64" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.uint64">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">uint64</span></span><a class="headerlink" href="#warp.uint64" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.float16">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">float16</span></span><a class="headerlink" href="#warp.float16" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.float32">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">float32</span></span><a class="headerlink" href="#warp.float32" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.float64">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">float64</span></span><a class="headerlink" href="#warp.float64" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.bool">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bool</span></span><a class="headerlink" href="#warp.bool" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-</section>
-<section id="vector-types">
-<h2>Vector Types<a class="headerlink" href="#vector-types" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2b">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2b</span></span><a class="headerlink" href="#warp.vec2b" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2ub">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2ub</span></span><a class="headerlink" href="#warp.vec2ub" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2s">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2s</span></span><a class="headerlink" href="#warp.vec2s" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2us">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2us</span></span><a class="headerlink" href="#warp.vec2us" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2i">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2i</span></span><a class="headerlink" href="#warp.vec2i" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2ui">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2ui</span></span><a class="headerlink" href="#warp.vec2ui" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2l">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2l</span></span><a class="headerlink" href="#warp.vec2l" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2ul">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2ul</span></span><a class="headerlink" href="#warp.vec2ul" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2h">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2h</span></span><a class="headerlink" href="#warp.vec2h" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2f">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2f</span></span><a class="headerlink" href="#warp.vec2f" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec2d">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec2d</span></span><a class="headerlink" href="#warp.vec2d" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3b">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3b</span></span><a class="headerlink" href="#warp.vec3b" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3ub">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3ub</span></span><a class="headerlink" href="#warp.vec3ub" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3s">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3s</span></span><a class="headerlink" href="#warp.vec3s" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3us">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3us</span></span><a class="headerlink" href="#warp.vec3us" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3i">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3i</span></span><a class="headerlink" href="#warp.vec3i" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3ui">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3ui</span></span><a class="headerlink" href="#warp.vec3ui" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3l">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3l</span></span><a class="headerlink" href="#warp.vec3l" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3ul">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3ul</span></span><a class="headerlink" href="#warp.vec3ul" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3h">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3h</span></span><a class="headerlink" href="#warp.vec3h" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3f">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3f</span></span><a class="headerlink" href="#warp.vec3f" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec3d">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec3d</span></span><a class="headerlink" href="#warp.vec3d" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4b">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4b</span></span><a class="headerlink" href="#warp.vec4b" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4ub">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4ub</span></span><a class="headerlink" href="#warp.vec4ub" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4s">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4s</span></span><a class="headerlink" href="#warp.vec4s" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4us">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4us</span></span><a class="headerlink" href="#warp.vec4us" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4i">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4i</span></span><a class="headerlink" href="#warp.vec4i" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4ui">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4ui</span></span><a class="headerlink" href="#warp.vec4ui" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4l">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4l</span></span><a class="headerlink" href="#warp.vec4l" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4ul">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4ul</span></span><a class="headerlink" href="#warp.vec4ul" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4h">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4h</span></span><a class="headerlink" href="#warp.vec4h" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4f">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4f</span></span><a class="headerlink" href="#warp.vec4f" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.vec4d">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vec4d</span></span><a class="headerlink" href="#warp.vec4d" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat22h">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat22h</span></span><a class="headerlink" href="#warp.mat22h" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat22f">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat22f</span></span><a class="headerlink" href="#warp.mat22f" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat22d">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat22d</span></span><a class="headerlink" href="#warp.mat22d" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat33h">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat33h</span></span><a class="headerlink" href="#warp.mat33h" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat33f">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat33f</span></span><a class="headerlink" href="#warp.mat33f" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat33d">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat33d</span></span><a class="headerlink" href="#warp.mat33d" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat44h">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat44h</span></span><a class="headerlink" href="#warp.mat44h" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat44f">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat44f</span></span><a class="headerlink" href="#warp.mat44f" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.mat44d">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mat44d</span></span><a class="headerlink" href="#warp.mat44d" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.quath">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quath</span></span><a class="headerlink" href="#warp.quath" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.quatf">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quatf</span></span><a class="headerlink" href="#warp.quatf" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.quatd">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quatd</span></span><a class="headerlink" href="#warp.quatd" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.transformh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transformh</span></span><a class="headerlink" href="#warp.transformh" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.transformf">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transformf</span></span><a class="headerlink" href="#warp.transformf" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.transformd">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transformd</span></span><a class="headerlink" href="#warp.transformd" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.spatial_vectorh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_vectorh</span></span><a class="headerlink" href="#warp.spatial_vectorh" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.spatial_vectorf">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_vectorf</span></span><a class="headerlink" href="#warp.spatial_vectorf" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.spatial_vectord">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_vectord</span></span><a class="headerlink" href="#warp.spatial_vectord" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.spatial_matrixh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_matrixh</span></span><a class="headerlink" href="#warp.spatial_matrixh" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.spatial_matrixf">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_matrixf</span></span><a class="headerlink" href="#warp.spatial_matrixf" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.spatial_matrixd">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_matrixd</span></span><a class="headerlink" href="#warp.spatial_matrixd" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-</section>
-<section id="generic-types">
-<h2>Generic Types<a class="headerlink" href="#generic-types" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Int">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Int</span></span><a class="headerlink" href="#warp.Int" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Float">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Float</span></span><a class="headerlink" href="#warp.Float" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Scalar">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Scalar</span></span><a class="headerlink" href="#warp.Scalar" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Vector">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Vector</span></span><a class="headerlink" href="#warp.Vector" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Matrix">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Matrix</span></span><a class="headerlink" href="#warp.Matrix" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Quaternion">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Quaternion</span></span><a class="headerlink" href="#warp.Quaternion" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Transformation">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Transformation</span></span><a class="headerlink" href="#warp.Transformation" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Array">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Array</span></span><a class="headerlink" href="#warp.Array" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-</section>
-<section id="scalar-math">
-<h2>Scalar Math<a class="headerlink" href="#scalar-math" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.min">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.min" title="Link to this definition">#</a></dt>
-<dd><p>Return the minimum of two scalars.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the element-wise minimum of two vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Return the minimum element of a vector <code class="docutils literal notranslate"><span class="pre">v</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.max">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.max" title="Link to this definition">#</a></dt>
-<dd><p>Return the maximum of two scalars.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the element-wise maximum of two vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Return the maximum element of a vector <code class="docutils literal notranslate"><span class="pre">v</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.clamp">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">clamp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.clamp" title="Link to this definition">#</a></dt>
-<dd><p>Clamp the value of <code class="docutils literal notranslate"><span class="pre">x</span></code> to the range [a, b].</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.abs">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">abs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.abs" title="Link to this definition">#</a></dt>
-<dd><p>Return the absolute value of <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sign">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sign</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.sign" title="Link to this definition">#</a></dt>
-<dd><p>Return -1 if <code class="docutils literal notranslate"><span class="pre">x</span></code> &lt; 0, return 1 otherwise.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.step">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">step</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.step" title="Link to this definition">#</a></dt>
-<dd><p>Return 1.0 if <code class="docutils literal notranslate"><span class="pre">x</span></code> &lt; 0.0, return 0.0 otherwise.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.nonzero">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">nonzero</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.nonzero" title="Link to this definition">#</a></dt>
-<dd><p>Return 1.0 if <code class="docutils literal notranslate"><span class="pre">x</span></code> is not equal to zero, return 0.0 otherwise.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sin">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.sin" title="Link to this definition">#</a></dt>
-<dd><p>Return the sine of <code class="docutils literal notranslate"><span class="pre">x</span></code> in radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.cos">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cos</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.cos" title="Link to this definition">#</a></dt>
-<dd><p>Return the cosine of <code class="docutils literal notranslate"><span class="pre">x</span></code> in radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.acos">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">acos</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.acos" title="Link to this definition">#</a></dt>
-<dd><p>Return arccos of <code class="docutils literal notranslate"><span class="pre">x</span></code> in radians. Inputs are automatically clamped to [-1.0, 1.0].</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.asin">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">asin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.asin" title="Link to this definition">#</a></dt>
-<dd><p>Return arcsin of <code class="docutils literal notranslate"><span class="pre">x</span></code> in radians. Inputs are automatically clamped to [-1.0, 1.0].</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sqrt">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sqrt</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.sqrt" title="Link to this definition">#</a></dt>
-<dd><p>Return the square root of <code class="docutils literal notranslate"><span class="pre">x</span></code>, where <code class="docutils literal notranslate"><span class="pre">x</span></code> is positive.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.cbrt">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cbrt</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.cbrt" title="Link to this definition">#</a></dt>
-<dd><p>Return the cube root of <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.tan">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">tan</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.tan" title="Link to this definition">#</a></dt>
-<dd><p>Return the tangent of <code class="docutils literal notranslate"><span class="pre">x</span></code> in radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.atan">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atan</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.atan" title="Link to this definition">#</a></dt>
-<dd><p>Return the arctangent of <code class="docutils literal notranslate"><span class="pre">x</span></code> in radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.atan2">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atan2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.atan2" title="Link to this definition">#</a></dt>
-<dd><p>Return the 2-argument arctangent, atan2, of the point <code class="docutils literal notranslate"><span class="pre">(x,</span> <span class="pre">y)</span></code> in radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sinh">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sinh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.sinh" title="Link to this definition">#</a></dt>
-<dd><p>Return the sinh of <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.cosh">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cosh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.cosh" title="Link to this definition">#</a></dt>
-<dd><p>Return the cosh of <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.tanh">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">tanh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.tanh" title="Link to this definition">#</a></dt>
-<dd><p>Return the tanh of <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.degrees">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">degrees</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.degrees" title="Link to this definition">#</a></dt>
-<dd><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> from radians into degrees.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.radians">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">radians</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.radians" title="Link to this definition">#</a></dt>
-<dd><p>Convert <code class="docutils literal notranslate"><span class="pre">x</span></code> from degrees into radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.log">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">log</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.log" title="Link to this definition">#</a></dt>
-<dd><p>Return the natural logarithm (base-e) of <code class="docutils literal notranslate"><span class="pre">x</span></code>, where <code class="docutils literal notranslate"><span class="pre">x</span></code> is positive.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.log2">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">log2</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.log2" title="Link to this definition">#</a></dt>
-<dd><p>Return the binary logarithm (base-2) of <code class="docutils literal notranslate"><span class="pre">x</span></code>, where <code class="docutils literal notranslate"><span class="pre">x</span></code> is positive.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.log10">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">log10</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.log10" title="Link to this definition">#</a></dt>
-<dd><p>Return the common logarithm (base-10) of <code class="docutils literal notranslate"><span class="pre">x</span></code>, where <code class="docutils literal notranslate"><span class="pre">x</span></code> is positive.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.exp">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">exp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.exp" title="Link to this definition">#</a></dt>
-<dd><p>Return the value of the exponential function <span class="math notranslate nohighlight">\(e^x\)</span>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.pow">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pow</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.pow" title="Link to this definition">#</a></dt>
-<dd><p>Return the result of <code class="docutils literal notranslate"><span class="pre">x</span></code> raised to power of <code class="docutils literal notranslate"><span class="pre">y</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.round">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">round</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.round" title="Link to this definition">#</a></dt>
-<dd><p>Return the nearest integer value to <code class="docutils literal notranslate"><span class="pre">x</span></code>, rounding halfway cases away from zero.
-This is the most intuitive form of rounding in the colloquial sense, but can be slower than other options like <a class="reference internal" href="#warp.rint" title="warp.rint"><code class="xref py py-func docutils literal notranslate"><span class="pre">warp.rint()</span></code></a>.
-Differs from <a class="reference external" href="https://numpy.org/doc/stable/reference/generated/numpy.round.html#numpy.round" title="(in NumPy v1.26)"><code class="xref py py-func docutils literal notranslate"><span class="pre">numpy.round()</span></code></a>, which behaves the same way as <code class="xref py py-func docutils literal notranslate"><span class="pre">numpy.rint()</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.rint">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">rint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.rint" title="Link to this definition">#</a></dt>
-<dd><p>Return the nearest integer value to <code class="docutils literal notranslate"><span class="pre">x</span></code>, rounding halfway cases to nearest even integer.
-It is generally faster than <a class="reference internal" href="#warp.round" title="warp.round"><code class="xref py py-func docutils literal notranslate"><span class="pre">warp.round()</span></code></a>. Equivalent to <code class="xref py py-func docutils literal notranslate"><span class="pre">numpy.rint()</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.trunc">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">trunc</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.trunc" title="Link to this definition">#</a></dt>
-<dd><p>Return the nearest integer that is closer to zero than <code class="docutils literal notranslate"><span class="pre">x</span></code>.
-In other words, it discards the fractional part of <code class="docutils literal notranslate"><span class="pre">x</span></code>.
-It is similar to casting <code class="docutils literal notranslate"><span class="pre">float(int(x))</span></code>, but preserves the negative sign when x is in the range [-0.0, -1.0).
-Equivalent to <code class="xref py py-func docutils literal notranslate"><span class="pre">numpy.trunc()</span></code> and <a class="reference external" href="https://numpy.org/doc/stable/reference/generated/numpy.fix.html#numpy.fix" title="(in NumPy v1.26)"><code class="xref py py-func docutils literal notranslate"><span class="pre">numpy.fix()</span></code></a>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.floor">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">floor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.floor" title="Link to this definition">#</a></dt>
-<dd><p>Return the largest integer that is less than or equal to <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.ceil">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">ceil</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.ceil" title="Link to this definition">#</a></dt>
-<dd><p>Return the smallest integer that is greater than or equal to <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.frac">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">frac</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.frac" title="Link to this definition">#</a></dt>
-<dd><dl class="simple">
-<dt>Retrieve the fractional part of x.</dt><dd><p>In other words, it discards the integer part of x and is equivalent to <code class="docutils literal notranslate"><span class="pre">x</span> <span class="pre">-</span> <span class="pre">trunc(x)</span></code>.</p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="vector-math">
-<h2>Vector Math<a class="headerlink" href="#vector-math" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dot">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">dot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.dot" title="Link to this definition">#</a></dt>
-<dd><p>Compute the dot product between two vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">dot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Compute the dot product between two quaternions.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.ddot">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">ddot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.ddot" title="Link to this definition">#</a></dt>
-<dd><p>Compute the double dot product between two matrices.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.argmin">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">argmin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></span><a class="headerlink" href="#warp.argmin" title="Link to this definition">#</a></dt>
-<dd><p>Return the index of the minimum element of a vector <code class="docutils literal notranslate"><span class="pre">v</span></code>. <a class="footnote-reference brackets" href="#id7" id="id1" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.argmax">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">argmax</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></span><a class="headerlink" href="#warp.argmax" title="Link to this definition">#</a></dt>
-<dd><p>Return the index of the maximum element of a vector <code class="docutils literal notranslate"><span class="pre">v</span></code>. <a class="footnote-reference brackets" href="#id7" id="id2" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.outer">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">outer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.outer" title="Link to this definition">#</a></dt>
-<dd><p>Compute the outer product <code class="docutils literal notranslate"><span class="pre">x*y^T</span></code> for two vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.cross">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cross</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.cross" title="Link to this definition">#</a></dt>
-<dd><p>Compute the cross product of two 3D vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.skew">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">skew</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.skew" title="Link to this definition">#</a></dt>
-<dd><p>Compute the skew-symmetric 3x3 matrix for a 3D vector <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.length">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">length</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.length" title="Link to this definition">#</a></dt>
-<dd><p>Compute the length of a vector <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">length</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Compute the length of a quaternion <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.length_sq">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">length_sq</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.length_sq" title="Link to this definition">#</a></dt>
-<dd><p>Compute the squared length of a 2D vector <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">length_sq</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Compute the squared length of a quaternion <code class="docutils literal notranslate"><span class="pre">x</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.normalize">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">normalize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.normalize" title="Link to this definition">#</a></dt>
-<dd><p>Compute the normalized value of <code class="docutils literal notranslate"><span class="pre">x</span></code>. If <code class="docutils literal notranslate"><span class="pre">length(x)</span></code> is 0 then the zero vector is returned.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">normalize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Compute the normalized value of <code class="docutils literal notranslate"><span class="pre">x</span></code>. If <code class="docutils literal notranslate"><span class="pre">length(x)</span></code> is 0, then the zero quaternion is returned.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transpose">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transpose</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.transpose" title="Link to this definition">#</a></dt>
-<dd><p>Return the transpose of the matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.inverse">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">inverse</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">2</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">2</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.inverse" title="Link to this definition">#</a></dt>
-<dd><p>Return the inverse of a 2x2 matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">inverse</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the inverse of a 3x3 matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">inverse</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the inverse of a 4x4 matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.determinant">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">determinant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">2</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">2</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.determinant" title="Link to this definition">#</a></dt>
-<dd><p>Return the determinant of a 2x2 matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">determinant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Return the determinant of a 3x3 matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">determinant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span></dt>
-<dd><p>Return the determinant of a 4x4 matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.trace">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">trace</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.trace" title="Link to this definition">#</a></dt>
-<dd><p>Return the trace of the matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.diag">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">diag</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.diag" title="Link to this definition">#</a></dt>
-<dd><p>Returns a matrix with the components of the vector <code class="docutils literal notranslate"><span class="pre">d</span></code> on the diagonal.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.get_diag">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">get_diag</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.get_diag" title="Link to this definition">#</a></dt>
-<dd><p>Returns a vector containing the diagonal elements of the square matrix <code class="docutils literal notranslate"><span class="pre">m</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.cw_mul">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cw_mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.cw_mul" title="Link to this definition">#</a></dt>
-<dd><p>Component-wise multiplication of two 2D vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cw_mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Component-wise multiplication of two 2D vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.cw_div">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cw_div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.cw_div" title="Link to this definition">#</a></dt>
-<dd><p>Component-wise division of two 2D vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">cw_div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Component-wise division of two 2D vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.vector">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vector</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">w</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.vector" title="Link to this definition">#</a></dt>
-<dd><p>Construct a 6D screw vector from two 3D vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">vector</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">arg_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">length</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Construct a vector of with given length and dtype.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.matrix">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">matrix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.matrix" title="Link to this definition">#</a></dt>
-<dd><p>Construct a 4x4 transformation matrix that applies the transformations as
-Translation(pos)*Rotation(rot)*Scale(scale) when applied to column vectors, i.e.: y = (TRS)*x</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">matrix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="o"><span class="pre">*</span></span><span class="n"><span class="pre">arg_types</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Construct a matrix. If the positional <code class="docutils literal notranslate"><span class="pre">arg_types</span></code> are not given, then matrix will be zero-initialized.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.identity">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">identity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">n</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.identity" title="Link to this definition">#</a></dt>
-<dd><p>Create an identity matrix with shape=(n,n) with the type given by <code class="docutils literal notranslate"><span class="pre">dtype</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.svd3">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">svd3</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">U</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">V</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.svd3" title="Link to this definition">#</a></dt>
-<dd><p>Compute the SVD of a 3x3 matrix <code class="docutils literal notranslate"><span class="pre">A</span></code>. The singular values are returned in <code class="docutils literal notranslate"><span class="pre">sigma</span></code>,
-while the left and right basis vectors are returned in <code class="docutils literal notranslate"><span class="pre">U</span></code> and <code class="docutils literal notranslate"><span class="pre">V</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.qr3">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">qr3</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">Q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">R</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.qr3" title="Link to this definition">#</a></dt>
-<dd><p>Compute the QR decomposition of a 3x3 matrix <code class="docutils literal notranslate"><span class="pre">A</span></code>. The orthogonal matrix is returned in <code class="docutils literal notranslate"><span class="pre">Q</span></code>,
-while the upper triangular matrix is returned in <code class="docutils literal notranslate"><span class="pre">R</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.eig3">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">eig3</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">Q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">d</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.eig3" title="Link to this definition">#</a></dt>
-<dd><p>Compute the eigendecomposition of a 3x3 matrix <code class="docutils literal notranslate"><span class="pre">A</span></code>. The eigenvectors are returned as the columns of <code class="docutils literal notranslate"><span class="pre">Q</span></code>,
-while the corresponding eigenvalues are returned in <code class="docutils literal notranslate"><span class="pre">d</span></code>.</p>
-</dd></dl>
-
-</section>
-<section id="other">
-<h2>Other<a class="headerlink" href="#other" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.lower_bound">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lower_bound</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">arr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.lower_bound" title="Link to this definition">#</a></dt>
-<dd><p>Search a sorted array <code class="docutils literal notranslate"><span class="pre">arr</span></code> for the closest element greater than or equal to <code class="docutils literal notranslate"><span class="pre">value</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lower_bound</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">arr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arr_begin</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arr_end</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span></dt>
-<dd><p>Search a sorted array <code class="docutils literal notranslate"><span class="pre">arr</span></code> in the range [arr_begin, arr_end) for the closest element greater than or equal to <code class="docutils literal notranslate"><span class="pre">value</span></code>.</p>
-</dd></dl>
-
-</section>
-<section id="quaternion-math">
-<h2>Quaternion Math<a class="headerlink" href="#quaternion-math" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quaternion">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quaternion</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quaternion" title="Link to this definition">#</a></dt>
-<dd><p>Construct a zero-initialized quaternion. Quaternions are laid out as
-[ix, iy, iz, r], where ix, iy, iz are the imaginary part, and r the real part.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quaternion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">z</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">w</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Create a quaternion using the supplied components (type inferred from component type).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quaternion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">r</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Create a quaternion using the supplied vector/scalar (type inferred from scalar type).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quaternion</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Construct a quaternion of type dtype from another quaternion of a different dtype.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_identity">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_identity</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.quatf" title="warp.quatf"><span class="pre">quatf</span></a></span></span><a class="headerlink" href="#warp.quat_identity" title="Link to this definition">#</a></dt>
-<dd><p>Construct an identity quaternion with zero imaginary part and real part of 1.0</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_from_axis_angle">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_from_axis_angle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">axis</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_from_axis_angle" title="Link to this definition">#</a></dt>
-<dd><p>Construct a quaternion representing a rotation of angle radians around the given axis.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_to_axis_angle">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_to_axis_angle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angle</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.quat_to_axis_angle" title="Link to this definition">#</a></dt>
-<dd><p>Extract the rotation axis and angle radians a quaternion represents.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_from_matrix">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_from_matrix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_from_matrix" title="Link to this definition">#</a></dt>
-<dd><p>Construct a quaternion from a 3x3 matrix.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_rpy">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_rpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">roll</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pitch</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">yaw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_rpy" title="Link to this definition">#</a></dt>
-<dd><p>Construct a quaternion representing a combined roll (z), pitch (x), yaw rotations (y) in radians.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_inverse">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_inverse</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_inverse" title="Link to this definition">#</a></dt>
-<dd><p>Compute quaternion conjugate.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_rotate">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_rotate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_rotate" title="Link to this definition">#</a></dt>
-<dd><p>Rotate a vector by a quaternion.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_rotate_inv">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_rotate_inv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_rotate_inv" title="Link to this definition">#</a></dt>
-<dd><p>Rotate a vector by the inverse of a quaternion.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_slerp">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_slerp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q0</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">q1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_slerp" title="Link to this definition">#</a></dt>
-<dd><p>Linearly interpolate between two quaternions.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.quat_to_matrix">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">quat_to_matrix</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.quat_to_matrix" title="Link to this definition">#</a></dt>
-<dd><p>Convert a quaternion to a 3x3 rotation matrix.</p>
-</dd></dl>
-
-</section>
-<section id="transformations">
-<h2>Transformations<a class="headerlink" href="#transformations" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transformation">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transformation</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">q</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transformation" title="Link to this definition">#</a></dt>
-<dd><p>Construct a rigid-body transformation with translation part <code class="docutils literal notranslate"><span class="pre">p</span></code> and rotation <code class="docutils literal notranslate"><span class="pre">q</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_identity">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_identity</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.transformf" title="warp.transformf"><span class="pre">transformf</span></a></span></span><a class="headerlink" href="#warp.transform_identity" title="Link to this definition">#</a></dt>
-<dd><p>Construct an identity transform with zero translation and identity rotation.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_get_translation">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_get_translation</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transform_get_translation" title="Link to this definition">#</a></dt>
-<dd><p>Return the translational part of a transform <code class="docutils literal notranslate"><span class="pre">t</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_get_rotation">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_get_rotation</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transform_get_rotation" title="Link to this definition">#</a></dt>
-<dd><p>Return the rotational part of a transform <code class="docutils literal notranslate"><span class="pre">t</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_multiply">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_multiply</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transform_multiply" title="Link to this definition">#</a></dt>
-<dd><p>Multiply two rigid body transformations together.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_point">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transform_point" title="Link to this definition">#</a></dt>
-<dd><p>Apply the transform to a point <code class="docutils literal notranslate"><span class="pre">p</span></code> treating the homogenous coordinate as w=1 (translation and rotation).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">p</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Apply the transform to a point <code class="docutils literal notranslate"><span class="pre">p</span></code> treating the homogenous coordinate as w=1.
-The transformation is applied treating <code class="docutils literal notranslate"><span class="pre">p</span></code> as a column vector, e.g.: <code class="docutils literal notranslate"><span class="pre">y</span> <span class="pre">=</span> <span class="pre">M*p</span></code>.
-Note this is in contrast to some libraries, notably USD, which applies transforms to row vectors, <code class="docutils literal notranslate"><span class="pre">y^T</span> <span class="pre">=</span> <span class="pre">p^T*M^T</span></code>.
-If the transform is coming from a library that uses row-vectors, then users should transpose the transformation
-matrix before calling this method.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_vector">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_vector</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transform_vector" title="Link to this definition">#</a></dt>
-<dd><p>Apply the transform to a vector <code class="docutils literal notranslate"><span class="pre">v</span></code> treating the homogenous coordinate as w=0 (rotation only).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_vector</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">m</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">4</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Apply the transform to a vector <code class="docutils literal notranslate"><span class="pre">v</span></code> treating the homogenous coordinate as w=0.
-The transformation is applied treating <code class="docutils literal notranslate"><span class="pre">v</span></code> as a column vector, e.g.: <code class="docutils literal notranslate"><span class="pre">y</span> <span class="pre">=</span> <span class="pre">M*v</span></code>
-note this is in contrast to some libraries, notably USD, which applies transforms to row vectors, <code class="docutils literal notranslate"><span class="pre">y^T</span> <span class="pre">=</span> <span class="pre">v^T*M^T</span></code>.
-If the transform is coming from a library that uses row-vectors, then users should transpose the transformation
-matrix before calling this method.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.transform_inverse">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">transform_inverse</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.transform_inverse" title="Link to this definition">#</a></dt>
-<dd><p>Compute the inverse of the transformation <code class="docutils literal notranslate"><span class="pre">t</span></code>.</p>
-</dd></dl>
-
-</section>
-<section id="spatial-math">
-<h2>Spatial Math<a class="headerlink" href="#spatial-math" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_adjoint">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_adjoint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">r</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">3</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.spatial_adjoint" title="Link to this definition">#</a></dt>
-<dd><p>Construct a 6x6 spatial inertial matrix from two 3x3 diagonal blocks.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_dot">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_dot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.spatial_dot" title="Link to this definition">#</a></dt>
-<dd><p>Compute the dot product of two 6D screw vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_cross">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_cross</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.spatial_cross" title="Link to this definition">#</a></dt>
-<dd><p>Compute the cross product of two 6D screw vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_cross_dual">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_cross_dual</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span><a class="headerlink" href="#warp.spatial_cross_dual" title="Link to this definition">#</a></dt>
-<dd><p>Compute the dual cross product of two 6D screw vectors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_top">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_top</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.spatial_top" title="Link to this definition">#</a></dt>
-<dd><p>Return the top (first) part of a 6D screw vector.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_bottom">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_bottom</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.spatial_bottom" title="Link to this definition">#</a></dt>
-<dd><p>Return the bottom (second) part of a 6D screw vector.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_jacobian">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_jacobian</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">S</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_parents</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_qd_start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">J_start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">J_out</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.spatial_jacobian" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.spatial_mass">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">spatial_mass</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">I_s</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="m"><span class="pre">6</span></span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_count</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">M_start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">M</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.spatial_mass" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-</section>
-<section id="utility">
-<h2>Utility<a class="headerlink" href="#utility" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mlp">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mlp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">weights</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bias</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">activation</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Callable</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">out</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.mlp" title="Link to this definition">#</a></dt>
-<dd><p>Evaluate a multi-layer perceptron (MLP) layer in the form: <code class="docutils literal notranslate"><span class="pre">out</span> <span class="pre">=</span> <span class="pre">act(weights*x</span> <span class="pre">+</span> <span class="pre">bias)</span></code>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>weights</strong> – A layer’s network weights with dimensions <code class="docutils literal notranslate"><span class="pre">(m,</span> <span class="pre">n)</span></code>.</p></li>
-<li><p><strong>bias</strong> – An array with dimensions <code class="docutils literal notranslate"><span class="pre">(n)</span></code>.</p></li>
-<li><p><strong>activation</strong> – A <code class="docutils literal notranslate"><span class="pre">wp.func</span></code> function that takes a single scalar float as input and returns a scalar float as output</p></li>
-<li><p><strong>index</strong> – The batch item to process, typically each thread will process one item in the batch, in which case
-index should be <code class="docutils literal notranslate"><span class="pre">wp.tid()</span></code></p></li>
-<li><p><strong>x</strong> – The feature matrix with dimensions <code class="docutils literal notranslate"><span class="pre">(n,</span> <span class="pre">b)</span></code></p></li>
-<li><p><strong>out</strong> – The network output with dimensions <code class="docutils literal notranslate"><span class="pre">(m,</span> <span class="pre">b)</span></code></p></li>
-</ul>
-</dd>
-<dt class="field-even">Note<span class="colon">:</span></dt>
-<dd class="field-even"><p>Feature and output matrices are transposed compared to some other frameworks such as PyTorch.
-All matrices are assumed to be stored in flattened row-major memory layout (NumPy default).</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.printf">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">printf</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.printf" title="Link to this definition">#</a></dt>
-<dd><p>Allows printing formatted strings using C-style format specifiers.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.print">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">print</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.print" title="Link to this definition">#</a></dt>
-<dd><p>Print variable to stdout</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.breakpoint">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">breakpoint</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.breakpoint" title="Link to this definition">#</a></dt>
-<dd><p>Debugger breakpoint</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.tid">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">tid</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.tid" title="Link to this definition">#</a></dt>
-<dd><p>Return the current thread index for a 1D kernel launch. Note that this is the <em>global</em> index of the thread in the range [0, dim)
-where dim is the parameter passed to kernel launch. This function may not be called from user-defined Warp functions.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">tid</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the current thread indices for a 2D kernel launch. Use <code class="docutils literal notranslate"><span class="pre">i,j</span> <span class="pre">=</span> <span class="pre">wp.tid()</span></code> syntax to retrieve the
-coordinates inside the kernel thread grid. This function may not be called from user-defined Warp functions.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">tid</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the current thread indices for a 3D kernel launch. Use <code class="docutils literal notranslate"><span class="pre">i,j,k</span> <span class="pre">=</span> <span class="pre">wp.tid()</span></code> syntax to retrieve the
-coordinates inside the kernel thread grid. This function may not be called from user-defined Warp functions.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">tid</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">Tuple</span><span class="p"><span class="pre">[</span></span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Return the current thread indices for a 4D kernel launch. Use <code class="docutils literal notranslate"><span class="pre">i,j,k,l</span> <span class="pre">=</span> <span class="pre">wp.tid()</span></code> syntax to retrieve the
-coordinates inside the kernel thread grid. This function may not be called from user-defined Warp functions.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.select">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.select" title="Link to this definition">#</a></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int8" title="warp.int8"><span class="pre">int8</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint8" title="warp.uint8"><span class="pre">uint8</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int16" title="warp.int16"><span class="pre">int16</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint16" title="warp.uint16"><span class="pre">uint16</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int64" title="warp.int64"><span class="pre">int64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">cond</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">cond</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code> then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">select</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">arr</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Select between two arguments, if <code class="docutils literal notranslate"><span class="pre">arr</span></code> is null then return <code class="docutils literal notranslate"><span class="pre">arg1</span></code>, otherwise return <code class="docutils literal notranslate"><span class="pre">arg2</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.atomic_add">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.atomic_add" title="Link to this definition">#</a></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically add <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.atomic_sub">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.atomic_sub" title="Link to this definition">#</a></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Atomically subtract <code class="docutils literal notranslate"><span class="pre">value</span></code> onto <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.atomic_min">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.atomic_min" title="Link to this definition">#</a></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_min</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the minimum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.atomic_max">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.atomic_max" title="Link to this definition">#</a></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">FabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">atomic_max</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">IndexedFabricArray</span><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">Any</span></span></em><span class="sig-paren">)</span></dt>
-<dd><p>Compute the maximum of <code class="docutils literal notranslate"><span class="pre">value</span></code> and <code class="docutils literal notranslate"><span class="pre">a[i,j,k,l]</span></code> and atomically update the array.</p>
-</dd></dl>
-
-<p>Note that for vectors and matrices the operation is only atomic on a per-component basis.</p>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.lerp">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lerp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.lerp" title="Link to this definition">#</a></dt>
-<dd><p>Linearly interpolate two values <code class="docutils literal notranslate"><span class="pre">a</span></code> and <code class="docutils literal notranslate"><span class="pre">b</span></code> using factor <code class="docutils literal notranslate"><span class="pre">t</span></code>, computed as <code class="docutils literal notranslate"><span class="pre">a*(1-t)</span> <span class="pre">+</span> <span class="pre">b*t</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lerp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Linearly interpolate two values <code class="docutils literal notranslate"><span class="pre">a</span></code> and <code class="docutils literal notranslate"><span class="pre">b</span></code> using factor <code class="docutils literal notranslate"><span class="pre">t</span></code>, computed as <code class="docutils literal notranslate"><span class="pre">a*(1-t)</span> <span class="pre">+</span> <span class="pre">b*t</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lerp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Linearly interpolate two values <code class="docutils literal notranslate"><span class="pre">a</span></code> and <code class="docutils literal notranslate"><span class="pre">b</span></code> using factor <code class="docutils literal notranslate"><span class="pre">t</span></code>, computed as <code class="docutils literal notranslate"><span class="pre">a*(1-t)</span> <span class="pre">+</span> <span class="pre">b*t</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lerp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Linearly interpolate two values <code class="docutils literal notranslate"><span class="pre">a</span></code> and <code class="docutils literal notranslate"><span class="pre">b</span></code> using factor <code class="docutils literal notranslate"><span class="pre">t</span></code>, computed as <code class="docutils literal notranslate"><span class="pre">a*(1-t)</span> <span class="pre">+</span> <span class="pre">b*t</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lerp</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd><p>Linearly interpolate two values <code class="docutils literal notranslate"><span class="pre">a</span></code> and <code class="docutils literal notranslate"><span class="pre">b</span></code> using factor <code class="docutils literal notranslate"><span class="pre">t</span></code>, computed as <code class="docutils literal notranslate"><span class="pre">a*(1-t)</span> <span class="pre">+</span> <span class="pre">b*t</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.smoothstep">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">smoothstep</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">edge0</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></span><a class="headerlink" href="#warp.smoothstep" title="Link to this definition">#</a></dt>
-<dd><p>Smoothly interpolate between two values <code class="docutils literal notranslate"><span class="pre">edge0</span></code> and <code class="docutils literal notranslate"><span class="pre">edge1</span></code> using a factor <code class="docutils literal notranslate"><span class="pre">x</span></code>,
-and return a result between 0 and 1 using a cubic Hermite interpolation after clamping.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.expect_near">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">expect_near</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">tolerance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Float" title="warp.Float"><span class="pre">Float</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.expect_near" title="Link to this definition">#</a></dt>
-<dd><p>Prints an error to stdout if <code class="docutils literal notranslate"><span class="pre">arg1</span></code> and <code class="docutils literal notranslate"><span class="pre">arg2</span></code> are not closer than tolerance in magnitude</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">expect_near</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">arg1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">arg2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">tolerance</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span></dt>
-<dd><p>Prints an error to stdout if any element of <code class="docutils literal notranslate"><span class="pre">arg1</span></code> and <code class="docutils literal notranslate"><span class="pre">arg2</span></code> are not closer than tolerance in magnitude</p>
-</dd></dl>
-
-</section>
-<section id="geometry">
-<h2>Geometry<a class="headerlink" href="#geometry" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.bvh_query_aabb">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bvh_query_aabb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">lower</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">upper</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bvh_query_t</span></span></span><a class="headerlink" href="#warp.bvh_query_aabb" title="Link to this definition">#</a></dt>
-<dd><p>Construct an axis-aligned bounding box query against a BVH object. This query can be used to iterate over all bounds
-inside a BVH. Returns an object that is used to track state during BVH traversal.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The BVH identifier</p></li>
-<li><p><strong>lower</strong> – The lower bound of the bounding box in BVH space</p></li>
-<li><p><strong>upper</strong> – The upper bound of the bounding box in BVH space</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.bvh_query_ray">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bvh_query_ray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">bvh_query_t</span></span></span><a class="headerlink" href="#warp.bvh_query_ray" title="Link to this definition">#</a></dt>
-<dd><p>Construct a ray query against a BVH object. This query can be used to iterate over all bounds
-that intersect the ray. Returns an object that is used to track state during BVH traversal.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The BVH identifier</p></li>
-<li><p><strong>start</strong> – The start of the ray in BVH space</p></li>
-<li><p><strong>dir</strong> – The direction of the ray in BVH space</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.bvh_query_next">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bvh_query_next</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">query</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">bvh_query_t</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.bvh_query_next" title="Link to this definition">#</a></dt>
-<dd><p>Move to the next bound returned by the query.
-The index of the current bound is stored in <code class="docutils literal notranslate"><span class="pre">index</span></code>, returns <code class="docutils literal notranslate"><span class="pre">False</span></code> if there are no more overlapping bound.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_point">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inside</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_point" title="Link to this definition">#</a></dt>
-<dd><p>Computes the closest point on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> with identifier <code class="docutils literal notranslate"><span class="pre">id</span></code> to the given <code class="docutils literal notranslate"><span class="pre">point</span></code> in space. Returns <code class="docutils literal notranslate"><span class="pre">True</span></code> if a point &lt; <code class="docutils literal notranslate"><span class="pre">max_dist</span></code> is found.</p>
-<p>Identifies the sign of the distance using additional ray-casts to determine if the point is inside or outside.
-This method is relatively robust, but does increase computational cost.
-See below for additional sign determination methods.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>point</strong> – The point in space to query</p></li>
-<li><p><strong>max_dist</strong> – Mesh faces above this distance will not be considered by the query</p></li>
-<li><p><strong>inside</strong> – Returns a value &lt; 0 if query point is inside the mesh, &gt;=0 otherwise.
-Note that mesh must be watertight for this to be robust</p></li>
-<li><p><strong>face</strong> – Returns the index of the closest face</p></li>
-<li><p><strong>bary_u</strong> – Returns the barycentric u coordinate of the closest point</p></li>
-<li><p><strong>bary_v</strong> – Returns the barycentric v coordinate of the closest point</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_point_no_sign">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_point_no_sign</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_point_no_sign" title="Link to this definition">#</a></dt>
-<dd><p>Computes the closest point on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> with identifier <code class="docutils literal notranslate"><span class="pre">id</span></code> to the given <code class="docutils literal notranslate"><span class="pre">point</span></code> in space. Returns <code class="docutils literal notranslate"><span class="pre">True</span></code> if a point &lt; <code class="docutils literal notranslate"><span class="pre">max_dist</span></code> is found.</p>
-<p>This method does not compute the sign of the point (inside/outside) which makes it faster than other point query methods.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>point</strong> – The point in space to query</p></li>
-<li><p><strong>max_dist</strong> – Mesh faces above this distance will not be considered by the query</p></li>
-<li><p><strong>face</strong> – Returns the index of the closest face</p></li>
-<li><p><strong>bary_u</strong> – Returns the barycentric u coordinate of the closest point</p></li>
-<li><p><strong>bary_v</strong> – Returns the barycentric v coordinate of the closest point</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_furthest_point_no_sign">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_furthest_point_no_sign</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_furthest_point_no_sign" title="Link to this definition">#</a></dt>
-<dd><p>Computes the furthest point on the mesh with identifier <cite>id</cite> to the given point in space. Returns <code class="docutils literal notranslate"><span class="pre">True</span></code> if a point &gt; <code class="docutils literal notranslate"><span class="pre">min_dist</span></code> is found.</p>
-<p>This method does not compute the sign of the point (inside/outside).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>point</strong> – The point in space to query</p></li>
-<li><p><strong>min_dist</strong> – Mesh faces below this distance will not be considered by the query</p></li>
-<li><p><strong>face</strong> – Returns the index of the furthest face</p></li>
-<li><p><strong>bary_u</strong> – Returns the barycentric u coordinate of the furthest point</p></li>
-<li><p><strong>bary_v</strong> – Returns the barycentric v coordinate of the furthest point</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_point_sign_normal">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_point_sign_normal</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inside</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">epsilon</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_point_sign_normal" title="Link to this definition">#</a></dt>
-<dd><p>Computes the closest point on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> with identifier <code class="docutils literal notranslate"><span class="pre">id</span></code> to the given <code class="docutils literal notranslate"><span class="pre">point</span></code> in space. Returns <code class="docutils literal notranslate"><span class="pre">True</span></code> if a point &lt; <code class="docutils literal notranslate"><span class="pre">max_dist</span></code> is found.</p>
-<p>Identifies the sign of the distance (inside/outside) using the angle-weighted pseudo normal.
-This approach to sign determination is robust for well conditioned meshes that are watertight and non-self intersecting.
-It is also comparatively fast to compute.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>point</strong> – The point in space to query</p></li>
-<li><p><strong>max_dist</strong> – Mesh faces above this distance will not be considered by the query</p></li>
-<li><p><strong>inside</strong> – Returns a value &lt; 0 if query point is inside the mesh, &gt;=0 otherwise.
-Note that mesh must be watertight for this to be robust</p></li>
-<li><p><strong>face</strong> – Returns the index of the closest face</p></li>
-<li><p><strong>bary_u</strong> – Returns the barycentric u coordinate of the closest point</p></li>
-<li><p><strong>bary_v</strong> – Returns the barycentric v coordinate of the closest point</p></li>
-<li><p><strong>epsilon</strong> – Epsilon treating distance values as equal, when locating the minimum distance vertex/face/edge, as a
-fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_point_sign_winding_number">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_point_sign_winding_number</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">inside</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">accuracy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">threshold</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_point_sign_winding_number" title="Link to this definition">#</a></dt>
-<dd><p>Computes the closest point on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> with identifier <code class="docutils literal notranslate"><span class="pre">id</span></code> to the given point in space. Returns <code class="docutils literal notranslate"><span class="pre">True</span></code> if a point &lt; <code class="docutils literal notranslate"><span class="pre">max_dist</span></code> is found.</p>
-<p>Identifies the sign using the winding number of the mesh relative to the query point. This method of sign determination is robust for poorly conditioned meshes
-and provides a smooth approximation to sign even when the mesh is not watertight. This method is the most robust and accurate of the sign determination meshes
-but also the most expensive.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> object must be constructed with <code class="docutils literal notranslate"><span class="pre">support_winding_number=True</span></code> for this method to return correct results.</p>
-</div>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>point</strong> – The point in space to query</p></li>
-<li><p><strong>max_dist</strong> – Mesh faces above this distance will not be considered by the query</p></li>
-<li><p><strong>inside</strong> – Returns a value &lt; 0 if query point is inside the mesh, &gt;=0 otherwise.
-Note that mesh must be watertight for this to be robust</p></li>
-<li><p><strong>face</strong> – Returns the index of the closest face</p></li>
-<li><p><strong>bary_u</strong> – Returns the barycentric u coordinate of the closest point</p></li>
-<li><p><strong>bary_v</strong> – Returns the barycentric v coordinate of the closest point</p></li>
-<li><p><strong>accuracy</strong> – Accuracy for computing the winding number with fast winding number method utilizing second-order dipole approximation, default 2.0</p></li>
-<li><p><strong>threshold</strong> – The threshold of the winding number to be considered inside, default 0.5</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_ray">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_ray</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">start</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">dir</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">t</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">sign</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">normal</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_ray" title="Link to this definition">#</a></dt>
-<dd><p>Computes the closest ray hit on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> with identifier <code class="docutils literal notranslate"><span class="pre">id</span></code>, returns <code class="docutils literal notranslate"><span class="pre">True</span></code> if a point &lt; <code class="docutils literal notranslate"><span class="pre">max_t</span></code> is found.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>start</strong> – The start point of the ray</p></li>
-<li><p><strong>dir</strong> – The ray direction (should be normalized)</p></li>
-<li><p><strong>max_t</strong> – The maximum distance along the ray to check for intersections</p></li>
-<li><p><strong>t</strong> – Returns the distance of the closest hit along the ray</p></li>
-<li><p><strong>bary_u</strong> – Returns the barycentric u coordinate of the closest hit</p></li>
-<li><p><strong>bary_v</strong> – Returns the barycentric v coordinate of the closest hit</p></li>
-<li><p><strong>sign</strong> – Returns a value &gt; 0 if the hit ray hit front of the face, returns &lt; 0 otherwise</p></li>
-<li><p><strong>normal</strong> – Returns the face normal</p></li>
-<li><p><strong>face</strong> – Returns the index of the hit face</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_aabb">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_aabb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">lower</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">upper</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">mesh_query_aabb_t</span></span></span><a class="headerlink" href="#warp.mesh_query_aabb" title="Link to this definition">#</a></dt>
-<dd><p>Construct an axis-aligned bounding box query against a <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a>.
-This query can be used to iterate over all triangles inside a volume.
-Returns an object that is used to track state during mesh traversal.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>id</strong> – The mesh identifier</p></li>
-<li><p><strong>lower</strong> – The lower bound of the bounding box in mesh space</p></li>
-<li><p><strong>upper</strong> – The upper bound of the bounding box in mesh space</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_query_aabb_next">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_query_aabb_next</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">query</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">mesh_query_aabb_t</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.mesh_query_aabb_next" title="Link to this definition">#</a></dt>
-<dd><p>Move to the next triangle overlapping the query bounding box.
-The index of the current face is stored in <code class="docutils literal notranslate"><span class="pre">index</span></code>, returns <code class="docutils literal notranslate"><span class="pre">False</span></code> if there are no more overlapping triangles.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_eval_position">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_eval_position</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.mesh_eval_position" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the position on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> given a face index and barycentric coordinates.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_eval_velocity">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_eval_velocity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_u</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">bary_v</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.mesh_eval_velocity" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the velocity on the <a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> given a face index and barycentric coordinates.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.hash_grid_query">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">hash_grid_query</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">point</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_dist</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><span class="pre">hash_grid_query_t</span></span></span><a class="headerlink" href="#warp.hash_grid_query" title="Link to this definition">#</a></dt>
-<dd><p>Construct a point query against a <a class="reference internal" href="runtime.html#warp.HashGrid" title="warp.HashGrid"><code class="xref py py-class docutils literal notranslate"><span class="pre">HashGrid</span></code></a>. This query can be used to iterate over all neighboring points within a fixed radius from the query point. Returns an object that is used to track state during neighbor traversal.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.hash_grid_query_next">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">hash_grid_query_next</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">query</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><span class="pre">hash_grid_query_t</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.hash_grid_query_next" title="Link to this definition">#</a></dt>
-<dd><p>Move to the next point in the hash grid query. The index of the current neighbor is stored in <code class="docutils literal notranslate"><span class="pre">index</span></code>, returns <code class="docutils literal notranslate"><span class="pre">False</span></code>
-if there are no more neighbors.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.hash_grid_point_id">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">hash_grid_point_id</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.hash_grid_point_id" title="Link to this definition">#</a></dt>
-<dd><p>Return the index of a point in the <a class="reference internal" href="runtime.html#warp.HashGrid" title="warp.HashGrid"><code class="xref py py-class docutils literal notranslate"><span class="pre">HashGrid</span></code></a>. This can be used to reorder threads such that grid
-traversal occurs in a spatially coherent order.</p>
-<p>Returns -1 if the <a class="reference internal" href="runtime.html#warp.HashGrid" title="warp.HashGrid"><code class="xref py py-class docutils literal notranslate"><span class="pre">HashGrid</span></code></a> has not been reserved.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.intersect_tri_tri">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">intersect_tri_tri</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">v0</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">v1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">v2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">u0</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">u1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">u2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.intersect_tri_tri" title="Link to this definition">#</a></dt>
-<dd><p>Tests for intersection between two triangles (v0, v1, v2) and (u0, u1, u2) using Moller’s method. Returns &gt; 0 if triangles intersect.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_get">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_get</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="runtime.html#warp.Mesh" title="warp.Mesh"><span class="pre">Mesh</span></a></span></span><a class="headerlink" href="#warp.mesh_get" title="Link to this definition">#</a></dt>
-<dd><p>Retrieves the mesh given its index. <a class="footnote-reference brackets" href="#id7" id="id3" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_eval_face_normal">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_eval_face_normal</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">face</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.mesh_eval_face_normal" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the face normal the mesh given a face index.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_get_point">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_get_point</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.mesh_get_point" title="Link to this definition">#</a></dt>
-<dd><p>Returns the point of the mesh given a index.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_get_velocity">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_get_velocity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.mesh_get_velocity" title="Link to this definition">#</a></dt>
-<dd><p>Returns the velocity of the mesh given a index.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mesh_get_index">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mesh_get_index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">index</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.mesh_get_index" title="Link to this definition">#</a></dt>
-<dd><p>Returns the point-index of the mesh given a face-vertex index.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.closest_point_edge_edge">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">closest_point_edge_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">p1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">q1</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">p2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">q2</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">epsilon</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.closest_point_edge_edge" title="Link to this definition">#</a></dt>
-<dd><p>Finds the closest points between two edges. Returns barycentric weights to the points on each edge, as well as the closest distance between the edges.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>p1</strong> – First point of first edge</p></li>
-<li><p><strong>q1</strong> – Second point of first edge</p></li>
-<li><p><strong>p2</strong> – First point of second edge</p></li>
-<li><p><strong>q2</strong> – Second point of second edge</p></li>
-<li><p><strong>epsilon</strong> – Zero tolerance for determining if points in an edge are degenerate.</p></li>
-<li><p><strong>out</strong> – vec3 output containing (s,t,d), where <cite>s</cite> in [0,1] is the barycentric weight for the first edge, <cite>t</cite> is the barycentric weight for the second edge, and <cite>d</cite> is the distance between the two edges at these two closest points.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="volumes">
-<h2>Volumes<a class="headerlink" href="#volumes" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_sample_f">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_sample_f</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">uvw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">sampling_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.volume_sample_f" title="Link to this definition">#</a></dt>
-<dd><p>Sample the volume given by <code class="docutils literal notranslate"><span class="pre">id</span></code> at the volume local-space point <code class="docutils literal notranslate"><span class="pre">uvw</span></code>.
-Interpolation should be <a class="reference internal" href="runtime.html#warp.Volume.CLOSEST" title="warp.Volume.CLOSEST"><code class="xref py py-attr docutils literal notranslate"><span class="pre">warp.Volume.CLOSEST</span></code></a> or <code class="xref py py-attr docutils literal notranslate"><span class="pre">wp.Volume.LINEAR.</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_sample_grad_f">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_sample_grad_f</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">uvw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">sampling_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">grad</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.volume_sample_grad_f" title="Link to this definition">#</a></dt>
-<dd><p>Sample the volume and its gradient given by <code class="docutils literal notranslate"><span class="pre">id</span></code> at the volume local-space point <code class="docutils literal notranslate"><span class="pre">uvw</span></code>.
-Interpolation should be <a class="reference internal" href="runtime.html#warp.Volume.CLOSEST" title="warp.Volume.CLOSEST"><code class="xref py py-attr docutils literal notranslate"><span class="pre">warp.Volume.CLOSEST</span></code></a> or <code class="xref py py-attr docutils literal notranslate"><span class="pre">wp.Volume.LINEAR.</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_lookup_f">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_lookup_f</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.volume_lookup_f" title="Link to this definition">#</a></dt>
-<dd><p>Returns the value of voxel with coordinates <code class="docutils literal notranslate"><span class="pre">i</span></code>, <code class="docutils literal notranslate"><span class="pre">j</span></code>, <code class="docutils literal notranslate"><span class="pre">k</span></code>.
-If the voxel at this index does not exist, this function returns the background value</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_store_f">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_store_f</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.volume_store_f" title="Link to this definition">#</a></dt>
-<dd><p>Store <code class="docutils literal notranslate"><span class="pre">value</span></code> at the voxel with coordinates <code class="docutils literal notranslate"><span class="pre">i</span></code>, <code class="docutils literal notranslate"><span class="pre">j</span></code>, <code class="docutils literal notranslate"><span class="pre">k</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_sample_v">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_sample_v</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">uvw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">sampling_mode</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.volume_sample_v" title="Link to this definition">#</a></dt>
-<dd><p>Sample the vector volume given by <code class="docutils literal notranslate"><span class="pre">id</span></code> at the volume local-space point <code class="docutils literal notranslate"><span class="pre">uvw</span></code>.
-Interpolation should be <a class="reference internal" href="runtime.html#warp.Volume.CLOSEST" title="warp.Volume.CLOSEST"><code class="xref py py-attr docutils literal notranslate"><span class="pre">warp.Volume.CLOSEST</span></code></a> or <code class="xref py py-attr docutils literal notranslate"><span class="pre">wp.Volume.LINEAR.</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_lookup_v">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_lookup_v</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.volume_lookup_v" title="Link to this definition">#</a></dt>
-<dd><p>Returns the vector value of voxel with coordinates <code class="docutils literal notranslate"><span class="pre">i</span></code>, <code class="docutils literal notranslate"><span class="pre">j</span></code>, <code class="docutils literal notranslate"><span class="pre">k</span></code>.
-If the voxel at this index does not exist, this function returns the background value.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_store_v">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_store_v</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.volume_store_v" title="Link to this definition">#</a></dt>
-<dd><p>Store <code class="docutils literal notranslate"><span class="pre">value</span></code> at the voxel with coordinates <code class="docutils literal notranslate"><span class="pre">i</span></code>, <code class="docutils literal notranslate"><span class="pre">j</span></code>, <code class="docutils literal notranslate"><span class="pre">k</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_sample_i">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_sample_i</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">uvw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.volume_sample_i" title="Link to this definition">#</a></dt>
-<dd><p>Sample the <a class="reference internal" href="#warp.int32" title="warp.int32"><code class="xref py py-class docutils literal notranslate"><span class="pre">int32</span></code></a> volume given by <code class="docutils literal notranslate"><span class="pre">id</span></code> at the volume local-space point <code class="docutils literal notranslate"><span class="pre">uvw</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_lookup_i">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_lookup_i</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.volume_lookup_i" title="Link to this definition">#</a></dt>
-<dd><p>Returns the <a class="reference internal" href="#warp.int32" title="warp.int32"><code class="xref py py-class docutils literal notranslate"><span class="pre">int32</span></code></a> value of voxel with coordinates <code class="docutils literal notranslate"><span class="pre">i</span></code>, <code class="docutils literal notranslate"><span class="pre">j</span></code>, <code class="docutils literal notranslate"><span class="pre">k</span></code>.
-If the voxel at this index does not exist, this function returns the background value.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_store_i">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_store_i</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">i</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/constants.html#None" title="(in Python v3.12)"><span class="pre">None</span></a></span></span><a class="headerlink" href="#warp.volume_store_i" title="Link to this definition">#</a></dt>
-<dd><p>Store <code class="docutils literal notranslate"><span class="pre">value</span></code> at the voxel with coordinates <code class="docutils literal notranslate"><span class="pre">i</span></code>, <code class="docutils literal notranslate"><span class="pre">j</span></code>, <code class="docutils literal notranslate"><span class="pre">k</span></code>.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_index_to_world">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_index_to_world</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">uvw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.volume_index_to_world" title="Link to this definition">#</a></dt>
-<dd><p>Transform a point <code class="docutils literal notranslate"><span class="pre">uvw</span></code> defined in volume index space to world space given the volume’s intrinsic affine transformation.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_world_to_index">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_world_to_index</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.volume_world_to_index" title="Link to this definition">#</a></dt>
-<dd><p>Transform a point <code class="docutils literal notranslate"><span class="pre">xyz</span></code> defined in volume world space to the volume’s index space given the volume’s intrinsic affine transformation.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_index_to_world_dir">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_index_to_world_dir</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">uvw</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.volume_index_to_world_dir" title="Link to this definition">#</a></dt>
-<dd><p>Transform a direction <code class="docutils literal notranslate"><span class="pre">uvw</span></code> defined in volume index space to world space given the volume’s intrinsic affine transformation.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.volume_world_to_index_dir">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">volume_world_to_index_dir</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">id</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.volume_world_to_index_dir" title="Link to this definition">#</a></dt>
-<dd><p>Transform a direction <code class="docutils literal notranslate"><span class="pre">xyz</span></code> defined in volume world space to the volume’s index space given the volume’s intrinsic affine transformation.</p>
-</dd></dl>
-
-</section>
-<section id="random">
-<h2>Random<a class="headerlink" href="#random" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.rand_init">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">rand_init</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">seed</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></span><a class="headerlink" href="#warp.rand_init" title="Link to this definition">#</a></dt>
-<dd><p>Initialize a new random number generator given a user-defined seed. Returns a 32-bit integer representing the RNG state.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">rand_init</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">seed</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></span></dt>
-<dd><p>Initialize a new random number generator given a user-defined seed and an offset.
-This alternative constructor can be useful in parallel programs, where a kernel as a whole should share a seed,
-but each thread should generate uncorrelated values. In this case usage should be <code class="docutils literal notranslate"><span class="pre">r</span> <span class="pre">=</span> <span class="pre">rand_init(seed,</span> <span class="pre">tid)</span></code></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.randi">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">randi</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.randi" title="Link to this definition">#</a></dt>
-<dd><p>Return a random integer in the range [0, 2^32).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">randi</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">min</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span></dt>
-<dd><p>Return a random integer between [min, max).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.randf">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">randf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.randf" title="Link to this definition">#</a></dt>
-<dd><p>Return a random float between [0.0, 1.0).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">randf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">min</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">max</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Return a random float between [min, max).</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.randn">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">randn</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.randn" title="Link to this definition">#</a></dt>
-<dd><p>Sample a normal distribution.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_cdf">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_cdf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">cdf</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></span></span><a class="headerlink" href="#warp.sample_cdf" title="Link to this definition">#</a></dt>
-<dd><p>Inverse-transform sample a cumulative distribution function.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_triangle">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_triangle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></span><a class="headerlink" href="#warp.sample_triangle" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a triangle. Returns sample barycentric coordinates.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_ring">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_ring</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_ring" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a ring in the xy plane.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_disk">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_disk</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_disk" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a disk in the xy plane.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_sphere_surface">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_sphere_surface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_sphere_surface" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a unit sphere surface.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_sphere">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_sphere</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_sphere" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a unit sphere.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_hemisphere_surface">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_hemisphere_surface</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_hemisphere_surface" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a unit hemisphere surface.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_hemisphere">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_hemisphere</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_hemisphere" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a unit hemisphere.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_square">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_square</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_square" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a unit square.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sample_unit_cube">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sample_unit_cube</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span><a class="headerlink" href="#warp.sample_unit_cube" title="Link to this definition">#</a></dt>
-<dd><p>Uniformly sample a unit cube.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.poisson">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">poisson</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">lam</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></span><a class="headerlink" href="#warp.poisson" title="Link to this definition">#</a></dt>
-<dd><p>Generate a random sample from a Poisson distribution.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>state</strong> – RNG state</p></li>
-<li><p><strong>lam</strong> – The expected value of the distribution</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.noise">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">noise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.noise" title="Link to this definition">#</a></dt>
-<dd><p>Non-periodic Perlin-style noise in 1D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">noise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Non-periodic Perlin-style noise in 2D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">noise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Non-periodic Perlin-style noise in 3D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">noise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyzt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec4f" title="warp.vec4f"><span class="pre">vec4f</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Non-periodic Perlin-style noise in 4D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.pnoise">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">px</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span><a class="headerlink" href="#warp.pnoise" title="Link to this definition">#</a></dt>
-<dd><p>Periodic Perlin-style noise in 1D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">px</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">py</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Periodic Perlin-style noise in 2D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">px</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">py</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Periodic Perlin-style noise in 3D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyzt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec4f" title="warp.vec4f"><span class="pre">vec4f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">px</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">py</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">pt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><span class="pre">float</span></a></span></span></dt>
-<dd><p>Periodic Perlin-style noise in 4D.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.curlnoise">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">curlnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xy</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">octaves</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">lacunarity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">gain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec2f" title="warp.vec2f"><span class="pre">vec2f</span></a></span></span><a class="headerlink" href="#warp.curlnoise" title="Link to this definition">#</a></dt>
-<dd><p>Divergence-free vector field based on the gradient of a Perlin noise function. <a class="footnote-reference brackets" href="#id7" id="id4" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">curlnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyz</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">octaves</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">lacunarity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">gain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span></dt>
-<dd><p>Divergence-free vector field based on the curl of three Perlin noise functions. <a class="footnote-reference brackets" href="#id7" id="id5" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a></p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">curlnoise</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">state</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">xyzt</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.vec4f" title="warp.vec4f"><span class="pre">vec4f</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">octaves</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">lacunarity</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">gain</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.float32" title="warp.float32"><span class="pre">float32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.vec3f" title="warp.vec3f"><span class="pre">vec3f</span></a></span></span></dt>
-<dd><p>Divergence-free vector field based on the curl of three Perlin noise functions. <a class="footnote-reference brackets" href="#id7" id="id6" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a></p>
-</dd></dl>
-
-</section>
-<section id="operators">
-<h2>Operators<a class="headerlink" href="#operators" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.add">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.add" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">add</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sub">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.sub" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">sub</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.bit_and">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bit_and</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></span><a class="headerlink" href="#warp.bit_and" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.bit_or">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bit_or</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></span><a class="headerlink" href="#warp.bit_or" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.bit_xor">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">bit_xor</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></span><a class="headerlink" href="#warp.bit_xor" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.lshift">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">lshift</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></span><a class="headerlink" href="#warp.lshift" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.rshift">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">rshift</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></span><a class="headerlink" href="#warp.rshift" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.invert">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">invert</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Int" title="warp.Int"><span class="pre">Int</span></a></span></span><a class="headerlink" href="#warp.invert" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mul">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.mul" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Transformation" title="warp.Transformation"><span class="pre">Transformation</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.mod">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">mod</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.mod" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.div">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.div" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">div</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.floordiv">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">floordiv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.floordiv" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.pos">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pos</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.pos" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pos</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pos</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">pos</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.neg">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">neg</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a></span></span><a class="headerlink" href="#warp.neg" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">neg</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Vector" title="warp.Vector"><span class="pre">Vector</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">neg</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Quaternion" title="warp.Quaternion"><span class="pre">Quaternion</span></a><span class="p"><span class="pre">[</span></span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">neg</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.Matrix" title="warp.Matrix"><span class="pre">Matrix</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><span class="pre">Any</span><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference internal" href="#warp.Scalar" title="warp.Scalar"><span class="pre">Scalar</span></a><span class="p"><span class="pre">]</span></span></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.unot">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span><a class="headerlink" href="#warp.unot" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int8" title="warp.int8"><span class="pre">int8</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint8" title="warp.uint8"><span class="pre">uint8</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int16" title="warp.int16"><span class="pre">int16</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint16" title="warp.uint16"><span class="pre">uint16</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int32" title="warp.int32"><span class="pre">int32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint32" title="warp.uint32"><span class="pre">uint32</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.int64" title="warp.int64"><span class="pre">int64</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">b</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.uint64" title="warp.uint64"><span class="pre">uint64</span></a></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">unot</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="n"><a class="reference internal" href="#warp.Array" title="warp.Array"><span class="pre">Array</span></a><span class="p"><span class="pre">[</span></span><span class="pre">Any</span><span class="p"><span class="pre">]</span></span></span></em><span class="sig-paren">)</span> <span class="sig-return"><span class="sig-return-icon">&#x2192;</span> <span class="sig-return-typehint"><a class="reference internal" href="#warp.bool" title="warp.bool"><span class="pre">bool</span></a></span></span></dt>
-<dd></dd></dl>
-
-<p class="rubric">Footnotes</p>
-<aside class="footnote-list brackets">
-<aside class="footnote brackets" id="id7" role="doc-footnote">
-<span class="label"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></span>
-<span class="backrefs">(<a role="doc-backlink" href="#id1">1</a>,<a role="doc-backlink" href="#id2">2</a>,<a role="doc-backlink" href="#id3">3</a>,<a role="doc-backlink" href="#id4">4</a>,<a role="doc-backlink" href="#id5">5</a>,<a role="doc-backlink" href="#id6">6</a>)</span>
-<p>Note: function gradients not implemented for backpropagation.</p>
-</aside>
-</aside>
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="sim.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">warp.sim</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="runtime.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Runtime Reference</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Kernel Reference</a><ul>
-<li><a class="reference internal" href="#scalar-types">Scalar Types</a><ul>
-<li><a class="reference internal" href="#warp.int8"><code class="docutils literal notranslate"><span class="pre">int8</span></code></a></li>
-<li><a class="reference internal" href="#warp.uint8"><code class="docutils literal notranslate"><span class="pre">uint8</span></code></a></li>
-<li><a class="reference internal" href="#warp.int16"><code class="docutils literal notranslate"><span class="pre">int16</span></code></a></li>
-<li><a class="reference internal" href="#warp.uint16"><code class="docutils literal notranslate"><span class="pre">uint16</span></code></a></li>
-<li><a class="reference internal" href="#warp.int32"><code class="docutils literal notranslate"><span class="pre">int32</span></code></a></li>
-<li><a class="reference internal" href="#warp.uint32"><code class="docutils literal notranslate"><span class="pre">uint32</span></code></a></li>
-<li><a class="reference internal" href="#warp.int64"><code class="docutils literal notranslate"><span class="pre">int64</span></code></a></li>
-<li><a class="reference internal" href="#warp.uint64"><code class="docutils literal notranslate"><span class="pre">uint64</span></code></a></li>
-<li><a class="reference internal" href="#warp.float16"><code class="docutils literal notranslate"><span class="pre">float16</span></code></a></li>
-<li><a class="reference internal" href="#warp.float32"><code class="docutils literal notranslate"><span class="pre">float32</span></code></a></li>
-<li><a class="reference internal" href="#warp.float64"><code class="docutils literal notranslate"><span class="pre">float64</span></code></a></li>
-<li><a class="reference internal" href="#warp.bool"><code class="docutils literal notranslate"><span class="pre">bool</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#vector-types">Vector Types</a><ul>
-<li><a class="reference internal" href="#warp.vec2b"><code class="docutils literal notranslate"><span class="pre">vec2b</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2ub"><code class="docutils literal notranslate"><span class="pre">vec2ub</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2s"><code class="docutils literal notranslate"><span class="pre">vec2s</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2us"><code class="docutils literal notranslate"><span class="pre">vec2us</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2i"><code class="docutils literal notranslate"><span class="pre">vec2i</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2ui"><code class="docutils literal notranslate"><span class="pre">vec2ui</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2l"><code class="docutils literal notranslate"><span class="pre">vec2l</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2ul"><code class="docutils literal notranslate"><span class="pre">vec2ul</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2h"><code class="docutils literal notranslate"><span class="pre">vec2h</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2f"><code class="docutils literal notranslate"><span class="pre">vec2f</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec2d"><code class="docutils literal notranslate"><span class="pre">vec2d</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3b"><code class="docutils literal notranslate"><span class="pre">vec3b</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3ub"><code class="docutils literal notranslate"><span class="pre">vec3ub</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3s"><code class="docutils literal notranslate"><span class="pre">vec3s</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3us"><code class="docutils literal notranslate"><span class="pre">vec3us</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3i"><code class="docutils literal notranslate"><span class="pre">vec3i</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3ui"><code class="docutils literal notranslate"><span class="pre">vec3ui</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3l"><code class="docutils literal notranslate"><span class="pre">vec3l</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3ul"><code class="docutils literal notranslate"><span class="pre">vec3ul</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3h"><code class="docutils literal notranslate"><span class="pre">vec3h</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3f"><code class="docutils literal notranslate"><span class="pre">vec3f</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec3d"><code class="docutils literal notranslate"><span class="pre">vec3d</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4b"><code class="docutils literal notranslate"><span class="pre">vec4b</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4ub"><code class="docutils literal notranslate"><span class="pre">vec4ub</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4s"><code class="docutils literal notranslate"><span class="pre">vec4s</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4us"><code class="docutils literal notranslate"><span class="pre">vec4us</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4i"><code class="docutils literal notranslate"><span class="pre">vec4i</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4ui"><code class="docutils literal notranslate"><span class="pre">vec4ui</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4l"><code class="docutils literal notranslate"><span class="pre">vec4l</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4ul"><code class="docutils literal notranslate"><span class="pre">vec4ul</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4h"><code class="docutils literal notranslate"><span class="pre">vec4h</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4f"><code class="docutils literal notranslate"><span class="pre">vec4f</span></code></a></li>
-<li><a class="reference internal" href="#warp.vec4d"><code class="docutils literal notranslate"><span class="pre">vec4d</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat22h"><code class="docutils literal notranslate"><span class="pre">mat22h</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat22f"><code class="docutils literal notranslate"><span class="pre">mat22f</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat22d"><code class="docutils literal notranslate"><span class="pre">mat22d</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat33h"><code class="docutils literal notranslate"><span class="pre">mat33h</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat33f"><code class="docutils literal notranslate"><span class="pre">mat33f</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat33d"><code class="docutils literal notranslate"><span class="pre">mat33d</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat44h"><code class="docutils literal notranslate"><span class="pre">mat44h</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat44f"><code class="docutils literal notranslate"><span class="pre">mat44f</span></code></a></li>
-<li><a class="reference internal" href="#warp.mat44d"><code class="docutils literal notranslate"><span class="pre">mat44d</span></code></a></li>
-<li><a class="reference internal" href="#warp.quath"><code class="docutils literal notranslate"><span class="pre">quath</span></code></a></li>
-<li><a class="reference internal" href="#warp.quatf"><code class="docutils literal notranslate"><span class="pre">quatf</span></code></a></li>
-<li><a class="reference internal" href="#warp.quatd"><code class="docutils literal notranslate"><span class="pre">quatd</span></code></a></li>
-<li><a class="reference internal" href="#warp.transformh"><code class="docutils literal notranslate"><span class="pre">transformh</span></code></a></li>
-<li><a class="reference internal" href="#warp.transformf"><code class="docutils literal notranslate"><span class="pre">transformf</span></code></a></li>
-<li><a class="reference internal" href="#warp.transformd"><code class="docutils literal notranslate"><span class="pre">transformd</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_vectorh"><code class="docutils literal notranslate"><span class="pre">spatial_vectorh</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_vectorf"><code class="docutils literal notranslate"><span class="pre">spatial_vectorf</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_vectord"><code class="docutils literal notranslate"><span class="pre">spatial_vectord</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_matrixh"><code class="docutils literal notranslate"><span class="pre">spatial_matrixh</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_matrixf"><code class="docutils literal notranslate"><span class="pre">spatial_matrixf</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_matrixd"><code class="docutils literal notranslate"><span class="pre">spatial_matrixd</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#generic-types">Generic Types</a><ul>
-<li><a class="reference internal" href="#warp.Int"><code class="docutils literal notranslate"><span class="pre">Int</span></code></a></li>
-<li><a class="reference internal" href="#warp.Float"><code class="docutils literal notranslate"><span class="pre">Float</span></code></a></li>
-<li><a class="reference internal" href="#warp.Scalar"><code class="docutils literal notranslate"><span class="pre">Scalar</span></code></a></li>
-<li><a class="reference internal" href="#warp.Vector"><code class="docutils literal notranslate"><span class="pre">Vector</span></code></a></li>
-<li><a class="reference internal" href="#warp.Matrix"><code class="docutils literal notranslate"><span class="pre">Matrix</span></code></a></li>
-<li><a class="reference internal" href="#warp.Quaternion"><code class="docutils literal notranslate"><span class="pre">Quaternion</span></code></a></li>
-<li><a class="reference internal" href="#warp.Transformation"><code class="docutils literal notranslate"><span class="pre">Transformation</span></code></a></li>
-<li><a class="reference internal" href="#warp.Array"><code class="docutils literal notranslate"><span class="pre">Array</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#scalar-math">Scalar Math</a><ul>
-<li><a class="reference internal" href="#warp.min"><code class="docutils literal notranslate"><span class="pre">min()</span></code></a></li>
-<li><a class="reference internal" href="#warp.max"><code class="docutils literal notranslate"><span class="pre">max()</span></code></a></li>
-<li><a class="reference internal" href="#warp.clamp"><code class="docutils literal notranslate"><span class="pre">clamp()</span></code></a></li>
-<li><a class="reference internal" href="#warp.abs"><code class="docutils literal notranslate"><span class="pre">abs()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sign"><code class="docutils literal notranslate"><span class="pre">sign()</span></code></a></li>
-<li><a class="reference internal" href="#warp.step"><code class="docutils literal notranslate"><span class="pre">step()</span></code></a></li>
-<li><a class="reference internal" href="#warp.nonzero"><code class="docutils literal notranslate"><span class="pre">nonzero()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sin"><code class="docutils literal notranslate"><span class="pre">sin()</span></code></a></li>
-<li><a class="reference internal" href="#warp.cos"><code class="docutils literal notranslate"><span class="pre">cos()</span></code></a></li>
-<li><a class="reference internal" href="#warp.acos"><code class="docutils literal notranslate"><span class="pre">acos()</span></code></a></li>
-<li><a class="reference internal" href="#warp.asin"><code class="docutils literal notranslate"><span class="pre">asin()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sqrt"><code class="docutils literal notranslate"><span class="pre">sqrt()</span></code></a></li>
-<li><a class="reference internal" href="#warp.cbrt"><code class="docutils literal notranslate"><span class="pre">cbrt()</span></code></a></li>
-<li><a class="reference internal" href="#warp.tan"><code class="docutils literal notranslate"><span class="pre">tan()</span></code></a></li>
-<li><a class="reference internal" href="#warp.atan"><code class="docutils literal notranslate"><span class="pre">atan()</span></code></a></li>
-<li><a class="reference internal" href="#warp.atan2"><code class="docutils literal notranslate"><span class="pre">atan2()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sinh"><code class="docutils literal notranslate"><span class="pre">sinh()</span></code></a></li>
-<li><a class="reference internal" href="#warp.cosh"><code class="docutils literal notranslate"><span class="pre">cosh()</span></code></a></li>
-<li><a class="reference internal" href="#warp.tanh"><code class="docutils literal notranslate"><span class="pre">tanh()</span></code></a></li>
-<li><a class="reference internal" href="#warp.degrees"><code class="docutils literal notranslate"><span class="pre">degrees()</span></code></a></li>
-<li><a class="reference internal" href="#warp.radians"><code class="docutils literal notranslate"><span class="pre">radians()</span></code></a></li>
-<li><a class="reference internal" href="#warp.log"><code class="docutils literal notranslate"><span class="pre">log()</span></code></a></li>
-<li><a class="reference internal" href="#warp.log2"><code class="docutils literal notranslate"><span class="pre">log2()</span></code></a></li>
-<li><a class="reference internal" href="#warp.log10"><code class="docutils literal notranslate"><span class="pre">log10()</span></code></a></li>
-<li><a class="reference internal" href="#warp.exp"><code class="docutils literal notranslate"><span class="pre">exp()</span></code></a></li>
-<li><a class="reference internal" href="#warp.pow"><code class="docutils literal notranslate"><span class="pre">pow()</span></code></a></li>
-<li><a class="reference internal" href="#warp.round"><code class="docutils literal notranslate"><span class="pre">round()</span></code></a></li>
-<li><a class="reference internal" href="#warp.rint"><code class="docutils literal notranslate"><span class="pre">rint()</span></code></a></li>
-<li><a class="reference internal" href="#warp.trunc"><code class="docutils literal notranslate"><span class="pre">trunc()</span></code></a></li>
-<li><a class="reference internal" href="#warp.floor"><code class="docutils literal notranslate"><span class="pre">floor()</span></code></a></li>
-<li><a class="reference internal" href="#warp.ceil"><code class="docutils literal notranslate"><span class="pre">ceil()</span></code></a></li>
-<li><a class="reference internal" href="#warp.frac"><code class="docutils literal notranslate"><span class="pre">frac()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#vector-math">Vector Math</a><ul>
-<li><a class="reference internal" href="#warp.dot"><code class="docutils literal notranslate"><span class="pre">dot()</span></code></a></li>
-<li><a class="reference internal" href="#warp.ddot"><code class="docutils literal notranslate"><span class="pre">ddot()</span></code></a></li>
-<li><a class="reference internal" href="#warp.argmin"><code class="docutils literal notranslate"><span class="pre">argmin()</span></code></a></li>
-<li><a class="reference internal" href="#warp.argmax"><code class="docutils literal notranslate"><span class="pre">argmax()</span></code></a></li>
-<li><a class="reference internal" href="#warp.outer"><code class="docutils literal notranslate"><span class="pre">outer()</span></code></a></li>
-<li><a class="reference internal" href="#warp.cross"><code class="docutils literal notranslate"><span class="pre">cross()</span></code></a></li>
-<li><a class="reference internal" href="#warp.skew"><code class="docutils literal notranslate"><span class="pre">skew()</span></code></a></li>
-<li><a class="reference internal" href="#warp.length"><code class="docutils literal notranslate"><span class="pre">length()</span></code></a></li>
-<li><a class="reference internal" href="#warp.length_sq"><code class="docutils literal notranslate"><span class="pre">length_sq()</span></code></a></li>
-<li><a class="reference internal" href="#warp.normalize"><code class="docutils literal notranslate"><span class="pre">normalize()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transpose"><code class="docutils literal notranslate"><span class="pre">transpose()</span></code></a></li>
-<li><a class="reference internal" href="#warp.inverse"><code class="docutils literal notranslate"><span class="pre">inverse()</span></code></a></li>
-<li><a class="reference internal" href="#warp.determinant"><code class="docutils literal notranslate"><span class="pre">determinant()</span></code></a></li>
-<li><a class="reference internal" href="#warp.trace"><code class="docutils literal notranslate"><span class="pre">trace()</span></code></a></li>
-<li><a class="reference internal" href="#warp.diag"><code class="docutils literal notranslate"><span class="pre">diag()</span></code></a></li>
-<li><a class="reference internal" href="#warp.get_diag"><code class="docutils literal notranslate"><span class="pre">get_diag()</span></code></a></li>
-<li><a class="reference internal" href="#warp.cw_mul"><code class="docutils literal notranslate"><span class="pre">cw_mul()</span></code></a></li>
-<li><a class="reference internal" href="#warp.cw_div"><code class="docutils literal notranslate"><span class="pre">cw_div()</span></code></a></li>
-<li><a class="reference internal" href="#warp.vector"><code class="docutils literal notranslate"><span class="pre">vector()</span></code></a></li>
-<li><a class="reference internal" href="#warp.matrix"><code class="docutils literal notranslate"><span class="pre">matrix()</span></code></a></li>
-<li><a class="reference internal" href="#warp.identity"><code class="docutils literal notranslate"><span class="pre">identity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.svd3"><code class="docutils literal notranslate"><span class="pre">svd3()</span></code></a></li>
-<li><a class="reference internal" href="#warp.qr3"><code class="docutils literal notranslate"><span class="pre">qr3()</span></code></a></li>
-<li><a class="reference internal" href="#warp.eig3"><code class="docutils literal notranslate"><span class="pre">eig3()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#other">Other</a><ul>
-<li><a class="reference internal" href="#warp.lower_bound"><code class="docutils literal notranslate"><span class="pre">lower_bound()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#quaternion-math">Quaternion Math</a><ul>
-<li><a class="reference internal" href="#warp.quaternion"><code class="docutils literal notranslate"><span class="pre">quaternion()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_identity"><code class="docutils literal notranslate"><span class="pre">quat_identity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_from_axis_angle"><code class="docutils literal notranslate"><span class="pre">quat_from_axis_angle()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_to_axis_angle"><code class="docutils literal notranslate"><span class="pre">quat_to_axis_angle()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_from_matrix"><code class="docutils literal notranslate"><span class="pre">quat_from_matrix()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_rpy"><code class="docutils literal notranslate"><span class="pre">quat_rpy()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_inverse"><code class="docutils literal notranslate"><span class="pre">quat_inverse()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_rotate"><code class="docutils literal notranslate"><span class="pre">quat_rotate()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_rotate_inv"><code class="docutils literal notranslate"><span class="pre">quat_rotate_inv()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_slerp"><code class="docutils literal notranslate"><span class="pre">quat_slerp()</span></code></a></li>
-<li><a class="reference internal" href="#warp.quat_to_matrix"><code class="docutils literal notranslate"><span class="pre">quat_to_matrix()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#transformations">Transformations</a><ul>
-<li><a class="reference internal" href="#warp.transformation"><code class="docutils literal notranslate"><span class="pre">transformation()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_identity"><code class="docutils literal notranslate"><span class="pre">transform_identity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_get_translation"><code class="docutils literal notranslate"><span class="pre">transform_get_translation()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_get_rotation"><code class="docutils literal notranslate"><span class="pre">transform_get_rotation()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_multiply"><code class="docutils literal notranslate"><span class="pre">transform_multiply()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_point"><code class="docutils literal notranslate"><span class="pre">transform_point()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_vector"><code class="docutils literal notranslate"><span class="pre">transform_vector()</span></code></a></li>
-<li><a class="reference internal" href="#warp.transform_inverse"><code class="docutils literal notranslate"><span class="pre">transform_inverse()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#spatial-math">Spatial Math</a><ul>
-<li><a class="reference internal" href="#warp.spatial_adjoint"><code class="docutils literal notranslate"><span class="pre">spatial_adjoint()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_dot"><code class="docutils literal notranslate"><span class="pre">spatial_dot()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_cross"><code class="docutils literal notranslate"><span class="pre">spatial_cross()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_cross_dual"><code class="docutils literal notranslate"><span class="pre">spatial_cross_dual()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_top"><code class="docutils literal notranslate"><span class="pre">spatial_top()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_bottom"><code class="docutils literal notranslate"><span class="pre">spatial_bottom()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_jacobian"><code class="docutils literal notranslate"><span class="pre">spatial_jacobian()</span></code></a></li>
-<li><a class="reference internal" href="#warp.spatial_mass"><code class="docutils literal notranslate"><span class="pre">spatial_mass()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#utility">Utility</a><ul>
-<li><a class="reference internal" href="#warp.mlp"><code class="docutils literal notranslate"><span class="pre">mlp()</span></code></a></li>
-<li><a class="reference internal" href="#warp.printf"><code class="docutils literal notranslate"><span class="pre">printf()</span></code></a></li>
-<li><a class="reference internal" href="#warp.print"><code class="docutils literal notranslate"><span class="pre">print()</span></code></a></li>
-<li><a class="reference internal" href="#warp.breakpoint"><code class="docutils literal notranslate"><span class="pre">breakpoint()</span></code></a></li>
-<li><a class="reference internal" href="#warp.tid"><code class="docutils literal notranslate"><span class="pre">tid()</span></code></a></li>
-<li><a class="reference internal" href="#warp.select"><code class="docutils literal notranslate"><span class="pre">select()</span></code></a></li>
-<li><a class="reference internal" href="#warp.atomic_add"><code class="docutils literal notranslate"><span class="pre">atomic_add()</span></code></a></li>
-<li><a class="reference internal" href="#warp.atomic_sub"><code class="docutils literal notranslate"><span class="pre">atomic_sub()</span></code></a></li>
-<li><a class="reference internal" href="#warp.atomic_min"><code class="docutils literal notranslate"><span class="pre">atomic_min()</span></code></a></li>
-<li><a class="reference internal" href="#warp.atomic_max"><code class="docutils literal notranslate"><span class="pre">atomic_max()</span></code></a></li>
-<li><a class="reference internal" href="#warp.lerp"><code class="docutils literal notranslate"><span class="pre">lerp()</span></code></a></li>
-<li><a class="reference internal" href="#warp.smoothstep"><code class="docutils literal notranslate"><span class="pre">smoothstep()</span></code></a></li>
-<li><a class="reference internal" href="#warp.expect_near"><code class="docutils literal notranslate"><span class="pre">expect_near()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#geometry">Geometry</a><ul>
-<li><a class="reference internal" href="#warp.bvh_query_aabb"><code class="docutils literal notranslate"><span class="pre">bvh_query_aabb()</span></code></a></li>
-<li><a class="reference internal" href="#warp.bvh_query_ray"><code class="docutils literal notranslate"><span class="pre">bvh_query_ray()</span></code></a></li>
-<li><a class="reference internal" href="#warp.bvh_query_next"><code class="docutils literal notranslate"><span class="pre">bvh_query_next()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_point"><code class="docutils literal notranslate"><span class="pre">mesh_query_point()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_point_no_sign"><code class="docutils literal notranslate"><span class="pre">mesh_query_point_no_sign()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_furthest_point_no_sign"><code class="docutils literal notranslate"><span class="pre">mesh_query_furthest_point_no_sign()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_point_sign_normal"><code class="docutils literal notranslate"><span class="pre">mesh_query_point_sign_normal()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_point_sign_winding_number"><code class="docutils literal notranslate"><span class="pre">mesh_query_point_sign_winding_number()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_ray"><code class="docutils literal notranslate"><span class="pre">mesh_query_ray()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_aabb"><code class="docutils literal notranslate"><span class="pre">mesh_query_aabb()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_query_aabb_next"><code class="docutils literal notranslate"><span class="pre">mesh_query_aabb_next()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_eval_position"><code class="docutils literal notranslate"><span class="pre">mesh_eval_position()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_eval_velocity"><code class="docutils literal notranslate"><span class="pre">mesh_eval_velocity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.hash_grid_query"><code class="docutils literal notranslate"><span class="pre">hash_grid_query()</span></code></a></li>
-<li><a class="reference internal" href="#warp.hash_grid_query_next"><code class="docutils literal notranslate"><span class="pre">hash_grid_query_next()</span></code></a></li>
-<li><a class="reference internal" href="#warp.hash_grid_point_id"><code class="docutils literal notranslate"><span class="pre">hash_grid_point_id()</span></code></a></li>
-<li><a class="reference internal" href="#warp.intersect_tri_tri"><code class="docutils literal notranslate"><span class="pre">intersect_tri_tri()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_get"><code class="docutils literal notranslate"><span class="pre">mesh_get()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_eval_face_normal"><code class="docutils literal notranslate"><span class="pre">mesh_eval_face_normal()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_get_point"><code class="docutils literal notranslate"><span class="pre">mesh_get_point()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_get_velocity"><code class="docutils literal notranslate"><span class="pre">mesh_get_velocity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mesh_get_index"><code class="docutils literal notranslate"><span class="pre">mesh_get_index()</span></code></a></li>
-<li><a class="reference internal" href="#warp.closest_point_edge_edge"><code class="docutils literal notranslate"><span class="pre">closest_point_edge_edge()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#volumes">Volumes</a><ul>
-<li><a class="reference internal" href="#warp.volume_sample_f"><code class="docutils literal notranslate"><span class="pre">volume_sample_f()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_sample_grad_f"><code class="docutils literal notranslate"><span class="pre">volume_sample_grad_f()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_lookup_f"><code class="docutils literal notranslate"><span class="pre">volume_lookup_f()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_store_f"><code class="docutils literal notranslate"><span class="pre">volume_store_f()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_sample_v"><code class="docutils literal notranslate"><span class="pre">volume_sample_v()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_lookup_v"><code class="docutils literal notranslate"><span class="pre">volume_lookup_v()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_store_v"><code class="docutils literal notranslate"><span class="pre">volume_store_v()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_sample_i"><code class="docutils literal notranslate"><span class="pre">volume_sample_i()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_lookup_i"><code class="docutils literal notranslate"><span class="pre">volume_lookup_i()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_store_i"><code class="docutils literal notranslate"><span class="pre">volume_store_i()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_index_to_world"><code class="docutils literal notranslate"><span class="pre">volume_index_to_world()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_world_to_index"><code class="docutils literal notranslate"><span class="pre">volume_world_to_index()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_index_to_world_dir"><code class="docutils literal notranslate"><span class="pre">volume_index_to_world_dir()</span></code></a></li>
-<li><a class="reference internal" href="#warp.volume_world_to_index_dir"><code class="docutils literal notranslate"><span class="pre">volume_world_to_index_dir()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#random">Random</a><ul>
-<li><a class="reference internal" href="#warp.rand_init"><code class="docutils literal notranslate"><span class="pre">rand_init()</span></code></a></li>
-<li><a class="reference internal" href="#warp.randi"><code class="docutils literal notranslate"><span class="pre">randi()</span></code></a></li>
-<li><a class="reference internal" href="#warp.randf"><code class="docutils literal notranslate"><span class="pre">randf()</span></code></a></li>
-<li><a class="reference internal" href="#warp.randn"><code class="docutils literal notranslate"><span class="pre">randn()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_cdf"><code class="docutils literal notranslate"><span class="pre">sample_cdf()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_triangle"><code class="docutils literal notranslate"><span class="pre">sample_triangle()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_ring"><code class="docutils literal notranslate"><span class="pre">sample_unit_ring()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_disk"><code class="docutils literal notranslate"><span class="pre">sample_unit_disk()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_sphere_surface"><code class="docutils literal notranslate"><span class="pre">sample_unit_sphere_surface()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_sphere"><code class="docutils literal notranslate"><span class="pre">sample_unit_sphere()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_hemisphere_surface"><code class="docutils literal notranslate"><span class="pre">sample_unit_hemisphere_surface()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_hemisphere"><code class="docutils literal notranslate"><span class="pre">sample_unit_hemisphere()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_square"><code class="docutils literal notranslate"><span class="pre">sample_unit_square()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sample_unit_cube"><code class="docutils literal notranslate"><span class="pre">sample_unit_cube()</span></code></a></li>
-<li><a class="reference internal" href="#warp.poisson"><code class="docutils literal notranslate"><span class="pre">poisson()</span></code></a></li>
-<li><a class="reference internal" href="#warp.noise"><code class="docutils literal notranslate"><span class="pre">noise()</span></code></a></li>
-<li><a class="reference internal" href="#warp.pnoise"><code class="docutils literal notranslate"><span class="pre">pnoise()</span></code></a></li>
-<li><a class="reference internal" href="#warp.curlnoise"><code class="docutils literal notranslate"><span class="pre">curlnoise()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#operators">Operators</a><ul>
-<li><a class="reference internal" href="#warp.add"><code class="docutils literal notranslate"><span class="pre">add()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sub"><code class="docutils literal notranslate"><span class="pre">sub()</span></code></a></li>
-<li><a class="reference internal" href="#warp.bit_and"><code class="docutils literal notranslate"><span class="pre">bit_and()</span></code></a></li>
-<li><a class="reference internal" href="#warp.bit_or"><code class="docutils literal notranslate"><span class="pre">bit_or()</span></code></a></li>
-<li><a class="reference internal" href="#warp.bit_xor"><code class="docutils literal notranslate"><span class="pre">bit_xor()</span></code></a></li>
-<li><a class="reference internal" href="#warp.lshift"><code class="docutils literal notranslate"><span class="pre">lshift()</span></code></a></li>
-<li><a class="reference internal" href="#warp.rshift"><code class="docutils literal notranslate"><span class="pre">rshift()</span></code></a></li>
-<li><a class="reference internal" href="#warp.invert"><code class="docutils literal notranslate"><span class="pre">invert()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mul"><code class="docutils literal notranslate"><span class="pre">mul()</span></code></a></li>
-<li><a class="reference internal" href="#warp.mod"><code class="docutils literal notranslate"><span class="pre">mod()</span></code></a></li>
-<li><a class="reference internal" href="#warp.div"><code class="docutils literal notranslate"><span class="pre">div()</span></code></a></li>
-<li><a class="reference internal" href="#warp.floordiv"><code class="docutils literal notranslate"><span class="pre">floordiv()</span></code></a></li>
-<li><a class="reference internal" href="#warp.pos"><code class="docutils literal notranslate"><span class="pre">pos()</span></code></a></li>
-<li><a class="reference internal" href="#warp.neg"><code class="docutils literal notranslate"><span class="pre">neg()</span></code></a></li>
-<li><a class="reference internal" href="#warp.unot"><code class="docutils literal notranslate"><span class="pre">unot()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    <script async="async" src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/interoperability.html b/docs/_build/html/modules/interoperability.html
deleted file mode 100644
index 0cf78d837..000000000
--- a/docs/_build/html/modules/interoperability.html
+++ /dev/null
@@ -1,659 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="Runtime Settings" href="../configuration.html" /><link rel="prev" title="Devices" href="devices.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Interoperability - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="interoperability">
-<h1>Interoperability<a class="headerlink" href="#interoperability" title="Link to this heading">#</a></h1>
-<p>Warp can interop with other Python-based frameworks such as NumPy through standard interface protocols.</p>
-<section id="numpy">
-<h2>NumPy<a class="headerlink" href="#numpy" title="Link to this heading">#</a></h2>
-<p>Warp arrays may be converted to a NumPy array through the <code class="docutils literal notranslate"><span class="pre">warp.array.numpy()</span></code> method. When the Warp array lives on
-the <code class="docutils literal notranslate"><span class="pre">cpu</span></code> device this will return a zero-copy view onto the underlying Warp allocation. If the array lives on a
-<code class="docutils literal notranslate"><span class="pre">cuda</span></code> device then it will first be copied back to a temporary buffer and copied to NumPy.</p>
-<p>Warp CPU arrays also implement  the <code class="docutils literal notranslate"><span class="pre">__array_interface__</span></code> protocol and so can be used to construct NumPy arrays
-directly:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">w</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">([</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">],</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">w</span><span class="p">)</span>
-<span class="nb">print</span><span class="p">(</span><span class="n">a</span><span class="p">)</span>
-<span class="o">&gt;</span> <span class="p">[</span><span class="mf">1.</span> <span class="mf">2.</span> <span class="mf">3.</span><span class="p">]</span>
-</pre></div>
-</div>
-</section>
-<section id="pytorch">
-<span id="pytorch-interop"></span><h2>PyTorch<a class="headerlink" href="#pytorch" title="Link to this heading">#</a></h2>
-<p>Warp provides helper functions to convert arrays to/from PyTorch. Please see the <code class="docutils literal notranslate"><span class="pre">warp.torch</span></code> module for more details. Example usage is shown below:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp.torch</span>
-
-<span class="n">w</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">([</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">],</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-
-<span class="c1"># convert to Torch tensor</span>
-<span class="n">t</span> <span class="o">=</span> <span class="n">warp</span><span class="o">.</span><span class="n">to_torch</span><span class="p">(</span><span class="n">w</span><span class="p">)</span>
-
-<span class="c1"># convert from Torch tensor</span>
-<span class="n">w</span> <span class="o">=</span> <span class="n">warp</span><span class="o">.</span><span class="n">from_torch</span><span class="p">(</span><span class="n">t</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>These helper functions allow the conversion of Warp arrays to/from PyTorch tensors without copying the underlying data.
-At the same time, if available, gradient arrays and tensors are converted to/from PyTorch autograd tensors, allowing the use of Warp arrays
-in PyTorch autograd computations.</p>
-<section id="example-optimization-using-warp-from-torch">
-<h3>Example: Optimization using <code class="docutils literal notranslate"><span class="pre">warp.from_torch()</span></code><a class="headerlink" href="#example-optimization-using-warp-from-torch" title="Link to this heading">#</a></h3>
-<p>An example usage of minimizing a loss function over an array of 2D points written in Warp via PyTorch’s Adam optimizer using <code class="docutils literal notranslate"><span class="pre">warp.from_torch</span></code> is as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span><span class="p">()</span>
-<span class="k">def</span> <span class="nf">loss</span><span class="p">(</span><span class="n">xs</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">ndim</span><span class="o">=</span><span class="mi">2</span><span class="p">),</span> <span class="n">l</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">atomic_add</span><span class="p">(</span><span class="n">l</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">xs</span><span class="p">[</span><span class="n">tid</span><span class="p">,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">**</span> <span class="mf">2.0</span> <span class="o">+</span> <span class="n">xs</span><span class="p">[</span><span class="n">tid</span><span class="p">,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">**</span> <span class="mf">2.0</span><span class="p">)</span>
-
-<span class="c1"># indicate requires_grad so that Warp can accumulate gradients in the grad buffers</span>
-<span class="n">xs</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="n">l</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="n">opt</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">optim</span><span class="o">.</span><span class="n">Adam</span><span class="p">([</span><span class="n">xs</span><span class="p">],</span> <span class="n">lr</span><span class="o">=</span><span class="mf">0.1</span><span class="p">)</span>
-
-<span class="n">wp_xs</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">from_torch</span><span class="p">(</span><span class="n">xs</span><span class="p">)</span>
-<span class="n">wp_l</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">from_torch</span><span class="p">(</span><span class="n">l</span><span class="p">)</span>
-
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="c1"># record the loss function kernel launch on the tape</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">loss</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">xs</span><span class="p">),</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">wp_xs</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">wp_l</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">wp_xs</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">500</span><span class="p">):</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">zero</span><span class="p">()</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">loss</span><span class="o">=</span><span class="n">wp_l</span><span class="p">)</span>  <span class="c1"># compute gradients</span>
-    <span class="c1"># now xs.grad will be populated with the gradients computed by Warp</span>
-    <span class="n">opt</span><span class="o">.</span><span class="n">step</span><span class="p">()</span>  <span class="c1"># update xs (and thereby wp_xs)</span>
-
-    <span class="c1"># these lines are only needed for evaluating the loss</span>
-    <span class="c1"># (the optimization just needs the gradient, not the loss value)</span>
-    <span class="n">wp_l</span><span class="o">.</span><span class="n">zero_</span><span class="p">()</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">loss</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">xs</span><span class="p">),</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">wp_xs</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">wp_l</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">wp_xs</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-    <span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="se">\t</span><span class="s2">loss: </span><span class="si">{</span><span class="n">l</span><span class="o">.</span><span class="n">item</span><span class="p">()</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-<section id="example-optimization-using-warp-to-torch">
-<h3>Example: Optimization using <code class="docutils literal notranslate"><span class="pre">warp.to_torch</span></code><a class="headerlink" href="#example-optimization-using-warp-to-torch" title="Link to this heading">#</a></h3>
-<p>Less code is needed when we declare the optimization variables directly in Warp and use <code class="docutils literal notranslate"><span class="pre">warp.to_torch</span></code> to convert them to PyTorch tensors.
-Here, we revisit the same example from above where now only a single conversion to a torch tensor is needed to supply Adam with the optimization variables:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">torch</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span><span class="p">()</span>
-<span class="k">def</span> <span class="nf">loss</span><span class="p">(</span><span class="n">xs</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">ndim</span><span class="o">=</span><span class="mi">2</span><span class="p">),</span> <span class="n">l</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">atomic_add</span><span class="p">(</span><span class="n">l</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">xs</span><span class="p">[</span><span class="n">tid</span><span class="p">,</span> <span class="mi">0</span><span class="p">]</span> <span class="o">**</span> <span class="mf">2.0</span> <span class="o">+</span> <span class="n">xs</span><span class="p">[</span><span class="n">tid</span><span class="p">,</span> <span class="mi">1</span><span class="p">]</span> <span class="o">**</span> <span class="mf">2.0</span><span class="p">)</span>
-
-<span class="c1"># initialize the optimization variables in Warp</span>
-<span class="n">xs</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randn</span><span class="p">(</span><span class="mi">100</span><span class="p">,</span> <span class="mi">2</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="n">l</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="c1"># just a single wp.to_torch call is needed, Adam optimizes using the Warp array gradients</span>
-<span class="n">opt</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">optim</span><span class="o">.</span><span class="n">Adam</span><span class="p">([</span><span class="n">wp</span><span class="o">.</span><span class="n">to_torch</span><span class="p">(</span><span class="n">xs</span><span class="p">)],</span> <span class="n">lr</span><span class="o">=</span><span class="mf">0.1</span><span class="p">)</span>
-
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">loss</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">xs</span><span class="p">),</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">xs</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">l</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">xs</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">500</span><span class="p">):</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">zero</span><span class="p">()</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">loss</span><span class="o">=</span><span class="n">l</span><span class="p">)</span>
-    <span class="n">opt</span><span class="o">.</span><span class="n">step</span><span class="p">()</span>
-
-    <span class="n">l</span><span class="o">.</span><span class="n">zero_</span><span class="p">()</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">loss</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">xs</span><span class="p">),</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">xs</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">l</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">xs</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
-    <span class="nb">print</span><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="se">\t</span><span class="s2">loss: </span><span class="si">{</span><span class="n">l</span><span class="o">.</span><span class="n">numpy</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<dl class="py function" id="module-warp.torch">
-<dt class="sig sig-object py" id="warp.torch.device_from_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">device_from_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">torch_device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.device_from_torch" title="Link to this definition">#</a></dt>
-<dd><p>Return the warp device corresponding to a torch device.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.device_to_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">device_to_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">wp_device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.device_to_torch" title="Link to this definition">#</a></dt>
-<dd><p>Return the torch device corresponding to a warp device.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.dtype_from_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">dtype_from_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">torch_dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.dtype_from_torch" title="Link to this definition">#</a></dt>
-<dd><p>Return the Warp dtype corresponding to a torch dtype.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.dtype_is_compatible">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">dtype_is_compatible</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">torch_dtype</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">warp_dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.dtype_is_compatible" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates whether the given torch dtype is compatible with the given warp dtype.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.from_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">from_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">t</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.from_torch" title="Link to this definition">#</a></dt>
-<dd><p>Wrap a PyTorch tensor to a Warp array without copying the data.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>t</strong> (<em>torch.Tensor</em>) – The torch tensor to wrap.</p></li>
-<li><p><strong>dtype</strong> (<em>warp.dtype</em><em>, </em><em>optional</em>) – The target data type of the resulting Warp array. Defaults to the tensor value type mapped to a Warp array value type.</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em>, </em><em>optional</em>) – Whether the resulting array should wrap the tensor’s gradient, if it exists (the grad tensor will be allocated otherwise). Defaults to the tensor’s <cite>requires_grad</cite> value.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The wrapped array.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">warp.array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.to_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">to_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.to_torch" title="Link to this definition">#</a></dt>
-<dd><p>Convert a Warp array to a PyTorch tensor without copying the data.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>a</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.array"><em>warp.array</em></a>) – The Warp array to convert.</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em>, </em><em>optional</em>) – Whether the resulting tensor should convert the array’s gradient, if it exists, to a grad tensor. Defaults to the array’s <cite>requires_grad</cite> value.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The converted tensor.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>torch.Tensor</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.stream_from_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">stream_from_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stream_or_device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.stream_from_torch" title="Link to this definition">#</a></dt>
-<dd><p>Convert from a PyTorch CUDA stream to a Warp.Stream.</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.torch.stream_to_torch">
-<span class="sig-prename descclassname"><span class="pre">warp.torch.</span></span><span class="sig-name descname"><span class="pre">stream_to_torch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">stream_or_device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.torch.stream_to_torch" title="Link to this definition">#</a></dt>
-<dd><p>Convert from a Warp.Stream to a PyTorch CUDA stream.</p>
-</dd></dl>
-
-</section>
-</section>
-<section id="cupy-numba">
-<h2>CuPy/Numba<a class="headerlink" href="#cupy-numba" title="Link to this heading">#</a></h2>
-<p>Warp GPU arrays support the <code class="docutils literal notranslate"><span class="pre">__cuda_array_interface__</span></code> protocol for sharing data with other Python GPU frameworks.
-Currently this is one-directional, so that Warp arrays can be used as input to any framework that also supports the
-<code class="docutils literal notranslate"><span class="pre">__cuda_array_interface__</span></code> protocol, but not the other way around.</p>
-</section>
-<section id="jax">
-<span id="jax-interop"></span><h2>JAX<a class="headerlink" href="#jax" title="Link to this heading">#</a></h2>
-<p>Interoperability with JAX arrays is supported through the following methods.
-Internally these use the DLPack protocol to exchange data in a zero-copy way with JAX.</p>
-<dl class="py function" id="module-warp.jax">
-<dt class="sig sig-object py" id="warp.jax.device_to_jax">
-<span class="sig-prename descclassname"><span class="pre">warp.jax.</span></span><span class="sig-name descname"><span class="pre">device_to_jax</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">wp_device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.jax.device_to_jax" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.jax.device_from_jax">
-<span class="sig-prename descclassname"><span class="pre">warp.jax.</span></span><span class="sig-name descname"><span class="pre">device_from_jax</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">jax_device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.jax.device_from_jax" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.jax.to_jax">
-<span class="sig-prename descclassname"><span class="pre">warp.jax.</span></span><span class="sig-name descname"><span class="pre">to_jax</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">wp_array</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.jax.to_jax" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.jax.from_jax">
-<span class="sig-prename descclassname"><span class="pre">warp.jax.</span></span><span class="sig-name descname"><span class="pre">from_jax</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">jax_array</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.jax.from_jax" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-</section>
-<section id="module-warp.dlpack">
-<span id="dlpack"></span><h2>DLPack<a class="headerlink" href="#module-warp.dlpack" title="Link to this heading">#</a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.device_to_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">device_to_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">wp_device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.device_to_dlpack" title="Link to this definition">#</a></dt>
-<dd><dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><em>DLDevice</em></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.dtype_to_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">dtype_to_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">wp_dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.dtype_to_dlpack" title="Link to this definition">#</a></dt>
-<dd><dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><em>DLDataType</em></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.dtype_from_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">dtype_from_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dl_dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.dtype_from_dlpack" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.device_from_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">device_from_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dl_device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.device_from_dlpack" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.shape_to_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">shape_to_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.shape_to_dlpack" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.strides_to_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">strides_to_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">strides</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.strides_to_dlpack" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.to_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">to_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">wp_array</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.to_dlpack" title="Link to this definition">#</a></dt>
-<dd><p>Convert a Warp array to another type of dlpack compatible array.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>np_array</strong> (<em>np.ndarray</em>) – The source numpy array that will be converted.</p></li>
-<li><p><strong>wp_array</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.types.array"><em>array</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p><strong>pycapsule</strong> – A pycapsule containing a DLManagedTensor that can be converted
-to other array formats without copying the underlying memory.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>PyCapsule</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.dtype_is_compatible">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">dtype_is_compatible</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dl_dtype</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">wp_dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.dtype_is_compatible" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.dlpack.from_dlpack">
-<span class="sig-prename descclassname"><span class="pre">warp.dlpack.</span></span><span class="sig-name descname"><span class="pre">from_dlpack</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pycapsule</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.dlpack.from_dlpack" title="Link to this definition">#</a></dt>
-<dd><p>Convert a dlpack tensor into a numpy array without copying.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>pycapsule</strong> (<em>PyCapsule</em>) – A pycapsule wrapping a dlpack tensor that will be converted.</p>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p><strong>np_array</strong> – A new numpy array that uses the same underlying memory as the input
-pycapsule.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>np.ndarray</p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="../configuration.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Runtime Settings</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="devices.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Devices</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Interoperability</a><ul>
-<li><a class="reference internal" href="#numpy">NumPy</a></li>
-<li><a class="reference internal" href="#pytorch">PyTorch</a><ul>
-<li><a class="reference internal" href="#example-optimization-using-warp-from-torch">Example: Optimization using <code class="docutils literal notranslate"><span class="pre">warp.from_torch()</span></code></a></li>
-<li><a class="reference internal" href="#example-optimization-using-warp-to-torch">Example: Optimization using <code class="docutils literal notranslate"><span class="pre">warp.to_torch</span></code></a><ul>
-<li><a class="reference internal" href="#warp.torch.device_from_torch"><code class="docutils literal notranslate"><span class="pre">device_from_torch()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.device_to_torch"><code class="docutils literal notranslate"><span class="pre">device_to_torch()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.dtype_from_torch"><code class="docutils literal notranslate"><span class="pre">dtype_from_torch()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.dtype_is_compatible"><code class="docutils literal notranslate"><span class="pre">dtype_is_compatible()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.from_torch"><code class="docutils literal notranslate"><span class="pre">from_torch()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.to_torch"><code class="docutils literal notranslate"><span class="pre">to_torch()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.stream_from_torch"><code class="docutils literal notranslate"><span class="pre">stream_from_torch()</span></code></a></li>
-<li><a class="reference internal" href="#warp.torch.stream_to_torch"><code class="docutils literal notranslate"><span class="pre">stream_to_torch()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#cupy-numba">CuPy/Numba</a></li>
-<li><a class="reference internal" href="#jax">JAX</a><ul>
-<li><a class="reference internal" href="#warp.jax.device_to_jax"><code class="docutils literal notranslate"><span class="pre">device_to_jax()</span></code></a></li>
-<li><a class="reference internal" href="#warp.jax.device_from_jax"><code class="docutils literal notranslate"><span class="pre">device_from_jax()</span></code></a></li>
-<li><a class="reference internal" href="#warp.jax.to_jax"><code class="docutils literal notranslate"><span class="pre">to_jax()</span></code></a></li>
-<li><a class="reference internal" href="#warp.jax.from_jax"><code class="docutils literal notranslate"><span class="pre">from_jax()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#module-warp.dlpack">DLPack</a><ul>
-<li><a class="reference internal" href="#warp.dlpack.device_to_dlpack"><code class="docutils literal notranslate"><span class="pre">device_to_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.dtype_to_dlpack"><code class="docutils literal notranslate"><span class="pre">dtype_to_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.dtype_from_dlpack"><code class="docutils literal notranslate"><span class="pre">dtype_from_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.device_from_dlpack"><code class="docutils literal notranslate"><span class="pre">device_from_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.shape_to_dlpack"><code class="docutils literal notranslate"><span class="pre">shape_to_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.strides_to_dlpack"><code class="docutils literal notranslate"><span class="pre">strides_to_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.to_dlpack"><code class="docutils literal notranslate"><span class="pre">to_dlpack()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.dtype_is_compatible"><code class="docutils literal notranslate"><span class="pre">dtype_is_compatible()</span></code></a></li>
-<li><a class="reference internal" href="#warp.dlpack.from_dlpack"><code class="docutils literal notranslate"><span class="pre">from_dlpack()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/profiling.html b/docs/_build/html/modules/profiling.html
deleted file mode 100644
index c3ca84e38..000000000
--- a/docs/_build/html/modules/profiling.html
+++ /dev/null
@@ -1,388 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="Interoperability" href="interoperability.html" /><link rel="prev" title="Devices" href="devices.html" />
-
-    <!-- Generated with Sphinx 7.2.2 and Furo 2023.08.19 -->
-        <title>Profiling - Warp 1.0.0-beta.2</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=112ea5a1" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.2</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.2</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Profiling</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="profiling">
-<h1>Profiling<a class="headerlink" href="#profiling" title="Link to this heading">#</a></h1>
-<p><code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> objects can be used to gain some basic insight into the performance of Warp applications:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedTimer</span><span class="p">(</span><span class="s2">&quot;grid build&quot;</span><span class="p">):</span>
-   <span class="bp">self</span><span class="o">.</span><span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">point_radius</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>This results in a printout at runtime to the standard output stream like:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">grid</span> <span class="n">build</span> <span class="n">took</span> <span class="mf">0.06</span> <span class="n">ms</span>
-</pre></div>
-</div>
-<p>The <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> object does not synchronize (e.g. by calling <code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code>)
-upon exiting the <code class="docutils literal notranslate"><span class="pre">with</span></code> statement, so this can lead to misleading numbers if the body
-of the <code class="docutils literal notranslate"><span class="pre">with</span></code> statement launches device kernels.</p>
-<p>When a <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> object is passed <code class="docutils literal notranslate"><span class="pre">use_nvtx=True</span></code> as an argument, the timing functionality is replaced by calls
-to <code class="docutils literal notranslate"><span class="pre">nvtx.start_range()</span></code> and <code class="docutils literal notranslate"><span class="pre">nvtx.end_range()</span></code>:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedTimer</span><span class="p">(</span><span class="s2">&quot;grid build&quot;</span><span class="p">,</span> <span class="n">use_nvtx</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s2">&quot;cyan&quot;</span><span class="p">):</span>
-   <span class="bp">self</span><span class="o">.</span><span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">point_radius</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>These range annotations can then be collected by a tool like <a class="reference external" href="https://developer.nvidia.com/nsight-systems">NVIDIA Nsight Systems</a>
-for visualization on a timeline, e.g.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">nsys</span> <span class="n">profile</span> <span class="n">python</span> <span class="n">warp_application</span><span class="o">.</span><span class="n">py</span>
-</pre></div>
-</div>
-<p>This code snippet also demonstrates the use of the <code class="docutils literal notranslate"><span class="pre">color</span></code> argument to specify a color
-for the range, which may be a number representing the ARGB value or a recognized string
-(refer to <a class="reference external" href="https://github.com/NVIDIA/NVTX/blob/release-v3/python/nvtx/colors.py">colors.py</a> for
-additional color examples).
-The <a class="reference external" href="https://github.com/NVIDIA/NVTX">nvtx module</a> must be
-installed in the Python environment for this capability to work.
-An equivalent way to create an NVTX range without using <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> is:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">nvtx</span>
-
-<span class="k">with</span> <span class="n">nvtx</span><span class="o">.</span><span class="n">annotate</span><span class="p">(</span><span class="s2">&quot;grid build&quot;</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s2">&quot;cyan&quot;</span><span class="p">):</span>
-   <span class="bp">self</span><span class="o">.</span><span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">point_radius</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>This form may be more convenient if the user does not need to frequently switch
-between timer and NVTX capabilities of <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code>.</p>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.ScopedTimer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">ScopedTimer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">active</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">print</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">detailed</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dict</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_nvtx</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'rapids'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">synchronize</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.ScopedTimer" title="Link to this definition">#</a></dt>
-<dd><p>Context manager object for a timer</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.11)"><em>str</em></a>) – Name of timer</p></li>
-<li><p><strong>active</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.11)"><em>bool</em></a>) – Enables this timer</p></li>
-<li><p><strong>print</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.11)"><em>bool</em></a>) – At context manager exit, print elapsed time to sys.stdout</p></li>
-<li><p><strong>detailed</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.11)"><em>bool</em></a>) – Collects additional profiling data using cProfile and calls <code class="docutils literal notranslate"><span class="pre">print_stats()</span></code> at context exit</p></li>
-<li><p><strong>dict</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.11)"><em>dict</em></a>) – A dictionary of lists to which the elapsed time will be appended using <code class="docutils literal notranslate"><span class="pre">name</span></code> as a key</p></li>
-<li><p><strong>use_nvtx</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.11)"><em>bool</em></a>) – If true, timing functionality is replaced by an NVTX range</p></li>
-<li><p><strong>color</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.11)"><em>int</em></a><em> or </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.11)"><em>str</em></a>) – ARGB value (e.g. 0x00FFFF) or color name (e.g. ‘cyan’) associated with the NVTX range</p></li>
-<li><p><strong>synchronize</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#bool" title="(in Python v3.11)"><em>bool</em></a>) – Synchronize the CPU thread with any outstanding CUDA work to return accurate GPU timings</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.ScopedTimer.elapsed">
-<span class="sig-name descname"><span class="pre">elapsed</span></span><a class="headerlink" href="#warp.ScopedTimer.elapsed" title="Link to this definition">#</a></dt>
-<dd><p>The duration of the <code class="docutils literal notranslate"><span class="pre">with</span></code> block used with this object</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.11)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="interoperability.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Interoperability</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="devices.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Devices</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Profiling</a><ul>
-<li><a class="reference internal" href="#warp.ScopedTimer"><code class="docutils literal notranslate"><span class="pre">ScopedTimer</span></code></a><ul>
-<li><a class="reference internal" href="#warp.ScopedTimer.elapsed"><code class="docutils literal notranslate"><span class="pre">ScopedTimer.elapsed</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=5929fcd5"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/runtime.html b/docs/_build/html/modules/runtime.html
deleted file mode 100644
index f26419fa3..000000000
--- a/docs/_build/html/modules/runtime.html
+++ /dev/null
@@ -1,2743 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="Kernel Reference" href="functions.html" /><link rel="prev" title="FAQ" href="../faq.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>Runtime Reference - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul class="current">
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="runtime-reference">
-<h1>Runtime Reference<a class="headerlink" href="#runtime-reference" title="Link to this heading">#</a></h1>
-<p>This section describes the Warp Python runtime API, how to manage memory, launch kernels, and high-level functionality
-for dealing with objects such as meshes and volumes. The APIs described in this section are intended to be used at
-the <em>Python Scope</em> and run inside the CPython interpreter. For a comprehensive list of functions available at
-the <em>Kernel Scope</em>, please see the <a class="reference internal" href="functions.html"><span class="doc">Kernel Reference</span></a> section.</p>
-<section id="kernels">
-<h2>Kernels<a class="headerlink" href="#kernels" title="Link to this heading">#</a></h2>
-<p>Kernels are launched with the <a class="reference internal" href="#warp.launch" title="warp.launch"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.launch()</span></code></a> function on a specific device (CPU/GPU):</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">simple_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">1024</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">,</span> <span class="n">c</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Kernels may be launched with multi-dimensional grid bounds. In this case threads are not assigned a single index,
-but a coordinate in an n-dimensional grid, e.g.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">complex_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="p">(</span><span class="mi">128</span><span class="p">,</span> <span class="mi">128</span><span class="p">,</span> <span class="mi">3</span><span class="p">),</span> <span class="o">...</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Launches a 3D grid of threads with dimension 128 x 128 x 3. To retrieve the 3D index for each thread use the following syntax:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">i</span><span class="p">,</span><span class="n">j</span><span class="p">,</span><span class="n">k</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Currently kernels launched on CPU devices will be executed in serial.
-Kernels launched on CUDA devices will be launched in parallel with a fixed block-size.</p>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Note that all the kernel inputs must live on the target device, or a runtime exception will be raised.</p>
-</div>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.launch">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">launch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">kernel</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">inputs</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">outputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adj_inputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adj_outputs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stream</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">adjoint</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">record_tape</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">record_cmd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.launch" title="Link to this definition">#</a></dt>
-<dd><p>Launch a Warp kernel on the target device</p>
-<p>Kernel launches are asynchronous with respect to the calling Python thread.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>kernel</strong> – The name of a Warp kernel function, decorated with the <code class="docutils literal notranslate"><span class="pre">&#64;wp.kernel</span></code> decorator</p></li>
-<li><p><strong>dim</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – The number of threads to launch the kernel, can be an integer, or a Tuple of ints with max of 4 dimensions</p></li>
-<li><p><strong>inputs</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a>) – The input parameters to the kernel</p></li>
-<li><p><strong>outputs</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a>) – The output parameters (optional)</p></li>
-<li><p><strong>adj_inputs</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a>) – The adjoint inputs (optional)</p></li>
-<li><p><strong>adj_outputs</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a>) – The adjoint outputs (optional)</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The device to launch on (optional)</p></li>
-<li><p><strong>stream</strong> (<em>Stream</em><em> | </em><em>None</em>) – The stream to launch on (optional)</p></li>
-<li><p><strong>adjoint</strong> – Whether to run forward or backward pass (typically use False)</p></li>
-<li><p><strong>record_tape</strong> – When true the launch will be recorded the global wp.Tape() object when present</p></li>
-<li><p><strong>record_cmd</strong> – When True the launch will be returned as a <code class="docutils literal notranslate"><span class="pre">Launch</span></code> command object, the launch will not occur until the user calls <code class="docutils literal notranslate"><span class="pre">cmd.launch()</span></code></p></li>
-<li><p><strong>max_blocks</strong> – The maximum number of CUDA thread blocks to use. Only has an effect for CUDA kernel launches.
-If negative or zero, the maximum hardware value will be used.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<section id="large-kernel-launches">
-<h3>Large Kernel Launches<a class="headerlink" href="#large-kernel-launches" title="Link to this heading">#</a></h3>
-<p>A limitation of Warp is that each dimension of the grid used to launch a kernel must be representable as a 32-bit
-signed integer. Therefore, no single dimension of a grid should exceed <span class="math notranslate nohighlight">\(2^{31}-1\)</span>.</p>
-<p>Warp also currently uses a fixed block size of 256 (CUDA) threads per block.
-By default, Warp will try to process one element from the Warp grid in one CUDA thread.
-This is not always possible for kernels launched with multi-dimensional grid bounds, as there are
-<a class="reference external" href="https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#features-and-technical-specifications-technical-specifications-per-compute-capability">hardware limitations</a>
-on CUDA block dimensions.</p>
-<p>Warp will automatically resort to using
-<a class="reference external" href="https://developer.nvidia.com/blog/cuda-pro-tip-write-flexible-kernels-grid-stride-loops/">grid-stride loops</a> when
-it is not possible for a CUDA thread to process only one element from the Warp grid
-When this happens, some CUDA threads may process more than one element from the Warp grid.
-Users can also set the <code class="docutils literal notranslate"><span class="pre">max_blocks</span></code> parameter to fine-tune the grid-striding behavior of kernels, even for kernels that are otherwise
-able to process one Warp-grid element per CUDA thread.</p>
-</section>
-<section id="runtime-kernel-specialization">
-<h3>Runtime Kernel Specialization<a class="headerlink" href="#runtime-kernel-specialization" title="Link to this heading">#</a></h3>
-<p>It is often desirable to specialize kernels for different types, constants, or functions.
-We can achieve this through the use of runtime kernel specialization using Python closures.</p>
-<p>For example, we might require a variety of kernels that execute particular functions for each item in an array.
-We might also want this function call to be valid for a variety of data types. Making use of closure and generics, we can generate
-these kernels using a single kernel definition:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">make_kernel</span><span class="p">(</span><span class="n">func</span><span class="p">,</span> <span class="n">dtype</span><span class="p">):</span>
-    <span class="k">def</span> <span class="nf">closure_kernel_fn</span><span class="p">(</span><span class="n">data</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">dtype</span><span class="p">),</span> <span class="n">out</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">dtype</span><span class="p">)):</span>
-        <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-        <span class="n">out</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">func</span><span class="p">(</span><span class="n">data</span><span class="p">[</span><span class="n">tid</span><span class="p">])</span>
-
-    <span class="k">return</span> <span class="n">wp</span><span class="o">.</span><span class="n">Kernel</span><span class="p">(</span><span class="n">closure_kernel_fn</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>In practice, we might use our kernel generator, <code class="docutils literal notranslate"><span class="pre">make_kernel()</span></code> as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func</span>
-<span class="k">def</span> <span class="nf">sqr</span><span class="p">(</span><span class="n">x</span><span class="p">:</span> <span class="n">Any</span><span class="p">)</span> <span class="o">-&gt;</span> <span class="n">Any</span><span class="p">:</span>
-    <span class="k">return</span> <span class="n">x</span> <span class="o">*</span> <span class="n">x</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">func</span>
-<span class="k">def</span> <span class="nf">cube</span><span class="p">(</span><span class="n">x</span><span class="p">:</span> <span class="n">Any</span><span class="p">)</span> <span class="o">-&gt;</span> <span class="n">Any</span><span class="p">:</span>
-    <span class="k">return</span> <span class="n">sqr</span><span class="p">(</span><span class="n">x</span><span class="p">)</span> <span class="o">*</span> <span class="n">x</span>
-
-<span class="n">sqr_float</span> <span class="o">=</span> <span class="n">make_kernel</span><span class="p">(</span><span class="n">sqr</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-<span class="n">cube_double</span> <span class="o">=</span> <span class="n">make_kernel</span><span class="p">(</span><span class="n">cube</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-
-<span class="n">arr</span> <span class="o">=</span> <span class="p">[</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">]</span>
-<span class="n">N</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">arr</span><span class="p">)</span>
-
-<span class="n">data_float</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">arr</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-<span class="n">data_double</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">arr</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">out_float</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-<span class="n">out_double</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">sqr_float</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">N</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">data_float</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">out_float</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">cube_double</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">N</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">data_double</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">out_double</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>We can specialize kernel definitions over warp constants similarly. The following generates kernels that add a specified constant
-to a generic-typed array value:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">make_add_kernel</span><span class="p">(</span><span class="n">key</span><span class="p">,</span> <span class="n">constant</span><span class="p">):</span>
-    <span class="k">def</span> <span class="nf">closure_kernel_fn</span><span class="p">(</span><span class="n">data</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">Any</span><span class="p">),</span> <span class="n">out</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">Any</span><span class="p">)):</span>
-        <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-        <span class="n">out</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">data</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">+</span> <span class="n">constant</span>
-
-    <span class="k">return</span> <span class="n">wp</span><span class="o">.</span><span class="n">Kernel</span><span class="p">(</span><span class="n">closure_kernel_fn</span><span class="p">,</span> <span class="n">key</span><span class="o">=</span><span class="n">key</span><span class="p">)</span>
-
-<span class="n">add_ones_int</span> <span class="o">=</span> <span class="n">make_add_kernel</span><span class="p">(</span><span class="s2">&quot;add_one&quot;</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">constant</span><span class="p">(</span><span class="mi">1</span><span class="p">))</span>
-<span class="n">add_ones_vec3</span> <span class="o">=</span> <span class="n">make_add_kernel</span><span class="p">(</span><span class="s2">&quot;add_ones_vec3&quot;</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">constant</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">)))</span>
-
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">)</span>
-<span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-
-<span class="n">a_out</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros_like</span><span class="p">(</span><span class="n">a</span><span class="p">)</span>
-<span class="n">b_out</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros_like</span><span class="p">(</span><span class="n">b</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">add_ones_int</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">a</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">a_out</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">add_ones_vec3</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">b</span><span class="o">.</span><span class="n">size</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">b</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">b_out</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-</section>
-<section id="arrays">
-<h2>Arrays<a class="headerlink" href="#arrays" title="Link to this heading">#</a></h2>
-<p>Arrays are the fundamental memory abstraction in Warp; they are created through the following global constructors:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="mi">1024</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="mi">1024</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">full</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="mi">1024</span><span class="p">,</span> <span class="n">value</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Arrays can also be constructed directly from <code class="docutils literal notranslate"><span class="pre">numpy</span></code> ndarrays as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">r</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="mi">1024</span><span class="p">)</span>
-
-<span class="c1"># copy to Warp owned array</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">r</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-
-<span class="c1"># return a Warp array wrapper around the NumPy data (zero-copy)</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">r</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">copy</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-
-<span class="c1"># return a Warp copy of the array data on the GPU</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">r</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Note that for multi-dimensional data the <code class="docutils literal notranslate"><span class="pre">dtype</span></code> parameter must be specified explicitly, e.g.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">r</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">rand</span><span class="p">((</span><span class="mi">1024</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span>
-
-<span class="c1"># initialize as an array of vec3 objects</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">r</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>If the shapes are incompatible, an error will be raised.</p>
-<p>Arrays can be moved between devices using the <code class="docutils literal notranslate"><span class="pre">array.to()</span></code> method:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">host_array</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-
-<span class="c1"># allocate and copy to GPU</span>
-<span class="n">device_array</span> <span class="o">=</span> <span class="n">host_array</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Additionally, arrays can be copied directly between memory spaces:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">src_array</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-<span class="n">dest_array</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">empty_like</span><span class="p">(</span><span class="n">host_array</span><span class="p">)</span>
-
-<span class="c1"># copy from source CPU buffer to GPU</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">copy</span><span class="p">(</span><span class="n">dest_array</span><span class="p">,</span> <span class="n">src_array</span><span class="p">)</span>
-</pre></div>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.array">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">array</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">typing.Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">strides</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">length</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ptr</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">capacity</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">copy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">owner</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ndim</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array" title="Link to this definition">#</a></dt>
-<dd><p>Constructs a new Warp array object</p>
-<p>When the <code class="docutils literal notranslate"><span class="pre">data</span></code> argument is a valid list, tuple, or ndarray the array will be constructed from this object’s data.
-For objects that are not stored sequentially in memory (e.g.: a list), then the data will first
-be flattened before being transferred to the memory space given by device.</p>
-<p>The second construction path occurs when the <code class="docutils literal notranslate"><span class="pre">ptr</span></code> argument is a non-zero uint64 value representing the
-start address in memory where existing array data resides, e.g.: from an external or C-library. The memory
-allocation should reside on the same device given by the device argument, and the user should set the length
-and dtype parameter appropriately.</p>
-<p>If neither <code class="docutils literal notranslate"><span class="pre">data</span></code> nor <code class="docutils literal notranslate"><span class="pre">ptr</span></code> are specified, the <code class="docutils literal notranslate"><span class="pre">shape</span></code> or <code class="docutils literal notranslate"><span class="pre">length</span></code> arguments are checked next.
-This construction path can be used to create new uninitialized arrays, but users are encouraged to call
-<code class="docutils literal notranslate"><span class="pre">wp.empty()</span></code>, <code class="docutils literal notranslate"><span class="pre">wp.zeros()</span></code>, or <code class="docutils literal notranslate"><span class="pre">wp.full()</span></code> instead to create new arrays.</p>
-<p>If none of the above arguments are specified, a simple type annotation is constructed.  This is used when annotating
-kernel arguments or struct members (e.g.,``arr: wp.array(dtype=float)``).  In this case, only <code class="docutils literal notranslate"><span class="pre">dtype</span></code> and <code class="docutils literal notranslate"><span class="pre">ndim</span></code>
-are taken into account and no memory is allocated for the array.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>data</strong> (<em>Union</em><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)"><em>list</em></a><em>, </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#tuple" title="(in Python v3.12)"><em>tuple</em></a><em>, </em><em>ndarray</em><em>]</em>) – </p></li>
-<li><p><strong>dtype</strong> (<em>Union</em>) – One of the built-in types, e.g.: <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.mat33</span></code>, if dtype is Any and data an ndarray then it will be inferred from the array data type</p></li>
-<li><p><strong>shape</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#tuple" title="(in Python v3.12)"><em>tuple</em></a>) – Dimensions of the array</p></li>
-<li><p><strong>strides</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#tuple" title="(in Python v3.12)"><em>tuple</em></a>) – Number of bytes in each dimension between successive elements of the array</p></li>
-<li><p><strong>length</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of elements of the data type (deprecated, users should use <cite>shape</cite> argument)</p></li>
-<li><p><strong>ptr</strong> (<a class="reference internal" href="functions.html#warp.uint64" title="warp.uint64"><em>uint64</em></a>) – Address of an external memory address to alias (data should be None)</p></li>
-<li><p><strong>capacity</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Maximum size in bytes of the ptr allocation (data should be None)</p></li>
-<li><p><strong>device</strong> (<em>Devicelike</em>) – Device the array lives on</p></li>
-<li><p><strong>copy</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the incoming data will be copied or aliased, this is only possible when the incoming <cite>data</cite> already lives on the device specified and types match</p></li>
-<li><p><strong>owner</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Should the array object try to deallocate memory when it is deleted</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether or not gradients will be tracked for this array, see <a class="reference internal" href="#warp.Tape" title="warp.Tape"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.Tape</span></code></a> for details</p></li>
-<li><p><strong>grad</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><em>array</em></a>) – The gradient array to use</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to allocate pinned host memory, which allows asynchronous host-device transfers (only applicable with device=”cpu”)</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.array.grad">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">grad</span></span><a class="headerlink" href="#warp.array.grad" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.array.requires_grad">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">requires_grad</span></span><a class="headerlink" href="#warp.array.requires_grad" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.array.vars">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">vars</span></span><a class="headerlink" href="#warp.array.vars" title="Link to this definition">#</a></dt>
-<dd></dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.zero_">
-<span class="sig-name descname"><span class="pre">zero_</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.zero_" title="Link to this definition">#</a></dt>
-<dd><p>Zeroes-out the array entires.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.fill_">
-<span class="sig-name descname"><span class="pre">fill_</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">value</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.fill_" title="Link to this definition">#</a></dt>
-<dd><p>Set all array entries to <cite>value</cite></p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>value</strong> – The value to set every array entry to. Must be convertible to the array’s <code class="docutils literal notranslate"><span class="pre">dtype</span></code>.</p>
-</dd>
-<dt class="field-even">Raises<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference external" href="https://docs.python.org/3/library/exceptions.html#ValueError" title="(in Python v3.12)"><strong>ValueError</strong></a> – If <cite>value</cite> cannot be converted to the array’s <code class="docutils literal notranslate"><span class="pre">dtype</span></code>.</p>
-</dd>
-</dl>
-<p class="rubric">Examples</p>
-<p><code class="docutils literal notranslate"><span class="pre">fill_()</span></code> can take lists or other sequences when filling arrays of vectors or matrices.</p>
-<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="n">arr</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">mat22</span><span class="p">)</span>
-<span class="gp">&gt;&gt;&gt; </span><span class="n">arr</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-<span class="go">array([[[0., 0.],</span>
-<span class="go">        [0., 0.]],</span>
-
-<span class="go">       [[0., 0.],</span>
-<span class="go">        [0., 0.]]], dtype=float32)</span>
-<span class="gp">&gt;&gt;&gt; </span><span class="n">arr</span><span class="o">.</span><span class="n">fill_</span><span class="p">([[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">],</span> <span class="p">[</span><span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">]])</span>
-<span class="gp">&gt;&gt;&gt; </span><span class="n">arr</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-<span class="go">array([[[1., 2.],</span>
-<span class="go">        [3., 4.]],</span>
-
-<span class="go">       [[1., 2.],</span>
-<span class="go">        [3., 4.]]], dtype=float32)</span>
-</pre></div>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.assign">
-<span class="sig-name descname"><span class="pre">assign</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.assign" title="Link to this definition">#</a></dt>
-<dd><p>Wraps <code class="docutils literal notranslate"><span class="pre">src</span></code> in an <a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a> if it is not already one and copies the contents to <code class="docutils literal notranslate"><span class="pre">self</span></code>.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.numpy">
-<span class="sig-name descname"><span class="pre">numpy</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.numpy" title="Link to this definition">#</a></dt>
-<dd><p>Converts the array to a <a class="reference external" href="https://numpy.org/doc/stable/reference/generated/numpy.ndarray.html#numpy.ndarray" title="(in NumPy v1.26)"><code class="xref py py-class docutils literal notranslate"><span class="pre">numpy.ndarray</span></code></a> (aliasing memory through the array interface protocol)
-If the array is on the GPU, a synchronous device-to-host copy (on the CUDA default stream) will be
-automatically performed to ensure that any outstanding work is completed.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.cptr">
-<span class="sig-name descname"><span class="pre">cptr</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.cptr" title="Link to this definition">#</a></dt>
-<dd><p>Return a ctypes cast of the array address.</p>
-<p>Notes:</p>
-<ol class="arabic simple">
-<li><p>Only CPU arrays support this method.</p></li>
-<li><p>The array must be contiguous.</p></li>
-<li><p>Accesses to this object are <strong>not</strong> bounds checked.</p></li>
-<li><p>For <code class="docutils literal notranslate"><span class="pre">float16</span></code> types, a pointer to the internal <code class="docutils literal notranslate"><span class="pre">uint16</span></code> representation is returned.</p></li>
-</ol>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.list">
-<span class="sig-name descname"><span class="pre">list</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.list" title="Link to this definition">#</a></dt>
-<dd><p>Returns a flattened list of items in the array as a Python list.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.to">
-<span class="sig-name descname"><span class="pre">to</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.to" title="Link to this definition">#</a></dt>
-<dd><p>Returns a Warp array with this array’s data moved to the specified device, no-op if already on device.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.flatten">
-<span class="sig-name descname"><span class="pre">flatten</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.flatten" title="Link to this definition">#</a></dt>
-<dd><p>Returns a zero-copy view of the array collapsed to 1-D. Only supported for contiguous arrays.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.reshape">
-<span class="sig-name descname"><span class="pre">reshape</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.reshape" title="Link to this definition">#</a></dt>
-<dd><p>Returns a reshaped array. Only supported for contiguous arrays.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>shape</strong> – An int or tuple of ints specifying the shape of the returned array.</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.view">
-<span class="sig-name descname"><span class="pre">view</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dtype</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.view" title="Link to this definition">#</a></dt>
-<dd><p>Returns a zero-copy view of this array’s memory with a different data type.
-<code class="docutils literal notranslate"><span class="pre">dtype</span></code> must have the same byte size of the array’s native <code class="docutils literal notranslate"><span class="pre">dtype</span></code>.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.contiguous">
-<span class="sig-name descname"><span class="pre">contiguous</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.contiguous" title="Link to this definition">#</a></dt>
-<dd><p>Returns a contiguous array with this array’s data. No-op if array is already contiguous.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.array.transpose">
-<span class="sig-name descname"><span class="pre">transpose</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">axes</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.array.transpose" title="Link to this definition">#</a></dt>
-<dd><p>Returns an zero-copy view of the array with axes transposed.</p>
-<p>Note: The transpose operation will return an array with a non-contiguous access pattern.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>axes</strong> (<em>optional</em>) – Specifies the how the axes are permuted. If not specified, the axes order will be reversed.</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<section id="multi-dimensional-arrays">
-<h3>Multi-dimensional Arrays<a class="headerlink" href="#multi-dimensional-arrays" title="Link to this heading">#</a></h3>
-<p>Multi-dimensional arrays can be constructed by passing a tuple of sizes for each dimension, e.g.: the following constructs a 2d array of size 1024x16:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">1024</span><span class="p">,</span> <span class="mi">16</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>When passing multi-dimensional arrays to kernels users must specify the expected array dimension inside the kernel signature,
-e.g. to pass a 2d array to a kernel the number of dims is specified using the <code class="docutils literal notranslate"><span class="pre">ndim=2</span></code> parameter:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">test</span><span class="p">(</span><span class="nb">input</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">ndim</span><span class="o">=</span><span class="mi">2</span><span class="p">)):</span>
-</pre></div>
-</div>
-<p>Type-hint helpers are provided for common array sizes, e.g.: <code class="docutils literal notranslate"><span class="pre">array2d()</span></code>, <code class="docutils literal notranslate"><span class="pre">array3d()</span></code>, which are equivalent to calling <code class="docutils literal notranslate"><span class="pre">array(...,</span> <span class="pre">ndim=2)`</span></code>, etc. To index a multi-dimensional array use a the following kernel syntax:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># returns a float from the 2d array</span>
-<span class="n">value</span> <span class="o">=</span> <span class="nb">input</span><span class="p">[</span><span class="n">i</span><span class="p">,</span><span class="n">j</span><span class="p">]</span>
-</pre></div>
-</div>
-<p>To create an array slice use the following syntax, where the number of indices is less than the array dimensions:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># returns an 1d array slice representing a row of the 2d array</span>
-<span class="n">row</span> <span class="o">=</span> <span class="nb">input</span><span class="p">[</span><span class="n">i</span><span class="p">]</span>
-</pre></div>
-</div>
-<p>Slice operators can be concatenated, e.g.: <code class="docutils literal notranslate"><span class="pre">s</span> <span class="pre">=</span> <span class="pre">array[i][j][k]</span></code>. Slices can be passed to <code class="docutils literal notranslate"><span class="pre">wp.func</span></code> user functions provided
-the function also declares the expected array dimension. Currently only single-index slicing is supported.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Currently Warp limits arrays to 4 dimensions maximum. This is in addition to the contained datatype, which may be 1-2 dimensional for vector and matrix types such as <code class="docutils literal notranslate"><span class="pre">vec3</span></code>, and <code class="docutils literal notranslate"><span class="pre">mat33</span></code>.</p>
-</div>
-<p>The following construction methods are provided for allocating zero-initialized and empty (non-initialized) arrays:</p>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.zeros">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">zeros</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype=&lt;class</span> <span class="pre">'float'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">**kwargs</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.zeros" title="Link to this definition">#</a></dt>
-<dd><p>Return a zero-initialized array</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>shape</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a><em> | </em><em>None</em>) – Array dimensions</p></li>
-<li><p><strong>dtype</strong> – Type of each element, e.g.: warp.vec3, warp.mat33, etc</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – Device that array will live on</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.zeros_like">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">zeros_like</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.zeros_like" title="Link to this definition">#</a></dt>
-<dd><p>Return a zero-initialized array with the same type and dimension of another array</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a>) – The template array to use for shape, data type, and device</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The device where the new array will be created (defaults to src.device)</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.full">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">full</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">typing.Any</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.full" title="Link to this definition">#</a></dt>
-<dd><p>Return an array with all elements initialized to the given value</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>shape</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a><em> | </em><em>None</em>) – Array dimensions</p></li>
-<li><p><strong>value</strong> – Element value</p></li>
-<li><p><strong>dtype</strong> – Type of each element, e.g.: float, warp.vec3, warp.mat33, etc</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – Device that array will live on</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.full_like">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">full_like</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">value</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.full_like" title="Link to this definition">#</a></dt>
-<dd><p>Return an array with all elements initialized to the given value with the same type and dimension of another array</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a>) – The template array to use for shape, data type, and device</p></li>
-<li><p><strong>value</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a>) – Element value</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The device where the new array will be created (defaults to src.device)</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.empty">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">empty</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">shape=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dtype=&lt;class</span> <span class="pre">'float'&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">**kwargs</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.empty" title="Link to this definition">#</a></dt>
-<dd><p>Returns an uninitialized array</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>shape</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a><em> | </em><em>None</em>) – Array dimensions</p></li>
-<li><p><strong>dtype</strong> – Type of each element, e.g.: <cite>warp.vec3</cite>, <cite>warp.mat33</cite>, etc</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – Device that array will live on</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.empty_like">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">empty_like</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.empty_like" title="Link to this definition">#</a></dt>
-<dd><p>Return an uninitialized array with the same type and dimension of another array</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a>) – The template array to use for shape, data type, and device</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The device where the new array will be created (defaults to src.device)</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.copy">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">copy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dest</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">src</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dest_offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">src_offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">count</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stream</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.copy" title="Link to this definition">#</a></dt>
-<dd><p>Copy array contents from src to dest</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dest</strong> (<a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a>) – Destination array, must be at least as big as source buffer</p></li>
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a>) – Source array</p></li>
-<li><p><strong>dest_offset</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Element offset in the destination array</p></li>
-<li><p><strong>src_offset</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Element offset in the source array</p></li>
-<li><p><strong>count</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of array elements to copy (will copy all elements if set to 0)</p></li>
-<li><p><strong>stream</strong> (<em>Stream</em><em> | </em><em>None</em>) – The stream on which to perform the copy (optional)</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.clone">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">clone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">src</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pinned</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.clone" title="Link to this definition">#</a></dt>
-<dd><p>Clone an existing array, allocates a copy of the src memory</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a>) – The source array to copy</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The device where the new array will be created (defaults to src.device)</p></li>
-<li><p><strong>requires_grad</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array will be tracked for back propagation</p></li>
-<li><p><strong>pinned</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a><em> | </em><em>None</em>) – Whether the array uses pinned host memory (only applicable to CPU arrays)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A warp.array object representing the allocation</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="matrix-multiplication">
-<h3>Matrix Multiplication<a class="headerlink" href="#matrix-multiplication" title="Link to this heading">#</a></h3>
-<p>Warp 2D array multiplication is built on NVIDIA’s CUTLASS library, which enables fast matrix multiplication of large arrays on the GPU.</p>
-<p>If no GPU is detected, matrix multiplication falls back to Numpy’s implementation on the CPU.</p>
-<p>Matrix multiplication is fully differentiable, and can be recorded on the tape like so:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">matmul</span><span class="p">(</span><span class="n">A</span><span class="p">,</span> <span class="n">B</span><span class="p">,</span> <span class="n">C</span><span class="p">,</span> <span class="n">D</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">loss_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="p">(</span><span class="n">m</span><span class="p">,</span> <span class="n">n</span><span class="p">),</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">D</span><span class="p">,</span> <span class="n">loss</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">loss</span><span class="o">=</span><span class="n">loss</span><span class="p">)</span>
-<span class="n">A_grad</span> <span class="o">=</span> <span class="n">A</span><span class="o">.</span><span class="n">grad</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>Using the <code class="docutils literal notranslate"><span class="pre">&#64;</span></code> operator (<code class="docutils literal notranslate"><span class="pre">D</span> <span class="pre">=</span> <span class="pre">A</span> <span class="pre">&#64;</span> <span class="pre">B</span></code>) will default to the same CUTLASS algorithm used in <code class="docutils literal notranslate"><span class="pre">wp.matmul</span></code>.</p>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.matmul">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">matmul</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">a</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">b</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">c</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">d</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">beta</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">allow_tf32x3_arith</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.matmul" title="Link to this definition">#</a></dt>
-<dd><p>Computes a generic matrix-matrix multiplication (GEMM) of the form: <cite>d = alpha * (a &#64; b) + beta * c</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>a</strong> (<em>array2d</em>) – two-dimensional array containing matrix A</p></li>
-<li><p><strong>b</strong> (<em>array2d</em>) – two-dimensional array containing matrix B</p></li>
-<li><p><strong>c</strong> (<em>array2d</em>) – two-dimensional array containing matrix C</p></li>
-<li><p><strong>d</strong> (<em>array2d</em>) – two-dimensional array to which output D is written</p></li>
-<li><p><strong>alpha</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – parameter alpha of GEMM</p></li>
-<li><p><strong>beta</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – parameter beta of GEMM</p></li>
-<li><p><strong>allow_tf32x3_arith</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – whether to use CUTLASS’s 3xTF32 GEMMs, which enable accuracy similar to FP32
-while using Tensor Cores</p></li>
-<li><p><strong>device</strong> – device we want to use to multiply matrices. Defaults to active runtime device. If “cpu”, resorts to using numpy multiplication.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-</section>
-<section id="data-types">
-<h2>Data Types<a class="headerlink" href="#data-types" title="Link to this heading">#</a></h2>
-<section id="scalar-types">
-<h3>Scalar Types<a class="headerlink" href="#scalar-types" title="Link to this heading">#</a></h3>
-<p>The following scalar storage types are supported for array structures:</p>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>bool</p></td>
-<td><p>boolean</p></td>
-</tr>
-<tr class="row-even"><td><p>int8</p></td>
-<td><p>signed byte</p></td>
-</tr>
-<tr class="row-odd"><td><p>uint8</p></td>
-<td><p>unsigned byte</p></td>
-</tr>
-<tr class="row-even"><td><p>int16</p></td>
-<td><p>signed short</p></td>
-</tr>
-<tr class="row-odd"><td><p>uint16</p></td>
-<td><p>unsigned short</p></td>
-</tr>
-<tr class="row-even"><td><p>int32</p></td>
-<td><p>signed integer</p></td>
-</tr>
-<tr class="row-odd"><td><p>uint32</p></td>
-<td><p>unsigned integer</p></td>
-</tr>
-<tr class="row-even"><td><p>int64</p></td>
-<td><p>signed long integer</p></td>
-</tr>
-<tr class="row-odd"><td><p>uint64</p></td>
-<td><p>unsigned long integer</p></td>
-</tr>
-<tr class="row-even"><td><p>float32</p></td>
-<td><p>single precision float</p></td>
-</tr>
-<tr class="row-odd"><td><p>float64</p></td>
-<td><p>double precision float</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<p>Warp supports <code class="docutils literal notranslate"><span class="pre">float</span></code> and <code class="docutils literal notranslate"><span class="pre">int</span></code> as aliases for <code class="docutils literal notranslate"><span class="pre">wp.float32</span></code> and <code class="docutils literal notranslate"><span class="pre">wp.int32</span></code> respectively.</p>
-</section>
-<section id="vectors">
-<span id="vec"></span><h3>Vectors<a class="headerlink" href="#vectors" title="Link to this heading">#</a></h3>
-<p>Warp provides built-in math and geometry types for common simulation and graphics problems.
-A full reference for operators and functions for these types is available in the <a class="reference internal" href="functions.html"><span class="doc">Kernel Reference</span></a>.</p>
-<p>Warp supports vectors of numbers with an arbitrary length/numeric type. The built in concrete types are as follows:</p>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>vec2 vec3 vec4</p></td>
-<td><p>2d, 3d, 4d vector of default precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>vec2f vec3f vec4f</p></td>
-<td><p>2d, 3d, 4d vector of single precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>vec2d vec3d vec4d</p></td>
-<td><p>2d, 3d, 4d vector of double precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>vec2h vec3h vec4h</p></td>
-<td><p>2d, 3d, 4d vector of half precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>vec2ub vec3ub vec4ub</p></td>
-<td><p>2d, 3d, 4d vector of half precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>spatial_vector</p></td>
-<td><p>6d vector of single precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>spatial_vectorf</p></td>
-<td><p>6d vector of single precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>spatial_vectord</p></td>
-<td><p>6d vector of double precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>spatial_vectorh</p></td>
-<td><p>6d vector of half precision floats</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<p>Vectors support most standard linear algebra operations, e.g.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-
-    <span class="c1"># basis vectors</span>
-    <span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">)</span>
-    <span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">0.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">)</span>
-
-    <span class="c1"># take the cross product</span>
-    <span class="n">c</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">cross</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">)</span>
-
-    <span class="c1"># compute</span>
-    <span class="n">r</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">dot</span><span class="p">(</span><span class="n">c</span><span class="p">,</span> <span class="n">c</span><span class="p">)</span>
-
-    <span class="o">...</span>
-</pre></div>
-</div>
-<p>It’s possible to declare additional vector types with different lengths and data types. This is done in outside of kernels in <em>Python scope</em> using <code class="docutils literal notranslate"><span class="pre">warp.types.vector()</span></code>, for example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># declare a new vector type for holding 5 double precision floats:</span>
-<span class="n">vec5d</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">types</span><span class="o">.</span><span class="n">vector</span><span class="p">(</span><span class="n">length</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Once declared, the new type can be used when allocating arrays or inside kernels:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># create an array of vec5d</span>
-<span class="n">arr</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">vec5d</span><span class="p">)</span>
-
-<span class="c1"># use inside a kernel</span>
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-
-    <span class="c1"># zero initialize a custom named vector type</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">vec5d</span><span class="p">()</span>
-    <span class="o">...</span>
-
-    <span class="c1"># component-wise initialize a named vector type</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">vec5d</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span>
-              <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">2.0</span><span class="p">),</span>
-              <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">3.0</span><span class="p">),</span>
-              <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">4.0</span><span class="p">),</span>
-              <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">5.0</span><span class="p">))</span>
-  <span class="o">...</span>
-</pre></div>
-</div>
-<p>In addition, it’s possible to directly create <em>anonymously</em> typed instances of these vectors without declaring their type in advance. In this case the type will be inferred by the constructor arguments. For example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-
-    <span class="c1"># zero initialize vector of 5 doubles:</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vector</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">,</span> <span class="n">length</span><span class="o">=</span><span class="mi">5</span><span class="p">)</span>
-
-    <span class="c1"># scalar initialize a vector of 5 doubles to the same value:</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vector</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span> <span class="n">length</span><span class="o">=</span><span class="mi">5</span><span class="p">)</span>
-
-    <span class="c1"># component-wise initialize a vector of 5 doubles</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vector</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span>
-                  <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">2.0</span><span class="p">),</span>
-                  <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">3.0</span><span class="p">),</span>
-                  <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">4.0</span><span class="p">),</span>
-                  <span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">(</span><span class="mf">5.0</span><span class="p">))</span>
-</pre></div>
-</div>
-<p>These can be used with all the standard vector arithmetic operators, e.g.: <code class="docutils literal notranslate"><span class="pre">+</span></code>, <code class="docutils literal notranslate"><span class="pre">-</span></code>, scalar multiplication, and can also be transformed using matrices with compatible dimensions, potentially returning vectors with a different length.</p>
-</section>
-<section id="matrices">
-<span id="mat"></span><h3>Matrices<a class="headerlink" href="#matrices" title="Link to this heading">#</a></h3>
-<p>Matrices with arbitrary shapes/numeric types are also supported. The built in concrete matrix types are as follows:</p>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>mat22 mat33 mat44</p></td>
-<td><p>2,3 and 4d square matrix of default precision</p></td>
-</tr>
-<tr class="row-even"><td><p>mat22f mat33f mat44f</p></td>
-<td><p>2,3 and 4d square matrix of single precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>mat22d mat33d mat44d</p></td>
-<td><p>2,3 and 4d square matrix of double precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>mat22h mat33h mat44h</p></td>
-<td><p>2,3 and 4d square matrix of half precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>spatial_matrix</p></td>
-<td><p>6x6 matrix of single precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>spatial_matrixf</p></td>
-<td><p>6x6 matrix of single precision floats</p></td>
-</tr>
-<tr class="row-odd"><td><p>spatial_matrixd</p></td>
-<td><p>6x6 matrix of double precision floats</p></td>
-</tr>
-<tr class="row-even"><td><p>spatial_matrixh</p></td>
-<td><p>6x6 matrix of half precision floats</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<p>Matrices are stored in row-major format and support most standard linear algebra operations:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-
-    <span class="c1"># initialize matrix</span>
-    <span class="n">m</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">mat22</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span>
-                 <span class="mf">3.0</span><span class="p">,</span> <span class="mf">4.0</span><span class="p">)</span>
-
-    <span class="c1"># compute inverse</span>
-    <span class="n">minv</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">inverse</span><span class="p">(</span><span class="n">m</span><span class="p">)</span>
-
-    <span class="c1"># transform vector</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">minv</span> <span class="o">*</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec2</span><span class="p">(</span><span class="mf">0.5</span><span class="p">,</span> <span class="mf">0.3</span><span class="p">)</span>
-
-    <span class="o">...</span>
-</pre></div>
-</div>
-<p>In a similar manner to vectors, it’s possible to declare new matrix types with arbitrary shapes and data types using <code class="docutils literal notranslate"><span class="pre">wp.types.matrix()</span></code>, for example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># declare a new 3x2 half precision float matrix type:</span>
-<span class="n">mat32h</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">types</span><span class="o">.</span><span class="n">matrix</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span><span class="mi">2</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-
-<span class="c1"># create an array of this type</span>
-<span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">mat32h</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>These can be used inside a kernel:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-    <span class="o">...</span>
-
-    <span class="c1"># initialize a mat32h matrix</span>
-    <span class="n">m</span> <span class="o">=</span> <span class="n">mat32h</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">2.0</span><span class="p">),</span>
-               <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">3.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">4.0</span><span class="p">),</span>
-               <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">5.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">6.0</span><span class="p">))</span>
-
-    <span class="c1"># declare a 2 component half precision vector</span>
-    <span class="n">v2</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec2h</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">))</span>
-
-    <span class="c1"># multiply by the matrix, returning a 3 component vector:</span>
-    <span class="n">v3</span> <span class="o">=</span> <span class="n">m</span> <span class="o">*</span> <span class="n">v2</span>
-    <span class="o">...</span>
-</pre></div>
-</div>
-<p>It’s also possible to directly create anonymously typed instances inside kernels where the type is inferred from constructor arguments as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-    <span class="o">...</span>
-
-    <span class="c1"># create a 3x2 half precision matrix from components (row major ordering):</span>
-    <span class="n">m</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">matrix</span><span class="p">(</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">2.0</span><span class="p">),</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">2.0</span><span class="p">),</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">2.0</span><span class="p">),</span>
-        <span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span><span class="mi">2</span><span class="p">))</span>
-
-    <span class="c1"># zero initialize a 3x2 half precision matrix:</span>
-    <span class="n">m</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">matrix</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">),</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="mi">3</span><span class="p">,</span><span class="mi">2</span><span class="p">))</span>
-
-    <span class="c1"># create a 5x5 double precision identity matrix:</span>
-    <span class="n">m</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">identity</span><span class="p">(</span><span class="n">n</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>As with vectors, you can do standard matrix arithmetic with these variables, along with multiplying matrices with compatible shapes and potentially returning a matrix with a new shape.</p>
-</section>
-<section id="quaternions">
-<span id="quat"></span><h3>Quaternions<a class="headerlink" href="#quaternions" title="Link to this heading">#</a></h3>
-<p>Warp supports quaternions with the layout <code class="docutils literal notranslate"><span class="pre">i,</span> <span class="pre">j,</span> <span class="pre">k,</span> <span class="pre">w</span></code> where <code class="docutils literal notranslate"><span class="pre">w</span></code> is the real part. Here are the built in concrete quaternion types:</p>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>quat</p></td>
-<td><p>Default precision floating point quaternion</p></td>
-</tr>
-<tr class="row-even"><td><p>quatf</p></td>
-<td><p>Single precision floating point quaternion</p></td>
-</tr>
-<tr class="row-odd"><td><p>quatd</p></td>
-<td><p>Double precision floating point quaternion</p></td>
-</tr>
-<tr class="row-even"><td><p>quath</p></td>
-<td><p>Half precision floating point quaternion</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<p>Quaternions can be used to transform vectors as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-    <span class="o">...</span>
-
-    <span class="c1"># construct a 30 degree rotation around the x-axis</span>
-    <span class="n">q</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">quat_from_axis_angle</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">),</span> <span class="n">wp</span><span class="o">.</span><span class="n">degrees</span><span class="p">(</span><span class="mf">30.0</span><span class="p">))</span>
-
-    <span class="c1"># rotate an axis by this quaternion</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">quat_rotate</span><span class="p">(</span><span class="n">q</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">0.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">))</span>
-</pre></div>
-</div>
-<p>As with vectors and matrices, you can declare quaternion types with an arbitrary numeric type like so:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">quatd</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">types</span><span class="o">.</span><span class="n">quaternion</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>You can also create identity quaternion and anonymously typed instances inside a kernel like so:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-    <span class="o">...</span>
-
-    <span class="c1"># create a double precision identity quaternion:</span>
-    <span class="n">qd</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">quat_identity</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-
-    <span class="c1"># precision defaults to wp.float32 so this creates a single precision identity quaternion:</span>
-    <span class="n">qf</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">quat_identity</span><span class="p">()</span>
-
-    <span class="c1"># create a half precision quaternion from components, or a vector/scalar:</span>
-    <span class="n">qh</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">quaternion</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">),</span>
-                       <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">),</span>
-                       <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">),</span>
-                       <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">))</span>
-
-
-    <span class="n">qh</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">quaternion</span><span class="p">(</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">vector</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">),</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">),</span><span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">0.0</span><span class="p">)),</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">float16</span><span class="p">(</span><span class="mf">1.0</span><span class="p">))</span>
-</pre></div>
-</div>
-</section>
-<section id="transforms">
-<span id="transform"></span><h3>Transforms<a class="headerlink" href="#transforms" title="Link to this heading">#</a></h3>
-<p>Transforms are 7d vectors of floats representing a spatial rigid body transformation in format (p, q) where p is a 3d vector, and q is a quaternion.</p>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>transform</p></td>
-<td><p>Default precision floating point transform</p></td>
-</tr>
-<tr class="row-even"><td><p>transformf</p></td>
-<td><p>Single precision floating point transform</p></td>
-</tr>
-<tr class="row-odd"><td><p>transformd</p></td>
-<td><p>Double precision floating point transform</p></td>
-</tr>
-<tr class="row-even"><td><p>transformh</p></td>
-<td><p>Half precision floating point transform</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<p>Transforms can be constructed inside kernels from translation and rotation parts:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-    <span class="o">...</span>
-
-    <span class="c1"># create a transform from a vector/quaternion:</span>
-    <span class="n">t</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span>
-            <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">),</span>
-            <span class="n">wp</span><span class="o">.</span><span class="n">quat_from_axis_angle</span><span class="p">(</span><span class="mf">0.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">degrees</span><span class="p">(</span><span class="mf">30.0</span><span class="p">)))</span>
-
-    <span class="c1"># transform a point</span>
-    <span class="n">p</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">transform_point</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">10.0</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">))</span>
-
-    <span class="c1"># transform a vector (ignore translation)</span>
-    <span class="n">p</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">transform_vector</span><span class="p">(</span><span class="n">t</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">10.0</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">))</span>
-</pre></div>
-</div>
-<p>As with vectors and matrices, you can declare transform types with an arbitrary numeric type using <code class="docutils literal notranslate"><span class="pre">wp.types.transformation()</span></code>, for example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">transformd</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">types</span><span class="o">.</span><span class="n">transformation</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>You can also create identity transforms and anonymously typed instances inside a kernel like so:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span> <span class="o">...</span> <span class="p">):</span>
-
-    <span class="c1"># create double precision identity transform:</span>
-    <span class="n">qd</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">transform_identity</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float64</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-<section id="structs">
-<h3>Structs<a class="headerlink" href="#structs" title="Link to this heading">#</a></h3>
-<p>Users can define custom structure types using the <code class="docutils literal notranslate"><span class="pre">&#64;wp.struct</span></code> decorator as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">struct</span>
-<span class="k">class</span> <span class="nc">MyStruct</span><span class="p">:</span>
-
-    <span class="n">param1</span><span class="p">:</span> <span class="nb">int</span>
-    <span class="n">param2</span><span class="p">:</span> <span class="nb">float</span>
-    <span class="n">param3</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Struct attributes must be annotated with their respective type. They can be constructed in Python scope and then passed to kernels as arguments:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">compute</span><span class="p">(</span><span class="n">args</span><span class="p">:</span> <span class="n">MyStruct</span><span class="p">):</span>
-
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="nb">print</span><span class="p">(</span><span class="n">args</span><span class="o">.</span><span class="n">param1</span><span class="p">)</span>
-    <span class="nb">print</span><span class="p">(</span><span class="n">args</span><span class="o">.</span><span class="n">param2</span><span class="p">)</span>
-    <span class="nb">print</span><span class="p">(</span><span class="n">args</span><span class="o">.</span><span class="n">param3</span><span class="p">[</span><span class="n">tid</span><span class="p">])</span>
-
-<span class="c1"># construct an instance of the struct in Python</span>
-<span class="n">s</span> <span class="o">=</span> <span class="n">MyStruct</span><span class="p">()</span>
-<span class="n">s</span><span class="o">.</span><span class="n">param1</span> <span class="o">=</span> <span class="mi">10</span>
-<span class="n">s</span><span class="o">.</span><span class="n">param2</span> <span class="o">=</span> <span class="mf">2.5</span>
-<span class="n">s</span><span class="o">.</span><span class="n">param3</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-
-<span class="c1"># pass to our compute kernel</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">compute</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">s</span><span class="p">])</span>
-</pre></div>
-</div>
-<p>An array of structs can be zero-initialized as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">MyStruct</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>An array of structs can also be initialized from a list of struct objects:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">([</span><span class="n">MyStruct</span><span class="p">(),</span> <span class="n">MyStruct</span><span class="p">(),</span> <span class="n">MyStruct</span><span class="p">()],</span> <span class="n">dtype</span><span class="o">=</span><span class="n">MyStruct</span><span class="p">)</span>
-</pre></div>
-</div>
-<section id="example-using-a-custom-struct-in-gradient-computation">
-<h4>Example: Using a custom struct in gradient computation<a class="headerlink" href="#example-using-a-custom-struct-in-gradient-computation" title="Link to this heading">#</a></h4>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-
-<span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">init</span><span class="p">()</span>
-
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">struct</span>
-<span class="k">class</span> <span class="nc">TestStruct</span><span class="p">:</span>
-    <span class="n">x</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span>
-    <span class="n">a</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-    <span class="n">b</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)</span>
-
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">test_kernel</span><span class="p">(</span><span class="n">s</span><span class="p">:</span> <span class="n">TestStruct</span><span class="p">):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="n">s</span><span class="o">.</span><span class="n">b</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">s</span><span class="o">.</span><span class="n">a</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">+</span> <span class="n">s</span><span class="o">.</span><span class="n">x</span>
-
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">loss_kernel</span><span class="p">(</span><span class="n">s</span><span class="p">:</span> <span class="n">TestStruct</span><span class="p">,</span> <span class="n">loss</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="n">v</span> <span class="o">=</span> <span class="n">s</span><span class="o">.</span><span class="n">b</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">atomic_add</span><span class="p">(</span><span class="n">loss</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="nb">float</span><span class="p">(</span><span class="n">tid</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span> <span class="o">*</span> <span class="p">(</span><span class="n">v</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="o">+</span> <span class="mf">2.0</span> <span class="o">*</span> <span class="n">v</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="o">+</span> <span class="mf">3.0</span> <span class="o">*</span> <span class="n">v</span><span class="p">[</span><span class="mi">2</span><span class="p">]))</span>
-
-
-<span class="c1"># create struct</span>
-<span class="n">ts</span> <span class="o">=</span> <span class="n">TestStruct</span><span class="p">()</span>
-
-<span class="c1"># set members</span>
-<span class="n">ts</span><span class="o">.</span><span class="n">x</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">)</span>
-<span class="n">ts</span><span class="o">.</span><span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">([[</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">3.0</span><span class="p">],</span> <span class="p">[</span><span class="mf">4.0</span><span class="p">,</span> <span class="mf">5.0</span><span class="p">,</span> <span class="mf">6.0</span><span class="p">]]),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="n">ts</span><span class="o">.</span><span class="n">b</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
-<span class="n">loss</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">test_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">ts</span><span class="p">])</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">loss_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="mi">2</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">ts</span><span class="p">,</span> <span class="n">loss</span><span class="p">])</span>
-
-<span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">loss</span><span class="p">)</span>
-
-<span class="nb">print</span><span class="p">(</span><span class="n">loss</span><span class="p">)</span>
-<span class="nb">print</span><span class="p">(</span><span class="n">tape</span><span class="o">.</span><span class="n">gradients</span><span class="p">[</span><span class="n">ts</span><span class="p">]</span><span class="o">.</span><span class="n">a</span><span class="p">)</span>
-</pre></div>
-</div>
-</section>
-</section>
-<section id="type-conversions">
-<h3>Type Conversions<a class="headerlink" href="#type-conversions" title="Link to this heading">#</a></h3>
-<p>Warp is particularly strict regarding type conversions and does not perform <em>any</em> implicit conversion between numeric types.
-The user is responsible for ensuring types for most arithmetic operators match, e.g.: <code class="docutils literal notranslate"><span class="pre">x</span> <span class="pre">=</span> <span class="pre">float(0.0)</span> <span class="pre">+</span> <span class="pre">int(4)</span></code> will result in an error.
-This can be surprising for users that are accustomed to C-style conversions but avoids a class of common bugs that result from implicit conversions.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Warp does not currently perform implicit type conversions between numeric types.
-Users should explicitly cast variables to compatible types using constructors like
-<code class="docutils literal notranslate"><span class="pre">int()</span></code>, <code class="docutils literal notranslate"><span class="pre">float()</span></code>, <code class="docutils literal notranslate"><span class="pre">wp.float16()</span></code>, <code class="docutils literal notranslate"><span class="pre">wp.uint8()</span></code>, etc.</p>
-</div>
-</section>
-</section>
-<section id="constants">
-<h2>Constants<a class="headerlink" href="#constants" title="Link to this heading">#</a></h2>
-<p>In general, Warp kernels cannot access variables in the global Python interpreter state. One exception to this is for compile-time constants, which may be declared globally (or as class attributes) and folded into the kernel definition.</p>
-<p>Constants are defined using the <code class="docutils literal notranslate"><span class="pre">wp.constant()</span></code> function. An example is shown below:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">TYPE_SPHERE</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">constant</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>
-<span class="n">TYPE_CUBE</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">constant</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>
-<span class="n">TYPE_CAPSULE</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">constant</span><span class="p">(</span><span class="mi">2</span><span class="p">)</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">collide</span><span class="p">(</span><span class="n">geometry</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">)):</span>
-
-    <span class="n">t</span> <span class="o">=</span> <span class="n">geometry</span><span class="p">[</span><span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()]</span>
-
-    <span class="k">if</span> <span class="p">(</span><span class="n">t</span> <span class="o">==</span> <span class="n">TYPE_SPHERE</span><span class="p">):</span>
-        <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;sphere&quot;</span><span class="p">)</span>
-    <span class="k">if</span> <span class="p">(</span><span class="n">t</span> <span class="o">==</span> <span class="n">TYPE_CUBE</span><span class="p">):</span>
-        <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;cube&quot;</span><span class="p">)</span>
-    <span class="k">if</span> <span class="p">(</span><span class="n">t</span> <span class="o">==</span> <span class="n">TYPE_CAPSULE</span><span class="p">):</span>
-        <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;capsule&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.constant">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">constant</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.constant" title="Link to this definition">#</a></dt>
-<dd><p>Function to declare compile-time constants accessible from Warp kernels</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>x</strong> – Compile-time constant value, can be any of the built-in math types.</p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="operators">
-<h2>Operators<a class="headerlink" href="#operators" title="Link to this heading">#</a></h2>
-<section id="boolean-operators">
-<h3>Boolean Operators<a class="headerlink" href="#boolean-operators" title="Link to this heading">#</a></h3>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>a and b</p></td>
-<td><p>True if a and b are True</p></td>
-</tr>
-<tr class="row-even"><td><p>a or b</p></td>
-<td><p>True if a or b is True</p></td>
-</tr>
-<tr class="row-odd"><td><p>not a</p></td>
-<td><p>True if a is False, otherwise False</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Expressions such as <code class="docutils literal notranslate"><span class="pre">if</span> <span class="pre">(a</span> <span class="pre">and</span> <span class="pre">b):</span></code> currently do not perform short-circuit evaluation.
-In this case <code class="docutils literal notranslate"><span class="pre">b</span></code> will also be evaluated even when <code class="docutils literal notranslate"><span class="pre">a</span></code> is <code class="docutils literal notranslate"><span class="pre">False</span></code>.
-Users should take care to ensure that secondary conditions are safe to evaluate (e.g.: do not index out of bounds) in all cases.</p>
-</div>
-</section>
-<section id="comparison-operators">
-<h3>Comparison Operators<a class="headerlink" href="#comparison-operators" title="Link to this heading">#</a></h3>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>a &gt; b</p></td>
-<td><p>True if a strictly greater than b</p></td>
-</tr>
-<tr class="row-even"><td><p>a &lt; b</p></td>
-<td><p>True if a strictly less than b</p></td>
-</tr>
-<tr class="row-odd"><td><p>a &gt;= b</p></td>
-<td><p>True if a greater than or equal to b</p></td>
-</tr>
-<tr class="row-even"><td><p>a &lt;= b</p></td>
-<td><p>True if a less than or equal to b</p></td>
-</tr>
-<tr class="row-odd"><td><p>a == b</p></td>
-<td><p>True if a equals b</p></td>
-</tr>
-<tr class="row-even"><td><p>a != b</p></td>
-<td><p>True if a not equal to b</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</section>
-<section id="arithmetic-operators">
-<h3>Arithmetic Operators<a class="headerlink" href="#arithmetic-operators" title="Link to this heading">#</a></h3>
-<div class="table-wrapper docutils container">
-<table class="docutils align-default">
-<tbody>
-<tr class="row-odd"><td><p>a + b</p></td>
-<td><p>Addition</p></td>
-</tr>
-<tr class="row-even"><td><p>a - b</p></td>
-<td><p>Subtraction</p></td>
-</tr>
-<tr class="row-odd"><td><p>a * b</p></td>
-<td><p>Multiplication</p></td>
-</tr>
-<tr class="row-even"><td><p>a / b</p></td>
-<td><p>Floating point division</p></td>
-</tr>
-<tr class="row-odd"><td><p>a // b</p></td>
-<td><p>Floored division</p></td>
-</tr>
-<tr class="row-even"><td><p>a ** b</p></td>
-<td><p>Exponentiation</p></td>
-</tr>
-<tr class="row-odd"><td><p>a % b</p></td>
-<td><p>Modulus</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Since implicit conversions are not performed arguments types to operators should match.
-Users should use type constructors, e.g.: <code class="docutils literal notranslate"><span class="pre">float()</span></code>, <code class="docutils literal notranslate"><span class="pre">int()</span></code>, <code class="docutils literal notranslate"><span class="pre">wp.int64()</span></code>, etc. to cast variables
-to the correct type. Also note that the multiplication expression <code class="docutils literal notranslate"><span class="pre">a</span> <span class="pre">*</span> <span class="pre">b</span></code> is used to represent scalar
-multiplication and matrix multiplication. The <code class="docutils literal notranslate"><span class="pre">&#64;</span></code> operator is not currently supported.</p>
-</div>
-</section>
-</section>
-<section id="graphs">
-<h2>Graphs<a class="headerlink" href="#graphs" title="Link to this heading">#</a></h2>
-<p>Launching kernels from Python introduces significant additional overhead compared to C++ or native programs.
-To address this, Warp exposes the concept of <a class="reference external" href="https://developer.nvidia.com/blog/cuda-graphs/">CUDA graphs</a>
-to allow recording large batches of kernels and replaying them with very little CPU overhead.</p>
-<p>To record a series of kernel launches use the <a class="reference internal" href="#warp.capture_begin" title="warp.capture_begin"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.capture_begin()</span></code></a> and
-<a class="reference internal" href="#warp.capture_end" title="warp.capture_end"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.capture_end()</span></code></a> API as follows:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="c1"># begin capture</span>
-<span class="n">wp</span><span class="o">.</span><span class="n">capture_begin</span><span class="p">()</span>
-
-<span class="k">try</span><span class="p">:</span>
-    <span class="c1"># record launches</span>
-    <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">100</span><span class="p">):</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">compute1</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-<span class="k">finally</span><span class="p">:</span>
-    <span class="c1"># end capture and return a graph object</span>
-    <span class="n">graph</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">capture_end</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>We strongly recommend the use of the the try-finally pattern when capturing graphs because
-<a class="reference internal" href="#warp.capture_begin" title="warp.capture_begin"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.capture_begin</span></code></a> disables Python garbage collection so that Warp objects are not
-garbage-collected during graph capture (CUDA does not allow memory to be deallocated during graph capture).
-<a class="reference internal" href="#warp.capture_end" title="warp.capture_end"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.capture_end</span></code></a> reenables garbage collection.</p>
-<p>Once a graph has been constructed it can be executed:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">wp</span><span class="o">.</span><span class="n">capture_launch</span><span class="p">(</span><span class="n">graph</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Note that only launch calls are recorded in the graph, any Python executed outside of the kernel code will not be recorded.
-Typically it is only beneficial to use CUDA graphs when the graph will be reused or launched multiple times.</p>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.capture_begin">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">capture_begin</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stream</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">force_module_load</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.capture_begin" title="Link to this definition">#</a></dt>
-<dd><p>Begin capture of a CUDA graph</p>
-<p>Captures all subsequent kernel launches and memory operations on CUDA devices.
-This can be used to record large numbers of kernels and replay them with low-overhead.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The device to capture on, if None the current CUDA device will be used</p></li>
-<li><p><strong>stream</strong> – The CUDA stream to capture on</p></li>
-<li><p><strong>force_module_load</strong> – Whether or not to force loading of all kernels before capture, in general it is better to use <code class="xref py py-func docutils literal notranslate"><span class="pre">load_module()</span></code> to selectively load kernels.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.capture_end">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">capture_end</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stream</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.capture_end" title="Link to this definition">#</a></dt>
-<dd><p>Ends the capture of a CUDA graph</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns<span class="colon">:</span></dt>
-<dd class="field-odd"><p>A handle to a CUDA graph object that can be launched with <a class="reference internal" href="#warp.capture_launch" title="warp.capture_launch"><code class="xref py py-func docutils literal notranslate"><span class="pre">capture_launch()</span></code></a></p>
-</dd>
-<dt class="field-even">Parameters<span class="colon">:</span></dt>
-<dd class="field-even"><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – </p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><em>Graph</em></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.capture_launch">
-<span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">capture_launch</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">graph</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stream</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.capture_launch" title="Link to this definition">#</a></dt>
-<dd><p>Launch a previously captured CUDA graph</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>graph</strong> (<em>Graph</em>) – A Graph as returned by <a class="reference internal" href="#warp.capture_end" title="warp.capture_end"><code class="xref py py-func docutils literal notranslate"><span class="pre">capture_end()</span></code></a></p></li>
-<li><p><strong>stream</strong> (<em>Stream</em><em> | </em><em>None</em>) – A Stream to launch the graph on (optional)</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="bounding-value-hierarchies-bvh">
-<h2>Bounding Value Hierarchies (BVH)<a class="headerlink" href="#bounding-value-hierarchies-bvh" title="Link to this heading">#</a></h2>
-<p>The <a class="reference internal" href="#warp.Bvh" title="warp.Bvh"><code class="xref py py-class docutils literal notranslate"><span class="pre">wp.Bvh</span></code></a> class can be used to create a BVH for a group of bounding volumes. This object can then be traversed
-to determine which parts are intersected by a ray using <a class="reference internal" href="functions.html#warp.bvh_query_ray" title="warp.bvh_query_ray"><code class="xref py py-func docutils literal notranslate"><span class="pre">bvh_query_ray()</span></code></a> and which parts are fully contained
-within a certain bounding volume using <a class="reference internal" href="functions.html#warp.bvh_query_aabb" title="warp.bvh_query_aabb"><code class="xref py py-func docutils literal notranslate"><span class="pre">bvh_query_aabb()</span></code></a>.</p>
-<p>The following snippet demonstrates how to create a <a class="reference internal" href="#warp.Bvh" title="warp.Bvh"><code class="xref py py-class docutils literal notranslate"><span class="pre">wp.Bvh</span></code></a> object from 100 random bounding volumes:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="n">rng</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">default_rng</span><span class="p">(</span><span class="mi">123</span><span class="p">)</span>
-
-<span class="n">num_bounds</span> <span class="o">=</span> <span class="mi">100</span>
-<span class="n">lowers</span> <span class="o">=</span> <span class="n">rng</span><span class="o">.</span><span class="n">random</span><span class="p">(</span><span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="n">num_bounds</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span> <span class="o">*</span> <span class="mf">5.0</span>
-<span class="n">uppers</span> <span class="o">=</span> <span class="n">lowers</span> <span class="o">+</span> <span class="n">rng</span><span class="o">.</span><span class="n">random</span><span class="p">(</span><span class="n">size</span><span class="o">=</span><span class="p">(</span><span class="n">num_bounds</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span> <span class="o">*</span> <span class="mf">5.0</span>
-
-<span class="n">device_lowers</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">lowers</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-<span class="n">device_uppers</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">uppers</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-
-<span class="n">bvh</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Bvh</span><span class="p">(</span><span class="n">device_lowers</span><span class="p">,</span> <span class="n">device_uppers</span><span class="p">)</span>
-</pre></div>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Bvh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Bvh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">lowers</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">uppers</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Bvh" title="Link to this definition">#</a></dt>
-<dd><p>Class representing a bounding volume hierarchy.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.Bvh.id">
-<span class="sig-name descname"><span class="pre">id</span></span><a class="headerlink" href="#warp.Bvh.id" title="Link to this definition">#</a></dt>
-<dd><p>Unique identifier for this bvh object, can be passed to kernels.</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.Bvh.device">
-<span class="sig-name descname"><span class="pre">device</span></span><a class="headerlink" href="#warp.Bvh.device" title="Link to this definition">#</a></dt>
-<dd><p>Device this object lives on, all buffers must live on the same device.</p>
-</dd></dl>
-
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>lowers</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of lower bounds <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code></p></li>
-<li><p><strong>uppers</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of upper bounds <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code></p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Bvh.refit">
-<span class="sig-name descname"><span class="pre">refit</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Bvh.refit" title="Link to this definition">#</a></dt>
-<dd><p>Refit the BVH. This should be called after users modify the <cite>lowers</cite> and <cite>uppers</cite> arrays.</p>
-</dd></dl>
-
-</dd></dl>
-
-<section id="example-bvh-ray-traversal">
-<h3>Example: BVH Ray Traversal<a class="headerlink" href="#example-bvh-ray-traversal" title="Link to this heading">#</a></h3>
-<p>An example of performing a ray traversal on the data structure is as follows:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">bvh_query_ray</span><span class="p">(</span>
-    <span class="n">bvh_id</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">uint64</span><span class="p">,</span>
-    <span class="n">start</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span>
-    <span class="nb">dir</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span>
-    <span class="n">bounds_intersected</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">bool</span><span class="p">),</span>
-<span class="p">):</span>
-    <span class="n">query</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">bvh_query_ray</span><span class="p">(</span><span class="n">bvh_id</span><span class="p">,</span> <span class="n">start</span><span class="p">,</span> <span class="nb">dir</span><span class="p">)</span>
-    <span class="n">bounds_nr</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">int32</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>
-
-    <span class="k">while</span> <span class="n">wp</span><span class="o">.</span><span class="n">bvh_query_next</span><span class="p">(</span><span class="n">query</span><span class="p">,</span> <span class="n">bounds_nr</span><span class="p">):</span>
-        <span class="c1"># The ray intersects the volume with index bounds_nr</span>
-        <span class="n">bounds_intersected</span><span class="p">[</span><span class="n">bounds_nr</span><span class="p">]</span> <span class="o">=</span> <span class="kc">True</span>
-
-
-<span class="n">bounds_intersected</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="n">num_bounds</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">bool</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-<span class="n">query_start</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">)</span>
-<span class="n">query_dir</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">normalize</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">))</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span>
-    <span class="n">kernel</span><span class="o">=</span><span class="n">bvh_query_ray</span><span class="p">,</span>
-    <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
-    <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">bvh</span><span class="o">.</span><span class="n">id</span><span class="p">,</span> <span class="n">query_start</span><span class="p">,</span> <span class="n">query_dir</span><span class="p">,</span> <span class="n">bounds_intersected</span><span class="p">],</span>
-    <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">,</span>
-<span class="p">)</span>
-</pre></div>
-</div>
-<p>The Warp kernel <code class="docutils literal notranslate"><span class="pre">bvh_query_ray</span></code> is launched with a single thread, provided the unique <a class="reference internal" href="functions.html#warp.uint64" title="warp.uint64"><code class="xref py py-class docutils literal notranslate"><span class="pre">uint64</span></code></a>
-identifier of the <a class="reference internal" href="#warp.Bvh" title="warp.Bvh"><code class="xref py py-class docutils literal notranslate"><span class="pre">wp.Bvh</span></code></a> object, parameters describing the ray, and an array to store the results.
-In <code class="docutils literal notranslate"><span class="pre">bvh_query_ray</span></code>, <a class="reference internal" href="functions.html#warp.bvh_query_ray" title="warp.bvh_query_ray"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.bvh_query_ray()</span></code></a> is called once to obtain an object that is stored in the
-variable <code class="docutils literal notranslate"><span class="pre">query</span></code>. An integer is also allocated as <code class="docutils literal notranslate"><span class="pre">bounds_nr</span></code> to store the volume index of the traversal.
-A while statement is used for the actual traversal using <a class="reference internal" href="functions.html#warp.bvh_query_next" title="warp.bvh_query_next"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.bvh_query_next()</span></code></a>,
-which returns <code class="docutils literal notranslate"><span class="pre">True</span></code> as long as there are intersecting bounds.</p>
-</section>
-<section id="example-bvh-volume-traversal">
-<h3>Example: BVH Volume Traversal<a class="headerlink" href="#example-bvh-volume-traversal" title="Link to this heading">#</a></h3>
-<p>Similar to the ray-traversal example, we can perform volume traversal to find the volumes that are fully contained
-within a specified bounding box.</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">bvh_query_aabb</span><span class="p">(</span>
-    <span class="n">bvh_id</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">uint64</span><span class="p">,</span>
-    <span class="n">lower</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span>
-    <span class="n">upper</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span>
-    <span class="n">bounds_intersected</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">bool</span><span class="p">),</span>
-<span class="p">):</span>
-    <span class="n">query</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">bvh_query_aabb</span><span class="p">(</span><span class="n">bvh_id</span><span class="p">,</span> <span class="n">lower</span><span class="p">,</span> <span class="n">upper</span><span class="p">)</span>
-    <span class="n">bounds_nr</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">int32</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>
-
-    <span class="k">while</span> <span class="n">wp</span><span class="o">.</span><span class="n">bvh_query_next</span><span class="p">(</span><span class="n">query</span><span class="p">,</span> <span class="n">bounds_nr</span><span class="p">):</span>
-        <span class="c1"># The volume with index bounds_nr is fully contained</span>
-        <span class="c1"># in the (lower,upper) bounding box</span>
-        <span class="n">bounds_intersected</span><span class="p">[</span><span class="n">bounds_nr</span><span class="p">]</span> <span class="o">=</span> <span class="kc">True</span>
-
-
-<span class="n">bounds_intersected</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">shape</span><span class="o">=</span><span class="p">(</span><span class="n">num_bounds</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">bool</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">)</span>
-<span class="n">query_lower</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">4.0</span><span class="p">,</span> <span class="mf">4.0</span><span class="p">,</span> <span class="mf">4.0</span><span class="p">)</span>
-<span class="n">query_upper</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">6.0</span><span class="p">,</span> <span class="mf">6.0</span><span class="p">,</span> <span class="mf">6.0</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span>
-    <span class="n">kernel</span><span class="o">=</span><span class="n">bvh_query_aabb</span><span class="p">,</span>
-    <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
-    <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">bvh</span><span class="o">.</span><span class="n">id</span><span class="p">,</span> <span class="n">query_lower</span><span class="p">,</span> <span class="n">query_upper</span><span class="p">,</span> <span class="n">bounds_intersected</span><span class="p">],</span>
-    <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda:0&quot;</span><span class="p">,</span>
-<span class="p">)</span>
-</pre></div>
-</div>
-<p>The kernel is nearly identical to the ray-traversal example, except we obtain <code class="docutils literal notranslate"><span class="pre">query</span></code> using
-<a class="reference internal" href="functions.html#warp.bvh_query_aabb" title="warp.bvh_query_aabb"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.bvh_query_aabb()</span></code></a>.</p>
-</section>
-</section>
-<section id="meshes">
-<h2>Meshes<a class="headerlink" href="#meshes" title="Link to this heading">#</a></h2>
-<p>Warp provides a <code class="docutils literal notranslate"><span class="pre">wp.Mesh</span></code> class to manage triangle mesh data. To create a mesh users provide a points, indices and optionally a velocity array:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">mesh</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Mesh</span><span class="p">(</span><span class="n">points</span><span class="p">,</span> <span class="n">indices</span><span class="p">,</span> <span class="n">velocities</span><span class="p">)</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Mesh objects maintain references to their input geometry buffers. All buffers should live on the same device.</p>
-</div>
-<p>Meshes can be passed to kernels using their <code class="docutils literal notranslate"><span class="pre">id</span></code> attribute which uniquely identifies the mesh by a unique <code class="docutils literal notranslate"><span class="pre">uint64</span></code> value.
-Once inside a kernel you can perform geometric queries against the mesh such as ray-casts or closest point lookups:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">raycast</span><span class="p">(</span><span class="n">mesh</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">uint64</span><span class="p">,</span>
-            <span class="n">ray_origin</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-            <span class="n">ray_dir</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-            <span class="n">ray_hit</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">)):</span>
-
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="n">t</span> <span class="o">=</span> <span class="nb">float</span><span class="p">(</span><span class="mf">0.0</span><span class="p">)</span>      <span class="c1"># hit distance along ray</span>
-    <span class="n">u</span> <span class="o">=</span> <span class="nb">float</span><span class="p">(</span><span class="mf">0.0</span><span class="p">)</span>      <span class="c1"># hit face barycentric u</span>
-    <span class="n">v</span> <span class="o">=</span> <span class="nb">float</span><span class="p">(</span><span class="mf">0.0</span><span class="p">)</span>      <span class="c1"># hit face barycentric v</span>
-    <span class="n">sign</span> <span class="o">=</span> <span class="nb">float</span><span class="p">(</span><span class="mf">0.0</span><span class="p">)</span>   <span class="c1"># hit face sign</span>
-    <span class="n">n</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">()</span>       <span class="c1"># hit face normal</span>
-    <span class="n">f</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>          <span class="c1"># hit face index</span>
-
-    <span class="n">color</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">()</span>
-
-    <span class="c1"># ray cast against the mesh</span>
-    <span class="k">if</span> <span class="n">wp</span><span class="o">.</span><span class="n">mesh_query_ray</span><span class="p">(</span><span class="n">mesh</span><span class="p">,</span> <span class="n">ray_origin</span><span class="p">[</span><span class="n">tid</span><span class="p">],</span> <span class="n">ray_dir</span><span class="p">[</span><span class="n">tid</span><span class="p">],</span> <span class="mf">1.e+6</span><span class="p">,</span> <span class="n">t</span><span class="p">,</span> <span class="n">u</span><span class="p">,</span> <span class="n">v</span><span class="p">,</span> <span class="n">sign</span><span class="p">,</span> <span class="n">n</span><span class="p">,</span> <span class="n">f</span><span class="p">):</span>
-
-        <span class="c1"># if we got a hit then set color to the face normal</span>
-        <span class="n">color</span> <span class="o">=</span> <span class="n">n</span><span class="o">*</span><span class="mf">0.5</span> <span class="o">+</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">(</span><span class="mf">0.5</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">,</span> <span class="mf">0.5</span><span class="p">)</span>
-
-    <span class="n">ray_hit</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">color</span>
-</pre></div>
-</div>
-<p>Users may update mesh vertex positions at runtime simply by modifying the points buffer.
-After modifying point locations users should call <code class="docutils literal notranslate"><span class="pre">Mesh.refit()</span></code> to rebuild the bounding volume hierarchy (BVH) structure and ensure that queries work correctly.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Updating Mesh topology (indices) at runtime is not currently supported. Users should instead recreate a new Mesh object.</p>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Mesh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Mesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">indices</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">velocities</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">support_winding_number</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Mesh" title="Link to this definition">#</a></dt>
-<dd><p>Class representing a triangle mesh.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.Mesh.id">
-<span class="sig-name descname"><span class="pre">id</span></span><a class="headerlink" href="#warp.Mesh.id" title="Link to this definition">#</a></dt>
-<dd><p>Unique identifier for this mesh object, can be passed to kernels.</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.Mesh.device">
-<span class="sig-name descname"><span class="pre">device</span></span><a class="headerlink" href="#warp.Mesh.device" title="Link to this definition">#</a></dt>
-<dd><p>Device this object lives on, all buffers must live on the same device.</p>
-</dd></dl>
-
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>points</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of vertex positions of type <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code></p></li>
-<li><p><strong>indices</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of triangle indices of type <a class="reference internal" href="functions.html#warp.int32" title="warp.int32"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.int32</span></code></a>, should be a 1d array with shape (num_tris, 3)</p></li>
-<li><p><strong>velocities</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of vertex velocities of type <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code> (optional)</p></li>
-<li><p><strong>support_winding_number</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If true the mesh will build additional datastructures to support <cite>wp.mesh_query_point_sign_winding_number()</cite> queries</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Mesh.refit">
-<span class="sig-name descname"><span class="pre">refit</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Mesh.refit" title="Link to this definition">#</a></dt>
-<dd><p>Refit the BVH to points. This should be called after users modify the <cite>points</cite> data.</p>
-</dd></dl>
-
-</dd></dl>
-
-</section>
-<section id="hash-grids">
-<h2>Hash Grids<a class="headerlink" href="#hash-grids" title="Link to this heading">#</a></h2>
-<p>Many particle-based simulation methods such as the Discrete Element Method (DEM), or Smoothed Particle Hydrodynamics (SPH), involve iterating over spatial neighbors to compute force interactions. Hash grids are a well-established data structure to accelerate these nearest neighbor queries, and particularly well-suited to the GPU.</p>
-<p>To support spatial neighbor queries Warp provides a <code class="docutils literal notranslate"><span class="pre">HashGrid</span></code> object that may be created as follows:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">grid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">HashGrid</span><span class="p">(</span><span class="n">dim_x</span><span class="o">=</span><span class="mi">128</span><span class="p">,</span> <span class="n">dim_y</span><span class="o">=</span><span class="mi">128</span><span class="p">,</span> <span class="n">dim_z</span><span class="o">=</span><span class="mi">128</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="n">points</span><span class="o">=</span><span class="n">p</span><span class="p">,</span> <span class="n">radius</span><span class="o">=</span><span class="n">r</span><span class="p">)</span>
-</pre></div>
-</div>
-<p><code class="docutils literal notranslate"><span class="pre">p</span></code> is an array of <code class="docutils literal notranslate"><span class="pre">wp.vec3</span></code> point positions, and <code class="docutils literal notranslate"><span class="pre">r</span></code> is the radius to use when building the grid.
-Neighbors can then be iterated over inside the kernel code using <a class="reference internal" href="functions.html#warp.hash_grid_query" title="warp.hash_grid_query"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.hash_grid_query()</span></code></a>
-and <a class="reference internal" href="functions.html#warp.hash_grid_query_next" title="warp.hash_grid_query_next"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.hash_grid_query_next()</span></code></a> as follows:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">sum</span><span class="p">(</span><span class="n">grid</span> <span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">uint64</span><span class="p">,</span>
-        <span class="n">points</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-        <span class="n">output</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-        <span class="n">radius</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
-
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="c1"># query point</span>
-    <span class="n">p</span> <span class="o">=</span> <span class="n">points</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span>
-
-    <span class="c1"># create grid query around point</span>
-    <span class="n">query</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">hash_grid_query</span><span class="p">(</span><span class="n">grid</span><span class="p">,</span> <span class="n">p</span><span class="p">,</span> <span class="n">radius</span><span class="p">)</span>
-    <span class="n">index</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="mi">0</span><span class="p">)</span>
-
-    <span class="nb">sum</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">()</span>
-
-    <span class="k">while</span><span class="p">(</span><span class="n">wp</span><span class="o">.</span><span class="n">hash_grid_query_next</span><span class="p">(</span><span class="n">query</span><span class="p">,</span> <span class="n">index</span><span class="p">)):</span>
-
-        <span class="n">neighbor</span> <span class="o">=</span> <span class="n">points</span><span class="p">[</span><span class="n">index</span><span class="p">]</span>
-
-        <span class="c1"># compute distance to neighbor point</span>
-        <span class="n">dist</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">length</span><span class="p">(</span><span class="n">p</span><span class="o">-</span><span class="n">neighbor</span><span class="p">)</span>
-        <span class="k">if</span> <span class="p">(</span><span class="n">dist</span> <span class="o">&lt;=</span> <span class="n">radius</span><span class="p">):</span>
-            <span class="nb">sum</span> <span class="o">+=</span> <span class="n">neighbor</span>
-
-    <span class="n">output</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="nb">sum</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The <code class="docutils literal notranslate"><span class="pre">HashGrid</span></code> query will give back all points in <em>cells</em> that fall inside the query radius.
-When there are hash conflicts it means that some points outside of query radius will be returned, and users should
-check the distance themselves inside their kernels. The reason the query doesn’t do the check itself for each
-returned point is because it’s common for kernels to compute the distance themselves, so it would redundant to
-check/compute the distance twice.</p>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.HashGrid">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">HashGrid</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dim_x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_y</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_z</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.HashGrid" title="Link to this definition">#</a></dt>
-<dd><p>Class representing a hash grid object for accelerated point queries.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.HashGrid.id">
-<span class="sig-name descname"><span class="pre">id</span></span><a class="headerlink" href="#warp.HashGrid.id" title="Link to this definition">#</a></dt>
-<dd><p>Unique identifier for this mesh object, can be passed to kernels.</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.HashGrid.device">
-<span class="sig-name descname"><span class="pre">device</span></span><a class="headerlink" href="#warp.HashGrid.device" title="Link to this definition">#</a></dt>
-<dd><p>Device this object lives on, all buffers must live on the same device.</p>
-</dd></dl>
-
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dim_x</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of cells in x-axis</p></li>
-<li><p><strong>dim_y</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of cells in y-axis</p></li>
-<li><p><strong>dim_z</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of cells in z-axis</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.HashGrid.build">
-<span class="sig-name descname"><span class="pre">build</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">points</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">radius</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.HashGrid.build" title="Link to this definition">#</a></dt>
-<dd><p>Updates the hash grid data structure.</p>
-<p>This method rebuilds the underlying datastructure and should be called any time the set
-of points changes.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>points</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of points of type <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code></p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The cell size to use for bucketing points, cells are cubes with edges of this width.
-For best performance the radius used to construct the grid should match closely to
-the radius used when performing queries.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-</section>
-<section id="volumes">
-<h2>Volumes<a class="headerlink" href="#volumes" title="Link to this heading">#</a></h2>
-<p>Sparse volumes are incredibly useful for representing grid data over large domains, such as signed distance fields
-(SDFs) for complex objects, or velocities for large-scale fluid flow. Warp supports reading sparse volumetric grids
-stored using the <a class="reference external" href="https://developer.nvidia.com/nanovdb">NanoVDB</a> standard. Users can access voxels directly
-or use built-in closest-point or trilinear interpolation to sample grid data from world or local space.</p>
-<p>Volume objects can be created directly from Warp arrays containing a NanoVDB grid, from the contents of a
-standard <code class="docutils literal notranslate"><span class="pre">.nvdb</span></code> file using <a class="reference internal" href="#warp.Volume.load_from_nvdb" title="warp.Volume.load_from_nvdb"><code class="xref py py-func docutils literal notranslate"><span class="pre">load_from_nvdb()</span></code></a>,
-or from a dense 3D NumPy array using <a class="reference internal" href="#warp.Volume.load_from_numpy" title="warp.Volume.load_from_numpy"><code class="xref py py-func docutils literal notranslate"><span class="pre">load_from_numpy()</span></code></a>.</p>
-<p>Volumes can also be created using <a class="reference internal" href="#warp.Volume.allocate" title="warp.Volume.allocate"><code class="xref py py-func docutils literal notranslate"><span class="pre">allocate()</span></code></a> or
-<a class="reference internal" href="#warp.Volume.allocate_by_tiles" title="warp.Volume.allocate_by_tiles"><code class="xref py py-func docutils literal notranslate"><span class="pre">allocate_by_tiles()</span></code></a>. The values for a Volume object can be modified in a Warp
-kernel using <a class="reference internal" href="functions.html#warp.volume_store_f" title="warp.volume_store_f"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.volume_store_f()</span></code></a>, <a class="reference internal" href="functions.html#warp.volume_store_v" title="warp.volume_store_v"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.volume_store_v()</span></code></a>, and
-<a class="reference internal" href="functions.html#warp.volume_store_i" title="warp.volume_store_i"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.volume_store_i()</span></code></a>.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Warp does not currently support modifying the topology of sparse volumes at runtime.</p>
-</div>
-<p>Below we give an example of creating a Volume object from an existing NanoVDB file:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># open NanoVDB file on disk</span>
-<span class="n">file</span> <span class="o">=</span> <span class="nb">open</span><span class="p">(</span><span class="s2">&quot;mygrid.nvdb&quot;</span><span class="p">,</span> <span class="s2">&quot;rb&quot;</span><span class="p">)</span>
-
-<span class="c1"># create Volume object</span>
-<span class="n">volume</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Volume</span><span class="o">.</span><span class="n">load_from_nvdb</span><span class="p">(</span><span class="n">file</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cpu&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Files written by the NanoVDB library, commonly marked by the <code class="docutils literal notranslate"><span class="pre">.nvdb</span></code> extension, can contain multiple grids with
-various compression methods, but a <a class="reference internal" href="#warp.Volume" title="warp.Volume"><code class="xref py py-class docutils literal notranslate"><span class="pre">Volume</span></code></a> object represents a single NanoVDB grid therefore only files with
-a single grid are supported. NanoVDB’s uncompressed and zip-compressed file formats are supported.</p>
-</div>
-<p>To sample the volume inside a kernel we pass a reference to it by ID, and use the built-in sampling modes:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">sample_grid</span><span class="p">(</span><span class="n">volume</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">uint64</span><span class="p">,</span>
-                <span class="n">points</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">),</span>
-                <span class="n">samples</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-
-    <span class="c1"># load sample point in world-space</span>
-    <span class="n">p</span> <span class="o">=</span> <span class="n">points</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span>
-
-    <span class="c1"># transform position to the volume&#39;s local-space</span>
-    <span class="n">q</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">volume_world_to_index</span><span class="p">(</span><span class="n">volume</span><span class="p">,</span> <span class="n">p</span><span class="p">)</span>
-
-    <span class="c1"># sample volume with trilinear interpolation</span>
-    <span class="n">f</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">volume_sample_f</span><span class="p">(</span><span class="n">volume</span><span class="p">,</span> <span class="n">q</span><span class="p">,</span> <span class="n">wp</span><span class="o">.</span><span class="n">Volume</span><span class="o">.</span><span class="n">LINEAR</span><span class="p">)</span>
-
-    <span class="c1"># write result</span>
-    <span class="n">samples</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">f</span>
-</pre></div>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Volume">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Volume</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">data</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume" title="Link to this definition">#</a></dt>
-<dd><p>Class representing a sparse grid.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>data</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of bytes representing the volume in NanoVDB format</p>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.Volume.CLOSEST">
-<span class="sig-name descname"><span class="pre">CLOSEST</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">0</span></em><a class="headerlink" href="#warp.Volume.CLOSEST" title="Link to this definition">#</a></dt>
-<dd><p>Enum value to specify nearest-neighbor interpolation during sampling</p>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.Volume.LINEAR">
-<span class="sig-name descname"><span class="pre">LINEAR</span></span><em class="property"><span class="w"> </span><span class="p"><span class="pre">=</span></span><span class="w"> </span><span class="pre">1</span></em><a class="headerlink" href="#warp.Volume.LINEAR" title="Link to this definition">#</a></dt>
-<dd><p>Enum value to specify trilinear interpolation during sampling</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.array">
-<span class="sig-name descname"><span class="pre">array</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.array" title="Link to this definition">#</a></dt>
-<dd><p>Returns the raw memory buffer of the Volume as an array</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.get_tiles">
-<span class="sig-name descname"><span class="pre">get_tiles</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.get_tiles" title="Link to this definition">#</a></dt>
-<dd><dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.array" title="warp.types.array"><em>array</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.get_voxel_size">
-<span class="sig-name descname"><span class="pre">get_voxel_size</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.get_voxel_size" title="Link to this definition">#</a></dt>
-<dd><dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a>[<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a>, <a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a>, <a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.load_from_nvdb">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_from_nvdb</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">file_or_buffer</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.load_from_nvdb" title="Link to this definition">#</a></dt>
-<dd><p>Creates a Volume object from a NanoVDB file or in-memory buffer.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Returns<span class="colon">:</span></dt>
-<dd class="field-odd"><p>A <code class="docutils literal notranslate"><span class="pre">warp.Volume</span></code> object.</p>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.Volume" title="warp.types.Volume"><em>Volume</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.load_from_numpy">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">load_from_numpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ndarray</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_world</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">voxel_size</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bg_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.load_from_numpy" title="Link to this definition">#</a></dt>
-<dd><p>Creates a Volume object from a dense 3D NumPy array.</p>
-<p>This function is only supported for CUDA devices.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>min_world</strong> – The 3D coordinate of the lower corner of the volume.</p></li>
-<li><p><strong>voxel_size</strong> – The size of each voxel in spatial coordinates.</p></li>
-<li><p><strong>bg_value</strong> – Background value</p></li>
-<li><p><strong>device</strong> – The CUDA device to create the volume on, e.g.: “cuda” or “cuda:0”.</p></li>
-<li><p><strong>ndarray</strong> (<em>array</em>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A <code class="docutils literal notranslate"><span class="pre">warp.Volume</span></code> object.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.Volume" title="warp.types.Volume"><em>Volume</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.allocate">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">allocate</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">min</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">voxel_size</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bg_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">points_in_world_space</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.allocate" title="Link to this definition">#</a></dt>
-<dd><p>Allocate a new Volume based on the bounding box defined by min and max.</p>
-<p>This function is only supported for CUDA devices.</p>
-<p>Allocate a volume that is large enough to contain voxels [min[0], min[1], min[2]] - [max[0], max[1], max[2]], inclusive.
-If points_in_world_space is true, then min and max are first converted to index space with the given voxel size and
-translation, and the volume is allocated with those.</p>
-<p>The smallest unit of allocation is a dense tile of 8x8x8 voxels, the requested bounding box is rounded up to tiles, and
-the resulting tiles will be available in the new volume.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>min</strong> (<em>array-like</em>) – Lower 3D coordinates of the bounding box in index space or world space, inclusive.</p></li>
-<li><p><strong>max</strong> (<em>array-like</em>) – Upper 3D coordinates of the bounding box in index space or world space, inclusive.</p></li>
-<li><p><strong>voxel_size</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Voxel size of the new volume.</p></li>
-<li><p><strong>bg_value</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em> or </em><em>array-like</em>) – Value of unallocated voxels of the volume, also defines the volume’s type, a <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code> volume is created if this is <cite>array-like</cite>, otherwise a float volume is created</p></li>
-<li><p><strong>translation</strong> (<em>array-like</em>) – translation between the index and world spaces.</p></li>
-<li><p><strong>device</strong> (<em>Devicelike</em>) – The CUDA device to create the volume on, e.g.: “cuda” or “cuda:0”.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.Volume" title="warp.types.Volume"><em>Volume</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Volume.allocate_by_tiles">
-<em class="property"><span class="pre">classmethod</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">allocate_by_tiles</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tile_points</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">voxel_size</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">bg_value</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">translation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Volume.allocate_by_tiles" title="Link to this definition">#</a></dt>
-<dd><p>Allocate a new Volume with active tiles for each point tile_points.</p>
-<p>This function is only supported for CUDA devices.</p>
-<p>The smallest unit of allocation is a dense tile of 8x8x8 voxels.
-This is the primary method for allocating sparse volumes. It uses an array of points indicating the tiles that must be allocated.</p>
-<dl class="simple">
-<dt>Example use cases:</dt><dd><ul class="simple">
-<li><p><cite>tile_points</cite> can mark tiles directly in index space as in the case this method is called by <cite>allocate</cite>.</p></li>
-<li><p><cite>tile_points</cite> can be a list of points used in a simulation that needs to transfer data to a volume.</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>tile_points</strong> (<a class="reference internal" href="#warp.array" title="warp.array"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.array</span></code></a>) – Array of positions that define the tiles to be allocated.
-The array can be a 2D, N-by-3 array of <a class="reference internal" href="functions.html#warp.int32" title="warp.int32"><code class="xref py py-class docutils literal notranslate"><span class="pre">warp.int32</span></code></a> values, indicating index space positions,
-or can be a 1D array of <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code> values, indicating world space positions.
-Repeated points per tile are allowed and will be efficiently deduplicated.</p></li>
-<li><p><strong>voxel_size</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Voxel size of the new volume.</p></li>
-<li><p><strong>bg_value</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em> or </em><em>array-like</em>) – Value of unallocated voxels of the volume, also defines the volume’s type, a <code class="xref py py-class docutils literal notranslate"><span class="pre">warp.vec3</span></code> volume is created if this is <cite>array-like</cite>, otherwise a float volume is created</p></li>
-<li><p><strong>translation</strong> (<em>array-like</em>) – Translation between the index and world spaces.</p></li>
-<li><p><strong>device</strong> (<em>Devicelike</em>) – The CUDA device to create the volume on, e.g.: “cuda” or “cuda:0”.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.Volume" title="warp.types.Volume"><em>Volume</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<div class="admonition seealso">
-<p class="admonition-title">See also</p>
-<p><a class="reference external" href="functions.html#volumes">Reference</a> for the volume functions available in kernels.</p>
-</div>
-</section>
-<section id="differentiability">
-<h2>Differentiability<a class="headerlink" href="#differentiability" title="Link to this heading">#</a></h2>
-<p>By default, Warp generates a forward and backward (adjoint) version of each kernel definition.
-Buffers that participate in the chain of computation should be created with <code class="docutils literal notranslate"><span class="pre">requires_grad=True</span></code>, for example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">a</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">1024</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">vec3</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>The <code class="docutils literal notranslate"><span class="pre">wp.Tape</span></code> class can then be used to record kernel launches, and replay them to compute the gradient of a scalar loss function with respect to the kernel inputs:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-
-<span class="c1"># forward pass</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">compute1</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">compute2</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">c</span><span class="p">,</span> <span class="n">d</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">loss</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">d</span><span class="p">,</span> <span class="n">l</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="c1"># reverse pass</span>
-<span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">l</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>After the backward pass has completed, the gradients with respect to the inputs are available via a mapping in the <a class="reference internal" href="#warp.Tape" title="warp.Tape"><code class="xref py py-class docutils literal notranslate"><span class="pre">wp.Tape</span></code></a> object:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># gradient of loss with respect to input a</span>
-<span class="nb">print</span><span class="p">(</span><span class="n">tape</span><span class="o">.</span><span class="n">gradients</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-</pre></div>
-</div>
-<p>Note that gradients are accumulated on the participating buffers, so if you wish to reuse the same buffers for multiple backward passes you should first zero the gradients:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">tape</span><span class="o">.</span><span class="n">zero</span><span class="p">()</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Warp uses a source-code transformation approach to auto-differentiation.
-In this approach, the backwards pass must keep a record of intermediate values computed during the forward pass.
-This imposes some restrictions on what kernels can do and still be differentiable:</p>
-<ul class="simple">
-<li><p>Dynamic loops should not mutate any previously declared local variable. This means the loop must be side-effect free.
-A simple way to ensure this is to move the loop body into a separate function.
-Static loops that are unrolled at compile time do not have this restriction and can perform any computation.</p></li>
-<li><p>Kernels should not overwrite any previously used array values except to perform simple linear add/subtract operations (e.g. via <a class="reference internal" href="functions.html#warp.atomic_add" title="warp.atomic_add"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.atomic_add()</span></code></a>)</p></li>
-</ul>
-</div>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.Tape">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">Tape</span></span><a class="headerlink" href="#warp.Tape" title="Link to this definition">#</a></dt>
-<dd><p>Record kernel launches within a Tape scope to enable automatic differentiation.
-Gradients can be computed after the operations have been recorded on the tape via
-<code class="docutils literal notranslate"><span class="pre">tape.backward()</span></code>.</p>
-<p class="rubric">Example</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-
-<span class="c1"># forward pass</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">compute1</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">,</span> <span class="n">b</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">compute2</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">c</span><span class="p">,</span> <span class="n">d</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">loss</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">d</span><span class="p">,</span> <span class="n">l</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="c1"># reverse pass</span>
-<span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">l</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Gradients can be accessed via the <code class="docutils literal notranslate"><span class="pre">tape.gradients</span></code> dictionary, e.g.:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nb">print</span><span class="p">(</span><span class="n">tape</span><span class="o">.</span><span class="n">gradients</span><span class="p">[</span><span class="n">a</span><span class="p">])</span>
-</pre></div>
-</div>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Tape.backward">
-<span class="sig-name descname"><span class="pre">backward</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">loss</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">grads</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Tape.backward" title="Link to this definition">#</a></dt>
-<dd><p>Evaluate the backward pass of the recorded operations on the tape.
-A single-element array <code class="docutils literal notranslate"><span class="pre">loss</span></code> or a dictionary of arrays <code class="docutils literal notranslate"><span class="pre">grads</span></code>
-can be provided to assign the incoming gradients for the reverse-mode
-automatic differentiation pass.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>loss</strong> (<em>wp.array</em>) – A single-element array that holds the loss function value whose gradient is to be computed</p></li>
-<li><p><strong>grads</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.12)"><em>dict</em></a>) – A dictionary of arrays that map from Warp arrays to their incoming gradients</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Tape.record_func">
-<span class="sig-name descname"><span class="pre">record_func</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">backward</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">arrays</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.Tape.record_func" title="Link to this definition">#</a></dt>
-<dd><p>Records a custom function to be executed only in the backward pass.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>backward</strong> (<em>Callable</em>) – A callable Python object (can be any function) that will be executed in the backward pass.</p></li>
-<li><p><strong>arrays</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)"><em>list</em></a>) – A list of arrays that are used by the function for gradient tracking.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Tape.reset">
-<span class="sig-name descname"><span class="pre">reset</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Tape.reset" title="Link to this definition">#</a></dt>
-<dd><p>Clear all operations recorded on the tape and zero out all gradients.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.Tape.zero">
-<span class="sig-name descname"><span class="pre">zero</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.Tape.zero" title="Link to this definition">#</a></dt>
-<dd><p>Zero out all gradients recorded on the tape.</p>
-</dd></dl>
-
-</dd></dl>
-
-<section id="jacobians">
-<h3>Jacobians<a class="headerlink" href="#jacobians" title="Link to this heading">#</a></h3>
-<p>To compute the Jacobian matrix <span class="math notranslate nohighlight">\(J\in\mathbb{R}^{m\times n}\)</span> of a multi-valued function <span class="math notranslate nohighlight">\(f: \mathbb{R}^n \to \mathbb{R}^m\)</span>, we can evaluate an entire row of the Jacobian in parallel by finding the Jacobian-vector product <span class="math notranslate nohighlight">\(J^\top \mathbf{e}\)</span>. The vector <span class="math notranslate nohighlight">\(\mathbf{e}\in\mathbb{R}^m\)</span> selects the indices in the output buffer to differentiate with respect to.
-In Warp, instead of passing a scalar loss buffer to the <code class="docutils literal notranslate"><span class="pre">tape.backward()</span></code> method, we pass a dictionary <code class="docutils literal notranslate"><span class="pre">grads</span></code> mapping from the function output array to the selection vector <span class="math notranslate nohighlight">\(\mathbf{e}\)</span> having the same type:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># compute the Jacobian for a function of single output</span>
-<span class="n">jacobian</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">((</span><span class="n">output_dim</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">output_buffer</span> <span class="o">=</span> <span class="n">launch_kernels_to_be_differentiated</span><span class="p">(</span><span class="n">input_buffer</span><span class="p">)</span>
-<span class="k">for</span> <span class="n">output_index</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">output_dim</span><span class="p">):</span>
-    <span class="c1"># select which row of the Jacobian we want to compute</span>
-    <span class="n">select_index</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">output_dim</span><span class="p">)</span>
-    <span class="n">select_index</span><span class="p">[</span><span class="n">output_index</span><span class="p">]</span> <span class="o">=</span> <span class="mf">1.0</span>
-    <span class="n">e</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">select_index</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-    <span class="c1"># pass input gradients to the output buffer to apply selection</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">grads</span><span class="o">=</span><span class="p">{</span><span class="n">output_buffer</span><span class="p">:</span> <span class="n">e</span><span class="p">})</span>
-    <span class="n">q_grad_i</span> <span class="o">=</span> <span class="n">tape</span><span class="o">.</span><span class="n">gradients</span><span class="p">[</span><span class="n">input_buffer</span><span class="p">]</span>
-    <span class="n">jacobian</span><span class="p">[</span><span class="n">output_index</span><span class="p">,</span> <span class="p">:]</span> <span class="o">=</span> <span class="n">q_grad_i</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">zero</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>When we run simulations independently in parallel, the Jacobian corresponding to the entire system dynamics is a block-diagonal matrix. In this case, we can compute the Jacobian in parallel for all environments by choosing a selection vector that has the output indices active for all environment copies. For example, to get the first rows of the Jacobians of all environments, <span class="math notranslate nohighlight">\(\mathbf{e}=[\begin{smallmatrix}1 &amp; 0 &amp; 0 &amp; \dots &amp; 1 &amp; 0 &amp; 0 &amp; \dots\end{smallmatrix}]^\top\)</span>, to compute the second rows, <span class="math notranslate nohighlight">\(\mathbf{e}=[\begin{smallmatrix}0 &amp; 1 &amp; 0 &amp; \dots &amp; 0 &amp; 1 &amp; 0 &amp; \dots\end{smallmatrix}]^\top\)</span>, etc.:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="c1"># compute the Jacobian for a function over multiple environments in parallel</span>
-<span class="n">jacobians</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">((</span><span class="n">num_envs</span><span class="p">,</span> <span class="n">output_dim</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">output_buffer</span> <span class="o">=</span> <span class="n">launch_kernels_to_be_differentiated</span><span class="p">(</span><span class="n">input_buffer</span><span class="p">)</span>
-<span class="k">for</span> <span class="n">output_index</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">output_dim</span><span class="p">):</span>
-    <span class="c1"># select which row of the Jacobian we want to compute</span>
-    <span class="n">select_index</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">output_dim</span><span class="p">)</span>
-    <span class="n">select_index</span><span class="p">[</span><span class="n">output_index</span><span class="p">]</span> <span class="o">=</span> <span class="mf">1.0</span>
-    <span class="c1"># assemble selection vector for all environments (can be precomputed)</span>
-    <span class="n">e</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">tile</span><span class="p">(</span><span class="n">select_index</span><span class="p">,</span> <span class="n">num_envs</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">grads</span><span class="o">=</span><span class="p">{</span><span class="n">output_buffer</span><span class="p">:</span> <span class="n">e</span><span class="p">})</span>
-    <span class="n">q_grad_i</span> <span class="o">=</span> <span class="n">tape</span><span class="o">.</span><span class="n">gradients</span><span class="p">[</span><span class="n">input_buffer</span><span class="p">]</span>
-    <span class="n">jacobians</span><span class="p">[:,</span> <span class="n">output_index</span><span class="p">,</span> <span class="p">:]</span> <span class="o">=</span> <span class="n">q_grad_i</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">num_envs</span><span class="p">,</span> <span class="n">input_dim</span><span class="p">)</span>
-    <span class="n">tape</span><span class="o">.</span><span class="n">zero</span><span class="p">()</span>
-</pre></div>
-</div>
-</section>
-<section id="custom-gradient-functions">
-<h3>Custom Gradient Functions<a class="headerlink" href="#custom-gradient-functions" title="Link to this heading">#</a></h3>
-<p>Warp supports custom gradient function definitions for user-defined Warp functions.
-This allows users to define code that should replace the automatically generated derivatives.</p>
-<p>To differentiate a function <span class="math notranslate nohighlight">\(h(x) = f(g(x))\)</span> that has a nested call to function <span class="math notranslate nohighlight">\(g(x)\)</span>, the chain rule is evaluated in the automatic differentiation of <span class="math notranslate nohighlight">\(h(x)\)</span>:</p>
-<div class="math-wrapper docutils container">
-<div class="math notranslate nohighlight">
-\[h^\prime(x) = f^\prime({\color{green}{\underset{\textrm{replay}}{g(x)}}}) {\color{blue}{\underset{\textrm{grad}}{g^\prime(x)}}}\]</div>
-</div>
-<p>This implies that a function to be compatible with the autodiff engine needs to provide an implementation of its forward version
-<span class="math notranslate nohighlight">\(\color{green}{g(x)}\)</span>, which we refer to as “replay” function (that matches the original function definition by default),
-and its derivative <span class="math notranslate nohighlight">\(\color{blue}{g^\prime(x)}\)</span>, referred to as “grad”.</p>
-<p>Both the replay and the grad implementations can be customized by the user. They are defined as follows:</p>
-<div class="table-wrapper colwidths-given docutils container" id="id1">
-<table class="docutils align-default" id="id1">
-<caption><span class="caption-text">Customizing the replay and grad versions of function <code class="docutils literal notranslate"><span class="pre">myfunc</span></code></span><a class="headerlink" href="#id1" title="Link to this table">#</a></caption>
-<colgroup>
-<col style="width: 100.0%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p>Forward Function</p></td>
-</tr>
-<tr class="row-even"><td><div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func</span>
-<span class="k">def</span> <span class="nf">myfunc</span><span class="p">(</span><span class="n">in1</span><span class="p">:</span> <span class="n">InType1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">inN</span><span class="p">:</span> <span class="n">InTypeN</span><span class="p">)</span> <span class="o">-&gt;</span> <span class="n">OutType1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">OutTypeM</span><span class="p">:</span>
-    <span class="k">return</span> <span class="n">out1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">outM</span>
-</pre></div>
-</div>
-</td>
-</tr>
-<tr class="row-odd"><td><p>Custom Replay Function</p></td>
-</tr>
-<tr class="row-even"><td><div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func_replay</span><span class="p">(</span><span class="n">myfunc</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">replay_myfunc</span><span class="p">(</span><span class="n">in1</span><span class="p">:</span> <span class="n">InType1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">inN</span><span class="p">:</span> <span class="n">InTypeN</span><span class="p">)</span> <span class="o">-&gt;</span> <span class="n">OutType1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">OutTypeM</span><span class="p">:</span>
-    <span class="c1"># Custom forward computations to be executed in the backward pass of a</span>
-    <span class="c1"># function calling `myfunc` go here</span>
-    <span class="c1"># Ensure the output variables match the original forward definition</span>
-    <span class="k">return</span> <span class="n">out1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">outM</span>
-</pre></div>
-</div>
-</td>
-</tr>
-<tr class="row-odd"><td><p>Custom Grad Function</p></td>
-</tr>
-<tr class="row-even"><td><div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func_grad</span><span class="p">(</span><span class="n">myfunc</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">adj_myfunc</span><span class="p">(</span><span class="n">in1</span><span class="p">:</span> <span class="n">InType1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">inN</span><span class="p">:</span> <span class="n">InTypeN</span><span class="p">,</span> <span class="n">adj_out1</span><span class="p">:</span> <span class="n">OutType1</span><span class="p">,</span> <span class="o">...</span><span class="p">,</span> <span class="n">adj_outM</span><span class="p">:</span> <span class="n">OutTypeM</span><span class="p">):</span>
-    <span class="c1"># Custom adjoint code goes here</span>
-    <span class="c1"># Update the partial derivatives for the inputs as follows:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">adjoint</span><span class="p">[</span><span class="n">in1</span><span class="p">]</span> <span class="o">+=</span> <span class="o">...</span>
-    <span class="o">...</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">adjoint</span><span class="p">[</span><span class="n">inN</span><span class="p">]</span> <span class="o">+=</span> <span class="o">...</span>
-</pre></div>
-</div>
-</td>
-</tr>
-</tbody>
-</table>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>It is currently not possible to define custom replay or grad functions for functions that
-have generic arguments, e.g. <code class="docutils literal notranslate"><span class="pre">Any</span></code> or <code class="docutils literal notranslate"><span class="pre">wp.array(dtype=Any)</span></code>. Replay or grad functions that
-themselves use generic arguments are also not yet supported.</p>
-</div>
-<section id="example-1-custom-grad-function">
-<h4>Example 1: Custom Grad Function<a class="headerlink" href="#example-1-custom-grad-function" title="Link to this heading">#</a></h4>
-<p>In the following, we define a Warp function <code class="docutils literal notranslate"><span class="pre">safe_sqrt</span></code> that computes the square root of a number:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func</span>
-<span class="k">def</span> <span class="nf">safe_sqrt</span><span class="p">(</span><span class="n">x</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
-    <span class="k">return</span> <span class="n">wp</span><span class="o">.</span><span class="n">sqrt</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>To evaluate this function, we define a kernel that applies <code class="docutils literal notranslate"><span class="pre">safe_sqrt</span></code> to an array of input values:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">run_safe_sqrt</span><span class="p">(</span><span class="n">xs</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">),</span> <span class="n">output</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">i</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">output</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">=</span> <span class="n">safe_sqrt</span><span class="p">(</span><span class="n">xs</span><span class="p">[</span><span class="n">i</span><span class="p">])</span>
-</pre></div>
-</div>
-<p>Calling the kernel for an array of values <code class="docutils literal notranslate"><span class="pre">[1.0,</span> <span class="pre">2.0,</span> <span class="pre">0.0]</span></code> yields the expected outputs, the gradients are finite except for the zero input:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">xs</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">([</span><span class="mf">1.0</span><span class="p">,</span> <span class="mf">2.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">],</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="n">ys</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros_like</span><span class="p">(</span><span class="n">xs</span><span class="p">)</span>
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">run_safe_sqrt</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="nb">len</span><span class="p">(</span><span class="n">xs</span><span class="p">),</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">xs</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">ys</span><span class="p">])</span>
-<span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">grads</span><span class="o">=</span><span class="p">{</span><span class="n">ys</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">xs</span><span class="p">)),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">)})</span>
-<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;ys     &quot;</span><span class="p">,</span> <span class="n">ys</span><span class="p">)</span>
-<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;xs.grad&quot;</span><span class="p">,</span> <span class="n">xs</span><span class="o">.</span><span class="n">grad</span><span class="p">)</span>
-
-<span class="c1"># ys      [1.   1.4142135   0. ]</span>
-<span class="c1"># xs.grad [0.5  0.35355338  inf]</span>
-</pre></div>
-</div>
-<p>It is often desired to catch nonfinite gradients in the computation graph as they may cause the entire gradient computation to be nonfinite.
-To do so, we can define a custom gradient function that replaces the adjoint function for <code class="docutils literal notranslate"><span class="pre">safe_sqrt</span></code> which is automatically generated by
-decorating the custom gradient code via <code class="docutils literal notranslate"><span class="pre">&#64;wp.func_grad(safe_sqrt)</span></code>:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func_grad</span><span class="p">(</span><span class="n">safe_sqrt</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">adj_safe_sqrt</span><span class="p">(</span><span class="n">x</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">adj_ret</span><span class="p">:</span> <span class="nb">float</span><span class="p">):</span>
-    <span class="k">if</span> <span class="n">x</span> <span class="o">&gt;</span> <span class="mf">0.0</span><span class="p">:</span>
-        <span class="n">wp</span><span class="o">.</span><span class="n">adjoint</span><span class="p">[</span><span class="n">x</span><span class="p">]</span> <span class="o">+=</span> <span class="mf">1.0</span> <span class="o">/</span> <span class="p">(</span><span class="mf">2.0</span> <span class="o">*</span> <span class="n">wp</span><span class="o">.</span><span class="n">sqrt</span><span class="p">(</span><span class="n">x</span><span class="p">))</span> <span class="o">*</span> <span class="n">adj_ret</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The function signature of the custom grad code consists of the input arguments of the forward function plus the adjoint variables of the
-forward function outputs. To access and modify the partial derivatives of the input arguments, we use the <code class="docutils literal notranslate"><span class="pre">wp.adjoint</span></code> dictionary.
-The keys of this dictionary are the input arguments of the forward function, and the values are the partial derivatives of the forward function
-output with respect to the input argument.</p>
-</div>
-</section>
-<section id="example-2-custom-replay-function">
-<h4>Example 2: Custom Replay Function<a class="headerlink" href="#example-2-custom-replay-function" title="Link to this heading">#</a></h4>
-<p>In the following, we increment an array index in each thread via <a class="reference internal" href="functions.html#warp.atomic_add" title="warp.atomic_add"><code class="xref py py-func docutils literal notranslate"><span class="pre">wp.atomic_add()</span></code></a> and compute
-the square root of an input array at the incremented index:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">test_add</span><span class="p">(</span><span class="n">counter</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span> <span class="nb">input</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">),</span> <span class="n">output</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)):</span>
-    <span class="n">idx</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">atomic_add</span><span class="p">(</span><span class="n">counter</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
-    <span class="n">output</span><span class="p">[</span><span class="n">idx</span><span class="p">]</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">sqrt</span><span class="p">(</span><span class="nb">input</span><span class="p">[</span><span class="n">idx</span><span class="p">])</span>
-
-<span class="k">def</span> <span class="nf">main</span><span class="p">():</span>
-    <span class="n">dim</span> <span class="o">=</span> <span class="mi">16</span>
-    <span class="n">use_reversible_increment</span> <span class="o">=</span> <span class="kc">False</span>
-    <span class="nb">input</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">dim</span> <span class="o">+</span> <span class="mi">1</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-    <span class="n">counter</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">int32</span><span class="p">)</span>
-    <span class="n">thread_ids</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">dim</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">int32</span><span class="p">)</span>
-    <span class="n">output</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">dim</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-    <span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-    <span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-        <span class="k">if</span> <span class="n">use_reversible_increment</span><span class="p">:</span>
-            <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">test_add_diff</span><span class="p">,</span> <span class="n">dim</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">counter</span><span class="p">,</span> <span class="n">thread_ids</span><span class="p">,</span> <span class="nb">input</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">output</span><span class="p">])</span>
-        <span class="k">else</span><span class="p">:</span>
-            <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">test_add</span><span class="p">,</span> <span class="n">dim</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">counter</span><span class="p">,</span> <span class="nb">input</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">output</span><span class="p">])</span>
-
-    <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;counter:    &quot;</span><span class="p">,</span> <span class="n">counter</span><span class="o">.</span><span class="n">numpy</span><span class="p">())</span>
-    <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;thread_ids: &quot;</span><span class="p">,</span> <span class="n">thread_ids</span><span class="o">.</span><span class="n">numpy</span><span class="p">())</span>
-    <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;input:      &quot;</span><span class="p">,</span> <span class="nb">input</span><span class="o">.</span><span class="n">numpy</span><span class="p">())</span>
-    <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;output:     &quot;</span><span class="p">,</span> <span class="n">output</span><span class="o">.</span><span class="n">numpy</span><span class="p">())</span>
-
-    <span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">grads</span><span class="o">=</span><span class="p">{</span>
-        <span class="n">output</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="n">dim</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-    <span class="p">})</span>
-    <span class="nb">print</span><span class="p">(</span><span class="s2">&quot;input.grad: &quot;</span><span class="p">,</span> <span class="nb">input</span><span class="o">.</span><span class="n">grad</span><span class="o">.</span><span class="n">numpy</span><span class="p">())</span>
-
-<span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s2">&quot;__main__&quot;</span><span class="p">:</span>
-    <span class="n">main</span><span class="p">()</span>
-</pre></div>
-</div>
-<p>The output of the above code is:</p>
-<div class="highlight-js notranslate"><div class="highlight"><pre><span></span><span class="nx">counter</span><span class="o">:</span><span class="w">     </span><span class="p">[</span><span class="mf">8</span><span class="p">]</span>
-<span class="nx">thread_ids</span><span class="o">:</span><span class="w">  </span><span class="p">[</span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="w"> </span><span class="mf">0</span><span class="p">]</span>
-<span class="nx">input</span><span class="o">:</span><span class="w">       </span><span class="p">[</span><span class="mf">1.</span><span class="w"> </span><span class="mf">2.</span><span class="w"> </span><span class="mf">3.</span><span class="w"> </span><span class="mf">4.</span><span class="w"> </span><span class="mf">5.</span><span class="w"> </span><span class="mf">6.</span><span class="w"> </span><span class="mf">7.</span><span class="w"> </span><span class="mf">8.</span><span class="p">]</span>
-<span class="nx">output</span><span class="o">:</span><span class="w">      </span><span class="p">[</span><span class="mf">1.</span><span class="w">  </span><span class="mf">1.4142135</span><span class="w">  </span><span class="mf">1.7320508</span><span class="w">  </span><span class="mf">2.</span><span class="w">  </span><span class="mf">2.236068</span><span class="w">  </span><span class="mf">2.4494898</span><span class="w">  </span><span class="mf">2.6457512</span><span class="w">  </span><span class="mf">2.828427</span><span class="p">]</span>
-<span class="nx">input</span><span class="p">.</span><span class="nx">grad</span><span class="o">:</span><span class="w">  </span><span class="p">[</span><span class="mf">4.</span><span class="w"> </span><span class="mf">0.</span><span class="w"> </span><span class="mf">0.</span><span class="w"> </span><span class="mf">0.</span><span class="w"> </span><span class="mf">0.</span><span class="w"> </span><span class="mf">0.</span><span class="w"> </span><span class="mf">0.</span><span class="w"> </span><span class="mf">0.</span><span class="p">]</span>
-</pre></div>
-</div>
-<p>The gradient of the input is incorrect because the backward pass involving the atomic operation <code class="docutils literal notranslate"><span class="pre">wp.atomic_add()</span></code> does not know which thread ID corresponds
-to which input value.
-The index returned by the adjoint of <code class="docutils literal notranslate"><span class="pre">wp.atomic_add()</span></code> is always zero so that the gradient the first entry of the input array,
-i.e. <span class="math notranslate nohighlight">\(\frac{1}{2\sqrt{1}} = 0.5\)</span>, is accumulated <code class="docutils literal notranslate"><span class="pre">dim</span></code> times (hence <code class="docutils literal notranslate"><span class="pre">input.grad[0]</span> <span class="pre">==</span> <span class="pre">4.0</span></code> and all other entries zero).</p>
-<p>To fix this, we define a new Warp function <code class="docutils literal notranslate"><span class="pre">reversible_increment()</span></code> with a custom <em>replay</em> definition that stores the thread ID in a separate array:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">func</span>
-<span class="k">def</span> <span class="nf">reversible_increment</span><span class="p">(</span>
-    <span class="n">buf</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span>
-    <span class="n">buf_index</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
-    <span class="n">value</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
-    <span class="n">thread_values</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span>
-    <span class="n">tid</span><span class="p">:</span> <span class="nb">int</span>
-<span class="p">):</span>
-    <span class="n">next_index</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">atomic_add</span><span class="p">(</span><span class="n">buf</span><span class="p">,</span> <span class="n">buf_index</span><span class="p">,</span> <span class="n">value</span><span class="p">)</span>
-    <span class="c1"># store which thread ID corresponds to which index for the backward pass</span>
-    <span class="n">thread_values</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span> <span class="o">=</span> <span class="n">next_index</span>
-    <span class="k">return</span> <span class="n">next_index</span>
-
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">func_replay</span><span class="p">(</span><span class="n">reversible_increment</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">replay_reversible_increment</span><span class="p">(</span>
-    <span class="n">buf</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span>
-    <span class="n">buf_index</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
-    <span class="n">value</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
-    <span class="n">thread_values</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span>
-    <span class="n">tid</span><span class="p">:</span> <span class="nb">int</span>
-<span class="p">):</span>
-    <span class="k">return</span> <span class="n">thread_values</span><span class="p">[</span><span class="n">tid</span><span class="p">]</span>
-</pre></div>
-</div>
-<p>Instead of running <code class="docutils literal notranslate"><span class="pre">reversible_increment()</span></code>, the custom replay code in <code class="docutils literal notranslate"><span class="pre">replay_reversible_increment()</span></code> is now executed
-during forward phase in the backward pass of the function calling <code class="docutils literal notranslate"><span class="pre">reversible_increment()</span></code>.
-We first stored the array index to each thread ID in the forward pass, and now we retrieve the array index for each thread ID in the backward pass.
-That way, the backward pass can reproduce the same addition operation as in the forward pass with exactly the same operands per thread.</p>
-<div class="admonition warning">
-<p class="admonition-title">Warning</p>
-<p>The function signature of the custom replay code must match the forward function signature.</p>
-</div>
-<p>To use our function we write the following kernel:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">test_add_diff</span><span class="p">(</span>
-    <span class="n">counter</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span>
-    <span class="n">thread_ids</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span>
-    <span class="nb">input</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">),</span>
-    <span class="n">output</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">float</span><span class="p">)</span>
-<span class="p">):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">idx</span> <span class="o">=</span> <span class="n">reversible_increment</span><span class="p">(</span><span class="n">counter</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="n">thread_ids</span><span class="p">,</span> <span class="n">tid</span><span class="p">)</span>
-    <span class="n">output</span><span class="p">[</span><span class="n">idx</span><span class="p">]</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">sqrt</span><span class="p">(</span><span class="nb">input</span><span class="p">[</span><span class="n">idx</span><span class="p">])</span>
-</pre></div>
-</div>
-<p>Running the <code class="docutils literal notranslate"><span class="pre">test_add_diff</span></code> kernel via the previous <code class="docutils literal notranslate"><span class="pre">main</span></code> function with <code class="docutils literal notranslate"><span class="pre">use_reversible_increment</span> <span class="pre">=</span> <span class="pre">True</span></code>, we now compute correct gradients
-for the input array:</p>
-<div class="highlight-js notranslate"><div class="highlight"><pre><span></span><span class="nx">counter</span><span class="o">:</span><span class="w">     </span><span class="p">[</span><span class="mf">8</span><span class="p">]</span>
-<span class="nx">thread_ids</span><span class="o">:</span><span class="w">  </span><span class="p">[</span><span class="mf">0</span><span class="w"> </span><span class="mf">1</span><span class="w"> </span><span class="mf">2</span><span class="w"> </span><span class="mf">3</span><span class="w"> </span><span class="mf">4</span><span class="w"> </span><span class="mf">5</span><span class="w"> </span><span class="mf">6</span><span class="w"> </span><span class="mf">7</span><span class="p">]</span>
-<span class="nx">input</span><span class="o">:</span><span class="w">       </span><span class="p">[</span><span class="mf">1.</span><span class="w"> </span><span class="mf">2.</span><span class="w"> </span><span class="mf">3.</span><span class="w"> </span><span class="mf">4.</span><span class="w"> </span><span class="mf">5.</span><span class="w"> </span><span class="mf">6.</span><span class="w"> </span><span class="mf">7.</span><span class="w"> </span><span class="mf">8.</span><span class="p">]</span>
-<span class="nx">output</span><span class="o">:</span><span class="w">      </span><span class="p">[</span><span class="mf">1.</span><span class="w">   </span><span class="mf">1.4142135</span><span class="w">   </span><span class="mf">1.7320508</span><span class="w">   </span><span class="mf">2.</span><span class="w">    </span><span class="mf">2.236068</span><span class="w">   </span><span class="mf">2.4494898</span><span class="w">   </span><span class="mf">2.6457512</span><span class="w">   </span><span class="mf">2.828427</span><span class="w">  </span><span class="p">]</span>
-<span class="nx">input</span><span class="p">.</span><span class="nx">grad</span><span class="o">:</span><span class="w">  </span><span class="p">[</span><span class="mf">0.5</span><span class="w">  </span><span class="mf">0.35355338</span><span class="w">  </span><span class="mf">0.28867513</span><span class="w">  </span><span class="mf">0.25</span><span class="w">  </span><span class="mf">0.2236068</span><span class="w">  </span><span class="mf">0.20412414</span><span class="w">  </span><span class="mf">0.18898225</span><span class="w">  </span><span class="mf">0.17677669</span><span class="p">]</span>
-</pre></div>
-</div>
-</section>
-</section>
-</section>
-<section id="custom-native-functions">
-<h2>Custom Native Functions<a class="headerlink" href="#custom-native-functions" title="Link to this heading">#</a></h2>
-<p>Users may insert native C++/CUDA code in Warp kernels using <code class="docutils literal notranslate"><span class="pre">&#64;func_native</span></code> decorated functions.
-These accept native code as strings that get compiled after code generation, and are called within <code class="docutils literal notranslate"><span class="pre">&#64;wp.kernel</span></code> functions.
-For example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">snippet</span> <span class="o">=</span> <span class="s2">&quot;&quot;&quot;</span>
-<span class="s2">    __shared__ int s[128];</span>
-
-<span class="s2">    s[tid] = d[tid];</span>
-<span class="s2">    __syncthreads();</span>
-<span class="s2">    d[tid] = s[N - tid - 1];</span>
-<span class="s2">    &quot;&quot;&quot;</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">func_native</span><span class="p">(</span><span class="n">snippet</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">reverse</span><span class="p">(</span><span class="n">d</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span> <span class="n">N</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span> <span class="n">tid</span><span class="p">:</span> <span class="nb">int</span><span class="p">):</span>
-    <span class="o">...</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">reverse_kernel</span><span class="p">(</span><span class="n">d</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span> <span class="n">N</span><span class="p">:</span> <span class="nb">int</span><span class="p">):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">reverse</span><span class="p">(</span><span class="n">d</span><span class="p">,</span> <span class="n">N</span><span class="p">,</span> <span class="n">tid</span><span class="p">)</span>
-
-<span class="n">N</span> <span class="o">=</span> <span class="mi">128</span>
-<span class="n">x</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="nb">int</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">reverse_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">N</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">x</span><span class="p">,</span> <span class="n">N</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>Notice the use of shared memory here: the Warp library does not expose shared memory as a feature, but the CUDA compiler will
-readily accept the above snippet. This means CUDA features not exposed in Warp are still accessible in Warp scripts.
-Warp kernels meant for the CPU won’t be able to leverage CUDA features of course, but this same mechanism supports pure C++ snippets as well.</p>
-<p>Please bear in mind the following: the thread index in your snippet should be computed in a <code class="docutils literal notranslate"><span class="pre">&#64;wp.kernel</span></code> and passed to your snippet,
-as in the above example. This means your <code class="docutils literal notranslate"><span class="pre">&#64;wp.func_native</span></code> function signature should include the variables used in your snippet,
-as well as a thread index of type <code class="docutils literal notranslate"><span class="pre">int</span></code>. The function body itself should be stubbed with <code class="docutils literal notranslate"><span class="pre">...</span></code> (the snippet will be inserted during compilation).</p>
-<p>Should you wish to record your native function on the tape and then subsequently rewind the tape, you must include an adjoint snippet
-alongside your snippet as an additional input to the decorator, as in the following example:</p>
-<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">snippet</span> <span class="o">=</span> <span class="s2">&quot;&quot;&quot;</span>
-<span class="s2">out[tid] = a * x[tid] + y[tid];</span>
-<span class="s2">&quot;&quot;&quot;</span>
-<span class="n">adj_snippet</span> <span class="o">=</span> <span class="s2">&quot;&quot;&quot;</span>
-<span class="s2">adj_a = x[tid] * adj_out[tid];</span>
-<span class="s2">adj_x[tid] = a * adj_out[tid];</span>
-<span class="s2">adj_y[tid] = adj_out[tid];</span>
-<span class="s2">&quot;&quot;&quot;</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">func_native</span><span class="p">(</span><span class="n">snippet</span><span class="p">,</span> <span class="n">adj_snippet</span><span class="p">)</span>
-<span class="k">def</span> <span class="nf">saxpy</span><span class="p">(</span>
-    <span class="n">a</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span>
-    <span class="n">x</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span>
-    <span class="n">y</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span>
-    <span class="n">out</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span>
-    <span class="n">tid</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
-<span class="p">):</span>
-    <span class="o">...</span>
-
-<span class="nd">@wp</span><span class="o">.</span><span class="n">kernel</span>
-<span class="k">def</span> <span class="nf">saxpy_kernel</span><span class="p">(</span>
-    <span class="n">a</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span>
-    <span class="n">x</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span>
-    <span class="n">y</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span>
-    <span class="n">out</span><span class="p">:</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
-<span class="p">):</span>
-    <span class="n">tid</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">tid</span><span class="p">()</span>
-    <span class="n">saxpy</span><span class="p">(</span><span class="n">a</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">out</span><span class="p">,</span> <span class="n">tid</span><span class="p">)</span>
-
-<span class="n">N</span> <span class="o">=</span> <span class="mi">128</span>
-<span class="n">a</span> <span class="o">=</span> <span class="mf">2.0</span>
-<span class="n">x</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">,</span> <span class="n">requires_grad</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
-<span class="n">y</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">zeros_like</span><span class="p">(</span><span class="n">x1</span><span class="p">)</span>
-<span class="n">out</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-<span class="n">adj_out</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="n">N</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">wp</span><span class="o">.</span><span class="n">float32</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">tape</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">Tape</span><span class="p">()</span>
-
-<span class="k">with</span> <span class="n">tape</span><span class="p">:</span>
-    <span class="n">wp</span><span class="o">.</span><span class="n">launch</span><span class="p">(</span><span class="n">kernel</span><span class="o">=</span><span class="n">saxpy_kernel</span><span class="p">,</span> <span class="n">dim</span><span class="o">=</span><span class="n">N</span><span class="p">,</span> <span class="n">inputs</span><span class="o">=</span><span class="p">[</span><span class="n">a</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">],</span> <span class="n">outputs</span><span class="o">=</span><span class="p">[</span><span class="n">out</span><span class="p">],</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
-
-<span class="n">tape</span><span class="o">.</span><span class="n">backward</span><span class="p">(</span><span class="n">grads</span><span class="o">=</span><span class="p">{</span><span class="n">out</span><span class="p">:</span> <span class="n">adj_out</span><span class="p">})</span>
-</pre></div>
-</div>
-</section>
-<section id="profiling">
-<h2>Profiling<a class="headerlink" href="#profiling" title="Link to this heading">#</a></h2>
-<p><code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> objects can be used to gain some basic insight into the performance of Warp applications:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedTimer</span><span class="p">(</span><span class="s2">&quot;grid build&quot;</span><span class="p">):</span>
-    <span class="bp">self</span><span class="o">.</span><span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">point_radius</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>This results in a printout at runtime to the standard output stream like:</p>
-<div class="highlight-console notranslate"><div class="highlight"><pre><span></span><span class="go">grid build took 0.06 ms</span>
-</pre></div>
-</div>
-<p>The <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> object does not synchronize (e.g. by calling <code class="docutils literal notranslate"><span class="pre">wp.synchronize()</span></code>)
-upon exiting the <code class="docutils literal notranslate"><span class="pre">with</span></code> statement, so this can lead to misleading numbers if the body
-of the <code class="docutils literal notranslate"><span class="pre">with</span></code> statement launches device kernels.</p>
-<p>When a <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> object is passed <code class="docutils literal notranslate"><span class="pre">use_nvtx=True</span></code> as an argument, the timing functionality is replaced by calls
-to <code class="docutils literal notranslate"><span class="pre">nvtx.start_range()</span></code> and <code class="docutils literal notranslate"><span class="pre">nvtx.end_range()</span></code>:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="k">with</span> <span class="n">wp</span><span class="o">.</span><span class="n">ScopedTimer</span><span class="p">(</span><span class="s2">&quot;grid build&quot;</span><span class="p">,</span> <span class="n">use_nvtx</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s2">&quot;cyan&quot;</span><span class="p">):</span>
-    <span class="bp">self</span><span class="o">.</span><span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">point_radius</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>These range annotations can then be collected by a tool like <a class="reference external" href="https://developer.nvidia.com/nsight-systems">NVIDIA Nsight Systems</a>
-for visualization on a timeline, e.g.:</p>
-<div class="highlight-console notranslate"><div class="highlight"><pre><span></span><span class="gp">$ </span>nsys<span class="w"> </span>profile<span class="w"> </span>python<span class="w"> </span>warp_application.py
-</pre></div>
-</div>
-<p>This code snippet also demonstrates the use of the <code class="docutils literal notranslate"><span class="pre">color</span></code> argument to specify a color
-for the range, which may be a number representing the ARGB value or a recognized string
-(refer to <a class="reference external" href="https://github.com/NVIDIA/NVTX/blob/release-v3/python/nvtx/colors.py">colors.py</a> for
-additional color examples).
-The <a class="reference external" href="https://github.com/NVIDIA/NVTX">nvtx module</a> must be
-installed in the Python environment for this capability to work.
-An equivalent way to create an NVTX range without using <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code> is:</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">nvtx</span>
-
-<span class="k">with</span> <span class="n">nvtx</span><span class="o">.</span><span class="n">annotate</span><span class="p">(</span><span class="s2">&quot;grid build&quot;</span><span class="p">,</span> <span class="n">color</span><span class="o">=</span><span class="s2">&quot;cyan&quot;</span><span class="p">):</span>
-    <span class="bp">self</span><span class="o">.</span><span class="n">grid</span><span class="o">.</span><span class="n">build</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">x</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">point_radius</span><span class="p">)</span>
-</pre></div>
-</div>
-<p>This form may be more convenient if the user does not need to frequently switch
-between timer and NVTX capabilities of <code class="docutils literal notranslate"><span class="pre">wp.ScopedTimer</span></code>.</p>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.ScopedTimer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.</span></span><span class="sig-name descname"><span class="pre">ScopedTimer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">name</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">active</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">print</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">detailed</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dict</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">use_nvtx</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">color</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'rapids'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">synchronize</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.ScopedTimer" title="Link to this definition">#</a></dt>
-<dd><p>Context manager object for a timer</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – Name of timer</p></li>
-<li><p><strong>active</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Enables this timer</p></li>
-<li><p><strong>print</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – At context manager exit, print elapsed time to sys.stdout</p></li>
-<li><p><strong>detailed</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Collects additional profiling data using cProfile and calls <code class="docutils literal notranslate"><span class="pre">print_stats()</span></code> at context exit</p></li>
-<li><p><strong>dict</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.12)"><em>dict</em></a>) – A dictionary of lists to which the elapsed time will be appended using <code class="docutils literal notranslate"><span class="pre">name</span></code> as a key</p></li>
-<li><p><strong>use_nvtx</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If true, timing functionality is replaced by an NVTX range</p></li>
-<li><p><strong>color</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> or </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – ARGB value (e.g. 0x00FFFF) or color name (e.g. ‘cyan’) associated with the NVTX range</p></li>
-<li><p><strong>synchronize</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Synchronize the CPU thread with any outstanding CUDA work to return accurate GPU timings</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.ScopedTimer.elapsed">
-<span class="sig-name descname"><span class="pre">elapsed</span></span><a class="headerlink" href="#warp.ScopedTimer.elapsed" title="Link to this definition">#</a></dt>
-<dd><p>The duration of the <code class="docutils literal notranslate"><span class="pre">with</span></code> block used with this object</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="functions.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">Kernel Reference</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="../faq.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">FAQ</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">Runtime Reference</a><ul>
-<li><a class="reference internal" href="#kernels">Kernels</a><ul>
-<li><a class="reference internal" href="#warp.launch"><code class="docutils literal notranslate"><span class="pre">launch()</span></code></a></li>
-<li><a class="reference internal" href="#large-kernel-launches">Large Kernel Launches</a></li>
-<li><a class="reference internal" href="#runtime-kernel-specialization">Runtime Kernel Specialization</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#arrays">Arrays</a><ul>
-<li><a class="reference internal" href="#warp.array"><code class="docutils literal notranslate"><span class="pre">array</span></code></a><ul>
-<li><a class="reference internal" href="#warp.array.grad"><code class="docutils literal notranslate"><span class="pre">array.grad</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.requires_grad"><code class="docutils literal notranslate"><span class="pre">array.requires_grad</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.vars"><code class="docutils literal notranslate"><span class="pre">array.vars</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.zero_"><code class="docutils literal notranslate"><span class="pre">array.zero_()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.fill_"><code class="docutils literal notranslate"><span class="pre">array.fill_()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.assign"><code class="docutils literal notranslate"><span class="pre">array.assign()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.numpy"><code class="docutils literal notranslate"><span class="pre">array.numpy()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.cptr"><code class="docutils literal notranslate"><span class="pre">array.cptr()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.list"><code class="docutils literal notranslate"><span class="pre">array.list()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.to"><code class="docutils literal notranslate"><span class="pre">array.to()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.flatten"><code class="docutils literal notranslate"><span class="pre">array.flatten()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.reshape"><code class="docutils literal notranslate"><span class="pre">array.reshape()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.view"><code class="docutils literal notranslate"><span class="pre">array.view()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.contiguous"><code class="docutils literal notranslate"><span class="pre">array.contiguous()</span></code></a></li>
-<li><a class="reference internal" href="#warp.array.transpose"><code class="docutils literal notranslate"><span class="pre">array.transpose()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#multi-dimensional-arrays">Multi-dimensional Arrays</a><ul>
-<li><a class="reference internal" href="#warp.zeros"><code class="docutils literal notranslate"><span class="pre">zeros()</span></code></a></li>
-<li><a class="reference internal" href="#warp.zeros_like"><code class="docutils literal notranslate"><span class="pre">zeros_like()</span></code></a></li>
-<li><a class="reference internal" href="#warp.full"><code class="docutils literal notranslate"><span class="pre">full()</span></code></a></li>
-<li><a class="reference internal" href="#warp.full_like"><code class="docutils literal notranslate"><span class="pre">full_like()</span></code></a></li>
-<li><a class="reference internal" href="#warp.empty"><code class="docutils literal notranslate"><span class="pre">empty()</span></code></a></li>
-<li><a class="reference internal" href="#warp.empty_like"><code class="docutils literal notranslate"><span class="pre">empty_like()</span></code></a></li>
-<li><a class="reference internal" href="#warp.copy"><code class="docutils literal notranslate"><span class="pre">copy()</span></code></a></li>
-<li><a class="reference internal" href="#warp.clone"><code class="docutils literal notranslate"><span class="pre">clone()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#matrix-multiplication">Matrix Multiplication</a><ul>
-<li><a class="reference internal" href="#warp.matmul"><code class="docutils literal notranslate"><span class="pre">matmul()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#data-types">Data Types</a><ul>
-<li><a class="reference internal" href="#scalar-types">Scalar Types</a></li>
-<li><a class="reference internal" href="#vectors">Vectors</a></li>
-<li><a class="reference internal" href="#matrices">Matrices</a></li>
-<li><a class="reference internal" href="#quaternions">Quaternions</a></li>
-<li><a class="reference internal" href="#transforms">Transforms</a></li>
-<li><a class="reference internal" href="#structs">Structs</a><ul>
-<li><a class="reference internal" href="#example-using-a-custom-struct-in-gradient-computation">Example: Using a custom struct in gradient computation</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#type-conversions">Type Conversions</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#constants">Constants</a><ul>
-<li><a class="reference internal" href="#warp.constant"><code class="docutils literal notranslate"><span class="pre">constant</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#operators">Operators</a><ul>
-<li><a class="reference internal" href="#boolean-operators">Boolean Operators</a></li>
-<li><a class="reference internal" href="#comparison-operators">Comparison Operators</a></li>
-<li><a class="reference internal" href="#arithmetic-operators">Arithmetic Operators</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#graphs">Graphs</a><ul>
-<li><a class="reference internal" href="#warp.capture_begin"><code class="docutils literal notranslate"><span class="pre">capture_begin()</span></code></a></li>
-<li><a class="reference internal" href="#warp.capture_end"><code class="docutils literal notranslate"><span class="pre">capture_end()</span></code></a></li>
-<li><a class="reference internal" href="#warp.capture_launch"><code class="docutils literal notranslate"><span class="pre">capture_launch()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#bounding-value-hierarchies-bvh">Bounding Value Hierarchies (BVH)</a><ul>
-<li><a class="reference internal" href="#warp.Bvh"><code class="docutils literal notranslate"><span class="pre">Bvh</span></code></a><ul>
-<li><a class="reference internal" href="#warp.Bvh.id"><code class="docutils literal notranslate"><span class="pre">Bvh.id</span></code></a></li>
-<li><a class="reference internal" href="#warp.Bvh.device"><code class="docutils literal notranslate"><span class="pre">Bvh.device</span></code></a></li>
-<li><a class="reference internal" href="#warp.Bvh.refit"><code class="docutils literal notranslate"><span class="pre">Bvh.refit()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#example-bvh-ray-traversal">Example: BVH Ray Traversal</a></li>
-<li><a class="reference internal" href="#example-bvh-volume-traversal">Example: BVH Volume Traversal</a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#meshes">Meshes</a><ul>
-<li><a class="reference internal" href="#warp.Mesh"><code class="docutils literal notranslate"><span class="pre">Mesh</span></code></a><ul>
-<li><a class="reference internal" href="#warp.Mesh.id"><code class="docutils literal notranslate"><span class="pre">Mesh.id</span></code></a></li>
-<li><a class="reference internal" href="#warp.Mesh.device"><code class="docutils literal notranslate"><span class="pre">Mesh.device</span></code></a></li>
-<li><a class="reference internal" href="#warp.Mesh.refit"><code class="docutils literal notranslate"><span class="pre">Mesh.refit()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#hash-grids">Hash Grids</a><ul>
-<li><a class="reference internal" href="#warp.HashGrid"><code class="docutils literal notranslate"><span class="pre">HashGrid</span></code></a><ul>
-<li><a class="reference internal" href="#warp.HashGrid.id"><code class="docutils literal notranslate"><span class="pre">HashGrid.id</span></code></a></li>
-<li><a class="reference internal" href="#warp.HashGrid.device"><code class="docutils literal notranslate"><span class="pre">HashGrid.device</span></code></a></li>
-<li><a class="reference internal" href="#warp.HashGrid.build"><code class="docutils literal notranslate"><span class="pre">HashGrid.build()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#volumes">Volumes</a><ul>
-<li><a class="reference internal" href="#warp.Volume"><code class="docutils literal notranslate"><span class="pre">Volume</span></code></a><ul>
-<li><a class="reference internal" href="#warp.Volume.CLOSEST"><code class="docutils literal notranslate"><span class="pre">Volume.CLOSEST</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.LINEAR"><code class="docutils literal notranslate"><span class="pre">Volume.LINEAR</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.array"><code class="docutils literal notranslate"><span class="pre">Volume.array()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.get_tiles"><code class="docutils literal notranslate"><span class="pre">Volume.get_tiles()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.get_voxel_size"><code class="docutils literal notranslate"><span class="pre">Volume.get_voxel_size()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.load_from_nvdb"><code class="docutils literal notranslate"><span class="pre">Volume.load_from_nvdb()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.load_from_numpy"><code class="docutils literal notranslate"><span class="pre">Volume.load_from_numpy()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.allocate"><code class="docutils literal notranslate"><span class="pre">Volume.allocate()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Volume.allocate_by_tiles"><code class="docutils literal notranslate"><span class="pre">Volume.allocate_by_tiles()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#differentiability">Differentiability</a><ul>
-<li><a class="reference internal" href="#warp.Tape"><code class="docutils literal notranslate"><span class="pre">Tape</span></code></a><ul>
-<li><a class="reference internal" href="#warp.Tape.backward"><code class="docutils literal notranslate"><span class="pre">Tape.backward()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Tape.record_func"><code class="docutils literal notranslate"><span class="pre">Tape.record_func()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Tape.reset"><code class="docutils literal notranslate"><span class="pre">Tape.reset()</span></code></a></li>
-<li><a class="reference internal" href="#warp.Tape.zero"><code class="docutils literal notranslate"><span class="pre">Tape.zero()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#jacobians">Jacobians</a></li>
-<li><a class="reference internal" href="#custom-gradient-functions">Custom Gradient Functions</a><ul>
-<li><a class="reference internal" href="#example-1-custom-grad-function">Example 1: Custom Grad Function</a></li>
-<li><a class="reference internal" href="#example-2-custom-replay-function">Example 2: Custom Replay Function</a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#custom-native-functions">Custom Native Functions</a></li>
-<li><a class="reference internal" href="#profiling">Profiling</a><ul>
-<li><a class="reference internal" href="#warp.ScopedTimer"><code class="docutils literal notranslate"><span class="pre">ScopedTimer</span></code></a><ul>
-<li><a class="reference internal" href="#warp.ScopedTimer.elapsed"><code class="docutils literal notranslate"><span class="pre">ScopedTimer.elapsed</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    <script async="async" src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/sim.html b/docs/_build/html/modules/sim.html
deleted file mode 100644
index 0c788ea15..000000000
--- a/docs/_build/html/modules/sim.html
+++ /dev/null
@@ -1,3661 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="warp.sparse" href="sparse.html" /><link rel="prev" title="Kernel Reference" href="functions.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>warp.sim - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul class="current">
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="warp-sim">
-<h1>warp.sim<a class="headerlink" href="#warp-sim" title="Link to this heading">#</a></h1>
-<p>Warp includes a simulation module <code class="docutils literal notranslate"><span class="pre">warp.sim</span></code> that includes many common physical simulation models, and integrators for explicit and implicit time-stepping.</p>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The simulation model is under construction and should be expected to change rapidly, please treat this section as work in progress.</p>
-</div>
-<section id="model">
-<h2>Model<a class="headerlink" href="#model" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">ModelBuilder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">up_vector</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">1.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">gravity</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-9.80665</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder" title="Link to this definition">#</a></dt>
-<dd><p>A helper class for building simulation models at runtime.</p>
-<p>Use the ModelBuilder to construct a simulation scene. The ModelBuilder
-and builds the scene representation using standard Python data structures (lists),
-this means it is not differentiable. Once <a class="reference internal" href="#warp.sim.ModelBuilder.finalize" title="warp.sim.ModelBuilder.finalize"><code class="xref py py-func docutils literal notranslate"><span class="pre">finalize()</span></code></a>
-has been called the ModelBuilder transfers all data to Warp tensors and returns
-an object that may be used for simulation.</p>
-<p class="rubric">Example</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-<span class="kn">import</span> <span class="nn">warp.sim</span>
-
-<span class="n">builder</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">sim</span><span class="o">.</span><span class="n">ModelBuilder</span><span class="p">()</span>
-
-<span class="c1"># anchor point (zero mass)</span>
-<span class="n">builder</span><span class="o">.</span><span class="n">add_particle</span><span class="p">((</span><span class="mi">0</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">),</span> <span class="p">(</span><span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">),</span> <span class="mf">0.0</span><span class="p">)</span>
-
-<span class="c1"># build chain</span>
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span><span class="mi">10</span><span class="p">):</span>
-    <span class="n">builder</span><span class="o">.</span><span class="n">add_particle</span><span class="p">((</span><span class="n">i</span><span class="p">,</span> <span class="mf">1.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">),</span> <span class="p">(</span><span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">),</span> <span class="mf">1.0</span><span class="p">)</span>
-    <span class="n">builder</span><span class="o">.</span><span class="n">add_spring</span><span class="p">(</span><span class="n">i</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="n">i</span><span class="p">,</span> <span class="mf">1.e+3</span><span class="p">,</span> <span class="mf">0.0</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span>
-
-<span class="c1"># create model</span>
-<span class="n">model</span> <span class="o">=</span> <span class="n">builder</span><span class="o">.</span><span class="n">finalize</span><span class="p">(</span><span class="s2">&quot;cuda&quot;</span><span class="p">)</span>
-
-<span class="n">state</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">state</span><span class="p">()</span>
-<span class="n">integrator</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">sim</span><span class="o">.</span><span class="n">SemiImplicitIntegrator</span><span class="p">()</span>
-
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">100</span><span class="p">):</span>
-
-   <span class="n">state</span><span class="o">.</span><span class="n">clear_forces</span><span class="p">()</span>
-   <span class="n">integrator</span><span class="o">.</span><span class="n">simulate</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">state</span><span class="p">,</span> <span class="n">state</span><span class="p">,</span> <span class="n">dt</span><span class="o">=</span><span class="mf">1.0</span><span class="o">/</span><span class="mf">60.0</span><span class="p">)</span>
-</pre></div>
-</div>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>It is strongly recommended to use the ModelBuilder to construct a simulation rather
-than creating your own Model object directly, however it is possible to do so if
-desired.</p>
-</div>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_builder">
-<span class="sig-name descname"><span class="pre">add_builder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">articulation</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xform</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">update_num_env_count</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">separate_collision_group</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_builder" title="Link to this definition">#</a></dt>
-<dd><p>Copies a rigid articulation from <cite>articulation</cite>, another <cite>ModelBuilder</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>articulation</strong> (<a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><em>ModelBuilder</em></a>) – a model builder to add rigid articulation from.</p></li>
-<li><p><strong>xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – offset transform applied to root bodies.</p></li>
-<li><p><strong>update_num_env_count</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – if True, the number of environments is incremented by 1.</p></li>
-<li><p><strong>separate_collision_group</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – if True, the shapes from the articulation will all be put into a single new collision group, otherwise, only the shapes in collision group &gt; -1 will be moved to a new group.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_body">
-<span class="sig-name descname"><span class="pre">add_body</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">origin=&lt;warp.types.transformf</span> <span class="pre">object&gt;,</span> <span class="pre">armature=0.0,</span> <span class="pre">com=array([0.,</span> <span class="pre">0.,</span> <span class="pre">0.]),</span> <span class="pre">I_m=array([[0.,</span> <span class="pre">0.,</span> <span class="pre">0.],</span>&#160;&#160;&#160;&#160;&#160;&#160;&#160; <span class="pre">[0.,</span> <span class="pre">0.,</span> <span class="pre">0.],</span>&#160;&#160;&#160;&#160;&#160;&#160;&#160; <span class="pre">[0.,</span> <span class="pre">0.,</span> <span class="pre">0.]]),</span> <span class="pre">m=0.0,</span> <span class="pre">name=None</span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_body" title="Link to this definition">#</a></dt>
-<dd><p>Adds a rigid body to the model.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>origin</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><em>Tuple</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em><em>, </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em><em>]</em>) – The location of the body in the world frame</p></li>
-<li><p><strong>armature</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Artificial inertia added to the body</p></li>
-<li><p><strong>com</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The center of mass of the body w.r.t its origin</p></li>
-<li><p><strong>I_m</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The 3x3 inertia tensor of the body (specified relative to the center of mass)</p></li>
-<li><p><strong>m</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Mass of the body</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – Name of the body (optional)</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the body in the model</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>If the mass (m) is zero then the body is treated as kinematic with no dynamics</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint">
-<span class="sig-name descname"><span class="pre">add_joint</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">joint_type</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_axes=[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_axes=[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint" title="Link to this definition">#</a></dt>
-<dd><p>Generic method to add any type of joint to this ModelBuilder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>joint_type</strong> (<a class="reference internal" href="runtime.html#warp.constant" title="warp.types.constant"><em>constant</em></a>) – The type of joint to add (see <a class="reference internal" href="#joint-types">Joint types</a>)</p></li>
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body (-1 is the world)</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>linear_axes</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.model.JointAxis"><em>JointAxis</em></a><em>]</em>) – The linear axes (see <a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><code class="xref py py-class docutils literal notranslate"><span class="pre">JointAxis</span></code></a>) of the joint</p></li>
-<li><p><strong>angular_axes</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.model.JointAxis"><em>JointAxis</em></a><em>]</em>) – The angular axes (see <a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><code class="xref py py-class docutils literal notranslate"><span class="pre">JointAxis</span></code></a>) of the joint</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_revolute">
-<span class="sig-name descname"><span class="pre">add_joint_revolute</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mode</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_lower</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-6.283185307179586</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_upper</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">6.283185307179586</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_revolute" title="Link to this definition">#</a></dt>
-<dd><p>Adds a revolute (hinge) joint to the model. It has one degree of freedom.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>axis</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The axis of rotation in the parent body’s local frame, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>target</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The target angle (in radians) of the joint</p></li>
-<li><p><strong>target_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joint target</p></li>
-<li><p><strong>target_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joint target</p></li>
-<li><p><strong>limit_lower</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The lower limit of the joint</p></li>
-<li><p><strong>limit_upper</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The upper limit of the joint</p></li>
-<li><p><strong>limit_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joint limit</p></li>
-<li><p><strong>limit_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joint limit</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-<li><p><strong>mode</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_prismatic">
-<span class="sig-name descname"><span class="pre">add_joint_prismatic</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mode</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_lower</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-10000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_upper</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_prismatic" title="Link to this definition">#</a></dt>
-<dd><p>Adds a prismatic (sliding) joint to the model. It has one degree of freedom.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>axis</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The axis of rotation in the parent body’s local frame, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>target</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The target position of the joint</p></li>
-<li><p><strong>target_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joint target</p></li>
-<li><p><strong>target_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joint target</p></li>
-<li><p><strong>limit_lower</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The lower limit of the joint</p></li>
-<li><p><strong>limit_upper</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The upper limit of the joint</p></li>
-<li><p><strong>limit_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joint limit</p></li>
-<li><p><strong>limit_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joint limit</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-<li><p><strong>mode</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_ball">
-<span class="sig-name descname"><span class="pre">add_joint_ball</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_ball" title="Link to this definition">#</a></dt>
-<dd><p>Adds a ball (spherical) joint to the model. Its position is defined by a 4D quaternion (xyzw) and its velocity is a 3D vector.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_fixed">
-<span class="sig-name descname"><span class="pre">add_joint_fixed</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_fixed" title="Link to this definition">#</a></dt>
-<dd><p>Adds a fixed (static) joint to the model. It has no degrees of freedom.
-See <a class="reference internal" href="#warp.sim.ModelBuilder.collapse_fixed_joints" title="warp.sim.ModelBuilder.collapse_fixed_joints"><code class="xref py py-meth docutils literal notranslate"><span class="pre">collapse_fixed_joints()</span></code></a> for a helper function that removes these fixed joints and merges the connecting bodies to simplify the model and improve stability.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_free">
-<span class="sig-name descname"><span class="pre">add_joint_free</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent=-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_free" title="Link to this definition">#</a></dt>
-<dd><p>Adds a free joint to the model.
-It has 7 positional degrees of freedom (first 3 linear and then 4 angular dimensions for the orientation quaternion in <cite>xyzw</cite> notation) and 6 velocity degrees of freedom (first 3 angular and then 3 linear velocity dimensions).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body (-1 by default to use the world frame, e.g. to make the child body and its children a floating-base mechanism)</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_distance">
-<span class="sig-name descname"><span class="pre">add_joint_distance</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">min_distance=-1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">max_distance=1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_distance" title="Link to this definition">#</a></dt>
-<dd><p>Adds a distance joint to the model. The distance joint constraints the distance between the joint anchor points on the two bodies (see <a class="reference internal" href="#fk-ik"><span class="std std-ref">Forward / Inverse Kinematics</span></a>) it connects to the interval [<cite>min_distance</cite>, <cite>max_distance</cite>].
-It has 7 positional degrees of freedom (first 3 linear and then 4 angular dimensions for the orientation quaternion in <cite>xyzw</cite> notation) and 6 velocity degrees of freedom (first 3 angular and then 3 linear velocity dimensions).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>min_distance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The minimum distance between the bodies (no limit if negative)</p></li>
-<li><p><strong>max_distance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The maximum distance between the bodies (no limit if negative)</p></li>
-<li><p><strong>compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The compliance of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Distance joints are currently only supported in the <a class="reference internal" href="#warp.sim.XPBDIntegrator" title="warp.sim.XPBDIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">XPBDIntegrator</span></code></a> at the moment.</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_universal">
-<span class="sig-name descname"><span class="pre">add_joint_universal</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis_0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis_1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_universal" title="Link to this definition">#</a></dt>
-<dd><p>Adds a universal joint to the model. U-joints have two degrees of freedom, one for each axis.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>axis_0</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The first axis of the joint, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>axis_1</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The second axis of the joint, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_compound">
-<span class="sig-name descname"><span class="pre">add_joint_compound</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis_0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis_1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">axis_2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_compound" title="Link to this definition">#</a></dt>
-<dd><p>Adds a compound joint to the model, which has 3 degrees of freedom, one for each axis.
-Similar to the ball joint (see <code class="xref py py-meth docutils literal notranslate"><span class="pre">add_ball_joint()</span></code>), the compound joint allows bodies to move in a 3D rotation relative to each other,
-except that the rotation is defined by 3 axes instead of a quaternion.
-Depending on the choice of axes, the orientation can be specified through Euler angles, e.g. <cite>z-x-z</cite> or <cite>x-y-x</cite>, or through a Tait-Bryan angle sequence, e.g. <cite>z-y-x</cite> or <cite>x-y-z</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>axis_0</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The first axis of the joint, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>axis_1</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The second axis of the joint, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>axis_2</strong> (<em>3D vector</em><em> or </em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis"><em>JointAxis</em></a>) – The third axis of the joint, can be a JointAxis object whose settings will be used instead of the other arguments</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_joint_d6">
-<span class="sig-name descname"><span class="pre">add_joint_d6</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">parent</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_axes=[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_axes=[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">name=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parent_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">child_xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">linear_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_compliance=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collision_filter_parent=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enabled=True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_joint_d6" title="Link to this definition">#</a></dt>
-<dd><p>Adds a generic joint with custom linear and angular axes. The number of axes determines the number of degrees of freedom of the joint.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>parent</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body</p></li>
-<li><p><strong>child</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the child body</p></li>
-<li><p><strong>linear_axes</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.model.JointAxis"><em>JointAxis</em></a><em>]</em>) – A list of linear axes</p></li>
-<li><p><strong>angular_axes</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.model.JointAxis"><em>JointAxis</em></a><em>]</em>) – A list of angular axes</p></li>
-<li><p><strong>name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – The name of the joint</p></li>
-<li><p><strong>parent_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the parent body’s local frame</p></li>
-<li><p><strong>child_xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform of the joint in the child body’s local frame</p></li>
-<li><p><strong>linear_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The linear compliance of the joint</p></li>
-<li><p><strong>angular_compliance</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The angular compliance of the joint</p></li>
-<li><p><strong>collision_filter_parent</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether to filter collisions between shapes of the parent and child bodies</p></li>
-<li><p><strong>enabled</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the joint is enabled</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added joint</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.collapse_fixed_joints">
-<span class="sig-name descname"><span class="pre">collapse_fixed_joints</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.collapse_fixed_joints" title="Link to this definition">#</a></dt>
-<dd><p>Removes fixed joints from the model and merges the bodies they connect. This is useful for simplifying the model for faster and more stable simulation.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_muscle">
-<span class="sig-name descname"><span class="pre">add_muscle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bodies</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">positions</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">f0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">lm</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">lt</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">lmax</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pen</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_muscle" title="Link to this definition">#</a></dt>
-<dd><p>Adds a muscle-tendon activation unit.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>bodies</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – A list of body indices for each waypoint</p></li>
-<li><p><strong>positions</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em><em>]</em>) – A list of positions of each waypoint in the body’s local frame</p></li>
-<li><p><strong>f0</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Force scaling</p></li>
-<li><p><strong>lm</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Muscle length</p></li>
-<li><p><strong>lt</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Tendon length</p></li>
-<li><p><strong>lmax</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Maximally efficient muscle length</p></li>
-<li><p><strong>pen</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the muscle in the model</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The simulation support for muscles is in progress and not yet fully functional.</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_plane">
-<span class="sig-name descname"><span class="pre">add_shape_plane</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">plane</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">1.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">width</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">length</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">body</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_plane" title="Link to this definition">#</a></dt>
-<dd><p>Adds a plane collision shape.
-If pos and rot are defined, the plane is assumed to have its normal as (0, 1, 0).
-Otherwise, the plane equation defined through the <cite>plane</cite> argument is used.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>plane</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The plane equation in form a*x + b*y + c*z + d = 0</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – The position of the plane in world coordinates</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – The rotation of the plane in world coordinates</p></li>
-<li><p><strong>width</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The extent along x of the plane (infinite if 0)</p></li>
-<li><p><strong>length</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The extent along z of the plane (infinite if 0)</p></li>
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The body index to attach the shape to (-1 by default to keep the plane static)</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The thickness of the plane (0 by default) for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_sphere">
-<span class="sig-name descname"><span class="pre">add_shape_sphere</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">radius</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_sphere" title="Link to this definition">#</a></dt>
-<dd><p>Adds a sphere collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The radius of the sphere</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the sphere is solid or hollow</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow sphere, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_box">
-<span class="sig-name descname"><span class="pre">add_shape_box</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hx</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hy</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">hz</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_box" title="Link to this definition">#</a></dt>
-<dd><p>Adds a box collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>hx</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The half-extent along the x-axis</p></li>
-<li><p><strong>hy</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The half-extent along the y-axis</p></li>
-<li><p><strong>hz</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The half-extent along the z-axis</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the box is solid or hollow</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow box, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_capsule">
-<span class="sig-name descname"><span class="pre">add_shape_capsule</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">radius</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">half_height</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_axis</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_capsule" title="Link to this definition">#</a></dt>
-<dd><p>Adds a capsule collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The radius of the capsule</p></li>
-<li><p><strong>half_height</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The half length of the center cylinder along the up axis</p></li>
-<li><p><strong>up_axis</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The axis along which the capsule is aligned (0=x, 1=y, 2=z)</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the capsule is solid or hollow</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow capsule, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_cylinder">
-<span class="sig-name descname"><span class="pre">add_shape_cylinder</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">radius</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">half_height</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_axis</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_cylinder" title="Link to this definition">#</a></dt>
-<dd><p>Adds a cylinder collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The radius of the cylinder</p></li>
-<li><p><strong>half_height</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The half length of the cylinder along the up axis</p></li>
-<li><p><strong>up_axis</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The axis along which the cylinder is aligned (0=x, 1=y, 2=z)</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the cylinder is solid or hollow</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow cylinder, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_cone">
-<span class="sig-name descname"><span class="pre">add_shape_cone</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">radius</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">half_height</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_axis</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_cone" title="Link to this definition">#</a></dt>
-<dd><p>Adds a cone collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The radius of the cone</p></li>
-<li><p><strong>half_height</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The half length of the cone along the up axis</p></li>
-<li><p><strong>up_axis</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The axis along which the cone is aligned (0=x, 1=y, 2=z)</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether the cone is solid or hollow</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow cone, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_mesh">
-<span class="sig-name descname"><span class="pre">add_shape_mesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mesh</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(1.0,</span> <span class="pre">1.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">has_ground_collision</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_mesh" title="Link to this definition">#</a></dt>
-<dd><p>Adds a triangle mesh collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>mesh</strong> (<a class="reference internal" href="#warp.sim.Mesh" title="warp.sim.model.Mesh"><em>Mesh</em></a><em> | </em><em>None</em>) – The mesh object</p></li>
-<li><p><strong>scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – Scale to use for the collider</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh is solid, otherwise it is a hollow surface with the given wall thickness</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow mesh, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_shape_sdf">
-<span class="sig-name descname"><span class="pre">add_shape_sdf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">body</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pos</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(0.0,</span> <span class="pre">0.0,</span> <span class="pre">0.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sdf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">(1.0,</span> <span class="pre">1.0,</span> <span class="pre">1.0)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1e-05</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_shape_sdf" title="Link to this definition">#</a></dt>
-<dd><p>Adds SDF collision shape to a body.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>body</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the parent body this shape belongs to (use -1 for static shapes)</p></li>
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The location of the shape with respect to the parent frame</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The rotation of the shape with respect to the parent frame</p></li>
-<li><p><strong>sdf</strong> (<a class="reference internal" href="#warp.sim.SDF" title="warp.sim.model.SDF"><em>SDF</em></a><em> | </em><em>None</em>) – The sdf object</p></li>
-<li><p><strong>scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – Scale to use for the collider</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shape</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact elastic stiffness</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact damping stiffness</p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The contact friction stiffness</p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of friction</p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The coefficient of restitution</p></li>
-<li><p><strong>is_solid</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the mesh is solid, otherwise it is a hollow surface with the given wall thickness</p></li>
-<li><p><strong>thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Thickness to use for computing inertia of a hollow mesh, and for collision handling</p></li>
-<li><p><strong>has_ground_collision</strong> – If True, the mesh will collide with the ground plane if <cite>Model.ground</cite> is True</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The index of the added shape</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_particle">
-<span class="sig-name descname"><span class="pre">add_particle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vel</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mass</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">radius=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">flags=&lt;warp.types.uint32</span> <span class="pre">object&gt;</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_particle" title="Link to this definition">#</a></dt>
-<dd><p>Adds a single particle to the model</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The initial position of the particle</p></li>
-<li><p><strong>vel</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The initial velocity of the particle</p></li>
-<li><p><strong>mass</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The mass of the particle</p></li>
-<li><p><strong>flags</strong> (<em>uint32</em>) – The flags that control the dynamical behavior of the particle, see PARTICLE_FLAG_* constants</p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em> | </em><em>None</em>) – The radius of the particle used in collision handling. If None, the radius is set to the default value (default_particle_radius).</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>Set the mass equal to zero to create a ‘kinematic’ particle that does is not subject to dynamics.</p>
-</div>
-<dl class="field-list simple">
-<dt class="field-odd">Returns<span class="colon">:</span></dt>
-<dd class="field-odd"><p>The index of the particle in the system</p>
-</dd>
-<dt class="field-even">Parameters<span class="colon">:</span></dt>
-<dd class="field-even"><ul class="simple">
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – </p></li>
-<li><p><strong>vel</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – </p></li>
-<li><p><strong>mass</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>radius</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em> | </em><em>None</em>) – </p></li>
-<li><p><strong>flags</strong> (<em>uint32</em>) – </p></li>
-</ul>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_spring">
-<span class="sig-name descname"><span class="pre">add_spring</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">control</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_spring" title="Link to this definition">#</a></dt>
-<dd><p>Adds a spring between two particles in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>i</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the first particle</p></li>
-<li><p><strong>j</strong> – The index of the second particle</p></li>
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The elastic stiffness of the spring</p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping stiffness of the spring</p></li>
-<li><p><strong>control</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The actuation level of the spring</p></li>
-</ul>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The spring is created with a rest-length based on the distance
-between the particles in their initial configuration.</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_triangle">
-<span class="sig-name descname"><span class="pre">add_triangle</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ka</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_drag</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_lift</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_triangle" title="Link to this definition">#</a></dt>
-<dd><p>Adds a triangular FEM element between three particles in the system.</p>
-<p>Triangles are modeled as viscoelastic elements with elastic stiffness and damping
-Parameters specified on the model. See model.tri_ke, model.tri_kd.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>i</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the first particle</p></li>
-<li><p><strong>j</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the second particle</p></li>
-<li><p><strong>k</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the third particle</p></li>
-<li><p><strong>tri_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ka</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_drag</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_lift</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The area of the triangle</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The triangle is created with a rest-length based on the distance
-between the particles in their initial configuration.</p>
-</div>
-<div class="admonition-todo admonition" id="id1">
-<p class="admonition-title">Todo</p>
-<ul class="simple">
-<li><p>Expose elastic parameters on a per-element basis</p></li>
-</ul>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_triangles">
-<span class="sig-name descname"><span class="pre">add_triangles</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ka</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_drag</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_lift</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_triangles" title="Link to this definition">#</a></dt>
-<dd><p>Adds triangular FEM elements between groups of three particles in the system.</p>
-<p>Triangles are modeled as viscoelastic elements with elastic stiffness and damping
-Parameters specified on the model. See model.tri_ke, model.tri_kd.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>i</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – The indices of the first particle</p></li>
-<li><p><strong>j</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – The indices of the second particle</p></li>
-<li><p><strong>k</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – The indices of the third particle</p></li>
-<li><p><strong>tri_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-<li><p><strong>tri_ka</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-<li><p><strong>tri_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-<li><p><strong>tri_drag</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-<li><p><strong>tri_lift</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The areas of the triangles</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a>[<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a>]</p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>A triangle is created with a rest-length based on the distance
-between the particles in their initial configuration.</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_tetrahedron">
-<span class="sig-name descname"><span class="pre">add_tetrahedron</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_lambda</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_damp</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_tetrahedron" title="Link to this definition">#</a></dt>
-<dd><p>Adds a tetrahedral FEM element between four particles in the system.</p>
-<p>Tetrahedra are modeled as viscoelastic elements with a NeoHookean energy
-density based on [Smith et al. 2018].</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>i</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the first particle</p></li>
-<li><p><strong>j</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the second particle</p></li>
-<li><p><strong>k</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the third particle</p></li>
-<li><p><strong>l</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the fourth particle</p></li>
-<li><p><strong>k_mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The first elastic Lame parameter</p></li>
-<li><p><strong>k_lambda</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The second elastic Lame parameter</p></li>
-<li><p><strong>k_damp</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The element’s damping stiffness</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The volume of the tetrahedron</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The tetrahedron is created with a rest-pose based on the particle’s initial configuration</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_edge">
-<span class="sig-name descname"><span class="pre">add_edge</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rest</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_edge" title="Link to this definition">#</a></dt>
-<dd><p>Adds a bending edge element between four particles in the system.</p>
-<p>Bending elements are designed to be between two connected triangles. Then
-bending energy is based of [Bridson et al. 2002]. Bending stiffness is controlled
-by the <cite>model.tri_kb</cite> parameter.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>i</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the first particle</p></li>
-<li><p><strong>j</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the second particle</p></li>
-<li><p><strong>k</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the third particle</p></li>
-<li><p><strong>l</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The index of the fourth particle</p></li>
-<li><p><strong>rest</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em> | </em><em>None</em>) – The rest angle across the edge in radians, if not specified it will be computed</p></li>
-<li><p><strong>edge_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>edge_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The edge lies between the particles indexed by ‘k’ and ‘l’ parameters with the opposing
-vertices indexed by ‘i’ and ‘j’. This defines two connected triangles with counter clockwise
-winding: (i, k, l), (j, l, k).</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_edges">
-<span class="sig-name descname"><span class="pre">add_edges</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">i</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">j</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">l</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rest</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_edges" title="Link to this definition">#</a></dt>
-<dd><p>Adds bending edge elements between groups of four particles in the system.</p>
-<p>Bending elements are designed to be between two connected triangles. Then
-bending energy is based of [Bridson et al. 2002]. Bending stiffness is controlled
-by the <cite>model.tri_kb</cite> parameter.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>i</strong> – The indices of the first particle</p></li>
-<li><p><strong>j</strong> – The indices of the second particle</p></li>
-<li><p><strong>k</strong> – The indices of the third particle</p></li>
-<li><p><strong>l</strong> – The indices of the fourth particle</p></li>
-<li><p><strong>rest</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – The rest angles across the edges in radians, if not specified they will be computed</p></li>
-<li><p><strong>edge_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-<li><p><strong>edge_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>] </em><em>| </em><em>None</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The edge lies between the particles indexed by ‘k’ and ‘l’ parameters with the opposing
-vertices indexed by ‘i’ and ‘j’. This defines two connected triangles with counter clockwise
-winding: (i, k, l), (j, l, k).</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_cloth_grid">
-<span class="sig-name descname"><span class="pre">add_cloth_grid</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vel</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_y</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cell_x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cell_y</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mass</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">reverse_winding</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_left</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_right</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_top</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_bottom</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ka</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_drag</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_lift</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">add_springs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spring_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spring_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_cloth_grid" title="Link to this definition">#</a></dt>
-<dd><p>Helper to create a regular planar cloth grid</p>
-<p>Creates a rectangular grid of particles with FEM triangles and bending elements
-automatically.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The position of the cloth in world space</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The orientation of the cloth in world space</p></li>
-<li><p><strong>vel</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The velocity of the cloth in world space</p></li>
-<li><p><strong>dim_x</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The number of rectangular cells along the x-axis</p></li>
-<li><p><strong>dim_y</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The number of rectangular cells along the y-axis</p></li>
-<li><p><strong>cell_x</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The width of each cell in the x-direction</p></li>
-<li><p><strong>cell_y</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The width of each cell in the y-direction</p></li>
-<li><p><strong>mass</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The mass of each particle</p></li>
-<li><p><strong>reverse_winding</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Flip the winding of the mesh</p></li>
-<li><p><strong>fix_left</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the left-most edge of particles kinematic (fixed in place)</p></li>
-<li><p><strong>fix_right</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the right-most edge of particles kinematic</p></li>
-<li><p><strong>fix_top</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the top-most edge of particles kinematic</p></li>
-<li><p><strong>fix_bottom</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the bottom-most edge of particles kinematic</p></li>
-<li><p><strong>tri_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ka</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_drag</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_lift</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>edge_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>edge_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>add_springs</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – </p></li>
-<li><p><strong>spring_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>spring_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_cloth_mesh">
-<span class="sig-name descname"><span class="pre">add_cloth_mesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vel</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vertices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_callback</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">face_callback</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ka</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_drag</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_lift</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">edge_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">add_springs</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spring_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">spring_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_cloth_mesh" title="Link to this definition">#</a></dt>
-<dd><p>Helper to create a cloth model from a regular triangle mesh</p>
-<p>Creates one FEM triangle element and one bending element for every face
-and edge in the input triangle mesh</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The position of the cloth in world space</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The orientation of the cloth in world space</p></li>
-<li><p><strong>vel</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The velocity of the cloth in world space</p></li>
-<li><p><strong>vertices</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em><em>]</em>) – A list of vertex positions</p></li>
-<li><p><strong>indices</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – A list of triangle indices, 3 entries per-face</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density per-area of the mesh</p></li>
-<li><p><strong>edge_callback</strong> – A user callback when an edge is created</p></li>
-<li><p><strong>face_callback</strong> – A user callback when a face is created</p></li>
-<li><p><strong>scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ka</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_drag</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_lift</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>edge_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>edge_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>add_springs</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – </p></li>
-<li><p><strong>spring_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>spring_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The mesh should be two manifold.</p>
-</div>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_soft_grid">
-<span class="sig-name descname"><span class="pre">add_soft_grid</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vel</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_y</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dim_z</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cell_x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cell_y</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cell_z</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_mu</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_lambda</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_damp</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_left</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_right</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_top</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">fix_bottom</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ka</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_drag</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_lift</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_soft_grid" title="Link to this definition">#</a></dt>
-<dd><p>Helper to create a rectangular tetrahedral FEM grid</p>
-<p>Creates a regular grid of FEM tetrahedra and surface triangles. Useful for example
-to create beams and sheets. Each hexahedral cell is decomposed into 5
-tetrahedral elements.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The position of the solid in world space</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The orientation of the solid in world space</p></li>
-<li><p><strong>vel</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The velocity of the solid in world space</p></li>
-<li><p><strong>dim_x</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The number of rectangular cells along the x-axis</p></li>
-<li><p><strong>dim_y</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The number of rectangular cells along the y-axis</p></li>
-<li><p><strong>dim_z</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – The number of rectangular cells along the z-axis</p></li>
-<li><p><strong>cell_x</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The width of each cell in the x-direction</p></li>
-<li><p><strong>cell_y</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The width of each cell in the y-direction</p></li>
-<li><p><strong>cell_z</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The width of each cell in the z-direction</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of each particle</p></li>
-<li><p><strong>k_mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The first elastic Lame parameter</p></li>
-<li><p><strong>k_lambda</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The second elastic Lame parameter</p></li>
-<li><p><strong>k_damp</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping stiffness</p></li>
-<li><p><strong>fix_left</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the left-most edge of particles kinematic (fixed in place)</p></li>
-<li><p><strong>fix_right</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the right-most edge of particles kinematic</p></li>
-<li><p><strong>fix_top</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the top-most edge of particles kinematic</p></li>
-<li><p><strong>fix_bottom</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Make the bottom-most edge of particles kinematic</p></li>
-<li><p><strong>tri_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ka</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_drag</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_lift</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.add_soft_mesh">
-<span class="sig-name descname"><span class="pre">add_soft_mesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">pos</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rot</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vel</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">vertices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_mu</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_lambda</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">k_damp</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_ka</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_drag</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">tri_lift</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.add_soft_mesh" title="Link to this definition">#</a></dt>
-<dd><p>Helper to create a tetrahedral model from an input tetrahedral mesh</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>pos</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The position of the solid in world space</p></li>
-<li><p><strong>rot</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The orientation of the solid in world space</p></li>
-<li><p><strong>vel</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em>) – The velocity of the solid in world space</p></li>
-<li><p><strong>vertices</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em><em>]</em>) – A list of vertex positions</p></li>
-<li><p><strong>indices</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – A list of tetrahedron indices, 4 entries per-element</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density per-area of the mesh</p></li>
-<li><p><strong>k_mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The first elastic Lame parameter</p></li>
-<li><p><strong>k_lambda</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The second elastic Lame parameter</p></li>
-<li><p><strong>k_damp</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping stiffness</p></li>
-<li><p><strong>scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_ka</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_drag</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>tri_lift</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.set_ground_plane">
-<span class="sig-name descname"><span class="pre">set_ground_plane</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">normal</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">offset</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.set_ground_plane" title="Link to this definition">#</a></dt>
-<dd><p>Creates a ground plane for the world. If the normal is not specified,
-the up_vector of the ModelBuilder is used.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-<li><p><strong>restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.ModelBuilder.finalize">
-<span class="sig-name descname"><span class="pre">finalize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.ModelBuilder.finalize" title="Link to this definition">#</a></dt>
-<dd><p>Convert this builder object to a concrete model for simulation.</p>
-<p>After building simulation elements this method should be called to transfer
-all data to device memory ready for simulation.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>device</strong> – The simulation device to use, e.g.: ‘cpu’, ‘cuda’</p></li>
-<li><p><strong>requires_grad</strong> – Whether to enable gradient computation for the model</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>A model object.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.sim.Model" title="warp.sim.model.Model"><em>Model</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.Model">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">Model</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.Model" title="Link to this definition">#</a></dt>
-<dd><p>Holds the definition of the simulation model</p>
-<p>This class holds the non-time varying description of the system, i.e.:
-all geometry, constraints, and parameters used to describe the simulation.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.requires_grad">
-<span class="sig-name descname"><span class="pre">requires_grad</span></span><a class="headerlink" href="#warp.sim.Model.requires_grad" title="Link to this definition">#</a></dt>
-<dd><p>Indicates whether the model was finalized with gradient computation enabled</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.num_envs">
-<span class="sig-name descname"><span class="pre">num_envs</span></span><a class="headerlink" href="#warp.sim.Model.num_envs" title="Link to this definition">#</a></dt>
-<dd><p>Number of articulation environments that were added to the ModelBuilder via <cite>add_builder</cite></p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_q">
-<span class="sig-name descname"><span class="pre">particle_q</span></span><a class="headerlink" href="#warp.sim.Model.particle_q" title="Link to this definition">#</a></dt>
-<dd><p>Particle positions, shape [particle_count, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_qd">
-<span class="sig-name descname"><span class="pre">particle_qd</span></span><a class="headerlink" href="#warp.sim.Model.particle_qd" title="Link to this definition">#</a></dt>
-<dd><p>Particle velocities, shape [particle_count, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_mass">
-<span class="sig-name descname"><span class="pre">particle_mass</span></span><a class="headerlink" href="#warp.sim.Model.particle_mass" title="Link to this definition">#</a></dt>
-<dd><p>Particle mass, shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_inv_mass">
-<span class="sig-name descname"><span class="pre">particle_inv_mass</span></span><a class="headerlink" href="#warp.sim.Model.particle_inv_mass" title="Link to this definition">#</a></dt>
-<dd><p>Particle inverse mass, shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_radius">
-<span class="sig-name descname"><span class="pre">particle_radius</span></span><a class="headerlink" href="#warp.sim.Model.particle_radius" title="Link to this definition">#</a></dt>
-<dd><p>Particle radius, shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_max_radius">
-<span class="sig-name descname"><span class="pre">particle_max_radius</span></span><a class="headerlink" href="#warp.sim.Model.particle_max_radius" title="Link to this definition">#</a></dt>
-<dd><p>Maximum particle radius (useful for HashGrid construction)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_ke">
-<span class="sig-name descname"><span class="pre">particle_ke</span></span><a class="headerlink" href="#warp.sim.Model.particle_ke" title="Link to this definition">#</a></dt>
-<dd><p>Particle normal contact stiffness (used by <a class="reference internal" href="#warp.sim.SemiImplicitIntegrator" title="warp.sim.SemiImplicitIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">SemiImplicitIntegrator</span></code></a>), shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_kd">
-<span class="sig-name descname"><span class="pre">particle_kd</span></span><a class="headerlink" href="#warp.sim.Model.particle_kd" title="Link to this definition">#</a></dt>
-<dd><p>Particle normal contact damping (used by <a class="reference internal" href="#warp.sim.SemiImplicitIntegrator" title="warp.sim.SemiImplicitIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">SemiImplicitIntegrator</span></code></a>), shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_kf">
-<span class="sig-name descname"><span class="pre">particle_kf</span></span><a class="headerlink" href="#warp.sim.Model.particle_kf" title="Link to this definition">#</a></dt>
-<dd><p>Particle friction force stiffness (used by <a class="reference internal" href="#warp.sim.SemiImplicitIntegrator" title="warp.sim.SemiImplicitIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">SemiImplicitIntegrator</span></code></a>), shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_mu">
-<span class="sig-name descname"><span class="pre">particle_mu</span></span><a class="headerlink" href="#warp.sim.Model.particle_mu" title="Link to this definition">#</a></dt>
-<dd><p>Particle friction coefficient, shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_cohesion">
-<span class="sig-name descname"><span class="pre">particle_cohesion</span></span><a class="headerlink" href="#warp.sim.Model.particle_cohesion" title="Link to this definition">#</a></dt>
-<dd><p>Particle cohesion strength, shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_adhesion">
-<span class="sig-name descname"><span class="pre">particle_adhesion</span></span><a class="headerlink" href="#warp.sim.Model.particle_adhesion" title="Link to this definition">#</a></dt>
-<dd><p>Particle adhesion strength, shape [particle_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_grid">
-<span class="sig-name descname"><span class="pre">particle_grid</span></span><a class="headerlink" href="#warp.sim.Model.particle_grid" title="Link to this definition">#</a></dt>
-<dd><p>HashGrid instance used for accelerated simulation of particle interactions</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.HashGrid" title="warp.HashGrid">HashGrid</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_flags">
-<span class="sig-name descname"><span class="pre">particle_flags</span></span><a class="headerlink" href="#warp.sim.Model.particle_flags" title="Link to this definition">#</a></dt>
-<dd><p>Particle enabled state, shape [particle_count], bool</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_max_velocity">
-<span class="sig-name descname"><span class="pre">particle_max_velocity</span></span><a class="headerlink" href="#warp.sim.Model.particle_max_velocity" title="Link to this definition">#</a></dt>
-<dd><p>Maximum particle velocity (to prevent instability)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_transform">
-<span class="sig-name descname"><span class="pre">shape_transform</span></span><a class="headerlink" href="#warp.sim.Model.shape_transform" title="Link to this definition">#</a></dt>
-<dd><p>Rigid shape transforms, shape [shape_count, 7], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_body">
-<span class="sig-name descname"><span class="pre">shape_body</span></span><a class="headerlink" href="#warp.sim.Model.shape_body" title="Link to this definition">#</a></dt>
-<dd><p>Rigid shape body index, shape [shape_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_shapes">
-<span class="sig-name descname"><span class="pre">body_shapes</span></span><a class="headerlink" href="#warp.sim.Model.body_shapes" title="Link to this definition">#</a></dt>
-<dd><p>Mapping from body index to list of attached shape indices</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.12)">dict</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_materials">
-<span class="sig-name descname"><span class="pre">shape_materials</span></span><a class="headerlink" href="#warp.sim.Model.shape_materials" title="Link to this definition">#</a></dt>
-<dd><p>Rigid shape contact materials, shape [shape_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>ModelShapeMaterials</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_shape_geo">
-<span class="sig-name descname"><span class="pre">shape_shape_geo</span></span><a class="headerlink" href="#warp.sim.Model.shape_shape_geo" title="Link to this definition">#</a></dt>
-<dd><p>Shape geometry properties (geo type, scale, thickness, etc.), shape [shape_count, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>ModelShapeGeometry</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_geo_src">
-<span class="sig-name descname"><span class="pre">shape_geo_src</span></span><a class="headerlink" href="#warp.sim.Model.shape_geo_src" title="Link to this definition">#</a></dt>
-<dd><p>List of <cite>wp.Mesh</cite> instances used for rendering of mesh geometry</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)">list</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_collision_group">
-<span class="sig-name descname"><span class="pre">shape_collision_group</span></span><a class="headerlink" href="#warp.sim.Model.shape_collision_group" title="Link to this definition">#</a></dt>
-<dd><p>Collision group of each shape, shape [shape_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)">list</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_collision_group_map">
-<span class="sig-name descname"><span class="pre">shape_collision_group_map</span></span><a class="headerlink" href="#warp.sim.Model.shape_collision_group_map" title="Link to this definition">#</a></dt>
-<dd><p>Mapping from collision group to list of shape indices</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.12)">dict</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_collision_filter_pairs">
-<span class="sig-name descname"><span class="pre">shape_collision_filter_pairs</span></span><a class="headerlink" href="#warp.sim.Model.shape_collision_filter_pairs" title="Link to this definition">#</a></dt>
-<dd><p>Pairs of shape indices that should not collide</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#set" title="(in Python v3.12)">set</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_collision_radius">
-<span class="sig-name descname"><span class="pre">shape_collision_radius</span></span><a class="headerlink" href="#warp.sim.Model.shape_collision_radius" title="Link to this definition">#</a></dt>
-<dd><p>Collision radius of each shape used for bounding sphere broadphase collision checking, shape [shape_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_ground_collision">
-<span class="sig-name descname"><span class="pre">shape_ground_collision</span></span><a class="headerlink" href="#warp.sim.Model.shape_ground_collision" title="Link to this definition">#</a></dt>
-<dd><p>Indicates whether each shape should collide with the ground, shape [shape_count], bool</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)">list</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_contact_pairs">
-<span class="sig-name descname"><span class="pre">shape_contact_pairs</span></span><a class="headerlink" href="#warp.sim.Model.shape_contact_pairs" title="Link to this definition">#</a></dt>
-<dd><p>Pairs of shape indices that may collide, shape [contact_pair_count, 2], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_ground_contact_pairs">
-<span class="sig-name descname"><span class="pre">shape_ground_contact_pairs</span></span><a class="headerlink" href="#warp.sim.Model.shape_ground_contact_pairs" title="Link to this definition">#</a></dt>
-<dd><p>Pairs of shape, ground indices that may collide, shape [ground_contact_pair_count, 2], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.spring_indices">
-<span class="sig-name descname"><span class="pre">spring_indices</span></span><a class="headerlink" href="#warp.sim.Model.spring_indices" title="Link to this definition">#</a></dt>
-<dd><p>Particle spring indices, shape [spring_count*2], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.spring_rest_length">
-<span class="sig-name descname"><span class="pre">spring_rest_length</span></span><a class="headerlink" href="#warp.sim.Model.spring_rest_length" title="Link to this definition">#</a></dt>
-<dd><p>Particle spring rest length, shape [spring_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.spring_stiffness">
-<span class="sig-name descname"><span class="pre">spring_stiffness</span></span><a class="headerlink" href="#warp.sim.Model.spring_stiffness" title="Link to this definition">#</a></dt>
-<dd><p>Particle spring stiffness, shape [spring_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.spring_damping">
-<span class="sig-name descname"><span class="pre">spring_damping</span></span><a class="headerlink" href="#warp.sim.Model.spring_damping" title="Link to this definition">#</a></dt>
-<dd><p>Particle spring damping, shape [spring_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.spring_control">
-<span class="sig-name descname"><span class="pre">spring_control</span></span><a class="headerlink" href="#warp.sim.Model.spring_control" title="Link to this definition">#</a></dt>
-<dd><p>Particle spring activation, shape [spring_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tri_indices">
-<span class="sig-name descname"><span class="pre">tri_indices</span></span><a class="headerlink" href="#warp.sim.Model.tri_indices" title="Link to this definition">#</a></dt>
-<dd><p>Triangle element indices, shape [tri_count*3], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tri_poses">
-<span class="sig-name descname"><span class="pre">tri_poses</span></span><a class="headerlink" href="#warp.sim.Model.tri_poses" title="Link to this definition">#</a></dt>
-<dd><p>Triangle element rest pose, shape [tri_count, 2, 2], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tri_activations">
-<span class="sig-name descname"><span class="pre">tri_activations</span></span><a class="headerlink" href="#warp.sim.Model.tri_activations" title="Link to this definition">#</a></dt>
-<dd><p>Triangle element activations, shape [tri_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tri_materials">
-<span class="sig-name descname"><span class="pre">tri_materials</span></span><a class="headerlink" href="#warp.sim.Model.tri_materials" title="Link to this definition">#</a></dt>
-<dd><p>Triangle element materials, shape [tri_count, 5], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.edge_indices">
-<span class="sig-name descname"><span class="pre">edge_indices</span></span><a class="headerlink" href="#warp.sim.Model.edge_indices" title="Link to this definition">#</a></dt>
-<dd><p>Bending edge indices, shape [edge_count*4], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.edge_rest_angle">
-<span class="sig-name descname"><span class="pre">edge_rest_angle</span></span><a class="headerlink" href="#warp.sim.Model.edge_rest_angle" title="Link to this definition">#</a></dt>
-<dd><p>Bending edge rest angle, shape [edge_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.edge_bending_properties">
-<span class="sig-name descname"><span class="pre">edge_bending_properties</span></span><a class="headerlink" href="#warp.sim.Model.edge_bending_properties" title="Link to this definition">#</a></dt>
-<dd><p>Bending edge stiffness and damping parameters, shape [edge_count, 2], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tet_indices">
-<span class="sig-name descname"><span class="pre">tet_indices</span></span><a class="headerlink" href="#warp.sim.Model.tet_indices" title="Link to this definition">#</a></dt>
-<dd><p>Tetrahedral element indices, shape [tet_count*4], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tet_poses">
-<span class="sig-name descname"><span class="pre">tet_poses</span></span><a class="headerlink" href="#warp.sim.Model.tet_poses" title="Link to this definition">#</a></dt>
-<dd><p>Tetrahedral rest poses, shape [tet_count, 3, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tet_activations">
-<span class="sig-name descname"><span class="pre">tet_activations</span></span><a class="headerlink" href="#warp.sim.Model.tet_activations" title="Link to this definition">#</a></dt>
-<dd><p>Tetrahedral volumetric activations, shape [tet_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tet_materials">
-<span class="sig-name descname"><span class="pre">tet_materials</span></span><a class="headerlink" href="#warp.sim.Model.tet_materials" title="Link to this definition">#</a></dt>
-<dd><p>Tetrahedral elastic parameters in form <span class="math notranslate nohighlight">\(k_{mu}, k_{lambda}, k_{damp}\)</span>, shape [tet_count, 3]</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_q">
-<span class="sig-name descname"><span class="pre">body_q</span></span><a class="headerlink" href="#warp.sim.Model.body_q" title="Link to this definition">#</a></dt>
-<dd><p>Poses of rigid bodies used for state initialization, shape [body_count, 7], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_qd">
-<span class="sig-name descname"><span class="pre">body_qd</span></span><a class="headerlink" href="#warp.sim.Model.body_qd" title="Link to this definition">#</a></dt>
-<dd><p>Velocities of rigid bodies used for state initialization, shape [body_count, 6], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_com">
-<span class="sig-name descname"><span class="pre">body_com</span></span><a class="headerlink" href="#warp.sim.Model.body_com" title="Link to this definition">#</a></dt>
-<dd><p>Rigid body center of mass (in local frame), shape [body_count, 7], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_inertia">
-<span class="sig-name descname"><span class="pre">body_inertia</span></span><a class="headerlink" href="#warp.sim.Model.body_inertia" title="Link to this definition">#</a></dt>
-<dd><p>Rigid body inertia tensor (relative to COM), shape [body_count, 3, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_inv_inertia">
-<span class="sig-name descname"><span class="pre">body_inv_inertia</span></span><a class="headerlink" href="#warp.sim.Model.body_inv_inertia" title="Link to this definition">#</a></dt>
-<dd><p>Rigid body inverse inertia tensor (relative to COM), shape [body_count, 3, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_mass">
-<span class="sig-name descname"><span class="pre">body_mass</span></span><a class="headerlink" href="#warp.sim.Model.body_mass" title="Link to this definition">#</a></dt>
-<dd><p>Rigid body mass, shape [body_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_inv_mass">
-<span class="sig-name descname"><span class="pre">body_inv_mass</span></span><a class="headerlink" href="#warp.sim.Model.body_inv_mass" title="Link to this definition">#</a></dt>
-<dd><p>Rigid body inverse mass, shape [body_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_name">
-<span class="sig-name descname"><span class="pre">body_name</span></span><a class="headerlink" href="#warp.sim.Model.body_name" title="Link to this definition">#</a></dt>
-<dd><p>Rigid body names, shape [body_count], str</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)">list</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_q">
-<span class="sig-name descname"><span class="pre">joint_q</span></span><a class="headerlink" href="#warp.sim.Model.joint_q" title="Link to this definition">#</a></dt>
-<dd><p>Generalized joint positions used for state initialization, shape [joint_coord_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_qd">
-<span class="sig-name descname"><span class="pre">joint_qd</span></span><a class="headerlink" href="#warp.sim.Model.joint_qd" title="Link to this definition">#</a></dt>
-<dd><p>Generalized joint velocities used for state initialization, shape [joint_dof_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_act">
-<span class="sig-name descname"><span class="pre">joint_act</span></span><a class="headerlink" href="#warp.sim.Model.joint_act" title="Link to this definition">#</a></dt>
-<dd><p>Generalized joint actuation force, shape [joint_dof_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_type">
-<span class="sig-name descname"><span class="pre">joint_type</span></span><a class="headerlink" href="#warp.sim.Model.joint_type" title="Link to this definition">#</a></dt>
-<dd><p>Joint type, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_parent">
-<span class="sig-name descname"><span class="pre">joint_parent</span></span><a class="headerlink" href="#warp.sim.Model.joint_parent" title="Link to this definition">#</a></dt>
-<dd><p>Joint parent body indices, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_child">
-<span class="sig-name descname"><span class="pre">joint_child</span></span><a class="headerlink" href="#warp.sim.Model.joint_child" title="Link to this definition">#</a></dt>
-<dd><p>Joint child body indices, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_X_p">
-<span class="sig-name descname"><span class="pre">joint_X_p</span></span><a class="headerlink" href="#warp.sim.Model.joint_X_p" title="Link to this definition">#</a></dt>
-<dd><p>Joint transform in parent frame, shape [joint_count, 7], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_X_c">
-<span class="sig-name descname"><span class="pre">joint_X_c</span></span><a class="headerlink" href="#warp.sim.Model.joint_X_c" title="Link to this definition">#</a></dt>
-<dd><p>Joint mass frame in child frame, shape [joint_count, 7], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_axis">
-<span class="sig-name descname"><span class="pre">joint_axis</span></span><a class="headerlink" href="#warp.sim.Model.joint_axis" title="Link to this definition">#</a></dt>
-<dd><p>Joint axis in child frame, shape [joint_axis_count, 3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_armature">
-<span class="sig-name descname"><span class="pre">joint_armature</span></span><a class="headerlink" href="#warp.sim.Model.joint_armature" title="Link to this definition">#</a></dt>
-<dd><p>Armature for each joint, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_target">
-<span class="sig-name descname"><span class="pre">joint_target</span></span><a class="headerlink" href="#warp.sim.Model.joint_target" title="Link to this definition">#</a></dt>
-<dd><p>Joint target position/velocity (depending on joint axis mode), shape [joint_axis_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_target_ke">
-<span class="sig-name descname"><span class="pre">joint_target_ke</span></span><a class="headerlink" href="#warp.sim.Model.joint_target_ke" title="Link to this definition">#</a></dt>
-<dd><p>Joint stiffness, shape [joint_axis_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_target_kd">
-<span class="sig-name descname"><span class="pre">joint_target_kd</span></span><a class="headerlink" href="#warp.sim.Model.joint_target_kd" title="Link to this definition">#</a></dt>
-<dd><p>Joint damping, shape [joint_axis_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_axis_start">
-<span class="sig-name descname"><span class="pre">joint_axis_start</span></span><a class="headerlink" href="#warp.sim.Model.joint_axis_start" title="Link to this definition">#</a></dt>
-<dd><p>Start index of the first axis per joint, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_axis_dim">
-<span class="sig-name descname"><span class="pre">joint_axis_dim</span></span><a class="headerlink" href="#warp.sim.Model.joint_axis_dim" title="Link to this definition">#</a></dt>
-<dd><p>Number of linear and angular axes per joint, shape [joint_count, 2], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_axis_mode">
-<span class="sig-name descname"><span class="pre">joint_axis_mode</span></span><a class="headerlink" href="#warp.sim.Model.joint_axis_mode" title="Link to this definition">#</a></dt>
-<dd><p>Joint axis mode, shape [joint_axis_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_linear_compliance">
-<span class="sig-name descname"><span class="pre">joint_linear_compliance</span></span><a class="headerlink" href="#warp.sim.Model.joint_linear_compliance" title="Link to this definition">#</a></dt>
-<dd><p>Joint linear compliance, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_angular_compliance">
-<span class="sig-name descname"><span class="pre">joint_angular_compliance</span></span><a class="headerlink" href="#warp.sim.Model.joint_angular_compliance" title="Link to this definition">#</a></dt>
-<dd><p>Joint linear compliance, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_enabled">
-<span class="sig-name descname"><span class="pre">joint_enabled</span></span><a class="headerlink" href="#warp.sim.Model.joint_enabled" title="Link to this definition">#</a></dt>
-<dd><p>Joint enabled, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_limit_lower">
-<span class="sig-name descname"><span class="pre">joint_limit_lower</span></span><a class="headerlink" href="#warp.sim.Model.joint_limit_lower" title="Link to this definition">#</a></dt>
-<dd><p>Joint lower position limits, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_limit_upper">
-<span class="sig-name descname"><span class="pre">joint_limit_upper</span></span><a class="headerlink" href="#warp.sim.Model.joint_limit_upper" title="Link to this definition">#</a></dt>
-<dd><p>Joint upper position limits, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_limit_ke">
-<span class="sig-name descname"><span class="pre">joint_limit_ke</span></span><a class="headerlink" href="#warp.sim.Model.joint_limit_ke" title="Link to this definition">#</a></dt>
-<dd><p>Joint position limit stiffness (used by SemiImplicitIntegrator), shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_limit_kd">
-<span class="sig-name descname"><span class="pre">joint_limit_kd</span></span><a class="headerlink" href="#warp.sim.Model.joint_limit_kd" title="Link to this definition">#</a></dt>
-<dd><p>Joint position limit damping (used by SemiImplicitIntegrator), shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_twist_lower">
-<span class="sig-name descname"><span class="pre">joint_twist_lower</span></span><a class="headerlink" href="#warp.sim.Model.joint_twist_lower" title="Link to this definition">#</a></dt>
-<dd><p>Joint lower twist limit, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_twist_upper">
-<span class="sig-name descname"><span class="pre">joint_twist_upper</span></span><a class="headerlink" href="#warp.sim.Model.joint_twist_upper" title="Link to this definition">#</a></dt>
-<dd><p>Joint upper twist limit, shape [joint_count], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_q_start">
-<span class="sig-name descname"><span class="pre">joint_q_start</span></span><a class="headerlink" href="#warp.sim.Model.joint_q_start" title="Link to this definition">#</a></dt>
-<dd><p>Start index of the first position coordinate per joint, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_qd_start">
-<span class="sig-name descname"><span class="pre">joint_qd_start</span></span><a class="headerlink" href="#warp.sim.Model.joint_qd_start" title="Link to this definition">#</a></dt>
-<dd><p>Start index of the first velocity coordinate per joint, shape [joint_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.articulation_start">
-<span class="sig-name descname"><span class="pre">articulation_start</span></span><a class="headerlink" href="#warp.sim.Model.articulation_start" title="Link to this definition">#</a></dt>
-<dd><p>Articulation start index, shape [articulation_count], int</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_name">
-<span class="sig-name descname"><span class="pre">joint_name</span></span><a class="headerlink" href="#warp.sim.Model.joint_name" title="Link to this definition">#</a></dt>
-<dd><p>Joint names, shape [joint_count], str</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#list" title="(in Python v3.12)">list</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_attach_ke">
-<span class="sig-name descname"><span class="pre">joint_attach_ke</span></span><a class="headerlink" href="#warp.sim.Model.joint_attach_ke" title="Link to this definition">#</a></dt>
-<dd><p>Joint attachment force stiffness (used by SemiImplicitIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_attach_kd">
-<span class="sig-name descname"><span class="pre">joint_attach_kd</span></span><a class="headerlink" href="#warp.sim.Model.joint_attach_kd" title="Link to this definition">#</a></dt>
-<dd><p>Joint attachment force damping (used by SemiImplicitIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_margin">
-<span class="sig-name descname"><span class="pre">soft_contact_margin</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_margin" title="Link to this definition">#</a></dt>
-<dd><p>Contact margin for generation of soft contacts</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_ke">
-<span class="sig-name descname"><span class="pre">soft_contact_ke</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_ke" title="Link to this definition">#</a></dt>
-<dd><p>Stiffness of soft contacts (used by SemiImplicitIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_kd">
-<span class="sig-name descname"><span class="pre">soft_contact_kd</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_kd" title="Link to this definition">#</a></dt>
-<dd><p>Damping of soft contacts (used by SemiImplicitIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_kf">
-<span class="sig-name descname"><span class="pre">soft_contact_kf</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_kf" title="Link to this definition">#</a></dt>
-<dd><p>Stiffness of friction force in soft contacts (used by SemiImplicitIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_mu">
-<span class="sig-name descname"><span class="pre">soft_contact_mu</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_mu" title="Link to this definition">#</a></dt>
-<dd><p>Friction coefficient of soft contacts</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_restitution">
-<span class="sig-name descname"><span class="pre">soft_contact_restitution</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_restitution" title="Link to this definition">#</a></dt>
-<dd><p>Restitution coefficient of soft contacts (used by XPBDIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.rigid_contact_margin">
-<span class="sig-name descname"><span class="pre">rigid_contact_margin</span></span><a class="headerlink" href="#warp.sim.Model.rigid_contact_margin" title="Link to this definition">#</a></dt>
-<dd><p>Contact margin for generation of rigid body contacts</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.rigid_contact_torsional_friction">
-<span class="sig-name descname"><span class="pre">rigid_contact_torsional_friction</span></span><a class="headerlink" href="#warp.sim.Model.rigid_contact_torsional_friction" title="Link to this definition">#</a></dt>
-<dd><p>Torsional friction coefficient for rigid body contacts (used by XPBDIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.rigid_contact_rolling_friction">
-<span class="sig-name descname"><span class="pre">rigid_contact_rolling_friction</span></span><a class="headerlink" href="#warp.sim.Model.rigid_contact_rolling_friction" title="Link to this definition">#</a></dt>
-<dd><p>Rolling friction coefficient for rigid body contacts (used by XPBDIntegrator)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.ground">
-<span class="sig-name descname"><span class="pre">ground</span></span><a class="headerlink" href="#warp.sim.Model.ground" title="Link to this definition">#</a></dt>
-<dd><p>Whether the ground plane and ground contacts are enabled</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="functions.html#warp.bool" title="warp.bool">bool</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.ground_plane">
-<span class="sig-name descname"><span class="pre">ground_plane</span></span><a class="headerlink" href="#warp.sim.Model.ground_plane" title="Link to this definition">#</a></dt>
-<dd><p>Ground plane 3D normal and offset, shape [4], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.up_vector">
-<span class="sig-name descname"><span class="pre">up_vector</span></span><a class="headerlink" href="#warp.sim.Model.up_vector" title="Link to this definition">#</a></dt>
-<dd><p>Up vector of the world, shape [3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>np.ndarray</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.up_axis">
-<span class="sig-name descname"><span class="pre">up_axis</span></span><a class="headerlink" href="#warp.sim.Model.up_axis" title="Link to this definition">#</a></dt>
-<dd><p>Up axis, 0 for x, 1 for y, 2 for z</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.gravity">
-<span class="sig-name descname"><span class="pre">gravity</span></span><a class="headerlink" href="#warp.sim.Model.gravity" title="Link to this definition">#</a></dt>
-<dd><p>Gravity vector, shape [3], float</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>np.ndarray</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.particle_count">
-<span class="sig-name descname"><span class="pre">particle_count</span></span><a class="headerlink" href="#warp.sim.Model.particle_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of particles in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.body_count">
-<span class="sig-name descname"><span class="pre">body_count</span></span><a class="headerlink" href="#warp.sim.Model.body_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of bodies in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.shape_count">
-<span class="sig-name descname"><span class="pre">shape_count</span></span><a class="headerlink" href="#warp.sim.Model.shape_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of shapes in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_count">
-<span class="sig-name descname"><span class="pre">joint_count</span></span><a class="headerlink" href="#warp.sim.Model.joint_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of joints in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tri_count">
-<span class="sig-name descname"><span class="pre">tri_count</span></span><a class="headerlink" href="#warp.sim.Model.tri_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of triangles in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.tet_count">
-<span class="sig-name descname"><span class="pre">tet_count</span></span><a class="headerlink" href="#warp.sim.Model.tet_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of tetrahedra in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.edge_count">
-<span class="sig-name descname"><span class="pre">edge_count</span></span><a class="headerlink" href="#warp.sim.Model.edge_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of edges in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.spring_count">
-<span class="sig-name descname"><span class="pre">spring_count</span></span><a class="headerlink" href="#warp.sim.Model.spring_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of springs in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.contact_count">
-<span class="sig-name descname"><span class="pre">contact_count</span></span><a class="headerlink" href="#warp.sim.Model.contact_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of contacts in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.muscle_count">
-<span class="sig-name descname"><span class="pre">muscle_count</span></span><a class="headerlink" href="#warp.sim.Model.muscle_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of muscles in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.articulation_count">
-<span class="sig-name descname"><span class="pre">articulation_count</span></span><a class="headerlink" href="#warp.sim.Model.articulation_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of articulations in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_dof_count">
-<span class="sig-name descname"><span class="pre">joint_dof_count</span></span><a class="headerlink" href="#warp.sim.Model.joint_dof_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of velocity degrees of freedom of all joints in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.joint_coord_count">
-<span class="sig-name descname"><span class="pre">joint_coord_count</span></span><a class="headerlink" href="#warp.sim.Model.joint_coord_count" title="Link to this definition">#</a></dt>
-<dd><p>Total number of position degrees of freedom of all joints in the system</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Model.device">
-<span class="sig-name descname"><span class="pre">device</span></span><a class="headerlink" href="#warp.sim.Model.device" title="Link to this definition">#</a></dt>
-<dd><p>Device on which the Model was allocated</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>wp.Device</p>
-</dd>
-</dl>
-</dd></dl>
-
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>It is strongly recommended to use the ModelBuilder to construct a simulation rather
-than creating your own Model object directly, however it is possible to do so if
-desired.</p>
-</div>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.Model.state">
-<span class="sig-name descname"><span class="pre">state</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">requires_grad</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.Model.state" title="Link to this definition">#</a></dt>
-<dd><p>Returns a state object for the model</p>
-<p>The returned state will be initialized with the initial configuration given in
-the model description.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="#warp.sim.State" title="warp.sim.model.State"><em>State</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.Model.count_contact_points">
-<span class="sig-name descname"><span class="pre">count_contact_points</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.Model.count_contact_points" title="Link to this definition">#</a></dt>
-<dd><p>Counts the maximum number of contact points that need to be allocated.</p>
-</dd></dl>
-
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.Model.flatten">
-<span class="sig-name descname"><span class="pre">flatten</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.Model.flatten" title="Link to this definition">#</a></dt>
-<dd><p>Returns a list of Tensors stored by the model</p>
-<p>This function is intended to be used internal-only but can be used to obtain
-a set of all tensors owned by the model.</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.sim.Model.soft_contact_max">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">soft_contact_max</span></span><a class="headerlink" href="#warp.sim.Model.soft_contact_max" title="Link to this definition">#</a></dt>
-<dd><p>Maximum number of soft contacts that can be registered</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.JointAxis">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JointAxis</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">axis</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_lower</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-inf</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_upper</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">inf</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mode</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.JointAxis" title="Link to this definition">#</a></dt>
-<dd><p>Describes a joint axis that can have limits and be driven towards a target.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.axis">
-<span class="sig-name descname"><span class="pre">axis</span></span><a class="headerlink" href="#warp.sim.JointAxis.axis" title="Link to this definition">#</a></dt>
-<dd><p>The 3D axis that this JointAxis object describes, or alternatively another JointAxis object to copy from</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>3D vector or <a class="reference internal" href="#warp.sim.JointAxis" title="warp.sim.JointAxis">JointAxis</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.limit_lower">
-<span class="sig-name descname"><span class="pre">limit_lower</span></span><a class="headerlink" href="#warp.sim.JointAxis.limit_lower" title="Link to this definition">#</a></dt>
-<dd><p>The lower limit of the joint axis</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.limit_upper">
-<span class="sig-name descname"><span class="pre">limit_upper</span></span><a class="headerlink" href="#warp.sim.JointAxis.limit_upper" title="Link to this definition">#</a></dt>
-<dd><p>The upper limit of the joint axis</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.limit_ke">
-<span class="sig-name descname"><span class="pre">limit_ke</span></span><a class="headerlink" href="#warp.sim.JointAxis.limit_ke" title="Link to this definition">#</a></dt>
-<dd><p>The elastic stiffness of the joint axis limits, only respected by <a class="reference internal" href="#warp.sim.SemiImplicitIntegrator" title="warp.sim.SemiImplicitIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">SemiImplicitIntegrator</span></code></a></p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.limit_kd">
-<span class="sig-name descname"><span class="pre">limit_kd</span></span><a class="headerlink" href="#warp.sim.JointAxis.limit_kd" title="Link to this definition">#</a></dt>
-<dd><p>The damping stiffness of the joint axis limits, only respected by <a class="reference internal" href="#warp.sim.SemiImplicitIntegrator" title="warp.sim.SemiImplicitIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">SemiImplicitIntegrator</span></code></a></p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.target">
-<span class="sig-name descname"><span class="pre">target</span></span><a class="headerlink" href="#warp.sim.JointAxis.target" title="Link to this definition">#</a></dt>
-<dd><p>The target position or velocity (depending on the mode, see <a class="reference internal" href="#joint-modes">Joint modes</a>) of the joint axis</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.target_ke">
-<span class="sig-name descname"><span class="pre">target_ke</span></span><a class="headerlink" href="#warp.sim.JointAxis.target_ke" title="Link to this definition">#</a></dt>
-<dd><p>The proportional gain of the joint axis target drive PD controller</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.target_kd">
-<span class="sig-name descname"><span class="pre">target_kd</span></span><a class="headerlink" href="#warp.sim.JointAxis.target_kd" title="Link to this definition">#</a></dt>
-<dd><p>The derivative gain of the joint axis target drive PD controller</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.JointAxis.mode">
-<span class="sig-name descname"><span class="pre">mode</span></span><a class="headerlink" href="#warp.sim.JointAxis.mode" title="Link to this definition">#</a></dt>
-<dd><p>The mode of the joint axis, see <a class="reference internal" href="#joint-modes">Joint modes</a></p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.Mesh">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">Mesh</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">vertices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">indices</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">compute_inertia</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">is_solid</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.Mesh" title="Link to this definition">#</a></dt>
-<dd><p>Describes a triangle collision mesh for simulation</p>
-<section id="example-mesh-creation-from-a-triangle-obj-mesh-file">
-<h3>Example mesh creation from a triangle OBJ mesh file:<a class="headerlink" href="#example-mesh-creation-from-a-triangle-obj-mesh-file" title="Link to this heading">#</a></h3>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
-<span class="kn">import</span> <span class="nn">warp</span> <span class="k">as</span> <span class="nn">wp</span>
-<span class="kn">import</span> <span class="nn">warp.sim</span>
-
-<span class="k">def</span> <span class="nf">load_mesh</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">filename</span><span class="p">,</span> <span class="n">use_meshio</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
-    <span class="k">if</span> <span class="n">use_meshio</span><span class="p">:</span>
-        <span class="kn">import</span> <span class="nn">meshio</span>
-        <span class="n">m</span> <span class="o">=</span> <span class="n">meshio</span><span class="o">.</span><span class="n">read</span><span class="p">(</span><span class="n">filename</span><span class="p">)</span>
-        <span class="n">mesh_points</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">m</span><span class="o">.</span><span class="n">points</span><span class="p">)</span>
-        <span class="n">mesh_indices</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">m</span><span class="o">.</span><span class="n">cells</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">data</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">int32</span><span class="p">)</span><span class="o">.</span><span class="n">flatten</span><span class="p">()</span>
-    <span class="k">else</span><span class="p">:</span>
-        <span class="kn">import</span> <span class="nn">openmesh</span>
-        <span class="n">m</span> <span class="o">=</span> <span class="n">openmesh</span><span class="o">.</span><span class="n">read_trimesh</span><span class="p">(</span><span class="n">filename</span><span class="p">)</span>
-        <span class="n">mesh_points</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">m</span><span class="o">.</span><span class="n">points</span><span class="p">())</span>
-        <span class="n">mesh_indices</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">m</span><span class="o">.</span><span class="n">face_vertex_indices</span><span class="p">(),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">int32</span><span class="p">)</span><span class="o">.</span><span class="n">flatten</span><span class="p">()</span>
-    <span class="k">return</span> <span class="n">wp</span><span class="o">.</span><span class="n">sim</span><span class="o">.</span><span class="n">Mesh</span><span class="p">(</span><span class="n">mesh_points</span><span class="p">,</span> <span class="n">mesh_indices</span><span class="p">)</span>
-
-<span class="n">mesh</span> <span class="o">=</span> <span class="n">load_mesh</span><span class="p">(</span><span class="s2">&quot;mesh.obj&quot;</span><span class="p">)</span>
-</pre></div>
-</div>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Mesh.vertices">
-<span class="sig-name descname"><span class="pre">vertices</span></span><a class="headerlink" href="#warp.sim.Mesh.vertices" title="Link to this definition">#</a></dt>
-<dd><p>Mesh 3D vertices points</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>List[Vec3]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Mesh.indices">
-<span class="sig-name descname"><span class="pre">indices</span></span><a class="headerlink" href="#warp.sim.Mesh.indices" title="Link to this definition">#</a></dt>
-<dd><p>Mesh indices as a flattened list of vertex indices describing triangles</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>List[<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Mesh.I">
-<span class="sig-name descname"><span class="pre">I</span></span><a class="headerlink" href="#warp.sim.Mesh.I" title="Link to this definition">#</a></dt>
-<dd><p>3x3 inertia matrix of the mesh assuming density of 1.0 (around the center of mass)</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>Mat33</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Mesh.mass">
-<span class="sig-name descname"><span class="pre">mass</span></span><a class="headerlink" href="#warp.sim.Mesh.mass" title="Link to this definition">#</a></dt>
-<dd><p>The total mass of the body assuming density of 1.0</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.Mesh.com">
-<span class="sig-name descname"><span class="pre">com</span></span><a class="headerlink" href="#warp.sim.Mesh.com" title="Link to this definition">#</a></dt>
-<dd><p>The center of mass of the body</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>Vec3</p>
-</dd>
-</dl>
-</dd></dl>
-
-<p>Construct a Mesh object from a triangle mesh</p>
-<p>The mesh center of mass and inertia tensor will automatically be
-calculated using a density of 1.0. This computation is only valid
-if the mesh is closed (two-manifold).</p>
-<dl class="field-list simple">
-<dt class="field-odd">param vertices<span class="colon">:</span></dt>
-<dd class="field-odd"><p>List of vertices in the mesh</p>
-</dd>
-<dt class="field-even">param indices<span class="colon">:</span></dt>
-<dd class="field-even"><p>List of triangle indices, 3 per-element</p>
-</dd>
-<dt class="field-odd">param compute_inertia<span class="colon">:</span></dt>
-<dd class="field-odd"><p>If True, the mass, inertia tensor and center of mass will be computed assuming density of 1.0</p>
-</dd>
-<dt class="field-even">param is_solid<span class="colon">:</span></dt>
-<dd class="field-even"><p>If True, the mesh is assumed to be a solid during inertia computation, otherwise it is assumed to be a hollow surface</p>
-</dd>
-</dl>
-<dl class="py method">
-<dt class="sig sig-object py" id="warp.sim.Mesh.finalize">
-<span class="sig-name descname"><span class="pre">finalize</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.Mesh.finalize" title="Link to this definition">#</a></dt>
-<dd><p>Constructs a simulation-ready <a class="reference internal" href="#warp.sim.Mesh" title="warp.sim.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a> object from the mesh data and returns its ID.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>device</strong> – The device on which to allocate the mesh buffers</p>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>The ID of the simulation-ready <a class="reference internal" href="#warp.sim.Mesh" title="warp.sim.Mesh"><code class="xref py py-class docutils literal notranslate"><span class="pre">Mesh</span></code></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>vertices</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a><em>]</em><em>]</em>) – </p></li>
-<li><p><strong>indices</strong> (<a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.List" title="(in Python v3.12)"><em>List</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.SDF">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">SDF</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">volume</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">I</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">array([[1.,</span> <span class="pre">0.,</span> <span class="pre">0.],</span> <span class="pre">[0.,</span> <span class="pre">1.,</span> <span class="pre">0.],</span> <span class="pre">[0.,</span> <span class="pre">0.,</span> <span class="pre">1.]],</span> <span class="pre">dtype=float32)</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mass</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">com</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">array([0.,</span> <span class="pre">0.,</span> <span class="pre">0.])</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.SDF" title="Link to this definition">#</a></dt>
-<dd><p>Describes a signed distance field for simulation</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.SDF.volume">
-<span class="sig-name descname"><span class="pre">volume</span></span><a class="headerlink" href="#warp.sim.SDF.volume" title="Link to this definition">#</a></dt>
-<dd><p>The volume defining the SDF</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.Volume" title="warp.Volume">Volume</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.SDF.I">
-<span class="sig-name descname"><span class="pre">I</span></span><a class="headerlink" href="#warp.sim.SDF.I" title="Link to this definition">#</a></dt>
-<dd><p>3x3 inertia matrix of the SDF</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>Mat33</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.SDF.mass">
-<span class="sig-name descname"><span class="pre">mass</span></span><a class="headerlink" href="#warp.sim.SDF.mass" title="Link to this definition">#</a></dt>
-<dd><p>The total mass of the SDF</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)">float</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.SDF.com">
-<span class="sig-name descname"><span class="pre">com</span></span><a class="headerlink" href="#warp.sim.SDF.com" title="Link to this definition">#</a></dt>
-<dd><p>The center of mass of the SDF</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p>Vec3</p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-<section id="joint-types">
-<span id="id2"></span><h3>Joint types<a class="headerlink" href="#joint-types" title="Link to this heading">#</a></h3>
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_PRISMATIC">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_PRISMATIC</span></span><a class="headerlink" href="#warp.sim.JOINT_PRISMATIC" title="Link to this definition">#</a></dt>
-<dd><p>Prismatic (slider) joint</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_REVOLUTE">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_REVOLUTE</span></span><a class="headerlink" href="#warp.sim.JOINT_REVOLUTE" title="Link to this definition">#</a></dt>
-<dd><p>Revolute (hinge) joint</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_BALL">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_BALL</span></span><a class="headerlink" href="#warp.sim.JOINT_BALL" title="Link to this definition">#</a></dt>
-<dd><p>Ball (spherical) joint with quaternion state representation</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_FIXED">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_FIXED</span></span><a class="headerlink" href="#warp.sim.JOINT_FIXED" title="Link to this definition">#</a></dt>
-<dd><p>Fixed (static) joint</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_FREE">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_FREE</span></span><a class="headerlink" href="#warp.sim.JOINT_FREE" title="Link to this definition">#</a></dt>
-<dd><p>Free (floating) joint</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_COMPOUND">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_COMPOUND</span></span><a class="headerlink" href="#warp.sim.JOINT_COMPOUND" title="Link to this definition">#</a></dt>
-<dd><p>Compound joint with 3 rotational degrees of freedom</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_UNIVERSAL">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_UNIVERSAL</span></span><a class="headerlink" href="#warp.sim.JOINT_UNIVERSAL" title="Link to this definition">#</a></dt>
-<dd><p>Universal joint with 2 rotational degrees of freedom</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_DISTANCE">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_DISTANCE</span></span><a class="headerlink" href="#warp.sim.JOINT_DISTANCE" title="Link to this definition">#</a></dt>
-<dd><p>Distance joint that keeps two bodies at a distance within its joint limits (only supported in <a class="reference internal" href="#warp.sim.XPBDIntegrator" title="warp.sim.XPBDIntegrator"><code class="xref py py-class docutils literal notranslate"><span class="pre">XPBDIntegrator</span></code></a> at the moment)</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_D6">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_D6</span></span><a class="headerlink" href="#warp.sim.JOINT_D6" title="Link to this definition">#</a></dt>
-<dd><p>Generic D6 joint with up to 3 translational and 3 rotational degrees of freedom</p>
-</dd></dl>
-
-</section>
-<section id="joint-modes">
-<span id="id3"></span><h3>Joint modes<a class="headerlink" href="#joint-modes" title="Link to this heading">#</a></h3>
-<p>Joint modes control the behavior of joint axes and can be used to implement joint position or velocity drives.</p>
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_MODE_LIMIT">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_MODE_LIMIT</span></span><a class="headerlink" href="#warp.sim.JOINT_MODE_LIMIT" title="Link to this definition">#</a></dt>
-<dd><p>No target or velocity control is applied, the joint is limited to its joint limits</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_MODE_TARGET_POSITION">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_MODE_TARGET_POSITION</span></span><a class="headerlink" href="#warp.sim.JOINT_MODE_TARGET_POSITION" title="Link to this definition">#</a></dt>
-<dd><p>The joint is driven to a target position</p>
-</dd></dl>
-
-<dl class="py data">
-<dt class="sig sig-object py" id="warp.sim.JOINT_MODE_TARGET_VELOCITY">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">JOINT_MODE_TARGET_VELOCITY</span></span><a class="headerlink" href="#warp.sim.JOINT_MODE_TARGET_VELOCITY" title="Link to this definition">#</a></dt>
-<dd><p>The joint is driven to a target velocity</p>
-</dd></dl>
-
-</section>
-</section>
-<section id="state">
-<h2>State<a class="headerlink" href="#state" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.State">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">State</span></span><a class="headerlink" href="#warp.sim.State" title="Link to this definition">#</a></dt>
-<dd><p>The State object holds all <em>time-varying</em> data for a model.</p>
-<p>Time-varying data includes particle positions, velocities, rigid body states, and
-anything that is output from the integrator as derived data, e.g.: forces.</p>
-<p>The exact attributes depend on the contents of the model. State objects should
-generally be created using the <a class="reference internal" href="#warp.sim.Model.state" title="warp.sim.Model.state"><code class="xref py py-func docutils literal notranslate"><span class="pre">Model.state()</span></code></a> function.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.particle_count">
-<span class="sig-name descname"><span class="pre">particle_count</span></span><a class="headerlink" href="#warp.sim.State.particle_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of particles</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.body_count">
-<span class="sig-name descname"><span class="pre">body_count</span></span><a class="headerlink" href="#warp.sim.State.body_count" title="Link to this definition">#</a></dt>
-<dd><p>Number of rigid bodies</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.particle_q">
-<span class="sig-name descname"><span class="pre">particle_q</span></span><a class="headerlink" href="#warp.sim.State.particle_q" title="Link to this definition">#</a></dt>
-<dd><p>Tensor of particle positions</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.particle_qd">
-<span class="sig-name descname"><span class="pre">particle_qd</span></span><a class="headerlink" href="#warp.sim.State.particle_qd" title="Link to this definition">#</a></dt>
-<dd><p>Tensor of particle velocities</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.particle_f">
-<span class="sig-name descname"><span class="pre">particle_f</span></span><a class="headerlink" href="#warp.sim.State.particle_f" title="Link to this definition">#</a></dt>
-<dd><p>Tensor of particle forces</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.body_q">
-<span class="sig-name descname"><span class="pre">body_q</span></span><a class="headerlink" href="#warp.sim.State.body_q" title="Link to this definition">#</a></dt>
-<dd><p>Tensor of body coordinates</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.body_qd">
-<span class="sig-name descname"><span class="pre">body_qd</span></span><a class="headerlink" href="#warp.sim.State.body_qd" title="Link to this definition">#</a></dt>
-<dd><p>Tensor of body velocities</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sim.State.body_f">
-<span class="sig-name descname"><span class="pre">body_f</span></span><a class="headerlink" href="#warp.sim.State.body_f" title="Link to this definition">#</a></dt>
-<dd><p>Tensor of body forces</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="runtime.html#warp.array" title="warp.array">array</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</dd></dl>
-
-</section>
-<section id="forward-inverse-kinematics">
-<span id="fk-ik"></span><h2>Forward / Inverse Kinematics<a class="headerlink" href="#forward-inverse-kinematics" title="Link to this heading">#</a></h2>
-<p>Articulated rigid-body mechanisms are kinematically described by the joints that connect the bodies as well as the relative relative transform from the parent and child body to the respective anchor frames of the joint in the parent and child body:</p>
-<a class="reference internal image-reference" href="../_images/joint_transforms.png"><img alt="../_images/joint_transforms.png" class="align-center" src="../_images/joint_transforms.png" style="width: 400px;" /></a>
-<div class="table-wrapper colwidths-given docutils container" id="id4">
-<table class="docutils align-default" id="id4">
-<caption><span class="caption-text">Variable names in the kernels from articulation.py</span><a class="headerlink" href="#id4" title="Link to this table">#</a></caption>
-<colgroup>
-<col style="width: 10.0%" />
-<col style="width: 90.0%" />
-</colgroup>
-<thead>
-<tr class="row-odd"><th class="head"><p>Symbol</p></th>
-<th class="head"><p>Description</p></th>
-</tr>
-</thead>
-<tbody>
-<tr class="row-even"><td><p>x_wp</p></td>
-<td><p>World transform of the parent body (stored at <a class="reference internal" href="#warp.sim.State.body_q" title="warp.sim.State.body_q"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_q</span></code></a>)</p></td>
-</tr>
-<tr class="row-odd"><td><p>x_wc</p></td>
-<td><p>World transform of the child body (stored at <a class="reference internal" href="#warp.sim.State.body_q" title="warp.sim.State.body_q"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_q</span></code></a>)</p></td>
-</tr>
-<tr class="row-even"><td><p>x_pj</p></td>
-<td><p>Transform from the parent body to the joint parent anchor frame (defined by <a class="reference internal" href="#warp.sim.Model.joint_X_p" title="warp.sim.Model.joint_X_p"><code class="xref py py-attr docutils literal notranslate"><span class="pre">Model.joint_X_p</span></code></a>)</p></td>
-</tr>
-<tr class="row-odd"><td><p>x_cj</p></td>
-<td><p>Transform from the child body to the joint child anchor frame (defined by <a class="reference internal" href="#warp.sim.Model.joint_X_c" title="warp.sim.Model.joint_X_c"><code class="xref py py-attr docutils literal notranslate"><span class="pre">Model.joint_X_c</span></code></a>)</p></td>
-</tr>
-<tr class="row-even"><td><p>x_j</p></td>
-<td><p>Joint transform from the joint parent anchor frame to the joint child anchor frame</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-<p>In the forward kinematics, the joint transform is determined by the joint coordinates (generalized joint positions <a class="reference internal" href="#warp.sim.State.body_q" title="warp.sim.State.body_q"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_q</span></code></a> and velocities <a class="reference internal" href="#warp.sim.State.body_qd" title="warp.sim.State.body_qd"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_qd</span></code></a>).
-Given the parent body’s world transform <span class="math notranslate nohighlight">\(x_{wp}\)</span> and the joint transform <span class="math notranslate nohighlight">\(x_{j}\)</span>, the child body’s world transform <span class="math notranslate nohighlight">\(x_{wc}\)</span> is computed as:</p>
-<div class="math-wrapper docutils container">
-<div class="math notranslate nohighlight">
-\[x_{wc} = x_{wp} \cdot x_{pj} \cdot x_{j} \cdot x_{cj}^{-1}.\]</div>
-</div>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sim.eval_fk">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">eval_fk</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_q</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_qd</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">mask</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">state</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.eval_fk" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the model’s forward kinematics given the joint coordinates and updates the state’s body information (<a class="reference internal" href="#warp.sim.State.body_q" title="warp.sim.State.body_q"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_q</span></code></a> and <a class="reference internal" href="#warp.sim.State.body_qd" title="warp.sim.State.body_qd"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_qd</span></code></a>).</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>model</strong> (<a class="reference internal" href="#warp.sim.Model" title="warp.sim.Model"><em>Model</em></a>) – The model to evaluate.</p></li>
-<li><p><strong>joint_q</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.array"><em>array</em></a>) – Generalized joint position coordinates, shape [joint_coord_count], float</p></li>
-<li><p><strong>joint_qd</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.array"><em>array</em></a>) – Generalized joint velocity coordinates, shape [joint_dof_count], float</p></li>
-<li><p><strong>mask</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.array"><em>array</em></a>) – The mask to use to enable / disable FK for an articulation. If None then treat all as enabled, shape [articulation_count], int</p></li>
-<li><p><strong>state</strong> (<a class="reference internal" href="#warp.sim.State" title="warp.sim.State"><em>State</em></a>) – The state to update.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sim.eval_ik">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">eval_ik</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">model</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">state</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_q</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_qd</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.eval_ik" title="Link to this definition">#</a></dt>
-<dd><p>Evaluates the model’s inverse kinematics given the state’s body information (<a class="reference internal" href="#warp.sim.State.body_q" title="warp.sim.State.body_q"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_q</span></code></a> and <a class="reference internal" href="#warp.sim.State.body_qd" title="warp.sim.State.body_qd"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_qd</span></code></a>) and updates the generalized joint coordinates <cite>joint_q</cite> and <cite>joint_qd</cite>.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>model</strong> (<a class="reference internal" href="#warp.sim.Model" title="warp.sim.Model"><em>Model</em></a>) – The model to evaluate.</p></li>
-<li><p><strong>state</strong> (<a class="reference internal" href="#warp.sim.State" title="warp.sim.State"><em>State</em></a>) – The state with the body’s maximal coordinates (positions <a class="reference internal" href="#warp.sim.State.body_q" title="warp.sim.State.body_q"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_q</span></code></a> and velocities <a class="reference internal" href="#warp.sim.State.body_qd" title="warp.sim.State.body_qd"><code class="xref py py-attr docutils literal notranslate"><span class="pre">State.body_qd</span></code></a>) to use.</p></li>
-<li><p><strong>joint_q</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.array"><em>array</em></a>) – Generalized joint position coordinates, shape [joint_coord_count], float</p></li>
-<li><p><strong>joint_qd</strong> (<a class="reference internal" href="runtime.html#warp.array" title="warp.array"><em>array</em></a>) – Generalized joint velocity coordinates, shape [joint_dof_count], float</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-<section id="integrators">
-<h2>Integrators<a class="headerlink" href="#integrators" title="Link to this heading">#</a></h2>
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.SemiImplicitIntegrator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">SemiImplicitIntegrator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">angular_damping</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.05</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.SemiImplicitIntegrator" title="Link to this definition">#</a></dt>
-<dd><p>A semi-implicit integrator using symplectic Euler</p>
-<p>After constructing <cite>Model</cite> and <cite>State</cite> objects this time-integrator
-may be used to advance the simulation state forward in time.</p>
-<p>Semi-implicit time integration is a variational integrator that
-preserves energy, however it not unconditionally stable, and requires a time-step
-small enough to support the required stiffness and damping forces.</p>
-<p>See: <a class="reference external" href="https://en.wikipedia.org/wiki/Semi-implicit_Euler_method">https://en.wikipedia.org/wiki/Semi-implicit_Euler_method</a></p>
-<p class="rubric">Example</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="n">integrator</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">SemiImplicitIntegrator</span><span class="p">()</span>
-
-<span class="c1"># simulation loop</span>
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">100</span><span class="p">):</span>
-    <span class="n">state</span> <span class="o">=</span> <span class="n">integrator</span><span class="o">.</span><span class="n">simulate</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">state_in</span><span class="p">,</span> <span class="n">state_out</span><span class="p">,</span> <span class="n">dt</span><span class="p">)</span>
-</pre></div>
-</div>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sim.XPBDIntegrator">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">XPBDIntegrator</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">iterations</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">2</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">soft_body_relaxation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.9</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">soft_contact_relaxation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.9</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_linear_relaxation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.7</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_angular_relaxation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.4</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rigid_contact_relaxation</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.8</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rigid_contact_con_weighting</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">angular_damping</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enable_restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.XPBDIntegrator" title="Link to this definition">#</a></dt>
-<dd><p>A implicit integrator using XPBD</p>
-<p>After constructing <cite>Model</cite> and <cite>State</cite> objects this time-integrator
-may be used to advance the simulation state forward in time.</p>
-<p class="rubric">Example</p>
-<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="n">integrator</span> <span class="o">=</span> <span class="n">wp</span><span class="o">.</span><span class="n">SemiImplicitIntegrator</span><span class="p">()</span>
-
-<span class="c1"># simulation loop</span>
-<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">100</span><span class="p">):</span>
-    <span class="n">state</span> <span class="o">=</span> <span class="n">integrator</span><span class="o">.</span><span class="n">simulate</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">state_in</span><span class="p">,</span> <span class="n">state_out</span><span class="p">,</span> <span class="n">dt</span><span class="p">)</span>
-</pre></div>
-</div>
-</dd></dl>
-
-</section>
-<section id="importers">
-<h2>Importers<a class="headerlink" href="#importers" title="Link to this heading">#</a></h2>
-<p>Warp sim supports the loading of simulation models from URDF, MuJoCo (MJCF), and USD Physics files.</p>
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sim.parse_urdf">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">parse_urdf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">urdf_filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">builder</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">floating=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">base_joint=None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density=1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stiffness=100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">damping=10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">armature=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape_ke=10000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape_kd=1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape_kf=100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape_mu=0.25</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape_restitution=0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">shape_thickness=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_ke=100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_kd=10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale=1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parse_visuals_as_colliders=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enable_self_collisions=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_inertial_definitions=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ensure_nonstatic_links=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">static_link_mass=0.01</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collapse_fixed_joints=False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.parse_urdf" title="Link to this definition">#</a></dt>
-<dd><p>Parses a URDF file and adds the bodies and joints to the given ModelBuilder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>urdf_filename</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – The filename of the URDF file to parse.</p></li>
-<li><p><strong>builder</strong> (<a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><em>ModelBuilder</em></a>) – The <a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><code class="xref py py-class docutils literal notranslate"><span class="pre">ModelBuilder</span></code></a> to add the bodies and joints to.</p></li>
-<li><p><strong>xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform to apply to the root body.</p></li>
-<li><p><strong>floating</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the root body is a free joint. If False, the root body is connected via a fixed joint to the world, unless a <cite>base_joint</cite> is defined.</p></li>
-<li><p><strong>base_joint</strong> (<em>Union</em><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>, </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.12)"><em>dict</em></a><em>]</em>) – The joint by which the root body is connected to the world. This can be either a string defining the joint axes of a D6 joint with comma-separated positional and angular axis names (e.g. “px,py,rz” for a D6 joint with linear axes in x, y and an angular axis in z) or a dict with joint parameters (see <a class="reference internal" href="#warp.sim.ModelBuilder.add_joint" title="warp.sim.ModelBuilder.add_joint"><code class="xref py py-meth docutils literal notranslate"><span class="pre">ModelBuilder.add_joint()</span></code></a>).</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shapes in kg/m^3 which will be used to calculate the body mass and inertia.</p></li>
-<li><p><strong>stiffness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joints.</p></li>
-<li><p><strong>damping</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joints.</p></li>
-<li><p><strong>armature</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The armature of the joints (bias to add to the inertia diagonals that may stabilize the simulation).</p></li>
-<li><p><strong>shape_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the shape contacts (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>shape_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the shape contacts (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>shape_kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The friction stiffness of the shape contacts (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>shape_mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The friction coefficient of the shape contacts.</p></li>
-<li><p><strong>shape_restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The restitution coefficient of the shape contacts.</p></li>
-<li><p><strong>shape_thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The thickness to add to the shape geometry.</p></li>
-<li><p><strong>limit_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joint limits (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>limit_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joint limits (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The scaling factor to apply to the imported mechanism.</p></li>
-<li><p><strong>parse_visuals_as_colliders</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the geometry defined under the <cite>&lt;visual&gt;</cite> tags is used for collision handling instead of the <cite>&lt;collision&gt;</cite> geoemtries.</p></li>
-<li><p><strong>enable_self_collisions</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, self-collisions are enabled.</p></li>
-<li><p><strong>ignore_inertial_definitions</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, the inertial parameters defined in the URDF are ignored and the inertia is calculated from the shape geometry.</p></li>
-<li><p><strong>ensure_nonstatic_links</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, links with zero mass are given a small mass (see <cite>static_link_mass</cite>) to ensure they are dynamic.</p></li>
-<li><p><strong>static_link_mass</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The mass to assign to links with zero mass (if <cite>ensure_nonstatic_links</cite> is set to True).</p></li>
-<li><p><strong>collapse_fixed_joints</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, fixed joints are removed and the respective bodies are merged.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sim.parse_mjcf">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">parse_mjcf</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">mjcf_filename</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">builder</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">xform=&lt;warp.types.transformf</span> <span class="pre">object&gt;</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">density=1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">stiffness=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">damping=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contact_ke=1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contact_kd=100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contact_kf=100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contact_mu=0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">contact_restitution=0.5</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_ke=100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">limit_kd=10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scale=1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">armature=0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">armature_scale=1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">parse_meshes=True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">enable_self_collisions=False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">up_axis='Z'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_classes=[]</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">collapse_fixed_joints=False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.parse_mjcf" title="Link to this definition">#</a></dt>
-<dd><p>Parses MuJoCo XML (MJCF) file and adds the bodies and joints to the given ModelBuilder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>mjcf_filename</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – The filename of the MuJoCo file to parse.</p></li>
-<li><p><strong>builder</strong> (<a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><em>ModelBuilder</em></a>) – The <a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><code class="xref py py-class docutils literal notranslate"><span class="pre">ModelBuilder</span></code></a> to add the bodies and joints to.</p></li>
-<li><p><strong>xform</strong> (<a class="reference internal" href="runtime.html#transform"><span class="std std-ref">transform</span></a>) – The transform to apply to the imported mechanism.</p></li>
-<li><p><strong>density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The density of the shapes in kg/m^3 which will be used to calculate the body mass and inertia.</p></li>
-<li><p><strong>stiffness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joints.</p></li>
-<li><p><strong>damping</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joints.</p></li>
-<li><p><strong>contact_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the shape contacts (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>contact_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the shape contacts (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>contact_kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The friction stiffness of the shape contacts (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>contact_mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The friction coefficient of the shape contacts.</p></li>
-<li><p><strong>contact_restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The restitution coefficient of the shape contacts.</p></li>
-<li><p><strong>limit_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The stiffness of the joint limits (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>limit_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The damping of the joint limits (used by SemiImplicitIntegrator).</p></li>
-<li><p><strong>scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The scaling factor to apply to the imported mechanism.</p></li>
-<li><p><strong>armature</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Default joint armature to use if <cite>armature</cite> has not been defined for a joint in the MJCF.</p></li>
-<li><p><strong>armature_scale</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – Scaling factor to apply to the MJCF-defined joint armature values.</p></li>
-<li><p><strong>parse_meshes</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – Whether geometries of type <cite>“mesh”</cite> should be parsed. If False, geometries of type <cite>“mesh”</cite> are ignored.</p></li>
-<li><p><strong>enable_self_collisions</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, self-collisions are enabled.</p></li>
-<li><p><strong>up_axis</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – The up axis of the mechanism. Can be either <cite>“X”</cite>, <cite>“Y”</cite> or <cite>“Z”</cite>. The default is <cite>“Z”</cite>.</p></li>
-<li><p><strong>ignore_classes</strong> (<em>List</em><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>]</em>) – A list of regular expressions. Bodies and joints with a class matching one of the regular expressions will be ignored.</p></li>
-<li><p><strong>collapse_fixed_joints</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, fixed joints are removed and the respective bodies are merged.</p></li>
-</ul>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>The inertia and masses of the bodies are calculated from the shape geometry and the given density. The values defined in the MJCF are not respected at the moment.</p>
-<p>The handling of advanced features, such as MJCF classes, is still experimental.</p>
-</div>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sim.parse_usd">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">parse_usd</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">source</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">builder</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_density</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">only_load_enabled_rigid_bodies</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">only_load_enabled_joints</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100000.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">250.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_kf</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">500.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_mu</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.6</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_restitution</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">default_thickness</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_limit_ke</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">100.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">joint_limit_kd</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">10.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">invert_rotations</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">verbose</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">ignore_paths</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">[]</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.parse_usd" title="Link to this definition">#</a></dt>
-<dd><p>Parses a Universal Scene Description (USD) stage containing UsdPhysics schema definitions for rigid-body articulations and adds the bodies, shapes and joints to the given ModelBuilder.</p>
-<p>The USD description has to be either a path (file name or URL), or an existing USD stage instance that implements the <a class="reference external" href="https://openusd.org/dev/api/class_usd_stage.html">UsdStage</a> interface.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>source</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>pxr.UsdStage</em>) – The file path to the USD file, or an existing USD stage instance.</p></li>
-<li><p><strong>builder</strong> (<a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><em>ModelBuilder</em></a>) – The <a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><code class="xref py py-class docutils literal notranslate"><span class="pre">ModelBuilder</span></code></a> to add the bodies and joints to.</p></li>
-<li><p><strong>default_density</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default density to use for bodies without a density attribute.</p></li>
-<li><p><strong>only_load_enabled_rigid_bodies</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, only rigid bodies which do not have <cite>physics:rigidBodyEnabled</cite> set to False are loaded.</p></li>
-<li><p><strong>only_load_enabled_joints</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, only joints which do not have <cite>physics:jointEnabled</cite> set to False are loaded.</p></li>
-<li><p><strong>default_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default contact stiffness to use, only considered by SemiImplicitIntegrator.</p></li>
-<li><p><strong>default_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default contact damping to use, only considered by SemiImplicitIntegrator.</p></li>
-<li><p><strong>default_kf</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default friction stiffness to use, only considered by SemiImplicitIntegrator.</p></li>
-<li><p><strong>default_mu</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default friction coefficient to use if a shape has not friction coefficient defined.</p></li>
-<li><p><strong>default_restitution</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default coefficient of restitution to use if a shape has not coefficient of restitution defined.</p></li>
-<li><p><strong>default_thickness</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The thickness to add to the shape geometry.</p></li>
-<li><p><strong>joint_limit_ke</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default stiffness to use for joint limits, only considered by SemiImplicitIntegrator.</p></li>
-<li><p><strong>joint_limit_kd</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#float" title="(in Python v3.12)"><em>float</em></a>) – The default damping to use for joint limits, only considered by SemiImplicitIntegrator.</p></li>
-<li><p><strong>invert_rotations</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, inverts any rotations defined in the shape transforms.</p></li>
-<li><p><strong>verbose</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, print additional information about the parsed USD file.</p></li>
-<li><p><strong>ignore_paths</strong> (<em>List</em><em>[</em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em>]</em>) – A list of regular expressions matching prim paths to ignore.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p><p>Dictionary with the following entries:</p>
-<div class="table-wrapper colwidths-given docutils container">
-<table class="docutils align-default">
-<colgroup>
-<col style="width: 25.0%" />
-<col style="width: 75.0%" />
-</colgroup>
-<tbody>
-<tr class="row-odd"><td><p>”fps”</p></td>
-<td><p>USD stage frames per second</p></td>
-</tr>
-<tr class="row-even"><td><p>”duration”</p></td>
-<td><p>Difference between end time code and start time code of the USD stage</p></td>
-</tr>
-<tr class="row-odd"><td><p>”up_axis”</p></td>
-<td><p>Upper-case string of the stage’s up axis (“X”, “Y”, or “Z”)</p></td>
-</tr>
-<tr class="row-even"><td><p>”path_shape_map”</p></td>
-<td><p>Mapping from prim path (str) of the UsdGeom to the respective shape index in <a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><code class="xref py py-class docutils literal notranslate"><span class="pre">ModelBuilder</span></code></a></p></td>
-</tr>
-<tr class="row-odd"><td><p>”path_body_map”</p></td>
-<td><p>Mapping from prim path (str) of a rigid body prim (e.g. that implements the PhysicsRigidBodyAPI) to the respective body index in <a class="reference internal" href="#warp.sim.ModelBuilder" title="warp.sim.ModelBuilder"><code class="xref py py-class docutils literal notranslate"><span class="pre">ModelBuilder</span></code></a></p></td>
-</tr>
-<tr class="row-even"><td><p>”path_shape_scale”</p></td>
-<td><p>Mapping from prim path (str) of the UsdGeom to its respective 3D world scale</p></td>
-</tr>
-<tr class="row-odd"><td><p>”mass_unit”</p></td>
-<td><p>The stage’s Kilograms Per Unit (KGPU) definition (1.0 by default)</p></td>
-</tr>
-<tr class="row-even"><td><p>”linear_unit”</p></td>
-<td><p>The stage’s Meters Per Unit (MPU) definition (1.0 by default)</p></td>
-</tr>
-</tbody>
-</table>
-</div>
-</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#dict" title="(in Python v3.12)">dict</a></p>
-</dd>
-</dl>
-<div class="admonition note">
-<p class="admonition-title">Note</p>
-<p>This importer is experimental and only supports a subset of the USD Physics schema. Please report any issues you encounter.</p>
-</div>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sim.resolve_usd_from_url">
-<span class="sig-prename descclassname"><span class="pre">warp.sim.</span></span><span class="sig-name descname"><span class="pre">resolve_usd_from_url</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">url</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target_folder_name</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">export_usda</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sim.resolve_usd_from_url" title="Link to this definition">#</a></dt>
-<dd><p>Downloads a USD file from a URL and resolves all references to other USD files to be downloaded to the given target folder.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>url</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – URL to the USD file.</p></li>
-<li><p><strong>target_folder_name</strong> (<a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a>) – Target folder name. If None, a timestamped folder will be created in the current directory.</p></li>
-<li><p><strong>export_usda</strong> (<a class="reference internal" href="functions.html#warp.bool" title="warp.bool"><em>bool</em></a>) – If True, converts each downloaded USD file to USDA and saves the additional USDA file in the target folder with the same base name as the original USD file.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Returns<span class="colon">:</span></dt>
-<dd class="field-even"><p>File path to the downloaded USD file.</p>
-</dd>
-<dt class="field-odd">Return type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)">str</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="sparse.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">warp.sparse</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="functions.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">Kernel Reference</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">warp.sim</a><ul>
-<li><a class="reference internal" href="#model">Model</a><ul>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder"><code class="docutils literal notranslate"><span class="pre">ModelBuilder</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_builder"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_builder()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_body"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_body()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_revolute"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_revolute()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_prismatic"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_prismatic()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_ball"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_ball()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_fixed"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_fixed()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_free"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_free()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_distance"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_distance()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_universal"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_universal()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_compound"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_compound()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_joint_d6"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_joint_d6()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.collapse_fixed_joints"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.collapse_fixed_joints()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_muscle"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_muscle()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_plane"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_plane()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_sphere"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_sphere()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_box"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_box()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_capsule"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_capsule()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_cylinder"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_cylinder()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_cone"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_cone()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_mesh"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_mesh()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_shape_sdf"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_shape_sdf()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_particle"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_particle()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_spring"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_spring()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_triangle"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_triangle()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_triangles"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_triangles()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_tetrahedron"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_tetrahedron()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_edge"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_edge()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_edges"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_edges()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_cloth_grid"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_cloth_grid()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_cloth_mesh"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_cloth_mesh()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_soft_grid"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_soft_grid()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.add_soft_mesh"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.add_soft_mesh()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.set_ground_plane"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.set_ground_plane()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.ModelBuilder.finalize"><code class="docutils literal notranslate"><span class="pre">ModelBuilder.finalize()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.sim.Model"><code class="docutils literal notranslate"><span class="pre">Model</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sim.Model.requires_grad"><code class="docutils literal notranslate"><span class="pre">Model.requires_grad</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.num_envs"><code class="docutils literal notranslate"><span class="pre">Model.num_envs</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_q"><code class="docutils literal notranslate"><span class="pre">Model.particle_q</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_qd"><code class="docutils literal notranslate"><span class="pre">Model.particle_qd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_mass"><code class="docutils literal notranslate"><span class="pre">Model.particle_mass</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_inv_mass"><code class="docutils literal notranslate"><span class="pre">Model.particle_inv_mass</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_radius"><code class="docutils literal notranslate"><span class="pre">Model.particle_radius</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_max_radius"><code class="docutils literal notranslate"><span class="pre">Model.particle_max_radius</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_ke"><code class="docutils literal notranslate"><span class="pre">Model.particle_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_kd"><code class="docutils literal notranslate"><span class="pre">Model.particle_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_kf"><code class="docutils literal notranslate"><span class="pre">Model.particle_kf</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_mu"><code class="docutils literal notranslate"><span class="pre">Model.particle_mu</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_cohesion"><code class="docutils literal notranslate"><span class="pre">Model.particle_cohesion</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_adhesion"><code class="docutils literal notranslate"><span class="pre">Model.particle_adhesion</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_grid"><code class="docutils literal notranslate"><span class="pre">Model.particle_grid</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_flags"><code class="docutils literal notranslate"><span class="pre">Model.particle_flags</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_max_velocity"><code class="docutils literal notranslate"><span class="pre">Model.particle_max_velocity</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_transform"><code class="docutils literal notranslate"><span class="pre">Model.shape_transform</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_body"><code class="docutils literal notranslate"><span class="pre">Model.shape_body</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_shapes"><code class="docutils literal notranslate"><span class="pre">Model.body_shapes</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_materials"><code class="docutils literal notranslate"><span class="pre">Model.shape_materials</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_shape_geo"><code class="docutils literal notranslate"><span class="pre">Model.shape_shape_geo</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_geo_src"><code class="docutils literal notranslate"><span class="pre">Model.shape_geo_src</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_collision_group"><code class="docutils literal notranslate"><span class="pre">Model.shape_collision_group</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_collision_group_map"><code class="docutils literal notranslate"><span class="pre">Model.shape_collision_group_map</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_collision_filter_pairs"><code class="docutils literal notranslate"><span class="pre">Model.shape_collision_filter_pairs</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_collision_radius"><code class="docutils literal notranslate"><span class="pre">Model.shape_collision_radius</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_ground_collision"><code class="docutils literal notranslate"><span class="pre">Model.shape_ground_collision</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_contact_pairs"><code class="docutils literal notranslate"><span class="pre">Model.shape_contact_pairs</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_ground_contact_pairs"><code class="docutils literal notranslate"><span class="pre">Model.shape_ground_contact_pairs</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.spring_indices"><code class="docutils literal notranslate"><span class="pre">Model.spring_indices</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.spring_rest_length"><code class="docutils literal notranslate"><span class="pre">Model.spring_rest_length</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.spring_stiffness"><code class="docutils literal notranslate"><span class="pre">Model.spring_stiffness</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.spring_damping"><code class="docutils literal notranslate"><span class="pre">Model.spring_damping</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.spring_control"><code class="docutils literal notranslate"><span class="pre">Model.spring_control</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tri_indices"><code class="docutils literal notranslate"><span class="pre">Model.tri_indices</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tri_poses"><code class="docutils literal notranslate"><span class="pre">Model.tri_poses</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tri_activations"><code class="docutils literal notranslate"><span class="pre">Model.tri_activations</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tri_materials"><code class="docutils literal notranslate"><span class="pre">Model.tri_materials</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.edge_indices"><code class="docutils literal notranslate"><span class="pre">Model.edge_indices</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.edge_rest_angle"><code class="docutils literal notranslate"><span class="pre">Model.edge_rest_angle</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.edge_bending_properties"><code class="docutils literal notranslate"><span class="pre">Model.edge_bending_properties</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tet_indices"><code class="docutils literal notranslate"><span class="pre">Model.tet_indices</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tet_poses"><code class="docutils literal notranslate"><span class="pre">Model.tet_poses</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tet_activations"><code class="docutils literal notranslate"><span class="pre">Model.tet_activations</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tet_materials"><code class="docutils literal notranslate"><span class="pre">Model.tet_materials</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_q"><code class="docutils literal notranslate"><span class="pre">Model.body_q</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_qd"><code class="docutils literal notranslate"><span class="pre">Model.body_qd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_com"><code class="docutils literal notranslate"><span class="pre">Model.body_com</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_inertia"><code class="docutils literal notranslate"><span class="pre">Model.body_inertia</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_inv_inertia"><code class="docutils literal notranslate"><span class="pre">Model.body_inv_inertia</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_mass"><code class="docutils literal notranslate"><span class="pre">Model.body_mass</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_inv_mass"><code class="docutils literal notranslate"><span class="pre">Model.body_inv_mass</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_name"><code class="docutils literal notranslate"><span class="pre">Model.body_name</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_q"><code class="docutils literal notranslate"><span class="pre">Model.joint_q</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_qd"><code class="docutils literal notranslate"><span class="pre">Model.joint_qd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_act"><code class="docutils literal notranslate"><span class="pre">Model.joint_act</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_type"><code class="docutils literal notranslate"><span class="pre">Model.joint_type</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_parent"><code class="docutils literal notranslate"><span class="pre">Model.joint_parent</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_child"><code class="docutils literal notranslate"><span class="pre">Model.joint_child</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_X_p"><code class="docutils literal notranslate"><span class="pre">Model.joint_X_p</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_X_c"><code class="docutils literal notranslate"><span class="pre">Model.joint_X_c</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_axis"><code class="docutils literal notranslate"><span class="pre">Model.joint_axis</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_armature"><code class="docutils literal notranslate"><span class="pre">Model.joint_armature</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_target"><code class="docutils literal notranslate"><span class="pre">Model.joint_target</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_target_ke"><code class="docutils literal notranslate"><span class="pre">Model.joint_target_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_target_kd"><code class="docutils literal notranslate"><span class="pre">Model.joint_target_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_axis_start"><code class="docutils literal notranslate"><span class="pre">Model.joint_axis_start</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_axis_dim"><code class="docutils literal notranslate"><span class="pre">Model.joint_axis_dim</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_axis_mode"><code class="docutils literal notranslate"><span class="pre">Model.joint_axis_mode</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_linear_compliance"><code class="docutils literal notranslate"><span class="pre">Model.joint_linear_compliance</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_angular_compliance"><code class="docutils literal notranslate"><span class="pre">Model.joint_angular_compliance</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_enabled"><code class="docutils literal notranslate"><span class="pre">Model.joint_enabled</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_limit_lower"><code class="docutils literal notranslate"><span class="pre">Model.joint_limit_lower</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_limit_upper"><code class="docutils literal notranslate"><span class="pre">Model.joint_limit_upper</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_limit_ke"><code class="docutils literal notranslate"><span class="pre">Model.joint_limit_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_limit_kd"><code class="docutils literal notranslate"><span class="pre">Model.joint_limit_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_twist_lower"><code class="docutils literal notranslate"><span class="pre">Model.joint_twist_lower</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_twist_upper"><code class="docutils literal notranslate"><span class="pre">Model.joint_twist_upper</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_q_start"><code class="docutils literal notranslate"><span class="pre">Model.joint_q_start</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_qd_start"><code class="docutils literal notranslate"><span class="pre">Model.joint_qd_start</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.articulation_start"><code class="docutils literal notranslate"><span class="pre">Model.articulation_start</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_name"><code class="docutils literal notranslate"><span class="pre">Model.joint_name</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_attach_ke"><code class="docutils literal notranslate"><span class="pre">Model.joint_attach_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_attach_kd"><code class="docutils literal notranslate"><span class="pre">Model.joint_attach_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_margin"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_margin</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_ke"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_kd"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_kf"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_kf</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_mu"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_mu</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_restitution"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_restitution</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.rigid_contact_margin"><code class="docutils literal notranslate"><span class="pre">Model.rigid_contact_margin</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.rigid_contact_torsional_friction"><code class="docutils literal notranslate"><span class="pre">Model.rigid_contact_torsional_friction</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.rigid_contact_rolling_friction"><code class="docutils literal notranslate"><span class="pre">Model.rigid_contact_rolling_friction</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.ground"><code class="docutils literal notranslate"><span class="pre">Model.ground</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.ground_plane"><code class="docutils literal notranslate"><span class="pre">Model.ground_plane</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.up_vector"><code class="docutils literal notranslate"><span class="pre">Model.up_vector</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.up_axis"><code class="docutils literal notranslate"><span class="pre">Model.up_axis</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.gravity"><code class="docutils literal notranslate"><span class="pre">Model.gravity</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.particle_count"><code class="docutils literal notranslate"><span class="pre">Model.particle_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.body_count"><code class="docutils literal notranslate"><span class="pre">Model.body_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.shape_count"><code class="docutils literal notranslate"><span class="pre">Model.shape_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_count"><code class="docutils literal notranslate"><span class="pre">Model.joint_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tri_count"><code class="docutils literal notranslate"><span class="pre">Model.tri_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.tet_count"><code class="docutils literal notranslate"><span class="pre">Model.tet_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.edge_count"><code class="docutils literal notranslate"><span class="pre">Model.edge_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.spring_count"><code class="docutils literal notranslate"><span class="pre">Model.spring_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.contact_count"><code class="docutils literal notranslate"><span class="pre">Model.contact_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.muscle_count"><code class="docutils literal notranslate"><span class="pre">Model.muscle_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.articulation_count"><code class="docutils literal notranslate"><span class="pre">Model.articulation_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_dof_count"><code class="docutils literal notranslate"><span class="pre">Model.joint_dof_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.joint_coord_count"><code class="docutils literal notranslate"><span class="pre">Model.joint_coord_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.device"><code class="docutils literal notranslate"><span class="pre">Model.device</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.state"><code class="docutils literal notranslate"><span class="pre">Model.state()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.count_contact_points"><code class="docutils literal notranslate"><span class="pre">Model.count_contact_points()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.flatten"><code class="docutils literal notranslate"><span class="pre">Model.flatten()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Model.soft_contact_max"><code class="docutils literal notranslate"><span class="pre">Model.soft_contact_max</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.sim.JointAxis"><code class="docutils literal notranslate"><span class="pre">JointAxis</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sim.JointAxis.axis"><code class="docutils literal notranslate"><span class="pre">JointAxis.axis</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.limit_lower"><code class="docutils literal notranslate"><span class="pre">JointAxis.limit_lower</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.limit_upper"><code class="docutils literal notranslate"><span class="pre">JointAxis.limit_upper</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.limit_ke"><code class="docutils literal notranslate"><span class="pre">JointAxis.limit_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.limit_kd"><code class="docutils literal notranslate"><span class="pre">JointAxis.limit_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.target"><code class="docutils literal notranslate"><span class="pre">JointAxis.target</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.target_ke"><code class="docutils literal notranslate"><span class="pre">JointAxis.target_ke</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.target_kd"><code class="docutils literal notranslate"><span class="pre">JointAxis.target_kd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JointAxis.mode"><code class="docutils literal notranslate"><span class="pre">JointAxis.mode</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.sim.Mesh"><code class="docutils literal notranslate"><span class="pre">Mesh</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sim.Mesh.vertices"><code class="docutils literal notranslate"><span class="pre">Mesh.vertices</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Mesh.indices"><code class="docutils literal notranslate"><span class="pre">Mesh.indices</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Mesh.I"><code class="docutils literal notranslate"><span class="pre">Mesh.I</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Mesh.mass"><code class="docutils literal notranslate"><span class="pre">Mesh.mass</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Mesh.com"><code class="docutils literal notranslate"><span class="pre">Mesh.com</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.Mesh.finalize"><code class="docutils literal notranslate"><span class="pre">Mesh.finalize()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.sim.SDF"><code class="docutils literal notranslate"><span class="pre">SDF</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sim.SDF.volume"><code class="docutils literal notranslate"><span class="pre">SDF.volume</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.SDF.I"><code class="docutils literal notranslate"><span class="pre">SDF.I</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.SDF.mass"><code class="docutils literal notranslate"><span class="pre">SDF.mass</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.SDF.com"><code class="docutils literal notranslate"><span class="pre">SDF.com</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#joint-types">Joint types</a><ul>
-<li><a class="reference internal" href="#warp.sim.JOINT_PRISMATIC"><code class="docutils literal notranslate"><span class="pre">JOINT_PRISMATIC</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_REVOLUTE"><code class="docutils literal notranslate"><span class="pre">JOINT_REVOLUTE</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_BALL"><code class="docutils literal notranslate"><span class="pre">JOINT_BALL</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_FIXED"><code class="docutils literal notranslate"><span class="pre">JOINT_FIXED</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_FREE"><code class="docutils literal notranslate"><span class="pre">JOINT_FREE</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_COMPOUND"><code class="docutils literal notranslate"><span class="pre">JOINT_COMPOUND</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_UNIVERSAL"><code class="docutils literal notranslate"><span class="pre">JOINT_UNIVERSAL</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_DISTANCE"><code class="docutils literal notranslate"><span class="pre">JOINT_DISTANCE</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_D6"><code class="docutils literal notranslate"><span class="pre">JOINT_D6</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#joint-modes">Joint modes</a><ul>
-<li><a class="reference internal" href="#warp.sim.JOINT_MODE_LIMIT"><code class="docutils literal notranslate"><span class="pre">JOINT_MODE_LIMIT</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_MODE_TARGET_POSITION"><code class="docutils literal notranslate"><span class="pre">JOINT_MODE_TARGET_POSITION</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.JOINT_MODE_TARGET_VELOCITY"><code class="docutils literal notranslate"><span class="pre">JOINT_MODE_TARGET_VELOCITY</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#state">State</a><ul>
-<li><a class="reference internal" href="#warp.sim.State"><code class="docutils literal notranslate"><span class="pre">State</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sim.State.particle_count"><code class="docutils literal notranslate"><span class="pre">State.particle_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.body_count"><code class="docutils literal notranslate"><span class="pre">State.body_count</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.particle_q"><code class="docutils literal notranslate"><span class="pre">State.particle_q</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.particle_qd"><code class="docutils literal notranslate"><span class="pre">State.particle_qd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.particle_f"><code class="docutils literal notranslate"><span class="pre">State.particle_f</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.body_q"><code class="docutils literal notranslate"><span class="pre">State.body_q</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.body_qd"><code class="docutils literal notranslate"><span class="pre">State.body_qd</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.State.body_f"><code class="docutils literal notranslate"><span class="pre">State.body_f</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-<li><a class="reference internal" href="#forward-inverse-kinematics">Forward / Inverse Kinematics</a><ul>
-<li><a class="reference internal" href="#warp.sim.eval_fk"><code class="docutils literal notranslate"><span class="pre">eval_fk()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.eval_ik"><code class="docutils literal notranslate"><span class="pre">eval_ik()</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#integrators">Integrators</a><ul>
-<li><a class="reference internal" href="#warp.sim.SemiImplicitIntegrator"><code class="docutils literal notranslate"><span class="pre">SemiImplicitIntegrator</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.XPBDIntegrator"><code class="docutils literal notranslate"><span class="pre">XPBDIntegrator</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#importers">Importers</a><ul>
-<li><a class="reference internal" href="#warp.sim.parse_urdf"><code class="docutils literal notranslate"><span class="pre">parse_urdf()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.parse_mjcf"><code class="docutils literal notranslate"><span class="pre">parse_mjcf()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.parse_usd"><code class="docutils literal notranslate"><span class="pre">parse_usd()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sim.resolve_usd_from_url"><code class="docutils literal notranslate"><span class="pre">resolve_usd_from_url()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    <script async="async" src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/modules/sparse.html b/docs/_build/html/modules/sparse.html
deleted file mode 100644
index 8b4e6efbd..000000000
--- a/docs/_build/html/modules/sparse.html
+++ /dev/null
@@ -1,758 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="../">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><meta name="viewport" content="width=device-width, initial-scale=1" />
-<link rel="index" title="Index" href="../genindex.html" /><link rel="search" title="Search" href="../search.html" /><link rel="next" title="warp.fem" href="fem.html" /><link rel="prev" title="warp.sim" href="sim.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 -->
-        <title>warp.sparse - Warp 1.0.0-beta.5</title>
-      <link rel="stylesheet" type="text/css" href="../_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="../_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="../_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="../_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="../index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="../index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="../_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="../_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="../search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="../installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="../faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul class="current">
-<li class="toctree-l1"><a class="reference internal" href="sim.html">warp.sim</a></li>
-<li class="toctree-l1 current current-page"><a class="current reference internal" href="#">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          <section id="warp-sparse">
-<h1>warp.sparse<a class="headerlink" href="#warp-sparse" title="Link to this heading">#</a></h1>
-<p>Warp includes a sparse linear algebra module <code class="docutils literal notranslate"><span class="pre">warp.sparse</span></code> that implements some common operations for manipulating sparse matrices that arise in simulation.</p>
-<section id="sparse-matrices">
-<h2>Sparse Matrices<a class="headerlink" href="#sparse-matrices" title="Link to this heading">#</a></h2>
-<p>Currently <cite>warp.sparse</cite> supports Block Sparse Row (BSR) matrices, the BSR format can also be used to represent Compressed Sparse Row (CSR) matrices as a special case with a 1x1 block size.</p>
-<dl class="py class" id="module-warp.sparse">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">BsrMatrix</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix" title="Link to this definition">#</a></dt>
-<dd><p>Untyped base class for BSR and CSR matrices.</p>
-<p>Should not be constructed directly but through functions such as <a class="reference internal" href="#warp.sparse.bsr_zeros" title="warp.sparse.bsr_zeros"><code class="xref py py-func docutils literal notranslate"><span class="pre">bsr_zeros()</span></code></a>.</p>
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.nrow">
-<span class="sig-name descname"><span class="pre">nrow</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix.nrow" title="Link to this definition">#</a></dt>
-<dd><p>Number of rows of blocks</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.ncol">
-<span class="sig-name descname"><span class="pre">ncol</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix.ncol" title="Link to this definition">#</a></dt>
-<dd><p>Number of columns of blocks</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.nnz">
-<span class="sig-name descname"><span class="pre">nnz</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix.nnz" title="Link to this definition">#</a></dt>
-<dd><p>Number of non-zero blocks: must be equal to <code class="docutils literal notranslate"><span class="pre">offsets[nrow-1]</span></code>, cached on host for convenience</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.offsets">
-<span class="sig-name descname"><span class="pre">offsets</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix.offsets" title="Link to this definition">#</a></dt>
-<dd><p>Array of size at least <code class="docutils literal notranslate"><span class="pre">1</span> <span class="pre">+</span> <span class="pre">nrows</span></code> such that the start and end indices of the blocks of row <code class="docutils literal notranslate"><span class="pre">r</span></code> are <code class="docutils literal notranslate"><span class="pre">offsets[r]</span></code> and <code class="docutils literal notranslate"><span class="pre">offsets[r+1]</span></code>, respectively.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="functions.html#warp.Array" title="warp.Array">Array</a>[<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.columns">
-<span class="sig-name descname"><span class="pre">columns</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix.columns" title="Link to this definition">#</a></dt>
-<dd><p>Array of size at least equal to <code class="docutils literal notranslate"><span class="pre">nnz</span></code> containing block column indices</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="functions.html#warp.Array" title="warp.Array">Array</a>[<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)">int</a>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py attribute">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.values">
-<span class="sig-name descname"><span class="pre">values</span></span><a class="headerlink" href="#warp.sparse.BsrMatrix.values" title="Link to this definition">#</a></dt>
-<dd><p>Array of size at least equal to <code class="docutils literal notranslate"><span class="pre">nnz</span></code> containing block values</p>
-<dl class="field-list simple">
-<dt class="field-odd">Type<span class="colon">:</span></dt>
-<dd class="field-odd"><p><a class="reference internal" href="functions.html#warp.Array" title="warp.Array">Array</a>[BlockType]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.scalar_type">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">scalar_type</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><span class="pre">Scalar</span></em><a class="headerlink" href="#warp.sparse.BsrMatrix.scalar_type" title="Link to this definition">#</a></dt>
-<dd><p>Scalar type for individual block coefficients. For CSR matrices, this is the same as the block type</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.block_shape">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">block_shape</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><span class="pre">Tuple</span></a><span class="p"><span class="pre">[</span></span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#warp.sparse.BsrMatrix.block_shape" title="Link to this definition">#</a></dt>
-<dd><p>Shape of the individual blocks</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.block_size">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">block_size</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a></em><a class="headerlink" href="#warp.sparse.BsrMatrix.block_size" title="Link to this definition">#</a></dt>
-<dd><p>Size of the individual blocks, i.e. number of rows per block times number of columns per block</p>
-</dd></dl>
-
-<dl class="py property">
-<dt class="sig sig-object py" id="warp.sparse.BsrMatrix.shape">
-<em class="property"><span class="pre">property</span><span class="w"> </span></em><span class="sig-name descname"><span class="pre">shape</span></span><em class="property"><span class="p"><span class="pre">:</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Tuple" title="(in Python v3.12)"><span class="pre">Tuple</span></a><span class="p"><span class="pre">[</span></span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">,</span></span><span class="w"> </span><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><span class="pre">int</span></a><span class="p"><span class="pre">]</span></span></em><a class="headerlink" href="#warp.sparse.BsrMatrix.shape" title="Link to this definition">#</a></dt>
-<dd><p>Shape of the matrix, i.e. number of rows/columns of blocks times number of rows/columsn per block</p>
-</dd></dl>
-
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_zeros">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_zeros</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">rows_of_blocks</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cols_of_blocks</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_type</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_zeros" title="Link to this definition">#</a></dt>
-<dd><p>Constructs and returns an empty BSR or CSR matrix with the given shape</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>bsr</strong> – The BSR or CSR matrix to set to zero</p></li>
-<li><p><strong>rows_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of rows of blocks</p></li>
-<li><p><strong>cols_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of columns of blocks</p></li>
-<li><p><strong>block_type</strong> (<em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em>) – Type of individual blocks. For CSR matrices, this should be a scalar type;
-for BSR matrices, this should be a matrix type (e.g. from <code class="xref py py-func docutils literal notranslate"><span class="pre">warp.mat()</span></code>)</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – Device on which to allocate the matrix arrays</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_set_zero">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_set_zero</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">bsr</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rows_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cols_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_set_zero" title="Link to this definition">#</a></dt>
-<dd><p>Sets a BSR matrix to zero, possibly changing its size</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>bsr</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – The BSR or CSR matrix to set to zero</p></li>
-<li><p><strong>rows_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the new number of rows of blocks</p></li>
-<li><p><strong>cols_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the new number of columns of blocks</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_set_from_triplets">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_set_from_triplets</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dest</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rows</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">columns</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">values</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_set_from_triplets" title="Link to this definition">#</a></dt>
-<dd><p>Fills a BSR matrix with values defined by coordinate-oriented (COO) triplets, discarding existing blocks.</p>
-<p>The first dimension of the three input arrays must match, and determines the number of non-zeros in the constructed matrix.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dest</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – Sparse matrix to populate</p></li>
-<li><p><strong>rows</strong> (<a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – Row index for each non-zero</p></li>
-<li><p><strong>columns</strong> (<a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em>]</em>) – Columns index for each non-zero</p></li>
-<li><p><strong>values</strong> (<a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><em>Union</em><em>[</em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>, </em><em>BlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>]</em><em>]</em><em>]</em>) – Block values for each non-zero. Must be either a one-dimensional array with data type identical
-to the <cite>dest</cite> matrix’s block type, or a 3d array with data type equal to the <cite>dest</cite> matrix’s scalar type.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_assign">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_assign</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dest</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">src</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_assign" title="Link to this definition">#</a></dt>
-<dd><p>Copies the content of the <cite>src</cite> matrix to <cite>dest</cite>, casting the block values if the two matrices use distinct scalar types.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dest</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – </p></li>
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>]</em><em>]</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_copy">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_copy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">scalar_type</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_copy" title="Link to this definition">#</a></dt>
-<dd><p>Returns a copy of matrix <code class="docutils literal notranslate"><span class="pre">A</span></code>, possibly chaning its scalar type.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>scalar_type</strong> (<em>Scalar</em><em> | </em><em>None</em>) – If provided, the returned matrix will use this scalar type instead of the one from <cite>A</cite>.</p></li>
-<li><p><strong>A</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_set_transpose">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_set_transpose</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dest</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">src</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_set_transpose" title="Link to this definition">#</a></dt>
-<dd><p>Assigns the transposed matrix <cite>src</cite> to matrix <cite>dest</cite></p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>dest</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Cols</em><em>, </em><em>Rows</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – </p></li>
-<li><p><strong>src</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – </p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_transposed">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_transposed</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_transposed" title="Link to this definition">#</a></dt>
-<dd><p>Returns a copy of the transposed matrix <cite>A</cite></p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>A</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – </p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_get_diag">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_get_diag</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">out</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_get_diag" title="Link to this definition">#</a></dt>
-<dd><p>Returns the array of blocks that constitute the diagonal of a sparse matrix.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>A</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>BlockType</em><em>]</em>) – the sparse matrix from which to extract the diagonal</p></li>
-<li><p><strong>out</strong> (<em>Optional</em><em>[</em><a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><em>BlockType</em><em>]</em><em>]</em>) – if provided, the array into which to store the diagonal blocks</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="functions.html#warp.Array" title="warp.Array">Array</a>[BlockType]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_set_diag">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_set_diag</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">diag</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rows_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cols_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_set_diag" title="Link to this definition">#</a></dt>
-<dd><p>Sets <cite>A</cite> as a block-diagonal matrix</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>A</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – the sparse matrix to modify</p></li>
-<li><p><strong>diag</strong> (<em>Union</em><em>[</em><em>BlockType</em><em>, </em><a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><em>BlockType</em><em>]</em><em>]</em>) – Either a warp array of type <code class="docutils literal notranslate"><span class="pre">A.values.dtype</span></code>, in which case each element will define one block of the diagonal,
-or a constant value of type <code class="docutils literal notranslate"><span class="pre">A.values.dtype</span></code>, in which case it will get assigned to all diagonal blocks.</p></li>
-<li><p><strong>rows_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the new number of rows of blocks</p></li>
-<li><p><strong>cols_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the new number of columns of blocks</p></li>
-</ul>
-</dd>
-</dl>
-<dl class="simple">
-<dt>The shape of the matrix will be defined one of the following, in that order:</dt><dd><ul class="simple">
-<li><p><cite>rows_of_blocks</cite> and <cite>cols_of_blocks</cite>, if provided. If only one is given, the second is assumed equal.</p></li>
-<li><p>the first dimension of <cite>diag</cite>, if <cite>diag</cite> is an array</p></li>
-<li><p>the current dimensions of <cite>A</cite> otherwise</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_diag">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_diag</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">diag</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rows_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">cols_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_diag" title="Link to this definition">#</a></dt>
-<dd><p>Creates and returns a block-diagonal BSR matrix from an given block value or array of block values.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>diag</strong> (<em>Union</em><em>[</em><em>BlockType</em><em>, </em><a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><em>BlockType</em><em>]</em><em>]</em>) – Either a warp array of type <code class="docutils literal notranslate"><span class="pre">A.values.dtype</span></code>, in which case each element will define one block of the diagonal,
-or a constant value of type <code class="docutils literal notranslate"><span class="pre">A.values.dtype</span></code>, in which case it will get assigned to all diagonal blocks.</p></li>
-<li><p><strong>rows_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the new number of rows of blocks</p></li>
-<li><p><strong>cols_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the new number of columns of blocks</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>[BlockType]</p>
-</dd>
-</dl>
-<dl class="simple">
-<dt>The shape of the matrix will be defined one of the following, in that order:</dt><dd><ul class="simple">
-<li><p><cite>rows_of_blocks</cite> and <cite>cols_of_blocks</cite>, if provided. If only one is given, the second is assumed equal.</p></li>
-<li><p>the first dimension of <cite>diag</cite>, if <cite>diag</cite> is an array</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_set_identity">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_set_identity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">rows_of_blocks</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_set_identity" title="Link to this definition">#</a></dt>
-<dd><p>Sets <cite>A</cite> as the identity matrix</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>A</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – the sparse matrix to modify</p></li>
-<li><p><strong>rows_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a><em> | </em><em>None</em>) – if provided, the matrix will be resized as a square matrix with <cite>rows_of_blocks</cite> rows and columns.</p></li>
-</ul>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_identity">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_identity</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">rows_of_blocks</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">block_type</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_identity" title="Link to this definition">#</a></dt>
-<dd><p>Creates and returns a square identity matrix.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>rows_of_blocks</strong> (<a class="reference external" href="https://docs.python.org/3/library/functions.html#int" title="(in Python v3.12)"><em>int</em></a>) – Number of rows and columns of blocks in the created matrix.</p></li>
-<li><p><strong>block_type</strong> (<em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Rows</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em>) – Block type for the newly created matrix – must be square</p></li>
-<li><p><strong>device</strong> (<em>Device</em><em> | </em><a class="reference external" href="https://docs.python.org/3/library/stdtypes.html#str" title="(in Python v3.12)"><em>str</em></a><em> | </em><em>None</em>) – Device onto which to allocate the data arrays</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>[<em>_MatrixBlockType</em>[<em>Rows</em>, <em>Rows</em>, <em>Scalar</em>] | <em>_ScalarBlockType</em>[<em>Scalar</em>]]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_scale">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_scale</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_scale" title="Link to this definition">#</a></dt>
-<dd><p>Performs the operation <code class="docutils literal notranslate"><span class="pre">x</span> <span class="pre">:=</span> <span class="pre">alpha</span> <span class="pre">*</span> <span class="pre">x</span></code> on BSR matrix <cite>x</cite> and returns <cite>x</cite></p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>) – </p></li>
-<li><p><strong>alpha</strong> (<em>Scalar</em>) – </p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a></p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sparse.bsr_axpy_work_arrays">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_axpy_work_arrays</span></span><a class="headerlink" href="#warp.sparse.bsr_axpy_work_arrays" title="Link to this definition">#</a></dt>
-<dd><p>Opaque structure for persisting <a class="reference internal" href="#warp.sparse.bsr_axpy" title="warp.sparse.bsr_axpy"><code class="xref py py-func docutils literal notranslate"><span class="pre">bsr_axpy()</span></code></a> temporary work buffers across calls</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_axpy">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_axpy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">beta</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">work_arrays</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_axpy" title="Link to this definition">#</a></dt>
-<dd><p>Performs the sparse matrix addition <code class="docutils literal notranslate"><span class="pre">y</span> <span class="pre">:=</span> <span class="pre">alpha</span> <span class="pre">*</span> <span class="pre">X</span> <span class="pre">+</span> <span class="pre">beta</span> <span class="pre">*</span> <span class="pre">y</span></code> on BSR matrices <cite>x</cite> and <cite>y</cite> and returns <cite>y</cite>.</p>
-<p>The <cite>x</cite> and <cite>y</cite> matrices are allowed to alias.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – Read-only right-hand-side.</p></li>
-<li><p><strong>y</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>] </em><em>| </em><em>None</em>) – Mutable left-hand-side. If <cite>y</cite> is not provided, it will be allocated and treated as zero.</p></li>
-<li><p><strong>alpha</strong> (<em>Scalar</em>) – Uniform scaling factor for <cite>x</cite></p></li>
-<li><p><strong>beta</strong> (<em>Scalar</em>) – Uniform scaling factor for <cite>y</cite></p></li>
-<li><p><strong>work_arrays</strong> (<a class="reference internal" href="#warp.sparse.bsr_axpy_work_arrays" title="warp.sparse.bsr_axpy_work_arrays"><em>bsr_axpy_work_arrays</em></a><em> | </em><em>None</em>) – In most cases this function will require the use of temporary storage; this storage can be reused across calls by passing an instance of <a class="reference internal" href="#warp.sparse.bsr_axpy_work_arrays" title="warp.sparse.bsr_axpy_work_arrays"><code class="xref py py-class docutils literal notranslate"><span class="pre">bsr_axpy_work_arrays</span></code></a> in <cite>work_arrays</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>[<em>_MatrixBlockType</em>[<em>Rows</em>, <em>Cols</em>, <em>Scalar</em>] | <em>_ScalarBlockType</em>[<em>Scalar</em>]]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py class">
-<dt class="sig sig-object py" id="warp.sparse.bsr_mm_work_arrays">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_mm_work_arrays</span></span><a class="headerlink" href="#warp.sparse.bsr_mm_work_arrays" title="Link to this definition">#</a></dt>
-<dd><p>Opaque structure for persisting <a class="reference internal" href="#warp.sparse.bsr_mm" title="warp.sparse.bsr_mm"><code class="xref py py-func docutils literal notranslate"><span class="pre">bsr_mm()</span></code></a> temporary work buffers across calls</p>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_mm">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_mm</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">z</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">beta</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">work_arrays</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_mm" title="Link to this definition">#</a></dt>
-<dd><p>Performs the sparse matrix-matrix multiplication <code class="docutils literal notranslate"><span class="pre">z</span> <span class="pre">:=</span> <span class="pre">alpha</span> <span class="pre">*</span> <span class="pre">x</span> <span class="pre">*</span> <span class="pre">y</span> <span class="pre">+</span> <span class="pre">beta</span> <span class="pre">*</span> <span class="pre">z</span></code> on BSR matrices <cite>x</cite>, <cite>y</cite> and <cite>z</cite>, and returns <cite>z</cite>.</p>
-<p>The <cite>x</cite>, <cite>y</cite> and <cite>z</cite> matrices are allowed to alias.
-If the matrix <cite>z</cite> is not provided as input, it will be allocated and treated as zero.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>x</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – Read-only left factor of the matrix-matrix product.</p></li>
-<li><p><strong>y</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><a class="reference external" href="https://docs.python.org/3/library/typing.html#typing.Any" title="(in Python v3.12)"><em>Any</em></a><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – Read-only right factor of the matrix-matrix product.</p></li>
-<li><p><strong>z</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>] </em><em>| </em><em>None</em>) – Mutable left-hand-side. If <cite>z</cite> is not provided, it will be allocated and treated as zero.</p></li>
-<li><p><strong>alpha</strong> (<em>Scalar</em>) – Uniform scaling factor for the <code class="docutils literal notranslate"><span class="pre">x</span> <span class="pre">*</span> <span class="pre">y</span></code> product</p></li>
-<li><p><strong>beta</strong> (<em>Scalar</em>) – Uniform scaling factor for <cite>z</cite></p></li>
-<li><p><strong>work_arrays</strong> (<a class="reference internal" href="#warp.sparse.bsr_mm_work_arrays" title="warp.sparse.bsr_mm_work_arrays"><em>bsr_mm_work_arrays</em></a><em> | </em><em>None</em>) – In most cases this function will require the use of temporary storage; this storage can be reused across calls by passing an instance of <a class="reference internal" href="#warp.sparse.bsr_mm_work_arrays" title="warp.sparse.bsr_mm_work_arrays"><code class="xref py py-class docutils literal notranslate"><span class="pre">bsr_mm_work_arrays</span></code></a> in <cite>work_arrays</cite>.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a>[<em>_MatrixBlockType</em>[<em>Rows</em>, <em>Cols</em>, <em>Scalar</em>] | <em>_ScalarBlockType</em>[<em>Scalar</em>]]</p>
-</dd>
-</dl>
-</dd></dl>
-
-<dl class="py function">
-<dt class="sig sig-object py" id="warp.sparse.bsr_mv">
-<span class="sig-prename descclassname"><span class="pre">warp.sparse.</span></span><span class="sig-name descname"><span class="pre">bsr_mv</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">A</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">x</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">y</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">alpha</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">beta</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">0.0</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">work_buffer</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#warp.sparse.bsr_mv" title="Link to this definition">#</a></dt>
-<dd><p>Performs the sparse matrix-vector product <code class="docutils literal notranslate"><span class="pre">y</span> <span class="pre">:=</span> <span class="pre">alpha</span> <span class="pre">*</span> <span class="pre">A</span> <span class="pre">*</span> <span class="pre">x</span> <span class="pre">+</span> <span class="pre">beta</span> <span class="pre">*</span> <span class="pre">y</span></code> and returns <cite>y</cite>.</p>
-<p>The <cite>x</cite> and <cite>y</cite> vectors are allowed to alias.</p>
-<dl class="field-list simple">
-<dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><ul class="simple">
-<li><p><strong>A</strong> (<a class="reference internal" href="#warp.sparse.BsrMatrix" title="warp.sparse.BsrMatrix"><em>BsrMatrix</em></a><em>[</em><em>_MatrixBlockType</em><em>[</em><em>Rows</em><em>, </em><em>Cols</em><em>, </em><em>Scalar</em><em>] </em><em>| </em><em>_ScalarBlockType</em><em>[</em><em>Scalar</em><em>]</em><em>]</em>) – Read-only, left matrix factor of the matrix-vector product.</p></li>
-<li><p><strong>x</strong> (<a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><a class="reference internal" href="functions.html#warp.Vector" title="warp.Vector"><em>Vector</em></a><em>[</em><em>Cols</em><em>, </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>] </em><em>| </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>]</em>) – Read-only, right vector factor of the matrix-vector product.</p></li>
-<li><p><strong>y</strong> (<a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><a class="reference internal" href="functions.html#warp.Vector" title="warp.Vector"><em>Vector</em></a><em>[</em><em>Rows</em><em>, </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>] </em><em>| </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>] </em><em>| </em><em>None</em>) – Mutable left-hand-side. If <cite>y</cite> is not provided, it will be allocated and treated as zero.</p></li>
-<li><p><strong>alpha</strong> (<em>Scalar</em>) – Uniform scaling factor for <cite>x</cite>. If zero, <cite>x</cite> will not be read and may be left uninitialized.</p></li>
-<li><p><strong>beta</strong> (<em>Scalar</em>) – Uniform scaling factor for <cite>y</cite>. If zero, <cite>y</cite> will not be read and may be left uninitialized.</p></li>
-<li><p><strong>work_buffer</strong> (<a class="reference internal" href="functions.html#warp.Array" title="warp.Array"><em>Array</em></a><em>[</em><a class="reference internal" href="functions.html#warp.Vector" title="warp.Vector"><em>Vector</em></a><em>[</em><em>Rows</em><em>, </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>] </em><em>| </em><a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar"><em>Scalar</em></a><em>] </em><em>| </em><em>None</em>) – Temporary storage is required if and only if <cite>x</cite> and <cite>y</cite> are the same vector. If provided the <cite>work_buffer</cite> array
-will be used for this purpose, otherwise a temporary allocation wil be performed.</p></li>
-</ul>
-</dd>
-<dt class="field-even">Return type<span class="colon">:</span></dt>
-<dd class="field-even"><p><a class="reference internal" href="functions.html#warp.Array" title="warp.Array">Array</a>[<a class="reference internal" href="functions.html#warp.Vector" title="warp.Vector">Vector</a>[Rows, <a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar">Scalar</a>] | <a class="reference internal" href="functions.html#warp.Scalar" title="warp.Scalar">Scalar</a>]</p>
-</dd>
-</dl>
-</dd></dl>
-
-</section>
-</section>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          <a class="next-page" href="fem.html">
-              <div class="page-info">
-                <div class="context">
-                  <span>Next</span>
-                </div>
-                <div class="title">warp.fem</div>
-              </div>
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-            </a>
-          <a class="prev-page" href="sim.html">
-              <svg class="furo-related-icon"><use href="#svg-arrow-right"></use></svg>
-              <div class="page-info">
-                <div class="context">
-                  <span>Previous</span>
-                </div>
-                
-                <div class="title">warp.sim</div>
-                
-              </div>
-            </a>
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer">
-      
-      
-      <div class="toc-sticky toc-scroll">
-        <div class="toc-title-container">
-          <span class="toc-title">
-            On this page
-          </span>
-        </div>
-        <div class="toc-tree-container">
-          <div class="toc-tree">
-            <ul>
-<li><a class="reference internal" href="#">warp.sparse</a><ul>
-<li><a class="reference internal" href="#sparse-matrices">Sparse Matrices</a><ul>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix"><code class="docutils literal notranslate"><span class="pre">BsrMatrix</span></code></a><ul>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.nrow"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.nrow</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.ncol"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.ncol</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.nnz"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.nnz</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.offsets"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.offsets</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.columns"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.columns</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.values"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.values</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.scalar_type"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.scalar_type</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.block_shape"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.block_shape</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.block_size"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.block_size</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.BsrMatrix.shape"><code class="docutils literal notranslate"><span class="pre">BsrMatrix.shape</span></code></a></li>
-</ul>
-</li>
-<li><a class="reference internal" href="#warp.sparse.bsr_zeros"><code class="docutils literal notranslate"><span class="pre">bsr_zeros()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_set_zero"><code class="docutils literal notranslate"><span class="pre">bsr_set_zero()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_set_from_triplets"><code class="docutils literal notranslate"><span class="pre">bsr_set_from_triplets()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_assign"><code class="docutils literal notranslate"><span class="pre">bsr_assign()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_copy"><code class="docutils literal notranslate"><span class="pre">bsr_copy()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_set_transpose"><code class="docutils literal notranslate"><span class="pre">bsr_set_transpose()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_transposed"><code class="docutils literal notranslate"><span class="pre">bsr_transposed()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_get_diag"><code class="docutils literal notranslate"><span class="pre">bsr_get_diag()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_set_diag"><code class="docutils literal notranslate"><span class="pre">bsr_set_diag()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_diag"><code class="docutils literal notranslate"><span class="pre">bsr_diag()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_set_identity"><code class="docutils literal notranslate"><span class="pre">bsr_set_identity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_identity"><code class="docutils literal notranslate"><span class="pre">bsr_identity()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_scale"><code class="docutils literal notranslate"><span class="pre">bsr_scale()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_axpy_work_arrays"><code class="docutils literal notranslate"><span class="pre">bsr_axpy_work_arrays</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_axpy"><code class="docutils literal notranslate"><span class="pre">bsr_axpy()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_mm_work_arrays"><code class="docutils literal notranslate"><span class="pre">bsr_mm_work_arrays</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_mm"><code class="docutils literal notranslate"><span class="pre">bsr_mm()</span></code></a></li>
-<li><a class="reference internal" href="#warp.sparse.bsr_mv"><code class="docutils literal notranslate"><span class="pre">bsr_mv()</span></code></a></li>
-</ul>
-</li>
-</ul>
-</li>
-</ul>
-
-          </div>
-        </div>
-      </div>
-      
-      
-    </aside>
-  </div>
-</div><script src="../_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="../_static/doctools.js?v=888ff710"></script>
-    <script src="../_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="../_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="../_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="../_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/objects.inv b/docs/_build/html/objects.inv
deleted file mode 100644
index 26f63c09488b3d0116eade8b8475df096048f880..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 5391
zcmV+q74YgKAX9K?X>NERX>N99Zgg*Qc_4OWa&u{KZXhxWBOp+6Z)#;@bUGkcVRCQ^
zBOq2~a&u{KZaN?_E-)@IEn;PKVJ<ZaBOp|0Wgv28ZDDC{WMy(7Z)PBLXlZjGW@&6?
zAZc?TV{dJ6a%FRKWn>_Ab7^j8AbM<RVlE1J+O3^QbK|zQ$M^aau6l3Vx;JfC>Q#2`
zw!5RmQ(H>E+!YHVK?ycV&;Te~`Sce9Ns;0po&znCYTNkz0UXSS7~YA<=hs4z{YAe2
zElh}%7r$JjJSq|@{}mTms2I=We5=yrfByI4)9l><Uv}HR;G)PBOKCUoJfb@mQo|6X
z1X~x1PU2VuX9SIzGFVF0mPaR<ZwV&~lKbC@+<q4>Y12$l5%YDyk_eEt8pIf|hM`tm
zWFULo6D|A!BB%sZAQKv801HAhZy{(M`cV)?MW(X~F~c$>2@x=Y=QLXB1c>SOYd~|F
zkX%xrNmeAM)1xz?=W7za>YpHW-ExA4TttD%KOB`*L7{)B;1EWTOvYTKL;=hCCk+(`
zmGw`a6e+;&F8uuK-Q)6hxq_8&$=)p<Zm*wFC`l3?>P7%C<`EgJ_km(MI6|Ttz+l*h
zDr}4f-SH2a1d{!O$h01FbS)P=4YDH5_g(_KXbs@W+Ge+^brUKg2{{;vA}Ap?DfvJ^
zECWLWvz-BZnc#CBl5H}GnJpG_{kKiQ3kgSaCBRV3k|cmriEi!`p1X}8wwQCF2N<FD
z6S0Iz1&#~aOdKWXTfy|e5FnGP(QP^Sf#Ovhj+mh(=Mo;-jtGg!+m6w9a9n1*1XVBz
zZ=Z@_kk4R-rz{K*J(U9yy;TV!EMh@ftbyQ~seoh=f^t(8AoKyJI8PvRqP3vpHRsGN
zP#_WQw!vFL#XcZpy+*?s{lLdu4m$ugB)KYto*B{2V5RAAcNIwoOMje8h$gl$bS<E^
z2`NjUNJ7#a%<zoD@{;I|SgOEQwSYz&YX1j~SN}@)OltaML$B2Ku!MrAW}4VS5y2%m
znnENQXA*3)_aI_BIFuI9niNGGT3;i&5tK?WrKr%oJ^}Yy5hEKg0ySQg<a)7g$8;O=
z;@>r*D?hc93bJMiQ+r3czkWIrC_OZ8O>)reNl3ryfSlrI5mWp8j7hU<Oa>uObE4Qf
zp=Z!LPfpL3M?3F5Q<|g_EYg$k^Qif%fLlFfn@cE?<_gx4TOvr<D{P>mAJD;26EswM
zCPSc;wV0;!FM2{Rm)iQhjwusjjG1pJ*9!-+A6TxO<TD;oJt#|N)7QHWIjK>v=r9RT
zJm(4D05iagEFxmRG_7E|-M-Q!0R+DpD$f4UOol=aOy3wyB0wZfOt+;t*pZ~55@e6I
zNJ?I55K-e*rBTqVArs5CEvMJfA%)O9z9)I!?@%8?Qo^$7;cpX~(oETEgrnwn*WaJ7
zAHLrX`DQ0bJUl*peSG-#_<X;3_#Qkx-u>s{@qTfA=OMYgeZGCTSv)OP{{gCe<OvIz
z`d*gy6A_9Hy3~6mfECvUzbm`fiHI=XJ1T(9_FLHmfxTo4D`%RZR&^wx$BImr7R73q
zc?qmw+XATWzAa&Uhqx;1iEX+qf$Fo5B!HdD_eRF26@Nv{*X3XlNL}sp4711-B4WL?
zP7*xr6*shdY;#o)^C!z9v<9yBABj+x5KlYQGOuNZz{x>$jve$&)#{+<Rlg20aoNY7
zD_J-hy*&AR1F)_Re+dZxJN-xpc0BY0cg#%BJ<9)qM<n@CkVp{Y&YKQCu@HsvWS{Yr
zkp#ocZ-4x?SUz2Uy}b$S=j$8TpwqAP_;mY#*7<(@*K!%$-G0~eo9A1!%-!Rc>(%Np
z@R2j+VX6<LhGNR_!035>)q_v&tQxf9iX^4$6xtM@>841Cz@&dk(ma8LhfV92uk?Mn
zPfd$t;YyD#TtL?ib}oTv&24UO9UP2{z^;a&7CsU?hX#Mo^vdlVO49u3l=NuHLMowJ
z&(wIELWF@=jcFnRYtK2RH}AGcHCga6K`n-wezzs?g<4BM=gyvuY(>+Y3*&839MXNP
zsDcgBE5V@MGI0e9a@3F=6=VZr)?5hw9;nI`UhtSxvR@%2;g%i>V!C~fz{6IZ$+h=b
zA)<=7@0b&7-?4zuEs{uzS2(JMpL+12y4w^jL;@-U=TNiT6z%30Mu)#X4G2^oi0R1i
zx2K`-0oQ$`fKhH)HQ-bowQrm0Y94u9N3~qouEv1ZXjH4sbTx-o)}x|F&?(jabvGb}
zmOwW)h&?)Xunr*S!$S@5j<@3y&~NC-m5vk#3cU<Kk?#KkZu(K{wV46o9)IX~6|KV$
z2=*CwRI!6~gvFt#N6LOeBptXDmC}FgE9D8|2db-TO24-bOwm1H4dH4|clbxw*Ep%s
z_KOM@1)rmiM`T>l1s>G9j5;2ZlX{o^dRym7>3UQ~1*_0~ig~^EnD9+vG?rn0JYH;k
z(0EQdl}2M5qWj`}D?JvbCiQtQMk8tfecH-R)4S946B8>nSOb)s$D73VOzFpE>+ntC
za;x#=35@{)%rxJEFU9bFxM2!D?P6I5CX}8-j_f}jqZ9>GLN7R>9N2mZ=dT3}v7fL)
zULJFChnwuYTvlHQVN!YRUkujfGwd03kP4zb9>ZYg>&ct_5Jtc&4QAL!P@;5nPsj8c
zvUFxbnLwCD9Ei-!$}d46>>nCLOj~oCP!;G!_e*_hZL*Txd8MKHA21r+FCA)e?^lS?
z{<{IR6eYYymc|MaJ5C02FBMHe3U3Dwdd|Y2@#DZpnenW(nEq%Q{~Y{aG#be<but=o
z3=r{=ZwI4sbh$=UQa6N%#-vD8!zCswr_SWkwVl<#PN4B=WfSQLHU@y+y?NvAkOR;W
zHe0(=>w4|mAvM1Ktiz9~PWYZO?jsoFM`QJgaawHc;eBX3(D-7z#dKmg%;|@|D2?Z%
z4Rz$3%kJTCDyO3uSa8e3F3kDT{J+ZWbPq{W!fjNpj6S3|n%dzMb>gh{5)FMaI&GDj
zX=s~$R&oyV?sdI-P6Xm-Xge3oH5kAA3OS8sk;~6u<ku@80zOw}8Te6%Sdb8k6iEWw
zv26hOK5uj09B~?COZ39lI$_-Bcov}r>=`yc8gqP(MSAdK3L3M{EUYeT09KPV0ISCu
z_&;EHDx{fH>R#0R2R6aZ@E8hONH6XzEb{!2rdx-HUItIro;cdk5K`Pv#y=rGDL|aQ
zOlY=I(1U#r3gjD@6VPKPddPzlzx*4{d<G-$R0MW=B$Sj}7DJm#O4R4i5u_imL<}aj
zK;rT;0uwQqfTrr|DgqNRn1H5g_^@ez<4o%tvOqNjOc*yd-H{}S-8QN*H_%CqQ*Mxx
z0y~=U5I%HUfelqi%<P(?A6OW3w?Ga1WEg!p!4Zv_V>Kk=La6QG<a@@0>fRfi9@?G_
z{7k8^8Xk?JBlT|f=VcpngVUM{yd?Zy2>2BuX#yRv_1~cc_y~SRp_2m4l>!;hs%zdr
zROt~1a-$yX0Vc;}OV>mx!QFEXzAz%^?@&U(8@**>AoM1n2eaP_qTDKjd@vc1Y?Dy5
z1Zr5fIC53%Sk$`CqJpcD52U9_6E}*^Q8|@}b%q7?+RLwawE((sK@--u*2ST~m||#5
zpn0y@G)gjPZ(+Xo>uSOjf{PIhLoiTVBf2wzSWgPfdQ#d`FF^=4UJ06S*@#wB(uFSQ
zAS~8Cc*NwZ2fn2#llhhkdKQsD79t+m3^74WSC@?k?wy&lWkl*e1D+C^&=6dU;nRD`
z()k~ci-%S4<@(NEVh|BU8lVX<gnWIxe|r4u;RcA^{0apY%hmP6*IOX=ZSnUlG;QCW
zZ=uoLKi=F1cZ>VQ3IndLpTFO(f~UvjVzqdD@R0a<d-wRYyh8#Y@$|e{-e0d4UxD27
z?a#-%zgAFd{q?Z;dHcM)2GljPR=ob`zx3k-GAAY1carQBo3K<j-`B{Kr+&Qktolh5
z)#rJPL|PuT!mx^n(glU5Jqn-_0H%NFUj5<F^94MHdB{@)6|<~fPj`+|jc|=qkiZyI
zA~Bfx0EUQB*R}*;d=q?bHerNC;*M$6$1H)7wpFCAJ1CVkkM;q)XqzlD{g!5#F+1_#
zcW2Q+RGNc;7o3rlqTx3@8KF_TWd@aDb1^yz6~4|`FsNj#S5`uL7ArFTsK^*;uY_Gw
z=KwBGr!t%=tQ6V~md16F2j1-XIyg(95@XlLk8vA8cBKyt8s(;5U*pIiPe_K!?2GmP
z{~d%raE^ybs661z?sT>HmL_`LA9|R8nA64OCR?O}GH0{Fv!p%=pHzV7*~fGcvD8xm
zK2_~VpvQfu5pK(usgStYZ-?*2UfP$%;Bd3e*HeOKWSvkCJK$DcJqpxU0kQh(iXhfs
zZ4rAI;vC%#c5>wCoCxYuFWx**-<J6n`LMpRV-~_)-m67m7sE}7cP71?gcI>>5@T0t
zj?rl$!^C5rm=&q+-FB#<&Ym|gv6E;^$8I>9Bogwoh#>7yE9hD4#?KlX!6GjKpXFvP
z^6HlcT&p_Zu@~;0`;>eH=0j!SFVoz;g%>Kk@Hao&uPWf}*m-6PEWkD_s#o)=?(kXW
zuOoPpnDu$wIK1SisJJj)m)T+Iq3Mn+84n7cHs{HiNlOs3L{XuqI84ZM=$RFAel{U<
z@a`O9_{v!+m7XGBGa<(vhi@nkqzF9(ocp4-%Ndl;VwcoxV__umkf<$$mr&L7tSAW{
z9~_FfQMD~W{Uu^=34C9AB=a$F58rf&!aEhlPsvn)?sv^5!;MuF?%g@G0=5@$d_>Z`
z+$4fj@bLJS3fAX)I5mV+(d7}x8O6=wq#AlKc#KlidO_Ka)Q`YGBQ_G~GF0x20oMEm
z_v$T`J*FU$u?yyHJS~AYo`!vlr*S{yX#uqH^lS#}E!OC+zDorNe<@fJ^<RN|nleu@
z*HcYMJ8SWW6ksL7gsW}6WcE|3F8;!*R8Kkj@X|-%r8*VkB@nEMJA3Y@V!X4}*EBz+
z=+`r7F!N+)Nd1*!_Ce2(skpoa*H1-{meSnD8~RHLy5mWq&XA}gW9BFU@v215G3ldF
zIUsyAswMS@zeq@OS)3y^VfpH;r^-GtCmMynl5e-^RO&D1elpT+SoajEw@RaR%<{r_
zRI(!8*3GMZ<5&efF%)D=Bf&E-nZ}6#Uk&sqX&fB79a7heyZ#C-hS_=8<joym$NNSm
z>MU=*O`o^vz|j4e1_AHcGZ5VKZYVSIK)=H-Y04I<yBl{=Uhc^Ov)N^}N(2N)Z%Ay=
zF`9hClMx2>`6_m#5PQ*Yel#%ie?NV>!Bi-BL&DIRqY;L|7*?k(pgNd1*%ChWD4uFY
z{qcR|>uAtEqtP5e4f)hAWP8pujOrc2cckuz?r-j@+&3<7ZV0ruT#qZ*u>BI91@>vB
z3h4;8I{fhh)gilhyR#b?uwfe#x+UomGkpnTP(5{6AtDqeBM>F1<jKx*?hy6x!zO(K
z@Ka(L^im}GN}&hdWy-nQN=2ckm8CUe@7-DMPyIJ$m4Evp#eIEBxzqPG_(5>KPI&kl
zj6I1B(kUldXqnK=FGk#wlSHy`#JN9BA^XQuK)6d2@CrPR^|aH|5;7!ayKlAU2wg1x
z_2ckLI5anFDFR)yhYyIJgg*vvADDs$?_9hF_9ZF_ZD&R!<ml)}E_^(Np13da6n^4<
zic@%+qCsDCr#ocSrmavc>t}6)+p$o$j65k@JyOTB<cx&<if%rVcxa4VxN{1xqmPgR
zCT^$+m3ak$L@$Wzsw#TTmEtK94t0{H4;w@U_*$z<AmZ3OzmSlx8Oe}%<%B??W?TW?
zwbDjoJi5i8is#6zx(|n_tU5Htqp+;Lfp3n*%$W=56?t+OLGOuD;5w~|Ef~q^d{s`Z
zIt#wmRgyNm4O+3_tDJRdmM>xAuIaUW>!RLT5q`sMOHA=INSR6QxsqkIn6#fpb!%7;
zP7Ef=ohMOJ2};u7tYYUdyr+hMUr|%<^0TvCdovmQ{;al${VDO&LEh=s5I|k*cHq+@
zQF{J}PXOk!&I7$0eJkh`?qz$>*jQ6BirRvyk*V}Br~d$D?E4R2cH4=gi3!<4e<yPL
zz5En9fv<zNL<XI@U4f=k*TFyQ&r1{VGH^?velU6h@4~k5P+L`x<7%Qp9lR7Nz^@Jz
z;AaO4@Vf&A^x=Ub1Fr`P#S%GRhPtDVR@z8nq6M2oDW$uS&ud6>4G=+r7z%8m0D}Sv
z6p&D$0Nb?y=b`{-qk!h)5}J)mX!W>+R*y?)^|*u@_7ZB>OR!nN)#DOeJubo3;}Tpw
zuAtetf>w_!X!W>)R*x&FVXvTey#kvRTs^M9)#D0WJ+8pj13pJ}n_k?z0%pnd&P5WP
zL&LnaiPLef&<3#=&AjLxjxO&PjlF35gt4Cl^P;htH_t|*6K-MV10$2|Dyy!wn5Q**
zH4rs_%&Xu3>WU>mPpz5PG(&kIRrFiLLo;=Czp)`MzSw`y^sGN-n?kr1Ya?vy4-R>D
zF<dOEQY_oZSwz>xW@9#bHo*Fq4!Ws+oH_rPOHe}AW&_J1_2nO{)Ze%tXY|8X(N7@>
zxAap=GP1D`h_*!<`O@sIV}85%x+R%So!ZiiS1K|c+6rCh0<(SReioCrX5#gaF4^C%
zf6U_7PwcfnxsBZ%-?-2P6kYu5Vo|z0{$v@o+vDYILo-*+S}N6VT~rJ6&)FeZnUqF|
zho6g^#r4Ji>w0i;!^6T@z?O@q^qMV6I;GeAeHF{$H?tS=@u~D>_2+NM(e=mm-!u1F
z$?WivCyZl5esUrII8dd6U8PP^a3TLVP&0d>PkHLAk?qE>U9RsAznJMF>5uKG<=Bt>
zrt;~(_Y_{f|5W|K#c}UE(5@Jd>T@%3eO0}#=#6(R`<Uvv%Zuu{%UP#~>BU;z&&8jv
z4LsAB3Yvv<HdNewTD?``VsZ4_*^yEEl$!RgJC$ytbS3DctGaPf?ZTS1KZo`*hYzI}
zjSsA|>Q(Q7YS?(M^|1;$Yn1<qit)5pTiq;M^&{U~|1z%iJ?&Qpo|)GVCRODh^@9v@
tCMhwF?H1;K=Tf_Le}B?F?Cfy#d`c{<Dzu#GDd{BSr|o{_?7t=fsYi=3Tdx2B

diff --git a/docs/_build/html/py-modindex.html b/docs/_build/html/py-modindex.html
deleted file mode 100644
index 9420f97f6..000000000
--- a/docs/_build/html/py-modindex.html
+++ /dev/null
@@ -1,322 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="search.html" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 --><title>Python Module Index - Warp 1.0.0-beta.5</title>
-<link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon no-toc" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="search.html" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon no-toc" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          
-<section class="domainindex-section">
-  <h1>Python Module Index</h1>
-  <div class="domainindex-jumpbox"><a href="#cap-w"><strong>w</strong></a></div>
-</section>
-<table class="domainindex-table">
-  <tr class="pcap">
-    <td></td><td>&#160;</td><td></td>
-  </tr>
-  <tr class="cap" id="cap-w">
-    <td></td><td><strong>w</strong></td><td></td>
-  </tr>
-  <tr>
-    <td><img src="_static/minus.png" class="toggler"
-              id="toggle-1" style="display: none" alt="-" /></td>
-    <td>
-        <code class="xref">warp</code></td><td>
-    <em></em></td>
-  </tr>
-  <tr class="cg-1">
-    <td></td>
-    <td>&#160;&#160;&#160;
-        <a href="modules/interoperability.html#module-warp.dlpack"><code class="xref">warp.dlpack</code></a></td><td>
-    <em></em></td>
-  </tr>
-  <tr class="cg-1">
-    <td></td>
-    <td>&#160;&#160;&#160;
-        <a href="modules/interoperability.html#module-warp.jax"><code class="xref">warp.jax</code></a></td><td>
-    <em></em></td>
-  </tr>
-  <tr class="cg-1">
-    <td></td>
-    <td>&#160;&#160;&#160;
-        <a href="modules/sparse.html#module-warp.sparse"><code class="xref">warp.sparse</code></a></td><td>
-    <em></em></td>
-  </tr>
-  <tr class="cg-1">
-    <td></td>
-    <td>&#160;&#160;&#160;
-        <a href="modules/interoperability.html#module-warp.torch"><code class="xref">warp.torch</code></a></td><td>
-    <em></em></td>
-  </tr>
-</table>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          
-          
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer no-toc">
-      
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    </body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/search.html b/docs/_build/html/search.html
deleted file mode 100644
index 1c6b91a7b..000000000
--- a/docs/_build/html/search.html
+++ /dev/null
@@ -1,291 +0,0 @@
-<!doctype html>
-<html class="no-js" lang="en" data-content_root="./">
-  <head><meta charset="utf-8"/>
-    <meta name="viewport" content="width=device-width,initial-scale=1"/>
-    <meta name="color-scheme" content="light dark"><link rel="index" title="Index" href="genindex.html" /><link rel="search" title="Search" href="#" />
-
-    <!-- Generated with Sphinx 7.2.6 and Furo 2023.08.19 --><title>Search - Warp 1.0.0-beta.5</title><link rel="stylesheet" type="text/css" href="_static/pygments.css?v=a746c00c" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo.css?v=135e06be" />
-    <link rel="stylesheet" type="text/css" href="_static/copybutton.css?v=76b2166b" />
-    <link rel="stylesheet" type="text/css" href="_static/styles/furo-extensions.css?v=36a5483c" />
-    <link rel="stylesheet" type="text/css" href="_static/custom.css?v=44d7c754" />
-    
-    
-
-
-<style>
-  body {
-    --color-code-background: #f8f8f8;
-  --color-code-foreground: black;
-  --admonition-title-font-size: 100%;
-  --admonition-font-size: 100%;
-  --color-api-pre-name: #4e9a06;
-  --color-api-name: #4e9a06;
-  --color-admonition-title--seealso: #ffffff;
-  --color-admonition-title-background--seealso: #448aff;
-  --color-admonition-title-background--note: #76b900;
-  --color-admonition-title--note: #ffffff;
-  
-  }
-  @media not print {
-    body[data-theme="dark"] {
-      --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-    }
-    @media (prefers-color-scheme: dark) {
-      body:not([data-theme="light"]) {
-        --color-code-background: #202020;
-  --color-code-foreground: #d0d0d0;
-  --color-admonition-title-background--note: #535353;
-  --color-admonition-title--note: #ffffff;
-  
-      }
-    }
-  }
-</style></head>
-  <body>
-    
-    <script>
-      document.body.dataset.theme = localStorage.getItem("theme") || "auto";
-    </script>
-    
-
-<svg xmlns="http://www.w3.org/2000/svg" style="display: none;">
-  <symbol id="svg-toc" viewBox="0 0 24 24">
-    <title>Contents</title>
-    <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 1024 1024">
-      <path d="M408 442h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8zm-8 204c0 4.4 3.6 8 8 8h480c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8H408c-4.4 0-8 3.6-8 8v56zm504-486H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zm0 632H120c-4.4 0-8 3.6-8 8v56c0 4.4 3.6 8 8 8h784c4.4 0 8-3.6 8-8v-56c0-4.4-3.6-8-8-8zM115.4 518.9L271.7 642c5.8 4.6 14.4.5 14.4-6.9V388.9c0-7.4-8.5-11.5-14.4-6.9L115.4 505.1a8.74 8.74 0 0 0 0 13.8z"/>
-    </svg>
-  </symbol>
-  <symbol id="svg-menu" viewBox="0 0 24 24">
-    <title>Menu</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-menu">
-      <line x1="3" y1="12" x2="21" y2="12"></line>
-      <line x1="3" y1="6" x2="21" y2="6"></line>
-      <line x1="3" y1="18" x2="21" y2="18"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-arrow-right" viewBox="0 0 24 24">
-    <title>Expand</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="feather-chevron-right">
-      <polyline points="9 18 15 12 9 6"></polyline>
-    </svg>
-  </symbol>
-  <symbol id="svg-sun" viewBox="0 0 24 24">
-    <title>Light mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="feather-sun">
-      <circle cx="12" cy="12" r="5"></circle>
-      <line x1="12" y1="1" x2="12" y2="3"></line>
-      <line x1="12" y1="21" x2="12" y2="23"></line>
-      <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line>
-      <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line>
-      <line x1="1" y1="12" x2="3" y2="12"></line>
-      <line x1="21" y1="12" x2="23" y2="12"></line>
-      <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line>
-      <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line>
-    </svg>
-  </symbol>
-  <symbol id="svg-moon" viewBox="0 0 24 24">
-    <title>Dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-moon">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none" />
-      <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z" />
-    </svg>
-  </symbol>
-  <symbol id="svg-sun-half" viewBox="0 0 24 24">
-    <title>Auto light/dark mode</title>
-    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor"
-      stroke-width="1.5" stroke-linecap="round" stroke-linejoin="round" class="icon-tabler-shadow">
-      <path stroke="none" d="M0 0h24v24H0z" fill="none"/>
-      <circle cx="12" cy="12" r="9" />
-      <path d="M13 12h5" />
-      <path d="M13 15h4" />
-      <path d="M13 18h1" />
-      <path d="M13 9h4" />
-      <path d="M13 6h1" />
-    </svg>
-  </symbol>
-</svg>
-
-<input type="checkbox" class="sidebar-toggle" name="__navigation" id="__navigation">
-<input type="checkbox" class="sidebar-toggle" name="__toc" id="__toc">
-<label class="overlay sidebar-overlay" for="__navigation">
-  <div class="visually-hidden">Hide navigation sidebar</div>
-</label>
-<label class="overlay toc-overlay" for="__toc">
-  <div class="visually-hidden">Hide table of contents sidebar</div>
-</label>
-
-
-
-<div class="page">
-  <header class="mobile-header">
-    <div class="header-left">
-      <label class="nav-overlay-icon" for="__navigation">
-        <div class="visually-hidden">Toggle site navigation sidebar</div>
-        <i class="icon"><svg><use href="#svg-menu"></use></svg></i>
-      </label>
-    </div>
-    <div class="header-center">
-      <a href="index.html"><div class="brand">Warp 1.0.0-beta.5</div></a>
-    </div>
-    <div class="header-right">
-      <div class="theme-toggle-container theme-toggle-header">
-        <button class="theme-toggle">
-          <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-          <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-          <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-          <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-        </button>
-      </div>
-      <label class="toc-overlay-icon toc-header-icon no-toc" for="__toc">
-        <div class="visually-hidden">Toggle table of contents sidebar</div>
-        <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-      </label>
-    </div>
-  </header>
-  <aside class="sidebar-drawer">
-    <div class="sidebar-container">
-      
-      <div class="sidebar-sticky"><a class="sidebar-brand" href="index.html">
-  
-  <div class="sidebar-logo-container">
-    <img class="sidebar-logo only-light" src="_static/logo-light-mode.png" alt="Light Logo"/>
-    <img class="sidebar-logo only-dark" src="_static/logo-dark-mode.png" alt="Dark Logo"/>
-  </div>
-  
-  <span class="sidebar-brand-text">Warp 1.0.0-beta.5</span>
-  
-</a><form class="sidebar-search-container" method="get" action="#" role="search">
-  <input class="sidebar-search" placeholder="Search" name="q" aria-label="Search">
-  <input type="hidden" name="check_keywords" value="yes">
-  <input type="hidden" name="area" value="default">
-</form>
-<div id="searchbox"></div><div class="sidebar-scroll"><div class="sidebar-tree">
-  <p class="caption" role="heading"><span class="caption-text">User's Guide</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="installation.html">Installation</a></li>
-<li class="toctree-l1"><a class="reference internal" href="basics.html">Basics</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/devices.html">Devices</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/interoperability.html">Interoperability</a></li>
-<li class="toctree-l1"><a class="reference internal" href="configuration.html">Runtime Settings</a></li>
-<li class="toctree-l1"><a class="reference internal" href="debugging.html">Debugging</a></li>
-<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
-<li class="toctree-l1"><a class="reference internal" href="faq.html">FAQ</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Core Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/runtime.html">Runtime Reference</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/functions.html">Kernel Reference</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Simulation Reference</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference internal" href="modules/sim.html">warp.sim</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/sparse.html">warp.sparse</a></li>
-<li class="toctree-l1"><a class="reference internal" href="modules/fem.html">warp.fem</a></li>
-</ul>
-<p class="caption" role="heading"><span class="caption-text">Project Links</span></p>
-<ul>
-<li class="toctree-l1"><a class="reference external" href="https://github.com/NVIDIA/warp">GitHub</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://pypi.org/project/warp-lang">PyPI</a></li>
-<li class="toctree-l1"><a class="reference external" href="https://discord.com/channels/827959428476174346/953756751977648148">Discord</a></li>
-</ul>
-
-</div>
-</div>
-
-      </div>
-      
-    </div>
-  </aside>
-  <div class="main">
-    <div class="content">
-      <div class="article-container">
-        <a href="#" class="back-to-top muted-link">
-          <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24">
-            <path d="M13 20h-2V8l-5.5 5.5-1.42-1.42L12 4.16l7.92 7.92-1.42 1.42L13 8v12z"></path>
-          </svg>
-          <span>Back to top</span>
-        </a>
-        <div class="content-icon-container">
-          <div class="theme-toggle-container theme-toggle-content">
-            <button class="theme-toggle">
-              <div class="visually-hidden">Toggle Light / Dark / Auto color theme</div>
-              <svg class="theme-icon-when-auto"><use href="#svg-sun-half"></use></svg>
-              <svg class="theme-icon-when-dark"><use href="#svg-moon"></use></svg>
-              <svg class="theme-icon-when-light"><use href="#svg-sun"></use></svg>
-            </button>
-          </div>
-          <label class="toc-overlay-icon toc-content-icon no-toc" for="__toc">
-            <div class="visually-hidden">Toggle table of contents sidebar</div>
-            <i class="icon"><svg><use href="#svg-toc"></use></svg></i>
-          </label>
-        </div>
-        <article role="main">
-          
-<noscript>
-<div class="admonition error">
-  <p class="admonition-title">Error</p>
-  <p>
-    Please activate JavaScript to enable the search functionality.
-  </p>
-</div>
-</noscript>
-
-<div id="search-results"></div>
-
-        </article>
-      </div>
-      <footer>
-        
-        <div class="related-pages">
-          
-          
-        </div>
-        <div class="bottom-of-page">
-          <div class="left-details">
-            <div class="copyright">
-                Copyright &#169; 2023, NVIDIA
-            </div>
-            Made with 
-            <a href="https://github.com/pradyunsg/furo">Furo</a>
-            
-          </div>
-          <div class="right-details">
-            <div class="icons">
-              <a class="muted-link " href="https://github.com/NVIDIA/warp" aria-label="GitHub">
-            <svg stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 16 16">
-                <path fill-rule="evenodd" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27.68 0 1.36.09 2 .27 1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.013 8.013 0 0 0 16 8c0-4.42-3.58-8-8-8z"></path>
-            </svg>
-        </a>
-              
-            </div>
-          </div>
-        </div>
-        
-      </footer>
-    </div>
-    <aside class="toc-drawer no-toc">
-      
-      
-      
-    </aside>
-  </div>
-</div><script src="_static/documentation_options.js?v=f0dcd5a7"></script>
-    <script src="_static/doctools.js?v=888ff710"></script>
-    <script src="_static/sphinx_highlight.js?v=dc90522c"></script>
-    <script src="_static/scripts/furo.js?v=32e29ea5"></script>
-    <script src="_static/clipboard.min.js?v=a7894cd8"></script>
-    <script src="_static/copybutton.js?v=fd10adb8"></script>
-    
-<script src="_static/searchtools.js"></script>
-<script src="_static/language_data.js"></script>
-<script src="searchindex.js"></script></body>
-</html>
\ No newline at end of file
diff --git a/docs/_build/html/searchindex.js b/docs/_build/html/searchindex.js
deleted file mode 100644
index 755aa80a0..000000000
--- a/docs/_build/html/searchindex.js
+++ /dev/null
@@ -1 +0,0 @@
-Search.setIndex({"docnames": ["basics", "configuration", "debugging", "faq", "index", "installation", "limitations", "modules/devices", "modules/fem", "modules/functions", "modules/interoperability", "modules/runtime", "modules/sim", "modules/sparse"], "filenames": ["basics.rst", "configuration.rst", "debugging.rst", "faq.rst", "index.rst", "installation.rst", "limitations.rst", "modules/devices.rst", "modules/fem.rst", "modules/functions.rst", "modules/interoperability.rst", "modules/runtime.rst", "modules/sim.rst", "modules/sparse.rst"], "titles": ["Basics", "Runtime Settings", "Debugging", "FAQ", "NVIDIA Warp Documentation", "Installation", "Limitations", "Devices", "warp.fem", "Kernel Reference", "Interoperability", "Runtime Reference", "warp.sim", "warp.sparse"], "terms": {"befor": [0, 1, 4, 8, 9, 11], "us": [0, 1, 2, 3, 4, 5, 6, 8, 9, 12, 13], "warp": [0, 1, 2, 5, 6, 7, 9, 11], "should": [0, 2, 4, 5, 7, 8, 9, 10, 11, 12, 13], "explicitli": [0, 7, 8, 11], "wp": [0, 1, 2, 4, 5, 6, 8, 9, 10, 11, 12], "init": [0, 1, 2, 4, 7, 10, 11], "method": [0, 2, 4, 6, 8, 9, 10, 11, 12], "follow": [0, 1, 3, 4, 5, 7, 8, 10, 11, 12, 13], "import": [0, 1, 4, 7, 10, 11], "print": [0, 1, 4, 7, 9, 10, 11, 12], "some": [0, 1, 3, 4, 5, 6, 9, 11, 13], "startup": [0, 2], "inform": [0, 1, 8, 12], "about": [0, 8, 12], "comput": [0, 3, 4, 7, 8, 9, 10, 12], "devic": [0, 1, 2, 3, 4, 8, 10, 11, 12, 13], "avail": [0, 2, 4, 5, 7, 8, 10, 11], "driver": [0, 2, 7], "version": [0, 3, 5, 8], "locat": [0, 1, 2, 8, 9, 11, 12], "ani": [0, 3, 6, 7, 8, 9, 10, 11, 12, 13], "gener": [0, 1, 3, 4, 5, 6, 8, 11, 12], "code": [0, 1, 3, 4, 7, 10, 11, 12], "e": [0, 1, 2, 4, 5, 6, 8, 9, 11, 12, 13], "g": [0, 1, 2, 4, 6, 8, 9, 11, 12, 13], "1": [0, 2, 3, 5, 6, 7, 8, 9, 10, 12, 13], "0": [0, 1, 2, 3, 7, 8, 9, 10, 11, 12, 13], "cuda": [0, 1, 3, 4, 5, 6, 10, 11, 12], "toolkit": [0, 2, 5, 8], "11": [0, 2, 5], "8": [0, 2, 5, 8, 11, 12], "12": [0, 5], "cpu": [0, 2, 3, 4, 7, 10, 11, 12], "amd64": [0, 2], "famili": [0, 2, 8], "25": [0, 2, 11, 12], "33": [0, 2], "step": [0, 4, 8, 9, 10, 12], "authenticamd": [0, 2], "nvidia": [0, 2, 11], "geforc": [0, 2], "rtx": [0, 2], "4080": 0, "sm_89": 0, "cach": [1, 2, 6, 8, 13], "c": [0, 2, 3, 5, 7, 9, 11, 12], "mmacklin": 0, "appdata": [0, 2], "local": [0, 2, 4, 5, 8, 9, 11, 12], "corpor": [0, 2], "In": [0, 1, 2, 3, 4, 7, 8, 9, 11, 12, 13], "ar": [0, 1, 2, 3, 4, 5, 6, 8, 9, 10, 11, 12, 13], "defin": [0, 2, 8, 9, 11, 12, 13], "python": [0, 2, 4, 5, 10, 11, 12], "annot": [0, 11], "decor": [0, 1, 8, 11], "def": [0, 1, 4, 7, 8, 10, 11, 12], "simple_kernel": [0, 11], "dtype": [0, 1, 4, 7, 8, 9, 10, 11, 12, 13], "vec3": [0, 2, 4, 8, 9, 11, 12], "b": [0, 7, 8, 9, 11, 12], "float": [0, 1, 2, 4, 7, 8, 9, 10, 11, 12], "get": [0, 7, 8, 11, 13], "thread": [0, 3, 4, 6, 9, 11], "index": [0, 4, 8, 9, 11, 12, 13], "tid": [0, 4, 6, 7, 9, 10, 11], "load": [0, 11, 12], "two": [0, 7, 8, 9, 11, 12, 13], "x": [0, 1, 2, 4, 8, 9, 10, 11, 12, 13], "y": [0, 1, 3, 4, 8, 9, 11, 12, 13], "dot": [0, 8, 9, 11], "product": [0, 8, 9, 11, 13], "between": [0, 3, 8, 9, 11, 12], "vector": [0, 2, 4, 8, 12, 13], "r": [0, 8, 9, 11, 12, 13], "write": [0, 4, 7, 11], "result": [0, 1, 7, 8, 9, 10, 11], "back": [0, 3, 10, 11], "memori": [0, 1, 2, 3, 4, 7, 9, 10, 11, 12], "becaus": [0, 7, 11], "nativ": [0, 3, 4], "all": [0, 2, 4, 7, 8, 9, 11, 12, 13], "input": [0, 1, 2, 4, 7, 9, 10, 11, 12, 13], "argument": [0, 1, 2, 7, 8, 9, 11, 12], "must": [0, 8, 9, 11, 13], "static": [0, 11, 12], "thi": [0, 1, 2, 3, 5, 6, 7, 8, 9, 10, 11, 12, 13], "allow": [0, 7, 8, 9, 10, 11, 12, 13], "fast": [0, 1, 9, 11], "execut": [0, 5, 7, 11], "essenti": 0, "speed": 0, "mai": [0, 1, 2, 5, 6, 8, 9, 10, 11, 12, 13], "run": [0, 2, 3, 4, 5, 6, 7, 11], "either": [0, 7, 8, 12, 13], "gpu": [0, 2, 4, 6, 10, 11], "thei": [0, 3, 7, 8, 11, 12], "cannot": [0, 1, 6, 8, 11], "access": [0, 2, 11], "arbitrari": [0, 3, 7, 8, 11], "global": [0, 4, 5, 8, 9, 11], "state": [0, 2, 4, 9, 11], "from": [0, 1, 2, 3, 4, 6, 7, 8, 9, 10, 11, 13], "environ": [0, 4, 7, 11, 12], "instead": [0, 1, 6, 8, 11, 12, 13], "read": [0, 8, 11, 12, 13], "data": [0, 3, 4, 6, 8, 10, 12, 13], "through": [0, 3, 4, 8, 10, 11, 12, 13], "paramet": [0, 8, 9, 10, 11, 12, 13], "have": [0, 1, 3, 5, 6, 8, 11, 12], "correspond": [0, 2, 7, 8, 9, 10, 11], "launch": [0, 1, 3, 4, 7, 8, 9, 10], "1024": [0, 4, 7, 11], "we": [0, 3, 7, 10, 11], "dim": [0, 3, 4, 7, 9, 10, 11], "number": [0, 1, 2, 4, 6, 7, 8, 9, 11, 12, 13], "insid": [0, 4, 5, 8, 9, 11], "can": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13], "retriev": [0, 9, 11], "each": [0, 1, 2, 4, 6, 7, 8, 9, 11, 12, 13], "builtin": [0, 3], "i": [0, 1, 2, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13], "1d": [0, 8, 9, 11], "2d": [0, 8, 9, 10, 11], "3d": [0, 4, 8, 9, 11, 12, 13], "4d": [0, 9, 11, 12], "grid": [0, 4, 8, 9, 12], "process": [0, 2, 3, 4, 8, 9, 11], "1024x1024": 0, "imag": 0, "could": [0, 5], "compute_imag": 0, "img": 0, "Then": [0, 12], "j": [0, 4, 9, 11, 12], "out": [0, 1, 2, 8, 9, 11, 13], "color": [0, 11], "valu": [0, 1, 4, 6, 8, 9, 10, 12, 13], "pixel": 0, "alloc": [0, 3, 4, 6, 7, 8, 10, 11, 12, 13], "expos": [0, 3, 7, 11, 12], "via": [0, 10, 11, 12], "wrap": [0, 10, 11], "an": [0, 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13], "underli": [0, 4, 8, 10, 11], "live": [0, 10, 11], "host": [0, 3, 7, 11, 13], "strongli": [0, 11, 12], "store": [0, 2, 8, 9, 11, 12, 13], "linear": [0, 8, 9, 11, 12, 13], "sequenc": [0, 11, 12], "int": [0, 2, 8, 9, 11, 12, 13], "matrix33": 0, "etc": [0, 1, 4, 6, 8, 11, 12], "similar": [0, 3, 7, 8, 9, 11, 12], "pytorch": [0, 4, 5, 7, 9], "uniniti": [0, 11, 13], "v": [0, 2, 4, 5, 8, 9, 11], "empti": [0, 7, 11, 13], "shape": [0, 2, 8, 9, 10, 11, 12, 13], "n": [0, 7, 8, 9, 11], "zero": [0, 4, 7, 8, 9, 10, 11, 12, 13], "quaternion": [0, 4, 12], "q": [0, 9, 11], "quat": [0, 11], "numpi": [0, 3, 4, 5, 7, 9, 11, 12], "automat": [0, 1, 3, 4, 8, 9, 11, 12], "transfer": [0, 11, 12], "specifi": [0, 7, 8, 9, 11, 12], "np": [0, 4, 10, 11, 12], "ones": [0, 3, 8, 11], "10": [0, 7, 9, 11, 12], "3": [0, 2, 4, 5, 8, 9, 10, 11, 12], "float32": [0, 7, 9, 10, 11, 12], "from_numpi": 0, "By": [0, 8, 11], "default": [0, 1, 4, 8, 9, 10, 11, 12], "extern": [0, 7, 11], "list": [0, 3, 6, 7, 8, 11, 12], "tupl": [0, 8, 9, 11, 12, 13], "creat": [0, 1, 6, 7, 8, 9, 11, 12, 13], "copi": [0, 3, 7, 8, 10, 11, 12, 13], "new": [0, 3, 6, 7, 8, 9, 10, 11, 12, 13], "howev": [0, 3, 8, 12], "possibl": [0, 2, 7, 8, 11, 12], "alia": [0, 7, 11, 13], "fals": [0, 1, 8, 9, 11, 12], "constructor": [0, 9, 11], "provid": [0, 4, 8, 9, 10, 11, 13], "contigu": [0, 11], "same": [0, 5, 6, 8, 9, 10, 11, 12, 13], "see": [0, 2, 4, 5, 6, 8, 9, 10, 11, 12], "interoper": [0, 4, 5, 7], "section": [0, 6, 8, 11, 12], "more": [0, 1, 3, 7, 8, 9, 10, 11, 12], "share": [0, 3, 5, 7, 8, 9, 10, 11], "framework": [0, 3, 4, 7, 9, 10], "To": [0, 1, 2, 4, 5, 6, 7, 8, 11], "bring": [0, 3], "view": [0, 4, 8, 10, 11], "device_arrai": [0, 11], "synchron": [0, 2, 4, 11], "ensur": [0, 2, 3, 4, 5, 8, 11, 12], "outstand": [0, 7, 11], "work": [0, 3, 5, 7, 11, 12, 13], "ha": [0, 1, 2, 4, 5, 7, 9, 11, 12], "finish": [0, 3], "where": [0, 3, 7, 9, 10, 11], "pass": [0, 1, 5, 6, 8, 9, 11, 13], "call": [0, 1, 2, 4, 6, 7, 8, 9, 10, 11, 12, 13], "return": [0, 3, 8, 9, 10, 11, 12, 13], "onto": [0, 9, 10, 13], "own": [0, 8, 11, 12], "func": [0, 8, 9, 11], "exampl": [1, 2, 3, 5, 6], "squar": [0, 8, 9, 11, 13], "freeli": 0, "within": [0, 2, 8, 9, 11, 12], "modul": [0, 4, 5, 8, 10, 11, 12, 13], "accept": [0, 8, 11], "definit": [0, 11, 12], "belong": [0, 12], "runtim": [0, 2, 3, 4, 5, 6, 7, 12], "dynam": [0, 3, 6, 11, 12], "librari": [0, 5, 9, 11], "ptx": [0, 1], "applic": [0, 1, 2, 7, 11], "restart": 0, "time": [0, 1, 4, 8, 10, 11, 12, 13], "note": [0, 1, 3, 4, 8, 9, 11], "trigger": 0, "first": [0, 2, 3, 4, 5, 8, 9, 10, 11, 12, 13], "regist": [0, 7, 12], "includ": [0, 4, 8, 11, 12, 13], "support": [0, 2, 4, 5, 6, 7, 8, 10, 11, 12, 13], "differenti": [0, 4, 8, 12], "differ": [0, 3, 4, 5, 7, 8, 9, 11, 12], "cpython": [0, 11], "math": [0, 1, 4, 11], "high": [0, 3, 4, 11], "level": [0, 1, 3, 11, 12], "shade": 0, "vec2": [0, 9, 11], "vec4": [0, 9, 11], "mat22": [0, 11], "mat33": [0, 11, 12], "mat44": [0, 11], "semant": 0, "so": [0, 3, 5, 7, 8, 10, 11, 12], "express": [0, 6, 8, 11, 12], "variabl": [0, 1, 2, 5, 8, 9, 10, 11], "rather": [0, 3, 12], "than": [0, 3, 5, 6, 7, 8, 9, 11, 12], "refer": [0, 8, 12], "unlik": 0, "infer": [0, 3, 9, 11], "sourc": [0, 3, 4, 10, 11, 12], "signatur": [0, 11], "extens": [0, 4, 11], "appropri": [0, 11], "invalid": 0, "2": [0, 1, 2, 5, 6, 8, 9, 10, 12], "valid": [0, 1, 2, 11, 12], "achiev": [6, 11], "good": [3, 6], "perform": [1, 2, 3, 4, 6, 8, 11, 13], "lambda": [6, 12], "comprehens": [3, 6, 11], "except": [2, 3, 6, 11, 12], "recurs": 6, "evalu": [6, 8, 9, 10, 11, 12], "eval": 6, "structur": [3, 6, 11, 12, 13], "set": [0, 2, 3, 4, 5, 6, 7, 8, 11, 12, 13], "dictionari": [1, 6, 8, 11, 12], "often": [2, 3, 11], "one": [2, 3, 5, 7, 8, 9, 10, 11, 12, 13], "best": [2, 11], "simpli": [2, 8, 11], "kernel": [2, 3, 4, 7, 8, 10], "built": [2, 4, 5, 7, 8, 11], "type": [1, 2, 3, 4, 6, 8, 10, 13], "function": [2, 4, 6, 7, 9, 10, 12, 13], "addit": [1, 2, 3, 7, 8, 9, 11, 12, 13], "format": [1, 2, 9, 10, 11, 13], "style": [2, 9, 11], "printf": [2, 9], "A": [2, 4, 6, 7, 8, 9, 10, 11, 12, 13], "f": [2, 4, 7, 8, 10, 11], "d": [2, 4, 8, 9, 11, 12], "onli": [1, 2, 3, 8, 9, 10, 11, 12, 13], "scalar": [2, 4, 8, 13], "For": [1, 2, 3, 5, 7, 8, 11, 13], "complex": [2, 6, 11], "difficult": 2, "understand": 2, "order": [2, 9, 11, 13], "oper": [1, 2, 3, 4, 7, 13], "lead": [2, 11], "bug": [2, 11], "help": [2, 7, 8], "diagnos": 2, "issu": [2, 6, 7, 8, 12], "simpl": [2, 11], "option": [1, 2, 5, 8, 9, 10, 11, 12, 13], "consol": 2, "config": [0, 1, 2, 5], "print_launch": [1, 2], "true": [1, 2, 5, 8, 9, 10, 11, 12], "It": [2, 7, 8, 9, 11, 12], "attach": [2, 12], "id": [2, 9, 11, 12], "debugg": [2, 9], "visual": [2, 5, 11, 12], "studio": [2, 5], "user": [1, 2, 3, 5, 6, 7, 8, 9, 11, 12], "compil": [1, 2, 4, 11], "mode": [1, 2, 5, 11], "line": [2, 4, 10], "symbol": [2, 12], "correctli": [2, 11], "after": [0, 1, 2, 5, 6, 9, 11, 12], "breakpoint": [2, 9], "insert": [2, 11], "origin": [2, 3, 4, 11, 12], "individu": [2, 3, 7, 8, 13], "still": [2, 7, 11, 12], "replai": 2, "inspect": 2, "The": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13], "central": 2, "": [1, 2, 5, 7, 8, 9, 10, 11, 12, 13], "home": 2, "directori": [1, 2, 4, 5, 6, 8, 12], "when": [2, 4, 5, 6, 7, 8, 9, 10, 11, 12], "initi": [1, 2, 4, 7, 8, 9, 10, 11, 12], "model": [2, 3, 4, 8], "3090": 2, "sm_86": 2, "2080": 2, "ti": 2, "sm_75": 2, "lukasw": 2, "gen": [1, 2], "bin": [1, 2, 5], "folder": [2, 12], "contain": [2, 4, 8, 9, 10, 11, 12, 13], "binari": [2, 4, 5, 9], "respect": [2, 5, 8, 11, 12, 13], "occasion": 2, "profil": [2, 4], "build": [0, 2, 3, 4, 8, 11, 12], "configur": [1, 2, 12], "arrai": [1, 2, 3, 4, 7, 8, 9, 10, 12, 13], "lie": 2, "violat": 2, "poorli": [2, 9], "form": [2, 7, 8, 9, 11, 12], "case": [1, 2, 3, 7, 9, 11, 12, 13], "detect": [2, 11], "put": [2, 12], "context": [2, 4, 6, 8, 11], "error": [1, 2, 5, 9, 11], "subsequ": [2, 5, 11], "silent": 2, "fail": 2, "which": [1, 2, 3, 5, 7, 8, 9, 11, 12, 13], "hard": 2, "If": [1, 2, 4, 5, 7, 8, 9, 10, 11, 12, 13], "suspect": 2, "enabl": [1, 2, 4, 5, 11, 12], "verify_cuda": [1, 2], "everi": [0, 1, 2, 6, 11, 12], "encount": [2, 12], "rais": [2, 9, 11], "narrow": 2, "down": 2, "problemat": [2, 7], "verifi": 2, "requir": [2, 3, 5, 8, 11, 12, 13], "signific": [2, 11], "overhead": [2, 11], "dure": [1, 2, 7, 9, 11, 12], "inspir": 3, "mani": [3, 8, 11, 12], "close": [3, 8, 11, 12], "both": [3, 8, 11], "These": [3, 10, 11], "map": [3, 7, 8, 10, 11, 12], "tradit": 3, "concept": [3, 7, 11], "implement": [3, 4, 9, 10, 11, 12, 13], "compar": [3, 9, 11], "smaller": 3, "subset": [3, 8, 12], "offer": [3, 7], "auto": [3, 11], "machin": [3, 4], "learn": 3, "intermedi": [3, 11], "represent": [3, 6, 11, 12], "make": [3, 4, 5, 7, 8, 9, 11, 12], "conveni": [3, 11, 13], "low": [3, 11], "routin": 3, "geometri": [3, 4, 11, 12], "mesh": [3, 4, 8, 9], "spars": [3, 4, 6, 8, 11], "volum": [3, 4, 8, 12], "point": [3, 4, 8, 9, 10, 11, 12], "cloud": 3, "usd": [3, 4, 5, 9, 12], "class": [3, 8, 9, 11, 12, 13], "citizen": 3, "full": [3, 11], "tensor": [3, 7, 8, 10, 11, 12], "base": [3, 9, 10, 11, 12, 13], "jax": [3, 4, 5], "design": [3, 4, 7, 8, 12], "well": [3, 7, 9, 11, 12], "mechan": [3, 8, 11, 12], "__cuda_array_interface__": [3, 10], "neural": 3, "network": [3, 9], "sens": [3, 9], "exist": [3, 9, 10, 11, 12, 13], "tool": [3, 5, 11], "problem": [3, 8, 11], "lot": 3, "sparsiti": 3, "condit": [3, 4, 9, 11], "logic": 3, "heterogen": 3, "workload": 3, "find": [3, 9, 11], "simul": [3, 8, 11, 12, 13], "graphic": [3, 4, 11], "sinc": [3, 11], "control": [1, 3, 4, 12], "No": [3, 11, 12], "our": [3, 11], "goal": 3, "cliff": 3, "expect": [3, 9, 11, 12], "consist": [3, 11], "behavior": [3, 11, 12], "unsupport": [3, 4], "don": [3, 5], "t": [3, 5, 7, 9, 10, 11, 12], "garbag": [3, 11], "collect": [3, 11], "One": [3, 8, 11], "common": [3, 8, 9, 11, 12, 13], "confus": 3, "necessari": 3, "answer": 3, "almost": 3, "never": [3, 8], "quit": 3, "expens": [3, 9], "avoid": [3, 6, 7, 11], "unless": [3, 8, 12], "natur": [3, 9], "take": [1, 3, 4, 9, 11], "care": [3, 11], "manual": [3, 5], "convers": [3, 10], "kernel_1": 3, "array_x": 3, "array_i": 3, "kernel_2": 3, "array_z": 3, "implicitli": 3, "need": [1, 3, 5, 10, 11, 12], "being": [3, 8, 9, 11], "asynchron": [3, 7, 11], "cpu_array_1": 3, "gpu_array_1": 3, "cpu_array_2": 3, "gpu_array_2": 3, "cpu_array_3": 3, "gpu_array_3": 3, "wrapper": [3, 11], "around": [3, 9, 10, 11, 12], "implicit": [3, 11, 12], "here": [3, 10, 11], "a1": 3, "a2": 3, "a3": 3, "non": [3, 8, 9, 11, 12, 13], "smooth": [3, 9, 11], "do": [3, 8, 11, 12], "singl": [3, 7, 8, 9, 10, 11, 12], "uniqu": [3, 7, 11], "gradient": [3, 8, 9, 10, 12], "known": 3, "subgradi": 3, "formal": 3, "convex": 3, "hull": 3, "direct": [3, 9, 10, 12], "deriv": [3, 8, 11, 12], "wai": [3, 4, 5, 7, 8, 9, 10, 11], "most": [3, 4, 9, 11, 12, 13], "handl": [3, 8, 11, 12], "pick": 3, "arbitrarili": [3, 7], "choos": [3, 8, 11], "h": [3, 11], "optim": 3, "particularli": [3, 11], "exploit": 3, "stochast": 3, "sensit": 3, "choic": [1, 3, 8, 12], "although": 3, "ye": 3, "4": [3, 5, 8, 9, 11, 12], "multipl": [1, 3, 4, 8, 9, 13], "name": [1, 3, 7, 8, 11], "convent": [3, 7], "alias": [3, 7, 11], "identifi": [3, 9, 11], "replac": [3, 8, 11], "isaacsim": 3, "while": [1, 3, 9, 11], "physic": [3, 4, 8, 12], "capabl": [3, 7, 11], "primarili": 3, "aim": 3, "develop": [3, 5, 8], "who": [3, 5], "fulli": [3, 8, 11, 12], "featur": [3, 4, 5, 7, 9, 11, 12], "engin": [3, 11], "also": [1, 2, 3, 4, 5, 8, 9, 10, 11, 13], "integr": [3, 4, 7], "great": 3, "auxiliari": 3, "task": 3, "reward": 3, "observ": 3, "reinforc": 3, "convert": [8, 9, 10, 11, 12], "jit": 4, "come": [4, 9], "rich": 4, "primit": 4, "easi": [4, 7, 8], "program": [1, 4, 9, 11], "spatial": [4, 11], "part": [4, 7, 8, 9, 11], "train": [], "pipelin": [4, 7], "other": [1, 4, 5, 7, 8, 10, 11, 12], "below": [1, 4, 8, 9, 10, 11], "easiest": [4, 5], "instal": [4, 11], "pypi": [4, 5], "pip": [4, 5], "lang": [4, 5], "pre": [4, 5], "packag": [4, 5], "window": [4, 5], "linux": [4, 5], "maco": [4, 5], "releas": [1, 4, 5, 8], "page": [4, 5], "your": [4, 5, 8, 11, 12], "extract": [4, 5, 9, 13], "archiv": [4, 5], "command": [4, 5, 11], "root": [4, 5, 8, 9, 11, 12], "length": [4, 8, 9, 11, 12], "random": [4, 10, 11], "given": [4, 8, 9, 10, 11, 12, 13], "num_point": 4, "distanc": [4, 9, 11, 12], "rand": [4, 11], "len": [4, 7, 10, 11], "github": [4, 6], "repositori": [4, 5], "script": [4, 7, 11], "show": [4, 7], "how": [0, 2, 4, 6, 7, 11], "file": [4, 5, 11], "sampl": [4, 8, 9, 11], "anim": 4, "output": [1, 4, 8, 9, 10, 11, 12], "render": [4, 5, 12], "pixar": 4, "usdview": 4, "blender": 4, "preview": 4, "recommend": [4, 7, 11, 12], "limit": [4, 8, 11, 12], "unit": [4, 9, 11, 12], "test": [2, 4, 8, 9, 11], "m": [4, 9, 11, 12], "registri": 4, "kit": 4, "pleas": [4, 5, 10, 11, 12], "detail": [1, 4, 10, 11], "resourc": [4, 7], "background": [4, 9, 11], "gtc": 4, "2022": 4, "present": [4, 11], "2021": 4, "siggraph": 4, "asia": 4, "cours": [4, 11], "technologi": 4, "been": [1, 4, 6, 9, 11, 12], "research": 4, "project": [4, 5, 8], "public": 4, "acceler": [4, 11, 12], "polici": 4, "parallel": [4, 9, 11], "xu": 4, "makoviychuk": 4, "narang": 4, "ramo": 4, "matusik": 4, "w": [4, 9, 10, 11, 12], "garg": 4, "macklin": 4, "disect": 4, "robot": 4, "cut": 4, "heiden": 4, "fox": 4, "gradsim": 4, "system": [4, 5, 7, 8, 11, 12], "identif": 4, "visuomotor": 4, "murthi": 4, "krishna": 4, "mile": 4, "florian": 4, "golemo": 4, "vikram": 4, "voleti": 4, "linda": 4, "petrini": 4, "martin": 4, "weiss": 4, "breandan": 4, "considin": 4, "et": [4, 12], "al": [4, 12], "you": [4, 5, 11, 12], "citat": 4, "misc": 4, "warp2022": 4, "titl": 4, "author": 4, "month": 4, "march": 4, "year": 4, "confer": 4, "howpublish": 4, "url": [4, 12], "http": [4, 12], "com": [4, 12], "under": [4, 7, 12], "nvscl": 4, "md": 4, "text": 4, "contact": [4, 12], "question": 4, "commerci": 4, "inquir": 4, "depend": [4, 12], "languag": [4, 6], "scope": [1, 4, 11], "current": [0, 1, 4, 6, 8, 9, 10, 11, 12, 13], "custom": [4, 12], "scopedtim": [5, 11], "cupi": 4, "numba": 4, "dlpack": 4, "debug": [1, 4], "bound": [4, 8, 9, 12], "check": [1, 4, 8, 9, 11, 12], "verif": [1, 4], "faq": 4, "doe": [1, 4, 6, 7, 8, 9, 11, 12], "relat": [4, 8], "taichi": 4, "what": [4, 11], "happen": [4, 11], "like": [4, 7, 8, 9, 11], "ab": [4, 9], "multi": [4, 9], "switch": [4, 7, 11], "over": [1, 4, 7, 8, 9, 10, 11], "isaacgym": 4, "physx": 4, "constant": [4, 8, 12, 13], "hash": [4, 9], "graph": [1, 4], "transform": [4, 12], "util": 4, "sim": [4, 5], "matric": [4, 8, 9], "fem": [4, 12], "integrand": 4, "workflow": [4, 7], "introductori": 4, "advanc": [4, 11, 12], "usag": [4, 9, 10], "space": [4, 9, 11, 12], "field": [1, 4, 9, 11, 12], "7": [5, 11, 12], "onward": [], "certain": [5, 11], "nvtx": [5, 11], "document": 5, "sphinx": 5, "furo": 5, "want": [5, 11], "themselv": [5, 11], "microsoft": 5, "2019": 5, "upward": [], "gcc": 5, "5": [5, 9, 11, 12], "higher": 8, "git": 5, "larg": [1, 5], "storag": [5, 11, 13], "clone": [5, 11], "build_lib": 5, "py": [2, 5, 8, 9, 11], "dll": 5, "core": [5, 8, 11], "cuda_path": 5, "otherwis": [5, 7, 8, 9, 10, 11, 12, 13], "without": [5, 10, 11, 12], "altern": [5, 7, 9, 12], "path": [0, 1, 5, 11, 12], "modif": 5, "reflect": 5, "symlink": 5, "ext": 5, "omni": 5, "upon": [5, 8, 11], "assign": [1, 7, 8, 11, 12, 13], "string": [1, 6, 7, 9, 11, 12], "There": [6, 7], "ordin": 7, "familiar": 7, "popular": 7, "target": [1, 7, 10, 11, 12], "specif": [1, 7, 11], "api": [4, 7, 11], "size": [1, 6, 7, 11, 13], "primari": [7, 11], "compat": [7, 10, 11], "softwar": 7, "simplifi": [7, 12], "omit": 7, "set_devic": 7, "chang": [1, 6, 7, 11, 12, 13], "thing": 7, "minim": [7, 10], "block": [6, 7, 9, 11, 13], "torch": [7, 10], "snippet": [7, 11], "explicit": [7, 8, 12], "anoth": [7, 8, 9, 10, 11, 12], "manag": [4, 5, 7, 11], "scopeddevic": [], "object": [7, 8, 9, 11, 12], "nest": [1, 7, 11], "restor": 7, "previou": [1, 7, 11], "exit": [7, 11], "handi": 7, "bigger": 7, "side": [7, 8, 11, 13], "effect": [1, 7, 11], "enclos": 7, "analog": 7, "invok": 7, "example_funct": 7, "arang": [7, 11], "from_torch": 7, "re": [7, 8], "unpredict": 7, "overlap": [7, 9], "synchronize_devic": 7, "fine": [1, 7, 11], "grain": 7, "latter": 7, "wait": 7, "complet": [7, 8, 11], "easili": 7, "hide": 7, "hood": 7, "repres": [7, 9, 11, 13], "align": [7, 8, 9, 12], "directli": [7, 8, 10, 11, 12, 13], "affin": [7, 9], "interop": [7, 10], "benefit": 7, "special": [7, 8, 13], "whether": [1, 7, 8, 10, 11, 12], "cucontext": 7, "pointer": [7, 11], "map_cuda_devic": 7, "foo": 7, "ctype": [7, 11], "c_void_p": 7, "context_ptr": 7, "wa": [7, 12], "made": [6, 7], "u": [7, 8, 9, 11, 12], "facilit": 8, "solv": [5, 8], "describ": [8, 11, 12], "equat": [8, 12], "pde": 8, "diffus": 8, "convect": 8, "fluid": [8, 11], "flow": [8, 11], "elast": [8, 12], "finit": [1, 8, 11], "element": [8, 9, 11, 12, 13], "galerkin": 8, "quickli": 8, "experi": 8, "variou": [1, 6, 8, 11], "formul": 8, "discret": [8, 11], "scheme": 8, "abil": 8, "bilinear": 8, "domain": [8, 11], "interpol": [8, 9, 11], "basi": [8, 9, 11, 12], "main": [8, 11], "instanc": [8, 11, 12, 13], "linear_form": 8, "diffusion_form": 8, "nu": 8, "grad": [8, 9, 10], "normal": [8, 9, 11, 12], "mean": [8, 11, 12], "usual": 8, "few": 8, "coordin": [8, 9, 11, 12, 13], "abstract": [8, 11], "actual": [8, 11], "discretefield": 8, "testfield": 8, "trialfield": 8, "functionspac": 8, "inner": 8, "shortcut": 8, "sever": 8, "posit": [8, 9, 11, 12], "y_mass_form": 8, "mass": 12, "vertic": [8, 12], "edg": [9, 11, 12], "nor": [8, 11], "launcher": 8, "standard": [1, 8, 10, 11, 12], "trial": 8, "viscos": 8, "typic": [8, 9, 11], "At": [8, 10, 11], "moment": [8, 12], "regular": [4, 8, 12], "triangl": [8, 9, 11], "tetrahedron": [8, 12], "equip": 8, "make_polynomial_spac": 8, "lagrang": 8, "polynomi": 8, "up": [5, 8, 11, 12], "geometrydomain": 8, "cell": [8, 11, 12], "boundari": 4, "right": [8, 9, 12, 13], "hand": [8, 13], "make_test": 8, "formula": 8, "let": 8, "degre": [8, 9, 11, 12], "freedom": [8, 12], "left": [8, 9, 12, 13], "make_tri": 8, "bsrmatrix": [8, 13], "pair": [1, 8, 12], "mix": 8, "solver": 8, "excerpt": 8, "example_diffus": 8, "outlin": 8, "procedur": 8, "geo": [8, 12], "grid2d": 8, "50": 8, "cell_siz": 8, "scalar_spac": 8, "forc": [0, 8, 11, 12], "term": 8, "rh": 8, "weakli": 8, "impos": [8, 11], "boundarysid": 8, "bd_test": 8, "bd_trial": 8, "bd_matrix": 8, "matrix": [8, 9, 12, 13], "assembl": [8, 11], "add": [8, 9, 11, 12], "bsr_axpi": [8, 13], "alpha": [8, 11, 13], "boundary_strength": 8, "beta": [8, 11, 13], "conjug": [8, 9], "zeros_lik": [8, 11], "bsr_cg": 8, "iter": [1, 7, 8, 9, 11, 12], "studi": 8, "permit": 8, "remain": 8, "ship": 8, "illustr": [0, 8], "homogen": [8, 9], "neumann": 8, "dirichlet": 8, "example_diffusion_3d": 8, "variant": 8, "example_convection_diffus": 8, "semi": [8, 12], "lagrangian": 8, "advect": 8, "example_diffusion_dg0": 8, "upwind": 8, "transport": 8, "example_diffusion_dg": 8, "discontinu": 8, "aith": [], "symmetr": [8, 9], "interior": 8, "penalti": 8, "example_stok": 8, "incompress": 8, "stoke": 8, "p_k": 8, "p_": 8, "k": [8, 9, 11, 12], "example_navier_stok": 8, "navier": 8, "example_mixed_elast": 8, "continu": 8, "picquadratur": 8, "mpm": 8, "bucket": [8, 11], "example_stokes_transf": 8, "_3d": [], "_usd": [], "seri": 11, "possibli": [8, 13], "distribut": [8, 9], "example_diffusion_mgpu": 8, "activ": [5, 8, 9, 11, 12], "subclass": 8, "geometrypartit": 8, "lineargeometrypartit": 8, "explicitgeometrypartit": 8, "accord": 8, "make_space_partit": 8, "spacepartit": 8, "translat": [8, 9, 11, 12], "wide": 8, "node": 8, "indic": [8, 9, 10, 11, 12, 13], "frontier": 8, "exterior": 8, "world": [8, 9, 11, 12], "null": [8, 9], "lookup": [8, 11], "look": [5, 8], "closet": [], "measur": 8, "area": [8, 12], "measure_ratio": 8, "maximum": [1, 6, 8, 9, 11, 12], "ratio": 8, "dimension": [8, 13], "neighbour": 8, "On": 8, "orient": [8, 12, 13], "outer": [8, 9], "equival": [8, 9, 11], "grad_out": 8, "div": [8, 9], "diverg": [8, 9], "jump": 8, "averag": [8, 9], "grad_jump": 8, "grad_averag": 8, "resolv": [8, 12], "callabl": [8, 9, 11], "none": [1, 8, 9, 10, 11, 12, 13], "nodal": 8, "accumulate_dtyp": 8, "float64": [8, 9, 11], "output_dtyp": 8, "deduc": 8, "bool": [8, 9, 10, 11, 12], "assum": [8, 9, 12, 13], "dg": 8, "kei": [8, 11], "match": [8, 11, 12, 13], "accumul": [8, 10, 11], "dest": [8, 11, 13], "fieldrestrict": 8, "restrict": [8, 11], "per": [1, 6, 7, 8, 9, 11, 12, 13], "interfac": [4, 10, 11, 12], "compos": [4, 8], "reference_cel": [], "prototyp": [], "reference_sid": [], "cellarg": [], "struct": 4, "quantiti": [], "sidearg": [], "cell_arg_valu": [], "cell_posit": [], "arg": [8, 11], "cell_lookup": [], "po": [9, 12], "guess": 8, "faster": [1, 9, 12], "cell_measur": [], "determin": [1, 8, 9, 11, 12, 13], "cell_measure_ratio": [], "its": [8, 9, 11, 12, 13], "cell_norm": [], "side_arg_valu": [], "boundary_side_index": [], "side_posit": [], "side_measur": [], "side_measure_ratio": [], "side_norm": [], "side_inner_cell_index": [], "side_index": [], "side_outer_cell_index": [], "cell_count": 8, "side_count": 8, "boundary_side_count": 8, "frontier_side_count": 8, "neighbor": [8, 9, 11], "partial": [8, 11], "dimens": [1, 6, 8, 9, 11, 12, 13], "element_count": 8, "reference_el": [], "protyp": [], "element_index_arg_valu": [], "str": [8, 11, 12, 13], "structinst": [], "element_arg_valu": [], "elementindexarg": [], "element_index": [], "elementindex": [], "elementarg": [], "element_measur": [], "element_measure_ratio": [], "element_posit": [], "element_norm": [], "element_lookup": [], "rule": [8, 11], "eval_arg_valu": [], "total_point_count": 8, "total": [1, 8, 12], "point_count": [], "point_coord": [], "qp_index": [], "th": [], "point_weight": [], "weight": 9, "partition_rank": 8, "partition_count": 8, "uniformli": [8, 9], "partion": 8, "cell_mask": 8, "count": [11, 12], "select": [8, 9, 11], "irregular": 8, "spread": 8, "codegen": [], "dof_mapp": 8, "dofmapp": 8, "ident": [8, 9, 11, 13], "reduc": [1, 8], "symmetrictensormapp": 8, "2x2": [8, 9], "resp": 8, "3x3": [8, 9, 12], "6": [5, 8, 9, 11, 12], "impli": [8, 11], "piecewis": 8, "reason": [8, 11], "chosen": 8, "construct": [8, 9, 10, 11, 12, 13], "geometry_partit": 8, "with_halo": 8, "substep": [], "touch": 8, "consid": [1, 8, 9, 12], "whole": [8, 9], "halo": 8, "make_space_restrict": 8, "space_partit": 8, "spacerestrict": 8, "nodes_per_el": 8, "value_dof_count": [], "spacearg": [], "node_count": 8, "node_posit": 8, "purpos": 13, "space_arg_valu": [], "trace": [8, 9], "lower": [8, 9, 11, 12], "make_field": 8, "unit_dof_valu": [], "dof": [], "rank": [], "vec_t": [], "element_node_index": [], "node_index_in_elt": [], "node_coords_in_el": [], "element_inner_weight": [], "coord": [], "vec3f": [8, 9, 11], "element_inner_weight_gradi": [], "element_outer_weight": [], "geo_partit": 8, "owned_node_count": 8, "exclud": 8, "interior_node_count": 8, "space_node_indic": 8, "partition_node_index": [], "space_node_index": [], "orthonorm": 8, "isomorph": 8, "nxn": 8, "l2": 8, "norm": 8, "half": [8, 11, 12], "frobeniu": 8, "tau": 8, "daviet": [], "bertail": [], "descoub": [], "2016": [], "make_restrict": 8, "space_restrict": 8, "evalarg": [], "properti": [8, 11, 12, 13], "dof_valu": 8, "set_node_valu": [], "node_index": [], "eval_inn": [], "0x000001c0ffaded40": [], "eval_grad_inn": [], "eval_out": [], "eval_grad_out": [], "int8": [9, 11], "uint8": [9, 11], "int16": [9, 11], "uint16": [9, 11], "int32": [9, 11, 12], "uint32": [9, 11, 12], "int64": [6, 9, 11], "uint64": [9, 11], "float16": [9, 11], "vec2b": 9, "vec2ub": [9, 11], "vec2u": 9, "vec2i": [8, 9], "vec2ui": 9, "vec2l": 9, "vec2ul": 9, "vec2h": [9, 11], "vec2f": [8, 9, 11], "vec2d": [9, 11], "vec3b": 9, "vec3ub": [9, 11], "vec3u": 9, "vec3i": [8, 9], "vec3ui": 9, "vec3l": 9, "vec3ul": 9, "vec3h": [9, 11], "vec3d": [9, 11], "vec4b": 9, "vec4ub": [9, 11], "vec4u": 9, "vec4i": 9, "vec4ui": 9, "vec4l": 9, "vec4ul": 9, "vec4h": [9, 11], "vec4f": [9, 11], "vec4d": [9, 11], "mat22h": [9, 11], "mat22f": [9, 11], "mat22d": [9, 11], "mat33h": [9, 11], "mat33f": [9, 11], "mat33d": [9, 11], "mat44h": [9, 11], "mat44f": [9, 11], "mat44d": [9, 11], "quath": [9, 11], "quatf": [9, 11], "quatd": [9, 11], "transformh": [9, 11], "transformf": [9, 11, 12], "transformd": [9, 11], "spatial_vectorh": [9, 11], "spatial_vectorf": [9, 11], "spatial_vectord": [9, 11], "spatial_matrixh": [9, 11], "spatial_matrixf": [9, 11], "spatial_matrixd": [9, 11], "min": [9, 11], "minimum": [5, 9, 12], "wise": [9, 11], "max": [9, 11], "clamp": 9, "rang": [7, 9, 10, 11, 12], "absolut": 9, "sign": [6, 9, 11, 12], "nonzero": 9, "equal": [9, 11, 12, 13], "sin": 9, "sine": 9, "radian": [9, 12], "co": 9, "cosin": 9, "aco": 9, "arcco": 9, "asin": 9, "arcsin": 9, "sqrt": [9, 11], "tan": 9, "tangent": 9, "atan": 9, "arctan": [], "atan2": 9, "sinh": 9, "cosh": 9, "tanh": 9, "log": 9, "log2": 9, "log10": 9, "exp": 9, "exponenti": [9, 11], "pow": 9, "power": 9, "round": [9, 11], "calcul": 12, "nearest": [9, 11], "integ": [1, 6, 9, 11], "halfwai": 9, "awai": 9, "intuit": 9, "colloqui": 9, "slower": 9, "rint": 9, "behav": 9, "even": [9, 11], "trunc": 9, "closer": 9, "word": 9, "discard": [9, 13], "fraction": 9, "cast": [9, 11, 13], "preserv": [9, 12], "neg": [9, 11, 12], "fix": [1, 6, 9, 11, 12], "floor": [9, 11], "largest": 9, "less": [1, 9, 10, 11], "ceil": 9, "smallest": [9, 11], "greater": [6, 9, 11], "ddot": 9, "doubl": [9, 11], "argmin": 9, "argmax": 9, "cross": [8, 9, 11], "skew": [8, 9], "length_sq": 9, "transpos": [9, 11, 13], "invers": [4, 9, 11], "4x4": 9, "diag": [9, 13], "compon": [9, 11], "diagon": [9, 11, 12, 13], "get_diag": 9, "cw_mul": 9, "multipli": [9, 11], "cw_div": 9, "divis": [9, 11], "6d": [9, 11], "screw": 9, "rot": [9, 12], "scale": [8, 9, 11, 12, 13], "appli": [8, 9, 11, 12], "rotat": [9, 11, 12], "column": [9, 13], "tr": 9, "svd3": 9, "sigma": 9, "svd": 9, "singular": 9, "qr3": 9, "qr": 9, "decomposit": 9, "orthogon": [8, 9], "upper": [8, 9, 11, 12], "triangular": [8, 9, 12], "eig3": 9, "eigendecomposit": 9, "eigenvector": 9, "eigenvalu": 9, "laid": 9, "ix": 9, "ii": 9, "iz": 9, "imaginari": 9, "real": [9, 11], "z": [9, 11, 12, 13], "suppli": [9, 10], "quat_ident": [9, 11], "quat_from_axis_angl": [9, 11], "axi": [8, 9, 11, 12], "angl": [9, 12], "quat_to_axis_angl": 9, "quat_from_matrix": 9, "quat_rpi": 9, "roll": [9, 12], "pitch": 9, "yaw": 9, "combin": 9, "quat_invers": 9, "quat_rot": [9, 11], "p": [9, 11], "quat_rotate_inv": 9, "quat_slerp": 9, "q0": 9, "q1": 9, "linearli": 9, "quat_to_matrix": 9, "rigid": [9, 11, 12], "bodi": [9, 11, 12], "transform_ident": [9, 11], "transform_get_transl": 9, "transform_get_rot": 9, "transform_multipli": 9, "togeth": 9, "transform_point": [9, 11], "treat": [9, 12, 13], "contrast": 9, "notabl": 9, "row": [9, 11, 13], "transform_vector": [9, 11], "transform_invers": 9, "spatial_adjoint": 9, "6x6": [9, 11], "inerti": [9, 12], "spatial_dot": 9, "spatial_cross": 9, "spatial_cross_du": 9, "dual": 9, "spatial_top": 9, "top": [9, 11, 12], "spatial_bottom": 9, "bottom": [8, 9, 12], "second": [9, 11, 12, 13], "spatial_jacobian": 9, "joint_par": [9, 12], "joint_qd_start": [9, 12], "joint_start": 9, "joint_count": [9, 12], "j_start": 9, "j_out": 9, "spatial_mass": 9, "i_": 9, "m_start": 9, "mlp": 9, "bia": [9, 12], "layer": 9, "perceptron": 9, "act": 9, "batch": [9, 11], "item": [9, 10, 11], "flatten": [9, 11, 12], "major": [9, 11], "layout": [9, 11], "stdout": [9, 11], "syntax": [9, 11], "l": [9, 10, 11, 12], "cond": 9, "arg1": 9, "arg2": 9, "arr": [7, 9, 11], "atomic_add": [6, 9, 10, 11], "atom": [9, 11], "fabricarrai": 9, "indexedfabricarrai": 9, "atomic_sub": 9, "subtract": [9, 11], "atomic_min": 9, "updat": [9, 10, 11, 12], "atomic_max": 9, "lerp": 9, "factor": [9, 12, 13], "smoothstep": 9, "edge0": 9, "edge1": 9, "smoothli": 9, "cubic": 9, "hermit": 9, "expect_near": 9, "toler": 9, "magnitud": 9, "bvh_query_aabb": [9, 11], "bvh_query_t": 9, "box": [9, 11, 12], "queri": [1, 8, 9, 11], "against": [9, 11], "bvh": [4, 9], "track": [9, 11], "travers": 9, "bvh_query_rai": [9, 11], "start": [6, 9, 11, 12, 13], "dir": [9, 11], "rai": 9, "intersect": [9, 11], "bvh_query_next": [9, 11], "move": [8, 9, 11, 12], "next": [9, 11], "mesh_query_point": 9, "max_dist": [9, 12], "face": [8, 9, 11, 12], "bary_u": 9, "bary_v": 9, "closest": [8, 9, 11], "found": [5, 9], "outsid": [9, 11], "rel": [9, 12], "robust": 9, "increas": 9, "cost": 9, "abov": [9, 10, 11], "watertight": 9, "barycentr": [9, 11], "mesh_query_point_no_sign": 9, "mesh_query_point_sign_norm": 9, "epsilon": 9, "pseudo": 9, "approach": [9, 11], "self": [9, 11, 12], "vertex": [8, 9, 11, 12], "1e": [9, 12], "mesh_query_point_sign_winding_numb": [9, 11], "accuraci": [9, 11], "threshold": [1, 9], "wind": [9, 12], "approxim": 9, "accur": [1, 9, 11], "suport_winding_numb": [], "correct": [9, 11], "dipol": 9, "mesh_query_rai": [9, 11], "max_t": 9, "hit": [9, 11], "along": [8, 9, 11, 12], "front": 9, "mesh_query_aabb": 9, "mesh_query_aabb_t": 9, "mesh_query_aabb_next": 9, "mesh_eval_posit": 9, "mesh_eval_veloc": 9, "veloc": [9, 11, 12], "hash_grid_queri": [9, 11], "hash_grid_query_t": 9, "withing": [], "radiu": [9, 11, 12], "hash_grid_query_next": [9, 11], "hash_grid_point_id": 9, "occur": [5, 9, 11], "coher": 9, "intersect_tri_tri": 9, "v0": 9, "v1": 9, "v2": [9, 11], "u0": 9, "u1": 9, "u2": 9, "moller": 9, "mesh_get": 9, "mesh_eval_face_norm": 9, "mesh_get_point": 9, "mesh_get_veloc": 9, "mesh_get_index": 9, "closest_point_edge_edg": 9, "p1": 9, "p2": 9, "q2": 9, "degener": 9, "volume_sample_f": [9, 11], "uvw": 9, "sampling_mod": 9, "volume_sample_grad_f": 9, "volume_lookup_f": 9, "voxel": [9, 11], "volume_store_f": [9, 11], "volume_sample_v": 9, "volume_lookup_v": 9, "volume_store_v": [9, 11], "volume_sample_i": 9, "volume_lookup_i": 9, "volume_store_i": [9, 11], "volume_index_to_world": 9, "intrins": 9, "volume_world_to_index": [9, 11], "xyz": 9, "volume_index_to_world_dir": 9, "volume_world_to_index_dir": 9, "rand_init": 9, "seed": 9, "32": [6, 9, 11], "bit": [6, 9, 11], "rng": [9, 11], "offset": [9, 11, 12, 13], "uncorrel": 9, "randi": 9, "randf": 9, "randn": [9, 10], "sample_cdf": 9, "cdf": 9, "cumul": 9, "sample_triangl": 9, "sample_unit_r": 9, "ring": 9, "xy": 9, "plane": [9, 12], "sample_unit_disk": 9, "disk": [9, 11], "sample_unit_sphere_surfac": 9, "sphere": [9, 11, 12], "surfac": [9, 12], "sample_unit_spher": 9, "sample_unit_hemisphere_surfac": 9, "hemispher": 9, "sample_unit_hemispher": 9, "sample_unit_squar": 9, "sample_unit_cub": 9, "cube": [8, 9, 11], "poisson": 9, "lam": 9, "param": 12, "nois": 9, "period": 9, "perlin": 9, "xyzt": 9, "pnois": 9, "px": [9, 12], "pz": 9, "pt": 9, "curlnois": 9, "free": [9, 11, 12], "curl": [8, 9], "three": [8, 9, 12, 13], "lower_bound": 9, "search": 9, "sort": 9, "arr_begin": 9, "arr_end": 9, "sub": 9, "bit_and": 9, "bit_or": 9, "bit_xor": 9, "lshift": 9, "rshift": 9, "invert": [9, 12], "mul": 9, "mod": 9, "floordiv": 9, "unot": 9, "footnot": 9, "backpropag": 9, "protocol": [10, 11], "temporari": [8, 10, 13], "buffer": [8, 10, 11, 12, 13], "__array_interface__": 10, "helper": [10, 11, 12], "shown": [10, 11], "autograd": 10, "loss": [10, 11], "written": [10, 11], "adam": 10, "ndim": [10, 11], "requires_grad": [8, 10, 11, 12], "100": [10, 11, 12], "opt": 10, "lr": 10, "wp_x": 10, "wp_l": 10, "tape": [10, 11], "record": [10, 11], "500": [10, 12], "backward": [1, 10, 11], "now": [10, 11], "popul": [8, 10, 13], "therebi": 10, "just": 10, "zero_": [10, 11], "tloss": 10, "declar": [10, 11], "them": [4, 8, 10, 11], "revisit": 10, "device_from_torch": 10, "torch_devic": 10, "device_to_torch": 10, "wp_devic": 10, "dtype_from_torch": 10, "torch_dtyp": 10, "dtype_is_compat": 10, "warp_dtyp": 10, "stream_from_torch": 10, "stream_or_devic": 10, "stream": [10, 11], "stream_to_torch": 10, "intern": [10, 11, 12], "exchang": 10, "device_to_jax": 10, "device_from_jax": 10, "jax_devic": 10, "to_jax": 10, "wp_arrai": 10, "from_jax": 10, "jax_arrai": 10, "device_to_dlpack": 10, "dldevic": 10, "dtype_to_dlpack": 10, "wp_dtype": 10, "dldatatyp": 10, "dtype_from_dlpack": 10, "dl_dtype": 10, "device_from_dlpack": 10, "dl_devic": 10, "shape_to_dlpack": 10, "strides_to_dlpack": 10, "stride": [10, 11], "to_dlpack": 10, "np_arrai": 10, "ndarrai": [10, 11, 12], "pycapsul": 10, "dlmanagedtensor": 10, "from_dlpack": 10, "gain": [9, 11, 12], "basic": 11, "insight": 11, "point_radiu": 11, "printout": 11, "took": 11, "06": 11, "statement": 11, "mislead": 11, "use_nvtx": [5, 11], "start_rang": 11, "end_rang": 11, "cyan": 11, "nsight": 11, "timelin": 11, "nsy": 11, "warp_appl": 11, "demonstr": 11, "argb": 11, "recogn": 11, "frequent": 11, "timer": 11, "dict": [8, 11, 12], "rapid": 11, "elaps": 11, "sy": 11, "cprofil": 11, "print_stat": 11, "append": [7, 11], "0x00ffff": 11, "associ": 11, "durat": [11, 12], "deal": 11, "intend": [11, 12], "interpret": 11, "complex_kernel": 11, "128": 11, "128x128x3": [], "serial": 11, "adj_input": 11, "adj_output": 11, "adjoint": 11, "record_tap": 11, "record_cmd": 11, "forward": [4, 11], "until": 11, "cmd": 11, "fundament": 11, "incompat": 11, "host_arrai": 11, "addition": 11, "src_arrai": 11, "dest_arrai": 11, "empty_lik": 11, "1024x16": 11, "16": [1, 11], "hint": 11, "array2d": 11, "array3d": 11, "slice": 11, "concaten": 11, "datatyp": 11, "pin": [8, 11], "kwarg": 11, "propag": 11, "src": [11, 13], "templat": 11, "dest_offset": 11, "src_offset": 11, "content": [11, 12, 13], "destin": 11, "least": [8, 11, 13], "big": 11, "ptr": 11, "capac": [8, 11], "owner": 11, "sequenti": 11, "address": 11, "resid": 11, "neither": [8, 11], "encourag": 11, "member": [6, 11], "taken": 11, "account": 11, "union": [11, 12, 13], "byte": 11, "success": [8, 11], "deprec": [8, 11], "devicelik": 11, "incom": 11, "alreadi": [1, 8, 11], "try": 11, "dealloc": 11, "delet": 11, "cutlass": 11, "fall": 11, "matmul": 11, "loss_kernel": 11, "a_grad": 11, "algorithm": 11, "allow_tf32x3_arith": 11, "gemm": 11, "3xtf32": 11, "fp32": 11, "resort": 11, "unsign": 11, "short": [6, 11], "long": [8, 11], "precis": 11, "numer": 11, "concret": [11, 12], "spatial_vector": 11, "algebra": [11, 13], "done": 11, "hold": [8, 11, 12], "vec5d": 11, "onc": [11, 12], "anonym": 11, "potenti": 11, "spatial_matrix": 11, "minv": 11, "manner": 11, "3x2": 11, "mat32h": 11, "v3": 11, "5x5": 11, "As": 11, "30": [5, 11], "qd": 11, "qf": 11, "qh": 11, "7d": 11, "ignor": [11, 12], "mystruct": 11, "param1": 11, "param2": 11, "param3": 11, "attribut": [8, 11, 12], "Or": 5, "strict": 11, "regard": 11, "respons": 11, "surpris": 11, "accustom": 11, "fold": 11, "type_spher": 11, "type_cub": 11, "type_capsul": 11, "collid": [11, 12], "capsul": [11, 12], "circuit": [6, 11], "secondari": 11, "safe": [8, 11], "strictli": 11, "modulu": 11, "maintain": 11, "geometr": 11, "raycast": 11, "ray_origin": 11, "ray_dir": 11, "ray_hit": 11, "got": 11, "modifi": [11, 13], "refit": 11, "rebuild": 11, "hierarchi": 4, "topologi": [6, 8, 11], "support_winding_numb": [9, 11], "num_tri": [8, 11], "datastructur": 11, "incredibli": 11, "sdf": [11, 12], "volumetr": [11, 12], "nanovdb": 11, "trilinear": 11, "nvdb": 11, "load_from_nvdb": 11, "give": 11, "open": [6, 8, 11], "mygrid": 11, "rb": 11, "commonli": 11, "mark": 11, "compress": [11, 13], "therefor": 11, "uncompress": 11, "zip": 11, "sample_grid": 11, "futur": [], "vdb": [], "openvdb": [], "houdini": [], "enum": 11, "classmethod": 11, "file_or_buff": 11, "load_from_numpi": 11, "min_world": 11, "voxel_s": 11, "bg_valu": 11, "dens": [8, 11], "corner": 11, "points_in_world_spac": 11, "enough": [11, 12], "inclus": 11, "those": [8, 11], "tile": [6, 11], "8x8x8": 11, "request": [6, 11], "unalloc": 11, "allocate_by_til": 11, "tile_point": 11, "repeat": 11, "effici": [4, 11, 12], "dedupl": 11, "particl": [11, 12], "dem": 11, "hydrodynam": 11, "sph": 11, "involv": 11, "interact": [11, 12], "establish": 11, "suit": 11, "hashgrid": [9, 11, 12], "dim_x": [11, 12], "dim_i": [11, 12], "dim_z": [11, 12], "sum": 11, "dist": 11, "conflict": 11, "doesn": 11, "itself": 11, "would": 11, "redund": 11, "twice": 11, "width": [11, 12], "particip": 11, "chain": [11, 12], "compute1": 11, "compute2": 11, "revers": 11, "wish": 11, "keep": [11, 12], "loop": [1, 7, 8, 11, 12], "mutat": 11, "previous": 11, "separ": [6, 11, 12], "unrol": [1, 11], "overwrit": 11, "whose": [11, 12], "record_func": 11, "reset": 11, "clear": [0, 11], "mathbb": 11, "entir": [1, 11], "mathbf": 11, "output_dim": 11, "input_dim": 11, "output_buff": 11, "launch_kernels_to_be_differenti": 11, "input_buff": 11, "output_index": 11, "select_index": 11, "q_grad_i": 11, "independ": 11, "begin": 11, "smallmatrix": 11, "end": [5, 11, 12, 13], "num_env": [11, 12], "precomput": 11, "reshap": 11, "prime": 11, "green": 11, "underset": 11, "textrm": 11, "blue": 11, "autodiff": 11, "in1": 11, "intype1": 11, "inn": 11, "intypen": 11, "outtype1": 11, "outtypem": 11, "out1": 11, "outm": 11, "func_replai": 11, "replay_myfunc": 11, "go": 11, "func_grad": 11, "adj_myfunc": 11, "adj_out1": 11, "adj_outm": 11, "goe": 11, "yet": [11, 12], "safe_sqrt": 11, "run_safe_sqrt": 11, "yield": 11, "4142135": 11, "35355338": 11, "inf": [11, 12], "desir": [8, 11, 12], "catch": 11, "nonfinit": 11, "caus": 11, "adj_safe_sqrt": 11, "adj_ret": 11, "plu": 11, "increment": [11, 12], "test_add": 11, "counter": [8, 11, 12], "idx": 11, "use_reversible_incr": 11, "thread_id": 11, "test_add_diff": 11, "els": [11, 12], "__name__": 11, "__main__": 11, "7320508": 11, "236068": 11, "4494898": 11, "6457512": 11, "828427": 11, "incorrect": 11, "know": 11, "alwai": 11, "entri": [11, 12], "frac": [9, 11], "henc": 11, "reversible_incr": 11, "buf": 11, "buf_index": 11, "thread_valu": 11, "next_index": 11, "replay_reversible_incr": 11, "phase": 11, "That": 11, "reproduc": 11, "exactli": 11, "operand": 11, "28867513": 11, "2236068": 11, "20412414": 11, "18898225": 11, "17677669": 11, "introduc": 11, "veri": 11, "littl": 11, "capture_begin": 11, "capture_end": 11, "captur": [1, 11], "capture_launch": 11, "reus": [8, 11, 13], "force_module_load": 11, "better": 11, "load_modul": 11, "rapidli": 12, "progress": 12, "modelbuild": 12, "up_vector": 12, "graviti": 12, "9": 12, "80665": 12, "scene": 12, "final": [11, 12], "builder": [8, 12], "anchor": 12, "add_particl": 12, "add_spr": 12, "semiimplicitintegr": 12, "clear_forc": 12, "dt": 12, "60": 12, "add_build": 12, "articul": [], "xform": 12, "update_num_env_count": 12, "separate_collision_group": 12, "collis": 12, "group": [11, 12], "add_bodi": 12, "armatur": 12, "i_m": 12, "frame": 12, "artifici": 12, "inertia": 12, "ad": 12, "center": 12, "kinemat": 4, "add_joint": 12, "joint_typ": 12, "parent": [6, 12], "child": [6, 12], "linear_ax": 12, "angular_ax": 12, "parent_xform": 12, "child_xform": 12, "linear_compli": 12, "angular_compli": 12, "collision_filter_par": 12, "jointaxi": 12, "ax": [11, 12], "angular": 12, "complianc": 12, "filter": 12, "add_joint_revolut": 12, "target_k": 12, "target_kd": 12, "limit_low": 12, "283185307179586": 12, "limit_upp": 12, "limit_k": 12, "limit_kd": 12, "revolut": 12, "stiff": 12, "damp": 12, "add_joint_prismat": 12, "10000": 12, "prismat": 12, "add_joint_bal": 12, "ball": 12, "add_joint_fix": 12, "add_joint_fre": 12, "add_joint_dist": 12, "min_dist": [9, 12], "add_joint_univers": 12, "axis_0": 12, "axis_1": 12, "univers": 12, "add_joint_compound": 12, "axis_2": 12, "compound": 12, "third": 12, "add_joint_d6": 12, "collapse_fixed_joint": 12, "verbos": [1, 12], "remov": [8, 12], "merg": 12, "connect": [8, 12], "add_muscl": 12, "f0": 12, "lm": 12, "lt": 12, "lmax": 12, "pen": 12, "muscl": 12, "tendon": 12, "waypoint": 12, "maxim": 12, "add_shape_plan": 12, "ke": 12, "100000": 12, "kd": [8, 12], "1000": [7, 12], "kf": 12, "mu": 12, "restitut": 12, "thick": 12, "has_ground_collis": 12, "extent": 12, "infinit": 12, "friction": 12, "coeffici": [12, 13], "ground": 12, "add_shape_spher": 12, "densiti": 12, "is_solid": 12, "05": 12, "solid": 12, "hollow": 12, "add_shape_box": 12, "hx": 12, "hy": 12, "hz": 12, "add_shape_capsul": 12, "half_height": 12, "up_axi": 12, "cylind": 12, "add_shape_cylind": 12, "add_shape_con": 12, "cone": 12, "add_shape_mesh": 12, "wall": 12, "add_shape_sdf": 12, "vel": 12, "flag": [8, 12], "particle_flag_": 12, "default_particle_radiu": 12, "subject": 12, "spring": 12, "actuat": 12, "rest": 12, "add_triangl": 12, "tri_k": 12, "tri_ka": 12, "tri_kd": 12, "tri_drag": 12, "tri_lift": 12, "viscoelast": 12, "add_tetrahedron": 12, "k_mu": 12, "k_lambda": 12, "k_damp": 12, "tetrahedr": [8, 12], "four": [6, 12], "tetrahedra": 12, "neohookean": 12, "energi": 12, "smith": 12, "2018": 12, "fourth": 12, "lame": 12, "pose": 12, "configru": [], "add_edg": 12, "edge_k": 12, "edge_kd": 12, "bend": 12, "bridson": 12, "2002": 12, "tri_kb": 12, "across": [8, 12, 13], "li": 12, "oppos": 12, "clockwis": [8, 12], "add_cloth_grid": 12, "cell_x": 12, "cell_i": 12, "reverse_wind": 12, "fix_left": 12, "fix_right": 12, "fix_top": 12, "fix_bottom": 12, "planar": 12, "cloth": 12, "rectangular": 12, "flip": 12, "place": 12, "add_cloth_mesh": 12, "edge_callback": 12, "face_callback": 12, "callback": 12, "manifold": 12, "add_soft_grid": 12, "cell_z": 12, "beam": 12, "sheet": 12, "hexahedr": [8, 12], "decompos": 12, "add_soft_mesh": 12, "set_ground_plan": 12, "readi": 12, "vari": 12, "descript": [1, 12], "constraint": 12, "were": 12, "particle_q": 12, "particle_count": 12, "particle_qd": 12, "particle_mass": 12, "particle_inv_mass": 12, "particle_radiu": 12, "particle_max_radiu": 12, "particle_k": 12, "particle_kd": 12, "particle_kf": 12, "particle_mu": 12, "particle_cohes": 12, "cohes": 12, "strength": 12, "particle_adhes": 12, "adhes": 12, "particle_grid": 12, "particle_flag": 12, "particle_max_veloc": 12, "prevent": 12, "instabl": 12, "shape_transform": 12, "shape_count": 12, "shape_bodi": 12, "body_shap": 12, "shape_materi": 12, "materi": 12, "modelshapemateri": 12, "shape_shape_geo": 12, "modelshapegeometri": 12, "shape_geo_src": 12, "shape_collision_group": 12, "shape_collision_group_map": 12, "shape_collision_filter_pair": 12, "shape_collision_radiu": 12, "broadphas": 12, "shape_ground_collis": 12, "shape_contact_pair": 12, "contact_pair_count": 12, "shape_ground_contact_pair": 12, "ground_contact_pair_count": 12, "spring_indic": 12, "spring_count": 12, "spring_rest_length": 12, "spring_stiff": 12, "spring_damp": 12, "spring_control": 12, "tri_indic": 12, "tri_count": 12, "tri_pos": 12, "tri_activ": 12, "tri_materi": 12, "edge_indic": 12, "edge_count": 12, "edge_rest_angl": 12, "edge_bending_properti": 12, "tet_indic": 12, "tet_count": 12, "tet_pos": 12, "tet_activ": 12, "tet_materi": 12, "k_": 12, "body_q": 12, "body_count": 12, "body_qd": 12, "body_com": 12, "body_inertia": 12, "body_inv_inertia": 12, "body_mass": 12, "body_inv_mass": 12, "body_nam": 12, "joint_q": 12, "joint_coord_count": 12, "joint_qd": 12, "joint_dof_count": 12, "joint_act": 12, "joint_child": 12, "joint_x_p": 12, "joint_x_c": 12, "joint_axi": 12, "joint_axis_count": 12, "joint_armatur": 12, "joint_target": 12, "joint_target_k": 12, "joint_target_kd": 12, "joint_axis_start": 12, "joint_axis_dim": 12, "joint_axis_mod": 12, "joint_linear_compli": 12, "joint_angular_compli": 12, "joint_en": 12, "joint_limit_low": 12, "joint_limit_upp": 12, "joint_limit_k": 12, "joint_limit_kd": 12, "joint_twist_low": 12, "twist": 12, "joint_twist_upp": 12, "joint_q_start": 12, "articulation_start": 12, "articulation_count": 12, "joint_nam": 12, "joint_attach_k": 12, "joint_attach_kd": 12, "soft_contact_margin": 12, "margin": 12, "soft": 12, "soft_contact_k": 12, "soft_contact_kd": 12, "soft_contact_kf": 12, "soft_contact_mu": 12, "soft_contact_restitut": 12, "xpbdintegr": 12, "rigid_contact_margin": 12, "rigid_contact_torsional_frict": 12, "torsion": 12, "rigid_contact_rolling_frict": 12, "ground_plan": 12, "contact_count": 12, "muscle_count": 12, "count_contact_point": 12, "obtain": [11, 12], "soft_contact_max": 12, "radii": [], "driven": 12, "toward": 12, "proport": 12, "drive": 12, "pd": 12, "joint_prismat": 12, "slider": 12, "joint_revolut": 12, "hing": 12, "joint_bal": 12, "spheric": 12, "joint_fix": 12, "joint_fre": 12, "joint_compound": 12, "joint_univers": 12, "joint_dist": 12, "joint_d6": 12, "d6": 12, "joint_mode_limit": 12, "joint_mode_target_posit": 12, "joint_mode_target_veloc": 12, "anyth": 12, "exact": 12, "particle_f": 12, "body_f": 12, "angular_damp": 12, "symplect": 12, "euler": 12, "variat": 12, "uncondition": 12, "stabl": 12, "small": 12, "en": 12, "wikipedia": 12, "org": 12, "wiki": 12, "implicit_euler_method": 12, "state_in": 12, "state_out": 12, "soft_body_relax": 12, "soft_contact_relax": 12, "joint_linear_relax": 12, "joint_angular_relax": 12, "rigid_contact_relax": 12, "rigid_contact_con_weight": 12, "enable_restitut": 12, "xpbd": 12, "urdf": 12, "mujoco": 12, "mjcf": 12, "parse_urdf": 12, "urdf_filenam": 12, "base_joint": 12, "shape_k": 12, "shape_kd": 12, "shape_kf": 12, "shape_mu": 12, "shape_restitut": 12, "shape_thick": 12, "parse_visuals_as_collid": 12, "enable_self_collis": 12, "ignore_inertial_definit": 12, "ensure_nonstatic_link": 12, "static_link_mass": 12, "01": 12, "pars": 12, "filenam": 12, "comma": 12, "rz": 12, "kg": 12, "stabil": 12, "tag": [8, 12], "geoemtri": 12, "link": 12, "parse_mjcf": 12, "mjcf_filenam": 12, "contact_k": 12, "contact_kd": 12, "contact_kf": 12, "contact_mu": 12, "contact_restitut": 12, "armature_scal": 12, "parse_mesh": 12, "ignore_class": 12, "xml": 12, "experiment": 12, "parse_usd": [5, 12], "default_dens": 12, "only_load_enabled_rigid_bodi": 12, "only_load_enabled_joint": 12, "default_k": 12, "default_kd": 12, "250": [7, 12], "default_kf": 12, "default_mu": 12, "default_restitut": 12, "default_thick": 12, "ignore_path": 12, "export_usda": 12, "usdphys": 12, "schema": 12, "rigidbodyen": 12, "jointen": 12, "prim": 12, "export": [], "download": 12, "usda": 12, "fp": 12, "stage": 12, "report": 12, "manipul": 13, "aris": 13, "bsr": 13, "csr": 13, "1x1": 13, "untyp": 13, "bsr_zero": 13, "nrow": 13, "ncol": 13, "nnz": 13, "og": [], "scalar_typ": 13, "FOr": [], "block_shap": 13, "block_siz": 13, "columsn": 13, "rows_of_block": 13, "cols_of_block": 13, "block_typ": 13, "bsr_set_from_triplet": 13, "fill": [11, 13], "coo": 13, "triplet": 13, "0x000001c0ff88a5c0": [], "0x000001c0ff88a650": [], "0x000001c0ff88b790": [], "bsr_assign": 13, "bsr_copi": 13, "bsr_set_transpos": 13, "bsr_transpos": 13, "bsr_get_diag": 13, "bsr_set_diag": 13, "bsr_diag": 13, "bsr_mm": 13, "bsr_mv": 13, "naiv": [], "percept": 4, "example_apic_fluid": 8, "at_nod": 8, "elementkind": 8, "kind": 8, "element_kind": 8, "geometry_element_count": [], "fill_element_mask": [], "mask": 12, "voigt": 8, "enumer": [], "off": [], "db16": [], "tangenti": [], "0x000002434ee4b160": [], "spring_k": 12, "spring_kd": 12, "0x0000024335a2dc00": [], "0x000002434ee3d0f0": [], "0x000002434ee3d120": [], "0x000001b3cd9386d0": [], "0x000001b3cd828a90": [], "0x000001b3cd828ac0": [], "0x000001b3cd829870": [], "o": [0, 1], "example_dir": [0, 1], "dirnam": [0, 1], "realpath": [0, 1], "__file__": [0, 1], "kernel_cache_dir": [0, 1], "join": [0, 1], "tmp": [0, 1], "warpcache1": [0, 1], "init_kernel_cach": 0, "warpcache2": 0, "fresh": 0, "clear_kernel_cach": 0, "tune": [1, 11], "enable_backward": [1, 8], "preced": 1, "prepend": 1, "fly": 1, "prior": 1, "verify_fp": 1, "boolean": 1, "implic": 1, "quiet": 1, "messag": 1, "disabl": [1, 11, 12], "affect": 1, "warn": [1, 5], "subdirectori": 1, "appdir": 1, "user_cache_directori": 1, "cache_kernel": 1, "recompil": 1, "cuda_output": 1, "prefer": 1, "cubin": 1, "ptx_target_arch": 1, "70": 1, "architectur": 1, "llvm_cuda": 1, "clang": 1, "llvm": 1, "nvtrc": 1, "set_module_opt": 1, "get_module_opt": 1, "max_unrol": 1, "amount": 1, "fast_math": 1, "use_fast_math": 1, "scale_2": 1, "test_debug": 2, "later": 5, "usdrender": 5, "copybutton": 5, "summar": 6, "discuss": 6, "256": [6, 11], "31": [6, 11], "fork": 6, "spawn": 6, "workaround": 6, "inc": 7, "get_cuda_devic": 7, "device_count": 7, "_": 7, "boundary_form": 8, "q_": 8, "serendip": 8, "s_k": 8, "leq": 8, "q_k": 8, "tight": 8, "de": 8, "degrad": 8, "overcom": 8, "temporarystor": 8, "persist": [8, 13], "set_default_temporary_stor": 8, "granular": 8, "div_out": 8, "temporary_stor": 8, "kernel_opt": 8, "fieldlik": 8, "pool": 8, "notr": 8, "overload": 8, "bounds_lo": 8, "bounds_hi": 8, "resolut": 8, "bounds_up": 8, "trimesh2d": 8, "tri_vertex_indic": 8, "tri": 8, "num_vertic": 8, "grid3d": 8, "tetmesh": 8, "tet_vertex_indic": 8, "num_tet": 8, "tet": 8, "frontiersid": 8, "interv": [8, 12], "gauss_legendr": 8, "gauss": 8, "legendr": 8, "endpoint": 8, "lobatto_gauss_legendr": 8, "lobatto": 8, "equispaced_clos": 8, "equispaced_open": 8, "regularquadratur": 8, "nodalquadratur": 8, "od": 8, "assumpt": 8, "thu": 8, "element_basi": 8, "colloc": 8, "make_polynomial_basis_spac": 8, "make_collocated_function_spac": 8, "elementbasi": 8, "collocatedfunctionspac": 8, "basisspac": 8, "basis_spac": 8, "space_topologi": 8, "spacetopologi": 8, "simplic": 8, "nonconforming_polynomi": 8, "simplex": 8, "embed": 8, "conform": 8, "skewsymmetrictensormapp": 8, "normalize_dirichlet_projector": 8, "projector_matrix": 8, "fixed_valu": 8, "projector": 8, "becom": 8, "idempot": 8, "project_linear_system": 8, "system_matrix": 8, "system_rh": 8, "normalize_projector": 8, "enforc": 8, "borrow_temporari": 8, "borrow": 8, "suffici": 8, "wil": [8, 13], "momori": 8, "borrow_temporary_lik": 8, "meant": [8, 11], "extend": 8, "space_arg": 8, "pattern": [8, 11], "valuat": 8, "element_node_indic": 8, "tracespacetopologi": 8, "ingredi": 8, "shapefunct": 8, "spacefield": 8, "invoc": 8, "destruct": 8, "detach": 8, "cbrt": 9, "arctang": 9, "logarithm": 9, "arg_typ": 9, "mesh_query_furthest_point_no_sign": 9, "furthest": 9, "reorder": 9, "reserv": 9, "octav": 9, "lacunar": 9, "max_block": 11, "hardwar": 11, "exce": 11, "abl": 11, "var": 11, "fill_": 11, "valueerror": 11, "cptr": 11, "op": 11, "collaps": 11, "permut": 11, "full_lik": 11, "teststruct": 11, "test_kernel": 11, "reenabl": 11, "benefici": 11, "default_rng": 11, "123": 11, "num_bound": 11, "device_low": 11, "device_upp": 11, "bvh_id": 11, "bounds_intersect": 11, "bounds_nr": 11, "query_start": 11, "query_dir": 11, "query_low": 11, "query_upp": 11, "nearli": 11, "recreat": 11, "raw": 11, "get_til": 11, "get_voxel_s": 11, "slide": 12, "Its": 12, "xyzw": 12, "improv": 12, "notat": 12, "children": 12, "add_ball_joint": 12, "tait": 12, "bryan": 12, "compute_inertia": 12, "load_mesh": 12, "use_meshio": 12, "meshio": 12, "mesh_point": 12, "mesh_indic": 12, "openmesh": 12, "read_trimesh": 12, "face_vertex_indic": 12, "x_wp": 12, "x_wc": 12, "x_pj": 12, "x_cj": 12, "x_j": 12, "x_": 12, "wc": 12, "cdot": 12, "pj": 12, "cj": 12, "eval_fk": 12, "fk": 12, "eval_ik": 12, "blocktyp": 13, "_matrixblocktyp": 13, "col": 13, "_scalarblocktyp": 13, "mat": 13, "bsr_set_zero": 13, "distinct": 13, "chane": 13, "constitut": 13, "bsr_set_ident": 13, "resiz": 13, "bsr_ident": 13, "newli": 13, "bsr_scale": 13, "bsr_axpy_work_arrai": 13, "opaqu": 13, "work_arrai": 13, "mutabl": 13, "uniform": 13, "bsr_mm_work_arrai": 13, "work_buff": 13, "conda": 4, "latest": 5, "sure": 5, "older": 5, "newer": 5, "oserror": 5, "libstdc": 5, "glibcxx_3": 5, "forg": 5, "libstdcxx": 5, "ng": 5, "toolchain": 5, "downgrad": 5, "deactiv": 5, "quadrilater": 8, "hexahedron": 8, "deform": 8, "displac": 8, "make_deformed_geometri": 8, "base_geo": 8, "deformation_spac": 8, "deformation_degre": 8, "deformation_field": 8, "deformation_field_expr": 8, "deform_geo": 8, "deformed_geo": 8, "scalar_space_degre": 8, "example_deformed_geometri": 8, "trianglemesh2d": 8, "deformation_gradi": 8, "quadmesh2d": 8, "quad_vertex_indic": 8, "quad": 8, "hexmesh": 8, "hex_vertex_indic": 8, "num_hex": 8, "hex": 8, "similarli": [8, 11], "func_n": 11, "__shared__": 11, "__syncthread": 11, "reverse_kernel": 11, "notic": 11, "readili": 11, "won": 11, "leverag": 11, "pure": 11, "bear": 11, "mind": 11, "stub": 11, "rewind": 11, "alongsid": 11, "adj_snippet": 11, "adj_a": 11, "adj_out": 11, "adj_x": 11, "adj_i": 11, "saxpi": 11, "saxpy_kernel": 11, "x1": 11, "invert_rot": 12, "usdstag": 12, "pxr": 12, "path_shape_map": 12, "usdgeom": 12, "path_body_map": 12, "physicsrigidbodyapi": 12, "path_shape_scal": 12, "mass_unit": 12, "kilogram": 12, "kgpu": 12, "linear_unit": 12, "meter": 12, "mpu": 12, "resolve_usd_from_url": 12, "target_folder_nam": 12, "timestamp": 12, "save": 12, "closur": 11, "might": 11, "varieti": 11, "particular": 11, "make_kernel": 11, "closure_kernel_fn": 11, "practic": 11, "sqr": 11, "sqr_float": 11, "cube_doubl": 11, "data_float": 11, "data_doubl": 11, "out_float": 11, "out_doubl": 11, "make_add_kernel": 11, "add_ones_int": 11, "add_on": 11, "add_ones_vec3": 11, "a_out": 11, "b_out": 11}, "objects": {"warp": [[9, 0, 1, "", "Array"], [11, 0, 1, "", "Bvh"], [9, 0, 1, "", "Float"], [11, 0, 1, "", "HashGrid"], [9, 0, 1, "", "Int"], [9, 0, 1, "", "Matrix"], [11, 0, 1, "", "Mesh"], [9, 0, 1, "", "Quaternion"], [9, 0, 1, "", "Scalar"], [11, 0, 1, "", "ScopedTimer"], [11, 0, 1, "", "Tape"], [9, 0, 1, "", "Transformation"], [9, 0, 1, "", "Vector"], [11, 0, 1, "", "Volume"], [9, 3, 1, "", "abs"], [9, 3, 1, "", "acos"], [9, 3, 1, "", "add"], [9, 3, 1, "", "argmax"], [9, 3, 1, "", "argmin"], [11, 0, 1, "", "array"], [9, 3, 1, "", "asin"], [9, 3, 1, "", "atan"], [9, 3, 1, "", "atan2"], [9, 3, 1, "", "atomic_add"], [9, 3, 1, "", "atomic_max"], [9, 3, 1, "", "atomic_min"], [9, 3, 1, "", "atomic_sub"], [9, 3, 1, "", "bit_and"], [9, 3, 1, "", "bit_or"], [9, 3, 1, "", "bit_xor"], [9, 0, 1, "", "bool"], [9, 3, 1, "", "breakpoint"], [9, 3, 1, "", "bvh_query_aabb"], [9, 3, 1, "", "bvh_query_next"], [9, 3, 1, "", "bvh_query_ray"], [11, 3, 1, "", "capture_begin"], [11, 3, 1, "", "capture_end"], [11, 3, 1, "", "capture_launch"], [9, 3, 1, "", "cbrt"], [9, 3, 1, "", "ceil"], [9, 3, 1, "", "clamp"], [11, 3, 1, "", "clone"], [9, 3, 1, "", "closest_point_edge_edge"], [11, 0, 1, "", "constant"], [11, 3, 1, "", "copy"], [9, 3, 1, "", "cos"], [9, 3, 1, "", "cosh"], [9, 3, 1, "", "cross"], [9, 3, 1, "", "curlnoise"], [9, 3, 1, "", "cw_div"], [9, 3, 1, "", "cw_mul"], [9, 3, 1, "", "ddot"], [9, 3, 1, "", "degrees"], [9, 3, 1, "", "determinant"], [9, 3, 1, "", "diag"], [9, 3, 1, "", "div"], [10, 5, 0, "-", "dlpack"], [9, 3, 1, "", "dot"], [9, 3, 1, "", "eig3"], [11, 3, 1, "", "empty"], [11, 3, 1, "", "empty_like"], [9, 3, 1, "", "exp"], [9, 3, 1, "", "expect_near"], [9, 0, 1, "", "float16"], [9, 0, 1, "", "float32"], [9, 0, 1, "", "float64"], [9, 3, 1, "", "floor"], [9, 3, 1, "", "floordiv"], [9, 3, 1, "", "frac"], [11, 3, 1, "", "full"], [11, 3, 1, "", "full_like"], [9, 3, 1, "", "get_diag"], [9, 3, 1, "", "hash_grid_point_id"], [9, 3, 1, "", "hash_grid_query"], [9, 3, 1, "", "hash_grid_query_next"], [9, 3, 1, "", "identity"], [9, 0, 1, "", "int16"], [9, 0, 1, "", "int32"], [9, 0, 1, "", "int64"], [9, 0, 1, "", "int8"], [9, 3, 1, "", "intersect_tri_tri"], [9, 3, 1, "", "inverse"], [9, 3, 1, "", "invert"], [10, 5, 0, "-", "jax"], [11, 3, 1, "", "launch"], [9, 3, 1, "", "length"], [9, 3, 1, "", "length_sq"], [9, 3, 1, "", "lerp"], [9, 3, 1, "", "log"], [9, 3, 1, "", "log10"], [9, 3, 1, "", "log2"], [9, 3, 1, "", "lower_bound"], [9, 3, 1, "", "lshift"], [9, 0, 1, "", "mat22d"], [9, 0, 1, "", "mat22f"], [9, 0, 1, "", "mat22h"], [9, 0, 1, "", "mat33d"], [9, 0, 1, "", "mat33f"], [9, 0, 1, "", "mat33h"], [9, 0, 1, "", "mat44d"], [9, 0, 1, "", "mat44f"], [9, 0, 1, "", "mat44h"], [11, 3, 1, "", "matmul"], [9, 3, 1, "", "matrix"], [9, 3, 1, "", "max"], [9, 3, 1, "", "mesh_eval_face_normal"], [9, 3, 1, "", "mesh_eval_position"], [9, 3, 1, "", "mesh_eval_velocity"], [9, 3, 1, "", "mesh_get"], [9, 3, 1, "", "mesh_get_index"], [9, 3, 1, "", "mesh_get_point"], [9, 3, 1, "", "mesh_get_velocity"], [9, 3, 1, "", "mesh_query_aabb"], [9, 3, 1, "", "mesh_query_aabb_next"], [9, 3, 1, "", "mesh_query_furthest_point_no_sign"], [9, 3, 1, "", "mesh_query_point"], [9, 3, 1, "", "mesh_query_point_no_sign"], [9, 3, 1, "", "mesh_query_point_sign_normal"], [9, 3, 1, "", "mesh_query_point_sign_winding_number"], [9, 3, 1, "", "mesh_query_ray"], [9, 3, 1, "", "min"], [9, 3, 1, "", "mlp"], [9, 3, 1, "", "mod"], [9, 3, 1, "", "mul"], [9, 3, 1, "", "neg"], [9, 3, 1, "", "noise"], [9, 3, 1, "", "nonzero"], [9, 3, 1, "", "normalize"], [9, 3, 1, "", "outer"], [9, 3, 1, "", "pnoise"], [9, 3, 1, "", "poisson"], [9, 3, 1, "", "pos"], [9, 3, 1, "", "pow"], [9, 3, 1, "", "print"], [9, 3, 1, "", "printf"], [9, 3, 1, "", "qr3"], [9, 3, 1, "", "quat_from_axis_angle"], [9, 3, 1, "", "quat_from_matrix"], [9, 3, 1, "", "quat_identity"], [9, 3, 1, "", "quat_inverse"], [9, 3, 1, "", "quat_rotate"], [9, 3, 1, "", "quat_rotate_inv"], [9, 3, 1, "", "quat_rpy"], [9, 3, 1, "", "quat_slerp"], [9, 3, 1, "", "quat_to_axis_angle"], [9, 3, 1, "", "quat_to_matrix"], [9, 0, 1, "", "quatd"], [9, 3, 1, "", "quaternion"], [9, 0, 1, "", "quatf"], [9, 0, 1, "", "quath"], [9, 3, 1, "", "radians"], [9, 3, 1, "", "rand_init"], [9, 3, 1, "", "randf"], [9, 3, 1, "", "randi"], [9, 3, 1, "", "randn"], [9, 3, 1, "", "rint"], [9, 3, 1, "", "round"], [9, 3, 1, "", "rshift"], [9, 3, 1, "", "sample_cdf"], [9, 3, 1, "", "sample_triangle"], [9, 3, 1, "", "sample_unit_cube"], [9, 3, 1, "", "sample_unit_disk"], [9, 3, 1, "", "sample_unit_hemisphere"], [9, 3, 1, "", "sample_unit_hemisphere_surface"], [9, 3, 1, "", "sample_unit_ring"], [9, 3, 1, "", "sample_unit_sphere"], [9, 3, 1, "", "sample_unit_sphere_surface"], [9, 3, 1, "", "sample_unit_square"], [9, 3, 1, "", "select"], [9, 3, 1, "", "sign"], [9, 3, 1, "", "sin"], [9, 3, 1, "", "sinh"], [9, 3, 1, "", "skew"], [9, 3, 1, "", "smoothstep"], [13, 5, 0, "-", "sparse"], [9, 3, 1, "", "spatial_adjoint"], [9, 3, 1, "", "spatial_bottom"], [9, 3, 1, "", "spatial_cross"], [9, 3, 1, "", "spatial_cross_dual"], [9, 3, 1, "", "spatial_dot"], [9, 3, 1, "", "spatial_jacobian"], [9, 3, 1, "", "spatial_mass"], [9, 0, 1, "", "spatial_matrixd"], [9, 0, 1, "", "spatial_matrixf"], [9, 0, 1, "", "spatial_matrixh"], [9, 3, 1, "", "spatial_top"], [9, 0, 1, "", "spatial_vectord"], [9, 0, 1, "", "spatial_vectorf"], [9, 0, 1, "", "spatial_vectorh"], [9, 3, 1, "", "sqrt"], [9, 3, 1, "", "step"], [9, 3, 1, "", "sub"], [9, 3, 1, "", "svd3"], [9, 3, 1, "", "tan"], [9, 3, 1, "", "tanh"], [9, 3, 1, "", "tid"], [10, 5, 0, "-", "torch"], [9, 3, 1, "", "trace"], [9, 3, 1, "", "transform_get_rotation"], [9, 3, 1, "", "transform_get_translation"], [9, 3, 1, "", "transform_identity"], [9, 3, 1, "", "transform_inverse"], [9, 3, 1, "", "transform_multiply"], [9, 3, 1, "", "transform_point"], [9, 3, 1, "", "transform_vector"], [9, 3, 1, "", "transformation"], [9, 0, 1, "", "transformd"], [9, 0, 1, "", "transformf"], [9, 0, 1, "", "transformh"], [9, 3, 1, "", "transpose"], [9, 3, 1, "", "trunc"], [9, 0, 1, "", "uint16"], [9, 0, 1, "", "uint32"], [9, 0, 1, "", "uint64"], [9, 0, 1, "", "uint8"], [9, 3, 1, "", "unot"], [9, 0, 1, "", "vec2b"], [9, 0, 1, "", "vec2d"], [9, 0, 1, "", "vec2f"], [9, 0, 1, "", "vec2h"], [9, 0, 1, "", "vec2i"], [9, 0, 1, "", "vec2l"], [9, 0, 1, "", "vec2s"], [9, 0, 1, "", "vec2ub"], [9, 0, 1, "", "vec2ui"], [9, 0, 1, "", "vec2ul"], [9, 0, 1, "", "vec2us"], [9, 0, 1, "", "vec3b"], [9, 0, 1, "", "vec3d"], [9, 0, 1, "", "vec3f"], [9, 0, 1, "", "vec3h"], [9, 0, 1, "", "vec3i"], [9, 0, 1, "", "vec3l"], [9, 0, 1, "", "vec3s"], [9, 0, 1, "", "vec3ub"], [9, 0, 1, "", "vec3ui"], [9, 0, 1, "", "vec3ul"], [9, 0, 1, "", "vec3us"], [9, 0, 1, "", "vec4b"], [9, 0, 1, "", "vec4d"], [9, 0, 1, "", "vec4f"], [9, 0, 1, "", "vec4h"], [9, 0, 1, "", "vec4i"], [9, 0, 1, "", "vec4l"], [9, 0, 1, "", "vec4s"], [9, 0, 1, "", "vec4ub"], [9, 0, 1, "", "vec4ui"], [9, 0, 1, "", "vec4ul"], [9, 0, 1, "", "vec4us"], [9, 3, 1, "", "vector"], [9, 3, 1, "", "volume_index_to_world"], [9, 3, 1, "", "volume_index_to_world_dir"], [9, 3, 1, "", "volume_lookup_f"], [9, 3, 1, "", "volume_lookup_i"], [9, 3, 1, "", "volume_lookup_v"], [9, 3, 1, "", "volume_sample_f"], [9, 3, 1, "", "volume_sample_grad_f"], [9, 3, 1, "", "volume_sample_i"], [9, 3, 1, "", "volume_sample_v"], [9, 3, 1, "", "volume_store_f"], [9, 3, 1, "", "volume_store_i"], [9, 3, 1, "", "volume_store_v"], [9, 3, 1, "", "volume_world_to_index"], [9, 3, 1, "", "volume_world_to_index_dir"], [11, 3, 1, "", "zeros"], [11, 3, 1, "", "zeros_like"]], "warp.Bvh": [[11, 1, 1, "", "device"], [11, 1, 1, "", "id"], [11, 2, 1, "", "refit"]], "warp.HashGrid": [[11, 2, 1, "", "build"], [11, 1, 1, "", "device"], [11, 1, 1, "", "id"]], "warp.Mesh": [[11, 1, 1, "", "device"], [11, 1, 1, "", "id"], [11, 2, 1, "", "refit"]], "warp.ScopedTimer": [[11, 1, 1, "", "elapsed"]], "warp.Tape": [[11, 2, 1, "", "backward"], [11, 2, 1, "", "record_func"], [11, 2, 1, "", "reset"], [11, 2, 1, "", "zero"]], "warp.Volume": [[11, 1, 1, "", "CLOSEST"], [11, 1, 1, "", "LINEAR"], [11, 2, 1, "", "allocate"], [11, 2, 1, "", "allocate_by_tiles"], [11, 2, 1, "", "array"], [11, 2, 1, "", "get_tiles"], [11, 2, 1, "", "get_voxel_size"], [11, 2, 1, "", "load_from_numpy"], [11, 2, 1, "", "load_from_nvdb"]], "warp.array": [[11, 2, 1, "", "assign"], [11, 2, 1, "", "contiguous"], [11, 2, 1, "", "cptr"], [11, 2, 1, "", "fill_"], [11, 2, 1, "", "flatten"], [11, 4, 1, "", "grad"], [11, 2, 1, "", "list"], [11, 2, 1, "", "numpy"], [11, 4, 1, "", "requires_grad"], [11, 2, 1, "", "reshape"], [11, 2, 1, "", "to"], [11, 2, 1, "", "transpose"], [11, 4, 1, "", "vars"], [11, 2, 1, "", "view"], [11, 2, 1, "", "zero_"]], "warp.dlpack": [[10, 3, 1, "", "device_from_dlpack"], [10, 3, 1, "", "device_to_dlpack"], [10, 3, 1, "", "dtype_from_dlpack"], [10, 3, 1, "", "dtype_is_compatible"], [10, 3, 1, "", "dtype_to_dlpack"], [10, 3, 1, "", "from_dlpack"], [10, 3, 1, "", "shape_to_dlpack"], [10, 3, 1, "", "strides_to_dlpack"], [10, 3, 1, "", "to_dlpack"]], "warp.fem": [[8, 0, 1, "", "BasisSpace"], [8, 0, 1, "", "BoundarySides"], [8, 0, 1, "", "Cells"], [8, 3, 1, "", "D"], [8, 0, 1, "", "DiscreteField"], [8, 0, 1, "", "DofMapper"], [8, 0, 1, "", "Domain"], [8, 0, 1, "", "ElementBasis"], [8, 0, 1, "", "ExplicitGeometryPartition"], [8, 0, 1, "", "Field"], [8, 0, 1, "", "FieldLike"], [8, 0, 1, "", "FrontierSides"], [8, 0, 1, "", "FunctionSpace"], [8, 0, 1, "", "Geometry"], [8, 0, 1, "", "GeometryDomain"], [8, 0, 1, "", "GeometryPartition"], [8, 0, 1, "", "Grid2D"], [8, 0, 1, "", "Grid3D"], [8, 0, 1, "", "Hexmesh"], [8, 0, 1, "", "LinearGeometryPartition"], [8, 0, 1, "", "NodalQuadrature"], [8, 0, 1, "", "PicQuadrature"], [8, 0, 1, "", "Polynomial"], [8, 0, 1, "", "Quadmesh2D"], [8, 0, 1, "", "Quadrature"], [8, 0, 1, "", "RegularQuadrature"], [8, 0, 1, "", "Sample"], [8, 0, 1, "", "Sides"], [8, 0, 1, "", "SkewSymmetricTensorMapper"], [8, 0, 1, "", "SpacePartition"], [8, 0, 1, "", "SpaceRestriction"], [8, 0, 1, "", "SpaceTopology"], [8, 0, 1, "", "SymmetricTensorMapper"], [8, 0, 1, "", "TemporaryStore"], [8, 0, 1, "", "Tetmesh"], [8, 0, 1, "", "Trimesh2D"], [8, 3, 1, "", "at_node"], [8, 3, 1, "", "average"], [8, 3, 1, "", "borrow_temporary"], [8, 3, 1, "", "borrow_temporary_like"], [8, 3, 1, "", "curl"], [8, 3, 1, "", "deformation_gradient"], [8, 3, 1, "", "degree"], [8, 3, 1, "", "div"], [8, 3, 1, "", "div_outer"], [8, 3, 1, "", "grad"], [8, 3, 1, "", "grad_average"], [8, 3, 1, "", "grad_jump"], [8, 3, 1, "", "grad_outer"], [8, 3, 1, "", "inner"], [8, 3, 1, "", "integrand"], [8, 3, 1, "", "integrate"], [8, 3, 1, "", "interpolate"], [8, 3, 1, "", "jump"], [8, 3, 1, "", "lookup"], [8, 3, 1, "", "make_collocated_function_space"], [8, 3, 1, "", "make_polynomial_basis_space"], [8, 3, 1, "", "make_polynomial_space"], [8, 3, 1, "", "make_restriction"], [8, 3, 1, "", "make_space_partition"], [8, 3, 1, "", "make_space_restriction"], [8, 3, 1, "", "make_test"], [8, 3, 1, "", "make_trial"], [8, 3, 1, "", "measure"], [8, 3, 1, "", "measure_ratio"], [8, 3, 1, "", "normal"], [8, 3, 1, "", "normalize_dirichlet_projector"], [8, 3, 1, "", "outer"], [8, 3, 1, "", "position"], [8, 3, 1, "", "project_linear_system"], [8, 3, 1, "", "set_default_temporary_store"]], "warp.fem.BasisSpace": [[8, 4, 1, "", "geometry"], [8, 2, 1, "", "node_positions"], [8, 4, 1, "", "shape"], [8, 4, 1, "", "topology"]], "warp.fem.DiscreteField": [[8, 4, 1, "", "dof_values"], [8, 2, 1, "", "make_deformed_geometry"], [8, 2, 1, "", "trace"]], "warp.fem.ElementBasis": [[8, 1, 1, "", "LAGRANGE"], [8, 1, 1, "", "NONCONFORMING_POLYNOMIAL"], [8, 1, 1, "", "SERENDIPITY"]], "warp.fem.FunctionSpace": [[8, 4, 1, "", "degree"], [8, 4, 1, "", "dimension"], [8, 1, 1, "", "dtype"], [8, 4, 1, "", "geometry"], [8, 2, 1, "", "make_field"], [8, 4, 1, "", "topology"], [8, 2, 1, "", "trace"]], "warp.fem.Geometry": [[8, 2, 1, "", "boundary_side_count"], [8, 2, 1, "", "cell_count"], [8, 2, 1, "", "side_count"]], "warp.fem.GeometryDomain": [[8, 0, 1, "", "ElementKind"], [8, 4, 1, "", "dimension"], [8, 2, 1, "", "element_count"], [8, 4, 1, "", "element_kind"]], "warp.fem.GeometryPartition": [[8, 2, 1, "", "boundary_side_count"], [8, 2, 1, "", "cell_count"], [8, 2, 1, "", "frontier_side_count"], [8, 2, 1, "", "side_count"]], "warp.fem.Polynomial": [[8, 1, 1, "", "EQUISPACED_CLOSED"], [8, 1, 1, "", "EQUISPACED_OPEN"], [8, 1, 1, "", "GAUSS_LEGENDRE"], [8, 1, 1, "", "LOBATTO_GAUSS_LEGENDRE"]], "warp.fem.Quadrature": [[8, 4, 1, "", "domain"], [8, 2, 1, "", "total_point_count"]], "warp.fem.SpacePartition": [[8, 2, 1, "", "interior_node_count"], [8, 2, 1, "", "node_count"], [8, 2, 1, "", "owned_node_count"], [8, 2, 1, "", "space_node_indices"]], "warp.fem.SpaceTopology": [[8, 1, 1, "", "dimension"], [8, 2, 1, "", "element_node_indices"], [8, 4, 1, "", "geometry"], [8, 2, 1, "", "node_count"], [8, 2, 1, "", "trace"]], "warp.fem.cache": [[8, 0, 1, "", "Temporary"]], "warp.fem.cache.Temporary": [[8, 4, 1, "", "array"], [8, 2, 1, "", "detach"], [8, 2, 1, "", "release"]], "warp.fem.field": [[8, 0, 1, "", "FieldRestriction"], [8, 0, 1, "", "SpaceField"], [8, 0, 1, "", "TestField"], [8, 0, 1, "", "TrialField"]], "warp.fem.operator": [[8, 3, 1, "", "operator"]], "warp.fem.space.shape": [[8, 0, 1, "", "ShapeFunction"]], "warp.jax": [[10, 3, 1, "", "device_from_jax"], [10, 3, 1, "", "device_to_jax"], [10, 3, 1, "", "from_jax"], [10, 3, 1, "", "to_jax"]], "warp.sim": [[12, 6, 1, "", "JOINT_BALL"], [12, 6, 1, "", "JOINT_COMPOUND"], [12, 6, 1, "", "JOINT_D6"], [12, 6, 1, "", "JOINT_DISTANCE"], [12, 6, 1, "", "JOINT_FIXED"], [12, 6, 1, "", "JOINT_FREE"], [12, 6, 1, "", "JOINT_MODE_LIMIT"], [12, 6, 1, "", "JOINT_MODE_TARGET_POSITION"], [12, 6, 1, "", "JOINT_MODE_TARGET_VELOCITY"], [12, 6, 1, "", "JOINT_PRISMATIC"], [12, 6, 1, "", "JOINT_REVOLUTE"], [12, 6, 1, "", "JOINT_UNIVERSAL"], [12, 0, 1, "", "JointAxis"], [12, 0, 1, "", "Mesh"], [12, 0, 1, "", "Model"], [12, 0, 1, "", "ModelBuilder"], [12, 0, 1, "", "SDF"], [12, 0, 1, "", "SemiImplicitIntegrator"], [12, 0, 1, "", "State"], [12, 0, 1, "", "XPBDIntegrator"], [12, 3, 1, "", "eval_fk"], [12, 3, 1, "", "eval_ik"], [12, 3, 1, "", "parse_mjcf"], [12, 3, 1, "", "parse_urdf"], [12, 3, 1, "", "parse_usd"], [12, 3, 1, "", "resolve_usd_from_url"]], "warp.sim.JointAxis": [[12, 1, 1, "", "axis"], [12, 1, 1, "", "limit_kd"], [12, 1, 1, "", "limit_ke"], [12, 1, 1, "", "limit_lower"], [12, 1, 1, "", "limit_upper"], [12, 1, 1, "", "mode"], [12, 1, 1, "", "target"], [12, 1, 1, "", "target_kd"], [12, 1, 1, "", "target_ke"]], "warp.sim.Mesh": [[12, 1, 1, "", "I"], [12, 1, 1, "", "com"], [12, 2, 1, "", "finalize"], [12, 1, 1, "", "indices"], [12, 1, 1, "", "mass"], [12, 1, 1, "", "vertices"]], "warp.sim.Model": [[12, 1, 1, "", "articulation_count"], [12, 1, 1, "", "articulation_start"], [12, 1, 1, "", "body_com"], [12, 1, 1, "", "body_count"], [12, 1, 1, "", "body_inertia"], [12, 1, 1, "", "body_inv_inertia"], [12, 1, 1, "", "body_inv_mass"], [12, 1, 1, "", "body_mass"], [12, 1, 1, "", "body_name"], [12, 1, 1, "", "body_q"], [12, 1, 1, "", "body_qd"], [12, 1, 1, "", "body_shapes"], [12, 1, 1, "", "contact_count"], [12, 2, 1, "", "count_contact_points"], [12, 1, 1, "", "device"], [12, 1, 1, "", "edge_bending_properties"], [12, 1, 1, "", "edge_count"], [12, 1, 1, "", "edge_indices"], [12, 1, 1, "", "edge_rest_angle"], [12, 2, 1, "", "flatten"], [12, 1, 1, "", "gravity"], [12, 1, 1, "", "ground"], [12, 1, 1, "", "ground_plane"], [12, 1, 1, "", "joint_X_c"], [12, 1, 1, "", "joint_X_p"], [12, 1, 1, "", "joint_act"], [12, 1, 1, "", "joint_angular_compliance"], [12, 1, 1, "", "joint_armature"], [12, 1, 1, "", "joint_attach_kd"], [12, 1, 1, "", "joint_attach_ke"], [12, 1, 1, "", "joint_axis"], [12, 1, 1, "", "joint_axis_dim"], [12, 1, 1, "", "joint_axis_mode"], [12, 1, 1, "", "joint_axis_start"], [12, 1, 1, "", "joint_child"], [12, 1, 1, "", "joint_coord_count"], [12, 1, 1, "", "joint_count"], [12, 1, 1, "", "joint_dof_count"], [12, 1, 1, "", "joint_enabled"], [12, 1, 1, "", "joint_limit_kd"], [12, 1, 1, "", "joint_limit_ke"], [12, 1, 1, "", "joint_limit_lower"], [12, 1, 1, "", "joint_limit_upper"], [12, 1, 1, "", "joint_linear_compliance"], [12, 1, 1, "", "joint_name"], [12, 1, 1, "", "joint_parent"], [12, 1, 1, "", "joint_q"], [12, 1, 1, "", "joint_q_start"], [12, 1, 1, "", "joint_qd"], [12, 1, 1, "", "joint_qd_start"], [12, 1, 1, "", "joint_target"], [12, 1, 1, "", "joint_target_kd"], [12, 1, 1, "", "joint_target_ke"], [12, 1, 1, "", "joint_twist_lower"], [12, 1, 1, "", "joint_twist_upper"], [12, 1, 1, "", "joint_type"], [12, 1, 1, "", "muscle_count"], [12, 1, 1, "", "num_envs"], [12, 1, 1, "", "particle_adhesion"], [12, 1, 1, "", "particle_cohesion"], [12, 1, 1, "", "particle_count"], [12, 1, 1, "", "particle_flags"], [12, 1, 1, "", "particle_grid"], [12, 1, 1, "", "particle_inv_mass"], [12, 1, 1, "", "particle_kd"], [12, 1, 1, "", "particle_ke"], [12, 1, 1, "", "particle_kf"], [12, 1, 1, "", "particle_mass"], [12, 1, 1, "", "particle_max_radius"], [12, 1, 1, "", "particle_max_velocity"], [12, 1, 1, "", "particle_mu"], [12, 1, 1, "", "particle_q"], [12, 1, 1, "", "particle_qd"], [12, 1, 1, "", "particle_radius"], [12, 1, 1, "", "requires_grad"], [12, 1, 1, "", "rigid_contact_margin"], [12, 1, 1, "", "rigid_contact_rolling_friction"], [12, 1, 1, "", "rigid_contact_torsional_friction"], [12, 1, 1, "", "shape_body"], [12, 1, 1, "", "shape_collision_filter_pairs"], [12, 1, 1, "", "shape_collision_group"], [12, 1, 1, "", "shape_collision_group_map"], [12, 1, 1, "", "shape_collision_radius"], [12, 1, 1, "", "shape_contact_pairs"], [12, 1, 1, "", "shape_count"], [12, 1, 1, "", "shape_geo_src"], [12, 1, 1, "", "shape_ground_collision"], [12, 1, 1, "", "shape_ground_contact_pairs"], [12, 1, 1, "", "shape_materials"], [12, 1, 1, "", "shape_shape_geo"], [12, 1, 1, "", "shape_transform"], [12, 1, 1, "", "soft_contact_kd"], [12, 1, 1, "", "soft_contact_ke"], [12, 1, 1, "", "soft_contact_kf"], [12, 1, 1, "", "soft_contact_margin"], [12, 4, 1, "", "soft_contact_max"], [12, 1, 1, "", "soft_contact_mu"], [12, 1, 1, "", "soft_contact_restitution"], [12, 1, 1, "", "spring_control"], [12, 1, 1, "", "spring_count"], [12, 1, 1, "", "spring_damping"], [12, 1, 1, "", "spring_indices"], [12, 1, 1, "", "spring_rest_length"], [12, 1, 1, "", "spring_stiffness"], [12, 2, 1, "", "state"], [12, 1, 1, "", "tet_activations"], [12, 1, 1, "", "tet_count"], [12, 1, 1, "", "tet_indices"], [12, 1, 1, "", "tet_materials"], [12, 1, 1, "", "tet_poses"], [12, 1, 1, "", "tri_activations"], [12, 1, 1, "", "tri_count"], [12, 1, 1, "", "tri_indices"], [12, 1, 1, "", "tri_materials"], [12, 1, 1, "", "tri_poses"], [12, 1, 1, "", "up_axis"], [12, 1, 1, "", "up_vector"]], "warp.sim.ModelBuilder": [[12, 2, 1, "", "add_body"], [12, 2, 1, "", "add_builder"], [12, 2, 1, "", "add_cloth_grid"], [12, 2, 1, "", "add_cloth_mesh"], [12, 2, 1, "", "add_edge"], [12, 2, 1, "", "add_edges"], [12, 2, 1, "", "add_joint"], [12, 2, 1, "", "add_joint_ball"], [12, 2, 1, "", "add_joint_compound"], [12, 2, 1, "", "add_joint_d6"], [12, 2, 1, "", "add_joint_distance"], [12, 2, 1, "", "add_joint_fixed"], [12, 2, 1, "", "add_joint_free"], [12, 2, 1, "", "add_joint_prismatic"], [12, 2, 1, "", "add_joint_revolute"], [12, 2, 1, "", "add_joint_universal"], [12, 2, 1, "", "add_muscle"], [12, 2, 1, "", "add_particle"], [12, 2, 1, "", "add_shape_box"], [12, 2, 1, "", "add_shape_capsule"], [12, 2, 1, "", "add_shape_cone"], [12, 2, 1, "", "add_shape_cylinder"], [12, 2, 1, "", "add_shape_mesh"], [12, 2, 1, "", "add_shape_plane"], [12, 2, 1, "", "add_shape_sdf"], [12, 2, 1, "", "add_shape_sphere"], [12, 2, 1, "", "add_soft_grid"], [12, 2, 1, "", "add_soft_mesh"], [12, 2, 1, "", "add_spring"], [12, 2, 1, "", "add_tetrahedron"], [12, 2, 1, "", "add_triangle"], [12, 2, 1, "", "add_triangles"], [12, 2, 1, "", "collapse_fixed_joints"], [12, 2, 1, "", "finalize"], [12, 2, 1, "", "set_ground_plane"]], "warp.sim.SDF": [[12, 1, 1, "", "I"], [12, 1, 1, "", "com"], [12, 1, 1, "", "mass"], [12, 1, 1, "", "volume"]], "warp.sim.State": [[12, 1, 1, "", "body_count"], [12, 1, 1, "", "body_f"], [12, 1, 1, "", "body_q"], [12, 1, 1, "", "body_qd"], [12, 1, 1, "", "particle_count"], [12, 1, 1, "", "particle_f"], [12, 1, 1, "", "particle_q"], [12, 1, 1, "", "particle_qd"]], "warp.sparse": [[13, 0, 1, "", "BsrMatrix"], [13, 3, 1, "", "bsr_assign"], [13, 3, 1, "", "bsr_axpy"], [13, 0, 1, "", "bsr_axpy_work_arrays"], [13, 3, 1, "", "bsr_copy"], [13, 3, 1, "", "bsr_diag"], [13, 3, 1, "", "bsr_get_diag"], [13, 3, 1, "", "bsr_identity"], [13, 3, 1, "", "bsr_mm"], [13, 0, 1, "", "bsr_mm_work_arrays"], [13, 3, 1, "", "bsr_mv"], [13, 3, 1, "", "bsr_scale"], [13, 3, 1, "", "bsr_set_diag"], [13, 3, 1, "", "bsr_set_from_triplets"], [13, 3, 1, "", "bsr_set_identity"], [13, 3, 1, "", "bsr_set_transpose"], [13, 3, 1, "", "bsr_set_zero"], [13, 3, 1, "", "bsr_transposed"], [13, 3, 1, "", "bsr_zeros"]], "warp.sparse.BsrMatrix": [[13, 4, 1, "", "block_shape"], [13, 4, 1, "", "block_size"], [13, 1, 1, "", "columns"], [13, 1, 1, "", "ncol"], [13, 1, 1, "", "nnz"], [13, 1, 1, "", "nrow"], [13, 1, 1, "", "offsets"], [13, 4, 1, "", "scalar_type"], [13, 4, 1, "", "shape"], [13, 1, 1, "", "values"]], "warp.torch": [[10, 3, 1, "", "device_from_torch"], [10, 3, 1, "", "device_to_torch"], [10, 3, 1, "", "dtype_from_torch"], [10, 3, 1, "", "dtype_is_compatible"], [10, 3, 1, "", "from_torch"], [10, 3, 1, "", "stream_from_torch"], [10, 3, 1, "", "stream_to_torch"], [10, 3, 1, "", "to_torch"]]}, "objtypes": {"0": "py:class", "1": "py:attribute", "2": "py:method", "3": "py:function", "4": "py:property", "5": "py:module", "6": "py:data"}, "objnames": {"0": ["py", "class", "Python class"], "1": ["py", "attribute", "Python attribute"], "2": ["py", "method", "Python method"], "3": ["py", "function", "Python function"], "4": ["py", "property", "Python property"], "5": ["py", "module", "Python module"], "6": ["py", "data", "Python data"]}, "titleterms": {"basic": [0, 1, 4, 8], "initi": 0, "kernel": [0, 1, 6, 9, 11, 12], "arrai": [0, 6, 11], "user": [0, 4], "function": [0, 3, 8, 11], "compil": 0, "model": [0, 12], "languag": [0, 3], "detail": 0, "built": 0, "type": [0, 9, 11, 12], "strong": 0, "unsupport": [0, 6], "featur": [0, 6], "debug": 2, "print": 2, "valu": [2, 11], "launch": [2, 11], "step": 2, "through": 2, "gener": [2, 9], "code": 2, "bound": [2, 11], "check": 2, "cuda": [2, 7], "verif": 2, "faq": 3, "how": 3, "doe": 3, "warp": [3, 4, 8, 10, 12, 13], "relat": 3, "other": [3, 9], "python": 3, "project": 3, "gpu": [3, 7], "program": 3, "e": 3, "g": 3, "numba": [3, 10], "taichi": 3, "cupi": [3, 10], "pytorch": [3, 10], "etc": 3, "support": 3, "all": 3, "when": 3, "should": 3, "i": 3, "call": 3, "wp": [3, 7], "synchron": [3, 7], "what": 3, "happen": 3, "you": 3, "differenti": [3, 11], "like": 3, "ab": 3, "x": 3, "multi": [3, 11], "switch": 3, "over": 3, "isaacgym": 3, "physx": 3, "nvidia": 4, "document": 4, "quickstart": 4, "exampl": [0, 4, 7, 8, 10, 11, 12], "addit": 4, "omnivers": 4, "learn": 4, "more": 4, "cite": 4, "licens": 4, "": 4, "guid": 4, "api": [], "refer": [4, 9, 11], "core": 4, "simul": 4, "instal": 5, "depend": 5, "build": 5, "from": [5, 12], "sourc": 5, "devic": 7, "default": 7, "scope": 7, "current": 7, "custom": [7, 11], "context": 7, "fem": 8, "integrand": 8, "workflow": 8, "introductori": 8, "advanc": [1, 8], "usag": 8, "particl": 8, "base": 8, "quadratur": 8, "partit": 8, "oper": [8, 9, 11], "integr": [8, 12], "geometri": [8, 9], "space": 8, "field": 8, "scalar": [9, 11], "vector": [9, 11], "math": 9, "quaternion": [9, 11], "transform": [9, 11], "spatial": 9, "util": 9, "volum": [6, 9, 11], "random": 9, "interoper": 10, "numpi": 10, "optim": 10, "us": [7, 10, 11], "from_torch": 10, "to_torch": 10, "jax": 10, "dlpack": 10, "profil": 11, "runtim": [1, 11], "dimension": 11, "matrix": 11, "multipl": [6, 7, 11], "data": 11, "matric": [11, 13], "struct": [6, 11], "convers": 11, "constant": 11, "boolean": 11, "comparison": 11, "arithmet": 11, "mesh": [11, 12], "hash": 11, "grid": 11, "jacobian": 11, "gradient": 11, "replai": 11, "grad": 11, "version": 11, "myfunc": 11, "1": 11, "2": 11, "graph": 11, "sim": 12, "todo": 12, "joint": 12, "mode": 12, "state": 12, "import": 12, "spars": 13, "chang": 0, "cach": 0, "directori": 0, "limit": [0, 6], "set": 1, "global": 1, "modul": 1, "full": 4, "tabl": 4, "content": 4, "process": 6, "scopeddevic": 7, "memori": 8, "manag": 8, "boundari": 8, "condit": 8, "interfac": 8, "larg": 11, "comput": 11, "hierarchi": 11, "bvh": 11, "rai": 11, "travers": 11, "creation": 12, "triangl": 12, "obj": 12, "file": 12, "forward": 12, "invers": 12, "kinemat": 12, "variabl": 12, "name": 12, "articul": 12, "py": 12, "conda": 5, "environ": 5, "high": 8, "order": 8, "curv": 8, "nativ": 11, "special": 11}, "envversion": {"sphinx.domains.c": 3, "sphinx.domains.changeset": 1, "sphinx.domains.citation": 1, "sphinx.domains.cpp": 9, "sphinx.domains.index": 1, "sphinx.domains.javascript": 3, "sphinx.domains.math": 2, "sphinx.domains.python": 4, "sphinx.domains.rst": 2, "sphinx.domains.std": 2, "sphinx.ext.intersphinx": 1, "sphinx.ext.todo": 2, "sphinx": 60}, "alltitles": {"Basics": [[0, "basics"]], "Initialization": [[0, "initialization"]], "Kernels": [[0, "kernels"], [6, "kernels"], [11, "kernels"]], "Example: Changing the kernel cache directory": [[0, "example-changing-the-kernel-cache-directory"]], "Arrays": [[0, "arrays"], [6, "arrays"], [11, "arrays"]], "User Functions": [[0, "user-functions"]], "Compilation Model": [[0, "compilation-model"]], "Language Details": [[0, "language-details"]], "Built-in Types": [[0, "built-in-types"]], "Strong Typing": [[0, "strong-typing"]], "Limitations and Unsupported Features": [[0, "limitations-and-unsupported-features"]], "Runtime Settings": [[1, "runtime-settings"]], "Global Settings": [[1, "global-settings"]], "Basic Global Settings": [[1, "basic-global-settings"]], "Advanced Global Settings": [[1, "advanced-global-settings"]], "Module Settings": [[1, "module-settings"]], "Kernel Settings": [[1, "kernel-settings"]], "Debugging": [[2, "debugging"]], "Printing Values": [[2, "printing-values"]], "Printing Launches": [[2, "printing-launches"]], "Step-Through Debugging": [[2, "step-through-debugging"]], "Generated Code": [[2, "generated-code"]], "Bounds Checking": [[2, "bounds-checking"]], "CUDA Verification": [[2, "cuda-verification"]], "FAQ": [[3, "faq"]], "How does Warp relate to other Python projects for GPU programming, e.g.: Numba, Taichi, cuPy, PyTorch, etc.?": [[3, "how-does-warp-relate-to-other-python-projects-for-gpu-programming-e-g-numba-taichi-cupy-pytorch-etc"]], "Does Warp support all of the Python language?": [[3, "does-warp-support-all-of-the-python-language"]], "When should I call wp.synchronize()?": [[3, "when-should-i-call-wp-synchronize"]], "What happens when you differentiate a function like wp.abs(x)?": [[3, "what-happens-when-you-differentiate-a-function-like-wp-abs-x"]], "Does Warp support multi-GPU programming?": [[3, "does-warp-support-multi-gpu-programming"]], "Should I switch to Warp over IsaacGym / PhysX?": [[3, "should-i-switch-to-warp-over-isaacgym-physx"]], "NVIDIA Warp Documentation": [[4, "nvidia-warp-documentation"]], "Quickstart": [[4, "quickstart"]], "Basic example": [[4, "basic-example"]], "Additional examples": [[4, "additional-examples"]], "Omniverse": [[4, "omniverse"]], "Learn More": [[4, "learn-more"]], "Citing": [[4, "citing"]], "License": [[4, "license"]], "Full Table of Contents": [[4, "full-table-of-contents"]], "User's Guide": [[4, null]], "Core Reference": [[4, null]], "Simulation Reference": [[4, null]], "Installation": [[5, "installation"]], "Dependencies": [[5, "dependencies"]], "Building from source": [[5, "building-from-source"]], "Conda environments": [[5, "conda-environments"]], "Limitations": [[6, "limitations"]], "Unsupported Features": [[6, "unsupported-features"]], "Structs": [[6, "structs"], [11, "structs"]], "Volumes": [[6, "volumes"], [9, "volumes"], [11, "volumes"]], "Multiple Processes": [[6, "multiple-processes"]], "Devices": [[7, "devices"]], "Default Device": [[7, "default-device"]], "Scoped Devices": [[7, "scoped-devices"]], "Example: Using wp.ScopedDevice with multiple GPUs": [[7, "example-using-wp-scopeddevice-with-multiple-gpus"]], "Current CUDA Device": [[7, "current-cuda-device"]], "Device Synchronization": [[7, "device-synchronization"]], "Custom CUDA Contexts": [[7, "custom-cuda-contexts"]], "warp.fem": [[8, "warp-fem"]], "Integrands": [[8, "integrands"]], "Basic workflow": [[8, "basic-workflow"]], "Introductory examples": [[8, "introductory-examples"]], "Advanced usages": [[8, "advanced-usages"]], "High-order (curved) geometries": [[8, "high-order-curved-geometries"]], "Particle-based quadrature": [[8, "particle-based-quadrature"]], "Partitioning": [[8, "partitioning"]], "Memory management": [[8, "memory-management"], [8, "id2"]], "Operators": [[8, "operators"], [9, "operators"], [11, "operators"]], "Integration": [[8, "integration"]], "Geometry": [[8, "geometry"], [9, "geometry"]], "Function Spaces": [[8, "function-spaces"]], "Fields": [[8, "fields"]], "Boundary Conditions": [[8, "boundary-conditions"]], "Interfaces": [[8, "interfaces"]], "Interoperability": [[10, "interoperability"]], "NumPy": [[10, "numpy"]], "PyTorch": [[10, "pytorch"]], "Example: Optimization using warp.from_torch()": [[10, "example-optimization-using-warp-from-torch"]], "Example: Optimization using warp.to_torch": [[10, "example-optimization-using-warp-to-torch"]], "CuPy/Numba": [[10, "cupy-numba"]], "JAX": [[10, "jax"]], "DLPack": [[10, "module-warp.dlpack"]], "warp.sim": [[12, "warp-sim"]], "Model": [[12, "model"]], "Todo": [[12, "id1"]], "Example mesh creation from a triangle OBJ mesh file:": [[12, "example-mesh-creation-from-a-triangle-obj-mesh-file"]], "Joint types": [[12, "joint-types"]], "Joint modes": [[12, "joint-modes"]], "State": [[12, "state"]], "Forward / Inverse Kinematics": [[12, "forward-inverse-kinematics"]], "Variable names in the kernels from articulation.py": [[12, "id4"]], "Integrators": [[12, "integrators"]], "Importers": [[12, "importers"]], "warp.sparse": [[13, "warp-sparse"]], "Sparse Matrices": [[13, "sparse-matrices"]], "Kernel Reference": [[9, "kernel-reference"]], "Scalar Types": [[9, "scalar-types"], [11, "scalar-types"]], "Vector Types": [[9, "vector-types"]], "Generic Types": [[9, "generic-types"]], "Scalar Math": [[9, "scalar-math"]], "Vector Math": [[9, "vector-math"]], "Other": [[9, "other"]], "Quaternion Math": [[9, "quaternion-math"]], "Transformations": [[9, "transformations"]], "Spatial Math": [[9, "spatial-math"]], "Utility": [[9, "utility"]], "Random": [[9, "random"]], "Runtime Reference": [[11, "runtime-reference"]], "Large Kernel Launches": [[11, "large-kernel-launches"]], "Runtime Kernel Specialization": [[11, "runtime-kernel-specialization"]], "Multi-dimensional Arrays": [[11, "multi-dimensional-arrays"]], "Matrix Multiplication": [[11, "matrix-multiplication"]], "Data Types": [[11, "data-types"]], "Vectors": [[11, "vectors"]], "Matrices": [[11, "matrices"]], "Quaternions": [[11, "quaternions"]], "Transforms": [[11, "transforms"]], "Example: Using a custom struct in gradient computation": [[11, "example-using-a-custom-struct-in-gradient-computation"]], "Type Conversions": [[11, "type-conversions"]], "Constants": [[11, "constants"]], "Boolean Operators": [[11, "boolean-operators"]], "Comparison Operators": [[11, "comparison-operators"]], "Arithmetic Operators": [[11, "arithmetic-operators"]], "Graphs": [[11, "graphs"]], "Bounding Value Hierarchies (BVH)": [[11, "bounding-value-hierarchies-bvh"]], "Example: BVH Ray Traversal": [[11, "example-bvh-ray-traversal"]], "Example: BVH Volume Traversal": [[11, "example-bvh-volume-traversal"]], "Meshes": [[11, "meshes"]], "Hash Grids": [[11, "hash-grids"]], "Differentiability": [[11, "differentiability"]], "Jacobians": [[11, "jacobians"]], "Custom Gradient Functions": [[11, "custom-gradient-functions"]], "Customizing the replay and grad versions of function myfunc": [[11, "id1"]], "Example 1: Custom Grad Function": [[11, "example-1-custom-grad-function"]], "Example 2: Custom Replay Function": [[11, "example-2-custom-replay-function"]], "Custom Native Functions": [[11, "custom-native-functions"]], "Profiling": [[11, "profiling"]]}, "indexentries": {"array (class in warp)": [[9, "warp.Array"], [11, "warp.array"]], "float (class in warp)": [[9, "warp.Float"]], "int (class in warp)": [[9, "warp.Int"]], "matrix (class in warp)": [[9, "warp.Matrix"]], "quaternion (class in warp)": [[9, "warp.Quaternion"]], "scalar (class in warp)": [[9, "warp.Scalar"]], "transformation (class in warp)": [[9, "warp.Transformation"]], "vector (class in warp)": [[9, "warp.Vector"]], "abs() (in module warp)": [[9, "warp.abs"]], "acos() (in module warp)": [[9, "warp.acos"]], "add() (in module warp)": [[9, "warp.add"]], "argmax() (in module warp)": [[9, "warp.argmax"]], "argmin() (in module warp)": [[9, "warp.argmin"]], "asin() (in module warp)": [[9, "warp.asin"]], "atan() (in module warp)": [[9, "warp.atan"]], "atan2() (in module warp)": [[9, "warp.atan2"]], "atomic_add() (in module warp)": [[9, "warp.atomic_add"]], "atomic_max() (in module warp)": [[9, "warp.atomic_max"]], "atomic_min() (in module warp)": [[9, "warp.atomic_min"]], "atomic_sub() (in module warp)": [[9, "warp.atomic_sub"]], "bit_and() (in module warp)": [[9, "warp.bit_and"]], "bit_or() (in module warp)": [[9, "warp.bit_or"]], "bit_xor() (in module warp)": [[9, "warp.bit_xor"]], "bool (class in warp)": [[9, "warp.bool"]], "breakpoint() (in module warp)": [[9, "warp.breakpoint"]], "bvh_query_aabb() (in module warp)": [[9, "warp.bvh_query_aabb"]], "bvh_query_next() (in module warp)": [[9, "warp.bvh_query_next"]], "bvh_query_ray() (in module warp)": [[9, "warp.bvh_query_ray"]], "cbrt() (in module warp)": [[9, "warp.cbrt"]], "ceil() (in module warp)": [[9, "warp.ceil"]], "clamp() (in module warp)": [[9, "warp.clamp"]], "closest_point_edge_edge() (in module warp)": [[9, "warp.closest_point_edge_edge"]], "cos() (in module warp)": [[9, "warp.cos"]], "cosh() (in module warp)": [[9, "warp.cosh"]], "cross() (in module warp)": [[9, "warp.cross"]], "curlnoise() (in module warp)": [[9, "warp.curlnoise"]], "cw_div() (in module warp)": [[9, "warp.cw_div"]], "cw_mul() (in module warp)": [[9, "warp.cw_mul"]], "ddot() (in module warp)": [[9, "warp.ddot"]], "degrees() (in module warp)": [[9, "warp.degrees"]], "determinant() (in module warp)": [[9, "warp.determinant"]], "diag() (in module warp)": [[9, "warp.diag"]], "div() (in module warp)": [[9, "warp.div"]], "dot() (in module warp)": [[9, "warp.dot"]], "eig3() (in module warp)": [[9, "warp.eig3"]], "exp() (in module warp)": [[9, "warp.exp"]], "expect_near() (in module warp)": [[9, "warp.expect_near"]], "float16 (class in warp)": [[9, "warp.float16"]], "float32 (class in warp)": [[9, "warp.float32"]], "float64 (class in warp)": [[9, "warp.float64"]], "floor() (in module warp)": [[9, "warp.floor"]], "floordiv() (in module warp)": [[9, "warp.floordiv"]], "frac() (in module warp)": [[9, "warp.frac"]], "get_diag() (in module warp)": [[9, "warp.get_diag"]], "hash_grid_point_id() (in module warp)": [[9, "warp.hash_grid_point_id"]], "hash_grid_query() (in module warp)": [[9, "warp.hash_grid_query"]], "hash_grid_query_next() (in module warp)": [[9, "warp.hash_grid_query_next"]], "identity() (in module warp)": [[9, "warp.identity"]], "int16 (class in warp)": [[9, "warp.int16"]], "int32 (class in warp)": [[9, "warp.int32"]], "int64 (class in warp)": [[9, "warp.int64"]], "int8 (class in warp)": [[9, "warp.int8"]], "intersect_tri_tri() (in module warp)": [[9, "warp.intersect_tri_tri"]], "inverse() (in module warp)": [[9, "warp.inverse"]], "invert() (in module warp)": [[9, "warp.invert"]], "length() (in module warp)": [[9, "warp.length"]], "length_sq() (in module warp)": [[9, "warp.length_sq"]], "lerp() (in module warp)": [[9, "warp.lerp"]], "log() (in module warp)": [[9, "warp.log"]], "log10() (in module warp)": [[9, "warp.log10"]], "log2() (in module warp)": [[9, "warp.log2"]], "lower_bound() (in module warp)": [[9, "warp.lower_bound"]], "lshift() (in module warp)": [[9, "warp.lshift"]], "mat22d (class in warp)": [[9, "warp.mat22d"]], "mat22f (class in warp)": [[9, "warp.mat22f"]], "mat22h (class in warp)": [[9, "warp.mat22h"]], "mat33d (class in warp)": [[9, "warp.mat33d"]], "mat33f (class in warp)": [[9, "warp.mat33f"]], "mat33h (class in warp)": [[9, "warp.mat33h"]], "mat44d (class in warp)": [[9, "warp.mat44d"]], "mat44f (class in warp)": [[9, "warp.mat44f"]], "mat44h (class in warp)": [[9, "warp.mat44h"]], "matrix() (in module warp)": [[9, "warp.matrix"]], "max() (in module warp)": [[9, "warp.max"]], "mesh_eval_face_normal() (in module warp)": [[9, "warp.mesh_eval_face_normal"]], "mesh_eval_position() (in module warp)": [[9, "warp.mesh_eval_position"]], "mesh_eval_velocity() (in module warp)": [[9, "warp.mesh_eval_velocity"]], "mesh_get() (in module warp)": [[9, "warp.mesh_get"]], "mesh_get_index() (in module warp)": [[9, "warp.mesh_get_index"]], "mesh_get_point() (in module warp)": [[9, "warp.mesh_get_point"]], "mesh_get_velocity() (in module warp)": [[9, "warp.mesh_get_velocity"]], "mesh_query_aabb() (in module warp)": [[9, "warp.mesh_query_aabb"]], "mesh_query_aabb_next() (in module warp)": [[9, "warp.mesh_query_aabb_next"]], "mesh_query_furthest_point_no_sign() (in module warp)": [[9, "warp.mesh_query_furthest_point_no_sign"]], "mesh_query_point() (in module warp)": [[9, "warp.mesh_query_point"]], "mesh_query_point_no_sign() (in module warp)": [[9, "warp.mesh_query_point_no_sign"]], "mesh_query_point_sign_normal() (in module warp)": [[9, "warp.mesh_query_point_sign_normal"]], "mesh_query_point_sign_winding_number() (in module warp)": [[9, "warp.mesh_query_point_sign_winding_number"]], "mesh_query_ray() (in module warp)": [[9, "warp.mesh_query_ray"]], "min() (in module warp)": [[9, "warp.min"]], "mlp() (in module warp)": [[9, "warp.mlp"]], "mod() (in module warp)": [[9, "warp.mod"]], "mul() (in module warp)": [[9, "warp.mul"]], "neg() (in module warp)": [[9, "warp.neg"]], "noise() (in module warp)": [[9, "warp.noise"]], "nonzero() (in module warp)": [[9, "warp.nonzero"]], "normalize() (in module warp)": [[9, "warp.normalize"]], "outer() (in module warp)": [[9, "warp.outer"]], "pnoise() (in module warp)": [[9, "warp.pnoise"]], "poisson() (in module warp)": [[9, "warp.poisson"]], "pos() (in module warp)": [[9, "warp.pos"]], "pow() (in module warp)": [[9, "warp.pow"]], "print() (in module warp)": [[9, "warp.print"]], "printf() (in module warp)": [[9, "warp.printf"]], "qr3() (in module warp)": [[9, "warp.qr3"]], "quat_from_axis_angle() (in module warp)": [[9, "warp.quat_from_axis_angle"]], "quat_from_matrix() (in module warp)": [[9, "warp.quat_from_matrix"]], "quat_identity() (in module warp)": [[9, "warp.quat_identity"]], "quat_inverse() (in module warp)": [[9, "warp.quat_inverse"]], "quat_rotate() (in module warp)": [[9, "warp.quat_rotate"]], "quat_rotate_inv() (in module warp)": [[9, "warp.quat_rotate_inv"]], "quat_rpy() (in module warp)": [[9, "warp.quat_rpy"]], "quat_slerp() (in module warp)": [[9, "warp.quat_slerp"]], "quat_to_axis_angle() (in module warp)": [[9, "warp.quat_to_axis_angle"]], "quat_to_matrix() (in module warp)": [[9, "warp.quat_to_matrix"]], "quatd (class in warp)": [[9, "warp.quatd"]], "quaternion() (in module warp)": [[9, "warp.quaternion"]], "quatf (class in warp)": [[9, "warp.quatf"]], "quath (class in warp)": [[9, "warp.quath"]], "radians() (in module warp)": [[9, "warp.radians"]], "rand_init() (in module warp)": [[9, "warp.rand_init"]], "randf() (in module warp)": [[9, "warp.randf"]], "randi() (in module warp)": [[9, "warp.randi"]], "randn() (in module warp)": [[9, "warp.randn"]], "rint() (in module warp)": [[9, "warp.rint"]], "round() (in module warp)": [[9, "warp.round"]], "rshift() (in module warp)": [[9, "warp.rshift"]], "sample_cdf() (in module warp)": [[9, "warp.sample_cdf"]], "sample_triangle() (in module warp)": [[9, "warp.sample_triangle"]], "sample_unit_cube() (in module warp)": [[9, "warp.sample_unit_cube"]], "sample_unit_disk() (in module warp)": [[9, "warp.sample_unit_disk"]], "sample_unit_hemisphere() (in module warp)": [[9, "warp.sample_unit_hemisphere"]], "sample_unit_hemisphere_surface() (in module warp)": [[9, "warp.sample_unit_hemisphere_surface"]], "sample_unit_ring() (in module warp)": [[9, "warp.sample_unit_ring"]], "sample_unit_sphere() (in module warp)": [[9, "warp.sample_unit_sphere"]], "sample_unit_sphere_surface() (in module warp)": [[9, "warp.sample_unit_sphere_surface"]], "sample_unit_square() (in module warp)": [[9, "warp.sample_unit_square"]], "select() (in module warp)": [[9, "warp.select"]], "sign() (in module warp)": [[9, "warp.sign"]], "sin() (in module warp)": [[9, "warp.sin"]], "sinh() (in module warp)": [[9, "warp.sinh"]], "skew() (in module warp)": [[9, "warp.skew"]], "smoothstep() (in module warp)": [[9, "warp.smoothstep"]], "spatial_adjoint() (in module warp)": [[9, "warp.spatial_adjoint"]], "spatial_bottom() (in module warp)": [[9, "warp.spatial_bottom"]], "spatial_cross() (in module warp)": [[9, "warp.spatial_cross"]], "spatial_cross_dual() (in module warp)": [[9, "warp.spatial_cross_dual"]], "spatial_dot() (in module warp)": [[9, "warp.spatial_dot"]], "spatial_jacobian() (in module warp)": [[9, "warp.spatial_jacobian"]], "spatial_mass() (in module warp)": [[9, "warp.spatial_mass"]], "spatial_matrixd (class in warp)": [[9, "warp.spatial_matrixd"]], "spatial_matrixf (class in warp)": [[9, "warp.spatial_matrixf"]], "spatial_matrixh (class in warp)": [[9, "warp.spatial_matrixh"]], "spatial_top() (in module warp)": [[9, "warp.spatial_top"]], "spatial_vectord (class in warp)": [[9, "warp.spatial_vectord"]], "spatial_vectorf (class in warp)": [[9, "warp.spatial_vectorf"]], "spatial_vectorh (class in warp)": [[9, "warp.spatial_vectorh"]], "sqrt() (in module warp)": [[9, "warp.sqrt"]], "step() (in module warp)": [[9, "warp.step"]], "sub() (in module warp)": [[9, "warp.sub"]], "svd3() (in module warp)": [[9, "warp.svd3"]], "tan() (in module warp)": [[9, "warp.tan"]], "tanh() (in module warp)": [[9, "warp.tanh"]], "tid() (in module warp)": [[9, "warp.tid"]], "trace() (in module warp)": [[9, "warp.trace"]], "transform_get_rotation() (in module warp)": [[9, "warp.transform_get_rotation"]], "transform_get_translation() (in module warp)": [[9, "warp.transform_get_translation"]], "transform_identity() (in module warp)": [[9, "warp.transform_identity"]], "transform_inverse() (in module warp)": [[9, "warp.transform_inverse"]], "transform_multiply() (in module warp)": [[9, "warp.transform_multiply"]], "transform_point() (in module warp)": [[9, "warp.transform_point"]], "transform_vector() (in module warp)": [[9, "warp.transform_vector"]], "transformation() (in module warp)": [[9, "warp.transformation"]], "transformd (class in warp)": [[9, "warp.transformd"]], "transformf (class in warp)": [[9, "warp.transformf"]], "transformh (class in warp)": [[9, "warp.transformh"]], "transpose() (in module warp)": [[9, "warp.transpose"]], "trunc() (in module warp)": [[9, "warp.trunc"]], "uint16 (class in warp)": [[9, "warp.uint16"]], "uint32 (class in warp)": [[9, "warp.uint32"]], "uint64 (class in warp)": [[9, "warp.uint64"]], "uint8 (class in warp)": [[9, "warp.uint8"]], "unot() (in module warp)": [[9, "warp.unot"]], "vec2b (class in warp)": [[9, "warp.vec2b"]], "vec2d (class in warp)": [[9, "warp.vec2d"]], "vec2f (class in warp)": [[9, "warp.vec2f"]], "vec2h (class in warp)": [[9, "warp.vec2h"]], "vec2i (class in warp)": [[9, "warp.vec2i"]], "vec2l (class in warp)": [[9, "warp.vec2l"]], "vec2s (class in warp)": [[9, "warp.vec2s"]], "vec2ub (class in warp)": [[9, "warp.vec2ub"]], "vec2ui (class in warp)": [[9, "warp.vec2ui"]], "vec2ul (class in warp)": [[9, "warp.vec2ul"]], "vec2us (class in warp)": [[9, "warp.vec2us"]], "vec3b (class in warp)": [[9, "warp.vec3b"]], "vec3d (class in warp)": [[9, "warp.vec3d"]], "vec3f (class in warp)": [[9, "warp.vec3f"]], "vec3h (class in warp)": [[9, "warp.vec3h"]], "vec3i (class in warp)": [[9, "warp.vec3i"]], "vec3l (class in warp)": [[9, "warp.vec3l"]], "vec3s (class in warp)": [[9, "warp.vec3s"]], "vec3ub (class in warp)": [[9, "warp.vec3ub"]], "vec3ui (class in warp)": [[9, "warp.vec3ui"]], "vec3ul (class in warp)": [[9, "warp.vec3ul"]], "vec3us (class in warp)": [[9, "warp.vec3us"]], "vec4b (class in warp)": [[9, "warp.vec4b"]], "vec4d (class in warp)": [[9, "warp.vec4d"]], "vec4f (class in warp)": [[9, "warp.vec4f"]], "vec4h (class in warp)": [[9, "warp.vec4h"]], "vec4i (class in warp)": [[9, "warp.vec4i"]], "vec4l (class in warp)": [[9, "warp.vec4l"]], "vec4s (class in warp)": [[9, "warp.vec4s"]], "vec4ub (class in warp)": [[9, "warp.vec4ub"]], "vec4ui (class in warp)": [[9, "warp.vec4ui"]], "vec4ul (class in warp)": [[9, "warp.vec4ul"]], "vec4us (class in warp)": [[9, "warp.vec4us"]], "vector() (in module warp)": [[9, "warp.vector"]], "volume_index_to_world() (in module warp)": [[9, "warp.volume_index_to_world"]], "volume_index_to_world_dir() (in module warp)": [[9, "warp.volume_index_to_world_dir"]], "volume_lookup_f() (in module warp)": [[9, "warp.volume_lookup_f"]], "volume_lookup_i() (in module warp)": [[9, "warp.volume_lookup_i"]], "volume_lookup_v() (in module warp)": [[9, "warp.volume_lookup_v"]], "volume_sample_f() (in module warp)": [[9, "warp.volume_sample_f"]], "volume_sample_grad_f() (in module warp)": [[9, "warp.volume_sample_grad_f"]], "volume_sample_i() (in module warp)": [[9, "warp.volume_sample_i"]], "volume_sample_v() (in module warp)": [[9, "warp.volume_sample_v"]], "volume_store_f() (in module warp)": [[9, "warp.volume_store_f"]], "volume_store_i() (in module warp)": [[9, "warp.volume_store_i"]], "volume_store_v() (in module warp)": [[9, "warp.volume_store_v"]], "volume_world_to_index() (in module warp)": [[9, "warp.volume_world_to_index"]], "volume_world_to_index_dir() (in module warp)": [[9, "warp.volume_world_to_index_dir"]], "bvh (class in warp)": [[11, "warp.Bvh"]], "closest (warp.volume attribute)": [[11, "warp.Volume.CLOSEST"]], "hashgrid (class in warp)": [[11, "warp.HashGrid"]], "linear (warp.volume attribute)": [[11, "warp.Volume.LINEAR"]], "mesh (class in warp)": [[11, "warp.Mesh"]], "scopedtimer (class in warp)": [[11, "warp.ScopedTimer"]], "tape (class in warp)": [[11, "warp.Tape"]], "volume (class in warp)": [[11, "warp.Volume"]], "allocate() (warp.volume class method)": [[11, "warp.Volume.allocate"]], "allocate_by_tiles() (warp.volume class method)": [[11, "warp.Volume.allocate_by_tiles"]], "array() (warp.volume method)": [[11, "warp.Volume.array"]], "assign() (warp.array method)": [[11, "warp.array.assign"]], "backward() (warp.tape method)": [[11, "warp.Tape.backward"]], "build() (warp.hashgrid method)": [[11, "warp.HashGrid.build"]], "capture_begin() (in module warp)": [[11, "warp.capture_begin"]], "capture_end() (in module warp)": [[11, "warp.capture_end"]], "capture_launch() (in module warp)": [[11, "warp.capture_launch"]], "clone() (in module warp)": [[11, "warp.clone"]], "constant (class in warp)": [[11, "warp.constant"]], "contiguous() (warp.array method)": [[11, "warp.array.contiguous"]], "copy() (in module warp)": [[11, "warp.copy"]], "cptr() (warp.array method)": [[11, "warp.array.cptr"]], "device (warp.bvh attribute)": [[11, "warp.Bvh.device"]], "device (warp.hashgrid attribute)": [[11, "warp.HashGrid.device"]], "device (warp.mesh attribute)": [[11, "warp.Mesh.device"]], "elapsed (warp.scopedtimer attribute)": [[11, "warp.ScopedTimer.elapsed"]], "empty() (in module warp)": [[11, "warp.empty"]], "empty_like() (in module warp)": [[11, "warp.empty_like"]], "fill_() (warp.array method)": [[11, "warp.array.fill_"]], "flatten() (warp.array method)": [[11, "warp.array.flatten"]], "full() (in module warp)": [[11, "warp.full"]], "full_like() (in module warp)": [[11, "warp.full_like"]], "get_tiles() (warp.volume method)": [[11, "warp.Volume.get_tiles"]], "get_voxel_size() (warp.volume method)": [[11, "warp.Volume.get_voxel_size"]], "grad (warp.array property)": [[11, "warp.array.grad"]], "id (warp.bvh attribute)": [[11, "warp.Bvh.id"]], "id (warp.hashgrid attribute)": [[11, "warp.HashGrid.id"]], "id (warp.mesh attribute)": [[11, "warp.Mesh.id"]], "launch() (in module warp)": [[11, "warp.launch"]], "list() (warp.array method)": [[11, "warp.array.list"]], "load_from_numpy() (warp.volume class method)": [[11, "warp.Volume.load_from_numpy"]], "load_from_nvdb() (warp.volume class method)": [[11, "warp.Volume.load_from_nvdb"]], "matmul() (in module warp)": [[11, "warp.matmul"]], "numpy() (warp.array method)": [[11, "warp.array.numpy"]], "record_func() (warp.tape method)": [[11, "warp.Tape.record_func"]], "refit() (warp.bvh method)": [[11, "warp.Bvh.refit"]], "refit() (warp.mesh method)": [[11, "warp.Mesh.refit"]], "requires_grad (warp.array property)": [[11, "warp.array.requires_grad"]], "reset() (warp.tape method)": [[11, "warp.Tape.reset"]], "reshape() (warp.array method)": [[11, "warp.array.reshape"]], "to() (warp.array method)": [[11, "warp.array.to"]], "transpose() (warp.array method)": [[11, "warp.array.transpose"]], "vars (warp.array property)": [[11, "warp.array.vars"]], "view() (warp.array method)": [[11, "warp.array.view"]], "zero() (warp.tape method)": [[11, "warp.Tape.zero"]], "zero_() (warp.array method)": [[11, "warp.array.zero_"]], "zeros() (in module warp)": [[11, "warp.zeros"]], "zeros_like() (in module warp)": [[11, "warp.zeros_like"]]}})
\ No newline at end of file
diff --git a/docs/conf.py b/docs/conf.py
index 8d433bae7..f8902f848 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -14,14 +14,13 @@
 import sys
 from datetime import date
 
-import warp as wp
-
-sys.path.insert(0, os.path.abspath(".."))
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
+import warp as wp  # noqa: E402
 
 # -- Project information -----------------------------------------------------
 
 project = "Warp"
-copyright = f"{date.today().year}, NVIDIA"
+copyright = f"2022-{date.today().year}, NVIDIA"
 author = "NVIDIA"
 
 version = wp.__version__
diff --git a/docs/modules/functions.rst b/docs/modules/functions.rst
index 5fe56ada3..bbf992796 100644
--- a/docs/modules/functions.rst
+++ b/docs/modules/functions.rst
@@ -91,6 +91,14 @@ Generic Types
 .. class:: Transformation
 .. class:: Array
 
+Query Types
+-------------
+.. autoclass:: bvh_query_t
+.. autoclass:: hash_grid_query_t
+.. autoclass:: mesh_query_aabb_t
+.. autoclass:: mesh_query_point_t
+.. autoclass:: mesh_query_ray_t
+
 
 Scalar Math
 ---------------
@@ -1278,7 +1286,7 @@ Geometry
 .. function:: bvh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> bvh_query_t
 
    Construct an axis-aligned bounding box query against a BVH object. This query can be used to iterate over all bounds
-   inside a BVH. Returns an object that is used to track state during BVH traversal.
+   inside a BVH.
 
    :param id: The BVH identifier
    :param lower: The lower bound of the bounding box in BVH space
@@ -1288,7 +1296,7 @@ Geometry
 .. function:: bvh_query_ray(id: uint64, start: vec3f, dir: vec3f) -> bvh_query_t
 
    Construct a ray query against a BVH object. This query can be used to iterate over all bounds
-   that intersect the ray. Returns an object that is used to track state during BVH traversal.
+   that intersect the ray.
 
    :param id: The BVH identifier
    :param start: The start of the ray in BVH space
@@ -1301,9 +1309,9 @@ Geometry
    The index of the current bound is stored in ``index``, returns ``False`` if there are no more overlapping bound.
 
 
-.. function:: mesh_query_point(id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32) -> bool
+.. function:: mesh_query_point(id: uint64, point: vec3f, max_dist: float32) -> mesh_query_point_t
 
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
+   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
 
    Identifies the sign of the distance using additional ray-casts to determine if the point is inside or outside.
    This method is relatively robust, but does increase computational cost.
@@ -1312,44 +1320,33 @@ Geometry
    :param id: The mesh identifier
    :param point: The point in space to query
    :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                  Note that mesh must be watertight for this to be robust
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
 
 
-.. function:: mesh_query_point_no_sign(id: uint64, point: vec3f, max_dist: float32, face: int32, bary_u: float32, bary_v: float32) -> bool
+.. function:: mesh_query_point_no_sign(id: uint64, point: vec3f, max_dist: float32) -> mesh_query_point_t
 
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
+   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
 
    This method does not compute the sign of the point (inside/outside) which makes it faster than other point query methods.
 
    :param id: The mesh identifier
    :param point: The point in space to query
    :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
 
 
-.. function:: mesh_query_furthest_point_no_sign(id: uint64, point: vec3f, min_dist: float32, face: int32, bary_u: float32, bary_v: float32) -> bool
+.. function:: mesh_query_furthest_point_no_sign(id: uint64, point: vec3f, min_dist: float32) -> mesh_query_point_t
 
-   Computes the furthest point on the mesh with identifier `id` to the given point in space. Returns ``True`` if a point > ``min_dist`` is found.
+   Computes the furthest point on the mesh with identifier `id` to the given point in space.
 
    This method does not compute the sign of the point (inside/outside).
 
    :param id: The mesh identifier
    :param point: The point in space to query
    :param min_dist: Mesh faces below this distance will not be considered by the query
-   :param face: Returns the index of the furthest face
-   :param bary_u: Returns the barycentric u coordinate of the furthest point
-   :param bary_v: Returns the barycentric v coordinate of the furthest point
 
 
-.. function:: mesh_query_point_sign_normal(id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32, epsilon: float32) -> bool
+.. function:: mesh_query_point_sign_normal(id: uint64, point: vec3f, max_dist: float32, epsilon: float32) -> mesh_query_point_t
 
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
+   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
 
    Identifies the sign of the distance (inside/outside) using the angle-weighted pseudo normal.
    This approach to sign determination is robust for well conditioned meshes that are watertight and non-self intersecting.
@@ -1358,18 +1355,13 @@ Geometry
    :param id: The mesh identifier
    :param point: The point in space to query
    :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                  Note that mesh must be watertight for this to be robust
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
    :param epsilon: Epsilon treating distance values as equal, when locating the minimum distance vertex/face/edge, as a
                    fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3
 
 
-.. function:: mesh_query_point_sign_winding_number(id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32, accuracy: float32, threshold: float32) -> bool
+.. function:: mesh_query_point_sign_winding_number(id: uint64, point: vec3f, max_dist: float32, accuracy: float32, threshold: float32) -> mesh_query_point_t
 
-   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given point in space. Returns ``True`` if a point < ``max_dist`` is found.
+   Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given point in space.
 
    Identifies the sign using the winding number of the mesh relative to the query point. This method of sign determination is robust for poorly conditioned meshes
    and provides a smooth approximation to sign even when the mesh is not watertight. This method is the most robust and accurate of the sign determination meshes
@@ -1380,36 +1372,24 @@ Geometry
    :param id: The mesh identifier
    :param point: The point in space to query
    :param max_dist: Mesh faces above this distance will not be considered by the query
-   :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                  Note that mesh must be watertight for this to be robust
-   :param face: Returns the index of the closest face
-   :param bary_u: Returns the barycentric u coordinate of the closest point
-   :param bary_v: Returns the barycentric v coordinate of the closest point
    :param accuracy: Accuracy for computing the winding number with fast winding number method utilizing second-order dipole approximation, default 2.0
    :param threshold: The threshold of the winding number to be considered inside, default 0.5
 
 
-.. function:: mesh_query_ray(id: uint64, start: vec3f, dir: vec3f, max_t: float32, t: float32, bary_u: float32, bary_v: float32, sign: float32, normal: vec3f, face: int32) -> bool
+.. function:: mesh_query_ray(id: uint64, start: vec3f, dir: vec3f, max_t: float32) -> mesh_query_ray_t
 
-   Computes the closest ray hit on the :class:`Mesh` with identifier ``id``, returns ``True`` if a point < ``max_t`` is found.
+   Computes the closest ray hit on the :class:`Mesh` with identifier ``id``.
 
    :param id: The mesh identifier
    :param start: The start point of the ray
    :param dir: The ray direction (should be normalized)
    :param max_t: The maximum distance along the ray to check for intersections
-   :param t: Returns the distance of the closest hit along the ray
-   :param bary_u: Returns the barycentric u coordinate of the closest hit
-   :param bary_v: Returns the barycentric v coordinate of the closest hit
-   :param sign: Returns a value > 0 if the hit ray hit front of the face, returns < 0 otherwise
-   :param normal: Returns the face normal
-   :param face: Returns the index of the hit face
 
 
 .. function:: mesh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> mesh_query_aabb_t
 
    Construct an axis-aligned bounding box query against a :class:`Mesh`.
    This query can be used to iterate over all triangles inside a volume.
-   Returns an object that is used to track state during mesh traversal.
 
    :param id: The mesh identifier
    :param lower: The lower bound of the bounding box in mesh space
@@ -1434,7 +1414,7 @@ Geometry
 
 .. function:: hash_grid_query(id: uint64, point: vec3f, max_dist: float32) -> hash_grid_query_t
 
-   Construct a point query against a :class:`HashGrid`. This query can be used to iterate over all neighboring points within a fixed radius from the query point. Returns an object that is used to track state during neighbor traversal.
+   Construct a point query against a :class:`HashGrid`. This query can be used to iterate over all neighboring points within a fixed radius from the query point.
 
 
 .. function:: hash_grid_query_next(query: hash_grid_query_t, index: int32) -> bool
diff --git a/docs/requirements.txt b/docs/requirements.txt
index 624e8cc23..c9b7984a5 100644
--- a/docs/requirements.txt
+++ b/docs/requirements.txt
@@ -1,3 +1,5 @@
 furo==2023.3.27
 sphinx==6.1.3
 sphinx_copybutton==0.5.2
+numpy==1.24.3
+black==23.12.0
diff --git a/tools/ci/building/build-linux-aarch64/build.sh b/tools/ci/building/build-linux-aarch64/build.sh
index 1087cf5dd..24c7e28d4 100755
--- a/tools/ci/building/build-linux-aarch64/build.sh
+++ b/tools/ci/building/build-linux-aarch64/build.sh
@@ -49,4 +49,3 @@ else
         $PYTHON "$SCRIPT_DIR/../../../../build_lib.py" --cuda_path=$CUDA --mode=$BUILD_MODE
     fi
 fi
-
diff --git a/tools/ci/building/build-windows-x86_64/build.bat b/tools/ci/building/build-windows-x86_64/build.bat
index ad5872fe3..0f715a1bd 100644
--- a/tools/ci/building/build-windows-x86_64/build.bat
+++ b/tools/ci/building/build-windows-x86_64/build.bat
@@ -13,7 +13,6 @@ call %PYTHON% -m pip install gitpython
 call %PYTHON% -m pip install cmake
 call %PYTHON% -m pip install ninja
 
-
 SET BUILD_MODE="release"
 for %%i in (%*) do (
     if "%%i"=="--debug" set BUILD_MODE="debug"
diff --git a/warp/__init__.py b/warp/__init__.py
index 3ef997ad6..0f146048e 100644
--- a/warp/__init__.py
+++ b/warp/__init__.py
@@ -59,7 +59,6 @@
 )
 from warp.context import set_module_options, get_module_options, get_module
 from warp.context import capture_begin, capture_end, capture_launch
-from warp.context import print_builtins, export_builtins, export_stubs
 from warp.context import Kernel, Function, Launch
 from warp.context import Stream, get_stream, set_stream, synchronize_stream
 from warp.context import Event, record_event, wait_event, wait_stream
diff --git a/warp/context.py b/warp/context.py
index d196b20f3..a56edfcd9 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -4075,7 +4075,7 @@ def print_function(f, file, noentry=False):  # pragma: no cover
     return True
 
 
-def print_builtins(file):  # pragma: no cover
+def export_functions_rst(file):  # pragma: no cover
     header = (
         "..\n"
         "   Autogenerated File - Do not edit. Run build_docs.py to generate.\n"
diff --git a/warp/stubs.py b/warp/stubs.py
index a3856efdd..d55f11cae 100644
--- a/warp/stubs.py
+++ b/warp/stubs.py
@@ -40,7 +40,8 @@
 from warp.types import spatial_matrix, spatial_matrixh, spatial_matrixf, spatial_matrixd
 
 from warp.types import Bvh, Mesh, HashGrid, Volume, MarchingCubes
-from warp.types import bvh_query_t, mesh_query_aabb_t, hash_grid_query_t
+from warp.types import bvh_query_t, hash_grid_query_t, mesh_query_aabb_t, mesh_query_point_t, mesh_query_ray_t
+
 
 from warp.types import matmul, adj_matmul, batched_matmul, adj_batched_matmul, from_ptr
 
@@ -69,7 +70,6 @@
 )
 from warp.context import set_module_options, get_module_options, get_module
 from warp.context import capture_begin, capture_end, capture_launch
-from warp.context import print_builtins, export_builtins, export_stubs
 from warp.context import Kernel, Function, Launch
 from warp.context import Stream, get_stream, set_stream, synchronize_stream
 from warp.context import Event, record_event, wait_event, wait_stream
@@ -868,7 +868,7 @@ def mlp(
 def bvh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> bvh_query_t:
     """
     Construct an axis-aligned bounding box query against a BVH object. This query can be used to iterate over all bounds
-       inside a BVH. Returns an object that is used to track state during BVH traversal.
+       inside a BVH.
 
        :param id: The BVH identifier
        :param lower: The lower bound of the bounding box in BVH space
@@ -881,7 +881,7 @@ def bvh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> bvh_query_t:
 def bvh_query_ray(id: uint64, start: vec3f, dir: vec3f) -> bvh_query_t:
     """
     Construct a ray query against a BVH object. This query can be used to iterate over all bounds
-       that intersect the ray. Returns an object that is used to track state during BVH traversal.
+       that intersect the ray.
 
        :param id: The BVH identifier
        :param start: The start of the ray in BVH space
@@ -900,11 +900,9 @@ def bvh_query_next(query: bvh_query_t, index: int32) -> bool:
 
 
 @over
-def mesh_query_point(
-    id: uint64, point: vec3f, max_dist: float32, inside: float32, face: int32, bary_u: float32, bary_v: float32
-) -> bool:
+def mesh_query_point(id: uint64, point: vec3f, max_dist: float32) -> mesh_query_point_t:
     """
-    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
+    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
 
        Identifies the sign of the distance using additional ray-casts to determine if the point is inside or outside.
        This method is relatively robust, but does increase computational cost.
@@ -913,66 +911,42 @@ def mesh_query_point(
        :param id: The mesh identifier
        :param point: The point in space to query
        :param max_dist: Mesh faces above this distance will not be considered by the query
-       :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                      Note that mesh must be watertight for this to be robust
-       :param face: Returns the index of the closest face
-       :param bary_u: Returns the barycentric u coordinate of the closest point
-       :param bary_v: Returns the barycentric v coordinate of the closest point
     """
     ...
 
 
 @over
-def mesh_query_point_no_sign(
-    id: uint64, point: vec3f, max_dist: float32, face: int32, bary_u: float32, bary_v: float32
-) -> bool:
+def mesh_query_point_no_sign(id: uint64, point: vec3f, max_dist: float32) -> mesh_query_point_t:
     """
-    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
+    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
 
        This method does not compute the sign of the point (inside/outside) which makes it faster than other point query methods.
 
        :param id: The mesh identifier
        :param point: The point in space to query
        :param max_dist: Mesh faces above this distance will not be considered by the query
-       :param face: Returns the index of the closest face
-       :param bary_u: Returns the barycentric u coordinate of the closest point
-       :param bary_v: Returns the barycentric v coordinate of the closest point
     """
     ...
 
 
 @over
-def mesh_query_furthest_point_no_sign(
-    id: uint64, point: vec3f, min_dist: float32, face: int32, bary_u: float32, bary_v: float32
-) -> bool:
+def mesh_query_furthest_point_no_sign(id: uint64, point: vec3f, min_dist: float32) -> mesh_query_point_t:
     """
-    Computes the furthest point on the mesh with identifier `id` to the given point in space. Returns ``True`` if a point > ``min_dist`` is found.
+    Computes the furthest point on the mesh with identifier `id` to the given point in space.
 
        This method does not compute the sign of the point (inside/outside).
 
        :param id: The mesh identifier
        :param point: The point in space to query
        :param min_dist: Mesh faces below this distance will not be considered by the query
-       :param face: Returns the index of the furthest face
-       :param bary_u: Returns the barycentric u coordinate of the furthest point
-       :param bary_v: Returns the barycentric v coordinate of the furthest point
     """
     ...
 
 
 @over
-def mesh_query_point_sign_normal(
-    id: uint64,
-    point: vec3f,
-    max_dist: float32,
-    inside: float32,
-    face: int32,
-    bary_u: float32,
-    bary_v: float32,
-    epsilon: float32,
-) -> bool:
+def mesh_query_point_sign_normal(id: uint64, point: vec3f, max_dist: float32, epsilon: float32) -> mesh_query_point_t:
     """
-    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space. Returns ``True`` if a point < ``max_dist`` is found.
+    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given ``point`` in space.
 
        Identifies the sign of the distance (inside/outside) using the angle-weighted pseudo normal.
        This approach to sign determination is robust for well conditioned meshes that are watertight and non-self intersecting.
@@ -981,11 +955,6 @@ def mesh_query_point_sign_normal(
        :param id: The mesh identifier
        :param point: The point in space to query
        :param max_dist: Mesh faces above this distance will not be considered by the query
-       :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                      Note that mesh must be watertight for this to be robust
-       :param face: Returns the index of the closest face
-       :param bary_u: Returns the barycentric u coordinate of the closest point
-       :param bary_v: Returns the barycentric v coordinate of the closest point
        :param epsilon: Epsilon treating distance values as equal, when locating the minimum distance vertex/face/edge, as a
                        fraction of the average edge length, also for treating closest point as being on edge/vertex default 1e-3
     """
@@ -994,18 +963,10 @@ def mesh_query_point_sign_normal(
 
 @over
 def mesh_query_point_sign_winding_number(
-    id: uint64,
-    point: vec3f,
-    max_dist: float32,
-    inside: float32,
-    face: int32,
-    bary_u: float32,
-    bary_v: float32,
-    accuracy: float32,
-    threshold: float32,
-) -> bool:
-    """
-    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given point in space. Returns ``True`` if a point < ``max_dist`` is found.
+    id: uint64, point: vec3f, max_dist: float32, accuracy: float32, threshold: float32
+) -> mesh_query_point_t:
+    """
+    Computes the closest point on the :class:`Mesh` with identifier ``id`` to the given point in space.
 
        Identifies the sign using the winding number of the mesh relative to the query point. This method of sign determination is robust for poorly conditioned meshes
        and provides a smooth approximation to sign even when the mesh is not watertight. This method is the most robust and accurate of the sign determination meshes
@@ -1016,11 +977,6 @@ def mesh_query_point_sign_winding_number(
        :param id: The mesh identifier
        :param point: The point in space to query
        :param max_dist: Mesh faces above this distance will not be considered by the query
-       :param inside: Returns a value < 0 if query point is inside the mesh, >=0 otherwise.
-                      Note that mesh must be watertight for this to be robust
-       :param face: Returns the index of the closest face
-       :param bary_u: Returns the barycentric u coordinate of the closest point
-       :param bary_v: Returns the barycentric v coordinate of the closest point
        :param accuracy: Accuracy for computing the winding number with fast winding number method utilizing second-order dipole approximation, default 2.0
        :param threshold: The threshold of the winding number to be considered inside, default 0.5
     """
@@ -1028,31 +984,14 @@ def mesh_query_point_sign_winding_number(
 
 
 @over
-def mesh_query_ray(
-    id: uint64,
-    start: vec3f,
-    dir: vec3f,
-    max_t: float32,
-    t: float32,
-    bary_u: float32,
-    bary_v: float32,
-    sign: float32,
-    normal: vec3f,
-    face: int32,
-) -> bool:
+def mesh_query_ray(id: uint64, start: vec3f, dir: vec3f, max_t: float32) -> mesh_query_ray_t:
     """
-    Computes the closest ray hit on the :class:`Mesh` with identifier ``id``, returns ``True`` if a point < ``max_t`` is found.
+    Computes the closest ray hit on the :class:`Mesh` with identifier ``id``.
 
        :param id: The mesh identifier
        :param start: The start point of the ray
        :param dir: The ray direction (should be normalized)
        :param max_t: The maximum distance along the ray to check for intersections
-       :param t: Returns the distance of the closest hit along the ray
-       :param bary_u: Returns the barycentric u coordinate of the closest hit
-       :param bary_v: Returns the barycentric v coordinate of the closest hit
-       :param sign: Returns a value > 0 if the hit ray hit front of the face, returns < 0 otherwise
-       :param normal: Returns the face normal
-       :param face: Returns the index of the hit face
     """
     ...
 
@@ -1062,7 +1001,6 @@ def mesh_query_aabb(id: uint64, lower: vec3f, upper: vec3f) -> mesh_query_aabb_t
     """
     Construct an axis-aligned bounding box query against a :class:`Mesh`.
        This query can be used to iterate over all triangles inside a volume.
-       Returns an object that is used to track state during mesh traversal.
 
        :param id: The mesh identifier
        :param lower: The lower bound of the bounding box in mesh space
@@ -1099,7 +1037,7 @@ def mesh_eval_velocity(id: uint64, face: int32, bary_u: float32, bary_v: float32
 @over
 def hash_grid_query(id: uint64, point: vec3f, max_dist: float32) -> hash_grid_query_t:
     """
-    Construct a point query against a :class:`HashGrid`. This query can be used to iterate over all neighboring points within a fixed radius from the query point. Returns an object that is used to track state during neighbor traversal.
+    Construct a point query against a :class:`HashGrid`. This query can be used to iterate over all neighboring points within a fixed radius from the query point.
     """
     ...
 
@@ -1863,8 +1801,6 @@ def atomic_min(a: Array[Any], i: int32, value: Any):
     Compute the minimum of ``value`` and ``a[i]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1875,8 +1811,6 @@ def atomic_min(a: Array[Any], i: int32, j: int32, value: Any):
     Compute the minimum of ``value`` and ``a[i,j]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1887,8 +1821,6 @@ def atomic_min(a: Array[Any], i: int32, j: int32, k: int32, value: Any):
     Compute the minimum of ``value`` and ``a[i,j,k]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1899,8 +1831,6 @@ def atomic_min(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any
     Compute the minimum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the minimum of ``a``.
     """
     ...
 
@@ -1991,8 +1921,6 @@ def atomic_max(a: Array[Any], i: int32, value: Any):
     Compute the maximum of ``value`` and ``a[i]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
@@ -2003,8 +1931,6 @@ def atomic_max(a: Array[Any], i: int32, j: int32, value: Any):
     Compute the maximum of ``value`` and ``a[i,j]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
@@ -2015,8 +1941,6 @@ def atomic_max(a: Array[Any], i: int32, j: int32, k: int32, value: Any):
     Compute the maximum of ``value`` and ``a[i,j,k]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 
@@ -2027,8 +1951,6 @@ def atomic_max(a: Array[Any], i: int32, j: int32, k: int32, l: int32, value: Any
     Compute the maximum of ``value`` and ``a[i,j,k,l]`` and atomically update the array.
 
     Note that for vectors and matrices the operation is only atomic on a per-component basis.
-
-    The gradient through this operation is accumulated for every element (or its component, in the case of vectors and matrices) of ``a`` that is equal to the maximum of ``a``.
     """
     ...
 

From 8260e416cf7a3126a87dd823c81ae14121aad125 Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Wed, 20 Dec 2023 11:20:14 -0800
Subject: [PATCH 47/50] Fix some possible device issues with unit testing

---
 examples/example_diffray.py          |  51 ++++----
 warp/fem/geometry/hexmesh.py         |  17 ++-
 warp/fem/geometry/quadmesh_2d.py     |  28 +++--
 warp/fem/geometry/tetmesh.py         |  27 +++--
 warp/fem/geometry/trimesh_2d.py      |  25 ++--
 warp/fem/space/partition.py          |  15 ++-
 warp/fem/utils.py                    |  15 ++-
 warp/tests/test_adam.py              | 174 +++++++++++++--------------
 warp/tests/test_bvh.py               |   2 +-
 warp/tests/test_examples.py          |  52 +++++---
 warp/thirdparty/unittest_parallel.py |   2 -
 11 files changed, 230 insertions(+), 178 deletions(-)

diff --git a/examples/example_diffray.py b/examples/example_diffray.py
index bca2c01f7..6a8ae9ee6 100644
--- a/examples/example_diffray.py
+++ b/examples/example_diffray.py
@@ -436,33 +436,36 @@ def __init__(self, stage=None, rot_array=[0.0, 0.0, 0.0, 1.0], verbose=False):
         )
 
     def ray_trace(self, is_live=False):
-        with wp.ScopedDevice("cuda:0"):
-            # raycast
-            wp.launch(
-                kernel=draw_kernel,
-                dim=self.num_rays,
-                inputs=[
-                    self.render_mesh,
-                    self.camera,
-                    self.texture,
-                    self.rays_width,
-                    self.rays_height,
-                    self.rays,
-                    self.lights,
-                    self.render_mode,
-                ],
-            )
-
-            # downsample
-            wp.launch(
-                kernel=downsample_kernel,
-                dim=self.num_pixels,
-                inputs=[self.rays, self.pixels, self.rays_width, pow(2, self.num_samples)],
-            )
+        # raycast
+        wp.launch(
+            kernel=draw_kernel,
+            dim=self.num_rays,
+            inputs=[
+                self.render_mesh,
+                self.camera,
+                self.texture,
+                self.rays_width,
+                self.rays_height,
+                self.rays,
+                self.lights,
+                self.render_mode,
+            ],
+            device=self.device,
+        )
+
+        # downsample
+        wp.launch(
+            kernel=downsample_kernel,
+            dim=self.num_pixels,
+            inputs=[self.rays, self.pixels, self.rays_width, pow(2, self.num_samples)],
+            device=self.device,
+        )
 
     def compute_loss(self):
         self.ray_trace()
-        wp.launch(loss_kernel, dim=self.num_pixels, inputs=[self.pixels, self.target_pixels, self.loss])
+        wp.launch(
+            loss_kernel, dim=self.num_pixels, inputs=[self.pixels, self.target_pixels, self.loss], device=self.device
+        )
 
     def get_image(self):
         return self.pixels.numpy().reshape((self.height, self.width, 3))
diff --git a/warp/fem/geometry/hexmesh.py b/warp/fem/geometry/hexmesh.py
index 09da0664e..9370b4ea1 100644
--- a/warp/fem/geometry/hexmesh.py
+++ b/warp/fem/geometry/hexmesh.py
@@ -1,11 +1,16 @@
 from typing import Optional
-import warp as wp
 
-from warp.fem.types import ElementIndex, Coords, Sample, OUTSIDE, make_free_sample
-from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary, borrow_temporary_like
+import warp as wp
+from warp.fem.cache import (
+    TemporaryStore,
+    borrow_temporary,
+    borrow_temporary_like,
+    cached_arg_value,
+)
+from warp.fem.types import OUTSIDE, Coords, ElementIndex, Sample, make_free_sample
 
+from .element import Cube, Square
 from .geometry import Geometry
-from .element import Square, Cube
 
 
 @wp.struct
@@ -493,7 +498,7 @@ def _build_topology(self, temporary_store: TemporaryStore):
             wp.copy(
                 dest=face_count.array, src=vertex_unique_face_offsets.array, src_offset=self.vertex_count() - 1, count=1
             )
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
             face_count = int(face_count.array.numpy()[0])
         else:
             face_count = int(vertex_unique_face_offsets.array.numpy()[self.vertex_count() - 1])
@@ -603,7 +608,7 @@ def _compute_hex_edges(self, temporary_store: Optional[TemporaryStore] = None):
                 src_offset=self.vertex_count() - 1,
                 count=1,
             )
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
             self._edge_count = int(edge_count.array.numpy()[0])
         else:
             self._edge_count = int(vertex_unique_edge_offsets.array.numpy()[self.vertex_count() - 1])
diff --git a/warp/fem/geometry/quadmesh_2d.py b/warp/fem/geometry/quadmesh_2d.py
index 4a8b0bf6e..6919f125e 100644
--- a/warp/fem/geometry/quadmesh_2d.py
+++ b/warp/fem/geometry/quadmesh_2d.py
@@ -1,12 +1,16 @@
 from typing import Optional
-import warp as wp
-
-from warp.fem.types import ElementIndex, Coords, Sample
-from warp.fem.types import OUTSIDE, make_free_sample
-from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary, borrow_temporary_like
 
+import warp as wp
+from warp.fem.cache import (
+    TemporaryStore,
+    borrow_temporary,
+    borrow_temporary_like,
+    cached_arg_value,
+)
+from warp.fem.types import OUTSIDE, Coords, ElementIndex, Sample, make_free_sample
+
+from .element import LinearEdge, Square
 from .geometry import Geometry
-from .element import Square, LinearEdge
 
 # from .closest_point import project_on_tet_at_origin
 
@@ -120,7 +124,7 @@ def cell_position(args: CellArg, s: Sample):
         )
 
     @wp.func
-    def cell_deformation_gradient(cell_arg: CellArg, s:Sample):
+    def cell_deformation_gradient(cell_arg: CellArg, s: Sample):
         """Deformation gradient at `coords`"""
         quad_idx = cell_arg.quad_vertex_indices[s.element_index]
 
@@ -135,7 +139,7 @@ def cell_deformation_gradient(cell_arg: CellArg, s:Sample):
         )
 
     @wp.func
-    def cell_inverse_deformation_gradient(cell_arg: CellArg, s:Sample):
+    def cell_inverse_deformation_gradient(cell_arg: CellArg, s: Sample):
         return wp.inverse(Quadmesh2D.cell_deformation_gradient(cell_arg, s))
 
     @wp.func
@@ -178,20 +182,20 @@ def side_position(args: SideArg, s: Sample):
         ] * args.cell_arg.positions[edge_idx[1]]
 
     @wp.func
-    def side_deformation_gradient(args: SideArg, s:Sample):
+    def side_deformation_gradient(args: SideArg, s: Sample):
         edge_idx = args.edge_vertex_indices[s.element_index]
         v0 = args.cell_arg.positions[edge_idx[0]]
         v1 = args.cell_arg.positions[edge_idx[1]]
         return v1 - v0
 
     @wp.func
-    def side_inner_inverse_deformation_gradient(args: SideArg, s:Sample):
+    def side_inner_inverse_deformation_gradient(args: SideArg, s: Sample):
         cell_index = Quadmesh2D.side_inner_cell_index(args, s.element_index)
         cell_coords = Quadmesh2D.side_inner_cell_coords(args, s.element_index, s.element_coords)
         return Quadmesh2D.cell_inverse_deformation_gradient(args.cell_arg, make_free_sample(cell_index, cell_coords))
 
     @wp.func
-    def side_outer_inverse_deformation_gradient(args: SideArg, s:Sample):
+    def side_outer_inverse_deformation_gradient(args: SideArg, s: Sample):
         cell_index = Quadmesh2D.side_outer_cell_index(args, s.element_index)
         cell_coords = Quadmesh2D.side_outer_cell_coords(args, s.element_index, s.element_coords)
         return Quadmesh2D.cell_inverse_deformation_gradient(args.cell_arg, make_free_sample(cell_index, cell_coords))
@@ -352,7 +356,7 @@ def _build_topology(self, temporary_store: TemporaryStore):
             wp.copy(
                 dest=edge_count.array, src=vertex_unique_edge_offsets.array, src_offset=self.vertex_count() - 1, count=1
             )
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
             edge_count = int(edge_count.array.numpy()[0])
         else:
             edge_count = int(vertex_unique_edge_offsets.array.numpy()[self.vertex_count() - 1])
diff --git a/warp/fem/geometry/tetmesh.py b/warp/fem/geometry/tetmesh.py
index 82f92d84b..6d4e6c899 100644
--- a/warp/fem/geometry/tetmesh.py
+++ b/warp/fem/geometry/tetmesh.py
@@ -1,13 +1,24 @@
 from typing import Optional
-import warp as wp
 
-from warp.fem.types import ElementIndex, Coords, Sample, make_free_sample
-from warp.fem.types import NULL_ELEMENT_INDEX, OUTSIDE
-from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary, borrow_temporary_like
+import warp as wp
+from warp.fem.cache import (
+    TemporaryStore,
+    borrow_temporary,
+    borrow_temporary_like,
+    cached_arg_value,
+)
+from warp.fem.types import (
+    NULL_ELEMENT_INDEX,
+    OUTSIDE,
+    Coords,
+    ElementIndex,
+    Sample,
+    make_free_sample,
+)
 
-from .geometry import Geometry
-from .element import Triangle, Tetrahedron
 from .closest_point import project_on_tet_at_origin
+from .element import Tetrahedron, Triangle
+from .geometry import Geometry
 
 
 @wp.struct
@@ -401,7 +412,7 @@ def _build_topology(self, temporary_store: TemporaryStore):
             wp.copy(
                 dest=face_count.array, src=vertex_unique_face_offsets.array, src_offset=self.vertex_count() - 1, count=1
             )
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
             face_count = int(face_count.array.numpy()[0])
         else:
             face_count = int(vertex_unique_face_offsets.array.numpy()[self.vertex_count() - 1])
@@ -497,7 +508,7 @@ def _compute_tet_edges(self, temporary_store: Optional[TemporaryStore] = None):
                 src_offset=self.vertex_count() - 1,
                 count=1,
             )
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
             self._edge_count = int(edge_count.array.numpy()[0])
         else:
             self._edge_count = int(vertex_unique_edge_offsets.array.numpy()[self.vertex_count() - 1])
diff --git a/warp/fem/geometry/trimesh_2d.py b/warp/fem/geometry/trimesh_2d.py
index 7e13fcf72..3125ffcfc 100644
--- a/warp/fem/geometry/trimesh_2d.py
+++ b/warp/fem/geometry/trimesh_2d.py
@@ -1,13 +1,24 @@
 from typing import Optional
-import warp as wp
 
-from warp.fem.types import ElementIndex, Coords, Sample, make_free_sample
-from warp.fem.types import NULL_ELEMENT_INDEX, OUTSIDE
-from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary, borrow_temporary_like
+import warp as wp
+from warp.fem.cache import (
+    TemporaryStore,
+    borrow_temporary,
+    borrow_temporary_like,
+    cached_arg_value,
+)
+from warp.fem.types import (
+    NULL_ELEMENT_INDEX,
+    OUTSIDE,
+    Coords,
+    ElementIndex,
+    Sample,
+    make_free_sample,
+)
 
-from .geometry import Geometry
-from .element import Triangle, LinearEdge
 from .closest_point import project_on_tri_at_origin
+from .element import LinearEdge, Triangle
+from .geometry import Geometry
 
 
 @wp.struct
@@ -365,7 +376,7 @@ def _build_topology(self, temporary_store: TemporaryStore):
             wp.copy(
                 dest=edge_count.array, src=vertex_unique_edge_offsets.array, src_offset=self.vertex_count() - 1, count=1
             )
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
             edge_count = int(edge_count.array.numpy()[0])
         else:
             edge_count = int(vertex_unique_edge_offsets.array.numpy()[self.vertex_count() - 1])
diff --git a/warp/fem/space/partition.py b/warp/fem/space/partition.py
index fc6126771..d9ed3bbc2 100644
--- a/warp/fem/space/partition.py
+++ b/warp/fem/space/partition.py
@@ -1,15 +1,18 @@
 from typing import Any, Optional, Union
 
 import warp as wp
-
+from warp.fem.cache import (
+    TemporaryStore,
+    borrow_temporary,
+    borrow_temporary_like,
+    cached_arg_value,
+)
 from warp.fem.geometry import GeometryPartition, WholeGeometryPartition
-from warp.fem.utils import compress_node_indices, _iota_kernel
 from warp.fem.types import NULL_NODE_INDEX
-from warp.fem.cache import cached_arg_value, TemporaryStore, borrow_temporary, borrow_temporary_like
+from warp.fem.utils import _iota_kernel, compress_node_indices
 
-from .topology import SpaceTopology
 from .function_space import FunctionSpace
-
+from .topology import SpaceTopology
 
 wp.set_module_options({"enable_backward": False})
 
@@ -272,7 +275,7 @@ def _finalize_node_indices(self, node_category: wp.array(dtype=int), temporary_s
 
         if device.is_cuda:
             # TODO switch to synchronize_event once available
-            wp.synchronize_stream(wp.get_stream())
+            wp.synchronize_stream(wp.get_stream(device))
 
         category_offsets.release()
 
diff --git a/warp/fem/utils.py b/warp/fem/utils.py
index 126f83132..a03db5472 100644
--- a/warp/fem/utils.py
+++ b/warp/fem/utils.py
@@ -1,10 +1,15 @@
 from typing import Any, Tuple
 
-import warp as wp
 import numpy as np
 
-from warp.utils import radix_sort_pairs, runlength_encode, array_scan
-from warp.fem.cache import borrow_temporary, borrow_temporary_like, TemporaryStore, Temporary
+import warp as wp
+from warp.fem.cache import (
+    Temporary,
+    TemporaryStore,
+    borrow_temporary,
+    borrow_temporary_like,
+)
+from warp.utils import array_scan, radix_sort_pairs, runlength_encode
 
 
 @wp.func
@@ -168,7 +173,7 @@ def compress_node_indices(
     if node_indices.device.is_cuda:
         unique_node_count_host = borrow_temporary(temporary_store, shape=(1,), dtype=int, pinned=True, device="cpu")
         wp.copy(src=unique_node_count_dev.array, dest=unique_node_count_host.array, count=1)
-        wp.synchronize_stream(wp.get_stream())
+        wp.synchronize_stream(wp.get_stream(node_indices.device))
         unique_node_count_dev.release()
         unique_node_count = int(unique_node_count_host.array.numpy()[0])
         unique_node_count_host.release()
@@ -217,7 +222,7 @@ def masked_indices(
     if offsets.device.is_cuda:
         masked_count_temp = borrow_temporary(temporary_store, shape=1, dtype=int, pinned=True, device="cpu")
         wp.copy(dest=masked_count_temp.array, src=offsets, src_offset=offsets.shape[0] - 1, count=1)
-        wp.synchronize_stream(wp.get_stream())
+        wp.synchronize_stream(wp.get_stream(offsets.device))
         masked_count = int(masked_count_temp.array.numpy()[0])
         masked_count_temp.release()
     else:
diff --git a/warp/tests/test_adam.py b/warp/tests/test_adam.py
index c104df306..d098d2410 100644
--- a/warp/tests/test_adam.py
+++ b/warp/tests/test_adam.py
@@ -26,32 +26,32 @@ def objective(params: wp.array(dtype=float), score: wp.array(dtype=float)):
 
 # This test inspired by https://machinelearningmastery.com/adam-optimization-from-scratch/
 def test_adam_solve_float(test, device):
-    wp.set_device(device)
-    params_start = np.array([0.1, 0.2], dtype=float)
-    score = wp.zeros(1, dtype=float, requires_grad=True)
-    params = wp.array(params_start, dtype=float, requires_grad=True)
-    tape = wp.Tape()
-    opt = warp.optim.Adam([params], lr=0.02, betas=(0.8, 0.999))
-
-    def gradient_func():
-        tape.reset()
-        score.zero_()
-        with tape:
-            wp.launch(kernel=objective, dim=len(params), inputs=[params, score])
-        tape.backward(score)
-        return [tape.gradients[params]]
-
-    niters = 100
-
-    opt.reset_internal_state()
-    for _ in range(niters):
-        opt.step(gradient_func())
-
-    result = params.numpy()
-    # optimum is at the origin, so the result should be close to it in all N dimensions.
-    tol = 1e-5
-    for r in result:
-        test.assertLessEqual(r, tol)
+    with wp.ScopedDevice(device):
+        params_start = np.array([0.1, 0.2], dtype=float)
+        score = wp.zeros(1, dtype=float, requires_grad=True)
+        params = wp.array(params_start, dtype=float, requires_grad=True)
+        tape = wp.Tape()
+        opt = warp.optim.Adam([params], lr=0.02, betas=(0.8, 0.999))
+
+        def gradient_func():
+            tape.reset()
+            score.zero_()
+            with tape:
+                wp.launch(kernel=objective, dim=len(params), inputs=[params, score])
+            tape.backward(score)
+            return [tape.gradients[params]]
+
+        niters = 100
+
+        opt.reset_internal_state()
+        for _ in range(niters):
+            opt.step(gradient_func())
+
+        result = params.numpy()
+        # optimum is at the origin, so the result should be close to it in all N dimensions.
+        tol = 1e-5
+        for r in result:
+            test.assertLessEqual(r, tol)
 
 
 @wp.kernel
@@ -63,32 +63,32 @@ def objective_vec3(params: wp.array(dtype=wp.vec3), score: wp.array(dtype=float)
 
 # This test inspired by https://machinelearningmastery.com/adam-optimization-from-scratch/
 def test_adam_solve_vec3(test, device):
-    wp.set_device(device)
-    params_start = np.array([[0.1, 0.2, -0.1]], dtype=float)
-    score = wp.zeros(1, dtype=float, requires_grad=True)
-    params = wp.array(params_start, dtype=wp.vec3, requires_grad=True)
-    tape = wp.Tape()
-    opt = warp.optim.Adam([params], lr=0.02, betas=(0.8, 0.999))
-
-    def gradient_func():
-        tape.reset()
-        score.zero_()
-        with tape:
-            wp.launch(kernel=objective_vec3, dim=len(params), inputs=[params, score])
-        tape.backward(score)
-        return [tape.gradients[params]]
-
-    niters = 100
-    opt.reset_internal_state()
-    for _ in range(niters):
-        opt.step(gradient_func())
-
-    result = params.numpy()
-    tol = 1e-5
-    # optimum is at the origin, so the result should be close to it in all N dimensions.
-    for r in result:
-        for v in r:
-            test.assertLessEqual(v, tol)
+    with wp.ScopedDevice(device):
+        params_start = np.array([[0.1, 0.2, -0.1]], dtype=float)
+        score = wp.zeros(1, dtype=float, requires_grad=True)
+        params = wp.array(params_start, dtype=wp.vec3, requires_grad=True)
+        tape = wp.Tape()
+        opt = warp.optim.Adam([params], lr=0.02, betas=(0.8, 0.999))
+
+        def gradient_func():
+            tape.reset()
+            score.zero_()
+            with tape:
+                wp.launch(kernel=objective_vec3, dim=len(params), inputs=[params, score])
+            tape.backward(score)
+            return [tape.gradients[params]]
+
+        niters = 100
+        opt.reset_internal_state()
+        for _ in range(niters):
+            opt.step(gradient_func())
+
+        result = params.numpy()
+        tol = 1e-5
+        # optimum is at the origin, so the result should be close to it in all N dimensions.
+        for r in result:
+            for v in r:
+                test.assertLessEqual(v, tol)
 
 
 @wp.kernel
@@ -103,41 +103,41 @@ def objective_two_inputs_vec3(
 
 # This test inspired by https://machinelearningmastery.com/adam-optimization-from-scratch/
 def test_adam_solve_two_inputs(test, device):
-    wp.set_device(device)
-    params_start1 = np.array([[0.1, 0.2, -0.1]], dtype=float)
-    params_start2 = np.array([[0.2, 0.1, 0.1]], dtype=float)
-    score = wp.zeros(1, dtype=float, requires_grad=True)
-    params1 = wp.array(params_start1, dtype=wp.vec3, requires_grad=True)
-    params2 = wp.array(params_start2, dtype=wp.vec3, requires_grad=True)
-    tape = wp.Tape()
-    opt = warp.optim.Adam([params1, params2], lr=0.02, betas=(0.8, 0.999))
-
-    def gradient_func():
-        tape.reset()
-        score.zero_()
-        with tape:
-            wp.launch(kernel=objective_two_inputs_vec3, dim=len(params1), inputs=[params1, params2, score])
-        tape.backward(score)
-        return [tape.gradients[params1], tape.gradients[params2]]
-
-    niters = 100
-    opt.reset_internal_state()
-    for _ in range(niters):
-        opt.step(gradient_func())
-
-    result = params1.numpy()
-    tol = 1e-5
-    # optimum is at the origin, so the result should be close to it in all N dimensions.
-    for r in result:
-        for v in r:
-            test.assertLessEqual(v, tol)
-
-    result = params2.numpy()
-    tol = 1e-5
-    # optimum is at the origin, so the result should be close to it in all N dimensions.
-    for r in result:
-        for v in r:
-            test.assertLessEqual(v, tol)
+    with wp.ScopedDevice(device):
+        params_start1 = np.array([[0.1, 0.2, -0.1]], dtype=float)
+        params_start2 = np.array([[0.2, 0.1, 0.1]], dtype=float)
+        score = wp.zeros(1, dtype=float, requires_grad=True)
+        params1 = wp.array(params_start1, dtype=wp.vec3, requires_grad=True)
+        params2 = wp.array(params_start2, dtype=wp.vec3, requires_grad=True)
+        tape = wp.Tape()
+        opt = warp.optim.Adam([params1, params2], lr=0.02, betas=(0.8, 0.999))
+
+        def gradient_func():
+            tape.reset()
+            score.zero_()
+            with tape:
+                wp.launch(kernel=objective_two_inputs_vec3, dim=len(params1), inputs=[params1, params2, score])
+            tape.backward(score)
+            return [tape.gradients[params1], tape.gradients[params2]]
+
+        niters = 100
+        opt.reset_internal_state()
+        for _ in range(niters):
+            opt.step(gradient_func())
+
+        result = params1.numpy()
+        tol = 1e-5
+        # optimum is at the origin, so the result should be close to it in all N dimensions.
+        for r in result:
+            for v in r:
+                test.assertLessEqual(v, tol)
+
+        result = params2.numpy()
+        tol = 1e-5
+        # optimum is at the origin, so the result should be close to it in all N dimensions.
+        for r in result:
+            for v in r:
+                test.assertLessEqual(v, tol)
 
 
 devices = get_test_devices()
diff --git a/warp/tests/test_bvh.py b/warp/tests/test_bvh.py
index d749b9a37..897bc6504 100644
--- a/warp/tests/test_bvh.py
+++ b/warp/tests/test_bvh.py
@@ -131,7 +131,7 @@ def test_bvh_query_ray(test, device):
     test_bvh(test, "ray", device)
 
 
-devices = get_test_devices(mode="basic")  # TODO: Fix the illegal address error when two CUDA devices are used
+devices = get_test_devices()
 
 
 class TestBvh(unittest.TestCase):
diff --git a/warp/tests/test_examples.py b/warp/tests/test_examples.py
index b0999d2c6..9e9f2c6a0 100644
--- a/warp/tests/test_examples.py
+++ b/warp/tests/test_examples.py
@@ -10,7 +10,7 @@
 import unittest
 
 import warp as wp
-from warp.tests.unittest_utils import get_unique_cuda_test_devices
+from warp.tests.unittest_utils import get_unique_cuda_test_devices, sanitize_identifier
 
 wp.init()
 
@@ -18,11 +18,15 @@
 # registers an example to run as a TestCase
 def add_example_test(cls, name, devices=None, options={}):
     def run(test, device):
+        # The copy() is needed because pop() is used to avoid passing extra args to Example()
+        # Can remove once all examples accept **kwargs and we no longer need to pop()
+        test_options = options.copy()
+
         try:
             module = importlib.import_module(f"examples.{name}")
 
-            torch_cuda_required = options.setdefault("torch_cuda_required", False)
-            options.pop("torch_cuda_required", None)
+            torch_cuda_required = test_options.setdefault("torch_cuda_required", False)
+            test_options.pop("torch_cuda_required", None)
             if torch_cuda_required and wp.get_device(device).is_cuda:
                 # Ensure torch has CUDA support
                 import torch
@@ -34,33 +38,41 @@ def run(test, device):
             test.skipTest(f"{e}")
 
         # create default USD stage output path which many examples expect
-        options.setdefault("stage", os.path.join(os.path.dirname(__file__), f"outputs/{name}.usd"))
+        test_options.setdefault(
+            "stage", os.path.join(os.path.dirname(__file__), f"outputs/{name}_{sanitize_identifier(device)}.usd")
+        )
+
+        try:
+            os.remove(test_options["stage"])
+        except OSError:
+            pass
 
-        num_frames = options.get("num_frames", 10)
-        options.pop("num_frames", None)
+        num_frames = test_options.get("num_frames", 10)
+        test_options.pop("num_frames", None)
 
         # Don't want to force load all modules by default for serial test runner
         wp.config.graph_capture_module_load_default = False
 
         try:
-            enable_backward = options.get("enable_backward", True)
+            enable_backward = test_options.get("enable_backward", True)
             wp.set_module_options({"enable_backward": enable_backward}, module)
-            options.pop("enable_backward", None)
+            test_options.pop("enable_backward", None)
 
-            wp.load_module(module, device=wp.get_device())
-            extra_load_modules = options.get("load_modules", [])
-            for module_name in extra_load_modules:
-                wp.load_module(module_name, device=wp.get_device())
-            options.pop("load_modules", None)
+            with wp.ScopedDevice(device):
+                wp.load_module(module, device=wp.get_device())
+                extra_load_modules = test_options.get("load_modules", [])
+                for module_name in extra_load_modules:
+                    wp.load_module(module_name, device=wp.get_device())
+                test_options.pop("load_modules", None)
 
-            e = module.Example(**options)
+                e = module.Example(**test_options)
 
-            # disable scoped timer to avoid log spam from time steps
-            wp.ScopedTimer.enabled = False
+                # disable scoped timer to avoid log spam from time steps
+                wp.ScopedTimer.enabled = False
 
-            for _ in range(num_frames):
-                e.update()
-                e.render()
+                for _ in range(num_frames):
+                    e.update()
+                    e.render()
         except Exception as e:
             test.fail(f"{e}")
         finally:
@@ -109,7 +121,7 @@ class TestSimExamples(unittest.TestCase):
     TestSimExamples,
     name="example_sim_cartpole",
     devices=cuda_test_devices,
-    options={"load_modules": ["warp.sim.integrator_euler"]},
+    options={"load_modules": ["warp.sim.collide", "warp.sim.integrator_euler", "warp.sim.articulation"]},
 )
 add_example_test(
     TestSimExamples,
diff --git a/warp/thirdparty/unittest_parallel.py b/warp/thirdparty/unittest_parallel.py
index a5eefca0d..d04d2533c 100644
--- a/warp/thirdparty/unittest_parallel.py
+++ b/warp/thirdparty/unittest_parallel.py
@@ -481,8 +481,6 @@ def set_worker_cache(args, temp_dir):
 
         wp.config.kernel_cache_dir = cache_root_dir
 
-        # wp.config.verify_cuda = True  # TESTING
-
         wp.build.clear_kernel_cache()
 
 

From 5c7af7d7a9f7257889a4904013f96594b943352e Mon Sep 17 00:00:00 2001
From: Eric Shi <ershi@nvidia.com>
Date: Wed, 3 Jan 2024 16:12:20 -0800
Subject: [PATCH 48/50] Fix assert_np_equal bug

---
 copyright.py                          |  2 +-
 examples/example_sim_cartpole.py      |  2 +-
 examples/example_sim_quadruped.py     |  2 +-
 examples/example_sim_rigid_chain.py   |  2 +-
 examples/example_sim_rigid_contact.py |  2 +-
 warp/codegen.py                       |  6 +++---
 warp/context.py                       | 12 ++++++------
 warp/tape.py                          |  2 +-
 warp/tests/test_matmul_lite.py        |  1 +
 warp/tests/unittest_utils.py          |  2 +-
 10 files changed, 17 insertions(+), 16 deletions(-)

diff --git a/copyright.py b/copyright.py
index cd852e9e5..e7a6a64e7 100644
--- a/copyright.py
+++ b/copyright.py
@@ -47,7 +47,7 @@
         s = f.read()
         f.close()
 
-        if s.startswith(header) is False:
+        if not s.startswith(header):
             s = header + s
 
             f = open(p, "wt")
diff --git a/examples/example_sim_cartpole.py b/examples/example_sim_cartpole.py
index 7f1d2779a..2e497f062 100644
--- a/examples/example_sim_cartpole.py
+++ b/examples/example_sim_cartpole.py
@@ -108,7 +108,7 @@ def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=T
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True, print=self.print_timers):
-            if self.use_graph is False or self.graph is None:
+            if not self.use_graph or self.graph is None:
                 for _ in range(self.sim_substeps):
                     self.state.clear_forces()
                     self.state = self.integrator.simulate(self.model, self.state, self.state, self.sim_dt)
diff --git a/examples/example_sim_quadruped.py b/examples/example_sim_quadruped.py
index 9fea24476..55ef8f00d 100644
--- a/examples/example_sim_quadruped.py
+++ b/examples/example_sim_quadruped.py
@@ -144,7 +144,7 @@ def __init__(self, stage=None, num_envs=1, enable_rendering=True, print_timers=T
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True, print=self.print_timers):
-            if self.use_graph is False or self.graph is None:
+            if not self.use_graph or self.graph is None:
                 for _ in range(self.sim_substeps):
                     self.state_0.clear_forces()
                     wp.sim.collide(self.model, self.state_0)
diff --git a/examples/example_sim_rigid_chain.py b/examples/example_sim_rigid_chain.py
index fd33ec03e..d5625fb6f 100644
--- a/examples/example_sim_rigid_chain.py
+++ b/examples/example_sim_rigid_chain.py
@@ -155,7 +155,7 @@ def __init__(self, stage):
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True):
-            if self.use_graph is False or self.graph is None:
+            if not self.use_graph or self.graph is None:
                 for _ in range(self.sim_substeps):
                     self.state_0.clear_forces()
                     self.integrator.simulate(self.model, self.state_0, self.state_1, self.sim_dt)
diff --git a/examples/example_sim_rigid_contact.py b/examples/example_sim_rigid_contact.py
index d3c84917e..84568dfed 100644
--- a/examples/example_sim_rigid_contact.py
+++ b/examples/example_sim_rigid_contact.py
@@ -144,7 +144,7 @@ def load_mesh(self, filename, path):
 
     def update(self):
         with wp.ScopedTimer("simulate", active=True):
-            if self.use_graph is False or self.graph is None:
+            if not self.use_graph or self.graph is None:
                 for _ in range(self.sim_substeps):
                     self.state_0.clear_forces()
                     wp.sim.collide(self.model, self.state_0)
diff --git a/warp/codegen.py b/warp/codegen.py
index 4dd04fec7..1cb1551bc 100644
--- a/warp/codegen.py
+++ b/warp/codegen.py
@@ -1381,12 +1381,12 @@ def emit_Ellipsis(adj, node):
         return
 
     def emit_NameConstant(adj, node):
-        if node.value is True:
+        if node.value:
             return adj.add_constant(True)
-        elif node.value is False:
-            return adj.add_constant(False)
         elif node.value is None:
             raise WarpCodegenTypeError("None type unsupported")
+        else:
+            return adj.add_constant(False)
 
     def emit_Constant(adj, node):
         if isinstance(node, ast.Str):
diff --git a/warp/context.py b/warp/context.py
index 7d1a3a4fb..ba3368de3 100644
--- a/warp/context.py
+++ b/warp/context.py
@@ -1090,7 +1090,7 @@ def typelist(param):
 
         # export means the function will be added to the `warp` module namespace
         # so that users can call it directly from the Python interpreter
-        if export is True:
+        if export:
             if hasattr(warp, key):
                 # check that we haven't already created something at this location
                 # if it's just an overload stub for auto-complete then overwrite it
@@ -1440,7 +1440,7 @@ def hash_recursive(module, visited):
                         ch.update(bytes(s, "utf-8"))
                     if func.custom_replay_func:
                         s = func.custom_replay_func.adj.source
-                        
+
                     # cache func arg types
                     for arg, arg_type in func.adj.arg_types.items():
                         s = f"{arg}: {get_type_name(arg_type)}"
@@ -3494,7 +3494,7 @@ def launch(
         device = runtime.get_device(device)
 
     # check function is a Kernel
-    if isinstance(kernel, Kernel) is False:
+    if not isinstance(kernel, Kernel):
         raise RuntimeError("Error launching kernel, can only launch functions decorated with @wp.kernel.")
 
     # debugging aid
@@ -3795,7 +3795,7 @@ def capture_begin(device: Devicelike = None, stream=None, force_module_load=None
     if force_module_load is None:
         force_module_load = warp.config.graph_capture_module_load_default
 
-    if warp.config.verify_cuda is True:
+    if warp.config.verify_cuda:
         raise RuntimeError("Cannot use CUDA error verification during graph capture")
 
     if stream is not None:
@@ -4210,7 +4210,7 @@ def export_stubs(file):  # pragma: no cover
 
             return_str = ""
 
-            if f.export is False or f.hidden is True:  # or f.generic:
+            if not f.export or f.hidden:  # or f.generic:
                 continue
 
             try:
@@ -4255,7 +4255,7 @@ def ctype_ret_str(t):
 
     for k, g in builtin_functions.items():
         for f in g.overloads:
-            if f.export is False or f.generic:
+            if not f.export or f.generic:
                 continue
 
             simple = True
diff --git a/warp/tape.py b/warp/tape.py
index 12769003b..194097c6d 100644
--- a/warp/tape.py
+++ b/warp/tape.py
@@ -82,7 +82,7 @@ def backward(self, loss: wp.array = None, grads: dict = None):
             if loss.size > 1 or wp.types.type_length(loss.dtype) > 1:
                 raise RuntimeError("Can only return gradients for scalar loss functions.")
 
-            if loss.requires_grad is False:
+            if not loss.requires_grad:
                 raise RuntimeError(
                     "Scalar loss arrays should have requires_grad=True set before calling Tape.backward()"
                 )
diff --git a/warp/tests/test_matmul_lite.py b/warp/tests/test_matmul_lite.py
index d3e00d5a8..6bb6a8824 100644
--- a/warp/tests/test_matmul_lite.py
+++ b/warp/tests/test_matmul_lite.py
@@ -293,6 +293,7 @@ def test_tape(test, device):
 
     tape.backward(loss=loss)
     A_grad = A.grad.numpy()
+    tape.reset()
 
     # test adjoint
     D.grad = wp.array2d(np.ones((m, n)), dtype=float, device=device)
diff --git a/warp/tests/unittest_utils.py b/warp/tests/unittest_utils.py
index 39f50fcaa..356f8836b 100644
--- a/warp/tests/unittest_utils.py
+++ b/warp/tests/unittest_utils.py
@@ -172,7 +172,7 @@ def assert_np_equal(result, expect, tol=0.0):
     b = expect.flatten()
 
     if tol == 0.0:
-        if (a == b).all() is False:
+        if not (a == b).all():
             raise AssertionError(f"Unexpected result, got: {a} expected: {b}")
 
     else:

From 71db0b8be60d776fb650d32139af324a31ba8bb8 Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Fri, 5 Jan 2024 00:51:48 -0500
Subject: [PATCH 49/50] Consistently use "aarch64" as the AArch64 architecture
 string

The official name for Arm's 64-bit instruction set architecture is
AArch64. Apple's implementation of it is called ARM64, and they will
return "arm64" for `platform.machine()`.
---
 build_llvm.py     | 39 ++++++++++-----------------------------
 warp/build_dll.py | 23 +++++++++++++++++++----
 2 files changed, 29 insertions(+), 33 deletions(-)

diff --git a/build_llvm.py b/build_llvm.py
index 975f96e7b..1f7fc331d 100644
--- a/build_llvm.py
+++ b/build_llvm.py
@@ -1,10 +1,9 @@
 import os
 import subprocess
 import sys
-import platform
 
 import warp
-from warp.build_dll import build_dll_for_arch, run_cmd
+from warp.build_dll import *
 
 # set build output path off this file
 base_path = os.path.dirname(os.path.realpath(__file__))
@@ -15,18 +14,6 @@
 llvm_install_path = os.path.join(llvm_project_path, "out/install/")
 
 
-# return a canonical machine architecture string
-# - "x86_64" for x86-64, aka. AMD64, aka. x64
-# - "aarch64" for AArch64, aka. ARM64
-def machine_architecture() -> str:
-    machine = platform.machine()
-    if machine == "x86_64" or machine == "AMD64":
-        return "x86_64"
-    if machine == "aarch64" or machine == "arm64":
-        return "aarch64"
-    raise RuntimeError(f"Unrecognized machine architecture {machine}")
-
-
 # Fetch prebuilt Clang/LLVM libraries
 def fetch_prebuilt_libraries():
     if os.name == "nt":
@@ -36,12 +23,12 @@ def fetch_prebuilt_libraries():
         packman = "./tools/packman/packman"
         if sys.platform == "darwin":
             packages = {
-                "arm64": "15.0.7-darwin-aarch64-macos11",
+                "aarch64": "15.0.7-darwin-aarch64-macos11",
                 "x86_64": "15.0.7-darwin-x86_64-macos11",
             }
         else:
             packages = {
-                "arm64": "15.0.7-linux-aarch64-gcc7.5",
+                "aarch64": "15.0.7-linux-aarch64-gcc7.5",
                 "x86_64": "15.0.7-linux-x86_64-ptx-gcc7.5-cxx11abi0",
             }
 
@@ -98,7 +85,7 @@ def build_from_source_for_arch(args, arch, llvm_source):
     python_bin = "python/Scripts" if sys.platform == "win32" else "python/bin"
     os.environ["PATH"] = os.path.join(base_path, "_build/target-deps/" + python_bin) + os.pathsep + os.environ["PATH"]
 
-    if arch == "arm64":
+    if arch == "aarch64":
         target_backend = "AArch64"
     else:
         target_backend = "X86"
@@ -302,15 +289,12 @@ def build_from_source(args):
         llvm_source = llvm_project_path
 
     # build for the machine's architecture
-    if machine_architecture() == "x86_64":
-        build_from_source_for_arch(args, "x86_64", llvm_source)
-    else:
-        build_from_source_for_arch(args, "arm64", llvm_source)
+    build_from_source_for_arch(args, machine_architecture(), llvm_source)
 
     # for Apple systems also cross-compile for building a universal binary
     if sys.platform == "darwin":
         if machine_architecture() == "x86_64":
-            build_from_source_for_arch(args, "arm64", llvm_source)
+            build_from_source_for_arch(args, "aarch64", llvm_source)
         else:
             build_from_source_for_arch(args, "x86_64", llvm_source)
 
@@ -375,15 +359,12 @@ def build_warp_clang(args, lib_name):
     if sys.platform == "darwin":
         # create a universal binary by combining x86-64 and AArch64 builds
         build_warp_clang_for_arch(args, lib_name + "-x86_64", "x86_64")
-        build_warp_clang_for_arch(args, lib_name + "-arm64", "arm64")
+        build_warp_clang_for_arch(args, lib_name + "-aarch64", "aarch64")
 
         dylib_path = os.path.join(build_path, f"bin/{lib_name}")
-        run_cmd(f"lipo -create -output {dylib_path} {dylib_path}-x86_64 {dylib_path}-arm64")
+        run_cmd(f"lipo -create -output {dylib_path} {dylib_path}-x86_64 {dylib_path}-aarch64")
         os.remove(f"{dylib_path}-x86_64")
-        os.remove(f"{dylib_path}-arm64")
+        os.remove(f"{dylib_path}-aarch64")
 
     else:
-        if machine_architecture() == "x86_64":
-            build_warp_clang_for_arch(args, lib_name, "x86_64")
-        else:
-            build_warp_clang_for_arch(args, lib_name, "arm64")
+        build_warp_clang_for_arch(args, lib_name, machine_architecture())
diff --git a/warp/build_dll.py b/warp/build_dll.py
index 65d8c9df5..81bf9cb8a 100644
--- a/warp/build_dll.py
+++ b/warp/build_dll.py
@@ -8,11 +8,24 @@
 import sys
 import os
 import subprocess
+import platform
 
 import warp.config
 from warp.utils import ScopedTimer
 
 
+# return a canonical machine architecture string
+# - "x86_64" for x86-64, aka. AMD64, aka. x64
+# - "aarch64" for AArch64, aka. ARM64
+def machine_architecture() -> str:
+    machine = platform.machine()
+    if machine == "x86_64" or machine == "AMD64":
+        return "x86_64"
+    if machine == "aarch64" or machine == "arm64":
+        return "aarch64"
+    raise RuntimeError(f"Unrecognized machine architecture {machine}")
+
+
 def run_cmd(cmd, capture=False):
     if warp.config.verbose:
         print(cmd)
@@ -183,7 +196,7 @@ def build_dll_for_arch(dll_path, cpp_paths, cu_path, libs, mode, arch, verify_fp
                 "-gencode=arch=compute_80,code=sm_80",  # Ampere
                 "-gencode=arch=compute_86,code=sm_86",
             ]
-            if arch == "arm64" and sys.platform() == "linux":
+            if arch == "aarch64" and sys.platform() == "linux":
                 gencode_opts += [
                     # SASS for supported mobile architectures (e.g. Tegra/Jetson)
                     "-gencode=arch=compute_53,code=sm_53",  # X1
@@ -357,11 +370,13 @@ def build_dll(dll_path, cpp_paths, cu_path, libs=[], mode="release", verify_fp=F
     if sys.platform == "darwin":
         # create a universal binary by combining x86-64 and AArch64 builds
         build_dll_for_arch(dll_path + "-x86_64", cpp_paths, cu_path, libs, mode, "x86_64", verify_fp, fast_math, quick)
-        build_dll_for_arch(dll_path + "-arm64", cpp_paths, cu_path, libs, mode, "arm64", verify_fp, fast_math, quick)
+        build_dll_for_arch(
+            dll_path + "-aarch64", cpp_paths, cu_path, libs, mode, "aarch64", verify_fp, fast_math, quick
+        )
 
-        run_cmd(f"lipo -create -output {dll_path} {dll_path}-x86_64 {dll_path}-arm64")
+        run_cmd(f"lipo -create -output {dll_path} {dll_path}-x86_64 {dll_path}-aarch64")
         os.remove(f"{dll_path}-x86_64")
-        os.remove(f"{dll_path}-arm64")
+        os.remove(f"{dll_path}-aarch64")
 
     else:
         build_dll_for_arch(dll_path, cpp_paths, cu_path, libs, mode, "x86_64", verify_fp, fast_math, quick)

From cb1802bea6a10a91a2f3a79134260fbe8dc9513e Mon Sep 17 00:00:00 2001
From: Nicolas Capens <ncapens@nvidia.com>
Date: Fri, 5 Jan 2024 00:54:54 -0500
Subject: [PATCH 50/50] Fix enabling Jetson CUDA compute for AArch64

---
 warp/build_dll.py | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/warp/build_dll.py b/warp/build_dll.py
index 81bf9cb8a..2f311158c 100644
--- a/warp/build_dll.py
+++ b/warp/build_dll.py
@@ -196,7 +196,7 @@ def build_dll_for_arch(dll_path, cpp_paths, cu_path, libs, mode, arch, verify_fp
                 "-gencode=arch=compute_80,code=sm_80",  # Ampere
                 "-gencode=arch=compute_86,code=sm_86",
             ]
-            if arch == "aarch64" and sys.platform() == "linux":
+            if arch == "aarch64" and sys.platform == "linux":
                 gencode_opts += [
                     # SASS for supported mobile architectures (e.g. Tegra/Jetson)
                     "-gencode=arch=compute_53,code=sm_53",  # X1
@@ -379,4 +379,6 @@ def build_dll(dll_path, cpp_paths, cu_path, libs=[], mode="release", verify_fp=F
         os.remove(f"{dll_path}-aarch64")
 
     else:
-        build_dll_for_arch(dll_path, cpp_paths, cu_path, libs, mode, "x86_64", verify_fp, fast_math, quick)
+        build_dll_for_arch(
+            dll_path, cpp_paths, cu_path, libs, mode, machine_architecture(), verify_fp, fast_math, quick
+        )